Untitled

Engineering Practice with Oilfield and
Drilling Applications
Wiley-ASME Press Series
Fabrication of Metallic Pressure Vessels
Owen Greulich, Maan H. Jawad
Engineering Practice with Oilfield and Drilling Applications
Donald W. Dareing
Flow-Induced Vibration Handbook for Nuclear and Process Equipment
Michel J. Pettigrew, Colette E. Taylor, Nigel J. Fisher
Vibrations of Linear Piezostructures
Andrew J. Kurdila, Pablo A. Tarazaga
Bearing Dynamic Coefficients in Rotordynamics: Computation Methods and Practical Applications
Lukasz Brenkacz
Advanced Multifunctional Lightweight Aerostructures: Design, Development, and Implementation
Kamran Behdinan, Rasool Moradi-Dastjerdi
Vibration Assisted Machining: Theory, Modelling and Applications
Li-Rong Zheng, Dr. Wanqun Chen, Dehong Huo
Two-Phase Heat Transfer
Mirza Mohammed Shah
Computer Vision for Structural Dynamics and Health Monitoring
Dongming Feng, Maria Q Feng
Theory of Solid-Propellant Nonsteady Combustion
Vasily B. Novozhilov, Boris V. Novozhilov
Introduction to Plastics Engineering
Vijay K. Stokes
Fundamentals of Heat Engines: Reciprocating and Gas Turbine Internal Combustion Engines
Jamil Ghojel
Offshore Compliant Platforms: Analysis, Design, and Experimental Studies
Srinivasan Chandrasekaran, R. Nagavinothini
Computer Aided Design and Manufacturing
Zhuming Bi, Xiaoqin Wang
Pumps and Compressors
Marc Borremans
Corrosion and Materials in Hydrocarbon Production: A Compendium of Operational and Engineering Aspects
Bijan Kermani and Don Harrop
Design and Analysis of Centrifugal Compressors
Rene Van den Braembussche
Case Studies in Fluid Mechanics with Sensitivities to Governing Variables
M. Kemal Atesmen
The Monte Carlo Ray-Trace Method in Radiation Heat Transfer and Applied Optics
J. Robert Mahan
Dynamics of Particles and Rigid Bodies: A Self-Learning Approach
Mohammed F. Daqaq
Primer on Engineering Standards, Expanded Textbook Edition
Maan H. Jawad and Owen R. Greulich
Engineering Optimization: Applications, Methods and Analysis
R. Russell Rhinehart
Compact Heat Exchangers: Analysis, Design and Optimization using FEM and CFD Approach
C. Ranganayakulu and Kankanhalli N. Seetharamu
Robust Adaptive Control for Fractional-Order Systems with Disturbance and Saturation
Mou Chen, Shuyi Shao, and Peng Shi
Robot Manipulator Redundancy Resolution
Yunong Zhang and Long Jin
Stress in ASME Pressure Vessels, Boilers, and Nuclear Components
Maan H. Jawad
Combined Cooling, Heating, and Power Systems: Modeling, Optimization, and Operation
Yang Shi, Mingxi Liu, and Fang Fang
Applications of Mathematical Heat Transfer and Fluid Flow Models in Engineering and Medicine
Abram S. Dorfman
Bioprocessing Piping and Equipment Design: A Companion Guide for the ASME BPE Standard
William M. (Bill) Huitt
Nonlinear Regression Modeling for Engineering Applications: Modeling, Model Validation, and Enabling Design of Experiments
R. Russell Rhinehart
Geothermal Heat Pump and Heat Engine Systems: Theory and Practice
Andrew D. Chiasson
Fundamentals of Mechanical Vibrations
Liang-Wu Cai
Introduction to Dynamics and Control in Mechanical Engineering Systems
Cho W.S. To
Engineering Practice with Oilfield and Drilling
Applications
Donald W. Dareing
University of Tennessee
This Work is a co-publication between ASME Press

and John Wiley & Sons, Inc.
© 2022 ASME
This Work is a co-publication between ASME Press and John Wiley & Sons, Inc.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted,
in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as
permitted by law. Advice on how to obtain permission to reuse material from this title is available at
http://www.wiley.com/go/permissions.
The right of Donald W. Dareing to be identified as the authors of this work has been asserted in accordance with law.
Registered Office
John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
Editorial Office
111 River Street, Hoboken, NJ 07030, USA
For details of our global editorial offices, customer services, and more information about Wiley products visit us at
www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Some content that appears in
standard print versions of this book may not be available in other formats.
Limit of Liability/Disclaimer of Warranty

While the publisher and authors have used their best efforts in preparing this work, they make no representations or
warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all
warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose.
No warranty may be created or extended by sales representatives, written sales materials or promotional statements for
this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential
source of further information does not mean that the publisher and authors endorse the information or services the
organization, website, or product may provide or recommendations it may make. This work is sold with the
understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained
herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers
should be aware that websites listed in this work may have changed or disappeared between when this work was
written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other
commercial damages, including but not limited to special, incidental, consequential, or other damages.
Library of Congress Cataloging-in-Publication Data applied for:
ISBN: 9781119799498
Cover Design: Wiley

Cover Image: © Puneet Vikram Singh, Nature and Concept photographer/Getty Images
Set in 9.5/12.5pt STIXTwoText by Straive, Pondicherry, India
10 9 8 7 6 5 4 3 2 1
To Kristin:
My wonderful companion, whose energy, integrity, support, and reliability go beyond measure.
vii
Contents
Preface xxi
Nomenclature xxiii
Part I Engineering Design and Problem Solving 1
1 Design and Problem Solving Guidelines 3

Design Methodology 3
Market Analysis 5
Operational Requirements 5
Product Development 6
Government Procurement Procedure 6
Petroleum Industry Procedure 6
Design Specifications 7
Specification Topics 7
Performance Requirements 7
Sustainability 7
Codes and Standards 8
Environmental 8
Social Considerations 9
Reliability 9
Cost Considerations 10
Aesthetics 10
Product Life Cycle 10
Product Safety and Liability 11
Engineering Ethics 11
Creating Design Alternatives 12
Tools of Innovation 12
Patents 13
Reference Books and Trade Journals 13
Experts in a Related Field 13
Brainstorming 13
Existing Products and Concepts 13
viii Contents
Concurrent Engineering 14
Feasibility of Concept 14
Evaluating Design Alternatives 14
Evaluation Metrics 15
Scoring Alternative Concepts 15
Starting the Design 16
Design for Simplicity 16
Identify Subsystems 17
Development of Oil and Gas Reservoirs 17
Design of Offshore Drilling and Production Systems 18
Connection of Subsystems 19
Torsion Loading on Multibolt Patterns 19
Make-Up Force on Bolts 21
Preload in Drill Pipe Tool Joints 24
Shoulder Separation 26
Possible Yielding in the Pin 26
Make-Up Torque 28
Bolted Brackets 29
Welded Connections 30
Torsion Loading in Welded Connections 30
Attachments of Offshore Cranes 32
Quality Assurance 33
Engineering Education 34
Mission Statement 34
Academic Design Specifications 34
Design of the Academic Program 35
Outcomes Assessment 35
Saturn – Apollo Project 35
Notes 36
References 36
2 Configuring the Design 37

Force and Stress Analysis 37
Beam Analysis 39
Shear and Bending Moment Diagrams 40
Bending Stresses 45
Beam Deflection and Boundary Conditions 47
Shear Stress in Beams 48
Neutral Axis 50
Composite Cross Sections 52
Material Selection 54
Mechanical Properties of Steel 54
Use of Stress–Strain Relationship in a Simple Truss 57
Statically Indeterminate Member 59
Modes of Failure 62
Contents ix
Material Yielding 62
Stress Concentration 62
Wear 63
Fatigue 63
Stress Corrosion Cracking 69
Brittle Fracture 69
Fluid Flow Through Pipe 70
Continuity of Fluid Flow 70
Bernoulli’s Energy Equation (First Law) 71
Reynolds Number 71
Friction Head for Laminar Flow 72
Turbulent Flow Through Pipe 72
Senior Capstone Design Project 74
Pump Selection 74
Required Nozzle Velocity 74
Nozzle Pressure 74
Pump Flow Rate Requirement 75
Vibration Considerations 77
Natural Frequency of SDOF Systems 80
Location of Center of Gravity 84
Moment of Inertia with Respect to Point A 84
Springs in Series, Parallel 85
Deflection of Coiled Springs 86
Free Vibration with Damping 86
Quantifying Damping 87
Critical Damping in Vibrating Bar System 88
Forced Vibration of SDOF Systems with Damping 89
Nonlinear Damping 93
Vibration Control 93
Other Vibration Considerations 94
Transmissibility 94
Vibration Isolation 95
Commonality of Responses 96
Application of Vibration Absorbers in Drill Collars 96
Natural Frequencies with Vibration Absorbers 97
Responses to Nonperiodic Forces 100
Dynamic Load Factor 102
Packaging 103
Vibrations Caused by Rotor Imbalance 105
Response to an Imbalanced Rotating Mass 105
Synchronous Whirl of an Imbalanced Rotating Disk 106
Balancing a Single Disk 109
Synchronous Whirl of Rotating Pipe 109
Stability of Rotating Pipe under Axial Load 110
Balancing Rotating Masses in Two Planes 112
x Contents
Refining the Design 113

Manufacturing 113
Manufacturing Drawings 114
Dimensioning 114
Tolerances 115
Three Types of Fits 116
Surface Finishes 117
Nanosurface Undulations 118
Machining Tools 119
Lathes 119
Drill Press 119
Milling Machines 120
Machining Centers 120
Turning Centers 120
References 121
Part II Power Generation, Transmission, Consumption 123
3 Power Generation 125

Water Wheels 125
Fluid Mechanics of Water Wheels 125
Steam Engines 127
Steam Locomotives 128
Power Units in Isolated Locations 130
Regional Power Stations 131
Physical Properties of Steam 131
Energy Extraction from Steam 132
First Law of Thermodynamics – Enthalpy 132
Entropy – Second Law 132
Thermodynamics of Heat Engines 133
Steam Turbines 135
Electric Motors 136
Internal Combustion Engines 137
Four Stroke Engine 137
Two Stroke Engines 138
Diesel Engines 139
Gas Turbine Engines 139
Impulse/Momentum 141
Energy Considerations 142
Engine Configurations 142
Rocket Engines 144
Rocketdyne F-1 Engine 144
Atlas Booster Engine 144
Gas Dynamics Within Rocket Engines 145
Contents xi
Rocket Dynamics 146

Energy Consumption in US 147
Solar Energy 148
Hydrogen as a Fuel 149
Hydroelectric Power 149
Wind Turbines 149
Geothermal Energy 149
Atomic Energy 150
Biofuels 150
Notes 150
References 150
4 Power Transmission 151

Gear Train Transmission 153
Water Wheel Transmission 153
Fundamental Gear Tooth Law 154
Involute Gear Features 154
Gear Tooth Size – Spur Gears 156
Simple Gear Train 158
Kinematics 158
Worm Gear Train 159
Planetary Gear Trains 160
Compound Gear Trains 161
Pulley Drives 162
Rope and Friction Pulleys 162
Belted Connections Between Pulley Drives 164
Fundamentals of Shaft Design 166
Shear Stress 167
Stress Analysis of Shafts 170
Twisting in Shafts Having Multiple Gears 171
Keyway Design 172
Mechanical Linkages 173
Relative Motion Between Two Points 173
Absolute Motion Within a Rotating Reference Frame 175
Scotch Yoke 177
Slider Crank Mechanism 178
Velocity Analysis 179
Acceleration Analysis 180
Four-Bar Linkage 181
Three Bar Linkage 184
Velocity Equation 185
Acceleration Equation 185
xii Contents

Geneva Mechanism 188
Flat Gear Tooth and Mating Profile 189
Cam Drives 191
Cam Drives – Linear Follower 191
Acceleration Polygon 193
Cam with Linear Follower, Roller Contact 194
Velocity Analysis – Rotating Reference Frame 195
Acceleration Analysis – Rotating Reference Frame 195
Velocity Analysis – Ritterhaus Model 196
Acceleration Analysis – Ritterhaus Model 196
Cam with Pivoted Follower 196
Power Screw 198
Hydraulic Transmission of Power 199
Kinematics of the Moineau Pump/Motor 202
Mechanics of Positive Displacement Motors 203
References 208
5 Friction, Bearings, and Lubrication 209

Rolling Contact Bearings 209
Rated Load of Rolling Contact Bearings 210
Effect of Vibrations on the Life of Rolling Contact Bearings 213
Effect of Environment on Rolling Contact Bearing Life 216
Effect of Vibration and Environment on Bearing Life 217
Hydrostatic Thrust Bearings 220
Flow Between Parallel Plates 220
Fluid Mechanics of Hydrostatic Bearings 222
Optimizing Hydrostatic Thrust Bearings 224
Pumping Requirements 224
Friction Losses Due to Rotation 225
Total Energy Consumed 226
Coefficient of Friction 227
Squeeze Film Bearings 228
Pressure Distribution Under a Flat Disc 228
Comparison of Pressure Profiles 230
Spring Constant of Hydrostatic Films 231
Damping Coefficient of Squeeze Films 231
Other Shapes of Squeeze Films 233
Squeeze Film with Recess 233
Squeeze Film Under a Washer 234
Spherical Squeeze Film 235
Nonsymmetrical Boundaries 236
Application to Wrist Pins 237
Contents xiii
Thick Film Slider Bearings 240

Slider Bearings with Fixed Shoe 240
Load-Carrying Capacity 243
Friction in Slider Bearings 243
Center of Pressure 244
Slider Bearing with Pivoted Shoe 245
Frictional Resistance 246
Exponential Slider-Bearing Profiles 247
Pressure Distribution for Exponential Profile 247
Pressure Comparison with Straight Taper Profile 248
Pressure Distribution for Open Entry 249
Exponential Slider Bearing with Side Leakage 250
Hydrodynamic Lubricated Journal Bearings 254
Pressure Distribution Around an Idealized Journal Bearing 254
Minimum Film Thickness in Journal Bearings 258
Friction in an Idealized Journal Bearing 259
Petroff’s Law 259
Sommerfeld’s Solution 260
Stribeck Diagram and Boundary Lubrication 261
Regions of Friction 261
Comparison of Journal Bearing Performance with Roller Bearings 263
Journal Bearing 263
Roller Contact Bearing (See Footnote 1) 263
Ball Bearing (See Footnote 1) 264
Note 264
References 264
6 Energy Consumption 267

Subsystems of Drilling Rigs 267
Draw Works in Drilling Rigs 269
Block and Tackle Hoisting Mechanism 270
Spring Constant of Draw Works Cables 270
Band Brakes Used to Control Rate of Decent 270
Rotary Drive and Drillstring Subsystem 272
Kelly and Rotary Table Drive 272
Friction in Directional Wells 272
Top Drive 273
Drillstring Design and Operation 275
Buoyancy 276
Hook Load 277
Definition of Neutral Point 277
Basic Drillstring: Drill Pipe and Drill Collars 279
xiv Contents
Physical Properties of Drill Pipe 279

Selecting Drill Pipe Size and Grade 281
Select Pipe Grade for a Given Pipe Size 281
Determine Maximum Depth for Given Pipe Size and Grade 282
Roller Cone Rock Bits 283
Polycrystalline Diamond Compact (PDC) Drill Bits 283
Natural Diamond Drill Bits 284
Hydraulics of Rotary Drilling 285
Optimized Hydraulic Horsepower 285
Field Application 288
Controlling Formation Fluids 290
Hydrostatic Drilling Mud Pressure 290
Annular Blowout Preventer 290
Hydraulic Rams 292
Casing Design 293
Collapse Pressure Loading (Production Casing) 295
Burst Pressure Loading (Production Casing) 295
API Collapse Pressure Guidelines 297
Plastic Yielding and Collapse with Tension 297
Summary of Pressure Loading (Production Casing) 298
Effect of Tension on Casing Collapse 298
Tension Forces in Casing 300
Design of 95 8 in. Production Casing 302
Design Without Factors of Safety 302
Directional Drilling 306
Downhole Drilling Motors 306
Rotary Steerable Tools 307
Stabilized Bottom-Hole Assemblies 308
Power Units at the Rig Site 310
References 310
Part III Analytical Tools of Design 313
7 Dynamics of Particles and Rigid Bodies 315

Statics – Bodies in Equilibrium 315
Force Systems 316
Freebody Diagrams 318
Force Analysis of Trusses 318
Method of Joints 319
Method of Sections 319
Kinematics of Particles 320
Linear Motion 320
Rectangular Coordinates 321
Polar Coordinates 322
Contents xv
Velocity Vector 325

Acceleration Vector 325
Curvilinear Coordinates 325
Navigating in Geospace 328
Tracking Progress Along a Well Path 328
Minimum Curvature Method 329
Dogleg Severity 331
Projecting Ahead 332
Kinematics of Rigid Bodies 333
Rigid Body Translation and Rotation 333
General Plane Motion 334
Dynamics of Particles 335
Units of Measure 335
Application of Newton’s Second Law 336
Static Analysis 336
Dynamic Analysis 337
Work and Kinetic Energy 337
Potential Energy 339
Drill Bit Nozzle Selection 341
Impulse–Momentum 342
Impulse–Momentum Applied to a System of Particles 343
Mechanics of Hydraulic Turbines 345
Performance Relationships 349
Maximum Output of Drilling Turbines 350
Dynamics of Rigid Bodies 351
Rigid Bodies in Plane Motion 352
Translation of Rigid Bodies 354
Rotation About a Fixed Point 354
Center of Gravity of Connecting Rod 355
Mass Moment of Inertia of Connecting Rod 356
General Motion of Rigid Bodies 356
Dynamic Forces Between Rotor and Stator 359
Interconnecting Bodies 361
Gear Train Start-Up Torque 361
Kinetic Energy of Rigid Bodies 363
The Catapult 364
Impulse–Momentum of Rigid Bodies 364
Linear Impulse and Momentum 365
Angular Impulse and Momentum 365
Angular Impulse Caused by Stabilizers and PDC Drill Bits 368
Accounting for Torsional Flexibility in Drill Collars 369
Interconnecting Bodies 370
Conservation of Angular Momentum 371
References 374
xvi Contents
8 Mechanics of Materials 375

Stress Transformation 376
Theory of Stress 377
Normal and Shear Stress Transformations 377
Maximum Normal and Maximum Shear Stresses 378
Mohr’s Stress Circle 381
Theory of Strain 383
Strain Transformation 384
Mohr’s Strain Circle 386
Principal Axes of Stress and Strain 386
Generalized Hooke’s Law 388
Theory of Plain Stress 388
Orientation of Principal Stress and Strain 389
Theory of Plain Strain 391
Pressure Vessel Strain Measurements 391
Analytical Predictions of Stress and Strain 391
Strain in the Spherical Cap 393
Conversion of Strain Measurements to Principal Strains and Stresses 393
Beam Deflections 396
Cantilever Beam with Concentrated Force 397
Cantilevered Beam with Uniform Load 398
Simply Supported Beam with Distributed Load 399
Statically Indeterminate Beams 400
Multispanned Beam Columns 402
Large Angle Bending in Terms of Polar Coordinates 403
Bending Stresses in Drill Pipe Between Tool Joints 405
Application to Pipe Bending in Curved Well Bores 408
Multispanned Beam in Terms or Polar Coordinates 410
Pulling Out of the Well Bore 410
Columns and Compression Members 411
Column Buckling Under Uniform Compression 411
Columns of Variable Cross Section 415
Tubular Buckling Due to Internal Pressure 416
Effective Tension in Pipe 417
Buckling of Drill Collars 418
Combined Effects of Axial Force and Internal/External Pressure 420
Buckling of Drill Pipe 420
Bending Equation for Marine Risers 424
Unique Features of the Differential Equation of Bending 424
Effective Tension 426
Buckling of Marine Risers 426
Tapered Flex Joints 429
Equation of Bending 430
Parabolic Approximation to Moment of Inertia 430
Solution to Differential Equation 432
Contents xvii
Application to Marine Risers 435

Torsional Buckling of Long Vertical Pipe 435
Boundary Conditions 436
Both Top and Bottom Ends Pinned 438
Simply Supported at Both Ends with no End Thrust 440
Force Applied to Lower End 441
Effect of Drilling Fluid on Torsional Buckling 442
Lower Boundary Condition Fixed 442
Operational Significance 442
Pressure Vessels 443
Stresses in Thick Wall Cylinders 443
Stresses in Thin-Wall Cylinders 444
Stresses Along a Helical Seam 444
Interference Fit Between Cylinders 445
Thin-Wall Cylinders 445
Surface Deflections of Thick-Wall Cylinders 447
Thick Wall Cylinder Enclosed by Thin Wall Cylinder 448
Thick Wall Cylinder Enclosed by Thick Wall Cylinder 448
Elastic Buckling of Thin Wall Pipe 449
Bresse’s Formulation 450
Application to Long Cylinders 451
Thin Shells of Revolution 452
Curved Beams 455
Location of Neutral Axis 455
Stress Distribution in Cross Section 456
Shear Centers 460
Unsymmetrical Bending 464
Principal Axis of Inertia 464
Neutral Axis of Bending 468
Bending Stresses 470
Beams on Elastic Foundations 471
Formulating the Problem 472
Mathematical Solution 473
Solution to Concentrated Force 474
Radial Deflection of Thin Wall Cylinders Due to Ring Loading 475
Formulation of Spring Constant 476
Equation of Bending for Cylindrical Arc Strip 477
Reach of Bending Moment 480
Bending Stress in Wall of a Multi Banded Cylinder 480
Criteria of Failure 482
Combined Stresses 482
Internal Pressure 483
xviii Contents
Applied Torque 483

Bending Moment 483
Failure of Ductile Materials 484
Visualization of Stress at a Point 485
Pressure Required to Yield a Cylindrical Vessel 486
Failure of Brittle Materials 487
Mode of Failure in Third Quadrant 489
References 489
9 Modal Analysis of Mechanical Vibrations 491

Complex Variable Approach 491
Complex Transfer Function 493
Interpretation of Experimental Data 493
Natural Frequency 494
Damping Factor 494
Spring Constant 495
Mass 495
Damping Coefficient 495
Two Degrees of Freedom 495
Natural Frequencies and Modes of Vibration 495
SDOF Converted to 2-DOF 497
Single Degree of Freedom 497
Other 2-DOF Systems 499
Undamped Forced Vibrations (2 DOF) 500
Undamped Dynamic Vibration Absorber 502
Base and Absorber Pinned Together 503
Multi-DOF Systems – Eigenvalues and Mode Shapes 507
Flexibility Matrix – Stiffness Matrix 508
Direct Determination of the Stiffness Matrix 511
Direct Determination of the Mass Matrix 512
Amplitude and Characteristic Equations 512
Parameters Not Chosen at Discrete Masses 514
Lateral Stiffness of a Vertical Cable 515
Building the Damping Matrix 516
Modal Analysis of Discrete Systems 516
Orthogonal Properties of Natural Modes 517
Proportional Damping 518
Transforming Modal Solution to Local Coordinates 519
Free Vibration of Multiple DOF Systems 520
Free Vibration of 2 DOF Systems 521
Suddenly Stopping Drill Pipe with the Slips 522
Critical Damping of Vibration Modes 524
Forced Vibration by Harmonic Excitation 526
Contents xix

Harmonic Excitation of 3 DOF Systems 527
Modal Solution of a Damped 2-DOF System 529
General Complex Variable Solution 530
Experimental Modal Analysis 532
Modal Response to Nonperiodic Forces 535
Natural Frequencies of Drillstrings 536
References 538
10 Fluid Mechanics 541

Laminar Flow 541
Viscous Pumps 541
Force to Move Runner 543
Capillary Tubes 544
Flow Through Noncircular Conduits 545
Elliptical Conduit 545
Rectangular Conduit 546
Unsteady Flow Through Pipe 547
Hydraulics of Non-Newtonian Fluids 551
Hydraulics of Drilling Fluids 551
Pressure Loss Inside Drill Pipe 551
Pressure Loss in Annulus 552
Oil Well Drilling Pumps 552
Drilling Hydraulics 554
Power Demands of Downhole Motors 556
Performance of Positive Displacement Motors (PDM) 557
Application of Drilling Turbines 560
Hydraulic Demands of Drilling Motors – Turbines 561
Fluid Flow Around Vibrating Micro Cantilevers 562
Mathematical Model 563
Fluid Pressure Formulation 564
Fluid Velocity Formulation 565
References 566
11 Energy Methods 569

Principle of Minimum Potential Energy 569
Stable and Unstable Equilibrium 569
Stability of Floating Objects 570
Stability of a Vertical Rod 572
Rayleigh’s Method 573
Multiple Degrees of Freedom 574
Structure Having Two Degrees of Freedom 574
Analysis of Beam Deflection by Fourier Series 576
Concentrated Load 577
Distributed Load 577
Axially Loaded Beam (Column) 578
xx Contents
Principle of Complementary Energy 579

Engineering Application 580
Castigliano’s Theorem 582
Chemically Induced Deflections 588
Microcantilever Sensors 588
Simulation Model 588
Molecular and Elastic Potential Energies 591
References 592
Index 593
xxi
Preface
Engineers are trained to understand the fundamental principles of mechanics and mathematics.
These tools provide a background of knowledge for making professional decisions. The tools of
engineering science apply across most engineering disciplines. The key to their application is visua-
lizing a reasonable mathematical model for the problem at hand. Freebody diagrams are helpful in
this regard. Mathematical solutions follow, leading to reasonable engineering results. Typically,
there is only one answer, so each problem is closed-ended.
On the other hand, design and problem-solving are open-ended. There are many possible solu-
tions and alternatives must be created. While each engineering design is different, the approach is
the same. An objective of this book is to explain the engineering design process and show how to
apply basic engineering tools.
The book contains three parts.
Part I Engineering Design and Problem-Solving

Part II Power Generation, Transmission, Consumption
Part III Analytical Tools of Design
Part I gives a systematic process for developing an engineering design. The application of engi-
neering tools is illustrated during the conceptual and preliminary activities of design. Concept eval-
uation and selection are explained. Visualizing a total device or any system in terms of its
subsystems is helpful in creating a design. Key considerations in finalizing a design are implement-
ing feedback from test results or other evaluation sources, finalizing a design and presentation of
final manufacturing drawings.
Every machine has (i) a prime mover or power source, (ii) mechanisms to transmit energy and
(iii) energy consumed by forming the final product, plus friction. Part II covers Power Generation,
Transmission, and Consumption.
Part III contains useful tools of engineering mechanics. Each selected topic goes beyond the tra-
ditional tools of design. Mathematical modeling and methods of solution are of historical signifi-
cance. Each topic is supplemented with key references for additional background information.
Physical responses of engineering systems are predictable through science and mathematics. This
one thing makes it possible to design modern structures and machinery to a high degree of relia-
bility. The first scientifically based engineered bridge is the Eads Bridge which spans the Mississippi
River at St. Louis. It was designed and constructed by James Eads. Construction began in 1867. It
was dedicated in 1874 and is still in use today.
xxii Preface
My goal in writing this book was to document the essence of engineering practice. The manu-
script is a condensation of lecture notes developed over years of teaching across the mechanical
engineering curriculum and industrial practice in the petroleum industry. It is written for under-
graduate and graduate students and as a reference for practicing engineers.
Donald W. Dareing
Professor Emeritus, Mechanical Engineering
University of Tennessee, Knoxville
Life Fellow Member, ASME
Knoxville, TN, USA
April 2021
xxiii
Nomenclature
a acceleration
BF buoyancy factor
c distance to outside beam surface, damping coefficient
ccr critical damping coefficient
E modulus of elasticity
Em energy per pound
F applied force, axial internal force at drill pipe-collar
interface above hydrostatic
FS safety factor
f friction force, vibration frequency
FB axial force in pipe (lower end)
Fcr critical buckling force
fn natural frequency, cps
G modulus of rigidity, angular momentum
h lubrication film thickness, enthalpy
hf friction head
H linear momentum, elevation
I area moment of inertia
Im impulse
J angular moment of inertia of a cross section,
angular mass moment of inertia
k, K local (modal) mechanical spring constant
K0 stress intensity factor
L length
m, M local (modal) mass, bending moment
N force
NR Reynolds number
p pressure
P unit force (force per area), power, diametral pitch
of gears
Q moment of area above shear surface, heat,
compressive force
Qeff = Q + (piAi − poAo) plus sign means compression
q roller bearing exponent
r radial position, frequency ratio (ω/ωn)
xxiv Nomenclature
S section modulus of a cross sectional area, Sommerfeld

number, entropy
t, T time, torque, period of oscillation
T eff = F B + wx + L − x A0 γ 0 − Ai γ i marine riser (x measured up from bottom)
T eff = F B + wx + L − x wm drill pipe (x measured up from bottom)
TR transmissibility
U, V principal axis of inertia of a cross section, V also indi-
cated shear force
V velocity, also total potential energy
w, W distributed load on a beam, weight of a discrete body
x, y, z reference frame
X, Y, Z reference frame
[X] modal matrix
x(t) local response
Z viscosity (cp)
Greek Symbols
δ displacement, log decrement

μ viscosity, coefficient of friction
ω rotational speed, circular frequency
ωn natural circular frequency
θ angular position
σ normal stress
τ shear stress
ɛ normal strain
γ shear strain
ζ damping factor
ν Poisson’s ratio
η(t) modal response
σa allowable design stress
σ yld yield strength
x
ζ=
L
F B + LA0 γ 0 − LAi γ i L2
β=
EI
w − A0 γ 0 + Ai γ i L3
α=
EI
TL
Θ=
EI
1
Part I
Engineering Design and Problem Solving
Engineering design is a logical sequence of activities that solves a problem or achieves a specified
objective. Every design project has a beginning and an end. They can be several years long, such as
putting a man on the moon and returning safely to earth, or it can be short, such as designing and
fabricating a water pump. Successful engineering designs require a clear objective – well thought
out and executed. Planning is critical. The design process may also be applied to management or
any problem situation.
In conducting design, it is important to understand the difference between “open-end problems”
and “closed-ended problems.” Engineering tools of design are usually closed ended and based on
fundamental laws of engineering science. The answer is unique. Use of engineering tools usually
follows certain steps:
1) Develop a mathematical model for the physical element under consideration.
2) Develop a freebody diagram of the element along with forces and moment considering the con-
straints placed on the element.
3) Solve the mathematical equations leading to a prediction of performance, usually expressed in
terms of stress, deflection, vibration, etc.
4) Judge the answer against experience, order of magnitude (believable), and uniformity of
dimensions.
On the other hand, open-ended problems have many possible solutions. Each must be generated
and evaluated before a design can start. Solving open-ended problems requires imagination and
creativity. Part I gives a process for solving open-ended problems, including steps in project work.
It also underscores important design principles that may be considered in moving through an engi-
neering project.
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
3
Design and Problem Solving Guidelines
Engineering design involves management of people, resources, money, and time. Success depends
on planning, resource, and time management. Time is usually the driver.
When discussing the importance of teaming with one company, the response was, “teaming isn’t
important – it is everything.” The very success of a company depends upon people skills and the
ability to work with others as a team member. Pete Carroll, while head football coach at the
University of Southern California, says, “Winning players don’t always win. It’s the winning plays
that win.”
Planning is a matter of thinking through the activities and tasks that will be necessary to achieve
the stated goal. This is somewhat experience dependent. For large projects, it may be useful to
divide tasks into major activities, such as design, fabrication, installation, and commission, which
are usually conducted in tandem. In other projects, where several activities are conducted simul-
taneously, major groupings may be needed. An example would be a military operation involving
various branches.
General Dwight Eisenhower, along with his staff, spent months developing a plan for the inva-
sion of Europe. His team of officers generated and evaluated various plans of attack. Eisenhower
once said “… the plan itself is not as important as the act of planning.” Thinking through the plan is
the key.
Plans need to be flexible. As new information is gathered along the way, the plan may need to be
modified. A good manager anticipates problems and deals with them early to avoid crises. A crisis is
a situation where a critical problem needs to be solved, but there is little time to solve it. A Gantt
chart can be useful in this regard.
Design Methodology
Basic steps for developing a product idea (or service) into a profitable venture are given in
Figure 1.1. The first few boxes indicate the importance of a preliminary market analysis and input
from customers to determine market reaction to a new product. Also, a preliminary market analysis
helps define and refine the attributes of the product. Initial feedback from customers is useful in
deciding whether to proceed with further development.
Design specifications are based on specific needs and expected performance. Design specifica-
tions represent the initial engineering baseline for generating design alternatives. In most cases,
design specification are legal statements of what is expected. They must be established accurately
and in concert with users of the future product.
4 Design and Problem Solving Guidelines
Operation requirements
Design specifications
Synthesis
Analysis
Concept A Concept B Concept C
Evaluate concepts
Preliminary
design
Build prototype
Test prototype
Fabrication
drawings
Manufacture
Figure 1.1 Design development process.
Design alternatives are typically generated by a team of professionals with special skills, such as
marketing, design, and manufacturing. This activity is sometimes called concurrent engineering
where the team considers every aspect of the product from technical feasibility to product life cycle
to manufacturing and marketing strategy. Feedback from potential customers is important. The
team also evaluates each design alternative and selects the best concept to advance. Depending
on the complexity of the concept, technical feasibility studies, requiring advanced computational
techniques, may be required during the refinements of design alternatives.
Since design is open ended, there are many possible solutions or design alternatives that satisfy a
given set of specifications. Once viable designs have been generated, they need to be ranked so
choice can be made. Choosing a preferred concept is based on trade-offs among evaluation metrics
identified for a given product; an evaluation method will be described later.
A preliminary design represents an update of the engineering baseline. The preliminary design
refines the preferred alternative. It advances the engineering baseline for the final design and fab-
rication phases.
Developing a final design may require the use of computer-aided-design (CAD), numerical anal-
ysis, and other analytical tools to refine dimensions. Prototype testing may also be desirable. Com-
puter simulations may alleviate the high cost of prototype testing.
The product configuration is again evaluated in the marketplace for customer feedback and
approval. This is accomplished through market surveys or market focus groups depending on
the nature of the product.
The next step is to interface CAD codes with manufacturing (CAM). This requires converting
design codes into machine tool codes. Depending on the product and the market, the ability to
reconfigure the machining and handling process in a timely manner may be important for “just
in time” delivery.
Market Analysis
The purpose of a market analysis is to identify what potential customers want in a new product, estab-
lish the size of the market, and determine what price the market is willing to pay for the product.
A market analysis will produce a set of product attributes, which more clearly define the main features
of the planned product. Using customer input and competitor product features, important features for
the new product can be identified and ranked as to their importance. This information identifies cus-
tomer preferences and competitive differentiation during the conceptual stage of product development.
New products can be either research driven, or market driven. Research-driven products stem
from ideas that spawn from basic or fundamental research. In this case, a new technique or device
may be the objective or a by-product of the study. The technique or device then becomes a solution
looking for a problem, so to speak. The market-driven product is developed in response to a definite
market need. In some cases, a market may be developed for a new idea.
Before investing much time and money, it is best to conduct a patent search to make sure the
product does not infringe on active patents. This exercise will also give useful information on
the state-of-the-art of products as applied to a given market. It may show the patent protection
period on a product has expired, offering the opportunity to enter the market with a competitor’s
product – with improvements.
In recent years, markets have become more demanding on product delivery. Customer needs
may change over a short period. Companies that can retool for “just in time manufacturing” in
response to this demand have an advantage. One tool company, that makes diamond drill bits
for oil and gas well drilling, built its business on making diamond drill bits overnight; each diamond
was handset. Each diamond bit was and still is tailored to suit a set of design specifications stipu-
lated by an oil company. The main reason for a quick response capability (or “just in time manu-
facturing”) is moderate demand for high cost of diamond drill bits. It is not good business to
stockpile high-cost products for a limited market application. Warehoused products may become
outdated. It is costly and risky.
Operational Requirements
Operational requirements or product attributes describe the expected functional performance of a

new product. Product description may come out of a business plan for a new product concept, a
government need for a new weapon, or an oil company’s need to develop an oil field in a given
geographic location.
Product Development
Top management may define the operational requirements for a product, based on a market anal-
ysis. Company engineers then develop a set of design specifications before proceeding. Product
design may be conducted within a company or contracted outside.
Government Procurement Procedure

The federal government has very strict guidelines for procuring products and services [1]. Govern-
ment procurement is normally through the General Service Administration (GSA). The need for a
product may come to the GSA from any government agency, which in turn coordinates the Federal
Product Description (FPD). An FPD describes the operational requirements and required functions.
For example, assume that the Joint Chief of Staff decides that the military needs a new type of
aircraft. They would make their request to the GSA and describe the aircraft in terms of expected
operational requirement, such as:
•• Range
Speed
•• Landing capabilities
Weapon weight
• Weather considerations.
The GSA would expand the requesting agency’s description of operational requirements.
Effective market research and analysis must be conducted to assure that user need is satis-
fied. During the market research and analysis phase, the preparing activity should advise
potential agency users that a FPD is being developed. Ask potential users to provide a state-
ment of their needs in essential functional or performance terms to the maximum practical
extent.
In addition to a clear description of operational requirements, FPDs will also develop a compre-
hensive list of design specifications for a new product. The tender document would be reviewed at
various levels with GSA before it is released to contract bidders.
Petroleum Industry Procedure

An economic analysis is conducted on each new oil reservoir to determine its profitability and the
best way to develop it. Following this, operational requirements are set before proceeding. Oper-
ational requirements may include such factors as
•• Ocean water depth

Size of reservoir
•• Oil, gas, water content of reservoir

Reservoir pressure
• Production rate (barrels per day).
Design specifications would document detailed engineering constraints on the design, such as
environmental conditions, ocean floor mud line load-bearing capabilities, material specifications,
expected loads, 100-year storm conditions, etc.
Considerable time is spent in gathering this information to establish operational requirements

and design specifications. Company engineers build a set of design specifications to form a “tender
document” for contractor bid preparation.
There are two contract approaches: turnkey and cost plus. Turnkey simply means that the con-
tractor will deliver a product at a fixed price. The contractor is responsible for every detail, including
identifying and satisfying all codes and standards relevant to the design. Since the price is fixed, oil
companies would be concerned about delivery dates. Missing the planned delivery date could
greatly increase future monetary returns and profit.
A cost plus is based on an agreed hourly rate. The equipment and supplies are additional costs to
the buyer. Usually there is a percentage tacked on to these costs. Company representatives are
directly involved in day-to-day decisions.
Design Specifications
Design specifications are an itemized set of constraints placed on a design. They identify product
performance expectations: what the product is supposed to do and how the product should perform.
They are contractual and represent the initial “engineering baseline” from which all concepts are
generated. They are an important part of a contract between customer and designer. Usually, the
customer signs-off on a set of specifications once they have been documented. Any changes, for any
reason, after the development work starts, will cause delays, and increase costs. The cost of making
changes is usually written into a contract.
Once operational requirements have been set, design specifications are documented. They may
be expanded by outside contractors in conjunction with company engineers. The contract usually
puts the burden of completeness on the contractor, such as all relevant Codes and Standards are the
responsibility of the design contractor.
Specification Topics
Design specifications are usually subdivided into key topics. Topics normally considered are dis-
cussed below.
Performance Requirements
Performance requirements identify specifics, such as loads, motions, flow rates, operating
pressures, and temperature limits, to name a few. In addition, the technical specifications
may include physical and chemical properties of materials to be used. Material properties
may include such items as yield strength and hardness. Weld procedures (including preheating)
and welder qualification requirements, special heat treatment and annealing may be specified.
Environmental and climate conditions may affect design. Examples are wind, oceanographic
conditions, such as wave height, wind-driven current velocities, and tidal currents.
Performance requirements define the physical constraints in the design. Depending upon the size
of the project, the specification document can be as small as a few pages or several volumes.
Sustainability
Sustainability means being good stewards of the resources on planet earth. A 1987 UN report
defines sustainable development as: “Meeting the needs of the present without compromising
the ability of future generations to meet their own needs.” This specification is relevant, ethical,
and makes good economic sense. Earlier business strategies were based on planned obsolescence,
where products were intentionally designed to wear out after a given period [2].
A few metrics relating to sustainability are:
a) design efficiency in terms of materials, weight, cost, energy consumption

b) quantity and type of waste from fabrication
c) minimal friction, wear, maintenance, reliability
d) use of renewable energy sources, such as wind and solar
e) environmental impact of a possible failure
f) environmental recovery plan in case of an unexpected failure/disaster
g) design for modular replacement instead of product throw-away
h) disposal at end-of-life cycle (computers, TVs)
i) automation verses human control (cause of many disasters)
j) redundancy in monitoring system performance.
Codes and Standards
During the midst of the industrial revolution of the nineteenth century, it became apparent that
mechanical components needed to be standardized to assure, for example, bolts made in one place
could fit together with nuts made in another place. The American Society of Mechanical Engineers
(ASME) took a leading role in standardizing mechanical components, such as pulleys, gears, and
key seats. Even shop drawing symbols needed to be standardized.
During the early days of the steam engine, it was common for steam boilers to explode causing
personal injury and death. In 1884, the ASME established a uniform test standard for boilers; this
was ASME’s first standard. This standard was followed by a boiler construction standard, which
was published in 1915. Such standards became the foundation of ASME’s current Boiler and Pres-
sure Vessel Code. Since the development of this Code, boiler disasters have been essentially
eliminated.
What is a Code? A code is a standard that has been mandated by one or more government
bodies. A code has the force of law behind it. When a standard is specified in a business contract,
it becomes a code, as well. Therefore, the words, codes, and standards, are sometimes used
interchangeably.
What is a Standard? Standards are a set of technical definitions and guidelines or a set of instruc-
tions for designers and manufacturers. Their use is strictly voluntary, and they do not have the force
of law. Standards serve as a vehicle of communications, defining quality, and establishing safety
criteria for producers and users.
Many professional organizations develop Standards; however, they must follow procedures
accredited by the American National Standards Institute (ANSI). These procedures must reflect
openness, transparency, balance of interest, and due process.
Many turnkey contracts make the contractor or design company responsible for applying all rel-
evant codes and standards to a design. While specific codes and standards are not listed in a tender
document, they are implied through legal contracts.
Environmental
Designers should consider the environmental impact of a new product throughout the product’s life
cycle, i.e. from fabrication to product disposal. Even in the early stages of offshore platform design,
disposal of a structure at the end of production life (about 20–30 years) is considered.
The most vulnerable or highest risk components in the design should be identified and environ-
mental consequences associated with possible failure considered. The designer needs to consider
the “what if” scenarios. If a failure occurs, how would it affect public health, public safety, public
image of company, property damage, as well as damage to the environment. Goodwill is an asset on
the balance sheet of any company.
Social Considerations
People, who will be affected by the implementation or use of a given product or project, should be
consulted and brought into design deliberations as early as possible – the sooner the better. This is
not only right but, by doing so, misunderstanding is alleviated, and public resistance will be reduced
as the project develops. People simply want to be consulted and given the opportunity to make
input on issues that affect their lives. If not given this opportunity, citizens may unite and work
against a worthwhile project.
Aside from human reaction, there are legitimate reasons for considering social factors in design,
such as safety and preservation of a culture. When oil companies began to develop oil fields in the
northern part of the North Sea, oil transportation from the offshore platforms had to be resolved.
Crude oil could be off-loaded directly onto tankers or transported to land by a pipeline and then
loaded onto tankers in a protected harbor for transport to refineries. The closest land point from
the platforms was to the small fishing village of Solom Voe, Shetland Islands. An extensive study
was conducted to design a port that would not destroy centuries-old lifestyle of the people living in
this area. There were also benefits to local economies.
Reliability
There are two methods of design which relate to the safety and reliability of products: (i) factor of
safety and (ii) statistical or reliability.
The factor of safety method is commonly used in engineering design. It is a time-proven design
tool and when used properly, safe and reliable designs are developed. Factor of safety (FS) is the
ratio of failure stress to an allowable design stress.
Material yield strength, σ yld, is often used as the failure stress. In this case
σ yld σ yld
FS = = 11
σ allowable σa
where σ a is the allowable stress level used throughout a design. When the factor of safety is
given, then
σ yld
σa = 12
FS
Designs are configured (size, dimension) according to the allowable stress.

Factors of safety are intended to cover all uncertainties not identified in a set of specifications.
In general, the higher the factor of safety, the higher the product weight and cost. So, it is impor-
tant to keep the FS as low as possible. Computation accuracy affects factor of safety too. Computer
software based on numerical techniques, such as finite element methods, provide very accurate
predictions of local stress in complex geometries, and thus reduce the uncertainty of stress
predictions.
The statistical or reliability approach seeks to establish design parameters so that the product
performs to an expected level of reliability [3]. Reliability is a statistical measure of performance.
For example, a product reliability of 0.9 means that there is a 90% chance that a given product will
perform its proper function without failure. The method requires statistical data on all random vari-
ables, such as strength, size, and weight. With this approach, products can be designed to a required
level of reliability.
Cost Considerations
Economics should be considered during the early stages of concept generation. Concepts that allow
for early return on investments may be critical. Oil companies want to begin oil production as early
as possible, maybe prior to the completion of the overall production system. Time is money and if it
takes three years to bring a reservoir online that’s three years without return on investment if early
production is not achieved.
Aesthetics
Product appearance is not usually a concern of engineers. However, aesthetics could be an impor-
tant marketing feature and should not be overlooked. Actual performance capabilities versus per-
ceived performance capabilities can be important. The customer may perceive a feature to be weak
or strong depending on the history of a product. Perception of certain product features should be
considered in some cases.
Aesthetics does not mean that a design must be ornamental or that geometry controls the shape
of the design or its components. Manufacturing complex geometries is impractical and costly. How-
ever, if a design is balanced (cost effective and functional), it is usually artful.
Product Life Cycle

Every product has a life cycle, which includes periods of:
1) Development
2) Market growth
3) Market maturity
4) Market decline
5) Product disposal
During the first phase, the product is developed and introduced to the market. This creates neg-
ative cash flow, so developing the product in a timely and cost-efficient manner is important. Dur-
ing the second phase, the product finds its way into the market and can generate cash flow while
establishing itself in the marketplace. Products then reach a level of maturity and are usually able to
capture a portion of the market and generate revenue. During this time, the product generates max-
imum return on investment. At some point in time, the market for a product declines and even-
tually vanishes. This could be the result of new products entering the market or simply the lack
of demand for a product. For these reasons, investors look for new products.
In setting design specifications, remember cost and safety override everything.
Product Safety and Liability

Currently, there are no federal product liability laws. Each state has developed its own legal
approach in this area. Under Tennessee’s product liability laws, there are four theories of recovery
[4, 5]:
Negligence – In legal terms, negligence is the failure to do what people of ordinary care and pru-
dence would do under the same or similar circumstances. When applied to design and manu-
facturing, the question becomes: are decisions, that affect product safety, being made
professionally, objectively, and fairly? Prior to 1970, many personal injury claims were filed
under the negligence theory of law. The plaintiff’s attorney had to establish the standard of
due care required of engineers and manufacturers.
Strict liability – The emphasis under this theory of recovery is on product defect and not on the
person or person’s negligence. The focus of the court is solely on the performance of the product.
In this case, the plaintiff needs to only prove that the product had an unreasonable dangerous
defect when it left manufacturing. The court recognizes three types of defects: manufacturing
defects, design defects, and warning defects. Design as used by the courts encompasses the entire
process by which a product is created and marketed. This theory is more favorable to the plaintiff
because it exposes the entire product development chain to liability, allowing the plaintiff to
choose the most advantageous defendant.
Fraud/misrepresentation – Fraud is a false representation of fact, by works, or conduct, which is
intended to and does deceive another, who then makes decisions based on the false information
and suffers legal injury. The elements required in an action of law are representation, falsity,
knowledge of falsity, intent to deceive the plaintiff, justifiable reliance by the plaintiff, and
damages. Examples are falsifying test data to deliver a product to the customer to meet a dead-
line, approving a design that does not meet required codes to eliminate redesign and fabrication
costs and misuse of Professional Registration approval.
Breach of warranty – Product does not satisfy the expressed warranty or implied warranty for
fitness.
Engineering Ethics
Ethics deals with the principles of human duty, moral principles, and rules of conduct. Engineering
ethics deals with the moral conduct of engineers in serving the public, their employers, and their
clients. What is at stake can be expressed in terms of public safety, public health, and the environ-
ment to mention a few. The challenge is to design and manufacture a product for profit without
undue risk to the general welfare of the public and environment. Ethics comes into play when a
company and its professional staff knowingly produce a product that has a high risk for personal
injury and/or damage to the environment. Unethical decisions are usually made for selfish or mon-
etary reasons.
Engineers make many ethical decisions independent of others. A good check list for these deci-
sions is:
•• Is it legal?
Is it fair?
• Is it morally right; can I live with the outcome?

The ASME has adopted the following Code of Ethics of Engineers for its members1
1) Engineers shall hold paramount the safety, health and welfare of the public in the performance
of their professional duties.
2) Engineers shall perform services only in the areas of their competence.
3) Engineers shall continue their professional development throughout their careers and shall pro-
vide opportunities for the professional and ethical development of those engineers under their
supervision.
4) Engineers shall act in professional matters for each employer or client as faithful agents or trus-
tees and shall avoid conflicts of interest or the appearance of conflicts of interest.
5) Engineers shall build their professional reputation on the merit of their services and shall not
compete unfairly with others.
6) Engineers shall associate only with reputable persons or organizations.
7) Engineers shall issue public statements only in an objective and truthful manner.
8) Engineers shall consider environmental impact in the performance of their professional duties.
Creating Design Alternatives
Once design specifications have been set, design alternatives can be generated. Since design is open
ended, i.e. there are many possible solutions, it is desirable to generate several design concepts,
evaluate them as a group before choosing the best direction for the design. The objective at this
point is to create design concepts that satisfy the specifications. Realistic concepts are ones that
are technically feasible and cost effective.
Innovation is a matter of synthesis and analysis of ideas. It requires time, focus, and effort of
thought. The quality of each concept depends on the ability to think conceptually.
Parker [6] lists some traits of the creative personality:
•• curiosity
risk taker
•• emotionally stable
uninhibited
•• imaginative, original
intuitive
•• high energy level

independent
•• task committed
sense of humor.
Innovation is an individual activity. The foundation of science is built on the genius of

individuals. Great ideas were not usually generated by committees. Once a fundamental
concept or hypothesis is presented, others will add value. Design innovation usually starts with
individual thought. Don’t be afraid to propose novel ideas. You may be the leader others are waiting
to follow.
Tools of Innovation
Ullman [7] discusses useful concept generation tools.
Creating Design Alternatives 13
Patents
Some new products are based on a modification or direct extension of established products, in
which case, the objective is to provide a higher quality or improved version of what is already
on the market. Legally, this is allowed, provided the patent right period has expired. US patent
law provides a 20-year life of a patent from the date the patent application was filed, and no shorter
than 17 years from issuance. A patent survey will show whether your idea is truly novel or is infring-
ing on an existing patent. A patent search early in the development process is worth the cost, time,
and effort. It could eliminate legal problems later.
Reference Books and Trade Journals

Reference books are a good source of information on existing design concepts or designs that are
currently in use. They may give analytical discussions of related designs. These discussions some-
time suggest alternatives based on direct extensions of current design concepts. Many trade journals
feature new products, which may spark new ideas.
Experts in a Related Field

Experts can sometimes provide insight for new concepts. Experts can be research and development
oriented in each technical field or they can be a company representative, such as a salesperson or a
customer service representative, who have detailed knowledge about a product type.
Brainstorming
Ideas are generated spontaneously. Successive ideas feed off the group discussion; “piggyback” is
sometimes used to describe this type of idea generation. All the ideas are reviewed and evaluated for
relevance and practicality. Features of the better ideas will merge and come together.
The rules for brainstorming are:
1) Record all the ideas generated.

2) Generate as many ideas as possible.
3) Think wildly. Impractical ideas sometime lead to a useful one.
4) Do not evaluate or criticize ideas while they are being generated.
Existing Products and Concepts

Many hours have gone into developing existing products and most have the benefit of being tested
in the marketplace. Existing products can be extrapolated into new product variations. This also
eliminates reinventing the wheel so to speak. Types of modifications that may be considered are:
•• Geometric modifications
Energy-flow modifications
– Change in the path
– Change in the form
• Materials used in the product.
Remember, there is no need to redesign every component in a new product. Use as many off-the-
shelf items as possible. Take advantage of current and established technology, such as gears, bear-
ings, and motors. On the other hand, some components may have to be tailored.
As part of the creative process, new ideas (untried techniques) must be proven technically feasible
before they can be accepted as viable alternatives. Feasibility studies may be required to show con-
cepts are fundamentally sound and will work. Some concepts may be evaluated experimentally,
analytically, or both. The results may expose risk and cost of untried concepts.
This is a screening activity to move the better ideas forward and remove alternatives that are not
technically sound. More detailed analyses may be required during the design stage of the develop-
ment process. Concept development may include iterations involving synthesis and analysis as
shown in Figure 1.1. Also, every aspect of product life cycle needs to be considered by the product
development team.
Concurrent Engineering
It is generally accepted that there are three basic activities for developing a profitable product. They
are marketing, designing, and manufacturing. Until the 1980s, these activities were conducted
sequentially, i.e. market data were passed on to designers who subsequently transmitted their
design to manufacturing. These three activities were conducted separately with essentially little
or no collaboration among the three disciplines.
World competition brought about competitive pricing and quick response to a dynamic market.
In response to this challenge, marketing, designing, and manufacturing issues are now being con-
sidered by a team. Marketing seeks and monitors customer input to the new product. Designers
seek the latest customer feedback from market analyst. Manufacturing issues are considered
throughout the conceptual work to avoid costly redesigns brought about by impractical or ineffi-
cient manufacturing requirements. Collaboration of marketing, design, and manufacturing
throughout the development process is commonly called Concurrent Engineering. It substantially
reduces development cost over the sequential operations approach.
Feasibility of Concept
Each concept must prove to be technically feasible before it is evaluated. This means that some
engineering may be required, and even drawings made to advance an idea beyond a hand sketch.
A preliminary study may be needed to show an idea is workable, realistic and satisfies all design
constraints or specifications, including cost constraints. In engineering practice, feasibility analyses
may require extensive computer calculations and/or laboratory testing.
Evaluating Design Alternatives
Innovation will produce several design alternatives, each of which satisfies the given set of design
specifications. The problem now is to choose the “best” concept among many alternatives.
Each of us make choices every day. Without realizing it, trade-offs are made leading up to a deci-
sion. The basic elements in making decisions are performance, cost, risk, and availability. Consider
the purchase of an automobile. The buyer has established mentally a set of performance criteria
(city driving, off road, mountain terrain, luxury, etc.), the price compatible with the family budget,
maintenance, or track record of the car (is it a new model), and availability. The final choice is a
trade-off among these four evaluation elements. The same rationale is used to purchase a suit or
buy a house.
Evaluating Design Alternatives 15
Evaluation Metrics
Four metrics (performance, cost, risk, and availability) are basic in choosing a preferred design con-
cept [8]. The following defines these metrics as they relate to mechanical design:
Performance – Capability to achieve needed operational characteristics, plus reliability.

Cost – Estimated cost of the design, including development and manufacturing costs.
Risk – Possibility that performance may not be met because of the design approach,
absence of testing, or some specific technical consideration.
Availability – Availability of a design depending upon the stage of development.
A procedure for scoring several alternatives is to divide each of these four metrics into key sub-
metrics and give each an appropriate weight. Each design alternative can then be scored under each
submetric. The scoring is strictly judgmental, so experience is important. The more experience, the
better the judgment of scoring. Once each concept has been scored for each submetric, the numbers
are totaled for a composite score for each concept. The total scores provide a means of comparing
each concept against the others.
This numerical evaluation scheme has two key objectives. The first is to have a way to quantify
one’s judgment against a fixed scale so that each design concept can be rated in the same manner,
thus, showing their truest level of merit in comparison with each other. The second objective is to
give a way to examine the rationale of the final scores by looking at the subelements of each concept
to see the strong and weak features of each. Since all design selections are the result of trade-offs, the
scoring system aids in selecting a preferred concept.
Scoring Alternative Concepts

Table 1.1 illustrates the evaluation of four design alternatives and a hypothetical score for each. All
constituents affected by the outcome of the project work together to set the weight given each met-
ric and scoring. Each of the four basic metrics will have sub-metrics, as well.
This example indicates that concept C is the best because it scored the highest among the four
alternatives. Other alternatives may have ranked higher in certain categories, but overall concept
C scored the highest. There may a discussion over individual scores in the various categories, but
the table provides a good basis for discussing the strengths and weaknesses of each alternative.
Results from this scoring method give a rationale for recommending a preferred concept and a jus-
tification for the choice.
Table 1.1 Evaluation summary (hypothetical numbers).
Weight Concepts
(%) A B C D
Performance 35 20 35 30 25
Cost 20 15 10 18 16
Risk 25 10 20 23 18
Availability 20 20 18 16 12
Total score 100 65 83 87 71

Once each alternative has been evaluated and ranked and the preferred design concept selected,
it is always a good idea to revisit the marketplace to see if the customers agree and to get further
feedback for the preliminary design. Keep in mind that each stage in the progression along the
development process represents an upgrade of the engineering baseline. The description and spe-
cifics of the preferred concept are an expansion of the initial set of design specifications and rep-
resent the most recent engineering baseline for the remaining steps in the development process.
Design specifications have therefore been greatly refined beyond the initial set. The design can
now move forward with a detailed description of the product.
Starting the Design
Mechanical devices typically contain a power source and a means of transmitting the power to
bring about a desired end effect.
A design may start with the end effect, which establishes the magnitude of loads throughout the
device, transmission linkages, and power requirements. This sets the overall size of the device and
nominal dimensions of the subparts. Force magnitudes, power requirements (output torque and
speed), pump requirements (flow rates and pressure demand), materials, and criteria of failure, fas-
teners should have been established earlier during the feasibility studies. Forces affect stress mag-
nitude and component dimensions. Types of stresses (bending, torsion, shear, etc.) and how they
combine should be visualized and understood. Are stresses the result of static or dynamic loads?
These initial calculations can be made by hand. More precise calculations, based on computer mod-
els, can be made later.
Design for Simplicity

Keep the design simple. Use off-the-shelf components where possible. Commercial items have
already gone through development and testing and have a proven performance record. Vendors
for commercial products are happy to discuss performance details and quote prices. Also, keep
the number of parts to a minimum. This increases product reliability. One company in East Ten-
nessee rewards individual performance on reduction of design components. In addition, always
consider how each part is to be fabricated. Curved shapes look good but are costly to make. Each
of these considerations will make the device more reliable and easier to maintain in the end. Under-
lying principles of good design practice are:
a) Minimize the number of parts. Reliability varies inversely with number of parts.
b) Keep the design simple, Complexity reduces reliability.
c) Use standard parts when possible; there are plenty of statistical performance data.
When configuring the design, it is helpful to consider whether starting the design from the out-
side-in or inside-out. If the design centers on a specific technical concept, such as a microcanti-
lever sensor, it may be helpful to start there, and work inside-out, i.e. let the cantilever be the
center point of the design and build outward from it. The sensor itself becomes the center point
of the device and will dictate other features, such as how it is to be held and how information is to
be retrieved.
Other projects may be constrained by space or how it links with other devices or subsystems. In
this case, it may be useful to start with the geometric constraint and work inward. For example, the
design of a house starts with the available space on a lot. Many times, the shape of the lot dictates
Development of Oil and Gas Reservoirs 17
the shape of the house, including the foundation, exterior walls, and roof. These subsystems, of
course, are based on the preferred concept, established earlier. The details of the inside, such as
heating and air conditioning, are designed after the size and shape of each room has been
determined.
Identify Subsystems
Consider the total design as a combination of subsystems. This is typically done in very large pro-
jects, but it is helpful in small projects, too. Breaking the whole design into subsystems helps vis-
ualize and organize how the whole design goes together. It also gives a clear division of
responsibility for different team members. Subsystems may include
•• Frame or fixture
Power unit, such as motor
•• Transmission linkages
“Use” mechanism, final form of design
• Controls.
This approach is common in industry especially for huge projects involving several subcontrac-
tors. It provides a clear division of effort and responsibility. Subsystems may also be broken down
into smaller units or sub-subsystems, etc.
The human body is a good example of subsystems. The body contains a frame (skeleton) or bone
structure with consideration of joints and flexibility. The skeleton supports other subsystems, such
as the lungs, heart, kidneys, brain, and all the plumbing. The gastrointestinal subsystem converts
raw fuel into useful energy. Medical doctors specialize in each subsystem.
Innovation continues during this activity and therefore teaming is important. Each team member
has something unique to add and the division of responsibility makes efficient use of each team
member. Team meetings allow ideas to be integrated and refined.
An objective at this point is to configure the design in terms of its shape and subsystems and how
they all fit together. This activity is best made by using classical engineering calculations; simple
hand calculations are fine. These initial calculations establish component sizes, which are refined
later. It is important to get a feel for the magnitude of loads, stress levels, and deflections as you
begin to work through the design.
Development of Oil and Gas Reservoirs
Significant advances in two technologies allowed oil companies to increase production over the past
50 years. They are (i) geophysical mapping of underground rock formations and (ii) directional dril-
ling, which allows navigation through multilayered formations to reach specific locations in deep
and complex reservoirs. A third technology, which has almost doubled oil and gas reserves, is hor-
izontal drilling and fracking. The later technology substantially increases oil and gas recovery.
While seismic surveys can map geologic formations and identify possible oil and gas traps, explor-
atory drilling must be performed to determine the existence of hydrocarbons and chemical com-
position. Once an oil reservoir has been delineated by directional drilling, an economic
evaluation is conducted to determine the best plan for developing the reservoir for maximum
recovery.
Oil companies typically identify business activities according to the following categories.
1) Geophysical surveying
2) Exploratory drilling
3) Production drilling
4) Production of oil and gas
5) Transportation of crude oil and gas
6) Refining crude oil and gas
7) Marketing.
Perhaps 80% of a company’s budget is directed at the first four (4) categories, which are classified
as upstream activities.
Design of Offshore Drilling and Production Systems

The design, fabrication, installation, and operation of offshore production system require an overall
plan that may take a few years to complete. The project begins only after an economic analysis has
been made, including current cost and future value of the asset.
In many cases, an offshore reservoir may extend across multiple sections licensed by different oil
companies. The proportional ownership is determined seismic survey maps. The oil company pos-
sessing the largest portion of the reservoir becomes the operator of the field. All companies pay their
share of development costs and benefit proportionately.
Planning is usually divided into the major activities mentioned above. One of the first activities is
documenting a thorough and complete understanding of design specifications. This may require
extensive data gathering on environmental, oceanographic, soil, and other specifics as mentioned
earlier. Results of these efforts are compiled in a sizable specifications document often called a ten-
der document. The tender document is used to gather additional information on contractor cap-
abilities as well as for contractor bids on certain portions of the project. The bidding process,
involving a few contractors is an opportunity, through discussions, to expand design specifications.
The design, fabrication, installation of offshore oil and gas production systems are multifaceted,
high risk, and costly. The total effort involves a team, including operator, design contractors, fab-
rication contractors, installation contractors, and drilling contractors to mention a few. An Engi-
neering Management contractor is often used to coordinate and interface each contractor’s
activities. The tow-out date of offshore platforms usually depends on the weather pattern for a spe-
cific location. For example, the tow-out and installation window for platforms in the North Sea
starts in May and ends in September. Delays in design or fabrication could result in costly losses
of early production. Upstream planning for this weather window is of the essence.
Major subsystems in offshore drilling and production platforms include
1) Base structure (steel jacket, concrete, tension leg, compliant structure).
2) Operating facilities, including personnel facilities, oil and gas processing for transportation
to land.
3) Drilling and production equipment.
4) Transportation, such as pipelines, of crude to onshore facilities.
Each of these subsystems has its own subparts, etc. Each can be viewed as specific designs, which
interface with the total system. For example, the drilling system typically includes
1) Derrick
2) Hydraulic system
3) Blow Out Preventer system

4) Drillstring including bottom hole equipment
5) Power system.
A few considerations in designing oil well drilling programs are
•• Location of entry point into the oil and gas reservoir relative to the rig.
Assessment of directional drilling equipment needed to reach the entry point as well as the
well path.
•• Formations to be encountered and anticipated drill bit types to penetrate these formations.
Anticipated formation pressures (normal as well as abnormal pressures).
•• Reservoir pressures to be controlled and blow out prevention concerns and strategies.
Casing programs from surface to total depth.
• Implementing optimum drilling practices.
Connection of Subsystems
An early consideration in any design is the attachment of subsystems. Bolt-type attachments allow
removal and replacement of subparts. Bolted connections can also be adjusted. Welded connections
can’t. It may be desirable in some cases to weld subsystems, such as the frame. Warpage and
machining of surfaces should be considered.
Torsion Loading on Multibolt Patterns

Bolts are often used to fasten beams to form a frame or support a given load, such as illustrated in
Figure 1.2. Four bolts are shown even though multiple bolts making up various patterns could be
used. The design objective is to determine the total shear force in each bolt to establish bolt size and
material strength.
The total shear force on each bolt is the vector sum of direct shear force and force caused by the
torsion moment at the support. The first force is simply the total shear force divided by the number
of bolts (assumes the connection is rigid). The direction of the shear force on each bolt is downward.
The second force is caused by a moment on the bolt pattern and is determined as follows.
V
Bolt
rn Beam
Fn
Figure 1.2 Bolted connection.

The total torque, T, applied to the bolted joint relates to shear force in each bolt by
T = F 1 r1 + F 2 r2 + F 3 r3 + 13
Assuming the shear force, F, taken by each bolt depends on distance from the centroid of the bolt
pattern,
F1 F2 F3 Fn
= = = = 14
r1 r2 r3 rn
Combining these two equations gives the force in each bolt.
Tr n
Fn = 15
r 21 + r 22 + r 23 +
The above formula is very similar to the shaft shear stress formula,
Tr
τ= 16
J
which states that shear stress is proportional to radius, r. The shear stress formula is modified as
follows to match the above bolt analysis.
Fn Tr n
τn = = 17
An r 2i Ai
Tr n An
Fn = 18
r 2i Ai
Only when all bolts are the same size,

Tr n
Fn = 19
r 2i
which agrees with Eq. (1.5). The direction of shear forces caused by torque on a bolt pattern is per-
pendicular to the r vectors as shown.
With reference to Figure 1.2, assume
V = 2000 lb
Length = 12 in. (from center of bolt pattern)
Bolt diameter = 0.25 in.
Bolt spacing = 2 in.
Direct shear force 2000

FV = = 500 lb
4
Torsion shear force Tr n T 24 000
Fn = 2 = = = 4243 lb
4r n 4r n 4 1 414
Total shear force F = F V + F n where each force is a vector
F = − 500j + 4243 i cos 45 − j sin 45
F = − 500j + 3000 i − j = 3000i − 3500j
F = 4610 lb (scalar magnitude)
The cross-sectional area of each bolt is A = π4 d2 = π4 0 25 2 = 0 0491 in 2 The maximum shear

4610
stress in the nth bolt is τn = = 93 885 psi. This shear stress level occurs in both inside bolts.
0 0491
Make-Up Force on Bolts

A consideration in bolted attachments is the level of pretightening. Often bolted connections are
subjected to externally applied forces, which may cause further extension of the bolt leading to pos-
sible separation of contacting surfaces.
Consider the two situations illustrated in Figure 1.3. The left drawing shows a bolt compressing a
spring of stiffness, ks. As the bolt is tightened, the shank of the bolt elongates while compressing the
spring. The extension of the bolt shank relates to bolt force by
PL P E B AB
δB = = where k B = 1 10
E B AB kB L
At the same time, the spring is compressed by
P
δS = 1 11
kS
The internal force, P, in the bolt and the spring are the same (Figure 1.4a). The force level is estab-
lished by the make-up torque.
If an external force, F, is applied to the connection, forces in the bolt (PB) and spring (PS) are no
longer equal (Figure 1.4b). The challenge is to determine the magnitude of these two forces in rela-
tion to the externally applied force, F.
δspring
δbolt
ks
Figure 1.3 Make-up force in bolted connection.
Figure 1.4 Force response to an external load. (a) (b) F
P PS
P PB
Consider the freebody diagram in Figure 1.4b. For equilibrium

F = PB − P S 1 12
Since there are two unknown forces in this one equation, it is necessary to consider deflections in
the bolt and spring.
F = P + ΔPB − P − ΔPS
F = ΔPB + ΔpS
where ΔPB is increase in bolt tension and ΔPS is reduction in spring force. F is the externally
applied force.
When load, F, is applied, the bolt stretch increases while the spring compression is relaxed by the
same amount.
ΔδB = ΔδS deflection equation 1 13
In terms of force changes
ΔPS ΔPB EA
= where k B = 1 14
kS kB L
By substitution
kB
F = ΔPS + ΔPS 1 15
kS
giving
ks
ΔPS = F 1 16
kB + kS
Also
kS
F = ΔPB + ΔPS 1 17
kB
kB
ΔPB = F 1 18
kB + kS
The resulting forces in the spring and bolt are
PS = P − ΔPS
PB = P + ΔPB
These relationships are shown in Figure 1.5.

kS
PS = P − F 1 19
kB − ks
kB
PB = P + F 1 20
kB + kS
The force required to separate the surfaces or created zero force in the spring is determined by
setting PS = 0
kB
F cr = P 1 + 1 21
kS
Force
PB
(Bolt)
P
ΔδS
δS
Displacement
δB
ΔδB
PS
P
(Spring)
Figure 1.5 Force changes in connection.
In this case, by substitution into Eq. (1.20)

kB
PB = P 1 +
kS
which is the same expression as Eq. (1.21) as expected.
When two plates are bolted together as shown in Figure 1.3b, the elastic compression in both plates
have been modeled as two springs in series. A common math model for this rather complex state of
strain is a truncated cone under uniaxial loading. Various formulations can be found in Ref. [9].
Example The arrangement of a weight (W) is held in place by a bolt tightened against a spring
(Figure 1.6). Assume the initial compression of the spring is 0.1 in. The frame experiences base
motion defined by
u t = u0 sin ωt 1 22
Frame
This motion may cause the spring to disengage
from the frame due to the acceleration of the u(t) = u0 sin ωt
weight. Ignoring the mass of the bolt and spring,
determine the frequency at which the bolt
becomes loose from the frame.
Other variables are quantified as
W – 100 lb
L – 2 in.
d – 3/16 in. (bolt diameter) w
kS – 2000 lb/in.
u0 – 1/16 in.
Figure 1.6 Spring-mounted weight.

To solve, we use Eq. (1.21) with the parameters below.

P = k s δs = 2000 0 1 = 200 lb initial spring force
2
π 3
AB = = 0 0276 in 2
4 16
30 × 106 0 0276
kB = = 414 000 lb in
2
100
F cr = u0 ω2 M = 0 0625 ωcr 2 = 0 0162 ωcr 2
386
Substituting the numbers gives
2 414 000
0 0162 ωcr = 200 1 +
2000
ωcr = 1602 rad s
f = 255 cps
Preload in Drill Pipe Tool Joints

In oil well drilling, joints of drill pipe are connected by
“tool joints” welded to the pipe body. At one end is the
pin joint, at the other end is a box joint (Figure 1.7).
As the total string of pipe is drilled down, new joints
of drill pipe are added. The connections are critical. If
joints are torqued to lightly, then bending stress may
be created within the pin leading to fatigue. Also, if
the shoulder in the connection loose, high pressure of
Pin the drill fluid could literally cut through the threads
causing a washout. Proper make up torque of each joint
Make-up torque
is therefore critical.
The American Petroleum Institute (API) recom-
Box
mended (or allowable) preload in the shoulders of both
box and pin for new pipe/tool joints is 50% of the force
required to yield either box or pin. This preload is gener-
ated by make-up torque. For used drill pipe tool joints, it
is 60%. This level of preload allows for additional pin
loading due to direct pull, such as drillstring weight
and over-pull.
After the tool joint is made up by the tongs (special tor-
queing tools), drill pipe is lifted vertically, and the slips
Slips
removed. At this point, both shoulder and internal pin
forces change from their make-up values. The questions
then become (i) What is the new shoulder contact force?
and (ii) How much direct pull would cause shoulder
separation?
The following explains how pin force and box shoulder
Figure 1.7 Tool joint make-up torque. force change when external pull forces are applied across
tool joints. The problem is statically indeterminate, so deflection equations are required along with
the static forces.
For the sake of discussion, let the pin and box be represented by a cylindrical bar and tube as
shown in Figure 1.8a. The tube is slightly longer than the bar. If the two heights are brought
together and fixed, the preload (tension) in the pin is equal in magnitude to the preload (compres-
sion) in the box (Figure 1.8b). The pin is extended by
FLp F
δp = = 1 23
EAp kp
The box is compressed by

FLb F
δb = = 1 24
EAb kb
where
F – preload in pin and box after make up, lb

Ap, Ab – cross-sectional area of pin and box at the pitch point, in.2
Lp, Lb – active length of pin and box (~0.75 in.)
kp, kb – effective spring constants in pin and box, lb/in.
E – modulus of elasticity, psi.
(a) (b) (c)

P
Δδ
δb
δp Box
Pin
(d)
P
Fb
Fp
Figure 1.8 Internal forces due to preload and pull.

Now assume that a pull force, P, is applied across the tool joint (Figure 1.8c). The pin is stretched
by Δδp while the box is relaxed by Δδb: note that Δδp = Δδb = Δδ Therefore, the corresponding
increase in the pin force is related to the corresponding reduction in the box force by
ΔF p ΔF b
= 1 25
kp kb
where ΔFp and ΔFb are changes in pin and box forces.
Using the freebody diagram shown in Figure 1.8d,
P = Fp − Fb 1 26
where Fp and Fb are new forces in pin and box resulting from the force, P, across the tool joint. These
forces can be expressed as
F p = F + ΔF p 1 27
F b = F − ΔF b 1 28
Bringing Eqs. (1.25)–(1.28) together gives
kp
ΔF p = P 1 29
kp + kb
Therefore, the new force in the pin due to both preload and direct pull is
kp
Fp = F + P 1 30
kp + kb
where
P – pull force across tool joint

F – initial preload in box/pin
Fp – new force in pin.
Similarly, the new force in the box is

kb
Fb = F − P 1 31
kp + kb
Two load conditions are of interest: (i) magnitude of P, which causes shoulder separation and
(ii) possible yielding in the pin caused by P.
Shoulder Separation
Shoulder separation occurs when Fb = 0. Using Eq. (1.31) with Fb = 0 gives
kp + kb
P= F 1 32
kb
which is the pull force across the tool joint required for shoulder separation.
Possible Yielding in the Pin

Rewriting Eq. (1.30) in terms of yield force in the pin
kp
F yld pin
=F+ P 1 33
kp + kb
The force across a tool joint causing yielding is
kp + kb
P= F yld − F 1 34
kp
This equation applies only if there is no shoulder separation.

Combining Eqs. (1.30) and (1.32) and assuming separation when (Δδ = δb) and F = 0,
Fp = P 1 35
as expected.
Example Consider an oil field drill pipe tool joint as shown in the figure. Assuming the make-up
contact force on the two shoulders of the pin and box is 472 000 lb, determine the shoulder force
after a hook force of 250 000 lb is applied across the tool joint when the slips are removed. Deter-
mine the magnitude of a pull force required to separate the shoulder (zero contact force).
The dimensions for calculating these areas are:
Box OD = 6 3/8 in.
Pin ID (bore) = 33/4 in.
Pitch diameter = 5.04 in. (first thread on pin, see API RP 7G).
From the given dimensions, Ap = 8.93 in.2 and Ab = 11.97 in.2 and the active pin and box lengths
(Lp = Lb ~ 3/4 in.) are the same. Then
kb Ab 11 97
= = = 0 573
kp + kb Ap + Ab 8 93 + 11 97
kp Ap 8 93
= = = 0 427
kp + kb Ap + Ab 8 93 + 11 97
Force in the pin is

kp
Fp = F + P 1 36
kp + kb
F p = 472 000 + 0 427 250 000 = 578 750 lb
Similarly, the new force in the box is

kb
Fb = F − P
kp + kb
F b = 472 000 − 0 573 250 000 = 472 000 − 143 250 = 329 000 lb 1 37
This represents the new shoulder force after the 250 000 lb load is applied.
The pull force required to separate the pin/box shoulder is (using Eq. (1.32))
1
P= 472 000 = 823 700 lb
0 573
This force is felt by the pin at separation. It is less than the 944 000 lb required to yield the tool
joint. This number, P, far exceeds the strength capacity of each of the four pipe grades (4½ in.),
which means that the drill pipe body would fail before tool joint shoulders separation.
Make-Up Torque
There are various formulas used to predict make-up torque, which account for friction in the
threads and shoulder. One that is commonly used to predict make-up torque in drill pipe joints
is called the screw jack formula.
F p Rt f
T= + + Rs f 1 38
12 2π cos θ
where
T – torque, ft-lb
p – lead of threads, inches (p = ¼ in. for 4 threads/in.)
Rt – average mean radius of threads, in.
f – coefficient of friction (~0.08)
Rs – mean radius of shoulder, in.
θ – ½ of included angle of thread (2θ = 60 )
F – contact force between mating shoulders, lb
The screw jack formula shows that total make-up torque is distributed among three areas. One
component of the torque drives the mating shoulders together creating the contact force. If there
were no friction in the connection, all the applied torque would create this force. From energy con-
siderations, the work done by a torque over one revolution is equal to the work to move an axial
force, F, over one threat pitch.
2πT 12 = Fp 1 39
The second component of torque is the torque required to overcome friction in the threads. The
third torque component overcomes friction in the shoulder.
If tool joints, connecting drill pipe, are made up too tight, the pin could be overstretched. If tool
joints are made up too low, threads can be exposed to washouts or bending fatigue. In practice, tool
joints are generally made up too low. According to one drilling engineer, “if tool joints are made up
too low, they will not make up downhole. On the other hand, drill collar connections will make
up downhole. If drill collar connections are difficult to break out, it usually means they were made
up too low initially.”
The make-up force or internal contact force between the two mating shoulders directly affects
the structural integrity and pressure sealing capacity of the tool joint. Consider, for example,
the torque components in a new NC50 (63 8 in. × 33/4 in.) tool joint using the following set of
numbers:
p = 0.25 in.
f = 0.08
θ = 30
Rt = 2.385 in.
Rs = 2.922 in.
Substituting these numbers into Eq. (1.28) gives

F 0 25 2 385 0 08
T= + + 2 922 0 08
12 2π 0 866
F
T= 0 0398 + 0 2203 + 0 233
12
F
T= 0 4931
12
This calculation shows that only 8% of make-up torque drives the shoulders together. The
remaining 92% is used to overcome friction. The coefficient of friction of thread dope is therefore
critical. Make-up torque recommendations [10] are based on a coefficient of friction of f = 0.08.
If the actual coefficient of friction is less than 0.08, contact force between the mating shoulders
will be higher than expected when the recommended API make-up torque is developed. This con-
dition could over stretch the pin or damage the shoulder area. If the actual coefficient of friction is
greater than 0.08, contact force between the mating shoulders will be lower than expected when the
recommended API make-up torque is developed. This condition allows bending stresses to reach
the pen causing fatigue damage and creates an inadequate pressure seal leading to a washout
through the threads.
Bolted Brackets
Things to consider in bracket design are indicated in Figure 1.9. The bracket supports a force, F,
which is resolved into vertical (FV) and horizontal (FH) components. Forces in both bolts (A and
B) are the superposition of several types of loads. First, the vertical force component is supported
evenly by both bolts. The horizontal component creates shear forces evenly in both bolts as well. In
addition, the horizontal component creates a moment (bFH) about point “a,” which adds to bolt
forces.
This moment creates a statically indeterminate problem, which requires deflection considera-
tions. Equations to be considered are
LB
δB
δA
B A
a
FH
FV
Figure 1.9 Bracket attached by two bolts.

bF H = F B lB + F A lA one statics equation, two unknowns 1 40

δA lA
= deflection equation 1 41
δB lB
The stretch in both bolts is
FAL FBL
δA = and δB = L is bolt length 1 42
EA EA
Combining Eqs. (1.41) and (1.42) gives a second equation for determining the forces in each bolt.
FA lA
= 1 43
FB lB
Substituting this result into Eq. (1.40) gives
2
lA
bF H = F B lB 1 + 1 44
lB
Welded Connections
Torsion Loading in Welded Connections
The analysis of welded connections is similar to bolted connections. In this case, total shear stress at
any point in a weld is the superposition of direct shear and torsion shear. Direct shear stress is deter-
mined by
F
τF = 1 45
Atotal
where F is total shear load on the joint and Atotal is the total throat weld area (tL). Local shear stress
due to applied torque, T, is assumed to follow classic shear stress predictions.
Tr
τT = 1 46
J0
where J0 is the polar moment of inertia of the weld pattern.
When the size and shape of each weld pattern is known, these equations can be used to determine
both shear stress components. The resulting shear at any point in the weld is the vector sum of both
stress components.
Consider for example a steel plate attached to a vertical post by two fillet welds (Figure 1.10). For
this weld pattern
dJ 0 = r 2 dA = h2 + x 2 dA 1 47
ℓ ℓ
2
J0 = h t dx + x 2 t dx 1 48
−ℓ −ℓ
2 3
J 0 = h2 t 2ℓ + ℓt 1 49
3
L
Combining parameters, l = 2 and A = Lt.
F
l x
τT
h r
c
τF
0
Weld area, A
L
Figure 1.10 Welded connection in torque.
3
2 L
J 0 = h2 A + t 1 50
3 2
1 2
J 0 = h2 A + L A one weld 1 51
12
A = tL t is measured across triangular weld
Maximum shear stress occurs at the end of the weld. This stress is the vector sum of two
components.
τmax = τT + τF vector summation 1 52

Tc F
τmax = + 1 53
2J 0 2A
Example Consider the specific numbers shown below.

L = 8 in.
t = 0.25 in.
h = 3 in.
F = 250 lb
T0 = 3000 ft-lb (assumed)
A = t L = 0.25(8) = 2 in.2 (Atotal = 2A = 4 in.2)
J 0 = 2 32 2 + 12 8 2 = 57 3 in.4 (both welds)
1 2
The stress components are
3000 12 5
τT = = 3141 psi torsion load
57 3
250
τF = = 62 5 psi direct load
4
The shear stress produced by the direct loading is minor compared to the shear stress produced by
the torque.
Attachments of Offshore Cranes

Cranes are essential in the operation of offshore structures. They load and unload necessary equip-
ment and supplies from supply boats. Off-loading cranes are often attached to a base structure, such
as a platform or floating vessel by weldments. A weldment with gusset plates is illustrated in
Figure 1.11.
The stresses in the weldment are due to a bending moment produced by a hook load in the crane.
The crane is rotated about a vertical axis after the equipment has been elevated to deck level. The
classic calculation is based on
Mc
σ= 1 54
I
Gusset plate
R
Gusset plate
θ
x
Figure 1.11 Offshore crane weldment.

By inspection we note that moment of inertia, Ix = Iy and each are principal moments of inertia.
They are both principal moments of inertia and are represented by a point in the Mohr inertia circle.
Furthermore, the moment of inertia about every diametrical axes is the same. Bending stress mag-
nitude varies by distance, c. The c value is maximum when bending occurs about the x or y axis. One
of the engineering challenges is to determine the moment of inertia. We choose to use the x axis as
the reference line.
The moment of inertia of the ring is determined as follows.
dI x = R sin θ 2 t ds 1 55
π
2
I x = 4R3 t sin 2 θ dθ 1 56
0
π
θ sin 2θ 2
I x = 4R3 t − 1 57
2 4 0
I x = πR3 t 1 58
The moment of inertia of two gusset plates with respect to the x axis (using the transfer formula) is
ta3 a 2
Ix = 2 + R+ at 1 59
12 2
Total moment of inertia with respect to the x axis is
ta3 a 2
I x = πR3 t + 2 + R+ at 1 60
12 2
If gusset plates are placed evenly around a cylindrical stand, bending moments of inertia about
any diameter are the same. This means the Mohr circle of inertia is a point.
Quality Assurance
Performance evaluation is necessary for nearly every product or service. Evaluation can be based on
quality assurance at the assembly line, laboratory testing, product performance in the marketplace,
and customer feedback, to mention a few. The metrics for any evaluation should be established
early. The metrics provide a baseline from which to evaluate performance. It is useful to include
the stake holders and end user in setting the metrics. Performance should be given a numerical
score and the rationale behind the score.
•• Functionality
Root cause analysis
•• Maintenance and reliability

Market response
• Assessment/feedback
Engineering Education
Engineering education is a good example of program assessment and feedback. A few years ago,
engineering programs were based on curricula containing certain required components, such as
mathematics, humanities, engineering science, and engineering design. In 2000, the Accreditation
Board for Engineering and Technology (ABET) initiated a different set of criteria based on objec-
tives, curricula, and a continuous improvement process.2 Each academic unit can design or modify
existing programs to satisfy these criteria. The flow diagram (Figure 1.12) illustrates such a feedback
process.
Mission Statement
The mission statement provides focus. It sets direction. It is much like the operational requirements
of Figure 1.1. For example
To provide a broad-based integration of courses and experience

That prepares its graduates to practice their profession successfully, to apply their skill to
solve current engineering problems collaboratively, and to help advance the knowledge
and engineering practice in their field.
This statement is reviewed from time to time with the constituents of the program.
Academic Design Specifications

Design specifications are set by the ABET. Each engineering program must satisfy each of these
components. These criteria offer the opportunity to redesign or modify existing undergraduate pro-
grams. There is a lot of flexibility in how programs are designed. It is very open ended. Typically,
well-established programs are modified to satisfy these criteria and continually improved.
Mission
Constituents
statement
ABET criteria
Feedback
Academic program
mechanism
B. S. graduate
Assessment
Figure 1.12 Feedback process of continuous improvement.

Design of the Academic Program

An academic program includes the integration of every component (students, curricula, faculty,
facilities, and resources) as spelled out by the eight ABET criteria. Curricula are at the center of
the academic program.
Perhaps the biggest improvement in engineering education has been the requirement that all
seniors must have a major design experience. While the basic engineering science courses are fun-
damentally the same, the use of computers, modern laboratory equipment, and classroom equip-
ment have brought about big changes in how courses are taught. In my opinion, the greatest
improvement to engineering curricula has been the Professional Component (Criteria 4), which
requires a major capstone design experience by each student prior to graduation. These projects
bring seniors face to face with the realities of applying engineering principles to solve practical engi-
neering problems. These experiences help engineering graduates step directly into an industrial
setting and immediately take on projects with confidence.
Graduate engineers are the product of this service. Just as industry evaluates the quality of its
service products, academia should do the same. The question becomes one of how prepared are
graduates to enter the work force. Have the graduates gained proficiency in the expected outcomes
as specified in Criteria 3 and 5?
Outcomes Assessment
What are the tools for measuring the results of the educational program? Evaluation and assess-
ment must be quantified in some way to measure improvements in the educational service from
year to year. The outcomes assessment process should involve all constituents involved or those
who have a stake in the outcome.
Once the outcome of the educational service has been evaluated, how are the conclusions fed
back into the educational process?
Saturn – Apollo Project

A few years ago, I had the pleasure of visiting with Albert C. Martin, Director of Launch Operations
for Stage II, Saturn/Apollo Project (Albert C. Martin, personal communication). We discussed the
steps leading up to each launch. Testing was a huge part of launch preparation.
During the early 1960s, the digital computer was just emerging as an engineering tool. Software,
as we have today, was not yet available for making numerical calculations of critical aspects of the
rocket. Thus, reliability of design, along with improvements, depended heavily on testing.
Test results were fed back to the design team for improvements. This cycle was made until each
team could say “… that everything that can possibly be done to assure success has been done.” Final
testing was conducted at the Kennedy Space Center, Cape Canaveral, prior to rocket assembly and
launch. Over 400 engineers and technicians conducted multiple tests on critical component of the
Saturn Stage II under the direction of Lyle C. Bjorn, Manager of Testing. There were no launch
failures during the Saturn/Apollo Project.
Important aspects of design in the Saturn/Apollo project were
Produce ability
Reliability
Safety.
A “Design Review and Change Board” reviewed results after each test. Information gained from
testing was evaluated for possible improvement to the overall second stage design. Before any
design changes were made, the Change Board scrutinized the recommendations. Representatives
from various contributors attended these meetings, including
Manufacturing
Design and engineering
Structures
Aerodynamics
Testing
Combustion
Quality control
Financial
Contracts
These types of meetings were held regularly since second-stage testing was conducted around the
clock and hundreds of data had to be evaluated. While every component was essential, the turbo
pump was central to the overall operation of each engine. To produce the necessary thrust, a high
rate of fuel burning was vital. That meant the turbo pumps operated at near 80 000 rpm. Bearings
that held the turbine shaft were pushed to the limit at this speed. Also, turbine blades were suscep-
tible to vibration and fatigue, and had to be monitored.
Since then, computer capability in both size and speed, along with computer software, make it
possible to accurately predict expected performance of critical components. Computer simulation
of complex systems (airplanes, for example) gives accurate predictions of expected performance,
essentially eliminating the need for many tests.
Notes
1 http://www.asme.org/Education/PreCollege/TeacherResources/Code_Ethics_Engineers.cfm.
2 Established by the Accreditation Board for Engineering and Technology (ABET)..
References
1 Federal Standardization Manual. (2000). U.S. General Service Administration, Federal supply
Service, http://www.fss.gov/pub.
2 Packard, V. (1960). The Waste Makers. David McKay Co.
3 Shigley, J.E. and Mischke, C.R. (1991). Mechanical Engineering Design, 5e. Marcel Dekker, Inc.
4 Peters, G.A. (1971). Product Liability and Safety. Coiner Publications, Ltd.
5 Heidelang, H. (1991). Safe Product Design in Law, Management and Engineering. Marcel Dekker, Inc.
6 Parker, J.P. (1989). Instructional Strategies for Teaching the Gifted. Allyn and Bacon, Inc.
7 Ullman, D.G. (1997). The Mechanical Design Process, 2e. McGraw-Hill Book Co.
8 Dareing, D.W. (2010). Engineering Design and Problem Solving. Tennessee Valley Publishing.
9 Shigley, J.E. and Mitchell, L.D. (1983). Mechanical Engineering Design. McGraw-Hill Book Co.
10 API (RP 7G) (1987). Recommended Practice for Drill Stem Design and Operating Limits. American
Petroleum Institute.
37
Configuring the Design
This chapter reviews classical analytical tools useful for determining shapes and sizes of product
components. Configuring a design refers to dimensions and interface between every part.
A good place to start the force analysis is at the utility end of the design. Magnitudes and type of
loading will come from the expected result or “use” point. Reactions are transferred through the
“transmission” linkages to the frame. Transmission loads will define equipment specifications
and size and shape to the transmission linkages (gears) the frame must accommodate. Defining
these forces, moments, or torque will also establish the input power required to drive the device.
Classical equations are good enough at this point.
Material yield strength, σ yld, is usually taken as the failure stress. The allowable design stress, σ a,
depends on the factor of safety.
σ yld
σa = 21
FS
The allowable force, Fa, is related to allowable design stress by
Fa
σa = 22
A
Three types of design situations may occur.
1) Fa = σ aA (area given, determine allowable load, Fa),

2) A = σFa (force given, determine structural dimensions, A),
3) σ a = F
A (force and dimensions given, determine material strength).
The same scenario applies for bending and torsion loads. Using elementary stress formulas and
classic mechanics Mc Tc
I , J is fine. Computer software can be used later to refine the design.
As components are being configured, always consider: how the part is to be made? A machinist is
a good source of information.
Force and Stress Analysis
An important aspect of design is determining external and internal forces and visualizing how
forces flow throughout the design. This will help to determine shapes and member sizes consistent
with material strengths and factors of safety. Freebody diagrams of subsystems are helpful.
38 Configuring the Design
F/2 t2
F
F/2
t1 F
S
d
S
w F
S
Figure 2.1 Simple stresses.
Example Consider the double strap connection with the following load conditions (Figure 2.1).
The problem is to determine
1) Thickness, t2, and width of outside straps.

2) Width of center strap.
3) Set-back location, d, of bolt.
Design specifications for this example are:
Connection load F = 5000 lb

Bolt size D = 3/8 in. (area = 0.1104 in.2)
σ yld = 75 000 psi (τyld = 43 275 psi) (straps)
σ yld = 120 000 psi (τyld = 69 240 psi) (bolt)
Factor of safety FS = 2
Thickness of middle strap t1 = ¼ in.
Step 1 – Check shear stress in bolt

S 2500
τ= = = 22 645 psi
A 0 1104
τyld 69 240
FS = = = 3 058
τ 22 645
Step 2 – Check normal stresses in middle strap and determine width, w.
Beam Analysis 39
σ yld
FS =
σa
75 000
σa = = 37 500 psi
2
F
σa =
w − D t1
5000
37 500 =
w − 0 375 0 25
w = 0.9083 in. (compare with standard strap iron width)
Step 3 – Check tear out shear stresses in middle strap
τyld
FS =
τa
43 275
τa = = 21 637 psi allowable shear stress
2
Equating to
S 2500 10 000
τa = = =
dt 1 d 0 25 d
gives
10 000
21 637 =
d
d = 0.462 in. (set-back distance, 0.5 in. OK)
Step 4 – Check normal stress in two outside straps and determine their thickness. Since both outside
straps carry ½F, each cross-sectional area must be ½ area of middle strap. Therefore
w = 0 9083 in
t 2 = 1 8 in
d = 0 462 in
In many situations, stresses resulting from various types of loading will be combined to form a
two-dimensional or three-dimensional state of stress. It is important to visualize loading situations
that produce the more general states of stress.
Appropriate criteria of failure must be used. The von Mises criteria of energy distortion are com-
monly used for static load situations. Other modes of failure are fatigue and corrosion or any com-
bination of static, fatigue, and corrosion. All possible modes of failure should be considered.
Beam Analysis
Beams are often used to support subsystems. They can experience various types of loading as they
provide a structural frame for designs. Bending is the most significant load because it can generate
relatively large stresses in comparison with direct pull or shear. For example, it is easier to break a
stick by bending than by direct pull. That’s because bending stresses are higher than axial stresses
for the same load. Consider a solid rectangular beam that is 2 by 4 . If 1000 lb is applied at the end
of a foot-long cantilever beam, the bending stress is 22 500 psi (stiffest axis) and 90 000 psi (flexible
axis). On the other hand, if the 1000 lb force is applied in tension, the tensile stress in the same
member is only 125 psi.
The notion of equilibrium has been applied in numerous scientific fields, especially engineering.
The acceptance and application of this science was slow to develop. As late as the early 1800s, many
engineering designs were still based on empirical and historical approaches. Even the first metal
bridge (cast iron) across the Severn River (1779) was developed empirically based on the shape
of stone arch bridges.
In the early days, many felt science was too mathematical for practical use. By mid-1800s, the
demand for longer bridges and larger (and safer) buildings brought about the scientific-based
design approach. The Eads Bridge (designed and constructed by James Eads) spans the Mississippi
River at St. Louis (Missouri) [1]. It was based on truss design. It was completed in 1867 and was the
first scientific-based engineering bridge. It is still in use today. Eiffel, the French engineer, applied
structural engineering principles to several bridges (1884) and structures, including the Eiffel
Tower (1889) and the Statue of Liberty.
The Raftsundet Bridge (Norway) [1], completed in 1998, has the longest (978 ft) beam span ever
for a bridge.
Shear and Bending Moment Diagrams

Beams are defined as structural members that carry transverse forces. Beams can be mounted in a
variety of ways, such as simple supports and fixed supports.
Shear and bending moment diagrams show how internal shear (V) and internal bending (M) are
distributed within beams, caused by external loading. Boundary conditions must also be consid-
ered. These diagrams show the location and magnitude of maximum shear and maximum bending.
The signs given to internal shear forces (V) and moments (M) are important because they help
establish the direction of normal and shear stresses. While somewhat arbitrary, the sign conven-
tions shown in Figure 2.2 are used here.
Example Consider the simply supported beam carrying a concentrated force at midpoint
(Figure 2.3). The size of the beam is not a factor except for its distributed weight, which may or
may not be considered. It is not considered in this example.
Shear, V Moment, M
M
F +V +M
x x
Figure 2.2 Sign convention for internal shear and bending.

Beam Analysis 41
Figure 2.3 Shear and bending moment diagrams. L
x F = 1000 lb
5 ft 5 ft
V 500 lb
2500 ft-lb
M
Mathematical expressions for shear and bending moments are
V = +500 lb (0–5 ft)

V = −500 lb (5–10 ft)
M(x) = 500x ft-lb (0–5 ft)
M(x) = (L − x)500 ft lb (5–10 ft)
with
Mmax = 2500 ft lb at x = 5 ft
Example Consider a beam supporting a distributed load (Figure 2.4). The total load distribution,
w, includes the weight of the beam along with a live load.
a
w = 100 lb/ft
x
a
L = 10 ft
V
1250 ft-lb
–500 lb
M
Figure 2.4 Example with distributed load.

Using a freebody diagram of the beam from the left end to section a–a, internal shear and bending
are expressed mathematically as
V x = 500 − 100x 23
M x = 500x − 50x 2
24
Maximum shear is located at x = 0 and x = 10 ft. Maximum bending moment is located by
dM
= 500 − 100 x = 0 25
dx
x = 5 ft
By substitution into Eq. (2.4)
2
M max = 500 5 − 50 5
M max = 1250 ft-lb
Shear and bending moment diagrams can also be constructed graphically. The method allows
both diagrams to be sketched while identifying key aspects, such as maximum and minimum
values, of each diagram. Equations that are useful in constructing shear and bending moment dia-
grams are developed from the diagram given in Figure 2.5.
Useful equations from this diagram are
dV
= w slope of the shear diagram equals the local load rate 26
dx
2
V2 − V1 = w x dx area under the load curve 27

1
dM
= V slope of the moment diagram equals the local shear 28
dx
y w(x) Figure 2.5 Freebody diagram of a beam element.
M M + dM
V V + dV
dx
Beam Analysis 43
M2 − M1 = V x dx area under the shear curve 29

1
Applying these equations to the problem in Figure 2.4 gives

dV
= − 100 lb ft slope of the shear diagram
dx
The freebody diagram of the total beam shows the shear at both supports is 500 lb. Taking point 1
at the left boundary, the shear at the midpoint (x = 5 ft) is
V 2 − 500 = − 5 100
V 2 = 0 at x = 5 ft
Taking point 1 at the left end and point 2 at the right end and applying Eq. (2.7)
V 2 − 500 = − 100 10
V 2 = − 500 lb
which agrees with the results from the freebody diagram.

Since the ends of the beam are assumed to be simply supported, moments at both ends are zero.
Noting that the slope of the moment diagram is equal to local shear force, and shear force is zero at
x = 5 ft, the maximum bending moment occurs at this point. Its magnitude is the area under the
shear diagram between locations 1 and 2.
1
M max = 5 500 = 1250 ft-lb
2
Note the moment is (+) plus causing the beam to bend upward.
If a concentrated force or moment is applied anywhere on the beam, the forces and moments in
Figure 2.6a and b apply.
If for example a concentrated force (F = −1000 lb) is applied at the center of the beam shown in
Figure 2.7, internal shear is reduced by 1000 lb across the concentrated force as shown in the shear
diagram. The maximum internal bending moment is determined graphically by
M max = M 5 = area 1 + area 2
1
M 5 = 500 5 + 500 5 = 3750 ft-lb
2
Figure 2.6 Stepping across concentrated force and F

moment. C
Vl Vr Ml Mr
y F = 1000 lb
w = 100 lb/ft
5 ft 5 ft
V
500 lb
3750 ft-lb
–1000 lb
Figure 2.7 Concentrated force with distributed force.
y 1000 ft-lb
w = 100 lb/ft
L = 10 ft
600 lb
100
V
1750 ft-lb
–400
M 800
750
1 ft
Figure 2.8 Local moment added to distributed loading.
The slope of the moment diagram is not zero at the midpoint because the shear is no longer zero.
There is a discontinuity in the moment diagram at x = 5 ft.
If a 1000 ft lb couple is applied at the center point, shear and bending moment diagrams would be
modified as shown in Figure 2.8.
A good check on each shear and moment diagram is reaction forces determined from a freebody
diagram of the loaded beam. They should match the shear forces and moments at the ends of the
two diagrams. The slope of the shear diagram is equal to the load rate, w, while the slope of the
moment diagram is equal to the local internal shear, which is equal to zero at x = 6 ft.
Beam Analysis 45
Bending Stresses
Bending stresses and deflections in beams are based on Euler’s model [2, 3] assuming a transverse
plane remains flat after bending, i.e. transverse planes do not warp (Figure 2.9). From this, normal
strain is linear with distance from the neutral axis of bending.
y dθ
ε= − 2 10
dx
Note, dx = ρ dθ, so
y
ε= − normal strain 2 11
ρ
y
σ = − E normal stress 2 12
ρ
Summing the differential forces (σ dA) across the beams cross section gives
1
σ dA = y dA = 0 2 13
ρ
A A
The integral on the right side defines the centroid of the area; therefore, the neutral axis of bend-
ing is the same as the centroid of the cross-sectional area.
The summation of the differential internal moments across the beam gives
E 2
M = y σ dA = y dA
ρ
A
EI
M= 2 14
ρ
Two important equations emerge from this derivation.
d2 y
EI = M for small deflections 2 15
dx 2
and by combining Eqs. (2.12) and (2.14)
My
σ= normal stress is compressive with + y 2 16
I
Figure 2.9 Euler’s theory of beam stress. ρ

y
dθ
n
n y
dx
This equation is used along with the moment diagram to determine maximum bending stress in a
beam. The bending diagram gives location and magnitude of the maximum bending moment in the
beam. Equation (2.16) defines the bending stress at this location.
The objective in structural design is to select the lightest beam that is strong enough to support
the applied loads. Material cost is directly related to weight of structure. The location and magni-
tude of the maximum bending moment is determined from the moment diagram. Cross-section
properties of many standard beams (H beam, I beam, T beams, channel beam, and angle iron)
are given by the American Institute of Steel Construction (AISC) [4]. Composite beams can also
be formed by combinations of standard beams, such as channel beams and angle iron beams.
Beam selection depends on the required section modulus based on the allowable stress. Bending
stress is the greatest at the point of maximum bending, as established by the bending moment
diagram.
M max c M max
σ max = = 2 17
I S
The allowable stress is found by
σ yld
σa = 2 18
FS
where FS is factor of safety. The required section modulus is determined from
M max
S= in 3 2 19
σa
It is the basis for beam selection.
Example Assume, for example, the required section modulus of a beam is S = 55 in.3 The lightest
Standard I beam satisfying this requirement is (S15 × 42.9). This beam is 15 in. high, weighs 42.9 lb/
ft, and has a cross-section area of A = 12.57 in.2 Since cost is directly related to beam weight, it is also
important to choose a beam having the lightest weight.
Note, weight per unit length and cross-sectional area are related by
wℓ = Aℓγ stl 2 20
1 ft2
w=A 492 lb ft3 = 3 417A
144 in 2
where A(in.2) and w(lb/ft).
Table 2.1 compares the relative strengths of different cross sections having the same area or
weight per foot. The section moduli are compared to those of a solid rod having a diameter of 4
in., whose section moduli is 6.28 in.3 The section moduli of the H beam and Standard I beam
Table 2.1 Comparison of section moduli for different shapes. A = 12.57 in.2, w = 42.95 lb/ft.
Shape Section modulus (S, in.3) Strength ratio
Circular rod (4 in dia) 6.28 1

Pipe (1 4 in. wall) 49.5 7.88
S15 × 42.9 59.6 9.49
W14 × 43 62.7 9.98
Beam Analysis 47
are significantly higher than those for the solid bar. The diameter of the pipe is 16.23 in. with a ¼-in.
wall thickness. While the weight in each case is the same, there is a wide difference in the
section moduli.
Beam Deflection and Boundary Conditions

Develop the mathematical deflection function for the cantilevered beam shown. What is the max-
imum deflection? Consider a uniformly loaded cantilevered beam with a concentrated force at its
free end (Figure 2.10).
The deflection function for this beam can be approached in two ways. First by starting with
d2 y
EI =M 2 21
dx 2
The local internal bending moment is
L−x 2
M x = − L−x F + w 2 22
2
d2 y L−x 2
EI = − L−x F + w 2 23
dx 2 2
Both F and w create negative curvature in the beam. By integration and applying boundary
conditions
dy 1 1
EI = − − L − x 2 F − L − x 3 w + C1 2 24
dx 2 6
1 1
EIy = − L − x 3F + L − x 4 w + C1 x + C2 2 25
6 24
Boundary conditions are
dy
y 0 =0 and =0 2 26
dx x=0
y F
x
EI
L
Figure 2.10 Deflection of cantilevered beam.

Giving
L2 L3 w L3 L4 w
C1 = F+ and C2 = − F− 2 27
2 6 6 24
The maximum deflection occurs at x = L
L3 L4 w
EIymax = − F+ 2 28
3 8
Note the total deflection at x = L is the effects of the two loads superimposed as one expression.
The minus sign is the downward direction. The maximum slope of the deflection occurs at x = L.
FL2 wL3
EIθmax = − +
2 6
The second approach starts with
d4 y
EI = −w 2 29
dx 4
By stepwise integration and applying boundary conditions
y 0 =0 2 30a
dy
=0 2 30b
dx x=0
d2 y
=0 2 30c
dx 2 x=L
d3 y
EI =F 2 30d
dx 3 x=L
This procedure leads to same expression for y(x) and the same maximum deflection as before.
Shear Stress in Beams

The theory of shear stresses in rectangular beams was developed by D.J. Jourawski while designing
wooden bridges (1844–1850) for the St. Petersburg–Moscow railroad line [5]. Jourawski recognized
that wood beams are weak in shear along fibers, and these shear stresses cannot be disregarded in
the design of wooden structures. He developed the mathematical formula that predicts shear stress
magnitude across beams and showed the maximum shear stress occurs at the neutral axis.
Shear stresses in most beam applications are small compared with bending stresses and are usu-
ally ignored. However, shear stress in wood structures can be critical. Trees often fail by shear in a
severe storm as manifested by gaps along center planes of the trunk or limbs. The theory behind
stress magnitudes and distribution can be found in Refs. [2, 5]. This theory is based on the early
work of Jourawski.
Shear stress at any location in the cross section of a beam is determined by
VQ
τ= 2 31
It
Beam Analysis 49
where
V – shear force, lb
I – cross-sectional area moment of inertia, in.4
Q – moment of the area outside of the plane of interest, in.3
For a rectangular cross section (Figure 2.11)

1
Q = yA = c + y c−y t 2 32
2
This equation shows that the maximum shear stress is located at the neutral axis.
After substitution, shear stress distribution in a rectangular cross section is
V 2
τ= c − y2 2 33
2I
The direction of this shear is shown in the left drawing; one component is horizontal, the other
transverse. Since
2
t 2c
I=
12
Vc2 3
τmax = = = 1 5 τave 2 34
2I 2A
Consider a 10 ft beam, simply supported with a 1000 lb load applied in the middle; its cross
section is 2 by 4 with the 4 side vertical.
Mc
σ max = = 5623 psi
I
3 V 3 500
τmax = = = 93 7 psi
2 A 2 8
Jourawski’s equation will now be applied to the “H” beam (not a standard beam) shown in
Figure 2.12. Numbers were selection for convenience of calculation.
y c
τ y
τmax
τave
Figure 2.11 Shear stress distribution.
1/4 6 in.
4 32.75
1/4
Q
t
4
1.03 24.5
Figure 2.12 Shear distribution across an H beam.
Flange Web
τ1 = VQ V Q1
It = I t1 τ2 = VQ V Q2
It = I t 2
1
Q1 = 4 125 4 × 6 = 6 1875 in 3 Q2 = 6.1875 in.3
t1 = 6 in. t2 = 0.25 in.
Q1 6 1875 Q2
t1 = 6 = 1 03 t 2 = 24 5
Neutral Axis
Qc = 4 125 6 × 0 25 + 2 4 × 0 25 = 8 187
t = 0 25
Qc
= 32 75
tc
A direct way to determine the cross-sectional moment of inertia is illustrated in Figure 2.13.
3 3
6 85 2 875 8
Ic = −2 = 61 73 in 4
12 12
The cross-sectional moment of inertia with respect to the neutral axis is I = 61.73 in.4 Assuming
the shear load, V = 1000 lb, the maximum shear stress is
1000
τmax = 32 75 = 530 5 psi
61 73
Beam Analysis 51
Figure 2.13 Format of cross-sectional moment of

6 in.
inertia.
27/8 in.
8.5 in.
8 in.
Example Consider a beam fabricated of two metal bars welded together as shown. Assume this
beam is to be used to support the load as shown in Figure 2.14. If plate material has a yield strength
of 75 000 psi and the weld material has a weld strength of 80 000 psi, determine the factor of safety
for the beam. Which dictates the overall FS, shear in the weld or bending in the beam?
Bending Stress
bh3 443
I= = = 21 33 in 4
12 12
Mc 48 000 2
σ= = = 4571 psi
I 21 33
75 000
FS = = 16 4
4571
2000 lb
2 ft
4 in.
4 in.
Weld
1/2 in.
Figure 2.14 Welded beam of rectangular bars.

Shear Stress
Q = 1 2 4 = 8 in 3
VQ 2000 8
τ= = = 750 psi
It 21 33 1
Shear strength of weld material is

τyld = 0 577σ yld = 0 577 80 000 = 46 160 psi
46 160
FS = = 61 5
750
Composite Cross Sections

A special application of the beam shear formula is in the design of attachments such as flanges and
straps. These additions may be added to increase stiffness or to construct a symmetrical cross sec-
tion, such as the composite channel beam. These additions can be attached by nails, bolts, glue,
welds, etc. as illustrated in Figure 2.15.
The design issue in each case is how they are fastened (nail, bolt, weld) and how they are spaced
to withstand the horizontal shear force. Buckling of straps may also be a consideration. The steps for
designing each fastener type are:
1) Determine the shear force in the joint or connection. This is done by establishing the maximum
shear stress at the fastener interface based on the maximum shear force as from the shear
diagram.
V max QJ
τJ = 2 35
IbJ
2) Determine the joint force, FJ, to be transmitted by the fastener.
V J = τ J AJ
AJ = bJ L
(a) (b) (c)
Figure 2.15 Composite beam cross sections.

Beam Analysis 53
F J = bJ τJ L 2 36
3) Design of connectors
a) Nail fastener (Figure 2.15a)

Fn – shear force limit per nail
FN N = FJ L 2 37
N FJ
= number of nails per length
L FN
b) Bolted connection (Figure 2.15b)

FB – limit force of bolt in shear
2F B N = F J L 2 38
2N FJ
= double bolt spacing
L FB
c) Welded connection assuming continuous weld along strip (Figure 2.15c)
τW – shear strength of weld
2τweld tL = F J L
FJ
t= weld thickness 2 39
2τweld
Example Consider a special beam made by bolting two 2 × 6 in. boards onto a 2 × 4 in. board to
form a special wood H beam (Figure 2.16). Assume this beam is simply supported and carries a
1000 lb at the center of a 12 ft long span. Determine the spacing for lag bolts if each bolt can support
a 500 lb in shear. Cross-section moment of inertia of the beam is
Figure 2.16 Wood H-beam.

Lag bolts
6
683 243
I= −2 = 256 − 21 3 = 234 7 in 4
12 12
QJ = 3 12 = 36 in 3
VQJ 500 36
τJ = = = 38 3 psi
It 234 7 2
Joint force over length L is
F J = τbL
F J = 38 3 2 L = 76 6L
Required number of lag bolts is
500N = 76 7L
L 500
= = 6 52 in spacing per bolt
N 76 7
or one bolt per 6.5 in.

A check on the maximum bending stress under the 1000 lb load is
M max c 500 72 4
σ max = = = 614 psi
I 234 7
which is within the limits of most wood.
Material Selection
One might ask, why not use the highest grade of material to minimize weight and cost? There are
two reasons: (i) brittle fracture during service and (ii) manufacturing. High grades of steel are sus-
ceptible to embrittlement at low temperatures and stress corrosion cracking. They are also more
difficult to machine and weld. As a rule, 80 000 psi yield is assumed to be the transition between
mild steel and high strength steel. Structural steel has yield strength of about 55 000 psi. This level of
strength allows for easier machining (drilling) and welding.
Mechanical Properties of Steel

Early forms of iron were weak in tension but strong in compression. This is due to a high level of
carbon and impurities during the smelting process. Charcoal (a derivative of hard wood) was first
used to reduce impurities. Coke (a derivative of coal) was developed by Abraham Darby in 1709.
This was significant because there is a greater supply of coal than hard wood. His coke also pro-
duced a good quality of cast iron.
One of the early uses of cast iron was in the construction of the Severn Bridge in 1779. The bridge
was designed to follow the shape of masonry arch bridges. The supporting arches were made of cast
Figure 2.17 Stress–strain relation.

σ
σyld
σpl
σel
ɛ
ɛpl 2000 μ
iron. The arched supporting frames were shaped by castings and required no machining or welding.
This structure marked the end of masonry arch bridges. Cast iron components performed well
because they were put in compression. This bridge is still in use.
Over time, steel became a higher quality of iron containing controlled amounts of carbon and
other chemicals to achieve a wide range of material strengths able to perform under high levels
of stress both in tension and compression. Low-strength steel or mild steel is ductile as shown
in the stress–strain test diagram in Figure 2.17. The key parameters in this diagram are:
σ pl – proportional limit
σ el – elastic limit
σ yld – yield strength
σ pl – strain at proportional limit
The relation between stress and strain is linear up to a point. Stress and strain are related by the
modulus of elasticity, E, defined by the slope of the straight line. Key points in the stress–strain
diagram are plastic limit, elastic limit, yield strength, ultimate strength, and fracture strength.
The plastic limit, elastic limit, and yield strength are approximately the same number. Yield
strength is often considered the point of material failure.
Strain is an important parameter and strain levels matter, too. Consider steel having a propor-
tional limit of 60 000 psi. At this stress level, corresponding strain is
σ pl 60 000
ε= = = 0 002 in in
E 30 000 000
or 2000 μ strain, a good reference number for a limit strain. Actual operating strains should be much
lower. If stress increases past the proportional limit, plastic elongation occurs leaving a permanent
set in the material after the load is removed. The removal path is parallel to, but offset from, the
loading path.
σyld
High strength
80 ksi
Low strength
2000 μ
Cast iron
Figure 2.18 Comparison of stress–stain.
High-strength steels do not display significant elongation prior to separation. In this case, the
yield strength is defined as the stress level producing a plastic set of 0.002 strain (2000 μ) in the
material after the load has been removed (Figure 2.18).
High-strength steel exhibits little or no pronounced elongation prior to failure. There is no dis-
tinct stress plateau at the yield point. Yield strength for high strength is arbitrarily defined by the
stress level that produces a 0.002 (or 2000 μ) permanent or plastic set after the load has been
removed.
The relation between stress and strain for uniaxial loading is
σ = Eε 2 40
τ = Gγ 2 41
where the modulus of elasticity, E, is determined experimentally. The modulus of rigidity, G, is a
calculated value.
E
G= 2 42
2 1+ν
Assuming a material has a modulus of elasticity of E = 30 × 106 psi and Poisson’s ratio of ν = 0.28,
its modulus of rigidity is G = 11.72 × 106 psi (Table 2.2).
Shear stress yield is also a calculated value. It is not a measured value.
τyld = 0 577σ yld 2 43
This formula is based on the Mohr’s energy of distortion criteria of yielding. Calculated shear
yield values using this equation agree with test results.
Table 2.2 Material properties.
Yield strength (ksi) Ultimate strength (ksi)

a
Axial Shear Axial Shear
Wrought iron 30 17.3 48 25

Structural steel 36 20.8 66
Cast iron (gray) 110 (C), 20(T) 41
Cast iron (malleable) 32 18.5 50
Stainless steel (304) 31 17.9 73
Steel, SAE 4340, heat treated 132 76.2 150 95
Aluminum (wrought) 41 24 62 38
Douglas fir (air dry) 8.1 1.1
Red oak (air dry) 8.4 1.8
a
The shear yield stress values are determined by τyld = 0.577σ yld, based on von Mises criteria for yielding.
Use of Stress–Strain Relationship in a Simple Truss
Example Consider a simple truss made of three members supported at fixed points A and B
(Figure 2.19). A side load of 10 kip is applied horizontally to point C. It is desired to find the
defection of point C.
Inputs to this problem are the material physical and dimensional properties of members AC
and BC.
y
B
A
2
45° 30°
q
10 kips x
C
n̂B n̂A
Figure 2.19 Deflections in simple truss.

Member AC Member BC
Aluminum (1) Steel (2)

E1 = 10 000 ksi E2 = 29 000 ksi
L1 = 10 ft L2 = 15 ft
A1 = 0.326 in.2 A2 = 0.508 in.2
The concurrent force system at point C gives
Fy = 0 2 44
F BC cos 30 + F AC cos 45 = 0 2 45
F AC = − 1 22F BC 2 46
Fx = 0 2 47
F BC sin 30 + 10 − F AC sin 45 = 0 2 48
By substitution
F BC = − 7 34 kip compression
Therefore, FAC = 8.95 kip (tension). These forces produce elongations in members AC and BC.
F AC LAC 8 95 10 12
δAC = = = 0 329 in extension 2 49
EAC AAC 10 000 0 326
F BC LBC 7 34 15 12
δBC = = = 0 0898 in compression 2 50
EBC ABC 29 000 0 508
These axial deflections are compatible with internal forces of each member.
The movement of point C is defined by the displacement vector, q, as shown in the figure. Unit
vectors in line with both members are also shown.
nA = 0 707i − 0 707j 2 51
nB = − 0 5i − 0 866j 2 52
The elongation of each member is related to the displacement vector, q, by
δB = nB q = − 0 5i − 0 866j ui + vj
− 0 0898 = − 0 5u − 0 866v
and
δA = nA q = 0 707i − 0 707j ui + vj
0 329 = 0 707u − 0 707v
Algebraic solution is
0 0635 = 0 3535u + 0 6123v
− 0 1645 = − 0 3535u + 0 353v
− 0 101 = 0 966v
− 0 101
v= = − 0 105 in
0 966
0 0898 − 0 866 − 0 105

u= = 0 362 in
05
q = 0 362i − 0 105j
Point C moves downward 0.105 in. and to the right by 0.362 in.
Statically Indeterminate Member
Example Consider a rigid bar pinned at the left end and supported by two flexible cables attached
at points A and B (Figure 2.20). The problem is to find the tension forces in both cables when a force,
F, is applied at the right end. There are three unknown forces involved, the vertical reaction at point
0 and tension forces in both cables, with only two equations of equilibrium. Therefore, cable
deflections must be considered.
The equations of statics give
M0 = 0 2 53
aT A + 2aT B − 3aF = 0 2 54
T A + 2T B = 3F 2 55
and
Fy = 0
T A + T B = F + R0 assumes R0 pulls downward 2 56
Deflection equations yield
T A LA T B LB
δA = δB = 2 57
EAA EAB
Also
δA = aθ δB = 2aθ so δB = 2δA
Giving
T B LB T A LA
=2 2 58
EAB EAA
0 a A a B a
δA
R0
δB
L
Figure 2.20 Pivoted rod with elastic cables.

We now have two statics equations and one deflection equation to solve for TA, TB, R0.
By substitution, Eq. (2.55) gives
AB LA
T A = 3F − 2 TA 2 59
AA LB
AB LA
TA 1 + 2 = 3F
AA LB
and by Eq. (2.58)
AB L A
TB = 2 TA
AA L B
Finally, the reaction at point 0 is
R0 = T A + T B − F 2 60
Example Consider a pressurized tube with caps held in place by a tight wire with diameter of
0.1 in. (Figure 2.21). The pressure tube has an inside diameter of 2 in. and a wall thickness of
0.04 in. The wire is tightened to 500 lb before internal pressure is applied. The 500 lb is captured
by the wedge. The problem is to determine the internal pressure that would cause leakage (contact
force between the cap and tube is zero).
The equilibrium equation for the freebody before the applied pressure is
F = T before pressure is applied
Since it is given that initial tension in the wire is 500 lb, the initial compression in the tube is also
500 lb.
When internal pressure is applied, the equation of statics becomes
T =P+F 2 61
Cap
p
Wire
Tube, F
Wire, T
Tube
Wedge
Figure 2.21 Pressure tube.

where
T = T + ΔT (new wire tension)

P = pAcap (force against cap)
F = F − ΔF (new compression in tube)
The problem now is statically indeterminate. By substitution

T + ΔT = pAcap + F − ΔF 2 62
Since T = F = 500 lb
ΔT = pAcap − ΔF 2 63
When pressure is applied, the change in position of both wire and tube is the same,
ΔδT = Δδw subscript T refers to wire tension, w refers to tube wall compression
or
ΔTL ΔFL AT Aw
= , ΔT = ΔF, and ΔF = ΔT
EAT EAw Aw AT
So
AT
ΔF = pAcap − ΔF
Aw
AT Aw
ΔF 1 + = pAcap and ΔT 1 + = pAcap 2 64
Aw AT
The areas are
π 2
AT = 01 = 0 0079 in 2 wire
4
AW = π2 0 04 = 0 2513 in 2 tube wall
π 2
Acap = 2 = 3 1416 in 2 cap
4
The ΔF and ΔT values are
pAcap 3 1416
ΔF = =p = 3 046p 2 65
1 0314 1 0314
pAcap 3 1416
ΔT = = p = 0 0958p 2 66
32 81 32 81
Therefore, the forces in the tube and wire after pressure is applied are
F = 500 − 3 046p
T = 500 + 0 0958p
At separation, F = 0, so
3 046p = 500
pcr = 164 15 psi 2 67
and
T = 500 + 15 7 = 515 7 lb
As a check, the tension in the wire should be due only to the pressure against the cap or
T = pAcap = 164 15 3 1416 = 515 7 lb 2 68
Modes of Failure
Material Yielding
Material yielding is a state of stress–strain where the material stretches with a small increase in load
leading to permanent distortion in a material. Yield strength is a common failure criterion in
design. Once yielding is reached, plastic deformation occurs, and low cycle fatigue is possible.
The criteria of failure commonly used to determine material yielding is the von Mises criteria based
on energy of distortion [3, 6]. For a biaxial state of stress this criterion shows, yielding occurs when
σ 2A − σ A σ B + σ 2B ≥ σ 2yld 2 69
where σ A and σ B are local principal stresses. Brittle materials behave differently. Failure occurs sud-
denly without yield. The Coulomb–Mohr Criteria [2, 6] is commonly used in this case.
Stress Concentration
Classical means of calculating stress (bending, torsion, pressure, axial, and combined stress) are
adequate in most cases to configure a reliable design considering factors of safety. However, in con-
sideration of safety and cost, such as in aerospace and drilling equipment, it is desirable to refine
stress calculations in critical areas of a design using finite element software or stress concentration
factors that have been established analytically or experimentally.
Consider, for example, a small hole in the center of a long strap or plate (Figure 2.22). Using the
theory of elasticity, Seely and Smith [7] show that local stress across a strap are defined by
σmax Figure 2.22 Plate with hole.
2a
σ0
2.5
2
Stress ratio
1.5
0.5
0
0 0.2 0.4 0.6 0.8 1
Distance ratio
Figure 2.23 Stress distribution away from the hole.
σ0 a2 a4
σx = 2+ 2 +3 4 2 70
2 r r
This formula has been confirmed by strain measurements and photoelastic method.
Equation (2.70) is plotted in Figure 2.23 to show how local stress varies across the width of the
plate starting at the edge of the hole and moving to the outside edge of the plate. The parameters in
the drawing are
σx
Stress ratio 2 71
σ0
a
Distance ratio 2 72
r
When r = a, the distance ratio is one and the stress ratio is 3, meaning that local stress at the hole is
three times the nominal stress, σ 0. Also, the distance ratio reduces to zero, meaning distance, r, is
very large and local stress approaches the nominal stress. The maximum stress is located at the edge
of the hole where σσmax
0
= 3. The stress concentration factor is 3 as σ max = 3σ 0.
Stress concentration factors for other geometric shapes have been established theoretically and
experimentally [7]. Always avoid sharp geometric changes, such as fillets, openings, and threads, as
they are points of stress concentration.
Wear
Mechanical wear is the result of friction between two sliding surfaces. Wear can be minimized or
even eliminated by proper bearing design and lubrication.
Fatigue
Metal fatigue results from cyclic stresses defined by
σ = σ m + σ A cos ωt 2 73
where
σ m – mean stress
σ A – stress amplitude
ω – frequency of stress cycle
Material failure depends on magnitude of both σ m and σ A. When σ m = 0, there is complete rever-
sal of stress. Endurance limits are determined experimentally under this condition. Tests are con-
ducted on multiple specimens, each being cycles at different stress levels of σ A until failure occurs.
The number of cycles to failure is documented at each stress level. The endurance limit is typically
defined as the stress level at which complete reversals do not cause failure beyond 106 cycles.
Test data for steel show that endurance limits depend on the ultimate strength. As a guide endur-
ance limit for steel having ultimate strength less than 200 ksi, assuming complete reversal, is
0.4–0.55 times ultimate strength.
In general, time-dependent stresses cycle about a mean or average stress level. In this case, the
Goodman diagram (Figure 2.24) is useful for determining fatigue limits for stress levels having a
nonzero mean stress. Mean strength is limited by yield strength, σ yld, of the material.
The Goodman diagram is constructed by marking the endurance limit on both the plus (a) and
minus (b) directions on the vertical axis. The mean stress is zero and consistent with complete stress
reversal. The ultimate strength sets the maximum possible value for the mean stress and amplitude
stress coordinates. This point is represented by point (c). Straight lines are drawn from a to c and b to
c. The lines establish the fatigue limit for the material a different mean stress. Yield strength limits
fatigue stress. The two lines a–c and c–b establish the fatigue limits for different mean stress levels.
Amplitudes and mean stress levels near yield strength levels usually lead to low cycle fatigue.
Cyclic fatigue limits in the compression range are insensitive to mean stress [6].
σult c
σyld
ss
x stre
Ma
Tension
a tres
s
ns
ea
M
s
es
str
in
M
σyld σmean
σult
Compression
Figure 2.24 Goodman diagram.

Example Two 10-in. steel bars (1.5 in. diameter) are connected by a special 90 coupling
(Figure 2.25). The steel has an ultimate strength of 120 ksi and a yield strength of 100 ksi. Equal
and opposite forces, F, are applied as shown but vary periodically from a maximum value of
300 lb to a minimum value of 100 lb.
(a) Determine the factor of safety in member A from failure by fatigue.
(b) Determine the factor of safety in member B from failure by fatigue.
(c) Determine the factor of safety if force, F, at the elbow is removed.
For this example, we assume
Se 0 5 Sult and Sult ≺200 ksi
Accordingly, the endurance limit is σ end = 60 ksi for this material based on complete reversal of
stress.
Cross-sectional properties of the bent rod are
π 4
I= 0 75 = 0 2485 in 4
4
J = 2I = 2 0 2485 = 0 497 in 4
Case a
The critical location in member A is at the elbow, which supports bending. Since the force, F, varies
with time, bending stress will also vary between the following two stress levels. For the stated load
condition, the stress varies between
M max c 300 10 0 75
σ max = = = 9054 psi 2 74
I 0 2485
and a minimum value of
M min c 100 10 0 75
σ min = = = 3018 psi 2 75
I 0 2485
over one cycle of periodic loading. The mean stress for this cycle is
1
σ mean = σ max + σ min = 6036 psi
2
The amplitude of the stress variation is
σ max − σ min
σ amp = = 3018 psi 2 76
2
Figure 2.25 Fatigue of analysis of bent rod. 10 in.
Member B
F
F
Member A
σamp
σend = 60 ksi
a
σmean
0
σult = 120 ksi
Figure 2.26 Modified Goodman diagram for normal stress.
These stresses are shown in the modified Goodman diagram (Figure 2.26) by the heavy dot on the
positive sloped line.
0c
The factor of safety is FS = 0a . This ratio is established as follows. Let the positive sloped line be
defined by
y = m0 x 2 77
and the negative slope line be defined by
y = − mx + b 2 78
The intersection of these lines defines the coordinates of point “c.” The “y” coordinate of the inter-
section point is
b
yc = 2 79
1 + mm0
In terms of the modified Goodman diagram for member A
b = 60 ksi
60
m= =05
120
3 018
m0 = =05
6 036
Giving
60
yc = = 30
1+1
and a factor of safety of
30
FSA = = 9 94
3 018
Following the same procedure for member B
FSB = 11 47
τamp
τend = 34.62 ksi
τmean
0
τult = 69.2 ksi
Figure 2.27 Modified Goodman diagram for shear fatigue.
Case b
Von Mises energy criteria of failure show that in the static stress case
τyld = 0 577σ yld static load criteria of failure
τyld = 57 7 ksi
Experiments show that endurance limits in complete stress reversal in shear also follow the same
criteria [8].
τend = 0 577σ end 2 80
For our example, the endurance limit in shear is

τend = 0 577 60 = 34 62 ksi complete reversal of stress
The modified Goodman diagram for shear is shown in Figure 2.27.
The cyclic shear stress levels in member B vary between
T max c 3000 0 75
τmax = = = 4527 psi 2 81
J 0 497
T min c 1000 0 75
τmin = = = 1509 psi 2 82
J 0 497
Using these numbers
τmean = 3018 psi

τamp = 1509 psi
These numbers define the location of the shear cycle by the heavy dot on the positive sloped line
in Figure 2.27.
Case c
A two-dimensional stress condition arises at the support if force, F, at the elbow is removed. The
bending stress in member A stays the same; however, member B experiences bending as well as
torsion creating normal and shear stresses at the support.
We return to von Mises criteria of static failure, which for the static case, yielding occurs when
σ von σ yld 2 83
where
σ 2von = σ 21 − σ 1 σ 2 + σ 22
and σ 1 and σ 2 are principal stresses.
1 σ 2
σ 1,2 = σ± + τ2 2 84
2 2
Applying von Mises’s criteria of failure to the particular state of stress shown above, material
yielding occurs when
σ 2yld = σ 2 + 3τ2 2 85
Experiments show this formula also applies to this particular two-dimensional stress state under
completely reversed cyclic stress levels provided [8]
σ 2end = σ 2 + 3τ2 2 86
where σ end is the endurance limit under uniaxial cyclic stress condition. This equation is rewritten
in the form of
2 2
σ τ
+3 =1
σ end σ end
σ τ
which is then potted as shown in Figure 2.28, where Ratio σ = σ end and Ratio τ = σ end .
If shear stress is zero, the limiting normal stress is the endurance limit. If the normal stress is zero,
the limiting shear stress level is
1
τ= σ end = 0 577σ end 2 87
3
0.7 Figure 2.28 Endurance limits for biaxial

0.6 stress state.
0.5
Ratio (τ)
0.4
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1
Ratio (σ)
Stress Corrosion Cracking

Three factors must exist simultaneously for stress corrosion cracking to occur: (i) corrosive envi-
ronment (such as the sea), (ii) local tension stress (could be residual stress caused by forming),
and (iii) hard metal (base metal or produced by quenching). The corrosive environment causes
irregularity in the surface to form and deepen a crack. Localized tension encourages crack growth.
Hard metal cannot blunt a crack and therefore is susceptible to sudden crack growth and brittle
fracture.
This type of failure can occur in offshore pipelines. In one case, a failure analysis showed that an
elbow in the pipeline had been formed by heating it to high temperatures and then quenched for
convenience of handling. In the process of forming the elbow, residual tension stresses were left in
the hardened material. Local tension stresses could also have been generated during pipeline
installation.
Brittle Fracture
Fracture mechanics deals with stresses at the tip of a crack. The crack could be visible at the surface
or internal to the structural member. The major concern is a sudden, catastrophic failure – one that
occurs without warning.
The Griffin theory [9] of brittle fracture states there are paths of weakness within the matrix of
materials where cracks propagate. His work was based on experimental testing of glass. He showed
that cracks form even though crystals within glass are much stronger than the nominal strength
of glass.
There are two approaches for predicting brittle failure:
1) Experimental – Charpy V-Notch tests are conducted on several test pieces, having the same an
initial crack, to determine impact energy levels required to cause the crack to suddenly extend
and separate the material. Tests are conducted at different temperatures to determine a materi-
al’s ability to support a crack. There is typically a temperature at which impact energy for failure
substantially drops (Figure 2.29).
This is the transition temperature, an important property of materials expected to operate at
low temperatures. High-strength materials (σ yld ≥ 80 000 psi) are particularly vulnerable to this
type of failure, especially at low temperatures.
2) Theoretical – Stress predictions are based on the theory of elasticity, which predicts stresses at
the end of an elliptical opening [6]. The elliptical opening can achieve the appearance of a crack
Figure 2.29 Charpy V-notch response.

Absorbed energy
Transition
temperature
Temperature
Figure 2.30 Critical crack length.
σy
σy
σy a a r
(see Figure 2.30). Stress predictions show that normal stresses away from the tip of a crack are
defined by
σ a
σy = 2 88
2r
This equation shows the normal stress is ∞ when r = 0. In many cases, failure does not occur in
the form of a crack propagation. Ductile materials can create a plastic region at the tip of the crack,
and thus blunt the crack inhibiting sudden propagation. Even in high-strength materials, a crack
may be stable. To quantify this phenomenon, the flowing equation is commonly used to determine
critical crack length.
K0
σy = K 0 = σ πa 2 89
2πr
where K0 is the stress intensity factor. A crack propagates suddenly when K0 = Kcr, a material con-
stant determined experimentally.
Fluid Flow Through Pipe
Fluid mechanics is fundamental to many machine components and production systems. The build-
ing blocks for fluid analysis are:
•• Continuity of fluid flow

Energy equation commonly known as Bernoulli’s equation
• Force analysis, including impulse momentum
Continuity of Fluid Flow

Continuity of flow is based on the conservation of mass, which requires that mass entering a control
volume is equal to mass leaving the control volume (Figure 2.31).
Qρ in = Qρ out 2 90
where
Q – flow rate (volume/time)

ρ – mass per volume
Em
2
V2
1
V1
z2
z1
Datum
Figure 2.31 Control volume – open flow.
If fluid is assumed incompressible, the volume flow rates are the same and
Q in = Q out = constant 2 91
VA in = VA out
where
V – average velocity
A – cross-sectional area of flow
Bernoulli’s Energy Equation (First Law)

The energy equation of fluid flow, assuming frictionless flow, is defined by the Bernoulli equation
V 21 p V2 p
+ 1 + y1 + Ein = 2 + 2 + y2 + h f 2 92
2g γ 2g γ
where each term has units of feet. Fluid friction losses are defined by the friction head, hf.
Each term can also be viewed as having units of ft-lb/lb. Ein is energy per lb (ft-lb/lb) being put
into the control volume between locations 1 and 2. This formula is used to determine pump power
required to move fluid from one location to another.
Reynolds Number
Reynolds number is a dimensionless number defined by
VD ρ VD γ
NR = = 2 93
μ μg
It can be viewed as the ratio of fluid kinetic energy divided by shear stress at the tubular surface.
This number helps identify whether flow is laminar or turbulent. If kinetic energy is high and wall
friction is low, flow is turbulent. On the other hand, if kinetic energy is low and wall friction is high,
flow is laminar. Experiments on Newtonian fluids show that flow ceases to be laminar when Rey-
nolds number is around 2000 and turbulent flow is fully developed when Reynolds number is
around 4000.
Friction Head for Laminar Flow

The friction head for laminar flow in pipe is determined experimentally using the Bernoulli equa-
tion assuming, V1 = V2 and the pipe is horizontal.
Δp
hf = 2 94
γ
From Hagen–Poiseuille
32 μ
Δp = L V 2 95
D2
By substitution
L 32 μ
hf = V 2 96
D γD
Noting that
VDρ
NR =
μ
and rearranging terms defines friction head for laminar pipe flow in terms of Reynolds gives
64 L V2
hf = 2 97
NR D 2g
Friction head is typically expressed by the Darcy–Weisbach empirical equation
L V2
hf = f 2 98
D 2g
This formulation shows the friction factor for laminar flow is
64
f =
NR
Note that this equation was derived mathematically.
Turbulent Flow Through Pipe

In the case of turbulent flow, the factor, f, depends on Reynolds number and pipe surface
roughness.
e
f = f N R, 2 99
D
This factor is determined experimentally and is given in the Moody diagram [10], which shows
that the friction factor follows laminar flow theory up to a Reynolds number of 2000. There is a
transition region between Reynolds numbers of 2000 and 4000. Above 4000 surface roughness is
a factor. Data show that at high values of e/D, friction factor becomes less dependent on the Rey-
nolds number.
Example Determine the pressure gradient along a pipe with the following flow conditions.
Pipe size – 4½ in. (16.6 lb/ft) (drill pipe)

ID – 3.826 in. (area = 11.497 in.2)
Flow rate – 400 gpm
Viscosity – 1 cp (water)
Density (γ) – 62.4 lb/ft3 (water)
Converting to uniform units:
ft3 gal 1 ft3

Q = 400 = 53 48 ft3 min
min min 7 48 gal
ft ft3 1 min 1 144 in 2
V = 53 48 = 11 16 fps
s min 60 s 11 5 in 2
1 ft2
Since
1 cp = 1 45 × 10 − 7 reyn
lb-s
The viscosity of water in units of ft2
is
lb-s 144 in 2 lb-s

μ = 1 45 × 10 − 7 = 208 8 2
in 2 ft2 ft
Pressure drop predictions for turbulent flow (Newtonian fluid) are determined as follows.
VDρ VDγ
NR = = = 3 43 × 105
μ μg
f = 0 038 Moody diagram, assuming e D = 0 01
L V2
hf = f Darcy – Weisbach
D 2g
p1 p2
− = h f Bernoulli equation
γ γ
Combining last two equations gives
Δp γ V2
=f 2 100
L D 2g
Substitution of all input data gives
Δp 62 4 11 162 lb ft2
= 0 038 12 = 14 383
L 3 826 2 32 2 ft
Δp lb ft2 1 ft2
= 14 383 = 0 1 psi ft
L ft 144 in 2
For a 10 000 ft pipe length, the pressure loss due to fluid flow would be 1000 psi.
System losses, though important, are only part of the design of hydraulic systems. In some cases,
the goal is to maximize the use of available mechanical/pump power available, such as in the
hydraulics of drilling fluids in oil well drilling. Positive displacement-type pumps may deliver as
much as 800 hydraulic horsepower (HHP) to the drilling fluid system. Here, drill pipe may be
two to three miles in length (5000–15 000 ft) and friction losses are significant. Drill bit nozzles
are selected to use up all HHP available at the drill bit for cleaning and removing cuttings. The
hydraulic system may need to power downhole drilling motors, another demand on available HHP.
Senior Capstone Design Project

A senior capstone design project at the University of Tennessee required a hydraulic system
designed to settle dust in an indoor equestrian arena. The size of the area is 140 ft by 240 ft. Spe-
cifications required that runoff of rainwater from the roof be stored and used instead of city water.
The student team identified four subsystems.
1) Collection – gutter, return line, storage tank

2) Delivery system – piping, pumps, and nozzles
3) Controls – valves, supplemental water well
Central to the final design was pump selection.
Pump Selection
The basis for choosing a pump is: (i) pressure at the nozzle to create enough velocity to propel a jet
of water capable of reaching across the length of the zone to be watered and (ii) enough flow rate to
saturate the zone with an appropriate amount of water per time (gpm). Both pump pressure (p, psi)
and flow rate (Q, gpm) define the performance requirements of the pump.
The two pump types considered were: positive displacement and impeller. As the following anal-
ysis shows, positive displacement-type pumps are best suited for this application because of better
flow rate control. However, because of costs, an impeller-type pump was less expensive and became
the focus of the analysis.
Required Nozzle Velocity

The initial step is to determine the exit velocity required from the nozzle to reach the full extent of
the zone. The path of a fluid projectile is defined by
gx 2
y x = x tan θ − 2 2 101
2 V cos θ
Two locations where y(x) is zero are x = 0 and x = L.
V2
L= sin 2θ 2 102
g
Assuming L = 81 ft and θ = 23 , then the required exit velocity from the nozzle is
Lg 81 32 2
V2 = = = 3626
sin 2θ sin 46
V = 60 2 ft s
Nozzle Pressure
Now turn to Bernoulli’s equation to determine nozzle pressure required to produce the require exit
velocity.
Applying Bernoulli’s energy equation between locations 1 and 2 gives
p1 V2 p V2
+ 1 = 2 + 2 units are in ft 2 103
γ 2g γ 2g
Since V1 is small by comparison with V2, it is dropped leaving
γ 2
p1 − p2 = Δp = V 2 104
2g 2
If the following units are chosen
p – psi
V – fps
γ – ppg
then Eq. (2.104) becomes
γ lb gal 7 48 gal 1 ft2 2

Δp lb in 2 = V ft s
2g 32 2 ft s2 ft3 144 in 2
7 48γ
Δp = V2 2 105
2 32 2 144 2
γV 2
Δp = p1 = p = 2 106
1240
Applying a nozzle coefficient of 0.95, V2 is replaced according to V = 0.95 V2, giving
γV 2
p= 2 107
1120
This equation shows that a nozzle pressure of
2
8 342 60 2
p= = 27 psi density of fresh water is 8 342 ppg
1120
is required to produce the 60.2 fps nozzle velocity.
Pump Flow Rate Requirement

Flow rate through the system, including the nozzle is constant. Exit velocity from the nozzle is
expressed by
VAn = Q 2 108
Accounting for units
1 ft2 1 ft3 1 min
V ft s A in 2 2
= Q gal min
144 in 7 48 gal 60 s
0 32Q
V= fps
An
2
γV 2 γ 0 32Q
p= =
1120 1120 An
giving
γQ2
p= plotted in Figure 2 32
10938A2n
where
Δp – psi
Q – gpm
γ – ppg (8.342 ppg, fresh water)
An – nozzle area (total flow area, TFA), in.2
which relates flow rate to the required pressure. Using the designated parameters,
An = 0.093 in.2 (diameter is 11 32 in.)

p = 27 psi
γ = 8.342 ppg
Then
10 938An 2 p 10 938 0 093 2 27

Q2 = = = 306 2
γ 8 342
Q = 17 5 gpm
The required flow rate for two nozzles is Q = 35 gpm with nozzle pressure of still 27 psi. In sum-
mary, a pump must deliver:
One zone: p = 27 psi at a flow rate of Q = 17.5 gpm

Two zones: p = 27 psi at a flow rate of Q = 35 gpm
Figure 2.32 shows the broader response of the 11 32 nozzle to different flow rates.
The diagram shows that the selected 1.5 hp pump was adequate to supply water to two nozzles at
the required delivery pressure and flow rate.
45
1 nozzle 2 nozzles
40
35
30 Pump performance
Pressure, psi
25
20
15
10
5 Friction losses
0
0 10 20 30 40 50 60 70 80
Flow rate, gpm
Figure 2.32 Performance of impeller-type pump.

Vibration Considerations
There is much to be learned from the theory of undamped single degree-of-freedom (SDOF) sys-
tems. Many mechanical vibration problems can be explained and resolved from this model. This is a
good place to start. However, in high-speed machinery, multi-degrees of freedom may need to be
considered.
Consider the single degree of freedom system of Figure 2.33. The elastic restoring force is the
horizontal beam, the mass is a combination of the discrete mass and part of the mass of the beam.
A simple harmonic is applied to the discrete mass to cause a forced vibration.
F t = F 0 cos ωt 2 109
The source of excitation is usually related to rotating components, such as the power source or
moving parts in the power transmission sequence. The frequency of the excitation is represented by
ω, which is different from the natural frequency ωn.
The equation of motion for the mass, m, is
mx + kx = F 0 cos t 2 110
The general solution to this equation has particular and complementary solutions. Considering
only the particular solution (the complementary solution decays with time), the response is
x t = X cos ωt 2 111
By substitution
F0
X= 2 112
k − m ω2
δst
X= 2 113
1 − r2
ω
r= frequency ratio 2 114
ωn
F0
δst = 2 115
k
F(t)
EI
L
2 in.
l
δ ½ in.
Figure 2.33 Forced vibration of SDF system.

15
10
Low side High side
5
Amplitude ratio, X/δ
0
0 0.5 1 1.5 2 2.5
–5
–10
–15
Frequency ratio, r
Figure 2.34 Frequency response of SDOF systems.
The response has the same frequency as the forcing function; however, the vibration amplitude
depends on the frequency ratio, r. The response amplitude is shown in Figure 2.34.
Two possible frequencies giving an amplitude of one (1) for δXst are r = 0 and r = 2.
Several practical observations can be made from this diagram. First and most important, large
vibration amplitudes occur when ω = ωn (r = 1). Under this condition, vibration displacements
and stress levels are largest; noise levels are usually high, too. When vibrations are excessive in
machinery, the cause is nearly always, frequency tuning or resonance. The solution to this problem
is simple – detune the system. This can be done by changing either the natural frequency of the
system or changing the driving frequency.
1) Change omega
2) Change mass (usually increase)
3) Change stiffness (usually decrease)
Note that on the low side of resonance, the amplitude ratio, δXst reaches 1.26 at a frequency ratio of
r = 0.455. This means that the dynamic effects could be analyzed by a quasi-static approach
with FS = 1.26.
On the high side of resonance, vibration amplitudes diminish rapidly to near zero with high
values of the frequency ratio. This indicates the best operating conditions are at high-frequency
ratios. It is shown that when r = 2, the amplitude ratio is 1 and the dynamic response is the same
as the static displacement cause by F0. The greatest reduction in vibration amplitude comes beyond
a frequency ratio of r = 2.
Another aspect of the high side of resonance is the phase change, which is reflected in the neg-
ative amplitude. Physically, the force and amplitude move in opposite directions, i.e. when the force
is applied downward, the mass is moving upward. When systems operate on the low side of reso-
nance, force and displacement are in phase.
Example Assume the numerical details of this vibrating structure (Figure 2.33) are
L = 40 in.
bh3 1
I= = in 4
12 48
E = 30 × 106 psi
W = 20 lb
The spring constant is determined from the lower diagram noting that the deflection of a canti-
lever beam is
Rℓ3 F
δ= , with R = 2 116
3EI 2
The string constant of the total beam with a centrally applied force is
F 6EI
k= = 3 2 117
δ ℓ
Taking a simplistic view of the system by letting ℓ = L
2 leads to a spring constant of
48EI 48 30 × 106 1
48
k= = = 468 7 lb in
L3 403
and a natural frequency of
1 468 7 386
fn = = 15 14 cps
2π 20
If we assume the mass is clamped over a 6-in. interval in the center, the elastic flexibility changes.
In this case ℓ = 20 − 3 = 17 in. and the spring constant changes as follows.
6EI 6 30 × 106 1
48
k= 3 = = 763 28 lb in
17 173
The mass of the beam should also be considered. The total weight of the beam is
Wt beam = volume × density
1 ft3
Wt beam = 40 in 3 490 lb ft3 = 11 34 lb
12 144 in 3
Assuming half of the beam weight contributes to total vibrating mass, natural frequency is pre-
dicted as
1 763 28 386
fn = = 11 48 cps 2 118
2π 25 67
A somewhat lower value, but more realistic, than the first prediction.
Several worthwhile points are made from this example:
1) Assuming a dynamic force amplitude of F0 = 5 lb, δst = 7635 3 = 0 0066 in Inertia effects amplify
this static displacement.
2) Reaction forces at the supports are also amplified through
RA ℓ 3 3EI
X= , RA = X 2 119
3EI ℓ3
At a dynamic amplitude of X = 0.25 in., RA = 95 lb.
3) Bending moment is greatest at the clamp.
Mc
σ= , M = RA ℓ = 95 17 = 1615 in -lb
I
1615 0 25
σ= = 19 380 psi
1 48
4) Time to reach 106 cycles
106
f n = 11 4 cps 11 48 = , T = 87 108 seconds
T
87 108
T= = 24 2 hours
3600
5) Vibration affects both noise level and fatigue.
6) If the driving force, F0, is generated by a rotating mass, then
N
f = cps, N = 60 11 48 = 689 rpm
60
The physical appearance of SDOF systems in an engineering setting may be different from the
one shown in Figure 2.33. It is up to the designer to develop the appropriate mathematical model.
Natural Frequency of SDOF Systems

Per the above analysis, frequency tuning (resonance) defines whether there will be a vibration prob-
lem. Possible resonance can be established by considering only two factors in a design: natural fre-
quency and frequency of excitation. Frequency of excitation is often related to rotary speed. Natural
frequency can be determined analytically for simple structures, but for complex structures, it is best
determined experimentally.
Engineering structures have many vibrations modes, and any one mode can be excited. The abil-
ity to calculate and measure natural frequencies becomes an important consideration in design.
Natural frequencies are determined by consider free vibration of undamped systems. To have a
vibration, there must be a spring and a mass. A simple spring–mass system (Figure 2.35) is often
used to represent and illustrate the natural frequency of SDOF systems even though the actual
mechanical arrangement may look completely different.
The spring can represent any elastic support to a mass. The equation of motion is
mx = W − k x + δst 2 120
The static displacement due to W is represented by δst. It does not appear in the dynamic analysis
because W = k δst. Both terms cancel out of the above equation leaving
mx + kx = 0 2 121
k
x+ x=0
m
x + ω2n x = 0 2 122
Figure 2.35 Freebody diagram for a SDOF

system.
k
δst
m F x(t)
where
k
ωn = rad s
m
The solution to this differential equation is
x t = A sin ωn t + B cos ωn t 2 123
Applying initial conditions, x(0) = B and x 0 = 0 gives

x t = B cos ωn t 2 124
Over one cycle of motion having a period of T seconds
ωn T = 2π 2 125
Therefore, the period of one oscillation of the vibration is
2π
T= 2 126
ωn
The natural frequency of the vibration is
1 ωn 1 k
fn = = = cps 2 127
T 2π 2π m
The natural frequency increases with structural stiffness but reduces with increase in mass. This
equation defines the natural frequency of any SDOF system regardless of its appearance. Several
different SDOF systems are shown in Figure 2.36. Natural frequencies can be determined directly
from the differential equation of motion. The coefficient to the dependent variable defines the nat-
ural circular frequency ω2n .
g
θ+ θ=0 pendulum 2 128
ℓ
ga
θ+ 2θ=0 r – radius of gyration 2 129
r
2T
x+ x = 0 mass on a tight wire 2 130
ma
a
L
θ cg
m
m m
T T EI
L
x x
Fluid level θ
k
δst R
x
0
W
Figure 2.36 Single degree of freedom systems.
T a+b
x+ x=0 tight wire with uneven spacing 2 131
m ab
3EI
y+ x = 0 cantilevered beam 2 132
mL3
g
x+ x=0 floating object 2 133
δst
2 k
θ+ θ=0 disk 2 134
3 M
The reader is encouraged to verify each equation of motion. Note that in each case, the motion is
simple harmonic and fits the elementary definition: acceleration is proportional and opposite of the
displacement. For example, the equation of motion for the rolling disk is
I 0 θ = − Rkx 2 135
Noting that
3
I0 = MR2 and x = Rθ
2
Then
2 k
θ+ θ=0
3 M
and
2 k
ω2 = 2 136
3 M
On another note, natural frequency of the spring/mass system in Figure 2.35 can also be
expressed in terms of static deflection δst. Since
W
m= and kδst = W
g
g
ωn = 2 137
δst
This equation is useful if δst is known or specified. For example, if δst = 0.25 in., when a structure
is fully loaded, then
386 in s2
ωn = = 39 29 rad s
0 25 in
and
f n = 6 25 cps
In some structural designs, such as balconies, platforms, and walkways, a maximum static deflec-
tion is specified. This specification also sets the natural frequency of the structure.
In many cases, natural frequencies are determined by setting up the differential equation of
motion. In other cases, natural frequencies are best determined experimentally.
Example Next, consider a semicircular rim pivoted at point O as shown in Figure 2.37. The
approach is to first determine its center of gravity and mass moment of inertia about the pivot point,
O. The equation of motion is the same as for a rigid body pendulum (second case in Figure 2.36).
I A θ + aMgθ = 0
aMg
ωn 2 =
IA
The objective is to relate IA and “a” to the variables describing the geometry of the semicircle.
M cg
R θ
y
Figure 2.37a Natural frequency of semicircular rim.

R
ψ
Figure 2.37b Moment of Inertia by Integration.
Location of Center of Gravity

π π
2 2
yM = 2 R sin ϕ R dϕm = 2R2 m sin ϕ dϕ 2 138

0 0
where
M – total mass
m – mass per length
π
yM = 2R2 m − cos ϕ 2
0 = 2R2 m 0 − 1 − 1 = 2R2 m 2 139
2R2 m 2R
y= = 2 140
0 5 2πRm π
From geometry of the semicircular rim
a = R−y
Moment of Inertia with Respect to Point A

Two methods will be given. The first is based on the transfer function. The second on direct
integration.
Method #1
I 0 = R2 M
I cg = I 0 − y2 M
I A = I cg + R − y 2 M 2 141
2
I A = 2RMR 1 − = 2RMa
π
Method #2 (see Figure 2.37b)
dI A = z2 μRdψ = z2 μR 2ϕ
π
4
I A = 2 2Rμ 2R sin ϕ 2 dϕ
0
π
ϕ sin 2ϕ 4
I A = 16R μ −
3
2 4 0
2
I A = 2RMR 1 − = 2RMa
π
The natural frequency of the half ring is
aMg g
ω2 = =
2RaM 2R
Springs in Series, Parallel

Mechanical springs may be arranged in series or in parallel (Figure 2.38). In the first case, the equiv-
alent spring constant is
k eq = k1 + k2 springs in parallel 2 142
In the second case

k1 k2
k eq = springs in series 2 143
k1 + k2
The spring constant of the beam
F 3EI
k1 = = 3 beam 2 144
δ L
EA
k2 = rod 2 145
L
GJ
kθ = torsion bar 2 146
L
Coil springs have much lower spring constants than rods in extension or twist.
(a) (b)
L m L
EI EI
k k
Figure 2.38 Springs in parallel and series. (a) Parallel. (b) Series.
R F Figure 2.39 Deflection of coiled springs.
Deflection of Coiled Springs

The linear deflection across the coiled spring is mainly to twisting in the coiled rod (Figure 2.39).
Twisting per unit circumferential length is
T ds
dθ = 2 147
GJ
Vertical movement of force, F, due to dθ is dδ = R dθ. Total axial displacement of force, F, is
2π
RT 2πR3 F
δ= Rdθ = checks with Castigliano method 2 148
GJ GJ
0
Spring constants can be expressed as

F GJ
k= = per one coil 2 149
δ 2πR3
F Gr 4
k= = per one coil
δ 4R3
F GJ
k= = per n coils 2 150
δ 2nπR3
F Gr 4
k= = per n coils
δ 4nR3
For example, if a coil is 4 in. across, made of ¼ in. wire and has 10 coils, its spring constant is
calculated to be
4
12 × 106 0 125
k= = 9 16 lb in 2 151
4 10 2 3
Free Vibration with Damping
Spring constants and masses can usually be determined directly through calculations. Damping
levels are difficult to quantify except in dampers containing fluids in laminar flow. Therefore,
damping coefficients are best determined experimentally.
Every vibrating system has a certain amount of friction, which affects response amplitude. In the
case of free vibration, dynamic behavior is predicted by
Free Vibration with Damping 87
d2 x dx
m +c + kx = 0 2 152
dt 2 dt
The damping coefficient, c, becomes an important parameter when predicting responses to free
or force vibrations. It is often difficult to calculate except in special cases. It is a parameter that is
usually determined experimentally. As in most cases, friction forces are modeled with linear damp-
ing or its equivalent. This assumption greatly simplifies the mathematical solution while giving rea-
sonable engineering results.
The solution to Eq. (2.152), assuming an underdamped system, is
x t = e − ζωn t A sin 1 − ς2 ωn t + B cos 1 − ς2 ω n t 2 153
where
ς= c
ccr , damping factor assuming under damping; ς ≺ 1
ccr = 2 km, critical damping coefficient
The first term, e − ςωn t , defines the decay of the free vibration. The other terms define the cyclic
motion.
Quantifying Damping
This section gives background equations for extracting damping factors from experimental free
vibration data.
The free vibration response is
x = e − ςωt A sin ωd t + B cos ωd t 2 154
where
ωd = 1 − ς2 ωn
k
ω = ωn =
m
The amplitude of successive vibrations is
X1 Xe − ςωt
= = eςωT d 2 155
X2 Ce − ςω t + T
giving
X1
ςωT d = ln 2 156
X2
Log decrement is defined by
X1
δ = ln 2 157
X2
or
δ = ςωT d 2 158
But
ωd T d = 2π 2 159
By substitution
2πςω 2πς
δ= = 2 160
ωd 1 − ς2
A Maker X: 3.90625 ms Y: –2.5454 V

4
V
Real
1
V
/div
–4
V
Start: 0 s Stop: 3.9961 s
Figure 2.40 Damped free vibration response of SDOF.
The damping factor, ζ, is typically small (≈0.004), so a reasonable approximation for δ is

δ = 2πζ 2 161
The damping, ζ, is determined once the log decrement has been established experimentally. The
log decrement is established from successive amplitude using Eq. (2.157).
A more accurate approach for determining the log decrement (δ) and the damping factor (ζ) is
1 X0
δ= ln 2 162
n Xn
where
X0 – amplitude at point “0”

Xn – amplitude at point “n”
n – number of cycles between points 0 and n
Using data from Figure 2.40

1 27
δ= ln = 0 0337
40 07
δ
ς= = 0 0054
2π
c
ς=
ccr
Keep in mind that the experimental value of the damping factor accounts for all energy dissipat-
ing mechanisms in the systems, such as material hysteresis, support friction, and air drag, to men-
tion a few. These friction effects have been collectively modeled as linear damping.
Critical Damping in Vibrating Bar System

In general, SDOF systems may look completely different from a simple spring mass system. In each
case, the equation of motion of free vibration system will be of the same form. Consider, for exam-
ple, the vibrating bar of Figure 2.41.
k θ
M 0
0′
a a
c
Figure 2.41 Critical damping in a rotational model.
Equation of motion is
I 0 θ + ca2 θ + ka2 θ = 0 2 163
where
1 1 2 1
I0 = ML2 = M 2a = Ma2
12 12 3
From the equation of motion, critical damping is
M
ca2 cr
=2 ka2 I 0 = 2a2 k
3
and the critical damping factor is
ca2 c
ς= = 2 164
ca cr M
2 k
3
Forced Vibration of SDOF Systems with Damping
The differential equation of motion is similar to Eq. (2.110) except a damping term has been added
(Figure 2.42).
d2 x dx
m +c + kx = F 0 cos ωt 2 165
dt 2 dt
The solution to this equation has particular and complementary components. We are interested
in only the particular solution as the complementary solution is transient and decays with time.
(a) (b)
F(t) m m
k c k c
y(t)
Figure 2.42 Single degree of freedom systems.

One approach is to assume the solution has the form of

x t = A sin ωt + B cos ωt
Substituting this expression into Eq. (2.165) yields two algebraic equations with unknowns, A
and B.
2ζrδst
A= 2 166
1 − r 2 2 + 2ζr 2
1 − r 2 δst
B= 2 167
1 − r 2 2 + 2ζr 2
where δst = F0
k, and
r – frequency ratio
ζ – damping factor
Combining these equations gives

x t = X cos ωt − ϕ 2 168
where
δst
X= 1
1 − r2 2 + 2ζr 2 2
2ζr
tan ϕ =
1 − r2
These equations are plotted in Figures 2.43 and 2.44. Notice that the maximum amplitude for
each level of damping occurs near r = 1 and is quantified by
X 1
= 2 169
δst max 2ζ
This equation shows that the damping factor (not damping coefficient) is a good indicator of the
effects of damping on response.
5
ζ = 0.1
Amplitude ratio, X/δ
3
ζ = 0.2
2
ζ = 0.4
1
0
0 1 2 3
Frequency ratio, r
Figure 2.43 Frequency response for damped – forced vibrations.

180
ζ = 0.1
150
Phase angle, degrees
ζ = 0.2
120
90
60 ζ = 0.4
30
0
0 0.5 1 1.5 2 2.5 3
Frequency ratio, r
Figure 2.44 Phase angle between force and displacement.
ωt
F0 F0
4
Amplitued, lb, in.
2
ϕ X
Amplitude,
0
–1 –0.5 0 0.5 1
–2
–4
–6
Time, s
Figure 2.45 Phase angle between force and response.
Several conclusions can be made from this analysis. The amplitude of vibration is greatest when
the driving frequency, ω, is the same as the natural frequency of the SDOF system. Under resonant
conditions bad things, such as material fatigue and noise, can happen. One way to alleviate vibra-
tion amplitude at resonance is to operate the system to the far right of resonance. Another approach
is to apply damping, but this adds cost.
The relation between force and displacement functions is represented in a vector diagram
(Figure 2.45). This figure is based on the following data.
ω = 10 rad/s (T = 0.629 seconds)

ϕ = 45 (phase angle)
F0 = 5 lb
X = 1 in.
The two vectors, F0 and X, on the left side are separated by the fixed angle, ϕ. As the vector pair
rotate with angular velocity, ω, their projections onto the vertical line define amplitudes vs. time.
F t = F 0 cos ωt
x t = X cos ωt − ϕ
The phase lag of X is explained in terms of energy balance between energy input by the force and
the energy dissipated by the damper. The differential work done by F over distance dx is
dx
dW = Fdx = F dt = F t V t dt 2 170
dt
The velocity of the mass is defined by
x t = V t = − Xω sin ωt − ϕ
Equation (2.170) shows that energy is put into a vibration when the applied force is in phase with
the velocity of the vibration. When pushing a child in a swing, a force applied at the bottom of the
motion, where the velocity is the greatest, puts the most energy into the back-and-forth motion.
By substitution
T
W = − F0X sin ωt − ϕ cos ωt ω dt 2 171

0
The integration over time, T, gives the work per cycle done by F(t). The amount of work required
to overcome the energy dissipated by the damper is expressed by
W cycle = πF 0 X sin ϕ 2 172
At resonance (r = 1), the phase angle becomes 90 so maximum work per cycle becomes
W cycle = πF 0 X 2 173
This is the amount of energy required to sustain the vibration at resonance.

From the frequency response diagram (Figure 2.43 and Eq. (2.169)), the vibration amplitude, X,
depends only on the damping factor and δst.
X 1
= 2 174
δst 2ζ
which converts to
F0
X= 2 175
cωn
Using this expression in Eq. (2.173) gives the work per cycle at resonance for a linearly damped
system.
W = πcωn X 2 2 176
Nonlinear Damping
This equation is commonly used to determine the equivalent damping of other types of damping,
which may be nonlinear. For example, if a damping force is defined by
F d = ± ax 2 2 177
and x(t) is approximated by x(t) = X cos ωt, x t = − ωX sin ωt, then the work per cycle of this
damping force is
T
dx
W = 2 ax 2 dt 2 178
dt
0
T
W = 2 aX 3 ω3 sin 3 ωt dt
0
8
W = a ω2 X 3 2 179
3
The equivalent linear damping coefficient is determined by equating work per cycles.
8 2 3
πceq ωX 2 = aω X 2 180
3
Giving
8
ceq = aωX 2 181
3π
The amplitude at resonance is predicted by
F 0 3π
X=
ω 8aωX
3πF 0
X= 2 182
8aω2
Vibration Control
Basic steps that can be taken to reduce severe vibrations are:
1) Eliminate the source of excitation.

2) Change the frequency of the force of excitation (usually linked to rotary speed).
3) Reduce the magnitude of the driving force.
4) Change the natural frequency by altering k or m. It is preferable to reduce natural frequency by
increasing m or reducing stiffness. This puts the frequency ratio, r, on the high side of resonance
where response amplitudes are lowest.
5) Apply damping
a) Some damping is inherent.

b) Apply dashpots (mechanical dampers). Damping is the costliest and least desirable approach.
Other Vibration Considerations

Transmissibility
Transmissibility is defined as the ratio of force transmitted to a foundation to the driving force. Total
transmitted force is the sum of force transmitted through the spring and dampener. One is displace-
ment dependent; the other is velocity dependent. Both vary with time. The total force transmitted to
the support (foundation, floor, etc.) can be greater than the magnitude of the driving force. This is of
practical concern in situations where nearby equipment may be affected.
The equation of motion for this case is generated from the freebody diagram given in
Figure 2.46a.
The transmitted force varies with time according to
F T t = kx + cx 2 183
Transmissibility is expressed as [11]
FT t
Tr =
F0
1
2 2
1 + 2ζr
Tr = 1 2 184
2 2
1 − r2 + 2ζr 2
Transmissibility defines how much of the driving force, F0, gets transmitted to the foundation. It
is significantly affected by damping and frequency ratio. A plot of this function is given in
Figure 2.47. Notice that damping reduces transmissibility up to a frequency ratio of 2. Beyond
this point, damping increases magnitude of the transmitted force.
In an industrial setting, it is common practice to install air conditioning units on roofs. Flat space
is available. Reciprocating and rotating components in these units may generate forces that can
cause structural-borne noise, which can be eliminated by properly mounting these units on soft
springs. This greatly reduces the transmitted force.
Example Consider a compressor unit weighing 200 lb mounted on a support frame, which
weighs 50 lb. An unbalanced rotating mass produces a vertical periodic force on the frame. Its rotat-
ing speed is 250 rpm producing a variable force of 20 lb. Determine the stiffness of the four spring
(a) (b) Figure 2.46 (a) Transmissibility. (b) Vibration

Isolation.
F = F0 cos ωt y = y0 cos ωt
m x
m x
kx cx k(x–y) c(x–y)
m
5
ζ = 0.1
Transmissibility, Isolation
3 ζ = 0.2
2
ζ = 0.3
0
0 1 2 3
Frequency ratio, r
Figure 2.47 Transmissibility vs. frequency.
mounts such that the magnitude of the dynamic force transmitted to the roof is 10% of the mag-
nitude of the driving force. Assume damping is negligible.
Applying Eq. (2.184) with zero damping in the springs gives
1
− Tr = minus because vibration on high side of resonance 2 185
1 − r2
0 1 1 − r2 = 1
ω
r = 3 32 =
ωn
2πN
ω=
60
2π 250
ωn = = 7 89 rad s
3 32 60
Recall
K 250 lb-s2
ωn = , where M =
M 386 in
giving, K = 40.3 lb/in. Therefore, each spring would have a spring constant of k = 10.1 lb/in. The
initial compression of each spring would be 6.2 in.
Vibration Isolation
Base motion is illustrated in Figure 2.46b. Response is similar to the transmissibility theory except
in this case the base is moving, and force is applied to the mass through the motions of the spring
and damper.
The differential equation of motion for base motion is
mx = − c x − y − k x − y 2 186
d2 x dx dy
m +c + kx = c + ky 2 187
dt 2 dt dt
The response ratio is
1
2 2
X 1 + 2ζr
= 1 2 188
y0 1 − r2 2
+ 2ζr 2 2
This equation defines how much of base motion gets transmitted to the mass, m.
Commonality of Responses
Notice that the equations for transmissibility and vibration isolation are the same. Therefore,
Figure 2.47 applies to both. Resonance still occurs when the two frequencies are tuned. Damping
has a slightly different effect. It reduces vibration amplitude up to ωωn = 2. Beyond this point,
damping is detrimental to the system.
Instruments are sometimes isolated from support frames, which may be subjected to severe vibra-
tions or shock forces. Electronic sensors within downhole drilling tools are good examples.
Application of Vibration Absorbers in Drill Collars

Vibration analyses show how drill collars respond to bit motion [12]. Drill pipe responds too, but it
is their higher modes that respond to typical rotary speeds. It is easier to resonant the drill collar
section than to develop the higher drill pipe modes. Therefore, the drill collar section is a critical
area in the drillstring to monitor [13].
Drill collar length is dictated by required weight on bit (WOB). In hard rock drilling areas, more
drill collars are carried to generate a high bit force. The irony of this practice is the natural frequency
of axial modes of vibration becomes close to frequencies of excitation associated with normal dril-
ling speeds. This situation creates frequency tuning almost automatically. The numbers for this sce-
nario are given below.
The natural frequencies of a drill collars (pinned at the bottom and free at the top) are deter-
mined from
n16 850
f na = , n = 1, 3, 5, … axial mode 2 189
4L
Note that collar size is not a factor, only length. For the first mode, n = 1 and
4212
f 1a = 2 190
L
If, for example, L = 600 ft, f1a = 7.02 cps.
Now compare this frequency to that associated with a three-roller cone bit rotating at 140 rpm.
The drill bit frequency in this case would be
3N 140
f = = = 7 cps 2 191
60 20
Actual drilling cases may vary, but the reality remains. Both natural frequency and driving fre-
quency are close enough to excite the first mode of the drill collars. The same condition could exist
for higher drill collar modes when positive displacement motors (PDMs) and turbines are used.
Drill collar response under this assumption, i.e. ignoring drill pipe participation, is shown in
Figure 2.48.
Dynamic force
Axial
x
displacement
F(t) =F0 sin ωt
Figure 2.48 Vibration mode shape of drill collars.
Natural Frequencies with Vibration Absorbers

Vibration absorbers offer a practical solution for changing natural frequency of drill collars. Drill
collar weight (length) can remain the same for a required WOB, while lowering natural frequency.
The natural frequencies of drill collars with vibra-
tion absorber is developed as follows. Starting with
the basic equation of motion for the distribu-
Drill pipe
ted model.
dX
=0
∂2 u ∂2 u AE E dx
EA 2 = m 2 , noting that c2 = = x=L
∂x ∂t m ρ
2 192
where m is mass per unit length, ρ is mass density,
and c is acoustic velocity of a compression in the col-
lars (16 850 ft/s). L
The formulation for determining natural frequen- Drill collars
cies of drill collars with a vibration absorber is given
below. The math model is shown in Figure 2.49.
∂2 u 1 ∂2 u
= 2 193
∂x 2 c2 ∂t 2 x
For the solution, assume
dX
EA = X(0)k
u x, t = X x sin ωt 2 194 dx
x=0
By substitution k
d2 X ω2
sin ωt = − X sin ωt 2 195 Figure 2.49 Math model for natural
dx 2 c2
frequencies with vibration absorber.
d2 X ω 2
2 + X=0 2 196
dx c
General solution is
ω ω
X x = A1 sin x + B1 cos x 2 197
c c
dX ω ω ω ω
= A1 cos x − B1 sin x 2 198
dx c c c c
Imposing boundary conditions
dX ω ω
=0 0 = A1 cos L − B2 sin L
dx x=L c c
and
dX ω
EA = kX 0 EA A1 = kB1
dx x=0 c
Combining these two equations gives

ω ω kL
L tan L = 2 199
c c EA
Let
kL ω
α= and β= L
EA c
Then
β tan β = α 2 200
The eigenvalue we seek is β = ωc L, which is implicit in the characteristic equation. This equation
was programed to obtain the relation between the two dimensionless numbers. The relationship α
and β is shown in Figure 2.50 and Table 2.3.
1.4
1.2 Second mode

First mode
α parameter, dimensionless
0.8
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
Eigenvalue, β = ωL/c
Figure 2.50 Eigenvalue of first mode with vibration absorber.

Table 2.3 Eigenvalues for first and second modes.
First mode Second mode
β α β α
0.453 785 0.221 326

0.488 692 0.259 842
0.523 598 0.302 299
0.558 505 0.348 992 3.176 497 0.110 917
0.593 411 0.400 261 3.211 403 0.224 554
0.628 318 0.456 499 3.246 31 0.341 192
0.663 225 0.518 167 3.281 216 0.461 136
0.698 131 0.585 801 3.316 123 0.584 712
0.733 038 0.660 029 3.351 029 0.712 273
0.767 944 0.741 594 3.385 936 0.844 198
0.802 851 0.831 375 3.420 842 0.980 9
0.837 757 0.930 422 3.455 749 1.122 83
0.872 664 1.039 999 3.490 656 1.270 483
Example Determine the natural frequency of the first and second modes (axial mode) of a drill
collar 6½ × 2 (102 lb/ft), 700 ft long for the following two cases:
a) No vibration absorber
b) A vibration absorber having a spring constant of 50 000 lb/in.
c) No vibration absorber
n16 800
fn = , n = 1, 3, 5
4L
16 850
f1 = = 6 cps first mode 2 201
4 700
3 16 850 12 638
f2 = = = 18 05 cps second mode 2 202
4L 700
d) With vibration absorber (k = 50 000 lb/in.)
ω ω kL
L tan L =
c c EA
kL ω
α= and β= L
EA c
β tan β = α
By substitution
kL 50 000 700 12
α= = = 0 467 dimensionless
EA 30 × 106 30
From tables
ω 16 850
β 0 628L = 0 628 ω1 = 0 628 = 15 12 rad s
First mode: c 700
f 1 = 2 41 cps
ω 16 850
β 3 28 L = 3 28 ω2 = 3 28 = 78 95 rad s
Second mode: c 700
78 95
f n2 = = 12 57 cps
2π
These natural frequencies are in the range of frequencies generated by roller cone rock bits asso-
ciated with rotational speeds drill collars and drilling motors. Frequencies from this type of drill bit
relate to rotary speed by
3N PDM N PDM
f PDM = = cps 2 203
60 20
N PDM = 20 12 57 = 251 4 rpm 2 204
This rotational speed is within the range of PDM motors.
It is worth noting that the natural frequency of the first mode using the lumped mass model is
1 k
f n1 = = 2 61 cps drill collars plus vibration absorber
2π M
where
k = 50 000 lb/in.
W = 700(102) = 71 400 lb
71 400 lb-s2 lb-s2
M= = 185 45
385 in in
The distributed model gave a natural frequency of 2.4 cps.
The vibration absorber reduced the natural frequency of the first mode from 6 to 2.6 cps.
This means typical rotary speeds are on the high side of resonance when vibration dampers where
vibration response is lowest.
Responses to Nonperiodic Forces
So far, forces of excitation have been periodic. However, there are many cases where applied forces
are not periodic. In this case, the solution is based on the principle of impulse momentum.
Consider the mass in Figure 2.51 is impacted by a force, F, over time, Δt.
This impulsive force brings about a sudden change in momentum according to
FΔt = m v2 − v1 2 205
At t = 0, the initial conditions are x = x0 and v = v0. After the force impact, the velocity at point 2
becomes
Im
v2 = + v0 2 206
m
where “Im” is impulse (FΔt). This increase in velocity is achieved without movement of the mass.
The response can now be viewed as a free vibration having an initial velocity of v2. After incorpor-
ating the initial conditions, the mass responds by
Figure 2.51 Impulse of an impact force. F
m F
dt
Figure 2.52 Response to impulse. F
dIm
t
dt
dx
τ t–τ
t
v0 I
x t = x 0 cos ωn t + sin ωn t + sin ωn t 2 207
ωn mωn
“I” in this case represents the differential impulse, Im. The first two terms show the effects of
initial conditions. The third term represents the effect of the impulse. In the case of damping
v0 + ζωn x 0 I
x t = e − ζωn t x 0 cos ωd t + sin ωd t + e − ζωn t sin ωd t 2 208
ωd m ωd
The effects of the initial conditions decay with time, leaving the effect of the impulse (last term).
The impulse–momentum concept is useful for obtaining the response of SDOF systems to con-
tinuous force such as shown in Figure 2.52. The response at any time, t, is the result of consecutive
differential impulses. A displacement response due to impulse, “I,” is also shown in the figure.
The total displacement at time t is the sum of all differential responses twisted together or convo-
luted together.
Assuming there is no damping, the total effect of each differential impulse, d Im = F(τ)dτ
becomes
dIm
dx = sin ωn t − τ 2 209
mωn
Fdτ
dx = sin ωn t − τ 2 210
m ωn
t
1
x t = F τ sin ωn t − τ dτ 2 211
m ωn
0
The integral can be viewed as the sum of several sine waves stacked side by side in a convoluted
fashion, thus the term convolution integral. It is also called Duhamel’s integral.
Dynamic Load Factor

One useful application is the response to a suddenly applied force, which remains constant with
time as shown in Figure 2.53.
Duhamel’s integral for this case gives
1 F0 t
x t = cos ωn t − τ 2 212
mωn ωn 0
or
F0
x t = 1 − cos ωn t 2 213
k
This equation can also be developed directly from Newton’s second law by solving
mx + kx = F 0
The maximum displacement of the mass occurs when ωnt = 2π at which
2F 0
x max = 2 214
k
F0
Figure 2.53 Suddenly applied constant force.

This result implies the maximum displacement in the spring (or

structure) is twice the displacement produced by F0 acting as a static m
force. This gives justification for a factor of safety of 2, which is
often used to account for dynamic effects on structures. k
Packaging h
Another practical example of this theory is packaging delicate
items. Assume that mass, m, is the item to be protected and is
Figure 2.54 Packaging model.
packed inside a box containing cushion material having spring con-
stant of k (Figure 2.54).
When the box strikes the floor, the cushion material begins to
compress. The velocity of the box and contents at point of contact is
x0 = 2hg 2 215
The constant force, F0, becomes, mg, the weight of the packaged item. The equation of motion
becomes
mx + kx = W 2 216
The solution to the problem can be found by solving the equation directly or using Duhamel’s
integral. Solving Eq. (2.216) directly and using initial conditions at first point of contact of the spring
x 0 =0
x 0 = 2hg
The complementary solution is
x c = A cos ωn t + B sin ωn t
The particular solution is
W
xp =
k
Applying the first initial condition
W
A= −
k
The second initial condition gives
2hg
B=
ωn
Bringing the terms together gives
2hg W
x t = sin ωn t + 1 − cos ωn t 2 217
ωn k
The first term is the response caused by the velocity of the mass during impact. The second term is
the response due to sudden application of the weight of the mass. This term is small compared with
the first term.
4
Displacement, in.
0
0 0.05 0.1 0.15 0.2
–2
–4
–6
Time, s
Figure 2.55 Time response of mass into packing material.
Example Assume the following conditions.
h = 24 in.
W = 0.5 lb
k = 1 lb/in.
05
m = Wg = 386 = 0 001 295 lb-s2 in
ωn = k
m = 1
0 001 295 = 27 78 rad s
f = 4.422 cps
T = 0.226 seconds
The displacement of the mass (m) into the packing (k), due only to the velocity at impact, reaches
a maximum value of 4.88 in. The corresponding impact force on the instrument is therefore 4.88 lb.
The dashed line in Figure 2.55 is a continuation of the sine function. It has no physical meaning
except to indicate separation of the mass from the packaging material.
Duhamel’s integral applies to any mechanical system. Consider a compound wheel given a tor-
que, which varies with time (Figure 2.56). One can determine the angular response of the wheel in
terms of the variables (M, Ig, T, t1, t2, k) after the couple (T) is removed at t2.
k Figure 2.56 Application to a rolling disk.
θ
r2
cg r1
0
T
T
T0
t
t1 t2
Since the acceleration of point 0 passes through cg, the equation of motion is
M0 = I0α 2 218
2
T − k r1 + r2 θ = I 0θ
I 0θ + k r1 + r2 2θ = T t 2 219
with
2
k r1 − r2
ω2n = 2 220
I0
The mathematical form is the same as for a simple spring–mass system. In terms of angular
parameters
t
1
θ t = T τ sin ωn t − τ dτ 2 221
I 0 ωn
0
Specific to this example

t1 t2
1 T0
θ t = τ sin ωn t − τ dτ + T 0 sin ωn t − τ dτ 2 222
I 0 ωn t1
0 t1
where T(t) beyond t2 is zero.
Vibrations Caused by Rotor Imbalance
Response to an Imbalanced Rotating Mass

A common source of excitation in machinery is rotational imbalance. The diagram of Figure 2.57
shows a model of an SDOF system driven by imbalance, em.
The acceleration of the rotating mass is
am = ao + am o 2 223
The x component of the acceleration of the rotating mass is

ωt
d2 x m
amx = − e ω2 cos ωt 2 224
dt 2 M–m
The x component of the force required to accelerate the mass, e
m, is
d2 x
fx = m − e ω2 cos ωt 2 225 k
dt 2 c
There is an equal but opposite force on the block. From the

freebody diagram, the equation of motion for the block is Figure 2.57 Rotating imbalance.
d2 x dx
M −m = − fx −c − kx 2 226
dt 2 dt
From which
d2 x dx
M 2 +c + kx = me ω2 cos ωt 2 227
dt dt
where M is the total mass of the system.
d2 x dx
M +c + kx = me ω2 ei ωt 2 228
dt 2 dt
Using the complex variable approach, we assume the solution to be
x = Xeiωt 2 229
Substituting this assumed solution into the equation of motion gives
XM r2
= 2 230
em 1 − r 2 + i2ςr
and
XM r2
= 2 231
em 2 2 1 2
1 − r2 + 2ςr
2ςr
tan φ = 2 232
1 − r2
where
ς= c
ccr
ccr = 2Mωn
A plot of Eq. (2.231) is shown in Figure 2.58.

The time history of the response is described by
x = X cos ωt − φ 2 233
The amplitude ratio depends on frequency ratio and damping factor. These effects are illustrated
in the frequency response curves. These results show the largest amplitudes at a frequency ratio of
one (1) as expected. The damping factor, ζ, limits the maximum response amplitude per
XM 1
= at r = 1 2 234
em 2ζ
Synchronous Whirl of an Imbalanced Rotating Disk

The physical situation is one in which an unbalanced disk is mounted on a shaft (Figure 2.59). The
center of rotation of the shaft is indicated by axis a–a. The geometric center of the disk is indicated
by line 0–0, and the location of the center of gravity is indicated by line c–c. Center distances are
shown in the left diagram.
14
12
ζ = 0.005
10
ζ = 0.05
Amplitude ratio, XM/em
6
ζ = 0.1
4
ζ = 0.5
2
0
0 0.5 1 1.5 2 2.5 3
Frequency ratio, r
Figure 2.58 Frequency response of imbalanced mass system.
0 ωt
c
e c c
0 0
a a a
Figure 2.59 Synchronous whirl of disk on a rotating shaft.
Synchronous whirl occurs when a given point on the outer edge of a disk always points inward. At
any point in time, the radius, R, and eccentricity, e, are fixed and each line rotates with angular
velocity, ω. The objective is to locate radial displacements R and e.
Following the equation of motion for rigid bodies
Macg = F 2 235
ac = a0 + ac 0 2 236
where
a0 = ix + jy
ac 0 = − ω2 e i cos ωt + j sin ωt
Giving
Mx + kx = Mω2 e cos ωt 2 237

My + ky = Mω e sin ωt 2
2 238
and
R = xi + yj
k – spring constant of shaft at location of the disc
Combining the solutions gives
er 2 ω
R= , where r = frequency ratio 2 239
1 − r2 ωn
This equation is interpreted as follows (Figure 2.60). On the low side of resonance, lines R and e
are inline, with the center of gravity, c, located outside of point O. Displacement, R, grows with
increase in the frequency ratio, r. On the high side of resonance, the position of c moves inside
arm, R as the distance R diminishes with increased frequency ratio. At very high frequency ratios,
the center of gravity, c, moves closer to axis a–a.
By comparison, the imbalance rotating mass of Figure 2.57, if ζ = 0 is
XM er 2
= 2 240
m 1 − r2
15
10
Center displacement ratio, R/e
0
0 0.5 1 1.5 2 2.5
–5
–10
–15
Frequency ratio, r
Figure 2.60 Response to rotation.

Imbalance
b
Trial weight, Wt ϕ
a
0
ϕ
Counterbalance
Figure 2.61 Imbalanced rotating disk.
Balancing a Single Disk

Consider a thin disk with an unknown imbalance, i.e. the location and magnitude of the mass are
unknown. It is desirable to add a single mass at a location, which balances the disk. Regarding
Figure 2.61, the shaded circle represents the unknown imbalance.
The steps to balance the disk are [11]:
Step 1 – Run the disk at any speed resulting in a measurable amplitude. The direction and ampli-
tude are indicated by vector 0–a.
Step 2 – Attach a trial weight, Wt, at any location on the disk and measure the new displacement
vector, 0–b, at the same speed as before. The amplitude 0–b is the result of both the unknown
imbalance and the effect of the added weight, Wt. The difference between the two vectors, a–b, is
the effect of the trial weight, Wt alone.
Step 3 – Move the trial weight counterclockwise by the angle ϕ. In this new location, the displace-
ment vector, a–b, will be parallel and opposite to vector, 0–a, as indicated by the dashed line.
Next, increase its weight (Wt) by the ratio of 0a
ab. The disk will then be balanced.
Due to damping, the phase angle of vector (0a) will lag the unknown position of the imbalance.
This is the basis of auto-tire balancing machines.
Synchronous Whirl of Rotating Pipe

Synchronous whirl of rotating pipe refers to motion where the inside surface of the deflected pipe is
always pointing inward toward the axis of the support bearings. The outside surface is always point-
ing outward away from the bearing axis.
Assuming there is no tension in the pipe, the equation of motion is
d4 y
EI − m ω2 y = 0 2 241
dx 4
d4 y
− α4 y = 0 2 242
dx 4
where
m ω4
α4 =
EI
The solution to Eq. (2.241) is
y = A sin αx + B cos αx + C sinh αx + D cosh αx 2 243
Boundary conditions for a simply supported beam are:
y 0 =0 2 244a
2
dy
=0 2 244b
dx 2 x=0
y L =0 2 244c
2
dy
=0 2 244d
dx 2 x=L
Constants B, C, D are zero leaving

y = A sin αL 2 245
For a nontrivial solution
αL = nπ, n = 1, 2, 3, K
m ω2 nπ 4
= 2 246
EI L
The critical rotary speed for the first mode (n = 1) is
π 4 EI
ω2crit = 2 247
L m
Stability of Rotating Pipe under Axial Load

A similar solution is given by Den Hartog [14] for a rotating beam (pipe) subjected to compressive
end loading, Q. If the shaft rotates slowly and jarred sidewise, the elastic stiffness will restore the
shaft to its original straight position. At some higher rotational speed, the shaft, if disturbed later-
ally, will not return. Under this condition, the shaft has buckled from the rotation. This scenario is
different from the whirling of unbalanced rotating shafts, which experience radial displacement
once the shaft begins to rotate.
The differential equation of bending for this case is similar to the previous case except that the
applied end force is compressive and the side-loading inertia is due to centrifugal forces based on
synchronous whirl.
d4 y Q d2 y mω2
+ − y=0 2 248
dx 4 EI dx 2 EI
The end force, Q, is compressive and carries a (+) sign. It can also be viewed as (Qeff)ave. The nota-
tion for the speed of rotation is ω rad/s.
Lateral buckling is imminent when
2
Q Q mω2 π2
+ + = 2 2 249
2EI 2EI EI L
Equation (2.249) is put into the form
2
Q ω
+ ≥1 2 250
Qcr ωcr
where
π 2 EI
Qcr =
L2
π 2 EI
ωcr =
L m
If Q is tension instead of compression, Q changes signs and the rotary speed required to create
whirling motion by the centrifugal force increases as a result. If the left side of Eq. (2.250) is greater
than one (1), the rotating column (pipe) is dynamically unstable.
Example Consider, for example, a 60 ft span of drill collars between two stabilizers. Assuming
pinned support at each stabilizer and
OD = 7 in.
ID = 2 in.
I = 117.07 in.4
E = 29 (10)6 psi
w = 120 lb/ft
m = 3.73 slugs/ft
N = 100 rpm
ω = 10.47 rad/s; ω = 2πN
60
WOB = 50 000 lb
then
π 2 29 106 117 07
Qcr = = 64 636 lb
12 60 2
and
π 2 29 106 117 07 1 ft2

Ωcr = = 6 78 rad s
60 3 73 144 in 2
Applying Eq. (2.250)

2
50 000 10 47
+ = 0 774 + 2 38 = 3 15
64 636 6 78
Based on these numbers, the collar section between the two stabilizers is unstable and will expe-
rience synchronous whirl. The compressive force contributes to this instability, but collar rotary
speed is the main culprit.
Balancing Rotating Masses in Two Planes

Consider the rotation of the two masses mounted on a shaft (Figure 2.62). The objective here is to
place one mass in plane 0 and a second mass in plane N such that the four masses will be balanced.
Equilibrium of the four masses, including the balancing masses, must satisfy the following two
equations
N
ai xαi M i ω2 Ri = 0 2 251
0
N
α i M i ω 2 Ri = 0 2 252
0
The input data for this case are given in Table 2.4.
Substituting the numbers in Table 2.4 into Eqs. (2.251) and (2.252) gives
− 6m1 R1 + 16mN RN sin θN i + 12m2 R2 + 16mN RN cos θN j = 0 moment 2 253

m0 R0 cos θ0 + m2 R2 + mN RN cos θN i + m0 R0 sin θ0 + m1 R1 + mN RN sin θN j = 0 force
2 254
From the moment equation
1 m 1 R1
tan θN = = 0 375 2 255
2 m 2 R2
y
m1
x m2
6 4
ω
6 6 4
0 1 2 N
Figure 2.62 Unbalanced rotating mass. system.
Table 2.4 Balancing rotating masses.
Planes containing the masses
0 1 2 N
ai 0 6 12 16
αi cos θ0 i + sin θ0 j j i cos θN i + sin θN j
mi m0 10 20 mN
Ri 6 6 4 6
y y
O 6 N
x x
6
–159.4°
–118.25°
mN = 10.65
m0 = 7.08
Figure 2.63 Location of masses in balancing planes. (z axis is out of paper, right-hand rule).
The arc tan of 0.375 has multiple angles. Only θN = − 159.4 satisfies Eq. (2.252), the moment
equation. By substitution, mNRN = 63.94 leading to mN = 10.65.
From the force equation
m1 R1 + mN RN sin θN
tan θ0 = = 1 86 2 256
m2 R2 + mN RN cos θN
The arctan has multiple units; however, only θ0 = − 118.25 satisfies the force equation. By sub-
stitution, m0R0 = 42.56 leading to m0 = 7.08. The solution is shown in Figure 2.63. The units of the
balancing masses are the same as the unbalanced masses.
Refining the Design
Once the design has been configured in terms of its subsystems, it is desirable to identify critical
areas for further analysis. This may require computer software to establish stress distributions
and points of high stress, temperature, etc. Testing components with instrumentation, such as
strain gages may be useful as check on computer models, especially if there are safety concerns.
Testing of certain subsystems in isolation may also be useful.
Final steps to a final design and manufacturing are shown in Figure 1.1. Fabrication drawings
communicate the final design and how it is to be fabricated.
Manufacturing
The fabrication of each component should be considered during the early stages of design. Consid-
eration should be given to how each part is to be made and the sequencing of fabrication. Parts may
be machined from a casting, which can reduce machining time and waste, or it may be machined
from bulk stock.
The machine tool cutting of metal began with tools powered by water wheels. One of the pro-
blems that James Watt faced in manufacturing his steam engines was boring large holes in an
engine block. The difficulty in manufacturing large cylinders with tightly fitted pistons was solved
by John Wilkinson, an ironmaster in Staffordshire. In 1774, he had invented a machine, powered by
water wheels, for boring cylinders with extreme accuracy. He used this method for boring canons.
This machining technology greatly improved the quality of the Boulton–Watt steam engine. The
steam engine, coupled with the development of electric motors during the 1800s, led to better
and more accurate technology for cutting metal.
These early milling and boring machines led to stepwise improvements in metal-cutting devices,
such as milling and boring machines capable of high-cutting speeds and quality surface finishes.
Manufacturing Drawings
Communication of the design is done through group discussions, oral presentations, and mechan-
ical drawings. It is critical that mechanical drawings are accurate and explain every dimension,
including tolerances.
The geometric configuration of a design, including all subsystems and components, is established
following the procedure discussed earlier. In many cases, components, such as bearings, motors,
and springs, can be obtained from numerous vendors. On the other hand, many design components
must be fabricated. The overall assembly usually contains a combination of both. Design informa-
tion is communicated to manufacturing through fabrication drawings, which convey specific infor-
mation about dimensions, tolerances, and surface finish as well as how the various parts are to be
assembled and fastened together. Fabrication drawings may be a part of a broader engineering
report giving back-up material, engineering analysis, quality control, quality testing, and feedback
process.
Dimensioning
Dimensioning standards have been established to clarify and unify how components are to be com-
municated from design to manufacturing. Dimensions on a drawing define the linear size of a part
in terms of length, width, and thickness and various cutouts that shape the part. The American
National Standards Institute (ANSI) defines dimensioning as “a numerical value expressed in
appropriate units of measure and indicated on a drawing and in other documents along with lines,
symbols, and notes to define the size or geometric characteristic of a part or part feature.” Proper
dimensioning conveys the intent of the designer.
Dimensioning practices are established by ANSI and emphasized by ASME [15] and other
authors [16–18]. Dimensioning practices are summarized below.
1) A dimension line is a thin line that shows the extent of a dimension.

2) Extension lines provide end limits for dimension lines. They should start about 1/16 in. from
the object and extend past the last dimension line about 1/8 in.
3) Dimension lines can be stacked with the first-dimension line about 3/8 in. from the object and
the remaining lines ¼ in. apart. Overall, dimensions are placed outside smaller dimensions.
4) Dimension lines are typically not placed inside object lines.
5) The dimension value is placed at a break in the dimension line. Numbers are always read from
the bottom.
6) Dimension numbers in the US are expressed in inches. There is no need to follow the numbers
by the (“) symbol. For decimal numbers, less than 1 in., the zero is omitted, for example, 0.25 in.
7) Dimension lines should be lighter than object lines in the drawing but visible. Use narrow
arrow heads.
8) Dimensions should not be given to hidden lines, if possible.
9) Place dimension lines on the view that best conveys the shape or contour of the part to the
machinist.
2×ϕ 1.1 0.75R
0.5R
6×ϕ 0.25
Figure 2.64 Dimensioned drawing.
10) Critical dimensions should give tolerances for that dimension.

11) Never have a dimension line cross another dimension line.
12) Dimension lines are drawn from center lines where necessary. Never use a center line as a
dimension line.
13) Always give the diameter of a hole, not its radius. The symbol, , precedes hole dimensions. For
example, 0.75, indicates a hole diameter of 3/4 in. Drill sizes are expressed in decimals.
14) The symbol, R, should be placed after the value. For example, a filet radius of 1 8 in. is indicated
by 0.125R.
The drawing in Figure 2.64 illustrates a few of these dimensioning principles.
Tolerances
Design may require components to operate with relative motion in areas of contact. The quality of
both geometries could affect friction and simply how parts fit together. The performance of slider
bearings and journal bearings depends on clearance between the moving parts. Assembly of sub-
parts may also be affected by clearances.
It is impossible for a part to be manufactured exactly to a prescribed dimension. This could be due
to machine tool flexibility, tool face interaction with work piece, interface dynamics, thermal
expansion, etc. It is therefore necessary to define, in the drawing, acceptable variations from the
A C
B D
Figure 2.65 Situations requiring dimensions with tolerance.
base dimension. This allowable variation is call tolerance. Large tolerances may affect functionality.
Small tolerances will affect cost thru precise manufacturing and parts inspection (and rejection).
Acceptable tolerances depend on the functionality of the mating surfaces (Figure 2.65). They con-
trol the manufacturing process, control variation between mating parts, and allow parts to be inter-
changeable. Smaller tolerances increase manufacturing cost exponentially.
The ANSI standard [16–18] defines tolerance as “the total amount by which a specific dimension
is permitted to vary. The tolerance is the difference between the maximum and minimum limits.”
Tolerance on a dimension can be expressed in three ways.
Bilateral tolerance – Tolerance is expressed by using allowable variation in both plus (+) and minus
(−) directions, for example, 1.125 ± 0.005. In this case, the upper limit is reached when the meas-
ured value is 1.13. The lower limit is reached when the measured value is 1.25. These two limits
must allow the proper function between mating parts, such as a sliding guide or a shaft and col-
lar fit.
Unilateral tolerance – In this case, a tolerance is allowed in one direction, either plus or minus. For
example
0 005 0
A=25 or A=25
0 − 0 005
Limits of size – A dimension in this case is set between to limits, such as
2 505
A=
2 495
Three Types of Fits

The fit between mating parts, as illustrated in Figure 2.65, can be classified into three types:
a) Clearance fit – One part easily fits into another part with a clearance gap. The shaft is always
smaller than the hole.
0 595 0 600
C= D=
0 593 0 602
Tolerance on shaft is 0.002

Tolerance on hole is 0.002
Minimum clearance is 0.600 − 595 = 0.005
Maximum clearance is 0.602 − 0.593 = 0.009 in.
Tightest fit is 0.005 in.
b) Force (interference) fit – One part is force fitted into the other.
0 503 0 500
C= D=
0 502 0 501
Tolerance on shaft is 0.001
Tolerance on hole is 0.001
Minimum clearance is 0.500 − 503 = −0.003
Tightest fit is 0.003 in. interference
Maximum clearance is 0.501 − 0.502 = −0.001 (loosest fit)
c) Transition fit – The loosest case provides clearance fit and the tightest case gives an interfer-
ence fit.
0 507 0 500
C= D=
0 502 0 505
Tolerance on shaft is 0.005 in.
Tolerance on hole is 0.005 in.
Minimum clearance is 0.500 − 0.507 = −0.007
Tightest fit is 0.007 in. interference
Maximum clearance is 0.505 − 0.002 = 0.503 in.
Loosest fit is 0.003 in. clearance
Used where accuracy is important but either a clearance or interference is permitted.
Surface Finishes
Surface finish can affect the performance of certain aspects of a design. Mechanical friction and the
performance of nearly every type of bearings (as will be shown later) depend directly on surface
finish. Rolling contact bearings (ball and cylindrical) require a surface finish of 20 μin. for long-term
performance. Slider and journal bearings require a surface finish of 0.003–0.005 in. for useful serv-
ice, too.
The texture of a machine surface can be important for several reasons. The finish affects the
appearance of the part, stress concentration, friction, and bonding. Surface finishes are quantified
in terms of surface roughness, the average of the surface undulations with respect to the nominal
surface plane. Typical surface roughness for the three machining operations mentioned above are:
Turning 0.5–6 μm (15–250 μin.)

Drilling 1.5–6 μm (60–250 μin.)
Milling 1.0–6 μm (30–250 μin.)
Per profilometer measurements, the rms values of highly polished surfaces are about 0.5–1 mil-
lionth of an inch (0.0127–0.254 μm). The peak-to-valley distance is even higher.
Nanosurface Undulations
Surface undulations (Figure 2.66) have been manufactured at the nanolevel using special tools.
These undulations have amplitudes of ±5 nm with a wavelength of ~200 nm. They were manufac-
tured in a silicon surface over a 5 μm × 5 μm area.
In considering nanosize dimensions, remember
Nanoscale – atomic to 100 nm

Microscale – 100 nm to 500 μm
Macroscale – 500 μm and above
1 Å = 0.1 nm
1 nm = 10−9 m
1000 nm = 1 μm = 10−6 m
Measurements of these undulations (Figure 2.67) show and quantify the geometry of the man-
ufactured surface pattern [19]. Keep in mind that the range of atomic attractions is less than a nan-
ometer (see Figures 11.18 and 11.19). Friction between contacting nanoundulations is probably due
to a combination of molecular and mechanical effects.
Figure 2.66 Manufactured undulations in a silicon surface. Source: Courtesy of Oak Ridge National
Laboratory.
Machining Tools 119
6
4
Vertical, nm
2
0
–2 0 200 400 600 800 1000 1200
–4
–6
Horizontal, nm
Figure 2.67 Image of surface topography. Source: Courtesy of Oak Ridge National Laboratory.
Machining Tools
Three principal machining processes are turning, drilling, and milling.
Lathes
Lathes are used to machine circular parts and drill holes. In this case, the work piece rotates and the
cutting tool is stationary and fixed to a platform that can be translated longitudinally and trans-
versely. Some of the cutting operations are:
•• Facing
Taper turning
•• Contour turning
Form turning
•• Chamfering
Cutoff
•• Threading (internal and external)

Boring (enlarging an existing hole)
•• Drilling
Knurling
These operations are conducted manually.
Drill Press
Several operations are made with drill presses. A hole is typically drill first, then it can be modi-
fied by:
•• Reaming
Tapping
•• Counter boring
Counter sinking
•• Centering
Spot facing
Each modification requires a special cutter head. In this operation, the cutter head rotates, and
the work piece is stationary.
Milling Machines
These machines use a spindle, which rotates a cutting tool against a translating work piece. The
rotating cutting tool can move vertically in z-direction. The work piece can move horizontally
in either x or y direction. It is manually operated.
Milling can be peripheral (cutting takes place on the side of the cutting tool) or face (cutting takes
place on the end of the cutting tool). In either case, the main purpose of milling is to produce a flat
surface on the work piece.
A few operations of peripheral milling are:
•• Slab milling
Slotting
•• Side milling
Straddle milling
A few operations of face milling are:
•• Conventional face milling

Partial face milling
•• End milling
Profile milling
•• Pocket milling
Surface contouring
Machining Centers
Machine centers are milling machines capable of conducting many automated sequential opera-
tions, including tool changes. The spindle rotates the cutting tool and can move vertically as well
as laterally. The work piece can move horizontally in both x and y directions. The motions of the
cutting tool and work piece are controlled by a computer code (CNC). Once set up, these centers are
fast and accurate in producing a single part with complex geometries.
Turning Centers
Turning centers operate similar to milling centers except that the work piece turns at a given cutting
speed. The cutting tool attachment is fixed but can rotate as in a terete to change cutters. Pro-
grammed computer codes control tool interchange.
Turn centers can turn a cylindrical surface, facing the surface, and drilling a transverse hole.
Computer-controlled machines produce accurately machined parts consistently and at a high rate.
They can machine complex geometries faster than can be done by hand-operated machines. Setup
time should be considered, but this cost is often offset by reduction of cost per part, when large
quantities are to be made.
References 121
References
1 Ressler, S. (2011). Understanding the World’s Greatest Structures. DVD Video. The Great Courses.
2 Timoshenko, S. (1955). Strength of Materials (Parts I and II), 3e. NY: D. Van Nostrand Co. Inc.
3 Hibbler, R.C. (2011). Mechanics of Materials, 8e. Prentice Hall.
4 (1989). Manual of Steel Construction, 9e. NY: American Institute of Steel Construction.
5 Timoshenko, S.P. (1953). History of Strength of Materials. New York: McGraw-Hill.
7 Seely, F.B. and Smith, J.O. (1956). Advanced Mechanics of Materials, 2e. New York: McGraw-Hill.
8 Timoshenko, S. and Young, D.H. (1968). Elements of Strength of Materials. D. Van Nostrand Co. Inc.
(see p 330).
9 Griffith, A.A. (1920). The phenomena of rupture and flow in solids. Philos. Trans. Royal Soc.
221A: 163.
10 Moody, L.F. (1944). Friction factors for pipe flow. Trans. ASME 66: 671–684.
11 Den Hartog, J.P. (1956). Mechanical Vibrations, 4e. New York: McGraw-Hill.
12 Dareing, D.W. and Livesay, B.J. (1968). Longitudinal and angular drill-string vibration with damping.
Trans. ASME J. Eng. Ind. 90, Series B (4): 671–679.
13 Dareing, D.W. (1984). Drill collar length is a major factor in vibration control. J. Pet. Technol.:
637–644.
14 Den Hartog, J.P. (1952). Advanced Strength of Materials, 297. McGraw-Hill Book Company.
15 ASME Y14.5 and Y14.100 (2018). Dimensioning, Tolerancing, and Engineering Drawing Practice
Package.
16 Puncochar, D.E. (2011). Interpretation of Geometric Dimensioning and Tolerancing, 3e.
Industrial Press.
17 Lowell W. Foster; Geometric Dimensioning and Tolerancing Per ASME Y14.5, Self-Published, 2019.
18 Drake, P.J. (1999). Dimensioning and Tolerancing Handbook. McGraw-Hill Education.
19 Binning, G. and Rohrer, H. (1985). Scanning tunneling microscopy. Physica 127B: 37–45.
123
Part II
Power Generation, Transmission, Consumption
Every machine has three functional aspects: (i) Power supply, (ii) power transmission and (iii) end
use. Machines are designed to perform to a given set of specifications as discussed in Part I. During
the process of achieving an end use, energy is consumed as indicated in the drawing.
Power Power
generation transmission End use
Fuel
Exhaust Friction
There are several options for a power source (electric motors, gasoline engines, diesel engines, gas
turbines). Output performance of each dictates which is best for a given application. Power is trans-
mitted by means of any one of several mechanisms (gears, pulleys, linkages, power screw, hydrau-
lics) to achieve a specified end effect. Part of input energy will be lost to friction or other
inefficiencies. Part II covers these aspects of machine design. The oil well drilling rig and its five
subsystems is used to illustrate this process.
125
Power Generation
Water Wheels
The water wheel was used as a source of power for many years. The evolution of the water wheel
was driven by population growth and the need for greater food output. The Egyptian Nora
(~700 BC) was used to lift water for irrigation. The Romans milled grain during the fourth century
AD and at the time of William the Conqueror, England had about 5000 grist mills. By 1790, there
were about 2000 grist mills in colonel America. By the time of the Civil War, there were some 55 000
water wheels in use, many powered manufacturing facilities. Power generated by the water wheel
(~10 hp) was transmitted through gear trains to achieve a required output torque and speed. Grist
mills typically have a gear ratio of 25 : 1 with the millstone having the higher speed and lower tor-
que. Wooden gear teeth were common.
Early water wheels were constructed by empirical methods which evolved and were improved
with experience (Figure 3.1). They worked and that is what mattered, first as a pump and a later
as a milling machine. Performance of water wheels is explained in terms of current engineering
mechanic.
Fluid Mechanics of Water Wheels

With reference to Figure 3.2 and Bernoulli’s energy equation:
y1 = y2 + E m 31
where
y1 – elevation of water entering the wheel, ft
y2 – elevation of water leaving the wheel, ft
Em – mechanical energy imparted to the wheel shaft, ft or (ft-lb per lb of fluid)
The energy equation reduces to

ft-lb
E m = H ft = H 32
lb
where H is the drop in elevation.
The power equation is
fl-lb ft3 lb ft-lb

P=H Q γ 3 = HQγ 33
lb s ft s
126 Power Generation
Figure 3.1 Lenoir Museum water wheel (Norris, Tennessee).
Expressed in terms of horsepower
HQγ
P= hp
550
Assuming
Q = 2 ft3/s
γ = 62.4 lb/ft3
H = 12 ft
Then power delivered by the water wheel is
12 62 4 2
P= = 2 7 hp assuming 100 hydraulic efficiency
550
In terms of rotary speed (rpm) and torque (ft-lb)

TN
P= hp
5252
Steam Engines 127
Flume
Water flow
Torque
1
Speed
Diameter
Torque
Water exits
wheel
2
Figure 3.2 Water wheel mechanics.
If the wheel turns a N = 1 revolution per 8 seconds or 7.5 rpm, then output torque is
2 7 5252
T= = 1891ft − lb N = 7 5 rpm
75
Power output is adjusted by flow rate, Q, with a side gate in the flume.
Steam Engines
The steam engines of Newcomen and Watt were based on steam vacuum, created by reducing the
temperature of steam trapped in an enclosed volume. These early engines were named “atmos-
pheric engines” because atmospheric pressure pushed the piston into a vacuum. James Watt
was aware of the advantages of using positive steam pressure but was concerned about the capa-
bility of boilers to withstand positive pressure. With time, especially during the application of steam
engines to river boats, positive steam pressure became common. Steam pressure level was kept
somewhat below 200 psi to minimize the risk of explosion.
Early steam engines were limited in their ability to contain high steam pressure because of ques-
tionable quality of design, manufacturing, and materials. However, to increase power output, steam
pressure had to be elevated to much higher levels. The fear of boiler explosions was a concern, espe-
cially when steam engines were used on steamboats. This concern was realized in the Sultana
steamboat disaster of 1865.
The American Society of Mechanical Engineers (ASME) was founded in 1880 and began to
address some of these issues. The first uniform code for testing boilers was presented in 1884.
The code evolved over the next several years and was eventually published in 1915. It included
all aspects of design, construction, and testing of boiler and pressure vessels. These standards
are found in the current ASME Boiler and Pressure Vessel Code, which has been adapted across
North America and in 60 countries around the world. As a result, steam boilers and pressure vessels
in power stations safely operate near 1000 psia.
Steam Locomotives
Steam in locomotive engines reach pressures of 200 psi and temperatures near 800 F. Water boils at
382 F at this pressure (200 psia). Heat is added at constant temperature to the water/steam com-
bination until all the water has been turned to steam. Beyond this point, the steam becomes super-
saturated and the temperature of the steam increases.
The bulk of a steam-driven locomotive is the boiler. This is where the fuel, usually coal, is burned
and steam is generated. Steam energy within the boiler is conveyed to the piston/cylinder chamber
to generate the driving force. This force is converted to torque on the wheels through a slider crank
arrangement. The steam chamber is relatively small compared to the boiler, but this is where
mechanical power is generated and transmitted (Figure 3.3). Energy is transmitted to a wheel
by means of a slider crank mechanism, which converts linear motion to rotational motion.
A steam locomotive includes a pump, a furnace and boiler, steam cylinder and a condenser. The
furnace and boiler occupy most of the volume of the total steam engine. This is where fuel (coal) is
burned and the heat gets transferred into the water/steam, which is moving through a long session
of pipes (boiler). Steam comes out of the boiler and enters a steam cylinder, a relatively small device
near the wheels.
In 1880, Ephraim Shay built a prototype of a geared steam engine with steam cylinders oriented
vertically. Reciprocating motion of the piston is converted to rotary motion by a slider crank
arrangement. This orientation allowed the use of bevel gears to increase torque to the wheels.
Two bevel gears are used. The larger bevel gear is attached to the axial of the train wheels. The
smaller gear is attached to the crankshaft (Figure 3.4). This gear arrangement reduced speed but
increased torque to the wheels. This design feature allowed the train to pull a heavier load, which
Rotation
Slider valve
Exhaust
Steam pressure
Figure 3.3 Steam cylinder and transfer mechanism.

Steam Engines 129
Steam cylinders
Rotation
Bevel gear pair
Crank shaft
Figure 3.4 Shay locomotive power system.
was especially important in removing logs in a mountainous terrain. Another feature of this engine
is the use of three steam cylinders side by side.
Shay engines were used by the Little River Railroad Lumber Company in timbering the Smoky
Mountains near Townsend, Tennessee during the early 1900s (Figure 3.5). Note the steam loader in
the rear car.
The miracle of the steam engine (and steam turbine) lies in the transformation of raw fuel (wood,
coal, oil, uranium) to mechanical torque and speed. Steam is the linkage that allows this to happen.
Steam is generated external to the cylinder.
Output power from the various electric motors or mechanical engines is delivered in terms of
torque and speed. Power is expressed in units of horsepower or Watts. In horsepower units
2πTN 2πTN TN
P= = = hp 34
550 60 33 000 5252
where
P – power, hp
T – torque, ft-lb
N – rotational speed, rpm
1 hp – 550 ft-lb/s
In SI units,
P watts = T N − m ω rad s 35
P watts = Tω Joules s
Figure 3.5 Shay engine designed for steep terrain. Source: Courtesy of Little River Railroad and Lumber
Company Museum, Townsend, Tennessee.
where
P – power, W
T – torque, N-m
ω – rotational speed, rad/s
1 hp – 745.7 W
Engine performance, which is determined experimentally, varies with torque and speed. Engines
are rated at maximum output conditions. Actual operating output may be different than the rated
output.
Power Units in Isolated Locations
Electrical power is sometimes needed in remote places away from stationary power plants. Exam-
ples are drilling operations, cruises ships, and locomotives. In each case, diesel engines drive elec-
tric generators which produce electrical power. Electrical power is then delivered through electric
cables to various points of operation.
For example, in early drilling operations power was taken directly from internal combustion
engines (ICE) and transmitted through chain and gear drives to an end use. This meant the ICE
was close to operations often creating safety issues. Currently, ICEs generate electric power through
generators. In recent years, power units are located away from operations, and electric power is
delivered through cables, which are simple, reliable, safer, and cheaper. Isolated power units of this
type are found on drilling rigs, cruise lines, and diesel locomotives.
The amount of deliverable power (kilowatts) is limited by space. Typically, portable power units
are mounted on skids for easy set up and positioning. The engine room in cruise ships contains
diesel engines and electric generators, while diesel engines and electric generator occupy the front
car in railroad locomotives.
Regional Power Stations
Regional power stations are fixed and usually located near rivers. Electrical output is measured in
terms of megawatts. Output from local power plants supplies electricity regionally and can be input
to a national power grid. The source of power is steam. Since space is usually not a problem, solid
fuels, such as coal, natural gas, and uranium are options.
Physical Properties of Steam

Thermodynamic properties of steam have been tabulated in steam tables [1]. The state of the mix-
ture of water and steam is illustrated in the following figure. The bell curve shows where water
begins to boil and where steam is completely saturated (Figure 3.6).
For example, consider a certain volume of water at room temperature (70 F) and atmospheric
pressure (14.7 psia). When heat is added to water, temperature begins to increase, and the volume
expands until the water starts to boil. Water has reached its saturation temperature or boiling tem-
perature of 212 F (p = 14.7 psia). If heat is continually added at a constant pressure, more and more
steam is formed, until all the water is converted to steam (saturated vapor). This expansion takes
place without an increase in temperature. Beyond the saturated vapor point, steam temperature
increases with added heat and becomes superheated steam.
If the heat is added at a higher pressure, say 100 psia, the saturation temperature increases. On
the other hand, if the pressure is lower than 14.7 psia, the saturation temperature reduces;
800
700
600
Temperature (ºF)
500 p = 420 psia Saturated vapor

400 Saturated liquid
300
p = 14.7 psia
200
100
0
0 0.5 1 1.5 2 2.5
Specific volume (ft3/lb)
Figure 3.6 Steam properties – (p, v, T) diagram.

therefore, it is easier to boil water at higher elevations. Desalination systems boil seawater in a vac-
uum to distill fresh water at a lower boiling temperature.
Energy Extraction from Steam

Liquid/steam changes take place continuously throughout the steam cycle of a steam engine. The
mechanical components of a steam engine, such as, boiler, piston/cylinder (or turbine), condenser,
and pump control the state of the steam. The whole process is one of energy conversions, starting
with combustion of raw fuel (coal), and ultimately converting heat into mechanical work
through steam.
First Law of Thermodynamics – Enthalpy

For a non-flow system, the first law is expressed as
W
Q− = u2 − u 1 36
J
The energy equation for an open flow system is
p1 v1 V2 z1 W p v2 V2 z2
Q + u1 + + 1 + = + u2 + 2 + 2 + 37
J 2gJ J J J 2gJ J
Enthalpy is a state variable that includes internal energy and pressure and is useful in tracking
energy conversions throughout a thermodynamic process. It is a component of the energy equation,
a formulation of the first law.
Ignoring the elevation change and combining internal energy and pressure energy gives
V 21 W V2
Q + h1 + = + h2 + 2 38
2gJ J 2gJ
where
Q – heat supplied between station 1 and 2 Btu/lb
h = u + pvJ Btu/lb (enthalpy, a state variable)
W – shaft work leaving apparatus, ft-lb/lb
J – 778 ft-lb/Btu
V – velocity, fps
g – acceleration due to gravity, 32.2 ft/s2
Entropy – Second Law

Entropy is a state variable used to determine the amount of heat transferred into and out of a
thermodynamic process. Since heat and work are related, entropy is used to determine energy taken
from heat engines.
For a reversible process [2]:
dQ
= Constant 39
T
The constant is defined as the difference between the entropy between points 2 and 1 in the
process:
800
700
600
Temperature (ºF)
500 Saturated
liquid Saturated vapor
400 p = 420 psia
300
p = 14.7 psia
200
100
0
0 0.5 1 1.5 2 2.5
Entropy (Btu/°F/lb)
Figure 3.7 Entropy chart for steam.
2
dQ
s 2 − s1 = 3 10
T
1
Expressing steam properties in terms of (T, p, s), allows heat flow in a thermodynamic process to
be determined by
Q1 − 2 = T ds 3 11
1
This is a huge advantage in determining heat flow into and out of a steam cycle. The state of steam
expressed in terms of temperature and entropy is shown in Figure 3.7.
The efficiency of a thermodynamic steam cycle becomes
Qin − Qout
efficency = 3 12
Qin
Thermodynamics of Heat Engines

A thermodynamic cycle is one in which water/steam is taken through various physical states and
returned to its original state. An example is a thermodynamic cycle (Rankin heat cycle) shown in
Figure 3.8. The key mechanical components in this cycle are pump, boiler, turbine, and condenser.
Its efficiency is determined by output work divided by heat going into the cycle. The letters on the
rectangular flow diagram indicate the phase of the water/steam as depicted on the above steam
properties diagram. Unlike the steam locomotive, water is recycled. This process requires a con-
denser, which lowers pressure and temperature and converts steam to water for pumping. In an
ideal condenser, pressure and temperature remain constant.
Utility power stations are based on a thermodynamic cycle; they are typically located near rivers.
Water from the rivers cool the condensers and remove heat from the steam.
Wout
3 4
Turbine
Qin Qout
Boiler Condenser
Pump
2 1
Win
T
Boiler
pressure 3
1 Condenser 4
pressure
Figure 3.8 Rankins heat cycle with T–s diagram.
This process is, perhaps, the only remaining external combustion engine in use today. The exter-
nal combustion generates steam, which drives steam turbines. This process allows the use of variety
of fuels, such as coal, natural gas, and uranium.
Utility power stations are usually located near rivers, which serve as heat sinks for James Watt’s
condensers. The thermodynamic cycle in this case is the same regardless of the fuel source (coal, oil,
or uranium). Coal and oil generated heat by combustion. Nuclear power stations generate heat by
nuclear fission, caused by splitting atoms. Fission produces a huge amount of heat that is used to
heat water/steam. Combustion is not involved.
The engines or power units are steam turbines. Steam is expanded through these turbines to
develop mechanical power in terms of torque and speed. This mechanical power drives an electric
generator, which supplies electric power to the cities. The energy conversions are raw fuel, steam
heat, mechanical turbine power, and electric power. Once the electrical power gets to your home it
goes through other power transformations depending on household needs.
These liquid/steam changes are taking place continuously throughout the steam cycle. The basic
mechanical components of a steam power plant (boiler, turbine, condenser, and pump) control the
state of the steam. The whole process is one of energy conversions, starting with combustion of raw
fuel (coal), and ultimately converting heat into mechanical work through steam.
Mechanical output comes off the steam turbine shaft in the form of torque and rotary speed. Typ-
ical output power is 50 MW at about 3500 rpm.
Steam Turbines 135
Steam Turbines Steam out
Steam turbines are mechanically simpler than steam

engines. Steam turbines simply rotate, while the
motions of piston, connecting rod and crankshaft Rotor velocity
are much more complicated. The rotor moves in
response to momentum changes of the steam through
its blades. Turbine shaft speed ranges between 3000
and 15 000 rpm with an output power of 20–100
MW. Typical operating conditions are 500 psia and
Fixed blade
300 C (572 F).
The physical arrangement of steam turbines is given
in Figure 3.9. Properties of steam continuously change
as energy is removed from the steam as manifested in
pressure drop, temperature, and density. Steam den- Rotor velocity
sity directly affects momentum as steam moves across
each moving turbine blade. The thermodynamic proc-
ess of steam through the turbines is assumed to be
isentropic.
Steam turbines generate power from steam, which
inter turbines under very high pressure (~1500 psia).
Nozzle
High steam pressure is converted into velocity by
means of nozzles. Nozzles can be converging-
diverging type (Laval nozzle) or simply fixed blades
Steam in
which direct the pressurized steam onto rotor blades.
To create a gradual extraction of energy from the Figure 3.9 Principle of the steam turbine
steam, multiple stages of rotors are used to bring steam (isentropic process).
down from high pressure to a lower pressure. Pressure
reduction is stepped down over each stage in a con-
trolled and calculated manner until the entire steam pressure has been used. This staged reduction
in pressure means that there is a progressive increase in volume requiring gradual increase in tur-
bine blade diameter.
Mechanical power delivered at the output shaft of steam turbines is generated by the change of
momentum change across the rotor blades [3]. Stator blades or fixed blades redirect steam onto the
rotor blades. Steam engines of the type used on locomotives rely on steam pressure to push against a
reciprocating piston. Rotary motion is achieved by a slider crank mechanism, made up of a con-
necting rod, piston, and crankshaft. Dynamic forces in turbines are the result of rotary motion,
while dynamic forces in reciprocating engines are caused by the combination of translation and
rotation of moving parts. Therefore, the steam turbine can operate at much higher rotary speeds
than reciprocating engines.
The main driver of the turbine is energy from steam, which enters the turbine at temperatures
around 1000 F and at pressures around 900 psig. By comparison, these operating temperatures and
pressures are much higher than those (near atmospheric pressure and 220 F) used in the early
steam engines by James Watt. Steam turbines are the primary generators of electric power.
Large electric generators used by the electric utility industry today have a stationary conductor.
A permanent magnet is attached to a rotating shaft and positioned inside a stationary conducting
(a) Figure 3.10 Combustion and pressure cycle of

Fuel, air mixture otto ICE engines.
Spark plug
Flame front
(b)
Pressure
Ignition
compression Power
Exhaust
Atm
Intake
Top position Bottom position
coil made up of a long, continuous piece of a wire conductor. When the magnet rotates, it generates
a small electric current in each section of wire. Every small current of individual sections adds up to
one current of considerable size. This is the basis of the generation of electric power whether its
electricity comes from a steam power utility plant or in a hydroelectric power station.
Electric Motors
Power transfer from engines was initially made by mechanical devices, such as chain drives, pull-
eys, and gears. This often created a hazardous working environment. In recent years, engine power
is transformed into electrically. Electrical power is very transportable and safe. As a result, electric
motors are used throughout industry.
During the early 1830s, Michael Faraday made an important discovery while studying magnet-
ism. Faraday discovered that electric current and an electrical potential are generated between the
ends of a conductor when the conductor is moved perpendicular across magnetic lines. His discov-
ery laid the foundation for modern-day electrical power used in homes and industry today. The
electric generator is a vital link in the energy chain. Electricity is very transportable and much safer
than mechanical linkages. Electric power can be applied locally to machinery or household equip-
ment and is easily conveyed throughout a city or country.
Following this discovery, Faraday built the first electromagnetic generator, which produced a
small direct current (DC) voltage and a large amount of current. At first, water wheels drove electric
generators. During the second half of the 1800s, electric generators were driven by steam engines.
By the turn of the century, steam turbines became the choice of prime movers for electric generators
and still are today. The steam turbine can produce high torque and horsepower because of the sim-
plicity of motion.
There are two general types of electric motors: alternating current (AC) and DC. AC electric
motors are divided further into single phase and three phase motors.
Single-phase motors are adequate for applications requiring up to about 5 hp. They draw more
current than three-phase motors, thus making three-phase motors a more efficient choice for indus-
trial applications. Single-phase motors are typically found in home products. AC motors are simple
in design, low cost, reliable, and come in numerous sizes and performance characteristics.
Three-phase motors require smaller wiring and less voltage making them safer and less expensive
to operate. Three phase motors are lighter in weight and more efficient than comparable single-
phase motors. More power is supplied to them than to a single-phase motor over the same period.
For this reason, they are commonly used in industry and manufacturing. However, three-phase
motors and controls are more complex and expensive.
There are three types of DC motors: brush motors, brushless motors, and stepper motors. Brush
motors are the most common because they are easy to build and cost effective. Carbon brushes,
which are used to transfer electric current to the rotating part, wear over time. DC brushless motors
overcome this wear problem but are costlier and require complicated drive electronics to operate. In
applications where speed and torque need to be controlled with high accuracy, brushed DC motors
are good choices. If higher performance and reliability are required, brushless DC motors may be
preferred.
Selection of the right motor depends on application. There are multiple choices; however, the
right choice has to match design requirements. In many cases, motor power is delivered at high
speed and low torque.
Internal Combustion Engines
The difference in the steam engine and ICEs is in the location of the combustion of the fuel or
energy source. Combustion of fuel to energize steam engines, occur external to the piston/cylinder
chamber, while combustion for gasoline engines takes place inside the cylinder. The internal com-
bustion process allows for a much higher power to weight ratio in engines. The external combustion
engine allows for the use of bulk fuels, such as coal, wood, and uranium.
Four Stroke Engine

The first ICE was built and tested by Nicolaus Otto (German) in 1876. The basic cycle of his engine
(called the Otto Cycle Engine) has four strokes: air intake, fuel injection and compression, spark
ignition and expansion of burned gas, and exhaust. The injected fuel was liquid (gasoline), which
was another product of the rock oil produced by Rockefeller. The Otto Cycle Engine (internal com-
bustion) was the first alternative to the steam engine, an external combustion engine.
The internal combustion has a much lower weight to power ratio (~2 lb/hp) than do steam
engines and is more adaptable to automobiles and other vehicles. They use petroleum fuels having
higher energy content than coal. Thermal energy is released and converted into mechanical energy
during the combustion process. The liquid fuel is mixed with air in a carburetor to form a combus-
tible mixture, which is compressed by the piston within the cylinder before being ignited by an
electric spark [4]. The explosion and expansion of the air-fuel mixture moves the piston down-
ward-generating torque through a connecting rod and crankshaft (Figure 3.10).
450
400
Torque, (ft-lb)
350
300
Torque (hp)
250
200
150
Power (hp)
100
50
0
0 500 1000 1500 2000 2500 3000
Rotary speed (RPM)
Figure 3.11 Performance diagram of six cylinder SI engine.
In 1885, Gottlieb Daimler and Wilhelm Mayback improved on the Otto engine. Their patented
engine is recognized as the foundation for the modern-day gasoline engine. The same year they
adapted their engine to a stagecoach and produced the first four-wheel automobile. Also, in
1885, Karl Benz designed and built the first complete automobile. By 1900 Benz & Company
became the world’s largest maker of cars.
Several pistons and cylinders are arranged to provide uniform torque. Proper balancing of each
group of cylinders is also necessary for smooth operation of the engine.
Figure 3.11 gives the performance data for a spark ignition engine (43 8 in. diameter, 53/4 in. stroke)
having a compression ratio of 5.5 : 1. Tests were made with a wide-open throttle. The reduction of
output torque at high speeds is due to internal friction. Gear transmissions are used to increase
torque at low speed.
In a gasoline engine, the fuel–air mixture, which is premixed in a carburetor, is drawn into a cyl-
inder, then, compressed to a ratio ranging from 4 : 1 to 10 : 1 before being ignited by a spark (Spark
Ignition, SI). The spark is sufficient to produce smooth combustion and thus the engine operates
smoothly. Gasoline engines require a lighter engine frame, making it preferable for light vehicles
and other mobile applications. High compression ratios are not possible because of pre-ignition.
Two Stroke Engines

The two-stroke engine accomplishes these same four events over two strokes instead of four. Start-
ing with the downward or first stroke, ignited gasoline (the power stroke) expands, and drives the
piston downward. The piston reaches a position where an exhaust valve is exposed and allows the
expended gas to escape. Still moving downward, the piston reaches another opening where a mix-
ture of fuel and air is blown into the cylinder by a scavenging fan, removing any remaining exhaust
gases while filling the cylinder with a new gas mixture.
When the piston moves upward (beginning of the second stroke), the intake valve closes and sub-
sequently the piston blocks the exhaust opening. Over the remaining part of the second stroke, the
gas mixture is compressed and ignited at the top position by means of a spark plug.
Alternatively, a crankcase-scavenged two stroke engine does not have a scavenging fan. The
crankcase is sealed to allow the intake fuel mixture to be compresses by the bottom part of the pis-
ton. Once the fuel mixture is trapped, it is compressed within the crankcase by the downward travel
(a) (b)
Pressure Ignition
Fuel Hot air

Power
Compression
Exhaust
Atm
Intake
Interior spray Top position Bottom position
Figure 3.12 Combustion and pressure cycle of diesel engines.
of the piston. When the piston moves upward to the intake valve, the gas mixture is forced into the
cylinder for further compression and ignition. The power stroke (downward stroke) is the same as
the first engine configuration. The crankcase scavenger replaces the fan scavenger.
Diesel Engines
The diesel engine carries the name of its inventor and developer, Rudolf Diesel (German). His idea
(1890) was to achieve a much higher compression than the Otto engine such that fuel ignition is
accomplished by temperature caused by the higher compression. The four strokes are the same,
except fuel is injected at the peak of compression, not during the compression cycle.
In a diesel engine, air is drawn into the cylinder alone and then compressed to a ratio ranging
from 14 : 1 to 25 : 1. The air at these compression ratios reach temperatures ranging between 700
and 900 C (compression ignition, CI). Since diesel fuels do not mix with air, they are injected
through an atomizer into the highly compressed air where mixing takes place [4]. Fuel droplets
must be mixed with heated air before ignition (Figure 3.12). There is a time lag between fuel injec-
tion and ignition measured terms of fractions of a second. Combustion could begin anywhere in the
combustion chamber; it is an uncontrolled process.
Because of the high compression ratio and uncontrollable combustion, diesel engines require a
more rugged frame.
Engines with high compression ratios are more efficient, meaning that diesel engines are more
efficient than gasoline engine. High compression means higher torque especially at low speeds.
Because of this diesel engines are used in heavy industries, such as trucking, shipping, construction,
oil drilling, and production. A key advantage in the diesel engine or gasoline engine is the low main-
tenance cost.
Gas Turbine Engines
In some ways, gas turbines are similar steam turbines in that energy is extract from high temper-
ature and high-pressure gases to rotate stages of turbine blades. Gases are caused to impinge on
rotors to bring about mechanical power in the form of torque and speed. Gas turbines have three
key components: (i) compressor, (ii) combustion chamber, and (iii) turbine. The compressor brings
in outside air and compresses it to a pressure of about 50–75 psia. The compressed air inters a com-
bustion chamber, where liquid fuel is injected and ignited (~650 C). This mixture then passed
through a turbine at high velocity and generates mechanical power (torque and rotational speed)
[5]. This power is often used to drive electric generators. In this configuration, the turbine,
compressor, and generator are mounted on one shaft. Part of the power generated by the turbine
is used to drive the compressor and part to drive the electric generator.
John Barber, an Englishman, developed the technical basis for the current gas turbine engine
around 1790. He invented and patented a gas turbine engine in 1791. It was not until 1903 that
Ægidius Elling, a Norwegian, built a gas turbine that generated more power (11 hp) than needed
to run the engine. Sir Frank Whittle later used his work to design and build the first jet engine for
aircraft propulsion (1930). The first successful use of his jet engine was in April 1937.
An axial gas pressure gradient toward the exhaust end of the engine creates the momentum
change between the gas entering and leaving the engine. This pressure gradient is amplified in
the compression chamber by combustion of the fuel. Per Newton’s third law, there is an equal
and opposite force applied to the engine structure which pushed the airplane forward.
Gas turbines are typically used to power aircraft and in this application, they are rated by how
much thrust they produce. Velocity to lift an aircraft is a key factor. Per to the law of linear impulse-
momentum:
TΔ t = m v2 − v1 3 13
Starting from rest at the end of a runway, the thrust required to reach a lift-off velocity is
Δt
v2 = T 3 14
m
Gas turbines are also used with propellers to form turbo-prop engines. In this case, the propeller is
attached to the power shaft in front of the compressor. Power is also taken off the main shaft to drive
mechanical devices, such as pumps.
Basically, gas turbine engines pull in air and discharge a heavier gaseous mixture made up of air
and burnt jet fuel at high velocities (Figure 3.13). The result is a thrust force somewhat like the force
propelling a pressurized toy balloon that has been released. The theory behind the thrust is based on
momentum change of the gases coming into and out of the engine.
Combustion
chamber
Fuel Power Exhaust

Air intake Compressor
Figure 3.13 Components of gas turbine engines.

The forward thrust to the jet housing is equal and opposite of the resultant force, F, applied to the
gases inside the engine (Newton’s third law). The higher the mass flow rate and exit velocity the
greater the thrust. To elevate both requires higher combustion rates, which create higher gas tem-
peratures within these engines.
The basic activities inside the engine are air intake, compression by the entry turbine blades,
injection of jet fuel with simultaneous ignitions, and escape of the gas mixture at high velocities.
The temperature in the combustion chamber rises to about 3 500 F in the hottest part of the flame.
These hot gases pass through the power turbine and impart power to the main shaft, driving the
compressor blades up front. The compression power is about 60% of total power leaving the com-
pression/combustion chamber. The remaining power is used to create the thrust of propulsion or
torque on a separate drive shaft.
The compressor turbine blades and the power turbine blades near the exit are on the same shaft
and rotate at the same revolutions per minute. Rotary speeds are in the range of 20 000 rpm. By
comparison typical rotary speeds of automobile engines are 2 000 rpm.
Gas turbines are also used to power electric generators using mechanical power off the main tur-
bine shaft. Gear trains reduce shaft speed and increase output torque.
The Boeing Company developed the first commercial jet airplane in the early 1950s. The Boeing
707 was certified for commercial service in 1958. This airplane was the first of subsequent 7 × 7
model airplanes and ushered in modern commercial jet travel.
Jet engines have greatly improved over the past 50 years. The goal of higher and higher thrust to
weight ratios, drives this technology.
General Electric recently presented their new high thrust jet engine (90115 B) capable of gener-
ating thrusts forces between 115 000 and 126 000 lb1 [6]. For comparison, this level of thrust is
equivalent to three 727 airplanes using all three engines. This new engine is so powerful that just
one of them could fly a Boeing 747 jetliner. The power capability of this engine is over 100 000 hp a
quantum leap in power from James Watt’s first steam engine.
Impulse/Momentum
The thrust produced by a gas turbine is determined from impulse–momentum. Consider the con-
trol volume in Figure 3.14
Vcv
1 2
Qfuel
V1
V2
Q1
Q2
Compressor Combustion Power

chamber
Figure 3.14 Control volume within a gas turbine.

Assuming the gas in the exit nozzle expands to ambient pressure and subsonic flight and the inlet
pressure is ambient, total internal forces applied to the gases within the control volume is
F i = V 2 Q2 − V 1 Q1 3 15
where the mass flow rate of the exhaust is
Q2 = Q1 + Qfuel
Since the mass flow rate of the fuel is small by comparison with air flow, it is neglected. Also
V 1 = vi + V CV
V 2 = ve + V CV
where
VCV – velocity of control volume (engine)

V1 – absolute velocity of air entering the control volume
V2 – absolute velocity of air leaving the control volume
vi – intake velocity relative to control volume
ve – exhaust velocity relative to control volume
By substitution
F = Q ve − vi total force on gas in control volume 3 16
where vi = VCV and thrust is opposite to F [5].
Thrust = Q ve − V CV
Energy Considerations
The power generated by the turbine near the exit is determined from the energy equation. This
energy is converted to torque and rotation speed of the shaft which becomes the drive shaft for
power take-off:
V 21 W power V2
Qfuel + h1 + = + h2 + 2 3 17
2gJ J 2gJ
The energy process through gas turbines is usually assumed to be isentropic.
Engine Configurations
Turboprop engines use power off the power shaft to rotate a propeller in front of the gas turbine
engine. An initial application of the gas turbine is in jet aircraft propulsion. Gas turbines can also be
Combustion
chamber
Power Exhaust
Fuel
Compressor
Figure 3.15 Turboprop engine configuration.
Combustion
chamber
Gear box
Power Exhaust
Fuel
Compressor
Figure 3.16 Helicopter configuration.
configured to produce torque on the main shaft for driving propellers (Figure 3.15). The turboprop
engine is a slight variation of the thruster concept. The turboprop engine is used in commercial
aircraft and in helicopters (Figure 3.16). This same configuration is also used to power electric
generators.
Gas turbines are often used for pumping oil or gas from offshore production facilities, as the tur-
bine can be set to operate at optimum impeller pump flow rate speeds (Figure 3.17). Linkage
between the power drive and impeller shaft is a direct connection.
A gas turbine is started with an electric motor which caused the compressor to create pressure
within the combustion chamber. Fuel is brought in to start the turbine, which in turn begins to
power the compressor.
Fluid out
Combustion
chamber
Power Exhaust
Fuel
Impeller pump Compressor
Figure 3.17 Impeller pump configuration.
Rocket Engines
The power system includes two solid busters and pivotal main engines that use liquid hydrogen as a
fuel. Liquid oxygen accelerated the combustion process in these engines. Thrust is like the jet pro-
pulsion of airplanes. In both cases, thrust is created by momentum changes into and out a combus-
tion chamber. With the jet engine, air is brought into the combustion chamber from the
environment, compressed and combined with ignited fuel to increase the momentum of the
exhaust. With the space rocket, fuel is ignited by liquid oxygen, which is carried with the engine
during and after lift-off. Since the magnitude of thrust depends on rate of change of momentum,
higher burning rates mean higher thrusts. Higher burning rates also mean higher temperatures.
Rocketdyne F-1 Engine

High-speed turbo pumps are essential to a successful lift off. In the case of the main engine of the
space shuttle, the turbo pumps are used to delivering liquid hydrogen and oxygen to the combustion
chamber at extremely high rates (Figure 3.18). To do this, the turbo pumps must rotate at speeds
near 80 000 rpm. These speeds usher in a whole host of mechanical issues from friction to vibra-
tions. This engine is considered the most powerful engine every made.
Rocket and jet engines were just immerging as new weapons during World War II. But there were
peaceful uses too. Space scientist, Werner von Braun, and his Redstone Arsenal team were already
at work, when on 4 October 1957, the Soviet Union successfully launched Sputnik I.
On 31 January 1958, the United States launched Explorer I, which carried a scientific instrument
that led to the discovery of the James van Allen magnetic radiation belts. In July 1958, Congress
passed the Space Act, which established the National Aeronautics and Space Administration
(NASA). The space race was on, and Congress was ready to pour money into research and
education.
Atlas Booster Engine

Figure 3.19 is a photo of an Atlas Booster Engine, one of three engines used to boost the Atlas Mis-
sile. It had a thrust of 165 000 lb with a 130 seconds burn time. The first Atlas launch was 19 July
Rocket Engines 145
Figure 3.18 Rocketdyne F-1 engine. Source: NASA.
Figure 3.19 Atlas F series booster engine. Source: Courtesy of Cape Canaveral Air Force Station Museum.
1958. These engines were made by Rocketdyne. These and other rocket engines are mounted on an
adjustable linkage (center of photo) which allow for navigation control.
Gas Dynamics Within Rocket Engines

Thrust produced by a rocket engine is determined from impulse momentum principles. Following
the procedure outlined above, total impulse required to bring about total change in momentum of
all gas particles across the control volume is (See Figure 3.20)
Vcv 2
Combustion
chamber V2
Qoxygen
Q2
Qfuel
Turbo-pump
1
V1 Diverging
nozzle
Figure 3.20 de Laval nozzle.
Thrust = Qfuel ve + pe − pa Ae 3 18
where
Qfuel – total mass flow rate of oxidizer and fuel
ve – velocity of exhaust relative to engine
pe – exhaust pressure
pa – ambient pressure
The de Laval nozzle expands and accelerates combustion gases causing exhaust gases to exit the
nozzle at hypersonic velocities.
A measure of performance of these engines is a parameter called Specific Impulse having units of
“pounds of thrust per pound of propellant burning per second.” The Specific Impulse of hydrogen/
oxygen rocket engines (~415) is as high as anything we know (Albert C. Martin, personal commu-
nication). Basic to these engines is the storage of the fuel at cryogenic temperatures and the turbo
pump which delivered the fuel at rotary speeds up to 80 000 rpm. Testing before each launch was
critical to the reliability of these engines. Lyle C. Bjorn, manager of testing for North American-
Rockwell at Cape Canaveral, managed some 400 engineers and technicians in preparing these
engines for each launch. In addition to fuel storage at cryogenic temperatures, special attention
was given to turbo pump especially, the contact bearings wear at these high speeds and vibration
modes in the turbine blades. There were no launch failures during the Saturn/Apollo project during
the 1960s.
The total weight (dead weight) of the Saturn V rocket was about 6.5 million lb. The first stage
contained five (5) Rocketdyne engines each capable of generating 1.5 million lb of thrust. These
engines were fueled by jet fuel which was oxidized by liquid oxygen. The second stage also con-
tained five (5) Rocketdyne engines each capable of generating 2 million lb of thrust. These engines
were fueled by liquid hydrogen which was oxidized by liquid oxygen.
Rocket Dynamics
The equation of motion a rocket for rockets ignoring drag is
T = Qfuel ve 3 19
where
Qfuel – rate at which mass is being exhausted from nozzle

ve – velocity of exhaust relative to rocket
Newton’s second law becomes [7]

dm dV
ve = M 3 20
dt dt
where
M – mass of rocket
V – velocity of rocket
The mass of the rocket is continuously changing as fuel is burned.

M t = M 0 − μt 3 21
dM
μ=
dt
dM dm
= =μ
dt dt
Combining everything gives
dV
M 0 − μt = μve 3 22
dt
The first stage launched the rocket down range about 500 mi and achieved an altitude of about 45
mi and a speed of 7000 mph. The second stage launched the remaining part of the rocket down
range another 500 mi to an altitude of 80 mi and a speed of 17 000 mph; great enough to put the
command module in earth orbit.
It is noteworthy that the Apollo 11 (1969) occurred 200 years after James Watt patented his steam
engine (1769). This launch also occurred 300 years after Sir Isaac Newton developed calculus and
established his laws of dynamics. Both are essential to propel, navigate, and monitor space flight.
Energy Consumption in US
In conclusion, consider the dependents of the US economy on fossils fuels. Figure 3.21 shows
sources and amount of energy consumed in the United States since 1800.2 Coal was the main energy
source for smelting iron and making steel during the eighteenth and nineteenth centuries. Even
during the twentieth century, coal remained a major fuel source for steam power plants.
Oil overcame coal as a major energy source during the mid-twentieth century due mainly to the
rapid growth in automotive and trucking industries. Around 1950 steam locomotives through my
small hometown of Miami, Oklahoma disappeared. They were replaced by diesel engines. At this
time oil began to outpace coal as indicated.
45
40
35
30
Quadrillion (Btu)
Oil
25
20
15
Coal
10
Natural gas
5
Wood
0
1800 1850 1900 1950 2000
Year
Figure 3.21 Energy consumption in the US since 1800. Source: Data from Energy Consumption in the US by
Source, 1800-2000 (Quadrillion Btu), U.S. Energy Information Administration.
The petroleum industry responded during the 1960s with research in drilling and production at
the same time the space program was also growing. Oil exploration and production expanded into
remote areas, especially offshore. Two technologies accelerated this expansion.
1) Geophysics greatly improved seismic mapping of promising formations through digital compu-
ters. Improvement came in quality and speed of mapping especially offshore.
2) Directional drilling allowed accurate navigation through multiple formations to reach complex
reservoir structures deep within the earth.
Lateral horizontal drilling into production zones along with hydraulic fracking greatly increased
oil production and oil recovery.
Industry also responded to the growing population by expanding production of goods and ser-
vices. The technology for engines and electric motors was in place along with mechanisms and con-
trols to improve manufacturing.
Currently nearly all electricity is generated by steam turbines. The technology and thermo
sciences are fully developed to optimize power output. Steam is the linkage between fuel (such
as coal and uranium) and electric power.
Unfortunately, both external and internal combustion engines produce carbon dioxide which
threatens the earth’s environment and climate. New energy sources will be needed to supplement
and perhaps eventually replace fossil fuels.
Solar Energy
Photovoltaic solar energy cells convert sunlight directly into electricity through the use of solar
induced electron–hole pairs. These solar cells are non-mechanical and are usually made from sil-
icon alloys. Photons carried by sunlight contain various amounts of energy corresponding to the
different wavelengths of the solar spectrum. Electrons are dislodged from the semiconductor’s
atoms and migrate to one surface, when solar energy is adsorbed. This migration of electrons, which
are negatively charged, leaves positive charges behind creating a voltage potential between positive
and negative surfaces. When these two surfaces (positive and negative) are connected to electrical
conductors, electricity flows. One cell produces about 1–2 W of power. Many photovoltaic cells are
typically linked together to generate power levels great enough to run most applications.
Hydrogen as a Fuel
Hydrogen fuel cells produce electrical power by bringing hydrogen and oxygen together in chem-
ical reaction. The process produces water and electricity. The mechanical power in hydrogen-fueled
vehicles is derived from this process. The only exhaust is water and heat. The process involves no
combustion or carbon exhaust and therefore is environmentally friendly. These energy cells are
efficient but expensive to build.
There is, however, a price to pay for the hydrogen. Hydrogen does not occur naturally in the
atmosphere. It must be separated from heavier molecules, such as water. Energy is required to bring
about this separation through the process of electrolysis. However, if energy is produced by hydro-
electric power stations, wind, or solar, hydrogen production and its use as an energy carrier is envi-
ronmentally friendly. Hydrogen has been called the perfect energy carrier.
Hydroelectric Power
Hydroelectric energy is generated from turbine-driven electric generators. The turbines are driven
by water flowing from a high level to a lower level. The difference in elevation represents the energy
potential to do work on the turbine. If the water is at the same level on both sides of the dam, there is
no energy potential, and the water can do no work on the turbine.
The elevation difference of the water across a dam is maintained by rain up stream. Rain is caused
by the sun, so the indirect energy from the sun is driving the turbine. Water wheels and ocean tidal
wave entrapment are also based on elevation potential.
Wind Turbines
There is enormous energy in hurricanes and tornadoes, and we have all witnessed their destructive
power. We have yet to harness this energy for constructive purposes. In fact, the potential of wind-
generated energy is much greater than the current consumption of energy worldwide. This is the-
oretical of course. The energy still has to be captured. It is interesting to know that there is much
energy potential all around us in the form of wind and solar. There has to be a means of econom-
ically capturing this energy for useful purposes.
Geothermal Energy
The earth’s crust is about 5 mi thick. Beyond that the rock becomes extremely hot. The hot core of
the earth is residual heat left from when it was formed some five billion years ago. There is plenty of
thermal energy within this core. So far the process of making this energy useful relies on drilling
into this very hot formation and pumping water into the hot rock. Supersaturated steam is produce
out of an adjacent well bore.
The closest distances to hot rock are in the caldera of old volcanoes. Some oil companies have
been very successful in producing thermal energy this way, especially on the west coast.
Atomic Energy
Atomic energy refers to the generation of high levels of heat through one of two types of reactions,
fission, and fusion.
Nuclear fission refers to a process of splitting atoms into small fragments. Heat is given off as a
source of energy, which is used in atomic power plants, to generate steam. The steam cycle to gen-
erate mechanical and electrical power is the same as for other types of fuels, such as coal and oil.
This process is also the basis of an atomic bomb. The use of nuclear fission for peaceful or destruc-
tive purposes depends on the rate of the atomic splitting process. In atomic power plants, the split-
ting process is highly controlled.
Nuclear fusion is the process of fusing two atomic nuclei together to form a heavier nucleus. Heat
energy is also generated in this process.
While there are no toxic gases exhausted to the atmosphere, the by-products of nuclear fission are
highly radioactive, creating a nuclear waste problem.
Biofuels
Biofuels are derived from near term “organisms.” Unlike fossil fuels, which have taken millions of
years to form, biomass has about a two gestation period. It is derived typically from plants such as
corn, soybeans, wheat, sugar beets, and sugar cane. A common type of biofuels is ethanol. It is made
through fermentation of certain biomasses such as sugar cane and corn. It must be blended with
standard fuels up to a maximum of around 10% to operate in standard cars. Engines have to be
modified to greater concentrations of biofuels with fossil fuels. Some cars are being manufactured
to run on any combination of fossil fuel and ethanol, even 100% ethanol.
Even though carbon dioxide is produced during combustion, biofuels are carbon neutral. This
means that carbon dioxide produced during combustion is adsorbed by new plant growth at the
same rate, and there is no net increase in carbon dioxide levels. Biofuels can help reduce the world’s
dependence on fossil fuel and carbon dioxide production.
Notes
1 www.boeing.com/commercial/787family/background.
2 http://www.eia.doe.gov/emeu/aer/eh/intro.html.
References
1 Keenan, J.H. and Keyes, F.B. (1938). Thermodynamic Properties of Steam, 1e. Wiley.
2 Stoever, H.J. (1953). Essentials of Engineering Thermodynamics. Wiley.
3 (1977). How Things Work. Granada Publishing Limited.
4 Obert, E.F. (ed.) (1950). Internal combustion engines. In: Analysis and Practice, 2e. Scranton, PA:
International Textbook Company.
5 John, J.E.A. and Haberman, W.L. (1988). Introduction to Fluid Mechanics. Englewood Cliffs, NJ:
Prentice Hall.
6 Eisenstein, P. (2004). Biggest jet engine. Popular Mechanics.
7 Housner, G.W. and Hudson, D.E. (1959). Applied Mechanics – Dynamics, 2e. Princeton, NJ: D. Van
Nostrand Co.
151
Power Transmission
Throughout antiquity stone was the primary building material. Stone was used to build structures,
such as the Pyramids of Misa (2600 BC), Mycenae (1600 BC), Parthenon (600 BC), and the famous
arches of Rome. Considering the density of stone (~160 lb/ft3), a 3-ft cube of stone weighs
W = 160 27 = 4320 lb
or about 2 tons. This is the approximate weight of each stone used to build the pyramids.
Archimedes (287 to 212 BC) identified three devices from antiquity that were used to trans-
mit force and motion. These devices were: (i) lever, (ii) pulley, and (iii) the screw. Later, three
other devices were added: (iv) wedge, (v) inclined plane, and (vi) wheel and axle (see
Figure 4.1).
These devices were used to remove, shape, and transport large blocks of stones from quar-
ries to a construction site. The wedge was a significant tool used by early builders to remove
cubic shaped stones. The sides were cut with chisels. Cracks underneath the stone’s base were
created by wedges and large hammers. Heavy stones were lifted in place by human effort
using these devices. Each represents a means for transmitting force to move huge blocks of
stone or other objects.
The wheel and axle were used for battle and transport of goods. Wheels can be attached to the axle
in two ways: rigidly fixed to the axle or freewheeling on bearings. The latter case eliminates skid-
ding on turns.
Even today, we find these devices used in machinery for transmitting force and power.
Settlers, during early days, made fences of split rails created by splitting logs with wooden wedges
and a wooden mallet. The wedge equation
F
N= 41
2 μ cos θ + sin θ
shows that if the half angle of the wedge is, say, 10 and the coefficient of friction is 0.3 then
N = 1 066F
152 Power Transmission
(a) F
Lever
W
a b
a
Wa = bF F= W
b
(b)
r
F
F W
F=W
Pulley
1
F=
W 2W
(c) T – Fr (d)
Screw F
F = (2μN cos θ + N sin θ)
F
N=
θ 2(μ cos θ + sin θ)
N
p 2πT – pW
Wedge
p W
F–
W r 2π
(e) (f)
W
Inclined plane
L
F μN
H F
θ
W N Wheel and axle
Say, μ = 0 FL=WH H
L=
F = W sin θ sin θ
Figure 4.1 (a–f ) Devices of antiquity.

Figure 4.2 Horse collar – a major breakthrough. Source: Courtesy of Little River Railroad and Lumber
Company Museum, Townsend, Tennessee.
In this example, the wedge generates an opening force, N, having a magnitude approximately
equal to the impact force, F, caused by the hammer.
For many years, oxen and horses were used to pull plows and wagons. Force generated by oxen
was transferred through a yoke. Earlier attempts to harness the horse included throat-girth harness
and breast-collar harness. The throat-girth harness (used as far back as ancient Chaldea, third Mil-
lennium BC) restricted breathing of the horse. It was not improved until the breast-strap harness
was developed in China (481 to 221 BC). In this harness, a strap or cinch was place around the belly
of the horse from which load was applied.
A breakthrough came with the horse collar, which puts pressure on the sternum
and transmits the pull force through the skeletal system of the horse without restricting
breathing. The horse collar was developed in China during fifth century AD and greatly
improved the capability of the horse. Studies show that a horse with a collar can apply
50% more force than an ox and at greater speed. The horse collar is essentially unchanged
since its inception.
Both horse and ox pull along straight lines and lack maneuverability, a limitation in the trans-
mission of animal power. Figure 4.2 shows two horses skidding a log into position for loading onto a
rail car. Force is transmitted to the log by means of horse collar and chain.
Gear Train Transmission
Water Wheel Transmission

In engineering practice, power is transformed into torque and speed to accomplish an end use. This
is done through a gear transmission or a pulley drive. Early designers of water wheels used gear
reduction drives (~1–25) to increase rotary speed of the millstones or saw blade.
Gears convert power from one set of torque and speed to another set. Often rotary speed is
reduced, and torque increased across a gear pair. This section gives basic features of gears and
reviews considerations in choosing gear tooth size for specific applications.
The shape of early gear teeth was not critical to the overall performance of a machine, such as the
water wheel, because gear speeds were low. Gear teeth were often made of wood having a tapered
flat surface of engagement. Gear teeth were made of apple wood, which is less likely to split during
operation. The flat contacting surface smooths out with use and performed quite well under water
wheel loads and speed.
As machinery began to run at higher and higher speeds, gear tooth profile became important. An
important feature of gear tooth geometry is that speed ratio between mating gears remains constant.
If not, the output speed of the driven gear would vary with time and become a source of machine
vibration.
Fundamental Gear Tooth Law

For the speed ratio to be constant, gear tooth profiles must follow the fundamental gear tooth law,
which states that the common normal to two mating surfaces must pass through the line of centers
at a fixed point, called the pitch point. This is essential to assure angular accelerations are zero
throughout the contact region. The proof of this law is well documented [1].
The speed ratio of two mating gears is expressed as
ω3 r2
= 42
ω2 r3
where r2 and r3 are the radii of the pitch circles for the two gears.
Many tooth profiles can satisfy the fundamental gear tooth law, but most have operational and
manufacturing limitations. Three profiles that have found practical application over the years are
involute [2], cycloid [1], and Wildhaber and Novikov [3, 4] profiles. Involute gears are by far the
most used and a brief summary is given below.
Other gear tooth geometries are possible provided they satisfy the fundamental gear tooth law [5].
Almost any geometry can be arbitrarily selected for one of the matching tooth pair. The challenge is
to find the tooth geometry that will mesh with the assumed profile while satisfying the fundamental
gear tooth law.
Involute Gear Features

The involute profile gets its name from how it is generated. An involute curve is generated by trac-
ing the path of the end of a tight string as it unwinds from a disc (Figure 4.3). The tip of the string
scribes an involute. Visualize a string wrapped around the base circle of gears 2 and 3. As the string
is unwound from the base circle on gear 2 onto the base circle on gear 3, any point (such as a knot)
would scribe involute curves on both gear pair. The normal to contacting tooth pairs is tangent to
both base circles and intersects the line of centers at the pitch point. The fundamental gear
tooth law is, therefore, satisfied. The radius of each base circle depends on the angle of the line
of action.
r B = r cos ϕ 43
Involute
ϕ P
Pressure line
ϕ
Base circle
rB Pitch circle
r
Figure 4.3 Involute gear tooth profile.
where r is the pitch radius and ϕ is the pressure angle. Pressure angles usually range between 20
and 25 , even though 14½ was once used. The pitch circle is different from the base circle. Its
radius is the distance to the pitch point from gear center.
By definition, diametral pitch (P) of a gear set is the number of teeth per inch of diameter of the
pitch circle.
N
P= 44
d
where d is diameter of the pitch circle and N is number of teeth on the gear.
It is a measure of the size of gear teeth in a gear set. For example, if a diametral pitch of 2 is spe-
cified for a gear pair for which the speed ratio is 1 : 2 and the pitch diameter of the pinion is 12 in.,
then the number of teeth on the pinion is
N p = P dp
N p = 2 12 = 24 teeth
The pitch diameter of the gear is 24 in. and
N g = 2dg
N g = 2 24 = 48 teeth
The addendum circle defines the outer edge of gear teeth, while the dedendum circle defines the
depth of tooth profile, both measured from the pitch circle.
1 1 25
The standard addendum and dedendum distances for interchangeable teeth are and ,
P P
respectively [1].
Circular pitch (p) is the arc length of a tooth plus tooth space.
πd π
p= = 45
N P
pP = π 46
Since p is the same for mating gears, P is the same.
The line of action, ϕ, is oriented from the pitch circle tangent line passing through the pitch point
at P.
Gear Tooth Size – Spur Gears

Pitch diameters establish the velocity and torque ratios in gear pairs regardless of tooth size. Tooth
size depends on force transmitted and geometric factors. Gear teeth must be large enough to trans-
mit the maximum force while keeping bending stress within allowable limits. Wilfred Lewis [6]
formulated a bending stress equation in terms of tooth form and load. His stress model is based
on treating gear teeth as cantilever beams (Figure 4.4).
The transmitted forced, W, is assumed to occur at the initial engagement point and with no other
tooth pairs in contact. Bending stress formula is developed below, using Lewis nomenclature.
Mc 6W t l
σ= = 47
I Ft 2
This formula has been modified to automatically account for standard gear tooth forms. The
resulting Lewis formula thus is written
W tP
σ= Lewis equation, Y is the form factor 48
K v FY
where
P – diametral pitch, teeth per diameter of pitch circle

Wt – tangent force
F – width of tooth
Y – form factor (tabulated by AGMA, also accounts for stress concentration)
Kv – dynamic factor
Various formulas are used to determine the dynamic factor. The Barth formula is stated below.
600
Kv = V is pitch line velocity in fpm 49
600 + V
F Figure 4.4 Lewis gear tooth model.

W
Wt
t
If gears are of high precision and there is no appreciable dynamic load, AGMA recom-
mends Kv = 1.
The AGMA also propose a modified version of the Lewis equation.
W tP
σ= 4 10
K v FJ
The parameter, J, is a geometric factor, which includes a modified form factor Y, a fatigue stress
concentration factor, Kf, and a load-sharing ratio, mN. J factors are also tabulated for various
gear forms.
Gear teeth are sized to transmit a specified force and maintain bending stress within specified
limits. Two controlling parameters are tooth thickness, which relates to diametral pitch, P, and
tooth width, F. As a rule, face width should be within the range 3p ≤ F ≤ 5p. If face width is too
long, the transmitted force may not be uniform across the tooth. This becomes a guide to be con-
sidered when determining cross-sectional geometry of gear teeth. In general, small gear teeth have a
large diametral pitch and are used to transmit power at high speeds and low torque. Larger gear
teeth are needed to transmit larger forces and thus have a smaller diametral pitch (low number
of teeth per inch of the pitch circle diameter).
Example Consider a simple speed reducer having the following design specifications.
Power transmitted – 50 hp
Input speed – 1000 rpm
Output speed – 500 rpm
Gear/Pinion material – UNS G10400 treated and drawn to 1000 F
Gear teeth – 20 full depth
Safety factor–3
The challenge is to determine the pitch diameter of both gear and pinion, the diametral pitch, and
tooth face width using the Lewis formula.
W tP
σa = 4 11
K v FY
The torque coming into the pinion is
TN 50 5252
Power = T= = 262 6 ft-lb
5252 1000
Following AGMA for coarse pitch, we choose P = 10 recommended for general use. Also, for 20
pressure line, the recommended minimum number of teeth is 18. Our choice for number of teeth
for the pinion is 60.
N N 60 π π
P= d= = = 6 in p= = = 0 314 16 in
d P 10 P 10
This fixes tooth size and diameter for the pinion. Based on this diameter,
262 6 12 in
Wt = = 1050 lb 4 12
3 1 ft
Numerical values in the Lewis formula are

σ yld 80 000
σa = = = 26 667 psi
FS 3
P = 10 arbitrarily chosen
W t = 1050 lb
Y = 0 494
Using these numbers and Kv = 1
W tP 1050 10
F= = = 0 797 in close to guideline 4 13
σaY 26667 0 494
AGMA recommends face width, F, between 3p ≤ F ≤ 5p.
Simple Gear Train

Kinematics
The diagram in Figure 4.5 shows pitch circles of a gear set. Assume the pinion (driving gear) is
centered at O1 and the driven gear is centered at O2. From the left diagram, the tangent velocity
of both circles at the contact point is the same giving
r1
ω2 = ω1 4 14
r2
From the right diagram, the tangent force (action reaction) is the same giving
r2
T2 = T1 4 15
r1
Since power in is T1ω1 and power out is T2ω2
r2 r1
Pout = T 2 ω2 = T 1 ω1 = T 1 ω1 4 16
r1 r2
Pout = Pin 4 17
ω2 T2
r2
r2
θ2
F
F
r1 θ1 ω1
r1
T1
Figure 4.5 Gear pair.

Gear pairs of this type are often used with electric motors to reduce rotary speed and increase
torque. Power is coming into the smaller gear and taken off the larger gear.
The starter motor and gear drive, found in every vehicle, are good examples of a simple gear train.
DC electric motors deliver moderate torque at high speeds and are ideal for this application. The
motor shaft, containing the pinion, drives the large gear, which elevates the torque and lowers the
speed. The higher torque to the engine drive shaft is necessary to overcome friction in the engine
and start-up inertia forces in each of the piston/connecting rods as well as the crank shaft. It is also
needed to initiate the compression and ignition strokes.
Worm Gear Train

A worm gear drive is useful when a large torque is required on the pinion gear shaft. In this case, the
worm gear rotates at a high speed and low torque, which is typical of electric motors. Figure 4.6
depicts a worm gear with input speed, ω1, and torque, T1. The torque demand, T2, on the pinion
gear is much greater than T1.
The speed and torque relationships are determined as follows.
The pitch line velocity of the pinion gear is V = r ω2. The pitch line velocity is due to the advance-
ment caused by the threads on the worm gear,
Δθ1 Δx
= 4 18
2π p
where
Δθ1 – angular displacement of worm gear

Δx – pitch line displacement of pinion
Dividing by Δt gives
p
V= ω1 4 19
2π
So
p
ω2 = ω1 4 20
2πr
p
T1
ω1
ω2
r
T2
Figure 4.6 Worm gear drive.

2πr
Since the number of teeth on the pinion gear is n =
p
ω1
ω2 = 4 21
n
showing that the output speed of the gear is a small fraction of the input speed of the worm gear.
Equating power-in to power-out gives
ω1
T2 = T 1 = nT 1 4 22
ω2
Worm gears are useful in cases where input power is working against a relatively large torque.
They are especially useful with electric motors, which produce power at high speed and low torque.
Worm gear drives can achieve speed and torque ratios in the order of 20 : 1. The axes of rotation of
the worm and gear are at 90 as shown. Worm gears can be self-locking, which is important in many
cases. Musical instruments are a good example. They occupy a smaller space than compound gear
trains.
Planetary Gear Trains

Planetary gear trains involve a gear (or gears) moving around another gear (or gears) as shown in
Figure 4.7. The outside gears are planet gears. The gears rotating about a fixed axis are sun gears.
This gear arrangement can produce a large step up (or down) in speed and torque.
Power can be supplied to planetary gear trains through gear a (link 2) and the rotating arm (link
5) separately or simultaneously. Both inputs affect output speed according to
ω4 = k1 ω5 + k2 ω2 4 23
The output speed of link 4 or gear d is the superposition of two inputs: links 5 (arm) and 2 (gear a).
Consider first the input from the arm (link 5).
The first gear constant (k1) in Eq. (4.23) is determined as follows. In this case, assume link 2 is
fixed. The driver becomes the arm (link 5). Assume the entire gear train is rotated clockwise,
including the arm (link 5). The motion of links 5, 2, and 4 is the same as shown in
Table 4.1. Next imagine that link 2 (gear a) rotates counterclockwise one turn back to its initial
fixed position; arm 5 is fixed during this motion. The resulting motions of the three links (2, 5, 4)
are shown in the second line. Adding each column gives the result of rotating the arm, link 5,
one turn.
b Figure 4.7 Planetary gear train.

c
Link 5
Link 4
Link 2 a
d
Table 4.1 Relative motions in planetary gear trains.
Member 5 2 4
Motion with 5 1 1 1
Motion relative to 5 0 −1 ac
−
bd
Total 1 0 ac
1−
bd
For one (1) rotation of the arm (link 5), gear 4 rotates
ac
k1 = 1 − 4 24
bd
Fixing arm (link 5) and rotating gear a (link 2) gives
ac
k2 = 4 25
bd
Planetary gear trains reduce speed (increase torque) at high ratios and occupy less space than
compound gear trains. They can be mounted in line with input and output shafts. They have better
service life and higher stability than compound gear trains. Speed and torque ratios of 10 : 1 are
common.
Compound Gear Trains

Now consider a typical application of a gear drive with input power being supplied by a 5 hp electric
motor rated at a speed of 1000 rpm (Figure 4.8). Using these two numbers establishes the output
torque of the motor.
2πTN 2πTN
P= = hp 4 26
550 60 33000
TN
P= hp
5252
where
P – hp
T – ft-lb
N – rpm
5 hp
2 hp 3 hp
ω2 T2
TA TB
T ω
T
ω T1 ω1
Figure 4.8 Gear train drive.

By substitution T = 26.26 ft-lb. Note the low-output torque and high-output speed of the elec-
tric motor.
Electric motors typically deliver power at high speed and low torque, so the first set of gears trans-
form power to a higher torque and lower speed (gear 2). This is done with a 1 : 4 gear pair (T2 = 4
T1 = 105 ft-lb and N2 = 250 rpm). The transmitted power can be consumed in several ways, such as
two tear drives A (2 hp) and B (3 hp). Gear 2 sets the rotary speed (NA = NB = N2 = 250 rpm). The
torque at gears A and B is determined from the power equation, P = Tω. T2 = 105 ft-lb and
N2 = 250 rpm
33 000 2
TA = = 42 ft-lb
2π250
33 000 3
TB = = 63 ft-lb
2π250
These torques drive something else.
The shaft section with the greatest torque is between gears 2 and A (T = 105 ft-lb). Assuming the
allowable shear stress in the shafting is τa = 20 000 psi, shaft diameter is
Tr 2T 2
τa = = 4 27
J πr 3
2 105 12
r3 = = 0 0401
π20000
r = 0.3423 in
d = 0 6847 in
Substituting and solving for r gives, r = 0.3423 in and a shaft diameter of d = 0.6847 in. The closest
standard shaft size would be d = 0.75 in.
Pulley Drives
Rope and Friction Pulleys

Another means of transferring power is through pulleys. We start with a simple rope and friction
pulley (Figure 4.9). This type of pulley is often found on fishing boats to pull in traps. The operator
pulls on one end of the rope to create a larger pull-in force, F0.
F Figure 4.9 Rope and friction pulley.

n
μN
θ
t
dθ
F + dF
F0
Pulley Drives 163
Consider the freebody diagram of a differential section of the rope against a pulley.
Fn = 0 Ft = 0 4 28
dθ dθ dθ
N = F + F + dF cos cos F = μN + F + dF cos
2 2 2
N = F dθ dF = − μN 4 29
Combining gives
dF
= − μ dθ 4 30
F
In this case, μ is the kinetic coefficient of friction as slippage is assumed between rope and pulley.
By direct integration and applying the boundary condition at θ = 0
F 0 = Feμθ θ = n2π + θ0 4 31
The rope usually wraps around the pulley several times to create high pulling force, F0. The
operators end provides a back pull, which affects the magnitude of the pulling force.
If, for example, the required pull-in force is F0 = 500 lb and assuming two complete wraps around
a pulley, i.e. θ = 2(2π) = 4π and μ = 0.2 then the operator applies
F0 500
F= = 0 2 4π = 40 5 lb 4 32
eμθ e
Power delivered to the incoming rope is Pout = rope velocity times force in rope. Total power con-
sumed by the drum is
P = Pout + friction losses 4 33
A cat head is an attachment (Figure 4.10) on a power shaft used to pull objects into a workspace.
Such a device is used around drilling rigs to lift and pull pipe or other pieces of equipment onto the
rig floor. Cat heads are also used to pull in and lift fishing traps onto boats. The shaft typically
rotates at a low rotary speed (rpm). A rope is wrapped around the cat head as shown and pulled
N Power source
F
Rope
Cat head
Figure 4.10 Pulling and lifting device.

with force F to move against a force P. A problem is to determine the magnitude of force, F to gen-
erate a force of P = 300 lb. Assume the rope makes two complete loops (plus 45 ) round the cat head
and the coefficient of kinetic friction between rope and pulley is μ = 0.2.
F = 300e − 0 2 2 125 2π
= 20 8 lb 4 34
This example illustrates the unique advantage in the use of cathead pulleys on oil rigs and fish-
ing boats.
Belted Connections Between Pulley Drives

Flexible connections, such as belts and chains, transmit power over a relatively long distance and
allow power transmission between shafts through pulleys (Figure 4.11). They are simpler and
cheaper than gear trains. Transmission ratios for speed and torque are the same as for gears except
for the possibility of slippage between belt and pulley. Since belts are long and stretch, they have the
added advantage of absorbing shock and damping torsional vibrations. Three common types of
belts are:
1) Flat belts – Flat belts were initially made of oak tanned leather. Modern flat belts are made of a
strong elastic core, such as steel or nylon, for strength in transmitting force (or torque). The core
is coated with special coating to enhance friction, while maintaining belt flexibility. Belt drives
are efficient and can be used over long distances between pulleys.
The application of the previous equation to belt drives assumes slippage is impending, so coef-
ficient of friction in this case is static coefficient of friction. To prevent slipping, the belt must be
pretensioned. With this in mind
T 1 = r1 F − Q 4 35
where force F is the transmitted force and Q is the pretension in the belt (Q ≺ F). Since
F = Qeμθ
The allowable torque and pretension are related by
T 1 = r 1 Q eμθ − 1
ω2
ω1 Q
T2
T1
r2
r1
F
Figure 4.11 Power transmission through pulleys.

Pulley Drives 165
where θ is the angle of contact between belt and pulley. The required pretension for a given
driving torque, T1, is
T1
Q= 4 36
r 1 eμθ − 1
where μ is the static coefficient of friction. Belt slippage occurs when
T1
Q≺ μθ
4 37
r1 e −1
T1 T2
This equation also applies to the driven pulley since = . Note that θ2 > θ1.
r1 r2
Example Using the following numbers determine the minimum preload, Q, required to transmit
3 hp at 800 rpm so the belt does not slip.
r1 = 2 in.
r2 = 4 in.
μ = 0.3
d = 14 in.
N1 = 800 rpm
The torque, T1, applied to pulley 1 is

P 3
T 1 = 5252 = 5252 = 19 7 ft-lb
N 800
T 1 = 19 7 12 = 236 3 in -lb
The belt angle, ϕ, is approximately
r2 − r1 2
tan ϕ = = = 0 143
d 14
ϕ = 8 13
The angle of contact on pulley 1 is θ1 = 180 − 16.26 = 163.7 . The angle of contact on pulley 2 is
θ2 = 180 + 16.26 = 196.26
163 7
θ1 = π = 2 86 rad
180
196 26
θ2 = π = 3 43 rad
180
By substitution, the minimum pretension is
236 3
Q= = 87 lb
2 e0 3 2 86 −1
This calculation is based on impending slippage because we used the coefficient of static friction.
The corresponding force, F is
T1
F= + Q = 205 3 lb
r1
Slippage between the belt and pulley 2 is
472 6
Q= = 65 7 lb
4 e0 3 3 43 −1
The smaller pulley dictates the required pretension because the angle of contact is smaller.
2) V-belts – V belts are made of fabric and cord, typically cotton, rayon, and nylon, which are
impregnated with rubber. The cord gives it strength, while the rubber gives flexibility and
enhances friction. They are used for shorter distances. They require grooved pulleys (sheaves).
While that are somewhat less efficient than blat belts, they can be used in a multiple drive. They
are also continuous and do not require a joining connection.
3) Timing belts – Timing belts contain rubber “teeth,” which provide a positive transfer of motion
and torque between pulleys. The surface of pulleys has recesses to engage the timing belt teeth.
They are made up of a wire core with a rubberized fabric bonded to it. The teeth made it possible
to run at any speed, fast or slow. Tooth engagement and disengagement may cause dynamic
fluctuations across the drive.
The relationships between torque and speed are the same as for gear pairs. However, in this case,
both pulleys rotate in the same direction.
Fundamentals of Shaft Design
Shafts are an important element in power transmission. As indicated previously, pulleys, gears, and
other linkages are attached to them, and it is through the shafts that torque gets transferred. Shafts
must be sized so that stress is kept within strength limits of the material. Torque in shafts produces
shear stress and shear strain.
When torque is applied (Figure 4.12), shear stress is generated on a small element as shown. One
component lies in the cross section, another also lies in a cross section some small distance to the
left. Two other shear stress components are also developed but in horizontal planes. The directions
of each stress component are consistent with the applied torque.
The sign convention for applied torque is plus (+) as shown, with the torque vector pointing out-
ward for the cross-sectional surface. The direction of the torque vector, T, follows the right-hand
rule, with fingers pointing in the direction of twist and the thumb in the direction of the vector. If
the torque vector points inward, then torque is considered negative (−).
Figure 4.12 Torsional shear stress.

r
θ
τ T
rθ
τ
2πr
Figure 4.13 Shear strain due to torque.
Shear Stress
Consider an imaginary cylinder within the shaft with radius, r. Let the outer surface be unrolled
into a flat surface (Figure 4.13)
where
L – length of shaft
r – radius of imaginary cylinder within the shaft
The twist in the shaft, due to torque, distorts this surface as indicated by ϕ. Simple shear strain is
developed from this diagram.
rθ
γ=ϕ= 4 38
L
It follows that
r
τ = Gγ = G θ 4 39
L
Integrating the shear stress across the face of the shaft
R R
θ 2
T= r τ dA = G r dA 4 40
L
0 0
θ
T=G J
L
TL
θ= 4 41
GJ
This equation defines the amount of twist over the length of a shaft and is used to find torsion
spring constants. The parameters are
θ – rotational twist of one end relative to the other

T – applied torque
L – length of shaft
G – modulus of rigidity
J – cross-section polar moment of inertia
Units must be compatible

By substitution, the magnitude of shear stress at any radial distance, r, is given by
Tr
τ= 4 42
J
Equation (4.42) shows the maximum shear stress is located at the outer surface, r = R. The direc-
tion of the shear strain and stress is consistent with the transmitted torque.
Example Consider a shaft with the following parameters

T = 1500 ft-lb
σ yld = 60 000 psi
τyld = 0.577(60 000) = 34 620 psi
G = 12 × 106 psi
FS = 1.5
L = 12 in.
τyld
τallowable = = 23 080 psi
FS
The shaft size under these conditions is

TR 2T
τa = =
J πR3
4 43
2T 2 1500 12
R3 = = = 0 4965
πτa π23 080
Giving R = 0.792 in. Shaft diameter (1.584 in.) would be rounded up to the lowest standard shaft
5
sizes, say D = 1 in diameter.
8
The expected twist over the active length of the shaft (12 in.) is
4
TL π 1 625
θ= , J= = 0 6846 in 4
GJ 2 2
1500 12 12
θ= = 0 026 29 rad = 1 51 4 44
12 × 106 0 6846
There are no normal stresses in the stress element shown in Figure 4.13. However, there could be
normal stresses developed simultaneously by other loads.
This derivation applies to circular shafts. Stresses in noncircular shafts, such as elliptical, trian-
gular, and other shapes, are discussed in Ref. [7].
Example Consider a case where bending and torque are applied simultaneous (Figure 4.14). In
this example, it is required to determine
a) stress condition at point A showing the stresses on an element;
b) principal stresses and maximum shear stress;
c) safety factor based on the energy of distortion criteria of failure. Yield strength of the material is
100 ksi.
The cross-sectional moment of inertia of the 1 diameter shaft is

π 4 π 4
I= R = 05 = 0 05 in 4
4 4
Its polar moment of inertia is J = 0.1 in.4
10 in.
A
x
1″ OD
z
8 in.
500 lb
Figure 4.14 Shaft under bending and torque.
The bending moment at location A is 5000 in.-lb. The torque applied to cross section A is 4000 in.-
lb. Corresponding bending and torsion stresses are
Mc 5000 0 5
σ= = = 50 000 psi
I 0 05
Tc 4000 0 5
τ= = = 20 000 psi
J 01
The stress element for point A is shown in Figure 4.15.
From Mohr’s stress circle
1
R = 252 + 202 2
= 32 4 45
There
σ 1 = 25 + 32 = 57 ksi 4 46
σ 2 = 25 − 32 = − 7 ksi 4 47
50 k ips
x
20 k ips
τ x axis
20
R σ1
σ2 25 2θ
σ
50
Figure 4.15 Mohr’s circle for normal and shear stress.

Applying the energy of distortion criteria of failure

1
σ = σ 21 − σ 1 σ 2 + σ 22 2
4 48
1
2
σ = 572 + 57 7 + − 7 2
= 60 8 ksi
Factor of Safety is
100
FS = = 1 64 4 49
60 6
Stress Analysis of Shafts
Example Consider the power train shown in Figure 4.16. An electric motor delivers power to
pulley B by way of a V belt. The rotary speed of pulley B is 500 rpm. The magnitude of the force
in the belt on one side of the pulley is 400 lb and the magnitude of the force on the other side is
200 lb. Determine the maximum normal stress in the shaft at location “a” during each revolution.
Assume the shaft is 1 in. in diameter and simply supported at the bearings.
Torque generate at pulley B is
T B = 5 400 − 200 = 1000 in -lb

This torque is opposed by output torque, T, which does useful work on something. Shear stress
produced in the shaft is
Tc
τ= 4 50
J
where
T = 1000 in.-lb
c = 0.5 in.
π
J= 0 5 4 = 0 0982 in 4
2
Pulley B
10 in.
ω T
a
1″ OD
Motor
Pulley A
Figure 4.16 “V” belt pulley transmission.

So
1000 0 5π
τ= = 5092 psi 4 51
0 0982
Bending stress at point “a” is
M a = 300 10 = 3000 in -lb
Mc 3000 0 5
σa = = = 30 550 psi 4 52
I 0 0491
Principal stress
σx σx 2
σ p1 = + + τ2xy 4 53
2 2
2
30 550 30 550
σ p1 = + + 50922 = 31 376 psi 4 54
2 2
Twisting in Shafts Having Multiple Gears

When a shaft contains several gears (Figure 4.17), internal torque between each gear is constant,
but different in magnitude. The torque coming into each gear is indicated by the torque vectors.
Internal torque between each gear is also shown along with the angular twist of each gear relative
to gear “a.”
20 ft-lb 55 ft-lb 10 ft-lb 15 ft-lb 10 ft-lb
T
+20
+
– –10
–25
–35
θ
2 c d e
a
1 b 3 4
θe/a
Figure 4.17 Shaft rotation of shaft between gears.

The twist of gear b relative to gear a is

TL 20 L
θa b = = 4 55
GJ GJ
The twist of gear “j” relative to gear a is
j
T i Li
θ j a = 4 56
1
GJ
Transverse loads generated by gears and pulleys create bending stresses in addition to shear
stress. All applied loads have to be considered in designing shafts as well as bearing supports.
Keyway Design
Keys, pins, and types of retainers are used to transmit torque from gears and pulleys. Examples of
such fasteners are:
•• Square keys
Round keys
•• Round pins passing through a hub and shaft

Tapered pins
Square keys are the most common means of transmitting torque to shafts. As a rule, the sides of
square keys are ¼ shaft diameter. Length is adjusted according hub length. Two keys may be used if
necessary.
Example Consider a gear-transmitting torque of 4200 lb-in. (Figure 4.18). The steel shaft has
yield strength of 75 ksi and a diameter of 1.5 in. A square key having side dimensions of
0.375 in. and yield strength of 65 ksi is chosen. Determine the factor of safety for the key if it is
1.5 in. long.
The shear force in the key is
T 4200
V= = = 5600 lb 4 57
0 75 0 75
F
a
V
Hub
Pitch circle
Figure 4.18 Torque transmission though keys.

The shear strength of the key material, per the von Mises energy of distortion criteria, is
τyld = 0 577σ yld = 37 5 ksi 4 58
The applied shear stress in the 38 in key is

5600
τ= = 9955 psi
0 375 1 5
The factor of safety of the key is
37 500
FS = = 3 77 4 59
9955
Mechanical Linkages
The study of mechanical linkages requires understanding of position, velocity, and acceleration of
key points as well as angular velocities and accelerations. Two factors affect the design and appli-
cation of linkages; (i) desired motion (kinematics) and (ii) inertia forces (dynamics) generated by
achieving the desired motion. Inertia forces radiate throughout machinery affecting bearings,
shafts, and frame.
Relative Motion Between Two Points

Consider two particles A and B having motions defined in relation to coordinates XY (fixed) and xy
(translating, not rotating). In the first case, assume the two points are not connected (Figure 4.19a).
The position vectors locating points A and B are
RA = X A i + Y A j Point A 4 60
RB = X B i + Y B j Point B 4 61
rA B = xi + yj 4 62
The position of point A is
RA = RB + r A B 4 63
(a) (b)
Y Y
y y
A A
x x
B B
X X
O O
Figure 4.19 Relative velocity and acceleration. (a) Point A independent of B. (b) Points A and B on same link
(special case of a).
By differentiation
dRA dRB dr A B
vA = = + 4 64
dt dt dt
vA = vB + vA B 4 65
In each case, the unit vectors do not change with time. Typically, the velocities and accelerations
of both points are known, and the relative velocity and acceleration are to be determined. For the
first case (Figure 4.19a), assume
vA = 200 − cos 30i + sin 30j mph
vB = 400 − cos 45i − sin 45j mph
Then the velocity of A with respect to B is
vA B = vA − vB 4 66
vA B = 109 64i + 382 8j mph
Relative acceleration is determined in a similarly starting with

aA = aB + aA B
When aA and aB are known

aA B = aA − aB
The second case (Figure 4.19b) shows two points fixed on a rigid body. The velocity and acceler-
ation of one point (say Point B) and angular velocity and acceleration of the rigid body are usually
known or can be determined from the linkage configuration. The objective is to determine the
velocity and acceleration of Point A. The following information is given for this example.
vB = 10 cos 60i + sin 60j fps 4 67
aB = 5 cos 45i − sin 45j fps2 4 68
The angular velocity of body AB is
ω = 2 rad s CCW
The angular acceleration of body AB is
α = 4 rad s2 CCW
L = 1 ft (fixed distance between A and B)
θ = 60
The velocity of point A is determined as follows.
vA = vB + vA B velocity
where
vA B = Lω − i sin θ + j cos θ 4 69
vA B = 2 − 0 866i + 0 5j fps
By substitution
vA = 10 i cos 60 + j sin 60 − 1 732i + j

vA = 3 268i + 9 66j
The acceleration of point A is determined from
aA = aB + aA B
The acceleration of A relative to B has two components, normal and tangent.

aA B = aA B n + aA B t 4 70
The normal component is
aA B n = ω2 L − cos θi − sin θj 4 71
aA B n = 4 − 0 5i − 0 866j
The tangent component is

aA B t = αL − sin θi + cos θj 4 72
aA B t = 4 − 0 866i + 0 5j fps2
Combining the terms gives
aA = 5 i cos 45 − j sin 45 + 4 − 0 5i − j0 886 + 4 − 0 866i + 0 5j

aA = − 1 93i − 5j
Absolute Motion Within a Rotating Reference Frame

In this arrangement, a point A moves in space and this motion is observed from a rotating x,y ref-
erence frame as shown in Figure 4.20. The rotating reference frame (xy) has angular acceleration, α,
as well as angular velocity, ω. Point B is the origin of the moving reference frame. Since the x,y
reference frame rotates as well as translates the derivatives of the unit vectors, i and j, are not zero.
The position of point A relative to the fixed XY frame is
rA = rB + r
By differentiation of the position vectors, the absolute velocity of point A is determined as follows.
dr
vA = vB + 4 73
dt
Since
r = ix + jy
dr di dj
=x + ix + y + jy 4 74
dt dt dt
Figure 4.20 Point moving within a rotating

A P
reference frame. Y
y
x
r
α
ω
B
O X
Also
di dj
= ωj and = − ωi
dt dt
By substitution
vA = vB + ω − iy + jx + v 4 75
vA = vB + ωxr 4 76
The interpretation of each term in this equation is important. First note that point A is the point of
interest. Point P is located within the xy frame and is coincident with point A at a given instant
of time.
vA – absolute velocity of A
vB – absolute velocity of B (origin of the xy frame)
ω – angular velocity of the rotating reference frame, xy
r – position of point P within the xy reference frame coincident with point A
ϖxr – absolute velocity of point P located within the xy frame and coincident with point A
v – velocity of point A relative to the xy reference frame vA P
If point B is fixed in space, the velocity of point A is

vA = ϖxr + v 4 77
or
vA = vP + vA P 4 78
Acceleration of point A is established by rate of change of velocity with time. By differentiation of

Eq. (4.78) and in consideration of the derivatives of the unit vectors
aA = aB + αxr + ωx ωxr + a + 2ωxv 4 79
where
aA – absolute acceleration of A
aB – absolute acceleration of B
α – angular acceleration of the xy reference frame
ω – angular velocity of rotating reference frame
αxr – tangent component of acceleration of a coincident point P in xy frame
ω x(ωxr) – normal component of acceleration of coincident point P in xy frame
r – radial position of point P (and A) relative to xy frame
a – acceleration of point A relative to the xy frame, axy
v – velocity of point A relative to the xy frame, vxy
The last term in Eq. (4.79) is the Coriolis component of acceleration. If point B is fixed, Eq. (4.79)
becomes
aA = aP t + aP n + axy + 2ωvxy 4 80
Example Consider a situation of a person (dot A) moving outwardly along the x axis, while a
merry-go-round is rotating with ω (Figure 4.21).
x
vxy
A A
aA
anP
aCC
ω
Figure 4.21 Coriolis component of point moving on a rotating frame.
Applying Eq. (4.81)

P – point on x axis coincident with dot A
A – person moving outward along x axis
N – 20 rpm
2π
ω= 20 = 2 09 rad s
60
r = 2 ft (radial distance from origin)
vxy = vA/P = 0.5 fps (assumed)
vA = vP + vA P
The acceleration of point A is
aA = anP + 2ωvA P 4 81
2
anP = ω r = 2 09 2 = 8 74 fps directed toward the origin
2 2
2ωvA/P = 2(2.09)0.5 = 2.09 fps2 (Coriolis component)

Mechanical linkages are used to impart specific motions to accomplish a given output or effect.
There are many possible configurations depending on the required output. The following covers
five common types of linkages. In practice, linkages may contain a combination of each [8, 9].
Scotch Yoke
The Scotch Yoke (Figure 4.22) is a mechanical analog for generating simple harmonic motion. The
motion of point B is defined by
x = r − r cos θ 4 82
x = r sin θθ = vA sin θ 4 83
x = r θ cos θθ + sin θθ 4 84
x = rω2 cos θ + rθ sin θ 4 85
If the crank rotates at a constant velocity, ω, then θ = 0 and
x = r ω2 cos θ
Acceleration
polygon
aB
A
θ
ω aA/B
aA
r
vA
vB/A
θ B
vB
Velocity x
polygon
Figure 4.22 Scotch yoke mechanism.
The displacement of the yoke is the same throughout. Taking B in the yoke as a coincident point
with point A and knowing the velocity of point A, the relative sliding velocity of point A within the
grove is determined from the relative velocity equation.
vA = vB + vA B 4 86
The velocity polygon is shown with the center coincident with the center of rotation of the crank.
Its construction starts with known velocity of point A (vA = rω). The direction of the other two com-
ponents is used to close the polygon.
vA B = vA − vB 4 87
The relation between these vectors is shown in Figure 4.22. Also

aB = aA + aB A 4 88
Here the acceleration of point A is known (aA = ω2r). The direction of the other two components
closes the polygon.
Slider Crank Mechanism

The history of the slider crank mechanism starts with James Watt’s steam engine (1769) and the
need to convert linear motion into rotary motion. Steam was used as the driving force to displace a
piston, while the connecting rod provided the means to generate pure rotation to a crank shaft. The
slider crank mechanism was used throughout the Industrial Revolution of the 1800s, most notably
the steam locomotive.
The slider crank mechanism illustrates all three types of rigid body dynamics: translation
(piston), rotation (crank shaft), and general (connecting rod) motion. With reference to
Figure 4.23 and assuming the crank rotates with constant velocity, it is desirable to find the velo-
cities and accelerations of points A and B as well as the angular velocity and acceleration of the
connecting rod.
Knowing the angular velocity (ω) of the crank shaft, the angular velocity of the connecting rod
and linear velocity of the piston are determined from a velocity polygon. This is needed to deter-
mine the angular acceleration of the connecting rod and the linear acceleration of the piston, point
B. The velocity polygon is shown (Figure 4.23) along with the slider crank to show the true direc-
tions of each vector with respect to the alignment of each link. The angular velocity of the connect-
ing rod is extracted from the relative velocity, vB A , or from the instantaneous center (IC).
Velocity Analysis
The velocity polygon is based on
vB = vA + vB A 4 89
The velocity of point A (vA = rω) is known. The velocity polygon is closed with the known direc-
tions of the other two components.
The x and y components of the velocity vectors in the polygon give
x vB = vA sin θ + vB A sin ϕ 4 90
y 0 = vA cos θ − vB A cos ϕ 4 91
Solving them together gives

vB = vA sin θ + sin ϕ 4 92
cos θ
vB A = vA 4 93
cos ϕ
IC
A
ω
ϕ vA
L
θ R
B
vB/A θ ϕ
vB
O
Slider crank
Velocity polygon
x
Figure 4.23 Angular velocity of connecting rod.

where angles ϕ and θ are related by

R
sin ϕ = sin θ 4 94
L
Once vB/A is determined,
vB A
ωAB = clockwise 4 95
L
The velocity of point B and angular velocity of link AB can also be determined by use of the
instantaneous center.
Acceleration Analysis
The acceleration polygon is drawn with origin at point O along with the slider crank to show the
true direction of each acceleration component with respect each linkage (Figure 4.24). The accel-
eration polygon is based on
aB = aA + aB A 4 96
The acceleration of point A is known (aA = ω R). Only the direction of point B is known. The
2
acceleration B relative to A has two components. The normal component is known

anB A = ω2AB L . Only the direction of atB A is known. Closure of the acceleration polygon is shown
in Figure 4.24.
x aB = aA cos θ + anB A cos ϕ − atB A sin ϕ 4 97

y 0 = aA sin θ − anB A sin ϕ − atB A cos ϕ 4 98
From Eq. (4.98)

sin θ
atB = aA − anB tan ϕ 4 99
A
cos ϕ A
A
ω
L
R B
ϕ aB θ ϕ
O
(aB/A)t
aA x
(aB/A)n
ϕ
Figure 4.24 Angular acceleration of connecting rod.

15 000
Connecting rod,
rad/s2
10 000
Piston, fps2
5000
Acceleration
0
0 60 120 180 240 300 360
–5000
–10 000
–15 000
Crank angle, °
Figure 4.25 Acceleration of connecting rod and piston.
where
anB A = ω2AB L 4 100
The tangent component, atB A = αAB L. The vector sum of both normal and tangents components
yields αAB.
The dynamic forces in the bearings at connections O, A, B are affected by the magnitude of the
accelerations of each of the three components (crank, connecting rod, piston). Angular acceleration
of the connecting rod and linear accelerations of the pistons are displayed in Figure 4.25 for one
complete rotation of the crank. The numbers are based on a crank radius of R = 3 in., a connecting
rod of L = 12 in., and a crank rotary speed of 2000 rpm.
The block represents a piston sliding inside a cylinder. The linear motion of the block is a function
of angular position of the crank defined by
x θ = R 1 − cos θ + L 1 − cos ϕ 4 101
where
r sin θ = L sin ϕ
These formulations are used to predict velocities and accelerations of the piston as continuous
functions of θ.
Four-Bar Linkage
This mechanism contains three moving links plus a fixed link represented by 00∗ (Figure 4.26). The
connecting rod, AB, is extended out to point C for generality. This linkage allows a multitude of
motions for point C depending on requirements of a design. Movement of any point can be deter-
mined by stepping the driving link, OA, through different angular positions, θ, and tracking other
points, including point C. Linear and angular velocities are determined from the velocity polygon as
shown. Velocities are required to determine linear and angular accelerations. They are needed to
determine normal acceleration components of the linkages.
IC
vA
B C
A
vA
vB/A O
O*
vB (aB)n
(aB/A)n aB aA
(aB/A)t (aB)t
Figure 4.26 Parameters to define a four-bar linkage.
The motion of each member (a, b, c, d) and any point on each member can be established as
follows. The relation between the lengths of each member and angular positions, defined by
(θ, ψ, ϕ), is determined from
c sin ϕ − b sin ψ = a sin θ 4 102
c cos ϕ + b cos ψ = d − a cos θ 4 103
Squaring both sides of each equation and then adding the results of the two gives
a2 + d2 − c2 − b2 − 2da cos θ
cos ϕ + ψ = 4 104
2bc
1
ϕ + ψ = cos − 1 a2 + d2 − c2 − b2 − 2da cos θ 4 105
2bc
where
a = OA b = AB c = O∗B d = OO∗
θ = slope of OA ϕ = interior slope of O∗B ψ = slope of AB
Typically, the angle θ is the independent variable. Angles ϕ and ψ are determined by solving
Eqs. (4.103) and (4.105) simultaneously.
Example Consider the geometry of a four-bar linkage having the following parameters
a = 4 in.
b = 8 in.
c = 10 in.
d = 12 in.
θ = 60
ω = 2π rad/s (N = 60 rpm)
With θ = 60 , we next establish values for and ψ by solving Eqs. (4.103) and (4.105) simultane-
ously. These values, determined by trial and error, are ψ = 44 and = 65 . Note that the sum of and
ψ is 109 according to Eq. (4.105).
Velocity Analysis
The unknowns are ωAB and vB. These parameters are determined from
vB = vA + vB A 4 106
This vector equation yields two scalar equations for finding the magnitudes of vB and vB/A.
vB − i cos 90 − ϕ − j sin 90 − ϕ = vA − i sin θ + j cos θ + vB A i sin ψ − j cos ψ

4 107
x − vB cos 25 = − vA sin θ + vB A sin ψ 4 108
y − vB sin 25 = vA cos θ − vB A cos ψ 4 109
x − vB cos 25 = − vA sin 60 + vB A sin 44
y − vB sin 25 = vA cos 60 − vB A cos 44
x − 0 91vB = − 0 866vA + 0 69vB A
y − 0 42vB = 0 5vA − 0 72vB A
Since
vA = aω = 4 2π = 25 13 ips
The two simultaneous equations become
x 0 9063 vB + 0 6947 vB A = 21 7626 4 110
y − 0 4226 vB + 0 7193 vB A = 12 565 4 111
which give vB = 7.33 ips and vB/A = 21.77 ips. The angular velocities of link AB and 0 ∗ B are
vB A 21 77
ωAB = = = 2 72 rad s
b 8
vB 7 33
ω0∗B = = = 0 733 rad s
c 10
The velocity polygon is shown in Figure 4.26.
Accelerations are important, especially in high-speed machinery, where inertia effects can generate
large bearing loads and other reactions. It is a good idea to quantify the inertia effects and forces
throughout any mechanism. The relative acceleration equations can be useful in this regard.
aB = aA + aA B 4 112
n t n t
aB + aB = aA + aB A + aB A 4 113
where
n
aB = ωO∗B 2 c
aB t = αO∗B c
aA = ω 2 a
n
aB A = ωAB 2 b
t
aB A = αAB b
The acceleration polygon is also shown in Figure 4.26.

The acceleration equation yields two scalar equations. They are used to find the magnitude of
(aB)t and (aB/A)t, which are then used to find angular accelerations. The directions of all vectors
are known from the drawings. The velocity and acceleration of point C can be determined from
similar polygons.
Three-Bar Linkage
A three-bar linkage is one in which two rotating members are connected by a sliding collar as
shown in Figure 4.27. One end of each member is pin connected to a rigid frame: the frame being
the third link. The link (OA) is connected to a collar and typically rotates at a constant angular
velocity, Ω. It is desired to determine the angular velocity (ω) and acceleration (α) of the second
link (y axis), on which the collar slides.
The first step is to determine the angular velocity (ω) of the x,y reference frame and relative veloc-
ity of the sliding collar. Velocity information is needed to complete the acceleration analysis. Keep
in mind that point A is located on the collar, while point P is coincident with A, but located on the
y axis.
Y
y
P
A x
ω,α
Ω vP
vA/P
vA
O X
B
Figure 4.27 Three-bar mechanism.

Velocity Equation
The velocity of point A on the collar is related to ω by
vA = vP + vA P 4 114
where
vA – absolute velocity of point A

vP – velocity of point P on the xy rotating frame coincident with point A
vA P – velocity of point A relative to the xy rotating frame
The velocity polygon, which is the vector representation of Eq. (4.114), is shown in Figure 4.27.
This vector equation separates into two scalar equations, which yield the magnitude of the sliding
velocity and the absolute velocity of P. The angular velocity of the y axis is then found from the
velocity of point P.
Acceleration Equation
The acceleration of point A relates to α by
aA = aP t + aP n + axy + 2ωxvxy 4 115
where
vxy – velocity of A relative to the y axis vA P or the sliding velocity of the collar on the rotating y
axis. It is determined from the velocity analysis
axy – acceleration of A relative to the xy reference frame. In this case, it is the sliding acceleration of
A on the y axis. Its direction is known, but not its magnitude
ap n – normal component of the acceleration of P (= ω2r). It is directed toward point B
aP t – tangent component of the acceleration of P. Its magnitude is αr
2ωxvxy – Coriolis acceleration component
This vector equation separates into two scalar equations whose solutions give the magnitude of
the sliding acceleration of A on the y axis (aA/P) and the tangent component of the absolute accel-
eration of point P. Angular acceleration, α, is determined from this tangent component.
Example Consider the example (Figure 4.28) with following parameters.
a = 8 in.
c = 12 in.
Y
y
P
Ω A
b ω x
βˆ a ĵ
iˆ
θ ϕ
X
αˆ c
Figure 4.28 Rotating reference frame.

θ = 30
Ω = 10 rad/s (constant)
Essential parameters used in solving this problem are shown in the diagram below. The xy ref-
erence frame rotates. The XY frame is fixed. Point P is coincident with point A and lies within to the
rotating frame, while point A moves with the collar along the y axis.
The problem is to determine the angular velocity (ω) and acceleration (α) of the xy reference
frame. Distance, b (or r), is determined by the law of cosine.
b2 = a2 + c2 − 2 ac cos 30 4 116
b = 6 46
Angle, ϕ, is determined from
b cos ϕ + 8 cos 30 = 12
to be ϕ = 38.2 .
Velocity Analysis
The velocity polygon is based on Eq. (4.114) (also see Figure 4.28). The velocity vectors are
vA = vA α sin θ − β cos θ 4 117
where vA = Ω a = 10 (8) = 80 ips
vP = − ivP
where
vP = b ω = 6 46 ω
vA P = vA − vP 4 118
By substitution
vA P = 80 α sin θ − β cos θ − − 6 46ωi 4 119
In terms of i, j,
α = i sin ϕ − j cos ϕ
β = i cos ϕ + j sin ϕ
By substitution
vA P = − 29 7 + 6 46ω i − 74 25j 4 120
From the velocity polygon, we see that
vA P x =0
− 29 74 + 6 43ω = 0
ω = 4.62 rad/s
Also
vA P = vA P y = − 74 24 ips 4 121
The minus sign means the collar is moving toward the origin. Bringing terms together
vA = − 29 8i − 74 3j
The two unknowns at this point are α and the linear acceleration, (aA)xy, the sliding acceleration of
the collar A on the y axis. Starting with the acceleration equation
where
aP t =rα
aP n = ω2 r
axy = aA xy
Components in the acceleration equation are

aP t = bα − i = − i6 46α 4 123
aP n = ω2 b − j = − 4 622 6 46j = − 137 9j 4 124
axy = axy j assumed direction
axy n
=0
Coriolis component
2 ω xvxy = 2ωvxy i = 2ωvA P i
= 2 4 62 74 24 = 686i 4 125
By substitution
aA = − 6 46α + 686 i − 138 − axy j 4 126
The absolute acceleration of point A or the collar is known.
aA = 800 i sin ϕ + θ + j cos ϕ + θ
aA = − 742 8i + 297 1j 4 127
Bringing all terms into Eq. (4.122) gives
x − 742 8 = − 6 46α + 686 giving α = 221 2 rad s2 CCW
y 297 1 = − 137 88 + axy giving axy = 435 ips2
The acceleration polygon is shown in Figure 4.29.
Figure 4.29 Acceleration polygon. y

P
(ap)n
aA
Coriolis
(ap)t
O B
(aA)xy
x
r1
θ ϕ
r2
Figure 4.30 Geneva mechanism – intermittent motion.
The direction of the acceleration of point A should be toward point 0 (as shown). A good check on
the polygon.
Geneva Mechanism
The Geneva Mechanism (Figure 4.30) is a special case of a three-bar mechanism. It is a mechanism
that is commonly used in watches. Driver 1 rotates at a constant angular velocity and imparts rota-
tion to the receiver 2, which experiences intermittent motion at a nonuniform speed.
The position of the Geneva wheel is determined by
sin θ
tan ϕ = c 4 128
− cos θ
r1
where c is the center distance. By differentiation
c
cos θ − 1
r1
ω2 = ω1 4 129
c 2 c
1+ − 2 cos θ
r1 r1
Differentiating again gives the expression for angular acceleration of the Geneva wheel.
2
c c
sin θ 1 −
r1 r1
α2 = ω21 2 4 130
2
c c
1+ − 2 cos θ
r1 r1
The receiving member can have more than four slots. An important feature is that the pin enters
the receiver tangent to the slot. These equations apply only when 45 ≥ θ ≥ − 45 .
As a comparison with the vector analysis, assume (c/r1 = 12/8 = 1.5 and θ = 30 ).
ϕ = 38.26 (using (4.128)) ϕ = 38.26
ω = 4.59 rad/s (using (4.129)) ω = 4.62 rad/s (Eq. (4.120))
α = 220.58 rad/s2 (using (4.130)) α = 217.4 rad/s2 (Eq. (4.127))
Flat Gear Tooth and Mating Profile

Another interesting kinematic situation is the profile of novel gear teeth. In this case, the motion
between two gears is given, i.e. the angular velocity ratio of two gears is constant. The question
becomes, knowing the profile of one gear tooth, what the profile of the mating gear tooth. Both
profiles must satisfy the fundamental gear tooth law. The attraction to such a gear set is reduced
Hertz contact stresses.
Consider, for example, a case where one tooth profile is flat, and the other profile is to be deter-
mined. The setup to the problem is illustrated in Figure 4.31 along with variables and reference
frames. The flat tooth profile is fixed to the rotating xy frame, while the profile to be determined
is fixed to the rotating x1y1 rotating frame. The solution strategy is to define the path of the contact
point, P, in terms of the x1y1 frame. Vector positions of point P in terms of each reference frame are
r = ix + jy and r 1 = αx 1 + βy1 4 131

The relation between these vectors is (d = a + b)
r1 = r − d 4 132
αx 1 + βy1 = i x − d cos θ + j y − d sin θ 4 133
The unit vectors are related by
i = α cos θ + ϕ − β sin θ + ϕ 4 134

j = α sin θ + ϕ + β cos θ + ϕ 4 135
By substitution
x 1 = x cos θ + ϕ + y sin θ + ϕ − d cos ϕ 4 136
y
y1
a cos θ
j β
x1
α
θ h M ϕ
O
O1
i x
a b
Figure 4.31 Kinematics of gear tooth profiles.

y1 = − x sin θ + ϕ + y cos θ + ϕ + d sin ϕ 4 137

Note that θ and ϕ are related by θ = b aϕ. These two equations define the profile we seek.
The first step is to define the parametric equation x(θ) and y(θ). With regard to the above figure,
1
x θ = cos α cos α − β h2 + a2 − 2ha sin θ 2
4 138
1
y θ = h + sin α cos α − β h2 + a2 − 2ha sin θ 2
4 139
where
h − a sin θ
tan β =
a cos θ
Alternatively,
y = tan θ x + h 4 140
Both coordinates define the location of P in the xy frame. The coordinates of point P in the x1y1
frame are determined by the transformation equations given above.
Example Consider the following parameters.
a = 6 in.
b = 3 in.
d = 9 in.
α = 0.17453 rad (10 )
h = 1 in.
The above equations were programed to obtain the mating tooth profile as viewed from the x1y1
reference from. Both profiles are shown Figure 4.32. The portion of each profile used for form teeth
on both gears depends on the diametral pitch of the gear system.
These equations were used to generate gear tooth pairs for a variety of flat surface orientations.
One such gear pair is shown in Figure 4.33.
2.5
P
2
y1 distance (in.)
1.5
0.5
b
M
0
–4 –3 –2 –1 0
x1 distance (in.)
Figure 4.32 Gear tooth profiles.

Cam Drives 191
Figure 4.33 Gear pair with flat teeth.
A companion study defines flat tooth profiles in hypoid gears [10].
Cam Drives
The purpose of a cam is to convert rotary motion into linear motion with specified displacements
over one complete rotation of the cam. The specified motion will normally have a low and a high
point with intermediate positions related to the shape of the cam.
Velocity and acceleration of the follower are also important as they affect contact forces and pos-
sible separation. Cams of a given shape are keyed to a shaft and rotated at constant velocities. The
shape of a cam is referenced from a rotating frame as shown. The challenge is to determine velocity
and acceleration of the follower at different angular position of the cam.
Consider the velocity and acceleration of cam and follower. Three cases will be discussed:
(i) linear displacement of a flat surface follower, (ii) linear displacement of a roller follower, and
(iii) angular displacement of a rocker arm follower.
In each of these examples, point P is chosen as a point within a rotating xy reference frame.
Cam Drives – Linear Follower

In this case, the follower experiences linear motion, which is dictated by the geometry of the cam.
Follower displacement is important as the purpose of cam drives is to convert rotary motion into a
prescribed displacement to the follower. Once cam geometry has been set, based on the desired
follower displacement, it is important to know the velocity and acceleration of the follower.
Example The first case is illustrated in Figure 4.34. The objective is to determine the velocity
and acceleration of the follower. Calculations are based on the following input parameters.
r = 4 in.
R = 6 in. (radius of curvature of cam surface at point of contact)
ω = 2 rad/s
θ = 30
ϕ = 25
Velocity Analysis
The equation used to find the vertical velocity of the linear follower is based on
vA = vP + vA P 4 141
y
x
P
r ϕ
βˆ
ω
θ
αˆ
R
y = y(x)
vA/P aP aA
vA
vP
CC
(aA/P)n
(aA/P)t
Velocity polygon Acceleration polygon
Figure 4.34 Cam and linear follower.
The component velocities are
vA = vA β
The magnitude vA is unknown, but its direction is known. The velocity of point P is
vP = ωr − sin θ + ϕ α + cos θ + ϕ β 4 142

vP = 2 4 − sin 55α + cos 55β = − 6 55α + 4 59β
The angle ϕ depends on cam geometry, θ, is the independent variable.

The velocity of point A relative to point P can be visualized by the principle of kinematic inver-
sion, i.e. the relative motion is the same regardless of which link is fixed. In this case, we choose to
Cam Drives 193
fix the cam and let the fixture rotate CW. The direction of point A relative to point P is tangent to the
cam surface with direction to the right; therefore
vA P = αvA P 4 143
The vector components are:
vP = 4 2 − 0 8192α + 0 5736β = − 6 55α + 4 59β ips 4 144

vA P = vA P α ips 4 145
vA = vA β 4 146
By substitution
vA β = − 6 55α + 4 59β + vA P α 4 147
Therefore
vA = 4 59 ips and vA P = 6 55 ips
The velocity polygon is shown in Figure 4.34. These velocities apply only for the given values of
and θ.
Acceleration Polygon
The acceleration polygon is based on
where
vxy = vA P = 6 55α
axy = aA P
The relative acceleration components can also be visualized by use of the principle of kinematic
inversion where the cam is fixed and other linkages, including the frame, rotate clockwise. The
normal component of point A relative to P results from the curvature of the cam. The tangent com-
ponent of A relative to P is tangent to the cam surface at the contact point.
aA – magnitude is unknown, but its direction is along the vertical center line
(aP)t – zero, because ω is constant
(aP)n = ω2r = (2)24 = 16 ips2
(aA/P)n = ω2R = (2)26 = 24 ips2 (relative to rotating xy reference frame)
(aA/P)t – magnitude unknown, direction tangent to contact surface
2ωxvA/P = 2 (2) 6.55 = 26.2 ips2 (Coriolis component)
By substitution
n t n
aA = aP + aA P + aA P + 2ωxvA P 4 149
Giving
aA β = 16 − 16 cos 55α − sin 55β + atA P α − 24β + 26 2β
which in turn gives aA = − 10 9in sec2 and atA P = 9 18 in s2
The acceleration polygon is shown in Figure 4.34.
Cam with Linear Follower, Roller Contact

The cam chosen for this discussion contains two circular arcs (Figure 4.35). The instant of time
under consideration assumes the follower is engaged with the cam at the beginning of the circular
arc having radius, R. Rotation of the cam is clockwise. We wish to determine velocity and acceler-
ation of the follower using: (i) Coriolis’ law and (ii) Ritterhaus model (shown as a slider crank),
which is kinematically equivalent of the cam over the circular arc travel.
Example Numerical input to the example is
•• R = 1½ in.
r = ½ in.
(a) (b)
vD vD/C
vD/P
vD
D
r vC
vP
D
P, in xy frame
y
ϕ
C x
R
C
ω θ
ω 0
θ
0 aC
aD
0 (aD/C)n
(aD/C)t
aD
(aD/P)t
aP
ϕ
(aD/P)n CC
Figure 4.35 Cam with linear follower, roller contact. (a) (b) Ritterhaus model.
Cam Drives 195
•• OC = 1.75 in.
N = 10 rpm ω =
2πN
=
2π10
= 1 05 rad s
•• θ = 45
60 60
From the law of cosines, OD = 2.81 in.
Velocity Analysis – Rotating Reference Frame

With reference to Figure 4.35a, the velocity of point A (on follower) is determined from
vD = vP + vD P
We know the direction and magnitude of point P. Only the direction of the other two is known.
The distance OD is determined by
OD = OC cos θ + CD cos ϕ
where
OC sin θ = CD sin ϕ
When θ = 45 , then ϕ = 38.2 and OD = 2.81 in. The velocity of point P then becomes
vP = 2 81 1 05 = 2 95 ips
From the velocity polygon
vD = vP tan ϕ = 2 95 tan 38 2 = 2 33 ips
vP 2 95
vD P = = = 3 75 ips
cos ϕ 0 786
Acceleration Analysis – Rotating Reference Frame

The acceleration of the linear slider is determined by both the Coriolis approach and the Ritterhaus
model. The acceleration polygon based on the Ritterhaus model is shown in the figure. The Coriolis
approach is given below.
The controlling equation is
aD = aP t + aP n + axy + 2ωxvxy 4 150
aD = aP t + aP n + axy t
+ axy n
+ 2ωxvxy 4 151
where point P is located on the cam coincident with point D. The values of each component are
listed below.
aD – direction is along a vertical line, magnitude unknown

aP t = 0 – cam has constant angular velocity
aP n = ODω2 – direction known, from point D to point O
axy t – direction known (perpendicular to CD), magnitude unknown
v2D P – direction from point D toward C
axy n =
R+r
2ω × vxy = 2ω CDω – magnitude known, direction along line DC but away from C
Using the input numbers

2
aP = aP n = ODω2 = 2 81 1 05 = 3 1 in s2
3 752
axy n
= aD P n = = 7 03 in s2
2
2ω × vxy = CC = 2 1 05 3 75 = 7 88 in s2 4 152
Bring the components together and using the acceleration polygon
CC − anD P 7 88 − 7 03
aD = aP − = 3 1− = 2 01 ips
cos 38 2 0 786
atD P = 1 08 sin 38 2 = 0 67 ips2
The solution is shown graphically in Figure 4.35.
Velocity Analysis – Ritterhaus Model

The velocity of the piston (point D) is determined from
vD = vC + vD C
where
vC = OCω = 1 75 1 05 = 1 84 ips
vC cos θ = vD C cos ϕ
cos 45
vD C = vC = 1 66 ips
cos 38 2
vD = vD C sin ϕ + vC sin θ = 1 66 0 618 + 1 84 0 707 = 2 33 ips close agreement
Acceleration Analysis – Ritterhaus Model

The acceleration of point D is determined from
aD = aC + aD C
aD = aC n + aD C n + aD C t
aD − β = OC ω2 α sin θ − β cos θ − CD Ω2 α sin ϕ + β cos ϕ + CDα − α cos ϕ + β sin ϕ

− aD β = 1 36 α − β − 0 85α − 1 08β + α − 1 57α + 1 24β
The solution comes from the two unit-vector components

− aD = − 1 36 − 1 08 + 1 24α
0 = 1 36 − 0 85 − 1 57α
which give
α = 0 372 rad s2 and aD = 1 98 ips2 close agreement
Cam with Pivoted Follower

Consider the cam arrangement shown in Figure 4.36. In this case, the coincident point P is the cen-
ter of the disc follower. Point A is coincident with P but located on an imaginary extension (dashed
line) of the cam; member 2. In this example, we assume the cam rotates with constant velocity, Ω
Cam Drives 197
vA/P
vA x
L
vP
y A,P
ω
Ω R
ρ
t
aA/P
n = 19.7
aA/P
aPt
CC = 25.6
0
aPn = 12.1
aA = 23.6
Figure 4.36 Cam with pivoted follower, roller contact.
counterclockwise. The problem is to determine both angular velocity (ω) and angular acceleration
(α) of the follower arm for this position.
Example Parameters used in this example are
R = 5.4 in.
L = 6 in.
ρ = 4.5 in. (determined from the geometry of the cam)
Ω = 20 rpm = 2.09 rad/s
With the rotation reference (xy) frame attached to cam follower

vA = vP + vA P 4 153
vA = 5 4 2 09 = 11 3 ips
The direction of relative velocity, vA/P, is tangent to the contact surface (see velocity polygon). The
relative velocity can be visualized by use of the principle of inversion; fix the cam and imagine the
remaining portion of the mechanism move clockwise about the cam giving vP/A. The direction of
the velocity of point P is perpendicular to the pivoting arm. For the sake of simplicity, velocities vP
and vA/P are scaled from the velocity polygon as
vP 8 5 ips
vA P 9 0 ips
85
ω= = 1 42 rad s
6
The acceleration equation is
aA = atP + anP + aA P + 2ϖxvA P 4 154
where
aA P = atA P + anA P 4 155
Equation (4.154) is the basis for the acceleration polygon shown in Figure 4.36. The known vector
components are shown as solid vectors. The two unknown vectors are shown as dashed vectors.
Their intersection closes the polygon producing the angular acceleration of the follower arm.
Note that point P is the coincident point located on the y axis and point A is located on the imag-
inary “extension” of the cam.
The magnitudes of the acceleration components are determined from
aA = anA = Ω2 R = 2 092 5 4 = 23 6 ips2 4 156
2
anP = Lω2 = 6ω2 = 6 1 42 = 12 1 ips2 4 157
atP only direction is known
anA P = Ω2 ρ = 2 092 4 5 = 19 66 ips2 4 158
atA P only direction is known
aCC = 2ωvA P = 2 1 42 9 = 25 6 ips2 4 159
The direction of aCC is based on the cross product of ωxvA P following the right-hand rule. It is
parallel to radius of curvature, ρ.
vP
The angular velocity of the arm is determined by ω = . The unknowns in the polygon are atP
L
and atA P. These components are represented by the dashed lines in the polygon. The angular accel-
t
aP 50
eration of the follower is determined by α = = 8 3 rad s2. Its direction is opposite to the
L 6
assumed direction in the drawing.
Power Screw
Power screws are used to convert rotary motion of a screw to linear motion of a nut. Threads are
usually acme type for applications of high load. Force is also converted from torque to linear force
on the nut. A machine component is typically attached to the nut to achieve a desired motion.
The relation between rotary torque and liner force is
2πT = pF 4 160
where
T – torque, in-lb
p – pitch of threads on screw drive, in
F – axial force to be moved, lb
The equation is based on the input energy supplied by torque, T, over on complete rotation of the
screw and output energy of displacing the force, F, acting over one thread pitch. The power for moti-
vating screw drives can come in through the nut or through the screw. Power transmission is often
achieved by a stepping motor or by a direct drive to the screw itself. Stepping motors are commonly
used when computers control platform motion. The mass of movable platforms affects input power
according to
2πT t = p Ma t 4 161
where T(t) varies with time. M is mass and a(t) is its linear acceleration.
The efficiency of acme lead screws is 20–25%. Much of the input power is lost to friction in the
threads. Also, there is backlash in the threads and this limits its application.
Hydraulic Transmission of Power
Downhole motors are used in oil well drilling to: (i) increase mechanical power to drill bits and
(ii) control the direction of drilling. These motors are located within the drillstring near a drill
bit (see Figure 4.37). They are powered by drilling fluid, which arrives at these motors under pres-
sure and a rate of flow. These motors essentially convert hydraulic power to mechanical power in
the form of bit torque and rotary speed.
The power in the circulating system goes through several transformations starting with mechan-
ical power delivered to the mud pumps [11]. The power chain includes:
•• Diesel engine – mechanical

Generator – electrical
•• Electric motor – electrical

Mud pump – mechanical to hydraulic
• Positive displacement motor (PDM)/turbine – hydraulic to mechanical
The mechanical power output of downhole motors/turbines starts with mechanical power deliv-
ered by diesel engines.
2πTN TN
PM = = hp 4 162
33 000 5252
where
PM – mechanical power, hp
T – torque, ft-lb
N – rotary speed, rpm
Hydraulic power is related to fluid pressure and flow rate by

pQ
PH = hp 4 163
1714
Figure 4.37 Energy transformations in drilling

operations.
Drill fluid pump
Fluid flow
Drill pipe
Rotation
Drilling motor
Drill bit
Bit torque
where
PH – hydraulic power, hp
p – fluid pressure, psi
Q – flow rate, gpm
The exchange of power across a motor/turbine is illustrated in Figure 4.38.

Applying Bernoulli’s energy equation across the motor gives
p1 p ft-lb
= 2 + E out 4 164
γ γ lb
The unit of each term is “ft” or ft-lb per lb of fluid flowing thought the control volume. To convert
this equation to units of power
Δp
γQ = Eout γQ energy per time 4 165
γ
Figure 4.38 Hydraulic power converted to mechanical power. Q

p1 1
PDM – turbine
p2 2
Drill bit
T, N
In terms of hydraulic and mechanical power

PH in = PM out 4 166
ΔpQ TN
= 4 167
1714 5252
assuming no friction losses.
The efficiency of a motor is determined by
PM
η= out
4 168
PH in
The performance of PDMs is monitored through rig data. Speed (N) is directly related to flow rate
(Q) and torque (T) is directly related to pressure drop (Δp). These data are available to the driller by
means a flow meter and the standpipe pressure gauge.
It has long been recognized that only a small portion of total power (~8000 hp) at the surface is
used at the drill bit for boring into a formation. For example, under normal rotary drilling opera-
tions (N = 100 rpm, T = 1500 ft-lb), horsepower delivered to drill bits for boring into a formation is
2πTN
Pbit = 4 169
33 000
2π 1500 100
Pbit = = 28 6 hp 4 170
33 000
a very small portion of total rig power. By using PDMs (Figure 4.39), bit rotary speed increases to,
say, 400 rpm, creating four times the power at the drill bit.
2π 1500 400
Pbit = = 114 2 hp 4 171
33 000
This increase in drilling power comes at a cost, which must be weighed against rig rate ($/d) and
increase in rate of penetration (ROP).
Figure 4.39 Downhole drilling motors. Source: Baker Hughes.
Kinematics of the Moineau Pump/Motor

The Moineau principle has been applied to pumping applications, especially in petrochemical
industries. As a pump, the Moineau pump is a positive displacement tool delivering fluids at a rate
proportional to input rotary speed. It is especially adapted to pipelines and flowlines because of its
tubular length and diameter. Since the early 1970s, the Moineau principle has been used as a dril-
ling motor, which is driven by hydraulic power. In this case, power is taken from fluid power and
converted to mechanical power in terms of rotary speed and torque [12].
The kinematics of the rotor relative to the fixed stator can be analyzed as a planet pitch circle
(rotor) and an annular pitch circle as shown in Figure 4.40. Of interest is the motion of a point
on the rotor pitch circle. The motion of a point P on the pitch circle scribes a hypocycloid in space
relative to the fixed stator. The path of the hypocycloid is defined by the following xy coordinates.
The vector position R of point P is
R = ai cos ϕ + aj sin ϕ + ri cos θ − rj sin θ 4 172
Gathering the x and y components gives
R = r cos θ + a cos ϕ i + − r sin θ + a sin ϕ j 4 173

From which
x = r cos θ + a cos ϕ 4 174
y = − r sin θ + a sin ϕ 4 175
Stator pitch circle
R ω
O′
a r
x
O ϕ θ
P(x,y)
Rotor pitch circle
Figure 4.40 Hypocycloid geometry.
The velocity and acceleration of point P can easily be determined by differentiation. The angles,
and θ, are related through the motion of the pitch circles and either or θ is known. If, for example,
r 3
the ratio of the radii of the pitch circles is = then the hypocycloid path is as shown in
R 4
Figure 4.41. If a circle is attached to the rotor pitch circle with point P as its center, an outline
of a stator geometry is generated representing the shape of a 3 : 4 stator “gear.”
Using a planetary gear table, the relation between ϕ and θ is
ϕ 1 1
= = = −3 4 176
θ R 4
1− 1−
r 3
Therefore, ϕ = 3θ.
Mechanics of Positive Displacement Motors

The geometry of rotors and stators can be configured to achieve a desired relation between flow rate
Q and rotational speed N. Output speed and flow rate are related by N = CQ. The following corre-
lates the design variables that affect the value of C.
n
C= 4 177
q
where
n – turns of output shaft (drill bit)

q – corresponding volume through motor
Rotor pitch circle
0
–5 –4 –3 –2 –1 0 1 2 5
–1
–4
–5
Figure 4.41 Hypocycloid path of point P for a 3 : 4 lobed motor.
Since the rotor and stator interact as a planetary gear train, the number of turns (n) of the output
shaft (equal to number of turns of the rotor) per one complete circular travel of the rotor center is
a
n= −1 4 178
b
where
a – number of lobes on the stator

b – number of lobes on the rotor
For example, if a stator has two lobes and the rotor one lobe, n = 1, then the output shaft makes
one turn for one circular path of the rotor center. Or, if a stator has 10 lobes and the rotor 9, the
output rotation would be 1/9 rotation for each circular path of the rotor, a much smaller rotation.
Consider the volume throughput (q) for one circular movement of the rotor center. Let the two
cylinders in Figure 4.42 represent the pitch cylinders of a rotor and stator. Let the length of both
pitch cylinders be the length of one stage of the stator, Ls. For the sake of visualization, assume:
•• center lines O and O are fixed in space

both pitch cylinders are torsional flexible
•• top ends of both rotor and stator are fixed

lower end of the stator makes one complete turn causing the lower end of the rotor to make
a/b turns
• rotation of both rotor and stator are linear with distance from the top fixed end.
Based on the principle of kinematic inversion, the volume between the rotor and stator over sta-
tor length, Ls, is displaced down the motor whether point O is viewed as making a circular path
Figure 4.42 Pitch cylinders of rotor and stator. 0′ 0

Rotor
Stator LR
LS
0′ 0
about point O , or both O and O are viewed as being fixed and the two pitch circles mesh as a gear
pair. This volume is
q = ALs 4 179
where
A – cross-sectional area of the space between rotor and stator cavity

Ls – length of one stage as measured on the stator
Note, this volume is different for each isolated cavity entrapped between rotor and stator. It repre-
sents the total volume over one stage (LS).
Combining Eqs. (4.178) and (4.179) with Eq. (4.177) gives
n a
−1
C= = b 4 180
q ALs
Since a = b + 1 for PDMs
1
C=
bALs
and since Lr = ab Ls
1
C=
aALr
The relation between rotary speed and flow rate, N = C Q becomes
1
N= Q 4 181
bALs
and noting that bLs = aLr
1
N= Q 4 182
aALr
ALr is the total entrapped volume between rotor and stator over length Lr. In consideration of
Eqs. (4.181) and (4.182), the smaller the denominator, the larger the output rotational speed for
a given flow rate, Q. Also, the larger the number of lobes, the lower the output speed. That’s
why multilobed motors operate at low speeds and high torque. Note that N and Q are proportional,
which is characteristic of PDMs and pumps.
The dimensional unit of the motor constant, C, is output revolution per gallon pumped through
the motor. If, for example, 450 gpm flow rate through a one to two lobed motor produces an output
rotational speed of 450 rpm,
Q = 450 gpm
b=1
the volume of fluid between rotor and stator over one stage of the stator is
ALS = 1 gal
This 1-gal volume can be achieved with either a large area and small stage length or a small area
and large stage length. Almost any flow rate and rotational speed relationship can be established by
various combinations of b, A, and LS.
Consider a motor design where ALS = 1 gal and the number of lobes on the rotor is b = 3. The
motor constant in this case is
1 1
C= = 4 183
bALS 3
and
1
N= Q
3
A flow rate of 450 gpm produces an output rotational speed of 150 rpm.
A high-speed motor could be designed by making the product bALS small. The smallest that “b”
can be is one. A one to two lobed motor with either a small stage length or a small void area or both
would be a high-speed motor. A low-speed motor could be designed by making the product bALS
large. Therefore, low-speed/high-torque motors are multilobed.
If a motor is to be designed to deliver a certain output speed (RPM) at a given flow rate (gpm), the
motor constant, C, is known and becomes a design specification. Motor geometry is determined as
follows.
N 1
C= = 4 184
Q bALs
To balance the units when ALs has units of in3 and Q (gal)
144 12
C=
bALs 7 48
or
231
bALs = 4 185
C
where
b – number of lobes on rotor

A – flow area through motor, in.2
Ls – length of stator, in.
Motor dimensions can be selected once the motor constant, C, have been specified.
Example Assume that a motor is specified to have a rotational speed of 300 rpm at a flow rate of
400 gpm. In this case
N 300 3
C= = = 4 186
Q 400 4
The motor has to be configured so that
4
bALs = 231 = 308
3
The selection of A and Ls depends on tool size and length. The final dimensions are made through
a series of design trade-offs.
Note that by measuring A and Ls of a motor one can determine the motor constant, C, using
Eq. (4.183) and estimate the relationship between output speed and flow rate through the motor
by assuming no slipping.
The conversion of fluid pressure to mechanical torque is the result of the fluid wedge at the fluid
entry or first stage. Fluid density and flow rate have no effect on torque generated by the motor.
PDMs are typically made up of several stages, and each stage contains an entrapped volume of fluid,
which travels through the motor. Even though these motors have multiple stages, no work is done
by the constant volume cavities, and only the top stage converts hydraulic pressure to torque.
Multiple stages, usually about five, perform as dynamic seals, so reasonable pressure differentials
across these motors can be developed. Performance tests show that even these back-up stages allow
a certain amount of fluid leakage. The extra stages do, however, create friction and affect the overall
efficiency.
Since the hydraulic horsepower consumed by the motor is equal to the output mechanical output
power (assuming no losses)
ΔpQ = TN 4 187
So, output torque is
Q
T= Δp
N
T = bALs Δp 4 188
This equation shows that output torque increases with number of lobes on the rotor (and stator).
Torque also increases with stage length, which increases the wedging effect of fluid entering
the motor.
It has long been recognized that out of the total mechanical power (~8000 hp) required to operate
a drilling rig, only a small fraction is used at drill bits. For example, under normal rotary drilling
operations, bit horsepower is
2πTN
Pbit = 4 189
33 000
2π 1500 80
Pbit = = 22 8 hp 4 190
33 000
The output speed of a 1 : 2 PDM is in the range of 400 rpm. Corresponding output power is
2π 1500 400
Pbit = = 144 hp 4 191
33 000
Bit power is increased by a factor of 4. This increase in bit power comes at a cost because motors
cost (Cpdm) is added into footage cost calculations. However, PDM cost is offset by increased ROP.
Note that mud density is not a factor in torque or power generation in PDMs. This is not the case
in power generation of turbines.
References
1 Ham, C.W., Crane, E.J., and Rogers, W.L. (1958). Mechanics of Machinery. NY: McGraw-Hill.
2 Shigley, J.E. and Mitchell, L.D. (1983). Mechanical Engineering Design, 4e. McGraw-Hill Inc.
3 Wildhaber, E. (1961). Gears with Circular Tooth Profile Similar to the Novikov System, VDI Berichte,
No. 47, Germany (first introduced by Ernest Wildhaber in 1910).
4 Novikov, M. (1956). L. USSR Patent No. 109,750.
5 Stadtfeld, D.H.J. and Saewe, J.K. (2015). Non-Involute Gearing, Function and Manufacturing
Compared to Established Gear Designs, 42–51. Gear Technology.
6 Mitchner, R.G. and Mabie, H.H. (1982). The determination of the Lewis form factor and the AGMA
geometry factor J for external spur gear teeth. ASME J. Mech. Des. 104 (1): 148–158.
7 Timoshenko, S. and Goodier, J.N. (1951). Theory of Elasticity, 2e. McGraw-Hill Book Co. Inc. (see
torque of non-circular cross sections).
8 Myszka, D.H. (2013). Machines and Mechanisms, Applied Kinematic Analysis, 4e. Pearson –
Prentice Hall.
9 Shigley, J.E. (1961). Theory of Machines. McGraw-Hill Book Co, Inc.
10 Dareing, D. W. and Chung-Moon Chen, C.-M. (1973). J. Eng. Ind. Trans. ASME, 95, B (4): 1171–1177.
11 Dareing, D.W. (2019). Oilwell Drilling Engineering. ASME Press.
12 R.J.L. Moineau (1932). Gear Mechanism, U.S. Patent 1 892 217.
209
Friction, Bearings, and Lubrication
In 1951 Vogelpohl [1] estimated that about 1 3 to 1 2 of energy produced is consumed by friction.
Considering all automobiles, trucks, and other modes of transportation plus machinery used in
manufacturing, it is not surprising the level of energy lost to friction is large. Also, equipment wears
out because of friction and must be replaced; another economic loss. Because of this, mechanical
friction and wear became a focus of research during the early 1960s. A study conducted (1966) in
Great Britain by Jost [2] concluded that by following good design and lubrication procedures, about
500-million-pound sterling could be saved per year. Studies in Europe and America have reached
similar conclusions. The general area of friction and wear became known the science of tribology, a
Greek word meaning a study of friction, lubrication, and wear between moving surfaces.
Rolling Contact Bearings
Shafts must be supported on some type of bearing to accommodate applied loads and allow for shaft
rotation. There are two broad categories of mechanical bearings: rolling contact bearings and thick
film–lubricated bearings. Within these two bearing categories, there are various alternatives. In
both cases, bearings are selected to support applied loads with minimum friction and wear.
Roller contact bearings are used in nearly every industry as well as jet engines and rocket engines.
They are reliable and forgiving. Failure usually comes in the form of pitting, but even then, they still
perform. Types of roller contact bearings are (Figure 5.1):
a) Cylindrical roller
b) Tapered roller
c) Deep groove ball
d) Angular contact ball
There is a great deal of history on performance and a comprehensive data bank for making sta-
tistical predictions on their performance.
Table 5.1 lists coefficients of friction for various bearings and materials.
Equipment literally “wears out” because of wear. In most cases, friction and wear produce heat,
which accelerates wear and if left unchecked can destroy equipment. Temperature is a measure of
heat and a good indicator of friction and wear. The objective in design is to minimize friction and
thus minimize mechanical wear and heat.
210 Friction, Bearings, and Lubrication
Figure 5.1 Rolling contact bearings.
Table 5.1 Comparison of friction.
Coefficient of friction (typical)
Dry friction
Generic 0.3
Steel on copper 0.15
Teflon on anything 0.1
Thick film bearings

Journal bearings 0.001
Slider bearings 0.001
Roller contact bearings
Cylindrical roller 0.001–0.003
Tapered roller 0.002–0.005
Deep groove ball 0.0015–0.003
Angular contact ball 0.0015–0.002
Rolling contact bearings normally fail by pitting caused by cyclic stresses beneath the surface of
either the roller or race. The criterion for bearing failure is a 6 mm2 (0.01 in.2) area pit in the surface
of either bearing race. These bearings are forgiving in that they will still perform even with a pit in
the surface, and their performance can be monitored easily with vibration and noise sensors.
Advantages of rolling contact bearing over thick film–lubricated bearings are simplicity and reli-
ability. On the other hand, hydrostatic bearings can be designed to support roofs of sports arenas
with minimal space in comparison to ball bearings.
Rated Load of Rolling Contact Bearings

In a typical industrial setting, rolling contact bearings are lubricated and, in many cases, sealed. The
life of these bearings is related to a steady load by
q
C
L= 51
P
where
L – expected bearing life in millions of revolutions

P – bearing load; assumed to be steady, lb
C – basic load rating of a given bearing size and type, lb (corresponding to 106 rev)
q – exponent (3.33 for roller bearings, 3.0 for ball bearings)
This empirical equation is based on laboratory testing and operating performance. Performance
data are displayed on a log–log graph (Figure 5.2).
The exponent, q, is determined as follows:
q
L C
= 52
106 P
q
L C C
Log = log = q log 53
106 P P
log L − log 106
q= 54
log C − log P
Note that q is not the slope of the performance line in Figure 5.2, but it’s reciprocal. This can be
checked with the q values in Figure 5.5.
If “L” is defined as life in millions of revolutions, then Eq. (5.1) applies. The performance line is
drawing through the data such that 90% of measured data is above the line and 10% is below the
line. This means that bearing performance based on this line has a reliability of 0.9. Rolling contact
bearing failure is defined as fatigue pitting on either the roller or race. Even under this condition,
roller bearings can still function.
Commercial bearings are rated by the value of “C,” which corresponds to a life of 106 cycles. This
coordinate is for convenience only as bearing loads are not expected to perform as this load level.
Equation (5.5) is used to select bearings for specific applications.
1
C = PLq 55
where P and L are design requirements and C is the bearing “dynamic load rating” listed in catalogs.
C
105
Bearing force, P, lb
104
103 q
106 107 108 109

Bearing life, L, revolutions
Figure 5.2 Bearing load rating.

There are many types of rolling contact bearings as indicated earlier, each having special features
to satisfy specific operating requirements.
Catalog ratings are based on radial load. In general, axial loads may also be applied, in which case
an equivalent radial load, Fe, is used to select an appropriate bearing. The Anti-Friction Bearing
Manufactures Association (AFBMA) equation for finding the equivalent radial load for ball bear-
ings is1
F e = VF r 56
or
F e = XVF r + YF a 57
where
Fe – equivalent radial load

Fr – applied radial load
Fa – applied axial load
V – rotation factor (1 for inner race rotation, 1.2 for outer race rotation)
X – radial factor
Y – thrust factor
Values of X and Y depend on bearing geometry, size, and number of balls. Two columns are listed
for each (Table 5.2). The pair that gives the highest value for Fe should be used.
Once a specific bearing model is chosen, the required load rating, C, is determined by Eq. (5.1)
and used to select a bearing from supplier catalogs.
Example A medium 300K Series Timken ball bearing:
Bearing number 305K (p. D22 Timken Catalog, see Footnote 1)

25 mm bore (0.9843 in., ID)
62 mm (2.4409 in., ID)
Dynamic load rating – 26 600 N (6000 lb)
Specifications requiring this bearing support a radial load of 600 lb. What is the expected life of
this bearing whose inner race rotates at 1000 rpm? Using Eq. (5.1)
3 3
C 6000
L= = = 1000 × 106 rev 58
F 600
Table 5.2 Equivalent radial load factors.
X10 Y1 X2 Y2
Radial contact ball bearings 1 0 0.5 1.4

Angular contact ball bearings (shallow angle) 1 1.25 0.45 1.2
Angular contact ball bearings (steep angle) 1 0.75 0.4 0.75
In terms of time, this bearing will last
109 rev 1h 1 d 1 yr
T= = 1 9 years
1000 rpm 60 min 24 h 365 d
Next assume the 600 lb radial load is supplemented with a 200-axial load. In this case,
F e = XVF r + YF a 59
F e = XF r + YF a = 0 5 600 + 1 4 200 = 580 lb

Since 580 ≺ 600 lb, the first solution applies. However, if the axial load is 300 lb
F e = XF r + YF a = 0 5 600 + 1 4 300 = 720 lb
then
3 3
C 6000
L= = = 578 7 × 106 rev
F 720
578 7 × 106 rev 1h 1 d 1 yr
T= = 1 1 years
1000 rpm 60 min 24 h 365 d
The axial load reduces the bearing life by 9.6 months.
Effect of Vibrations on the Life of Rolling Contact Bearings

Radzimovsky and Dareing [3] predict the effect of dynamic forces on the life of rolling contact bear-
ings. Dynamic forces are assumed to be of the form
2πt
P t = Pm + Pv sin 5 10
T
where
Pm – mean value of load variation, lb

Pv – amplitude of variable component of load variation, lb
T – period of cyclic load, s
t – time, s
To derive the equivalent load, it may be assumed that if a certain load acts for a certain fraction of
the life duration, which the bearing expects to have under this load, then the same fraction of the
life of the bearing is consumed. For another load, having different magnitude, only the remaining
part of the bearing capacity may be utilized. Experience supports this assumption. Therefore, over
one load cycle
T
TPqe = P t q dt 5 11
0
1
T q
q
1 Pv 2πt
P e = pm 1− sin dt 5 12
T Pm T
0
From which
1
T q
q
1 2πt
β= 1 − α sin dt 5 13
T T
0
where
Pv
α=
Pm
Pe
β=
Pm
Pe is defined as a steady load that will cause fatigue failure in the bearing after the same number of
revolutions (or hours, if speed is constant) as a given unsteady load defined by Eq. (5.10).
When contact roller bearings are properly lubricated, q = 3.33 for cylindrical roller bearings and
q = 3 for ball bearings.
For ball bearings, the mathematical relation between these two parameters is
1
β = 1 + 1 5α2 3
5 14
For cylindrical bearings, the relationship must be determined numerically.

The effect of dynamic loading on bearing life is defined by the life reduction coefficient, K.
Lunsteady Le
K= = 5 15
Lsteady L
K and β are related as follows:

q
C
Le Pe 1
= = q 5 16
L C Pe
P P
1
K= 5 17
βq
The relation between K and β (Le and Pe) is illustrated in Figure 5.3.
C
105
Pe, Le
Bearing force, P, lb
104 P, L
103
106 107 108 109

Bearing life, L, revolutions
Figure 5.3 Effect of dynamic loading on bearing life.

Example Consider the following set of numbers:
C = 10 700 lb (load rating for deep groove radial ball bearing)

Pmin = 900 lb
Pmax = 5100 lb
N = 1800 rpm
Therefore
Pm = 3000 lb
Pv = 2100 lb
Expected bearing life under steading loading conditions is
10 700 3000 3 106 rev

L= = 421 hours
1800 60 rev h
From these input data,
Pm = 3000 lb
Pv = 2100 lb
2100
α= =07
3000
β = 1.202 (Table 5.3)
K = 0.576 (Table 5.3)
The equivalent load is

Pe = 1 202 3000 = 3606 lb
Table 5.3 Equivalent load and life reduction factors.
Roller bearings Ball bearings
α βr Kr βb Kb
0.0 1.000 0 1.000 1.000 1.000

0.1 1.004 2 0.986 1.005 0.985
0.2 1.019 9 0.936 1.020 0.943
0.3 1.045 8 0.861 1.043 0.881
0.4 1.980 3 0.773 1.073 0.806
0.5 1.121 6 0.683 1.112 0.727
0.6 1.168 2 0.596 1.155 0.649
0.7 1.218 7 0.518 1.202 0.576
0.8 1.271 9 0.449 1.252 0.510
0.9 1.327 0 0.390 1.303 0.451
1.0 1.383 5 0.340 1.357 0.400
The bearing life under this dynamic load condition is

Le = 421 × 0 576 = 242 h
Representing a 179/421 = 0.4252 or a 42.52% reduction in bearing life.
Effect of Environment on Rolling Contact Bearing Life

One of the most severe environments for rolling contact bearings occurs in oil well drilling. Since
the first roller bit patented by Hughes [4], the roller cone bit has gone through many improvements.
The design challenges are bearing life and cutter wear. Figure 5.4 shows a typical bit cone cross
section and the bearing arrangement. Initially, these bearings were open to drilling fluids, which
contain barite and bentonite, additives for density and viscosity. Both additives are abrasive.
The ball bearing carries little or no load but is there to hold the cone in place. The balls are
inserted through a hole that leads to the bearing cavity. Assuming weight-on-bit (WOB) is 45
000 lb, each of the three bearing legs supports 15 000 lb. This load generates lateral, thrust, and
moment loads within the bearing arrangement. The small pin or sleeve bearing is required to sup-
port the thrust load with some help from the ball bearing. Lateral loads are carried by the cylindrical
bearing and the sleeve bearing. This may be the best arrangement for the space, but in an industrial
application, cylindrical and sleeve bearings would not be use in parallel. Sleeve bearings require
radial space, which in this case causes nonuniform loading to the cylindrical rollers.
For many years, these bearings were unsealed and operated in a corrosive and abrasive environ-
ment caused by drilling fluid made up of: (i) a base fluid, usually water, (ii) barite to increase the
fluids density, and (iii) bentonite to give the fluid gel strength. Under this condition, bearing life is
about 8–10 hours.
In recent years, roller bearing life has increased to over 100 hours due to reliable dynamic seals
and lubricant reservoirs and improvements in bearing design.
Bearings in roller bit bearings have typically been open to the drilling mud plus high loads
required for acceptable rates of penetration (ROP). Also, the overall bearing assembly comprising
Cylindrical roller bearing

Nbit
Sleeve bearing
Nrel
1/3 WOB
Figure 5.4 Bearings mounted inside a roller cone.

Bit weight × 1000 lb
100
q = 2.8 (Air)
q = 1.4
(Drilling mud)
10
10 100 1000
Roller-bearing life, h
Figure 5.5 Life expectancy of unsealed bit bearing (N = 60 rpm).
cylindrical, ball and journal bearings within each bit cone is not conducive to prolonged bearing
life. The cylindrical roller bearing usually fails first because of misalignment and abrasion.
Tests conducted by King [5] on 7 7 8 in. W7R drill bits in barite mud show that q = 1.4, while the
same drill bit shows q = 2.8 in air (Figure 5.5). The values of q are strikingly different from those for
well-lubricated ball and cylindrical bearings. These tests were conducted under steading external
loading.
Effect of Vibration and Environment on Bearing Life

Dareing and Radzimovsky [6] expanded the analysis by King [5] to include other values of the expo-
nent, q, to cover the effects of drilling mud. The definitions of the dimensionless parameters are still
the same. Figure 5.6 relates, β, α, and q.
q = 3.33
1.4
1.35 3.0
1.3 2.5
Equivalent load ratio, β
1.25
2.0
1.2
1.75
1.15
1.5
1.1
1.2
1.05
1 q = 1.0
0 0.2 0.4 0.6 0.8 1
Load amplification ratio, α
Figure 5.6 Equivalent steady loads yielding same bit life.

The results are also presented as life reduction due to dynamic loading. A life reduction coeffi-
cient, K, is used to show reduction in bearing life caused by unsteady loading. K is the ratio of bear-
ing life corresponding loading to bearing life corresponding to steady loading.
Lunsteady
K= 5 18
Lsteady
This ratio can also be related to β through
1
K= 5 19
βq
Figure 5.7 and Table 5.4 relate K, α, and q.

As an example, consider the following case:
α = 0.25 (variable force component is 25% of the average force component)

Pm = 40 000 lb
q = 1.4 (bearing cavity open to drilling mud)
According to Figure 5.7,
Pe
β= = 1 01
Pm
This means that the effect of the 25% vibration amplitude is the same as increasing mean bit force
by 400 lb, an insignificant affect.
Alternatively, the life reduction coefficient, K, for this example is
Lunsteady
K= = 0 99 see Table 5 4
Lsteady
1 q = 1.0
0.9 1.2
1.5
0.8
1.75
Life reduction factor, K
0.7
2.0
0.6
0.5 2.5
0.4 3.0
0.3 q = 3.33
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1
Load amplification ratio, α
Figure 5.7 Bearing life reduction due to unsteady loading.

Table 5.4 Life reduction factors, K.
K, life reduction factors
α q = 3.33 3 2.75 2.5 2.25 2 1.75 1.5 1.25 q=1
0 1 1 1 1 1 1 1 1 1 1
0.1 0.980 968 0.985 222 0.988 113 0.990 713 0.993 018 0.995 025 0.996 729 0.998 128 0.999 219 1
0.2 0.927 925 0.943 396 0.954 105 0.963 878 0.972 657 0.980 393 0.987 035 0.992 542 0.996 874 1
0.3 0.851 058 0.881 058 0.902 392 0.922 291 0.940 545 0.956 939 0.971 265 0.983 334 0.992 965 1
0.4 0.762 356 0.806 452 0.838 831 0.869 853 0.899 051 0.925 925 0.949 965 0.970 653 0.987 486 1
0.5 0.671 915 0.727 273 0.769 301 0.810 744 0.850 884 0.888 889 0.923 827 0.954 689 0.980 427 1
0.6 0.586 415 0.649 351 0.698 694 0.748 77 0.798 723 0.847 458 0.893 634 0.935 668 0.971 763 1
0.7 0.509 306 0.576 369 0.630 517 0.686 978 0.744 934 0.803 212 0.860 209 0.913 826 0.961 448 1
0.8 0.441 753 0.510 204 0.566 93 0.627 543 0.691 435 0.757 575 0.824 35 0.889 389 0.949 387 1
0.9 0.383 566 0.451 467 0.509 022 0.571 849 0.639 658 0.711 744 0.786 777 0.862 508 0.935 369 1
1 0.333 895 0.4 0.457 135 0.520 65 0.590 592 0.666 663 0.748 039 0.833 038 0.916 869 1
Assuming the life expectancy for the steady case is 22 hours, then the life expectancy for the
unsteady force case is
Lunsteady = 22 0 99 = 21 78 hours
The effect of the unsteady bit force is a reduction in bit life of 13 minutes.
If the force variation is 100% instead of 25%, bit life would be reduced from 22 to 19 hours, a small
but significant decrease in bit life.
Now assume that the bearing cavity is for which the sealed bearing cavity has a wear exponent of
q = 2.5 and a 50 hours life under a steady bit load of 40 000 lb. A 40 000 lb-vibration variation gives
α = 1 and a K value of 0.52. Bit life under this condition is
Lunsteady
K= = 0 52
Lsteady
Lunsteady = 50 0 52 = 26 hours
In this case, bearing life has been cut in about half.

These calculations show that the life of sealed bearings is more sensitive to dynamic forces than is
the life of unsealed bearings. As bit bearing seals are improved, there is more incentive to eliminate
bit force vibrations.
Hydrostatic Thrust Bearings
There are three basic types of thick film–lubricated bearings:
1) Hydrostatic thrust bearings

2) Hydrostatic squeeze films
3) Hydrodynamic lubricated bearings
If properly designed and maintained, all three types of bearings would experience no wear. Some
bearings of these types have operated over a period of 30 years with no measurable wear. Friction-
generated heat and temperature are important considerations though. Each of these three types of
thick film–lubricated bearings are discussed in the remaining sections of this chapter.
Lubricant viscosity directly affects the performance of each bearing type. Viscosity is expressed in
units of centipoise (Systems International [SI] system) or reyn (English). Viscosities of several fluids
are given in Table 5.5.
Flow Between Parallel Plates

First, consider pressure-induced flow between two parallel plates (Figure 5.8). Pressure at point 1 is
greater than at point 2. The pressure difference drives the fluid from left to right. The velocity profile
across the film and flow rate through the film are determined as follows.
From the freebody of a differential fluid volume
p dyb − p + dp dyb − τ dxb + τ + d τ dxb = 0
dp dτ
= 5 20
dx dy
Table 5.5 Viscosities of common fluids.
cp reyn × 106
Honey 1500 218

Castor oil 1000 145
Glycerin 500 72
SAE 60 550 80
SAE 30 165 24
SAE 10 65 9.5
Kerosene 2 0.290
Mercury 1.5 0.217
Water 1.0 0.145
Gasoline 0.6 0.087
Air 0.018 0.0026
−7
Note: 1 cp = 1.45 × 10 reyn.
y
b
h
x
L
1 2
τ+dτ
dy p p+dp
dx
Figure 5.8 Flow between parallel plates.
Assuming the lubricant is Newtonian

du
τ=μ
dy
dp d2 u
=μ 2 5 21
dx dy
By integration
du 1 dp
= y + C1 5 22
dy μ dx
1 dp 2
uy = y + C1 y + C2 5 23
2μ dx
Imposing boundary conditions, u(0) = 0, u(h) = 0 gives
1 dp 2
uy = y − yh 5 24
2μ dx
If the pressure gradient is (+), flow is to the left. If the pressure gradient is (−), flow is to the right.
Now consider the flow rate through the parallel film of width b.
h
Q = b u y dy 5 25
0
h
b dp
Q= y2 − hy dy 5 26
2μ dx
0
3
bh dp
Q= − 5 27
12μ dx
Pressure gradient must be minus for positive flow to the right. This equation will now be adapted
to radial flow.
Fluid Mechanics of Hydrostatic Bearings

Hydrostatic bearings rely on external pumps for fluid film pressure. Fluid is circulated through
these bearings at specified pressure and flow rate to establish and maintain a specified operating
film thickness, h. A simple hydrostatic bearing is shown in Figure 5.9.
Modifying Eq. (5.27), for this case b = 2πr giving
πrh3 dp
Q= − 5 28
6μ dr
Integrating Eq. (5.28) and noting that p(r2) = 0, gives
6μQ r2
p= ln 5 29
πh3 r
Setting r = r1 gives
6μQ r2
p0 = ln 5 30
πh3 r1
The flow rate, Q, required to create pressure, p0 is
p0 πh3
Q= r2 5 31
6μ ln
r1
By substitution
r2
ln
p r = p0 r 5 32
r2
ln
r1
W Figure 5.9 Hydrostatic thrust bearing.
p0, Q
p0
p(r)
r1
r2
The load-carrying capacity is determined by integrating the pressure profile.

2
W = p0 πr 21 + 2πrp r dr 5 33
1
By substitution for p(r)

r2
2
ln
W= p0 πr 21 + 2πp0 r r
r 2 dr 5 34
1
ln
r1
2
2 r2
W = πp0 r 21 + r 2 r ln r dr 5 35
ln 1
r1
2
2 r2 r2 r2
W = πp0 r 21 + r 2 2 ln r + 4 5 36
ln 1
r1
2 r 22 r 21 r2 r2
W = πp0 r 21 + r2 − ln − 1 5 37
ln 4 2 r1 4
r1
Bringing these terms together gives
πp0 r 22 − r 21
W= r2 5 38
2 ln
r1
This equation relates supply pressure (recess pressure) to the load, W.
r2
2W ln
r1
p0 = 5 39
π r 22 − r 21
Equations (5.31) and (5.39) are the basis of hydrostatic bearing design. The expected load, W, on
the bearing would dictate the required pressure, p0, from the external pump. The corresponding
flow rate, Q, required to establish cavity pressure, p0, is determined from Eq. (5.31). The hydraulic
horsepower from the pump is determined from the product of p0 and Q. This type of bearing finds
application in the support and movement of heavy structures, such as roofs of sports arenas and
space telescopes.
Bearing design requires consideration of space and load. Unknowns in the design are flow rate,
delivery pressure, and viscosity.
Example Consider the following information
W = 10 000 lb (design load)

h = 0.006 in. (required minimum film thickness)
μ = 15.63 × 10−7 reyn (lubricant viscosity)
r2 = 6 in. (outer radius)
r1 = 1.5 in. (inner radius)
The problem is to determine the delivery pressure, p0, and flow rate, Q, and horsepower, P,
required to satisfy these design constraints. Using Eq. (5.39), p0 = 261.5 psi.
6
2 10 000 ln
15
p0 = = 261 5 psi
π 62 − 1 52
Using Eq. (5.31),
3
π 261 5 0 006
Q= = 13 65 in 3 s
−7 6
6 15 63 × 10 ln
15
Pump power required to deliver this pressure and flow rate is
P = p0 Q = 261 5 13 65 = 3570 in -lb s
1 ft 1 hp
P = 3570 = 0 541 hp
12 in 550 ft-lb s
The power input to the pump depends on mechanical efficiency of the pumping unit.
HHP hydraulic
η= 5 40
HP mechanical
HHP
HP =
η
This example shows that a very large force (10 000 lb) can be supported by a relatively low pres-
sure (261.5 psi). The required pumping power (0.541 hp) is relatively low. Analyses show that roll-
ing contact bearings would be very large (and costly) by comparison with hydrostatic
lubricating units.
We will now examine the best design for minimum power consumption.
Optimizing Hydrostatic Thrust Bearings

The above example addresses the pumping power required to support a 10 000 lb load while main-
taining a film thickness of 0.006 in., which is normal for this type of bearing. Calculations show that
a standard 3/4 horsepower motor and pump (above the 0.541 hp mentioned above) would be
required in this case.
Circular thrust bearings may also rotate and in this case, energy is lost in viscous friction. Both
pumping requirements and rotational friction are considered in optimizing the design of thrust
bearings.
In optimizing thrust bearings, there are two factors to consider [7]. The first is the radius of the
recess (r1) with respect to the bearing size (r2). The second is the operating film thickness. Both
affect energy consumption. The objective is to configure the bearing to minimize hydraulic power
and thus the amount of mechanical power required to support a given load.
Pumping Requirements
The formula for hydraulic horse power is HHP (or P) = p0Q. By substitution
p0 πh3
P = p0 r2 5 41
6μ ln
r1
Making a further substitution for p0,

r2 2
2W ln
πh3 r1
P= r2 5 42
6μ ln π r 22 − r 21
r1
This expression defines the pump power output necessary to support the required load. It is
rewritten as
r2
8π ln
r1 W 2 h3
Pp = 2 5 43
r1 2 A 12μ
1−
r2
2
W h3
Pp = C 5 44
A 12μ
where
A = πr 22
and
r2
8π ln
r1
C= 5 45
2 2
r1
1−
r2
This equation is plotted to show how pumping power is affected by ratio of recess radius (r1) to
outside radius (r2) (see Figure 5.10).
This plot shows that the coefficient, C, and pumping power are minimized when the radius ratio,
r1/r2, is about 0.5.
Friction Losses Due to Rotation

In many cases, hydrostatic bearings are used as thrust bearings to support axial forces applied to a
rotating shaft. The rotation produces friction in the film. Therefore, two energy components are
required to sustain a given film thickness. Power is required to pump through the film from the
80
70
60
Power constant, C
50
40
30
20
10
0
0 0.2 0.4 0.6 0.8 1
Radius ratio, r1/r2
Figure 5.10 Best recess for minimum pumping power.

recess to the outside. Power is also required to overcome friction brought about by film shear caused
by rotation. The total power required to operate a bearing at a given film thickness is the sum of
both the pumping and film friction losses.
Friction losses are due to fluid shear across the film. Local shear force in a thin film is the shear
force times a differential area, which in this case is a circular ring with radius, r, times width, dr.
dF = τ dA
U
dF = μ dA, dA = 2πr dr
h
U
dF = μ2π dr , where U = rω ips 5 46
h
The moment of all forces due to local shear stress is the sum or integral of all shear forces over the
thin-film area.
2πμω 2
dM = r dF = r dr
h
2
2πμω 3
M= r dr
h
1
The resulting moment is
πμω 4
M= r − r 41 5 47
2h 2
Friction power loss is
P f = ωM 5 48
πμω2 4
Pf = r 2 − r 41
2h
2
2πN πμ 4
Pf = r − r 41 in -lb s 5 49
60 2h 2
Total Energy Consumed

The total energy consumed by the bearing is the sum of pumping losses and rotational friction
losses.
PT = P p + P f 5 50
Example Consider the following example.
W = 100 000 lb
r1 = 5 in.
r2 = 8 in.
μ = 29.24 cp = 42.4 × 10−7 reyn
N = 750 rpm
Using Eq. (5.39), p0 = 774 psi and from Eq. (5.31), Q = 43.92 in.3/s or
1 ft3 7 48 gal 60 s
Q = 43 92 in 3 s = 11 41 gpm 5 51
144 12 in 3
1 ft3 1 min
Power lost to pumping is (from Eq. (5.39))
2
100 000 h3
Pp = 31 8 in -lb s 5 52
201 12 42 4 × 10 − 7
3
1000h 1 ft 1 hp
Pp = 5 53
64 6 12 in 550 ft-lb s
3
1000h
PP = hp 5 54
42 6
Power lost to rotational friction is
2π750 2
π42 4 × 10 − 7 4 142
Pf = 8 − 54 = in -lb s 5 55
60 2h h
142 1 0 0216
Pf = = hp 5 56
h 12 550 h
Bringing these numbers together into Eq. (5.50) gives the total power consumed by the bearing.
3
1000h 0 0216
PT = + hp 5 57
42 6 h
This equation is plotted in Figure 5.11, which shows that total energy required to operate this
hydrostatic bearing is minimized with a film thickness of h = 0.004 in. Film thickness, h, is con-
trolled by flow rate, Q, according to Eq. (5.31).
Coefficient of Friction
The coefficient of friction is determined as follows:
F
f = 5 58
W
60
50
Energy consumed, hp
40
30
20
10
0
0 0.002 0.004 0.006 0.008 0.01 0.012
Lubricant film thickness, in.
Figure 5.11 Optimizing film thickness against energy losses.

where
M
F= assuming h = 0 004 in 5 59
r ave
2πN πμ 4
M= r − r 41
60 2h 2
M = 45 41 in -lb r ave = 6 5 in
45 41
F= = 6 99 lb
65
Therefore
6 99
f = 0 0007
10 000
Squeeze Film Bearings
Hydrostatic squeeze films generate their film pressure from the motion of two approaching sur-
faces. Pressure builds in the film because the fluid cannot be squeezed out instantaneously. Pressure
distribution in the film depends on shape and size of the bearing.
Pressure Distribution Under a Flat Disc

Pressure distribution in a circular squeeze film is derived from continuity of viscous flow between
parallel surfaces (Figure 5.12). Flow within the squeeze film is also developed from Eq. (5.28). In
this case, Q varies with r according to
Q r = Vπr 2 5 60
where V is velocity of approach.
2πrh3 dp
Vπr 2 = − 5 61
12μ dr
The solution to Eq. (5.61) gives the pressure distribution within a circular squeeze film.
6μV
dp = − r dr
h3
3μV
p = − 3 r2 + C1
h
Applying the boundary condition, p(R) = 0,
3μV 2
pr = R − r2 parabolic 5 62
h3
The force–velocity relationship is determined by integrating the pressure across the circular face.
dW = p r 2 πr dr
Figure 5.12 Circular squeeze film.
dr
W R
R
3μV
W= 3 R2 − r 2 2πr dr
h
0
3πμR4
W= V 5 63
2h3
The average fluid pressure is
W
pave =
πR2
If the applied force, W, is constant, the velocity of approach, V, changes with the inverse of film
thickness, h3.
In this case Eq. 5.63 is written as
dh 2W 3
V= = h 5 64
dt 3πμR
and
3πμR dh
dt = 5 65
2W h3
giving a means for tracking film thickness vs. time.
The time interval between two film thicknesses (h1 to h2) is
3πμR4 1 1
t= − 5 66
4W h22 h21
0.012
0.01
Film thickness, in.
0.008
0.006
0.004
0.002
0
0 0.005 0.01 0.015 0.02 0.025
Time, s
Figure 5.13 Film thickness change with time.
where h1 is the reference film thickness. Figure 5.13 shows how film thickness changes with time
assuming
W = 10 000 lb
μ = 15.63 × 10−7 reyn
R = 6 in.
h1 = 0.01 in.
The film change from 0.01 to 0.005 in. takes about 0.015 second (see Figure 5.13).
Comparison of Pressure Profiles

Pressure profiles for both hydrostatic and squeezed films are based on
bh3 dp
Q= − 5 67
12μ dr
dp Q 12μ
= − 5 68
dr b h3
The flow rates for both are:
dp 1
Hydrostatic film Q = constant b = 2πr = − C1 5 69
dr r
dp
Squeeze film Q = πr 2 V b = 2πr = − rC 2 5 70
dr
For the same value of p0, the squeeze film carries the larger load. Both pressure profiles are shown
in Figure 5.14.
W V
Q, p0
p0
Figure 5.14 Comparison of hydrostatic and squeeze film.
Spring Constant of Hydrostatic Films

Hydrostatic films operate at a given film thickness, h, and have a certain rigidity with respect to h.
Starting with
6μQ r2
p0 = ln 5 71
πh3 r1
πp0 r 22 − r 21
W= 5 72
2 ln rr21

1
W = 3μQ r 22 − r 21 5 73
h3
dW 1
k= = − 9μQ r 22 − r 21 4 5 74
dh h
The minus sign simply means that the slope of the function, W(h), is negative. We only seek the
magnitude of the slope. Equation (5.74) shows that k h14. Stiffness of the film is very high for small
film thicknesses.
Damping Coefficient of Squeeze Films

Damping in squeeze films can be important for systems having high natural frequencies.
From Eq. (5.64)
W 3πμR4
c= = damping coefficient, force velocity 5 75
V 2h3
90
Damping coefficient, lb/ips × 10−3

80
70
60
50
40
30
20
10
0
0 0.002 0.004 0.006 0.008 0.01 0.012
Film thickness, in.
Figure 5.15 Damping coefficient vs. film thickness.
V
V
ri ρ
r0 ri
r0
Figure 5.16 Squeeze film geometries.
r dr
ρ
V
h0
r0
Figure 5.17 Dome-shaped squeeze film.

Using the previous numbers for a squeeze film
W = 10 000 lb
μ = 15.63 × 10−7 reyn
R = 6 in.
Damping coefficients were determined and plotted against film thickness (Figure 5.15).
Damping coefficients are lowest at the thicker films. For this example, c = 10 000 lb/ips at a
film thickness of 0.01 in. Recall, critical damping for an single degree of freedom (SDOF)
system is
ccr = 2 km 5 76
c
For an underdamped system, ζ = ≺1. For the above case
ccr
10 000
≺1
2 km
The spring constant, k, and mass, m, would have to be large for an underdamped system.
Other Shapes of Squeeze Films

This section develops pressure distributions for three other shapes.
1) Flat plate with a cavity

2) Washer
3) Spherical film
In each case (Figures 5.16 and 5.17), pressure profile is established by the flow equa-
tion (Eq. (5.61)).
Squeeze Film with Recess

The pressure distribution in the thin film is
3μV 2
pr = r − r2 5 77
h3 0
The pressure in the cavity is assumed constant
3μV 2
pi = r − ri 2 5 78
h3 0
The load-carrying capacity of this film is
W = W1 + W2 5 79
where
W 1 = πr 2i pi
r2
3πμV r 4o r4
W2 = 2πrp r dr = 3 − r 2i r 20 + i 5 80
h 2 2
r1
Combining expressions gives
3πμV 4
W= r 0 − r 4i 5 81
2h3
The average pressure is

W
pave = 5 82
πr 20
Squeeze Film Under a Washer

The boundary conditions for this problem are zero pressure on the inside and outside of the washer.
We anticipate a maximum pressure somewhere within the films. The location of this point is indi-
cated by radius ρ. Fluid on the outside of ρ flows outward. Fluid on the inside of ρ flows inward.
Flow equation to the right of ρ is
πrh3 dp
πr 2 − πρ2 = − 5 83
6μ dr
Flow equation to the left of ρ is
πrh3 dp
πρ2 − πr 2 = 5 84
6μ dr
Note that the same equation applies to both sides of ρ. The equation is used for outside flow know-
ing that fluid pressure is zero at r = r0. By direct integration
6μV r 20 r 2 r
pr = − − ρ2 ln 5 85
h3 2 2 r0
Using this equation and knowing that r = ri, p = 0 gives
r 20 − r 2i
ρ2 = ri 5 86
2 ln
r0
and
r
ln
3μV r0
pr = 3 r 20 − r 2 − r 20 − r 2i ri 5 87
h ln
r0
Note that maximum pressure occurs at r = ρ. The applied load, W, is determined from
dW = 2πr drp r
By integration
3πμV 2 1 2r 2i
W= r 0 − r 2i 1+ ri + r2 − r2 5 88
2h3 ln 0 i
r0
dh
V= 5 89
dt
3πμ 2 1 2r 2i dh
dt = r − r 2i 1+ ri + r2 − r2 5 90
2W 0 ln 0 i h3
r0
Spherical Squeeze Film

This type of film symbolized cases, such as a hammer face, coming down in a viscous film. Experi-
ments of dropping steel balls onto a film cause triangular shape indentation of the film surface indi-
cating high local pressure at the center location (Figure 5.17).
The fluid flow equation is the same as in the previous cases; however, in the spherical dome
example
h = h0 + ρ − ρ2 − r 2 5 91
Using the binomial expansion with ρ >> r gives
r2
h h0 + 5 92
2ρ
By substitution
3
r2
2πr h0 +
2ρ dp
πr 2 V = − 5 93
12μ dr
r dr
dp = − 6μV 3 5 94
r2
h0 +
2ρ
Integrating and applying the boundary condition at r = r0, (p = 0) gives
1 1
p r = 3μρV 2 − 2 5 95
2
r r 20
h0 + h0 +
2ρ 2ρ
Load W relates to the design variables by
r0
W= 2πrp r dr 5 96
0
Substituting for p(t) and integrating
r 20
h0 +
1 ρ
W = 6πμV ρ2 − 2 5 97
h0 r2
h0 + 0
2ρ
Nonsymmetrical Boundaries
The general case of fluid flow in films is shown in Figure 5.18. The square element has dimensions
of dx, dy, and film thickness, h.
Case 1 – First consider the trapped film is being squeezed by the top surface moving downward with
velocity V. Fluid is entering and leaving the control volume (film) in the x and y directions. The
downward movement of the top surface also causes fluid to enter the control volume from the
top. To satisfy continuity of flow
Qin = Qout
dQx dQy
Qx + Qy + V dx dy = Qx + dx + Qy + dy 5 98
dx dy
Recall
h3 dp h3 dp
Qx = − dy and Qy = − dx 5 99
12μ dx 12μ dy
By substitution, the continuity equation becomes
h 3 d2 p d2 p
V= − + 5 100
12μ dx 2 dy2
or
12μV d2 p d2 p
− = + 5 101
h3 dx 2 dy2
In polar coordinates, the Poisson equation becomes
∂2 p 1 ∂p 1 ∂2 p 12μV
+ + 2 = − 5 102
∂r 2 r ∂r r ∂θ
2
h3
This form is useful in arc-shaped geometries. For the circular squeeze film, pressure is independent
of θ, so the Poisson equation becomes
1 d dp 12μV
r = − 5 103
r dr dr h3
V Qx
y
Qy
dQy
Qy +
dy
Film thickness, h
x dQx
Qx +
dx
Figure 5.18 Two-dimensional fluid flow.

Figure 5.19 Squeeze film between journal and F, v

bearing.
h0 h
Bearing
θ
Journal
r sin θ
dh0
dh0 r
dp r 12μV 6μV
= − = − 3 r 5 104
dr 2 h3 h
the same expression as Eq. (5.61).
Case 2 – There are situations where fluid enters a film having noncircular boundaries as previously
discussed. Also, boundary pressures may not be uniform. The differential equation in this case is
derived from the control volume of Figure 5.18 as well. The continuity equation (Eq. (5.101)) still
applies except the film is assumed to be constant (V = 0).
∇2 p = 0 Laplace equation 5 105

There are many example solutions to both Poisson and Laplace equations [8]. Problems having
noncircular boundaries and various internal shapes can also be solved by finite difference methods.
Application to Wrist Pins

An important application of squeeze films is the lubrication of wrist pins within pistons.
A kinematic analysis of the slider crank mechanism shows that the relative angular motion of
the pin at the end of the connecting rod is a rocking motion. The wrist pin does not make a complete
revolution within the pistons; therefore, hydrodynamic lubrication is unlikely. However, the pres-
sure force on the piston is cyclic allowing the pin to be lubricated with squeeze film action during
the compression phase with the film recovering during the intake phase. The goal is to maintain a
required minimum film thickness during the compression stroke.
During the power phase, the lubrication film is squeezed when the pin (journal) moves into the
film displacing fluid (see Figure 5.19). The volume that is displaced through section h when the
journal moves dh0 is
d vol = br sin θdh0 5 106

Using
bh3 dp
Q= − 5 107
12μ r dθ
In this case, b is the length of the bearing. Film thickness is defined by
h = c − e cos θ
c e
h = mr 1 − n cos θ ; m= , n= 5 108
r c
where
h – film thickness at position, θ

c – radial clearance between journal and bearing (concentric position)
e – center distance between bearing and journal (see Figure 5.30)
r – radius of journal
n – eccentricity ratio
d vol dh0
Q= = br sin θ = brV sin θ see Figure 5 19 5 109
dt dt
Substituting Eqs. (5.109) and (5.108) into Eq. (5.107) gives
12μV sin θ
dp = − dθ 5 110
m2 r 1 − n cos θ 3
giving
12μV 1
pθ = 2 +C 5 111
m r 2n 1 − n cos θ
3
π 1
Assuming p = 0 at θ = , gives C = −
2 2n
6μV 1
pθ = 2 −1 5 112
m3 rn 1 − n cos θ
Local pressure in the film depends on V, n, and θ.

When n = 0 or the pin is concentric with the bearing, Eq. (5.111) is indeterminate. Applying L’Ho-
pital’s rule shows
12μV cos θ
pθ = n=0 5 113
m3 r
Equation (5.112) is determinate for other values of n. Note that when n = 1, the film thickness is
zero at θ = 0 and pressure is ∞, as expected.
The relation between bearing load and these parameters is determined from equilibrium
considerations.
dW = brd θp θ cos θ 5 114
When n = 0 (concentric case)
π
2
24μVb μVb
W= cos 2 θ dθ = 6π n=0 5 115
m3 m3
0
However, in general (using Eq. (5.113)) and being patient with the integration, we find that
μVb
W= K 5 116
m3
where
1
2 1+n 2
n
K = 12 tan − 1 + 5 117
1 − n2
3
2 1−n 1 − n2
At n = 0, K = 12[2tan−11] = 6π
The velocity of approach is
6π
n
0 1
dh
V= −
dt
But h = mr(1 − n) and dh = − mr dn. By substitution
dn
V = mr
dt
Wm3 1
V=
μb K
Substituting into Eq. (5.116) for V and separating variables gives
t2 n2 1
12brμ 1+n 2
2 n
dt = tan − 1 + dn 5 118
m2 W 1−n 1 − n2
3
2 1 − n2
t1 n1
Integration gives
1 n2
24brμ 1+n 2
n
Δt = 2 tan − 1 5 119
mW 1−n 1 − n2
1
2
n1
Table 5.6 Time history of film thickness, inches.
h1 (in.) h2 (in.) Δh (in.) havg (in.) navg K vavg (ips) Δt (s)
0.000 30 0.000 25 0.000 05 0.000 275 0.084 20 0.008 6 0.005

0.000 25 0.000 20 0.000 05 0.000 225 0.25 25.4 0.006 7 0.007
0.000 20 0.000 15 0.000 05 0.000 175 0.417 40 0.004 3 0.011 7
0.000 15 0.000 10 0.000 05 0.000 125 0.584 58 0.002 9 0.016 9
0.000 10 0.000 05 0.000 05 0.000 075 0.74 125 0.002 1 0.036 5
0.000 05 0.000 03 0.000 02 0.000 040 0.866 330 0.000 52 0.038
0.115
These equations were developed for a 180 bearing. They can be applied to 360 provided the
negative pressures on the opposite side are small compared to the positive pressures.
Example Consider a piston having wrist pin parameters
Radial clearance = 0.0003 in.

Bearing length = 1¼ in.
Viscosity, μ = 15.63 × 10−7 reyn
P (unit load) = 125 psi
Wrist pin diameter = 0.875 in.
The problem is to determine the time required to change film thickness from 0.0003 to 0.000 03
in. Compare this time with the time or the expansion stroke for the engine operating at 3000 rpm.
Actually W = W(t) and is not a constant. The calculating procedure is as follows: W1 will be
known. Assume t2 and find W2 from W(t). Calculate Δt and compare with t2. Repeat until t2
and Δt agree, etc.
For an engine running at 3000 rpm, the duration of the power stroke is 0.01 second. According to
Table 5.6, the time to reduce film thickness from 0.0003 to 0.000 03 in. is 0.104 second. This time is
greater by factor of 10 over the power stroke at 3000 rpm. The bearing should operate properly
under the given assumptions.
Thick Film Slider Bearings
Slider Bearings with Fixed Shoe

Three factors are required for pressure to develop in hydrodynamic lubrication films: (i) viscous
fluid, (ii) sliding velocity, and (iii) a fluid wedge. These three conditions are met with a slider bear-
ing (Figure 5.20). In this example, the shoe is fixed with a slight taper. The runner has a velocity of
U. The lubricant is a viscous fluid. The analysis in this case establishes the pressure profile over the
active part of the bearing and the load-carrying capacity of the bearing.
Assuming a Newtonian fluid and no side leakage, the Navier–Stokes equation for this 2-D case is
dp d2 u
=μ 2 5 120
dx dy
as before. However, in the slider-bearing case, the film is tapered.
y
Leading edge
Velocity
Trailing edge induced
α
h2
h1 h
x
U
x Pressure
induced
L
Figure 5.20 Flow within a slider-bearing film.
The velocity profile, u(y), across the film is now affected by the pressure gradient of the runner.
The velocity distribution across the film at any x location is modified by the boundary conditions; u
(0) = 0 and u(h) = −U. These boundary conditions yield
1 dp 2 y
u= y − yh − 1 − U 5 121
2μ dx h
This equation shows that velocity is the result of pressure-induced flow (first term) and velocity-
induced flow (second term). The two flow components are shown in Figure 5.20.
For continuity of flow at any x location
Q= ub dy 5 122
0
1 dp h3 1
Q=b − − hU 5 123
2μ dx 6 2
Since flow, Q, is the same at every location, x,
dQ
=0 5 124
dx
A simplified form of Reynold’s equation of lubrication is obtained by differentiating Eq. (5.203).
d h3 dp d 1
= − hU 5 125
dx 12μ dx dx 2
The solution to Eq. (5.125) defines the pressure distribution, p(x), for a given slider-bearing
geometry.
Pressure is assumed to be constant across the thickness of the film. Pressure distribution with x is
found by direct integration.
h3 dp 1
= − Uh + C1 5 126
12μ dx 2
At some point, x, the pressure gradient, dp

dx, is zero. The film thickness at this point will be defined
as h∗. Therefore, C1 = ½Uh∗. Equation (5.126) can now be expressed as
dp h∗ − h
= 6μU 5 127
dx h3
The symbol h∗ is a constant of integration. The other constant of integration comes from
Eq. (5.127). Both are established from pressure boundary conditions. Once the bearing film shape
has been defined, by h(x), Eq. (5.127) can be integrated with respect to x to determine the pressure
distribution, p(x) along the length the film. This function is used to relate bearing load to other
design parameters.
The expression for film thickness, h(x), is typically defined by
h x = h1 + αx 5 128
where
h2 − h1
α=
L
The mathematical steps required to integrate Eq. (5.127) are a bit unwieldy. They can be found in
references, such as Radzimovsky [9]. The resulting pressure function resolves into a simple
expression.
x x
μU 6α 1 −
px = − L L 5 129
L x 2
α − 2a a − α
L
where a = h1/L. In dimensionless terms, Eq (5.129) becomes
Note that Ref. [9] has the slider moving to the right and the shoe angle α with a negative slope.
The equations apply directly to Figure 5.20 provided the angle α is given a minus sign and a = h1/L.
Location parameter, x, is consistent with Figure 5.20.
In dimensionless terms
x x
pL 6α 1 −
= − L L 5 130
μU x 2
α − 2a a − α
L
A plot of Eq. (5.130) is shown in Figure 5.21.
The maximum value of film pressure occurs at x/L ~ 0.32 from the trailing edge, where x = 0.
45
40
35
pL/µU × 10−5
30
25
20
15
10
5
0
0 0.2 0.4 0.6 0.8 1
x/L
Figure 5.21 Pressure distribution in slider-bearing films.

Load-Carrying Capacity
The load-carrying capacity of slider bearings is determined by integrating under the pressure curve.
L
W= p x b dx 5 131
0
Again, the integration steps require patience but eventually lead to

1 a−α 2α
W = 6μbU ln + 5 132
α2 a 2a − α
If the applied load is expressed in terms of load per area (P) of bearing
W = PLb 5 133
Then, Eq. (5.132) is expressed in dimensionless terms as
−1
μU α2 a−α 2α
= ln + 5 134
PL 6 a 2a − α
This equation shows that the load capacity increases with a reduction in a = h1/L and a reduction
in angle, α.
Friction in Slider Bearings

Friction in slider bearings is derived from
du
τ=μ 5 135
dy
Starting with the velocity distribution across the oil film
du 1 dp U
= 2y − h + 5 136
dy 2μ dx h
The differential shear force within the film is
du
dF = μdA 5 137
dx
1 dp μU
dF = b 2y − h + dx 5 138
2 dx h
Friction force on the runner (y = 0) is
1 dp μU
dF R = b − h+ dx 5 139
2 dx h
Using Eq. (5.130) and integrating from x = 0 to x = L give
4 a−α 6
F R = − μUb ln + 5 140
α a 2a − α
Similarly, friction on the fixed shoe is
2 a−α 6
F h = μUb ln + 5 141
α a 2a − α
FR − 2α 2a − α ln a −a α − 3α2
f = = 5 142
W 3 2a − α ln a −a α + 6α
where
h2 − h1
α=
L
h1
a=
L
Center of Pressure
The location of the center of pressure is determined from the first moment of the pressure curve.
L
AW = xbp x dx 5 143
0
“A” is x distance to the applied load, W, measured from the leading edge. Width of bearing is “b.”
Using the previous expression for p(x) and integrating, we find that
a−α
A a − α 3a − α ln − 2 5α2 + 3aα
= a 5 144
L a−α
α α − 2a ln − 2α2
a
The location of the center of pressure does not depend on W, U, or viscosity, μ. It depends on the
angle of inclination, α, and the quantity, a, which contains the minimum film thickness, h1.
Example Consider a slider bearing having the following parameters.
Length (L) – 3.0 in.

Width (b) – 2.5 in.
Load (W) – 3450 lb
Velocity (U) – 100 in./s
Inclination (α) – (−)0.000 266 7 rad
Viscosity (μ) – 10 × 10−6 reyn
We wish to determine:
1) Minimum film thickness

2) Friction force on moving member
3) Coefficient of friction
4) Power loss in the bearing due to viscous friction
The unit pressure, P, is

W 3450
P= = = 460 psi
bL 25 3
Equation (5.134) is put in the form
PLα2 a−α 2α
= ln + 5 145
6μU a 2a − α
The left side is equal to

2
PLα2 460 3 − 0 000 266 7
= = 0 0163
6μU 6 10 10 − 6 100
So
a−α 2α
ln + = 0 0163
a 2a − α
The shoe angle is given as α = −0.000 266 7. The left side is implicit in “a.” By trial and error, “a” is
found to be 0.000 333.
h1 = αL = 0 000 333 3 = 0 001 in
The friction force on the runner under these conditions (using Eq. (5.140)) is 5.97 lb. The coef-
ficient of friction is
FR 5 97
f = = = 0 001 73
W 3450
Loss of power in friction is
FRU
H= = 0 905 hp
550 12
Slider Bearing with Pivoted Shoe

Equation (5.144) shows that the location of center of pressure depends only on minimum film thick-
ness, h1, and angle of inclination, α. It is independent of load (W), viscosity (μ), and speed (N). Ana-
lyses also show that load capacity is greatest and minimum film thickness is maximum when h2 ~ 2
h1. Under this condition, A 5 9L (Figure 5.22).
Therefore, if a bearing pad is designed to pivot on a point located 5 9 of its length from the leading
edge, we would expect the h1 and h2 (and α) to be established automatically giving greatest load and
W Pressure
profile
α h2
h1 5/9 L
x
U a c
Figure 5.22 Slider bearing with pivoted shoe.

film thickness. This feature makes the pivoted slider bearing very attractive since it automatically
sets the best angle, thus greatly reducing manufacturing costs.
It is useful in the analysis of pivoted slider bearing to use the following expression.
h2
m= −1
h1
Adjusting Eq. (5.104) to reflect this expression gives
6L2 2m
W = μUb ln 1 + m − 5 146
m2 h21 m+2
6μUbL2 1 2m
W= ln 1 + m − 5 147
h21 m2 m m+2
6μUbL2
W= KW 5 148
h21
where
1 2
KW = ln 1 + m −
m2 m m+2
Because KW is insensitive to values of m, some designers use an average value of KW = 0.025 to
simplify calculations.
Frictional Resistance
Expressing the friction force on the runner in terms of m gives
μUbL 4 6
F0 = ln 1 + m − 5 149
h1 m 2+m
μUbL
F0 = KF 5 150
h1
where
4 6
KF = ln 1 + m −
m 2+m
F0
f = 5 151
W
Combining terms gives
h1 1 K F h1
f = = Kf 5 152
L 6 KW L
where
KF
Kf =
KW
Exponential Slider-Bearing Profiles
The shape of the film does not markedly influence pressure distribution. In some cases, the choice
of film geometry greatly simplifies mathematical solution but still predicts useful film performance.
The exponential film (Figure 5.23) is a good example [10].
The film geometry of an exponential slider is defined by
h x = h1 esx 5 153
h2 = h1 esL 5 154
1 h2
s= ln 5 155
L h1
Note that h1, h2, and L define the exponential profile.
Pressure Distribution for Exponential Profile

Returning to Reynold’s equation
d h3 dp U dh
= − 5 156
dx 12μ dx 2 dx
The solution is determined directly by substituting for h(x), integrating with respect to x, satis-
fying boundary pressure conditions, and solving for p(x). An alternate approach is to define h∗
as the film thickness where the film pressure is maximum, dp dx = 0. In this case
h3 dp U ∗
= h −h 5 157
12μ dx 2
dp h∗ − h
= 6μU 5 158
dx h3
Substituting for h(x), integration, and imposing p(0) = 0 boundary condition
6μU − 1 h∗
px = 2 1 − e − 2sx + 1 − e − 3sx 5 159
h1 s 2 3h1
p (x)
y (x) = h1 esx
h1
h2
U x
Figure 5.23 Exponential slider-bearing profile.

2500
2000
Film pressure, psi
1500
1000
500
0
0 0.5 1 1.5 2 2.5 3 3.5 4
Distance from trailing edge, in.
Figure 5.24 Pressure distribution in an exponential slider bearing.
Using the boundary condition, p(L) = 0 gives
h∗ 3 1 − e − 2sL
= 5 160
h1 2 1 − e − 3sL
Consider an exponential slider bearing having the following dimensions.
h1 = 0.001 in.
h2 = 0.002 in.
L = 4 in.
μ = 24 × 10−6 reyn (165 cp – SAE 30 Oil)
U = 100 ips
Figure 5.24 shows that maximum pressure is achieved at x = 1.4 in. The load-carrying capacity of
the film is determined by integrating under the pressure curve.
Pressure Comparison with Straight Taper Profile

Pressure distribution for both straight and exponential profiles is given in Figure 5.25. The above
input data were used for both. It is interesting that profile shape does not have a significant effect on
film pressure distribution, a conclusion found by considering various profiles.
3000
2500
Film pressure, psi
Straight
2000
Exponential
1500
1000
500
0
0 0.5 1 1.5 2 2.5 3 3.5 4
Location (x), in.
Figure 5.25 Comparison of pressure profiles.

The load-carrying capacity is defined by
ℓ
W= p x b dx 5 161
0
Friction is
ℓ
F= τ x b dx 5 162
0
du
τ=μ 5 163
dx y=0
du 1 dp U
= 2y − h + 5 164
dy 2μ dx h
Coefficient of friction is
F
f = 5 165
W
Pressure Distribution for Open Entry

In this case, the slider-bearing shoe extends well beyond the previous length, L. The exponential
expression for the fixed shoe still applies
h x = h1 esx 5 166
where (for the above example)
1 h2 1
s= ln = ln 2 = 0 1733 5 167
L h1 4
However, in this case, the profile of the shoe extends beyond the distance, L (see Figure 5.26).
Imposing p(∞) = 0 boundary conditions on Eq. (5.160),
y y(x) = h1 esx
h1
U
x
Figure 5.26 Exponential slider bearing with front opening.

7000
6000
Film pressure, psi
5000
4000
3000
2000
1000
0
0 2 4 6 8 10 12 14
Distance from trailing edge, in.
Figure 5.27 Pressure distribution in an exponential bearing with open leading edge.
h∗ 3
= 5 168
h1 2
This occurs at an x location of
h∗
= esx 5 169
h1
1 h∗ 1
x= ln = 0 4055 = 2 34 in agrees with Figure 5 27 5 170
s h1 0 1733
h∗ 3
Replacing = in Eq. (5.159) gives
h1 2
6μU − 1 1
px = 2 1 − e − 2sx + 1 − e − 3sx 5 171
h1 s 2 2
3μU − 2sx
px = e − e − 3sx 5 172
sh21
Applying the same input data as before, except the leading edge is open and extends beyond 4 in.,
gives a pressure profile as shown in Figure 5.27. In this case, the maximum pressure (6100 psi)
increases dramatically over the previous case (2400 psi).
Exponential Slider Bearing with Side Leakage

The exponential profile also provides a direct mathematical approach accounting for side leakage.
A sketch of a 3-D control volume of a film is shown in (Figure 5.28). While the film geometry is the
same as for a 2-D film, the pressure distribution and film velocities are different. The pressure pro-
file varies as p = p(x, z).
Fluid velocity components also vary with x and y.
u = u x, y
w = w x, y
The pressure distribution within the film is determined by solving Reynolds equation accounting
for flow in the side (z) directions.
h2
a
x
a
h1
z L
Figure 5.28 Control volume of 3-D film.
∂ h3 ∂p ∂ h3 ∂p dh
+ = − 6U 5 173
∂x μ ∂x ∂z μ ∂z dx
Since fluid film pressure is symmetrical with respect to the x axis, it is reasonable to assume the
pressure function, p(x, z), can be represented by a series expansion containing even cosine terms.
∞
nπz
p x, z = X n cos 5 174
n = 1, 3, 5
2a
here Xn(x) is an unknown function of x. Note that Eq. (5.174) satisfies two of the boundary condi-
tions, i.e. zero pressure at z = ± a.
After substitution, Reynolds equation becomes
∞
dh nπ 2 nπz dh
h3 X n + 3h2 X n − h3 X n cos = − 6μU 5 175
n = 1, 3, 5
dx 2a 2a dx
If the right side of Eq. (5.175) is multiplied by one, which in turn is expanded into a Fourier series,
the x and z variables can be separated.
∞
dh nπ 2 dh 4 n−1 nπz
h3 X n + 3h2 X − h3 X n + 6μU −1 2 cos =0 5 176
n = 1, 3, 5
dx n 2a dx nπ 2a
From which
dh 3 nπ 2 dh 4 n−1
h Xn − h3 X n = − 6μU −1 2 5 177
dx 2a dx nπ
This is an ordinary differential equation with variable coefficients, which depend on the geom-
etry of the lubricating film. Its solution is not simple. Hayes [11], however, solved this equation for
the flat shoe.
The exponential shape of the bearing shoe, on the other hand, is amenable to a direct solution.
The exponential shape is defined by
h x = h1 esx 5 178
where
1 h2
s= ln
L h1
L – length of bearing

3 nπ 2 6μUs 4 n−1
Xn + X − Xn = − −1 2 e − 2sx 5 179
L n 2a h21 nπ
or
X n + 3sX n − α2 X n = βe − 2sx 5 180
where
nπ
α=
2a
6μUs 4 n−1
β= − 2 −1 2
h1 nπ
The total solution to Eq. (5.180) (including both particular and complementary solutions) is
β
X n x = e − 2sx An sinh λx + Bn cosh λx − e − 2sx
3
5 181
2s2 + α2
where
1
9 2
λ = s2 + α 2
4
Imposing boundary conditions of Xn(0) = 0 and Xn(L) = 0
β sinh λx − 1Ls
e − 2sx e 2 − cosh λL − e − 2sx − cosh λx
3 1
Xn x = 5 182
2s2 + α2 sinh λL
This expression completes the pressure function defined by Eq. (5.174).
pL β sinh λx − 1Ls nπ z
e − 2sx e 2 − cosh λL − e − 2sx − cosh λx
3 1
= cos
μU 1, 3, 5
2s + α
2 2 sinh λL 2 a
5 183
where
nπ
α=
2a
6μUs 4 n−1 μU
β= − 2 −1 2 = β
h1 nπ L
Ls 24 n−1 h2
β = − −1 2 where Ls = ln
h21 nπ h1
1
9 2 2
λ= s + α2
4
1 h2
s= ln
L h1
1
2 2
x 9 2 h2 nπ 2 L
λx = ln + 5 184
L 4 h1 2 a
Table 5.7 Pressure profile for an exponential bearing with side leakage.
pL
dimensionless
μU p (psi)
x z=0 z = 0.8a z=0 z = 0.8a
0.0 0 0 0 0
0.5 1.748 0.792 1048 475
1.0 2.456 1.029 1473 618
1.5 2.558 1.033 1535 620
2.0 2.318 0.923 1391 554
2.5 1.886 0.755 1131 453
3.0 1.337 0.551 802 331
3.5 0.704 0.311 422 187
4.0 0.0 0.0 0 0
As an application of Eq. (5.183), consider a slider bearing operating under the following
parameters.
h1 = 0.001-in.
h2 = 0.002-in.
L = 4-in. (bearing length)
2a = 4-in. (bearing width)
μ = 165 cp or 24 × 10−6 reyn (SAE 30 lubricant)
U = 100 ips
The shape of the fixed shoe is of an exponential shape as defined above.

The pressure distribution is listed in Table 5.7 giving pressure along the spine of the pressure
profile (z = 0) and for z = 0.8a.
For comparison, check the pressure distribution for the same exponential bearing without side
leakage (Figure 5.24). We see the pressure distribution along the spine of Figure 5.29 is substantially
lower than the maximum pressure distribution without leakage.
1600
1400
1200
z=0
Film pressure, psi
1000
800
z = 0.8 a
600
400
200
z=a
0
0 1 2 3 4
Bearing length, in.
Figure 5.29 Pressure distribution in exponential film.

Hydrodynamic Lubricated Journal Bearings
Testing by Towers [12] showed that two bearing surfaces can be separated by a thin film of
liquid under certain operating conditions of load, speed, and viscosity. His test device included
a fixed sleeve and a rotating shaft to simulate wheel bearings in locomotives. The sleeve had a
hole drilled in the top for injecting different greases. Once grease had been injected, Towers
plugged the hole with a wooden plug. During his test, he noticed that the wooden plug kept
popping out and oil would leak out of the hole. Following this observation, he measured film
pressure distributions around journal bearings and concluded that this pressure is great
enough to support bearing loads. Reynolds [13] applied the principles of fluid mechanics to
formulate differential equations, which predict fluid film pressure distributions. This pioneer-
ing work of both Tabor and Reynolds established the foundation for analytical advances,
which has provided a sound basis for the practical engineering design of thick film–lubricated
bearings.
Pressure Distribution Around an Idealized Journal Bearing

The mechanism for developing a thin film in journal bearings is the same as for slider bear-
ings. The variation in film thickness is shown in Figure 5.30. In this case, the runner is
replaced by the surface of the journal and the shoe is replaced by the outer bearing
surface.
The radial clearance between the bearing surface and the journal surface is designated as
“c.” The radius of the journal is r and the radius of the bearing is r + c. The distance between
the center of the bearing and the center of the journal is “e” and is called eccentricity of the
bearing.
Figure 5.30 Mathematical model of idealized

W
journal bearing.
θ Bearing
Journal
ϕ
e ω
h(ϕ)
r
hmin
x
U Journal surface
Bearing surface
2πr
Figure 5.31 Film thickness in journal bearing.
The film between bearing and journal surfaces is laid out in Figure 5.31. The distance x is related
to ϕ by
x=r
The film thickness, h, becomes
h ϕ = c + e cos ϕ
h ϕ = c 1 + n cos ϕ 5 185
where
e
n = c (eccentricity ratio, also attitude)
c – radial clearance between journal and bearing (concentric position)
e – center distance between bearing and journal
ϕ – independent variable
h – film thickness at position, i
r – radius of journal
The derivation of the pressure distribution around the journal starts as before
d 3 dp dh
h = 6μU 5 186
dx dx dx
Changing the independent variable to rϕ
d dp dh
h3 = 6μUr 5 187
dϕ dϕ dϕ
Integration gives
dp 1 k
= 6μUr 2 − 3 5 188
dϕ h h
where k is a constant of integration. Substituting for h gives
dp 6μrU 1 k
= 2 − 3 5 189
dϕ c2 1 + n cos ϕ c 1 + n cos ϕ
Reynolds [13] derived this equation in 1886 and found an approximate solution in the form of a
Fourier series. The solution was useful only for small eccentricity ratios. The exact solution was
later developed by Sommerfeld [14] and recaptured by Radzimovsky [9]. Following Sommerfeld’s
work, the expression that defines the pressure surrounding a journal for an idealized bearing is
6μrU n 2 + n cos ϕ sin ϕ

p ϕ =p 0 + 5 190
c2 2 + n2 1 + n cos ϕ 2
p(0) is film pressure at ϕ = 0 . This pressure can be determined by using a known pressure at the
inlet point, i.e. pi at ϕi. The location of the inlet point is not necessarily at ϕ = 0.
Example It is desired to find the pressure distribution around the journal.
Diameter of bearing d = 1.5 in.

Length of bearing b = 2.5 in.
Speed of journal N = 3000 rpm
Radial clearance c = 0.001 in.
Expected Tave of the film t = 165 F
Oil pressure at inlet pi = 45 psi
Location of inlet hole ϕi = 315
Attitude (n = e c, eccentric ratio) n = 0.8
Lubricating oil SAE 20 (μ = 2.15 × 10−6 reyn)
Note that p(0) can be determined from Eq. (5.190) by the known pressure, pi at ϕi. By substitution,
the pressure at ϕ = 0 is p(0) = 556 psi. Film pressure can now be determined at any ϕ from
2 + n cos ϕ sin ϕ
p ϕ = 556 + 691 5 191
1 + n cos ϕ 2
The locations of the maximum and minimum pressure are determined from
3n
cos ϕ = − 5 192
n2 +2
ϕmax = 155 5 pmax = 5476 psi
ϕmin = 205 4 pmin = − 4364 psi
The attitude angle, θ, is not given in the problem statement. It can be shown that for an ideal
journal bearing the attitude angle, θ, is 90 from the applied load, W.
The minus pressure over part of the film as shown in Figure 5.32 is ignored. Only the positive
pressure is considered in formulating the load capacity of the bearing film.
6000
4000
Film pressure, psi
2000
0
0 30 60 90 120 150 180 210 240 270 300 330 360
–2000
–4000
–6000
Angular position, ϕ °
Figure 5.32 Journal bearing pressure distribution.

The load-carrying capacity of journal bearings is based on a minimum required film thickness. The
relation between the design variables was established by Sommerfeld [14] and captured by Radzi-
movsky [9].
r 2 μN 2 + n2 1 − n2
= 5 193
c P 12π 2 n
where
n – eccentricity ratio (also attitude)

μ – viscosity, reyn
N – rotary speed, rev/s
P – average journal pressure, psi
r – radius of journal, in.
c – radial clearance, in.
The left side of this equation is called the Sommerfeld number in recognition of his pioneering
work on the lubrication of journal bearings. This function is shown graphically in Figure 5.33.
Example Consider a journal bearing with the following operating conditions:

μ = 0.145 × 10−6 reyn = 1 cp (water)
r = 1 in.
b = 2 in. (bearing length)
c = 0.010 in.
N = 1000 rpm
W = 1 lb (P = 0.25 lb/in.2)
r 2 μN
S=
c P
1 2
0 145 × 10 − 6 100
S= = 0 0967 01
0 010 0 25 6
From Eq. (5.193) or Figure 5.33

n = 0.185
e
c = 0 185
0.8
Attitude, n
0.6
0.4
0.2
0
0 0.05 0.1 0.15 0.2 0.25
Sommerfeld number, S
Figure 5.33 Sommerfeld number vs. attitude (eccentricity ratio).

e = 0.001 85
Minimum film thickness is
5/8″ dia
h0 = c – e = 0.010 − 0.00185
h0 = 0.0085
Minimum Film Thickness in Journal Bearings

Experiments conducted Karelitz and Keyon [15], McKee [16], and Stanton [17] have shown that the
smallest practical limit for fluid film lubrication is about 0.000 05 in. (1.27 μm). However, for large
commercial bearings the minimum film thickness should be kept above 0.0001 in. (2.54 μm). To
accommodate possible particles of contamination in the lubricant, elastic, and thermal distortions
in the bearing structure, the lower limit might possibly be set at 0.001 in. (25.4 μm).
Karelitz’s testing arrangement centered on rotating a shaft against a flat test block. Even with a
lubricant, the block showed wear as indicated by a dial gage. At some point in the experiment, wear
stopped, indicating a thin film had formed separating the two sliding surfaces.
Measurements on the wear cavity were:
a = 0.02–0.06-in.
θ = 3.7 –10.9
Width of the test block was 1 in.

Details of the circular wear profile in the test block were used to calculate the minimum film
thickness, which was determined to be approximately 0.00005 in. (1.27 μm). Measurements of
the surface asperities showed that the peak-to-valley distance of the asperities was of the same order
of magnitude.
The minimum allowable film thickness in journal bearings depends on factors, such as:
(i) smoothness or finish of the sliding surfaces, (ii) elastic rigidity of the bearing structure,
(iii) thermal distortion, and (iv) size of contamination particles in the lubricant.
Film thickness common in electric motors is hmin = 0.00075 in. (0.019 mm) at median speeds of
500–1500 rpm. For larger shafts, having running speeds between 1500 and 3600 rpm, the minimum
film thickness ranges between 0.003 and 0.005 in. (0.076–0.13 mm). For small automotive and avi-
ation engine bearings, with high finished surfaces, hmin ranges between 0.0001 and 0.0002 in.
(0.0025–0.005 mm). The smallest allowable film thickness in journal bearings depends on factors,
such as: (i) smoothness or finish of the sliding surfaces, (ii) elastic rigidity of the bearing structure,
(iii) thermal distortion, and (iv) size of contamination particles in the lubricant.
Friction in an Idealized Journal Bearing

Friction acting on a journal is due to fluid shear stress at the surface of the journal. Shear stress is
related to rate of shear as explained earlier. Before discussing Sommerfeld’s solution, consider fric-
tion torque in concentric journal rotation.
Petroff’s Law
Petroff [18] gives a baseline for friction in journal bearings showing that when the journal is con-
centric within a bearing, friction is due to simple shear in the film. In this case, shear stress on the
surface of the bearing is
du U 2πrN
τ=μ μ =μ 5 194
dr c c 60
where
τ – shear, psi
−7
μ – lb-s
viscosity, reyn in 2 (1 cp = 1.45 × 10 reyn)
N – rotary speed, rpm
c – radial clearance between journal and bearing, in.
U – surface velocity of journal, in./s
Friction torque produced by shear stress on the bearing surface is
4π 2 r 3 bμN μbr 2 N
T = τ 2πrb r = = 0 658 5 195
c60 m
Expressing bearing load in terms of average pressure,
W
P= 5 196
2rb
Representing the friction force by F = fW and T = Fr, then
T = rfW = rfP2rb = 2r 2 fbP 5 197

Equating torque expressions gives
μN r
f = 0 349 5 198
P c
r μN r 2
f = 2π 2 5 199
c P c
where
μ – viscosity, reyn
N – rotary speed, rev/s
P – average journal pressure, psi
c – radial clearance, in.
f – coefficient of friction
This expression for the coefficient of friction is known as Petroff’s law. It is based on the ideal
condition of concentric rotation of a journal inside a bearing.
Sommerfeld’s Solution
Sommerfeld developed journal friction for nonconcentric journals also based on fluid shear stress at
the journal’s surface. His expression for the coefficient of friction is
Fj
f = 5 200
W
Substituting for the friction force on the journal gives
r 1 + 2n2
f = 5 201
c 3n
Since the friction term on the left is a function only of attitude (n), it is also a function of the
Sommerfeld number. This relation is shown in Figure 5.34.
Example Consider a journal bearing having the following operating conditions.
W = 3000 lb
N = 2000 rpm
r = 2 in. (journal radius)
μ = 4 × 10−6 reyn (viscosity at 100 F)
c = 0.01 in. (radial clearance)
L = 2 in. (bearing length)
The objective is to determine the minimum operating film thickness and coefficient of friction.
Both unknowns depend on the Sommerfeld number.
r 2 μN
S= 5 202
c P
2 2
4 × 10 − 6 2000
S= = 0 0142 5 203
0 01 375 60
The eccentricity ratio can be determined from Eq. (5.201) or from Figure 5.34, from which n ~
0.85. Already we know this is a highly loaded bearing and expect a relatively small minimum film
thickness and high friction coefficient. Noting that
hmin = c − e = 0 01 1 − 0 85 = 0 0015 in
4
(r/c) f
3
Petroff
Sommerfeld
2
Lightly loaded
1 bearing
0
0 0.05 0.1 0.15 0.2 0.25
Sommerfeld number, S
Figure 5.34 Petroff friction and lightly loaded bearings.

The coefficient of friction for this bearing is

r 1 + 2n2 1 + 2 0 85
f = = = 1 0588
c 3n 3 0 85
from which f = 0.0053. Friction torque on the journal is
T j = fWr = 0 0053 3000 2 = 31 8 in -lb
Power loss, then, is

T jN 31 764 2000
Pf = = = 1058 8 in -lb s
60 60
or 0.16 hp.
The rate at which heat is being generated is
1058 8 1
Qf = = 0 1134 Btu s
12 778
Stribeck Diagram and Boundary Lubrication

While Sommerfeld’s work gave much needed insight into the performance of journal bearings, at
very high loads (small Sommerfeld numbers) a point is reached where the film thickness is of the
same order of magnitude at the surface roughness (peaks and valleys) and the theory of thick film
lubrication no longer applies. Under this condition, the peaks between bearing surfaces begin to
make intermittent contact.
Stribeck [19] conducted friction tests under conditions of heavily loaded bearings and found there
is a certain value of the Sommerfeld number for which boundary lubrication begins. The lowest
point in these curves marks the limit of thick film lubrication. Further loads increase friction, indi-
cated by the upward trend of the friction curve. Under this condition, friction and heat increase and
usually lead to bearing failure.
Ways to improve the performance of journal bearings focus on ways to move the minimum fric-
tion point to the left. Running in bearings at low loads and speeds has been recommended for new
machinery. Animal-based oil performs better than mineral oil.
Stribeck recognized this effect and studied ways to shift the minimum friction point farther to the
left. Factors that affect the location of the minimum point are run-in time and lubricant chemistry.
Recently, Barnhill [20] showed that a significant shift in the minimum friction point can be
achieved by use of ionic liquid (IL) additives.
Regions of Friction
Fuller [7] shows four regions of bearing friction (Figure 5.35). Coefficient of friction is plotted
against ZN
P , where
Z – viscosity (cp)
N – rotary speed (rpm)
P – force over projected area; length (L) × diameter (D) (psi)
This ratio is called the Hersey number. It is not dimensionless as the Sommerfeld number but is a
mixture of units for convenience.
Zone 4 Figure 5.35 Plot of coefficient of friction vs.

ZN/P.
Friction
Zone 3
Zone 2
Zone 1
fmin
Z N/P
1) Thick film region – Bearing surfaces are completely separated by a liquid film. When the lubri-
cant is free of abrasive particles, wear is prevented. The coefficient of friction can be as low as
0.001 or even less. This region begins at a Hersey number within the range of 30–40 [7]. This
number could be lower depending on surface finish and run-in time.
2) Thin-film lubrication region – This region represents the lower limit of complete separation by a
film. Film thickness varies from around 0.0002 in. down to 0.000 05 in. Bearing rigidity and sur-
face smoothness are important too.
3) Mixed-film region – Loading is so severe that complete surface separation cannot be achieved by
the lubricant. Most of the surface experiences rubbing without the benefit of complete separa-
tion of the peaks and valleys of the surface roughness. At best, lubrication is achieved by surface
separation by a few layers of fluid molecules. The coefficient of friction under this condition may
range from 0.02 to 0.08.
4) Boundary-film region – There is no fluid film because of either low viscosity or velocity or
because of high bearing load. Film thickness can only be described in terms of a molecule.
The coefficient of friction may range between 0.08 and 0.15.
To the right of the minimum point, bearings operate under thick film conditions. Under this con-
dition, a thick film completely separates the bearing surfaces. If the load on the bearing is increased,
the operating point moves to the left until thermal equilibrium is again established. Additional
increases in load will lower the coefficient of friction until the minimum point has been reached.
It is best to operate journal bearings at this low friction point. Further increases in load will cause
boundary-film lubrication causing friction and heat to increase.
Two factors can move the minimum friction point to the left. One is surface smoothness, and the
other is the “oiliness” of the lubricant. Surface smoothness is sometimes achieved by “running in”
bearings, i.e. initially running new machine at lower-than-specified operating speeds. This allows
the bearing surfaces to polish themselves.
In some designs, cylindrical roller bearings have been replaced with journal-type bearings. As
explained earlier, friction in journal bearings depends on the Hersey number, ZN P .
If the cylindrical bearing is replaced by a journal bearing, local stresses are greatly reduced, but it
is unlikely that a lubricant film is developed because of the high load and low relative rotation of the
cone on the pin.
Consider, for example, a drill bit journal bearing having the following conditions
Z – 105 cp (same as SAE W30 motor oil at 100 F)

Nrel – 200 rpm
P – 3000 psi
The Hersey number in this case is 7. Since this Hersey number is less than 30, journal surfaces are
at best boundary lubricated. Nonetheless sliding friction in the journal bearing is greatly reduced; a
reliable dynamic seal is essential, however.
Comparison of Journal Bearing Performance with Roller Bearings

For this comparison, we choose a 2 in. ID (inside diameter) journal rotating at 2000 rpm with an
applied load of 1000 lb.
Journal Bearing
Assuming a 2-in. journal with a 1 in. length gives P = 500 psi as the average bearing load and μ = 16
(10)6 reyn
r 2 μN
S= 5 204
c P
1 2
16 × 10 − 6 2000
S= = 0 0108
0 01 500 60
The eccentricity ratio can be determined from Eq. (5.193) or from Figure 5.33, from which
n = 0.91. Already we know that this is a highly loaded bearing, and we expect a relatively small
minimum film thickness and high friction coefficient. Noting that
hmin = c − e = 0 01 1 − 0 91 = 0 0009 in
The coefficient of friction for this bearing is
2
r 1 + 2n2 1 + 2 0 91 2 86
f = = = = 0 973
c 3n 3 0 91 2 79
From which the coefficient of friction is: f = 0.0097.
Journal bearing perform better at high speeds as shown at N = 4000 rpm.
Roller Contact Bearing (See Footnote 1)

Selecting a 1.9685 in. ID roller bearing having a dynamic load capacity of 18 300 lb gives
(using Eq. (5.1))
10
C 3
L= 5 205
F
10
18 300 3
L= = 1601 × 106 rev 5 206
2000
The amount of time to reach this level of revolutions at 2000 rpm is
1 min 1 h 1 d 1 yr
h = 1601 × 106 rev 5 207
2000 rev 60 min 24 h 365 d
h = 1 52 year
Ball Bearing (See Footnote 1)

Selecting a 1.9685 in. ID ball bearing having a dynamic load capacity of 15 600 lb gives
3
15 600
L= = 474 6 × 106 rev 5 208
2000
The amount of time to reach this level of revolutions at 2000 rpm is
1 min 1 h 1 d 1 yr
h = 474 6 × 106 rev 5 209
2000 rev 60 min 24 h 365 d
h = 0 445 year
Note
1 http://www.timken.com/catalogs.
References
1 Vogelpohl, G. (1951). Scientific Lubrication 3: 9.
2 Peter Jost, H. (1966). Lubrication (Tribology), A Report on the Present Position and Industry’s Needs,
80 pp. Her Majesty’s Stationery Office.
3 Radzimovsky, E.I. and Dareing, D.W. (1964). Influence of Load Variation Upon Life Duration of Roller-
Contact Bearings. Ukranian Technical-Economical Institute, Scientific Notes.
4 Hughes, B. (2009). Hughes Two–Cone Drill Bit historic mechanical engineering landmark 1909–2009.
American Society of Mechanical Engineers (ASME). The Woodlands, Texas.
5 King, G.R. (1959). Effect of fluid environment on rock-bit bearing performance. AAODC Paper.
6 Dareing, D.W. and Radzimovsky, E.I. (1965). Effect of dynamic bit forces on bit bearing life. Trans.
AIME, SPE J. 5 (4): 272–276.
7 Fuller, D.D. (1984). Theory and Practice of Lubrication for Engineers, 2e. Wiley.
8 Timoshenko, S. and Goodier, J.N. (1951). Theory of Elasticity, 2e. New York: McGraw-Hill (see
page 258).
9 Radzimovsky, E.I. (1959). Lubrication of Bearings. New York: Ronald Press.
10 Cameron, A. (1966). The Principles of Lubrication. Longmans.
11 Hayes, D.F. (1958). Plane sliders of finite width. Am. Soc. Lubr. Eng. 1 (2) (Lubrication Science &
Technology): 233–240.
12 Towers, B. (1883). First report on fiction experiments. Proc. Inst. Mech. Eng.(London) 34.
13 Reynolds, O. (1886). On the theory of lubrication and its application to mr. beauchamp tower’s
experiments, including an experimental determination of the viscosity of olive oil. Phil. Trans. Roy.
Soc. London 177 (Pt. I): 157–234.
References 265
14 Sommerfeld, A. (1904). Zur Hydrodynamischen Theorie der Schmiermitelreibung. Zeitschrift fur

Mathematik und Physik 50.
15 Karelitz, G.B. and Kenyon, J.N. (1937). Oil-film thickness at transition from semifluid to viscous
lubrication. Trans. ASME 59: 239–246.
16 McKee, S.A. (1928). The effect of running-in on journal bearing performance. Mech. Eng. 50: 528–533.
17 Stanton, T.E. (1922). On the characteristics of cylindrical journal lubrication at high values of
eccentricity. Proc. Roy. Soc. Lond. Ser. A 102.
18 Petroff, N. (1883). Friction in machines and the effect of lubrication (in Russian). Engineering Journal
St. Petersburg, Nos. 1,2,3 and 4.
19 Stribeck, R. (1902). Die Wesentlichen Eigenschaften der Gleit- und Rollenlager” [the main
characteristics of the sliding and roller bearings]. Zeitschrift des Vereins Deutscher Ingenieure [Journal
of the Association of German Engineers] 36 (B and 46): 1341–1348, 1432–38, and 1463–70.
20 Barnhill, W.C. (2016). Tribological testing and analysis of ionic liquids as candidate anti-wear
additives for next-generation engine lubricants. Master thesis. University of Tennessee.
267
Energy Consumption
Machines are designed to perform certain functions to achieve a desired outcome. Energy is con-
sumed in friction and other inefficiencies in the process. The overall efficiency of a machine is the
ratio of useful output to input. Industry relies on special machines to accomplish such tasks as
machining, moving objects, and pumping fluids.
There are many examples of energy consumption. The oil well drilling rig (Figure 6.1) and proc-
ess are used here to illustrate various avenues for energy consumption. The “end effect,” in this
case, is a well bore for reaching and producing oil from a subsurface hydrocarbon reservoir. The
path of a well bore can be vertically downward or along a preplanned well path extending down-
ward and laterally to reach a reservoir located at some lateral distance from the drilling site [1].
A considerable amount of mechanical power is required to drill an oil well. Diesel engines and
generators are mounted on skid units located some distance away from the central drilling activity.
This is arranged for safety and efficiency as electrical power is more transportable than direct
mechanical drives. Portable power units are necessary because drilling operations are usually
located in remote areas where public utilities are not available. The irony of the drilling operation
is while power units supplies some 8000 hp to drilling rigs, only a very small percentage (~50 hp) is
used to drill into rock to make a well bore. What, then, happens to the rest of the power and how is it
consumed? The answer lies in the subsystems that make up the total rig.
Subsystems of Drilling Rigs
There are five basic subsystems in a drill rig. Each subsystem can be viewed as a separate machine,
each having a power source, various means of transmitting power, and an end use.
• Hoisting – The hoisting system includes the derrick, crown block, traveling block, and a power-
driven drum. The crown block and traveling block provide a huge mechanical advantage for lift-
ing tubulars (~300 000 lb) in and out of boreholes. This pulley arrangement reduces the force on
the fast end (drum) of the cable to around 30 000 lb.
• Rotary drive and drillstring – The drillstring extends from the rig floor to the drill bit. Its two basic
tubular components are drill pipe and drill collars. This very long pipe transmits torque from the
surface to the drill bit and provides a conduit for drill mud. The rotary table transmits torque to
drill pipe by means of a square or hexagonal pipe called the Kelly, which allows the drillstring to
advance during the drilling process while maintaining rotary torque. A large amount of energy is
lost to friction along the drillstring.
268 Energy Consumption
Figure 6.1 Oil well drilling rig (Northern Alberta, Canada, 1963).
• Hydraulic or circulation – The hydraulic system is central to the rotary drilling method. Surface
mud pumps circulate drilling fluid down the drillstring to the drill bit for hole cleaning purposes.
Cuttings are carried back to the surface for inspection and analysis. Drill fluid is continually mon-
itored and maintained before being recirculated down hole for bit cleaning. A significant amount
of energy is lost to fluid friction over several thousand feet of pipe.
• Well control – The progression of tools available to defend against a well blowout are drilling
mud, annular preventer, and hydraulic rams. The annular preventer and hydraulic rams make
up the Blowout Preventer (BOP) stack. Flow rate returns are continually monitored for possible
formation fluid invasion into the well bore. When this happens, the well is taking a kick from the
formation and bad things can happen if this is allowed to continue. Well control measures are
taken to control formation fluid invasion.
• Mechanical/electrical power (~8000–10 000 hp) – Diesel is the prime source of power around a
drilling rig. These engines power electric generators which supply power to motors around the rig
for lifting pipe and tools in and out of the well bore, drive the rotary table, and power mud pumps.
While this level of power is needed to around a drilling rig, ironically, only about 50 hp is used to
break rock formations to advance the well bore.
Draw Works in Drilling Rigs
The structural framework of drilling rigs supports the hoisting system. The very top of a drilling rig
supports the crown block. This, plus the traveling block, makes up a block and tackle arrangement
with a huge mechanical advantage for pulling heavy pipe weight in and out of a well bore. Typical
pipe weight is around 300 000 lb. This weight is supported by the hoisting systems during regular
drilling and when the drillstring is pulled out of the well bore to replace bottom-hole equipment,
such as a drill bit. A typical length of a joint of drill pipe is 30 ft. A drillstring is disconnected in 90 ft
sections and leaned inside the derrick during a “tripping” operation. As a 90 ft section is removed,
the remaining portion of the drillstring is hung by slips during the disconnect.
The “end use” in this case is hoisting and lowering pipe and equipment in and out of a well bore.
This involves several steps, but still energy is required to elevate the drillstring mass. This energy is
not retrieved when lowering the drillstring back into the well bore. Much of it is lost in friction
between pipe and well bore, and some is lost in controlling the speed of the drillstring mass during
reentering steps.
Special pulley arrangement is a block and tackle arrangement (Figure 6.2). This arrangement has
a pulley (top) which is free to rotate about a fixed axis. This top pulley is the crown block. The mov-
ing pulley is the traveling block. This arrangement of pulleys has a huge mechanical advantage: pull
Crown block
Deadline
Fastline
ΔxA
A Five lines (each side)
FA
Δxb Traveling block
B
FB = W
Figure 6.2 Oil rig draw works.

force vs. lifting load. The weight of pipe suspended in a well bore may about 320 000 lb, while the
pull force on to a drum is 30 000 lb.
A band brake is used to hold and adjust drillstring position during various activities, such as
installing equipment and addition drill pipe. During tripping in and out of the well bore, the band
brake is used to set pipe in “slips” inserted within the rotary table. In each case, energy is lost in the
band brakes.
Block and Tackle Hoisting Mechanism

There is a huge mechanical advantage in this hoisting subsystem or energy transmission mechan-
ism. Assuming there are 10 lines between crown block and traveling block and applying the prin-
cipal of virtual work (Figure 6.2)
Δx A = 10Δx B 61
vA = 10vB 62
Equating work done by both forces
F A Δx A = F B Δx B 63
1
FA = FB 64
10
This means that a force of 30 000 lb is required to lift a 300 000 lb load. This load represents the
buoyed weight of the drillstring, friction along the drillstring, and possible force to dislodge stuck
pipe. The sole purpose of the draw works is to insert or pull pipe in and out of the well bore.
Spring Constant of Draw Works Cables

The axial spring constant of the draw works affects the dynamic behavior of the entire drillstring. It
is a boundary constraint in mathematical models. This spring constant is determined on site by
establishing a reference point on the Kelly. Raising the drill bit off the bottom transfers the bit force
to the draw works.
Measure how much the mark travels as the draw works picks up part of the weight on bit (WOB),
say 40 000 lb. Based on the number of strands of the wire cable passing around the crown and trav-
eling block, determine the amount the Kelly moves up. The difference between the theoretical
travel and the actual travel is the stretch of the pully system. The spring constant of the pulley
system is the ratio of the 40 000 lb and the stretch. The spring constant is typically in the range
of 40 000–50 000 lb/in.
Band Brakes Used to Control Rate of Decent

The mechanical arrangement for band brakes similar to the ones used in drilling rigs is shown in
Figure 6.3. The relation between T1 and T2 is
T 2 = T 1 eμθ
where
θ – angle of contact between belt and pulley
μ – coefficient of friction
FA Fast line
Impending motion
Drum
Brake torque
1 F
2
b a
Figure 6.3 Drum brake.
The relation between fast line force, FA, and T1 is
R T 2 − T 1 = rF A
By substitution
RT 1 eμθ − 1 = rF A
Applying force, F, adjusts and locks the brake. The tension in the brake band at point 1 is
r FA
T1 = 65
R eμθ − 1
The tension in the band at point 2 is
r eμθ F A
T2 = 66
R eμθ − 1
These equations show that T 2 T 1. The force, F, applied to the end of the brake handle is deter-
mined by equilibrium of the handle, i.e.
aF = bT 1 67
Combining Eqs. (6.5) and (6.7) gives
b r FA
F= 68
a R eμθ − 1
Rotary Drive and Drillstring Subsystem
Kelly and Rotary Table Drive

The rotary table is a means of transmitting torque to the drillstring (Figure 6.4). It is used to support
the freely hanging drillstring during tripping out or going into the well bore. During drilling, a spe-
cial bushing, called the Kelly bushing, is used to transmit torque to the drillstring. The Kelly is a
special pipe having a square or hexagonal external shape for the purpose of transmitting torque.
The Kelly slides through the rotary table as well bore deepens. Drilling fluid flows through the cen-
ter of the Kelly. A special sub, called a Kelly saver sub, is attached to the lower end of the Kelly to
protect the lower pin-end from multiple connections with every section of drill pipe.
The elements in this subsystem are (i) rotary drive, (ii) Kelly, (iii) drillstring, and (iv) bottom-hole
assembly (BHA). The BHA contains a drill bit but may also include other tools such as stabilizers,
motors and measurement-while-drilling (MWD) instruments. Again, we see this subsystem has a
power source, means of transmitting power, and an “end usage.” The end usage in this case is mak-
ing a hole in subterranean rock formations. Power is transmitted to a drill bit by rotating the entire
drillstring. Two type of rotating units may be used, rotary table or a top drive.
Much energy is lost to friction during drillstring rotation and pulling the drilling out of the
well bore.
Friction in Directional Wells

Directional wells may extend laterally several thousand feet. The well plan may call for defined
paths to be drilled to reach a specific target in a reservoir. Complex well paths can be drilled
Kelly (hexagon)
Kelly bushing
Drive pin
Rotary table
Bowl
Master bushing
Figure 6.4 Kelly being driven by a rotary table.

Table 6.1 Coefficient of friction.
Well #1 Well #2 Well #3a
Pick-up 0.28 0.31 0.40

Slack-off 0.27 0.31 0.41
Rotating 0.27 0.29 0.41
a
Well #3 had several severe dog legs in the lower portion of the build zone.
and navigated by rotary steerable tools (RSTs) and MWD monitoring. Friction forces created by
complex well paths can create high drillstring torque and pull out forces. These loads can some-
times exceed the capability of the surface equipment and strength of the drillstring. During direc-
tional drilling operations, drill pipe is usually limited by rotating torque and not by direct pull. One
practice is to allow rotary torque to reach 80% of make-up torque.
Johancsik et al. [2] measured axial pipe force and rotary torque directly below the Kelly. This
information, along with a computer model of drillstring friction, was used to determine the coef-
ficient of friction in three different wells (9790, 15 573, and 12 200 ft). The coefficient of frictions for
each well is given in Table 6.1.
Through experience, it has been determined that friction forces can be greatly reduced by rotating
the pipe and maintaining mud circulation during tripping. Top drives have been designed for this
purpose.
Top Drive
Modern drill rigs are now using top drives instead of rotary tables. In this case, power is supplied at
the top end of the drillstring by use of an electric motor. The Kelly and rotary table are eliminated;
however, the drillstring is still suspended in slips as before during pipe removal and insertion. The
swivel and motor are mounted to move along guide rails directing pipe in and out of the well bore
(Figures 6.5 and 6.6). Single joints of drill pipe are added as before, but pipe connections are made
by robotic mechanisms.
Pull-out or slack-off friction forces can sometimes be excessively high to the point of getting the
pipe stuck. The standard practice to minimize pipe sticking especially while tripping is to use top
drives to keep drillstrings rotating during tipping. Experience shows that pipe rotation, i.e. shearing
contact friction forces in the tangent direction, reduces contact friction forces in the longitudinal
direction.
Rotary power is typically transferred to drillstrings through the Kelly by means of the rotary table.
Power is delivered to rotary tables by electric motors in the form of torque and rotary speed. Fol-
lowing conventional methods of rotary drilling, a single 30 ft joint of drill pipe is added to a drill-
string by
•• setting the drillstring in the slips

breaking out the Kelly saver sub
•• attaching the Kelly onto a single 30 ft joint which has been place in the mouse hole
raising the Kelly assembly and attaching the new single onto the drillstring which is hanging in
the slips
•• removing the slips

circulating back to bottom and continue drilling
Hook-block Guide rails
Counterbalance
cylinder
S-pipe
Motor dolly
assembly
Standard
swivel
Swivel
links
Pressure
filter
Drilling motor and
transmission
assembly
Standard
Motor mud hose
alignment
cylinder
Main shaft Fluids
service loop
Pipehandler
Electrical
service loop
Figure 6.5 Schematic of top drive. Source: Courtesy NOV Inc.
The top drive replaces the Kelly, Kelly bushing, and rotary table. Power is transmitted to drill-
strings by an electric motor which travels with the top drive assembly. Triples or 90 ft sections
of drill pipe, which have been previously racked in the derrick, can be drilled down without inter-
rupting drilling. The ability to handle 90 ft sections of drill pipe while tripping in and out has dis-
tinct advantages over the conventional rotary table approach. Major advantages are
•• reduced connection time with both drilling and tripping;

ability to circulate and rotate out of the hole to reduce drag on the drillstring, reducing the
chances of getting stuck;
Figure 6.6 Top drive as mounted on a drilling rig. Source: Courtesy of NOV.
•• ability to drill through bridges and tight spots without picking up the Kelly (adding pipe);
helps control tool face orientation in directional drilling by capturing the trapped torque in drill-
strings over a 90 ft interval as opposed to a 30 ft interval – reduces downhole motor orientation
activity;
•• reduces the number of connections required by the rig crew and thus improves rig safety;
motor drive can be calibrated against tong torque gauges as a means to quantify rotary torque.
Top drives greatly reduce friction in high-angle directional drilling by allowing pipe rotation
while pulling pipe out of the well bore. They are essential in the drilling of extended
reach wells.
Five basic subsystems are
1) Drilling motor and swivel assembly
2) Guide dolly assembly
3) Pipe handler assembly
4) Counterbalance system
5) Top drive control system
A drilling motor and swivel assembly contain an electric motor, a gear drive, and a main shaft
which connects directly onto a standard swivel. This assembly is supported by the swivel and trav-
eling block and mounted on a guide dolly. The guide dolly is constrained to move up and down on
vertical guide rails that are rigidly attached to the rig. This assembly also contains an air brake capa-
ble of developing 35 000 ft-lb of static braking torque at the output shaft. One motor assembly can
develop 30 000 ft-lb of continuous torque at speed up to 175 rpm and can generate intermittent tor-
que as high as 41 500 ft-lb. This assembly can generate torque high enough to make up tool joints
(Figure 6.6).
A pipe handler connects and disconnects a stand of pipe from the drilling motor assembly.
Mechanical pipe handlers are designed to break out pipe in the derrick at any height with torque
capability up to 60 000 ft-lb.
The counterbalance system provides a 6 in. cushioned stroke to prevent damage to tool joint
threads while making or breaking connections. Each of these operations can be controlled from
a console located at the rig floor.
Drillstrings transmit mechanical rotary power from the surface to drill bits, serve as a conduit for
drilling fluid, apply force to drill bits, and affect hole direction. A major portion of drillstrings is
made up of drill pipe while the bottom portion (roughly 700 ft) is a heavier pipe called drill collars.
The bottom portion (~150 ft) of the drill collar section contains tools (drill bit, positive displacement
motors [PDMs], turbine, MWD, and stabilizers) called the BHA. Even though the total length of a
drillstring may be several thousand feet long, the BHA affects everything: rate of penetration, foot-
age cost, and well bore direction.
Drillstring Design and Operation

Drillstrings transmit mechanical rotary power from the surface to drill bits, serve as a conduit for
drilling fluid, apply force to drill bits, and affect hole direction. A major portion of drillstrings is
made up of drill pipe, while the bottom portion (roughly 700 ft) is a heavier pipe called drill collars.
The bottom portion (~150 ft) of the drill collar section contains tools (drill bit, PDM, turbine, MWD,
and stabilizers) called the BHA. Even though the total length of a drillstring may be several
thousand feet long, the BHA affects everything: rate of penetration, footage cost, and well bore
direction.
Drillstrings are subjected to many operational loads such as direct pull, torsion, and bending.
Each has a static and a dynamic component. They are designed from static load considerations fol-
lowing the state-of-the-art as given in API RP 7G [3]. Since 1960 much has been learned about the
dynamic behavior of drillstrings through downhole measurements and analytical studies. Vibra-
tions are known to affect the performance of downhole equipment. Well bore curvature and local
dog-legs also affect fatigue life.
Buoyancy
In practice WOB is established by slacking off the desirable bit force from the hook load at the sur-
face. This puts drill collars in compression while drill pipe is in tension. A basic question is how long
should the drill collar section be to prevent drill pipe buckling? This is important because of the
large difference in structural stiffness between drill pipe and drill collars. From a structural point
of view, stiffness attracts bending moment and resulting stresses can be damaging, especially under
stress reversals caused by drillstring rotation.
Hook Load
Applying Archimedes’ principle to drill pipe hanging freely from the draw works predicts a hook
load of
H = W −B 69
where
H – hook load
W – air weight of total string
B – weight of the drill mud displaced by the string as stated by Archimedes’ principle
In this equation, W is a body force and B is a surface force. Equation (6.9) can also be written as
B γm
H = W 1− = W 1− 6 10
W γ stl
H = WBF 6 11
where
W – air with of pipe
BF – buoyancy factor
Assuming a 12 ppg mud, the buoyancy factor is 0.817 as calculated:
7 48 gal
γ m = 12 lb gal = 89 76 lb ft3 6 12
1 ft3
so
89 76
BF = 1− = 0 817
490
The buoyed weight of a drillstring (including tools, etc.) is total air weight multiplied by the buoy-
ancy factor.
The magnitude of buoyancy forces can be quite high. For example, consider 5½ in. (19.2 lb/ft)
drill pipe having a cross-section area of 4.9624 in.2 If 10 000 ft of 5½ in. drill pipe hangs freely in
γ m = 12 ppg mud, the hydrostatic force pushing up at the lower open end is determined as follows.
p = Lγ m = 0.052 (10 000) (12) = 6240 psi
F = pA
F = 6240 lb/in.2 (4.9624 in.2)
F = 30 965 lb
It would appear that this force is great enough to buckle the drill pipe several times. However, this
does not happen.
Definition of Neutral Point

Assume the bottom end of an open-ended pipe (Figure 6.7a) is given a virtual sidewise displace-
ment. The static equivalence of the hydrostatic pressure forces on the pipe below section a–a is
(a) (b)
pA
0
a
a Hydrostatic
B n n
Lnp
W
W
WOB
Figure 6.7 (a, b) Stability of vertical pipe under hydrostatic loading.
represented by the vectors B and pA. The force B is equal to the weight of fluid displaced by the pipe
below section a–a. The vector pA is the product of the local fluid pressure and pipe area. The solid
vector, W, represents the air weight of the drill pipe below cross section a–a. W is a body force, while
B is a surface force.
Moments about point “0” show there is a restoring moment that always moves the pipe back
toward the vertical, assuming the density of the pipe is greater than the density of the fluid. In other
words, an open-ended drill pipe will not buckle from hydrostatic pressure alone regardless of well
bore depth.
Now assume the drill pipe is attached to the top of a drill collar section (Figure 6.7b) and the
compression at the top of the drill collars is equal to the local hydrostatic pressure. According to
the discussion above, the drill pipe will not buckle under this condition.
A good definition for the neutral point is the point in the drillstring, where compressive stress is
equal to the local hydrostatic pressure. If the hydrostatic pressure point is located within the drill
collars, the drill pipe will not buckle. Drill pipe will buckle when the internal force at the lower end
reaches a critical level, which is somewhat higher than the hydrostatic force level.
The location of the neutral point depends on bit force (WOB). If bit force is zero, the neutral is
located at the drill bit. As bit force is increased, the neutral point moves up the drill collars.
In practice, the neutral point is kept within the drill collars as a safety measure to avoid buckling
the drill pipe. The relation between WOB and distance to neutral point is determined by
WOB = wLnp − wm Lnp 6 13
or
wm
WOB = wLnp 1 − 6 14
w
γm
WOB = wLnp 1 − 6 15
γ
where
γ – steel density (490 lb/ft3 or 65.5 ppg)
γ m – drilling mud density (ppg)
The distance to the neutral point (or point of hydrostatic compression) from the drill bit is deter-
mined by
WOB
Lnp = 6 16
wBF
Note the location of the neutral point does not depend on collar length, only WOB.
Drill collar length is generally computed by
Lnp = 0 85Lc 6 17
which means that the neutral point is 85% of drill collar length. The extra length represents a safety
factor against drill pipe buckling.
WOB
Lc = 6 18
0 85wBF
Basic Drillstring: Drill Pipe and Drill Collars

Consider a simple drillstring made up of drill pipe and drill collars. If the drillstring is hanging freely
from the derrick, the neutral point is located at the bottom of the drill collars. As WOB is added, the
neutral point moves up the string and reaches the top of the drill collar section when
WOB = w1 L1 − A1 L1 γ 6 19
Since A1L1γ = wfL1 (where wf is weight per length of fluid displaced by the collars),
WOB = w1 − w f L1 = w1 L1 BF 6 20
Additional bit forces move the neutral point up into the drill pipe. Letting (x) represent distance
from top of drill collars to the neutral point, then
WOB = w1 L1 + w2 x − wf 1 L1 − wf 2 x 6 21
WOB = w1 L1 BF + w2 xBF 6 22
WOB = w1 L1 + w2 x BF 6 23
The neutral point location from the bottom of the drill collars is Lnp = L1 + x.
Physical Properties of Drill Pipe

Dimensions of three sizes of standard drill pipe are listed in Table 6.2. Each of the listed pipe sizes
are made in different weights as dictated by the inside diameter (ID) dimensions. Wall thickness for
each can easily be determined as well as the weights (lb/ft) using the density of steel (490 lb/ft) and
cross-sectional area. However, pipe weight is expressed in terms of nominal weight which includes
the weight of the connectors. For example, 5 in. (19.50 lb/ft) drill pipe has an ID of 4.276 in. and a
plain tubular weight of 17.93 lb/ft. Drill pipe is identified by size, weight, grade, and class [3].
Table 6.2 Dimensions of standard drill pipe.
OD (in.) ID (in.) Plain weight (lb/ft) Nominal weight (lb/ft)
41 2 3.958 12.24 13.75

3.825 14.98 16.6
3.640 18.69 20.00
3.50 21.36 22.82
5 4.408 14.87 16.25
4.276 17.93 19.50
4.000 24.03 25.60
51 2 4.892 16.87 19.20
4.778 19.81 21.90
4.670 22.54 24.70
•• Size – Nominal outside diameter of pipe

Weight – Weight per foot of drill length, including weight of connections
•• Grade – Yield strength of pipe material (E75, X95, G105, S135, numbers refer to yield strength)
Class – New, premium, used
Load and torque capacity of the four grades can be calculated as shown below. Considering 5
(19.50) and grade E75 drill pipe, the pull capacity is 395 595 lb.
Pull capacity is determined as follows:
F yld
σ yld = 6 24
Area
π 2
Area = 5 − 4 2762
4
F yld = 5 2760 75 000
F yld = 395 595 lb
Shear strength, τyld, is a determined from
τyld = 0 577σ yld 6 25
which is based on the von Mises energy of distortion criteria of failure. σ yld is an experimental value,
while τyld is a calculated value.
The maximum allowable torque is determined as follows:
T yld
τyld = 6 26
S
J
S= torsion section modulus
c
T yld = 11 415 75 000 0 577
T yld = 493 984 in -lb
T yld = 41 165 ft-lb checks with API Tables
Tension and torque limits for the other three grades (X95, G105, S135) can be determined by the
above equations as well.
Selecting Drill Pipe Size and Grade

Drill pipe is subjected to various types of loads during any drilling operation. Common loads are
direct pull, torque, bending, and internal pressure. Drill pipe selection (size, grade) is typically
based on direct pull type loading with a factor of safety to account for uncertainty. Other types
of loading, such as dog-leg bending, should be checked for stress reversal type fatigue, especially
if the dog legs are severe.
Normally, drillstrings are designed to an allowable load of 90% of yield strength of drill pipe, so
the allowable load is
H a = 0 9H yld 6 27
where
Hyld – direct pull causing material yielding
Ha – maximum allowable pull force, i.e. buoyed weight of drillstring plus margin of overpull (MOP)
The corresponding factor of safety (FS) is
H yld
FS = = 1 11 6 28
Ha
The level of the safety factor (FS) is subject to modification depending on application. High FS
means higher weight and cost, while low FS could lead to unexpected failure. Factors of safety are
used to cover uncertainty in the design, such as material properties, applied loads, and accuracy of
stress models.
The allowable pull force relates to hook load and MOP by
H a = H + MOP 6 29
where
H – working hook load (based on buoyed weight of drillstring at maximum depth)
MOP – margin of overpull, in case of stuck pipe or excessive friction
The allowable load (Ha) is the basis for selecting drill pipe size and grade for a given class of pipe.
Bringing these equations together gives
H a = 0 9H yld = wp Lp + wc Lc BF + MOP 6 30
This equation will be applied to the following two examples.
Select Pipe Grade for a Given Pipe Size
Example As an example, consider the following:

Total depth – 14 000 ft
Mud weight at total depth (TD) – 15 ppg
MOP – 100 000 lb
Drill pipe – 5 (19.50); Premium class

Drill collars – 63/4 in. outside diameter (OD) × 213 16 in. ID, 750 ft
The problem is to select the lowest grade of drill pipe that will reach TD and accommodate the spe-
cified MOP.
The static hook load (working hook load) at TD is
H = wp Lp + wc Lc BF 6 31
H = 19 5 13 250 + 100 750 0 771
H = 257 032 lb
The selected grade must have the strength to support
H a = H + MOP 6 32
H a = 257 032 + 100 000
H a = 357 032 lb
To account for the FS pipe grade selection must satisfy
H yld ≥ 1 1 357 032
H yld ≥ 396 702 lb
Grade X95 (Premium Class) has a yield load limit of 394 612 lb. Grade G105 has a yield load limit
of 436 150 lb. Grade G105 would be the proper choice for drill pipe grade because 436 150 is greater
than the required 396 702 lb yield strength. In this case, the FS increases from FS = 1.1 to 436 150/
357 032 = 1.22.
Determine Maximum Depth for Given Pipe Size and Grade
Example Now consider a situation requiring maximum depth capability of a given pipe size,
class, and grade. Other specifications are
Drill pipe size – 4½ in. (16.6 lb/ft), Grade E75, Premium Class
MOP – 57 000 lb
Drill collar – 6½ in. OD × 2½ in. ID, 700 ft long
Mud weight – 12 ppg
Recall
H a = 0 9H yld
H a = H + MOP
H = 0 9H yld − MOP 6 33
The maximum allowable depth is based on the allowable hook load, H. From the API RP 7G
Standard [3], the pull strength for the specified pipe Grade and Class is
H yld = 260 165 lb
The working hook load, H, is

H = 0 9 260 165 − 75 000 = 159 149 lb
Maximum hole depth is determined from
H = wdp Ldp + wc Lc BF 6 34
where
H = 159 149 lb
BF = 0.817
wdp = 16.6 lb/ft
wc = 96 lb/ft
Lc = 700 ft
Substituting these numbers into the above equation gives Ldp = 7687 ft. The maximum depth
capability is
TD = Ldp + Lc = 8 387 ft 6 35
Roller Cone Rock Bits

There are many different types of roller cone bits and each has unique features and benefits. They
can be divided into milled tooth bits and tungsten carbide insert cutters. Bearing performance has
been extended over the years through design and seal improvements.
Three components of roller cone bits are cutters, bearings, and bit body. The cutters are fixed to
cones which rotate on bearings relative to the bit body. Cutter elements are either machined
directly on the roller cones (Figure 6.8a) or are tungsten carbide inserts which are pressed into
the cone surfaces (Figure 6.8b). Cutters on milled tooth bits are longer, and the cones are offset
to create dragging action making these bits aggressive in soft formations. Insert bits are designed
for hard formations and break up the rock by crushing action rather than chipping and gouging
action.
(a) (b)
Figure 6.8 (a, b) Milled tooth and insert roller drill bits.
(a)
(b)
Figure 6.9 (a) Polycrystalline diamond compact (PDC) cutter. (b) PDC drill bit.
Roller cone rock bits are made with many design features and for a wide range of formation hard-
ness. The selection of any of these bits depends on bit cost, expected rate of penetration, and bit life.
Bit performance prediction is based on a database of bit records and is a statistical prediction.
Polycrystalline Diamond Compact (PDC) Drill Bits

Polycrystalline diamond compact (PDC) bits are drag bits containing multiple cutters made of syn-
thetic diamonds. The size of synthetic diamonds is about 175 diamonds per carat. These tiny dia-
monds are bonded to form a disc shape (about the size of a nickel) and backed by a thick layer of
tungsten carbide substrate. The synthetic diamond layer is about 0.025 in. thick. The tungsten car-
bide backing is about 0.115 in. thick (Figure 6.9a). This cutting structure is then bonded onto a
metal stud which is pressed into a steel body, which forms the drill bit. These cutters are also
embedded directly into the bit body (Figure 6.9b). When used in tungsten carbide matrix bits, they
are bonded onto a short disc which is then bonded into the matrix.
Rock bits of all types fail rock by developing high shear stresses under cutters. PDC cuter develop
shear stresses efficiently provided they can be forced into the rock. The application areas for PDC
drill bits are soft to medium hard formations. PDC bits were introduced in the early 1970s and
became a good companion bit for downhole motors and turbines, which operate at high rotary
speeds.
Natural Diamond Drill Bits

Natural diamond drill bits were introduced in the late 1940s. These bits allow exploration of deep
reservoirs which usually mean harder and more abrasive formations. Individual diamonds are set
in the matrix of the bit so that about one-third of the diamond is exposed and two-third is buried
within the matrix (Figure 6.10). The average depth of cut is generally one-third of the exposure of
the stone.
Industrial diamonds are handset in a machined graphite mold. A special mixture of matrix gran-
ules is placed between the mold and blank, then, heated in a furnace until the matric material
Control diameter
Figure 6.10 Leading face of a natural diamond drill bit. Source: Used by permission from Baker Hughes.
melts, capturing the diamonds onto the blank. A shank is then threaded onto the blank and welded
to form the diamond drill bit.
Diamond drill bits are used in hard formations typically found in deep wells (15 000 ft and
deeper).
Hydraulics of Rotary Drilling
The hydraulic system or circulating system is central to the rotary drilling method. Drilling fluid is
necessary to remove rock chips from underneath drill bits and carry them back up to the surface for
examination and disposal. Efficiency of cuttings removal depends on the amount of hydraulic
horsepower (HHP) in terms of fluid pressure and flow rate delivered to the drill bit nozzles in drill
bits and convert this energy into kinetic energy through specially selected nozzles. The effect of
nozzle size on pressure drop across drill bits is shown in Table 6.3.
Bit nozzles transform available energy into kinetic energy which is discharged under the bit for
cleaning and cuttings removal (Figure 6.11). Flow rate selection and nozzle sizes are important for
making the best use of available HHP at the lower end of a drillstring. Only a portion of the HHP
supplied by mud pumps at the surface (Figure 6.12) gets to the drill bit. About one-third is lost due
to fluid friction as the drill muds moves down the drillstring.
Optimized Hydraulic Horsepower

Another source of lost energy is friction in fluid flow. This is emphasized in friction losses in pump-
ing drill fluids in an oil well.
When drilling fluid is pumped through a drilling system, energy is lost due to fluid friction. This
friction loss is taken from the HHP and can be viewed as parasitic. Losses occur in surface equip-
ment, inside, and out of the drillstring. Energy loss across drill bit nozzles is not considered para-
sitic. Parasitic losses are substantial and must be considered in designing the overall circulating
system. A major portion of the total HHP input by the mud pumps is lost to friction. The remaining
Table 6.3 Pressure drop across bit nozzles (12 ppg drilling fluid density).
Nozzle size 11, 11, 11 11, 11, 12 11, 12, 12 12, 12, 12
TFA (sq in.) 0.2784 0.2961 0.3137 0.3313
Q (gpm) Δp (psi) Δp (psi) Δp (psi) Δp (psi)
300 1274 1126 1003 900

310 1360 1203 1071 961
320 1449 1281 1142 1024
330 1541 1363 1214 1088
340 1636 1447 1289 1155
350 1734 1533 1366 1224
360 1834 1622 1445 1295
370 1938 1713 1526 1368
380 2044 1807 1610 1443
390 2153 1903 1696 1520
400 2265 2002 1784 1599
410 2379 2103 1874 1680
420 2497 2207 1967 1763
430 2617 2314 2061 1848
440 2740 2423 2158 1935
450 2866 2534 2258 2024
TFA means total flow area

Nozzle sizes are expressed in 32 of an inch (for example 11/32 in.)
Figure 6.11 Fluid jets for bottom hole cleaning.

Rotation
Force
Bit
nozzle
Hook force
Swivel
Standpipe
Kelly
Rotary hose
Mud pump
BOP stack
Shale shaker
Drill pipe
Mud pits
Tailings L
Drill collars
Figure 6.12 Hydraulic system.
part is used for bottom hole cleaning and power for downhole motors or turbines. Parasitic losses
occur in
a) Surface equipment
b) Inside drill pipe
c) Inside drill collars
d) Annular space around drill collars
e) Annular space around drill pipe
Chapter 10 gives the analytical tools to predict pressure losses throughout the circulating system.
Bit cleaning, however, depends upon HHP that is created underneath the drill bit (HHPbit = ΔpbitQ).
This section develops the rationale for maximizing HHP for best bottom hole cleaning.
The difference between the power supplied by the pump and the power consumed by friction is
what is left over at the bottom of the drillstring. Similar calculations at other flow rates define a
curve showing how the available bottom hole HHP varies with flow rate.
System pressure losses are related to flow rate, Q, by
ps = CQ1 8 6 36
Mathematically, the available HHP reaches a maximum at a particular flow rate [4]. The condi-
tion for maximum available hydraulic power is predicted as follows:
1
HHPa = p Q − CQ2 8 6 37
1714 p
d HHPa
= pp − 2 8CQ1 8 = 0 6 38
dQ
By substitution of Eq. (6.36)
pp = 2 8ps 6 39
which means that HHPa is maximum when

1
ps = p 6 40
28 p
or when system pressure losses are about one-third of the maximum allowable pump pressure. The
flow rate at which this condition occurs depends on the hydraulics of the circulating system. Sub-
stituting Eqs. (6.36) into (6.40) gives
pp
= CQ1opt8 6 41
28
1
pp 18
Qopt = 6 42
2 8C
Using C = 0.0142 and assuming
Pp = 3000 psi maximum pump pressure
Then
1
3000 18
Qopt = = 513 gpm
2 8 0 0142
1
ps = 3000 = 1071 psi
28
3000 − 1071 513
HHPa = = 577 hp maximum
1714
These results are shown in Figure 6.13.
Field Application
A practical application of this information is explained below. Assume you go to the rig and find the
following drilling data:
•• Standpipe pressure is 2500 psi

Mud density is 12 ppg
•• Bit nozzle sizes are 11, 11, 12 (total flow area [TFA] = 0.2961 in.2)
Flow rate 360 gpm
•• Maximum allowable pump pressure, however, is 2800 psi

A bit change is imminent
The question is – What would be your recommendation for changes in these drilling parameters,
if any? The first step is to determine the value of “C” for the hydraulic system at current depth.
3500
3000
Pump output
2500
Pressure, psi
2000
1500
Parasitic losses
1000
500
0
0 200 400 600 800 1000
Flow rate, gpm
1600
1400
1200
Hydraulic horsepower, hp
1000
Pump Parasitic
800
600
Available
400
200
0
0 200 400 600 800 1000
Flow rate, gpm
Figure 6.13 System hydraulics.
Using Table 6.3, we see that the pressure drop across the drill bit is 1622 psi. Knowing the stand-
pipe pressure, then the parasitic losses are ps = 2500 − 1622 = 878 psi. Using Eq. (6.36)
ps 878
C= = = 0 022
Q1 8 3601 8
This number defines the total losses throughout the circulating system. Turn now to the drilling
parameters that will optimize bit cleaning.
With a maximum allowable pump pressure of 2800 psi, then system losses at maximum hydrau-
lics is
2800
ps = = 1000 psi
28
So bit pressure drop is 2800 − 1000 = 1800 psi. Optimum flow rate is
1
2800 18
Qopt = = 387 pgm
2 8 0 022
Using Table 6.3 and to select nozzle sizes would suggest staying with the same nozzle size but
increase flow rate to 387 gpm. Standpipe pressure should read 2800 psi.
Controlling Formation Fluids
Well control refers to drilling techniques for keeping formation fluids from erupting at the surface
[1]. There are three levels of well-control: (i) drilling mud weight, (ii) annular preventer, and
(iii) mechanical rams within a blowout preventer stack (BOP). Maintaining mud pressure slightly
higher than formation pressures is non-disruptive and is the first line of defense against a blowout.
The annular preventer is used to shut in a well and to remove a kick. Mechanical rams (pipe, shear,
blind) are used as a last resort.
It is apparent that the consequences of a well blow out are not good. There is loss of the well,
equipment, time, and possibly loss of life and personnel injury. There is also damage to the envi-
ronment and perceived losses in reputation and good will. Since the early days of rotary drilling
much has been learned about the early detection and control of formation pressures while drilling.
Hydrostatic Drilling Mud Pressure

Mud weight reduces rate of penetration exponentially. Pressure
The practical selection of mud weight depends on for-
mation pressure, fracture gradient of the formation and
drilling rate. In normal pressure areas, mud weight can
be kept constant down to total depth (Figure 6.14). Fracture
However, formation pressure is not normal and mud
weight may have to be increased. Mud weight is limited
by formation strength at the end of the last casing. This
pF
limit is established by conducting a leak-off test after a
casing has been cemented in place.
Hydrostatic pressure is the result of gravity and is
always in the drilling mud whether the pumps are run-
Drilling mud
ning. Total pressure is the algebraic sum of hydrostatic
and dynamic pressures. This means that both hydro-
static and dynamic pressures can be analyzed sepa- Formation
rately and then added together to determine pressure
at any point in the circulating system.
pf pm
Annular Blowout Preventer
The annular preventer is located at the top of the BOP
stack (Figure 6.15). A key component of an annular Figure 6.14 Invasions of formation fluids.
Controlling Formation Fluids 291
Figure 6.15 Annular preventer. Source: Used by permission from Baker Hughes.
preventer is an elastomer (rubber) ring which deforms around drill pipe sealing off the annular
space. The elastomer ring is activated mechanically. The pressure limit of annular preventers is
around 1500 psi.
Because drilling mud is not heavy enough to equalize formation pressure down hole, back pres-
sure is generated in the standpipe when the well is shut in. In this case, standpipe pressure increases
until bottom hole pressure is equal to formation pressure. Bottom hole formation pressure is deter-
mined from shut-in pressure by
p f = psi + 0 052γ m L 6 43
where
pf – formation breakdown pressure, psi
psi – shut-in pressure (standpipe pressure), psi
γ m – drilling mud weight, ppg
L – hole depth, ft
The constant, 0.052 compensates for mixed units.
The magnitude of the well kick is equal to the shut-in standpipe pressure, psi. The new mud
weight required to balance the formation pressure is
γ new = γ m + Δγ m 6 44
where
psi
Δγ m = 19 23 6 45
L
Assuming
•• shut-in standpipe pressure is 300 psi (magnitude of kick)

mud weight inside drill pipe is 12 ppg
• hole depth is 8000 ft

then formation pressure is
p f = 0 052 12 8000 + 300 = 5292 psi
The mud weight required to balance this formation pressure is

5292 = 0 052 γ m + Δγ m 8000 psi
γ m + Δγ m = 12 72 ppg
The well took a 0.72 ppg kick and the 12 ppg mud must be brought up to 12.72 ppg to establish a
balanced pressure condition. Final mud weight would be 13.22 ppg to have a 0.5 ppg overbalance
while drilling.
Hydraulic Rams
In cases where the magnitude of the well kick is very high (10 000 psi), the well may be shut in with
ram type closing devices (Figures 6.16 and 6.17).
Types of Rams:
•• Pipe ram – Closes around pipe to trap annulus pressure

Blind ram – Closes well when pipe is not at this location
• Shear ram – Severs pipe or casing and seals well bore
Figure 6.16 Double ram unit. Source: Courtesy of NOV.

Casing Design 293
Drill pipe
Force
Closing ram
Figure 6.17 Schematic of pipe ram.
Figure 6.18 Blowout preventer (BOP) stack. Source: Courtesy of NOV.
Ram units are stacked as shown in Figure 6.18 to provide back-up units and accommodate dif-
ferent pipe sizes. The total unit is known as the BOP stack. Annular preventers are attached to the
top of BOP stacks. The total height of BOP stacks may be 20–30 ft, which accounts for drilling rig
floors being some 30–40 ft above ground level.
The end use of these devices is to control formation fluids which could migrate into a well bore
while drilling and create a blowout at the surface.
Casing Design
A casing program is planned from the bottom up, i.e. the smallest or inside casing depends on size of
production tubing or rate of production expected from the reservoir. The size and number of casing
tubulars depend on formation pressure and fracture strength of the layered formations [1].
Cementing casing is an important and costly part of setting casing in a well bore. Cement around
casing provides support and prevents drilling mud and gas from moving upward from outside of the
casing. It is a vital aspect of well control.
Basic casing strings typical of most oil and gas wells

are (Figure 6.19)
•
BOP stack
Conductor is a piece of large pipe (~30 in. OD)
installed at the surface to keep the upper part of
the hole from caving in. The installation of conduc-
tor pipe is referred to as “spudding” in a well. It can Conductor
be installed prior to setting up the drill rig on land
wells and may or may not be cemented in place.
– Holds back unconsolidated surface formations
– Contains flow line for mud returns Surface casing
– Contains diverter above flow line when shallow
gas is probable
– Up to 30 in diameter
– Set between 300 and 1000 ft
• Surface casing is set to protect fresh water sands. It is

required by law. It contains a well head flange that Intermediate casing
supports the BOP stack and later, production
“Christmas Tree” equipment. Also, surface casing
must support the other casing strings. The annular
space must be cemented back to the surface to
prevent possible gas returns outside of the
BOP stack:
Production casing
– Holds back unconsolidated shallow formations
– Protects shallow zones from drilling mud and dee-
per formation contamination
– Serves as a base for the BOP stack
– Casing seat must hold deeper formation pressure
– Supports the suspended weight of other casing Figure 6.19 Types of casing strings.
strings
– Up to 133 8 in. diameter
– Set at depths from 1000 to 5000 ft (~2000 ft is typical)
– Cemented back to the surface
• Intermediate casing (or liner) is set for the same reasons:

– Controls open-hole conditions that may prevent a well from being drilled safely to the reservoir
– Prevents lost circulation
– Isolates troublesome shale, shallow gas, or water flows
– Reduces amount of uncased hole, especially in deep wells
• Production casing is set through production zone (except in open-hole conditions). It is cemented
from the casing shoe back to the surface:
– Designed to hold maximum shut-in pressure.
The objective in casing design is to select the lowest grade and weight of casing, to withstand
specified burst and collapse pressure loads, in order to minimize overall casing cost.
Casing Design 295
Collapse Pressure Loading (Production Casing)

Collapse pressure is the difference between the outside and inside pressures according to
Δp = p0 0 + γ 0 x − pi 0 − γ i x 6 46
It is common practice to assume that the most severe collapse pressure loading condition occurs
when the casing is completely evacuated, and external pressure is the result of hydrostatic mud
pressure. External pressure for this case increases linearly with depth and is based on the weight
of drilling mud when casing is run. This condition is usually assumed in casing design because it
gives a conservative design against collapse. The collapse load equation then becomes
p x = 0 052γ o x 6 47
where
p – pressure, psi
γ o – mud density (γ m), ppg (when casing was run-in)
x – vertical distance, ft (downward from surface)
The collapse pressure loading line is illustrated in Figure 6.20.
Burst Pressure Loading (Production Casing)

Burst loading is based on reservoir pressure (pf) with gas trapped inside the casing from bottom to
top. Assuming the formation at the casing shoe can support bottom-hole pressure, then the internal
pressure distribution is defined by
pi x = p f − 0 052γ g L − x 6 48
where γ g is density of gas.
pc
x
p c = 0.052 γm x
L
Figure 6.20 Collapse pressure loading line.

Here, pf is formation or reservoir pressure. At the surface, the shut-in pressure is formation pres-
sure minus the effect of the gas pressure.
In terms of pressure gradients, Eq. (6.48) becomes
pi x = pf − 0 052γg L − x 6 49
The pressure gradient term has units of psi/ft.

The external pressure is usually assumed to follow drilling mud hydrostatic pressure.
po = 0 052γ m x 6 50
The resulting burst pressure, pb, then becomes the difference between the inside pressure and the
outside pressure in the casing:
pb = pi − po 6 51
pb = p f − 0 052γ g L − x − 0 052γ m x 6 52
For the sake of simplicity, the gas pressure gradient is sometimes ignored (a conservative assump-
tion) leaving
pb = p f − 0 052γ m x 6 53
Taking formation pressure as bottom hole mud pressure

p f = 0 052γ m L 6 54
Then the burst loading line becomes (Figure 6.21)

pb = 0 052γ m L − x 6 55
Surface
p
x
pb = 0.052γm(L–x)
L
Burst pressure
Figure 6.21 Burst pressure loading line.

Casing Design 297
API Collapse Pressure Guidelines

The API standard on casing design [5] defines three regions of casing properties leading to collapse.
1) Plastic collapse pressure – The equation defining the minimum collapse pressure in this case is
Ac
pP = f ymn − Bc − C c 6 56
D t
pP – pressure for plastic collapse
fymn – specified minimum yield strength
AcBcCc – empirical constants depending on casing grade and D/t ratio
2) Transition collapse pressure – The equation defining the minimum collapse pressure in this
case is
Fc
pT = f ymn − Gc 6 57
D t
pT – pressure for transition collapse

FcGc – empirical constants depending on casing grade and D/t ratio
3) Elastic collapse pressure – The equation defining elastic collapse pressure is
1
pE = 46 95 × 106 2 6 58
D t D t−1
Plastic Yielding and Collapse with Tension

Tension also effects plastic yielding and collapse in casing. Tension and compression combine to
create a biaxial state of stress (Figure 6.22), which is evaluated for yielding using the von Mises
criteria of failure:
σ ≥ σ yld 6 59
where σ yld is material yield strength as determined from uniaxial testing and
2
σ = σ 2a − σ a σ θ + σ 2θ 6 60
According to Eq. (6.59), yielding occurs when von Mises stress, σ ≥ σ yld.
2 σa
σ 2a − σ a σ θ + σ 2θ ≥ σ yld 6 61 p
In dimensionless units, yielding occurs when

2 2
σa σa σθ σθ
− + ≥1 6 62
σ yld σ yld σ yld σ yld
σθ
For example, if
σ θ = − 40 000 psi Figure 6.22 Biaxial state
of stress.
σ a = 20 000 psi
σ yld = 50 000 psi
then
σθ
= − 0 8 80
σ yld
σa
= 0 4 40
σ yld
By substituting these numbers into Eq. (6.62), the left side is equal to 1.12, therefore this biaxial
stress state will cause plastic yielding.
Strength properties of casing sizes are conveniently display in API tables. Collapse and burst
strength capabilities are based on API equations outlined earlier. Casing size is listed in terms
of their outside diameter. Weight changes reflect changes in inside diameter. Grade or material
strength may vary for each casing size and weight. Grade of material effects collapse, burst, and
tension limits. A complete set of casing performance data can be found in Ref. [5].
Summary of Pressure Loading (Production Casing)

Collapse pressure loading is illustrated in Figure 6.22 showing the pressure distribution increasing
directly with depth according to
pc = 0 052γ m x psi 6 63
The maximum collapse pressure loading is at the bottom end.
Burst pressure load is the difference in internal pressure and the external pressure. An extreme
case would exist if formation gas pressure is exerted internally from bottom to top and internal pres-
sure being partially resisted by hydrostatic mud pressure outside the casing. The resulting burst
pressure for this assumption is
pb = p f − 0 052γ m x 6 64
pb = 0 052γ m L − x 6 65
The maximum burst pressure occurs at the well head.
The collapse and burst pressure loads are illustrated in Figure 6.23 by the “X” loading diagram.
A casing string designed strictly for collapse would be strongest at the bottom and weakest at the
top. While this unusual casing string would not collapse, it would be susceptible to burst type failure
in the event the well is shut in during a blowout. A casing string designed for burst alone would
most likely collapse if the casing were evacuated.
A proper casing design is one which can resist both collapse pressure load and burst pressure
load. In addition, it is tapered and has the strongest section near the top and bottom with the weak-
est casing section near the middle.
The selection of a factor of safety may vary with the level of experience in a given area and the
confidence level for each casing load prediction. Small factors of safety not only give lighter and
therefore cheaper casing strings but also carry a higher risk element.
Effect of Tension on Casing Collapse

Tension along the axis of casing also affects the collapse resisting capability of casing. The collapse
pressures given in the casing performance tables are based on a uniaxial state of stress assuming
casing is subjected to external pressure only. This pressure creates circumferential or hoop stresses.
Casing Design 299
Figure 6.23 Typical loading for casing design. Pressure
Burst
Depth
Collapse
Casing in general can be loaded with a direct pull creating longitudinal stresses simultaneously
with circumferential stress producing a biaxial state of stress. According to the von Mises criteria of
failure, local yielding in casing will develop when the maximum energy of distortion created by
biaxial stress state is equal to the maximum energy of distortion at yielding under a uniaxial state
of stress. Equating these two energies gives
σ 2yld = σ 2T − σ T σ C + σ 2C 6 66
or
2 2
σT σT σC σC
− + =1 6 67
σ yld σ yld σ yld σ yld
where
σ T – tension stress, psi
σ C – collapse stress, psi
σ yld – yield strength, psi
Equation (6.67) can also be written as
2 2
T T pC pC
− + =1 6 68
T yld T yld Pc,yld pc,yld
This equation establishes the relation between collapse pressure and tension which produce
yielding. The plot of this equation (Figure 6.24) shows how much the collapse pressure has to
be derated for a given amount of direct pull. The portion of this diagram that is applicable to casing
design is the portion in the fourth quadrant. These coordinates can be tabulated for accuracy and
convenience.
For example, if the direct pull on 95 8, 40 lb/ft, N80 casing at a given location is 110 000 lb, then
T 110 000
= = 0 12
T yld 916 000
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
–0.1
–0.2
–0.3
Pressure ratio, p/pyld
–0.4
–0.5
–0.6
–0.7
–0.8
–0.9
–1
Tension ratio, T/Tyld
Figure 6.24 Yielding condition for biaxial state of stress.
According to Figure 6.24, the casing’s resistance to collapse pressure has to be derated by a factor
of 0.93. At this location, collapse strength is derated from 3090 to 2874 psi.
Tension Forces in Casing

The neutral point here is defined as the point in the casing where axial force or stress is zero. This
point is different from the neutral point in drill collars, which is based on buckling. Tension cal-
culations are usually based on casings being completely surrounded by drilling mud (inside and
out). In reality, this is not true, but the calculation is conservative.
Based on this assumption, the internal force distribution along casing is defined by (see
Figure 6.25).
F x = γA L − x − Lγ m A 6 69
where x is measure from the top and w = γA
F = w L − x − wm L 6 70
This equation is used to define the magnitude of the true axial force in casing. Accordingly,
at x = 0
F = W BF 6 71
and at x = L
F = − wm L 6 72
Casing Design 301
Figure 6.25 Internal force in casing. H
Lnp
x
F(x) F=0
w (L – x)
L γm A
The location of the point where F = 0 (σ a = 0)

0 = γAL − γALnp − γ m AL 6 73
0 = w − wm L − wLnp
wm
Lnp = 1 − L 6 74
w
Lnp = BF L 6 75
If casing is empty and plugged at the bottom, the force distribution would be completely different
from what is predicted by Eq. (6.70). Equation (6.75) defines the location of the neutral point, the
point of zero stress. The internal tension above the neutral point is determined from
F x = w Lnp − x 6 76
Casing collapse pressure strength has to be derated to account for the effect of tension. The
p T
strength ellipse in Figure 6.24 relates the pressure ratio, , to the tension ratio, . The collapse
pyld T yld
strength is derated from pyld to p which becomes the allowable collapse pressure, pa before FS is
applied.
Design of 95 8 in. Production Casing
Example As an example, consider the design of a production casing having the following
specifications:
•• Depth of well – 9500 ft

Casing size – 95 8 in.
• Mud weight – 12 ppg

The objective of the design is to select a distribution of casing having lowest weight and grade.
The design starts at the bottom by considering collapse loading (Figure 6.26). The maximum col-
lapse pressure load is shown as 5928 psi (p = 0.052(12)9500 = 5928 psi).
Design Without Factors of Safety

For the sake of simplicity, FS are not considered at this point. They will be included later. From
Table 6.4, the lowest weight and grade combination having a collapse strength equal to or greater
than 5928 psi is 47 lb/ft (S-95), which has a collapse strength of pc = 7100 psi (burst strength is
pb = 8150 psi).
Each casing section will be tabulated as the design moves upward.
Section #1 9500 − x 47 lb/ft (S95)
At what point can casing weight and grade be reduced? Casing 43.5 lb/ft (S95) has a collapse
strength of pc = 5600 psi (burst strength of pb = 7510 psi).
The location at which the applied collapse pressure is 5600 psi is determined from the pressure
load function.
pc
x pc = 0.052 (12) x
Lnp = 7762 ft
9500 ft
5928 psi
Figure 6.26 Collapse loading (95 8 casing, 12 ppg mud).

Table 6.4 Strength capability of 95 8 in. casing (non API grades).
Nominal Collapse Burst Body tension Coupling tension

weight (lb/ft) Grade pressure (psi) pressure (psi) (lb × 1000) (lb × 1000)
36.00 H40 1740 2540 410

K53 2020 3520 564 489
S80 2980 3520 564 605
40.00 K55 2570 3950 630 561
S80 4230 3950 630 694
C75 2980 5390 859 694
N80 3090 5750 916 737
SS95 4230 5750 916 837
S95 4230 6820 1088 858
C95 3330 6820 1088 847
43.50 C75 3750 5930 942 776
N80 3810 6330 1005 825
SS95 5600 6330 1005 936
S95 5600 7510 1193 959
C95 4130 7510 1293 948
P110 4430 8700 1381 1106
47.00 C75 4630 6440 1018 852
N80 4750 6870 1086 905
SS95 7100 6870 1086 1027
S95 7100 8150 1289 1053
C95 5080 8150 1289 1040
P110 5310 9440 1493 1213
Source: Condensed from Lone Star Steel casing tables, Ref. [6].
pc x = 0 052 12 x = 5600 psi

5600
x= = 8974 ft
0 052 12
This locates the beginning of Section #2.
Section #1 9500 − 8974 47 (S95)

Section #2 8974 − x 43.5 (S95)
The next lowest weight is 40 (S95) having collapse strength of 4230 psi (burst strength is 6820 psi).
The beginning of the third section then is
4230
x= = 6778
0 052 12
The status of the casing design is now
Section #1 9500 − 8974 47 (S95)

Section #2 8974 − 6778 43.5 (S95)
Section #3 6778 − x 40 (S95)
The next lightest casing weight is 36 (S80) having a collapse strength of 2980 psi. This casing
weight and grade is selected for Section #4, which starts at
2980
x= = 4776 ft
0 052 12
The status of the casing design is now
Section #1 9500 − 8974 47 (S95)

Section #2 8974 − 6778 43.5 (S95)
Section #3 6778 − 4776 40 (S95)
Section #4 4776 − x 36 (S80)
Each of these sections is shown in Figure 6.27. This process could continue along the collapse
load line, but at this point, burst loading becomes the important consideration. The two dots in
Section #4 indicate that 36 (S80) can support collapse loading from 4776 ft and upward. At the same
time, this casing can withstand the burst loading up to the upper dot.
The fourth casing weight and grade has a burst strength of pb = 3520 psi. The length of this
section is controlled by the burst loading line. The location where the burst load is equal to
3520 psi is determined as (using Eq. (6.65))
pc = 0.052 (12) x
Lnp = 7762 ft
4
4750
9500 ft
3
5928 psi
Figure 6.27 Collapse and burst loading.

pb 3520
x = L− = 9500 − = 3859
0 052 12 0 052 12
which locates the upper dot. From this point upward, burst pressure controls the design.
The casing string has four different casing weights (Figure 6.29):
Section #1 9500 − 8974 ft 47 (S95)

Section #2 8974 − 6778 ft 43.5 (S95)
Section #3 6778 − 4776 ft 40 (S95)
Section #4 4776 − 3859 ft 36 (S80) (burst controls)
Continuing upward on the burst pressure line (Figure 6.28), casing weight of 40 lb/ft gives
several grade options. We choose grade S95 because its burst strength is 6820 psi which is greater
than the maximum burst load (5928 psi) at the surface. Adding Section #5 completes the casing
design.
Interval (ft) Weight (grade) (lb/ft) Weight (lb)
Section #1 9500 − 8974 ft 47 (S95) 24 722

Section #2 8974 − 6778 ft 43.5 (S95) 95 526
Section #3 6778 − 4776 ft 40 (S95) 80 080
Section #4 4776 − 3859 ft 36 (S80) 33 012 (burst controls)
Section #5 3859 − 0 40 (S95) 154 360 (burst controls)
Total 387 700 lb
5928 psi
pb = 0.052(12)(L – x)
x
5
3859 ft 6820 psi

4
4776 ft
3520 psi
9500 ft
Figure 6.28 Burst load line and casing sections.

Directional Drilling
Directional well paths depend on shape of reservoir and location of a drilling and production plat-
form. As many as 40 directional wells may be drilled from one platform. A platform may be located
offshore or in environmentally sensitive land locations. With multiple wells from one platform, well
paths of each must be documented with respect to the others. The point of entry into the reservoir is
also important.
There are three basic types of well paths (see Figure 6.29). Well paths may be more complicated
depending on shape of the reservoir. Broken and faulted reservoirs may require entry from the side
or even from below.
Downhole Drilling Motors

Downhole motors were introduced in the late 1960s as a means of deflecting a direction well away
from vertical without using whipstocks. A bent sub was placed on top of the motor to bring about
lateral drilling (Figure 6.30a). Tool face is within the bend. It was oriented in a desired plane of
drilling to bring about a directional change. This arrangement began to replace whipstocks. Tool
face, hole direction, and inclination were measured with a downhole sensor lowered on a wire line.
Measurements were taken intermittently until the desired hole angle was achieved. PDM emerged
as a preferred directional tool during this time.
The advantage of using PDMs in directional drilling was huge. It replaced the time-consuming,
multiple steps of setting a whipstock and surveying to establish hole direction. The bent sub and
PDM did the same thing faster. At first, surveys were made by placing a survey tool in a side pocket
and using a hard wire to bring hole direction and inclination information back to the surface. Later
survey instruments were inserted and retrieved from within the drill collars to track progress. Once
the angle had been established, a stabilized BHA was used to maintain hole direction to the target.
Figure 6.29 Three basic types of

directional wells.
Type 1 Type 2 Type 3

Bent sub
L1 Tool face
Motor
α
L2
Drill bit
L3
Bent sub Bent housing Tilted bearing

(a) (b) (c)
Figure 6.30 (a–c) Effect of bend location on rate of build.
When the bit drilled off course, a bent sub and PDM were used again to bring the hole back on
course.
As a rule, a bent sub and motor turn a well path about 75% of the bent sub angle over 30 ft of
drilling. This means that a 2 bent sub creates a rate of build of about 1½ per 30 ft or a dog leg
severity of 5 per 100 ft. The maximum allowable bent sub angle is limited by the mount of offset
at the bit. If this offset is large, it is difficult to get the BHA into the well bore. As a result, the max-
imum bent sub angle is limited to 2½.
Bit torque is transmitted across motors, and this torque creates a rotational reaction of the tool
face, which depends on hole depth, friction, bit torque, and drillstring size. This reactive torque is a
major consideration in setting and maintaining tool face orientation during a correction run.
Eventually, the bent sub was replaced by a bend within the motor housing (Figure 6.30b). The
bent housing configuration brought about a more responsive direction change. This configuration
was subsequently replaced by a tilted bearing assembly (Figure 6.30c) which located the tool face
even closer to the drill bit. This configuration produced even a greater rate of directional change,
Δθ α
which can be expressed by = . By reducing the distance, L, the rate of directional change is
Δs L
increased. By moving the bend angle close to the drill bit, the angle, α could be reduced and still
achieve the same rate of build.
The realigned bearing or tilted bearing assembly gave another advantage, too. The motor housing
itself could be used to drill ahead as well as for making hole corrections. By rotating the motor hous-
ing, the drill bit would be continually disoriented, and drilling could proceed with motor power.
Rotation of the motor housing at low speeds simply disoriented the drill bit. When hole corrections
were needed, the tool face in the housing was set, and the correction was made by motor power.
This, along with MWD, eliminated costly tripping time.
Rotary Steerable Tools

A vision for many years has been a technology that would allow directional drilling to proceed
along a specified path without stopping to make directional changes. Essentially, the drill bit would
Figure 6.31 Schematic of rotary steerable tool.
Internal support
Internal shaft
F, adjustable
Tool face
Internal support L4
α Drill bit
advance through layers of various formations much like airplanes flying and navigated along a spe-
cified flight path, with the ability to alter the course in the event of strong crosswinds or bad
weather. In each case three elements are essential:
•• Power to move forward

Navigation capabilities to define position and well path
• Control of the well path

Technology is now in place to direct drill bits in three-dimensional space. Tool face monitoring
and control is central. The ability to monitor bit location and path through MWD and then bring
about directional change is equally important. Each of these tasks is now achieved by a RST system.
The only limitation appears to equipment reliability.
The schematic in Figure 6.31 illustrates an internal shaft which is deflected by an internal force,
F. The direction and magnitude of the force established tool face orientation and rate of change in
hole angle. Note that the parameters, α and L4 have the same role as the ones in Figure 6.33; how-
ever, the magnitude and orientation of α is adjusted while drilling. Many mechanical designs for
this purpose have been patented. The author proposes another method to orient tool face and mag-
nitude of angle α, which affects dog-leg severity.
Stabilized Bottom-Hole Assemblies

Downhole motors are commonly used to deflect a well at a kick-off point toward a target. The
deflection takes place in the plane of the target line. A bent sub/motor assembly is used until hole
angle reaches about 6–8 ; a stabilized building assembly is not effective below these angles. In hard
rock, a bent sub/motor assembly may have to establish even higher angles before a building assem-
bly is place in the hole. Building assemblies are desirable as early as possible because tool face ori-
entation is not a consideration. However, bent sub/motor assemblies have been used to build up to
45 from the kick-off point.
Bent sub/motor assemblies are used to
•• Kick off
Establish rate of build
• Make hole corrections

After hole angle is achieved, drilling proceeds with a stabilized assembly: building, holding, or
dropping. These assemblies differ primarily in the number and location of stabilizers.
Stabilizers are special tools that centralize drill collars at specific locations to establish a desired
side load to drill bits. Stabilizer blades are roughly 2 ft long and allow fluid to flow upward in the
annulus with minor flow restriction. Stabilized assemblies can build, hold, or drop hole angle,
depending on stabilizer number and spacing.
Prediction of bit side forces produced by various stabilizer assemblies is a guide for predicting
directional performance. Ultimately, stabilizer assemblies have to be tried, tested, and usually mod-
ified in a given area until they perform as needed. A good data base of BHA performance in a given
area is useful in this regard.
Building assemblies – Building assemblies develop an upward side force on drill bits. A near-bit
stabilizer acts as a fulcrum point for the collar weight to pry against (Figure 6.32a). The mechan-
ical advantage of this assembly is greatest when the first stabilizer is close to the drill bit. A second
stabilizer is positioned some distance to establish a large free span distance to increase the
strength of building assemblies. Positioning of the second stabilizer depends on drill collar size.
Stiffer or larger drill collars will allow a longer free span distance.
Holding assemblies – Stabilizers are positioned within drill collars so that relatively small side
forces (usually less than 200 lb) are developed at the bit. When drill bits having long shanks
Drill collars
30 – 90 ft
30 ft
Stabilizer
30 ft
40 – 60 ft
15 – 20 ft
Near bit stabilizer
Drill bit
(a) (b)
Figure 6.32 Bottom hole assemblies. (a) Building. (b) Holding.

Figure 6.33 Diesel-driven generator.
or large gauge protection areas are used, it is desirable to minimize the tilt angle with stabilizer
placement to alleviate bit binding. A typical holding assembly is illustrated in Figure 6.32b. Cal-
culations show that hole angle directly affects the magnitude of the side force for each of these
assemblies. Side forces increase with hole angle; examples are given in the Appendix.
Dropping assemblies – Unlike building and holding assemblies, dropping assemblies have no near-
bit stabilizers. The drill bit stands alone and is pushed downward by the weight of drill collars
suspended between the bit and the first stabilizer which is roughly 60 ft from the bit. The first
stabilizer in this case acts much like a hinge allowing the suspended collars to swing downward
like a pendulum.
Power Units at the Rig Site
The power source for drilling rigs is a skid unit containing a diesel engine and an electric generator
(Figure 6.33). The output power can be as high as 10 000 hp (7 457 000 W or 7.457 MW). These
power units are necessary because drilling rigs are usually located in remote area where public uti-
lities are not available.
Electric power is safer and easier to transmit to various power needs around the rig.
As a power unit, the “end usage” of the power source is electric power. This power is consumed by
(i) hoisting, (ii) rotary system, (iii) drilling mud pumps, and (iv) well control equipment.
References
2 Johancsik, C.A., Friesen, D.B., and Dawson, R. (1984). Torque and drag in directional wells –
prediction and measurement. J. Pet. Technol. SPE 11380: 987–992.
References 311
3 API Standard (1987). Recommended Practice for Drill Stem Design and Operating Limits (API RP 7G),
12e. American Petroleum Institute.
4 Kendall, H.A. and Goins, W.C. (1960). Design and operation of jet-bit programs for maximum
horsepower, maximum impact force and maximum jet velocity. Trans. AIME 219 (01).
5 API Committees. (2008). Technical report on equations and calculations for casing, tubing, and line
pipe used as casing or tubing; and performance properties tables for casing and tubing. ANSI/API
Technical Report 5C3. 1 ed.
6 Lone Star Steel Company (1984). Casing and Tubing Technical Data. Dallas, TX.
313
Part III
Analytical Tools of Design
Engineering Design is about predicting future performance – through science and mathematics.
Analytical tools are available for designing structures and machines to a high level of reliability.
Insightful application of these tools is essential for safe and reliable products.
Designs from antiquity were based on experience, tradition, geometry, esthetics, and empirical
formulas and probably a lot of testing. The material of construction was stone, quarried by hand,
moved to the construction site, and lifted in place. Each step required strenuous manual labor
using, levers, ramps, and pulleys. Even though Newtonian physics and thermodynamic principles
emerged during the 1700s, it was not until the mid-1800s that science-based engineering design was
accepted. Also, during this time steel became a structural material.
Engineering science is now commonly used to develop machines and structures to satisfy design
specifications. Life expectancy and performance can be predicted with a high degree of accuracy
using analytical models and numerically based computer software.
Part III brings together the mechanics of many analytical tools commonly used in design. Refer-
ences are given for further investigation into these topics.
315
Dynamics of Particles and Rigid Bodies
Dynamics is a science that deals with bodies in motion. The subject stems from Newton’s laws of
motion, which were first published in 1687. An important aspect of Newton’s mathematics is the
concept of rate of change, such as speed and acceleration. This led to the principles of calculus. The
subject of dynamics is viewed per the following breakdown:
Statics – Study of bodies at rest or in a state of static force equilibrium, having zero acceleration.
Kinematics – Study of motion of objects without regard to the forces that cause the motion.
Kinetics – Study of motion of objects caused by an imbalance of applied forces.
All three, stem from Newton’s laws of motion, which apply to both discrete particles and rigid
bodies.
Statics – Bodies in Equilibrium
Statics refers to bodies at rest. Equilibrium refers to a system of forces that allows this to happen. If a
body is at rest, then the forces applied to the body are in equilibrium. Statics is a special case of the
second law of Newtonian physics, i.e. linear and angular accelerations of a body are both zero.
When a body is in equilibrium the force system must balance.
F=0 71
M0 = 0 point 0 is arbitrary 72
There are different types of force systems: concurrent forces, biaxial forces, coplanar forces, and
general three-dimensional (3-D) force systems. It is helpful to recognize the type of force system in
setting up the equations. Freebody diagrams are most helpful.
In any engineering situation, an analysis based on elementary mechanics is a good starting point.
This provides a means of scoping out the problem by defining magnitudes of forces, stresses, flow
rates, etc. A more refined analysis with computer software can follow, if necessary. This approach is
useful in evaluating the technical feasibility of design concepts and configuring design possibilities.
What are the magnitudes of external and internal forces? How do the forces flow through a design
configuration and its subsystems? Understanding load magnitudes (internal and external) is
essential in design innovation.
316 Dynamics of Particles and Rigid Bodies
Force Systems
A concurrent force is one where all forces intersect at a point. The force vectors can be either two or
three dimensional. In either case, the sum of all the force vectors must be zero for equilibrium:
Fn = 0 73
Three unknown force components can be determined from this vector equation. The three forces
acting on the pin A in the simple structure (Figure 7.1b) form a concurrent force system. Summing
forces in the x direction gives
− T cos θ + Q = 0
Summing forces in the y direction gives
T sin θ − F = 0
Solving these equations together with θ = 30 and F = 1000 lb, gives T = 2000 lb and Q = 8660 lb.
Members that are pinned at each end (AB, Figure 7.1a) are examples of biaxial (or collinear) force
systems. The end connections do not transmit moment, therefore, applied forces at each end are
collinear. This condition is useful in determining the direction of forces as illustrated above.
A coplanar force system is one in which side forces and/or moments are applied to a structural
element. The force vectors are applied in one plane and moments vectors are perpendicular to the
plane. In which case
Fn = 0 74
Ma = 0 75
where
n – number of force vectors

a – fixed point
Consider a freebody diagram left of section a–a in Figure 7.2. It is desired to find the magnitude of
internal shear, V, and internal moment, M. Applying Eqs. (7.4) and (7.5)
F A − V − wx = 0
V = F A − wx
(a) (b)
B θ A A
X
Q
L
F F
Figure 7.1 Concurrent force system.

a M w
V EI
FA x FB
a
L
Figure 7.2 Internal force balance.
and
x
M + wx − FAx = 0
2
1
M = F A x − wx 2
2
The reactions at points A and B are found by applying Eqs. (7.4) and (7.5) to the total beam.
In general, force vectors and moment vectors can be three dimensional. In which case, equilib-
rium is established when
Fn = 0 76
Ma = 0 77
Consider the bent rod of Figure 7.3. The problem is to determine the force and moment reactions
at point O. The six unknowns are Ox, Oy, Oz, Mx, My, Mz. Equation (7.6) gives three scalar equations
whose solution establishes the three reaction forces:
x
0
ℓ
z F
Figure 7.3 Three-dimensional force balance.

Ox + F cos α = 0
Oy + F cos β = 0
Oz + F cos γ = 0
where cos α, cos β, and cos γ are directional cosines of force F.
Equation (7.7) gives
M0 + r × F = 0
where
r = Li + 0j + lk
F = F i cos α + j cos β + k cos γ
i j k
M0 = F L 0 l
cos α cos β cos γ
The solution to the three scalar equations gives the moment reactions at point 0.
Freebody Diagrams
Freebody diagrams are isolated portions of a structure used to determine external or internal forces
and moments. Internal reactive forces (and moments) are needed to determine stresses, buckling,
or other responses that may affect the structures performance, including possible failure.
Control volumes are imaginary boundaries defined to study fluid flow parameters such as flow
rate, pressure drop, hydraulic horsepower, and friction.
Force Analysis of Trusses

A truss is a structure made up of a network of triangular sections. The overall stiffness of the struc-
ture is generated by the combination of each triangular section. The first of all metal trusses was
designed and built in 1840 by Squire Whipple [1]. Compression members were cast iron. The chords
and diagonals (tension members) were made of wrought iron. Prior truss bridges were made
of wood.
Whipple’s metal truss bridges, built across several locations along the Erie Canal, were the first
scientifically designed truss bridges. His force analysis assumes that the connection points are fric-
tionless pins, bending moments at the joints are ignored. While this is not exactly true in rigid con-
nections, the assumption yields reasonable forces in truss members. This assumption greatly
simplifies the force analysis by allowing forces in each beam to be treated as collinear, i.e. internal
forces are in-line with the axis of the element. Each member is assumed to be axially loaded with
zero bending.
Consider the trusses shown in Figure 7.4. Force of magnitude in each element is needed to con-
duct a stress and buckling analysis the structure. Whipple gives two approaches: (i) method of joints
and (ii) method of sections.
The first step is to view the total structure as a freebody with unknown forces, RA and RB. The
magnitudes of these two forces are determined from statics:
6 ft 6 ft
A B
1 5
RA 2 RB 4 ft
ϕ
4
3
6 ft
10 kips
Figure 7.4 Truss example.
MA = 0
− 12RB + 9 10 = 0 78
RB = 7 5 kip
Fy = 0
RA + RB = 10 79
RA = 2 5 kip
The next step is to determine the magnitude of the forces in each of the seven (7) members. This
information is needed to size each member for bridge performance and safety. Dimensions for this
example are selected so that sin ϕ = 3 5 and cos ϕ = 4 5 for convenience of calculation.
Method of Joints
This method of determining forces in each structural element is based on freebody diagrams of each
joint. Forces at each joint form a concurrent force system with each force coming together at each
joint as shown at joint B in Figure 7.5.
Fy = 0
RB − F 4 cos ϕ = 0 7 10
F 4 = 9 375 kip tension
Fx = 0
F 5 − F 4 sin ϕ = 0 7 11
F 5 = 5 625 kip compression
Forces in each element are determined from freebody diagrams of each joint, remembering that
the force system at each joint is concurrent. Whipple sometimes used cables for members in
tension.
Method of Sections
The method of sections creates freebody diagrams by cutting across the structure with an imaginary
line as shown in the left portion of Figure 7.5. Forces in each member are determined using all three
of the equations of statics:
Method of sections Method of joints
A B
F1 F5
F2 F4
RA = 2.5 kips ϕ RB = 7.5 kips
F3
10 kips
Figure 7.5 Application of freebody diagrams.
Fy = 0
RA − F 2 cos ϕ = 0
4 7 12
2 5 − F2 = 0
5
MA = 0
4F 3 − F 2 3 cos ϕ + 4 sin ϕ = 0
7 13
F 3 = 1 2F 2
F 3 = 3 75 kip tension
Fx = 0
− F 1 − F 2 sin ϕ + F 3 = 0 7 14
Since each force prediction is plus (+), the assumed directions of F1, F2, F3 are correct.
This method is most useful for finding internal loads in members some distance away from sup-
port points.
Kinematics of Particles
Linear Motion
Linear motion refers to motion of a particle along a straight line. Position, velocity, and acceleration
are related through calculus. If s = s(t) is known, then
ds
vt = 7 15
dt
dv
at = 7 16
dt
dv ds dv
a= = v
ds dt ds
a ds = v dv 7 17
This expression is useful when a(t) is known.
Assuming acceleration, a, is constant gives
1 2
a s 2 − s1 = v − v21 7 18
2 2
If an object is thrown upward with an initial velocity of v1, the height of travel is
1
− g s2 − 0 = 0 − v21
2
v21
s2 = 7 19
2g
If a(t) is not constant and a known function of time, it is best to use dv = a(t)dt.
v 2 = v1 + a t dt 7 20
1
s2 = s1 + v t dt 7 21
1
These two equations are useful in tracking velocity and position with time.
Rectangular Coordinates
A particle traveling in two-dimensional space is defined by tracking both x(t) and y(t) components.
Projectiles are good examples. Ignoring air friction, the path of a projective is determined as follows
(Figure 7.6).
x Component of the Path
ax = 0
ay = –g
vo
Figure 7.6 Projectile path.

ax = 0
vx = C 1
x t = C1 t + C2 7 22
y Component of the Path
ay = − g
vy = − gt + C3
t2
y t = −g + C3 t + C4 7 23
2
Initial conditions are
x 0 =0 C2 = 0
vx = v0 cos θ C1 = v0 sin θ
y 0 =0 C4 = 0
vy = v0 sin θ C3 = v0 cos θ
Applying initial conditions gives

x t = v0 cos θ t 7 24
2
t
y t = −g + v0 sin θ t 7 25
2
Eliminating the time variable gives the path in terms of y = y(x)
g x2
y x = x tan θ − 2 7 26
2 v0 cos θ
The projectile touches ground at a horizontal distance of

2 sin θ cos θv20
L=
g
v20
L= sin 2θ
g
from the launch point. Assuming θ = 80 and v0 = 100 fps, L = 1062 ft.
Polar Coordinates
Some problems can best be defined in terms of independent variables r and θ as shown in Figure 7.7.
In the left case, the point moves in a circle with constant radius R. Angular position θ(t) may vary
with time. In this case, the position of a point on the circle is defined by s(t) = R θ(t).
Its tangent velocity is equal to
dθ
v=R = Rω 7 27
dt
Its vector is tangent to the circle. Acceleration of the point has two components, a normal com-
ponent and a tangent component. At any point on the circle
y e𝜃
y
er
R r
θ θ
x x
r is constant r can vary with time
Figure 7.7 Polar coordinates.
a = − Rω2 er + Rαeθ 7 28
The position of the particle may be determined by arc travel, s(t), in which case:
s t = Rθ
s t = Rω
s t = Rα
If θ(t) is not continuous, then position, velocity, and acceleration can be determined like the
straight line motion formulas discussed earlier.
In general, both the angular and radial positions can change with time (Figure 7.7, right draw-
ing). In this case
r = rer 7 29
der
v=r + rer 7 30
dt
as shown in Figure 7.8.
Since
der deθ
= θeθ and = − θer
dt dt
v = rθeθ + rer or v = vθ eθ + vr er 7 31
By differentiation
a = rθ − θer + eθ rθ + θr + rθeθ + er r
2
a = r − rθ er + rθ + 2rθ eθ 7 32
Consider the mechanism shown in Figure 7.9 where the collar is constrained to follow along a
path defined by
r = Aeaθ 7 33
By differentiation:
Figure 7.8 Differential change in unit vectors.

−∆θer
eθ
−∆θeθ
∆θ
er
3
y-coordinate
2
r
1
θ
0
–4 –3 –2 –1 0 1 2 3
–1
–2
x-coordinate
Figure 7.9 Exponential cam drive.
r = A a eaθ θ 7 34
r = Aa e θ + θae
aθ aθ
7 35
Assuming r and θ are related by
r = 2e0 1θ
and the arm rotates with constant angular velocity θ = 2 rad/s. The problem is to determine the
velocity vector and acceleration vector of the collar for θ = 60 or 1.047 rad.
For this case,
A = 2 in
a=01
r = 2 22 in
vr = r = Aaeaθ θ = 0 4442 in s
r = Aa θaeaθ = Aa2 θe0 1θ = 0 0444 rad s2
vθ = rθ = 2 22 2 = 4 44 in s2
Velocity Vector
Applying Eq. (7.41) gives
v = vθ eθ + vr er
v = 4 44eθ + 0 444er
Acceleration Vector
Applying Eq. (7.42) gives
2
a = r − rθ er + 2rθ eθ
a = 0 0444 − 2 22 2 2 er + 2 0 444 2 eθ
a = − 8 835er + 1 776eθ
Curvilinear Coordinates
In this case, a particle moves along a prescribed path, y(x) (Figure 7.10). Its location along this path
ds
is defined by the arc distance, s(t). Tangent velocity is determined by . Tangent acceleration is
dt
d2 s
determined by 2 . The total acceleration has a normal component, too, as explained below:
dt
v = vet 7 36
dv deT dv
a= =v + eT 7 37
dt dt dt
but
Δs
ΔeT = eN 7 38
ρ
deT v
= eN 7 39
dt ρ
so
Figure 7.10 Curvilinear motion. y

et
s
en
y(x)
x ρ
x
v2
a= eN + aeT 7 40
ρ
where
d2 y
1 dx 2
= curvature 7 41
ρ 2
3
2
dy
1+
dx
Example Consider the acceleration components of a point moving along a path defined by
y
a
ρ
an
at
y(x)
S
x
y = 3x 2 − 2x
at coordinates (x = 2, y = 8). Assuming the velocity and acceleration of a moving point at this coor-
dinate are s = 30 fps and s = 20 fps2, its acceleration components are
v2
aN = = 0 0059 302 = 5 32 fps2
ρ
where
1 6
= = 0 0059
ρ 1 + 102
3
2
The tangent component of acceleration is given as
at = 20 fps2
The direction of the normal acceleration component is always toward the center of curvature of
the path.
Figure 7.11 gives another two-dimensional space path, which could represent the terrain on a hill
or roller coaster. The path is represented by one half of a sin wave:
πx
y x = H sin
L
Figure 7.11 Sinusoidal ramp.

y
H
s
x
L
Assuming a constant velocity of 60 mph (88 ft/s), the acceleration at the top of the hill (x = L/2)
would be
v2
an =
R
W
y
v
Hcr
The radius of curvature of the road at this point is

dy π π πH
=H cos = 0 =0
dx x = L2 L 2 L
d2 y π 2 π π 2
= −H sin = −H
dx 2 x = L2 L 2 L
Therefore, from Eq. (7.52)

1 π 2
= −H
R L
The minus sign means that the y(x) at L/2 is cupping down away from the +y axis.
The acceleration of a vehicle at the top of the hill is (note at = 0)
v2 π 2
a = an = = Hv2 fps2
R L
Now consider the contact force between the tires and road of a 2000 lb vehicle going over the peak
point at 60 mph.
Applying the second law,
W − N = ma
Solving for N gives
W π 2
N =W− Hv2
g L
An interesting situation occurs when N = 0 or the wheels start to leave the road. That condition is
reached when
π 2
Hv2 =g
L
Assuming a wavelength L = 1000 ft and vehicle velocity is 60 mph (88 fps), the critical rise (Hcr) in
the hill is
π 2
H cr 88 2 = 32 2 ft s2
1000
H cr = 421 3 ft
Under this condition, passengers would also feel no force with the seat and a sense of zero gravity.
Navigating in Geospace
The early 1960s marked the beginning of major research efforts to improve drilling technology.
A visionary goal was to navigate drill bits along a specified path and intersect a given target at
any subterranean location. Also, during this time, major efforts were under way to explore outer
space requiring space capsules to be placed in specified orbits or intersect specified space targets.
Requirements placed on directional drilling and space travel are ironically similar. Both require
engines of thrust, navigation, and the ability to make course changes.
Even though there are similarities, navigating through geospace offers the challenge of unpre-
dictable forces, while space travel navigates through a well-defined variable gravity field and envi-
ronmental drag. Both technologies have matured so that trajectories and targets are achievable with
high degrees of accuracy and reliability.
The key to drill bit navigation is the ability to monitor drill bit parameters, while drilling several
thousand feet into the earth, allowing well path monitoring and control. Accurately navigating drill
bit in geospace is vital in developing complex oil and gas reservoir structures and for capping dan-
gerous fluid eruptions. Multiple directional production wells are often drilled from one offshore
platform.
Tracking Progress Along a Well Path

The path into the earth’s crust that drill bits make is calculated incrementally. Three parameters are
needed to determine incremental steps of the path, Δx, Δy, Δz. They are (i) well bore inclination,
which is determined by a plumb line, (ii) azimuth of the tangent, which is determined by a compass,
and (iii) incremental distance drilled, which is determined by drill pipe movement down the
well bore.
The tangent to the well path at any location is defined by
a = sin α sin εi + sin α cos εj − cos αk 7 42

This unit vector is referenced from the right hand x, y, z axes as shown in Figure 7.12.
Three methods have been developed for determining well path shape while drilling. They are
Station 1 Δy
y (North)
ϵ
Δx
CL
Δz
x (East)
α
z
Station 2
Figure 7.12 Incremental steps between two survey stations.
1) Average angle method

2) Radius of curvature method
3) Minimum curvature method
Only the minimum curvature method is discussed here.
Minimum Curvature Method

This method is also called the circular arc method because the well path between stations 1 and 2 is
assumed to be an arc of a circle. The arc, in general, lies in a space plane (Figure 7.13) which is
defined by unit vectors tangent to the well path at survey stations 1 and 2.
The two unit vectors are defined in terms of hole angle (α) and direction (ε):
a1 = sin α1 sin ε1 i + sin α1 cos ε1 j + cos α1 k 7 43
a2 = sin α2 sin ε2 i + sin α2 cos ε2 j + cos α2 k 7 44
Arc distance between stations 1 and 2 is the course length, CL, and can be related to radius of
curvature of the arc by
cos β = a1 a2 vector dot product 7 45
Since
Rβ = CL = s 7 46
CL 180
R= 7 47
β π
The latitude, departure, and vertical distances between stations 1 and 2 are the components of
vector, B. This vector is
B = A1 + A2 7 48
where
β
A1 = R tan a1 7 49
2
β
A2 = R tan a2 7 50
2
y (North)
R
ß
1
x (East)
2
a1 a2
1
ß/2
A1
B
A2 2
Figure 7.13 Variables used in minimum curvature method.
Example As an example, consider the survey data as follows:

α1 = 40 α2 = 38
ε1 = 25 ε2 = 29
It is desired to determine incremental displacements of the well bore between location 1 and 2
given the following information. Assume a drilling interval of 93 ft:
a1 = sin 40 sin 25i + sin 40 cos 25j − cos 40k
a1 = sin 39 sin 29i + sin 38 cos 29j − cos 38k
a1 = 0 2717i + 0 5826j − 0 766k

a2 = 0 2985i + 0 5385j − 0 788k
cos β = 0 9984
β = 3 2394
93 180
R= = 1645 ft
3 2394 π
A1 = 1645 0 0283 a1 = 46 55a1
A1 = 46 55 0 2717i + 0 5826j − 0 766k
A2 = 1645 0 0283 a2 = 46 55a2

A2 = 46 55 0 2985i + 0 5385j − 0 788k
B = 46 55 0 5702i + 1 121j − 1 554k
B = 26 54i + 52 18j − 72 34k
The components of vector B define the incremental steps:

Δx = 26 54 ft
Δy = 52 18 ft
Δz = − 72 34 ft
The severity of the curvature of this section of the well bore is
3 239
Deg ft = = 0 0348 ft = 3 48 100 ft
93 ft
This is considered a high curvature.
Dogleg Severity
Dogleg severity (DLS) is a term used to quantify the severity of well path curvature, either inten-
tional or unintentional. DLS affects bending in drilling tools and casing as well as drill string fric-
tion, applied weight-on-bit (WOB), and fatigue. The rate of change of hole angle with respect to
distance drilled is (see Figure 7.14):
s = βR 7 51
π
s=β R 7 52
180
β 180
= ft 7 53
s πR
DLS is defined as the change in hole angle over a given drilled length. It is expressed in degrees/
100 ft of drilled hole:
β
DLS = 100 7 54
s
By substitution
18 000
DLS = 100 ft 7 55
πR
DLS defines the curvature of the well bore and affects bending of
all tubulars passing through the dogleg.
Vectors which define DLS are, in general, 3-D vectors. Once the
angle, β, has been established between two survey stations:
β
CL
R= 7 56
β rad
and R
CL DLS π
β rad = 7 57 s
100 180
18 000 1
R= ft 7 58
π DLS Figure 7.14 Dogleg severity.
For example
R = 1500 ft DLS = 3 8 100 ft

R = 150 ft DLS = 3 8 100 ft
The vector method provides a mathematical basis for calculating DLS in 3D space. The DLS based
on the minimum curvature method (previous example) is
18 000 1
DLS = = 3 48 100 ft
π 1645
Projecting Ahead
The change in well bore inclination or direction required to turn a well toward a target cannot be
achieved instantaneously. These changes have to be achieved over an interval of drilling while stay-
ing within a specified maximum dogleg severity. The actual angle change will be greater than the
estimated angle because of the footage drilled continually shortens the distance to the target. It is
therefore important to make corrections as early as possible.
Each survey station, in a way, is a new kick-off point. If the last survey station has a unit target
vector, a, then it is essential to be able to turn the hole in the plane of drilling toward the target and
stay within maximum dogleg severity limits. The situation depicted in Figure 7.15 shows that the
target can still be reached with a correction run having a radius of curvature, R. However, the max-
imum DLS must not the exceeded.
If drilling continues along direction, a, a greater dogleg will be required to turn the hole toward
the target. However, if drilling continues along direction a and the hole reaches point m, the hole
cannot be turned toward the target. Point “m” is a critical point and prior to reaching this critical
point, the well would have to be plugged back and sidetracked.
Planned well path
β
Current location
R
a Target
Figure 7.15 Needed correction to reach target.

Kinematics of Rigid Bodies 333
Kinematics of Rigid Bodies
Kinematics analysis of rigid bodies gives acceleration of the center of gravity and angular acceler-
ation. These two accelerations directly affect magnitude and direction of forces at connection joints.
Rigid bodies can experience three types of motion:
1) Translation – Any line within the body remains parallel.

2) Rotation – Every point in the body describes a circle.
3) General – Combination of translation and rotation.
The problem in each case is to determine the velocity and acceleration of any point in a rigid body
compatible with physical constraints. Knowing the velocity and acceleration of one point in a rigid
body along with angular velocity and angular accelerations is enough information to determine the
velocity and acceleration of any other point in the body.
Rigid Body Translation and Rotation

An example of all three types of motions in one system is the slider crank mechanism. Another
example is illustrated in the mechanism shown in Figure 7.16. A spring-mass contains a disc which
can be rigidly attached to the bar or it can be pinned at the bar attachment. In either case, the equa-
tion of motion is
I 0 θ + a2 kθ = 0 7 59
The natural circular frequency can be pulled directly from this equation, but question pertains to
the mass moment of inertia with respect to the pivot point, O for the two cases.
Case #1 – If the disc is pinned to the bar, the disc translates as a rigid body and its mass can be
viewed as a discrete mass of M. In this case,
I 0 = a2 m + M 7 60
Case #2 – If the disc is fixed to the bar, the disc rotates about fixed point, O, and the moment of
inertia is
r2
I O = a2 m + a2 M + M 7 61
2
There can be a significant difference between the two natural frequencies.
k M
O
r
m
a a
Figure 7.16 Mass moment of inertia with disc.

Case #3 – If the pivot support is replaced with a spring, the system vibrates at two degrees of free-
dom system with general plane motion.
General Plane Motion

Another example of general plane motion is Figure 7.17. The velocity of points A and B is known
and has horizontal directions. Accelerations of both points are assumed to be zero. The problem
here is to determine the angular velocity, ω, of the wheel. The relative velocity equation given ear-
lier is one approach. The instantaneous center method is used below.
The instantaneous center (IC) is a point at which the velocity of zero. Linear velocities at any
point in the body can be determined by multiplying radial distance by angular velocity. Angular
velocity can be determined by dividing linear velocity by radial distance.
The governing equations in this example are
vA = d + r A ω 7 62
vB = r B − d ω 7 63
Example For example, assume
r B = 6 in
r A = 4 in
V A = 5 fps
V B = 2 fps
Then
VA d + rA
= 7 64
VB rB − d
B
VB = 2 fps
6 in.
IC
d
4 in.
A
VA = 5 fps
Figure 7.17 Use of instantaneous centers.

5 d+4
=
2 6−d
5 6−d = 2 d + 4
d = 3 14 in
5 fps 12 in
ω= = 8 4 rad s CCW 7 65
7 14 in 1 ft
Dynamics of Particles
Dynamics refers to the motion of particles and rigid bodies when an imbalance of forces is applied
to them. The engineering problem is one of finding the resulting motion caused by an imbalance of
forces. The subject of dynamics is usually separated into kinetics and kinematics. Kinetics deals
with motion response caused by a force imbalance and is based on Newtonian physics, F = ma.
Kinematics is a study of motions without regard to forces. Kinetics is covered in this section.
Units of Measure
Two units of measure used in mechanics are Systems International (SI) and English System. Both
are used in nearly every industry around the world. There is a move toward the SI system interna-
tionally because its units are subdivided into tenths, thousands, micro, etc. Within each system,
there are base units and derived units. The derived unit in each case is based on Newton’s second
law. Base units for both systems are given in Table 7.1.
Force (Newton) in the SI system is the derived unit, the rest are base units. Mass (slugs) in the
English system is the derived unit, and the rest are base units. Once the base units are arbitrarily set,
corresponding derived units must satisfy the second law, F = ma. For example
F N = M kg a m s2 SI
Therefore, force (N) has units of kg(m)/s2. Similarly, mass is the derived unit in the English
system:
F lb = m slugs a ft s2
Therefore, mass (slugs) has units of lb(s2)/ft.

The acceleration at sea level due to gravity in each system as determined from Newton’s gravi-
tation law is
g = 32 2 ft s2 English
Table 7.1 Comparison of units of measure.
Force Mass Length Time
SI Na kg meter t, s
a
English lb slug feet t, s
a
Refers to derived units, the others are base units.
g = 9 81 m s2 SI
International Bureau of Weights and Measures in Sevres, France defines the standard kilogram
by the mass of platinum volume having a diameter of 39 mm and a length of 39 mm:
1 lb – weight of 0 4536 kg at sea level
1 lb – 4 4482 N
Application of Newton’s Second Law

The simplest application of Newton’s law is a particle in free fall:
W = Mg 7 66
Mass of an object is constant in any gravity field. Its weight, however, depends on local gravity.
When mass and g (at a given location in space) are known, then weight can be determined from this
equation. On the other hand, when weight is known near the surface of the earth, its mass can also
be determined from Eq. (7.66).
In any design, dynamic forces should be considered along with static forces. Total loads and stres-
ses are affected by both. Consider the two masses (Figure 7.18) connected by a cable and pulley.
Here, we need to know the total force in the cable in order to determine its required size and
strength.
Static Analysis
A static analysis establishes whether block A slides (break friction) or not. It will also provide a
reference load in the cable to compare with the predicted dynamic force.
Example Consider block A
T = f + 100 sin 30
N = 100 cos 30 = 86 6 lb
The friction force is
f = Nμ = 86 6 0 2 = 17 32 lb assuming μ = 0 2
A
T
f
100 lb N B
30º
200 lb
Figure 7.18 Internal dynamic loads.
Static tension in the cable required to break friction is

T = 17 32 + 50 = 67 32 lb
Under static conditions, a 200 lb tension in the cable is enough to break friction and cause both
masses to move and accelerate.
Dynamic Analysis
Freebody of mass A Freebody of mass B

M A a = T − f − 100 sin 30 M B a = 200 − T 7 67
100 200
a = T − f − 50 a = 200 − T 7 68
g g
Also
N = 100 cos 30
N = 86 6 lb
f = μN = 0 2 86 6 = 17 32
Adding both equations:
300
a = 150 − 17 32
32 2
32 2
a = 132 68 = 14 24 ft s2
300
Tension force in the cable is
200
T = 200 − 14 24 = 111 55 lb
32 2
The dynamic tension (111.55 lb) is much greater than the static tension (67.32 lb) required to
break friction.
Work and Kinetic Energy

Energy is a useful tool in engineering analysis. Energy is not a vector, so we wind-up with one
energy equation accounting for all mechanical work and changes in kinetic energy over a given
travel distance. The basic formula stems from Newton’s second law. Starting with
d2 x d dx dx
F=m 2 = m 7 69
dt dt dt dx
2 2
F dx = m v dv 7 70
1 1
The left integral represents the work done by force F between location 1 and location 2. The right
integral represents the change in kinetic energy between the two points:
1
Work = m V 22 − V 21 7 71
2
While the force does work on a mass to change its kinetic energy, the reverse is also true. The
kinetic energy of a mass can do work on something:
Work = KE 2 − KE1 7 72
Work can be performed by gravity, applied force, friction, and springs.
The arrangement of Figure 7.19 shows a mass being pulled downward by gravity toward a spring
of constant, k. The problem here is to determine the maximum compression (δmax) of the spring
when the mass moves into it and is stopped. The work of friction must be considered, as well as
gravity and the elastic compression of the spring. The kinetic energies at locations 1 and 2 are zero
so according to Eq. (7.72), total work on the system during the travel between locations 1 and 2
is zero:
1
Work = W l + δmax sin θ − f l + δmax − kδ2max = 0 7 73
2
1 2
kδ − δmax W sin θ − f − W sin θ − f l = 0 7 74
2 max
Example Assuming
W = 100 lb
k = 500 lb in
l = 15 in
θ = 45
μ=01
The friction force, f is
f = μN = 0 1 W cos θ = 0 1 100 cos 45 = 7 07 lb
The quadratic equation then becomes
250δ2max − 63 64δmax − 954 6 = 0
ℓ 2
f
δmax
N W
k
Figure 7.19 Work and kinetic energy.

The solution is
63 64 ± 63 642 − 4 250 − 954

δmax = = 2 08 in minus solution is trivial
2 250
Ignoring the middle term gives δmax = 1.95 in.
Potential Energy
Potential energy is also a useful concept. Potential energy may be stored in mechanical springs or
related to elevation of mass. Loss in potential represents work done by the potential energy source.
Stated mathematically
Work by the potential energy source = PE 1 − PE 2 7 75
For example, assuming the elevation of a mass (Figure 7.19) is measured by distance, y, from a
datum line. The work done by gravity is
W gravity = PE 1 − PE 2 = W y1 − y2 7 76
where y1 y2. By substitution into Eq. (7.72):

PE 1 − PE 2 = KE 2 − KE 1 7 77
giving
PE 1 + KE 1 = PE 2 + KE 2 7 78
which shows that total energy (potential plus kinetic) is conserved.
Next consider simple harmonic motion of a spring-mass system (right drawing, Figure 7.20).
Assuming at position 1, the spring is not stretched, then the potential energy at position 1 is
PE 1 = Wy1
W W
y1 y1
y2
y2
Datum Datum
Figure 7.20 Potential energy.

while the potential energy at position 2 is
1 2
PE2 = Wy2 + k y1 − y2
2
Still, Eq. (7.78) applies, allowing the velocity of the mass at position 2 to be determined.
The advantage of the conservation of energy is that we track energy, a scalar quantity, and not
force vectors. One energy equation accommodates only one unknown, however. For example,
assume we wish to fine the velocity of the mass at any position y2 starting with zero velocity at
y1. From Eq. (7.78)
1 2 1
Wy1 + 0 = Wy2 + k y1 − y2 + mV 22 7 79
2 2
1 2 1 W 2
W y1 − y2 − k y1 − y2 = V 7 80
2 2 g 2
k 2
V 22 = 2g y1 − y2 − y − y2 7 81
m 1
No vectors are involved.
Example Now consider the application of a safety belt tethered to a steel structure and attached
to a harness worn by a steel worker. Assume the belt is nylon 5 ft long and has a spring rate of 50 lb/
in. We examine the impact force applied to a 200 lb worker resulting from an accidental fall.
The change in potential from location 1 to the end and extension of the safely belt (location 2) is
60 in. plus δ, the stretch of the belt. Since the KE and both location is zero then the potential energy
at both locations must be the same:
PE1 = PE 2
where
PE1 = Wy1 7 82
1 2
PE2 = Wy2 + kδ 7 83
2
Equating the two equations and noting that (y1 − y2) = 60 + δ ft gives
1
200 60 + δ = 50δ2
2
400 60 + δ = 50δ2
δ2 − 8δ − 480 = 0 7 84
−b ± b2 − 4ac
δ= 7 85
2a
δ = 26 27 in
Force applied to the worker by the safety belt is
F = kδ = 50 26 27 = 1314 lb a very large and perhaps harmful force
Drill Bit Nozzle Selection

One of the most useful applications of conservation of energy is in fluid flow. Bernoulli’s equation,
which is based on the conservation of energy, accounts for energy going in and out of a control
volume. Consider the fluid flow of drilling fluid flow through drill bit nozzles. The practical drilling
problem is to determine nozzles (usually three nozzles) that would convert available hydraulic
horsepower into maximum kinetic energy for cleaning cuttings from under a drill bit. The total
cross-sectional area of flow is called total flow area or TFA.
Bit pressure drop depends on TFA of all nozzles in a drill bit. Applying Bernoulli’s energy equa-
tion between sections 1 and 2 (see Figure 7.21) gives
p1 V2 p V2
+ 1 = 2 + 2 7 86
γ 2g γ 2g
Since V1 is small by comparison with V2
γ 2
p1 − p2 = Δp = V V 2 is replaced by V 7 87
2g
In oil field units
7 48γ
Δp = V2 7 88
2 32 2 144
γV 2
Δp = 7 89
1240
where
Δp − psi
V − fps (note the unit change)
γ − ppg
Applying a nozzle coefficient of 0.95 [2] gives
γV 2
Δp = 7 90
1120
The continuity flow equation is
0 32Q
V= fps 7 91
An
Figure 7.21 Fluid flow through bit nozzles. Bit nozzle

(TFA)
1
Δp
2
Dissipated energy
By substitution
γQ2
Δp = 7 92
10 938An 2
where
Δp − psi (scales directly to mud weight)

Q − gpm
γ − ppg (pressure scales directly with γ)
An − total nozzle area (TFA), in.2
which predicts pressure drop (Δpbit) across bit nozzles.

Consider, for example, the pressure drop across three (3), 11 32 nozzles (TFA = 0.2784 in.2) is
Δpbit = 1887 psi assuming mud weight of γ = 10 ppg and a flow rate of 400 gpm. It is desirable
to use up available horsepower for best bottom hole cleaning. This is accomplished by selecting
the best flow rate and then solving for total nozzle flow area.
ΔpQ
Dividing bit hydraulic horsepower HHP = by the cross-sectional area of the hole gives
1714
hydraulic horsepower per square inch of hole or HSI. An HSI of 5–7 is considered a high level of
hydraulic horsepower across roller cone drill bits.
Impulse–Momentum
The principle of impulse–momentum is derived directly from Newton’s second law. Assuming lin-
ear motion
d dx dv
F=m =m 7 93
dt dt dt
2 2
F dt = m dv 7 94
1 1
After integration
2
F dt = m V 2 − V 1
1
Im = mV 2 − mV 1 7 95
Im = H 2 − H 1 7 96
The letter “H” represents linear momentum.

The impulse–momentum method is useful when the applied force is known as a function of time.
A good example is the take-off and landing of aircraft. In this case, the thrust capability of jet
engines is a better parameter than a horsepower rating. Impulse and momentum are vector quan-
tities and are related as follows (Figure 7.22).
Figure 7.22 Linear impulse and momentum. F

F(t)
Smooth surface
F(t)
V1 V2
T
1 2
Figure 7.23 Thrust from gas turbine. T, kips

Lift-off
50
t, s
5 t
For example, assume a lift-off requirement of an airplane is 150 mph. Weight of airplane = 100 kip.
Maximum engine thrust capability is 50 000 lb. Assume total engine thrust at take-off varies according
to Figure 7.23.
Impulse in this example is the area under the triangle and rectangle. Since 60 mph = 88 fps
1 100 88
5 50 + t − 5 50 = 150 −0
2 g 60
t = 16 17 seconds
The airplane would achieve the 150-mph lift-off speed in 16.17 seconds.
Remember, impulse and momentum are vector quantities, so in planar motion, both may have x
and y components. Jet and rocket engines are rated in terms of thrust, which is easy to convert into
changes in momentum and velocity.
Impulse–Momentum Applied to a System of Particles

Linear momentum is useful for determining the total force required to change direction of fluid
flow by a diverter, such as a turbine blade. The general discussion is made by considering flow
within a control volume (Figure 7.24), which shows a fluid plug (A1Δs1) moving into the control
volume past section 1.
Control volume
V2
1
ΔS1
V1 ΔS2
2
Figure 7.24 Resultant force required to deflect fluid.
The same amount of fluid (A2Δs2) exits the control volume past section 2. The sum of total change
in momentum of all fluid particles internal to the control volume is unchanged except for volume of
fluid entering and leaving the control volume:
Fluid momentum in = V 1 Δs1 A1 ρ
Fluid momentum out = V 2 Δs2 A2 ρ
Applying impulse–momentum principles to the system of fluid particles gives
F i + f i Δt = V 2 Δs2 A2 ρ2 − V 1 Δs1 A1 ρ1 7 97
Internal forces, f i between elemental masses cancel because they exist in equal and opposite pairs
leaving only the sum of the external forces. Also, to satisfy continuity of flow
Δs2 Δs1
A2 ρ = A 1 ρ = Qm
Δt Δt
Therefore,
F i = F = Qm V 2 − V 1
This force represents the total force applied to the control volume to bring about the change in
momentum between sections 2 and 1. According to Newton’s third law, there is an equal reaction
(R) on the engine frame and airplane:
R = Qm V 2 − V 1 7 98
where Qm is mass flow rate through the control-volume. V 1 and V 2 are absolute fluid velocities
(vectors) into and out of the control volume.
If the control volume moves with a given velocity, the reaction equation remains the same keep-
ing in mind Qm is mass flow rate through the control volume, while V 1 and V 2 are absolute velocities
of the fluid entering and leaving the control volume. A major application of this theory is the pre-
diction of torque and power generated by fluid across turbine blades.
Mechanics of Hydraulic Turbines

Downhole turbines were introduced into drilling programs before PDMs. They generate power at
a much higher rotary speed. Turbine speed is approximately double that of PDMs. Assuming an
output speed of 800 rpm and bit torque of 1500 ft-lb
2π 1500 800
Pbit = = 228 4 hp
33 000
a factor of 8 increase over normal rotary speeds of 100 rpm. However, maximum power of drilling
turbines is developed at one rotational speed, and this optimum speed is sometimes difficult to
monitor and control.
Downhole drilling turbines develop torque fundamentally differently from positive displacement
motors (PDMs). Torque is developed by fluid momentum, while torque in a PDM is developed by
direct application of fluid pressure. Output speed of a turbine depends directly on output torque,
while output speeds of PDMs depend on flow rate and is independent of torque. Maximum power is
developed at a particular rotor speed and herein lies a monitor/control issue. Monitoring and oper-
ating turbines at their optimum power speed is necessary for best use of drilling turbines. Since
turbine torque is the result of fluid momentum, torque output depends on both drilling mud density
and flow rate.
Each turbine stage has a fixed set of blades (stator) and a rotating set of blades (rotor). Stator
blades direct drilling fluid onto rotor blades where output torque is generated. Output torque, out-
put power, and overall pressure drop of a turbine is a direct multiple of the number of stages in the
turbine.
When the rotor blade is stationary, fluid momentum changes the most, and the torque is the
greatest. When the rotor blade is moving, fluid momentum is not changed as much so torque drops
off. The faster the rotor turns, the lower the torque generated by the rotor blades. At a high speed of
rotor rotation, fluid leaving the stator blades cannot catch the rotor blade, so its momentum does
not change. This speed is the runaway speed, and at this speed, no torque is developed on the rotor
blades.
In practice, turbine rotational speed is not preset; output speed responds to applied torque or bit
weight. Torque is applied, and rotational speed automatically adjusts to the torque. Both stall torque
and runaway speed change with mud weight and flow rate. Turbines are rated at maximum power.
Whether maximum power is developed by downhole turbines depends on the ability to monitor
and control bit torque.
The mechanics of turbine power is explained by considering momentum changes across the con-
trol volume, abcd (Figure 7.25). The rate of change of fluid momentum across cd and ab requires a
force vector, R.
R = Qm V 2 − V 1 7 99
where
V 1 − absolute velocity of fluid leaving stator blade boundary ab

V 2 − absolute velocity of fluid leaving rotor blade boundary cd
Qm − mass flow rate through control volume
R − resultant force on fluid in control volume
U − tangent velocity of rotor blade (center of blade)
Stator blade
UH
β
a b
V1 u
θ
U Rotor blades
z
c d
β
u
V2/b
U
V2
Figure 7.25 Rotor and stator turbine blades.
The component of this vector in the tangent direction is
Rθ = Qm U − 2V cos β 7 100
or
2u
Rθ = Q m U − 7 101
tan β
Since this is the force acting on the fluid in the control volume, the fluid force applied to the
blade is
2u
F θ = Qm −U 7 102
tan β
This force produces output torque. Total torque generated by one turbine stage is
2u
T = rQm −U 7 103
tan β
where Qm is now the total mass flow rate through the turbine and U is the tangent speed of the rotor
and r is the average radial distance. Also
1−ε Q 1−ε Q
u= = 7 104
Aλ λ A
where
u – axial velocity of fluid entering (and leaving) control volume

ε – coefficient of fluid loss around outside of rotor
λ – contraction coefficient due to flow area occupied by blade thickness
A – cross-sectional area of space occupied by rotor (or stator) blade
In consideration of these factors, the torque generated by one turbine stage is

2 1−ε Q
T = rρQ −U 7 105
Aλ tan β
where Qm has been replaced by ρQ. U relates to rotational speed by
2πN
U = rω = r
60
The actual rotational speed, N, depends on applied torque. In practice, rotational speed of the drill
bit depends on the torque developed by the drill bit (or WOB).
From Eq. (7.105), maximum torque occurs when U = 0 or the rotor blade is stalled:
γ 2 1 − ε Q2
T max = r stall torque, T st 7 106
g Aλ tan β
On the other hand, output torque is zero when
2 1−ε Q
U= 7 107
Aλ tan β
This velocity is the runaway speed of the turbine. In terms of angular velocity of the turbine,
60 2 1 − ε Q
N rs = runaway speed 7 108
2πr Aλ tan β
When the blade tip angle (β) decreases from 30 to 20 , maximum torque and runaway speed
increase by a factor of 1.59.
Output mechanical horsepower is formulated in terms of rotor/stator design parameters, drilling
fluid density and flow rate.
Starting with
γ 2 1−ε Q
T=r Q −U 7 109
g Aλ tan β
and
U
Pout = Tω = T 7 110
r
Power output is defined by
γ 2 1−ε Q
Pout = Q −U U 7 111
g Aλ tan β
Maximum power occurs at

d Pout 2 1−ε Q
= − 2U = 0 7 112
dU Aλ tan β
or
1−ε Q
U= 1 2 runaway speed 7 113
Aλ tan β
which gives
γ 1−ε 2 3
Pmax = Q mechanical output 7 114
g Aλ tan β
Note that maximum power is achieved at ½ runaway speed.
Pressure drop across one turbine stage is determined by applying Bernoulli’s equation to the con-
trol volume shown in Figure 7.26:
p1 p ft-lb
= Eout + 2 Bernoulli,
γ γ lb
Δp
P= γQ conversion to power
γ
At maximum power output,
Pmax = ΔpQ 7 115
and
γ 1−ε 2 2
Δp = Q 7 116
g Aλ tan β
Total pressure drop across multirotor turbines creates a downward force that must be supported
by thrust bearings.
The above equations predict the performance for drilling turbines. Actual performance is estab-
lished experimentally as explained earlier. Mechanical output is typically less than hydraulic power
input to turbines. Efficiency of the power transformation is defined by
HPout
η= 7 117
HHPin
u
Rotor r
1
2
Hub
Figure 7.26 Control volume of flow through a turbine rotor.

Performance Relationships
If performance information is known for one set of operating parameters, the performance at other
operating parameters can be calculated using
Q2
N2 = N1 7 118
Q1
2
Q2 γ2
T2 = T1 7 119
Q1 γ1
2
Q2 γ2
Δp2 = Δp1 7 120
Q1 γ1
3
Q2 γ2
HP2 = HP1 7 121
Q1 γ1
These formulas are based on the preceding derivation.
In many cases, performance data is given for one mud weight or a given flow rate. The above
equations are useful to predict performance at different mud weights or flow rates. For example,
consider the data in Table 7.2. Assume that speed (N1 = 1010 rpm), torque (T1 = 1430 ft lb), pressure
drop (Δp1 = 1450 psi), and output power (HP1 = 275 hp) are known at a flow rate of 475 gpm. Fur-
thermore, it is desired to determine these parameters at a flow rate of 400 gpm with density staying
the same at 12 ppg. The conversion is as follows:
400
N2 = 958 = 807 rpm
475
2
400
T2 = 2390 = 1695 ft-lb
475
2
400
Δp2 = 2241 = 1589 psi
475
3
400
HP2 = 436 = 260 hp exponent
475
These results compare exactly with the numbers in Table 7.2. Converting the data to a different
mud weight is accomplished using the above equations.
Table 7.2 Performance date for 63/4 in. turbine (12 ppg).
Pump Bit Power Torque Pressure Thrust force

rate (gpm) speed (rpm) output (hp) (ft-lb) drop (psi) (1000 lb)
250 504 64 662 621 10

300 605 110 953 894 14
350 706 175 1297 1217 19
400 807 260 1695 1589 25
450 906 371 2145 2011 31
500 1009 509 2648 2483 39
Maximum Output of Drilling Turbines

The torque and rotary speed are related graphically as a straight line (Figure 7.27). Both stall torque
and runaway speed change with mud weight and flow rate. Mechanical power delivered by turbines
can be derived from the torque-speed line as follows:
T st
T = T st − N 7 122
N rs
where
Tst − stall torque

Nrs − runaway speed
The mechanical output power is

TN
P= hp 7 123
5252
When Eq. (7.149) is substituted in to this equation, the result is a parabola that reaches a max-
imum value at ½ stall torque and ½ runaway speed (Figure 7.27):
T st N
P= N 1− hp 7 124
5252 N rs
This is a parabolic function that reaches a maximum value at one-half stall torque and one-half
runaway speed. Because turbine speed depends on applied torque, mechanical power delivered by
turbines depends on applied torque. When applied torque changes, output speed changes, and
power is delivered at particular combinations of torque and speed. A means of monitoring turbine
speed is necessary to make maximum use of turbine power under drilling operations.
Consider the performance data given in Table 7.2 for a 63/4 in turbine operating in 12 ppg mud.
Assuming a flow rate of 400 gpm:
•• Bit speed is 807 rpm

Output torque is 1695 ft lb
•• Maximum mechanical power is 260 hp

Pressure drop across the turbine is 1589 psi
Torque
Stall torque
Power
Runaway speed
Output speed
Figure 7.27 Performance curves for drilling turbines.

According to the equations given above:
•• Runaway speed is 1614 rpm

Stall torque is 3390 ft lb
The amount of hydraulic horsepower consumed under the above conditions is

ΔpQ 1589 400
HHP = = = 371 hp
1714 1714
TN 1695 807
HP = = = 260 hp
5252 5252
Maximum efficiency of the turbine then is
HP 260
η= = =07
HHP 371
or 70%. The efficiency curve is also parabolic because HHP is constant and HP is parabolic. Effi-
ciency is maximum when mechanical power is maximum.
Dynamics of Rigid Bodies
Planar motion of rigid bodies is considered here. A rigid body can be viewed as a composite of many
particles connected into one body. In this case, the equations of motion of rigid bodies are deter-
mined from the sum of the effects of all particles. Applying the second law to each particle and then
summing all forces both internal and external gives
F = maG 7 125
where F is the vector sum of all external forces applied to the rigid body and aG is the acceleration
of the center of gravity of the rigid body. The sum of all internal forces is zero since they occur in
equal and opposite pairs. This single vector equation produces two scalar equations for planar
motion.
A second equation is developed from consideration of rate of change of angular momentum and
moments measured about an axis (point P) perpendicular to the plane of motion (Figure 7.28).
Consider the axis passing through point P, attached to the moving body. By summing all
moments applied to the body about P and equating them to all angular inertia forces gives
M P = I P α − ym aP x + xm aP y 7 126
This general expression reduces to
MP = IPα 7 127
when any one of the following three conditions are met.
aP = 0
d = 0 for aP
aP passes through the center of gravity, G
Each are useful in setting up equations of motion for certain cases.
maG
–x
y G
–y
maP α
Figure 7.28 Rigid body in planar motion.
Rigid Bodies in Plane Motion

Equation (7.126) can be transformed into another very useful form, by kinematic relationships:
M P = I G α − ym aG x + xm aG y
In this case, unlike IP, IG is a property of the body. This means that point P can be located either
within the body or outside the body provided moments of both the external forces and inertia forces
are taken with respect to the same point P (Figure 7.29). In this case, the moment equation can be
viewed as equating a force diagram to a kinetic diagram
MP = P 7 128
where
MP is the summation of all moments with respect to P (left diagram)
P is the sum of all kinetic effects (right diagram)
The location of point P is arbitrary and is factored into both sides of the equation as shown in
Figure 7.29.
This useful concept is based on the force and kinetic equivalence. This equality can easily be
visualized by equating a free body force diagram to a free body kinetic diagram (Figure 7-29).
Two reference points are illustrated in Figure 7.30: (i) Point P is fixed on the plane of rolling and
(ii) point P is on the roller in contact with the rolling plane. The advantage of the first case is that the
friction force on the roller has no moment with respect to point P so
2rF = I G α + r rα m
1
2rF = r 2 mα + r 2 mα
2
3
F = rmα
4
4F
α=
3rm
The friction force is also removed by use of the second approach:
3
2rF = r 2 mα
2
M IG α
F1
m aG
G
r1
F2
r2 d
P P
Figure 7.29 Force – kinetic equivalence.
F F
G G
P d
Figure 7.30 Reference point “P” alternatives.
4F
α=
3rm
These diagrams graphically illustrate the above equations showing that
MP = P 7 129
Giving
M P = I cg α + dmacg 7 130
and
F = macg 7 131
These two equations of dynamics will now be applied to rigid bodies in plane motions under the
following conditions:
•• Plane translation
Pure rotation
• General motion
12 ft
3 ft ma
10 ft
50 lb
20 lb
FA FA
Figure 7.31 Rigid body translation.
Translation of Rigid Bodies

Bodies in pure translation have no rotational component; any line in the body stays parallel to its
initial orientation. The force and kinetic diagrams are useful in determining reactions throughout
(Figure 7.31):
Fy = 0; FA + FB = 50 (y component of Eq. (7.125))
M G = 0; 6F B + 3 20 − 6F A = 0
Combining gives, FA = 30 lb and FB = 20 lb. The x component of Eq. (7.125) gives
F x = ma
50
20 = a
32 2
a = 12 88 ft s2
Rotation About a Fixed Point

In this case, each point in the rigid body travels along a circular path around a center of rotation.
A special case of rigid body dynamics is a bar pinned at one end with freedom to rotate or oscillate
(Figure 7.32). Each point in the rigid body travels along a circular path about the pinned end.
Motion could be induced by aP = 0 an applied torque by the force of gravity. Reaction forces at
the support and motion of the body and angular motion, θ, are to be determined.
The three unknowns in this type of problem are force reaction components at the support and
angular acceleration, α.
The equation of motion is (noting that aP = 0)
M0 = I0θ 7 132
L
− WL + w sin θ = I 0 θ 7 133
2
which defines the angular acceleration of the body. A useful form of

the equation is
1 L 0
θ+ WL + w sin θ = 0 7 134
I0 2
The solution of this differential equation defines the angular L
motion, θ, as a function of time, t. In its present form, the equation m
is nonlinear. However, for small angular displacements, it is writ-
ten as θ
L w
θ+ W+ θ=0 7 135
I0 2
The natural circular frequency is
R
L w
ω2 = W+ M
I0 2
An interesting aspect of this pendulum is how the disc is
mounted on the bar. If the disc is fixed to the bar and rotates with Figure 7.32 Dynamics of a
the bar, then two component pendulum.
1 1 2
I0 = mL2 + L2 M + RM 7 136
3 2
However, if the disc is pinned to the bar by a bearing, then
1
I0 = mL2 + L2 M 7 137
3
and the disc translates as a rigid body.
Center of Gravity of Connecting Rod

A simple method of obtaining this information is as follows. The center of gravity can be deter-
mined by supporting the connecting rod (Figure 7.33) and measuring the forces, RA and RB. The
location of the center of gravity is
aW = RB L 7 138
RB
a= L 7 139
W
RA
W RB
a
L
Figure 7.33 Model for determining center of gravity.

Mass Moment of Inertia of Connecting Rod

It is difficult to analytically calculate the mass moment of inertia of
connection rods because of its complex geometry. This information
is typically determined experimentally. One method is based on the
L–a period of oscillation of a simple pendulum as shown in Figure 7.34.
The period of a rigid body pendulum is determined from the equa-
tion of motion:
cg
d2 θ
I0 − L − a Wθ = 0 7 140
dt 2
L−a W 1
ω2 7 141
I0 s2
Since ω is rad/s and there are 2π rad per cycle
ω
f = cps 7 142
2π
θ θ
The period of each oscillation is
Figure 7.34 Oscillation of a
connecting rod. 1
T= seconds 7 143
f
Example Consider
W = 6 lb
L – a = 6 − 2 = 4 in
T = 1 5 seconds determined experimentally
Back calculations give I0 from the period of oscillation:
2π
ωn = = 4 19 rad s
T
L−a W
I0 =
ω2
46
I0 = = 1 3671 in -lb-s2
17 556
The mass moment of inertia with respect to the center of gravity is determined from the transfer
formula:
I cg = I 0 − L − a 2 m 7 144
I cg = 1 3671 − 42 386
6
= 1 1183 in -lb-s2
General Motion of Rigid Bodies

In general, rigid bodies may move in planar motion with a combination of translation and rotation.
Movement from one position to another can be viewed as separate motions of translation and then
rotation.
Example A disc translates and rotates simultaneously as shown in Figure 7.35.
The spool has a mass of 8 kg and a radius of gyration of kg = 0.35 m with respect to its center.
Determine the spools angular acceleration and the force in the cord if a force of 100 N is applied
as shown.
100 N
0.2 m 0.5 m
100 N T MaG
IG α
=
78.48 N
Figure 7.35 General motion of a rigid body.
Equations of motion are
TG = IGα
0 2 100 − 0 5T = I G α 7 145
and
F = MaG
100 + T − 78 48 = MaG 7 146
Note that
2
I G = Mk 2G = 8 kg 0 35 m = 0 98 kg-m2
The relation between aG and α is
aG = 0 5α
Combining Eq. (7.145)

20 − 0 5T = 0 98α
And Eq. (7.146)
21 52 + T = 8 0 5 α
gives
α = 10 32 rad s2
aG = 5 16 m s2
T = 19 77 N
The static tension, T, and force, F (left cable), are T = 22.42 N and F = 56.06 N. The increase in
applied force from 56.06 to 100 N produced the upward acceleration and angular acceleration caus-
ing tension T to drop from 22.42 to 19.77 N.
Example In another example, consider a disc that is attached to a rail accelerating horizontally at
3 fps2. Determine the reactive force components at support A and the angular acceleration of the
disc in its current position.
Here a disc is pinned to a runner having an acceleration of 3 fps2. The disc weighs 15 lb having a
mass moment of inertia with respect to the center of gravity, G, of
1 2
IG = mr
2
It is desired to find the five unknowns, Ax, Ay, aGy, aGx, α.
The freebody diagram and kinetic diagrams, shown in Figure 7.36, yield
F y = maGy 7 147
Ay − W = maGy
F x = maGx 7 148
aA = 3 fps2
r = 2 ft
cg
Ay
A
Ax macg
W = 15 lb
Ig α
Figure 7.36 Example of general motion.

Ax = maGx
M A = I G α − rmaGx 7 149
0 = I G α − rmaGx
rm
α= aGx
IG
Kinematic relationships
aG = aA + aG A 7 150
aGx i + aGy j = 3i + aGA n j − aG A ti 7 151
aGx = 3 − rα
aGy = 0
Dynamic Forces Between Rotor and Stator

PDMs are commonly used in bottom-hole assemblies to provide greater power to drill bits and to
help navigate drill bits in directional drilling operations. Internal components are shown in
Figure 7.37.
These motors (and pumps) contain a rotor and a stator each possessing special gear-shaped geo-
metries which mess together in such a way to form cavities, which progress in the longitudinal
direction. These motors have multiple cavities or stages. The number of lobes establishes the rela-
tion between volume flow rate and rotor speed. The stator always has one more lobe than the rotor.
In general, the high lobes perform with lower flow rate, the higher the torque. Moineau pumps are
often used in petrochemical production facilities.
Side loads are generated by the rotor against the elastomer molding of the stator as the rotor
moves with planetary motion within the stator. These side loads can cause (i) premature failure
of the elastomer lining at high rotor speeds and (ii) excite lateral modes of vibration over the drill
e
Stator
cg Fr
Rotor
Figure 7.37 Pitch circles of rotor and stator.

Table 7.3 Planetary relationships.
Arm, e Rotor, r Stator, s
With e 1 1 1
Relative to e 0 rs −1
−
rr
rs
Total 1 1− 0
rr
collar section, depending on force frequency and magnitude. Figure 7.37 shows the pitch circles of
both rotor and stator. The force, Fr, represents a radial force interactive force between rotor and
stator liner. It is the reactive force on the stator that can excite lateral modes of vibration. It passes
through the center of gravity of the rotor according to the earlier discussion of rigid bodies.
A common lobe arrangement is the 1 : 2 motor/pump. In this case, the diameter of the pitch circle
of the rotor is one-half that of the stator. As the rotor travels with planetary motion within the sta-
tor, the center of the rotor progresses along a circular path. In general, the relation between arm
rotation of (e) and rotation of the rotor is
ωr rs
= 1−
ωe rr
as shown by Table 7.3.
The darker circle represents the cross section of the rotor lobe. The rotor moves with general type
motion, i.e. simultaneous rotation and translation. At any one location, the acceleration of the cen-
ter of gravity of the rotor is
aG = ω2 e 7 152
where
e = pitch radius of the rotor

ω = angular velocity of the rotor
Note that any cross section of the 1 : 2 rotor travels within a transverse plane per the following
equations:
x = r cos θ + cosϕ
y = − r − sin θ + sin ϕ
These equations show that the path of center of the rotor lobe is a straight line; however, the path
of the center of the pitch circle is a circle of radius r.
Consider the following parameters:
ns
= 2 lobe ratio
nr
N = 200 rpm
2πN
ω= rad s
60
W r = 200 lb
r r = pitch radius of rotor 2 in
m1 aG1
IG1 α1 F1 M y
F
IG2 α2
cg
cg x
+M
z
d1
F2 d2 m2 aG2
Figure 7.38 Interconnecting bodies.
From the kinematics of the rotor/stator,

ns
ω= 1− ωe 7 153
nr
meaning that vector, e, rotates counterclockwise with angular velocity of ω. The acceleration of the
rotor center is acg = ω2e = ω2rr. The force, Fr = Macg or
2
200 2π200 2
Fr = = 454 lb
32 2 60 12
The force vector rotates with ωe = 20.94 rad/s or a frequency of 3.33 cps. The force magnitude and
frequency are both important in determining possible resonance of lateral modes of vibration.
Interconnecting Bodies
In many mechanical situations, rigid bodies are interconnected. Each rigid body can be analyzed
separately. However, the analysis is greatly simplified by applying the equations of motion to the
assembly (Figure 7.38). The connecting forces are not involved in the equations because they are
interactive and cancel out. The equations of motion are applied as follows:
MP = I Gi αi + di mi aGi 7 154a
F ix = mi aGi x 7 154b
F iy = mi aGi y 7 154c
where the subscript “i” refers to each body.

The advantage in this approach is that internal forces are not involved in the calculation and can
greatly simplify a dynamic analysis. The last three scalar equations, however, allow for only three
unknowns. If there are more than three, then the system then part of the system may have to be
isolated.
Gear Train Start-Up Torque

As an example, consider a simple gear set driven by a motor as shown in Figure 7.39. Assume that
each gear has relatively large angular moment of inertia which has to be considered during the
x
T2, ω2, α2
F
T1, ω1, α1
r1 I1 I2
y
P
r2
TM F
Motor
T0 (external)
Figure 7.39 Start-up torque.
start-up of the gear drive. These inertias will affect the startup time required to reach the desired
transmitted torque. The angular acceleration during the start-up period will now be discussed by
applying the above equations.
From Eq. (7.154a) and including the motor armature as part of I1,
T M + T o − r 1 + r 2 F = I 1 α1 − I 2 α2 7 155
Moments are taken with respect to point P.
From kinematics,
r 1 θ1 = r 2 θ 2 7 156
r 1 α1 = r 2 α2 7 157
Accounting for directions,
r1
α2 = α1 7 158
r2
Bringing these equations together gives
r1
T M + T 0 − r1 + r2 F = I 1 − I 2 α1 7 159
r2
From a freebody diagram of gear 1,
T M − r 1 F = I 1 α1
T M − I 1 α1
F=
r1
The progression of substitution and simplification follows:
r1 r1
TM + T0 − 1 + T M − I 1 α1 = I1 − I 2 α1
r2 r2
r2 r1 r1
T o + I 1 α1 − T M + I 1 α1 = I1 − I 2 α1
r1 r2 r2
2
r1 r1
TM = T0 + I1 + I2 α1 7 160
r2 r2
If inertia is not considered, then

r1
TM = T0 7 161
r2
Motor limitations may require that gear inertia be overcome first before applying the external
load. This would also allow the large gears to serve as flywheels. Eq. (7.160) could be used to deter-
mine time to reach rated speed of the motor.
Kinetic Energy of Rigid Bodies

Considering rigid bodies are made up of many particles, then Eq. (7.72) also applies except total
kinetic energy is equal to
1 2 1
KE = Mv + I g ω2 7 162
2 g 2
Work may be produced by either torque or forces or both.
W = FΔx = TΔθ 7 163
Friction plays an important role in maintaining pure rolling of a disc. However, friction in this
case does no work provided there is no sliding. Recall
dx
dW = F dx = F dt = FV dt 7 164
dt
Forces at points of zero velocity do no work.
Example Consider the disc and spring arrangement in Figure 7.40. The potential energies in this
problem are gravity and the elastic energy of the spring:
h1
h2 s 1
θ
Figure 7.40 Principal of work and kinetic energy.

2
PE1 = h1 W + ½ k δst
2
PE2 = h2 W + ½ k s + δst
The kinetic energies are
KE1 = 0 disc starts at rest
1 1
KE2 = MV 22 + I G ω22 = 0 maximum compression of spring
2 2
Since rolling friction does zero work
PE1 + KE 1 = PE 2 + KE 2 7 165
In this case,
PE1 = PE 2
2 2
h1 W + ½ k δst = h2 W + ½ k s + δst
2 2
W s sin θ = ½ k s + δst − ½ k δst
This equation is quadratic. If, however, δst = 0, then
1
W sin θ = ks
2
and
2W sin θ
s= 7 166
k
In this case, the friction force at the contact point does zero work.
The Catapult
Catapults of antiquity stored elastic energy in flexible beams, such as in composite bows made up of
animal sinew (tension), wood, and bone (compression). This energy is released suddenly to propel a
warhead. In modern times, catapults are used on aircraft carriers to assist in launching aircraft.
When the composite bow reached its limits, new elastic systems were needed to increase range
and size of projectiles. Eventually, the bow was replaced with torsion springs. Each torsion spring
was a bundle of sinew strands with a wooden arm thrust through its center. Each sinew strand was
thread size. The torsion spring was lighter and more powerful (energy storage – 20 times greater
than steel per pound).
When the trigger was pulled of, the energy stored in the torsion springs was used to accelerate the
projectile, but a significant portion was also used to accelerate the wooded throwing arm. The for-
ward motion of the arms was eventually arrested by the sling snapping taut. Since the tightening of
the sling also contributed to the forward motion of the projectile, nearly 100% of the elastic energy
was transmitted to kinetic energy of the projectile. A U-joint at the support point gave the mech-
anism more flexibility in aiming the projectile. Catapults could throw 50 lb weights. These improve-
ments accelerated the arms race at that time.
Impulse–Momentum of Rigid Bodies

The equations of motion for rigid bodies can be extended to yield angular impulse and angular
momentum expressions (Figure 7.41). Starting with
Position 1 Position 2
M
G
G
F1
F2
Figure 7.41 Impulse–momentum of rigid bodies.
F = maG 7 167
F dt = m vcg,2 − vcg,1 7 168
F dt = H 2 − H 1 7 169
and
M P = I G α + dmaG 7 170
dω dvG
MP = IG + dm 7 171
dt dt
M P dt = I G ω2 − ω1 + m d2 vG2 − d1 vG1 7 172
Linear Impulse and Momentum

Equation (7.169) provides two algebraic equations:
F x dt = H x2 − H x1 7 173
F y dt = H y2 − H y1 7 174
where H is linear momentum.
Angular Impulse and Momentum

The angular impulse–momentum equation provides a third equation:
M P dt = GP2 − GP1 7 175
which is rewritten as
GP2 = M P dt + GP1 7 176
where G is angular momentum reference from an arbitrary point P.

M
1 2
P
f
t1 t2
N
2
IG ω 1 M dt IG ω2
1
W
m vG1 m vG2
2
f dt
N 1
Figure 7.42 Impulse–momentum (two equations).
The application of these expressions is illustrated with the situation in Figure 7.42. Here a disc
rolls on a horizontal flat surface with an applied moment, M applied to it. At time, t, the disc has
angular velocity, ω1. The problem is to determine, ω2 at time t2. We have three momentum equa-
tions to use.
Linear impulse–momentum (two equations):
mvG1 + f dt = mvG2 7 177
and
0+ N − W dt = 0 7 178
The second equation tells us that N = W. The first equation gives
f dt = mv2
1
We cannot say at this point that slippage is impending.

Angular impulse–momentum (one equation):
I 0 ω1 + M − rf dt = I 0 ω2 7 179
1
Mt 2 − r f dt = I 0 ω2
1
Mt 2 − rmv2 = I 0 ω2
but v2 = rω2, so
Mt 2 = I 0 + r 2 m ω2 7 180
and
Mt 2
ω2 = 7 181
I G + r2 m
Example The disc shown in Figure 7.43 accelerates to the right due to force, F. The magnitude of
this force increases with time according to F = t + 10 N.
The problem is to determine the angular velocity of the disc after five seconds. Assume the disc
starts from rest. The disc weighs 981 N and has a mass moment of inertia about its center of
IG = 12.25 kg m2.
The angular impulse–momentum equation is
GP2 = M P dt + GP1 7 182
By substitution,
5
I P ω1 + 1 15F t dt = I P ω2 7 183
0
5
I P ω1 + 1 15 t + 10 dt = I P ω2 7 184
0
1 2
0.4 m
0.75 m
P
A FA
NA
Figure 7.43 Impulse–momentum (one equation).

With ω1 = 0,
5
1 2
1 15 t + 10t = I P ω2
2 0
1 15 12 5 + 50 = 12 25 + 0 752 100 ω2
71 87 = 68 5ω2
ω2 = 1 05 rad s
Angular Impulse Caused by Stabilizers and PDC Drill Bits

Tooth breakage on PDC drill bits is possible due to large impulsive forces caused by momentum
changes.
Consider a disc rotating at a constant angular velocity about its geometric center (Figure 7.44).
Over a very short time interval, Δt = t2 − t1, the edge of the disc is caught by a fixed pin as shown. At
time t2, the disc suddenly rotates about its instantaneous center, point “a.”
Over this short time interval (Δt), angular momentum with respect to point “a” is conserved,
there is no angular impulse. However, linear momentum with respect to the center of gravity is
not conserved. Based on the conservation of angular momentum:
G2 = G1 7 185
from which
I G + Mr 2 ω2 = I G ω1 7 186
and
IG 1
ω2 = ω1 = ω1 7 187
I G + Mr 2 3
assuming a solid disc. The velocity of the center of gravity is now v02 = rω2. Considering the com-
ponent of linear impulse to the right:
Imx = H x2 − H x1 7 188
ω1
ω2
Tool edge
r
v02
F(t)
a Well bore
Figure 7.44 Cutter impulsive force.

giving
2
F t dt = vo2 M − 0 = rω2 M 7 189

1
This change in linear momentum takes place over a very short period of time. The magnitude of
the maximum impulsive force can be approximated by
2
F t dt = F ave Δt 7 190
1
Dividing the impulse over time Δt gives the average impact force:
rω2 M
F ave = 7 191
Δt
Example Consider a rigid disc weighing 40 000 lb, having a 7-in. diameter and rotating at a speed
of 100 rpm. These numbers were chosen to simulate drill collars. Adjusting the units for substitu-
tion into Eq. (7.191)
35
r= = 0 2917 ft
12
40 000
M= = 1242 slugs
32 2
1 2π100
ω2 = = 3 49 rad s
3 60
Putting these numbers into Eq. (7.191) gives
0 2917 3 49 1242
F ave = = 79 496 lbs
0 05
Here, Δt is arbitrarily chosen to be 0.05 second. A smaller Δt increases the predicted average
force. The peak impact force on the pin would be much higher. Cutter damage and failure is pos-
sible from this level of force.
Accounting for Torsional Flexibility in Drill Collars

Now consider the torsion flexibility of drill collars of length, L (Figure 7.45). Assume the drill collars
rotate clockwise as in drilling. Once a cutter contacts an obstruction, creating impact force, F(t), a
G
shear wave starts to travel to the right with acoustic velocity c = 10 600 ft/s c = ρ = 10 600 fps .
Over a short span of time, the leading edge of the shear wave has traveled a distance, x = ct. The
shape of the impulse force, F(t), will be sustained within the collars as illustrated. The space interval
Δx is related to the time interval, Δt by Δx = cΔt. As the shear wave travels to the right, linear
momentum is changed according to
Δt
F t dt = rω2 mΔx m is mass per unit length 7 192

0
In terms of the average impulse force,

Figure 7.45 Torsion shear wave set up by cutter

impact.
Force c
Time
Δt
F
Δx
F ave Δt = rω2 mcΔt 7 193

F ave = rω2 mc 7 194
Assuming the drill collars weigh 114 lb/ft (m = 3.54 slugs/ft) and applying the previous numbers
to Eq. (7.194) gives
F ave = 0 291 3 49 3 54 10 600 = 37 980 lb
This number is substantially less than for the rigid disc because not all of the linear momentum of
the collars is changed simultaneously.
Interconnecting Bodies
Interconnecting bodies can also be analyzed as one unit using the equations of motion for inter-
connecting bodies explained earlier. In terms of impulse–momentum:
M A dt = GA 2 − GA 1 7 195
I Gi ω2 + di m i V 2 = M A dt + I Gi ω1 + di m i V 1 7 196
Example Consider the arrangement shown in Figure 7.46. Relevant numbers are given in the
figure.
For this example, Eq. (7.196) reduces to
I A ω2 + rmB V B2 = M A Δt + I A ω1 + rmB V B1
The moment, MA = 0.2(58.86) = 11.77 N - m
Using the numbers in the figure and noting that VB = 0.2ω the equation yields
ω2 = 65 17 rad s
ω1 = 10 rad/s Ay
ω2
0.2 m
A Ax
I = 0.4 kg-m2
196.2 N
B B
mB = 6 kg
58.86 N
Figure 7.46 Interconnecting system.
and
V B2 = 13 m s
when Δt = 3 seconds
This approach gives a direct solution without dividing the total system into two separate freebody
diagrams.
Conservation of Angular Momentum

Gear pairs can also be analyzed as interconnecting bodies using angular impulse-momentum
principles.
M 0 dt = G0 2 − G0 1 7 197
If angular impulse is zero (left side), then angular momentum is conserved.
G0 2 = G0 1 7 198
Example Consider the conservation of angular momentum during the activation a flywheel
(Figure 7.47). The gear train, clutch, and flywheel are considered as a group of interconnecting
bodies.
A simple gear train transmits power from an electric motor through a gear pair to location A.
Initial power from the motor is 3 hp delivered at 1000 rpm. Determine the torque and speed deliv-
ered to point A. If the clutch is engaged to active a flywheel, determine the output speed of the motor
immediately after the clutch is engaged assuming the two torques are unchanged. How much time
is required to bring the system up to its operating speed? Assume mass polar moments of inertia are
Gear 2 Reactive torque, TR
Armature, JA
Gear 1
Flywheel, JF
Tin Clutch
Figure 7.47 Angular momentum changes during flywheel engagement.
J A = 0 1 lb-s2 in armature
J 1 = 0 05 lb-s2 in gear 1
J 2 = 0 2 lb-s2 in gear 2
J F = 0 4 lb-s2 in flywheel
R1 = 2 in pitch radius of gear 1
R2 = 4 in pitch radius of gear 2

Determine:
a) Torque and speed at point A.

b) Rotational speeds after clutch has been engaged.
c) Time required to bring the system back up to its original speed
Part a
The torque delivered by the motor is
TM N 3 5252
HP = TM = = 15 756 ft-lb = 189 1 in -lb 7 199
5252 1000
The reactive torque is TR = 2Tin = 31.512 ft-lb at N2 = 500 rpm.
Part b
This part of the problem will be addressed using the equation of motion for the gear pair shown in
Figure 7.39. The dynamic behavior of the gear pair is described by:
2
r1 r1 dω
TM = T0 + I1 + I2
r2 r2 dt
In terms of impulse-momentum
2 2
r1 r1 r1
TM − T o Δt = J A + J 1 + J F + J 2 Ω0 − J A + J 1 + J2 ω
r2 r2 r2
The change in motion caused by the activation of the flywheel takes place instantaneously such
that impulse is zero. This means that angular momentum is conserved and
2 2
r1 r1
JA + J1 + JF + J 2 Ω0 = J A + J 1 + J2 ω 7 200
r2 r2
where Ω0 is new motor speed at time zero.

J A + J 1 + 0 25J 2
Ω0 = ω1 7 201
J A + J 1 + J F + 0 25J 2
By substituting the numbers
0 1 + 0 05 + 0 25∗0 2
Ω0 = ω1
0 1 + 0 05 + 0 4 + 0 25∗0 2
ω1
Ω0 = or 0 333 1000 = 333 rpm
3
The flywheel has the effect of lower the motor speed from 1000 to 333 rpm immediately after the
clutch is engaged.
Part c
At time zero (immediately after clutch engagement), assuming motor power is constant (30 hp),
motor torque is
HP 5252 3 5252
T 0 = = = 52 52 ft-lb input torque at t = 0 7 202
N 333
Please note motor performance curve will define the true torque–speed relationship. We now
apply Eq. (7.160) to determine time to reach original speed, ω1 with the flywheel engaged.
2
r1 r1 dΩ
TM = T0 + J1 + J2 7 203
r2 r2 dt
Assuming the motor power remains the same, i.e. 30 hp then by
TΩ
HP = = 3 hp 7 204
5252
Torque and speed are related by
TΩ = 15 756 T − ft-lb, Ω − rpm 7 205
TΩ = 19 800 T − in-lb, Ω − rad s

By substitution
1 dΩ
T t − T R = J A + J 1 + J F + 0 25J 2 7 206
2 dt
which reduces to
19 800 1 dΩ
− T R = J A + J 1 + J F + 0 25J 2 7 207
Ω 2 dt
This equation is nonlinear and best solved numerically. However, if we assume speed recovery is
made without the reactive torque the equation simplifies to a form that is easily solved analytically:
dΩ 19 800 19 800
Ω = = = 33 × 103 7 208
dt J A + J 1 + J F + 0 25J 2 06
1 dΩ2
= 33 × 103
2 dt
Ω2 = 33 × 103 2 t + C
When t = 0, Ω = Ω0, so C = Ω20 . Solving for t when Ω = ω1 gives
ω21 − Ω20 1 − 0 0123 2

t= = ω1 7 209
66 × 103 66 × 103
2π
ω1 = 1000 = 104 72 rad s
60
By substitution, t = 0.164 second. The recovery time is quite short for this set of numbers. How-
ever, for larger gears, time of recovery would be greater.
References
1 Whipple, S. (1883). An Elementary Treatise on Bridge Design, 4e. New York (USA): D. Van Nostrand.
2 Schuh, F. (1977). Drilling Equations. Petroleum Engineering Publishing Company.
3 Garrett, W.R. and Rollins, H.M. (1926). Steering wheel for rock bits, Presented at the 9th Annual
Petroleum Mechanical Engineering Conference, Los Angeles, Calif.
375
Mechanics of Materials
Romans arches were used to form bridges, aqueducts, and buildings, including the Coliseum.
Arches accommodate uniform loads quite well, but concentrated load, especially at the center
point, cause instability and possible collapse [1]. Instability was solved by building wall-type exten-
sion, called spandrels, outside of arches. The Romans also discovered how to make concrete to
solidify the spandrels. Horizontal support at the base of arches was also essential. Stone arches were
used throughout Europe until the early 1700s.
Iron became the new structural material replacing masonry and wood by the early 1700s. The
British iron maker, Abraham Darby (1678–1717), began to produce better iron by using coke, a
derivative of coal, instead of charcoal, a derivative of wood. His iron business (1709) was in Coal-
brookdale, England. In those days, there was a much greater supply of coal than timber, plus Darby
produced a higher quality of iron.
In 1777 his grandson Darby III was contracted to build a bridge spanning 100 ft across the Severn
River. The bridge, which was completed in 1779, was made completely of cast iron. Each member
was casts into desired shapes, thus eliminating costly hand work required to chisel and carve stone.
The cast iron shape was an arch which closely resembled the geometry of Roman arches, putting
the cast iron components in compression. Historically, this bridge marked the end of arch bridges
made of stone and timber.
Two factors began to change the design of engineering structures: (i) stronger steel and
(ii) scientific tools of analysis. Scientific reason began to replace the empirical approach. The first
science-based engineering structure is the Eads Bridge built over the Mississippi River at St. Louis.
This bridge was designed and constructed by James Eads between 1867 and 1874. It was the world’s
longest arch bridge and was made of steel. The bridge is still in use today.
Engineering design deals with loads, stresses, deflections, and failure of structural members. The
challenge in design is to configure member sizes and shapes that can withstand external loads with-
out failing. As a rule, cost relates to weight, so the lighter the structure, the lower the cost. Weight is
reduced through engineering analysis. Cost increases with member complexity and number of
components. Design is all about making equipment cost-effective, and this is done through achiev-
ing functionality through simplicity.
Equipment can contain simple components under direct tension/compression, shear, or twist. In
general beams, relative long members that carry transverse loads, can develop high stresses, and
require special analyses. In simple terms, it is easier to break a stick by bending than to pull it apart.
This section summarizes important aspects of stress analysis used in engineering design.
376 Mechanics of Materials
Stress Transformation
When a rod or strap is pulled (Figure 8.1), the average normal stress across the section is simply
σ = AF as determined during a standard tension test. Considering the forces on an inclined plane,
we see that a normal force (N) and a shear force (V) are required for equilibrium:
Fn = 0 Ft = 0
N − P cos θ = 0 V + P sin θ = 0
N = P cos θ V = − P sin θ
Corresponding stresses (normal and shear) on the inclined plane are
N P
σn = = 1 + cos 2θ normal stress 81
An 2A
V P
τn = = − sin 2θ shear stress 82
An 2A
The sign convention of each is set by the direction of the n,t axis with θ measured counterclock-
wise from the x-axis. The stress values in Figure 8.2 have been unitized, i.e.
σn 1
= 1 + cos 2θ normal stress
P A 2
t
N
V n
P θ
A, cross section x
θ θ
Figure 8.1 Stresses across a strap.
Normal
Shear
0.5
Stress, psi
0
–90 –45 0 45 90
–0.5
Plane orientation, degrees
Figure 8.2 Effect of plane orientation.

τn 1 t
= − sin 2θ shear stress
P A 2 n
τn σn
The equations show how normal and shear stresses
on any plane vary with θ. The maximum normal stress θ
σ x
occurs on the transverse plane (θ = 0). The maximum
shear stress occurs on planes inclined at 45 . The stress
condition in Figure 8.1 can also be viewed from a small
element as shown in Figure 8.3, where forces are in
Figure 8.3 Stress element.
equilibrium – not stress.
Theory of Stress σy
In general, local stress is defined by three stress components, σ x, σ y, τxy

τxy which are applied to this small element. This condition σx
(Figure 8.4) defines stress at a point even though surfaces are σx
required to yield forces, which must be in equilibrium. Stresses a
are shown in their “plus” direction.
From moments about point “a,” τxy = τyx, showing that shear stres- σy
ses on perpendicular planes have the same magnitude. These stress Figure 8.4 Stress components
components may be established analytically or experimentally. at a point (σ z = 0).
Normal and Shear Stress Transformations

Normal and shear stresses on inclined planes vary as before. Forces associated with these stresses
are in equilibrium. Equilibrium equations expressed in terms of the stress components are
given below:
Fn = 0 83
σ n = σ x cos 2 θ + σ y sin 2 θ + 2τxy sin θ cos θ
In terms of double angle:

σx + σy σx − σy
σn = + cos 2θ + τxy sin 2θ
2 2
84 t
Ft = 0 85 n
σn
σx − σy τnt
τnt = − sin 2θ + τxy cos 2θ 86
2 θ
σx x
The sign convention for each stress is plus (+) as
shown in Figure 8.5. τxy
To illustrate the use of Eqs. (8.4) and (8.6), consider the
stress condition shown in Figure 8.6.
The stress components are σy
σ x = 80 MPa Figure 8.5 Stress on an inclined surface.

σ y = − 100 MPa
σy = 100 MPa τxy = − 60 MPa
We wish to determine normal and shear stresses

on a plane whose normal axis is 42 counterclock-
σx = 80 σx = 80 MPa
wise from the x-axis (Figure 8.7).
Substituting these stress components into
Eqs. (8.4) and (8.6), gives
τxy = 60 MPa
σ n = − 60 26 MPa
σy = 100 τnt = − 95 78 MPa
Figure 8.6 Plane stress at a point (MPa). These stresses are shown on perpendicular sur-
faces in Figure 8.8.
The normal stress, σ n, and shear stress, τnt, both
vary with angle, θ. The variations of both
n (Figure 8.9) show orientations where normal stres-
ses are maximum and minimum. Likewise, shear
σn
stress reaches maximum values at certain inclina-
42º
80 tions, θ. Normal and shear stress components for
θ x the 42 orientation are identified by the
τnt heavy dots.
60
Maximum Normal and Maximum Shear
100 Stresses
Maximum normal and maximum shear stresses
Figure 8.7 Orientation of a biaxial state of stress.
and their orientation are of interest in design.
Material failure may depend on both, depending
on criteria of failure. The location and magnitude
of maximum stresses are determined by calculus.
Starting with the normal stress,
𝜎n = 60.3 MPa σn = + cos 2θ + τxy sin 2θ
2 2
87
+ 42° dσ n
=0 88
dθ
gives
𝜏nt = 95.8 MPa
2τxy
tan 2θp = 89
σx − σy
By substitution
– 48°
σx + σy σx − σy 2
σ p1,p2 = ± + τ2xy
𝜎n = 40.3 MPa 2 2
8 10
Figure 8.8 Stress components on an inclined
plane at 42 . Similarly, to determine maximum shear and
starting with
150
100
Shear
50
Stress, MPa
42
0
–180 –90 0 90 180
–50
Normal
–100
–150
Angle of inclined plane, degrees
Figure 8.9 Stress components vs. surface orientation.
σx − σy 8000 psi
τnt = − sin 2θ + τxy cos 2θ 8 11
2
dτnt 4000 psi
=0 8 12
dθ
10 000 psi
σx − σy
tan 2θτ = − 8 13
2τxy
σx − σy 2
τmax = ± + τ2xy 8 14
2
Observations:
Figure 8.10 Example stress components.
1) Shear stresses are zero on planes of max and min
normal stress (principal planes)
2) Principal planes are 45 from planes of max shear stress
3) The two principal planes are 90 apart; perpendicular to each other
Example Consider the stress condition given in Figure 8.10.

σ x = 10 000 psi
σ y = − 8000 psi
τxy = − 4000 psi
Substituting these numbers into Eq. (8.10) gives, σ p1, p2 = − 8849 and +10 850 psi (Figure 8.11).
2τxy
tan 2θp =
σx − σy
2 − 4000
tan 2θp = = − 0 444
10 000 − − 8000
Two values of θp satisfy this equation
15
10
5
Tangent, 2θ
0
–90 –45 0 45 90 135 180 225 270
–5
–10
–15
Angle, 2θ
Figure 8.11 Tangent function.
p2
t
9850 n
8849 psi
1000
–12º 33º
x
p1
10 850 psi
Figure 8.12 Maximum normal and shear stresses.
θp = − 12
and
θp = 78
To determine which principal stress applies to which plane, substitute, say θp = −12 into
σ n = σ x cos 2 θ + σ y sin 2 θ + 2τxy sin θ cos θ
gives σ p = 10 850 psi. Both principal stresses are shown on the element in Figure 8.12.
Now consider maximum shear stress:
σx − σy
tan 2θτ = −
2τxy
10 000 − − 8000
tan 2θτ = − = 2 25
2 − 4000
2θmax = 66
θmax = 33
τmax = − 9850 psi
σ n = 1000 psi
These stresses are also shown in Figure 8.12.
Mohr’s Stress Circle

In 1882, Otto Mohr, a German engineer, recognized that transformation of stresses at a point can be
represented by a special circle [2]. Mohr’s circle gives a graphical representation of the transforma-
tion equations and is easy to remember. Normal stresses are located on the horizontal axis
(Figure 8.13). Shear stresses are located on the vertical axis. Stress components at a point are located
on the circle as follows. σ x and σ y are marked on the horizontal axis. The average of these stresses is
the center of the circle. Shear stress is measured from these two normal stress points. The coordi-
nates, σ x, τxy determine the radius of the circle. In this figure, σ x σ y. When τxy is plus (moment is
CCW), it is plotted down in the Mohr diagram:
σx σy
τxy is plus and moment is CCW. It is plotted down in the Mohr diagram
Both principal stresses can be formulated from the diagram:
σn = + cos 2θ + τxy sin 2θ 8 15
2 2
σx + σy σx − σy 2
σ p1,p2 = ± + τ2xy 8 16
2 2
2τxy
tan 2θp = 8 17
σx − σy
τ
τmax
y axis
σp2 σp1
σ
2θ
σy τxy
x axis
σx
Figure 8.13 Mohr’s stress circle.

Maximum shear stress is formulated from the diagram:

σx − σy
2
σx − σy 2
τmax = ± + τ2xy 8 19
2
σx − σy
tan θτ = − 8 20
2τxy
Example The use of Mohr’s circle is illustrated with the following example. Assume the
components of stress at a point are σ x = + 8 ksi, σ y = − 6 ksi, and τxy = + 4 ksi (Figure 8.14).
The objective is to determine the principal normal stresses, maximum shear stress, σ p1, σ p2, τmax,
and the orientation of the elements on which they act.
The center of the circle is
σx + σy 8−6
= =1
2 2
The radius of the circle is
σx − σy 1
2 8+6
R= + τ2xy = + 42 = 8 06
2 2
The x axis is drawn between (1, 0) and (8, −4).
4
tan 2θp = 8 21
7
2θp = 29 74
θp = 14 87
From the circle, the principal normal stresses are

σ p1 = 9 06 ksi
σ p2 = − 7 06 ksi
τ 8.06
6 ksi y axis
9.06
8 ksi 29.74º
σ
–7.06 4
4 ksi
x axis
6 8
Figure 8.14 Solution by Mohr’s circle (ksi stress units).

n
p2
1 ksi
45 º
7.06 ksi
14.87° 8.06 p1
9.06 ksi
x
x axis
Figure 8.15 Plane of maximum shear stress.
Also, τmax = − 8.06 ksi and acts on a plane oriented 45 from the principal axis. Principal planes
and planes of maximum shear are shown in Figure 8.15.
Theory of Strain
The theory of strain is developed independently from the theory of stress. Both stress and strain are
related by Hooke’s law, which will be cover later.
There are two types of strain, normal and shear, and they both relate to the two types of stress,
normal and shear as previously discussed.
Normal strain is simple elongation (or compression) as shown in Figure 8.16. By definition
δ
ε= dimensionless 8 22
L
In general
du
ε= 8 23
dx
(a) (b)
δ
L
δ
Figure 8.16 Two types of strain. (a) Normal strain. (b) Shear strain.
ϕ
α
Because internal tension/compression may not be uniform.

The magnitude of normal strain is roughly
σ yld 50 × 103
ε= = = 1 67 × 10 − 3
E 30 × 106
Normal strain is commonly expressed in micro units,
ε = 1670 μ strain
ε = 1670 × 10 − 6
A normal strain of 2000 μ strain is a fairly large strain and a good reference number.
Shear strain is a measure of the amount of distortion as opposed to stretching:
δ
γ= = tan α ≈ α 8 24
L
Figure 8.16 shows shear strain is a measure of distortion indicated by angle, α. A more general
condition of shear distortion is shown in the side drawing. In this case
γ=α+β 8 25
Both angles add to define the total distortion. Shear strain is also expressed by angle, ϕ.
π
γ= −ϕ 8 26
2
The sign convention for shear strain is
π
ϕ≺ shear strain is plus
2
π
ϕ shear strain is negative
2
Strain Transformation
y
t Strain at a point includes both normal and shear strains
simultaneously. We will now transform these strains on
εt Υnt
an element oriented by θ from the x-axis (Figure 8.17).
Normal strains, referenced from the n,t reference
εn frame, can be expressed in terms of εx, εy, γ xy by
n
εn = εx cos 2 θ + εy sin 2 θ + γ xy sin θ cos θ 8 27

θ
x
In terms of double angle
Figure 8.17 Strain transformation.
εx + εy εx − εy γ xy
εn = + cos 2θ + sin 2θ 8 28
2 2 2
Maximum normal strains are determined by use of calculus.
εx + εy εx − εy 2 γ xy 2
εp1 , εp2 = ± + 8 29
2 2 2
γ xy
tan 2θp =
εx − εy
Shear strains, reference from the n,t coordinate reference frame, can be expressed in terms of εx,
εy, γ xy by
γ nt εx − εy γ xy
= − sin 2θ + cos 2θ 8 30
2 2 2
Maximum strain at a point is determined from calculus.
γ max εx − εy 2 γ xy 2
= ± + 8 31
2 2 2
εx − εy
tan 2θγ max = − 8 32
γ xy
Note the mathematical similarity between strain and stress transformations.
σ n = σ x cos 2 θ + σ y sin 2 θ + 2τxy sin θ cos θ 8 33

γ xy
εn = εx cos 2 θ + εy sin 2 θ + 2 sin θ cos θ 8 34
2
σx − σy
2
γ nt εx − εy γ xy
= − sin 2θ + cos 2θ 8 36
2 2 2
The transformation equations for stress and strain are mathematically the same except shear
strain is divided by 2. This means, analytical expressions for maximum normal and shear stresses
also apply to strain. Mohr’s circle also applies to shear strain except the shear strain axis is divided
by 2. The sign convention for both stress and stain are compared in Figure 8.18.
(a) (b)
y y cw – plot up
Plot up
+ Plot down +
ccw – plot down
x x
Figure 8.18 Sign convention for shear strain. (a) Stress. (b) Strain.
y t Example Consider
εt
εx = 800μ
εy = − 1000μ
γ xy = − 600μ
What are the normal strains and shear strain on an ele-

ment oriented −30 from the x-axis?
x
By substitution into Eqs. (8.34) and (8.36)
30° εn εn = 800 cos 2 − 30 − 1000 sin 2 − 30
γnt n − 600 sin − 30 cos − 30
Figure 8.19 Strain transformation. εn = 610μ

εt = 800 cos 2 60 − 1000 sin 2 60 − 600 sin 60 cos 60
εt = − 810μ
γ nt = − 800 + 1000 sin − 60 + − 600 cos − 60
γ nt = 1259μ
These strains are shown in Figure 8.19.
Mohr’s Strain Circle

The use of Mohr’s strain circle to determine principal strains and maximum shear strains is illus-
trated below. Local strain components are
εx = 1200μ
εy = − 600μ
γ xy = 900μ
Mohr’s circle for this example is shown in Figure 8.20.

Again, planes of maximum stress and strain are the same. No shear stress and no shear strain
(Figure 8.21).
450
tan 2θ = =05
900
2θ = 26 6 θ = 13 3
Principal Axes of Stress and Strain

From the theories of stress and strain, we know that there is no shear stress on the principal planes;
therefore, there is no shear strain as shear stress and shear strain are related through
τ = Gγ 8 37
This means that the principal planes of normal stress are the same as the principal planes of nor-
mal strain (Figure 8.22).
γ/2
1006.2
y axis
300 1306.2
ε
–700.6
26.56° 450
x axis
Figure 8.20 Mohr’s strain circle.
p2 εave = 300μ
p2 n
εp2 = –700.6μ
εp1 = 1306.2μ t 45°
p1 p1
x x
13.28° γmax = –2012.4μ
Figure 8.21 Planes of maximum normal and shear strains.
ɛy
σy
σx ɛx
Figure 8.22 Comparison of normal stress and strain.

Generalized Hooke’s Law
The theories of stress and strain were developed independent of each other. Both variables are
related through Hooke’s law. Laboratory tests show that when a load is applied to a test rod,
the rod stretches elastically. The relation between applied stress and resulting strain in a uniaxial
state of stress is expressed by
σ
ε= 8 38
E
The strain has the same direction as the stress.
Tests show that stress in one direction not only produces strain in the same direction but also a
shrinking normal strain in transverse directions. The amount of shrinking is related to the strain in
the direction of the load by
εT transverse
ν= − = − 8 39
εA axial
This ratio is called Poisson’s ratio and is a property of the material. It is determined from a simple
uniaxial state of stress. Equations (8.38) and (8.39) are basic to the generalized Hooke’s law for a
triaxial state of stress.
1
εx = σx − ν σy + σz 8 40
E
1
εy = σy − ν σx + σz 8 41
E
1
εz = σz − ν σy + σx 8 42
E
Be aware that in general, the ratio of transverse strains is not equal to Poisson’s ratio. The rules for
finding Poisson’s ratio must follow the procedure described above.
Hooke’s law includes the effects of direct stress plus Poisson’s effect due to stresses in the other
two transverse directions. Two common engineering situations will now be addressed: (i) plane
stress and (ii) plane strain.
Theory of Plain Stress

Under plane stress, σ z = 0; however εz 0 according to Figure 8.25a. An example is zero stress on a
free surface. Stresses lie in the plane of the surface, but stress normal to the surface is zero. Applying
Hooke’s law gives
1
εx = σ x − νσ y 8 43
E
1
εy = σ y − νσ x 8 44
E
ν
εz = − σx + σy 8 45
E
By substitution gives
ν
εz = − εx + εy 8 46
1−ν
Note that
εz ν εx + εy 8 47
Rearranging Eqs. (8.43) and (8.44) gives expressions for stress in terms of strain:
E
σx = εx + νεy 8 48
1 − ν2
E
σy = εy + νεx 8 49
1 − ν2
Shear stress is not involved in either equation but relates to shear strain by a separate equation:
τxy = Gγ xy 8 50
The modulus of elasticity (E) is determined experimentally from a simple tension test. The mod-
ulus of rigidity (G), however, is determined mathematically by
E
G= 8 51
2 1+ν
From this example, using E = 30 × 106 and ν = 0.25, with
E 30
G= = 106 = 12 × 106
2 1+ν 2 1 + 0 25
Orientation of Principal Stress and Strain

Consider the state of stress shown in Figure 8.23. The center diagram shows principal stresses. The
right diagram shows principal strain. The planes on orientation are the same, indicating zero shear
stress and zero shear strain.
•• Maximum normal stress is 8.71 ksi

Maximum normal strain is 330μ
•• Principal directions for both stress and strain are the same
Maximum shear stress is 6.71 ksi
• Maximum shear strain is 559μ
13.28°
8 ksi 330μ
4.71 ksi
4 ksi
3 ksi 230μ
8.71 ksi
Figure 8.23 Plane stress and corresponding strain.

Example Strain components at a point in a structure are given as

εx = 2000μ
εy = 1000μ
γ xy = 1000μ
Using Mohr’s strain circle (Figure 8.24) to determine the principal stains and principal stresses
assuming plane stress conditions.
The radius, R, of the circle is
1
R = 5002 + 5002 2
= 707 1
Principal strains are

ε1 = 1500 + 707 = 2207μ
ε2 = 1500 − 707 1 = 793μ
Shear strain is zero on these planes.
Principal stresses for plane stress are determined from
E
σ p1 = εp1 + νεp2
1 − ν2
30 × 106
σ p1 = 2207 + 0 28 × 793 10 − 6
1 − 0 282
σ p1 = 79 000 psi
30 × 106
σ p2 = 793 + 0 28 × 2207 10 − 6
1 − 0 282
σ p2 = 45 900 psi
Shear stress is zero on these planes.
γ/2
1000 μ
εp2 1500 μ εp1

ε
2θ
500 μ
x axis
2000 μ
Figure 8.24 Mohr’s strain circle.

(a) (b)
σy εy
σx εx
σz = 0, εz ≠ 0 εz = 0, σz ≠ 0
Figure 8.25 Elements of plain stress and plane strain. (a) Plane stress. (b) Plane strain.
Theory of Plain Strain

Plain strain exists when an element is not allowed to move in the z-direction (εz = 0), creating stress
in the z-direction (Figure 8.25b). When εz = 0, generalized Hooke’s law gives
E
σx = 1 − ν εx + νεy 8 52
1 + ν 1 − 2ν
E
σy = 1 − ν εy + νεx 8 53
1 + ν 1 − 2ν
E
σz = ν εx + εy 8 54
1 + ν 1 − 2ν
Still
τxy = Gγ xy 8 55
Pressure Vessel Strain Measurements

Analytical Predictions of Stress and Strain
One of the laboratory experiments at the University of Tennessee involved attaching strain rosettes
to a pressure vessel to gain experience in the use of strain gauges. The experiment allowed strain
measurements to be checked against theoretically calculated strains (and stresses) as predicted by
thin-wall vessel theory. The pressure vessel had a mean diameter of 5.585 in. and a wall thickness of
0.04 in. (~20-gauge sheet metal having thickness of 0.0375 in.). The sheet metal was rolled into a
cylinder and welded longitudinally. Spherical caps, of the same thickness, were welded to the cyl-
inder. The radius to thickness ratio was
r mean
= 69 81 10
t
so the pressure vessel was well within thin-wall theory range. The cylinder and end caps were
formed of steel having yield strength of 60 000 psi. Thin wall stress equations are
r
σθ = p cylinder 8 56
t
r
σ a = p cylinder 8 57
2t
r
σ s = p sphere 8 58
2t
Before pressuring the tank, it is important to know the pressure limit or the pressure which would
cause material yielding. Using the von Mises’ criteria of failure
σ ≥ σ yld yielding occurs 8 59
2
σ = σ 2A − σ A σ B + σ 2B 8 60
2
2 r 2 1 1
σ = p 12 − +
t 2 2
r
60 = p 0 866
t
p = 992 5 psi
Cylinder – Yielding occurs when p = 992.5 psi.
Sphere – Yielding occurs when p = 1,719 psi.
From this, we see that a pressure limit of 400 psi is well within the yield pressure limit. It is also a
good idea to get a feel for expected levels of strain before starting to gather strain data. Recall that in
a standard uniaxial stress–strain test the strain corresponding to a yield stress of 60 000 psi is
60 × 103
ε= = 0 002 in in or 2000μ micro strain
30 × 106
Therefore, we would expect strain data in the range of 1000μ or less.
Assume the vessel is subjected to 400 psi. From the thin wall equations:
σ θ = σ p1 = 27 924 psi
σ a = σ p2 = 13 962 psi
Assuming E = 30 × 106 and ν = 0.29, the expected principal strain levels are
1
εp1 = σ p1 − νσ p2 = 795 8μ
E
1
εp2 = σ p2 − νσ p1 = 195 5μ
E
These points are located on the Mohr’s strain circle shown in Figure 8.26. The diagram also pro-
jects the expected strains to be measured by the strain rosette.
γ
2
300
c axis
ɛp2 = 195.47
495.6
x axis ɛ
30° ɛp1 = 795.83
60°
a axis
b axis
Figure 8.26 Mohr’s circle.

c b
90°
a
θa
x axis
Figure 8.27 Strain rosette orientation.
Consider the strains measured by the strain rosette oriented by 15 as shown in Figure 8.27. The
three gauges are oriented as follows:
θa = 15
θb = 60
θc = 105
The expected normal strain values in each of the three individual gauges with vessel pressure of
400 psi are
εa = 495 6 − 300 cos 30 = 235 8μ
εb = 495 6 + 300 cos 60 = 645 6μ
εc = 495 6 + 300 cos 30 = 755 4μ
which are determined directly from Mohr’s strain circle.
Strain in the Spherical Cap

Since stress (and strain) in the spherical caps is the same in all directions, Mohr’s strain circle for at
any point in the end caps is simple a point. This also means that strain measurements by a three-
gauge rosette are the same in each of the three gauges. Also, there is no shear stress or strain in the
spherical cap.
According to Eq. (8.43), the strain in each rosette gauge should be
1 r
ε= 1−ν p 8 61
E 2t
For the example above, assuming ν = 0.29,
1 69 81
ε= 0 71 400 = 330μ
30 2
Conversion of Strain Measurements to Principal Strains and Stresses

Strain gauges are used to experimentally determine local strain at a point. A single gauge may be
used in a uniaxial loaded member. Only two gauges, 90 apart, are required to determine strain at a
point if aligned with known principal strain/stress directions. However, strain at a point usually
requires a strain rosette containing three gauges because principal strain directions are unknown.
The basic construction of a strain gauge is shown in Figure 8.28. The “Grid” is composed of a
metallic wire with a known electrical resistance. Manufactured strain gauges have a known relation
Solder
Grid
Lead wires
Figure 8.28 Strain gauge schematic.
Lead wires Figure 8.29 Strain gauge used as the fourth resistance
in a Wheatstone bridge.
R2
Eout
R3 R4
Ein
between their change in elongation and associated change in electrical resistance. This relation is
known as the “gauge factor,” or GF. The “lead wires” may be sent directly to an Ohmmeter to meas-
ure resistance; however, the lead wires are more commonly used in a Wheatstone bridge config-
uration. The “backing” typically contains an adhesive for locking the gauge to the material of
interest. Thus, as the material is strained, the wires in the grid are strained by the same amount.
Wheatstone bridge circuits are commonly used in conjunction with strain gauges because they
allow small changes in resistance to be measured accurately. This is important because elongation-
induced strains in the gauge result in very small resistance changes as compared to the gauge’s ini-
tial resistance. The Wheatstone bridge is shown in Figure 8.29. Resistance R1 (not shown) in the
bridge is the strain gauge. The output voltage, E0, may be measured to determine the change in
resistance of the strain gauge. When the output voltage equals 0, the bridge is said to be balanced,
meaning the resistances in opposite legs are equal, and there is no measurable strain [3].
Strain rosettes may come in different gauge arrangements (Figure 8.30).
They differ by the orientation of the gauges.
Rosette a θa = 0 θb = 45 θc = 90
Rosette b θa = 0 θb = 90 used when principal directions known
Rosette c θa = 0 θb = 60 θc = 120
Two things to note about strain gauges: (i) they do not measure shear strain and (ii) they do not
measure stress. They only measure normal strain in one direction. But knowing three normal
(a) (b) (c)
c b b c
b
a a a
Figure 8.30 Strain rosette arrangements. (a) Rosette a. (b) Rosette b. (c) Rosette c.
strains in three arbitrary directions allows us to calculate shear strain and principal stresses at
a point.
If normal strains are measured in three arbitrary directions, they can be used to find strains (εx, εy,
γ xy) associated with any xy coordinate system. From the theory of strain discussed earlier
εa = εx cos 2 θa + εy sin 2 θa + γ xy sin θa cos θa 8 62

εb = εx cos 2 θb + εy sin 2 θb + γ xy sin θb cos θb 8 63
εc = εx cos θc + εy sin θc + γ xy sin θc cos θc
2 2
8 64
Theta, θ, is the angular orientation of each strain gauge.

Strain measurements on a free surface are useful for defining strain at a given location. Once the
state of strain is determined, stresses are calculated from Hooke’s law.
In matrix form
εa εx
εb = A εy 8 65
εc γ xy
Strain referenced from the xy frame are determined from inverting Eq. (8.65).
−1
εx εa
εy = A εb 8 66
γ xy εc
Once εx, εy, γ xy have been determined, εp1, εp2, γ max can be determined analytically or by using
Mohr’s strain circle as discussed above. Principal stresses are determined from Hooke’s law, accord-
ing to the theory of plane stress as previously discussed.
Example Consider a biaxial state of plane stress (σ z = 0) is shown in Figure 8.31, along with a
strain rosette. Determine the strains (εa, εb, εc) that would be indicated in the strain rosette.
Given
E = 100 GPa
ν = 0 28
E
G= = 39 06 GPa
2 1+ν
36 MPa y
24 MPa c 90° b
72 MPa
45°
x
a
Figure 8.31 Strain rosette.
Approach: Determine (εx, εy, γ xy), then determine (εa, εb, εc).
1
εx = σ x − νσ y
E
1
εx = 72 − 0 28 36 = 619 2μ
100 × 103
εa = εx = + 619 2 μ strain
1
εy = σ y − νσ x = 158 4μ strain
E
1
εx = 36 − 0 28 72 = 158 4μ
100 × 103
τxy − 24
γ xy = = = − 614 4μ strain
G 39 06 103
Using
εn = εx cos 2 θ + εy sin 2 θ + γ xy sin θ cos θ

εb = + 81 6μ strain
εc = + 696μ strain
Beam Deflections
Shear and bending moment diagrams were discussed earlier along with shear and bending stresses.
This section revisits beam behavior by looking at beam displacements. The earlier discussion
showed that
EI
=M 8 67
ρ
Recall from calculus, local curvature of any function, y(x) is determined from
d2 y
1 dx 2
= 8 68
ρ dy 2
3
2
1+ dx
In applying this expression, we assume beam deflections are small so, dy

dx ≈ 0 . Applying this
assumption to Eq. (8.68) leaves
1 d2 y
8 69
ρ dx 2
In some cases, such as pipelines suspended in the ocean, this assumption is invalid and the total
expression for curvature must be used.
By substitution
d2 y
EI =M 8 70
dx 2
This equation is linear and amendable to mathematical solutions.
Cantilever Beam with Concentrated Force

Consider a cantilever beam with concentrated load at the end (Figure 8.32).
M x = − L−x F 8 71
d2 y F
= − L−x 8 72
dx 2 EI
By direct integration
dy F 1
= L − x 2 + C1
dx EI 2
F
yx = − L − x 2 + C1 x + C2
6EI
dy
Applying boundary conditions: = 0 and y(0) = 0.
dx x=0
FL2
C1 = −
2EI
FL3
C2 =
6EI
Combining everything
F 3
yx = − L−x + 3L2 x − L3
6EI
Figure 8.32 Deflection of y

cantilever beam. F
x
x EI
L
which reduces to
Fx 2
yx = − 3L − x 8 73
6EI
The maximum displacement and slope occur at x = L.
FL3
ymax = −
3EI
The slope along the beam is
F
θ x = − − x 2 + 2x 3L − x 8 74
6EI
with maximum slope at x = L.
FL2
θmax = −
2EI
Cantilevered Beam with Uniform Load

If the applied load is distributed as shown in Figure 8.33, then
d2 y 1 w 2
EI 2 = −w L−x L−x = − L−x 8 75
dx 2 2
The solution to this equation is determined by direct integration.

3
dy w L−x
EI = + C1
dx 2 3
4
w L−x
EIy x = − + C1 x + C2
2 12
Boundary conditions for this beam are
dy w L3
EI = 0 yielding C1 = −
dx x=0 2 3
y Figure 8.33 Deflection of a uniformly

loaded beam.
w
v0
x
m0 x EI
L
w L4
y 0 =0 yielding C2 = +
2 12
After applying the boundary conditions, the final solution becomes
1 w
yx = − L − x 4 − L4 + 4L3 x 8 76
24 EI
Expanding the first term by the binomial series and collecting terms gives
wx 2 2
yx = − x + 6L2 − 4Lx
24EI
The maximum deflection and slope occur at x = L.
1 wL4 wL3
ymax = − and θmax = −
8 EI 6EI
Simply Supported Beam with Distributed Load

The differential equation of bending can also be casts into another form as follows:
Since dV dM
dx = w and dx = V ,
d3 y dM
EI = =V using Eq 8 70 8 77
dx 3 dx
giving
d4 y dV
EI = =w 8 78
dx 4 dx
Applying this equation to the beam loading shown in Figure 8.34 gives
d4 y
EI = −w 8 79
dx 4
By direct integration
d3 y
EI = − wx + C1
dx 3
Figure 8.34 Simply supported beam with y

uniform load.
w
x
EI
x
L
d2 y wx 2
EI 2 = − + C1 x + C2
dx 2
dy wx 3 1
EI = − + x 2 C1 + C2 x + C3
dx 2 2
x4 x3 x2
EIy x = − w + C1 + C2 + C3 x + C4 8 80
24 6 2
Boundary conditions for this beam are
y 0 =0
d2 y
=0
dx 2 x=0
y L =0
d2 y
=0
dx 2 x=L
The first two boundary condition yield, C4 and C2 both equal zero. The second two boundary
conditions produce
L4 L3
0 = −w + C1 + C3 L
24 6
and
L2
0 = −w + C1 L
2
Solving both equations simultaneously gives
wL wL3
C1 = and C3 = −
2 24
Collecting terms gives
w 4 wL 3 wL3
EIy x = − x + x − x 8 81
24 12 24
wx
yx = − L3 − 2 Lx 2 + x 3
24 EI
The maximum displacement and slope are x = L/2.
5wL4
ymax = − occurs at x = L 2 8 82
384EI
wL3
θmax = occurs at x = 0 and x = L 8 83
24EI
Statically Indeterminate Beams

When a static analysis defines all forces, the procedure for determining shear and moment dia-
grams and forces are determined as explained above. In some cases, beam forces cannot be deter-
mined by a static analysis alone and may require the use of deflection equations.
Figure 8.35 Statically indeterminate beam. y

V0 w
x
0 EI
M0 x R
L
Consider the beam shown in Figure 8.35, with a fixed boundary at the left end and a simple sup-
port at the right end. Three unknowns are M0, V0, R. These external loads are needed before shear
and bending moment diagrams (as well as shear and bending stresses) can be established. Three
equations are needed. Two come from a force analysis.
Fy = 0 V 0 + R − wL = 0 8 84
2
L
M 0 = 0 LR − w + M0 = 0 8 85
2
The third comes from the deflection equation:
d2 y w 2
EI = L−x R− L−x 8 86
dx 2 2
By direct integration:
dy 1 w
EI = − L − x 2R + L−x 3
+ C1
dx 2 6
R w
EIy x = L−x 3 − L−x 4
+ C1 x + C2 8 87
6 24
Appling boundary conditions:
y L = 0 gives C 2 = − C1 L
RL3 wL4
y 0 = 0 gives C2 = − +
6 24
With these constants, Eq. (8.87) defines beam displacements in terms of the reaction force, R. The
force, R, is determined from a third boundary condition,
dy
=0
dx x=0
Giving
3
R= wL 8 88
8
Substituting R into Eqs. (8.84) and (8.85) gives
5 1 2
V0 = wL and M0 = wL 8 89
8 8
Multispanned Beam Columns

The effect of stabilizer location on magnitude and direction of drill bit side force can be calculated
by treating the stabilized section as a multispanned beam column. One method is outlined in
Ref. [4]. The goal is to produce the needed side force on drill bits to maintain or affect a desired
well bore direction. This example is based on the following input data:
Hole size – 8½ in.
Drill collar size – 63/4 × 2 13/16 in.
Drilling mud weight – 12 ppg
Inclination of the well bore – 30
WOB – 40 000 lb
Two stabilizers used, one at 20 ft the other at 60 ft from the drill bit. Two other support points are
at the drill bit and the contact point at the tangent point at the upper end, which is to be determined.
The output gives an accurate prediction of drill collar deflection from the drill bit to point of
contact with the wall. Collar deflections between drill bit and the first stabilizer are shown in
Figures 8.36 and 8.37.
0.1
First stabilizer Second stabilizer
0
0 10 20 30 40 50 60 70 80 90 100
Drill collar displacement, in.
–0.1
–0.2
–0.3
–0.4
–0.5
–0.6
–0.7
–0.8
–0.9
Well bore contact
–1
Distance from drill bit, ft
Figure 8.36 Deflections over a drill collar span.
0.02
0.015
Drill collar displacement, in.
0.01
0.005
Bit First stabilizer
0
0 5 10 15 20 25
– 0.005
– 0.01
– 0.015
– 0.02
– 0.025
– 0.03
– 0.035
Distance from drill bit, ft
Figure 8.37 Drill collar deflections between drill bit and first stabilizer.
Table 8.1 Transverse side forces.
Location Transverse force (lb)
Drill bit 181

First stabilizer 1088
Second stabilizer 1481
Wall contact 471a
a
Shear is actually minus at the upper tangent location.
Predicted side forces developed at drill bit, the stabilizers and wall contact are listed in Table 8.1.
Since the side force at the drill bit is only 181 lb while weight on bit (WOB) is 40 000 lb, the sta-
bilizer arrangement is considered a holding assembly.
Large Angle Bending in Terms of Polar Coordinates

In some cases, bending is deflected by circular guides, such as pipe in a curved bore hole. It is con-
venient in this case to set-up equations of bending in polar coordinates. This avoids the complexity
of nonlinearities associated with rectangular coordinates.
The differential element of length ds = R dθ is used to establish the necessary bending equations
in terms of polar coordinates (Figure 8.38).
Summing forces in the “n” direction gives
dV T
=q+ 8 90
ds R
Summing forces in the “t” direction gives
dT V
= − 8 91
ds R
Summing moments about the center of curvature of the well bore gives
dM d Tw dT
− +R =0 8 92
ds ds ds
By substitution
dM dw
−V −T =0 8 93
ds ds
d2 M dV d dw
2 = + T
ds ds ds ds
Well bore
R centerline
dθ
θ
T + dT
M + dM
V + dV
V
w(s) q
T
M
Figure 8.38 Equations of bending in terms of polar coordinates.
d2 M T d dw
2 = q + R + ds T ds
ds
The curvature of the displaced beam is expressed in terms of beam deflection (see Timoshenko
and Woinowsky-Krieger [5], pp. 503 and 504)
1 1 w d2 w
= + 2 + 8 94
ρ R R ds2
Assuming Euler bending
1
M = EI 8 95
ρ
and combining the above equations gives
T d dw EI d2 w d4 w
q+ + T = 2 2 + EI 8 96
R ds ds R ds ds4
d4 w EI d2 w d dw T
EI 4 + 2 2 − T =q+ 8 97
ds R ds ds ds R
In general, T can vary with s; however, in many cases, T can be assumed constant over stepwise
pipe intervals. In which case, Eq. (8.97) simplifies to
d4 w 2
2d w
− λ =ζ 8 98
ds4 ds2
where
T 1
λ2 = − 2
EI R
Rq + T
ζ=
REI
In this discussion q, the distributed load is assumed. The solution to the differential equation is
1
w s = − χs2 + As + B + C cosh λs + D sinh λs 8 99
2
Letting
T
χ=
EI
RT −
R
Once w(s) has been determined, shear and bending moment can be established.
Shear force along the pipe
dM dw
V= −T 8 100
ds ds
Bending moment along the pipe is
1 w d2 w
M = EI + 2 + 8 101
R R ds2
In this case, λ2 is assumed to be plus. In some cases, this coefficient could be minus making the
bending equation:
d4 w 2
2d w q
4 +λ = 8 102
ds ds2 EI
For example, T could be a compressive force. The second term now has a (+) sign and this leads to
a complimentary solution with sines and cosines.
Bending Stresses in Drill Pipe Between Tool Joints

Bending caused by local dog legs can be significant. Consider the schematic (Figure 8.39) showing
tool joints lying flat against the inside of a curved well bore. A drill pipe section of length, L, is
assumed to be under uniform tension, T. This condition creates high localized bending at each tool
joint as explained below. In this case, pipe deflection, which yields bending moments and stress are
best determined by use of polar coordinates.
Consider a joint of drill pipe bent within a dog leg. The magnitude of bending stresses depends
both on local hole curvature and tension in the drill pipe. When curved pipe is pulled, the middle
portion tends to straighten out and shift more curvature toward tool joints. Bending stresses are
greatest near tool joints. Total stress is the sum of stress caused by direct pull and bending stress.
The analysis starts with
d4 w 2
2d w
− λ =ζ 8 103
ds4 ds2
where
T 1
λ2 = −
EI R2
Rq + T
ζ=
REI
θ T
Tool joint
s
Drill pipe
w(s)
Center line of well

T
Figure 8.39 Bending of a pipe section between tool joints.
R – radius of centerline
Furthermore, let
Rq + T
χ=
RT − EI
R
In applying Eq. (8.103), we assume side-loading, q, is negligible.

1
2
Letting
T
χ=
RT − EI
R
T 1
λ2 = − 2
EI R
and applying the following boundary conditions:
w 0 = w0 w ℓ = w0
dw dw
=0 =0
ds s=0 ds s=ℓ
establish the four constants A, B, C, and D.
1) B + C = 0
2) Al + B + C cosh λl + D sinh λl = 12 χl2
3) A + Dλ = 0
4) A + Cλ sinh λl + Dλ cosh λl = χl
Combining (1)–(3) gives
1 1
Al 1 − sinh λl − C 1 − cosh λl = χl2 8 105
λl 2
Combining (3) and (4) gives
A 1 − cosh λl + Cλ sinh λl = λl
or
Al 1 − cosh λl + Cλl sinh λl = λl2 8 106
Equations (8.105) and (8.106) are two algebraic equation with A and C as unknowns:
a11 a12 A b1
=
a21 a22 C b2
where
a11 = l 1 − λl1 sinh λl

a12 = cosh λl − 1
a21 = l(1 − cosh λl)
a22 = λl sinh λl
Using Cramer’s rule with l = 30 ft, λ = 0.0153 and λl = 5.502
A = 0 015
C = 0 9895
1
D = − A = − 0 9815
λ
B = − C = 0 9895
Displacement along the pipe.
1
w s = − χs2 + As + B + C cosh λs + D sinh λs with χ = 8 334 × 10 − 5
2
From
1 w d2 w
M = EI + 2 +
R R ds2
Bending moment at s = 0 is
1 w0
M max = EI + 2 −χ + Cλ2 EI 8 107
R R
20 000
15 000
Bending stress, psi
10 000
5000
0
0 5 10 15 20 25 30
Distance along pipe, ft
Figure 8.40 Bending stress along a section of 5 in. (16.25 lb/ft) drill pipe (R = 1000 ft, T = 100 000 lb).
Example Parameters for this example are
λ = 0.0153
χ = 8.334 × 10−5
C = 0.9895
R = 1000 ft = 12 000 in.
I = 13 in.4
E = 30 × 106 psi
C = 2.5 in.
By substitution and dropping wo/R2
M max c
σ max =
I
1
σ max = Ec − χ + Cλ2
R
1
σ max = 30 × 106 2 5 − 8 334 × 10 − 5 + 0 9895 0 0153 2
12 000
σ max = 17 372 psi
These equations were programed to show how bending stress varies over a 30 ft pipe joint when
the radius of hole curvature is R = 1000 ft. Tension is assumed to be 100 000 lb. The highest bending
occurs at the tool joints 17 372 psi (see Figure 8.40).
Application to Pipe Bending in Curved Well Bores

A major challenge is analysis of bending and friction on drill strings in doglegs. Several para-
meters factor into the complexity of this problem: location of points of contact, magnitude contact
forces and resulting friction, shape of the dogleg, tension, and flexural stiffness of pipe to mention
a few.
RB
θ R
RC
RA
Figure 8.41 Model of pipe with three points of contact.
Consider a pipe making contact at three points as shown in Figure 8.41. This problem is
similar to a straight beam supported at three points. It is statically indeterminate and requires
the solution to beam deflection as well as statics. The same is true of the model in the figure.
In general, the center contact point will not be in the center between points A and B
requiring solutions to sections on both sides of point C with matching boundary conditions
at C.
For the sake of simplicity, we assume point C is midway between A and B, to take advantage
of symmetry. The problem in this case becomes a cantilever beam problem, using polar
coordinates.
The bending equations for the CB section are given as follows:
1
2
The boundary conditions that apply are
w 0 = δ w ℓ = −δ
dw 1 d2 w w 1
=0 = 0, which leads to + = −
ds s=0 ρ s=ℓ ds2 s=ℓ R s=ℓ R
where ℓ is arc length between points C and B. Distance, δ, is inward (+) radial displacement of point
C and outward (−) radial displacement of point B.
The boundary conditions quantify the constants, A, B, C, D. The moment at point C is determined
by Eq. (8.101). Forces V and RB are determined by statics. Reaction RC is equal to 2 V.
Multispanned Beam in Terms or Polar Coordinates

Another approach for solving the multispanned beam problem is explained by Huang et al. [6].
Results are given for strait beams, but the method applies equally well to polar coordinates.
A stepwise approach was used by Dareing and Ahlers [7] and Rocheleau and Dareing [8]
predict drill pipe deflection between tool joints as well as contact forces and friction. The model
in both papers assume contact is made by each tool joint within a curved portion of the well
bore. The solutions give insight as to how bending stiffness enters the friction calculations.
The basic equations for (i) pulling out of the well bore and (ii) going into the well bore are
explained below.
In both papers, pipe is analysis by considering deflections of segments between tool joints. It is
also assumed that tool joints contact the well bore (Figure 8.42).
Pulling Out of the Well Bore

The differential equation of bending for each pipe section is
d4 w 2
2d w
4 −λ =ζ 8 109
ds ds2
where
T 1
λ2 = −
EI R2
Rq + T
ζ=
REI
θn
Pipe deflection wn
ℓ2
Centerline of well bore
ℓ1
w3
w2
w1
Figure 8.42 Multiple spanned pipe in a dogleg.

Table 8.2 Pull out force over a 90 turn; R = 200 ft (4½ in. 16.5 lb/ft drill pipe).
Top tension
Back pull (lb) Free-free (lb) Fixed-fixed (lb) No stiffness (lb)
2 000 6 162 5 544 4 234

4 000 9 421 8 773 7 438
6 000 12 794 12 361 10 642
8 000 16 164 15 849 13 846
10 000 19 614 19 337 17 050
12 000 23 117 22 826 20 254
The general solution to Eq. (8.109) is

1
w s = − χs2 + As + B + C cosh λs + D sinh λs
2
Following the above procedure given in Ref. [8], pullout forces corresponding to six (6) different
pull back forces were calculated, and the results given in Table 8.2.
Columns and Compression Members
Column Buckling Under Uniform Compression

Column is a structure member which carries compressive loads. It can be a member in a truss or an
independent vertical support. Leonhard Euler (1707–1783), a Swiss mathematician, was the first to
study the buckling of columns. He derived the critical buckling equation,
C π 2
P= 8 110
4 ℓ
stating
2
Therefore, unless the load P to be borne be greater than C4 πℓ , there will be absolutely no fear
of bending; on the other hand, it the weight P be greater, the column will be unable to resist
bending. Now when the elasticity of the column and likewise its thickness remain the same,
the weight P which it can carry without danger will be inversely proportional to the square of
the height of the column; and a column twice as high will be able to bear only one-fourth of
the load [9].
The constant, C, represents the elastic property and cross-section dimensions, which trans-
late into
π 2
Pcr = EI 8 111
L
Euler called attention to the difference between compression failure and failure by buckling.
Buckling is a condition, under which a critical force, Qcr, is reached causing the column to
become unstable, i.e. the column takes a different configuration. Columns are usually assumed
to have uniform compression throughout. End supports, or boundary conditions may vary as
shown in Figure 8.43.
Consider a column simply supported as shown in Figure 8.44. The differential equation of bend-
ing from the freebody diagram of the lower end is
d2 y
EI + Qy = 0 8 112
dx 2
d2 y
+ β2 y = 0 8 113
dx 2
where
Q
β2 =
EI
Here, Q, is a compressive load. The solution to Eq. (8.112) is
y = A sin βx + B cos βx 8 114
For simply supported type boundary conditions, y(0) = 0 giving B = 0.
y = A sin βx 8 115
Applying y(L) = 0 gives
0 = A sin βL 8 116
Q Q Q Q
L′
L′
L
Pinned Fixed Free-fixed Pinned-fixed

L′ = 0.7 L L′ = ½ L L′ = 2 L
Figure 8.43 Columns with different end constraints.

x For a nontrivial solution

Q βL = n π, n = 1, 2, 3 8 117
For the first buckling mode, n = 1 and
Q π 2
= 8 118
EI L
giving
y π 2
Qcr = EI 8 119
L L
Q In terms of compressive stress, Eq. (8.119) becomes
π2
σ cr = E 8 120
M L r 2
where L/r is the slenderness ratio of the column cross
section and r is the radius of gyration of the smallest
y moment of inertia.
Figure 8.45 shows how critical buckling stress increases
Q Q with reduction in slenderness ratio. However, for short
Figure 8.44 Simply supported columns. compressive members, material yielding will occur prior
to buckling, as indicated by the horizontal line (assuming
σ yld = 60 ksi; L/r = 70.2). This condition establishes a
limit to Euler’s buckling equation.
Testing of columns show failure follows along a curved line between compression failure and
Euler buckling (Figure 8.46). The transition between these two failure modes is defined by three
slenderness ratio ranges:
•• Compression block range

Intermediate range
• Slender range
140
120
Critical stress, ksi
100
80
60 Yield strength, 60 ksi
40
20
0
0 50 100 150 200 250 300
Slenderness ratio, L/r
Figure 8.45 Euler buckling curve.

Compression Intermediate
block range range
100
80
Critical stress, ksi
60
40
Euler buckling
20
0
0 50 100 150 200 250
Slenderness ratio, L/r
Figure 8.46 Failure theory ranges.
As a guideline, the compression block theory covers L/r < 40. Euler buckling theory applies to
L/r > 140. Failure over the intermediate range is defined by empirical formulas, depending on
material type (steel, aluminum, wood). These empirical formulas are defined in column
codes [10–12].
The above equations apply to simple supports. The buckling equations also apply for other
boundary conditions if L is replaced by L (see Figure 8.43).
Example Consider a section of heavy pipe (6 × 2 , 85 lb/ft) supported vertically between pinned
connections separated by 60 ft. The minimum yield strength for the pipe is σ yld = 110 ksi.
π 4
I= R − R4i = 62 83 in 4
4 0
A = π R20 − R2i = 25 13 in 2
r = 1 58 in radius of gyration
L r = 455 69 slenderness ratio
Using Eq. (9.120)
π 2
σ cr = 30 × 106 = 1426 psi
455 69
This stress-load level is well within the yield limits of the collar material so Euler buckling theory
would apply. The corresponding collar force is Fcr = 1426(25.13) = 35 832 lb. This force level is
within typical drill bit force levels, so buckling is possible.
Next consider a larger pipe size (8 × 2 , 160 lb/ft; σ yld = 100 ksi).
I = 200 27 in 4
A = 47 12 in 2
r = 2 06 in radius of gyration
L r = 349 5 slenderness ratio
Using Eq. (9.146)

π2
σ cr = 30 × 106 = 2424 psi
349 5 2
This stress-load level is well within the yield limits of the collar material so Euler buckling theory
would apply. The corresponding collar force is Fcr = 2424(25.13) = 60 915 lb.
Columns of Variable Cross Section

Columns of antiquity indicate that designers had an empirical understanding of the effect of cross-
section properties of column stability. Stone columns of the Greco–Roman era show columns hav-
ing a more robust cross section over the center portion.
This section develops the critical force for columns having a variable cross section. The cross-
sectional inertia is assumed to vary linearly from each end to a maximum value in the center of
twice the end values:
2x
I = I0 1 + 8 121
L
Following the Rayleigh method of approximation, we assume a mode shape of
πx
y x = y0 sin 8 122
L
By the principle of minimum potential energy
V =U+Ώ 8 123
where
L
1 2
U= EI y dx 8 124
2
0
1 2
Ω = −P y dx 8 125
2
0
d2 V
For stability, dy20
≥0
L
2
EI y dx
0
Pcrit = L
8 126
2
y dx
0
By substitution of the assumed deflection function, y(x) and the expression for the moment of
inertia I(x), is [13]
1 70 π 2 EI 0
Pcrit = 8 127
L2
See Timoshenko and Gere [14] for a comprehensive discussion of elastic stability.
Tubular Buckling Due to Internal Pressure

An interesting and important aspect of column buckling is found in the mechanics of long pipe,
such as drill pipe, marine risers, and production tubing, used in the oil industry. The bending of
these pipes is greatly affected by fluid pressure inside and out. This phenomenon is discussed in
the following sections.
Consider the buckling of pipe with internal pressure acting alone as illustrated in Figure 8.47. The
surface pressure force is shown in the left drawing. The right element represents the statically
equivalent force system which is obtained by adding and subtracting pressure forces over the inside
area across an imaginary plane. The net effect of the pressure within an enclosed volume is zero,
leaving the end force vectors as shown. The equivalent force system is easily incorporated into the
bending equation.
This is a special case of Euler’s equation. The solution, using Eq. (8.119) and noting that
Q = piAi is
π 2
pi cr Ai = EI 8 128
L
where
pi − pressure inside pipe, psi

Ai − inside area of pipe, in.2
The pipe is assumed to be rigidly and simply supported and not capped. The pipe in this case
cannot be a compression block, but burst pressure limits may have to be considered.
Consider a section pipe (4 × 3.5 ) 100 ft long simply supported and axially constrained at each
end. Applying Eq. (8.128) gives
π 2 EI
pcr = 8 129
L Ai
π 2 30 × 106 5 20
pcr = = 111 13 psi
1200 9 621
This pressure is equivalent to a buckling force of
F cr = 111 3 9 621 = 1069 lb
pAi
p
pAi
Figure 8.47 Statically equivalent constant pressure effects.

Effective Tension in Pipe

Pressure applied to the outside surface acts to stabilize pipe. Figure 8.48 shows the statically equiv-
alent force system (right drawing) accounts for external pressure, internal pressure, and internal
tension in the walls of the pipe.
The net effect of both internal and external pressures combined with an axial force is an effective
tension defined by
T eff = T − pi Ai − po Ao 8 130
where
T is the actual internal force (plus is tension and a minus T is compressive force)
Ai = π4 d2i
A0 = π4 d20
This term is called effective tension because it occupies the space in the differential equation of
bending reserved for tension. It varies with vertical location in regard to drill pipe.
A common application of effective tension is in the bending vertical pipe in a hydrostatic situ-
ation. This term shows that axial tension and external pressure tend to stabilize while internal pres-
sure encourages buckling.
In this case, the inside and outside hydrostatic pressures are equal, but vary with depth. The pres-
sure terms can be replaced by
p0 A0 = zγ m A0
pi Ai = zγ m Ai
giving
p0 A0 − pi Ai = zγ m Ac 8 131
where Ac is the cross-sectional area of the pipe.
Noting that the volume displaced by a unit length of pipe is
ΔL Ac γ m = ΔLwm 8 132
T p0A0
piAi
p0
pi
T
T
Figure 8.48 Statically equivalent internal/external pressure system.

The pressure can now be written as
p0 A0 − pi Ai = zwm 8 133
where wm is the weight of drill fluid displaced by a unit length of pipe. Effective tension, in the
equation of drill pipe where drilling mud is the same inside and out becomes
T eff = T + wm z 8 134
where z is measure from the top downward. If the pipe is hanging freely, then
T eff top
=H hook load
At the lower end, with the pipe open,
T eff bottom
= T + Lwm = − Lwm + Lwm = 0 8 135
If the pipe is not open at the lower end but is subjected to an internal force of FB (tension), then
effective tension at location L is
T eff = F B + Lwm 8 136
Effective compression is also a useful parameter. It is defined as
Qeff = Q + pi Ai − p0 A0
where plus numbers mean compression. For the case of drilling mud inside and outside the pipe
Qeff = Q − L − x wm
where x is measured from the bottom upward. Q is the actual internal compression in the pipe.
Buckling of Drill Collars

Example Consider the buckling of drill collars between a drill bit and the first stabilizer as
depicted in Figure 8.49. Dimensions of the drill collars are
OD = 6 ½ in
ID = 2 in
w = 102 lb ft
L = 90 ft
We wish to determine the WOB that will buckle the 90 ft section under the following two
conditions:
a) Air
b) 11 ppg drilling mud (BF = 0.832)
Boundary conditions are assumed to be pinned at both drill bit and stabilizer locations. The actual
boundary condition at the stabilizer is somewhere between fixed and pinned so the solution will be
an approximation. Also, axial compression in the collars between these two locations varies line-
arly. We simplify this by using the average compression between the two locations [15].
x
Q2 wℓBF
Stabilizer
EI
ℓ
Q1
Qave
Qeff
Drill bit
WOB
WOB
Figure 8.49 Buckling of drill collars between drill bit and stabilizer.
The effective compression at any location, x, in the drill collars is

Qeff = Q − wx − wx − L − x wm
But the actual force at the drill bit is

Q = WOB + Lwm
By substitution, the effective compression at any location in the drill collars is
Qeff = WOB + xBFw
Buckling will occur when

π 2
Qcr = EI
L
π
I= 3 254 − 14 = 86 84 in 4
4
For the first case (a), buckling in air
1
Qeff ave
= WOB − wL
2
1 π 2
WOB − wL = EI
2 cr L
π 2 1
WOB cr = EI + wL
L 2
π 2 1
WOB cr = 30 × 106 86 84 + 102 90 = 26 634 lb
90 × 12 2
For the second case (b), buckling in drilling fluid

π 2 1
WOB cr = EI + wL BF
L 2
WOB cr = 22 044 + 3818
WOB cr = 25 862 lb
By comparison, buckling of a slick assembly (no stabilizers) based on Lubinski’s solution [16] is
1
2 EI 3
WOB cr = 1 94 wBF 3
144
1
2 30 × 106 86 84 3
WOB cr = 1 94 102 × 0 832 3 = 9835 lb
144
which is much lower than the stabilized assembly.
Combined Effects of Axial Force and Internal/External Pressure

Using the previous notation for effective compressive force when the axial force is compressive,
Qeff = Q + pi Ai − po Ao effective compression 8 137
where +Q is compression. By substitution
d2 y
EI + Qeff y = 0 8 138
dx 2
d2 y
+ β2 y = 0 8 139
dx 2
where
Qeff
β2 =
EI
and following the earlier solution, buckling occurs when
π 2
Q + pi Ai − po Ao cr ≥ EI 8 140
L
where the left term is effective compression. In this case, all loads are uniform with x. Note that
external pressure stabilizes the beam while internal pressure destabilizes the beam.
Buckling of Drill Pipe
Lateral buckling consideration starts with the differential equation of bending, which contains the
effective tension term:
T eff = F B + wx + L − x wm 8 141
For this case, we assume that the bottom force, FB, is composed of two terms
F B = − F − wm L note both terms are compressive 8 142
where
x
wmL − fluid force caused by local hydrostatic pressure;
recall wmL = γ mLAc
F − internal force beyond local hydrostatic pressure (stress)
The reason for this is how the lower end of drill pipe is axi-
ally loaded at its point of connection with drill collars. If the
neutral point is located at the top of the drill collars, then the
internal stress in the drill pipe at this conjuncture is hydro-
static or wmL. Length L is the vertical distance from the rig L
floor to the top of the drill collars. Under hydrostatic open-
ended loading, drill pipe will not buckle. The analysis seeks
to determine the additional force, F, beyond hydrostatic that
buckles drill pipe connected to the drill collars.
The effective tension term becomes
T eff = − F + w − wm x 8 143
y
The mathematical model for this analysis is shown in
F+Lwm
Figure 8.50.
d4 y d dy Figure 8.50 Mathematical model of

EI − w − wm x − F =0 8 144 drill pipe.
dx 4 dx dx
Boundary condition imposed on the solution are
y 0 =0
dy
0 =0
dx
y L =0
d2 y
L =0
dx 2
In dimensionless form, Eq. (8.144) is written as
d4 y d2 y d2 y dy
4 − αζ 2 +β −α =0 8 145
dζ dζ dζ 2 dζ
where
w − wm L3
α=
EI
FL2
β=
EI
Following the derivation in Ref. [17] critical values of β, which contain buckling force, F, are
given in Table 8.3 for the first four modes of buckling. The critical buckling force for the first mode
defines the on-set of buckling.
Values of α are limited to 4000 because of the high degree of numerical precision required to exe-
cute the calculations. The relation between α and βcr for higher values of α can be determined as
follows.
Table 8.3 Buckling parameters.
α βcr1 βcr2 βcr3 βcr4
0 20.19 59.68 118.90 197.86

1 20.84 60.21 119.42 198.37
2 21.50 60.73 119.94 198.88
5 23.44 62.31 121.50 200.40
10 26.66 64.94 124.12 202.95
20 32.99 70.22 129.36 208.07
50 51.06 86.18 145.31 223.51
100 77.92 113.25 172.67 249.60
200 121.61 166.59 230.50 303.21
500 218.06 297.87 403.85 474.41
1000 339.59 465.36 634.66 740.10
2000 529.83 729.55 997.20 1162.37
4000 827.87 1146.48 1569.43 1830.50
Table 8.4 Relation between ϕ values and α.
α ϕ1 ϕ2 ϕ3 ϕ4
1 20.89 60.21 119.42 198.37

2 13.54 38.26 75.56 125.82
5 8.02 21.31 41.55 68.54
10 5.74 13.99 26.74 43.73
20 4.48 9.53 17.56 28.24
50 3.76 6.35 10.71 16.47
100 3.62 5.26 8.02 11.59
200 3.56 4.87 6.74 8.87
500 3.46 4.73 6.41 7.53
1000 3.40 4.65 6.35 7.40
2000 3.34 4.60 6.28 7.32
4000 3.29 4.55 6.23 7.26
When the values of βcr vs. α (Table 8.4) are plotted on log–log scale, they become straight and
parallel with a slope of about 2/3. A straight line on log–log graph can be expressed as
2
ln βcr = ln α + ln ϕ 8 146
3
which simplifies to
2
βcr = ϕα3 8 147
giving
βcr F cr
ϕcr = 2 = 1 2 8 148
α3 EI 3 w − wm 3
Pipe length, L, does not show up in this equation. The buckling force becomes independent of
pipe length as would be expected for long pipe. The same conclusion applies to other boundary
conditions at x = 0. According to Table 8.4, ϕ1 is 3.29 for the first mode; ϕn for higher buckling
modes are also given in this table.
The critical buckling force equation for long pipe can now be predicted by
1
EI 3 2
F cr = ϕ w − wm 3 8 149
144
where Fcr is a force over and above local hydrostatic compression at x = 0. Also the section modulus
is assumed to be in units of lb-in.2 The critical buckling forces for the first four modes can be deter-
mined from
1
EI 3 2
F cr1 = 3 29 w − wm 3 8 150a
144
1
EI 3 2
F cr2 = 4 55 w − wm 3 8 150b
144
1
EI 3 2
F cr3 = 6 23 w − wm 3 8 150c
144
1
EI 3 2
F cr4 = 7 26 w − wm 3 8 150d
144
Example Consider 5½ in. (19.2 lb/ft) drill pipe along with the following parameters:
E = 30 10 6 psi
I = 16 8045 in 4
A = 4 9624 in 2
γ m = 10 ppg
γ stl = 490 lb ft3
L = 1000 ft
Applying these numbers to
w − wm L3 BFwL3
α= = 8 151
EI EI
gives
0 847 19 2 10003 144 in 2
α= = 4645 2
30 10 6 16 8045 1 ft2
Since α is greater than 4000, critical buckling parameter βcr is determined by Eq. (8.147), where
ϕ1, ϕ2, ϕ3, ϕ4 are 3.29, 4.55, 6.23, and 7.26, respectively. The critical buckling parameter for the first
mode is
2
βcr 1 = 3 29 4645 2 3 = 916
Applying
F cr L2
βcr =
EI
gives
30 10 6 16 8045 1 ft2
F cr 1 = 916 = 3207 lb 8 152
10002 144 in 2
The application of the higher modes of drill pipe buckling can only be viewed as approximate
since in reality the pipe makes contact with the well bore, but even this could be helpful in hori-
zontal drilling where the vertical portion of drill pipe is sometimes put in compression.
Mode shapes and bending moment distribution are given in Ref. [17].
Bending Equation for Marine Risers
The mechanics of buckling of marine risers is similar to that of drill pipe except for the effective
tension term, since in general fluid densities of drilling mud and sea water are different. These fluid
forces have a significant effect on marine riser bending. The differential equation of bending con-
taining these forces will now be discussed.
Unique Features of the Differential Equation of Bending

The differential equation of bending for long vertical pipe is developed from the differential element
shown in Figure 8.51. The solid vectors represent the actual applied forces and body force of gravity,
while the dashed vectors represent the statically equivalent hydrostatic forces both inside and out-
side the riser pipe. The independent variable, x, is measured from the lower end upward.
The parameters in the drawing are defined below:
Fx – actual internal force in pipe, lb

FB − internal force at location at the bottom end, lb
V – shear force, lb
M – bending moment, lb-ft
θ – slope of deflection at x location, rad
P = Aopo − Aipi = (Aoγ o − Aiγ i)(L − x) lb
bo = Aoγ o lb/ft
bi = Aiγ i lb/ft
γ o − weight density of external fluid, lb/ft3
γ i − weight density of internal fluid, lb/ft3
γ − weight density of pipe material
w − air weight of pipe per length(γAc)
q – side load per unit length, lb/ft
T
θ + dθ
Fx + dFx
P + dP
V + dV M + dM
q b0
L
dx w + bi dx
V
M
x P Fx
FB
Figure 8.51 Freebody diagram of a differential marine riser section.
Summation of forces in the x-direction gives

dF x
=w 8 153
dx
F x = wx + F B 8 154
Summation of forces in the y-direction gives
dV d dy
= Ao γ o − Ai γ i L − x +q 8 155
dx dx dx
where q represents transverse loads applied over length, dx. The effect of an addition uniform con-
2
stant pressure, p, throughout the pipe can be included by adding − pAi dx
d y
2 to the right side of
Eq. (8.155). This term is important in analyzing bending of production tubing. Summation of
moments gives
dM dy
− Fx −V = 0 8 156
dx dx
Combining Eqs. (8.154)–(8.156) gives
d2 M d dy dV
− Fx − =0 8 157
dx 2 dx dx dx
2
d y
Applying Euler’s equation, M = EI dx 2 , the differential equation of bending for riser pipes with
variable internal tension (or compression) and exposed to inside and outside hydrostatic
pressure is [4]
d4 y d dy
EI − F B + wx + L − x Ao γ o − Ai γ i =q 8 158
dx 4 dx dx
The variable coefficient in the second term is special and unique to this equation. This term
includes fluid pressure inside and out as well as the weight of the pipe. Each of these parameters
vary linearly with vertical distance along the pipe. This term creates a challenge in solving the dif-
ferential equation and is called “effective tension” a unique feature of the mathematical solution to
this bending problem.
Effective Tension
The effective tension term is clearly shown within the brackets of Eq. (8.158). When uniform inter-
nal pressure is included, the effective tension becomes
T eff = F B + wx + L − x A0 γ 0 − Ai γ i − Ai p 8 159
Effective tension is defined for mathematical convenience. It occupies the space in the differential
equation set aside for the actual tension, FB + wx. Effective tension alone accounts for the effects of
hydrostatic pressure inside and outside riser pipes as well as uniform pressure throughout the pipe
from top to bottom. If, for example, a vertical pipe is suspended in air, then
T eff = F B + wx 8 160
This is the true internal tension.

If a vertical pipe is suspended in a given fluid with density γ m (both inside and outside) effective
tension becomes
T eff = F B + wx + L − x wm 8 161
where wm = Acγ m, having units of force per length, which represents the weight of the fluid (drilling
mud) displaced per unit length of pipe.
When the variable terms are combined
T eff = F B + wm L + x w − wm 8 162
Buckling of Marine Risers

The difference between the buckling of marine risers in the ocean and drill pipe buckling is that
fluid outside risers is sea water while fluid inside risers is drilling fluids. Drilling fluids are signif-
icantly greater that sea water. The differential equation of marine riser bending accounting for
these effects on effective tension is [4]
d4 y d dy
EI 4 − F B + wx + L − x Ao γ o − Ai γ i =0 8 163
dx dx dx
The dimensionless form of this equation is
d4 y d2 y d2 y dy
4 − αζ 2 − β 2 −α =0 8 164
dζ dζ dζ dζ
where
x
ζ= and
L
w − A0 γ 0 + Ai γ i L3
α=
EI
β=
EI
The results of the drill pipe buckling analysis, discussed earlier, is now applied to marine risers. In
this case, the expressions for α and β account for the difference between fluid densities inside and
outside risers. The sign on β is adjusted to account for the sign difference in the third term in
Eqs. (8.145) and (8.164).
β= −
EI
where FB is the actual internal force in the riser at the mud line connection (x = 0).
The relation between the critical value of βcr and α, according to Eq. (8.147) becomes
2
F B + LA0 γ 0 − LAi γ i L2 w − A0 γ 0 + Ai γ i L3 3
− =ϕ 8 165
EI cr EI
Solving for FB gives the critical internal force at x = 0 causing risers to buckle.
1
EI 3 2
F B,cr = − ϕ w − A0 γ 0 + Ai γ i 3 − LA0 γ 0 − LAi γ i 8 166
144
Note that in this formula FB is the actual force at the lower end of the riser and can be plus
(tension) or minus (compression).
As a check we apply this formula to the drill pipe buckling problem where
FB = − (F + wmL) and γ i = γ 0 = γ m
Recall that Lwm = Lγ mAc or wm = γ mAc. By substitution
1
EI 3 2
− F + wm L = − ϕ w − wm 3 − LAc γ m 8-167
144
1
EI 3 2
F cr = ϕ w − wm 3
144
In this case (drill pipe model) the force F, is the compressive force above local hydrostatic at the
top of the drill collars.
For values of α greater than 4000, the equation below can be used.
2 2
βcr = ϕ1 α3 = 3 29α3 1st Mode 8-168
Accounting for units
w − A0 γ 0 + Ai γ i L3 144in2
α=
EI ft 2
F B,cr + LA0 γ 0 − LAi γ i L2 144in2
βcr = −
EI 1ft 2
For values of α less than 4000, ϕ should be taken from Table 8.5.
Table 8.5 Relation between ϕ1 values and α (first mode).
α ϕ1
1 20.89
2 13.54
5 8.02
10 5.74
20 4.48
50 3.76
100 3.62
200 3.56
500 3.46
1000 3.40
2000 3.34
4000 3.29
Example Critical forces, FB, were calculated using equation (8.166) and plotted in Fig 8-52 for the
operational data listed below.
Riser OD - 22 in.
Riser ID - 20.75 in.
A0 - 2.64 ft2
A1 - 2.35 ft2
w - 142.81 lb/ft
I - 2398 in4
E - 29 × 106 psi
γ0 - 64 lb/ft3 (sea water)
γi - variable, ppg (7.48 gal/ft3)
L - variable, ft
This figure gives riser tension (compression) at the bottom connection, at impending buckling.
It is curious that each line passes through a common point.
If A0γ 0 = Aiγ i, then Eq. (8.166) becomes
500
400
5000 ft 4000
Critical buckling force, kip
300
3000 ft
200
2000 ft
100
1000 ft
0
8 10 12 14 16 18
–100
–200
Mud weight, ppg
Figure 8.52 Critical buckling force, (FB)cr, for marine risers (first mode).
1
EI 3 2
F B,cr = − 3 29 w 3 8 169
144
In this case, the critical buckling force is independent of L. For the example case, this point cor-
responds to
2 6398
γi = 64 = 70 897 lb ft3 or 9 48 ppg
2 3483
The critical buckling force in this special case is
1
29 × 106 2398 3
2
F B,cr = − 3 29 142 8 3 = − 70 5 kip
144
A separate calculation showed that the critical buckling force for each case in air is FB,cr is 70.52
kip (compression). This number is determined from Eq. (8.166) with γ i = γ o = 0.
Length, L, is not a factor assuming α 4000. Note that for lengths of 2000 ft and 1000 ft, ϕ is
slightly higher that 3.29 and in the range of 3.56 so these lines are slightly affected.
Numbers that are minus, represent the amount of compressive force at the bottom end, that ini-
tiate buckling. Plus numbers define the magnitude of tension at the lower end that correspond to
the onset of buckling. This data shows the effects of higher mud weight on buckling especially in
long riser pipes. Top tension in each case is determined from Ttop = FB + wL.
In anticipation of threatening weather, such as hurricanes, risers are sometimes disconnected
from mud line equipment creating a hydrostatic condition inside and outside the riser pipe γ i = γ o.
This operational step eliminates the possibility of axial buckling.
Tapered Flex Joints
It is fundamental that structural stiffness attracts bending moment. This makes the top and bottom
connections critical areas in the design and operation of risers. Ball joints eliminate bending, but
the angle across ball joints could create problems for drill strings and production tubulars and
equipment. Most of the riser is relatively flexible.

x F
δ
Bending (static and dynamic) over much of the riser
H is controlled by top tension.
Tapered flex-joints provide a more balanced transi-
tion between the flexible riser and fixed mudline
equipment. Taper flex joints are typically truncated
y(x) cones with constant inside diameter, but the outside
diameter varies linearly from top to bottom (see
Figure 8.53). The method of analysis developed by
Dareing [4] allows tapered stress joints to be analyzed
by solving only one differential equation, which pro-
L vides a closed form solution to bending deflections
and stresses.
x
Equation of Bending
Consider a tapered cantilever beam subjected to a
y
transverse force, H and a pull full, V at the free end
as shown in Figure 8.53. The differential equation of
Figure 8.53 Tapered flex joint (add delta). bending for the beam is
d2 y
EI x − Fy = L − x H − Fδ 8 170
dx 2
Gravity and hydrostatic forces are dropped because their effects are small by comparison with
forces applied at the top (V and H). The solution to this equation is straightforward if there is
no pull force, but this force is central to riser operations.
The solution to Eq. (8.170) is also complicated by the variable coefficient, I(x), which accounts for
the variation of the cross-sectional moment of inertia. By approximating the straight-tapered stress
joint with a tubular having a parabolic moment of inertia distribution, the bending equation can be
cast into the Euler type differential equation that is amenable to a closed form solution.
Parabolic Approximation to Moment of Inertia

If the distribution of the moment of inertia along the column is assumed to be of the form
2
h−x
I x = Ip 8 171
a
where
L
h=
Ip
1− I SJ
and
Ip
a=h
I SJ
where
Ip − moment of inertia of riser pipe

ISJ − moment of inertia of bottom end of tapered stress joint
the differential equation of bending becomes

2
EI p 2d y
h − x − Fy = L − x H − Fδ 8 172
a2 dx 2
Because I(x) is assumed to vary parabolic, the associated stress joint will be referred to as a “par-
abolic stress joint.” Table 8.6 compares the parabolic moment of inertia with the moment of inertia
of a straight taper over a 30 ft length.
Table 8.6 Comparison of wall thickness and moment of inertia over a 30 ft flex joint (23.5 in. OD base, 22 in.
OD at top and having a 20.5 in. ID throughout).
Wall thickness (in.) Moment of inertia (in.4)
Joint height (ft) Parabolic Straight Parabolic Straight
0 1.500 1.500 6301 6301

1 1.473 1.475 6164 6174
2 1.446 1.450 6027 6048
3 1.419 1.425 5892 5923
4 1.392 1.400 5759 5798
5 1.365 1.375 5627 5674
6 1.339 1.350 5497 5551
7 1.313 1.325 5369 5429
8 1.286 1.300 5241 5308
9 1.260 1.275 5116 5187
10 1.234 1.250 4992 5067
11 1.208 1.225 4869 4948
12 1.183 1.200 4748 4830
13 1.157 1.175 4629 4712
14 1.132 1.150 4511 4596
15 1.106 1.125 4394 4480
16 1.081 1.100 4279 4365
17 1.056 1.075 4166 4250
18 1.032 1.050 4054 4136
19 1.007 1.025 3943 4023
20 0.983 1.000 3835 3911
21 0.959 0.975 3727 3800
22 0.935 0.950 3621 3689
23 0.911 0.925 3517 3579
(Continued)
Table 8.6 (Continued)
Wall thickness (in.) Moment of inertia (in.4)
Joint height (ft) Parabolic Straight Parabolic Straight
24 0.887 0.900 3414 3470

25 0.864 0.875 3313 3361
26 0.841 0.850 3213 3254
27 0.818 0.825 3115 3147
28 0.795 0.800 3018 3040
29 0.772 0.775 2923 2935
30 0.750 0.750 2830 2830
Solution to Differential Equation

The solution to Eq. (8.172) has particular and complimentary components:
y x = yP x + yC x 8 173
The particular solution is
H
yP x = − L − x +δ 8 174
F
To determine the complementary solution, it is necessary to perform certain substitutions. By
definition
a2 F
γ2 = 8 175
EI P
The differential equation giving the complementary solution is
2 d2 y
h−x + γ2y = 0 8 176
dx 2
This is a linear differential equation containing a variable coefficient. This equation is of the
Euler type.
The earlier approximation for I(x), using a parabolic function, was made to take advantage of the
classic method of solving the Euler type differential equation.
The solution starts by eliminating the variable coefficient as follows:
h − x = ez 8 177
Differentiating, gives
− dx = ez dz
and
dz 1
= − z 8 178
dx e
Note that
dy dy dz 1 dy
= = − z 8 179
dx dz dx e dz
d2 y 1 d2 y dy
2 = − e2z − 8 180
dx dz2 dz
and
2
h−x = e2z 8 181
Combining Eqs. (8.179) and (8.180) with Eq. (8.176) gives
d2 y dy
− − γ2y = 0 8 182
dz2 dz
The variable coefficient is gone, and the solution to Eq. (8.182) can be obtained by assuming
y = CeDz. It follows that
D2 − D − γ 2 CeDz = 0 8 183
which leads to
1
D1,2 = 1± 1 + 4γ 2 8 184
2
The complementary solution is therefore
D1 D2
yC x = C1 eD1 z + C2 eD2 z = C 1 h − x + C2 h − x 8 185
The total solution to Eq. (4.75) is
D1 D2 H
y x = C1 h − x + C2 h − x − L−x +δ 8 186
F
Equation (8.186) defines the deflection in the tapered flex joint. This closed form solution sim-
plifies the simultaneous solution of stress joints coupled with marine risers as well as with offshore
pipe lines and flow lines.
Example Consider the tapered flex joint defined in Table 8.6. The pull force and horizontal force
at the top are assumed to be FB = 75 000 lb and H = 50 000 lb. Parameters that enter into the deflec-
tion and bending stress calculations are
L
h= = 90 94
Ip
1− I SJ
Ip
a=h = 60 94
I SJ
a2 F 60 942 75000 144 in 2
γ2 = =
EI p 29 10 6 2830 1ft2
γ = 0 699
Applying boundary conditions

y 0 =0 8 187a
dy
0 =0 8 187b
dx
y L =δ 8 187c
gives
H
hD1 C1 + hD2 C2 + δ = L 8 188
FB
H
D1 h D1 − 1 C 1 D2 h D2 − 1 C 2 = 8 189
FB
D1 D2
h−L C1 + h − L C2 = 0 8 190
The solution to these three algebraic equations gives: C1 = 0.0856, C2 = −100.079 and δ = 0.417 ft.
The stress distributions for his special case are given in Table 8.7. These numbers apply to an
assumed stand-alone tapered flex joint with assumed vertical and horizontal forces at the top. These
equations will be used to interface tapered flex joints with uniform riser pipes.
Table 8.7 Deflection and stress in tapered flex joint (corresponds to data in Table 8.6; H = 50 000 lb, FB = 75
000 lb).a
Deflection y Bending moment Bending Axial Total tension

(x) (ft) (ft-lb) stress (psi) stress (psi) stress (psi)
0.0000 1 468 742 32 865 723 33 588

0.0006 1 418 785 32 331 737 33 068
0.0023 1 368 914 31 778 750 32 528
0.0051 1 319 127 31 203 764 31 967
0.0091 1 269 424 30 607 779 31 386
0.0141 1 219 803 29 988 794 30 782
0.0203 1 170 262 29 344 809 30 153
0.0275 1 120 801 28 674 826 29 500
0.0357 1 071 418 27 977 842 28 820
0.0449 1 022 110 27 251 860 28 111
0.0551 972 877 26 495 878 27 373
0.0663 923 717 25 705 897 26 602
0.0785 874 627 24 881 917 25 798
0.0915 825 605 24 019 937 24 957
0.1054 776 650 23 118 959 24 077
0.1202 727 759 22 175 981 23 156
0.1358 678 930 21 187 1 005 22 191
0.1522 630 160 20 150 1 029 21 179
0.1694 581 446 19 061 1 055 20 116
Table 8.7 (Continued)
Deflection y Bending moment Bending Axial Total tension

(x) (ft) (ft-lb) stress (psi) stress (psi) stress (psi)
0.1873 532 786 17 917 1 082 18 999

0.2058 484 175 16 712 1 110 17 822
0.2249 435 612 15 442 1 140 16 582
0.2447 387 092 14 103 1 172 15 274
0.2649 338 612 12 687 1 205 13 892
0.2857 290 168 11 189 1 240 12 429
0.3069 241 756 9 602 1 276 10 878
0.3284 193 370 7 917 1 316 9 232
0.3502 145 007 6 125 1 357 7 482
0.3723 96 662 4 216 1 401 5 617
0.3945 48 328 2 179 1 448 3 627
0. 0 0 1 498 1 498
a
Straight taper geometry used in stress calculations.
Application to Marine Risers

The following example (Figure 8.54) illustrates the effect of flex joints by comparison with a plain
riser pipe. Calculations [4] of the two deflections are based on the following input data:
Ocean depth is 1000 ft

Riser dimensions are 22 in. OD by 20.5 in. ID
Pull force at the bottom of riser is 150 000 lb
Top lateral rig offset is 50 ft
Mud wt (12 ppg) and sea water wt (64 lb/ft3)
Cross section inertia of riser is 2830 in.4
Cross section inertia at bottom of flex joint is 6301 in.4
Top dimension of flex joint matches riser dimensions
Only the lower 65 ft is shown in the figure.
Torsional Buckling of Long Vertical Pipe
The governing differential equations of bending stem from Huang and Dareing [18]. Figure 8.55
illustrates the model and shows the coordinates used in the analysis. There are two differential
equations of bending caused by torsion buckling. One equation relates to bending in the xz plane.
The other relates to bending in the yz plane.
d4 x d3 y d dx
EI 4 − T 3 − T eff =0 8 191a
dz dz dz dz
70 Figure 8.54 Comparison of riser deflections.
60
Plain riser
50
Plain riser
Distance, ft
40
30
20
Flex joint
10
0
0.0 0.5 1.0 1.5 2.0
Displacement, ft
T d4 y d3 x d dy
EI 4 + T 3 − T eff =0 8 191b
dz dz dz dz
x
The symbol, T, without the subscript, represents applied tor-
y que, which is assumed constant over the pipe length. Effective
tension, Teff, however, is assumed to vary along the pipe.
Force, FB represents the actual internal force in the pipe at
z = L. Effective tension varies along the pipe and is defined
z
L in terms of the independent variable, z.
T eff = F B + w L − z + zwm 8 192
The first term is the actual tension in the pipe at location z,

while the last term represents the contribution of hydrostatic
fluid pressure at location z. Assuming fluid density is the same
both inside and outside the pipe, the force FB at the lower
exposed end of the pipe is FB = − Lwm giving
T T eff = L − z w − wm 8 193
FB
Going back to the more general case and rearranging
Figure 8.55 Torsion Eq. (8.192)
stability model.
T eff = F B + wm L + w − wm L − z 8 194
Boundary Conditions
Boundary conditions for pinned conditions at both top and bottom are
x 0 =y 0 =0 8 195a
x L =y L =0 8 195b
2
dx dy
EI 0 =T 0 8 195c
dz2 dz
d2 y dx
EI 2 0 = −T 0 8 195d
dz dz
d2 x dy
EI L =T L 8 195e
dz2 dz
d2 y dx
EI L = −T L 8 195f
dz2 dz
Equations (8.191a) and (8.191b) are now combined using the complex variable
u = x z + iy z 8 196
Multiplying Eq. (8.191b) by the imaginary number “i” and adding the resulting equation to
Eq. (8.191a) gives
d4 u d3 u d du
EI 4 − iT 3 − F B + wm L + w − wm L − z =0 8 197
dz dz dz dz
Now replace z with the dimensionless number:
L−z
ς= 8 198
L
Noting that dz = − Ldζ and dividing through by EI gives
d4 u d3 u d du
4 − iΘ − β + αζ =0 8 199
dζ dζ 3 dζ dζ
which expands into
d4 u d3 u d2 u d2 u du
4 − iΘ 3 −β 2 − αζ −α =0 8 200
dζ dζ dζ dζ 2 dζ
where
w − wm L3 wBFL3 wm
α= = , BF = 1 −
EI EI w
F B + wm L L2
β=
EI
TL
Θ=
EI
The force, FB, is the actual force at the lower end of the pipe. It is replaced by
F B = F − wm L
Recall
γ m ΔLAc = wm ΔL
γ m Ac = wm weight per unit length of pipe
So
wm L = γ m Ac L
which is the hydrostatic force acting on the lower end of an open pipe. The dimensionless β is now
expressed as
F B + wm L L2 FL2
β= =
EI EI
Here F can be tension (+) or compression (−) above hydrostatic compression at the bottom of
pipe. When drill pipe is attached to the top of the drill collars, this implies the neutral point is
located at the top of the collars.
Both Top and Bottom Ends Pinned

The boundary conditions transform into
u 0 =0 8 201a
u 1 =0 8 201b
2
du du
2 0 = iΘ 0 8 201c
dζ dζ
d2 u du
1 = iΘ 1 8 201d
dζ 2 dζ
Assuming the solution to the differential equation of the form
u ς = a0 + a1 ς + a2 ς2 + + a n ςn + 8 202
where the a’s are complex constants for a given set of boundary conditions, and ζ is the independent
real variable. u(ζ) is a complex function as defined earlier.
Substituting Eq. (8.202) into Eq. (8.200), gives the recurrence equation:
Θ β n−3 α
an = i an − 1 + an − 2 + an − 3 , n≥4 8 203
n n n−1 n n−1 n−2
The complex constants are dependent on α and β, which are real numbers. Using Eq. (8.202) and
its derivatives, the boundary conditions yield
a0 = 0 8 204a
∞
an = 0 8 204b
n=1
Θ
a2 = i a1 8 204c
2
∞ ∞
n n − 1 an = iΘ nan 8 204d
n=1 n=1
These conditions show that a0 = 0 and a2 depends on a1. From these conditions and by successive
application of the recurrence equation (Eq. (8.203)), the coefficient an can be expressed as a linear
combination of a1 and a3 or
an = F n + iJ n a1 + H n + iLn a3 8 205
where Fn, Hn, Jn, Ln are real numbers. It is apparent that for n = 0, 1, 2, 3 and using boundary con-
ditions Eqs. (8.204a) and (8.204c):
F0 = 0 J0 = 0 H0 = 0 L0 = 0
F1 = 1 J1 = 0 H1 = 0 L1 = 0
F2 = 0 J2 = Θ 2 H2 = 0 L2 = 0
F3 = 0 J3 = 0 H3 = 1 L3 = 0
The remaining values for Fn, Hn, Jn, Ln (n ≥ 4) are determined from the recurrence equations
derived from Eqs. (8.203) and (8.205). The starting values are used to obtain higher-order values
of Fn, Jn, Hn, Ln by repeated application of Eqs. (8.206a)–(8.206d), which are each functions of
Θ, α, and β.
Θ β n−3 α
Fn = − Jn − 1 + Fn − 2 + Fn − 3 8 206a
n n n−1 n n−1 n−2
Θ β n−3 α
Jn = + Fn − 1 + Jn − 2 + Jn − 3 8 206b
n n n−1 n n−1 n−2
Θ β n−3 α
Hn = − Ln − 1 + Hn − 2 + Hn − 3 8 206c
n n n−1 n n−1 n−2
Θ β n−3 α
Ln = + Hn − 1 + Ln − 2 + Ln − 3 8 206d
n n n−1 n n−1 n−2
Next consider the second and fourth conditions. Substituting Eq. (8.205) into the boundary con-
ditions Eqs. (8.204b) and (8.204d), the following two homogeneous equations are obtained:
Fn + i J n a1 + Hn + i L n a3 = 0 8 207a
and
n n − 1 Fn + Θ nJ n + i n n − 1 Jn − Θ nF n a1
8 207b
+ n n − 1 Hn + Θ nLn + i n n − 1 Ln − Θ nH n a3 = 0
For nontrivial solutions of a1 and a3, the determinant of the coefficients in Eqs. (8.207a) and
(8.207b) must be zero. The expansion of the determinate produces a real part and an imaginary
part. Both parts contain the eigenvalue. The real part of the expanded determinate is
Fn n n − 1 Hn + Θ nLn − Jn n n − 1 Ln − Θ nH n
− Hn n n − 1 Fn + Θ nJ n + Ln n n − 1 Jn − Θ nF n = 0
8 208a
and the imaginary part is
Jn n n − 1 Hn + Θ nLn + Fn n n − 1 Ln − Θ nH n
− Ln n n − 1 Fn + Θ nJ n − Hn n n − 1 Jn − Θ nF n = 0
8 208b
Let the left sides of Eqs. (8.208a) and (8.208b) be denoted by P and Q, respectively, then
P Θ, α, β = 0 8 209a
Q Θ, α, β = 0 8 209b
Both are real value functions containing the eigenvalue Θ. Assuming α and β are given only one
equation is needed. We choose P(Θ, α, β) = 0 from which to determine the eigenvalues. This is done
by trial.
Simply Supported at Both Ends with no End Thrust

Eigenvalues generated from Eq. (8.209a) are listed in Table 8.8. Pipe length, L, is reflected through
the dimensionless number, α. Critical torque is determined from the dimensionless number, Θ.
Note that as α gets smaller Θ approaches 2π as expected.
The first example is designed to compare the results with Greenhill’s solution. Both pipes are
assumed suspended in air for which β = 0 and buoyancy factor, BF = 1. Input values for these cal-
culations are
OD − 5½ in
ID − 4 892 in
w − 19 2 lb ft
I − 18 27 in 4
E − 30 × 106 psi
Numerical results are plotted in Figure 8.56 along with results from Greenhill’s equation. Pipe
length and critical torque were extracted from Table 8.8.
Critical torques predicted for the drill pipe case (linear tension) is greater than torques predicted
Greenhill’s first mode. These two curves converge for short pipe as tension approaches zero. The
linear tension case does have a second mode, but it is somewhat higher than for the Greenhill
Table 8.8 Eigenvalues for torsional buckling.
α Θ
1 6.5
2 6.6
10 7.5
20 7.9
40 8.3
50 8.37
100 8.552
200 8.66
500 8.783
1000 8.865
2000 8.935
350 000
300 000
Critical torque, ft-lb
250 000
200 000
Greenhill second mode
150 000
100 000
Linear tension
50 000 First mode
0
0 200 400 600 800
Pipe length, ft
Figure 8.56 Comparison of results with Greenhill’s classic solution.
second mode. In each case, there is no force applied to the pipe at the lower end. However, internal
tension, which increases linearly in vertical pipe, increases critical torque as expected.
Force Applied to Lower End

The math model allows for a force FB at the lower end. This force appears in the β term. Two cases
were considered: F = 5000 lb and F = −3000 lb. Buoyancy factor is zero (no fluid) in both cases.
Results show moderate amounts of tension does not affect critical torque especially for long pipe.
It loses its effect compared to linear increase in tension toward the surface.
However, if pipe is put in compression by a relatively small amount (−3000 lb), critical torque
reduces significantly (Figure 8.57).
This is due in part to the instability of the axial mode. A separate calculation of drill pipe buckling
[17] shows the critical buckling force for the axial mode is
1
EI 3 2
F cr = 3 29 wBF 3
144
250 000
200 000
Critical torque, ft-1b
F = + 5000 lb
150 000
100 000
F=0
50 000 F = –3000 lb
0
0 200 400 600 800
Pipe length, ft
Figure 8.57 Effects of applied force at lower end.

F cr = 3683 lb
This calculation is based on conditions, i.e. BF = 1. Assuming a 12 ppg drilling fluid (BF = 0.817),
the critical buckling force is Fcr = 3219 lb.
Effect of Drilling Fluid on Torsional Buckling

Calculations showed that drilling mud density has negligible effect on torsion buckling. However, it
does affect instability caused by axial forces.
Lower Boundary Condition Fixed

The data above was developed for a pinned-pinned boundary condition. We now consider the case
where the lower end is fixed and the top end pinned. This more accurately represents the rigidity at
the drill pipe – drill collars interface.
In this case, the initial values for F, J, H and L are
F0 = 0 J0 = 0 H0 = 0 L0 = 0
F1 = 0 J1 = 0 H1 = 0 L1 = 0
F2 = 1 J2 = 0 H2 = 0 L2 = 0
F3 = 0 J3 = 0 H3 = 0 L3 = 0
A comparison to critical torques for both sets of boundary conditions is given in Figure 8.58. Crit-
ical torque for the fixed-pinned case is greater than for the pinned-pinned case because of its rigid-
ity. However, this effect becomes less important for long pipe as indicated.
Operational Significance
Three events take place in drill pipe directly above the drill collars. The effects of each can jointly or
independently lead to lateral instability of the drill string, which could affect stick–slip torsional
vibrations and overall friction along drill strings:
1) Buckling of axial mode

2) Torsional buckling
3) Drill pipe whirl
250 000
200 000
Critical torque, ft-lb
150 000
Fixed-pinned
100 000
Pinned-pinned
50 000
0
0 200 400 600 800
Pipe length, ft
Figure 8.58 Comparison of critical torques for two sets of boundary conditions.
200
Critical torque, kip-ft
150
100
Tyld = 55826 ft-lb
50
F = –3000 lb
0
0 200 400 600 800
Pipe length, ft
Figure 8.59 Effect of neutral point in drill pipe on critical torque.
Buckling of axial modes are predicted by

1
EI 3 2
F cr = ϕ BFw 3 first mode
144
For 5½ in. drill pipe (19.2 lb/ft) is predicted by
1
30 × 106 18 27 3
2
F cr = 3 29 19 2 3 = 3683 lb
144
A 3000 lb axial compressive force is slightly lower than this critical force of 3683 lb. The result is a
lower torsional buckling torque as shown.
The mode shape of torsional buckling can be visualized as a coil trapped inside a well bore.
Yield torque 44 074 ft lb (E75)

55 826 ft lb (X95)
61 703 ft lb (G105)
79 332 ft lb (S135)
The yield torque or maximum possible torque that can be applied to this particular drill pipe is 55
286 ft lb (Figure 8.59). Critical torques are within this limit under typical drilling operations. Tor-
sional buckling is also enhanced by axial compression.
Pressure Vessels
Stresses in Thick Wall Cylinders

The theory of hoop and radial stresses in thick wall cylinders is derived from the theory of elasticity
and is documented in the literature [19]. These stresses are defined in terms of internal and external
pressures by Lames’ equations (Figure 8.60).
pi a2 − p0 b2 − a2 b2 p0 − pi r 2
σθ = 8 210
b2 − a2
Figure 8.60 Hoop and radial stress distributions (p0 = 0).
σθ
a
σr
b
pi a2 − p0 b2 + a2 b2 p0 − pi r 2
σr = 8 211
b2 − a2
σθ ℓt Stress distributions for the special case of p0 = 0 are shown in

Figure 8.60.
R The stress state within the cylinder’s cross section is biaxial if
axial stresses are not included. The maximum stress levels for
both hoop and radial are located on the inside surface.
p2Rℓ
Stresses in Thin-Wall Cylinders
σθ ℓt The Lame equations show that for Dt ≥ 20, the maximum hoop
ℓ stress is only 5% higher than the average hoop stress across the
cylinder wall [20]. The thin wall model gives a simple view of
how hoop stresses are developed (see the freebody diagram in
Figure 8.61 Hoop stresses in thin
wall pipe. Figure 8.61).
Summing forces in the radial direction gives
R D
σθ = p= p 8 212
t 2t
This simple formula gives reasonable engineering results in many cases.
Many pressure vessels can be analyzed using thin-walled equations. However, axial stresses
caused by end loads create an additional stress component in the vessel walls. This axial stress com-
ponent is
R
σa = p 8 213
2t
or one half of the hoop stress. Both hoop and axial stress components are the principal stresses. If
the end cap is spherical, then the stress at each point in the cap is
R
σϕ = p ϕ is arbitrary 8 214
2t
Each point in the spherical cap experiences the same stress level in all directions. A state of hydro-
static stress exists in the spherical cap and Mohr’s stress diagram is a point.
Stresses Along a Helical Seam

Pressure vessels are sometimes manufactured by rolling flat sheet metal into a helical shape and
then welding the seam as shown in Figure 8.62.
Figure 8.62 Stress along a pipe steam. t
σθ
σa
x
σn
θ
θ Seam n
Normal stress to the seam and shear stress in the seam is determined from Eqs. (8.4) and (8.6) by
setting τxy = 0 (x and y are principal axes)
σn = + cos 2θ 8 215
2 2
σx − σy
τnt = − sin 2θ 8 216
2
where
R
σx = σa = p axial stress
2t
R
σy = σθ = p hoop stress
t
By substitution
1 R
σn = p 3 − cos 2θ 8 217
4 t
1 R
τnt = p sin 2θ 8 218
4 t
Following the earlier sign convention, θ in the above figure is minus. For simplicity, we assume
the hoop stress Rt p = 10 000 psi and θ = 45 then by substitution
σ n = 7500 psi and τnt = − 2500 psi
Interference Fit Between Cylinders

Thin-Wall Cylinders
First consider the effect of an interference fit between two thin wall cylinders. This situation is
depicted by the expansion of the outer cylinder, δ0, and compression of the inner cylinder, δi, result-
ing in hoop stress on both. As a result, pressure is developed between the two contacting surfaces
(Figure 8.63).
The radial displacement of thin-walled cylinders due to either external or internal pressure is
determined from
a2
δ= p 8 219
hE
Outside surface, Figure 8.63 Radial displacements of

inner ring interference surfaces.
δi
Inside surface,
outer ring δ0
Interference, c
Contact Surface
where
a − average radius of cylinder

h − wall thickness
δ − radial displacement of cylindrical surface
The basis of this simple equation relates to pressure and stress as follows:
2π a + δ − 2πa δ
εθ = = 8 220
2πa a
a
σ θ = Eεθ and σ θ = p
h
Combining these equations gives the deflection equation relating δ to pressure, p (Eq. (8.219)).
For the interference case, radial deflection of each cylinder is
a2i
δi = p inside cylinder
hi E
a20
δ0 = p outside cylinder 8 221
h0 E
The given interference between the two cylinders relates to these displacements by
c = δ i + δ0
a2i a2
c= + 0 p 8 222
hi E h0 E
Assume dimensions as
a0 = 9 92 in h0 = 0 125 in c = 0 08 in
ai = 10 in hl = 0 125 in
Then the pressure between the two cylinders over the contacting surfaces is
102 9 922 p 1587

0 08 = + = p 8 223
0 125 0 125 E 30 × 106
p = 1512 psi
This means that the hoop stress in the cylinder walls is
a 9 92
σθ = p= 1512 = 120 ksi 8 224
h 0 125
Surface Deflection of Thick-Walled Cylinders

A similar problem exists when a thin wall cylindrical encloses a thick wall cylinder with an inter-
ference fit. In this case the radial displacement of the outside surface is formulated as follows.
At any radial location, ρ, there is no shear on the principal planes. Strain in the tangent
direction is
σ θ − νσ r
εθ = 8 225
E
This equation applies to any location, ρ, including both inside and outside surfaces of a thick-
walled cylinder. The radial displacement at any location, ρ within a thick-walled cylinder is
determined from
2π ρ + δ − 2πρ δ
εθ = =
2πρ ρ
where
δ σ θ − νσ r
= 8 226
ρ E
The stress condition at the outer surface of the left drawing in Figure 8.64 is
b2 + a 2
σθ = − p 8 227
b2 − a2
σr = − p 8 228
By substitution into Eq. (8.226), the radial deflection of the outside surface of the thick wall cyl-
inder is
pb b2 + a2
δ= − −ν 8 229
E b2 − a2
The radial displacement is inward.

Next consider the right drawing in Figure 8.64.
b2 + a2
σθ = p
b2 − a2
σr = − p
and by substitution the radial deflection of the inside surface is
pa b2 + a2
δ= +ν
E b 2 − a2
The radial displacement is outward

σr
σθ p
σr
b b σθ
a a p
Figure 8.64 Radial displacement in thick wall cylinders.
Thick Wall Cylinder Enclosed by Thin Wall Cylinder

Example Consider the previous case of two thin wall cylinders except in this case the inside cyl-
inder is thick-walled. As before interference between the two cylinder relates to these displace-
ments by
c = δi + δ0 8 230
where
pb b2 + a2
δi = − −ν inward 8 231
E b2 − a2
and
a2ave
δ0 = p 8 232
h0 E
Assume dimensions as
a0 = 10 in h0 = 0 125 in c = 0 08 in
a = 5 in b = 10 08 in aave = 10 0625 in
Applying c = δi + δ0
p10 08 10 082 + 52 p 10 062

0 08 = − 0 25 + 8 233
E 10 082 − 52 E 0 125
p
0 08 = 14 139 + 810
E
p
= 0 000 097
E
p = 2913 psi
compared with p = 1512 psi for the two thin wall
cylinders with the same interference. The difference
is the stiffness of the outer surface (ring) within the
thick wall cylinder.
2 in. 3 in.
Thick Wall Cylinder Enclosed by Thick Wall

Cylinder
Now consider the case where a sleeve pressed fitted
around a solid 2-in. diameter shaft (Figure 8.65).
Figure 8.65 Sleeve shrunk on a solid bar. The outside diameter of the sleeve is 3 in. Determine
the radial interference that will cause the sleeve to yield. Base your answer in the maximum shear
stress criteria of failure. Yield strength of the sleeve material is 60 ksi. Accordingly, τyld = 30 ksi.
The outside cylinder has a Dt ratio of 1.667 which is much less than 20 must be treated as a thick
wall cylinder. The approach is to first determine the relation between the interference and contact
pressure between the shaft and sleeve.
The radial deflection of the outside surface of the shaft is
− pb b2 + a2 −p p
δshaft = −ν = 1 1 − 0 25 = − 0 75 inward
E b2 − a2 E E
where a = 0 and b = 1 in. As before, pressure, p, is a plus number and the minus means radial
deflection of the shaft surface is inward.
The expression for the radial deflection of the inside surface of the sleeve is obtained by applying
Lames’ equation to the outer cylinder which now has an inside pressure of p and zero outside
pressure:
pa b2 + a2 p 1 52 + 12 p
δsleeve = +ν = 1 + 0 25 = 2 85
E b2 − a2 E 1 52 − 1 1 E
The interference is the sum of the two radial deflections.
c = δshaft + δsleeve
p p p
c = 0 75 + 2 85 = 3 6
E E E
p = 8 33 × 106 c
Now we return to the stress components at the inside surface of the sleeve.
b2 + a 2
σθ = p = 2 6p
b2 − a2
σr = − p
The maximum shear stress is

1
τmax = σ θ − σ r = 1 8p
2
With a material yield shear strength of 30 ksi
30 000 = 1 8 8 33 × 106 c
Giving c = 0.002 in. and p = 16 660 psi. The maximum shear stress is 29 988 psi. Corresponding
hoop stress is σ θ = 2.6(16 660) = 43 318 psi.
This example shows that a small interference between a shaft and thick-walled cylinder creates
very high contact pressure and shear stress.
Elastic Buckling of Thin Wall Pipe

One way to view the buckling of thin-walled cylinders is to consider the circumference as a straight
beam with end loads. This model is suggested by the shape of the primary mode as depicted in
Figure 8.66. Assuming Euler’s equation applies,
Cylinder
Buckle mode
Figure 8.66 Loaded circular ring.
2
4π d
F cr = EI = σ θ A = p tℓ four quarter-lobes 8 234
L 2t
2
4π t3 ℓ d
E = p tℓ 8 235
πd 12 2t cr
After simplification,
8 1
pcr = E 3 8 236
3 d t
This solution treats the circumference of the cylinder as a straight beam.
Bresse’s Formulation
Bresse [21] gives a more realistic model and solution to the buckling of circular rings. His formu-
lation was developed to predict buckling of straps of unit length. Figure 8.67 shows one fourth of the
ring of radius, r. The radial deflection, w(s), is expressed in polar coordinates. Curvature of the
deflection arc relative to the initial curvature (1/r) is
1 d2 w w
= + 2 8 237
ρ ds2 r
Assuming Euler bending
d2 w w
EI + 2 = −M 8 238
ds2 r
p w
Cylinder
Buckle mode P
r θ
P
Figure 8.67 Freebody diagram of segment of cylinder.

The bending moment at location s or θ is M = rwp. Therefore,
d2 w w
EI 2 + r2 = − rwp p is force per arc length of strap 8 239
ds
Collecting terms
d2 w 1 rp
+ + w=0 8 240
ds2 r2 EI
d2 w
+ k2 w = 0 8 241
dθ2
where
r3p
k2 = 1+
EI
w θ = C1 sin kθ + C2 cos kθ 8 242
Applying boundary conditions, w(0) = 0 and w(π/2) = 0 leads to
C2 = 0
and
π
0 = C 1 sin k
2
Leading to a nontrivial solution of
π
k =π
2
Instability or buckling occurs when k = 2.
pr 3
1+ =4 8 243
EI
3EI t3
pcr = 3 with I =
r 12
E t 3
pcr = viewed as pressure since strap as unit width
4 R
1
pcr = 2E Bresse's equation 8 244
D t 3
Compared with Eq. (8.236), the Bresse formula predicts a smaller buckling pressure.
Application to Long Cylinders

When applying Bresse’s equation to long cylinders E is modified as follows. In the bending of plates
and shells, longitudinal (z axis) strain at a point is zero. Consider plate bending as shown in
Figure 8.68.
1
εz = σ z − νσ x 8 245
E
y Figure 8.68 Bending of flat plate (z axis out

of paper).
Mz Mz
x
In the bending of plates and shells, strain in the z direction is zero. The means
σ z = νσ x out of paper 8 246
By substitution
1
εx = σ x − ν2 σ x 8 247
E
E
σx = εx
1 − ν2
This affects the buckling of long tubes and requires that E be replaced by 1 −E ν2 in Euler bending.
Therefore, Bresse’s formula becomes
2E 1
pcr = compare with Eq 8 244 8 248
1 − ν2 d t 3
Consider a long, thin-walled pipe with outside diameter of 8 in. and a wall thickness of 0.25 in.:
d
t = 0 825 = 32. If the yield strength of pipe material is 75 000 psi, determine the critical buckling
pressure. Will the pipe fail by buckling or material yielding? Assume E = 30 × 106 and ν = 0.25.
2 × 30 × 106 1
pcr = = 1953 psi 8 249
1 − 0 252 32 3
Since
2σt d
p= , σθ = p 8 250
d 2t
Circumferential stress in pipe wall is
d 8
σθ = p = 1953 = 31 248 psi 8 251
2t cr 2 0 25
The pipe will buckle before reaching the yield stress level of 75 000 psi.
Thin Shells of Revolution

Thin shells of revolution are generated by rotating the geometry of the meridian curve about the
axis of symmetry (Figure 8.69).
A freebody of a surface element gives
σm σt p
+ = 8 252
rm rt t
Figure 8.69 Thin shell of revolution. y

x
σm
P y = y(x)
σt
0t
0m ϕ
F P F
This equation contains two unknowns, σ m and σ t. The stress, σ m, is determined independently
from equilibrium of global forces in the y direction.
Fy = 0
p πx 2 − 2πxt σ m cos ϕ = 0
Therefore,
x p
σm = 8 253
2t cos ϕ
where
dy
tan θ = and ϕ = 90 − θ
dx r
The radius of curvature, rm, is determined from the curvature of the meridian. The tangent radius,
rt relates to x by x = rt cos ϕ. Both radii are defined by
d2 y
1 dx 2
= 3 8 254
rm dy 2 2
1+ dx
x
rt =
cos ϕ
In the diagram
r m = 0m P and r t = 0t P
Example Consider a shell of revolution as shown in Figure 8.70. Its profile is defined by
dy d2 y
y = 8 − 0 5x 2 = −x = −1
dx dx 2
Internal pressure is p = 1500 psi. We wish to find the stress, σ m and σ t at y = 6 in. The first step is to
find σ m.
x p
σm = 8 255
2t cos ϕ
The elements in this equation are
x = 2 in y = 6 in
t = 0 125 in
p = 1500 in
dy
= −2 θ = 63 4 ϕ = 26 6
dx x=2
The meridian stress is

2 1500
σm = = 13 420 psi 8 256
2 0 125 cos 26 6
We now turn to the tangent stress.
σm σt p
+ = 8 257
rm rt t
Elements in this equation are
d2 y
1 dx 2 −1 −1
= 3 = 3 = 8 258
rm dy 2 2
1 + −2 2 2 11 18
1+ dx
r m = 11 18 in
8
y = 8 – 0.5x2
(2, 6)
F
F
Ot
Om ϕ
θ x
4
Figure 8.70 Surface of revolution shell.

Curved Beams 455
Also
x 2
rt = = = 2 24 in 8 259
sin θ sin 63 4
By substitution, the tangent stress is
p σm 1500 13 420
σt = − rt = − 2 24 psi 8 260
t rm 0 125 11 18
σ t = 21 815 psi
As a check on this number, hoop stress in the wall of a cylinder with R = 2 in. and t = 0.125 in.
R 2
σθ = p= 1500 = 24 000 psi in the same ballpark
t 0 125
Curved Beams
Winkler [5] developed the theory of curved beams in 1858. Bending stresses in curved beams are
based on transverse planes remaining planes after bending as in the case of straight beams, i.e. there
is no warping in the plane. With reference to Figure 8.71, normal strain at location y is
Δθ y Δθ r n − r
ε= = 8 261
Δϕ r Δϕ r
For a given applied moment, M
Δθ
=C constant 8 262
Δϕ
where
Δϕ − differential arc length
Δθ − angular rotation of transverse plane (caused
by bending moment, M)
And since
σ = Eε c
n M
rn − r e y
σ=E C 8 263
r
n r
Location of Neutral Axis
c Δθ
rn
As with the straight beam theory, the sum of nor-
mal forces across the face of the transverse plane R
Δϕ
is zero:
σ r dA = 0 8 264
A
Figure 8.71 Curved beam parameters.
Substitution for σ yields the location of the neutral axis:

rn
− 1 dA = 0
r
A
A
rn = dA
8 265
r
A
The neutral axis is not the centroid axis of the cross section as with the straight beam.
Stress Distribution in Cross Section

The stress distribution across the beam depends on the magnitude of the bending moment.
Moments of internal stresses are equal to the external moment, M:
M = y σ dA = r n − r σ dA 8 266
A
There are several steps leading to the stress distribution equation. The steps include
rn − r 2
M = EC dA 8 267
r
A
r 2n − 2r n r + r 2
M = EC dA
r
A
dA
M = EC r 2n − 2r n A + r n − y dA 8 268
r
A A
But R = rn + e and R − e = rn. Further substitution and collecting terms give

M = EC R − r n A 8 269
Replacing C with
σr
C= from Eq 8 263
E rn − r
n By substitution
y σr
M= R − rn A 8 270
rn − r
n My
σ= 8 271
M Ae r n − y
ϕ rn
which defines the stress distribution across the beams
section (Figure 8.72). Stress distribution and magnitude
depend on A, e, and rn.
Example To illustrate the procedure, consider the
Figure 8.72 Stress distribution in a curved beam with rectangular cross section shown in
curved beam.
Figure 8.73. We assume the following parameters for this
example:
Curved Beams 457
Figure 8.73 Rectangular cross section.

b
h c c
r0 n n
dr
R
rn
ri r
h = 2 in.
r0 = 3 in.
ri = 1 in.
R = 2 in.
b = 0.75 in.
By substitution
h
rn = = 1 82 in
ln rroi
e = 2 − 1 82 = 0 18 e = R − rn
M = 1000 ft-lb
A = 1 5 in 2
By substitution
My
σ=
Ae r n − y
12 000y
σ=
1 5 0 18 1 82 − y
The plot of this equation is shown in Figure 8.74.
Bending stress is a maximum at y = rn − ri = 1.82 − 1 = 0.82.
12 000 0 82
σ max = = 36 440 psi
1 5 0 18 1
By comparison, the straight beam formula gives a maximum stress of
Mc bh3
σ max = I= = 0 5 in 4
I 12
1000 12 1
σ max = = 24 000 psi
05
40.0
30.0
20.0
Bending stress, ksi
10.0
0.0
0 0.5 1 1.5 2 2.5 3
–10.0
–20.0
–30.0
Radial location, in.
Figure 8.74 Bending stress across section.
The exact solution to the curved beam problem was developed by Golovin [22] based on the the-
ory of elasticity. The exact solution shows there is also a radial stress component not considered by
the Winkler solution.
For a curved beam having a rectangular cross section, it can be shown that the normal stress com-
ponent can be represented by
M
σθ = m 8 272
a2
where
a – inside radius of curved beam

b – outside radius of curved beam
m – numerical factor as shown in Table 8.9
The numbers show that the Winkler model gives very accurate results.
Example The hook shown in Figure 8.75 is to carry a 2000 lb load. Determine the factor of safety
against material yielding of 80 000 psi using the von Mises criteria of failure.
a = 2 in inside radius
c = 3 5 in outside radius
b1 = 1 in inside width
b2 = 0 5 in outside width
Table 8.9 Coefficient “m.”
Winkler model [5] Exact solution [22]
b/a Linear stress (straight beam) Inside Outside Inside Outside
1.3 ±66.67 +72.98 −61.27 +73.05 −61.35

2 ±6.0 +7.725 −4.863 +7.755 −4.917
3 ±1.5 +2.285 −1.095 +2.292 −1.13
Curved Beams 459
3.5 in.
rn
2
0.5 1
F
A
Section A–A
Figure 8.75 Bending stresses in a curved hook.
The elements in the stress equation are listed below:

My
σ= bending stress, y distance from neutral axis
Ae r n − y
b1 + b2
A= c − a = 1 125 in 2 cross sectional area
2
dA b1 c − b2 a c
= ln − b1 + b2 = 0 4327 in
r c−a a
A
rn = dA
= 2 6 in radial distance to neutral axis of bending
r
The determination of rn is often a difficult task for many cross-section shapes. These values can be
found for various shapes in tabular form [23, 24]:
a 2b1 + b2 + c b1 + 2b2
R= = 2 67 in radial distance to centroid
3 b1 + b2
e = R − r n = 0 07 in
y = r n − a = 0 6 in location of max bending stress
Substituting these numbers into the bending stress equation gives the maximum stress at the
inside surface of
2000 2 6 0 6
σ max = = 19 810 psi
1 125 0 07 2 6 − 0 6
Total stress
2000
σ = σ b + σ a = 19 810 + = 21 587 psi
1 125
Factor of safety
80 000
FS = = 3 71
21 587
Shear Centers
Beams made up of thin members as shown in Figure 8.76 are susceptible to twisting as well as bend-
ing [23]. There is, however, a longitudinal bending axis through which transverse bending loads
must pass to avoid twisting of the beam. The intersection of the longitudinal bending axis and trans-
verse plane is called the shear center, indicated by point 0. The shear center is a fixed point in the
beams cross section and is determined as follows.
When a transverse force is applied to a beam, it produces shear stresses in each section of the
beams cross section. The vector sum or resultant of the internal shear forces (V1, V2, V2) are equal
to the total internal shear force generated by the external shear force, V.
V= Vi = VR 8 273
1, 2, 3
VR is the resultant of all three shear forces. The line of action VR is determined from the principle
of moments, i.e.
eV R = di V i = Ti 8 274
1, 2, 3 1, 2, 3
The reference point is arbitrary, but it is convenient to choose point “a” in this case. Distance, e,
locates the line of action of VR. It also locates point 0, a point where VR has zero torque. This point is
called the shear center. This means that if the external shear force is applied through the shear cen-
ter, it will produce zero twist on the beam. If it is applied away from the shear center, VR and V
produce a twisting couple as shown.
VR
V
0 e
a
VR C
Figure 8.76 Shear center.

Shear Centers 461
Bending axis and shear centers are of special interest in beams made up of thin parts. The location
of shear centers is important in the use of angle and channel beams as well as other nonsymmetrical
cross-section geometries containing thin members. Beams of this type are often used in flooring,
roofing, and even aircraft wings. To avoid twisting these types of beams the line of action of applied
load or loads must pass through the shear center. The following explains how to locate shear
centers.
The earlier discussion of shear stress showed that shear stress in beams is determined by
VQ
τ y = 8 275
Ib
Following its derivation, shear stress occurs in vertical and horizontal planes. The maximum
shear stress typically occurs at the neutral axis of bending.
An assumption in applying this formula to beams made up of thin members is: shear stresses lie
within a plane transverse to thin members regardless of orientation. These stresses are still gener-
ated by the gradient of the bending moment.
If any section is horizontal as shown in Figure 8.77, it is viewed as having the ability to contain
shear flow within the walls of the thin members. Shear flow in the transverse plane is determined as
follows.
(a)
F1
dx
F3
F2
τ
z
(b)
b b
z
VR
e V1 dz
b–z
C
h
0
Longitudinal t
bending axis V2
Point a
V3
Figure 8.77 (a) Shear flow in horizontal member. (b) Shear flow around channel cross section.
dx
t
F2 t
b
b
z F3
F1 τ(z)
z
τ(z)
dz
V
Figure 8.78 Shear flow.
Shear flow for a channel beam made up of three thin members is shown in Figure 8.78:
b
V 1
V1 = b − z tht dz 8 276
It 2
0
Vthb2
V1 = 8 277
4I
Noting that the moment of the resultant is equal to the sum of the moments of the parts, both
respect to point “a” gives
h
eV R = V 1 + V 2 = hV 1
2
Vthb2
eV R = h 8 278
4I
Since VR = V
1
2b
e= 1
8 279
1+ 6 aw + a f
where
aw = wh
af = bt
Angle iron is another common beam with cross section shown in Figure 8.78. Its shear center is
determined as follows.
The local shear stress in a cross section at any angle from vertical is determined the same as for a
vertical section. The shear flow, V2 in the thin member is
b−z
Q = b−z sin θ 8 280
2
dV = τt dz
b
V 1
V2 = b−z t b + z sin θ t dz 8 281
It 2
0
Shear Centers 463
Vt 2 3
V2 = b 8 282
I 6
1 b3
I= z sin θ 2 t dz = t 8 283
2 3
2
V2 = V 8 284
2
where θ = 45 . When θ is zero, the shear force along the thin leg is zero. In this case, the angle strip
is horizontal and at the neutral axis of bending; therefore, normal forces are zero.
The horizontal force components of V1 and V2 cancel out. The resultant of the vertical compo-
nents is
2 2
2 2
VR = V + V 8 285
2 2
or
VR = V 8 286
The resultant of the two shear forces in each section passes through point 0 (Figure 8.79). The
twisting moment of the resultant with respect to point 0 is zero; therefore, point 0 is the shear cen-
ter. If the transverse force is applied away from the shear center, say at the centroid C, then the
resulting internal shear force, VR, along with the external shear load, V, will cause the angle iron
to twist.
Example Consider a 5-mm plate of steel formed into the semicircular shape as shown in
Figure 8.80. The distance, e, to the shear center is determined as follows.
Local shear at any location, θ, is
VQ
τ θ = 8 287
It
where
θ
Q= y dA
0
y = r cos θ
dA = r dθ t
θ V1 = 2 V
2
Q = t r cos θ dθ = tr sin θ
2 2
0
C
so
V2 = 2 V
V 2 2
τ θ = r t sin θ 8 288 VR
It
Moments with respect to point O of the shear flow,
dVs = τ(θ)dA is
Figure 8.79 Angle iron beam with
dM = r dV s = r r dθ τ θ t internal shear.
V 2 4
dM = t r sin θ dθ
τ It
Vt
r M = 2 r4
θ I
VR
π 3
I= r t
dθ 2
After substitution, and integration, the internal moment
e of all differential shear forces is
t 4
M= rV 8 289
π
Equating this to the moment of the internal shear forces
to the moment of the resultant gives
Figure 8.80 Semicircular beam cross eV R = M
sections.
and noting that the magnitude of VR = V, gives
4
e= r 8 290
π
Assuming r = 250 mm then e = 318 mm.
Unsymmetrical Bending
Previous discussions of bending stresses in beams assumed that applied loads were applied in the
planes of principal axis of inertia. Beam cross sections were usually symmetrical, and loads were
applied along an axis of symmetry. There are cases in design where loads are not applied in planes
of symmetry such as shown in Figure 8.81, producing unsymmetrical bending. Unsymmetrical
bending is defined as bending caused by applying loads that do not lie in or are parallel to principal
centroidal axis of inertia [23]. Principal centroidal axis are designated by the U, V axis. The external
load, F, may or may not pass through the centroid of the cross section. If the is thin walled, it is
assumed that the external passed through the shear center.
Principal Axis of Inertia

Area moment of inertia is a property of any cross section of a beam. In our discussion of symmet-
rical bending, moments of inertia were obtained with respect to the neutral or centroid axis of the
cross section. This axis is the location of zero bending and bending stresses increase linearly from
the neutral axis. This is also the case in unsymmetrical bending, but the neutral axis is not as
obvious.
Consider the transformation of the moments of inertia from x,y axis to x ,y axis (refer to
Figure 8.82)
The radial distance from point 0 to differential area dA is
r = xi + yj = x 1 α + y1 β 8 291
Note that
Figure 8.81 Unsymmetrical bending. V

V F
U U
F
ϕ
Figure 8.82 Transformation axes for area moments of y

inertia. y'
ĵ dA
x'
θ
x
î
α = i cos θ + j sin θ
β = − i sin θ + j cos θ
i = α cos θ − β sin θ
j = α sin θ + β cos θ
Substituting for i and j and collecting terms

x 1 = x cos θ + y sin θ 8 292
y1 = − x sin θ + y cos θ 8 293
By definition
I x1 = y21 dA, I y1 = x 21 dA, I x1y1 = x 1 y1 dA 8 294

A A A
Substituting for y1 gives
I n = I x cos 2 θ + I y sin 2 θ + − 2I xy sin θ cos θ 8 295
It is helpful to note that area moments of inertia (Ix, Iy, Ixy) are calculated with respect to given
coordinates, say x and y. Moments of inertia with respect to x y each change with orientation, θ.
There is a remarkable similarity of the transformation equations for stress, strain, and area moment
of inertia:
y Y Figure 8.83 Angle iron cross section.

V U
b
θ
X
C
y
t
x
O
σ n = σ x cos 2 θ + σ y sin 2 θ + 2τxy sin θ cos θ given by 8 33

γ xy
εn = εx cos 2 θ + εy sin 2 θ + 2 sin θ cos θ given by 8 34
2
I n = I x cos 2 θ + I y sin 2 θ + − 2I xy sin θ cos θ given by 8 295
Notice the similarity between the three equations. Each of these equations has the same form.
This means that the earlier discussion of stress (and strain) applies directly to inertia transforma-
tions. Mohr’s circle also applies and gives a convenient way determine principal axis of inertia.
Because of the minus sign attached to the product of inertia in Eq. (8.295), positive values of product
of inertia are plotted up and negative values of product of inertia are plotted down.
Consider the cross section of the angle iron shown in Figure 8.83. It is desired to find the area
moment of inertia about any axis passing through the center of gravity, point c. The approach is
to find the moments of inertia, IX, IY, IXY and then use Mohr’s inertia circle to find IN oriented from
the x-axis by angle θ.
Using moment of inertia transfer equations:
b b
x2bt = bt x=y= 8 296
2 4
2
b3 t b b3 t b3 t
Ix = + bt = Iy = I xy = 0 8 297
12 2 3 3
2
b3 t b 5b3 t
I x = I X + y2 A I X = − 2bt = 8 298
3 4 24
5b3 t
I y = I Y + x2A I Y = 8 299
24
b b b3 t
I xy = I XY + xyA I XY = − 2bt = − 8 300
4 4 8
Example Assuming b = 4 in. and t = ¼ in.

I##, in.4 Y axis
Iv = 1.33 Iu = 5.33
3.33
I#, in4
2θ
–2
X axis
Figure 8.84 Mohr’s circle for inertia.
I X = I Y = 3 33 in 4
I XY = − 2 in 4
Using Mohr’s circle for area moment of inertia (Figure 8.84),
I u = 5 33 in 4 θ = 45
I v = 1 33 in 4
Example Consider another cross-sectional area shown in Figure 8.85. Here we wish to determine
the principal axis of inertia and principal area moments of inertia with respect to these axes.
Cross product of inertia
1
I xy 1
=4 − 2 4 = − 8 in 4
4
I xy 2
=0
1
I xy 3
=4 2 −4 = −8
4
Area moment of inertia with respect to the x axis:
3
1 1 1 1
Ix 1 =4 + 42 4 = + 16 16 in 2
4 12 4 16 12
1 7 53
Ix 2 = = 8 79 in 4
4 12
Ix 3 = 16 in 4
Area moment of inertia with respect to the y axis:

1 43 1 16
Iy 1
= + 22 4 = + 4 = 5 33 in 4
4 12 4 12
U y
8” 23.4°
x
¼”
4”
Figure 8.85 Principal Axis of Inertia.
Iy 2
0
Iy 3
= 5 33 in 4
I x = 40 79 41 in 4
I y = 10 67 11 in 4
I xy = − 16 in 4 plotted down
Using Mohr’s circle of inertia (Figure 8.86):

16
tan 2θ =
15
2θ = 46 8 θ = 23 4
I V = 26 + 22 = 48 in 4
I U = 26 − 22 = 4 in 4
Neutral Axis of Bending

When a force, P, is applied at angle ϕ measured from a principal axis of inertia, the neutral axis of
bending is determined by
Iu
tan α = tan ϕ 8 301
Iv
is measured from a principal axis of inertia
α is measured from the other principal axis of inertia
Force P does not have to be applied through the centroid, C; however, ϕ is still measured from a
principal axis of inertia.
Example Consider the symmetrical beam cross section in Figure 8.87.
We wish to determine the orientation of the neutral axis of bending.
The location of the centroid the composite area with respect to the baseline is
I##
22
y axis
48
26 41
I#
11 2θ
4
16
x axis
Figure 8.86 Mohr’s circle of inertia (numbers have been rounded).
y,V
1”
N 1
ϕ = 30°
4”
F
x,U
α
y = 1.75”
2 1”
N
4”
Figure 8.87 Neutral axis of bending.
1
4+4 y= 4+3 4
2
7
y= = 1 75 in
4
The moment of inertia with respect to the x axis (passing through cg) is
1 43
Ix 1 = + 1 25 2 4 = 11 58 in 4
12
4 13
Ix 2 = + 1 25 2 4 = 6 58 in 4
12
Ix = Ix 1 + Ix 2 = 18 17 in 4
413 1 3
Iy = + 4 = 5 67 in 4
12 12
So
7
y= = 1 75 in
4
I x = 1 817 in 4 Ix = IU
4
I y = 5 66 in Iy = IV
I xy = 0 y, V are axis of symmetry
18 17
tan α = tan 30
5 66
α = 61 64
Bending Stresses
Normal stresses caused by unsymmetrical bending are determined from [23]
Muv Mvu
σ= + 8 302
Iu Iv
where
Iu, Iv − moments of internal with respect to principal axis U and V

u, v − coordinates to point where stress is to be determined
Mu = M cos ϕ
Mv = M sin ϕ
Example These equations will now be applied to the unsymmetrical beam shown in Figure 8.88.
The bending moment diagram shows a maximum bending of 3200 ft lb along the middle portion.
The bending components about the principal axes are shown below:
1 3 1 3
Iu = bh = 68 = 256 in 4
12 12
8 63
Iv = = 144 in 4
12
Neutral axis is oriented per
256
tan α = tan 30 = 1 028
144
or α = 45.79 from the U principal axis.
The bending moment, M, is resolved into components about the U and V axis.
M u = 3200 12 cos 30 = 33 300 in -lb
M u = 3200 12 sin 30 = 19 200 in -lb
No sign has been attached to either of these components. It is somewhat simpler to observe that
both bending components induce compressive stresses at point A.
30° F
M A
30°
U 8 in.
𝛼
6 in.
16 ft
800 lb
4 ft
V 800
M
3200 ft-lb
x
Figure 8.88 Shear and bending moment diagrams.
Substituting the numbers gives
33 000 4 19 200 3
σA = + = 916 psi compression
256 144
The neutral axis angle, α, does not enter into this calculation directly.
Beams on Elastic Foundations
This analysis applies to long beams support by elastic foundations, such as a railroad track
lying across a series of railroad ties. Winkler [9] developed the mathematics to this problem. This
work was motivated by the need to understand bending stresses in railroads during the growth of
steam locomotion in Europe. Loading in the beam can be concentrated or distributed as in regu-
lar beams.
Formulating the Problem

The equations of bending are developed with reference to the freebody diagram shown in
Figure 8.89.
The shear forces and moments are shown in their positive directions. The elastic foundation force
is q dx with q being the foundation force per unit length, dx. From this diagram
dM
=V 8 303
dx
dV
=q 8 304
dx
d2 y
EI = −M 8 305
dx 2
Combining these equations gives
d4 y
EI = −q 8 306
dx 4
The magnitude of force, q (force/length) depends on deflection, y.
q = ky 8 307
where
k – distributed spring constant (force/length of beam per depth of deflection, force/length2)
One way to experimentally determine k of a uniform support is to load a flat surface onto the
uniform support with force, F, and measure displacement into support. The equation of equilib-
rium is
F = qΔx = kyΔx
F pΔxb pb
k= = = lb in 2 8 308
Δxy Δxy y
where
F – applied force
p – pressure applied to the top surface, psi
b – width of surface
Δx – incremental length
y – depth of penetration into surface
Δx Figure 8.89 Differential element of beam on

elastic foundation.
x
M M + dM
y
V V + dV
q Δz
Another type of support is analogous to railroad tracks lying on ties. Assuming tie spacing is
ℓ then
pℓb = Ky
where K is the spring constant of each individual tie being pushed into the surface. Since q = bp,
ℓq = Ky and
K
q= y 8 309
ℓ
So
K
k eq =
ℓ
By substitution for q in Eq. (8.306)
d4 y
EI = − ky 8 310
dx 4
d4 y
EI + ky = 0 8 311
dx 4
d4 y
+ 4β4 y = 0 8 312
dx 4
where
k
β4 =
4EI
Mathematical Solution
The solution to Eq. (8.312) is assumed to be y(x) = Cesx. From substitution, there are four values of
“s” that satisfy this equation:
s4 + 4β4 = 0 8 313
s4 = − 4β4
s4 = 4β4 − 1 8 314
In the complex plane eiπ = − 1 therefore

s4 = 4β4 eiπ
But since we need at least four roots
s4 = 4β4 einπ
where n = 1, 3, 5, 7. Therefore
nπ
s= 2βei 4 n = 1, 3, 5, 7 8 315
These four roots are expressed in terms of complex coordinates as
s1 = β 1 + i
s2 = β − 1 + i
s3 = β − 1 − i
s4 = β 1 − i
All four values of “s” satisfy Eq. (8.312) and therefore the sum of each is also a solution. Therefore,
y x = C1 eβ 1 + i x + C 2 eβ −1 + i x
+ C 3 eβ −1−i x
+ C4 eβ 1 − i x
Expanding further
y x = eβx C1 eiβx + C 4 e − iβx + e − βx C2 eiβx + C3 e − iβx 8 316
Recall
eiβx = cos βx + i sin βx

e − iβx = cos βx − i sin βx
By substitution, the terms in the first and second brackets are
C 1 eiβx + C 4 e − iβx = C 1 + C4 cos βx + i C1 − C4 sin βx 8 317

and
C2 eiβx + C 3 e − iβx = C 2 + C3 cos βx + i C2 − C3 sin βx 8 318

Since y(x) is a real number, C1 and C4 must be complex conjugates as well as C2 and C3. Bringing
all this together gives the general solution
y x = eβx A cos βx + B sin βx + e − βx C cos βx + D sin βx 8 319
Solution to Concentrated Force

The general solution is now applied to a beam loaded by a concentrated force. As in the case of open
beams, the general solution does not apply across the concentrated force, so we look for the solution
to the right side of load P. With respect to the x, y coordinates (Figure 8.90), positive values of x
predict huge displacement at +∞, therefor constants A and B must be zero, leaving
y x = e − βx C cos βx + D sin βx
The problem is to determine the deflection, slope, and internal shear and moments caused by the
concentrated force, P. The origin of the x axis is located at the point of application of the concen-
trated force. Beam deflection is measured positive in the downward direction.
The remaining equation has two arbitrary constants, C and D. These constants are determined by
using the slope and shear at x = 0.
dy
=0
dz x=0
P y(x)
Figure 8.90 Deflection of a beam on an elastic foundation.

d3 y P
EI =V = −
dx 3 x=0 2
Using the slope at x = 0, C = D leaving
y x = Ae − βx sin βx + cos βx A is arbitrary 8 320

The shear condition at x = 0 gives
Pβ
A=
2k
Collecting terms
Pβ − βx
yx = e cos βx + sin βx 8 321
2k
This equation shows that displacements diminish as x increases to (+) infinity. Keep in mind that
Eq. (8.321) was developed for displacement to the right of the concentrated load, P. It does not apply
across the concentrated load, P, but we can apply it to the left side knowing that displacements are
symmetrical about loading P.
Other parameters of interest are slope, shear, and bending moment along the beam. These para-
meters are determined from
Pβ
y= F1 deflection 8 322
2k
Pβ2
y = − F 2 slope 8 323
k
P
EIy = − F 3 moment 8 324
4β
P
EIy = F4 shear 8 325
2
where
F 1 = e − βx cos βx + sin βx 8 326

− βx
F2 = e sin βx 8 327
F 3 = e − βx cos βx − sin βx 8 328
− βx
F4 = e cos βx 8 329
In each case, displacement, slope, shear, and bending are symmetrical on each side of load P. If
there are multiple concentrated loads applied to the beam, local parameters are determined by
superposition. Other loading conditions are given.
Radial Deflection of Thin Wall Cylinders Due to Ring Loading

One of the most powerful applications of this theory is predicting bending and circumferential
stresses in thin wall cylinders resulting from uniform circumferential loads. In this case, radial dis-
placement, u, is the dependent variable and the distributed “spring” effect comes from the elastic
circumferential strain.
Formulation of Spring Constant

From earlier discussions, hoop stress is related to external pressure surrounding a thin wall cylin-
der by
a
σθ = p 8 330
h
where
h – wall thickness
a – mean radius of cylinder
p – applied pressure (internal or external)
When external pressure is applied, thin wall cylinders shrink by radial displacement, u, where
(+) displacement is in the inward direction (Figure 8.91). The hoop strain across the wall (assumed
to be uniform) is
2π a − u − 2πa u
εθ = = − minus means compression 8 331
2πaa a
Noting that σ θ = Eεθ, then
a u
p=E 8 332
h a
then
a2
u= p 8 333
hE
As explained previously, the spring constant for elastic foundations is force per length of beam per
deflection into the elastic foundation (k = lb/in. per inch of deflection with units of force/length2).
For the cylindrical strip (Figure 8.92), the elastic compression of the cylinder acts as the elastic foun-
dation. The formulation of the expression for k is given below.
Consider a small arc of b = aΔθ and length L. When external pressure, p, is applied, the distrib-
uted load, w, is
2a
σθ
ΔL p
σθ
h
Figure 8.91 Compressive hoop stress due to external pressure.

b x
p
t
u
t
Δθ
a
L
Figure 8.92 Arc section of a thin-walled cylinder.
wL = pbL 8 334
where
w – distributed load, lb/in.

L – length of strip, in.
p – pressure, psi
b – arc length
The distributed load produces an inward deflection of

a2
u= p 8 335
hE
The spring constant, k, is defined as
w
k= lb in 2 8 336
u
By substitution
bhE hE
k= or k= per unit arc length 8 337
a2 a2
The arc distance, b, is a small arbitrary distance and cancels out later.
Equation of Bending for Cylindrical Arc Strip

The above formulation assumes zero bending and was developed for the sole purpose of developing
the expression for the spring constant, k. We now examine deflections caused by nonuniform exter-
nal loading, i.e. w constant.
The equation of bending for the strip beam (Figure 8.92) is developed from the differential ele-
ment shown in Figure 8.93. Recall from beam the
d2 u
M t = EI t 8 338
dx 2
t
b
Mt h σθ
u a
σx
t Δθ
σθ
a
Figure 8.93 Element of cylindrical surface.
where
Mt – bending moment per length, b, about the t axis

It – moment of inertia per length, b, with respect to the t axis
In this case, EIt is modified as follows. Each of the two sides of the strip beam lies in radial planes
and hoop strains are not allowed when only bending is considered. Because of this,
σx σθ
εx = −ν 8 339
E E
σθ σx
εθ = −ν =0 8 340
E E
From the second equation:
σ θ = νσ x 8 341
Therefore, from the first equation:
E E z E d2 u
σx = εx = = − z 8 342
1 − ν2 1 − ν2 ρ 1 − ν2 dx 2
where z is distance from neutral axis of bending of the cylinder wall.
The internal moment is
h
2
Mt = σ x z dz internal moment per unit arc length

− h2
By substitution and integration:
Eh3 d2 u d2 u
Mt = = D 8 343
12 1 − ν2 dx 2 dx 2
Eh3
D= 8 344
12 1 − ν2
Applying
dM dV
=V and =q
dx dx
gives
d2 M
=q 8 345
dx 2
Turning now to the equation of bending in terms of cylinder dimension,
d4 u hE
D + 2u=0 8 346
dx 4 a
d4 u bhE
D + 2 2u=0 8 347
dx 4 ha
d4 u 12 1 − ν2
4 + u=0 8 348
dx h 2 a2
d4 u
+ 4β4 u = 0 8 349
dx 4
where
3 1 − ν2
β4 =
a2 h 2
While Eq. (8.349) has the same mathematical form as Eq. (8.312), the mathematical expressions
for β are different. All the previous examples of different loading apply directly to thin-walled cylin-
ders, where loading is radial and symmetrical around the centroid axis of the cylinder as indicated
in Figure 8.94. The circumferential loading (P) is analogous to a point load on a beam except in the
pipe case P has units of force per circumferential length. The general solution is the same as in the
previous case.
See Timoshenko and Woinowsky-Krieger [5] and Den Hartog [13] for a variety of other applica-
tions of this theory.
P
u x
Figure 8.94 Thin cylinder with ring load (strap).

Reach of Bending Moment

The stiffness of a strip beam along the length of a thin-walled cylinder is typically much higher than
for ordinary beams on an elastic foundation. As a result, wall deflections and bending stresses occur
over a rather short distance away from the load. Both decay at a rate dictated by e−βx. The value of
this term becomes negligible when βL = 3π2 . Assuming ν = 0.3, the minimum length of the deflected
surface is
3π 3 1 − ν2
L= , where β4 =
2β a2 h 2
By substitution and assuming ν = 0.3
3π h
L= = 3 67a 8 350
2β a
1
which gives L = 0.82a for ha = 20 . If cylinder radius is a = 8 in., then the span of significant bending
away from the load is 6.56 in. This observation is borne out by the following example.
Bending Stress in Wall of a Multi Banded Cylinder
Example Consider a situation where three collars are shrunk onto a cylinder as shown in
Figure 8.95. Strain gauges on each collar indicate circumferential strains in each collar as
900μ collar#1
500μ collar#2
1000μ collar#3
Each collar has a rectangular cross section of ¼ in. thick by 1 in. wide. Dimensions of the cylinder
are given in the figure. Question: What is the bending stress in the cylinder wall directly under col-
lar #2?
Stress in each strap is relate to circumferential strain by
σ θ = εθ E 8 351
0.25
#1 #2 #3
18”
4” 4”
Figure 8.95 Local bending in strapped cylinder.

Contact pressure is related to stress by

r
p = εθ E 8 352
t
r
p = εθ E
t
Radial force per circumferential length is
t
P = pw = εθ Ew lb in w is width of each band is 1 in
r
where w is width of the strap.
0 25
P= 30 × 106 1 εθ = 0 833 × 106 εθ lb in
9
Using this formula, the force per circumferential length is
Collar#1 P1 = 750 lb in
The bending moment distribution caused by one collar is
P P − βx
Mt = − F3 = − e cos βx − sin βx in -lb in 8 353
4β 4β
P
Mt = C βx in -lb in 8 354
4β
where
3 1 − ν2
β4 =
r 2 t2
3 1 − 0 252
β4 = = 0 5556
92 0 252
Therefore, β = 0.8633 1/in. A plot of F3 is shown in Figure 8.96. Note, Cβx = − F3.
The active distance of bending is
h 0 25 1
= =
a 9 36
4 714 h 1
L= a = 3 626a = 0 604a
13 a 36
L = 0 604 9 = 5 44 in
This number is consistent with Figure 8.96.
The total moment under strapped #2 is, by superposition
P1 P2 P3
M total = Cβ4 + C β0 + C β4 8 355
4β 4β 4β
1
M total = P1 + P3 Cβ4 + P2 Cβ0
4β
C β0 = − 1 Cβ4 = + 0 0206 4β = 4 0 8569 = 3 43
1.2
0.8
F3, dimensionless
0.6
0.4
0.2
0
0 1 2 3 4 5 6 7 8
–0.2
–0.4
x distance, in.
Figure 8.96 Distribution of bending from plane of application.
Substituting variables
1
M total = 750 + 833 0 0206 + 417 − 1 = − 112 in -lb per inch of circumference
3 43
Notice the short reach of the bending moment (F3 function).
Mx c
σ bending =
I
where
c = 0.125 in.
253
I = 0 12 = 0 0013 in 3 per inch circumference
Giving the total bending stress due to all three straps

112 0 125
σ bending = = 10 770 psi
0 0013
Criteria of Failure
Classic formulas are used to predict stress magnitudes associated with several different design con-
figurations and loads. A design may be subjected to different combinations of these loads and stres-
ses. In this case, local stress at a given point is the superposition of each stress type assuming stress
levels are within the elastic limit. It is important to visualize each of these stress types and how to
combine them. The combined stresses identify σ xσ yτxy in critical areas of a design. This information
can then be used to determine principal stresses as previously discussed.
Combined Stresses
The stress condition at any point (biaxial or tri axial) in a design can be the combined effects of
simple stress, direct pull, bending, torsion, pressure, etc.
Example Consider the case where pipe is loaded by torsional shear, internal pressure plus bend-
ing. Using the following pipe data and loads:
•• Pipe is 4-in. OD and 3.75 ID

Internal pressure is 1000 psi
•• Applied torque is 2000 ft lb

Bending moment is 5000 ft lb
• Material yield is 60 000 psi
The challenge here is to determine the factor of safety against yielding. First, calculate the three
stress components.
Moments of inertia are needed to determine torsion shear and bending moment stresses:
π 4 π 4
I= r − r 4i = 2 − 1 8754 = 2 8591 in 4
4 0 4
J = 2I = 5 7183 in 4
A = π r 20 − r 2i = π 22 − 1 8752 = 1 5217 in 2
Internal Pressure
t 0 125 r0
= = 0 0625 or = 16 10
r0 2 t
The pipe can be considered a thin-walled cylinder:
2 + 1 875
r mean = = 1 94
2
r mean
= 15 5
0 125
r mean
σθ = p = 15 5 1000 = 15 500 psi
t
Applied Torque
Tc 2000 12 2
τ= = = 8394 psi
J 5 7183
Bending Moment
5000 12 2
σb = = 41 971 psi
2 8591
Assuming the pipe is support axially so that the internal pressure does not create axial stress, the
state of stress is
σ x = σ b = 41 971 psi
σ y = σ θ = 15 500 psi
τxy = 8394 psi
τ
τmax
y axis
σ1
2θ σx
σ
σy
σ2
τxy
x axis
Figure 8.97 Mohr’s circle for the example.
These stresses are shown in Mohr’s circle (Figure 8.97), which also indicates principal stresses
and maximum shear for the combined state of stress:
σ p1 = 44 409 psi
σ p2 = 13 077 psi
τmax = 15 666 psi
The orientation of the principal axis from the x axis is θ = 11.2 counterclockwise.
The combined effect of three different types of loading (internal pressure, bending, and torsion)
produced a biaxial state of stress having principal stress as shown.
Failure of Ductile Materials

Many stress conditions, such as bending, truss analysis and simple tension or shear, allow the direct
application of uniaxial stress data to determine load limits and determine factors of safety. How-
ever, many loading situations produce combinations of stresses which are superimposed to produce
two or three-dimensional states of stress. Also, two and three-dimensional states of stress also may
be produced simply due to complex geometry and these areas are often critical.
The maximum normal stress criteria states that material yielding occurs when the maximum prin-
cipal stress equals the yield strength as determined from uniaxial test data. Experience shows that
this theory is not accurate. An example is creatures that live in the deepest oceans. They experience
extreme hydrostatic pressure but survive.
As per the maximum shear stress criteria, material yielding occurs when the maximum shear
stress at a point is greater than the maximum shear yield of the material:
τmax ≥ τyld 8 356
Since the maximum shear stress at yield (per uniaxial test data) is ½ axial yield then
τyld = 0 5σ yld 8 357
and yielding occurs in a biaxial or tri-axial state of stress when

τmax ≥ τyld = 0 5σ yld 8 358
This criterion is close to reality.

Von Mises’ criteria of energy distortion more accurate and is the one most commonly used to
determine material yielding. It states that material yielding occurs when the energy of distortion
in a biaxial or triaxial stress state is equal to or greater than the energy of distortion of in a uniaxial
test at yielding [23]. This criterion leads to the following set of equations commonly used in design.
2 2 2
σ1 − σ2 + σ2 − σ3 + σ3 − σ1 ≥ 2σ 2yld 8 359
When the stress state is biaxial (σ 3 = 0), the criteria of failure can be expressed as
σ 21 − σ 1 σ 2 + σ 22 ≥ σ 2yld 8 360
Letting σ represent von Mises stress
σ = σ 21 − σ 1 σ 2 + σ 22 8 361
Then yielding occurs when

σ ≥ σ yld 8 362
This equation is often represented by ellipse on a stress diagram. The von Mises criteria is typ-
ically used to establish shear yield from normal yield strength determined from uniaxial test data.
Consider a state of pure shear. The Mohr’s circle has a radius of τ giving, σ 1 = τ and σ 2 = − τ. By
substitution into Eq. (8.361)
σ = 3τ 8 363
When yielding occurs
3τyld = σ yld 8 364
Therefore,
τyld = 0 577σ yld 8 365
This formula is commonly used to establish shear yield strength. It is rarely a measured value.
Appling the von Mises criteria to the previous example gives
1
σ = 44 412 − 44 41 13 08 + 13 082 2
= 39 53 ksi
The factor of safety for this stress state is

60
FS = = 1 518
39 53
Visualization of Stress at a Point
Example A pipeline is subjected to the following loads (Figure 8.98):
•• Internal fluid pressure of 150 000 psi

Bending moment of 4000 ft lb
• Torque of 6000 ft lb
pi a
M T
5 in. OD
4 in. ID
Figure 8.98 State of stress in pipe.
Draw the stress element for point “a” and show the stress components on the element caused by
the applied loads:
π r 4o − r 4i π
I= = 2 54 − 24 = 18 11 in 4
4 4
J = 2I = 36 22 in 4
It is important to visualize stress components at any given point in an engineering structure.
A stress situation is shown in the figure associated resulting from three types of loading: internal
pressure, bending, and torque. The magnitude of stress from each load can be determined from
classic formulas for each load:
Mc 4000 12 2 5
σB = = = 6626 psi bending
I 18 11
R 25
σθ = p = 1500 = 7500 psi pressure
t 05
Tc 6000 12 2 5
τ= = = 4970 psi torque
J 36 22
The stress element at point “a” is shown in Figure 8.99.
Principal stresses and strains can be determined from this element.
Pressure Required to Yield a Cylindrical

Vessel
Another application of von Mises criterion of
failure is the determination of pressure limit on
σθ
a given pressure vessels. Consider a cylinder ves-
sel having an OD of 5.625 in. and a wall thickness
of 0.04 in. (18-gauge sheet metal). The mean
diameter is 5.587 in. The ends are capped with
σB σB spherical shape caps having the same wall thick-
ness as the cylinder. Assume a yield strength of
σ yld = 60 000 psi and Poisson’s ratio of ν = 0.29.
τ Stress components in the body of the cylin-
der are
σθ r
σ z = p axial stress 8 366
2t
Figure 8.99 Stress element at point “a.”
r
σθ = p hoop stress 8 367
t
Substituting these expressions into Eq. (8.361) yields a von Mises stress of
r
σ = 0 866 p 8 368
t
Equating to the yield strength of 60 000 psi gives a maximum allowable pressure of 9921.5 psi. The
internal pressure also creates a third normal stress component, σ r, at the inside surface, which is
usually relative small by comparison.
The corresponding strains, εz, εθ, are determined Hooke’s equations for biaxial stress.
Failure of Brittle Materials

Rock formations in the earth are subjected to complex biaxial or triaxial states of stress. Formation
rocks are subjected to overburden as well as horizontal loads. When a well bore penetrates through
rock formations, well bore pressure is applied to rock as well as overburden. Formation fluid pres-
sures further complicate the analysis of rock stresses.
Commonly used criteria of failure for brittle materials are
•• Maximum normal stress theory

Coulomb–Mohr theory
• Modified Mohr theory
Each of these theories is illustrated in Figure 8.100.

Note
Sut – ultimate tension strength as determined from uniaxial tests
Suc – ultimate compressive strength as determined from uniaxial tests
The maximum normal stress criteria of failure theory is adequate for normal stress conditions in
the first quadrant. In cases where there is a combination of tension and compression (second and
σB
Maximum normal stress
Sut
σA
σB
Coulomb–Mohr theory σA = –1
Modified Mohr theory
Hydrostatic stress
Suc
Figure 8.100 Criteria of failure of brittle materials.

fourth quadrant) the other two criteria are more appropriate. Consider the biaxial stress state for the
fourth quadrant. The Coulomb–Mohr criteria is shown as a line between Sut, ultimate strength in
tension and Suc, the ultimate strength in compression.
τ = c + σ tan ψ 8 369
where
τ − maximum shear stress at rupture

c – cohesive strength of rock
σ − compressive stress normal to the plane of rupture (compression is a + number)
ψ − angle of internal friction
If principal stresses are arranged so that σ 1 σ 2 σ 3, then the critical stress limits are σ A and σ B.
These two applied stresses are related to the two strengths by
σA σB
− =1 8 370
Sut Suc
In this equation, Suc is a positive number. For example, using numbers for sandstone:
Suc = 20 000 psi, Suc = 500 psi
σA
σ B = Suc −1 8 371
Sut
This equation predicts failure in the fourth quadrant.
σA
σ B = 20 000 −1 8 372
500
At σ A = 500, σ B = 0 and at σ B = − 20 000, σ A = 0. A biaxial state of stress (σ 1 and σ 3) located within
this line does not produce failure; however, a biaxial state of stress outside this line does produce
failure per this criterion.
Example Consider a design where Gray cast iron is used. Local stress in a critical area is defined
by 40 ksi (compression) and 5 ksi (tension) (Figure 8.101). Determine the factor of safety against
static failure using the modified Mohr criteria of failure:
σB σ uc = 110 ksi
20
σA σ ut = 20 ksi
The factor of safety is the ratio of the line of intersection with the
20,20
line of stress:
5,40 d= 52 + 402 = 40 3
D= 8 812 + 70 512 = 71 06
Factor of safety is
8.81,70.51
71 06
FS = = 1 76
40 3
110
Experimental data show that actual points of failure fall outside
Figure 8.101 Factor of safety of the Coulomb–Mohr criteria.
for gray cast iron application.
References 489
σB
One point is the case, where σA = − 1. An example is the twisting of classroom chalk which
causes failure on a 45 plane. Under this torsion loading, the principal normal stresses are equal
but opposite and oriented on planes 45 from the axis of twist. The modified Mohr line is drawn
from this point. The data of Coffin [25] and Grassi and Cornet [26] fall along this line. The modified
Mohr criteria, then, appears to be more accurate than the Coulomb–Mohr criteria.
Mode of Failure in Third Quadrant

The mode of rock fracture due to compressive principal stresses (third quadrant) is different. The
intrinsic line of failure is indicated in Figure 8.101. The Maximum Normal Stress Criteria of Failure
does not apply in this case. Consider the case where all principal normal stresses are equal. Such a
condition is indicated by the corner point identified by “hydrostatic stress.” Experience shows that
hydrostatic stress alone does not cause failure.
References
1 Ressler, S. (2011). Understanding the World’s Greatest Structures, DVD Video, The Great Courses.
2 Mohr, O. (1882). Civil ingenieur, p. 113; also, see Mohr, O. (1906). Abhandlungen aus dem Gebiete
der technischen Mechanik, Berlin, p. 219.
3 Perry, C.C. and Lissner, H.R. (1955). The Strain Gage Primer. McGraw – Hill Book Company, Inc.
4 Dareing, D.W. (2012). Mechanics of Drillstrings and Marine Risers. ASME Press.
5 Timoshenko, S. and Woinowsky-Krieger, S. (1959). Theory of Plates and Shells, 2e, 466. New York:
McGraw-Hill.
6 Huang, T., Dareing, D.W., and Beran, W.T. (1980). Bending of tubular bundles attached to marine
risers. Trans. ASME J. Energy Resour. Technol. 102: 24–29.
7 Dareing, D.W. and Ahlers, C.A. (1990). Tubular bending and pull out forces in high curvature well
bores. ASME J. Energy Resour. Technol. 112: 84–89.
8 Rocheleau, D.N. and Dareing, D.W. (1992). Effect of drag forces on bit weight in high curvature well
bores. Trans. ASME J. Energy Resour. Technol. 114 (3): 175–180.
9 Timoshenko, S.P. (1953). History of Strength of Materials. New York: McGraw-Hill; Also, see, Winkler,
E. (1867). Die Lehre von der Elastizitat und Feitigheit, Prague, p 182.
10 (1989). Manual of Steel Construction, 9e. NY: American Institute of Steel Construction.
11 (1986). Specifications for Aluminum Structures. Washington, DC: Aluminum Association Inc.
12 American Institute of Timber Construction (1985). Timber Construction Manual. NY: Wiley.
13 Den Hartog, J.P. (1952). Advanced Strength of Materials. NY: McGraw – Hill.
14 Timoshenko, S. and Gere, J.M. (1961). Theory of Elastic Stability, 2e. New York: McGraw-Hill.
16 Lubinski, A. (1950). A study of the buckling of rotary drilling string. API Drill. Prod. Prac. 17:
178–214.
17 Huang, T. and Dareing, D.W. (1968). Buckling and lateral vibration of drill pipe. Trans. ASME J. Eng.
Ind. 90, Series B (4): 613–619.
18 Huang, T. and Dareing, D.W. (1966). Predicting the stability of long vertical pipe transmitting torque
in a viscous media. Trans. ASME, J. Eng. Ind. 88, Series B (2): 191–200.
19 Lame’ and Clapeyron: “Memoire sur l’equilibre interieur des corps solides homogenes; Memoirs
presents par dever sasvans” (1833) 4 (also see Timoshenko and Goodier, Theory of Elasticity)
21 Bresse, M. (1866). Cours de mecanique applique, Paris, Part I, p. 334 (also see Ref. 22).
22 Golovin, Trans. Inst. Tech. St. Petersburg, 1881 (also see Timoshenko and Goodier, Theory of
Elasticity)
23 Seely, F.B. and Smith, J.O. (1956). Advanced Mechanics of Materials, 2e. New York: McGraw-Hill.
24 Riley, W., Sturges, L., and Morris, D. Mechanics of Materials, 6e. Wiley.
25 Coffin, L.F. (1950). The flow and fracture of brittle materials. Trans. ASME J. Appl. Mech. 17: 233–248.
26 Grassi, R.C. and Cornet, I. (1949). Fracture of grey cast iron tubes under biaxial stress. Trans. ASME J.
Appl. Mech. 71: 178–182.
491
Modal Analysis of Mechanical Vibrations
Mechanical vibrations became an important design consideration during the early 1900s when
machinery began to move at higher and higher speeds. Initially, mechanical vibrations were mod-
eled as quasi-static events, i.e. structural responses were determined by treating dynamic forces as
static even though they varied with time. The quasi-static approach is a practical and useful design
tool provided the forcing frequency is much less than the fundamental frequency of the structure as
discussed in Chapter 2. Fatigue analyses of many offshore structures are based on this assumption
as ocean wave frequencies are much lower than natural frequencies of the structure. However, if
the driving frequency is close to a natural frequency of a structure, the mass of the structure plays a
significant role in dynamic response.
Vibrations, in general, are not good for machinery. They can produce noise and damage machine
components through low-cycle or high-cycle fatigue. It is good practice to determine natural fre-
quencies during the final phase of a design, and compare them with frequencies anticipated from
engines, pumps, etc. Forcing frequencies are usually related to rotary speed. The final check on
resonance comes from testing.
Many vibration problems can be solved by applying the fundamental results of single degree of
freedom (SDOF) models. However, high-speed machinery may require consideration of higher
modes of vibration. In these cases, modal analysis is most useful. This method develops solutions
in terms of each vibration mode, thus making it easy to visualize the overall behavior in terms of the
contribution of each mode.
Complex Variable Approach
The complex frequency response is a convenient way to solve differential equations of motion. It
will be used throughout this chapter to explain vibration response of multiple degree of freedom
systems. The complex variable approach will be first applied to the SDOF problem. In this case,
the forcing function (F0 cos ωt) is replaced with the imaginary function, F0eiωt.
d2 x dx
m 2 +c + kx = F 0 eiωt 91
dt dt
The right side of the equation is shown below:
d2 x dx
m +c + kx = F 0 cos ωt + i sin ωt 92
dt 2 dt
492 Modal Analysis of Mechanical Vibrations
The complex force is introduced to expedite the solution.

Equation (9.2) is also written as
F 0 iωt
x + 2ζωn x + ω2n x = ω2n e 93
k
where
c
ς=
ccr
ccr = 2 km
The real part of the complex solution to Eq. (9.3) is the solution. The solution to Eq. (9.3) is
assumed to be of the form
x t = Xeiωt 94
where x(t) is the complex response.
By substitution, Eq. (9.4) satisfies Eq. (9.1) provided
δst
X= 95
1 − r 2 + i 2ζr
ω
r=
ωn
Therefore,
δst
x t = eiωt 96
1 − r2 + i 2ζr
which can also be written
1 − r 2 − i2ζr
x t = δst eiωt 97
1 − r 2 2 + 2ζr 2
where δst = F0
k. The solution can also be written as
x t = Xeiωt = Xe − iϕ eiωt = Xei ωt − ϕ 98
which expands into

x t = X cos ωt − ϕ + i sin ωt − ϕ 99
The solution we seek is the real part of the complex displacement function:
x t = X cos ωt − ϕ 9 10
The complex amplitude, X, is conveniently represented in a Nyquist plot (Figure 9.1). The real
and imaginary parts are identified in the complex plane.
It follows that
X 1 − r2
Re = 9 11
δst 1 − r 2 2 + 2ζr 2
X − 2ζr
Im = 9 12
δst 1 − r 2 2 + 2ζr 2
Giving Re[X]
δst Re
X= 1 9 13
1 − r2 2
+ 2ζr 2 2
ϕ
and
Im[X]
2ζr
tan ϕ = 9 14 X
1 − r2
These expressions agree with earlier discussions of
SDOF systems in Chapter 2.
Now, we have another way of expressing vibration –Im
response. Both pairs of equations give response ampli-
tude, X, and phase angle ϕ. Figure 9.1 Vibration response represented
in complex plane.
Complex Transfer Function

The complex transfer function is defined as
X 1 1 − r 2 − i 2ςr
H= = 9 15
F0 k 1 − r 2 2 + 2ςr 2
which is the complex amplitude divided by the amplitude of the applied force.
This is a useful parameter because transfer functions can be determined experimentally. The cal-
culus of the Re H function, shows that its maximum occurs at
r = 1−ζ 9 16
with a maximum value, of
1 1
Re H = 9 17
max k 4ς 1 − ς
The minimum of Re H occurs at
r=1+ζ 9 18
with a value, of
1 1
Re H = − 9 19
min k 4ς 1 + ς
The frequency ratio difference between these two points is
Δr = 2ζ 9 20
1
giving a useful measure of the damping factor, ζ. Note that when r = 0, Re H = , a useful meas-
k
ure of the stiffness parameter, k.
1
The maximum value of the Im H function occurs at r = 1 and with a magnitude of minus .
2ζk
Interpretation of Experimental Data

Consider the plot shown in Figure 9.2. Assume it is established experimentally. Steps for extracting
damping factor, natural frequency, mass, and stiffness from this data are
0.03
Real part of H, Re[H]
0.02
0.01
0
0 5 10 15 20 25 30
–0.01
–0.02
Frequency of excitation, cps
0
Imaginary part of H, Im[H]
0 5 10 15 20 25 30
–0.01
–0.02
–0.03
–0.04
Frequency of excitation, cps
Figure 9.2 Experimentally determined complex transfer function.
Step 1 – 2ζ = Δr (calculate ζ)
1
Step 2 – Determine from imaginary part, (calculate k)
2kζ
k
Step 3 – Determine natural frequency, ωn = , (calculate m)
m
c
Step 4 – Determine damping factor, ζ = , (calculate c)
ccr
Natural Frequency
f n = 15 2 cps
as read directly from Re[H].
Damping Factor
2ζ = Δr
16 5 − 14
Δr = = 0 1645
15 5
0 1645
ζ= = 0 0822
2
Spring Constant
1
= 0 038
2ζk
1
k= = 160 lb in
2 0 0822 0 038
Mass
1 k
fn =
2π m
2 k
2π 15 2 =
m
160
m= = 0 0175 lb-s2 in
9121
Thus,
W = mg = 0 0175 386 = 6 77 lb
Damping Coefficient
The damping coefficient can now be determined:
ccr = 2 km
ccr = 2 160 0 0175 = 3 35 lb ips
Since
c
ζ= = 0 0822
ccr
c = 0 0822 3 35 = 0 2751 lb ips
All basic parameters defining the SDOF system have now been quantified from the experimen-
tally generated complex transfer function. The phase angle, ϕ, lag between the driving force, F, and
displacement, x(t), can be determined for any frequency ratio, r, by Eq. (9.12).
Two Degrees of Freedom
Systems with two degrees of freedom require two dependent variables to determine its vibration
behavior. The response can be a free vibration or a forced vibration as in the case of SDOF systems.
Two degrees of freedom systems have two distinct modes or shapes of vibration with each mode
having its own natural frequency. The first step in analyzing vibrations is to determine these modal
characteristics.
Natural Frequencies and Modes of Vibration

Consider the two-mass structure shown in Figure 9.3. Damping is not a consideration in determin-
ing mode shapes and natural frequencies.
For the sake of simplicity, both masses are assumed to be the same and
both springs have the same elastic constant. The differential equations of
k motion for each mass are derived from separate freebody diagrams giving
mx 1 + 2kx 1 − kx 2 = 0 9 21
x1 m mx 2 − kx 1 + kx 2 = 0 9 22
These equations are put in matrix form for convenience:
k m 0 x1 2k −k x1 0
+ = 9 23
0 m x2 −k k x2 0
x2 m The solutions to these coupled differential equations are assumed to be of
the form:
Figure 9.3 Schematic x 1 = X 1 cos ωt 9 24
of a 2-DOF system.
x 2 = X 2 cos ω t 9 25
Upon substitution:
2k − ω2 m −k X1 0
= 9 26
−k k − ω2 m X2 0
ω2 m
Defining, ξ = k , then
2−ξ −1 X1 0
= 9 27
−1 1−ξ X2 0
This is called the amplitude equation. For a nontrivial solution,
2−ξ −1
=0 9 28
−1 1−ξ
which expands into the characteristic equation:
ξ2 − 3ξ − 1 = 0 9 29
This equation has two roots, called eigenvalues: ξ1 = 0.382 and ξ2 = 2.618 giving
k
ω1 = 0 618 first mode 9 30
m
k
ω2 = 1 618 second mode 9 31
m
By substituting the eigenvalues into the amplitude equation:

X2
= 1 618 first mode shape 9 32
X1 1
X2
= − 0 618 second mode shape 9 33
X1 2
Both mode shapes are displayed in Figure 9.4.

These results show the 2-DOF system has two natural

modes of vibration. Each mode can respond separately
or simultaneously. If the two masses are given initial
displacements defining the first mode (or second
mode), only that mode vibrates. However, in general
both modes can participate in a vibration response 1 1
simultaneously.
If the lower mass is also attached to ground through a
spring of constant k, then both mode shapes and corre-
sponding natural frequencies change to ω1 = k
m and
1.618 –0.618
ω2 = 1 732 k
m.
Figure 9.4 Modes shapes and eigenvalues.
SDOF Converted to 2-DOF
Consider the system containing a rigid bar with mass distribution (μ), a discrete mass (m) and two
springs of stiffness k1 and k2. As shown, the system is SDOF. However, if the pin connection is
removed, the system becomes 2-DOF. Equations of motion are developed for both cases for ana-
lytical comparison. The mass of the bar is considered in both cases (Figure 9.5).
Single Degree of Freedom

Only one equation is required.
M0 = I0α
k1 m k2
µ
a a a a
0 θ
SDOF
y P
cg
2-DOF
d
Figure 9.5 Conversion to 2-DOF system.

By expansion
a2 θk 1 − 2a 2 θk 2 = I b + a2 m θ
where the moment of inertia of the bar is with respect to point 0 is.
2a
2 3 2a 16 3
Ib = 2 x 2 μ dx = μx 0
= μa
3 3
0
and μ is mass per unit length of the bar.

The equation of motion for this case is
4a2 k 2 − a2 k 1
θ+ θ=0 9 34
I b + a2 m
Natural circular frequency is
4a2 k 2 − a2 k 1
ω2n = 9 35
I b + a2 m
4k 2 − k 1
ω2n =
4
M+m
3
Two Degrees of Freedom

When the pin support is removed, the system becomes a 2-DOF problem and two equations of
motion are required, a moment and a force equation. As explained in Chapter 7, there are two ways
to set up the moment equation. One approach is
M P = I cg θ + d m + μ4a acg 9 36
The second
M P = I P θ + d m + μ4a aP 9 37
The second moment equation is somewhat simpler when noting

am = d 4μa + m
am = d M + m M is mass of bar
and
I P = I 0 + a2 m
In this case, the moment equation becomes
M P = I P θ + amy
The moment equation expands into
ak 1 y − aθ − k 2 y + 2aθ 2a = I P θ + amy 9 38
The force equation is
F y = m + M acg 9 39
− k 1 y − aθ − k 2 y + 2aθ = m + M y + dθ 9 40
Two dependent variables, y and θ, are required. Rearranging terms the two equations of
motion are
I P θ + amy + k 1 + 4k 2 a2 θ − k 1 − 2k 2 ay = 0 moment
and
amθ + m + M y − k 1 − 2k 2 aθ + k1 + k 2 y = 0 force
In this case the solutions are assumed to be in the form of
y t = Y cos ωt
θ t = Θ cos ωt
The procedure for finding natural modes and mode shapes is the same as given above.
Other 2-DOF Systems

There are many configurations of 2-DOF systems. The designer must envision an appropriate
model for each situation. Typically, two dependent variables are required to describe the modes
and their corresponding natural frequencies. Figure 9.6 illustrates several 2-DOF configurations.
x1 x2
L
x1
k x2
m
k M k I
θ1 θ2
x1 x2
T T
I
m1 m2 I
k
θ
k θ2
θ1
r2
r1
M k R
m
x
Figure 9.6 Examples of 2-DOF systems.

k After setting up an appropriate math model, the

θ equations of motion are developed based on the
r2 kinetics of rigid and discrete bodies as discussed ear-
r1 lier. As an example, consider the 2-DOF system in
x1 = r1 θ Figure 9.7.
M Equations of motion are
k M = I0θ
I 0 θ = r1 x2 − r1 θ k − r2 r2 θ k disc 9 41
m
x2
F = mx 2
Figure 9.7 Combination of rotation and linear mx 2 = − x 2 − r 1 θ k discrete mass
motion.
9 42
Since θ = x1
r1 , I 0 = k2g M, kg is radius of gyration, the two equations of motion become
2 2
kg r2
M x1 + 1 + kx 1 − kx 2 = 0 9 43
r1 r1
and
mx 2 + kx 2 − kx 1 = 0 9 44
Determination of the modal matrix and characteristic equation continues as before, starting with
x 1 = X 1 cos ωt
x 2 = X 2 cos ωt
The substitution of these two equations leads to amplitude equations and eigenvalues as before.
Undamped Forced Vibrations (2 DOF)
Consider a scenario (Figure 9.8) with a periodic force, F(t), applied to the top mass. Assume the 2-
DOF system is undamped.
The differential equations of motion are
mx 1 + 2kx 1 − kx 2 = F 0 cos ωt 9 45
and
mx 2 − kx 1 + kx 2 = 0 9 46
For a solution to both equations, assume
x 1 t = X 1 cos ωt 9 47
x 2 t = X 2 cos ωt 9 48
These equations indicate both masses vibrate with the frequency of excitation (ω), but with dif-
ferent amplitudes. This frequency is not necessarily one of the natural frequencies (ω1 or ω2).
By substitution
2−ξ −1 X1 δst
= 9 49
−1 1−ξ X2 0
where
F0
δst = static displacement k
k
ωm
2
ξ= ω is driving frequency in this case
k x1 m F = F0 cos ωt
By Cramer’s rule the amplitudes of the responses are
δst −1 k
0 1−ξ
X1 = 9 50
2−ξ −1 x2 m
−1 1−ξ
2−ξ δst
Figure 9.8 Forced vibration for 2-DOF
−1 0 system.
X2 = 9 51
2−ξ −1
−1 1−ξ
The solutions to Eqs. (9.45) and (9.46) are thus defined by substituting X1 and X2 into Eqs. (9.47)
and (9.48). The amplitudes of each mass vary with the frequency of the force of excitation. The
vibration amplitude of each mass vs. driving frequency is shown in Figures 9.9 and 9.10. Resonance
occurs when the denominator is zero.
The denominator of both X1 and X2 expressions is zero when ξ = ξ1 or ξ = ξ2 per the characteristic
equation given earlier for this 2 DOF system. Large vibration amplitudes are anticipated when
ω2 m
= 0 382
k
and
ω2 m
= 2 618
k
15
10
Amplitude ratio, X1/δ
0
0 0.5 1 1.5 2 2.5 3 3.5
–5
–10
–15
Eigenvalues, ξ
Figure 9.9 Frequency response of mass 1.

15
10
Amplitude ratio, X2 / δ
0
0 0.5 1 1.5 2 2.5 3 3.5
–5
–10
–15
Eigenvalues, ξ
Figure 9.10 Frequency response of mass 2.
or
k
ω = 0 618 = ω1 frequency of first mode 9 52
m
k
ω = 1 618 = ω2 frequency of second mode 9 53
m
Either mode of vibration, defined earlier, can be excited provided the exciting frequency is tuned
to one of the natural frequencies.
Undamped Dynamic Vibration Absorber

The discussion of dynamic absorbers is an extension of the previous model with slight modifications
to the parameters [1]. The vibration model is shown in Figure 9.11.
The model depicts a base SDOF system (M and K), which is being excited by F(t); its natural fre-
quency is
K
Ωn = 9 54
M
From our previous discussion of forced vibrations of SDOF systems, resonance is expected when
ω = Ωn. The attachment of (m and k), which is the dynamic absorber, brings about a very interesting
result.
If a machine or a component is operated at a constant speed and resonant vibrations exists,
response vibrations can be suppressed or eliminated by use of a dynamic absorber. The added
spring, k and mass, m, represent the dynamic absorber.
The dynamic absorber has a natural frequency of
k
ωn = 9 55
m
These are not modal frequencies but are frequencies

of the two-separate spring-mass subparts.
The analysis below shows how the absorber elimi-
nates the resonant vibration of mass, M, and absorbs K
the applied dynamic force. Note that the absorber is
an attachment to mass, M.
x1 M F = F0 cos ωt
Base and Absorber Pinned Together
When the absorber is pinned to the base mass, the dif-
ferential equation of motion for the base mass, M, is
Mx 1 + K + k x 1 + kx 2 = F 0 cos ωt 9 56 k Dynamic absorber
The differential equation of motion for the attached
absorber is x2 m
mx 2 + k x 2 − x 1 = 0 9 57
Vibration response of the two-mass system is Figure 9.11 Math model of dynamic
assumed to be of the form: absorber.
x 1 = X 1 cos ωt 9 58
x 2 = X 2 cos ωt 9 59
By substitution and collecting terms
k ω2 k
X1 1 + − 2 − X 2 = δst 9 60
K Ωn K
ω2
X1 = X2 1 − 9 61
ω2n
where
F0
δst =
K
Simultaneous solution to these two equations is
ω2
X1 1− ω2a
= 9 62
δst 1− ω2
1+ k
− ω2
− k
ω2a K Ω2n K
X2 1
= 9 63
δst 1− ω2
1+ k
− ω2
− k
ω2a K Ω2n K
X1
From these equations, mass M has zero response, = 0, when ωa = ω but mass, m, responds
δst
with amplitude
K F0
X 2 = − δst = − 9 64
k k
The main mass, M, has zero response, x1(t) = 0. The force in the damper spring varies per, −F0 cos ωt,
ω
which is equal and opposite to the external force, F0. This response is true for any value of . Since we
Ωn
are interested in suppressing resonance of the main mass, M, consider the case of ωa = Ωn for which
k K k m
= and = 9 65
m M K M
1.6
1.4
1.2
Frequency ratio, ω/ωa
0.8
0.6
0.4
0.2
0
0 0.1 0.2 0.3 0.4 0.5 0.6
Mass ratio, μ = m/M
Figure 9.12 Separation of the natural frequencies.
Letting
k
μ=
K
Then
2
ω
X1 1− ωa
= 9 66
δst ω
2
ω
2
1− ωa 1 + μ− Ωn −μ
X2 1
= 9 67
δst ω
2
ω
2
1− ωa 1 + μ− Ωn −μ
To summarize, if the main mass is vibrating at resonance, Ωn = ω. The absorber is chosen such
that ωa = ω = Ωn (resonant condition). Question: what should be the size of k and m of the
absorber? As a guide, consider the frequency spacing between the two natural frequencies of
the now 2-DOF system. The denominator in Eqs. (9.66) and (9.67) will help with the answer.
The two resonant frequencies surrounding the null response are found by setting the denomina-
tor of Eq. (9.67) equal to zero.
2 2
ω ω
1− 1 + μ− −μ = 0 9 68
ωa ωa
Expanding and collecting terms,

4 2
ω ω
− 2+μ +1=0 9 69
ωa ωa
The two roots to this equation are
2
ω μ μ2
= 1+ ± μ+ 9 70
ωa 2 2
k m
The plot of this equation is shown in Figure 9.12. If, for example the ratio μ = = is taken as
K M
ω ω
0.2, then the new 2 DOF will resonate at = 0 8 and = 1 25. These two frequency ratios are
ωa ωa
6.0
4.0
2.0
0.0
0 0.5 1 1.5 2 2.5
–2.0
–4.0
–6.0
6.0
4.0
2.0
0.0
0 0.5 1 1.5 2 2.5
–2.0
–4.0
–6.0
Figure 9.13 Frequency responses with dynamic absorber.
also shown in Figure 9.13 by the dashed lines. Keep in mind that dynamic absorbers are designed
ω
for one operating frequency, = 1 0.
ωa
These diagrams show that the separation of resonance points depends on the mass or stiff-
ness ratio.
Example Steps for determining the magnitude of k and m are illustrated an example with the
parameter as follows:
F 0 = 50 N
M = 2 kg
ω = 100 rad s
X 2 ≤ 1 cm requirement
The problem is to determine k and m of the absorber.
Step 1 F0
X2 =
k
50 N 100 cm
k= = 5000 N m
1 cm 1 m
Step 2 ωa = ω
k
= ω2
m
k 5000
m= 2 = = 0 5 kg
ω 10 000
k 50
Since μ = = = 0 25 . According to Eq. (9.70), resonant frequencies are separated by
m 200
ω ω
= 1 29 and = 0 77.
ωa ωa
Dynamic absorber theory was explained in terms of a linear spring-mass model. This model, how-
ever, applies to many practical geometric configurations. The torsion model is like the discrete mass
linear model discussed above so the results in the above apply directly when linear parameters are
converted to angular parameters. The conversion to other configurations is not as obvious. The goal
is to put the model of any SDOF system into the form of a simple spring-mass system.
Example Consider the system shown in Figure 9.14. The primary system is the disc, while the
linear spring-mass is the absorber.
I P θ = − 2rθk 1 2r + 2rF 0 cos ωt 9 71

3 2 x1 x1
r m1 + 2r k 1 2r = 2rF 0 cos ωt
2 r r
3
m1 x 1 + 2k 1 x 1 = F 0 cos ωt
4
3
Its linear equivalence is M = m1 and K = 2k1.
4
k1 Figure 9.14 Dynamic absorber applied to a

F(t) rolling disc.
m1
m
x1
k x2
r
P
F(t)
k1
a x1 m1
0
k
Absorber
m x2
Figure 9.15 Damper applied to pivoted rod.
Example Consider the system shown in Figure 9.15. The equation of motion for the primary
system (the bar) is
I 0 θ = − ak1 aθ + 2aF 0 cos ωt 9 72

2
L m1 x 1
+ ak1 x 1 = 2aF 0 cos ωt
3 a
which translates into
2 1
m1 x 1 + k 1 x 1 = F 0 cos ωt 9 73
3 2
2m1 k1
Therefore, the linear equivalences are M = and K = .
3 2
The equations of motion for the vibration absorber system are
Mx 2 + K + k x 1 − kx 2 = 2F t 9 74
and
mx 2 + k x 2 − x 1 = 0 9 75
The math model of Figure 9.11 and results as explained earlier apply directly to this case.
Multi-DOF Systems – Eigenvalues and Mode Shapes
In general, free vibration equations of motion for a multi-DOF system in are expressed in matrix
form by
m x + k x = 0 9 76
Assuming the solution has the form of

x t = X cos ωt
gives
k − ω2 m X = 0 amplitude equation 9 77
For a nontrivial solution,
k − ω2 m = 0 characteristic equation 9 78
The characteristic equation gives natural frequencies of each mode in the system. Mode shapes
are determined from the amplitude equation.
The result from each equation depends on the stiffness [k] matric and the mass matric [m]. Each
of these matrices is a by-product of the equations of motion for each mass as illustrated with the 2
DOF systems. However, both matrices can be determined directly without developing the differen-
tial equations of motion as explained below.
Flexibility Matrix – Stiffness Matrix

Displacements in engineering structures can be determined through classic bean analysis or using
computer software. These tools provide a convenient way to build a flexibility matrix. Consider a
cantilever beam supporting two concentrated forces. Deflections, x1 and x2 can be determined by
use of a flexibility matrix.
x1 α11 α12 F1
= 9 79
x2 α21 α22 F2
By inversion, the stiffness matrix is
F1 k 11 k 12 x1
=
F2 k 21 k 22 x2
where αij is the displacement at location “i” due to a unit force at location “j.” As per Maxwell’s law
of reciprocity [2], αij = αji. The flexibility elements can easily be determined from beam deflection
theory.
The relationship between the flexibility matrix and the stiffness matrix is determined as follows:
F = k x 9 80
x = α F 9 81
Substituting for {F} gives

x = α k x 9 82
Therefore,
α k = I 9 83
or
−1
k = α 9 84
showing the stiffness matrix is the inverse of the flexibility matrix. In consideration of the top draw-
ing and Eq. (9.81), if F2 = F3 = 0, then
x1 α11
x2 = α21 F1
x3 α31
Figure 9.16 Flexibility (αij) and stiffness (kij) 1 lb

elements.
α31
α11 α21
1 2 3
k11
1 in.
k21
Letting F1 = 1 (Figure 9.16)

x 1 = α11 x 2 = α21 x 3 = α31
According to Eq. (9.80), elements in the stiffness matrix are developed as follows:
F1 k 11 k 12 k 13 x1
F2 = k 21 k 22 k 23 x2 9 85
F3 k 31 k 32 k 33 x3
If locations x2 and x3 are fixed then Eq. (9.85) reduces to

F1 k 11
F2 = k 21 x1
F3 k 31
Furthermore if x1 is given a unit displacement the applied force at location 1 is
F 1 = k11
and
F 2 = k21 reaction force
F 3 = k31 = 0 reaction force
Differential equations of motion may be developed using the flexibility matrix directly:
x = − α m x 9 86
k x = − k α m x
Since [k][α] = [I]
m x + k x = 0 9 87
Example Consider the flexibility and stiffness matrix for the system shown in Figure 9.17.
Following the above procedure
1 5 −1
α = 9 88
k −1 1
k 1 1
k = 9 89
4 1 5
k
k
m
a a m
cg
x1
x2
Figure 9.17 Two degree of freedom system.
The differential equations of motion are given in matrix form as

1 0 x1 k 1 1 x1 0
m + = 9 90
0 1 x2 4 1 5 x2 0
yielding
k
mx 1 + x1 + x2 = 0
4
and
k
mx 2 + x 1 + 5x 2 = 0
4
For comparison, we now develop the differential equations of motion from rigid body analysis
from Chapter 7 (see Figure 9.18).
k
k
m
a a
m
cg
x1 x2
kxcg
kx2
Figure 9.18 Rigid body analysis.

From force equation
F = maG
1 1
− x 1 + x 2 k − kx 2 = 2m x 1 + x 2
2 2
1 3
m x 1 + x 2 + kx 1 + kx 2 = 0 9 91
2 2
From moment equation
MG = IGα G is mass center
x1 − x2
+ akx 2 = 2a2 m
2a
m x 1 − x 2 = − kx 2 = 0 9 92
Notice that equations 9.91 and 9.92 look completely different from the ones developed by the
direct method. However, if the two are added and then subtracted, we get the same two equations.
Direct Determination of the Stiffness Matrix

In general, the stiffness matrix is defined in general by
k 11 k 12 k 13 etc
k 21 k 22
k = 9 93
k 31 k 32
k ij etc
where kij is the force at point “i” as a result of a unit displacement at point “j.” Similar, mij is the
force at point “i” due to a force required to produce a unit acceleration at point “j.”
Example A tight cable with three masses equally spaced by “a” distance as shown in Figure 9.19.
The cable is tensioned by force T.
From the diagram
T T
k 11 = 2 , k 21 = − , k 31 = 0
a a
m m m
T T
a a a a
k11
2
1 k21 3
Figure 9.19 Stiffness matrix of 3-DOF system.

Accordingly, the stiffness matrix for the system is

2 −1 0
T
k = −1 2 −1 9 94
a
0 −1 2
Direct Determination of the Mass Matrix

The mass matrix can also be established in a similar manner. Consider the matrix equation:
F1 m11 m12 m13 x1

F2 = m21 m22 m23 x2
F3 m31 m32 m33 x3
If masses 2 and 3 are fixed and mass 1 given a unit acceleration
f1 m11 m12 m13 1

0 = m21 m22 m23 0
0 m31 m32 m33 0
f 1 = m11 1 f 2 = 0, f 3 = 0
but f1 = m(1), therefore m(1) = m11(1) and m11 = m.
Following this procedure, the mass matrix fills out as follows:
1 0 0
m =m 0 1 0 9 95
0 0 1
Amplitude and Characteristic Equations

The amplitude equation is
2−ξ −1 0 X1 0
−1 2−ξ −1 X2 = 0 9 96
0 −1 2−ξ X3 0
m ω2 a
where ξ = T . The characteristic equation is
2−ξ 2−ξ 2 −2 = 0 9 97
The eigenvalues from this equation are
T
ξ1 = 2 − 2, ω21 = 2− 2 9 98
ma
T
ξ1 = 2, ω21 = 2 9 99
ma
T
ξ1 = 2 + 2, ω21 = 2+ 2 9 100
ma
The modal matrix is

1 1 1
X1 X2 X3 = 2 0 − 2 9 101
1 −1 1
The first column is the shape of the first mode, etc. Mode shapes are useful for visualizing the total
response to various dynamic loads applied to a spring mass system. In general, total motion involves
all modes, however, any one mode can stand out especially, when the frequency of excitation is
close to the natural frequency of a given mode.
Example Consider the system shown in Figure 9.20.
1 1 1
X 1, X 2X 3 = 1 83 0 445 − 1 25 9 102
2 247 − 0 802 0 555
The corresponding eigenvalues are

ξ1 = 0 198
ξ2 = 1 55
ξ3 = 3 25
where
mω2
ξ=
k
Therefore,
k k k
ω1 = 0 445 , ω2 = 1 072 , ω3 = 1 803
m m m
m x1
m x2
m x3
ξ1 = 0.198 ξ2 = 1.55 ξ3 = 3.25
Figure 9.20 Mode shapes and frequencies.

Parameters Not Chosen at Discrete Masses

In some 2-DOF systems, there may be more than one discrete mass and only two dependent vari-
ables are required such as shown in Figure 9.21. In this example, two of the discrete masses are not
located at the points of displacements (x1 and x2). This system is viewed as a rigid body composed of
a rigid weightless bar containing three discrete masses.
The stiffness matrix can be determined as in the previous example. In this case, give point 1 a unit
displacement and determine the required force, f1. At the same time determine the reaction force,
f2, at point 2. The results are
10
f 1 = k11 = k 9 103
9
2
f 2 = k21 = k 9 104
9
Next fix point 1 and give point 2 a unit displacement:
13
f 2 = k22 = k 9 105
9
2
f 1 = k12 = k 9 106
9
The stiffness matrix thus becomes
k 10 2
k = 9 107
9 2 13
The mass matrix is determined in a similar manner (Figure 9.22). First fix point 2 and give point 1
a unit acceleration to determine m11 and m21. This assessment requires dynamic principles as
follows.
x1 x2
2m m m
a a a
k k k
Figure 9.21 Two degree of freedom system.
f11
2m m m f21
a a a
Figure 9.22 Model for mass matrix.

With point 2 fixed and applying T = m i ai r i

2 1
f 11 3 a = 2m 2a + m a 9 108
3 3
f 11 = m 9 109
Applying F= mia1 to the freebody gives
2 1 5
f 11 + f 21 = 2m
+m = m 9 110
3 3 3
5 2
f 21 = m−m = m 9 111
3 3
Therefore,
2
m11 = m and m21 = m 9 112
3
Next fix point 1 and give point 2 a unit acceleration. Reactions are
5
f 22 = m22 = m 9 113
3
2
f 12 = m12 = m 9 114
3
The mass matrix thus becomes
TT
m 3 2
m = 9 115
3 2 5
0
Lateral Stiffness of a Vertical Cable

A long vertical cable containing three discrete
masses is shown in Figure 9.23. The tension at the m 1 k11
top is greater than the tension at the bottom due
to the distribution of the three masses. Determine
the stiffness matrix for lateral modes of vibration.
The difference between this example and the k21
m 2
example in Figure 9.19 is variable tension between
the masses. There is some similarity between this
example and vertical drill pipe, but in this case bend- a
ing flexibility included. Bending stiffness, however,
could be modeled as torsion springs at the connection m 3
points. The mass matrix is the same, however, the
stiffness matrix is modified as explained below. Ten-
sions between the three masses will be repre-
sented by
4
T0 tension between points 0 and 1 Top tension
T1 tension between points 1 and 2 T1 = T0 − m/g TB
T2 tension between points 2 and 3 T2 = T0 − 2m/g
T3 tension between points 3 and 4 T3 = T0 − 3m/g Figure 9.23 Lateral vibration of hanging pipe.
Tension in the fourth member is also the pull at the bottom end. If it is desired to maintain a
bottom pull of TB the top pull force has to be TT = TB + 3m/g.
1
k11 = T0 + T1 9 116
a
1
k 21 = − T1 9 117
a
k 31 =0 9 118
1
k 22 = T1 + T2 9 119
a
1
k 32 = − T2 9 120
a
1
k 33 = T3 + T4 k 13 = 0 9 121
a
Building the Damping Matrix

In the general case, damping must be included in the equations of motion:
m x + c x + k x = F 9 122
The damping matrix is represented by
c11 c12 c13 c14
c21 c22 c23
c = 9 123
c31 c32 cij
c41 etc
The damping matrix can also be developed in a manner similar to the stiffness and mass matrices.
In this case, cij is the force at point “i” due to a force required to produce a unit velocity at point “j.”
Modal Analysis of Discrete Systems
The modal analysis method uncouples the differential equations of motion and puts them into the
form of a SDOF equation, whose solution was discussed earlier [3]. The transformation is based on
x t = X η t 9 124
where
{x(t)} – local coordinates

[X] – modal matrix
{η (t)} – modal coordinates
The expansion of Eq. (9.124) for a 3-DOF system is
x1 X 11 X 12 X 13 η1
x2 = X 21 X 22 X 23 η2 9 125
x3 X 31 X 32 X 33 η3
x1 X 11 X 12 X 13
x2 = X 21 η1 t + X 22 η2 t + X 23 η3 t 9 126
x3 X 31 X 32 X 33
x1
x2 =X 1 η1 t + X 2 η2 t + X 3 η3 t 9 127
x3
where X(j) represents the jth mode. The displacement of the ith coordinate expressed in terms of
contributions from each mode is
N
xi = X ij η j t 9 128
j=1
where
i – ith location
j – jth mode
Xij – mode amplitude at ith location within the jth mode
The overall vibration of a multisystem can be visualized in terms of each modal responses.
Orthogonal Properties of Natural Modes

Consider a discrete mass system whose displacements are defined by {x(t)}. The equations of motion
for each mass in terms of local coordinates are
m x + c x + k x = F 9 129
Using Eq. (9.128),
m X η + c X η + k X η = F 9 130
Multiplying each term by the transposed modal matrix gives
T T T T
X m X η + X c X η + X k X η = X F 9 131
The orthogonality property of the modes with respect to mass and stiffness matrices gives
X Tr m X s =0 9 132
X Tr m X r = Mr modal mass 9 133
X Tr k X s =0 9 134
X Tr k X r = Kr modal stiffness 9 135
The modes are not necessarily orthogonal with respect to the damping matrix. This is true only if
damping is proportional. Assuming proportional damping
X Tr c X s =0 9 136
X Tr c X r = Cr modal damping 9 137

In each case, modal mass, damping, and stiffness are diagonal matrices. This orthogonality prop-
erty of the modes allows the equations of motion to be converted into modal coordinates, ηi(t). The
converted equation becomes
M η + C η + K η = Q 9 138
where modal mass, damping, and stiffness matrices are diagonal, and the modal force is
Q = XT F 9 139
Because of the orthogonality properties of the natural modes, the modal mass and stiffness matri-
ces are diagonal. If damping is proportional, the modal damping matrix is also diagonal. The dif-
ferential equation defining the response of each mode is
M i ηi + Ci ηi + K i η = Qi i = 1, 2, 3, 4, … 9 140
where
T
Qi = X i F 9 141
The solution to Eq. (9.140) gives the time history of each mode. Theory discussed earlier for SDOF
problems applies directly to modal responses to various forces, whether periodic or not.
The results in terms of modal coordinates transfer directly back to local coordinates according to
Eq. (9.128).
Proportional Damping
The local damping matrix converts into a diagonal matrix only if [c] is proportional to [m] and [k]:
c =αm +βk 9 142
The proof that this type of local damping leads to a diagonal matrix is as follows. For the off-
diagonal terms in the modal damping matrix to be zero:
X Ts c X r = 0 9 143
By substitution
X Ts α m + β k X r = αX Ts m X r + βX Ts k X r 9 144
Since
X Ts m X r = 0 9 145a
X Ts k X r = 0 9 145b
then
X Ts c X r = 0 9 146
Therefore,
X Tr c X r = Cr 9 147
Examples are illustrated in Figure 9.24.

(a) (b) (c)
k c k c k c
m m m
c c
k k k
m m m
c c
Figure 9.24 Examples of damped systems.
Case a
2 −1 c
c =c , α = 0, β= proportional damping
−1 1 k
Case b
1 0 c
c =c , α= , β=0 proportional damping
0 1 m
Case c
2 −1
c =c not proportional damping
−1 2
Transforming Modal Solution to Local Coordinates

The transformation of local coordinates into modal coordinates resolves multidegree of freedom
systems into the form:
M η + C η + K η = 0 9 148
as explained above and greatly simplifies the mathematics. The solution to this equation is well
documented [1]. Once a solution is obtained in terms of modal coordinates, it is transformed back
to local coordinates according to
x t = X η t 9 149
where
{x(t)} – local coordinates

[X] – modal matrix
k
{η(t)} – modal coordinates
This equation shows how each mode contributes to the total response, a
useful aspect of modal analysis.
x1 m
For the 2-D example shown in Figure 9.25, the reverse transformation
requires the following matrices:
k c 1 1
X = modal matrix
1 −1
1 0 1 −1 2 −1
x2
m =m c =c k =k
m 0 1 −1 1 −1 2
2 0 0 0 2 0
M =m C =c K =k
0 2 0 4 0 6
c c
α= − β=
m k
All modal matrices are diagonal, and damping is proportional.
In some cases, off diagonal terms are ignored for the sake of simplicity.
Figure 9.25 Damping In many cases, this assumption gives reasonable engineering solutions.
between two masses.
Free Vibration of Multiple DOF Systems
Free vibration of multidegrees of freedom systems can be visualized as the summation of all vibra-
tion modes. The level of participation of each mode depends on initial conditions. The approach to
free vibration response follows the same approach outlined above. Initial conditions are expressed
in terms of local coordinates, which must be transformed into modal coordinates.
The modal equation when there are no external forces is
M η + C η + K η = 0 9 150
giving
M i ηi + Ci ηi + K i ηi = 0 9 151
Assuming zero damping the solution to each modal response is
ηi t = Ai cos ωi t + Bi sin ωi t 9 152
The arbitrary constants are determined from initial conditions given in terms of local coordinates
and transformed to modal coordinates per
x 0 = X η0 9 153
x 0 = X η0 9 154
Transferring to modal coordinates,
X Ti m x 0
ηi 0 = ; k could also be used 9 155
Mi
X Ti m x 0
ηi 0 = ; k could also be used 9 156
Mi
Once the solution is obtained in modal coordinates, it is transformed back into local coordinates
for the true response.
The overall response to a given set of initial conditions is defined by Eq. (9.128).
Free Vibration of 2 DOF Systems

Example Consider the free vibrations of the 2-degree system in Figure 9.3. Assume the following
initial conditions for x1 and x2. The initial position of x1(0) = x10, while x 2 0 = x 1 0 = x 2 0 = 0.
x 10
x 0 = 9 157
0
0
x 0 = 9 158
0
Modal mass and modal stiffness matrics are
3 618 0
M =m 9 159
0 1 382
1 386 0
K =k 9 160
0 3 618
Modal equations are
M 1 η1 + K 1 η1 = 0
M 2 η2 + K 2 η2 = 0
The solutions to both equations are
η1 0
η1 t = η1 0 cos ω1 t + sin ω1 t 9 161
ω1
η2 0
η2 t = η2 0 cos ω2 t + sin ω2 t 9 162
ω2
From Eqs. (9.155) and (9.156)
X T1 m x 0 x 10
η1 0 = =
M1 3 618
X T2 m x 0 x 10
η2 0 = =
M2 1 382
X T1 m x 0
η1 0 = =0
M1
X T2 m x 0
η2 0 = =0
M2
Therefore,
x 10
η1 t = cos ω1 t
3 618
x 10
η2 t = cos ω2 t
1 382
The response of each mass is the modal summation:

x 1 t = η1 t + η2 t 9 163
x 2 t = 1 618η1 t − 0 618 η2 t 9 164
If the masses are displaced initially in a mode shape, only that mode vibrates in free vibration.
Suddenly Stopping Drill Pipe with the Slips

Drill pipe is lowered into a well bore by block and tackle. Downward movement is controlled by a
band brake. The rate of stopping affects magnitude of stress waves set up in the pipe. Sometimes the
slips are clamped around the pipe causing an instantaneous stop. Axial vibration response resulting
from setting the slips too early is discussed below. The problem is to determine the response of pipe
(modeled as multiple springs and masses) moving downward with velocity v0 after point 0 is sud-
denly stopped (Figure 9.26).
Ignoring damping, the equation of motion of the system is
mi x i + k x i 0 = 0 9 165
where
1 0 0 0 2 −1 0 0
0 1 0 0 −1 2 −1 0
m =m and k =k
0 0 1 0 0 −1 2 −1
0 0 0 0 0 0 −1 1
Values of m and k according to the approxima-

Pipe tion are
0 wl EA
m= and k=
g l
k
The modal matrix, X, is determined as before
along with the natural frequency, ωi of each mode.
m 1
Transforming from local coordinates to modal
coordinates using
k
x i = X ηi
v0
m 2 uncouples the modal differential equation:
M ηi + K ηi = 0 9 166
k
Solutions to this equation, which describes
m 3 the time history of each mode separately, are of
the form:
k ηi t = Ai cos ωi t + Bi sin ωi t 9 167

In terms of initial conditions
m 4
ηi 0
ηi t = ηi 0 cos ωt + sin ωt
Figure 9.26 Suddenly stopping pipe with the slips. ωi
Initial conditions in terms of local coordinates are

xi 0 = 0 9 168
xi 0 = v0 1 9 169
These initial conditions transform into modal coordinates by reverse transformation of
x i = X ηi
T
X m x = XT m X η = M η
This equation yields
X iT m x o
ηi 0 = =0 9 170
Mi
and
X iT m x o X Ti m v0
ηi 0 = = 9 171
Mi Mi
By substitution
ηi 0
ηi t = sin ωi t 9 172
ωi
Note: X Ti is a horizontal 1 × 4 matrix. When multiplied into the square matrix gives a 1 × 4 flat
matrix.
Response in terms of local coordinates is
xi = X i ηi t 9 173
i = 1, 2, 3, 4
Example Consider the 3-DOF system of Figure 9.20. The modal matrix, [X] and modal frequen-
cies are known. Using this model:
1 0 0 1
1 1 83 2 247 m 0 1 0 v0 1
X T m v0 0 0 1 1
η1 0 = 1 = 9 174
Mi M1
5 077mvo 5 077mv0
η1 0 = = = 0 54v0 9 175
M1 9 398m
1 0 0 1
1 0 445 − 0 802 m 0 1 0 v0 1
X T2
m v0 0 0 1 1
η2 0 = = 9 176
M2 M2
0 652mvo 0 652mv0
η2 0 = = = 0 354v0 9 177
M2 1 84m
1 0 0 1
1 − 1 25 0 555 m 0 1 0 v0 1
X T3
m v0 0 0 1 1
η3 0 = = 9 178
M3 M3
0 305mvo 0 305mv0
η3 0 = = = 0 106v0 9 179
M3 2 87m
Therefore
0 54v0
η1 t = sin ω1 t
ω1
0 354v0
η2 t = sin ω2 t
ω2
0 106v0
η3 t = sin ω3 t
ω3
Recall model mass, Mi is
T
Mi = X i m Xi
m 0 0 1 1
M1 = 1 1 83 2 247 0 m 0 1 83 =m 1 1 83 2 247 1 83
0 0 m 2 247 2 247
M 1 = 9 398m
M 2 = 1 841m
M 3 = 2 87m
The response of each mass will contain contributions from each of the three modes.
The final solution is
x1 1 1 1
x2 = 1 83 η1 t + 0 445 η2 t + − 1 25 η3 t 9 180
x3 2 247 − 0 802 0 555
Critical Damping of Vibration Modes

Damping characteristics of each mode are determined from the free vibration equation:
M i ηi + Ci ηi + K i ηi = 0 9 181
Similar to the SDOF equation in local coordinates:
Ci Ci
ζi = = 9 182
Ci,cr 2M i ωi
Ci,cr = 2 K i M i 9 183
Consider the model in Figure 9.27.
c c c
m m m
T T
a a a a
1 2 3
Figure 9.27 Damped vibration model.
As given earlier the modal matrix for this system is
1 1 1
X1 X2 X3 = 2 0 − 2 9 184
1 −1 1
The eigenvalues and natural frequencies are
T
ξ1 = 2 − 2, ω21 = 2− 2
ma
T
ξ1 = 2, ω21 = 2
ma
T
ξ1 = 2 + 2, ω21 = 2+ 2
ma
The damping matrix is
1 2 1 1 0 0 1 1 1
C =c 1 0 −1 0 1 0 2 0 − 2 9 185
1 − 2 1 0 0 1 1 −1 1
4 0 0
C =c 0 2 0 9 186
0 0 4
Similarly, the modal mass matrix is
4 0 0
M =m 0 2 0 9 187
0 0 4
The damping factor for each of the three modes is
Ci Ci
ζi = = 9 188
Ci,cr 2M i ωi
c 1 c 1 c 1
ζ1 = , ζ2 = , ζ3 = 9 189
2m ω1 2m ω2 2m ω3
Showing for this example ζ 1 ζ2 ζ 3. The maximum amplitude for each modal response occurs
at rr = 1 and is defined by
Qr0 1
ηr = 9 190
K r 2ζ r
Forced Vibration by Harmonic Excitation
Consider separate forces are applied simultaneously to each mass. Modal responses are obtained
from the solution to
Qr
ηr + 2ζ r ωr ηr + ω2r ηr = 9 191
Mr
where
ζ r – mode damping factor, = Cr

C r,cr
Cr,cr = 2Mrωr
Qr = X(r)T{F}
and assuming the force amplitude at each point is periodic:

F = F 0 cos ωt 9 192
By substitution
Qr = Qr0 cos ωt 9 193
where
T
Qr0 = X r F0
Following the solution of a SDOF system,

Qr0
Kr
ηr t = 1 cos ωt − ϕr 9 194
2 2 2
1 − r 2r + 2ζ r r r
where
2ζ r r r
tan ϕr =
1 − r 2r
ω
rr =
ωr
ϕr represents the phase lag of each mode with respect to the modal force, Qr.
Complex Variable Approach

In terms of complex solution, the complex modal force is
T
Qr = X r F 0 eiωt 9 195
The complex variable solution to Eq. (9.191) is

Qr0 K r
ηr t = eiωt 9 196
1 − r 2r + i 2ζ r r r
In terms of local coordinates

N
x = X η = X r ηr 9 197
r=1
For example, the expression for a 3-DOF system is
x1
x2 = X 1 η1 + X 2 η2 + X 3 η3 9 198
x3
The response at point i is

N
xi = X ir ηr 9 199
r=1
where X(r) is the shape of the rth mode. Xir is mode amplitude at location i of the rth mode. By
substitution
N
Qr0 K r
xi t = X r eiωt 9 200
r=1 1 − r 2r + i 2ζ r r r
N T
X r F0 1
r=1
Kr 1 − r 2r + i 2ζ r r r
The complex amplitude of the response is

N T
XrX r F0 1
Xi = 9 202
r=1
Kr 1 − r 2r + i 2ζ r r r
From which the complex amplitude at each location is
N
X 1r X r T F0 1
X1 = , for i = 1 9 203
r=1
Kr 1 − r 2r + i 2ζ r r r
N
X 2r X r T F0 1
X2 = , for i = 2, etc 9 204
r=1
Kr 1 − r 2r + i 2ζ r r r
Note that when rr = 1 for any mode, that mode responds with resonance.
Harmonic Excitation of 3 DOF Systems

Consider the 3-DOF system of Figure 9.28. Assume separate forces applied simultaneously to each
mass and linear damping is parallel to each spring. Modal responses are obtained from the
solution to
Qr
η + 2ζ r ωr ηr + ω2r ηr = 9 205
Mr
F1 m x1
m x2
m x3
Figure 9.28 Mode shapes.
where
ζ r – mode damping factor, = Cr

C r,cr
Cr, cr = 2Mrωr
Qr = X(r)T{F}
and assuming the force amplitude at each point varies according to
F = F r 0 cos ωt 9 206
By substitution
Qr = Qr0 cos ωt 9 207
where
T
Qr0 = X r F0
Following the solution of SDOF differential,

Qr0
Kr
ηr t = 1 cos ωt − ϕr 9 208
2 2 2
1 − r 2r + 2ζ r r r
where
2ζ r r r
tan ϕr =
1 − r 2r
ω
rr =
ωr
ϕr represents the phase lag of each mode with respect to the modal force, Qr.
Modal Solution of a Damped 2-DOF System

The following damped system is used to illustrate the modal analysis approach (Figure 9.29).
The modal matrix for this arrangement is
1 1
X =
1 −1
The local mass, stiffness and damping matrices are given below:
1 0 1 −1 1 −1
m =m c =c k =k
0 1 −1 1 −1 1
The modal matrices are

1 1 1 0 1 1 2 0
M =m =m
1 −1 0 1 1 −1 0 2
1 1 2 −1 1 1 2 0
K =k =k
1 −1 −1 2 1 −1 0 6
1 1 1 −1 1 1 0 0
C =c =c
1 −1 −1 1 1 −1 0 4
Modal force is
1 1 1
Q = F 0 cos ωt
1 −1 0
Modal equations are

M 1 η1 + C1 η1 + K 1 η1 = F 0 cos ωt
M 2 η2 + C2 η2 + K 2 η2 = F 0 cos ωt
Modal solutions then become
F0 K
η1 t = 1
1 cos ωt − ϕ1 k
1 − r1 2 2 + 2ζ 1 r 1 2 2
9 209 F1 = F0 cos ωt
x1 m
F0 K
η2 t = 2
1 cos ωt − ϕ2
1 − r22 2
+ 2ζ 2 r 2 2 2
k c
9 210
Bring it all together x2 m
x 1 t = X 11 η1 t + X 12 η2 t
9 211
x 1 t = η1 t + η2 t
and
x 2 t = X 21 η1 t + X 22 η2 t
9 212 Figure 9.29 Damped forced vibration.
x 2 t = η1 t − η2 t
General Complex Variable Solution

In terms of complex solution, the complex modal force is
T
Qr = X r F 0 eiωt 9 213
The complex variable solution to Eq. (9.205) is

Qr0 K r
ηr t = eiωt 9 214
1 − r 2r + i 2ζ r r r
In terms of local coordinates

N
x = X η = X r ηr 9 215
r=1
For example, the expression for a 3-DOF system is
x1
x2 = X 1 η1 + X 2 η2 + X 3 η3 9 216
x3
The response at point i is

N
x = X η = X r ηr
r=1
N
xi = X ir ηr 9 217
r=1
where X(r) is the shape of the rth mode. Xir is mode amplitude at location i of the rth mode. By
substitution [4]
N
Qr0 K r
r=1 1 − r 2r + i 2ζ r r r
N T
X r F0 1
r=1
Kr 1 − r 2r + i 2ζ r r r
The complex amplitude of the response is

N T
XrX r F0 1
Xi = 9 220
r=1
Kr 1 − r 2r + i 2ζ r r r
From which the complex amplitude of each mass is

N
X 1r X r T F0 1
X1 = , for i = 1 9 221
r=1
Kr 1 − r 2r + i 2ζ r r r
N
X 2r X r T F0 1
X2 = , for i = 2, etc 9 222
r=1
Kr 1 − r 2r + i 2ζ r r r
Note that when rr = 1 for any mode, that mode responds with resonance.
Consider, for example a 3-DOF system where a periodic force is applied to only one the mass at
location number one (1):
F 1 = F 01 cos ωt 9 223
Using the complex notation, the force matrix becomes
F 01
F = 0 eiωt 9 224
0
Since
T
Q = X F 9 225
By Eq. (9.219)
N T
X ir X r F0 1
xi t = eiωt 9 226
r=1
Kr 1 − r 2r + i 2ζ r r r
Assuming a harmonic force is applied only to mass number one (1)
F 01
T
Q = X 0 eiωt 9 227
0
then
Qr = X Tr F 01 eiωt
If each mode is normalized to the amplitude at location number on (1), then [4]
N
X ir F 01 1
xi t = eiωt 9 228
r=1
K r 1 − r 2
r + i 2ζ r r r
Maximum displacement amplitudes are defined by
N
X1 1 1
= X 1r DTF 9 229
F 01 r=1
K r 1 − r 2r + i 2ζ r r r
N
X2 1 1
= X 2r CTF 9 230
F 01 r=1
K r 1 − r 2r + i 2ζ r r r
N
X3 1 1
= X 3r CTF 9 231
F 01 r=1
K r 1 − r 2r + i 2ζ r r r
Expanding each of these equations gives

X1 1 1 1 1 1 1
= X 11 + X 12 + X 13
F 01 K 1 1 − r 21 + i 2ζ 1 r 1 K 2 1 − r 22 + i 2ζ 2 r 2 K 3 1 − r 23 + i 2ζ 3 r 3
9 232
X2 1 1 1 1 1 1
= X 21 + X 22 + X 23
F 01 K 1 1 − r 1 + i 2ζ 1 r 1
2 K 2 1 − r 2 + i 2ζ 2 r 2
2 K 3 1 − r 3 + i 2ζ 3 r 3
2
9 233
X3 1 1 1 1 1 1
= X 31 + X 32 + X 33
F 01 K 1 1 − r 21 + i 2ζ 1 r 1 K 2 1 − r 22 + i 2ζ 2 r 2 K 3 1 − r 23 + i 2ζ 3 r 3
9 234
Each of these transfer functions has real and imaginary parts.
Experimental Modal Analysis
Sometimes it is useful to determine modal mass, damping, and stiffness as well as modal frequen-
cies and mode shapes experimentally. This is especially useful when shape and geometry of an elas-
tic system is complex and difficult to model mathematically. The analytical basis for obtaining
vibration parameters experimentally is explained below.
If this information is established experimentally, then it is possible to extract mode shapes, modal
parameters (mass, stiffness, and damping), and natural frequencies of each mode. This is possible
because the reach of each term in each of the three equations as a short reach as shown in Fig-
ures 9.30 and 9.31. Modal information can be transferred back to local coordinates as discussed
earlier.
For simplicity, consider the 2-DOF system in Figure 9.8 with the addition of dampers parallel to
the springs. Specific parameters for this example are
m1 = m = 10/386 lb s2/in.
m2 = m = 10/386 lb s2/in.
k1 = k = 50 lb/in.
k2 = k = 50 lb/in.
c1 = c = 0.05 lb/ips
c2 = c = 0.05 lb/ips
The modal matrix from earlier discussion is

1 1
X =
1 618 − 0 618
The response to a periodic force applied to mass #1 is
X1 1 1 1 1
= 1 + 1 9 235
F 01 K 1 1 − r 1 + i 2ζ 1 r 1
2 K 2 1 − r 2 + i 2ζ 2 r 2
2
X2 1 1 1 1
= 1 618 − 0 618 9 236
F 01 K 1 1 − r 21 + i 2ζ 1 r 1 K 2 1 − r 22 + i 2ζ 2 r 2
Local mass, stiffness, and damping matrices are
1 0 2 −1 2 −1
m =m , k =k , c =c
0 1 −1 1 −1 1
Experimental Modal Analysis 533
0.25
0.2
0.15
0.1
Direct transfer function,
0.05
Re[X1/F01]
0
0 2 4 6 8 10 12 14 16 18 20
–0.05
–0.1
–0.15
–0.2
–0.25
Frequency of excitation, f (cps)
0
–0.05 0 2 4 6 8 10 12 14 16 18 20
–0.1
–0.15
Direct transfer function,
–0.2
Im[X1/F01]
–0.25
–0.3
–0.35
–0.4
–0.45
–0.5
Frequency of excitation, f (cps)
Figure 9.30 Direct transfer function.
Elements in the modal mass, stiffness, and damping matrices are determined from
T 3 618 0
M = X m X =m ; M 1 = 0 094 and M 2 = 0 0358
0 1 382
T 1 382 0
K = X k X =k ; K 1 = 69 1 and K 2 = 181
0 3 618
T 1 382 0
C = X c X =c ; C1 = 0 07 and C 2 = 0 18
0 3 618
0.25
0.2
0.15
Cross Transfer Function,
0.1
Re[X2/F01]
0.05
-0.05 0 2 4 6 8 10 12 14 16 18 20
-0.1
-0.15
-0.2
-0.25
Frequency of Excitation, f (cps)
0.1
0
0 2 4 6 8 10 12 14 16 18 20
-0.1
Cross Transfer Function,
-0.2
Im[X2/F01]
-0.3
-0.4
-0.5
-0.6
-0.7
-0.8
Frequency of Excitation, f (cps)
Figure 9.31 Cross transfer function.
The critical damping factor for the first and second mode are
C1
Ccr1 = 2M 1 ω1 = 5 09 lb ips ζ1 = = 0 0136
Ccr1
C2
Ccr2 = 2M 2 ω2 = 5 09 lb ips ζ2 = = 0 0355
Ccr2
The natural frequencies of the first and second modes are
k K1
ω1 = 0 618 = = 27 15 rad s or f 1 = 4 32 cps
m M1
k K2
ω2 = 1 618 = = 71 08 rad s or f 2 = 11 3 cps
m M2
Modal Response to Nonperiodic Forces 535
Modal Response to Nonperiodic Forces
The modal analysis method is a powerful tool in analyzing the response of multi-DOF systems to a
variety of applied forces such as
1) Nonperiodic forces
2) Step, ramp functions
3) Base motion
4) Imbalance rotating mass
These applications show the true value of the modal analysis method.
A tight cable contains two discrete masses that is stretched by force T as depicted in
Figure 9.32. The objective is to determine system response to a nonperiodic force. An assumed
force history is also shown in the figure. One of the special features of modal analysis is its
ability to adapt SDOD solutions to multidegrees of freedom problems, which in this case is a 2-
DOF system.
In this case, the amplitude and characteristic equations are
1 1
X = and 2−ξ 2 −1 = 0
1 −1
mω2 a
The eigenvalues are ξ1 = 1 and ξ2 = 3, where ξ = T .
As before, modal response is determined from
Qr
ηr + 2ζ r ωr ηr + ω2r ηr =
Mr
However, in this case, the external force is not periodic, but defined by the graph. For the sake of
simplicity, we consider an undamped system such that
Qr
ηr + ω2r ηr = 9 237
Mr
m m
T T
a a a
F(t)
F(t)
2F
t1 t2
t
Figure 9.32 Response of 2-DOF to non-periodic force.

Applying Duhamel’s method [5], modal response is determined from

t
1 t−τ
ηr t = Qr τ − i eiωr dτ 9 238
M r ωr
0
where the modal force, is
F1 t
Qr = X Tr applies for any F 1 t 9 239
0
In this case, F1 is a square function defined by the diagram in Figure 9.32. By substitution,
t
1 F t t−τ
ηr t = XT − i eiωr dτ 9 240
M r ωr r 0
0
Modal response by both vibration modes is defined by η1(t) and η2(t). Local response is defined by
N
xi = X ir ηr 9 241
r=1
Natural Frequencies of Drillstrings
Natural frequencies of drill strings can be determined by solving the differential equation of motion
for the continuous pipe system. The differential equation of motion for axial vibration is
∂2 u ∂2 u
AE = m 9 242
∂x 2 ∂t 2
∂2 u 1 ∂2 u
= 9 243
∂x 2 c2 ∂t 2
where
u(x,t) – axial displacement of any point within the pipe

A – cross-section area of pipe
E – modulus of elasticity
m – mass per unit length of pipe
c – acoustic velocity of compression(tension) wave (16 850 ft/s in steel pipe)
This equation applies over both drill pipe and drill collar sections.
u1(x, t) – drill collar displacements

u2(x, t) – drill pipe displacements
The mathematical modal considered here is shown in Figure 9.33.

The determination of mode shapes and corresponding natural frequencies proceeds as follows.
Boundary conditions are
Natural Frequencies of Drillstrings 537
u1 0, t = 0 drill bit 9 244a

u1 L1 , t = u2 L1 , t interface; pipe collars k
9 244b
M
du1 du2
A1 E L 1 , t = A2 E L1 , t
dx dx 9 244c
interface; pipe collars
∂ 2 u2 ∂u2
M + A2 E + ku2 = 0 drawworks
∂t 2 ∂x Drill pipe
9 244d
The fourth boundary condition is derived by direct
application of the second law to the top mass. L
Natural frequencies are determined, following the
method of separation of variables as applied in Ref.
[6] starting with
u x, t = X x sin ωt 9 245
x
Figure 9.34 shows how natural frequencies change
with hole depth for the following set of drillstring para- Drill collars L1
meters. Multiple roots to the characteristic equation
give the frequencies of each of the modes of vibration:
Figure 9.33 Axial vibration models of
Drill pipe size (5 in. OD by 4.276 in. ID)
drillstrings.
Drill collar size (7 in. OD by 2.937 in. ID)
Length (L1) = 550 ft
Acoustic velocity = 16 850 ft/s
Spring constant of drawwork cables (k) = 640 000 lb/ft (53 333 lb/in.)
Mass of swivel and traveling block (M) = 466 slugs (15 000 lb)
Drillstring length (L) is a variable.
Considering a well depth of 5000 ft, the natural frequency of the fourth mode is 6.13 cps and a
critical rotary speed of 122.6 rpm (Table 9.1). The fourth mode was resonant with a roller cone bit at
this speed. Note also that the frequency band width of the response is quite large so any of the axial
modes can easily be excited.
On the other hand, a rotary speed of 80 rpm would fall between critical speeds of the second and
third axial modes so a roller cone bit should operate relatively smooth at this speed.
According to Figure 9.34, the frequency bandwidth of any mode is in the order of Δf ~ 0.1 cps (ΔN
~ 2 rpm). Since the exciting frequency is related directly to bit rotation, rotary speed has to be main-
tained within a ±1.0 rpm variance in order for energy to be fed into the mode. It takes time to put
energy into a mode because of drill string length. On the other hand, the frequency band width of
drill collar response is much larger, plus the drill collars are much shorter making it easier to excite
drill collar modes as demonstrated in shallow test wells. The frequency band width at 20 000 ft is
much smaller, and it is more difficult for bit speeds to latch onto these modes.
Figure 9.34 also shows how natural frequency changes with depth. For example, at a rotary speed
of 120 rpm (6 cps), mode number changes about every 1500 ft.
Natural frequencies of marine risers and modal responses are discussed in Refs. [7, 8].
Table 9.1 Natural frequencies of axial modes [6].
Length – 5000 ft Length – 20 000 ft
Natural Critical rotary Natural Critical rotary

Mode frequency (cps) speeda (rpm) frequency (cps) speeda (rpm)
1 1.72 34.4 0.414 8.3

2 3.39 67.8 0.84 16.8
3 4.87 97.4 1.26 25.2
4 6.13 122.6 1.67 33.4
5 7.26 145.2 2.09 41.8
6 8.30 166.0 2.50 50.0
7 9.77 195.4 2.92 58.4
a
Based on three cycles per bit rotation.
Source: Based on Dareing [6].
20 000
15 000
Drillstring length, ft
10 000
5000
0
0 2 4 6 8 10
Natural frequency, cps
Figure 9.34 Natural frequencies of axial modes (critical rotary speed is 20 times frequency). Source: Based on
Dareing [6].
References
1 Den Hartog, J.P. (1953). Mechanical Vibrations, 4e. New York: McGraw-Hill.
2 Langhaar, H.L. (1962). Energy Methods in Applied Mechanics. New York: Wiley.
3 Timoshenko, S., Young, D.H., and Weaver, W. Jr. (1974). Vibration Problems in Engineering, 4e. Wiley.
4 Craig, R.R. Jr. (1981). Structural Dynamics. New York: Wiley.
References 539
5 Thomson, W.T. and Dahleh, M.D. (1998). Theory of Vibrations with Applications. New York:
Prentice Hall.
7 Dareing, D.W. and Huang, T. (1976). Natural frequencies of marine risers. J. Pet. Technol. 28: 813–818.
8 Dareing, D.W. and Huang, T. (1979). Marine riser vibration response determined by modal analysis.
ASME J. Energy Resour. Technol. 101: 159–166.
541
10
Fluid Mechanics
Two types of fluid flow are turbulent and laminar. Turbulent fluid flow problems require the use of
empirical formulas based on experimental data. Laminar flow is more amenable to mathematical
predictions, and there are many problems where this theory applies. Two areas of flow of fluids
where laminar theory applies are thin films and capillary tubes.
Laminar Flow
Laminar flow predictions depend on rheology of the fluid. A Newtonian fluid is one in which shear
stress is proportional to shear rate as expressed by
du
τ=μ 10 1
dy
where the constant, μ, defines the viscosity of the fluid. The rheology of non-Newtonian fluids can
often be modeled by Bingham or Power law models [1, 2].
Viscous Pumps
Viscous pumps move fluids between parallel surfaces by simple shear. Fluid flow is induced by one
surface moving parallel to the other. The pump is usually working against back pressure, so there is
a negative pressure-induced effect to be considered. A simple pump configuration and flow patterns
are illustrated in Figure 10.1. In this case
du U
τ=μ ≈μ 10 2
dy h
The velocity gradient across a thin film produced by the runner, is Uh. The shear stress is constant
across the film.
Back pressure also affects the velocity profile. The freebody diagram of a fluid element within the
thin film gives
d2 u dp
μ 2 =
dy dx
1 dp 2
uy = y + C1 y + C2
2μ dx
542 Fluid Mechanics
Fixed shoe
h
x
U
p pmax
x
Pressure induced
h Velocity induced
U
Figure 10.1 Mechanics of viscous pumps.
By integration and applying boundary conditions:

u 0 =U and u h =0
gives
1 dp 2 y
uy = y − hy − U −1 10 3
2μ dx h
For fluid pumps, film thickness, h, is constant [3]. The overall flow through the film from left to
right is the sum of both velocity and pressure-induced flows. Flow through the film (from left to
right) is determined by integration of the velocity profile. At any location x
h
Q = b u y dy
0
1 h3 dp
Q= Uh − flow rate per unit width 10 4
2 12μ dx
It is useful to define
1
Q = α hU 10 5
2
where “α” is a measure of the effect of back pressure on positive flow. It is also a measure of the
volumetric efficiency of the pump.
Laminar Flow 543
Noting that
dp p
= max 10 6
dx ℓ
Equating Eqs. (10.4) and (10.5)
1 1 h3 pmax
α hU = hU − 10 7
2 2 12μ ℓ
pmax 6μU
= 2 1−α 10 8
ℓ h
giving
pmax h2
α = 1− 10 9
ℓ 6μU
Again, α is used for the convenience of calculating flow, Q.
When
pmax = 0, α=1
pmax 6μU
= 2 , α=0
ℓ h
Example Consider the following operating conditions.

pmax = 200 psi
ℓ = 5 in
U = 10 ips
μ = 100 cp 100 × 1 45 × 10 − 7 = 1 45 × 10 − 5 reyn
h = 0 002 in
Without back pressure, flow rate is 0.01 in3/sec. With back pressure flow rate delivered by the
viscous pump is
1 0 002 3 200
Q= 10 0 002 − −5
2 12 1 45 × 10 5
Q = 0 01 − 18 39 × 10 − 4 = 0 008 16 in 3 s
For this case,

0 008 16
α= = 0 816
0 01
Force to Move Runner

The force, Fr, required to move the runner to the right is
ℓ
du
Fr = μ dx 10 10
dy y=0
0
544 Fluid Mechanics
η du 1 dp U
= − h− 10 11
dy y=0 2μ dx h
1/3
By Eq. (10.10)
h ℓ
F r = pmax + μU
2 h
α The mechanical power to move the runner at veloc-
0 2/3 1
ity, U, is
Figure 10.2 Mechanical efficiency vs.
PM = F r U 10 12
flow rate.
ℓ
PM = 4 − 3α μU 2 10 13
h
The hydraulic power delivered by the pump is
PH = pmax Q 10 14
ℓ
PH = μU 2 3α 1 − α 10 15
h
Pump efficiency [3]
PH p Q
η= = max 10 16
PM PM
3α 1 − α
η= 10 17
4 − 3α
Figure 10.2 shows how mechanical efficiency varies with flow rate indicated by α.
Maximum pump efficiency (33.3%) occurs at α = 23. Recall that α is the volumetric efficiency of
the pump, 1 Q
2Uh = α, with a value of 66.6% at maximum pump efficiency.
For the above example α = 0.816 and η = 29%, which is less-than optimum pump efficiency
of 33.3%.
A useful configuration of a viscous pump is shown
in Figure 10.3.
Qout Qin
Capillary Tubes
pmax p0
Summing forces on the fluid plug (Figure 10.4) and
assuming a Newtonian fluid gives
du
dpπr 2 = − 2πrμ dx 10 18
dr
ω du r dp
= − 10 19
R dr 2μ dx
h
The velocity distribution of fluid moving within
pipe follows:
1 dp 2
ur = R − r2 10 20
Figure 10.3 Viscous pump.
4μ dx
Laminar Flow 545
which shows that fluid velocity is r

maximum in the center (r = 0) and
zero at the inside wall of the pipe
(r = R). The derivation assumes New- τ u(r)
tonian fluid having viscosity of μ.
Using this expression for the velocity p p-dp x
profile and integrating across the
inside pipe area gives the Hagen–
Poiseuille equation [4, 5]. τ
π dp 4
Q= R 10 21 dx
8μ dx
Rearranging and letting Q = VA Figure 10.4 Freebody diagram of fluid plug inside pipe.
dp 8μ
= AV 10 22
dx πR4
The pressure drop in pipe for laminar flow conditions is
Δp 32μ
= 2 V 10 23
L D
Flow equations that relate flow rate to pressure gradient for non-Newtonian fluids (Bingham,
Power law models) are given in Ref. [1].
Flow Through Noncircular Conduits

The Navier–Stokes equations for flow through noncircular conduits reduces to
∂2 w ∂2 w 1 dp
+ = 10 24
∂x 2 ∂y 2 μ dz
This is Poisson’s equation which applies to a variety of problems, such as twisting of noncircular
shafts [6], squeeze films, flow through porous media, and others. There are many solutions readily
available.
Elliptical Conduit
For example, consider flow through an elliptical conduit (Figure 10.5). Some solutions fall into
place with ease. For example
1 dp a2 b 2 x2 y2
w x, y = + 2 −1
2μ dz a + b2
2 a 2
b
10 25 y
satisfies Eq. (10.24) and boundary conditions, w = 0,
around the outer edges. The flow rate through the
b
elliptical conduit is determined from a
x
Q= w x, y dx dy
z
While the fluid friction at the conduit walls is
found by Figure 10.5 Elliptical conduits.
546 Fluid Mechanics
dw
τ=μ
dn
where
dw ∂w dx ∂w dy
= +
dn ∂x dn ∂y dn
dw ∂w ∂w
= cos ϕ + sin ϕ
dn ∂x ∂y
dy
tan 90 − ϕ = −
dx
Also, when a = b, Eq. (10.25) simplifies to
− 1 dp 2
wr = R − r2 10 26
4μ dz
which agrees with Eq. (10.20).
Rectangular Conduit
The velocity distribution in a rectangular conduit (Figure 10.6) does not have a direct solution as
described above. The solution requires the method of separation of variables. Poisson’s equation
(Eq. (10.24)) still applies.
Observing the velocity w(x, y) will be symmetrical with respect to the y axis, we assume the solu-
tion to be of the form
∞
nπx
w x, y = cos Yn y 10 27
n = 1, 3, 5
2a
This function automatically satisfies boundary conditions, w(a, y) = 0 and w(−a, y) = 0. Yn(y)
defines how w(x, y) varies with y.
The constant, F, on the right side of the Eq. (10.24) is expressed in terms of a Fourier series to
factor out the cosine term (see Figure 10.7).
Since this function is even only, the cosine part the series is needed:
nπx
F x = an cos n = 1, 3, 5, …
2a
10 28
where
y 2a
1 nπx
a an = F cos dx 10 29
2a 2a
− 2a
and
b
1 dp
x F= 10 30
μ dz
This integration gives

z
2F 2a nπx a nπx 2a
an = sin − sin
Figure 10.6 Flow through rectangular conduits. 2a nπ 2a 0 2a a
Laminar Flow 547
−a +a
Figure 10.7 Even function.
4F nπ
an = sin n = 1, 3, 5
nπ 2
4F n−1
an = −1 2 10 31
nπ
Therefore,
∞
4 n−1 nπx
F x = F −1 2 cos 10 32
1, 3, 5
nπ 2a
Substituting Eqs. (10.27) and (10.28) into Eq. (10.24) gives

n2 π 2 4 n−1
Yn − 2
Yn = F −1 2 10 33
4a nπ
The solution to this equation is
nπy nπy 16Fa2 n−1

Y n y = An sinh + Bn cosh − 3 3 −1 2 10 34
2a 2a nπ
Observing that w(x, y) is symmetrical with respect to the x axis, An must be zero. Bn is determined
from w(x, b) = 0 and w(x, –b) = 0.
1 16Fa2 n−1
Bn = −1 2
cosh nπb
2a
n3 π 3
1 dp
Keep in mind that F = μ dz . Bringing everything together
nπy
16a2 dp 1 n−1 cosh nπx
w x, y = − −1 2 1− 2a
cos n = 1, 3, 5, … 10 35
π 3 μ dz n3 cosh nπb
2a
2a
When a = b, the solution gives the velocity distribution across a square conduit.
Unsteady Flow Through Pipe

Consider a situation where a pressure gradient is suddenly applied between entrance and exit of a
pipe. Every location within the pipe experiences the same pressure gradient. The fluid starts at rest
and time is required to develop the steady-state velocity profile given by Eq. (10.20).
548 Fluid Mechanics
The Navier–Stokes equation [7] for this case, describing start-up flow in the z direction is
∂v ∂p 1 ∂ ∂v
ρ = − +μ r 10 36
∂t ∂z r ∂r ∂r
The equation is expressed in terms of polar coordinates. Following the work of Szymanski [8], the
solution is expressed as the sum of a steady state apart and an unsteady part.
v r, t = vs r + vu r, t 10 37
The steady-state solution comes from the solution to
1 dp 1 d dv
= r 10 38
μ dz r dr dr
Using boundary conditions of
dv
=0 C1 = 0
dr r=0
1 dp 2
v R =0 C2 = − R
4μ dz
Therefore,
1 dp 2
vs r = − R − r2 10 39
4μ dz
The unsteady part of the solution is determined from
∂vu 1 ∂ ∂vu
ρ =μ r 10 40
∂t r ∂r ∂r
Using separation of variables and letting vu(r, t) = R(r)T(t), Eq. (10.40) becomes
1 d dR
ρRT = μ r T 10 41
R dr dr
μ
Separating the variable and letting ν = ρ
1 d λ2
1 T r dr r dR
dr
= = 0 10 42
ν T R
− λ2
Choosing (−λ2)
T + νλ2 T = 0 10 43
and
1 d dR
r + λ2 R = 0 10 44
r dr dr
d2 R dR
r2 2 +r + λ2 r 2 R = 0 10 45
dr dr
Let, x = λr and by substitution
Laminar Flow 549
d2 R dR
x2 +x + x2R = 0 10 46
dx 2 dx
The general expression for the Bessel equation is
d2 R dR
x2 2 +x + x 2 − n2 R = 0
dx dx
In our case, n = 0 and Eq. (10.46) is a Bessel equation of order 0. Its solution is the Bessel function:
R r = AJ 0 x 10 47
where J0(x) is a Bessel function of the first kind, whose expression is [9]
x2 x4 x6 k x 2k
J0 x = 1 − + − + −1 + k = 1, 2, 3, … 10 48
22 24 2 2
26 3 2
22k k 2
The plot of this function is shown in Figure 10.8.

dT
+ νλ2 T = 0 10 49
dt
dT
= − νλ2 dt 10 50
T
ln T = − νλ2 t + C 10 51
− νλ2 t
T t = Ce 10 52
Bringing everything together, the total solution becomes
1 dp 2
R − r 2 + AJ 0 λr e − νλ t
2
u x, t = − 10 53
4μ dx
1.2
0.8
0.6
Bessel function, J0
0.4
0.2
0
0 1 2 3 4 5 6 7 8
–0.2
–0.4
–0.6
Argument, x
Figure 10.8 Bessel function of the first kind.

550 Fluid Mechanics
Two unknowns in this equation are the arbitrary constant, A, and the eigenvalue, λ. These con-
stants are determined as follows:
uu R, t = 0 velocity is zero at the surface always
This means that J0(λR) = 0. The first root of this function is 2.4048; therefore, λR = 2.4048 or
λ = 2 4048
R . The second requirement is the initial condition uu(R, 0) = 0.
1 dp 2
0= − R + AJ 0 0 10 54
4μ dx
Since J0(0) = 1
1 dp 2
A= R 10 55
4μ dx
The total solution is
1 dp
R2 − r 2 − R2 J 0 λr e − νλ
2
u r, t = − t
10 56
4μ dx
This equation is put in dimensionless form
u r 2 r − 2 4048 2 τ
= 1− − J 0 2 4048 e 10 57
um R R
ν R2 dp
where τ = R2
t and um = − 4μ dx .
The velocity profiles shown in Figure 10.9 correspond to the following values of τ
(∞, 0.5, 0.3, 0.2, 0.1, 0.05). The velocity profile at τ = ∞, corresponds to the steady-state condition
discussed earlier. At τ = 0.05, the velocity profile shows the greatest velocity gradient and shear
stress being greatest at r = R, the boundary radius, with very little velocity gradient over the center
portion.
1
0.9
0.8
Velocity ratio, u/um
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
–1 –0.5 0 0.5 1
Radial position, r/R
Figure 10.9 Transition from unsteady to steady flow.

Hydraulics of Non-Newtonian Fluids
The hydraulics of oil well drilling is central to the rotary drilling method. A special drilling fluid is
pumped down drilling pipe to supply jetting action under drill bits to clean and cool them and then
carry the cuttings back to the surface for analysis and disposal. Drilling hydraulics is designed to
optimize hydraulic horsepower under the drill bit for best cleaning and drilling. Drilling cost in
terms of feet drilled per hour (rate of penetration, ROP).
Drilling hydraulics offers several aspects of fluid flow including non-Newtonian fluid flow, opti-
mizing hydraulics for maximum available hydraulic horsepower at the end of several thousand feet
of pipe, and the application of downhole drilling motors and turbines [2].
Hydraulics of Drilling Fluids

Drilling fluids, called drilling mud, are specially designed for variable density and gel strength prop-
erties. Mud density is needed to provide sufficient downhole pressure to control formation fluid;
barite is added for this purpose. Bentonite is added to give gel strength for suspending cutting pro-
duced by drill bits. These additives produce non-Newtonian fluids whose rheology are typically
describes by the Bingham Plastic model or the Power Law model. Much work has been done over
the years to relate the drilling variables into a mathematical model. It is essential to have the tech-
nology to predict fluid losses, which occur mainly inside drill pipe, which can be several thousand
feet long.
Pressure Loss Inside Drill Pipe

The equation for predicting pressure losses inside drill pipe (or any tubular) is [10, 11] becomes
Δp PV 0 2 γ 0 8 Q1 8
= 7 7 10 − 5 10 58
L D4 8
where
D – ID of pipe, in.
PV – plastic viscosity, cp
γ – mud weight, ppg
L – pipe length, ft
Δp – pressure drop, psi
This equation is commonly used within the petroleum industry.

Example Assume the following:

Plastic viscosity – 10 cp
Drill pipe size – 4½ in. (16.6 lb/ft) (ID = 3.826 in.)
Flow rate, Q – 400 gpm
Substituting directly into Eq. (10.58) gives

Δp
= 0 0634 psi ft
L
Pressure loss over 10 000 ft of pipe is predicted to be 634 psi.
552 Fluid Mechanics
Pressure Loss in Annulus

Pressure loss in the annulus is predicted by [10, 11]
Δp PV 0 2 γ 0 8 Q1 8
= 7 7 10 − 5 3 18 10 59
L Dh − Dp D h + Dp
Oil Well Drilling Pumps

Mud pumps have two distinct parts: (i) fluid end and (ii) power end (Figure 10.10). The fluid end
contains fluid manifold, intake and exhaust valves, pistons, and liner. The power end is a mechan-
ical drive containing a slider crank mechanism, which converts rotary motion and torque to linear
piston motion and force.
Mud pumps are positive displacement pumps. Volume output is directly related to the number of
strokes of the crank. Output capacity is normally expressed in terms of volume per stroke. Volume
output depends on liner size and stroke length. Actual flow rate output depends on volumetric effi-
ciency of the pump and suction intake.
Fluid ends are limited by fluid pressure in the manifold. Triplex pumps are limited to a maximum
discharge pressure of approximately 5500 psi. Higher manifold pressures can cause fatigue cracks
and premature fatigue failure.
Power ends are limited by force transmitted from the piston through the slider crank mechanism.
Bearings usually limit the maximum allowable force on the piston, and this force depends on fluid
pressure and piston area. Allowable fluid pressure in large liners is low (high volume). Allowable
pressure in small liners is typically high (small volume).
While pressure and volume output change with liner size, power input remains the same.
Hydraulic horsepower is related mathematically to pressure and flow rate by
Power = pQ
which is adjusted to horsepower units by
144 in 2 1 ft3 1 min 1 hp

HHP = pQ hp
1 ft2 7 48 gal 60 s 550 ft-lb s
pQ
HHP = hp
1714
p,Q
Liner
T, N
Motor
Manifold
Power end Fluid end
Figure 10.10 Schematic of mud pumps.

where
HHP – hydraulic horsepower, hp.

p – fluid pressure, psi
Q – volume flow rate, gpm
Performance of mud pumps is measured in terms of hydraulic horsepower output. Mud pumps
produce maximum hydraulic horsepower with different combinations of output pressure and flow
rate (Table 10.1). The constant horsepower curve is shown in Figure 10.11.
The rectangle that touches the horsepower curve represents pressure – flow rate capability the
pump. The horizontal line indicates the maximum operating pressure set by the manufacturer.
Drilling contractors usually operate below the maximum pressure. Flow rates listed in
Table 10.1 are limits for each of the liner sizes.
Table 10.1 Typical triplex mud pump performance features.
Liner (in.) pmax (psi) Fmax (lb) Flow rate (gpm) HHP (hp)
5½ 5558 132 048 481 1560

6 4665 131 879 573 1560
6½ 3981 132 090 681 1571
7 3423 131 717 778 1554
7½ 2988 132 010 894 1559
Figure 10.11 Pump performance vs.

liner size.
5½ in.
Pressure, p
Manufacturer 6in. liner
Contractor
6½ in.
Flow rate, Q
554 Fluid Mechanics
Drilling Hydraulics
The purpose of a hydraulic analysis is to establish available hydraulic horsepower at the bot-
tom of a drillstring to make best use of what is available to drive motors (or turbines), clean
drill bits, and carry cutting out of the well bore.
Example Consider the follow operating conditions [2].
Hole size – 97 8in

Hole depth – 8500 ft
Plastic viscosity – 10 cp
Drill pipe size – 4½ in. (16.6 lb/ft); ID = 3.826 in.
Drill pipe length – 7900 ft
Drill collar size – 8 × 3 in.
Drill collar length – 600 ft
Pump data:
6-in. liner
12-in. stroke
Liner pressure rating is 4670 psi
Operating pressure limit is 3000 psi
Surface equipment – Type 4
Step # 1 Calculate parasitic pressure losses for a flow rate of 400 gpm.
Step # 2 Calculate parasitic pressure losses for a range of flow rates (Figure 10.12).
Step # 3 Calculate and plot the hydraulic horsepower loss curve for each flow rate
(Table 10.2).
Step # 4 Superimpose the three hydraulic horsepower curves to obtain the available power curve
(Figure 10.13).
3500
3000
2500
Pressure, psi
2000
1500
Parasitic losses
1000
500
0
0 200 400 600 800 1000
Flow rate, gpm
Figure 10.12 Effect of flow rate on parasitic losses.

Table 10.2 Pressure losses and available hydraulic horsepower.
Flow p, p,
rate, pipe collar p, pipe p, collar Surface System
Q ID ID annulus annulus equipment losses Pump System Available
(gpm) (psi) (psi) (psi) (psi) (psi) (psi) HHP HHP HHP
100 42 10 1.4 1.7 1.7 57 175 3 172

150 87 21 2.9 3.5 3.4 117 263 10 252
200 145 35 4.8 5.8 5.7 197 350 23 327
250 217 53 7.2 8.7 8.6 294 438 43 395
300 301 74 10.0 12.1 11.9 409 525 72 454
350 398 97 13.2 16.0 15.7 540 613 110 502
400 506 123 16.8 20.3 20.0 686 700 160 540
450 625 153 20.8 25.1 24.7 848 788 223 565
500 755 184 25.2 30.4 29.9 1025 875 299 576
550 897 219 29.9 36.1 35.5 1217 963 391 572
600 1049 256 34.9 42.2 41.5 1424 1050 498 552
650 1211 296 40.3 48.7 47.9 1644 1138 624 514
700 1384 338 46.1 55.7 54.8 1879 1225 767 458
750 1567 383 52.2 63.1 62.0 2127 1313 931 382
800 1760 430 58.6 70.8 69.7 2389 1400 1115 285
850 1963 479 65.4 79.0 77.7 2665 1488 1321 166
900 2176 531 72.5 87.6 86.1 2953 1575 1551 24
950 2399 585 79.9 96.5 94.9 3255 1663 1804 −142
1000 2631 642 87.6 105.9 104.1 3570 1750 2083 −333
1600
1400
Hydraulic horsepower, hp
1200
1000
Pump Parasitic
800
600
Available
400
200
0
0 200 400 600 800 1000
Flow rate, gpm
Figure 10.13 Available hydraulic horsepower at bottom of drill string.

556 Fluid Mechanics
Note the difference between the maximum operating pump pressure (3000 psi) and system losses
at a given flow rate represents the pressure available at the bottom of the drill bit.
Pressure drop predictions listed in Table 10.2 were obtained using Eqs. (10.58) and (10.59).
The straight line shows how the hydraulic horsepower supplied by the pump varies with flow
rate. This line is based on the maximum operating pump pressure of 3000 psi. For example, the
fluid power supplied by the pump at 400 gpm is
3000 400
HHPP = = 700 hp
1714
while the fluid power consumed by friction at 400 gpm is
686 400
HHP f = = 160 hp
1714
which agrees with Table 10.2.

Note that the flow rate limit of one pump is 573 gpm, which is slightly higher than the optimum
flow rate.
The example shows that the maximum available hydraulic power of 590.85 hp is achieved at a
flow rate of 525 gpm (see Figure 10.13). The system pressure loss at 525 gpm is 1071 psi, which
is 1/2.8 of the maximum allowable operating pressure of 3000 psi. This also means that 589
HHP (1929 psi at 525 gpm) is available for motor, turbine, or drill bit.
The annular velocity between drill pipe and well bore is determined from
Q 144
V an =
π
D2h − D2dp 7 48
4
Q
V an = 24 51 10 60
D2h − D2dp
where

Dh – well bore diameter, in.
Ddp – OD of drill pipe, in.
Both HHPa and Van are shown in a composite diagram (Figure 10.14). The selection of the best
flow rate is a compromise between highest HHPa and annular velocity requirements.
Power Demands of Downhole Motors
Analytical tools are useful for configuring a motor for a given set of operational specifications. How-
ever, after a tool is built its true performance is determined by laboratory testing. Test results estab-
lish how a given motor will perform in the field. A typical test arrangement is shown schematically
in Figure 10.15.
700
600
500
Horsepower, hp
400
300
200
100
0
0 200 400 600 800 1000
Flow rate, gpm
250
Vav , fpm
Figure 10.14 Composite diagram of HHP and annular velocity.
Fluid, usually water, is pumped through the tool and each of the above parameters are measured.
The dynamometer generates torque on the drive shaft and sensors measures each parameter for
several torque settings. Performance tests are made for each size and model.
Performance of Positive Displacement Motors (PDM)

Typical performance data for positive displacement motors (PDMs) is shown in Figure 10.16. Pres-
sure differentials for various dynamometer torque settings are recorded at constant flow rates.
Applied torque and rotational speeds are recorded. When torque is zero, rotational speed is a max-
imum. Pressure differential is required to overcome internal friction but is lowest under no torque
conditions. When torque is increased, rotational speed drops off slightly because of fluid leakage
through each stage. Higher torque requires higher differential pressures and causes a reduction
in rotational speed due to leakage.
When applied torque is taken to the extreme, PDMs stall out. This means that the pressure level is
great enough to force fluid through motors without rotation. Fluid leakage is 100%. Even though
torque is a maximum under stall conditions, power output is zero because the output shaft does not
turn. Therefore, operating a PDM near stall conditions produces little or no power to the drill bit.
558 Fluid Mechanics
1 2
T, N
Q
p1 p2 Dynamometer
PDM – turbine
Force
Figure 10.15 Performance testing parameters.
Stall torque
Constant flow rate, Q

Torque, T (ft-lb)
Δp
Maximum operating pressure
10% reduction in Q
Speed, N (rpm)
Figure 10.16 Typical performance diagram for Moineau type motors.
PDMs are normally rated at torque and pressure levels corresponding to about a 10% leakage
through the motor.
PDMs operate between recommended maximum and minimum flow rates. The maximum flow
rate is based on a maximum allowable whirl velocity of the rotor. The rotor is constantly undergoing
accelerations which generate forces on rubber undulations in stators. This could cause premature
failure of the rubber. Minimum flow rate is based on inefficiency, which is typical at low flow rates.
Based on these guidelines, laboratory performance data can be approximated as shown in
Figure 10.17. The numbers are typical of a 1 : 2 PDM. The performance of multiple lobed PDMs
deviates somewhat from this model. Performance data for a variety of motor sizes are given in
Table 10.3.
It is important to note that maximum recommended operating torque and pressure levels are not
stall conditions. Stall torque is approximately 50% higher than recommended operating torque.
The capability to monitor performance of PDMs from the surface is a distinct advantage. Pressure
drop across these motors can be monitored by standpipe pressure and output speed can be mon-
itored by pump flow rate.
Figure 10.17 Performance diagrams for a 63/4 Δpmotor

Moineau motor. (psi)
580
Q, gpm
200 475
HHP
160.7
67.68
Q, gpm
200 475
Table 10.3 Positive displacement motor performance data.
Pump
rate (gpm) Bit speed (rpm)
Tool Max Torque Power
size (min) (max) (min) (max) diff. (psi) (ft-lb) range (HP) HHP
33/4 75 85 280 700 580 385 20–51

43/4 100 240 245 600 580 585 27–67
6 /4
3 200 475 205 485 580 1500 59–138 (67.7–160.7)
8 245 635 145 380 465 2090 59–152
9½ 395 740 195 365 695 3890 145–271
560 Fluid Mechanics
PDMs are a spin-off from the Moineau pump [12]. This pump, sometimes called a progressive
cavity pump, is still in use today for transporting: food, oil, sewage, and viscous chemicals. Fluid is
trapped within cavities (characteristic of the Moineau pump), which progress axially from
the intake to the outlet end. Pressure within each cavity may vary because of leakage between
cavities
PDMs have a direct relationship between output speed and through-put flow rate. This charac-
teristic makes it possible to monitor output speed by monitoring flow rate:
N = CQ 10 61
where N is rotational speed (rpm), Q is flow rate through the motor (gpm), and C is a constant
depending on geometry of rotor and stator. The proportionality constant, C, is determined from
maximum rotational speed and corresponding flow rate data from the manufacture’s performance
information:
N N max
C= = 10 62
Q Qmax
For the 63/4 motor performance data given in Table 10.3,
485
C= = 1 02
475
Rotational speed and flow rate are related by
N = 1 02Q
The motor’s rotational speed, therefore, can be monitored by the flow rate.
Application of Drilling Turbines

The successful application of downhole turbines depends on the amount of hydraulic horsepower
available to drive a turbine. It is therefore important to establish the amount of available hydraulic
horsepower at the bottom of the drillstring. The available hydraulic horsepower varies with flow
rate. Hydraulic horsepower consumed by turbines also varies with flow rate and increases with
the third power of flow rate.
For example, consider the turbine having performance data as listed in Table 10.4. The amount of
hydraulic horsepower consumed by this turbine depends on both flow rate, Q, and pressure drop,
Δp, according to
Table 10.4 Hydraulic horsepower consumed 63 4 in. turbine (12 ppg) (HHP is
developed from Eq. (10.64)).
Flow rate (gpm) Pressure drop (psi) Hydraulic horsepower (hp)
250 621 90.6

300 894 156.5
350 1217 248.5
400 1589 370.8
450 2011 527.9
500 2483 724.3
QΔp
HHP = 10 63
1714
Since pressure drop across turbines is independent of the amount of mechanical power delivered,
turbines consume the calculated hydraulic horsepower level, whether the drill bit is off or on bot-
tom. Mechanical output power varies considerably with bit torque, however.
When this type of data is plotted, the power consumption curve for turbines can be compared
with the available hydraulic horsepower to see if there is enough power to drive the turbine. To
work turbines to the fullest, turbines should be run at high flow rates. This usually means that
all the available horsepower is consumed by turbines, and there is little or no hydraulic horsepower
left to clean the drill bit. In some cases, open flow diamond bits can be cleaned and cooled by high
flow rates without regard to HHP when used with turbines.
Hydraulic Demands of Drilling Motors – Turbines

Typical PDM power consumption is shown in Figure 10.18 along with the available HHP curve to
show how much power is consumed by a PDM. The difference between to two curves represents the
HHP available for bit cleaning. This difference is the basis for selecting nozzle size for best bottom-
hole cleaning. This type of diagram is useful for determining best flow rate, too. Flow rate selection
also should consider annular velocity requirements.
Successful application of downhole turbines depends on the amount of hydraulic horsepower
available at the bottom of the drillstring. Hydraulic analysis of the total circulating system is critical.
Figure 10.19 shows an available hydraulic power curve and turbine power consumption curve,
assuming the turbine operates at maximum power output. Note that the power consumption curve
is basically different from the PDM power curve.
Turbines usually operate at high flow rates to make best use of the turbine’s power capability.
This usually means that all the available hydraulic horsepower is consumed by turbines, and there
is little or no hydraulic horsepower left for bit cleaning. This condition is represented by the inter-
section of the turbine power with the available power curve. For this reason, open drill bits (no
nozzles) are often used with turbines. In this case, bit cleaning and cooling are accomplished by
high flow rates.
600
500
Hydraulic horsepower
Available
400
300
200
PDM
100
0
0 200 400 600 800
Flow rate, gpm
Figure 10.18 Typical PDM power consumption.

562 Fluid Mechanics
700
600
Hydraulic horsepower
500
Available
400
300
Turbine
200
100
0
0 200 400 600 800
Flow rate, gpm
Figure 10.19 Power consumption of drilling turbine.
Fluid Flow Around Vibrating Micro Cantilevers
Measurements of fluid viscosity have enormous applications in areas ranging from clinical diagnos-
tics to industrial and engineering applications. Since determination of absolute viscosities is a chal-
lenging problem, most applications exploit relative variation in viscosities. Viscosity
measurements, quite often, require bulky experimental apparatus and relatively large volumes
of samples. Measuring viscosities of very small volume samples using miniature devices are very
attractive. Although resonance responses of cantilevers, the size of a meter, is insensitive to viscos-
ity of surrounding medium, fluids significantly influence the resonance response of a microfabri-
cated cantilever.
Consider the Navier–Stokes equations for flow through a control volume located above a canti-
lever surface (Figure 10.20). The solution is given in closed form and links viscosity directly to the
damping factor, which can easily be extracted from experimental frequency response data [13].
∆z
2a
Figure 10.20 Microcantilever with coordinate system.
Mathematical Model
Key assumptions made in the mathematical development of fluid flow adjacent to microcantilevers
surfaces are
1) Fluid flow is laminar.

2) Inertia forces within the flow pattern are small in comparison with viscous forces and are there-
fore not included in the hydrodynamic analysis. This is consistent with Reynolds basic equation
of lubrication and squeeze film vibration damping theory. The mass of the fluid on and near the
surfaces of microcantilevers does, however, affect the natural frequency of cantilevers because of
the small dimensions and mass density of microcantilevers. The effective mass is best deter-
mined experimentally. It is a separate determination and not an integral factor in our hydrody-
namic analysis.
3) Fluid flow pattern depends only on the instantaneous velocity of the cantilever surface. This is
consistent with assumption (2). This assumption allows viscous forces and viscosity to be related
to surface velocity, damping coefficients, and damping factors in a simple and easy to use
formula.
4) Actual hydrodynamic fluid flow near a cantilever surface is approximated with flow within a
rectangular control volume.
5) Fluid flow is two-dimensional. There is no fluid flow along the length of the cantilever.
The coordinate system used in the development is illustrated in Figure 10.21. A typical surface
section of length, Δz, and width, 2a is taken as the base of the control volume. The pressure func-
tion, p(x, y), and velocity function, u(x, y) vary with x and y and are assumed constant over the unit
distance, Δz.
The predicted damping coefficient, c, will have units of force per velocity per unit length of the
microcantilever. The control volume under consideration is defined by a rectangle have dimensions
of (2a × h × Δz). Fluid flows into and out of this control volume to maintain continuity of flow.
Fluid is assumed to be incompressible.
The simplified Navier–Stokes equations governing fluid flow within the control volume are
∂p ∂2 u ∂2 u
=μ + 10 64
∂x ∂x 2 2y2
∂p ∂2 v ∂2 v
=μ + 2 10 65
∂y ∂x 2 2y
Figure 10.21 Cross section of micro cantilever with

control volume.
Control
volume
y
h
a a
x
Cantilever cross section

564 Fluid Mechanics
These equations, plus the continuity equation,

∂u ∂v
+ =0 10 66
∂x ∂y
are satisfied along with appropriate boundary conditions, which are assumed to be
p a, y = 0
p − a, y = 0
p x, h = 0
πx
p x, 0 = ps cos ps is stagnation pressure
2a
Velocity related boundary conditions are
u x, h = 0
u x, 0 = 0
u 0, y = 0
These boundary conditions approximate the actual flow pattern above and around the top sur-
face. They give reasonable results as will be shown.
Fluid Pressure Formulation

The simplified Navier–Stokes equations and continuity equation can be combined to produce the
Laplace equation with pressure, p(x, y), as the dependent variable:
∂2 p ∂2 p
+ =0 10 67
∂x 2 ∂y2
Using the separation of variables technique and applying the first two pressure boundary condi-
tions, gives
πx πy πy
p x, y = cos C cosh + D sinh 10 68
2a 2a 2a
The remaining two boundary conditions at y = 0 and y = h yield
πh
πx πy cosh πy
p x, y = ps cos cosh − 2a sinh 10 69
2a 2a πh 2a
sinh
2a
The shape of the predicted pressure profile depends on the ratio, h/a. By trying different values of
h/a, the pressure slope at y = h becomes essentially zero when h/a = 5. This means that the pressure
above the cantilever’s surface blends into the atmospheric pressure at this height (h = 5a) above the
cantilever. For example, if the width of a microcantilever is 30 μm, then film pressure extends up to
a height of about 5 × 15 or 75 μm with significant pressure occurring out to about y = ½h or 37.5 μm
(see Figure 10.22).
Fluid Velocity Formulation 1

When the pressure function is applied to the x 0.9
component of the Navier–Stokes equations, we
obtain 0.8
∂2 u ∂2 u p π πx 0.7
Vertical position, y/h

+ = − s sin
∂x 2 ∂y 2 μ 2a 2a 10 70 0.6
πy πy
cosh − D sinh
2a 2a 0.5
where 0.4
πh
cosh 2a 0.3
D= πh
sinh 2a
0.2
To solve Eq. (10.70), we start by assuming
0.1
πx
u x, y = sin gy 10 71
2a 0
0 0.2 0.4 0.6 0.8 1
The substitution of Eq. (10.71) into Eq. (10.70)
Pressure ratio, p/ps
produces the following ordinary differential
equation: Figure 10.22 Pressure profile at x = 0 (h/a = 5).
dg 2
π 2 p π πy πy
2 − g= − s cosh − D sinh
dy 2a μ 2a 2a 2a
10 72
Through trial and error, we found that the following function satisfies Eq. (10.72) and is therefore
a particular solution:
1 ps πy πy
gp y = − y sinh − D cosh 10 73
2 μ 2a 2a
The complimentary solution to
d2 g π 2
2 − g=0 10 74
dy 2a
is
πy πy
gc y = E cosh + F sinh 10 75
2a 2a
The total solution to Eq. (10.72) is the sum of gp and gc or
πy πy 1 ps πy πy
g y = E cosh + F sinh − y sinh − D cosh 10 76
2a 2a 2 μ 2a 2a
The constants of integration, E and F, are determined by applying the following boundary
conditions:
g 0 =0
g h =0
566 Fluid Mechanics
1 The first boundary condition gives, E = 0. The

second boundary condition gives,
0.9
1 ps
0.8 F= h 1 − D2
2 μ
0.7
Vertical position, y/h
Applying these results to Eqs. (10.76) and

0.6 (10.71) gives
0.5 1 ps πx πy
u x, y = sin h 1 − D2 sinh
0.4 2 μ 2a 2a
0.3 πy πy
− y sinh − D cosh 10 77
2a 2a
0.2
This equation can be formatted in terms of
0.1
useful dimensionless groupings as follows:
0
0 0.05 0.1 0.15 0.2 0.25 1 ps a πx h π h y
u x, y = sin 1 − D2 sinh
Velocity ratio 2 μ 2a a 2 a h
h y π h y π h y
Figure 10.23 Velocity profile at x = a (h/a = 5). − sinh − D cosh
a h 2 a h 2 a h
10 78
Equation (10.78) is plotted in terms of dimensionless parameters (see Figure 10.23), where the
velocity ratio is defined by
2μu
velocity ratio = 10 79
ps a
This graph shows how the x component of the velocity varies from the cantilever surface (y = 0) to
the top edge of the control volume (y = h) along the line, x = a. A value of h/a = 5 was also chosen
for this plot, which shows the velocity gradient at y = h to be essentially zero as expected. Also, we
note that the maximum horizontal velocity occurs at y/h equal to 0.125 or ⅛th of h. If, for example,
the cantilever width is 50 μm, the maximum velocity occurs at 0.125(5 × 25) = 16.625 μm above the
cantilever surface.
References
1 Dareing, D.W. and Dayton, R.D. (1992). Non-Newtonian behavior of power lubricants mixed with
ethylene glycol. STLE Tribol. Trans. 35 (1): 114–120.
3 Cameron, A. (1966). The Principles of Lubrication. Longmans Green and Co., Ltd.
4 Hagen, G. (1839). On the motion of water in narrow cylindrical tubes. Poggendorff’s Ann. Phys. Chem.
46: 423.
5 Poiseuille, J.L.M. (1843). Recherches Experimentales sur le Mouvement des Liquides dans les Tubes
de Tres Petits Diametres. Acad. Sci. Savants Etrangers 9.
6 Timoshenko, S. and Goodier, J.N. (1951). Theory of Elasticity, 2e. New York: McGraw-Hill.
7 Schlichting, H. (1960). Boundary Layer Theory. McGraw-Hill Book Company.
References 567
8 Szymanski, F. (1932). Quelques solutions exactes des equations de l’hydrodynamicque de fluide

visqueux dans le cas d’um tube cylindrique. J. de math. pures et appliquees, Series 9 11: 67; see also
Intern. Congr. Appl. Mech. Stockholm I, 249 (1930).
9 Reddick, H.W. and Miller, F.H. (1956). Advanced Mathematics for Engineers, 3e. NY: Wiley.
10 Moore, P.L. (1974). Drilling Practice Manual. Tulsa, Oklahoma: Penn Well Books.
11 Kendall, H.A. and Goins, W.C. (1960). Design and operation of jet-bit programs for maximum
horsepower, maximum impact force and maximum jet velocity. Trans. AIME 219.
12 Moineau, R.J.L. (1932). Gear mechanism. US Patent 1,892,217, 27 December.
13 Dareing, D.W., Yi, D., and Thundat, T. (2007). Vibration response of microcantilevers bounded by
confined fluid. Ultramicroscopy 107: 1105–1110.
569
11
Energy Methods
There are two approaches to engineering mechanics: (i) Newtonian mechanics and (ii) energy
methods. Both approaches lead to the same results; however, in some cases, energy methods
can give a more direct solution. Earlier chapters focused on Newtonian mechanics. Newtonian
mechanics requires the use of force and acceleration vectors. Energy methods requires only the
tracking of energy, a scalar quantity.
Principle of Minimum Potential Energy
This principle states that: “A motionless conservative system is in stable equilibrium if its potential
energy is a relative minimum” [1]. A conservative system is one in which the change in potential
energy is independent of the path taken in going from on configuration to another.
Total potential energy is defined as
V =U+Ω 11 1
where
V – total potential energy

U – potential of internal forces (elastic energy)
Ω – potential of external forces (applied forces)
Both potentials can be viewed as work one performs to achieve the virtual displacement.
The concept of this principle is illustrated in Figure 11.1, which depicts a bead on a curved wire in
three different positions. In each of the three positions, intuitively we note the bead is in equilib-
rium. However, only the first position represents stable equilibrium. In this case, V(x), is a function
of one independent variable; the system has one degree of freedom.
Stable and Unstable Equilibrium

Let us examine the change in potential about x = a when the bead is given a virtual displacement, h.
Expanding V(x) in a Taylor Series about point “a” gives
h2 h3
V a + h = V a + hV a + V a + V a + 11 2
2 3
570 Energy Methods
V
Unstable
Unstable
Stable
h
x
a
Figure 11.1 Bead on a curved wire.
The change in potential energy, ΔV, due to the virtual displacement, h, is
ΔV = V a + h − V a
h2 h3 hm − 1
ΔV = hV a + V a + V a + + Vm−1 a + 11 3
2 3 m−1
which may be written
1 2 1
ΔV = δV + δV+ + δ m−1 V + 11 4
2 m−1
where δk is the kth variation of the total potential energy, V.
δk V = hk V k a
If h is sufficiently small, ΔV has the same sign as the first nonzero term in the series. If the value of
V is a relative minimum, the first nonzero term is an even term, because an odd term reverses its
sign when h is reversed. Therefore, for V to have relative minimum at x = a.
1) The first nonzero variation δkV is of even order

2) δkV 0, i.e. Vk(a) 0
Condition (1) implies that δV = 0. In other words, for the illustration (Figure 11.1)
dV d2 V
First location − =0 and 0 stable equilibrium
dx dx 2
dV d2 V
Second location − =0 and ≺ 0 unstable equilibrium
dx dx 2
dV d2 V
Third location − = 0 and =0 unstable equilibrium
dx dx 2
Stability of Floating Objects

Consider a wood block floating in water (Figure 11.2). Since the density of wood is less than the
density of water, part of the block is above the water line and part below. The center of gravity
Water line
c b
B
mc
B
Water line
c θ
b
W
Water line
V
c
b
θ
𝜋
B 0
2
Figure 11.3 Potential energy of a

Figure 11.2 Stabilizing moment on a floating block. floating block.
of the block is indicated by point “c,” while the center of buoyancy is indicated by point “b.” From
experience and intuition, we expect the block to float horizontally as shown. Any effort to rotate the
block brings about a corrective action that tends to put the block horizontal again. What causes this
to happen is the shifting of vector B away from vector W, which brings about a corrective couple
(torque) as shown.
Another parameter shown in the rotated figure is the metacenter, “mc.” The distance between c
and mc is the metacentric height. The metacentric height varies with rotation angle, but over small
angles of θ its change is small and often assumed constant. The height is a measure of how fast the
block (or ship) will return to its horizontal position. Floating objects with long metacentric heights
return faster.
The potential energy of this floating object is illustrated in Figure 11.3. When the block is flat, it is
stable. When the block is vertical, it is unstable.
572 Energy Methods
P Stability of a Vertical Rod

As another example, consider the problem shown in
Figure 11.4 pinned at the lower end but constrained
by a torsional spring, kθ. The bar of length, L, is rigid.
V θ =U θ +Ω θ 11 5
L (1– cos θ)
θ where
L 1
L/2 (1– cos θ) U= k θ θ2 11 6
2
W Ω = − P + W 2 L 1 − cos θ 11 7
L/2
Note that in this example “you” do plus work on the
elastic spring and minus work on gravity and concen-
kθ trated force, P.
Expanding
Figure 11.4 Stability of a vertical column. θ2 θ 4 θ6

cos θ = 1 − + − + 11 8
2 4 6
And combining terms
θ2 θ4 θ 6
V θ = k − PL − WL 2 + PL + WL 2 − + 11 9
2 4 6
V θ = k − PL − WL 2 θ + PL + WL 2 θ3 − θ5 + 11 10
V θ = k − PL − WL 2 + PL + WL 2 3θ2 − 5θ4 + 11 11
δV 0 = 0 11 12
δ2 V 0 = k − PL + WL 2 first nonzero term is even 11 13

kθ W
For stability, V (0) − . 0, therefore, Pcr =
L 2
The plot in Figure 11.5 shows the effects of the magnitude of applied force, P, on total potential
energy, V, is calculated from Eq. (11.5).
5
4 P = 5 lb
Potential energy, (in.lb)
3
P = 10 lb
2
1
P = 15 lb
0
–10 –5 0 5 10
–1
P = 20 lb
–2
Displacement angle, θ (°)
Figure 11.5 Potential energy of vertical column.

Input data for the potential energy calculations are
L = 24 in
W = 10 lb
k θ = 500 in -lb rad
The potential energy curves show a strong return to the neutral position for the lower end loads.
When P = 20 lb, the potential curvature is negative, indicating unstable equilibrium at θ = 0.
The predicted critical force for this set of numbers is
kθ W 500 10
Pcr = − = − = 15 83 lb 11 14
L 2 24 2
This condition is represented by the horizontal line.
Rayleigh’s Method
This approach is based on assuming a reasonable displacement function, y(x). Consider, for exam-
ple, a cantilever beam supported transversely at the top by a spring. A force of magnitude P is
applied at the top as shown (Figure 11.6).
Following the Rayleigh method, we assume the cantilever will buckle into a configuration
approximated by
πx
y = y0 1 − cos 11 15
2L
The unknown parameter is y0. It is important that the assumed displacement satisfies the bound-
ary conditions.
The total internal energy of the system is
L
1 1 2
U= EI y
2
dx + ky 11 16 x
2 2 0 P
0 y0
where
k
π 2 πx
y = y0 cos
2L 2L
The potential energy of the external force, P, is
L
1 2 L
Ω= − P y dx 11 17
2 EI
0
π πx
y = y0 sin
2L 2L
L
1 π πx 2
Ω= − P y0 sin dx y
2 2L 2L
0
1 2 π π Figure 11.6 Rayleigh’s method of

Ω Py predicting buckling.
2 0 2L 4
574 Energy Methods
Total potential energy then is

V =U+Ω
1 π 3 π 1 1 π π
V= EIy20 + ky20 − Py20 11 18
2 2L 4 2 2 2L 4
dV
= 0 at y0 = 0; therefore, the system is in equilibrium at y0 = 0
dy0
d2 V
is plus only when
dy20
π 2 8L
P ≺ EI +k
2L π2
Therefore, the system is unstable when
π 2 8L
P ≥ EI +k 2
2L π
So, the critical value of P is
π 2 8L
Pcr = EI +k 11 19
2L π2
π 2
If the spring is removed, the buckling force becomes Pcr = EI, which agrees with the earlier
2L
discussing of Euler buckling.
Multiple Degrees of Freedom
The application of the principle of minimum potential energy generally follows the steps below.
1) Set up generalized coordinates at joints, taking advantage of symmetry whenever possible.

2) Express elongation of each member in terms of the generalized coordinates.
3) Write expressions of U and Ω in terms of the generalized coordinates; V = U + Ω.
∂V
4) = 0, x i − generalized coordinate
∂x i
5) Solve for generalized coordinate.
6) Substitute to obtain elongation (e) of each member.
EA
7) Axial force in each member is then determined from T = e.
L
Structure Having Two Degrees of Freedom

Example Consider a structure made of three members as shown in Figure 11.7. The problem is to
determine displacement components u,v due to a horizontal force of 100 lb. The cross-section
properties are:
EA =C
a
1
EA b = C
2
3
EA c = C
2
u
100 lb
v
a c
b H
45° 60°
Figure 11.7 Deflection components.
This problem is a statically indeterminate but is solved conveniently with the energy method.
Internal elastic energy is expressed in terms of the two displacement components.
1
U= e2i k i 11 20
2
where
ei = q ni
q = − ui + vj
1 1 1
na = i+j nb = − i+ 3j nc = i − 3j
2 2 2
E i Ai
ki =
Li
The axial displacements of each of the three members are
1
ea = − u+v
2
1
eb = − u + 3v
2
1
ec = u − 3v
2
By substitution
2 2 2
1 1 1 1 1 1
U= ka − u+v + kb − u+ 3v + kc u − 3v 11 21
2 2 2 2 2 2
The potential energy of the external force, 100 lb is
Ω = − 100u 11 22
Total potential energy is
V =U+Ω 11 23
For equilibrium
∂V 1 1 1
= ka u + v + kb u + 3v + k c u − 3v − 100 = 0 11 24
∂u 2 4 4
576 Energy Methods
∂V 1 1 1
= ka u + v + kb 3u + 3v + k c − 3u + 3v = 0 11 25
∂v 2 4 4
where
C C
ka = =
La H 2
C C 3C
kb = = 2 =
2Lb 2 3H 4H
3C 3C 3 3C
kc = = 2 =
2Lc 2 3H 4H
Giving
H
0 785u − 0 021v = 100
C
− 0 021u + 1 649v = 0
Simultaneous solutions give
H H
u = 127 3 and v = 1 62
C C
By substitution
H
ea = − 91 1 and T a = k a ea = − 64 5 lb
C
H
eb = − 64 9 and T b = kb eb = − 28 2 lb
C
H
ec = 62 and T c = k c ec = + 80 5 lb
C
Analysis of Beam Deflection by Fourier Series

The principle of minimum potential energy along with the use of Fourier series will now be applied
to beams. The first case will be a simply supported beam supporting a concentrated force
(Figure 11.8). The pinned boundaries suggest the use of the sine (or odd function) in the series.
Recall the internal energy (strain) is expressed in terms of beam deflection by a Fourier series in
the form of
L 2
1 d2 y
U = EI dx 11 26
2 dx 2
0
a Figure 11.8 Fourier series solution for

concentrated beam load.
F
x
EI
L
y
∞
nπx
yx = an sin 11 27
n=1
L
∞
nπ 2 nπx
y = − an sin 11 28
n=1
L L
By substitution
L
1 π 4 nπ
U = EI n4 a2n sin2 x dx 11 29
2 L L
0
∞
EIπ 4
U= n4 a2n 11 30
4L3 1
Concentrated Load
For the work done by the external force, F
Ω = − Fy a
where
∞
nπa
ya = an sin 11 31
1
L
Plus (+) y is taken downward, so displacement y is (+). Total potential energy is
V =U+Ω
∞
EIL nπ 4 nπa
V= a2n − Fan sin 11 32
1
4 L L
dV
For minimum potential energy = 0, which gives
dan
4
2F L nπa
an = sin 11 33
EIL nπ L
and
∞ 4
2F L nπa nπx
yx = sin sin 11 34
EIL 1
nπ L L
Distributed Load
The approach in representing beam deflection with a Fourier series for distributed load follows the
same process as for the concentrated load (Figure 11.9). The expression for the strain energy
remains the same. However, the work done by the external forces differs. Again, based on pinned
boundary conditions
∞
nπx
yx = an sin
1
L
578 Energy Methods
Figure 11.9 Fourier series solution for distributed

w(x) loading.
x
EI
x dx
L
y
The potential energy of the external distributed load is

dΩ = − w x dxy x
which leads to
L
∞
EIL nπ 4 nπx
V= a2n − an w x sin dx 11 35
1
4 L L
0
dV
For stationary potential energy, = 0, which gives
dan
L
4
2 L nπx
an = w x sin dx 11 36
LEI nπ L
0
L
∞ 4
2 L nπx nπx
yx = w x sin dx sin 11 37
LEI 1
nπ L L
0
Axially Loaded Beam (Column)

We now consider the Euler buckling problem as shown in Figure 11.10.
The post buckle displacement, e, of the applied force, P, is determined as follows.
ds2 = dx 2 + dy2
2
dy
ds = 1+ dx
dx
dx
P
P x
ds EI
Lʹ e
L
y
Figure 11.10 Axial buckling load.

L L
2
dy
L= ds = 1+ dx 11 38
dx
0 0
L L
2
dy 1 2
L≈ 1+ dx ≈ L + y dx
dx 2
0 0
L
2
e = L−L = y dx
0
L
1 2
ΩP = − P y dx 11 39
2
0
∞
nπ nπx
y = an cos
1
L L
π2 P
ΩP = − n2 a2n
4L
The total potential energy is
π 4 EI π2P
V= n4 a2n − n2 a2n 11 40
4L3 4L
∂V
For stationary potential energy, = 0.
∂an
π 4 EI π2P
2n4 an − 2n2 an = 0 11 41
4L3 4L
∂2 V
For stability, 0; therefore for n = 1, stability occurs only when
∂a2n
π 2 EI
P
L2
So
π 2 EI
Pcr = 11 42
L2
which is Euler’s buckling equation.
Principle of Complementary Energy
This principle is derived from the theory of stationary potential energy, which requires that
∂V
=0 11 43
∂x i
for equilibrium. This means equilibrium is established when
580 Energy Methods
∂U ∂Ω
= − 11 44
∂x i ∂x i
The work of external forces over a displacement of dxi is
dΩ = − P1 dx 1 + P2 dx 2 + P3 dx 3 + P4 dx 4 + 11 45
Following the chain rule of differentiation
∂Ω ∂Ω
dΩ = dx 1 + dx 2 + 11 46
∂x 1 ∂x 2
From Eqs. (11.45) and (11.46)
∂Ω
Pi = − 11 47
∂x i
It follows from Eq. (11.44) that
∂U
= Pi 11 48
∂x i
This equation is not restricted to small deflections nor limited to systems with linear elastic
properties.
A.M. Legendre (1752–1833) showed that Eq. (11.48) can be transformed into a conjugate form [1]
∂γ
= xi 11 49
∂Pi
where γ is complementary energy per volume as illustrated in Figure 11.11. Complementary energy
is the shaded area.
Force displacement relationships typically are given in terms of stress and strain. In this case
dγ = dσ ε
P x
dγ = d
A L
Engineering Application
The application of Eq. (11.49) to trusses is summarized below. P
The total complementary energy for a truss is
γ= γi , γ i − complementary energy of ith member

ϒ
11 50
dP
Let P be an external force acing on a joint and x be the com-
ponent of the displacement of the joint in the direction of U
P. Then
∂γ i x
x= 11 51 x
∂P
where γ i = f(Ni) and Ni – tension in ith member. Figure 11.11 Complementary
energy.
dγ = dPx
γ= x dP
It is not necessary to determine γ i for each member because we use only Eq. (11.51) as
shown below.
Accounting for energy throughout the structure
dγ i ∂N i
x= 11 52
dN ∂P
where
dγ i
= ei according to Eq 11 49 11 53
dN
ei – extension of ith member due to loads (Ni) in the ith member.
∂N i
x= ei = ei ni 11 54
∂P
For statically determinant trusses, the total force in any member is the superposition of the effects
of all forces applied throughout the structure.
N i = a1 P1 + a2 P2 + a3 P3 + + aP 11 55
By differentiation
∂N i
= a = ni 11 56
∂P
The interpretation of ni is therefore the tension in ith member if P = 1 and all other loads are
removed.
30 in.
40 in. 2 3 4
5 6
P, kip
Figure 11.12 Truss problem.

582 Energy Methods
Table 11.1 Deflection analysis (area = 0.5 in.2, P = 4 kip).
Member Ni (kip) Stress (psi) Strain (in./in.) Length (in.) ei (in.) ni (kip) eini
1 6 12 0.000 216 30 0.006 48 1.5 0.009 72

2 −5 −10 −0.000 13 50 −0.006 25 −1.25 0.007 813
3 0 0 0 40 0 0 0
4 5 10 0.000 125 50 0.006 25 1.25 0.007 813
5 −3 −6 −2.7E−05 30 −0.000 81 −0.75 0.000 608
6 −3 −6 −2.7E−05 30 −0.000 81 −0.75 0.000 608
x= eini = 0.010 328
Example Consider the truss in Figure 11.12 where the stress–strain relation is nonlinear and
defined by
1
σ = 200 000ε3
1
Ni ei 3
= 200 000 11 57
A L
3
Ni 1
ei = L 11 58
A 2 × 105
The problem is to determine the linear deflection of the point of application of force P and in the
direction of force P. The approach follows
∂N i
x= ei = ei ni 11 59
∂P
A force analysis based on the method of joints gives according to Table 11.1, x = 0.010 328 in.
Castigliano’s Theorem
Materials commonly used in engineering structures have a linear relation between stress and strain,
which is typically expressed by
σ = Eε
In this case, γ = U and the conjugate pair now becomes
∂U ∂U
= Pi = xi 11 60
∂x i ∂Pi
The second expression leads to Castigliano’s theorem, which states:
If an elastic system is mounted so that rigid-body displacement of the entire system are
impossible and certain external forces (P1, P2, P3, …) act on the system, the displacement
component xi, of the point of application of the force, Pi in the direction of force Pi is deter-
mined by the equation
∂U
= xi 11 61
∂Pi
One useful application is finding deflection steel trusses.

∂U i
δ= 11 62
i
∂P
where Ui is strain energy in the ith member summed throughout the truss. The force in the ith
member is defined by
N i = a1 P1 + a2 P2 + a3 P3 + + ai Pi 11 63
where the loads, Pi, are applied throughout the truss. If we want to determine deflection, δ, in the
direction, P, applying Castigliano’s law
∂U i ∂N i
δ= 11 64
i
∂N i ∂P
But the elongation in each structural member is

∂U i N i Li N 2 Li
ei = = , where U i = i 11 65
∂N i EAi 2EAi
while
∂N i
= ni
∂P
the force in the ith member is due to a unit force applied to point where P is applied with all other
forces set equal to zero. Therefore,
∂N i
δ= ei = ei ni 11 66
∂P
the same expression as before. Since materials used in most engineering structures have linear
stress–strain properties, these equations become very useful. Mathematical expressions for elastic
energy are readily available as shown below. Strain energy expressions for members with uniform
loadings are
N 2L M2L T2L
U axial = U bending = U torque =
2EI 2EI 2GJ
Consider a frame or structural member subjected to axial, torsion, shear, and moment loads. The
total strain energy in the member is the sum of each strain energy component. If there are several
members to the structural frame, the total strain energy in the frame is
N2 M2 T2
U= + + dx 11 67
all-members
2EA 2EI 2GJ
The integral means that each loading can vary along each structural member. Length L is
replaced by dx. Applying Eq. (11.64)
∂N ∂M ∂T
N M T
δ= ∂P + ∂P + ∂P dx 11 68
all-members
EA EI GJ
Nn Mm Tt
δ= + + dx 11 69
all-members
EA EI GJ
584 Energy Methods
where
m – moment “load in the ith member due to a unit load (or couple) acting alone on the
n – tension system at the point where the displacement δ is to be determined” (called
(compression) dummy variable)
t – torque
are loads in the ith member due to a unit load (or couple) P acting alone on the system at point
where δ is to be determined.
M – moment “load in the ith member due to the actual loads

N – tension (compression) (or couples) acting on the system”
T – torque
The displacement, δ, will be a linear (or angular) displacement depending on the type of loading.
Several examples of the application of Eq. (11.69) are given below.
Example Consider the split circle hoop fixed at one end and displaced by a force, F, normal to the
plane of the hoop (Figure 11.13). The problem is to find the displacement, δ, in the direction of the
applied force, P. This problem will be solved using Eq. (11.69).
Moment and torque are defined by
M = PR sin θ 11 70
T = PR 1 − cos θ 11 71
2π
Mm Tt
δ= + R dθ 11 72
EI GJ
0
The dummy variables, m and t, are simply

y
m = R sin θ
δ
t = R 1 − cos θ
After substitution
2π
R2 sin2 θ R2 1 − cos θ 2 P θ R
δ=R P+ P dθ
EI GJ
0
11 73 x
2π
sin 2 θ 1 − 2 cos θ + cos 2 θ
δ = R3 P+ P dθ z
EI GJ
0
Recall
2π
2π
θ sin 2θ
sin 2 θ dθ = − =π
2 4 0 Figure 11.13 Transverse displacement of a
0
split ring.
2π
2π
θ sin 2θ
cos 2 θ dθ = + =π
2 4 0
0
and
πr 4 πr 4
I= and J=
4 2
Collecting terms and integrating gives
2PR3 2 3
δ= + 11 74
r4 E G
Example In this case, we wish to determine the twist angle at point b where a torque (T) is
applied (Figure 11.14). Applying Eq. (11.69)
ℓ ℓ
Mm Tt
δ= + dx
1, 2
EI GJ
0 0
ℓ ℓ
Tt Mn
δ= dx + dx
GJ EI
0 0
T=M=T and t=m=1

Bring everything together
Tt Mm
θb = l+ l 11 75
GJ EI
0 l a
y
x
T
b
Figure 11.14 Bent rod.

586 Energy Methods
T 1 T 1
θb = l+ l
GJ EI
By substitution
2Tℓ 1 1
θb = + 11 76
πr 4 G E
Example Here we want to find displacement at the end of a cantilevered beam due to a uniform
load, w (Figure 11.15). This is a good example of the dummy load concept. While there is no applied
force at the displacement point, dummy force, n, is still applied producing dummy moment, m.
Applying Eq. (11.69)
L
Mm
δ= dx 11 77
EI
0
where
m = − L−x 1
L−x
M= − L−x w 11 78
2
By substitution
L
L−x 3
δ= w dx
2EI
0
L4 w
δ= 11 79
8EI
which agrees with the solution in Chapter 2.
x
EI
n
L
Figure 11.15 Beam deflection.
Example Here is a good example of multiple sections in a structure (Figure 11.16). Total strain
energy is sum of both sections. The applied force, F, is parallel to the x axis.
z Bent wire
T M
2
r
θ 1 s
y
r
x F
Figure 11.16 Multiple sections.
Equation (11.70) for this example is

Mm Tt
δ= + ds 11 80
1, 2, 3
EI GJ
Member 1
m1 = 1 s M 1 = Fs
Member 2
m2 = 2r sin θ M 2 = 2rF sin θ
t 2 = r 2 cos θ − 1 T 2 = Fr 2 cos θ − 1
Bring all terms together
r π π
s Fs 4Fr 2 r2F
δ= ds + sin θrdθ +
2
2 cos θ − 1 2 rdθ 11 81
EI EI GJ
0 0 0
The integrals are expanded as follows

r
F 2 Fr 3
s ds =
EI 3EI
0
π
π
4Fr 3 4Fr 3 θ sin 2θ 2πFr 3
sin 2 θdθ = − =
EI EI 2 4 0 EI
0
π π
Fr 3 2 Fr 3
2 cos θ − 1 dθ = 1 − 4 cos θ + 4 cos 2 θ dθ
GJ GJ
0 0
π
Fr 3 θ sin 2θ
= θ − 4 sin θ + 4 +
GJ 2 4 0
3πFr 3
=
GJ
588 Energy Methods
Bringing everything together
Fr 3 1 3πFr 3
δ= + 2π + 11 82
EI 3 GJ
Chemically Induced Deflections
The theory and application of energy methods is an established engineering tool. Its main applica-
tions occur in predicting deflections and loads in engineering structures as previously described.
Recent studies show how it is also used to predict deflections at the molecular and nano levels.
Microcantilever Sensors
Recently, microcantilevers have attracted much attention due to their potential as an extremely
sensitive sensor platform for chemical and biological detection. Microcantilever sensors have many
advantages over the competing technologies because of high sensitivity and easy production at low
cost. Adsorption-induced cantilever bending is observed when the adsorption is confined to a single
side of a cantilever. The adsorption-induced cantilever bending is ideally suited for measurements
in air or under solution.
The exact molecular mechanism involved in adsorption-induced stress is not completely under-
stood. Typically, deflections of microcantilevers are explained in terms of energy transfer between
surface free energy and elastic energy associated with structural bending of the cantilevers [2–4].
Surface stress involved in adsorption-induced stress is often calculated by Stoney’s equation [5].
Ed2
Pt = 11 83
6r
where
P – tension force per unit cross-sectional area of film

t – thickness of film
E – modulus of elasticity of beam material
d – thickness of beam
r – radius of curvature of deflected beam
Dareing and Thundat [6] take a different approach by explaining the mechanism of bending in
terms of atomic and elastic energies.
Their simulation relates atomic (or molecular) interactive potential of the adsorbent to elastic
energy in the cantilever. Total energy is the sum of this atomic potential plus elastic energy. Beam
deflection is determined by minimizing the total potential energy expression in terms of beam
curvature.
Simulation Model
The bending model is based on energy potential in the first layer of atoms attached to one surface of
a cantilever and elastic potential in the microcantilever itself. Atoms are situated against the can-
tilever surface as shown in Figure 11.17.
2 3
1 a
b
2c n n
Figure 11.17 Arrangement of atoms (or small molecules) on cantilever surface.
This assumption that the first atomic layer on the beams surface plays a dominant role in micro-
cantilever deflections is supported by the experimental works of Martinez et al. [7] and Schell-
Sorokin and Tromp [8], who measured changes in curvature in cantilevered thin plates due to
adsorption of submonolayer of different atoms in ultra-high vacuum conditions. Per the model,
atoms in the attached film are attracted and repulsed per the Lennard-Jones potential formula [9].
−A B
wr = + 12 11 84
r6 r
dw 6A 12B
F= = 7 − 13 11 85
dr r r
where r is the spacing between atoms. Part of this potential is transferred into the cantilever as elas-
tic strain energy causing the beam to deflect. The equilibrium configuration of the cantilever is
determined by minimizing the total potential function, which is made up of the Lennard-Jones
potential and the elastic energy in the cantilever. Both potential components will now be expres-
sions in terms of the local curvature of the beam.
Assuming the following Lennard-Jones constants,
A = 10 − 77 J m6 = 10 − 5 nN-nm nm6
B = 10 − 134 J m6 = 10 − 8 nN-nm nm12
then atomic potential and force of attraction, vs. separation distance, vary as shown in Figures 11.18
and 11.19.
0
0 0.5 1 1.5
–0.0005
–0.001
Energy, nN-nm
–0.0015
–0.002
–0.0025
re
–0.003
Separation, nm
Figure 11.18 Potential energy between two atoms.

590 Energy Methods
rmax
0.04
re
0.02
0
0 0.5 1 1.5
–0.02
Force, nN
–0.04
–0.06
–0.08
–0.1
–0.12
–0.14
Separation, nm
Figure 11.19 Resultant force between two atoms.
Figure 11.20 shows the curvature of the beam over a distance, b, which is taken as the atomic
space between two atoms on the surface of the beam. Three atoms are involved in the atomic poten-
tial expression. Here, we assume that the adsorbate distribution is uniform over the surface since
the surface is chemically homogeneous. Since the distribution of surface atoms is assumed to be
uniform, the curvature will be uniform along the cantilever. The curvature can be established
by considering the elastic energy in the beam over the length, d, and the atomic potential between
the three atoms, one on the cantilever surface and two in the attached film or coating. It is assumed
R
½z
½(b – z)
n
n
Figure 11.20 Position of surface atoms on deflected beam.

in this simulation that the second and higher layers of atoms play a minor role in deflecting
the beam.
Other assumptions made in creating this model are typical of those used in beam theory, i.e. a
cross-sectional plane before bending remains a plane after bending. In addition, the near-surface
layer of atoms 2 and 3 also remain in the same plane after beam deformation. Based on these
assumptions, the energy potential in the near surface layer of atoms can be expressed in terms
of beam curvature. Note that the movement of 2 toward 3 is
z=ϕ c+a 11 86
b = Rϕ
so
b
z= c+a 11 87
R
Molecular and Elastic Potential Energies

The atomic potential, Us, in terms of beam curvature (1/R) according to the Lennard-Jones expres-
sion is
−A B −A B
Us = 6 + 12 +2 3 + 6 11 88
b−z b−z 1 2 1 2
b−z +a 2
b−z +a 2
4 4
The elastic bending potential, Ub, over the atomic length, b, is
2
1
U b = 1 2EI b 11 89
R
For stable equilibrium, the total potential energy
U = Us + Ub 11 90
must be a relative minimum, which is determined from
dU
c =0 11 91
d
R
The resulting equation is
A a B a
EI c 6 1+ 13 12 1 +
= b7 c − b c
c2 R c a 7 c a 13
1− 1+ 1− 1+
R c R c
A3 c a a B c a a
7 1− 1+ 1+ 13 3 1 − 1+ 1+
+2 b 2 R c c − b R c c
7
1 c a 2 a 2 4 1 c a 2 a 2
1− 1+ + 1− 1+ +
4 R c b 4 R c b
11 92
592 Energy Methods
The values of c/R, which satisfy this expression, define the curvature of the microcantilever beam.
This curvature is constant along the beam or the deflection of the beam is a circular arc. The trans-
verse deflection of the end of the microcantilever can be determined from simple trigonometry
using the known beam curvature.
δ = R 1 − cos θ 11 93
We have applied the above simulation to a typical microcantilever beam of 200 μm length with a
beam cross section of, 30 μm × 1 μm. The material modulus is 1.79 × 1011 Pa. The thickness of
atomic monolayer, a = 0.5 nm, and the spacing of the atoms on the cantilever surface, b = 0.5
nm. The Lennard-Jones constants, A = 10−77 J m6 and B = 10−134 J m12.
c
The computed curvature for this set of conditions is, = 3 × 10 − 7 . This curvature is positive
R
meaning that the beam curves upward cupping the adsorbate monolayer. The corresponding angle,
θ = 0.000 119 7 rad or 0.006 86 . The corresponding transverse deflection of the end of the beam is
predicted as 11.97 nm or approximately 12 nm.
References
1 Langhaar, H.L. (1962). Energy Methods in Applied Mechanics. New York: Wiley.
2 Ibach, H. (1997). Surf. Sci. Rep. 29: 193–263.
3 Dahmen, K., Lehwald, S., and Ibach, H. (2000). Bending of crystalline plates under the influence of
surface stress – a finite element analysis. Surf. Sci. 446: 161–173.
4 Raiteri, R., Grattarola, M., Butt, H.-J., and Skladal, P. (2001). Micromechanical cantilever-based
biosensors. Sens. Actuat. B 79: 115–126, Elsevier.
5 Stoney, G.G. (1909). Proc. R. Soc. Lond. 82: 172.
6 Dareing, D.W. and Thundat, T. (2005). Simulation of adsorption-induced stress of a microcantilever
sensor. J. Appl. Phys. 043526: 97.
7 Martinez, R.E., Augustyniak, W.M., and Golovchenko, A. (1990). Phys. Rev. Lett. 64: 1035.
8 Schell-Sorokin, A.J. and Tromp, R.M. (1990). Phys. Rev. Lett. 64: 1039.
9 Israelachivili, J.N. (1991). Intermolecular and Surface Forces, 2e. Academic Press (Lennard-Jones
constants, A and B).
593
Index
a Drill Collar Buckling 418

Amplitude Equation 492, 496 Drill Pipe Buckling 420
Atomic Force Microscope 118 Marine Riser Buckling 426
Annular Velocity 557 Beam on Elastic Foundation 471
Annular Pressure Drop 552 Mathematic Formulation 473
Available Hydraulic Horsepower 289 Concentrated Force Load 474
Optimum Flow Rate 289 Deflection 475
Nozzle Selection 286 Slope 475
Archimedes 151 Shear Force 475
Early devices 152 Bending Moment 475
Buoyancy 278 Application to Cylinder 475
Radial Stiffness 476
b Ring (strap) Loading 479
Beams 39 Radial Deflection 475
Shear Diagram 40 Surface Slope 475
Moment Diagram 41 Radial Shear 475
Bending Stresses 45 Bending in Wall 475
Shear Stresses 48
Deflections 47, 576, 397 c
Multi-spanned Beam – Columns 402 Configuring the Design 37
Straight Beams 402 Closed Ended Problems 1
Circular Dog Legs 403, 410 Capillary Tubes 544
Boundary Conditions 47, 396 Characteristic Equation 512
Composite Beams 52 Creativity 12
Statically Indeterminate 59, 60, 400 Codes and Standards 8
Polar Coordinates - High Angle Bending Castigliano Theory 582
403, 405 Columns 411
Beam – Columns 418, 420, 424 429, 435 Variable Cross Section 415
Buckling from Internal Pressure 416 Concept Feasibility 14
Effective Tension 417 Cost Effective 14
Drill Pipe 420 Technically Feasible 14
Drill Collars 418 Concept Evaluation 15
Marine Risers 426 Performance 15
594 Index
Concept Evaluation (cont’d) Thick Wall over Thick Wall 448

Risk 15 Ring Loading on Cylinders 479
Availability 15 Multiple Rings 480
Cost 15 Circumferential Buckling 449
Cat head pulley 163 Radial deflections due to Ring Loading 477
Coefficient of Friction 336 Bending due to Multi Bands 480
Static 336 Brasses’ Buckling Equation 450
Dynamic 337 Long Cylinders 451
Connections 19 Short Cylinders 450
Bolted Connections 19, 21 Thin Cylinders of Revolution 452
Welded Connections 30 Chemical Induced Deflections 588
Drill Pipe Tool Joints 24 Mathematical Model 591
Make-up torque 28 Microcantilever Sensors 588
Composite Beams 52 Combined Stresses 485
Welded Attachments 53 Curved Beams 455
Bolted Attachments 53 Neutral Axis of Bending 455
Nailed Attachments 53 Stress Distribution 456
Cam Drives 191 Charpy V – Notch test 69
Linear Follower (Flat Surface) 192 Critical Crack length 70
Linear Follower (Roller Contact) 194 Columns 578
Ritterhaus Model 196 Euler buckling formula 411
Pivoted Follower 196 Tapered columns 429
Casing Design 293 Compression member range 414
Intermediate Casing 294 Intermediate column range 414
Production Casing 294, 302 Euler Column range 414
Collapse Pressure Loading 295
Burst Pressure Loading 295 d
Internal Tension 300 Damping Coefficient 87
Design Procedure 302 Critical Damping 87
Joint Couplings 303 Damping Factor 87
Leak - off Test 290 Experimental Measurement 493, 87
Critical Pressure Gradient 290 Damping Matrix 517
Production Casing Design 302 Critical Modal Damping 534
Cylinders 443 Modal Damping Factor 534
Thin Wall – Criteria 444 Design Methodology 3
Hoop Stress 444 Open Ended Problems 1, 3
Axial Stress 444 Operational Requirements 5
Radial Deflections 445 Concept Evaluation 14
Pressure Vessels 444 Starting the Design 16
Helix Seam 445 Subsystems 17, 19
Spherical Ends 444 Design Specifications 7
Thick Wall Cylinders 443 Creating Alternatives 12
Lame’ Formula 443 Drilling Rig 267
Radial Deflections 448 Hoisting (Crown Block, Travelling Block) 269
Interference Fit 445 Hydraulic Circulating System 285
Thin Wall over Thin Wall 446 Rotary System 272
Thin Wall over Thick Wall 448 Rotary Table 272
Index 595
Top Drive 273 Directional Drilling 306

Well Control System 290 Navigating in Geo Space 328
Drilling mud density 290 Directional Control 307
Annular Preventer 291 Positive Displacement Motors 202, 561
Blow Out Preventer (BOP) 292 Drilling Turbines 345, 560
Power Units 310 Rotary Steerable Tools 308
Drilling and Production 267, 17 Stabilized Assemblies 309
Drill Bits 284 Dog Leg Severity 331
Roller Cone 284 Maximum bending stress in Dog Legs 408
Polycrystalline Diamond Compact (PDC) 284 Pipe Friction 411, 272
Natural Diamond Drill Bit 284 Dynamics of Particles 335
Bit Nozzle Selection 341 Equations of Motion 336
Drillstring 268 Kinetic energy 337
Design 276, 279, 281 Potential energy 339
Wellbore friction 272 Impulse – momentum 342
Buoyancy 277 Conservation of momentum 371
Neutral Point 277 Dynamics of Rigid Bodies 351
Vibration 536 Equations of Motion 352
Natural frequencies 537 Translation 354
Drill Pipe 281 Pure Rotation 354
Lateral Buckling 420 General Motion 356
Lateral Vibrations 420 Interconnecting Bodies 361
Torsional Buckling 435 Start-up power 361
Selection 281 Conservation of Momentum 371
Size, Grade 281 Impact Forces on Drill Collars 368
Bending Stress in Dog Legs 408 Rotor Loads on PDM Stator 359
Joint Connections 24
Make up Torque 28 e
Drill Collars 278, 96 Energy Consumption in U. S. 147
Neutral Point 278 Effective Tension/Compression 417, 420
Vibration Absorber 96 Environmental 8
Lubinski – Drill Collar Buckling 420 Ethics 11
Stabilized Bottom Hole Assembly Enthalpy 132
(BHA) 309, 402 Entropy 132
Deflections between Stabilizers 402 Electric Motors 136
Side Loads at Drill Bit 403 Eigenvalues 496
Buckling Between Stabilizers 418 Energy Method in Mechanic 569
Drill Collar Length 277 Castigliano’s Theorem 582
Static Considerations 279 Dummy load 584
Vibration Considerations 96 Complimentary Energy 579
Natural Frequency with Absorber 97 Legendre Transformation 580
First and Second Drill Collar Modes 98 Variational Calculus 570
Drill Bit Nozzles 287 Stability Test 570
Nozzle Selection 289, 341 Multiple Degrees of Freedom 574
Diesel Engines 139 Principle of Minimum Potential Energy 569
Compression Ratio 139 Chemically Induced Deflections 588
Diesel-Electric Generators 310 Energy Alternatives 147
596 Index
f Generalized Formulation 388

Force Analysis 316 Plane Stress 388
Freebody Diagrams 318 Plane Strain 391
Trusses – Joint Method 319
Trusses – Section Method 319
i
Failure Modes 62
Four Stroke 137
Material Yielding 55
Two Stroke 138
Metal Fatigue 63
Diesel Engines 139
Brittle Fracture 69
Impeller Pumps 76, 144
Stress corrosion cracking 69
Performance Features 76
Deflection 47
Senior Capstone Project 74
Instability 569
Offshore Pipeline Application 144
Factor of Safety 37
Interference fit between cylinders 445
Friction Head 72
Thin wall cylinders 445
Bernoulli Energy Equation 71
Thin wall – thick wall 448
Darcy – Weisbach 72
Thick wall – thick wall 448
Flat Gear Tooth System 189
Kinematics 189
j
Mathematical Formulation 190
Journal Bearings 254
Fourier Series 576
Pressure Distribution 254
Minimum Potential Energy 577
Load Capacity 257
Fatigue Analysis 64
Goodman Diagram 64
Minimum Film Thickness 258
Combined Normal and Shear 68
Boundary Lubrication 261
Formation Pressure 290
Hersey Diagram 262
Normal Pressure 290
Fossil Fuel Consumption 147 k
g Rectangular Coordinates 321
Gas Turbines 139 Polar Coordinates 322
Mechanics of Fluid Flow 140 Curvilinear Coordinates 325
Thrust – Impulse Momentum 141 Relative Velocity 174
Power Generation 142 Relative Acceleration 174
Applications 140 Translating Reference Frame 173
Jet Engine 140 Rotating Reference Frame 175
Turbo Prop 143 Coriolis Component 177
Helicopter 143 Kinematics of Rigid Bodies 333
Impeller Pump 144 Velocity analysis 179
Instantaneous center 179
h Acceleration analysis 180
Drilling Hydraulics 285 Angular acceleration 181
Hydrostatic Pressure 290
Heat Cycle 134 l
Hydrostatic Trust Bearings 220 Lubrication 220
Fluid Mechanics 222 Hydrostatic Bearings 220
Optimization 224 Fluid Flow Equations 220
Hooke’ Law 383 Velocity Profile 222
Index 597
Pressure Profile 222 Cast Iron 56

Load Caring Capacity 223 Endurance Limit 63, 68
Friction 226 Manufacturing 113
Optimum Design 227 Drawings 114
Spring Constant 231 Dimensioning 114
Squeeze Films 228 Tolerances 115
Pressure Profile 229 Surface Finish 117
Velocity of Approach 229 Market Analysis 5
Applied Force 229 Mohr’s Circle 381
Damping Constant 231 Stress Circle 381
Disc with recess 233 Strain Circle 386
Washer 234 Inertia Circle 467
Circular Arc Surface 235 Mechanical Linkages 173
General 2-Dimension Shape 236 Scotch Yoke 177
Poisson Equation 236 Slider Crank 178
Laplace Equation 237 Four Bar Linkage 181
Wrist Pin Application 237 Three Bar Linkage 184
Slider Bearings 240 Geneva Mechanism 188
Fixed shoe 240 Marine Risers 424
Velocity Profile 241 Equation of Bending 426
Pressure Profile 242 Effective Tension 426
Load Capacity 243 Buckling 426
Friction 243 Modal Analysis 507
Pivoted Shoe 245 Local coordinates 516
Pivot Point Location 245 Modal Coordinates 516
Exponential Profile 247 Modal Transformation 519
Pressure Distribution 247 Stiffness matrix 504
Load Capacity 249 Mass Matrix 512
Comparison with Flat Profile 248 Damping Matrix 516
Open Entry 249 Proportional Damping 518
Exponential Profile with side leakage 250 Modal Matrix 512
Journal Bearings 254 Mode Shapes 513
Pressure Distribution 256 Orthogonality of natural modes 517
Load Capacity 257 Modal Response – Multiple DOF 516
Petroff’s Law 259 Undamped Free Vibration 520
Sommerfeld Number 260 Damped Free Vibration 524
Stribeck Analysis 261 Undamped Forced Vibration 528
Hersey Diagram 262 Damped Forced Vibration 529
Friction verses ZN/P 262 Complex Variable Approach 530
Minimum Film Thickness 258 Experimental Modal Analysis 532
Direct Transfer Function 532
m Cross Transfer Function 532
Material Properties 55, 57 Microcantilever Sensors 562
Yield Strength 55, 56 Minimum Potential Energy 569
High Strength Steel 56 Stable Equilibrium 569
Low Strength Steel 56 Unstable Equilibrium 569
Brittle Materials 56 Critical Buckling Force 572
598 Index
Molecular Potential 589 Laminar 541

Lennard-Jones Model 590 Turbulent 541
Molecular Attraction 590 Power Transmission 151
Moineau Mechanism 202 Archimedes Devices 152
Kinematics of Rotor/Stator 202 Horse Collar 153
Mechanics of Torque, Speed 203 Lewis Gear Tooth Model 156
Stage Length 205 Flat Tooth Gear Profile 155
Rotor – stator 206 Simple Gear Set 158
Positive Displacement Pump 202 Torque, Speed Conversions 158
Positive Displacement Motor (PDM) 207 Worm Gear 159
Mechanical Power Output 557 Planetary Gears 160
Performance Data 558 Compound Gear Train 161
Hydraulic Transmission 199
n Downhole Drilling Motors (PDM) 202
Non-Periodic Forces 100, 535 Power Output 201
Convolution Integral 101 Monitoring Performance 203
Multi-Degrees of Freedom 530 Lobe Ratio 204
Modal Analysis 516 Stages of lobes 205
Navigating in Geo-Space 328 Pressure Drop verses Torque 207
Non-Circular Conduits 545 Flow Rate verses Speed 203
Elliptical 545 Hydraulic Drilling Turbines 345
Rectangular 546 Power Output 348
Non-Newtonian Fluids 551 Monitoring Performance 350
Rheology Models 551 Pulley Drives 162
Drilling Fluid Model 551 Belt Connections 164
Internal Pressure Drop 551 Friction Drives 163
Annulus Pressure Drop 552 Belt Pretension 164
Bit Nozzle Pressure Drop 556 Power Screw 198
Nano Surface Undulations 118 Pressure Vessels 443
Manufactured 118 Hoop Stress 444
Measurements 119 Axial Stress 444
Spherical Cap 444
o
Thin Shells of Revolution 452
Open ended problems 1
Radial Deflection 476, 479
Stresses in a Helical Weld 444
Optimum Hydraulics 555, 285, 289
Flow Rate 557 q
Optimum Thrust Bearing Design 224 Quality Assurance 33
Oilwell Drilling Pump 552 Engineering Education 34
Triplex Pump Performance 553 Apollo/Saturn Project 35
p r
Phase Angle 91 Reliability 9
Problem Solving Process 4 Rotor Imbalance 105
Performance Requirements 7 Single rotor 109, 108
Product Life Cycle 10 Shaft Balance in Two Planes 112
Pipe Flow 541, 545, 547, 551 Stability of rotating pipe 110
Index 599
Rocket Engines 144 Beam – Column 424

Thrust 146 Surface Finish 117
Turbopump 146 Micro Dimensions 117
Rocketdyne F-1 Engine 145 Nano Dimensions 118
Atlas Booster Engine 145 Steam Locomotives 127
Rocket Dynamics 147 Shay Engine 129
Renewable Energy 148 Steam loader 130
Refining the design 113 Steam Turbines 135
Drawings 114 Shaft Design 166
Tolerances 115 Keyway 172
Rolling Contact Bearings 209 Squeeze Film Bearings 228
Types 209 Fluid Mechanics 229
Load Ratings 210 Interior recess 231
Vibration Effects 213 Washer geometry 232
Environment Effects 216 Spherical Geometry 232
Comparison with Hydrostatic Bearings 224 Wrist Pin 237
Comparison with Journal Bearings 263 Non-Symmetrical Boundaries 236
Rotary Table 272 Slider Bearings - Fixed Shoe 240
Kelly Drive 272 Design Parameters 241
Power Transmission 272 Slider Bearings - Pivoted Shoe 245
Slider Bearings - Exponential Profiles 247
s Comparison with Straight profile 248
Specifications 7 Side Leakage Solution 250
Performance requirements 7 Stress Transformation 377
Sustainability 7 Strain Transformation 384
Social Considerations 9 Strain Measurements 393
Safety and Liability 11 Strain Gauges 394
Codes and Standards 8 Strain Rosettes 395
Environmental 8 Strain Conversions 395
Cost Considerations 10 Conversion to Stress 393
Scoring Design Alternatives 15 Shear Centers 460
Subsystems in Design 19 Shear Flow 461
Bolted Connections 19, 21 Calculation of Shear Centers 462
Welded Connections 30 Angle Iron Beam 463
Statically Indeterminate Systems 400, 25, 59 Semicircular Cross Section 463
Single Degree of Freedom 77 Twisting of Open Channels 460
Natural frequency 80, 82 Stable – Unstable Equilibrium 569
Spring Arrangements 85 Stress Concentration 62
Coil Spring Constant 86
Mass units 335 t
Critical Damping 87 Tapered Flex Joints 429
Vibration Response 77, 89 Mathematical Model 430
Stability of Pipe 416 Differential Equation 431
Euler Buckling 411 Closed Form Solution 432
Fluid pressure Effects 416 Application to Marine Risers 435
Synchronous Whirl 106, 109 Torsional Buckling of Long Pipe 435
Rotating pipe under axial load 110 Mathematical Model 436
600 Index
Torsional Buckling of Long Pipe (cont’d) Gravitational system 335

Boundary Conditions 438, 442 System International (SI) 335
Simply supported, no thrust 440 Undamped Vibration Absorber 502
Applied Thrust Force 441
Lower End Fixed 442 v
Operational Significance 442 Vibration Resonance 77
Transmissibility 94 Single Degree of Freedom 78, 80
Vibration Isolation 94 Multi Degrees of Freedom 530
Theory of Stress 378 Modal Response 531
Stress Transformation 378 Vibration Control 93
Mohr’s Stress Circle 381 Mass, Stiffness, Damping Affects 93
Theory of Strain 383 Viscous damping 87
Strain Transformation 384 Undamped vibration absorber 502
Mohr’s Strain Circle 386 Viscous Pumps 541
Theories of Failure 482 Pressure Induced Flow 542
Von Mises Distortion Energy 484 Velocity Induced Flow 542
Brittle Material 487 Optimum Output 544
Coulomb – Mohr Criteria 487 Viscosity 541
Modified Coulomb – Mohr Criteria 487 Newtonian Fluid 220
Transfer Functions 531 Non-Newtonian Fluids 541
Direct Transfer Function 533 Bingham Plastic Model 541
Cross Transfer Function 534 Exponential Model 541
Top Drive 273 Plastic Viscosity 551
Operation 274 Gel Strength 551
Triplex Mud Pump 552
Output Performance 253 w
Liner Size 253 Water wheels 125
Turbine Performance 560, 345 Mechanics 125
Fluid Mechanics 345 Power Output 126
Mathematical Predictions 347 Power Transmission 127
Optimum speed 350 Well Control 290
Stall Torque 350 Formation Pressure 290
Runaway Speed 350 Normal – abnormal 290
Performance Conversions 349 Pressure Gradient 295
Mud Weight 295
u Hydrostatic Pressure 295
Unsymmetrical Bending 464 Annular Preventer 291
Principal Axes of Inertia 464 Blowout Preventer (BOP) 293
Mohr’s Inertia Circle 467 Shut-in Standpipe Pressure 291
Bending Stresses 470 Formation Pressure 290
Neutral Axis of Bending 468 Required Mud Weight 291
Unsteady Flow 547 Whipple truss analysis 318
Szymanski’s Solution 548 Method of joints 319
Units of Measure 335 Method of sections 319

Untitled

Uploaded by

Copyright:

Available Formats

You might also like

Untitled

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Untitled

Uploaded by

Copyright:

Available Formats

Engineering Practice with Oilfield and

This Work is a co-publication between ASME Press

Limit of Liability/Disclaimer of Warranty

Library of Congress Cataloging-in-Publication Data applied for:

Cover Design: Wiley

Set in 9.5/12.5pt STIXTwoText by Straive, Pondicherry, India

Part I Engineering Design and Problem Solving 1

1 Design and Problem Solving Guidelines 3

2 Configuring the Design 37

Refining the Design 113

Part II Power Generation, Transmission, Consumption 123

3 Power Generation 125

Rocket Dynamics 146

4 Power Transmission 151

Acceleration Analysis 187

5 Friction, Bearings, and Lubrication 209

Thick Film Slider Bearings 240

6 Energy Consumption 267

Physical Properties of Drill Pipe 279

Part III Analytical Tools of Design 313

7 Dynamics of Particles and Rigid Bodies 315

Velocity Vector 325

8 Mechanics of Materials 375

Application to Marine Risers 435

Applied Torque 483

9 Modal Analysis of Mechanical Vibrations 491

Complex Variable Approach 526

10 Fluid Mechanics 541

11 Energy Methods 569

Principle of Complementary Energy 579

Part I Engineering Design and Problem-Solving

S section modulus of a cross sectional area, Sommerfeld

δ displacement, log decrement

Engineering Design and Problem Solving

Design and Problem Solving Guidelines

Concept A Concept B Concept C

Figure 1.1 Design development process.

Operational requirements or product attributes describe the expected functional performance of a

Government Procurement Procedure

Petroleum Industry Procedure

•• Ocean water depth

•• Oil, gas, water content of reservoir

• Production rate (barrels per day).

Considerable time is spent in gathering this information to establish operational requirements

a) design efficiency in terms of materials, weight, cost, energy consumption

Designs are configured (size, dimension) according to the allowable stress.

Product Life Cycle

Product Safety and Liability

• Is it morally right; can I live with the outcome?

Creating Design Alternatives

•• high energy level

Innovation is an individual activity. The foundation of science is built on the genius of

Reference Books and Trade Journals

Experts in a Related Field

1) Record all the ideas generated.

Existing Products and Concepts

• Materials used in the product.

Evaluating Design Alternatives

Performance – Capability to achieve needed operational characteristics, plus reliability.