Professional Documents
Culture Documents
Untitled
Untitled
Untitled
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
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.
ISBN: 9781119799498
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
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
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
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 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
Greek Symbols
Part I
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
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.
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
4 Design and Problem Solving Guidelines
Operation requirements
Design specifications
Synthesis
Analysis
Evaluate concepts
Preliminary
design
Build prototype
Test prototype
Fabrication
drawings
Manufacture
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.
Operational Requirements 5
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
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.
•• 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.
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.
Design Specifications 7
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,
8 Design and Problem Solving Guidelines
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:
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.
Design Specifications 9
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
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.
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.
Design Specifications 11
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?
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.
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
•• task committed
sense of humor.
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.
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:
•• Geometric modifications
Energy-flow modifications
– Change in the path
– Change in the form
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.
14 Design and Problem Solving Guidelines
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.
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:
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.
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
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.
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.
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.
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.
18 Design and Problem Solving Guidelines
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.
•• 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.
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.
V
Bolt
rn Beam
Fn
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
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.
δspring
δbolt
ks
P PS
P PB
22 Design and Problem Solving Guidelines
PS = P − ΔPS
PB = P + ΔPB
Force
PB
(Bolt)
P
ΔδS
δS
Displacement
δB
ΔδB
PS
P
(Spring)
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.
2 414 000
0 0162 ωcr = 200 1 +
2000
ωcr = 1602 rad s
f = 255 cps
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
Pin
(d)
P
Fb
Fp
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
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.
kp + kb
P= F yld − F 1 34
kp
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
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.
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
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.
Connection of Subsystems 31
F
l x
τT
h r
c
τF
0
Weld area, A
L
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.
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.
32 Design and Problem Solving Guidelines
Gusset plate
R
Gusset plate
θ
x
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
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
• Assessment/feedback
34 Design and Problem Solving Guidelines
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
This statement is reviewed from time to time with the constituents of the program.
Mission
Constituents
statement
ABET criteria
Feedback
Academic program
mechanism
B. S. graduate
Assessment
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?
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
36 Design and Problem Solving Guidelines
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
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.
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.
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.
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
38 Configuring the Design
F/2 t2
F
F/2
t1 F
S
d
S
w F
S
Example Consider the double strap connection with the following load conditions (Figure 2.1).
The problem is to determine
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)
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
40 Configuring the Design
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.
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
x F = 1000 lb
5 ft 5 ft
V 500 lb
2500 ft-lb
M
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
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
dM
= V slope of the moment diagram equals the local shear 28
dx
M M + dM
V V + dV
dx
Beam Analysis 43
V 2 − 500 = − 100 10
V 2 = − 500 lb
Vl Vr Ml Mr
44 Configuring the Design
y F = 1000 lb
w = 100 lb/ft
5 ft 5 ft
V
500 lb
3750 ft-lb
–1000 lb
y 1000 ft-lb
w = 100 lb/ft
L = 10 ft
600 lb
100
V
1750 ft-lb
–400
M 800
750
1 ft
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
n
n y
dx
46 Configuring the Design
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.
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.
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
dy
y 0 =0 and =0 2 26
dx x=0
y F
x
EI
L
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.
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
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.
50 Configuring the Design
1/4 6 in.
4 32.75
1/4
Q
t
4
1.03 24.5
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
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.
Shear Stress
Q = 1 2 4 = 8 in 3
VQ 2000 8
τ= = = 750 psi
It 21 33 1
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
V J = τ J AJ
AJ = bJ L
F J = bJ τJ L 2 36
3) Design of connectors
FN N = FJ L 2 37
N FJ
= number of nails per length
L FN
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
6
54 Configuring the Design
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
F J = τbL
F J = 38 3 2 L = 76 6L
500N = 76 7L
L 500
= = 6 52 in spacing per bolt
N 76 7
M max c 500 72 4
σ max = = = 614 psi
I 234 7
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.
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.
56 Configuring the Design
σyld
High strength
80 ksi
Low strength
2000 μ
Cast iron
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.
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.
Material Selection 57
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
Member AC Member BC
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
Material Selection 59
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
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
where
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
62 Configuring the Design
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
2a
σ0
Material Selection 63
2.5
2
Stress ratio
1.5
0.5
0
0 0.2 0.4 0.6 0.8 1
Distance ratio
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
64 Configuring the Design
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
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
Member B
F
F
Member A
66 Configuring the Design
σamp
σend = 60 ksi
a
σmean
0
σult = 120 ksi
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
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
Material Selection 67
τamp
τmean
0
τult = 69.2 ksi
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].
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
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.
68 Configuring the Design
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
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
0.4
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1
Ratio (σ)
Material Selection 69
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
Transition
temperature
Temperature
70 Configuring the Design
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 mechanics is fundamental to many machine components and production systems. The build-
ing blocks for fluid analysis are:
where
Em
2
V2
1
V1
z2
z1
Datum
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
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.
72 Configuring the Design
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
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.
Example Determine the pressure gradient along a pipe with the following flow conditions.
1 cp = 1 45 × 10 − 7 reyn
lb-s
The viscosity of water in units of ft2
is
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
74 Configuring the Design
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.
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.
gx 2
y x = x tan θ − 2 2 101
2 V cos θ
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
Fluid Flow Through Pipe 75
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
γQ2
p= plotted in Figure 2 32
10938A2n
76 Configuring the Design
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,
Then
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:
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
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.
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
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.
Vibration Considerations 79
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
80 Configuring the Design
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.
k
δst
m F x(t)
where
k
ωn = rad s
m
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
82 Configuring the Design
a
L
θ cg
m
m m
T T EI
L
x x
Fluid level θ
k
δst R
x
0
W
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
Vibration Considerations 83
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
R
ψ
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
I 0 = R2 M
I cg = I 0 − y2 M
I A = I cg + R − y 2 M 2 141
Vibration Considerations 85
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
(a) (b)
L m L
EI EI
k k
Figure 2.38 Springs in parallel and series. (a) Parallel. (b) Series.
86 Configuring the Design
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
88 Configuring the Design
Real
1
V
/div
–4
V
Start: 0 s Stop: 3.9961 s
k θ
M 0
0′
a a
c
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
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)
where δst = F0
k, and
r – frequency ratio
ζ – damping factor
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
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
ωt
F0 F0
4
Amplitued, lb, in.
2
ϕ X
Amplitude,
0
–1 –0.5 0 0.5 1
–2
–4
–6
Time, s
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.
92 Configuring the Design
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
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
W = πcωn X 2 2 176
Vibration Control 93
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
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
m x
m x
kx cx k(x–y) c(x–y)
m
Vibration Control 95
5
ζ = 0.1
Transmissibility, Isolation
3 ζ = 0.2
2
ζ = 0.3
0
0 1 2 3
Frequency ratio, r
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
96 Configuring the Design
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.
Dynamic force
Axial
x
displacement
∂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
98 Configuring the Design
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
1.4
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
β α β α
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
100 Configuring the Design
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.
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
Responses to Nonperiodic Forces 101
m F
dt
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.
102 Configuring the Design
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.
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
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.
104 Configuring the Design
4
Displacement, in.
0
0 0.05 0.1 0.15 0.2
–2
–4
–6
Time, s
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.
θ
r2
cg r1
0
T
T
T0
t
t1 t2
Vibrations Caused by Rotor Imbalance 105
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
am = ao + am o 2 223
d2 x
fx = m − e ω2 cos ωt 2 225 k
dt 2 c
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
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
0 ωt
c
e c c
0 0
a a a
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
108 Configuring the Design
ac 0 = − ω2 e i cos ωt + j sin ωt
Giving
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
Imbalance
b
Trial weight, Wt ϕ
a
0
ϕ
Counterbalance
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.
d4 y
EI − m ω2 y = 0 2 241
dx 4
d4 y
− α4 y = 0 2 242
dx 4
110 Configuring the Design
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
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
Vibrations Caused by Rotor Imbalance 111
2
Q Q mω2 π2
+ + = 2 2 249
2EI 2EI EI L
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
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
y
m1
x m2
6 4
ω
6 6 4
0 1 2 N
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
Refining the Design 113
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.
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
114 Configuring the Design
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.
0.5R
6×ϕ 0.25
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
116 Configuring the Design
A C
B D
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
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
Refining the Design 117
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:
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.
118 Configuring the Design
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
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
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
•• Drilling
Knurling
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
120 Configuring the Design
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
•• 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.
6 Shigley, J.E. and Mitchell, L.D. (1983). Mechanical Engineering Design. McGraw-Hill Book Co.
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
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.
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
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.
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
126 Power Generation
HQγ
P= hp
550
Assuming
Q = 2 ft3/s
γ = 62.4 lb/ft3
H = 12 ft
12 62 4 2
P= = 2 7 hp assuming 100 hydraulic efficiency
550
Flume
Water flow
Torque
1
Speed
Diameter
Torque
Water exits
wheel
2
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
128 Power Generation
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
Steam cylinders
Rotation
Bevel gear pair
Crank shaft
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
130 Power Generation
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.
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
Regional Power Stations 131
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 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.
800
700
600
Temperature (ºF)
300
p = 14.7 psia
200
100
0
0 0.5 1 1.5 2 2.5
Specific volume (ft3/lb)
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.
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
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)
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
Wout
3 4
Turbine
Qin Qout
Boiler Condenser
Pump
2 1
Win
T
Boiler
pressure 3
1 Condenser 4
pressure
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
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
Internal Combustion Engines 137
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.
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.
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)
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.
(a) (b)
Pressure Ignition
Compression
Exhaust
Atm
Intake
Interior spray Top position Bottom position
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.
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
140 Power Generation
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
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
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
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
By substitution
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
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
Gas Turbine Engines 143
Combustion
chamber
Power Exhaust
Fuel
Compressor
Combustion
chamber
Gear box
Power Exhaust
Fuel
Compressor
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.
144 Power Generation
Fluid out
Combustion
chamber
Power Exhaust
Fuel
Impeller pump Compressor
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.
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.
Vcv 2
Combustion
chamber V2
Qoxygen
Q2
Qfuel
Turbo-pump
1
V1 Diverging
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
Energy Consumption in US 147
where
M – mass of rocket
V – velocity of rocket
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.
148 Power Generation
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
Energy Consumption in US 149
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.
150 Power Generation
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
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
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 4.2 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.
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.
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.
r B = r cos ϕ 43
Gear Train Transmission 155
Involute
ϕ P
Pressure line
ϕ
Base circle
rB Pitch circle
r
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
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.
156 Power Transmission
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.
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
t
Gear Train Transmission 157
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
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
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.
ω2 T2
r2
r2
θ2
F
F
r1 θ1 ω1
r1
T1
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.
Dividing by Δt gives
p
V= ω1 4 19
2π
So
p
ω2 = ω1 4 20
2πr
p
T1
ω1
ω2
r
T2
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.
Link 5
Link 4
Link 2 a
d
Gear Train Transmission 161
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.
5 hp
2 hp 3 hp
ω2 T2
TA TB
T ω
T
ω T1 ω1
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
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
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.
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
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
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.
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 (−).
θ
τ T
Fundamentals of Shaft Design 167
rθ
τ
2πr
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
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.
10 in.
A
x
1″ OD
z
8 in.
500 lb
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
Factor of Safety is
100
FS = = 1 64 4 49
60 6
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
10 in.
ω T
a
1″ OD
Motor
Pulley A
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
T
+20
+
– –10
–25
–35
θ
2 c d e
a
1 b 3 4
θe/a
•• Square keys
Round keys
F
a
V
Hub
Pitch circle
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
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.
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
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).
174 Power Transmission
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
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 = 3 268i + 9 66j
The acceleration of point A is determined from
aA = aB + aA B
α
ω
B
O X
176 Power Transmission
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
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).
Mechanical Linkages 177
x
vxy
A A
aA
anP
aCC
ω
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
x = r ω2 cos θ
178 Power Transmission
Acceleration
polygon
aB
A
θ
ω aA/B
aA
r
vA
vB/A
θ B
vB
Velocity x
polygon
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
Here the acceleration of point A is known (aA = ω2r). The direction of the other two components
closes the polygon.
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
IC
A
ω
ϕ vA
L
θ R
B
vB/A θ ϕ
vB
O
Slider crank
Velocity polygon
x
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
A
ω
L
R B
ϕ aB θ ϕ
O
(aB/A)t
aA x
(aB/A)n
ϕ
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, °
where
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.
182 Power Transmission
IC
vA
B C
A
vA
vB/A O
O*
vB (aB)n
(aB/A)n aB aA
(aB/A)t (aB)t
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.
Mechanical Linkages 183
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.
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.
Acceleration Analysis
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.
184 Power Transmission
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
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
Velocity Equation
The velocity of point A on the collar is related to ω by
vA = vP + vA P 4 114
where
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.
a = 8 in.
c = 12 in.
Y
y
P
Ω A
b ω x
βˆ a ĵ
iˆ
θ ϕ
X
αˆ c
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
vP = − ivP
where
vP = b ω = 6 46 ω
vA P = vA − vP 4 118
By substitution
In terms of i, j,
α = i sin ϕ − j cos ϕ
β = i cos ϕ + j sin ϕ
By substitution
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
Mechanical Linkages 187
The minus sign means the collar is moving toward the origin. Bringing terms together
vA = − 29 8i − 74 3j
Acceleration Analysis
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
aA = aP t + aP n + axy + 2ωxvxy 4 122
where
aP t =rα
aP n = ω2 r
axy = aA xy
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
Coriolis
(ap)t
O B
(aA)xy
188 Power Transmission
x
r1
θ ϕ
r2
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 .
Mechanical Linkages 189
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))
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
y
y1
a cos θ
j β
x1
α
θ h M ϕ
O
O1
i x
a b
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.
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.)
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.
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
192 Power Transmission
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
vA = vA β
The magnitude vA is unknown, but its direction is known. The velocity of point P is
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
Therefore
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
aA = aP t + aP n + axy + 2ωxvxy 4 148
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
194 Power Transmission
•• 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.
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
where point P is located on the cam coincident with point D. The values of each component are
listed below.
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
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
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
aD = aC n + aD C n + aD C t
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
counterclockwise. The problem is to determine both angular velocity (ω) and angular acceleration
(α) of the follower arm for this position.
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
vP 8 5 ips
vA P 9 0 ips
85
ω= = 1 42 rad s
6
The acceleration equation is
where
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
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
Hydraulic Transmission of Power 199
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.
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:
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
Fluid flow
Drill pipe
Rotation
Drilling motor
Drill bit
Bit torque
where
PH – hydraulic power, hp
p – fluid pressure, psi
Q – flow rate, gpm
PDM – turbine
p2 2
Drill bit
T, N
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).
202 Power Transmission
R ω
O′
a r
x
O ϕ θ
P(x,y)
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.
0
–5 –4 –3 –2 –1 0 1 2 5
–1
–4
–5
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
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:
• 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
Hydraulic Transmission of Power 205
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
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
206 Power Transmission
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
Motor dimensions can be selected once the motor constant, C, have been specified.
Hydraulic Transmission of Power 207
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
208 Power Transmission
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
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.
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.
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
210 Friction, Bearings, and Lubrication
Dry friction
Generic 0.3
Steel on copper 0.15
Teflon on anything 0.1
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.
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
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
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.
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
X10 Y1 X2 Y2
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
2πt
P t = Pm + Pv sin 5 10
T
where
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
214 Friction, Bearings, and Lubrication
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
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
Therefore
Pm = 3000 lb
Pv = 2100 lb
Pm = 3000 lb
Pv = 2100 lb
2100
α= =07
3000
β = 1.202 (Table 5.3)
K = 0.576 (Table 5.3)
α βr Kr βb Kb
Sleeve bearing
Nrel
1/3 WOB
100
q = 2.8 (Air)
q = 1.4
(Drilling mud)
10
10 100 1000
Roller-bearing life, h
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.
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
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
1
K= 5 19
βq
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
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
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.
dp dτ
= 5 20
dx dy
cp reyn × 106
y
b
h
x
L
1 2
τ+dτ
dy p p+dp
dx
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
222 Friction, Bearings, and Lubrication
Pressure gradient must be minus for positive flow to the right. This equation will now be adapted
to radial flow.
p0, Q
p0
p(r)
r1
r2
Hydrostatic Thrust Bearings 223
W = p0 πr 21 + 2πrp r dr 5 33
1
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
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.
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
Hydrostatic Thrust Bearings 225
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.
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
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
πμω 4
M= r − r 41 5 47
2h 2
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
PT = P p + P f 5 50
W = 100 000 lb
r1 = 5 in.
r2 = 8 in.
μ = 29.24 cp = 42.4 × 10−7 reyn
N = 750 rpm
Hydrostatic Thrust Bearings 227
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.
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
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.
Q r = Vπr 2 5 60
where V is velocity of approach.
By substitution into Eq. (5.28)
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
Squeeze Film Bearings 229
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
230 Friction, Bearings, and Lubrication
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
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).
bh3 dp
Q= − 5 67
12μ dr
dp Q 12μ
= − 5 68
dr b h3
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.
Squeeze Film Bearings 231
W V
Q, p0
p0
W 3πμR4
c= = damping coefficient, force velocity 5 75
V 2h3
90
V
V
ri ρ
r0 ri
r0
r dr
ρ
V
h0
r0
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.
In each case (Figures 5.16 and 5.17), pressure profile is established by the flow equa-
tion (Eq. (5.61)).
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
234 Friction, Bearings, and Lubrication
r2
3πμV r 4o r4
W2 = 2πrp r dr = 3 − r 2i r 20 + i 5 80
h 2 2
r1
3πμV 4
W= r 0 − r 4i 5 81
2h3
πrh3 dp
πr 2 − πρ2 = − 5 83
6μ dr
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
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
Squeeze Film Bearings 235
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
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
r 20
h0 +
1 ρ
W = 6πμV ρ2 − 2 5 97
h0 r2
h0 + 0
2ρ
236 Friction, Bearings, and Lubrication
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
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).
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
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
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
240 Friction, Bearings, and Lubrication
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.
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.
dp d2 u
=μ 2 5 120
dx dy
as before. However, in the slider-bearing case, the film is tapered.
Thick Film Slider Bearings 241
y
Leading edge
Velocity
Trailing edge induced
α
h2
h1 h
x
U
x Pressure
induced
L
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
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
242 Friction, Bearings, and Lubrication
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
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
Coefficient of Friction
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.
We wish to determine:
W Pressure
profile
α h2
h1 5/9 L
x
U a c
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
Coefficient of Friction
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 247
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.
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
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.
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.
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.
Load-Carrying Capacity
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
y y(x) = h1 esx
h1
U
x
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).
h2
a
x
a
h1
z L
∂ 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
252 Friction, Bearings, and Lubrication
or
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
Exponential Slider-Bearing Profiles 253
Table 5.7 Pressure profile for an exponential bearing with side leakage.
pL
dimensionless
μU p (psi)
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
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.
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.
Journal
ϕ
e ω
h(ϕ)
r
hmin
Hydrodynamic Lubricated Journal Bearings 255
x
U Journal surface
Bearing surface
2πr
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
256 Friction, Bearings, and Lubrication
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.
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 °
Load-Carrying Capacity
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
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.
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
e = 0.001 85
5/8″ dia
h0 = c – e = 0.010 − 0.00185
h0 = 0.0085
a = 0.02–0.06-in.
θ = 3.7 –10.9
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.
r – radius of journal, in.
U – surface velocity of journal, in./s
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
μ – viscosity, reyn
N – rotary speed, rev/s
P – average journal pressure, psi
r – radius of journal, in.
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.
260 Friction, Bearings, and Lubrication
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.
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
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.
262 Friction, Bearings, and Lubrication
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.
Hydrodynamic Lubricated Journal Bearings 263
Consider, for example, a drill bit journal bearing having the following conditions
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.
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.
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
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
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.
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.
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
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.
Draw Works in Drilling Rigs 269
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.
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
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.
Δx A = 10Δx B 61
vA = 10vB 62
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.
T 2 = T 1 eμθ
where
θ – angle of contact between belt and pulley
μ – coefficient of friction
Draw Works in Drilling Rigs 271
FA Fast line
Impending motion
Drum
Brake torque
1 F
2
b a
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
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
b r FA
F= 68
a R eμθ − 1
272 Energy Consumption
Kelly (hexagon)
Kelly bushing
Drive pin
Rotary table
Bowl
Master bushing
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
•• 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
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
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
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
276 Energy Consumption
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.
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
Rotary Drive and Drillstring Subsystem 277
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.
(a) (b)
pA
0
a
a Hydrostatic
B n n
Lnp
W
W
WOB
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
Rotary Drive and Drillstring Subsystem 279
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
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.
•• 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
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
Rotary Drive and Drillstring Subsystem 281
Tension and torque limits for the other three grades (X95, G105, S135) can be determined by the
above equations as well.
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
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.
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
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
(a) (b)
Figure 6.8 (a, b) Milled tooth and insert roller drill bits.
284 Energy Consumption
(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.
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).
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.
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)
Force
Bit
nozzle
Hydraulics of Rotary Drilling 287
Hook force
Swivel
Standpipe
Kelly
Rotary hose
Mud pump
BOP stack
Shale shaker
Drill pipe
Mud pits
Tailings L
Drill collars
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
288 Energy Consumption
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
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:
•• Bit nozzle sizes are 11, 11, 12 (total flow area [TFA] = 0.2961 in.2)
Flow rate 360 gpm
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
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
290 Energy Consumption
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.
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.
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
292 Energy Consumption
where
psi
Δγ m = 19 23 6 45
L
Assuming
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:
Drill pipe
Force
Closing ram
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.
294 Energy Consumption
•
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
• 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
pc
x
p c = 0.052 γm x
L
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
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
Surface
p
x
pb = 0.052γm(L–x)
L
Burst pressure
Fc
pT = f ymn − Gc 6 57
D t
1
pE = 46 95 × 106 2 6 58
D t D t−1
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
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].
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
300 Energy Consumption
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
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.
F x = γA L − x − Lγ m A 6 69
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
Lnp
x
F(x) F=0
w (L – x)
L γm A
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.
302 Energy Consumption
Example As an example, consider the design of a production casing having the following
specifications:
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
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
304 Energy Consumption
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
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
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):
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.
5928 psi
pb = 0.052(12)(L – x)
x
5
9500 ft
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.
Bent sub
L1 Tool face
Motor
α
L2
Drill bit
L3
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.
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:
•• Kick off
Establish rate of build
Drill collars
30 – 90 ft
30 ft
Stabilizer
30 ft
40 – 60 ft
15 – 20 ft
Drill bit
(a) (b)
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.
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
1 Dareing, D.W. (2019). Oilwell Drilling Engineering. ASME Press.
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
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.
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
315
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 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.
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
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
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
a M w
V EI
FA x FB
a
L
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
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.
6 ft 6 ft
A B
1 5
RA 2 RB 4 ft
ϕ
4
3
6 ft
10 kips
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:
320 Dynamics of Particles and Rigid Bodies
A B
F1 F5
F2 F4
RA = 2.5 kips ϕ RB = 7.5 kips
F3
10 kips
Fy = 0
RA − F 2 cos ϕ = 0
4 7 12
2 5 − F2 = 0
5
F 2 = 3 125 kip compression
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
F 1 = 1 875 kip compression
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
Combining these two equations gives
Kinematics of Particles 321
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
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
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
Kinematics of Particles 323
y e𝜃
y
er
R r
θ θ
x x
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:
324 Dynamics of Particles and Rigid Bodies
eθ
−∆θeθ
∆θ
er
3
y-coordinate
2
r
1
θ
0
–4 –3 –2 –1 0 1 2 3
–1
–2
x-coordinate
r = A a eaθ θ 7 34
r = Aa e θ + θae
aθ aθ
7 35
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
Kinematics of Particles 325
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
s
en
y(x)
x ρ
x
326 Dynamics of Particles and Rigid Bodies
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
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
Kinematics of Particles 327
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
d2 y π 2 π π 2
= −H sin = −H
dx 2 x = L2 L 2 L
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.
Station 1 Δy
y (North)
ϵ
Δx
CL
Δz
x (East)
α
z
Station 2
y (North)
R
ß
1
x (East)
2
a1 a2
1
ß/2
A1
B
A2 2
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
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
332 Dynamics of Particles and Rigid Bodies
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.
β
Current location
R
a Target
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:
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.
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
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.
B
VB = 2 fps
6 in.
IC
d
4 in.
A
VA = 5 fps
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
g = 32 2 ft s2 English
SI Na kg meter t, s
a
English lb slug feet t, s
a
Refers to derived units, the others are base units.
336 Dynamics of Particles and Rigid Bodies
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
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.
Dynamics of Particles 337
Dynamic Analysis
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
338 Dynamics of Particles and Rigid Bodies
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
ℓ 2
f
δmax
N W
k
The solution is
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
PE 1 = Wy1
W W
y1 y1
y2
y2
Datum Datum
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
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
Dynamics of Particles 341
p1 V2 p V2
+ 1 = 2 + 2 7 86
γ 2g γ 2g
γ 2
p1 − p2 = Δp = V V 2 is replaced by V 7 87
2g
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
γV 2
Δp = 7 90
1120
The continuity flow equation is
0 32Q
V= fps 7 91
An
By substitution
γQ2
Δp = 7 92
10 938An 2
where
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
Smooth surface
F(t)
V1 V2
T
1 2
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.
Control volume
V2
1
ΔS1
V1 ΔS2
2
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:
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.
Dynamics of Particles 345
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
Stator blade
UH
β
a b
V1 u
θ
U Rotor blades
z
c d
β
u
V2/b
U
V2
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
Dynamics of Particles 347
where
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
γ 2 1−ε Q
Pout = Q −U U 7 111
g Aλ tan β
348 Dynamics of Particles and Rigid Bodies
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
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).
Torque
Stall torque
Power
Runaway speed
Output speed
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
MP = IPα 7 127
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.
352 Dynamics of Particles and Rigid Bodies
maG
–x
y G
–y
maP α
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
Dynamics of Rigid Bodies 353
M IG α
F1
m aG
G
r1
F2
r2 d
P P
F F
G G
P d
4F
α=
3rm
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
354 Dynamics of Particles and Rigid Bodies
12 ft
3 ft ma
10 ft
50 lb
20 lb
FA FA
M G = 0; 6F B + 3 20 − 6F A = 0
F x = ma
50
20 = a
32 2
a = 12 88 ft s2
M0 = I0θ 7 132
L
− WL + w sin θ = I 0 θ 7 133
2
Dynamics of Rigid Bodies 355
RA
W RB
a
L
I cg = I 0 − L − a 2 m 7 144
I cg = 1 3671 − 42 386
6
= 1 1183 in -lb-s2
100 N
0.2 m 0.5 m
100 N T MaG
IG α
=
78.48 N
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
aG = 0 5α
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
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
e
Stator
cg Fr
Rotor
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
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
Dynamics of Rigid Bodies 361
m1 aG1
IG1 α1 F1 M y
F
IG2 α2
cg
cg x
+M
z
d1
F2 d2 m2 aG2
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
x
T2, ω2, α2
F
T1, ω1, α1
r1 I1 I2
y
P
r2
TM F
Motor
T0 (external)
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
Dynamics of Rigid Bodies 363
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
h1
h2 s 1
θ
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.
Position 1 Position 2
M
G
G
F1
F2
F = maG 7 167
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
F x dt = H x2 − H x1 7 173
F y dt = H y2 − H y1 7 174
which is rewritten as
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
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):
and
0+ N − W dt = 0 7 178
f dt = mv2
1
I 0 ω1 + M − rf dt = I 0 ω2 7 179
1
Dynamics of Rigid Bodies 367
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
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
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
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
giving
2
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
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.
Time
Δt
F
Δx
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
Dynamics of Rigid Bodies 371
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
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.
M 0 dt = G0 2 − G0 1 7 197
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
372 Dynamics of Particles and Rigid Bodies
Armature, JA
Gear 1
Flywheel, JF
Tin Clutch
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
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
Dynamics of Rigid Bodies 373
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
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
TΩ
HP = = 3 hp 7 204
5252
1 dΩ
T t − T R = J A + J 1 + J F + 0 25J 2 7 206
2 dt
374 Dynamics of Particles and Rigid Bodies
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
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.
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
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
Normal
Shear
0.5
Stress, psi
0
–90 –45 0 45 90
–0.5
Plane orientation, degrees
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
Fn = 0 83
σ n = σ x cos 2 θ + σ y sin 2 θ + 2τxy sin θ cos θ
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
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
Theory of Stress 379
150
100
Shear
50
Stress, MPa
42
0
–180 –90 0 90 180
–50
Normal
–100
–150
Angle of inclined plane, degrees
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
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
380 Mechanics of Materials
15
10
5
Tangent, 2θ
0
–90 –45 0 45 90 135 180 225 270
–5
–10
–15
Angle, 2θ
p2
t
9850 n
8849 psi
1000
–12º 33º
x
p1
10 850 psi
θp = − 12
and
θp = 78
To determine which principal stress applies to which plane, substitute, say θp = −12 into
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
Theory of Stress 381
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.
τ
τmax
y axis
σp2 σp1
σ
2θ
σy τxy
x axis
σx
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
τ 8.06
6 ksi y axis
9.06
8 ksi 29.74º
σ
–7.06 4
4 ksi
x axis
6 8
n
p2
1 ksi
45 º
7.06 ksi
14.87° 8.06 p1
9.06 ksi
x
x axis
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.
384 Mechanics of Materials
ϕ
α
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
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
γ max εx − εy 2 γ xy 2
= ± + 8 31
2 2 2
εx − εy
tan 2θγ max = − 8 32
γ xy
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.
386 Mechanics of Materials
y t Example Consider
εt
εx = 800μ
εy = − 1000μ
γ xy = − 600μ
γ/2
1006.2
y axis
300 1306.2
ε
–700.6
26.56° 450
x axis
p2 εave = 300μ
p2 n
εp2 = –700.6μ
εp1 = 1306.2μ t 45°
p1 p1
x x
13.28° γmax = –2012.4μ
ɛy
σy
σx ɛx
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.
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
•• Principal directions for both stress and strain are the same
Maximum shear stress is 6.71 ksi
13.28°
8 ksi 330μ
4.71 ksi
4 ksi
3 ksi 230μ
8.71 ksi
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
γ/2
1000 μ
x axis
2000 μ
(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.
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
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
c b
90°
a
θa
x axis
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.
Solder
Grid
Lead wires
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
Generalized Hooke’s Law 395
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
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
396 Mechanics of Materials
36 MPa y
24 MPa c 90° b
72 MPa
45°
x
a
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
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
Beam Deflections 397
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.
FL2
C1 = −
2EI
FL3
C2 =
6EI
Combining everything
F 3
yx = − L−x + 3L2 x − L3
6EI
x
x EI
L
398 Mechanics of Materials
which reduces to
Fx 2
yx = − 3L − x 8 73
6EI
FL3
ymax = −
3EI
F
θ x = − − x 2 + 2x 3L − x 8 74
6EI
FL2
θmax = −
2EI
d2 y 1 w 2
EI 2 = −w L−x L−x = − L−x 8 75
dx 2 2
dy w L3
EI = 0 yielding C1 = −
dx x=0 2 3
x
m0 x EI
L
Beam Deflections 399
w L4
y 0 =0 yielding C2 = +
2 12
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
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
x
EI
x
L
400 Mechanics of Materials
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
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
402 Mechanics of Materials
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
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.
Beam Deflections 403
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.
dV T
=q+ 8 90
ds R
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
404 Mechanics of Materials
Well bore
R centerline
dθ
θ
T + dT
M + dM
V + dV
V
w(s) q
T
M
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
Beam Deflections 405
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.
d4 w 2
2d w
− λ =ζ 8 103
ds4 ds2
where
T 1
λ2 = −
EI R2
Rq + T
ζ=
REI
406 Mechanics of Materials
θ T
Tool joint
s
Drill pipe
w(s)
R – radius of centerline
Furthermore, let
Rq + T
χ=
RT − EI
R
1) B + C = 0
2) Al + B + C cosh λl + D sinh λl = 12 χl2
3) A + Dλ = 0
Beam Deflections 407
4) A + Cλ sinh λl + Dλ cosh λl = χl
1 1
Al 1 − sinh λl − C 1 − cosh λl = χl2 8 105
λl 2
A 1 − cosh λl + Cλ sinh λl = λl
or
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
A = 0 015
C = 0 9895
1
D = − A = − 0 9815
λ
B = − C = 0 9895
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
408 Mechanics of Materials
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).
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.
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).
RB
θ R
RC
RA
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
w s = − χs2 + As + B + C cosh λs + D sinh λs 8 108
2
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.
410 Mechanics of Materials
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
Table 8.2 Pull out force over a 90 turn; R = 200 ft (4½ in. 16.5 lb/ft drill pipe).
Top tension
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
412 Mechanics of Materials
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
• 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
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
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
Columns and Compression Members 415
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.
416 Mechanics of Materials
π 2
pi cr Ai = EI 8 128
L
where
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
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
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
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
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.
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].
Columns and Compression Members 419
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.
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.
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.
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
Buckling of Drill Pipe 421
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 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.
422 Mechanics of Materials
α ϕ1 ϕ2 ϕ3 ϕ4
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
Buckling of Drill Pipe 423
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
424 Mechanics of Materials
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].
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.
T
θ + dθ
Fx + dFx
P + dP
V + dV M + dM
q b0
L
dx w + bi dx
V
M
x P Fx
FB
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
426 Mechanics of Materials
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
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
d4 y d dy
EI 4 − F B + wx + L − x Ao γ o − Ai γ i =0 8 163
dx dx dx
Bending Equation for Marine Risers 427
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
F B + LA0 γ 0 − LAi γ i L2
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).
F B + LA0 γ 0 − LAi γ i L2
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
428 Mechanics of Materials
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.
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
Tapered Flex Joints 429
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.
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
430 Mechanics of Materials
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.
2
h−x
I x = Ip 8 171
a
where
L
h=
Ip
1− I SJ
and
Ip
a=h
I SJ
Tapered Flex Joints 431
where
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).
(Continued)
432 Mechanics of Materials
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
Tapered Flex Joints 433
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
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
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
434 Mechanics of Materials
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
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
436 Mechanics of Materials
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
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
Torsional Buckling of Long Vertical Pipe 437
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
438 Mechanics of Materials
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.
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
Torsional Buckling of Long Vertical Pipe 439
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
440 Mechanics of Materials
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.
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
Torsional Buckling of Long Vertical Pipe 441
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
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.
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
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.
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:
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.
Pressure Vessels 443
200
Critical torque, kip-ft
150
100
Tyld = 55826 ft-lb
50
F = –3000 lb
0
0 200 400 600 800
Pipe length, ft
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
pi a2 − p0 b2 − a2 b2 p0 − pi r 2
σθ = 8 210
b2 − a2
444 Mechanics of Materials
σθ
a
σr
b
pi a2 − p0 b2 + a2 b2 p0 − pi r 2
σr = 8 211
b2 − a2
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.
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)
σx + σy σx − σy
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
δi
Inside surface,
outer ring δ0
Interference, c
Contact Surface
where
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
a 9 92
σθ = p= 1512 = 120 ksi 8 224
h 0 125
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
b2 + a2
σθ = p
b2 − a2
σr = − p
pa b2 + a2
δ= +ν
E b 2 − a2
σr
σθ p
σr
b b σθ
a a p
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
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.
Cylinder
Buckle mode
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
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
The solution to Eq. (8.240) is
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.
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.
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
454 Mechanics of Materials
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
r m = 11 18 in
8
y = 8 – 0.5x2
(2, 6)
F
F
Ot
Om ϕ
θ x
4
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.
456 Mechanics of Materials
A
rn = dA
8 265
r
A
The neutral axis is not the centroid axis of the cross section as with the straight beam.
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
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
458 Mechanics of Materials
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.
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
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
3.5 in.
rn
2
0.5 1
F
A
Section A–A
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
460 Mechanics of Materials
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
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.
462 Mechanics of Materials
dx
t
F2 t
b
b
z F3
F1 τ(z)
z
τ(z)
dz
V
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.
464 Mechanics of Materials
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.
r = xi + yj = x 1 α + y1 β 8 291
Note that
Unsymmetrical Bending 465
U U
F
ϕ
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
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:
466 Mechanics of Materials
b
θ
X
C
y
t
x
O
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
Iv = 1.33 Iu = 5.33
3.33
I#, in4
2θ
–2
X axis
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
U y
8” 23.4°
x
¼”
4”
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
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
Unsymmetrical Bending 469
I##
22
y axis
48
26 41
I#
11 2θ
4
16
x axis
y,V
1”
N 1
ϕ = 30°
4”
F
x,U
α
y = 1.75”
2 1”
N
4”
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
470 Mechanics of Materials
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
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.
Beams on Elastic Foundations 471
30° F
M A
30°
U 8 in.
𝛼
6 in.
16 ft
800 lb
4 ft
V 800
M
3200 ft-lb
x
33 000 4 19 200 3
σA = + = 916 psi compression
256 144
The neutral axis angle, α, does not enter into this calculation directly.
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.
472 Mechanics of Materials
d4 y
EI = −q 8 306
dx 4
The magnitude of force, q (force/length) depends on deflection, y.
q = ky 8 307
where
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
M M + dM
y
V V + dV
q Δz
Beams on Elastic Foundations 473
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
y x = C1 eβ 1 + i x + C 2 eβ −1 + i x
+ C 3 eβ −1−i x
+ C4 eβ 1 − i x
474 Mechanics of Materials
Expanding further
Recall
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)
d3 y P
EI =V = −
dx 3 x=0 2
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
b x
p
t
u
t
Δθ
a
L
wL = pbL 8 334
where
d2 u
M t = EI t 8 338
dx 2
478 Mechanics of Materials
t
b
Mt h σθ
u a
σx
t Δθ
σθ
a
where
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
Eh3 d2 u d2 u
Mt = = D 8 343
12 1 − ν2 dx 2 dx 2
Eh3
D= 8 344
12 1 − ν2
Beams on Elastic Foundations 479
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
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.
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”
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.
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
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.
Criteria of Failure 483
Example Consider the case where pipe is loaded by torsional shear, internal pressure plus bend-
ing. Using the following pipe data and loads:
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
484 Mechanics of Materials
τ
τmax
y axis
σ1
2θ σx
σ
σy
σ2
τxy
x axis
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.
Since the maximum shear stress at yield (per uniaxial test data) is ½ axial yield then
τyld = 0 5σ yld 8 357
Criteria of Failure 485
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
σ = σ 21 − σ 1 σ 2 + σ 22 8 361
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
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
• Torque of 6000 ft lb
486 Mechanics of Materials
pi a
M T
5 in. OD
4 in. ID
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.
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.
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
Hydrostatic stress
Suc
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
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.
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.
15 Dareing, D.W. (2019). Oilwell Drilling Engineering. ASME Press.
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)
20 Shigley, J.E. and Mitchell, L.D. (1983). Mechanical Engineering Design. McGraw-Hill Book Co.
490 Mechanics of Materials
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
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.
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
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
492 Modal Analysis of Mechanical Vibrations
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
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
Complex Variable Approach 493
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.
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
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
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
Two Degrees of Freedom 495
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).
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.
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
X2
= − 0 618 second mode shape 9 33
X1 2
M0 = I0α
k1 m k2
µ
a a a a
0 θ
SDOF
y P
cg
2-DOF
d
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
M P = I cg θ + d m + μ4a acg 9 36
The second
M P = I P θ + d m + μ4a aP 9 37
F y = m + M acg 9 39
− k 1 y − aθ − k 2 y + 2aθ = m + M y + dθ 9 40
Two Degrees of Freedom 499
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.
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
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.
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
Undamped Forced Vibrations (2 DOF) 501
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
15
10
Amplitude ratio, X2 / δ
0
0 0.5 1 1.5 2 2.5 3 3.5
–5
–10
–15
Eigenvalues, ξ
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.
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
Undamped Forced Vibrations (2 DOF) 503
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
504 Modal Analysis of Mechanical Vibrations
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
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
6.0
4.0
Amplitude ratio, X1/δ
2.0
0.0
0 0.5 1 1.5 2 2.5
–2.0
–4.0
–6.0
Frequency ratio, ω/ωa
6.0
4.0
Amplitude ratio, X2/δ
2.0
0.0
0 0.5 1 1.5 2 2.5
–2.0
–4.0
–6.0
Frequency ratio, ω/ωa
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
506 Modal Analysis of Mechanical Vibrations
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.
m1
m
x1
k x2
r
P
Multi-DOF Systems – Eigenvalues and Mode Shapes 507
F(t)
k1
a x1 m1
0
k
Absorber
m x2
Example Consider the system shown in Figure 9.15. The equation of motion for the primary
system (the bar) is
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
k − ω2 m X = 0 amplitude equation 9 77
508 Modal Analysis of Mechanical Vibrations
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.
F = k x 9 80
x = α F 9 81
x1 α11
x2 = α21 F1
x3 α31
Multi-DOF Systems – Eigenvalues and Mode Shapes 509
1 2 3
k11
1 in.
k21
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
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
510 Modal Analysis of Mechanical Vibrations
k
k
m
a a m
cg
x1
x2
k
k
m
a a
m
cg
x1 x2
kxcg
kx2
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
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.
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
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
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
Multi-DOF Systems – Eigenvalues and Mode Shapes 513
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
k k k
ω1 = 0 445 , ω2 = 1 072 , ω3 = 1 803
m m m
m x1
m x2
m x3
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
f11
2m m m f21
a a a
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
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
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
Modal Analysis of Discrete Systems 517
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.
X Tr m X s =0 9 132
X Tr k X s =0 9 134
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
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
k c k c k c
m m m
c c
k k k
m m m
c c
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
where
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 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
Free Vibration of Multiple DOF Systems 521
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).
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
522 Modal Analysis of Mechanical Vibrations
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
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
524 Modal Analysis of Mechanical Vibrations
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
c c c
m m m
T T
a a a a
1 2 3
1 1 1
X1 X2 X3 = 2 0 − 2 9 184
1 −1 1
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
4 0 0
M =m 0 2 0 9 187
0 0 4
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
526 Modal Analysis of Mechanical Vibrations
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
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
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.
x1
x2 = X 1 η1 + X 2 η2 + X 3 η3 9 198
x3
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
xi t = X r eiωt 9 201
r=1
Kr 1 − r 2r + i 2ζ r r r
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.
F1 m x1
m x2
m x3
where
F = F r 0 cos ωt 9 206
By substitution
Qr = Qr0 cos ωt 9 207
where
T
Qr0 = X r F0
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.
Forced Vibration by Harmonic Excitation 529
1 0 1 −1 1 −1
m =m c =c k =k
0 1 −1 1 −1 1
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
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
530 Modal Analysis of Mechanical Vibrations
x1
x2 = X 1 η1 + X 2 η2 + X 3 η3 9 216
x3
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
xi t = X r eiωt 9 218
r=1 1 − r 2r + i 2ζ r r r
N T
X r F0 1
xi t = X r eiωt 9 219
r=1
Kr 1 − r 2r + i 2ζ r r r
Note that when rr = 1 for any mode, that mode responds with resonance.
Forced Vibration by Harmonic Excitation 531
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
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
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
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.
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
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
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)
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
534 Modal Analysis of Mechanical Vibrations
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)
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
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
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 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
This equation applies over both drill pipe and drill collar sections.
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].
538 Modal Analysis of Mechanical Vibrations
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.
6 Dareing, D.W. (2019). Oilwell Drilling Engineering. ASME Press.
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
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
542 Fluid Mechanics
Fixed shoe
h
x
U
p pmax
x
Pressure induced
h Velocity induced
U
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
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
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
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
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
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
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.
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
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
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
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
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].
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
Q – flow rate, gpm
L – pipe length, ft
Δp – pressure drop, psi
Δp PV 0 2 γ 0 8 Q1 8
= 7 7 10 − 5 3 18 10 59
L Dh − Dp D h + Dp
Power = pQ
p,Q
Liner
T, N
Motor
Manifold
where
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.
Liner (in.) pmax (psi) Fmax (lb) Flow rate (gpm) HHP (hp)
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].
3500
3000
2500
Pressure, psi
2000
1500
Parasitic losses
1000
500
0
0 200 400 600 800 1000
Flow rate, gpm
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
1600
1400
Hydraulic horsepower, hp
1200
1000
Pump Parasitic
800
600
Available
400
200
0
0 200 400 600 800 1000
Flow rate, gpm
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
686 400
HHP f = = 160 hp
1714
Q 144
V an =
π
D2h − D2dp 7 48
4
Q
V an = 24 51 10 60
D2h − D2dp
where
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.
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.
Power Demands of Downhole Motors 557
700
600
500
Horsepower, hp
400
300
200
100
0
0 200 400 600 800 1000
Flow rate, gpm
250
Vav , fpm
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.
1 2
T, N
Q
p1 p2 Dynamometer
PDM – turbine
Force
Stall torque
Δp
Maximum operating pressure
10% reduction in Q
Speed, N (rpm)
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
Power Demands of Downhole Motors 559
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.
580
Q, gpm
200 475
HHP
160.7
67.68
Q, gpm
200 475
Pump
rate (gpm) Bit speed (rpm)
Tool Max Torque Power
size (min) (max) (min) (max) diff. (psi) (ft-lb) range (HP) HHP
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.
Table 10.4 Hydraulic horsepower consumed 63 4 in. turbine (12 ppg) (HHP is
developed from Eq. (10.64)).
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.
600
500
Hydraulic horsepower
Available
400
300
200
PDM
100
0
0 200 400 600 800
Flow rate, gpm
700
600
Hydraulic horsepower
500
Available
400
300
Turbine
200
100
0
0 200 400 600 800
Flow rate, gpm
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.
Fluid Flow Around Vibrating Micro Cantilevers 563
Mathematical Model
Key assumptions made in the mathematical development of fluid flow adjacent to microcantilevers
surfaces are
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
Control
volume
y
h
a a
x
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.
∂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 Flow Around Vibrating Micro Cantilevers 565
∂2 u ∂2 u p π πx 0.7
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
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.
2 Dareing, D.W. (2019). Oilwell Drilling Engineering. ASME Press.
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
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.
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
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.
h2 h3
V a + h = V a + hV a + V a + V a + 11 2
2 3
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
570 Energy Methods
V
Unstable
Unstable
Stable
h
x
a
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
1 2 1
ΔV = δV + δV+ + δ m−1 V + 11 4
2 m−1
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.
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
Water line
c b
B
mc
B
Water line
c θ
b
W
Water line
V
c
b
θ
𝜋
B 0
2
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
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
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
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 (°)
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
The application of the principle of minimum potential energy generally follows the steps below.
EA =C
a
1
EA b = C
2
3
EA c = C
2
Multiple Degrees of Freedom 575
u
100 lb
v
a c
b H
45° 60°
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
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
∞
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
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
x
EI
x dx
L
y
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
ds2 = dx 2 + dy2
2
dy
ds = 1+ dx
dx
dx
P
P x
ds EI
Lʹ e
L
y
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.
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
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
Member Ni (kip) Stress (psi) Strain (in./in.) Length (in.) ei (in.) ni (kip) eini
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
Principle of Complementary Energy 583
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
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.
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
Recall
2π
2π
θ sin 2θ
sin 2 θ dθ = − =π
2 4 0 Figure 11.13 Transverse displacement of a
0
split ring.
Principle of Complementary Energy 585
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
0 l a
y
x
T
b
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
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.
Principle of Complementary Energy 587
z Bent wire
T M
2
r
θ 1 s
y
r
x F
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
Fr 3 1 3πFr 3
δ= + 2π + 11 82
EI 3 GJ
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
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.
Chemically Induced Deflections 589
2 3
1 a
b
2c n n
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
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 11.20 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
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
−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
Engineering Practice with Oilfield and Drilling Applications, First Edition. Donald W. Dareing.
© 2022 John Wiley & Sons, Inc. Published 2022 by John Wiley & Sons, Inc.
594 Index
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