Strand Erling PDF

Faculty of Science and Technology
MASTER’S THESIS
Study program/ Specialization:

Spring semester, 2014
Petroleum Engineering / Drilling Technology
Open
Writer:
Erling Strand …………………………………………

(Writer’s signature)
Faculty supervisors:
Mesfin Belayneh
Kjell Kåre Fjelde
Thesis title:
Hydraulic Calculations Using Discovery Web for Visualization
Credits (ECTS): 30
Key words:
Pages: 82
Integrated Operations
Well Control + enclosure: 18
Real-time monitoring
Discovery Web
Rheology and Hydraulics Stavanger, 12.06.2012
Drilling Scenario Simulations
Abstract
Advanced real-time monitoring systems are useful tools for safe and cost effective well
operation practices. As one moves into deeper water, higher pressure and higher temperature,
the drilling operations only becomes more challenging. Since the operational drilling window
is getting narrower, implementation of good IO technology (i.e. real-time data, technology
and people) are necessary to increase safe operations, increase productivity, enhance HSE and
reduce NPT.
This thesis is divided into a theoretical and a simulation part. The theoretical part presents
major downhole drilling problems related to hydraulics, prevention and remedial actions. The
theoretical part emphasizes on describing different rheological models for hydraulic
calculations. The simulation part presents the real-time monitoring system Discovery Web
application developed by Kongsberg Oil & Gas Technologies. In the simulation part, the
rheology models are used for hydraulic calculation in different drilling scenarios. Pressure at
different locations in the well (Pump pressure (bar), BHP (bar) and ECD (sg)) has been
calculated. Different events and unwanted situations are considered and different visualization
views have been provided to demonstrate how different well parameters will develop. The
input parameters to the models have typically been ROP, flow rate, rheology data and mud
density.
In order to illustrate the applicability of the implemented models, a case study is presented
while drilling a vertical well from 4000m. The following drilling scenarios have been
simulated and the results are briefly discussed. These are:
− Connection scenario
− Kick scenario during drilling
− Kick scenario during connection
− Pack-off scenario and sensitivity of pack-off
− Lost circulation scenario
− ROP vs cutting concentration scenario
− Hydraulics and rheology model comparison scenario
− Washout scenario
By building a monitoring panel in Discovery Web, based on the proper rheological models
and hydraulic calculations, this thesis have been used to show how models and simulations
can be combined in Discovery Web. The experience has been that it was easy to implement
the model using Discovery Web Formulas. Another strength of the software is that it is very
easy to visualize and import data from real wells in this application. Hence, it is a very good
tool for comparing models with real data. This real-time data handling capability and
visualization flexibility is considered as one of the major strengths. It has also been shown
how to embed events into the models, which later can be used as a basis for developing
training scenarios and demonstrations in teaching. The results show how the models
implemented can be the first step in introducing models in combination with real-time data for
monitoring and handling drilling problems using this application.
2
Acknowledgement
This thesis was written for the Department of Petroleum Engineering at the University of
Stavanger in collaboration with Kongsberg Oil & Gas Technologies.
I would like to use this opportunity to thank the several people that have helped me in the
work with this thesis. I would first like to express my appreciation to the Discovery Web team
for allowing me to take on this assignment as well as the material and intellectual support that
they have contributed in this process.
My deepest gratitude goes to my Discovery Web mentor Jan Kåre Igland whom have
supported and educated me in drilling applications. His engagement in discussions and
tutoring skills has been a most valuable asset in this process.
Finally, I would like to thank my advisors at the University of Stavanger Mesfin Agonafir
Belayneh and Kjell Kåre Fjelde. They have provided excellent academic guidance and
advisory through the process. Their engagement in discussions and knowledge within Drilling
Technology has been a major motivation staying on the right track while working on this
thesis.
3
Table of Contents
Abstract .................................................................................................................................................... 2
Acknowledgement ................................................................................................................................... 3
Table of Contents ..................................................................................................................................... 4
Table of Figures ....................................................................................................................................... 5
Abbreviations ........................................................................................................................................... 6
Nomenclature ........................................................................................................................................... 7
1. Introduction .......................................................................................................................................... 9
1.1 Background .................................................................................................................................... 9
1.2 Problem formulation .................................................................................................................... 11
1.3 Objective of the thesis ................................................................................................................. 13
2. Hydraulic Models ............................................................................................................................... 14
2.1 Rheology ...................................................................................................................................... 14
2.1.1 Newtonian fluids .................................................................................................................. 15
2.1.2 Non-Newtonian fluids. ......................................................................................................... 15
2.1.3 Comparison and error calculation of rheology prediction ................................................... 20
2.2 Hydraulics .................................................................................................................................... 21
2.2.1 Pressure losses ...................................................................................................................... 21
2.2.2 Bit pressure losses ................................................................................................................ 23
2.2.3 Cuttings concentration.......................................................................................................... 23
2.3 Modeling in Discovery Web ........................................................................................................ 25
2.3.1 Bingham frictional model..................................................................................................... 25
2.3.2 Herschel-Bulkley frictional model ....................................................................................... 27
2.3.3 Robertson-Stiff frictional model .......................................................................................... 29
3. Drilling Parameters, Problem Detection and Remedial Action ......................................................... 31
3.1 Drilling parameters ...................................................................................................................... 31
3.2 Typical equipment and wellbore problems.................................................................................. 34
3.3 Stuck pipe .................................................................................................................................... 38
3.4 Lost circulation ............................................................................................................................ 45
4. Examples of other Software Tools ..................................................................................................... 48
4.1 Sekal DrillScene .......................................................................................................................... 48
4.2 eDrilling Solutions ....................................................................................................................... 49
4.3 DrillBench ................................................................................................................................... 51
5. Architecture of Discovery Web ......................................................................................................... 53
6. Discovery Web Real-time ECD Control Design ............................................................................... 57
6.1 Simulation based on our built data .............................................................................................. 57
6.1.1 Connection scenario ............................................................................................................. 58
6.1.2 Kick scenario during drilling................................................................................................ 59
6.1.3 Kick scenario during connection .......................................................................................... 60
6.1.4 Pack-off scenario and sensitivity of pack-off ....................................................................... 61
6.1.5 Lost circulation scenario ...................................................................................................... 65
6.1.6 ROP vs. Cuttings concentration scenario ............................................................................. 69
6.1.7 Comparison of different rheology models for hydraulic calculations .................................. 70
6.1.8 Washout scenario ................................................................................................................. 71
6.2 Simulation based on real well based data .................................................................................... 73
7. Summary and Discussion ................................................................................................................... 74
8. Conclusion ......................................................................................................................................... 77
9. Future Work ....................................................................................................................................... 78
References .............................................................................................................................................. 80
Appendix A ............................................................................................................................................ 83
Appendix B ............................................................................................................................................ 84
Appendix C ............................................................................................................................................ 87
4
Table of Figures
Figure 1 - Pore pressure diagram illustrating the drilling window. ......................................... 10
Figure 2 - Bingham-plastic model. .......................................................................................... 16
Figure 3 - Herschel-Bulkley fluid rheogram. ........................................................................... 17
Figure 4 - Robertson-Stiff fluid rheogram. .............................................................................. 19
Figure 5 - Comparison of rheology prediction. ....................................................................... 20
Figure 6 - Pressure drops during circulation. ........................................................................... 22
Figure 7 - Illustration of drill collar without centralizer sticking in a well [26]. ..................... 42
Figure 8 - Well path for stuck point derivations. ..................................................................... 43
Figure 9 - Illustration showing total loss of circulation. .......................................................... 45
Figure 10 - DrillScene: Drilling through a depleted region. Time frame is 2 minutes [31]. ... 48
Figure 11 - DrillScene: MPD solution based on back-pressure [30]. ...................................... 49
Figure 12 - eDrilling: Pump rate [32]. ..................................................................................... 50
Figure 13 - eDrilling: Calculated bottomhole ECD [32]. ........................................................ 50
Figure 14 - DrillBench: Kick during connection [34]. ............................................................ 51
Figure 15 - SiteCom setup [35]. ............................................................................................... 53
Figure 16 - Workflows of SiteCom [35]. ................................................................................. 54
Figure 17 - Basic drilling console - DrillingTime Playback UiS [37]. .................................... 55
Figure 18 - Calculation model Discovery Web Formula and Arithmetic Smart Agent [37]. .. 55
Figure 19 - Discovery Web Formula [37]................................................................................ 56
Figure 20 - Well schematics based on our built data. .............................................................. 57
Figure 21 - Connection scenario. ............................................................................................. 58
Figure 22 - Kick scenario during drilling. ............................................................................... 59
Figure 23 - Kick scenario during connection. .......................................................................... 60
Figure 24 - Illustration of pack-off scenario. ........................................................................... 61
Figure 25 - Pack-off scenario 1 (10% reduction in annulus diameter). ................................... 62
Figure 26 - Pack-off scenario 2 (16% reduction in annulus diameter) .................................... 63
Figure 27 - Pack-off scenario 3 (20% reduction in annulus diameter). ................................... 64
Figure 28 - Illustration of total loss of circulation at 9 5/8" casing shoe. ................................ 65
Figure 29 - Lost circulation scenario 1 (Partial loss at 9 5/8" casing shoe). ............................ 66
Figure 30 - Lost circulation scenario 2 (Total loss at 9 5/8" casing shoe). .............................. 67
Figure 31 - Lost circulation scenario 3 (Total loss at 4120 m). ............................................... 68
Figure 32 - ROP vs. Cuttings concentration scenario. ............................................................. 69
Figure 33 - Comparison of different rheology models for hydraulic calculations. ................. 70
Figure 34 - Illustration of washout scenario in drillpipe (Intellipipe). .................................... 71
Figure 35 - Washout scenario in Intellipipe. ............................................................................ 72
Figure 36 - Temperature effect on density [39]. ...................................................................... 78
5
Abbreviations
(B) - Bingham
(H-B) - Herschel-Bulkley
(R-S) - Robertson-Stiff
AV - Apparent Viscosity
BHA - Bottom Hole Assembly
BHP - Bottom Hole Pressure
CTFV - Critical Transport Fluid Velocity
CTV - average Cuttings Travel Velocity
D&E - Drilling and Evaluation
DDM - Derrick Drilling Machine
E&P - Exploration & Production
ECD - Equivalent Circulating Density
ERD - Extended Reach Drilling
FIT - Formation Integrity Test
FLOWIN - Flow Rate In
FLOWOUT - Flow Rate Out
FRT - Free Rotating Torque
HPHT - High Pressure High Temperature
HWDP - Heavy Weight Drill Pipe
IO - Integrated Operations
IRIS - International Research Institute of Stavanger
KOG - Kongsberg Oil & Gas
LC - Lost Circulation
LCM - Loss Circulation Material
LOT - Leak-Off Test
LPM - Liter Per Minute
LWD - Logging While Drilling
MPD - Managed Pressure Drilling
MWD - Measurement While Drilling
MWIN - Mud Weight In
MWOUT - Mud Weight Out
NCS - Norwegian Continental Shelf
NPT - Non-Productive Time
OBM - Oil Based Mud Systems
OH - Open Hole
OPC - Standardized Alarm System
PDC - Polycrystalline Diamond Compact
POOH - Pull Out Of Hole
PUW - Pick-Up Weight
PV - Plastic Viscosity
PVT - Pressure Volume Temperature
RIH - Run In Hole
ROP - Rate Of Penetration
RPM - Revolutions Per Minute
SCFF - Subcritical Fluid Flow
SOW - Slack-off Weight
SPP - Stand Pipe Pressure
TD - Target Depth
TOC - Top of Cement
UBO - Underbalanced Operations
WOB - Weight On Bit
YP - Yield Point
6
Nomenclature
a - Frictional fractions parameters, dimensionless
A - Pipe Area, m!
A - R-S model parameter similar to K, lbf. sec ! /100 ft ! or dyne. sec ! /100 cm!
A1 - Area of drill pipe 1, m!
A2 - Area of drill pipe 2, m!
An - Total nozzle area, m!
b - Frictional fractions parameters, dimensionless
B - Robertson-Stiff model parameter similar to n, dimensionless
C - Orifice coefficient (C: 0.80 for non-jet bit, C: 0.95 for jet bit), dimensionless
C - Robertson-Stiff model correction factor, 1/sec !
Ca - Herschel-Bulkley model parameter, dimensionless
Cc - Herschel-Bulkley model parameter, dimensionless
d - Density, sg or ppg
d - Diameter of pipe, in or m
𝑑! - Hydraulic diameter, in or m
𝐷!!"# - Diameter between annulus and pipe, in or m
D1 - Outside pipe diameter, in or m
D2 - Inside casing diameter, in or m
DN - Nozzle diameter, in
Dp - Inside pipe diameter, in or m
E - Young’s E-Modulus, Pa
ECD - Equivalent Circulating density, sg
F - Force, Pa
𝑓! - Fanning friction factor, dimensionless
f - Friction factor, dimensionless
fa - Friction factor to the annulus, dimensionless
fp - Friction factor to the pipe, dimensionless
g - Gravity, m/𝑠 !
h - Height, m
K - Consistence index, lbf. sec ! /100 ft ! or dyne. sec ! /100 cm!
k - Surface roughness, m
L - Length, ft
𝐿!" - Drill collar length, ft or m
n - Flow behavior index, dimensionless
NRe - Reynolds number, dimensionless
NRe Cr - Critical value of Reynolds number, dimensionless
P - Pressure, psi
Pp - Pump pressure, psi or bar
Q - Flow rate, gallon/min or lpm
R - Build or sail section radius, m
𝑅! - Length of b in Figure 7, m
𝑅! - Well radius, m
ROP - Rate of Penetration, m/h
s - Source (phase transfer between phases)
t - Time, s or hr
u - Fluid velocity, m/s or ft/s
va - Annular average velocity, ft/sec
7
Vavg - Average velocity, ft/sec or m/s
VCr - Critical velocity, ft/sec or m/s
vp - Pipe average velocity, ft/sec
w - Weight of pipe, kg/m
xint - Factor for interpolation, dimensionless
y - Herschel & Bulkley model parameter, dimensionless
z - Herschel & Bulkley model parameter, dimensionless
α - Angle, degree
β - Buoyancy, dimensionless
𝜀 - Surface roughness coefficient, dimensionless
𝜌 - Density, lbm/gal, sg
φ - Porosity, precent
𝛾 - Shear rate, 1/sec
𝛾∗ - Shear rate value corresponding to geometric mean of the shear stress, 𝜏 ∗ , 1/sec
𝛾!"# - Maximum shear stress value of data, 1/sec
𝛾!"# - Minimum shear stress value of data, 1/sec
𝜃! - Reading from rheometer at 3RPM
𝜃! - Reading from rheometer at 6RPM
𝜃!"" - Reading from rheometer at 100RPM
𝜃!"! - Reading from rheometer at 300RPM
𝜇 - Viscosity, cP
𝜇! - Plastic viscosity, cP
𝜏 - Shear stress, lbf/100 ft !
𝜏∗ - Shear stress value corresponding to the geometric mean of the shear rate, 𝛾 ∗
𝜏!"#! - Calculated shear stress, lbf/100 ft !
𝜏!"# - Maximum shear stress value of data, lbf/100 ft !
𝜏!"#$ - Measured shear stress, lbf/100 ft !
𝜏!"# - Minimum shear stress value of data, lbf/100 ft !
𝜏! - Yield stress, lbf/100 ft !
𝜏! - Yield point, lbf/100 ft !
%C - Cuttings concentration, precent
∂ - Partial derivative, dimensionless
∆ - Gradient
∆𝐿 - Segment length, m
∆𝑃 - Pressure, kPa or bar
∆𝑃! - Frictional pressure drop, bar
∆𝑃!" - Annulus pressure drop, bar
∆𝑃!" - Drill bit pressure drop, bar
∆𝑃!" - Nozzle pressure drop, bar
∆𝑃!" - Downhole mud motor pressure drop, bar
∆𝑃!" - Total pressure drop, bar
!"
!"
- Gradient pressure, psi/ft
8
1. Introduction
Presently the oil industry is showing advancement in technology and methods to solve
operational problems that the conventional methods cannot handle. The industry is facing
drilling challenges such as deep-water, depleted formations, high pressure and high
temperature formations, gas hydrate formations and extended reach drilling environments.
New technologies allow us to explore and to produce from more and more challenging oil
fields.
Application of integrated operation (IO) practices is important for reaching these strategic
goals. IO practices utilize appropriate technologies, competent people, real-time data
(gathering and processing) that are crucial for proper decision-making. The overall advantage
of IO is to reduce non-productive time (NPT), increase productivity, increase safe operations,
and decrease the risk of HSE accidents.
This thesis presents the real-time data application tool Discovery Web, where the focus is on
implementing rheological models for pressure calculations and demonstration, and
visualization of different drilling scenarios.
1.1 Background
Exploration and production (E&P) comprises of all the activities that are involved whitin
exploration, development and production of hydrocarbons. When drilling through different
formations one may encounter undesired well instability problems, such as well fracturing
and well collapse. Well fracturing is due to well pressure exceeding the formation fracture
pressure. This may result in formation fracture and huge mud losses into the formation. Well
collapse occurs when the well pressure is lower than the collapse pressure of the formation.
The problem results in bridging and pack-off. If this happens, it may cause drillstring
mechanical sticking, and increase well pressures. This will be reflected in an increase in pump
pressure (SPP). In the worst-case scenario, the drillstring needs to be cut and a sidetrack
procedure initiated. Also, when drilling through a reservoir, well pressure lower than the
formation pressure, may cause undesired influx of formation fluid into the well. This as a
result causes a kick and in the worst-case scenario it may cause a blowout. The Macondo
project can be referred to as an example. The drilling problems are still challenging for the oil
industry and are huge cost factors [1].
Wellbore condition evaluation is based on detailed process models that are capable of
predicting not just only downhole hydraulics, but also thermal and mechanical effects during
a drilling operation. Nevertheless, how advanced these models are, there will always be
uncertainties in input parameters, modeled effects and outcome. Exploration in deep water,
HPHT wells limits the drilling window even more. This makes it almost impossible to ensure
precise predictions under any circumstances [2].
Figure 1 is a typical figure found in drilling programs, which shows the safe operational
drilling window. During operation, it is important to maintain the well pressure within the
allowable drilling window. In this project the behavior of ECD with respect to the flow
behavior will be monitored in Discovery Web.
9
Figure 1 - Pore pressure diagram illustrating the drilling window.
The dynamic circulation pressure is determined by the static mud weight and the dynamic
friction loss as given in Eq. 1:
∆𝑃! (1)
𝐸𝐶𝐷 = 𝜌 +
𝑔ℎ
where ECD is equivalent circulation density, g is acceleration due to gravity, h is true vertical
depth, ρ is static mud density and ΔPf is friction loss. In some cases, there will also be a
choke pressure like some of the managed pressure drilling (MPD) systems, as well as when
circulating out a kick in a conventional well.
Drilling is the process that is used to connect the reservoir to surface, recovering potential
hydrocarbons. During drilling, the drill bit intrudes several geological formations on its way
down to target depth (TD). Knowledge about these formations is a key for drilling to the
desired depth. By analyzing wireline logs, MWD, LWD, cores and cuttings, it is possible to
determine the formations different properties. Wells are drilled using rotating bits and drilling
fluid as circulation fluid. The drilling fluids have many functions, but the most important is to
maintain the pressure in the borehole. Thus, maintaining the well pressure within the drilling
window, avoiding fracture and collapse of formation. In addition one must also stay above
pore pressure. Fluids are therefore a key element in the drilling process. It is therefore
important to accurately predict the rheology and hydraulics at all times while drilling a well.
This thesis presents hydraulic calculations and visualization using Kongsberg Oil & Gas
Technologies real-time monitoring program, Discovery Web. Discovery Web is a web-based
browser that helps us reach out to all the involved by implementing a visualization and
collaboration tool for multi disciplinary target groups [3].
10
1.2 Problem formulation
During drilling operations today, many operators are facing increased non-productive time
(NPT). The NPT are due to drilling incidents (pack-off, kick, poor hole cleaning, excessive
torque, fracture, collapse and lost circulation). If preventive measurements had been taken,
many of these problems could have been avoided, or at least reduced the impacts affecting the
drilling operations. There have been a few presented methodologies that are governing this
proactive measurement [2]. Falconner presented in 1989 a work process that has the
advantage of simplicity and does not require advanced computer systems for analyzing real-
time data, but instead require special friction tests to be performed at regular intervals (for
example between each connection). The method records pick-up weight (PUW), slack-off
weight (SOW) and free rotation torque (FRT) while drilling, and comparing the measured
data with simulations performed in advance based on torque and drag charts from the planned
well path. If there are deviations from the pre-calculated data, the rate of penetration (ROP)
may be reduced or hole cleaning procedures may be initiated to improve the downhole
conditions [4]. Niedemayr developed another method in 2010, which performs automatic
analysis of all bottom hook-loads and torque measurements. By using this system it is
necessary to implement an external mechanical friction model to obtain the required torque
and drag charts [2, 5].
The hydrodynamic force exerted on the inner and outer part of drilling string is caused by the
rheology and circulation of drilling fluid, which again influences the drag force on the
drillstring. This makes it necessary to perform friction tests with no circulation or low
circulation rates to obtain comparable results. However, reduction of pump rate while taking
PUW and SOW is not always the case, and essential data that could have been used to
determine the downhole conditions are lost. Cayeux et al [6] developed in 2009 a continuous
monitoring system using an embedded torque and drag model coupled with hydraulic
calculations. By performing systematic analysis of all possible off-bottom weights and torque
and with any flow rate, this model makes it possible to monitor friction in other conditions
than drilling, like for instance running in hole (RIH), pulling out of hole (POOH) or back-
reaming to casing shoe [2, 6]. The model was tested out on 5 fields and a total of 15 different
wells. It warned in almost all cases about the evolution of poor downhole conditions prior to
actual drilling incidents. It was only in a few specific instances that the monitoring system
failed in showing advanced warning signs. Dependent on the different conditions from fast to
slow changing conditions, the warning signs was visible between 30 minutes to 1 day before
the actual incident, respectively [2].
Early symptom detection, armed with real-time calibrated process models, will help us to
manipulate hydraulic parameters and avoid undesired events. By introducing the model to
multiple drilling problem symptoms, the possibility for decreasing the NPT increases. The
combined use of real-time data from Discovery Web and standard hydraulic calculations
makes us able to regulate the ECD, by manipulation of rheology, geometry, flow rate, mud
densities and ROP. Drilling is challenging when the window between the pore pressure and
fracturing pressure becomes narrower. If the well pressure crosses the operational window,
drilling problems such as kick, loss circulation or stuck pipe can occur.
11
The operational challenges include:
• Deep water:
o Here the operational window is narrow. In addition, the formation can have
higher temperature and pressure, which strongly affects the property of drilling
fluids.
• Depleted reservoirs:
o Here the formation pressure and the in-situ stresses of the formation are much
lower than normal. Hole instability and formation damage may be a risk when
drilling with a too high mud weight.
• Drilling in methane hydrate:

o Drilling in gas hydrated formation leads to dissociation of gas out of hydrates.
This as a result changes the properties of the drilling fluid and may cause a
minor kick, which again may affect the drilling platform.
• Drilling extended reach horizontal wells:

o In horizontal wells, the window between the fracture and collapse pressure are
narrower. Horizontal wells may be exposed to well instability problems if one
cannot properly manage the well pressure within the drilling window.
The overall problem occurs within these problematic areas and therefore increases the
operational costs and creates impacts on health and environment. The oil and gas industry still
continues to fight against borehole and string related issues. The well pressure with respect to
the formation pressure and strength determines what kind of problems that can occur. In
average, wellbore instability problems alone increased the total drilling budget by 10% [7].
Non-productive time (NPT) has usually a direct link to i.e. the time spent due to lost
circulation, stuck pipe and tripping in and out. These problems cost operators a significant
amount of money. There are also other forms of NPT that are invisible NPT. For example, the
presence of vibration may cause a reduction in drilling rate. In addition, the vibration may
also cause drill bit damage, which results in tripping in and out for changing the bit.
Unwanted formation influx into the well is also a cost factor when having to circulate the kick
out of the well. Optimization procedures do not only help reduce the visible NPT but it will
also have a large impact on the invisible NPT as well. Studies from several drilling wells
shows that NPT cost contributes to about 25-30% of the total drilling cost [8].
This thesis will analyze the mentioned problems based on literature review material and by
building a monitoring panel in Discovery Web. This thesis therefore addresses issues such as:
1. How can integrated operations (real-time data) help reduce the possible problems,
which indirectly reduces the NPT?
2. How sensitive the hydraulic parameters are with regard to the different types of events
or unwanted problems that can occur?
3. Can Discovery Web be a useful tool in monitoring live wells?
4. Can Discovery Web be a useful tool for professionals and university level educational
training?
5. Can real-time data contribute in improving drilling performances?
12
1.3 Objective of the thesis
The objective of this thesis is to show how models and simulations can be used in handling
and monitoring possible problems that arises in operational challenging environments. This
thesis provides the development of hydraulic visualization scenarios using Discovery Web;
how we can use Discovery Web Formulas for calculating well pressures and how this could
be visualized to demonstrate important pressure effects in the well.
Ø Provide an overview of the following:

• Drilling parameters that are measured during drilling and what they represent.
• Typical drilling problems that can occur.
• Symptoms of the different drilling problems (trend changes in drilling parameters).
• Examples where software tools have been used for well diagnostics and literature
review.
Ø Describe the functionality of Discovery Web.
Ø Implement different rheology models in Discovery Web for hydraulic calculations.
Ø Develop hydraulic visualization views for different drilling scenarios using Discovery
Web.
Ø Discuss the potential in using Discovery Web, how it can be used for engineering and
education purposes, and potential recommendations for future work.
13
2. Hydraulic Models
2.1 Rheology
Rheology is a Greek word that comes from the words reo, meaning to float, and logy meaning
science. Rheology deals with the study of deformation and flow of matter (mainly liquids and
in some cases solids and soft solids). In short, the drilling fluid flow property is characterized
by their rheological properties, which are a function of composition of the drilling fluid,
temperature and pressure. Drilling fluid helps to remove cuttings from the wellbore by
keeping the cuttings in suspension during drilling. Other characteristics are minimizing
friction during pumping, minimizing impact on the formation as we drill and being able to
separate the cuttings at surface. It is important analyze fluid flow velocity profiles, fluid
viscosity, frictional pressure losses, ECD, and annular hole cleaning. It is the basis for all
analyses of wellbore hydraulics [9, 10].
Flow properties for drilling mud is often characterized by the following rheology properties
[9]:
− Plastic viscosity (PV)

This part of the flow resistance is caused by mechanical friction between the particles
in the mud, between the particles and the liquid phase and the liquid elements
themselves. Plastic viscosity depends on the liquid viscosity, and the particle
concentration, size, and shape in the mud.
− Yield limit (YP)

Flow resistance transpires when attractive forces between the particles occur as a
result of electrostatic forces. Yield limit will vary with the shear rate and decrease
with increasing shear rate. The property is called shear thinning.
− Gel strength
Gel strength is related to the attractive forces between particles when the fluid is at
rest, and is measured as a function of time. It expresses the liquid thixotropic
properties, which means that shear stress is not fixed to a specific rate, but changes
with shear time.
− Apparent viscosity (AV) and Funnel viscosity (“Marsh Funnel”)

Apparent viscosity and funnel viscosity will give an estimate of the total fluid
viscosity. The combination of the three parameters mentioned above (PV, YP and gel
strength) will affect the total viscosity. This measurement is only used as a control
parameter for drilling fluids. To explain the cause of the change, the other rheological
parameters need to be determined.
The fluids can be divided into two groups according to their rheological properties; these are
Newtonian fluids and non-Newtonian fluids, respectively.
14
2.1.1 Newtonian fluids
Newtonian liquids have a viscosity, which is independent of shear rate. They are simple and
clean liquids containing no particles larger than molecules. For instance liquids such as water,
oil, and glycol behave as Newtonian fluids [9]. Given as Eq. 2 the shear stress is directly
proportional to shear rate:
𝜏 = 𝜇 ∙ 𝛾 (2)
where τ is shear stress, µ is viscosity and γ is shear rate.
2.1.2 Non-Newtonian fluids.

Unlike the Newtonian fluids, the viscosity for non-Newtonian fluids depends on shear rate.
These are divided into three main categories: Plastic liquids, pseudo plastic fluids and dilatant
fluids. It follows that the assortment of drilling fluids will be either plastic or pseudo plastic
fluids. In short, the main difference between plastic and pseudo plastic fluids are that plastic
fluids have yield strength and a pseudo plastic does not. Still, both are simultaneously shear
thinning, i.e. AV decreases with increasing shear rate. Two examples of plastic and pseudo
plastic fluids; water with added bentonite, and water containing polymers [9]. The following
rheology data set given in Table 1 is used as an example for how the different rheology
parameters may be determined by using both graphics and equations. The fluid is made out of
water, bentonite, polymer and barite [11]. The main goal will be to determine the rheological
model that is best fitted to describe the given Fann data in Table 1.
RPM Reading [°]

600 54.50
300 43.50
200 37.50
100 32.00
6 23.00
3 20.50
Table 1 - Fann data [11].
2.1.2.1 Bingham-Plastic model
The Bingham model best describes liquids that have a yield point, and includes suspension of
solids. The model is widely used to describe the condition of drilling fluid. Nevertheless, it is
not suitable for viscosity and pressure loss calculations. The model is based upon two
measurements that are performed by a Fann viscometer, respectively at 600 and 300 rpm. It is
from these measurements possible to calculate the different rheological properties. However,
it does not represent the most accurate behavior of drilling fluid at the bit (very high shear
rate) and in the annulus (very low shear rates).
To describe a fluid in the best possible way, good mathematical models needs to be
developed; perhaps one of the most famous of these is the Bingham-plastic model. It follows
from Figure 2 that the equation for shear stress (𝜏) is given by Eq. 3 [9]:
𝜏 = 𝜏! + 𝜇! ∙ 𝛾 (3)
where the yield point, 𝜏! (YP) and plastic viscosity, 𝜇! (PV) can either be read from a graph
similar to Figure 2 or calculated by using Eq. 4 and Eq. 5.
15
Figure 2 - Bingham-plastic model.
The slope of the curve in Figure 2 represents the plastic viscosity (𝜇! ).
𝜇! 𝑐𝑃 = 𝜃!"" − 𝜃!"" (4)
Curve intersection with the shear stress y-axis gives the yield strength in Eq. 5.
𝜏! 𝑙𝑏𝑠/100𝑓𝑡 ! = 𝜃!"" − 𝜇! = 2 ∙ 𝜃!"" − 𝜃!"" (5)
Using Eq. 4 and Eq. 5 and values from Table 1 the parameters 𝜇! (PV) and 𝜏! (YP) can be
determined.
𝜇! = 54.50 − 43.50 = 11 𝑐𝑃

𝜏! = 43.50 − 𝜇! = 32.50 𝑙𝑏𝑓/100 𝑓𝑡 !
2.1.2.3 Herschel-Bulkley model

The Herschel-Bulkley model is a modified version of the power-law model and is the model
that normally describes the measured data best. By defining a third parameter, yield stress
(𝜏! ), it is possible to get better results at low shear rates. The model is defined by Eq. 6 [9,
12]:
𝜏 = 𝜏! + 𝐾(𝛾)! (6)
or
log (𝜏 − 𝜏! ) = log (𝐾) + 𝑛 𝑙𝑜𝑔(𝛾) (7)
In comparison to Bingham, the model is using three parameters to describe the rheological
behavior; therefore an initial calculation of 𝜏! is required for calculation of the other
parameters (Eq. 8).
𝜏 ∗! − 𝜏!!" ∙ 𝜏!"# (8)

𝜏! =
2 ∙ 𝜏 ∗ − 𝜏!"# − 𝜏!"#
where 𝜏 ∗ is the shear stress value, corresponding to the geometric mean of the shear rate, 𝛾 ∗
and is calculated by interpolation.
16
𝛾∗ = 𝛾!"# ∙ 𝛾!"# (9)
Using Eq. 8 and Eq. 9 and values from Table 1. The parameters 𝜏 ∗ , 𝛾 ∗ and 𝜏! may be
determined.
𝛾 ∗ = 72.25 𝑠𝑒𝑐 !!
𝜏 ∗ = 28.26 𝑙𝑏𝑓/100 𝑓𝑡 !
𝜏! = 20.14 𝑙𝑏𝑓/100 𝑓𝑡 !
Figure 3 and Table 2 shows the final results. A trend line was obtained using regression
techniques.
100
Shear Stress [lb f/100 ft2]
10
τ -‐ τ0 = 0,9350γ 0,5331
R² = 0,99675
1
1 10 100 1000 10000
Shear rate [sec-‐1 ]
Figure 3 - Herschel-Bulkley fluid rheogram.
From
Figure 3, the Herschel-Bulkley parameters are as follows:
𝑛 = 0.5331
𝐾 = 0.9350 𝑙𝑏𝑓/100 𝑓𝑡 !
𝛾 [𝑠𝑒𝑐 !! ] 𝜏 [𝑙𝑏𝑓/100 𝑓𝑡 ! ]

1021,80 57,72
510,90 46,11
340,60 41,06
170,30 34,60
10,22 23,36
5,11 22,37
Table 2 - Shear stress calculated as function of Herschel-Bulkley parameters.
17
2.1.2.4 Robertson-Stiff model
Robertson-Stiff model was developed as a more general model to describe the rheology
behavior of drilling fluids and cement slurries. The model is given by Eq. 10 [13]:
𝜏 = 𝐴(𝛾 + 𝐶)! (10)

or
log 𝜏 = log 𝐴 + 𝐵 log (𝛾 + 𝐶) (11)
where A and B are model parameters similar to n and K in the Herschel-Bulkley model.
Parameter C is the shear rate correction factor, so that the term (𝛾 + 𝐶) is considered the
effective shear rate. Thus, 𝜏 is plotted against (𝛾 + 𝐶) on log-log coordinates, B is the slope
and A is the intercept where (𝛾 + 𝐶) = 1. Eq. 12 represents the yield stress for the Robertson-
Stiff model.
𝜏! = 𝐴𝐶 ! (12)
𝛾!"# ∙ 𝛾!"# − 𝛾 ∗! (13)

𝐶=
2 ∙ 𝛾 ∗ − 𝛾!"# − 𝛾!"#
where 𝛾 ∗ is the shear rate value corresponding to the geometric mean of the shear stress, 𝜏 ∗ ,
and is calculated by interpolation.
𝜏∗ = 𝜏!"# ∙ 𝜏!"# (14)
Again by using the data from Table 1, Eq. 13 and Eq. 14, the parameters 𝜏 ∗ , 𝛾 ∗ and 𝐶 may be
determined by calculations and interpolation.
𝜏 ∗ = 35.66 𝑙𝑏𝑓/100 𝑓𝑡 !

𝛾 ∗ = 195.65 𝑠𝑒𝑐 !!
𝐶 = 52.01 𝑠𝑒𝑐 !!
18
Figure 4 and Table 3 shows the results. A trend line was obtained by using regression
techniques.
Robertson-‐Stiff
100
10
τ = 6.379(γ+C) 0,3130
R² = 0,99872
1
1 10 100 1000 10000
Shear rate + C [sec-‐1 ]
Figure 4 - Robertson-Stiff fluid rheogram.
From Figure 4 the Robertson- Stiff parameters are as follows:
𝐴 = 6.379 𝑙𝑏𝑓. 𝑠𝑒𝑐 ! /100 𝑓𝑡 !

𝐵 = 0.3130
𝛾 [𝑠𝑒𝑐 !! ] 𝜏 [𝑙𝑏𝑓/100 𝑓𝑡 ! ]

1021,80 56,70
510,90 46,32
340,60 41,38
170,30 34,63
10,22 23,25
5,11 22,63
Table 3 - Shear stress calculated as a function of Robertson-Stiff parameters.
19
2.1.3 Comparison and error calculation of rheology prediction
2.1.3.1 Comparison of the rheology prediction

An Excel sheet has been created to compare and evaluate the different rheology models
against the experimental data set in Table 1. Figure 5 illustrates a comparison of shear stress
and shear rate data for the given rheology models.
Comparison of shear stress and shear rate data

70
60
50
Experimental data
40
Herschel-‐Bulkley
30
Robertson-‐Stiff
20
Bingham-‐Plastic
10
0
0 200 400 600 800 1000 1200
Shear rate [sec-‐1 ]
Figure 5 - Comparison of rheology prediction.
2.1.3.2 Error calculation of rheology prediction

By performing an error analysis, we can determine which model best represents the
experimental data set in Table 1. Table 4 shows that the Herschel-Bulkley model gives the
lowest error and has therefore been considered the best fit for the following simulations.
Rheology Model: Error %
Bingham-Plastic Model 1,382
Herschel-Bulkley Model 0.344
Robertson-Stiff Model 0,452
Table 4 - Error analysis for rheology prediction.
20
2.2 Hydraulics
2.2.1 Pressure losses

While drilling a well it is important to always be aware of the pressure losses within the
system, both at surface and downhole. The downhole static pressures can be calculated by
using the pumped mud weight, while additional pressure losses caused by circulation can be
calculated using the relationship between pump rates and drilling fluid rheological properties.
The downhole static pressure has however no influence on the pump pressure required to
circulate drilling fluid. The mud pumps are located on the same deck as where the mud is
circulated in return so that the drilling fluid is approximately in static equilibrium between the
pump outlet and return flow from the wellhead (fixed platform) or on top of the standpipe
manifold (semi-submersible rig). The frictional pressure drop and the nozzle pressure drop
makes up most of the pump pressure, and it is therefore extremely important to calculate these
before planning the drilling program. Yet, other pressure drops in the fluid flow should also
be estimated [14].
The total pressure drop provided by the mud pumps is determined by:
− Drill pipe frictional pressure drop (∆𝑃! )

o Liquid rheology properties.
o Lengths and inner diameters of the pipes and BHA components from the mud
pumps to the drill bit.
o Also note that some BHA components will have additional pressure losses due
to motor and MWD. BHA components can also have a smaller diameter than
regular drill pipe causing some additional pressure loss.
− Downhole mud motor pressure drop (∆𝑃!" )

o Some of the pressure energy in the drilling mud is often used to operate
downhole mud motors and sometimes turbines for rotating the drill bit or
acquire energy for downhole measurement systems (MWD). The mud pumps,
in form of higher pump pressure, must supply this energy.
− Nozzle pressure drop (∆𝑃!" )

o A large portion of the dynamic pressure energy is transferred to velocity
energy, which is used for flushing and partial breakage of the rock in the
borehole.
− Drill bit pressure drop (∆𝑃!" )

o Drill bit pressure drop is established when flow from the nozzles and the front
of the drill bit passes the edge. This is different dependent on the drill bit type.
For a roller cone there is ample space to flow around the cones, therefore this
pressure drop is often negligible and can be set as equal to zero. However for a
PDC drill bit, the bit body is seated firmly against the bottom of the borehole.
This allows for a thin layer flow of drilling fluid and the frictional pressure
drop can be significant.
21
− Annulus pressure drop (∆𝑃!" )
o The annulus can cause an additional pressure drop that represents lost energy,
which must be taken from the pump pressure. When cuttings are mixed
together with drilling mud, the average density is increased, and the static
pressure in the annulus between the drill string and the borehole wall increases
slightly due to this density increase.
o There will also be a large pressure loss around the BHA components due to the
variation in annulus geometry. BHA components differ in outer diameter and
may lead to a reduced flow area.
Figure 6 - Pressure drops during circulation.

The total pressure drop (∆𝑃!" ) illustrated in Figure 6 and shown in Eq. 15 is the sum of these
individual contributions, and is equal the pump pressure (𝑃! ), which must be supplied for by
the mud pumps.
𝑃! = ∆𝑃! + ∆𝑃!" + ∆𝑃!" + ∆𝑃!" + ∆𝑃!" = ∆𝑃!" (15)
Pump pressure is mainly determined by the frictional pressure losses. The mud density is a
part of the frictional pressure loss calculation models and is influencing the pump pressure
indirectly through the friction models. The hydrostatic component created by the cuttings will
cause a different hydrostatic pressure in drillpipe versus annulus, which will be reflected in
the pump pressure. Hence, the size of the hydrostatic pressure component is directly
dependent on the depth and must be considered. When calculating the frictional pressure
losses it might be efficient to subdivide the drillstring and annulus into shorter segments. This
is because any change in flow regime, wellbore geometry or fluid properties will affect the
frictional pressure loss.
22
The frictional pressure loss ∆𝑃! is calculated from Eq. 16:
2 (16)
∆𝑃! = 𝑓 𝜌𝑢 ! ∆𝐿
𝑑! !
where 𝑑! is hydraulic diameter, 𝑓! is friction factor, 𝜌 is fluid density, u is fluid velocity and
∆𝐿 is segment length [15]. It may be convenient to notice that that the friction will quadruple
if the rate doubles.
From the following procedure it is possible to calculate the frictional pressure drop [13]:
1. Determine rheological properties and choose the best-fit rheological model.
2. Based on the chosen rheology model, calculate the Reynolds number.
3. By comparing the calculated Reynolds number, determine the following flow regime.
4. Calculate the fanning friction factor.
5. Use the correct formula to determine the pressure loss.
2.2.2 Bit pressure losses

As drilling fluid flows though nozzles, the pressure loss is based on change in kinetic energy.
In oil field units, the pressure loss across the nozzle can be calculated from Eq. 19 [16]:
𝜌 𝑄 ! (17)
𝑃= !
2959.41 𝐶 ! 𝐴𝑛
where P [kPa] is pressure drop, 𝜌 [kg/l] is specific gravity of drilling fluid, Q [l/min] is flow
rate, An [ 𝑖𝑛! ] is total nozzle area, and C is orifice coefficient (C: 0.80 for non-jet bit and C:
0.95 for jet bit).
2.2.3 Cuttings concentration

During drilling, a real-time analysis of downhole and surface measurements can give
indications of poor hole cleaning. However, it is not always that intuitive to know how the
cuttings are settling throughout the entire borehole section, this is because the transportation
of cuttings and the formation of cuttings beds are largely influenced by a series of actions (i.e.
reciprocation and circulation rate) performed during this operation. Larsen et al [17] have
developed a model that is based on empirical correlations that enables a drilling engineer to
select the proper hydraulics to ensure problem-free drilling in high angle wellbores (from 55
to 90° from vertical). The model predicts the required critical transport fluid velocity (CTFV),
the average cuttings travel velocity (CTV) and the annular cuttings concentration under a
given set of drilling conditions. Under development of this model, 7000 tests were simulated
to show how CTFV and the Subcritical Fluid Flow (SCFF) would affect the annular cuttings
concentration. CTFV is defined as the minimum fluid velocity required to maintain a
continuously upward movement of the cuttings.
If cuttings start to accumulate in the wellbore, the annular fluid velocity is lower than the
CTFV. SCFF is defined as any flow rate corresponding to an annular velocity below the
CTFV. Following hydraulic drilling parameters were evaluated to investigate SCFF and
CTFV: flow rate, inclination, mud density, mud rheology, cuttings size, drillpipe eccentricity
and ROP. The predictions presented in Larsen et al can easily be read from charts, hand
calculated, or programmed on a computer [17, 18].
23
By implementing a transient cuttings-transport model, it is possible to get an updated
prognosis of the distribution of cuttings in suspension and in beds along the annulus, thus
giving us a more correct measurement for the cuttings concentration. Cayeux et al [18]
implemented a cuttings-transport model where the transport of cuttings is governed by the
mass-conservation equation. This equation can be written as Eq. 20.
𝜕𝜌 (18)
+ ∆ ∙ 𝜌𝑣 = 0
𝜕𝑡
A transient cuttings-transport model makes it possible to better predict downhole conditions

that evolve over time. Effects related to change in operational parameters are taken into
account to represent the simulation as realistic as possible. By real estimation of downhole
conditions, it is possible to provide better operational recommendations to avoid stuck pipe
and pack-off incidents. Thus, by adjusting the hydraulic drilling parameters, such as
drillstring rotational speed, flow rate and ROP one can avoid the formation of cuttings bed or
deliver a proper method of removing them. For more theory on transient flow models, see
Appendix A [18, 19].
Nevertheless, a transient flow model would have been too advanced to implement using only
Excel formulas. Hence, the steady state model shown in Eq. 21 was considered instead. The
equation is based on a no-slip model [20].
𝜋 ! (19)
(𝑅𝑂𝑃) ∙𝐷 ∙ (1 − 𝜙)
%C = 4 !!"#
𝜋 !
Q + (𝑅𝑂𝑃) ∙ 𝐷 ∙ (1 − 𝜙)
4 !!"#
where φ is reservoir porosity and Dhole is the hole size. Since the annulus volume and the flow
rate is known; the time it takes between cuttings is generated at bottom and when they are
observed at shaker can be determined. In order to get the transient behavior that reflects that
the response evolves over time, the following time interpolation has been implemented:
𝑡𝑖𝑚𝑒 − 𝑡𝑖𝑚𝑒𝑛𝑒𝑤𝑟𝑜𝑝 (20)

𝑥𝑖𝑛𝑡 =
𝑡𝑖𝑚𝑒𝑏𝑜𝑡𝑡𝑜𝑚𝑠𝑢𝑝
where xint is the interpolation coefficient, time is the cumulative time that runs through the
operation, timenewrop is the time when the new ROP is initiated, and timebottomsup is the
calculated time it takes to transport the cuttings from bottom to surface.
The following limits needs to be established for transient flow determination:
𝑥𝑖𝑛𝑡 = 0 if time = timenewrop

𝑥𝑖𝑛𝑡 = 1 if time = timenewrop + timebottomsup
Eq. 21 determines the transient cuttings concentration at the shaker during the bottoms-up
circulation.
%C!"#$%&'$! = 1 − 𝑥𝑖𝑛𝑡 ∙ %𝐶!"# + 𝑥𝑖𝑛𝑡 ∙ %𝐶!"# (21)
The transient model takes into consideration the increasing mud weight in the annulus as we
circulate. After the bottoms-up circulation the new mud weight is established in the annulus.
24
2.3 Modeling in Discovery Web
The following rheology models have been implemented into Discovery Web for creation of
hydraulic visualization scenarios.
2.3.1 Bingham frictional model

The majority of these formulas are taken from Data Drilling Handbook [16]. For more
formula details see Appendix B.
2.3.1.1 Frictional pressure loss calculation inside the drillstring

Determination of average velocity and critical velocity value for drillpipe:
Average velocity:
𝑄 (22)
𝑉!"# = 𝜋
(𝐷 ! )
4 !
where 𝑉!"# is average velocity, 𝐷! is string inside diameter and Q is fluid flow rate.
Critical velocity:
2.48 (23)
𝑉!" = 𝜇 + 𝜇! ! + 73.57 ∙ 𝜏! ∙ 𝐷! ! ∙ 𝑑
𝐷! 𝑑 !
Flow regime determination:

(With critical Re = 2100 for a Bingham fluid.)
If 𝑉!"# < 𝑉!" the flow is laminar.

If 𝑉!"# > 𝑉!" the flow is turbulent.
Frictional pressure loss calculation inside the drillstring:
• Laminar flow:
𝑑𝑝 𝑄 ∙ 𝜇! 𝜏! (24)
= ! +
𝑑𝐿 612.95𝐷! 13.26𝐷!
• Turbulent flow:
𝑑𝑝 𝑑 !.! 𝑄!.! 𝜇! !.! (25)

=
𝑑𝐿 901.63𝐷! !.!
25
2.3.1.3 Frictional pressure loss calculation inside the annulus
Determination of average velocity and critical velocity value for annular flow:
Average velocity:
𝑄 (26)
𝑉!"# = 𝜋
(𝐷 − 𝐷! ! )
!
4 !
Critical velocity:
3.04 (27)
𝑉!" = 𝜇 + 𝜇! ! + 40.05 ∙ 𝜏! ∙ (𝐷! − 𝐷! )! ∙ 𝑑
𝐷! − 𝐷! 𝑑 !
where 𝑉! is critical fluid velocity, 𝐷! is annulus outside diameter, 𝐷! is annulus inside

diameter (outside string), 𝜇! is plastic viscosity, 𝜏! is yield value, and d is fluid specific
gravity.

(With critical Re = 2100 for a Bingham fluid.)
If 𝑉!"# < 𝑉!" the flow is laminar.

If 𝑉!"# > 𝑉!" the flow is turbulent.
Frictional pressure loss calculation inside the annulus:
• Laminar flow:
𝑑𝑝 𝑄 ∙ 𝜇! 𝜏! (28)
= !
+
𝑑𝐿 408.63(𝐷! + 𝐷! )(𝐷! − 𝐷! ) 13.26(𝐷! − 𝐷! )
• Turbulent flow:
𝑑𝑝 𝑑 !.! 𝑄!.! 𝜇! !.! (29)

=
𝑑𝐿 706.96(𝐷! + 𝐷! )!.! (𝐷! − 𝐷! )!
26
2.3.2 Herschel-Bulkley frictional model
The majority of these formulas are taken from the doctoral thesis “Analysis of drilling fluid
rheology and tool joint effect to reduce errors in hydraulics calculations” [13]. For more

Determination of Reynolds number and critical Reynolds number value for drillpipe:
Reynolds number:
𝐷! !
!!! (30)
2 3𝑛 + 1 𝜌𝑣!
2
𝑁!" =
𝑛 𝐷! ! 3𝑛 + 1
!
𝜏! +𝐾
2𝑣! 𝑛𝐶!
Critical Reynolds numbers value:

! (31)
4(3𝑛 + 1) !!!
𝑁!! !" =
𝑛 𝑦
𝐼𝑓 𝑁!" < 𝑁!" !" the flow is laminar.

If 𝑁!" > 𝑁!" !" the flow is turbulent.
Frictional pressure loss calculation inside the drillstring:
• Laminar flow:
!
𝑑𝑝 4𝐾 𝜏! 3𝑛 + 1 8𝑄 (32)
= +
𝑑𝐿 14400𝐷! 𝐾 𝑛𝐶! 𝜋𝐷!!
• Turbulent flow:
𝑑𝑝 𝑓! 𝑄 ! 𝜌 (33)
=
𝑑𝐿 144𝜋 ! 𝐷!!
27
Determination of Reynolds number and critical Reynolds number value for annular flow:
Reynolds number:
!!! 𝐷! − 𝐷! ! (34)
4 2𝑛 + 1 𝜌𝑣!
𝑁!" = 2
𝑛 𝐷! ! 2 2𝑛 + 1
!
𝜏! +𝐾
2𝑣! 𝑛𝐶!
Critical Reynolds number value:

! (35)
8 2𝑛 + 1 !!!
𝑁!" !" =
𝑛𝑦
𝐼𝑓 𝑁𝑅𝑒 < 𝑁!" !" the flow is laminar.

If 𝑁!" > 𝑁!" !" the flow is turbulent.
• Laminar flow:
!
(36)
𝑑𝑝 4𝐾 𝜏! 16(2𝑛 + 1) 𝑄
= + ! !
𝑑𝐿 14400 𝐷! − 𝐷! 𝐾 𝑛𝐶! (𝐷! − 𝐷! ) 𝐷! 𝐷!
𝜋 2 − 2
• Annular flow:
𝑑𝑝 𝑓! 𝑄 ! 𝜌 (37)
=
𝑑𝐿 144𝜋 ! 𝐷! − 𝐷! 𝐷!! − 𝐷!! !
28
2.3.3 Robertson-Stiff frictional model
The majority of these formulas are taken from doctoral thesis “Analysis of drilling fluid
rheology and tool joint effect to reduce errors in hydraulics calculations” [13]. For more

Determination of Reynolds number and critical Reynolds number value for drillpipe:
Reynolds number:
! (38)
89100𝜌𝑣!!!! 0.416𝐷!
𝑁 !" =
𝐴 1
3+
𝐵
Critical Reynolds numbers value for laminar flow:
𝑁!" !" !"# = 3470 − 1370𝐵 (39)
Critical Reynolds numbers value for turbulent flow:
𝑁!" !" !"#$ = 4270 − 1370𝐵 (40)
If 𝑁!" < 𝑁!" !" !"# the flow is laminar.

If 𝑁!" > 𝑁!" !" !"#$ the flow is turbulent.
If 𝑁!" !" !"# < 𝑁!" < 𝑁!" !" !"#$ the flow is transient and a interpolation has been
introduced to ensure that all models are smooth and continuous. If not, this can cause
problems for well flow models.
Frictional pressure loss calculation inside the drillpipe:
• Laminar flow:
! (41)
𝐶
𝑑𝑝 1 + 3𝐵 0.2𝑣! + 𝐷
= 8.33 ∙ 10!! ×2!!! ×𝐴 6 !
𝑑𝑙 𝐵 !!!
!
𝐷!
• Turbulent flow:
𝑑𝑝 𝑓! 𝑣!! 𝜌 (42)
=
𝑑𝐿 25.81𝐷!
29
Determination of Reynolds number and critical Reynolds number value for annular flow:
Reynolds number:
! (43)
109000𝜌𝑣!!!! 0.0208 𝐷! − 𝐷!
𝑁 !" =
𝐴 1
2+
𝐵
Critical Reynolds numbers value for laminar flow:
𝑁!" !" !"# = 3470 − 1370𝐵 (44)
Critical Reynolds numbers value for turbulent flow:
𝑁!" !" !"#$ = 4270 − 1370𝐵 (45)
If 𝑁!" < 𝑁!" !" !"# the flow is laminar.

If 𝑁!" > 𝑁!" !" !"#$ the flow is turbulent.
If 𝑁!" !" !"# < 𝑁!" < 𝑁!" !" !"#$ the flow is transient and a interpolation has been
introduced to ensure that all models are smooth and continuous. If not, this can cause
problems for well flow models.
• Laminar flow:
! (46)
𝐶
𝑑𝑝 1 + 2𝐵 0.2𝑣! + 𝐷 − 𝐷!
= 8.33 ∙ 10!! ×4!!! ×𝐴 8 !
𝑑𝑙 𝐵 !!!
𝐷! − 𝐷! !
• Annular flow:
𝑑𝑝 𝑓! 𝑣!! 𝜌 (47)
=
𝑑𝐿 25.81 𝐷! − 𝐷!
All flow models require continuous models. It is therefore very important to include a
transient interpolation in all friction models to ensure that the flow models do not go unstable.
• Transient flow interpolation:
𝑁 !" − 𝑁!" !" !"# (48)

𝑥𝑖𝑛𝑡 =
𝑁!" !" !"#$ − 𝑁!" !" !"#
𝑑𝑝 𝑑𝑝 𝑑𝑝 (49)
= 1.0 − 𝑥𝑖𝑛𝑡 ∗ + 𝑥𝑖𝑛𝑡 ∙
𝑑𝐿 𝑑𝐿 !"# 𝑑𝐿 !"#$
30
3. Drilling Parameters, Problem Detection and
Remedial Action
Measurement and recording of hydraulic parameters and the quality of these are essential for
our interpretation and understanding of wellbore conditions. Measurement gauges are
constantly evolving, but the hydraulic parameters remain the same. This section discusses the
various parameters we measure during a drilling operation and what they represent.
Recording of wellbore parameters includes measuring, reading and storing of data. On the
Norwegian shelf it is a requirement that there should be two independent systems for
measuring parameters. In practice, this can be used somewhat differently on the different
installations. However, it should be interpreted and applied so that the entire supply chain
should have two independent systems. A similarity can be drawn towards primary and
secondary barrier when considering well barriers [21]. This means that there should be two
sets of sensors, transmitters and measurement gauges to prevent unexpected errors, thus
giving us an equivalent maximum coverage on a back-up system [22].
3.1 Drilling parameters

The parameters are recorded and used directly for carrying out the drilling operation. In order
to process and analyze the necessary borehole parameters, we start by measuring the
following quantities:
• Applied torque for make-/break-up of pipe connections [kNm]
• Tensile force - weight of drillstring/weight on drill bit [tons]
• Fluid balance for drilling fluid in/out of borehole [m! ]
• Pump pressure from the mud pumps [bar]
• Applied torque on drillstring [kNm]
• Pipe tally [m] [pieces (joints) run]
• Mud weight and ECD [sg]
• Rate of penetration [m/h]
• Height of DDM [m]
• Flow rate [lpm]
In addition to these we also measure the temperature and gas content in the return drilling
mud.
3.1.1 Applied torque for make-/break-up of pipe and pipe connection

The applied torque mentioned above, whether it is for rotation of the drill string or make-
/break-up of pipe connections it is obtained by measuring the force applied directly to the
drillstring. The way it is performed is that the torque on the tong is measured with a hydraulic
transducer in the tong line. We are gauging the tension in the tong line and not the torque. It is
therefore necessary to calibrate the measurement by multiplying the length of the tong arm
times the tension in the tong line. It is important to note that the tong line must be
perpendicular to the tong arm when the measurement is performed, a slight angle will reduce
the actual value. However, a few degrees will not make an appreciable difference [22]. A
correct measurement will lead to proper make-up torque; avoiding running loose (little torque
applied) and pin breakage (to much torque applied). When drilling, torque is measured
continuously, if applied torque is increasing this may be the result of pack-off/bridging.
31
3.1.2 Tensile force - weight on drill string/ weight on drill bit
To get a direct reading from the weight indicator on the weight of the drillstring, the tensile
force in the drilling line must be calibrated for the number of times it has been cut and then
again adjusted for any known weight of surface equipment such as e.g. travelling block and
DDM. Any deviation from this measurement is correlated for by use of the neutral weight and
provides again the weight on the drill bit as a secondary direct reading [22]. Note that when
we not are drilling, the tensile force at the top of the drillstring will be equal to the buoyed
weight of the entire drillstring. In deep wells, the tensile force may approach pipe tensile
strain. For normal drilling operations, the optimum conditions are met when the bottom part
of the heavy weight drillpipe (HWDP) is in compression and the drill bit carries the weight,
the top part of the HWDP and the entire drillstring above are in tension and carried by the
drawworks. A good weight indication sensor should give information on high over-pull when
POOH, which may be the result of pack-off [14].
Today, most of the measurements are performed downhole using measurement while drilling
(MWD). A sensor is mounted above the downhole motor (because the mud pulses used for
data transmission in the drillstring cannot go through the motor). Vibration and noise from the
drill bit makes it difficult to make precise measurements, this is due to the sensor and other
electronics are subjected to large vibrational loads. Having equipment that is reasonably
robust and reliable eliminates this. Another issue that has not been completed satisfactorily is
the transmission of signals to surface. The transfer rate is too slow to get up all the
information. Storing some information downhole and just sending up the information needed
reduces this problem. To handle this, the sensor needs to be equipped with a mini computer,
which is driven by mud flow through a small turbine [14]. However, new technology makes it
possible to bring more and more information back to surface [23]
3.1.3 Volume control for drilling mud in/out of borehole

By using volume control for all active mud pits in the surface system and by measuring the
return volume from the borehole, it is possible to record the fluid balance within the
circulation system. The way this is being practiced is that there are placed level sensors in all
active pits and in the mud return line. For good well control practices it is crucial to always be
aware of changes in total fluid volume. Direct measurement of return flow is the first
measurement that can provide indication of well control or circulation problems [22].
In addition to the pump rate, the volume control gives indications of a variety of conditions. A
sudden and large reduction in the return flow of drilling mud indicates that the formation may
be fractured, and that mud flows into the formation downhole (lost circulation).
Certainly, there will always be some net loss of drilling fluid because we are supposed to fill a
borehole that is getting longer while drilling, but this is compensated for by the cuttings mass
that is suspended in the return flow. If the amount of returned mud begins to increase
exponentially, this may suggests inflow of liquid or gas downhole, and a kick may be under
development [14].
32
3.1.4 Pump pressure
The drilling mud is pumped by one or several mud pumps, each of which can deliver about
2000 lpm at a pumping pressure of up to 386 bar (figures will obviously depend on the design
of the system). It is worth noticing that a higher pressure increases the load on the equipment
greatly, resulting in excessive wear and frequent interruptions. Pump pressure is measured
from a sensor that is placed on the high-pressure manifold which the mud line runs through on
the drill floor, and another one placed on the pressure side of the mud pumps. When
performing these measurements, the theoretical pressure calculations should be adjusted for
the placement of each sensor point. It is because we actually measure the pressure loss
(friction) within the circulation system. The mud goes from the mud pumps through a fixed
pipe (standpipe) on the drill floor, through a flexible high-pressure hose on top of the DDM
connected to a rotary coupling attached directly to the uppermost drillpipe [14, 22]. Between
the two above-mentioned measurement points there will be a noticeable difference because
of:
• The sensor located at the pressure manifold on the drilling deck does not measure the
pressure loss between the mud pump and the manifold; instead the sensor located on
the mud pumps measures this.
• Between the mud pump and manifold on the drill floor there is a significant height
difference. In the largest installations in the North Sea this height may be in excess of
30 meters and more. This means that the sensor measures hydrostatic pressure in
addition to the pressure loss in the system.
3.1.5 Flow rate

The flow rate has a close connection to the pump pressure. The flow rate is directly measured
on a mechanical gauge mounted on the outlet of the mud pumps. Before the signal is sent to
the computer screen, it is converted to the desired output (in our case lpm). To do this
conversion, we need an input for cylinder volume (stroke length and diameter). In real terms
the mud pumps will have a lower efficiency than the theoretical calculated value. Normally,
this efficiency will vary as a percentage between at 95-98% [22]. Flow rate is one of the
factors in addition to rheology that affects hole cleaning. If proper hole cleaning of the
wellbore is not achieved, several drilling problems may occur. Drilling problems such as lost
circulation, stuck pipe, pack-off and more will be discussed later.
3.1.6 DDM height and ROP

Sometimes it is necessary to know the length of the drillpipe above the drill floor that you
have to work with in case of an accident, or when it is time to do a new connection of
drillpipe. This height can be measured mechanically, optically or with a pressure sensor,
which may be directly readable. Together with the measured length of the drillstring there is
also made calculations to reveal the depth of the wellbore, and provide the distance from the
drill bit to bottom, during tripping in/out of hole [22]. When change in elevation per unit of
time is recorded, it is common to convert the value to drilling speed (ROP) in meters per hour.
Drilling speed is the rate that the drill bit moves through the formation, it is dependent on the
formation type, but tend to vary between 2 to 30 meters/hour.
Other factors affecting drilling speed are drill bit type, rotary speed, weight on bit (WOB),
well pressure (depth) and hydraulics (flow and pressure of drilling fluid). ROP management
ensures good hole cleaning. If there is an indication of lower cuttings concentration compared
to what we should anticipate, considered the chosen ROP, there might be development of
cutting beds, and jamming of the string might occur [14, 22].
33
3.1.7 Mud weight and equivalent circulating density (ECD)
Drilling fluid density is often referred to as mud weight. By regulating the mud weight, we
have the ability to manipulate and balance the borehole pressure, keeping us inside the
drilling window. While drilling a well, the well pressure should be kept within two limiting
values. The pressure must be greater than the formation pore pressure to prevent influx of
formation fluids (kick), and at the same time, be lower than the formation fracture pressure to
prevent fracturing formation (lost circulation). It is the interval between these two yielding
values that creates the basis for the drilling window (Figure 1). This drilling window differs
surely from well to well, dependent on depth and rock characteristics. Hematite and barite are
used to increase the hydrostatic gradient in the wellbore, thus by regulating the mud weight.
Eq. 52 gives the hydrostatic gradient.
𝑃!!"#$%&!"#$ 𝑏𝑎𝑟 = 0.0981 ∙ 𝐷𝑒𝑝𝑡ℎ !"# 𝑚 ∙ 𝑀𝑢𝑑𝑤𝑒𝑖𝑔ℎ𝑡 [𝑠𝑔] (50)
When working within the drilling window, it is not enough to simply keep track of the
hydrostatic pressure. When drilling, drilling mud is circulated through the drillpipe, and
cuttings are transported back to surface. This tells us that the mud weight can be divided into
three different categories. The static mud weight as given in the Eq. 52, a differential pressure
that is applied to the wellbore to get the fluid to flow, and cuttings that are suspended within
drilling mud on its way to surface. The sum of all these three has been given the definition
equivalent circulating density (ECD) and the expression is given in Eq. 1. To sum it up, the
mud weight is divided into static mud weight and the frictional pressure loss in the annulus.
The generated cuttings have an increasing effect on the ECD by increasing the static and the
frictional contribution. If the ECD is higher than expected, one might expect that cuttings are
packed around the bit. The annular space is blocked and the cuttings are creating additional
friction. This condition could occur due to poor hole cleaning.
3.2 Typical equipment and wellbore problems

The data collected and the equipment that has been used to complete the job can be seen as a
status report of the wellbore. Just as a motor or another form of electrical equipment is being
monitored, whether it is increased temperature in a bearing, decreased oil pressure in a
lubrication system, experience helps us predict how long the machinery can run without any
critical problems. Experience can help us interpret the measured parameters and give us
warnings on undesired events. For example, if the measurement of torque is flicking. These
operational parameters can be compared with the hydraulic parameters used for drilling
operations, and can show trends that can tell us what might go wrong.
When looking at different issues that arise in the wellbore and operational equipment, the
different parameters we measure and record changes through time. They can change gradually
over time or abruptly, if we for instance obtain a change in the drilled formation. This means
that the hydraulic parameters change if there e.g. is an alteration in force, fluid flow or
rotation. Trends in the measured parameters can be directly related to downhole problems. It
is true that MWD equipment in the drillstring may fail by accident or that an internal leak
may occur, but the majority of all downhole problems do not come without a warning.
Just look at the Macondo accident, here it was shown in hindsight that it was the combination
of lack of understanding the trends and the failed equipment that caused the accident [1].
Indication of upcoming problems can therefore be detected by studying error in trends or
change in the trends for one or more of the measured parameters. This section focuses upon
wellbore failures and the related equipment respectively.
34
Nevertheless, we must note the importance of conditioning and maintenance of surface
equipment, since successful drilling is dependent on continual preservation of the equipment
on the drilling unit. For example, the failure of surface equipment may lead to incidents such
as loss of pull force, drillstring rotation and pump capacity, which over a short period of time
can lead to severe wellbore problems. Incidents such as stuck pipe, lost circulation and well
control issues. These problems are usually the result of adverse trends over time [22].
3.2.1 Failure of surface equipment

In case of unexpected changes in the parameters that have no obvious explanation, it is
recommended to first check the surface equipment. This is in fact a golden rule before pulling
the drillstring out of the wellbore due to downhole problems. You should always eliminate the
possible sources of error in the surface equipment. Obviously, each situation must be
considered individually. One does not make an effort to check the surface equipment if there
is an influx of hydrocarbons (kick) or if there is only a few seconds differing from free to
stuck pipe [22]. Just to mention a few examples of common situations, and what to look for:
• Measurement of reduced surface volume: The reason for this can be as simple as
transferring drilling mud between pits, without informing the right personnel. It could
also be a sensor malfunction, or that there is a leakage out of the active system. In the
worst case it could be a kick or lost circulation.
• It is common to check general sensor and/or instrument measurement if there is an
unexpected change in parameter.
• If the pump pressure should decrease unexpectedly during drilling. This cause might
be as simple as a valve leak in the pump system. The liquid portion of the pump must
be checked (valves and pistons) or in worst case, it may be due to washouts and the
string needs to be POOH.
• Loss of pressure during a pressure test or a leak test against formation (FIT or LOT): It
is recommended to check all valves that are in contact with the cement pump system
first. In worst case, the formation is not strong enough to handle the maximum
acceptable differential pressure, and a leakage to the formation has occurred.
• A consequence of strong vibration/fluctuation of pump pressure is control of the
pulsation bafflers on the mud pumps.
It is worth noticing the importance of having two independent systems for controlling
hydraulic parameters (the rig's own system and mud loggers). If something happens
unexpectedly there should be an inspection of pumps, lines, valves, sensors and all data
should be checked against a secondary system [22].
35
3.2.2 Wellbore problems, detection and remedial action
Technical drilling issues can be related to everything from weakness in formation properties
to drillstring failure, or a combination of both. This category includes wrong choice of
equipment and equipment not suited for the specific job. For example, during operations in
HPHT wells, the downhole equipment must be qualified and tested against the needed
properties of such wells. High temperatures affect both mechanical and electronic equipment,
and may in worst-case cause equipment failure. However, by establishing good routines,
proper equipment may be selected. It is through experience we can study and better interpret
the hydraulic parameters to keep us away from unnecessary problems. Experienced data is
reflected in the following tables to come. These are examples of drilling problems that not
only explain what to look for, but also various causes and suggestions for how these can be
resolved. The following tables are based on [22] and are slightly modified.
The tool we have for determining wellbore problems is precisely the interpretation and
comprehension of drilling parameters referred to in 3.1.1. The events given below addresses
problems that can occur, some are simple, some are more complex and can have complex
causes. There are dozens of variables and conditions in a drilling process shown in trends, so
a 100% correct recipe on problem solving is almost impossible, but technological innovations
can help us to move a step closer.
Event Tearing-off drillstring.

Characteristics Immediate loss of pull weight, torque and pump pressure.
Cause Incorrect dimensioning of pipes and pipe connections.
Cracking due to wear and overload.
Incorrect make-up torque of pipe connections.
Washouts not detected in time.
Action POOH for possible fishing job.
Event Washouts in drillstring.

Characteristics Continuous loss of pump pressure. May have a sudden loss of pressure when a
leakage occurs.
Cause Cracking due to wear and overload.
Incorrect dimensioning of pipes and pipe connections.
Action Stop the mud pumps immediately.
Eliminate the possibility of leaks in surface equipment. If doubt, pull out of hole.
Pull out "wet" of the hole, observe continuously for leaks.
Check the entire drill string.
If there is a high resistance pulling the pipe, use no or minimal flow rate.
Replace the component above / below the leak point.
Event Wear of downhole motor and "stalling".

Characteristics "Stalling” i.e. the rotor stops.
Harder to get started with a control sequence.
Several pieces of elastomer from the stator in the return flow.
Cause The limit of the engine operating time is reached.
In advanced long-range wells, downhole motors are often pushed to the limit of
performance, so that the life expectancy might be somewhat reduced.
Action Assess the need for steering, or whether one can continue in rotation mode.
Assess risk of further rotary drilling with regard to being able to loose rotator in
the well.
POOH to replace downhole motor.
36
Event Stalling.
Characteristics Pump pressure increases rapidly.
Tool face turns over.
No ROP.
Cause Wear of motor.
Jammed stabilizer or drill bit.
Formation change, drill bit hits new formation too aggressively.
Action Quick response is required:
- Shut down the mud pumps.
- Pull off bottom.
- Start up pumps.
- Adjust tool face and resume steering.
Event Drill bit related problems.

Characteristics The new drill bit delivers poor or no ROP or too high torque has to be applied.
Abrupt increase in pump pressure.
Abrupt reduction/stop in ROP.
Gradual reduction of ROP, trouble steering.
Cause Wrong selection of drill bit.
Plugged nozzle(s).
Formation conditions (e.g. hard rock)
Drill bit: The roller cone bit may have lost a cone, or junk in the hole may have torn
off some of the cutters on the PDC drill bit.
Wear of drill bit.
Action Continue drilling ahead with same rate as long as the pressure is acceptable.
Consider reducing pump rate.
When drilling ahead, a sudden pressure drop may indicate that the plugs have been
washed away.
Consider the event, control eventual negative development in torque and pump
pressure. If it is not formation related, POOH with drillstring for replacement of bit.
Optimization of drilling parameters and steering intervals. May be combined wear

of stabilizers and drill bit. Consider the total life of the drill bit, requirements of
wellbore path, and the remaining length to TD.
Event Failure of MWD-signals.

Characteristics Poor/incomplete signals.
“Noise” in signals.
Total loss of signals.
Cause The pressure drop in the drillstring is insufficient (too low).
Failure of surface equipment.
Failure of MWD tools, surface or downhole.
Action Can control the pressure loss by calculating the total loss through MWD and the
drillstring under the current pumping rate and mud conditions. If the calculation
shows less than the requirement. Increase the pumping rate so that the required
pressure drop is achieved (if possible).
POOH with drill string to replace the drill bit nozzles or reset the sensors choke
setting.
Check pressure bafflers on the mud pumps.
Check for leaks or loose parts in the mud line valves.
MWD operator will check the receiver / computer. If failure downhole, POOH to
replace MWD.
37
3.3 Stuck pipe
Stuck pipe events are costing the industry hundreds of millions of dollars each year. Stuck
pipe incidents are unplanned events, which require the drilling contractors and operators to
make quick decisions in order to minimize or ease the sticking condition. There are multiple
conditions that can cause stuck pipe, one is not just stuck; there are a number of problems and
combinations of problems that lead up to this problem. These problems include mechanical
sticking, junk sticking, under gauge sticking, sloughing-hole sticking, lost-circulation
sticking, differential sticking and blow out sticking. Thus there are some similarities between
the events, each stuck pipe event has a distinctive set of conditions involving a combination of
the well geometry, geology, conditions, depth and sticking mechanisms [22, 24].
It is important to perform correct planning and selection of the appropriate equipment to

minimize stuck pipe incidents. When a stuck pipe situation occurs, one must act quickly and
correctly in order to get free. It is important to identify the situation and the mechanism
causing it. In almost all cases, there have been trends showing warning signs. The industry
should focus on stuck pipe events in planning, covering both proper reaction and cleanup
(fishing) [22].
The mechanisms behind stuck pipe incidents is divided into three main groups:
1) Pack-off/ Bridging.
− Pack-off: Cuttings or caved in solids that wraps around the drillstring.
− Bridging: Medium to large pieces of hard formation that jams the pipe.
2) Differential sticking.
3) Wellbore geometry.
Behind each of these main mechanisms there are several causes that lead to stuck pipe. The
tables below shows a systematic set of these causes. The tables takes basis in [22] and are
slightly modified. The illustrations are taken from [25]. Take notice of how the hydraulic
parameters can give indications of problems, and how they can be used to help, prevent and
cure a stuck pipe situation.
3.3.1 Pack-off / Bridging

Poor hole cleaning
− Cuttings do not stay in suspension, instead they drop to the bottom or

at the low side of the pipe and causes pack-off.
Avoid getting stuck: Free stuck pipe: Fig

• ROP management to ensure optimal hole cleaning. • Establish circulation and try to
• Maintain correct mud specifications and annular reciprocate.
velocities. • Use viscous pills dependent on hole
• Monitor cuttings volume changes in shakers. characteristics (depth, angle, size).
• Recognize increased over-pull. • Try to rotate string in order to get the
• If allowed, always reciprocate and rotate pipe while cuttings in suspension.
circulating.
• Use recommended viscous sweeps.
• Consider back reaming and use regular wiper trips.
38
Reactive formations
− Reactive shale is swelling up and creating clay balls and mud rings
around the drillstring.
Avoid getting stuck: Free stuck pipe:

• Plan regular wiper trips, be prepared to stop and • Establish circulation.
clean wellbore. • Concentrate on working drillstring
• Avoid long periods without circulation in problem downwards.
areas. • Gradually apply freeing force.
• Watch out for potential surge and swab pressures. • Once circulation is established,
• Prepare for back reaming while tripping. increasing MW might be beneficial.
• Recognize changes in mud properties.
Unconsolidated formation
− Poorly consolidated formations fall out of the borehole and results in

pack-off or bridging.

• Prepare to ream during a stand. • Establish circulation.
• Control ROP. • Disturb bridge by working string
• Use solids removal equipment. downwards.
• Prepare to use shaker screen blinding. • Increase force gradually, and ensure
• Clean out of hole before drilling ahead. that the cavings are removed before
• Avoid excessive periods of circulation in these further drilling.
formations.
• Wipe each connection and avoid excessive swab and
surge pressures.
• Prepare for loose fill when RIH.
Fractured formations
− Fractured formations fall into the wellbore and causes pack-off or

bridging.

• Clean-out excess fill before drilling ahead. • Reciprocate to disturb bridge.
• Hole conditioning and preparation for LCM. • If annulus is not packed off, apply
• Minimize surge pressure and restrict tripping speed. large forces in the beginning of
• RIH with JAR on BHA. procedures for getting free.
• Prepare to wash and ream when tripping in.
39
Cementing
− Cement blocks fall out the borehole and results in bridging.

• Minimize rat hole below casing shoe. • Work drillstring up and down to break
• When drilling out the rat hole and cement, always down the cement blocks.
ream section carefully before drilling further. • Acid may be used to break down the
• Always be aware when tripping through casing cement.
shoe and past cement plugs. • Gradually start to increase freeing
forces.
Cementing
− Wet cement results in pack-off

• Treat the mud if green cement is suspected. • Immediately apply maximum force
• Know theoretical TOC and start circulation above. while working string upwards and start
• Control the returns of cement at shakers. to jar.
• Do not rely on weight indicator • Attempt to start circulation
• Restrict ROP when cleaning out cement.
Junk
− Junk from BHA or surface results in bridging.

• Use maintained equipment in good condition. • Work and jar drill string up and down.
• Inspect equipment (MWD and BHA) regularly. • Gradually increase freeing forces.
• Be careful when working around rotary table.
• Cover the hole, use fishnets.
• Install wiper on drillpipe when possible.
40
3.3.2 Wellbore geometry
Key seating
− Key seating: Short section with high dogleg (high angle change) where
the drillstring is worn around the track in the borehole wall. The hole
diameter gets the profile as a keyhole, the diameter of the track is too
small for the BHA.

• Minimize rotation of pipe. • Reciprocate drill string while jarring or by
• Minimize dogleg severity. rotating.
• Consider design and configuration of BHA. • When free, gradually work pipe upwards.
• Minimize the rat hole length below casing. • Avoid jamming pipe in key seat.
• Consider string reamer or installing a wiper. • Rotate pipe with minimum tension through key
• Attempt to cure before further drilling. seat.
• May use back reaming.
• Minimize the risk of wall sticking.
Under gauge
− Reduction in hole diameter and ovality in the borehole cross-sectional

area. Abrasive formation wears down the diameter of the drill bit. Core
drilling can stipulate a hole diameter slightly less than the nominal
diameter. A BHA equipped with motor creates an oval hole diameter,
while a BHA used for rotation does not do this. In all cases the BHA
and drill bit may be stuck.

• Always use gauge bit and stabilizers. • Jar bit upwards using maximum force.
• Ream to bottom if a problem is suspected. • Work drillstring upwards.
• Never force bit to bottom.
• Proper bit selection.
• Be careful running PDC after tricone bits.
Mobile formation
− Overburden or tectonic stresses creates movement in plastic clay or

salt formations so that some parts of the hole are smaller in diameter
than the nominal diameter.

• Regular wiper trips. • Establish circulation; concentrate working string
• Condition mud prior to penetrating salt. downwards, with gradually increasing force.
• Consider eccentric PDC bits. • If squeezing salt, pump freshwater pill while
• Increase mud weight. working pipe, do not delay pumping water pill.
• Minimize open hole time. • OBM systems: Use water input spacer ahead of
• Importance of minimizing reaction time. pill.
• Use maximum pull on pipe while circulating
pill. Increase mud weight afterwards.
41
3.3.3 Differential sticking
The combination of a static drillstring in contact with a permeable formation and the
development of filter cake provide a pressure differential that holds the string firmly against
the borehole wall. High overbalance enhances this effect. In practice differential sticking
usually occurs when the pipe is stationary during a connection or when taking survey. First
indication of sticking is when there is full circulation through the pipe and no up/down
mobility or rotary abilities, other than pipe torque and stretch. When solids in the mud no
longer are suspended, they settle out, increasing the sticking force. The sticking force may be
calculated by using the definition of differential pressure and the drill collar contact area [22].
Figure 7 - Illustration of drill collar without centralizer sticking in a well [26].
Eq. 51 calculates the sticking force:
𝐹 = 𝜇 ∙ ∆𝑃 ∙ 𝐴 (51)
where ∆𝑃 is the difference between the outside mud pressure and the pore pressure inside the
rock. By the absence of centralizers in the calculation, assuming it is only drill collars that is
in contact with the mud cake, the contact length can be determined and the sticking force
calculated by Eq. 52 [26]:
(𝑅! − 𝑡)! − 𝑅!! − (𝑅! − 𝑅! )! (52)

𝐹 = 𝜇 ∙ ∆𝑃 ∙ 2𝑅! cos !! ∙ 𝐿!"
2𝑅! (𝑅! − 𝑅! )
where 𝑅! is well radius, 𝑅! is b shown in Figure 7, t is the mud cake thickness and 𝐿!" is
length of drill collar. The formula for a deviated well is shown in Eq. 53 [27]. The force
needed to pull the drillstring is expressed as the sum of the pipe weight, the drag force and the
differential sticking force.
𝐹 = 𝛽𝑤ℎ 𝑐𝑜𝑠𝛼 + 𝜇𝑠𝑖𝑛𝛼 + 𝜇 ∙ ∆𝑃 ∙ 𝐴 (53)
42
Figure 8 - Well path for stuck point derivations.
Sometimes it is not possible to work the pipe free. In the worst-case scenario it is therefore
necessary to determine the stuck point for eventual cutting of pipe or drill collar. Models can
determine the point of stuck. For stuck point determination, a tensile test needs to be
performed to measure an additional force F and the corresponding elongation, ΔL. Aadnøy et
al [27] derived a model for a well geometry shown in Figure 8. For a complex geometry one
can develop a model based on torque and drag and for any loadings.
Aadnøy et al [27] derived a model based on the following assumptions:

− Negligible BHA (much stiffer than drillpipe).
− Only one size of drillpipe.
− No friction applied (force is acting directly on the stuck point).
If drag is included, some of the pulling force will transform to friction. However when
pulling, the pull-rate is usually slow and nearly static, so we can easily assume that the drag
forces in a vertical well are neglected. A golden rule is that pull force is assumed unaffected
of friction for vertical wells. It is therefore reasonable to say that for straight section in
deviated wells, friction is assumed neglected or dealt with by measuring the drag forces
before the drillstring got stuck. Conversely, for curved section the friction is taken into
account. Similarly to Figure 8 assume now that our well consists only of a vertical section,
build section and sail section down to TD. The build section with a radius, R builds up to an
angle, 𝛼. Then the sail section is drilled holding this angle down to the stuck point. Assume
now that drag is neglected, and the pipe is pulled slowly with a force dF and a length dl.
43
The depth estimation down to the stuck point is given by Eq. 54 [27]:
𝑑𝑙 1 (54)
𝑙 = 𝐴𝐸𝑒 !" − (𝑒 !" − 1) 𝑙! + 𝑅𝛼
𝑑𝐹 2
Eq. 54 is only valid for a constant drillpipe size, and for a well consisting of the three
following sections; vertical, build-up and sail. Surely it is possible to assume two drillpipe
sizes, referring to indexes, top pipe size is index 1 and bottom pipe size is index 2. Eq. 55
applies for the depth to stuck point if two different drillpipe sizes are used [27]. For example
using drillpipe in combination with HWDP.
𝑑𝑙 𝐴! !" 1 𝐴! (55)
𝑙 = 𝐴! 𝐸𝑒 !" − 𝑒 −1 𝑙! + 𝑅𝛼 − 𝑙! −1
𝑑𝐹 𝐴! 2 𝐴!
When the situation is at point, quick and proper evaluation of hydraulic drilling parameters
and reaction may be essential to prevent an eventual catastrophe. Nevertheless, either if you
are lucky enough to get the pipe free or a preferable sidetrack is initiated, there are a few good
practices to follow in order to bring the situation back to normal drilling. They are as follow;
proper hole cleaning and removal of cuttings, activation of the jar, POOH and replace
components of the BHA that may have been damaged due to large impact forces.
44
3.4 Lost circulation
Lost circulation is another problem that also lead to increased NPT. An illustration of a lost
circulation scenario is shown in Figure 9. During a drilling operation there will be a
continuous change in the mud system. This is because the system is continuously given
additives such as basic fluids, solids, and chemicals to compensate for the gradually
increasing hole volume. A certain amount of liquid may be bound to the cuttings or lost to
permeable zones within the wellbore. This is a mass balance equation that we easily can keep
track of. Lost circulation is therefore a measurable loss of the total drilling fluid to the
formations.
Figure 9 - Illustration showing total loss of circulation.
45
Table 5 shows that the severity of the downhole problem can be divided and graded according
to the amount of lost mud [22]:
Grade: Type: Description:

1 Seepage loss Continuous loss of up to 3𝑚! per hour under normal drilling
conditions. This is an indication that we are on the threshold of
what the exposed formations can withstand.
2 Partial loss Continuous loss of more than 3𝑚! per hour under normal drilling
conditions, but we do still have returns from the wellbore
annulus. This indicates that there are conditions in the wellbore
which means that we are in excess of the tolerance limit of the
formations.
3 Total loss The tolerance limit of the formation is clearly exceeded. We have
no return flow from the annulus. Lost circulation may also cause
the mud level in the annulus to decrease (tens of meters is not
uncommon).
Table 5 - Severity of lost circulation [22].
As seen in Table 5 there is a development scale, and it is then often the case that the first two
conditions (1st and 2nd) may be a warning of the 3rd (total loss circulation). The problems
that can result in lost circulation can be divided into two main categories; Natural and
Induced. Induced lost circulation occurs when the pressure exerted to the wellbore exceeds
the maximum pressure the wellbore can resist (fracture pressure). Natural lost circulation
occurs when the drill bit is penetrating formations with large pores, high permeability, leaky
faults and natural fractures. Also here the formation is exposed to a fluid pressure in the
wellbore, which exceeds the tolerance limit of loss [22, 28].
Pressure induced fractures may be the result of different causes. For example, the mud weight
may simply be too high for the exposed formations, or alternatively the ECD too high. As
mentioned earlier, ECD is a sum of the mud weight plus the friction in the annulus and the
cuttings suspended in the mud. It is important that we understand that the greatest ECD is at
the drill bit. It is the friction in the annulus that makes the contribution to the ECD
measurement. Most of the total friction force in the drilling mud within the annulus is
between the BHA and the wellbore. Here, the distance between the annulus and the pipe is at
its smallest, ergo higher fluid velocity and higher friction. For a particular point in the well,
the fluid pressure in the borehole reaches its maximum as the drill bit passes, it will diminish
as the components with large diameter is passing [22].
Pressure waves transferred within the drilling fluid may also lead to total or partial lost
circulation. For example, the pressure in the wellbore that builds up during pack-off of
cuttings or cavings from the borehole wall may exceed the fracture pressure in the underlying
formation causing lost circulation. Similarly, a shut in pressure could fracture weaker zones.
Drilling through formations with low formation pressure may cause circulation losses. As we
drill, the formation pressure and the fracture pressure will follow each other, so that when the
drill bit intrudes a low-pressure zone, it will have a lower fracture gradient than the overlying
zone. If we have been drilling with a drilling fluid that provides a borehole pressure greater
than the fracture pressure of the low-pressure zone, the formation would fracture [22].
46
Similarly, naturally fractured and high permeability formations also have the same basic
mechanism, that the wellbore pressure is higher than the strength of the weaker zone. Natural
fracturing in a formation will have an additional weakness because the initial fracture is
already there; the question is whether or not the pressure will exceed the fracture opening
pressure. Likewise a fault that has not been sealed will also act as a trap for circulation losses.
High permeable zones are often poorly consolidated, for instance sand zones where the
cementations between the sand grains are weak. Drilling with high mud weigh to control
unstable clay further up will often lead to drilling into the sand zone with too high mud
weight. Some formations may have cavities. Drilling through these cavities with an excess
pressure of the mud relative to the formation fluid pressure can result in total or partial lost
circulation. Prevention of circulation loss should be included in the planning processes and in
the operational procedures that focuses on minimizing the risk for failures. By giving the
proper concern, one also focuses on minimizing the risk of stuck pipe, because lost circulation
often have a high risk of escalating into a stuck pipe situation [22].
For an optimized and efficient drilling operation, it is important to consider the following
issues during the design phases: [22]:
− Design the casing so that it isolates the low-pressure zones and estimated loss zones as
soon as possible after penetration.
− Good hole cleaning is essential. The more cuttings, the higher ECD.
− Plan for a minimum mud weight to control the known formation pressure. The weight
must be balanced against the desire for high mud weight to keep the other formations
from caving in.
− Drilling mud must have a certain rheology to purge the borehole properly.
− Excessively viscous mud increases friction and also provides higher ECD.
− Pre-treat the drilling mud with loss circulation material (LCM) before penetrating the
known loss zone. Additives may be helpful to keep a zone with little or partial loss
under control.
− Avoid pressure waves, plan for quiet tripping into / out of the borehole, "soft" start and
stop all movements before making connections. Break up the mud gel strength when
circulation starts by initiating rotation first. Start the mud pumps gently and increase
incrementally with full return before the next increase.
− If possible, eliminate annulus restrictions; unstable formations should be back-reamed
at fixed intervals.
− Control the ROP to keep down the amount of cuttings in the annulus.
− Assemble a BHA that is able to manage LCM in the drilling mud, especially in terms
of MWD/LWD equipment.
− Have clear procedures for how the rig crew deals with LCM on the surface. It can be
associated a high risk, having to stop circulation completely for a period time, due to
plugging of surface equipment.
47
4. Examples of other Software Tools
Calibrated models for optimizing oil and gas production has been around for some time now,
and usage of process models to control drilling processes is getting more and more common.
A few examples of integrated drilling control systems are presented in this chapter [29].
4.1 Sekal DrillScene

The advanced monitoring system DrillScene is based on 20 years of modeling research and
development in Stavanger at the International Research Institute of Stavanger (IRIS). The
software system is based upon a continuous comparison of transient flow model predictions
against real-time measurements at the rig site. The model can be used for giving warnings
about possible unwanted events that is developing in the well and may in that context be a
proactive tool [29].
4.1.1.1 Drilling Simulation Environment for Testing Drilling Automation Techniques

The purpose of the DrillScene wellbore simulator is to generate response from a “real” well,
submitted to the drilling actions simulated in a virtual rig. Simulation of realistic well incident
responses is therefore integrated into the wellbore simulator. Such incidents include for
example influx, losses, pack-off, formation collapse and various stuck pipe situations. The
advantage of such a simulator is the possibility to play through the simulation in fast-forward
mode, pause or use multiverse-like capabilities. Figure 10 illustrates drilling trough an
unexpected depleted region causing fracture and loss [30, 31].
Figure 10 - DrillScene: Drilling through a depleted region. Time frame is 2 minutes [31].
One of the drilling methods used for simulations are MPD. The MPD solution is based on
back-pressure, which means that a back-pressure pump is used at surface to control the well
pressure. Figure 11 shows how the annulus needs to be sealed to be able to apply pressure at
the surface. This makes it possible to continuously control the downhole pressure by changing
the surface back-pressure. The back-pressure in the system is generated by a choke that
48
creates a pressure drop when drilling fluids are circulated through it. However, if there is zero
or too low circulation through the well, an additional pump is used to create the required
pressure drop by circulating the drilling fluid directly through the choke [30].
Figure 11 - DrillScene: MPD solution based on back-pressure [30].
4.2 eDrilling Solutions

eDrilling has a company mission that contributes to increased value creation, safety and
profitability for its customers through solutions for planning, training, optimization and
control of drilling operations. eDrilling has been around for a few years now and is an
innovative system for real-time drilling simulation, 3D visualization and control from a
remote drilling expert center. The concept processes all available real-time drilling data (both
surface and downhole) in combination with real-time modeling to monitor and optimize the
drilling processes most effectively. By implementing this information into the model it is
possible to visualize the wellbore in 3D real-time. eDrilling has been implemented in an
Onshore Drilling Center on the Ekofisk field in Norway. The system has for instance given
very early warnings on ECD and friction related problems [32].
4.2.1.1 Supervision of ECD with automatic diagnosis embedded

During drilling, the eDrilling system will continuously be monitoring the ECD and compare it
with calculations executed by an advanced dynamic pressure and temperature model. The
model will continuously be calibrated where it is considered sufficiently reliable for that
purpose. Figure 12 and Figure 13 shows an illustration of the flow model calculations using
input from real drilling data. The simulations are based on washing/reaming (12-14 hours),
drilling 18 meters (14-16.8 hours), back reaming, and tripping out (last half hour). At this
stage, data was recorded while replaying data after operations had finished. If accurate pore
and fracture pressures were available, the calculated ECD, both at the bottom and at other
positions along the OH section, would have been compared continuously to the given pressure
profiles. If getting close to or exceeding boundaries a visual notification would have been
triggered in the 3D view [32].
49
Figure 12 - eDrilling: Pump rate [32].
Figure 13 - eDrilling: Calculated bottomhole ECD [32].
50
4.3 DrillBench
DrillBench is a software tool for design and evaluation of all drilling operations. It is designed
significantly for wells with narrow margins (i.e. narrow drilling window). DrillBench
considers both transient and steady state conditions, and enables engineers to model wells
with diverse and extreme complexities that has been verified extensively against actual data.
The model can be used to model all types of wells and scenarios (HPHT, deep-water, ERD
and MPD). The interface is user friendly and designed for both drilling engineers and drillers.
Schlumberger is now the current owner of DrillBench [33].
4.3.1.1 Dynamic Modeling in drilling of the Gullfaks C-5A Well

During drilling of new wells in the Gullfaks field, Statoil in the early 2000 observed an
increased trend in amount of drilling problems in the Shetland formation. Underbalanced
drilling was initially considered as a remedial action to this problem (drilling with a
hydrostatic pressure that is lower than the formation pressure). In general, this helps to
increase productivity by reducing formation damage in the reservoir section and solves
problems such as low ROP, kick/losses due to pressure depletion. At Gullfaks there was a
problem related to small margins and it was difficult to drill conventionally. In the planning
phase, the operational window was defined by means of steady state modeling. To ensure that
the bottomhole pressure could be controlled, various combinations of influx rates and choke
pressures were simulated. It was also important to ensure proper hole cleaning during the
operations. By modeling the 7” liner, they were able to quantify the surge/swab pressures that
again created the basis for the 7” liner running procedure. In order to evaluate the dynamics of
the system, a number of scenarios were developed by introducing transient simulations (for
instance if drilling through a high productivity fracture). This would assist the rig team in
detecting drilling problems and making correct choke adjustment to maintain bottomhole
pressure [34].
Figure 14 - DrillBench: Kick during connection [34].
51
Gullfaks is a mature field entering the final production stage. The Shetland formation, which
makes up the cap rock of the Gullfaks field, has by high-pressure water injection been
pressured up to reservoir pressure. This has re-activated old faults along the annulus of poorly
cemented water injector wells. By increasing the pressure, which is below the cap rock, the
Shetland formation has become extremely difficult to drill using normal overbalanced drilling
techniques. It was when performing a re-entry job in well C-5A in 2002 that the well took a
high-pressure kick, resulting in production shutdown.
It was assumed that the high-pressure water injection was the cause. From the kick data the
pore pressure and fracture gradient was calculated to be 1.84 sg and 1.90 sg, respectively.
This makes it an almost impossible drilling window to drill in a conventional manner since
the difference between the ECD and the static mud weight can be larger than the margin, and
it will be impossible to avoid kick or losses. An underbalanced method was therefore
considered. Figure 14 shows that a kick will be taken if the well is drilled in overbalanced
condition when performing a connection, which again can lead to fracturing of the formation.
Figure 14 is not considering the pressure build-up during conventional well control. Dynamic
modeling was successfully used for the planning and preparation phase for Gullfaks C-5A,
which was the first underbalanced well drilled offshore in Norway by Statoil. It was by the
use of dynamic modeling that it was possible to identify potential transient events and create
remedial actions. The transient modeling results were also used as support material in training
of rig crews, increasing the general understanding of flow dynamics during both overbalanced
and underbalanced drilling [34].
52
5. Architecture of Discovery Web
SiteCom is Kongsberg Oil & Gas Technologies solution to enable real-time monitoring and
analysis of all data from well operations. A typical system setup topology is presented in
Figure 15 below. The SiteCom Suite delivers in addition to real-time monitoring, historical
data storage for post-well analyses and training. At the well site it also provides an aggregator
for data gathering from all the different service companies [35].
Figure 15 - SiteCom setup [35].

By using an aggregated system for collecting data from all service companies, operators are
able to gather real-time data from multiple data sources. Data is taken from the rig systems in
many different formats (OPC, WITS0, NMEA, WITSML) and then aggregated into the
onshore drilling databases. By using a single link to shore (that replaces systems from mud
logger, LWD & drilling systems) it is possible to use a standard application for analyzing and
processing of data. This enables secure and flexible data, and standardized processes and
workflows. Nevertheless, this helps us to create corporate standards independent of contractor
that gives the client end-user quality data using only one interface (Discovery Web). This is
because all data is transmitted, hosted and managed completely within the client domain, and
the data management is independent of rig site service provider. Figure 16 illustrates how
SiteCom aggregates all data sources at all rigs and transfers the data back to the central
database (<WITSML/>), which the client end-user can connect to from his interface at home
or at the office via a web browser and the Discovery Web application [35].
53
Figure 16 - Workflows of SiteCom [35].
The client-end user uses Discovery Web that has a standard flexible interface hosted in
Internet Explorer that allows consoles to be easily configured to monitor all key drilling and
evaluation (D&E) processes to maximize efficiency and minimize NPT. To help improve the
D&E processes, Discovery Web is implemented with the following instruments [35]:
− Real-time drilling dashboard.
− Access to all rigs and wells with just a single click.
− Pumps ‘n’ Pits (Information about pits and pumps, flow in/out, PVT gain/loss and
basic kick detection).
− Drilling calculations (Drilling parameter cross-plots and drilling diagnostics).
− Very rich feature set (i.e. Chat, Print, Export, Widgets, Formulas, Maps).
A basic drilling console illustration from Discovery Web is shown in Figure 17. As
mentioned the application can be accessed and viewed from anywhere. Custom templates can
be easily designed and distributed with the following benefits [36]:
− Enable a full overview over well site status.
− Improve client access to real-time data.
− Improve collaboration between teams and across disciplines.
− Easily distribute information.
− Reduce overall costs.
− Reduce the need for training.
− Helps to overcome contractual and geographical hurdles.
The combination of time-indexed, depth-indexed, real-time and historical data makes it a

powerful tool in data gathering and processing. Data is displayed in widgets and can be
viewed as bar graphs, log graphs, cross plots, image plots, lithology plots and circular gauges.
Sources for data is everything from MWD, LWD, mud, cement, mud logging to weather and
positioning [36].
54
Figure 17 - Basic drilling console - DrillingTime Playback UiS [37].
In addition to real-time monitoring, the feature Discovery Web Formulas gives you the ability
to create formulas (Figure 19). This makes it possible to include models that can be compared
with real data. A formula consists of static parameters (static values), inputs (mnemonics) and
outputs (formulas). Outputs from a formula can be visualized and used in the most common
widgets (visualization elements) similar to log data, including log widgets, history table
widgets and most single-value widgets. This is why outputs can utilize inputs and static
parameters within the equation using predefined mathematical functions. Syntax for output
formulas is taken from Microsoft Excel, where the supported Microsoft Excel functions are
provided for. Figure 18 shows the calculation model used for Discovery Web Formulas and
the Arithmetic Smart Agent. A graphical representation of an output equation is also provided
for in Figure 19 and displays a rendered graphical display of the completed equation in a
standard mathematical notation format [36].
Figure 18 - Calculation model for Discovery Web Formula and Arithmetic Smart Agent [37].
55
Figure 19 - Discovery Web Formula [37].
56
6. Discovery Web Real-time ECD Control Design
6.1 Simulation based on our built data
Figure 20 shows the standard base-case that has been considered in the Discovery Web
simulations. One is assuming drilling an 8 ½” hole from a platform rig through a reservoir
section at 4000 meter with a 9 5/8” casing set at 3500 meters. The drillpipe is considered
having an outer and inner diameter of 5” and 4.27” respectively. BHA outer and inner
diameter is 6.5” and 2.5”. At 4000 meters, the length of drillpipe is 3800 meters and length of
BHA is 200m. By assuming an average ROP of 40 m/h over a time interval of 4 hours, the
following simulations are performed during drilling of 160 meters reservoir section.
Following assumptions are taken into considerations:

- Vertical well.
- Negligible pressure and temperature effects on the drilling fluid rheology and mud density.
- ROP value.
- Constant mud weight.
- Cutting particles are suspended in the drilling mud when there is no circulation.
- Cuttings are transported with the fluid velocity.
- Constant motor, MWD and surface equipment pressure effects.
- Implementation of Herschel-Bulkley rheology model.
- Appendix B: Rheology models.
- Appendix C: Rheology models and hydraulic calculations implemented in Discovery Web.
- Further assumptions will be explained through the different scenarios.
Figure 20 - Well schematics based on our built data.
57
6.1.1 Connection scenario
After drilling one stand (30 meter), the mud pumps have to be disconnected and a new stand
has to be connected to the drillstring. Since there is no circulation, all friction in the system
will disappear. This will be noticed as a drop in pump pressure from maybe 250-300 bars to 0.
The drop in bottomhole pressure in the annulus will be approximately 6 bars for a 12 ¼” hole
and 20-25 bars for an 8 ½” hole. The connection will last 5-10 minutes and we do not expect
large temperature changes in the well. For the connection simulation scenario following
parameters were considered. These are: mud weight in is 1.60 sg, flow rate in is 2000 lpm and
average ROP is 40 m/h. Figure 21 shows that the pump pressure (SPP) varies from 293 to 0
bar during a connection. Bottomhole pressure (BHP) increases linearly due to the increased
hydrostatic column and drops 40 bars during a connection. ECD varies from 1.60 to 1.695 sg.
Figure 21 - Connection scenario.
58
6.1.2 Kick scenario during drilling
When drilling through a reservoir section, proper mud weight selection is essential to avoid a
kick. Good simulation tools and real-time data can help in early detection of downhole
problems. For the kick simulation scenario the following parameters were considered. The
mud weight and flow rate is 1.6 sg and 2000 lpm, respectively. The simulation is shown in
Figure 22. The kick occurred during drilling with insufficient mud weight to balance
formation fluid. In time period 10.30-11.00, the ECD drops below the pore pressure curve,
and a kick influx is taken into the well. We observe that the ECD becomes lower than the
pore pressure.
Kick
influx
Figure 22 - Kick scenario during drilling.
59
6.1.3 Kick scenario during connection
In this scenario assume that an engineer has designed the ECD just a little bit higher than the
formation pressure without the knowledge of safety margin considerations. The engineer did
not consider that during a connection the dynamic part of the ECD would disappear (friction).
This scenario is designed to illustrate the possible incident in a reservoir. For this kick during
connection scenario, the following parameters were considered. ROP is 40 m/h and flow rate
is 2000lpm. At around 4128 ft a connection was made and the mud weight was reduced below
the reservoir pressure. Figure 23 shows that the kick influx will only be taken within the
connection period. To prevent that the connection pressure drops below the pore pressure
gradient, a safety margin between the static mud weight and pore pressure should have been
considered. If an additional safety margin of 0.02 sg had been taken into account, the
connection pressure drop would not have resulted in a kick.
Kick
influx
Figure 23 - Kick scenario during connection.

60
6.1.4 Pack-off scenario and sensitivity of pack-off
This scenario assumes that during drilling of unstable formations such as unconsolidated
sandstone, brittle shale may fall into the wellbore and cause pack-off or bridging. It is
assumed that the hole packs-off gradually behind the drill bit while drilling from 4000 m to
4160 m. Figure 24 shows an ideal pack-off illustrating the reduction of annular flow capacity
by 10-20%. This section simulates the sensitivity of pack-off by gradually reducing the
annular capacity. The annulus section above the pack-off point at 4000 m is considered
unchanged (standard 8.5” hole). The following simulations have been executed for a 4 hours
interval to illustrate the dramatic pressure build-up a pack-off can cause. During an actual
drilling operation, the drilling operation would naturally stop earlier.
Figure 24 - Illustration of pack-off scenario.

Model input parameters used for the pack-off scenario:
- Mud weight in 1.60 sg
- Flow rate in 2000 lpm
- Average ROP 40 m/h
- Simulation time 4 hours
- BHA OD 6.5”
61
6.1.4.1 10% Pack-off effect
The simulation effect of 10% pack-off is presented in Figure 25. In Track 4 the outer diameter
of the annulus is reduced from 8.5" to 7.65". This results in an increased ECD that approaches
the fracture gradient. Accordingly, the pack-off effect also increases the pump pressure (SPP)
and the bottomhole pressure (BHP), but the impacts does not show that clearly for a 10%
pack-off.
Figure 25 - Pack-off scenario 1 (10% reduction in annulus diameter).
62
The simulation effect of 16% pack-off is presented in Figure 26. In Track 4 the outer diameter
of the annulus is reduced from 8.5" to 7.14". The 16% pack-off effect results in a sudden
increase in SPP, BHP and ECD. Consequently, the formation fractures instantaneously when
the hole packs off. The increase in SPP for the constant flow rate is symptom for the
deterioration downhole due to pack off.
Fracture
Figure 26 - Pack-off scenario 2 (16% reduction in annulus diameter)
63
This simulation assumes that the annular constriction is reduced by 20% due to pack-off. In
Track 4 the outer diameter of the annulus is reduced from 8.5" to 6.8". The result of the
simulation shows that pack-off increases the well pressure dramatically. As shown on Figure
27, until the clock is 9.00am, the ECD will not exceed the fracture gradient. It might not look
like this, but this is due to scaling limits in Track 7. The linearly increasing ECD trend is due
to the gradually increased pack-off length. For example at 9.00 am the pack-off length is 40
m, and at 12.00 am the pack-off length is 160 m. Again, it is important to emphasize that a
drilling operation would never be continued as shown in this simulation, since the pack-off
would have caused problems and led to a halt in the operation.
Fracture
Figure 27 - Pack-off scenario 3 (20% reduction in annulus diameter).

64
6.1.5 Lost circulation scenario
During a drilling operation there will be a continuous change in the mud system. The system
is continuously given additives such as basic fluids, solids, and chemicals to compensate for
change in hole size, lost mud and volumes occupied by cuttings. A certain amount of liquid
may be bound to the cuttings or lost to permeable zones within the wellbore. This is a mass
balance equation that we easily can control. Lost circulation is a measurable loss of the total
drilling fluid to the formations.
In this scenario, there are performed three different simulations, illustrating both partial and
total loss to formation at the 9 5/8” casing shoe and total loss at 4120m. The operation is
ongoing as normal until the point where mud is lost to the formation. An illustration for total
loss at the 9 5/8” casing shoe is shown in Figure 28.
Figure 28 - Illustration of total loss of circulation at 9 5/8" casing shoe.
Parameters used for simulation of the lost circulation scenario:

- Average ROP 40 m/h
65
6.1.5.1 Partial loss at 9 5/8” casing shoe
Figure 29 shows the simulation result of partial loss to the formation below the 9 5/8” casing
shoe. Drilling fluid is assumed to be partially lost into the formation at 9.00pm. In Track 4
flow rate out (FLOWOUT) decreases from 2000 lpm to 0 lpm in steps of 500 lpm. This
results in an ECD that is trying to compensate for the decrease in flow rate (ECD decreases).
However, since we are drilling deeper and deeper, the pressure in bar does not decrease since
the hydrostatic column is increasing. However, the case is that we are loosing mud and this is
reflected in the ECD, which is given in sg, and more easily reflects the loss of frictional
forces.
Figure 29 - Lost circulation scenario 1 (Partial loss at 9 5/8" casing shoe).
66
6.1.5.2 Total loss at 9 5/8” casing shoe
It is normal that we start loosing mud gradually as shown in Figure 29. Figure 30 illustrates
total loss below the 9 5/8” casing shoe. In this case we loose all the mud at once. When the
flow rate out drops from 2000 lpm to 0 lpm in Track 4, this results in a strong reduction in the
ECD at 9:00 am. Consequently, this also reduces the SPP and BHP, since the friction is lost in
parts of the well.
Figure 30 - Lost circulation scenario 2 (Total loss at 9 5/8" casing shoe).
67
6.1.5.3 Total loss at 4120 m
Figure 31 is similar to the simulation performed for total loss of circulation at the 9 5/8"
casing shoe. The difference is that that the total loss of mud is at 4120 m.
Figure 31 - Lost circulation scenario 3 (Total loss at 4120 m).
68
6.1.6 ROP vs. Cuttings concentration scenario
The simulation result of the ROP and cutting concentration effect scenario is illustrated in
Figure 32, and shows how ramping up the average ROP from 20 to 80 m/h affects the mud
weight and the cuttings concentration. In the model, we have assumed that the cuttings are
transported with the liquid velocity. The sensor for measuring the cuttings concentration is
located right above the bit and increases momentarily in Track 3 as the ROP increases. The
sensor for measuring the mud weight out is located at the surface, shown in Track 4, and does
not respond until there have been a bottoms-up circulation. In Track 7, the ECD formula is
including the transient effect that takes into consideration the increasing mud weight in the
annulus as we circulate. After the bottoms-up circulation the ECD stabilizes at a constant
value, which reflects that the cuttings is present in the whole well.
Figure 32 - ROP vs. Cuttings concentration scenario.

69
6.1.7 Comparison of different rheology models for hydraulic calculations
The different models were compared with respect to pressure predictions. These are Bingham
(B), Robertson-Stiff (R-S), and Herschel-Bulkley (H-B). Figure 33 shows the simulation
results. In Track 7, H-B and R-S give similar results, while B predicts lower pressure than the
others in the annulus. For SPP in Track 5, H-B and B gave similar results, while R-S predicts
lower pressure in drillpipe. This might fit with the theoretical part where it was said that B is
not the best model to describe the annulus. Generally, it is said that three-parameter models
are better than two-parameter models. This is also reflected in Track 7, where R-S and H-B
model seems to fit quite well. H-B was chosen for the simulations, since the model did not
show any large deviations compared to the other models.
Figure 33 - Comparison of different rheology models for hydraulic calculations.

70
6.1.8 Washout scenario
Today, technology develops faster than before, which clearly is a mechanism that helps us to
improve the way we operate. One of these new technologies is Intellipipe [23], a pipe that
includes sensors along the drillstring to measure e.g. pressure and temperature at different
locations in the wellbore in real-time. By using this equipment it will for instance be possible
to detect a leak in the drillstring. In this scenario, we have placed three sensors, respectively at
1000 m, 2000 m and 3000 m. Unknowingly, a leak has occurred at 1500 m, and by using real-
time monitoring it should be possible to locate the approximate position of the leak. Figure 34
shows an illustration of a washout scenario in drillpipe (Intellipipe).
Figure 34 - Illustration of washout scenario in drillpipe (Intellipipe).
Model input parameters used for simulating the washout scenario:

- No reciprocation
71
6.1.8.1 Washout scenario
The washout scenario simulates loss of drilling fluid at 1500 m through a washout in drillpipe.
The drilling fluid is lost from drillpipe to annulus from 0 lpm to 2000 lpm. At 2000 lpm loss
rate, the well below 1500 m is dead, and is only affected by the weight of the hydrostatic
column. Figure 35 illustrates that a washout simulation scenario is established at 10.00am.
From Track 3-5 it is clear that the washout has occurred between Sensor 1 (1000 m) and
Sensor 2 (2000 m). In Track 3, Sensor 1 takes no affect of the washout due to the fact that it is
located above the loss point. Since Sensor 2 and 3 are located below the loss point, they will
experience a loss in pressure. Sensor 3 more than Sensor 2 due to the distance from the leak
point (would therefore sense that a large portion of the well is dead). Reduction of the annular
friction will cause the ECD to drop, experiencing only the hydrostatic column from the loss
point.
Figure 35 - Washout scenario in Intellipipe.

72
6.2 Simulation based on real well based data
In this work there have been several attempts to link the implemented model to a real well and
create similar scenarios. However, in order to do this, additional parameters need to be more
easily available in Discovery Web for real well consideration. This includes geometry data, as
for instance hole diameter, casing depths, drillpipe and BHA components. It will also be very
important to have the rheological data for the mud being used available from the vendors. In
addition, if more advanced cuttings transport models were to be implemented [17],
information about the cuttings size could also be a possible input parameter.
Due to the similarity of the self-made scenario, in terms of manually entering the parameters,
it is decided not to retrieve data from old wells to rebuild similar scenarios. Nevertheless, it is
determined that in the future it will be possible to run the model through a real-time system,
and automatically retrieve the necessary parameters. Further follow-up work to this task will
be given in the future work chapter.
73
7. Summary and Discussion
Since the introduction of petroleum exploration and exploitation the drilling technology is
showing advancement. However, as the energy demand increases, the industry is expanding
exploration activities in very challenging environments. These are to mention, deep-water,
depleted formations, gas hydrated formations, HPHT, arctic and extended reach formations.
Drilling with conventional methods in challenging environments limits the operations and
may cause undesired problems. It is not possible to say it too often, the E&P challenges are
only increasing. By increased costs, falling production and stricter environmental regulations,
it is essential to be aware of the challenges and reach strategic goals. During drilling
operations today, many operators are facing increased NPT, due to drilling incidents (pack-
off, kick, poor hole cleaning, fracture, collapse and lost circulation). Early symptom detection,
armed with real-time calibrated process models, will help us to manipulate hydraulic drilling
parameters and avoid unexpected events. Wellbore condition evaluation is based on detailed
process models that are capable of predicting downhole hydraulic, thermal and mechanical
affects during a drilling operation. By introducing the model to multiple drilling problem
symptoms, the possibility for decreasing the NPT only increases [2, 3, 35].
In order to solve the conventional drilling problems, the industry is introducing new advanced
drilling methods such as MPD, UBO, casing and liner drilling. For instance, the company
Reel Well [38] has introduced a new drilling solution that aims of drilling wells that reaches
beyond 20 km. All these technologies use control and monitoring systems that are based on
real-time measured data.
It is no doubt that the introduction of good IO tools will help the ongoing drilling operations.
Information is internally transferred through network systems to remote users, operators or
managing systems in order to tie together different specialists, improve capacity and reduce
costs. By using an aggregated system for collecting data from all service companies, operators
are able to gather real-time data from multiple data sources. Data is taken from the rig
systems in many different formats and then aggregated into the onshore drilling databases
[35].
Discovery Web is a web-based browser that will help us to reach out to all the people
involved by implementing a visualization and collaboration tool for multi-disciplinary target
groups. Having developed a model that simulates different drilling scenarios, using only static
values, mnemonics and outputs, I am confident that Kongsberg Oil & Gas Technologies
brings an equally good tool to the marked. In addition to real-time monitoring, the feature
Discovery Web Formulas gives you the ability to create your own formulas. By this it is
possible to create models that can be compared with real data in order to evaluate the
predictive capability of the model. Outputs from a formula can be visualized and used in the
most common widgets similar to log data. In addition, when having models running along
with real-time data, deviating trends may be an indicator of unwanted events taking place in
the well. There must be a reason why the model does not fit the data anymore. Hence, they
can in theory function as warning indicators. By building a monitoring panel in Discovery
Web, based on the proper rheological models and hydraulic calculations, this thesis have been
used to show how models and simulations can be combined in Discovery Web. The next step
would be to compare the developed model against a real well. In addition, the model
implemented can also be made more advanced [3, 35].
74
The experience has been that it was easy to implement the model using Discovery Web
Formulas. Another strength of the software is that it is very easy to visualize and import data
from real wells in this application. Hence, it is a very good tool for comparing models with
real data. This real-time data handling capability and visualization flexibility is considered as
one of the major strengths.
Proper training combined with new advanced tools is a key element in preventing undesirable
incidents. No matter how good and advanced the models are, it is important to not only be
critical with respect to data, but also know how to be able to interpret the data in the best
possible way. Nevertheless how advanced new tools are, it does not help if the people do not
know how the tools work and how to use them. By allowing new students to play around with
powerful tools such as Discovery Web (by building models and interpreting data), it will raise
the awareness around drilling incidents and enable students to easily pick up and understand
new information when they start working after graduation.
In this thesis, three hydraulic simulation models were implemented in Discovery Web. These
are: Bingham, Robertson-Stiff and Herschel-Bulkley. In order to illustrate the applicability of
the implemented models, a case study was presented which was based on drilling a well from
4000 m. Based on the input of rheology, geometry, flow rate, mud densities and ROP, the
following drilling scenarios have been simulated.
• Connection scenario
o Stop in flow rate (in) reduces the SPP, BHP and ECD.
o SPP is reduced from 293 to 0 bar during a connection.
o BHP increases linearly due to the increased hydrostatic column when drilling
deeper, and drops 40 bars during a connection.
o ECD varies from 1.60 - 1.70 sg during a connection.
• Kick scenario during drilling

o Unexpected high pore pressures cause this type of kick.
o In this case, the ECD is lower than pore pressure.
o The planned mud weight is insufficient.
• Kick scenario during connection

o During a connection the well pressure caused by the static mud weight is lower
than pore pressure and is inducing a kick.
o One should always plan for having a static mud weight above pore pressure
when including safety margins (swab or riser margin).
• Pack-off scenario and sensitivity of pack-off

o Decrease in outer diameter of annulus increases the SPP, BHP and ECD.
o In the simulation, pack-off length increases while drilling (after 1 hour the
pack-off length is 40 m and after 4 hours the pack-off length is 160 m). In a
real situation formation will break down when exceeding fracture pressure, but
here we have demonstrated the pressure effects.
o 10% pack-off: ECD varies from 1.61 - 1.72 sg, SPP 0 - 300 bar.
o ECD drops to 1.61 sg (MW 1.60 sg) due to the effect of cuttings concentration
in the annulus.
75
• Lost circulation scenario and sensitivity of lost circulation
o Decrease in flow rate (out) reduces the SPP, BHP and ECD.
o While drilling deeper, the BHP in bar can increase even if we have losses,
since the hydrostatic pressure is increased due to deeper wells. However, the
ECD will be reduced.
• ROP vs. Cuttings concentration scenario

o Increase in ROP will increase the cuttings concentration in the annulus, which
again will increase the static mud weight and ECD.
o The full response on the ECD will first be seen when the new cuttings
concentration has been circulated bottoms-up. This was reflected in the
simulation. This is also reflected in the sensor located at the bottom, while the
mud weight out will first increase after one bottoms-up circulation.
o In this case, no slip conditions were assumed. However, more advanced
cuttings transport models should be considered, taking into account slippage
(which will affect concentration profile) and bed build up [17].
o The ECD increases 0.014 sg when changing the ROP from 20 to 80 m/h.
• Hydraulics and rheology model comparison

o The three rheological models for hydraulic calculations were compared. These
are: Bingham, Robertson-Stiff and Herschel-Bulkley.
o For the given rheological and hydraulic data, the models gave similar result for
SPP and ECD. However, some discrepancies could be observed.
o SPP for Robertson-Stiff is 90 bars lower than for Bingham and Herschel-
Bulkley, this is probably due to a lower frictional loss in drill pipe for this
model. Bingham gave a slightly lower annular frictional pressure drop.
o We chose to use Herschel-Bulkley since this model gave the least variations
compared to the others. In addition, a three-parameter model is usually more
accurate than a two-parameter model, like Bingham.
• Washout scenario
o The purpose was to demonstrate the advantage of the new technology
Intellipipe.
o The simulation shows that it is possible to use drillpipe with implemented
pressure sensors at different locations in the annulus to detect location of
washout in drillpipe.
o Three sensors are considered, and the washout takes place between the two
uppermost sensors. Sensor 1 takes no affect of the washout due to the fact that
it is located above the loss point. Sensor 2 and 3 are located below the loss
point and will experience a loss in pressure. Sensor 3 will see the effect more
than Sensor 2 due to the increased distance to the leak point.
o Reduction of the annular friction will cause the ECD to drop, experiencing
only the hydrostatic column of mud from the leak point.
76
8. Conclusion
Good real-time monitoring and control systems are one of the major factors for a successful
drilling operation. By implementing good IO technology, several drilling related problems
such as excessive torque and drag, kick, stuck pipe and lost circulation can be reduced and
avoided. Real-time monitoring systems run both measurement and model based predictions in
parallel. Whenever the measurement system fails, the model-based prediction can be used to
monitor the downhole condition as illustrated in this thesis with Discovery Web. In addition,
when we run models and measurements in parallel, the models can be calibrated to fit the real
data. However, when discrepancies in trends of the measured data and the model-based
predictions occur it can be a sign of detecting worsening downhole conditions and a first
warning of an unexpected event. They can also be used for forward prediction based on a
current real-time calibrated model state.
One example can be to have measured hydraulic data displayed along with models for
predicting well pressures. Such models can be very complex when integrating fluid
mechanics, solid mechanics and thermodynamics. Here relatively simple models for well
pressure predictions were implemented and it was shown how one could visualize those data.
In addition, it was shown how different parameters such as e.g. ROP and flow rate would
affect the trends. It was also of importance to demonstrate different events/unwanted
situations that can occur, and how that would affect the trend lines / pressure development.
Since the introduction of real-time monitoring systems there have been several studies that
have been carried out to investigate the behavior of hydraulic drilling parameters and how
they can be used in early symptom detection.
This thesis comes to the conclusion that:

- Discovery Web is a very good tool for showing and taking into use real-time data from
real wells.
- Knowledge about drilling problems and real-time monitoring systems are a
combination that certainly should be more practiced by the universities to raise the
awareness on decreasing NPT.
- It is shown that one can use Discovery Web Formulas to enter formulas and models,
and how we can visualize them. Later these models can be compared against actual
data when all input well data is available. In this case, it was focused on rheological
models for hydraulic calculations.
- It is a tool that is very flexible in terms of visualization processing and it is made
suggestions for how this can be visualized.
- It has also been shown how to embed events into the models, which later can be used
as a basis for developing training scenarios and demonstrations in teaching.
- The results show how the models implemented can be the first step in introducing
models in combination with real-time data for monitoring and handling drilling
problems using this application. In terms of gathering information, the data needs to be
of high quality and the equipment equally reliable when drilling in challenging
environments.
- Advanced real-time monitoring systems can be used to drill difficult wells and aid in
avoiding bad incidents that will have negative impacts on the operation. Future wells
will be of the difficult type, hence there should be a need for such tools in order to be
able to exploit complex hydrocarbon reserves.
77
9. Future Work
The following bullet points includes recommendations for future work:
• Implement pressure and temperature effects in density calculation and rheological
models.
NORSOK D-010 [21] defines HPHT as a “well with expected shut-in pressure
exceeding 690 bar, (10,000 psi), and a static bottomhole temperature higher than 150
°C”. HPHT has an impact on the mud properties and could cause well control issues.
The surface mud weight should be adjusted correspondingly for thermal and pressure
effects on the effective fluid density in the well. Temperature is the most critical
parameter since it will lower the specific gravity of the mud. Similarly the pressure
will increase the specific gravity. Figure 36 illustrates the temperature effect on
density for a simulated scenario [39] .
Figure 36 - Temperature effect on density [39].
• Implementing Matlab as a second programming language

The idea is that the Excel sheet will function as a very simplified steady state
hydraulic model calculating pressures in a well. A steady state model gives only a
snapshot of the well condition at a given time. However, by introducing Discovery
Web we were able to do the simulations over time. Excel is a great tool if you need to
have the values and formulas for inputs and outputs in tables in front of you. Matlab is
best for “sophisticated” math, especially regarding large data sets; like matrix algebra
and differential equations. If one were to implement a transient flow model, Matlab
would have been the right tool due to the introduction of space discretization grids.
Where local calculation of hydrostatic pressure and frictional losses will take into
account increasing pressure and temperature versus depth, instead of assuming fixed
values in the whole well.
78
• Improve graphical representation
By taking real-time monitoring a step further, one should make good graphical tools
that clearly indicate when something is about to go wrong. For example, by creating a
"speedometer look" for the mud weight trend, where the arrow was located in a green
area if everything was ok and switched over to the red marked area if the mud weight
was too high, or too low.
• Implement an MPD system

The code implemented in Discovery Web already contains the ability to manipulate
the choke opening and activate a back-pressure pump. Furthermore, it could be
possible to run simulations that show the optimal adjustments to maintain a constant
well pressure during a connection. For a MPD system the ECD formula can be written
as Eq. 56.
∆𝑃! + 𝐶ℎ𝑜𝑘𝑒 𝑝𝑟𝑒𝑠𝑠𝑢𝑟𝑒 (56)

𝐸𝐶𝐷 = 𝜌 +
𝑔ℎ
By starting up the back-pressure pump, one could avoid the drop in ECD during a
connection, thus keeping us inside drilling window in challenging environments.
79
References
1. BP, 2010, Deepwater Horizon Accident Investigation Report, [Report] [cited
03.03.2014], Available from:
https://http://www.itslearning.com/file/download.aspx?FileID=1055334&FileVersionI
D=-1.
2. Cayeux, E., et al., 2012, Early Symptom Detection Based on Real-Time Evaluation of
Downhole Conditions: Principles and Results from several North Sea Drilling
Operations, [Conference Paper] SPE-150422-MS, presented at SPE Intelligent Energy
International, Utrecht, The Netherlands, 27-29 March 2012.
3. Kongsberg Oil & Gas Technologies, 2012, Discovery Web - an integrated IO /

collaboration and visualization platform, [PowerPoint presentation prepared for:
SBBU].
4. Falconer, I.G., J.P. Belaskie, and F. Variava, 1987, Applications of a Real Time
Wellbore Friction Analysis, [Conference Paper] SPE-18649-MS, presented at
SPE/IADC Drilling Conference, New Orleans, Louisiana, February 28-March 3, 1989.
5. Niedermayr, M., et al., 2010, Case Study--Field Implementation of Automated Torque-

and-Drag Monitoring for Maari Field Development, [Conference Paper] SPE-128243-
MS, presented at IADC/SPE Drilling Conference and Exhibition, New Orleans,
Louisiana, USA, 2-4 February 2010.
6. Cayeux, E. and B. Daireaux, 2009, Early Detection of Drilling Conditions

Deterioration Using Real-Time Calibration of Computer Models: Field Example from
North Sea Drilling Operations, [Conference Paper] SPE-119435-MS, presented at
SPE/IADC Drilling Conference and Exhibition, Amsterdam, The Netherlands, 17-19
March 2009.
7. Aadnoy, B.S. and S. Ong, Introduction to special issue on Borehole Stability,

Petroleum Science and Engineering, 2003, 38(3-4): p. 79-82.
8. Hovda, S., et al., 2008, Potential of Ultra High-Speed Drill String Telemetry in Future
Improvements of the Drilling Process Control, [Conference Paper] SPE-115196-MS,
presented at IADC/SPE Asia Pacific Drilling Technology Conference and Exhibition,
Jakarta, Indonesia, 25-27 August 2008.
9. Strand, S., 1998, Øvinger i bore- og brønnvæsker, Høgskolen i Stavanger, Stavanger.
10. Kolle, G. and R. Mesel, 1998, Brønnvæsker: for VK1 brønnteknikk, Vett & Viten,
Nesbru, ISBN: 82-412-0322-5.
11. Okafor, M.N. and J.F. Evers, 1992, Experimental Comparison of Rheology Models for
Drilling Fluids, [Conference Paper] SPE-24086-MS, presented at Western Regional
Meeting, Bakersfield, California, March 30-April 1, 1992.
80
12. Kok, M.V. and T. Alikaya, Effect of Polymers on the Rheological Properties of
KCl/Polymer Type Drilling Fluids, Energy Sources, 2005, 27(5): p. 405-415.
13. Ochoa, M.V., 2006, Analysis of Drilling Fluid Rheology and Tool Joint Effect to
Reduce Errors in Hydraulics Calculations, [PhD Thesis], Texas A&M University, p.
99.
14. Skaugen, E., 1997, Kompendium i boring, Høyskolen i Stavanger, Stavanger.
15. Azar, J.J. and G.R. Samuel, 2007, Drilling engineering, [Book], PennWell Corp.,
Tulsa, Oklahoma, USA.
16. Gabolde, G. and J.-P. Nguyen, 2006, Drilling data handbook, [Book], Editions
Technip, Paris. ISBN: 2-7108-0871-4.
17. Larsen, T.I., A.A. Pilehvari, and J.J. Azar, Development of a New Cuttings-Transport
Model for High-Angle Wellbores Including Horizontal Wells, SPE Drilling &
Completion, 1997, 12(02): p. 129-136. SPE-25872-PA.
18. Cayeux, E., et al., 2013, Real-Time Evaluation of Hole Cleaning Conditions Using a
Transient Cuttings Transport Model, [Conference Paper] SPE-163492-MS, presented
at SPE/IADC Drilling Conference and Exhibition, Amsterdam, The Netherlands, 5-7
March 2013.
19. Evje, S. and K.K. Fjelde, Hybrid Flux-Splitting Schemes for a Two-Phase Flow
Model, Journal of Computational Physics, 2002, 175(2): p. 674-701.
20. V. C. Kelessidis, G.M., A. Koutroulis, T. Michalakis, Significant parameters affecting

efficient cuttings transport in horizontal and deviated wellbores in coil tubing drilling:
A critical review, 1st International Symposium of the Faculty of Mines (ITU) on Earth
Sciences and Engineering, 2002, 19(04): p. 213-227. SPE-81746-PA.
21. Standards Norway, Well integrity in drilling and well operations, in NORSOK D-010
Rev. 4. 2013.
22. Vett & Viten, E.A., 2008, Borekunnskap: modul 0401, Nesbru.
23. NOV, 2013, Intelliserv Product Catalog, [Web Page], [cited 04.06.2014], Available
from:
http://www.nov.com/uploadedFiles/Business_Groups/Downhole/IntelliServ/LandingP
age/IntelliServ-Product-Catalog-D392004875-MKT-001-Rev-01.pdf.
24. Segura, J. and Weatherford International, Drillpipe Cutting at Ultrahigh Pressure

Proven for Remediating Deepwater Stuck-Pipe Hazards, SPE Drilling & Completion,
2011, 26(04): p. 569-577. SPE-139511-PA.
25. Nakhleh, J.T.Ø., 2009, Drilling problem scenarios for Simulation and Training, [BSc
Thesis], Institute of Petroleum Technology, University of Stavanger, p. 55.
26. Belayneh, M., 2013, MPB 110 Geology and Petro-physics lecture note, [Lecture].
81
27. Aadnoy, B.S., K. Larsen, and P.C. Berg, 1999, Analysis of Stuck Pipe in Deviated
Boreholes, [Conference Paper] SPE-56628-MS, presented at SPE Annual Technical
Conference and Exhibition, Houston, Texas, USA, 3-6 October 1999.
28. Wang, H., et al., Best Practice in Understanding and Managing Lost Circulation
Challenges, SPE Drilling & Completion, 2008, 23(02): p. 168-175. SPE-95895-PA.
29. Sekal, 2014, DrillScene Advanced Monitoring, [Web Page], [cited 11.03.2014],
Available from: http://sekal.com/?id=1017.
30. Cayeux, E., et al., 2012, Advanced Drilling Simulation Environment for Testing New
Drilling Automation Techniques, [Conference Paper] SPE-150941-MS, presented at
IADC/SPE Drilling Conference and Exhibition, San Diego, California, USA, 6-8
March 2012.
31. Iversen, F., et al., Drilling Automation: Potential for Human Error, SPE Drilling &
Completion, 2013, 28(01): p. 45-59. SPE-151474-PA.
32. Rommetveit, R., et al., 2008, eDrilling used on Ekofisk for Real-Time Drilling
Supervision, Simulation, 3D Visualization and Diagnosis, [Conference Paper] SPE-
112109-MS, presented at SPE Intelligent Energy Conference and Exhibition,
Amsterdam, The Netherlands, 25-27 February 2008.
33. Schlumberger, 2014, Drillbench Dynamic Drilling Simulation Software, [Web Page],
[cited 05.04.2014], Available from:
http://www.software.slb.com/products/foundation/Pages/drillbench.aspx.
34. Rommetveit, R., et al., 2004, Use of Dynamic Modeling in Preparations for the
Gullfaks C-5A Well, [Conference Paper] SPE-91243-MS, presented at SPE/IADC
Underbalanced Technology Conference and Exhibition, Houston, Texas, USA, 11-12
October 2004.
35. Kongsberg Oil & Gas Technologies, 2014, Drilling Software Business Unit, [PDF
handout].
36. Kongsberg Oil & Gas Technologies, 2012, Welcome to Discovery Web, [Help
Manual], [cited 19.02.2014], Available from:
http://kidemo.intellifield.no/DWeb/Help/Discovery_Web.htm - Topics/Welcome.htm.
37. Kongsberg Oil & Gas Technologies, 2014, Discovery Web Interface, [Pictures],
Screenshots from program.
38. Reel Well AS, 2014, JIP: ERD beyond 20 km, [Web Page], [cited 27.05.2014],
Available from: http://www.reelwell.no/Extended-Reach-Drilling/JIP-ERD-beyond-
20-km.
39. Belayneh, M., 2014, Chapter 6: Well Control, PET-525, [Lecture].
82
Appendix A
Conservation laws
When dealing with mass balance there are three fundamental laws that also apply to well and
pipe flow, they are as follow: conservation of mass, momentum and energy. If we consider
conservation of mass in a pipe segment it can be expressed in the following manner [19]:
Mass flux in = Mass flux out

Mass at new time level (n+1) = mass at old time level (n) + [(mass flux in)-(mass flux out)]
By introducing the drift flux model, which is a simplified version of the more fundamental
two fluid models, it is possible to describe one and two phase flow in pipe at a transient level.
The model is derived from the Navier Stokes equations and combines the momentum
equations for the mixture and an additional equation expresses the slippage between gas and
liquid. This surely makes the model dependent on the type of two-phase flow pattern (bubble,
dispersed bubble, slug and annular flow). The model describes 1D flow.
Transient flow model

Consider a well segment, where t is time and z is the flow direction through this segment.
The following nonlinear partial differential equations for conservation of mass and
momentum across this segment can be expressed as follows [19]:
Conservation of liquid mass:

𝜕 𝜕 (57)
𝐴𝜌! 𝛼! + 𝐴𝜌! 𝛼! 𝑣! = 𝑠!
𝜕𝑡 𝜕𝑧
Conservation of gas mass:

𝜕 𝜕 (58)
𝐴𝜌! 𝛼! + 𝐴𝜌! 𝛼! 𝑣! = 𝑠!
𝜕𝑡 𝜕𝑧
Conservation of mixture momentum:

𝜕 𝜕 𝜕 ∆𝑝!"#$ (59)
𝐴(𝜌! 𝛼! 𝑣! + 𝜌! 𝛼! 𝑣! ) + 𝐴 𝜌! 𝛼! 𝑣!! + 𝜌! 𝛼! 𝑣!! +𝐴 𝑝 = −𝐴(𝜌!"# 𝑔 + )
𝜕𝑡 𝜕𝑡 𝜕𝑧 ∆𝑧
where A is area, 𝜌! is phase densities (liquid I = l, gas i = g), 𝑣! is phase velocities, p is

pressure, 𝑠! is source (inflow, leakage, phase transfer between phases), g is gravity constant,
𝛼! is phase volume fractions taking values between 0 and 1, 𝜌!"# = 𝜌! 𝛼! + 𝜌! 𝛼! , 𝑣!"# =
𝛼! 𝑣! + 𝛼! 𝑣! , 𝜇! is phase viscosities, 𝜇!"# = 𝜇! 𝛼! + 𝜇! 𝛼! , 𝑑!"# is outer diameter in annulus
and 𝑑!" is inner diameter in annulus and corresponds to outer diameter of drill string.
To be able to solve the equations, which contains a certain number of unknowns it might be
necessary to add some closure laws, to ensure that the number of equations is the same as the
number of unknowns [19].
83
Appendix B
1. Rheology and Hydraulics Equations for Bingham Plastics Model [13].
Pipe Flow Annular Flow
0.408𝑄 0.408𝑄
𝑣! = 𝑣! =
𝐷!! 𝐷!! − 𝐷!!
𝜇! = 𝜃!"" − 𝜃!""
𝜏! = 𝜃!"" − 𝜇!
5𝜏! 𝐷! 5𝜏! 𝐷! − 𝐷!
𝜇! = 𝜇! + 𝜇! = 𝜇! +
𝑣! 𝑣!
928𝐷! 𝑣! 𝜌 757 𝐷! − 𝐷! 𝑣! 𝜌
𝑁!" = 𝑁!" =
𝜇! 𝜇!
If 𝑁!" < 2100 If 𝑁!" < 2100

16 16
𝑓! = 𝑓! =
𝑁!" 𝑁!"
37,100𝜌𝜏! 𝐷!! 37,100𝜌𝜏! 𝐷!!

𝐻𝑒 = 𝐻𝑒 =
𝜇!! 𝜇!!
928𝐷! 𝑣! 𝜌 𝐷! = 0.816 𝐷! − 𝐷!
𝑁!" =
𝜇!
928𝐷! 𝑣! 𝜌
𝑁!" =
𝜇!
Hedstrøm number. 0.0791

𝑓!,! = !.!"
𝑁!"
𝑑𝑝 𝑓! 𝑣!! 𝜌 𝑑𝑝 𝑓! 𝑣!! 𝜌
= =
𝑑𝐿 25.81𝐷! 𝑑𝐿 25.81 𝐷! − 𝐷!
𝑑𝑝
∆𝑝 = ∆𝐿 𝑑𝑝
𝑑𝐿 ∆𝑝 = ∆𝐿
𝑑𝐿
156𝜌𝑞 !
∆𝑝!"##$%&,!"# = !
𝐷!! ! + 𝐷!! ! + 𝐷!! !
84
2. Rheology and Hydraulics Equations for Herschel-Bulkley Model [13].

!
𝜏 ∗ − 𝜏!"# 𝜏!"#
𝜏! =
2𝜏 ∗ − 𝜏!"# − 𝜏!"#
𝛾∗ = 𝛾!"# 𝛾!"#
log (𝜏 − 𝜏! ) = log (𝐾) + 𝑛 𝑙𝑜𝑔(𝛾)
0.408𝑄 0.408𝑄
𝑣! = 𝑣! =
𝐷!! 𝐷!! − 𝐷!!
𝐷! ! 𝐷! − 𝐷! !
!!!
2 3𝑛 + 1 𝜌𝑣!!!! 4 2𝑛 + 1 𝜌𝑣!
2 2
𝑁!" = 𝑁!" =
𝑛 𝐷! ! 3𝑛 + 1
!
𝑛 𝐷! ! 2 2𝑛 + 1
!
𝜏! +𝐾 𝜏! +𝐾
2𝑣! 𝑛𝐶! 2𝑣! 𝑛𝐶!
! !
4 3𝑛 + 1 !!! 8 2𝑛 + 1 !!!
𝑁!" !" = 𝑁!" !" =
𝑛𝑦 𝑛𝑦
𝑙𝑜𝑔𝑛 + 3.93
𝑦=
50
1.75 − 𝑙𝑜𝑔𝑛
𝑧=
7
Laminar if: 𝑁!" < 𝑁!" !"
!
!
𝑑𝑝 4𝑘 𝜏! 3𝑛 + 1 8𝑄 𝑑𝑝 4𝑘 𝜏! 16(2𝑛 + 1) 𝑄
= + = +
𝑑𝐿 14400𝐷! 𝑘 𝑛𝐶! 𝜋𝐷!! 𝑑𝐿 14400 𝐷! − 𝐷! 𝑘 𝑛𝐶! (𝐷! − 𝐷! )
𝜋
𝐷! !
−
𝐷! !
2 2
Turbulent if: 𝑁!" > 𝑁!" !"
!!
𝑓! = 𝑦 𝐶! 𝑁!" 𝑓! = 𝑦 𝐶!∗ 𝑁!" !!
1 𝜏! 1 𝜏!
𝐶! = 1 − 𝐶!∗ = 1 − !
2𝑛 + 1 ! 1+𝑛 2 2𝑛 + 1 𝑄
𝜏! + 𝑘
3𝑛 + 1 𝑄 𝑛 𝐷! 2 − 𝐷! 2 𝜋 𝐷! 2 !− 𝐷! 2 !
𝜏! + 𝑘 𝐷!
𝑛𝜋( )!
2
𝑑𝑝 𝑓! 𝑄 ! 𝜌 𝑑𝑝 𝑓! 𝑄 ! 𝜌
= =
𝑑𝐿 144𝜋 ! 𝐷!! 𝑑𝐿 144𝜋 ! 𝐷! − 𝐷! 𝐷!! − 𝐷!! !
𝑑𝑝
∆𝑝 = ∆𝐿
𝑑𝐿
156𝜌𝑄 !
∆𝑝!"##$%&,!"# = !
𝐷!! ! + 𝐷!! ! + 𝐷!! !
85
3. Rheology and Hydraulics Equations for Robertson-Stiff Mode [13].

!
𝛾!"# 𝛾!"# − 𝛾 ∗
𝐶=
2𝛾 ∗ − 𝛾!"# − 𝛾!"#
log (𝜏) = log (𝐴) + 𝐵 𝑙𝑜𝑔(𝛾)
0.408𝑄 0.408𝑄
𝑣! = 𝑣! =
𝐷!! 𝐷!! − 𝐷!!
! !
89100𝜌𝑣!!!! 0.416𝐷! 109000𝜌𝑣!!!! 0.0208 𝐷! − 𝐷!
𝑁 !" = 𝑁 !! =
𝐴 1 𝐴 1
3+𝐵 2+𝐵
Laminar if: 𝑁!" ≤ 3470 − 1370𝐵
!
𝐶
! 𝑑𝑝 1 + 2𝐵 0.2𝑣! + 𝐷 − 𝐷!
𝐶 = 8.33𝐸 − 4×4!!! ×𝐴 8 !
𝑑𝑝 1 + 3𝐵 0.2𝑣! + 𝐷 𝑑𝑙 𝐵 !!!
= 8.33𝐸 − 4×2!!! ×𝐴 6 ! 𝐷! − 𝐷! !
!!!
𝑑𝑙 𝐵
𝐷! !
Turbulent if: 𝑁!" ≥ 4270 − 1370𝐵
𝑙𝑜𝑔𝐵 + 3.93
𝑎=
50
1.75 − 𝑙𝑜𝑔𝐵
𝑏=
7
𝑎 𝑎
𝑓! = ! 𝑓! = !
𝑁!" 𝑁!"
𝑑𝑝 𝑓! 𝑣!! 𝜌 𝑑𝑝 𝑓! 𝑣!! 𝜌
= =
𝑑𝐿 25.81𝐷! 𝑑𝐿 25.81 𝐷! − 𝐷!
𝑑𝑝
∆𝑝 = ∆𝐿
𝑑𝐿
156𝜌𝑄 !
∆𝑝!"##$%&,!"# = !
𝐷!! ! + 𝐷!! ! + 𝐷!! !
86
Appendix C
1. Excel Spreadsheet - FANN Rheology Data
Excel Spreadsheet using Fann rheology data for determination of the different rheological
model parameters.
Input:
Fann data
Measured Shear rate Shear rate Shear Stress
Reading shear stress (RPM) (Oil Field) (Oil Field)
θ600 54,50 600 1021,8 58,15
θ300 43,50 300 510,9 46,41
θ200 37,50 200 340,6 40,01
θ100 32,00 100 170,3 34,14
θ6 23,00 6 10,218 24,54
θ3 20,50 3 5,109 21,87
Output - Herschel-Bulkley:
Graphical fitting of experimental data:
Regression used to fit of experimental data:

Herschel-‐Bulkley Model
Shear Stress (corr) Shear rate Shear Stress

(Herschel-‐Bulkley) (Oil Field) (Herschel-‐Bulkley) error
38,01 1021,80 57,72 0,001
26,28 510,90 46,11 0,001
19,88 340,60 41,06 0,004 Calculated values
14,01 170,30 34,60 0,002 𝜏! 28,26
4,40 10,22 23,36 0,008 𝜏∗ 20,14
1,74 5,11 22,37 0,004 K 0,9350
Total error: 0,344 n 0,5331

87
Output - Robertson-Stiff Model:
Regression used to fit of experimental data:

Robertson-‐Stiff Model
Shear rate (corr) Shear Stress Shear Stress
(Robertson-‐Stiff) (Oil Field) (Robertson-‐Stiff) error
1073,81 58,15 56,70 0,004
562,91 46,41 46,32 0,000
392,61 40,01 41,38 0,006 Calculated values
222,31 34,14 34,63 0,002 AA 6.379
62,23 24,54 23,25 0,009 A 3054.25
57,12 21,87 22,63 0,006 B 0.313
Total error: 0,452 C 52.01

88
Output - Bingham Model:
Regression and formula used to fit of experimental data:

Bingham Model
Shear rate (corr) Shear Stress Shear Stress
(Bingham) (Oil Field) (Bingham) error
1021,8 58,15 60,96 0,008
510,9 46,41 42,32 0,011
340,6 40,01 37,43 0,011
170,3 34,14 31,55 0,013
10,2 24,54 26,03 0,010
5,1 21,87 25,85 0,030
Total error: 1,382

Formula Graphic
𝜇! 11,00 16,53
𝜏! 32,50 25,67
89
2. Implemented models in Discovery Web
Contains the Discovery Web code that is implemented into the simulations.
2.1. Bingham Frictional Model
Static Inputs: Variable Input:

Parameter:
Value:
Unit: Parameter: Value: Unit:
YP 11,00 lb/100ft2 Q 2000 lpm
PV 32,50 cP d 1,60 sg
BackPPump 0,00 lpm ROP 40 m/h
VcutTransp 1,02 m/s TIME 4 h
ChokeOD 3,00 inch
ChokeOpen 1,00 0-‐1
AnnCsgL1 3500,00 m
AnnCsgL2 500,00 m
AnnCsgID1 5,00 inch
AnnCsgID2 5,00 inch
AnnCsgID3 6,50 inch
DPL2 200,00 m
DPID1 4,27 inch
DPID2 2,50 inch
BitNoz 6,00
BitNozID 12,00
Motor
13,80 bar
MWD 13,80 bar
SurfEq 2,00 bar
dcuttings 2,50 sg
PackOffAnnL1 0,00 (0-‐1) %
CutCons 0,01 (0-‐1) %
InitialWellDepth 4000,00 m
Output
Parameter: Value: Unit: Discovery Web Formula:
dout 1,61 sg ((1-‐CutCons)*din+dcuttings*CutCons)*8.33
AnnCsgOD1 8,53 inch AnnCsgOD1*(1-‐PackOffAnnL1)
DPL1 3960,00 m (AnnCsgL1+AnnCsgL2-‐DPL2)+(ROP*TIME)
DPV1 3,61 m/s Q/1000/60/(3.14/4*(DPID1*0.0254)^2)
DPV2 10,53 m/s Q/1000/60/(3.14/4*(DPID2*0.0254)^2)
DPVCri1 1,15 (2.48/d/DPID1*(PV+SQRT(PV^2+73.57*YP*DPID1^2*d)))/60
DPVCri2 1,32 (2.48/d/DPID2*(PV+SQRT(PV^2+73.57*YP*DPID2^2*d)))/60
DPFricLam1 20,33 bar (((DPL1*Q*PV)/(612.95*DPID1^4))+((YP*DPL1)/(13.26*DPID1)))/100
DPFricLam2 6,09 bar (((DPL2*Q*PV)/(612.95*DPID2^4))+((YP*DPL2)/(13.26*DPID2)))/100
DPFricTurb1 105,72 bar (DPL1*d^0.8*Q^1.8*PV^0.2/(901.63*DPID1^4.8))/100
DPFricTurb2 69,73 bar (DPL2*d^0.8*Q^1.8*PV^0.2/(901.63*DPID2^4.8))/100

90

DPFricLamTurb1 105,72 bar IF(DPV1<DPVCri1,DPFricLam1,DPFricTurb1)
DPFricLamTurb2 69,73 bar IF(DPV2<DPVCri2,DPFricLam2,DPFricTurb2)
WellDepth 4160,00 m InitialWellDepth+(ROP*TIME)
AnnCsgL3 160,00 m (ROP*TIME)
LossShoe1 0,00 m IF(LossShoe=1,IF(WellDepth>=4060,AnnCsgL1,0),0)
LossTD1 0,00 m IF(LossTD=1,IF(WellDepth>=4120,AnnCsgL1,0),0)
VelAnn1 1,38 m/s (Q*4)/(60*1000*3.14*((AnnCsgOD1*0.0254)^2-‐(AnnCsgID1*0.0254)^2))
(PV+SQRT(PV^2+40.05*YP*(AnnCsgOD1-‐
LamTurb1 1,18 AnnCsgID1)^2*dout))*3.04/(AnnCsgOD1-‐AnnCsgID1)/d/60
(AnnCsgL1*Q*PV/(408.63*(AnnCsgOD1+AnnCsgID1)*(AnnCsgOD1-‐
AnnFricLam1 17,58 AnnCsgID1)^3)+YP*AnnCsgL1/13.26/(AnnCsgOD1-‐AnnCsgID1))/100
(AnnCsgL1*dout^0.8*Q^1.8*PV^0.2/(706.96*(AnnCsgOD1+AnnCsgID1)^
AnnFricTurb1 26,57 1.8*(AnnCsgOD1-‐AnnCsgID1)^3))/100
(AnnCsgL2*dout^0.8*Q^1.8*PV^0.2/(706.96*(AnnCsgOD2+AnnCsgID2)^
AnnFricTurb2 3,91 1.8*(AnnCsgOD2-‐AnnCsgID2)^3))/100
IF(LossTD3>=120,((120*dout^0.8*Q^1.8*PV^0.2/(706.96*(AnnCsgOD3+
AnnCsgID3)^1.8*(AnnCsgOD3-‐
AnnCsgID3)^3))/100),((AnnCsgL3*d^0.8*Q^1.8*PV^0.2/(706.96*(AnnCsg
AnnFricTurb3 5,55 OD3+AnnCsgID3)^1.8*(AnnCsgOD3-‐AnnCsgID3)^3))/100))
IF(LossShoe1=3500,0,IF(LossTD1=3500,0,IF((VelAnn1>LamTurb1),AnnFric
AnnFric1 26,57 bar Turb1,AnnFricLam1)))
AnnFric2 3,91 bar IF(LossTD2=500,0,IF((VelAnn2>LamTurb2),AnnFricTurb2,AnnFricLam2))
AnnFric3 5,55 bar IF((VelAnn3>LamTurb3),AnnFricTurb3,AnnFricLam3)
FlowAChoke 0 lpm Q+BackPPump
CumLAnn 4160,00 m AnnCsgl1+AnnCsgL2+AnnCsgL3
CumLDP 4160,00 m DPL1+DPL2
ChokeID 1,200 (0-‐1) ChokeOD-‐(ChokeOD*ChokeOpen)
ChokeArea 2,543 in2 3.14/4*(ChokeOD-‐ChokeID)^2
BitNozTFA 0,6623 inch2 (BitNozID/32)^2*3.14/4*BitNoz
BitNozPloss 54,62 bar (d*Q^2/2959.41/0.95^2/BitNozTFA^2)/100
SumAnnFric 36,03 bar IF(Q>0,(AnnFric1+AnnFric2+AnnFric3),0)
SumEqFric 29,60 bar IF(Q>0,(Motor+MWD+SurfEq),0)
SumDPFric 175,45 bar (DPFricTurb1+DPFricTurb2)
ChokePress 0 bar (dout*FlowAChoke^2/(2959.41*0.95^2)/ChokeArea^2)/100
PumpPress 299,42 bar SumAnnFric+BitNozPloss+SumEqFric+SumDPFric+ChokePress
ECDCalc 1,6983 sg ((dout*CumLAnn*0.0981)+SumAnnFric+ChokePress)/(CumLAnn*0.0981)
Hydrostatic 693.1 bar ECDCalc*CumLAnn*0.0981
91
2.2. Herschel-Bulkley Frictional Model
Static Input: Variable Input:

Parameter: Value: Unit: Parameter: Value: Unit:
BackPPump 0,00 lpm Qin 2000 lpm
ChokeOD 3,00 inch Qloss 50 lpm
ChokeOpen 1,00 0-‐1 din 1,60 sg
AnnCsgL1 3500,00 m ROP 40 m/h
AnnCsgL2 500,00 m TIME 4 h
AnnCsgID1 5,00 inch
AnnCsgID2 5,00 inch
AnnCsgID3 6,50 inch
DPL2 200,00 m
DPID1 4,27 inch
DPID2 2,50 inch
BitNoz 6,00
BitNozID 12,00
Motor 13,80 bar
MWD 13,80 bar
SurfEq 2,00 bar
dcuttings 2,50 sg
PackOffAnnL3 0 (0-‐1) %
Porosity 0,2
Washoutdepth 1500
m
Sensor1Depth 1000 m
Sensor2Depth 2000 m
Sensor3Depth 3000 m
Output -‐ Above WashoutDepth

d 13,33 ppg 0.26417*Qin
Q 528,34 gpm 8.33*din
Ro 18,87 degree (2*R3-‐R6)
To (FANN) 20,14 0.511*R0
n (FANN) 0,533 unitless 3.32*log((R600-‐R0)/(R300-‐R0))
lbf.^n/100f
K (FANN) 0,935 t 0.511*((R300-‐R0)/(511^n))
(ROP*0.000277778*((PI()/4)*(0.0254*AnnCsgOD3)^2)*(1-‐
Porosity))/((Qin*1.66667*10^-‐
5)+(((ROP*0.000277778)*((PI()/4)*(0.0254*AnnCsgOD3)^2)*(1-‐
CutCons 0,010 Porosity))))
dout 13,40 ppg ((1-‐CutCons)*din+dcuttings*CutCons)*8.33
AnnCsgOD1 8,53 inch 8.53*(1-‐PackOffAnnL1)

92

DPV1 11,82 ft/s 0.408*Q/DPID1^2
DPV2 34,49 ft/s 0.408*Q/DPID2^2
1-‐
(1/(2*n+1))*(T0/(T0+K*(((3*n+1)*0.002228*Q)/(n*PI()*((DPID1/1
Cc1 0,7595 unitless 2)/2)^3))^n))
1-‐
(1/(2*n+1))*(T0/(T0+K*(((3*n+1)*0.002228*Q)/(n*PI()*((DPID2/1
Cc2 0,8570 unitless 2)/2)^3))^n))
(2*(3*n+1)/n)*(7.48*d*DPV1^(2-‐
n)*(0.5*(DPID1/12))^n)/((T0*((DPID1/12)*0.5/DPV1)^n)+K*((3*n+
ReDP1 3107 unitless 1)/(n*Cc1))^n)
(2*(3*n+1)/n)*(7.48*d*DPV2^(2-‐
n)*(0.5*(DPID2/12))^n)/((T0*((DPID2/12)*0.5/DPV2)^n)+K*((3*n+
ReDP2 16014 unitless 1)/(n*Cc2))^n)
y 0,073 unitless (log(n)+3.93)/50
z 0,289 unitless (1.75-‐log(n))/7
NReDPCr 2584 unitless ((4*(3*n+1))/(n*y))^(1/(1-‐z))
(4*K)/(14400*(DPID1/12))*((T0/K)+(((3*n+1)/(n*Cc1))*((8*Q*0.00
DPFricLam1 0,026 psi/ft 2228)/(Pi()*(DPID1/12)^2)))^n)
(4*K)/(14400*(DPID2/12))*((T0/K)+(((3*n+1)/(n*Cc2))*((8*Q*0.00
DPFricLam2 0,057 psi/ft 2228)/(Pi()*(DPID2/12)^2)))^n)
fDPTurb1 0,00775 unitless y*(ReDP1*Cc1)^(-‐z)
fDPTurb2 0,005 unitless y*(ReDP2*Cc2)^(-‐z)
DPFricTurb1 0,132 psi/ft (fDPTurb1*d*7.48*(0.002228*Q)^2)/(144*Pi()^2*(DPID1/12)^5)
DPFricTurb2 1,154 psi/ft (fDPTurb2*d*7.48*(0.002228*Q)^2)/(144*Pi()^2*(DPID2/12)^5)
DPL1 3960 m (AnnCsgL1+AnnCsgL2-‐DPL2)+(ROP*TIME)
IF(ReDP1<NReDPCr,(DPFricLam1*WashoutDepth/0.3048),(DPFricT
DPFricLamTurb1 650 psi urb1*WashoutDepth/0.3048))
VelAnn1 4,51 ft/s (0.408*Q)/(AnnCsgOD1^2-‐AnnCsgID1^2)
1-‐(1/(1+n))*T0/(T0+K*((2*(2*n+1)/(n*((AnnCsgOD1*0.5/12)-‐
(AnnCsgID1*0.5/12))))*((0.002228*Q)/(Pi()*((AnnCsgOD1*0.5/12)
Ca1 0,649 unitless ^2-‐(AnnCsgID1*0.5/12)^2))))^n)
1-‐(1/(1+n))*T0/(T0+K*((2*(2*n+1)/(n*((AnnCsgOD2*0.5/12)-‐
1-‐(1/(1+n))*T0/(T0+K*((2*(2*n+1)/(n*((AnnCsgOD3*0.5/12)-‐
(4*(2*n+1)/n)*(7.48*dout*VelAnn1^(2-‐n)*(0.5*((AnnCsgOD1-‐
AnnCsgID1)/12))^n)/((T0*(((AnnCsgOD1-‐
ReAnn1 756 unitless AnnCsgID1)/12)*0.5/VelAnn1)^n)+K*(2*(2*n+1)/(n*Ca1))^n)
AnnCsgID3)/12))^n)/((R0*(((AnnCsgOD3-‐
NReAnnCr 4959 unitless ((8*(2*n+1))/(n*y))^(1/(1-‐z))

93

(4*K)/(14400*((AnnCsgOD1-‐
AnnCsgID1)/12))*((T0/K)+(((16*(2*n+1))/(n*Ca1*((AnnCsgOD1-‐
AnnCsgID1)/12)))*((Q*0.002228)/(Pi()*((AnnCsgOD1/12)^2-‐
AnnFricLam1 0,040 psi/ft (AnnCsgID1/12)^2))))^n)
(4*K)/(14400*((AnnCsgOD2-‐
(4*K)/(14400*((AnnCsgOD3-‐
fAnnTurb1 0,012 unitless y*(Ca1*ReAnn1)^-‐z
(fAnnTurb1*7.48*dout*(0.002228*Q)^2)/(144*Pi()^2*((AnnCsgO
AnnFricTurb1 0,037 psi/ft D1-‐AnnCsgID1)/12)*((AnnCsgOD1/12)^2-‐(AnnCsgID1/12)^2)^2)
AnnCsgL3 160,00 m ROP*TIME
IF(ReAnn1<NReAnnCr,(AnnFricLam1*WashoutDepth/0.3048),(Ann
AnnFricLamTurb1 195 psi FricTurb1*WashoutDepth/0.3048))
AnnFricLamTurb2 0 psi
IF(Washoutdepth>0,0,IF(ReAnn3<NReAnnCr,(AnnFricLam3*AnnCs
AnnFricLamTurb3 0 psi gL3/0.3048),(AnnFricTurb3*AnnCsgL3/0.3048)))
Output -‐ Below WashoutDepth

QoutLoss 528,3 lpm 0.26417*(Qin-‐Qloss)
DPV1Loss 11,8 ft/s 0.408*QoutLoss/DPID1^2
DPV2Loss 34,5 ft/s 0.408*QoutLoss/DPID2^2
1-‐
(1/(2*n+1))*(T0/(T0+K*(((3*n+1)*0.002228*QoutLoss)/(n*PI()*((D
Cc1Loss 0,7595 unitless PID1/12)/2)^3))^n))
1-‐
(1/(2*n+1))*(T0/(T0+K*(((3*n+1)*0.002228*QoutLoss)/(n*PI()*((D
Cc2Loss 0,8570 unitless PID2/12)/2)^3))^n))
(2*(3*n+1)/n)*(7.48*dout*DPV1Loss^(2-‐
n)*(0.5*(DPID1/12))^n)/((T0*((DPID1/12)*0.5/DPV1Loss)^n)+K*((3
ReDP1Loss 3107 unitless *n+1)/(n*Cc1Loss))^n)
(2*(3*n+1)/n)*(7.48*dout*DPV2Loss^(2-‐
n)*(0.5*(DPID2/12))^n)/((T0*((DPID2/12)*0.5/DPV2Loss)^n)+K*((3
ReDP2Loss 16014 unitless *n+1)/(n*Cc2Loss))^n)
(4*K)/(14400*(DPID1/12))*((T0/K)+(((3*n+1)/(n*Cc1Loss))*((8*Qo
DPFricLam1Loss 0,0263 psi/ft utLoss*0.002228)/(Pi()*(DPID1/12)^2)))^n)
(4*K)/(14400*(DPID2/12))*((T0/K)+(((3*n+1)/(n*Cc2Loss))*((8*Qo
DPFricLam2Loss 0,0570 psi/ft utLoss*0.002228)/(Pi()*(DPID2/12)^2)))^n)
fDPTurb1Loss 0,0077 unitless y*(ReDP1Loss*Cc1)^(-‐z)
fDPTurb2Loss 0,0047 unitless y*(ReDP2Loss*Cc2)^(-‐z)

94

(fDPTurb1Loss*dout*7.48*(0.002228*QoutLoss)^2)/(144*Pi()^2*(
DPFricTurb1Loss 0,1320 psi/ft DPID1/12)^5)
(fDPTurb2Loss*dout*7.48*(0.002228*QoutLoss)^2)/(144*Pi()^2*(
DPFricTurb2Loss 1,1538 psi/ft DPID2/12)^5)
IF(ReDP1Loss<NReDPCr,(DPFricLam1Loss*(DPL1-‐
WashoutDepth)/0.3048),(DPFricTurb1Loss*(DPL1-‐
DPFricLamTurb1Loss 1065,6 psi WashoutDepth)/0.3048))
IF(ReDP2Loss<NReDPCr,(DPFricLam2Loss*DPL2/0.3048),(DPFricTu
DPFricLamTurb2Loss 757,1 psi rb2*DPL2/0.3048))
VelAnn1Loss 4,5134 ft/s (0.408*QoutLoss)/(AnnCsgOD1^2-‐AnnCsgID1^2)
1-‐(1/(1+n))*T0/(T0+K*((2*(2*n+1)/(n*((AnnCsgOD1*0.5/12)-‐
(AnnCsgID1*0.5/12))))*((0.002228*QoutLoss)/(Pi()*((AnnCsgOD1*
Ca1Loss 0,6491 unitless 0.5/12)^2-‐(AnnCsgID1*0.5/12)^2))))^n)
1-‐(1/(1+n))*T0/(T0+K*((2*(2*n+1)/(n*((AnnCsgOD2*0.5/12)-‐
1-‐(1/(1+n))*T0/(T0+K*((2*(2*n+1)/(n*((AnnCsgOD3*0.5/12)-‐
(4*(2*n+1)/n)*(7.48*dout*VelAnn1Loss^(2-‐n)*(0.5*((AnnCsgOD1-‐
AnnCsgID1)/12)*0.5/VelAnn1Loss)^n)+K*(2*(2*n+1)/(n*Ca1Loss))
ReAnn1Loss 755,6 unitless ^n)
AnnCsgID3)/12))^n)/((R0*(((AnnCsgOD3-‐
(4*K)/(14400*((AnnCsgOD1-‐
AnnCsgID1)/12))*((T0/K)+(((16*(2*n+1))/(n*Ca1Loss*((AnnCsgOD
1-‐
AnnCsgID1)/12)))*((QoutLoss*0.002228)/(Pi()*((AnnCsgOD1/12)^2
AnnFricLam1Loss 0,0396 psi/ft -‐(AnnCsgID1/12)^2))))^n)
(4*K)/(14400*((AnnCsgOD2-‐
2-‐
(4*K)/(14400*((AnnCsgOD3-‐
3-‐
fAnnTurb1Loss 0,0122 unitless y*(Ca1Loss*ReAnn1Loss)^-‐z

95

(fAnnTurb1Loss*7.48*dout*(0.002228*QoutLoss)^2)/(144*Pi()^2*
((AnnCsgOD1-‐AnnCsgID1)/12)*((AnnCsgOD1/12)^2-‐
AnnFricTurb1Loss 0,0369 psi/ft (AnnCsgID1/12)^2)^2)
IF(QoutLoss>0,IF(ReAnn1Loss<NReAnnCr,(AnnFricLam1Loss*(Ann
CsgL1-‐WashoutDepth)/0.3048),(AnnFricTurb1Loss*(AnnCsgL1-‐
AnnFricLamTurb1Loss 259,7 psi WashouthDepth)/0.3048)),0)
IF(QoutLoss>0,IF(ReAnn2Loss<NReAnnCr,(AnnFricLam2Loss*AnnC
AnnFricLamTurb2Loss 65,8 psi sgL2/0.3048),(AnnFricTurb2*AnnCsgL2/0.3048)),0)
AnnFricLamTurb3Loss 0,0 psi
Output -‐ Final

WellDepth 4160 m InitialWellDepth+(ROP*TIME)
FlowAChoke 0 lpm Qin+BackPPump
CumLAnn 4160 m AnnCsgL1+AnnCsgL2+AnnCsgL3
CumLDP 4160 m DPL1+DPL2
BitNozPloss 54,62 bar (din*(QoutLoss/0.26417)^2/2959.41/0.95^2/BitNozTFA^2)/100
(AnnFricLamTurb1Loss+AnnFricLamTurb2Loss+AnnFricLamTurb1)/1
SumAnnFric 35,87 bar 4.5038
SumEqFric 29,60 bar IF(Qin>0,((Motor+MWD+SurfEq)),0)
IF(QoutLoss>0,(DPFricLamTurb1+DPFricLamTurb1Loss+DPFricLamT
SumDPFric 170 bar urb2Loss),DPFricLamTurb1)/14.5038
ChokePress 0 bar (din*FlowAChoke^2/(2959.41*0.95^2)/ChokeArea^2)/100
PumpPress 294 bar SumAnnFric+BitNozPloss+SumEqFric+SumDPFric+ChokePress
(((dout/8.33)*CumLAnn*0.0981)+SumAnnFric+ChokePress)/(CumL
ECDCalc 1,697 sg Ann*0.0981)
BHP 692.5 bar ECDCalc*0.0981*CumLAnn
IF(ReAnn1<NReAnnCr,(AnnFricLam1*Sensor1Depth)/(14.50378*0.3
048),(AnnFricTurb1*Sensor1Depth/(14.50378*0.3048)))+((dout/8.3
Sensor1Target 166,8 bar 3)*0.0981*Sensor1Depth)
Sensor1Total 166,8 bar 3)*0.0981*Sensor1Depth)
Sensor2Total 333,5 bar Sensor2+Sensor2Loss+((dout/8.33)*0.0981*Sensor2Depth)
Sensor3Total 500,3 bar Sensor3+Sensor3Loss+((dout/8.33)*0.0981*Sensor3Depth)
96
2.3. Robertson-Stiff Frictional Model
Static Input: Variable Input:

Parameter: Value: Unit: Parameter: Value: Unit:
BackPPump 0,00 lpm Qin 2000 lpm
VcutTransp 1,02 m/s din 1,6 sg
ChokeOD 3,00 inch ROP 40,0 m/h
ChokeOpen 1,00 0-‐1 TIME 4 h
AnnCsgL1 3500,00 m
AnnCsgL2 500,00 m
AnnCsgID1 5,00 inch
AnnCsgID2 5,00 inch
AnnCsgID3 6,50 inch
DPL2 200,00 m
DPID1 4,27 inch
DPID2 2,50 inch
BitNoz 6,00
BitNozID 12,00
Motor 13,80 bar
MWD 13,80 bar
SurfEq 2,00 bar
dcuttings 2,50 sg
CutCons 0,01 (0-‐1) %
InitialWellDepth 4000 m
Output
d 13,33 ppg 8.33*din
Q 528,34 gpm 0.26417*Qin
C 52,01
B 0,3130
A 3054,246
dout 13,40 ppg ((1-‐CutCons)*din+dcuttings*CutCons)*8.33
DPV1 11,823 ft/s 0.408*Q/(DPID1^2)
DPV2 34,490 ft/s 0.408*Q/DPID2^2
ReDP1 8251,17 unitless 89100*DPV1^(2-‐B)*d/A*(0.0416*DPID1/(3+1/B))^B
ReDP2 42475 unitless 89100*DPV2^(2-‐B)*d/A*(0.0416*DPID2/(3+1/B))^B
NReLamCr 3041 unitless 3470-‐1370*B
NReDPTurbCr 3841 4270-‐1370*B
8.33/10000*2^(2+B)*(A/478.789033)*((3+1/B)*((0.2*60*DPV1+C/6*
DPFricLam1 0,03524 psi/ft DPID1)/DPID1^(1/B+1)))^B
8.33/10000*2^(2+B)*(A/478.789033)*((3+1/B)*((0.2*60*DPV2+C/6*
DPFricLam2 0,09402 psi/ft DPID2)/DPID2^(1/B+1)))^B
97

a1 0,06851 (LOG(B)+3.93)/50
b2 0,32206 (1.75-‐LOG(B))/7
fDPTurb1 0,00375 unitless a1/ReDP1^b2
fDPTurb2 0,00221 unitless a1/ReDP2^b2
DPFricTurb1 0,0634466 psi/ft (fDPTurb1*d*DPV1^2)/(25.81*DPID1)
DPFricTurb2 0,5440832 psi/ft (fDPTurb2*d*DPV2^2)/(25.81*DPID2)
DPL1 3960 m (AnnCsgL1+AnnCsgL2-‐DPL2)+(ROP*TIME)
IF(ReDP1<NReLamCr,(DPFricLam1*DPL1/0.3048),IF(ReDP1>NReDPTur
DPFricLamTurb1 824 psi bCr,(DPFricTurb1*DPL1/0.3048),0))
IF(ReDP2<NReLamCr,(DPFricLam2*DPL2/0.3048),IF(ReDP2>NReDPTur
DPFricLamTurb2 357 psi bCr,(DPFricTurb2*DPL2/0.3048),0))
109000*VelAnn1^(2-‐B)*dout/A*(0.0208*(AnnCsgOD1-‐
ReAnn1 1602,26 unitless AnnCsgID1)/(2+1/B))^B
8.33/10000*4^(1+B)*(A/478.789033)*((2+1/B)*((0.2*60*VelAnn1+C/
AnnFricLam1 0,040871 psi/ft 8*(AnnCsgOD1-‐AnnCsgID1))/(AnnCsgOD1-‐AnnCsgID1)^(1/B+1)))^B
8.33/10000*4^(1+B)*(A/478.789033)*((2+1/B)*((0.2*60*VelAnn2+C/
8.33/10000*4^(1+B)*(A/478.789033)*((2+1/B)*((0.2*60*VelAnn3+C/
NReAnnTurbCr 3841,14 4270-‐1370*B
fAnnTurb1 0,00636 unitless a1/ReAnn1^b2
AnnFricTurb1 0,01906 psi/ft (fAnnTurb1*dout*VelAnn1^2)/(25.81*(AnnCsgOD1-‐AnnCsgID1))
AnnCsgL3 160,00 m ROP*TIME
xint3 -‐0,128 (ReAnn3-‐NReLamCr)/(NReDPTurbCr-‐NReLamCr)
IF(ReAnn1<=NReLamCr,(AnnFricLam1*AnnCsgL1/0.3048),IF(ReAnn1>=
AnnFricLamTurb1 469 psi NReAnnTurbCr,(AnnFricTurb1*AnnCsgL1/0.3048),0))
AnnFricLamTurb2 68 psi NReAnnTurbCr,(AnnFricTurb2*AnnCsgL2/0.3048),0))
NReAnnTurbCr,(AnnFricTurb3*AnnCsgL3/0.3048),(((1-‐
xint3)*(AnnFricLam3*AnnCsgL3/0.3048))+((AnnFricTurb3*AnnCsgL3/0
AnnFricLamTurb3 49 psi .3048)*xint3))))
WellDepth 4160,00 m InitialWellDepth+(ROP*TIME)
FlowAChoke 0 lpm Qin+BackPPump
CumLAnn 4160,00 m AnnCsgL1+AnnCsgL2+AnnCsgL3
CumLDP 4160,00 m DPL1+DPL2

98

BitNozPloss 54,62 bar (din*Qin^2/2959.41/0.95^2/BitNozTFA^2)/100
IF(Qin>0,((AnnFricLamTurb1+AnnFricLamTurb2+AnnFricLamTurb3)/14
SumAnnFric 40,42 bar .5038),0)
SumEqFric 29,60 bar IF(Qin>0,((Motor+MWD+SurfEq)),0)
SumDPFric 81,45 bar (DPFricLamTurb1+DPFricLamTurb2)/14.5038
ChokePress 0 bar ((dout/8.33)*FlowAChoke^2/(2959.41*0.95^2)/ChokeArea^2)/100
PumpPress 209,81 bar SumAnnFric+BitNozPloss+SumEqFric+SumDPFric+ChokePress
(((dout/8.33)*CumLAnn*0.0981)+SumAnnFric+ChokePress)/(CumLAn
ECDCalc 1,7077 sg n*0.0981)
BHP 696.9 bar ECDCalc*0.0981*CumLAnn
99
3. Log input data
This is an example of what a manipulated log that is implemented into Discovery Web looks
like. The following example is taken from the connection scenario.
TIME,BLOCKCOMP,MWIN,FLOWIN,FRAC_EQMD,POR_EQMD,CumTime
yyyy-MM-dd"T"HH:mm:ss.fffzzz,m,sg,L/min,sg,sg,h
2014-02-28T08:02:00.000+01:00,28.98,1.60,2000.00,1.75,1.50,0
2014-02-28T08:03:00.000+01:00,28.13,1.60,2000.00,1.75,1.50,0.01667
2014-02-28T08:04:00.000+01:00,26.88,1.60,2000.00,1.75,1.50,0.03333
2014-02-28T08:05:00.000+01:00,26.22,1.60,2000.00,1.75,1.50,0.04999
2014-02-28T08:06:00.000+01:00,25.25,1.60,2000.00,1.75,1.50,0.06665
2014-02-28T08:07:00.000+01:00,24.62,1.60,2000.00,1.75,1.50,0.08331
2014-02-28T08:08:00.000+01:00,23.83,1.60,2000.00,1.75,1.50,0.09997
2014-02-28T08:09:00.000+01:00,23.48,1.60,2000.00,1.75,1.50,0.11663
2014-02-28T08:10:00.000+01:00,22.74,1.60,2000.00,1.75,1.50,0.13329
2014-02-28T08:11:00.000+01:00,22.19,1.60,2000.00,1.75,1.50,0.14995
2014-02-28T08:12:00.000+01:00,21.24,1.60,2000.00,1.75,1.50,0.16661
2014-02-28T08:13:00.000+01:00,20.19,1.60,2000.00,1.75,1.50,0.18327
2014-02-28T08:14:00.000+01:00,18.92,1.60,2000.00,1.75,1.50,0.19993
2014-02-28T08:15:00.000+01:00,18.10,1.60,2000.00,1.75,1.50,0.21659
2014-02-28T08:16:00.000+01:00,17.26,1.60,2000.00,1.75,1.50,0.23325
2014-02-28T08:17:00.000+01:00,16.05,1.60,2000.00,1.75,1.50,0.24991
2014-02-28T08:18:00.000+01:00,15.18,1.60,2000.00,1.75,1.50,0.26657
2014-02-28T08:19:00.000+01:00,14.31,1.60,2000.00,1.75,1.50,0.28323
2014-02-28T08:20:00.000+01:00,12.86,1.60,2000.00,1.75,1.50,0.29989
2014-02-28T08:21:00.000+01:00,12.22,1.60,2000.00,1.75,1.50,0.31655
2014-02-28T08:22:00.000+01:00,11.57,1.60,2000.00,1.75,1.50,0.33321
2014-02-28T08:23:00.000+01:00,11.12,1.60,2000.00,1.75,1.50,0.34987
2014-02-28T08:24:00.000+01:00,10.36,1.60,2000.00,1.75,1.50,0.36653
2014-02-28T08:25:00.000+01:00,9.27,1.60,2000.00,1.75,1.50,0.38319
2014-02-28T08:26:00.000+01:00,7.76,1.60,2000.00,1.75,1.50,0.39985
2014-02-28T08:27:00.000+01:00,6.58,1.60,2000.00,1.75,1.50,0.41651
2014-02-28T08:28:00.000+01:00,5.27,1.60,2000.00,1.75,1.50,0.43317
2014-02-28T08:29:00.000+01:00,4.03,1.60,2000.00,1.75,1.50,0.44983
2014-02-28T08:30:00.000+01:00,2.81,1.60,2000.00,1.75,1.50,0.46649
2014-02-28T08:31:00.000+01:00,1.87,1.60,2000.00,1.75,1.50,0.48315
2014-02-28T08:32:00.000+01:00,1.47,1.60,2000.00,1.75,1.50,0.49981
2014-02-28T08:33:00.000+01:00,0.85,1.60,0.00,1.75,1.50,0.51647
2014-02-28T08:34:00.000+01:00,4.44,1.60,0.00,1.75,1.50,0.53313
2014-02-28T08:35:00.000+01:00,4.21,1.60,0.00,1.75,1.50,0.54979
2014-02-28T08:36:00.000+01:00,8.46,1.60,0.00,1.75,1.50,0.56645
2014-02-28T08:37:00.000+01:00,31.90,1.60,0.00,1.75,1.50,0.58311
2014-02-28T08:38:00.000+01:00,32.46,1.60,833.33,1.75,1.50,0.59977
2014-02-28T08:39:00.000+01:00,31.12,1.60,1000.00,1.75,1.50,0.61643
2014-02-28T08:40:00.000+01:00,31.12,1.60,2000.00,1.75,1.50,0.63309
2014-02-28T08:41:00.000+01:00,31.12,1.60,2000.00,1.75,1.50,0.64975
100

Strand Erling PDF

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Strand Erling PDF

Uploaded by

Copyright:

Available Formats

Faculty of Science and Technology

Study program/ Specialization:

Erling Strand …………………………………………

Hydraulic Calculations Using Discovery Web for Visualization

• Drilling in methane hydrate:

• Drilling extended reach horizontal wells:

Ø Provide an overview of the following:

Ø Describe the functionality of Discovery Web.

Ø Implement different rheology models in Discovery Web for hydraulic calculations.

− Plastic viscosity (PV)

− Yield limit (YP)

− Apparent viscosity (AV) and Funnel viscosity (“Marsh Funnel”)

where τ is shear stress, µ is viscosity and γ is shear rate.

2.1.2 Non-Newtonian fluids.

RPM Reading [°]

𝜇! 𝑐𝑃 = 𝜃!"" − 𝜃!"" (4)

𝜏! 𝑙𝑏𝑠/100𝑓𝑡 ! = 𝜃!"" − 𝜇! = 2 ∙ 𝜃!"" − 𝜃!"" (5)

𝜇! = 54.50 − 43.50 = 11 𝑐𝑃

2.1.2.3 Herschel-Bulkley model

𝜏 ∗! − 𝜏!!" ∙ 𝜏!"# (8)

Figure 3 - Herschel-Bulkley fluid rheogram.

𝛾 [𝑠𝑒𝑐 !! ] 𝜏 [𝑙𝑏𝑓/100 𝑓𝑡 ! ]

𝜏 = 𝐴(𝛾 + 𝐶)! (10)

log 𝜏 = log 𝐴 + 𝐵 log (𝛾 + 𝐶) (11)

𝛾!"# ∙ 𝛾!"# − 𝛾 ∗! (13)

𝜏∗ = 𝜏!"# ∙ 𝜏!"# (14)

𝜏 ∗ = 35.66 𝑙𝑏𝑓/100 𝑓𝑡 !

Figure 4 - Robertson-Stiff fluid rheogram.

From Figure 4 the Robertson- Stiff parameters are as follows:

𝐴 = 6.379 𝑙𝑏𝑓. 𝑠𝑒𝑐 ! /100 𝑓𝑡 !

𝛾 [𝑠𝑒𝑐 !! ] 𝜏 [𝑙𝑏𝑓/100 𝑓𝑡 ! ]

2.1.3.1 Comparison of the rheology prediction

Comparison of shear stress and shear rate data

Figure 5 - Comparison of rheology prediction.

2.1.3.2 Error calculation of rheology prediction

Rheology Model: Error %

Bingham-Plastic Model 1,382

Herschel-Bulkley Model 0.344

Robertson-Stiff Model 0,452

Table 4 - Error analysis for rheology prediction.

2.2.1 Pressure losses

− Drill pipe frictional pressure drop (∆𝑃! )

− Downhole mud motor pressure drop (∆𝑃!" )

− Nozzle pressure drop (∆𝑃!" )

− Drill bit pressure drop (∆𝑃!" )

Figure 6 - Pressure drops during circulation.

𝑃! = ∆𝑃! + ∆𝑃!" + ∆𝑃!" + ∆𝑃!" + ∆𝑃!" = ∆𝑃!" (15)

2.2.2 Bit pressure losses

2.2.3 Cuttings concentration

A transient cuttings-transport model makes it possible to better predict downhole conditions

𝑡𝑖𝑚𝑒 − 𝑡𝑖𝑚𝑒𝑛𝑒𝑤𝑟𝑜𝑝 (20)

The following limits needs to be established for transient flow determination:

𝑥𝑖𝑛𝑡 = 0 if time = timenewrop

%C!"#$%&'$! = 1 − 𝑥𝑖𝑛𝑡 ∙ %𝐶!"# + 𝑥𝑖𝑛𝑡 ∙ %𝐶!"# (21)

2.3.1 Bingham frictional model

2.3.1.1 Frictional pressure loss calculation inside the drillstring

Flow regime determination:

If 𝑉!"# < 𝑉!" the flow is laminar.

Frictional pressure loss calculation inside the drillstring:

𝑑𝑝 𝑑 !.! 𝑄!.! 𝜇! !.! (25)

where 𝑉! is critical fluid velocity, 𝐷! is annulus outside diameter, 𝐷! is annulus inside

Flow regime determination: