Ucalgary 2012 Islam Arshad

University of Calgary
PRISM: University of Calgary's Digital Repository
Graduate Studies The Vault: Electronic Theses and Dissertations
2012-10-03
Geomechanical Productivity and Injectivity Modeling

of Multifractured Horizontal Wells
Islam, Arshad
Islam, A. (2012). Geomechanical Productivity and Injectivity Modeling of Multifractured Horizontal

Wells (Unpublished master's thesis). University of Calgary, Calgary, AB.
doi:10.11575/PRISM/26287
http://hdl.handle.net/11023/275
master thesis
University of Calgary graduate students retain copyright ownership and moral rights for their
thesis. You may use this material in any way that is permitted by the Copyright Act or through
licensing that has been assigned to the document. For uses that are not allowable under
copyright legislation or licensing, you are required to seek permission.
Downloaded from PRISM: https://prism.ucalgary.ca
UNIVERSITY OF CALGARY
Geomechanical Productivity and Injectivity Modeling of Multifractured Horizontal Wells
by
Arshad Islam
A THESIS
SUBMITTED TO THE FACULTY OF GRADUATE STUDIES
IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE
DEGREE OF MASTER OF SCIENCE
DEPARTMENT OF CHEMICAL AND PETROLEUM ENGINEERING
CALGARY, ALBERTA
SEPTEMBER, 2012
© Arshad Islam 2012

Abstract
The advances in hydraulic fracturing technology and horizontal well completions have
led in recent years to rapid rise in exploitation and development of tight gas and shale
plays all over the world, and particularly in North America. The popularity of new field
technology has in fact raised many new questions. In particular, for forecasting the
productivity and EUR of multifractured horizontal wells, it is not clear if conventional
reservoir simulation concepts can be adapted for modeling or if extra physics must be
included to obtain realistic solutions.
This work presents various methods to model multifractured horizontal wells in tight gas
sands using a conventional reservoir simulator coupled with geomechanics. Two actual
wells in the same formation but fractured with different techniques (i.e., X-link gelled
water fracs and un-gelled water (slick water) fracs) are studied. Detailed investigation of
the role of fracture conductivity, effects of initial permeability level, net pay thickness,
assumed size of the stimulated reservoir volume (SRV), pressure or stress dependent
permeability of the SRV and formation, and virgin reservoir were carried out by history
matching the rate and cumulative production. It was established that i) history match is
not possible without use of stress or pressure dependent permeability and ii) permeability
dependence on pressure inside stimulated reservoir volume must be larger than in the rest
of the formation. It was also observed that the standard method for using the same
geomechanical data both in uncoupled reservoir and coupled geomechanical model will
give incorrect results in terms of production.
A new method based on uniaxial deformation theory is proposed to more accurately
ii
approximate the geomechanical effects in conventional reservoir simulators without
running a fully coupled simulator. The production results from the uncoupled reservoir
modeling using the new method for correcting the permeability data for poroelastic
effects were remarkably similar to rigorously coupled geomechanical modeling. This
work will be of importance to engineers in analyzing and forecasting production
performance of multifractured horizontal completions using numerical models. It will
allow engineers to use uncoupled (conventional) reservoir modeling as a practical
approximation of more complex coupled geomechanical models.
Same wells were used to model the injection process. History matching of the field
bottomhole injection pressure using uncoupled and fully coupled geomechanical models
showed the importance of including the geomechanics. Sensitivity study of several
history matching parameters such as fracture permeability reduction factor, limiting
length of fracture propagation and Biot’s constant was performed and its effects on
injection pressure were discussed. Possibility of shear failure in the SRV during injection
was also studied, using the Mohr-Coulomb failure criterion. No shear failure was
detected when intact rock shear strength parameters were used, but significant shear
regions were generated when failure envelope represented fractured or weakened rock.
The main contributions of this work are i) better understanding of the role of the
geomechanical effects in both production and injection modeling, ii) the demonstration of
the need for coupled geomechanical modeling in injection, and iii) presentation of
techniques for approximating these effects in uncoupled reservoir simulation. In addition,
valuable insights were gained into the mechanics of fracturing and reservoir behavior
during production.
iii
Acknowledgements
I wish to express my sincere appreciations from the bottom of my heart to all those who
supported me to complete this thesis. My deepest appreciation goes to my supervisor, Dr.
Antonin Settari for his continuous support, patience, encouragement and insightful
guidance throughout the whole study and research. I would like to thank him for his trust
in me and giving me the opportunity to work with him. I have enjoyed every single
moment of working with him and years of studies and research under his guidance.
I would like to thank Dr. Roberto Aguilera, Dr. Christopher Clarkson, and Mohammad
Ali Bagheri for being the members of the supervisory and examination committee of
MSc. Program. I am grateful to management of Apache Petroleum Corporation for
providing us field and research data. I also wish to thank Mr. Vikram Sen for helping us
and Taurus Reservoir Solutions Ltd. for the generous donation of the TRS® and GeoSim®
software and University of Calgary, the Department of Chemical and Petroleum
Engineering and Graduate Studies at the University of Calgary for providing excellent
research environment and facilities. Last but not the least my colleagues in my research
group and Ayaz Mehmood, Faisal Iqbal, Faraz Rasheed, Omair Shafiq, Syed Saad Alam,
and Wajih Naeem – Thank you all for your support, advice, friendship and making my
stay memorable and enjoyable.
Finally I would like to express my gratitude to my family for their endless understanding,
support, encouragement and love during the past and future tough and easy times. I am
really thankful to my Father who taught me how to live and how to be strong on the
curvy road of life and my passionate loving Mother.
iv
Dedication
To my Father, Muhammad Islam and my Mother, Rabia Islam,
for their years of love, affection, support and encouragement.
v
Table of Contents
Abstract ............................................................................................................................... ii
Acknowledgements ............................................................................................................ iv
Dedication ............................................................................................................................v
Table of Contents ............................................................................................................... vi
List of Tables .......................................................................................................................x
List of Figures and Illustrations ........................................................................................ xii
List of Symbols, Abbreviations and Nomenclature ......................................................... xvi
CHAPTER ONE: INTRODUCTION ..................................................................................1

1.1 BACKGROUND .......................................................................................................1
1.2 RESEARCH OBJECTIVES ......................................................................................3
1.2.1 Production Modeling of Multifractured Horizontal Wells ................................3
1.2.2 Injection Modeling of Multifractured Horizontal Wells ...................................4
1.3 THESIS ORGANIZATION ......................................................................................5
CHAPTER TWO: LITERATURE REVIEW ......................................................................8

2.1 PRESSURE / STRESS DEPENDENT ROCK PROPERTIES .................................9
2.1.1 Experimental Work ...........................................................................................9
2.1.2 Mathematical Correlations ..............................................................................13
2.2 EFFECTS OF PRESSURE / STRESS DEPENDENT PERMEABILITY ON
GAS PRODUCTION .............................................................................................20
2.3 CONVENTIONAL FRACTURE MODELS ...........................................................23
2.3.1 Two – Dimensional (2-D) Fracture Models ....................................................24
2.3.1.1 PKN Fracture Model ..............................................................................24
2.3.1.2 KGD Fracture Model .............................................................................26
2.3.2 Three – Dimensional (3-D) Fracture Models ..................................................27
2.4 UNCONVENTIONAL FRACTURE MODELS .....................................................28
2.5 SUMMARY .............................................................................................................31
vi
CHAPTER THREE: SUMMARY OF THE FIELD DATA .............................................33
3.1 BASIC DATA..........................................................................................................33
3.2 STIMULATION DATA ..........................................................................................36
3.3 PRODUCTION DATA............................................................................................39
3.4 GEOMECHANICAL DATA ..................................................................................41
CHAPTER FOUR: MODELING OF STATIC FRACTURES .........................................45

4.1 REPRESENTATION OF FRACTURES IN RESERVOIR SIMULATOR ............47
4.1.1 Transmissibility Modifications ........................................................................47
4.1.2 Element of Symmetry Model ..........................................................................50
4.1.3 Grid Refinement ..............................................................................................51
4.2 FRACTURE DENSITY ..........................................................................................51
4.3 FRACTURES PARAMETERS ESTIMATION .....................................................53
4.4 SUMMARY .............................................................................................................57
CHAPTER FIVE: HISTORY MATCHING USING PSEUDO CONTINUUM

APPROACH – UNCOUPLED MODELS................................................................59
5.1 ROLE OF FRACTURE CONDUCTIVITY ............................................................61
5.1.1 Baseline Propped Fracture Conductivity .........................................................62
5.1.2 Realistic Propped Fracture Conductivity.........................................................63
5.1.3 Infinite Fracture Conductivity – An Analytical Approach ..............................67
5.2 EFFECT OF NET PAY THICKNESS (NPT) .........................................................71
5.3 EFFECTS OF STIMULATED RESERVOIR VOLUME (SRV) ...........................73
5.4 HISTORY MATCHING USING MANUAL CHANAGE OF RESERVOIR
PERMEABILITY WITH TIME ............................................................................81
5.5 SUMMARY .............................................................................................................84
CHAPTER SIX: HISTORY MATCHING AND MODELING OF

GEOMECHANICAL EFFECTS USING PRESSURE / STRESS DEPENDENT
PERMEABILITY .....................................................................................................85
6.1 DISCUSSION ON NEW INFORMATION ............................................................88
6.2 PRESSURE / STRESS DEPENDENT PERMEABILITY CORRELATION ........91
vii
6.3 HISTORY MATCHING USING PRESSURE DEPENDENT
PERMEABILITY ..................................................................................................92
6.3.1 Effects of Shape of Pressure Dependent Permeability Curves (Stress
Factor, S) ..........................................................................................................92
6.3.2 Effects of Stimulated Reservoir Volume (SRV) .............................................96
6.3.3 History Matching Using Pressure Dependent Permeability outside SRV .......99
6.3.4 Effect of Native Reservoir Permeability – Well B ........................................103
6.3.5 History Matching Using Pressure Dependent Permeability both Inside and
Outside SRV ..................................................................................................106
6.4 APPROXIMATION OF GEOMECHANICAL EFFECTS IN
CONVENTIONAL RESERVOIR SIMULATOR USING UNIAXIAL
DEFORMATION THEORY ...............................................................................110
PERMEABILITY BOTH INSIDE AND OUTSIDE SRV – COUPLED
PRODUCTION GEOMECHANICAL MODELS ..............................................116
PERMEABILITY BOTH INSIDE AND OUTSIDE SRV – COUPLED
PRODUCTION GEOMECHANICAL MODELS – VARIABLE MEAN
TOTAL STRESS CASES ....................................................................................120
6.7 SUMMARY ...........................................................................................................125
CHAPTER SEVEN: HISTORY MATCHING OF FIELD INJECTION PRESSURE

USING UNCOUPLED AND FULLY COUPLED GEOMECHANICAL
INJECTION MODELS ...........................................................................................127
7.1 MODELING OF FRACTURE PROPAGATION .................................................128
7.2 THEORY OF SIMPLIFIED FRACTURE MODEL .............................................129
7.3 HISTORY MATCHING OF FIELD INJECTION PRESSURE USING
UNCOUPLED INJECTION MODELS ..............................................................131
7.3.1 Effects of Permeability Reduction Factor ( ............................................134
7.3.2 Effects of Limiting Length of Fracture Propagation .....................................138
7.4 HISTORY MATCHING OF FIELD INJECTION PRESSURE USING A
FULLY COUPLED GEOMECHANICAL INJECTION MODEL .....................141
viii
7.4.1 Effects of Limiting Length of Fracture Propagation .....................................144
7.4.2 Effects of Biot’s Constant ( )........................................................................148
7.5 FAILURE PREDICTIONS....................................................................................151
7.5.1 Tensile Failure ...............................................................................................152
7.5.2 Shear Failure ..................................................................................................152
7.5.3 Mohr – Coulomb Shear Failure Criterion .....................................................153
7.6 SUMMARY ...........................................................................................................165
CHAPTER EIGHT: RESULTS AND CONCLUSIONS ................................................166
REFERENCES ................................................................................................................169
APPENDIX A: SELECTED DATA FILES ....................................................................181
ix
List of Tables
Table 3.1 Rock and fluid properties of well A and B (input to reservoir simulator) ........ 36
Table 3.2 Proppant types in well A and B ........................................................................ 39
Table 3.3 Stress and rock properties for geomechanical modeling of well A and B ........ 43
Table 4.1 Baseline permeability of sand used in well A and B ........................................ 54
Table 5.1 Fracture properties of Case 5.1.1 – 2 for well A............................................... 63
Table 5.2 Fracture properties of Case 5.1.1 – 2 for well B ............................................... 63
Table 5.3 Dimensionless fracture conductivity of Cases 5.1.1 - 2 for well A and B........ 69
Table 5.4 Fracture properties of Case 5.2 for well A and B ............................................. 71
Table 5.5 SRVand initial permeability multiplier of Cases 5.3.1 – 6 for well A and B ... 75
Table 5.6 Fracture properties of Case 5.4 for well A and B ............................................. 82
Table 6.1 Fracture properties of Case 5.3.4 and 6.1 for well A ........................................ 88
Table 6.2 Fracture properties of Case 5.3.4 and 6.1 for well B ........................................ 89
Table 6.3 Stress factor (S) of Cases 6.3.1.1 - 4 for well A and B ..................................... 93
Table 6.4 Fracture properties of Case 6.3.2 for well A..................................................... 96
Table 6.5 Fracture properties of Case 6.3.2 for well B ..................................................... 97
Table 6.6 Fracture properties of Case 6.3.3.2 for well A and B ..................................... 101
Table 6.7 Fracture properties of Cases 6.3.4.1 - 2 for well B ......................................... 104
Table 6.8 Fracture properties of Case 6.3.5 both for well A and B ................................ 107
Table 6.9 Fracture properties of Cases 6.6 both for well A and B ................................. 122
Table 7.1 Initial and modified stresses for injection cases – Well A and B ................... 132
Table 7.2 Data for transmissibility calculations in injection cases – Well A and B ....... 133
Table 7.3 Permeability reduction factor of Cases 7.3.1.1-4 both for well A and B........ 135
Table 7.4 Parameters varied in injection Cases 7.3.2.1 – 2 – Well A and B .................. 139
x
Table 7.5 Parameters varied in coupled injection Cases 7.4.1.1 – 4 – Well A ............... 144
Table 7.6 Parameters varied in injection cases 7.4.2.1 – 2 – Well A ............................. 149
Table 7.7 Mohr-Coulomb circles input parameters – Block 1 and 2 – Well A .............. 159
xi
List of Figures and Illustrations
Figure 1.1 Natural gas resource triangle (from Holditch, 2006)......................................... 2
Figure 3.1 Gas compressibility factor and viscosity of well A and B .............................. 35
Figure 3.2 Proppant concentration of well A from field treatment report ........................ 37
Figure 3.3 Proppant concentration of well B from field treatment report ........................ 37
Figure 3.4 Injection rates of well A and B from field treatment report ............................ 38
Figure 3.5 Bottomhole field injection pressure - well A and B ........................................ 38
Figure 3.6 Field gas rate and cumulative production of well A........................................ 40
Figure 3.7 Field gas rate and cumulative production of well B ........................................ 40
Figure 3.8 Horizontal stress trajectories across western Canada sedimentary basin
(from Bell and Babcock, 1986 as reproduced by Mossop and Shetsen (1994)) ....... 44
Figure 4.1 Well completion and fracture spacing – well A and B .................................... 52
Table 4.2 Calculated fracture properties of sands in well A ............................................. 56
Table 4.3 Calculated fracture properties of sands in well B ............................................. 57
Figure 5.1 Flow chart of simulation cases – Chapter 5..................................................... 61
Figure 5.2 Comparison of field data and simulation results of Cases 5.1.1 - 2 - well A .. 66
Figure 5.3 Comparison of field data and simulation results of Cases 5.1.1 - 2 - well B .. 66
Figure 5.4 Dimensionless fracture conductivity of Cases 5.1.1 - 2 – Well A and B ........ 70
Figure 5.5 Comparison of field data and simulation results-Cases 5.1.2 & 5.2 -well A .. 72
Figure 5.6 Comparison of field data and simulation results-Cases 5.1.2 & 5.2 - well B . 72
Figure 5.7 Stimulated reservoir volume (SRV) schematic – Complex fracture

geometry representation ............................................................................................ 75
Figure 5.10 Comparison of field data and simulation results of Cases 5.3.1 - 3-well B .. 78
xii
Figure 5.11 Comparison of field data and simulation results of Cases 5.3.4 - 6 -well B . 78
Figure 5.12 Areal view of fracture in x-y plane – Flow pattern representation................ 79
Figure 5.13 Time dependent permeability multipliers-History matched case–Well A .... 82
Figure 5.14 History matching of well A using time dependent permeability ................... 83
Figure 5.15 History matching of well B using time dependent permeability ................... 83
Figure 6.1 Flow chart of simulation cases – Chapter 6..................................................... 87
Figure 6.2 Comparison of field data and simulation results of Case 6.1 - well A ............ 90
Figure 6.3 Comparison of field data and simulation results of Case 6.1 - well B ............ 90
Figure 6.4 Pressure dependent permeability curves of Cases 6.3.1.1 – 4 - well A & B ... 94
Figure 6.5 Comparison of field data and simulation results of Cases 6.3.1.1-4-well A ... 94
Figure 6.6 Comparison of field data and simulation results of Cases 6.3.1.1-4-well B ... 95
Figure 6.7 Comparison of field data and simulation results of Case 6.3.2 - well A ......... 98
Figure 6.8 Comparison of field data and simulation results of Case 6.3.2 - well B ......... 98
Figure 6.9 Comparison of field data and simulation results of Case 6.3.3.1 - well A .... 100
Figure 6.10 Comparison of field data and simulation results of Case 6.3.3.1 - well B .. 100
Figure 6.11 History matched case using PDPM outside SRV - well A.......................... 102
Figure 6.12 History matched case using PDPM outside SRV - well B .......................... 102
Figure 6.13 Comparison of field data and simulation results of Case 6.3.4.1-2-well B . 105
Figure 6.14 History matched case using PDPM outside SRV and modified reservoir
permeability - well B .............................................................................................. 106
Figure 6.15 Pressure and effective stress dependent permeability curves at stress
factor of 2.7 and 3.9 - well A and B........................................................................ 108
Figure 6.16 History matched case using PDPM both inside and outside SRV- well A . 108
Figure 6.17 History matched case using PDPM both inside and outside SRV - well B 109
Figure 6.18 Comparison of field data and simulation results of uncoupled and coupled
cases: Constant mean total stress - well A .............................................................. 117
xiii
cases: Constant mean total stress - well B .............................................................. 117
Figure 6.20 Coupling of reservoir and geomechanical simulator – An overview .......... 118
Figure 6.21 Plot of mean effective stress and fluid pressure – well A ........................... 119
Figure 6.22 Permeability multipliers and mean effective stress for constant and
variable mean total stress cases – well A ................................................................ 120
factor of 2.7, 3.9, 6 and 8.4 - well A ....................................................................... 121
factor of 2.7, 3.9, 6 and 8.4 - well B ....................................................................... 122
cases: Variable mean total stress - well A............................................................... 123
cases: Variable mean total stress - well B ............................................................... 124
Figure 7.1 Flow chart of simulation cases – Chapter 7................................................... 128
Figure 7.2 Pressure and effective stress dependent permeability curves at stress factor
of 6 (uncoupled injection cases) - well A and B ..................................................... 134
Figure 7.3 Transmissibility multipliers – uncoupled injection cases - well A................ 136
Figure 7.4 Comparison of simulation and field BHIP of Case 7.3.1.1-3- well A........... 137
Figure 7.5 Comparison of simulation and field BHIP of Case 7.3.1.1-3 - well B .......... 137
Figure 7.6 Comparison of simulation and field BHIP of Case 7.3.2.1- well A .............. 139
Figure 7.7 Comparison of simulation and field BHIP of Case 7.3.2.1-2- well B ........... 140
Figure 7.8 Effective stress dependent permeability curve - coupled injection - well A . 143
Figure 7.9 Transmissibility multipliers for coupled injection cases - well A ................. 143
Figure 7.10 Comparison of simulation and field BHIP of Cases 7.4.1.1-4 - well A ...... 145
Figure 7.11 Proppant concentration, BHIP and rate - well A ......................................... 147
Figure 7.12 History matched case – Coupled simulation - well A ................................. 148
Figure 7.13 Comparison of simulation and field BHIP of Cases 7.4.1-4 - well A ......... 149
xiv
Figure 7.14 Minimum effective stress and BHIP at different Biot’s constant numbers
– Coupled cases - well A ........................................................................................ 151
Figure 7.15 Schematics of tensile and shear failure under normal loading .................... 153
Figure 7.16 Mohr-Coulomb failure criterion for critical stress state – A graphical
representation (from Faejer et al., 2008) ................................................................. 154
Figure 7.17 Fluid pressure after 236 days and selected grid block locations – Well A.. 156
Figure 7.18 Minimum effective stress block 1 – 4 – history matched case – Well A .... 157
Figure 7.19 Total in-situ stresses and BHIP for block 1 and 2 –– Well A ..................... 157
Figure 7.20 Mohr – Coulomb failure envelope for Co= 46570psi – Well A .................. 159
Figure 7.21 Stress level in Fracture # 4 after 237mins of injection – Co= 46570 psi –
YZ cross section – Well A ...................................................................................... 160
XY areal view – Well A.......................................................................................... 161
Figure 7.23 Mohr – Coulomb failure envelope for Co = 9314 psi – Well A .................. 162
Figure 7.24 Stress level in Fracture # 4 after 237 mins of injection – Co= 9314 psi –
Figure 7.26 Mohr – Coulomb failure envelope for Co= 4657psi –– Well A .................. 163
xv
List of Symbols, Abbreviations and Nomenclature
cross sectional area
API American Petroleum Institute
fracture area in y-direction
matrix area in y-direction
half aperture distance (fracture half width)
constant defining conductivity / porosity
BHIP bottomhole injection pressure
BHP bottomhole flowing pressure
formation volume factor of phase l
conversion constant
bulk compressibility
fracture conductivity
fracture compressibility at zero stress
grain compressibility
average pore compressibility
CO2 carbon dioxide
pore compressibility
solid grain (matrix) compressibility
constant depending on grain packing
uniaxial compressive strength
xvi
maximum pore compressibility
d average fracture spacing
elastic modulus
dimensionless fracture conductivity
flow factor
shear modulus or modulus of rigidity
acceleration of gravity
poroelastic constant
formation net pay thickness or pay zone
fracture height
root mean square value of height distribution
IPM initial permeability multiplier
formation permeability
gas apparent permeability
bulk modulus of grain
fracture permeability
Kh horizontal formation permeability
effective fluid permeability
matrix permeability
microcracks permeability
dimensionless parameter that describes
conductivity for fluid seepage in plugged pores
xvii
relative permeability of phase l
solid grain modulus
baseline permeability of sand
Kv vertical formation permeability
specific permeability to water
pressure dependent formation permeability
permeability at initial pressure or effective stress
Klinkenberg permeability (no gas slippage)
permeability at 1000 net confining pressure
crack length per unit area
fracture half length
mass of sand in formation
number of thickness
net overburden pressure
net confining pressure
PDPM pressure dependent permeability multipliers
overburden pressure
external pressure
fluid pressure
initial fluid pressure
fracture opening or closing pressure
fluid pressure in fracture at a distance x
xviii
internal pore or fluid pressure
equivalent pressure due to cementation and permanent
deformation of grains
confining pressure
maximum net overburden pressure
minimum net overburden pressure
PVT pressure volume and temperature
fluid pressure at the wellbore
effective elastic modulus
arithmetic mean of gas pressure in core
flow rate of phase l
local fluid flow rate
constant, permeability reduction factor
radius of curvature
wellbore radius
effective wellbore radius
stress factor
skin
combined standard deviation of peak heights
SRV stimulated reservoir volume
shear strength or Cohesion
transmissibility
Tcf trillion cubic feet
xix
transmissibility of fracture in y-direction
matrix transmissibility in y-direction
transmissibility multiplier in y-direction
transmissibility multiplier in z-direction
transmissibility multiplier
ultimate tensile strength or tensile strength
t time
volume of sand (downhole volume of sand)
X-link cross link
downhole volume
fracture volume
volume of total fracturing fluid injected
volume of fluid leaked off from fracture face into formation
grid width
fracture width
total fracture length in y-direction (y = 2Lf)
shear strain
Biot’s constant
constant from curve fitting experimental data
failure plane orientation angle
empirical coefficient
strain in corresponding direction
xx
volumetric strain
poroelastic constant
l phases i.e., gas, oil, and water
lame’s 1st parameter
fracture fluid viscosity
viscosity of fluid of phase l
formation bulk density of the overburden
density of sand
external stress
maximum horizontal total stress
minimum horizontal total stress
initial minimum horizontal stress
minimum horizontal stress at corresponding time
stress in corresponding direction
total mean stress
initial total mean stress
vertical total stress
effective stress
minimum horizontal effective stress
mean effective stress
maximum effective stress
minimum effective stress
xxi
effective stress in x-direction
effective stress in y-direction
or vertical effective stress
initial effective stress
initial effective stress in x-direction
initial effective stress in y-direction
tangential effective stress
shear Stress
Poisson’s ration
friction angle
formation porosity
pressure dependent formation porosity
sand porosity
instantaneous porosity
porosity at initial pressure or effective stress
difference operator
pressure drop
grid size in y-direction
increase in effective stress
Subscripts
i x, y, and z - direction
ij xy, xz, and yz – direction
xxii
CHAPTER ONE: INTRODUCTION
1.1 BACKGROUND
Holditch (2006) defined the tight gas reservoirs as
“Reservoirs that cannot be produced at economic flow rates nor recover economic
volumes of natural gas unless the well is stimulated by a large hydraulic fracture
treatment, by a horizontal wellbore, or by use of multilateral wellbores.”
The importance of unconventional reservoirs increased with advancements in horizontal
well drilling and hydraulic fracturing technology over the past decade. Tight gas
reservoirs produce at low rate over longer period of time due to low permeability.
Compared with conventional reservoir much closer well spacing is needed for tight gas
reservoir to get economic recovery. The concept of resource triangle used by Masters
(1979) gives us very good idea about importance of unconventional resources. Figure 1
illustrates that high quality and highly permeable hydrocarbons, which occur in small
volumes, lay on top of triangle while abundant hydrocarbons having very small
permeability lay at the bottom of triangle. The inadequate gas price and existing
technology make it difficult to recover unconventional hydrocarbons at the base of the
triangle at full scale. The ease to recover the hydrocarbons with current technology
decreases downward through the triangle. The history of tight gas sand exploitation traces
back to 1950’s and 1960’s and the development accelerated with the advent of hydraulic
fracturing technology and developments in multilateral well completions. Increased gas
1
prices and improved technology has led to development of several tight gas sand plays all
over the world, most notably in North America.
Limited knowledge of geology and engineering of unconventional gas resources, natural
gas policies and unfavourable market conditions and shortage of expertise currently
obstructs the exploitation and development of tight gas reservoirs world-wide except in
North America. 58 Tcf of gas has been produced from tight gas reservoirs in U. S.
throughout the year 2000 and it is estimated that 20% of gas production in U.S. is
currently from tight gas reservoirs as reported by Holditch (2006). According to Rogner
(1996) as reported by Kawata and Fujita (2001) there are about 7405 Tcf of gas in place
in tight gas sandstones all over the world.
Figure 1.1 Natural gas resource triangle (from Holditch, 2006)
Rapid growth of global energy consumption, increasing demand for fuel and faster
decline of conventional energy resources shift the focus of global energy producers to
2
development of unconventional resources. To fully exploit the unconventional resources
attention must be paid to understanding of the geology and engineering. New methods
and techniques should be developed based on better understanding of the theory to
maximize recovery.
1.2 RESEARCH OBJECTIVES
The current project is part of a research consortium which studies new simulation
methods for the development and exploitation of unconventional natural resources. The
emphasis of the work is on better understanding and modeling of reservoir engineering
and geomechanical aspects of stimulation and production of tight gas and shale
reservoirs. The research is divided into two main sections.
1.2.1 Production Modeling of Multifractured Horizontal Wells
The main purpose of this project is to study different fracturing techniques used in
hydraulic fracturing, and the effect of fracturing and geomechanical behavior of the
reservoir on well productivity. The study utilized stimulation and production data for two
horizontal wells in a tight gas play. The objectives of this study are discussed below:
 Model production performance of both wells with different fracture properties.
 Estimate fracture parameters using downhole injected proppant volume.
 Investigate effects of fracture parameters (fracture length, conductivity and
width), spacing, fracture density and assumed size of Stimulated Reservoir
Volume (SRV) on gas rate and gas production.
3
 Carry out preliminary history matching for both wells with available data from the
field using manual change of reservoir permeability with time.
 Investigate the effects of pressure / stress dependent permeability on results, both
inside and outside stimulated reservoir volume.
 Carry out history matching using pressure / stress dependent permeability using
both uncoupled and fully coupled geomechanical models.
 Approximate the poroelastic effects (i.e., geomechanical effects) in a conventional
reservoir simulator without running fully coupled geomechanical simulator.
Apache Canada, which participates in the consortium, provided field data for the purpose
of this study. They provided data for 2 different wells, which will be referred to here as
Well A and Well B. The wells were drilled in the same formation but at different
locations. As both wells were completed in tight gas sand reservoir, they were stimulated
by hydraulic fracturing to get maximum recovery.
Two different kinds of fracturing techniques were used in each well, and they will be
represented by following names.
 X-link Gelled Water Frac with CO2 (Hybrid)-----------------------Well A
 Slick Water Frac (Ungelled Water Frac) with CO2------------------Well B
1.2.2 Injection Modeling of Multifractured Horizontal Wells
In this part of project the objective is to study the injection (stimulation) process itself.
The main tasks are:
4
 To model and history match field injection bottom hole pressure using both
uncoupled and fully coupled geomechanical models.
 To study in detail the effects of Biot’s constant and permeability reduction factor,
as history matching parameters, on field injection pressure.
 Carry out a sensitivity study of these parameters on injection pressure.
 Investigate the shear failure mechanisms and the extent of the Stimulated
Reservoir Volume (SRV) as a function of rock mechanics data.
1.3 THESIS ORGANIZATION
The major tasks, which have been completed in the course of this study, are described in
Chapters organized as follows:
1. Literature Review is summarized in Chapter 2.
2. Chapter 3 gives brief summary of field data that includes basic data required for
reservoir study, stimulation data from field, field production data and
geomechanical data.
3. Chapter 4 discusses modeling of static fractures using field stimulation data,
estimation of fracture parameters and representation of these fractures in
simulator.
4. Fracture conductivity can significantly affects the production results. The effects
of fracture conductivity along with other parameters such as net pay thickness and
5
stimulated reservoir volume on gas rate cumulative production are discussed in
Chapter 5.
5. In Chapter 6, the effects of pressure and stress dependent permeability and shape
of permeability curves on production are discussed. Use of pressure and stress
dependent permeability both inside and outside stimulated reservoir volume is
essential for history matching, and is implemented using both uncoupled and fully
coupled geomechanical models. A new method based on uniaxial deformation
theory is proposed to approximate poroelastic effects and model geomechanical
effects in conventional reservoir simulator (without running fully coupled
geomechanical model).
6. History matching of field injection pressure with simulation results using
uncoupled and fully coupled injection model is given in Chapter 7. Simplified
fracture models, fracture propagation modeling, effects of permeability reduction
factor, and Biot’s constant are also discussed in great detail in Chapter 7.
7. Summary of results and conclusions is discussed in Chapter 8.
In addition, research was carried out on two related topics. These are not included here
partially because of space limitations, and also because they require some additional
work. However, they are briefly described below. We intend to publish this work
separately.
 First, fracture density and location of fractures (fracture placement) plays a very
important role in production after stimulation treatment. It is often the case that
some fractures are not contributing to production for various reasons. Cases
6
having different sets of fractures of several combinations were simulated and
categorized in classes based on equal effects on gas rate and cumulative
production.
 Second, the flow of gas in shale is more complex compared to conventional
reservoirs, and additional physics such as slippage desorption, and diffusion may
play a large role. A study of these phenomena on a pore scale has been carried out
jointly with other researches, and shows that the effects can be significant if
permeability is small. The effects are more dominant and considerable in shale
gas reservoirs but initial investigation revealed that these factors are equally
important for tight gas sands.
Finally, the use of units must be discussed. In this work, we use the field (Imperial)
system of units unless stated otherwise. The reason is that the large amount of data we are
dealing with is in these units which are customary in the petroleum business.
7
CHAPTER TWO: LITERATURE REVIEW
The process of creating a single fracture or a fracture system in porous medium by
injecting fluid at very high pressure is termed hydraulic fracturing (well stimulation). In
the process, fluid pressure must overcome the compressive stresses and tensile strength of
the rock, which defines the fracturing pressure during the fracturing job. In a typical
hydraulic fracturing process, two wings of a crack (fracture) are created initially by
injecting fracturing fluid at very high rate and pressure. The clean fluid which is called
pad is injected first in order to extend crack growth deep into the formation and create a
sufficient width for slurry injection. Specially designed slurry containing proppants of
different types is pumped following the pad. The fracturing fluid may contain borate,
guar gum, x-linked gels, alcohols and various other fluids in different combinations to
carry proppant and further extend the fracture and create a highly conductive path.
Chemicals used during fracturing degrade at reservoir conditions which help to recover
injected fluid during flow back.
Klepper No. 1 was the first well to be hydraulically fractured in Hugoton gas field in
western Kansas in 1947. The concept of hydraulic fracturing has evolved from already
existing practices in petroleum industry at that time such as water injection, acidizing of
oil and gas wells and cementing. The process was adopted very rapidly after its initial
successful jobs and performances so that more than 10,000 jobs had been performed by
the end of 1955. Clark (1949) was the first person to introduce hydraulic fracturing to
petroleum industry by name of “Hydrafrac”. Hubbert and Willis (1957) established that
8
the fracture orientation is perpendicular to the direction of the least principal compressive
stress, which in turn can be assumed from the faulting of a region. The minimum
principal stress is usually horizontal in tectonically relaxed areas described by normal
faulting and fractures produced in such areas should be vertical. The minimum principal
stress may be vertical in tectonically active area described by folding or thrust faulting
and fractures produced should be horizontal. They were the first to point out that
fractures produced should be perpendicular to least principal stress irrespective of fluid
type and supported their argument by experiment. Harrison et al. (1953) strongly favored
theory presented by Hubbert and Willis.
In the beginning the process was used as a recovery process in most of the wells.
Nowadays hydraulic fracturing is an essential part of well completion and plays a very
important role in exploitation and development of unconventional reservoirs such as tight
gas sands, shale gas, shale oil and heavy oil. Recently it is being used in applications such
as waterfracs, waterflood-induced fracturing, produced water re-injection, drilling waste
disposal and induced fracturing in thermal recovery methods.
2.1 PRESSURE / STRESS DEPENDENT ROCK PROPERTIES
2.1.1 Experimental Work
The non-rigid and deformable nature of porous media and its effect on flow properties
was recognised as early as 1928 (Meinzer, 1928; Jacob, 1940, 1950). Earlier laboratory
data on reservoir rocks properties were obtained by ignoring overburden pressure and this
was continued until the importance of pressure / stress dependence of rock properties was
9
recognized. Carpenter and Spencer (1940) conducted experiments on oil-bearing
consolidated sandstones and showed that change in porosity with overburden pressure is
small. Fatt and Davis (1952) in their experiments on eight typical consolidated oil bearing
sandstones showed permeability changes with overburden pressure. They concluded that
most of the decrease takes place from zero to 3000 psi range of overburden pressure and
reduction in permeability generally increases as initial permeability increases. The effect
of overburden pressure on pore volume compressibility and relative permeability is small
and can usually be ignored but permeability dependence on overburden pressure is very
strong and reduction in permeability is larger for low permeability rocks (Fatt, 1953).
Hall (1953) conducted experiments on samples from an under-saturated reservoir and
established that total or effective compressibility of any reservoir rock is the results of
two separate factors, i.e., an expansion of the individual rock grains as the surrounding
fluid pressure decreases and the additional formation compaction because of increase in
effective overburden pressure due to pressure decline. McLatchie et al. (1958) and Wyble
(1958) performed experiments on various oil-bearing sandstones over a range of effective
burden pressure of 0 – 8000psi and 0 – 5000psi respectively. They concluded that at
certain given effective overburden pressure the percentage reduction in permeability is
larger for low permeability rocks. Anisotropy in stress is also stress-dependent and
permeability reduction under uniform hydrostatic loading in laboratory is always greater
than that under non uniform loading, i.e., permeability reduction is always overestimated
or reported permeability is always underestimated (Gray and Fatt, 1963). Vairogs et al.
(1971) performed experiments on 10 core samples of initial permeability ranging from
10
0.04 mD to 199 mD over wide range of net confining pressure up to 20,000 psi. They
made the following observations from experiments:
 The effect of stress on tight rock is far greater than for highly permeable rocks
and the reason being is smaller pore radii in tighter formation. The decrease in
flow capacity due to increasing confining pressure for smaller pore radii is much
larger than in larger pores.
 The presence of fractures and shale streaks further accelerate permeability
reduction due to increasing confining pressure.
Thomas and Ward (1972) performed experiments on a variety of tight sandstone core
samples. Significant decrease in permeability with increasing net burden pressure was
observed and most of the decrease took place at pressure to 3000 psi. According to them
relative permeability and porosity does not significantly changes with net overburden
pressure while permeability decreases tremendously with increasing net overburden
pressure in presence of fractures. Brighenti (1989) studied core samples of permeability
range of 0.0009 – 200 mD and different mineralogical characteristics in the laboratory.
He found correlations between normalized permeability and applied stress (net confining
pressure) from experimental data and results from those correlations were in good
agreement with Jones and Owens (1980) and Ostensen (1983) model for stress-dependent
permeability of tight gas sandstones. He reported that reduction in permeability is larger
for low permeability core samples and highlighted that the effect is more pronounced in
presence of fractures.
11
Natural fractures or fissures are considered to be the main source of hydrocarbon
transport in low permeability reservoirs from super low permeability matrix blocks
(Warpinski, 1991; Buchsteiner et al., 1993; Lorenz, 1999). Matrix rock permeability is
typically < 1 μD under in-situ conditions for low permeability reservoirs. Fissures or
natural fractures exist in many of these formations and their presence is critical to the
recovery of hydrocarbon from these reservoirs (Warpinski, 1991). Permeability of these
reservoirs increases one to two orders of magnitude during injection due to decrease of
effective stress, and decreases during drawdown or production due to increase of
effective stress. Reyes and Osisanya (2002) performed experiments on core samples of
four shale types (Wapanuka, Wilcox, Atoka and Catoosa formations in Oklahoma) for
studying effects of in-situ stresses on porosity and permeability. They developed
empirical correlations from experimental data and found that porosity and permeability
are both function of effective stress and must be considered during well development
programs from very beginning. Lei et al. (2007) performed experiments on five core
samples over the range of 2 – 25 MPa effective stress. According to them sensitivity of
permeability to variations in stress is due to the significant change in the bearing
skeletons, solid particles, and pore throats. Pores with oval or polygonal shapes are more
resistant to deformation while pores with flat shapes or silt like structure are more stress
sensitive. Permeability is mainly controlled by pore throats of radii less than 1 μm in tight
formations.
12
2.1.2 Mathematical Correlations
Dobrynin (1962) developed mathematical equations for physical properties such as
porosity, density, permeability, resistance and velocity of elastic waves and verified the
results with experimental data from consolidated sandstones over the range of 0 – 20,000
psi net overburden pressure. The changes of these physical rock properties were due to
changes in pore space of the rocks. Pore compressibility and permeability expressions as
function of net over burden pressure were given by
where = Porosity
= Permeability
= Maximum pore compressibility
= Net overburden pressure =
= Overburden pressure
= Internal pore pressure or fluid pressure
= Empirical coefficient
= Pore compressibility
13
Maximum pore compressibility can be determined by extrapolation of the experimental
curves of pore compressibility to zero pressure using Cartesian co-ordinates. The
empirical coefficient  depends on pore size distribution and pore compressibility; it can
be found by solving Equation (2.4).
Gangi (1978) theoretically derived mathematical equations for permeability and porosity
variation with pressure in porous medium using Hertzian theory of deformation of
spheres and his result is given as;
where = Constant depending on grain packing
= Permeability at initial pressure
= Porosity at initial pressure
= External pressure =
= Radius of spherical grains
= Equivalent pressure due to cementation and permanent deformation of
grains
= Effective elastic modulus of the grains
= Bulk modulus of the grains
14
= Poisson’s ratio of the grains
The theoretical curves generated using the above correlations are found to be in good
agreement with experimental data but this expression needs information such as
cementation of grains, Poisson’s ratio of rock, bulk grain modulus and approximation of
the constant for grain packing.
Jones and Owens (1980) conducted compressibility and flow tests on more than 100 tight
gas sand core samples. They found the following correlations for permeability, gas
slippage and water permeability:
where = Stress factor
= Confining pressure
= Gas apparent permeability
= Klinkenberg permeability (no gas slippage)
= Arithmetic mean of gas pressure in core during flow of gas
= Specific permeability to water
According to them flow in tight gas sandstones is mainly controlled by slit like micro
cracks rather than round shape capillary tubes and they showed that the cubic root of
permeability varies linearly with log of confining stress. They pointed out that routine
permeability measurement (at ambient pressure) should be corrected to 1,000 psi
15
confining pressure, which is the main reason for the appearance of 1,000 in the above
equation. Rocks having lower permeability are affected more by net effective stress than
high permeable rocks. Pore volume compressibility (porosity change with pressure) is
only very slightly affected by increasing confining pressure and hence this nonlinearity
can be ignored.
Jennings et al. (1981) proposed a model to predict permeability and electrical
conductivity of formations. The model assumes that rock composed of irregular and
small pores which are interconnected by thin cracks and microfractures can be modeled
by rectangular slits rather than by round shape capillary tubes. In their model
permeability is related to confining pressure by third order polynomials.
where = Coefficients of cubic equation relating permeability to net
confining pressure
= Net confining pressure
The A coefficients in the above equation can be found by curve fitting of polynomial to
experimental data.
Walsh (1981) developed an alternative asperity-controlled model for fracture
permeability vs. effective stress. This model is based on experimental data on the shape
of surface roughness. By approximating the Gaussian asperity height distribution with an
exponential, Walsh theoretically derived Jones and Owens (1980) correlation. He
confirmed results from his theoretical model given below by Equation (2.12) with
available experimental data from literature.
16
where = Root mean square value of height distribution
= Permeability at initial effective stress
= Half aperture distance or half fracture width
= Effective stress
= Initial effective stress
The first term in Equation (2.12) is the effect of aperture and second term is the effect of
tortuosity. Walsh suggested that based on experimental results the second term in the
above equation can be neglected except for very high effective stress. The analysis
suggested that K1/3 should be linearly related to .
Ostensen (1983) established the existence of network of grain-boundary cracks in tight
gas sandstones. He developed stress dependent permeability model based on flow
through micro cracks in which crack width is controlled by elastic deformation of surface
asperities.
Where = Microcrack permeability
= Crack length per unit area
= Combined standard deviation of peak heights
= Elastic modulus
17
= Radius of curvature
= Poisson’s ratio
= External stress
His model strongly supports the theory that the permeability in tight gas sandstones is
controlled by flow through micro cracks. He proved (using available data from tight gas
sandstones) that square root of permeability varies linearly with log of stress and his
correlation is better than the one presented by Jones and Owens (1980). He was of the
opinion that if micro cracks controls permeability in tight gas sandstones, then the same
could be true for pore compressibility (Porosity). Plot of total porosity with log of
confining pressure should give straight line, which was confirmed with available data.
McKee et al. (1988) have derived relationship for permeability, porosity, and density as
functions of effective stress by using Carman-Kozeny equation. The major assumptions
are incompressible solid matrix and volume changes resulting from compression were
ascribed to pore spaces. Eight core samples, i.e., four coals, one granite, two sandstones,
and one clay, were used to verify applicability of correlations and excellent matches were
found. The resulting relations are:
where = Average pore compressibility
= Increase in effective stress = =
= Pressure drop =
18
= Permeability at initial effective stress
= Porosity at initial effective stress
and s is a constant for linearly elastic materials and rocks (Walsh, 1981).
Buchsteiner et al. (1993) developed equations for fracture permeability and permeability
anisotropy vs. effective stress for fissured or natural fractures dominant reservoirs, which
are stress sensitive. An increase in effective stress due to initial large drawdown results in
high initial rates termed “flush” production provided by rapid drainage of fracture and
fracture network. Flush production substantially reduces permeability, thus inducing
permeability anisotropy and sometimes altering direction of maximum and minimum
permeabilities as a result of depletion.
where = Conductivity / Porosity
Liu and Civan (1995) developed correlation both for permeability and porosity using
Carman-Kozeny equation and the pore compressibility concept, which is given as;
where = Fracture compressibility at zero stress
= Constant determined by curve fitting experimental data
= Grain compressibility
19
= instantaneous porosity during particulate processes
= Dimensionless parameter that describes conductivity for fluid seepage in
plugged pores
= Flow factor, it is defined as the fraction unplugged pore throats
As mentioned earlier, Reyes and Osisanya (2000) developed empirical correlations for
porosity and permeability from experimental dataset of four shale core samples. Their
correlations are in exponential form and depend on the average pore compressibility of
the system. Lei et al. (2007) correlated permeability and effective stress with quadratic
polynomial relation from experimental results on five core samples.
2.2 EFFECTS OF PRESSURE / STRESS DEPENDENT PERMEABILITY ON
GAS PRODUCTION
Vairogs et al. (1971) were the first to study the effects of stress dependent permeability
on gas production. They confirm the conclusions made by preceding authors through
laboratory experiments. They developed a reservoir simulator that incorporated stress
effects for single phase isothermal gas flow to well in a circular reservoir. They assumed
that reduction in permeability caused by changes in radial stresses is small as compared
with that due to the tangential and vertical stresses, and the change in permeability is
therefore a function of average effective (confining) stress defined as
where = Mean effective stress
= Tangential effective stress
20
= Vertical effective stress
They showed that significant decrease in gas production rate was modeled due to
reduction in permeability as gas pressure declines. The authors also concluded that
increasing pressure drawdown will not help to obtain larger recovery in tight gas
reservoirs. Vairogs and Rhoades (1973) have done pressure transient analysis on gas
wells. They showed that drawdown tests in stress sensitive formations are not reliable.
Permeability obtained from semilog analysis for drawdown tests is not consistent and
build up tests give reasonably good results both for early time and late time.
Ostensen (1986) in his paper emphasized that the effect of effective stress on
permeability in tight gas reservoir cannot be ignored and permeability data obtained
under uniformly stressed conditions in laboratory is not actual representation of in-situ
conditions because horizontal stress is usually lower than vertical. He concluded that the
permeability reduction is overestimated or the reported permeability is underestimated by
factor of two or more depending upon cracks directions if measured under uniform
hydrostatic conditions. His finding agrees with the conclusions of Gray and Fatt (1963).
Stress-dependent permeability decreases initial steady state production by 30% for typical
tight gas sand reservoir. Average decrease in stress dependent permeability mainly
controls the late time steady state production.
Warpinski (1991) concluded that low permeability or tight formations require special
care to recover hydrocarbons. Problems that can significantly reduce production from
these fissured formations are leak off, damage, and complex fracturing. Leakoff during
stimulation may be with constant or pressure sensitive permeability and becomes more
severe when natural fractures begin to dilate and results in high leakoff into formation.
21
Liquid damage is transient and will eventually recover but it may take weeks or even
months. Mechanical damage is due to shear slippage and it is permanent. Complex
fracturing will cause high treatment pressure and generally reduce effective half length.
Special care must be taken when formations are initially perforated and fractures are
created otherwise anticipated production may not be achieved.
Significant reduction in production rates in stress sensitive reservoirs is due to rapid
decrease in near-wellbore permeability because of increase in effective stress as a result
of pressure depletion. Thus, reducing bottomhole pressure to increase recovery may not
help because the high pressure drawdown near wellbore will eventually decreases
permeability. The best way to increase production in tight formations is to maintain high
reservoir pressure (Buchsteiner, 1993). Lorenz (1999) reported that although large
fractures may dominate flow at well bore, they must have support from densely populated
smaller and more stress sensitive fractures. Conductivity for fractures width less than 10
μm drops more rapidly than cubic of aperture.
Davies and Davies (2001) indicated in their study that reservoirs characterized by stress-
dependent permeability are dynamic permeability reservoir models. Stress dependent
permeability, both in unconsolidated and consolidated reservoirs, is controlled mainly by
pore geometry, and the rate of change of permeability with stress is a direct function of
rock type. According to their conclusions sand with large pores and initial permeability is
more stress sensitive for unconsolidated sandstones. In consolidated sandstones, sands
with long narrow slit like pores are more stress sensitive than those with long capillary
tube like pores. They showed that stress dependent permeability has significant effects on
the performance of both well and reservoir as a whole.
22
Lei et al. (2007) performed production analysis on low permeability core samples and
concluded that near wellbore permeability decreases up to 30%, which results in
production rate decrease of up to 21%. They emphasized that optimum bottomhole
flowing pressure should be determined from reservoir conditions and rock stress
sensitivity to get maximum recovery. Pressure / stress dependent permeability is equally
important in injection modeling such as for water flooding, water reinjection, etc. Wu et
al. (20008) showed that assuming constant permeability in stress sensitive reservoirs
overestimates pressure response at the wellbore as it ignores the effect of fracture
expansion on fracture permeability. They concluded that for high injection pressure
operation neglecting permeability dependence on pressure will lead to large errors.
2.3 CONVENTIONAL FRACTURE MODELS
Knowledge of fracture evolution and creation (i.e., fracture length, width and height) and
how these properties affect the fracture performance is critical to fracture modeling and
simulation. Fracture propagation in fracture simulation software packages, currently
available in the market, is estimated by solving the material balance of injected fluid. The
total amount of fracturing fluid injected minus the volume of fluid leaked off from the
fracture faces into the formation is the amount of fracturing fluid remaining in fracture
that actually constitutes fracture volume. The fracture volume from material balance of
injected fracturing fluid is written as:
where = Fracture volume
23
= Volume of total fracturing fluid injected
= Volume of fluid leaked off from fracture face into formation
Fracture volume is calculated from Equation (2.21) at any time, which is then used to
estimate fracture parameters using different assumptions about fracture geometry.
2.3.1 Two – Dimensional (2-D) Fracture Models
Two dimensional fracture models only solve for two fractures dimensions, i.e., fracture
width and length (or radius) while fracture height is assumed constant (generally equal to
net pay thickness or pay zone thickness, or determined by stress contrast). Two
dimensional models such as PKN and KGD model are discussed here.
2.3.1.1 PKN Fracture Model
PKN fracture model was first proposed by Perkins and Kern (1961), and was later
improved by Nordgren (1972). PKN fractures model takes the following assumptions to
derive relations for fracture length and width.
 They considered vertically limited fracture, i.e., constant fracture height usually
equal to net pay thickness.
 Fracture in each vertical cross section deforms independently in vertical planes
perpendicular to the main fracture planes, i.e., plane strain conditions exist.
 Fracture vertical cross section is elliptical and maximum fracture width is at the
center of a cross section (ellipse).
24
 The fluid pressure in the fracture is constant in each vertical cross section
(perpendicular to fracture propagation). Fluid pressure is then only a function of
distance from the well, and at the fracture tip is equal to total earth stress
perpendicular to plane of the fracture, i.e., minimum horizontal stress.
Perkins and Kern (1961) derived the equations for fracture length and width vs. time both
for Newtonian and Non Newtonian fluid in laminar and turbulent flow regimes
neglecting fluid loss and rate of change of fracture volume. Nordgren (1972) in his work
later concluded that length at early times must be determined by their formula not
Carter’s formula while width predicted by Perkins and Kern is good approximation to his
solution. Large time fracture length formula is identical to Carter’s formula and width
formula gives superior results to the Perkins and Kern width formula.
PKN fracture width at any location x along the fracture length at a time t is given as
where = Fracture height
= Poisson’s ratio
(x) = Fluid pressure in fracture at a distance x from the well
= Minimum principal horizontal stress
= Shear modulus or Modulus of rigidity
Pressure gradient for restricted, vertical, and elliptical fracture from injection of
Newtonian fluid in laminar flow is
25
Where = Local fluid pressure
= Local fluid flow rate
= fracturing fluid viscosity
= Local fracture width
= coordinate along fracture length
2.3.1.2 KGD Fracture Model
KGD model solves the 2-D vertical crack problem in a horizontal plane. Khristianovic
and Zheltov (1955) were first to investigate the fracture width assuming constant width in
each vertical crossection which was later improved by Geertsma and de Klerk (1969).
KGD model assumes plane strain condition in a horizontal plane, constant fracture height
and fracture tip which closes smoothly.
The maximum fracture width for bi-wing linearly propagating fracture at wellbore (x=0)
at any time t is
Where = Fracture half length
= Fluid pressure at the wellbore (i.e., the bottomhole pressure)
Khristianovic and Zheltov (1955) expressed pressure gradient for laminar fracture fluid
flow and linearly propagating fracture by
26
A very small region is assumed at the very tip of fracture, where there is no fluid flow
and there is infinite flow resistance in that region. Most of the pressure drop occurs at the
tip of the fracture and therefore pressure and flow rate are constant in majority of the
fracture except for small region near fracture tip (Khristianovic and Zheltov, 1955).
Carter’s model is not used presently for fracture propagation modeling but Carter’s leak
off model is still used in fracture modeling. Although PKN and KGD models (2-D
models) have been replaced by fully 3-D and pseudo 3-D models in fracture modeling
these days, they are still available in many fracture modeling software packages as these
models serve as the basis for current 3-D models discussed next.
2.3.2 Three – Dimensional (3-D) Fracture Models
Real fractures are three-dimensional, which means that fracture height, width and length
all grow with respect to time and space. 3-dimensional models allow fracture height to
change at any location along the fracture length and correspondingly the width and its
vertical profile also change throughout the injection period.
In fully three – dimensional fracture models (Clifton and Abou-Sayed, 1981; Cleary,
1983) fracture develops vertically in plane perpendicular to minimum in – situ stress
assuming that the medium through which it can propagate, i.e., formation and the upper
and lower shale, is infinite. The fracture shape and width is obtained from a 3-
dimensional solution of frac opening equations on a half-space. Pseudo-three-
dimensional fracture models (P3D) usually assume that fracture length is relatively large
compared to fracture height so that all cross-sections deform independently of each other
in the lateral direction, and the propagation of the vertical tips can be obtained by a series
27
of 2-D solutions. Lumped 3-D fracture models are based on further simplification of the
P3D fracture models. These models assume that the fracture height growth is governed
by KGD fracture model and the fracture length growth is governed by PKN fracture
model for fractures in well-contained reservoir where fracture length is larger than
fracture height. The corresponding lumped fracture model is called a Lumped Lateral
Model. Analogously, for fractures in poorly-contained reservoir where fracture height is
larger than fracture length, one can assume that the height growth is governed by PKN
fracture model and the fracture length growth is governed by GDK model, resulting in a
Lumped Vertical Model.
All of the fracturing models discussed above are primarily concerned with the created
fracture geometry. However, for production modeling the important part of the solution is
the propped fracture geometry, which must be obtained by simultaneous solution of the
proppant transport during the fracturing job. The propped length and height of the
fracture can be significantly smaller than the created fracture dimensions, and it is a
function of not only the created fracture geometry, but a number of other parameters. In
our work, we will be primarily concerned with the propped fracture dimensions and its
conductivity.
2.4 UNCONVENTIONAL FRACTURE MODELS
The presence of microcracks and natural fractures in tight formation has been known for
some time, but the advancements in hydraulic fracturing technology and interdisciplinary
research in this area helped to understand the complex nature of these reservoirs.
28
Interactions of natural fractures or fissures with single planar fracture form a complex
fracture networks with increased permeability around the main fracture in naturally
fractured reservoir such as tight sands and shales. This area of increased permeability is
often referred to as the “Stimulated Reservoir Volume” (SRV) and this concept is largely
based on the understanding and analysis of microseismic events (Olson et al., 2012). The
branching of hydraulic fractures due to crossing, bypassing, intersecting, arresting due to
shear slippage or dilating, and diversion along natural fractures or fissures give rise to
development of a complex fracture network, modeling of which is referred to as an
Unconventional Fracture Model (UFM). Interactions of natural fractures with the
hydraulic fractures and their contributions to production in unconventional reservoirs are
still a current topic of debate and discussion. Laboratory experiments are a way to
investigate the physics of complex hydraulic fracturing but one potential problem is the
scaling of laboratory results to real field operations (Olson et al., 2012).
Sampath and Keighin (1982) made pore casts from tight gas sand cores. They reported
that thin film inter-granular pores (less than 3 μm across) often observed in sandstones
and tight sands have large number of pores smaller than 1 μm. Walls (1982) performed
experimental study on 10 core samples from Spirit River Formation in Alberta and 2
samples from Cotton Valley Formation in Texas. He reported that pore structure in tight
sands is very narrow and slit like aperture between pores acts like thin cracks, which
provides major connectivity and allow fluids through it. Brower and Morrow (1985)
produced pore casts tight sands and concluded that flow through tight gas sands is
controlled by network of interconnected polyhedral sheet pores, which are somewhat like
the surface of a randomized honeycomb.
29
Warpinski and Teufel (1987) have provided direct evidence of complex fracture
geometry from mineback studies and illustrated the complex nature of actual fractures
created in naturally fractured or fracture dominated reservoir. They reported that geologic
discontinuities – faults, joints, bedding planes, and stress anisotropy have significant
effects on fracture height, length, leakoff, treatment pressure, and proppant transport. The
effects of these discontinuities on hydraulic fracture strongly depend on treatment
pressure, in-situ stresses, orientations of fissures, and permeability. A hydraulic fracture
model that simulates a complex fracture network, referred to as Unconventional Fracture
Model (UFM), had been recently been developed assuming presence of natural fractures
in formation (Weng et al., 2011). They concluded that stress anisotropy, natural fractures,
and interfacial friction play an important in creating fracture network and their results can
be used to improve fracture stimulation design.
Lin and Zhu (2012) presented a semi-analytical model using a source and sink method to
estimate well performance in complex fracture geometry by assigning fracture network of
natural fractures to multiple hydraulic fractures. Each fracture (either hydraulic or
natural) is considered as a series of line sources and they analytically solved the problem
using principle of superposition. Olson et al. (2012) carried out experiments in the
laboratory to examine the effects of cemented natural fractures on hydraulic fracture
propagation. They showed that oblique embedded fractures more likely divert the
hydraulic fracture as compared to orthogonal fractures. They demonstrated in their
experiments that interactions of hydraulic and natural fractures are due to hydraulic
fracture bypassing natural fracture by propagating around it, hydraulic fracture arresting
into a natural fracture and diverting along it, and combination of both.
30
Joint deformation, magnitude and orientation of original and altered stresses, and natural
factures spacing are sources of fracture complexity and interconnectivity (Urbancic,
2012). Wu et al. (2012) presented a method for computing so called “stress shadow”
which may be present during stimulation in multifractured horizontal wells. Stress
shadow is an effect of in-situ stress increase seen when performing a given stage of the
treatment caused by additional stresses created by previous stimulation in the adjacent
treatment stages in a complex fracture network. Stress shadow can significantly affect
fracture width, fracture network pattern, alter fracture propagation path, and proppant
placement. However, its role is still controversial because in many treatments the effect is
absent while in others is larger than expected (Vermylen and Zoback, 2011).
To sum up all the discussion, one can quote one sentence by Warpinski et al. (1993):
“The more complex the reservoir, the more complex the fracturing.”
2.5 SUMMARY
The following conclusions are drawn from above literature review and are now well
accepted in petroleum industry.
 The permeability of rock decreases with increasing confining pressure. The
reduction in permeability is 10 – 100 times larger for very tight rocks than for
permeable ones. One possible explanation of larger reduction in permeability is
smaller pore radii in tight formations. The flow capacity of an aperture depends
on a high power of its dimensions such fourth power for round capillaries and
third power for rectangular slits. Flow in tight gas sandstones (very low
31
permeability rocks) is mainly controlled by slit like apertures (micro cracks)
rather than round shape capillaries (Jones and Owens, 1980; Jennings et al., 1981;
Ostensen, 1983).
 The effect of confining pressure on pore volume compressibility and relative
permeability is not substantial and can be ignored for practical calculations.
 The reduction of permeability is further accelerated in presence of fractures, shale
streaks or other heterogeneity.
 The permeability reduction is always overestimated by conducting tests under
uniform hydrostatic loading in laboratory than non-uniform loading. True triaxial
tests at in situ conditions are better approximations.
 Pressure / stress dependent permeability must be incorporated in reservoir
engineering studies of tight formations because constant permeability assumption
for these formations is always misleading.
 Classical bi-wing planar fractures alone cannot explain fracture geometry and
productivity for unconventional reservoirs due to presence of natural and shear
fractures / fissures and geological discontinuities. Unconventional fracture models
are the more realistic and correct way of modeling complex fracture networks in
tight formations.
32
CHAPTER THREE: SUMMARY OF THE FIELD DATA
Wells under investigation are both dry gas wells of very low permeability completed in
the Cadomin formation. British Petroleum Canada, which was originally the participant
in the research consortium and agreed to provide the data, sold its Canadian gas holdings
to Apache Corporation who ultimately provided field data for this study. This chapter is
organized in several sections discussing basic data required for reservoir study,
stimulation treatment data from field, production data and geomechanical data. The
objectives of this chapter are to discuss and represent given data, highlight missing data,
and discuss methods used to generate necessary missing data by using best available
correlations from literature.
3.1 BASIC DATA
Basic data is representation of data that is required for reservoir study such as drainage
area, permeability, porosity, reservoir temperature, pressure, relative permeability, water
saturation and PVT data. Some of the above data was not available at the time of start of
the project so it was decided to use correlations to generate it. The following data was
missing from the information provided by the company:
 Drainage area
 Net pay thickness
 Connate water saturation
 PVT data
33
Drainage area of 10,000 x 6,000 ft2 (1377 acres) is assumed for current project. Initially
net pay thickness of 80 ft and connate water saturation of 25% was assumed based on
discussion with various sources. To generate PVT data for both wells the following step
by step process was followed using correlations.
 Correlation proposed by Sutton (1985) is used to calculate pseudo critical
properties of natural gas.
 Standing and Katz compressibility chart has been used for long time for natural
gas compressibility factor estimation. Several empirical correlations were
developed over years to calculate compressibility factor. Yarborough and Hall
(1974) presented an equation of state based on Carnahan-Starling equation of
state that accurately represents the Standing and Katz compressibility factor chart.
Dranchuck et al. (1974) developed a correlation for reduced density based on the
Benedict-Webb-Rubin type equation of state. Dranchuck and Abu-Kassem (1975)
derived an expression for reduced density that can be used to calculate
compressibility factor. All of the above three methods are applied to calculate z-
factor and it is observed that z-factor estimated by the above three methods give
essentially the same value for range of pressure of interest in this project. It was
then decided to use Yarborough and Hall (1973) method to calculate gas
compressibility factor.
 Carr et al. (1954) developed a graphical correlation for estimation of viscosity of
natural gas. Lee et al. (1966) proposed semi-empirical correlation in terms of gas
density, reservoir temperature, and molecular weight of gas for calculating
34
viscosity of natural gas. The viscosity of natural gas in this project is calculated
using the correlation presented by Lee et al. (1966).
Figure 3.1 represents natural gas compressibility factor and viscosity calculated by
correlations as discussed.
1.00 0.022
Compressibility factor (Z-factor)
Gas viscosity
0.97 0.019
Compressibility factor
Gas Viscosity (cp)

0.94 0.016
0.91 0.013
0.88 0.01
0 800 1600 2400 3200 4000
Pressure (psia)
Figure 3.1 Gas compressibility factor and viscosity of well A and B
The fluid and rock properties that are used as an input to reservoir simulator, TRS®
(provided by Taurus Reservoir Solutions Ltd.), are summarized in Table 3.1 for both
wells. As mentioned in the Introduction, field units system is used in this project unless it
is stated otherwise.
35
Table 3.1 Rock and fluid properties of well A and B (input to reservoir simulator)
Property Value (well A) Value (well B)

Total vertical depth, ft 5246 5308
Formation thickness, ft 80
Porosity, fraction 0.035 0.03
Permeability (Kh), mD 0.015 0.010
Permeability ratio (Kv/Kh), fraction 0.10
Reference pressure, psi 2000 2200
Reference depth, ft 5290 5350
Depth of gas oil contact, ft 15000
Depth of water oil contact, ft 15000
Formation compressibility, 1/psi 3.50E-06
Initial water saturation, percent 25
Reservoir temperature, °F 190
Gas specific gravity, fraction 0.613
Rate constraint, MMscf/d 7.0 5.5
Pressure constraint (minimum BHP), psi 300 275
3.2 STIMULATION DATA
Trican Well Service Ltd. performed the stimulation operations on these wells. One well
was fractured using hybrid fracturing technique, while the other was fractured by slick
water energized with CO2. Two different mesh sizes of proppant, i.e., 40/70 and 50/140
were used in each well in different quantities. Figures 3.2 and 3.3 represent the surface
and bottomhole proppant concentrations pumped in the injection field treatment. Figure
3.4 and 3.5 shows the bottomhole pumping rates and injection pressure in the two wells.
Treatment in Well B was pumped at higher rate and received larger total volume.
36
400
Inline - Well-A
Bottomhole - Well-A
300
Concentration (Kg/m3)
200
100
0
0 50 100 150 200 250
Elapsed Time (mins)
Figure 3.2 Proppant concentration of well A from field treatment report
300
Inline - Well-B
Bottomhole - Well-B
Concentration (kg/m3)
200
100
0
0 50 100 150 200 250 300
Elapsed Time (mins)
Figure 3.3 Proppant concentration of well B from field treatment report
37
15
Bottomhole Injection Rate - Well-A
Bottomhole Injection Rate - Well-B
12
Rate (m3/min)
0
0 50 100 150 200 250 300
Elapsed Time (mins)
Figure 3.4 Injection rates of well A and B from field treatment report
9000
Well A Well B
Bottomhole Injection Pressure (psi)
6000
3000
0 50 100 150 200 250 300
Time (mins)
Figure 3.5 Bottomhole field injection pressure - well A and B

38
The type, quantity and placement of proppant into formation play a very important role in
fracturing. Rushing and Sullivan (2003) concluded that slick water fractures performance
improves by using smaller proppant size (40/70 as opposed to 20/40) in Bossier tight gas
sands. Liu and Sharma (2005) proved through experimental study that proppant transport
is affected by the ratio of the proppant size to fracture width. The lateral flow of proppant
is significantly hindered as proppant size approaches the fracture width, particularly for
x-linked fluids, and proppant settling velocity is also reduced. This may eventually cause
tip screen-out or proppant bridging. The proppant is also distributed non-uniformly across
fracture width. Proppant particles are usually concentrated at the center of the flow
channel and thus travel faster than average fluid velocity because fluid velocity is
maximum at its center. Currently, it is believed in the industry that smaller proppant size
significantly improves initial production rates and effective fracture half length. The type
and amount of proppant used in each well is summarized in Table 3.2.
Table 3.2 Proppant types in well A and B
Property Value (well A) Value (well B)
40/70 mesh size sand, tonnes 50.6 30.0

50/140 mesh size sand, tonnes 31.3 63.2
3.3 PRODUCTION DATA
Average daily gas rate and cumulative gas production was the only complete data that
was available. Figure 3.6 and 3.7 represents the gas rate and cumulative production for
Well A and B respectively.
39
16000 1200
Avg daily Gas Rate
Cumulatve Gas Produciton
Cumulative Gas Production (MMscf)

12000 900
Avg Daily Gas Rate (Mscfd)
8000 600
4000 300
0 0
8-Jun-06 11-Dec-06 15-Jun-07 18-Dec-07 21-Jun-08 24-Dec-08 28-Jun-09 31-Dec-09
Figure 3.6 Field gas rate and cumulative production of well A
12000 1200
Avg daily Gas Rate
Cumulatve Gas Produciton

9000 900
Avg Daily Gas Rate (Mscfd)
6000 600
3000 300
0 0
17-Aug-06 9-Feb-07 4-Aug-07 27-Jan-08 21-Jul-08 13-Jan-09 8-Jul-09 31-Dec-09
Figure 3.7 Field gas rate and cumulative production of well B

40
3.4 GEOMECHANICAL DATA
One of the most important requirements for geomechanics modeling is the knowledge of
in-situ stresses and of rock mechanics data, i.e., Poisson’s ratio, elastic modulus and
possibly other data (such as shear failure parameters). Data necessary to describe the
stress field is the variation of the principal stresses with depth and horizontal stress
orientation. It is well established that usually the vertical stress ( ) is larger than
maximum and minimum horizontal stresses ( and ) at large depth, although
exceptions may exist at shallow depths or in highly tectonic areas. The vertical stress at
depth (z) from the surface (in absence of tectonic events) is caused by the weight of the
overburden sediments and can be obtained as
where = Formation bulk density of the overburden, which can be obtained from well
logs
The propagation of hydraulic fractures and secondary fracture system is controlled by
minimum horizontal stress ( ) and the maximum stress. There are no reliable methods
for predicting in porous rocks without field data. The most common means of stress
orientation measurements are borehole breakouts, downhole telemetry, differential strain
analysis, anelastic strain relaxation, tiltmeter measurements, and microseismic
monitoring of hydraulic fractures. The methods used to measure stress magnitude are
micro-frac or mini-frac testing, and leak-off tests but estimates can also be obtained from
stress logs.
41
Anderson et al. (1973) proposed a relation to compute using Poisson’s ratio and
reservoir pore pressure.
where = Minimum horizontal stress
= Vertical stress
= Poisson’s ratio
= Biot’s constant
= Reservoir pore pressure
Biot’s constant plays an important role in poroelasticity.
where = = Solid grain (matrix) compressibility
= = Bulk compressibility
= Solid grain modulus
= Bulk modulus of grain or specimen
Newberry et al. (1985) modified Equation (3.2) for low porosity micofractured rocks.
Large amounts of data on fracture gradient have been compiled for US by Breckels and
Eekelen (1982) and for Canada by Bell and Babcock (1986). The stress orientation is
strongly affected by tectonic events so that is why data reported by authors show
considerable variations and deviations from the above equations. Therefore, log derived
stress data is more reliable for giving stress variation with depth rather than absolute
stress values of .
42
The is determined for this study from mini-frac analysis report and is estimated
from fracture study, reported by Apache Corporation, taking ratio of maximum to
minimum horizontal stress equal to 1.33 while vertical stress is determined from
calibrated density logs for current project. The horizontal stress values are also confirmed
from calibrated stress logs data. Static elastic modulus, Poisson’s ratio and uniaxial
compressive strength were measured in laboratory. Table 3.3 represents stress and rock
properties that are used in geomechanical modeling both for well A and B.
Table 3.3 Stress and rock properties for geomechanical modeling of well A and B
Property Value
Maximum horizontal stress gradient, psi/ft 1.13
Minimum horizontal stress gradient, psi/ft 0.85
Vertical Stress gradient, psi/ft 1.24
Formation density, kg/m3 2570
Static elastic modulus, psi 7.99E6
Poisson’s ratio 0.125
Uniaxial compressive strength, psi 46570
Friction angle, degrees 30
0.68
1.33
Bell et al. (1994) presented horizontal stress trajectories for western Canada sedimentary
basin in form of chart. The horizontal stress directions determined from breakouts for
western Canada sedimentary basin are given in Figure 3.8. It is assumed here that all
principal stresses are in direction of rectangular coordinate system. In our project well is
completed in x-direction which means that minimum horizontal stresses are also in this
direction. Maximum horizontal stresses are in y-direction while vertical stresses are in z-
direction in rectangular coordinate system.
43
Figure 3.8 Horizontal stress trajectories across western Canada sedimentary basin
(from Bell and Babcock, 1986 as reproduced by Mossop and Shetsen (1994))
44
CHAPTER FOUR: MODELING OF STATIC FRACTURES
The fractures created by hydraulic fracturing treatments, after being propped, do not
propagate with time, and are therefore known as static fractures. These fractures are
created to increase the productivity or injectivity of well. Static or stationary fractures are
filled with proppant injected at very high rate and pressure to create a conductive channel
in porous media of high porosity and high permeability. Permeability of this channel is
not constant and changes with time as stresses changes in wellbore / reservoir due to
pressure depletion. The properties of these highly conductive channels depend on the
type of stimulation treatments such as slickwater, x-linked gelled, hybrid, reverse hybrid,
acid, propped and un-propped fractures, proppant type, size, strength and quantity.
Palisch et al. (2010) defined slickwater treatment, which is also referred as waterfrac, as
“a fracture treatment that utilizes a large volume of water to create an adequate fracture
geometry and conductivity to obtain commercial production from low perm, large net pay
reservoirs.” The key benefits of slickwater frac job are reduced gel damage, lower costs,
and ability to reuse load and produced water, i.e., less environmental footprints
(Mayerhofer and Meehan, 1998; Sharma et al., 2004; Sharma et al., 2005; Liu et al.,
2006; Palisch et al., 2010) and complex fracture geometry (Palisch et al., 2010). Smaller
effective fracture half lengths due to proppant settling in low viscosity fluids, reduced
ability to transport proppant deeper into formation, large amount of water required for
proppant job, and narrower hydraulic fracture width are main concerns with slickwater
frac job (Rushing and Sullivan, 2003; Sharma et al., 2004; Sharma et al., 2005; Liu et al.,
45
2006; Palisch et al., 2010). Hybrid fracture treatment is not very different from slickwater
job except use of x-linked gels during the proppant pumping stage. The key benefits of
these types of job are better proppant transport which results in deeper placement of
proppant pack, larger fracture width (Sharma et al., 2005), and longer effective fracture
half length (Rushing and Sullivan, 2003; Sharma et al., 2004; Liu et al., 2006).
Susceptible gel damage, undesirable vertical fracture growth, and possibility of tip
screenout are major concerns tied to hybrid frac job (Sharma et al., 2005; Liu et al.,
2006). In reverse hybrid fracs, the fractures are created using a polymer gel pad followed
by slickwater for proppant placement (Sharma et al., 2005; Liu et al., 2006). Two
significant advantages of reverse hybrid frac job are larger fracture width allowing easier
proppant transport due to use of x-linked fluid pad and deeper placement of the proppant
into the formation due to viscous fingering as reported by (Sharma et al., 2005; Liu et al.,
2006). Fractures created by acid treatment are open channels and have very high porosity
(i.e., 1.0) while porosity of propped fractures is in the range of 0.32 – 0.40. The
permeability of an open fracture for single phase laminar flow between parallel plates is
Where = fracture width, ft
= Fracture permeability, mD
Turbulent flow, surface roughness, irregular surface shape, tortuosity, multiphase flow
and presence of different proppants all decrease fracture permeability and should be
considered in fracture permeability calculations. Settari et al. (1990) represented a
fracture in modeling productivity of fractured wells in single phase flow by introducing
46
transmissibility modifications of single grid system for both fracture and reservoir to
effectively minimize the grid effect and maximize the stability of numerical simulation.
The fundamental feature of the transmissibility modifications method is to combine the
fracture transmissibility with the transmissibility of reservoir blocks in the fracture plane,
without representing fracture by separate grid blocks. Ji et al. (2004a) discussed various
methods for modeling static fractures in reservoir simulation. According to their
conclusions transmissibility modification method can be used to model infinite and finite
conductivity fractures both in single and multiphase flow. The method can also be
extended to model dynamic fractures propagation in water injection, reservoir damage
and coupled geomechanical problems.
This chapter presents modeling of static fractures using concept of transmissibility
modifications in the reservoir simulator, TRS® provided by Taurus Reservoir Solutions
Ltd. Field injection data was analysed and used to model static fractures for production
modeling. The representation of fractures in reservoir simulator, grid refinement,
selection of fracture density, fracture spacing and estimation of fracture other parameters
will also be discussed in this chapter.
4.1 REPRESENTATION OF FRACTURES IN RESERVOIR SIMULATOR
4.1.1 Transmissibility Modifications
Transmissibility describes flow rate per unit pressure drop between two adjacent blocks.
Transmissibility has two parts, one part is flow properties part and other part is geometric
part. The flow part of transmissibility depends on the pressure and saturation of system.
47
As pressure changes in reservoir these properties changes and they are modified in
simulator accordingly. This part of transmissibility is also called the mobility of fluid.
From Darcy’s law
Where = Flow rate of phase l
= Effective fluid permeability = *
= Absolute permeability
= Relative permeability of phase l
A = Cross sectional area
= Viscosity of fluid of phase l
= Formation volume factor of phase l
= Change in pressure
= distance between centers of blocks between which the transmissibility is
computed
Mobility of fluid changes as pressure and saturation changes; this is automatically
updated in each time step by the simulator itself. Geometric part of transmissibility
remains constant if reservoir absolute permeability and grid size is not changing during
48
simulation. If permeability is considered stress/pressure dependent, this part of
transmissibility is also changing as stress/pressure changes. Here we will be discussing
only the geometric part of transmissibility for the case of a horizontal well with vertical
fractures. As the well is completed in x-direction, the one directional linear flow from
formation into the fracture will be in y-z plane. Transmissibility of non fractured blocks
is calculated by Equation (4.3). Transmissibility of blocks having fractures will be
modified using method given by Settari et al. (1990) and Ji et al. (2004a).
Substituting Equation (4.4) and (4.5) into (4.6)
Similarly for Z-direction,
Where = Transmissibility of fracture in y-direction
= Fracture area in y-direction =
= Fracture permeability
49
= Grid size in y-direction
= Matrix transmissibility in y-direction
= Matrix area in y-direction =
= Matrix permeability
= Fracture conductivity =
= Fracture width
= Transmissibility multiplier in y-direction
= Transmissibility multiplier in z-direction
Equation (4.8) and (4.9) are used to calculate transmissibility multipliers in y and z-
directions respectively and multipliers are represented in form of table in reservoir
simulator for fractured blocks.
4.1.2 Element of Symmetry Model
To reduce computational time, the line of symmetry for each well is taken by dividing
whole reservoir model into two halves in the x direction and performing the simulations
and analysis on a half model. Since fracture density is not constant for our project, the
line of symmetry is taken along the middle fracture by dividing the total number of
fractures into two halves. This implies that we must model only an odd number of
fractures (i.e., 3, 5, 7, etc.). The porosity, transmissibility and well index values for this
specific fracture are divided by 2 to take into account the symmetry. The output of the
simulator such as gas rates and cumulative production is then multiplied by factor of 2.0.
50
Another symmetry one could take an advantage of is in the y direction (solving only for
one wing of all the fractures), but this was not employed for our study.
4.1.3 Grid Refinement
Coarse grids are initially constructed for both wells, which are refined around each
fracture at later stage. Smaller grid sizes are chosen near the well bore to capture effects
of changing pressure and gas saturation, and larger grid sizes were used away from the
well. Grids are refined around each well completion (set of perforations) and fractures in
order to get better pressure profile near the vicinity of the well bore and of the fractures.
Irregular global refined grids of unequal size are used. Grid sizes in all directions i.e. x, y,
z - directions are changed from case to case to accommodate the fracture length due to
change in fracture density, net pay height and stimulated reservoir volume.
4.2 FRACTURE DENSITY
Fracture density plays an important part in production of multi fractured horizontal wells.
The larger is the number of active fractures, the higher will be the production. Initially we
were provided with very limited data about the frac jobs and assumed that a multi stage
fracturing job was done originally to create fractures. There was no information about
number of stages and details of the fracturing job. Therefore initially it was decided to
assume a total 9 numbers of fractures with assumed fracture spacing, which can be
changed at a later time. Fracture density has significant influence on production and
analysis of results and current fracture density may not be actual representation of in-situ
51
conditions, but the effects of fracture density on history matching will be discussed in
appropriate sections of subsequent chapters.
Later it was found that the wells were stimulated with a single treatment pumped into an
open hole. This was a very unusual, experimental type of treatment. Fracture spacing in
such a treatment is not controllable, but the microseismic monitoring showed that it was
successful in creating multiple fractures with spacing of about 50-60 m. This information
was used later to build the injection models in Chapter 7.
In the simulation models, the well is completed in x-direction in Cartesian co-ordinate
system. The orientation of the fractures from the geomechanical data given in Table 3.3 is
in y-z planes. The fracture height is growing in y-z plane; width in x-direction, while
length in x-y plane. Total length of horizontal section of wellbore is assumed 3300 ft
(1000m), which is an average wellbore length for horizontal wells in that area. The
schematic wellbore and fracture spacing is presented in Figure 4.1.
Figure 4.1 Well completion and fracture spacing – well A and B
52
Where Lf = Fracture half length
d = Average fracture spacing between 2 adjacent fractures, (assumed = 300 ft)
Line of symmetry is taken at fracture # 5 and reservoir model is constructed for a half of
the model comprising of ½ of fracture # 5 and fractures 6 to 9.
4.3 FRACTURES PARAMETERS ESTIMATION
Detailed modeling of fractures for each well based on wells injection data was beyond the
scope of this project. Modeling of fractures using correct amount and type of proppant,
fluid, type of treatment and injection rates and volumes is important and useful for
production analysis, but due to the constraints and uncertainty about the individual
fracture data in the open-hole treatment, it was decided to use simple equations for
fractures parameters estimations based on amount of proppant injected in each well. The
following assumptions are taken for fractures parameters calculations
 Fracture height is equal to pay zone or net pay thickness.
 Propped fracture width of 0.25 and 0.125 inches is assumed for well A and B
respectively.
 It was assumed that the proppants are contained within the main fractures only.
Complex fractures network of secondary fractures in stimulated reservoir volume
is assumed to be created, and these fractures will have conductivity (i.e., are
partially open) but will have no proppants in them.
 Porosity of each sand type is 37%.
 Density of both sand types is 2.65 gm/cc.
53
 Base Line permeability for each type of sand is taken from StimLab data sheet.
Values calculated based on closure stress and bottomhole temperature are given in
Table 4.1.
Table 4.1 Baseline permeability of sand used in well A and B
Property Mesh size Value (well A) Value (well B)
40/70 47800 42500

Permeability, mD
50/140 20000 18000
Sand of two different mesh types are used in each well. Permeability and conductivity of
each sand type in each well is calculated separately by assuming that amount of sand
injected downhole is contained only in the main fracture. The overall permeability and
conductivity of fractures is calculated based on fraction of sand injected in each particular
fracture in a well.
Putting Equation (4.13) into (4.12) and rearranging
Where = Volume of sand (downhole volume occupied by sand), ft3
54
= Conversion constant = 0.035315 (gm*ft3/kg*cc)
= Mass of sand in formation, Kg
= Density of sand, gm/cc
= Downhole volume, ft3
= Sand porosity
= Volume of all fractures, ft3
= Fracture width, inches
= Total fracture length in y-direction, ft
= Fracture half length, ft
= Number of fractures
= Fracture conductivity, mD-ft
= Baseline permeability of sand, mD
Sand of mesh size 40/70 in well A was used for the above sample calculation. The
calculated fracture parameters for each sand type both for well A and B are given in
Tables 4.2 and 4.3.
Sample Calculation for 40/70 type mesh sand in well A:
ft3
ft3
55
ft
= 35.68 ft
mD-ft
Table 4.2 Calculated fracture properties of sands in well A
Property – Well A Value – 40/70 Value – 50/140
Mass of sand in formation, Kg 50600 31300

Fraction of sand in formation 0.618 0.382
Baseline permeability of sand, mD 47800 20000
Downhole volume of sand, ft3 674 417
Total volume of all fractures, ft3 1070 662
Fracture half length, ft 35.68 22.07
Fracture conductivity, mD-ft 995.83 416.67
Closure stress, psi 3100
Fracture height, ft 80
Number of fractures (n) 9
Fracture width, inches 0.25
Sand porosity, fraction 0.37
Density of sand, gm/cm3 2.65
Overall average fracture permeability, mD 37175.58
Overall average fracture conductivity, mD-ft 774
Total fracture half length, ft 58
The overall fracture permeability is calculated by addition of permeability of different
sand types used in treatment. Permeability of the mixture is obtained by weighting
56
baseline permeability of each sand type by the fraction of that sand type in total injected
sand volume.
Table 4.3 Calculated fracture properties of sands in well B
Property – Well B Value – 40/70 Value – 50/140
Mass of sand in formation, Kg 30000 63200

Fraction of sand in formation 0.322 0.678
Baseline permeability of sand, mD 42500 18000
3
Downhole volume of sand, ft 340 842
3
Total volume of all fractures, ft 635 1337
Fracture half length, ft 42.31 89.13
Fracture conductivity, mD-ft 442.71 187.50
Closure stress, psi 3825
Number of fractures (n) 9
Sand porosity, fraction 0.37
3
Density of sand, gm/cm 2.65
Overall average fracture permeability, mD 25886.27
Overall average fracture conductivity, mD-ft 270
Total fracture half length, ft 131
4.4 SUMMARY
 Static fracture is modeled in reservoir simulator by combining the fracture
transmissibility with the transmissibility of reservoir blocks in the fracture plane,
57
without representing fracture by separate grid blocks. This method is known as
the transmissibility modification method.
 Refined grids are used near wellbore and around fracture blocks. Reservoir is
modeled by taking element of symmetry model.
 Total 9 fractures are assumed initially with average fracture spacing of 300 ft.
 Fracture parameters, i.e., half length, average permeability and conductivity are
calculated based on the amount of proppant injected into well and fracture.
58
CHAPTER FIVE: HISTORY MATCHING USING PSEUDO
CONTINUUM APPROACH – UNCOUPLED MODELS
History matching could be very simple or complex depending on nature of problem,
types and number of independent variables involves in it. There are no well defined
procedures and methods to get a history match. There is always more than one method by
which comparable results can be achieved. Then the question arises: which method to
choose and which to discard? The correct approach is to investigate the problem in detail
and put every possible scenario on the table for its solution. One should list all possible
and known variables that are involved and reduce their number by measured laboratory
and estimated data. Thorough investigation of remaining parameters and considering
extra physics of the problem, if needed, will also narrow down the number to a set of
variables that are critical to solution.
Different approaches to model the complex fracture geometry are being developed. Some
involve detailed solution of injection creating a single planar fracture (SPF), which then
interacts with the pre-determined network of fractures (Olson and Wu, 2012; Wu et al.,
2012). Another model is the creation of shear fracture network simultaneously with a
tensile fracture using pseudo-continuum technique (Nassir et al., 2012) or displacement
discontinuity methods (Tao et al., 2009). The approach considered in this study assumes
that the creation/enhancement of fracture system can be represented by pseudo-
continuum with dynamically changing permeability (as a tensor) as a function of
effective stress. The matrix permeability can also be made a function of effective stress
59
and all these functions have a hysteresis in the injection-production cycle. This method is
an extension of the techniques used in tight gas (Settari et al., 2002b, 2009) and has been
applied to Eagleford, Barnett and other shale plays. This method can be used for the
optimization of the treatment because it includes modeling the creation of the stimulated
reservoir volume as well as production modeling. It can be used to optimize the number
of stages and sets of perforations within the stage, treatment size for a stage, use of
different fluids, etc. It has been also used to investigate effects of stress shadowing, well
spacing, and can be further simplified to an uncoupled modeling. In this Chapter of the
thesis, we do not deal with the simulation of the fracturing (stimulation process) itself.
Therefore the size of the SRV and the induced fracture inside each SRV are matching
parameters.
This chapter presents investigation of the role of fracture conductivity, net pay thickness,
permeability multipliers and stimulated reservoir volume on gas rate and cumulative
production. Simulations are run for different cases and results are then compared with
field data. The sensitivity of the results to any particular parameter and deviation of
results from field data are discussed in each section. Assumptions made for some cases to
simplify problem are discussed in that particular section. Figure 5.1 represents flow chart
of simulation runs performed in this chapter. Every previous case serves as base case for
next simulation run.
60
Field Production History Matching
Baseline Propped Fracture Permeability

Case – 5.1.1
Infinite Fracture Conductivity
– An Analytical Approach
Realistic Propped Fracture Permeability
Case – 5.1.2
Effect of Net Pay Thickness (NPT)

Case – 5.2
Effects of Stimulated Reservoir Volume (SRV)
History Matching using Manual Change of

Reservoir Permeability with Time
Figure 5.1 Flow chart of simulation cases – Chapter 5
5.1 ROLE OF FRACTURE CONDUCTIVITY
The conductivity of fracture can be calculated from fracture permeability and width, and
it is represented by the following equation.
Where = Fracture conductivity
= Fracture width
Since in this project fractures are propped, fracture conductivity is controlled by proppant
permeability and propped fracture width. Propped fracture permeability and width are
61
referred to as fracture permeability and fracture width respectively throughout this thesis
unless mentioned otherwise. American Petroleum Institute (API) developed standard
procedures for measuring the conductivity of proppants and these procedures are
documented in API RP-61. Later Much and Penny (1987) recommended minor changes
in API test that resulted in significant improvements in results, which was approved by
International organization for Standardization (ISO) in 2006 as ISO 13503-5. Palisch et
al. (2007) presented comparison between the two tests for high and low quality ceramic
proppant and reported that loss in conductivity measured by modified test was as much as
85% depending on proppant quality and test conditions. Industry has been using the API
modified test since that time and the test is usually referred as “long term” conductivity
test. “Short term” conductivity test is referred to the original procedure of RP-61.
Baseline or Laminar or Reference conductivities reported by service companies and
proppant manufacturers are measured using the long term tests.
5.1.1 Baseline Propped Fracture Conductivity
For our wells, baseline conductivity is calculated using Equation (5.1) by taking proppant
permeability from StimLab database, and assuming fracture width. Case 5.1.1 in Tables
5.1 and 5.2 represents the case where baseline conductivity for fracture is used in
simulation for wells A and B. The other properties in Tables 5.1 and 5.2 are calculated
using equations given in Chapter 4.
62
Table 5.1 Fracture properties of Case 5.1.1 – 2 for well A
Values Values
Property – Well A
(Case 5.1.1) (Case 5.1.2)
Net pay thickness, ft 80
Fracture half length, ft 58
Fracture conductivity, mD-ft 774 39
25081 1301
250081 13001
Table 5.2 Fracture properties of Case 5.1.1 – 2 for well B
Values Values
Property – Well B
(Case 5.1.1) (Case 5.1.2)
13501 651
135001 6501
5.1.2 Realistic Propped Fracture Conductivity
Even though the modified API test (using increasing testing time and variable
temperatures) gives good indication of proppant performance, it still largely
overestimates conductivity values. Several authors have worked in this area and
investigated the factors that could potentially decreases conductivity, which is critical for
reservoir simulation of productivity. Barree et al. (2003), Dedurin et al. (2006), and
Palisch et al. (2007) estimated realistic fracture conductivity by investigating the factors
63
that led to conductivity reduction and explained it in detail. The following are some
factors that according to these authors play a major role in defining actual fracture
conductivity and should be taken in to account in fracturing design treatment.
 Non Darcy flow
 Multiphase flow
 Relative permeability effects
 Proppant concentration
 Gravity segregation and capillary effects
 Filter cake at fracture faces
 Gel damage through proppant pack
 Proppant embedment into fracture face
 Spalling of formation into fracture
 Fines migration
 Cyclic stresses
 Sieve size distribution and proppant types
 Long term proppant degradation, non-uniform proppant distribution and elevated
temperature effects
It has been shown that all these effects (which are not considered in API test procedure)
are cumulative and can reduce fracture conductivity by as much as 99% under realistic
conditions. Schubarth et al. (2006) supported these conclusions in their study.
Montgomery and Steanson (1985) stated that four factors, namely propped fracture area,
propped fracture conductivity, reservoir permeability and drainage area mainly control
64
improvements in productivity. They showed that conductivity degradation continues with
extended time up to 9 months. Conductivity of proppants was continuously reducing until
it reached equilibrium due to particle crushing and rearrangements. Different proppants
may be affected by the above factors in a different way but all of them typically
experience 2 – 3 orders of magnitude of reduction in conductivity. Therefore, estimating
realistic fracture conductivity is the key to realistic predictions of stimulation
effectiveness.
In light of the above discussion, it was decided to use multiplier of 0.05 to take all this
factor into account, i.e., fracture conductivity was reduced by 95%. Case 5.1.2 in Tables
5.1 and 5.2 represent the case where realistic fracture conductivity is used in simulation
for wells A and B by using 5% of baseline fracture conductivity. Transmissibilities for
fracture flow (see chapter 4) are calculated for this case using the modified values of
conductivity. Simulations were run for both cases, i.e., Case 5.1.1 and 5.1.2 and results
are presented in Figures 5.2 and 5.3 for both wells.
65
16 1200
Field Avg Gas Rate Gas Rate - C-5.1.1
Gas Rate - C-5.1.2 Cumulative Gas - C-5.1.1
Cumulative Gas - C-5.1.2 Field Cumulative Gas
Cumulative Gas Prodcution (MMscf)

12 900
Avg Daily Gas Rate (MMscfd)
8 600
4 300
0 0
12-Jun-06 14-Dec-06 18-Jun-07 20-Dec-07 23-Jun-08 26-Dec-08 29-Jun-09 1-Jan-10
Figure 5.2 Comparison of field data and simulation results of Cases 5.1.1 - 2 - well A
12 1200
Gas Rate - C-5.1.2 Field Cumulative Gas
Cumulative Gas - C-5.1.1 Cumulative Gas - C-5.1.2

9 900
6 600
3 300
0 0
16-Aug-06 8-Feb-07 3-Aug-07 26-Jan-08 21-Jul-08 13-Jan-09 8-Jul-09 1-Jan-10
Figure 5.3 Comparison of field data and simulation results of Cases 5.1.1 - 2 - well B
66
The predicted gas rate and cumulative gas is much smaller than field results for both
wells, but the difference between the results of the two cases is very small for well A
while it is a bit larger for well B. Since fracture width for both wells is kept constant
throughout and only permeability is changed, it means that changing proppant
permeability does not affect the results significantly. It can be concluded from these runs
that fracture for current project are acting almost as “infinite conductivity fractures” (the
term will be discussed next). Sharp decline in gas rate in early time for either well is due
to large pressure drop near the wellbore and around fractures.
5.1.3 Infinite Fracture Conductivity – An Analytical Approach
Dimensionless fracture conductivity or capacity, Fcd, is the ratio of the ability of the
fracture to transmit fluids down to fracture and into wellbore to the ability of the
formation to deliver fluids into the fracture. It is defined as
Where = Dimensionless fracture conductivity
= Fracture width
= Formation permeability
= Fracture half length
Dimensionless fracture conductivity has been in use since the earliest days of hydraulic
fracturing. McGuire and Sikora (1960) plotted productivity index against relative
67
conductivity, dimensionless fracture conductivity, for various fracture half length. They
showed that productivity increases 2 to 6 times in range of 103 – 104 of relative
conductivity while the increase is negligible for larger values of relative conductivity.
Prats (1961) concluded that the effective wellbore radius is always equal to a half of
fracture half length for an infinite conductivity fracture and for Fcd > 10 increase in
effective wellbore radius is small and fracture behaves as an infinite conductive fracture.
Dedurin et al. (2006) in their paper confirmed his conclusions. It is typical that
dimensionless fracture conductivity has higher values than predicted by Prats for steady
state production. Cinco-Ley and Smaniego-V (1981) performed pressure transient
analysis on vertically fractured well producing at constant flow rate in an infinite,
isotropic, and homogenous reservoir and showed that for dimensionless time greater than
1, the Fcd value must be 30. They and Barree et al. (2003) argued that for Fcd > 30
fractures can be represented as an infinite conductivity fractures. Pearson (2001)
suggested Fcd value of 10 for infinite conductivity based on Prats’ work and industrial
experience and concluded that realistic conductivity should be utilized when calculating
Fcd. It is an important parameter in the design of hydraulic fracture treatments and it is a
tool that helps engineers in predicting benefits of fracture simulation.
An infinite conductivity fracture of half-length Lf can be, in pseudo-steady state
conditions, replaced by an effective well bore radius or skin using the following relation,
Skin from Equation (5.3)
68
Where = Effective wellbore radius
= Wellbore radius
= Skin
Positive skin effect means there is an increase in pressure drop, which indicates extra
flow resistance near the wellbore while negative skin effect means decrease in pressure
drop, which is characterized by flow enhancement near wellbore. Effective wellbore
radius increases with negative skin, i.e., hydraulic & acid fracturing while it decreases
with positive skin, i.e., mud invasion, drilling fluid loss, and near wellbore formation
damage.
Dimensionless fractures conductivity concept will be used here to plot dimensionless
effective wellbore radius against Fcd. It is proved from simulation results that fractures
are almost infinite conductive fractures. Criterion suggested by Prats, Dedurin and
Pearson (i.e., fractures are of infinite conductivity if Fcd ≥ 10) is used here to analytically
predict whether fractures have finite or infinite conductivity. For this purpose
dimensionless fracture conductivity, effective wellbore radius, and skin is calculated
using Equations (5.2, 5.3 and 5.4) and is given in Table 5.3 for baseline and realistic
fracture conductivity cases for both wells. The dimensionless effective wellbore radius is
plotted against dimensionless fracture conductivity in Figure 5.4.
Table 5.3 Dimensionless fracture conductivity of Cases 5.1.1 - 2 for well A and B
Property Values – Well A Values – Well B
Wellbore radius, ft 0.25

Fracture width, inches 0.25 0.125
Reservoir permeability, mD 0.015 0.010
69
Fracture half length, ft 58 131
Case 5.1.1
Fracture permeability, mD 37175.60 25886.30
Skin, - 4.75 - 5.56
Dimensionless fracture conductivity 890 205
Case 5.1.2
Fracture permeability, mD 1858.80 1294.32
Skin, - 4.75 - 5.56
Dimensionless fracture conductivity 44.51 10.30
Well A Well B
1
Dimensionless effective wellbore radius
(rwe/Lf)
0.1
0.01
0.01 0.1 1 10 100 1000
Dimensionless fracture conductivity (Fcd)
Figure 5.4 Dimensionless fracture conductivity of Cases 5.1.1 - 2 – Well A and B
Dimensionless effective wellbore radius for each well is equal to 0.5 and dimensionless
fracture conductivity is large than 10 for all cases. This analysis confirms our earlier
conclusion that fractures in this project is infinite conductivity fractures. Case 5.1.2 for
both wells will be the base case for next simulation runs.
70
5.2 EFFECT OF NET PAY THICKNESS (NPT)
Net pay thickness used in earlier cases was an assumed value of 80 ft discussed before.
Sponsor company provided us with more information on well and geology in a recent
meeting. It was decided to run a case using updated value of net pay thickness of 100 ft.
Increasing net pay thickness means larger pore volume and eventually higher production.
Case 5.2 shown below is considered as base case for further cases and investigation.
Summary of fracture parameters and transmissibility multipliers both for well A and B
for Case 5.2 is given in Table 5.4.
Table 5.4 Fracture properties of Case 5.2 for well A and B
Property (Case 5.2) Values – Well A Values – Well B

1301 651
13001 6501
71
16 1200
Gas Rate - C-5.2 Cumulative Gas - C-5.1.2
Cumulative Gas - C-5.2 Field Cumulative Gas
12 900

8 600
4 300
0 0
Figure 5.5 Comparison of field data and simulation results-Cases 5.1.2 & 5.2 -well A
12 1200
Gas Rate - C-5.2 Field Cumulative Gas
Cumulative Gas - C-5.1.2 Cumulative Gas - C-5.2

9 900
6 600
3 300
0 0
Figure 5.6 Comparison of field data and simulation results-Cases 5.1.2 & 5.2 - well B
72
Figures 5.5 and 5.6 represent comparison of field data and simulation results of current
case for well A and B respectively. Production in early time is mostly from the fracture
and therefore base case and current case results are similar in early time. Increase in net
pay thickness means larger pore volume hence more gas in place and larger production
compared to the base case.
5.3 EFFECTS OF STIMULATED RESERVOIR VOLUME (SRV)
Hydraulic fracturing technology has improved immensely since the adoption of this
process. The full understanding and optimization of hydraulic fracture simulation will not
be possible until accurate methods are available to directly image the fracture
characteristics. The only methods available for in-situ imaging are the use of downhole
receiver and tiltmeter arrays for detection and monitoring of fracture induced
microseismic events and deformations. Microseismic events (MS) are primarily shear
movements of a limited area of rock surface. Whether MS can also detect tensile events is
still somewhat controversial. A pure shear movement emits both compressional waves (p-
waves) and shear waves (s waves) and detection of p-wave and s-wave arrivals at
observation well provide the information to accurately locate the source point of the
event. A 3-dimensional location for any particular event can be found using single
vertical array of receivers and locations of all these events are then used to generate an
image of a growing fracture. Exact understanding of causes of these events which is
necessary to interpret the measurements is still evolving as these shear failure events can
be found hundreds of feet away from fractures and even outside the stimulated zone.
73
It is now believed that apart from major fractures there is also a network of small
fractures and secondary fractures being created during hydraulic fracturing treatments.
The leak-off of fracturing fluid into the formation perpendicular to fracture face is
contributing to their development, but they can extend beyond the fluid penetration. The
region in which the secondary fracturing was created is often called Stimulated Reservoir
Volume (SRV). These secondary fractures could be the opening of already existed
natural fractures in the formation or it could be newly created network of open fractures
due to shear failure of rock. These secondary fractures are highly conductive open
channels because these are open fractures. It is possible that early time production of our
wells could be due to the contribution of main fracture and network of these open
fractures and late time production curves are comparatively flat due to the closure of
these fractures.
To measure the effects of these secondary fractures on production, few cases assuming
the presence of SRV were run and results were then compared with base case. The
volume around the main fracture and wellbore where open secondary fractures are
concentrated (i.e., the SRV) is shown schematically on Figure 5.7. In left side (a) of the
figure the schematic of fracture and SRV created downhole in reservoir is shown, while
the right side (b) shows the representation of the fracture and SRV by a rectangular
region in reservoir simulator. To model increase of permeability inside SRV, a multiplier
which we will call an initial permeability multiplier (IPM) is used. It means that
permeability of rock inside SRV is enhanced by some factor due to fracture development.
This permeability can then be decreased as a function of the depletion of the reservoir, as
will be discussed in Chapter 6.
74
Figure 5.7 Stimulated reservoir volume (SRV) schematic – Complex fracture
geometry representation
Calculated transmissibility and assumed values of SRV and initial permeability
multipliers for all sub-cases in each well are presented in Table 5.5. 6,18 25, 35, and 40 in
Table 5.5 represent IPM and numbers in bracket followed IPM values represent modified
reservoir permeability after multiplying IPM. The method of transmissibility calculation
and formulation has been given in chapter 4.
Table 5.5 SRVand initial permeability multiplier of Cases 5.3.1 – 6 for well A and B
Property (Cases 5.3.1 - 6) Values – Well A Values – Well B

Case 5.3.1
218 109
2168 1084
SRV (one wing), ft3 40x3000x100 40x3000x100
Initial permeability multiplier 6 (0.09, 0.009) 6 (0.06, 0.006)
Case 5.3.2
218 109
2168 1084
75
Case 5.3.3
218 109
2168 1084
Case 5.3.4
73 20
723 187
Case 5.3.5
73 20
723 187
Case 5.3.6
53 17
521 164
Comparison between simulation results and field data for Cases 5.3.1 to 5.3.6 are
presented in Figure 5.8 and 5.9 for well A and in Figure 5.10 and 5.11 for well B.
76
16 1600
Field Avg Gas Rate Gas Rate - C-5.2
Gas Rate - C-5.3.1 Gas Rate - C-5.3.2
Cumulative Gas - C-5.2 Cumulative Gas - C-5.3.1

12 1200
8 800
4 400
0 0
16 2800
Gas Rate - C-5.3.5 Gas Rate - C-5.3.6
Field Cumulative Gas Cumulative Gas - C-5.3.4
12 2100

8 1400
4 700
0 0
77
12 1500
Gas Rate - C-5.3.1 Gas Rate - C5.3.2
Cumulative Gas - C-5.2 Cumulative Gas - C-5.3.1
1200
9

900
600
3
300
0 0
Figure 5.10 Comparison of field data and simulation results of Cases 5.3.1 - 3-well B
12 2800
Gas Rate - C-5.3.5 Gas Rate - C5.3.6

9 2100
6 1400
3 700
0 0
Figure 5.11 Comparison of field data and simulation results of Cases 5.3.4 - 6 -well B
78
We will first discuss the first three cases, i.e., Cases 5.3.1 - 3 given in Figure 5.8 for well
A and in Figure 5.10 for well B. In these three cases the width of the SRV is varied from
40 to 100 ft while the same initial permeability multiplier is used to enhance permeability
of reservoir rock inside SRV. The results of these cases are quite similar in early time but
difference is quite large for late time especially for the first two cases. The flow regions
can be divided into three segments. Region I represent is flow through propped fracture
that transfers fluid into wellbore, region II is the SRV which is transmitting reservoir
fluids to fractures, and region III is the flow in the reservoir itself that is feeding to the
SRV. The 2-dimensional top view of all the three regions around a fracture is shown in
Figure 5.12.
Figure 5.12 Areal view of fracture in x-y plane – Flow pattern representation
By definition of hydraulic diffusivity, the greater is the permeability, the larger is the rate
of pressure propagation. Pressure propagation will be faster in region I and II and
comparatively slower in region III. Early time higher production is mainly from SRV.
79
When pressure propagation reaches SRV boundary, then production declines due to
smaller reservoir permeability. The time at which production curves deviate from each
other, such as between Cases 5.3.2 and 5.3.3, indicates the time taken by pressure to
reach the boundary of region III. These cases give us some indication that for higher early
time production, we either need larger initial permeability multiplier or larger SRV width
or combination of both.
Case 5.3.4 is run to achieve early time history match by varying both width of SRV and
initial permeability multiplier. Early time history match is achieved with parameters
given in Table 5.5 and results are in Figure 5.9 and 5.11 both for well A and B
respectively. The increase in permeability multiplier for well A is 3 times and 6 times for
well B as compared to Case 5.3.3 and with little increase in SRV width. Now the task is
to select an optimum length of SRV and also match the results at late time. For this
purpose two more cases are run, i.e., Case 5.3.5 and 5.3.6 by varying both length of SRV
and initial permeability multiplier. The summary of these cases is given in Table 5.5
while results for those cases are presented in Figure 5.9 for well A and 5.11 for well B.
All the parameters in Case 5.3.4 and 5.3.5 are same except for the length of SRV but
there is huge difference in results. The same argument as discussed before can be applied
here to explain the difference in results and decrease in production. If one analyzes the
final case, i.e., Case 5.3.6 for any well, it can be said that although results are improved in
early time, simulation over predicts results at late time. The reason is that the
permeability multiplier is constant throughout the simulation, but in reality permeability
inside the SRV and outside in the formation decreases with reservoir pressure depletion
80
(or with increase in net effective stress). Thus permeability at late time is much smaller
than permeability in early time due to reservoir deformation.
It is therefore concluded that history match with field data is not possible by using
constant permeability multiplier or constant permeability overall. Pressure / stress
dependent permeability is necessary to obtain a history match, which will be the topic of
discussion in subsequent chapters. As Case 5.3.6 provide a reasonably good match with
field data, this case will be used as a base case for further runs.
5.4 HISTORY MATCHING USING MANUAL CHANAGE OF RESERVOIR
PERMEABILITY WITH TIME
This case is run to authenticate our conclusions of need for using pressure / stress
dependent permeability. There are different ways to change permeability with respect to
time. One of them is to use time dependent transmissibility. As permeability is contained
in the transmissibility equation, modifying transmissibility while keeping other
parameters constant means modifying permeability with respect to time. Field data
analysis also gives us some indications that permeability is initially higher and decreases
continuously with time. It is vital to first test this concept, which can be done in the TRS
software by changing the permeability at specified time intervals manually. Different
permeability multiplier, value smaller than 1, is applied both inside and outside SRV at
specified time intervals to get history match, which is represented in Figure 5.13 for Well
A. The true pressure / stress dependent permeability effects will be considered and
modeled in Chapter 6.
81
Table 5.6 Fracture properties of Case 5.4 for well A and B
Property (Case 5.4) Values – Well A Values – Well B

Total number of fractures 9
Initial Permeability Multiplier 25 (0.375, 0.0375) 40 (0.06, 0.006)
Figure 5.14 and 5.15 represent comparison of field data and results from simulation for
current case both for well A and B respectively. It is apparent that it is possible to
duplicate the history very accurately. This method is of course purely empirical, but
provides motivation for including geomechanics.
1
Inside SRV
outside SRV
0.8
Permeability Multipliers
0.6
0.4
0.2
0
Figure 5.13 Time dependent permeability multipliers-History matched case–Well A
82
16 1200
Cumulative Gas - C-5.4 Field Cumulative Gas

12 900
8 600
4 300
0 0
Figure 5.14 History matching of well A using time dependent permeability
12 1200
Field Cumulative Gas Cumulative Gas - C-5.4
9 900

6 600
3 300
0 0
Figure 5.15 History matching of well B using time dependent permeability

83
5.5 SUMMARY
 Using baseline conductivity in simulation and other calculations will always
overestimate results; therefore realistic proppant conductivity should be used.
 Fractures have infinite conductivity for both wells for this project, as it was
proved both by running simulations and comparing production curves and
dimensionless fracture conductivity calculations.
 Net pay thickness has no effects on production in early time but production
increases at late time due to larger reservoir size.
 Stimulated reservoir volume has significant effects on results. Permeability inside
SRV must be larger than formation permeability by factor greater than 2.0.
 Early time history match can be achieved by using permeability multiplier in SRV
and constant permeability in outside formation but it is difficult to match the late
time results using the same parameters.
 History match is impossible without the use of pressure / stress dependent
permeability. The last case run verified this conclusion and it proved (from results
in Figure 5.14 and 5.15) that history match is only possible by introducing some
form of time changing permeability.
It has been established in this chapter that history match without the use of pressure /
stress dependent permeability is not possible. History matching using pressure / stress
dependent permeability using both uncoupled reservoir and a fully coupled
geomechanical model will be discussed in detail in Chapter 6.
84
CHAPTER SIX: HISTORY MATCHING AND MODELING OF
GEOMECHANICAL EFFECTS USING PRESSURE / STRESS
DEPENDENT PERMEABILITY
It was proved and concluded in Chapter 5 that history match is not possible without the
use of time dependent transmissibility or permeability. A function is needed that defines
permeability as function of effective stress or pressure. Stress dependent permeability
always means effective stress dependent permeability in this thesis and pressure
dependent permeability means that the independent variable is pressure, which is
equivalent to the assumption that total stresses are constant. Average (mean) effective
stress is defined as
where = Mean effective stress
= Total mean stress
= Biot’s constant
= Fluid pressure
= Vertical stress
= Maximum horizontal stress
= Minimum horizontal stress
85
In Equation (6.1) is part of total mean external stress carries by solid matrix
(framework) and the remaining part, , is carried by the fluid. The equation can also
be written for one dimensional system in which mean stress will be replaced by
corresponding stress values. Permeability is in general a function of effective stress
which means that it is function of both total stress and fluid pressure. However, for
simulation it can be also treated approximately as a function of total stress only (which is
rare) or fluid pressure only depending on which is kept as an independent variable.
where = Permeability
= Initial total mean stress
Equation (6.3) represents a case where permeability is function of both total mean stress
and fluid pressure and in (6.4) it is a function of fluid pressure only.
This chapter presents the correlation used to generate pressure / stress dependent
permeability data, history matching using this data, stress sensitivity of formation, effect
of shape of pressure / stress dependent permeability multipliers on results, comparison of
results between uncoupled and fully coupled geomechanical models and approximation
of geomechanical effects in conventional reservoir simulator using uniaxial deformation
theory. Cases run for detailed investigation of various parameters and results are
compared to field data. The conclusions and discussions are presented at end of each
particular section. Figure 6.1 represents flow chart of simulation runs performed in this
chapter. Every previous case serves as base case for next simulation run.
86
Field Production History Matching using
Pressure / Stress Dependent Permeability
Case – 6.1 – 15 Fractures – Uncoupled

Production Models
Effects of Shape of Pressure Dependent

Permeability Curves – Uncoupled
Production Models – Cases – 6.3.1.1–4
Effects of (SRV) – Uncoupled

Production Models – Cases – 6.3.2
ase – 5.2
History Matching using Pressure History Matching using Pressure

Dependent Permeability outside SRV Dependent Permeability both Inside
Cases – 6.3.3.1 – 2 and Outside SRV – Cases – 6.3.5
History Matching using Stress Dependent Permeability both Inside and

Outside SRV – Coupled Production Geomechanical Models – Constant
Mean Total Stress Cases (Case – 6.5)
Approximation of Geomechanical Effects in Conventional

Reservoir Simulator using Uniaxial Deformation Theory
History Matching using Stress Dependent Permeability both Inside and

Outside SRV – Uncoupled and Coupled Production Geomechanical
Models – Variable Mean Total Stress Cases (Case – 6.6)
87
6.1 DISCUSSION ON NEW INFORMATION
In a recent meeting with sponsored company representatives some more information
about in-situ fracture density and fracture spacing was given. According to microseismic
report from Apache Corporation, average in-situ fracture spacing is 55 m (180 ft) while
fracture half-length is 250 m (820 ft). It was decided to rework all the fracture parameters
calculation shown in Chapter 4 in light of the new data. For the same amount of volume,
the larger the number of fractures, the smaller is fracture length given by Equation (4.14).
A new base case to compare other runs with in this Chapter was created by choosing
Case 5.3.4 as a starting point (the reason for choosing this case is early time relatively
good match). Changing fracture density changes fracture half length and as a result
smaller initial permeability multiplier is needed to get the same results as Case 5.3.4. The
purpose of this new base case is to get the same results as the Case 5.3.4, so that the
current base case uses new updated information. Calculated rock and fracture properties
that are used as input in reservoir simulator for the new base case 6.1 are summarized in
Table 6.1 and 6.2 for well A and B respectively.
Table 6.1 Fracture properties of Case 5.3.4 and 6.1 for well A
Values Values
Property – Well A
(Case 5.3.4) (Case 6.1)
Fracture conductivity, mD-ft 39
SRV (one wing), ft3 120x3000x100
88
Number of fractures, n 9 15
73 131
723 1301
Table 6.2 Fracture properties of Case 5.3.4 and 6.1 for well B
Values Values
Property – Well B
(Case 5.3.4) (Case 6.1)
3
SRV (one wing), ft 140x3000x100
Number of fractures, n 9 15
20 55
187 543
Figures 6.2 and 6.3 represent the comparison of the old base case and Case 6.1 and also
with field data.
89
16 2800
Field Avg Gas Rate Gas Rate - C-5.3.4 Gas Rate - C-6.1
Field Cumulative Gas Cumulative Gas - C-5.3.4 Cumulative Gas - C-6.1

12 2100
8 1400
4 700
0 0
Figure 6.2 Comparison of field data and simulation results of Case 6.1 - well A
12 2800
Field Avg Gas Rate Gas Rate - C-5.3.4 Gas Rate - C-6.1
Field Cumulative Gas Cumulative Gas - C-5.3.4 Cumulative Gas - C-6.1

9 2100
6 1400
3 700
0 0
Figure 6.3 Comparison of field data and simulation results of Case 6.1 - well B
90
6.2 PRESSURE / STRESS DEPENDENT PERMEABILITY CORRELATION
The permeability dependence of rock and fractures on stress and fluid pressure in pores is
well documented in literature. The permeability of oil or gas bearing formations depends
on effective stress. The dependence is larger for consolidated sandstones. The detailed
discussions on pressure / stress dependent permeability are given in Chapter 2 along with
mathematical correlations derived by various authors. For the modeling of geomechanical
effects in this Chapter, the correlation presented by Jones and Owens (1980) was selected
and is given by Equation (2.8). According to their correlation data must be corrected for
1,000 psi effective pressure (stress) and stress factor, S, can be found by plotting
laboratory data on semi-log plot. Since for our project there is no laboratory data
available and stress dependent permeability curves are used as history matching
parameter, Equation (2.8) needs to be modified as follows.
where = Permeability at corresponding mean effective stress
= Permeability at initial mean effective stress
= Stress factor
= Mean effective stress
= Initial mean effective stress
Initial and corresponding mean effective stress is found using Equation (6.1). For
pressure dependent permeability condition stated in Equation (6.4) is used to generate
data that depends on fluid pressure by keeping the total mean stress constant throughout.
91
It is assumed that stresses in the reservoir are not changing due to change in pressure and
remain constant during production. Taking this assumption and using Equations (6.1) and
(6.5) pressure dependent permeability multipliers are calculated for different values of the
stress factor S.
6.3 HISTORY MATCHING USING PRESSURE DEPENDENT PERMEABILITY
6.3.1 Effects of Shape of Pressure Dependent Permeability Curves (Stress Factor, S)
Using Case 6.1 as a base case, few sub-cases were run to see the effects of the shape of
pressure dependent permeability multipliers, which is defined by the stress factor (S), on
gas rate and cumulative production. As fracture conductivity is taken as the realistic
fracture conductivity by considering all factors explained in Chapter 5, no pressure
dependent permeability multiplier is applied to blocks representing fractures. The initial
permeability is found at reference pressure of 2000 and 2200 psi for well A and B
respectively. Pressure dependent permeability curves of different shapes are generated
using range of values of stress factor by using Equation (6.5). The larger the value of S,
the larger is the permeability reduction and reservoir is more stress sensitive. It is
important to note that the same pressure dependent function is used in the whole reservoir
without distinguishing between high and low stress dependent regions. Summary of
assumed and calculated rock and fracture properties, which are discussed in more detail
in Chapter 4, for wells A and B are presented in Table 6.1 and 6.2 respectively. All the
properties that are discussed in base case (i.e., Case 6.1) will remain the same and in
92
addition pressure dependent multipliers with different stress factors are used. Stress
factors for sub Cases 6.3.1.1 to 6.3.1.4 are given in Table 6.3.
Table 6.3 Stress factor (S) of Cases 6.3.1.1 - 4 for well A and B
Property (Cases 6.3.1.1 - 4) Values – Well A Values – Well B
Case 6.3.1.1
Stress factor (Pi, psi = 2000 (A), 2200 (B)) 0.33
Case 6.3.1.2
Case 6.3.1.3
Case 6.3.1.4
Stress factor (Pi, psi = 2000 (A), 2200 (B)) 3.5 3.8
Pressure dependent permeability functions generated from stress factor values given in
Table 6.3 are presented in Figure 6.4. Figures 6.5 and 6.6 show the comparison of sub
cases in this current section with base case (i.e., Case 6.1) and field data followed by a
discussion on the results.
93
1.0
0.8
0.6 S=0.33 - Well A

S=1.7 - Well A
K(p)/Ki
S=3.3 - Well A
S=3.5 - well A
0.4
S=3.8 - Well A
S=0.33 - Well B
S=1.7 - Well B
0.2
S=3.3 - Well B
S=3.5 - Well B
S=3.8 - Well B
0.0
0 400 800 1200 1600 2000 2400
Fluid Pressure (psi)
Figure 6.4 Pressure dependent permeability curves of Cases 6.3.1.1 – 4 - well A & B
16 2800
Gas Rate - C-6.3.1.1 Gas Rate - C-6.3.1.2
Cumulative Gas - C-6.3.1.1 Cumulative Gas - C-6.3.1.2

12 2100
8 1400
4 700
0 0
Figure 6.5 Comparison of field data and simulation results of Cases 6.3.1.1-4-well A
94
12 2800
9 2100

6 1400
3 700
0 0
Figure 6.6 Comparison of field data and simulation results of Cases 6.3.1.1-4-well B
The trend of the results shown in Figure 6.5 and 6.6 is similar except the last case. It is
showing very little deviation in very early time but the gap in results is starting to
increase after that. As the value of stress factor increases, permeability multiplier function
decreases leading to smaller gas rate and production. In all these cases permeability
multipliers are decreasing from 1 at initial pressure to some smaller value at minimum
bottom hole pressure constraint for well flowing pressure. The overall average reservoir
pressure is also declining resulting in smaller production due to corresponding overall
decrease of permeability. There is considerable difference in permeability multipliers of
well A and B at the same pressure; difference is larger at higher pressure. Now question
would be which curve to choose to define stress sensitivity of reservoir for this field?
Keeping in mind the unconventional nature of tight gas sands and high stress sensitivity
95
of tight reservoirs, Case 6.3.1.3 for well A and Case 6.3.1.4 for well B is chosen as the
base case for further investigation. Although both wells are producing from the same
formation and stress sensitivity will not be that different for each of them, different cases
have been selected as an initial guess.
6.3.2 Effects of Stimulated Reservoir Volume (SRV)
The effects of stimulated reservoir volume were already studied in detail in Chapter 5.
Case 5.3.6 presented in Table 5.5 gives very good match in early time. The purpose to
run a case in this section is to apply the SRV size assumed in Case 5.3.6 to base case,
which is selected for this section. It is important to mention that pressure dependent
function is applied both inside and outside SRV for the current simulation run. Table 6.4
and 6.5 summarized fracture and rock properties and also present comparison between
data sets for current case and base case for well A and B respectively.
Table 6.4 Fracture properties of Case 6.3.2 for well A
Values Values
Property – Well A
(Case 6.3.1.3) (Case 6.3.2)
Initial permeability multiplier 10 (0.15, 0.015)
Number of fractures, n 15
131
1301
96
Stress factor (Pi, psi = 2000 (A)) 3.3
Table 6.5 Fracture properties of Case 6.3.2 for well B
Values Values
Property – Well B
(Case 6.3.1.4) (Case 6.3.2)
55
543
Stress factor (Pi, psi = 2200 (B)) 3.8
Comparison of results between Case 6.3.2 and corresponding base case for each well is
presented in Figures 6.7 and 6.8. The average fracture spacing is 180 ft, so to avoid any
fracture interference and to get maximum width created by microseismic events; the SRV
width should be smaller than average fracture spacing. Although the production of this
current case is smaller than in the base case, the shape of curve both at early time and late
resembles to field data. The SRV is chosen from Case 5.3.6 in which good match at early
time was achieved. Case 6.3.2 is selected as base case for both wells and will be used to
compare results.
97
16 1600
Field Avg Gas Rate Gas Rate - C-6.3.1.3
Cumulative Gas - C-6.3.1.3 Cumulative Gas - C-6.3.2

12 1200
8 800
4 400
0 0
Figure 6.7 Comparison of field data and simulation results of Case 6.3.2 - well A
12 1600
Cumulative Gas - C-6.3.1.3 Cumulative Gas - C-6.3.2

9 1200
6 800
3 400
0 0
Figure 6.8 Comparison of field data and simulation results of Case 6.3.2 - well B
98
6.3.3 History Matching Using Pressure Dependent Permeability outside SRV
The effects of varying stress factor and stimulated reservoir volume was studied so far,
and both parameters had significant impacts on results. Until this point all the cases in
this section are run by applying pressure dependent permeability multiplier to the whole
reservoir. The secondary fractures in SRV are believed to be open initially but closing
rapidly in early time of production to some residual value. So permeability dependence
on pressure will be larger inside the SRV than outside of it. Cases in this section are run
by applying pressure dependent multipliers to the whole reservoir except inside SRV.
This may seem contradictory to what was stated in earlier chapters and sections. But
these cases were investigated and included here to highlight the importance of using
correct approach to get history match. High production both in early time and late time is
expected by removing pressure dependent permeability multipliers inside SRV so
constant (higher) permeability is used inside SRV.
Same fracture and rock properties given in Table 6.4 and 6.5 are used in this case. The
first case in this section, which is Case 6.3.3.1, is exactly similar to base case, i.e., Case
6.3.2 except there is no pressure dependent permeability inside the SRV in the former.
Figures 6.9 and 6.10 present result of field data, base case, and Case 6.3.3.1 for well A
and B respectively.
99
16 1200
Gas Rate - C-6.3.3.2 Field Cumulative Gas
Cumulative Gas - C-6.3.2 Cumulative Gas - C-6.3.3.1

12 900
8 600
4 300
0 0
Figure 6.9 Comparison of field data and simulation results of Case 6.3.3.1 - well A
12 1500
Gas Rate - C-6.3.3.1 Field Cumulative Gas
Cumulative Gas - C-6.3.2 Cumulative Gas - C-6.3.3.1
1200
9

900
600
3
300
0 0
Figure 6.10 Comparison of field data and simulation results of Case 6.3.3.1 - well B
100
As both wells are drilled in same formation, one would expect that the permeability
dependence on pressure or stress should be the same. Case 6.3.3.2 is run both for well A
and well B. Its purpose is to use same stress factor in both wells and improve results for
well B in late time because it was deviating at late time as shown in Figure 6.10. It is
important to recall that pressure dependent permeability multipliers (PDPM) are applied
to the whole reservoir excluding SRV. New set of parameters of fracture and rock for
well A and B are given in Table 6.6. The results of this case are presented in Figures 6.11
and 6.12 for well A and B respectively.
Table 6.6 Fracture properties of Case 6.3.3.2 for well A and B
Property (Case 6.3.3.2) Values – Well A Values – Well B

109 27
1084 261
Stress factor outside SRV, S 3.8 3.8
101
16 1200
Field Cumulative Gas Cumulative Gas - C-6.3.3.2

12 900
8 600
4 300
0 0
Figure 6.11 History matched case using PDPM outside SRV - well A
12 1200

9 900
6 600
3 300
0 0
Figure 6.12 History matched case using PDPM outside SRV - well B
102
Excluding permeability multipliers inside the SRV improved both gas rate and
cumulative production as shown in Figures 6.11 and 6.12. One interesting observation
from these figures is the similar production at very early time. Larger permeability inside
SRV than in the rest of the formation results in higher production in early time; at late
time production is controlled by region outside SRV so using permeability multiplier in
that region actually brings down production and results in good match.
Initial permeability multiplier of 12 and 25 is used to get match for well A and B
respectively while the microseismic events length (SRV length) was decreased by 100 ft
to get history match, which means that decrease in SRV length is compensated by
increasing initial permeability multiplier. Late time results are improved for well B by
implementing those changes, as can be confirmed by comparing Figures 6.10 and 6.12.
6.3.4 Effect of Native Reservoir Permeability – Well B
Simulation results shown in Figure 6.11 of Case 6.3.3.2 for well B are 3.80% larger than
field data at the end. This seems a reasonable match for current field data but difference
will grow if the simulation is run for longer time. So far all history matching parameters
were investigated but the model still needs further improvements and tunings. Since well
B is deeper than well A, it could be possible that native reservoir permeability initially
selected for well B is different. For this purpose few cases were run and effects of native
reservoir permeability on production were investigated and evaluated. The new reservoir
permeability is decreased by factor of 2, which gives 0.005 mD. Case 6.3.3.2 is selected
103
as base case for these simulation runs. The only difference between Case 6.3.3.2 and
Case 6.3.4.1 is use of smaller (by factor of 2) native reservoir permeability.
Case 6.3.4.2 is run by taking Case 6.3.4.1 as base case in which microseismic length and
initial permeability multiplier is modified to get history match. Fracture properties and
summary of rock properties used in all of the above discussed cases are presented in
Table 6.7. Figure 6.13 represents comparison of results between base case and sub cases
from this section. Figure 6.14 shows history matched case result with field data using
pressure dependent permeability outside SRV and modified native reservoir permeability.
Starting with comparison of base case and Case 6.3.4.1 using smaller permeability will
decrease production, which is obvious and it is shown in Figure 6.13. The entire
production curve shifts downward, which is then pushed upward by increasing
microseismic length from 400 to 500 ft and using larger permeability multiplier for Case
6.3.4.2. Initial permeability capture early time production as well as late time but major
part of late time production is increased by increasing microseismic length.
Table 6.7 Fracture properties of Cases 6.3.4.1 - 2 for well B
Values Values Values

Property – Well B
(Case 6.3.3.2) (Case 6.3.4.1) (Case 6.3.4.2)
Stress factor outside SRV, S 3.8
25 25 50
Initial permeability multiplier
(0.25, 0.025) (0.125, 0.0125) (0.25, 0.025)
Reservoir permeability, mD 0.01, 0.001 0.005, 0.0005
104
27 53 27
261 521 261
SRV (one wing), ft3 140x400x100 140x400x100 140x500x100
12 1200
9 900

6 600
3 300
0 0
Figure 6.13 Comparison of field data and simulation results of Case 6.3.4.1-2-well B
The results for history matched case model by decreasing native reservoir permeability
given in Figure 6.14 are much improved compared to previous results shown in Figure
6.12.
From all these simulation runs and exercises, it is deduced that similar results can be
achieved by varying various parameters but question remains about the validity of the
various scenarios that were used to force a match which do not account for all physics
actually happening in reservoir. History matching so far excluded pressure dependent
105
permeability in SRV. Pressure dependent permeability inside SRV cannot be ignored
and will be considered next.
12 1200
9 900

6 600
3 300
0 0
Figure 6.14 History matched case using PDPM outside SRV and modified reservoir
permeability - well B
6.3.5 History Matching Using Pressure Dependent Permeability both Inside and
Outside SRV
In section 6.3.4 pressure dependent permeability was used outside SRV only and the
purpose of simulation runs in this section is to get history match with field data using
pressure dependent permeability both inside and outside SRV. Current case represent
more realistic approach than cases run in section 6.3.4 as smaller permeability multipliers
are applied inside SRV and larger outside SRV.
106
Fracture and rock properties for Case 6.3.5 and previously history matched case, i.e.,
Case 6.3.3.2 and 6.3.4.2 for well A and B respectively, are summarized in Table 6.8. The
pressure dependent curves for history matched model case both for well A and B with
modified stress factor are presented in Figure 6.15. The results of current case and field
data are shown in Figures 6.16 and 6.17 for well A and B respectively.
Table 6.8 Fracture properties of Case 6.3.5 both for well A and B
Values Values
Property – Well A
(Case 6.3.3.2) (Case 6.3.5)
Stress factor Inside SRV, S No Multipliers 3.9
109 27
1084 261
Values Values
Property – Well B
(Case 6.3.4.2) (Case 6.3.5)
Reservoir permeability, mD 0.01, 0.001 0.005, 0.0005
Stress factor Inside SRV, S No Multipliers 3.9
27 14
261 131
107
Mean Effective Stress (Psia)
3500 4000 4500 5000 5500 6000
1
S=2.7 - Well A S=2.7 - Well B
0.8
0.6
K(p)/Ki
0.4
0.2
0
0 300 600 900 1200 1500 1800 2100 2400
Pressure (Psia)
factor of 2.7 and 3.9 - well A and B
16 1200
12 900

8 600
4 300
0 0
Figure 6.16 History matched case using PDPM both inside and outside SRV- well A
108
12 1200
9 900

6 600
3 300
0 0
Figure 6.17 History matched case using PDPM both inside and outside SRV - well B
The results shown in Figures 6.16 and 6.17 do not exactly matches at mid time of
production for both wells. The results obtained from current cases are not as good as
shown in Figures 6.11 for well A and 6.14 for well B but current cases represent more
realistic approach and should be closer to reality than previous cases. It is important to
highlight that in all preceding sections pressure dependent permeability multipliers are
calculated using mean effective stress where total mean stress is kept constant and only
fluid pressure is varied, concept defined by Equation (6.4). This implies that stresses near
wellbore and in reservoir are not changing during production and remain constant
throughout and effectives stress changes as fluid pressure changes.
109
6.4 APPROXIMATION OF GEOMECHANICAL EFFECTS IN
CONVENTIONAL RESERVOIR SIMULATOR USING UNIAXIAL
DEFORMATION THEORY
Coupled geomechanical modeling in reservoir simulation is the correct way to represent
effective stress dependent porosity and permeability, which calculates stresses of porous
media using concept of deformations. Complexity of models and cost of geomechanical
modeling make it very difficult to run it all the time. Geomechanical effects can be
modeled in conventional reservoir simulators (uncoupled model) without running
geomechanical model by reasonable approximations. Most commercial reservoir flow
simulators offer options to include pressure dependent porosity and permeability. Since
rock data is measured in laboratory as function of effective stress, there is always
confusion how to convert lab data that only depends on pressure for conventional
reservoir simulation. Settari et al. (2005) presented approximations for porosity (rock
compressibility) coupling that is mainly dependent on type of deformations and
permeability coupling by introducing formulations both for global and local stress
changes caused by pressure depletion.
Although it is true that there is more complex deformation near wellbore or fractures,
uniaxial deformation theory serves as a reasonable approximation when reservoir is
producing at uniform average pressure. Then the stress changes can be calculated as a
function of pressure depletion.
Hook’s law for homogenous and isotropic solid materials in three dimensions is (Fjær et
al., 2008)
110
where = Elastic modulus, it is the measure of resistance of elastic deformation
= Strain in corresponding direction where i = x, y, z
= Stress in corresponding direction
= Poisson’s ratio, it is ratio of lateral expansion to longitudinal contraction
= Modulus of rigidity or shear modulus, it the measure of resistance of shear
deformation =
Porous media also contains fluid in its pores, which has a large impact on the physical
and mechanical behavior of rock, which will be considered in next phase. The first linear
theory of isothermal Poroelasticity was developed by Maurice A. Biot (1941), a Belgian-
American physicist. According to Biot, the presence of the pore fluid adds extra terms to
the strain energy of the material. He assumed that pore fluid cannot produce any shearing
strain and effect of pore fluid must be the same on all components of strain. Hence
Equation (6.6) for isotropic and isothermal linearly poroelastic material using positive
compressive stress and pressure becomes (Fjær et al., 2008)
111
where = Fluid pressure
= Poroelastic constant
For linearly isotropic material the stress–strain behavior of material is fully described
only by 2 constants i.e. Young’s Modulus (E) and Poisson’s ratio (υ). The others
constants can be calculated from these two constants. The poroelastic constant is given by
where = Bulk drained compressibility
= Grain (solid rock or rock matrix) compressibility
= Bulk modulus of framework or frame modulus =
= Solid grain modulus =
Bulk modulus, which is the measure of resistance of sample rock to hydrostatic
compression, can be represented in following form.
where = Volumetric strain =
= lame’s 1st parameter =
Then last term in Equation (6.7) is
Where = Biot’s constant =
Putting Equation (6.10) into (6.7) it becomes
112
The following assumptions are now made to derive relations for stress changes in
poroelastic media based on uniaxial deformation theory. The properties of rock in each
direction are constant, i.e., isotropic material.
 The rock material behaves as linearly elastic material.
 Entire reservoir undergoes uniform pressure change.
 Poisson ratio is same in all direction due to isotropy of material.
 Reservoir is deforming uniaxially in vertical direction.
 Vertical stress is constant, i.e., change in vertical stress is zero ( or = 0).
 No lateral strains, i.e., = =0
 No shear strain, i.e., =0
 No arching effects.
Applying it to Equation (6.11) and neglecting shear stress it is reduced to
Solving for first two terms in equation (6.12) and simplifying terms
Where = Poroelastic constant =
113
This expression means that changes in stress in x-direction and y-direction do not depend
on the modulus of rock but only on Poisson’s ratio, Biot’s constant and change in fluid
pressure. Poroelastic constant varies between 0 and . The maximum value of Poisson
ratio is normally 0.5, fluids, and minimum is zero, brittle rocks. For the same type of
rock, changes in stress depend only on the change in pore pressure of fluid.
Equation (6.13) in form of in-situ minimum and maximum horizontal stress can be
written as
If total mean stress is kept constant shown in Equation (6.15) then permeability becomes
function of pressure defined by Equation (6.4). The data from lab can be used directly by
modifying it. As changes in mean effective stress are equal to change in fluid pressure
then equation simplifies to Equation (6.16).
Lets instead assume that each grid block deform uniaxially independent of all other grid
blocks. If fluid pressure in particular grid block changes from to then change in
pressure is
Effective mean stress when total mean stress is changing with pressure at any particular
pressure is
Change in total mean stress is equal to
114
Putting Equation (6.14) into (6.2)
Substituting Equation (6.19) into (6.18) results in
Mean effective stress, which of function of both total mean stress and fluid pressure
defined by Equation (6.18) is given as
Equation (6.21) represents more rigorous approach as compared to Equation (6.16) by
taking into account poroelastic effects using uniaxial deformation theory. If permeability
is a function of total mean stress and fluid pressure then permeability multipliers
corrected for geomechanical effects can be calculated using Equation (6.22). Equation
(6.15) is a function of fluid pressure only and total mean stress is constant and Equation
(6.22) is function of both fluid pressure and varying total mean stress. The mean effective
stress for variable total mean stress calculated by Equation (6.22) will be smaller by
factor of than for constant total mean stress calculated by Equation (6.15).
115
BOTH INSIDE AND OUTSIDE SRV – COUPLED PRODUCTION
GEOMECHANICAL MODELS
To emphasize the importance of use of proper data set for any type of simulation, some
cases are run. Case 6.3.5 both for well A and B is considered as base for current case.
Fully coupled geomechanical models are run using effective stress dependent
permeability shown in Figure 6.14. The difference between Case 6.3.5 and current case is
that former is uncoupled reservoir flow model while the latter is a fully coupled
geomechanical model. Permeability multipliers are used in form of table in reservoir
simulators. In uncoupled model dependent variable is fluid pressure while in coupled
geomechanical model dependent variable is effective mean stress. In current case
permeability generated for uncoupled model is used in coupled geomechanical simulation
without correcting data set for poroelastic effects. Same permeability multipliers are
applied in input file for coupled simulation without modifying data set and table of mean
effective stress and multipliers are generated by changing fluid pressure given by
Equation (6.15).
Fracture and rock properties of Case 6.3.5, i.e., base case are given in Table 6.8 both for
well A and B are used here for current case, i.e., Case 6.5. The mean effective stress
dependent permeability curves shown in Figure 6.15 as dashed lines are used in coupled
geomechanical models for both wells. The results of Case 6.5 and Case 6.3.5 for both
wells and their comparison with field data are presented in Figures 6.18 and 6.19.
116
16 1600
Field Avg Gas Rate Gas Rate-C-6.3.5-Uncoupled-CMS*
Gas Rate-C-6.5-Coupled-CMS Field Cumulative Gas
Cumulative Gas-C-6.3.5-Uncoupled-CMS Cumulative Gas-C-6.5-Coupled-CMS
* CMS - Constant total mean stress

12 Stress 1200
8 800
4 400
0 0
Figure 6.18 Comparison of field data and simulation results of uncoupled and
coupled cases: Constant mean total stress - well A
12 1600
Field Avg Gas Rate Gas Rate-C-6.3.5-Uncoupled-CMS*
Gas Rate-C-6.5-Coupled-CMS Field Cumulative Gas
Cumulative Gas-C-6.3.5-Uncoupled-CMS Cumulative Gas-C-6.5-Coupled-CMS
* CMS - Constant total mean stress

9 Stress 1200
6 800
3 400
0 0
coupled cases: Constant mean total stress - well B
117
Running coupled model with the same data set results in higher production for both wells
as shown in Figures 6.18 and 6.19. Change in pressure given in Equation (6.17) is always
negative in depletion (production) case so mean effective stress calculated by Equation
(6.22) will always be smaller than calculated by Equation (6.15) for production case and
larger for injection case. In coupled reservoir simulation, fluid pressure serves an input to
geomechanical model which then calculates total stresses as function of change in
pressure, Poisson’s ratio and Biot’s constant. The reservoir model then takes input as
effective stress from geomechanical model and computes permeability multipliers from
that corresponding effective stress. The above discussion can be summarized in the
following flow chart.
Figure 6.20 Coupling of reservoir and geomechanical simulator – An overview
The plot of effective mean stress calculated by Equation (6.22) for variable total mean
stress, and by Equation (6.15) for constant total mean stress is given in Figure 6.21.
118
5688
Constant Mean Total Stress - Eq (6.15) - Well A
Variable Mean Total Stress - Eq (6.22) - Well A
5188 σm’ = 5188 @ P = 500 psia

Mean Effective Stress (Psi)
4688
σm’ = 4331 @ P = 500 psia
4188
3688
0 500 1000 1500 2000
Pressure (Psi)
Figure 6.21 Plot of mean effective stress and fluid pressure – well A
The effective stress is smaller for variable mean total stress case than constant mean total
stress case due to poroelastic effects. It means that total stresses are smaller in porous
media due to poroelasticity. Plotting permeability multipliers calculated using Equation
(6.5) both for constant and variable mean total stress cases against mean effective stress
and fluid pressure will further clarify why the results in coupled simulation case are
higher than in the uncoupled case. Total mean stresses calculated by geomechanical
simulator are not constant but always changing due to poroelasticity, which results in
smaller values at corresponding pressure. So permeability multiplier will be larger for
coupled case and hence higher gas rate and cumulative production. For the same value of
pressure, permeability multiplier for variable mean total stress case is always larger than
for the constant mean total stress case due to smaller values of mean effective stress as
shown in Figure 6.22.
119
1
S=2.7 - Well A - Constant Total Mean Stress
S=2.7 - Well A - Variable Total Mean Stress
0.8 S=3.9 - Well A - Constant Total Mean Stress

S=3.9 - Well A - Variable Total Mean Stress
0.6
’
K(p)/Ki
0.534 @ σm = 4331
0.4
’
0.368 @ σm = 4331
’
0.216 @ σm = 5188
0.2
’
0.075 @ σm = 5188
0
3688 4188 4688 5188 5688
Figure 6.22 Permeability multipliers and mean effective stress for constant and
variable mean total stress cases – well A
Although all the discussion and examples are for well A, same conclusions apply to well
B.
BOTH INSIDE AND OUTSIDE SRV – COUPLED PRODUCTION
GEOMECHANICAL MODELS – VARIABLE MEAN TOTAL STRESS
CASES
Based on discussion in previous section, it is necessary to correct the data set in order to
get realistic history match and forecast. Permeability multipliers computed using variable
total mean stress will be the correct way of modeling geomechanical effects in reservoir
120
simulation. For this purpose first of all permeability multipliers are recalculated using
Equation (6.22) and compared with previous multipliers used in section 6.3.5. The
modified set of permeability multipliers with larger stress factor for current case and
previous cases both for well A and B are presented in Figures 6.23 and 6.24.
Mean Effective tress (Psia)

3688 4188 4688 5188 5688
1
S=2.7-CMS* S=6.0-VMS** S=3.9-CMS S=8.4-VMS
S=2.7-CMS S=6.0-VMS S=3.9-CMS S=8.4-VMS
0.8 * CMS - Constant total mean stress

** VMS-Variable total mean stress
Outside SRV
0.6
K(p)/Ki
Inside SRV
0.4
0.2
0
0 500 1000 1500 2000
Pressure (Psi)
factor of 2.7, 3.9, 6 and 8.4 - well A
Simulation is run using modified set of data for both wells using uncoupled and fully
coupled simulators. Case 6.6 is the case for both wells in which permeability multipliers
are calculated by considering poroelastic effects. Fracture and rock properties used in
current case and their comparison with previously history matched cases are presented in
Table 6.9. Simulation results are presented in Figures 6.25 and 6.26 for well A and B
respectively.
121
3553 4053 4553 5053 5553 6053
1
0.8 * CMS - Constant total mean stress

** VMS-Variable total mean stress
Outside SRV
0.6 Inside SRV
K(p)/Ki
0.4
0.2
0
0 500 1000 1500 2000 2500
Pressure (Psi)
factor of 2.7, 3.9, 6 and 8.4 - well B
Table 6.9 Fracture properties of Cases 6.6 both for well A and B
Values Values
Property – Well A
(Case 6.3.5) (Case 6.6)
27
261
Stress factor Inside SRV, S 3.9 8.4
122
Values Values
Property – Well B
(Case 6.3.4.5) (Case 6.6)
Reservoir permeability, mD 0.005, 0.0005
14
131
Stress factor Inside SRV, S 3.9 8.4
16 1200
12 900

8 600
Field Avg Gas Rate Gas Rate-C-6.6-Uncoupled-VMS*
Gas Rate-C-6.6-Coupled-VMS Field Cumulative Gas
Cumulative Gas-C-6.6-Uncoupled-VMS Cumulative Gas-C-6.6-Coupled-VMS

4 300
* VMS - Variable total mean stress
0 0
coupled cases: Variable mean total stress - well A
123
12 1200
9 900

6 600
Field Avg Gas Rate Gas Rate-C-6.6-Uncoupled-VMS*
Gas Rate-C-6.6-Coupled-VMS Field Cumulative Gas
Cumulative Gas-C-6.6-Uncoupled-VMS Cumulative Gas-C-6.6-Coupled-VMS
* VMS - Variable total mean stress

3 300
0 0
coupled cases: Variable mean total stress - well B
Results of both uncoupled and coupled geomechanical production models are remarkably
similar as shown in Figures 6.25 and 6.26. This shows that proper care should be taken
while running coupled geomechanical model with a data set from an uncoupled model.
Stresses are changing constantly so constant total mean stress assumption is not an
accurate assumption; it must be replaced with variable total mean stress. Poroelastic
effects can be modeled in conventional reservoir simulator by assuming uniaxial
reservoir deformation. Matrix permeability enhances during fracturing, which is
represented here by initial permeability multiplier (IPM). IPM of 50 and 100 is used for
124
well A and B respectively to get history match. These numbers are consistent with those
reported by Settari et al. (2002b) and Settari et al. (2009).
6.7 SUMMARY
 Stimulated reservoir region is much larger than propped fracture length.
 The dimensions of Stimulated reservoir volume (SRV) are different for x-linked
and slick water fracturing wells – shorter for x-linked fracture job.
 One possible way to get history match is the use of a constant permeability
multipliers inside SRV and pressure dependent permeability (K (p)) outside SRV.
 The two wells seem to have somewhat different permeability.
 Accurate way to get history match is to use pressure dependent permeability
multipliers both inside and outside SRV.
 Permeability of the region inside SRV has larger pressure dependence than region
outside SRV due to presence of complex network of microfractures.
 Moving from uncoupled to fully coupled geomechanical modeling without
correcting data sets for poroelastic effects will result in higher production. Data
sets must be correct for geomechanical aspects.
 Constant mean total stress assumption is not realistic approach as horizontal
stresses are decreasing with pressure depletion. Such assumption will always
result in overestimated production.
125
 New method based on uniaxial deformation theory is proposed to approximate
poroelastic effects in the reservoir and it predicts variation of mean total stress
depending on Poisson’s ratio, Biot’s constant and fluid pressure.
 The method proves to be a good approximation to capture poroelastic effects and
can be used in conventional reservoir simulation without running fully coupled
geomechanical simulation.
 Good history match is achieved in uncoupled simulation when data set is
corrected for poroelastic effects and results are confirmed by comparing it with
results from coupled geomechanical simulation.
126
CHAPTER SEVEN: HISTORY MATCHING OF FIELD INJECTION
PRESSURE USING UNCOUPLED AND FULLY COUPLED
GEOMECHANICAL INJECTION MODELS
It was evident in prior sections that history matching of field production data did not give
a unique set of parameters. Models with a different set of parameters can give almost the
same results and different sub-sets of parameters with different set of assumptions can
predict the same results. Injection models can be used to verify the number of parameters
involved in production history matching by matching well injection pressure during
treatment with simulation results. Field bottomhole injection pressure (BHIP) both for
well A and B is given in Figure 3.5. Well B which is deeper than A has a higher
bottomhole injection pressure than well A. Breakdown pressure, maximum required to
initiate fracture in formation, is also higher for well B.
This chapter represents simplified modeling of fracture propagation (dynamic fractures),
and history matching of field injection pressure using uncoupled and fully coupled
geomechanical injection models. The sensitivity study of various parameters such as
permeability enhancement/reduction factor, length of fracture propagation, and Biot’s
constant is performed in this chapter. Limitations of the uncoupled and coupled
geomechanical injection models are also discussed in this chapter. Figure 7.1 represents
flow chart of simulation runs performed in this chapter. Every previous case serves as
base case for next simulation run.
127
History Matching of Field Injection Pressure using
Uncoupled and Coupled Geomechanical Injection Models
Uncoupled Geomechanical Injection Coupled Geomechanical Injection

Models Models
Effects of Permeability Reduction Effects of Limiting Length of Fracture

Factor (Rfa) – Cases – 7.3.1.1 – 3 Propagation – Cases – 7.4.1.1 – 4
Effects of Limiting Length of Fracture Effects of Biot’s Constant (α) –

Propagation – Cases – 7.3.2.1 – 2 Cases – 7.4.2.1 – 2
Failure Predictions
Mohr – Coulomb Failure Criterion (Shear Fracturing)
7.1 MODELING OF FRACTURE PROPAGATION
Fracture propagation modeling is an important part of physics and must be considered in
injection modeling of wells. Classical modeling of fracture geometry is well established
and documented in literature of rock mechanics and stimulation (Economides and Nolte,
2000). Coupling of fracture propagation (fracture dynamics) and fluid flow is
computationally very expensive (Ji et al., 2004b, 2009). The modeling of “complex”
fracturing discussed in Chapter 2 (Weng et al., 2011) is also expensive. However, proper
representation of dynamic propagation in which the fracture is directly coupled into a
reservoir simulator is important for many applications. Settari (1990, 2002a) has shown
128
that the fully coupled treatment of fracture mechanics, reservoir modeling and
geomechanics is important for better understanding of the unconventional fracturing
applications (waterflooding at fracturing condition, etc.) as well as for tight gas fracturing
treatments such as waterfracs and will be described briefly here. Such coupling, described
next, can be achieved in a simplified fashion which makes the modeling computationally
efficient.
7.2 THEORY OF SIMPLIFIED FRACTURE MODEL
The method is based on modifications of transmissibility in the reservoir flow model as
discussed in Chapter 4. For injection scenarios transmissibilities are modified
dynamically. To model dynamic fracturing process, a transmissibility multiplier function
is assigned to a line (or plane) of grid blocks assumed for fracture propagation extending
around the well. The multiplier function is a table that can be derived from simple 2-D
analytical fracture models which approximates the actual fracture. In an uncoupled
modeling, transmissibility multipliers are a function of fracture injection pressure, while
in a coupled (geomechanical) system they are a function of a minimum effective stress.
The multipliers are calculated based on the estimation of a 2D crack opening in a cross-
section by Equation (7.2) (Sneddon, 1969), then calculating the fracture permeability by
Equation (7.1) as a function of the net pressure in the fracture, and finally calculating the
multiplier as described in Equation (7.3).
129
Where = Fracture permeability
= constant, permeability reduction factor
= Fracture width
= Fracture treatment pressure (fluid pressure)
= Fracture opening or closing pressure
= Fracture half height
= Poisson’s ratio
= Elastic modulus
= Transmissibility multiplier
= Matrix permeability
= grid width
Hf in Equation (7.2) is the fracture half-height based on the 2-D Perkins-Kern geometry
assumption of vertical fracture with smooth closure at the top and bottom (Perkins and
Kern, 1961). Permeability reduction factor ( ) is a constant and it is one of the history
matching parameters. Fracture opening or closing pressure ( ) is normally taken as
initial minimum horizontal stress acting perpendicular to fracture face. may be
actually greater (or lower) than the initial minimum horizontal stress due to poroelastic
and thermo elastic effects. For uncoupled modeling it can be assumed that the stress
changes have stabilized and fracture opening or closing pressure is equal to the average
130
(adjusted) at minimum horizontal stress during the treatment. So transmissibility in
Equation (7.3) is function of fracture treatment pressure only for uncoupled modeling.
Where = Initial minimum horizontal stress (or adjusted value)
7.3 HISTORY MATCHING OF FIELD INJECTION PRESSURE USING
UNCOUPLED INJECTION MODELS
Hydraulic fracturing alters the initial in-situ stresses particularly around wellbore and
fractures. Although estimation of stress changes due to injection pressure using uniaxial
deformation theory is not accurate representation, it serves as a reasonable approximation
for this project. The Concept used in Chapter 6 for approximation of geomechanical
effects in an uncoupled model and equations developed for production cases, i.e.,
Equation (6.15) and (6.22) can be used also for injection modeling. For injection cases
change in pressure is always positive given by Equation (6.15). So effective mean stress
calculated by Equation (6.22) considering poroelastic effects is always larger than
calculated by Equation (6.15), which ignores poroelasticity. This means that
pressurization during fracturing increases stresses and stresses should be modified
accordingly. Changes in total stresses in x, and y – direction are calculated by the same
method as discussed for production modeling. History matching field injection pressure
using uncoupled injection models is different from coupled simulation because there is no
stress solution in uncoupled model so any changes in stresses due to poroelastic effects
131
must be incorporated manually in reservoir simulator for either permeability multipliers
or transmissibility calculation. Modified stresses are used in uncoupled model assuming
that hydraulic fracture increases in-situ stresses near wellbore and around fractures due to
poroelasticity and permeability enhancement in that region must be predicted by using
these modified stresses. For this purpose simple approach is used; minimum horizontal
stress was estimated from field injection pressure data and values are compared with
initial in-situ stress values calculated by data given in Table 3.3. The factor by which
both values are different is taken as multiplying factor to calculate modified in-situ stress
(stresses after fracturing). Initial and modified in-situ stresses in each corresponding
direction for wells A and B are presented in Table 7.1.
Fracture propagation in reservoir simulator is represented by modifying transmissibility
of fracture plane against corresponding injection pressure. Transmissibilities are
calculated using Equation (7.3) which is used in a form of table vs. injection pressure in
reservoir simulator blocks and fracture is allowed to grow without restricting its
geometry, i.e., half length and width. The data used to calculate transmissibility for wells
A and B is presented in Table 7.2.
Table 7.1 Initial and modified stresses for injection cases – Well A and B
Well A Well B
Initial Modified Initial Modified
value value value value
Vertical Stress, psi ( ) 6572 8400 6646 9090
Maximum Horizontal Stress, psi ( ) 5989 7800 6056 8500
Minimum Horizontal Stress, psi ( ) 4505 6300 4556 7000
132
Table 7.2 Data for transmissibility calculations in injection cases – Well A and B
Property Well A Well B

Fracture half height, ft 50
Pressure constraint, psia 10,000
Poison ratio, fraction 0.125
Grid block width, ft 2
Minimum horizontal stress, psia 6300 7000
Elastic modulus, psia 7.99E6 7.99E6
Matrix permeability, mD (Vertical) 0.015 (0.0015) 0.005 (0.0005)
In Chapter 6, it was shown with field examples that pressure / stress dependent
permeability increases due to poroelastic effects. In section 6.6 of Chapter 6 stress
dependent matrix permeability multipliers outside SRV are defined by stress factor of 6
for history matched cases. Permeability multipliers for uncoupled injection cases are
calculated using Equation (6.5) while effective mean stress is determined from Equation
(6.22). Modified in-situ stresses given in Table 7.1 are used for permeability multipliers
calculations. Calculated permeability multipliers both for wells A and B are presented in
Figure 7.2. Injection model was setup in reservoir simulator and run for given injection
period by setting maximum bottomhole injection pressure of 10,000 psi.
133
Mean Effective Stress (psi)
2000 3000 4000 5000 6000
50
Injection - Well A - S = 6
Injection - Well B - S = 6
Injection - Well A - S = 6
40 Injection - Well B - S = 6
30
K(p)/Ki
20
10
0
2000 4000 6000 8000 10000
factor of 6 (uncoupled injection cases) - well A and B
7.3.1 Effects of Permeability Reduction Factor (
Permeability reduction factor plays an important role to calculate transmissibility of grid
blocks in fracture plane. Fracture permeability given by Equation (7.1) becomes an
intrinsic permeability of smooth open fracture if reduction factor is taken as 1. Fracture
permeability in reality is much smaller than this value due to tortuosity, asperities
interlocking, rock chipping at fracture face, unequal and rough surface of rock faces, and
fracture degradation. All these factors tend to decrease fracture permeability by order of
magnitude and are represented by the permeability reduction factor. So fracture
permeability calculated by Equation (7.1) with a factor < 1 represents more realistic
situation. Three injection cases are run to match field injection pressure with simulation
134
injection pressure by varying only the fracture permeability reduction factor. Fracture
propagation was allowed in y-direction without restricting or confining its half length.
Transmissibility given by Equation (7.3) is dependent on permeability reduction factor so
transmissibility values are calculated using data set given in Table 7.2 for each case with
the corresponding reduction factor. Matrix permeability multipliers are same in all
direction, i.e., x, y and z, and given in Figure 7.2 for these cases, and are applied in all
grid blocks except blocks in fracture plane because fracture transmissibility multipliers
are used for those blocks. The purpose of running these cases is to see the effects of
permeability reduction factor Rfa on injection pressure. Reduction factors for each case
run for wells A and B are given in Table 7.3.
Table 7.3 Permeability reduction factor of Cases 7.3.1.1-4 both for well A and B
Property Cases Well A Well B

Case 7.3.1.1 0.1
Permeability
Case 7.3.1.2 0.01
Reduction Factor
Case 7.3.1.3 0.001
Transmissibility multipliers are calculated for the above three cases using corresponding
values of permeability reduction factor by Equation (7.3) and are plotted in Figure 7.3;
similar graph can be constructed for well B. Transmissibility in x-direction is equal to 1.0
as fracture is growing only in y-z plane. Simulation results of bottomhole injection
pressure for all the three cases and field injection pressure are presented in Figure 7.4 and
7.5 for well A and B respectively.
135
1.0E+12
1.0E+09
Transmissibility Multipliers
1.0E+06
TmultY-Rfa=0.1
TmultZ-Rfa=0.1
TmultY-Rfa=0.001
1.0E+03 TmultZ-Rfa=0.001
TmultY-Rfa=0.01
TmulyZ-Rfa=0.01
1.0E+00
6000 7000 8000 9000 10000
Figure 7.3 Transmissibility multipliers – uncoupled injection cases - well A
Results shown in Figures 7.4 and 7.5 show that decreasing reduction factor pushes the
injection pressure upward. Smaller reduction factor means smaller transmissbility
mutlipliers and hence frcature propagation becomes more difficult due to larger pressure
drop down in the fracture. Smaller matrix permeability and fracture propagation hinder
fluid movement and pressure builds up. The injection pressure curves from simulation
show almost constant pressure trend throughout but slowly increasing with time.
136
8000
BHIP - Field - Well A
BHIP - C-7.3.1.1
BHIP - C-7.3.1.2
7500 BHIP - C-7.3.1.3
7000
6500
6000
0 50 100 150 200 250
Time (mins)
Figure 7.4 Comparison of simulation and field BHIP of Case 7.3.1.1-3- well A
8500
Field - BHIP - Well B
BHIP - C-7.3.1.1
BHIP - C-7.3.1.2
BHIP - C-7.3.1.3
8000
7500
7000
6500
0 50 100 150 200 250 300
Time (mins)
Figure 7.5 Comparison of simulation and field BHIP of Case 7.3.1.1-3 - well B
137
7.3.2 Effects of Limiting Length of Fracture Propagation
It is evident from the results of previous cases that decreasing the reduction factor will
not help us in history matching injection pressure. A mechanism that would create larger
pressure increase with time is required. One method is to restrict or confine fracture
propagation in length (half length), which means controlling fracture propagation by
modifying transmissibility of grid blocks in fracture plane only within an assumed
fracture half length. In this way fracture half length cannot exceed a specified number of
blocks for which transmissibilities are modified. Case 7.3.1.2 is selected as base case for
simulation runs in this section. Several simulation cases are run both for well A and B
using different values of pre determined fracture half length. Reservoir parameters used
in simulation run are same as base case. Permeability multipliers given in Figure 7.2 are
used everywhere except in fractured blocks for Case 7.3.2.1 and 7.3.2.2. Permeability
multipliers defined by stress factor of 3.0 are also used in fracture plane grid blocks only
for Case 7.3.2.2 – well B and are applied in x-direction while value of 1.0 is used for
multipliers in other two directions.
The parameters used for history matching of injection pressure for wells A and B are
summarized in Table 7.4. Comparison of simulation results of current cases and
bottomhole field injection pressure are presented in Figure 7.6 and 7.7 for well A and B
respectively followed by the discussion of the results.
138
Table 7.4 Parameters varied in injection Cases 7.3.2.1 – 2 – Well A and B
Well A Well B
Property
(Case 7.3.2.1) (Case 7.3.2.1)
Permeability reduction factor 0.01
Permeability multipliers in x, y and z-direction S = 6.0
Well B
(Case 7.3.2.2)
Permeability multipliers in x - direction S = 3.0
Permeability multipliers in y and z - direction S = 6.0
8000
BHIP - C-7.3.1.2
BHIP - C-7.3.2.1
7500
7000
6500
6000
0 50 100 150 200 250
Time (mins)
Figure 7.6 Comparison of simulation and field BHIP of Case 7.3.2.1- well A
In Case 7.3.1.2 of Figure 7.6, which is constructed for well A, fracture propagation was
allowed in y-direction without imposing any constraint on it but in Case 7.3.2.1 length of
139
fracture propagation was confined to its half length of 75 ft. Pressure rises approximately
by 300 psi after 70 minutes of injection in Case 7.3.2.1, which means that it takes 70
minutes for fracture to propagate to 75 ft (half length). Field pressure afterward keeps
increasing but change in computed pressure is very small. This is the best history match
that can be achieved with uncoupled injection simulation for well A.
10500
BHIP - Field - Well B
BHIP - C-7.3.1.2
BHIP - C-7.3.2.1
BHIP - C-7.3.2.2
9500
Bottomhole Pressure (psi)
8500
7500
6500
0 50 100 150 200 250 300
Time (mins)
Figure 7.7 Comparison of simulation and field BHIP of Case 7.3.2.1-2- well B
Results of Cases 7.3.1.2, 7.3.2.1 and 7.3.2.2 for well B are presented in Figure 7.7. In
Case 7.3.1.2 fracture propagation is allowed in y-direction as explained earlier but in
Case 7.3.2.1 fracture propagation is restricted to half length of 150 ft. The huge increase
in pressure to value of 10,000 psi in Case 7.3.2.1 is due to pressure build up as soon as
fracture propagation reaches its confinement limit. Small matrix permeability
perpendicular to fracture face also restricts fracture leak-off in x-direction and fracturing
140
pressure increases its value to its limit set to 10,000 psia in the simulator. To decrease
pressure build up during injection period and avoid operating at maximum constraint, one
can increase the permeability of matrix in x-direction so that leak-off can increase in x-
direction and delay fracture propagation in y-direction. For this reason Case 7.3.2.2 was
run for well B, in which permeability multipliers are applied in x-direction for fractured
blocks with S value of 3.0. Multipliers were used only in x-direction without effecting
permeability in other directions. This method somewhat improves the results, which can
be seen in Figure 7.7 – Case 7.3.2.2, but simulation still did not match field pressure
curves.
At this point Case 7.3.2.1 for well A and Case 7.3.2.2 for well B represent best history
matched cases for uncoupled injection scenarios. This depicts the limitation of uncoupled
model to history match injection pressure.
7.4 HISTORY MATCHING OF FIELD INJECTION PRESSURE USING A
FULLY COUPLED GEOMECHANICAL INJECTION MODEL
This section of chapter deals with history matching field injection pressure using a fully
coupled injection geomechanical model. In coupled geomechanical simulation, because
stresses are continuously computed, properties are updated at each time step in reservoir
simulator by taking as an input effective stress from geomechanical part of the simulator.
Therefore there is no need to modify stress data to correct for poroelastic effects. To run a
fully coupled geomechanical simulation the original in-situ stress data given in Table 7.1
is used to calculate transmissibility and permeability multipliers, which is function of
141
effective stresses. Transmissibility of grid blocks in fracture plane is a function of
minimum horizontal effective stress given by Equation (7.6) while matrix permeability
multiplier is function of mean effective stress in coupled geomechanical simulation
defined by Equation (6.3).
Where = Minimum horizontal effective stress
= Minimum horizontal total stress
= Biot’s constant
Minimum horizontal effective stress in Equation (7.6) is function of both time and space;
at any particular point in reservoir block it is constantly changing with time during
pressurization. Biot’s constant of 1.0 was initially used for effective stress calculation.
Run time for coupled simulation is very large and consequently a detailed study for each
parameter was not possible due to time constraints. The sensitivity study and calculations
shown here are performed for well A. Only conclusions and end results are then applied
to well B to get a history match. Because the results and conclusions obtained for well A
hold also true for well B, detailed discussion is not presented here. Permeability
multipliers calculated using original in-situ stresses are given in Figure 7.8 while
transmissibility multipliers for different fracture permeability reduction factors R fa are
presented in Figure 7.9 for well A. Same type of curves can be generated for well B using
the same equation and data set given in Table 7.1.
142
25,000
1,000
Injection - Well A - S=6
20,000 800
K(p)/Ki 600
15,000
K(p)/Ki
400
10,000
200
5,000 0
0 500 1000 1500 2000 2500 3000 3500 4000
0
0 1000 2000 3000 4000
Figure 7.8 Effective stress dependent permeability curve - coupled injection - well A
1.00E+08
1.00E+06
Transmissibility Multipliers
TmultY - Rfa=0.00001
TmultZ - Rfa=0.00001 1.00E+04
TmultY - Rfa=0.000007
TmultZ - Rfa=0.000007
TmultY - Rfa = 0.0000572
TmultZ - Rfa = 0.0000572 1.00E+02
1.00E+00
-3500 -3000 -2500 -2000 -1500 -1000 -500 0
Minimum Horizontal Effective Stress (psi)
Figure 7.9 Transmissibility multipliers for coupled injection cases - well A
143
Sensitivity study was performed on two history matching parameters, i.e., maximum
length of fracture propagation and Biot’s constant. Permeability multipliers and
transmissibility for fractured blocks were as presented in Figures 7.8 and 7.9.
7.4.1 Effects of Limiting Length of Fracture Propagation
Few simulation cases are run using different values of permeability reduction factor and
confining length of fracture propagation. For this purpose, it was assumed that rock
behaves as a perfectly elastic material which does not exhibit hysteresis during loading
and unloading. A base case here (Case 7.4.1.1) was the run allowing unlimited fracture
propagation in y-direction and modifying Biot’s constant value in the geomechanical
simulator to 0.65. Summary of history matching parameters varied in cases run in this
section are presented in Table 7.5.
Table 7.5 Parameters varied in coupled injection Cases 7.4.1.1 – 4 – Well A
Property – Well A Case 7.4.1.1 Case 7.4.1.2 Case 7.4.1.3 Case 7.4.1.4
Biot’s constant 0.65

Permeability reduction
0.00001 0.000007 0.00000572
factor Rfa
Fracture half length, ft Not restricted 100 150 135
In all the cases same Biot’s constant is used. The reason for using such a small value is
that, as will be shown in the next section, the use of high values of  (close to 1) creates
large poroelastic stresses, which prevent fracture from initiating in the model. The
objective here is to restrict or confine fracture propagation to an assumed fracture half
144
length by restricting fracture transmissibility modifiers to this length in the simulator. In
all cases the fracture height was assumed to be equal to pay zone thickness. Both
permeability multipliers and transmissibility are function of effective stress in coupled
modeling and entered in form of tables. All the numbers in Table 7.5 are used as an initial
guess for initial simulation for sensitivity study, which serves as base cases for further
simulation runs. Simulation results for all these cases and field injection pressure are
presented in Figure 7.10 for well A.
8500
BHIP - C-7.4.1.1
BHIP - C-7.4.1.2
BHIP - C-7.4.1.3
BHIP - C-7.4.1.4
7500
6500
5500
0 50 100 150 200 250
Time (mins)
Figure 7.10 Comparison of simulation and field BHIP of Cases 7.4.1.1-4 - well A
Fracture propagation was allowed in y-direction in base case (7.4.1.1) and there is a slow
continuous pressure build observed for that case due to uninterrupted fracture
propagation. The fracturing pressure rise is due to the pressure drop down in the fracture.
In case 7.4.1.2 the same parameters were used except fracture propagation was restricted
145
to half length of 100 ft. Pressure started building up after 70 minutes of injection, which
was the time taken by fracture to reach half length of 100 ft shown in Figure 7.10. The
restriction at the tip of fracture preventing fracture growth is causing pressure build up
(increment) upon further injection. It was revealed while analyzing the results in detail
that the downward dip between 100 – 110 minutes is due to numerical errors. Since this
case did no matches field pressure therefore no further investigation was carried out. In
Case 7.4.1.3 both permeability reduction factor and fracture half length were changed.
The effect of permeability reduction is discussed in detail in uncoupled simulation
section; decreasing its value shifts pressure injection curve upward which can be
observed in Figure 7.10 by comparing Cases 7.4.1.2, 7.4.1.3 and7.4.1.4. Limiting fracture
half length was increased to 150 ft from 100 ft. Very little bump in injection pressure is
now observed to be on the low side which is an indication that if all other parameters are
kept constant then the correct fracture half length will fall in between 100 to 150 ft to get
a reasonable match.
Case 7.4.1.4 was run by reducing further the permeability reduction factor (Rfa) and using
fracture half length of 135 ft as shown in Figure 7.10. Although simulation does not
exactly match field injection pressure, it is the best history match so far. It is concluded
that injection history match requires some mechanism to constrain fracture propagation at
a late stage. This issue was not pursued further; however, the coupled cases show much
improvement and are much better than the uncoupled simulations.
Discussion on the late time mismatch:
It is important to point out that history matching of field injection pressure after 190
minutes of injection cannot be achieved through our simulation results. It was observed
146
by thorough investigation of field treatment report that the injected proppant
concentration was increased after 190 minutes to approximately three times of the overall
average concentration. The snap shot of field treatment report is presented in Figure 7.11;
increase in BHIP after 190 minutes is due to increase in proppant concentration at that
time. As our simulation study does not include coupling of fracture propagation
simulation with proppant transport, modeling of fracture propagation based on downhole
variable proppant concentration is not possible here and beyond the scope of this study.
Fracture modeling in our case was performed based on total amount of downhole injected
slurry into formation. In light of above discussionTreatment Data Charts

it will be fair assumption to compare
simulation results COMPANY

BP CANADA ENERGY with field data for first 190 minutes of injection only. Gelled Water Frac
CO2 Crosslinked
BP Hz Noel a-A88-J 93-P-1 Cadomin (Gas) June 8, 2006
Wellhead Treating Bottomhole

70000
60000
Pressure ( kPa )
50000
40000
30000
20000
10000
0
0 50 100 150 200 250
Wellhead Inline Bottomhole

500
450
Concentration ( kg/m³ )
400
350
300
250
200
150
100
50
0
0 50 100 150 200 250
Blender Clean Blender Slurry Combined

14.0
Figure 7.11 Proppant concentration, BHIP and rate - well A
12.0
Rate ( m³/min )
Figure10.0
7.12 represents simulation and field bottomhole injection pressure for first 190
8.0
6.0
minutes of injection.
4.0
2.0
147
0.0
0 50 100 150 200 250
Elapsed Time ( min )
7500
BHIP - C-7.4.1.4
7000
6500
6000
0 50 100 150 200
Time (mins)
Figure 7.12 History matched case – Coupled simulation - well A
7.4.2 Effects of Biot’s Constant ( )
Sensitivity study was performed to see the effects of Biot’s constant on injection pressure
and here we only report the results for well A. Summary of the values of Biot’s constant,
used in two sub-cases run in this section is given in Table 7.6. History matched case from
previous section, i.e., Case 7.4.1.4 is used as a base case for these runs. All other
parameters are kept constant. Comparison of simulation results for the sub cases and the
base case together with the field injection pressure is presented in Figure 7.13.
148
Table 7.6 Parameters varied in injection cases 7.4.2.1 – 2 – Well A
Property – Well A Case 7.4.2.1 Case 7.4.2.2
Biot’s constant 0.75 1.0

11000
10000

BHIP - C-7.4.1.4
BHIP - C.7.4.2.1
9000
BHIP - C-7.4.2.2
8000
7000
6000
0 50 100 150 200 250
Time (mins)
Figure 7.13 Comparison of simulation and field BHIP of Cases 7.4.1-4 - well A
It is evident from Figure 7.13 that increasing Biot’s constant changes the results
tremendously. In the case of = 1 (Case 7.4.2.2) the model failed to initiate fracture
because the poroelastic stress component caused by injection pressure was too high and
the total stress increased above the injection pressure limit (set at 10,000 psia). The result
is a rapid, almost instant increase of pressure up to the injection pressure limit, which
149
plots on the graph as a vertical line close to time=0 followed by a horizontal line. The
smaller the Biot constant the slower is the increase in total stress and it is less difficult to
fracture the rock. Note that for fracture initiation (and propagation) minimum effective
stress must be negative; in other words injection pressure should be higher than minimum
horizontal total stress. Bottomhole injection pressure the wellbore grid block for fracture
# 2 (2nd fracture from the line of symmetry) is plotted vs. minimum horizontal effective
stress in Figure 7.14. It is evident from the curves that it is easier to create or propagate
fracture for smaller values of Biot’s constant. No fractures are created in Case 7.4.2.2
because the minimum horizontal effective stress is positive for all pressure values and
continuous injection with a BHIP limit will force a reduction in the injection rate in the
model. That is why injection pressure jumps to maximum value of 10,000 psi in very
early time and remains constant. It is therefore concluded that Biot’s constant equal to 1.0
is not reasonable assumption and should be less than 1.0. Biot’s constant can be
calculated from Equation (6.10) if rock bulk and grain modulus are known, which can be
measured in the laboratory. In absence of laboratory data approximate value can be
estimated by history matching which in our case led to the value of 0.65, but such
estimates must be used with necessary caution. Case 7.4.1.4 represents best history
matched case in coupled geomechanical simulation cases.
150
10000
Biot's Constant=0.65 - C-7.4.1.4
8000
6000
4000
2000
-3000 -2000 -1000 0 1000 2000 3000
Minimum Horizontal Effective Stress (psi)
Figure 7.14 Minimum effective stress and BHIP at different Biot’s constant
numbers – Coupled cases - well A
7.5 FAILURE PREDICTIONS
The simulations of the injection process presented in this Chapter showed that one must
assume a substantial stress-dependent enhancement of permeability around the primary
single plane fracture (SPF) to history match the injection pressures. Subsequently, the
permeability of the SRV will decline during production as was described in detail in
Chapter 6. Often it is postulated that the creation of this SRV is due to shear fracturing,
i.e., creating shear failure. Coupled modeling provides us with the tool to investigate
under what conditions shear fracturing occurs and what would be the extent of the SRV if
it was caused purely by shear failure. This aspect is examined in the present Section.
151
7.5.1 Tensile Failure
When tensile stress across a plane exceeds critical limit then tensile failure occurs. This
critical limit is called the tensile strength or ultimate tensile strength (UTS). It is a
characteristic property of rock and does not depend on size of specimen. It is opposite of
compressive strength and has units of stress. For linearly elastic material it can be
determined from stress strain curve in laboratory testing. Here all the compressive
stresses are denoted by positive sign. Tensile failure will occur when minimum effective
stress exceeds tensile strength of rock, is given as:
Where = Minimum effective stress
= Ultimate tensile strength or tensile strength
The tensile failure criterion is applied to determine the propagation of the main fracture
(SPF) through grid blocks. In some rare instances, tensile failure can also occur in the
reservoir around the SPF (e.g. due to thermal effects, Tran et al., 2012).
7.5.2 Shear Failure
When shear stresses along a plane in a specimen exceed shear strength of material then
shear failure occurs. The two sides of failure surface move relative to each other and
friction forces, which depend on normal forces acting on specimen, oppose this relative
movement. So shear strength of material / rock indirectly depends on the normal stress
acting over the failure plane.
152
There are different shear criterions available in literature such as Tresca, Mohr-Coulomb
and Griffith. For this study Mohr-Coulomb criterion is selected to predict failure
mechanism during injection. The objective of this study is to determine which type of
failure occurred during injection or pressurization.
(a) (b)
Figure 7.15 Schematics of tensile and shear failure under normal loading
7.5.3 Mohr – Coulomb Shear Failure Criterion
Mohr–Coulomb theory is a mathematical model describing the response of brittle
materials to shear stress as well as normal stress. It is a general and frequently used
criterion which postulates a linear relationship between shear stress at failure and normal
effective stress acting on failure plane and is given as:
Where = Shear stress at failure
= Shear strength or Cohesion
= Friction angle
153
= Effective normal stress
When friction angle is zero ( ) then Equation (7.7) becomes Tresca Criterion, so
Tresca criterion is special case of Mohr-Coulomb. Mohr-Coulomb criterion in graphical
form is presented in Figure 7.16.
Figure 7.16 Mohr-Coulomb failure criterion for critical stress state – A graphical
representation (from Faejer et al., 2008)
The distance of point of intersection and origin is denoted by and is called attraction. It
is related to other Mohr-Coulomb parameters by following relation.
and in Figure 7.16 is maximum and minimum effective stress. Mohr-Coulomb
criteria states that specimen fails under this criterion when circle touches failure line,
which means that failure criteria for some plane(s) is fulfilled. The value of intermediate
effective stress has no influence on failure criteria as it lies between maximum and
minimum effective stress so pure shear failure is independent of intermediate stress.
154
Figure 7.16 represents a case in which failure occurred in specimen referring it to critical
stress state. The shear and corresponding effective normal stress from Figure 7.16 is
Where = Failure plane orientation angle
gives orientation of failure plane with respect to maximum effective stress and it is
related to friction angle by the following relation.
If cohesion is known then uniaxial compressive strength, , can be calculated by given
relation.
To investigate if tensile or shear failure will occur; time-history of pressure and stresses
was extracted for specific grid blocks from a coupled simulation run. A program was
written in MATLAB® to present failure prediction of these grid blocks in graphical form
based on Mohr-Coulomb failure criterion. For this purpose, the history matched case, i.e.,
Case 7.4.1.4 was used. All the fractures behave the same way and pressure propagation is
also almost the same for all fractures. Therefore only one fracture is selected for analysis
which is fracture # 4 (4th fracture from line of symmetry) and conclusions are drawn from
this analysis will apply to all sets of fractures. Four grid blocks were selected and marked
as shown in Figure 7.17, which represents layer view of fracture in x-y plane. The well is
completed in x-direction and block 1 represents the perforation location.
155
Block 2
Block 3
Block 1 (Well) Block 4
Y
X
(Well )
Figure 7.17 Fluid pressure after 236 days and selected grid block locations – Well A
Minimum effective stress with respect to time for all four blocks is shown in Figure 7.18.
All the stress values are taken from the geomechanical module output for history matched
case and selected grid blocks. Minimum effective stress in block 1 falls to negative value
in very early time, which is an indication that fracture is created within this block.
156
3000
Block - 1 Block - 2
Block - 3 Block - 4
2000
Minimum Effective Stress (psi)
1000
-1000
-2000
-3000
0 50 100 150 200 250
Time (mins)
Figure 7.18 Minimum effective stress block 1 – 4 – history matched case – Well A
10000
8000
Total Stress(s) & BHIP (psi)
6000
Sx - Block 1
Sy - Block 1
Sz - Block 1
4000 Pressure - Block 1
Sx - Block 2
Sy - Block 2
Sz - Block 2
Pressure - Block 2
2000
0 50 100 150 200 250
Time (mins)
Figure 7.19 Total in-situ stresses and BHIP for block 1 and 2 –– Well A
157
Stress in block 2 is constant for first 40 minutes of injection then started falling off until
fracture is created at 68 minutes after injection (the point when minimum effective stress
becomes negative). Stresses both for block 3 and 4 are positive which indicates that no
tensile fractures are created in these blocks (as expected). Total in-situ stresses, i.e.,
stresses in x, y, and z-direction and bottomhole fluid injection pressure for block 1 and 2
are presented in Figure 7.19. Pressure and minimum total stress, i.e., Sx in block 2, is
constant until 47 minutes of injection and difference between them is positive, which is
an indication that no fractures are initiated till this time. But difference becomes negative
after 67 minutes of injection when pressure becomes larger than minimum horizontal
total stress, and fracture initiation or opening starts after this point.
Mohr-Coulomb failure criterion is used to construct the failure envelope and predict
failure based on minimum and maximum effective stress values from the coupled
simulation output and rock geomechanical data given in Table 3.3. As discussed earlier
fracture initiation or opening was observed only in block 1 and 2 for the whole injection
period. So failure envelope was constructed for only those blocks. Stresses values and
other properties used to construct Mohr-Coulomb are given in Table 7.7. Circle 1, 6 in
Table 7.7 represents 1st value of stresses for block 1 and 2 respectively. Circle 1 is for
block 1 and circle 6 is for block 2. Mohr – Coulomb failure envelope for stress values in
Table 7.7 is presented in Figure 7.20 for well A. Similar envelope can be constructed
from coupled simulations output for well B.
158
Table 7.7 Mohr-Coulomb circles input parameters – Block 1 and 2 – Well A
= 30°, = 46570 psi , = 13443 psi

Block 1 Block 2
Time (mins)
(psi) (psi) (psi) (psi)
11.51 2670 - 1461 4569 2504 Circle – 1, 6
70.30 2600 -1626 3168 -218.50 Circle – 2, 7
155.10 2265 -2329 2349 -2329 Circle – 3, 8
162.62 2439 -1924 2590 -1786 Circle – 4, 9
237.39 2237 -2296 2328 -2303 Circle – 5, 10
Mohr - Coulomb Failure Envelope

18000
Circle 1-1
Circle 2-1
16000 Circle 3-1
Circle 4-1
Circle 5-1
14000 Circle 6-2
Circle 7-2
Circle 8-2
12000
Circle 9-2
Circle 10-2
10000
Shear Stress
8000
6000
4000
2000
-2000
-2.5 -2 -1.5 -1 -0.5 0 0.5
Effective Normal Stress 4
x 10
Figure 7.20 Mohr – Coulomb failure envelope for Co= 46570psi – Well A
It is evident from Figure 7.20 that there is no shear failure during injection because
Mohr’s circle is way below the failure line. It is obvious that the dominant failure
159
mechanism in these blocks is tensile because the SPF penetrated them. Although the data
shown covers only some selected times, we also repeated the same exercise for all time
steps and confirmed that no shear failure occurred. Complete spatial map of the failure
can be obtained by plotting “stress level” S L, a feature offered in GeoSim®
(Geomechanical Simulator) which represents the ratio of the size of the Mohr circle at
any point to the circle at critical state, i.e., when the circle touches the failure line. Stress
level therefore ranges between 0 and 1. When SL < 1 there is no shear failure, and 1
represent when circle touches failure line, shear failure. Stress level for fracture # 4 after
237 minutes of injection (end of injection) is shown in Figure 7.21 for YZ crossection
through the fracture plane followed by XY areal view for the same fracture block and
injection time in Figure 7.22. There is no shear failure anywhere.
Block 2 Well
YZ cross section – Well A
160
Well
XY areal view – Well A
Two more cases were run by reducing the uniaxial compressive strength by 5 and 10
times to a base value given in Table 7.7 while keeping other parameters such friction
angle, elastic modulus and Poisson’s ratio the same as in the previous case. Mohr-
Coulomb circles for these cases are shown in Figure 7.23 and 7.26 while stress level is
shown in Figure 7.24, 7.25, 7.27 and 7.28. Circle – 10 and circle – 8 almost touch the
failure line in Figure 7.23, and stress level values both for block – 1and 2 are equal to
0.98 - very close to failure. By reducing the uniaxial compressive strength further by
factor of 2, majority of circles in Figure 7.26 touches the failure line, indicating shear
failure. The results can be confirmed both in Figure 7.27 and 7.28 as stress level are
much larger than 1, it is also important to point out that simulation for last case was run
for 141 mins as compared to other cases, which were run for 237 mins.
161
6000
Circle 1-1
Circle 2-1
Circle 3-1
5000 Circle 4-1
Circle 5-1
Circle 6-2
Circle 7-2
4000 Circle 8-2
Circle 9-2
Circle 10-2
Shear Stress
3000
2000
1000
-1000
-5000 -4000 -3000 -2000 -1000 0 1000 2000 3000 4000 5000
Effective Normal Stress
Figure 7.23 Mohr – Coulomb failure envelope for Co = 9314 psi – Well A
Block 2 Well
162
Well

4000
Circle 1-1
Circle 2-1
3500 Circle 3-1
Circle 4-1
Circle 5-1
Circle 6-2
3000
Circle 7-2
Circle 8-2
Circle 9-2
2500
Circle 10-2
Shear Stress
2000
1500
1000
500
-500
-3000 -2000 -1000 0 1000 2000 3000 4000 5000
Effective Normal Stress
Figure 7.26 Mohr – Coulomb failure envelope for Co= 4657psi –– Well A
163
Block 2 Well
Well
164
From above simulation runs, Mohr-Coulomb failure analysis and knowledge of presence
of micro cracks and heterogeneities in tight sands as recorded by microseismic, it is
concluded that the high original value of uniaxial compressive strength (which does not
allow any shear events) is unlikely. Reducing the C0 to account for weak planes and
natural fractures then will predict possibility of shear fracturing and shear-generated SRV
creation.
7.6 SUMMARY
 The fracture permeability factor (Rfa) in injection model has little effect on the
slope of pressure. However, it shifts the pressure curve upward.
 Confining the fracture propagation helped to increase the pressure and thus nearly
matches the field data.
 Decreasing value of Biot’s constant results in a decrease of effective stresses. For
larger values of Biot’s constant it is very difficult to fracture the formation.
 Correct value of Biot’s constant should be used as it affects the results, especially
in matching injection pressure. This should be done using laboratory
measurement of bulk and grain modulus or triaxial testing.
 No shear events were predicted in injection cases if a high value of uniaxial
compressive strength of rock is used. However, shear failure is detected if a lower
(effective) value of UCS is used.
 Microseismic events detected in the field during stimulation can be attributed to
opening of existing micro cracks in reservoir or creating shear failure.
165
CHAPTER EIGHT: RESULTS AND CONCLUSIONS
The following main conclusions are drawn from the studies presented in previous
chapters:
For production history matching:
 The method for calculating stress changes presented in Chapter 6 proves to be a
good approximation to capture poroelastic effects and can be used in conventional
reservoir simulation without running fully coupled geomechanical simulation.
 Moving from uncoupled to fully coupled geomechanical modeling without
correcting data sets for poroelastic effects will results in higher production. Data
sets must be corrected for stress changes in the reservoirs. Constant mean total
stress assumption is not a realistic approach and will always results in
overestimated production. Under depletion, the total mean stress is decreasing due
to poroelasticity.
 Using baseline propped fracture conductivity in simulation and other calculations
will always over estimate results; so realistic proppant conductivity should be
used.
 Net pay thickness has little effects on production in early time but affects
production at late time.
 Size and permeability of Stimulated Reservoir Volume (SRV) has significant
effects on results. Permeability inside the SRV must be larger than formation
permeability, and both are stress- or pressure-dependent during production.
166
 An accurate history match is impossible without use of pressure / stress dependent
permeability.
 Length of Stimulated Reservoir Volume (SRV) is generally much larger than
propped fracture length.
 The dimensions of SRV are different for x-linked and slick water fracturing wells
– shorter for x-linked fracture job.
 A realistic and accurate way to obtain a history match is the use of pressure
dependent permeability multipliers both inside and outside SRV.
 Permeability of the region inside SRV has larger pressure dependence than region
outside SRV due to presence of complex network of induced fractures.
 The two wells seem to have somewhat different virgin reservoir permeability.
For modeling of the fracturing process:
 Permeability reduction factor (Rfa) in injection model has no effects on results. It
shifts the pressure curve upward
 Confining the length of fracture propagation causes to increase of pressure in later
part of the job and thus improves the matches to field pressure.
 Decreasing the value of Biot’s constant results in decrease of effective stresses.
For larger values of Biot’s constant it is very difficult to fracture the formation.
 No shear events are detected in injection cases when a high value of uniaxial
compressive strength of the rock (UCS) is assumed, representative of intact rock.
167
 Shear failure regions around the fractures are predicted when the UCS is lowered
to represent media with pre-existing fractures or planes of weakness.
 Microseismic events detected on field can be attributed to opening of existing
micro cracks in reservoir or secondary shear fracturing. Therefore the increase of
permeability in the SRV can be a combination of shear fractures and stress
sensitivity of the matrix.
Our conclusions are broadly in agreement with laboratory experiments and petro physical
data on tight gas sands, and with the interpretation of microseismic data. They are also
supported by theoretical correlations for stress-sensitive permeability and results of other
geomechanical modeling.
168
REFERENCES
American Petroleum Institute: “Recommended Practices for Evaluating Short Term Proppant
Pack Conductivity”, API RP 61, Oct. 1989.
Anderson, R. A., Ingram, D. S., and Zanier, A. M. 1973. Determining Fracture Pressure
Gradients from Well Logs. J. Pet Tech 25 (11): 1259-1268. SPE-4135-PA.
Barree, R. D., Cox, S. A., Barree, V. L. et al. 2003. Realistic Assessment of Proppant Pack
Conductivity for Materials Selection. Paper SPE 84306 presented at the SPE Annual
Technical Conference and Exhibition, Denver, Colorado, 5-8 October.
Bell, J. S., and Babcock, E. A. 1986. The Stress Regime of the Western Canadian Basin and
Implications for Hydrocarbons Production. Bull. Of Cdn. Pet. Geo. 34 (3): 364-378.
Bell, J. S., Price, P. R., and McLellan, P. J. 1994. In-situ stress in the western Canada
Sedimentary Basin. In Geological atlas of the western Canada sedimentary Basin.
Mossop, G.D., and Shetsen, I.: Canadian Society of Petroleum Geologists and Alberta
Research Council, 439 – 446.
Biot, M.A. 1941. General Theory of Three Dimensional Consolidation. J. Applied Physics 12:
155-164.
Breckels, I. M., and Van Eekelen, H. A. M. 1982. Relationship between Horizontal Stress and
Depth in Sedimentary Basins. J. Pet Tech 34 (9): 2191-2199. SPE-10336-PA.
Brighenti, G. 1989. Effect of Confining Pressure on Gas Permeability of Tight Sandstones.

Paper IS-1989-024 presented at the ISRM International Symposium, Pau, France, Aug.
30 – Sept. 2.
Brower, K. R., and Morrow, N. R. 1985. Fluid Flow in Cracks as Related to Low-Permeability
Gas Sands. SPE Journal 25 (2): 191-201. SPE-11623-PA.
169
Buchsteiner, H., Warpinski, N. R., and Economides, M. J.1993. Stress Induced Permeability
Reduction in Fissured Reservoirs. Paper SPE 26513 presented at the SPE Annual
Technical Conference and Exhibition, Houston, Texas, 3 – 6 October.
Carpenter, C. B., and Spencer, G. B. 1940. Measurements of Compressibility of Consolidated

Oil Bearing Sandstones. U.S. Bureau of Mines, R. I. 3450.
Carr, N. L., Kobayashi, R., and Burrows, D. B. 1954. Viscosity of Hydrocarbon Gases under
Pressure. J. Pet Tech 6 (10): 47-55. SPE-297-G.
Cater, R. D. Derivation of the General Equation for Estimating the Extent of the Fractured
Area. Appendix to: Howard, G. C., and Fast, C. R. 1957. Optimum Fluid
Characteristics for Fracture Extension. Drilling and Production Practice, API 261-270.
Cinco-Ley, H., and Samaniego-V, F. 1981. Transient Pressure Analysis for Fractured Wells. J.
Pet Tech 33 (9): 1749-1766. SPE-7490.
Clark, J. B. 1949. A Hydraulic Process Increasing the Productivity of Wells. Trans. AIME,
186, 1-8.
Cleary, M. P., Keck, R. G., and Mear, M. E. 1983. Microcomputer Models for the Design of
Hydraulic Fractures. Paper SPE 11628 presented at the SPE/DOE Low Permeability
Gas Reservoirs Symposium, Denver, Colorado, 14-16 March.
Cleary, M. P., and Lam, K. Y. 1983. Development of Fully a Three-Dimensional Simulator for
Analysis and Design of Hydraulic Fracturing. Paper SPE 11631 presented at the
SPE/DOE Low Permeability Gas Reservoirs Symposium, Denver, Colorado, 14-16
March.
Clifton, R. J., and Abou-Sayed, A. S. 1981. A Variational Approach to the Prediction of the
Three Dimensional Geometry of the Hydraulic Fractures. Paper SPE 9879 presented at
the SPE/DOE Low Permeability Gas Reservoirs Symposium, Denver, Colorado, 27-29
May.
170
Davies, J. P., and Davies, D. K. 2001. Stress Dependent Permeability: Characterization and
Modeling. SPE Journal 6 (2): 224-235. SPE-71750-PA.
Dedurin, A. V., Majar, V. A., Voronkov, A. A. et al. 2006. Designing Hydraulic Fractures in
Russian Oil and Gas Fields to Accommodate Non-Darcy and Multiphase Flow. Paper
SPE 101821 presented at the SPE Russian Oil and Gas Technical Conference and
Exhibition, Moscow, 3-6 October.
Dobrynin, V. M. 1962. Effect of Overburden Pressure on Some Properties of Sandstones. SPE

Journal 2 (4): 360-366. SPE-461-PA.
Dranchuck, P. M., and Abu-Kassem, J. H. 1975. Calculation of Z Factors for Natural Gases
Using Equations of State. J. Cdn. Pet. Tech. 14 (3): 34-36. SPE-75-03-03.
Dranchuck, P. M., Purvis, R. A., and Robinson, D. B., 1974. Computer Calculation of Natural
Gas Compressibility Factors Using the Standing and Katz Correlation. Paper SPE 73-
112 presented at the SPE Annual Technical Meeting, Edmonton, 8-12 May.
Economides, M. J., and Nolte, K .G. 2000. Reservoir Simulation, Third Ed. John Wiley &
Sons, Inc.
Fatt, I. 1953. The Effect of Overburden Pressure on Relative Permeability. Trans. AIME, 198,
325-326.
Fatt, I., and Davis, D. H. 1952. Reduction in Permeability with Overburden Pressure. Trans.
AIME, 195, 329.
Fjær, E., Holt, R. M., Horsrud, P. et al. 2008. Petroleum Related Rock Mechanics, 2nd Edition.
Elsevier. 21-36.
Gangi, A. F. 1978. Variations of Whole and Fractured Porous Rock Permeability with
Confining Pressure. Int. J. Rock. Mech. Min. Sci. & Geomech. Abstr. 15: 249-257.
171
Geertsma, J., and de Klerk, F. 1969. A Rapid Method of Predicting Width and Extent of
Hydraulically Induced Fractures. J. Pet Tech 21 (12): 1571-1581. SPE-2458-PA.
Gray, D. H., Fatt, I., and Bergamini, G. 1963. The Effect of Stress on Permeability of
Sandstone Cores. SPE Journal 3 (2): 95-100. SPE-531-PA.
Hall, H. N. 1953. Compressibility of Reservoir Rocks. Trans. AIME, 198, 309-311.
Harrison, E., Kieschnick, W. F., and McGuire, W. J. 1953. The Mechanics of Fracture
Induction and Extension. Trans. AIME, 201, 252-321.
Holditch, S. A. 2006. Tight Gas Sands. Distinguished Author Series, J. Pet Tech 58 (6): 86-93.
SPE-103356.
Hubbert, M. K., and Willis, D. G. 1957. Mechanics of Hydraulic Fracturing. Trans. AIME,
210, 153-168.
International Organization for Standardization: “Procedures for Measuring the Long Term
Conductivity of Proppants”, ISO 13503-5, July, 2006.
Jacob, C. E. 1940. On the Flow of Water in an Elastic Artesian Aquifers. Trans. AGU, 574-
586.
Jacob, C. E. 1950. Engineering Hydraulics, New York: John Wiley & Sons, Inc. 321-386.
Jennings, J. B., Carroll, H. B., and Raible, C. J. 1981. The Relationship of Permeability to
Confining Pressure in Low Permeability Rock. Paper SPE 9870 presented at the
SPE/DOE Low Permeability Gas Reservoirs Symposium, Denver, Colorado, 27-29
May.
Ji, L., Settari, A., Orr, D. W. et al. 2004a. Methods for Modelling Static Fractures in Reservoir
Simulation. Paper SPE 2004-260 presented at the SPE Canadian International
Petroleum Conference, Calgary, Alberta, 8-10 June.
172
Ji, L. Settari, A. Sullivan, R. B. et al. 2004b. Methods for Modeling Dynamic Fractures in
Coupled Reservoir and Geomechanics Simulation. Paper SPE 90874 presented at the
SPE Annual Technical Conference and Exhibition, Houston, Texas, 26 – 29 September.
Ji, L. Settari, A. Sullivan, and R. B. 2009. A Novel Hydraulic Fracturing Model Fully Coupled
with Geomechanics and Reservoir Simulation. SPE Journal 14 (3): 423-430. SPE-
110845-PA.
Jones, F. O. and Owens, W. W. 1980. A Laboratory Study of Low-Permeability Gas Sands. J.

Pet Tech 32 (9): 1631-1640. SPE-7551-PA.
Kawata, Y. and Fujita, K. 2001. Some Publications of Possible Unconventional Hydrocarbon

Availability Until 2100. Paper SPE 68755 presented at the SPE Asia Pacific Oil and
Gas Conference, Jakarta, 17-19 April.
Khristianovic, S. A., and Zheltov, Y. P. 1955. Formation of Vertical Fractures by Means of

Highly Viscous Liquid. Proc., Fourth World Pet. Cong., Rome, Sec. 2, 579-586.
Lee, A. L., Gonzalez, M. H., and Eakin, B. E. 1966. The Viscosity of Natural Gases. J. Pet
Tech 18 (8): 997-1000. SPE-1340-PA.
Lei, Q., Xiong, W., Yuan, J. et al. 2007. Analysis of Stress Sensitivity and Its Influence on Oil
Production from Tight Reservoirs. Paper SPE 111148 presented at the SPE Eastern
Regional Meeting, Lexington, Kentucky, 17 – 19 October.
Lin, J., and Zhu, D. 2012. Predicting Well Performance in Complex Fracture Systems by Slab
Source Method. Paper SPE 151960 presented at SPE Hydraulic Fracturing Technology
Conference, The Woodlands, Texas, 6-8 February.
Liu, X., and Civan, F. 1995. Formation Damage by Fines Migration Including Effects of Filter
Cake, Pore Compressibility, and Non-Darcy Flow - A Modeling Approach to Scaling
from Core to Field. Paper SPE 28980 presented at the SPE International Symposium on
Oilfield Chemistry, San Antonio, Texas, 14 – 17 February.
173
Liu, Y., and Sharma, M. M. 2005. Effect of Fracture Width and Fluid Rheology on Proppant
Settling and Retardation; An Experimental Study. Paper SPE 96208 presented at the
SPE Annual Technical Conference and Exhibition, Dallas, Texas, 9-12 October.
Liu, Y., Gadde, P. B., and Sharma, M. M. 2006. Proppant Placement using Reverse-Hybrid
Fracs. Paper SPE 99580 presented at the SPE Gas Technology Symposium, Calgary,
Alberta, 15 – 17 May.
Lorenz, J. C. 1999. Stress-Sensitive Reservoirs. J. Pet Tech: 61-63.
Masters, J.A. 1979. Deep Basin Gas Trap, Western Canada. AAPG Bulletin 63 (2): 152-181.
Mayerhofer, M. J., and Meehan, D. J. 1998. Waterfracs - Results from 50 Cotton Valley Wells.
Paper SPE 49104 presented at the SPE Annual Technical Conference and Exhibition,
New Orleans, Louisiana, 27 – 30 September.
McGuire, W. J. and Sikora, V. J. 1960. The Effect of Vertical Fractures on Well Productivity.
J. Pet Tech 12 (10): 72-74. SPE-1618-G.
McKee, C. R., Bumb, A. C., and Koenig, R. A. 1988. Stress-Dependent Permeability and
Porosity of Coal and Other Geologic Formations. SPE Form Eval 3 (1): 81-91. SPE-
12858-PA.
McLatchie, A. S., Hemstock, R. A., and Young, J. W. 1958. The Effect of Compressibility of
Reservoir Rock and Its Effects on Permeability. J. Pet Tech 10 (6): 49-51. SPE-894-G.
Meinzer, O. E. 1928. Compressibility and Elasticity of Artesian Aquifers. Econ. Geol. 23: 263-
271.
Montgomery, C. T. and Steanson, R. E. 1985. Proppant Selection: The Key to Successful

Fracture Stimulation. J. Pet Tech 37 (12): 2163-2172. SPE-12616.
Mossop, G. D. and Shetsen, I. (comp.). 1994. Geological atlas of the Western Canada
Sedimentary Basin; Canadian Society of Petroleum Geologists and Alberta Research
174
Council, URL <http://www.ags.gov.ab.ca/publications/wcsb_atlas/atlas.html>, [August
27, 2012].
Much, M. G. and Penny, G. S. 1987. Long Term Performance of Proppants under Simulated
Reservoir Conditions. Paper SPE 16415 presented at the SPE Low Permeability
Reservoirs Symposium, Denver, Colorado, 18-19 May.
Nassir, M., Settari, A., and Wan, R. G. 2012. Prediction and Optimization of Fracturing in
Tight Gas and Shale Using a Coupled Geomechanical Model of Combined Tensile and
Shear Fracturing. Paper SPE 152200 presented at the SPE Hydraulic Fracturing
Technology Conference, The Woodlands, Texas, 6-8 February.
Newberry, B. M., Nelson, R. F., and Ahmed, U. 1985. Prediction of Vertical Hydraulic
Fracture Migration Using Compressional and Shear Wave Slowness. Paper SPE 13895
presented at the SPE/DOE Low Permeability Gas Reservoirs Symposium, Denver,
Colorado, 19-22 March.
Nordgren, R. P. 1972. Propagation of a Vertical Hydraulic Fracture. SPE Journal 12 (4): 306-
314. SPE-3009-PA.
Olson, J. E., Bahorich, B., and Holder, J. 2012. Examining Hydraulic Fracture – Natural
Fracture Interaction in Hydrostone Block Experiments. Paper SPE 152618 presented at
SPE Hydraulic Fracturing Technology Conference, The Woodlands, Texas, 6-8
February.
Olson, J. E., and Wu, K. 2012. Sequential vs. Simultaneous Multizone Fracturing in Horizontal
Wells: Insights From a Non-Planar, Multifrac Numerical Model. Paper SPE 152602
presented at the SPE Hydraulic Fracturing Technology Conference, The Woodlands,
Texas, 6-8 February.
Ostensen, R. W. 1983. Microcrack Permeability in Tight Gas Sandstone. SPE Journal 23 (6):
919-927. SPE-10924-PA.
175
Ostensen, R. W. 1986. The Effect of Stress Dependent Permeability on Gas Production and
Well Testing. SPE Formation Evaluation 1 (3): 227-235. SPE-11220-PA.
Palisch, T., Duenckel, R., Bazan, L. et al. 2007. Determining Realistic Fracture Conductivity
and Understanding its Impact on Well Performance – Theory and Field Examples.
Paper SPE 106301 presented at the SPE Hydraulic Fracturing Technology Conference,
College Station, Texas, 29-31 January.
Palisch, T. T., Vincent, M. C., and Handren, P. J. 2010. Slickwater Fracturing: Food for
Thought. SPE Prod & Oper 25 (3): 327-344. SPE-115766-PA.
Pearson, C. M. 2001. Dimensionless Fracture Conductivity: Better Input Vales Make Better
Wells. J. Pet Tech 53 (1): 59-63. SPE-60184.
Perkins, T. K., and Kern, L. R. 1961. Widths of Hydraulic Fractures. J. Pet Tech 13 (9): 937-
949. SPE-89-PA.
Prats, M. 1961. Effect of Vertical Fractures on Reservoir Behavior – Incompressible Fluid

Case. SPE Journal 1 (2): 103-118. SPE-1575-G.
Reyes, L., and Osisanya, S. O. 2002. Empirical Correlation of Effective Stress Dependent
Shale Rock Properties. J. Cdn. Pet. Tech. 41 (12). SPE-02-12-02.
Rogner, H. H. 1996. An assessment of World Hydrocarbon Resources. IIASA, WP-96-26,

Luxemburg, Austria, May 1996.
Rushing, J. A., and Sullivan, R. B. 2003. Evaluation of a Hybrid-Frac Stimulation Technology

in the Bossier Tight Gas Sand Play. Paper SPE 84394 presented at the SPE Annual
Technical Conference and Exhibition, Denver, Colorado, 5-8 October.
Sampath, K., and Keighin, C. W. 1982. Factors Affecting Gas Slippage in Tight Sandstones of
Cretaceous Age in the Uinta Basin. J. Pet Tech 34 (11): 2715-2720. SPE-9872-PA.
176
Schubarth, S. K., Spivey, J. P., and Huckabee, P. T. 2006. Using Reservoir Modeling To
Evaluate Stimulation Effectiveness in Multilayered Tight Gas Reservoirs: A Case
History in the Pinedale Anticline Area. Paper SPE 100574 presented at the SPE Gas
Technology Symposium, Calgary, Alberta, 15 –175 May.
Settari, A., Puchyr, P. J., and Bachman, R. C. 1990. Partially Decoupled Modeling of
Hydraulic Fracturing Process. SPE Prod Eng 5 (1): 37-44. SPE-16031.
Settari, A. Sullivan, R. B., Walters, D. A. et al. 2002a. 3-D Analysis and Prediction of
Microseismicity in Fracturing by Coupled Geomechanical Modeling. Paper SPE 75714
presented at the SPE Gas Technology Symposium, Calgary, Alberta, 28 April – 2 May.
Settari, A. Sullivan, R. B., and Bachman, R. C. 2002b. The Modeling of the Effect of Water
Blockage and Geomechanics in Waterfrac. Paper SPE 77600 presented at the SPE
Annual Technical Conference and Exhibition, San Antonio, Texas, Sept. 29 – Oct 2.
Settari, A., Bachman, R. C., and Walters, D. A. 2005. How to Approximate Effects of
Geomechanics in Conventional Reservoir Simulation. Paper SPE 97155 presented at
the SPE Annual Technical Conference and Exhibition, Dallas, Texas, 9-12 October.
Settari, A., Sullivan, R. B., Rother, R. et al. 2009. Comprehensive Coupled Modeling Analysis
of Stimulations and Post-Frac Productivity – Case Study of the Wamsutter Field. Paper
SPE 119394 presented at the SPE Hydraulic Fracturing Technology Conference, The
Woodlands, Texas, 19–21 January.
Sharma, M. M., Gadde, P. B., Sullivan, R. B. et al. 2004. Slick Water and Hybrid Fracs in the
Bossier: Some Lessons Learnt. Paper SPE 89876 presented at the SPE Annual
Technical Conference and Exhibition, Houston, Texas, 26 – 29 September.
Sharma, M. M., Gadde, P. B., Liu, Y. et al. 2005. Selecting Strategies for Improving
Performance in Tight Gas Sands. Gas Tips 11 (4): 8-11.
177
Sneddon, I. N., and Lowengrud, M. 1969. Crack Problems in the Classical Theory of
Elasticity, 20-30. New York: SIAM Series in Applied Mathematics, John Wiley &
Sons.
Sutton, R. P. 1985. Compressibility Factors for High-Molecular-Weight Reservoir Gases.

Paper SPE 14265 presented at the SPE Annual Technical Conference and Exhibition,
Las Vegas, Nevada, 22 – 25 September.
Tao, Q., Ehlig-Economides, C. A., and Ghassemi, A. 2009. Investigation of Stress-Dependent

Fracture Permeability in Naturally Fractured Reservoir Using a Fully Coupled
Poroelastic Displacement Discontinuity Model. Paper SPE 124745 presented at the SPE
Annual Technical Conference and Exhibition, New Orleans, Louisiana, 4-7 October.
Thomas, R. D., and Ward, D. C. 1972. Effect of Overburden Pressure and Water Saturation on
Gas Permeability of Tight Sandstone Cores. J. Pet Tech 24 (2): 120-124. SPE-3634-
PA.
Urbancic, T., Baig, A., and Goldstein, S. 2012. Assessing Stimulation of Complex Natural
Fractures as Characterized Using Microseismicity: An Argument the Inclusion of Sub-
Horizontal Fractures in Reservoir Models. Paper SPE 152616 presented at SPE
Hydraulic Fracturing Technology Conference, The Woodlands, Texas, 6-8 February.
Vairogs, J., Hearn, C. L., Dearing, D. W. et al. 1971. Effect of Rock Stress on Gas Production
from Low Permeability Reservoirs. J. Pet Tech 23 (9): 1161-1167. SPE-3001-PA.
Vairogs, J., and Rhoades, V. W. 1973. Pressure Transient Tests in Formations Having Stress
Sensitive Permeability. J. Pet Tech 25 (8): 965-970. SPE-4050-PA.
Vermylen, J. P., and Zoback, M. D. 2011. Hydraulic Fracturing, Microseismic Magnitudes,

and Stress Evolution in the Barnett Shale, Texas, USA. Paper SPE 140507 presented at
SPE Hydraulic Fracturing Technology Conference, The Woodlands, Texas, 24-26
January.
178
Walls, J. D. 1982. Tight Gas Sands-Permeability, Pore Structure, and Clay. J. Pet Tech 34 (11):
2708-2714. SPE-9871-PA.
Walsh, J. B. 1981. Effect of Pore Pressure and Confining Pressure on Fracture Permeability.
Int. J. Rock. Mech. Min. Sci. & Geomech. Abstr. 18: 429-435.
Warpinski, N. R., and Teufel, L. W. 1987. Influence of Geologic Discontinuities on Hydraulic

Fracture Propagation. J. Pet Tech 39 (2): 209-220. SPE-13224-PA, SPE – 17011, and
SPE – 17074.
Warpinski, N. R. 1991. Hydraulic Fracturing in Tight, Fissured Media. J. Pet Tech 43 (2): 146-
151, 208-209. SPE-20154-G.
Warpinski, N. R., Lorenz, J. C., Branagan, P. T. et al. 1993. Examination of a Cored Hydraulic
Fracture in a Deep Gas Well. SPE Prod & Fac 8 (3): 150-158. SPE-22876-PA, SPE –
26302, and SPE – 26946.
Weng, X., Kresse, O., Cohen., C. et al. 2011. Modeling of Hydraulic-Fracture-Network

Propagation in a Naturally Fractured Formation. SPE Prod & Oper 26 (4): 368-380.
SPE-140253-PA.
Wu, Y. S., Rutqvist, J., Karasaki, K. et al. 2008. A Mathematical Model for Rock
Deformation’s Effect on Flow in Porous and Fractured Reservoirs. Paper SPE 08-142
presented at the 42nd US and 2nd Canadian Rock Mechanics Symposium (UCRMS), San
Francisco, California, June 29 – July 2.
Wu, R., Kresse, O., Weng, X. et al. 2012. Modeling of Interaction of Hydraulic Fractures in
Complex Fracture Networks. Paper SPE 152052 presented at SPE Hydraulic Fracturing
Technology Conference, The Woodlands, Texas, 6-8 February.
Wyble, D. O. 1958. Effect of Applied Pressure on the Conductivity, Porosity and Permeability
of Sandstones. J. Pet Tech 10 (11): 57-59. SPE-1081-G.
179
Yarborough, L., and Hall, K. R. 1974. How to Solve Equation-of-State for Z-factors. Oil and
Gas Journal: 86-88.
180
APPENDIX A: SELECTED DATA FILES
GeoSim requires three input files to run a coupled problem. These files have different
extensions and are used for different parts of model.
Case – 6.6.dat : The Input File for Flow Model

'
' **** TERASIM RESERVOIR SIMULATOR ****
'
' ==== ELEMENT OF SYMMETRY MODEL ====
' Stimulated Well - no water (comparison run)
' Gas Reservoir with Inactive Bottom Water
' Cartesian COORDS
'
ATITLE= DRY GAS WELL A
' ====================================================================================
' MAIN CONTROL
' ====================================================================================
'
' KEYWORD VALUE
' ------- -------
GEOM= 3-D ' AREAL FOR SINGLE LAYER, 3-D FOR MULTI-LAYER
UNITS= FIELD ' FIELD OR SI
NINE= OFF ' LEAVE OFF FOR FRAC PROBLEMS
INIT= ON ' INITIALIZATION: INIT = ON OFF
NIN= 0 ' RESTART CONTROL: NIN = (LAST COMPLETED TS)
ALIGN= OFF ' USED ONLY FOR NESTED GRID PROBLEMS
IREG= OFF ' NO. OF EQUILIBRIUM REGIONS FOR INITIALIZATION
AREG= OFF ' NO. OF ACCOUNTING REGIONS
MROCK= OFF ' MULTIPLE ROCK TYPES
PFRAC= OFF ' PRESSURE DEPENDENT WATER RELATIVE PERMEABILITIES
HORIZ-VERT ' ENTER BOTH HORIZONTAL AND VERTICAL PERMS
GAS= DRY ' BLACK OIL MODE ON OR OFF
WATER= ON
OIL= NONE
CNARRAY
PXMULT = ON
PYMULT = ON
PZMULT = ON
END
' ====================================================================================
' LINK TO GEOSIM - GEOMECHANICAL SIMULATOR
' ====================================================================================
GMECH = EXPL
'
' DIMENSIONS
'
NWMAX=1 ' MAXIMUM NUMBER OF WELLS
NLMAX=8 ' MAXIMUM NUMBER OF COMPLETIONS (PERFORATIONS) PER WELL
END
181
' ====================================================================================
‘RESERVOIR DESCRIPTION - (GRID LISTS)
' ====================================================================================
' ==== NUMBER OF GRID CELLS: FUNDAMENTAL GRIDS ====
' X-DIREC Y-DIREC(V) Z-DIREC(W)
N1=110 N2=43 N3=17
'
' ==== X - DIREC GRID DIMENSIONS (N=1, NX) ====
'
+I=1 2 3 5 10 10 30 30 30 30 10 10 5 4 2 4 5 10 10 30 30 30 30 10 10 5 4 2 4 5 10 10 30 30 30 30 10 10 5 4 2 4 5
+I=43 10 10 30 30 30 30 10 10 5 4 2 4 5 10 10 30 30 30 30 10 10 5 4 2 4 5 10 10 30 30 30 30 10 10 5 4 2 4 5 10 10
+I=84 30 30 30 30 10 10 5 4 2 4 5 10 10 30 30 150 150 350 500 500 1000 1000 1000 1000 1000 1000 1000
END
'
' ==== y - DIREC GRID DIMENSIONS (N=2, NY) ====
'
+J=1 1000 1000 500 200 100 60 22 22 18 12 10 10 10 8 5 4 4 4 4 4 2 2 2 4 4 4 4 4 5 8 10 10 10 12 18 22 22 60 100
+J=40 200 500 1000 1000
END
'
' ==== Z - DIREC GRID DIMENSIONS (N=3, NZ) ====
'
+K=1 14 10 10 5 4 2 2 2 2 2 2 2 4 5 10 10 14
END
'
' ==== DEPTH TO THE TOP OF RESERVOIR (ONE FOR EACH GRID FOR TOP LAYER) ====
'
I=1/* J=1/* K=1/1 5250.0
END
'
' ==== POROSITY OF FORMATION (FRACTION) ====
'
+K=1/* 0.035
'
' ---- POROSITY OF PROPPED FRACTURED GRID BLOCKS (FRACTION) ----
'
I=1/1 J=15/29 K=1/* 0.37
I=14/14 J=15/29 K=1/* 0.37
I=27/27 J=15/29 K=1/* 0.37
I=40/40 J=15/29 K=1/* 0.37
I=53/53 J=15/29 K=1/* 0.37
I=66/66 J=15/29 K=1/* 0.37
I=79/79 J=15/29 K=1/* 0.37
I=92/92 J=15/29 K=1/* 0.37
'
' ---- ELEMENT OF SYMMETRY - POROSITY MULTIPLIER ----
'
MULT I=1/1 J=1/* K=1/* 0.5
END
'
' ==== PERMEABILITY (HORIZONTAL & VERTICAL) OF FORMATION (mD) ====
'
+K=1/* 0.015 0.0015
'
' ---- PERMEABILITY MULTIPLIERS OF SRV REGION (ONE WING-120X300X100 ft3)-INITIAL
PERMEABILITY MULITPLIER(IPM) ----
' HORZ VERT
MULT I=1/6 J=5/39 K=1/* 50.0 50.0
MULT I=9/19 J=5/39 K=1/* 50.0 50.0
MULT I=22/32 J=5/39 K=1/* 50.0 50.0
182
MULT I=35/45 J=5/39 K=1/* 50.0 50.0
MULT I=48/58 J=5/39 K=1/* 50.0 50.0
MULT I=61/71 J=5/39 K=1/* 50.0 50.0
MULT I=74/84 J=5/39 K=1/* 50.0 50.0
MULT I=87/97 J=5/39 K=1/* 50.0 50.0
END
'
' ==== TRANSMISSIBILITY MODIFIERS FOR FRACTUREDE BLOCKS ====
' Tx TY TZ
I=1/1 J=15/29 K=1/* 1.0 27.0 261.0
I=14/14 J=15/29 K=1/* 1.0 27.0 261.0
I=27/27 J=15/29 K=1/* 1.0 27.0 261.0
I=40/40 J=15/29 K=1/* 1.0 27.0 261.0
I=53/53 J=15/29 K=1/* 1.0 27.0 261.0
I=66/66 J=15/29 K=1/* 1.0 27.0 261.0
I=79/79 J=15/29 K=1/* 1.0 27.0 261.0
I=92/92 J=15/29 K=1/* 1.0 27.0 261.0
'
' ---- ELEMENT OF SYMMETRY - POROSITY MULTIPLIER ----
'
MULT I=1/1 J=1/* K=1/* 1.0 0.5 0.5
'
' ---- PRESSURE DEPENDENT PERMEABILITY MULITPLIERS INSIDE SRV - VARIABLE MEAN TOTAL STRESS CASE ----
'
STRMULT STRESS PRODUCTION
'
I=1/6 J=5/14 K=1/*
I=1/6 J=30/39 K=1/*
I=2/6 J=15/29 K=1/*
I=9/19 J=5/14 K=1/*
I=9/19 J=30/39 K=1/*
I=9/13 J=15/29 K=1/*
I=15/19 J=15/29 K=1/*
I=22/32 J=5/14 K=1/*
I=22/32 J=30/39 K=1/*
I=22/26 J=15/29 K=1/*
I=28/32 J=15/29 K=1/*
I=35/45 J=5/14 K=1/*
I=35/45 J=30/39 K=1/*
I=35/39 J=15/29 K=1/*
I=41/45 J=15/29 K=1/*
I=48/58 J=5/14 K=1/*
I=48/58 J=30/39 K=1/*
I=48/52 J=15/29 K=1/*
I=54/58 J=15/29 K=1/*
I=61/71 J=5/14 K=1/*
I=61/71 J=30/39 K=1/*
I=61/65 J=15/29 K=1/*
I=67/71 J=15/29 K=1/*
I=74/84 J=5/14 K=1/*
I=74/84 J=30/39 K=1/*
I=74/78 J=15/29 K=1/*
I=80/84 J=15/29 K=1/*
I=87/97 J=5/14 K=1/*
I=87/97 J=30/39 K=1/*
I=87/91 J=15/29 K=1/*
I=93/97 J=15/29 K=1/*
'
' ---- STRESS FACTOR (S) = 8.4 STARTING FROM 2000 PSI ----
'
183
' Effective Kx Ky Kz
' Stress
3688 1.00000000 1.00000000 1.00000000
3731 0.87880844 0.87880844 0.87880844
3774 0.76903837 0.76903837 0.76903837
3817 0.66989236 0.66989236 0.66989236
3859 0.58061636 0.58061636 0.58061636
3902 0.50049720 0.50049720 0.50049720
3945 0.42886013 0.42886013 0.42886013
3988 0.36506659 0.36506659 0.36506659
4031 0.30851214 0.30851214 0.30851214
4074 0.25862450 0.25862450 0.25862450
4117 0.21486170 0.21486170 0.21486170
4159 0.17671037 0.17671037 0.17671037
4202 0.14368419 0.14368419 0.14368419
4245 0.11532228 0.11532228 0.11532228
4288 0.09118791 0.09118791 0.09118791
4331 0.07086710 0.07086710 0.07086710
4374 0.05396740 0.05396740 0.05396740
4417 0.04011671 0.04011671 0.04011671
4459 0.02896224 0.02896224 0.02896224
4502 0.02016941 0.02016941 0.02016941
4545 0.01342093 0.01342093 0.01342093
'
' ---- PRESSURE DEPENDENT PERMEABILITY MULITPLIERS OUTSIDE SRV - VARIABLE MEAN TOTAL STRESS CASE ----
'
STRMULT STRESS PRODUCTION
'
I=1/* J=1/4 K=1/*
I=1/* J=40/* K=1/*
I=7/8 J=5/39 K=1/*
I=20/21 J=5/39 K=1/*
I=33/34 J=5/39 K=1/*
I=46/47 J=5/39 K=1/*
I=59/60 J=5/39 K=1/*
I=72/73 J=5/39 K=1/*
I=85/86 J=5/39 K=1/*
I=98/* J=5/39 K=1/*
'
' ---- STRESS FACTOR (S) = 6.0 STARTING FROM 2000 PSI ----
'
' Effective Kx Ky Kz
' Stress
3688 1.00000000 1.00000000 1.00000000
3731 0.91237313 0.91237313 0.91237313
3774 0.83093231 0.83093231 0.83093231
3817 0.75532405 0.75532405 0.75532405
3859 0.68521320 0.68521320 0.68521320
3902 0.62028185 0.62028185 0.62028185
3945 0.56022833 0.56022833 0.56022833
3988 0.50476632 0.50476632 0.50476632
4031 0.45362397 0.45362397 0.45362397
4074 0.40654308 0.40654308 0.40654308
4117 0.36327836 0.36327836 0.36327836
4159 0.32359670 0.32359670 0.32359670
4202 0.28727653 0.28727653 0.28727653
4245 0.25410718 0.25410718 0.25410718
4288 0.22388828 0.22388828 0.22388828
4331 0.19642926 0.19642926 0.19642926
4374 0.17154876 0.17154876 0.17154876
184
4417 0.14907421 0.14907421 0.14907421
4459 0.12884134 0.12884134 0.12884134
4502 0.11069375 0.11069375 0.11069375
4545 0.09448251 0.09448251 0.09448251
END
'
' ==== INTIAL CONDITIONS (INIAILIZATION) ====
'
' INITIAL PRESSURE
END
' GAS AND WATER SATURATION
END
PREF=2000. DREF=5300. DGOC=15000. DWOC=15000.
END
' ====================================================================================
' PVT DATA
' ====================================================================================
' PRESSURE EGT VISGT
' (PSI) (SCF/RB) (CP)
14.7 4.4892 0.01352
147 45.317 0.01358
279 86.898 0.01368
411 129.19 0.01380
676 215.82 0.01409
808 260.05 0.01426
1073 350.03 0.01465
1205 395.63 0.01486
1470 487.59 0.01534
1602 533.74 0.01559
1866 625.81 0.01614
2028 681.59 0.01650
2205 741.73 0.01691
2557 858.74 0.01777
2734 915.18 0.01821
2910 970.02 0.01866
3087 1023.1 0.01912
3248 1070.2 0.01955
3513 1143.9 0.02025
3627 1174.3 0.02054
4000 1268.5 0.02152
8000 1916.5 0.03056
12000 2255.7 0.03743
16000 2481.2 0.04311
20000 2649.7 0.04805
END
' ====================================================================================
' PHYSICAL PROPERTIES DATA
' ====================================================================================
' ==== DENSITIES OF OIL-GAS-WATER: (LBM/FT**3) ====
'
DENO= 52.10 DENG= 0.04688 DENW= 62.4
'
' ==== COMPRESSIBILITIES OF FORMATION-OIL-WATER: (ONE/PSI) ====
'
CFOR=3.500E-06 CO=8.38E-06 CW=3.10E-06
'
' ==== WATER INPUT FORM VOL FACT; REF PRES (FORM-WATER): (PSI ====
'
BWI= 1.026 PFOR= 500.0 PW= 500.0
'
185
' ==== VISCOSITY OF WATER (CP) VISCOSITY PRESSURE VARIANCE: (CP/PSI) ====
'
VWI= 0.25
'
' ==== OIL VISCOSITY PRESSURE VARIANCE ABOVE THE BUBBLE POINT: (CP/PSI) ====
'
CVO=8.72D-05 ' =(VO2-VO1)/(P2-P1)
'
' ==== WATER VISCOSITY PRESSURE VARIANCE: (CP/PSI) ====
'
CVW=0.0000000 ' NORMALLY SET TO ZERO
'
=====================================================================================
' RELATIVE PERMEABILITY & CAPILLARY PRESSURE
' ====================================================================================
' GAS RELATIVE PERMEABILITY CURVE FITTING QUADRATIC
GQUAD=OFF ' NORMALLY OFF
' SW KRW KRG PCOG
' ==== ==== ==== ====
0.25 0 1.000 1000.3
0.3 0.007 0.793 271.80
0.35 0.016 0.69 110.54
0.4 0.029 0.608 60.298
0.45 0.047 0.52 38.098
0.5 0.072 0.453 26.232
0.55 0.105 0.38 19.377
0.6 0.148 0.320 14.533
0.65 0.201 0.26 9.6890
0.7 0.266 0.205 4.8404
0.8 0.438 0.111 4.4760
0.9 0.679 0.039 3.9670
1 1.000 0.000 0.0000
END
' ====================================================================================
' RECURRENT DATA
' ====================================================================================
'
TIME= 15/06/2006 START
'
' ====================================================================================
' SIMULATOR CONTROLS
' ====================================================================================
' ==== CONVERGENCE TOLERENCE ====
'
RTOL 0.01 ' RESIDUAL COVERGENCE TOLERANCE (RANGE .0001 TO .5)
PTOL .100000E-00 ' PRESSURE COVERGENCE TOLERANCE (RANGE .1 TO 100.)
STOL .500000E-02 ' SATURATION CONVERGENCE TOLERANCE (RANGE .0001 TO .5)
'
' ==== AUTOMATIC TIME STEP CONTROLS ====
'
PNORM= 500.00 ' MAX GRID BLOCK PRESSURE CHANGE (PSI/KPA)
BPNORM= 500.00 ' MAX GRID BLOCK BUBBLE POINT PRESSURE CHANGE (PSI/KPA)
SNORM= 0.2 ' MAX GRID BLOCK SATURATION CHANGE (FRACTION)
'
' ==== TIME STEP CONTROLS ====
'
TS = 0.5 ' FIRST TIMESTEP SIZE (DAYS)
M-TS = 10.0 ' MAXIMUM TIMESTEP SIZE (DAYS)
N-TS = 99999 ' MAXIMUM NUMBER OF TIMESTEPS (SET TO LARGE VALUE)
'
186
' ==== IMPLICIT SWITCHING CONTROLS (THRESHOLDS) ====
'
PNIMP= 2000 ' BUBBLE POINT THRESHOLD (USUALLY 1/4 OF BPNORM)
SNIMP= 0.1 ' SATURATION THRESHOLD (USUALLY 1/4 OF SNORM)
'
' ==== SOLUTION CONTROLS ====
'
IMPLICIT ' FORMULATION (RE) SETTING: IMPLICIT EXPLICIT OFF (START WITH EXPLICIT)
NEIGH=OFF ' IMPL/EXPL WELL CELLS: NEIGH= ON PLUS (USUALLY OFF)
INFO= OFF ' GENERAL OUTPUT DETAIL: INFO= ON OFF PLUS (USUALLY OFF)
'
' ==== OUTPUT CONTROLS ====
'
R-OUT=ON
W-OUT=ON
RESTART=OFF
FORECAST=1
CPLOT=ON
PR+W PR+G PR+P PR-B PR+V PR+K PR+C
W-LONG
'
' ====================================================================================
' WELL DESCRIPTION
' ====================================================================================
WELL=PROD START NEW
PRODUCER
DIREC=1
WR= 0.2500 CC=0.37000 FF=1.000000 SKIN= 0.0000
' HORIZONTAL WELL COMPLETION - PERFORATION LOCATIONS
LAYER=1 I=1 J=22 K=9 GRID=1 LMULT = 1000.0 HDOFF= 0.0
'
CONSTRAINT = 1 GAS = 7000000.00000 ' SCF/DAY
CONSTRAINT = 2 PRESS = 300.0000 Min ' PSI
'
LIST=SWITCH
BACK=OFF
WCUT= OFF
'
' ==== TIME AND ACTION ====
TIME= 1.0 CONTINUE
TIME= 4.0 CONTINUE
TIME= 15.0 CONTINUE
TIME= 30.0 CONTINUE
TIME= 60.0 CONTINUE
TIME= 85.0 CONTINUE
TIME= 166.0 CONTINUE
TIME= 1296.0 STOP
187
Case – 6.6.fem : The Input File for Stress - Strain Model
Stress - Strain Model - Case 6.6
Metric units, 3D , 1 material, linearly elastic
COM ( NOTE: above TITLE must be 2 lines )
COM
COM ***********************************
COM * Taurus Reservoir Solutions Ltd. *
COM * FEM3D *
COM * 3-D STRESS STRAIN MODEL *
COM * VERSION 3.11 *
COM * 06-DECEMBER-2004 *
COM ***********************************
COM
COM RECORD B
COM
COM --- IUNIT,INPUT UNITS : 0 = metric (kPA, m, deg C)
COM | 1 = English (psia, ft, deg F)
COM |
COM |
COM V
COM *----*
1
COM *----*
COM RECORD C
COM
COM ,--Ngrids TOTAL NUMBER OF GRIDS AND SUB-GRIDS
COM | ,--REFERENCE DEPTH TO THE TOP OF GRID
COM | | (depth is negative downward)
COM | | ,--A_type, analysis geometry type:
COM | | | 0=3D_Solid, 1=Radial
COM | | | ,--Mat_cnt, NUMBER OF MATERIALS
COM | | | | ,--Mat_hyst, STRESS-STRAIN HYSTERESIS
COM | | | | | 0=no, 1=yes (all mat.types)
COM | | | | | ,--Iter_flg, SOLVER OPTIONS:
COM | | | | | | 0=direct solver;1=ILU iterative solver
COM | | | | | | 3=CG solver
COM | | | | | | ,--Th_comp_flg, Thermal compaction
COM | | | | | | | and hardening option(uses Temp_ref
COM | | | | | | | of .gii file in calcs):
COM | | | | | | | 0=no; 1=thermal compaction
COM | | | | | | | ,--Read_flg, Grid read option:
COM | | | | | | | | 0=read dx,dy,dz data;
COM | | | | | | | | 1,2=read CPG coord by nodes
COM | | | | | | | | 1=reserv.grid only
COM | | | | | | | | 2=read total grid
COM | | | | | | | | 3=read faulted FEM grid by
COM | | | | | | | | FaultPrep preprocessor
COM | | | | | | | | ,--Bore Radius for Radial analyses
COM | | | | | | | | | ,---G_type, grid type
COM | | | | | | | | | | 1=Continuous/Refined,
COM | | | | | | | | | | 2= Faulted
COM | V | | | | | | V |
COM V m/ft V V V V V V m/ft V
COM *---------|----*----*----*----*----*----*----*----*
1 -5250 0 1 0 1 0 0 0 1
COM *---------|----*----*----*----*----*----*----*----*
COM
COM RECORD C1
COM
188
COM Read always - PRIMARY GRID SIZE
COM ,--Nx,NUMBER OF Primary GRID BLOCKS IN X DIRECTION
COM | ,--Ny,NUMBER OF Primary GRID BLOCKS IN Y DIRECTION
COM | | ,--Nz,NUMBER OF PRIMARY GRID BLOCKS IN Z DIRECTION
COM | | |
COM | | |
COM V V V
COM *----*----*
110 43 40
COM *----*----*
COM RECORD D
COM
COM Defines relative position of the reservoir grid in the
COM primary FEM grid
COM ,--N_low_dx, NUMBER OF X-GRID BLOCKS BEFORE PRIMARY RES.GRID
COM | ,--N_low_dy,NUMBER OF Y-GRID BLOCKS BERORE PRIMARY RES. GRID
COM | | ,--N_low_dz,NUMBER Z-GRID BLOCKS BEFORE PRIMARY RES.GRID
COM | | |
COM V V V
COM *----*----*
0 0 13
COM *----*----*
COM
COM ,--N_high_dx, NUMBER OF X-GRID BLOCKS AFTER RES.GRID
COM | ,--N_high_dy,NUMBER OF Y-GRID BLOCKS AFTER RES. GRID
COM | | ,--N_high_dz,NUMBER Z-GRID BLOCKS AFTER RES.GRID
COM | | |
COM V V V
COM *----*----*
0 0 10
COM *----*----*
COM RECORD E
COM
COM *** PRIMARY GRID SIZE DATA ***
COM ( read the following 3 lines only if Read_flg = 0)
COM ( otherwise no data )
COM ---- ARRAY OF VALUES SPECIFYING LENGTHS OF GRID BLOCKS IN X DIRECTION - (Nx values ) ----
COM <---------------------------- m/ft --------------------------------------->
COM
2 3 5 10 10 30 30 30 30 10 10 5 4 2 4 5 10 10 30 30 30 30 10 10 5 4 2 4 5 10 10 30 30 30 30 10 10 5 4 2 4 5 10
10 30 30 30 30 10 10 5 4 2 4 5 10 10 30 30 30 30 10 10 5 4 2 4 5 10 10 30 30 30 30 10 10 5 4 2 4 5 10 10 30 30
30 30 10 10 5 4 2 4 5 10 10 30 30 150 150 350 500 500 1000 1000 1000 1000 1000 1000 1000
COM
COM ---- ARRAY OF VALUES SPECIFYING LENGTHS OF GRID BLOCKS IN Y DIRECTION - (Ny values ) ----
COM <---------------------------- m/ft --------------------------------------->
COM
1000 1000 500 200 100 60 22 22 18 12 10 10 10 8 5 4 4 4 4 4 2 2 2 4 4 4 4 4 5 8 10 10 10 12 18 22 22 60 100
200 500 1000 1000
COM
COM ---- ARRAY OF VALUES - LENGTHS OF GRID IN Z DIRECTION - (Nz values, from top down ) ----
COM <---------------------------- m/ft --------------------------------------->
COM
1000 1000 1000 500 500 300 200 200 100 100 40 30 30 14 10 10 5 4 2 2 2 2 2 2 2 4 5 10 10 14 30 30 40 100 100
200 200 300 500 500
COM
COM *** GROUP I - STRESS INITIALIZATION ***
COM
COM RECORD I0
COM
COM ,--FEM_ref_Depth: REFERENCE DEPTH FOR STRESS INITIALIZATION
189
COM | (z is negative downward)
COM | ,--Mode_instr: MODE OF STRESS INITIALIZATION
COM | | 0 - read stresses at FEM_ref_depth + gradients
COM | | 1 - read stresses by layers
COM | | 2 - initialization by variable gradient (using rock, fluid
COM V | saturations and K0 stress ratio) (project specific)
COM m/ft V
COM -----|----*
5300 0
COM -----|----*
COM
COM **** UNIFORM STRESS INITIALIZATION ****
COM Read this data only if Mode_instr = 0 was specified
COM
COM RECORD I1
COM
COM ,--FEM_Ref_Sx: INITIAL TOTAL STRESS IN X-DIR (compression= +)
COM | ,--FEM_Ref_Sy: INITIAL TOTAL STRESS IN Y-DIR
COM | | ,--FEM_Ref_Sz: INITIAL TOTAL STRESS IN Z-DIR
COM | | |
COM V V V
COM ----- kPa / psia ----
COM -----|---------|---------|
4505 5989 6572
COM -----|---------|---------|
COM
COM RECORD I2
COM
COM ,--FEM_Grad_Sx:INITIAL TOTAL X-STRESS GRADIENT WITH DEPTH (increase down=+)
COM | ,--FEM_Grad_Sy: INITIAL TOTAL Y-STRESS GRDIENT WITH DEPTH
COM | | ,--FEM_Grad_Sz:INITIAL TOTAL Z-STRESS GRADIENT WITH DEPTH
COM | | |
COM V V V
COM --- kPa/m / psia/ft ---
COM -----|---------|---------|
0.0 0.0 0.0
COM -----|---------|---------|
COM
COM RECORD F
COM
COM *** STATIC/DYNAMIC LOAD/B.C. EVENT CONTROL ***
COM
COM (Notice: have at least one load event in order to read in fixity B.C.)
COM
COM *** STATIC AND/OR DYNAMIC LOADS ***
COM Read the number of static and transient load events
COM ,--num_static: Number of Quasi-Static (initialization) Load Events
COM | ,--Indic_transient: Indicator for transient solution:
COM | | 0=initialization only (do not read transient loads)
COM | | 1=do transient stress solve, read transient loads
COM V V
COM *----*
0 1
COM *----*
COM *** GROUP F2 - TRANSIENT (DYNAMIC SIMULATION) LOAD EVENT DATA ***
COM (Read if Indic_transient >0)
COM
COM RECORD F2 - Transient Load Event Data
COM
COM ---- ZDispfl option to zero the displacement array at the beginning ----
190
COM | of the transient coupled simulation (1=zero , 0=retain)
COM | ,--Num_Disp_BC: number of generalized displacement BC sets
COM | | ,--Num_Face_Loads: number of face regions with normal forces
COM | | | ,N_Tab_Loads: number of transient load tables
COM | | | | ,N_frac_loads: Number of dynamic/static frac loaded regions
COM | | | | | (+ dynamic loading; - static loading)
COM | | | | | | ,--Special fixity key (0: no additional fixities; 1: need file xx.fix
COM | | | | | | | (form: total # of fixed nodes, I,J,K,Fi,Fy,Fz)
COM V V V V V V V
COM *----*----*----*----*----*----*
1 0 0 0 0 0 0
COM *----*----*----*----*----*----*
COM
COM RECORD F2A
COM
COM DISPLACEMENT FIXITY KEYS FOR SIDES (FACES) OF THE MODEL :
COM
COM INPUT 6 LINES FOR SIDES I=1,..,6 IN THE FOLLOWING ORDER:
COM I=1 - TOP (z = Lz)
COM I=2 - BOTTOM (z = 0)
COM I=3 - LEFT (y = 0)
COM I=4 - RIGHT (y = Ly)
COM I=5 - FRONT (x = Lx)
COM I=6 - BACK (x = 0)
COM
COM ,--Anl_fix_key(1,I): x-direction displacement on face I:0=free,1=fixed
COM | ,--Anl_fix_key(2,I): y-direction displacement: 0=free, 1=fixed
COM | | ,--Anl_fix_key(3,I): z-direction displacement: 0=free, 1=fixed
COM | | |
COM V V V
COM *----*----*
0 0 0
0 0 1
0 1 0
0 1 0
1 0 0
1 0 0
COM *----*----*
COM
COM *** GROUP M - MATERIAL PROPERTIES (CONSTITUTIVE MODELS) ***
COM
COM **** MATERIAL PROPERTIES ****
COM Read as many sets of data as there is material types,total of Mat_cnt
COM Each set consists of record M0 plus one of the Records M1,M2 or M3
COM
COM RECORD M0
COM
COM ,--Mat_type: type of material (1=linearly elastic, 2=hyperbolic, 3=tables,
COM | 4=hypo-plastic, 6=Barton-Bandis, 11=lin.el./EP
COM | ,--KEY1 (=ISIGE for Mat_type=2,
COM | | =independent variable for tables for Mat_type=3:
COM | | 0= min eff stress
COM | | 1= mean eff stress
COM | | 2= depth
COM | | ,--KEY2 (=ISIGB for Mat_type=2)
COM | | | ,--KEY3 (=IDILATE for Mat_type=2, OPTION FOR MODELING DILATION:
COM | | | | 0= NO DILATION
COM | | | | 1= SHEAR-INDUCED DILATION
COM | | | | 2= STRESS-RATIO-INDUCED DILATION
COM V V V V
191
COM *----*----*----*
1 0 0 0
COM *----*----*----*
COM
COM RECORD M1
COM
COM Input for Mat_type = 1 - linearly elastic material
COM --------------------------------------------------
COM
COM ,--Mat_E: YOUNG'S MODULUS
COM | ,--Mat_nu: POISSON RATIO
COM | | ,--Mat_cte: COEFF. OF THERMAL EXPANSION (LINEAR)
COM | | | ,--Mat_ke: MATRIX (GRAIN) MODULUS (1/compress.)
COM | | | | ,--Mat_Density: DENSITY (mass/volume)
COM | | | | |
COM | | | | |
COM V V V V V
COM kPa - 1/deg C kPa kg/m3
COM psia - 1/deg F psia lb/cuft
COM -----|---------|---------|---------|---------|
7.9E6 0.125 0.0 1.0E12 160
COM -----|---------|---------|---------|---------|
COM
COM *** GROUP S - SOLUTION CONTROLS ***
COM
COM RECORD S1
COM
COM **** FEM Solution Strategy ****
COM
COM a) If all Material types are elastic (Material type 1,2,3,4 or 6)
COM read the following data:
COM
COM In this case iterations can be used within time step to
COM solve problems with dilation
COM
COM ,--Solution_Type: linear/nonlinear elastic solution strategy
COM | Set to 1 - Constant K matrix during time step
COM | ,--Max_Iterations:
COM | | 0 = no iteration during time/load step (for linear and
COM | | nonlinear elasticity without dilation)
COM | | n > 0 = use max of n iterations during solution to
COM | | eliminate out of balance force caused by dilation
COM v v
COM *----*
1 0
COM *----*
COM RECORD S2
COM
COM ITERATIVE SOLVER PARAMETERS (Read only if Iter_flg = 1)
COM
COM a) Iterative solver specification flags (options)
COM
COM ,--Iter_iadd: SPD flag - To add amount to the main diagonal to force
COM | positive definitness (if problem with diag. dominance)
COM | 0 = no, >0 = number of adds
COM | ,--Iter_idrop: ILU flag - Type of ILU decomposition
COM | | 0 = level-based ILU (fixed fill pattern, no dropping)
COM | | 1 = drop tolerance ILU (drop terms based on the orig.
COM | | diagonal test
COM | | 2 = drop tolerance ILU, updated diag. test (preferred)
192
COM | | ,--Iter_level: Number of extra diagonals used for ILU
COM | | | (applies only for fixed fill option, Iter_drop = 0)
COM | | |
COM V V V
COM *----|----|
0 0 0
COM *----|----|
COM
COM b) Iterative solver specification data
COM
COM ,--Riter_dgadd: Amount added to the main diagonal
COM | (applies only if SPD flag > 0,i.e., Iter_idgadd >0 )
COM | Value must be > 0 if applicable.
COM | ,--Riter_dropt: Drop tolerance value for Drop ILU options
COM | | (applies only if ILU flag = 1 or 2, i.e., Iter_idrop=1,2)
COM | | Value must be > 0 if applicable.
COM | | ,--Riter_dgtol: Diagonal tolerance for checking if matrix
COM | | | is diag. dominant (enter 0.0)
COM | | | ,--Riter_ctol: Absolute convergence tolerance
COM | | | | (for each component of the solution)
COM | | | | ,--Riter_reduce: Residual norm
COM | | | | | reduction factor from the initial
COM | | | | | value
COM | | | | |
COM | | | | | NOTE: Soln is converged when the
COM | | | | | FIRST tolerance is satisfied
COM V V V V V
COM *----|---------|---------|---------|---------|
0 0 0 1.0d-20 1.0d-20
COM *----|---------|---------|---------|---------|
COM
END
EOF
Case – 6.6.gii : The Input Interface

COM **************************************************
COM * Taurus Reservoir Solutions Ltd. *
COM * GEOINT *
COM * INTERFACE BETWEEN RES AND STRESS MODEL *
COM * VERSION 3.14 *
COM * 26-APRIL-2007 *
COM **************************************************
COM RECORD A
COM
COM ,--IUNIT,INPUT UNITS : 0 = metric (kPA, m, deg C)
COM | 1 = English (psia, ft, deg F)
COM | ,--ASIZE, size of memory for arrays (integer words)
COM | | ,--ASOLVERSIZE, size of solver memory (integer words)
COM | | |
COM V V V
COM *---------*---------*
1 50e9 30e9
COM *---------*---------*
COM RECORD B
COM
COM **** NUMERICAL CONTROLS OF THE COUPLING ****
COM
COM ,--Ptol, PRESSURE DP TOLERANCE FOR CONVERGENCE
COM | (global iter between reserv. and stress model)
193
COM | ,--Accmax, MAX. ACCELERATION FACTOR ( > 1)
COM | | ,--N_noacc, NO OF T.S. WITHOUT
COM | | | ACCELERATION AFTER A RATE
COM | | | CHANGE IN THE HOST MODEL
COM | | | ,--N_solve, frequency of stress
COM | | | | soln in EXPLICIT mode:
COM | | | | 0= every t.s.
COM | | | | n= every nth t.s.
COM | | | | -1= every TCHG
COM | | | | -n= every nth t.s.+TCHG
COM | | | | ,--Key_resm, no of
COM | | | | | regions for res.perm.
COM | | | | | modifiers >=0
COM | | | | | ,--Key_resm_p, no of
COM | | | | | | regions for res.poros.
COM | | | | | | modifiers >=0
COM | | | | | | ,--Igappr, geom. iter. method (1,2,3) ...
COM | | | | | | | 0,1 = standard (Settari+Mourits)
COM | | | | | | | 2 = modified standard
COM | | | | | | | 3 D.Tran
COM V | | | | | | ,--Factor1, constr.type constant
COM kPa | | | | | | | for igappr=3
COM psia V V V V V V V
COM -----|---------|----*----*----*----*----*---------|
1.0 1.0 10 0 0 0 0 0 .0
COM -----|---------|----*----*----*----*----*---------|
COM
COM RECORD B1
COM
COM **** INITIALIZATION DATA - OUTSIDE OF RESERVOIR GRID, NULL CELLS ****
COM **** NORMALIZING STRESS TABLES ****
COM
COM ,--Press_ref, REFERENCE PRESSURE at depth=0 (used to set initial pressure
COM | in null cells of reservoir grid)
COM | ,--Temp_ref, REFERENCE TEMPERATURE (used for thermal
COM | | compaction in FEM3D)
COM | | ,--wGrad, PORE FLUID GRADIENT to set pressure in null cells
COM | | | p(m) = depth(m)*wGrad
COM | | | ,--Norm_perm, if=1 normalize stress dependent
COM | | | | permeability in Record J to initial stress
COM | | | | ,--Norm_por, if=1 normalize stress
COM V V V | | dependent porosity in Record K to
COM kpa deg C kPa/m | | initial stress (not yet implemented)
COM psia deg F psia/ft V V
COM -----|---------|---------|----*----*
0.0 0.0 0.0
COM -----|---------|---------|----*----*
COM
COM RECORD C
COM
COM ***** PRINTOUT OPTIONS AND CONTROLS *****
COM
COM ,--IOUTPR: type of printout of FEM solution:
COM | 0 - use FEM print routines (in element order)
COM | 1 - print 3-D arrays of variables consistent with host grid (read 16 keys of stresses, displacements...)
COM | 2 - Same as IOUTPR=1, and read additional keys to print mechanical properties
COM |
COM | ,--IOUTCA: controls postprocessing files for TERAPRO/SEISMIC:
COM | | 0 - no files
COM | | 1 - write TERAPRO data to .gc* with freq. of arrays
194
COM | | +-2 - write SEISMIC data to .seis with freq. of arrays
COM | | +-3 - write BOTH with freq. of arrays
COM | | +2 or +3 means write reservoir arrays into .seis over FEM grid
COM | | -2 or -3 means write reservoir arrays into .seis only over res grid
COM | | ,--idisp_avg: Method for averaging z-displacement for TERAPRO:
COM | | | 0 - compute block center values
COM | | | 1 - compute top of block values
COM | | | ,--Key_var_seis: control on content of SEIS file
COM | | | | 0 - Heriot Watt set of arrays (Appendix E)
COM | | | | 1 - as 0 plus Sz and strainz (V. Sen)
COM | | | | 2 - as 1 plus vert.displ (averaged according to idisp_avg)
COM | | | | 3 - as 2 but suppress writing reservoir arrays
COM | | | |
COM V V V V
COM *----*----*----*
1 1 1 0
COM *----*----*----*
COM
COM Read these 2 print controls (Records D and E) ONLY if IOUTPR > 0
COM
COM RECORD D
COM
COM ,--IPLNPR: choice of 2-D planes to print 3-D arrays:
COM | 0 - x-y (areal), 1 - x-z (crossections)
COM V
COM *----*
0 0
COM *----*
COM
COM RECORD E
COM
COM KEYPRT(1-16) keys to print arrays:
COM 0=no, 1=yes (total stresses), 2=yes (eff. stresses)
COM
COM ,--print stress Sx
COM | ,--print stress Sy
COM | | ,--print stress Sz
COM | | | ,--print shear stress Txy
COM | | | | ,--print shear stress Txz
COM | | | | | ,--print shear stress Tyz
COM | | | | | | ,--principal max. stress Smax
COM | | | | | | | ,--principal min. stress Smin
COM | | | | | | | | ,--Displacement Ux
COM | | | | | | | | | ,--Displacement Uy
COM | | | | | | | | | | ,--Displacement Uz
COM | | | | | | | | | | | ,--Hysteresis keys
COM | | | | | | | | | | | | ,--Pressure
COM | | | | | | | | | | | | | ,--Stress level
COM | | | | | | | | | | | | | | ,--Volume.
COM | | | | | | | | | | | | | | | strain
COM | | | | | | | | | | | | | | | Savg
COM V V V V V V V V V V V V V V V V
COM *----*----*----*----*----*----*----*----*----*----*----*----*----*----*----*
1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1
COM *----*----*----*----*----*----*----*----*----*----*----*----*----*----*----*
COM RECORD F
COM
COM **** MONITORING AND DEBUG PRINTOUT - Use with caution !!! ****
COM
COM ,--Monitor, PRINT CONTROL:
195
COM | 0 = NO EXTRA PRINTOUT
COM | 1 = SOME DEBUG FOR FIRST 6 NODES
COM | 2 = 1 + POROSITIES
COM | 3 = 2 + STRESSES AND STRAINS AFTER EACH ITER.
COM | ,--Mode_prt, ARRAY PRINT CONTROL MODE:
COM | | 0 = CONTROLLED BY HOST MODEL ARRAY FREQUENCY
COM | | 1 = INDEPENDENT, BASED ON IPRTSTR FREQUENCY
COM | | 2 = PRINT BOTH AT HOST FREQUENCY AND IPRTSTR FREQUENCY
COM | | ,--IPRTSTR, FREQUENCY OF PRINTING STRAIN/STRESS ARRAYS
COM | | | (used if Mode_prt = 1 or 2)
COM | | | ,--Monit_sol, KEY TO MONITOR PERFORMANCE OF FEM SOLVERS:
COM | | | | 0 = no print
COM | | | | 1 = print timings and other info
COM | | | |
COM V V V V
COM *----*----*----*
0 0 0 0
COM *----*----*----*
COM RECORD G
COM
COM **** PRINT CONTROLS FOR MONITORING STRESSES ALONG A WELLBORE ****
COM
COM ,--Imonit_str: output level:
COM | 0 - none, 1 - t.s.summary, 2 - as 1 + stresses+p
COM | ,--IPLT_STR: postprocessing plot/link files
COM | | 0=no
COM | | 1=write .trs (p,T+stresses) and .trd (displacements) files
COM | | for Mmonit_str elements (blocks) specified in Record I
COM | | (external file names xxxx.trsn and xxxx.trdn where
COM | | n is the number of the block in the input order)
COM | | 2=as =1+write fort57 with link data for stand alone FEM3D)
COM | | ,--Mmonit_str: number of blocks for which the postprocessing
COM | | | files .trs and .trp are written (2 files/block, max=10)
COM | | | ,--IPRT_RESET: reset print controls after fracture
COM | | | | re-initiation: 0 = no, 1 = print arrays
COM | | | |
COM V V V V
COM *----*----*----*
0 1 0 0
COM *----*----*----*
COM RECORD H
COM
COM **** REGION TO SEARCH FOR MINIMUM STRESS ****
COM (should define a line within the grid)
COM
COM ,--Im1 - first I-index (in FEM grid)
COM | ,--Im2 - last I-index (in res. indexing over FEM grid)
COM | | ,--Jm1
COM | | | ,--Jm2
COM | | | | ,--Km1
COM | | | | | ,--Km2
COM | | | | | |
COM | | | | | |
COM V V V V V V
COM *----*----*----*----*----*
1 110 1 43 1 40
COM *----*----*----*----*----*
COM
END
EOF
196

Ucalgary 2012 Islam Arshad

Uploaded by

Document Information

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

Ucalgary 2012 Islam Arshad

Uploaded by

Copyright:

Available Formats

University of Calgary

PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

Geomechanical Productivity and Injectivity Modeling

Islam, A. (2012). Geomechanical Productivity and Injectivity Modeling of Multifractured Horizontal

Geomechanical Productivity and Injectivity Modeling of Multifractured Horizontal Wells

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF MASTER OF SCIENCE

DEPARTMENT OF CHEMICAL AND PETROLEUM ENGINEERING

© Arshad Islam 2012

productivity and EUR of multifractured horizontal wells, it is not clear if conventional

included to obtain realistic solutions.

give incorrect results in terms of production.

A new method based on uniaxial deformation theory is proposed to more accurately

effects were remarkably similar to rigorously coupled geomechanical modeling. This

work will be of importance to engineers in analyzing and forecasting production

performance of multifractured horizontal completions using numerical models. It will

allow engineers to use uncoupled (conventional) reservoir modeling as a practical

approximation of more complex coupled geomechanical models.

showed the importance of including the geomechanics. Sensitivity study of several

history matching parameters such as fracture permeability reduction factor, limiting

techniques for approximating these effects in uncoupled reservoir simulation. In addition,

supported me to complete this thesis. My deepest appreciation goes to my supervisor, Dr.

MSc. Program. I am grateful to management of Apache Petroleum Corporation for

software and University of Calgary, the Department of Chemical and Petroleum

stay memorable and enjoyable.

curvy road of life and my passionate loving Mother.

To my Father, Muhammad Islam and my Mother, Rabia Islam,

for their years of love, affection, support and encouragement.

CHAPTER ONE: INTRODUCTION ..................................................................................1

CHAPTER TWO: LITERATURE REVIEW ......................................................................8

CHAPTER FOUR: MODELING OF STATIC FRACTURES .........................................45

CHAPTER FIVE: HISTORY MATCHING USING PSEUDO CONTINUUM

CHAPTER SIX: HISTORY MATCHING AND MODELING OF

CHAPTER SEVEN: HISTORY MATCHING OF FIELD INJECTION PRESSURE

CHAPTER EIGHT: RESULTS AND CONCLUSIONS ................................................166

APPENDIX A: SELECTED DATA FILES ....................................................................181

Table 3.2 Proppant types in well A and B ........................................................................ 39

Table 4.1 Baseline permeability of sand used in well A and B ........................................ 54

Table 5.1 Fracture properties of Case 5.1.1 – 2 for well A............................................... 63

Table 5.2 Fracture properties of Case 5.1.1 – 2 for well B ............................................... 63

Table 6.4 Fracture properties of Case 6.3.2 for well A..................................................... 96

Table 6.5 Fracture properties of Case 6.3.2 for well B ..................................................... 97

Figure 1.1 Natural gas resource triangle (from Holditch, 2006)......................................... 2

Figure 3.5 Bottomhole field injection pressure - well A and B ........................................ 38

Table 4.2 Calculated fracture properties of sands in well A ............................................. 56

Table 4.3 Calculated fracture properties of sands in well B ............................................. 57

Figure 5.1 Flow chart of simulation cases – Chapter 5..................................................... 61

Figure 5.7 Stimulated reservoir volume (SRV) schematic – Complex fracture

Figure 5.13 Time dependent permeability multipliers-History matched case–Well A .... 82

Figure 6.1 Flow chart of simulation cases – Chapter 6..................................................... 87

Figure 7.1 Flow chart of simulation cases – Chapter 7................................................... 128

cross sectional area

API American Petroleum Institute

fracture area in y-direction

matrix area in y-direction

half aperture distance (fracture half width)

constant defining conductivity / porosity

BHIP bottomhole injection pressure

BHP bottomhole flowing pressure

formation volume factor of phase l

fracture compressibility at zero stress