Decline Curve Enhanced Shale Oil Recovery Based On Data: Analysis and Eagle Ford Shale

DECLINE CURVE ANALYSIS AND ENHANCED SHALE OIL RECOVERY BASED ON
EAGLE FORD SHALE DATA
By
Dieudonne K. Delaihdem
RECOMMENDED:
_..,Dr. Abhijit Dandekar
~-
Dr. Abhijit Dandekar
Chair, Department of Petroleum Engineering
APPROVED:
DECLINE CURVE ANALYSIS AND ENHANCED SHALE OIL RECOVERY BASED ON
EAGLE FORD SHALE DATA
THESIS
Presented to the Faculty
of the University of Alaska Fairbanks
in Partial Fulfillment of the Requirements
for the Degree of
MASTER OF SCIENCE
By
Dieudonne K. Delaihdem, B.S.
Fairbanks, Alaska
December 2013
ABSTRACT
Transient and fracture dominated flow regimes in tight permeability shale reservoirs with hydraulically
fractured horizontal wells impose many unconventional challenges. These include execution of
appropriate shale decline curve analysis and the optimization of hydrocarbons recovery. Additionally,
short production profiles available are inadequate for accurate production decline analysis.
This research assessed the effectiveness of Arps’ decline curve analysis and recently established methods-
-power law exponential analysis, logistic growth analysis, Duong’s method and the author’s approach--to
predict future production of horizontal wells in the Eagle Ford Shale. Simulation models investigated
history matching, enhanced shale oil recovery, and drainage area beyond stimulated reservoir volume.
Traditional Arps’ hyperbolic method sufficiently analyzed past production rates, but inaccurately
forecasted cumulative productions. The recent decline models show slight variations in their past
performance evaluations and forecasting future production trends. The technique proposed and used in
this work enhanced the successful application of Arps’ hyperbolic decline from 32.5% to 80%.
Simulation results indicate 4.0% primary oil recovery factor and 5.8% enhanced shale oil recovery factor
using CO2 miscible injection. Based on pressure observed outside of the stimulated reservoir volume,
limited to the range of data used in this study, drainage area outside stimulated reservoir volume is not
significant.
vii
DEDICATION
To God is the Glory

Praises to the Lord; His Love Endures Forever
ix
TABLE OF CONTENTS
Signature Page……………………………………………………………………………………..i
Title Page …………………………………………………………………………………………iii
ABSTRACT..................................................................................................................................... v
DEDICATION ............................................................................................................................... vii
TABLE OF CONTENTS................................................................................................................ ix
LIST OF FIGURES ...................................................................................................................... xiii
LIST OF TABLES ....................................................................................................................... xxii
ACKNOWLEDGEMENT .......................................................................................................... xxiv
Chapter 1 Introduction ..................................................................................................................... 1
1.1 Shale Oil Reservoirs ............................................................................................................ 1
1.2 Shale Oil- What is it? ........................................................................................................... 2
1.3 Shale Play Development ...................................................................................................... 3
1.4 Purpose and Scope of Study ................................................................................................ 4
Chapter 2 Literature Review ............................................................................................................ 6
2.0 Introduction .......................................................................................................................... 6
2.1 Geology ................................................................................................................................ 7
2.2 Decline Curve Analysis ..................................................................................................... 11
2.2.1 Arps Decline Curve Analysis ..................................................................................... 12
2.2.2 Power Law Exponential Analysis .............................................................................. 15
2.2.3 Logistic Growth Analysis .......................................................................................... 16
2.2.4 Duong’s Method ........................................................................................................ 17
2.3 Analyses of the Eagle Ford Shale ...................................................................................... 19
2.3.1 Reservoir Fluids ......................................................................................................... 19

x
2.3.2 Natural Fractures ........................................................................................................ 20
2.3.3 Reservoir Stimulation and Wellbore Completions .................................................... 22
2.3.4 Simulation .................................................................................................................. 25
2.4 Improving Shale Oil Recovery by CO2 Injection and CO2 as Frac Fluid .......................... 29
Chapter 3 Eagle Ford Production History Assessment Using Various Decline Curve Analysis
Techniques ..................................................................................................................................... 36
3.0 Introduction ........................................................................................................................ 36
3.1 Methods.............................................................................................................................. 37
3.1.1 Data Preparation before the Analysis......................................................................... 37
3.1.2 Arps’ Decline Models Applications ........................................................................... 37
3.1.3 Power Law Exponential (PLE) Analysis Application................................................ 38
3.1.4 Logistic Growth Analysis (LGA) Application ........................................................... 39
3.1.5 Duong’s Decline Curve Analysis Method ................................................................. 40
3.1.6 Logistic Arps’ Hyperbolic Approach (LAHA) .......................................................... 41
3.2 General Results .................................................................................................................. 42
3.2.1 Data Preparation before the Analysis ......................................................................... 42
3.2.2 Arps’ Decline Curve Analysis ................................................................................... 45
3.2.3 Power Law Exponential Analysis (PLE) ................................................................... 48
3.2.4 Logistic Growth Analysis (LGA) .............................................................................. 51
3.2.5 Duong’s Decline Curve Analysis Method ................................................................. 52
3.2.6 Logistic Arps Hyperbolic Approach (LAHA) ........................................................... 57
3.3 Discussion of General Results ........................................................................................... 60
Chapter 4 Eagle Ford Shale Oil Production Decline Analysis Specific Results ............................ 63
4.0 Introduction ........................................................................................................................ 63

xi
4.1 Burleson County ................................................................................................................ 63
4.1.1 Introduction ................................................................................................................ 63
4.1.2 Specific Results of Shale Oil Production Decline Analysis in Burleson County ...... 65
4.2 Leon County....................................................................................................................... 70
4.2.1 Introduction ................................................................................................................ 70
4.2.2 Specific Results of Shale Oil Production Decline Analysis in Leon County............. 71
4.3 Brazos County.................................................................................................................... 76
4.3.1 Introduction ................................................................................................................ 76
4.3.2 Specific Results of Shale Oil Production Decline Analysis in Brazos County .......... 77
4.4 Lee County ......................................................................................................................... 81
4.4.1 Introduction ................................................................................................................ 81
4.4.2 Specific Results of Shale Oil Production Decline Analysis in Lee County ............... 82
4.5 Gonzales County ................................................................................................................ 84
4.5.1 Introduction ................................................................................................................ 84
4.5.2 Specific Results of Shale Oil Production Decline Analysis in Gonzales County ...... 86
4.6 Karnes County ................................................................................................................... 91
4.6.1 Introduction ................................................................................................................ 91
4.6.2 Specific Results of Shale Oil Production Decline Analysis in Karnes County.......... 92
4.7 Dimmit County .................................................................................................................. 98
4.7.1 Introduction ................................................................................................................ 98
4.7.2 Specific Results of Shale Oil Production Decline Analysis in Dimmit County ........ 98
4.8 Zavala County .................................................................................................................. 102
4.8.1 Introduction .............................................................................................................. 102
4.8.2 Specific Results of Shale Oil Production Decline Analysis in Zavala County ........ 103
xii
4.9 Chapter Summary ............................................................................................................ 108
Chapter 5 Shale Oil Reservoir Simulations ................................................................................. 109
5.0 Introduction ...................................................................................................................... 109
5.1 Methods: Production Decline and Future Performance Simulations ............................... 110
5.1.1 Building the Reservoir Simulation Models .............................................................. 110
5.1.2 The Reservoir Simulation Initialization Outputs ..................................................... 113
5.2 Results: Production Decline and Future Performance Simulation ................................... 126
5.3 Methods: Enhanced Shale Oil Recovery Simulation ....................................................... 133
5.3.1 Results: Enhanced Shale Oil Recovery Simulation ................................................. 137
5.4 Pressure Profiles across Lease Boundaries ...................................................................... 146
5.4.1 Methods.................................................................................................................... 147
5.4.2 Results: Pressure Profiles across Lease Boundaries ............................................... 149
Chapter 6 Discussion ................................................................................................................... 151
6.1 Decline Curve Analysis of the Eagle Ford Shale Oil Production .................................... 151
6.2 Production Decline and Forecast, ESOR and Pressure Profiles Simulation .................... 154
6.2.1 Production Decline and Forecast Simulation ........................................................... 154
6.2.2 Enhanced Shale Oil Recovery (ESOR) Simulation ................................................. 154
6.2.3 Pressure Profile Simulation ...................................................................................... 156
Chapter 7 Conclusions and Recommendations ............................................................................ 157
7.1 Conclusions ...................................................................................................................... 157
7.2 Recommendations ............................................................................................................ 159
REFERENCES ............................................................................................................................ 161
NOMENCLATURE .................................................................................................................... 165
Appendix ...................................................................................................................................... 168

xiii
LIST OF FIGURES
Figure 2.1: Map view of southern Texas showing Eagle Ford Shale resource play, counties,
scheduled oil and gas and permitted wells locations (TRRC, 2012). ...................................... 8
Figure 2.2: Map view of the Eagle Ford Shale hydrocarbon phases and windows of thermal
maturity with arrow labels indicating the researched counties’ locations (Wang and Liu,
2011). ....................................................................................................................................... 9
Figure 2.3: Cretaceous stratigraphic column from south Texas showing stratigraphic location of
the Eagle Ford shale (Fan et al., 2011). ................................................................................. 10
Figure 2.4: Micro-fractures in Eagle Ford Shale cores (Stegent et al., 2010). ............................... 21
Figure 3.1(A, B): Rate and semi-log rate versus time of the Simms lease #03914 production
profile, in which the use of the blue marked trend is justified compared to the erratic (red
marked) sections. ................................................................................................................... 43
Figure 3.2 (A, B): Rate versus time plots for Cannon #09607 lease and CEF #09608, presented as
examples for erratic production profiles. ............................................................................... 44
Figure 3.3 (A, B): (A) Arps’ hyperbolic decline analysis for the erratic production profiles of the
CEF #09608 and Cannon #09607 lease wellbores resulted in abnormally high Arps’ decline
exponent; i.e., b=3.05 for (A) and b=1.84 (B) respectively. .................................................. 44
Figure 3.4: Four diagnostic plots used to determine Arps’ exponential or harmonic decline. ...... 46
Figure 3.5: Arps’ hyperbolic analysis showing cumulative production and rate match between the
actual and the modeled data. .................................................................................................. 47

xiv
Figure 3.6: Percentage application of Arps’ decline curve analyses for Forty Profiles in the Eagle
Ford. ....................................................................................................................................... 47
Figure 3.7: The percentage of abnormal decline exponents is more than twice that of the normal
decline exponents obtained from Arps’ hyperbolic model. ................................................... 48
Figure 3.8: PLE diagnostic plot for the Simms lease achieved R2≥0.97 only when initial and
erratic data sets were filtered out. .......................................................................................... 49
Figure 3.9: PLE diagnostic plot for the Easterling well achieved R2≥0.97 only when erratic data
sets and no production sections were filtered out. ................................................................. 49
Figure 3.10: Power law exponential analysis showing cumulative production and rate match
between the actual and the modeled data for Giesenschlag W.H. well in Burleson County. 50
Figure 3.11: Power law exponential analysis showing not very good cumulative production and
rate match between the actual and the modeled data for Hullabaloo well in Brazos County.50
Figure 3.12: Logistic growth analysis showing cumulative production and rate match between the
actual and the modeled data for Giesenschlag W.H. well in Burleson County. .................... 51
Figure 3.13 (A, B): Duong’s method diagnostic plots for a production profile to determine a, m,
q1 and q∞. .............................................................................................................................. 52
Figure 3.14 (A, B): Different results are obtained in fitting past cumulative production profile of a
well (Hullabaloo) with Duong’s method when the rate versus time function does or does not
pass through the origin. .......................................................................................................... 53

xv
Figure 3.15 (A, B): Different results are obtained in fitting past cumulative production profile of a
well (Donaho Unit) with Duong’s method when rate versus the time function does or does
not go through the origin........................................................................................................ 54
Figure 3.16 (A, B): No significant differences are obtained in fitting past cumulative production
profile of a well (Giesenchlag W.H.C) with Duong’s method when rate versus the time
function does or does not pass through the origin. ................................................................ 54
Figure 3.17 (A-C): Duong’s method failed to match the production history for unfiltered Cannon
# 09607 lease.......................................................................................................................... 56
Figure 3.18: Duong’s method showing cumulative production and rate match between the actual
and the modeled data for Giesenschlag W.H. well in Burleson County. ............................... 57
Figure 3.19 (A, B): Percentage of normal decline exponent (0<b<1) improves with the use of the
new approach (LAHA). ......................................................................................................... 58
Figure 3.20: Comparison of decline exponents (b) using Arps' hyperbolic and LAHA. ............... 59
Figure 3.21: Expected ultimate recoveries (EUR) generated by LAHA are expectedly lower than
that of Arps’ hyperbolic decline for cases where b>1 under Arps’ hyperbolic method became
0<b<1 under LAHA. .............................................................................................................. 60
Figure 4.1: Profile of Burleson County wells’ production history................................................. 66
Figure 4.2: EUR generated for three horizontal wells' production profiles Burleson County. ...... 67
Figure 4.3: Remaining reserves forecasted to 2 barrels/day for three Burleson county horizontal
wells' production profiles. ...................................................................................................... 68

xvi
Figure 4.4: Remaining production time to reach 2 barrels/day for three Burleson county
horizontal wells' production profiles. ..................................................................................... 69
Figure 4.5: Profile of Leon County wells’ production history. ...................................................... 72
Figure 4.6: EUR generated for three horizontal wells' production profiles in Leon County. ........ 74
Figure 4.7: Remaining reserve forecasts for three production profiles in Leon County. ............... 75
Figure 4.8: Remaining production life to reach 2 barrels/day for 3 wells’ in Leon County. ......... 76
Figure 4.9: Profile of Brazos County wells’ production history. ................................................... 78
Figure 4.10: EUR generated for two horizontal wells’ production profiles in Brazos County. ..... 79
Figure 4.11: Remaining reserve forecasts for two production profiles in Brazos County. ............ 80
Figure 4.12: Remaining production life to reach 2 barrels/day for 2 wells in Brazos County....... 81
Figure 4.13: EUR generated for one horizontal well’s production profile in Lee County............. 83
Figure 4.14: Remaining reserve forecasts for one production profile in Lee County. ................... 83
Figure 4.15: Remaining production life to reach 2 barrels/day for one well in Lee County. ........ 84
Figure 4.16: Profile of Gonzales County wells’ production history. ............................................. 86
Figure 4.17: EUR generated for eleven horizontal wells’ production profiles in Gonzales County.
............................................................................................................................................... 88
Figure 4.18: Remaining reserve forecasts for eleven production profiles in Gonzales County. .... 89
xvii
Figure 4.19: Remaining production life to reach 2 barrels/day for 11 wells in Gonzales County. 90
Figure 4.20: Profile of Karnes County wells’ production history. ................................................. 93
Figure 4.21: EUR generated for ten horizontal wells’ production profile in Karnes County. ....... 95
Figure 4.22: Remaining reserve forecasts for ten production profiles in Karnes County. ............. 96
Figure 4.23: Remaining production life to reach 2 barrels/day for ten wells in Karnes County. .. 97
Figure 4.24: Profile of Dimmit County wells’ production history................................................. 99
Figure 4.25: EUR generated for four horizontal wells’ production profiles in Dimmit County. . 100
Figure 4.26: Remaining reserve forecasts for four production profiles in Dimmit County. ........ 101
Figure 4.27: Remaining production life to reach 2 barrels/day for four wells in Dimmit County.
............................................................................................................................................. 102
Figure 4.28: Profile of Zavala County wells’ production history. ............................................... 104
Figure 4.29: EUR generated for six horizontal wells’ production profiles in Zavala County. .... 105
Figure 4.30: Remaining reserve forecasts for six production profiles in Zavala County. ........... 106
Figure 4.31: Remaining production life to reach 2 barrels/day for six wells in Zavala County. . 107
Figure 5.1: Two-Phase P-T envelope of the synthetic Eagle Ford Shale oil. .............................. 114
Figure 5.2: Liquid phase volume fraction versus pressure. ......................................................... 115
Figure 5.3: Vapor phase volume fraction versus pressure. .......................................................... 116
xviii
Figure 5.4:Vapor Z-Factor versus pressure.................................................................................. 116
Figure 5.5: Equilibrium ratio (K Value=vapor/liquid) versus pressure. ...................................... 117
Figure 5.6: Eagle Ford Shale synthetic crude solution gas-oil ratio with pressure. ..................... 118
Figure 5.7: Eagle Ford synthetic crude oil formation volume factor with pressure. .................... 118
Figure 5.8: Eagle Ford synthetic crude oil viscosity versus pressure at reservoir temperature. .. 119
Figure 5.9: Eagle Ford synthetic crude oil two-phase formation volume factor with pressure. .. 119
Figure 5.10: Relative permeability curves krw & krow vs Sw plotted by using Table 5.4. ........ 122
Figure 5.11: Relative permeability curves- krg & krog vs. Sg plotted by using Table 5.4.......... 122
Figure 5.12: Relative permeability curves krg & krog vs. Sl plotted by using Table 5.4. ........... 123
Figure 5.13: Base reservoir model of the Burleson County lease well. ....................................... 124
Figure 5.14: Base reservoir model of the Dimmit County well-Hutch. ....................................... 125
Figure 5.15: 3D view of the Hutch well showing the horizontal well and the hydraulic fractures.
............................................................................................................................................. 125
Figure 5.16: History match between simulation output (blue and gold lines) and the past
production data (red and green circles) for Giesenschlag-Groce well in the Burleson County.
............................................................................................................................................. 126
xix
Figure 5.17: Comparison of EUR forecasted to 2 barrels/day, between numerical simulation
model (NSM) and various analytical decline curve analysis techniques for Giesenschlag-
Groce well. ........................................................................................................................... 127
production data (red and green circles) for Hutch well in the Dimmit County. .................. 128
model (NSM) and various analytical decline curve analysis techniques for Hutch well. .... 129
Figure 5.20: Initial pressure of the reservoir for the Hutch well model in Dimmit County. ....... 129
Figure 5.21: Pressure drawdown close to the wellbore, hydraulic fractures and within reservoir
after 6 months of production of the Hutch well. .................................................................. 130
after 12 months of production of the Hutch well. ................................................................ 130

xx
Figure 5.27: Reservoir model (5P) consisting of 5 hydraulically fractured horizontal producers for
primary shale oil depletion. This model is later converted to 3-2/P-I with 3 producer and 2
injectors for ESOR ............................................................................................................... 134
Figure 5.28: Reservoir model (7P) consisting of 7 hydraulically fractured horizontal producers for
primary shale oil depletion. This model is later converted to 4-3/P-I with 4 producer and 3
injectors for ESOR. .............................................................................................................. 135
Figure 5.29: Initial pressure condition for the reservoir model (7P) consisting of 7 hydraulically
fractured horizontal producers for primary shale oil depletion. This model is later converted
to 4-3/P-I with 4 producer and 3 injectors for ESOR........................................................... 135
Figure 5.30: Production rates profile of reservoir simulation under primary depletion drive and
miscible gas injection displacement process (5P and 3-2/P-I well patterns) ....................... 138
miscible gas injection displacement process (7P and 4-3/P-I well patterns). ...................... 138
Figure 5.32: Production Gas-Oil-Ratios of the miscible gas injectants under primary depletion
drive and miscible displacement process. ............................................................................ 142
Figure 5.33: Uniform reservoir pressure of 3200 psi at the onset of primary oil recovery .......... 144
Figure 5.34: The reservoir pressure depleted from 3200 to 1850 psi uniformly by 2015 during the
primary depletion drive. Note the wells are all producers, well cone pointing up. .............. 144
xxi
Figure 5.35: Next pressure change in the reservoir was not until 2038 at the onset of miscible gas
injection due to injector bottom hole pressure of 4000 psi. Note: cones pointing down are the
injectors. ............................................................................................................................... 145
Figure 5.36: Pressure wave is highest at the injectors and least at the producers creating a pressure
gradient. This pressure waves moves the oil towards the producers. .................................. 145
Figure 5.37: Pressure profile did not change much between 2040 and 2100. Almost an abrupt
sharp pressure drop can be observed in the spaced zone between producers and injectors. 146
Figure 5.38: Simulation model to investigate bottom hole pressure in adjoining leases during
depletion............................................................................................................................... 148
Figure 5.39: Pressure profiles of all observation wells at distances away from the ends of
stimulated reservoir volume (SRV). .................................................................................... 150

xxii
LIST OF TABLES
Table 2.1: Porosity and permeability values used in Eagle Ford Shale simulation models ........... 25
Table 2.2: Eagle Ford pressure gradients ....................................................................................... 26
Table 2.3: Eagle Ford Shale parameters previously used in simulation ........................................ 26
Table 2.4: Equation of State parameters and Compositions of a Synthetic Eagle Ford Shale Oil
(Orangi et al., 2011) ............................................................................................................... 28
Table 4.1: Burleson County wells’ completion and production history ........................................ 65
Table 4.2: Leon County wells’ completion and production history............................................... 71
Table 4.3: Brazos County wells’ completion and production history ............................................ 77
Table 4.4: Lee County wells’ completion and production history ................................................. 82
Table 4.5: Gonzales County wells’ completion and production history ........................................ 85
Table 4.6: Karnes County wells’ completion and production history ........................................... 92
Table 4.7: Dimmit County wells’ completion and production history .......................................... 98
Table 4.8: Zavala County wells’ completion and production history .......................................... 103
Table 5.1: Reservoir fluids, depths, lateral lengths and stimulation data (from www.rrc.state.tx.us
June 10, 2012) ...................................................................................................................... 112
Table 5.2: Initial and final reservoir model parameters of the Eagle Ford Shale (EFS) .............. 112
Table 5.3: Miscible gas injectants and multiple contact miscibility pressures ............................ 120
Table 5.4 (a-b): Relative permeability tables generated using CMG Builder. ............................. 121
Table 5.5: Primary depletion, ESOR cumulative oil produced, and recovery factors ................. 139
Table 5.6: Miscible gas injectants properties at reservoir operating condition during ESOR. (Note:
Viscosity ratio (µ) and fraction of rates (q) are relative to the highest property values of CO2)
............................................................................................................................................. 141
xxiii
Table 5.7: Pressure responses across stimulated reservoir volume into non-stimulated reservoirs
............................................................................................................................................. 149
xxiv
ACKNOWLEDGEMENT
I would like to thank my advisor Dr. Mohabbat Ahmadi, and other members of my graduate study
committee; Dr. Catherine Hanks and Dr. Abhijit Dandekar for all their support and contributions.
Without their intellectual advice and feedback during the entire study this work would not have
been academically successful. I appreciate their input of talent and time in carefully evaluating
every stage of my work.
I would also like to thank Dr. Jim Erdle of CMG for making available their simulation software
for me to use even while on vacation in Ghana. I am deeply grateful to CMG support staff for
their assistance on numerous occasions.
I extend my thanks also to the staff and students of the Petroleum Engineering department at
UAF.
Finally, I express my sincere gratitude to my wife and children, parents and siblings, and my
Christian fellowship brothers for believing strongly in me and for their prayerful supports.
Finally, my praise and worship for this accomplishment belong to GOD the Invincible, Author of
Life, Alpha and Omega.
Dieudonne K. Delaihdem
1
Chapter 1 Introduction
1.1 Shale Oil Reservoirs
Hydrocarbon accumulations in petroleum reservoirs around the world migrated from very fine-
grained, dark-gray or black organic-rich sedimentary source rocks, referred to as organic-rich
shales. For decades, organic-rich shale formations have been regarded as source rocks from
which hydrocarbons originated and migrated into sandstone and limestone of various reservoir
qualities. Oil- and gas-prone shales form when massive amounts of organic debris deposition
occur in swamps, lakes, marine environments, followed by rapid burial without decay (Passey et
al., 2010). Subsequently, over geologic time, these organic constituents convert into hydrocarbons
under the effect of temperature and pressure changes in the subsurface due to burial. Thus, these
organic-rich shales undergo the necessary geologic processes from diagenesis to catagenesis to
convert dead organic contents into useful hydrocarbons. The Eagle Ford Shale, located within the
Maverick Basin in south Texas, is a perfect example of organic-rich shale, and it is a world-class
source rock for a number of conventional petroleum systems such as the Austin Chalk and the
East Texas oilfields (Sondhi, 2011). The geochemical evidence of Eagle Ford Shale sourcing the
Austin Chalk is the presence of similar kerogen type II found in both reservoirs (Martin et al.,
2011). The Eagle Ford Shale oil, however, is generated from kerogen type I, I/II and III, which
implies that the overall hydrocarbon composition of the Eagle Ford Shale and Austin Chalk may
be significantly different.
The decline in the conventional hydrocarbon reserves, especially in the United States, motivated
researchers to explore organic-rich shales as potential hydrocarbon reservoirs. Geoscientists
recognized the reservoir potential of the ultra-low permeability organic-rich shale formations
saturated with and/or still “cooking” petroleum. Petroleum engineers have undertaken to develop
2
technologies for economic and commercial development of these resources. Technological
innovations over the years, coupled with increasing demands and rising prices of petroleum
products, generated considerable interest in developing shale oil and shale gas reservoirs,
particularly in the United States and Canada (Mullen, 2010).
Shale oil and conventional oil reservoirs require different development strategies for economic oil
recovery. Low permeability shale reservoirs require extensive hydraulic fracture stimulation
treatments at the onset of economic oil recovery (Miskimins, 2008). On the other hand,
conventional reservoirs have relatively good permeability, and therefore produce at economic
rates without hydraulic fractures. Understanding petrophysical and geomechanical properties is
essential for optimum stimulation treatment of shale reservoirs.
1.2 Shale Oil- What is it?
Shale oil is crude oil produced from, and still residing in, organic-rich, dark-gray, black shale
formations (source rocks) that are in the oil window of thermal maturity (Chaudhary, 2011).
Shale reservoir rocks are characterized by nano-Darcy permeability and micro porosity and are
classified as ultra-low or tight permeability formations. The nano-Darcy permeability, the non-
communicating micro porosity and lack of structural conduits prevented the migration of the
hydrocarbons generated in these shales into conventional reservoirs. As a result, economic
recovery is particularly challenging without the current innovations in petroleum technology
(Medeiros et al., 2008). Shale oil should not be confused with oil shale, organic-rich shale which
contains kerogen (solid organic matter) that has not yet reached the thermal maturity for
generating any hydrocarbon phases (Chaudhary, 2011). Shale oil is liquid hydrocarbon (crude oil)
3
trapped in micro porous source rocks due to lack of conventional formation permeability and
natural migration pathways; consequently, shale oil requires unconventional recovery processes.
1.3 Shale Play Development
Economic development of a shale oil play depends on the shale’s total organic content (TOC),
petrophysical characteristics, and reservoir properties (Mullen, 2010). A TOC in excess of 3% is
usually a good cut-off-grade (McFarland, 2010). The Eagle Ford TOC is 2.0% to 6.5%
(Vassilellis et al., 2010). Rock-Eval pyrolysis, vitrinite reflectance, and burial history of shale
deposits are used to evaluate the type of kerogen present and level of thermal maturity in order to
determine the petroleum potential of the shale play (McFarland, 2010). Vitrinite reflectance (VF)
between 0.6% and greater than 1.35% in an oil-prone shale correlates with a hydrocarbon
generation window, which is 60 to 120 degrees Celsius (McFarland, 2010). Another important
factor in determining the viable development of shale oil is the areal extent and thickness of the
organic-shale deposits (McFarland, 2010).
Successful exploitation of shale oil reservoirs requires sufficient liquid-conductive fractures in the
otherwise impervious shale. This enhances fluid flow from the oil-saturated matrix into horizontal
wellbores. Thus, shale formations are normally impermeable except for the presence of natural
fractures. Natural fractures form as a result of tectonic activities, overburden pressure and/or
pressurized liquid expulsion (Berg and Gangi, 1999). Consequently, effective characterization of
shale reservoirs requires knowledge of the geology (lithology and mineralogy) of a formation in
order to understand and predict the petrophysical parameters such as natural fracture density and
distribution, hardness, and brittleness index (Mullen, 2010). Major engineering challenges,
4
therefore, include adequate reservoir characterization for the appropriate wellbore stimulation
design (Vassilellis et al., 2010). The challenges include numerical analysis of natural fracture
networks, determination of rock geomechanical properties, and proper selection of hydro-
fracturing fluids (Vassilellis et al., 2010).
Shale plays’ geomechanical and geochemical attributes vary, and individualized developmental
approaches may be required for different shale reservoirs (Rickman, et al., 2008). The Eagle Ford
is not only mineralogically unique but also produces three hydrocarbon types, requiring different
development strategies within the play (Mullen, 2010).
1.4 Purpose and Scope of Study
Decline curve analysis of past production data is an essential tool petroleum reservoir engineers
use to estimate reserves, evaluate well performance, improve well completion effectiveness, and
determine reservoir properties. In the first part of this study, the available production decline
techniques assessed the past wells’ performance, forecasted the future performance rates and
estimated reserves for selected wells in the Texas Eagle Ford Shale play. These various
techniques include Arp’s empirical decline curve analysis (DCA), logistic growth analysis
(LGA), power law exponential analysis, and Duong’s method. In addition, a new approach,
involving the use of logistic growth analysis cumulative production data in Arps’ hyperbolic
decline model, to eliminate the dominance of decline exponents greater than one was proposed
and used. The techniques evaluated production profiles for forty hydraulically fractured
horizontal wells in the Eagle Ford for expected ultimate recovery (EUR) and future performance.
In the second part of the study, numerical simulation models (NSM) with realistic researched
data: reservoir rocks, fluids, wellbore and hydraulic fracture were used to history match the past
5
production of two wells in the Eagle Ford Shale play and generate future performance
predictions. Furthermore, NSM accomplishes other goals such as:
1. Compares and contrasts the practical effects of miscible gas injectants such as CO2, CH4,
CO2-CH4 1:1 mixture and other combinations of CO2, CH4, N2, C2H6, C3H8, to enhance
shale oil recovery beyond primary depletion,
2. Explores the technical feasibility of enhanced shale oil recovery (ESOR) by miscible gas
injection displacement process using different numbers and arrangements of, and
3. Defines the drainage area of several horizontal wells in a reservoir of similar modeling
parameters as for ESOR using observation wells outside of the stimulated reservoir
volumes (SRV).
6
Chapter 2 Literature Review
2.0 Introduction
The Eagle Ford Shale play is both gas and liquid hydrocarbon-rich (Stegent et al., 2010 and
Vassilellis et al., 2010). The hydrocarbon resources of the Eagle Ford are roughly 3.35 billion
barrels of oil and 20.80 trillion standard cubic feet of natural gas (Government-Report, 2011).
The Eagle Ford Shale is an unconventional hydrocarbon resource; thus, economic recovery is not
possible without the implementation of advanced petroleum recovery technology (Holditch,
2003).
The Eagle Ford Shale is of Upper Cretaceous (Cenomanian-Turonian) age (Tuttle, 2010). It is the
source of hydrocarbon accumulations in the overlying Austin Chalk and the massive East Texas
Woodbine sands fields (Fan et al., 2011 and Tuttle, 2010). It is currently one of the most
attractive source reservoirs in Texas due to its relatively high total organic content (TOC) of 2.0-
6.5% (Fertl and Rieke III, 1980; Mullen, 2010; Vassilellis et al., 2010).
Three distinct hydrocarbon phases are produced from the Eagle Ford Shale and, depending upon
the development site (county, lease and depth); they include dry gas, gas condensate, and oil
(Mullen, 2010). As a result, hydrocarbon production requires reservoir engineering techniques
tailored to the phases of hydrocarbon production (Mullen, 2010).
Production rates of the Eagle Ford oil wells decline rapidly after the first two or three months of
high liquid withdrawal (TRRC-Website, 2012). About 40% of the Eagle Ford wells’ cumulative
production could be achieved within five years despite the projected 30-year longevity for most
of the wells (Hart, 2011). Thereafter, the wells will experience low recovery of fluids unless there
7
is additional intervention such as re-fracturing of the formation and/or application of enhanced oil
recovery (EOR) techniques (e.g., gas injection) (Hart, 2011).
The oil recovery factor for the Eagle Ford Shale play during primary energy reservoir depletion
will be roughly 5%. However, improvement in fracking technology, fracturing of wells multiple
times, longer laterals, closely spaced well intervals, and miscible displacement (especially using
CO2 solvents) all can improve shale oil recovery factors (Hart, 2011).
2.1 Geology
The Eagle Ford Shale play, located in the Maverick Basin of southeast Texas, is a SW-NE
trending source and reservoir rock. It stretches approximately 400 miles long and 50 miles wide,
from the Mexican border in the SW into the NE of Texas (Wang and Liu, 2011 and Vassilellis et
al., 2010) (Figures 2.1 and 2.2). The Eagle Ford Shale formed in the Upper Cretaceous
(Cenomanian-Turonian) due to major anoxic extinction of microorganisms and rapid
sedimentation in restricted marine basins (Tuttle, 2010 and Martin et al., 2011).
The Eagle Ford Shale covers almost 11 million acres. The stratigraphic column in Figure 2.3
shows the Eagle Ford Shale underlain unconformably by the Buda Limestone and overlain by the
Austin Chalk. The gross thickness of the formation ranges from 20 to 500 ft, while the formation
top varies between subsea elevations of 2500 ft to 14000 ft (Mullen, 2010; Nwabuoku, 2011 and
Wang and Liu, 2011). The rejuvenated mid-Cretaceous Sabine uplift and subsequent erosion of
the Eagle Ford Shale after it had been deposited account for the reduction in the thickness of the
8
formation to about 50 ft in the NE from the average gross thickness of 350 ft in the SW (Martin et
al., 2011).
Figure 2.1: Map view of southern Texas showing Eagle Ford Shale resource play, counties,
scheduled oil and gas and permitted wells locations (TRRC, 2012).
9
Leon
Brazos
Burleson
Lee
Gonzales
Karnes
Zavala
Dimmit
Figure 2.2: Map view of the Eagle Ford Shale hydrocarbon phases and windows of thermal
maturity with arrow labels indicating the researched counties’ locations (Wang and Liu, 2011).
10
Figure 2.3: Cretaceous stratigraphic column from south Texas showing stratigraphic location of
the Eagle Ford shale (Fan et al., 2011).
Hydrocarbon accumulations within the Eagle Ford Shale are closely related to local stratigraphy.
The stratigraphic sequence consists of two main clastic depositional environments. These include
a lower transgressive sequence of black, organic-rich, laminated shale and an upper regressive
sequence of quartz-rich siltstone, bentonites, limestone, and calcareous black shale. Both of these
sequences have natural fractures, sealing faults, and variable thicknesses (Martin et al., 2011).
The increased permeability associated with natural fractures augments oil and gas storability in
localized sweet spots, and sealing faults confine and trap hydrocarbon accumulations in the
fractures while also preventing migration out of the source rock (Mullen, 2010 and Martin et al.,
11
2011). Consequently, the amount of recoverable hydrocarbon varies across the Eagle Ford Shale
play due to the sequence stratigraphy, facies changes, and natural fracture distribution.
2.2 Decline Curve Analysis
Analysis of past production decline to predict future production performance is valuable to oil
and gas industry operators and financial resource institutions. Arps (1945) developed methods to
analyze decline trends in conventional reservoirs with great success. Hydrocarbon recovery from
unconventional shale reservoirs is increasingly crucial, especially in the US and Canada; it is
crucial to evaluate current production trends, predict future production performance and evaluate
the effectiveness of different hydraulic fracture stimulations and completion designs. The
traditional Arps decline models have not successfully estimated reserves or future production in
tight permeability reservoirs (Duong, 2010). One of the assumptions in Arps’ empirical equations
is the existence of a boundary-dominated flow regime, which is observed historically for most
conventional reservoirs. This assumption does not apply to tight permeability shales, which are
dominated by long transient flow regimes. As a result, traditional Arps’ decline relations may
result in decline exponent (b) greater than one (i.e., b>1) for shale reservoirs. The effect of b>1 is
unrealistic infinite cumulative production forecast with time. Thus, Arp’s hyperbolic models tend
to overestimate the expected ultimate recovery, despite closely fitting the past production rate
profiles.
To ensure accurate prediction mechanisms, researchers have proposed other analytical models to
predict future production and estimate reserves for shale reservoirs (Duong, 2010). The recent
12
decline analysis techniques used to specifically address production forecasting in shale reservoirs
include:
1. Power Law Exponential (PLE) Analysis by Ilk et al., 2008,
2. Logistic Growth Analysis (LGA) by Clark, 2011, and
3. Duong method by Duong, 2011.
McNeil et al. (2009) recommended that erratic production profiles be filtered by elimination of
data points that deviate from a reasonable decline trend.
2.2.1 Arps Decline Curve Analysis
Arps (1945) developed the mathematical relations for three types of graphical representation of
production decline for conventional reservoirs. These empirical equations define the historical
exponential, hyperbolic, and harmonic decline types observed for different qualities of traditional
reservoirs. The basic concept of decline analysis involves fitting a trendline through a well’s
historical performance on a semi-log plot and extrapolating that line to estimate future production
performance, assuming the past trend will not change under constant operational conditions.
Mathematically, it is the concept of loss-ratio (1/D) and the derivative of the loss-ratio (b), where
D and b are the decline parameter and decline exponent, respectively, expressed as follows (Arps,
1945):
(2.1)
(2.2)
Equation 2.1 translates into three diagnostic equations corresponding to the three decline types
(exponential, harmonic and hyperbolic). When b=0, the decline type is exponential or constant
13
percentage which, graphically, is a straight line fitting historical performance on a semi-log plot
of rate versus time or rate versus cumulative production (Arps, 1945). The empirical equations for
exponential decline are:
(2.3)
and
(2.4)
where
D =Decline constant, 1/day or 1/time
q = Production rate at time t, stb/day or stb/time
= Initial rate, stb/day or stb/time, and
Np = Cumulative production at time t, STB
When b=1, decline is harmonic and a straight line on the semi-log plot of rate versus cumulative
production. The empirical equations for harmonic decline are:
(2.5)
and
(2.6)
where;
14
Di =Initial decline constant, 1/day or stb/time
q = Production rate at time t, stb/day or stb/time
Np = Cumulative production at time t, STB
Hyperbolic decline is for the limit 0<b<1. The hyperbolic decline plot is a curve (concave
upwards) on the semi-log rate-time, since decline exponents change with time, in contrast to the
constant percentage decline (Clark, 2011). Low productivity wells exhibit hyperbolic-harmonic
decline behavior (Clark, 2011). Arps’ hyperbolic modeling equations are as follows:
(2.7)
(2.8)
In case of adequate production data, a horizontal shale well’s performance is commonly analyzed
using Arps’ hyperbolic rate decline (Duong, 2010 and Shelley et al., 2012). As expected in
fracture-dominated, extremely tight formations, initial production rates are extremely high,
followed by rapid decline ruled by the unusually long transient flow regimes (Medeiros at al.,
2008 and Duong, 2010). The benefit of using the hyperbolic model for shales is that a reasonable
match is usually obtained for wellbore performances that are characterized by long transient flow
regimes (Duong, 2010). The downside is that cumulative production becomes infinite due to
matching data with b>1. The transient flow regime behavior portrays an infinite-acting reservoir
because the pressure response profile in the reservoir constantly moves outward, apparently never
able to reach the finite dimensional boundaries, resulting in an erroneous EUR (Clark, 2011).
15
2.2.2 Power Law Exponential Analysis
This model, proposed by Ilk et al., 2008, models the loss-ratio and its derivative for production
decline by a power law relationship observed in transient linear and bilinear flow regimes. In
contrast to Arps’ exponential decline model, it represents decay as a power law function instead
of a constant decline (Clark, 2011). Ilk et al. (2008) stated that the production profile usually
follows exponential rate decline, aside from the early-time rapid decline rate; therefore, PLE
analysis can estimate reserves. According to McNeil et al. (2009), the first one or two data points
from the production profile may not count towards having a good data production fit.
With the power law exponential loss-ratio method, Ilk et al. (2008) define the decline parameter
(D) and production rate (q) respectively as
∞ (2.9)
∞
(2.10)
With further modification by Ilk et al., (2008), the PLE Equation 2.10 becomes
∞ (2.11)
where,
= Decline constant “intercept” at 1 time unit, (t=1, )
∞ = Decline constant at “infinite time” ∞ )
= Decline constant ( = /n),
n = Time exponent, dimensionless
q = Production rate, stb/d
= Rate “intercept”, (t=0, stb/d)

16
t =production time (days)
The model parameters , D∞, and n are required to make data adjustments for early-time
transient flow and late-time constant decline behavior. Also the tn term (Equations 2.10 and
2.11) matches early-time transient flow regime, while ∞ models late-time decline behavior
(Clark 2011). According to Ilk et al. (2008), Equation 2.9 approximates the loss-ratio by a
decaying power law relation where ∞ is constant for late-time behaviors.
To estimate the EUR, decline parameters (initial rate (qi) and decline rate (D)) are determined
from the semi-log plot of rate versus time using the PLE-generated profile. This diagnostic plot
must linearly fit a trendline of least square regression R2≥0.97 to estimate expected ultimate
recovery EUR and generate future production curves.
2.2.3 Logistic Growth Analysis
Logistic growth analysis (LGA) is an empirical mathematical technique used to analyze
numerous physical trends such as population (Clark 2011). LGA is able to analyze well
production performance and predict future reserves, especially in shale oil reservoirs (Clark,
2011). Clark (2011) clearly explained the use of LGA with original concepts and modifications
necessary for its application to tight permeability formations.
The recommended approach starts with the least square regression of cumulative production (Q
(t)) versus time (t) by means of the mathematical expression:
(2.12)
17
where,
K = Carrying capacity (maximum physically recoverable oil) (stb)
a = Constant of tn when half the oil has been recovered, (day), and
n= Exponential parameter (hyperbolic exponent) (Clark, 2011).
Subsequently, the corresponding equation for production rate (q) versus time (t) is given as
(2.13)
2.2.4 Duong’s Method
Duong’s method addresses the fracture-dominated flow in shale reservoirs due to the lack of
matrix permeability. Traditional decline curve analysis (DCA) is based on a drainage area with
relatively good matrix permeability, which subsequently establishes pseudo-radial and boundary-
dominated flows (BDF). However, in a shale reservoir, the drainage zone consists primarily of
natural fractures and the stimulated reservoir volume (SRV); thus, it is a fracture-dominated flow
reservoir and no BDF is established due to nano-Darcy matrix permeability. Duong (2011) stated
that with negligible matrix contribution compared to that of fluid flow in fractures, the
determination of EUR based on the traditional concept of drainage area is inaccurate.
Duong (2011) outlines the detailed procedure for evaluating and forecasting cumulative
production using his model. It involves two diagnostic plots by means of the empirical equations
below (Equations 2.14 and 2.15).

18
(2.14)
where,
a = Intercept constant, day-1
= Cumulative oil production, STB
m = Slope (which must be greater than unity for unconventional tight formations)
q = Oil rate, stb/day
t = Production time, days
A log-log plot of the above relation (q/Np vs. t) gives a straight line with a negative slope, -m,
and an intercept, a, which are two of the four unknown parameters in Duong’s decline curve
analysis method. Note that the slope is negative, but m is always a positive value. The value of m
must be greater than one (1) to indicate unconventional tight reservoirs.
Another unknown parameter q1- oil rate at Day 1- is determined using the relationship (Duong
2011):
(2.15)
(2.16)
Note: (2.17)
where,
= Oil rate at Day 1
= Oil rate at infinite time

19
t (a, m) = Time function based on Equation 2.17
If the resulting linear plot for any production profile using Equation 2.15 fails to pass through the
origin, Duong (2010) proposed the use of Equation 2.16. The modification term , defined as
the production rate at infinite time, is an anomaly related to wellbore operating conditions (Duong
2010). The two diagnostic plots are used to determine variables a (Intercept constant, day-1), m
(Slope of q/Np versus t), q1 (Oil rate at Day 1), and q∞ (Oil rate at infinite time).
Rate-time and cumulative production-time relationships to estimate forecast at abandonment are
given as (Duong 2011):
(2.18)
(2.19)
2.3 Analyses of the Eagle Ford Shale
2.3.1 Reservoir Fluids
The Eagle Ford Shale reservoir has dry gas, wet gas, and liquid hydrocarbon depending upon the
depth and temperature of the reservoir. The maturity windows corresponding to the liquid, wet
gas and dry gas zones across the entire formation are shown in colors of green, orange, and red,
20
respectively, in Figure 2.2. The phases’ distribution in the Eagle Ford source rock reservoir is not
gravity segregated as associated with traditional reservoirs due to the different maturity windows
for in-situ, self-sourcing shale reservoir hydrocarbon phases. The oil-rich zone is in shallower
depths than the less dense gas strata in the Eagle Ford Shale instead of the conventional gas cap
and underlying liquid in multi-phase reservoirs. Thus, in a multi-phase conventional reservoir;
gravity controls the segregation of the fluids. On the other hand, tight permeability shale reservoir
multi-phases are zoned according to hydrocarbon maturity windows depending on reservoir depth
and temperature. The liquid-rich, wet gas, and dry gas regions occur at estimated depths of 8000
ft, 10000 ft, and 14000 ft, respectively (Chaudhary et al., 2011).
2.3.2 Natural Fractures
Cores from the Eagle Ford Shale indicate the presence of mineralized vertical natural fractures
(Orangi et al., 2011). Orangi et al. (2011) also confirmed the presence of micro fractures. Micro-
fracturing of rock formations occurs naturally by liquid pore pressure (Berg and Gangi, 1999).
Also, increasing pore pressure typically associated with the conversion of high density kerogen to
light hydrocarbon fluids creates natural fractures in source reservoir rocks (Berg and Gangi,
1999). In this model, natural fractures are due to the differential pore volume resulting from
kerogen conversion into crude.
The brittle and calcareous zones of the Eagle Ford Shale are easily fractured and could result in
localized sweet spots. These oil-saturated natural fractures extend between 10 ft and 30 ft within
brittle, carbonate rich (dolomitic shale) inter-bedded shale layers rich in calcite, chert, and silt
(Fertl and Rieke III, 1980). Brittle, naturally fractured zones within the organic-rich source rock
21
are targets for optimal well placement, hydraulic fracture staging, and completion designs (Fertl
and Rieke III, 1980).
According to Stegent et al. (2010), outcrop- scale natural fractures are rare in the calcareous
portion of the Eagle Ford Shale. He pointed out evidence of dense micro-fractures, as shown in
Figure 2.4.
Figure 2.4: Micro-fractures in Eagle Ford Shale cores (Stegent et al., 2010).
22
2.3.3 Reservoir Stimulation and Wellbore Completions
Shale reservoirs are economically viable as a result of technological advancements in petroleum
engineering tools and methodologies. These advanced technologies include horizontal wellbore
drilling with multistage and transverse hydraulic fracturing (Wei and Economides, 2005; Mullen,
2010). Drilling horizontal wellbores along the direction of the least principal in-situ stress ensures
propagation of transverse fractures (Nwabuoku, 2011).
Hydraulic fracturing increases reservoir contact with the horizontal wellbores, decreases wells’
drawdown, and ultimately increases conductivity to wellbores (Medeiros et al., 2008). Near
wellbore stimulation by multiple transverse fractures effectively creates reservoir contacts for
ultra-low permeability reservoirs with horizontal wellbores (Vincent, 2011). The ultra-low
permeability Eagle Ford Shale reservoirs, therefore, require horizontal wellbores stimulated with
optimal hydraulic fracturing techniques (Medeiros et al., 2007).
Stegent et al. (2010) concluded that successful hydraulic fracture treatment and strategic wellbore
completions of the Eagle Ford Shale reservoir required a thorough understanding of its
petrophysical properties. Comprehensive petrophysical analysis and reservoir quality assessment
may involve gathering the following information:
1. Geochemical data to assess hydrocarbon maturity and kerogen types
2. Shale or gamma ray logs to calculate TOC, kerogen, free gas volume, formation shaliness
and brittleness
3. Electric logs and resistivity logs for fluid saturation determination
4. Sonic logs to assess stress orientations, Poisson’s ratio, and Young Modulus, and
23
5. Borehole images and dipmeter logs to identify structures and sedimentary features and
their orientations (Mullen, 2010 and Stegent et al., 2010).
The Eagle Ford Shale consists of 38-88% clay, with ~ 50% of that clay being smectite, a clay
mineral susceptible to swelling. As a result, potential clay swelling during development is a
considerable challenge (Hsu and Nelson, 2002).
As of late 2008, horizontal wells of the Eagle Ford Shale reservoir typically had 12-20 hydraulic
fractures in multiple stages along the lateral leg (Martin et al., 2011). Review of some of the
earlier horizontal wells drilled within the Eagle Ford Shale had as many as 4 clusters per 14-16
stages. Hydraulic fracture stages have been spaced at equal intervals of about 250 ft over a lateral
length of about 4000 ft (Nwabuoku, 2011).
Optimal placement of hydraulic fractures needs to be based on shale log data which can be used
to identify the brittle and TOC rich zones. This may result in uneven hydraulic fracture intervals,
contrary to the common approach of equal spacing of hydraulic fracture stages (Mendoza at al.,
2011). Evaluation of optimal fracture staging and perforation clusters per stage is essential for
effective shale reservoir drainage; too closely spaced fractures can lead to undesired fracture
geometry and less productive Stimulated Reservoir Volume (SRV), while largely spaced fracture
stages will generate fractures with less reservoir contact (Nwabuoku, 2011). Closely spaced
adjacent transverse fractures may create zones of attraction that can cause interference between
successive fracture stages through stress shadowing (Roussel et al., 2012). Stress shadowing can
result in disruption of transverse fracture patterns, re-stimulation of previously generated
transverse fractures, and failure to initiate new fractures (Roussel et al., 2012). Roussel et al.
24
(2012) observed that fracture spacing greater than 200-300 ft propagates fractures perpendicular
to the lateral in brittle zones. Ductile regions along the wells’ trajectories, however, can
accommodate closely spaced fracture stages with only a small amount of proppant and smaller
fracture widths (Roussel et al., 2012).
The effective drainage area in most tight permeability reservoirs is confined, usually to a
rectangular region containing systemic complex fractures. This drainage area consists of both
natural fractures, if present, and the SRV generated by hydraulic fracturing of the reservoir unit
(Medeiros et al., 2008).
Perkins and Kern (1961) observed that competent reservoir rocks with high brittleness (or with
low ductile characteristics) respond in elastic fashion to applied pressure. As a result, the
dimensions (width and length) of fractures propagated in hydraulic fracture operations depend on
the intensity of applied pressure (fracturing fluid injection rate) and the fluid’s viscosity.
Hydraulic fractures at depths of producing intervals in the Eagle Ford Shale propagate
perpendicularly to minimum horizontal stresses, resulting in mostly vertical fractures (Wei and
Economides, 2005). Consequently, horizontal wells’ orientations are key considerations for
drilling, stimulation, and completion designs, in order to create the most efficient network of
transverse fractures (Wei and Economides, 2005).

25
2.3.4 Simulation
This section reviews the simulation parameters used to characterize reservoir models of the Eagle
Ford Shale play. Table 2.1 shows the different ranges of porosity and absolute permeability
values that different authors incorporated in previous reservoir simulations models. The different
values correspond to different parts of the Eagle Ford Shale.
Table 2.1: Porosity and permeability values used in Eagle Ford Shale simulation models
Absolute Matrix Permeability Hydrocarbon zone/Sources
Porosity (%) (Nano-Darcy) referenced
5.0-14.0 40-1300 Oil/Wang and Liu, 2011
3.0-10.0 3-405 Oil and Gas/Martin et al., 2011
3.4-14.6 10 Gas/Vassilellis, et al., 2010
9.0-12.0 420 Oil and Gas/Mendoza et al., 2011
Oil and Gas/Stegent and Ingram, 2011
8.0-18.0 20-1200 and Mullen, 2010
6.0 100 Oil/Chaudhary et al., 2011
9.0 25-500 Oil/Orangi et al., 2011
Wang and Liu (2011) conducted a sensitivity analysis of the following reservoir and wellbore
parameters with respect to the wells’ performance: natural fracture permeability, reservoir-pay
volume (pay thickness and matrix porosity), natural fracture spacing, stimulated reservoir volume
(SRV) half-length, SRV enhanced permeability, and matrix permeability. They concluded that
well performance was least sensitive to matrix permeability and most sensitive to natural fracture
permeability.
26
The Eagle Ford Shale also has variable pressure regimes and is over-pressured in some places
(Mullen, 2010). Table 2.2 shows some of the reported pressure gradients from the Eagle Ford
Shale.
Table 2.2: Eagle Ford pressure gradients
Pressure gradient (psi/ft) Sources referenced

0.62 Wang and Liu, 2011
0.43-0.65 Vassilellis, et al., 2010
0.55-0.85 Chaudhary et al., 2011
Table 2.3 contains reservoir temperatures, water saturation and rock compressibility parameters
found in the literature.
Table 2.3: Eagle Ford Shale parameters previously used in simulation

Water
Temperature Saturation Rock compressibility Hydrocarbon zone/
(oF) (%) (10-6 psi-1) Sources referenced
315 Oil and Gas/
Mendoza et al., 2011
Oil and Gas/
7-31 Stegent and Ingram, 2011 and
Mullen, 2010
150-350 30 5 Oil/Chaudhary et al., 2011
273 20-100 Oil/Orangi et al., 2011
The natural fracture distribution in the Eagle Ford Shale is particularly complex and difficult to
represent in simulation grid models (Wang and Liu, 2011). Wang and Liu’s (2012) simulation
27
model used a composite fracture width of 0.001 ft, incorporating natural, re-activated and
hydraulic fractures.
Waters et al. (2009) pointed out that the pressure drop within a low permeability shale reservoir is
extremely small due to the nano-Darcy scale of the matrix permeability. The greatest decline
occurs near wellbores and fracture zones. Depending on fracture density, shale reservoirs may
only deplete to about 50% of initial pressure in 60 years, or show minimal pressure depletion in
after 10 years of production (Waters et al., 2009).
Hydrocarbon fluid flow in horizontal wellbores in fractured media in ultra-low permeability
formations goes through a long transient flow regime (Medeiros et al., 2008). The transient
decline may last the entire productive life of the hydraulically fractured horizontal wellbore
(Medeiros et al. 2007). Adequate representation of the transient flow regime in simulation models
of shale reservoirs requires logarithmic, locally-refined grid sizes. This provision captures the
impact of high pressure depletion and saturation changes near the matrix-fracture boundary and
the horizontal wellbore fracture interface (Rubin, 2010).
Orangi et al. (2011) provided synthetic Eagle Ford crude oil data generated and recombined from
typical stock tank oil and separator compositions. The Eagle Ford Shale synthetic crude’s
properties and compositions are shown in Table 2.4 (Orangi et al., 2011). The synthetic crude
compositional model was developed based on the Peng-Robinson equation of state (EOS) (Orangi
et al. 2011).
28
Table 2.4: Equation of State parameters and Compositions of a Synthetic Eagle Ford Shale Oil
(Orangi et al., 2011)
Temperature
Gas density
Gas gravity
Saturation
(RB/STB)
Viscosity
pressure
pressure
(scf/stb)
(Ib/ft3)
Initial
Fluid
GOR
(psi)
(psi)
API
(cp)
(oF)
Boi
Oil 500 41 0.06906 0.905 6300 237 2053 1.3534 0.58
Composition
Pc Tc Acentric Molecular Specific (mole
Component (psia) (oF) Factor Weight Gravity fraction)
N2 492.3 -232.3 0.04 28.01 0.808 0.00073
CO2 1071.3 88.4 0.225 44.01 0.8159 0.01282
C1 673.1 -116.1 0.013 16.04 0.35 0.31231
C2 708.4 90.6 0.0986 30.07 0.48 0.04314
C3 617.4 206.7 0.1524 44.1 0.5077 0.04148
IC4 529.1 275.5 0.1848 58.12 0.5631 0.0135
NC4 550.7 306.3 0.201 58.12 0.5844 0.03382
IC5 483.5 369.7 0.2223 72.15 0.6248 0.01805
NC5 489.5 386.6 0.2539 72.15 0.6312 0.02141
NC6 439.7 455.3 0.3007 86.18 0.6641 0.04623
C7+ 402.8 601.7 0.3739 114.4 0.7563 0.16297
C11+ 307.7 764.9 0.526 166.6 0.8135 0.12004
C15+ 241.4 909.8 0.6979 230.1 0.8526 0.10044
C20+ 151.1 1155.8 1.0456 409.2 0.9022 0.07306
Orangi et al. (2011) again provided a pressure-volume-temperature (PVT) plot for the synthetic
crude of the Eagle Ford Shale (EFS) (Figure 2.5). It is not known whether the plot of formation
volume factor (FVF) versus pressure was generated by correlation or by simulation.

29
Pressure Oil FVF Oil Formation Volume factor (FVF)-Bo vs Pressure

(psi) (rb/stb)
1.4
6000 1.3534
Formation Volume Factor - FVF (rb/stb)

3500 1.358 1.35
3000 1.365 1.3
2500 1.372
2000 1.38 1.25
1800 1.355 1.2
1600 1.335
1.15
1400 1.31
1200 1.29 1.1
1000 1.26 1.05
750 1.23
1
500 1.2 0 1000 2000 3000 4000
250 1.16 Pressure (psi)
14.7 1.05
Figure 2.1: Pressure versus FVF at reservoir temperature of EFS. The table and corresponding
graph in this figure were approximated from the pressure-FVF curve provided by Orangi et al.
(2011).
2.4 Improving Shale Oil Recovery by CO2 Injection and CO2 as Frac Fluid
The oil recovery factor for the Eagle Ford Shale play during primary drive reservoir depletion is
roughly 5% (Hart, 2011). However, improvements in fracking technology such as repeated
fracturing of wells, longer laterals, closely spaced wells, and miscible displacement using CO 2
solvents can lead to enhanced shale oil recovery (Hart, 2011).
Liquid CO2 has minimal near-wellbore effects. It is non-damaging to fractures, generating clean
fractures when used as hydraulic fracturing fluid, either as an additive or a sole proppant-laden
frac fluid (Lillies and King, 1982). Sinal and Lancaster (1987) stated that the main purpose of
CO2 in stimulation “frac jobs” is the ability to produce cleaner hydraulic fractures that eliminate
30
formation damage such as near-wellbore relative permeability reduction, especially in tight and
low pressure formations.
Eagle Ford Shale play development can benefit from using liquid CO2 with proppants in
hydraulic fracturing in a number of ways. CO2 as a hydro-fracturing fluid can reduce or eliminate
effects of clay swelling (Gupta and Bobier, 1998; Hsu and Nelson, 2002 and Stegent and Ingram,
2011). Furthermore, CO2 enhances the sand-propped integrity of fractures and reduces fracture
closure caused by reservoir pressure depletion, because CO2 flow-back to the wellbore is not
dependent on reservoir pressure (Lillies and King, 1982). Using CO2 also eliminates water flow-
back, a characteristic of conventional fracturing fluids.
The main drawback in CO2 hydraulic fracturing is that of its phase behavior-- it must remain a
liquid phase to ensure stimulation “fracking” of the formation to the desired fracture geometry.
Unlike the conventional high viscosity fracturing fluids, CO2 lacks the ability to transport
proppants well into the fractures before vaporizing (Lillies and King, 1982; Settari et al., 1986
and Sinal and Lancaster, 1987). Unless viscosifying agents are incorporated, 100% liquid CO 2
fracking generates less stimulated volume and low conductivity fractures. Because of this, its
usefulness in tight oil reservoirs may be limited (Settari at al., 1986 and Sinal and Lancaster
1987).
However, CO2 also enhances oil recovery in tight permeability formations because of its high
injectivity (Lillies and King, 1982). CO2 typically attains miscibility at most reservoir pressures
and temperatures (Lillies and King, 1982).

31
Holm, (1986) explained miscibility in reservoirs to mean the formation of a homogeneous, single
phase fluid without any interfaces (Figure 2.6) when two or more reservoir fluids and injected
fluid combine at a certain physical condition. The presence of two or more phases creates
separation of the interfaces, resulting in high interfacial tension. Both first-contact and multi-
contact miscible displacement eliminate interfacial tension between the oil and the injected fluid,
which removes or lowers residual oil saturation (Holm, 1986).
Figure 2.2: Illustration of miscible and immiscible (Holm, 1986).

32
CO2 flooding enhances light and medium crude oil recovery from tight reservoir units such as
shale plays due to its favorable phase behavior such as CO2 solubility and development of multi-
contact miscibility of CO2-oil systems depending on reservoir parameters (Gui et al. 2008).
Extraction of intermediate hydrocarbon components by CO2 initiates miscibility between CO2 and
the in-situ oil; the enriched CO2 dissolution as more reservoir oil is contacted leads to miscible or
near miscible displacement of the crude by the CO2 (Malik and Islam, 2000).
CO2 flooding displaces oil from reservoirs by vaporizing and/or condensing gas drive. This
phenomenon is multi–contact miscible displacement, where CO2 eventually becomes miscible
with oil in the reservoir after it sufficiently vaporizes intermediate hydrocarbons. Miscibility
develops when injected gas and reservoir fluid come into contact above threshold minimum
miscibility pressure (MMP) (Holm, 1986 and Rahmatabadi, 2011). MMP is a function of
reservoir conditions (i.e. temperature) and injection gas and oil compositions.
CO2 flooding can be a miscible or near-miscible process (Gui et al., 2008; Bui, 2010 and
Rahmatabadi, 2011). If injected gas and reservoir oil mix to produce a single phase fluid in the
reservoir, then the gas is said to have a first contact miscible (FCM) displacement of the reservoir
oil (Rahmatabadi, 2011). On the other hand, multiple contact miscible (MCM) displacement takes
place when miscibility develops through mass transfer of the gas; i.e., CO2 seeps through the
reservoir to contact fresh oil (Rahmatabadi, 2011).
However, MMP is often impractical to reach, since the depth of many hydrocarbon reservoirs
may be relatively shallow and characterized by lower pressures relative to the threshold for
dynamic mixing of CO2 and crude oil in the reservoir (Bui, 2010). In these cases, CO2 causes
33
viscosity reduction and swelling of the reservoir oil by near-miscible displacement process due to
phase behavior changes when CO2-oil systems form in the reservoir.
Malik and Islam (2000) and Nielson (1989) mentioned that CO2 and reservoir oil are not first
contact miscible; however, the CO2 creates a miscible front for the vaporized hydrocarbon
components. Miscibility occurs dynamically beyond the minimum miscibility pressures between
1500 and 6000 psia (Nielson, 1989). At these pressures, the density of the CO2 increases
sufficiently for solubility with the oil hydrocarbon components (Nielson, 1989). Dynamic
miscibility is achievable in crude oil with at least 25o API gravity (Malik and Islam, 2000 and
Nielson, 1989).
Depending on the reservoir oil composition, dissolution of CO2 in crude oil during miscible
displacement can cause problems with asphaltene precipitation in the reservoir (Ghoodjani and
Bolouri, 2012 and Nielson, 1989). As a result, plugging and clogging in reservoirs by asphaltene
deposition may lead to permeability impairment (Ghoodjani and Bolouri, 2012).
Successful enhanced oil recovery by gas injection partly depends on the reservoir wettability.
Organic-rich shales deposited in swamps, lakes, and marine environments are primarily water-
wet. When petroleum diagenesis progresses, the wettability converts to strongly oil-wet and
weakly water-wet as a result of the mineralogy and organic content (Odusina et al., 2011). This
was supported by a Nuclear Magnetic Resonance (NMR) study which found that organic-rich
shale formations are mainly oil-wet (Odusina et al., 2011).

34
CO2 high injectivity into tight reservoirs makes the choice of CO2 injection for enhanced oil
recovery effective compared to secondary recovery scheme such as water flooding (Holm, 1986).
Also, CO2 injection is preferred to Nitrogen (N2) and hydrocarbon gas injection because of CO2’s
greater relative ability to reduce rock-fluid interfacial tension and oil viscosity (Ghoodjani and
Bolouri, 2011; Ghoodjani and Bolouri, 2012).
Pure CO2, though expensive, has a viscosity and density closer to that of crude oil than natural
gas does (Holm, 1986). This is one of CO2’s advantages over methane in gas injection enhanced
oil recoveries. Also, CO2 attains minimum miscibility pressure at lower reservoir pressures and
can be used in shallower reservoirs than natural gas can (Holm, 1986).
Another matter to consider is that CO2 gas viscosity is low compared to oil, which makes CO2
extremely mobile. Consequently, early breakthrough and bypassing of oil may occur, especially
in shale reservoirs with high induced conductivity (Gui et al., 2008).
The optimum CO2 injection rate depends on the price of oil, original oil in place (OOIP), span of
injection (short or long term), oil viscosity, reservoir size, and reservoir heterogeneity (Ghoodjani
and Bolouri, 2012). To avoid CO2 leaking, especially in the case of both EOR and geologic
sequestration, injection pressure should not exceed the reservoir fracturing pressure (Ghoodjani
and Bolouri, 2012). However, relatively high OOIP, high oil prices, and short term application
may require high CO2 injection rates in order to maximize oil recovery (Ghoodjani and Bolouri,
2012).
35
Typical successful examples of using CO2 in tight permeability reservoirs to enhance oil recovery
include:
1. The Saskatchewan ultra-low permeability limestone and dolomitic shale member had a
predicted 3.64% OOIP recovery factor over 100 years of primary recovery (Wang et al.,
2010). However, enhanced oil recovery during CO2 flooding increased the predicted
recovery factor to over 34%; and
2. Oil production from Elm Coulee Field in Montana from the Bakken Formation of the
Williston Basin had a 5-10% primary recovery factor. CO2 flooding increased oil
recovery factor for the 8-12 ft thick upper shale and the underlying sandy dolomitic layer
by 20% of the OOIP (Shoaib and Hoffman, 2009).

36
Chapter 3 Eagle Ford Production History Assessment Using Various Decline Curve
Analysis Techniques
3.0 Introduction
This chapter discusses the methods and results of different decline curve analysis techniques the
author has used to analyze past production profiles of forty horizontal wells in the Eagle Ford.
These techniques include Arps’ decline analyses relations, power-law exponential (PLE) analysis,
logistic growth analysis (LGA), Duong method and, finally, my own proposed method, Logistic
Arps’ Hyperbolic Approach (LAHA). The analyses cover production profiles of oil-producing
horizontal wellbores selected from eight different counties in the Eagle Ford Shale play. The
selection of the counties is based on their locations in the oil window of thermal maturity,
proximity to one another, extension across most of the oil-prone zones and, finally, publicly
available field development and production data from the Texas Rail Road Commission (TRRC)
website. Counties include Leon, Brazos, Burleson, Lee, Gonzales, Karnes, Dimmit and Zavala
Counties (Figure 2.2).
The objectives of my analyses are to assess the ability of the different decline analysis methods to
match the wells’ past production rates and cumulative production, and generate the future
production trends. The minimum abandonment rate used was 2 STB/day, since many of the wells
are currently operating close to this rate. Most of these wells have short production histories,
which imply decline trends may not have been well established by the wellbores or the reservoirs
at the time of the analyses. The production profiles of wells made publicly available by TRRC
varied from 9-50 months between 2008 and January 2013. Despite this limitation, the importance
of generating early future production forecasts for the shale reservoirs cannot be overemphasized.
37
The results are categorized into two segments. The general results are observations regarding the
respective effectiveness of the different decline curve analysis techniques. The county-well
specific results are considered case studies for the individual counties and wells, and are
discussed separately in Chapter 4. For detailed results, refer to Appendix 1 and 2.
3.1 Methods
3.1.1 Data Preparation before the Analysis
At the start of every analysis, I examined the original data set and plotted the production rates
against time. This made possible the identification of reasonable decline trends established by the
well’s performance data. Data points that did not follow a relatively smooth decline trend were
eliminated and not incorporated in the analysis. Not all the profiles have observable sections of
relatively smooth decline to isolate for the analyses. Consequently, I could not define the decline
paths for all of the forty production profiles, especially for cases where the data is extremely
erratic.
3.1.2 Arps’ Decline Models Applications
I applied the Arps’ decline curve relations, namely exponential, harmonic, and hyperbolic, to
analyze the forty production profiles of the Eagle Ford Shale.
Arps’ exponential or harmonic declines involve using past production data to generate decline
diagnostic plots: rate versus time, cumulative production and the semilog rate versus time, or
cumulative production with the constraint of least square regression R2≥0.95. I generated similar
diagnostic plots for all forty profiles, which I interpreted for either exponential or harmonic
38
decline. Using the decline parameters obtainable from the equation of the trendline, I proceeded
to calculate expected ultimate recovery (EUR) and future production.
The Arps’ hyperbolic decline relation applications involved the use of Microsoft Office’s Excel
multivariable solver tool to determine the decline rate (Di), initial production rate (qi) and the
decline exponent (b) using model Equation 2.7.
(2.7)
With these parameters (Di, qi, and b) established for the best fits for each of the production
declines, I plotted the production rate against time resulting from Arps’ hyperbolic model. I also
calculated the expected ultimate recoveries, the remaining recoverable reserves, and time to reach
abandonment for all the forty oil-producing wellbores using Equation 2.8.
(2.8)
3.1.3 Power Law Exponential (PLE) Analysis Application
I evaluated the forty production profiles from Eagle Ford Shale oil wells according to the power
law exponential decline model. First, I applied Equation 2.11 using the solver tool in Microsoft
Excel to determine values for (rate intercept at time t=0, stb/d), (decline constant at
infinite time )), (decline constant ( = /n), ), and n (time exponent,
dimensionless), such that the least square regression is maintained between the real and the PLE
generated data.
(2.11)
39
Subsequently, diagnostic semi-log rate versus time was plotted to obtain the exponential decline
rate ( ) and the initial production rate ( ). I calculated the EUR, remaining reserve, and time to
reach the abandonment rate of 2 barrels/day using exponential decline relations. This diagnostic
plot must linearly fit a trendline of least square regression R2≥0.97 for optimal benefit in
estimating expected ultimate recovery EUR and generating future production curves.
Alternatively, EUR at the well’s abandonment rate can be evaluated from integral q with respect
with time t. Substituting Equation 2.11 into gives cumulative production (Np) at
abandonment:
The solution to the above integral is beyond the scope of this research and has, therefore, not been
considered.
3.1.4 Logistic Growth Analysis (LGA) Application
To apply the logistic growth model, I used the least square regression method to determine K (the
carrying capacity or maximum physically recoverable barrels of oil), a (the constant of time (t) to
the nth (tn) when half the oil has been recovered), and n (the exponential parameter or hyperbolic
exponent equivalent), by means of Equation 2.12. I also generated the plots of cumulative
production versus time to verify a close match between the actual and the model profiles for the
forty production profiles in the Eagle Ford.

40
(2.12)
With the values K, n, and a determined above, I calculated the model’s production rate decline
and generated profiles of rate versus time using Equation 2.13.
(2.13)
Finally, I calculated the expected ultimate recoveries, the remaining recoverable reserves, and
time to reach abandonment for the forty wells.
3.1.5 Duong’s Decline Curve Analysis Method
For all the forty production profiles researched in the Eagle Ford, I generated two diagnostic plots
to determine the initial variables a, m, q1 and q∞, where,
= Oil rate at Day 1 (stb)
= Oil rate at infinite time (stb)
These diagnostic plots were obtained using the relationships expressed in Equations 2.14, 2.15
and 2.16. I plotted the ratios of production rate and cumulative production (q/Np) to time (t) to
determine the slope (m) and the intercept (a), and also, the production rate against the time
function t(a,m), defined by Equation 2.17. I model both Equation 2.15 and 2.16 to investigate
the effects of q∞ (the production at infinite time), which, in practice, no reservoir will attain.
41
(2.14)
(2.15)
(2.16)
(2.17)
where,
q = Oil rate, stb/day
= Oil rate at Day 1
= Oil rate at infinite time
t (a, m) = Time function based on Equation 2.17
3.1.6 Logistic Arps’ Hyperbolic Approach (LAHA)
This section explains my version of the decline curve analysis technique. The motivation to
investigate this is as a result of the critical observation that the logistic growth analysis technique
has robustly fitted the past cumulative production and the production rates decline trends for all
forty wells. The logistic Arps’ hyperbolic approach (LAHA) seeks to maintain the benefits of the
traditional Arps’ hyperbolic relation, which include the good fit to the past production rates. In
42
addition, LAHA results are expressed in terms of conservative decline curve analysis parameters
to which industry analysts and investors are accustomed.
After the analyses of all forty production profiles, I used the smoother reasonable decline trend
generated from the logistic growth analysis (LGA) model as input for Arps’ hyperbolic decline
relation. By the least square regression method, I obtained new sets of traditional decline curve
analysis parameters, namely the decline rate (Di), the initial rate (qi) and the decline exponent (b)
for the new decline profile generated by LGA. Further, I match the resultant curve with the
original data output by LGA and Arps’ hyperbolic decline models. The expectation of my
approach is to normalize the values of decline exponents b>1, which subsequently will leads to
generating forecasts of realistic values similar to those of LGA. Therefore, I again calculated the
EUR, remaining reserves, and abandonment time resulting from my approach.
3.2 General Results
3.2.1 Data Preparation before the Analysis
Figure 3.1 presents an example of the numerous cases where at least one offset data point was
removed from the profile so that a reasonable decline trend could be achieved and analyzed. This
profile (Figure 3.1) is of the production data of the Simms lease horizontal wellbore, and it shows
some offset data at the beginning and towards the end on the rate versus time plot (Figure 3.1 A).
The semilog rate versus time plot of the same production profile in Figure 3.1 B also supports the
need to eliminate the most erratic data points; the equation of the trendline, with the highest
degree of least square regression, cannot be validated without reasonably excluding sections of
the data.
43
(A) Rate vs Time (B) Semi-log Rate vs Time

60 100
Daily Oil Rate (Barrels)/Day)

40
20
0 1
0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200
Time (t) in days Time (t) in days
Used section of production profile Used section of production profile

Filtered data points not used in analysis Filtered data points not used in analysis
Figure 3.1 (A, B): Rate and semi-log rate versus time of the Simms lease #03914 production
profile, in which the use of the blue marked trend is justified compared to the erratic (red marked)
sections.
Some of the profiles are extremely erratic, making it difficult to make sense of declining
production trends. Examples of these kinds of profiles are presented in Figure 3.2 (A, B). The
analyses these production profiles resulted, for example, in unusual parameter values such as the
Arps’ decline exponent associated with Arps’ hyperbolic model Figure 3.3 (A, B). Further
examples are reported with respect to the specific techniques later in the text.
44
(A) Rate vs Time (Cannon) (B) Rate vs Time (CEF)

600 600
Oil Rate q (Barrels)/Day)

500
400 400
300
200 200
100
0 0
0 200 400 600 800 0 200 400 600 800
Time (t) days Time (t) days
Actual Production Rate Actual Production Rate
Figure 3.2 (A, B): Rate versus time plots for Cannon #09607 lease and CEF #09608, presented
as examples for erratic production profiles.
(A) Rate vs Time for Cannon (B) Rate vs Time for CEF
600 600
500
400 400
300
200 200
100
0 0
0 200 400 600 800 0 200 400 600 800
Time (t) days Time (t) days
Actual Production Rate Arps Hyperbolic Actual Production Rate Arps Hyperbolic
Figure 3.3 (A, B): (A) Arps’ hyperbolic decline analysis for the erratic production profiles of the
CEF #09608 and Cannon #09607 lease wellbores resulted in abnormally high Arps’ decline
exponent; i.e., b=3.05 for (A) and b=1.84 (B) respectively.
45
3.2.2 Arps’ Decline Curve Analysis
Figure 3.4 shows four diagnostic plots I generated in order to verify Arps’ exponential or
harmonic decline for one of the production profiles in Burleson County (Giesenschlag W.H.C
Unit). Similar kinds of diagnostic plots were created for each of the forty production profiles.
Figure 3.5 is also one example of the forty results obtained for Arps’ decline curve analysis using
the hyperbolic relation.
The pie chart (Figure 3.6) summarizes the relative extent to which the different types of Arps’
decline curve analysis successfully modeled the forty production profiles of the Eagle Ford Shale
oil reservoirs. Normalized 6.45% and 29.03% of the production profiles are observed to exhibit
exponential and harmonic decline respectively. A higher percentage was observed for Arps’
harmonic decline when a least square regression R2≥0.85 was imposed to classify Arps’
exponential and harmonic decline. This is consistent with preliminary findings presented in
Agboada and Ahmadi (2013). On the other hand, Arps’ hyperbolic relation fitted the rate profiles
of all forty past production profiles (Figure 3.7). Thus, the Arps’ hyperbolic model closely fits
the production rates well but the corresponding cumulative production is not very close to the
actual history cumulative production. Relative to Arps’ exponential and harmonic decline models,
the Arps’ hyperbolic relation had 20.97% normal decline exponents such as 0<b<1, and 43.55%
abnormal decline exponent b>1(Figure 3.6).

46
Rate vs Time Rate vs Cumulative Production

120 120
100 100
Rate (Barrels)S/Day)
Rate (Barrels)/Day)
80 80
60 60
40 40
20 20
0 0
0 500 1,000 1,500 2,000 0 10,000 20,000 30,000
Time (t) days Cumulative Production (BARRELS)
Actual Production Rate

Log (Rate) vs Time Log(Rate) vs Cumulative

Production
100
100
Rate (Barrels)/Day)
Rate (Barrels)/Day)
y= 4.12E+01e-1.60E-03x
R² = 8.94E-01
10
10
y = 1.67E+02e-1.40E-04x
R² = 9.57E-01
1 1
0 500 1,000 1,500 2,000 0 10,000 20,000 30,000
Time (t) days Cumulative Production (BARRELS)

Figure 3.4: Four diagnostic plots used to determine Arps’ exponential or harmonic decline.
47
Production Rates and Cumulative Production vs Time

120 30,000
100 25,000
Cumulative Production Np(Barrels)

80 20,000
60 15,000
40 10,000
20 5,000
0 0
0 200 400 600 800 1,000 1,200 1,400 1,600 1,800 2,000
Time t(days)
Actual Cumulative Arps' Hyperbolic Cumulative Actual Production Rate Arps' Hyperbolic Rate
Figure 3.5: Arps’ hyperbolic analysis showing cumulative production and rate match between the
actual and the modeled data.
Percentages of Arps' Decline Curve Models for Forty Profiles in the Eagle
Ford
6.45%
43.55% 29.03% Exponential Decline @ R^2=0.95

Harmonic Decline @ R^2=0.95
Hyperbolic 0<b<1
20.97%
b>1
Figure 3.6: Percentage application of Arps’ decline curve analyses for Forty Profiles in the Eagle
Ford.
48
Ratio of the Normal to the Abnormal Arps' Hyperbolic Decline

Exponent
32.5%
67.5%
Normal Decline exponent 0<b<1 Abnormal Decline exponent b>1
Figure 3.7: The percentage of abnormal decline exponents is more than twice that of the normal
decline exponents obtained from Arps’ hyperbolic model.
The fittingness of all the forty production profiles by Arps’ hyperbolic model resulted specifically
in 67.5% abnormal decline exponents (b>1) and 32.5% normal decline exponents 0<b<1) (Figure
3.7). Thus, regardless of the values of b, an excellent fit was observed for the forty production
profiles. The problem associated with b>1 becomes significant when future performance and
reserves are estimated. When b>1, the cumulative production becomes erroneously bigger than
expected for some of the wells depending on how much the b value deviates from normal (b=1).
3.2.3 Power Law Exponential Analysis (PLE)
Figure 3.8 and Figure 3.9 show PLE diagnostic plots for two wells in Leon County where I
eliminated data necessary to achieve R2≥0.97. Figure 3.10 illustrates the outcome of modeling
one of the forty production profiles using the PLE. The PLE, however, did not fit all the
production profiles of the forty wells in this fashion as one can see in Figure 3.11.
49
Rate vs Time
100
y = 30.731e-0.002x
Rate (Barrels)/Day)
R² = 0.9754
10
1
0 200 400 600 800 1000 1200 1400
Time (t) (days)
Actual Data PLE Analysis Plot Data points filtered out from analysis
Figure 3.8: PLE diagnostic plot for the Simms lease achieved R2≥0.97 only when initial and
erratic data sets were filtered out.
Rate vs Time
100
y = 11.841e-0.002x
R² = 0.9824
10
1
0 100 200 300 400 500 600 700 800 900 1000
Time (t) in days
Actual data used in analysis PLE Analysis Plot Data points filtered from analysis
Figure 3.9: PLE diagnostic plot for the Easterling well achieved R2≥0.97 only when erratic data
sets and no production sections were filtered out.
50

120 30,000
Cumulative Production (Barrels))

100 25,000
Rate (Barrels)/Day)
80 20,000
60 15,000
40 10,000
20 5,000
0 0
0 200 400 600 800 1,000 1,200 1,400 1,600 1,800 2,000
Time (t) (days)
Actual Production Rate PLE Model Plot Actual Cumulative Production PLE Cumulative
Figure 3.10: Power law exponential analysis showing cumulative production and rate match
between the actual and the modeled data for Giesenschlag W.H. well in Burleson County.

120 50,000
45,000
100
Cumulative Production (Barrels)
40,000
35,000
80
30,000
60 25,000
20,000
40
15,000
10,000
20
5,000
0 0
0 200 400 600 800 1,000 1,200 1,400 1,600
Time (t) in days
Actual Production Rate PLE Rate Actual Cumulative PLE Cumulative
Figure 3.11: Power law exponential analysis showing not very good cumulative production and
rate match between the actual and the modeled data for Hullabaloo well in Brazos County.
51
3.2.4 Logistic Growth Analysis (LGA)
The results of the logistic growth analyses (LGA) for forty oil cumulative productions have
closely fitted the past production profiles. Figure 3.12 is one example of the outcome of a good
fitting of the cumulative productions, as well as the corresponding production rates, generated
from the logistic growth analysis, to the actual production profile.
Production Rates and Cumulatives using Logistic growth analysis
100 30,000
25,000

Daily Oil Rate (Barrels/Day)
20,000
50 15,000
10,000
5,000
0 0
0 200 400 600 800 1,000 1,200 1,400 1,600 1,800 2,000
Time (t) in days
Actual Production Rate LGA Rate Actual Production Cumulative LGA Cumulative
Figure 3.12: Logistic growth analysis showing cumulative production and rate match between
the actual and the modeled data for Giesenschlag W.H. well in Burleson County.
52
3.2.5 Duong’s Decline Curve Analysis Method
The Duong’s method diagnostic plots determined the required parameters such as a (the intercept
constant, day-1), m (the slope), (Oil rate on day 1, stb) and (Oil rate at infinite time, stb)
for all forty wells investigated. In the diagnostic plot presented in Figure 3.13, the values of
a=13.533, m= 1.67, q1=0.0032 and q∞=11.432 have been defined by the equation of the line of
least square regression. Note that in this example, the value for q ∞ will be zero if the line passes
through the origin, but this is not the case for all forty production profiles, unless it is made to do
so by Excel tools.
Figure 3.13 (A, B): Duong’s method diagnostic plots for a production profile to determine a, m,
q1 and q∞. ( = Oil rate at Day 1, = Oil rate at infinite time, t= Production time, days, t (a, m)
= Time function based on Equation 2.17)
53
The outcomes of Duong’s method fitting the past production history are consistently better when
the value of q∞ is zero compared to when it is nonzero. In Figure 3.14 (A, B), the cumulative
production corresponding to q∞ = -1.3 does not fit the actual production history as much as it does
for Figure 3.14 (A, B), corresponding to q∞ = 0. Another example of a similar result is presented
in Figure 3.15 (A, B). There are other cases where no significant difference in the cumulative
trends between history and the Duong’s method profiles exist. The rates at infinite time are -0.1
and 0.0 (zero) for Figure 3.16 (A, B) respectively; i.e., both rates are approximately zero; hence
their cumulative production profiles are similar.
(B) Cumulative Production vs

(A) Cumulative Production vs
Time
Time
50000 50000
40000 40000
30000 30000
20000 20000
10000 10000
0 0
0 500 1000 1500 0 500 1000 1500
Actual Cumulative Production

Duong Model Actual Cumulative Production Duong Model
Figure 3.14 (A, B): Different results are obtained in fitting past cumulative production profile of
a well (Hullabaloo) with Duong’s method when the rate versus time function does or does not
pass through the origin.
54
(A) Cumulative Production vs (B) Cumulative Production

Time vs Time
40000 40000

30000 30000
20000
20000
10000
10000
0
0 100 200 300
0
0 100 200 300
Actual Cumulative Production Actual Cumulative Production

Duong Mode Plot Duong Model Plot
Figure 3.15 (A, B): Different results are obtained in fitting past cumulative production profile of
a well (Donaho Unit) with Duong’s method when rate versus the time function does or does not
go through the origin.
(A) Cumulative Production vs (B) Cumulative Production

Time vs Time
30000 30000
20000 20000
10000 10000
0 0
0 400 800 1200 1600 0 400 800 1200 1600
Actual Cumulative Production Actual Cumulative Production

Duong Model Plot Duong Model Plot
Figure 3.16 (A, B): No significant differences are obtained in fitting past cumulative production
profile of a well (Giesenchlag W.H.C) with Duong’s method when rate versus the time function
does or does not pass through the origin.
55
The results for an entire process of evaluating one of the wells (Cannon Unit) in the Karnes
County using Duong’s method are presented in Figure 3.17 (A, C). The model’s rate versus time
plot is seen overlying the actual data, but does not fit it very well (Figure 3.17 (A, C). In Figure
‎3.17 (B), the rate at infinity (q∞) is a nonzero digit of -27.663 stb/d. In the end, there is not a good
overlap between the actual cumulative production data and that of Duong’s method (Figure ‎3.17
(C)). Figure 3.18 is one example of the outcome of a fitting of the cumulative productions, as
well as the corresponding production rates, generated from the Duong’s method, to the actual
production profile.
56
Figure 3.17 (A-C): Duong’s method failed to match the production history for unfiltered Cannon
# 09607 lease.
57
Production Rates and Cumulatives using Duong's Method

100 30,000

25,000
80
Daily Oil Rate (Barrels/Day)
20,000
60
15,000
40
10,000
20
5,000
0 0
0 200 400 600 800 1,000 1,200 1,400 1,600 1,800 2,000
Time (t) in days
Actual Data Duong's Method Prouction Rate Actual Cumulative Production Duong's Method Cumulative
Figure 3.18: Duong’s method showing cumulative production and rate match between the actual
and the modeled data for Giesenschlag W.H. well in Burleson County.
3.2.6 Logistic Arps Hyperbolic Approach (LAHA)
When the combined real and extended LGA cumulative production data was fed into Arps’
hyperbolic, it improved the normal hyperbolic decline (0<b<1) from 32.5% to 80%, and thereby
reduced the number of b>1 from 67.5% to 20% (Figure 3.19 (A, B). Recall, Arps’ hyperbolic
relation achieves reasonable reserve estimate and future production forecast when the decline
exponent is within the range 0<b<1. Otherwise when b>1, reserve estimates and future
production forecasts become unrealistically optimistic.

58
(A) Normal and Abnormal (B) Normal and Abnormal

Arps' Hyperbolic Decline Arps' Hyperbolic Decline
Exponent Exponent Using LAHA
20%
32.5%
67.5% 80%
Normal Decline exponent 0<b<1 Normal Decline exponent 0<b<1

Abnormal Decline exponent b>1 Abnormal Decline exponent b>1
Figure 3.19 (A, B): Percentage of normal decline exponent (0<b<1) improves with the use of the
new approach (LAHA).
Figure 3.20 illustrates the effects of the proposed LAHA approach. The number of production
profiles with decline exponent 0<b<1 increased from 13 to 32, as presented in Figure 3.20 and
Appendix Table A.1. The blue and the red bars on Figure 3.20 correspond to the normal and
abnormal decline exponent values for all forty production profiles. The green bars represent the
normal decline exponents of previously abnormal decline exponents (i.e. red bars). They show
that LAHA has been successful in eliminating high decline exponents within a certain range of
b>1, such as b<1.5.

59
Arps' Hyperbolic and Logistic Hyperbolic Decline Exponent (b)

Comparison
5
Decline Exponent (b)
3
b=1 line
2
20
36
10
11
12
13
14
15
16
17
18
19
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
37
38
39
40
1
2
3
4
5
6
7
8
9
Production Profiles
Decline Decline LAHA (0<b<1)

exponent (0<b<1) exponent (b>1) Improved Decline Exponent
Figure 3.20: Comparison of decline exponents (b) using Arps' hyperbolic and LAHA.
In addition, LAHA lowered the EUR values wherever the decline exponents were shown to
improve from abnormal b>1 to normal 0<b<1. EUR results for Arps’ hyperbolic decline
exponents b>1 and LAHA decline exponents (0<b<1) are presented side by side in Figure 3.21
and Appendix Table A.2.

60
EUR Comparison between Arps' Hyperbolic and Logistic Arps'

Hyperbolic Approach (LAHA)
6E+5
8 1 1
9 . 6 1
. .
. 0 . .
7 6
5E+5 8 E 0 3
E E
E + E E
+ +
+ 6 + +
6 9
5 9 6
4E+5
EUR (Barrels)
3E+5
2E+5
1E+5
0E+0
18
31
10
11
12
13
14
15
16
17
19
20
21
22
23
24
25
26
27
28
29
30
32
33
34
35
36
37
38
39
40
1
2
3
4
5
6
7
8
9
Production Profiles
Decline Decline LAHA (0<b<1)

exponent (0<b<1) exponent (b>1) Improved Decline Exponent
Figure 3.21: Expected ultimate recoveries (EUR) generated by LAHA are expectedly lower than
that of Arps’ hyperbolic decline for cases where b>1 under Arps’ hyperbolic method became
0<b<1 under LAHA.
3.3 Discussion of General Results
The consequences of not filtering the data include unexpected results such as:
61
1. Duong’s method diagnostic plot for q1 and q∞ determination may fail to pass through the
origin, resulting in unmatched cumulative production
2. Arps’ hyperbolic decline may cause high incidence of decline exponent b>1 resulting in
wrong reserves future production forecasts or
3. The PLE may fail to achieve the best least square regression fit (R2>0.97) resulting in
abnormal rate as production time approaches infinity.
Arps’ hyperbolic decline model fits reasonably well for all forty past production rate profiles
regardless of the values of the decline exponent (b). The problem of abnormal decline exponent
(i.e., b>1) becomes significant when estimating future performance and reserves. Thus, the
consequences of an abnormal decline exponent are unrealistic overestimation of cumulative
production and infinite recoverable reserves. Mathematically, when b>1 is used for cumulative
production calculations, the result is infinite over times when the production rate approaches zero.
PLE estimated the least EUR and remaining reserves to abandonment for all forty profiles
analyzed. Compared to the other models, the general observation is that PLE is an augmented
Arps’ exponential model and provided pessimistic forecasts. A better fit is observed when a
trendline of R2>0.97 is passed through the late-time profile (after ignoring the early months’
production) of the new production data points.
The strength of LGA against the other models is the determination of the maximum possible
recoverable oil (K), which ensures that EUR does not grow infinitely with time. It is consistently
more conservative than the optimistic models such as Arps’ hyperbolic (includes harmonic) and
Duong’s method.
62
Duong’s method is particularly sensitive to the level of noise in the production trend data. Even
with filtered profiles, Duong’s model history-matched the cumulative shale oil production trends
the least, compared to the other decline models, for most of the forty production profiles. As a
consequence, the forecasted EUR, the remaining reserves, and time to abandonment are
sometimes extremely high or low compared to the values of other models. When Excel is used to
force the trendline to go through the origin, a slightly closer match is obtained. However, it gives
a lower least square regression, changes the curvature of the cumulative production trend, and
provides questionably high forecasts. The closer the rate at infinity (q ∞) is to zero, the better the
match between past and modeled cumulative production data.
Despite the problems associated with EUR when decline exponent b>1, Arps’ model is still
applied frequently in the industry. To maintain the useful tradition and conserve the original
parameters of decline analysis, I proposed and tested a new technique: Logistic Arps hyperbolic
approach (LAHA). In this approach, closely matched production profiles obtained from LGA
with reasonable EUR have been used to investigate Arps’ hyperbolic decline. The expectation is
to find a solution to the occurrence of erroneous b>1 and to eliminate the unrealistically high
EUR. The method has been successful in tackling both the abnormal decline exponent (b) and the
high EUR when b is greater than unity by up to 0.5. Decline exponents b>1.5 also show
significant improvement, but not enough to bring the decline exponent to normal (0<b<1).
63
Chapter 4 Eagle Ford Shale Oil Production Decline Analysis Specific Results
4.0 Introduction
This chapter discusses the performance history and the results of the decline curve analyses for
forty production profiles from the eight counties evaluated in Chapter 3. The completion details
of these wells are not publicly available from Texas Railroad Commission online data resource.
The historical performance and detailed results of the individual production profiles are discussed
county by county and, to some extent, well by well. The exact methods for application of the
different decline analysis techniques have already been discussed in Chapter 3. The detailed
results for the expected ultimate recoveries (EUR), remaining reserves, and time to abandonment
are tabulated and presented in Appendix Table A.2. The relative locations of the various counties
for the wells investigated are shown in the arrow labels on Figure 2.1 and 2.2.
4.1 Burleson County
4.1.1 Introduction
Three wells with long production histories were examined in this county located up-dip in the oil
generation window of the Eagle Ford Shale formation (Figure 2.2). The wells are identified by
their lease names: Giesenschlag-Groce, AB Childers and Giesenschlag W.H. Hydraulic fracture
stage and perforation cluster specifications are not available publicly at the Texas Railroad
Commission free webpage.
Well #2H in the Giesenschlag-Groce lease had fifty-four months of production data from
September 2008 to February 2013 (Table 4.1). The wellbore produced initially at a monthly
average rate of 216 barrels/day for 27 days in the first month, but declined to an average rate of
64
108 barrels/day by January 2009. A total of 65,298 barrels of oil was produced over the entire
production period between September 2008 and February 2013.
Well #3H from AB Childers well produced 28,000 barrels of oil within fifty-six months (August
2008-February 2013). Initial and latest rates were 66.5 and 8.4 barrels/day, respectively (Table
4.1).
The Giesenschlag W.H lease well #3H was completed in May 2008 and its production profile
available were fifty-eight months long as of February 2013. The latest production rate reported, as
of February 2013, was 3.6 barrels/day, compared to an initial rate of 96.8 barrels/day in May
2008. It had cumulative oil production of 28,134 barrels (Table 4.1). The past production rates
versus time for the three wells are presented in Figure 4.1.
65
Table 4.1: Burleson County wells’ completion and production history
(Barrels/day)
(Barrels/day)
Recompleted
Current rate
Completion
Cumulative
production
Number of
Horizontal
Initial rate
Producing
MD /TVD
length (ft)
Depth (ft)
until now
(Months)
(Barrels)
fracture
County
date &
stages
Name
Well
date
life
GIESENSCHLA
10890 /
TVD Not
W. H.
G
5/21/2008 (not 998 availabl 58 53.1 26489 3.6

available e
)
9214
CHILDER
Burleson
/TVD Not
AB
7/30/2008 (not 635 availabl 56 66.5 28000 8.4

S
available e
)
GIESENSCHLA
G-GROCE
Not
13558 /
8/3/2008 3640 availabl 54 216 65298 20.6
9016
e
4.1.2 Specific Results of Shale Oil Production Decline Analysis in Burleson County
The production rate profiles of all three wells in Burleson County are similar especially the
Giesenschlag W.C and AB Childer wells’ production declines (Figure 4.1).

66
History: Rate versus time for the 3 wells of Burleson County

250
200
150
Rate (stb/d)
100
50
0
0 200 400 600 800 1000 1200 1400 1600 1800
Time (Days)
Giesenschlag W.H AB Childers Giesenchlag Groce
Figure 4.1: Profile of Burleson County wells’ production history.
PLE analyses have predicted the least EUR, remaining reserves, and time to reach abandonment
for all three wells (Figures 4.2-4.4 and Appendix Table A.2). Of the three wells, Arps’
exponential decline curve analysis modeled only the AB Childers successfully, but the other
techniques analyzed and generated future production trends for all three wells. LGA production
performance predictions are always relatively less compared to Arps’ hyperbolic model over
predictions. The Duong’s method predictions are questionably high in the case of the AB
Childers well when compared to the other decline techniques. LAHA has successfully reduced
the original predictions of Arps’ hyperbolic decline curve analysis, especially for the AB Childers
and Giesenschlag Groce wells which had decline exponents greater than unity. LAHA improved
the Arps’ hyperbolic decline exponent from 1.01 to 0.93 and 1.12 to 0.91 for AB Childers and
Giesenschlag-Groce respectively (Appendix Table A.1).

67
EUR Forecasts Using Various Decline Analysis Methods for Three

Horizontal Wells' Production Profiles, Burleson County
120000
100000
EUR (Barrels of Oil)
80000
60000
40000
20000
0
GIESENSCHLAG AB GIESENSCHLAG-
W. H. " CHILDERS GROCE
Wells' Lease ID
PLE Arps Arps Arps LGA LAHA Duong

Exponential Harmonic Hyperbolic
Figure 4.2: EUR generated for three horizontal wells' production profiles Burleson County.
68
Remaining Reserves Forecasted to 2 Barrels/Day
70000
60000
50000
40000
Barrels of Oil
30000
20000
10000
0
Wells' Lease ID
PLE Arps Arps Arps Hyperbolic LGA LAHA Duong

Exponential Harmonic
Figure 4.3: Remaining reserves forecasted to 2 barrels/day for three Burleson county horizontal
wells' production profiles.
69
Remaining Production Time to Reach 2 Barrels/Day
14000
12000
10000
8000
Times (Days)
6000
4000
2000
0
Wells' Lease ID
PLE Arps Arps Arps LGA LAHA Duong

Figure 4.4: Remaining production time to reach 2 barrels/day for three Burleson county
horizontal wells' production profiles.
70
4.2 Leon County
4.2.1 Introduction
Leon County is set in District 05 of the Eagle Ford Shale and encloses the up-dip oil sector of the
play (Figures 2.1 and 2.2). It is located right above the Brazos County. Three horizontal wells’
production profiles were retrieved from the public domain of the TRRC and analyzed (Table
4.2).
The horizontal wellbore (#3H) of the Simms lease was completed in December 2009; it
comprises thirty-nine months of history as of February 2013. The well’s horizontal section is
between 7333 ft and 9141 ft measured depth (MD), and its production interval is between 7260 ft
and 8885 ft MD. It had an initial production rate of approximately 120 barrels/day for 10 days in
December 2009, which declined sharply to approximately 38 barrels/day in January 2010. The
rate as of February 2013 was roughly 4.0 barrels/day, and the cumulative oil produced was
18,244 barrels (Table 4.2).
The Easterling lease well #2H was completed and brought online in July 2010. It produced at the
rate of 14.5 barrels/day in August 2010, but as of February 2013, the rate had declined to 2.8
barrels/day (Table 4.2). Thirty-two months’ cumulative production as of February 2013 is only
4,929 barrels of oil. The records also indicate no production for the months of May-August 2012
and have late-time erratic production rate profiles indicating tempered operating conditions that
can introduce errors into the decline curve analysis.

71
The Donaho Unit wellbore was completed in May 2012 and had only 10 months of history data at
the time of this analysis. This well had an initial flowing rate of 188 barrels/day but has since
declined to 60.5 barrels/day, with a cumulative production of 34622 barrels (Table 4.2).
Table 4.2: Leon County wells’ completion and production history
Producing life
(Barrels/day)
(Barrels/day)
Recompleted
Current rate
Completion
Cumulative
production
Number of
Horizontal
Initial rate
MD /TVD
length (ft)
Depth (ft)
until now
(Months)
(Barrels)
fracture
County
date &
stages
Name
Well
date
SIMMS
9141 /
12/21/2009 1881 2 39 120 18244 4.0
6989
EASTERLING
9200 / Not
Leon
7/7/2010 1875 32 14.5 4929 2.8

7300 available
DONAHO
UNIT
13632 / Not
5/17/2012 4231 10 188 34622 60.5
7746 available
4.2.2 Specific Results of Shale Oil Production Decline Analysis in Leon County
The comparison between the past productions profiles of all three wells from Leon County is
illustrated in Figure 4.5. A continuous early-time section of the Simms well’s production profile
(29 of the 39 months’ data) was analyzed successfully with all the decline analysis techniques,
with the exception of Arps’ exponential model. The first two data points and later section
72
comprising seven months are very erratic and questionable to include in the analysis. Arps’
hyperbolic model calculated b=1.23 for the Simms well, which became b=0.89 with LAHA
(Appendix Table A.1). Forecasts of expected ultimate recoveries are very close, but Arps’
hyperbolic model predicts 3648 barrels more than Duong’s Method does (Figures 4.6-4.8 and
Appendix Table A.2). The overestimation by Arps’ hyperbolic approach may be due in part to
having a decline exponent greater than one.
History: Rate versus time for the 3 wells of Leon County

500
450
400
350
300
Rates (stb/d)
250
200
150
100
50
0
0 200 400 600 800 1000 1200
Time (Days)
Simms Easterling Donaho Unit
Figure 4.5: Profile of Leon County wells’ production history.
A section of the Easterling well’s production profile also analyses quite well with all of the
techniques except Arps’ exponential (Figures 4.6-4.8 and Appendix Table A.2). PLE gives the
73
lowest, hyperbolic the highest values of EUR and remaining reserves. The differences in the
forecasted expected ultimate recoveries are minimal, but Arps’ hyperbolic and Duong’s method
predictions are the most optimistic. The Arps’ hyperbolic decline exponent of the actual data was
b=1.42, resulting again in the highest future performance predictions. With LAHA it became
b=0.97 and the expected ultimate recovery became very close to that of LGA and Arps’ harmonic
models.
The Donaho Unit well’s profile analyzed very well with all the techniques except Arps’
exponential and the harmonic. The original decline exponent for Arps’ hyperbolic is normal
(b=0.5), but LAHA was still used (results shown in Figures 4.6-4.8 and Appendix Table A.1
and A.2). The PLE also matched the cumulative production, but the Duong’s method diagnostic
plot deviated slightly from the real data, resulting in high EUR compared to the rest of the models
(Figure 4.6). Arps’ exponential and the harmonic fell short of criteria (R2≥0.95) and so did not
apply.
74
EUR Forecasts Using Various Decline Analysis Methods for Three

Horizontal Wells' Production Profiles, Leon County
100000
90000
80000
70000
60000
50000
40000
30000
20000
10000
0
SIMMS EASTERLING DONAHO
UNIT
Wells' Lease ID

Figure 4.6: EUR generated for three horizontal wells' production profiles in Leon County.
75
60000
50000
40000
Barrels of Oil
30000
20000
10000
0
UNIT
Wells' Lease ID

Figure 4.7: Remaining reserve forecasts for three production profiles in Leon County.
76
5000
4000
Times (Days)
3000
2000
1000
0
UNIT
Wells' Lease ID

Figure 4.8: Remaining production life to reach 2 barrels/day for 3 wells’ in Leon County.
4.3 Brazos County
4.3.1 Introduction
Brazos County is located approximately 90 miles northwest of Houston and shares boundaries
with Leon and Burleson Counties on the up–dip segment of the Eagle Ford Shale oil zone. Two
production profiles corresponding to oil wells from the Reser-Sanders Unit and Hullabaloo leases
are posted online at the TRRC website. The Hullabaloo well has a historic cumulative production
of 54,481 barrels of oil for 52 months; it started production at the rate of 193.8 barrels/day, and as
of February 2013 it had declined to 13.5 barrels/day (Table 4.3 and Figure 4.9). The Reser-
Sanders Unit well started producing at a rate of about 23 barrels/day, which declined to 2.3
77
barrels/day in February 2013 (Table 4.3 and Figure 4.9). Cumulative oil production for 49
months is only 7,907 barrels.
Table 4.3: Brazos County wells’ completion and production history
fracture stages
Producing life
(Barrels/day)
(Barrels/day)
Recompleted
Current rate
Completion
Cumulative
production
Number of
Horizontal
Initial rate
MD /TVD
length (ft)
Depth (ft)
until now
(Months)
(Barrels)
County
date &
Name
Well
date
SANDERS
RESER-
Not
UNIT
15452 /
2/6/2009 3520 availa 49 22.9 7907 2.3
11763
-ble
Brazos
HULLABALOO
Not
12200 /
11/4/2008 2255 availa 52 193.8 54481 13.5
9518
-ble
4.3.2 Specific Results of Shale Oil Production Decline Analysis in Brazos County
The production rate profiles of these wells are dissimilar, as presented in Figure 4.9. Examination
of the two profiles suggested that the Hullabaloo well is located in a richer oil zone than the
Reser-Sanders Unit well (Table 4.3 and Figure 4.9). This observation is also supported by the
forecasts in Appendix Tables 2 and Figures 4.10-4.12.

78
History: Rate versus time for 2 wells, Brazos County

250
200
150
Rate (stb/d)
100
50
0
0 200 400 600 800 1,000 1,200 1,400 1,600 1,800
Time (days)
Reser-Sanders Unit Hullabaloo
Figure 4.9: Profile of Brazos County wells’ production history.
The Reser-Sanders Unit well’s production profile cannot be modeled with Arps’ exponential or
harmonic analyses. The smooth and seemingly undisturbed rate profile of the Hullabaloo well
was analyzed by all the decline curve analysis methods. Arps’ hyperbolic decline exponents are
b=1.45 for Hullabaloo and b=1.11 for Reser-Sanders. These decline exponents did not conform to
standards (0<b<1). Consequently, the EUR for Hullabaloo is overestimated above all the other
models’ predictions; specifically, it is 57,794 barrels more than the Duong’s Method prediction,
which is the second highest (Appendix Tables 2 and Figure 4.10). The effect of b=1.11 for
Reser-Sander is not obvious, especially with respect to the EUR (Figure 4.10). However, with
the application of LAHA, the decline exponent became normalized to b=0.99 and b=0.87 for the
Hullabaloo and Reser-Sanders wells, respectively. As a result, the predictions for both wells
79
improved greatly. The Hullabaloo well’s profiles indicated possible Arps’ exponential and
harmonic declines, but neither could be applied to the Reser-Sander Unit well. Arps’ exponential
and the PLE forecasted very close to each other. Duong’s method compares fairly well for
estimating the EUR for the Reser-Sanders Unit well; it is in its own class and not particularly
similar to any other model results for the Hullabaloo well. LAHA reduced the optimistic EUR of
152627 barrels of oil forecasted by Arps’ hyperbolic decline analysis to a more realistic 86418
barrels of oil.
EUR Forecasts Using Various Decline Analysis Methods for Two

Horizontal Wells' Production Profiles, Brazos County
100000
1
80000
5
2
6
60000 2
7
40000
20000
0
RESER- HULLABALOO
SANDERS
UNIT
Wells' Lease ID

Figure 4.10: EUR generated for two horizontal wells’ production profiles in Brazos County.
80
60000
50000
1
0
8
4
0
40000 2
Barrels of Oil
30000
20000
10000
0
RESER- HULLABALOO
SANDERS
UNIT
Wells' Lease ID

Figure 4.11: Remaining reserve forecasts for two production profiles in Brazos County.
81
14000
12000
10000
Times (Days)
8000
6000
4000
2000
0
RESER- HULLABALOO
SANDERS
UNIT
Wells' Lease ID

Figure 4.12: Remaining production life to reach 2 barrels/day for 2 wells in Brazos County.
4.4 Lee County
4.4.1 Introduction
Lee County is located adjacent to Burleson County. A horizontal well (E1H) from the Fenn
Ranch unit lease was completed in December 2008 and produced at initial rates of 82.8
barrels/day, which declined to approximately 16.1 barrels/day in February 2013. Cumulative oil
recovery was 46,957 barrels over the fifty-one month period. The horizontal production interval
is approximately 1331ft long and has nine hydraulic fracture stages (Table 4.4).
82
Table 4.4: Lee County wells’ completion and production history
(Barrels/day)
(Barrels/day)
Recompleted
Current rate
Completion
Cumulative
production
Number of
Horizontal
Initial rate
Producing
MD /TVD
length (ft)
Depth (ft)
until now
(Months)
(Barrels)
fracture
County
date &
stages
Name
Well
date
life
FENN
10325 /
Lee
RANCH 12/20/2008 1331 9 51 82.8 46957 16.1

8101
UNIT
4.4.2 Specific Results of Shale Oil Production Decline Analysis in Lee County
In terms of the EUR predictions, the PLE again forecasted the least volume of recoverable
reserves; LGA falls behind Duong’s method by 10,497 barrels, while Arps’ hyperbolic and
subsequent LAHA have extremely high predictions (Figures 4.13-4.15, Appendix Tables 2).
The Arps’ hyperbolic decline exponent b=3.01 resulted in one of the highest volumetric
cumulative predictions for all forty production profiles (Figures 4.13-4.15, Appendix Tables 2).
LAHA significantly lowered the EUR from the original forecast by Arps’ hyperbolic, but could
not normalize the high decline exponent of b=3.01.

83
EUR Forecasts Using Various Decline Analysis Methods for One

Horizontal Well Production Profiles, Lee County
1400000
1200000
1000000
800000
600000
400000
54052 101329 111826
200000
0
FENN RANCH UNIT
Well's Lease ID

Figure 4.13: EUR generated for one horizontal well’s production profile in Lee County.
1400000
1200000
Barrels of Oil
1000000
800000
600000
400000
82414 92911
200000 9877
0
FENN RANCH UNIT
Well's Lease ID
Figure 4.14: Remaining reserve forecasts for one production profile in Lee County.
84
350000
300000
4
250000
4 3
Times (Days)
9 3
200000 3 0
3 0
150000 1 2
5
100000
50000
1509 16561 12457
0
FENN RANCH UNIT
Well's Lease ID

Figure 4.15: Remaining production life to reach 2 barrels/day for one well in Lee County.
4.5 Gonzales County
4.5.1 Introduction
Gonzales County is located south of the up-dip northern section of the Eagle Ford Shale. Eleven
oil producing horizontal wells’ production profiles were retrieved for decline analysis. The wells’
lease names of the respective production profiles are Mostyn, Holmes, Georg, Parr, Bozka, S
Duderstadt, Cinco Ranch, Koenning Unit, Gonzo Hunter, Perkins, and Otto. Table 4.5 is a
summary of some of the vital statistics of these wells’ completion and production histories.
85
Table 4.5: Gonzales County wells’ completion and production history
until now (Months)

fracture stages
Producing life
(Barrels/day)
(Barrels/day)
Recompleted
Current rate
Completion
Cumulative
production
Number of
Horizontal
Initial rate
MD /TVD
length (ft)
Depth (ft)
(Barrels)
County
date &
Name
Well
date
Not Not Not Not Not Not
MOSTYN Not availa- availa avai- availa availa- Not availa-
available ble -ble lable -ble ble available ble
14155
HOLMES
6/15/2011 / 7217 7013 23 20 443.3 116014 61.8
10555
GEORG
5/27/2011 / 6724 3383 12 22 350 16194 1.4
13055
PARR
5/27/2011 / 7441 4975 17 22 490.2 19448 14.6
14240
BOZKA /
4/1/2011 10533 3146 11 23 351.3 41824 25.8
15890 Not
S.
/ avai-
Gonzales
DUDERSTADT
3/8/2011 11693 4240 lable 22 203.9 85671 44.8
15084
CINCO RANCH /
3/23/2011 10029 5066 15 24 359.3 48756 32.6
KOENNING
availa- availa avai- availa availa- Not availa-
UNIT
6/16/2010 ble -ble lable -ble ble available ble
Not
GONZO
13750 avai-
HUNTER
8/3/2010 / 9537 3290 lable 27 323.2 68385 28.3
PERKINS Not availa- availa avai- availa availa- Not availa-
available ble -ble lable -ble ble available ble
OTTO Not availa- availa avail availa availa- Not availa-
available ble -ble -able -ble ble available ble
86
4.5.2 Specific Results of Shale Oil Production Decline Analysis in Gonzales County
Figure 4.16 shows the variation in rate profiles of the various horizontal wells in this one county.
The various initial production rates, in particular, may be due to local differences in natural
fracture density and probable existence of sweet spots. Differences in shale petrophysical
properties (mineralogy, organic content and geomechanics) within the same formation may also
produce such variations. Koenning Unit and Mostyn wells’ production profiles started at the
lowest rates, while Parr, Perkins and Holmes had the highest initial rates. The rest of the wells
(Georg, Bozka, S Duderstadt, Cinco Ranch, Koenning Unit, Gonzo Hunter, Perkins, and Otto)
have initial production rates ranging between 320 and 360 barrels/day. The Gonzo Hunter and
Holmes production profiles are also very erratic.
History: Rate versus time for 11 wells of Gonzales County

600
500
Mostyn
Holmes
400
Georg
Rate (stb/d)
Parr
300 Bozka
S. Duderstadt
Cinco Ranch
200
Koenning Unit
Gonzo Hunter
100 Perkins
Otto
0
0 100 200 300 400 500 600 700 800 900
Time (Days)
Figure 4.16: Profile of Gonzales County wells’ production history.

87
Figures 4.17-4.19 and Appendix Table A.2 show the results of assessing the eleven production
profiles with the decline models. Analysis of these eleven wells provided the following
comments:
1. PLE forecasted in all cases the lowest EUR, remaining reserves, and time to reach
abandonment.
2. Arps’ exponential decline modeled Koenning Unit only, and it forecasted similarly to
PLE. Both Arps’ exponential and PLE calculations may underestimate the volumetric oil
in place.
3. The forecasts by Arps’ harmonic for five of the wells, where diagnostic plots have
R2≥0.95, are comparable to the other predictions except in the case of the Perkins lease
well, where the model predicts an unrealistically long time to reach 2 barrels/day (Figure
4.19). The cause of having such a high prediction of abandonment time is not clear from
the original data. The diagnostic verification for Arps’ harmonic decline analysis was
obtained for 21 out of 25 months of the production profile for Perkins.
4. The Arps’ hyperbolic decline exponents for the various production profiles are shown in
Appendix Table A.1. The cases where b>1 were the Mostyn, Bozka, S. Duderstadt,
Koenning Unit, Hunter Gonzo and Otto production profiles, and the EUR forecasted for
these cases soared above the other predictions.
5. The LGA model had decent history matches in all cases, therefore producing reasonable
and comparable predictions of EUR, remaining reserves, and time to reach 2 barrels/day.
6. Duong’s method did not have good history matches, but its forecasts were comparable to
the other models in most cases.
7. LAHA reduced the high values of the Arps’ hyperbolic when b>1. The corresponding b
values with LAHA are shown in Appendix Table A.1.

88
EUR Forecasts Using Various Decline Analysis Methods for Eleven

Horizontal Wells' Production Profiles, Gonzales County
200000
180000
160000
140000
120000
100000
80000
60000
40000
20000
0
MOSTYN
GEORG
BOZKA
RANCH
HOLMES
DUDERSTADT
PARR
KOENNING
HUNTER
OTTO
PERKINS
CINCO
GONZO
UNIT
S.
Wells' Lease ID

Figure 4.17: EUR generated for eleven horizontal wells’ production profiles in Gonzales County.
89
120000
100000
80000
Barrels of Oil
60000
40000
20000
0
MOSTYN
RANCH
GEORG
BOZKA
HOLMES
PARR
DUDERSTADT
HUNTER
KOENNING
OTTO
PERKINS
CINCO
GONZO
UNIT
S.
Wells' Lease ID
Figure 4.18: Remaining reserve forecasts for eleven production profiles in Gonzales County.
90
25000
4
2
5
4
2
20000 9
0
7
2
0
15000
Times (Days)
10000
5000
0
BOZKA
RANCH
MOSTYN
GEORG
HOLMES
PARR
DUDERSTADT
KOENNING
OTTO
HUNTER
PERKINS
CINCO
GONZO
UNIT
S.
Wells' Lease ID

Figure 4.19: Remaining production life to reach 2 barrels/day for 11 wells in Gonzales County.
91
4.6 Karnes County
4.6.1 Introduction
Karnes County is approximately 50 miles southeast of San Antonio in one of the prolific
hydrocarbon zones of the Eagle Ford Shale play. It has all three thermal maturity windows:
liquid, condensate and dry gas. Supposedly, wells producing more than 1000 barrels/day are
located here, but such information cannot be confirmed through the publicly available database
on the TRRC website.
Ten oil production profiles were retrieved from the TRRC website. The wells’ lease names are
Berry, Coates "A", Muenchow Unit, Cannon, Brysch Jonas, Kathryn Kealey, CEF, Gilley Unit,
Yosko Unit, and Carter Unit. Available wells’ completion and production data are summarized in
Table 4.6.
92
Table 4.6: Karnes County wells’ completion and production history
fracture stages
Producing life
(Barrels/day)
(Barrels/day)
Recompleted
Current rate
Completion
Cumulative
production
Number of
Horizontal
Initial rate
MD /TVD
length (ft)
Depth (ft)
until now
(Months)
(Barrels)
County
date &
Name
Well
date
16752 /
BERRY
12/5/2010 12078 4166 10 26 515.4 128040 13.1
15610 /
COATES
1/28/2011 9707 5550 20 26 317 110478 59.3
MUENCHOW 15767 /
UNIT 6/20/2011 10943 4397 12 20 502 45621 21
17105 /
CANNON
4/3/2011 11900 5120 n/a 24 507 179226 151
08/03/2011
BRYSCH
& 18540 /
JONAS UNIT
03/26/2012 12158 6141 20 19 567.8 120058 118.3
KATHRYN Not Not Not
Karnes
KEALEY Not availa- avail avai-

UNIT Available ble a-ble lable 24 324.5 125725 72.9
Not
CEF 16652 / avai-
6/28/2011 11095 5564 lable 20 547.7 150377 248.8
Not
GILLEY
16969 / avai-
UNIT
5/4/2011 12144 919 lable 22 56.4 28923 18.4
Not
YOSKO
17335 / avai-
UNIT
4/8/2011 12200 5010 lable 22 741.8 166501 94.6
Not
CARTER
16878 / avai-
UNIT
6/13/2011 5864 4343 lable 21 666 95695 110.3
4.6.2 Specific Results of Shale Oil Production Decline Analysis in Karnes County
Figure 4.20 shows a great disparity in the initial production rate for the various profiles for the
various horizontal wells in this county. With the exception of Yosko, Gilley, and Muenchow Unit
wells, the production profiles are very erratic.

93
History: Rate versus time for 10 wells, Karnes County

900
800
700
600 Berry
Coates
Muenchow Unit
Rates (stb/d)
500
Cannon
Jonas
400
Kathryn Kealey
CEF
300 Gilley Unit
Yosko Unit
Carter Unit
200
100
0
0 100 200 300 400 500 600 700 800
Time (Days)
Figure 4.20: Profile of Karnes County wells’ production history.
The results of the ten production profiles analyzed by the various decline models are presented in
Figures 4.21-4.23 and Appendix Table A.2. From these results the following comments can be
made:
1. The PLE did apply to the production profile of all wells’ production profiles in this
county at the R2≥ 0.97 acceptance level. The EUR and production forecasts are still low
compared to the other models.

94
2. Arps’ exponential decline modeled the Kathryn Kealey well’s production profile only. Its
forecast was not very different from that of the PLE.
3. Arps’ harmonic diagnostic proofs were obtained only for Muenchow, Kathryn Kealey
and Yosko Unit wells.
4. CEF well future production performance are extremely high with Arps’ hyperbolic
techniques because the decline exponent, b=3.05, is greater than unity. In this instance,
the PLE and LGA results seem reasonable to use because the EUR and production
forecasts generated by the other models are all erroneously high. A close look at the
original data suggests that the erratic nature of the production trend, which is difficult to
filter, may be the reason for the associated problems analyzing its decline and estimating
reserves.
5. The LGA model had very decent history matches in all cases, producing reasonable
values of EUR, remaining reserves, and time to reach 2 barrels/day compared to the other
models.
6. Duong’s method does not have good history matches, but its forecasts are comparable to
those of the other models in most cases. In some cases, positive values of q ∞, greater than
the abandonment rate of 2 barrels/day, predicted excessively long times to reach
abandonment.
7. LAHA adjusted the EUR obtained for Arps’ hyperbolic with abnormal decline exponents
that are by to 0.5 higher than b=1. LAHA could not sufficiently reduce decline exponent
values greater than b=1.5. The production trends of most of the wells (Cannon, Brysch
Jones, CEF, and Carter Units) do not follow reasonable decline trends; with or without
data filtering, neither Arps’ hyperbolic nor LAHA predict future performances
95
accurately. It is my guess that the production profiles reported for these wells individually
may instead have multiple producers.
EUR Forecasts Using Various Decline Analysis Methods for Ten

Horizontal Wells' Production Profiles, Karnes County
2000000
5
8 8 3 1 22
1800000 3
6 4 0 6 03
9
5 9 2 3 01
0
7 2 9 0 15
1600000 1
4 8 7 3 17
9
9 5 7 0 36
4
1400000 0 1 4 73
2
0 2
1 2
1200000
9
1000000
800000
600000
400000
200000
0
BERRY
MUENCHOW
CEF
BRYSCH
CANNON
COATES
YOSKO
CARTER
KATHRYN
GILLEY
JONAS
KEALEY
UNIT
UNIT
UNIT
UNIT
UNIT
UNIT
Wells' Lease ID

Figure 4.21: EUR generated for ten horizontal wells’ production profile in Karnes County.
96

500000
4 12 9 5
2 1 1
8 91 3 3
450000 8587
9 9 67 1
96 3 1
4038
5 4 36 0
95 4 1
9626
9 2 00 6
63 4 4
1773
5 1 15 5
38 4 6
400000 9737
4 1 57 7
36 9
400 7 36 0
1 2 4
6
5
4
350000 2
Barrels of Oil
300000
250000
200000
150000
100000
50000
0
BERRY
MUENCHOW
CEF
BRYSCH
CARTER
COATES
CANNON
KATHRYN
YOSKO
GILLEY
JONAS
KEALEY
UNIT
UNIT
UNIT
UNIT
UNIT
UNIT
Wells' Lease ID

Figure 4.22: Remaining reserve forecasts for ten production profiles in Karnes County.
97
50000
1 7 5 6 1 1
2 1 73
9 1 4 0 9 4
0 7 68
45000 5 0 8 3 1 0
0 7 32
2 3 2 5 1 0
8 1 89
6 6 0 5 7 9
9 6 57
3 1 9 1 9 9
2 4 2
40000 4 3 4 3
3
6 9
7
35000
Times (Days)
30000
25000
20000
15000
10000
5000
0
BERRY
CEF
MUENCHOW
BRYSCH
CANNON
COATES
YOSKO
CARTER
KATHRYN
GILLEY
JONAS
KEALEY
UNIT
UNIT
UNIT
UNIT
UNIT
UNIT
Wells' Lease ID

Figure 4.23: Remaining production life to reach 2 barrels/day for ten wells in Karnes County.
98
4.7 Dimmit County
4.7.1 Introduction
Located in the southern segment of the Eagle Ford Shale formation, Dimmit County is situated in
oil-prone and condensate-rich zones of the play. Production profiles of horizontal wells in the
Hutch, Voltz Unit, Fog Mount and JBGS leases are available for analysis from the TRRC public
webpage. Available statistics of these wells, including completion date, initial and current rates,
and cumulative oil production are presented in Table 4.7.
Table 4.7: Dimmit County wells’ completion and production history

fracture stages
Producing life
(Barrels/day)
(Barrels/day)
Recompleted
Current rate
Completion
Cumulative
production
Number of
Horizontal
Initial rate
MD /TVD
length (ft)
Depth (ft)
until now
(Months)
(Barrels)
County
date &
Name
Well
date
11405 /
HATCH 3/16/2010 3789 17 36 217.1 63219 22.4
7292
Not Not
VOTLZ 13597 /
12/25/2010 availa- availa- 24 120 33024 19
UNIT 6948
Dimmit
ble ble
FOG 09/06/2010 Not
12275 /
MOUN & 3145 availa- 26 99.9 35376 22.1
6058
T 01/15/2011 ble
Not Not
13820 /
JBGS 10/13/2010 availa- availa- 25 77.1 62743 33.5
6782
ble ble
4.7.2 Specific Results of Shale Oil Production Decline Analysis in Dimmit County
Figure 4.24 shows the rate profile histories for the four horizontal wells. The similarity of the
profiles for the Voltz and Fog Mount wells may be an indication of similar petrophysical
99
characteristics or completion technique. The JBGS well’s production profile is very erratic,
especially at early-time performance stages.
History: Rate versus time for 4 wells, Dimmit County

300
250
200
Rate (stb/d)
Hatch
150
Voltz
Fog Mount
JBGS
100
50
0
0 200 400 600 800 1000 1200
Time (Days)
Figure 4.24: Profile of Dimmit County wells’ production history.
PLE results are the least for each lease; Arps’ hyperbolic and Duong’s method share the highest
predicted future performance and reserves (Figures 4.25-4.27, Appendix Tables 2). Arps’
hyperbolic projected high EUR, remaining reserves and abandonment time for the Fog Mount
well, due to excess b=1.76 >1. Arps’ exponential cannot model any of the production profiles,
and Arps’ harmonic was applied to the Hutch lease well only.
100
EUR Forecasts Using Various Decline Analysis Methods for Four

Horizontal Wells' Production Profiles, Dimmit County
150000
125000
100000
75000
50000
25000
0
MOUNT
HATCH
JBGS
VOTLZ
UNIT
FOG
Wells' Lease ID

Figure 4.25: EUR generated for four horizontal wells’ production profiles in Dimmit County.
101
200000
180000
160000
140000
Barrels of Oil
120000
100000
80000
60000
40000
20000
0
MOUNT
HATCH
JBGS
VOTLZ
UNIT
FOG
Wells' Lease ID

Figure 4.26: Remaining reserve forecasts for four production profiles in Dimmit County.
102
50000
45000
40000
35000
30000
Times (Days)
25000
20000
15000
10000
5000
MOUNT
HATCH
JBGS
VOTLZ
UNIT
FOG
Wells' Lease ID

Figure 4.27: Remaining production life to reach 2 barrels/day for four wells in Dimmit County.
4.8 Zavala County
4.8.1 Introduction
Zavala County shares boundaries with Dimmit County. It is situated mainly in the oil generation
thermal window of the Eagle Ford play. A total of six wells’ production profiles were
investigated. The wells’ lease names are K.M. Ranch, Mustang Ranch C, Alpha Ware, Felps,
Howett and Addison Avery. Available information about these wells is summarized in Table 4.8
below.
103
Table 4.8: Zavala County wells’ completion and production history
Producing life
(Barrels/day)
(Barrels/day)
Recompleted
Current rate
Completion
Cumulative
production
Number of
Horizontal
Initial rate
MD /TVD
length (ft)
Depth (ft)
until now
(Months)
(Barrels)
fracture
County
date &
stages
Name
Well
date
6/27/2011 Not
K.M 12627 /
& 5758 avail 20 577 55503 91.5
RANCH 6587
08/08/2012 a-ble
MUSTANG 11684 / 117.
6/30/2010 5824 22 29 57848 55.8
RANCH C 5573 2
ALPHA 11800 / 247.
7/5/2011 5458 21 20 42606 34.2
Zavala
WARE 5693 2
9723 / 111.
FELPS 6/1/2011 4717 15 21 41336 44.8
5616 8
Not
11007 / 101.
HOWETT 10/16/2011 3870 avail 17 52740 51.8
6535 1
a-ble
ADDISON 9435 /
4/27/2011 4162 12 26 93.5 30768 21.1
AVERY 5290
4.8.2 Specific Results of Shale Oil Production Decline Analysis in Zavala County
Figure 4.28 shows the rate profiles of all six wells. The K.M Ranch shows two episodes of high
initial production, most likely due to recompletion of the well. The Alpha Ware, Mustang Ranch,
Felps, and Addison Avery production profiles show established sections of reasonable decline
trends.
Six production profiles assessed with the decline models have their results presented in Figures
4.29-4.31 and Appendix Table A.2. From the analyses, the following comments can be made:
1. PLE predictions again forecasted the lowest EUR, remaining reserves, and time to reach
2 barrels/day in all six cases.

104
History: Rate versus time for Six wells, Zavala County

600
500
400
K.M Ranch
Rate (stb/d)
Mustang Ranch
300
Alpha Ware
Felps
Howett
200
Addison Avery
100
0
0 100 200 300 400 500 600 700 800 900
Time (Days)
Figure 4.28: Profile of Zavala County wells’ production history.
2. Arps’ exponential decline modeling of Addison Avery well’s production profile
calculated a similar EUR value to those calculated by the PLE, hyperbolic, and the LGA
models.
3. Arps’ harmonic predictions apply to the K.M Ranch and Addison Avery wells only. It
estimated results comparable to those of the other models.
4. Two severe cases of high predictions were observed for the Mustang Ranch C and Alpha
Ware wells due to high decline exponent values of b=4.50 and b=1.99, respectively. This
could come as a result of changing operating conditions. The profiles are difficult to filter
without eliminating all data.

105
5. Duong’s method has fairly high production predictions for the Felps, Howett, and
Addison Avery wells, because the diagnostic plots of Equation 3.10 fail to pass through
the origins.
6. LAHA reduced the exaggerated values of Arps’ hyperbolic decline model when b>1.
EUR Forecasts Using Various Decline Analysis Methods for Six

Horizontal Wells' Production Profiles, Zavala County
400000
5 2
350000 9 8
1 1 5
8 7
3 7 7
1 2
300000 8 0 3
2 6
4 8 7
5 9
5 5 3
6 6
250000 5 6 4
2 6
6 2 6
8 2
4
200000 2 6
8
150000
100000
50000
0
RANCH
MUSTANG
ALPHA
HOWETT
ADDISON
WARE
FELPS
RANCH C
AVERY
K.M
Well Lease ID

Figure 4.29: EUR generated for six horizontal wells’ production profiles in Zavala County.
106
300000
1 2
3 8
8 7
250000 5 4 2
9 61 5
0
8 46 7
6
1 75 1
8
2 89 2
5
Barrels of Oil
200000 0 37 7
8
7 67 8
4 3
2
9
2
150000
100000
50000
0
RANCH
MUSTANG
HOWETT
ALPHA
ADDISON
WARE
FELPS
RANCH C
AVERY
K.M
Well Lease ID

Figure 4.30: Remaining reserve forecasts for six production profiles in Zavala County.
107
30000
1 3
2 31 0
6 7
3 51 6 7 0
2 40 8 2 6
6 93 2 2 7
1 51 3 0
8 61 8
6 49
20000 7 23
6 0
9 7
Times (Days)
10000
0
RANCH
MUSTANG
HOWETT
ALPHA
ADDISON
WARE
FELPS
RANCH C
AVERY
K.M
Well Lease ID

Figure 4.31: Remaining production life to reach 2 barrels/day for six wells in Zavala County.
108
4.9 Chapter Summary
The observations in this chapter about the individual wells and their respective counties do not
suggest any commonality for the production performance of the horizontal wells. The production
data analyses results are very unique for each well. A few of the wells in the same county have
similar past production trends; this is a good indication for similar shale mineralogy and
completion design that can only be confirmed if petrophysical data and completion statistics of
the wells are available. The expected ultimate recoveries, remaining reserves and the lifespan of
the individual wells indicate wide variability of shale oil distribution within the play. The
average initial rates for producing hydraulically fractured horizontal wells show vast disparity
from county to county. The Karnes county wells have consistently reported the highest initial
production rates for its individual producing wells and may be identified as the most liquid
prolific county. For the other counties, the average initial rates are relatively low due to huge rate
gaps between the individual wells. These same trends are noticeable in the overall estimates for
total recoverable reserves, which indicate the relative liquid hydrocarbon richness among the
counties.
109
Chapter 5 Shale Oil Reservoir Simulations
5.0 Introduction
Numerical Simulation Model (NSM) analysis of all forty horizontal wells could not be done due
to insufficient data available, since the reliability of simulation output depends on quality and
quantity of wells’ and reservoirs’ input data. As a consequence, only two Eagle Ford Shale (EFS)
wells were simulated for production performance history matching and cumulative production
forecast. Future productions forecasted by NSM were compared with the analytical results for
these two wells. The chapter also included enhanced shale oil recovery (ESOR) using different
configurations of producers and injectors with the reservoir model and parameters of one of the
wells that was history matched. The ESOR covers the miscible gas injection displacement
processes by CO2, 1:1 mixture of CO2-CH4, and an enriched gas mixture. The injectant methane
(CH4) was also initially considered, but due to low reservoir pressure, it turned out rather as an
immiscible gas displacement process.
History matching the production data for each well began with reservoir parameters modification
within an acceptable range of values reported in the literature (Chaudhary, 2011 and Orangi et al.,
2011). The properties modified include fracture density, porosity, and permeability, matrix
porosity and permeability, relative permeability curve end points and Corey exponents, critical
and residual saturations, and initial liquid saturations and pressures. The early part of each well’s
performance was dominated by fracture fluid loading and linear flow regimes, so the remarkably
sensitive parameters of primary fracture geometry, porosity, and permeability were modified in
order to model and history-match the initial high oil rate. On the other hand, the matrix
parameters were adjusted until a match was obtained for the later low rates. It should be noted
110
that the use of generalized reservoir parameters from the literature and their further adjustment,
during the history matching for the very short production periods of the EFS, subject the
outcomes to high degrees of uncertainty.
The simulation also covers assessment of pressure waves and hydrocarbon fluids withdrawal
from non-stimulated reservoir lease zones adjacent to stimulated reservoir areas. Assuming any
shale oil reservoir, pressure depletion was simulated to investigate the extent hydrocarbon fluids
withdrawal goes beyond the geometric confines of the simulated reservoir volume.
5.1 Methods: Production Decline and Future Performance Simulations
5.1.1 Building the Reservoir Simulation Models
Two single-well simulation models were built using the data provided earlier (Table 2.4) using
the Gilman and Kazemi dual permeability option in CMG (GEM) software. Initial reservoir
model parameters were based on the knowledge of 3-10% range and an average of 6% porosity,
matrix permeability, water saturation, compressibility, reservoir temperature and pressure values
quoted in Tables 2.1 through 2.3. Eagle Ford Shale synthetic crude compositional model,
developed based on the Peng-Robinson equation of state (EOS) by Orangi et al. (2011), was
incorporated into CMG Winprop. The gas-oil ratio (GOR) is 500 scf/stb and API gravity of 41o
for the compositional model. Subsequently, Two-phase P-T diagram and pressure-dependent
phase properties were generated.
Formation volume factor (FVF) of the Eagle Ford Shale, as in Figure 2.5, provided by Orangi et
al., (2011) was also used to perform reservoir fluid analysis using Winprop. Corresponding
pressure-volume-temperature (PVT) values and diagrams at initial composition were generated

111
by CMG software. A compositional model was used, meaning that the simulator calculated all
phase properties at any pressure for any cell using the given initial oil/gas phase composition.
Using Winprop, the minimum miscibility pressure (MMP) for multiple contacts miscibility
(MCM) for CO2 gas, methane, CO2-CH4 gas mixtures, and enriched gas mixtures that would be
injected for miscible displacement of EFS oil were analyzed.
The tuned-up EOS model of the EFS crude was incorporated into the Builder-GEM, CMG
Compositional Software, to build the individual single-well, dual-permeability models needed to
simulate horizontal wells of the Giesenschlag-Groce and Hutch leases.
Stimulated reservoirs of two horizontal wells were modeled using researched data summarized
below in Tables 5.1 and 5.2. Table 5.2 includes the initial and final reservoir model parameters,
literature reviewed and history matched respectively. These two wells, Giesenschlag-Groce and
Hutch, have been previously analyzed with empirical decline models in Chapter 4. They have
been stimulated with 18 and 17 perforations respectively. The models have initial water saturation
of 30%. The relative permeability parameters used were provided by Chaudhary (2011) and
Orangi et al. (2011) presented in Chapter 2 (Tables 2.4 and 2.5). By sensitivity analysis, the
reservoir model parameters were modified until the wells’ production histories were matched.
History matching, future production trends, and potential miscible gas injection displacement for
enhanced shale oil recovery (ESOR) were modeled and simulated using the Hutch well past
performance data. The Giesenschlag-Groce well was only simulated for history matching and
generating future production curves. The wells’ numerical simulation models estimated future
112
production to abandonment at 2 stb/d and these were compared with the analytical decline models
results in Chapter 4.
Table 5.1: Reservoir fluids, depths, lateral lengths and stimulation data (from www.rrc.state.tx.us
June 10, 2012)
County/ Top Depth Height Lateral Fracture Thickness Fracture GOR API Acres
Well (TVD) (TVD) (ft) Length Interval Interval Stages (scf/stb)
(ft) (ft) (ft) (MD) (ft)
Burleson
Giesenschlag 9780- Not
-Groce 8877 9016 139 3640 13420 8877-9016 available 510 42.1 459
Dimmit 7616- Not
Hatch 7145 7292 147 3694 11310 available 17 506 36 3489
Table 5.2: Initial and final reservoir model parameters of the Eagle Ford Shale (EFS)
113
Dimmit County
Reservoir and Hydraulic Fracture Burleson County Well Well
Parameters
Initial Final Initial Final
Matrix porosity (fraction)- Variable 0.1 0.04 0.1 0.05
Natural fracture porosity (fraction)- Variable 0.001 0.001 0.001 0.001
Shale Matrix Absolute Permeability (md) by
History Matching- Variable 0.0001 0.00009 0.0001 0.0001
Natural Fracture Permeability (md) by History
Matching- Variable 0.001 0.0004 0.001 0.0007
Hydraulic Fracture Permeability (md)
calculated from given conductivity 83300 83300
Hydraulic Fracture Spacing (ft) 200 220
Hydraulic Fracture Half Length (ft) by 500 500
Hydraulic Fracture Width (ft) 0.001 0.001
Hydraulic Fracture Conductivity (md-ft) 83.3 83.3
Hydraulic Fracture Stages 18 17
Reservoir Thickness (ft) from TRRC website 139 147
Shale Compressibility (1/psi) 5.00E-06 5.00E-06
Initial Water Saturation 0.3 0.3
Initial Reservoir Pressure (psi) 3800 3200
Reservoir Temperature (deg. F) 241 212
5.1.2 The Reservoir Simulation Initialization Outputs
A two-phase P-T (pressure-temperature) diagram was obtained for the compositional model of
the Eagle Ford synthetic crude oil and presented in Figure 5.1. It shows the critical point where
oil and gas phases coexist in equilibrium, and the enclosure of the 2-phase boundary is the phase
envelope, where proportionate amounts of oil and gas mixtures exist. The critical pressure and
temperature are 1500 psia and 800 oF respectively. The cricondentherm and the cricondenbar are
895oF and 2280 psia, respectively.

114
P-T Envelope for synthetic Eagle Ford shale oil

2500
2250
2000
Pressure (psia)
1750
1500
1250
1000
750
500
250
0
0 100 200 300 400 500 600 700 800 900 1000
Temperature (deg F)
2-Phase boundary Critical Point
Figure 5.1: Two-Phase P-T envelope of the synthetic Eagle Ford Shale oil.
Winprop also analyzed and generated other pressure-temperature dependent phase properties
associated with the analysis of the synthetic Eagle Ford Shale (EFS) crude include liquid and
vapor (gas) phase and are presented in Figures 5.2- 5.5. These include liquid and vapor volumes
versus pressure curves Figures 5.2 and 5.3, illustrating liquid and vapor saturations before and
after the bubble point pressure of 2054.85 psia. Also, Figure 5.4 is the plot of gas compressibility
with pressure showing increasing gas compressibility at reduced pressures below the saturation
pressure of 2054.85 psia. Finally, Figure 5.5 is the vapor liquid equilibrium (VLE) ratio
alterations with pressure for the different components of the EFS synthetic crude. The
compositions shown in the legend are those of the EFS synthetic crude (Table 2.4). Equilibrium
ratios are very important in compositional model reservoir simulations for the calculation of the
vapor and liquid equilibrium composition at various reservoir pressure-temperature (PT) flash
115
conditions. Figure 5.5 separates the highly volatile and the non-volatile components of the EFS
synthetic crude. It also helps to explain the preference of CO2, CH4 and the enriched gas mixtures
as the best miscible displacement injectant in enhanced shale oil recovery.
Liquid Volume vs. Pressure

Liquid Phase Volume %
90
70
50
30
10
0 1000 2000 3000 4000 5000 6000
Pressure (psia)
237 degrees F
Figure 5.2: Liquid phase volume fraction versus pressure.
116
Vapor Volume vs. Pressure

Vapor Phase Volume %
80
60
40
20
0
0 1000 2000 3000 4000 5000 6000
Pressure (psia)
237 degrees F
Figure 5.3: Vapor phase volume fraction versus pressure.
Gas Compressibility (Z) vs. Pressure

1.00
Vapor Z-Factor
0.80
0.60
0.40
0.20
0.00
0 1000 2000 3000 4000 5000 6000
Pressure (psia)
237 degree F
Figure 5.4:Vapor Z-Factor versus pressure.
117
Equilibrium ratios of the EFS synthetic crude compositions

1.0E+3
CO2
1.0E+2
N2
1.0E+1
K value @ temperature 237 0F
CH4
1.0E+0
C2H6
1.0E-1 C3H8
1.0E-2 IC4
1.0E-3 NC4
1.0E-4 IC5
1.0E-5 NC5
FC6
1.0E-6
C7+
1.0E-7
C11+
1.0E-8
C15+
0 500 1000 1500 2000
Pressure (psia) C20+
Figure 5.5: Equilibrium ratio (K Value=vapor/liquid) versus pressure.
Further outputs of the Eagle Ford Shale (EFS) synthetic crude fluid analysis using Winprop
include PVT tables and figures presented in Figures 5.6 through 5.9. The highlighted values are
the bubble point parameters of the EFS synthetic crude compositions. These PVT data analysis
results illustrate the regression between data provided by Orangi et al., (2011) and Winprop
simulation.
118
Final Exp.
Pressure GOR Pressure GOR
(psia) scf/stb (psia) scf/stb
6000 504.637 6000 500
3500 504.637 Solution Gas-Oil Ratio (GOR, Rs) vs. Pressure

3000 504.637 600
GOR (Rs) scf/stb

2500 504.637 500
2054.85 504.637
400
2000 492.671
1800 449.983 300
1600 408.64 200
1400 368.538 100
1200 329.553 0
1000 291.52 0 2000 4000 6000
750 244.761
Pressure (psi)
500 197.916
250 146.704
Exp. GOR Final GOR
14.7 0
Figure 5.6: Eagle Ford Shale synthetic crude solution gas-oil ratio with pressure.
Oil Formation Volume Factor

Final (FVF, Bo) vs. Pressure
Formation Volume Factor- FVF (rb/stb)
FVF Experimental 1.40

Pressure rb/stb Pressure FVF rb/stb 1.35
(psia) (Bo) (psia) (Bo)
6000 1.30 6000 1.3534 1.30
3500 1.34 3500 1.366 1.25
3000 1.35 3000 1.37 1.20
2500 1.37 2500 1.374
2054.85 1.377 2000 1.38 1.15
2000 1.37 1800 1.355 1.10
1800 1.35 1600 1.335 1.05
1600 1.33 1400 1.31
1400 1.31 1200 1.29 1.00
1200 1.29 1000 1.265 0 2000 4000 6000
1000 1.27 750 1.23 Pressure (psi)
750 1.25 500 1.2
500 1.22 250 1.16 Experimental Final FVF (Bo)
250 1.19 14.7 1.05 FVF (Bo)
14.7 1.05
Figure 5.7: Eagle Ford synthetic crude oil formation volume factor with pressure.
119
Oil
Pressure Viscosity Oil Viscosity vs. Pressure
(psia) (cp) 0.140
6000 0.132544 0.135
3500 0.116668 0.130
Viscosity (cp)
3000 0.11307
0.125
2500 0.109293
2054.85 0.105758 0.120
2000 0.106071 0.115
1800 0.107182 0.110
1600 0.108247 0.105
1400 0.109246 0.100
1200 0.110155 0 2000 4000 6000
1000 0.110942 Pressure (psi)
750 0.111692
500 0.112059
Oil Viscosity
250 0.111756
(cp)
14.7 0.108883
Figure 5.8: Eagle Ford synthetic crude oil viscosity versus pressure at reservoir temperature.
Pressure Oil GOR Z- Gas Bt

Formation Volume Factor
(psia) FVF (Rs) Factor FVF (rb/
(Bo, Bt) vs. Pressure
(Bo) scf/stb Bg stb)
rb/stb (rb/scf)
3.00
Formation volume factor (FVF) rb/stb
2.80
6000 1.30 504.64 1.055 0.00020 1.30
2.60
3500 1.34 504.64 0.891 0.00030 1.34
3000 1.35 504.64 0.875 0.00034 1.35 2.40
2500 1.37 504.64 0.867 0.00041 1.37 2.20
2054.85 1.38 504.64 0.868 0.00050 1.38 2.00
2000 1.37 492.67 0.868 0.00051 1.38 1.80
1800 1.35 449.98 0.873 0.00057 1.38 1.60
1600 1.33 408.64 0.878 0.00065 1.39 1.40
1400 1.31 368.54 0.885 0.00075 1.41 1.20
1200 1.29 329.55 0.893 0.00088 1.45 1.00
1000 1.27 291.52 0.902 0.00107 1.50 0 1500 3000 4500 6000
750 1.25 244.76 0.915 0.00145 1.63 Pressure (psi)
500 1.22 197.92 0.930 0.00222 1.90 Oil FVF (Bo)
250 1.19 146.70 0.948 0.00452 2.81
Total FVF (Bt)
14.7 1.05 0 0.983 0.07984 41.34
Figure 5.9: Eagle Ford synthetic crude oil two-phase formation volume factor with pressure.
120
After regression using Winprop, the calculated saturation pressure for the oil of GOR 500 scf/stb
is 2054.85 psia at 237oF; it is comparable to Orangi et al.’s (2011) experimental value of 2053
psia (Table 2.5). Also in this simulation, Winprop analyzed minimum miscibility pressure
(MMP) for CO2 gas and reservoir fluids, which was achieved at 3111 psia by the vaporizing gas
drive phenomenon explained earlier in the literature review (Holm, 1986 and Rahmatabadi,
2011). Similarly, miscibility for 100% CH4 was achieved at 6234 psia, which is higher than the
reservoir pressure of 3200 psia and, therefore, excludes its consideration from the miscible gas
displacement process in this reservoir. Two miscibility pressures, multiple-contact-miscibility
(MCM) and first-contact-miscibility (FCM) pressures for a 1:1 CO2-CH4 gas mixture were
achieved at 3111 psia and 3210 psia, respectively, by vaporizing gas drive. Finally, an enriched
gas mixture (arbitrarily composed using one of the CMG user manual’s illustrations), comprising
0.1% CO2, 2.1% N2, 86.12% CH4, 5.9% C2H6, 3.6% C3H8, 1.7% i-C4, and 0.5% i-C5, attained
MCM at 2964 psia. Thus, a summary of the minimum miscibility pressures obtained from
Winprop analysis is provided in Table 5.3:
Table 5.3: Miscible gas injectants and multiple contact miscibility pressures
Gas Injected Multiple Contact Minimum miscibility pressure (psia)
CO2 3111
CH4 6246
CO2-CH4 3111
Enriched Gas 2964
The final oil-water and oil-gas relative permeability data and curves generated by Builder-CMG
reservoir simulation software-are provided in Tables 5.4(a-b) and Figures 5.10 – 5.12.
121
Note: Sw, Sg, Sl, Krw, Krow, Krg and Krog are petroleum engineering standard abbreviations
associated with reservoir fluids saturations and relative permeability, which are defined in the
nomenclature.
Table 5.4 (a-b): Relative permeability tables generated using CMG Builder.
Sw Krw Krow Sg Krg Krog
0.30 0.00 0.35 0.05 0.00 0.35
0.33 0.00 0.27 0.07 0.00 0.27
0.35 0.00 0.21 0.09 0.01 0.21
0.38 0.00 0.15 0.12 0.01 0.15
0.41 0.00 0.11 0.14 0.03 0.11
0.43 0.00 0.07 0.16 0.04 0.08
0.46 0.01 0.05 0.18 0.06 0.05
0.49 0.02 0.03 0.20 0.08 0.04
0.52 0.03 0.02 0.23 0.10 0.02
0.55 0.05 0.01 0.25 0.13 0.01
0.58 0.08 0.01 0.27 0.16 0.01
0.60 0.11 0.00 0.29 0.19 0.00
0.63 0.15 0.00 0.31 0.23 0.00
0.66 0.19 0.00 0.33 0.26 0.00
0.69 0.25 0.00 0.36 0.31 0.00
0.72 0.32 0.00 0.38 0.35 0.00
0.74 0.40 0.00 0.40 0.40 0.00
0.77 0.49 0.00 (b)
0.80 0.60 0.00
(a)
122
Relative Permeability Curves

0.6
kr-Relative Permeability
0.5
0.4
0.3
0.2
0.1
0
0.3 0.4 0.5 0.6 0.7 0.8
Sw
Krw Krow
Figure 5.10: Relative permeability curves krw & krow vs Sw plotted by using Table 5.4.
Relative Permeability Curves

0.4
0.3
0.2
0.1
0
0.05 0.12 0.19 0.26 0.33 0.4
Sg
krog krg
Figure 5.11: Relative permeability curves- krg & krog vs. Sg plotted by using Table 5.4.
123
Relative permeability Curves

0.4
0.3
0.2
0.1
0
0.6 0.67 0.74 0.81 0.88 0.95
Sl
krg krog
Figure 5.12: Relative permeability curves krg & krog vs. Sl plotted by using Table 5.4.
The base model of the Giesenchlag-Groce single horizontal well has Cartesian grid dimensions of
18x5x1 for a total of 90 grid block cells (Figure 5.13). Grid block thicknesses of 200 ft x 200 ft x
139 ft result in a modeled volume of 5.0x108 ft3. It includes eighteen hydraulic fracture stages of
200 ft equidistant spacing. The base grid only captures the hydraulic fracture stages without
taking into consideration shale petrophysical variations and permeability anisotropy. However, in
order to capture the effects of near wellbore and fracture large pressure depletion and saturation
changes, smaller size grid blocks were introduced near the wellbore using CMG Builder local
grid refinement tool (Chaudhary, 2011). As a result, the total grid presented in Figure 5.13 is
2484 cells. The calculated original oil in place (OOIP) is 3.11x106 STB.
124
Shale Oil Rerservoir Model

Porosity 2008-09-01 K layer: 1
File: CMGBuilder_A-Z_Hstory M
1,000
0 1,000 2,000 3,000 User: CO2
Date: 7/25/2013
Scale: 1:6777
Y/X: 1.00:1
Axis Units: ft
Giesenschlag-Groce well in Burleson County

0.1
3600ft production interval 0.1
0
0
0.1
0.1
Well-1
0.1
0.1
-1,000
-1,000
0.1
Horizontal Induced 0.1
500ft Base and Local 200ft wellbore fractures
fracture fracture 0.1
refinement grids
half stage 0.00 435.00 870.00 feet 0.1
length interval 0.00 135.00 270.00 meters
0.1
0 1,000 2,000 3,000
Figure 5.13: Base reservoir model of the Burleson County lease well.
The Dimmit County well (Hutch) was modeled with Cartesian grid dimensions of 221x5x1 for a
total of 1105 grid block cells Figures 5.14 and 5.15. Grid block thicknesses of 17 ft x 200 ft x
147 ft resulted in a modeled volume of 2.3x108 ft3. It has 17 hydraulic fracture stages and a total
of 3366 grid cells including the effects of local grid refinement capture the effects of near
wellbore and fracture large pressure depletion and saturation changes. The calculated OOIP is
2.75x106 STB.
The homogenous reservoir model with abrupt top and bottom is not representative of the EFS. To
accurately represent the EFS will require importing geologic tops and bottoms (the horizons) into
125
the model which are not available at the time of research. Previous EFS models used a similar
homogenous system to investigate fracture conductivity; unlike in this case that it was used to
history match very short past production performance. The results of the reservoir modeling and
simulation investigate the technical feasibility of the concept and may be subject to reservations.
Hutch well in Dimmit County

3740ft producing interval
221ft
Horizontal Induced
fracture
wellbore fractures
stage
interval
Figure 5.14: Base reservoir model of the Dimmit County well-Hutch.
Figure 5.15: 3D view of the Hutch well showing the horizontal well and the hydraulic fractures.
126
5.2 Results: Production Decline and Future Performance Simulation
Figure 5.16 illustrates the suitable history match between simulation output and the historical oil
production rates and cumulative oil production profiles for the Giesenchlag-Groce horizontal well
of Burleson County Figure 5.13. The numerical simulation model (NSM) forecasted the
cumulative productions to abandonment at 2stb/d using the same production history matched
model (Figure 5.17). In this figure, the expected ultimate recovery (EUR) forecasted by the
numerical simulation model (NSM) is not particularly close to any of the decline curve analysis
techniques’ results for the Giesenschlag-Groce well in the Burleson County.
History match for the Giesenschlag-Groce well in Burleson County
History matched oil rate

Actual oil rate
Actual cumulative production
History matched cumulative production
production data (red and green circles) for Giesenschlag-Groce well in the Burleson County.
127
Simulation and Analytical Production Forecasts to 2 barrels/day

140000
120000
100000
80000
60000
40000
20000
0
GIESENSCHLAG-
GROCE
Well Lease ID
PLE Arps Arps Arps Hyperbolic LGA Logistic- Duong NSM


model (NSM) and various analytical decline curve analysis techniques for Giesenschlag-Groce
well.
Figure 5.18 shows the decent history match between simulation outputs and the historical oil
production rates and cumulative oil production profiles for the Hutch horizontal well of Dimmit
County Figure 5.14. The numerical simulation model (NSM) forecasts of cumulative productions
associated with the closely matched production history are presented for abandonment at 2stb/d
and 5stb/d respectively in Figure 5.19. In this figure, the expected ultimate recovery (EUR)
forecasted by the numerical simulation model (NSM) is close to the analytical forecasts from
Arps’ harmonic, Duong’s method, and LAHA decline curve analysis.
Pressure variations over the period of history matching the Hutch well reservoir are illustrated in
Figures 5.20 through 5.26. The pressure drawdown, from the observations at six months
intervals in these figures, is highest close to the horizontal wellbore and hydraulic fractures
128
compared to the rest of the reservoir leading to possible phase saturation and multiphase flow
near the wellbore and the hydraulic fractures.
History match for the Hutch well in Dimmit County
History matched oil rate

Actual oil rate
Actual cumulative production
History matched cumulative production
production data (red and green circles) for Hutch well in the Dimmit County.
129
Simulation and Analytical Production Forecasts to 2 barrels/day

140000
120000
100000
80000
60000
40000
20000
0
HATCH
Well Lease ID
PLE Arps Arps Arps Hyperbolic LGA Logistic- Duong NSM


model (NSM) and various analytical decline curve analysis techniques for Hutch well.
Figure 5.20: Initial pressure of the reservoir for the Hutch well model in Dimmit County.
130
after 6 months of production of the Hutch well.
131
132
133
5.3 Methods: Enhanced Shale Oil Recovery Simulation
Oil recovery from shale oil reservoirs by primary depletion is very low. As discussed in the
literature review (Chapter 2), it is possible to enhance shale oil recovery by miscible gas
injection displacement process. This section investigates the following depletion scenarios:
1. Five and seven horizontal wells, in patterns 5/P and 7/P, with different hydraulic fracture
half-lengths but equivalent stimulated reservoir volume (SRV) scenarios were modeled to
investigate the effects of the number of horizontal wells and hydraulic fracture half-
lengths on oil recovery by primary energy depletion. Both patterns have the same size,
shape, and total fracture half-length, while the difference is the number of wells and
individual fracture half-length.
2. Three horizontal wells (producers) interspaced with two horizontal wells (injectors) (3-
2/P-I), and four horizontal wells (producers) interspaced with three horizontal wells
(injectors) (4-3/P-I) were modeled by converting two producers of 5/P and three
producers of 7/P, respectively. 3-2/P-I and 4-3/P-I scenarios investigated the effects of
the number of producers and injectors and hydraulic fracture half-length on enhanced
shale oil recovery.
3. Injection of different gas (es) to investigate their respective effectiveness for enhanced
shale oil recovery.
A reservoir of 450 acres was used to simulate the three scenarios. The logic was to test well
geometries and spacing versus production volumes with and without enhanced shale oil recovery
(ESOR) mechanisms. 5/P and 7/P horizontal wells, hydraulically fractured with approximate half-
lengths of 375 ft and 525 ft, respectively, fitted well into the 450-acre reservoir (Figures 5.27 and
134
5.28). Reservoir rock and fluid properties were kept the same as that of the previous history
match in the Dimmit County (Table 5.2 and Figure 5.14).
Figure 5.27 modeled reservoir as 5/P and 3-2/P-I with two viewpoints; to simulate oil recovery
(a) without and (b) with gas injection (primary drive depletion and ESOR scenarios). The primary
drive depletion model (5/P) and the ESOR model (3-2/P-I) have hydraulic fractures with
approximate half-lengths of 525 ft.
Similarly, 7P and 4-3/P-I well patterns were modeled, with hydraulic fractures with approximate
half-lengths of 375 ft for the primary drive depletion model and ESOR. Adjacent wells’
stimulated reservoir volumes (SRV) were modeled 50 ft apart (Figure 5.28). Figure 5.29
illustrates the initial pressure condition of the simulated reservoir model.
3740ft producing interval
Figure 5.27: Reservoir model (5P) consisting of 5 hydraulically fractured horizontal producers
for primary shale oil depletion. This model is later converted to 3-2/P-I with 3 producer and 2
injectors for ESOR
135
50ft spacing between adjacent SRV
Figure 5.28: Reservoir model (7P) consisting of 7 hydraulically fractured horizontal producers
for primary shale oil depletion. This model is later converted to 4-3/P-I with 4 producer and 3
injectors for ESOR.
Figure 5.29: Initial pressure condition for the reservoir model (7P) consisting of 7 hydraulically
fractured horizontal producers for primary shale oil depletion. This model is later converted to 4-
3/P-I with 4 producer and 3 injectors for ESOR.
136
Seventeen hydraulic fracture clusters were integrated into the model based on completion
information available for this the Hutch lease well on the TRRC webpage. However, to lessen the
computation time required to simulate the entire reservoir with seventeen clusters, only one
hydraulic fracture per well was simulated, and the results multiplied by seventeen.
The compositional model was built with the same reservoir parameters as the Hutch lease
wellbore in the Dimmit County (Table 5.2). Each of the given patterns was simulated in turn;
first, with primary drive mechanism, and second, with gas injection. Primary depletion was
simulated to 2 barrels/day/producer with both 5/P and 7/P configurations before the
commencement of gas injection. Thus, 5/P and 7/P depleted the field to 10 barrels/day and 14
barrels/day, respectively, before the start of gas injection.
The reservoir was depleted under a minimum bottom hole pressure of 1850 psia. When the
primary depletion reached the production limits of 2 barrels/day/producer, then two wells of the
5/P pattern and three wells of the 7/P pattern operated as injectors as described for Figures 5.27
and 5.28. The gas injection continued until the production limit of 2 barrels/day/producer was
reached, again under enhanced shale oil recovery. The production constraint of 2
barrels/day/producer, in a field-wide sense, implies 10 barrels/day for a 5/P well arrangement.
Similarly, the field-wide rate constraints for the 7/P well pattern are 14 barrels/day. Subsequently,
final rate limit for ESOR is 6 barrels/day for 3-2/P-I, when the reservoir depletes to 2
barrels/day/producer. For the 4-3/P-I well pattern, the rate is 8 barrels/day for the limit of 2
barrels/day/producer.
137
Operational conditions were arbitrarily chosen for ESOR, including 1000 rcf/day of gas injected
continuously at a bottom-hole pressure of 4000 psia. These conditions did not have the goal of
recovery factor optimization or any economic significance, but were used to investigate the
technical viability of the miscible gas injection to enhance shale oil recovery.
5.3.1 Results: Enhanced Shale Oil Recovery Simulation
The production rate profiles for both types of reservoir and wells’ configurations using the
different gas injectants simulated under primary depletion drive and miscible gas displacement
process are presented in Figures 5.30 and 5.31. These plots show different recovery patterns for
the different miscible gas injectants due to differences in respective gas injection properties at the
reservoir conditions of temperature and pressure. The miscible gas injectants displacement
process occurred at slightly higher rates for the 7 wells’ pattern compared to the 5 wells’
configuration (Figure 5.30 and 5.31).
The peaks are highest for miscible CO2 injectant (Figure 5.30 and 5.31). The reason is that
supercritical CO2 has a comparable density as the in-situ hydrocarbon fluids, and therefore, it is
able to interact and displace more residual oil due to a more efficient mass transfer activity. This
argument is justifiable by the discussion of the contents of Table 5.6.

138
Reservoir Simulation under Primary Depletion Drive and Miscsible Gas

Injection Displacement Process (5 Wells)
50
(barrels/day/producer)
Oil Production Rate
40
30
20
10
0
11/18/2010 12/30/2040 2/11/2071 3/26/2101
Time (Year)
CO2 Enriched Gas CO2-CH4 Production Under Primary Drive
miscible gas injection displacement process (5P and 3-2/P-I well patterns)
Reservoir Simulation under Primary Depletion Drive and Miscsible Gas

Injection Displacement Process (7 Wells)
50
(barrels/day/producer)
Oil Production Rate
40
30
20
10
0
11/18/2010 12/30/2040 2/11/2071 3/26/2101
Time (Year)
CO2 CO2-CH4 Enriched Gas Production Under Primary Drive
miscible gas injection displacement process (7P and 4-3/P-I well patterns).
The simulation results of primary depletion and miscible gas injection displacement process in
terms of cumulative oil production and recovery factors are summarized in Table 5.5. CO2
139
injection has slightly more enhanced shale oil recovery than the other gas injectants depending on
the wells’ pattern, production limit rate, and simulation period. For the 7 wells (4-3/P/I) reservoir
depletion to 2 barrels/day/producer, the individual gases injected have the recovery factors of
CO2 =5.8%, CO2-CH4=5.44 and enriched gas=5.37% (Table 5.5). The corresponding
observations for the 5 wells (3-2/P-I) are CO2 =5.4%, CO2-CH4=5.1 and enriched gas=5.0%
(Table 5.5).
In terms of total recovery factors and depletion to 2 barrels/day/producer, the 7 well patterns (4-
3/P-I) performs slightly better than the 5 well pattern (3-2/P-I) (Table 5.5). The ultimate
justification of these slight advantages depends on economic evaluation for the different
scenarios.
Table 5.5: Primary depletion, ESOR cumulative oil produced, and recovery factors
Primary
Configuration
Depletion ESOR
Cumulative Primary Cumulative
Years of Oil Depletion Years of Oil ESOR
Production Primary Produced Recovery Gas ESOR Produced Recovery
Limit Depletion (STB) Factor Injected Depletion (STB) Factor
CO2 1069648.5 5.4
2 Barrels
04/30/2010 Enriched 11/30/2043
5 Wells per
- 827200.3 4.2 Gas - 997942.5 5.0
Pattern day per
11/30/2043 CO2- 12/1/2100
producer
CH4 1008066.0 5.1
CO2 1162522.2 5.8
2 Barrels
04/30/2010 Enriched 1/31/2038
7 Wells per
- 804701.1 4.0 Gas - 1068135.8 5.4
Pattern day per
1/31/2038 CO2- 12/1/2100
producer
CH4 1082924.3 5.4
140
Presence of natural fractures would reduce the volumetric sweep efficiency during EOR. The
interesting features about the enhanced recovery process are the differences in the recovery
patterns considering the different peaks/humps of each injectant’s (Figures 5.30 through 5.31).
These differences can be explained by considering the operating conditions of injection wells, the
producing GOR (Figure 5.32), the Z-factors (measure of the gas compressibility), density,
viscosity of the injectants, and subsequently, the ratios of bottom hole injection rates presented in
Table 5.6. With the operating conditions of 4000 psia bottom hole injection pressure (BHIP) and
gas injection rate of 1000 rcf/day, Table 5.6 pointed to the significant differences between CO2
and the other miscible gas injectants. These differences include low compressibility of CO 2
compared to the compressibility of the other injectants, and higher density and viscosity of CO2 in
comparison to the others. The effects of these significant differences include high volumetric
injection of the CO2 at relatively high injection rates compared to the volumes and rates of
injection of miscible CO2-CH4 and enriched gas. Despite injecting much more CO2 both at the
microscopic and macroscopic scale, the GOR plot in Figure 5.32 indicates only a slight
difference between the production of CO2 and the other injectants of less volumetric injection.
This means that a lot of miscible CO2 tends to remain in the reservoir due to its solubility in oil
compared to the others. This favorable behavior of CO2, results in its low viscosity, later
breakthrough, oil swelling and reduction in oil viscosity all of which in favor of higher oil
recovery.
141
Table 5.6: Miscible gas injectants properties at reservoir operating condition during ESOR.
(Note: Viscosity ratio (µ) and fraction of rates (q) are relative to the highest property values of
CO2)
Z- Density Viscosity Bg 1/Bg µ-gas/µ-CO2 q-gas/q-CO2

Gas
Factor gm/cc cp (rcf/scf) (scf/rcf) ( Ratio) (Fraction)
CO2 0.5950 0.6572 0.2488 0.0028243 354.1 100.0 1.00

CH4 0.9740 0.1464 0.0166 0.0046236 216.3 6.7 0.61
CO2-
CH4 1.0459 0.1178 0.0148 0.0049648 201.4 5.9 0.57
Enric
hed
Gas 0.9416 0.1808 0.0193 0.0044699 223.7 7.7 0.63
Producing Gas-Oil-Ratios for the miscible gas injectants under primary depletion drive and
miscible displacement process is illustrated in Figure 5.32 indicating a slight difference between
the volumes of injectants per barrels of oil produced. This again confirms that at any given time
CO2 is more in the reservoir resulting in higher volumetric efficiency than its counterparts.
The breakthrough time for CO2 injection was 5 years; it was 2.5-3 years for the other miscible
gas injectants. These variations in breakthrough time are due to the favorable high viscosity and
density as well as high solubility in oil. Thus, less viscous and less dense gases have high
mobility and can therefore easily bypass the oil in the reservoir and are easily producible through
the fractures.
142
Figure 5.32: Production Gas-Oil-Ratios of the miscible gas injectants under primary depletion
drive and miscible displacement process.
The minimum miscibility pressures (MMP) of CO2 and CO2¬-CH4 are the same (Table 5.3), but
methane is mostly immiscible and has low density than the hydrocarbon fluids. As a result, it
segregates and overrides the oil instead of mixing to reduce viscosity and increase swelling factor
of the oil.
The enriched gas had the lowest minimum miscibility pressure, but its oil recovery behavior was
the poorest. Also, considering (Table 5.6), enhanced shale oil recovery by the enriched gas
should not have been the lowest since it has a slightly favorable compressibility, density,
143
viscosity than CO2-CH4 and CH4. This behavior can be attributed to injecting less volume of
enriched gas which is only 63% of volume of pure injected CO2.
The pressure profiles (Figure 5.33 through 5.37) observed in the reservoirs from start to end of
simulation can explain the general low recovery factors. The average reservoir pressure depleted
very fast from 3200 psi to 1850 psi (BHP) within 5 years (Figures 5.33 and 5.34). Reading from
Figures 5.30 and 5.31, this period marked the end of the early years of fast production rate
decline and the beginning of the very low production rates, which lasted the long transient flow
periods till abandonment.
At the onset of miscible gas injection (Figure 5.35), pressure gradient returned to the reservoir
leading to the peaks observed in Figures 5.30 and 5.31 associated with increased production of
mobilized oil. By 2040, the peak is fully reached (Figure 5.31), and since the pressure wave
could not penetrate the 50 ft interval between the stimulated zones surrounding producers and
injectors, there was a sharp decline in the production rate again. It is observed that the pressure
regimes in the reservoir in 2040 and 2100 are similar (Figures 5.36 and 5.37); hence pressure
depletion rate declined leading to long production plateau for most part of the enhanced shale oil
recovery (ESOR).
144
Figure 5.33: Uniform reservoir pressure of 3200 psi at the onset of primary oil recovery
Figure 5.34: The reservoir pressure depleted from 3200 to 1850 psi uniformly by 2015 during the
primary depletion drive. Note the wells are all producers, well cone pointing up.
145
Figure 5.35: Next pressure change in the reservoir was not until 2038 at the onset of miscible gas
injection due to injector bottom hole pressure of 4000 psi. Note: cones pointing down are the
injectors.
Figure 5.36: Pressure wave is highest at the injectors and least at the producers creating a
pressure gradient. This pressure waves moves the oil towards the producers.
146
Figure 5.37: Pressure profile did not change much between 2040 and 2100. Almost an abrupt
sharp pressure drop can be observed in the spaced zone between producers and injectors.
5.4 Pressure Profiles across Lease Boundaries
Oilfield demarcation into leases limits an owner’s acreage in terms of practical development, but
subsurface fluid flow behavior cannot obey legal and physical boundaries unless shale reservoirs
are completely impermeable. As a result, lease boundaries’ integrity with respect to pressure
depletion in the subsurface need to be examined. To address this issue, the drainage area was
investigated by measurements of pressure profiles across the boundary of the stimulated reservoir
volume (SRV) and into adjacent leases using a simulation model.

147
5.4.1 Methods
I modeled a hypothetical field development scenario with the reservoir parameters as those
previously used in the ESOR. It comprised as many as seven producing horizontal wells to create
a substantial stimulated reservoir volume and six lateral observation wells in contiguous non-
stimulated shales (Figure 5.38). Thus, wells 1-7 are perforated and hydraulically fractured
producers; wells 8-10 are observation wells oriented parallel to the large stimulated reservoir
zone, at 150 ft, 450 ft and 750 ft away from the lease boundary. Wells 11-13 are observation
wells oriented perpendicular to the producing wells at distances 150 ft 450 ft and 750 ft
respectively to the right of the large stimulated reservoir zone. The observation wells monitor
bottom hole pressure responses across adjoining stimulated reservoir regions. The Cartesian grid
dimensions of the reservoir is 60x75x1 for a total of 24611 grid block cells with local grid
refinement near wellbores and hydraulic fractures. Figure 5.37. Grid block thicknesses of 100 ft
x 100 ft x 147 ft resulted in a modeled volume of 6.62x109 ft3. No assumptions with regards to
geological complexities were incorporated in the reservoir model. This model was simulated for
60 years and the wells’ pressure profiles were plotted as wells’ bottom-hole pressures (BHP)
versus time.
148
Figure 5.38: Simulation model to investigate bottom hole pressure in adjoining leases during
depletion.
149
5.4.2 Results: Pressure Profiles across Lease Boundaries
The pressure drops vary based the distances of the observation wells away from the tip of the
SRV (Figure 5.38). The closest to the SRV boundaries are observation wells 8 and 11 at 150 ft
away and there is negligible difference in their measured BHP (Figure 5.39). Similar
insignificant differences are observed at equal distances for the parallel and perpendicular
observation wells 9, 12 and 10, 13. However, the most important observation is that the pressure
drawdowns in the non-stimulated zone never attain the saturation pressure of 2054 psia of the
reservoir. It remained under saturated, and it should be noted that volumetric oil withdrawal
associated with large pressure depletion will be insignificant in this circumstance. Table 5.7
shows the ranges of pressure depletion occurring successively over the entire simulation period of
60 years. Considering the low compressibility of oil, the drainage of fluid away from hydraulic
fracture tips is not significant.
Table 5.7: Pressure responses across stimulated reservoir volume into non-stimulated reservoirs
Wells'
distance Days Successive Pressure drawdown (psi)
Observation from before
Wells' ID# limit first pressure 1st 1st 1st 1st 1st 1st
of SRV communication January January January January January January
(ft) 2015 2020 2030 2040 2050 2060
8 150 0.08 225.7 402.8 580.2 672.4 732.8 777.0

9 450 92 6.5 41.4 142.8 231.8 302.9 361.1
10 750 745 0.1 2.1 22.8 60.8 104.9 150.2
11 150 0.74 246.6 421.8 592.8 682.7 740.5 782.8
12 450 122 7.8 46.7 151.1 240.1 310.4 366.8
13 750 716 0.1 2.4 24.9 63.5 106.0 147.1
150
Figure 5.39: Pressure profiles of all observation wells at distances away from the ends of
stimulated reservoir volume (SRV).
151
Chapter 6 Discussion
6.1 Decline Curve Analysis of the Eagle Ford Shale Oil Production
Production profile analyses by the traditional Arps’ DCA and various new models have one
unique purpose: to evaluate and address the past production performance and future prediction
challenges associated with low permeability and transient flow regimes in shale and other tight
formations. The outcome of the application of each analytical model to the forty production
profiles from wells in the Eagle Ford Shale oil trend varies according to the individual models’
fundamental concepts and empirical equations. This section discusses the strengths and
weaknesses identified in the use of each decline curve analysis technique.
Traditional methods of Arps’ decline curve relations assessed the production profiles to varying
degrees based on the decline type (exponential, hyperbolic, or harmonic). Based on R2 ≥ 0.95,
6.45% and 29.03% of the wells follow exponential and harmonic decline, respectively.
The Arps’ hyperbolic decline model, regardless of the values of the decline exponent (b), closely
fitted the past production rate profiles for all forty wells. However, the success of the hyperbolic
technique depends on accurate reserves estimation and future production predictions; this
accuracy is better when the decline exponent (b) is 0<b<1 than when b>1. The percentage success
of the hyperbolic decline model (0<b<1) to accurately match the observed decline history is
32.5% for the forty wells investigated. The percentage majority (67.5%) of the forty wells
production decline assessment using the Arps’ hyperbolic relation resulted in b>1. This result
may be due to inconsistencies in the actual production trend due to possible changing operating
conditions. The unusually high percentage of b>1 compromises the reliability of most of the
152
reserves and future production calculations regarding the use of the Arps empirical decline curve
analysis equations.
When Arps’ hyperbolic decline model is used in combination with other decline models, such as
the logistic growth analysis, Duong’s method, and the power law relation techniques, more
realistic values of b and subsequent cumulative forecasts may be obtainable. I verified this
assertion. The results obtained from the combination of the PLE model data and Arps’ hyperbolic
relations are slightly less accurate to the logistic Arps’ hyperbolic approach (LAHA). When
Duong’s method is combined with Arps’ hyperbolic decline concept, similar percentage
improvement of normal decline exponent is observed as in the case of LGA, but the Duong’s
method still matches the cumulative production less fittingly than the other techniques.
The logistic growth analysis (LGA) technique, one of the established alternatives to the
traditional Arps’ methods, seems to match cumulative production profiles regardless of operating
conditions. However, a suitable cumulative production match does not necessarily translate into
an acceptable production rate match. Consequently, the erratic production profiles may affect the
outcomes of LGA; hence, data filtering is necessary to accurately estimate reserves and generate
future performance curves. The LGA decline exponents (n), equivalent to Arps’ hyperbolic
decline exponents (b), persistently have values 0≤n≤1 (i.e., 95%), except for two: Karnes County
well #09667 of the Berry lease and Gonzales County well #15526 of the Georg lease. LGA gives
more conservative estimates than do the optimistic decline model types (Arps’ hyperbolic,
harmonic, and Duong’s method).

153
The PLE predictions are similar to Arps’ exponential decline whenever both apply. PLE
consistently estimated the lowest values of EUR, future production, and lifetime remaining. It is,
therefore, the most conservative approach to generating production forecasts and calculating shale
oil reserves.
Duong’s method did not match the majority of the wells’ past cumulative production because of
its high sensitivity to inconsistent decline trends. Nevertheless, Duong’s decline analysis method
generated volumes comparable to that of other models, especially when the random data points
are accurately identified and filtered out.
Finally, the Logistic-Arps’ Hyperbolic Approach (LAHA), is the author’s proposal to combine
two established techniques, the new logistic growth analysis (LGA) model and the traditional
Arps’ hyperbolic decline relation, to improve the percentage success of the Arps’ method. The
expectation of LAHA is to translate the advantages of the LGA method by increasing the
percentage of 0<b<1 for the Arps’ hyperbolic decline curve analysis, thus improving future
production prediction and EUR.
The selection of one technique over the other must be based on longer production data and the
feedback from continuous applications of the available techniques. There is not yet a perfect
decline model for the Eagle Ford Shale oil, but all of them can be used if their individual
limitations are recognized and managed properly until longer production histories become
available.
154
6.2 Production Decline and Forecast, ESOR and Pressure Profiles Simulation
6.2.1 Production Decline and Forecast Simulation
In this section, I discussed the simulation results of the two wells compositional models in the
Burleson and Dimmit Counties in comparison to the results obtained from decline curve analysis
methods.
Production decline simulation using the two horizontal wells’ production profiles resulted in
satisfactory history matches. Comparison of expected ultimate recovery (EUR) between the two
wells’ simulation results and the corresponding analytical methods do not show similar trends.
Thus, the EUR for the Burleson County well’s production profile is not particularly close to any
of the decline models’ results. On the other hand, prediction based on simulations of the well in
Dimmit County are similar to the forecasts obtained using the Arps’ harmonic, Duong’s method,
and LAHA decline techniques.
However, conducting simulations of only two of the 40 horizontal wells production profiles
evaluated analytically was insufficient for a fair performance comparison between the various
decline curve analyses techniques. It should also be noted that history matching, especially for a
short production history and a small model, is not unique. This suggests that comparisons of the
simulation and analytical results be done with caution.
6.2.2 Enhanced Shale Oil Recovery (ESOR) Simulation
This section discusses the results of using different wells’ configuration and gas injections to
enhance shale oil recovery. As noted earlier in the literature review, the oil recovery factor for the
155
Eagle Ford Shale play during primary energy reservoir depletion is roughly 5% (Hart 2011). In
this research, simulation results showed recovery factors improved from 4.0% under primary
depletion drive to as much as 5.8% for ESOR using CO2 injectant and the seven wells pattern (4-
3/P-I). The recovery factor is very low compared to significant recovery factor improvements by
miscible gas injection displacement processes associated with conventional reservoirs. The
difference is due to broad range of mobility of hydrocarbon phases in conventional reservoir;
whereas in shale reservoirs, the mobility is restricted to a much narrower saturation range because
of the nano- Darcy permeability of the matrix. Most of the gas injectants for the miscible
displacement take the path of hydraulic fracture conduits instead of infiltrating the micro pores to
mobilize the residual oil of the shale reservoirs. This leads to poor volumetric sweep efficiency.
The results of the miscible gas injection displacement process vary for the different gas injectants,
but the best scenario is for well pattern (4-3/P-I), CO2 gas injection and depletion to 2
barrels/day/producer. The likelihood for this to occur is CO2 relative high ability to attain
solubility and multiple contact miscibility with reservoir oil compared to other gases. Different
interactions between the in-situ crude and the different gas compositions resulted in different rate-
time profiles and breakthrough time. CO2 took the most time to breakthrough compared to the
other gases injected. These differences are the result of differences in phase behavior-solubility
and first or multiple contact miscibility with the in-situ crude-of the various gas compositions.
Depending on well configuration and rate constraints, depletion under primary drive took
different time periods. Enhanced shale oil recovery (ESOR) simulation could not deplete the
reservoir to the production limit of 2 barrels/day/producer 60 years after primary drive depletion.
Since shale reservoirs are expected to advance to stripper status and could last 100 or more years,
156
ESOR to lower production limits could be worthwhile, depending on its economic viability. After
more than 60 years of ESOR, 3-2/P-I and 4-3/P-I well configurations flow at 2.4 and 3.0
barrels/day/producer respectively. Thus, with ESOR by gas injection, the Eagle Ford Shale oil
reservoir may produce above 2 barrels/day/producer. However, whether ESOR is economically
realistic will be determined by crude oil prices, plus maintenance and operating costs.
The 4-3/P-I well pattern generally recovers more oil than the 3-2/P-I well arrangement under
ESOR to 2 barrels/day/producer (Table 5.5). This may be as a result of increased displacement or
sweep efficiency by the gases in the 4-3/P-I arrangement. This could be due to better volumetric
sweep and higher pressure gradient between producer/injector pairs.
6.2.3 Pressure Profile Simulation
Pressure profiles across the extent of the stimulated reservoir volume (SRV) and adjacent leases
were modeled and simulated in order to investigate the drainage zone. The simulation result could
apply to any tight nano-Darcy permeability reservoir such as the shale reservoirs. Shale reservoir
leases are like sharing a piece of a globe (pie), and it is important to investigate the level of
encroachment going on in the subsurface, with regards to hydrocarbon fluid depletion, from one
piece to the other. The pressure profile simulation results show that the total pressure depletion
outside of the SRV over a long period of 60 years is about 780 psi. This drop in pressure is
insignificant considering the fact that oil compressibility is very low and, in unsaturated oil
reservoir, a little fluid withdrawal can lead to high pressure drops. Hence, it can be concluded that
the volumetric oil flow from outer leases adjacent to the SRV is not significant.
157
Chapter 7 Conclusions and Recommendations
7.1 Conclusions
The purpose of this research was to evaluate the effectiveness of various established decline curve
analysis methods of evaluating past production performance and predicting future performance of
the Eagle Ford Shale oil reservoirs. Both analytical and simulation techniques modeled
production profiles and were compared.
Furthermore, since the Eagle Ford Shale oil recovery factor is low, this research explored the
technical feasibility of enhanced shale oil recovery (ESOR) by miscible gas injection
displacement process.
In addition to the above two objectives, this thesis investigates pressure profiles across lease
boundaries and analyzes possible drainage area outside of the stimulated reservoir volume (SRV).
Based on the above objectives, the following conclusions were deduced from this study. These
conclusions are limited to the data and the results from this study.
1. The actual production profiles from production wells of the Eagle Ford Shale are often
erratic; it is difficult to discern a reasonable decline trend without filtering out off-trend
data. This may affect the results of the decline curve analyses; filtering may be biased and
introduce errors in future estimates.
2. Arps’ exponential and PLE analysis are conservative approaches for estimating reserves
and forecasting future performance. Arps’ exponential decline relation rarely applies at
R2 ≥ 0.95. Arps’ hyperbolic, Logistic growth analysis, Duong’s method and Logistic
158
Arps hyperbolic approach consistently provided high estimates, and these methods may
be said to be the optimistic approach to generating future productions. The Power law
exponential technique, on the other hand, consistently provided the least EUR estimates,
and the method may be said to be the pessimistic approach to generating future
productions
3. Arps’ hyperbolic DCA method proves unsuccessful to estimate reserves and generate
future production trends correctly as a consequence of unusual high Arps’ decline
exponent (b). Forecasting transient production with Arps’ hyperbolic equations led to a
severe overestimation of EUR, and for that matter, remaining reserves.
4. Duong model did not always fit the cumulative production despite closely fitting most of
the production rates. The EUR mostly converges to finite/reasonably recoverable
reserves. There were 5 cases where it forecasted the highest.
5. Logistic growth analysis maintained the greatest success in modeling the past production
performance, both production rates and the cumulative fitted closely. It also consistently
generated the conservative EUR of all the optimistic DCA techniques.
6. Logistic Arps Hyperbolic Approach (LAHA) is my innovation. It successfully improved
the occurrence of 0<b<1 associated with Arps’ hyperbolic from 32.5% to 80%, thereby
reducing b>1 from 67.5% to 20%. EUR forecasted by LAHA were relatively realistic
predictions compared to the high (unrealistic) value associated with the original Arps’
159
hyperbolic. It is close to the forecasts of LGA, and it is reported in terms or parameters to
which industry, analyst and investors are accustomed.
7. With CO2 gas injection displacement process, the 4.0% primary recovery factor increased
to 5.8%. Favorable performance of the CO2 compared to the other miscible gas injectants
is due to its higher density, viscosity and solubility which reduce the oil viscosity and
causes it to swell. More CO2 remains in the reservoir increasing its volumetric efficiency
compared to the other gases. Optimization scenarios and economic analyses were not
involved. If proven economically viable, stripping shale oil reservoirs are good
candidates for ESOR by CO2 injection.
8. Finally, based on pressure drop, drainage area of SRV does not significantly extent into
the adjacent non-stimulated shale zones of extremely low permeability. Integrity of lease
boundaries cannot be said to be compromised unless where the formation is naturally
fractured. The volumetric oil withdrawal outside of the SRV is insignificant considering
the low mobility of oil. This could be different for gas wells with low viscosity.
7.2 Recommendations
Further work to be done may include the following:
1. Production data publicly available at the time this research was conducted was
insufficient for accurate determination of wells’ individual performance in terms of
decline analysis and production forecast. It is strongly recommended that any future work
include more data from a larger number of wells to ensure statistically significant results.
2. Initial rates, decline rates, and EUR should be analyzed together with the wellbore
dynamics and geometry, i.e., perforation clusters, hydraulic fracture intervals, fracture
160
half-lengths, and lateral length of horizontal wells, whenever available, to assess their
interdependency.
3. Since reservoir simulation models depend on accurate static and dynamic reservoir and
hydraulic fracture data, the study strongly recommends that future simulation input
parameters be considerably augmented beyond the very basic ones available at the time
of this research.
4. Reservoir simulation of production decline should include a larger number of wells. This
creates a broader range of well behavior to assess past and future performance and
evaluate the difference between simulation results and other analytical methods.
5. Miscible gas displacement process to enhance shale oil recovery should include
optimization scenarios such as gas mobility control measures to avoid early bypass and
breakthroughs. These should include sensitivity of the inter-spacing of the ends of the
hydraulic fractured zones and polymer augmented miscible gas injection.

161
REFERENCES
Agboada, D., and Ahmadi, M. (2013). Production Decline and Numerical Simulation Model
Analysis of the Eagle Ford Shale Oil Play. SPE 165315-P, 1-30.
Arps, J. J. (1945). Analysis of decline curves. Transactions of the American Institute of Mining,
Metallurgical and Petroleum Engineers, 160, 228.
Berg, R. R., and Gangi, A. F. (1999). Primary Migration by Oil-Generation Microfracturing in

Low-Permeability Source Rocks: Application to the Austin Chalk, Texas. AAPG Bulletin , 83
(5), 727-756.
Bui, L. H. (2010). Near Miscible Carbon-dioxide Application to Improve Oil Recovery. Kansas
City: University of Kansas.
Chaudhary, A. (2011). Shale Oil Production Performance From A Stimulated Reservoir Volume.
College Station: Texas AandM University.
Chaudhary, A. S., Ehlig-Economides, C., and Wattenbarger, R. (2011). Shale Oil Production
Performance from a Stimulated Reservoir Volume. SPE 147596 , 1-21.
Clark, A. J. (2011). Decline Curve Analysis in Unconventional Resource Play using Logistic
Growth Models (2011 ed.). The University Texas at Austin.
Duong, A. (2010). An Unconventional Rate Decline Approach for Tight and Fractured-
Dominated Gas Wells. SPE 137748 , 1-15
Duong, A. (2011). Rate-Decline Analysis for Fractured-Dominated Shale Reservoirs. SPE

137748 , 377-387.
Fan, L. et al., 2011. An Integrated Approach for Understanding Oil and Gas Reserves Potential in
Eagle Ford Shale Formation. SPE 148751, p. 15.
Fertl, W. H., and Rieke III, H. H. (1980). Gamma Ray Spectral Evaluation Techniques Identify
Fractured Shale Reservoirs and Source-Rock Characteristics. SPE 8454 , 1-10.
Ghoodjani, E., and Bolouri, S. (2012). Numerical and Analytical Optimization of Injection Rate
During CO2-EOR and Sequestration Processes. Orlando: Carbon Management Technology
Conference.
Ghoodjani, E. and Bolouri, S., (2011). Experimental Study of CO2-EOR and N2-EOR with Focus
on Relative Permeability Effect. Petroleum and Environmental Biotechnology, pp. 1-5.
Government-Report. (2011, July 12). Government Report Details Vast Shale Wealth. Retrieved
July 06, 2013, from http://eaglemap.com/news/bid/61474/Government-Report-Details-Vast-
Shale-Wealth
162
Gui, P., Jia, X., Cunha, J., and Cunha, L. (2008). Economic Analysis for Enhanced CO2 Injection
and Sequestration Using Horizontal Wells. JCPT, 47 (11), 34-37.
Gupta, D., and Bobier, D. (1998). The History and Success of Liquid CO2 and CO2/N2
Fracturing System. SPE 40016, 1-8.
Hart, N. (2011, September 30). How Long Will Eagle Ford Shale Wells Produce? Retrieved from
Eagle Ford Shale Blog Web site: http://eaglefordshaleblog.com/2011/09/30/how-long-will-eagle-
ford-shale-wells-produce
Holditch, A. S. (2003). The Increasing Role of Unconventional Reservoirs in the Future of the Oil
and Gas Business. JPT Archives, 55 (11), 34-37.
Holm, L. (1986). Miscibility and Miscible Displacement. SPE 15794, 1-2.
Hsu, S.C., and Nelson, P. P. (2002). Characterization of Eagle Ford Shale. In Engineering
Geology (pp. 169-183). Elsevier Science B.V.
Ilk, D., Rushing, J., Perego, A. & Blasingame, T., 2008. Exponential vs. Hyperbolic Decline in
Tight Gas Sands: Understanding the Origin and Implications for Reserve Estimates Using Arps'
Decline Curves. SPE 116731, 1- 23.
Lillies, A. T., and King, S. R. (1982). Sand Fracturing With Liquid Carbon Dioxide. SPE 11341,
1-8.
Malik, Q. M., and Islam, M. (2000). CO2 Injection in the Weyburn Field of Canada:
Optimization of Enhanced Oil Recovery and Greenhouse Gas Storage with Horizontal Wells.
SPE 59327, 1-16.
Martin, R., Baihly, J., Malpani, R., Lindsay, G., Atwood, K. W., and Schlumberger. (2011).
Understanding Production from Eagle Ford-Austin Chalk System. SPE 145117, 1-28.
McFarland, G. (2010). Petroleum Geochemistry and Shale Plays. Richardson: Obsidian Energy
Company, LLC.
McNeil, R., Jeje, O. and Renaud, A., 2009. Application of the Power Law Loss-Ratio Method of
Decline Analysis. Canadian International Petroleum Conference (CIPC) 2009-159, pp. 1-7.
Medeiros, F., Kurtoglu, B., Ozkan, E., and Kazemi, H. (2007). Analysis of Production Data from
Hydraulically Fractured Horizontal Wells in Tight, Heterogeneous Formations. SPE 110848, 1-
11.
Medeiros, F., Ozkan, E., and Kazemi, H. (2008). Productivity and Drainage Area of Fractured
Horizontal Wells in Tight Gas Reservoirs. SPE 108110, 1-10.
Mendoza, E., Aular, J., and Sousa, L. (2011). Optimizing Horizontal-Well Hydraulic-Fracture
Spacing in the Eagle Ford Formation, Texas. SPE 143681, 1-16.
163
Miskimins, J.L. (2008). Design and Life Cycle Considerations for Unconventional Reservoir
Wells. Paper SPE 114170 presented at the SPE Unconventional Reservoirs Conference,
Keystone, Colorado, USA, 10-12 February.
Mullen, J. (2010). Petrophysical Characterization of the Eagle Ford Shale in South Texas. SPE
138145, 1-19.
Nielson, J. (1989). Recovery of oil from oil-bearing formation by continually flowing pressurized
heated gas through channel alongside matrix. US Patent 4856587, 1-4.
Nwabuoku, K. C. (2011). Increasing Lateral Coverage in Eagle Ford Horizontal Shale

Completion. SPE 147549, 1-12.
Odusina, E., Dr. Sondergeld, C., and Dr. Raj, C. (2011). NMR Study of Shale Wettability.
SPE147371, 1-15.
Orangi, A., Nagarajan, N., Honarpour, M., and Rosenzweig, J. (2011). Unconventional Shale Oil
and Gas-Condensate Reservoir Production, Impact of Rock, Fluid, and Hydraulic Fractures. SPE
140536, 1-15.
Passey, Q., Bohacs, K., Esch, W., Klimentidi, R., and Sinha, s. (2010). From Oil-Prone Source
Rock to Gas-Producing Shale Reservoir - Geologic and Petrophysical Characterization of
Unconventional Shale Gas Reservoirs. SPE 131350, 1-29.
Perkins, T., and Kern, L. (1961). Widths of Hydraulic Fractures. Journal of Petroleum
Technology-SPE 89, 13 (9), 1-13.
Rahmatabadi, K. A. (2011). Advances in Calculation of Minimum Miscibility Pressure. Austin:

University of Texas Austin (http://www.pge.utexas.edu/theses11/ahmadi.pdf).
Rickman, R. et al., 2008. A Practical Use of Shale Petrophysics for Stimulation Design
Optimization: All Shale Plays Are Not Clones of the Barnett Shale. SPE 115258-MS, 1-11.
Roussel, N. P., Manchanda, R., and Sharma, M. M. (2012). Implications of Fracturing Pressure
Data Recorded during a Horizontal Completion on Stage Spacing Design. SPE 152631, 1-14.
Rubin, B. (2010). Accurate Simulation of Non Darcy Flow in Stimulated Fractured Shale
Reservoirs. SPE 132093-MS, 1-16.
Settari, A., Bachman, R., and Morrison, D. (1986). Numerical Simulation of Liquid CO2
Hydraulic Fracturing. Petroleum Society of CIM (Paper No: 86-37-67), 399-418.
Shelley, R., Saugier, L., Al-Tailji, W., Guliyev, N., Shah, K., and Godwin, J. (2012). Data-Driven
Modeling Improves the Understanding of Hydraulic Fracture Stimulated Horizontal Eagle Ford
Completions. SPE 152121, 1-11.
Shoaib, S., and Hoffman, T. B. (2009). CO2 Flooding the Elm Coulee Field. SPE 123176, 1-11.
164
Sinal, M. L., and Lancaster, G. (1987, Sept.-Oct.). Liquid CO2 Fracturing: Advantages and
Limitations. JCPT-87-05-01, pp. 26-30.
Sondhi, N., 2011. Petrophysical Characterization of the Eagle Ford Shale, Norman, Oklahoma:
Mewbourne School Of Petroleum And Geological Engineering.
Stegent, N., and Ingram, S. (2011). Hydraulic Fracture Stimulation Design Considerations and
Production Analysis. SPE 139981, 1-10.
Stegent, N., Wagner, A., Mullen, J., and Borstmayer, R. (2010). Engineering a Successful
Fracture-Stimulation Treatment in the Eagle Ford Shale. SPE 136183.
TRRC-Website. (2012, May 22). Railroad Commission of Texas. Retrieved June 10, 2012, from
http://www.rrc.state.tx.us/eagleford/index.php
Tuttle, S. (2010, September 24). Oil Resource Plays-Examples and Technology. Society of
Independent Professional Earth Scientists (SIPES): Houston Chapter, pp. 1-37.
Vassilellis, G., Li, C., Seager, R., and Moos, D. (2010). Investigating the Expected Long-Term
Production Performance of Shale Reservoirs. SPE 138134, 1-12.
Vincent, M. C. (2011). Optimizing Transverse Fractures in Liquid-Rich Formations. SPE 146376,

1-16.
Wang, J., and Liu, Y. (2011). Well Performance Modeling in Eagle Ford Shale Oil Reservoir.
SPE 144427, 1-9 (1).
Wang, X., Luo, P., Er, V., and Huang, S. (2010). Assessment of CO2 Flooding Potential for
Bakken Formation, Saskatchewan. CSUG/SPE 137728, 1-14.
Waters, G., Dean, B., Downie, R., Kerrihard, K., Austbo, L., and McPherson, B. (2009).
Simultaneous Hydraulic Fracturing of Adjacent Horizontal Wells in the Woodford Shale. SPE
119635, 1-22.
Wei, Y., and Economides, M. (2005). Transverse Hydraulic Fractures from a Horizontal Well.
SPE 94671, 1-12.
165
NOMENCLATURE
a = Constant of tn when half the oil has been recovered, (variable used in LGA)
a = Intercept constant, day-1 (variable used in Duong method)
b = Arps’ decline exponent
Bo = Oil formation volume factor
BHP = Bottom hole pressure
CH4 = Methane
CO2 = Carbon dioxide
d = Day
DCA = Decline curve analysis
= Decline constant “intercept” at 1 time unit, (t=1, ) (variable used in PLE)
= Decline constant at “infinite time” ) (variable used in PLE)
= Decline constant ( = /n), (variable used in PLE)
D = Arps’ decline constant, 1/day or 1/time
EOS = Equation of state
EOR = Enhanced oil recovery
ESOR =Enhanced shale oil recovery
EUR = Expected ultimate recovery
FCM = First contact miscibility
FVF = Formation volume factor
GOR = Gas oil ratio
K = Carrying capacity (maximum physically recoverable oil), (variable used in LGA)
Krg = Relative permeability to gas

166
Krog = Relative permeability to oil in a gas-oil system
Krow = Relative permeability to oil in a oil-water system
Krw = Relative permeability to water
LAHA = Logistic Arps’ hyperbolic approach
LGA = Logistic growth analysis
MCM = Multiple contact miscibility
MMP = Minimum miscibility pressure
m = Slope (variable used in Duong method)
NSM = Numerical simulation model
N2 = Nitrogen
n = Time exponent, dimensionless (variable used in PLE)
n = LGA hyperbolic exponent (variable used in LGA)
Ng =Corey gas exponent
No = Corey oil exponent
OOIP = Original oil in place
PLE = Power law exponential
P-T = Pressure-temperature
PVT = Pressure-volume-temperature
q = Production oil rate at time t, stb/day or stb/time
= Rate “intercept” t=0, stb/d (variable used in PLE)
= Oil rate at Day 1(variable used in Duong method)
= Oil rate at infinite time (variable used in Duong method

167
Sl = Total liquid saturation
Sg = Gas saturation
Sw = Water saturation
STB = Stock tank barrel (same as stb)
stb/d = Stock tank barrel/day
SRV = Stimulated reservoir volume
TOC = Total organic content
TRRC = Texas railroad commission
t (a, m) = Time function based on Equation 3.11 (variable used in Duong method)
Z-factor = Gas compressibility

168
Appendix
Table A.1: Table of decline exponents (b) when Arps’ hyperbolic and LAHA models are applied
LAHA (0<b<1)
Production Decline Decline Improved
Profile Well's Well Lease Exponent Exponent Decline
Counts County Names (0<b<1) (b>1) Exponent
Giesenschlag
BURLESON
1 W. H. " 0.89
AB
2 Childers 1.01 0.93
Giesenschlag-
3 Groce 1.12 0.91
4 Simms 1.23 0.89
LEON
5 Easterling 1.42 0.97

6 Donaho Unit 0.50
7 Berry 0.57
8 Coates 0.72
Muenchow
9 Unit 1.06 0.76
10 Cannon 1.86
KARNES
Brysch
11 Jonas Unit 1.56
Kathryn
12 Kealey Unit 1.25 0.84
13 CEF 3.05
14 Gilley Unit 0.85
15 Yosko Unit 0.59
16 Carter Unit 1.57
17 Hatch 1.07 0.91
DIMMIT
18 Votlz Unit 0.97

Fog
19 Mount 1.76
20 JBGS 1.05 0.43
169
Table A.1: Continued:

LAHA (0<b<1)
Production Decline Decline Improved
Profile Well's Well Lease Exponent Exponent Decline
Counts County Names (0<b<1) (b>1) Exponent
21 Mostyn 1.22 0.94
22 Holmes 0.97
23 Georg 0.07
24 Parr 0.72
GONZALES
25 Bozka 1.52 0.89

26 S. Duderstadt 1.09 0.91
27 Cinco Ranch 0.98
Koenning
28 Unit 1.20 0.91
29 Gonzo Hunter 1.20 0.56
30 Perkins 0.62
31 Otto 1.19 0.94
K.M
32 Ranch 1.41 0.97
Mustang
33 Ranch C 4.50
ZAVALA
Alpha
34 Ware 1.99
35 Felps 1.07 0.56
Howett
36 0.58
Addison
37 Avery 1.21 0.78
Reser-
BRAZOS
Sanders
38 Unit 1.11 0.87
Hullabaloo
39 1.45 0.99
Fenn
LEE Ranch
40 Unit 3.01
170
Table A.2: EUR, remaining reserves and well’s life to 2stb/day for Eagle Ford horizontal wells
Reserves in
Barrels
& Time to
Well's Well's Lease Reach 2 Arps’ Arps’ Arps’
County Names Barrels/day PLE Exponential Harmonic Hyperbolic LGA LAHA Duong
EUR 8915 - 31606 33162 28795 29416 30976

GIESENSCHLAG Remaining
W. H. " Reserves 907 - 5117 6673 2306 2927 4487
Remaining
Life 326 - 1796 1654 863 1090 1234
BURLESON
EUR 16722 16828 28971 29391 22014 27091 56199

AB Remaining
CHILDERS Reserves 5864 5970 18113 18533 11156 16233 45341
Remaining
Life 1437 1443 4874 5003 2876 4466 12285
EUR 45996 - 108595 121748 85119 99202 111905

GIESENSCHLAG- Remaining
GROCE Reserves 12598 - 49131 62284 25655 39738 52441
Remaining
Life 2052 - 10727 14026 5205 8701 8800
EUR 16590 - 24044 30047 21734 22906 26399

Remaining
SIMMS Reserves 1836 - 9290 15293 6980 8152 11645
Remaining
Life 668 - 2603 4125 1887 2291 2123
EUR 4513 - 5686 7204 5248 5692 7057
LEON
Remaining
EASTERLING Reserves 949 - 2122 3640 1684 2128 3493
Remaining
Life 291 - 725 1212 586 752 900
EUR 45959 - - 65300 66938 76359 122874
DONAHO Remaining
UNIT Reserves 17165 - - 36506 38144 47565 94080
Remaining
Life 878 - - 3035 3435 4937 4391
EUR 37971 - 107057 123119 85566 99506 95714

Remaining
HATCH Reserves 12247 - 47095 63157 25604 39544 35752
Remaining
Life 1606 - 9293 11557 4521 7729 2080
EUR 35333 - - 61152 50099 61893 84619

VOTLZ Remaining
UNIT Reserves 4192 - - 30011 18958 30752 53478
Remaining
DIMMIT
Life 632 - - 5523 3342 6194 17794
EUR 36842 - 59641 215465 55588 54195 62045

FOG Remaining
MOUNT Reserves 7681 - 30480 186304 26427 25034 32884
Remaining
Life 742 - 6257 45950 4516 10909 4153
EUR 72692 - - 127734 90817 77105 131864

Remaining
JBGS Reserves 25448 - - 80490 43573 29861 84620
Remaining
Life 1086 - - 13410 5832 2903 11463
EUR 54052 - 60422 1348154 101329 1017639 111826

FENN Remaining
LEE
RANCH Reserves 9877 - 41507 1329239 82414 998724 92911

UNIT Remaining
Life 1509 - 8959 449331 16561 330025 12457
171
Table A.2: Continued

Reserves in
Barrels
& Time to
EUR 113208 - - 151792 125904 125086 128690

Remaining
BERRY
Reserves 858 - - 39442 13554 12736 16340
Remaining Life 444 - - 2900 1982 1850 2874
EUR 129078 - - 194380 162457 179139 249748

Remaining
COATES
Reserves 27148 - - 92450 60527 77209 147818
Remaining Life 1117 - - 9860 5861 8433 18797
EUR 48885 - 71910 89673 53138 57098 57860

MUENCHOW Remaining
UNIT Reserves 8791 - 31816 49579 13044 17004 17766
Remaining Life 195 - 5122 7185 2228 3100 672
EUR 238393 - 671093 8657490 672319 349323 644186

Remaining
CANNON
Reserves 130145 - 505544 8491941 506770 183774 478637
Remaining Life 3830 - 59201 2008923 38291 12350 7787
EUR 114286 - 264765 983609 427641 144368 233354

BRYSCH
Remaining
JONAS
Reserves 56269 - 180422 899266 343298 60025 149011
UNIT
Remaining Life 2210 - 23566 177192 36450 4522 1338
EUR 74118 73881 181062 308584 109331 145672 187824

KARNES
KATHRYN Remaining
KEALEY Reserves 23226 22989 130170 257692 58439 94780 136932
UNIT
Remaining Life 1332 1262 18189 43213 5471 11653 5719
EUR 266087 - 802971 1630304019 910953 479021 2315763
CEF Remaining
Reserves 155230 - 659561 1630160609 767543 335611 2172353
Remaining Life 3892 - 74578 548209395 74199 24844 44016
EUR 19250 - - 35558 303815 1117057 87412

GILLEY Remaining
UNIT Reserves 8763 - - 25071 293328 1106570 76925
Remaining Life 1291 - - 4896 76385 382972 13478
EUR 166491 - 161924 224487 203877 213767 260796

YOSKO Remaining
UNIT Reserves 16309 - 11742 74305 53695 63585 110614
Remaining Life 850 - 11812 5801 4947 6302 4041
EUR 115278 - 272867 1012169 432453 158008 247024

CARTER Remaining
UNIT Reserves 36553 - 194142 933444 353728 79283 168299
Remaining Life 1688 - 25889 191179 37713 6448 3217

172

Reserves in Barrels
Well's Well's Lease & Time to Reach 2 Arps’ Arps’ Arps’
EUR 43801 - 86403 119258 62536 82340 108520

MOSTYN Remaining Reserves 11130 - 53732 86587 29865 49669 75849
Remaining Life 1346 - 10118 17558 4655 9171 9778
EUR 126992 - 162311 170451 100998 126946 142343

HOLMES Remaining Reserves 63944 - 99263 107403 37950 63898 79295
Remaining Life 960 - 14023 8813 2500 5883 2944
EUR 10890 - - 11033 11140 10926 11295

GEORG Remaining Reserves 2145 - - 2288 2395 2181 2550
Remaining Life 338 - - 365 466 484 926
EUR 11017 - - 14437 12352 12466 15313

PARR Remaining Reserves 1778 - - 5198 3113 3227 6074
Remaining Life 311 - - 1086 759 791 1702
EUR 27919 - 54416 142638 42643 51054 49172

BOZKA Remaining Reserves 7755 - 28386 116608 16613 25024 23142
Remaining Life 988 - 5291 24970 2613 4503 1275

GONZALES
EUR 56522 - 147211 196735 115295 139452 119047

S.
DUDERSTADT Remaining Reserves 14559 - 72417 121941 40501 64658 44253
Remaining Life 1065 - 11412 17007 5008 9683 721
EUR 40546 - - 90723 63144 70705 75906

CINCO
RANCH Remaining Reserves 3197 - - 45687 18108 25669 30871
Remaining Life 503 - - 7246 2999 4663 1042
EUR 27696 27113 50445 65699 37065 46753 57770

KOENNING
UNIT Remaining Reserves 8171 7588 30920 46213 17579 27266 36225
Remaining Life 1017 1067 6322 9909 3112 5481 8907
EUR 64796 - - 141216 62834 64044 73870

GONZO
HUNTER Remaining Reserves 16455 - - 92875 14493 15703 25529
Remaining Life 508 - - 16906 2314 2554 4459
EUR 50388 - 83345 65357 59918 62989 73930

PERKINS Remaining Reserves 2485 - 35442 17454 12015 15086 26027
Remaining Life 546 - 45202 2353 2265 2924 1617
EUR 24932 - 63274 91498 53775 62347 57675

OTTO Remaining Reserves 6897 - 26594 54818 17095 25667 20995
Remaining Life 947 - 5436 9945 3131 5201 850

173

Reserves in
Barrels
& Time to
EUR 24706 - 50994 114960 45298 51049 44888

K.M Remaining
RANCH Reserves 7534 - 28352 92318 22656 28407 22246
Remaining
Life 632 - 4841 17169 3150 4870 606
EUR 146185 - 376235 5981256282 138455648 2872696626 1708562

MUSTANG Remaining
RANCH C Reserves 108957 - 327445 5981207492 138406858 2872647836 1659772
Remaining
Life 6695 - 47698 2326186769 35495642 1103119307 66823
EUR 52147 - 108824 478005 89691 77868 168091

ALPHA Remaining
WARE Reserves 27584 - 84261 453442 65128 53305 143528
ZAVALA
Remaining
Life 1722 - 13624 1772208 6539 6959 9579
EUR 52147 - - 155600 73713 76968 273177

Remaining
FELPS Reserves 19374 - - 126288 44401 47656 243865
Remaining
Life 1631 - - 20525 5009 5585 21664
EUR 68404 - - 98311 87739 87975 176165

Remaining
HOWETT Reserves 29655 - - 47087 36515 36751 124941
Remaining
Life 1095 - - 4627 3669 3706 16945
EUR 33879 33879 75666 100021 48495 59793 220122

ADDISON Remaining
AVERY Reserves 16800 16800 58587 82942 31416 42714 203043
Remaining
Life 2033 2033 11143 17164 4862 7451 30067
EUR 5511 - - 8022 8673 9073 9891

RESER- Remaining
SANDERS Reserves 644 - - 621 1272 1672 2490
UNIT
Remaining
BRAZOS
Life 320 - - 1988 965 1260 2384
EUR 44298 40664 72974 152627 69446 84623 94833

Remaining
HULLABALOO Reserves 11076 10690 43000 108402 25221 40398 50608
Remaining
Life 1467 1862 9629 13469 5166 9223 7499

Decline Curve Enhanced Shale Oil Recovery Based On Data: Analysis and Eagle Ford Shale

Uploaded by

Copyright:

Available Formats

Decline Curve Enhanced Shale Oil Recovery Based On Data: Analysis and Eagle Ford Shale

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

Decline Curve Enhanced Shale Oil Recovery Based On Data: Analysis and Eagle Ford Shale

Uploaded by

Copyright:

Available Formats

DECLINE CURVE ANALYSIS AND ENHANCED SHALE OIL RECOVERY BASED ON

EAGLE FORD SHALE DATA

_..,Dr. Abhijit Dandekar

EAGLE FORD SHALE DATA

Presented to the Faculty

of the University of Alaska Fairbanks

in Partial Fulfillment of the Requirements

for the Degree of

Dieudonne K. Delaihdem, B.S.

To God is the Glory

Title Page …………………………………………………………………………………………iii

DEDICATION ............................................................................................................................... vii

LIST OF FIGURES ...................................................................................................................... xiii

LIST OF TABLES ....................................................................................................................... xxii

ACKNOWLEDGEMENT .......................................................................................................... xxiv

Chapter 1 Introduction ..................................................................................................................... 1

1.1 Shale Oil Reservoirs ............................................................................................................ 1

1.2 Shale Oil- What is it? ........................................................................................................... 2

1.3 Shale Play Development ...................................................................................................... 3

1.4 Purpose and Scope of Study ................................................................................................ 4

Chapter 2 Literature Review ............................................................................................................ 6

2.0 Introduction .......................................................................................................................... 6

2.1 Geology ................................................................................................................................ 7

2.2 Decline Curve Analysis ..................................................................................................... 11

2.2.1 Arps Decline Curve Analysis ..................................................................................... 12

2.2.2 Power Law Exponential Analysis .............................................................................. 15

2.2.3 Logistic Growth Analysis .......................................................................................... 16

2.2.4 Duong’s Method ........................................................................................................ 17

2.3 Analyses of the Eagle Ford Shale ...................................................................................... 19

2.3.1 Reservoir Fluids ......................................................................................................... 19

2.3.2 Natural Fractures ........................................................................................................ 20

2.3.3 Reservoir Stimulation and Wellbore Completions .................................................... 22

2.3.4 Simulation .................................................................................................................. 25

3.0 Introduction ........................................................................................................................ 36

3.1.1 Data Preparation before the Analysis......................................................................... 37

3.1.2 Arps’ Decline Models Applications ........................................................................... 37

3.1.3 Power Law Exponential (PLE) Analysis Application................................................ 38

3.1.4 Logistic Growth Analysis (LGA) Application ........................................................... 39

3.1.5 Duong’s Decline Curve Analysis Method ................................................................. 40

3.1.6 Logistic Arps’ Hyperbolic Approach (LAHA) .......................................................... 41

3.2 General Results .................................................................................................................. 42

3.2.1 Data Preparation before the Analysis ......................................................................... 42

3.2.2 Arps’ Decline Curve Analysis ................................................................................... 45

3.2.3 Power Law Exponential Analysis (PLE) ................................................................... 48

3.2.4 Logistic Growth Analysis (LGA) .............................................................................. 51

3.2.5 Duong’s Decline Curve Analysis Method ................................................................. 52

3.2.6 Logistic Arps Hyperbolic Approach (LAHA) ........................................................... 57

3.3 Discussion of General Results ........................................................................................... 60

4.0 Introduction ........................................................................................................................ 63

4.1 Burleson County ................................................................................................................ 63

4.1.1 Introduction ................................................................................................................ 63

4.2 Leon County....................................................................................................................... 70

4.2.1 Introduction ................................................................................................................ 70

4.3 Brazos County.................................................................................................................... 76

4.3.1 Introduction ................................................................................................................ 76

4.4 Lee County ......................................................................................................................... 81

4.4.1 Introduction ................................................................................................................ 81

4.5 Gonzales County ................................................................................................................ 84

4.5.1 Introduction ................................................................................................................ 84