Gopher PDF

A COMPARISON OF HYDRAULIC FRACTURE MODELING WITH DOWNHOLE
AND SURFACE MICROSEISMIC DATA IN A STACKED FLUVIAL PAY SYSTEM
by
Nur Azlinda Mohammad

A thesis submitted to the faculty and the Board of Trustees of the Colorado
School of Mines in partial fulfillment of the requirements for the degree of Master of
Science (Petroleum Engineering).
Golden, Colorado
Date ____________________
Signed: __________________________
Nur Azlinda Mohammad
Approved: __________________________
Dr. Jennifer L. Miskimins
Thesis Advisor
Golden, Colorado
Date ____________________
Approved: ____________________
Dr. Ramona Graves
Professor and Department Head,
Petroleum Engineering
ii
ABSTRACT
The work in this project concentrates on comparing the results of hydraulic
fracture mapping and modeling using two microseismic monitoring techniques:
downhole and surface arrays. There were three objectives achieved in this project in order
to obtain the resulting comparisons. The first objective was to develop detailed post-
treatment models of the hydraulic fracturing treatments in the subject well, Well D1,
which was monitored with downhole microseismic data. The second objective was to
develop detailed post-treatment models of the hydraulic fracturing treatments in the
subject well, Well S1, which was monitored with surface microseismic data. The final
objective of this project was to determine the match characteristics of the downhole and
surface microseismic data to hydraulic fracture models developed for both Wells D1 and
S1. Comparisons of the match characteristics of the multiple inputs were then developed.
In this study, GOHFER™, a fully three-dimensional fracture simulator was used
to build hydraulic fracture stimulation models that were integrated with microseismic
events detected during actual hydraulic fracture treatments on Wells D1 and S1.
GOHFER™ uses data from actual hydraulic fracturing treatments (pressure, slurry rate,
and proppant concentration) and log-derived data to supplement reservoir and mechanical
properties as input. Another software called Transform™ was employed to analyze the
microseismic event datasets, as well as, to perform the final integration process where
actual microseismic events were overlapped with the simulated fracture geometries
created in GOHFER™. Comparisons of the match results are then made between the two
techniques: downhole microseismic monitoring and surface microseismic monitoring.
iii
Input data for this project were obtained from five wells in the Greater Natural
Buttes, Uinta basin, Utah. Two wells, Wells D1 and S1, were stimulated with hydraulic
fracturing treatments where each well was monitored by geophone arrays that were
placed at different locations. Downhole geophone receivers were employed in an
observation well, Well D2, to monitor microseismic activities in the treatment well, Well
D1. For the other treated well, Well S1, geophone receivers were placed on the ground
surface surrounding the treatment well to monitor microseismic events in the subsurface.
Multiple log data were provided for the treated wells including the sonic log from which
rock mechanical properties were calculated.
Ten fracture models were built using GOHFER™ for five stages each from Wells
D1 and S1. Pressure matching procedures were performed to achieve the final simulated
fracture geometries. The integration process took place by utilizing data from two
sources: microseismic mapping during actual hydraulic fracturing treatment; and fracture
profiles from simulated fracture models. Results from the integration process show good
agreement for most stages in the downhole-monitored well, whereas comparisons of
surface microseismic mapping measurements with the simulated fracture geometries
yield questionable results especially regarding microseismic event locations with respect
to depth.
iv
TABLE OF CONTENTS
ABSTRACT....................................................................................................................... iii
LIST OF FIGURES ........................................................................................................... ix
LIST OF TABLES.......................................................................................................... xvii
ACKNOWLEDGMENTS ............................................................................................. xviii
DEDICATION................................................................................................................. xix
CHAPTER 1 INTRODUCTION .........................................................................................1
1.1 Purpose of This Work ..................................................................................1
1.2 Research Objectives.....................................................................................2
1.3 Study Area (Field Location) ........................................................................2
1.4 Data Set........................................................................................................4
1.5 Recent Works and Applications...................................................................8
CHAPTER 2 LITERATURE REVIEW ............................................................................10
2.1 Geological Overview .................................................................................10
2.1.1 Depositional System ......................................................................10
2.1.2 Stratigraphy....................................................................................12
2.1.2.1 Mancos Shale Formation .................................................12
2.1.2.2 Mesaverde Group.............................................................15
2.1.2.3 North Horn Formation .....................................................15
2.1.2.4 Wasatch Formation ..........................................................16
2.1.2.5 Green River Formation ....................................................16
2.1.3 Producing Formation .....................................................................16
2.1.4 Petroleum System and Reservoir Characterization........................18
v
2.2 Microseismic Overview .............................................................................19
2.2.1 Microseismic Events......................................................................20
2.2.2 Downhole-monitored Microseismic ..............................................24
2.2.3 Surface-monitored Microseismic...................................................25
2.2.4 Software Used for Microseismic Analyses....................................29
2.3 Fracture Geometry Modeling.....................................................................29
2.3.1 2-Dimensional Models...................................................................29
2.3.2 Pseudo-3-Dimensional Models......................................................32
2.3.3 3-Dimensional Models...................................................................33
2.3.4 Net-Pressure Analysis....................................................................34
CHAPTER 3 HYDRAULIC FRACTURE MODELING .................................................36
3.1 GOHFER™ Software ................................................................................39
3.2 Modeling Process.......................................................................................41
3.2.1 Input ...............................................................................................42
3.2.2 Output ............................................................................................51
3.3 Pressure Matching Process ........................................................................54
CHAPTER 4 MICROSEISMIC EVENTS AND HYDRAULIC FRACTURE

MODEL RESULTS ...........................................................................................................56
4.1 Input Data For Hydraulic Fracture Model .................................................58
4.1.1 Logs and Log-derived Properties...................................................58
4.1.2 Formation Zone Setup....................................................................62
4.1.3 Pumping Schedule .........................................................................66
4.2 Input Data For Microseismic Model..........................................................78
4.2.1 Microseismic Event Locations For Well D1 .................................79
vi
4.2.2 Microseismic Event Locations For Well S1 ..................................94
4.3 GOHFER™ Pressure Matches ................................................................111
4.3.1 Matched Parameters.....................................................................112
4.3.2 Advanced Parameters...................................................................114
4.4 Hydraulic Fracture Model Results – Pressure Matches and Fracture

Geometries ...............................................................................................115
4.4.1 Well D1 – Stage 1 ........................................................................116
4.4.2 Well D1 – Stage 2 ........................................................................118
4.4.3 Well D1 – Stage 4 ........................................................................122
4.4.4 Well D1 – Stage 5 ........................................................................126
4.4.5 Well D1 – Stage 7 ........................................................................130
4.4.6 Well S1 – Stage 2.........................................................................130
4.4.7 Well S1 – Stage 3.........................................................................140
4.4.8 Well S1 – Stage 4.........................................................................140
4.4.9 Well S1 – Stage 5.........................................................................143
4.4.10 Well S1 – Stage 9.........................................................................146
4.5 Crosswell Tomograms .............................................................................149
4.6 Data Integration .......................................................................................155
CHAPTER 5 DATA INTEGRATION RESULTS AND DISCUSSION........................157
5.1 Discussion of Results from Downhole-monitored Well (D1) ................ 157
5.1.1 Well D1 - Stage 1.........................................................................158
5.1.2 Well D1 - Stage 2.........................................................................161
5.1.3 Well D1 - Stage 4.........................................................................165
5.1.4 Well D1 - Stage 5.........................................................................167
vii
5.1.5 Well D1 - Stage 7.........................................................................170
5.2 Discussion of Results from Surface-monitored Well (S1) ..................... 174
5.2.1 Well S1 - Stage 2 .........................................................................174
5.2.2 Well S1 - Stage 3 .........................................................................178
5.2.3 Well S1 - Stage 4 .........................................................................182
5.2.4 Well S1 - Stage 5 .........................................................................185
5.2.5 Well S1 - Stage 9 .........................................................................188
5.3 Downhole Methodology vs. Surface Methodology .................................190
CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS ....................................201
6.1 Conclusions..............................................................................................201
6.2 Recommendations and Future Work .......................................................204
NOMENCLATURE ........................................................................................................207
REFERENCES ................................................................................................................209
CONTENTS OF THE DVD-ROM..................................................................................214
DVD-ROM ............................................................................................................ POCKET
viii
LIST OF FIGURES
Figure 1.1: Flow diagram of the project objectives .............................................................3
Figure 1.2: Map location of the Uinta Basin........................................................................4
Figure 1.3: Well locations for downhole microseismic monitoring ....................................6
Figure 1.4: Well locations for surface microseismic monitoring ........................................6
Figure 2.1: Figure of late Cretaceous paleogeography ......................................................13
Figure 2.2: Stratigraphic column of the Uinta basin and its adjacent basin ......................14
Figure 2.3: An example of the Price River formation at the Kenney reservoir outcrop....17
Figure 2.4: Figure of vibrator truck creating vibration on ground surface ........................21
Figure 2.5: Particle and wave motions of the primary and secondary waves....................21
Figure 2.6: The transmission technique to create a 3-D map ............................................23
Figure 2.7: The emission technique that maps the distribution of the microseism............23
Figure 2.8: Side view of the settings for the downhole arrays...........................................25
Figure 2.9: 3-Component downhole geophone..................................................................26
Figure 2.10: Single and staked receiver geophones for downhole arrays………………..26
Figure 2.11(a): Surface microseismic mapping layouts ....................................................28
Figure 2.11(b): 3-D seismic output from the microseismic mapping process...................28
Figure 2.12: Example of microseismic events in Well D1 ................................................30
Figure 2.13: Map-view example of the downhole monitored wells (Well D1 and D2) ....31
Figure 2.14: Log-log slope interpretation for the idealized net pressure data ...................35
Figure 3.1: Wellbore stability plot for Well S1 .................................................................37
Figure 3.2: Gamma ray log for the downhole monitored well, Well D1...........................38
Figure 3.3: Actual pressure, slurry rate, and proppant concentration curves for Stage 9..40
ix
Figure 3.4: Work flow of GOHFER™ actual job data match ...........................................43
Figure 3.5: An example of the grid structure feature of GOHFER™ ...............................44
Figure 3.6: Fracture orientation determined by the principal stresses, σ1, σ2, and σ3 ........... 46
Figure 3.7: Example of strain where rock sample is deformed and compressed...............47
Figure 3.8: Plot of Young’s Modulus (Stress vs. Strain)...................................................49
Figure 3.9: Poisson’s Ratio, the ratio of lateral to axial strain...........................................50
Figure 3.10: The main screen layout of the LOGCALC™ ...............................................52
Figure 3.11: Example of WinParse™ view .......................................................................53
Figure 3.12: The HTGraph™ displaying the actual pressure, slurry rate, and
proppant concentration curves acquired during a hydraulic fracturing job ...........55
Figure 4.1: Flow chart of the integration process where fracture geometry output
from GOHFER™ models are overlapped with microseismic events …………....57
Figure 4.2: Input data imported into GOHFER™ for Stage 1 of Well D1........................67
Figure 4.7: Input data imported into GOHFER™ for Stage 2 of Well S1 ........................72
Figure 4.10: Input data imported into GOHFER™ for Stage 5 of Well S1 ......................75
Figure 4.11: Input data imported into GOHFER™ for Stage 9 of Well S1 ......................76
Figure 4.12: The Pumping Schedule tab in WinGOHFER™…........................................77
Figure 4.13: Microseismic events (blue dots) that occurred in Stage 1 of Well D1.........80
x
Figure 4.14: Microseismic events (red dots) that occurred in Stage 2 of Well D1............81
Figure 4.15: Microseismic events (aqua dots) that occurred in Stage 3 of Well D1 .........82
Figure 4.16: Microseismic events (orange dots) that occurred in Stage 4 of Well D1......83
Figure 4.17: Microseismic events (green dots) that occurred in Stage 5 of Well D1........84
Figure 4.18: Microseismic events (grey dots) that occurred in Stage 6 of Well D1..........85
Figure 4.19: Microseismic events (pink dots) that occurred in Stage 7 of Well D1..........86
Figure 4.20: Actual hydraulic fracturing data for Stage 1 of Well D1 ..............................87
Figure 4.27: Microseismic events (blue dots) that occurred in Stage 2 of Well S1 ..........95
Figure 4.28: Microseismic events (red dots) that occurred in Stage 3 of Well S1 ............96
Figure 4.29: Microseismic events (aqua dots) that occurred in Stage 4 of Well S1..........97
Figure 4.30: Microseismic events (orange dots) that occurred in Stage 5 of Well S1 ......98
Figure 4.31: Microseismic events (green dots) that occurred in Stage 6 of Well S1 ........99
Figure 4.32: Microseismic events (grey dots) that occurred in Stage 7 of Well S1 ........100
Figure 4.33: Microseismic events (pink dots) that occurred in Stage 8 of Well S1 ........101
Figure 4.34: Microseismic events (light green dots) that occurred in Stage 9 of
Well S1.................................................................................................................102
Figure 4.35: Actual hydraulic fracturing data for Stage 2 of Well S1.............................103
xi
Figure 4.43: Graph of matched GOHFER pressure (aqua) in comparison to

bottomhole pressure (black) and the actual pressure (red) measured
during the fracturing job of Stage 1 for Well D1.................................................119
Figure 4.44: Graph showing the matched GOHFER pressure (aqua) in comparison
to the actual pressure (red) measured during the fracturing job of Stage 1
for Well D1 ..........................................................................................................120
Figure 4.45: GOHFER output showing proppant concentration (lb/ft2) inside the
fracture of Stage 1 for Well D1............................................................................121
to bottomhole pressure (black) and the actual pressure (red) measured
for Well D1 ..........................................................................................................124
for Well D1 ..........................................................................................................128
xii
to bottomhole pressure (black) and the actual pressure (red) measured during
the fracturing job of Stage 5 for Well D1 ............................................................131
for Well D1 ..........................................................................................................132
for Well D1 ..........................................................................................................135
for Well S1...........................................................................................................138
fracture of Stage 2 for Well S1 ............................................................................139
for Well S1...........................................................................................................141
for Well S1...........................................................................................................144
for Well S1...........................................................................................................147
xiii
for Well S1...........................................................................................................150
Figure 4.68: Map view of two crosswell tomogram profiles between Wells S1-S2
and S1-S3 .............................................................................................................152
Figure 4.69: The two crosswell tomogram profiles (Profile 1 on the left-hand
side and Profile 2 on the right-hand side) with Well S1 as the axis ....................153
Figure 4.70: Profile 1 of the crosswell tomogram between Wells S1 and S2 .................154
Figure 4.71: Profile 2 of the crosswell tomogram between Wells S3 and S1 .................154
Figure 4.72: Side view of both crosswell tomogram profiles showing the location
of the microseismic events with respect to Profiles 1 and 2 ................................155
Figure 5.1: Profile view of microseismic events (blue dots) that occurred during
Stage 1 hydraulic fracturing treatment in Well D1..............................................159
Figure 5.2: Map view of microseismic events (blue dots) that occurred during
Figure 5.3: Profile view of microseismic events (red dots) that occurred during
Figure 5.4: Map view of microseismic events (red dots) that occurred during
Figure 5.5: Profile view of microseismic events (orange dots) that occurred during
Figure 5.6: Map view of microseismic events (orange dots) that occurred during
Figure 5.7: Profile view of microseismic events (green dots) that occurred during
Figure 5.8: Map view of microseismic events (green dots) that occurred during
xiv
Figure 5.9: Profile view of microseismic events (pink dots) that occurred during
Figure 5.10: Map view of microseismic events (pink dots) that occurred during
Figure 5.11: Profile view of microseismic events (blue dots) that occurred during
Stage 2 hydraulic fracturing treatment in Well S1...............................................175
Stage 2 hydraulic fracturing treatment in Well S1 along with the
simulated fracture profile in the background .......................................................178
Figure 5.14: Profile view of microseismic events (red dots) that occurred during
Figure 5.17: Profile view of microseismic events (aqua dots) that occurred during
Figure 5.18: Map view of microseismic events (aqua dots) that occurred during
Figure 5.20: Profile view of microseismic events (orange dots) that occurred during
xv
Figure 5.23: Profile view of microseismic events (green dots) that occurred during
Figure 5.26: The microseismic moment magnitude with respect to true vertical depth..194
Figure 5.27: Hearing distance of microseismic events for Well D1................................195
xvi
LIST OF TABLES
Table 4.1: Well D1 Input Data of Hydraulic Fracture Model............................................61
Table 4.2: Well S1 Input Data of Hydraulic Fracture Model ............................................62
Table 4.3: Microseismic Event Locations for Well D1 .....................................................94
Table 4.4: Microseismic Event Locations for Well S1....................................................111
Table 5.1: Comparisons of Fracture Geometries Between Fracture Model and

Fracture Mapping.................................................................................................197
xvii
ACKNOWLEDGMENTS
First and foremost, I would like to especially thank my advisor, Dr. Jennifer
Miskimins, who has walked with me the many extra miles in order to ensure the
completion of this work. Thank you for giving me the opportunity to learn and grow
intellectually, and providing me a great learning environment and experiences while
working in this project. There have been many ups and downs prior to the completion of
this work, but you have been there for me on multiple occasions, and for that, I thank you
for not giving up on me. There are no words that could describe my appreciation to you.
Along the line, I would take this opportunity to thank the three important people
who also served on my committee. Dr. Donna Anderson, your help and guidance
throughout the learning process (which started way back in Massadona 2006) made me
fall in love with geology without any boundaries. You made geology so interesting, even
an engineer couldn’t resist it. Mr. Robert Kidney, thank you for your patience in
entertaining my multiple weekly emails. The meetings and email attachments sure helped
me find the path to take while working in this project. Dr. Manika Prasad, thank you for
always being available when I had the rush to ask random questions on logs, geophysics,
or places to go in San Francisco. I truly appreciate the open-door policy and perhaps one
of these days, you will finally get the chance to push me off the whitewater raft into the
river.
A special thanks goes to Dr. Bob Barree (from Barree & Associates) for the time
he spent to guide me in building and calibrating my hydraulic fracture models. I would
also like to thank Ms. Amelia Webster (from Transform Software) who went beyond the
xviii
scope of her duty in helping me to effectively use the software and help me integrate the
input data provided in this project. Not to forget, the new head of department for the
Petroleum Engineering Department who will be signing my silver diploma, Dr. Ramona
Graves, thank you for your ongoing support and words of wisdom that you usually share
with me at the hallway.
To the operating companies, EOG Resources, Inc. and Anadarko Petroleum
Corp., I am indebted for the support given in terms of input data that were provided to
make this project possible. I am also very thankful to the FAST (Fracturing, Acidizing,
Stimulation Technology) Research Consortium run by Dr. Jennifer Miskimins who
generously funded the growth of this research project. Also, I would like to thank Ms.
Denise Winn-Bower from the Petroleum Engineering Department who tirelessly assisted
me in multiple occasions, even those that are beyond her call of duty.
A special note of thanks goes to my family and friends in the two continents of
Asia and North America (and the Caribbean as well). Thank you for the never ending
support which made it possible for me to proceed to the ‘Finish’ line. This work has been
a long time coming.
xix
DEDICATION
I would like to dedicate this work to my family back in Malaysia. To my mother,
Hasnah Abu Hassan, who never stops believing in me and my endless dreams. My
brother, Asyraff Syazwan, the smart little boy in the family whom I love so dearly and
dreams to become a great scientist. My father, Mohammad Adnan, who never tires of
supporting me and providing me with great assistance. My sister, Julia Inne, who took
one for the team to look after the family while I’m gone. My better half, Paul Okonkwo,
deserves special mention for his faith, patience, and confidence in me and my
capabilities. Lastly, this is dedicated to my six-year advisor, Dr. Jennifer Miskimins, who
never failed to provide me countless support, gives me stability and answers to my many
questions, and patiently watched me bloom from an 18-year old freshman at this college
to becoming the person that I am today. Without the support, faith, and love from my
loved ones, this work wouldn’t come out to be what it is today.
xx
1
CHAPTER 1
INTRODUCTION
This chapter introduces the purpose and scope of this study, research
objectives, the area of study, the data sets available and used, and ends with
potential applications.
1.1 Purpose of This Work
This project focuses on analyzing processed surface and downhole
microseismic data and constructing hydraulic fracture models to determine the
match characteristics of the fracturing treatments. The microseismic data was
acquired from passive seismic events recorded while monitoring hydraulic
fracture treatments. These works are crucial in verifying the reliability of surface
microseismic technology in detecting and measuring microseismic events
resulting from stress changes corresponding to hydraulic fracture treatments
occurring in the subsurface, despite vertical limitation concerns. For comparison
purposes, downhole microseismic data sets from a more conventional method of
fracture mapping is included in this project to enable the analyses of the surface
fracture mapping method with respect to the more common and conventional
downhole method of microseimic imaging. Methodologies to model hydraulic
fracturing treatments using both microseismic techniques are developed in this
work.
2
1.2 Research Objectives
This project promotes the significance of multiple inputs integration in a
fracture mapping and modeling task. The following are the main project
objectives:
1. Develop detailed post-treatment models of the hydraulic fracturing
treatments in the subject wells with downhole microseismic data;
2. Develop detailed post-treatment models of the hydraulic fracturing
treatments in the subject wells with surface microseismic data; and,
3. Determine the match characteristics of the downhole microseismic
with the surface microseismic data and the hydraulic fracturing
treatments models as an integration of multiple engineering and
geologic inputs.
As shown in Figure 1.1, an integration of post-treatment hydraulic
fracturing models with microseismic data is performed for both surface and
downhole monitoring methods to meet Objectives 1 and 2, respectively. Upon the
completion of both objectives, match characteristics for the surface and downhole
monitoring methods are analyzed to meet Objective 3.
1.3 Study Area (Field Location)
The study area is located in the Uinta Basin in eastern Utah. There are
several hydrocarbon producing fields within the Uinta Basin including the Greater
Natural Buttes (GNB) Field where the data for this project originated from
(Figure 1.2). Additional information about the study area is provided in Chapter 2.
3
Surface
Monitoring
Of
Hydraulic
Fractures
Objective #1 Objective #2
Hydraulic Downhole Hydraulic Surface

Fracture Monitored Fracture Monitored
Treatment Microseismic Treatment Microseismic
Models Data Models Data
(Well D1) (Well D2) (Well S1) (Well S2 & S3)
Match Match
Characteristics Characteristics
For Downhole For Surface
Data Data
Objective #3
Conclusions
And
Recommendations
Figure 1.1: Flow diagram of the project objectives as an integration process of

multiple inputs.
4
Figure 1.2: Map location of the Uinta Basin, located in the northeastern part of
Utah (USGS, 2002).
1.4 Data Set
A data set has been provided by EOG Resources, Inc. and Anadarko
Petroleum Corp. for the purposes of this study. Details regarding the geology
background of the field are acquired through literature review. This project uses a
set of data from five producing wells in the GNB field where the wells are located
in two adjacent sections (Figures 1.3 and 1.4). Available information is
summarized as follows:
• For the downhole microseismic monitored well, D1, the data available is
as follows:
o Well D1 well data (hydraulically fractured well)
Open hole and cased hole log suites; and,

5
Pump curves from hydraulic fracturing treatments for
Stages 1 through 7.
o Well D2 well data (monitoring well)
Downhole seismic stimulation monitoring (Stages 1
through 7).
• For the surface-monitored hydraulically fracturing well, S1:
o Well S1 well data (hydraulically fractured well)
Open hole and cased hole log suites;
Schlumberger’s Sonic Scanner™ and Ultrasonic Borehole
Imager™ (UBI) logs;
Pump curves from hydraulic fracturing treatments for
Stages 2 through 9; and,
900 feet of core over four intervals of reservoir cut with oil
-based mud.
o Well S2 well data (monitoring well)
Schlumberger’s Sonic Scanner™ and Formation Micro
Imaging™ (FMI) logs.
o Well S3 well data (monitoring well)
Schlumberger’s Dipole Sonic Scanner™.

6
Well
645’ D2
Well
1 mile
D1
• Log suites
• Sonic
•Log suites Scanner
•9 Stages
Pump Curves
Figure 1.3: Well locations for downhole microseismic monitoring

(modified from Kidney, 2007).
• FMI Log •Sonic Scanner

• Sonic Well • 8 Stages
Scanner S2 Pump Curves
• UBI Log
2180’
• 900’ core
1 mile
Well
S1
1580’
•Dipole Sonic Well

Log
S3
Figure 1.4: Well locations for surface microseismic monitoring

(modified from Kidney, 2007).
7
o Surface seismic stimulation monitoring for Well S1 treatments,
Stages 2 through 9.
o Crosswell seismic data
Profile 1 1580 feet separation between Well S1 and S2
(refer to Figure 1.4); and,
Profile 2 2180 feet separation between Well S1 and S3.
o 3-Dimensional seismic covering 80 miles square area
Final migrated data; and,
Volume velocity and curvature attributes for fault/fracture
mapping.
• Software:
o GOHFERTM software is used to simulate the stages of hydraulic
fracturing activity for both surface and downhole monitoring wells.
This software is provided by Barree & Associates through the
FAST Consortium at the Colorado School of Mines.
o Transform™ Software Version 2.2 is used to analyze the
microseismic data in 3-Dimensions and 4-Dimensions with respect
to time. The same software is also used in the input-data
integration process. This software is provided by Transform™
Software through the FAST Consortium at the Colorado School of
Mines.
8
1.5 Recent Work and Applications
Microseismic technology is a developing technology that is still
undergoing significant technological developments throughout the years since it
was initially introduced commercially in the 1980’s. The geoscientists in the
industry are making improvements that are beneficial to the quality of this
mapping technology and are acquiring better and more reliable microseismic data.
Some recent works that have been published are as follow:
• Applicability of surface microseismic monitoring approach in the absence
of offset wells (Abbott et al., 2007);
• Passive seismic monitoring to overcome environment and lack of area
issues of active seismic (Duncan, 2004); and,
• Microseismicity is controlled by structural mechanisms associated with
the fracture and the reservoir. Also, shear stress occurring at fracture tip
provides a mechanism for microseisms to reflect length and possible
height of the fracture (Warpinski et al., 2001).
The microseismic mapping methodology is commonly used to monitor the
growth of hydraulic fractures that are created through a hydraulic fracturing
treatment as well to monitor reservoir performance, for instance, in reservoirs that
undergo secondary recovery process. Fracture mapping (microseismic mapping),
provides a direct measurement of fracture geometry created from a hydraulic
fracturing treatment. Hydraulic fracture models, on the other hand, give the ability
to predict how changes to a fracture treatment would alter fracture geometry.

9
Combining both the direct measurements from microseismic mapping and the
geometries obtained from the hydraulic fracture models, calibrated fracture
models with good predictive capabilities can be created.

10
CHAPTER 2
LITERATURE REVIEW
This chapter summarizes the geologic environment of the studied field
(GNB) and the three major technologies associated with this project:
microseismic monitoring technology, fracture diagnostic methods, and hydraulic
fracture modeling. Also, the area of interest within GNB as well as the monitoring
techniques used during the hydraulic fracturing treatments on the two treated
wells is emphasized in this literature review section.
2.1 Geological Overview
The development of the Uinta basin was initiated during the early Tertiary
(Paleocene and/or Eocene) time (Osmond et al., 1968). This basin was originally
part of a Cretaceous foreland basin. It is asymmetrical and bordered by the Uinta
Mountains on the north side, the Wasatch Mountains on the west side, and the San
Rafael Swell on the southwest side (Spencer, 1987 and Osmond et al., 1968). In
recent times, the Uinta basin took shape as a result of the uplift of its margin
which resulted in a height increment of 3,000 to 6,000 ft (Slim, 2007).
2.1.1 Depositional System
The depositional environment of the studied section in the Uinta basin is a
fluvial-deltaic type of deposit which is mainly indicated by the heterogeneous

11
nature of the stacked sands throughout the formations of interest. According to
Galloway and Hobday (1996), a fluvial environment is described as a “continental
depositional environment where sediments are carried and transported by running
water to be later deposited into lacustrine (lake) or marine (ocean) basins”. There
are two types of stream systems that can be developed on different slopes with
sections that have high sinuosity (meandering streams) and low sinuosity (braided
streams).
Some of the fluvial types of deposits mentioned by Galloway and Hobday
(1996) include channel fill (lag; different types of bars), channel-margin deposits
(crevasse splays; levees), and floodbasin deposits (floodplain, interchannel lakes).
The probable type of deposition for the producing sand bodies studied in this
project is from point-bar depositions along a meandering river (high sinuosity
environment) (Kidney, 2007).
The basin is an asymmetric structure that underwent subsidence due to the
Laramide orogeny during the late Cretaceous-Eocene (Osmond et al., 1968). It
has a steep northern flank and a gentle southern flank where the basin is filled by
lacustrine and fluvial sedimentary rocks (Gutierrez, 2007). The uppermost
Cretaceous and lowermost Tertiary strata dip at 40 to 60 towards the north
(Gutierrez, 2007). Also, Morgan et al. (2005) concluded that this area is
dominated by 30 to 50 NE regional dip into the basin that are interrupted by a few
E-W- and NW-SE-trending faults. Most faults that are exposed at surface in the
southwest of Uinta Basin are grouped into two types: the Duchesne-graben type
12
with shallow and hingeline type of faults, and the oblique-slip faults (Morgan et
al., 2005).
2.1.2 Stratigraphy
The stratigraphic sections that contribute to the petroleum system of the
Uinta basin are discussed briefly in this section. Five formations extend from the
older marine Mancos shale to the younger Green River formation. The Mesaverde
group was sourced from eroded sediment from the Sevier orogenic belt (Figure
2.1) to the west of the basin (Slim, 2007). According to Hintze (1988) and
Johnson and Roberts (2003), the sediments were “shed to the central and eastern
Utah during repetitive cycles of regressions and transgressions of the Campanian
sea”. The stratigraphic column for the Uinta basin, the Douglas Creek Arch, and
the Piceance basin is shown in Figure 2.2.
2.1.2.1 Mancos Shale Formation
In general, the Mancos Shale is formed by a thick sequence of marine
shale that interfingers with a number of sandstone tongues reflecting several
episodes of deltaic depositional environment along the margin of the Cretaceous
Interior Seaway on the western side of the basin (Gutierrez, 2007). This formation
has varying thicknesses from approximately 1,700 ft on the west side of the Uinta
basin to approximately 5,000 ft in the middle section of the basin (Gutierrez,
2007). The major members of this formation are the Tununk Shale, the Ferron
Sandstone, the Lower Blue Gate Shale Member, the Emery Sandstone, and the
13
Upper Blue Gate Shale member (Montgomery et al, 2003). The Mancos shale
lithology ranges from black organic-rich shale to a combination of siltstones and
sandy siltstones (Gutierrez, 2007).
Figure 2.1 Late Cretaceous paleogeography leading to the formation of the Uinta
basin after the movement along the Sevier orogenic belt
(modified from Blakey, 2007).
14
Wasatch
Price River
Figure 2.2: Stratigraphic column of the Uinta basin and its adjacent basin, the
Piceance basin, showing the two producing formations of this basin. The Douglas
Creek Arch, which separates the Uinta and Piceance basins, is also shown
(USGS, 2002).
15
2.1.2.2 Mesaverde Group
The Mesaverde Group is a “part of a wedge of alluvial fan and plain and
deltaic sandstones deposited as a belt along the western margin of the
epicontinental sea as it regressed” (Gutierrez, 2007). The group ranges in between
2,200 to 2,900 ft of thickness with the Star Point Formation, the Blackhawk
Formation, the Castlegate Formation, and the Price River Formation serving as
the group members of the Mesaverde (Gutierrez, 2007).
The Price River Formation is the formation in particular from which the
two treatment wells (Well D1 and S1) studied in this project produce. This
formation mainly consists of sandstone and siltstone (Guiseppe and Heller, 1998).
Due to its fluvial environment depositional system, many sandstone sheets
ranging from 20 to 100 ft thick are separated by a number of siltstone intervals
from 7 to 30 ft thick (Gutierrez, 2007).
2.1.2.3 North Horn Formation
The North Horn Formation in the Uinta basin section separates the
underlying Mesaverde Group from the overlying Wasatch Formation. This
formation consists of a distinct reddish sandstone, siltstone, and limestone beds.
The lithology of this formation ranges from “thin lacustrine shale and lime
wackestones overlain by variegated floodbasin mudstones and fine-grained fluvial
sandstone lenses with thin granule horizons” (Gutierrez, 2007).

16
2.1.2.4 Wasatch Formation
The Wasatch Formation changes thickness across the basin: the beds thin
from approximately 3,000 ft thick on the west side of GNB to approximately
1,000 ft on the east side of the area where the margin of the basin approaches
(Osmond, 1992). According to Gutierrez (2007), the lower part of this formation
consists of “varicolored shale, paludal limestones, carbonaceous shale and coals,
and conglomeratic sandstones”. Overlying these beds are “alluvial red shale and
siltstones enclosing scattered, lenticular channel-form sandstones” (Gutierrez,
2007). Prominent color changes between the beds helps to define contacts within
the formation.
2.1.2.5 Green River Formation
As a result of a lacustrine depositional environment, the Green River
Formation consists of “dark brown marlstone (oil-shale) with interbedded light
gray, oil stained water-bearing sandstones” (Gutierrez, 2007) that thickens from
2,200 ft on the southeast corner of the basin to approximately 3,800 ft in the
northeast corner. The lower part of this formation is “lacustrine ostracodal
coquina limestones, gray to tan with some oil staining” (Gutierrez, 2007) that are
of great color contrast to the underlying Wasatch Formation red beds.
2.1.3 Producing Formation
The main gas producing formations from the Uinta basin are the Wasatch
(lacustrine deposition) and Price River (fluvial deposition) formations (refer to

17
Figure 2.2). The producing formation of interest in this project is the Price River
formation from the late Cretaceous time. This reservoir is identified as a tight gas
sand reservoir with permeability in the milli- to microdarcy range and porosity as
low as six to twelve percent (Kidney, 2007).
Figure 2.3 shows an outcrop location of the Price River formation in
Colorado. In general, the lenticular shapes of sand lenses are observed at the
outcrop and are assumed to be presented in a similar manner in the subsurface.
The behavior of these discontinuous sand bodies stacking on top of each other
agrees well with the fluvial point-bar type of depositional environment. These
discontinuous characteristics require great planning of the field and reservoir
development to ensure optimum recovery of the hydrocarbon in place.
Figure 2.3: An example of the Price River formation at the Kenney reservoir
outcrop, Colorado where the lenticular shape of sand lenses (shown in yellow) is
observed (Kidney, 2007). The red box indicates sand body discontinuation trend
seen in fluvial type environment.
18
2.1.4 Petroleum System and Reservoir Characterization
The petroleum system of the Greater Natural Buttes Field was described
by Cuzella and Stancel (2006) and is summarized below:
• Trapping mechanism: Continuous gas accumulation
Discontinuous lenticular sandstone reservoir
No distinct gas-water contact
• Producing formation: Wasatch (gross thickness: 1,000-3,600 ft)
Price River (gross thickness: 3,800-5,600 ft)
• Reservoir lithology: Sandstone and shale interbedded
• Source beds: Mesaverde organic mudstones and coals
• Gas characteristics:
o Specific Gravity: 0.63
o Water-Gas-Ratio (average value): 100 bbl/mcgf
o British Thermal Unit: 1,100
o Cumulative Production (up to 2005): 1.16 TCFG
The characteristics of the Price River reservoir in particular are
summarized below, according to Cuzella and Stancel (2006):
• Porosity (average): 11%
• Permeability (average): 0.01 millidarcy
• Water Saturation (average): 0.43%
• Reservoir Pressure Gradient: 0.43 to 0.52 psi/ft
• Type of Drive: Solution gas drive

19
2.2 Microseismic Overview
Seismic and microseismic imaging technologies are commonly classified
into two categories: active and passive sources. Surface reflection seismic uses
active seismic sources. Microseismic monitoring (fracture mapping) uses passive
seismic sources caused by the rock breaking due to the injected fluid causing
pressure changes. Also, geologic features such as fault slippage can cause the
microseismic events as well. Microseismic monitoring techniques can be broadly
grouped by the location of geophones being downhole in a wellbore or located on
the ground level surface.
2.2.1 Microseismic Events
A seismic event can be generated from two types of sources: active
seismic and passive seismic. An active seismic source is an intentionally induced
seismic events for the purpose of gathering subsurface images to map the
reservoir in two dimensions (2-D) or three dimensions (3-D). The source for
active seismic can be as simple as dynamite or can also be a more sophisticated
technology such as specialized air guns or Vibroseis trucks to send seismic waves
through the ground formations (Figure 2.4). The signal waves reflected by the
different types of formations that have different rock properties and
configurations are then recorded and analyzed to construct a 2-D or 3-D model of
the geologic structure of the reservoir.
Passive seismic, on the other hand, as described by its name, does not
require intentional sources like active seismic does. This seismic imaging is done
20
using sources induced internally in a subject reservoir such as those from
hydraulic fracturing treatments rather than the specialized air guns, Vibroseis
trucks or dynamite. Smaller seismic events in the subsurface with low amplitudes
are called microseismic.
A microseismic event occurs when the rock underneath the ground surface
is fractured due to stress failure. From this failure, two body waves are released
from the “crack”. Pressure waves (P-waves) travel at the greatest velocity within
solids. Thus, they are the first waves to appear on a seismogram. They are also
called primary waves. P-waves are fundamentally pressure disturbances that
propagate through a material by alternately compressing and expanding the rock
and the particle motion is parallel to the direction of wave propagation. On the
contrary, the shear waves, which are also known as the secondary waves (S-
waves), are transverse waves that travel more slowly than P-waves. Therefore, the
S-waves appear later than P-waves on a seismogram. In S-waves, the motion of a
particle is perpendicular to the direction of wave propagation as shown in Figure
2.5. Detection of both P-waves and S-waves arrivals may provide information to
locate the source point of an event.
A magnitude is an assigned single number to quantify the amount of
seismic energy released by an earthquake or another seismic event. A local
magnitude (ML) value of two or greater is identified as an earthquake, where a
magnitude of three is potentially damage causing to surface structures
(Microseismic Inc., 2007). A local magnitude between two and negative two is
considered a micro-earthquake. It is generated by tectonic stresses and represents

21
a potential passive seismic source for mapping subsurface structures and
occurrences (Microseismic Inc., 2007).
Figure 2.4: Example of a vibrator truck creating vibration on the ground

surface where the sound waves are transmitted through the formations
and reflected to the surface and later captured by the seismographic
recording truck (University of Akron, 2009).
Figure 2.5: Particle and wave motions of the primary (P-wave) and secondary
(S-wave) waves (University of Oregon, 2008).
22
Passive microseismic imaging is divided into two categories, the first one
being the transmission technique and the second is the emission technique (Neale
and Duncan, 2003). The transmission technique observes the travel time of
seismic signals originating from micro-earthquakes to create a 3-D map as shown
in Figure 2.6 (Duncan, 2004). This technique is designed to capture bigger
earthquakes with a local magnitude of zero or greater, and it works similar to
conventional 3-D seismic imaging. Meanwhile, an ML of between negative two
and negative four is defined as microseismicity (Neale, 2007). Such microseismic
events may occur very frequently over a period of time.
The second passive microseismic imaging technique, emission, targets the
microseismic activity itself induced by the fracture treatment as the imaging
targets. This technique observes and records direct arrivals of the seismic waves
from the events and maps the distribution of the hypocenter position, the location
where the energy stored in the strain rock is released as pictured in Figure 2.7
(Duncan, 2004). The emission technique is able to capture small microseismic
events with magnitude ranging from -1 to -3 (Neale, 2007).
As the technology has evolved in the oil and gas industry, microseismic
monitoring has become one of the common methods used in subsurface hydraulic
fracture mapping along with surface and downhole tiltmeters. Similar to
tiltmeters, the microseismic mapping method is also able to provide fracture
dimensions such as height, length, and azimuth.

23
Figure 2.6: The transmission technique observes the travel time of seismic signals
originating from micro-earthquakes to create a 3-D map, similar to conventional
3-D seismic imaging (Duncan, 2004).
Figure 2.7: The emission technique targets the microseismic activity induced by
the fracture treatment (as the imaging targets). This technique observes and
records direct arrivals of the seismic waves from the events and maps the
distribution of the hypocenter position, the location where the energy stored in the
strain rock is released (Duncan, 2004).
24
In addition, microseismic mapping technology gives better resolution
along with fracture dimensions in “real-time”. The microseismic events occur as
formation rocks break, and the events signals are potentially strongest at the tips
of the fractures. Therefore, microseismicity fracture mapping provides fracture
dimensions in terms of “created fracture length” (Warpinski, 2009). In addition to
the total fracture length acquired in this mapping process, the microseismic data
can also be used to build a calibrated model to help compute effective fracture
length based on the measured total fracture length (Warpinski, 2009).
2.2.2 Downhole-monitored Microseismic
The location of the geophone receivers plays an important role in the
accuracy of data collected from seismic activity. The traditional location of the
geophones used in microseismic monitoring is in the borehole of an offset well
near the treatment well (Figure 2.8). Downhole monitoring is currently the most
common method used to produce 3-D images of hydraulic fractures.
The equipment lowered down the borehole consists of 3-component
geophones that are attached to wireline from the surface. These 3-component
instruments are designed to receive waves from the X, Y, and Z direction in the
well bore (Figure 2.9) (Shemeta et. al., 2007). Having geophones stacked on top
of each other (Figure 2.10) can help to reduce the signal-to-noise ratio, resulting
in better wave analyses and more accurate calculations (Shemeta et. al., 2007).
The downhole geophone receivers are usually run in an offset well near the
treatment well.
25
Observation Distance
Depends on Seismic Attenuation
Treatment Well
Observation Well
Perforated
Interval
Figure 2.8: A side view of the settings for the downhole microseismic mapping
method which requires the observation well to be adjacent to the treatment well.
The loss or weakening of energy as a seismic wave propagates through the earth
is called seismic attenuation. Such attenuation affects observation distance of
microseisms (modified from Roundtree, 2009).
Downhole monitoring technology can be accurate and reliable for
hydraulic fracture mapping purposes if an accurate velocity model of the reservoir
is available. The major drawbacks of downhole monitoring are the loss of
production from the offset well that is being shut in for the geophones and the
cost of drilling a monitoring well if a nearby wellbore is not available (Duncan,
2004).
2.2.3 Surface-monitored Microseismic
A relatively new technique is microseismic imaging by monitoring the
activity from the ground surface. With this method, a number of geophones are
26
Figure 2.9: 3-Component downhole geophone capturing wave movement in X, Y,

and Z directions (Albright et al., 1994).
Figure 2.10: Single and stacked receivers technology implemented in the

downhole microseismic monitoring (Shemeta et al, 2007).
27
laid on the ground surface around the treatment well. The geophones are
connected to each other using lines that are initially attached close to the
treatment well, and extent outwardly with “arms” that range from a few hundred
feet to 10,000 feet or more as pictured in Figure 2.11a (Abbott et. al., 2007). The
geophones placed on the surface monitor seismic activity before the fracturing
treatment begins, during the treatment, and after the treatment has been
completed. This is done to have an initial recording of the background noise
before the reservoir is treated (Figure 2.11b).
Today, there has been a relatively limited amount of work done on surface
microseismic monitoring technology as compared to downhole techniques.
Overcoming vertical limitations is one of the main issues in this monitoring
method. As the well gets deeper into the formation, there are more challenges of
capturing the microseismic events that are happening near the well bore, leading
to a bigger range of error and decreasing the interpretational accuracy. With
increasing vertical depth, the small amplitude waves generated by the seismic
event are too “weak” to be transmitted all the way up to the surface through
thousands feet of medium. The signals could have been released by the event, but
might not make their way to the surface to be recorded.
Eliminating the noise from the surface environment is also a big concern
(Abbott et. al., 2007). By having the geophones laid out on the ground surface,
the receivers can also pick up noise from the environment. For instance, any well
pump nearby, trucks coming in and out of the location, or any other sort of major
28
noise caused by interaction on the ground surface can degrade the seismic signal
(Warpinski, 2007).
Geophone
“arms”
Monitored
well
(a)
(b)
Figure 2.11: Surface microseismic mapping layouts (a) geophones array (“arms”)
extending radially on the ground surface (Abbott et al., 2007) and (b) 3-D seismic
output from the microseismic mapping process (Abbott et al., 2007).
29
2.2.4 Software Used for Microseismic Analyses
In order to perform microseismic analyses for this project, Transform™
software was used to upload, view, and analyze the microseismic data obtained
from the wells. This software offers flexibility in visualization with innovative
and user-friendly 2-D and 3-D visualization of a broad spectrum of the well
production data, hence, improving the understanding of the area of interest.
Figures 2.12 and 2.13 show the general layout of the 3-D viewing windows in
Transform™ once the microseismic data are uploaded into the system.
2.3 Fracture Geometry Modeling
To date, there are a variety of fracture modeling software and simulation
packages available to model fracture geometries during a hydraulic fracturing
process. The three main classes of models that have been developed over time are
the 2-dimensional (2-D), the pseudo-3-dimensional (P3D), and the well-
developed 3-dimensional (3-D) models. All are discussed in Sections 2.3.1-2.3.3
including the unique characteristics and requirements for each model.
2.3.1 2-Dimensional Models
The most common 2-D fracture models known in the industry are the
Kristianovich, Geertsma, and De Klerk (KGD) model and the model created by
Perkins and Kern which later was modified by Nordgren (PKN) (Miskimins, 2004
and 2005). According to Green (2006), when using these models fracture height
30
Monitoring
Well
Stage 7
Stage 6
Stage 5
Stage 4
Stage 3
Stage 2
Stage 1
Treatment
Well
Figure 2.12: An example of microseismic events in Well D1 shown at seven

different stages that are represented by different colors using Transform™
software with Stage 1 occurring at the deepest depth and Stage 7 is at the
shallowest depth.
31
Treatment Well
Monitoring
Well
Microseismic
events for all
stages
Figure 2.13: A map-view example of the downhole monitored wells

(Well D1 and D2) using Transform™. The different colors of
microseismic events represent the different stages of hydraulic
fracturing that took place.
32
needs to remain constant along the fracture length and can be set using
lithological boundaries. However, the two models differ in the fracture extension
occurrence concept. The PKN model states that fracture extension occurs by
rectangular extension whereas KGD chose the radial/circular extension. The
application of these models is mainly for reservoirs that have “high stress contrast
between neighboring formations where the contrasts follows lithologic
boundaries” (Green, 2006). However, the 2-D models have limited application
since such scenarios seldom apply to real reservoirs.
In general, the application of these models in heterogenous reservoirs
demands substantial manipulations to take place by making estimations of
fracture heights. These estimations are usually field-measured values or based on
previous experiences and results. Since this method requires user-input,
inaccurate fracture height estimation will cause over- or under-prediction of
height, ignoring the effects of leak-off. In this case, the fracture under
investigation will result in out-of-zone growth, later causing completion and
productivity issues (Green, 2006).
2.3.2 Pseudo-3-Dimensional Models
Pseudo-3-dimensional (P3D) models differ from the 2-D models such that
they do not require an estimate of fracture height but instead require “an input of
the minimum horizontal stress in the proposed fracture zone and bounding layers”
(Green, 2006). According to Green, a simplified depiction of fluid flow in the
fracture is implemented in the P3D models to shorten the calculation time by

33
estimating 2-D fluid flow and the pressure-width relation. As a result of
innovative inventions and improvements in computer power, the P3D models are
no longer preferred and are being substituted by the later fully 3-D models.
2.3.3 3-Dimensional Models
The new and improved 3-D models produce relatively better results with
the closest approximation to the actual hydraulic fracture growth. These models,
however, require accurate stress contrast and other reservoir data. The positive
side of the 3-D models is “the calculation of fluid flow and pressure along the
fracture uses a fully 2-D model of fluid flow to calculate the pressure” (Green,
2006). This type of calculation provides an accurate width at any point. However,
these new models are lacking on the “suitably detailed input data” to assist in
precise evaluation and future development (Green, 2006). Such input data
requires additional costs and time to the operating companies, which can be
inconvenient to the process. Fine-tuning the methodology in 3-D models by
performing thorough post-fracture evaluation will possibly provide a simple yet
effective finding to field applications. Some of the more common 3-D models that
are currently used in the industry are as follows (Miskimins, 2004 and 2005):
• FRACPRO – a model originally developed by Professor Mike Cleary at
MIT and Resource Engineering Systems, Cambridge, Mass, sponsored by
GRI;
• E STIMPLAN – a model originally developed by Mike Smith at NSI,
Tulsa;
34
• MFRAC – a model developed by Bruce Meyer of Meyer & Associates,
Natrona, Pennsylvania; and,
• GOHFER™ – a fully 3-D simulator developed by Dr. Robert Barree as
part of a PhD program at the Colorado School of Mines.
2.3.4 Net-Pressure Analysis
The net pressure (Pnet) is one of the important components that is
calculated during a hydraulic fracturing process. It is defined as the pressure in the
fracture (Pfracture) minus the closure pressure (Pclosure) (Gidley et al, 1989) as shown
in Equation 2.1:
Pnet = Pfracture - Pclosure (2.1)
From this net pressure response, the type of hydraulic fracture growth is
able to be determined. Figure 2.14 shows an idealized log-log plot of the net
pressure. Such plots can be used to analyze the behavior of fracture growth at the
subsurface and indicate its direction. From this plot, Zone 1 is interpreted as area
with restricted height growth and unrestricted extension. At 0 log-log slope, Zone
2 represents the opening of natural fractures where height growth is stable. In
Zone 3a, extension growth is restricted with two fracture wings whereas in Zone
3b, extension growth is restricted with one fracture wing. Finally, in Zone 4,
fracture height growth is presumed unstable.

35
Figure 2.14: Log-log slope interpretation for the idealized net pressure data
(Gildey et al., 1989).
36
CHAPTER 3
HYDRAULIC FRACTURE MODELING
Hydraulic fracturing treatments are generally performed to stimulate wells
in order to increase well productivity. Some common goals of hydraulic fracture
stimulations are: to maximize effective fracture length and number of zones
producing; to achieve maximum conductivity; to minimize treatment cost; to
minimize formation damage; and to add to or accelerate the hydrocarbon recovery
process.
A successfully working hydraulic fracture model must be able to assist in
explaining the results acquired from a specific job that has been pumped or be
able to predict the results of a specific job that is to be designed. For both cases of
pre-treatment and post-treatment results, computer models need precise
characterization of the studied reservoir, rock properties, and the stress state of the
area, as well as, detailed information on materials to be pumped.
For this study, hydraulic fracture treatment data from two wells, D1 and
S1 are analyzed. Well S1 is the single well that was treated with hydraulic
fracture stimulations and has available pump stage data along with surface
monitored microseismic events data for the eight stages. The pump curves
provided for this project were generated during the fracturing treatment from
Stage 2 to Stage 9 (Figure 3.1). Data from Stage 1 for Well S1 are not available
for this project. Figure 3.2 shows the gamma ray log along the depth of interest
37
Figure 3.1: Wellbore stability plot for Well S1 with open and cased hole logs
including data on the gamma ray, neutron-density, pore pressure gradient and the
lithologic section of the well. The boxes represent different stages of the
hydraulic fracture treatment starting from the bottom of the well
(modified from Kranz et al, 2007).
38
0 API unit 200
Stage 7
Stage 6
Stage 5
Stage 4
Stage 3
Stage 2
Stage 1
Figure 3.2: Gamma ray log for the downhole monitored well, D1, with
approximate lithologic representation. The boxes represent different stages of the
hydraulic fracture treatment starting from the bottom of the well. The yellow
color filling in the log represents sandstones whereas the dark brown represents
shale. The orange color, on the other hand, represents a mix of sand and shale.
39
for downhole monitored well, D1. Seven stages of hydraulic fracture treatment
were performed on this well within the 2,020 ft gross depth interval, resulting in
seven datasets of downhole monitored microseismic events.
As mentioned in Chapter 1, the Grid Oriented Hydraulic Fracture
Extension Replicator (GOHFER™) software is incorporated into this project to
enable the process of modeling the fracture job. According to the creator of the
software, Barree (2007), GOHFER™ “attempts to honor the expected
discontinuous nature of rocks with bedding places, planes of weakness or
incipient failure, and pre-existing natural fractures and fissures”. The modeling
process using this software is discussed extensively in the following sections.
Figure 3.3 depicts an example of the three curves associated with each
stage pumped, tubing pressure, slurry rate, and proppant concentration, using data
from Stage 9 of Well S1. From these input curves, pressure history matching is
done for each hydraulic fracture treatment stage.
3.1 GOHFER™ Software
GOHFER™ software is one of the more commonly used software
packages in the oil and gas industry to model complex hydraulic fractures in the
subsurface, such as the subject tight gas sand reservoir in the Uinta basin. This
software uses a finite difference grid structure to describe the entire reservoir
which allows vertical and lateral variations, single and multiple perforated
intervals, and bi-wing asymmetric fractures in order to accurately model even the
most complex reservoirs. For this project, the ability to model bi-wing
40
asymmetric fractures is critical to precisely represent the fracture patterns in the
producing fluvial-deposited sand bodies.
T ubing Pressure (psi) A Slurry Rat e (bpm) B

A Proppant Concentrat ion (lb/gal) C B C
7000 60 7
6
6000 50
5
5000 40
4000 30
3000 20
2
2000 10
1
1000 0 0
14:10 14:15 14:20 14:25 14:30 14:35 14:40 14:45
7/14/2008 7/14/2008
Time
Customer: Job Date: Ticket #:
GohWin v1.6.5
Well Desc: UWI: 10-Nov-08 13:00
Figure 3.3: Actual pressure, slurry rate, and proppant concentration curves
for Stage 9 using GOHFER™ software. These data were available for all eight
stages on Well S1 and for all seven stages of Well D1.
At each node, reservoir properties such as porosity, permeability, and pore
pressure are assigned by directly uploading the LAS log files into the system.
Other inputs such as formation zone thickness as well as rock-mechanical
properties derived from the logs are also able to be assigned.
Besides simulating fractures using data from the treatment that took place
on the treatment well, an accurate lithologic representation of the well is also
created based on the input from logs. By modeling the hydraulic fracture
treatment, the length and the height of the fracture can be estimated. It is common
41
in a fracture analysis to use microseismic to confirm fracture growth. For this
project, the downhole-monitored and surface-monitored microseismic data are
compared to the simulated fracture models. The software work flow and list of
important input data needed to complete the initial construction of the model is
further discussed in the next section.
After inserting input data and running the simulation using the software,
there are two important requirements that must be met in order to produce a
unique match for the model. By modeling, an attempt is made to match both the
pressure behavior as well as the fracture geometry. Pressure matching is done by
getting the simulated pressure curve to match with the actual pressure data from
the field. Only then, the simulated fracture geometry can be matched to the actual
fracture geometry.
3.2 Modeling Process
The general work flow for actual job data matching using post-treatment
models is shown in Figure 3.4. The seven steps shown cover the process of
building a GOHFER™ model using actual oil/gas field data after the hydraulic
fracture treatment took place. Step 5 is an iterative process that needs to be
repeated until the desired match is achieved between actual pressure and the
GOHFER™ simulated pressure. The results of the pressure matching process for
this project using Figure 3.4 are discussed in Chapter 5.

42
3.2.1 Input
The grid oriented feature of GOHFER™ is one of the important key
factors that contributes to this project. The rectangular grid structure is used to
describe the entire reservoir, serving the same function as a reservoir simulator,
and it allows the assignment of complex and detailed descriptions of the particular
fractured intervals. For each node, reservoir properties such as permeability,
porosity, and pore pressure, and mechanical properties such as Young’s Modulus,
Poisson’s Ratio, Biot’s constant, and tectonic stress are assigned (Figure 3.5).
The list of input required for a working model to be created using GOHFER™
are listed below, but are not limited to (Barree, 2009):
• Rock elastic properties:
o Poisson’s Ratio, Young’s Modulus, Biot’s constant
• Reservoir pressure variations:
o Depletion, overpressure, underpressure (lateral and vertical
variations)
• In-situ stresses:
Computed from derived rock properties
• Leakoff magnitude and distribution of natural fractures:
o Acquired from injection diagnostic tests
• Fluid and proppant schedule:
o Constructed to arrive at desired length and conductivity with
required height development

43
GOHFER
Actual Job Data
Match Work Flow
Create new
1 GohWin™ file
Create
WinGOHFER™
simulation Create Log File using
LOGCALC™ to
analyze LAS files
Complete Actual
Simulation Input
Process
Define grids, adjust &

a initialize node sizes &
nudge
Formation zone set up,
depth intervals, b
formation rock type Perforation diameter,
c number of holes
Define wellbore, enter perforated
actual pumping d
schedule
Save & Run

2 Simulation
Perform PDAT
3 (if data available)
View pressure plots

4 using GohGraph™, set
stage locations
Adjust reservoir & grid

5 properties to match
pressures
View pressure match

results using HTG™,
6 repeat step 5 if needed
View simulated fracture

7 geometry using
WinParse™
Figure 3.4: Work flow of GOHFER™ actual job data match for
post-treatment model.
44
Figure 3.5: An example of the grid structure feature of GOHFER™ where

reservoir and mechanical properties are individually assigned for each node. The
left track shows the perforated section of the well, colored in brown, and a log
track of gamma ray in yellow. The right track generally represents the scale for
reservoir or mechanical properties used where in this example it represents the
scale for Young’s Modulus values in MMpsi.
For the reservoir and mechanical properties, values are usually derived
from log files that are provided by the service company for the specified treatment
wells. The LOGCALC™ function in the software is used specifically to perform
these value derivations. Properties such as stress, strain, Young’s Modulus, and
Poisson’s Ratio and their impacts on the fracture model are further discussed as
follows:
45
Stress:
The total fracture closure stress equation as used in GOHFER™ is listed
as (Barree, 2009):
ν
Pc =
(1 −ν )
[D γ
tv ob ]
− α v (Dtvγ p + Poff ) + α h (Dtvγ p + Poff ) + ε x E + σ t (3.1)
where Pc = closure pressure, psi
ν = Poisson’s Ratio
Dtv = true vertical depth, feet
γob = overburden stress gradient, psi/ft
γp = pore fluid gradient, psi/ft
αv = vertical Biot’s poroelastic constant
αh = horizontal Biot’s poroelastic constant
Poff = pore pressure offset, psi
εx = horizontal tectonic strain, microstrains
E = Young’s Modulus, million psi
σt = regional horizontal tectonic stress, psi
For effective use of Equation 3.1, variable interaction with each source of data
should be studied in order to fully understand the complexity of estimating a
“physically consistent and reasonably accurate stress profile” (Barree, 2009).
A good comprehension of stress distribution is important to understand
fracture growth and geometry. The distribution of stress is responsible in
controlling some factors such as the orientation of a fracture, fracture height

46
containment, treating pressure magnitude, and changes in treating pressure during
the fracturing job. Hydraulically induced fracture orientation in the subsurface is
mainly dominated by the stress difference that exists between the three principal
stresses: overburden pressure (σ1), maximum horizontal pressure (σ2), and
minimum horizontal pressure (σ3) as shown in Figure 3.6. According to Barree
(2009), the magnitude of the stresses especially the minimum stress will
determine the fluid pressure needed to open a fracture. On the other hand, stress
variations between different rock layers control the magnitude of treating pressure
and pressure behavior with time by controlling the height of the fracture created
(Barree, 2009).
σ1
σ1 > σ 2 > σ 3
σ3
σ2
Figure 3.6: Fracture orientation is determined by the relationship of principal
stresses, σ1, σ2, and σ3, as shown in this figure. Fracture growth will occur
perpendicular to the minimum principal stress (σ3) since it is the path of least
resistance (modified from Barree, 2009).
47
Strain:
Strain is defined as the deformation of rock when a force or stress is
applied to rock, changing its shape, size and/or length. The strain equation is
simply the difference between the new and original length divided by the original
length of the measured section (Figure 3.7). Strain is a dimensionless parameter
that can represents a positive or negative value representing compression and
extension, respectively.
ε = (L2 – L1) / L1 (3.2)
∆L
Compressed
Original L1 Sample
Length L2
Length
Figure 3.7: An example of strain where rock sample is deformed and compressed
due to force or stress applied (modified from Barree, 2009).
Young’s Modulus:
Young’s Modulus indicates the “stiffness” of rock. This parameter
measures the amount of stress required to cause deformation of rock, hence,
indicating the hardness or “stiffness” of the rock. Young’s Modulus is simply
calculated by taking the ratio of stress over strain (Figure 3.8). A higher modulus
value represents stiffer material because it needs more stress applied to yield a
given strain. For example, a very hard and well cemented carbonate or sandstone
48
may have modulus value of 10 to 12 million psi (Barree, 2009). On the other
hand, a very soft, friable and unconsolidated sandstone may have modulus value
of one million psi or less (Barree, 2009).
The task of determining elastic properties for a rock sample is a difficult
process. It is usually done by physically taking static measurements on cores and
using dynamic log measurements to calibrate. According to Warpinski et al.
(1998), laboratory measurements have shown that static measurements of cores
yields lower values than dynamic moduli values. These discrepancies are
contributed by matrix weaknesses, in-situ reservoir confining stress, and the
poroelastic nature of rocks themselves (Economides and Nolte, 1981). In addition,
getting exact dynamic measurements is difficult because a small amount of error
in measurement could lead to large discrepancies in values.
For Young’s Modulus values that are measured in a laboratory on an
unconfined sample, the measured value can potentially be very different that
Young’s Modulus values that are measured on the same sample but at reservoir
stress conditions. Lama and Vutukuri (1978) stated that stress history and
saturation conditions do affect the measurement of Young’s Modulus. Hence, lab
core measurements are problematic to work with in achieving practical and useful
rock elastic properties.
However, decent measurements are achievable. According to Pantoja
(1998), careful and close observations are needed to develop suitable correlations
for in-situ moduli that are based on static measurements using appropriate scale
which affects the final values and final application of the data.
49
Stress
Strain
Figure 3.8: Young’s Modulus is defined as the ratio of stress over strain. The
higher the slope is, the “stiffer” the material is. In this figure, the red slope line
represents a higher modulus value for “stiffer” material. On the other hand, the
green slope line represents a lower modulus value for material that is less stiff
(modified from Barree, 2009).
Poisson’s Ratio:
Poisson’s Ratio is another common rock mechanical parameter that
contributes to fracture geometry. By definition, Poisson’s Ratio is the ratio of
lateral strain, εx to axial strain, εz under conditions of axial loading as shown in
Figure 3.9 (Barree, 2009). In other words, the ratio of Equation 3.3 over Equation
3.4 produces Poisson’s Ratio value as shown in Equation 3.5. The numerical
value of this parameter ranges from 0 to 0.5 where a 0 value indicates no lateral
strain occurring when a sample is loaded and a value of 0.5 indicates that the
particular sample expands laterally as much as it compacts axially. A typical
Poisson’s Ratio value for rocks ranges between 0.15 to 0.35.

50
εz = ∆L/L1 (3.3)
εx = ∆D/D1 (3.4)
ν = εx / εz (3.5)
Where εz = Strain in z-direction (axial strain)

εx = Strain in x-direction (lateral strain)
∆L = Change in length
L1 = Initial length
∆D = Change in diameter
D1 = Initial diameter
ν = Poisson’s Ratio
εz
L1
D1 εx
Figure 3.9: Poisson’s Ratio is simply the ratio of lateral to axial strain under
conditions of axial loading. Smaller ratio value indicates small lateral strain
occurring as sample is loaded (modified from Barree, 2009).
The LOGCALC™ feature in the software helps compute dynamic
parameters such as the previously mentioned Young’s Modulus and Poisson’s
Ratio sourced from well logs. This feature is most beneficial when core data are
not available. A set of well logs are used as source curves to be fed into
LOGCALC™ and the computed parameters are displayed as output curves

51
(Figure 3.10). The general structure of the input and output curves are listed as
(Barree, 2009):
• Input (source) curves:
o Any number of input LAS-type source files;
o CSV-type files with core data, or externally generated values;
o All source curves are used as inputs to various calculations;
o Output curves can also be used as sources; or,
o Multiple sources can be input to each calculation.
• Output curves:
o Can be generated by copying input, averaging multiple inputs,
summing inputs;
o List of outputs required for GOHFER™ model generation are pre-
defined with specific calculations;
o Order of computations is set to minimize work load;
o Each generated curve may be used as input to later calculations; or,
o New output curves can be generated by the user.
3.2.2 Output
The main simulated outputs sourced from GOHFER™ are viewed through
another feature called WinParse™. There are several options on parameters that
can be displayed using WinParse™. The most common and useful parameter is
the proppant concentration (Figure 3.11). From this figure, proppant concentration
within the fracture can be studied. The asymmetric, bi-wing features of the
52
fracture are also displayed to assist approximating fracture length on both sides of
the wellbore since not all fractures will grow symmetrically.
Generated Input
Output Curve (Source) Curves
Figure 3.10: The main screen layout of the LOGCALC™ displaying the input
curves and the generated output curves. The input log curves are displayed on the
right side of the panel and the generated output curve is displayed on the left side
of the source curves.
53
Perforated Holes
Figure 3.11: An example of WinParse™ view showing the distribution of

proppant concentration in this asymmetric fracture. In asymmetric growth,
fracture wings extended out from both sides of the well are modeled so that the
effects of lateral pressure gradiaents, dip of beds, and/or changes in rock
properties can be honored to represent the hydraulic fracture growth more
accurately. As mentioned in Chapter 2, one of the appealing features of
GOHFER™ is its ability to model asymmetric bi-wing fracture on a wellbore.
The lateral-scaled axis represents the length of the fracture and the vertical-scaled
axis represents the depth section of the wellbore where the fracture is being
initiated.
Some other parameters that can be displayed using WinParse™ are (but
not limited to):
• Net pressure;
• Viscosity;
54
• Fracture pressure;
• Fracture width;
• Injection rate; and,
• Leak-off rate.
Another feature of GOHFER™ that serves the purpose of viewing output
results from the fracture simulation is the HTGraph™. Figure 3.12 shows an
output example from the simulation of Stage 2 for Well D1. From this graph, six
major curves are displayed for comparison and matching purposes. The six major
curves are the actual data from the actual hydraulic fracturing job performed in
the field as shown in Figure 3.3: tubing pressure, slurry rate, and proppant
concentration; and the simulated GOHFER™ output curves: surface pressure,
slurry rate, and proppant concentration.
3.3 Pressure Matching Process
The pressure matching process refers to the task of matching the simulated
GOHFER™ pressure curve with the actual pressure curve. The key to a unique
simulation output is when both pressures and geometries are matched. Hence, a
unique fracture model result is successfully obtained when the two pressure
curves are matched to overlap each other and also when the fracture geometry
obtained from the GOHFER™ simulated result is matched to the actual geometry.
The ability to obtain as much input data as possible in building a hydraulic
fracture model will reduce the needs for making assumptions. A typical problem
in a hydraulic fracture simulation is attempting to match the simulator result to the

55
actual field data. It is common to modify the input data in order to manipulate the
model. This is when advanced parameters are used which are later discussed in
the next chapter.
T ubing Pressure (psi) A Slurry Rat e (bpm) B

Proppant Concentrat ion (lb/gal) C GOHFER Surface Pressure (psi) A
A GOHFER Slurry Rat e (bpm) B GOHFER Surface Prop Conc (lb/gal) C B C
7000 100 6.0
6500 90 5.5
6000 5.0
80
5500 4.5
70
5000 4.0
4500 60 3.5
4000 50 3.0
3500 40 2.5
3000 2.0
30
2500 1.5
20
2000 1.0
1500 10 0.5
1000 0 0.0
11:40 11:45 11:50 11:55 12:00 12:05 12:10 12:15
1/30/2009 1/30/2009
Time
Customer: Job Date: Ticket #:
GohWin v1.6.5
Well Desc: UWI: 09-Jul-09 12:43
Figure 3.12: The HTGraph™ is used to display the actual pressure, slurry rate,
and proppant concentration curves acquired during a hydraulic fracturing job.
Another set of data that is also displayed on the same screen window is the
simulated data for the pressure, slurry rate, and proppant concentration. These
simulated data are noted as GOHFER™ Surface Pressure, GOHFER™ Slurry
Rate, and GOHFER™ Surface Proppant Concentration.
56
CHAPTER 4
MICROSEISMIC EVENTS AND HYDRAULIC FRACTURE
MODEL RESULTS
This chapter describes the process of inputting the actual hydraulic
fracturing data for Wells D1 and S1 into the GOHFER™ software, the final
matches determined between the actual and simulated data, and the comparison of
these models with the microseismic events as displayed in Transform™. First, the
pumping treatment schedules from the actual hydraulic fracturing treatments were
uploaded into the GOHFER™ software. Five models were then built for each
subject well, Well D1 and Well S1, using the actual pumping data and LAS log
files. A match of the simulated GOHFER™ pressure data as well as the simulated
flow rate and proppant concentration to the actual treatment pressure was then
obtained to the actual data. Once the pressures were matched for each stage from
both Wells D1 and S1, the fracture geometry profiles were overlaid with the
microseismic events captured during the actual hydraulic fracturing treatments.
Performing this integration step allows analyses be made pertaining to the use of
microseismic data in confirming fracture growth to replicate the actual hydraulic
fracture treatment. Also, comparisons are made between the downhole-monitored
treated well and the surface-monitored treated well. The integration process of
these multiple inputs is described by the flow chart in Figure 4.1.

57
Actual hydraulic Recorded microseismic

fracturing job data events
(pressure, slurry rate, (X, Y, and Z
proppant concentration) locations)
GOHFER™ Transform™
Actual and simulated Microseismic event

pressure matched, locations displayed
simulated fracture around treatment
geometry well
Microseismic events and

modeled fracture geometry
are overlapped and
compared
Determine
Match
Characteristics
Figure 4.1: Flow chart of the integration process where fracture geometry output
from GOHFER™ models are overlapped with microseismic events.
58
4.1 Input Data For Hydraulic Fracture Model
Building a working hydraulic fracture model requires a large number of
input data. In this section, a list of input data ranging from log-derived
parameters to actual pumping schedule data are described as well as their effects
on the model.
4.1.1 Logs and Log-derived Properties
As described in Chapter 3, the GOHFER™ model is heavily dependent on
log file input. For this project, a number of Log ASCII Standard (LAS) files were
used to construct the hydraulic fracture models for both Wells D1 and S1. The
LAS files include some important logs such as the resistivity, gamma ray, caliper,
bulk density, neutron porosity, neutron density, and most importantly the dipole
sonic log which provides information on transit compressional travel time
(DTCO) and transit shear travel time (DTSM).
Uploading these log files into LOGCALC™, the log-processing
component of GOHFER™, resulted in crucial log-derived properties that were
later fed into the formation zone setup in WinGOHFER™. Some of the derived
properties are:
• Poisson’s Ratio
o With the availability of full wave sonic or dipole sonic logs,
this property is calculated with a simple equation (Equation
4.1) that is based on the ratio of the shear-to-compressional
wave arrival times:

59
ν = (R – 2) / (2R – 2) (4.1)
where R = (DTSM2/DTCO2) and is unitless.
In this project, Poisson’s Ratio values are calculated from the
available sonic log.
• Young’s Modulus
o If Poisson’s Ratio is calculated using Equation 4.1, Young’s
Modulus can also be calculated directly provided that a bulk
density log and sonic logs are available. The equation used to
calculate this property is shown in Equation 4.2:
E = 13447 ρb [(3R – 4) / (DTCO2*R*(R – 1))] (4.2)
where ρb = bulk density in g/cm3 and R = (DTSM2/DTCO2) is
unitless.
For this project, Young’s Modulus values are derived from the
provided sonic and density logs.
• Biot’s Constant
o Horizontal Biot’s constant is set at 1 in all cases in
GOHFER™. This constant represents the interaction of pore
fluid pressure on horizontal total stress. Due to the pore fluid
being in direct communication with the frac fluid, the pressure
response should be at a ratio of 1:1 and is not translated
through any rock deformation.
o Vertical Biot’s constant represents the poroelastic constant that
describes the stress that counteracts the reservoir gradient in

60
the lateral direction. The estimation of this property is
configured by Equation 4.3:
αv = 0.6 + (M*PHIE) (4.3)
where M is defaulted at 1 and PHIE is the shale-corrected
effective porosity from a neutron-density crossplot.
• Process Zone Stress (PZS)
o This parameter is a directly measured pressure taken from the
extension pressure and the closure pressure in a fracture
injection test. PZS is not related to only one property since it
includes effects of fluid lag, intact rock strength (tensile
strength), and other non-linear stress dissipations around the tip
of the fracture. Due to the unavailability of the mini-frac
analysis (fracture injection test) in this project, the PZS is
estimated from log data using Equation 4.4:
PZS_VSHALE = a1* VSHALE + a2 (4.4)
where a1 = 500 and a2 = 500.
This is an acceptable alternative way to produce PZS value
when the mini-frac data is not available (Barree, 2008). Due to
the lack of such data, this value might not accurately represent
the actual value as would be predicted by the mini-frac test.
• Total Stress
o This value is calculated from other parameters such as
Poisson’s Ratio, Young’s Modulus, Pore Pressure Horizontal

61
Biot’s constant, Vertical Biot’s constant, tectonic strain, and
tectonic stress. The equation used is the same as shown in
Equation 3.1 in Chapter 3.
Tables 4.1 and 4.2 represent the log-derived data for both Wells D1 and
S1. For each well, an average value of Young’s Modulus, Poisson’s Ratio, and
porosity is calculated for each hydraulic fracture stage. The values are averaged at
every 10 feet of the gross height interval for presentation purposes. Also, the net
height shown in these two tables refers to the perforated interval for each stage.
For Well D1, the hydraulic fracturing job treatment was performed with
seven separate stages. However, due to the poor quality of some treatment data,
Stages 3 and 6 are eliminated from this project. On the other hand, Well S1 is
hydraulically fractured at nine separate stages. For the same reason as Well D1,
Stages 6, 7, and 8 of Well S1 are not included in this project.
Table 4.1: Well D1 Input Data of Hydraulic Fracture Model
Hydraulic Perforated Gross Net Poro Young’s Poisson’s

Fracture Depth Height Height* -sity Modulus Ratio
Stage (feet) (feet) (feet) (%) (MMpsi) (unitless)
Stage 1 8273 - 8548 275 14 9.6 4.907 0.265
Stage 2 7975 - 8188 213 12 6.9 6.480 0.221
Stage 3** 7703 - 7890 187 12 11.6 5.301 0.274
Stage 4 7434 - 7653 219 12 8.6 6.649 0.231
Stage 5 7146 - 7345 199 12 7.7 5.812 0.263
Stage 6** 6856 - 7071 215 12 8.9 5.755 0.272
Stage 7 6528 - 6750 222 14 7.7 6.153 0.248
*Net height refers to perforated height interval for each stage
**Eliminated stage from this project
62
Table 4.2: Well S1 Input Data of Hydraulic Fracture Model
Hydraulic Perforated Gross Net Poro Young’s Poisson’s

Fracture Depth Height Height* -sity Modulus Ratio
Stage (feet) (feet) (feet) (%) (MMpsi) (unitless)
Stage 2 7950 - 8071 121 12 6.5 6.631 0.200
Stage 3 7716 - 7913 197 12 7.1 5.701 0.241
Stage 4 7460 - 7663 203 14 7.0 6.932 0.203
Stage 5 7213 - 7417 204 14 6.3 6.525 0.233
Stage 6** 6947 - 7142 195 14 6.9 6.193 0.240
Stage 7** 6721 - 6878 157 12 7.0 6.248 0.238
Stage 8** 6416 - 6647 231 12 6.7 6.340 0.243
Stage 9 6223 - 6371 148 12 7.5 5.784 0.251
*Net height refers to perforated height interval for each stage
**Eliminated stage from this project
4.1.2 Formation Zone Setup
One of the important features in WinGOHFER™ is the Formation Zone
Setup and Grid section as previously shown in Figure 3.5. In this section, user
input data and log-derived data are the major components to the construction of a
GOHFER™ model. The input data are listed as:
• User input
o Zone Setup – Noting the producing and non-producing
formation intervals.
o Perforation Diameter – A perforation diameter of 0.34 inches is
used for both wells.
o Holes Perforated – Number of total perforations shot at each
perforated grid. Each grid represents a true vertical depth of 10
feet. The number of perforations ranges between two to three
shots per foot.

63
• Log input and log-derived input
o Effective Porosity – This value is read from the utilized LAS
logs. Effective porosity affects fluid leakoff calculations within
the fracturing simulator.
o Permeability – This value is read from the utilized LAS logs
using correlations embedded in LogCALC™. This is done due
to the lack of permeability value from logs or mini-frac
analysis. This property represents reservoir permeability with
respect to fluid flow.
o Poisson’s Ratio – This property describes the deformation
characteristics of the formation material. The value is derived
from the sonic log.
o Young’s Modulus – This property describes the stiffness of the
formation. The value can also be calculated from the sonic log
and bulk density log.
o Pore Pressure Gradient – This is where a value for fluid
pressure gradients is entered for the pore pressure calculation.
o Pore Pressure Offset – This property describes the amount of
overpressure or depletion that is present within the reservoir
layers.
o Pore Pressure – this property describes the present status of the
reservoir pressure. This value is a combination of the data from
the Pore Pressure Gradient and Pore Pressure Offset tabs.

64
o Horizontal Biot’s Constant – This property describes the
poroelastic constant that used to predict the counteracting stress
of the pore pressure against the overburden gradient.
o Vertical Biot’s Constant – This property describes the
poroelastic constant that describes the stress that counteracts
the reservoir gradient in the lateral direction.
o Process Zone Stress – This value includes the effects of fluid
lag, tensile strength of rocks, and other non-linear stress
dissipations around the tip of the fracture.
o Total Stress – This value is calculated using values from
Poisson’s Ratio, Young’s Modulus, Pore Pressure, Horizontal
Biot’s Constant, Vertical Biot’s Constant, Tectonic Strain, and
Tectonic Stress tabs that are previously derived from the logs.
o Fissure Opening Pressure – This property describes the amount
of pressure offset above the total stress that is required to open
natural fractures (can be measure from pump in tests).
o Tectonic Strain – This property describes the amount of lateral
tectonic movement that the reservoir structure is under. A
range of tectonic strain values from 100 to 300 microstrains are
used for each stage of each well (Barree, 2008). Such variation
could vary potentially be due to depth, geology structure such
as fault, as well as asymmetric sediments accretion that were

65
deposited gradually in the fluvial-deltaic type of environment
(Escartin et. al., 1949).
o Tectonic Stress – This property describes a constant shift in
tectonic stress.
o Transmissibility Multiplier – This property describes the ability
of the fluid to transmit pressure along a fracture length. The
multiplier is applied to the parallel plate solution for fluid flow
down the fracture to determine the amount of fracture frictional
pressure. Thin fluids like water will results in very narrow
fractures that show low values of the transmissibility
multiplier. Thicker fluids, on the other hand, will show higher
value closer to 1.
o Percent Dolomite – This value represents the percentage of
reservoir rock that is dolomite.
o Percent Limestone – This value represents the percentage of
the reservoir rock that is limestone.
o Proppant Holdup Factor – This value accounts for the relative
velocity of the proppant with respect to the fluid velocity. This
relatively velocity ratio is calculated internally in GOHFER™
based on the fracture width, local shear rate, and local leakoff
values.
Figures 4.2 to 4.11 represent input data that were imported into
GOHFER™ for each model. These figures are generated using WinParse™ from
66
GOHFER™. The data included are Poisson’s Ratio, Young’s Modulus,
Permeability, Horizontal Biot’s Constant, Vertical Biot’s Constant, Porosity,
Percent Dolomite, Percent Limestone, Process Zone Stress, Pore Pressure, Total
Stress, and Critical Fissure Opening Pressure.
4.1.3 Pumping Schedule
Another input section in WinGOHFER™ is the Pumping Schedule, where
additional information on the wellbore as well as pumping information from the
actual hydraulic fracturing job, is inserted to construct the model. The four tabs in
this section are labeled “Wellbore”, “Design”, “Injection Test”, and “Actual” as
shown in Figure 4.12.
In the “Wellbore” tab, wellbore information such as the Perforation
Coefficient of Discharge, treatment tubing diameter, and wellbore fluid are
inserted. The “Perforation Coefficient of Discharge” is the value that is used in
the calculation of perforation friction. This value is assumed to be the starting
value at the onset of the fracture stimulation job, and a value varying from 0.5 to
0.7 was used for the models in this thesis. As proppant is injected into the
wellbore, the value of this coefficient will increase, as well as the perforation
diameter due to erosion.
Under the “Wellbore String” section in this tab, a value of four inches was
used as the treatment tubing diameter. Another user input required under this tab
is the “Wellbore Fluid” condition that gives the status of the wellbore contents
prior to injection.
Figure 4.2: Input data imported into GOHFER™ for Stage 1 of Well D1.
67
68
69
70
71
Figure 4.7: Input data imported into GOHFER™ for Stage 2 of Well S1.
72
73
74
75
76
77
Figure 4.12: The Pumping Schedule tab (upper right) in WinGOHFER™

displaying wellbore information along with designed and actual pumping
schedules (tabs on the lefthand side).
The second tab in the Pumping Schedule section is the “Design” tab which
allows the fluid and proppant types to be chosen from a drop-down menu.
Multiple stages are inserted while designing a fracturing treatment to represent the
different stages that take place during the fracturing job with respect to proppant
concentration and time.
Another important tab to ensure a working model is the “Actual” tab,
where the actual pumping schedule data are imported into WinGOHFER™. In
78
this section, the fluid and proppant type along with proppant concentration and
slurry rate information from the actual job are uploaded into the system. The
proppant type used in all the models is the Jordan Unimin 20/40 whereas the
fracturing fluid used in is “YF 125.1 HTD 200 J-475 5” fluid for both wells.
Two of the extended parameters included in this tab are “Friction
Coefficient Factor” and “Perforation Factor”. The Friction Coefficient Factor
allows the overall pipe friction to be adjusted to match the observed treating
pressured. This value is important in the process of pressure matching the actual
and simulated pressures. A value greater than 1.0 increases total pipe friction
whereas a value less than 1.0 reduces pipe friction. Value ranges from 0.3 to 1.5 is
used in these models. The Perforation Factor, on the other hand, determines the
percentage of perforation holes that are open. A value of 1 represents that all
holes are open and a value of 0.5 represents that only half of the perforations are
open. For these models, a Perforation Factor of 1 is used for all stages. Both
factors are entered per stage.
4.2 Input Data For Microseismic Model
For the microseismic model, a number of Comma Separated Values
(CSV) files were provided for each stage for both Wells D1 and S1. The X, Y,
and Z coordinates for each microseismic event are included in the files. These
files are imported into the Transform™ software along with the actual pumping
schedule data.
79
4.2.1 Microseismic Event Locations for Well D1
As previously stated in Chapter 1, Well D1 was monitored by downhole
microseismic technology set up in an adjacent monitoring well, Well D2. Figures
4.13 through 4.19 display the microseismic events that were detected during each
treatment stage in Well D1. The different colors of the microseismic events
represent different stages of hydraulic fracturing treatment done on each well
whereas the varying sizes of the event spheres are based on the magnitude of the
microseismic events (ie. the larger the magnitude is, the bigger the microseismic
event sphere is). These figures are the output of the CSV files that were
previously fed into the Transform™ software.
Figures 4.20 through 4.26 display the actual data from the hydraulic
fracturing job including the tubing pressure measured at surface, slurry flow rate,
and proppant concentration. These three sets of datastreams are used to build the
hydraulic fracture models as well as imported into Transform™ to be displayed
along with the microseismic event histogram and microseismic event count. A
general observation shows that the microseismic events at each stage lasted
throughout the fracturing job with most events distributed during early- and
middle-time during the job.
Table 4.3 represents the estimations of the length, height, and azimuth of
the microseismic events that were detected in each fractured stage. The fracture
lengths are measured in terms of total length instead of half-length so that fair
numerical comparisons can be made between fracture geometries from the
downhole-monitored well to geometries from the surface-monitored well.

Well
Well D2
D1
8029 ft
Stage 1 Microseismic
Events
8426 ft
Figure 4.13: Microseismic events (blue dots) that occurred in Stage 1 of Well D1. The microseismic events were
between depths of 8029 ft to 8426 ft.
80
Well
Well D2
D1
7920 ft
8195 ft
Events
Figure 4.14: Microseismic events (red dots) that occurred in Stage 2 of Well D1. The microseismic events were
81
Well
Well D2
D1
7372 ft
7813 ft Events
Figure 4.15: Microseismic events (aqua dots) that occurred in Stage 3 of Well D1. The microseismic events were
82
Well
D1
7284 ft
Events
7909 ft Well
D2
Figure 4.16: Microseismic events (orange dots) that occurred in Stage 4 of Well D1. The microseismic events were
83
Well
Well D2
D1
7069 ft
7383 ft
Events
Figure 4.17: Microseismic events (green dots) that occurred in Stage 5 of Well D1. The microseismic events were
84
Well
Well D2
D1
6685 ft
Events
7163 ft
Figure 4.18: Microseismic events (grey dots) that occurred in Stage 6 of Well D1. The microseismic events were
85
6483 ft
Events
Well
7058 ft Well D2
D1
Figure 4.19: Microseismic events (pink dots) that occurred in Stage 9 of Well D1. The microseismic events were
86
Figure 4.20: Actual hydraulic fracturing data (slurry rate, surface pressure, and proppant concentration) for Stage 1 of
Well D1 including the microseismic event count and histogram.
87
88
89
90
91
92
93
94
Table 4.3: Microseismic Event Locations for Well D1
Microseismic
Perforated Perf. Microseismic Microseismic Activity and
Stage Height Height Activity Activity Fracture
Interval (feet) Total Length Height Azimuth
(feet) (feet) (feet) (degree angle)
1 8273 - 8548 275 700 397 113
2 7975 - 8188 213 900 275 107
3 7703 - 7890 187 600 441 95
4 7434 - 7653 219 1250 500 100
5 7146 - 7345 199 850 314 99
6 6856 - 7071 215 1250 478 100
7 6528 - 6750 222 1150 278 102
4.2.2 Microseismic Event Locations for Well S1
In contrast to Well D1, Well S1 was monitored by surface microseismic
technology without any additional monitoring wells. Figures 4.27 through 4.34
display the microseismic events that were detected during each treatment stage in
Well S1. These figures are also the output of CSV files that were previously
entered into the Transform™ software. Figures 4.35 through 4.42 display the
actual data from the hydraulic fracturing job including the tubing pressure
measured at surface, slurry flow rate, and proppant concentration along with the
microseismic event histogram and microseismic event count. A general
observation shows that the microseismic events at each stage are mostly
concentrated during middle- and late-time of job. Table 4.4 represents the
estimations of the length, height, and azimuth of the microseismic events that
were detected in each fractured stage. The total length measured for the fractures
based on microseismic activities are estimated from the tip of the fracture wing on
the West side to the tip of the fracture wing on the East side.
7528 ft
Events
8071 ft
Figure 4.27: Microseismic events (blue dots) that occurred in Stage 2 of Well S1. The microseismic events were
between depths of 7528 ft to 8071 ft
95
7329 ft
Events
8123 ft
Figure 4.28: Microseismic events (red dots) that occurred in Stage 3 of Well S1. The microseismic events were
96
7171 ft
Events
7803 ft
Figure 4.29: Microseismic events (aqua dots) that occurred in Stage 4 of Well S1. The microseismic events were
97
7025 ft
7472 ft
Events
Figure 4.30: Microseismic events (orange dots) that occurred in Stage 5 of Well S1. The microseismic events were
98
6495 ft
Events
7083 ft
Figure 4.31: Microseismic events (green dots) that occurred in Stage 6 of Well S1. The microseismic events were
99
6501 ft
Events
7168 ft
Figure 4.32: Microseismic events (grey dots) that occurred in Stage 7 of Well S1. The microseismic events were
100
6262 ft
Events
7100 ft
Figure 4.33: Microseismic events (pink dots) that occurred in Stage 8 of Well S1. The microseismic events were
101
5471 ft
Events
6836 ft
Figure 4.34: Microseismic events (light green dots) that occurred in Stage 9 of Well S1. The microseismic events
were between depths of 5471 ft to 6836 ft.
102
Well S1 including the microseismic event count and histogram.
103
104
105
106
107
108
109
110
111
Table 4.4: Microseismic Event Locations for Well S1
Microseismic
Perforated Perf. Microseismic Microseismic Activity and
Stage Height Height Activity Activity Fracture
Interval (feet) Total Length Height Azimuth
(feet) (feet) (feet) (degree angle)
2 7950 - 8071 121 750 543 120
3 7716 - 7913 197 800 794 120
4 7460 - 7663 203 1100 632 110
5 7213 - 7417 204 1100 447 110
6 6947 - 7142 195 800 588 120
7 6721 - 6878 157 1100 667 110
8 6416 - 6647 231 650 838 125
9 6223 - 6371 148 700 872 110
4.3 GOHFER™ Pressure Matches
As mentioned in Section 3.3, the pressure matching process refers to the
task of matching the simulated GOHFER™ pressure curve with the actual
pressure curve. Typically, it is problematic to attempt matching the simulator
result to the actual field data in a hydraulic fracture simulation. Obtaining as much
input data as possible to build an accurate hydraulic fracture model reduces the
needs for making assumptions. However, in order to get the final pressure match,
some modifications are made to the user-input and log-derived parameters, as
well as advanced parameters to obtain unique results as discussed in the next
section. These modifications are made with the understanding that all data have
some inherent error. Modifications are made in a systematic way to honor the
most accurate data first.

112
4.3.1 Matched Parameters
Throughout the pressure matching process, few parameters are modified in
the Formation Zone Setup and Grid section in GOHFER™. The most common
parameter that is modified in this process is the tectonic strain. Whenever the
computed stress is above or below the measured closure pressure, the differences
in the values are attributed to tectonic strain. Thus, changes are made by
increasing the tectonic strain to lift the simulated pressure curve and vice versa.
Another modification is also made on the Process Zone Stress (PZS) parameter.
Since the PZS accounts for the effect of fluid lag, rock tensile strength, and other
non-linear stress dissipations around the tip of the fracture, is it not related to only
one single property. Also, since the value for this parameter in this work is
lacking in detailed information due to the unavailability of a mini-frac analysis,
the PZS is derived from logs by using the LogCALC™ function in GOHFER™.
In the Pumping Schedule tab, there are two parameters that can be
changed in this pressure matching process. They are the Perforation Coefficient of
Discharge and Friction Coefficient Factor. As previously mentioned in Section
4.1.3, the Perforation Coefficient of Discharge is the value that is used to calculate
perforation friction. This value is assumed to be the starting value at the onset of
the fracture stimulation job. As proppant is injected into the wellbore, the value of
this coefficient will increase, as well as the perforation diameter due to erosion
effects. Equation 4.5 represents one of the methods to calculate total friction
pressure in terms of slurry flow rate:
Total ∆P (qi) = Kperf*q2 + Knear WB*q1/2 (4.5)

113
Where Kperf = perforation friction coefficient (psi/BPM2)
Knear WB = near-wellbore friction coefficient (psi/BPM1/2)
P = friction pressure (psi)
q = slurry flow rate (barrel per minute)
The first term on the right-hand side of Equation 4.5 represents the delta pressure
contributed by perforation friction, whereas the second term on the same equation
represents the friction pressure due to tortuosity and misalignment of the fracture
and the wellbore. For the Kperf term in particular, it is calculated using Equation
4.6:
Kperf = (0.2369*ρ) / (N2*D4*C2) (4.6)
Where ρ = density (pound per gallon)
N = number of perforation (unitless)
D = diameter of perforations (inch)
C = perforation coefficient of discharge (unitless)
From Equation 4.6, it is known that a higher value of perforation
coefficient of discharge (closer to 1.0) will result in a lower total friction pressure
value and vice versa. A value of 0.55 is set as the default in the software. This
coefficient is crucial in determining the match to actual pressure during the early
times of the treatment. The Friction Coefficient Factor, on the other hand, allows
the overall pipe friction to be adjusted to match the observed treating pressure. A
value greater than 1.0 increases total pipe friction whereas a value less than 1.0
reduces pipe friction. This parameter is adjusted accordingly to match the general
114
shape of the actual pressure, especially during late times due to high shut-in
pressures simulated in most of the WinGOHFER™ models for the stages.
4.3.2 Advanced Parameters
Under the Customer and Reservoir Information tab in WinGOHFER™,
there is a section called Advanced Reservoir Parameters. These parameters assist
in matching the pressure responses to phenomena such as natural fracture clusters
or the ongoing changes in lateral reservoir parameters. Manipulating the models
by entering input values for the advanced reservoir parameters is an important
factor in the pressure matching process due to the additional leak-off issues
caused by such reservoir changes. The advanced parameters are:
• Pressure Dependent Modulus Stiffness Factor (MSF);
o Used to determine the magnitude of modulus change created by the
opening of natural fractures.
o Used in an exponential equation such that as the pressure exceeds the
fissure opening pressure, the modulus will change exponentially.
o Normal range for MSF value is 0 to 0.01 with positive values causing
the apparent Young’s modulus to increase, whereas entering negative
values will cause Young’s modulus to decrease with fissure opening.
A value of zero indicates no modulus stiffening.
• Pressure Dependent Leakoff (PDL);
o Used to determine the magnitude of leakoff change created by the
opening of natural fractures.

115
o Used in an exponential equation such that as the pressure exceeds the
fissure opening pressure, the leakoff will change exponentially.
o Normal value range for PDL is from 0 to 0.01 where only positive
values are allowed. A value of zero indicates no PDL behavior and
only matrix leakoff.
• Relative Permeability Ratio;
o Used to account for the change in permeability from the reservoir fluid
to the invading fracturing fluid.
• Transverse Storage Coefficient (TSC);
o Used to define how much fluid is moved from the main planar
hydraulic fracture to a presumed system of fractures transverse to the
main fracture. Both fluid and proppant are moved into “storage” by
this parameter.
o Values ranging from 0 to 0.003 are recommended. This parameter is
generally used in conjunction with PDL.
• Tortuosity Pre-Factor;
o This value can be entered if a step down analysis is performed to
determine tortuosity and number of open holes.
4.4 Hydraulic Fracture Model Results - Pressure Matches and Fracture

Geometries
In this section, simulated results from the GOHFER™ hydraulic fracture
models are presented for both Well D1 and Well S1. Also included in this section
are the discussions on the pressure-matching processes performed to obtain the

116
results for each stage. The final fracture geometries from both wells are also
presented.
As presented in the next subchapters, the fracture geometries simulated
from the hydraulic fracture models do varies in terms of height containment and
half-length. Such variations are contributed by several factors such as the
difference in rock mechanical properties for different rock type (ie. Young’s
Modulus and Poisson’s Ratio values of sandstone vs. shale) as well as reservoir
heterogeneity in both lateral and vertical directions. For example, hydraulic
fracture is well-contained in areas where a thick block of shale formation is
present above and/or below the perforated interval. Since there are tendencies for
the fracture to decrease in width and/or stop to propagate once it approaches shale
bodies, shale tends to be a good fracture barrier in between the treated stages in a
well. However, when the shale is consist of thin and laminate layer, fracture
height containment could potentially be poor as the fracture will not completely
stop propagating as it goes through the thin shale layer. In fact, the width of the
fracture will decrease as it propagates through the shale layer and can increase
again when the fracture makes contact on the sand body overlying and/or
underlying the shale layer.
4.4.1 Well D1 - Stage 1
Actual treatment data including the pressure, slurry rate, and proppant
concentration, as well as, the simulated data for Stage 1 of Well D1 are shown in
Figures 4.43 and 4.44 (Figure 4.43 also includes the simulated bottomhole
117
pressure curve). Match adjustments were made in this stage by increasing the
tectonic strain to 105 microstrains and decreasing the friction coefficient factor to
be 0.3 during late-time stage to match the pressure data. At early-time, the
simulated pressure data could only be matched to the actual pressure treatment
data by adjusting the perforation coefficient of discharge to 0.8. This lowers the
friction caused by perforations. Advanced parameters values were adjusted to
0.003/psi for the MSF, PDL, and TSC in this stage.
After close observation, it was identified that there was an issue regarding
the injected slurry rate during late-time. This problem occurs at every stage for
this downhole-monitored well. It is typical for pressures to drop as they approach
late-time stages in the simulation following the supposedly decreasing pumping
rate and proppant concentration injection trends. Nonetheless, the actual treatment
pressure data seemed to “peak” with sudden and rapid increases which remain
high.
The slurry rate increment as proppant injection is stopped is not a normal
scenario during a hydraulic fracturing job because the slurry volume is usually
decreased and later stopped with the termination of proppant injection as an
indication of the completion of the job for that particular stage. These odd flow
rate and pressure behaviors affect the results of the simulated data when the late-
time pressure for all Well D1 stages display a “step” shape as the simulated
pressure curve approaches the shut-in time. When the simulated surface pressure
is compared to the simulated bottomhole pressure (Figure 4.43), the “step” shape
in the late-time pressure for the bottomhole pressure is not as visible.

118
One possible explanation to the “step” shape behavior is that within the
wellbore, the bottomhole pressure has overcome the overburden pressure value in
the reservoir. Therefore, the rock layer gets shifted upwards since the overburden
pressure is not able to apply the same or higher pressure to prevent horizontal
fracture component growth. Such deformation can possibly be captured by using
tiltmeters or radioactive tracers.
In Figure 4.45, the final matched fracture geometry for Stage 2 of Well S1
is shown. The fracture height is very well contained within the perforated interval.
The proppant is highly concentrated within the shaly sandstone rock layers in the
perforated area especially at the region near wellbore. The fracture total length is
610 ft and the simulated fracture stayed within the 1500 ft grid size built for this
model.
4.4.2 Well D1 – Stage 2
Figure 4.46 shows the pressure, flow rate, and proppant concentration data
of the actual treatment for Stage 2 along with the simulated matched data. Figure
4.47 shows the simulated surface pressure curve without the bottomhole pressure
information for more detailed comparisons. To match the simulated pressure to
actual pressure, the tectonic strain was increased to 300 microstrains to shift the
simulated pressure upwards. Also, the perforation coefficient of discharge value
was raised to 0.7 to lower the total pressure. The friction coefficient factor was
increased to as high as 1.5 at early-time and is kept at this higher value throughout
the fracturing job. The MSF and PDL parameters were changed to 0.002/psi
whereas the TSC was changed to 0.003/psi.

T ubing P ressure (psi) A Slurry Rat e (bpm) B
P roppant Concent rat ion (lb/gal) C GOHFER Surface P ressure (psi) A
GOHFER Slurry Rat e (bpm) B GOHFER Surface P rop Conc (lb/gal) C
A GOHFER Bot t om Hole P ressure (psi) A B C
9000 120 20
8500
110 18
8000
100
7500 16
7000 90
6500 14
80
6000
70 12
5500
5000 60 10
4500
50 8
4000
40
3500 6
3000 30
2500 4
20
2000
10 2
1500
1000 0 0
09:18 09:22 09:26 09:30 09:34 09:38 09:42 09:46
1/30/2009 1/30/2009
Time
Customer: Job Date: T icket #:
GohWin v1.6.5
Well Desc: UWI: 22-Aug-09 16:45
Figure 4.43: Graph showing the matched GOHFER™ pressure (aqua) in comparison to bottomhole pressure (black)
and the actual pressure (red) measured during the fracturing job of Stage 1 for Well D1. The “step” shape mentioned
is seen at time 09:40 in this graph in the observed tubing pressure does not significantly appear in the bottomhole
pressure.
119
A GOHFER Slurry Rat e (bpm) B GOHFER Surface P rop Conc (lb/gal) C B C
6000 120 20
5500 110 18
100
5000 16
90
4500 14
80
4000 70 12
3500 60 10
3000 50 8
40
2500 6
30
2000 4
20
1500 10 2
1000 0 0
09:18 09:22 09:26 09:30 09:34 09:38 09:42 09:46
1/30/2009 1/30/2009
Time
GohWin v1.6.5
Figure 4.44: Graph showing the matched GOHFER™ pressure (aqua) in comparison to
the actual pressure (red) measured during the fracturing job of Stage 1 for Well D1.
Fairly good matches are observed.
120
Perforations
Figure 4.45: GOHFER™ output showing proppant concentration (lb/ft2) inside the fracture of Stage 1 for Well D1. The
vertical scale on the right-hand side represents depth and the Gamma Ray log on the left-hand side. The horizontal scale
at the bottom is for fracture length. The blue dashed line represents fracture half-length. The proppant concentration scale
is shown as the color bar at the bottom.
121
122
The final matched fracture geometry for this stage is shown in Figure 4.48.
Again, the fracture appears to be very well contained in terms of its height. The
proppant concentration is evenly distributed within the fracture, except for some
higher distributions surrounding the perforation holes. The fracture total length is
805 ft after excluding the high permeability “stringers” within the fracture
geometry. Overall, the fracture remained within the 2000 ft grid size built for this
model.
In Figure 4.49, actual treatment data as well as simulated matched data
including the simulated bottomhole pressure data for Stage 4 are displayed. Figure
4.50 displays the same information with the exclusion of the simulated
bottomhole pressure data. In order to match the pressure data, the tectonic strain
was increased to approximately 300 microstrains whereas the perforation
coefficient of discharge was changed to 0.65. The Friction Coefficient Factor
when proppant concentration is 1 pound per gallon is set to be lower than 1.0 at
0.8 and for proppant concentration at 2 pound per gallon, the coefficient factor is
set to be 1.3 to increase the friction inside the pipe. Advanced parameters
including the MSF are set to 0.003/psi meanwhile a PDL value of 0.002/psi was
entered for this model. Also, as described in Section 4.4.1, the “step” shape during
late-time is visible in these figures.

T ubing P ressure (psi) A
Slurry Rat e (bpm) B
P roppant Concent rat ion (lb/gal) C
GOHFER Surface P ressure (psi) A
GOHFER Slurry Rat e (bpm) B
GOHFER Surface P rop Conc (lb/gal) C
9000 120 20
8500 19
110 18
8000
100 17
7500 16
7000 90 15
6500 14
80 13
6000
70 12
5500 11
5000 60 10
4500 9
50 8
4000
40 7
3500 6
3000 30 5
2500 4
20 3
2000
10 2
1500 1
1000 0 0
11:40 11:45 11:50 11:55 12:00 12:05 12:10 12:15
1/30/2009 1/30/2009
Time
GohWin v1.6.5
Well Description: UWI: 22-Aug-09 17:05
is seen at time 12:09 in this graph in the observed tubing pressure does not significantly appear in the bottomhole pressure
123
7000 120 20
19
6500 110 18
6000 100 17
16
5500 90 15
14
5000 80 13
4500 70 12
11
4000 60 10
9
3500 50 8
3000 40 7
6
2500 30 5
4
2000 20 3
1500 10 2
1
1000 0 0
11:40 11:45 11:50 11:55 12:00 12:05 12:10 12:15
1/30/2009 1/30/2009
Time
GohWin v1.6.5
Well Description: UWI: 22-Aug-09 17:04
Figure 4.47: Graph showing the matched GOHFER™ pressure (aqua) in comparison to the actual pressure (red)
measured during the fracturing job of Stage 2 for Well D1. Fairly good matches are observed except for the 150 psi
pressure differences during proppant injection.
124
Perforations
125
126
Figure 4.51 displays the simulated final fracture geometry from Stage 4.
The proppant concentration is fairly distributed within the fractured region, along
with numerous high permeability streaks. A higher amount of proppant is
concentrated at the bottom of the perforated interval as well as the upper portion
of the perforated interval. The fracture total length for this stage is 800 ft. The
fracture also grew out-of-zone into the shale region above and below the
producing interval since the shale are only present in thin layers, hence reducing
the fracture width but did not terminate the fracture.
The actual treatment data and GOHFER™ simulated data are displayed in
both Figures 4.52 and 4.53 (Figure 4.52 also includes the simulated bottomhole
pressure). Increasing the tectonic strain to 140 microstrains helped to match the
actual and simulated pressure curves. Also, the perforation coefficient of
discharge was decreased to 0.47 to raise the total pressure at early-time. The
friction coefficient factor, on the other hand, was set to less to 0.4 during late-time
as an attempt to imitate the late-time and shut-in pressure of the actual treatment.
The “step” shape during late-time pressure can be seen for this stage. Also, the
advanced parameters play some role in this stage where the value of MSF was set
at 0.004/psi, the PDL was at 0.0025/psi, and the TSC was entered as 0.002/psi.
Final fracture geometry for this stage is displayed in Figure 4.54. The height
containment for this fracture is poor with out-of-zone growth of about 180 ft
above and below the perforated producing interval. There are some high proppant
9000 120 20
8500 110 18
8000
7500 100
16
7000 90
6500 14
80
6000
70 12
5500
5000 60 10
4500 50 8
4000
3500 40
6
3000 30
2500 4
20
2000
10 2
1500
1000 0 0
13:10 13:15 13:20 13:25 13:30 13:35 13:40 13:45
12/15/2008 12/15/2008
Time
GohWin v1.6.5
is seen at time 13:39 in this graph in the observed tubing pressure does not significantly appear in the bottomhole
pressure.
127
7000 120 20
6500 110 18
6000 100
16
5500 90
14
5000 80
4500 70 12
4000 60 10
3500 50 8
3000 40
6
2500 30
4
2000 20
1500 10 2
1000 0 0
13:10 13:15 13:20 13:25 13:30 13:35 13:40 13:45
12/15/2008 12/15/2008
Time
GohWin v1.6.5
128
High permeability streaks
Perforations
129
130
concentration areas at the middle of the perforated zone, in correspondence with
some thin sandstone layers. The fracture total length is estimated to be 615 ft.
Figures 4.55 and 4.56 show the GOHFER™ simulated data that are
matched to the actual treatment data (Figure 4.45 also includes the simulated
bottomhole pressure). The pressure matching process was done by adjusting
several parameters: tectonic strain was increased to 100 microstrains, MSF was
set at 0.005/psi, PDL was set at 0.003/psi, and TSC was set at 0.002/psi. To match
the early- and late-time stage, the friction coefficient factor was set to 0.5 and 0.1
respectively, whereas the perforation coefficient of discharge was kept at the same
default value of 0.55.
Figure 4.57 shows the resulted final fracture geometry for Stage 7.
Fracture height containment is very poor in this stage, especially at the region
above the perforated interval where shale is present in thick layers. Fracture total
length is 770 ft with the highest proppant concentration near a few perforated
sand layers. The fracture also grew into the lower zone of the perforated interval
into some shale and shaly sand layers.
4.4.6 Well S1 – Stage 2
The pressure, pumping flow rate, and proppant concentration of the actual
treatment and the simulated matched data for Stage 2 of Well S1 are displayed in
Figure 4.58. At the early-time, the simulated pressure data was matched to the
T ubing P ressure (psi) A P roppant Concent rat ion (lb/gal) C
Slurry Rat e (bpm) B GOHFER Surface P ressure (psi) A
GOHFER Slurry Rat e (bpm) B GOHFER Surface P rop Conc (lb/gal) C
9000 120 20
8500
110 18
8000
100
7500 16
7000 90
6500 14
80
6000
70 12
5500
5000 60 10
4500
50 8
4000
40
3500 6
3000 30
2500 4
20
2000
10 2
1500
1000 0 0
14:57:30 15:01:30 15:05:30 15:09:30 15:13:30 15:17:30 15:21:30 15:25:30
1/30/2009 1/30/2009
Time
GohWin v1.6.5
is seen at time 15:20 in this graph in the observed tubing pressure does not significantly appear in the bottomhole pressure.
131
T ubing P ressure (psi) A P roppant Concent rat ion (lb/gal) C
Slurry Rat e (bpm) B GOHFER Surface P ressure (psi) A
7000 120 20
6500 110 18
6000 100
16
5500 90
14
5000 80
4500 70 12
4000 60 10
3500 50 8
3000 40
6
2500 30
4
2000 20
1500 10 2
1000 0 0
14:57:30 15:01:30 15:05:30 15:09:30 15:13:30 15:17:30 15:21:30 15:25:30
1/30/2009 1/30/2009
Time
GohWin v1.6.5
132
Perforations
133
7000 120 20
6500 110 18
6000 100
16
5500 90
14
5000 80
4500 70 12
4000 60 10
3500 50 8
3000 40
6
2500 30
4
2000 20
1500 10 2
1000 0 0
09:54 10:00 10:06 10:12 10:18 10:24 10:30 10:36 10:42
12/15/2008 12/15/2008
Time
GohWin v1.6.5
is seen at time 10:39 in this graph in the observed tubing pressure does not significantly appear in the bottomhole pressure.
134
6000 120 20
5500 110 18
100
5000 16
90
4500 14
80
4000 70 12
3500 60 10
3000 50 8
40
2500 6
30
2000 4
20
1500 10 2
1000 0 0
09:54 10:00 10:06 10:12 10:18 10:24 10:30 10:36 10:42
12/15/2008 12/15/2008
Time
GohWin v1.6.5
Fairly good matches are observed except for the high pressure peak during shut-in.
135
Perforations
136
137
actual pressure treatment data by keeping the perforation coefficient of discharge at
its default value of 0.55.
The initial simulated pressure curve and actual pressure curve show a
difference of approximately 500 psi. However, adjusting the tectonic strain to the
amount of 200 microstrains shifted the simulated curve upwards to match the
actual curve. At the late-time stage, the friction coefficient factors are decreased to
0.2 to match the general shape of the actual pressure near shut in time.The proppant
concentration injected along with the slurry flow rate is at the increment of one
pound per gallon at every five minutes of the fracturing job. The slurry rate was
kept at a constant value throughout the process at approximately 50 barrels per
minute.
The only advanced parameters adjusted are the PDL and TSC from zero to
0.001/psi to slightly increase the stiffness of the rock and the tortuosity storage
factor. In Figure 4.59, the output from GOHFER™ is shown as the simulated
fracture geometry in terms of proppant concentration distribution inside the
fracture. As can be seen, the proppant are highly concentrated near perforations
holes surrounding the wellbore, as well as, at the upper zone above the perforated
interval at 7780 ft to 7830 ft. This out-of-zone growth occurred as the fracture
grows into the shale layer overlaying the perforated sand section, which result in
proppant transported to non-productive areas. The fracture does not display much
height containment as the fracture grew 200 ft above its perforated interval as well
as 300 ft lower than its perforated range. The fracture total length is 570 ft.
measured during the fracturing job of Stage 2 for Well S1. Good matches are shown throughout the
treatment process for this stage.
138
Perforations
Figure 4.59: GOHFER™ output showing proppant concentration (lb/ft2) inside the fracture of Stage 2 for Well S1. The
139
140
Figure 4.60 shows the surface pressure, pumping flow rate, and proppant
concentration of the actual treatment and the simulated matched data for Stage 3 of
Well S1. To match the simulated pressure to the actual pressure, the tectonic strain
was increased to 140 microstrains. The perforation coefficient of discharge was
lowered to 0.50 and at the late-time stage, the friction coefficient factor was also
lowered to 0.6 to decrease the friction in the pipe. The slurry rate for this stage was
constant at 50 barrels per minute whereas the proppant concentration steadily
increased at one pound per gallon every five minutes. The MSF, PDL, and TSC
were all set to 0.001/psi to match the characteristics at early-time.
Figure 4.61 shows the final simulated fracture geometry in terms of
proppant concentration. In contrast to the fracture geometry seen in Stage 2 (Figure
4.59), this fracture shows more height containment within the perforated interval.
The total length for this fracture is 1385 ft. This can be an advantage to the
production as the fracture is able to tap more hydrocarbon resources further away
from the wellbore. Also, large amounts of proppant are concentrated at the
shallower perforated sections. A reasonable explanation to this situation is the
possibility of proppant buildup near the wellbore area which later grew into the
above bounding shale layers.
The surface pressure, pumping flow rate, and proppant concentration of the
actual treatment and the simulated matched data for Stage 4 of Well S1 are shown
measured during the fracturing job of Stage 3 for Well S1. Good matches are shown throughout the
treatment process for this stage
141
Perforations
142
143
in Figure 4.62. For this stage, adjustment to 180 microstrains was made to shift
the simulated curve upward while attempting to match the pressure data. At the
early-time, the simulated pressure data was matched to the actual pressure
treatment data without making any additional adjustments. Also, no advanced
parameters were changed in this particular stage. The perforation coefficient of
discharge is kept at its default value of 0.55. During the middle- and late-time
stages, the friction coefficient of factor was set to 0.8 and 0.5 respectively to
decrease the friction in the pipe so that the pressure data matched.
The final fracture geometry for Stage 4 is shown in Figure 4.63. Similar
to fracture geometry in Stage 2 of this well (Figure 4.59), the fracture exhibits
poor height containment within the perforated interval. The proppant is
transported 200 ft beyond the upper perforated interval and approximately 160 ft
deeper than the lower perforated interval. There are a lot of high streak “stringers”
that branched out of the main fracture body in which some of the streaks seemed
to grow out of the 1600 ft grid size setup. The fracture total length for this fracture
is measured at 965 ft. The proppant concentration distribution follows a general
trend where the proppant amount increases downwards in the interval with the
highest concentration at the bottom of the fracture in Stage 4.
Actual treatment data, the surface pressure, slurry rate, and proppant
concentration, as well as the GOHFER™ simulated data for Stage 5 are shown in
Figure 4.64. The only adjustment made in this stage to match the pressure data
measured during the fracturing job of Stage 4 for Well S1. Good matches are shown throughout the treatment process
for this stage except for the 250 psi difference in between time 12.02 to 12.15.
144
Perforations
145
146
was changing the pressure coefficient of discharge to 0.6 and decreasing the
friction coefficient factor to be less than 1.0 during early- and late-time stages.
Figure 4.65 represents the fracture geometry for Stage 5 from the
hydraulic fracturing job performed on Well S1. Although there are a number of
“stringers” in this fracture profile, the proppant concentration is highly distributed
near the wellbore region. In fact, the proppant are evenly and highly distributed
along the entire perforated depth interval. It also appears that the fracture grew
downwards for approximately 150 ft beyond the lower perforated interval and
some proppant are settling at the bottom of the fracture. The fracture total length
for this fracture is 660 ft.
In Figure 4.66, the surface pressure, pumping flow rate, and proppant
concentration from the actual treatment data along with the simulated matched
data are displayed for Stage 9. The pressure matching process for this stage was
by far the simplest one. The only modifications made were increasing the tectonic
strain to 115 microstrains and adjusting the friction coefficient factor at the late-
time stage to 0.1 in order to match the late-time pressure as the shut in time
approaches. The perforation coefficient of discharge was kept at the default value
of 0.55 and no advanced parameter values were changed.
Figure 4.67 displays the final simulated fracture geometry for this stage.
The fracture grew upwards and downwards away from the perforated interval by
approximately 300 ft in both directions. The area that has the most proppant
measured during the fracturing job of Stage 5 for Well S1. Fairly good matches are shown throughout
the treatment process for this stage.
147
Perforations
148
149
concentration is along the perforated depth interval surrounding the wellbore. The
fracture total length is 672 ft. Although this is among the shortest fracture total
length measured for this well, the proppant seemed to be evenly distributed within
the interval rather than settling at any specific zone in this stage.
4.5 Crosswell Tomograms
As mentioned in Chapter 1, the list of input data provided for this project
includes two sets of crosswell tomogram data. These data were also imported into
Transform™ software for viewing purposes and potential analysis. Figure 4.68
shows the map view of the crosswell tomograms locations with Profile 1 between
Wells S1 and S2, and Profile 2 between Wells S1 and S3.
Figure 4.69 shows a side view of both tomogram profiles with Well S1 as
the axis. Figures 4.70 and 4.71 display the individual tomogram for Profile 1 and
Profile 2, respectively. As seen in Figure 4.72, the majority of the microseismic
events detected in Well S1 are on the opposite side of the two tomogram profiles
provided for this project. Hence, the tomograms are less than ideal to be
incorporated into this work as no useful correlation can be made in between the
crosswell seismic profiles and the microseismic events. These profiles are
potentially useful for other studies such as determining sand body termination in
the stacked fluvial reservoir environment in order to better plan infill drilling
within the current area spacing for production optimization.

7000 120 20
6500 110 18
6000 100
16
5500 90
14
5000 80
4500 70 12
4000 60 10
3500 50 8
3000 40
6
2500 30
4
2000 20
1500 10 2
1000 0 0
14:10 14:15 14:20 14:25 14:30 14:35 14:40 14:45
7/14/2008 7/14/2008
Time
GohWin v1.6.5
measured during the fracturing job of Stage 9 for Well S1. Fairly good matches are shown throughout
the treatment process for this stage
150
Perforations
151
152
Well S2
Well S1
Well S3
Figure 4.68: Map view of two crosswell tomogram profiles between Wells S1-S2
and S1-S3. The colored dots (located in the dotted orange circle) are the
microseismic events detected for Well S1.
153
Figure 4.69: The two crosswell tomogram profiles (Profile 1 on the left-hand side
and Profile 2 on the right-hand side) with Well S1 as the axis.
154
Figure 4.70: Profile 1 of the crosswell tomogram between Wells S1 and S2.
Figure 4.71: Profile 2 of the crosswell tomogram between Wells S3 and S1.
155
Figure 4.72: Side view of both crosswell tomogram profiles showing the location
of the microseismic events with respect to Profiles 1 and 2.
4.6 Data Integration
Using the simulated fracture geometries acquired from the GOHFER™
models and the known microseismic event locations, the integration process can
be performed. Fracture geometries from each stage are imported into the
Transform™ software to overlay them with the microseismic activities that
occurred during each hydraulic fracturing stage. By overlaying the microseismic
events on top of the fracture profile, the match characteristics of the

156
microseismicity with respect to the simulated fractures can be analyzed to achieve
Objective 3 of this project. Data integration results are presented in Chapter 5.

157
CHAPTER 5
DATA INTEGRATION RESULTS AND DISCUSSION
The results for each stage from the two wells are presented in this chapter.
Overlaying the simulated fracture profiles from GOHFER™ with the
microseismic events for each stage leads to the analyses and conclusions on the
matching characteristics between the microseismic with respect to the hydraulic
fracture. These results are discussed extensively in this chapter.
5.1 Discussion of Results From Downhole-monitored Well (D1)
As previously stated in Chapter 4, only five out of seven stages from the
treatment of Well D1 are studied in this project. These five stages are Stages 1, 2,
4, 5, and 7. General observations of the microseismic event locations in this well
show fairly good agreement with the simulated fracture geometry from
GOHFER™. Also, the microseisms show a preferential growth trend towards the
east side of the wellbore, advancing towards the monitoring well. Probable
explanations for such behavior are discussed in the following sections. In the next
sections, the results for the individual stages from Well D1 are displayed and
discussed.
158
As shown and discussed in Chapter 4, fracture modeling of the fracturing
treatment was conducted to ensure the match of the observed pressure data and to
approximately match the fracture geometry. The fracture geometry predicted by
GOHFER™ is plotted with the microseismic data as shown in Figure 5.1. In
Stage 1, Well D1 was stimulated from 8273 ft to 8548 ft at 120o phasing, 3 shots
per foot. From Figure 5.1, it is observed that the microseisms show a fair
representation of the simulated fracture geometry.
This microseismic mapping result indicated a fracture total length of 700
ft with a fracture height of 397 ft. The mapped microseisms show a slightly
asymmetric shape towards the SE. Based on this figure, there is some upward
growth of the microseismic activity past the perforated interval beyond the top of
the fracture. Using the option to “Activate Time Playback” in the Transform™
software, each microseism occurrence with respect to time can be viewed as the
well is being hydraulically fractured. For this stage, most microseismic activities
occurred at the center of the wellbore during the early time and later move away
from the wellbore. The time playback also shows that most of the earlier
microseisms occurred at the deepest level of the perforated section.
The fracture geometry predicted by the fracture model shows a fracture
total length of 610 ft and a fracture height of 285 ft. In comparison to the
microseismic mapping measurements, the simulated fracture half-length is 15%
shorter than the microseismic mapped total length whereas the simulated fracture
height is 40% less than the microseismic mapped fracture height. According to
Figure 5.1: Profile view of microseismic events (blue dots) that occurred during Stage 1 hydraulic fracturing
treatment in Well D1 overlapped with the GOHFER™ modeling results. The right and left side of this figure points
towards the SW and SE, respectively.
159
160
Liu et. al. (2009), it is common when observed fractures give longer half-length
measurements in comparison to the half-lengths predicted by fracturing models.
Also, Liu et. al. (2009) stated that fracture height measured from a fracturing
model is usually less than the actual mapped fracture height. This is because
height growth is often considerably less than expected based on the stress
contrasts, presumably due to relatively complicated vertical fracture mechanics
across the various layered materials. In addition, such behavior could also be
contributed to by other factors such as thin high stress zone layers or high leakoff
into the formation. However, for this stage, this is not the case when the model
fracture height is actually less than the measurement mapped by microseismic
events.
It is known that a dominant source mechanism of hydraulic fracture
induced seismicity is shear slip. Such slip is induced by elevated pore pressure
that reduces normal stress along pre-existing fractures (Pearson, 1981). According
to Green and Sneddon (1950), slip can also occur near the tips of created fractures
due to large shear stresses generated by tensile opening. Since shear slip can be
triggered by small pore-pressure increases relative to crack-opening pressure, it is
a fair assumption to expect the microseismic events to extend into the rock
beyond the opened hydraulic fracture lengths and widths (Evans et al., 1999).
Hence, the measured fracture half-length discrepancies in between the
microseimic mapped events and the simulated fracture from the fracture model
could be explained from this phenomenon.

161
Another observation on the microseisms captured using the downhole
methodology is the width of the microseismic “cloud” that presumably surrounds
the fractured area in the stimulated section. Figure 5.2 shows the map view of
Well D1 along with the microseismic events that occurred during Stage 1. The
width of the cloud is approximated at about 150 ft with microseismic events
tightly clustered within the cloud. A good visual comparison is made later in this
chapter when the surface-monitored microseismic events are brought into
discussion. Also, note that all simulated fracture geometries from the fracture
models are displayed as symmetrical fractures.
Results from the overlapping of the simulated fracture geometry with
microseismic events for Stage 2 of the treatment are shown in Figure 5.3. In Stage
2, Well D1 was stimulated from 7975 ft to 8188 ft at 120o phasing, 3 shots per
foot. From Figure 5.3, it is observed that the microseismic events agree well with
the simulated fracture geometry.
The microseismic mapping results indicated a fracture total length of 900
ft with a fracture height of 275 ft. The mapped fracture is asymmetric to the SE
direction. The activation of time playback displays microseisms occurring mostly
in the center during early time and started to occur sporadically with time before
they start growing to the SE side.
The fracture geometry from the model gives a fracture total length of 805
ft with fracture height of 260 ft. In comparison to the microseismic mapping

162
results, the fracture total length from the model is 95 ft less than the microseismic
mapping result whereas the fracture height from the model is higher than previous
method by 15 ft. Reasons for such discrepancies are possibly related to several of
the reasons mentioned in Section 5.1.1.
Stage 1 hydraulic fracturing treatment in Well D1. The right side of the figure is
the East side whereas the left side of the figure is the West side.
Figure 5.3: Profile view of microseismic events (red dots) that occurred during Stage 2 hydraulic fracturing treatment
in Well D1 overlapped with the GOHFER™ modeling results. The right and left side of this figure points towards the
SW and SE, respectively.
163
164
Another observation made from the overlapping of the fracture geometry
profile from the fracture model and the microseismic mapping results is the
microseismic cloud width. As shown in Figure 5.4, the microseisms mapped in
Stage 2 are also sporadically close to one another with highest microseisms
density on the SE side of the wellbore. Also, the width of the microseismic cloud
is approximately 200 ft.
165
foot. From Figure 5.5, it is observed that the microseismic events show a fair
agreement with the simulated fracture geometry.
in the center during early time and starting to grow downward with time before
they settle to the SE side.
ft with a fracture height of 470 ft. To numerically compare the microseismic
mapping results to the measurement from the model, fracture total length as well
as fracture height from the model are both less than the microseismic mapping
result. However, based on Figure 5.5, a few high permeability streaks at depth
7450 ft and 7500 ft did overlay with the farthest microseismic events to the right
of the wellbore. Also, height-wise, the microseismic events do cover a good
portion of the simulated fracture height, making the height measurement
discrepancy (30 ft) a minor issue.
Figure 5.6 shows the map view of the mapped microseisms. In this figure,
it is clearly seen how the microseisms are very dense on the SE side of the
wellbore. The microseismic events cloud is very long compared to the

Figure 5.5: Profile view of microseismic events (orange dots) that occurred during Stage 4 hydraulic fracturing
166
167
previous stages and the cloud width is about 150 ft. Again, the microseismic
events are distributed very close to each other with the exception of event
activities on the west side with two microseism outliers.
168
agreement with the simulated fracture geometry with preferential microseisms
occurrence on the east side of Well D1.
in the center during early time and started to move away from the wellbore and
later settle to the SE side.
ft with fracture height of 425 ft. In comparison to the microseismic mapping
results, the fracture total length of the model is significantly less than the total
length mapped by the microseisms whereas the fracture height from the model is
significantly more than the microseismic mapping result. The simulated fracture
geometry did not show as good height containment as indicated by the
microseismic events. Also, the simulated fracture indicated shorter facture total
lengths that did not match with the microseismic event locations. Hence, this is
not as good a comparison of the two results as in the previous stages.
The map view of the microseismic events for Stage 5 is displayed in
Figure 5.8. In this figure, the microseismic events are shown to be tightly spaced
to one another. Hence, the microseismic cloud width is as small as 100 ft. Again,
it is obvious that the fracture is visually asymmetric to the East side of the
wellbore with few events detected past the observation well, Well D2.
Figure 5.7: Profile view of microseismic events (green dots) that occurred during Stage 5 hydraulic fracturing
169
170
171
agreement with the simulated fracture geometry in terms of fracture half-length,
but poor agreement when it comes to fracture height.
ft with a fracture height of 278 ft. The mapped fracture is very asymmetric to the
SE direction. The activation of time playback shows that the microseisms are
symmetric halfway into the treatment, and then later grow in great numbers to the
East of the wellbore.
ft with a fracture height of 675 ft. In comparison to the microseismic mapping
results, the fracture total length of the model is significantly less than the total
length mapped by the microseisms whereas the fracture height from the model is
significantly more than the microseismic mapping result. Similar to the fracture
simulated in Stage 5 of the fracture model, the simulated fracture geometry did
not show as good height containment as presented by the microseismic events.
However, some of the high-permeability fracture streaks in the fracture created in
Stage 7 did overlap with microseismic events, even the farthest event away from
the wellbore. Hence, the microseism locations are still in considerable agreement
with the fracture total length from the model.
The map view of the microseismic events for Stage 7 is displayed in
Figure 5.10. In this figure, the microseismic events are shown to be tightly spaced
to one another especially on the East side. Hence, the microseismic cloud width is
as small as 100 ft. Again, the fracture is visually asymmetric to the east side of the
wellbore with few events detected passed the observation well, Well D2.
Figure 5.9: Profile view of microseismic events (pink dots) that occurred during Stage 7 hydraulic fracturing treatment in Well D1
overlapped with the GOHFER™ modeling results. The right and left side of this figure points towards the SW and SE, respectively.
172
173
Figure 5.10: Map view of microseismic events (pink dots) that occurred during
174
5.2 Discussion of Results From Surface-monitored Well (S1)
The same procedure performed on Well D1 (downhole-monitored) is also
applied to Well S1 (surface-monitored). The simulated fracture geometry profiles
resulting from the fracture models are overlapped with microseismic events
mapped using the surface geophone receivers. Results from the overlapping
process are displayed and discussed in the next few sub-sections. The five stages
studied in this project for Well S1 are Stages 2, 3, 4, 5, and 9. General observation
of the microseismic events shows fewer events detected by the surface receivers
in comparison to the number of event detected using the downhole receivers.
Also, there are some difficulties in determining fracture half-length for the
fractures mapped using this methodology due to the lack of the bi-wing fracture
shape. Hence, fracture half-length measurement is not determined for the studied
stages but instead, the total length of the fracture is used for comparisons. In the
next five sections, the results for the individual stages from Well S1 are displayed
and discussed.
The fracture geometry predicted by GOHFER™ is plotted with the
microseismic data as shown in Figure 5.11. In Stage 2, Well S1 was stimulated
from 7950 ft to 8071 ft at 180o phasing, 2 shots per foot. From Figure 5.11, it is
observed that the number of microseisms shown is very small with only seven
microseismic events detected.

Figure 5.11: Profile view of microseismic events (blue dots) that occurred during Stage 2 hydraulic fracturing
treatment in Well S1 overlapped with the GOHFER™ modeling results. The right and left side of this figure points
175
176
The total length of fracture estimated from microseismic mapping is 750 ft
with a possible fracture half-length of 375 ft. The fracture height is estimated to
be 543 ft. The microseismic events tend to grow towards the SE direction, away
from the wellbore. Using the “Activate Time Playback” function in the
Transform™ software, the microseismic event occurrences are viewed with
respect to treatment time. Early time events occurred close to wellbore whereas
late time microseismic events occurred away from the wellbore.
The fracture geometry predicted by the fracture model gives a fracture
total length of 570 ft and a fracture height of 420 ft. Due to a possible noisy
environment surrounding the geophone receivers located on the surface,
uncertainties related to travel time picks are a significant source of errors for
events recorded on a surface array. However, this error can be partially lessened
by using a large, redundant array with thousands of receivers (Eisner et. al.,
2009). According to Eisner et. al. (2009), location uncertainties for microseismic
events recorded using surface arrays are much poorly constrained in vertical
direction than horizontal direction. Hence, the general rule is stated that “depth
estimation from a surface array is not as robust as from a downhole array” (Eisner
et. al., 2009). This statement appears to be true for Well S1.
Figure 5.12 shows the map view of the microseismic events from Stage 2.
Figure 5.13 shows the microseismic events and the simulated fracture geometry in
a partial map view for fracture geometries comparison. From Figure 5.12, the
microseisms cloud is shown to be asymmetric. In this case, it is difficult to

177
determine the width of the microseismic events cloud due to the small number of
events and the lack of a cloud shape for this fracture.
Stage 2 hydraulic fracturing treatment in Well S1. The right side of the figure is
178
Stage 2 hydraulic fracturing treatment in Well S1 along with the simulated
fracture profile in the background.
from 7716 ft to 7913 ft at 120o phasing, 3 shots per foot. From Figure 5.14, it is
observed that the microseisms show more activity than the previous stage.
Total fracture length for this stage is 800 ft and the fracture height is 794 ft
based on the microseismic events. Almost all of the microseimic activity extends
to the SE direction of the wellbore. Fracture total length and height from the
model is measured as 1385 ft and 262 ft. The activation of time playback shows
deeper events occurred during early time and the microseisms later grew upward.
Figure 5.14: Profile view of microseismic events (red dots) that occurred during Stage 3 hydraulic fracturing
179
180
Figures 5.15 and 5.16 show the microseismic events from a map view,
with the simulated fracture profile included in the latter figure for fracture
geometry comparison purposes. From Figure 5.15, the microseisms are clearly
seen to occur only on the SE and south side of the wellbore. If a microseismsic
cloud width were to be estimated, it could be up to 600 ft, which is very large.
181
Figure 5.16 Map view of microseismic events (red dots) that occurred during
182
from 7460 ft to 7663 ft at 180o phasing, 2 shots per foot. In Figure 5.17, only ten
microseismic events were captured by the surface receivers. For a big treatment
job with high volume that lasted for almost an hour, the amount of microseismic
events detected is particularly small.
Total fracture length for this stage is 1100 ft whereas the fracture height is
632 ft. Almost all of the microseimic activity extends to the east direction of the
wellbore. Again, the measurement of fracture half-length is not determined for
this stage due to no typical bi-wing fracture pattern being shown. Fracture total
length and height from the model is measured as 965 ft and 590 ft.
The activation of time playback for this stage shows that microseismic
events did not occur during the early and middle of the treatment process. The
first microseism occurrence is a few minutes after the proppant concentration is
maximized at 5 pounds per gallon. Reasons for such a phenomenon are not
known. Shallower events occurred earlier and as time progresses, the microseisms
started to grow downward. Similar to results in the previous stages, the
microseismic event locations are measured a few hundred feet lower than the
perforated interval.
geometry comparison purposes. From Figure 5.18, the microseisms are

Figure 5.17: Profile view of microseismic events (aqua dots) that occurred during Stage 4 hydraulic fracturing
183
184
seen to occur mainly on the east side of the wellbore. No microseismic cloud
width was estimated due to the sporadic nature of the ten microseismic events.
185
from 7213 ft to 7417 ft at 180o phasing, 2 shots per foot. In Figure 5.20, a good
number of microseismic events are captured by the surface receivers.
447 ft. Most of the microseimic activity extends to the east and SE direction of the
wellbore. Fracture total length and height from the model is measured as 660 ft
and 433 ft. The activation of time playback for this stage shows that shallower
events occurred early and the microseisms later grew downward towards the east.
Figure 5.20: Profile view of microseismic events (orange dots) that occurred during Stage 5 hydraulic fracturing
186
187
geometry comparison purposes. From Figure 5.21, the microseisms are clearly
seen to occur only on the SE and south side of the wellbore. The microseismic
event locations appear to very sporadic and not clustered near each other.
188
from 6223 ft to 6371 ft at 180o phasing, 2 shots per foot. In this stage, only eight
microseismic events were captured by the surface receivers.
619 ft. Most of the microseimic activity occurred only to the east side of the
wellbore. On the other hand, fracture total length and height from the model is
measured as 672 ft and 482 ft, respectively.

Figure 5.23: Profile view of microseismic events (green dots) that occurred during Stage 9 hydraulic fracturing
189
190
The activation of time playback for this stage shows that microseismic
events started to occur at the center of the wellbore. Shallower events occurred
earlier, and as time progressed, the microseisms started to grow downward. In
contrast to results from previous stages, the simulated fracture profile is vertically
covered by half the height of the mapped microseisms. In other words, there is
better agreement in depth between the microseismic mapping and the fracture
model.
geometry comparison purposes. In Figure 5.24, the microseisms are observed to
occur only on the SE side of the wellbore. The microseismic event locations are
less sporadic and give a microseismic event cloud width of about 200 ft.
5.3 Downhole Methodology vs. Surface Methodology
After a discussion of results for both Wells D1 and S1, comparisons are
made based on the observations of the downhole and surface monitoring
methodologies. One very apparent difference between the two methods is the
microseismic events distribution.
For Well D1 that was monitored using downhole geophone arrays, there
are approximately more than 30 microseismic events detected per stage for the
studied stages included in this project. This is possibly due to the advantageous
location of the geophone receivers being downhole, hence, reducing the travel
distance in between the source of the microseisms and the receivers. Also, the
191
events detected are tightly clustered to one another with high event density close
to the wellbore and expanded to the east and SE of the wellbore.
192
From a general observation pertaining to the microseismic events recorded
for Well D1, there is a potential distance bias due to the location of the
observation well (Well D2) being located on the east side of Well D1 (hence,
capturing more events on the east side of Well D1). However, such bias could
have potentially been eliminated with the addition of another observation well
situated on the west side of Well D1 to enable possible events on the other half of
the fracture wing to be captured. This solution will provide the ability to detect
microseisms at the optimum level. Hence, more accurate conclusions can be
deduced pertaining to fracture asymmetry/symmetry as the distance bias is
eliminated. Also, note that other probable causes to fracture asymmetric shape
193
mapped by microseismic mapping could be contributed by reservoir heterogeneity
(eg. shorter half-lengths as fluvial sand bodies approach boundary and are
terminated) or a phenomenon called stress shadowing where an existing fracture
can impose stress on a nearby propagating fracture, providing fracture
containment.
Figure 5.26 shows the plot of the microseismic magnitude against the
distance between the event and toolstring. The magnitude scale is measured from
the amplitude of the seismic event where magnitude is related to seismic energy
and each unit of magnitude has an energy increase of 32 times (Pinnacle
Technologies, 2008). For instance, a magnitude of -1 releases 32 times more
energy than a magnitude of -2. The average event magnitude in this well is low at
approximately -3.7 (Pinnacle Technologies, 2008). Hence, this resulted in a
hearing distance of about 720 ft as shown in Figure 5.27. The hearing distance is
plotted as a red circle around the observation well in this figure. From past
experiences of the monitoring company, the distance between Wells D1 and D2
of 700 ft is near the edge of the hearing distance for these formations in the Uinta
basin (Pinnacle Technologies, 2008). The anticipated fracture azimuth would
bring the eastern fracture wings closer to Well D2 and should allow for an
adequate mapping of the fracture geometries in the eastern wings. As a
consequent, the fracture wings on the west side would be too far away from Well
D2 and could not be completely mapped. From this information, it can be
concluded that the asymmetric feature of the microseismic events recorded for
this well is mainly caused by the bias location of the observation well, Well D2.
194
Figure 5.26: The microseismic moment magnitude (y-axis) with respect to true
vertical depth and tool-event distance (x-axis) (Pinnacle Technologies, 2008).
195
Well D1 Well D2
Figure 5.27: This plot shows the hearing distance for all Stages 1 to 7 in Well D1.
The average hearing distance for this well is approximately 720 feet and shown as
a red circle with circle radius of 720 ft. From this figure, it is seen that the
microseismic events on the west side of Well D1 is farther and beyond the scope
of the hearing distance in this formation. Hence, the reason of location bias for the
monitoring well is a good explanation to the asymmetric fracture shape as
monitored by microseismic mapping (Pinnacle Technologies, 2008).
As displayed in Figures 5.1 to 5.10 in Chapter 5, the microseismic events
monitored on Well D1 show good agreement between the fracture geometries
mapped from the microseismic mapping as well as the geometries produced from
the fracture models with very few event outliers. This can be observed aerially
from map view as well as vertically from profile view. Based on this observation,
the measurement of fracture azimuth is obtained for each stage, giving an average
of 102o. By overlaying the microseismic events with the simulated fracture
profiles from GOHFER™, Stage 2 from Well D1 shows the best agreement in
196
between the two data. Figure 5.3 shows that the fracture heights from the
microseismic mapping and the fracture model are almost the same value. This is a
very successful representation of the integration process discussed in Chapter 4.
The revision of the other four stages gives comparable fracture geometry
measurements (half-length and height) to the results produced from the
microseismic mapping.
On the other hand, the microseismic events captured by the surface
geophone arrays provide a different view on the integration process. From the
map views provided in Figures 5.12, 5.15, 5.18, 5.21, and 5.24, it is obvious that
the microseismic event locations are widely spread among each other. In other
words, the surface receivers give a good areal coverage since the receivers are
located on the surface, surrounding the treatment well. However, good areal
coverage and the microseismic data they captured for the surface-monitored
stages neither provide a clear measurement of fracture azimuth nor give a good
estimation of fracture half-length. The layout of the microseismic events cloud is
hard to be determined, thus, reducing the certainty in deducing the fracture
geometries.
Table 5.1 shows the comparisons of fracture geometries between the
fracture model (GOHFER™) and the fracture mapping (microseismic mapping).
The average percent differences of total hydraulic fracture length between the
models and the mapping for Well D1 and Well S1 are 34.07% and 31.73%,
respectively. In terms of fracture total length, the average values for both methods
are comparable with 2.34% difference between the two values. On the other hand,
197
the average percent difference of fracture height for Well D1 is 27.28% whereas
Well S1 gives an average percent difference in fracture height of 54.22% which is
significantly larger than the average value from Well D1. From these percentages,
it is shown that the fracture heights measured using the surface tools gives a larger
percentage difference between the fracture model values and the microseismic
mapped values. Hence, using the downhole tools is a better option in measuring
the vertical height growth of a hydraulic fracture. Also, the fracture height
comparison supports the previous statement by Eisner et al. (2009) indicating that
the surface geophone arrays do not provide a robust of depth estimation in
comparison to the downhole geophone arrays.
Table 5.1: Comparisons of Fracture Geometries between Fracture Model

(GOHFER™) and Fracture Mapping (Microseismic Mapping)
Micro-
GOHFER™ GOHFER™ seismic Micro- Percent Percent
Model Model Mapping seismic Difference Difference
Stage Total Fracture Total Mapping In Total In
Fracture Height Fracture Fracture Fracture Fracture
Length (feet) Length Height Length Height
(feet) (feet) (feet) (%) (%)
1 610 285 700 397 -14.75 -39.30
2 805 260 900 275 -11.80 -5.77
Well D1
4 800 470 1250 500 -56.25 -6.38

5 615 425 850 314 -38.21 26.12
7 770 675 1150 278 -49.35 58.81
Average: 34.07 27.28
2 570 420 750 543 -31.58 -29.29
3 1385 262 800 794 42.24 -203.05
Well S1
4 965 590 1100 632 -13.99 -7.12

5 660 433 1100 447 -66.67 -3.23
9 672 482 700 619 -4.17 -28.42
Average: 31.73 54.22
198
After reviewing the overlapping results for both wells, it is fair to
conclude that the discrepancies between the mapped microseismic events and the
simulated fracture models, particularly for fracture height for both methodologies,
demonstrate the difficulty to obtain sufficient information to obtain completely
matched results. Also, since microseisms do not necessarily occur on the
hydraulic fracture itself (but can in fact be some distance to the side and/or ahead
of fracture), it is difficult to determine completely accurate geometries of the
fracture. According to Warpinski et al. (1999) in their study of the B-Sand at the
M-site in the Piceance basin of Colorado, analyses show that microseismic events
can occur as much as: 12-15 ft ahead of the tip of the fracture and 15-20 ft normal
to the tip of the fracture because of the large stress concentrations at the fracture
tip; and several tens of feet normal to the body of the fracture due to leakoff
effects (increases in pore pressure).
Other factors that can contribute to microseismic event location
uncertainty are the distance from microseism to receiver and the azimuth of
microseismic signals. From the same study done by Warpinski et al. (1999),
accuracy of the azimuth to the microseism is about ±5o whereas uncertainty in the
distance from the receiver to the microseism varies considerably depending upon
the number of levels on which the S-wave arrival is accurately detected, but it is
approximately on the order of ±50 ft. Given the uncertainties in microseism
locations, the microseismic results in this research and other situations should be
considered as an envelope that “surrounds” the fracture.

199
The strength of microseisms recorded during a treatment is influenced by
several factors such as:
• Injection rates, pressure, along with fluid type, proppant
concentration, and temperature of injected fluid;
• controlling stress and pressure changes in the formation;
• Environmental factors associated with the formation where the
fracture is created (which includes the stiffness of the rock,
tectonic stresses, and pre-existing fractures/fault); and,
• Production history of the region (resulting in pore-pressure and
stress changes).
Also, microseism viewing distance does vary by formation and basin. For
example, viewing distance in the Barnett shale is approximately 4000 ft whereas
the viewing distance in the Rocky Mountain region is approximated at 1500 ft.
Unfortunately, noisy environments surrounding geophone receivers, both surface
and downhole, can possibly reduce the viewing distance of microseismic events.
Noisy environments can occur for many reasons. Some examples are:
• Monitoring from an observation well that is on the same pad as the
treatment well;
• Monitoring in a field that has ongoing drilling near receivers;
• Monitoring from an observation well that is intersected by the
created hydraulic fractures; and,

200
• Monitoring in a field with closely spaced wells undergoing
artificial lift.
In addition to noise, the quality of the microseismic signal is also impacted
by signal attenuation as a result from geometry spreading and absorption along
the travel path. With increasing travel distances between source and receivers
(especially for surface geophone arrays), signal strength will decrease which
ultimately limits the range over which microseisms can be detected as signal
strength approaches background noise level.
As introduced in Chapter 1, the main focus of this work is to provide an
objective comparison between the results of modeling methodologies in
conjunction with microseismic integration for the two techniques: the
conventional downhole microseismic monitoring and the unconventional surface
monitoring. Both methodologies use similar physical modus operandi which
requires a source of microseismic events (ie. shear slippage), geophone receivers
(downhole/surface), and accurate velocity model to ensure correct determination
of microseismic event locations, hence mapping out the possible geometry of a
hydraulic fracture in the subsurface.
The 27 figures in this chapter help to conclude the outcome of this project
in order to meet the third project objective which is to determine the match
characteristics of the downhole and surface microseismic data to the hydraulic
fracture models build for both Wells D1 and S1. Conclusions and a list of
recommendations to further study these methodologies and increase efficiency of
microseismic and fracture modeling are further discussed in the final chapter.
201
CHAPTER 6
CONCLUSIONS AND RECOMMENDATIONS
In a tight gas sand reservoir such as the one studied in this project,
hydraulic fracturing treatments are required in order to acquire and sustain
commercial production. Although the exact distance at which leakoff-induced
microseisms might occur is difficult to determine due to factors such as unknown
permeability of natural fractures in the formation and uncertainties from the
velocity model, the analyses demonstrated in this thesis provide guidance in
interpreting the microseismic data in conjunction with hydraulic fracture
modeling and show that the microseisms map out an envelope in which the
hydraulic fracture is embedded. Conclusions regarding mapping and associated
modeling from this research project, along with recommendations for future work,
are provided in the following sections.
6.1 Conclusions
After constructing and pressure matching the fracture models for Wells D1
and S1 and integrating the resulting fracture profiles with the microseismic events
captured by surface and downhole geophone arrays, the following conclusions are
made:
• The results of input integration for Well D1 (downhole-monitored)
showed good agreement between the microseismic mapping

202
measurements and the fracture geometries simulated by GOHFER™.
Stage 2 from the Well D1 treatment gives the best agreement in terms
of fracture height and fracture total length with percent difference of
11.80% for fracture total length and 5.77% for fracture height when
the fracture geometries from the microseismic mapping were
compared to the values from the fracture models.
• The results of input integration for Well S1 (surface-monitored)
showed questionable and poor agreement between the microseismic
mapping measurements and the fracture geometries simulated by
GOHFER™.
• Downhole microseismic methods are the best option to determine
vertical growth patterns of seismicity induced by hydraulic fracturing
processes since depth estimations from surface arrays are not as strong
as depth estimations from downhole arrays. Values calculated in Table
5.1 supports this conclusion where the average percent difference of
fracture height for Well S1 is 54.22% whereas in Well D1, the value is
only 27.28%.
• Surface geophone receiver arrays are capable of providing good areal
coverage surrounding the treated wellbore as they are able to visually
show the location of microseisms that occurred up to a thousand feet
away from the wellbore. However, some of the stages in Well S1
resulted in small amounts of detected microseismic events. Thus, the
measurements of fracture height and half-length bear more

203
uncertainties due to the lack of microseisms detected by the surface
receivers.
• By combining complimentary strengths and weaknesses of fracture
mapping and modeling, calibrated fracture models are produced from
this work. Fracture models give the ability to predict how changes to a
fracture treatment would alter fracture geometry. Fracture mapping
(microseismic mapping), on the other hand, provides a direct
measurement of fracture geometry from a hydraulic fracturing
treatment but cannot be used to predict what could be happening under
a different set of conditions. Combining both the direct measurements
and the fracture models, calibrated fracture models with good
predictive capabilities can be created.
• Most microseismic occurrences during the Well D1 hydraulic
fracturing treatment were centered near the wellbore and expanded
preferentially towards the southeast of the wellbore. Such asymmetric
fracture growth could be contributed to by several factors including:
bias location of the observation well (microseismic events advancing
towards Well D2), reservoir heterogeneity (termination of sand bodies
in fluvial environment), or possibly limited viewing distance as
(smaller magnitude) microseismic events occur further away from the
wellbore towards the northwest direction. This last option, limited
viewing distance, is likely the reason in this case based on the hearing
distance radius shown in Figure 5.27.

204
6.2 Recommendations and Future Work
In this section, recommendations are made in order to improve further
studies using surface and downhole microseismic techniques in conjunction with
hydraulic fracture modeling. Some of the important steps that can be taken to
enhance the quality of data and minimize uncertainties are suggested as follows:
• Employing both surface and downhole geophone receiver arrays on the
same well would be a beneficial effort as well as a great experimental
process to obtain a fair comparison in between the two methodologies. By
having receiver tools in both locations, microseismic events captured by
both arrays can be compared in terms of locations, fracture heights, and
fracture half-lengths. At the very least, the accuracy of the surface-
monitored microseisms can be compared to the conventional downhole-
monitored data that is known to be well tested to date.
• Performing mini-frac analysis before the actual pumping treatment would
be a good addition to any hydraulic fracturing process as values such as
closure stress gradients and permeability can be obtained through G-
function analysis using data from the mini-frac analysis.
• A hybrid of downhole and surface tools is a practical idea that could
essentially increase the accuracy of microseismic events detected. Two
good combinations would be:
o Using downhole geophone arrays with surface tiltmeter tools; and,
o Using surface geophone arrays along with downhole tiltmeter
tools.
205
The microseismic data provides a detailed map of micro-earthquakes that
are induced by the hydraulic fracturing process. The tiltmeter, on the other
hand, provides an integrated measure of the actual rock deformation.
Separately, both tools provide important information about fracture
growth, the process, and the resultant geometries. Jointly, however, the
combination of the tools can provide a much more complete or composite
view of the fracture.
• Since perfectly straddling the downhole toolstring along the stimulated
interval would be very difficult, the toolstring should be placed as close as
possible to the target depth as this will minimize the depth error and
increase signal-to-noise ratio.
• For both surface and downhole arrays, stacking the seismograms is a good
move to overcome low signal-to-noise ratio (Shemeta, et al, 2007). A
typical downhole receiver array consists of 12 levels of receivers spaced at
50 ft apart meanwhile a stacked array has 18 receivers. By digitally
stacking the signals from the receivers, noise cancellation can be obtained,
thus improving signal quality by reducing noise, improving hodogram
linearity, and clearer phase arrivals. As a result, enhanced microseismic
maps are gained by improving data quality of low-amplitude events and
extracting data from events (more and smaller events) that might not be
analyzable without the stacking. When the signal-to-noise ratio is
improved, weak signals are better identified and more accurate arrival
time determination are made possible.

206
• Other recommendations such as a matched filter technique to enhance
signal from induced microseismic (Hulsey et al, 2009), using perforation
timing measurements for velocity calibration (Warpinski et al, 2003), and
using source parameters technique to better describe the fracturing process
by determining the actual failure mechanism (Mayerhofer et al, 2000) are
also suggested for further review.

207
NOMENCLATURE
αv = Vertical Biot’s poroelastic constant (unitless)

αh = Horizontal Biot’s poroelastic constant (unitless)
C = Perforation coefficient of discharge (unitless)
CSV = Comma Separated Values
D = Diameter of perforations (inch)
D1 = Initial diameter (inch)
Dtv = True vertical depth (feet)
DTCO = Transit compressional travel time (microsecond/ft)
DTSM = Transit shear travel time (microsecond/ft)
∆D = Change in diameter (inch)
∆L = Change in length (feet)
E = Young’s Modulus (million psi)
FMI = Formation Micro Imaging
GNB = Greater Natural Buttes
γob = Overburden stress gradient (psi/ft)
γp = Pore fluid gradient (psi/ft)
ε = Strain (microstrain)
εx = Regional horizontal strain (microstrain)
εz = Strain in z-direction, axial strain (microstrain)
ν = Poisson’s Ratio (unitless ratio)
Kperf = Perforation friction coefficient (psi/BPM2)
Knear WB = Near-wellbore friction coefficient (psi/BPM1/2)
L1 = Original length (feet)
L2 = Final length (feet)
LAS = Log ASCII Standard
ML = Local magnitude
MSF = Pressure Dependent Modulus Stiffness Factor (1/psi)
N = Number of perforation (unitless)
Psi = Pound per square inch
Pnet = Net pressure (psi)
Pfracture = Pressure in fracture (psi)
Pclosure = Closure pressure (psi)
Poff = Pore pressure offset (psi)
Pfriction = Friction pressure (psi)
PDL = Pressure Dependent Leakoff (1/psi)
PZS = Process Zone Stress (psi)
PHIE = Effective porosity (decimal)
q = Slurry flow rate (barrel per minute)
ρ = Density (pound per gallon)
σt = Regional horizontal tectonic stress (psi)
σ1 = Overburden pressure (psi)
σ2 = Maximum horizontal pressure (psi)
σ3 = Minimum horizontal pressure (psi)
208
TSC = Transverse Storage Coefficient (1/psi)

UBI = Ultrasonic Borehole Imager
VSHALE = Volume of shale (ft3/ft3)
209
REFERENCES
Abbott, D., Neale, C., Lakings, J., et al. (2007): “Hydraulic Fracture Diagnostics
in The Williams Fork Formation, Piceance Basin, Colorado, Using Surface
Microseismic Monitoring Technology”. Society of Petroleum Engineers Paper
No.108142, p.1-10.
Albright, J.N., Rutledge, J.T., Phillips, W.S. et al, (1994): “Subsurface Fracture
Mapping Using Microearthquakes Detected During Primary Oil Production,
Clinton County, Kentucky”. Paper SPE 28384 presented during the 1994 SPE
Annual Technical Conference and Exhibition, New Orleans, Louisiana, 25-28
September.
Barree, R.D. (2009): “GOHFER™ Fracture Theory Text PDF’s”. Accessed on

July 15, 2009 at: http://gohfer.corelab.com/default.aspx?q=Documents.
Barree, R.D. (2008): Personal Communication.
Blakey, R.C. (2007): “Colorado Plateau Stratigraphy and Geology and Global
and Regional Paleogeography”. Northern Arizona University Website, Accessed
on November, 2007 at: http://jan.ucc.nau.edu/~rcb7/RCB.html
Cuzella, J.J. and Stancel, S.G. (2006): “Greater Natural Buttes Field, Uinta
Basin, Utah”. The Mountain Geologists, Vol. 43, No. 3, July 2006, p. 213-218.
Duncan, P.M. (2004): “Passive Seismic Seeing Active Interest”. Hart’s E & P, p.
47-48.
Economides, M.J., and Nolte, K.G. (1981): “Reservoir Stimulation”.

Schlumberger Educational Services, Houston, Texas, p.2-1 to 2-21.
Eisner, L., Duncan, P.M., Heigl, W.M., and Keller, W.R. (2009): “Uncertainties
in Passive Seismic Monitoring”. The Leading Edge, May 2009, p. 258-263.
Escartin, J., Cowie, P.A., Searle. R.C., et. al., (1949): “Quantifying Tectonic
Strain and Magmatic Accretion at a Slow Spreading Ridge Segment, Mid-Atlantic
Ridge, 290N”. Journal of Geophysical Research, American Geophysical Union,
Washington, DC.
Evans, K.F., Cornet, F.H., Hashida, T., et al, (1999): “Stress and Rock Mechanics
Issues of Relevance to HDR/HWR Engineered Geothermal System: Review of
Developments During the Past 15 Years”. Geothermics, No. 28, p. 455-474.
Galloway, W.E., and Hobday, D.K. (1996): “Terrigenous Clastic Depositional

Systems (2nd edition)”. Springer-Verlag Berlin Heidelberg, p. 489.
210
Gidley, J.L., Holditch, S.A., Nierode, D.E. and Veatch, R.W.J. (1989): “Recent
Advances in Hydraulic Fracturing”. Henry L. Doherty SPE Monograph Series, v.
12, Richardson, Texas, SPE, p. 452.
Green, C.A. (2006): “Hydraulic Fracture Model Sensitivity Analyses of Massively

Stacked Lenticular Reservoirs in the Mesaverde Formation, Southern Piceance
Basin, Colorado”. Master’s thesis, Colorado School of Mines, Golden, Colorado.
Green, A.E., and Sneddon, I.N. (1950): “The Distribution of Stress in the
Neighborhood of a Flat Elliptic Crack in an Elastic Solid”. Proc. Camb. Phil.
Soc. 46, p. 159.
Guiseppe, A.C., and Heller, P.L. (1998): “Long-term River Response to Regional
Doming in the Price River Formation, Central Utah”. Geology, v. 26, No. 3, p.
1661-1678.
Gutierrez, G. (2007): “Analyses of Faults and Fractures, and Their Impact on a

Tight Gas Sand Resource Play in the Uinta Basin”. Master’s thesis, Colorado
School of Mines, Golden, Colorado, p. 2-50.
Hintze, L.F. (1988): “Geologic History of Utah”. Brigham Young University

Geology studies. Special publication No. 7, Brigham Young University, Utah, p.
202.
Hulsey, B.J., Eisner, L., Thornton, M.P., and Jurick, D. (2009): “Application of
Relative Location Technique from Surface Arrays to Microseismicity Induced by
Shale Fracturing”. Paper acquired through personal communication with Leo
Eisner, 2009.
Johnson, R.C., and Roberts, S.B. (2003): “The Mesaverde Total Petroleum
System, Uinta-Piceance Province, Utah and Colorado”. USGS Digital Data
Series DDS-69-B, Chapter 7, p. 65.
Kidney, R.L. (2007): Personal communication.
Kranz, B., Donald, A., Kozlowski, K., and Higgins, S. (2007): “Well S1
Mechanical Earth Model”. Personal communication.
Lama, R.D. and Vutukuri, V.S. (1978): "Handbook of Mechanical Properties of

Rocks". Trans Tech Publications, Clausthal, Germany 1978, 2, 86-89, 145, p. 234.
Liu, X., Zhou, Z.Q., Li, X.W., et al, (2006): “Understanding Hydraulic Fracture
Growth in Tight Oil Reservoirs by Integrating Microseismic Mapping and
Fracture Modeling”. Paper SPE 102372 presented at the 2006 SPE International
Oil & Gas Conference and Exhibition, Beijing, China, 5-7 December.
211
Mayerhofer, M.J., Walker, R.N., Urbanic, T., and Rutledge, J.T. (2000): “East
Texas Hydraulic Fracture Imaging Project: Measuring Hydraulic Fracture
Growth of Conventional Sandfracs and Waterfracs”. Paper SPE 63034 presented
at the 2000 SPE Annual Technical Conference and Exhibition, Dallas, Texas, 1-4
October.
Microseismic Inc. (2007): “Surface Based Passive Seismic Monitoring For

Improved Reservoir Management”. Power Point Presentation from Microseismic
Inc. Website at: http://microseismicinc.com/.
Miskimins, J.L. (2004): Class note to PEGN 426 – Well Stimulation, Colorado
School of Mines, Golden, Colorado.
Miskimins, J.L. (2005): Class note to PEGN 426 – Well Stimulation, Colorado
School of Mines, Golden, Colorado.
Morgan, C.D., Mcclure, K.P., Bereskin, S.R., and McPherson, M.m (2005): “A
Preliminary Discussion of Fault Styles in the Southwest Uinta Basin, Utah”.
RMS AAPG Poster, Laramie, Wyoming.
Montgomery, S.L., David, E.T., and Barker, C.E. (2003): “Analog for Fluvial-
Deltaic Reservoir Modeling: Ferron Sandstone of Utah”. AAPG Studies in
Geology 50.
Neale, C. (2007): Personal communication.
Neale, C. and Duncan, P.M. (2003): “Passive Seismic Uses Nonactive Noise”.
The American Oil and Gas Reporter, p.68-72.
Osmond, J.C., Locke, R., Dille, A.C., Praetorious, W., and Wilkins, J.G. (1968):
“Natural Gas in Uinta Basin, Utah”. AAPG Special Volumes, v. A007, p. 174-
198.
Osmond, J.C., (1992): “Greater Natural Buttes Gas Field, Uinta County, Utah”.
Hydrocarbon and mineral resources of the Uinta Basin, Utah and Colorado: Utah
Geological Association Guidebook 20.
Pantoja, D. (1998): "Petrophysical Analysis of Vuggy Carbonates and Thin

Bedded Sands and Shales for Improved Reservoir Management". Master’s thesis,
Colorado School of Mines, Golden, Colorado, p. 226.
Pearson, C. (1981): “The Relationship between Microseismicity and High Pore

Pressures during Hydraulic Stimulation Experiments in Low Permeability
Granitic Rocks”. J. Geophys. Res. Vol. 86, September, p. 7855-7864.
212
Pinnacle Technologies, Inc. (2008): “Fracture Mapping Results for Well D1”.
Personal communication.
Roundtree, R. (2009): Personal communication.
Shemeta, J.E., Maxwell, S.C., Warpinski, N.R., et al, (2007): “Stacking

Seismograms to Improve Microseismic Images”. Paper SPE 108103 presented at
the 2007 SPE Rocky Mountain Oil & Gas Technology Symposium, Denver,
Colorado, 16-18 April.
Slim, M.I. (2007): “Borehole-Image Log Interpretation and 3D Facies Modeling

in the Mesaverde Group, Greater Natural Buttes Field, Uinta Basin, Utah”.
Master’s thesis, Colorado School of Mines, Golden, Colorado, p.18.
Spencer, C.W. (1987): “Hydrocarbon Generation as a Mechanism For

Overpressuring in Rocky Mountain Region”. AAPG Bulletin, v. 71, p. 368-388.
U. S. Geological Survey (2002), “Assessment of Undiscovered Oil and Gas

Resources of the Uinta-Piceance Province of Colorado and Utah”. USGS Fact
Sheet FS-026-02.
University of Akron Website, Accessed on March 29, 2009 at:

http://www3.uakron.edu/geology/Foos/Energy/3oil10.jpeg
University of Oregon Website, Accessed on March 20, 2008 at:

http://physics.uoregon.edu/~jimbrau/BrauImNew/Chap07/FG07_06.jpg
Warpinski, N.R., Wright, T.B., Uhl, J.E., et al. (1999): “Microseismic Monitoring
of the B Sand Hydraulic Fracture Experiment at the DOE/GRI Multi-Site
Project”. SPE Journal 4 (3), September, p. 242-250.
Warpinski, N.R., Peterson, R.E., Branagan, P.T., Engler, B.P. and Wolhart, S.L.
(1998): “In-Situ Stress and Moduli: Comparison of Values Derived from Multiple
Techniques”. Paper SPE 49190, presented at the 1998 SPE Annual Technical
Conference and Exhibition, New Orleans, Lousiana, 27-30 September, p. 569-
583.
Warpinski, N.R., Wolhart, S.L., and Wright, C.A. (2001): “Analysis and
Prediction of Microseismicity Induced by Hydraulic Fracturing”. Paper SPE
71649 prepared for the 2001 SPE Annual Technical Conference and Exhibition,
New Orleans, Lousiana, 30 September-3 October.
Warpinski, N.R., Sullivan, R.B., Uhl, J.E. et al, (2003): “Improved Microseismic
Fracture Mapping Using Perforation Timing Measurements for Velocity
Calibration”. Paper SPE 84488 presented at the 2003 SPE Annual Technical
Conference and Exhibition, Denver, Colorado, 5-8 October.
213
Warpinski, N.R. (2007): Personal communication.
Warpinski, N.R. (2009): “Integrating Microseismic Monitoring With Well

Completions, Reservoir Behavior, and Rock Mechanics”. Paper SPE 125239
presented at the 2009 SPE Tight Gas Completions Conference, San Antonio,
Texas, 15-17 June.
Willis, R.B., Fontaine, J., Paugh, L., and Griffin, L., (2005): “Geology and
Geometry: A Review of Factors Affecting the Effectiveness of Hydraulic
Fractures”. Paper SPE 97993 prepared for the 2005 SPE Eastern Regional
Meeting, Morgantown, West Virginia, 14-16 September.
Zimmer, U., Maxwell, S.C., Waltman, C., and Warpinski, N.R. (2007):
“Microseismic Monitoring Quality-Control (QC) Reports as an Interpretative
Tool for Nonspecialists”. Paper SPE 110517 presented at the 2007 SPE Annual
Technical Conference and Exhibition, Anaheim, California, 11-14 November.
214
DVD-ROM CONTENTS
APPENDIX A:
• WELL D1-LAS LOG FILES
o WELL D1-LAS LOG
• WELL S1-LAS LOG FILES
o WELL S1-MAIN PEX 006LUP-GenV12
o WELL S1-MEM
• WELL D1-PUMP CURVES
o STAGE 1 PUMP CURVES
• WELL S1-PUMP CURVES
• WELL D1-MICROSEISMIC EVENTS
o STAGE 1
o STAGE 2
o STAGE 3
o STAGE 4
o STAGE 5
o STAGE 6
o STAGE 7
• WELL S1-MICROSEISMIC EVENTS
o STAGE 2
o STAGE 3
o STAGE 4
o STAGE 5
o STAGE 6
o STAGE 7
o STAGE 8
o STAGE 9
• WELL D1-OTHER DATA
• WELL S1-OTHER DATA
215
APPENDIX B:
• WELL D1-GOHFER™ MODELS
o STAGE 1-DOWNHOLE
o STAGE 2-DOWNHOLE
o STAGE 4-DOWNHOLE
o STAGE 5-DOWNHOLE
o STAGE 7-DOWNHOLE
• WELL S1-GOHFER™ MODELS
o STAGE 2-SURFACE
o STAGE 3-SURFACE
o STAGE 4-SURFACE
o STAGE 5-SURFACE
o STAGE 9-SURFACE
• WELL D1-TRANSFORM™ MODEL
• WELL S1-TRANSFORM™ MODEL

Gopher PDF

Uploaded by

Copyright:

Available Formats

Gopher PDF

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

Gopher PDF

Uploaded by

Copyright:

Available Formats

A COMPARISON OF HYDRAULIC FRACTURE MODELING WITH DOWNHOLE

AND SURFACE MICROSEISMIC DATA IN A STACKED FLUVIAL PAY SYSTEM

Nur Azlinda Mohammad

Science (Petroleum Engineering).

Nur Azlinda Mohammad

Dr. Jennifer L. Miskimins

Dr. Ramona Graves

Professor and Department Head,

The work in this project concentrates on comparing the results of hydraulic

fracture mapping and modeling using two microseismic monitoring techniques:

develop detailed post-treatment models of the hydraulic fracturing treatments in the

In this study, GOHFER™, a fully three-dimensional fracture simulator was used

techniques: downhole microseismic monitoring and surface microseismic monitoring.

placed at different locations. Downhole geophone receivers were employed in an

rock mechanical properties were calculated.

agreement for most stages in the downhole-monitored well, whereas comparisons of

surface microseismic mapping measurements with the simulated fracture geometries

LIST OF FIGURES ........................................................................................................... ix

LIST OF TABLES.......................................................................................................... xvii

ACKNOWLEDGMENTS ............................................................................................. xviii

CHAPTER 1 INTRODUCTION .........................................................................................1

1.1 Purpose of This Work ..................................................................................1

1.2 Research Objectives.....................................................................................2

1.3 Study Area (Field Location) ........................................................................2

1.4 Data Set........................................................................................................4

1.5 Recent Works and Applications...................................................................8

CHAPTER 2 LITERATURE REVIEW ............................................................................10

2.1 Geological Overview .................................................................................10

2.1.1 Depositional System ......................................................................10

2.1.2.1 Mancos Shale Formation .................................................12

2.1.2.2 Mesaverde Group.............................................................15

2.1.2.3 North Horn Formation .....................................................15

2.1.2.4 Wasatch Formation ..........................................................16

2.1.2.5 Green River Formation ....................................................16

2.1.3 Producing Formation .....................................................................16

2.1.4 Petroleum System and Reservoir Characterization........................18

2.2.1 Microseismic Events......................................................................20

2.2.2 Downhole-monitored Microseismic ..............................................24

2.2.3 Surface-monitored Microseismic...................................................25

2.2.4 Software Used for Microseismic Analyses....................................29

2.3 Fracture Geometry Modeling.....................................................................29

2.3.1 2-Dimensional Models...................................................................29

2.3.2 Pseudo-3-Dimensional Models......................................................32

2.3.3 3-Dimensional Models...................................................................33

2.3.4 Net-Pressure Analysis....................................................................34

CHAPTER 3 HYDRAULIC FRACTURE MODELING .................................................36

3.1 GOHFER™ Software ................................................................................39

3.2 Modeling Process.......................................................................................41

3.2.1 Input ...............................................................................................42

3.2.2 Output ............................................................................................51

3.3 Pressure Matching Process ........................................................................54

CHAPTER 4 MICROSEISMIC EVENTS AND HYDRAULIC FRACTURE

4.1 Input Data For Hydraulic Fracture Model .................................................58

4.1.1 Logs and Log-derived Properties...................................................58

4.1.2 Formation Zone Setup....................................................................62

4.1.3 Pumping Schedule .........................................................................66

4.2 Input Data For Microseismic Model..........................................................78

4.2.1 Microseismic Event Locations For Well D1 .................................79

4.3 GOHFER™ Pressure Matches ................................................................111