Geomechanical Review of Hydraulic Fracturing Technology

By: Julius Bankong Arop

Bachelor of Engineering in Civil and Environmental Engineering
University of Nigeria Nsukka, 2011
SUBMITTED TO THE DEPARTMENT OF CIVIL ENGINEERING IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF
MASTER OF ENGINEERING IN CIVIL AND ENVIRONMENTAL ENGINEERING AT THE
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
JUNE 2013
2013 Julius Bankong Arop. All Rights Reserved.
The author hereby grants MIT permission to reproduce and to distribute publicly paper and
electronic copies of this thesis document in whole or in part in any medium now known or
hereafter created.

Signature of
Author:___________________________________________________________________________
Department of Civil and Environmental Engineering
May 21, 2013
Certified
by:______________________________________________________________________________
John Germaine
Senior Research Associate of Civil and Environmental
Engineering
Accepted
by:______________________________________________________________________________
Heidi M. Nepf
Chair, Departmental Committee for Graduate Students

Geomechanical Review of Hydraulic Fracturing Technology

By: Julius Bankong Arop
Submitted to the Department of Civil and Environmental Engineering on May 21st, 2013 in Partial
Fulfillment of the Requirements for the Degree of Master of Engineering in Civil and Environmental
Engineering at the Massachusetts Institute of Technology
Abstract
Hydraulic fracturing as a method for recovering unconventional shale gas has been around for
several decades. Significant research and improvement in field methods have been documented in
literature on the subject. The heterogeneous nature of shale has made hydraulic fracturing design
to be unique for particular site conditions. Actual methods of carrying out fracturing operations and
design decisions are also different for various companies in the industry. Hence, there are no
standards for decisions in processes such as: formation testing, fracture modeling, choice of
fracturing fluid or propping agent selection. This has led to different interpretations of pressure
tests and proprietary fracture designs that have not been evaluated for adequacy against any
recognized scale.
The goal of this thesis is to do an appraisal of hydraulic fracturing in theory and practice. A review is
done of the early theoretical work upon which most of the current hydraulic fracturing literature is
based. Effort is also made to thoroughly cover the core aspects of fracture modeling and practical
operations with a view to shedding light on the strength and drawbacks of current methodologies.
The thesis focuses on the geo-mechanics of the process thus less emphasis is laid on post fracturing
operations. It is hoped that this will help establish the basis for a standard framework to guide
fracturing design. Finally, the ambiguity of nomenclature in oil and gas circles has led to
considerable confusion in conducting academic work. For this reason, effort was made in the thesis
to clearly define the various terminology.

Thesis Supervisor: John Germaine

Title: Senior Research Associate of Civil and Environmental Engineering
2

Acknowledgements
My special gratitude goes to the one who gave me life and divined that I would start this thesis and
complete it, God Almighty. My heartfelt thanks go to my parents, Surveyor Emmanuel Arop and
Mrs. Margaret Arop for their unflinching support and sound advice. I count myself blessed to have
such wonderful parents. Thank you Mum and Dad.
My deepest accolades go to my supervisor, Dr John Germaine for his patience, understanding,
motivation and profound wealth of experience which guided me in seeing this work to the finish. I
am grateful not just for your tutorship but for your friendship. I have learnt a great deal from you,
in more ways than you know. Thank you Dr G.
My siblings Mercy, Elizabeth, Emmanuel, Margaret and John for their cherished words of
encouragement and uplifting jokes.I love you all.
I want to also thank Mr Augustine Okoye for all his helpful counsel since I arrived in the United
States. I wish to thank Engineer Ernest Onyia for his invaluable recommendations for source
materials at the start of my thesis. My appreciation also goes to my friends: Obehi, Loko P, O.T,
Pam Pam, Matt Pires, Fidele, Awa, Deborah, Dimitris, Sar, Catherine and Emma Okafor for all their
kind words of encouragement.
Finally, I want to thank the resource people of onepetro.org. Your immense depository of
petroleum engineering literature has been of great help to me in writing this thesis.

TABLE OF CONTENTS
Chapter 1: Development of Unconventional Resources 17
1.1 Introduction. 17
1.2 The Hydraulic Fracturing Process.. 17
1.3 Well Development...21
1.3.1 Rotary Drilling. 21
1.3.2 Horizontal Drilling 24
1.3.3 Completions- Casing and Cement Design.. 26
Chapter 2: Formation Evaluation. 31
2.1 Formation Testing.. 31
2.1.1 Formation Integrity Tests (FIT).

31

2.1.2 Field Procedure for Formation Integrity Tests 32

2.2 Permeability, Fracture Conductivity, and Fracture Length 35
2.2.1 Obtaining Design Parameters from Pressure buildup tests: An introduction 35
2.2.2 Parameter Determination from Pressure tests: Method of Lee and Holditch (1979) 36
2.2.3 Parameter Determination from Pressure tests: Method of Barnum (1990) 41
2.2.4 Permeability from Productivity Index Test... 45
2.3 Well Condition Ratio. 46
2.4 Applicability of Wells for fracturing.. 46
2.4.1 From Buildup and Well Condition Ratio.. 46
2.4.2 General Criteria/Limitations on Well Selection for Fracturing Treatment. 48
2.5 Core Analysis. 50
2.5.1 Qualitative Visual Analysis.. 51
2.5.2 Routine, Quantitative Analysis.. 52
2.5.3 Special Core Analysis... 56
2.6 Logging Analysis.. 57
2.6.1 Electrical logs.. 58
2.6.2 Lithology logs59
4

2.6.3 Logging While Driling (LWD). 59

2.6.4 General Log Characteristics of Shale.. 60
Chapter 3: Fracture Geometry. 61
3.1 In-situ stress determination.. 62
3.1.1 Pump-in/Shut-in Tests. 64
3.1.2 Pump-in/Flowback Tests 67
3.1.3 Step-Rate Injection Tests (SRT).. 69
3.2 Mini-Fracture Tests and G-function Analysis..69
3.2.1 Standard Leak Off Tests (LOT). 70
3.2.2 Extended Leak Off Tests (XLOT). 73
3.2.3 Case Study Comparison of LOTs and XLOTs.. 75
3.2.4 G-function Derivative Analysis 78
3.3 2D Geometry Models.. 80
3.3.1 Howard and Fast Fracture Area/Extent 80
3.3.2 The PKN and KGD Geometries: Fracture Width. 84
3.3.3 Fractures in Horizontal Wells. 87
3.3.4 Fracture Geometry in Layered Formations: Fracture Height 88
3.3.5 Economics of 2D Models: Optimum Fracture Half-length. 88
3.4 Other Geometry Models.. 89
3.4.1 Pseudo 3D Models. 89
3.4.2 Fully 3D Models.. 92
3.5 Direct Mapping Techniques for Hydraulic Fracturing.. 94
3.5.1 Tiltmeters (Surface and Downhole) 94
3.5.2 Micro-seismic mapping 96
3.6 Net Pressure Matching and Calibration of Fracture Growth Models... 98
3.6.1 Net Pressure History Matching.100
3.6.2 Fracture Model Calibration.103

Chapter 4: Fracturing Materials 105

4.1 Fracturing Fluids. 105
4.2 Additives 106
4.2.1 Viscosifiers 107
4.2.2 Surfactants and Alcohols... 108
4.2.3 Fluid-Loss Additives. 109
4.2.4 Clay Stabilizers 111
4.2.5 Gel Breakers. 112
4.2.6 Bactericides/Biocides 112
4.2.7 pH Control. 113
4.3 Proppants. 113
4.3.1 Proppant Size and Concentration. 114
4.3.2 Proppant Shape. 116
4.3.3 Proppant Stress.... 117
4.3.4 Proppant Embedment Pressure 119
4.3.5 Fracture Capacity (conductivity).. 120
4.3.6 Proppant Selection: Method of Rixe et al (1963).. 123
4.4 Case Study Example of typical Hydraulic Fracturing Stimulation. 125
4.4.1 Introduction. 125
4.4.2 Preliminary Slickwater Fracture Design... 126
4.4.3 Starting Treatment 126
4.4.4 Proppant Selection and Placement 127
4.4.5 Finishing Treatment 127
Chapter 5: Conclusion/Recommendations.. 129
Tables 132
Figures. 150
Appendix

271

Bibliography. 283
6

LIST OF TABLES
Table 1.1: Types of Completions and conditions for selection (After Azar and Samuel , 2007)... 132
Table 1.2 : API designated classes of Portland cement (After Azar and Samuel , 2007). 135
Table 2.1: Core recovery types and application)... 136
(After Unalimser and Funk (1998), and Keelan,1972)
Table 2.2: Permeability of well samples, showing effect of confining pressures.. 137
( After Al-Multhana et al. 2008)
Table 3.1 : Plot types used for analysis, After Lee et al, 2004....

138

Table 3.2 : Classification of pressure tests performed at the casing shoe: PITs.. 139
(After Addis et al, 1998)
Table 3.3: Shut-in pressures, minimum and maximum stress gradients from ELOTs performed

140

on the North West Shelf, Australia (After Addis et al, 1998)

Table 3.4: Shut-in pressures, minimum and maximum stress gradients from ELOTs performed 140
on the Norwegian North Sea (After Addis et al, 1998)
Table 3.5: Oil reservoir Data for 3D Model Application (After Settari and Cleary, 1984). 141
Table 3.6: Pumping Data for 3D Model Application (After Settari and Cleary, 1984).... 142
Table 3.7: Fluid Treatment Data for 3D Model Application (After Settari and Cleary, 1984).... 143
Table 3.8: General net pressure matching parameters (After Weijers et al, 2005). 143
Table 4.1: Fracturing Fluids and General Conditions for Use (After Xiong et al, 1996).

144

Table 4.2: Suggested Fracture Treatment Types for Different Formations 145
(After Kennedy et al, 2012)
Table 4.3: Effect of different size proppant on a fracture geometry and fracture characteristic..146
(After Cohen et al, 2013)
Table 4.4 : Pumping schedule with several proppant sizes (After Cohen et al, 2013) ..147
Table 4.5: Embedment Pressure Values for various formations.

148

(After Howard and Fast, 1970

Table 4.6 : Preliminary slickwater fracture design (After Fontaine et al, 2008) 149

LIST OF FIGURES
Figure 1.1- Shale Distribution in the United States (After Arthur et al, 2009). 150
Figure 1.2 Process Flow Diagram for a Hydraulic Fracturing Treatment 151
(After Arthur et al, 2009)
Figure 1.3 Water storage frac tanks - After Arthur et al (2009) 152
Figure 1.4 Picture of frac head (After Travelin Terriers, 2013 www.travelinterriers.com)............ 153
Figure 1.5 Hydraulic fracturing site (After Hibbeler and Rae ,2005).

154

Figure 1.6 Double well casing to protect groundwater sources (After Marathon Oil Corp, 2012). 155
Figure 1.7 Horizontal Drilling in Shale Strata (After King et al ,2010) 155
Figure 1.8 Drilling Rig Components (After King et al, 2012) 156
Figure 1.9 Vertical and directional wells (After Azar and Samuel, 2007). 157
Figure 1.10 Different types of drill bits (After Devereux ,2012) . 157
Figure 1.11 Basic elements to drill a well (After Azar and Samuel, 2007) 158
Figure 1.12 Section of rotary drilling rig (After Azar and Samuel ,2007).. 159
Figure 1.13 Block and Tackle Hoisting system (After Azar and Samuel , 2007). 160
Figure 1.14 Fluid circulation system (After Azar and Samuel , 2007) 161
Figure 1.15 Rotary system of drilling rig (After Azar and Samuel ,2007)162
Figure 1.16 Floor of a drilling rig showing rotary table (After Devereux, 2012) 163
Figure 1.17 Typical blowout preventer (BOP) arrangement (After Azar and Samuel ,2007).. 164
Figure 1.18 Picture of BOP (After Devereux , 2012). 165
Figure 1.19 Blowout preventer positioned on a well (After Devereux, 2012).166
Figure 1.20 Horizontal drilling by jetting (After Devereux , 2012).... 167
Figure 1.21 Horizontal drilling by using a whipstock (After Devereux ,2012). 168
Figure 1.22 Horizontal drilling via steerable motor (After Devereux , 2012)169
Figure 1.23 Optimization of casing string along depth (After Azar and Samuel , 2007) 170
Figure 1.24 A conductor pipe (After Azar and Samuel , 2007) 171
Figure 2.1 Guide lines on PIT graph (After Postler et al ,1997). 172
Figure 2.2 Schematic of PIT equipment (After Postler et al ,1997). 172
Figure 2.3 - Description of dimensionless fracture conductivity (After Jones and Britt, 2009) 173

Figure 2.4 Three conventional test configurations for determination of vertical permeability.. 174
(After Ehlig- Economides et al, 2006)
Figure 2.5 Horner plot, pressure buildup test data from a South Texas gas well.. 175
(After Lee and Holditch,1979)
Figure 2.6 Millheim-Cinchowicz plot, pressure buildup test data from South Texas gas well. 176
( After Lee and Holditch, 1979)
Figure 2.7 Synthetic drawdown curves for modified M-C method, constant rate case... 176
(After Lee and Holditch, 1979)
Figure 2.8 Modified M-C plots, field buildup tests (After Lee and Holditch , 1979) 177
Figure 2.9 - Extrapolation of linear region, modified M-C method (After Lee and Holditch ,1979) 177
Figure 2.10 Typical partial perforation pressure buildup response 178
(After Barnum et al, 1990)
Figure 2.11 Pressure buildup showing effect of wellbore damage and after-production 179
(After Howard and Fast, 1970)
Figure 2.12 Bottomhole pressure versus buildup time (After Gladfelter et al, 1955) 179
Figure 2.13 Drawdown plot of dimensionless real gas potential versus log of dimensionless 180
time (After Vairogos et al, 1973)
Figure 2.14- Vertical permeability correlation chart (After Barnum et al, 1990) 181
Figure 2.15 Nomogram for end of first straight-line region (After Barnum et al, 1990) 182
Figure 2.16 Effect of fracture capacity on post-fracturing conductivity. 183
( After Howard and Fast ,1970)
Figure 2.17 Effect of in-situ stress contrast on fracture containment.. 184
(After Heydarabaradi et al, 2010)
Figure 2.18 Containment of a fracture opposite overburden (After Heydarabaradi et al, 2010).. 184
Figure 2.19 Well blowout

185

(After Lucasher Gusher http://www.sjgs.com/gushers.html#spindletop, 2013)

Figure 2.20 - Full closure core catcher (After Unalmiser and Funk, 1998) 186
Figure 2.21 Composite photograph of core from various depositional environments ... 187
(After Keelan,1982)

Figure 2.22 SEM photograph of Illite clay in pore space. (After Keelan , 1982). 188
Figure 2.23 NMR relaxation and distribution (After Unalmiser and Funk ,1998).. 188
Figure 2.24 CT scan 1 (After Al-Muthana et al ,2008)... 189
Figure 2.25 CT scan 2 (After Al-Multhana et al, 2008) 189
Figure 2.26 CT scan 3 (After Al-Multhana et al, 2008) 189
Figure 2.27 Effect of calibration on grain density (After Al-Multhana et al, 2008). 190
Figure 2.28 Effect of drying time on porosity and permeability (After Al-Multhana et al, 2008). 190
Figure 2.29 Schematic diagram of shear and compressional wave propagation 191
(After Timur A. Turk)
Figure 2.30 Sonic log versus depth (After Glover, 1986).. 191
Figure 2.31 Pre- and post-fracture temperature logs showing thermal conductivity effects 192
(After Jones and Britt, 2009)
Figure 2.32 A vertical well in the Woodford shale showing bedding planes and high angle. 193
fractures using an LWD electrical image (After Quinn et al, 2008)
Figure 2.33 LWD electrical images obtained from horizontal well in Barnett shale 194
(After Quinn et al, 2008)
Figure 2.34 Correlation between Total Organic content (TOC) log and Gamma Ray (GR). 195
(After Orlandi et al, 2011)
Figure 3.1 Normal leakoff plot of log Pressure Derivative versus time log -log plot.. 196
(After Barree et al, 2009)
Figure 3.2 Tip extension plot of Pressure derivative versus time (After Barree et al, 2009).. 197
Figure 3.3 Pressure and Flow data from two cycles of Mini-frac tests. 198
in an offshore location at the Timor sea (After Zoback, 2012)
Figure 3.4 Pressure-time records of two tests in shaly dolomite near Anna, Ohio. 199
(After Zoback and Haimson, 1982)
Figure 3.5 Results from tests in a well in South Carolina (After Zoback and Haimson, 1982)... 200
Figure 3.6 Double tangent method plot I (After Amadei and Stephansson, 1997) 201
Figure 3.7 Double tangent method plot II (After Amadei and Stephansson, 1997) 201
Figure 3.8 Double tangent method plot III (After Amadei and Stephansson, 1997) 202

10

Figure 3.9 Muskat Method/Exponential pressure decay 202

(After Amadei and Stephansson, 1997)
Figure 3.10- McLennan and Roegiers inflection method. 203
(After Amadei and Stephansson, 1997)
Figure 3.11 Horner plot (After Jones and Britt, 2007) .203
Figure 3.12 - Method of Turnbridge/Bilinear pressure decay rate method 204
(After Amadei and Stephansson, 1997)
Figure 3.13 - Pump-in flowback closure stress test (After Jones and Britt, 2009) 205
Figure 3.14 Step-rate test (SRT) result, showing additional derivative plot belo.. 206
(After Jones and Britt, 2009)
Figure 3.15 Step rate test (SRT) at different depths (After Zoback, 2012) 207
Figure 3.16 A schematic mini-frac test /XLOT showing pressure as a function of volume.. 208
or time (at constant flowrate) (After Zoback, 2010)
Figure 3.17 Extended pressurization cycles of XLOTs (After Lee et al, 2004) 209
Figure 3.18 A standard LOT pressure versus volume plot showing leak-off pressure. 210
(After Addis et al, 1998)
Figure 3.19 An idealized pressurization cycle from a PIT ( After Addis et al, 1998) 210
Figure 3.20 An Ideal example of a pressure record from a hydraulic fracturing stress test 211
showing the features sought in an ELOT (After Addis et al, 1998)
Figure 3.21 Pressure records for the six ELOTS performed in the North West Shelf.. 211
Australia (After Addis et al, 1998)
Figure 3.22 A comparison of leak-off pressures from standard LOTs (+) with minimum stress. 212
estimates from ELOTs for the North West Shelf of Australia (After Addis et al, 1998)
Figure 3.23 Leak-off pressures from standard LOTs (+) from vertical wells plotted with 213
minimum stress estimates from E(X)LOTS in the Oseberg Field, Norwegian sea.
(After Addis et al, 1998)
Figure 3.24 Leak-off pressures from standard LOTs (+) from vertical and inclined wells.. 213
plotted with minimum stress estimates from E(X)LOTS in the Oseberg Field,
Norwegian sea. (After Addis et al, 1998)

11

Figure 3.25 Pressure Dependent Leakoff G-function plot (After Barree et al, 2009).. 214
Figure 3.26 Tip extension G-function plot (After Barree et al, 2009).. 214
Figure 3.27 - Storage G-function plot (After Barree et al, 2009) 215
Figure 3.28 Howard and Fast assumed fracture geometry (After Holditch et al, 1987). 216
Figure 3.29 Perkins-Kern-Nordgrens assumed fracture geometry (After Holditch et al, 1987) 217
Figure 3.30 Vertical and Horizontal plane strain condition, with PKN geometry . 218
(showing wellbore) below (After Valko and Economides, 1995)
Figure 3.31 Geertsma-de Klerk-Daneshy (KGD) assumed fracture geometry.. 219
(After Holditch et al, 1987)
Figure 3.32 The KGD Geometry showing wellbore - (After Valko and Economides, 1995). 219
Figure 3.33- Radial fracture assumed geometry (After Holditch et al, 1987) 220
Figure 3.34 Transverse vertical fractures from a horizontal well.. 221
(After Valko and Economides, 1995)
Figure 3.35 Turning fractures in a horizontal well from longitudinal initiation 222
to the transverse direction (After Valko and Economides, 1995)
Figure 3.36 Vertical Fracture initiated from an arbitrarily oriented horizontal well at.. 222
an angle, a from the minimum horizontal stress (After Valko and Economides, 1995)
Figure 3.37 Vertical fracture profile through a three-layer formation with dissimilar 223
properties (After Valko and Economides, 1995)
Figure 3.38 Fracture height growth with associated width reduction due to adjoining. 223
layer of high stiffness (After Valko and Economides, 1995)
Figure 3.39 A T-shape fracture (proppant screenout) (After Valko and Economides, 1995)224
Figure 3.40 Cumulative gas volume produced with time for different fracture half-lengths. 225
(After Hareland and Rampersad, 1994)
Figure 3.41 Gas production decline with time for different fracture half-length.. 225
(After Hareland and Rampersad, 1994)
Figure. 3.42 Schematic of a Pseudo 3D Hydraulic Fracturing model (After Xiaowei Weng, 1992)226
Figure. 3.43 Schematic of the 2D flow in the fracture (After Xiowei Weng, 1992).. 226
Figure. 3.44 Approximation of the outer flow field at the element of interest by a local ... 227

12

radial flow (After Xiaowei Weng, 1992)

Figure. 3.45 Comparison of fracture shapes predicted by different models 227
(After Xiaowei Weng, 1992)
Figure. 3.46 Example of StimPlan Output (After Barree et al, 2009).. 228
Figure 3.47 Configuration of a fully 3D physical system (After Settari and Cleary, 1984). 228
Figure. 3.48 Fracture shapes for different degrees of containment 229
(After Settari and Cleary, 1984)
Figure. 3.49 Comparison of computed fracturing pressures for different fracture geometries 229
(After Settari and Cleary, 1984)
Figure 3.50 An example of Fracpro-PT Output (After Barree et al, 2009) 230
Figure 3.51 Fracture Diagnostic techniques (After Cipolla and Wright, 2002) 231
Figure. 3.52 Principle of tiltmeter fracturing mapping (After Cipolla and Wright, 2002) 232
Figure. 3.53 Surface tiltmeter in a site at Western Missouri (After Jones and Britt) 232
Figure 3.54 Surface deformation for hydraulic fractures of different orientations at a depth of.. 233
3000 ft (After Wright et al, 1998)
Figure 3.55 Picture of tiltmeter used to predict eruptions on Mount Loa, Hawaii.. 234
(After Wilkipedia ,2013)
Figure 3.56 Conceptual drawing of actual fracture dimensions and estimated fracture 235
dimensions using downhole tiltmeter fracture mapping (After Wright et al, 1999)
Figure 3.57 Theoretical downhole tiltmeter pattern of deformation . 236
(After Jones and Britt, 2009)
Figure 3.58 Principle of microseismic fracture mapping (After Cipolla and Wright, 2002) 237
Figure 3.59 Microseismic event location from Fisher (2005) (After Jones and Britt, 2009) 237
Figure 3.60 Multiple component geophone array with P and S wave arrivals ..238
(After Jones and Britt, 2009)
Figure 3.61 Example of a Hodogram..

239

(www.Crewes.org/Researchlinks/Explorerprogram/Holdogram/Holdogram.html)
Figure 3.62 Unconventional gas well history matching (After Britt et al, 2010)

240

Figure. 3.63 Nolte-Smith interpretation guide from Britt et al (1994) 241

13

(After Jones and Britt, 2009)

Figure 3.64 Width decoupling due to containment caused by layers (After Barree et al, 2009)242
Figure. 3.65 Composite layering and width decoupling Weijers et al, 2005.. 242
Figure. 3.66 Observed net pressure (black) and match with model net pressure (green). 243
(After Weijers et al, 2005)
Figure. 3.67 FracPRO PT model showing width profile and proppant concentration in fracture..244
Figure. 4.1 Composition of a Proppant Laden fracture Fluid ( After Arthur et al, 2009).. 245
Figure 4.2- Schematic diagram showing effect of adding fluid-loss additives 245
(After Hawsey et al , 1961)
Figure 4.3- Plot of fluid loss versus time ( After Hawsey et al ,1961) ..246
Figure 4.4- Fracture area and injection rate versus fluid loss coefficient. 247
(After Hawsey et al, 1961)
Figure 4.5- Correction factor for Spurt loss using nomograph ( After Hawsey et al, 1961) ..248
Figure 4.6 Animation of Hydraulic fracturing (After Marathon Oil Corp, 2013) 249
Figure 4.7 Cross-sectional photos of three proppants under stress in the conductivity cell. 249
(courtesy of Stim-Lab consortium) (After Palisch et al, 2010)
Figure 4.8 Effect of proppant settling velocities (After Bivins et al , 2005) .. 250
https://slb.com/~/media/Files/resources/oilfield_review/ors05/sum05/03_new_fibers.pdf
Figure 4.9 Production rate for different proppant sizes - After Cohen et al (2013) 251
Figure 4.10 Cumulative production for different proppant size (After Cohen et al , 2013). 251
Figure 4.11- Krumbein and Sloss chart of sphericity and roundness ... 252
(After Kullman (2011) South Dakota School of Mines)
Figure 4.12- Comparison of commercial proppants on the Krumbein and Sloss char.253
(After Kullman (2011) South Dakota School of Mines)
Figure 4.13 A schematic of pressure during shut-in and flow (After Sookprasong , 2010) 254
Figure 4.14- Embedment pressure test technique (After Howard and Fast ,1970). 254
Figure 4.15 Picture of mounted core assembly for fracture capacity test (After Rixe et al ,1963) 255
Figure 4.16 Diagram of mounted core assembly for fracture capacity test. 255
(After Howard and Fast, 1970)

14

Figure 4.17 Test apparatus for fracture flow capacity test (After Rixe et al , 1963) 256
Figure 4.18 Nitrogen flow system for measuring fracture flow capacity (After Rixe et al, 1963).. 257
Figure 4.19 Scanning Electron Microscope (SEM) photograph (404X) of 40/80 mesh.. 258
lightweight ceramic proppant fines after a wet, hot crush test at 10,000 psi.
(After Terracina et al , 2010)
Figure 4.20 Enveloping of proppant fines with resin coating as seen by a CAT scan of CRCS wet, 258
hot crush test at 10,000 psi. Grain to grain bonding with resin is also shown.
(After Terracina et al ,2010)
Figure 4.21 Proppant flowback from the fracture into the wellbore can occur with.259
uncoated proppant or procured RCS (After Terracina et al, 2010)
Figure 4.22 SEM photo (651X) of CRCS grain to grain bonding that eliminates 259
proppant flowback by forming a consolidated proppant pack in the fracture
(After Terracina et al ,2010)
Figure 4.23 Proppant embedment into the fracture face reduces fracture width and. 260
conductivity (After Terracina et al (2010)
Figure 4.24 SEM photograph (514X) of formation fines spalling (in red circle).. 260
due to grain embedment (After Terracina et al , 2010)
Figure 4.25 Approximate production increase due to fracturing (After Rixe et al ,1963). 261
Figure 4.26 Generalized selection chart for maximum fracture capacity obtainable 262
at a depth of 7000 ft (After Rixe et al,1963)
Figure 4.27 Propping agent selection curve for sand at well depth of 7000ft.. 263
( After Rixe et al, 1963)
Figure 4.28 Propping agent selection curve for rounded nutshells at well depth of 7000ft... 264
( After Rixe et al, 1963)
Figure 4.29 Propping agent selection curve for rounded nutshells at well depth of 7000ft. 265
( After Rixe et al, 1963)
Figure 4.30 Potter County, PA Marcellus Well log (After Fontaine et al, 2008).

266

Figure 4.31- Treatment initiation: The effect of acid upon wellbore entry .. 267
(After Fontaine et al ,2008)

15

Figure 4.32- The use and effect of proppant slugs ( After Fontaine et al ,2008)267
Figure 4.33- Illustration of near wellbore complexity (tortuosity) versus simplicity 268
(After Fontaine et al,2008)
Figure 4.34- Use of wellbore sweeps between proppant stages (After Fontaine et al ,2008) 268
Figure 4.35- The effect of a sweep (After Fontaine et al, 2008).. 269
Figure 4.36 - Effects of Proppant Concentration and Mesh size on Net Pressure. 269
(After Fontaine et al, 2008)
Figure 4.37- Finishing the treatment (After Fontaine et al, 2008) 270

16

CHAPTER ONE DEVELOPMENT OF UNCONVENTIONAL RESOURCES

1.1 Introduction
Hydraulic fracturing is a method of obtaining resources like natural gas and crude oil from
unconventional reservoirs such as shale. Shale is termed unconventional because its low
permeability causes difficulty in extracting resources by merely drilling into the formation.
Hydraulic fracturing is not a new technique, as the first fracturing test was done in 1947 and was
commercialized in 1950 (King, 2012). Another key technology, germane to shale gas development is
horizontal drilling, which started in the 1930s. The popularity of hydraulic fracturing as a means to
obtain natural gas has been heightened by the large distribution of shale gas deposits in the United
States. In Fig.1.1, the shale gas regions in the United States are shown. The Society of Petroleum
Engineers (SPE) have estimated that in the last 60 years, about 2.5 million hydraulic fracturing
operations have taken place worldwide; 1 million in the United States alone and tens of thousands
of horizontal wells have been drilled (King, 2010). In this chapter, an overview of hydraulic
fracturing is done, with a look at the well development process.
1.2 The Hydraulic Fracturing Process
The hydraulic fracturing treatment follows the actual drilling and completion of the well. It starts by
transportation of equipment to the site, then rigging them up. Rigging up means making sure all
the necessary iron connections are in place, between the frac head on the well, frac pumps,
manifold trailer and additives equipment (Arthur et al, 2009). Fig 1.2 shows a process flow chart of
the hydraulic fracturing process. Specialized equipment are required for stimulation treatment,
which include: storage tanks, chemical trucks, and a variety of pipes and fittings. Fracture tanks are
large trailer tanks, designed to hold several hundred barrels of fresh water which is used as base
fluid for water-based (slickwater) fracture treatments (see Fig 1.3). Additives are transported to the
17

site in flatbed trucks which contain pumps that enable the pumping of additives to blenders. Acid is
usually transported to the fracturing site by an acid transport truck, which can hold up to 5000
gallons of acid. If the fracturing sites are closeby, acid can also be transported from site to site via
backside pump. Proppants are held in large tanks called sand storage tanks, which feed proppants
(usually sand) to the blender through large conveyor belts. These storage tanks may hold as much
as 350,000 to 450,000 pounds of proppant (Arthur et al, 2009)
A blender retrieves fresh water from the frac tanks using suction pumps and blends the water with
the proppant in a hopper. Fluids and proppant are combined with additives at design
concentrations to form a slurry, which is pressurized and transferred to frac pumps. The high
pressure frac pumps, transfer the slurry via positive displacement pumps to a manifold trailer. The
manifold acts as a transfer station, and pumps the fluid through ground lines to the frac head (see
picture in Fig 1.4) (Arthur et al, 2009). A typical massive frac site is shown in Fig 1.2d. Notice the
large pit in the background were pumped out fracturing fluid is stored. Pits like these have raised
huge environmental concerns about groundwater contamination. Note also the large number of
fluid and proppant storage tanks, high rate blending equipment e.t.c. To set up a fracturing
operation is expensive even where logistics and availability are good (Hibbeler and Rae, 2005).
Initial drilling is the same as for a conventional reservoir. A borehole is drilled vertically, then a
casing is placed before cement and mud is pumped to place a barrier between the borehole and
adjacent formation (see Fig 1.5). Drilling of the well is now continued, to an adequate depth within
the producing reservoir, called the kick-off point then the well bore is deviated gradually until it
curves horizontally (see Fig 1.6) and drilled a distance of typically 1000 ft to more than 5000 ft
(Arthur et al, 2009).

18

Hydraulic fracturing is done in isolated intervals along the horizontal well (Fig 1.7) because it is
impossible to apply pressure along the entire length of the well bore due to distance constraints
(1000 to 5000ft). It is controlled by volume that an operator can pump into the hole at pressure.
These intervals are isolated using packers. Perforations are created in the wellbore within the
interval bounded by packers, using a perforating tool. In shale gas development, fracture treatment
is done in several stages. Initial stages could involve just pumping freshwater into the wellbore,
thereafter using an acid flush to clean cement and mud from the wellbore to ensure fluid flow is
not impeded during fracture treatment (Arthur et al, 2009). In some fracture treatments, acid is
pushed through the perforated interval to help breakdown the formation surrounding the
wellbore.
The fracturing fluid is pumped through the perforated intervals at high pressures in order to create
fractures in the surrounding formation (pay zone). Hard particles commonly known as proppants,
are added to the fracturing fluid and pumped into the formation after the fractures have been
created. The proppant size and concentration is increased in stages over the entire course of one
treatment. The propping agents hold open the newly created fractures, to facilitate hydrocarbon
recovery. The design of fracture treatment is a complex task, which involves analysis, planning,
experience and rigorous observation of different stages in the entire process. Cipolla and Wright
(2007), outline the following questions considered in fracturing operation:
1. Do fractures effectively cover the pay zone?
2. Are fractures confined to the pay zone?
3. Does the fracture grow into gas or water-bearing zone?
4. What is the optimum number of fracture treatment stages, and the best treatment size to
cover thick pay zones?

19

5. Is the final fracture conductivity sufficient for target production? What is the optimum
proppant?
These are a few of the many questions that engineers face in designing fracturing stimulation
treatments.
1.3 Well development
Well development is an integral part of the hydraulic fracturing process. Well development is
broadly divided into: the drilling stage and the completions stage. For a successful fracturing
operation it is important that drilling equipment are properly maintained and that their rated
capacity is not exceeded. A drilling rig is the most visible part of the drilling operation, however
what is important is the underground activity (King et al, 2012). In Fig 1.8, a drilling rig schematic is
shown, with its visible equipment. As reported by King et al (2012), the main considerations in
selection of drilling rigs are:
1. Noise, which can be minimized by using electric rigs ;
2. Dust: If air drilling is used, control of air and cuttings is required
3. Appearance and time on location (usually 2 to 5 weeks): Most rigs for unconventional well
drillings are from 50 ft to over 100 ft tall, which is visually undesirable and take more time
to set up. Lower profile rigs are preferred on shallower wells but the trade-off is that larger
rigs are faster in operation.
4. Water and mud storage: requiring determination of size of pits or steel tanks. Also, storage
considerations for chemicals that would be mixed with the mud.
5. Pressure control equipment: The equipment has to undergo regular servicing and
inspection.

20

Completions involve the final stages of the well development process, which include casing and
cementing design.
1.3.1 Rotary Drilling
Rotary drilling primary elements
The rotary drilling process for a vertical or directional hole (Fig 1.9) involve the following elements:
1. Application of a force downward on a drill bit (see picture in Fig 1.10)
2. Rotation of the drill bit
3. Circulation of fluid, known as drilling fluid (liquid, gas or gasified liquid), from the surface
through the tubular (drill string), and back to the surface through the annular space, which
is the area between drill string and borehole wall or casing (see Fig 1.11) (Azzar and Samuel,
2007).
Rotary Drilling Systems
According to Azzar and Samuel (2007), drilling for oil and natural gas requires two major
components: manpower and hardware systems. The hardware systems that make up a drilling rig
are:
1. A power generation system
2. A hoisting system
3. A drilling fluid and circulation system
4. A rotary system
5. Well blowout control systems
6. A drilling data acquisition system

21

In this section, we would look at the functions of these systems and their working components. A
section of a drill rig is shown in fig 1.12.
Hoisting system
The core function of the hoisting system is to hoist the drill and casing strings during drilling and
casing operations (Azzar and Samuel, 2007). The hoisting system is supported on a derrick or mast,
which is the most visible part of a drilling rig, standing around 140 ft above the drilling floor. The
derrick has a set of sheaves at the top over which steel rope can pass (Devereux, 2012). The main
components of the hoisting system are: the drawworks (drum and brake), the crown block, the
travelling block, the hook, drilling (wire) line and elevator as shown in the schematic Fig 1.13.
Drilling fluid and circulation system
The function of the fluid circulating system in rotary drilling is to ensure the movement of drilling
fluid from the surface to the hole bottom and back to the surface. The major components of this
system include: mud pumps/air compressors, high-pressure surface connections, drill string, drill
bit, return annulus, mud pits and mud treatment equipment (Azzar and Samuel, 2007). A schematic
fluid circulation system is shown in Fig 1.14.
Rotary system
This systems primary function is to achieve rotation of the drill bit. It is comprised of the following:
drill pipe, drill collar, rotary table, swivel, kelly bushing and drive, and kelly. In modern rigs, the
rotary table is replaced with a top drive motor to induce drill bit rotation, especially in offshore
drilling (see Fig 1.15 and Fig 1.16). The swivel supports the drill stem, enables rotation to take place
and conveys drilling fluid to the drill string. The kelly is the first part of the drill stem beneath the

22

swivel, it allows drilling ahead at a total depth equal to its length before a new section of drill pipe
is required to be added (Azzar and Samuel, 2007).
Well blowout control system
During drilling, the formation fluid can begin to flow into the wellbore, this is termed a kick. If this
continues, a blowout can occur which poses severe danger to the fracturing operation, drilling crew
and equipment. The well blowout control systems requirements are to safely permit:
1. Shutting in the well at the surface.
2.

Controlling the removal of formation fluids from the wellbore

3. Pumping high density mud into the hole

4. Tripping the drill pipe into or out of the hole
The basic components of the blowout preventer (BOP) stack are: annular preventer, ram
preventers, spools, internal preventers, casing head, flow and choke lines and fittings, kill lines and
connections, mud and gas handling facilities and accumulators (Azzar and Samuel, 2007). A
schematic arrangement is shown in Fig 1.17 and a pictorial arrangement in Fig 1.18. Typically, a BOP
stack would have at least one annular preventer and two ram preventers. Below the rams are pipes
that extend to the side, called side outlets which allow flow out or into the annulus during well
killing operations. One section of the side outlets connects to the standpipe manifold to direct flow
to the annulus, it is called the kill line. The opposite side is called the choke line, it controls the flow
in and out of the manifold. The moment a driller detects that a kick is in progress, one of the BOP
stack preventer units will be closed to seal the well annulus (Devereux, 2012).

23

Blowout prevention equipment is crucial to rig selection. The BOP stack and system must have a
pressure rating equal to or greater than the bursting pressure of casing string and wellhead (Azzar
and Samuel, 2007). A BOP in position at an offshore facility is shown in Fig 1.19.
Drilling data and acquisition system
This system consists of devices to monitor, analyze, display, record, and retrieve information about
drilling operations. The parameters of interest include: drilling rate, hook load, hole depth, pump
pressure, flow rate, torque, rotary speed, mud density, temperature, salinity and flow properties,
mud tank level, pump strokes, weight on drill bit, hoisting speed (Azzar and Samuel 2007).
The monitoring equipment are used to detect drilling problems such as lost circulation and well
kicks. Drilling rate charts show points of drilling breaks which are useful in giving an idea of changes
in rock lithology and formation pressures. A rapid increase in hook load may indicate that a lost
circulation zone has been reached. Excessive torque readings may indicate a high concentration of
drilled cuttings in the annulus or bit failure. A drastic increase in pit level may indicate the intrusion
of formation fluids into the wellbore, hence indicating a kick and the danger of blowout occurring.
Rotary speed, flow rate and mud properties have to be maintained to achieve optimum drilling
conditions (Azzar and Samuel, 2007).
1.3.2 Horizontal Drilling
Horizontal drilling simply means directing the drill bit to follow a horizontal path, oriented at
approximately 90 from the vertical, through the reservoir rock (Azzar and Samuel, 2007). Over the
years, hydraulic fracturing has been performed on vertical, deviated and horizontal wells. However
the coupling of horizontal wells and hydraulic fracturing have been proven to improve well
performance in oil and gas reservoirs (Britt et al, 2010). This is due to the fact that it enhances the

24

recovery of hydrocarbons and reduces the number of vertical wells to develop fields of interest.
Horizontal wells have found application in the Barnett shale, Marcellus shale and other shale plays
were fracturing operations have been conducted for several years.
To drill horizontally it is required to deviate the wellbore. Several methods exist for deviating the
wellbore. They involve developing a side force at the bit, with a magnitude and direction sufficient
to guide the bit to the pre-designed path (Devereux, 2012). These techniques include: jetting,
whipstock, steerable motors, rotary drilling assemblies and rotary steerables. The process of
deviating the wellbore is known as kicking off the well.
Jetting
This is the fastest and cheapest way to kick off the well. It involves the use of tricone drilling bits,
which have nozzles between the three drilling cones. There is one large nozzle and two smaller
nozzles. The Drilling fluid flows out of the larger nozzle with greater force than the others, hence a
pocket would be washed into the rock in that direction (see Fig 1.20). By aligning the bit in the
desired direction and increasing pump rate without rotating the bit, the well is deviated (Devereux,
2012). From the description, it can be seen that this technique would be more effective in
unconsolidated formations.
Whipstock
When the kick off point is reached during drilling, the drillstring and bit is removed from the hole
and a wedge, called a whipstock is placed into the hole. The drillstring is re-inserted into the hole.
Forced against the side of the whipstock, it begins to deviate as drilling continues (see Fig 1.21). The
drill string will bend to allow the bottomhole assembly (BHA) to go around the curved hole
(Devereux, 2012).

25

Steerable motors
This is a downhole motor, which has an adjustable bend in its lower portion (see Fig 1.22). Thus, the
entire drillstring can be rotated to the desired direction. Steerable motors are commonly used to
initially kick off the well from its vertical path, but after the deviation angle exceeds 60, it becomes
difficult to make the assembly slide i.e. drill without rotating the drill string. Steerable motors that
drill while still rotating the drill string exist, but are very costly (Devereux, 2012).
1.3.3 Completions Casing and Cement design
According to Azzar and Samuel (2007), the completion process begins when the drill bit first enters
the pay zone. This is a very important part of the drilling process as materials such as drilling mud
which may be adequate in other intervals may not be acceptable in the pay zone. In the initial
completion design, the selection of the pay zone is the first step. It is based on several engineering
considerations, which include:
1. Prospect development economies
2. Porosity and permeability requirements
3. Hydrocarbon type and saturation requirements
4. Recoverable hydrocarbon volumes
5. Pressure support
6. Reservoir stability
7. Availability of technology for cost-effective production of reserves
8. Ability to plug and abandon reservoir
9. Environmental factors and other risks

26

After considering the above factors, the drilling engineer decides on the appropriate completion
type. Completion types are broadly classified into: open hole completions and closed hole
completions.
In the open hole completions, drilling is carried out from the surface and terminated at the top of
the pay zone then running and cementing the casing. Thereafter, the pay zone is drilled with nondamaging drilling fluid and mud. In this completion type, the formation has contact with the
wellbore thus allowing production or injection along every section of the interval in contact (Azzar
and Samuel ,2007).
For the closed hole completions, a casing string is first run to prevent borehole collapse and to
isolate the productive formation. The size of the casing string is optimized along the drilling depth
(see Fig 1.23) (Azzar and Samuel, 2007), In table 1.1, the completion types generally used are
shown. Also shown are estimates of reservoir variables corresponding to the appropriate
completion type.
Casing design
There are several types of casings, depending on the required function. In deep wells, as many as 6
different casings can be used to perform specific functions at different stages of drilling and
completing the well (Devereux, 2012).
Conductor pipe
This is generally the first casing that is run in the well. It may be driven into the ground, using a pile
driver or cemented inside an already drilled hole (see fig 1.24). According to Devereux (2012), it is
used primarily to achieve the following:

27

1. As a closed circulation system, to allow drilling fluid return back to the surface as drilling
progresses. This is useful so that mud returned can be treated, drilling cuttings removed
and the mud re-used.
2. Prevents unconsolidated surface formation from being eroded away by the drilling fluid.
3. Sometimes used to support the weight of the wellhead and BOPs.
Surface casing
This is run through the conductor pipe and is set deep enough for the formation at the shoe to
withstand pressure from a kicking formation several depths below (Devereux, 2012). It is cemented
in place over the entire length of string. It has the following purposes:
1. Allows a BOP to be connected so that the well can be drilled deeper
2. Protect freshwater sources from contamination by the drilling fluids
3. Preventing loose or weak formation from falling into the wellbore
As put by Azar and Samuel (2007) From a safety standpoint it represents the second line of
defense, sealing the well and handling any high pressure flow
Intermediate casings
These connect the surface casing to the production casing. They are used primarily to increase the
pressure integrity of the well as it is deepened and to protect any directional work done. They are
usually cemented in place, before higher mud weights are used. Intermedite casings are sometimes
referred to as liners, when they are not hung from the surface but from some down hole point
(Azar and Samuel, 2007).

28

Production casing
These are the final casing run in the well. In essence, they are long-term pressure vessels because
they contain the tubing that serve as passageways for the hydrocarbon flow from the reservoir to
the surface. If the tubing leaks, the production casing must be able to withstand the resulting
pressure (Devereux, 2012)
The main design considerations for a casing string include: tension, burst pressure, collapse, driving
forces, temperature, combined axial and internal forces, corrosion and connections (Devereux,
2012).
Cementing
Cementing is the final aspect of well completions. It is a generally rushed operation because of the
haste to start producing the well. This sometimes leads to spending time and money trying to
correct errors that initial care would have avoided (Azar and Samuel, 2007). Cement outside the
casing is designed to meet various needs, which include:
1. Supporting the weight of the casing.
2. Preventing fluids from moving upwards inside the cement.
3. Preventing fluids from migrating between the cement/formation and casing/cement
interfaces.
4. Serves as corrosion protection for the casing.
5. Protects the casing against movement of formation.
6. Allows for the perforation of the production casing, without the inner cement breaking-off
due to the shock wave.

29

Most cements used in the oil and gas industry are a type of Portland cement. Portland cement
has the four main constituents: tricalcium silicate (C3Si), dicalcium silicate (C2Si), tetracalcium
aluminoferrite (C4AlF) and tricalcium aluminate (C3Al). Even when made from the same
components, not all cements will react in the same way when mixed with water due to
dissimilarities in cement grind, additives, water impurities and the manufacturing process (Azar
and Samuel, 2007). The American Petroleum Institute (API) has established standard
specifications for oil well cement. This designates 8 classes of cement, classified according to
their behavior in deeper, high temperature down hole conditions. These classes are shown in
table 1.2. The API class G cement has been found to be the most useful when its properties are
modified during mixing (Devereux, 2012). In carrying out cement design, the following factors
are considered: density, thickening time, compressive strength, temperature rating, rheology,
chemical additives.

30

CHAPTER 2 FORMATION EVALUATION

A formation consists of rock layers (strata) that have similar properties. A formation that contains
recoverable hydrocarbons is called a reservoir. The region of the reservoir that is accessible to the
well bore or fracture is usually called the payzone. Formation evaluation is important, especially in
low permeability reservoirs because of the presence of alternating layers with various properties. It
is necessary to define these properties, which include: thickness, fluid saturation, porosity, Youngs
Modulus, in-situ stress, permeability, formation conductivity etc (Holditch et al, 1987). This is done
by carrying out specialized tests and techniques that would be described in this chapter.
2.1 Formation Testing
2.1.1 Formation integrity Tests (FIT)
The formation integrity test is carried out to confirm the strength of formation and well casing shoe
by increasing the bottom hole pressure to a design pressure. There is a lot of confusion in industry
nomenclature, as formation integrity tests (FITs) are sometimes called Leak off tests (LOTs). The
LOTs, also known as pressure integrity tests (PITs) are used to determine the fracture gradient of a
formation (from stress estimates), as described in chapter three. However, FITs are conducted to
show that the formation below the casing shoe will not fail while drilling subsequent sections with a
higher bottom hole pressure. Simply put, the FIT is a pressure test applied to the formation directly
below a casing shoe. It is generally conducted soon after drilling resumes after an intermediate
casing string has been set. The purpose of the test is to determine the maximum pressures that
may be safely applied without the risk of formation breakdown. The results of the test are used to
design the mud program for the subsequent hole section and to set safe limits on casing shut-in
or choke pressures for well-control purposes Schlumberger (2012).

31

Another difference between PITs and FITs is that stress estimates are not obtained with data from
formation integrity tests (FIT) since fracture initiation does not actually occur. In the FIT, pressure is
applied to a pre-defined value and no leak off occurs. The FIT indicates that the maximum well bore
pressure did not exceed the least principal stress or was not sufficient to initiate a fracture of the
well bore wall in an open hole test (Zoback, 2012). In well planning and development, the FITs are
normally conducted before LOTs.
A straightforward way of calculating the pressure required for a formation integrity test (prior to
performing the test) is using the formula:
Pressure required for FIT (psi) = (Target FIT (ppg) Current mud weight (ppg)) x 0.052 x TVD of shoe
in ft

.(2.1)

Where, TVD = True Vertical Depth; and the Target FIT (ppg) is the equivalent mud weight of the
required FIT pressure.
It is worthy to note that no standard methodology exists in the industry for conducting FITs or LOTs.
In the next section, we would outline the field procedure of the formation integrity test, done by .
This would also, cover the build-up section of the test which is the formation integrity test.
2.1.2 Field Procedure for Formation Integrity Tests
Postler(1997) writes extensively on field procedures for LOTs. Similarities exist between the way
LOTs and FITs are conducted, hence some parts of this section would make reference Postlers
paper.

32

Reference Guidelines before the test (Postler, 1997)

A test graph is prepared before testing. The horizontal axis is labeled in bbl increments and the
vertical axis in 100-psi increments. Then guidelines are drawn on the graph for reference during the
test and to aid in interpretation. These guidelines are shown in fig. 2.1 (the fig. will be re-drawn to
illustrate the FIT).
a. Predicted FIT Pressure: shown in the figure as a horizontal line. This value is based on data
from offset wells and the local overburden or pore pressure gradients. This line acts as a
guide during the test; A plot beneath this line is most likely not leak-off and pumping
should continue.
b. Minimum Leak-Off Pressure: This horizontal line is equal to the predicted FIT pressure
equivalent mud weight (EMW) given as Target FIT* in eqn (2.1) minus 1/2ppg. The ppg
accounts for inaccuracies in the predicted FIT pressure; pressure, volume or mud weight
measurements; and mud gelation effects.
c. Maximum allowable Pressure: A horizontal line that indicates an upper limit based on
equipment limitation or lost circulation experience.
Test Procedure
Department of Infrastructure, Energy and Resources, Tasmania (2008) present a description of the
FIT field procedure. The method described below is termed the hesitation method of conducting
formation integrity tests. It involves pumping and waiting for the pressure to stabilize before
repeating the procedure until the maximum test procedure is achieved. During the test, it is
ensured that the maximum pressure does not exceed any of the following:
1. Actual Leak-off pressure.

33

2. The pressure specified in the drilling program (typically 80% of casing burst pressure)
3. The wellhead test pressure
4. The blow-out preventer, BOP test pressure
5. A maximum pressure gradient of 0.8 psi/ft at the casing shoe (i.e. Required FIT Pressure/
TVD of casing shoe).
A typical schematic diagram of equipment for the test is shown in Fig. 2.2.
The FIT procedure is as follows:
1. The formation of interest is drilled to approximately 10 ft
2. The drill string is used to pull the bit back in casing shoe.
3. It is then ensured the hole is filled and the annular BOP closed around the drill pipe.
4. Rig up the pump to the drill pipe. A pressure gauge of appropriate range (typically 0 to 1500
psi) is mounted at the top of the pump unit manifold.
5. Mud is pumped slowly until pressures begin to increase. Volume pumped will start from
this point.
6. 0.125 to 0.25 bbl of drilling fluid is pumped , then waiting is done for 2 minutes or more
depending on how much time is required for the pressure to stabilize.
7. The volume pumped is recorded, and the bleed back stabilized pressure.
8. Steps 6 and 7 are repeated and a plot of pressures versus cumulative mud volume is done.
9. The procedure is continued until either the final stabilized pressure deviates from the
predicted FIT pressure or the required maximum pressure is reached. The stabilized
pressure should be kept below the FIT pressure.
10. The well is kept closed in to be sure that a constant (stabilized) pressure has been attained.
11. The pressure is slowly released and volume recovered in tank is recorded.
34

2.2 Permeability, Fracture Conductivity and Fracture Length

The main aim of hydraulic fracturing is to initiate and sustain a stable fracture with high quality
conductivity to maximize well productivity and recovery. There are three core design parameters
that are germane to successful hydraulic fracturing treatment: permeability, fracture half-length
and fracture conductivity. The relationship between these variables is given by:

.(2.2)

Where Cr is the dimensionless fracture conductivity, k is the formation permeability, is the

fracture permeability, w is the fracture width and Xf is the fracture half-length. It is shown in Fig 2.3.

Equation 2.2 describes the relationship between the fractures ability to transport fluids to the
wellbore and the reservoirs ability to flow fluids to the fracture (Jones and Britt, 2009). This sums
up the essence of hydraulic fracturing. The fracture width, w is calculated from the fracture halflength and is given in detail in chapter 3.
Permeability is of critical importance in determining wells applicable for hydraulic fracturing. Since
the main reason fracturing is done, is to extract deposits of natural gas or crude oil that would not
flow naturally to the well bore. The permeability of the formation also affects the formation
breakdown pressure in hydraulically fractured wells. It is based on this effect that it can be
determined from the pressure buildup data of pressure tests. Experimental evidence supports the
fact that permeable rock has a lower breakdown pressure than impermeable rock under similar
conditions. Also, asides from showing a lower leak-off (breakdown) pressure, a Pressure Integrity
test (PIT) in a highly permeable formation shows a non-linear pressure build-up due to fluid losses
(Postler et al, 1997).

35

2.2.1 Obtaining the design parameters from Pressure build-up test: An Introduction
The three conventional test configurations for the pressure buildup test are shown in Fig 2.4. The
main feature of this test is that only a limited portion of the well is open to flow. It is sometimes
called the limited entry model test. The dimension, hw in the figure is the perforated (flow) interval
length.
Lee and Holditch (1979) carried out a theoretical investigation of pressure transient testing for
formation evaluation in low permeability gas reservoirs. Most current methods of analysis are
based on this investigation. In summary, Lee and Holditch reviewed the pseudo radial method of
Russell and Truitt (1964) and the linear flow method of Millheim and Cichowicz (1968), then
propose a modification to the method of Milloheim and Cichowicz. In this section, we would review
the work done by Lee and Holditch investigation before describing another method suggested by
Barnum et al, 1990.
The Lee and Hoolditch method doesnt consider the effects of partial perforation (limited entry),
but examines different fluid flow regimes for unbounded and bounded reservoirs. The Barnum
model considers the partial perforation model but is done for a single partial perforation
configuration. The Barnum proposition is also only applicable when the first straight line on the plot
is significant (high wellbore storage) which is mostly in low permeability reservoirs.
2.2.2 Parameter Determination from Pressure tests by the method of Lee and Holditch (1979)
This investigation was done to obtain permeability, half-length and fracture conductivity in low
permeability hydraulically fractured reservoirs. Data for this paper was based on pressure testing of
a South Texas gas well. Lee and Holditch considered pseudo-radial, linear flow, modified MillheimCichowicz method and Simulator history matching.

36

Pseudo-radial Flow
This was pioneered by Russell and Truitt (1964). Pseudo-radial flow is defined by Lee and Holditch
(1979) as one in which sufficient time occurs during buildup or drawdown such that the bottomhole
pressures varies linearly with flow time for drawdown tests or (th + t)/t in a semi-log plot
(Horner plot) for buildup tests. In an unbounded reservoir(infinitely acting) reservoir, the analysis
depends on the skin factor, given by Van Everdingen and Hurst to be:
S = 1.151 [ (P1 Pwf)/m log (k/ rw2) + 3.23]

.(2.3)

Where, rw is the wellbore radius; is the porosity; is the viscosity; Ct is the total fluid

compressibility (psi-1), measured at initial pressure; Pwf is the flowing well bore pressure and the

value of P1 is obtained from the straight line portion of the pressure buildup curve 1 hour after
shutting in, k is the permeability.
Fracture half-length, Xf (ft) for infinitely conductive vertical fractures is given by:
Xf = 2rwe-s

..(2.4)

Taking the log of both equations and equating to eliminate s, we obtain

Log Xf = 1/2[log(/c rw2 ) + (Pwf - P1)/m -2.63]

(2.5)

Thus, from the pressure buildup test, a Horner plot can be made and the slope, m determined (see
Fig 2.5). The formation permeability can thereafter be obtained from the following:
K = 162.6qB/mh

..(2.6)

37

q is the gas flow rate(Mscf i.e. 103 standard cubic ft), B is the gas formation volume factor evaluated
at the initial pressure, (bbl/ Mscf) and the half-length can further be determined from equation 2.4.
Lee and Holditch outlined several practical problems with the Pseudo-radial flow method. The most
important of which was that this method is not very useful in low permeability reservoirs because
the time required to obtain the straight line where the slope is related to formation permeability
can be too long, taking months or years. Other problems are that it assumes infinite fracture
conductivity which isnt always reasonable and finally, It may be difficult to obtain the proper slope
because of boundary effects that become appreciable as a result of delay in pressure transient.
For moderate to high permeability reservoirs, the pseudo radial method is extremely useful. Pratt
(1961) showed that for such reservoirs, pseudo radial flow can be used to model radial flow
behavior and vice versa (Jones and Britt, 2009)
Linear Flow
Consideration of linear flow was started by Millheim and Cinchowicz, who showed that when the
principal flow into a fracture is linear i.e. at the start of the test (earliest times), the pressure-time
characteristics for a pressure buildup test is given by:
Pi Pws = m (( + )

..(2.7)

Pws Pwf

..(2.8)

When, tp >> ,

Where Pi is the initial reservoir pressure, psi; Pwf is the flowing bottomhole pressure, psi and Pws
is the shut-in bottomhole pressure in psi, is producing time in hr.
38

A plot of bottomhole pressure versus square root of time would result in a straight line with slope,
m related to the fracture half-length and formation permeability as follows:
qB
hm

Xf = 4.064

2
)

(2.9)

Figure 2.6 shows an example plot of this method, with data from a pressure buildup test.
Limiting assumptions of this method include:
1. An independent estimate of formation permeability must be made in order to obtain Xf. This
value may be obtained from prior pressure test data from offset wells.
2. An impractically high value of fracture conductivity is assumed, so that fluid flow per unit area of
the fracture at the well bore is the same as that at the tip (uniform flux).
3. The linear flow mechanism can only suffice for earliest-time data, without distortion caused by
wellbore storage, which is not always the case.
Modified Millheim Cinchowicz method
Lee and Holditch propose a modification of the Millheim Cichowicz (M-C) method for finite
conductivity vertical fractures. For drawdown tests, this is done by plotting the dimensionless
pressure versus the square root of dimensionless time, as shown in Fig. 2.7. The data used for the
figure was published by Cinco et al (1978). The dimensionless pressure, PD is given by:

PD =

()
141.2

(2.10)

and dimensionless time, tD is given by:

tD =

0.00024
2

.(2.11)
39

From the fig. 2.7, it can be observed that at low values of Cr, the plot of PD vs tD is non-linear at

the earliest times of pressure buildup. Larger values of Cr however, have more linear plots at their
earliest times. Generally, at tD < 0.15, transient flow affects pressure-time behavior, at 0.15 <tD

<0.34 true linear portion of the plots are seen and at tD > 0.34, pseudo radial flow is established
(Lee and Holditch, 1979).

For both build-up and drawdown tests, the modified M-C method is outlined as follows:
1. Using an independent estimate of k, a plot of PD vs is done (see fig. 2.8).

2. The linear portion of that plot (fig. 2.7 for drawdown and fig 2.8 for buildup) is extrapolated
as a straight line to = 0 (see Fig. 2.9) and the value of the intercept is recorded.

3. The slope and intercept values of buildup/drawdown data is then compared with standard
charts containing Cr curves, in order to obtain an estimate of the Cr.
4. For a selected value of PD, the matching values of and can be obtained and
substituted in the following equation to obtain the half-length,
0.0002637

= (

)1/2 [ ]

..(2.12)

The modified M-C method proposed by Lee and Holditch still suffers the disadvantage of requiring
a first estimate of the permeability, k to obtain the fracture length and conductivity. However, it
provides better estimates of fracture half-length and conductivity from early-time pressure buildup
data.

40

2.2.3 Parameter Determination from Pressure Buildup tests by the method of Barnum et al (1990)
Barnum et al, 1990 carried out work on analysis of pressure buildup tests laying emphasis on the
configuration type, were the perforation interval was in the middle of the formation (see Fig 2.4). It
was based on plots of bottomhole pressure against the log of time. This results in a three-region
curve as shown in case 3 of fig 2.10. The first region is a straight line with a slope inversely
proportional to the permeability-thickness of the perforated interval. The second region is a
concave upward curve representing the transition between the first and third time region. The
third/final region is a typical straight line chart. In the various cases of figure 2.10, p* and pi can be
seen. P* is the extrapolation of the first straight line to infinite shut-in time. The dimensionless
pressure group, (p* - pi)/m can be correlated with the vertical permeability and would be shown
later. Where, m is the slope of the first straight line (Barnum et al, 1990). In the partial perforation
model, there is a near-wellbore restriction to flow due to wellbore damage. Wellbore damage is
caused by the infiltration of mud, cement, filtrate etc which alters the pressure buildup curve at an
early shut-in time, as shown in fig 2.11. The result is an extra drop in the pressure distribution at
the wellbore, in what is referred to as the skin effect (Howard and Fast, 1970). Saidowski (1979)
showed that the total or apparent skin factor, St based on the Horner analysis of the final straight
line, could be represented as follows:
St = (h/hw)S +Sp

.(2.13)

Where, S= actual skin factor due to apparent damage caused by wellbore restriction,
Sp = pseudo skin factor resulting from partial perforation,
h = total formation thickness, and
hw = perforated interval length.

41

The methods of calculating actual skin friction from empirical relations, exist in literature. For
instance, Van Everdingen and Hurst give an equation for S as given already in equation 2.3:
S = 1.151 [ (p1 pwf)/m log (k/c rw2) + 3.23]
Where, rw is the wellbore radius, pwf is the flowing well bore pressure, measured just before
shutting in, and the value of p1 is obtained from the straight line portion of the pressure buildup
curve 1 hour after shutting in. If the buildup curve is not straight after one hour, the straight line
must be extrapolated backwards, as shown in Fig 2.12;

k is an independent estimate of

permeability (Howard and Fast, 1970).

The actual skin friction, S is constant and can range from +1 to -3. Other methods of obtaining S
based on the dimensionless real gas potential obtained from pressure drawdown curves are shown
in Fig 2.13 (Vairogos and Rhoades, 1973). The actual skin friction is used to simulate wellbore
damage (stimulation).
The pressure drop in the skin can be obtained from the following:
Pskin = m x 0.97s

..(2.14)

The efficiency of the completion can be obtained by comparing the actual productivity index, J and
the ideal (no skin). The ratio of these quantities is the flow efficiency, given by:

Flow Efficiency

. ..............(2.15)

The pseudo skin factor, Sp is a function of the horizontal to vertical permeability ratio and can be
estimated from the following equation:
Sp = (h/hw -1) [In(h/rw) (kh/kv) -2]

(2.16)
42

Authors like Streltsova-Adams and McKinley (1981) developed techniques to obtain the vertical
permeability based on the start time of the second straight line region. Others like Bilhartz and
Ramey (1977) suggested methods based on the end time of the first line and the beginning time of
the second line. In 1986, Yeh and Reynolds developed a type curve fitting technique for vertical
permeability determination. The 1990 method proposed by Barnum et al would be discussed here.
It is a method proposed when the first straight line region exists for significant time periods. This
case is obtained in formations with adequately low permeability (such as shale), such that the flow
from the entire interval is significantly delayed. There are three pressure buildup patterns that
exhibit a first straight line characteristic of the perforated interval, as seen in Fig 2.10. Under
constant flow conditions: In Case 1, the buildup response would be only the first straight line, this
occurs when the vertical permeability is very low or when the flow time is short. Case 2 occurs
when the vertical permeability is higher, hence longer flow times. In Case 3, for much higher
permeability and even longer (relative) flow times, the three regions will occur.
It is worthy to note that the permeability determination of Barnum et al, 1990 is based on the
Hantush partial perforation model which has the following assumptions:
1. An infinite reservoir of constant thickness with impermeable upper and lower boundaries.
2. Single phase flow with constant fluid properties.
3. Homogenous rock properties, with constant horizontal and vertical permeabilities
throughout the formation of interest.
The dimensionless pressure group, (p* - pi)/m is correlated with the vertical permeability, flow time
and thickness of the perforated interval, as given in fig 2.14 for centered completions of 10,20, 40,
50, 67 and 80 percent (percentage of the total interval that is perforated). A plot of the
dimensionless group versus dimensionless time, tDV, given as follows:
43

tDV =

0.0002637
h2

..(2.17)

where tp is the producing time in hours, Ct is the total compressibility, psi-1 is the viscosity in cp.

is the porosity. This is very similar to the dimensionless time given in equation 2.11, the difference
is the replacement of fracture half-length, Xf with perforation interval length, hw.
Analysis Procedure
1. Construct a Horner plot of the pressure buildup data (as shown in fig. 2.11)
2. Compare the resulting plot with the buildup patterns of figure 2.11 in order to locate
straight line regions.
3. Determine if a particular straight line is the first or final region by applying the following
two checks:
-

Compare the total skin factor, St based on this line (is calculated from eqn 2.3) with the
pseudo skin factor, Sp estimated from equation 2.4. If St Sp (i.e. negative S), then the
line is most likely the first straight line region. If St Sp (i.e. positive S), the line is
probably the final straight line region.

Compare p* from this line with pi. If p*>pi, then the first straight line region is most
likely present, but if p* = pi, the line is probably the final straight line region.

4. After the checks in step 3, if a straight line region is recognized, and a portion of the
transition region is present then perform an additional check on the validity of this
selection with the nomogram in Fig. 2.15. If the end of the line is not apparent, the latest
time (tp + t)/t, should be the end of the straight line indicated by the end of the
nomogram.

44

5. Calculate (p* -pi)/m for this first straight line using fig 2.14 with the corresponding
completion percentage curve, to estimate the dimensionless time, tDV. The vertical
permeability can then be obtained from equation 2.17.
When the first straight line does not exist, other techniques by the previously mentioned authors:
BilHartz and Ramey; Streltsova-Adams and McKinley should be used. A method of permeability
determination based on the consideration of time for different flow regimes (early radial,
hemispherical spherical, infinitely acting radial), proposed by Ehlig-Economides et al, 2006 is also
recommended.
2.2.4 Permeability from Productitvity Index test
The productivity index test, is basically the flowback test. In this test, the pressure after shut-in is
released in a controlled manner and monitored. The flowback test is carried out after the pressure
buildup test. Details of the flowback test are in chapter 3. The productivity index permeability is
obtained from the following equation:

KPI = [ ( )]/0.007073h(Pws Pwf) ]

..(2.18)

Where, rw is the radius of the wellbore(ft), re is the external radius(ft), h is the formation thickness,
Pws (psi) is the static reservoir pressure obtained from the pressure buildup data. It is given by the
following, for infinite homogeneous reservoirs:

Pws = Pi 162.6

log

..(2.19)

45

2.3 Well Condition Ratio

The well condition ratio is an important parameter in determining the applicability of wells for
fracturing. It is the ratio of the permeability from a productivity index test to the buildup test. If a
reduction in permeability exists near the well, the permeability from a productivity index test will
be lower than that measured from a pressure buildup test. Thus if a well condition ratio, CR =
KPI/KBU < 1 then there is a permeability reduction but if >1 then there is an increase in effective
permeability. An increase in effective permeability could result from an earlier fracture treatment
or an acid stimulation (Howard and Fast, 1970).
2.4 Applicability of Wells for Fracturing
The choice of wells as candidates for fracturing is determined from considering various factors. This
section first considers the effect of well condition ratio; buildup data interpretation and formation
fluid carrying capacity.
2.4.1 From Buildup and Well Condition Ratio
Low permeability Effect on fracturing response
A high reservoir pressure with no significant near bore permeability reduction is no guarantee that
a well would produce at commercially acceptable rates. This is because the well may still be too
tight to produce at economical rates without fracturing stimulation. Data by Wilsey and Bearden
(1953) show that if the slope of the pressure buildup curve is greater than about 50
psi/cycle/Bbl/day (i.e pressure vs production graph) producing rate, when plotted on a semi-log
graph, and if considerable reservoir pressure exists, then that well would respond favorably to
reservoir fracturing treatment (Howard and Fast, 1970) .

46

Determination of State of Pressure Depletion

The static reservoir pressure can be obtained from the pressure buildup data. If this pressure is low,
then there is no need to carry out stimulation since most of the primary reserves have been already
been recovered from the formation. Fracture treatment on low pressure reservoirs would in a short
period lead to pressure decline, thus making it not economically viable (Howard and Fast, 1970).
Possible Increase in productivity from fracturing using condition ratio
The possible increase in productivity of a well can be determined by comparing the well condition
ratio obtained before and after stimulation using identical treatment. If a condition ratio is changed
from 0.1 to 1.0 after treatment then a ten times increase in stabilized production can be expected
from the well after fracturing treatment. A higher condition ratio means less reduction in nearbore
permeability which improves well productivity. Condition ratios obtained using fracturing fluids like
surfactants can at maximum improve the condition ratio to 1.0. In order to obtain a condition ratio
>1.0 then the rock has to be subjected to a physical change for instance by using acid treatment.
The lower the natural formation permeability, the greater the condition ratio that could be
achieved from treatment. Moderate permeability formations after any degree of stimulation could
only have condition ratios 2.0. While low permeability formations can have condition ratios as
high as 5.0 (Howard and Fast, 1970).
Effect of Formation Fluid-carrying Capacity
The formation fluid-carrying capacity (formation thickness x formation permeability) significantly
affects formation response to different sizes of fracturing treatments. Fig 2.16 shows a formations
response to treatment with two fractures of different fluid-carrying capacities (200md-ft and
5000md-ft). From these plots, the higher the formation capacity, the higher the capacity of

47

fractures (fracture permeability x fracture height) needed to ensure adequate fracturing treatment.
Refer to Fig 4.25, to see a plot of the fracture capacity normalized to the formation capacity against
the stabilized productivity ratio. The fracture penetration is represented as a percentage of the
drainage radius (i.e radius of fracture areal extent). As the ratio of fracture to formation capacity
increases, deeper penetrating fractures would lead to an increase in stabilized well productivity
(Howard and Fast, 1970).
2.4.2 General Criteria/Limitations on Well Selection for fracturing treatment
Proximity to Water and Gas Contacts
The impact of hydraulic fracturing on natural water bearing zones is a huge debate in the industry.
There are many accounts of the deleterious impact of fracturing on drinking water, told by land
owners and residents of lands adjacent to fracturing sites. Although there is little scientific
publications corroborating or refuting these claims, companies prefer to reject a potential
fracturing candidate entirely or reduce the size of treatment due to proximity of water contacts.
Systems mitigating the effects of water zone penetration exist using relative permeability modifiers
in the fracturing fluid or employing specialized proppant systems (dos Santos et al, 2009). However,
the use of these techniques almost inevitably requires the production/use of large amounts of
water during fracturing which can negate oil/gas production (Martin A.N et al, 2010). During oil
production the proximity of a gas contact influences suitability of a well for fracturing treatment, as
fractures could penetrate upward into a gas contact in the same way.
Containment within Payzone
It is important that fractures are designed considering the in-situ stress contrast, whether the
overlying layers (outside pay zone) would be sufficient barriers to fracture propagation or not. This

48

is to ensure that the fracture doesnt cross into water bearing strata (as enunciated in the first
section) and also to make for economical fracture treatment. This is shown in fig 2.17. In essence,
the stress gradient has to be considered when fracturing since the fracture direction would depend
on the in-situ stress i.e. consideration of the overburden stress is important to know the extent of
fracture propagation. See Fig. 2.18, in this diagram the effective stress gradient is assumed to vary
linearly downwards. (Heydarabadi et al., 2010).
Condition of Well Equipment
Due to the significantly high pressures expected during fracturing, it is necessary that the well
equipment should be able to adequately withstand such pressures. Well bore tubulars (i.e. casing
and tubing), gas lift mandrels and valves, sliding side doors, subsurface safety valves and flow
control devices. Packers have to be tested to ensure that fracturing pressure doesnt cause blowout
(see Fig 2.19). Hence, if the equipment do not meet the minimum requirements for the design
hydraulic fracture, stimulation cannot proceed (Martin A.N et al, 2010).
Skin factor considerations
A high positive skin factor justifies fracturing treatment since it indicates that the formation
permeability is higher than the near wellbore permeability. A negative skin factor however doesnt
eliminate the need for fracturing altogether. It means the formation permeability is less than
permeability surrounding the wellbore, and could indicate the presence of natural fractures in the
formation. If an acid treatment does not improve productivity, then hydraulic fracturing can be
considered (Heydarabadi et al., 2010).

49

Production History of Offset wells

A well with significantly lower production rate than offset wells in a field would most probably
benefit from fracturing treatment. Since its low production rate might be caused by reduction in
effective permeability around the wellbore (Howard and Fast, 1970).
2.5 Core Analysis
Core analysis simply means obtaining actual samples of rock and testing to know the physical and
chemical properties. It is imperative for a comprehensive formation evaluation. Coring analysis,
provides the only direct method of measuring reservoir petrophysical properties. In spite of this,
variation in data quality, sensitivity of results to different test methods, and the reluctance of
companies to share expertise and experience have caused poor data quality and errors in
interpretation of core analysis (McPhee, 2012). Best field and laboratory practices have to be
followed to obtain high quality results from core analysis.
Because most formations consist of layered strata, it is important that core obtained be cut for
different layers. In productive (pay) zones, the interest is normally in those properties that
determine amount of oil and gas present, and reservoir production rates. Such properties include:
permeability, porosity and water saturation. For the non-productive layers interest is more in the
formation geomechanical properties and stress distribution in order to calculate fracture geometry.
These parameters include: poissons ratio, Youngs modulus and fracture toughness. The method of
core recovery chosen depends on the specific rock type. Core recovery types can be shown in the
table 2.1. A full closure catcher used to obtain whole cores is shown in Fig. 2.20.
Coring Analysis techniques can be divided into three broad groups: 1. Qualitative, Visual Analysis; 2.
Routine, Quantitative Analysis; 3. Special Core Analysis (Holditch et al, 1987).

50

2.5.1 Qualitative, Visual Analysis

This involves all the techniques for visual description of the core. It involves: core photography,
scanning electron microscopy (SEM), X-ray diffraction, thin section analysis. Other core imaging
techniques include: nuclear magnetic resonance (NMR), computerized tomography (CT), micro-CT,
X-ray CT, acoustic and electrical resistivity.
Core photography
Core photographs are best taken under natural and ultraviolet (UV) light. Oil fluorescence is
revealed under UV light and hence the location of hydrocarbons. Multiple-footage photographs as
seen in Fig 2.21 are useful for the observation of changes in core bedding that indicates differences
in depositional environment, permeability, etc. (Keelan, 1982).
Scanning Electron Microscopy (SEM)
This enables electron photographs to be taken of pore space and magnified up to 40,000 times.
Samples are coated with an ultra-thin layer of electrically conductive materials and then
bombarded with electrons, leading to a secondary electron emission that yields a visible image (see
Fig. 2.22). The two most important uses of SEM are to recognize the clay type, and to observe
microporosity in clay linings and carbonates (Keelan, 1982).
X-ray diffraction
This is a method widely used in the identification of minerals, based on their unique diffraction
response to X-rays. It finds application in identifying clay minerals during coring analysis. However,
some materials such as amorphous chert cannot be identified by X-ray diffraction techniques
(Keelan, 1982).

51

NMR
This method has been in use as a petrophysical core analysis method since 1953. Its use has been
extended from prediction of pore size and wettability to estimation of permeability, water
saturation and residual oil saturation. It makes use of the radio frequency response of protons in a
magnetic resonance field to determine spin-lattice relaxation time, t1 and spin-spin relaxation time,
t2 (see Fig 2.23). These relaxation signals are linked to fluids within the pore network and correlated
with standard pore-size measurements, such as mercury injection (Unalmiser and Funk, 1998). The
results are processed further to obtain the aforementioned core parameters.
CT scans
These are useful in obtaining digital image of core samples, and also serve as saturation monitoring
tools. With accurate calibration, CT axial slice and longitudinal scout images provide digital data
that can be used to obtain density, fracture orientation, porosity, pore volume connectivity, and
net-to gross ratios (Unalmiser and Funk, 1998).
Fig 2.24 to Fig 2.26 show CT scans of three different samples obtained from a single well. Figure
2.24 displays the CT scan image of a non-uniform sample, which is not visible from outside using
the core photograph alone. The sample shown in Fig 2.25 has large vugs (cavities). The vugs can be
seen on the surface (dark spots) and internally with the CT scan. The sample in fig 2.26 is uniform
and an ideal candidate for core analysis studies. Both non-uniform and vuggy samples are not good
for further core analysis studies (Al-Multhana, 2008).
2.5.2 Routine, Quantitative Analysis
These are done to obtain first-order estimates of porosity, permeability, fluid saturation and
lithology. These variables are useful for determining oil and gas in place (OIP) measured in tcf (tons
52

per cubic ft). It is the value of oil/gas per volume of rock, as shown in the formula below (McPhee,
2012):

OIP = GRV [ ](1 )

.(2.20)

Where GRV is the gross rock volume and G is the gross factor in the net to gross ratio (N/G). B is the
formation volume factor obtained from PVT experiments. N is the net, is the porosity and Sw is

the water saturation obtained from routine, quantitative analysis and logging.

Routine, quantitative analysis is normally done 24 to 48 hours after a core is cut, because results
may immediately decide on whether a well would be completed or abandoned (Holditch, 1987).
Permeability
The standard method of obtaining permeability in routine core analysis is by allowing dry gas,
usually nitrogen, helium or air to flow through the samples. It has the following advantages over
using liquid permeability: reduced fluid-rock interaction, easier to execute, faster and less
expensive. In liquid producing reservoirs however, the validity of the gas permeability method is
being questioned (Unalmiser and Funk, 1998) Another shortcoming of using dry gas to obtain
permeability is that it has to be corrected for the Klinkeberg effect, also known as gas slippage. This
effect is due to variation in permeability measurements with the type of gas used and the mean
existing pressures in the core when measurement was done. It is more evident in samples from low
permeability formations (Keelan, 1972). Modern permeability testing include corrections for both
the Klinkeberg effect and the Forcheimer inertial coefficient (Unalmiser and Funk, 1998)
Equations for obtaining linear and radial permeability measurements and details of equipment used
are available in the American Petroleum Institute (API) Recommended Practice (RP) 40. However,

53

generally rock permeability in the laboratory is given by the following formula, based on Darcys
law:

K=

245

.................(2.21)

Where K is the absolute permeability (md), Q is the flow rate (cc/min), L is the sample length
(inches), A is the cross-sectional area (cm2), is fluid viscosity (cp), and is the differential

pressure across the sample (psi).

Several conditions can affect permeability measurements, they include: drying time, confining
pressure and core fluid/relative permeability effect (caused by the presence of more than one fluid
in pore spaces). Table 2.2 shows the effect of confining pressure on permeability measurements of
a well sample.
From the table, it is seen that the permeability values increased by about 70 to 80% when the
confining pressure was less than 200 psig. Specifications for testing standards recommend that
confining pressure should be greater than 200 psig (Al-Multhana et al, 2008). It is important to note
though, that natural overburden pressures can be as high as 6000 psi and thus, reservoir
permeability would invariably be different from those obtained from typical laboratory tests.
Keelan (1972) reports that Shale cores, unconsolidated, fractured samples and low permeability
(<0.1md) formations show high sensitivity to overburden pressures. He further states that,
Permeability reduction increases with overburden and reduction values of as little as 7 to as great
as 100 percent of initial values have been reported for overburden pressures up to 5000 psi

54

Porosity
Porosity is the ratio of void volume to total (bulk) volume. It is obtained by measurement of either
two of the three variables: pore volume (PV), bulk volume(BV) and grain volume (GV). Since,
Porosity,

and BV = PV + GV

....(2.22)

It is important that standard calibration of temperature and barometric pressure is done when
measuring grain density for GV determination. A dolomite sample (shown in Fig 2.27) had a grain
density that reduced by about 0.005g/cc after calibration (Al Multhana et al., 2008). Porosity
measurement like permeability is also sensitive to drying time as shown in fig 2.28. The type of
porosity test to be carried out depends on the formation being sampled, for instance in vuggy
formations special procedures are required.
Saturation
Measurement of residual fluid saturation was originally done by: 1. Using high powered vaccum
distillation to recover oil and water; 2. Distillation extraction, which divides the extraction process
into two parts. Firstly water is distilled, then oil is extracted using suitable solvents (Keelan, 1972).
Currently, fluid tracer studies, displaced-miscible fluid analyses (reducing damage to clays) and
improved geochemical techniques are used to obtain saturation (Unalmiser and Funk, 1998). The
calculation of saturation from electrical properties is treated in section 2.5.3, under electrical
properties.
Fluid saturations are normally reported as a percent of the pore volume, and the accuracy of
measurements is largely determined by conditions during sample recovery.

55

2.5.3 Special Core Analysis

The special core analysis is divided into two stages. In the first stage, the measurement of
permeability and porosity are repeated using other techniques and further coring analysis is done
to obtain measurement of capillary pressure, relative permeability, electrical properties, and cation
exchange capacity (CEC) (Holditch, 1987). A parameter of interest that influences most of the
properties in the first phase is the wettability of the sample. It is a measure of the preferred
inclination of a fluid, i.e water or oil to spread on the rock surface (Unalmiser and Frank, 1998). It
combines the interaction of the rock surface, fluid interfaces and pore shape (Morrow, 1991).
Category 2 of the special core analysis involves measurement of formation geomechanical
properties like Poissons ratio, Youngs modulus, and fracture toughness (Holditch, 1987).
Capillary pressure
Capillary pressure is used to characterize the reservoir by indicating water saturation, size of pore
channels and differentiating productive from non-productive intervals (Keelan, 1982). Laboratory
techniques for determining capillary pressure include: porous plate, centrifugal testing, mercury
injection and water vapor desorption (Unalmiser and Frank, 1998).
Electrical Properties
According to McPhee (2012), a fundamental set of equations was defined by Archie (1942), that
gave the relationship between porosity (), formation resistivity (Rt), formation water resistivity
(Rw), core resistivity, Ro and water saturation of reservoir rocks (Sw). The Archie Water Saturation
Model, based on the electrical properties of the rock is given below:

..(2.23)

56

...................(2.24)

Combining equations (2.23) and (2.24), give:

1 (1/n)
]

= [

.....................(2.25)

The porosity exponent, also called cementation factor, m and saturation exponent, n are
obtained from formation factor, (F) and resistivity index (I) tests on the core.
Youngs Modulus
The Youngs modulus is obtained from coring analysis by conducting triaxial compression tests.
However, logging techniques have been advanced to obtain the Youngs modulus from velocity
measurements obtained during logging analysis. This would be outlined in section 2.6
2.6 Logging Analysis
Logging operations is a very important part of formation evaluation. Extensive work has been done
on improvement of logging tools and monitoring programs. Well logging can be performed at any
stage of a wells development: drilling, completions, production or abandonment. This section
presents a brief overview of the typical logs. Holditch (1987), recommends the following
precautions for openhole logs:
1. A low water-loss mud system should be used.
2. The hole should be drilled at a moderate rate of penetration to reduce the possibility of
washouts occurring.
3. Calibration of the logging tools and adequate maintenance has to be done before logging
operation.

57

Logs can be broadly grouped into: electrical logs, lithology logs and logging-while-drilling (LWD). It
must be mentioned here that LWD is not necessarily a group of logging operations but a
condition of logging. Thus, electrical or lithology logs can be LWDs depending on the particular
constraints surrounding each individual operation.
2.6.1 Electrical logs:
These include: sonic logging, resistivity logging, neutron porosity logging, density logging, image
logging. A popular electrical log is the sonic log. Sonic logging involves the measurement of the
travel time of an acoustic wave through the formation. It is used principally to calibrate seismic
data and to obtain formation porosity (Glover, 1986). The integration of logging and core analysis in
tight gas reservoir (TGR) characterization is of utmost importance, especially in defining porosity.
Due to heterogeneity, formations may contain micro fractures, that can cause secondary porosity
(Orlandi et al, 2011)
Asides porosity, sonic logs are also used to estimate rock geomechanical properties. A density log
and a full wave form (shear and compressional waves, see Fig. 2.29) are recorded after running a
sonic log. Holditch (1987), outlines the following equations that can be used with data from sonic
logs to calculate the rock mechanical properties:

Poissons ratio, u

2
0.5 1

2
1

.(2.26)

Shear Modulus, G = 1.34x1010 b/ts2

..........(2.27)

Youngs Modulus, E = 2G(1+u)

..(2.28)

58

Where is the shear wave travel time (sec/m), and is the compressional wave travel time

(sec/m) and b is the bulk density (g/cm3).

The typical responses of a sonic log to varying soil layers are depicted in Fig 2.30.
2.6.2 Lithology Logs
As the name implies, these logs are useful in describing the different layers encountered as the
borehole is drilled and also to identify geometry of fractures present. They are used in pre-fracture
and post-fracture formation evaluation. Types of lithology logs include: temperature logs, gamma
ray logs and spontaneous potential (SP) logs. Temperature logs are shallow investigative tools, used
to infer fracture height, but are inadequate for use in deviated boreholes. A comparison of prefracture and post-fracture temperature logs is also useful in determining changes to formation, well
bore and temperature gradient after completion operations (Jones and Britt, 2009). Fig 2.31 shows
temperature logs depicting conductivity and fluid movement effects, before and after fracturing.
Gamma ray log is another widely used lithology log. Fracture azimuth is determined using a
shielded gamma ray log and gyroscope, with the fracture geometry being traced with radioactive
tracer (Jones and Britt, 2009).
2.6.3 Logging while Drilling (LWD)
The meaning of this logging technique is evident in its name. It simply means, carrying out logging
at the same time drilling is done. LWD is very popular with development of unconventional
reservoirs such as the Barnett shale. Early horizontal drilling in Barnett shale, made use of LWD,
particularly gamma ray logs (Quinn et al, 2008). An example of an LWD gamma ray log is shown in
Fig 2.32. It shows bedding planes and fractures using an LWD electrical image. Over the years,
major advances in LWD operations have been made. Real-time, visually clear LWD images
59

describing fracture orientation are now possible, such as those shown in Fig 2.33 (depicting drillinginduced tensile fractures).
2.6.4 General Log Characteristics of Gas Shale
Orlandi et al (2011), lists the following as typical log behavior in gas shale reservoirs:
1. High gamma ray (GR) activity because of its high uranium content. The presence of uranium
can be associated with organic matter. Hence Schmoker (1981) proposed a relationship
between the total organic carbon (TOC) and gamma ray activity. A plot of TOC log and GR
log is shown in Fig. 2.34.
2. Considerable resisitivity due to the presence of kerogen and non-conductive gas.
3. Lower bulk density than surrounding rock, because of organic content in the shale rock
matrix.
4. Increase in travel time of compressional waves because of the presence of organic matter.
Acoustic compressional waves have low velocity in kerogen.

60

CHAPTER 3 FRACTURE GEOMETRY

In the design of hydraulic fractures, most procedures to optimize well productivity begin with the
fracture size. There are several approaches proposed to obtain the optimum fracture size, these
have been documented by vast technical literature on the subject. Limitations in the different
hydraulic fracture design methods are inherent in their assumptions of fracture geometry,
dependence on fracture fluid/reservoir properties, layered formations and other factors like stress
intensity. According to Robert Kennedy et al (2012), challenges in fracture geometry when
fracturing unconventional reservoirs include: fracture azimuth and dip, not creating expected
length, brittle and ductile rocks complex and simple networks, well bore axis (vertical or
horizontal drilling).
In all cases however, knowledge of existing in-situ stress tensors is essential to developing a
fracture propagation model which describes the methods of obtaining a desired hydraulic fracture
geometry definitely including the fracture (half) length, width, height and fracture complexity.
There are new techniques of creating complex fracture networks with low viscosity fluid and multistage fracturing methods.
G. Hareland and P.R Rampersad (1994) propose the following factors to be considered in optimizing
hydraulic fracture design for low permeability gas reservoirs:
1. The relationship between fracture dimensions and reservoir production.
2. The suitability of materials required for the fracture operation (propping agents,
fracturing fluid etc).
3. The relationship between cost of fracture operation and reservoir properties (yield)

61

3.1 In-situ stress Determination

Since the classical Hubert and Willis (1957) paper on hydraulic fracture mechanics, it has been
proven both theoretically and empirically that hydraulic fractures propagate in a plane
perpendicular to the least principal stress (Zoback and Haimson, 1981). In deep formations (
300m) because of considerable overburden stress, this least principal stress is usually the minimum
horizontal stress hmin. Geologically, this is true in normal and strike-slip faulting environments but
in reverse faulting environments, the least principal stress is the vertical stress. The magnitude of
this minimum horizontal stress can be obtained from the pressure in the fracture, immediately
after pumping has stopped called the instantaneous shut-in pressure (ISIP). The maximum
horizontal stress, hmax is also useful in well bore stability analysis such as determination of
optimal mud weights, well trajectories, casing set points and determination of possibility of shear
failure on pre-existing faults. Despite its importance in geomechanics, the Shmax is the most
difficult stress tensor to obtain accurately, particularly as it cannot be measured directly (Zoback,
2012).
As stated previously, the most important pre-fracture stimulation parameter to be obtained is the
minimum in-situ horizontal stress, also taken for the sake of simplification as the closure pressure.
The limitations of this assumption will be explained in a later section. The closure pressure is
defined as the fluid pressure required in commencing the opening of a fracture. Closure pressure is
of critical importance in obtaining reasonable predictions of hydraulic fracture geometry for a given
net pressure and to evaluate proppant strength requirements (Jones and Britt, 2007). Field in-situ
stress measurement is not a straightforward technique. Many variables that affect measurement
are involved, for instance: the effects of perforations and propagation of the fracture beyond the
zone that is being tested. However, by monitoring bottom hole treating pressure (BHTP), injection
62

fall-off and flowback tests, accurate values of in-situ stress can be measured (Holditch et al, 1987).
The apparent drawback of using methods such as logs and core analysis to approximate in-situ
fracture closure stresses is that they must all be calibrated by fracturing rock. Hence when a
definitive value is required, injection tests are more suitable in obtaining closure stress values;
there are three fundamental testing methods used (Jones and Britt, 2007):
1. Pump-in/shut-in tests (Also known as injection fall-off tests)
2. Pump-in/flowback tests (Also known as injection flowback tests)
3. Step-rate injection tests (used to measure fracture extension pressure)
The ideal formation evaluation would be one where the values of in-situ stresses obtained from
injection tests and those calculated from logs and core analysis all result in a consistent stress
profile (Holditch et al, 1987). It is worthy to note that there are several different methods of
measuring rock in-situ stress, including: relief methods, jacking methods, borehole break out
methods, strain recovery methods, acoustic emission methods, fault-slip data analysis, earthquake
focal mechanisms etc but for the purpose of this thesis and considering the prevailing in-situ stress
determination methods for pre-fracturing, only hydraulic methods will be considered. The hydraulic
methods are also the most reliable for determining in-situ stress in deep (> 50m) formations
(Amadei and Stephansson, 1997), hence its emphasis is justified.
Several variations occur for the two pump-in tests outlined above. In both cases, the closure
pressure is obtained by a change in the pressure decline properties as the fracture closes (J. Jones
et al, 2007). Theoretically, the pump-in/shut-in test can be taken to be a special case of the pumpin/flowback test provided the flow rate is set to zero (Hsiao C. et al, 1990). The pump-in/flowback
test can be carried out separately or concurrently with the pump-in/shut-in test. The pumpin/flowback test is operationally more difficult to perform than the pump-in/shutdown test
63

because of the need to maintain a constant and correct flowback rate as the pressure declines. If
the pump-in/flowback test is done properly however it can be compared with results from the
pump-in/shut-in tests to show consistency (Middlebrook et al, 1997). Generally, the pump-in/shutin test is used for obtaining fluid loss to the formation and fracture geometry, while the pumpin/flowback test is used for obtaining closure stress (Hsiao and Tsay, 1990). This is especially true
for low permeability formations like shale, where the pump-in/flowback test is more suitable for
closure stress determination because extensive pressure decline monitoring is required to achieve
fracture closure pressure (J. jones et al, 2007). Summary of plots for shut-in and flowback tests
ranging from normal pressure versus time plots to G-function plots are shown in table 3.1.
3.1.1 Pump-in/Shut-in Tests
This test is conducted by pumping in a fluid at a rate sufficient to create a fracture, then shutting in
the well and allowing the pressure to decline to below closure pressure. The instantaneous shut-in
pressure, ISIP is the pressure obtained at the instant when pumping is stopped and the well is shutoff. It is the pressure at which a fracture stops propagating and closes immediately after pump
shut-off. It is always larger than the closure pressure because a fracture cannot start to close
instantly when pumping is terminated (J. Jones et al, 2007).
Conventionally, in a pump-in/shut-in test, a log-log plot of the pressure difference versus shut-in
time is used to identify the existence of linear or bi-linear flow. A one-half slope, before fracture
closure normally indicates the existence of linear flow , as depicted in the P/t slope in Fig. 3.1.
While a one-quarter slope indicates bi-linear flow as depicted in the P/t slope in Fig. 3.2. In the
case of linear flow, the closure pressure can be obtained from the (linear/log-log) plot of the
bottom hole treating pressure (BHTP) against square root of elapsed time, evaluated for the
different pumping cycles. It is the point at which the graph deviates from a straight line. Also, for a
64

bi-linear flow the fracture closure pressure can be determined from a (linear/log-log) plot of the
bottom hole treating pressure (BHTP) versus one fourth root of the shut-in time (Hsiao and Tsay,
1990). This value is checked for different cycles to ensure reasonable consistency.
In a pump-in/shut-in test, ISIP can be obtained from normal plots of pressure versus time, log-log
plots and semi-log plots of pressure versus time. The reason for the use of these different plotting
techniques will be discussed later. The normal plot of pressure versus time has the advantage of
giving accurately the length of shut-in time, since real time values are plotted (Jones and Britt,
2007). In a normal plot, the ISIP value is obtained from the point in which the graph deviates from a
straight line (Fig. 3.3)
There are however, a number of factors that complicate the determination of hmin, these
include:
1. Multiple Shut-In Pressures
In cases such as reverse faulting environments, where the least principal stress is the vertical
stresss, a vertical fracture forms at the well bore at an azimuth perpendicular to hmin (assuming
no natural fissures like joints or sub-horizontal bedding planes are present), and rotates into a
horizontal plane perpendicular to the vertical stress as it propagates away from the well bore.
Hence, the plane pressure decreases as fracture plane rotates. This effect is frequently observed at
shallow depths ( 300m) and doesnt influence results significantly as long as proper interpretation
of pressure-time history is done (Zoback and Haimson, 1982). This is shown in Fig. 3.4 which show
pressure records of pump-in tests taken in shaly dolomite near Anna, Ohio.

65

2. Decrease in shut-in pressures

Another factor observed, is the decrease in shut-in pressures as fractures are propagated. These
have been explained by Hickman and Zoback in 1982 to be due to viscuous effects of fracturing
fluid. Thus, the shut-in pressure that should be used as a measure of Shmin are the final values of
shut-in pressure (Zoback and Haimson, 1982). This is shown in Fig. 3.5, which are results from a well
drilled in granitic rock near Monticello, South Carolina.
3.

Gradual pressure changes upon shut-in (Indistinct ISIP)

In this case, the pressure change after shut-in is too gradual to obtain any distinct measurement of
shut-in pressure from the normal plots of pressure vs time. This is caused primarily by leak beyond
the intervals, past the packers or further propagation after pumping stops, (Zoback and Haimson,
1982). Various methods have been proposed by investigators to solve this problem. Amadei and
Stephansson (1997) explain these various methods of interpretation of shut-in pressures using data
obtained at the Aspo Hard Rock Laboratory for nuclear waste disposal in southeast Sweden. These
are shown in figs 3.6 to 3.12. Klasson (1989) suggested a method of using the inflection point after
shut-in to determine ISIP. This makes used of double tangents to obtain points of intersection from
the pressure decay curves. This method is depicted in figs. 3.6 to 3.8. It can be seen that the
drawing of the tangents depend on personal judgement which is a drawback.
An alternative method of interpretation of shut-in pressure is called the exponential pressure decay
method or Muskat method. The basic assumption of this technique is that pressure decay after
fracture closure approaches an asymptotic value in an exponential manner ( Aamodt and
Kuriyagawa, 1983). The shut-in pressure is obtained by plotting In(P-P) against time, where P is
the asymptotic pressure (apparently the rock formation pore pressure). This is shown in fig. 3.9.

66

Another graphical method utilizing inflection points for shut-in pressure determination is one
proposed by McLennan and Roegiers (1983). It involves a semi-log plot of pressure vs log (t+t)/t
where t is time of pressurization and t is the length of shut in time. This is shown in fig. 3.11. This
plot, also called Horner plot can be used to identify radial flow and estimate reservoir pressure
(Jones and Britt, 2007). See fig. 3.11.
As reported by Amadei and Stephansson (1997), the bilinear pressure decay rate method of
Turnbridge (1989) has become the most reliable and most popular method for the determination of
shut-in pressures. The fundamental concept of this method is that the shut-in pressure decay curve
can be expressed as a flow rate out of the borehole, since the drop in pressure following shut-in
must be related to the quantity of fluid flowing out the system into the hydraulic fracture, and to
the quantity of fluid from other sources of leakage (Turnbridge, 1989). Turnbridge showed
mathematically that the plot of rate of pressure decay versus pressure should consist of two line
segments and the intersection of these segments give an estimate of the ISIP, as shown in Fig. 3.12.
More rigorous methods to single out the shut-in pressure by using nonlinear regressional analysis
to obtain the best fitting linear curve have been done by Lee and Haimson (1989).

3.1.2 Pump-in/Flowback tests

In a pump-in/flowback test, injection is followed by a flowback at a constant rate through a
flowback manifold. This flowback rate is maintained with adjustable choke valves and a low-rate
flowmeter. The primary purpose of the flowback is to establish a rate that is approximately equal to
the fluid leak off rate to the formation. At this flowback rate, a typical reverse curvature occurs in
the pressure decline at closure pressure. A trial and error procedure is then carried out to obtain
the ideal flowrate by performing the first flowback (at 1 to 2 bbl/min) and changing the rate until an
s- shaped pressure decline curve is obtained. The pump-in/flowback test is carried out again to
67

confirm repeatability (Jones and Britt, 2007). A major advantage of the pump-in/flowback test
which will be re-emphasized here is that in shale, the created fracture may not close in a
reasonable amount of time due to its low permeability or the fluid loss properties of the drilling
mud (conducting the shut-in test is not practical), hence the flowback test can be used to induce
the pressure decline in a controlled manner (Lee et al, 2004).
An extra derivative plot of dp/dt can be used to carry out additional analysis for the pumpin/flowback tests. Since, the closure is identified as the point at which the rate of decline
accelerates, closure would be identified with the maximum point on the derivative plot (Jones and
Britt, 2007). See fig. 3.13. Because of the operational difficulty in maintaining a constant (and
correct) flow rate in the pump-in/flowback test it is often used as a last resort for determining
fracture closure pressure.
Raeen A.M et al, (2001) using field data obtained from different offshore locations in Norway
showed how pump-in/flowback tests reduce estimates of minimum horizontal stress significantly
and the conventional extended leak-off tests (XLOT) or mini-fracs overestimate the minimum
horizontal stress by at least 20 bars. He proposed a system stiffness approach for interpreting
pump-in/ flowback tests, based on the following, In a flowback test we may estimate the system
stiffness by dividing the change in system pressure by the corresponding flowback volume, subject
to the condition that the flowback rate is much larger than the leak-off rate to the formation. In
practical situations this is often ensured by low formation leak-off due to tight formations or an
effective mud cake. By flowing back as fast as possible, the error due to fracture leak-off may be
minimized. The fracture stiffness, S in the form of an ellipsoid, for

an open fracture and

considering flowback (non-propagating) conditions is given as:

=

3
(1
16

2 )3

..(3.1)

68

Where R = fracture radius, E is Youngs Modulus and u is Poissons ratio. These are properties of
shale. Raaren noted that the fracture width isnt contained in the equation hence the fracture
stiffness will be constant during flowback, assuming the fracture closes by reducing its width with
the length constant.

3.1.3 Step-rate Injection tests (SRT)

In a step-rate test, the injection is performed at constant, incrementally increasing rates and the
final injection pressure for each step rate is plotted separately against pump rate as shown in fig.
3.14. A typical SRT, showing a plot of final pressures versus pump rate for a number of (fixed flow)
step rates is illustrated in fig. 3.15. Each rate is usually maintained for fixed periods of time, usually
1 to 2 minutes.
The primary goal of the SRT is to obtain the fracture extension pressure, which is obtained as the
break point in the bottom hole final pressure vs pump rate plot shown in figs 3.14 and 3.15.
Alternatively, the fracture extension pressure can be determined by shutting down after each rate
step and obtaining the ISIP. Then plotting the ISIP from each rate step as a function of pump rate
and identifying the inflexion point. The fracture extension pressure is also a good upper bound on
the closure pressure (Jones and Britt, 2007)
Every data collection program should include SRT because it is useful not just in determination of
the fracture extension pressure but also in obtaining the fracture closure pressure.

3.2 Mini-fracture tests and G-function Analysis

Mini-fracture tests, commonly known as mini-fracs or extended leak off test (XLOT) are performed
to obtain important understanding of the geomechanics of hydraulic fracturing stimulation. These
tests are carried out using the planned fracturing fluids, pumped at the planned pump rates but
69

scaled down fluid volumes. Important information on fracture geometry, in-situ stress contrast, and
fluid loss coefficient/fracture fluid efficiency can be obtained from properly designed and executed
mini-fracs (Jones and Britt, 2007). It is important to note that both of the previously described
pump-in/shut-in test and pump-in/flowback tests are variants of the XLOT. One basic difference is
that the shut-in and flowback tests do not use the fracture fluid for injection.
A schematic pressure-time history illustrating an extended leak-off test (XLOT) or mini-frac is shown
in Fig. 3.16. The linear portion of the plot shows a constant pumping rate (pressure versus time)
and a fixed well bore volume. At the pressure where there is a distinct deviation from linearity
(referred to as the leak off point or LOP), a hydraulic fracture is created. This is explainable because
at constant pump rate, the well bore pressure will not decrease unless there is a considerable
increase in the system volume into which injection is taking place. Also, the pressure in the well
bore must be sufficient to propagate the fracture far from the well bore to cause a system volume
increase large enough to affect fluid pressure. Thus, there must be a hydraulic fracture
propagating away from the well bore perpendicular to the least principal stress in the well bore
region, once there is a noticeable change in the pressurization rate (Zoback, 2012). It is reasonable
therefore to say that a distinct leak off point, LOP is approximately equal to the least principal
stress, assuming well bore resistance (caused by high flow rate and high viscosity) and tortuosity
effects between created fracture and well bore are ignored (Fig 3.17). This fact, is the reason why in
typical oil-field practice, leak off tests are stopped after LOP instead of conducting complete,
extended leak-off test (Zoback, 2012). However, a comparison of LOTs and XLOTs will be done later
in the section to adjudge the appropriateness of this practice.
3.2.1 Standard Leak off tests
The industrys ambiguous nomenclature has defined LOTs in several ways. However, there are two
main types of leak off tests (LOTs): Pressure Integrity Test (PIT) and Formation Integrity Test (FIT).
70

The PIT is a leak off test in which the pressure is increased until the pressure rate decreases, which
indicates that leak off has occurred (Addis,1998). In the FIT, pressure is applied to a pre-defined
value and no leak off occurs. The FITs merely indicate that the maximum well bore pressure did not
exceed the least principal stress or was not sufficient to initiate a fracture of the well bore wall in an
open hole test (Zoback, 2012). Table 3.2 gives a classification of pressure tests, making reference to
Figs 3.18 to 3.20. Leak off tests are performed immediately beneath cemented casing in order to
test the integrity of the set cement and determine the drilling fluid density for the next drilling
operations. The well is shut-in at the beginning of the test, and fluid is pumped into the well bore to
gradually increase the pressure that the formation experiences. The test is normally stopped
shortly after the LOP is reached. For over 40 years, LOTs have been used to determine stress for
drilling, planning, sealing capacity of faults, mud weight design, well bore stability etc. The fact that
LOTs are simple, cheap and similar to hydraulic fracturing tests make them an attractive test option
(Gang et al, 2009).
As reported by Gang et al, 2009 in saturated rocks with low permeability, the pore pressure is often
assumed to be unaffected by the state of stress and the Terzaghis effective stress is applicable to
tensile cracks.
From classical theory of elasticity, considering stresses around a circular opening,

71

Qnat= hmin + hmax + 2(hmin - hmax )cos 2a

(3.2)

At a = 45,
Qnatmax = 3hmin hmax

.(3.3)

Fracture will initiate if,

Q T

.(3.4)

Considering pore pressure and leak-off pressure,

Q = Plo + Pp

.(3.5)

Combining equations 3.3 to 3.5

hmax = 3hmin Plo Pp + T

.(3.6)

For fracture re-opening, Plo becomes Pr and the tensile strength is no more significant , since
fracture initiation has already occurred (Gang et al, 2009),
hmax = 3hmin Pr - Pp

....(3.7)

Where Qnat is natural hoop stress, Pp is the pore pressure, T is the tensile strength of the
borehole formation. Plo is the leak-off pressure. Unlike in hydraulic fracturing techniques, the leak-

72

off pressure instead of the shut-in pressure is taken to be equal to the minimum horizontal stress,
hmin = Plo
It should be noted that the above formula has found application only in shallow formations (2km)
where both stress and temperatures are low, and in vertical boreholes for leak-off tests yielding
horizontal fractures.
The drawbacks of the LOTs lie in the assumption that the fracture initiation pressure (hmin) is
equal to the leak-off pressure since the pressure required to equilibrate the fracture normal stress
is the shut-in pressure. Also, the ISIP that can be used to estimate the hmin is mostly obtained in
the 2nd or 3rd cycle of pressurized loading and not in the first cycle as assumed by LOTs (Gang et al,
2009).

3.2.2 Extended Leak-off tests

The execution of extended leak-off tests (XLOTs) is similar to LOTs, the main difference being the
additional pressurization cycles, this is better shown in fig. 3.17. The XLOTs share a similar
theoretical framework with the hydraulic fracture stress measurements. Considering, an ideal
poroelastic rock, when a fracture is created and oriented coaxially with the hole, the magnitude
and orientation of Shmin in the plane perpendicular to the holes axis can be obtained from the
shut-in/closure pressure (Addis et al, 1998). The magnitude and orientation of the Shmax can be
obtained from the fracture orientation.

Addis et al, 1998 modified the Haimson-Furst equation and present it as follows:
For fracture initiation,
hmax = 3hmin + T kPlo (2-k)Pp

....(3.8)

For Fracture re-opening,

73

hmax = 3hmin kPr (2-k)Pp

....(3.9)

Where k is a poro-elastic constant

A limitation of the XLOTs is that the orientation of the fracture isnt known, hence the direction of
the horizontal stresses cannot be determined (Gang et al, 2009).

Comments on LOTs, XLOTs and Hydraulic Fracturing

There have been several papers written comparing the XLOTs and LOTs, with a view of stating their
inherent differences and comparing their functionality. Gang et al, 2009 gave his critique of LOTs
(PITs) as follows:
Comparison of LOTs and Hydraulic fracturing
1. LOTs are carried out without a downhole packer, which is essential in hydraulic fracturing
to ensure a well-defined seal-off zone.
2. In deep wells, a large volume of drilling fluid (30 to 200m3) is used to pressurize the bottom
hole. This isnt done in hydraulic fracturing to eliminate problems due to pressurizing large
fluid volumes.
3. Neglecting shear stresses can cause serious errors especially in the inversion process of
LOTs.
The inversion process referred to here is with reference to inclined wells.
4. Mud compressibility, casing expansion and leakage of the casing cement can influence
pressure records from LOTs thus undermining their usefulness in stress field predictions.
5. If the open-hole portion of the well exceeds a few meters in length, the leak off test data is
not suitable as a good estimate for stress due to the influence of pre-existing cracks. The
test will most likely re-open the preexisting fractures than generate a new one. A hydraulic

74

fracturing method specialized in stress measurements for such situations is called a HTPF
(Hydraulic test for preexisting fractures)
Some of the above listed limitations like number 1 can be attributed to XLOTs.
Gang et al (2004) compared XLOTs and Hydraulic Fracturing as follows:
1. Non-Newtonian fluid is used in XLOTs which make the test interpretation more difficult
than hydraulic fracturing tests which use water or brine.
2. Rotation of fractures can occur as a result of fast pumping rate of large volumes of water in
XLOTs.
3. Lack of fracture orientation data is still a drawback in XLOTs, consequently making them
less reliable than hydraulic fracturing tests.
It is worth noting that the volume of fluid pumped in leak-off testing is significantly smaller than
hydraulic fracturing and normally limited both horizontally and vertically to a few borehole radii.
We can deduce from this that because of the small volume pumped during leak off tests, the
fracture created is not sufficient to measure far field stresses. However, leak off tests still gain
popularity in practice for the Oil and Gas industry. Normally, hydraulic fracturing is performed in
rocks with considerable permeability and leak-off testing (LOTs and XLOTs) is commonly performed
in shales which possess substantially lower permeability (Lee et al, 2004).
3.2.3 Case study Comparison of LOTs and XLOTs
A more quantitative comparison of the XLOTs (ELOTs) and LOTs was done by Addis et al, (1998 )
based on data obtained from six XLOTs conducted by the Australian petroleum industry on the
North West Shelf of Australia. These are represented in tables 3.2 and Fig. 3.21. Also, data from
standard LOTs and XLOTs conducted on wells drilled by Norsk Hydro in the Oseberg field of the
Norwegian North Sea were analyzed, as seen in table 3.4 and fig. 3.23 and fig. 3.24.

75

As reported by Addis et al, (1998), in the pressure vs time plots of figure 3.21, tests in wells 1 to 3
are considered good quality tests (Ideal cases of XLOTs and LOTs are represented in Figs 3.18 to
3.240) because they indicate:
i.

Well defined peaks in the first pressurization cycle proving new fracture initiation

ii.

Well constrained shut-in curves indicative of negligible fluid leakage to the formation

iii.

Approximately equal shut-in pressures for each cycle depicting consistent fracture
orientation, with regards to stress fields

iv.

Distinct pressure rebound between pressurization cycles

v.

Well-defined fracture re-opening pressure, clearly shown in the third cycle of

pressurization.

Test on wells 4 to 6 are adjudged poor since they have indistinct peaks in the first pressurization
cycle, which may be reflective of pre-existing fractures rather than generating a new fracture.
These tests arent entirely unusable as the minimum stress may still (but with some uncertainty) be
estimated from the lowest shut-in pressure.
Data Interpretation from tests
The North West Shelf of Australia
Table 3.2 shows the shut-in pressures from XLOTs in wells 1 to 3 and the lowest shut-in pressures of
tests 4 to 6. The minimum stress has an average gradient (i.e. average between the six wells in
table 3.3) of 14.25kPa/m (0.63 psi/ft) that is fairly consistent through the good quality (1 to 3) and
poor quality (4 to 6) XLOTs. It was obtained from shut-in pressures using the double tangent
method (shown in figs 3.6 to 3.8) for the pressure vs square root of time plots of the Norwegian
data. Since this minimum stress value is significantly lower than the vertical stress (21.95Kpa/m or
0.95 psi/ft) as seen in stress plots of fig. 3.22, we can confidently assume the minimum stress acts
in the horizontal direction. The scattered values in fig. 3.22 between the overburden stress profile
76

and the minimum stress profile are the leak off pressures obtained from other standard LOTs from
vertical wells in the same area. It is distinctively seen that the minimum stress form a lower bound
to the leak-off stress pressures. The maximum stress values of table 3.2 were obtained from
equation 3.8, assuming a poro-elastic constant, k of unity.

The Oseberg field in the Norwegian North Sea

The data for the Norwegian North Sea are represented in the tables 3.4 and figs 3.23 and 3.24. The
wells tested in fig 3.23 are vertical (<10 degree inclination). The average minimum stress gradient,
16.72Kpa/m (0.738psi/ft) which shows an acceptable range of consistency (16.46KPa/m to
17.23KPa/m) forms a lower bound to the leak-off pressures obtained from the standard LOTs. The
leak off pressures occur between the overburden stress and the minimum stress (broken lines on
plot) as occurred in the data from the North West Shelf. However, in figure 3.24 where data from
wells inclined at up to 66 degrees are recorded it is clear that a substantial number of leak off
pressures now fall below the minimum horizontal stress gradient. This shows that leak off pressure
values that match fracture initiation pressures, which usually decrease in inclined wells. It can also
be inferred that that the leak-off pressure gradient decreases by 0.15kPa/m for every 10 degree of
well inclination, assuming effects of wall orientation are neglected (but can affect leak off
pressures), (Addis et al, 1998).
In most situations, the XLOTs have been shown to be better estimates of the minimum horizontal
stress than standard LOTs. Generally, the LOTs and XLOTs are performed in shale and mudstone
formations which typically have the highest stresses and fracture gradients. Transferring such data
for stress field predictions in formations with low stress gradients such as reservoir sandstones or
limestones is not advisable (Addis et al, 1998).

77

3.2.4 G-Function Derivative Analysis

As previously stated, one of the parameters that can be obtained from the mini-frac test is the
fracturing fluid efficiency. This is defined as the fracture volume (at the end of pumping) divided by
the total slurry volume pumped. It is a measure of fluid leakoff, (Jones and Britt, 2007). The total
volume lost between tp and tp +tc is the volume of the fracture at tp. Dividing this volume by the
total volume injected gives efficiency.
Where tp is the elapsed time between fracture initiation and shut-in, tc is the time between shut-in
and when the fracture closes. ti is the time between start of pumping and fracture initiation. tp + tc
is the total time between fracture initiation and closure. From a plot of fluid efficiency versus the
fracture closure time (tc/ti),
Jones and Britt (2007) give this relationship:
Vf/VL = ef/(1-ef)
(3.10)
where ef is the fluid efficiency, Vf is the fracture volume and VL is the fluid loss volume during
injection. (1-

ef) gives the volume of fluid lost to the formation while pumping. Thus VL *(1- ef)

gives the total slurry volume pumped. is a new variable which is used in G-function analysis
described later, it is given by:
= /4G0p

(3.11)

Where G0 is the pressure difference function equal to 1.57-0.238 ef , p is the match pressure and
K is a correction to the fluid-loss coefficient that accounts for additional fluid loss during pumping
such as spurt loss or losses due to pre-existing fractures (Jones and Britt, 2007). Parameters from
78

the fluid efficiency equations also find use in G-function analysis as will be shown further. Fluid
efficiency can also be used exclusively for preparing fracturing treatment design schedules.
The G-function derivative analysis is basically a manipulation of dimensionless time. It is another
method of analyzing pressure records from the mini-frac test to obtain better estimates of
minimum stress. It is useful in representing non-ideal decline behavior by creating plots of pressure
vs G-function and dp/dg as a function of g function. G-function is also defined as the representation
of the elapsed time after shut-in, normalized to the duration of fracture extension, (Barree et al,
2009). Barree et al, (2009) presents multiple analysis techniques (involving G-function) for holistic
interpretation of mini-frac tests. These multiple analysis methods are done for different cases,
including: pressure dependent leakoff, fracture tip extension and variable fracture storage as
shown in figs. 3.25 to figs. 3.27. Barree et al made a derivation of G-function, assuming high fluid
efficiency in low permeability formation.
The dimensionless pumping time used in the G-function is defined as:
tD = (t-tp)/tp

....(3.12)

Where t = tp +tc: total time from start of fracture initiation to closure

Considering low leak-off, the dimensionless time (tD) is used to compute an intermediate function,
g(tD) = 4/3 [(1+tD ) 1.5 - tD 1.5)]

..(3.13)

The G-function used in the diagnostic plots shown in Figure 3.2 L N is derived from the
intermediate function as shown below:
G(tD) = 4/p [g(tD) - g0 ]

.(3.14)

Where g0 is the dimensionless fluid loss volume function at shut-in (ie when t=tp and (tD =0). All
derivatives are calculated using a central difference function of pressure and G-function
(normalized shut-in time).
79

3.3 2D Geometry Models

3.3.1 Howard and Fast Fracture Area/Extent
According to G. C. Howard and C.R Fast (1970), fracturing fluid and reservoir rock properties have
significant effects on fracture penetration. R.D Carter derived an equation for calculating the area
of a hydraulically induced horizontal or vertical fracture. Among the assumptions he made in
deriving his equation are: 1. Uniform width of the fracture 2. Linear flow of fracturing fluid from
fracture and perpendicular direction of flow from the fracture face 3. The velocity of flow into the
formation at a point is dependent on the duration of flow (time-dependent flow). Howard and
Fasts assumed fracture geometry is shown in Fig. 3.28
G.C Howard et al (1970) publish Carters derivation as follows:
Volume rate at which fluid flows linearly from fracture into the formation is given by,
()

iL (t) = 20

dAf

.....(3.15)

Af is the area of fracture, v is the velocity of flow and iL is the volume rate

Since Af is a function of time, and the value of v corresponding to a given element dAf at time is

v(t-),thus:

iL (t) = 20 ( )dAf/dt

...(3.16)

The rate at which the fracture volume is increased is:

Qf = W dAf/dt

..(3.17)

80

Where W is the fracture width

The injection rate of fluid, I is the sum of the rate at which fluid flows to the formation,iL and the
fracture volume increase,Qf. That is:
.....(3.18)

I = iL + Qf
Subs. Eqns 3.16 and 3.17 into 3.18,

I = = 20 ( )dAf/dt + W dAf/dt

...(3.19)

Solving using Laplace transformation,

Af =

4 2

( 2 . () +

1)

Where x = 2C

..(3.20)

..(3.21)

This equation clearly shows that the pump rate is directly proportional to the fracture extent.
The coefficient, C is is the fracturing fluid coefficient. It is sometimes referred to as the fluid leak
off. It is determined by the fracture fluid properties and the reservoir fluid and rock characteristics.
The value of C is obtained from the flow mechanism during injection of fracturing fluid. A high
coefficient means high fluid loss, while a low coefficient means low fluid loss. Low fluid loss
properties during treatment mean a larger fracture area for a given volume and injection rate since
there is less fluid lost ineffectively to the formation during treatment. Thus, C relates the fracture
extent to the aforementioned properties, G.C Howard et al (1970)

81

Linear Flow Mechanisms

According to G.C Howard et al (1970), there are three types of linear flow mechanisms defined by
the fracturing fluid coefficient, C. Thus, the value of C depends on the mechanism encountered. The
value of C in mechanisms 1 and 2 can be obtained from reservoir data and existing charts. The third
mechanism however has to be obtained experimentally because it involves fluid loss. It is important
to note here that for fractures in horizontal wells, the flow is linear between the fracture and the
reservoir, but inside the fracture near the well, the flow changes from linear to radial, this results in
a considerable reduction in flow rate from each individual transverse fracture compared to fracture
in a vertical well (Valko et al, 2005). Thus, these three linear flow mechanisms can only be applied
to a horizontal well, when considering the flow to the reservoir normal to the fracture. Hydraulic
fracture in a horizontal well will be treated in a later section.
The three linear flow mechanisms are:
Mechanism 1- The effect of fracturing fluid viscosity and relative formation permeability, C1
This occurs under conditions of injecting a highly viscous fracturing fluid at constant pressure, given
by,
Linear flow velocity, v =

= /2 =

C1 = 0.0469/ , ft/min0.5

1/

..(3.22)

..(3.23)

Where, k is the formation permeability (darcies), p is the difference in pressure between the fluid
at the formation face and the far-field fluid in the formation (psi), t is the time (mins), is the
viscosity at bottom hole conditions(centipoise, temperature effects accounted for), is the

porosity.

82

Mechanism 2- The effect of fracturing fluid viscosity and compressibility effects of reservoir fluids,
C2
This occurs when the fracturing liquid has similar physical properties with those of the reservoir
fluid. In this case, the fluid flow rate is determined by both the coefficient of compressibility of the
injected fluid and the existing reservoir fluids.
Linear flow velocity, v == c/ = 2/
C2 = 0.0374 c/, ft/min0.5

..(3.24)
..(3.25)

Where n is the porosity and c is the fluid compressibility (measured in psi-1).

Mechanism 3- The effect of wall building fracturing fluids, C3
As reported by, G.C Howard et al (1970), it is possible to determine experimentally from actual
formation cores, the fluid loss characteristics of fracture fluids that contain fluid loss additives, by
plotting the empirically obtained cumulative filtrate volume vs the square root of flow time. The
straight line relation obtained is given by,
V = m + Vsp

..(3.26)

m is the slope of the straight line relation of the plot. Obtaining C from this mechanism is explained
in detail in section 4.2.3 of the next chapter. The intercept, Vsp is the spurt loss. This increases the
fracture width, W shown in eqns 3.19 to 3.21.
The flow velocity is obtained by differentiating eqn 3.16 with respect to t and dividing by Ac which
is the cross-sectional area of the medium through which the fracture takes place (packers). Ac is the
diameter of the well x length between the packers.
83

Linear flow velocity, v =

C3 = 0.0164

,
A

..(3.27)

..(3.28)

3.3.2 The PKN and KGD Geometries: Fracture width

In the past, hydraulic design procedures were mostly based on the Perkins and Kern (PK) and the
Khristianovich and Zheltov (KZ) models, excluding numerical simulation methods. In these two
models, the fracture width and length are functions of the continuity, elasticity and fluid flow
equations (G.R Hareland and P.R Rampersad, 1994) Both of these 2D propagation models assume
constant height along the fracture length. The principal difference in assumptions being plain strain
on the vertical plane for the PK model and plane strain on the horizontal plane for the KZ models.
The PK method is suitable for long fractures (xf/hf 1), while the KZ model is more appropriate for
short fractures (xf/hf 1), (Van Eekelen, 1982). The two models are mutually exclusive, elegant
methods and cannot be used interchangeably, Valko et al(2005)
Perkins and Kern Model
Perkins and Kern originally developed their 2D model for hydraulic fracture prediction, which was
further modified by Norgden and became the PKN model. Peter Valko and Michael
Economides,(2005) describe the PKN condition, saying the vertical plane strain along a fracture with
length greater than height (xf/hf 1) allows vertical planes to slide against each other. This is
shown in fig 3.29. Thus, in the PKN model, the formation stiffness is concentrated in vertical planes
perpendicular to the direction of the fracture propagation, Van Eekelen (1982). The fracture cross
section in this plane is assumed elliptical and the formation stiffness in the horizontal plane is
84

discarded, as depicted in fig 3.30. The fracture height is assumed to be constant. The half-length is
obtained by procedures discussed in section 2.2.2 of chapter 2.
Valko at al(2005) present the PKN width equation as:

(4)
3.57

(3.29)

where Wmax = maximum fracture width at the well bore, = viscosity of fluid injected, Xf =
fracture length, I = injection rate and E = modulus of elasticity of rock (Youngs Modulus).
Considering the shape factor, of the well bore, the average fracture width is:
1

(4)

= 2.24

(3.30)

Valko further coupled the PKN width equation by material balance, considering constant fluid
injection rate and no leak off. The net pressure in the well bore was obtained as:

= 1.52

1
5

( )
4 2 4
6

(3.31)

where = net pressure at well bore, t = time , hf = fracture height.

In a different review of the PKN model, Jones and Britt (2009) present relations as follows:
4
4 (1)

+
4
4

= ( )

4 1
+ 4 (4)

(3.32)

(3.33)

Where E is the plain strain modulus, [E/(1-u2)], E is the Youngs modulus and Kic is the apparent
fracture toughness.
85

The Khristianovich and Zheltov model

The Khristianovich and Zheltov model was further developed by Geertsma and deKlerk to become
the KGD model. In the KGD model, the formation stiffness is concentrated in the horizontal plane,
Van Eekelen (1982) Due to the horizontal plane strain condition, the resulting fracture cross-section
is rectangular, (Valko, 2005) with the formation stiffness in the vertical plane discarded as shown in
figs 3.31 and 3.32.
Valko et al(2005) present the KGD width equation as:

= 3.22

1
4

( )

(3.34)

Considering the shape factor, of the well bore, the average fracture width is:
1

(4)

= 2.53 hf

..(3.35)

Considering constant injection rate and no leak off, the net pressure is given by:

= 1.09

1
3

( )

1
3

( )

(3.36)

Comparing the average widths of the PKN and KGD models,

The ratio of the average KGD width to the PKN model is

2 (4)

0.95 .(3.37)

The equations of average fracture width and net pressure for the PKN and KGD models were
derived considering Newtonian behavior of fracture fluids. However for design consideration, most
fracturing fluids exhibit non-Newtonian behavior (G. Hareland et al, 1994), according to the power
law:

86

t =

.(3.38)

A suitable way of adding Newtonian behavior to the width equations is to include another equation
that links the equivalent Newtonian viscosity, e to the flow rate, thus the PKN and KGD width
equations will be in terms of the power law width parameters.
Radial Width Equation
In the case of a radially propagating fracture, that is xf =hf/2 = R. Fig 3.33 depicts radial fracture. A
modified PKN or KGD equation can be written as follows:

(4)
2.24

.(3.39)

In another paper, Jones and Britt, 2009 give the following relations for a radial fracture, where

1
4

( )

(4)

(3.40)

.(3.41)

3.3.3 Fractures in horizontal wells

In horizontal wells, transverse fractures are relatively more difficult to achieve than longitudinal
fractures. However, for shale gas formations normally characterized by low permeability,
transverse fractures in horizontal wells have greater production benefits (Valko et al, 2005)
Transverse vertical fractures move along the path of least resistance, which is normal to the
minimum horizontal stress as shown in fig. 3.34. In horizontal wells or deviated wells, there are
effects in the immediate vicinity around the bore that lead to the transverse fractures taking
tortuous paths before eventually becoming normal to the horizontal stress (fig. 3.35). These effects
87

are increased by the presence of natural fractures in the formation and the deviation of the
horizontal well at an angle from the minimum horizontal stress, shown in fig. 3.36
3.3.4 Fracture Geometry in Layered Formations (Fracture Height)
The simplified geometry proposed by the PKN and KGD models are not adequate when considering
multi-layered formations. This is because the different layers have varying properties such as:
elasticity, fracture toughness, ductility, permeability and interface bonding (Van Eekelen, 1982)
Thus different fracture widths are expected at the different layers (see fig. 3.37). Also, the PKN and
KGD models assume a constant fracture height, which is estimated based on barrier action of
rock layers above and below the zone of interest (Van Eekelen, 1982). This estimated fracture
height further affects the calculation of length, width and proppant transport.
Proppant screenout is the inhibition of proppant transport to the target formation. It can be caused
by the existence of layers of higher stiffness, between target formation layers as shown in 3.38. This
can cause the termination of treatment, as depicted in fig. 3.39 (Valko et al, 2005).
To account for the stiffness effects of multi-layered formations and design fracture heights
accordingly, pseudo-3D models have been developed.
3.3.5 Economics of 2D models: Optimum Fracture half- Length
This dimension is of interest to the oil and gas prospecting industry because the goal of hydraulic
fracture design is to maximize the post treatment performance and subsequent benefits at the
lowest treatment costs (Valko et al, 2005).

Fracture conductivity, width and height are all

determined with respect to a fixed fracture length (G. Hareland et al). Thus, the propped width,
fracture permeability, assumed fracture half-length and the reservoir permeability are used to
estimate post treatment performance. This enables the prediction of future incremental benefits,
88

which when discounted to the present gives the net present value (NPV) of the revenue. The
treatment costs are determined from the estimates of required fracturing fluid volume, proppant
quantity, injection rate etc and subtracted from the present value to obtain the NPV for the
particular fracture half-length (Valko et al., 2005).
The fracture half-length cannot be arbitrarily made to be long, because longer lengths require
higher treatment costs. Also, longer lengths do not always increase the NPV. From the figs 3.40, a
comparison of gas produced from different fracture half-lengths is done. The maximum revenue is
not strictly the shortest or longest production duration. The best production period is based on
the NPV which is a function of fracture length (Hareland et al, 1994). In the low permeability,
shale gas reservoirs the longer fracture length may start with a high rate of production but
eventually decline after a short period, this is shown in fig. 3.41. Thus, a unique fracture half-length
has to be determined for every specific case, considering the reservoir characteristics. (Hareland et
al,1994)
Typically half lengths range from less than 100m for a high permeability formation to more than
500m for a low permeability reservoir (Valko et al, 2005).

3.4 Other Geometry Models

3.4.1 Pseudo 3D Models
Abundant technical publications exist within the petroleum industry that describe pseudo 3D
models, simplified 3D models and fully 3D models. Simplified 3D models identify the fracture shape
based on the rock characteristics but do not fully solve the fluid flow problem, which is solved in
the more complex full 3D models (Holditch et al, 1987). Pseudo 3D (P3D) models are useful
approximations, because of the prohibitive complexity of the 3D models. In P3D models, the
89

pressure drop, width and equilibrium height are calculated along the fracture length (Barree et al,
2009). Several variations of P3D models exist but the fundamental difference between the P3D and
2D models is that the fracture height isnt constant along the length.
A typical hydraulic simulation involves the solution of the elastic continuity equation, which relates
pressure to crack opening and the constitutive equations that describe fluid flow. Early
development and testing of a Pseudo 3D model was proposed by Settari and Cleary (1986). Initial
P3D models assumed a 1D fluid flow for long fractures (high fracture length/width ratio), although
incorporation of 2D flow in P3D models has since been advanced (Xiaowei). Typically, in a P3D
model, the fracture is divided into vertical sections and each section is taken to be a 2D vertical
crack as shown in Fig. 3.42 (Xiaowei,1992). Xiaowei Weng (1992) presents a study comparing the
original P3D models (that assume 1D flow); a modified P3D model (considering 2D flow) and the full
3D models. Xiaowei, illustrates the equations of a conventional P3D model as follows:
The crack opening pressure, is related to the minimum stress and fracture width, by:

P(x,z) (z) =

G(z, z0)

db
0

(3.42)

Where p = fluid pressure, = minimum horizontal stress, b = fracture width, hu and hL = upper
and lower heights, and G = influence function.
Assuming 1D flow, the pressure drop is a function of the vertical, z co-ordinate and the continuity
equation across the P3D model section is given by,
jqz + qL + jb
jz
jt

=0

(3.43)

where qz = flow-rate in the z direction. qL is the leak-off rate. qL is given as,

qL = 2/t-t

90

where C = fluid leak-off coefficient and t is the fluid arrival time. Equations 3.33 and 3.34 can be
used to determine the fracture width, pressure distribution and height growth rates.
The pressure gradient/flow relation is given below:
- jp = 2K (

4+2 n
)

qzn b-(2n + 1)

(3.44)

jz
where n and K are the power-law and fluid-consistency indices respectively, considering NonNewtonian behavior
Xiaowei further proposed a modified Pseudo 3D Model that assumes a 2D continuity equation for
fluid flow across the sections. This is shown below:
jqx + jqz + qL + jb
jx
jz
jt

=0

(3.45)

where qx = flow-rate in the x direction

He considered a modified radial flow field for a penny shaped fracture, as shown in Fig. 3.43 and Fig
3.44. The inner region exhibits horizontal flow, while the outer region shows radial flow. The
direction of radial flow shown is perpendicular to the fracture boundary. Also, the streamlines
diverge significantly for the penny shaped (h=2L) than for the near- well bore region of elongated
fractures (L>h)
Xiaowei gives the continuity equation for the assumed radial flow in the outer region as:
(1/r)(j/jr)(rqr) + qL + (jb/jt) = 0

.(3.46)

Where r = radial distance from imaginary source and qr = radial flow rate. Revising this equation by
changing the derivative with respect to r to one with respect to z, he obtained,

91

j/jz(rqr) + zqL + z(jb/jt) = 0

.(3.47)

The pressure gradient/flow relation now changes to:

- jp = 2K (
jz

4+2 n
)

qrn-1 - qz
b

.(3.48)

-(2n + 1)

where qz = qrsinQ and Q is the angle of the streamlines measured from the horizontal axis.
The method of collocation using Chebyshev Polynomials,as done by Settari and Cleary (1986) was
used by Weng to solve for width in equations (3.33) and (3.36).
A comparison between the conventional 2D, original P3D, Xiaowei-Modified P3D and fully 3D
models was done. The fully 3D models were the Terra Tek and University of Texas models. Results
of the predicted fracture geometries using these different models are shown in Fig. 3.45. From the
results it is seen that University of Texas fully 3D models predicts narrower (smaller height) but
longer (extended fracture half-length) than the P3D and the 2D models. The vertical KGD model
predicts a shorter fracture but with greater height.
Currently available Pseudo 3D hydraulic fracturing simulators include: MFRAC, StimPlan, eStimPlan, Frac Cade (Baree etal, 2009). An example of a StimPlan output is shown in Fig. 3.46.
3.4.2 Fully 3D Models
Fully 3D models to simulate hydraulic fracturing treatments have been developed. The fully 3D
models have the capacity to describe more accurate fracture geometry but are more complex to
execute. Fully 3D models formulated include one by Settari and Cleary (1984) and Van Den Hoek
(1993). In the model developed by Settari and Cleary (1984), a set of equations were formulated,
considering:

92

1. A 3D, two-phase flow in the reservoir.

2. Two phase Non-Newtonian fracture flow and heat transfer, considering losses due to overburden stress.
3. Vertical and lateral fracture propagation
4. Proppant transport, wellbore hydraulics, PVT relations and rheology characteristics of
reservoir and fracture fluids.
The physical system of the Settari and Cleary model is shown in fig. 3.47. Settari and Cleary (1984)
make an additional comparison of their model against the 2D models of fracture geometry using
several examples. A drawback of the 3D model is its sensitivity to the many required input
parameters and consequently, the difficulty in obtaining all the data necessary for the models
successful application. Typical data required for the 3D model application is shown in Tables 3.5,
3.6 and 3.7.
The essential feature of the 3D model is its ability to predict fracturing pressures, see figs. 3.48 and
fig. 3.49. The pressure response varies considerably, with almost constant pressure in the 3D high
containment case. In the model example illustrated, the properties of the adjacent strata were
unknown, so high containment and low containment are hypothetical upper and lower bounds
of over-burden stress from layers surrounding the pay zone (Settari and Cleary , 1984).
Currently available 3D simulators, include: GOHFER, N-Stim Plan, FracPRO and Terra Frac. An
example of a FracPRO output is shown in Fig. 3.50.

93

3.5

Direct mapping techniques for hydraulic fracturing

The two main technologies, available in the petroleum industry of directly mapping hydraulic
fractures in order to obtain fracture geometry are the microseismic mapping and downhole
tiltmeters. Each technology measures different features of the hydraulic fracturing process (Du et
al, 2008). These two diagnostic tools are carried out at distances from offset wellbores and show
the big picture of far-field fracture growth. One shortcoming of these methods is that they map
the entire hydraulic fracture growth zone but give no information on the smaller details like the
effective propped-fracture length or conductivity. The resolution of these technologies is inversely
proportional to their offset distances from the fracture (Cipolla and Wright, 2002). The placement
locations of these direct mapping techniques are shown in fig. 3.51
3.5.1 Tiltmeters (Surface and downhole)
Tiltmeters are highly sophisticated, extremely accurate biaxial instruments that use bubble
sensors and carpenters levels to determine the change in angle of a surface (Jones and Britt,
2009). The principle of tiltmeters is quite simple. The creation of hydraulic fractures causes allround deformation to the surrounding earth. This induced tilt (deformation) can be measured by
downhole tiltmeters placed in nearby wellbores or surface tiltmeters located on the ground
surface, assuming that the deformation is elastic, as shown in fig. 3.52. Surface tiltmeters measure
the fracture orientation while downhole tiltmeters are used to obtain fracture geometry.
Surface tiltmeter
A typical tiltmeter array consists of 12 to 16 instruments installed in shallow holes (10 to 40 ft deep)
and evenly spaced at radial distances ranging from a few hundred feet to as far as 1 mile around
the treatment well, depending on the depth of treatment zone and anticipated fracture

94

dimensions. Tiltmeters are usually placed at a depth that is 0.4 x overall well depth (Cipolla and
Wright, 2002). Each tiltmeter is packed into position using sand to protect the device from the
effects of weather and noise interference as seen in fig. 3.53 (Jones and Britt, 2009).
The magnitude of induced deformation at the surface is so small typically of order one ten
thousandth of an inch (0.00001 in) because of the distance away from the fracture; hence it is
impossible to measure deformation directly. Instead, the surface tiltmeters measure the gradient of
the displacement or the tilt field which is a function of fracture azimuth, dip, depth to fracture
center and fracture volume, as seen in fig. 3.54. The induced deformation is almost totally
independent of reservoir mechanical properties and in-situ stress state. For instance, a north-south
growing vertical fracture of given dimensions produces almost the same deformation pattern
whether the fracture is in a low stiffness rock like diatomite, high modulus hard carbonate or
unconsolidated sandstone (Wright et al, 1998). Surface tiltmeters are extremely sensitive and can
measure changes in tilt of a surface with an accuracy of about 1 x 10-7 radians (Jones and Britt,
2009).
Downhole tiltmeter
Downhole tiltmeter mapping technology was developed to circumvent the limitations of the
surface tiltmeter by giving estimates of the fracture dimensions. The downhole tiltmeters have the
same operational principle as the surface tiltmeters, but instead of being at the surface, the
tiltmeters are positioned by wireline in one or multiple offset wellbores at the depth of the
hydraulic fracture. Typically, the array consists of 7 to 12 tiltmeters coupled to the borehole with
standard oil-field centralizer springs (see Fig. 3.55) (Wright et al, 1999). Downhole tiltmeters
provide a map of the deformation of the Earth adjacent to the hydraulic fracture. Thus, what is

95

obtained is an estimate of an ellipsoid that best approximates the fracture dimensions, as seen in
fig. 3.56
Typically, downhole tiltmeters are located closer to the fracture than the surface tiltmeter and
hence more sensitive to fracture dimensions (Cipolla and Wright, 2002). The closer the downhole
tiltmeter to the fracture, the better the quality of data obtained to determine fracture height
(Jones and Britt, 2009). Fig. 3.57 shows a downhole deformation pattern from tiltmeters in an
offset well, at 100 ft from the injection (stimulated) well. As seen in the figure, the downhole array
tilts in a continuous fashion, similar to surface tiltmeter records but the arrays span the same depth
interval as the zone being fractured. The total interval covered by a downhole tilt array ranges from
300 ft to >1000ft, depending on the design conditions (Wright et al, 1998). Conventionally, surface
and downhole tiltmeter analysis are done separately but techniques have been proposed to
combine them for evaluating fracture geometry during drill cuttings disposal (Griffin et al, 2000)
The greatest advantage of both surface and downhole tiltmeter fracture mapping is that for a given
fracture geometry, the induced deformation field is almost completely independent of formation
properties. Also, the required degree of formation description is lower in tiltmeter mapping than
microseismic mapping (velocity profiles, attenuation thresholds etc) as will be described in a later
section. Complex fracture growth would yield independent fractures at different orientations or
depths but in tiltmeter mapping a simpler analysis is required (Wright et al, 1998).
3.5.2 Micro-seismic mapping
Microseismic theory and mapping is based on earthquake seismology. Similar to earthquakes, but
at a much higher frequency (200 to 2000Hz), microseismic events emit elastic P waves
(compressional) and S waves (shear waves) (Jones and Britt, 2009). During hydraulic fracture, there
is an increase in formation stress proportional to the net fracturing pressure, and an increase in
96

pore pressure due to fracturing fluid leak off. The increase in stresses at the fracture tip and pore
pressure increments causes shear slippages to occur, as shown in fig. 3.58. Microseismic technology
thus uses earthquake seismology methodologies to detect and locate these hydraulic fracturing
induced shear slippages, which resemble micro-earthquakes. Fig. 3.59 shows a schematic of the
hydraulic fracture creation and the tip and leakoff processes develop compressional and shear
wave arrivals. Microsesisms or micro-earthquakes occur with fracture initiation and are observed
with receivers placed on an offset wellbore like with the downhole tiltmeters.
Microseismic mapping technology involves installing an array of tri-axial geophone or
accelerometer receivers into an offset well at approximately the depth of the fracture (like in
downhole tiltmeters) see fig 3.51, orienting the receivers (geophones), recording seismic data,
finding micro-earthquakes in the data and locating them. Locating the earthquake events requires
the determination of compressional (P) and shear (S) wave arrivals and consequent acoustic
interpretation of the velocity of the P-S waves, as shown in fig. 3.60 (Du et al, 2008). The fig. shows
a three component (tri-axial) geophone with P and S wave arrivals. Employing the arrival times on
the X, Y and Z components, both location and direction of events can be obtained (Jones and Britt,
2009). Standard microseismic mapping use P-S arrival time separation for distance location.
Horizontal and vertical plane hodograms are used to determine azimuth and inclination (Warpinski
et al, 1995). [The hodogram is used to obtain information on the direction travelled by a wave
before it is detected at the receiver, the motion of the receiver itself is plotted on the hodogram,
see fig. 3.61 (Crewes.org)].
The distance,d is computed, using the standard equation,
=

..(3.49)

97

Where Vp and Vs are the p-wave and s-wave velocities and tp and ts are the p-wave and s-wave
arrival times respectively. This equation assumes that the formation is homogeneous and isotropic
with constant acoustic velocities (Warpinski et al, 1995)
The micro-earthquake events are located using either regression algorithms or grid search
methods, assuming that the velocity profile is known. The velocity profile is typically obtained from
dipole sonic log measurements and its accuracy can be improved using perforation timing
measurements (Du et al, 2008). Du et al, (2008) also propose a method of further improving the
velocity profile by joint inversion using both downhole tiltmeter and microseismic mapping. The
resulting microseismic data can then be used to obtain fracture azimuth and fracture geometry.
Further studies to integrate microseismic mapping, fracture modeling and reservoir simulation to
optimize well performance has been done by Liu et al (2009) to optimize well performance in the
Chanqing field, China. An example of practical model calibration using results from microseismic
mapping and net pressure matching is given at the end of this chapter.

3.6. Net Pressure Matching and Calibration of Fracture Growth Models

Despite the efforts invested in the improvement of fracture models (pseudo3D models, fully 3D
models etc), most field investigations reveal that fracture geometry obtained during stimulation,
differ significantly from the model predictions. Barree et al, 2009 summarizes some of these
disparities in common field observations as follows:
1. Less fracture widths than predicted.
2. High net treating pressures are frequently encountered.
3. Height containment is often better than predicted by models.
98

4. Shear failure occurs in the rock masses.

Weijers et al (2005) explains these differences in model predictions and field observations as
follows:
1. Poor measurement of the critical input parameters (Youngs modulus, permeability and
fracture stress) of the models, especially the fracture closure stress.
2. Lack of a complete understanding of all the physical mechanisms involved in hydraulic
fracturing growth.
3. The high net treating pressures encountered in field observations as said by Barree et al,
above is caused by complex fracture growth (multiple hydraulic fractures) which results in
competition between opening fractures. Also, plastic deformation and creation of microfractures in the process zone cause increased fracture growth resistance at the tip
leading to higher observed net treating pressures.
A detailed look at Weijers et al explanation above, reveals the importance of adapting the basic
hydraulic fracturing models to fit the specific formation being stimulated. In other words, to
calibrate the models, based on observation of the actual fracture. Fracture models show how
changes in a fracture treatment should affect the fracture geometry, but as described earlier has a
vague relationship with reality. Direct mapping can measure the fracture geometry of a specific
treatment but cannot predict what changes can occur if the treatment is altered. The synergy of
direct measurements with fracture models gives calibrated fracture models with better prediction
capabilities (Weijers et al, 2005). Thus, the model can now be used to make predictions when
parameters are varied during fracture injection.
This is a two-fold process; it means firstly modifying the models to match parameters observed
through pressure tests i.e. net observed pressures. Secondly, calibrating the model by adjusting
99

calculated geometry to match fracture dimensions seen via direct mapping (microseismic
mapping or downhole tiltmeters). They might be obtained either in real-time or post-fracture
stimulation.
3.6.1 Net Pressure History Matching
The net pressure is a very important parameter in fracture growth prediction. It can be measured
directly in the field to interpret the fracturing process. The net treating pressure as defined by the
relationship in equation (3.32), differs most times from the observed net pressures (Jones and Britt,
2009). The process of changing model inputs and assumptions to calculate a model pressure that
matches the observed net pressure response is called net pressure history matching (Weijers et al,
2005). The fig 3.62 shows bottomhole pressure (BHP) history matching for modeled and observed
cases in unconventional gas wells. The diagram shows a good correlation between the plots of the
actual BHP and simulated BHP (using model after history matching). The observed net treating
pressure is given by the following equation:
, = +

...(3.50)

It is worthy to note that the Bottom Hole treating pressure (BHTP) is the sum of the first three
terms of the equation.
Weijers et al, (2005) presents the determination of terms of the equation (3.46) as follows:

is obtained from surface pressure measurements throughout the treatment.

the hydrostatic head is obtained from calculation of quantity of fluid injected and
knowing the density of fluid components

100

The three main friction components are: wellbore friction, perforation friction and
near-wellbore friction. Wellbore friction is determined from flow loop tests and
near-wellbore friction is obtained from step rate tests (SRT)

The fracture closure stress in the pay zone is obtained from pressure decline
analysis, usually through LOTs or XLOTs.

Although several companies have develop software for calculating BHTP from surface pressure, to
date no technique has been developed to accurately explain all variables affecting friction
pressures (Jones and Britt, 2009)
The earliest attempt to understand the varying behavior of observed net treating pressure was
done by Nolte and Smith (1981). They developed a diagnostic framework for pressure monitoring,
as shown in fig 3.63. The figure shows a log-log plot of fracturing pressure vs pump time. The
interpretation of the plot is based on the work of Perkins and Kern (1961) and Nordgren (1972),
which show that net pressure is proportional to time raised to an exponent (Jones and Britt, 2009).
A summary of the interpretation of the different modes in the Nolte-Smith plot is given by Jones
and Britt, 2009 below:
Mode I Confined-height extension
A log-log net pressure to pump time slope of 1/ 8 to 1/4 implies that the fracture is propagating
with confined height and unrestricted extension, that fluid loss is linear-flow dominated (See
section 3.1.1). Also, the injection rate and fluid viscosity are fairly constant. These assumptions are
supported by the PKN model.
Mode II Constant height growth

101

This flat pressure-time slope indicates that the predicted net pressure increase is reduced by either
stable height growth or increased fluid loss. As expected, this mode of net pressure increase is
temporary, because the balance of fracture growth and fluid-loss which maintains this constant
pressure with time is easily disturbed when the height growth reaches a low stress zone or vice
versa (formation inhomogeneity). This leads to mode III or IV.
Mode III Restricted extension
This mode is a region of positive unit slope (1:1 log-log slope). It shows a build-up of pressure due
to a restriction in fracture growth. This could be caused by a high stiffness barrier action.
Mode IV Uncontrolled height growth
The negative slope is as a result of rapid and unstable height growth into a region of low closure
stress. If a negative slope is observed at the start of the hydraulic fracturing treatment, it indicates
a lack of height confinement. This shows that the fracture will grow radially, and future treatments
should be designed using a radial model.
The Nolte interpretation guide is a powerful tool for explaining what causes the observed net
pressure, which is a first step to pressure matching. However, the main interest of pressure
matching is not only in the interpretation of pressure behavior but the adjustment of the model to
fit the observed pressure. Table 3.7 shows the eight general parameters adequate to create a net
pressure history match for a fracture treatment
Some of these parameters would be referred to, in the fracture model calibration example in
section 3.6.2. These parameters are adjusted in the model, so that the calculated net pressure
matches the observed net pressure. In so doing the predicted fracture geometry using the same
parameters would be similar to field fracture growth.
102

3.6.2 Fracture Model Calibration

Net pressure matching alone (to describe fracture growth) is an indirect diagnostic technique,
hence the need to complement its results with direct diagnostic technology like microseismic
mapping. Two of the matching parameters listed in table 3.7 are not considered in current fracture
models. These are width decoupling and composite layer effect. They are explained with figs. 3.64
and fig 3.65. The consequence of changing the width decoupling parameter in the model is that the
fracture width becomes the half-length. The result of accounting for the composite layer effect in
the model is that the fracture height is exchanged for the half-length. (Weijers et al, 2005).
An advantage of calibrating the model is that the user better understands the reservoir that he/she
is completing thus understanding properties like fracturing fluid effects in the pay zone and during
proppant transport, permeability, closure stress, level of far-field fracture complexity, composite
layering effects etc. This knowledge can then be applied to treatment decisions, for instance
knowing that due to tip effects, multiple fractures would be created as fracture propagates one
could realize that certain fracture lengths are not achievable. It may require increasing drilling fluid
density or to use fracture using another means like changing fracture azimuth or horizontal
completions (Lehman et al, 2002).
Practical Model Calibration (Weijers et al, 2005)
An example of a model calibrated with field measurements is illustrated in figs. 3.66 and 3.67. In fig
3.66, a plot of net pressure versus time is shown from a propped fracture treatment. The high initial
net pressure during the breakdown (at 180 min) was matched to significant tip effects. To match
the increasing net pressure, seen throughout most of the fracture treatment, fracture growth
complexity was assumed. In this case, the closure stress contrast of the barrier-payzone was high
but still lower than the 2000 psi of net pressure observed at the peak of the net pressure plot (see
103

fig 3.66). The closure stress contrast is the difference in stress between the pay zone and the
surrounding barriers. The higher the closure stress contrast, the higher the net pressure required
for out-of-zone fracture growth. As shown in the top section of fig. 3.67 in the width profile,
significant out-of-zone fracture growth still occurred later in this case. A fracture half-length and
total fracture height of 250 ft was obtained using the FracPRO PT 3D simulator.
Results from microseismic mapping, however showed a fracture height of only 130 ft and a
fracture half-length of 700ft which differs greatly from the un-calibrated pressure matching
aforementioned results. The net pressure history matching had to be carried out again to account
for directly observed geometry. To achieve the directly mapped results, it was determined that a
closure stress gradient in excess of 1 psi/ft had to exist in the shales surrounding the pay zone.
Since this wasnt practical, the composite layering effect was added to the model to account for the
observed level of material confinement. The immediate result of this confinement layer was an
increase in net pressure, hence both tip effects and fracture complexity had to be reduced to
maintain a net pressure match. Also due to longer fracture length (700 ft) the fracture surface area
in the pay zone increased considerably, so the reservoir permeability and wallbuilding coefficient
had to be decreased in order to maintain a match of leakoff characteristics. The bottom portion of
fig. 367 shows the final fracture geometry of the calibrated model. The resulting calibrated model
can now be run in predictive mode to evaluate alternative designs (Weijers et al, 2005).

104

CHAPTER 4 FRACTURING MATERIALS

Hydraulic fracturing is carried out using two broad classes of fracturing materials: fracturing fluid
and proppants. Fracturing fluid is a generic term, that involves both the base fluid (water, oil, acid
etc.) and additives. Additives are chemicals added to influence the overall properties of the
fracturing fluid. Propping agents or proppants are materials used to stop the fracture from closing
after treatment. They effectively prop or hold the fracture open to enable hydrocarbon recovery.
The comprehensive design and selection of these fracturing materials is paramount to successfully
achieve desired fracturing objectives. A lot of factors are considered in the choice of fracturing
fluid, additives and propping agents. The process of selection though is a subjective process: factors
considered range from job experiences, formation evaluation, laboratory test results and so forth.
In this chapter, the function of conventional fracturing fluids and additives are outlined; and
selection approaches for both fracturing fluids and proppants is discussed.
4.1 Fracturing Fluids
The effectiveness of a hydraulic fracturing operation is controlled by several variables, but only a
few are easily controlled, these are: the fracturing fluid properties, the injection rate and the
quality of propping agents (Pye and Smith, 1973). The fracture fluid design is an essential part of
the hydraulic fracturing stimulation treatment. Fracture fluids can be classified into four main
divisions based on their fluid bases: oil-based, water-based, foam-based and alcohol-based, see
table 4.1 below (Xiong et al, 1996). For decades, oil based fluids were in use and skepticism
followed water-based fluids. It was thought that pumping water into water sensitive formations (oil
reservoirs) would obstruct oil flow. However success of water based fluid experiments led to their
eventual wide usage with two-thirds of wells now being fractured with water based fluids (Howard
et al, 1970). Water base fluids have considerable advantage over other bases, including: non105

inflammability; higher specific gravity thus lower hydraulic horse power (HHP) for treatment; low
viscosity making it easier to pump, finally low cost and ease of availability.
This has found even more application with gas prospecting, as the fluids currently being used for
hydraulic fracture treatments in the Marcellus shale are water based or mixed slick water fracturing
fluids (Arthur, 2008). Slick water fracturing fluids consist mainly of water mixed with friction
reducing additives like potassium chloride. As reported by Robert Kennedy et al (2012) High rate
slick water fracturing can induce tensile fractures as well as shear existing fractures in the brittle
shale formation with low horizontal stress anisotropy. Slick water fracturing has become the norm
in Barnett and Marcellus shale plays. The disadvantage of slickwater fracturing fluids is that due to
its low viscosity, it is not an efficient carrier of proppants. Hence, though successful in several US
shale developments, it has not been suitable for all cases (R kennedy, 2012). In cases where
slickwater is inadequate, hybrid fracs have been proposed. A hybrid frac is a combination of
slickwater (to create the fractures) and another more viscous fracturing fluid (solely for proppant
transport) (King, 2010). Careful selection of fluids and proppants (sand constituents) is necessary
based on the reservoir properties of the specific shale formation. As reported by R. Kennedy, No
two shales are alike (King, 2010). The table 4.2 outlines suggested fracture treatment types and
pump rates for dry gas, wet gas and oil.
4.2 Additives
Additives are chemicals added to the fracturing fluid to achieve specific target properties of the
fracturing fluid. They constitute between 0.1 to 0.5 % of the total fracture fluid volume as shown in
Fig. 4.1. The figure shows volumes of additives in a hypothetical 2,500,000 gallon fracture
treatment, a size typically used in Marcellus shale horizontal well development, (Arthur et al, 2008).

106

General properties that should be possessed by a fracturing fluid include: low leak off rate, the
capacity to transport a propping agent and low pumping friction loss. Low leak off rate allows the
fluid to create the fracture and influences the extent of the fracture area (Howard et al, 1970). To
achieve low leak-off rate, fluid-loss additives are used. Capacity for proppant transport is influenced
by density, viscosity and velocity of fluid flow. The viscosity is the most critical parameter in
proppant transport as would be explained in later sections. Additives called viscofiers are used to
enhance viscosity.
The type of additive used, also depends on the base fracturing fluid. For instance, water base fluid
more than other base fracturing fluids; require surfactants to reduce interfacial tension and
resistance to return flow after treatment. Conversely, additives for friction-loss reduction in
fracturing are less needed for water-base fluids since it naturally has a friction-reducing advantage.
The same considerations apply to other base fluids, with additives chosen to supplement their
inherent limitations.
In a broad sense, additives serve the following uses: 1. Enhance fracture creation 2. Reduce
formation damage. Additives that enhance fracture creation include viscofiers, temperature
stabilizers, pH-control agents and fluid loss control materials. Those that reduce formation damage
are gel breakers, biocides, surfactants, clay stabilizers and gases (P.C Harris, 1988).
4.2.1 Viscosifiers
As evident in the name, viscofiers are used to increase the viscosity of fracturing fluid. The most
popular of these fluid thickening agents is Guar gum (Howard et al, 1970). Chemical modification
of water base polymers obtained from guar has produced a wide range of derivatives with useful
properties (shown in Table 4.1) such as: hydroxypropyl guar (HPG) and carboxymethyl
hydroxypropyl guar (CMHPG). Hydroxylethyl cellulose and carboxymethyl hydroxyethyl cellulose
107

(CMHEC) are derivatives from cellulose, another natural source. These derivatives provide viscosity
for fracturing wells with formation temperatures from 60 to more than 400 degrees Fahrenheit. To
prevent loss of viscosity due to decomposition of these derivatives at high temperatures (above
225 degrees Fahrenheit) , chemical stabilizers like methanol or thiosulphate are added. For lower
temperatures (below 150 degrees Fahrenheit) , aqueous solutions of these derivatives, known as
base gels are used. (P.C Harris, 1988). When high viscosity is required, cross linking a polymer is
normally done using transition metal cations which is more cost effective than merely increasing
the concentration of the polymer solution.
R. Kennedy, 2012 asserts that Low viscosity slickwater fluids generate fractures of lesser width and
therefore greater fracture length, theoretically increasing the complexity of the created fracture
network for a better reservoir to well bore connectivity As stated previously though, slickwater is
not an efficient proppant carrier due to its low viscosity. There is a tendency of proppants settling,
which can be overcome by increasing the pump rate to increase flow velocities. Hybrid fracs can
also be used, combining both slickwater and crosslinked fluid systems (highly viscous) to increase
proppant transport capacity (Brannon and Bell, 2011).
4.2.2 Surfactants and Alcohol
This is another group of commonly used additives in a water-based fracturing fluid, known as
surface active agents or surfactants. There are several uses of surfactants. These include: 1.
Reducing interfacial tension and hence capillary pressure 2. Providing a foam stabilizing action 3.
Reducing compatibility problems between fracturing fluids and reservoir fluids.
Decreasing the capillary pressure is useful in low permeability formations to reduce the pressure
needed in causing flow back of fracturing fluid since less fluid will be retained in the pore spaces of

108

the reservoir. In fracturing fluid recovery, the addition of gases like nitrogen and CO2 are also useful
(P.C Harris, 1988). The foam stabilizing action is effective in gas wells.
Investigators like McLeod and Coulter (1966) presented cases for the use of alcohol to lower
interfacial tension. Its use in fracturing however has been limited because it is cost prohibitive.
Another drawback to alcohol usage is that it impedes fluid-loss control and attainment of desired
viscosity (Howard and Fast, 1970)
4.2.3 Fluid-Loss Additives
These materials are added to the base fracturing fluid to keep the fluid within the fracture, hence
stopping it from leaking off into the rock matrix. This is of utmost importance, because loss of fluid
through uncontrolled leak-off will cause an increase in proppant concentration around the well
bore, which if allowed can create a proppant bridge and completely stop fracture propagation
(P.C Harris, 1988). Fluid-loss control in fracturing operations cannot be over-emphasized. Low fluid
loss (low leak-off rate) would mean larger and deeper fractures for a given volume of fracturing
fluid and injection rate.
Hawsey and Jacocks (1961) describe the way the fluid loss additive functions, as follows: The fluidloss additive which is largely insoluble, disperses into micron-size (m) particles when added to the
fracturing fluid. As fracturing takes place, some fluid is lost immediately to the formation, called
spurt loss. After spurt loss, the fluid-loss additive deposits a thin film on the face (interior walls)
of the newly created fracture, hence preventing further fluid-loss to the formation. It continues this
as fracture propagates and new fracture area is exposed. This thin film is sometimes called a filter
cake. The thin film would remain in place, as long as there is pressure in the fracture. When
flowback begins, it re-disperses and flows out through the large gaps between propping agents. The
spurt loss is an important part of the fracturing operation, since a good fluid-loss additive would not
109

only coat the fracture face but prevent excessive spurt loss from occurring. Spurt loss is defined as
the fluid loss per area, before the formation of a filter cake, and is very significant in naturally
fractured reservoirs (Jones and Britt, 2009). It is directly proportional to reservoir permeability.
In their 1961 paper on the use of fluid-loss additives in hydraulic fracturing of oil and gas wells,
Hawsey and Jacocks (1961) present a test that demonstrates fluid-loss properties of a fracturing
fluid. It was developed by an API sub-committee for measuring fracturing fluid efficiency. A
pressure differential of 1000 psig was maintained in two different cells at a temperature of 125F.
The results of the test are recorded as fluid lost (in cubic centimeters) through a filter (core plug) in
a 25 minute interval. Lease oil, an oil-base fracturing fluid was used in the two cells. The lease oil in
cell 1 had fluid loss additives, but in cell 2 lease oil was used without any additives. Using the oilbase fracturing fluid, lease oil without additives signifies high fluid-loss to the filter medium (acting
as formation) and lease oil with the additives is meant to depict low fluid-loss. The additive used in
this experiment, is known commercially as Adomite Mark II. Natural sand was added to the
fracturing fluid to serve as a proppant. The schematic diagram of the experiment is shown in Fig
4.2.
A plot of fluid loss against square root of time for the cell with fluid-loss control (additive) is shown
in Fig 4.3. The spurt loss, which is the amount of fluid lost before fluid-loss control is obtained by
extrapolating the fluid-loss line back to the y axis i.e. when time = 0. In this case, the slope, m = 0.5
and spurt volume, Vsp = 4.0cc. The fluid-loss coefficient, C can be obtained as follows:
C = 16.4 (m/a)

.(4.1)

Where, a is the area of the filter(sq cm), and m is the slope of the fluid-loss curve. It gives a
measurement of the fracturing fluid efficiency after spurt loss has ended. Thus, the lower the fluidloss coefficient the higher the fluid efficiency. It should be noted, that this method of obtaining the
110

fluid-loss coefficient, C was explained in chapter 3, under section 3.3.1 Mechanism 3 The effect
of wall building fracture fluids
In this example, the filter area was 20.2 sq cm and a fluid-loss coefficient of 0.4 is obtained. In Fig
4.2d, a plot of fluid-loss coefficient versus fracture area and injection rate for a fluid volume of
15,000 gal is shown. Refer back to equation 3.20 to see governing equation of plot. A fluid loss
coefficient of 5.0 (typical value without additive) would give a theoretical fracture area of 30,000 sq
ft. If a fluid-loss additive is used and the fluid loss coefficient is reduced to 0.4, as shown in this
case, then for an injection rate of 25 bbl/min, the theoretical fracture area increases to 160,000 sq
ft . The fracture area has been increased over five times. However, this theoretical area has to be
corrected for spurt loss. Fig 4.5 shows a nomograph for this purpose. In this example, for a fluidloss coefficient of 0.4, injection rate of 25 bbl/min, spurt loss of 4.0 cc, injection volume of 15,000
gal we obtain a factor of 73 percent from the nomograph. Hence, instead of a theoretical area of
160,000 ft, we have a spurt loss-corrected area of (0.73x160,000) = 116,800ft (Hawsey and Jacocks,
1961).
Despite the usefulness of fluid-loss additives, several investigators have shown that they severely
reduce fracture proppant conductivity and formation permeability hence affecting well
productivity. Pye and Smith (1973), conduct formation damage test using fluid-loss additives. They
conclude that at FLA/sand ratios exceeding 0.03, their model shows reduction in conductivity of
proppants, severe enough to make the fracturing treatment ineffective.
4.2.4 Clay Stabilizers
These are additives used to improve compatibility between the formation and fracturing fluid. Most
formations contain clay minerals that are susceptible to swelling and migration. Clay damage is
extremely important in low permeability, low pressure reservoirs as it affects capillary pressures
111

(Anderson et al, 2006). Fracturing fluids must provide a high electrovalent strength, so that clays
contacted would not experience ionic shock. Clay stabilizers like inorganic salts such as KCl, NaCl,
NH4Cl, or CaCl2 are used to prevent shocking the clays. Other stabilizers such as polymeric clay
stabilizers can attach anions on the clay surface to control migration of fines (P.C Harris, 1988). This
is useful, as invasion of fines hampers proper proppant placement.
4.2.5 Gel Breakers
These additives are useful for flowback and cleanup after the fracturing operation. It is important to
ensure good fracture conductivity. Gel breakers oxidize the backbone of the polymer molecules,
allowing the polymer to be produced out of the fracture. Enzyme breakers such as hemicellulose
are used at temperatures below 120F with a pH less than 8.5. Other oxidizing breakers like
ammonium and sodium persulfate are used at higher temperatures, 150F, or at lower
temperatures with an activator (Jones and Britt, 2009).
4.2.6 Bactericides/Biocides
Biocides are additives for controlling bacteria growth and are often a necessity for water-base
fluids. Investigators like Hawsey et al (1964) have shown that pumping untreated water into a
reservoir can trigger bacteria growth (Howard and Fast, 1970). Aerobic bacteria can destroy the
viscosity of a fracturing fluid within a few hours. If anaerobic bacteria is introduced by a fracturing
fluid, it can produce hydrogen sulphide (H2S) within the reservoir (P.C Harris, 2008). Biocides used
to control both types of bacteria include quarternary amines, amide-type chemicals and chlorinated
phenols (Howard and Fast, 1970).

112

4.2.7 pH Control
Fluid pH affects various properties of the fracturing fluid. These include: initial polymer gelation
rate, crosslinking characteristics, gel break properties, bacteria control, viscosity stability and other
properties. The typical pH range for fracturing fluid is from 3 to 10 (virtually the entire pH range).
Buffers made from mixing weak acids with weak bases are used to maintain the desired pH (P.C
Harris, 1988).
Jack and Britt (2009), group the selection criteria for fracturing fluid into the following:
1. Safety and environmental compatibility.
2. Compatibility with formation and additives.
3. Simple preparation and quality control.
4. Low pumping pressure.
5. Appropriate viscosity.
6. Low fluid loss.
7. Flowback and cleanup (for high conductivity).
8. Economics.
Hongje Xiong et al (1996) propose an approach to select fracturing fluids and additives for fracture
treatments using fuzzy logic. This is contained in the appendix.
4.3 Proppants
Proppants are used to prevent closure of the created fractures, in order to enable the flow of
hydrocarbons from the reservoir to the wellbore (see Fig 4.6). The goal of proppants is to maximize
fracture conductivity (i.e. flow path for hydrocarbons), thus the magnitude of fracture conductivity
is a measure of proppant performance. Proppants are necessary because of the tendency of
113

fractures to heal (close) naturally after fracture creation. There are several types of proppants
available commercially, including: sand, ceramic (glass beads), resin coated ceramic, aluminium
alloys, nutshells, plastics (see Fig 4.7). These proppants are differentiated mainly by their specific
gravity and strength. The cost of ceramic proppants range from 5 to 10 times that of sand (Jones
and Britt, 2009). In this section, the key parameters that guide proppant design and selection are
described.
4.3.1 Proppant size and concentration
The proppant size is an important consideration for design and depends on the degree of stress,
target conductivity, and achievable fracture width. The testing of proppant size distribution is a
quality control procedure done through sieve analysis. The American Petroleum Institute (API) has
two publications showing tests for sands, intermediate strength and high strength proppants: RP 56
1983 and RP 60 1989. According to API recommendations, for proppant size classification, more
than 90% of the tested sample should fall between the designated sieve size, like 12/20, 20/40,
30/50 etc. Not more than 1% should be smaller than the smallest sieve size and less than 0.1%
should be larger than the largest sieve mesh. Generally, fines are unacceptable as they reduce
fracture conductivity, the maximum tolerable fines for proppants is 1% (i.e percentage that passes
the BS #200 sieve) (Jones and Britt, 2009). The percentages given by Jones and Britt do not add up.
It is recommended that classification of proppants can be done simply when 90% of the tested
sample falls between the designated sieve sizes and fines tolerance limit should be observed.
The size and concentration of the proppants influences proppant placement in several ways
(Cohen et al, 2013):
1. Larger proppants settle closer to the wellbore, due to their higher settling velocity (see Fig
4.8).
114

2.

Proppant bridging occurs more easily in large proppants.

3. The higher the proppant concentration the more the apparent viscosity of the slurry,
increasing the average width of the supported fracture and reducing the length.
4. Smaller proppants are transported a further distance and increase the chance of tip
screenout.
Table 4.3 shows the effect of different proppant sizes on fracture geometry and fracture properties.
The smallest proppant (80/100 mesh sand) maximizes the fracture length and consequently the
Effective Stimulated Volume (ESV). While the largest proppant (20/40 mesh sand) banks (or
bridges) in the vicinity of the wellbore (point 0) and maximizes the average propped conductivity
(Cohen et. al, 2013). From this, it is seen that the largest proppant would be the preferred option if
it can be transported far enough, since it would increase fracture height and hence conductivity.
Cohen et al, (2013) also show a graph comparing production rate for fractures stimulated with
proppants of different sizes (Fig 4.9). Initial production increases with proppant size but the
production decline is faster. While for the smaller proppant size, the initial production is low but
production decline is slower. Cohen et al, (2013) explain that the initial production is dependent on
the pressure differential around the borehole caused by the conductivity. This is why it favors larger
proppant sizes. In terms of production duration, the production rate of the larger proppant size
depends only on the formation matrix permeability thus it declines faster. While that of the smaller
proppant depends on both the formation matrix permeability and the conductivity of the fracture
network, hence the longer it takes to decline.
Cohen et al, 2003 propose a method of gradually increasing the proppant size during injection for a
single treatment. The pumping schedule for the combined-size treatment is shown in table 4.4. A

115

cumulative production plot comparing this combined-size treatment with single-size treatment is
shown in Fig 4.10.
Proppant concentration is achieved by using pelletized spacer materials. The spacers should have
the following qualities:
1. Similar specific gravity as the proppant
2. Easily transportable
3. Essentially insoluble in fracturing fluid, but soluble to post-fracturing injected solvents for
easy removal
4. Resistant to breakage during pumping and ease of storage and handling in the field.
Three examples of spacers are: Urea (NH2CONH2), Hydrocarbon resin (soluble in naphthalenic and
aliphatic hydrocarbons) and Sodium bisulfate (Howard and Fast, 1970).
4.3.2 Proppant Shape
This consists of two main descriptions: roundness and sphericity. The roundness is a measure of the
smoothness of the proppant, while the sphericity is how well it resembles a sphere (Jones and Britt,
2009). Krumbein and Sloss (1963) developed a chart of sphericity and roundness which is the most
widely accepted till date (See Fig 4.11). API RP 60 working with the Krumbein and Sloss chart
recommend that sand should have a minimum roundness and sphericity of 0.6, while ceramic
proppants should have a minimum value of 0.7 (see Fig 4.12). At high stresses, say 4000psi the
more spherical the proppant is, the more the permeability but at lower stresses, the more angular
the proppant is the higher the permeability. Angular proppants tend to crush under high stresses,
thus generating fines which lead to a reduction in conductivity (Jones and Britt, 2009).

116

4.3.3 Proppant Stress

The stress in which the proppant would be subjected, is a critical factor to consider when selecting
propping agents. The proppants have to be chosen, so that they do not crush under field closure
stresses. Over the years, the traditional way of calculating closure stress, has been the minimum
horizontal stress minus the bottom hole flowing pressure (measured at the start of production).
The following sets of equations provide the theoretical framework for calculating the proppant
stress.
The effective stress, is given by
=

(4.2)

Where is the overburden pressure (normally estimated as 1 psi/ft), is the reservoir (fluid)

pressure, is the Biot-Willis poro-elastic contant (which is usually taken as 1)

The flowing bottom hole pressure, is the pressure at the bottom of the wellbore when
production starts, it is given as:
=

(4.3)

Where ISIP is the instantaneous shut-in pressure, is the minimum horizontal stress

Theoretically, the minimum horizontal stress is given as follows:

=

(
1

) + +

(4.4)

Where is the Poissons ratio, is the tectonic stress (if known),

The traditional method of calculating proppant stress, is using:

117

(Effective stress = Total stress pore pressure)

.(4.5)

Jones and Britt, (2009) present a modified Eatons equation for theoretically calculating proppant
stress as follows:
= +

.(4.6)

Where is the final net pressure.

In this equation, the inclusion of the final net treating pressure ( ) is done in order to
emphasize the reduction of the flowing bottom whole pressure ( ) as well drawdown takes

place (see schematic Fig 4.13). This view is shared by Sookprasong (2010), whose proposed
proppant stress equation however uses reservoir pressure, instead of ( ) as shown above.

Practically, the minimum horizontal stress is obtained from the pressure decline characteristics of
the LOTs and XLOTs described in Chapter 3. From the equation above, Jones and Britt (2009)
highlight the following:
1. As the reservoir pressure reduces, the stress on the proppant decreases.
2. As well drawdown occurs, reduces at the wellbore and the stress on the proppant
will increase

The pressure in the fracture is higher with increasing distance from the wellbore. The maximum
stress a proppant would be subjected to would normally take place at the initial stages of
production, during or immediately after fracturing fluid clean-up when hydrocarbon starts to flow,
especially in gas wells. Quality assurance is sometimes done using proppant crushing tests (API RP
56 1983) (Jones and Britt, 2009).

118

4.3.4 Proppant Embedment Pressure

Embedment pressure is defined by Rixe et. al (1963) as a measure of the maximum pressure
required to embed a steel ball to a given depth in rock. This gives a direct indication of the
resistance of the formation to embedment by a propping material and the effect of the rock on
proppant deformation. It is a kind of indentation test and is a measure of the rock strength. A test
procedure for determining the embedment pressure is described by Howard and Fast (1970). A
steel ball point 0.05 inches in diameter (which simulates the proppant) is attached to the upper
platen of a hydraulic testing machine which moves, and loads the rock specimen (3.5 inches
diameter) hydraulically (see Fig 4.14). Where possible, the rock core should be obtained from the
well to be hydraulically fractured. The steel ball point is embedded to a depth of 0.0125 in and a
strain recorder is used to observe the results. The load at the target embedment, is recorded
and at least two more indentations are made on the test specimen, about 0.5 inches apart. Finally,
the diameter of the indentation, is measured under a microscope and the embedment pressure
is calculated as:

Embedment pressure (psi) =

4
2

...(4.7)

The embedment pressures for 22 formations are shown in table 4.5. McGlothin and Huitt (1959)
present another method of calculating rock embedment strength. They suggest crushing proppants
between plates with similar penetration hardness (Howard and Fast, 1970).
An obvious drawback in the test described by Howard and Fast, is the inherent assumption that the
steel ball point has the same physical characteristics as the proppant. Steel varies significantly from
sand, ceramic, nutshells e.t.c.

119

4.3.5 Fracture capacity (conductivity)

As earlier stated, the fracture conductivity is a measure of proppant performance. Proppant
selection is deemed successful only when it can achieve substantial fracture conductivity. The
fracture flow capacity (conductivity) depends on the fracture width, proppant distribution and
proppant concentration. The post-fracture width is controlled by proppant size used for
stimulation, while proppant concentration is controlled by spacers. Proppant distribution is not an
easily controlled parameter (Howard and Fast, 1970).
Rixe et al (1963), describe a test procedure for determining fracture conductivity of a given
proppant. The proppant is placed between two rock cores and subjected to similar overburden
pressure and temperature. Pictures of proppants in conductivity cells are shown in Fig 4.7. The rock
cores (3.5 inches diameter and 2 inches long) are mounted in steel cups with a low temperature
melting point alloy, in such a way that 0.25 inches of the smooth core face extends above the top
edge of the cup. This is shown pictorially in Fig 4.15 and schematically in fig 4.16. A hole is drilled
axially in the upper half of the core to intersect a shallow hole in the center of the lower half, as
shown in the aforementioned figures. The overburden stress is simulated using a hydraulic ram, see
setup shown in Fig 4.17 and Fig 4.18. The temperature is controlled by placing the rock cores in a
heated box. Fracture capacity is determined by allowing nitrogen gas to flow from the hole in the
upper half of the core through the simulated propped fracture (Howard and Fast, 1970). The flow
capacity is calculated based on Darcys law and considering radial gas flow. To take into account the
time effect, the test is carried out for a 30-day period with flow rates recorded at 7-day intervals. It
has been discovered that 30 days is the time required for the fracture to stabilize and hence, for
long-term fracture capacity to be stimulated (Fixe et al, 1963). The use of the conductivity and
embedment pressure for proppant selection is described in section 4.3.6

120

Although the test described above has been used over the years for successful fracturing
treatments, it has certain drawbacks. Firstly, the field conditions of proppants are not dry as
assumed by the test. The drawbacks of the test are made more evident by considering several
other factors that affect the effectiveness of fracture conductivity:
1. Production and migration of proppant fines
2. Proppant flowback
3. Proppant embedment
4. Multiphase flow
5. Non-Darcy flow considerations
Production and migration of fines
Fines are the little particles that break off the surface of proppants as they are subjected to closure
stress. These fines have a significant impact on reduction of fracture conductivity. Several
investigators like Coulter et al (1972) show that the presence of just 5% fines can reduce
conductivity by as much as 62%. These results have been corroborated by Lacy et al (1997) showing
a 54% reduction with 5% fines present. The decrease in conductivity is made worse when the fines
migrate to the wellbore. Standard crushing tests like API RP 56, API RP 58, API RP 60 and ISO 135031 make use of dry proppants subjected to the closure stress for only 2 minutes which doesnt
simulate the field wet, hot condition. Modified crush test procedures have been proposed by
Freeman et al (2009) and Diep (2009) to better simulate field conditions (see Fig. 4.19 and Fig 4.20).
In Fig 4.20, the resin coating prevents the migration of fines in the curable resin coated sand (CRCS)
proppant (Terracina et al, 2010).

121

Proppant flowback
Proppant flowback is the movement (flow) of proppants back to the wellbore (see Fig 4.21). The
higher the pump velocity, the more the chance of flowback occurring. Proppant flowback and pack
re-arrangement is the main cause of well production decline, equipment damage and well lockdown for repairs. It reduces conductivity at the wellbore and decreases connectvitity to the
reservoir (Terracina et al, 2010). The use of resin coated proppants has helped to reduce the
flowback tendency of proppants, this is shown in the highly magnified SEM photograph of a CRCS
grain-to grain bonding in Fig 4.22.
Proppant Embedment
This causes a loss in conductivity due to the interaction between the formation and the proppant at
the face of the fracture. It is considered a problem in partially consolidated to unconsolidated
formations, but it also occurs a little in hard rock (Jones and Britt, 2009). Fig 4.23 demonstrates how
proppant embedment can reduce post-fracture width and hence conductivity. As shown, more
than 1/3 of the grain diameter is lost due to embedment. Another result of embedment is the
creation of formation fines through spalling (see Fig. 4.24), which can then migrate and further
reduce conductivity (Terracina et. al, 2010). The advantage of embedment is that it would help
prevent flowback.
Non-Darcy Flow considerations
When the flow is high (turbulent), Darcys equation becomes inadequate. Non-Darcy flow reduces
the permeability of the proppant pack (Jones and Britt, 2009).

122

4.3.6 Proppant selection - Method of Rixe et al(1963)

Rixe et al (1963) proposed the following procedure for proppant selection.
Step 1: Determination of the fracture flow capacity needed to achieve desired level of well
production. The required fracture flow capacity depends on the formation flow capacity and the
fracture penetration into the reservoir. Fig 4.25 depicts the relationship between the ratio of
fracture to formation capacity (also called contrast) to the increase in production ratio, while
considering fracture penetration. These plots are based on data that assume Darcy flow of
homogeneous fluid through the formation. From the plot it is seen that the higher the formation
flow capacity, then relatively higher fracture flow capacity is required to achieve maximum
production (i.e high fracture to formation flow contrast). Quantitatively, a contrast of 10 and a
fracture penetration of 30% a 4-fold increase in production is expected, but if the contrast is
increased to 100, it would yield a 6-fold increase in production. Also, from the plot it is observed
that the deeper the fractures, the more significant the contrast effect on production increase.
Step 2: Determination of embedment pressure. This could be done as described in section 4.3.4.
Step 3: Selection of type of propping agent required by using charts such as the one shown in Fig
4.26. This chart was prepared from correlating results from embedment pressure tests and fracture
capacity tests on different propping agents and various formations. The maximum fracture capacity
for a given embedment pressure and propping agent must lie in the region below the curve of the
particular propping agent.
These curves vary with depth in which proppant performance was evaluated. For the figure shown,
it was at 7000 ft.

123

Step 4: Selection of the size and concentration of the propping agents from the embedment
pressure, using charts in Fig 4.27 to Fig 4.29.
Example Calculation
Well depth . 7000ft
Formation .. Tensleep, Wyoming
Embedment Pressure (table 4.5) 196,600psi
Planned increase in well productivity. 6-fold
Formation capacity . 160 md-ft
Planned fracture penetration. 30 percent of drainage radius
1. From Fig 4.25, it is seen that to effect a 6-fold increase for a 30% fracture penetration, then a
fracture to formation contrast of 50 is required.
2. Hence, the fracture capacity required = 160 X 50 = 8000md-ft.
3. Plotting the fracture capacity and embedment pressure on Fig 4.26, shows that it is above the
curve for sand, hence sand is not suitable. However it is below, the nut shells straight line plot, so it
is adequate.
4. Using the plot in Fig 4.28 for rounded nut shells, we conclude the selection as follows: A fracture
capacity of 8000md-ft can be achieved in the Tensleep, Wyoming formation with an embedment
pressure of 196,600 psi using a -12 to 20 mesh rounded nutshell at 0.024 Ib/ft concentration.
Notice that in Fig 4.27, the chart for natural sand. At low embedment pressure (50 psi) the smaller
size sand proppants, type C (-20 to 40 mesh) clearly has a lower fracturing capactity than the larger
124

size proppants type B (-10 to 20). However as the embedment pressure increases to about 200 to
250 thousand psi, both curves B and C gradually come together almost intersecting. This means
that at high embedment pressure, the fracture capacity (conductivity) of both large and small size
proppants is nearly equal. This does not occur in other materials like nutshells and aluminium (see
Fig 4.28 and 4.29) which all have fracture capacity proportional to proppant size. Thus, it is most
likely caused by a material property of natural sand. It is proposed here that it occurs as a result of
the tendency of large sand particles to crush at high pressures. Another possibility could be that at
high pressures, there is greater proppant embedment in large size proppants (see fig 4.23), hence,
their fracture conductivity reduces bringing it closer to conductivity values of small size proppants.
4.4 Case study example of typical Hydraulic Fracturing Stimulation (Fontaine et al, 2008)
4.4.1 - Introduction
This section is on the stimulation treatment of a Marcellus well in Potter county, Pennsylvania
operated by Guardian Exploration LLC. It was presented in a paper by Fontaine et al (2008). This
Case study is chosen because it describes in detail how a typical fracturing design is routinely
adjusted to fit observed field conditions. Also, the treatment was done by using ideas that were
successful in other shale plays like Barnett shale of Fort Worth basin and Fayetteville of Arkansas.
This is however rarely encountered as shales differ greatly with location. The well log is shown in
Fig 4.30. The Marcellus was at a depth of 5220 to 5310 ft. The well casing had a 4-4.5 inch outside
diameter, weighing 11ppf and rated at 6700 psi working pressure. With regards to stratigraphy, the
Marcellus shale was underlain by the Onondanga lime which proved to be a good lower barrier to
fracture growth but there was no significant upper barrier.

125

4.4.2 - Preliminary slickwater fracture design

The preliminary fracture design is shown in table 4.6. The design consisted of pumping 545,000
gallons of 5 PPT (parts per trillion) slickwater base fluid at a rate of 50BPM (barrels per minute).
100,000 pounds of 80/100 mesh proppant, 350,000 pounds of 30/50 mesh proppant and 137,500
pounds of 20/40 mesh proppant were to be placed at concentrations ranging from 0.5 to 3 PPA
(pounds proppant added). Well bore sweeps were pumped at intervals of proppant usage, in order
to reduce near wellbore proppant settlement. Drilling fluid sweeps are fluids of adequate density
and viscosity, used to clean wellbore of cuttings (in this case remnant proppants). They erode the
top of proppant dunes that are deposited near the wellbore and prevent them from building up.
Additives used include the following: friction reducing polyacrylamide, bacteriacides and a nonfoaming microemulsion surfactant. To achieve 50 BPM, a 6500 HHP (hydraulic horsepower) was
used, maintaining a pressure gradient of about 1 psi/ft.
4.4.3 - Starting treatment
At the start of treatment, the well perforation was done at the lowest portion of the organic rich
section of the formation. Acid was not used initially, to reduce the risk of formation damage. An
initial downhole rate of approximately 25 BPM with a corresponding surface treating pressure of
over 5000 psi as seen in Fig 4.31. After about 45 minutes, it was observed that the bottom hole
pressure (BHP) had increased to about 8000 psi, thus suggesting a pressure gradient of 1.5 psi/ft
which was too high. This showed that the perforations were not transmitting the fluid, thus
pumping was stopped. Then spot acid was used over the perforations to breakdown the formation
further. A steady downhole rate of 42 BPM and 6200-6300 psi BHP was later achieved after acid
treatment. Proppant slugs (80/100 mesh sand at 0.1 ppa concentration) were pumped to further
reduce the treating pressure to 5500 psi (since 6400 was maximum allowable) and stabilize the
126

downhole rate to the target 50 BPM (see Fig 4.32). It was believed that the initial high bottom hole
pressure was because of high perforation friction and the possible existence of multiple competing
natural

fractures,

thus

causing

near

well

bore

tortuosity

(see

Fig.4.33).

4.4.4 Proppant selection and placement

The slickwater base fluid used was a thin fluid that exhibited apparent viscosity only when pumped.
This thinness helped in reducing fluid friction, but had the drawback of limited proppant transport.
At 50BPM, its apparent viscosity was equivalent to 30 centipoise guar gel viscosity. The proppant
treatment involved the placement of 587,500 pounds of proppants with 545,000 gallons of fluid.
The percentage of proppant composition was 17% 80/100 mesh, 60% 30/50 mesh and 23% 20/40
mesh. As stated previously, to prevent build- up of proppants around the wellbore, well sweeps
were used (Fig 4.34). According to WoodWorth (2007), small volume sweeps such as the ones used
in this treatment are the most effective. The sweep highlighted in figure 4.34 is highlighted in
Figure 4.4f. Looking closely, we can see that the pressure was gradually increasing from 5000 psi to
5600 psi because of proppant build-up. The introduction of the sweeps helped to mitigate this
increase. The proppant placement started with the smaller size (80/100) proppants before finally
the largest size (20/40) as seen in the schedule of table 4.6. The effect of increasing proppant size
and concentration on the net pressure is shown in Fig 4.35.
4.4.5 - Finishing treatment
Fig 4.36 shows the pressure plot for finishing treatment. When proppant addition is stopped, there
is a loss of hydrostatic pressure in the wellbore that increases the surface treating pressure. This
pressure increase is 795 psi. It was ensured that the wellbore rating of 6700 psi was not exceeded
during the flush of the treatment. In Fig 4.37, the portion of the plot highlighted with the box
occurs immediately after shutting down the pumps. It is normally called water hammer and
127

shows that there is a good wellbore to formation existing, a rough measure of conductivity. This
was made possible because of previous steps to minimize near wellbore tortuosity using acid
treatment.

128

Chapter 5 CONCLUSION/RECOMMENDATIONS
Hydraulic fracturing is a very sophisticated operation and so far has been carried out via technical
expertise, sound experience and the rule-of-thumb approach. Fracture mechanics itself is a
complex field and require lots of mathematical approximations to estimate fracture geometry.
During formation testing, pressure transient tests are conducted to calculate formation
permeability, fracture conductivity and fracture half-length estimates. However, as stated in the
classical Lee and Holditch paper the permeability is not obtained directly from pressure tests even
with the modified Millheim Cinchowicz method. It is obtained as an estimate from other drawdown
tests. This trend prevails up till date with permeability estimates obtained as external data from
pressure tests. It is recommended that a unified system of testing be designed, were permeability
can be obtained simultaneously from pressure tests as well as the other core parameters:
conductivity and half-length.
Mini-fracs also known as Extended leak-off tests (XLOTs) have been shown to give better estimates
of the minimum horizontal stress than the leak-off tests (LOTs) because of the additional
pressurization cycles. Experience has shown that taking the value of the minimum horizontal stress
in the first cycle were the LOTs terminate is overly conservative. A better estimate is in the second
or third cycle, which is contained in the XLOTs. However due to the relative ease of conducting the
LOT, it is more popular in the industry. This practice does not give the best estimate of the
minimum horizontal stress and should be discouraged.
The major difference in the two variants of XLOTs, which are Pump-in/shut-in and Pumpin/flowback tests and the actual XLOTs is that in the shut-in and flowback tests the planned
fracturing fluid is not used. The pump-in/shut-in have been found to be useful to estimate closure
pressure but in shales which take too much time to close, the flowback test is preferred. However,
129

research by Raen show that the flowback test underestimates the closure pressure. It has been
established that the Step-rate test (SRT) which gives the fracture extension pressure is a good
upper bound for the closure pressure. Thus, further work can be done on development of standard
charts that could mark the upper bound (with SRTs) and lower bound (with flowback tests) of
closure pressure in order to bound the validity of the results from shut-in tests or XLOTs.
The early 2D models: PKN and KGD are useful simple approximations to determining the average
fracture width while assuming constant fracture height. To account for complexity in fracture
growth, the pseudo 3D models and fully 3D models have been proposed. The drawbacks of the 3D
model is its reliance on the accuracy of many input parameters in order to make useful predictions.
An error in just one parameter would significantly affect 3D model based simulation results. A
solution to this could be to build a database from results of past successful hydraulic fracturing
stimulation. From this database an operator would have a profound idea of shale characteristics at
depth in a region even before beginning the formation testing or fracture modeling. To create this
solution would require the co-operation of fracturing companies that have been in the business for
a long time.
It is a widely known fact that during actual stimulation, the observed net pressure is different from
the predicted bottom hole pressure (BHTP). Thus, effort has been made to use the pressure
observed as stimulation progresses to change model inputs, so as to have a better estimate of
fracture dimensions. This process termed net pressure matching is ubiquitous in the industry. There
has however been no effort to quantitatively explain the influence that changes in friction pressure
has on the observed net pressure. Further research could be done on this area, with a view to
bridging the distance between the net observed pressure and model predicted bottom hole
pressure. A standard guide for the complementary use of both the direct mapping techniques

130

(tiltmeter and microseismic mapping) and net pressure matching is needed. This would remove the
black box status that most oil and gas professionals have attached to this process.
Fracturing fluid and additive selection is a highly proprietary procedure. Some suggested methods
of standardizing the fluid selection process is in appendix B. Proppant selection has been an area of
interest to oil and gas industry professionals over the years. The theory behind proppant stress
determination is in section 4.3.3 of the thesis. It is recommended that the closure pressure should
be obtained from SRTs, which give upper bound estimates. If these are used for proppant stress
determination they would give values of proppant stress that are reasonably conservative.
Standard crushing tests like American Petroleum Institute (API) RP 56, API RP 58, API RP 60 and ISO
13503-1 make use of dry proppants subjected to the closure stress for only 2 minutes which
doesnt simulate the field wet, hot condition. Modified crush test procedures have been proposed
by Freeman et al (2009) and Diep (2009) to better simulate field conditions. These test procedures
should be improved, standardized and adopted by the API, in order to improve proppant selection.
Most fracture capacity tests make use of nitrogen gas. Natural gas is made up mainly of methane,
hydrogen, carbon dioxide, nitrogen, hydrogen sulphide and other impurities. More investigation
into the use of an adequate mixture of carbon dioxide and nitrogen for the embedment pressure
test is recommended (eliminating methane and other inflammable components). The aim would be
to better simulate the fluid flow of natural gas. This hybrid mixture would improve assessment of
fracture conductivity and make for better proppant selection.

131

TABLES
Table 1.1 Types of Completions and conditions for selection (After Azar and Samuel , 2007)

132

133

Table 1.2 - API designated classes of Portland cement (After Azar and Samuel , 2007)

134

Table 2.1 Core recovery types and application After Unalimser and Funk (1998), and Keelan
(1972)
Core recovery Description

Typical formation

types
Conventional

It is the most common method used. Small plug-type Relatively

(Plug) Analysis

samples are obtained at selected core intervals. homogeneous

Normally 1-1.5 inches in length.

Whole

formations

core A full diameter core is obtained (see Fig 2.20). It is Heterogeneous

(Full diameter useful in those formations that the small samples carbonates
or continuous)

(like in the plug method) are not representative.

samples

and
containing

fractures and vugs.

Horizontal

Conventional cores up to 90 ft in length are cut in

coring

horizontal wells using downhole motors.

Sidewall

Small cores drilled, punched or recovered from Soft sand regions (as

coring

projectiles fired into the wall of the drill hole. It found in the United
provides samples from specific locations of the States Gulf Coast)
logged well after hole has been drilled. The use of the
Rotary tool prevents damage to hard formations, that
can be cause using percussion sidewall coring.
Sidewall values have a large margin of error in hard,
low permeability formations and unconsolidated
sands

Improved

This method has evolved from normal use of only Unconsolidated

135

Rubber Sleeve rubber sleeve for recovery to a combination of non- formations

Core method

intrusive bits, full-closure core catchers and sleeve

inner barrels. Rock samples are obtained by: 1.
Punching when cores are completely unconsolidated
or 2. Initially freezing the rock prior to removal of the
confining sleeve, then core drilling with liquid
nitrogen.

Sponge Coring

It is an economical method for assessing residual oil

saturation and less expensive than pressure coring

Gel Coring

Displaceable formation specific gel is used in the core

barrel. This minimizes flushing during coring, fluid
migration during retrieval and fluid loss at the
surface. It also provides mechanical support for the
core.

Oriented Core

It provides orientation to core to determine direction

and degree of formation dip, tilt of fractures, and
directional permeability.

136

Table 2.2 Permeability of well samples, showing effect of confining pressure ( After AlMulthana et al. 2008)
Sample ID

Pressure > 200 psig

Pressure < 200 psig

Dry Weight (gm)

Permeability (md)

Dry Weight (gm)

Permeability (md)

18

182.20

367

182.33

581

19

183.59

771

183.70

985

22

198.89

179

198.97

258

25

211.75

0.16

211.75

0.79

26

208.24

0.09

208.21

0.86

27

205.25

0.09

205.21

0.84

178

187.44

1329

188.00

1476

179

186.37

1060

186.46

1201

180

183.60

1714

183.69

1915

Table 3.1 Plot types used for analysis, After Lee et al, 2004.

Purpose

Operation

Quantities

Scales

Comments

Overview,

Display

Pressure and

Linear-Linear

Both pressure

Quality

Volume vs

and volume are

Time

displayed as a
function of time

Fracture initiation

Pump-in, fluctuating

Pressure vs

Linear-Linear

Fracture extends

137

flowrate

volume

into the far-field

(decreasing

when pressure

pressure)

decreases while
still pumping

Fracture initiation

Pump-in,

Pressure vs

Constant flow rate

time

Linear-Linear

This plot assumes

pump rate is
constant

Fracture closure

Pump-in, flow-back

Pressure vs

Linear-Linear

volume

in non-permeable

A flow-back is
often required to

zone

ensure fracture
closure

Fracture closure

Pump-in, Shut-in

Pressure vs

Linear-Linear

Slope change is

in permeable

square root of

expected when

zone

time

fracture closes

Fracture closure

Pump-in, Shut-in

Pressure vs G

Linear-Linear

function

in permeable

This is most
appropriate for

zone

extremely long
pump-in times

Fracture closure
in permeable

Pump-in, Shut-in

Pressure vs

Log-Log

time

zone

In all the tests above, the plots show a change in system compliance as the fracture is opened or

138

closed, characterized as a change in slope in the pressure vs time or pressure vs volume data (Lee
et al 2004)

Table 3.2 Classification of pressure tests performed at the casing shoe: PITs (After Addis et al,
1998)

Table 3.3 Shut-in pressures, minimum and maximum stress gradients from ELOTs performed on
the North West Shelf, Australia (After Addis et al, 1998)

139

Table 3.4 Shut-in pressures, minimum and maximum stress gradients from ELOTs performed on
the Norwegian North Sea (After Addis et al, 1998)

140

Table 3.5 Oil reservoir Data for 3D Model Application (After Settari and Cleary, 1984)

141

Table 3.6 Pumping Data for 3D Model Application (After Settari and Cleary, 1984)

142

Table 3.7 Fluid Treatment Data for 3D Model Application (After Settari and Cleary, 1984)

Table 3.8 General net pressure matching parameters ( After Weijers et al, 2005)
Parameter

Mainly Affects

During

Formation Permeability

Slope decline

Breakdown injection

Wallbuilding Coefficient

Slope decline

Minifrac tests

Pressure-dependent leakoff

Slope decline

Prop frac

Fracture Complexity

Level

All injections

Stress Contrast (Pay-barrier)

Level

All injections

143

Tip Effects

Level

All injections

Proppant Drag

Level

All injections

Compliance change during tip Level

TSO

screen-out (TSO)
Composite Layering

Geometry

All injections

Width decoupling

Geometry

All injections

Table 4.1 Fracturing Fluids and General Conditions for Use (After Xiong et al, 1996).
Fluid Base

Fluid type

Main Composition

Water-based

Linear fluids

Gelled water, HPG, HEC, Short

CMHPG, CMHEC, etc

Oil-based

Generally Used For

fractures,

low

temperatures

Crosslinked

Crosslinker + HPG, HEC or Long

fractures,

fluids

CMHEC, etc

temperatures

Linear fluids

Oil, Gelled oil

Water

sensitive

high

formations,

long fractures
Water

external Emulsifier +Oil + Water

Good for fluid loss control

polyemulsion
Foam-based

Acid-based foam

Acid+foamer+ N2

Low pressures, water sensitive

formations

Water-based

Water + Foamer+CO2 or N2

Low pressure formations

foam

144

Alcohol

based Methanol + Foamer + N2

foam
Alcohol based

Low pressure formations with

water blocking problems

Linear system

Gelled water+ alcohol

Removal of water blocking

Crosslinked

Crosslinked system + alcohol

problems

system

Table 4.2- Suggested Fracture Treatment Types for Different Formations (After Kennedy et al,
2012)
FracFluid type

Formation

Pump rate

Slickwater/linear

Dry Gas or Low High 100+ BPM

gel

liquid

Conductivity

Play Application

Infinite to gas

Barnett,
Marcellus,
Fayettevillle

Hybrid Frac

Gas condensate

Low 60 to 80 More conductive Eagle, Ford, Utica

BPM

Cross linked Frac

Oil bearing

frac

Low 40 to 60 Highly conductive Bakka,

BPM

frac

Niobara,

Eagle, Ford

145

Table 4.3 Effect of different size proppant on a fracture geometry and fracture characteristics
(After Cohen et al, 2013)

146

Table 4.4 Pumping schedule with several proppant sizes (After Cohen et al, 2013)

147

Table 4.5 Embedment Pressure Values for various formations (After Howard and Fast, 1970)

148

Table 4.6 Preliminary slickwater fracture design (After Fontaine et al, 2008)
Stage Description

PPA (Ib/gal)

PAD
80/100

0.5

Sweep
80/100

0.75

Sweep
30/50

1.00

Sweep
30/50

1.50

Sweep
30/50

2.00

Sweep
30/50

2.50

Sweep
20/40

2.50

Sweep
20/40

3.00

Slurry rate (BPM)

Clean volume (gal)

50

100000

50

80000

50

5000

50

80000

50

5000

50

50000

50

5000

50

50000

50

5000

50

50000

50

5000

50

50000

50

5000

50

25000

50

5000

50

25000

149

FIGURES
Figure 1.1 Shale distribution in the United States ( After Arthur et al ,2009)

150

Figure 1.2 Process Flow Diagram for a Hydraulic Fracture Treatment (After Arthur et al, 2009)

151

Figure 1.3 Water storage frac tanks - After Arthur et al (2009)

152

Figure 1.4 Picture of frac head (After Travelin Terriers, 2013 www.travelinterriers.com)

153

Figure 1.5 Hydraulic fracturing site (After Hibbeler and Rae ,2005)

154

Figure 1.6 Double well casing to protect groundwater sources (After Marathon Oil Corp, 2012)

Figure 1.7 Horizontal Drilling in Shale Strata (After King et al, 2010)

155

Figure 1.8 Drilling Rig Components (After King et al, 2012)

156

Figure 1.9 Vertical and directional wells (After Azar and Samuel, 2007)

Figure 1.10 Different types of drill bits (After Devereux ,2012)

157

Figure 1.11 Basic elements to drill a well (After Azar and Samuel, 2007)

158

Figure 1.12 Section of rotary drilling rig (After Azar and Samuel ,2007)

159

Figure 1.13 Block and Tackle Hoisting system (After Azar and Samuel , 2007)

160

Figure 1.14 Fluid circulation system (After Azar and Samuel , 2007)

161

Figure 1.15 Rotary system of drilling rig (After Azar and Samuel ,2007)

162

Figure 1.16 Floor of a drilling rig showing rotary table (After Devereux, 2012)

163

Figure 1.17 Typical blowout preventer (BOP) arrangement (After Azar and Samuel ,2007)

164

Figure 1.18 Picture of BOP (After Devereux , 2012)

165

Figure 1.19 Blowout preventer positioned on a well (After Devereux, 2012)

166

Figure 1.20 Horizontal drilling by jetting (After Devereux , 2012)

167

Figure 1.21 Horizontal drilling by using a whipstock (After Devereux ,2012)

168

Figure 1.22 Horizontal drilling via steerable motor (After Devereux , 2012)

169

Figure 1.23 Optimization of casing string along depth (After Azar and Samuel , 2007)

170

Figure 1.24 A conductor pipe (After Azar and Samuel , 2007)

171

Figure 2.1 Guide lines on PIT graph (After Postler et al ,1997)

Figure 2.2 Schematic of PIT equipment (After Postler et al ,1997)

172

Figure 2.3 - Description of dimensionless fracture conductivity (After Jones and Britt, 2009)

173

Figure 2.4 Three conventional test configurations for determination of vertical permeability
(After Ehlig-Economides et al, 2006)

174

Figure 2.5 Horner plot, pressure buildup test data from a South Texas gas well (After Lee and
Holditch,1979)

175

Figure 2.6 Millheim-Cinchowicz plot, pressure buildup test data from South Texas gas well (
After Lee and Holditch, 1979)

Figure 2.7 Synthetic drawdown curves for modified M-C method, constant rate case (After Lee
and Holditch, 1979)

176

Figure 2.8 Modified M-C plots, field buildup tests (After Lee and Holditch , 1979)

Figure 2.9 - Extrapolation of linear region, modified M-C method (After Lee and Holditch ,1979)

177

Figure 2.10 Typical partial perforation pressure buildup response (After Barnum et al, 1990)

178

Figure 2.11 Pressure buildup showing effect of wellbore damage and after-production (After
Howard and Fast, 1970)

Figure 2.12 Bottomhole pressure versus buildup time (After Gladfelter et al, 1955)

179

Figure 2.13 Drawdown plot of dimensionless real gas potential versus log of dimensionless time
(After Vairogos et al, 1973

180

Figure 2.14- Vertical permeability correlation chart (After Barnum et al, 1990)

181

Figure 2.15 Nomogram for end of first straight-line region (After Barnum et al, 1990)

182

Figure 2.16 Effect of fracture capacity on post-fracturing conductivity ( After Howard and Fast
,1970)

183

Figure 2.17 Effect of in-situ stress contrast on fracture containment (After Heydarabaradi et al,
2010)

Figure 2.18 Containment of a fracture opposite overburden (After Heydarabaradi et al, 2010)

184
a. Constant Stress gradient
b. Overstressed burden

Figure 2.19 Well blowout (After Lucasher Gusher

http://www.sjgs.com/gushers.html#spindletop, 2013)

185

Figure 2.20 - Full closure core catcher (After Unalmiser and Funk, 1998)

186

Figure 2.21 Composite photograph of core from various depositional environments (After
Keelan,1982)

187

Figure 2.22 SEM photograph of Illite clay in pore space. (After Keelan , 1982)

Figure 2.23 NMR relaxation and distribution (After Unalmiser and Funk ,1998)

188

Figure 2.24 CT scan 1 (After Al-Muthana et al ,2008)

Figure 2.25 CT scan 2 (After Al-Multhana et al, 2008)

Figure 2.26 CT scan 3 (After Al-Multhana et al, 2008)

189

Figure 2.27 Effect of calibration on grain density (After Al-Multhana et al, 2008)

Figure 2.28 Effect of drying time on porosity and permeability (After Al-Multhana et al, 2008)

190

Figure 2.29 Schematic diagram of shear and compressional wave propagation (After Timur A.
Turk)

Figure 2.30 Sonic log versus depth (After Glover, 1986)

191

Figure 2.31 Pre- and post-fracture temperature logs showing thermal conductivity effects (After
Jones and Britt, 2009)

192

Figure 2.32 A vertical well in the Woodford shale showing bedding planes and high angle
fractures using an LWD electrical image (After Quinn et al, 2008)

The red bars in the rose plot (circle) show fracture strikes while the green bars indicate azimuth of
sedimentary rock (Quinn et al, 2008).

193

Figure 2.33 LWD electrical images obtained from horizontal well in Barnett shale. (After Quinn
et al, 2008)

The left image shows intersecting and non-intersecting natural fractures as well as drilling-induced
tensile fractures on the top and bottom of the well. The right image shows drilling-induced tensile
fractures clearly visible along the bottom of the wellbore. This is less clear at the top of the well.
Also, there are many small, healed natural fractures intersecting the wellbore that have been
opened in the tensile section of the wellbore (compare static normalized image and dynamic
normalized image). This fractures were caused by drilling (Quinn et al, 2008).

194

Figure 2.34 Correlation between Total Organic content (TOC) log and Gamma Ray (GR) (After
Orlandi et al, 2011)

195

Figure 3.1 Normal leakoff plot of log Pressure Derivative versus time log -log plot (After Barree
et al, 2009)

196

Figure 3.2 Tip extension plot of Pressure derivative versus time (After Barree et al, 2009)

197

Figure 3.3 Pressure and Flow data from two cycles of Mini-frac tests in an offshore location at
the Timor sea (After Zoback, 2012)

In tests like this one, there is no significant difference between the shut-in pressure and closure
pressures on each of the two cycles. The pumping pressure is only about 100 psi higher than the
shut-in pressure. Also, the test records show that the surface pressure (which is roughly the
downhole pressur plus the hydrostatic pressure from the surface to measurement depth) varies by
less than 2% (Zoback, 2012).

198

Figure 3.4 Pressure-time records of two tests in shaly dolomite near Anna, Ohio. (After Zoback
and Haimson, 1982)

Test #37 represents initiation and extension of a horizontal bedding plane fracture. In test #39 a
vertical fracture initiated at the well bore and turned into a horizontal plane. This interpretation
was confirmed by an impression packer (Haimson, 182b)

199

Figure 3.5 Results from tests in a well in South Carolina (After Zoback and Haimson, 1982)

This shows change in shut-in pressure as hydraulic fractures are propagated (Haimson and Zoback,
1982)

200

Figure 3.6 Double tangent method plot I (After Amadei and Stephansson, 1997)

Figure 3.7 Double tangent method plot II (After Amadei and Stephansson, 1997)

201

Figure 3.8 Double tangent method plot III (After Amadei and Stephansson, 1997)

Figure 3.9 Muskat Method/Exponential pressure decay (After Amadei and Stephansson, 1997)

202

Figure 3.10- McLennan and Roegiers inflection method (After Amadei and Stephansson, 1997)

Figure 3.11 Horner plot (After Jones and Britt, 2007)

203

Figure 3.12 - Method of Turnbridge/Bilinear pressure decay rate method (After Amadei and
Stephansson, 1997)

204

Figure 3.13 - Pump-in flowback closure stress test (After Jones and Britt, 2009)
Pump-off
(leads to
curve
reversal)

205

Figure 3.14 Step-rate test (SRT) result, showing additional derivative plot below (After Jones
and Britt, 2009)

206

Figure 3.15 Step rate test (SRT) at different depths (After Zoback, 2012)

207

Figure 3.16 A schematic mini-frac test /XLOT showing pressure as a function of volume, or time
(at constant flowrate) (After Zoback, 2010)

208

Figure 3.17 Extended pressurization cycles of XLOTs (After Lee et al, 2004)

209

Figure 3.18 A standard LOT pressure versus volume plot showing leak-off pressure (After Addis
et al, 1998)

Figure 3.19 An idealized pressurization cycle from a PIT ( After Addis et al, 1998)

210

Figure 3.20 An Ideal example of a pressure record from a hydraulic fracturing stress test
showing the features sought in an ELOT (After Addis et al, 1998)

Figure 3.21 Pressure records for the six ELOTS performed in the North West Shelf, Australia
(After Addis et al, 1998)

211

Figure 3.22 A comparison of leak-off pressures from standard LOTs (+) with minimum stress
estimates from ELOTs for the North West Shelf of Australia (After Addis et al, 1998)

212

Figure 3.23 Leak-off pressures from standard LOTs (+) from vertical wells plotted with minimum
stress estimates from E(X)LOTS in the Oseberg Field, Norwegian sea. (After Addis et al, 1998)

Figure 3.24 Leak-off pressures from standard LOTs (+) from vertical and inclined wells plotted
with minimum stress estimates from E(X)LOTS in the Oseberg Field, Norwegian sea. (After Addis
et al, 1998)

213

Figure 3.25 Pressure Dependent Leakoff G-function plot (After Barree et al, 2009)

Figure 3.26 Tip extension G-function plot (After Barree et al, 2009)

214

Figure 3.27 - Storage G-function plot (After Barree et al, 2009)

215

Figure 3.28 Howard and Fast assumed fracture geometry (After Holditch et al, 1987)

216

Figure 3.29 Perkins-Kern-Nordgrens assumed fracture geometry (After Holditch et al, 1987)
Axis of borehole

hf

Xf

217

Figure 3.30 Vertical and Horizontal plane strain condition, with PKN geometry (showing
wellbore) below (After Valko and Economides, 1995)

218

Figure 3.31 Geertsma-de Klerk-Daneshy (KGD) assumed fracture geometry (After Holditch et al,
1987)

Figure 3.32 The KGD Geometry showing wellbore - (After Valko and Economides, 1995)

219

Figure 3.33- Radial fracture assumed geometry (After Holditch et al, 1987)

220

Figure 3.34 Transverse vertical fractures from a horizontal well (After Valko and Economides,
1995)

221

Figure 3.35 Turning fractures in a horizontal well from longitudinal initiation to the transverse
direction (After Valko and Economides, 1995)

Figure 3.36 Vertical Fracture initiated from an arbitrarily oriented horizontal well at an angle, a
from the minimum horizontal stress (After Valko and Economides, 1995)

222

Figure 3.37 Vertical fracture profile through a three-layer formation with dissimilar properties
(After Valko and Economides, 1995)

Figure 3.38 Fracture height growth with associated width reduction due to adjoining layer of
high stiffness (After Valko and Economides, 1995)

223

Figure 3.39 A T-shape fracture (proppant screenout) (After Valko and Economides, 1995)

224

Figure 3.40 Cumulative gas volume produced with time for different fracture half-lengths (After
Hareland and Rampersad, 1994)

Figure 3.41 Gas production decline with time for different fracture half-length (After Hareland
and Rampersad, 1994)

225

Figure. 3.42 Schematic of a Pseudo 3D Hydraulic Fracturing model (After Xiaowei Weng, 1992)

Figure. 3.43 Schematic of the 2D flow in the fracture (After Xiowei Weng, 1992)

226

Figure. 3.44 Approximation of the outer flow field at the element of interest by a local radial
flow (After Xiaowei Weng, 1992)

Figure. 3.45 Comparison of fracture shapes predicted by different models (After Xiaowei Weng,
1992)

227

Figure. 3.46 Example of StimPlan Output (After Barree et al, 2009)

Figure 3.47 Configuration of a fully 3D physical system (After Settari and Cleary, 1984)

228

Figure. 3.48 Fracture shapes for different degrees of containment (After Settari and Cleary,
1984)

Figure. 3.49 Comparison of computed fracturing pressures for different fracture geometries
(After Settari and Cleary, 1984)

229

Figure 3.50 An example of Fracpro-PT Output (After Barree et al, 2009)

230

Figure 3.51 Fracture Diagnostic techniques (After Cipolla and Wright, 2002)

231

Figure. 3.52 Principle of tiltmeter fracturing mapping (After Cipolla and Wright, 2002)

Figure. 3.53 Surface tiltmeter in a site at Western Missouri (After Jones and Britt)

232

Figure 3.54 Surface deformation for hydraulic fractures of different orientations at a depth of
3000 ft (After Wright et al, 1998)

233

Figure 3.55 Picture of tiltmeter used to predict eruptions on Mount Loa, Hawaii (After
Wilkipedia ,2013)

234

Figure 3.56 Conceptual drawing of actual fracture dimensions and estimated fracture
dimensions using downhole tiltmeter fracture mapping (After Wright et al, 1999)

The fracture mapping resolution depends on offset well location and distance relative to fracture
dimensions (Wright et l, 1999)

235

Figure 3.57 Theoretical downhole tiltmeter pattern of deformation (After Jones and Britt, 2009)

236

Figure 3.58 Principle of microseismic fracture mapping (After Cipolla and Wright, 2002)

Figure 3.59 Microseismic event location from Fisher (2005) (After Jones and Britt, 2009)

237

Figure 3.60 Multiple component geophone array with P and S wave arrivals (After Jones and
Britt, 2009)

238

Figure 3.61 Example of a Hodogram

(www.Crewes.org/Researchlinks/Explorerprogram/Holdogram/Holdogram.html)

239

Figure 3.62 Unconventional gas well history matching (After Britt et al, 2010)

240

Figure. 3.63 Nolte-Smith interpretation guide from Britt et al (1994) (After Jones and Britt,
2009)

241

Figure 3.64 Width decoupling due to containment caused by layers (After Barree et al, 2009)

Figure. 3.65 Composite layering and width decoupling Weijers et al, 2005

242

The first figure (left) shows the typical mechanism of fracture growth confinement due to increased
closure stress in the layers above and below the target zone. The middle figure is a mechanism
based on the 2D KGD model, and would result in perfect confinement at the layer interface. The
last figure (right) shows composite layering effect due to partial debonging of the layer interfaces
(Weijers et al, 2005).

Figure. 3.66 Observed net pressure (black) and match with model net pressure (green) (After
Weijers et al, 2005)

Note: The model net pressure response from both the calibrated and uncalibrated models are
almost identical.

243

Figure. 3.67 FracPRO PT model showing width profile and proppant concentration in fracture

In top section, estimated fracture for net pressure history match using typical model assumptions
is shown. In bottom section, matching of both net pressure history and directly observed fracture
geometry using additional containment effects such as composite layering is shown (After
Weijers et al, 2005)

244

Figure. 4.1 Composition of a Proppant Laden fracture Fluid ( After Arthur et al, 2009)

Figure 4.2- Schematic diagram showing effect of adding fluid-loss additives (After Hawsey et al ,
1961)

245

Figure 4.3- Plot of fluid loss versus time ( After Hawsey et al ,1961)

246

INJECTION RATE (BBL/MIN)

Figure 4.4- Fracture area and injection rate versus fluid loss coefficient (After Hawsey et al, 1961)

247

Figure 4.5- Correction factor for Spurt loss using nomograph ( After Hawsey et al, 1961)

248

Figure 4.6 Animation of Hydraulic fracturing (After Marathon Oil Corp, 2013)

Figure 4.7 Cross-sectional photos of three proppants under stress in the conductivity cell
(courtesy of Stim-Lab consortium) (After Palisch et al, 2010)

249

Figure 4.8 Effect of proppant settling velocities (After Bivins et al , 2005)

https://slb.com/~/media/Files/resources/oilfield_review/ors05/sum05/03_new_fibers.pdf

250

Figure 4.9 Production rate for different proppant sizes - After Cohen et al (2013)

Figure 4.10 Cumulative production for different proppant size (After Cohen et al , 2013)

251

Figure 4.11- Krumbein and Sloss chart of sphericity and roundness (After Kullman (2011) South
Dakota School of Mines)

252

Figure 4.12- Comparison of commercial proppants on the Krumbein and Sloss chart (After
Kullman (2011) South Dakota School of Mines)

253

Figure 4.13 A schematic of pressure during shut-in and flow (After Sookprasong , 2010)

Figure 4.14- Embedment pressure test technique (After Howard and Fast ,1970)

254

Figure 4.15 Picture of mounted core assembly for fracture capacity test (After Rixe et al ,1963)

Figure 4.16 Diagram of mounted core assembly for fracture capacity test (After Howard and
Fast, 1970)

255

Figure 4.17 Test apparatus for fracture flow capacity test (After Rixe et al , 1963)

256

Figure 4.18 Nitrogen flow system for measuring fracture flow capacity (After Rixe et al, 1963)

257

Figure 4.19 Scanning Electron Microscope (SEM) photograph (404X) of 40/80 mesh lightweight
ceramic proppant fines after a wet, hot crush test at 10,000 psi. (After Terracina et al , 2010)

Figure 4.20 Enveloping of proppant fines with resin coating as seen by a CAT scan of CRCS wet,
hot crush test at 10,000 psi. Grain to grain bonding with resin is also shown. (After Terracina et al
,2010)

258

Figure 4.21 Proppant flowback from the fracture into the wellbore can occur with uncoated
proppant or procured RCS (After Terracina et al, 2010)

Figure 4.22 SEM photo (651X) of CRCS grain to grain bonding that eliminates proppant flowback
by forming a consolidated proppant pack in the fracture (After Terracina et al ,2010)

259

Figure 4.23 Proppant embedment into the fracture face reduces fracture width and
conductivity. After Terracina et al (2010)

Figure 4.24 SEM photograph (514X) of formation fines spalling (in red circle) due to grain
embedment (After Terracina et al , 2010)

260

Figure 4.25 Approximate production increase due to fracturing (After Rixe et al ,1963)

261

Figure 4.26 Generalized selection chart for maximum fracture capacity obtainable at a depth of
7000 ft (After Rixe et al,1963)

262

Figure 4.27 Propping agent selection curve for sand at well depth of 7000ft ( After Rixe et al,
1963)

263

Figure 4.28 Propping agent selection curve for rounded nutshells at well depth of 7000ft ( After
Rixe et al, 1963)

264

Figure 4.29 Propping agent selection curve for rounded nutshells at well depth of 7000ft ( After
Rixe et al, 1963)

265

Figure 4.30 Potter County, PA Marcellus Well log (After Fontaine et al, 2008)

Onandanga Lime (good

barrier to fracture

Marcellus Shale @
5220 to 5310 ft depth

Depth (ft)

266

Figure 4.31- Treatment initiation: The effect of acid upon wellbore entry (After Fontaine et al
,2008)

Figure 4.32- The use and effect of proppant slugs ( After Fontaine et al ,2008)

Proppant slugs are additives

added to stabilize bottomhole
pressure, by reducing effects, if
any of fractures near the
wellbore.
267

Figure 4.33- Illustration of near wellbore complexity (tortuosity) versus simplicity. (After Fontaine
et al,2008)

Figure 4.34- Use of wellbore sweeps between proppant stages (After Fontaine et al ,2008)

268

Figure 4.35- The effect of a sweep (After Fontaine et al, 2008)

Figure 4.36 - Effects of Proppant Concentration and Mesh size on Net Pressure (After Fontaine et
al, 2008)

269

Figure 4.37- Finishing the treatment (After Fontaine et al, 2008)

270

APPENDIX
Principle of Fracturing Fluid Selection Using Fuzzy Logic (Xiong et al, 1996)
Table A1 gives a list of fracturing fluids and common conditions for their use.
Table A1

In the fuzzy logic system proposed by Xiong et al, 1996, the system first determines base fluid,
viscosifying method and energization method. Next, a choice is made of the 3 to 5 best
combinations of the possible fluids. Thereafter the system determines polymer type and loading,
crosslinker, gas type and other additives. It simultaneously checks the compatibility of the fluid and
additives with formation fluids. The possible combinations of viscosifying method, base fluid and
energizer is shown in table A2

271

Table A2 The Possible combinations of base fluid, viscosifying method and energization method

In the tables A3 to A9, the importance of the fuzzy variables to the decisions of the system is given
as a weighting factor. These factors are extracted from expertise (experience). They can be
adjusted for different situations in a fuzzy system. For instance, to rank the viscosification methods
for a particular formation case, the parameters considered are: temperature, fracture length,
height and treatment size (see fig A1). The membership functions are translated to weighting
factors and used to rank best choice of viscosification method. The rules that guide the ranking of
best choice are in figs. A13 to A18. The fuzzy system is a considerably subjective process because of
its reliance of weighting factors borne out of human experience/judgement.

272

Fig A1 Procedure to select fluids and additives

273

Viscosification method
Crosslinked fluids are usually used in the higher temperature formations and/or in large size
treatments. Linear fluids are preferred for low temperature and medium or small size treatments.
Polyemulsions are used when moderate viscosity is needed without the need for high temperature
stability.
Table A3 Parameters for Crosslinked, Polyemulsion and Linear fluids

Base fluids
Table A4 - Parameters for Water, Oil and Alcohol fluids

274

Table A5 Parameters for foam/energized/normal fluids

Polymer Type
There are different polymer types available for water base fluids in the industry. Guar and HPG are
popular. HEC is selected when fluid cleanup is important. Cost, residue and fracture length are also
important.
Table A6 Parameters for Polymer Type selection

275

Polymer concentration
This varies with treatment size and formation temperature (see table A-7 below)
Table A7 Parameters for Polymer Loading

Gas Quality: For foams, usually 70 to 75% quality is used depending on the pressure treatment size.
For energized fluid, the gas quality is about 15%.
Gas type: For foamed or energized fluid, N2 and CO2 are not compatible with oil based fluids. N2 is
normally not used in high temperature deep formations. The parameters to choose gas types are
shown in table A8 below:
Table A8 Parameters for gas type (N2 and CO2)

276

Crosslinker: The three commonly used crosslinkers are: borate, zirconium and titanium. Their
selection depends on the parameters shown in table A9.
Table A9 Parameters for crosslinker

Examples
In the table below are two example situations to apply the fuzzy logic system:
Table A-10

No 1 : In this case, we have a deep and thick formation containing gas. It has a high temperature
and normal pressure. The formation is also not water sensitive. Due to the above observations of

277

formation depth and temperature, a suitable choice proposed by Xiong et al (1996) is 60 Ibm
crosslinked CMHPG water gel (see table A-11 below)
Table A-11 Recommendations for example no 1

No 2: This is a low temperature, low permeability and low pressure gas formation, which like the
first case is not water sensitive. Three possible recommendations are made for this system (see
table A-12 below)

278

Table A-12 Recommendations for example no 2

Rules for the fuzzy evaluator

Table A 13 Rules for selecting viscosification methods

279

Table A 14 Rules for selecting base fluids

Table A 15 Rules for selecting energization methods

280

Table A 16 Rules for selecting gas type

Table A 17 Rules for selecting polymer type

281

Table A 18 Rules for selecting crosslinker type

282

Bibliography
Addis M.A et al (1998) A Comparison of Leak-off Test and Extended Leak-off Test Data for Stress Estimation.
Presented at SPE/ISRM Eurock 98 held in Trondheim, Norway, 8 to 10 July. OnePetro.org.
Al-Muthana S.A, Ma S.M, Okasha Taha (2008), Best Practices in Conventional Core Analysis-A Laboratory
Investigation. Presented at the 2008 SPE Saudi Arabia section Technical Symposium held in
Alkhobar, Saudi Arabia, May 10 to 12, 2008. SPE 120814. OnePetro.org.
Amadei Bernard and Ove Stephansson, (1997) Rock Stress and Its Measurement. 1st edition.

Cambridge

University Press. London: Chapman and Hall

Arthur et al (2009) Evaluating Implications of Hydraulic fracturing in Shale Gas Reservoirs. Prepared for
presentation at the 2009 SPE Americas E&P Environmental & Safety Conference, held in San
Antonio, Texas, USA from 23 to 25 March, SPE 121038. OnePetro.org.
st

Azar J.J and Samuel G.R (2007) Drilling Engineering. 1 Edition. Oklahoma: Penwell corporation. ISBN-13:
978-1-59370-072-0.
Barree R.D, Barree V.L and Craig D.P (2007) Holistic Fracture Diagnostics: Consistent interpretation of Prefrac Injection Tests Using Multiple Analysis Methods. Presented at the 2008 SPE Rocky Mountain oil
and gas technology symposium, Denver, 16 to 18 April. SPE 107877. OnePetro.org.
th

Barree & Associates (2009) Modeling Fracture Geometry. Retrieved on February 28 , 2013.

http://barree.net/C2%20-%20Modeling%20Fracture%20Geometry.pdf
Britt L.K, Jones J.R and Miller II (2010) Defining Horizontal Wwell Objectives in Tight and Unconventional Gas
Reservoirs. In Canadian Unconventional Resources & International Petroleum Conference, Calgary
Alberta, Canada, from 19 to 21 October. CSUG/SPE 137839. OnePetro.org.

283

Cipolla C. and Wright C.A (2002) Diagnostic Techniques to Understand Hydraulic Fracturing: What? Why and
How? First presented at the 2000 SPE/CERI Gas technology symposium, Calgary, Canada, 3 to 5
April. OnePetro.org.
Cohen et al (2013), Optimum Fluid and Proppant Selection for Hydraulic Fracturing of Shale Gas Reservoirs: A
Parametric study Based on Fracturing-to Production Simulation. Presented at the SPE Hydraulic
Fracturing Technology Conference held in the Woodlands, Texas, USA, February 4 to 6 2013. SPE
163876. OnePetro.org.
rd

Crewes Hodogram Explorer 1.0 (2013), Hodogram. Retrieved on April 23 at 3:00pm.

(www.Crewes.org/Researchlinks/Explorerprogram/Holdogram/Holdogram.html)

Detournay E. and Carbonell R. (1997) Fracture-Mechanics Analysis of the Breakdown Process in Minifracture
or Leakoff Test. First presented at the 1994 SPE/ISRM Rock Mechanics in Petroleum Engineering
Conference, held in Delft, The Netherlands, 29 to 31 August. SPE 28076. OnePetro.org.
nd

Devereux Steve (2012), Drilling Technology in Non-Technical Language. 2

edition. Oklahoma: Penwell

corporation. ISBN: 978-1-59370-264-9.

Du J. et al, Joint Inversion of Downhole Tiltmeter and Microseismic Data and its Application to Hydraulic
Fracture Mapping in Tight Gas Sand Formation. Presented at the 42

nd

US Rock Mechanics

nd

symposium and 2 US-Canada Rock Mechanics symposium, held in San Francisco, June 29 to July 2,
2008, ARMA 08-344. OnePetro.org.
Ehlig-Economides C.A, Nduonyi M, Abiazie J. (2006), Test Design for Vertical Permeability from a
Conventional Pressure Buildup Test. Presented at the First International Oil Conference and
st

nd

Exhibiton held in Cancun, Mexico, August 31 to September 2 2006. SPE 103680. OnePetro.org.
Fontaine J, Johnson N, Schoen D (2008), Design, Execution and Evaluation of a Typical Marcellus Shale
SlickWater Stimulation: A Case History. Presented at the 2008 SPE Eastern Regional/AAPG Eastern

284

Section Joint meeting, held in Pittsburgh, Pennsylvania, USA, Octobetr 11 to 15, 2008. SPE 117772.
OnePetro.org.
Gang Li, Lorwongngam A., Roegiers J. C (2009) Critical Review of Leak-off Test as a Practice for the
rd

th

Determination of In-situ Stresses. Presented at the 43 US Rock mechanics symposium, and 4 USCanada Rock Mechanics symposium, held in Ashville, North Carolina June 28 to July 1, ARMA 09-3.
OnePetro.org.
Gladfelter R.E, Tracy G.W, Wilsey L.E (1955) Selecting Wells which will respond to Production-Stimulation
Treatment. Presented by G.W Tracy at the Spring meeting of the Mid-continent District, Division of
production, Amarillo, March 1955. OnePetro.org.
Hareland G. and Rampersad P.R (1994) Hydraulic Fracturing Design Optimization in Low Permeability Gas
Reservoirs. Presented at the III Latin America/Carribean Petroleum Engineering Conference held in
Buenos Aires, Argentina, 27 to 29 April. SPE 27033. OnePetro.org.
Harris P.C (1988), Fracturing-Fluid Additives. Journal of Petroleum Technology, October 1988 ed. SPE
Distinguished Author series.
Hawsey J.D, Jacocks C.L (1961), The Use of Fluid-Loss Additives in Hydraulic Fracturing of Oil and Gas Wells.
Presented at the 32

nd

Annual California Regional Meeting of the Society of Petroleum Engineers

nd

rd

(SPE) of AIME in Bakersfield, California, November 2 to 3 . SPE-244. OnePetro.org.

Heyderbaradi F.R et al (2010), Criteria for Selecting a Candidate Well for Hydraulic Fracturing. Presented at
th

the 34 Annual SPE International Conference and Exhibition held in Tinapa, Calabar, Nigeria, 31 July
to 7 August 2010. SPE 136988. OnePetro.org.
Hibbeler J. and Rae P. (2005) Simplifying Hydraulic Fracturing: Theory and Practice. Prepared for
presentation at the 2005 SPE technical conference and exhibition, Dallas, Texas, USA, from 9 to 12
October, SPE 97311. OnePetro.org.

285

Holditch S.A, Robinson B.M, Whitehead W.S, (1987) Prefracture and Postfracture Formation Evaluation
Necessary to Characterize the Three Dimensional Shape of the Hydraulic Fracture. SPE Formation
Evaluation, December 1987. OnePetro.org.
st

Howard G.C and Fast C.R (1970), Hydraulic Fracturing. 1 Edition. Dallas, Texas: Millet the Printer. American
Institute of Mining, Metallurgical and Petroleum Engineers, Inc. (AIME) and Society of Petroleum
Engineers (SPE.)
Hsiao C. and Tsay F. (1990) Evaluation of Fracture Parameters Using Pump-in/Flow-back Test. International
technical meeting jointly hosted by Petroleum Society of CIM and Society for Petroleum Engineers
(SPE), held between June 10 to June 13. OnePetro.org.
st

Jones J.R and Britt L.K (2009) Design and Appraisal of Hydraulic Fractures 1 Edition. Texas: Society of
Petroleum Engineers (SPE). ISBN: 978-1-55563-143-7
Keelan D.K (1971) A Critical Review of Core Analysis Techniques. Presented at the 22

nd

Annual Technical

meeting of the Petroleum society of CIMM, Banf, Alberta, June 1971. OnePetro.org.
Keelan D.K (1982), Core Analysis for aid in Reservoir Description. Journal of Petroleum technology, June
1982 ed. Society of Petroleum Engineers (SPE) of AIME Distinguished Author Series. OnePetro.org.
Kennedy R. et al, (2012) Optimized Shale Resource Development: Proper Placement of Wells and Hydraulic
Fracture Stages. Prepared for presentation at the Abu Dhabi International Petroleum Conference
and Exhibition held in Abu Dhabi, UAE, 11 to 14 November. SPE 162534. OnePetro.org.
King G.E (2012) What every Representative, Environmentalist, Regulator, Reporter, Investor, University
Researcher, Neighbor and Engineer Should Know About Estimating Frac Risk and Improving Frac
Performance in Unconventional Gas and Oil Wells. In the SPE Hydraulic Fracturing Technology
Conference held in the Woodlands, Texas, USA from 6 to 8 February, SPE 152596. OnePetro.org.

286

Lee D, Birchwood R, Bratton T (2004) Leak-Off Test Interpretation and Modeling with Application to
Geomechanics. Presented at the Gulf Rocks 2004, the 6

th

North American Rock Mechanics

Symposium held in Houston, Texas, June 5 9, ARMS/NARMS 04-547. OnePetro.org.

Lee W.J and Holditch S.A (1979), Fracture Evaluation with Pressure Transient Testing in Low Permeabiltiy Gas
Reservoirs. Part 1: Theoretical Background. SPE 7929. OnePetro.org.
Lehman L. et al (2002), Calibrating Fracture Models with Direct Diagnostics: A Necessary but Humbling
Experience. Presented at the SPE Asia Pacific Oil and Gas conference held in Melbourne, Australia,
8 to 10 October, SPE 77904. OnePetro.org.
Martin A.N and Economides M.J (2010) Best Practices for Candidate Selection, Design and Evaluation of
Hydraulic Fracture Treatments. Presented at the SPE Production and Operations Conference and
Exhibition held in Tunis, Tunisia, 8 to 10 June 2010. SPE 135669. OnePetro.org.
McPhee C. (2012) The Core Analysis Elephant in the Formation Evaluation Room. Presented at the SPE
Annual Technical Conference held at San Antonio, Texas, USA, October 8 to 10. SPE 158087.
OnePetro.org.
Middlebrook M.L, Aud W.W, Harkrider J.D, Hansen J.T (1997), An Evolving Approach in the Analysis of Stress
Test Pressure Decline Data First presented at the 1995 SPE Rocky Mountain meeting/ Low
permeability reservoir symposium held in Denver, Colorado, 20 to 22 March. OnePetro.org.
Orlandi M, Bartelucci P, Chelini V (2011) Unconventional Reservoir Characterization Methods Using Core and
Well Logging Data: Shale Gas and Tight Gas Sand Examples. Presented at the 10

th

Offshore

Mediterranean Conference and Exhibition in Ravenna, Italy, March 23 to 25. OnePetro.org.

Palisch et al (2010), How to Use and Misuse Proppant Crush Tests: Exposing the Top 10 Myths. Presented at
the SPE hydraulic fracturing technology conference, held at the Woodlands, Texas, USA, January 19
to 21, 2009. SPE 119242. OnePetro.org.

287

Postler D.P (1997) Pressure Integrity Test Interpretation. Presented at the 1997 SPE/IADC conference held
in Amsterdam, The Netherlands, between 4 to 6 March. SPE /IADC 37589. OnePetro.org.
Pye D.S and Smith (1973), Fluid Losss Additive Seriously Reduces Fracture Proppant Conductivity and
Formation Permeability. Presented at the 48

th

Annual meeting of the Society of Petroleum

rd

Engineers of AIME to be held in Las Vegas, Nevada, USA between September 30 to October 3 ,
1973. SPE 4680. OnePetro.org.
Quinn T.H, Dwyer J, Wolfe C, Morris S, Coope D (2008), Formation Evaluation Logging While Drilling (LWD)
in Unconventional Reservoirs for Production Optimization. Presented at the 2008 SPE Eastern
Regional/AAPG Eastern Section Joint meeting, held in Pittsburgh, Pennsylvania, USA, Octobetr 11 to
15, 2008. SPE 119227. OnePetro.org.
Raeen A.M, Brudy M, SPE, Statoil ASA (2001), Pump in/Flowback Tests Reduce the Estimate of Horizontal Insitu Stress Significantly Presented at the 2001 SPE Annual technical Conference and Exhibition held
in New Orleans, Louisiana, 30 September to 3 October. SPE 71367. OnePetro.org.
Rixe F.H, Fast C.R, Howard G.C (1963) Selection of Propping Agents for Hydraulic Fracturing. Presented by
F.H Rixe at the Spring meeting of the Rocky Mountain District, API Division of Production, April 1963.
OnePetro.org.
Settari A. and Cleary M.P (1986) Development and testing of a Pseudo-Three-Dimensional Model of Hydraulic
fracture Geometry Society of Petroleum Engineers (SPE) Production Engineering, 1986.
OnePetro.org.
Settari A. and Cleary M.P (1984) Three Dimensional Simulation of Hydraulic Fracturing Journal of Petroleum
Technology, 1984 ed. Society of Petroleum Engineers (SPE). OnePetro.org.

288

Sookprasong P.A (2010), In-situ closure stress on Proppant in the Fracture: A Controversial New Way of
Thinking. Presented at the SPE Tight Gas Completions Conference held in San Antonio, Texas, USA,
2 to 3 November 2010. SPE 136338. OnePetro.org.
Terracina J.M, Turner J.M, Collins D.H, Spillars S.E (2010) Proppant Selection and Its Effects on the Results of
Fracturing Treatments Performed in Shale Formations. Presented at the SPE Annual Technical
Conference and Exhibition held in Florence, Italy, 19 to 22 September 2012. SPE 135502.
OnePetro.org.
Timur (Turk) A. (1980) Acoustic Logging. Chapter 51, Petroleum Engineering Handbook.
Unalmiser S. and Funk J.J (2008) Engineering Core Analysis. SPE Distinguished Author Series, December
1981. SPE 36780. OnePetro.org.
Valko Peter and Economides M.J (1995), Hydraulic Fracture Mechanics 1st Edition. London: John Wiley and
Sons. ISBN: 0471956643, 9780471956648
Van Eekelen, H.A.M (1982) Hydraulic Fracture Geometry: Fracture Containment in Layered Formations.
Society of Petroleum Engineers (SPE) Journal, June 1982. OnePetro.org.
Warpinski et al (1995), Microseismic Mapping of Hydraulic Fractures using Multi-level Wireline Receivers.
Presentation at the 2005 SPE annual technical conference and exhibition, Dallas, Texas, USA, from 22
to 25 October, SPE 30507. OnePetro.org.
Weijers L., Wright C., Mayerhofer M, Cipolla C. (2005), Developing Calibrated Fracture Growth Models for
Various Formations and Regions Across the United States. Presented at the 2005 SPE Annual
Technical Conference held in Dallas, Texas, USA, October 9 to 12. SPE 96080. OnePetro.org.
Weng Xiaowei (1992) Incorporation of 2D Fluid Flow into a Pseudo-3D Hydraulic Fracturing Simulator. SPE
production engineering, November 1992. OnePetro.org.

289

Wilkipedia. (2013) Tiltmeters- Wilkipedia, the free encyclopedia.

http://en.wikipedia.org/wiki/Tiltmeter. Retrieved on May 1st at 9:25 pm.

Wright C.A, Davis E.J, Wang G, Weijers L. (1999), Downhole Tiltmeter Fracture Mapping: A New Tool for
Direct Measurement of Hydraulic Fracturing Growth. Rock Mechanics for industry, Amadei, Krantz,
Scott & Smeallie (eds) Balkema, Rotterdam, ISBN 9058090523. OnePetro.org.
Wright C.A et al, (1998), Downhole tiltmeter fracture mapping: A new tool for directly measuring Hydraulic
fracture Dimensions. This paper was presented at the 1998 SPE Annual Technical Conference and
Exhibition held in New Orleans, Louisisana, September 27 to 30, SPE 49193. OnePetro.org.
Xiong Hongjie, Davidson B, Saunders B, Holditch S.A (1996) A Comprehensive Approach to Select Fracturing
Fluids and Additives for Fracture Treatments. Presented at the 1996 SPE Annual Technical
Conference and Exhibition, held in Denver, Colorado, USA, October 6 to 9. SPE 36603. OnePetro.org.
Yang Mei, Xicai L, Di J, Economides M.J (2013), Hydraulic Fracture Design Flaws: Proppant Selection.
Presented at the SPE Western Regional & AAPG Pacific Section meeting, 2013 Joint technical
Conference held in Monterey, California, USA, 19 to 25 April 2013. SPE 165328. OnePetro.org.
Yang Mei, Economides M.J (2012) Proppant Selection for Hydraulic Fracture Production Optimization in
Shale Plays. Presented at the 2012 SPE Eastern Regional meeting held in Lexington, Kentucky, USA,
3 to 5 October 2012. SPE 161355. OnePetro.org.
th

Youtube (2012) Animation of Hydraulic Fracturing. Marathon Oil Corporation. Published April 26 , 2012.

https://www.youtube.com/watch?v=VY34PQUiwOQ. Retrieved on 15th April, 2013.

Zoback Mark D. (2010) Reservoir Geomechanics. Paperback edition. New York: Cambridge University Press.
ISBN: 978-0-521-14619-7

290

Zoback M.D and Haimson B.C (1981), Status of the Hydraulic Fracturing Method for In-situ Stress
Measurements Workshop on Hydraulic Fracturing Sress Measurements hosted by the University of
Wisconsin, held at Monterey, California in December 1981. OnePetro.org.

291