/

University of Southern California Dissertations and Theses

/

The study of mass transfer in gas shales and the optimization of hydraulic stimulation processes via additives

(USC Thesis Other)

The study of mass transfer in gas shales and the optimization of hydraulic stimulation processes via additives

PDF

Transcript (if available)

Content The Study of Mass Transfer in Gas Shales and the Optimization of Hydraulic Stimulation Processes via Additives By Junyi Xu A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (CHEMICAL ENGINEERING) December 2013 Copyright 2013 Junyi Xu ii Dedication I would like to dedicate this dissertation to my parents, Hongchuan Xu and Xiaochu Yang, for their endless support, encouragement and sacrifices. iii Acknowledgements First and foremost, I would like to express my deepest gratitude to my advisors, Dr. Kristian Jessen and Professor Theodore Tsotsis for their invaluable guidance, motivation and encouragement throughout the course of this work. Their wisdom and commitment inspired and empowered me. Without their immense knowledge, insightful comments and suggestions, it would not have been possible for me to overcome all the challenges along the way. Their contributions to me will be embedded in my heart forever. I am also very grateful to my dissertation committee member Professor Doug Hammond. His advice and support provided me the opportunity to better appreciate the new findings. I would also like to thank Professor Iraj Ershaghi and Dr. Katherine Shing who served on my qualifying exam committee. I will forever cherish their invaluable guidance and help. I would like to express my indebted gratitude and special thanks to Mr. Kyle Mork for the internship opportunity in Energy Corporation of America. The field experience I gained has been serving as an important source for providing directions for my lab-scale research. In particular, I would also like to thank Chad Perkins and Pete Sullivan for their technical mentorship. I would also like to extend my gratitude to Mr. John Mork and Mr. Matt Flavin for their encouragement and help. I would like to thank my former colleagues at USC, Dr. Bobby Liu and Dr. Qiyao Feng who helped me a lot in my research journey. I also want to thank my current group members Mr. Yu Wang and Ms. Mini Basabdatta for their assistance at various stages of this research. Equally important, I would like to thank my officemate Mr. Devang Dasani, for his encouragement and help. I am extremely grateful to the Mork Family Department of Chemical Engineering and Materials Science for giving me the opportunity to pursue my dream. I would like to thank Ms. Tina Silva, Ms. Karen Woo, Ms. Aimee Barnard, Ms. Idania Takimoto, and Mr. Martin Olekszyk and the rest of the staff in MFD for their kind help on numerous occasions. iv I am also indebted to all my friends for their invaluable friendship and encouragement during the course of my life. Finally and most importantly, I would like to thank my parents, Hongchuan Xu and Xiaochu Yang for their endless love, encouragement and support. Without their sacrifices, it would have been impossible for me to fulfill my dream. v Table of Contents Dedication ii Acknowledgements iii List of Tables viii List of Figures x Abstract xvi Chapter 1: Introduction 1 1.1 Importance of Unconventional Natural Gas 1 1.2 The Shale Gas Industry 6 1.3 Current Practice for Gas Shale Production 8 1.4 Dissertation Objectives 10 1.5 Organization of Dissertation 11 Chapter 2: Characterization of Gas-shale Materials 12 2.1 Overview of Characterization Techniques 12 2.2 Specific Surface Area and Pore Structure Analysis 15 2.3 Scanning Electron Microscope (SEM) 34 2.4 Elemental Analysis and Mineralogy of Gas-shale Materials 37 2.5 Permeability and Porosity Measurements 46 2.6 Summary and Discussion 50 Chapter 3: The Study of Mass Transfer in Gas Shales via Depletion Experiments 54 3.1 Experimental Approach and System Configuration 56 3.2 Design of Depletion Experiments 60 3.3 Depletion Experiments 64 3.3.1 Loading Time Experiments 64 3.3.2 Short-Term Methane Depletion Experiments 66 3.3.3 Short-Term Argon Depletion Experiments 74 3.3.4 Blank (without a Core) Depletion Experiments 79 vi 3.3.5 Long-Term Methane Depletion Experiments 86 3.3.6 Data Analysis and Preliminary Discussion 89 3.4 Modeling 94 3.5 Summary and Conclusions 105 Chapter 4: Optimization of Hydraulic Stimulation Processes via Additives 106 4.1 Introduction and Overview of Additives 106 4.1.1 Emulsions 107 4.1.2 Viscoelastic Surfactants 109 4.1.3 Polymers 111 4.2 Experimental Study of a Novel Surfactant Additive 114 4.3 Lab-scale Wettability Alteration Experiments 115 4.4 Lab-scale High-Pressure Flow-Back Experiments 120 4.4.1 Experimental Approach 120 4.4.2 Experimental Results 124 4.5 Analysis for Field-scale Implementation 126 4.6 Investigation of Surfactant X under Actual Frac Fluid Conditions 130 4.6.1 Contact Angles Measurements 130 4.6.2 Spontaneous Imbibition Experiments 131 4.6.3 Potential Interactions of the Surfactant X with the Proppants 133 4.7 Summary and Conclusions 137 Chapter 5: Summary of Dissertation and Recommendations for Future Work 139 4.8 Dissertation Summary 139 4.9 Recommendations for Future Work 141 Bibliography 143 Appendix A-1 151 Appendix A-2 155 Appendix A-3 158 Appendix B 159 vii Appendix C 163 Appendix D 168 viii List of Tables Table 1.1: Distribution of Worldwide Unconventional Gas Reservoirs 3 Table 2.1: Sample Depth and Description 13 Table 2.2: Specific Surface Area Comparison 19 Table 2.3: Pore Volume Comparison via HK and BJH 25 Table 2.4: Vertical Permeability Measured by He Flow-through Experiments 48 Table 2.5: Sample # and Depth for Directional Permeability Measurement 49 Table 2.6: Correlation of Shale Matrix Permeability with Various Other Characteristics Measured Using Various Other Techniques 50 Table 2.7: Correlation of Shale Total Gas Content with Various Other Characteristics Measured Using Various Other Techniques 51 Table 3.1: Component Names for Figure 3.1 56 Table 3.2: Parameters Summary for Methane Short Term Depletion 66 Table 3.3: Summary of Parameters for Argon Depletion 74 Table 3.4: Parameters Summary of Long Term Methane Depletion Experiments 86 Table 3.5: Depleted Void Volume of Short Term 2000 psig Argon Depletion 90 Table 3.6: Depleted Void Volume of Short Term 2000 psig Methane Depletion 90 Table 4.1: Commonly Used Fracturing Fluid Additives and Their Function 106 Table 4.2: Physical Attributes of Conventional Emulsions and Dilute Microemulsions 107 Table 4.3: Overview of Microemulsion (ME) 108 Table 4.4: Summary of Contact Angle Experiments 116 ix Table 4.5: Variability in Contact Angle Measurements within the Same Sample 116 Table 4.6: Contact Angle Measurements on Samples Exposed to Surfactant Solution after 43 Hours 118 Table 4.7: Component for Figure 4.13 120 Table 4.8: Forced Imbibition Data Summary 125 Table 4.9: Cost Estimate for Use of Proposed Additive in Well Stimulation -Case 1 127 Table 4.10: Extra Cost Associate with Flow-back Water 127 Table 4.11: Cost Estimate for Use of Proposed Additive in Well Stimulation -Case 2 128 Table 4.12: Cost estimate for Use of Proposed Additive in Well Stimulation -Case 3 129 Table 4.13: Summary of Contact Angle Measurements with Frac Fluid 131 Table 4.14: Dimensions and Other Characteristics of the Samples Used in the Spontaneous Imbibition Experiments 132 x List of Figures Figure 1.1: Commercial Shale Gas Production in the USA 6 Figure 1.2: Shale-Gas Production Growth by EIA 7 Figure 2.1: Micrometrics ASAP 2010 Instrument and PC 16 Figure 2.2: BET Specific Surface Area vs. Sample Depth 18 Figure 2.3: Lost Gas (percent) vs. Sample Depth 18 Figure 2.4: Total Gas Content (scf/ton) vs. Depth 19 Figure 2.5: Average Pore Diameter vs. Depth 20 Figure 2.6: Cumulative Pore Volume from BJH Adsorption 23 Figure 2.7: Cumulative Intrusion from Mercury Injection Measurements 24 Figure 2.8: Micropore (d<2nm) Volume vs Depth 26 Figure 2.9: Mesopore (2<d<50nm) Volume vs Depth 26 Figure 2.10: Macropore (50<d<500nm) Volume vs Depth 27 Figure 2.11: Total Pore Volume vs Depth 27 Figure 2.12: Cumulative Pore Volume from BJH Desorption (Depth 7721.5-7750.5ft) 28 Figure 2.13: dV/dD vs Pore Diameter from BJH Desorption (Depth 7721.5 – 7750.5ft) 29 Figure 2.14: dV/dD vs Pore Diameter from BJH Desorption (Depth 7841.5ft - 7891.5ft) 30 Figure 2.15: Cumulative Pore Volume from HK Method (Depth 7772.5ft – 7823.5ft) 31 xi Figure 2.16: Cumulative Pore Volume from HK Method (Depth 7852.5ft – 7891.5ft) 32 Figure 2.17: Vertical (left) and Horizontal (right) SEM Images from 7772.5 ft (2000x) 34 Figure 2.18: Vertical View of Samples via Different Treatments (1000X) 35 Figure 2.19: EDX Analysis and SEM Image for Vertical View Sample with Reported Permeability at Depth 7891.5 ft (2000X): k = 2.54E-04mD 38 Figure 2.20: EDX Analysis and SEM Image for Vertical View Sample with Reported Permeability at Depth 7862.5 ft (2000X): k = 1.58E-04mD 38 Figure 2.21: EDX Analysis and SEM Image for Vertical View Sample with Reported Permeability at Depth 7885.5 ft (2000X): k = 4.65E-05mD 39 Figure 2.22: EDX Analysis and SEM Image for Vertical View Sample with Reported Permeability at Depth 7732.5 ft (2000X): k = 1.77E-09mD 39 Figure 2.23: EDX Analysis and SEM Image for Vertical View Sample with Reported Permeability at Depth 7742.5 ft (2000x): k = 1.16E-08mD 40 Figure 2.24: EDX Analysis and SEM Image for Vertical View Sample with Reported Permeability at Depth 7721.5 ft (2000x): k = 1.22E-08mD 40 Figure 2.25: TOC Analysis Results versus Sample Depth 41 Figure 2.26: Total Clay Content versus Sample Depth via XRD 41 Figure 2.27: Calcite Content versus Sample Depth via XRD 42 Figure 2.28: Quartz Content versus Sample Depth via XRD 42 Figure 2.29: Plagioclase Content versus Sample Depth via XRD 43 Figure 2.30: Pyrite Content versus Sample Depth via XRD 43 Figure 2.31: Minerals vs Depth 44 xii Figure 2.32: Relative Clay vs Depth 44 Figure 2.33: Matrix Permeability vs Depth 46 Figure 2.34: Shale Core Plug Porosity vs. Depth 47 Figure 2.35: Horizontal Permeability Measured by He Flow-through Experiments 47 Figure 2.36: Permeability vs Net Stress 48 Figure 3.1: High Pressure Depletion System 57 Figure 3.2: Core-Holder and Data Acquisition Sub-system 59 Figure 3.3: Shale Core Sample for Depletion Experiments 61 Figure 3.4: Flow rates of Methane Experiments 65 Figure 3.5: Flow rates of Argon Experiments 65 Figure 3.6: Pressure Curve of Short-Term 2000-0 psig Methane Depletion 67 Figure 3.7: Pressure Curve of Short Term 2000-500 psig Methane Depletion 67 Figure 3.8: Pressure Curve of Short Term 2000-750 psig Methane Depletion 68 Figure 3.9: Pressure Curve of Short Term 2000-1000 psig Methane Depletion 68 Figure 3.10: Pressure Curve of Short Term 2000-1250 psig Methane Depletion 69 Figure 3.11: Pressure Curve of Short Term 2000-1500 psig Methane Depletion 69 Figure 3.12: Flow rate of Short Term 2000-0 psig Methane Depletion 70 Figure 3.13: Flow rate of Short Term 2000-500 psig Methane Depletion 70 Figure 3.14: Flow rate of Short Term 2000-750 psig Methane Depletion 71 Figure 3.15: Flow rate of Short Term 2000-1000 psig Methane Depletion 71 Figure 3.16: Flow rate of Short Term 2000-1250 psig Methane Depletion 72 xiii Figure 3.17: Flow rate of Short Term 2000-1500 psig Methane Depletion 72 Figure 3.18: Flow Rates Comparison of Short Term Methane Depletion Experiments 73 Figure 3.19: Pressure Change of Short Term 2000-0 psig Argon Depletion 75 Figure 3.20: Pressure Change of Short Term 2000-500 psig Argon Depletion 75 Figure 3.21: Pressure Change of Short Term 2000-1000 psig Argon Depletion 76 Figure 3.22: Flow Rate of Short Term 2000-0 psig Argon Depletion 76 Figure 3.23: Flow Rate of Short Term 2000-500 psig Argon Depletion 77 Figure 3.24: Flow Rate of Short Term 2000-1000 psig Argon Depletion 77 Figure 3.25: Flow Rate Comparison of Short Argon Depletion Experiments 78 Figure 3.26: Pressure Curve of 2000-0 psig Methane Blank Depletion 79 Figure 3.27: Pressure Curve of 2000-500 psig Methane Blank Depletion 80 Figure 3.28: Pressure Curve of 2000-750 psig Methane Blank Depletion 80 Figure 3.29: Pressure Curve of 2000-1000 psig Methane Blank Depletion 81 Figure 3.30: Pressure Curve of 2000-1250 psig Methane Blank Depletion 81 Figure 3.31: Pressure Curve of 2000-1500 psig Methane Blank Depletion 82 Figure 3.32: Flow Rate of 2000-0 psig Methane Blank Depletion 82 Figure 3.33: Flow Rate of 2000-500 psig Methane Blank Depletion 83 Figure 3.34: Flow Rate of 2000-750 psig Methane Blank Depletion 83 Figure 3.35: Flow Rate of 2000-1000 psig Methane Blank Depletion 84 Figure 3.36: Flow Rate of 2000-1250 psig Methane Blank Depletion 84 xiv Figure 3.37: Flow Rate of 2000-1500 psig Methane Blank Depletion 85 Figure 3.38: Flow Rates Comparison of All Methane Blank Depletion Runs 85 Figure 3.39: Pressure Change during 2000-0 psig Long Term Depletion 87 Figure 3.40: Pressure Change during 2000-500 psig Long Term Depletion 87 Figure 3.41: Pressure Change during 2000-1000 psig Long Term Depletion 88 Figure 3.42: Flow Rate Comparison of 2000 psig Set 92 Figure 3.43: Cumulative Production of the Entire Depletion Lifespan of 2000 psig Set 93 Figure 3.44: Illustration of the BPM for Our Simulation Purpose 95 Figure 3.45: Comparison of BPM with Experimental Data for 2000-0 psig Long Term 103 Figure 3.46: Comparison of BPM with Experimental Data for 2000-500 psig Long Term 104 Figure 3.47: Comparison of BPM with Experimental Data for 2000-1000 psig Long Term 104 Figure 4.1: Well performance on back production from fracture treatments when using mutual solvent (2-BE) and without using mutual solvent 110 Figure 4.2: Well performance on back production from fracture treatments when using the environmentally acceptable mutual solvent A and then 2-BE 110 Figure 4.3: Water (Without Chemical Treatment) 112 Figure 4.4: Water (Treatment With 2% FC 722) 113 Figure 4.5: Contact Angle vs. Concentration of FC 754-glass Tube 113 Figure 4.6: Contact angle vs. Concentration of FC 722-glass tube 113 xv Figure 4.7: Illustration of contact angle measurement experiments 117 Figure 4.8: Spontaneous Imbibition for DI and 30 ppm Surfactant X 118 Figure 4.9: Spontaneous Imbibition for DI and 60 ppm Surfactant X 119 Figure 4.10: Comparison of Contact Angle before and After Surfactant X Treatment 119 Figure 4.11: High-Pressure Flow Back Experimental Setup 121 Figure 4.12: Main Experimental System 123 Figure 4.13: Spontaneous Imbibition Results for 3 Fluid Compositions 132 Figure 4.14: Spontaneous Imbibition for Frac Fluid/30ppm and Frac Fluid/60ppm 133 Figure 4.15: Static Adsorption Experiment Setup 134 Figure 4.16: Adsorption of Surfactant X onto the 100 Mesh Sand 135 Figure 4.17: Flow through Sand Experimental Setup 136 Figure 4.18: Adsorption Loss at 0.66m/s Fluid Velocity 137 xvi Abstract Gas shales contain an abundance of natural gas. Low-permeability (“tight”) shale is commonly found domestically in the Appalachian Basin but it has yet to be fully explored. A major limitation to the efficient extraction of this “tight gas” is the current lack of knowledge of the physics and mass transfer properties involved in production from such low-permeability porous materials. Without an improved fundamental understanding and appropriate modeling and simulation, forecasting of natural gas production from these formations will be inaccurate and unreliable, and operators must continue relying on “trial and error” in their development strategies. With the demand for natural gas continuing to surge, an urgent need, therefore, exists today to improve the efficient drilling and completion of new wells as well as for re-completion of existing wells in order to meet this increased demand. As a result, we not only need to drill and complete more wells, we also need to improve production efficiency on a per well basis. In this study we have characterized the properties of gas shales and the fundamental mass transfer mechanisms associated with the gas production process; we have developed, in addition, a new method to reduce water-blockage related formation damage. A set of techniques have been employed systematically during our shale characterization work, which provide us with ways to better understand the unique physical and chemical properties of these shale materials. These formations, for example, are characterized by a very low and anisotropic permeability, with the horizontal permeability being much higher than the permeability in the vertical direction. Our experiments using BET with these shale samples have revealed that they possess multiple levels of porosity, ranging from microporous and mesoporous regions (characterizing the sample’s matrix) to macropore/microfracture domains. We have also designed and carried out high-pressure gas depletion experiments in order to study the gas production behavior of these shale samples and the various mass transfer mechanisms that prevail. Furthermore, we have utilized the experimental knowledge generated to form the basis of a simple continuum-type model to study gas depletion from these materials. Our high-pressure system was designed in such a way as to mimic xvii the actual formation’s pressure conditions, so that we can independently control the pressure parameters to study the resulting production behavior of the shale core sample. The analysis of the experimental data indicates that the gas that depletes during the early stage is free-gas from the macropore/microfracture domains (as well as the macrofractures that may also be present in these small size, lab-scale samples). The gas depleted at longer times is via a desorption-diffusion-viscous flow process from the shale matrix. We use numerical simulations to analyze the model in order to understand and interpret the experimental behavior we observe during our lab-scale, high-pressure depletion experiments. The combined experimental and numerical studies show that for the shale core sample investigated, the production rate (after the initial production period), as well as the ultimate production potential are both highly dependent on the porous structure of the formation. During hydraulic fracturing, water-blockage related formation damage often takes place caused by water trapped in the formation that blocks the path-way for gas molecules to escape. As part of the study, therefore, we have systematically investigated the effectiveness of a novel surfactant, selected from an extensive survey among commercially available surfactants, in reducing capillarity of the formation. Besides employing contact angle measurements and spontaneous imbibition experiments to study the surfactant, we have also carried out forced imbibition experiments, with which we are able to study the fluid flow-back behavior under the presence of the novel surfactant. We have also conducted static and dynamic experiments in order to investigate the surfactant loss due to the presence of commonly used proppants. Our lab-scale investigation have shown that via the addition of the surfactant, the capillarity of the formation was significantly reduced; this then means that use of this particular surfactant in the field may result in less water-blockage related formation damage. 1 Chapter 1: Introduction 1.1 Importance of Unconventional Natural Gas Natural gas is one of the most important conventional energy resources in the world. As early as 1821, commercial natural gas was available in the United States. In the 1800’s and early 1900’s, lighting street-lamps was the primary function of natural gas [Ground Water Protection Council, 2009]. However, with the ever increasing global demand for hydrocarbon resources, natural gas became more and more important. Natural gas is a gas mixture, which contains methane and other light hydrocarbons such as ethane, propane and butane [Chemistry and Technology of Fuels and Oils. 2000, NaturalGas.org Accessed: September 2008]. Due to the chemical composition and properties of natural gas, it has become one the most efficient and cleanest fossil fuel resources [EIA 2007]. During burning, natural gas emits only half the CO 2 that coal emits, and it also emits lower levels of other air pollutants [EIA 1999]. These properties of natural gas make it a very popular fuel for the generation of electricity [NaturalGas.org Accessed: May 2010] in recent years. For example, in June 2013, the U.S. Federal Energy Regulatory Commission (FERC) reported that 43% of the total 1162 gigawatts (GW) of installed capacity in the U.S. was generated from natural gas [Reuters 2013, Accessed July 2013]. Natural gas is also one of the most favorite choices of fuel among a wide range of industries. For example, the five industries of metals, pulp and paper, food processing, chemicals and petroleum refining, account for around 3/4 of the current industrial natural gas use [EIA 2002], and together they employ over four million people in the U.S. [U.S. Bureau of Labor Statistics, 2007]. Natural gas can also be used to produce a variety of products, such as plastics, chemicals, and fertilizers. In order to meet the economic demand for such gas, the production of unconventional natural gas is becoming progressively more critical. In the past, most of the natural gas industry was focusing on producing gas from conventional reservoirs; this is changing today with the industry focusing more attention, in recent years, on unconventional gas 2 reservoirs/formations. These unconventional gas formations have not, as yet, been appraised by both operators and researchers systematically. For example, unconventional gas resources - including tight sands, coalbed methane, and gas shales - constitute some of the largest components of the remaining natural gas resources in the United States. Both better understanding of formations’ geology and advances in production technologies have enabled numerous operators in the United States to unlock the potential of these challenging resources, boosting production levels to an estimated 30% of the US natural gas production. Around the world, unconventional gas resources are also widespread, but with few notable exceptions, they have not received close attention from natural gas operators. This is partially due to the lack of geologic and engineering information on those unconventional resources; in addition, policies regarding natural gas and market conditions have also been unfavorable for development in many countries. Furthermore, there is a chronic shortage of technical expertise in the specific technologies that are needed in order to develop these resources economically. As a result, only limited Exploration & Production (E & P) has taken place to date outside of North America. However, a number of countries such as Mexico, Russia, and especially China, have begun to extend their footprints into the unconventional gas territory. Though there is no general agreement on the exact size of unconventional gas resources, many of those who have attempted to estimate the volumes of gas in place within unconventional gas reservoirs agree at least on one aspect: that this is, indeed, a large resource [Working Document of the NPC Global Oil & Gas Study]. Table 1.1 (Source: Working Document of the NPC Global Oil & Gas Study) below, is from the study of [Kawata et al., 2001] and shows estimates of the worldwide unconventional gas resources. Based on what has happened in the United States in the last decade, there is good reason to expect that unconventional gas production will increase significantly around the world in the coming decades. 3 Table 1.1: Distribution of Worldwide Unconventional Gas Reservoirs There are a number of reasons which will contribute, in the future, to an increase in the unconventional gas production. These reasons can be summarized [Working Document of the NPC Global Oil & Gas Study] as follows: • A significant number of geologic basins around the world contain unconventional gas reservoirs. For example, Rogner [1997] estimates that in the world there are around o 9,000 Tcf of gas in place in coalbed methane formations, o 16,000 Tcf of gas in place in shale-gas reservoirs, and o 7,400 Tcf of gas in place in tight-gas sands. • Any reasonable recovery efficiency leads one to the conclusion that there will be an ample opportunity in the future to develop unconventional gas worldwide. 4 • Tight-gas sand development in the United States, critical to the future U.S. gas supply, has totaled over 4 Tcf/year, and is supported by ongoing technological development. • The technology developed in the United States over the past 3 to 4 decades is now available for application around the world. • New technology is rapidly becoming a worldwide commodity through efforts of major service companies. • The global need for energy, particularly natural gas, will continue to be an incentive for worldwide unconventional gas resource development. • Tight-gas sands, gas shales, and coalbed methane are already critical to North America today, and will be an important energy source worldwide during the 21st Century. The above discussion clearly indicates that unconventional natural gas is a very critical energy resource for the future. In order to realize the potential of unconventional natural gas, we need to know more about the complex subsurface structures that store the gas, and to develop better means to recover it. Unconventional natural gas is produced from low-permeability reservoirs/formations that produce mainly dry natural gas. Such low- permeability reservoirs include sandstone-type formations, low-permeability carbonates, gas shales, and coalbed methane reservoirs which are all being produced nowadays [Working Document of the NPC Global Oil & Gas Study]. In unconventional gas shale and coalbed methane reservoirs, the gas is often generated from the reservoir rock itself [Ground Water Protection Council, 2009]. However, due to the low permeability of such reservoirs, operators need to stimulate the newly drilled wells in order to introduce extra permeability into the formations. Hydraulic fracturing of a reservoir is currently the preferred stimulation method for gas shales. The following is a brief overview of the three basic types of unconventional reservoirs: 1. Tight Gas –Tight gas is the type of gas which can be found in a very tight formation underground; the gas usually is trapped within very low-permeability hard rocks, or it can also be stored in low-permeability and low-porosity sandstone or limestone formations 5 [NatrualGas.org, Accessed May 2010]. According to the Energy Information Administration (EIA), 309.58 Tcf of technically recoverable tight gas exists in the U.S as of January 2009, which represents over 17% of the total recoverable natural gas in the United States. 2. Shale Gas– This gas is produced from low-permeability shale formations that are also the source for the natural gas. The shale-gas can be stored as free-gas both in the macroporous regions and/or the micro/mesoporous regions of the shale [Frantz and Jochen, 2005], and it can also be stored as adsorbed gas within the shale formation [Geology.com, Accessed May 2010]. Due to their extremely low permeability, hydraulically fracturing needs to be introduced in order to stimulate production in such formations. Although shale formations contain large amounts of natural gas, due to present technological limitations, operators are only able to produce ~10% of the gas in place. Despite the difficulty to produce the gas, the current potential of shale-gas reservoirs as an important natural gas resource is still very promising. As of November 2008, FERC estimated that there is 742 Tcf of technically recoverable shale gas in the United States [NatrualGas.org, Accessed May 2010]. 3. Coalbed Methane – This type of gas is produced from coal seams which act as source and reservoir of the natural gas [ALL Consulting and the Montana Board of Oil and Gas Conservation, 2004]. The natural gas in these formations can be generated via two processes: 1) biogenic action of indigenous microbes on the coal, 2) thermogenic alterations of coal. The Potential Gas Committee estimates that there is 163.0 Tcf of technically recoverable coalbed methane in the United States [NatrualGas.org, Accessed May 2010]. The above mentioned unconventional natural gas has been and will continue to be a major contributor for the future North American gas production - with unconventional gas production expected to increase from 42% of total US gas production in 2007 to as much as 64% in 2020 [Energytomorrow.com, Accessed May 2010]. 6 1.2 The Shale-Gas Industry Operators have drilled more than 28,000 shale gas wells in the United States dating back to the 1800’s [Hill et al., 2000], with a shallow well drilled in 1821 marking the beginning of the commercial natural gas industry [Peebles, 1980]. As pointed out by [Hill et al., 2000, Working Document of the NPC Global Oil & Gas Study]: “The Devonian Antrim Shale of the Michigan basin, the most active United States natural gas play in the 1990’s, became commercially productive in the 1980’s, as did the Mississippian Barnett Shale of the Fort Worth basin and the Cretaceous Lewis Shale of the San Juan basin”. Shale-gas production started to surge after 2000, as reported by [DOE/EIA-0216 (98)]: “In 1998, shale-gas reservoirs supplied 1.6% of the total United States dry-gas production, and amounted to 2.3% of the proven natural gas reserves.” Figure 1.1: Commercial Shale-Gas Production in the USA The following basins in the United States have been marked as active producing areas of shale-gas: Appalachian, Michigan, Illinois, Fort Worth and San Juan, see Figure 1.1 [Faraj et al., 2004]. The production increase, starting with the year 2000, is quite dramatic, as Figure. 1.2 [Sieminski, 2013] indicates. 7 Figure 1.2: Shale-Gas Production Growth by EIA. 1.3 Current Practice for Shale-Gas Production As mentioned above, the very low-permeability of gas shales presents a “bottleneck” in the commercial production of gas from such formations. However, three factors have come together in recent years to make shale-gas production economically viable: 1) Improved technology in horizontal drilling, 2) Improved technology in hydraulic fracturing, 3) A huge gap between supply and demand of natural gas. The combination of horizontal drilling and fracturing techniques make the cost-effective production of shale-gas possible [Ground Water Protection Council, 2009]. From the rig- counts of horizontal rigs, one is able to observe the transition from conventional onto unconventional production. According to EIA [2008], in the late 1990’s, about 40 drilling rigs, or 6% of total, were drilling horizontally. The number of rigs drilling horizontal wells has grown to 519, as of May 2008, or 28% of the total. And as reported by [WTRG Economics, Accessed July 2013], as of July 5 th , 2013, the number of horizontal rigs in 8 U.S. increased to 1068 rigs while the total U.S. rig counts is 1757, which means the horizontal rigs amount to 61% of the total number of rigs. In fact, even with the help of horizontal drilling and hydraulic fracturing, the shale-gas industry is still trying to reach a more ‘cost-effective’ state. With 13,902 gas wells drilled as of May 10, 2010 [Railroad Commission of Texas, 2010], the Barnett Shale has become one of the most prominent shale-gas plays in the U.S. [Working Document of the NPC Global Oil & Gas Study]. Parshall [2008] also pointed out that the commercial efforts in the Barnett shale have been named a show-case for the modern tight-reservoir development activities in the United States. However, even for the Barnett shale wells, using currently available technologies, the average recovery rate per well is around 7% of the total gas in place, which is far below a potentially achievable 20% recovery factor [Working Document of the NPC Global Oil & Gas Study]. In order to push the shale-gas reservoir performance to a higher level, researchers have put their focus on the following technology aspects [Working Document of the NPC Global Oil & Gas Study]: “ 1) Fracture modeling and analysis; 2) New fracturing fluids and proppants; 3) Hydraulic fracturing methods used in horizontal wells; 4) Stimulation methods used in naturally fractured formations; 5) Micro-seismic fracture mapping and post-fracture diagnostics; 6) Data collection and availability during drilling, completions, stimulations, and production; 7) Reservoir characterization through laboratory measurements; 8) Reservoir imaging; 9) Produced-water handling, processing, and disposal; For the purpose of cost-effective production of shale gas, all the above mentioned processes need to be integrated, and no single process can be optimized alone without the 9 basic understanding of the characteristics of the formation. Hence, understanding the fundamental processes involved in the gas storage and transport, is critical for field operators to fully leverage the advantages of technologies and to increase shale-gas production more efficiently and effectively. 1.4 Dissertation Objectives Shale gas is considered to be a very promising alternative to the conventional hydrocarbons as a future clean and efficient energy source. However, there are a number of technical barriers which have prevented natural gas production from gas shales from becoming economically viable. Such barriers include: 1) Lack of basic understanding of the characteristics of gas shale reservoirs including lab-scale experimentation/characterization of relevant shale samples. 2) Lack of lab-scale investigation to reduce formation damage during hydraulic stimulation processes 3) Lack of modeling/simulations which link reservoir behavior with lab-scale data to promote shale-gas production. In other words, it would be very difficult to improve the production potential of shale-gas formations without overcoming the above barriers. Therefore, the main objective of the research described in this Thesis is to meet the technical challenges associated with gas production from gas shales by specifically targeting all the above technical barriers. By combining various techniques, we aim to characterize both the micro-structure of Appalachian Basin shale rock and the dynamics during release of the gas found within it. Furthermore, parameters for the mathematical models will be extracted and numerical simulations will be used to validate our findings. By providing the industry with more reliable models, predictive tools can be developed to reduce exploration and development costs, and also to increase recovery on a per-well (and per field) basis. To this end, a lab-scale study has been designed and conducted in 10 order to better address the water-blockage related formation damage issue during the hydraulic stimulation processes. 1.5 Organization of Dissertation In Chapter 2, a systematic shale sample characterization protocol is described, which includes employing a combination of various techniques to study the physical and chemical properties of Marcellus Shale samples. The characterization results are discussed and a cross-reference system with all found characteristics of those samples is also provided. In Chapter 3, both the general experimental methodology and the design of a high- pressure depletion system (HPDS) to study mass transfer mechanisms during gas production are described. A detailed description of the functionality of the HPDS and its operational procedure are provided. Both short-term and long-term depletion experiments with HPDS are presented, and the production behavior of these experiments is analyzed. A simple numerical model is constructed to describe the mass transfer mechanisms based on the physical characteristics of the sample and the production behavior observed from these depletion experiments. The numerical model is then employed to predict the production behavior and to compare with the production data obtained from the long- term depletion experiments. In Chapter 4, a lab-scale investigation aiming at reducing formation capillarity with a novel surfactant is presented. Both contact angle experiments and spontaneous imbibition experiments with and without the novel surfactant are presented and compared. A forced imbibition system is utilized to analyze the fluid flow-back behavior in the presence of the surfactant. The static and dynamic experiments to study surfactant loss to proppant are also presented in this Chapter. These experimental results are then discussed to provide new insight for potential field trials. Finally, the findings and contributions of this dissertation are summarized and recommendations for future work are presented in Chapter 5. 11 Chapter 2: Characterization of Gas-Shale Materials 2.1 Overview of Characterization Techniques A major limitation for producing gas from gas-shale materials is the lack of knowledge of the physics and mass transfer properties of these low-permeability porous media. Without such knowledge, modeling of gas production from such formations will be inaccurate and unreliable, and operators must then rely on trial and error. Unfortunately, no single technique alone can provide a complete characterization of the structure of gas shales. However, by combining various techniques we aim to characterize, in this Chapter, both the micro-structure of Appalachian Basin shale rock and the behavior of the gas found within it. In order to provide a deeper insight into the shale’s structure, we need to use different techniques, each for a specific aspect, and then combine all the results to form a more complete understanding of the shale samples. Hence, in our study we have utilized a number of techniques to gather the characterization data that we need. The relevant data include: • Specific surface area (SSA) and pore size distribution (PSD) via gas sorption (BET) • Topography of rock sample via Scanning Electron Microscopy (SEM) • Elemental information via Energy-Dispersive X-ray Spectroscopy (EDX) • Matrix permeability data • Matrix porosity data • Micro-fracture information via SEM • Mineralogy information via X-ray Diffraction (XRD) • Shale core permeability under varying confining pressure via high-pressure flow experiments • Shale core porosity under varying confining pressure via high-pressure depletion experiments 12 The technical information generated using all the above techniques helps to provide additional insight into the fundamental mechanisms that control the gas production from shale. The permeability and porosity data, for example, directly relate to the gas mass transfer characteristics of the shale. And the permeability measurements reveal the efficiency of flow-paths within the formation that allow for gas to propagate towards a production well from the matrix. In fact, for a gas producing well, the gas molecules need to desorb from the pore surface, diffuse through the porous structure and flow through the fracture networks in order for the gas to be produced. Since the mechanism of gas travelling through the fracture networks is viscous flow, gas desorption and diffusion in the porous structures are likely to be the rate limiting processes. As a result, the matrix porous structure directly affects the long term production performance of a well. If we can establish a connection between the characteristics of the porous structure and the production behavior observed in the laboratory, we may be able to guide the operational work in the field. Hence, high-pressure laboratory depletion experiments are valuable in terms of evaluating the long-term production performance of a shale-gas well. (These high-pressure depletion experiments will be addressed in a greater detail in Chapters 3 of this Thesis). Although permeability and overall porosity measurements are important in terms of understanding the overall mass transfer mechanisms qualitatively, they are still not adequate for building a complete physical model to investigate mass transfer mechanisms quantitatively. In order to obtain a more detailed physical model, PSD tests have been carried out and have been interpreted in terms of a pore scale model, which then forms the basis of evaluating the mass transfer characteristics in the porous structures as well as the sorption behavior of the shale [Seaton et al., 1989; Liu et al., 1992; Ghassemzadeh et al., 2000]. During the experiments, equilibrium and dynamic sorption were performed with pure Nitrogen. The observed behavior provides important data required in order to validate our modeling efforts. SEM provides a direct visualization of the porous structure of the shale samples, and EDX allows one to obtain the elemental composition of samples. SEM and EDX data can 13 be combined with mineral information and organic composition to better understand the physiochemical properties of gas shales. In this study, the samples for characterization were supplied by the Energy Corporation of America (ECA). We have studied a total of 18 samples from a Marcellus well from ECA each from a different depth. Table 2.1 provides a summary of the various samples analyzed from the various depths, the characterization techniques utilized and the type of information generated. Table 2.1: Sample Depth and Description Experiment type Information generated Physical state of sample sample depth(feet) BET Specific surface area Ground sample 7721.5 7813.5 7732.5 7823.5 7742.5 7832.5 7750.5 7841.5 7762.5 7852.5 Pore size distribution 7772.5 7862.5 7781.5 7873.5 7792.5 7885.5 7802.5 7891.5 SEM and EDX Micro fracture networks Shale cube 7721.5 7813.5 7732.5 7823.5 7742.5 7832.5 Pore size distribution 7750.5 7841.5 7762.5 7852.5 7772.5 7862.5 sample composition 7781.5 7873.5 7792.5 7885.5 7802.5 7891.5 XRD Mineralogy Shale cube 7721.5 7813.5 7732.5 7823.5 7742.5 7832.5 7750.5 7841.5 7762.5 7852.5 7772.5 7862.5 7781.5 7873.5 7792.5 7885.5 7802.5 7891.5 14 Sections 2.2 through 2.5 below provide a detailed description of the various techniques we have utilized and the data we have obtained from using these techniques. Section 2.6 provides a summary and a discussion of all the data and provides a road-map for our lab- scale experiments. 2.2 Specific Surface Area and Pore Structure Analysis The most widely used theory to characterize sample surface area is the BET [Brunauer et al., 1938] method. The surface area can be determined by the following BET equation: 0 m 0 V 1 - 1 - P P C C C V P P V P m         + = ）（ , (2.2-1) with ( 0 P P ) falling in the range of 0.05 to 0.3. In the BET equation, m V is the monolayer coverage, mol/g, V is the total gas adsorbed at a given pressure, mol/g, 0 P is the vapor pressure, Pa, P is the adsorption pressure at equilibrium with the surface at 77.4 K, Pa, and C is the BET constant. A plot of ）（ P P V P - 0 versus 0 P P yields a straight line, and the slope of the line =         C C m V 1 - while the intercept of the line = C V m 1 . m V can then be calculated from the slope, and the surface area can then be calculated by the following form: A N V S A m = , (2.2-2) where S is the surface area of adsorbate, m 2 /g, N A is the Avogadro number and A is the cross-sectional area of the molecule. The sample’s average pore diameter and the cumulative volume of pores and the pore size distribution in the mesopore (>2 nm) and macropore regions (>50 nm) were determined by the BJH method [Barrett et al., 1951] from the adsorption and desorption data. The BJH method is based on the assumption that the pores are cylindrical and the 15 pore radius equals to the sum of the Kelvin radius of the pore and the thickness of the adsorbed film. This can be expressed as: t k p r r r + = , (2.2-3) where p r is the pore radius, k r is the Kelvin radius of the pore and t r is the thickness of the adsorbed film. The Kelvin radius (in the unit of Å) can be calculated by: ) log( 15 . 4 0 P P r k = , (2.2-4) while the thickness (in the unit of Å) of the adsorbed film for liquid N 2 can be calculated by: m a t W W r τ = , (2.2-5) where a W is the weight adsorbed, m W is the weight for monolayer coverage and τ is the thickness of one layer. We have also employed the Horvath-Kawazoe (HK) method [Horvath et al., 1983] to characterize micropores (<2nm). The HK method is a widely used method for determining pore-size distribution in microporous materials. The assumption of the HK method is that the pressure at which pore filling occurs may be obtained from the mean free energy change of the adsorbate molecule as it is transferred from the bulk gas phase to the adsorbed phase, and can take the following form [Dombrowski et al., 2006]:                 − +         − −         −                 − + =         3 0 9 0 3 0 9 0 3 0 0 3 1 9 1 3 1 9 1 * ) 2 ( ln sf sf sf sf sf ff f sf s HK H d H d d d H kTd A N A N P P σ σ σ σ σ , (2.2-6) with 16 sf d σ 85 . 0 0 = , (2.2-7) Here, HK P is the condensation pressure; 0 P is the bulk saturation pressure; s N is the number of atoms of adsorbent per unit surface area, f N is the number of atoms of adsorbate per unit surface area; sf A is the dispersion constant representing adsorbent- adsorbate interactions; ff A is the dispersion constant representing adsorbate-adsorbate interactions; H is the pore width in the slit pore model, and sf σ is the mean diameter of adsorbent and adsorbate molecule. Nitrogen absorption and desorption tests, using a Micrometrics ASAP 2010 Instrument (Figure 2.1), were used to characterize the obtained core sample (the core samples was ground into a fine powder prior to the analysis). The analysis procedure basically consists of two parts, a) the degassing stage and b) the analysis parts. In the degassing step each ground sample (with a mass ranging from 100 mg to 1 g) is pretreated at 110 o C in vacuum under 500 mmHg for at least 16 hr. This step is done prior to N 2 adsorption to remove all moisture from the pores. The analysis is fully automated and controlled by a personal computer (PC) that collects and processes the measurements. Figure 2.1: Micrometrics ASAP 2010 Instrument and PC 17 Figure 2.2 shows the measured BET specific surface area versus depth for all our 18 samples from the various depths while the specific surface areas for each of the pore types (micropore, mesopore and macropore) are listed in Table 2.2. The data in Figure 2.2 indicate that for the most part the surface area of samples increases with depth, i.e., with samples from greater depths (e.g., 7881.5 and 7891.5 ft) having higher surface areas than the samples from the shallower depths. The data in Table 2.2 indicate the increase in the specific surface area with depth is primarily due to the increase in the surface area in the micropore region. Figure 2.3 shows the “desorbed gas” analysis results with the same samples (These tests were performed at CoreLab; the measurements are made using the U.S. Bureau of Mines (USBM) standard lost-gas method to measure the lost-gas and the gas content, shown in Figure 2.4, [GRI-95/0496, 1996]). The fraction of lost gas (percent) appears to be again an increasing function of sample depth, however, the trend is not as clear as with the SSA data from BET. Figure 2.4 shows the gas content data (scf/ton) for the same samples, again performed by CoreLab. (The gas content measurement requires one to measure both the lost gas and the remaining gas. The lost gas is estimated by projecting the first few hr of desorption measurements back to the time at which desorption begins, which is done by putting the core in a canister and measuring how much gas coming out. And residual gas can be measured by grinding the sample into fine power, and measuring the gas desorption of the powers, a process that may require two or more months to be completed). Again the trends in Figure 2.4 are very similar to those observed in Figures 2.2 and 2.3. The data in Figures 2.2 – 2.4 are all consistent in showing that for this particular formation the amount of gas stored increases with increasing depth. As a result, the expectation is that the deeper depths will have a higher production potential, and this hypothesis should hold true for the nearby locations to the well from which the cores were taken unless there is a dramatic change on the geological structure. Data of this kind are very helpful in terms of potentially helping the operators to pinpoint the “sweet spots” for placing their wells in order to get access to the most profitable zones, but also to provide the information needed in order to evaluate the long-term production potential for the formation. 18 Figure 2.2: BET Specific Surface Area vs. Sample Depth Figure 2.3: Lost Gas (percent) vs. Sample Depth 19 Table 2.2: Specific Surface Area Comparison Figure 2.4: Total Gas Content (scf/ton) vs. Depth Specific Surface Area for Meso & Macropores (m 2 /g) Specific Surface Area for Micropores (m 2 /g) Micropore Specific Surface Area/Total Specific Surface Area (%) Depth(ft) 7721.5 7732.5 7742.5 7750 7762.5 7772.5 7781.5 7.83 1.27 13.95604396 7792.5 16.97 7.45 30.50778051 7802.5 14.47 7.61 34.46557971 7813.5 12.855 4.815 27.24957555 7823.5 14.06 3.89 21.67130919 7832.5 15.91 6.87 30.15803336 7841.5 17.93 8.99 33.39524517 7852.5 19.76 13.77 41.06770057 7862.5 14.48 12.68 46.68630339 7873.5 13.005 27.335 67.76152702 7885.5 20.145 20.145 50 7891.5 20 Figure 2.5 shows the average pore diameter (includes micropores, mesopore and macropores) of the samples. The average diameter decreases with increasing depth, which is consistent with the data in Figure 2.2 that show the samples from the deeper depths have a higher specific surface area. Matrix permeability measurements (by CoreLab) and further discussed in Sec. 2.5 indicate that the matrix permeability increases with increasing depth, which is contradictory to the decreasing average pore size data of Figure 2.5. This suggests then that the pore structure at the deeper depths may be more interconnected or that macroporosity is present that is not accessed by the BET measurements. Figure 2.5: Average Pore Diameter vs. Depth 21 Figure 2.6 (a, b and c) shows measurements of the cumulative pore volume versus pore diameter calculated from the adsorption data based on the BJH method for three different samples (from depths 7742.5, 7781.5, and 7852.5 ft). Figure 2.7 (a, b and c) reports mercury injection data with the same samples (carried out at CoreLab). These two sets of data (per sample) using two different technique are qualitatively similar and indicate that most of the mesopore/macropore volume is occupied by pores with average size of <100 nm. The various cumulative pore volumes for all 18 samples measured by the BET technique and analyzed by the HK and BJH methods corresponding to the micropore (d<2nm), the mesopore (2 nm<d<50nm) and macropore (50 nm<d<500nm) regions as well the total core porosity based on the measured total pore volume are shown in Table 2.3. Figure 2.8 plots the micropore volume vs. depth, Figure 2.9 plots the mesopore volume vs. depth, Figure 2.10 plots the macropore volume vs. depth, while Figure 2.11 plots the total pore volume vs. depth. One notices the volume of micropores increase significantly (the volume almost doubles, from around 0.008 cm 3 /g at 7721.5 ft to around 0.015 cm 3 /g at 7891.5 ft -- the reason for the increase in the amount of micropores is highly related to the increase in the amount of illite type of clay of our samples (for further details see Section 2.4 of this Chapter) with depth while the volume of mesopores stays relatively the same, ~0.020 cm 3 /g, on the average. The volume of the macropores also increases with depth, but its magnitude is not comparable with the other two types of pores (as the mercury injection data also indicate, see Figure 2.7), as a result, the increase in the total pore volume is primarily attributed to the increasing micropore volume. These data along with the specific surface area data in Table 2.2 strongly indicate that gas stored in the micropores is a key gas source for production especially for the samples from the deeper depths, together with the gas found in the mesopores. Depletion from the shale matrix can thus be described by the bidisperse pore model (BPM) [Ruckenstein et al., 1971] which will be further described in our modeling section in Chapter 3. Figures 2.12 show the cumulative pore volume (based on the BJH analysis of the data, defined in these Figures as the volume occupied by pores with a diameter which is larger than the given pore diameter corresponding to that volume) for the mesopores and the macropores for a select number of the shale samples (due to the similarity in terms of 22 behavior among the remaining samples, the rest of figures for other depths are included in Appendix A-1). Figure 2.13 and 2.14 show the PSD (in terms of dV/dD vs. pore diameter) for the mesopores and the macropores for a number of the shale samples (the PSD for the remaining samples can be found in Appendix A-2). Figures 2.15 and 2.16 show the cumulative pore volume (based on the HK analysis of the data, defined in these Figures as the volume occupied by pores with a diameter which is smaller than the given pore diameter corresponding to that volume) for the micropores of a number of the samples (the rest can be found in Appendix A-3). Looking at Figures 2.12 we observe there is a steep change in the slope at a pore size equals to 4 nm, moreover, from Figures 2.13-2.14, it is obvious that this is the prevailing pore diameter for the mesopores. Looking at Figures 2.15 and 2.16, we observe with depth goes lower, the density of finer micropores increase, which is also an indication that with deeper depth, more gas is going to be stored in the microporous region. 23 (a) Depth 7742.5ft (b) Depth 7781.5ft (c) Depth 7852.5ft Figure 2.6: Cumulative Pore Volume from BJH Adsorption 24 (a) Depth 7742.5ft (b) Depth 7781.5ft (c) Depth 7852.5ft Figure 2.7: Cumulative Intrusion from Mercury Injection Measurements 25 Table 2.3: Pore Volume Comparison via HK and BJH HK micropore volume BJH mesopore and macropore volume Cumulative micropore (<2nm) volume (cc/g) Cumulative mesopore(2- 50nm) volume (cc/g) Cumulative macropore(50 -500nm) volume(cc/g) Total pore volume (cc/g) Core porosity (%) Depth(ft) 0.0078 0.013599 0.002146 0.023545 5.768525 7721.5 0.01083 0.000739 0.011569 2.834405 7732.5 0.010866 0.000922 0.011788 2.88806 7742.5 0.020722 0.000829 0.021551 5.279995 7750 0.022027 0.000793 0.02282 5.5909 7762.5 0.023632 0.001987 0.025619 6.276655 7772.5 0.043501 0.00043 0.043931 10.763095 7781.5 0.014953 0.000927 0.01588 3.8906 7792.5 0.0081 0.025529 0.000973 0.034602 8.47749 7802.5 0.0081 0.024667 0.000597 0.033364 8.17418 7813.5 0.0062 0.018399 0.001931 0.02653 6.49985 7823.5 0.0063 0.021651 0.00206 0.030011 7.352695 7832.5 0.0089 0.019114 0.004288 0.032302 7.91399 7841.5 0.0083 0.022375 0.00275 0.033425 8.189125 7852.5 0.0126 0.019301 0.004085 0.035986 8.81657 7862.5 0.0108 0.017554 0.002172 0.030526 7.47887 7873.5 0.0156 0.012594 0.003593 0.031787 7.787815 7885.5 0.0148 0.021214 0.002464 0.038478 9.42711 7891.5 26 Figure 2.8: Micropore (d<2nm) Volume vs. Depth Figure 2.9: Mesopore (2<d<50nm) Volume vs. Depth 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 7700 7750 7800 7850 7900 Pore V olume (cc/g) Depth (ft) 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0.05 7700 7750 7800 7850 7900 Pore V olume (cc/g) Depth (ft) 27 Figure 2.10: Macropore (50<d<500nm) Volume vs. Depth Figure 2.11: Total Pore Volume vs. Depth 0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004 0.0045 0.005 7700 7750 7800 7850 7900 Pore V olume (cc/g) Depth (ft) 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0.05 7700 7750 7800 7850 7900 Pore V olume (cc/g) Depth (ft) 28 (a) Depth 7721.5 ft (b) 7732.5 ft (c) Depth 7742.5ft (d) Depth 7750 ft Figure 2.12: Cumulative Pore Volume from BJH Desorption (Depth 7721.5-7750.5ft) 29 (a) Depth 7721.5 ft (b) 7732.5 ft (c) Depth 7742.5ft (d) Depth 7750.5ft Figure 2.13: dV/dD vs. Pore Diameter from BJH Desorption (Depth 7721.5 – 7750.5ft) 30 (a) Depth 7862.5 ft (b) 7873.5 ft (c) Depth 7885.5ft (d) Depth 7891.5ft Figure 2.14: dV/dD vs. Pore Diameter from BJH Desorption (Depth 7841.5ft - 7891.5ft) 31 (a) Depth 7772.5ft (b) Depth 7802.5ft (c) Depth 7813.5ft (d) Depth 7823.5ft Figure 2.15: Cumulative Pore Volume from HK Method (Depth 7772.5ft – 7823.5ft) 32 (a) Depth 7862.5ft (b) Depth 7873.5ft (c) Depth 7885.5ft (d) Depth 7891.5ft Figure 2.16: Cumulative Pore Volume from HK Method (Depth 7852.5ft – 7891.5ft) 33 2.3 Scanning Electron Microscope (SEM) Scanning microscopy is a local chemical and textural characterization technique for solid materials which is based on the interaction of these materials with a focused electron beam with energies between 0.5 and 35 kV. Electrons and radiation emitted back from the sample as the result of the impact of the electron beam (secondary electrons, scattered electrons, and X-rays) are used to form images showing the various properties of the material (including topography, heterogeneities in composition, and local elemental composition) [Lynch et al., 2003]. A key observation from our imaging efforts with these shale samples is that there are significant differences between the SEM images in the vertical (along the axis) and horizontal (perpendicular to the axis) core faces. For example, Figure 2.17 shows the vertical and horizontal SEM images (2000x) for a sample from the depth of 7772.5ft. The most noticeable aspect that distinguishes them is that there are many more pores showing in the horizontal image than in the vertical image. These differences in the SEM images are also consistent with significant differences between the vertical and horizontal permeabilities measured with 1 cm 3 shale cubes by our Group [Roychaudhuri et al., 2013], and discussed in further detail in Sec. 2.5. (We note that other available horizontal and vertical SEM images show the same phenomena as what Figure 2.17 presents, and can be found in Appendix A). So, in conclusion horizontal and vertical SEM images are useful in helping us understand the mechanisms of gas flow through the matrix as well the microfracture network of the formation. 34 Figure 2.17: Vertical (left) and Horizontal (right) SEM Images from 7772.5 ft (2000X) We have also found the SEM technique to be useful in studying the effectiveness of acid treatment processes during a frac job. For example, Figure 2.18 shows the SEM images of three different samples from the same depth of 7792.5ft: A reference untreated sample (a), a sample that was exposed to a 3vol.% HCl solution for 1 hr (b), and a sample exposed to fresh water (c). From these three images, it is clear that treatment in an acidic (3vol.% HCl) solution results in the dissolution of some of the minerals, and thus creates additional porosity and permeability. Treatment in the fresh-water environment appears to also dissolve some of the minerals, lthough significantly less so than the acidic solution. (a) Sample without Treatment (c) Sample with Fresh Water Treatment Figure 2.18: Vertical V (a) Sample without Treatment (b) Sample with Low pH Fluid Treatment ater Treatment Figure 2.18: Vertical View of Samples after Different Treatments (1000X) 35 ow pH Fluid Treatment (1000X) 36 2.4 Elemental Analysis and Mineralogy of Gas-Shale Materials Energy Dispersive X-ray Spectroscopy (also known as EDS, EDX or XEDS) is an X-ray micro-analytical technique that can provide information on the chemical composition of a sample for elements with atomic number (Z) >3 [AMMRF 2012]. During the analysis, an electron beam in either a SEM or a transmission electron microscopy (TEM) instrument is focused on the sample resulting in X-ray emissions from the electrons from the primary beam penetrating the sample and interacting with the atoms from which it is made Figures 2.19-2.21 are EDX analysis and SEM vertical images of three samples with the highest matrix permeability (see Figure 2.33 in Sec. 2.5). Figures 2.22-2.24 are EDX analysis and SEM vertical images of three samples with the lowest matrix permeabilities. The EDX results clearly show that the samples with the higher carbon content have higher matrix permeabilities. (The rest of the EDX figures present similar result and have been documented in Appendix B). The EDX results are consistent with the total organic content (TOC) data for these samples, which are shown in Figure 2.25 for all 18 samples. Although EDX analysis measures the elemental composition, it is not able to fully describe the mineralogical content of the samples. Hence, in this study we also use X-ray Diffraction (XRD) to help us better understand the sample’s mineralogy. In an X-ray diffraction measurement, a sample is mounted onto a goniometer and is gradually rotated while being bombarded with X-rays, producing a diffraction pattern of regularly spaced spots known as reflections. The two-dimensional images taken at different rotations are converted into a three-dimensional model of the density of electrons within the sample using the mathematical method of Fourier transforms, combined with chemical data already known for the sample. A complete list of minerals and their abundance within the sample is then generated. Figures 2.26 – 2.30 show the clay content, calcite content, quartz content, plagioclase content and the pyrite content for various samples from different depths. From these Figures, it is clear that there are definite trends that can be recognized. For example, the clay content decreases with depth, the calcite, quartz, and pyrite contents all increase with 37 depth, while the plagioclase content seems not to have a strong depth dependence. Compared to the EDX and TOC data, we note that the carbon content shares a similar trend with calcite, quartz and pyrite content. The various mineral contents of Figures 2.26 -2.30 from tis section are summarized into Figure. 2.31 (provided by CoreLab). And Figure 2.32 (also provided by CoreLab) presents the amount of different clay types vs. depth. Comparing Figure 2.31 and Figure 2.32, one notices that although the total clay content decreases with depth, the ratio of clay groups (illite & mica/chlorite) increases. Kuila et al. [2013] proposed a theory that the amount of illite type clay is dominant in microporous regions, meaning that the more abundant the illite type clay group is, the more micropores can be found. If we combine our micropore volume from section 2.2 with Figure 2.32, we notice the same observation as made by Kuila et al. [2013]. As a result, from a field operator’s perspective, if there is an increasing amount of illite, it is likely that the formation will have a higher amount of micropores, which further suggests that there will be a higher ratio of adsorbed gas to free gas. 38 Figure 2.19: EDX Analysis and SEM Image for Vertical View Sample with Reported Permeability at Depth 7891.5 ft (2000X): k = 2.54E-04 mD Figure 2.20: EDX Analysis and SEM Image for Vertical View Sample with Reported Permeability at Depth 7862.5ft (2000X): k=1.58E-04 mD 39 Figure 2.21: EDX Analysis and SEM Image for Vertical View Sample with Reported Permeability at Depth 7885.5 ft (2000X): k = 4.65E-05 mD Figure 2.22: EDX Analysis and SEM Image for Vertical View Sample with Reported Permeability at Depth 7732.5 ft (2000X): k = 1.77E-09 mD 40 Figure 2.23: EDX Analysis and SEM Image for Vertical View Sample with Reported Permeability at Depth 7742.5 ft (2000X): k = 1.16E-08 mD Figure 2.24: EDX Analysis and SEM Image for Vertical View Sample with Reported Permeability at Depth 7721.5 ft (2000X): k = 1.22E-08 mD 41 Figure 2.25: TOC Analysis Results vs. Sample Depth Figure 2.26: Total Clay Content vs. Sample Depth via XRD 42 Figure 2.27: Calcite Content vs. Sample Depth via XRD Figure 2.28: Quartz Content vs. Sample Depth via XRD 43 Figure 2.29: Plagioclase Content vs. Sample Depth via XRD Figure 2.30: Pyrite Content vs. Sample Depth via XRD 44 Figure 2.31: Minerals vs. Depth Figure 2.32: Relative Clay vs. Depth 45 2.5 Permeability and Porosity Measurements Matrix permeability and porosity data for these samples are available from Corelab and directional permeabilities of intact shale core samples were previously measured by our Group [Roychaudhuri et al., 2013]. The matrix permeability measurements are based on the GRI approach [GRI-95/0496, 1996] that uses the pulse-decay method with ground samples (20/35 mesh size). The porosity measurements are made via the GRI approach using helium with core plugs (CoreLab did not specify the dimensions or other information about these core plugs). The directional permeability of intact shale samples was measured with He flow-through experiments [Roychaudhuri et al., 2013]. Experiments were carried out with whole (not ground) shale samples, with a volume of approximately 1 cm 3 . The samples were aligned so that flow would occur either along their horizontal (high-permeability) or their vertical (low-permeability) direction. A steady-state permeation method was adopted for measuring the permeability of samples along their horizontal direction, while a transient technique was utilized for measuring the permeability of samples along the vertical direction [Roychaudhuri et al., 2013]. Figure 2.33 shows the matrix permeability vs depth. Figure 2.34 is the core plug porosity vs. depth. It is obvious in Figure 2.33 that there is a general trend (if one ignores the three outliers, indicated in red circle, for samples from depths of 7750.50, 7762.50 and 7772.50 ft) in the variation of the matrix permeability with respect to depth. The deeper the sample, the higher the matrix permeability appears to be. In addition, Figure 2.34 shows the core-plug porosity vs. depth appears to have a very similar trend as matrix permeability. Figure 2.35, which is reproduced here from the study of [Roychaudhuri et al., 2013], shows the horizontal permeabilities vs. depth, while Table 2.4 (which is also reproduced here from the same study) reports the vertical permeabilities for the shale core samples. (For the correspondence between Sample # and depth, see Table 2.5). Comparing the directional permeability data in Figure 2.35 and Table 2.4, it is obvious that the permeabilities in the horizontal direction are over two orders of magnitude higher than the permeabilities in the vertical direction, something which is also consistent with our 46 SEM directional images (Figure 2.17 and SEM images in Appendix A) qualitatively. In Figure 2.36, which shows the core-plug permeability vs confining pressure for the core sample from 7880.5ft, from the figure, one observes that the permeability does not change with net-stress, a somewhat unexpected result. Our lab scale observation might be due, however, to the fact that we conducted the core plug permeability vs. confining pressure measurements after the completion of the high-pressure depletion experiment, by which time the core may no longer be “shrinkable”. Figure 2.33: Matrix Permeability vs. Depth 47 Figure 2.34: Shale Core Plug Porosity vs. Depth Figure 2.35: Horizontal Permeability Measured by He Flow-Through Experiments 48 Figure 2.36: Permeability vs. Net Stress Table 2.4: Vertical Permeability Measured by He Flow-Through Experiments 0 0.5 1 1.5 2 2.5 3 0 500 1000 1500 2000 2500 3000 Permeability (md) Net Stress (psig) 49 Table 2.5: Sample # and Depth for Directional Permeability Measurement Sample # Depth (ft) 1 7721.5 2 7732.5 3 7742.5 4 7750 5 7762.5 6 7772.5 7 7781.5 8 7792.5 9 7802.5 10 7813.5 11 7823.5 12 7832.5 13 7841.5 14 7852.5 15 7862.5 16 7873.5 17 7885.5 18 7891.5 50 2.6 Summary and Discussions In order to improve the understanding of the transport and structural characteristics of the shale samples, we need to take into consideration all the results generated by the various techniques that have been utilized. An attempt to do so is shown in Table 2.6., where we summarize all our finding of potential correlations between matrix permeabilities and the various structural characteristics of the shale samples from various depths. Table 2.6: Correlation of Shale Matrix Permeability with Various Other Characteristics Measured Using Various Other Techniques Note: ↑ means increasing,↓ means decreasing, — means non-changing, e.g., Matrix permeability ↑ & average pore diameter↓ indicates that average pore diameter decreases with depth while matrix permeability increases matrix permeability ↑ BET,BJH and HK specific surface area (Figure 2.2) ↓ average pore diameter (Figure 2.5) ↑ micropore volume (Figure 2.8) ↑ mesopore volume (Figure 2.9) — macropore volume (Figure 2.10) ↑ total pore volume (Figure 2.11) ↑ GRI Approach lost gas (Figure 2.3) ↑ total gas content (Figure 2.4) ↑ TOC (Figure 2.25) ↑ core plug porosity (Figure 2.32) ↑ SEM horizontal view images (Figure 2.17) — vertical view images (Figure 2.17) — EDX carbon content (Figures 2.19-2.24) ↑ XRD clay content (Figure 2.26) ↓ calcite content (Figure 2.27) ↑ quartz content (Figure 2.28) ↑ plagioclase content (Figure 2.29) ↑ pyrite content (Figure 2.30) ↑ 51 Some of the relationships discovered are to be expected (e.g., increasing permeability with porosity) while others are unexpected (e.g., increasing permeability while the average pore diameter decreases). For others, furthermore, it is not entirely clear whether the various correlations are simply coincidental or that they offer further insight into the various phenomena which are occurring. Table 2.7 shows similar correlations between the gas content and the other sample characteristics. Table 2.7: Correlation of Shale Total Gas Content with Various Other Characteristics Measured Using Various Other Techniques Note: ↑ means increasing,↓ means decreasing, — means non-changing, e.g., Total gas content ↑ & average pore diameter↑ indicates that average pore diameter increases with depth while total gas content increases total gas content ↑ BET,BJH and HK specific surface area (Figure 2.2) ↓ average pore diameter (Figure 2.5) ↑ micropore volume (Figure 2.8) ↑ mesopore volume (Figure 2.9) — macropore volume (Figure 2.10) ↑ total pore volume (Figure 2.11) ↑ GRI Approach lost gas (Figure 2.3) ↑ matrix permeability (Figure 2.4) ↑ TOC (Figure 2.25) ↑ core plug porosity (Figure 2.32) ↑ SEM horizontal view images (Figure 2.17) — vertical view images (Figure 2.17) — EDX carbon content (Figures 2.19-2.24) ↑ XRD clay content (Figure 2.26) ↓ calcite content (Figure 2.27) ↑ quartz content (Figure 2.28) ↑ plagioclase content (Figure 2.29) ↑ pyrite content (Figure 2.30) ↑ 52 Again some of the common trends are to be expected (e.g., increasing gas content with porosity and surface area), while the reasons for the common trends are not so obvious for others, and may be coincidental. Nevertheless, cross-referencing in such a way the properties measured via the use of the various techniques provides an opportunity to better understand the mechanisms that determine gas storage and transport in such materials and may serve as a guideline for both researchers as well as field operators to better appreciate the phenomena occurring in their own shale samples even if only a fraction of the measurement tools mentioned in this chapter are available to them. Since we want to better appreciate the mass transfer mechanisms, we will describe a sequence of lab-scale depletion experiments in Chapter 3 to fulfill this objective. For these experiments we utilize a cylindrical core sample from a well located at depth 7880.5 ft (with bulk volume of 12.9 cc and density 2.45g/cm 3 ). Though we do not have previous characterization data for this depth, with the help of the data presented in Chapter 2, we are able to draw the following qualitative and quantitative conclusions about its characteristics: Qualitative characteristics of sample at a depth of 7880.5 ft: a. The micropore volume should be comparable to the mesopore volume. b. The sample should have significant surface area associated with micropores, as a result, a significant amount of adsorbed gas should be expected. c. The sample’s horizontal permeability should be about two orders of magnitude higher than its vertical permeability. d. The sample should have three types of porous regions, specifically: micropores (d<2nm), mesopores (2nm<d<50nm) and macropores (50nm<d<500nm) & microfractures (500nm<d<100 micron – these are observed visually). In addition, the following characteristics can be estimated quantitatively: 53 a. Bulk volume (12.9 cc, the diameter and height of the core is measured by a caliper) and bulk density 2.45g/cm 3 (which is measured by dividing the total weight of the core by the 12.9 cc bulk volume). b. Average mesopore diameter: ~4nm (see previous discussions and Figure 2.14. c. Considering the pore volume with depth data (see Table 2.3), the cumulative micropore volume for 7880.5ft should be within the range from 0.006 cm 3 /g to 0.015 cm 3 /g, mesopore volume for 7880.5ft should be within the range from 0.01 cm 3 /g to 0.025 cm 3 /g, and the cumulative total pore volume should be within the range from 0.022 cm 3 /g to 0.04 cm 3 /g. d. As a result, the micropore porosity for the 7880.5ft sample can be calculated to be from 1.5% to 3.7% while the mesopore porosity for the 7880.5 ft sample can be calculated to be from 2.4% to 6.1%. 54 Chapter 3: The Study of Mass Transfer in Gas Shales via Depletion Experiments To date, the study of shale gas well production performance has mostly focused on building models based on historical production data analysis while long-term shale-gas well performance characteristics are generally not well understood [Anderson et al., 2010]. As pointed out in the literature [Anderson et al., 2010], production analysis has proved to be a valuable reservoir characterization tool due to its practical, reasonably reliable and inexpensive nature. However, some of the more popular techniques, such as those proposed by [Arps 1945; Palacio et al., 1993; Doublet et al., 1994; Fetkovich et al., 1996; Agarwal et al. 1998] are designed for conventional reservoirs and primarily vertical wells. In terms of unconventional reservoirs, Wattenbarger et al. [1998] proposed a model for modeling tight gas with fractures, while Anderson et al. [2010] further improved the work by Wattenbarger et al. [1998] aiming at modeling a fractured shale- gas formation; Ambrose et al. [2010] proposed a methodology for estimating gas in place and sorption behavior based on its porous characteristics. All these approaches were established under reasonable assumptions and provide operators opportunities to select better operation strategies for the shale gas formation. Although the aforementioned analysis tools are widely used by operators to forecast production performance of shale gas wells, they are not able to reveal the mechanisms that determine well performance. In order to understand the specific production behavior associated with a shale gas well, the fundamental mass transfer characteristics of the gas shale formation need to be studied and understood. It is well-known, that in order for any adsorbed gas to reach a production well, the gas molecules must be desorbed from the pores first, and migrate through the various porous regions within the shale matrix and then transport within the fracture networks of the formation to be able to finally reach the wellbore. Since gas flow through fracture networks usually occurs via viscous flow that is significantly more efficient compared to transport mechanisms such as surface flow and Knudsen diffusion, understanding how these latter types of mass transfer as well as the desorption process limit the production rate becomes critical. Furthermore, if we fully understand the factors that limit the production rate, we can then build models to guide 55 the shale-gas operators towards an optimal depletion strategy. Accordingly, high-pressure, lab-scale depletion experiments are valuable in terms of understanding the mass transfer mechanisms during the production processes. In this chapter we report our findings from high-pressure depletion experiments and analyze in detail (experimentally and numerically) the characteristics of the long-term depletion behavior. The experimental results indicate the long-term production performance of our shale core sample is a process involving diffusion in the mesopore region and gas desorption and migration in the micropore region. This production mechanism is then confirmed by our numerical modeling efforts in this Chapter. 56 3.1 Experimental Approach and System Configuration Our high-pressure depletion system is designed to mimic the real shale-gas production system at relevant field conditions. Since the gas transport mechanisms within the shale formation dominate the long-term production performance and have not yet been studied well, our lab-scale, high-pressure experiments will focus on studying and analyzing the long-term gas production of a shale core sample and its production behavior. Our study also includes the initial production characteristics of the shale core in order to provide more data for our long-term production analysis. In order to study the dynamic properties of the gas-shale samples and to interpret well production behavior, we have constructed a high-pressure lab-scale experimental apparatus, shown in Figure 3.1, and further described in the accompanying Table 3.1. All components installed in the system have a pressure rating of at least 5,000 psig. Table 3.1: Component Names for Figure 3.1 A Argon G Argon Buffer Tank M Pressure Gauge B Methane H Pressure Regulator N Confining Pressure Pump C Air I Vacuum Gauge O Pressure Transducer D Gas Booster J Vacuum Pump P Back Pressure Regulator E Pressure Gauge K Pressure Gauge Q Computer F CH 4 Buffer Tank L Core-holder R Auto Control Valve 57 Figure 3.1: High-Pressure Depletion System. C C C C B B B B D D D D F F F F E E E E H H H H L L L L N N N N J J J J r el i ef r el i ef r el i ef r el i ef P P P P I I I I G G G G A A A A K K K K M M M M Q O O O O R R R R 58 The above set-up includes five different experimental sub-systems: 1) Vacuum and ventilation system, including I and J. A vacuum pump, monitored by a vacuum gauge, is used to evacuate the system prior to loading. Then, sub- system 1 can be isolated from the rest of the system. 2) The high-pressure gas generation system, including A, B, C, D, and E. Methane (B) or Argon (A) feeds into the booster (D). The booster is operated by air (C) to boost the gas to its desired pressure. The high pressure gas is then delivered to and stored in the storage system (sub-system 3). 3) The high-pressure storage system, including F, G, H and K. Gas from sub-system 2 is stored in the CH 4 or Argon buffer tank. When an experiment is initiated, the pressure regulator (H) is set to deliver a selected pressure to the experimental system (sub-system 4). The exact delivery pressure can be monitored by the pressure gauge (K). 4) The core-holder section, including L, M and N. The core-holder (L) is a stainless steel, Hassler-type core-holder capable of withstanding a pressure of up to 10,000 psig, and is used to hold the gas-shale core-plug under investigation. Confining pressure is supplied via a hydraulic pump (N). Gas can be sealed in the core once the auto control valve is closed. Figure 3.2 shows the details of the core-holder section. The experimental data are then recorded and analyzed via a data acquisition system (sub-system 5, below). 5) The data acquisition analysis system, including O and Q. Pressure data can be collected via a pressure transducer (O) and flow data can be calculated via the pressure change. The data acquisition and analysis are controlled by a computer. (Q). 59 Figure 3.2: Core-Holder and Data Acquisition Sub-system H ydr aul i c Pum p H ydr aul i c Pum p H ydr aul i c Pum p H ydr aul i c Pum p Back P r essur e Back P r essur e Back P r essur e Back P r essur e R egul at or R egul at or R egul at or R egul at or M M M M C om put er Pressur e Pressur e Pressur e Pressur e Transducer Transducer Transducer Transducer C or e H ol der C or e H ol der C or e H ol der C or e H ol der Vi t on Sl eeve Vi t on Sl eeve Vi t on Sl eeve Vi t on Sl eeve C or e C or e C or e C or e Stainless Steel Cap Val ve Val ve Val ve Val ve 2 2 2 2: : : : Aut o C ont r ol Val ve Aut o C ont r ol Val ve Aut o C ont r ol Val ve Aut o C ont r ol Val ve V al ve V al ve V al ve V al ve 1 1 1 1 60 3.2 Design of Depletion Experiments As discussed previously, shale-gas needs to transport through both the fracture networks as well as the shale matrix in order to be produced. The long-term production performance of a given well, based on current stimulation processes, has not yet been studied in detail. In Chapter 2, we found that the shale sample, from which our shale sample plug has been extracted (see Figure 3.3), has the following two unique characteristics: 1) Based on BET analysis of powder samples, the shale matrix in question has three categories of porosity (porous regions) each representing a given range of the pore-size distribution (pore diameter d): a. The macropores (50nm<d<500nm) for which the pore volume is negligible compare to the total pore volume. b. The mesopore region (2nm<d<50nm) accounting for approximately 50~60% of the total pore volume. c. The micropore region (d<2nm) which, for our sample, has comparable pore volume to that of the mesopore region. 2) The sample is characterized by a strong heterogeneity in permeability, with the horizontal permeability being at least two orders of magnitude larger than the vertical permeability. Since we have identified three types of porous regions in our shale core sample (which may also contain additional microfractures – see Figure 3.3 -- and macroporosity beyond (d>500 nm) that identified by the BET analysis of powders), as a result, we expect that the gas initially stored in these regions will be produced from the core sample at different rates during the depletion experiments. Gas production from the three regions can be classified as: 1) Gas in the macropores & microfractures system, which travels very fast from the fractures to the wellbore. This type of gas will be produced first once the well is put into production. 61 2) Free gas in the mesopores. Gas from these pores usually travels via diffusion to the fracture network and from there to the production wells. Accordingly, gas from the mesopores will arrive later to the production well than any gas that is produces by viscous flow alone. 3) Adsorbed gas in the micropores has to desorb from the surface of the micropores first and then travel via surface diffusion to the mesopores followed by a combination of Knudsen diffusion and continuum diffusion in the mesopores to reach the fracture networks. The time-scale for such gas to be produced is generally much longer. Figure 3.3: Shale Core Sample for Depletion Experiments Dictated by the range of mass-transfer phenomena at play, two types of experiments have been performed in this study: Those intended to study type #1 gas depletion behavior, defined as the short-term depletion experiments, and those intended to study type #2 and # 3 gas depletion behaviors, defined as long-term depletion experiments. 62 In this work we study the role of the following three key experimental parameters: a. Initial Pore Pressure: In our experiments, we can precisely control the initial pore pressure by supplying the shale core sample with gas at our target pressure via the use of the high-pressure gas charging system. b. Confining Pressure: The application of confining pressure (via the aid of sub- system 4, see Figures 3.1 and 3.2) in our experiments is intended to simulate, in the simplest possible form, the amount of stress that is being exerted on the formation. In our experimental system we are able to independently control and adjust the initial core pressure, the depletion pressure -- see below -- as well as the confining pressure. Thus, we are able to investigate the production behavior of shale samples under different net-stress. c. Downstream (or Back) Pressure: The downstream pressure in our experiments can be maintained constant via a back-pressure regulator, and is intended to mimic the operational conditions (well-bore pressure) as selected by the producer. The procedures (experimental protocol) for the depletion experiments described in this Chapter are listed below (with reference to Figure 3.2): 1. The core is placed into the core-holder, as illustrated in Figure 3.2. Before placing the core into the core-holder, it was wrapped with Viton shrink-tubing in order to facilitate insertion into the Hassler (Viton) sleeve and to protect the core’s integrity during depletion experiments which are accompanied with rapid changes in net stress. 2. Valves 1 and 2 are opened, the back-pressure regulator (BPR) is closed, and the shale core as well as the downstream BPR section are loaded with gas at the desired pressure, e.g., 2000 psig. Note, that the confining pressure is increased in parallel with the pore pressure to prevent damage to the core and Viton sleeve. 3. After the loading phase is complete, valves 1 and 2 are closed. Then the BPR is regulated to release the gas until the pressure transducer shows the desired back- pressure, e.g., 1000 psig. This is the initial pressure of the storage tank that collects the depleted gas. Then the BPR is fully closed, so in this case, the BPR 63 section acts simply as a storage tank: The BPR pressure changes as the gas from the core arrives in the tank (see step 4 below). The volume of the BPR section is measured prior to the depletion experiments, as discussed below. 4. Valve 2 is opened to connect the core to the BPR storage volume and to record pressure changes in the BPR via a pressure transducer. The change, in real time, of the pressure in the tank allows one to calculate the flow rate of gas out of the core. The procedure for measuring the internal system volume, also known as the “dead- volume”, prior to the depletion experiments, is as follows: 1) Open valves 1 and 2. Load the gas to obtain a certain pressure in the system with the BPR fully closed. 2) Close valves 1 and 2 to stop the loading process. 3) Fully open the BPR to let the gas out and record the mass flow rate for the degassing process by a mass flow meter (MFM) at the downstream. 4) Integrate the mass flow rate vs. time to obtain the total volume of gas depleted at standard conditions. 5) Use the equation of state to calculate the physical volume of the degassed section including the BPR section volume + dead volume of tubing and fittings of the flow path from the core in the core-holder to the BPR. 6) Open valve 1 and close valve 2, and load a selected pressure of gas into the system with the BPR fully closed. 7) Close valve 1 to stop the loading of the BPR section. 8) Repeat steps (3-5) to obtain the physical volume of the BPR section. Subsequently, the dead volume of tubing and fittings of the flow path in the core- holder section can also be calculated. 9) Repeat the process with different loading pressures, average the results to have the more accurate internal volume information. Once the dead volume and BPR section volume are determined, we can now perform and interpret the depletion experiments. It is important to ensure that the volume of the BPR 64 section is much larger than the dead volume of the core-holder section to allow for accurate measurements during the depletion process. 3.3 Depletion Experiments 3.3.1 Loading Time Experiments Prior to the initiation of the depletion experiments, the effect of the gas loading time was investigated. In these experiments, after the loading, the downstream depletion pressure is kept at 0 psig and the amount of gas produced was measured via a mass flow meter. Loading experiments using both Methane as well as inert Argon gas were carried out. Figure 3.4 shows the flow rate measured for the Methane experiments while Figure 3.5 shows the flow rates for the Argon experiments. Three loading times were investigated, namely 3 hr of loading (noted as run 1 in the Figures), overnight loading (run 2), and two days of loading (run 3). As Figures 3.4 and 3.5 indicate there is no difference in behavior between the 3 hr and the two days loading. As a result, the 3 hr loading was used for all the short-term depletion experiments. For the long-term experiment, in order to ensure a fully saturation of the core, we introduced two weeks loading for the entire long term set and we run depletion experiments until there is no pressure change over a period of 500 sec. 65 Figure 3.4: Flow Rates of Methane Experiments Figure 3.5: Flow Rates of Argon Experiments 0 20 40 60 80 100 0 200 400 600 800 1000 Flow rate (sccm) Time (s) Argon Experiments run1 run2 run3 0 100 200 300 400 500 600 700 0 200 400 600 800 Flow rate (sccm) Time (s) Methane Experiments run 1 run2 run3 66 3.3.2 Short-Term Methane Depletion Experiments We start by reporting the observations of a series of short-term methane depletion experiments. Table 3.2 shows the confining pressure, the initial pore pressure, and the back pressures used in these experiments. All experiments share the same confining pressure and initial pore pressure, and the only variable that changes is the back-pressure. Figures 3.6 - 3.11 show the pressure profiles in the depletion chamber for each of these runs. For these experiments, the depletion chamber is closed and its pressure rises as a result of the depletion process and this allows us to monitor the rate of depletion indirectly. As these Figures indicate, the pressure change in the depletion chamber is a small fraction of the initial core pressure, <5% but typically much less than that. Figures 3.12 - 3.17 show the depletion flow rates for each of the runs while Figure 3.18 contains all the depletion rate runs for comparison purposes (note that for the short-term depletion experiments, we stopped the run when there are no pressure changes in 3 consecutive sampling times (over a 0.3 sec period). Table 3.2: Parameter Summary for Methane Short-Term Depletion Short Term Methane Depletion Set 1 Parameters Confining Pressure (psig) 3000 3000 3000 3000 3000 3000 Core Pressure (psig) 2000 2000 2000 2000 2000 2000 Back-Pressure (psig) 0 500 750 1000 1250 1500 67 Figure 3.6: Pressure Curve of Short-Term 2000-0 (psig) Methane Depletion Figure 3.7: Pressure Curve of Short-Term 2000-500 (psig) Methane Depletion 0 20 40 60 80 100 120 140 0 2 4 6 8 10 Pressure (psig) Time (s) 490 500 510 520 530 540 550 560 570 580 590 0 1 2 3 4 Pressure (psig) Time (s) 68 Figure 3.8: Pressure Curve of Short-Term 2000-750 (psig) Methane Depletion Figure 3.9: Pressure Curve of Short-Term 2000-1000 (psig) Methane Depletion 1000 1010 1020 1030 1040 1050 1060 0 0.5 1 1.5 2 2.5 3 Pressure (psig) Time (s) 700 720 740 760 780 800 820 840 0 0.5 1 1.5 2 2.5 Pressure (psig) Time (s) 69 Figure 3.10: Pressure Curve of Short-Term 2000-1250 (psig) Methane Depletion Figure 3.11: Pressure Curve of Short-Term 2000-1500 (psig) Methane Depletion 1240 1245 1250 1255 1260 1265 1270 1275 1280 1285 0 0.5 1 1.5 2 Pressure (psig) Time (s) 1540 1545 1550 1555 1560 1565 0 0.5 1 1.5 Pressure (psig) Time (s) 70 Figure 3.12: Flow Rate of Short-Term 2000-0 (psig) Methane Depletion Figure 3.13: Flow Rate of Short-Term 2000-500 (psig) Methane Depletion 0 10 20 30 40 50 60 0 1 2 3 4 Flow Rate (scc/s) Time (s) 0 5 10 15 20 25 30 35 40 45 0 1 2 3 4 5 6 Flow Rate (scc/s) Time (s) 71 Figure 3.14: Flow Rate of Short-Term 2000-750 (psig) Methane Depletion Figure 3.15: Flow Rate of Short-Term 2000-1000 (psig) Methane Depletion 0 10 20 30 40 50 60 70 0 0.5 1 1.5 2 2.5 Flow Rate (scc/s) Time (s) 0 10 20 30 40 50 60 70 0 0.5 1 1.5 2 2.5 Flow Rate (scc/s) Time (s) 72 Figure 3.16: Flow Rate of Short-Term 2000-1250 (psig) Methane Depletion Figure 3.17: Flow Rate of Short-Term 2000-1500 (psig) Methane Depletion. 0 10 20 30 40 50 60 70 0 0.5 1 1.5 2 Flow Rate (scc/s) Time (s) 0 10 20 30 40 50 60 0 0.2 0.4 0.6 0.8 1 1.2 Flow Rate (scc/s) Time (s) 73 Figure 3.18: Flow Rates Comparison of Short-Term Methane Depletion Experiments 0 10 20 30 40 50 60 70 0 1 2 3 4 5 Flow Rate (scc/s) Time (s) Methane 2000-0 (psig) Methane 2000-500 (psig) Methane 2000-750 (psig) Methane 2000-1000 (psig) Methane 2000-1250 (psig) Methane 2000-1500 (psig) 74 3.3.3 Short-Term Argon Depletion Experiments A similar set of short-term Argon depletion experiments was also completed. Table 3.3 shows the confining, initial pore, and initial depletion pressures used in these experiments. All experiments, again, share the same confining pressure and initial pore pressure, and the only variable that changes is the downstream pressure. Figures 3.19 to 3.21 show the pressure profiles in the depletion chamber for each of these runs while Figures 3.22 - 3.24 show the depletion flow rates for each of the runs and Figure 3.25 provides a comparison of the depletion rates. Table 3.3: Summary of Parameters for the Argon Depletion Experiments. Short-Term Argon Depletion Parameters Confining Pressure (psig) 3000 3000 3000 Core Pressure (psig) 2000 2000 2000 Back-Pressure (psig) 0 500 1000 75 Figure 3.19: Pressure Change of Short-Term 2000-0 (psig) Argon Depletion Figure 3.20: Pressure Change of Short-Term 2000-500 (psig) Argon Depletion 0 20 40 60 80 100 120 140 0 2 4 6 8 Pressure (psig) Time (s) 500 520 540 560 580 600 620 640 660 0 1 2 3 4 5 6 Pressure (psig) Time (s) 76 Figure 3.21: Pressure Change of Short-Term 2000-1000 (psig) Argon Depletion Figure 3.22: Flow rate of Short-Term 2000-0 (psig) Argon Depletion 0 5 10 15 20 25 30 35 40 45 0 2 4 6 8 Flow Rate (scc/s) Time (s) 1000 1010 1020 1030 1040 1050 1060 1070 1080 1090 1100 0 0.5 1 1.5 2 2.5 3 Pressure (psig) Time (s) 77 Figure 3.23: Flow rate of Short-Term 2000-500 (psig) Argon Depletion Figure 3.24: Flow rate of Short Term 2000-1000 (psig) Argon Depletion 0 10 20 30 40 50 60 0 1 2 3 4 Flow Rate (scc/s) Time (s) 0 10 20 30 40 50 60 70 0 0.5 1 1.5 2 2.5 Flow Rate (scc/s) Time (s) 78 Figure 3.25: Flow rate Comparison of Short-Term Argon Depletion Experiments 0 10 20 30 40 50 60 70 0 2 4 6 8 Flow Rate (scc/s) Time (s) Argon 2000-0 (psig) Argon 2000-500 (psig) Argon 2000-1000 (psig) 79 3.3.4 Blank (without a Core) Depletion Experiments We have also completed a series of Methane depletion experiment without a core inside the core-holder. In these experiments, the core-holder was pressurized to a certain initial pressure and then allowed to deplete into the downstream depletion chamber whose pressure was measured in order to calculate the rates of depletion. The values of the initial pressures and the initial pressures in the depletion chamber are the same with those utilized during the short-term Methane depletions runs and shown in Table 3.2. Figures 3.26 to 3.31 show the pressure profiles in the depletion chamber for each of these runs, while Figures 3.32 to 3.37 show the corresponding depletion flow rates. A summary of all blank runs are provided in Figure 3.38. A comparison of the depletion flow rate observed when the core is present and when it is absent allows one to determine the impact that the presence of the core (particularly the gas stored in the fracture network and the macroporosity) has on the observed dynamic behavior. Figure 3.26: Pressure Curve of 2000-0 (psig) Methane Blank Depletion 0 50 100 150 200 250 300 350 0 2 4 6 8 10 Pressure (psi) Time (s) 80 Figure 3.27: Pressure Curve of 2000-500 (psig) Methane Blank Depletion Figure 3.28: Pressure Curve of 2000-750 (psig) Methane Blank Depletion 400 450 500 550 600 650 700 750 0 2 4 6 8 10 Pressure (psi) Time (s) 700 750 800 850 900 950 0 2 4 6 8 Pressure (psi) Time (s) 81 Figure 3.29: Pressure Curve of 2000-1000 (psig) Methane Blank Depletion Figure 3.30: Pressure Curve of 2000-1250 (psig) Methane Blank Depletion 980 1000 1020 1040 1060 1080 1100 1120 1140 0 2 4 6 8 Pressure (psi) Time (s) 1220 1240 1260 1280 1300 1320 1340 1360 0 1 2 3 4 5 Pressure (psi) Time (s) 82 Figure 3.31: Pressure Curve of 2000-1500 (psig) Methane Blank Depletion Figure 3.32: Flow Rate of 2000-0 (psig) Methane Blank Depletion 0 10 20 30 40 50 60 70 0 2 4 6 8 10 12 Flow Rate (scc/s) Time (s) 1480 1490 1500 1510 1520 1530 1540 1550 1560 0 1 2 3 4 Pressure (psi) Time (s) 83 Figure 3.33: Flow Rate of 2000-500 (psig) Methane Blank Depletion Figure 3.34: Flow Rate of 2000-750 (psig) Methane Blank Depletion 0 10 20 30 40 50 60 70 0 2 4 6 8 10 Flow Rate (scc/s) Time (s) 0 10 20 30 40 50 60 70 0 2 4 6 8 Flow Rate (scc/s) Time (s) 84 Figure 3.35: Flow Rate of 2000-1000 (psig) Methane Blank Depletion Figure 3.36: Flow Rate of 2000-1250 (psig) Methane Blank Depletion 0 10 20 30 40 50 60 70 80 0 1 2 3 4 5 6 Flow Rate (scc/s) Time (s) 0 10 20 30 40 50 60 70 0 1 2 3 4 5 Flow Rate (scc/s) Time (s) 85 Figure 3.37: Flow Rate of 2000-1500 (psig) Methane Blank Depletion Figure 3.38: Flow Rates Comparison of All Methane Blank Depletion Runs 0 10 20 30 40 50 60 70 0 1 2 3 4 Flow Rate (scc/s) Time (s) 0 10 20 30 40 50 60 70 80 0 2 4 6 8 10 12 Flow Rate (scc/s) Time (s) Blank 2000-0 (psig) Blank 2000-500 (psig) Blank 2000-750 (psig) Blank 2000-1000 (psig) Blank 2000-1250 (psig) Blank 2000-1500 (psig) 86 3.3.5 Long-Term Methane Depletion Experiments The short-term depletion experiments described above are intended to capture the dynamics of the gas flow from the microfractures and the macroporosity of the core. We have carried out additional long-term depletion experiments, the experimental conditions for which are shown in Table 3.4. In these experiments the run was terminated if the pressure in the depletion chamber did not change over a period of 500 s. Figures 3.39 to 3.41 report the pressure profiles in the depletion chamber for each of these runs. These long-term depletion experiments are intended to probe the dynamics of transport in the core’s matrix as well as those dynamics associated with the desorption process in order to understand its impact during shale gas production. Table 3.4: Parameter Summary of Long-Term Methane Depletion Experiments Long Term Depletion Experiments Confining Pressure (psig) 3000 3000 3000 Pore Pressure (psig) 2000 2000 2000 Back-Pressure (psig) 0 500 1000 87 Figure 3.39: Pressure Curve of 2000-0 (psig) Long-Term Methane Depletion Figure 3.40: Pressure Curve of 2000-500 (psig) Long-Term Methane Depletion 0 20 40 60 80 100 120 140 160 180 0 20 40 60 80 100 Pressure (psi) Time (hr) 490 500 510 520 530 540 550 560 570 580 590 0 1000 2000 3000 4000 5000 6000 Pressure (psig) Time (s) 88 Figure 3.41: Pressure Curve of 2000-1000 (psig) Long-Term Methane Depletion 1015 1020 1025 1030 1035 1040 1045 1050 1055 1060 1065 1070 0 1000 2000 3000 4000 5000 Pressure (psig) Time (s) 89 3.3.6 Data Analysis and Preliminary Discussion Table 3.5 shows the analysis of the results of the experiments described in Sec. 3.3.3. Based on the pressure profiles in the downstream depletion chamber shown above in Figures 3.19-3.21, we calculate the total amount of gas (in scc under standard ambient temperature and pressure conditions of P=14.504 psi and T=298.15 K) that arrives into the chamber from the high-pressure upstream chamber (core holder). Knowing the empty volume in the high-pressure chamber (empty volume in the upstream without core), also known as the “dead volume of the core holder section” (refer to section 3.2), we can then calculate the total amount of gas (scc) stored in the dead volume and subtract it from the amount of gas that was depleted from the high-pressure chamber into the low-pressure chamber. This gas is noted in Table 3.5 as “Gas Depleted from the Core”. Note that the dead volume measurement is carried out by replacing the shale core with a same dimension stainless steel insert, in this case, we do not introduce extra dead volume into the system after inserting the shale core into the core-holder. We then calculate the volume that this gas would occupy under the initial pressure and temperature conditions of our experiment using the following relationship RT z V P RT z V P 2 2 2 1 1 1 = , (3.3.6 -1) where 1 and 2 denotes the initial and final state of the experiment, respectively. The gas compressibility factors (z) can be estimated via equation of state [Maloney, 2008]. This volume is now listed as “Volume of Gas at Initial Pressure Conditions” in Table 3.5. For Argon this volume does not change with the initial depletion pressure, which is consistent with the idea that Argon is an inert gas that does not adsorb in the core. This pore volume corresponds to a porosity of 3.1%. We have estimated in Chapter 2 the (micropore + mesopore + macropore) porosity to be 8.33%; we believe that the porosity of 3.1% from our short term depletion experiment should only represent the macropore and microfracture porosity, though strictly speaking, this amount may include a small fraction of the mesopore porosity as well since there is no way to 100% distinguish mesopore depletion and macropore depletion from each other. 90 Table 3.5: Depleted Void Volume of Short-Term 2000 psig Argon Depletion Table 3.6 summarizes a similar analysis of the data from the methane depletion experiments described in Sec. 3.3.2 above. Note that the “Volume of Gas at Initial Pressure Conditions” calculated from the methane data is consistent with that calculated from the Ar data at the high initial pore pressure conditions. However, as the initial pore pressure decreases the “Volume of Gas at Initial Pressure Conditions” increases beyond the value calculated form the Ar data, which is indicative that some of the gas arriving to the low-pressure chamber under such conditions may be either adsorbed gas or gas stored in the mesoporous and/or microporous regions in the core. We note that the relative error of the volume estimations should be within +/- 1% since the pressure gauge manufacturer reports all of their measurement devices have over 99% in measurement accuracy. Table 3.6: Depleted Void Volume of Short-Term 2000 psig Methane Depletion Short-Term 2000 psig Argon Depletion (with +/- 1% in error) Pore Pressure (psig) 2000 Back Pressure (psig) 1000 500 0 Confining Pressure (psig) 3000 Gas Depleted from the Core (scc) 24.9 39.4 50.2 Volume of Gas at Initial Pressure Conditions (cc) 0.39 0.41 0.4 Average Void Volume Depleted (cc) 0.4 Short-Term 2000 psig Methane Depletion (with +/- 1% in error) Pore Pressure (psig) 2000 Back Pressure (psig) 1500 1250 1000 750 500 0 Confining Pressure (psig) 3000 Core Depleted (scc) 14.2 23.4 33.2 40.9 51.9 68.1 Void Volume (cc) 0.39 0.39 0.42 0.43 0.44 0.46 Average Void Volume (cc) 0.42 91 As Figures 3.39-3.41 indicate there are two distinct regions in the depletion chamber pressure profile curves. They include a first region where the pressure rises fast, followed by a second region where the change in pressure is considerably lower. Figure 3.42 shows the depletion rates as a function of time as a (log 10 - log 10 ) plot for the three runs, discussed above, with an initial loading pressure of 2000 psig. Both the flow rates and time are shown in terms of (log 10 - log 10 ) plot in order to clearly demonstrate the substantial differences in the initial depletion behavior and the depletion behavior at longer times. It is clear that the initial production rates, corresponding to gas stored in the microfractures and/or macroporosity of the shale sample, are substantially higher than the production rates at later times where the gas produced is either desorbed from the shale sample or is “free-gas” from the mesoporous and/or microporous regions. Figure 3.43 shows the cumulative production (dead volume depletion + core depletion) for the three runs with initial loading pressure of 2000 psig and varying initial depletion chamber pressure. It is clear from the Figure that the pressure where the core depletes has a substantial impact on the amount of gas that depletes, with substantial more gas being depleted for the cases with the lower initial depletion chamber pressure. Note, for example, in Figure 3.43 that for the case of an initial depletion pressure of 0 psig it takes a much longer time for the system to reach steady state, and that the final cumulative gas produced is close to 4 times higher than the gas produced for an initial depletion chamber pressure of 1000 psig. From the above results one can distinguish three production stages associated with the lab-scale core depletion experiments: 1) Stage 1: Gas at this stage is produced very fast (in less than a few seconds in the experiments reported here with the small-size core), likely via viscous flow; this gas is mostly stored in the macropores & microfractures of the core samples. 2) Stage 2: The majority of the gas depleted at this stage is stored in the bulk-state in the mesoporous region travels via Knudsen through the mesopores to reach the macropores and microfractures and then deplete out of the core. This stage is in between the viscous flow stag potentially overlapping with both. 3) Stage 3: This stage (for this small size shale sample) lasts for nearly 100 hr. The gas in this stage must come from the surface of the micropores/mesopores (where it desorbs from) or from the dense organic matrix of the sample itself. In a real world setting all three types of gas are likely to participate as well. One key observation of our lab-scale experiments is that nearly half the gas that is held by the shale sample is in the adsorbed/absorbed state and that the rate by which this gas depletes as well the total amount of the gas that it eventually depletes is a strong function of the back-pressure (well pressure as controlled by the producer). In what follows, a model is developed that aims to describe the basic phenomena that take place during the lab-scale depletion experiments and to analyze the experimental data that have been generated. Figure 3.42: 0.000001 0.00001 0.0001 0.001 0.01 0.1 1 10 100 0.1 1 Flow Rates (scc/s) between the viscous flow stage (see above) and the desorption stage (see below) potentially overlapping with both. Stage 3: This stage (for this small size shale sample) lasts for nearly 100 hr. The gas in this stage must come from the surface of the micropores/mesopores (where orbs from) or from the dense organic matrix of the sample itself. In a real world setting all three types of gas are likely to participate as well. One key scale experiments is that nearly half the gas that is held by the mple is in the adsorbed/absorbed state and that the rate by which this gas depletes as well the total amount of the gas that it eventually depletes is a strong function of the pressure (well pressure as controlled by the producer). In what follows, a model is developed that aims to describe the basic phenomena that take place during the scale depletion experiments and to analyze the experimental data that have been Figure 3.42: Flow Rates Comparison of the 2000 (psig) Set 10 100 1000 10000 1000001000000 Time (s) 2000-1000 (psig) 2000-500 (psig) 2000-0 (psig) 92 e (see above) and the desorption stage (see below) Stage 3: This stage (for this small size shale sample) lasts for nearly 100 hr. The gas in this stage must come from the surface of the micropores/mesopores (where orbs from) or from the dense organic matrix of the sample itself. In a real world setting all three types of gas are likely to participate as well. One key scale experiments is that nearly half the gas that is held by the mple is in the adsorbed/absorbed state and that the rate by which this gas depletes as well the total amount of the gas that it eventually depletes is a strong function of the pressure (well pressure as controlled by the producer). In what follows, a simple model is developed that aims to describe the basic phenomena that take place during the scale depletion experiments and to analyze the experimental data that have been 1000000 1000 (psig) 93 Figure 3.43: Cumulative Production of the Entire Depletion Lifespan of the 2000 (psig) Set 0 20 40 60 80 100 120 140 160 0.1 10 1000 100000 Cumulative Production (cc) Time (s) 2000-1000 (psig) 2000-500 (psig) 2000-0 (psig) 94 3.4 Modeling The main focus of our model will be to describe gas transport during depletion stages 2 and 3 above, so we assume here that the gas within the macrofractures of the shale sample where pore viscous flow prevails instantly equilibrates with the pressure of the depletion chamber. In real world settings, depending on the distance that the gas has to travel through the fracture network to the production well, this may not be a valid assumption, and one must then also account for the viscous flow phenomena in the macrofracture network. Our model, see Figure 3.44, accounts for two types of pore regions: One in which gas is stored via adsorption (or absorption) and transport occurs either through surface flow (if the gas is adsorbed) or transport through the dense structure (if the gas is absorbed). Please note that the mass balance equations describing both phenomena are indistinguishable. In the second region, gas is stored in a bulk state and gas transport occurs via both viscous (Poiseuille-type) as well as slip-flow (Knudsen transport). This second region, as shown in Figure 3.44, surrounds the first region and acts as a conduit of the gas stored (adsorbed and/or absorbed) in the shale matrix to reach the macrofracture network (and in a real setting eventually the production well). The above model is known as the bidispersed pore model (BPM), and was originally developed to account for transport and adsorption (Ruckenstein 1971, Do 1998, Shi and Durucan 2003, Shi and Durucan 2005, Yang et al. 2006, Dadwhal et al. 2009) in adsorbent and catalysts particles. It strictly applies to such particles which are characterized by a bimodal pore size distribution, and which are typically made via shaping and compressing into pellets of microparticulate powders. BET analysis of the shale sample used in the depletion experiments (and discussed in Ch. 2 of this Thesis) also indicates the presence of a bimodal pore size distribution with both a microporous and a mesoporous region being present. However, it is unlikely that the shale sample matrix physically resembles the idealized structure of Figure 3.44. In the context of applying the BPM to the lab-scale depletion experiments, the diameter of the microparticles must be interpreted as the characteristic length scale of the microporous region, which is the average distance a molecule must travel via surface diffusion before encountering a mesopore/macropore. And the diameter of the particle must be viewed as 95 the characteristic length scale of the mesopore/macropore region which is the average distance a molecule must travel via transitional flow (Bulk + Knudsen) before encountering and depletion into the microfracture network. Figure 3.44: Illustration of the BPM for Our Simulation Purpose. 96 Below we first discuss the mass balance equations that describe the phenomena that occur in the microporous and the mesoporous/macroporous regions. These equations are then converted into their dimensionless form and the important dimensionless groups are identified. The models are then utilized to analyze the experimental data. For the microporous region gas transport is via surface diffusion with the flux (mol/cm 2 .s) described by the following Equation: I I s s r C D J ∂ ∂ − = , (3.4-1) with s D (cm 2 /s) being the surface diffusivity, I C (mol/cm 3 ) the gas concentration in the microporous region, I r (cm) the radial coordinate in the “microparticle”, see Figure 3.44. Surface diffusion is typically an activated process, so surface diffusivity is described via an Arrhenius-type equation:       − = RT E D D s s s exp 0 , (3.4-2) with s E (J/mol) being the activation energy for surface diffusion. For the mesoporous region, as represented by the “pellet” in the BPM, transitional flow is the prevailing transport mechanism and the flux J T (mol/cm 2 .s) is described by the following equation: M M M i M M M i T r P B C r C D J ∂ ∂ − ∂ ∂ − = 0 τ ε τ ε , (3.4-3) where M C (mol/cm 3 ) is the gas concentration, P M (atm) is the gas pressure in the mesopores, ε is the “macroparticle” porosity, i τ is the tortuosity, M D (cm 2 /s) the Knudsen diffusivity, B 0 (mol/atm.s) the viscous term, μ the gas viscosity (g/cm.s) in the mesopores, and M r (cm) the radial coordinate in the “macroparticle”. Often during the application of the BPM, the pore space is considered to consist of straight, 97 nonintersecting, cylindrical pores with a large aspect ratio (length/diameter), in which case M D and B 0 are described by the following two Equations: M RT r D M π 8 3 2 = , (3.4-4) μ μ B r B = = 8 2 0 , (3.4-5) where r is the (average) pore radius, M (g/mol) the molecular weight, and T (K) the temperature. Note that approximating the mesopore pore structure by a simple pore model of this kind is not a necessity for the further application of BPM for the analysis of the lab-scale depletion experiments. Assuming that the following equation of state (EOS) applies in the “macroparticle” M M P zRTC = , (3.4-6) where R (82.057 cm 3 atm/K/mol) is the universal gas constant, z is the gas compressibility factor then Equation (3.4-3) can be also written as: M M M i M M M i T r zC C RT B r C D J ∂ ∂ − ∂ ∂ − = ) ( ) ( 0 τ ε τ ε , (3.4-7) Assuming a spherical geometry for the “microparticle” and the “macroparticle” the mass balance equations are as follows: For the “microparticle”: [ ] s I I I I J r r r t C 2 2 1 ∂ ∂ − = ∂ ∂ , (3.4-8) 98 For the “macroparticle”: [ ] T M M M I M J r r r t C t C 2 2 1 - 1 ∂ ∂ − = ∂ ∂ + ∂ ∂ ）（ ε ε , (3.4-9) where is I C is the volumetric mean concentration in the “microparticle” described by the following Equation: I I R I I I dr C r R C I ∫ = 0 2 3 3 , (3.4-10) Substituting equations (3.4-1) and (3.4-7) into (3.4-8) and (3.4-9), we then have: For the “microparticle”:       ∂ ∂ ∂ ∂ = ∂ ∂ I I I I I s I r C r r r D t C 2 2 , (3.4-11) For the “macroparticle”:       ∂ ∂ ∂ ∂ +       ∂ ∂ ∂ ∂ = ∂ ∂ + ∂ ∂ M M M M M M t M M M M M t M I M r zC C r r r RT B r C r r r D t C t C ) ( ） - 1 （ 2 2 0 2 2 τ ε τ ε ε ε , (3.4-12) Equations (3.4-11) and (3.4-12) are complimented with the following set of boundary conditions (BC): The BC for the microparticle are: r I = R I , M M S M I bC bC C C f C + = = 1 ） ( , (3.4-13) r I = 0 , 0 = ∂ ∂ I I r C , (3.4-14) 99 where S C is the saturation capacity (mol/cm 3 )b is the adsorption affinity (cm 3 /mol) and the key assumption in Eqn. 3.4-13 is that methane adsorption in these materials is described by a Langmuir isotherm, which has been verified at USC by Mr. Wang Yu in his experimental studies with these materials. The BC for the macroparticle are: r M = R M , b M C C = , (3.4-15) r M = 0 , 0 = ∂ ∂ M M r C , (3.4-16) where b C (mol/cm 3 ) is the bulk concentration surrounding the “pellet”, which is actually the concentration in the microfractures that is assumed equal to the initial pressure of the depletion chamber. The initial conditions for the model, we assume that the “microparticle” and the “pellet” are equilibrated at the initial pore pressure (core-holder), P 0 . As a result, at time t=0, the initial conditions of the model maintained at gas pressure P i can be written as: t = 0, zRT P C C M M / 0 0 = = , ) ( 0 0 M I I C f C C = = (3.4-17) In order to solve the above set of equations dimensionless we define the following dimensionless variables: 0 I I C C x = (3.4-18) 0 M M C C y = (3.4-19) 0 M b b C C y = (3.4-20) 100 I I R r = α (3.4-21) M M R r = β (3.4-22) t R D I s 2 = τ (3.4-23) The dimensionless Equations can then be written for the “microparticle” and the “pellet” in the form:       ∂ ∂ ∂ ∂ = ∂ ∂ α α α α τ x x 2 2 1 , (3.4-24) with α α xd C C x I I ∫ = = 1 0 2 0 3 , (3.4-25) and       ∂ ∂ ∂ ∂ +       ∂ ∂ ∂ ∂ = ∂ ∂ + ∂ ∂ β β β β σ β β β β γ τ λ τ ) ( ) ( 1 1 2 2 2 2 zy y y G y x y , (3.4-26) where 0 0 ） - 1 （ M I C C ε ε λ = , (3.4-27) 2 2 M I s t M R R D D τ γ = , (3.4-28) 2 2 0 M I s t M R R D BRT C τ σ = , (3.4-29) 101 Note that G(y) in (3.4-26) actually reflects the viscosity change with concentration y since viscosity if a function of pressure. The dimensionless BC for the pellet and the microparticle are then given by: 1 = β , b y y = , (3.4-30) β = 0 , 0 = ∂ ∂ β y , (3.4-31) and 1 = α , ) ( 1 1 0 0 0 0 0 y f bC bC C y bC y bC C C C x S S I I = + + = = , (3.4-32) 0 = α , 0 = ∂ ∂ α x , (3.4-33) Since we do not have TGA results to calculate the adsorption affinity for this specific sample, we assume we have the same adsorption affinity mol cm b / 3 . 631 3 = as my colleague Mr. Yu Wang found out from his TGA experiments on samples from nearby depths. The corresponding initial conditions are given in dimensionless form by τ = 0, y =1, x = 1 , (3.4-34) Note that equations (3.4-28) and (3.4-29), involving the dimensionless parameters γ and σ , can be reorganized to identify relevant scaling groups: Knudsen micro M M t s I M I s t M D R D R R R D D τ τ τ τ γ =         = = −1 2 2 2 2 , (3.4-35) bulk micro M M t s I M I s t M RT B C R D R R R D BRT C τ τ τ τ σ =         = = −1 0 0 2 2 2 2 0 , (3.4- 36) 102 where τ micro is the micropore diffusion time scale, τ Knudsen is the Knudsen diffusion time scale in the mesopore and τ bulk is the bulk diffusion time scale in the mesopore. As a direct result, the dimensionless parameters γ andσ represent the following physical time-scale rations: mesopore in scale time diffusion knudsen The scale time diffusion micropore The = γ , (3.4-37) mesopore in scale time diffusion bulk The scale time diffusion micropore The = σ , (3.4-38) The above set of equations can be solved by an orthogonal collocation method (Appendix C), and we use matlab to do fit parameters γ andσ for all of our three long term experimental data at the same time. Note that λ in (3.4-27) can be estimated to be 12.31 as described in the following: we can choose 3% as the mesopore porosity from chapter 2, we can calculate 0 M C as 0.0068mol/cm 3 from equation (3.4-17) for 2000psig initial pressure, we can subtract the free gas depletion amount from the total core depletion amount to obtain the amount of adsorbed gas, assuming the adsorbed gas occupies 100% micropore volume, we can divide the adsorbed gas by micropore volume to have 0.0025 mol/cm 3 as the initial gas concentration in micropore 0 I C . Figures 3.45-3.47 show the comparison between the BPM and the long term experimental observations (stage 2 and 3) for the 2000-0psig, 2000-500psig and 2000-1000psig depletion processes. The comparison indicate BPM provides a fair fit for the experimental data, the deviation may be because we do not have the TGA results to obtain adsorption affinity for this particular core and we have to use the data from nearby depth. And the parameters γ andσ (with 95% confidence interval provided) are fitted to be: And if we assume 4nm is the mesopore diameter and 1.2nm is the micropore di the micropore and mesopore diffusivity can be estimated from 10 1.05 10 98 . 0 [ 8 8 D s − − × × = which are in agreement with 2004; Yuan et al. 2013]. Figure 3.45: Comparison of BPM with Experimental Data for 2000 0 5 10 15 20 25 30 35 0 0.5 Production (cc) ] 10 4.36 10 03 . 4 [ -5 5 - × × = γ , ] 10 3.86 10 61 . 3 [ -7 7 × × = − σ , is the mesopore diameter and 1.2nm is the micropore di he micropore and mesopore diffusivity can be estimated from γ ,σ to be: sec / ] 2 8 cm and ] 10 1.67 10 54 . 1 [ 5 5 cm D M − − × × = with the diffusivities reported by [Leythaeuser 1982; Figure 3.45: Comparison of BPM with Experimental Data for 2000-0 (psig) 0.5 1 1.5 2 2.5 3 Time (s) x 100000 2000-0 (psig) Experiment 2000-0 (psig) BPM 103 (3.4-41) (3.4-42) is the mesopore diameter and 1.2nm is the micropore diameter, sec / 2 cm , 1982; Schloemer Long Term 3 104 Figure 3.46: Comparison of BPM with Experimental Data for 2000-500 (psig) Long Term Figure 3.47: Comparison of BPM with Experimental Data for 2000-1000 (psig) Long Term 0 1 2 3 4 5 6 7 8 0 1000 2000 3000 4000 5000 6000 Production (cc) Time (s) 2000-500 (psig) Experiment 2000-500 (psig) BPM 0 0.5 1 1.5 2 2.5 3 0 500 1000 1500 Production (cc) Time (s) 2000-1000 (psig) Experiment 2000-1000 (psig) BPM 105 3.5 Summary and Conclusions In this chapter, we have described the design of a high-pressure depletion system (HPDS) for the purpose of investigating the mass transfer mechanisms of a shale core sample. The HPDS was constructed with the ability to both evaluate the short term production and long term production behaviors of the sample. A comparison between methane and argon short-term depletion experiments indicates that viscous flow is the prevailing mass transfer mechanism for early times of the depletion process. These experiments also allowed us to estimate the porosity (including macropores and microfractures within the core sample) of the sample. For long-term depletion processes, gas desorption and diffusion are identified to be the rate limiting factor in the production performance. The experimental observations from long term depletion experiments indicate that there desorption of gas from the shale surfaces contribute significantly to the cumulative production when the back pressure is low (less than 500psig). Combining the data we generated from our sample characterization efforts in Chapter 2 and the depletion data from presented in this chapter, we identified three type of porous structures (micropore structure, mesopore structure and macropore & microfracture structure) that contribute to three distinct stages of the gas production: During the initial production, all the free gas stored within the macropore & microfracture structure is being produced via viscous flow, this process is an fast (almost instant) event compared to the entire production period. After the initial production by viscous flow, diffusion (knudsen diffusion and/or bulk diffusion) in mesopores begins to contribute and as a result, the concentration in the mesopore starts to drop. When the concentration within the mesopores decreases, gas stored in the micropores begins to desorb gradually and migrates to the mesopores via surface diffusion. A bidisperse pore model (BPM) was formulated and found to provide for a fair description of the desorption-diffusion processes of the long term production. And the diffusivities are found to be consistent with the literatures. 106 Chapter 4: Optimization of Hydraulic Stimulation Processes via Additives 4.1 Introduction and Overview of Additives Hydraulic fracturing is the process that is today most commonly utilized during shale-gas operations to improve production rates by contacting a larger volume of the shale play. During hydraulic fracturing, operators utilize various additives along with water and proppants to facilitate the shale-gas production. However, hydraulic fracturing is also a common cause of water-related formation damage. As a result, an improved understanding of the function of common additives is critical in terms of avoiding and/or minimizing the damage to the formation during hydraulic stimulation. Table 4.1 (from [Wylde et al., 2011]) shows the commonly used fracturing fluid additives and their function. The additives most commonly used today in shale-gas exploration can be generally categorized into three groups: Emulsions, Viscoelastic Surfactants, and Polymers. Table 4.1: Commonly Used Fracturing Fluid Additives and Their Function 107 4.1.1 Emulsions Emulsions can be of three different types, as described in Table 4.2 (from [Rickman et al., 2010]). Among these different types of emulsions, microemulsions (ME’s) are the most effective additives, as documented via both lab-scale and field-scale tests, in terms of boosting gas production by reducing water-blockage-type of formation damage (capillarity). Table 4.3 below summarizes the important characteristics of ME’s in terms of their composition, functionality, and describes key lab-scale and field scale observations of their performance. Table 4.2: Physical Attributes of Conventional Emulsions and Dilute Microemulsions 108 Table 4.3: Overview of Microemulsions (ME) ME used in hydraulic fracturing fluids Composition “Optically clear, thermodynamically stable blend of biodegradable solvent, surfactant, co-solvent and water. The diluted ME is a nano- emulsion whose stability is dictated by kinetic processes” from [Rickman et al., 2010]. Functional Objectives Increase shale-gas recovery as well as productivity by lowering the capillary pressure as well as capillary end effect Lab Observations “Capillary pressure is reduced by as much as 50%” from Penny et al. [2005, 2006, 2007], “and the relative permeability to gas has been improved, reservoir damage factor has also been reduced by a factor of 2” from Penny et al. [2006, 2007]. Field Tests “Tests from nine different oil and/or gas basins (D.J, San Juan, Unitah, Raton, Green River, Pinedale, Big Horn, Fort Worthshow and Williston) indicate that 30% of the wells have achieved a 350% production improvement and over 68% had lower lifting cost” from [Penny et al. [2005]. “200 wells in Barnett shale have been treated with ME and analyzed, results showed that there is an over 50% on average load recoveries and a 30-40% increase in gas production” from Penny et al. [2006] “Over 300 wells in Barnett, Fayetteville and Appalachian basin shale, some 200 wells in San Juan and other low permeability basins have been treated with various fluid combinations of ME, and results showed 50% to 100% increase in load recoveries and gas production” from Penny et al. [2007]. 109 4.1.2 Viscoelastic Surfactants Viscoelastic surfactants are another important class of additives. They have been reported [Houston et al., 2009] capable of reducing the surface tension and the interfacial tension (IFT), and to change the contact angle, all resulting in improved flow-back and load recoveries via a reduction of capillary forces. For example, typical load recoveries of 30- 60% are reported when using a surfactant compared to 20% without surfactants. There are several types of viscoelastic surfactants, as summarized by Zhang et al. [2010]: “Viscoelastic surfactants including viscoelastic anionic surfactants (VAS), viscoelastic surfactants (VES), [Samuel 1997, 1999, 2000, 2003], viscoelastic surfactant polymers [Shashkina et al., 2005; Maestro et al., 2004; Couillet et al., 2005].” The advantages of viscoelastic surfactants include: higher recovery [Wylde et al., 2011], no residue during fracturing [Zhang et al., 2010], decreased viscosity of the flow-back fluid and increased gas production [Tang et al., 2002; Paktinat et al., 2011; Wylde et al. 2011]. The following are a couple of examples of field tests where the use of viscoelastic surfactants provided positive results in terms of increased gas production. a. Montney Formation (British Columbia and Alberta). Figures 4.1 and 4.2, from [Wylde et al. 2011], describe the field results which indicate that there is a significant increase in production due to the use of the surfactants [Wylde et al., 2011]. b. Sulige gas field (Inner Mongolian, China) – “heterogeneous sandstone tight-gas formation with almost no natural productivity. The use of the viscoelastic surfactant was reported to increase the gas production of two wells by 12 and 14%, respectively”, by [Zhang et al., 2010]. 110 Figure 4.1: Well Performance on Back Production from Fracture Treatments When Using Mutual Solvent (2-BE) and without Using Mutual Solvent. Figure 4.2: Well Performance on Back Production from Fracture Treatments when Using the Environmentally Acceptable Mutual Solvent A and then 2-BE. 111 4.1.3 Polymers Polymers are most commonly used during hydraulic fracturing either as friction reducers (FR) or as wettability alteration agents. For example, stimulation fluids with FR pumped at high rates with low sand concentrations have been very effective at treating the Marcellus Shale formation [e.g., Houston et al,. 2009l; Bang et al., 2010]. The reported advantages of FR include effective reduction of the power required to move large volumes of fluid, and a reduced volume of the necessary stimulation fluids. Several field trials and tests have been carried out, and the results show that the formation damage can be reduced [Houston et al., 2009; Bang et al., 2010]. Polymers as wettability alteration additives that have been reported include FC 722 [Tang et al., 2002; Kewen et al., 2000], FC 754 [Kewen et al., 2000], FC 759 [Tang et al., 2002] and FC 4430 [Bang et al., 2010]. The function of their fluoro-chemical group is for water repellant, as indicated in Figures 4.3 and 4.4 (from [Kewen et al., 2000]). Lab tests on Berea sandstone and Kansas Chalk [Tang et al., 2000; Kewen et al., 2000] with two of these surfactants (FC 222 and FC 754) show substantial impacts on the measured contact angles for both water and oil, as Figures 4.5 and 4.6 indicate [Tang et al., 2000]. These lab-scale and field-scale tests of wettability alteration polymers provided a number of important observations [Tang et al., 2000; Kewen et al., 2000; Bang et al., 2010] as follows: a. Wettability of the rock was altered from strongly water-wet to intermediate gas- wet by FC 754 and FC722. b. These compounds alter the wetting characteristics of the materials permanently due to their stabilities. c. “Chemical Treatment showed an improvement factor of greater than 2 for surfactant concentrations ranging from 0.1 and 2%” [Bang et al., 2010]. d. Surfactant adsorbs on the rock surface, the amount varies with surfactant concentrations. e. Numerical simulations show well productivity can be increased by as much as 40- 50% with surfactant treating [Bang et al., 2010]. 112 Figure 4.3: Water (Without Chemical Treatment). Figure 4.4: Water (Treatment With 2% FC 722). 113 Figure 4.5: Contact Angle vs. Concentration of FC 754 Glass Tube Figure 4.6: Contact Angle vs. Concentration of FC 722 Glass Tube 114 4.2 Experimental Study of a Novel Surfactant Additive The objective of this part of the study is to evaluate novel surfactant additives which have the ability to strongly adhere onto the rock surfaces and to change the wetting characteristics. The goal is to be able to alter the characteristics of the gas reservoir from water-wet to intermediate-wet or even gas-wet in order to mitigate potential water blockage issues, and thus to improve production performance. After a systematic literature search, we chose for further study a low-cost surfactant, surfactant X manufactured by Company Y (because of IP restrictions, we are unable to provide additional information at this point and time), which showed the potential to meet our goals. Surfactant X is reported by its manufacturer to function at very low concentrations (less than 100 ppm), which makes its use as an additive in shale-gas operations promising from a cost standpoint. Surfactant X is reported by the manufacturer to migrate to the surface even at very low concentrations to provide the surface with oil and water repellency. Based on the Company Y, surfactant X can be used to reduce the wettability of solid materials. 115 4.3 Lab-Scale Wettability Alteration Experiments In our research, a series of contact angle experiments as well as spontaneous imbibition experiments were performed in close collaboration with my USC colleague Ms. Roychaudhuri as we worked together on developing approaches toward analyzing the wettability alteration characteristics of surfactant X and she carried out the experiments. As described by [Roychaudhuri et al., 2013]:“The contact angles were measured using a Ram é-Hart Model 290-F1 automated Goniometer/Tensiometer. Shale samples with a surface area of approximately 1 cm 2 were prepared from the horizontal direction and were dry-polished using 600 grit SiC polishing paper on a rotating polisher. Spontaneous imbibition experiments have been performed with 1 cm 3 cubes of shale with sample dimensions (1 cm x 1 cm x 1 cm). Only one face of the cube, perpendicular to the bedding plane, was exposed to the water while the other five faces were coated using a high-strength epoxy (resistant up to ~250 o C and ~17,236 kPa) to make them impermeable to water. The experimental set-up used for the measurements is a Mettler Toledo AT 201 microbalance equipped with a densimeter kit containing 200 ml of DI water. The sample weight and temperature were recorded by a PC and analyzed based on the Archimedes principle. Because of the small size of the samples, we assume that the driving force for water uptake is capillarity and all other effects, including gravity, are ignored.” Table 4.4 shows the summary of the contact angle experiments as reported by [Roychaudhuri et al., 2013]. (Note that Sample #14 is from depth 7852.5ft while Sample #17 is from depth 788.5 ft). Figure 4.7 illustrates the experimental process. “The surfactant X, shows a beneficial effect by substantially increasing both the static and dynamic contact angles for the two shale samples, with the increase in contact angle being greater for higher surfactant concentrations” [Roychaudhuri et al., 2013]. Furthermore, Table 4.4 indicates that “concentrations as low as 30 ppm are sufficient to increase the contact angle by as much as 40-50% for the samples from the two depths” [Roychaudhuri et al., 2013]. 116 Based on [B. Roychaudhuri et al. 2013]:“It should be noted that contact angle measurements are a very sensitive indicator of the sample’s surface composition and roughness, and inherent to the technique, thus, lies the challenge of its usefulness to characterize highly heterogeneous materials like shales. We have found the repeatability to be good for sample surfaces that lie side-by-side at the exact same position in the core with variations in measurements that are typically less 10%, and more commonly much less than that (see Table 4.5). This observation lends credence to using the technique to study the impact of exposure to surfactants using the same sample surface. The same is not true, however, for sample surfaces taken a few cm apart, vertically, in the same core; measurements can then vary by more than 10 degrees.” Table 4.4: Summary of Contact Angle Experiments Table 4.5: Variability in Contact Angle Measurements within the Same Sample X X X 117 Figure 4.7: Illustration of Contact Angle Measurement Experiments Figure 4.8, “compares the imbibition behavior of two neighboring samples (cube 1 and cube 2) from sample #17 (7885.5 ft)” [Roychaudhuri et al., 2013]. Cube 1 was treated with DI water, cube 2 was with DI water (containing 30 ppm of the surfactant X) [Roychaudhuri et al., 2013]. From Figure 4.8, “we observe that the shale sample (cube 2) exposed to the surfactant solution imbibes less water than the sample exposed to DI water” [Roychaudhuri et al., 2013]. As described by Roychaudhuri et al. [2013]: “The experiment was then repeated with four additional neighboring samples from sample #17 (cube 3 and 4 are neighboring samples, and so are cubes 5 and 6). In these imbibition experiments cubes 3 and 5 were exposed to pure DI water for 43 hr, while cubes 4 and 6 were exposed to a 60 ppm of surfactant (surfactant X) solution.” Figure 4.9 reports the observed imbibition behavior, a fairly good repeatability of the imbibition behavior can be found among the neighboring samples shown in Figure 4.8. It is clearly indicated by Roychaudhuri et al. [2013] that “The water up-take is consistently reduced when the surfactant is used. Contact angle measurements using DI water were performed on cubes 1, and 2, upon the completion of the spontaneous imbibition experiments.” The observed contact angles are reported in Table 4.6 [Roychaudhuri et al., 2013]. “The static and 1 cm 3 Shale Cube Drop 1: 5 Microliter DI Water Drop 2: 5 Microliter 30ppm Surfactant X solution Drop 3: 5 Microliter 60ppm 118 receding contact angles for cube 2 exposed to surfactant is significantly higher than that of the neighboring cube 1 exposed to DI water only, which indicates that this surfactant, even at very low ppm-level concentration, creates lasting changes in the wettability of the shale samples” as described by Roychaudhuri et al. [2013]. Figure 4.10 [Roychaudhuri et al., 2013] presents the change in the static contact angles before and after the surfactant treatment . Figure 4.8: Spontaneous Imbibition for DI and 30 ppm Surfactant X Table 4.6: Contact Angle Measurements on Samples Exposed to Surfactant Solution after 43 hr. X 119 Figure 4.9: Spontaneous Imbibition for DI and 60 ppm Surfactant X Figure 4.10: Comparison of Contact Angle before and after Surfactant X Treatment 120 4.4 Lab-Scale High-Pressure, Flow-Back Experiments During hydraulic fracturing, proppants such as sand are used to ensure that the fractures that are created remain open following the stimulation to keep the high-permeability path ways for gas flow open. However, the water used during the hydraulic fracturing process shows a tendency to imbibe into the formation and, hence, to block the pathways created for the gas, which affects the production rate of a well. If our hypothesis that surfactant X will change the wettability holds true under formation conditions, then this should result in a reduced water blockage and should, hence, result in a higher production rate. High- pressure, flow-back experiments are valuable in terms of understanding the changes in the amount of flow-back water that the presence of surfactant X may generate. Such experiments have, therefore, been carried out as part of our research. 4.4.1 Experimental Approach In order to study the wettability altering ability of Surfactant X and to interpret the flow- back behavior, I have modified (jointly with Ms. Roychaudhuri) our high-pressure depletion experimental apparatus in order to carry out this type of experiments, as shown in Figure 4.11. The individual components, shown in Figure 4.11 (all with a minimum pressure rating of 5000 psig), are listed in Table 4.7. Table 4.7: Components for Figure 4.11 A Argon G Argon Buffer Tank M Pressure Gauge B Methane H Pressure Regulator N Confining Pressure Pump C Air I Pressure Gauge D Gas Booster J Syringe Pump E Pressure Gauge K Computer F CH 4 Buffer Tank L Core-holder 121 Figure 4.11: High-Pressure, Flow Back Experimental Set-up C C C C B B B B D D D D F F F F E E E E H H H H L L L L N N N N G G G G A A A A I I I I M M M M J J J J K K K K 122 Very similar to our High-Pressure Depletion System, our High Pressure Flow-back experimental system consists of four different sub-systems: 1) The fluid injection sub-system, including J and K. A syringe pump, controlled by a computer, is used to inject the desired fluid into the core. This sub-system can be isolated from the rest of the setup. 2) The high-pressure gas booster sub-system, including A, B, C, D, and E. Methane (B) or Argon (A) feeds into the booster (D). The booster is operated by air (C) to boost the gas to a desired pressure. The high-pressure gas is then stored in the storage system (sub-system 3). 3) The high-pressure storage sub-system, including F, G, H and I. Gas from sub- system 2 is stored in the CH 4 (or Argon) buffer tank. When an experiment is initiated, the pressure regulator (H) is set to deliver a selected pressure to the experimental system (sub-system 4). The exact delivery pressure can be monitored by a pressure gauge (I). 4) The core-holder section, including L, M and N. The core-holder (L) is a stainless steel Hassler-type core-holder capable of withstanding pressure at least 5000 psig and holds the gas-shale core plug under investigation. Confining pressure is supplied via a hydraulic pump (N). (Figure 4.12 shows the technical details of the core-holder and flow-back experimental section). The experimental data are then recorded and analyzed via a computer. 123 Figure 4.12: Main Experimental System H ydr aul i c Pum p H ydr aul i c Pum p H ydr aul i c Pum p H ydr aul i c Pum p Val ve Val ve Val ve Val ve 4 4 4 4 C or e H ol der C or e H ol der C or e H ol der C or e H ol der Vi t on Sl eeve Vi t on Sl eeve Vi t on Sl eeve Vi t on Sl eeve C or e C or e C or e C or e V al ve V al ve V al ve V al ve 3 3 3 3 V al ve V al ve V al ve V al ve 1 1 1 1 V al ve V al ve V al ve V al ve 2 2 2 2 124 4.4.2 Experimental results The flow-back experiments with the modified high-pressure depletion set-up were performed to test Surfactant X in close collaboration with Ms. Roychaudhuri using the system shown in Figure 4.12 according to the following experimental procedure: 1) Apply 3000 psig of confining pressure to the core-holder: 2) Inject 2000 psig of gas into the core: a. Close valves 2 and 3, b. Open valves 1 and 4. Gas with desired pressure will then flow into the core. 3) Empty-out gas in the line: a. Close valve 1, so that the gas cylinder will be isolated. b. Close valve 4, so that the gas can be sealed within the core-holder, c. Open valves 2 and 3, so that gas in the lines can be emptied out. 4) Fill the line with the fluid: a. Keep valves 2 and 3 open. Fluid will be injected to fill the line via valve 2. Pump the fluid until we see it coming out of valve 3; this is to make sure that the fluid occupies all the dead volume. As a result, all new fluid we inject into the system will be imbibed by the core. 5) Inject fluid into the core: a. Keep valve 2 open, close valve 3. b. Open valve 4 to inject the desired fluid with constant pressure; injection will stop once the core is saturated, with the volume of fluid injected recorded in the computer. 6) Fluid flow-back: a. Stop supplying pressure to the core from the syringe pump. b. Open valve 4, fluid in the core will flow-back to the syringe pump. c. How much fluid comes back is recorded in the computer. 7) Perform steps (1) to (6) for both DI water alone, and for DI water solution with 60 ppm of Surfactant X in order to compare the differences in the amount of flow- back we get. 125 A summary of the data extracted from a flow-back experiment is provided in Table 4.8. The core was selected from a depth of 7880.5 ft, which is the same core used in the high- pressure depletion experiments. Our flow-back experiments show that we can increase the load recovery by more than 30% when Surfactant X is added to the injected fluid. This, in turn, indicates that 30% less water will remain in the formation to clog the pathways of the shale gas, and hence an increase in the production of gas is expected. Table 4.8: Forced Imbibition Data Summary In the above experiments the surfactant (surfactant X) was dissolved in ethanol prior to being dissolved in the DI water. This approach presents challenges in terms of its field implementation, however. In our research we have studied, therefore, whether there is a better way to add the surfactant X into the fracturing fluid. The most obvious choice is to add the surfactant to the commonly utilized friction reducer (FR), as it is a polymer compound itself and surfactant X should dissolve easily in certain polymers. We have investigated, therefore, the solubility of Surfactant X in the friction reducer as well. Our results indicate that the solubility (under shaking for 2 hr) of surfactant X in the friction reducer is 65 mg/cc of FR at 50 o C. By comparison, the solubility of surfactant X in ethanol (at room temperature – heating ethanol under field conditions is not recommended because it is a flammable substance) is 10 mg/cc ethanol. Porosity (without confining pressure) 13.7 Forced Imbibition 1 DI Injected, cc 0.841 Weight Retained, g 0.407 Recovery, % 51.6 Porosity (without confining pressure) 13.6 Forced Imbibition 1 60 ppm Surfactant X Injected, cc 0.8374 Weight Retained, g 0.124 Recovery, % 85.2 126 4.5 Analysis for Field-Scale Implementation As indicated by the experiments above, the addition of the Surfactant X shows a good potential for improving the hydraulic stimulation process through reduced water-blocking. Since the use of this additive involves an additional expense, we provide a preliminary estimate of the costs alongside with the potential benefits to be realized in this section. In our study, we first calculate the current net return per well, the cost of using the proposed additive per well, and the expected net return per well with and without the use of the additive. We based our estimates on the use of ethanol as a solvent, thus the costs estimated represent a likely worst case scenario (see discussion above about using the FR in lieu of using the ethanol). The estimate of ultimate recovery (EUR) for an active Marcellus Shale well averages 3.5 Bcf (ftp://ftp.eia.doe.gov/natgas/ usshaleplays.pdf) and we use this estimate in our calculations. Assuming an average gas price during the lifespan of a well at $2.5/mmbtu (this price is close to the historic low for natural gas, thus the estimates for returns are very conservative) then the ultimate gross income per well is calculated to be $ 8.75MM. With an estimated average cost per well in the Marcellus Shale of $4.5MM, the income (before taxes and other expenses) is $4.25MM per well representing a return of 94%. The cost for implementing the use of surfactant includes the costs for the Surfactant X itself as well as the ethanol (we assume here that no extra fees will be incurred by the service company for implementing the use of surfactant). We consider here three different cases: (1) Adding surfactant only during the pad stage at a concentration level of 60 ppm; (2) Adding the surfactant during the pad stage at a concentration level of 60 ppm, followed by continuous addition for the remainder of the frac-job at a concentration level of 30 ppm; (3) Adding the surfactant throughout the frac-job at a concentration level of 30 ppm. For each of these cases we assume that the surfactant is dissolved in ethanol (at its maximum solubility of 10 mg/cc) prior to the mixing. Table 4.9 below summarizes the cost of using the additive including the ethanol solvent for Case 1 above. Table 4.10 shows the cost incurred by handling (including treatment) of the additional flow-back water (in these calculations we assume a 10% increase in flow- 127 back water which is 1/3 of the effect we observe in our lab-scale experiments). Based on the calculations shown in these two Tables, the total added costs for Case 1 are $ 64,484. Table 4.9: Cost Estimate for Use of Proposed Additive in Well Stimulation-Case 1 Pad volume per stage (including alcohol) 50000.0 gallons 193500.0 liters Additive per stage (60 ppm) 11.6 kg 25.8 lbs Cost of X 111.0 $/lbs Cost of X per stage 2864 $ Cost of X per well (10 stages) 28640 $ Solubility of surfactant X in 80 proof 10 mg/cc 0.01 kg/liter 0.022 lbs/liter Volume of solvent (80 proof) per stage 1173 liters Volume of solvent (80 proof) per well 11730 liters 3031 gallons Cost of solvent (80 proof) 2.4 $/gallon Cost of solvent (80 proof) per well (10 stages) 7272 $ Extra cost of stimulation per well (10 stages) 35912 $ Table 4.10: Extra Cost Associated with Flow-back Water Current amount of injected fluid during frac job per stage 300,000 gallon 7143 bbl Increase of flow-back in percentage 10 % Cost in terms of treating and handling flow- back water 4 $/bbl Extra cost for flow-back per stage 2857.2 $ Extra cost for flow-back water treatment per well (10 stages) 28572 $ 128 Table 4.11 below summarizes the cost of using the additive and solvent for Case 2. The cost for handling the additional flow-back water is the same as in Case 1, and is listed in Table 4.10. The total estimated costs for Case 2 are $ 118,207. Table 4.11: Cost Estimate for Use of Proposed Additive in Well Stimulation-Case 2 Pad volume per stage (including alcohol) 50000.0 gallons 193500.0 liters Additive per stage (60 ppm) 11.6 kg 25.8 lbs Cost of X 111.0 $/lbs Cost of X per stage 2864 $ Cost of X per well (10 stages) 28640 $ Solubility of X in 80 proof 10 mg/cc 0.01 kg/liter 0.022 lbs/liter Volume of solvent (80 proof) per stage 1173 liters Volume of solvent (80 proof) per well 11730 liters 3031 gallons Cost of solvent (80 proof) 2.4 $/gallon Cost of solvent (80 proof) per well(10 stages) 7272 $ Cost of pad treating per well (10 stages) 35912 $ Slurry volume per stage (including alcohol) 250,000 gallons 967500.0 liters Additive per stage (30 ppm) 29.0 kg 64.4 lbs Cost of X 111.0 $/lbs Cost of X per stage 7148.4 $ Cost of X per well (10 stages) 71484 $ Solubility of X in 80 proof 10 mg/cc 0.01 kg/liter 0.022 lbs/liter Volume of solvent (80 proof) per stage 2927 liters Volume of solvent (80 proof) per well 29270 liters 7563 gallons Cost of solvent(80 proof) 2.4 $/gallon Cost of solvent (80 proof) per well(10 stages) 18151 $ Cost of slurry treating per well (10 stages) 89635 $ 129 Table 4.12 below summarizes the costs of the additive and ethanol for Case 3. Again, the cost incurred by handling the additional flow-back water is the same as in Case 1. The total costs for Case 3 amount to $ 136,164. Table 4.12: Cost Estimate for Use of Proposed Additive in Well Stimulation Frac-fluid per stage (including alcohol) 300000.0 gallons 1161000.0 liters Additive per stage (30 ppm) 34.8 kg 77.3 lbs Cost of X 111.0 $/lbs Cost of X per stage 8580 $ Cost of X per well (10 stages) 85800 $ Solubility of X in 80 proof 10 mg/cc 0.01 kg/liter 0.022 lbs/liter Volume of solvent (80 proof) per stage 3514 liters Volume of solvent (80 proof) per well 35140 liters 9080 gallons Cost of solvent (80 proof) 2.4 $/gallon Cost of solvent (80 proof) per well(10 stages) 21792 $ Total cost of stimulation per well (10 stages) 107592 $ Assuming that the 30% increase in flow-back, observed at lab-scale, will increase the field-scale EUR by only 10%, the new income before tax per well will be $9.63MM. The net returns for the three cases above are then calculated to be 111%, 108%, and 107.5%, respectively. 130 4.6 Investigation of Surfactant X under Actual Frac-Fluid Conditions Since the preliminary estimates described above indicate that the use of the surfactant X has the potential to substantially improve the net return on investment, additional experiments were undertaken to verify that the surfactant will function adequately under the true field-scale conditions. (We note that all the previous investigations were performed with alcohol as the solvent for surfactant X). To do that, we have carried contact angle measurements as well as spontaneous imbibition experiments using real frac-fluids provided by our industrial collaborators the Energy Corporation of America (ECA), as detailed below. 4.6.1 Contact Angle Measurements For these measurements, we have used two different shale samples (same procedure as described in Section 4.2) from the same depth (7885.5 ft) side-by-side with which we have conducted DI water/surfactant X contact angle measurements previously (see Table 4.4). The contact angle measurements were then made with three different fluids: (1) an ECA frac-fluid; (2) the same ECA frac-fluid with 30 ppm surfactant X; (3) the same ECA frac-fluid with 60 ppm surfactant X. Table 4.13 shows the contact angle measurement results (the analysis method is the same as Figure 4.7 illustrates). And, very similar to the observation in Section 4.2, as Table 4.13 indicates, the shale samples are highly heterogeneous materials and there is certain level (less than 10 degrees) of scatter in the measured values even at the same sample depth. However, from Table 4.13, we are still able to observe the same beneficial effect of surfactant X, as in Section 4.2, of substantially increasing both the static and dynamic contact angles for the two shale samples under the actual frac-fluid, with the increase in contact angle being greater for higher surfactant concentrations. The results, again, validate our conclusion in Section 4.2 that even a low-concentration (as low as 30ppm) surfactant X solution has the ability to significantly alter the wettability. 131 Table 4.13: Summary of Contact Angle Measurements with the Frac-Fluid. 4.6.2 Spontaneous Imbibition Experiments Three identical shale samples (see Table 4.14 for dimensions and other characteristics) have been used to study spontaneous imbibition of real frac-fluid with and without additive. These samples were selected from the core at depth 7885.5 ft, for which we have also conducted spontaneous imbibition experiments using the ethanol/surfactant X solution (see Section 4.2). The three different fluid compositions used in the above contact angle measurements were also used for the spontaneous imbibition experiments. Figure 4.13 shows the fluid uptakes as a function of time for the three fluids investigated, while Figure 4.14 compares two experiments with the ECA frac-fluid containing 30 ppm and 60 ppm of the surfactant X, respectively. When comparing the results in Figure 4.13 with those in Figure 4.8 (Section 4.2) showing the spontaneous imbibition results for the Di water/ surfactant X solution, it is clear that the water uptake curve reaches a plateau in a much shorter time with the frac-fluid/surfactant X solutions than for the ethanol/surfactant X solution. Furthermore, the effect is more significant the higher the surfactant X concentration in the frac-fluid is. These findings suggest that the presence of the surfactant X surfactant either dissolved in the frac-fluid or in DI water causes the sample shale samples to take in much less fluid under spontaneous imbibition conditions. Compare to DI water alone, the surfactant X dissolved in the real frac-fluid exhibits a higher wettability alteration ability, as manifested by the significantly smaller volumes of fluid that are imbibed in the shale samples. 7885.5 ft Sample 1 Sample 2 Advancing (Deg) Static (Deg) Receding (Deg) Advancing (Deg) Static (Deg) Receding (Deg) Frac fluid 53.4 50.3 21.3 62.3 55.8 17.1 30 ppm 62.5 56.2 22.2 70.7 62.1 17.8 60 ppm 81.2 76.2 29.7 76.7 71.4 18.3 132 Table 4.14: Dimensions and Other Characteristics of the Samples Used in the Spontaneous Imbibition Experiments 7885.5ft Sample 1 Sample 2 Sample 3 Fluid info ECA frac-fluid 30 ppm 60 ppm Length (mm) 12.05 11.81 12.04 Breadth (mm) 11.81 11.84 11.75 Thickness (mm) 11.67 11.72 11.37 Wt Initial, g 4.2067 4.2194 4.1936 Wt+Epoxy, g 4.7708 4.7427 4.7388 Porosity (%) 5.01 5.03 4.92 Figure 4.13: Spontaneous Imbibition Results for 3 Fluid Compositions 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 0.02 0.022 0 5 10 15 Weight Change (g) Time (h) Comparison of Imbibition Frac Fluid Frac Fluid/30 (ppm) X Frac Fluid/60 (ppm) X 133 Figure 4.14: Spontaneous Imbibition for Frac Fluid/30ppm and Frac Fluid/60ppm 4.6.3 Potential Interactions of the Surfactant X with the Proppants. One concern related to the use of the surfactant under real field conditions is the potential loss via adsorption on the surface of the sand commonly used as a proppant during well stimulation. If adsorption onto the proppant indeed occurs, a higher initial surfactant concentrations must be utilized in order to overcome such losses and to maintain the effectiveness of the additive. To investigate the potential loss of surfactant to proppant (sand), we have designed and carried out both static and flow-through experiments. The purpose for the static experiments is to determine what the maximum loss of the surfactant X to the sand could potentially be. The flow-through experiments were performed to estimate the real additive loss under the actual hydraulic stimulation conditions. The static experiments (illustrated in Figure 4.15) have been performed with 100 cc of a 60 ppm surfactant X solution (in frac-fluid) and 11.6 g of 100-mesh white sand provided by ECA. Accordingly, the sand to fluid ratio is 1 lb/gal, which the 0 0.0002 0.0004 0.0006 0.0008 0.001 0.0012 0.0014 0.0016 0.0018 0.002 0 2 4 6 8 Weight Change (g) Time (h) Comparison of 30ppm and 60ppm Frac Fluid/30 (ppm) X Frac Fluid/60 (ppm) X 134 maximum sand/fluid ratio for 100-mesh white sand used in the field. At time t=0, 11.6 g of the 100-mesh white sand was added to the 60 ppm surfactant X solution, the mixture was then capped (in order to prevent evaporation of the fluid) and put into a temperature bath at 50 o C. The concentration of additive was then measured by a Waters 486 GPC/HPLC system. A 4 week adsorption experiment (with fluid samples collected at 0 hr, 8 hr, 18 hr, 30 hr, 48 hr, 3 days, 4 days, 5 days, 6 days, 7 days, 14 days, 21 days, and 28 days) was carried out. Figure 4.16 shows the adsorption of surfactant X on the 100-mesh white sand as a function of time. From Figure 4.16, one observes that the adsorption curve has a very similar behavior as what Paria et al. [2005; 2007] found in their investigation of the adsorption of a non-ionic surfactant onto sand. However, a much longer time is required to reach equilibrium for the surfactant X compared to their nonyl- phenyl ethoxylates surfactant. The final saturated amount of surfactant X on 100-mesh sand is 0.095 mg/g, which is the maximum possible loss from a 60 ppm surfactant X solution to the 100-mesh sand. This result also implies that in order to mitigate the loss of surfactant, operators could lower their shut-in time, if applicable. Figure 4.15: Static Adsorption Experiment Set-up E-1 Water Bath with 50 Deg. C Sample collector Concentration measurement system (Gel Permeation Chromatography) T 100mesh sand with initial 60ppm FS-1400 solution 135 Figure 4.16: Adsorption of Surfactant X onto the 100 Mesh Sand We have also performed flow-through experiments in order to verify the potential loss dynamics of the surfactant under real frac-job conditions. A flow-through experiment (as illustrated in Figure 4.17) system has been designed to fulfill this objective. A column (1 in in diameter and 7 in in length) was packed with 100-mesh sand; both sides of the column were equipped with 40-mesh stainless steel sieves in order to prevent potential leaking of sand. A fluid booster was used to pump the surfactant X from a source tank into the sand column with a constant velocity 0.66 m/s (superficial velocity calculated with respect to the total cross-section of the sand column). The produced fluid was returned to the source tank. The concentration in the source tank was measured (0 min, 90 min and 180 min) with a Waters 486 GPC/HPLC system. Since the fluid velocity in the wellbore during a frac-job varies from 0.35 m/s to 16.8 m/s (based on the typical range of 2-95 bpm (barrel per min) of injection rates during the frac-job), the 0.66 m/s velocity reflects the lower range of the injection speed. And if there is little or no loss of surfactant at lower fluid velocity, a higher velocity will further mitigate any possible loss 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 5 10 15 20 25 30 Adsorption (mg-X/g-sand) Adsorption Time (days) 136 of surfactant to the sand compare to low-velocity conditions. Figure 4.18 shows the flow- through experimental results. From Figure 4.18, we observe that there is no concentration change over a period of 180 min. Considering that a typical, per stage, frac-job is around 90 min, this result indicates that operators should not see any significant loss of surfactant X during the 90 min per stage frac-job period. Figure 4.17: Flow-through Sand Experimental Set-up Sand Column Fluid Booster Water Bath at 50 Deg. C Sample 137 Figure 4.18: Adsorption Loss at 0.66 m/s Fluid Velocity 4.7 Summary In this Chapter, we have reviewed some current practices related to mitigation of water blockage effects following well stimulation with water-based fracturing fluids. Hydraulic fracturing requires large amounts of water-based fluids, usually 2-5 million gal/per well, in order to create the expected extra permeability. As a result of the formation’s hydrophilic nature, a large portion of the frac-fluid tends to imbibe into the matrix and blocks the escape route for shale-gas. In order to mitigate the water-blockage, a detailed investigation of a novel surfactant has been performed. Our contact angle experiments indicate that surfactant X (solvent = ethanol) increases the contact angle of the Marcellus Shale sample by as much as 40-50% with concentration as low as 30 ppm compared to DI water alone. The spontaneous imbibition experiments with the same surfactant X solution show that the Marcellus shale imbibes 40% less water, as compared to DI water without the surfactant. Moreover, our lab-scale, flow-back experiment with 60 ppm surfactant X solution (solvent = ethanol) indicates the flow-back volume can be increased by as much as 30% compared to the DI water. Based on these experimental results, we analyzed the per-well return for three hypothetical scenarios with surfactant X under the worst economical and operational conditions, and found that a 10% increase in return can -0.01 -0.005 0 0.005 0.01 0 50 100 150 200 Adsorption (mg-FS/g-sand) Adsorption Time (min) 138 be achieved. We also conducted lab-scale experiments with actual frac-fluid compositions to verify the effectiveness of surfactant X. Our lab-scale contact angle and spontaneous imbibition experiments with surfactant X in actual frac-fluids show consistent behavior compare to the data obtained from the DI water with additive (ethanol as solvent). We also observed that surfactant X in frac-fluid decreases the imbibition liquid volume by as much as 95% (with 60 ppm of surfactant in the frac-fluid) compared to DI water. We performed static and flow-through experiments to help estimating the potential loss of surfactant X to the proppant during a frac-job. Our static and flow-through experiments indicate that there will be no loss of surfactant during the frac-job, which further leads to the conclusion that surfactant X treatment is a promising approach towards tackling the water blocking issue, and that it will be worthwhile to conduct further field trials. 139 Chapter 5: Summary of Dissertation and Recommendations for Future Work 5.1 Dissertation Summary In this dissertation, we characterized the structural and other related properties relevant to gas production mechanisms for samples from the Marcellus Shale. We also carried out high-pressure depletion experiments in order to evaluate the mass transfer efficiency and the resulting production performance of a shale core sample from the same formation. These experiments were complemented with detailed numerical study of the observed behavior. We also conducted a detailed experimental study in order to evaluate the use of novel surfactants in terms of their ability of reducing the capillarity of the shale samples, and thus improving mass transfer efficiency. Our characterization tests with the shale samples revealed that they possess high heterogeneity. During these studies we investigated two key characteristics: of such shales, namely their permeability and porosity, properties which both relate directly to the mass transfer mechanism of gas in these materials. The study on the permeability of these shales indicates that the samples exhibit a dramatic difference in the values of their horizontal (perpendicular to the core axis) and vertical (along the core axis) permeabilities. Specifically, their horizontal permeability, on the average, tends to be two orders of magnitude higher than the permeability in the vertical direction. The differences among the horizontal and directional permeabilities, signifying a directional permeability heterogeneity, are consistent also with electron microscopy (SEM) images of the same samples: the surfaces perpendicular to the horizontal direction appear to be visibly more porous compared to the surfaces perpendicular to the core axis A important corollary from these observations, of potential value to operators in the Marcellus Shale, is that creating artificial fractures in the vertical direction is, likely, to have a significantly higher impact on mass transfer efficiency during gas production from the formation. Our study of the pore size distribution of the shale samples, via BET, indicates that, based on the IUPAC classification, these materials possess three different levels of porosities: a microporous region, with pores with d p <2 nm, a mesoporous region with pores with 2 140 nm< d p <50 nm, and a region that contains macropore and/or microfractures (d p >50 nm) region. BET analysis indicates that the microporous and mesoporous regions occupy comparable fractions of the pore volume compare, while the macropore and microfracture region represents a significantly smaller part (<10%) of the total pore volume (these samples also contain visible, via the naked eye and/or SEM, fractures of the order of tens to hundreds of microns which are not detectable by BET). The presence of substantial microporous and mesoporous regions is consistent with high-pressure depletion experiments – see further discussion below -- that desorption from the shale matrix and diffusion through these two regions of the shale pore space plays an important role during gas production from such formations. The mass transfer mechanisms during gas production from the shale samples were also investigated via the use of a High-Pressure Depletion System (HPDS). Both the short- term and long-term production behavior of a cylindrical shale core sample was experimentally evaluated. The HPDS is designed as to mimic the actual formation pressure conditions, whereby we can independently control the three different pressures: the initial pore pressure, the back-pressure (depletion pressure), as well as the confining pressure. The short-term experimental data indicate that the gas produced comes from the macropore/microfracture region (as well as the larger fractures present in the sample). This is known as the “free gas” and transports principally via viscous flow at the very beginning of the depletion. The long-term depletion experiments reveal that a significant fraction (approximately 40%) of the total gas produced (and thus of the gas in place) comes from the micropore/mesopore regions (otherwise known as the shale matrix) and requires much longer time to be produced. The mass transfer mechanism during the long- term gas production has been found to be the desorption from the micropore region and diffusion in the micropore/mesopore regions; gas molecules in the micropores transport to the mesopore region via surface diffusion, and these gas molecules then transport through the mesopore region via both viscous (Poiseuille-type) and/or slip-flow (Knudsen). We also find that the Bidisperse Pore Model (BPM) fit the experimental production data obtained from our long term depletion experiments. 141 We have also studied the effectiveness of a novel surfactant (selected from a search of currently available commercial surfactants) on reducing the capillarity of the formation in order to mitigate water-blockage type gas shale formation damage. Contact angle experiments indicate very low concentration (usually 30 ppm to 60 ppm) of the novel surfactant solutions, were able to increase the contact angle by 40 - 50%. Complementary spontaneous imbibition experiments indicate that the surfactant treatment is able to decrease the gas shale water imbibition amount by as much as 95% (with 60 ppm of surfactant in the frac-fluid) compared to a DI water treatment alone. Forced imbibition experiments with the surfactant being present show a 30% increase in the quantity of the flow-back fluid, which signifies a potential for significant mitigation of water-blockage type of formation damage. Static and dynamic surfactant loss experiments on sand, which is used as a common proppant, further indicate that there will be no loss of surfactant during the frac-job. 5.2 Recommendations for Future Work The present study has made progress in understanding the mass transfer mechanisms during gas production of the Marcellus Shale formation, and in identifying a novel surfactant which appears effective in altering the formation wettability in order to mitigate water-blockage type of formation damage. Some recommendations for additional future work are described below: • A more comprehensive characterization of the shale samples with high resolution SEM as well as FIB/SEM will offer additional insight into the causes for the observed differences in permeability in the horizontal and vertical directions. And a broader range of samples from other wells from the Marcellus region need to be characterized in detail and compared with our current findings. This will provide a more comprehensive understanding of the entire formation. 142 • High-Pressure depletion experiments also need to be carried out using larger cores in order to not only increase the accuracy of the experimental results, but also improve the understanding of potential compaction effects resulting by the increasing of net-stress particularly during the short-term period of the depletion process. This will require upgrades to the current High-Pressure Depletion System to handle larger size of core samples. • Although we have found that the BPM describes the long term production behavior, as well as the time scales of gas transport in micropore region and mesopore region of our core sample, more experiments with different cores and various confining pressure are also needed to be conducted via the High-Pressure Long Term Depletion experiments in order to further evaluate the prediction performance of the BPM. • Last, but not least, field tests are needed in order to investigate the wettability alteration ability of the proposed surfactant under the actual formation conditions. The field tests results would allow the further optimization efforts with the chosen novel surfactant, and may suggest further lab-scale experimental evaluation of other surfactants with similar wettability alteration characteristics. 143 Bibliography Adam Sieminski (2013), Outlook for shale gas and tight oil development in the U.S., Joint Forum on US Shale Gas & Pacific Gas Markets May 14, 2013. Agarwal, Ram, et al. (1998). "Analyzing well production data using combined type curve and decline curve analysis concepts." SPE Annual Technical Conference and Exhibition. ALL Consulting and the Montana Board of Oil and Gas Conservation (2004). Coal Bed Methane Primer: New Source of Natural Gas – Environmental Implications. Prepared for the U.S. Department of Energy, National Petroleum Technology Office. February. Ambrose, Raymond, et al. (2010)."New pore-scale considerations for shale gas in place calculations." SPE Unconventional Gas Conference. Ambrose, W. A., E. C. Potter, & R. Briceno (2008), An Unconventional Future for Natural Gas in the United States. http://www.geotimes.org/feb08/article.html?id=feature_gas.html. February. AMMRF 2012, http://www.ammrf.org.au/myscope/analysis/eds/, Accessed July 2013. Anderson, David, et al. (2010). "Analysis of production data from fractured shale gas wells." SPE unconventional gas conference. Asps, J. J. (1945). "Analysis of decline curves." Transactions of the American Institute of Mining, Metallurgical and Petroleum Engineers 160: 228. Bang, Vishal, et al. (2010). "A new solution to restore productivity of gas wells with condensate and water blocks." SPE Reservoir Evaluation & Engineering 13.2: 323-331. Barrett E.P., Joyner L.S., Halenda P.P. (1951), The Determination of Pore Volume and Area Distribution in Porous Substances-I, Computations from Nitrogen Isotherms, J. Am. Chem., 73, 373. 144 B. Roychaudhuri, T. T. Tsotsis, K. Jessen (2013), An Experimental Investigation of Spontaneous Imbibition in Gas Shales, Journal of Petroleum Science and Engineering, 2013. Brunauer S., Emmett P.H. and Teller E. (1938), Adsorption of gases in multimolecular layers., J. Am. Chem. Soc., 60, 309. Chemistry and Technology of Fuels and Oils (2000), Volume 36, Number 2, pp 82-88. March. Couillet, Isabelle, et al. (2005). "Synergistic effects in aqueous solutions of mixed wormlike micelles and hydrophobically modified polymers." Macromolecules38.12: 5271-5282. Dadwhal, Megha, et al. (2009) "Adsorption of arsenic on conditioned layered double hydroxides: column experiments and modeling." Industrial & Engineering Chemistry Research 48.4: 2076-2084. Do Duong, D. (1998) Adsorption analysis: equilibria and kinetics. Vol. 2. Imperial College Pr. DOE/EIA-0216 (1998), U.S. crude oil, natural gas, and natural gas liquids reserves— 1998 annual report: U.S Department of Energy, Energy Information Administration Report. DOE/EIA-0216, 150 p. Dombrowski, Robert J., and Christian M. Lastoskie. (2002) "A two-stage Horvath- Kawazoe adsorption model for pore size distribution analysis." Studies in Surface Science and Catalysis 144: 99-106. Doublet, L. E., et al. (1994). "Decline curve analysis using type curves--analysis of oil well production data using material balance time: application to field cases."International Petroleum Conference and Exhibition of Mexico. EIA (1999), Natural Gas 1998: Issues and Trends. http://www.eia.doe.gov/pub/oil_gas/natural_gas/analysis_publications/natural_gas_1998 _issues_trends/pdf/it98.pdf. April 1999. 145 EIA (2002) Manufacturing Energy Consumption Survey. http://www.eia.doe.gov/emeu/mecs/mecs2002/data02/shelltables.html. March 2005. EIA (2007). International Energy Outlook 2007, Chapter 4: Natural Gas. http://tonto.eia.doe.gov/FTPROOT/forecasting/0484(2007).pdf. May 2007. EIA (2008), Is U.S. Natural Gas Production Increasing? http://tonto.eia.doe.gov/energy_in_brief/print_pages/natural_gas_production.pdf Accessed: May 2010. EIA (2010), Natural Gas Basic Statistics. http://www.eia.doe.gov/basics/quickgas.html. Accessed: May 2010. Energytomorrow.com, Shale gas, http://energytomorrow.com/Shale_Gas.aspx Accessed: May 2010. Faraj B, Williams H, Addison G, McKinstry B, et al: (GTI Canada) (2004), “Gas Potential of Selected Shale Formations in the Western Canadian Sedimentary Basin,” GasTIPS 10, number 1 (Winter): 21–25. Fetkovich, M. J., E. J. Fetkovich, and M. D. Fetkovich (1996). "Useful concepts for decline curve forecasting, reserve estimation, and analysis." SPE Reservoir Engineering 11.1: 13-22. Frantz, J.K., and Jochen, V (2005), Schlumberger. Shale Gas White Paper. 05-OF299. Schlumberger Marketing Communications. October. G. Horvath and K. Kawazoe (1983), Journal of Chemical Engineering of Japan, 16(1983)470. Geology.com. Marcellus Shale- Appalachian Basin Natural Gas Play: New Research Results Surprise Everyone on the Potential of this Well-known Devonian Black Shale. http://geology.com/articles/marcellus-shale.shtml. Accessed: May 2010. 146 Ghassemzadeh, J., Xu, L., Tsotsis, T.T., Sahimi, M. (2000) “Statistical Mechanics and Molecular Simulation of Adsorption in Microporous Materials: Pillared Clays and Carbon Molecular Sieve Membranes”. J. Phys. Chem. B, 104, 3892. GRI-95/0496 (1996), Gas Research Institute, April 1996. Ground Water Protection Council (2009), Modern Shale Gas Development in the United States: A Primer, April. Hill, D. G., and C. R. Nelson (2000), Gas productive fractured shales—an overview and update: GasTIPS, v. 6, no. 2, p. 4–13. Houston, Nathan, et al. (2009). "Fracture-stimulation in the marcellus shale-lessons learned in fluid selection and execution." SPE Eastern Regional Meeting. JB Curtis (2002), AAPG Bulletin, v. 86, no. 11 (November), pp. 1921–1938. John Lynch et al. (2003). Physico-Chemical Analysis of Industrial Catalysts: A Practical Guide to Characterisation. pp 119-120. Kawata, Yuko, and Kazuo Fujita (2001), "Some predictions of possible unconventional hydrocarbons availability until 2100." paper SPE 68755. Kewen, Li, and Firoozabadi Abbas. (2000). "Experimental study of wettability alteration to preferential gas-wetting in porous media and its effects." SPE Reservoir Evaluation & Engineering 3.2: 139-149. Kuila, Utpalendu, and Manika Prasad. (2013) "Specific surface area and pore‐size distribution in clays and shales." Geophysical Prospecting. Leythaeuser, Detlev, Rainer G. Schaefer, and Arif Yukler. (1982) "Role of diffusion in primary migration of hydrocarbons." Am. Assoc. Pet. Geol., Bull.;(United States) 66.4. Liu, H., Zhang, L. and Seaton N.A. (1992) “Determination of the connectivity of porous solids from nitrogen sorption measurements -- II. Generalisation” Chemical Engineering Science, Vol. 47(17-18), December, pp. 4393-4404. 147 Lucas, Gloria, et al. (2011). "Understanding Unconventional Gas Reservoir Damages."OTC Brasil. Maestro, Alicia, et al. (2004). "Formation and disruption of viscoelastic wormlike micellar networks in the mixed surfactant systems of sucrose alkanoate and polyoxyethylene alkyl ether." The Journal of Physical Chemistry B 108.37: 14009-14016. Maloney, James O. (2008). Perry's chemical engineers' handbook. Vol. 7. New York: McGraw-Hill. NaturalGas.org. Overview of Natural Gas. Background. www.naturalgas.org/overview/background.asp. Accessed: September 2008. NaturalGas.org. Electrical Generation Using Natural Gas. http://www.naturalgas.org/overview/uses_eletrical.asp. Accessed: May 2010. Paktinat, Javad, et al. (2011). "High brine tolerant polymer improves the performance of slickwater frac in shale reservoirs." North American Unconventional Gas Conference and Exhibition. Palacio, J. C., and T. A. Blasingame. (1993). "Decline curve analysis using type curves: Analysis of gas well production data." paper SPE 25909: 12-14. Paria, Santanu, C. Manohar, and Kartic C. Khilar (2005)."Kinetics of adsorption of anionic, cationic, and nonionic surfactants." Industrial & engineering chemistry research 44.9: 3091-3098. Paria, Santanu, and Pak K. Yuet. (2007). "Adsorption of non-ionic surfactants onto sand and its importance in naphthalene removal." Industrial & engineering chemistry research 46.1: 108-113. Peebles, M. W. H. (1980), Evolution of the gas industry: New York, New York University Press, 235 p. 148 Penny, Glenn, John Pursley, and David Holcomb. (2005). "The application of microemulsion additives in drilling and stimulation results in enhanced gas production." SPE Production Operations Symposium. Penny, Glenn, Terrell Dobkins, and John Pursley. (2006). "Field study of completion fluids to enhance gas production in the Barnett shale." SPE Gas Technology Symposium. Penny, Glenn, and John Pursley. (2007). "Field studies of drilling and completion fluids to minimize damage and enhance gas production in unconventional reservoirs." European Formation Damage Conference. Railroad Commission of Texas (2010). Newark, East (Barnett Shale) Field. http://www.rrc.state.tx.us/data/fielddata/barnettshale.pdf Accessed: May 2010 Reuters 2013 :http://www.reuters.com/article/2013/06/19/utilities-ferc-power- idUSL2N0EV12G20130619 , Accessed July 2013. Rickman, Richard, and Omkar Jaripatke. (2010). "Optimizing Micro-emulsion/Surfactant Packages for Shale and Tight Gas Reservoirs." SPE Deep Gas Conference and Exhibition. Roen, J. B. (1993), Introductory review—Devonian and Mississippian black shale, eastern North America, in J. B. Roen and R. C. Kepferle, eds., Petroleum geology of the Devonian and Mississippian black shale of eastern North America: U.S. Geological Survey Bulletin 1909, p. A1–A8. Rogner, Hans-Holger (1997), "An assessment of world hydrocarbon resources."Annual review of energy and the environment 22.1 (1997): 217-262. Ruckenstein, E., A. S. Vaidyanathan, and G. R. Youngquist. (1971). "Sorption by solids with bidisperse pore structures." Chemical Engineering Science 26.9 : 1305-1318. Samuel, Mathew, et al. (1997). "Polymer-free fluid for hydraulic fracturing." SPE Annual Technical Conference and Exhibition. Samuel, M. M., et al. (1999). "Polymer-free fluid for fracturing applications." SPE Drilling & Completion 14.4: 240-246. 149 Samuel, Mathew, et al. (2000). "Viscoelastic surfactant fracturing fluids: applications in low permeability reservoirs." SPE Rocky Mountain Regional/Low-Permeability Reservoirs Symposium and Exhibition. Samuel, Mathew M., et al. (2003). "Compositions containing aqueous viscosifying surfactants and methods for applying such compositions in subterranean formations." U.S. Patent No. 6,637,517. 28 Oct. Schloemer, S., and B. M. Krooss. (2004) "Molecular transport of methane, ethane and nitrogen and the influence of diffusion on the chemical and isotopic composition of natural gas accumulations." Geofluids 4.1: 81-108. Seaton, N. A., Walton J.P.R.B and Quirke N. (1989) “ A new analysis method for the determination of the pore size distribution of porous carbons from nitrogen adsorption measurements”, Carbon, Vol 27 (6),pp.853-861 Shashkina, Julia A., et al. (2005). "Rheology of viscoelastic solutions of cationic surfactant. Effect of added associating polymer." Langmuir 21.4: 1524-1530. Shi, J. Q., and S. Durucan. (2003) "A bidisperse pore diffusion model for methane displacement desorption in coal by CO 2 injection." Fuel 82.10: 1219-1229. Shi, Ji-Quan, and Sevket Durucan. (2005) "Gas Storage and Flow in Coalbed Reservoirs: Implementation of a Bidisperse Pore Model for Gas Diffusion in Coal Matrix." SPE Reservoir Evaluation & Engineering 8.2 : 169-175. Tang, Guo-Qing, and Abbas Firoozabadi. (2002). "Relative permeability modification in gas/liquid systems through wettability alteration to intermediate gas wetting."SPE Reservoir Evaluation & Engineering 5.6: 427-436. U.S. Bureau of Labor Statistics (2007), Monthly Labor Review, November. http://www.bls.gov/opub/mlr/2007/11/contents.htm. Wattenbarger, Robert, et al. (1998). "Production analysis of linear flow into fractured tight gas wells." SPE Rocky Mountain Regional/Low-Permeability Reservoirs Symposium. 150 Working Document of the NPC Global Oil & Gas Study. Topic Paper #29. http://www.npc.org/study_topic_papers/29-ttg-unconventional-gas.pdf , Accessed: April 2010. WTRG Economics: http://www.wtrg.com/rotaryrigs.html, Accessed July 2013. Wylde, Jonathan, and Bill O'Neil. (2011). "Environmentally Acceptable Replacement of 2-Butoxyethanol with a High Performance Alternative for Fracturing Applications." SPE International Symposium on Oilfield Chemistry. Yang, Li, et al. (2006) "Adsorption of arsenic on layered double hydroxides: effect of the particle size." Industrial & engineering chemistry research 45.13: 4742-4751. Yuan et al. (2013). “Experimental study and modeling of methane adsorption and diffusion in shale”, FUEL. Zelenev, Andrei, et al. (2010). "Microemulsion-Assisted Fluid Recovery and Improved Permeability to Gas in Shale Formations." SPE International Symposium and Exhibiton on Formation Damage Control. Zhang, Jin, et al. (2010). "Development and Field Pilot of a Novel Viscoelastic Anionic- Surfactant (VAS) Fracturing Fluid." SPE Western Regional Meeting. 151 Appendix A-1 Other Available Cumulative Pore Volume Data for Mesopore and Macropore (a) Depth 7762.5ft (b) Depth 7772.5ft (c) Depth 7781.5ft (d) Depth 7792.5ft Cumulative Pore Volume from BJH Desorption (Depth 7762.5 – 7792.5ft) 152 (a) Depth 7802.5ft (b) Depth 7813.5ft (c) Depth 7823.5 ft (d) Depth 7832.5ft Cumulative Pore Volume from BJH Desorption (Depth 7802.5 – 7832.5ft) 153 (a) Depth 7841.5ft (b) Depth 7852.5ft (c) Depth 7862.5ft (d) Depth 7873.5ft Cumulative Pore Volume from BJH Desorption (Depth 7841.5ft - 7873.5ft) 154 (a) Depth 7885.5ft (b) Depth 7891.5ft Cumulative Pore Volume from BJH Desorption (Depth 7885.5ft - 7891.5ft) 155 Appendix A-2 Other Available PSD (in terms of dV/dD vs. pore diameter) for Mesopore and Macropore (a) Depth 7762.5ft (b) Depth 7772.5ft (c) Depth 7781.5ft (d) Depth 7792.5ft dV/dD vs Pore Diameter from BJH Desorption (Depth 7762.5 – 7792.5ft) 156 (a) Depth 7802.5ft (b) Depth 7813.5ft (a) Depth 7823.5ft (b) Depth 7832.5ft dV/dD vs Pore Diameter from BJH Desorption (Depth 7802.5 – 7832.5ft) 157 (a) Depth 7841.5ft (b) Depth 7852.5ft dV/dD vs Pore Diameter from BJH Desorption (Depth 7841.5 – 7852.5ft) 158 Appendix A-3 Other Available Cumulative Pore Volume Data for Micropore (a) Depth 7832.5ft (b) Depth 7841.5ft (c) Depth 7852.5ft Cumulative Pore Volume from HK Method (Depth 7832.5ft – 7852.5ft) 159 Appendix B Other Available Vertical and Horizontal SEM Images Vertical (left) and Horizontal (right) SEM Images of 7721.5 ft (2000X) Vertical (left) and Horizontal (right) SEM Images of 7732.5 ft (2000X) 160 Vertical (left) and Horizontal (right) SEM Images of 7742.5 ft (2000X) Vertical (left) and Horizontal (right) SEM Images of 7750.5 ft (2000X) 161 Vertical (left) and Horizontal (right) SEM Images of 7781.5 ft (2000X) Vertical (left) and Horizontal (right) SEM Images of 7792.5 ft (2000X) 162 Vertical (left) and Horizontal (right) SEM Images of 7802.5 ft (2000X) 163 Appendix C Other Available EDX Data EDX for Vertical View Sample with Reported Matrix Permeability at 7750.5 ft (2000X): k = 2.82E-05mD EDX for Vertical View Sample with Reported Matrix Permeability at 7772.5 ft (2000X): k = 3.81E-05mD 164 EDX for Vertical View Sample with Reported Matrix Permeability at 7792.5 ft (2000X): k = 3.93E-08mD EDX for Vertical View Sample with Reported Matrix Permeability at 7802.5 ft (2000X): k = 1.55E-06mD 165 EDX for Vertical View Sample with Reported Matrix Permeability at 7813.5 ft (2000X): k = 3.59E-06mD EDX for Vertical View Sample with Reported Matrix Permeability at 7832.5 ft (2000X): k = 1.36E-07mD 166 EDX for Vertical View Sample with Reported Matrix Permeability at 7841.5 ft (2000X): k = 1.43E-05mD EDX for Vertical View Sample with Reported Matrix Permeability at 7852.5 ft (2000X): k = 1.47E-05mD 167 EDX for Vertical View Sample with Reported Matrix Permeability at 7873.5 ft (2000X): k = 9.25E-06mD 168 Appendix D Orthogonal Collocation Method The following orthogonal collocation method to solve BPM follows the very detailed work by [Do 1998]. The dimensionless set of equations can be written as: Dimensionless equations: 2 2 2 α α α τ ∂ ∂ + ∂ ∂ = ∂ ∂ x x x , (C-1)       ∂ ∂ ∂ ∂ +       ∂ ∂ ∂ ∂ +       ∂ ∂ ∂ ∂ +       ∂ ∂ ∂ ∂ =       ∂ ∂ + ∂ ∂ = β β β β σ β β β β σ β β β β σ β β β β γ α λ τ α y y H y G y y y H y G y y y H y G y x y ) ( ) ( 1 ) ( ) ( ) ( 1 ) ( ) ( ) ( 1 1 3 3 2 2 2 2 2 1 2 2 2 2 1 , (C-2) ) ( 1 y H , ) ( 2 y H and ) ( 3 y H are functions of y base on the fact that y is a function of pressure. And σ is a function of y base on the fact that viscosity is a function of pressure. First, we make the following transformation (symmetry at the centers of the micro- particle and pellet): 2 α = v , (C-3) 2 β = u , (C-4) 169 As a result, equations (C-1), (C-2) will become:       ∂ ∂ + ∂ ∂ = ∂ ∂ v x v x v x 6 4 2 2 τ , (C-5)       ∂ ∂ ∂ ∂ +       ∂ ∂ + ∂ ∂ +       ∂ ∂ ∂ ∂ +       ∂ ∂ + ∂ ∂ +       ∂ ∂ ∂ ∂ +       ∂ ∂ + ∂ ∂ +       ∂ ∂ + ∂ ∂ =       ∂ ∂ + ∂ ∂ u y u y H y G u u y u y u u y u y H y G u u y u y u u y u y H y G u u y u y u u y u y u v x y ) ( ) ( 4 6 4 ) ( ) ( 4 6 4 ) ( ) ( 4 6 4 6 4 6 3 2 2 2 2 2 1 2 2 2 2 1 σ σ σ γ λ τ , (C-6) The boundary conditions for the above equations are: 1 = v , ) ( 1 y F x = , (C-7) 1 = u , 0 C C y y b b = = , (C- 8) We choose M interior collocation points in the microparticle, and evaluate equation (C-5) at the interior collocation point k to get: l M l kl k x C x ∑ + = = ∂ ∂ 1 1 μ τ , (C-9) For , , , 2 , 1 M k K = where μ μ μ kl kl kl A vB C 6 4 + = , (C- 10) The concentration at the exterior surface of the microparticle is in equilibrium with the concentration in the mesopore, that is: 170 ) ( 1 y F x M = + (C- 11) Next we evaluate equation (C-6) at the interior collocation point i along the pellet co- ordinate and get: l i M l kl k i x C x , 1 1 , ∑ + = = ∂ ∂ μ τ , (C-12) For M k , , 2 , 1 K = and 1 , , 2 , 1 + = N i K Here k i x , is the adsorbed concentration at the k point along the microparticle co-ordinate and the point i along the pellet co-ordinate. Note that this equation is valid up to the point N+1 in the pellet co-ordinate. The adsorbed concentration at the exterior surface of the microparticle at the collocation i, 1 , + M i x can be calculated from (C-11) and takes the following form: ) ( 1 , i M i y F x = + , (C-13) Next, we evaluate the mass balance equation along the pellet co-ordinate (C-6) at the i-th collocation point:                           + +                           + +                           + +       = + ∂ ∂ ∑ ∑ ∑ ∑ ∑ ∑ ∑ ∑ ∑ ∑ ∑ + = + = + = + = + = + = + = + = + = + = + = + 1 1 , 1 1 3 , 1 1 , 3 1 1 , 1 1 2 , 1 1 , 2 1 1 , 1 1 1 , 1 1 , 1 1 1 , 1 1 1 , 1 , 1 ) ( ) ( 4 ) ( ) ( ) ( ) ( ) ( 4 ) ( ) ( ) ( ) ( ) ( 4 ) ( ) ( ) ( 6 N j j j i N j j i j i i N j j j i i i i N j j j i N j j i j i i N j j j i i i i N j j j i N j j i j i i N j j j i i i i N j j j i M l i M i y A y H y G A u y C y H y G y y A y H y G A u y C y H y G y y A y H y G A u y C y H y G y y C x A y σ σ σ γ λ τ μ , (C-14) For , , , 2 , 1 N i K = where j i j i i j i A B u C , , , 6 4 + = , (C-15) 171 Equation (C-14) is only valid for N interior collocation points. The equation for the (N+1)-th interpolation point is the concentration at the exterior surface of the pellet, which takes the following form: 0 1 C C y y b b N = = + , (C-16) Equations (C-12) and (C-14) form a set of (N+1)(M+1) equations in terms of (N+1)(M+1) unknowns to be solved, where ) , , 2 , 1 ; 1 , , 2 , 1 ( , M k N i x k i K K = + = and ) , , 2 , 1 ( N i y i K =

Abstract (if available)

Abstract Gas shales contain an abundance of natural gas. Low-permeability ("tight") shale is commonly found domestically in the Appalachian Basin but it has yet to be fully explored. A major limitation to the efficient extraction of this "tight gas" is the current lack of knowledge of the physics and mass transfer properties involved in production from such low-permeability porous materials. Without an improved fundamental understanding and appropriate modeling and simulation, forecasting of natural gas production from these formations will be inaccurate and unreliable, and operators must continue relying on "trial and error" in their development strategies. With the demand for natural gas continuing to surge, an urgent need, therefore, exists today to improve the efficient drilling and completion of new wells as well as for re-completion of existing wells in order to meet this increased demand. As a result, we not only need to drill and complete more wells, we also need to improve production efficiency on a per well basis. ❧ In this study we have characterized the properties of gas shales and the fundamental mass transfer mechanisms associated with the gas production process

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Investigation of gas transport and sorption in shales

PDF

Investigation of gas sorption and mass transfer in gas shales

PDF

Spontaneous countercurrent and forced imbibition in gas shales

PDF

Modeling and simulation of multicomponent mass transfer in tight dual-porosity systems (unconventional)

PDF

Mass transfer during enhanced hydrocarbon recovery by gas injection processes

PDF

The study of CO₂ mass transfer in brine and in brine-saturated Mt. Simon sandstone and the CO₂/brine induced evolution of its transport and mechanical properties

PDF

A study of diffusive mass transfer in tight dual-porosity systems (unconventional)

$Coarse-scale simulation of heterogeneous reservoirs and multi-fractured horizontal wells$

PDF

Coarse-scale simulation of heterogeneous reservoirs and multi-fractured horizontal wells

PDF

Modeling and simulation of complex recovery processes

PDF

Investigation of adsorption and phase transition phenomena in porous media

PDF

Impact of exposure to brine/CO₂ on the mechanical and transport properties of the Mt. Simon sandstone

PDF

Process intensification in hydrogen production via membrane-based reactive separations

PDF

The use of carbon molecule sieve and Pd membranes for conventional and reactive applications

PDF

A high efficiency, ultra-compact process for pre-combustion CO₂ capture

PDF

Performance prediction, state estimation and production optimization of a landfill

$Isotopic insights to human-impacted hydrologic systems: from urban watersheds to hydraulically fractured basins$

PDF

Isotopic insights to human-impacted hydrologic systems: from urban watersheds to hydraulically fractured basins

PDF

A study of dispersive mixing and flow based lumping/delumping in gas injection processes

PDF

Methanol synthesis in a membrane reactor

PDF

A flow-through membrane reactor for destruction of a chemical warfare simulant

PDF

Methanol synthesis in the membrane reactor