Shale Analytics: Data-driven Analytics in Unconventional Resources

288 Pages·2017·5.12 MB·English

by  Shahab D. Mohaghegh

Shahab D. Mohaghegh
Shale Analytics
Data-Driven Analytics in Unconventional
Resources
123
Shahab D.Mohaghegh
Petroleum andNatural GasEngineering
West Virginia University
Morgantown,WV
USA
and
Intelligent Solutions, Inc.
Morgantown,WV
USA
ISBN978-3-319-48751-9 ISBN978-3-319-48753-3 (eBook)
DOI 10.1007/978-3-319-48753-3
LibraryofCongressControlNumber:2016955428
©SpringerInternationalPublishingAG2017
Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpart
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
orinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilar
methodologynowknownorhereafterdeveloped.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfrom
therelevantprotectivelawsandregulationsandthereforefreeforgeneraluse.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authorsortheeditorsgiveawarranty,expressorimplied,withrespecttothematerialcontainedhereinor
foranyerrorsoromissionsthatmayhavebeenmade.
Printedonacid-freepaper
ThisSpringerimprintispublishedbySpringerNature
TheregisteredcompanyisSpringerInternationalPublishingAG
Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland
Foreword
It is an honor and a pleasure to be asked to write the Foreword to this much
anticipated book on the soft-computing, data-driven methodologies applied across
unconventional reservoirs so as to harness the power of raw data and generate
actionableknowledge.Wearetakenalongawell-documentedbutstillbumpyroad
that starts with an introduction to the shale revolution and draws salient compar-
isons between the traditional modeling of these unconventional resources and the
non-deterministic and stochastic workflows prevalent in all industries that strive to
analyze vast quantities of raw data to address and solve business problems.
We are enlightened as to an array of analytical methodologies that have suc-
cessfully proven to be not only pertinent in the oil and gas industry but also
computer resource friendly. Methodologies drawn from artificial intelligence and
data mining schools of thought, such as artificial neural networks, fuzzy logic,
fuzzy cluster analysisand evolutionary computing, the last ofwhich isinspired by
the Darwinian Theory of Evolution through Natural Selection.
The book inspires geoscientists entrenched in first principles and engineering
concepts to think outside the box shaped by determinism, and to marry their
experienceandinterpretations ofthedatawiththeresults generatedbydata-driven
advanced analytical workflows. The latter approach enables ALL the data to be
exploitedandopensthedoortohiddenrelationshipsandtrendsoftenmissedbythe
formerapproachexecutedinisolation.Weareencouragedasengineerstonotonly
to break out of the silos and put ALL our data into the context of experience and
interpretation but also to allow the data to do the talking and ask questions of our
traditional workflows.
The convincing conversational tone is enhanced by some in-depth case studies
that have proven successful to oil and gas operators from a business value
proposition perspective. The Shale Production Optimization Technology (SPOT)
chapteraloneisvalidationenoughtoprovidecredibleendorsementofthestrategic
and tactical business insight into hydraulic fracturing practices essential to exploit
the unconventional resources. The book provides plausible and cogent arguments
for a top-down modeling methodology. We are introduced to a very convincing
vii
viii Foreword
presentation of material that details a “formalized, comprehensive, multivariate,
full-field, and empirical reservoir model”.
Wearefortunatetohavethisbookpublishedatsuchanopportunetimeintheoil
and gas industry as it climbs slowly out of a big trough left by the erratic price
fluctuationsofthepasttwoyearssince2014.Operatorsandservicecompaniesalike
will see incredible value within these pages. The authors have diligently provided
an index to very important soft-computing workflows to ensure ALL the hard and
soft data detailed in the early chapters are mined to surface powerful knowledge.
This original knowledge can help design best practices for drilling and completing
the wells in unconventional reservoirs. It also sheds light on different modeling
practices for field re-engineering and well forecasting in reservoirs that necessitate
innovative workflows as traditional interpretive approaches honed in the conven-
tional reservoirs prove inadequate.
Thestrengthofthemessageslogicallylaidoutinthisbookintroducethereader
to a collection of soft-computing techniques that address some critical business
issues operators in the unconventional reservoirs are facing on a daily basis:
quantifying uncertainty in the shale, forecasting production, estimating ultimate
recovery, building robust models and formulating best practices to exploit the
hydrocarbons.
We are indebted to Shahab Mohaghegh for his original thought, passion and
innovation across the many years as he evangelizes the application of the
ever-increasing popularity of data-driven methodologies. As one of the earliest
pioneersinthisfield,wearegratefulthathehasputpentopaperandprovidedthe
industry with a very valuable book.
Cary, USA Keith R. Holdaway FGS
Advisory Industry Consultant
SAS Global O&G Domain
Contents
1 Introduction... .... .... ..... .... .... .... .... .... ..... .... 1
1.1 The Shale Revolution.... .... .... .... .... .... ..... .... 2
1.2 Traditional Modeling .... .... .... .... .... .... ..... .... 4
1.3 A Paradigm Shift .. ..... .... .... .... .... .... ..... .... 4
2 Modeling Production from Shale... .... .... .... .... ..... .... 7
2.1 Reservoir Modeling of Shale .. .... .... .... .... ..... .... 9
2.2 System of Natural Fracture Networks.... .... .... ..... .... 10
2.3 System of Natural Fracture Networks in Shale. .... ..... .... 13
2.4 A New Hypothesis on Natural Fractures in Shale... ..... .... 14
2.5 Consequences of Shale SNFN . .... .... .... .... ..... .... 16
2.6 “Hard Data” Versus “Soft Data”.... .... .... .... ..... .... 18
2.7 Current State of Reservoir Simulation and Modeling
of Shale . .... .... ..... .... .... .... .... .... ..... .... 19
2.7.1 Decline Curve Analysis.... .... .... .... ..... .... 20
2.7.2 Rate Transient Analysis.... .... .... .... ..... .... 21
2.8 Explicit Hydraulic Fracture Modeling.... .... .... ..... .... 22
2.9 Stimulated Reservoir Volume.. .... .... .... .... ..... .... 24
2.10 Microseismic . .... ..... .... .... .... .... .... ..... .... 27
3 Shale Analytics .... .... ..... .... .... .... .... .... ..... .... 29
3.1 Artificial Intelligence .... .... .... .... .... .... ..... .... 33
3.2 Data Mining.. .... ..... .... .... .... .... .... ..... .... 33
3.2.1 Steps Involved in Data Mining .. .... .... ..... .... 34
3.3 Artificial Neural Networks .... .... .... .... .... ..... .... 35
3.3.1 Structure of a Neural Network... .... .... ..... .... 36
3.3.2 Mechanics of Neural Networks Operation.. ..... .... 38
3.3.3 Practical Considerations During the Training
of a Neural Network.. .... .... .... .... ..... .... 41
xi
xii Contents
3.4 Fuzzy Logic.. .... ..... .... .... .... .... .... ..... .... 55
3.4.1 Fuzzy Set Theory .... .... .... .... .... ..... .... 57
3.4.2 Approximate Reasoning.... .... .... .... ..... .... 59
3.4.3 Fuzzy Inference.. .... .... .... .... .... ..... .... 60
3.5 Evolutionary Optimization .... .... .... .... .... ..... .... 62
3.5.1 Genetic Algorithms... .... .... .... .... ..... .... 63
3.5.2 Mechanism of a Genetic Algorithm... .... ..... .... 64
3.6 Cluster Analysis... ..... .... .... .... .... .... ..... .... 66
3.7 Fuzzy Cluster Analysis... .... .... .... .... .... ..... .... 68
3.8 Supervised Fuzzy Cluster Analysis.. .... .... .... ..... .... 70
3.8.1 Well Quality Analysis (WQA)... .... .... ..... .... 71
3.8.2 Fuzzy Pattern Recognition.. .... .... .... ..... .... 74
4 Practical Considerations. ..... .... .... .... .... .... ..... .... 83
4.1 Role of Physics and Geology.. .... .... .... .... ..... .... 84
4.2 Correlation is not the Same as Causation. .... .... ..... .... 84
4.3 Quality Control and Quality Assurance of the Data . ..... .... 86
5 Which Parameters Control Production from Shale .... ..... .... 91
5.1 Conventional Wisdom ... .... .... .... .... .... ..... .... 92
5.2 Shale Formation Quality.. .... .... .... .... .... ..... .... 93
5.3 Granularity... .... ..... .... .... .... .... .... ..... .... 98
5.4 Impact of Completion and Formation Parameters... ..... .... 98
5.4.1 Results of Pattern Recognition Analysis ... ..... .... 99
5.4.2 Influence of Completion Parameters .. .... ..... .... 102
5.4.3 Important Notes on the Results and Discussion... .... 106
5.5 Chapter Conclusion and Closing Remarks .... .... ..... .... 106
6 Synthetic Geomechanical Logs. .... .... .... .... .... ..... .... 109
6.1 Geomechanical Properties of Rocks . .... .... .... ..... .... 109
6.1.1 Minimum Horizontal Stress. .... .... .... ..... .... 110
6.1.2 Shear Modulus .. .... .... .... .... .... ..... .... 110
6.1.3 Bulk Modulus... .... .... .... .... .... ..... .... 110
6.1.4 Young’s Modulus .... .... .... .... .... ..... .... 111
6.1.5 Poisson’s Ratio .. .... .... .... .... .... ..... .... 112
6.2 Geomechanical Well Logs .... .... .... .... .... ..... .... 112
6.3 Synthetic Model Development . .... .... .... .... ..... .... 113
6.3.1 Synthetic Log Development Strategy.. .... ..... .... 115
6.3.2 Results of the Synthetic Logs ... .... .... ..... .... 116
6.4 Post-Modeling Analysis .. .... .... .... .... .... ..... .... 124
7 Extending the Utility of Decline Curve Analysis... .... ..... .... 127
7.1 Decline Curve Analysis and Its Use in Shale.. .... ..... .... 127
7.1.1 Power Law Exponential Decline. .... .... ..... .... 129
7.1.2 Stretched Exponential Decline... .... .... ..... .... 130
Contents xiii
7.1.3 Doung’s Decline . .... .... .... .... .... ..... .... 130
7.1.4 Tail-End Exponential Decline (TED).. .... ..... .... 132
7.2 Comparing Different DCA Techniques... .... .... ..... .... 134
7.2.1 Is One DCA Technique Better Than the Other? .. .... 136
7.3 Extending the Utility of Decline Curve Analysis in Shale . .... 140
7.3.1 Impact of Different Parameters on DCA Technique ... 140
7.3.2 Conventional Statistical Analysis Versus Shale
Analytics.. ..... .... .... .... .... .... ..... .... 142
7.3.3 More Results of Shale Analytics. .... .... ..... .... 144
7.4 Shale Analytics and Decline Curve Analysis .. .... ..... .... 151
8 Shale Production Optimization Technology (SPOT).... ..... .... 153
8.1 Dataset .. .... .... ..... .... .... .... .... .... ..... .... 153
8.1.1 Production Data.. .... .... .... .... .... ..... .... 153
8.1.2 Hydraulic Fracturing Data.. .... .... .... ..... .... 154
8.1.3 Reservoir Characteristics Data... .... .... ..... .... 154
8.2 Complexity of Well/Frac Behavior.. .... .... .... ..... .... 155
8.3 Well Quality Analysis (WQA) . .... .... .... .... ..... .... 164
8.4 Fuzzy Pattern Recognition .... .... .... .... .... ..... .... 175
8.5 Key Performance Indicators (KPIs).. .... .... .... ..... .... 183
8.6 Predictive Modeling ..... .... .... .... .... .... ..... .... 197
8.6.1 Training, Calibration, and Validation of the Model.... 197
8.7 Sensitivity Analysis ..... .... .... .... .... .... ..... .... 201
8.7.1 Single-Parameter Sensitivity Analysis . .... ..... .... 202
8.7.2 Combinatorial Sensitivity Analysis ... .... ..... .... 208
8.8 Generating Type Curves.. .... .... .... .... .... ..... .... 211
8.9 Look-Back Analysis..... .... .... .... .... .... ..... .... 220
8.10 Evaluating Service Companies’ Performance .. .... ..... .... 224
9 Shale Numerical Simulation and Smart Proxy .... .... ..... .... 229
9.1 Numerical Simulation of Production from Shale Wells.... .... 229
9.1.1 Discrete Natural Fracture Modeling... .... ..... .... 230
9.1.2 Modeling the Induced Fractures . .... .... ..... .... 231
9.2 Case Study: Marcellus Shale .. .... .... .... .... ..... .... 233
9.2.1 Geological (Static) Model .. .... .... .... ..... .... 233
9.2.2 Dynamic Model.. .... .... .... .... .... ..... .... 234
9.2.3 History Matching. .... .... .... .... .... ..... .... 235
9.3 Smart Proxy Modeling ... .... .... .... .... .... ..... .... 237
9.3.1 A Short Introduction to Smart Proxy.. .... ..... .... 237
9.3.2 Cluster Level Proxy Modeling... .... .... ..... .... 238
9.3.3 Model Development (Training and Calibration) .. .... 240
9.3.4 Model Validation (Blind Runs).. .... .... ..... .... 247
10 Shale Full Field Reservoir Modeling .... .... .... .... ..... .... 251
10.1 Introduction to Data-Driven Reservoir Modeling
(Top-Down Modeling) ... .... .... .... .... .... ..... .... 253
xiv Contents
10.2 Data from Marcellus Shale.... .... .... .... .... ..... .... 255
10.2.1 Well Construction Data.... .... .... .... ..... .... 255
10.2.2 Reservoir Characteristics Data... .... .... ..... .... 256
10.2.3 Completion and Stimulation Data .... .... ..... .... 258
10.2.4 Production Data.. .... .... .... .... .... ..... .... 258
10.3 Pre-modeling Data Mining.... .... .... .... .... ..... .... 259
10.4 TDM Model Development .... .... .... .... .... ..... .... 260
10.4.1 Training and Calibration (History Matching)..... .... 260
10.4.2 Model Validation. .... .... .... .... .... ..... .... 262
11 Restimulation (Re-frac) of Shale Wells .. .... .... .... ..... .... 267
11.1 Re-frac Candidate Selection ... .... .... .... .... ..... .... 268
11.2 Re-frac Design .... ..... .... .... .... .... .... ..... .... 272
Bibliography .. .... .... .... ..... .... .... .... .... .... ..... .... 279
Chapter 1
Introduction
WithoutData,YouAreJustanotherPersonwithanOpinion.
W.E.Deming(1900–1993).
When W.E. Deming uttered these words in 1980s, “world new nothing about the
impact that shale will have in changing the energy landscape of the twenty-first
century.” Nevertheless, no other quotation can so vividly describe the state of our
scientificandengineeringknowledgeofthephysicsofcompletion,stimulation,and
interactionbetweeninducedandnaturalfracturesduringtheoilandgasproduction
from shale. Many wellintentionedengineers andscientists that areinvolvedinthe
day-to-dayoperationsofshalewellsdeveloptheintuitionrequiredtomakethebest
ofwhatisavailabletotheminordertoincreasetheefficiencyandtherecoveryfrom
theshalewells.However,afullunderstandingofthephysicsandthemechanicsof
the storage and transport phenomena and the production operation in shale has
remained elusive to a large extent.
There are many opinions and speculations on what exactly happens as we
embarkupondrilling,completing,andhydraulicallyfracturingshalewells.Muchof
theopinionsandspeculationsarehardlyeversupportedbyfactsanddata,andinthe
occasionsthat theyare,ithardlyever transcendstheanecdotalevidencesthathave
been heard of, or been seen. However, the fact remains that as an industry, we
collect a significant amount of data (field measurements—facts) during the opera-
tionsthatresultinoilandgasproductionfromshale.Itishardtoimaginethatthese
collected data do not contain the knowledge we need in order to optimize pro-
duction and maximize recovery from this prolific hydrocarbon resource.
It took our industry professionals several years to come to the inevitable con-
clusion that our conventional modeling techniques that were developed for car-
bonateandcoalbedmethaneformationscannotsubstituteourlackofunderstanding
of the physics and the mechanics of completion and production from shale wells.
Butnowthatthisfacthasbecomeself-evident,maybethespeculativeresistanceto
the required paradigm shift to move to a data-driven solution can finally be
overcome.
This book is dedicated to scratching the surface of all that is possible in the
application of Petroleum Data Analytics to reservoir management and production
©SpringerInternationalPublishingAG2017 1
S.D.Mohaghegh,ShaleAnalytics,DOI10.1007/978-3-319-48753-3_1
2 1 Introduction
operation of shale, what we have chosen to call “Shale Analytics.” Here we
demonstrate how the existing data collected from the development of shale assets
can help in developing a better understanding of the nuances associated with
operating shale wells. How to learn from our experiences in order to positively
impact future operations. How to create a system of continuous learning from the
data that is generated on a regular basis. In other words, using this technology we
canmakesurethateverybarrelofoilandeveryMCFofgasthatisproducedfrom
our shale wells not only brings return to our investment, but also enriches our
understanding of how this resource needs to be treated for maximum return.
1.1 The Shale Revolution
More than 1.8 trillion barrels of oil are trapped in shale in Federal lands in the
western United States in the states of Colorado, Utah, and Wyoming, of which
800 billion is considered recoverable—three times the proven reserves of Saudi
Arabia. The INTEK assessment for EIA found 23.9 billion barrels of technically
recoverable shale oil resources in the onshore Lower 48 States. The Southern
CaliforniaMonterey/Santosplayisthelargestshaleoilformationestimatedtohold
15.4 billionbarrelsor64 %ofthetotalshaleoilresourcesfollowedbyBakkenand
Eagle Ford with approximately 3.6 billion barrels and 3.4 billion barrels of oil,
respectively [1].
Unlike the conventional oil and gas resources that are concentrated in certain
parts of the world, shale resources (shale gas and shale oil) are widely distributed
throughout the world. As such, it has the potential to democratize the energy
production throughout the world. Figure 1.1 shows a map of the world along with
the countries that have been studied in order to see if they have shale resources.
Among the countries that have been studied (shown in white background color in
Fig. 1.1), every one of them has been blessed with this organic rich resource play.
Shale oil production has been estimated to extend to about 12 % of the global oil
supply by 2035.1 This may result in overall lower energy prices and consequently
higher economic growth rate all over the world. Figure 1.2 shows the impact of
shale on the United States’ proven oil and gas reserves.
Until recently only 19 % of U.S. power generation was based on natural gas
whilemorethan50 %wasbasedoncoal.Naturalgasreleaseslessthan50 %ofthe
greenhouse gases than coal and just by switching from coal to gas considerable
environmental impact can be expected with no impact on day-to-day life styles.
Being able to reduce greenhouse gas emission by 50 % with no major impact on
average people’s life style or economic well-being would have been more like an
environmentalist dream than reality only a decade ago.
1http://www.pwc.com/gx/en/industries/energy-utilities-mining/oil-gas-energy/publications/shale-
oil-changes-energy-markets.html.