&EPA
United States
Environmental Protection
Agency
EPA 600/R-l 1/046 I May 2011 I www.epa.gov
Proceedings of the Technical Workshops
for the Hydraulic Fracturing
Study: Well Construction and Operation
U.S. Environmental Protection Agency
-------
EPA 600/R-11/046
May 2011
United States
Environmental Protection
Agency
Proceedings of the Technical Workshops
for the Hydraulic Fracturing Study:
Well Construction & Operations
Office of Research and Development
US Environmental Protection Agency
Washington, DC
-------
Table of Contents
List of Figures ii
List of Tables iii
Introduction 1
Workshop Participants 3
Agenda 5
Summary and Abstracts from Theme 1: Well Construction 9
Summary of Presentations for Theme 1: Well Construction 10
Summary of Discussions Following Theme 1: Well Construction Presentations 12
Abstracts for Theme 1: Well Construction 15
Overview of the Well Construction Sessions 16
Public Water Sources and Hydraulic Fracturing-A State Drinking Water Perspective 20
Well Completion Methods for Aquifer Protection 21
Well Planning and Construction Techniques 25
Production Casing Design Considerations 29
Shell's Well Construction Practices in the Marcellus Shale 33
Multi-well Pad, Tight Gas, Directional Drilling Program Protects Aquifers 36
Casing Perforating Overview 40
Summary and Abstracts from Theme 2: Fracture Design and Stimulation 42
Summary of Presentations for Theme 2: Fracture Design and Stimulation 43
Summary of Discussions Following Theme 2: Fracture Design and Stimulation Presentations .. 45
Abstracts for Theme 2: Fracture Design and Stimulation 49
The Distribution of Natural Fractures above a Gas Shale: Questions about Whether Deep Fracture Fluid Leaks
into Groundwater Outside the Realm of Faulty Borehole Construction 50
Evaluation of Well Records and Geophysical Logs for Determining the Presence of Freshwater, Saltwater, and
Gas above the Marcellus Shale, South-Central New York 56
Fracture Design in Horizontal Shale Wells- Data Gathering to Implementation 61
Hydraulic Fracturing in Coal Bed Methane Development, Raton Basin, Southern Colorado, USA 69
Fracture Design and Stimulation -Monitoring 71
A Case History of Tracking Water Movement Through Fracture Systems in the Barnett Shale 79
Measurements and Observations of Fracture Height Growth 81
Sustainable Fracturing Rationale to Reach Well Objectives-The Impact of Uncertainties and Complexities on
Compliance Assurances 87
Design and Rationale for a Field Experiment using Tracers in Hydraulic Fracture Fluid 88
Review of Stimulation Fluid Retention Mechanisms and Likelihood of Fluid Communication with Shallow
Potential Aquifers in Shale Gas Development 90
Summary and Abstracts from Theme 3: Well Integrity 96
Summary of Presentations from Theme 3: Well Integrity 97
Summary of Discussions Following Theme 3: Well Integrity Presentations 98
Abstracts for Theme 3: Well Integrity 101
Assessment Methods for Well Integrity during the Hydraulic Fracturing Cycle 102
Pre and Post Well Integrity Methods for Hydraulically Fractured/Stimulated Wells 107
Case Study for Well Integrity over a Full Life Cycle 122
Risks to Drinking Water from Oil and Gas Wellbore Construction and Integrity: Case Studies and Lessons
Learned 136
Revisiting the Major Discussion Points of the Technical Presentation Sessions 148
Glossary of Terms 150
-------
List of Figures
Figure 1. Location of study area in south-central New York, water wells that penetrated
saltwater and (or) gas, and gas wells that penetrated freshwater and (or) gas above the
Marcellus Shale 59
Figure 2. Geophysical logs and reported freshwater zone and gas show for the 760-860 ft depth
interval in gas well 19484, Cortland County, New York 60
Figure 3 62
Figure 4 65
Figure 5 65
Figure 6. Simplified Location Schematic 76
Figure 7. Inside the treatment monitoring van 76
Figure 8. Treatment Chart - Pressure, Rate and Prop Concentration 77
Figure 9. Additive Chart 77
Figure 10. Fluid Tracking, Numeric Value, and Stage Summary Screen 78
Figure 11. Blender schedule 78
Figure 12. Mineback photograph of complex fracture 81
Figure 13. Mineback photograph of offsets & splitting 82
Figure 14. Measured stress profile in Mesaverde 82
Figure 15. Overview of DOE/GRI M-Site hydraulic fracture diagnostics field test site 84
Figure 16. Example Haynesville shale microseismic data 84
Figure 17. A compendium of microseismic fracture diagnostic results in the Barnett shale
relative to known aquifers 85
Figure 18. Ultrasonic Imager (a) tool design and (b) transducer position (Smolen, 1996) 110
Figure 19. Illustrative Example of USIT Log Run on Injection Well Ill
Figure 20. Example SBT Log 112
Figure 21. MIT Tool 113
Figure 22. Magnetic Thickness Tool (MTT) 115
Figure 23. Example of Multi-Finger Caliper Survey 115
Figure 24. WFL Measurement Technique 115
Figure 25. Example of WFL Log 116
Figure 26. Example Borax-PNL Log 117
Figure 27. Ultrasonic Leak Detection Tool 117
Figure 28. Example of a Horizontal and Vertical Well (API, 2009) 119
-------
List of Tables
Table 1 27
Table 2. Parameters analyzed during baseline water survey 35
Table 3. Acoustic Properties of Materials (Smolen, 1996) 112
-------
Introduction
The Hydraulic Fracturing Study
In its Fiscal Year 2010 budget report, the U.S. House of Representatives Appropriation
Conference Committee identified the need for a study of the potential impacts of hydraulic
fracturing (HF) on drinking water resources. The Committee directed EPA scientists to
undertake a study of HF to better understand any potential impacts of hydraulic fracturing on
drinking water and ground water. The EPA produced a draft study plan, which focuses on the
key stages of the HF water lifecycle: water acquisition, chemical mixing, well injection, flowback
and produced water, and wastewater treatment and waste disposal. This plan was submitted to
the Science Advisory Board (SAB) in February 2011 and the peer review of the plan was held on
March 7 and 8, 2011. At the time these technical workshop proceedings were developed, the
SAB had not given its official review to EPA.
EPA has included stakeholder concerns in the planning process of the study from its inception,
engaging stakeholders in a dialogue about the study through a series of webinars and facilitated
public meetings held between May and September 2010. EPA also held four technical
workshops in February and March 2011 to explore the following focus areas: Chemical &
Analytical Methods (February 24-25), Well Construction & Operations (March 10-11), Fate &
Transport (March 28-29) and Water Resource Management (March 29-30).
The technical workshops centered around three goals: (1) inform EPA of the current technology
and practices being used in hydraulic fracturing, (2) identify research related to the potential
impacts of hydraulic fracturing on drinking water resources, and (3) provide an opportunity for
EPA scientists to interact with technical experts. EPA invited technical experts from the oil and
natural gas industry, consulting firms, laboratories, state and federal agencies, and
environmental organizations to participate in the workshops. EPA will use the information
presented in this document to inform research that effectively evaluates the relationship
between HF and drinking water.
An initial report of results from the EPA's study is expected by late 2012, with an additional
report expected in 2014.
About the Proceedings
These proceedings provide an overview of the twenty-four presentations given on well
construction and operations at the Technical Workshop for the U.S. EPA Hydraulic Fracturing
Study held on March 10-11, 2011. This workshop consisted of three sessions or themes: Theme
1-Well Construction; Theme 2-Fracture Design and Stimulation; and Theme 3-Well Integrity.
The proceedings include abstracts of the presentations and a summary of the discussions that
took place during the workshop. The presentations from this workshop are not part of the
proceedings document, but may be accessed at http://epa.gov/hydraulicfracturing.
-------
This is the second of four technical workshops on topics relating to the EPA Hydraulic Fracturing
Study. The other three workshops are: Chemical and Analytical Methods (Feb. 24-25), Fate and
Transport (Mar. 28-29), and Water Resources Management (Mar. 29-30). Proceedings will be
available separately for the other three workshops.
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA. Any mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
Editorial Team for the Proceedings
The attendees at the Well Construction and Operations workshop were selected based on
information submitted to EPA during the attendee nomination process. Presenters, a workshop
lead, and theme leads were selected from the pool of attendees, once again, based on the
information submitted to EPA during the attendee nomination process. The workshop lead,
Scott Anderson of the Environmental Defense Fund, assisted EPA in finalizing details for the
workshop and served as the lead editor of the proceedings document. The theme leadsBob
Whiteside of Texas World Operations for Theme 1, Tim Beard of Chesapeake Energy
Corporation for Theme 2, and Jim Bolander of Southwestern Energy for Theme 3served as
editors for their respective themes.
-------
Workshop Participants
Ahmed
Steve
Scott
Timothy
Chrystal
John
James
Jeanne
Grant
Susan
Alan
Scott
David
Brian
Jill
Carolyn
Michael
Terry
Jay
Bill
Richard
Gang
Patrick
Brad
John
Fred
Lloyd
Charles
Anthony
Mark
Carey
Jeff
Bruce
Barbara
Name
Abou-Sayed
Acree
Anderson
Beard
Beasley
Bolakas
Bolander
Briskin
Bromhal
Burden
Byrnes
Cline
Cramer
D'Amico
Dean
Debrick
Eberhard
Engelder
Foreman
Godsey
Hammack
Han
Handren
Hansen
Harju
Hauchman
Hetrick
Hillenbrand
Ingraffea
Jensen
Johnston
Jollie
Kobelski
Kutchko
Affiliation
Advantek International
US Environmental Protection Agency
Environmental Defense Fund
Chesapeake Energy Corporation
US Environmental Protection Agency
Stantec Consulting Corporation
Southwestern Energy
US Environmental Protection Agency
US Department of Energy
US Environmental Protection Agency
Chesapeake Energy Corp.
Internal Revenue Service
ConocoPhillips
US Environmental Protection Agency
US Environmental Protection Agency
Devon Energy
Halliburton Energy Services
Pennsylvania State University
Williams Producton RMT
Geo Logic Environmental Services, LLC
US Department of Energy
Hess Corporation
Denbury Resources Inc
Devon Energy Corporation
Energy & Environmental Research Center,
University of North Dakota
US Environmental Protection Agency
Newfield Exploration Company
US Environmental Protection Agency
Cornell University
Utah Department of Environmental Quality
US Environmental Protection Agency
US Environmental Protection Agency
US Environmental Protection Agency
US Department of Energy
3
-------
Matthew
Hal
Jeff
Carl
Briana
Gary
Greg
Jon
Richard
Steve
Robert
Randall
David
Daniel
Gregory P.
Mark
Talib
Cody
D. Steven
Norm
Robert
JohnH.
Nathan
Bryce
Name
Land is
Macartney
McDonald
Miller
Mordick
Norris
Oberley
Olson
Patterson
Platt
Puls
Ross
Russell
Soeder
Stanton
Stebbins
Syed
Teff
Tipton
Warpinski
Whiteside
Williams
Wiser
Yeager
Affiliation
US Environmental Protection Agency
Pioneer Natural Resources USA, Inc.
US Environmental Protection Agency
US Environmental Protection Agency
Natural Resources Defense Council
US Environmental Protection Agency
US Environmental Protection Agency
The University of Texas at Austin
El Paso Exploration and Production
US Environmental Protection Agency
US Environmental Protection Agency
US Environmental Protection Agency
QEP Energy Co.
US Department of Energy
US Geological Survey
CNX Gas Company LLC
TSA, Inc.
Shell Exploration and Production Company
Newfield Exploration Mid-Continent Inc.
Pinnacle - A Halliburton Service
Texas World Operations
US Geological Survey
US Environmental Protection Agency
Energy Corporation of America
-------
Agenda
Technical Workshops for the Hydraulic Fracturing Study
Well Construction & Operations March 10-11,2011
US EPA Conference Center
One Potomac Yard (South Building)
2777S. Crystal Drive
Arlington, VA 22202 Room S-1204 and 1206
March 10, 2011
7:30 am Registration
8:00 am Welcome
Jeanne Briskin, Hydraulic Fracturing Study Task Force Lead, EPA Office of Research &
Development
Scott Anderson, Environmental Defense Fund, Workshop Lead
Pat Field, Facilitator, Consensus Building Institute
Theme 1: Well Construction
8:15 am Technical Presentation Session 1: Considerations for Aquifer Protection
Overview of the Well Construction Sessions, Bob Whiteside, Texas World Operations
Public Water Sources and Hydraulic Fracturing -A State Drinking Water Perspective, Mark
Jensen, Utah Department of Environmental Quality
Well Completion Methods for Aquifer Protection, Bill Godsey, GeoLogic Environmental
Services, LLC
9:30 am Break
9:40 am Technical Presentation Session 2: Well Design
Well Planning and Construction Techniques, Carolyn Debrick, Devon Energy
Production Casing Design Considerations, Brad Hansen, Devon Energy
Well Construction Practices in the Marcellus, Cody Teff, Shell Exploration and Production
Company
10:55 am Break
-------
_,f
if
11:05 am Technical Presentation Session 3: Drilling and Completion
Multi-Well Pad, Tight Gas, Directional Drilling Program Protects Aquifers, Jay Foreman,
Williams Production
Casing Perforating Overview, Brad Hansen, Devon Energy
Cementing, Cement Quality Evaluation/Logs and Zonal Isolation for Hydraulically Fractured
Wells, TalibSyed, ISA, Inc.
12:20 pm Lunch
Theme 2: fracture Design and Stimulation
1:30 pm Technical Presentation Session 4: Geologic Assessment
The Distribution of Natural Fractures Above a Gas Shale: Questions About Whether Deep
Fracture Fluid Leaks into Groundwater Outside the Realm of Faulty Borehole Construction,
Terry Engelder, Pennsylvania State University
Evaluation of Well Records and Geophysical Logs to Determine the Presence of Freshwater,
Saltwater, and Gas above the Marcellus Shale, South-Central New York, John Williams, US
Geological Survey
2:30 pm Break
2:40 pm Technical Presentation Session 5: Fracture Propagation
Fracture Design in Horizontal Shale Wells - Data Gathering to Implementation, Tim Beard,
Chesapeake Energy
Evaluating Hydraulic Fracture Propagation in a Shallow Sandstone Interval, David Cramer,
ConocoPhillips
Hydraulic Fracturing in Coalbed Methane Development, Raton Basin, Southern Colorado,
Hal Macartney, Pioneer Natural Resources USA, Inc.
3:55 pm Break
-------
4:05 pm Technical Presentations Session 6: Monitoring
Monitoring a Frac Treatment- How Do We Know What is Going On?, Mike Eberhard,
Halliburton Energy Services
A Case History of Tracking Water Movement Through Fracture Systems in the Barnett
Shale, Patrick Handren, Denbury Resources
Measurements and Observations of Fracture Height Growth, Norman Warpinski, Pinnacle -
A Halliburton Service
5:20 pm Revisit the Major Discussion Points of the Technical Presentation Sessions
Scott Anderson, Environmental Defense Fund, Workshop Lead
Bob Whiteside, Texas World Operations, Theme Lead - Well Construction
Tim Beard, Chesapeake Energy Corporation, Theme Lead - Fracture Design and
Stimulation
5:45 pm Adjourn for the Day
March 11, 2011
Continuation of Theme 2: Fracture Design & Stimulation
8:00 am Technical Presentation Session 7: Verifying Zonal Isolation
Sustainable Fracturing Rationale to Reach Well Objectives - The Impact of Uncertainties
and Complexities on Compliance Assurances, Ahmed Abou-Sayed, Advantek International
Design and Rationale for a Field Experiment using Tracers in Hydraulic Fracture Fluid,
Daniel Soeder, US Department of Energy, National Energy Technology Laboratory
Review of Stimulation Water Retention Mechanisms and Likelihood of Fluid
Communication with Shallow Aquifers, Scott Cline, Unaffiliated
9:15 am Break
Theme 3: Well Integrity
9:25 am Technical Presentation Session 8: Pre- and Post-Hydraulic Fracturing Well Integrity Test
Methods
Assessment Methods for Well Integrity during the Hydraulic Fracturing Cycle, Jim Bolander,
Southwestern Energy
Pre & Post Well Integrity Methods for Hydraulically Fractured/Stimulated Wells, Talib Syed,
TSA, Inc.
10:25 am Break
-------
10:35 am Technical Presentation Session 9: Case Studies
Case Study for Well Integrity over a Full Life Cycle, Lloyd Hetrick, Newfield Exploration
Company
Risks to Drinking Water from Oil and Gas Wellbore Construction and Integrity: Case Studies
and Lessons Learned, Briana Mordick, Natural Resources Defense Council
11:35 am Revisit the Major Discussion Points of the Technical Presentation Session
Scott Anderson, Environmental Defense Fund, Workshop Lead
Tim Beard, Chesapeake Energy, Theme Lead - Fracture Design and Stimulation
Jim Bolander, Southwestern Energy, Theme Lead - Mechanical Integrity
12:00 pm Closing Discussions
Susan Burden, EPA Office of Research & Development
Scott Anderson, Environmental Defense Fund, Workshop Lead
-------
Summary and Abstracts from Theme 1: Well Construction
-------
Summary of Presentations for Theme 1: Well Construction
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA. Any mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
Technical Presentations
The first set of technical presentations in this theme addressed considerations for aquifer
protection.
Bob Whiteside, Texas World Operations, described the three primary categories of well
completion. Case 1 wells have surface casings cemented continuously from the surface down into
water-bearing formations with waters having greater than 10,000 mg/L total dissolved solids
(IDS). The casing and cement would extend through any waters with less than 10,000 mg/L total
dissolved solids. Intermediate casing would be set through the surface casing to a greater depth.
Case 2 wells have surface casing cemented continuously from the surface down into water-
bearing formations with waters having greater than 10,000 mg/L IDS, but have no intermediate
casing. In some situations, ground water protection can be enhanced through the use of an
external casing packer. Case 3 wells have surface casings cemented continuously from the surface
to depths within water bearing formations with water of less than 10,000 mg/L IDS. Mr.
Whiteside noted that very few incidents of ground water contamination have been associated
with Case 1 and Case 2 wells. Shallow Case 3 wells pose more of a challenge, but safety and
aquifer protection can be ensured through proper planning, identification of artificial
penetrations, careful geological study, and close attention to fracture procedures.
Mark Jensen, Utah Department of Environmental Quality, discussed the Utah Drinking Water
Source Protection Program which is designed to help public water suppliers protect their drinking
water wells, springs, and intakes. A source water protection plan delineates protection zones,
inventories potential contamination sources, and develops plans to address current and future
sources of potential contamination. In Utah, source protection zones are delineated based on
ground water travel time or hydrogeologic boundaries. Land management strategies within the
protection zones are developed and implemented by the public water system (PWS) and would
be one part of a source protection plan that could consider hydraulic fracturing projects in the
area. Mr. Jensen emphasized the importance of collaboration between the PWS, state agencies,
and other groups.
Bill Godsey, Geo Logic Environmental Services, LLC, gave an overview of how conducting due
diligence while following established industry standards on well construction and operations can
ensure aquifer protection during HF operations. Conducting HF in a manner that protects drinking
water resources relies on identification of aquifers and water wells, identification of potential
migration pathways, and appropriate casing and cementing programs. Knowledge of adjacent oil
and gas fields, as well as coal, lignites, and other mineral resources, is also important. Mr. Godsey
concluded that HF can be conducted safely, ensuring aquifer protection, when there is
appropriate site characterization and planning. Mr. Godsey also provided information the
locations, areal extent, and general size of on major and minor aquifers in East Texas.
10
-------
The second set of technical presentations addressed well design.
Carolyn Debrick, Devon Energy, discussed Devon's well planning and construction techniques
used in the Haynesville Shale in East Texas. Ms. Debrick explained that the goal of drilling from an
engineering perspective is to have good production for the entire life of the well. A properly
designed well will meet this goal while performing in an environmentally safe manner. For
example, casing should be designed to handle the loads incurred during drilling and the operating
life of the well. Ms. Debrick emphasized the importance of successful primary cementation of
surface casing. She noted that a remedial cement job is expensive and does not provide the same
level of isolation as a primary cement job.
Brad Hansen, Devon Energy, described production casing and design considerations for safe and
productive wells. He discussed three primary casing design factors: ensuring mechanical integrity,
optimizing well cost, and providing well completion field personnel with the important design
specification of maximum allowable loads. Before HF is conducted, it is important to know the
maximum allowable fracture pressure. For this calculation, additional tension loads must be
considered, such as those due to cooling of the production casing and to increased internal
pressure, which may cause a ballooning effect on the production casing.
Cody Teff, Shell Exploration and Production Company, provided an example of Shell's well
planning and construction techniques in the development of the Marcellus Shale to both identify
and protect potential subsurface drinking water sources. Shell uses geologic information and
seismic interpretation to identify and manage hazards that could compromise the integrity of the
wellbore.
The third set of technical presentations addressed well drilling and completion.
Jay Foreman, Williams Production, discussed various engineering and regulatory controls
intended to produce natural gas safely, economically, and in an environmentally sound manner.
Proper procedures are intended to allow drilling and completion operations to be performed
without endangering drinking water supplies. These procedures include cementing the casing and
conductor pipe in place, evaluating wireline logs prior to HF, and closely monitoring downhole
pressure.
Brad Hansen, Devon Energy, provided an overview of casing perforation. The primary objective of
perforations is to create effective flow communication between the cased wellbore and a
productive reservoir. The perforating gun consists of four components: a carrier, a shaped
charge, the detonator cord, and the detonator. The impact pressure, which ranges from 10 to 15
million psi, overcomes the strength of casing and formation and forces material away radially,
creating holes through the casing and pathways into the reservoir formation.
Talib Syed, TSA, Inc., discussed important aspects of well design such as cementing, cement
quality evaluation and logs, and zonal isolation techniques. These are key factors for assuring
wellbore integrity, which is important to ensure that production occurs in a controlled, safe
11
-------
manner and to prevent fluids from possibly migrating into underground sources of drinking water
(USDWs).
Summary of Discussions Following Theme 1: Well Construction
Presentations
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA. Any mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
External casing packer. The presenters explained the importance of using external casing packers
to create an external seal between the casing and the sides of the wellbore, as several transient
pressures are created when a well is fractured. According to the presenters, the external casing
packer provides another mechanism for preventing USDW contamination in Case 2 wells (as
described by Mr. Whiteside). The presenters clarified the term "permanence" with respect to a
seal. A presenter indicated that a good seal is above the fracture zone and will last in the range of
20 to 30 years. According to the presenter, the lifespan of a seal can be increased to 36 or 38
years if swellable packers are used.
Surface casing determination. A participant asked about determining the depth of surface casing
and inquired about the process of safely putting surface casing through an aquifer or ground
water supply. The presenters explained that the location of protected water is determined
through maps and other sources of data and can vary depending on the state definition of
"useable" water. One participant cautioned that the depth of usable water may not be the same
as the depth of the bottom of the formation. Many states have specific depth requirements for
surface casings.
Diagnostic tools to check for construction features. A participant asked whether there are
diagnostic tools to check casing, proper cementation, and other design features, and if industry
runs them routinely. The presenters stated that protecting water is industry's responsibility and
that they receive oversight from state agencies. According to the presenters, there is no current
standard for running logs; this is dictated by local geology, the exploration/development phase of
the field, and other criteria. Participants claimed that there are ways to drill wells in a manner
that is protective of ground water and there is a movement in Texas for all drilling plans to be
approved by a professional engineer.
Ground water modeling in fractured bedrock. A participant asked how ground water models are
developed for fractured bedrock. In Utah, there are no requirements for the use of specific
models, though the ground water modeling method must be applicable to the area of interest. A
presenter indicated that locations of faults and the boundaries of aquifers are key components of
hydrogeologic mapping. Analytical methods, such as calculated ground water travel time, are also
used.
12
-------
Contamination and drinking water well construction. The presenters clarified that well head
protection is especially a problem for domestic water supply wells. According to the presenters,
most contamination complaints concern domestic supply wells. Presenters claimed that there are
often inherent problems with these wells. The presenters stated that results of most modeling
studies and chemical analyses do not indicate contamination from subsurface migration. Instead,
they stated their belief that contamination is more often linked to well head protection issues
and cannot be attributed to oil and gas activities.
Fractures leaving the intended zone. Several participants asked about the number of times
fractures have left the intended zone. They expressed concern about containment and assurance
of fracture control. The presenters stated that shale in East Texas is approximately 100 ft thick,
and that it is common for the first well to be fractured downward with respect to the target
formation. Fracturing upward is done when there are many wells in the area. However, the
presenters stated that operators are highly confident that fractures will not leave production
zones. In the Eagle Ford Shale, depleted production zones above and below the gas reservoir act
as barriers for pressure transmission and encourage lateral leakoff within the depleted zone.
According to one presenter, the most significant concern is the impact to the wellbore as
fractures propagating upward move closer to the well. Participants emphasized that staying
within the intended zone is also important for production. One participant recommended core
studies as a source of information on rock mechanics to provide fundamental information
regarding fracture propagation.
Shallow gas sources and drilling with air. A participant asked about addressing the issue of
shallow gas sources within or below a drinking water aquifer. The presenters stated that well
design, drilling techniques and cement together can establish zonal isolation, which is especially
important if there are shallow gas zones. There is concern about potential flow of the liquid
components of the cement during cementation as the cement hardens; this flow was considered
by participants to be one of the causes of channels and microannuli that can compromise the
integrity of cement around the well. Guidance is available, for example from the American
Petroleum Institute (API), on surface casing, cementing and drilling with air. Air drilling is
sometimes conducted when installing surface casing and eliminates the need for drilling muds at
relatively shallow depths. One participant noted that, in areas with a history of natural migration
of gas, an operator's goal should be to not exacerbate any existing problems. A participant added
that drilling with air does not seem to be exacerbating any gas migration issues in Pennsylvania.
Longer lateral sections of the well. Participants stated that longer lateral sections of the well allow
greater gas recovery at lower cost and less surface impact. The presenters explained the
importance of a longer lateral. According to the presenters, a longer lateral is a more effective
use of surface facilities since it allows for more resource recovery without additional or bigger
surface facilities and also reduces the number of wells that must be installed to drain the
reservoir.
Best practices and information collection. Several participants asked about the identification and
implementation of best practices. The presenters explained that some operators do meet
occasionally to talk about wells and share best practices. A participant asked whether there is a
13
-------
standard or recommended process to collect information in a given zone and, although
participants acknowledged the benefits to a standard process to ensure contamination does not
occur, specific information to be collected was not mentioned. Other participants emphasized the
importance of looking specifically at the available data and identifying data that need to be
collected. Sources of information include the Society of Petroleum Engineers (SPE), the API, the
American Association of Drilling Engineers (AADE), and state agencies. Participants noted that
well design is constantly evolving and service companies are constantly developing new products.
Microannuli. Several participants asked about gas migration through microannuli over the life of
the well. The presenters stated that pressure is drawn down and is lower at the wellbore during
production. The objective is to draw gas toward and through the perforations so that gas will flow
into the well; therefore, gas should not flow out through microannuli in the wellbore cement. The
presenters noted the importance of taking remedial action during the drilling and construction
phase to address any cement problems like microannuli. Several participants asked about the
growth of microannuli over time. The presenters stated that they are not aware that microannuli
grow over time. The presenters emphasized that the goal of maximum production is an incentive
to fix any problems immediately as they occur. A participant stated that microannulus flow does
not necessarily mean "micro" scale flow. Participants wondered whether that flow would impact
ground water. According to one participant, it is not flow rate within microannuli, but rather
leakage through the cement out of the wellbore, that is important.
14
-------
Abstracts for Theme 1: Well Construction
Abstracts were submitted to U.S. EPA by the presenters for use in this proceedings document.
Not all presenters submitted abstracts of their presentations.
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA. Any mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
15
-------
Overview of the Well Construction Sessions
Bob Whiteside, P.E.
Texas World Operations/Signa Engineer Corp
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Introduction
The extraction of hydrocarbons from shale and other low permeability formations using
hydraulic fracturing technology has lead to the development of many new oil and gas reserves
and many new environmental questions. Newspaper and television reports highlight water
contamination cases, wildlife losses and surface water incidents on what seems like a daily
basis. Environmental agencies, defenders of the environment, academicians and oil and gas
professions are all searching for answers.
A portion of the of problem lies within the definitions and the terms of what is considered
ground water and/or underground sources of drinking water. The Underground Injection
Control Regulations (40 CRF 140-148) dealt with the same issues by defining a USDW as any
formation containing water with less than 10,000 mg/l Total Dissolved Solids. While the UIC
regulations deal with very limited numbers and specific type of wells, the oil and gas industry
extracts hydrocarbons from where they are located within the subsurface. While most
production wells are well above the USDW definition, a growing number of wells exist within
the lower salinity formations. Some examples are wells in Wyoming which produce from
formations containing 5000 mg/l TDS or coal methane wells located in formations that have
much lower TDS's.
Eight presenters will give papers dealing with a range of topics which include regulatory
concerns, well completion methods, casing design, cementing practices and testing methods.
The session is constructed to give the listener a sense of what is currently being done within the
oil and gas industry to protect ground water and introduce so of the latest techniques to
enhance protection of human health and the environment. After each three presentation set, a
question and answer period will follow. All participants in the workshop are encouraged to ask
questions and seek answers during those times.
Well Settings in Texas
Well completion can be easily broken into three primary categories:
Case 1: Wells that have surface casings cemented at depths containing waters
greater than 10,000 mg/l TDS with intermediate casing set deeper
Case 2: Wells that have surface casing cemented at depths containing waters
greater than 10,000 mg/l TDS with no intermediate casing.
16
-------
Case 3: Wells whose surface casings are cemented at depths less than 10,000
mg/l IDS
Case I Well Considerations
The deeper wells incorporate traditional designs and completions, which are adequate for
ground water protection. Groundwater in the Class 1 scenario is protected by multiple layers of
casing and cement. When standard API, SPE and industry standards are incorporated into the
well design; little, if any, additional consideration is required to adequately protect
groundwater.
In Texas, the Railroad Commission of Texas (RRC) prescribes the method of cementing, the
number of centralizers, the excess quantity of cement required and other design
considerations. The minimum and maximum depth the surface casing must be set at is
prescribed by the Texas Commission on Environmental Quality (TCEQ). Operators are required,
by law, to apply for and receive a drilling permit from the RRC and a letter entitled "Depth of
Usable-Quality Water to Be Protected" issued by the TCEQ Surface Casing Team, Waste Permits
Division before drilling can begin.
After the surface casing is cemented and generally 5 to 10 feet of new borehole has been
drilled, a Formation Integrity Test (FIT) is performed. The FIT is a hydrostatic pressure test that
is designed to determine if the surface casing cement job has adequate strength to drill further
and if the formation in which the casing is terminated has sufficient strength to withstand any
pressure event that might occur while drilling. If the wellbore passes the FIT, the well can safely
be drilled deeper.
Once drilling has progressed through geological formations which lack sufficient strength to
withstand expected production pressures or are too weak to support further drilling
operations, an intermediate casing is set and cemented in place. Generally only the lower
sections of these casings are cemented. Texas regulations require intermediate casing to be
cemented from the bottom of the casing to a height above ... hydrocarbon or geothermal
resource fluids ... (TAC, Title 16, Part I, 3, §3.7). The intermediate casing and cement provides
additional layers to protect groundwater and decreases the probability of hydraulically
fracturing into groundwater formations.
Case 2 Well Considerations
The shallower wells only have cemented surface casing covering the TCEQ-described useable
water. Therefore, additional design elements have been added to reduce risk and avoid ground
water incidents.
These wells have traditional surface casing and cementing designs. In some cases, enhanced
ground water protection is achieved by means of an external casing packer (ECP). An ECP is an
17
-------
inflatable packer consisting of an inflation bladder, a deformable set of steel slats and an outer
rubber covering. The ECP is screwed onto the bottom of the surface casing and run in the hole
with the casing. Once the casing is set at its maximum depth, cement is pumped through the
casing, around the outside of the casing, up the annulus and eventually exits the top of the
wellbore. At the end of the cement column being pumped down the inside of the casing is a
wiper plug to separate the cement from the displacement fluid. The wiper plug activates the
inflation ports within the ECP body which allows fluid and pressure to enter and inflate the
packer. Once the packer is inflated, a permanent mechanical seal is formed between the
bottom of the casing and wellbore in the confining layer below the aquifer.
Examination of bond logs within the Eagle Ford field has shown a number of wells with "gas
cut" cement. When gas is entrained in the cement slurry during emplacement, channeling and
contamination of the slurry can result in poor bonding. Hydraulic fracturing pressures can
further degrade the cement column and, in extreme cases, impact the cement behind the
surface casing. Incorporating an inflatable ECP in the production casing is one way to reduce
the risk to ground water. The packer is inserted into the production casing with a mechanical
port collar immediately above the packer. The ECP placed so that it will inflate and seal at the
junction of the production formation and the formation above. The ECP is inflated by means of
a tool run on a workstring. After inflating the packer, the port collar is opened to allow the
annulus between the wellbore and the casing to be cemented.
Case 3 Well Considerations
The Case 3 wells are located in or adjacent to useable ground water (under any definition).
Therefore, there is no way to protect useable ground water. The only thing that can be done is
a comprehensive ground water study of all existing water wells in the area that are at a depth
within 500 feet of the top of the production zone. A full suite of tests should be performed by a
certified lab for metals, salts, and organics before any drilling or fracturing is performed. If the
ground water is already contaminated by natural causes, an aquifer exemption should be
issued.
Follow up testing should be required after fracturing activities. An area of review of
approximately 3 miles should be a minimum with all water wells tested. If municipal water
wells are involved, a reasonable "off-limits" distance should be applied (approximately 5 mile
radius). The regulators and the operators should look closely at performing smaller frac jobs to
limit height and more stages to limit fracture growth. The nearest water wells should be
sampled within 50 to 60 days of the frac activity to determine impact and on a quarterly basis
for a period of no less than 2 years. If water quality parameters in the water wells do not
change, the operator should be safe in the assumption that impacts to the aquifer have not
occurred as result of hydraulic fracturing. If any of the water quality parameters have changed,
the state regulatory program may want to reconsider the value of the aquifer or require the
operator to provide water from other sources.
Oil and gas production is always a matter of economics. If the operators feel there is no penalty,
18
-------
then they will drill shallow wells to make easy money. The real question is: "Do we need
shallow production or do we save water resources?" In many cases, shallow wells may be
profitable because the natural concentrations of salts, metals or other compounds are already
elevated. Shale plays are growing in number every day. Do we have to produce all of them just
because we can? The only true evaluation is the value of the hydrocarbon versus the value of
the water. Currently operators are not really being forced to make those decisions but that
does not means they shouldn't be forced to. With today's gas prices and an increasing gas
supply, I believe that wells that produce gas with oil should be rated as more valuable to the
energy market than just a well that produces "cheap gas" which may endanger the
environment.
The Texas definition for "useable-quality water" is the same as Ohio and the DIG definition for
underground source of drinking water (USDW) - ground water with 10,000 ppm TDS or less
must be protected. Under those restrictions, it will generally be better for the operators to drill
deeper and the ground water is no longer an issue.
Conclusion
Very few, if any, incidents of ground water contamination have been reported from the wells
listed as Case I and Case 2. These wells generally are inherently safe because of the depth of
useable water protected by surface casing and cement. Inexpensive enhancements can be
added where the depth to useable waters is relatively close.
Case 3 wells pose a different challenge to the drilling engineer and the regulator. These wells
can be made safe by careful planning, additional geological study, and close attention to
fracture procedures. However, each well field must be considered on an individual basis. Texas
uses the "aquifer exemption" regulations to determine if drilling and production from these
fields can be conducted in a manner which is protective of human health and the environment.
19
-------
Public Water Sources and Hydraulic Fracturing - A State
Drinking Water Perspective
Mark E. Jensen
Utah Department of Environmental Quality / Division of Drinking Water
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
The Utah Drinking Water Source Protection program is designed to help the public water
suppliers to protect their wells, springs, and intakes. Source protection zones in Utah are
delineated based on groundwater travel time or hydrogeologic boundaries. This delineation
method requires site-specific hydrogeologic and source construction information. About 58
public wells, springs and tunnels are located in oil and gas fields, but over 200 public water
sources are located in coal deposit areas. Land management strategies within the protection
zones are developed and implemented by the public water systems, and the public water
systems would be involved in potential hydraulic fracturing projects that may impact their
water sources.
20
-------
Well Completion Methods for Aquifer Protection
William E. Godsey, PG
Geo Logic Environmental Services, LLC
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Introduction
Hydraulic fracturing of highly variable hydrocarbon producing geologic formations can be
conducted safely and in an environmentally protective manner using well established
petroleum industry standards. Geologic, environmental and engineering characteristics have
utilized numerous fracturing media and techniques that are used in a variety of applications.
The industry standards, in conjunction with appropriate due diligence and inquiry in the area of
the targeted area for hydraulic fracturing, can result in successful well completions and
groundwater protection.
Purpose
The purpose of this presentation is to outline how appropriate due diligence can be undertaken
concurrent with leasing, site acquisition, permitting and development of prospect areas to
identify and mitigate potential pathways of frac fluids other than intended target zones.
Identification of potential pathways for fluids will allow for drilling, completion and hydraulic
fracturing and can identify potential areas of concern and provide the engineering and design
of the well bore construction team the opportunity to prevent negative consequences,
regardless of the depth of the wells.
Location
This presentation is applicable to any location where hydraulic fracturing is conducted.
Examples and illustrations are taken from the State of Texas where numerous geologic and
geographic settings exist as does a long history of hydraulic fracturing throughout hundreds of
oil and gas fields and the completion of tens of thousands of wells. Examples of aquifer diversity
and extent are illustrated from Texas and Oklahoma.
Methods
The methods utilized in this presentation include literature review, personal interviews and
experience as a state regulator, as an oil and gas operator, as a consultant to industry, local,
federal and state government, water supply corporations, mining companies and legal entities
as expert witness. Graphical representations taken from data published by state agencies were
used to illustrate specific site circumstances.
21
-------
Appropriate Due Diligence
In many cases, once the prospect is developed and the leases are taken, it is up to the drilling
and completion departments of the companies to drill and complete the well. The engineers
responsible for the casing and cementing of the well have numerous factors to consider for the
proper design of the well. Not only does the well have to be designed properly for the target
zone to be completed and stimulated, but other factors must also be assessed. Among the
factors to consider are near surface conditions and well pad stability. Well pad stability and
near surface wash out is usually managed by setting of conductor casing. The uppermost
aquifer and base of usable quality drinking water must be isolated and protected. Fresh water
intervals are usually protected by the surface casing. Pressurized zones, or formations which
produce oil and gas, between the surface casing and the total depth of the well must be
isolated, too. An intermediate casing can be used to provide additional fresh water protection
or isolation of productive and/or pressurized zones.
Research by others has shown that the fracture influence in deep shale gas is limited to a few
hundred feet from the well bore. Fracturing in shallow coal beds for methane is a separate
mechanism1 from deep shale gas fracturing; however, the investigation for potential pathways
is the same.
Aquifer Identification
Beginning at the surface, inquiries as to the types of aquifers present and the use of these
aquifers is advised. The classification and definition of groundwater varies from state to state.
Therefore, it is crucial to understand, the nature and areal extent of the hydrogeologic
conditions of the area. Not all freshwater-bearing aquifers are utilized. Some of the aquifers
have objectionable characteristics such as high iron or sulfate concentrations which render an
objectionable taste, unless treated before consumption. Identification of large capacity
municipal supply wells is suggested as these wells generally supply a large number of people as
opposed to a single family.
Some aquifers are so massive that the water quality changes with depth as does the use of the
water. Some fresh water aquifers are even known to produce hydrocarbons naturally and the
same formation can be so extensive as to have water quality become brackish-to-saline and
produce oil and gas as well, such as the Wilcox Formation.
While general HF operations are similar for coal beds and shale, the details differ for HF operations in these
different geologic settings. For example, in the presence of typical fracturing fluids, coal tends to swell which
reduces permeability through fractures and, therefore, reduces production. To control coal swelling, the
approach to fracturing coal beds can include the use of different fracturing fluid mixtures or gas-based (nitrogen or
carbon dioxide gas) fracturing fluids. Fracture design can also be different because coal has a distinct natural
system of fractures (cleats and joints), can have different stress and strain regimes, and can require significant
dewatering prior to gas production. (Explanation provided by The Cadmus Group)
22
-------
Water Well Inventory
Determining the number, types and utilization of water wells in the area of a well being
hydraulically fractured can be cumbersome if sometimes nearly impractical for a number of
reasons. Experience has shown from investigations of water well complaints that there are a
number of issues that repeatedly come forth. The issues include a lack of information by the
well owner about the well construction and age of the well, the company or person that drilled
the well or other pertinent information. In many cases, there are other factors such as poor
well head protection, poor casing quality and a lack of sanitation around the well. Other
problems include poor drainage around the well and close proximity to septic systems,
especially in rural areas not serviced by sanitary sewer systems.
Identifying wells that could be potentially impacted by fluids from a hydraulically fractured well,
should they escape would be beneficial for any investigation. In most cases it is practical only to
identify large capacity municipal water supply wells prior to beginning drilling. This information
is usually available from state agencies.
Adjacent Oil and Gas Fields
Some areas where hydraulic fracturing may take place will involve penetration through
shallower oil and gas fields. Deeper penetrations may exist through zones where fracturing is to
take place. In either situation, evaluation of potential pathways for migration should take place
to avoid conditions where fluid migration may occur. These zones have been proven to be
effectively isolated by casing and cement in numerous applications. Identification of producing
zones that occur at depths shallower than the target zone and especially immediately above
the target zone is advisable. Examination of penetrations through the target zone to assure
appropriate isolation is suggested as well.
Areas where oil and gas exploration have taken place also contain previous well bores which
have been plugged and abandoned or drilled as "dry holes." These well bores should be
identified and evaluated as potential pathways prior to development of the target area. In most
cases, these wells have been identified and are known by state regulators and are mapped
accordingly.
Coal, Lignite and Other Mineral Resources
In numerous oil and gas producing areas, other mineral assemblages are also present. One of
the most common mineral resources encountered is coal and lignite in near surface deposits.
Both coal and lignite are known to produce methane naturally. When these mineral beds are
highly fractured naturally and water moves through the units, minerals such as pyrite and other
forms of iron and sulfur can form in the fractures giving the water an objectionable quality.
Where these minerals are present in sufficient quantities, mining may have occurred in
underground or near-surface operations. These activities should be noted when drilling in areas
where coal and lignite resources are found. In some areas of Texas, drilling occurs in active
mine areas and is compatible with mining activity.
23
-------
Summary and Conclusions
Hydraulic fracturing can be conducted safely and aquifers can be protected when appropriate
site investigation is conducted. There are many sources of information available for review.
Once aquifers are identified and potential pathways for potential exposure are identified,
appropriate casing and cementing designs can be implemented to address the specific site
conditions. Hydraulic fracturing of geologic formations varies from region to region.
Groundwater conditions and quality vary from region to region and protection/isolation
techniques are available to address these variables. There are numerous approaches to
hydraulic fracturing that involve various propping materials and delivery fluids. The key to
successful hydraulic fracturing is identification of aquifers, location of potential pathways and
appropriate casing and cementing programs to assure the frac materials remain in the target
zone.
24
-------
Well Planning and Construction Techniques
Carolyn S. Debrick, PE
Devon Energy
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
This paper will focus on the well planning and construction techniques Devon uses in the
Haynesville Shale. It will briefly cover issues that are related to designing and drilling the well
safely and protecting subsurface drinking water sources. The goal as a Devon drilling engineer is
to design the well for the maximum volume of fluids and pressure to be encountered while
drilling as well as for the entire life of the well, and to do so in an environmentally safe manner.
Background
The Haynesville Shale is located in East Texas and North Louisiana. This paper will focus on
Devon's well design and construction in the East Texas Haynesville Shale. This paper discusses
an area covering six counties in East Texas. The depth of the shale varies from 10,500 ft to
13,500 ft. Pore pressure varies from an equivalent mud weight (EMW) of 15 pounds per gallon
(ppg) to 18 ppg. Bottom hole temperatures range from 270 degrees F to 350 degrees F.
Measured well depths range from 15,000 ft to 20,000 ft.
General Data Gathering
Before the well can be designed various data needs to be gathered and interpreted. The data
gathered is as follows: pore pressure, fracture pressure, fresh water zones, temperature
gradients, squeezing or unstable formations, depleted zones, disposal zones, sensitive shales,
shallow gas hazards, presence of H2S or C02, geologic targets, well interference data, minimum
hole size required, production casing size required, completion design and fluids, topographic
surface restrictions, and regulatory requirements.
General Well Design
The pore pressure and fracture gradient chart with the geological data is the basis for the entire
well design. This determines how many casing seats will be required and consequently what
diameter casing size is set at surface. The information is displayed in terms of EMW. The
geological cross section is also included as some formations have higher or lower shoe
strengths. There are empirical methods of determining pore pressure from logs and seismic
data but the best information is from offset wells. Information on reservoir depletion due to
production is also gathered and an estimated bottom hole pressure due to hydrocarbon
withdrawal is determined. In East Texas, disposal wells that inject fluids may have higher than
normal pore pressures. Information from the disposal well operators is gathered, and
subsequently, a fracture gradient chart is created. Fracture gradient is a function of overburden
and pore pressure but varies depending on the age of the rock and the in-situ stresses.
25
-------
Equations generated over the years by industry experts can be used but actual offset leak off
test data from the surrounding area is best.
For the Texas Haynesville Shale, Devon Energy uses more than three different well designs
depending on where geographically the well is drilled. In Panola County, there is a high density
of producing wells and some formations are depleted to an EMW of 2 ppg. In addition there are
disposal zones that are charged above normal pressures. These factors impact the number of
casing strings required to drill the well and the surface casing setting depth.
Directional Plan: Anti-Collision
Part of the early well planning process is to assess possible hazards such as potential collisions
with existing well bores. Devon Energy looks at all offset wells including producing wells,
abandoned wells, disposal wells, and any water wells.
Often Devon drills multiple wells off an existing pad or platform. In this case, survey data from
each existing well to compare to the well that is being planned.
There are various surveying tools used to measure and determine the path of a well bore. Each
of these tools has a degree of inaccuracy. This inaccuracy varies depending on the type of tool,
(i.e. gyro vs single shot or magnetic). This uncertainty is translated into an ellipse referred to
as "the ellipse of uncertainty". It is assumed that the actual well bore can lie anywhere within
the ellipse. The size of the ellipse of uncertainty depends on the type of tool run. Each type of
tool has been assigned an "error factor" by experts that help determine the size of the ellipse.
The anti-collision calculations take into account this "error" and adjust the ellipses accordingly.
We can then examine the survey data combined with ellipses of uncertainty to asses any
possible risk of collision.
Vertical wells can be legally surveyed using rudimentary angle only devices. While these basic
surveys satisfy the legal requirements for surveying, they do not provide adequate information
to track the well bore for anti collision purposes. If a well does not have adequate survey data
we will survey the well bore in question to gather the necessary data to run anti collision
calculations.
Drinking Water Source Identification and Surface Casing Setting Depth
In Texas, the Texas Commission for Environmental Quality (TCEQ) maintains data on drinking
water protection zones and water wells. TCEQ defines the location of the base of underground
source of drinking water (USDW) and identify water source wells within one-quarter mile. The
USDW in East Texas is typically at the base of the Wilcox formation. This depth can be as
shallow as 250 ft or as deep as 1650 ft or more. Surface casing is to be set within 200 ft below
these zones.
Devon not only considers the depth of drinkable water when determining surface casing setting
depth but also considers what maximum pressure that can be held at the surface in a well
control event and not break down the shoe. This pressure is referred to as maximum
26
-------
anticipated surface pressure (MASP). With shallow shoe depths this pressure is very low (see
Table 1 below).
Table 1.
Shoe Depth
(feet below
surface)
600
1000
1500
Fracture
Equivalent
Mud Weight
fib/gal, or ppg]
11.5
12.5
13.5
Planned Mud
Weight
(Ib/gal, or
PPg]
10
10
10
Maximum Anticipated
Surface Pressure
(lb/in2 or psi)
47
130
234
In Panola County, the depth of the base of the Wilcox is between 250 ft to 350 ft. Based on
these depths and TCEQ requirements the surface casing would be set 450 to 550 ft below
ground surface respectively. Devon requests an exception to set surface casing deeper for
safety considerations. Devon sets surface casing shoes in East Texas deeper than prescribed by
the regulator for two reasons: 1) to provide additional USDW's protection; and 2) to provide
sufficient kick tolerance for drilling ahead.
Casing Design Considerations
The goal of the casing design is to provide a safe and reliable design. The design depends on the
loads that may occur. Major design considerations are drilling loads, casing running loads,
fracture stimulation loads, connection selection, buckling, corrosion issues, temperature
related issues, and compressive loads on surface. Any part of the casing that is not cemented is
subjected to dynamic well conditions and casing movement due to temperature, pressure, and
fluid gradient changes. The selection of the top of cement is based on these considerations.
Connection selection is critical. Most casing failures occur in the connection. Bending,
compression, tensile, and fatigue life when rotated are considered.
For surface casing the connection needs to also be able to support the weight of all the casing
strings and the applied loads associated with the well life. If the compressive loads exceed the
safe rating of the connection, a base plate is installed on the surface casing head.
Cementing Surface Casing
Obtaining a good primary cement job is critical to Devon. Remedial cement jobs are costly
and typically do not provide the same level of isolation. In East Texas we utilize Class A cement
for surface casing. This cement can develop compressive strengths at lower temperatures.
Typically 300 ft of 15.0 ppg neat cement with no fillers is placed on bottom and followed by a
lighter weight 12.6 ppg cement with extenders and accelerators to achieve minimum
compressive strength before drill out. The hydrostatic density of the cement column when the
cement is in its fluid state must not exceed the formation fracture strength. Casing is
centralized with bow type centralizers - one every joint for first 4 joints and one every third to
surface. Haynesville field practices to ensure a good cement job include: conditioning mud
27
-------
before tripping to run casing, running a spacer before the cement, and moving pipe while
cementing. Cement is circulated to surface on all these jobs or a top out job is conducted.
Once the cement is set and the shoe is drilled out, a formation leak off test or integrity test is
conducted. This is to ensure that a good cement job was accomplished and that the shoe and
formation at the shoe has sufficient strength to drill to the next casing seat.
In addition, due to the current and potential for future disposal wells in East Texas, we bring
cement on the next casing string into the surface casing shoe if the formations in the open hole
can hold the hydrostatic column of cement.
Casing and Cementing Horizontal Production Casing
Additional design and planning is required for the Haynesville due to the long measured depth
of the well, the close tolerances of casing to hole diameter, and the high mud weights. Devon
runs a tapered 5.5" by 4.5" casing string in 6-3/4" hole and the typical mud weight at total
depth is 15 to 16 ppg. In terms of cementing, the same field practices apply here as with surface
casing. However, mud and cement rheology are critical in this situation. Prior to pulling out of
the hole to run casing a good practice is to condition the mud to as low plastic viscosity and
yield point as possible. Surge and swab is run to determine the casing running speed.
Calculations are also made on the cement job with the cement and mud rheologies to
determine the maximum pump rate which is usually low. We add rheology improving products
in the cement as well as expander and strength retrogression products. However we still can
lose returns while running casing in the hole or cementing. If there is risk that cement will not
reach inside the intermediate we run a swell packer just above the intermediate casing shoe.
This packer will swell to the casing internal diameter in a maximum of 2 weeks.
This swell packer provides isolation between the two casings. The fracture treatment pressures
for the Haynesville Shale can be as high as 13,000 psi. Back pressure is held on the casing when
possible for safety reasons during the fracture treatment job. In addition the swell packer
provides isolation from any gas in the open hole.
Cementing Intermediate Casing
Typically the intermediate casing string has pay zones behind pipe. When possible we bring
cement inside the surface casing shoe. If this is not possible we run a cement bond log prior
to perforating and stimulating any zone in this casing string.
28
-------
Production Casing Design Considerations
Brad Hansen
Devon Energy
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
This abstract presents information to consider in the design of a safe and effective production
casing string for well production and also as a conduit for a fracture stimulation. The
presentation discusses casing design factors and casing design loads. Pipe performance is
discussed as well as material selection. A description of the various types of casing connections
is given. Also, additional considerations that should be addressed if the well will be hydraulically
fractured down casing are discussed.
There are three major requirements to be considered in designing production casing:
1) Ensure the well's mechanical integrity
2) Optimize well costs
3) Provide operations personnel with the maximum allowable loads
Many factors enter into the production casing design. These include the mud weights required
to drill the well and balance the formation pressures, the fracture gradients, casing seat depths,
casing sizes, the directional plan, the cement program and the temperature profiles. Also, the
type of fracture fluid and proppant to be used, maximum proppant concentration, and the
calculation for the maximum anticipated hydraulic fracture surface pressure should be
considered. The types, composition, and volumes of the anticipated production must also be
considered. This information is used to determine the planned loads over the life of the well.
Once these expected loads are determined, the pipe selection can be made that will meet or
exceed the minimum design factors required by the designer. The design factor is the pipe
rating divided by the anticipated load. This design factor must meet or exceed the minimum
design factor that the designer has set. Most pipe ratings are based on the yield strength of the
pipe. To determine the yield strength of a given material, a specimen is machined and put into
a load cell where tension is pulled and the strain measured on the sample until it fails. A stress-
strain curve is then generated. The yield strength using the API method is defined as the stress
at a strain of 0.5% elongation. This yield strength is less than the ultimate strength of the
sample.
There are two main design cases for internal yield pressure of production casing. One is
modeled with a tubing leak near the surface with the shut-in tubing pressure added to the
packer fluid weight as an internal load. The shut-in tubing pressure is estimated from the
bottom hole pressure minus the weight of the gas in the tubing. The weight of the gas in the
tubing is calculated both at static and at flowing temperatures (sometimes called a hot shut-in)
29
-------
The other internal yield pressure case is injection down casing such as during a hydraulic
fracture stimulation. The internal pressure is modeled by the applied surface pressure and the
fluid gradient based on the fluid being pumped. This is analogous to a hydraulic fracture screen-
out downhole since fluid friction down the casing is not subtracted from the internal pressure
profile. The external casing pressure profile is modeled with the mud gradient from surface to
the top of cement. Then the gradient from the cement mix water from that point to the outer
casing shoe. From the outer casing shoe to total depth (TD), the external pressure profile is the
pore pressure profile.
Production casing collapse loads assumes zero pressure on the inside of the pipe and a final
mud weight gradient on the outside of the casing.
Rated internal yield pressure of casing is calculated using the Barlow Equation below:
P = 0.875 * [2*Yp*T]/D
P= internal yield pressure or burst strength (psi)
Yp = yield strength of the pipe (example P110 is 110,000 psi)
T = nominal wall thickness (inches)
D = nominal outer diameter of pipe (inches)
Per API, the calculated number is rounded to the nearest 10 psi. The 0.875 factor in the above
equation represents the allowable manufacturer's tolerance of minus 12.5% on wall thickness
per API specifications.
Collapse ratings on API tubulars are derived from four different equations based on the outside
diameter / thickness ratio and the yield strength of the pipe.
Axial strength of the pipe body is calculated from the formula below:
Fy = (n/4)*(D2-d2)Yp
Fy = tension strength (Ibs. rounded to the nearest 1,000)
Yp = yield strength of pipe ( psi)
D = OD of pipe (inches)
d = ID of pipe (inches)
Calculations for joint strength can be found in API bulletin 5C3. Published joint strength of API
connections is based on the ultimate strength of the pipe and not the yield strength. Most, but
not all premium connections are based on the yield strength of the connection.
API Spec 5CT is the Specification for Casing and Tubing. The different grades of API pipe specify
a minimum and maximum yield strength. A maximum hardness is also specified from grades
designed for sour service.
30
-------
The chemical composition of the different grades of API casing is also specified. Grades
designed to work in sour service have more stringent chemical requirements.
Sour service is defined by the National Association of Corrosion Engineers or NACE, as an
environment where the partial pressure of H2S exceed 0.05 psia. The total pressure must also
exceed 65 psia for a gas well and 265 psia for an oil well. The NACE standard MR0175 and the
ISO standard 15156 specify material to be used in sour service. In summary, API casing grades
H40, J55, K55, M65, L80, C90 and T95 are good for all temperatures. N80 is good above 150
degrees F, P110 is good above 175 degrees F and Q125 is good above 225 degrees F.
Casing connections represent less than 3% of the pipe length yet account for more than 90% of
pipe failures. Also, the connection represents 10% to 50% of the total tubular cost.
API connections STC (short thread and coupled) and LTC (long thread and coupled) each have 8
threads per inch and have rounded crests and roots. On LTC, the tread section is longer so it will
have better scalability and tensile strength than STC.
A buttress connection is another API connection that has 5 threads per inch. It is not symmetric
for the load and stab flanks.
There are several types of premium connections available, but most fall into one of the
following categories:
A metal to metal seal thread and coupled connection generally has the internal yield, collapse,
and tension ratings equal to the pipe body.
An integral joint connection has half the leak paths of thread and coupled connections. Also,
the connection outer diameter (OD) is significantly smaller than a coupled connection. It also
features a metal to metal seal. The joint strength of an integral joint connection is usually 70 to
80% of the pipe body.
A flush joint connection is approximately the same OD as the pipe body. Its joint strength is
usually only 45 to 60 % of the pipe body strength in tension.
Prior to the hydraulic fracturing of a well, the maximum allowable surface fracture pressure
must be calculated. The fluid gradients inside and outside the pipe are needed to make this
calculation. Not only must the burst (internal yield) pressure of the pipe be considered when
making this calculation but also the effect of the internal hydraulic fracturing pressure and
hydraulic fracture injection rate on tension. The internal pressure during the hydraulic fracture
causes a ballooning effect on the production casing that adds to the tension load. During the
fracture, the production casing is cooled2 by the injection of fracture fluids, which also adds to
Fracture fluids stored at the surface will be near surface temperature, which is generally a much cooler
temperature than the bottom hole temperature.
31
-------
the tension load of the production casing. These additional tension loads must be taken into
consideration when determining the maximum allowable hydraulic fracture pressure.
32
-------
Shell's Well Construction Practices in the Marcellus Shale
Cody Teff
Shell E&P Company
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Introduction
This paper is about the well planning and construction techniques that Shell uses in the
development of the Marcellus Shale to:
Identify potential subsurface drinking water sources (private and municipal)
Protect them through hazard avoidance (target zone planning) and zonal isolation
It will include location selection, directional planning, casing selection and design, cement slurry
design, and integrity testing prior to fracture stimulation.
Background
The Marcellus Shale is a laterally contiguous shale deposit that covers parts of Pennsylvania,
West Virginia, Ohio, New York, and Virginia. This shale has been proven to contain
commercially viable natural gas resources and thus has entered the development phase of a
hydrocarbon resource. As part of the development technique, hydraulic fracturing is utilized to
enable production of natural gas at commercial rates. This technique allows for high
conductivity fractures to be created, or natural fractures to be supplemented, that allow higher
flow surface areas to connect to a wellbore. In order for fracturing operations to commence a
wellbore must be drilled to the appropriate location with the appropriate equipment to allow
integrity and control during the fracturing operation.
Drinking Water Source Identification
The process for the identification of sub-surface drinking water sources starts when a well
location to be drilled has been identified. This process begins with a spatial and title review to
identify offsetting land owners or potential users of sub-surface water for consumption or other
use. Once the spatial and title review have been completed a survey is conducted, via
registered mail, to determine if people within 1,000' of the proposed drilling location have sub-
surface water source wells. If sub-surface water wells are present then information about the
depth of the well is gathered. Additionally a request to conduct a base line survey (Table 2),
including gathering a sample, is requested from the owner or user of the sub-surface water
well.
Well Directional Planning
The directional planning of oil and gas wells are undertaken to hit subsurface targets and avoid
subsurface hazards that could compromise the integrity of the wellbore or the ability to reach
the final objective. There are two basic approaches that are used to identify hazards, the first is
33
-------
by geologic interpretation and the second is by seismic interpretation. Geologic interpretation
uses the latest data available to continually improve the model. The basic inputs are surface
data, including topography and outcrop information. As information on new wells is gathered
this model is improved to grow the understanding and improve future interpretations. The
seismic interpretation is based on the acquisition of seismic data, or the acoustic response of
subsurface features to a surface event. Examples, not all inclusive, of subsurface hazards to
potentially avoid when performing directional planning would be faults, shallow gas, and
shallow water flows.
Zonal Isolation
Zonal isolation is generally accomplished through the use of steel casing that is cemented in
place for the purpose of structure and annular isolation. The steel casing is typically designed
for the anticipated operating loads to which it will be exposed, including running, future well
construction activities (including hydraulic fracturing operations), and production operations.
The annular area outside the steel casing is cemented to support the pipe and help control
some of the loads (e.g. buckling). The cement is also used to control flow in the pipe annulus.
The cement is engineered for specific properties, including but not limited to, set up time,
compressive strength, and viscosity.
Conclusion
All of the practices represented in this paper are built on a back bone of Health, Safety,
Security, and Environmental (HSSE) management. These practices comply with regulations and
incorporate best industry practices. Hazard identification and management are core to the
sustainability of our operations. At Shell, safety is a deeply held value that is demonstrated by
our pursuit of "Goal Zero", or the goal to have a zero incident work environment.
34
-------
Table 2. Parameters analyzed during baseline water survey
Test
pH (Lab)
Alkalinity
Ihloride
Hardness
Sulfate
Total Dissolved Solids (TDS)
Total Suspended Solids (TSS)
MBAS/Surfactants
Nitrate-Nitrogen
Turbidity
Specific conductance
Barium
Calcium
Iron
Magnesium
Potassium
Sodium
Arsenic
Holding Time (with Preservative)
Immediate
14 days
! days
6 months
28 days
14 days
ays
48 hours
Immediate
gjjjjg
6 months
^H
6 months
6 months
6 months
6 months
Chromium 6 months
Lead 6 months
Mercury 28 days
Silver 6 months
Bromide
Strontium
Oil and grease
Benzene
Toluene
Ethylbenzene
Xylene
Ethylene glycol
Total coliform
E. coli
Fecal coliforms
pH (Field)
Methane (% in atmosphere; well head space)
Methane (% of LEL; well head space)
Ethane 7 days
Propane
28 days
14 days
14 days
14 days
davs
35
-------
Multi-well Pad, Tight Gas, Directional Drilling Program Protects
Aquifers
Jay Foreman
Williams Production RMT
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Engineering and regulatory controls are in place in the oil & gas industry specifically in Colorado to
ensure that completions operations for natural gas wells are conducted safely, cost effectively, and in an
environmentally sound manner. This includes making sure that hydraulic fracture treatments go into
and stay in the targeted zones where they will help stimulate oil or gas production. As a Completions
Manager for the Piceance Asset, my job is to manage the engineering and operations activity required to
turn a drilled well into one that produces natural gas safely and economically. Once a well is drilled, the
casing is cemented in place and the drilling rig leaves the site. A completions engineer then examines
the logs and specifics of the wellbore to design the perforating and stimulation procedures required to
extract natural gas from subsurface rock strata. Since the reservoir rock is so impermeable, natural gas
will not typically flow out of the reservoir at economic rates. Therefore, hydraulic fracturing (frac'ing) is
required to "stimulate" the formation to produce natural gas. Proper well construction is critical to
isolating the subterranean layers during completion and production operations. Not only is this
important for keeping the frac treatments in the reservoir but for keeping unwanted water out of the
gas zones. All aspects of drilling and completing wells are regulated by the Colorado Oil and Gas
Conservation Commission (COGCC) and, if applicable, the Bureau of Land Management (BLM). In this
document and the associated presentation given at the EPA's Well Construction Operations Workshop,
these two agencies' regulations will be collectively referred to as State regulations. I will discuss well
construction, cement design, and zonal isolation relative to our directional, multi-well pad development
program on the Western Slope of Colorado. Water wells in the area are typically less than 250' deep yet
our surface casings are set at well below this depth with cement circulated to surface. Some pads have
over 20 wells on 7-1/2' centers so directional work begins as shallow as 100'.
Conductor pipe (+/-4S') and surface casings (>10% of the TD of the permitted well depth) are set before
the production interval is drilled. The drilling mud in the annulus of the 9-5/8" diameter surface casing is
displaced with cement engineered to meet State requirements. After the cement has cured and
developed the required compressive strength, a smaller drill bit is used to drill out the bottom of the
surface casing and drill the well to the permitted depth. Once a well is drilled and conditioned, the drill
pipe is removed and 4-1/2" diameter production casing is run to the total depth (TD) of the well. At that
point, the drilling mud in the hole is circulated for several hours to "condition" the hole and prepare it
for cementing operations. Production casing of a specified grade and weight (wall thickness) is run such
that its resulting burst pressure rating is greater than the anticipated operating pressures during the
subsequent completion. Casing design is an area of engineering expertise and is addressed extensively
by State regulations. The American Petroleum Institute has developed strict specifications for the
manufacture of oilfield tubulars including casing. The design burst pressure of the casing string in a
wellbore is duly noted by the operating company's staff and contractors and the maximum allowable
treating pressure (max pressure) is established and known by all who work on the well for the
36
-------
remainder of its productive life. The pressure is not to be exceeded in order to protect the mechanical
integrity of the wellbore.
The well is cemented with specially engineered cement with known density, thickening time, fluid loss,
free water, and compressive strength. These parameters are based on bottom hole circulating and static
temperatures and must meet State requirements. It is worth noting that manufacturing and testing
specifications for completion cements are much more stringent than for cements used in construction
projects. The API has established guidelines for all aspects of the cement and cementing operations to
which operating and service companies adhere. The volume of cement pumped on the primary cement
job is determined by isolation requirements for the zones to be completed as well as applicable State
requirements for height of cement in the casing annulus. The cement is pumped down the casing and
displaced with water behind a wiper plug. Once in place, the cement cures, getting hard and building
compressive strength.
After the drilling rig leaves the well and sufficient time has passed to ensure proper cement compressive
strength development, a cement bond log (CBL) is run to evaluate the quality and height of cement fill in
the casing annulus. It is evaluated by the completions engineer to determine if the completion
operations can be safely and effectively completed on the well. If not, the completion procedure will be
altered to include remediation procedures to repair the primary cement job or to even exclude some
zones from completion. In either case, the objective is to prevent undesirable communication between
zones in the annulus.
Regulators are notified of the start of completions operations. The CBL is submitted to regulators prior
to the first frac and any deficiencies with annular fill or cement bond quality are discussed. Calibrated
pressure gauges are used to determine if pressure exists on the annulus of the production casing. If the
annular pressure limit set by regulatory agencies is exceeded, they must be notified and remediation
plans developed, sundried, approved, and executed.
The geologist and completions engineer evaluate the wireline logs to determine how the subsurface
zones will be grouped together into frac stages, perforated and hydraulically fractured. The completion
procedure is written, capturing critical information such as perforation depths, plug specifications and
depths, frac job volumes, rates, proppant concentrations and volumes. All completion procedures note
the casing specifications and the max pressure which is not to be exceeded. Treatments are engineered
to minimize waste of materials while maximizing production. This can be based on fracture simulation
programs and/or on field experience considering production results and the market costs of services and
materials. Before any perforating begins, the production casing is pressure tested to the maximum
allowable pressure. The test is recorded and submitted to the State. A well that fails the pressure test
must be remediated in accordance with plans submitted to and approved by regulatory agencies.
The day of the first frac arrives and the frac crew gets the necessary equipment rigged up to the
wellhead. Prior to any hydraulic fracturing treatment, all personnel on location are gathered for a prejob
safety meeting where each person's job responsibilities are reviewed. A headcount is taken and
emergency egress procures are reviewed. These meetings even go so far as to designate a driver and
note the closest medical facility in the unlikely event that someone is injured during the operation.
During this meeting, it is clearly stated who is to "control" the job. One service company supervisor/
engineer and one company representative will be in complete control of the location during the job.
37
-------
Max Pressure is discussed with all pump operators. Radio communication between all critical crew
members is confirmed to ensure job control is maintained at all times.
Specifically engineered pressure gauges (typically 15,000 psi working pressure) and backup gauges are
placed on the high pressure treating lines near the wellhead to monitor the treating pressure. Once
every person is in their place, the high pressure pumps and lines (typically either 10,000 or 15,000 psi
working pressure) are primed and pressure tested to the wellhead above max pressure for the job. Each
pump has its own pressure gauge and "trip out" that is checked . This safety device will automatically
shut the pump down if treating pressure exceeds the preset limit. A "global kickout" is also set on the
control computer in the treatment control van where the two people in charge of the job can safely and
comfortably monitor and control all aspects of the job. The global kickout will automatically shut down
all pumps should treating pressure reach max pressure. The treatment design is programmed into the
computer control system. All blending and pumping equipment on location can be run from the control
van by the computer system or manually overridden as job requirements or well response determines.
Low pressure/ high accuracy pressure gauges are used to monitor the pressure on the annulus of the
production casing during the treatment. Should annulus pressure rise beyond predetermined limits, the
job will be aborted immediately and the situation evaluated. Regulators will be notified of any such
event.
Once all safety systems are checked, the wellhead valves are opened and the pumping begins. A "pad"
volume greater than wellbore volume is pumped at treating rate until the treating pressure stabilizes. At
that point, the injection is stopped and an Instantaneous shut-in pressure (ISIP) reading is taken.
Calculations are made from this wellhead pressure to determine how many perforations are open and
accepting frac fluid. In addition, a formation frac gradient is calculated and compared to the anticipated
frac gradient from the prejob simulation or experience in the area. If all perforations are open, the
treatment proceeds. If not, the completion engineer may be consulted and the job can be redesigned if
necessary.
During the remainder of the treatment, wellhead treating pressure, casing annulus pressure, and
equipment performance are all closely monitored. While all are important, the most critical parameters
are the actual wellhead treating pressure vs. anticipated treating pressure and max pressure. The
pressure trends throughout the job give clues to what is occurring downhole. Dramatic increases in
pressure may indicate that zones are "screening out" and no longer accepting fluid. Dramatic decreases
in treating pressure may indicate an equipment problem either on surface or downhole. Subtle pressure
changes can give clues to chemical performance as well as fracture growth in the formation. Should
treating pressure rise and approach the Max Pressure, the two people in charge of the job will discuss
and slow injection rate to reduce pipe and perforation friction and the resulting wellhead treating
pressure. In some cases, rate changes in response to the well pressure are sufficient to allow the entire
job design to be pumped and flushed. In other cases, the well "refuses" to accept the designed
treatment and the job must be terminated early to prevent exceeding max pressure and damaging the
wellbore or surface equipment.
38
-------
Treatment placement can be validated in several ways.
The first is by monitoring the wellhead pressure during the treatment.
Secondly, the actual treating rates, pressures, and concentrations can be loaded back into the
fracture design simulator. The simulator can then be "calibrated" to match the actual treatment
data. Once a "good" match is obtained, the fracture dimensions of the simulation can be
evaluated.
Thirdly, tracers can be run in the treatment and post-frac logs run to determine if unanticipated
fracture height growth had occurred.
Lastly, microseismic monitoring from a nearby wellbore can give direct indication of the fracture
height, length, and azimuth. With this service, arrays of sensitive geophones "listen" for the
minute sounds of the hydraulic fracture growing through the reservoir rock. The location of
these events can be calculated and plotted as dots on a three dimensional display. All the events
recorded during the treatment are plotted vs. time for a representation of how the fracture
grew throughout the treatment. These costly, non-routine monitoring projects are done for
specific engineering purposes. The results can be used to further refine the fracture simulator
and enhance the geologist's and completion engineer's general knowledge of fracture growth in
the area.
Finally, the annulus pressure of the production casing is monitored during the treatment to ensure no
communication occurs up the backside.
The plug/perf/frac process is repeated until all the frac stages are completed. The well is then cleaned
out, production tubing landed, and the well is turned to production. The final analysis of the success or
failure of the fracturing process comes from the production results.
39
-------
Casing Perforating Overview
Brad Hansen
Devon Energy
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
This paper provides a general overview of casing perforating. The primary objective of a
perforating gun is to provide effective flow communication between the cased wellbore and a
productive reservoir. To achieve this, the perforating gun "punches" a pattern of perforation
through the casing and cement sheath and into the productive formation.
In the early days (1932) perforating was performed with a bullet gun. Today the bullet gun has
been almost completely replaced with the shaped-charge perforator. The shaped charge
consists of a case or container, the main explosive material, and a liner.
The perforating gun consists of four components, a conveyance for the shaped charge such as a
hollow carrier, the individual shaped charge, the detonator cord, and the detonator.
The three main explosives used in a shaped charge are RDX (Cyclotrimethylene Trinitramine),
HMX (Cyclotetramethylene Trinitramine), and HNS (Hexanitrosilbene). The main difference
between these explosives is their temperature stability. RDX is commonly used in environments
less than 330 degrees F. HMX is used for temperatures up to 400 degrees F and HNS is suited
for temperatures up to 520 degrees F. Each shaped charge generally contains between 3 and 60
grams of explosives.
A shaped charge perforating gun detonates almost instantaneously when the electrical charge
is sent from the perforating truck. The detonation creates a jet that has a velocity of 25,000 to
30,000 ft/second. The impact pressure caused by the jet is approximately 10 to 15 million psi.
This pressure overcomes the casing and formation strength and forces material radially away
from the jet axis.
Most perforating guns punch holes with diameters of 0.23" to 0.72". The typical perforating
guns have penetrations of 6" to 48". Most guns shoot from 4 to 12 shots per foot. Perforating
guns come with different pressure and temperature ratings.
The length of the actual perforation downhole is a function of the standoff of the perforating
gun from the casing. Less standoff generally means a longer perforation tunnel, while more
standoff results in a shorter perforation tunnel. Phasing is the angle difference between
successive perforations. Typically, perforating guns come with either 60, 90, 120, 180 or 0
degrees phasing. 60 degrees is a common phasing for a well that will be hydraulically fractured.
40
-------
The API RP 19B (replacing API RP 43 in September 2006) is the recognized standard for
evaluating perforator performance. However, many perforator performance tables are still
published with the older API RP 43 test data given.
The two main types of carriers are the hollow carrier and the expendable shaped charge gun.
The hollow carrier holds the shaped charges in a heavy wall tube that is sealed from wellbore
fluids and pressure. Most of the debris from shooting this type of gun is retrieved when the gun
is pulled from the well. Sometimes expendable shaped charge guns are used. This type of gun
allows a larger charge to be run than a similar OD hollow carrier gun. The charge itself is sealed
from the wellbore environment. Much of the debris is left in the well and falls into the rathole
on vertical wells.
Wireline pressure control equipment is run above the wellhead so that the perforating gun can
be run in and out of the well when the well has pressure on it. This pressure equipment is
commonly known as a lubricator. Lubricators are sized by ID and working pressure. This
equipment consists of a wellhead connection, the wireline blowout preventer (BOP), the riser
and the control head. It may also have full opening valves, pump in subs, tool catchers and
other equipment in the run. The control head is the uppermost point of the lubricator system
where the wireline enters. Well pressure is controlled with packing, pack-off rubbers, grease
injection or a combination of all three. The riser section is used to allow the full wireline tool
string to be raised above the wellhead valve before and after the operations.
Depth control for perforating is usually accomplished with a gamma ray/casing collar locator
log. Short joints are also run in the production casing to assist in the correlation. The distance
from the top shot to the casing collar locator is measured before running the perforating
system into the wellbore to ensure the perforations are placed where they were intended.
Gamma ray logs measure the natural radioactivity of the formations. The gamma ray log can be
recorded in open holes as well as cased holes which make it an ideal log for correlating
different gamma ray signatures between wells. Nearly all gamma radiation encountered in the
earth is emitted by the radioactive potassium isotope (atomic weight 40) and by the radioactive
elements of the uranium and thorium series.
Some horizontal completions today are completed with an openhole system below an
intermediate casing string. These wells have external casing packers that form a seal between
the production casing and the formation. They also have hydraulic or ball drop actuated sliding
sleeves to open successive sleeves to perform multiple fracture stimulations without the need
to rig up wireline and set plugs and perforate new intervals. Perforating is not required to
provide effective communication between the cased borehole and the productive formation
with these types of systems.
41
-------
Summary and Abstracts from Theme 2: Fracture Design and
Stimulation
42
-------
Summary of Presentations for Theme 2: Fracture Design and Stimulation
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA. Any mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
Technical Presentations
The first set of technical presentations in this theme addressed geologic assessment with regard
to fracture design and stimulation.
Terry Engelder, Pennsylvania State University, raised questions regarding the possibilities of fluid
and gas migration or leaks into ground water through natural interconnected deep fracture
systems in and around the Marcellus Shale. He assessed many parameters in his consideration of
whether the physics of fluid flow at depths of 6,000 to 8,000 feet indicates the possibility of
leakage of fracture fluid between the Marcellus and ground water. His conclusion is that the
possibility of fracture fluid leakage from the deep Marcellus to the water table is remote. Dr.
Engelder emphasized that without a pressure drive, or hydraulic head, there can be no leakage
(fluid flow or leakage occurs only when fluids move from higher pressure zones to lower pressure
zones).
John Williams, U.S. Geological Survey (USGS), discussed the use of well records and geophysical
logs to determine the presence of fresh water, saltwater, and gas stratigraphically above the
Marcellus Shale. He described three databases: USGS's National Water Information System
(NWIS), the New York Department of Environmental Conservation's Water Well Database, and
the New York State Museum's Empire State Oil and Gas Information System (ESOGIS). Mr.
Williams discussed considerations for future drilling that would greatly expand existing
information and better support assessments of fluid movement and hydraulic fracturing
activities. These considerations include consistent characterization and complete reporting of
fresh water, saltwater, and gas occurrence; measurement of specific conductance of water
produced during drilling; geophysical logging prior to surface-casing installation; and compilation
and integration of information from gas and water wells.
The second set of technical presentations addressed fracture propagation.
Tim Beard, Chesapeake Energy, discussed fracture design in horizontal wells in shale gas plays,
which is a relatively new application of HF. The goal of HF in shales is to maximize the "stimulated
reservoir volume" (SRV), i.e., maximizing the area of reservoir rock that is fractured, filled with
proppant, and connected to the wellbore to enable maximum hydrocarbon production. Local and
regional in-situ stress data and reservoir properties are required information for developing a
fracture design. Typically, drilling occurs perpendicular to the maximum principal stress in the
targeted reservoir. Many diagnostic tools (e.g., microseismic monitoring, tiltmeters, etc.) are used
to evaluate downhole stimulation. Failure to appropriately design a given HF treatment can result
in poor well stimulation and lower production potential according to Mr. Beard.
43
-------
David Cramer, ConocoPhillips, discussed fracture propagation in shallow reservoirs, using an
example case of a water-flooded oil reservoir in south Texas. Conditions are favorable for
propagating horizontal fractures in shallow reservoirs. Treatment pressure response allows for
the estimation of fracture geometry. Mr. Cramer emphasized the economic incentive for limiting
fracture propagation within the target zones and described methods for controlling fracture
growth.
Hal Macartney, Pioneer Natural Resources USA, Inc., discussed HF in the development of coal
bed methane in the Raton Basin, Colorado. This field contains approximately 2,400 wells and
produces 200 million cubic feet of gas per day from coal beds. Many of these wells have been
fractured, and Mr. Macartney shared his belief that there has been no evidence of contamination
to USDWs resulting from these operations. Mr. Macartney attributed this success to horizontal
fracture propagation with very little height growth as seen by direct pressure measurement in
open zones above the fractured coal beds, the lack of natural fractures that extend out of the
coalbed target zone, sound cement and casing design, and close monitoring of fracture pressures
and fluid volumes.
The third set of technical presentations in this theme addressed monitoring.
Mike Eberhard, Halliburton Energy Services, discussed the monitoring, calibration, and oversight
activities that must take place during well construction, as well as before, during, and after actual
HF operations. Mr. Eberhard presented images of HF sites and monitoring equipment, and
described the monitoring techniques used by Halliburton. He emphasized the importance of
proper well construction, as well as knowledge of rock mechanical properties and other fluid and
geological conditions.
Patrick Handren, Denbury Resources, described a microseismic evaluation of wells in the Barnett
Shale. He provided background information on microseismic monitoring techniques and
presented case studies of two wells in the Barnett. Data from the microseismic surveys were used
to calculate average SRV, fracture height, and the area covered by the fracture network. These
data were then used to partially predict fluid movement from fracturing a third well. The results
of this study indicate that increased well density increases the complexity of fluid movement
according to Mr. Handren. In addition, while lateral fluid movement is not limited to the acreage
covered by the calculated stimulated reservoir volume, Mr. Handren stated these techniques do
allow for some estimate and prediction of fluid movement.
Norman Warpinski, Pinnacle-A Halliburton Service, described data and information showing that
layered sedimentary sequences can restrict vertical fracture growth. This information included
mineback studies, core observations, microseismic mapping, and tiltmeter data. Mr. Warpinski
described how adjacent zones of significantly different stress conditions can limit the vertical
growth of shallow hydraulic fractures. Mr. Warpinski concluded that hydraulic fractures
consistently remain thousands of feet below ground water aquifers.
44
-------
The final set of technical presentations in this theme addressed verifying zonal isolation.
Ahmed Abou-Sayed, Advantek International, discussed the complexities and quantitative
uncertainties associated with fracture stimulation and design. Key uncertainties include created
fracture shape and interaction with layering, faults, and other fractures. Dr. Abou-Sayed
emphasized the importance of the stress field azimuth related to well orientation and fracture
extent and conditions. He noted the utility of novel pressure transient test interpretations for
fracture identification and concluded that multiple fractures in single wells must be well designed
and require close monitoring.
Daniel Soeder, U.S. Department of Energy, National Energy Technology Laboratory, described a
proposed field experiment using tracers in HF fluid. A tracer study could address two key HF
issues, perception of risk and lack of field data. This type of study also could potentially provide
information for other studies investigating geochemistry and fluid fate and transport. Mr. Soeder
described the properties of environmental and introduced tracers and described the design and
proposed locations for the proposed experiment.
Scott Cline (unaffiliated) discussed the mechanisms that affect stimulation water retention in gas-
bearing shales. While fluid leak off into the fracture face and clay adsorption and swelling may
account for some fluid retention, Dr. Cline believes that retention is primarily affected by capillary
forces and stranding in narrow fracture branches; proppant packs and gravity may also affect
fluid retention. Dr. Cline concluded that these mechanisms, combined with other aspects of HF
operations and the local geology, indicate that there is a low risk of ground water contamination
by HF fluids.
Summary of Discussions Following Theme 2: Fracture Design and
Stimulation Presentations
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA. Any mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
Horizontal fractures. A participant asked about horizontal fracture propagation. A presenter
stated that fractures propagate horizontally due to the orientation of the stresses in the rock; the
fact that fractures tend to stop moving vertically at shallow depths and move horizontally is
because minimum stress is in the horizontal direction in shallow geology. A participant indicated
that in general, shallower rock tends to be more plastic, while deeper rock is more brittle and
easier to fracture. However, characteristics of shales do vary from formation to formation.
Participants stated that both stress and rock properties affect the orientation of fractures, though
participants disagreed on the relative importance of these parameters.
Natural hydraulic fractures. A participant asked for clarification on how natural hydraulic
fractures occur. The presenters explained that a natural hydraulic fracture forms from gas or
water. During maturation, organic matter is converted from fatty acids and lipids to kerogens,
45
-------
and further maturation leads to the creation of oil and gas. This reaction increases volume, which
increases the pressure in the rock. If this elevated pressure exceeds the least stress in the rock
(which tends to be horizontal), cracks develop in the rock. A presenter stated that these cracks
will generally be densest near the source rock (the oil and gas reservoir) and noted that
compressibility of gas makes gas more effective at driving cracks than water. In addition, the
presenters stated that joints do not interconnect as they propagate outward, which gives the
rock a low bulk permeability. Participants stated that the degree of mineralization will control
conductivity more than the extent of fractures.
Lack of evidence of contamination. Several participants noted that there has been no definitive
evidence of contamination in water wells from HF, based on monitoring results before and after
drilling.
Availability ofdownhole data. A participant asked whether any downhole data are available for
rock properties in the Marcellus Shale. A presenter responded that while most of this information
is proprietary, some available horizontal Fullbore Formation Microlmager (FMI) results provide
subsurface joint information in the Marcellus. According to the presenter, data from the
Haynesville Shale suggest a different stress field orientation at the time of joint propagation.
Depth of water wells. A participant asked about the quality of data on water well depths. Another
participant responded that, in general, water well depth data are very good, while data on depth
to water may be less reliable. The presenter indicated that well depths are generally obtained
from drilling records.
Chemical indicators of contamination. A participant asked about the chemical most likely to
indicate ground water contamination from a HF treatment. Participants responded that the
answer would depend on the fluid system in use, as well as local hydrogeologic processes. In
most cases, however, some participants suggested that the most useful indicator might be
potassium chloride. Potassium chloride may be found in the base fluid used in fracturing
treatments. According to the participants, potassium chloride generally is not present in shallow
ground water so it can could be used as an indicator. A few participants expressed skepticism that
unique tracers could be assigned to drilling companies or individual HF jobs.
Microseismic surveys. Many of the data points in the tiltmeter study presented by Norm
Warpinski represent relatively shallow (< 2,000 ft) layers because these are areas that receive
more monitoring. A participant noted that microseismic data do not indicate which areas have
received proppant. Another participant asked about the limitations of microseismic monitoring in
very shallow layers. Microseismic arrays are able to capture activity that takes place above the
array (usually placed at 400-1,000 ft). However, most of the time, the fracture stops in the
middle of the array. A participant claimed that microseismic monitoring would be able to identify
a shallow fault, but only if it were activated by the fracturing.
Participants also discussed the uncertainty in microseismic data for measuring fracture extent
(length and height) and orientation. This depends on several factors, but in ideal settings the
vertical and horizontal uncertainties can be as small as 20-25 ft for length and 10-15 ft for
46
-------
height, according to one participant. The largest uncertainty is the angle or orientation of the
fracture (~3-4°). Participants also discussed shear and tensile fracturing. While some hydraulic
fractures are tensile fractures, a participant indicated that microseismic monitoring detects
mainly shear events. One participant noted that these shear events may not always begin closest
to the wellbore and move outward. Other participants noted that all producing reservoirs are
naturally microseismically active. However, a few participants stated that microseismic
monitoring provides an accurate picture of the stress changes in the formation.
A participant asked about using microseismic surveys to plan well placement. One participant
described a microseismic program that ran for two years, spanning 20 wells and 30,000 acres.
The participant noted that well interaction does not seem to substantially affect production. In
this case, wells that were "watered out" (where water production dominates over gas
production) eventually did as well as or better than before in terms of gas production. The
participant concluded that the fracture interaction increases the fracture matrix feeding into all
of the wells.
Fracture height growth. A participant asked if the Tully Limestone acts as an upper barrier to
fractures in the Marcellus Shale. Another participant responded that generally this is not the
case, because the Tully is fairly thin. A participant suggested that the Tully might act as a barrier if
it were thicker. Another participant stated that the Onondaga Formation, which underlies the
Marcellus, does act as a barrier. A participant asked about the relationship between fracture
height growth and the low rate of return of fracturing fluid in the Marcellus Shale. Participants
suggested that low water recovery is common (and preferred) in the Marcellus due to greater
fracture complexity; the volume of the stimulated reservoir increases with increasing fracture
complexity, rather than length of the fractures, allowing a greater volume of gas to move into the
fractures. Participants indicated that the water recovery rate is affected by capillary trapping and
the thickness of the reservoir formation. A participant noted that fluid movement must be driven
by a pressure differential and is impeded by impermeable zones in the subsurface.
Public opinion and HF tracer studies. A participant asked if the proposed tracer test would be
likely to satisfy all concerned parties. The presenter clarified that the tracer study should be able
to satisfy all parties with regard to accuracy of results although those results may only apply to
the study location. The presenter noted that one study in one location cannot be used to draw
conclusions about other areas, but this study would be a first step. A participant noted that 20
years of EPA and other studies have been unable to alleviate the public's mistrust of Class I
Underground Injection Control (UIC) wells. Another participant stated that some people in New
York State are looking to the EPA's current study as a definitive determination on the safety of
HF.
Pumping tests. A participant called attention to a graph used in Mr. Cline's presentation (on slide
10) that showed pumping pressure over time during a HF treatment. The participant noted that,
as shown in the graph, a formation integrity test (FIT) is not the same as a leakoff test (LOT). The
participant added that fracture closure pressure (FCP) is the closest value to the least principal
stress.
47
-------
Fracture complexity. A participant asked about how fracture models account for fracture
complexity, since fracture models appear to describe a single fracture that propagates outward. A
participant stated that while multiple or branching fractures do occur, points of weakness in the
formation control which fractures will dominate in growth. According to the participant, all of the
fracture branches will grow simultaneously, and longer fractures require less pressure to
propagate further. One participant added that fracture activity in naturally fractured reservoirs is
different from fracture activity in homogeneous, less fractured formations.
Shallow fracturing. A participant asked about the shallowest fracturing in the Marcellus Shale.
There was one HF test in Otsego County, NY at 2,000 ft deep. However, this was not a large
volume slickwater HF job, and activities were discontinued after the test. Participants described
current and past shallow HF operations in other parts of the country. These locations included
Alaska, where fracturing for a DIG Class I well took place at the base of permafrost (approximately
2,000 ft); the Huron Shale near the West Virginia/Kentucky border; and Oak Ridge National Lab in
Tennessee, where, according to one participant, uncased monitoring wells became a conduit for
fluid movement and led to contamination. A participant added his understanding that New York
is planning to limit large volume HF operations to a depth of 2,000 ft or deeper (or 1,000 ft below
ground water supplies, where applicable).
48
-------
Abstracts for Theme 2: Fracture Design and Stimulation
Abstracts were submitted to U.S. EPA by the presenters for use in this proceedings document.
Not all presenters submitted abstracts of their presentations.
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA. Any mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
49
-------
The Distribution of Natural Fractures above a Gas Shale:
Questions about Whether Deep Fracture Fluid Leaks into
Groundwater Outside the Realm of Faulty Borehole
Construction
Terry Engelder
The Pennsylvania State University
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Extended Abstract
One concern about production of shale gas is the possibility of hydraulic fracture fluid leaking
upward along an interconnected network of fractures to contaminate groundwater. In the
Appalachian Basin, isotopic studies of stray methane provide unambiguous evidence for
leakage from gas storage fields through the Marcellus gas shale and into groundwater. The
question is whether this is a case of faulty borehole construction where methane is leaking
along poor cement jobs outside of casing or whether this is a case of methane traveling toward
the water table along natural pathways, most likely consisting of unhealed faults or fractures.
Elsewhere in the Marcellus gas fields, preliminary data isotopic studies by the Pennsylvania
Department of Environmental Protection (PA DEP) indicate that thermogenic methane comes
from Upper Devonian sands that are not immediately charged by gas from the Marcellus. While
migration of natural gas is common, particularly at shallow depths, the migration of deep
fracture fluid remains undetected in the Appalachian Basin. The question is whether the physics
of fluid flow at depths of 6000' to 8000' permits leakage of fracture fluid between the Marcellus
and groundwater with a probability that should concern the public. A back-of-the-envelope
analysis suggests that the physics of the Earth reduces the probably of leakage to a level where
the risk should be acceptable to a nation that consumes natural gas at the rate of nearly 25
trillion cubic feet (Tcf) per year.
If leakage occurs outside the realm of faulty borehole construction, unhealed fractures are the
most likely pathway. The plausibility of rapid leakage along fractures depends on evidence for
the pervasive development of interconnected fractures between gas shales (> 6,000 feet) and
fresh groundwater (< 1,000 feet). Although continuous fracture imaging in the borehole might
aid in sorting out whether fractures are interconnected from depth to the surface, borehole
coverage may be insufficient for a conclusive answer. This leaves outcrop mapping as the other
means of direct observation to resolve the extent of fracture interconnectivity.
For this discussion, natural fractures fall into either of two failure classes: shear failure leading
to faults and tensile failure leading to macroscopic cracks called joints. Faults grow during direct
shear failure under unusually high stress or grow as reactivated joints and bedding planes.
Joints propagate within a spectrum failure conditions depending on their crack driving stress.
If propagation takes place at depth under high fluid pressure working against crack normal
50
-------
compressive stress, the joint is a natural hydraulic fracture. If propagation takes place in the
near surface under relaxation stresses accompanying exhumation, the joint is an exhumation-
related fracture. Natural hydraulic fractures and exhumation-related fractures are end
members of a spectrum that may include fold-related jointing where both pore pressure and
bed-parallel stretching lead to an effective stress state favoring propagation.
Outcrop observation is only effective to the extent that the operator has a strategy for
distinguishing between deep-formed fractures and exhumation-related surface fractures. The
most common fractures in a gas shale are the natural hydraulic fractures which occur in a
plume emanating from the gas shale but rapidly dissipating above gas shale. Exhumation-
related fractures have a much different morphology and are easily distinguished from NHF. A
plume-like distribution of joints above gas shales of the Appalachian Basin is consistent with
fluid-drive propagation mechanism where high pressure fluid bubbles from gas shales as a
consequence of thermal maturation but rapidly looses pressure as it migrates up section. The
implication is that fractures driven by fluid pressure are not uniformly distributed up to the
surface but rather are concentrated near the top of gas shale.
Distribution of Natural Fractures
It is commonly assumed that if a rock contains fractures (i.e., faults, fluid-driven joints, and
exhumation-related joints), they are a natural pathway for contamination of ground water.
Outcrops are often densely populated with fractures and it is assumed that rocks in the
subsurface look the same way throughout the 6000' to 8000' of overburden above a gas shale
such as the Marcellus. In fact, many joints in outcrop are exhumation-related fractures that
propagate in the near surface and are not found at significant depths, (> 100s of feet). Outcrops
over the Marcellus of the Appalachian Plateau portion of the Appalachian Basin consist of
clastic and carbonate rocks varying in age from Devonian to Permian. Faults are exceedingly
rare on scales greater than the size of tectonic wedges, particularly above the Frasnian section.
Tectonic wedges are most common in large-channel sandstones where bedding slip can occur
on crossbeds. Fluid-driven joints are most common in gas shales but in sections overlying these
gas shales, they lack the requisite interconnectivity to be effective conduits even in the
presence of a pressure drive. Without an interconnected pathway of joints, the physical
principle governing the rate of leakage between the Marcellus and groundwater is the equation
for fluid flow in porous media, Darcy's Law.
Darcy's Law
The rate of fracture fluid leaking into ground water by flow through the overburden between
the Marcellus and near surface rocks is understood using Darcy's Law. Although flow along
natural pathways including joints and faults may be more appropriately represented by parallel
plate flow, lack of interconnectivity of these joints and faults means that Darcy's Law is the
better model for flow in the bulk rock. Fluid flow (Q) in a porous media can occur only if a
pressure drop (Pa-Pb) 0-e-, a differential hydraulic head) develops between two points with
the entrance point (i.e., fracture fluid in the Marcellus) being at a higher hydraulic head than
the exit point (i.e., fresh groundwater). The rate of fluid flow is governed by the magnitude of
the pressure drop. The relationship between rate of flow and pressure drop is expressed in an
51
-------
equation with four variables including the viscosity of the fluid (fj), the permeability of the rock
(k), the length of the flow path (L), and the cross section of the flow (A). Of course, the rate of
flow approaches zero if the permeability, cross section, and pressure drive become very small
or viscosity and flow path length become very large.
-kA.
juL
Path Length (L)
The least ambiguous variable in Darcy's Law is the length of the flow path. All else being equal,
the rate of leakage of fracture fluid from deep (6000' to 8000') gas shales of the Appalachian
Basin is as much as four to five times less than leakage of fracture fluid from conventional gas
reservoirs stimulated at a depth of 1500', for example, in the Pavilion gas field of Wyoming.
Pressure Drive (Pa - Pb)
Leakage of fracture fluid will take place only if pressure drive is present and sustained for a
period long enough to drive fluid from the Marcellus to groundwater. There are three major
sources for a pressure drive: pressure during wellbore stimulation, a topographic pressure
drive, and maturation-related abnormal pressure. A pressure drive is the most critical part of
Darcy's Law in terms of risk to groundwater.
Maturation-Related Pressure Drive (Pa - Pb)
While it might be argued that overpressure gas also creates a pressure drive from the Marcellus
to groundwater, this pressure drive was incapable of draining the Marcellus gas shale over
periods of as much as 260 million years ago (Ma). If fracture fluid is injected into the gas and
maintained at gas pressure, the gas and water would separate with the gas making its way to
the top of the pressurized column. Theoretically the top of the pressurized gas-water could
drive its way to groundwater. Long before the column with gas on top got to groundwater, the
column would have broken into a hydrostatically pressurized regime. Such break through would
immediately relax the pressure drive and flow would stop long before fracture fluid was driven
upward to place groundwater at risk. The probability of a sustained maturation-related
pressure drive causing groundwater contamination is very, very low.
Regional Flow (L) and Hydrodynamic Pressure Drive (Pa - Pb)
One natural pressure drive arises from topographically-driven hydrodynamic flow.
Hydrodynamic flow is driven by the pressure drop between groundwater under topographic
highs and ground water under topographic lows. The depth of penetration for hydrodynamic
flow is largely governed the geometry of the most permeable units but an important secondary
governor is the lateral distance between source (topographic high) and sink (topographic low)
and the vertical distance of flow as governed by topography. The largest volume of
underground flow is short circuited by local topography where depth of penetration is less than
the topography. Some groundwater is driven deeper in the section and flows further out into
the basin from topographic highs. In this latter case, the volume of flow is less and the time of
52
-------
flow between source and sink is commensurately longer. In the Appalachian Basin, penetration
to the Marcellus at 7000' is nearly an order of magnitude greater than the local topography.
Modeling suggests that penetration to 7000' from a 300' topography drive has a time constant
of 100,000 years or longer. This means that the probability of regional flow leading to a leakage
up and into groundwater is remote on time scales that really matter to the EPA debate about
hydraulic fracturing. Because there is no indication from the Appalachian Basin that such long
wave-length flow paths upset the density stratification of the basin, the probability of a
topographic drive causing leakage between Marcellus and groundwater is again very, very low.
Density Stratification Reduces Effectiveness of Natural Pressure Drives (Pa -
Within the Appalachian Basin, groundwater is stratified by density. Freshwater is found from
the top of the water table to depths of as much as 1000 feet. Below the freshwater layer,
groundwater becomes progressively more saline with waters in the vicinity of the Marcellus
approaching oil field brines. This high salinity may have developed by very long term (1-10
million years) groundwater circulation down section to the Silurian Salina Formation which is
salt rich. In a one-dimensional flow model density stratification is stable without the possibility
of a pressure drive to upset this stability. Flow between fracture fluid in the Marcellus and fresh
groundwater would upset this density stratification. In the hundreds of thousands of water
wells drilled in the state of Pennsylvania there is no evidence of fresh water wells gradually
becoming saline, the only sign that a pressure drive associated with 50 years of hydraulic
fracturing in PA has upset the regional density stratification. Density stratification indicates that
rate of regional flow carrying fracture fluid to groundwater is very, very low.
Pressure Drive Reduction Upon Flowback (Pa - Pb)
Flowback immediately following well stimulation relieves any pressure drive that was
momentarily developed between fracture fluid that is injected into the deep Marcellus and the
layer of fresh groundwater at depths of less than 1000 feet. Without a pressure drive there can
be no direct leakage between fracture fluid in the Marcellus and groundwater several thousand
feet above. Any man-made pressure drive during hydraulic fracture stimulation is not held in
place long enough to put groundwater at risk.
Distribution of Stimulated Fractures (k)
Recent studies indicate that stimulation may extend laterally as much as 2000 feet from the
borehole (Mayerhofer - Pinnacle) and as much as 1000 feet above the borehole (Fisher -
Pinnacle). In the Marcellus this leaves as much as 6000 feet between unstimulated rock and
groundwater. This thickness of rock would be exceedingly difficult for fracture fluid to
penetrate without large and sustained pressure drive which, of course, is lost with the onset of
flowback after maximum of 1000 feet of penetration.
The Inward Pressure Drive by Gas Depletion (Pa - Pb}
Once gas production starts, reservoir pressure drops. If a pressure drive develops subsequent
to the initiation of production, the pressure drive will cause flow from the rock formation and
53
-------
into the Marcellus reservoir. For fracture fluid at the extremes of stimulated fractures, flow is
back along the fractures and into the production tubing.
Thermal Maturation (k)
The Marcellus reached maximum thermal maturation about the Middle Permian (260 Ma). At
that time, generation was sufficiently rapid to cause the development of overpressure (> 0.7
psi/ft) and in some cases the high pressures drove natural hydraulic fractures. Exhumation and
thermal cooling commenced with the onset of rifting in the Triassic (perhaps 220 Ma). Despite
exhumation, much of the northern Appalachian Basin still holds overpressured gas. Even with a
network of natural fractures, the Marcellus has not leaked a sufficient quantity of gas to reach a
hydrostatic pressure. In fact, joints in shale are so planar that, when pressed together under
confining stress, these joints fail to provide a sufficiently larger permeability over the matrix
permeability to permit economic gas production without propping using sand of 100 mesh or
less. Apparently, gas pressures in natural hydraulic fractures don't prop these joints sufficiently
to enhance bulk permeability despite the presence of overpressures.
Permeability of Black Shale (k)
Black shales including the Marcellus are seal-quality rocks with a permeability of 100 to 500
nanodarcies. A porous sandstone can have a permeability of a darcy (9.8 x 10"13 m2). Because
permeability is found in the numerator of Darcy's law, lower permeabilities lead to reduced
flow rates which means that shale matrix will not serve as a path for leakage of fracture fluid.
Permeability of Joints (k)
Laboratory experiments show that joints unfilled by any mineralization are permeable relative
to rock matrix. However, in order to affect the bulk permeability of the rock, these open joints
have to interconnect. Otherwise, the bulk permeability of the rock is close to the matrix
permeability. In the Appalachian Basin interconnected joints are common in gas shales like the
Marcellus. The population of interconnected joints trails off with distance above gas shale and
groundwater. While the presence of joints allow the possibility of fracture fluid leakage as long
as the joints are propped open, lack of connectivity reduces the bulk permeability of the
overburden to that of the intact rock.
Permeability of Faults (k)
In oil basins, faults are some of the most effective seal rocks, much less permeable than matrix
sandstone. This is particularly true for faults that cut shales where the clay smear mechanism
may render a fault gouge that is less permeable than the progenitor shale. Faults leak after
earthquake-related slip but in an area that is not prone to earthquakes as is the case for the
Appalachian Basin, faults are rarely open conduits. Flexural slip folding causes bedding slip
surfaces that are coated with fibers known as slickelites which have virtually no permeability.
Viscosity of Fluid within the Black Shale
Black shales including the Marcellus are very impermeable rocks relative to many other
lithologies. Commonly, the permeability is on the order of 100 to 500 nanodarcies. This means
54
-------
that a low viscosity fluid such as natural gas is held in place for considerable lengths of time.
Because viscosity is found in the denominator of Darcy's law, increasing the viscosity of actively
moving fluid would reduce the flow rate. Injection of a high viscosity hydraulic fracture water
with additives just makes a good seal all that more effective. A rock that has not leaked natural
gas in geological time is unlikely to leak a more viscous fracture fluid on an anthropomorphic
time scale.
Capillary Forces (k)
Capillary forces are inversely proportional to the size of pore throats in a water-wet shale. To
the extent that fracture fluid converts gas shale to a water-wet rock, capillary forces may
become important in reducing leakage from gas shale.
Unknown Effects
In a fully developed section, enough water is injected to cause a regional extension of 1% in the
direction of the maximum horizontal stress at the depth of the Marcellus. A strain discontinuity
will develop at the top of the layer of injection. The effect of this strain discontinuity of regional
permeability patterns is unknown but a strain discontinuity seems unlikely to affect the nature
of the section over the stimulated zone.
Conclusion
I have identified at least 16 parameters that govern potential leakage of fracture fluid between
the Marcellus and groundwater. Most parameters favor the protection of groundwater. In
assessing risk to each parameter, the overall risk is the product of each multiplied serially. The
sixteen parameters together make a powerful case that leakage of fracture fluid from the deep
Marcellus to the water table is remote. This conclusion is consistent with the 2009
Groundwater Protection Council study.
55
-------
Evaluation of Well Records and Geophysical Logs for
Determining the Presence of Freshwater, Saltwater, and Gas
above the Marcellus Shale, South-Central New York
John H.Williams
U. S. Geological Survey
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Records of water wells in the National Water Information System (NWIS) and records and
geophysical logs of gas wells in the Empire State Oil and Gas Information System (ESOGIS) were
evaluated to provide a preliminary determination of the presence of freshwater, saltwater, and
gas above the Marcellus Shale in south-central New York. This work expands the geographic
extent of the well-record evaluation of Williams (2011) that included Chemung, Tioga, and
Broome Counties to include Cortland, Chenango, Otsego, and Delaware Counties (fig. 1). In
total, these counties form the core of the Marcellus and Utica shale-gas fairway in New York.
Water-well records stored in the NWIS, which is maintained by the New York Water Science
Center of the U. S. Geological Survey (USGS), were retrieved for the 7-county study area online
from http://waterdata.usgs.gov/nwis/inventory. The NWIS contains records for nearly 4,000
water wells in the study area. Many of the water wells recorded in the NWIS and almost all of
the 65 sites at which saltwater and (or) gas zones were penetrated were inventoried as part of
glacial-drift aquifer investigations in the 1960s and 1970s (Randall, 1972; MacNish and Randall,
1982; Randall and others, 1988). These investigations were focused on the glaciated valleys of
the Susquehanna and Chemung Rivers in New York. The presence of saltwater in water wells
was reported by drillers or well owners based on taste tests, or was determined by water-
quality analyses that indicated a chloride concentration of greater than 250 mg/L. The presence
of gas was reported by drillers or well owners or was observed during field inventory.
Gas-well records and geophysical logs stored in the ESOGIS, which is maintained by the
Reservoir Characterization Group at the New York State Museum (NYSM), were retrieved for
the 7-county study area online from http://esogis.nysm.nysed.gov/. The ESOGIS contains
records for about 600 gas wells in the study area. The gas-well records in the ESOGIS are for
single- and multiple-well sites, wells whose confidential status had not expired (typically 2
years), and permitted but uncompleted wells. The density of the gas-well distribution
generally decreases in a northeast direction across the study area from more than 200 wells in
Chemung County to less than 20 wells in Otsego County. Because the formations above the
Marcellus Shale generally have not been the focus of gas exploration, many of the gas-well
records contain little or no information on the stratigraphic interval of interest. Penetration of
water and (or) gas zones above the Marcellus Shale was reported for 112 gas-well sites. Water
flows were reported by gas-well drillers as freshwater or saltwater presumably based on taste
tests. Water flows commonly were rated by the gas-well drillers in inches of the stream
56
-------
discharging from an open pipe into a mud pit while drilling with air. Reportedly, a 1-inch stream
roughly equates to a flow of 10 to 20 gal/min, and a 2-inch stream roughly equates to a flow of
40 to 50 gal/min. Gas-flow rates generally were not quantified; those that were rated were
reported in MCF (1,000 cubic feet), which presumably equates to the flow rate per day.
Geophysical logs for 150 gas wells in the study area are stored in the ESOGIS in Log Ascii
Standard (LAS) format. No geophysical logs were collected prior to the installation of steel
surface casing, which typically is set to a depth of 500 to 1000 feet below land surface. Most
of the log suites did include gamma measurements from the bottom of the well to near land
surface. Neutron porosity and density logs were commonly collected along with the gamma
logs. Only a few temperature, focused resistivity, or induction logs and no fluid resistivity logs
were available for the interval above the Marcellus Shale. Again, because the formations above
the Marcellus Shale generally have not been the focus of gas exploration, limited geophysical
logging has been completed on this stratigraphic interval.
The well records and geophysical logs were reviewed to obtain information on well
completions, geologic formations penetrated by the wells, and the presence of freshwater,
saltwater, and gas above the Marcellus Shale in the study area. The spatial and stratigraphic
distributions of freshwater, saltwater, and gas above the Marcellus Shale were investigated. To
aid in the evaluation, Geographic Information System (GIS) coverages and histograms of the
well data were created with ESRI ArcGIS software; and geophysical log composites were
created with WellCad software.
The evaluation of the well records and geophysical logs provide a preliminary but incomplete
determination of freshwater, saltwater, and gas above the Marcellus Shale in the study area
(Figure 1). The evaluation indicates that freshwater aquifer zones are log-normally distributed
with depth and that freshwater circulates to a greater depth in the uplands than in the valleys.
The base of the freshwater aquifer appears to be about 850 ft below land surface in upland
settings but only about 300 ft below land surface in valley settings. At depths greater than 300
ft in valley settings, groundwater in the Upper Devonian bedrock, and in a few areas in the
glacial drift, is salty. Williams and others (1998) found saltwater at similar depths in the glacial
drift and Upper Devonian bedrock during an inventory of water wells in the glaciated valleys of
Bradford, Tioga, and Potter Counties across the border in Pennsylvania. Water-quality analyses
from these wells indicated that the shallow saltwater is characterized by elevated
concentrations of chloride, barium, strontium, and radium and low concentrations of sulfate.
Gas is present locally in the glacial drift, Upper Devonian bedrock, Tully Limestone, and
Hamilton Group above the Marcellus Shale. The frequency of gas zones in the Upper Devonian
bedrock generally increases with depth. The highest rates of gas flow above the Marcellus Shale
appear to be associated with the Tully Limestone. Pockets of gas are locally present above the
base of the freshwater aquifer with gas and freshwater occurring in close vertical proximity
(Figure 2). Reported gas shows from targeted zones below the Marcellus Shale were correlated
with distinct cooling anomalies on temperature logs suggesting that such logs could be an
effective tool for delineating gas above the Marcellus if available.
57
-------
Consistent and complete reporting of freshwater, saltwater, and gas during the drilling of future
Marcellus and Utica shale-gas wells would greatly expand existing information. Field
measurement of specific conductance of water produced during drilling would enhance the
quantitative value of the gas-well records. Consideration should be given for a two-phase open-
hole logging program that includes collection of caliper, induction, fluid resistivity, and
temperature logs in addition to nuclear logs for the depth interval above the Marcellus. In
such a program, the uppermost part of the well would need to be logged prior to the
installation of surface casing. Compilation and integration of information from gas wells and
from water wells that are inventoried and water-quality sampled during gas development and
ongoing county- and basin-wide programs (Hetcher-Aguila, 2005; Hetcher-Aguila and Eckhardt,
2006; Nystrom, 2007; and Nystrom, 2008) would provide an important database for
understanding and protecting the freshwater aquifers in the Marcellus and Utica shale-gas
fairway.
References Cited
Hetcher-Aguila, K.K., 2005, Ground-water quality in the Chemung River Basin, New York, 2003:
U.S. Geological Survey Open-File Report 2004-1329, 19 p.
Hetcher-Aguila, K.K., and Eckhardt, D.A.V., 2006, Ground-water quality in the upper
Susquehanna River Basin, New York, 2004-05: U.S. Geological Survey Open-File Report
2006-1161, 21 p.
MacNish, R.D., and Randall, A. D., 1982, Stratified-drift aquifers in the Susquehanna River basin,
New York: New York State Department of Environmental Conservation Bulletin 75, 68 p.
Nystrom, E.A., 2007, Ground-water quality in the Delaware River Basin, New York, 2001 and
2005-06: U.S. Geological Survey Open-File Report 2007-1098, 37 p.
Nystrom, E.A., 2008, Ground-water quality in the Mohawk River Basin, New York, 2006: U.S.
Geological Survey Open-File Report 2008-1086, 33 p.
Randall, A.D., 1972, Records of wells and test borings in the Susquehanna River basin, New
York: New York State Department of Environmental Conservation Bulletin 69, 92 p.
Randall, A.D., Snavely, D.S., Holecek, T.P., and Waller, R.M., 1988, Alternate sources of large
seasonal ground-water supplies in the headwaters of the Susquehanna River basin, New
York: U. S. Geological Survey Water-Resources Investigations Report 85-4127, 121 p.
Williams, J. H., Taylor, L. E., and Low, D. J., 1998, Hydrogeology and groundwater quality of the
glaciated valleys of Bradford, Tioga, and Potter Counties, Pennsylvania: Pennsylvania
Topographic and Geologic Survey Water Resources Report 68, 89 p.
Williams, J.H., 2011, Evaluation of well logs for determining the presence of freshwater,
saltwater, and gas above the Marcellus Shale in Chemung, Tioga, and Broome Counties,
New York: U.S. Geological Survey Scientific Investigations Report 2010-5224, 27 p.
58
-------
V.,, -t
I II
.
. .T in.,-(-.»-worwt
i
'"'
'toic.no
-
i jiprj -i
'
i
Figure 1. Location of study area in south-central New York, water wells that penetrated
saltwater and (or) gas, and gas wells that penetrated freshwater and (or) gas above the
Marcellus Shale
59
-------
Depth
1ft 300ft
760
780
800
820
840
860
Strata
1
u.
>
1
g
*
Gas
i
Freshwa
Gamma Resistivity Neutron
Freshwater zone o API units 20030 Ohm-m 3000 0 Percent 30
Figure 2. Geophysical logs and reported freshwater zone and gas show for the 760-860 ft depth
interval in gas well 19484, Cortland County, New York
60
-------
Fracture Design in Horizontal Shale Wells - Data Gathering to
Implementation
Tim Beard
Chesapeake Energy Corporation
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Introduction
Hydraulic fracturing has been used in the petroleum industry since the late 1940s. However,
the hydraulic fracturing of horizontal shale wells is a relatively new practice. Although relatively
"new," the hydraulic fracturing of horizontal wells is still governed by the same physics as a
conventional reservoir. The biggest differences between hydraulic fracturing operations in a
more conventional and shale reservoir are the type of fluids utilized and the volume of fluid and
sand pumped. The increase in fluid and sand volume in shale wells is primarily due to the need
to maximize stimulated reservoir volume (SRV) in the relatively low permeability formation.
The goal of hydraulically fracturing a typical shale play is to contact as much of the reservoir
rock as possible with proppant-filled fractures. The total volume contained between all propped
fractures along the wellbore represents the SRV. To maximize the SRV, there are many variables
that must be considered prior to drilling a horizontal shale well.
This abstract will focus on general fracture design in horizontal shale plays across the U.S. with
an emphasis on the data taken into consideration for each frac job and a brief discussion of
how that data is obtained and used. Additional discussion will be focused on frac modeling and
the validity of frac barriers. Finally, a brief discussion of the diagnostics used to determine frac
placement will be included.
Planning to Hydraulically Fracture a Horizontal Shale Well
Prior to drilling, companies must gather local and regional in-situ stress data (usually by drilling
a pilot hole and running logs), and make economic and land decisions concerning the
orientation, length, and placement of the lateral prior to drilling a horizontal well. With the
obtained stress data and reservoir properties, evaluation and design of the horizontal well and
stimulation is performed comprising some of the key analyses and tasks briefly described
below.
Orientation and Lateral Length
One of the first variables that is considered when drilling a horizontal shale well is the
maximum and minimum principle stress orientation in the target formation. These data are
typically estimated from wireline logs in a pilot hole. The maximum and minimum principle
stress directions are typically consistent throughout a given geographic area. Therefore, a few
61
-------
pilot holes are all that are necessary to determine the principle stress directions for a given
region within a play development area. Shale wells are typically drilled perpendicular to
maximum principle stress (Figure 3). Drilling a well perpendicular to maximum principle stress
provides an orientation where the hydraulically induced fractures can propagate normal to the
wellbore during the hydraulic fracturing process. The fractures will propagate in the direction of
maximum principle stress because they preferentially open against the minimum principle
stress. Simply stated, horizontal shale wells are drilled to create the maximum amount of
transverse fractures - thereby attempting to maximize production.
Maximum Principle Stress Direction
~
Lateral Placement
Figure 3
Lateral length is a variable that allows the operator the option of creating more (or less)
transverse fractures. The longer the lateral, typically the greater the number of perforation
clusters and the greater the number of hydraulic fracturing stages. However, maximum
practical lateral length is limited by increasing potential production difficulties that are faced in
longer laterals. Ultimately, lateral length is driven by economics associated with drilling costs,
completion efficiency, wellbore failure risk, etc. Both lateral length and the azimuth in which
the well is drilled are often affected by lease boundary considerations.
Horizontal Placement
Where the lateral portion of the wellbore is vertically positioned or "landed" is critical to
optimum stimulation and fracture geometry, and resulting well production. There are
numerous theories in the industry about where in the zone of interest the lateral should be
horizontally drilled, but a common denominator is to target the highest quality rock with
consideration given to the stress profile and predicted fracture geometries. Landing the lateral
in the upper to middle portion of the targeted, preferred rock allows for the optimization of
proppant placement in slickwater applications. From a production perspective, it is best to land
the lateral slightly lower in section and drill at a slight incline through the formation, if the
62
-------
formation dip allows for this approach. This "toe up" drilling practice promotes less liquid hold-
up or build-up across the lateral.
Data Gathering
Once the lateral is drilled, the planning of the actual hydraulic fracturing takes into account
many variables obtained from data gathered in each wellbore (or in pilot holes) by logging, and
in some cases, analysis of core samples. Some, but not all, of the variables that are involved in
the fracture design include:
Porosity and Permeability
Brittleness vs. Ductility
Young's Modulus
Poisson's Ratio
Thickness
Barriers
Depth
In-Situ Stress
Lithology
Stress Anisotropy
Natural Fractures
Gas or Liquids Reservoir
Temperature
Reservoir Pressure
Young's Modulus and Poisson's ratio are typically calculated from the shear and compressional
data estimated from dipole-sonic log response. These values are then used to calculate the in-
situ stress of the rock using several possible stress equations. A stress equation that is
applicable in many transverse isotropic shales plays is:
oHmin= (Eh/Ev)(vv/(l-Vh))(Ov-aPp) + aPp + (Eh/(l-vh2))shmin+ (Ehvh/(l-vh2))shmax
Where: oHmin= Minimum Horizontal Stress
Eh = Horizontal Young's Modulus
Ev= Vertical Young's Modulus
vv= Vertical Poisson's Ratio
Vh= Horizontal Poisson's Ratio
ov= Vertical Stress
a = Biot's Coefficient
Pp = Pore Pressure
£hmin= Minimum Horizontal Strain
£hmax= Maximum Horizontal Strain
63
-------
This equation recognizes that shales are anisotropic. With lower vh in organic rich shales and
greater Eh, the difference in oHmin between shale and sandstone/limestone decreases and often
reverses. This leads to a minimum stress in shales and the bounding sandstone/limestone
become barriers. The equation above has also replaced the stectonicterm that has been used in
the past, to incorporate lateral strain ((Eh/(l-vh2))shmin+ (Ehvh/(l-vh2))shmax). For stiff
sandstone/limestone interbedded with slightly less stiff shale, the tectonic strain creates
greater stress in the stiffer beds and less stress in the shales. This equation is the best fit for
pump-in data in the field.
Data Verification and Calibration
Pump-in tests are done on regionally representative wells to obtain actual stress values and
validate estimated stresses obtained from the above equation. A typical pump-in test is done
by pumping into a well at a rate high enough to fracture the rock with a small volume of fluid,
followed by a time period of hours to measure closure. This closure pressure provides the
actual oHmin- After-closure analysis can also be performed by observing a well post-closure to
determine permeability, pore pressure, etc. Core data are also a valuable tool in elastic
properties measurement and calibration of wireline-interpreted elastic moduli.
Fracture Modeling
Estimation of fracture geometry is modeled using an analytical fracture modeling simulator.
Rock mechanical properties and fluid loss data (permeability, porosity, pressure,
compressibility, fracturing fluid properties, etc.) are principal inputs into fracture modeling.
After entering the directional survey of the wellbore, an iterative process of comparing and
contrasting models using differing variables is performed with the goal of designing the
"optimum" hydraulic fracture for the given set of reservoir properties. An "optimum" fracture
design is one that:
1) Fractures the height of the pay interval
2) Creates a sufficiently conductive propped fracture half length that fits the well and
perforation cluster spacing, with some overlap.
3) Minimizes well interference
4) Takes into consideration the numerous variables, and accounts for the role played by
each parameter to achieve the largest SRV and ultimately the greatest production.
Fracture length and height are two primary outputs of fracture modeling software. The
example model (Figure 4) below shows a fracture half length of ~1,200' and a fracture height of
~100'. As can be seen, the fracture is contained in a lower stress region of the overall stress
column. Barriers exist above and below the primary zone of interest, confining the fracture to
the lower stress interval.
64
-------
MIV-
Width I'rotilo
Width Contours
Figure 4
The model below (Figure 5) also shows a fracture that is contained by a lower stress interval
with higher stress intervals above and beneath. It can be seen that the fracture half length is
~800' and the fracture height is ~250'. A number of factors control the height growth of a
fracture, but the relative difference between the stresses in and around the fracture is the most
important factor. Fractures tend to remain in low stress vertical regions that effectively "lock
in" or "trap" the fracture and keep it from breaking into higher stress rock. Staying in the
reservoir rock is highly desired because remaining in the zone of interest maximizes the
operators production and minimizes the wasting of frac energy on non-productive rock.
Stress
Width Prutik-s
Wulth Contours
Width i in.)
Length i ft :>
Figure 5
65
-------
Perforation Clusters and Stage Spacing
The number of perforation clusters per stage and the spacing of the clusters are area and shale
specific. In the majority of shale plays the perforation clusters are 50-100' apart. This spacing of
perforation clusters is very dependent on a number of variables. More permeability and
porosity typically allows for greater spacing between clusters. The greater the number of
natural fractures, typically the greater the spacing between clusters. A lower stress anisotropy
(which typically leads to greater frac complexity), typically results in a greater distance between
clusters. In more ductile shales, the distance between perforation clusters will be shortened.
Similarly, in a hydrocarbon liquids-rich play, where greater conductivity is typically desired, the
distance between perforation clusters will be shortened.
Stage spacing typically correlates with perforation cluster spacing. In the majority of the shale
plays 4-6 perforation cluster per stage is normal. The greater the number of perforation
clusters, the less likely it is that each cluster will get adequately treated. Thus, limiting the
number of clusters per stage typically leads to more stimulated reservoir volume. A typical
stage length is 250-500 ft.
Fluid Selection
Many variables are involved in fracture fluid chemistry design (i.e., brittleness vs. ductility,
highly anisotropic vs. low anisotropy, rate that can be achieved, fluid-rock sensitivity, etc.). Prior
to pumping any fluid systems, fluid-rock core measurements are used to determine the fluid
additives necessary in each play to prevent formation damage from drilling or fracture fluids.
The majority of the shale plays in North America are treated with a large percentage of
"slickwater". Slickwater is predominantly fresh water with additives (typically ~11 chemical
additives) that constitute less than 1 percent by volume of the liquid pumped. Slickwater is
frequently the fracture fluid of choice due to the lack of damage to the formation and its ability
to increase fracture complexity within the shales, as compared to more viscous linear or
crosslinked gels. Light gels are often used at the end of a stage to transport higher sand
concentrations. In hydrocarbon liquids-rich plays, more gels are typically utilized to carry higher
concentrations of coarser-grained proppant, allowing greater fracture conductivity.
Based on the nature of the induced fracture geometries, the volumes of fluids pumped, and the
position of fractured intervals within the geologic column, Chesapeake Energy, the American
Petroleum Institute and the American Natural Gas Alliance estimate that the risk of
contamination to groundwater from hydraulic fracture stimulation of deep shale
unconventional gas is extremely small to non-existent in most settings. However, we do realize
that there are employees who routinely work around hydraulic fracturing additives and while
safety is paramount in our industry, there is always the potential for an accidental surface spill.
It was with the concern for our employees and the potential for spills in mind that we forged
our "Green Frac" program.
Chesapeake Energy's Green Frac program was initiated in 2009 to determine if it was possible
to improve the overall environmental "footprint" of the additives used in our hydraulic
66
-------
fracturing operations. A primary goal was to eliminate any additive that was not absolutely
critical to successful completion and operation of our wells. For those deemed critical, materials
have been selected that pose lower risk to personnel and to the environment in the event of an
accidental surface discharge. To date, we have either eliminated, have found more desirable
substitutes, or are in the process of successfully testing substitutes for the majority of additives
historically used in hydraulic fracturing of unconventional shales.
Proppant Selection
Proppant selection is based on such factors as; the particular stresses to which the proppants
will be subjected, the amount of fracture flow conductivity required, propped fracture length
designed, and complexity estimated. Different proppants fit different plays and wells within
plays. A 100-mesh sand is frequently used in the early portion of many hydraulic fracturing
stages for diversion, etching, and as a propping agent. Larger 40/70- and 40/80-mesh proppants
are presently the predominant proppants used in gas shales. Still larger 30/50- and 20/40-mesh
proppants are used in some areas for conductivity enhancement. The larger proppants are
especially important in liquids-rich environments. Resin-coated proppants are being used to
"tail-in" for sand flow back mitigation and in areas where proppant strength and greater
conductivity are needed. Similarly, ceramic proppants are being used for greater conductivity
and strength. Optimum proppant selection is critical to well performance. If a sub-optimal
proppant program is implemented that does not fit the application, production can be greatly
curtailed.
Execution
Equipment for a "typical" multistage-stage fracture stimulation consists of 10-20 2,000-
horsepower pumps, a blender, 2-4 sand storage bins, a hydration unit, a chemical truck, and 20-
30 workers. After having considered all of the variables, a fit-for-purpose fracture design is
pumped. With proper pre-job data gathering and the proper consideration given to the
numerous parameters, the job is optimized for the given shale well.
Diagnostics
Microseismic monitoring, tiltmeters, gamma emitting agents, chemical tracers, production logs,
temperature sensitive or acoustic fiber optics are all tools that can and are being used to
evaluate what is happening downhole during and after the fracture stimulation job. These tools
provide better understanding of hydraulic fracturing, and improve the hydraulic fracturing
process. These topics will be discussed in detail by other authors at this workshop.
Summary
Planning and executing an "optimum" hydraulic fracture requires a multidisciplinary
approach to gathering data, evaluating the data and estimating reservoir and fracture
properties, and designing and executing a fracture stimulation program.
Using properly-gathered data, hydraulic fracture models can accurately predict vertical
barriers and the resulting fracture geometry.
67
-------
Failure to appropriately design a given hydraulic fracture treatment can result in a sub-
optimal to poor well stimulation and lower production potential, risking the millions of
dollars invested in the well up to the point of stimulation.
While the hydraulic fracturing of horizontal shale wells is relatively "new", this highly
engineered practice follows the same basic practices and science-based principals
successfully used by the industry since the late 1940's and implemented in tens of
thousands of vertical wells since that time.
68
-------
Hydraulic Fracturing in Coal Bed Methane Development, Raton
Basin, Southern Colorado, USA
Hal Macartney
Pioneer Natural Resources USA, Inc.
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Pioneer Natural Resources USA, Inc. operates a natural gas field in the Colorado portion of the
Raton Basin, a field containing approximately 2400 wells and producing 200 million cubic feet of
gas per day from coal beds. Many of these wells were hydraulically fractured by Pioneer's own
personnel and equipment. There have been no instances of damage to underground sources of
drinking water from these operations, and no more than 20' of height growth in induced
fractures.
The Raton Basin is located between the Rocky Mountains and the high plains to the east, and it
straddles the Colorado-New Mexico state line. The target formations for coal bed methane are
the Tertiary-aged Raton and Cretaceous-aged Vermejo, both characterized by intermittent thin
coals, sands, silts and shales. Both of these formations are at the surface in portions of the
basin. Naturally occurring gas seeps are common, and coal is actively mined for industrial
consumption.
Pioneer's coal bed wells are vertical and produce from depths from 450' to 3500', and from as
many as 20 coal seams varying from l'-8' in thickness. Coals are hydraulically fractured in
stages using a coiled tubing tool which enables multiple stimulations in one hole-entry. Cased
boreholes are pre-perforated in all the target coals and stimulation proceeds up from the
lowest, with each zone isolated for its treatment.
Pressures are closely monitored during the frac in three critical areas:
1. In the tubing delivering the fluids and pressure to the frac tool
2. In the open space above the frac tool, inside the casing
3. In the well-head at the surface, outside casing and inside surface casing
Tubing pressure(l) indicates the delivered pressure to the rock underground and is used to
gauge job performance in breaking down the formation and delivering fluid and sand into it.
Casing pressure (2) monitors any fluid communication from the treatment zone to open
perforations above the top packer; any such pressures terminate pumping.
The well-head pressure (3) indicates if any fluid or pressure has migrated behind casing to the
surface.
69
-------
The casing pressure(2), gives us practical and unequivocal evidence of how high our fractures
are growing; perforated zones that are too close will communicate. From experience, 20' is the
safe margin for interval spacing and therefore the upper limit of height growth. It is estimated
from performance, volumetrics, and computer models, that our lateral fracture growth is from
120-200'.
In a typical hydraulic fracture stage will use 150 barrels (6300 gallons) of foamed fluid,
consisting of 70% nitrogen, 30% water (recycled water produced from coal bed wells), 60lbs. of
a natural guar gelling compound, 4 gallons of an organic enzyme to break down the gel, and 15
gallons of a mild detergent to create foam. Around 8000 Ibs. of sand proppant is placed for
every foot of coal stimulated.
Analysis of data from 2273 Pioneer frac jobs since late 2001 shows that more than 12,000
individual hydraulic fracture stages were executed. Of these, approximately 10% were
interrupted before the end of the pumping because of high pressures (inability to initiate or
finish pumping sand), materials or mechanical difficulties, or because of pressure loss. These
last events have dropped to near zero in recent years with broader interval selection.
To date, with more than 12,000 stages pumped, there have been no instances where Pioneer's
hydraulic fracture fluids or pressures impacted underground sources of drinking water. This is
due to a number of factors. Mechanically, the fractures propagate horizontally with very little
height growth and frac volumes and energy rapidly dissipate in the formation. Geologically, the
coals and sands are discontinuous and lack through-going natural fractures. Operationally, real-
time monitoring of frac pressures and fluid volumes informs us of out-of-zone loss and results
in early shut-in. Finally, there is a competent seal all the way to surface provided by cement and
casing.
Pioneer continues to model and improve its hydraulic fracture processes, applying experiences
gained in the Raton Basin to its operations in other active plays.
70
-------
Fracture Design and Stimulation -Monitoring
Mike Eberhard
Halliburton Energy Services
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Dcoi:570405.2* This abstract provides a general overview only and is applicable to a majority of the
hydraulic fracturing treatments currently being pumped. It is not intended to address all
situations/scenarios that may occur.
As the previous sections have shown there is considerable work that goes on before a fracture
treatment is pumped. Two points that bear repeating concern (1) the importance of proper well
construction and (2) the availability of information about conditions to be expected during the
treatment. It is through the well construction process that drinking water aquifers are
protected, producing formations are isolated, casing is protected from corrosive fluids, etc. In
addition, since the fracture treatment is carefully designed beforehand and expected pressures
and other parameters are established, the casing and tubulars will have been designed to
handle the treatment and subsequent well production without compromising the integrity of
the well.
There has also been discussion about what goes into the design of a hydraulic fracture
treatment, i.e., knowledge of the mechanical rock properties of the formation to be treated as
well as adjacent bounding layers, reservoir properties of the target formation, information
about the fluid systems to be used and how the formation will interact with these fluids. From
this information the operator and pumping service company can set up the hydraulic fracture
treatment and know what will be pumped, what equipment will be required, and what is to be
expected during the actual treatment.
What Do You Need to Know before Showing up on Location
The first step in setting up a fracture treatment job is to know the expected treatment rate and
pressures. These two parameters are based on several factors discussed more thoroughly
within this workshop, but for this section it is important to note that they are calculable. For a
given formation there is a pressure which when applied will cause the rock to fracture. This
pressure is often referred to in terms of a gradient (fracture gradient - fg). Knowing the fracture
gradient, the actual bottom hole treating pressure (BHTP) required to fracture the rock can be
calculated for a given depth:
BHTP = fg * depth + excess pressure (1)
71
-------
In this equation excess pressure is the additional pressure required to extend a hydraulic
fracture; i.e., net extension pressure, process zones stress, etc. These excess pressures are
typically significantly lower than the pressure required to fracture the rock.
Once the BHTP is known then an expected wellhead treating pressure (WHIP) can be calculated
by accounting for additional pressures that occur while treating a well:
WHIP = BHTP + Ppipe + Pperf - Phyd (2)
In this equation Ppipe is the friction pressure resistant to flow down the wellbore during
pumping operations and is fluid and rate dependent; Pperf is the pressure drop across the
perforations; and Phyd is the hydrostatic pressure of the fluid in the wellbore and is also fluid
dependent.
Once the expected BHTP and WHTP are determined, the proper casing string or tubular
configuration can be designed to handle the pressures experienced while treating the
formation. The WHTP is also used to calculate the hydraulic horsepower (number of trucks;
HHP) required to pump the job at the desired treatment rate from the following equation:
HHP = (WHTP* Rate)/40.8 (3)
The next step in setting up a job is to know what will be pumped, e.g., the additives required
and the rates at which the additives are to be used, proppant type and volume, etc. For some
jobs this requires pre-job testing to determine whether the fluid system intended for use in the
fracture treatment is compatible with the base fluid being supplied on location. This is an
important step since it also establishes what will be required for the fluid system to perform as
desired. Once this information is known then a final treatment design is determined and
communicated to the field location for execution. This information is then put together in
tabular form, giving the operator and service company a ready guide for setting up the job. An
example of a typical pump schedule is included in the appendix.
Rigging Up the Pumping and Monitoring Equipment
The care that is taken in designing a fracture treatment job carries over to the implementation
of the job, beginning with the set-up for the job. After the equipment, personnel, and materials
are on location a safety meeting is held. During this safety meeting items such as well site
concerns, proper PPE, rig-up concerns, etc. are reviewed to ensure that appropriate steps are
being taken to ensure safety on the job site. The time it takes to rig up the pumping equipment
and surface treating lines can vary from a couple of hours to a couple of days depending on the
treatment. During this time there is also quality control work going on to ensure that the
fracturing fluid will perform as expected and that the correct materials are on location in the
appropriate quantities.
After all the surface equipment has been rigged up there is another safety meeting. During this
safety meeting details of the job are reviewed, including the maximum WHTP, expected WHTP,
72
-------
pump rate, overall job schedule, who is responsible for what, etc. After the safety meeting all
surface piping is pressure tested to a predetermined maximum pressure. At this time the pop-
off valves on the surface lines are tested to make sure they work at the desired pressure and
the pressure kick-outs on the high-pressure pumps are also tested to insure they work properly.
In addition, the pumps used for liquid additives are bucket tested to ensure that they are
functional and are calibrated properly. The proposed pumping schedule is loaded into the on-
site computer system to assist the fracturing treatment operator in running the job as close to
design as possible. While computers are capable of actually running the treatment, at this time
most service companies still rely on a team in the treatment van to control the actual fracturing
treatment with the assistance of the computers.
Pumping the Treatment
Once everything has been calibrated and pressure tested there is generally one last review
between the operator's representative and service company representative to go over the
treatment parameters. Once everyone is in agreement, the wellhead is opened up and the high
pressure pumps are brought on line. At this time fluid is being pumped down the wellbore at a
slow rate as pressure starts to increase. The rate and pressure are increased to the anticipated
WHTP where the formation should fracture (breakdown). This is one of the first points where
the actual treatment can be calibrated to the job design. If breakdown does not occur within a
reasonable pressure compared to what is expected then the treatment is shut down and
possible causes are investigated.
There are several points on the surface where rates, pressures, and densities are monitored
and recorded during a treatment. (A simplified location schematic showing where the different
treatment monitoring occurs is provided in the appendix.) For example, highly accurate
transducers are placed at several different locations in the surface lines and equipment to
monitor real-time pressure data, a variety of different flowmeters are used (depending on the
material being metered) to record treatment rates and additive rates, and densometers are
used to measure the density of the fluid being pumped downhole. Examples of some of the
data being monitored and recorded include: WHTP, annular pressure, downhole slurry pump
rate, clean fluid rate, wellhead proppant concentration, and individual additive rates, along
with an extensive amount of mechanical information about the equipment on location. All the
information from these multiple sources is collected and displayed by state-of-the-art computer
systems located in treatment control vans. Most of the time, these data are transmitted using
hard wires connecting the computer to the monitoring device.
It is also important to note that in addition to monitoring there are also mechanical devices
which are used during a fracture treatment to provide additional safety for the wellhead. Two
of these devices are pressure pop-off valves on surface lines and pressure kick-outs on the high
pressure pumps.
While pumping the treatment both the operator and service company continually monitor the
computer screens displaying information about the treatment as it is being pumped. The main
concern is pressure. Both the operator and the service company want to make sure the
73
-------
maximum WHIP is not exceeded to protect the wellbore from any possible damage. (It is
important to understand that it is inefficient to have to repair wellbores so every effort is made
to prevent them from being damaged.) Some variations in pressure are normally seen during a
fracture treatment. These variations are interpreted to determine their causes and significance;
there are constant decisions being made about what the status of the treatment is and what to
do as the treatment proceeds. An example of a treatment chart can be found in the appendix of
this abstract.
Close attention is also paid to the annulus. In many cases the annulus is monitored with a gauge
for any pressure increase in excess of normal fluid cool-down and heat-up, in other cases the
annular valve is open and any fluid flow up the annulus can be seen at the wellhead and
appropriate steps can be taken to address the fluid flow in the annulus.
Since any additive used in a hydraulic fracturing treatment serves a specific purpose, it is
important that these additives are run at their designed concentrations. As mentioned earlier
all additive rates are monitored during the treatment to insure they are run correctly. (An
example of an additive rate chart is shown in the appendix.) In addition, overall job treatment
information is displayed in the treatment control van in real-time to assist the operator and
service company in understanding how the treatment is progressing. This allows for spot checks
throughout the treatment process to compare the physical inventories of volumes of additives
pumped with those calculated to again insure the treatment is being pumped as planned.
In addition, during the pumping operation there is continual monitoring of the surface lines,
equipment, and wellhead to make sure there are no leaks. If a leak does develop, it is either
isolated if possible or the treatment is shut down and the leak fixed before pumping is
resumed.
The majority of hydraulic fracture treatments are pumped as planned or with changes that are
based on the way the treatment is proceeding. On occasion, the formation may be difficult to
fracture stimulate, resulting in a rapid pressure increase while pumping; this is called a screen-
out. Even if there is a rapid increase in pressure relative to normal increases in pressure due to
pumping, the system is still compressible so there is still time to react. As the pressure
increases, the fracture treatment operator will start bringing pumps off-line to counteract the
rapid pressure increase. In a worst case scenario, if the pressure increases too fast then the
pump kick-out will activate and shut down the treatment.
After the Fracture Treatment
After the well has been treated the equipment used in the fracture treatment is rigged back
down. At this time there is another safety meeting to discuss any possible issues that may be
associated with this rig down. A final physical inventory of materials still on location is
conducted to determine the actual volume of materials that was pumped during the treatment.
During the rig-down of the pumping equipment steps are taken to prevent any spills and
surface contamination. Finally, the operator is provided with a post job report that provides
74
-------
details of the treatment, a summary of what occurred during the time on location, and what
was pumped into the well.
Appendix
Nomenclature and Terminology
Treatment Rate (bpm) - the downhole rate that fluid is entering the formation
Hydraulic Horsepower (hhp) - horsepower being applied to the formation while pumping
Wellhead Treating Pressure (psi) - the surface pressure at the wellhead during pumping
Max Pressure (psi) -the maximum WHTP that will be allowed
Bottom Hole Treating Pressure (psi) - pressure being applied to the formation including net
pressure
Frac Gradient (psi/ft) - pressure at which fluid will cause the formation rock to part
Pipe Friction Pressure (psi) -friction pressure of the fluid being pumped down the wellbore
Perf Friction Pressure (psi) - pressure drop across the perforations
Hydrostatic Pressure (psi) - pressure the fluid column exerts on the formation
Net Pressure (psi) -excess pressure overfrac pressure required to extend the fracture
Instantaneous Shut-in Pressure (psi) - a pressure used to calibrate the frac gradient
Clean Volume (gal or bbl) -volume of fluid pumped without proppant
Dirty Volume (gal or bbl) - volume of fluid pumped with proppant
Proppant Concentration (Ib/gal) - the amount of proppant added to one gal of fluid
Proppant - small diameter material used to keep the fracture open
Solid Additive (Ib/Mgal) - a solid chemical added to the fluid system for a specific purpose
Liquid Additive (gal/Mgal) - any liquid chemical added to the fluid system for a specific purpose
Pop-off-a mechanical device activates at a preset pressure to prevent damage to surface and
downhole tubular
Kick-outs - mechanical or electrical devices that activate at a preset pressure to disengage high
pressure pumps
High Pressure Pumps - Positive displacement pumps used for pumping downhole
Centrifugal Pumps - used on the low pressure equipment to mix and move fluid
Additive Pumps - used to inject liquid additives; different types based on the additive type and
additive rate
Pressure Transducer - device used to measure and transmit pressure data
Flowmeter - used to measure and transmit fluid flow rates; different types depending on
application
Annulus - Area between two concentric casing strings or tubular strings
75
-------
Figure 6. Simplified Location Schematic
High pressure
Low pressure
1 -Annul us pressure gauge
2- Surface treating pressure gauge
3- Down hole densometer
4- Slurry flow meter
5- Clean flow meter
6- Gel flow meter
7- Liquid add itwe flow meters
High pressure pumps
Proppant blender
Pregel blender
Liquid additive tanks
Proppant storage
Flowmeter
() Stop valve
(/) Pressure transducer
Kick-out
Dens | Densometer
PO| Pop-off
Figure 7. Inside the treatment monitoring van
76
-------
A
8(X
70C
60C
50C
4EK
30C
20C
IOC
10-
10-
10-
>0-
10-
1)
Mum Proppant Cone lib pah C1 BH Proppant Cone
)
3
1
I
3
)
9
J
3
c
-15
-14
-13
-12
-11
-10
-9
-8
-4
-3
-1
-0
Figure 8 Treatment Chart -- Pressure, Rate and Prop Concentration
A
9-
8-
7-
6-
5-
4-
:
Additives
BA-4
i
3L. BULK { 1 00003 Cone (gal Mgal) A
--,
^,.M,,tl.«l-»"^V^HM^
U^VU
«*M
,r- ^ J. , , -1^4" ^
V^H^
A 1 A A 1 A A 4
»
A
A
..
: : y 1 -S V-"
06:50 07:00 07:10 07:20 07:30 07:40 07:50 08:00
: :4 :ou : ;* :on
Time
U.
:
B
-35
-30
-25
-20
-15
-10
-5
Figure 9. Additive Chart
77
-------
Steqe Tim« mn 4 45
Slag. Pure Tm«. min 4 4b
Stag* Puro Time La* mm 244
Ptennod Volume 90! 14900
Actual Vdurt* gat 9630
Sloge Sunmtr
Clean Volume /
2190
TOSH
1MM
.
IB;
113.
Stage a! WH 126
Treating Pressure
Slurry Rat*
Calc'd BH Pres
Hydrostatic Pressure
Slurry Proppant Cone
BH Proppant Cone
Job CMan Vol
Stage Clean Vot
Job Proppant
Stage Proppant
Proppant In FormatKwi
Proppant in Wellbore
BA-40L
CAT-3 Cone
CAT-4 Cone
CL-23 Cone
CL-37
FR-66 Cone
Vlcon NF Cone
WG-ISCooc
4995
61 39
6416
3066
0
3566
1792196
9738
2133101
0
:-:i4-4e
1450
0
0
0
0
0
1.010
2008
0
09:33:36
p»l
bpm
psi
psi
gal
ft
i
,0
gal/Mgal
gali'Mgal
gal'Mgal
gal/Mgal
gal.Mgal
gai'Mgal
gal'Mgal
Figure 10. Fluid Tracking, Numeric Value, and Stage Summary Screen
Company:
I raw.
1- ii i m utli. u
ihrl'-lim
C X\ 130 '
HALLIBURTON
I . jliiin
IKS\U.I I'U
nm on...»-a
-*K. -.,-
7317 ««
'
i :
i r
i .<
> .
11.
limn omm »«4
< n«
i
lltl* <*
ItM*
:
KM M III III
:««H«HI MMI:
1:1*
tarfDMlfBlMd 50 JO W
rni-«i r nwm |
KMI r«d. J'.fc-.
T... Il
! Ml, |
\UXPRI':SMRr.: 5lH.li fa ic^
ra».. T p.«
r>rt/«w>l MM
p.rf/«»«: «.»
I .k P«4 F>W IJWI 11
1.1 V> ^ B~. In. lp.il- l.M«
Kill IU1KI Hill
II.' 1 1 *.
OWrt ft
a
VIAUVII M 1 HI MM At. Mini in 1
I*1M^ Kali* lyliMUi Li
1* tit
i>l Im kn 1
|i«ii i«»T
-I "-I.--.!
Figure 11. Blender schedule
78
-------
A Case History of Tracking Water Movement Through Fracture
Systems in the Barnett Shale
Pat Handren
Denbury Resources, Inc
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Staged fracturing has been successfully carried out in the Barnett Shale for the past several
years. Over the course of these years diagnostic work has been performed to assess the
geometry of the complex network of fractures created during the pumping process. One of the
tools used to "measure" the parameters of fracture azimuth, length, height and width of what
has been termed as stimulated reservoir volume (SRV) is micro-seismic detection. The observed
SRV coverage has proved very useful for predicting potential areas of communication between
wells with increased density spacing. This tool has provided information that has led to
improved stimulation efforts by many operators.
In the case study area there were two wells that were observed with micro-seismic mapping.
The mapping is based on the detection of small seismic events that occur during the fracturing
process. The locations of the events are determined by analyzing the signal received at the
monitoring tools. Essentially the event is triangulated by looking at the strength of the signals
received by receivers oriented in different directions. This method of analysis provides a "map"
of where the fractures could potentially be occurring.
The first well was located to the north of the case study well. Four fracture stimulations were
done, consisting of approximately 34,000 barrels of water and 340,000 pounds of sand for each
stage. The average of the stages was an azimuth of north 45 degrees east and an SRV per stage
of about 900,000,000 cubic feet (~21,000 acre-ft). The average height of the SRV was 350',
which means each stimulation stage covered about 60 acres of area. Of interest in this well was
the observed growth during the fourth stage. This was the least contained of all the stages, but
it was still limited by the lithology change from the Barnett Shale into the Marble Falls
formation.
The second well was located to the south of the case study well. Two fracture stimulations were
performed, consisting of approximately 20,000 barrels of water and 380,000 pounds of sand for
each stage. The average of the stages was an azimuth of north 34 degrees east and an SRV per
stage of about 490,000,000 cubic feet (~11,000 acre-ft). The average height of the SRV was
400', which means each stimulation stage covered about 28 acres of area. One of the most
important things observed with the Micro-seismic mapping in both wells is that the height
growth was well contained to within the Barnett Shale interval.
79
-------
Even with the tools available to perform fracture diagnostics operators are still faced with
challenges that are difficult to predict. As well density increases it becomes increasingly
probable that wells will communicate either through previously created fractures or through
adjacent wellbores and then into previously created fractures. The occurrence of this type of
communication will be reviewed for a well that was fractured in 2009.
A typical Barnett Shale well in the area has 9 5/8" surface casing set at 850' and is cemented to
surface to protect fresh water sands. The well is then drilled to a true vertical depth of about
6700' with a lateral length of approximately 3000'. After drilling to total depth, a 4 1/2"
production casing string is run to bottom and is cemented in place with cement to 5400' or
higher. The top of cement depth is verified with a cement bond log run on electric line. The well
is then ready for stimulation. Each stimulation stage is preceded with perforating of three
intervals.
The case study well had plans for six staged fracture stimulation. Each stage was scheduled to
be pumped at a fluid rate of 100 barrels per minute with an average fluid volume of 17,000
barrels of fluid and 250,000 pounds of sand. If the wells previous evaluated with micro-seismic
mapping gave any insight into SRV based on fracture treatment volume, then the estimated
SRV would be approximately 410,000,000 cubic feet (~9,400 acre-ft). If an average height is
assumed at 375 feet, then the average are covered by each stage would be 25 acres. This would
mean that all six stages covered a total of 150 acres. Over the course of performing the
stimulations in the well communication was achieved to wells spread over more than 600 acres.
The farthest well that was "hit" by water from one of the stimulations was 1,500' away. A total
of six wells were affected by water from the study well's stimulations.
The basic conclusions drawn from the both the micro-seismic mapping and the observations
made from the study well is that the stimulations stay reasonably contained within the Barnett
Shale interval. As well density increases the complexity of subsequent fracture stimulation
interaction with adjacent wells increases. If fracture azimuth is known, then a reasonable
estimate of well to well interaction can be predicted.
80
-------
Measurements and Observations of Fracture Height Growth
Norman R. Warpinski
Pinnacle - A Halliburton Service
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Hydraulic fracturing is a process that is necessary for economic extraction of natural gas and oil
from unconventional resources such as tight gas sands and gas shales. It is a process that is well
understood in its overall behavior and development, but is difficult to quantify in many of the
details because of both geologic and mechanistic uncertainty. For example, fine details of the
layering are impossible to resolve using the borehole tools available today, and features
between wells are difficult, if not impossible, to distinguish unless their scale is extremely large.
The mechanistic uncertainty follows from the poor description of the reservoir and the geologic
features within, but also from the computational difficulties associated with a complex
interaction problem in a heterogeneous material.
Nevertheless, thousands of papers have been written in the petroleum literature to study
hydraulic fracturing, and these have provided a wealth of understanding about the behavior of
fractures in different environments. These papers have provided field evidence, mineback and
coring evidence, laboratory testing, analytical models, numerical models, and a host of other
results that have guided the understanding, development and optimization of the fracturing
process. What we may be missing in the fine details can be accounted for in overall generalized
findings about the fracturing process.
Geology, Geology, Geology
It should be obvious from the literature that we
only have a limited ability to direct fracture
growth; Mother Nature does not let go easily.
The best example is fracture azimuth (the
direction a fracture propagates), which is
dictated by the in situ stress that exists at the
hydraulic fracture location and is very difficult
to alter. Fractures will propagate in the same
direction all across a field. A second general
finding is that the layered earth sequence
makes vertical fracture height growth difficult,
thus generally promoting the growth of length
over height. Height growth is inefficient due to
the variable layer properties, the large number
,. . , ., ... . ., Figure 12. Mineback photograph of complex
of interfaces, the rapidly varying stress that can
fracture
81
-------
occur vertically, and the potential for a large number of energy-dissipative mechanisms that can
occur in such an environment.
Figure 12, for example, shows a mineback photograph of a hydraulic fracture that has very
complex behavior that is largely due to geologic
factors, such as the stress state at this location and
the interfacial properties. Fractures are not single
planar features that extend long distances; they
are a series of interconnected fracture segments
that have many internal terminations and
interactions with the local geologic conditions
(Warpinski and Teufel 1987).
Figure 13 shows a second example of the
complexity that can occur as hydraulic fractures
intersect natural fractures and other geologic
discontinuities (e.g., interfaces). There are many
offsets and some splits that occur as part of this
interaction process, the details of which are largely
driven by the local stress state and the material
properties in conjunction with the treatment
conditions. In many instances, natural fractures,
faults, and interfaces have been observed to
terminate fracture growth, thus providing a complete
containment feature.
The in situ stress has a dominant role in all of these
processes, but also directly affects vertical hydraulic
fracture growth. Fractures are impeded from growing
vertically by higher stress layers. This might appear to
be an unusual case because stresses decrease as the
depth becomes shallower, but measurements have
shown that large stress contrasts exist in sedimentary
basins at all depths.
Figure 14 shows an example of the results from a
stress measurement program at the DOE funded
multi-well experiment in the Mesaverde formation
located in the Piceance basin (e.g., Warpinski and
Teufel 1989). The stress measurements made in
reservoir rocks (sandstones) are shown in blue,
whereas the non reservoir shales, mudstones, and
siltstones are shown in red. The stress contrasts are
Figure 13. Mineback photograph of offsets
& splitting.
8500
. m\ 2,000 psi
* m \ / stress
| ." \f contrast
» \
i Lithostatic
V (1.07psi/ft)
i\
\
0 2000 4000 6000 8000 10000
Stress (psi)
Source: DOE Multiwell experiment & DOE/GRI M-Site test
Figure 14. Measured stress profile in
Mesaverde.
82
-------
often in the range of 1,000 - 2,000 psi. While the overall trend is one of decreasing stress with
shallower depth, the large variations make it unlikely that fractures would grow very far across
such a section. Fractures that grow out of zone and propagate vertically upward would quickly
hit another low stress layer and tend to grow laterally in it. Should the pressure overcome the
next higher stress layer above it, then the fracture would grow and again hit a lower stress
layer, and also result in preferential lateral growth. Repeated crossing of these layers is an
inefficient process that soon uses up the fluid and energy.
All of these processes and mechanisms have been verified in laboratory testing and modeling.
We now have the laboratory equipment to study layered and fracture rocks and the
computational tools to study fracture behavior in a discontinuous medium. As noted above, the
exact details may be difficult to determine because of the poor understanding of the geologic
details, but the overall behavior is very clear.
Diagnostics Tell the Story
While all of the mechanisms discussed above provide the understanding of what is occurring as
fractures propagate, it is the advent of far-field diagnostic technologies that have given us a full
picture of the propensity of fractures to propagate laterally. Although tiltmeter deformation
measurements have been applied more often and longer, it is microseismic technology that has
been the most revealing.
Microseisms are small earth movements that occur in the vicinity of a hydraulic fracture due to
inflation of that fracture and leakoff of high pressure fluid into the formation. These two
mechanisms cause changes in both stress and pressure that can induce complex shear slippage
processes. These microseisms emit seismic energy that can be detected at receiver arrays
located in adjacent wells, and the waveform data, in conjunction with a velocity model, can be
processed to extract microseismic locations. The sum of these locations yields a map of where
the activity is occurring which describes the fracture.
One common question is that of validation. How can we be sure that the microseismic data is
representative of the true fracture behavior? The answer to that question is in the results from
several validation experiments, the most extensive of which was the DOE/GRI funded M-Site
test in Colorado. (Warpinski et al. 1998) Figure 15 shows a side view representation of the
testing results from M-Site, in which several approaches were taken to verify the microseismic
data. There were two monitor wells with seismic receivers to capture microseismicity, but there
were also tiltmeters cemented in place in one of the wells to measure the earth deformation
and compare the mechanical behavior with the microseismic behavior to verify fracture height.
In addition, intersection wells were drilled to verify fracture azimuth and examine the fractures
in core or with imaging logs, but one of those intersection wells was drilled prior to fracturing
and instrumented with pressure gages. During fracturing, the time at which the hydraulic
fracture intercepted this well could be determined by an observed increase in pressure, thus
providing a fracture length at that time which could be compared to the microseismic length.
All parameters - length, height, and azimuth - exhibited close agreement between the
microseismic results and the verification technologies.
83
-------
Monitorwell:
cemented-in
Gamma Ray Tiltmeter heights
intersection
wells
/ /
Microseismicdata
Monitorwell:
wireline
receivers
Cemented-in receivers Intersection wells
Source: DOE/GRI M-Site test
Figure 15. Overview of DOE/GRI M-Site hydraulic fracture diagnostics field test
site.
While only a very limited number of industrial fracture monitoring projects have been
published, there are many thousands that have already been done and these provide a
comprehensive record of the behavior of fractures in these sedimentary environments. Figure
16 shows a case of a Haynesville shale
fracture (Pope et al. 2009) where there is
some extensive height growth - on the order
of 600 ft. This degree of height growth does
occur in some of these deep shale reservoirs
and the monitoring provides information that
can be used to optimize the process as much
as is possible. Any amount of height growth
out of zone is undesirable because it wastes
fluid, horsepower, chemicals, and time. The
point of hydraulic fracturing is to stimulate
the reservoir, not the unproductive rocks
around it. Monitoring provides information
that can be used to figure out ways to
minimize this behavior.
Figure 16. Example Haynesville shale
microseismic data.
SPE 125079 GMX Resources
84
-------
Since one monitoring test proves nothing and one can always use the best examples, a more
compelling result can be demonstrated by showing all of the fracturing results in a basin in a
correlated plot. Figure 17 shows the results of nearly 2400 fractures in the Barnett shale prior
to mid-2010 - everything that was monitored up to that time (Fisher 2010). The plot has been
sorted by depth, with deeper wells on the left. The perforation depth is shown, along with the
top and bottom of the hydraulic fracture as measured by the microseismicity. Although difficult
to see and read, the data are also colored by county. In addition to the fracturing results, the
deepest water well in each county, as obtained from the USGS web site, is also plotted at the
top.
Perf/Frae TVOs (line)
Figure 17. A compendium of microseismic fracture diagnostic results in the Barnett shale
relative to known aquifers.
These results show that fracturing does not intrude on the aquifers. There is a limit to how
much a fracture can grow vertically, even in the most advantageous conditions. There is
considerable variability in fracture height in this plot, with much of it due to intersections of
faults. However, even the most extreme cases do not extend vertically anywhere close to the
aquifers. Similar results have been compiled for the Woodford and the Marcellus shale and
those plots look similar.
The fractures that have been compiled in Figure 17 are for relatively deep injections, but there
are many reservoirs that are much shallower. One might expect that fracturing to surface
would be common in shallow reservoirs, but Mother Nature again conspires against vertical
fracture growth by reversing the stress field at shallow depths. Hydraulic fractures at depths
greater than ~2,000 ft are mostly vertical, but at depths less than ~1,500 ft, they are either
horizontal or mostly horizontal (a vertical component in some layers) due to the overburden
stress being generally greater than the horizontal stresses at shallow depths. There is a wealth
85
-------
of tiltmeter data on ~10,000 fractures that details how fractures have primarily vertical
components at depth, but have a larger percentage of the fracture growing horizontally in
shallow environments.
Summary
There are over seventy years of experience in conducting hydraulic fractures, a multitude of
fracture models, thousands of petroleum engineering papers on the subject, many years of
studying fractures using minebacks, corethroughs, laboratory experimentation and numerical
analysis, and most recently the application of fracture diagnostic measurements in thousands
of projects across North America. All of this knowledge and information has provided a sound
understanding of the basic principles and general behavior of hydraulic fracturing.
Vertical propagation of a hydraulic fracture across layers is very inefficient and it is difficult to
obtain extensive vertical growth. Fracture heights of several hundred feet are common, and
they may occasionally exceed 1,000 ft in a few deep reservoirs. However, there has never been
an observed case of a hydraulic fracture propagating thousands of feet vertically to intersect an
aquifer. In shale projects where large fluid volumes are injected, the thousands of diagnostic
measurements have consistently shown that fractures remain thousands of feet deeper than
the aquifers.
Fractures do occasionally intersect faults, but the diagnostic information shows that vertical
growth is also limited when this occurs. Some of the largest measured heights occur in cases
where a fault has been intersected, but growth is equally likely to be downward as upward and
it is typically only about twice the height of a normal fracture.
Shallow hydraulic fractures are not observed to grow vertically because of the changing stress
state. Less than about 1500 ft, the overburden stress is the least principal stress and this causes
fractures to be primarily horizontal at shallow depths. Some vertical components may occur,
but they are typically very limited.
References
Pope, C, Peters, B., Benton, T., and Palisch, T. 2009. Haynesville Shale: One Operator's Approach
to Well Completions in this Evolving Play. Paper SPE 125079, SPE Annual Technical
Conference and Exhibition, New Orleans, Louisiana, 4-7 October.
Fisher, M.K. 2010. Data Confirm Safety of Well Fracturing. American Oil and Gas Reporter. July.
Warpinski, N.R., Branagan, P.T., Peterson, R.E., and Wolhart, S.L. 1998. An Interpretation of M-
Site Hydraulic Fracture Diagnostic Results. Paper SPE 39950 SPE Rocky Mountain
Regional/Low Permeability Reservoirs Symposium, Denver, Colorado, 5-8 April.
Warpinski, N.R. and Teufel, L.W. 1987. Influence of Geologic Discontinuities on Hydraulic
Fracture Propagation. Journal of Petroleum Technology, 39: 209-220.
Warpinski, N.R. and Teufel, L.W. 1989. In Situ Stresses in Low-Permeability, Nonmarine Rocks.
Journal of Petroleum Technology, 41: 405-414.
86
-------
Sustainable Fracturing Rationale to Reach Well Objectives - The
Impact of Uncertainties and Complexities on Compliance
Assurances
Ahmed Abou-Sayed
Advantek International
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
The presentation will discuss lessons learned; extract best practices and guidelines applied to
injection of fluids and slurries during fracturing and exploration and production (E&P)
associated streams disposal (wastes, produced water, drill cuttings, and solids/proppant flow-
back). Fracture generation, propagation and multiplication during multiple injections in same
well, batch injections and re-fracturing is covered. Design requirements, monitoring and
assurance of containment for environmentally safe injections are covered. Results from major
worldwide injection projects are viewed from operator's and regulator's perspectives.
87
-------
Design and Rationale for a Field Experiment using Tracers in
Hydraulic Fracture Fluid
Daniel J. Soeder
U.S. Department of Energy, National Energy Technology Laboratory
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
The economic recovery of natural gas from organic-rich shales requires the use of horizontal
boreholes and staged hydraulic fracturing. Many questions have been raised about the
potential threat this production method may pose to groundwater. Field-based measurements
to gather hydrologic and geophysical data from a representative hydraulic fracture treatment in
the shale could help ascertain the movement of hydraulic fracture fluid in the ground, and
determine how close it might come to contaminating drinking water supply aquifers.
Geophysical field data collected by microseismic methods show the extent and dimensions of
hydraulic fractures created in lateral boreholes as a stimulation technique for shale gas
production. The data indicate that hydraulic fractures do not approach closer than several
thousand feet below the freshwater aquifers above the Barnett Shale of the Fort Worth Basin,
and the Marcellus Shale of the Appalachian Basin, the two major shale gas production areas in
the U.S. (Fisher, 2010). Nevertheless, there is still a degree of uncertainty concerning the
potential effects that such fracturing treatments might have on groundwater. In particular, the
possible migration of fracturing fluids from the target production formation into drinking water
supply aquifers remains a hotly-debated topic. The absence of rigorous data to support either
side in this argument has left the general public confused, concerned, and in some cases
frightened.
The proposed field experiment would begin by collecting representative groundwater samples
for baseline analysis along the planned trajectory of the horizontal borehole prior to drilling.
Structural features will be located by a seismic survey during site characterization, and
additional groundwater sampling points will be installed over structures such as faults, which
might provide conduits for hydraulic fracture fluids to move out of the stimulation zone and
into aquifers. Soil gas samples will also be collected from locations above the laterals and
analyzed for any traces of natural methane or radon gas potentially released by the fracture
treatment. Prior to hydraulic fracturing, a conservative tracer will be placed in the fracturing
fluid. Microseismic and other advanced geophysics will be run above the laterals during the
hydraulic fracturing process to map the length and orientation of the induced fractures. A series
of groundwater samples will be collected before, during and after the drilling and hydraulic
fracturing operations, and analyzed for the tracer. Groundwater sampling will be carried out at
regular intervals for a few weeks to months after the hydraulic fracturing to determine if there
is any upward movement of fluids over time.
88
-------
After completion of the hydraulic fracturing, a vertical borehole will be drilled down to the
Marcellus Shale, and continue through it to the underlying Oriskany Sandstone. The drilling will
be paused at water-bearing formations, such as sandstones and limestones, to collect
formation water samples. The samples will be analyzed for the tracer, to determine if it has
contaminated any of the deeper saltwater aquifers. Water samples will also provide data on the
chemistry of natural formation brines in the basin, and determine if the brines in the Marcellus
are chemically related to other formation waters. Data collected from this experiment should
provide insights into the location of hydraulic fractures in relation to aquifers, the potential for
the upward movement of hydraulic fracturing fluid to contaminate groundwater, and the
geochemistry of Appalachian Basin formation waters in comparison to the Marcellus Shale.
Reference: Fisher, Kevin, 2010: Data Confirm Safety of Well Fracturing, The American Oil & Gas
Reporter, July 2010
89
-------
Review of Stimulation Fluid Retention Mechanisms and
Likelihood of Fluid Communication with Shallow Potential
Aquifers in Shale Gas Development
Scott Cline, PhD Petroleum Engineering
Stanley, NY
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
What happens to injected fluid in the subsurface and whether or not it is a risk to potential
sources of drinking water has previously been reviewed by others including EPA in the 2004
study of possible effects on USDW from hydraulic fracturing of shallow coal bed reservoirs (EPA,
2004). However, as opposed to coal bed methane development where stimulation volumes
were relatively small, fluid recoveries high, and depths shallow; the development of shales and
other "tight" formations involve large fluid volume stimulations in long horizontal wells with
typically very low fluid recoveries but at generally deeper target depths. It is therefore prudent
to revisit distinguishing and relevant aspects of the fluid retention subject.
Relevant factors to consider include explaining low overall fluid recoveries occurring with most
horizontal well stimulations in low permeability, assessing if there are any potential induced or
natural paths with conductivity from the target formation to the shallow aquifers and if so is
there a persistent driving force toward the shallow aquifers. While this abstract addresses
primarily Devonian and Ordovician shales, the concepts are applicable to many other low
permeability naturally fractured formations.
With continued increase in temperature and pressure during rapid burial and dewatering,
organic matter within shale was converted into natural gas. Because permeability was low and
burial rapid, pore space could not expand sufficiently to accommodate the gas generated. The
gas generation thus resulted in an incremental pore pressure to such a magnitude that the rock
cracked in a massive network of natural fractures. (Engelder, 2008) As more hydrocarbons were
generated, the cracks continued to grow until they opened into full scale joints or natural
hydraulic fractures which culminated in significant over-pressuring as the gas was unable to
escape the relatively closed system. Although other overlying organic rich shales have similar
fracture mechanisms, such complex joint systems do not likely extend conductively very far
vertically. Certainly other tectonic related faulting occurs in the subsurface but it is rarely
systematic (Engelder, 2009).
When these natural fractures are then subsequently hydraulically fracture stimulated in
horizontal wells, the wells typically exhibit good gas production but poor or slow fluid recovery.
The fluid recovery factor is usually significantly less than 50% and in horizontal wells in low
permeability formations such as Marcellus, the recovery of total produced water, which
90
-------
includes both injected and natural formation water, is often less than 20% of total injected fluid
volume. The relative contributions to fluid entrapment are still not completely understood or
quantifiable but include:
1. Fluid "leak-off" into the fracture face. In formations with low permeability and low
water saturation this fluid is permanently phase trapped by capillary forces related to
pore size, size distribution, and wettability.
2. Hydrophilic clay adsorption and clay swelling restricting flow.
3. Narrow fracture branches (shear fractures as opposed to tensile fractures) trap fluid by
capillary forces and stranding especially beyond the zone of production influence.
4. Fluid in proppant packs may be unable to move as fluid prefers going around the packs.
5. Fluid may move by gravity to the bottom of the fractures and unable to move as
reservoir flow dynamics slow with depletion. (Daneshy, 2010)
Permanent fluid retention by the formation is possible via the fluid leak-off mechanism since
some fracture fluid is injected and imbibed into the reservoir rock surrounding the fracture
(Penny et al, 2005). These low permeability facies have extremely small effective pore throat
radii and/or micro-fracture widths, high irreducible wetting phase saturations and significant
capillary pressures that generally increase as permeability decreases (Kalfayan, 2008). As the
well is produced, the produced gas must overcome the capillary pressure at the formation-
fracture interface. When the gas does succeed in breaking through the fluid at the interface,
the gas flows through the point of least resistance leaving a large portion of injected water
phase trapped in the reservoir rock and at the reservoir-fracture interface. While low
permeability limits the leak-off penetration, the larger surface area of the off-balance fractures
compensates for this effect and can allow a substantial volume of fluid to become trapped in
the formation. Some authors believe that this leak-off effect in low permeability formations is
limited and accounts for less than 25% lost fluid but indeterminate because of the uncertainly
of complex fracture surface area (Daneshy, 2010).
Although Marcellus shale is generally characterized by relatively high quartz contents (60%) and
relatively low clay content (muscovite-illite 30%, kaolinite 2.4%) relative to some other shale,
hydrophilic clay content in shales may contribute to water retention in shale through
adsorption and pore throat constrictions due to swelling (Boyce and Carr, 2009). Nevertheless
this mechanism is likely to contribute somewhat to injected water retention and could result in
permanent retention.
Recently, Daneshy and others have theorized that the dominant factor responsible for water
retention in naturally fractured shale is simply the interaction of three types of induced
fractures: tensile, sliding shear and twisting (Daneshy, 2010). Tensile fractures that grow
perpendicular to the least in-situ stress and that have historically formed the basis of standard
design models are the easiest to initiate and close easily when injection pressure ceases. This
type of fracture would probably not retain significant water except by the leak-off mechanism
91
-------
previously discussed since most water would be easily expelled upon closure of unpropped
fractures after stimulation pressures are released.
In contrast, the sliding shear and twisting type fractures are created by shear forces resulting
from higher pressure and both tend to close less readily than tensile fractures. These sliding
shear and twisting fractures can also act to prevent tensile fracture closure. These shear
fractures are often created by interaction between planes of weakness due to the pre-existing
natural fracturing and they act to locally divert the fracture into complex branching patterns
where significant fluid can be stranded in complex fracture networks (Dahi-Taleghani, 2009).
When the width of the branch fractures is narrow enough, the capillary forces can also
contribute to keeping the water permanently trapped in place.
Another proposed water trapping mechanism occurs when the fluid is trapped in proppant
packs while gas flows around the pack instead of through them (Daneshy, 2010). Since the
fracturing fluid is higher density than the gas, any mobile water will displace to the bottom of
the fracture and within a short period of time, the local flow velocity is not sufficient to lift the
fracturing fluid. This leaves the water trapped at the fracture bottom and may partially explain
why water recovery is much more efficient early in production while fluids are moving at higher
rates (Daneshy, 2010).
In addition to retention mechanisms, the fracturing pressures which could potentially drive
fluid from the target shale formation toward an aquifer are only applied for short periods of
time (hours for each stage), while the required travel time for fluid to flow from target to a
shallow aquifer under continually hydraulically induced pressure gradients is measured in
years. This is the inverse of an injection disposal well where injection pressure is continual and
ever-increasing. Calculations done by ICF showed that the maximum rate of seepage under
continual injection pressure in the absence of fractures through strata lying above the target
shale zone under ideal flow conditions would be substantially less than 10 ft/day or 5 in/hr of
pumping time during sustained stimulation pressures (ICF, 2009). Even if the water could be
moved, the developable shale formations are typically separated from potential shallow
aquifers by thousands of feet of relatively impermeable laminated strata requiring decades or
more of continual pressure application to move fluid that distance. Continually decreasing well
pressure during production also implies that eventually there would be little energy to move
fluid through the low permeability shale and thus trapped for geologic time. Water outside the
zone of influence would have neither permeability pathway nor sustainable pressure
differential to move fluid either.
Additionally, the volume of fluid used to fracture a well could only fill a small percentage of the
void space between the deeper target formation and the shallow aquifer. The already highly
diluted chemicals (typically about 1,000 gallons per million gallons of fresh water) would be
further diluted by the formation water and the void space above. Assuming an average of 10%
porosity above the target zone, the void volume for each 1,000 foot column below the aquifer
would be greater than 32 million gallons per acre (ICF, 2009). Obviously, the deeper the target
92
-------
zone, the higher the void volume and dilution factor. A typical Marcellus well with 160 acre
drainage (8 wells per 1280 spacing unit) at 3,000 feet of separation from the USDW would
contain approximately 15.4 billion gallons of saline water thus introducing a significant dilution
factor on the already highly diluted stimulation fluid. Presumably some of the chemicals in the
additives used in hydraulic fracturing fluids would also be adsorbed to and bound to the
organic-rich shales or decompose with time, temperature and fluid interactions within the hot
saline formation waters.
From a fluid flow perspective, any chance of flow toward an aquifer through the remote chance
of a conductive open fault to the shallow aquifer or through an unplugged wellbore would be
reversed when the horizontal well is produced subsequent to fracture stimulation. Any residual
fluid would be further flushed by flow back toward the production zone and into the well bore
as pressures decline in the reservoir during production. In any event an experienced stimulation
engineer would recognize any unusual occurrence as an anomalous change in injection rate and
pressure thus forcing stimulation cessation. The historical experience of hydraulic fracturing in
tens of thousands of wells is consistent with the analytical conclusion.
Even though stranded water is likely either immobile or directed toward the producing well, it
is prudent to examine the potential height of induced fractures. Prior attempts to address
fracture height containment were focused on formation elastic properties and the theory that
the higher Young's modulus can act as a barrier to fracture propagation. More recent
experiments indicate that elastic moduli contrasts are insufficient to stop vertical growth
although they may help in redirecting and changing fracture width and conductivity. Daneshy
and others have argued that blunting at the fracture tip, especially in naturally fractured shales,
is a more plausible explanation for observed fracture height containment that typically exceeds
modeling predictions (Gu, 2008). The mechanism is not yet well modelled but has been
observed in coal mines and indicated by microseimic mapping (Daneshy, 2010). Other authors
postulate that in-situ stress in layered formations (cookie effect) is the more controlling
fractures resulting in shear dampening. (Lewis, date unknown)
While the mechanisms for controlling fracture height have been widely discussed and
investigated, direct evidence that induced fracture heights are limited is now abundant through
micro-seismic monitoring. This was recently illustrated convincingly by Fisher for both
Marcellus and Barnett shale stimulations where an extensive micro-seismic database
demonstrated consistently large separation between the deepest groundwater sources and the
shallowest induced fracture (Fisher, 2010). This is not surprising given theories that shear
failure (slippage) results in blunting of the fracture tip thus limiting vertical growth via fracture
reorientation near an interface. However, while such newer theories minimize the historical
Young's modulus contrast contribution, it is interesting that the Fisher data show Marcellus
fracture growth essentially confined within roughly the bounds of the underlying Onondaga
limestone and overlying Tully which had been historically theorized as bounding layers because
of high elastic modulus. Whether this is simply coincidence is unknown.
93
-------
In any event, induced fractures do not appear to extend far above the target zones and the
injected water that does not return to the surface through production is likely trapped by a
combination of capillary, geo-mechanical-proppant stranding and adsorption mechanisms
which render the injected water essentially immobile. Unlike shallow coal seams where fluids
may be injected in close proximity to the aquifers and thus may migrate through the aquifer,
the deeply buried shales have extremely low permeability and low vertical fluid flow potential.
Once the stimulation pressures, lasting only a matter of hours to days are released, fluid flow
within the zone of influence is toward the wellbore pressure sink so that even if water was
mobile it would move preferentially toward the wellbore.
While development depth is a factor to consider in high volume stimulations, operators indicate
that that large volume fracture stimulation of shales at depths less than approximately 3,000
feet are unlikely and even if small stimulations eventually occur at shallower depths, the
induced fractures begin to curve from vertical to horizontal because the least principle stress
direction rotates to vertical at depths less than approximately 2000 feet. Some states such as
NY have already proposed a site specific review of large volume stimulations (>80,000 gallons)
whenever the target formation is less than 2000 feet deep or within 1000 feet of the deepest
potential fresh water supply. Extensive ongoing water well testing is also proposed to monitor
any changes in the drinking water sources.
References
Boyce, Matthew and Carr, Timothy: "Lithostratigraphy and Petrophysics of the Devonian
Marcellus Interval in West
Virginia and Southwestern Pennsylvania,"
http://www.mapwv.gov/UnconventionalResources/marcellusLithoAndPetroPaper.pdf (October 18,
2009).
Dahi-Taleghani, Arash and Olson, Jon E.; "Numerical Modeling of Multi-Stranded Hydraulic
Fracture Propagation:
Accounting for the Interaction Between Induced and Natural Fractures" SPE 124884 This paper
was prepared for presentation at the 2009 SPE Annual Technical Conference and
Exhibition held in New Orleans, Louisiana, USA, 4-7 October 2009.
Daneshy, Ali: "Why Care About Fluid Recovery," E&P Magazine online,
http://epmag.eom/Magazine/2010/6/item60863.php, (June 6, 2010).
Daneshy, A.: "The Factors Controlling the Vertical Growth of Hydraulic Fractures," SPE Hydraulic
Fracturing
Technology Conference Proceedings, The Woodlands, (January 19-21, 2009)
Engelder, Terry: "Unconventional Natural Gas Reservoir Could Boost U.S. Supply," Penn State
Live (January 17, 2008).
Engelder, Terry, Lash, Gary,. Uzcategui, Redescal S; "Joint sets that enhance production from
Middle and Upper Devonian gas shales of the Appalachian Basin," AAPG Bulletin; July
2009; v. 93; no. 7; p. 857-889.
EPA. "Evaluation of Impacts to Underground Sources of Drinking Water by Hydraulic Fracturing
of Coalbed Methane Reservoirs", EPA 816-R-04-003, June 2004
94
-------
Fisher, Kevin. "Data Confirm Safety of Well Fracturing," American Oil and Gas Reporter, (July
2010).
Gu, Hongre and Siebrits, Eduard: "Effect of Formation Modulus Contrast on Hydraulic Fracture
Height Containment," SPE Production and Operations Journal, Volume 23, No. 2, pp
170-176, (May, 2008)
ICF Report to NYSERDA, "Well Permit Issuance and High Volume Fracturing to Develop the
Marcellus Shale and Other Low Permeability Gas Reservoirs," Agreement 9679, (August
7, 2009).
Kalfayan, Leonard et al: "Optimizing Surfactants to Improve Stimulation Flowback in Tight Gas
Wells," World Oil (November 2008).
Lewis, Rick: "What Makes a Good Gas Shale,"
http://spemc.org/resources/presentation_022510.pdf, Oklahoma City (date unknown).
Penny, Glenn, Pursley, John, Holcomb, John: "The Application of Microemulsion Additives in
Drilling and Stimulation Results in Enhanced Gas Production," SPE paper 94274 (April
2005.)
95
-------
Summary and Abstracts from Theme 3: Well Integrity
96
-------
Summary of Presentations from Theme 3: Well Integrity
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA. Any mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
Technical Presentations
The first set of technical presentations in this theme addressed pre- and post-HF well integrity
assessment methods.
Jim Bolander, Southwestern Energy, introduced Theme 3. Mr. Bolanderthen discussed the
factors that affect mechanical integrity of HF wells, as well as methods and techniques used to
assess internal and external mechanical integrity at different stages in the life of the well. He
described two main causes of mechanical integrity failures: cement channeling and casing leaks.
Mr. Bolander concluded that proper planning, assessment, and remediation throughout the life
of the well are key to maintaining mechanical integrity.
Talib Syed, ISA, Inc., described wellbore design and monitoring techniques that are used to
ensure well integrity before, during, and after stimulation. He discussed factors to consider in
casing design, methods for cement evaluation, as well as methods for evaluating internal and
external mechanical integrity, including the Ultrasonic Imaging Tool (USIT) and traditional logs.
Mr. Syed emphasized the importance of proper casing design, cement placement, and
continuous monitoring through the life of the well. He also recommended that wells for
refracturing be carefully selected and closely monitored.
The final set of technical presentations described case studies for mechanical integrity.
Lloyd Hetrick, Newfield Exploration Company, presented a hypothetical case study in an
unconventional shale play with multiple zones. Mr. Hetrick discussed potential mechanisms for
mechanical integrity failures at each stage of well construction, stimulation, and production. He
also described methods that could be used to assess mechanical integrity at each of these
steps, as well as potential remediation actions. He called attention to adjacent wells and the
impacts they could have on well integrity.
Briana Mordick, Natural Resources Defense Council, presented two case studies of risks to
drinking water from oil and gas wellbore construction. In Bainbridge Township, Ohio, a poor
cement job, the decision to proceed with fracturing despite the poor cement job, and an
overpressured annulus (due to shutting in the well) led to gas migration into a drinking water
aquifer. The gas then entered homes through domestic water wells completed in that aquifer,
resulting in an explosion in one house. In Garfield County, Colorado, natural gas and other
contaminants were found in a local creek, which lead to an investigation of water quality in the
area. While some of the contamination was likely caused by faulty cementing in gas wells, other
sources may have contributed to the contamination. Ms. Mordick described the challenges
97
-------
inherent in determining sources of contamination and recommended that future studies
investigate contamination risks from both drilling fluids and fracturing fluids.
Summary of Discussions Following Theme 3: Well Integrity Presentations
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA. Any mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
Corrective action. Options for corrective actions on wells depend on when the problem occurs
and what type of problem it is. According to some participants, if there is an issue during a
pressure test in the pre-fracturing stage, operators have the ability to shut down operations
and perform remediation. The following are corrective action options suggested by
participants:
For a shallow mechanical issue, operators can pull the casing and replace it. If it is a
cementing issue, there are remedial cementing options ("squeezing cement"), though
this adds additional risk due to squeeze perforations in the casing.
If a problem occurs during fracturing, the operator can immediately shut down the well.
In a horizontal well, stages can be isolated.
A participant asked for more information on shutting in a well. A participant responded that
when fracturing stops or a well is shut in, pressure immediately decreases, lowering the
potential for fluid flow. The participant added that keeping a well shut in for a period of time
lowers the pressure further, though this may lead to casing problems. Participants stated that
monitoring annular pressures over the life of the well is essential. A participant recommended
that monitoring, recording, and reporting should not stop during or after the well is shut in. A
participant suggested that the pressure fall-off curve can provide important information about
the fracture treatment.
Use of logs and other tools. A participant asked if logs are required by regulatory agencies, or if
operators only run logs when they suspect a problem. Participants explained that, depending
on the state, certain issues (such as casing leaks) must be reported. The operator must work
with the state to develop a remediation plan, which generally includes using logs to understand
the nature of the problem. Participants indicated that in other situations, logs are run
proactively. For example, in Arkansas, a new regulation requires monitoring cement placement
during the cement pumping phase and monitoring annular pressures during fracture treatment.
According to a participant, the use of cement evaluation tools is also a standard process in new
development areas. The UltraSonic ImagerTool (USIT) is widely used in the Alaskan North Slope
in various kinds of production wells. The participant stated that the USIT is especially useful in
water-alternating-gas enhanced recovery wells, which can have well integrity issues not
detectable with conventional MITs.
Pressure test slope interpretation. A participant asked if a negative net pressure always means a
fracture has gone out of zone. Another participant stated that the net pressure plot is a key tool
in the field, and it is essential for understanding the qualitative analysis of pressure test slope
98
-------
changes. This participant explained that a slightly positive slope indicates fracture extension,
though a negative slope occurs at the very beginning of height growth. According to the
participant, a negative slope after the initial height growth indicates the fracture is potentially
growing out of zone; in severe cases, a vertical slope will indicate a pressure out or screen out.
In these situations, however, the participant stated that pressure and volume constraints will
halt fracture growth. A participant added that if fracture height is very large, the pressure
signature may no longer indicate fracture growth. Another participant stated that not every
negative pressure response indicates uncontrolled vertical height growth.
Property transfer. A participant asked what information is available to operators who are
acquiring new properties and wells. Other participants responded that operators have access to
well-by-well records from the previous operator. In addition, information is available from the
state. Operators also visit the site. A participant stated that, in general, the purpose of these
well record reviews and inspections is determining the potential value of the property, not
searching for defects in the wells.
Water supply wells and water quality. A participant asked whether nitrates in drinking water
wells could be due to faulty water well construction. A participant stated that this is possible,
though it is more likely that the presence of nitrates is due to natural recharge or infiltration
processes (from agricultural sources). Other participants emphasized the importance of good
construction of water wells. They noted that water wells often take water from across many
zones and are not subject to much regulation; in addition, contamination is often introduced
into water wells due to their poor construction. One participant noted that ground water is
often contaminated before HF activities take place, though the public generally does not realize
this. For example, benzene, toluene, ethylbenzene, and xylenes (BTEX) and endocrine
disrupters may come from nearby waste water treatment plants or gas stations. Another
participant noted the long history of naturally-occurring natural gas in water wells in Garfield
County, Colorado. A participant noted that, because of these concerns, it is in everyone's best
interest to gather baseline water quality data; it protects both residents and industry and
provides important data for regulators and scientists. A participant recommended that EPA
involve hydrogeologists in the study and take into account chemicals that are already present in
ground water.
Neighboring fields and fracture contact. A participant asked if operators working on
neighboring fields communicate with each other. Responses indicated that operators do
communicate about their activities and coordinate with each other, especially in areas with a
high potential for fracture contact. For example, operators may shut in wells to create a
pressure barrier to other fractures. One participant noted that in the Barnett and Haynesville,
operators may postpone fracturing activities to avoid impacts to certain procedures being
conducted at neighboring sites. Another participant stated that his company may delay well
completions within a single field to avoid interference. In addition, participants indicated that
drilling programs are planned so that new wells do not interfere with currently producing wells.
Participants added that state regulators are aware that fracture contact occurs and that
operators manage the situations within the industry. However, operators do not have a right of
99
-------
refusal for fracturing in neighboring areas (i.e. operators cannot force other operators to refrain
from drilling nearby). According to participants, fracture contact and well communication are
well-to-well issues, not well-to-surface issues.
Small operators. A participant asked whether small operators use the same best practices as
larger companies. Participants indicated that best practices are generally used industry wide,
though they may be inconsistently applied by both large and small companies. One participant
stated that smaller operators have more at stake in a single HF job, because shut down or
litigation problems would have a larger effect on the company. Another participant added that
small companies approach operations in a wide variety of ways, both positive and negative.
Abandoned wells. Multiple participants referenced large numbers of abandoned wells with
unknown locations in Pennsylvania and West Virginia. The participants indicated that this is of
great concern to operators, because it creates the potential of fracturing into an old well that
may or may not be properly plugged.
The Garfield County, Colorado case study (reference to Mordick, "Risks to Drinking Water from
Oil and Gas Wellbore Construction and Integrity: Case Studies and Lessons Learned"). A
participant asked about the compositions of fluids analyzed in the Garfield County case study.
The presenter clarified that, in most cases, the produced gas and fluid were representative of
the Mesa Verde Formation, and little evidence of cross flow between the Wasatch and the
Mesa Verde Formations was found.
Multi-well pads and well spacing. A participant noted that well pads typically have multiple
wells and recommended that fracture modeling and the planned EPA study take this into
account. According to this participant, well integrity is important, but it is essential to consider
groups of wells that are operating close together spatially and temporally. The participant
emphasized that the fourth dimension, time, is very important. Other participants indicated
that operators are aware of these issues and take them into account in modeling. In addition,
participants stated that the same best practices generally apply to single and multiple wells.
Another participant added that multi-well pads reduce the environmental footprint of drilling,
though another participant noted that it greatly increases well density and the chance for
fracture interaction between the multiple, closely located wells.
Regulatory issues. A participant described the Colorado Oil and Gas Conservation Commission's
regulatory response to concerns about HF. The participant noted that state agencies and
regulators are continuously addressing new challenges. Another participant noted that state
regulations were updated in response to both the Garfield County, Colorado, and Bainbridge
Township, Ohio, incidents reference in the presentation given by Briana Mordick.
100
-------
Abstracts for Theme 3: Well Integrity
Abstracts were submitted to U.S. EPA by the presenters for use in this proceedings document.
Not all presenters submitted abstracts of their presentations.
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA. Any mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
101
-------
Assessment Methods for Well Integrity during the Hydraulic
Fracturing Cycle
James Bolander
Southwestern Energy
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Introduction
The objective will be to evaluate well integrity of casing and cement during the drilling and
completion phases surrounding hydraulic fracturing. Critical processes will be evaluated using
passive monitoring techniques (pressure and volume measurements) and direct mechanical
techniques to determine effectiveness of casing and cement to protect drinking water
resources.
As defined by the EPA Draft Study Plan, drinking water resources will include "any body of
water, ground or surface, that could currently, or in the future, produce an appropriate
quantity and flow rate of water to serve as a source of drinking water for public or private
water supplies."
The primary focus of this assessment will be to concentrate on well integrity during drilling and
completion activities associated with running and cementing of production casing operations,
completion activities including the hydraulic fracturing process and post-frac activities. Many of
the solutions discussed are based on a conventional cased and cemented completion; however,
most of the methods discussed will be applied over any type of well configuration.
The purpose of this paper will be to discuss assessment methods and will not expand into
remedial solutions to meet hydraulic fracturing or producing well criteria.
Pre-frac Evaluation
The first step to evaluating well integrity of the production string will be to monitor and
interpret the pressures and volumes associated with the primary cement job. Key issues to
review include test results of the field blend samples (if applicable), actual cement slurry
density, cement slurry volumes, pump pressure, fluid return volumes, displacement volumes
and lift pressure. Based on the well design, the amount and type of cement will be determined
to achieve zonal isolation and sufficient coverage for isolation above the zone(s) to be
completed. Knowing the design parameters (estimated TOC, hydrostatic pressure and
displacement volumes) are key in the on-site monitoring of the treatment. Ensuring that the
cement blend is correct and that the correct dry cement / mix water ratio is followed is a critical
factor to ensuring the proper quality of cement. Monitoring return volumes and lift pressure
will be the first indication of adequate coverage of the productive horizon, any hydrocarbon
strata or any strata containing protected water. Monitoring the displacement volume will allow
102
-------
the estimation of the cement quality at the casing shoe. Monitoring and evaluating these key
components of the cement job will assist in planning of the initial steps of a well's completion.
After the production casing has been set and cemented a priority needs to be confirmation of
the wellbore integrity prior to moving forward with perforating and the hydraulic fracture
processes.
This confirmation process involves measuring the presence and quality of the cement bond or
seal between the casing and the formation and confirmation of the mechanical / pressure
integrity of the casing or tubing.
Confirmation of cement presence and quality can be obtained using various wireline tools
which can confirm the presence, height, bond and overall quality of the cement. Based on the
results of the pressure and volume monitoring of the cement job, different steps may be
chosen to confirm that an adequate seal is present.
Case #1 - Proper density, proper returns, lift pressure and displacement observed during
primary cementing. If design was sufficient for isolation and field conditions are known, a
temperature log may be run which can determine and confirm the top of cement (TOC)
measuring the heat change of cement during the setting phase. Based on average curing time,
this log should be run within the first 8-24 hours of pumping. Another wireline log option
would be a conventional cement bond log (CBL). The CBL operates on an acoustic principle: it
transmits a signal and measures the time travel from a set distance from transmitter to
receiver. Understanding the travel time of free pipe and empirical standards based on pipe size
and cement type are key in understanding the quality of cement bond and isolation that is
present, as well as the TOC. It is recommended to allow the cement to set a minimum of 48
hours prior to running the CBL. If necessary, pressure can be applied to the casing during the
CBL procedure if a micro-annulus is observed between the casing and cement sheath.
Case #2 - Returns, lift pressure or displacement does not correlate with design criteria. Risk is
insufficient coverage or channeling which could jeopardize proper isolation of protected water.
If there are no shallow horizons which require coverage and sufficient cement height was
designed, a conventional CBL may be sufficient to determine if adequate bonding above your
zone of interest is present to maintain pressure control for hydraulic fracturing. If there are
concerns about top of cement and quality, a radial ultrasonic tool (CET, USIT, CAST-V) log may
be run. The radial ultrasonic tool uses a high-frequency sonic pulse which will give a full 360°
interpretation of cement quality. In addition, the ultrasonic tool also measures casing
parameters such as diameter and thickness to confirm casing design specifications.
Once top of cement (TOC) and quality have been verified, and are considered adequate for
zonal isolation and hydraulic fracturing activities, casing integrity will be addressed. Several
studies have indicated that a minimum of 10 feet of zonal isolation is required dependent upon
hole and casing size
103
-------
Casing integrity will be confirmed with a surface-applied pressure test. Based on design criteria
(Casing parameter- burst and maximum anticipated treating pressure) the casing and tree will
be tested to a pressure greater than the maximum anticipated treating pressure (MATP) with
an appropriate safety factor (Burst Safety Factor ~1.3 and/or not less than 500 psi greater than
MATP). The pressure test is conducted using a high pressure pump truck and water. With the
frac tree valve closed, the tree and casing are tested for an average test time of 30 minutes. The
pressure will be monitored and if a pressure drop is observed (10% range), the casing will be
removed from service until such time the casing demonstrates full pressure integrity.
If during the pre-frac assessment process, casing and cement integrity are deemed to be
insufficient, the well should be removed from service until remedial operations have been
completed to restore integrity. Once remedial operations have been completed, repeat the well
integrity assessment to determine casing and cement integrity to confirm adequate pressure
and zonal integrity will be achieved to perform hydraulic fracture operations and well
production operations.
During Hydraulic Fracturing Treatment
Continuous monitoring of key parameters during the frac treatment (surface injection rate and
pressure and annuli pressures) is important in the continued monitoring of well integrity. These
key frac parameters are important in the evaluation of the post frac analysis (height, length and
conductivity) they are also important in the monitoring of well integrity.
Surface injection pressure is a component in the calculation of net pressure (BHTP - PC), which
is an important monitoring tool to determine if there is a loss of well integrity during the frac
treatment. A negative slope of the net pressure plot is indicative of excessive frac height
growth. This could be attributed to break out of zone and/or confining layer (discussed in
previous Workshop Theme) or loss of cement integrity during pumping. If there is a loss in
cement integrity, a corresponding spike in annular pressure may be observed.
In addition, monitoring of annular pressures may also indicate a breach in the casing which
could result in potential exposure of protected water.
If during a hydraulic fracturing treatment, there is reason to suspect any potential breach in the
production casing, production casing cement or isolation of any sources of protected water,
cease pumping and perform diagnostic testing on the well as is necessary to determine if
breach actually occurred and if remedial operations are required to restore well integrity.
During the frac job process, additional assessment methods may include evaluation of
microseismic events near the wellbore which may indicate a loss of cement integrity.
Other evaluation techniques such as use of tracers (chemical and radioactive) are important in
the planning and execution of the hydraulic frac treatment but will be discussed in the next
section.
104
-------
Post -frac Evaluation
Similar to pre-frac assessment, post-frac evaluation involves both passive monitoring
techniques and direct measurements.
Passive monitoring during the post-frac period includes continuous monitoring of well
production rate and pressure data, and fluid and gas compositional data in the flowback and
production stages.
Monitoring of rate and pressure data:
o Monitor flow rate changes that are anomalous to the wells typical behavior
which may include the following: change in gas/liquid rate which could indicate
an influx from an external source due to a breach in the casing or tubing.
o Flowing pressure changes can also be affected due to influx from an external
source and should be consistent with rate changes.
o Monitoring of annular pressures is important throughout the life of the well from
initial flowback until abandonment. Changes should be noted and corrective
actions taken, if necessary. As stated in API Guidance document HF1, "maximum
and minimum allowable annular pressures should be assigned to all annuli and
these should consider the gradient of the fluid in each. These limits establish the
safe working range of pressures for normal operation in the well's current
service and should be considered "do not exceed" limits."
Fluid and gas compositional analysis may also be utilized to monitor for changes in
characteristics. An example would be influx of fluids from an external source which
could change the flowback/produced fluids base characteristics such as total dissolved
solids (IDS). In addition, regular fluid compositional analysis recorded on a well can aid
in the determination of scaling and corrosion tendencies.
Continuous monitoring of pressure, flow and gas/liquid is an important tool in the maintenance
of a well. In addition to monitoring the above parameters, regular inspection of the wellhead
assembly and equipment removed from a well during a workover operation to inspect for leaks
and/or corrosion/erosion damage.
Mechanical methods of evaluating well integrity may involve the running of tracer logs after the
hydraulic fracture treatment or the running of mechanical and/or electromagnetic inspection
tools to evaluate the condition of the tubing and casing in the well. Additional logs may be run
which can detect flow behind pipe or a production log which confirms flow pattern within
wellbore.
To aid in the post frac analysis of the effectiveness of a well's hydraulic fracture treatment, the
job may be traced using radioactive tracers throughout the treatment to confirm the placement
of the fluids and proppant during the job. A multiple isotope gamma-ray (GR) tool is run in the
well after the treatment to measure the location of the isotopes to confirm placement within
the perforated interval. The tool is limited to measurements near the wellbore (<2') which can
also be used to determine any channeling behind the casing during the fracture treatment
which could compromise well integrity.
105
-------
Another tracer method is the use of chemical tracers in the hydraulic fracturing fluids. Specific
chemical tracers can be placed in the frac fluid at different stages to confirm flowback of fluids
from different stages. This confirmation can be used to determine if all frac stages are
contributing and can also be used to fingerprint flowback fluids, if necessary.
During the life cycle of the well, regular maintenance may be required which includes workover
operations in which tubing and packer installed in the well will be pulled out. Visual inspection
of the equipment is important as mentioned above to document the condition of the
equipment. In addition, mechanical inspection logs may be run to verify the condition of the
casing. A mechanical multi-finger caliper log can be run which physically measures the internal
diameter of the casing and its condition. Depending upon casing ID, the caliper tool may record
as many as (64) measurements of the internal diameter measuring changes in ID which would
detect corrosion pitting and possible holes or splits in the pipe. Electromagnetic flux and
ultrasonic tools can be run which will measure the changes in internal diameter as well as
casing thickness.
These inspection tools can be run throughout the life of the well to document changes of the
casing's condition over time. Understanding the condition of the casing over time is important
in the planning of future operations such as refracturing and/or recompletions in the well to
maintain well integrity over the well's life cycle.
Conclusion
There are many techniques available from passive monitoring to use of mechanical tools to
monitor the integrity of the well throughout the well's life cycle. Proper planning and
documentation is important to maintain well integrity and ultimately protection of the
environment.
References
American Petroleum Institute, API Guidance Document HF1, "Hydraulic Fracturing Operations -
Well Construction and Integrity Guidelines", October 2009
U.S. Environmental Protection Agency, "Draft Plan to Study the Potential Impacts of Hydraulic
Fracturing on Drinking Water Resources", February 2011
Various Vendor Technical Brochures and Technical Specification Sheets on various wireline
tools (Schlumberger, Baker Hughes (Atlas) and Protechnics)
106
-------
Pre and Post Well Integrity Methods for Hydraulically
Fractured/Stimulated Wells
TalibSyed, P.E.
ISA, Inc.
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Wellbore integrity is important to ensuring that reservoir formation fluids are brought to the
surface in a controlled and safe manner, and do not migrate into overlying fresh water
aquifers/underground sources of drinking water (USDWs). This paper will look into wellbore
design and monitoring techniques that are critical in assuring that wellbore integrity is
maintained in conjunction with hydraulic fracturing/stimulation completion practices.
The subsurface zone or formation containing hydrocarbons produces into the well, and that
production is contained within the well all the way to the surface. This containment is what is
meant by the term "well integrity". NORSOK D-010 defines well integrity as "Application of
technical, operational and organizational solutions to reduce risk of uncontrolled release of
formation fluids throughout the life-cycle of a well". Wellbore integrity as related to hydraulic
fracturing can be divided into three areas: pre-hydraulic fracturing design and completion
aspects to ensure wellbore integrity; techniques to verify that wellbore integrity is maintained
post-hydraulic fracturing; and the potential impact on long-term wellbore integrity (casing and
cement) from re-fracturing stimulations.
Well Design and Construction
Casing Setting and Design
As is required in all engineering designs, surface equipment and down-hole tubular are
designed for the anticipated operating pressures. This design requirement results in the proper
selection of appropriate casing and tubing grade and weight to avoid wellbore collapse. There is
a higher risk of compromising the casing integrity during drilling operations. The following
points should be considered in casing design (NORSOK 2004):
Planned well trajectory and bending stresses induced by doglegs and curvature
Maximum allowable setting depth with regards to kick margin
Estimated pore pressure development
Estimated formation strength
Estimated temperature gradient
Drilling fluids and cement program
Estimated casing wear
Setting depth restrictions due to formation evaluation requirements
Isolation of weak formations, potential loss circulation zones, sloughing and caving
107
-------
Metallurgical considerations
Potential for H2S and C02
Equivalent circulating density (ECD) and surge/swab effects due to narrow clearances
Geo-tectonic forces applicable
The casing is exposed to different loading conditions during various well operations (landing,
cementing, drilling, production). It has to be designed to withstand tensile, burst, and collapse
loads. Since it is impossible to predict the magnitude of these loads during the life of the casing,
the design is based on a worst-case scenario. The casing rating also deteriorates with time
(wear and tear). Therefore, safety factors are used to make sure that the casing could withstand
expected loading conditions.
Collapse pressure is mainly due to the fluid pressure outside the casing (due to drilling fluid or
cement slurry). Overpressure zones could also subject the casing to high collapse pressure. The
casing's critical collapse strength is a function of its length, diameter, wall thickness, Poisson's
Ratio etc. Burst loading is due to the fluid pressure inside the casing. Severe burst pressure
occurs if there is a kick during drilling operations. The tensile stress originates from pipe weight,
bending load and shock load. The axial force due to pipe weight is its weight in air less the
buoyancy force. Bending force results when the casing is run in deviated wells where the upper
portion of the casing is in tension and the lower portion is in compression. Shock load is
generated by setting of the slips and application of hoisting brakes. The sudden stoppage when
casing is run generates stress waves along the casing string.
In addition to the three loading conditions described above, casing design should also consider
the likelihood of buckling, piston and thermal effects. Buckling results when the casing is
unstable (e.g. partially cemented). The casing string will exhibit a helical configuration below
the neutral point, resulting in rapid wear at the neutral point and eventually lead to casing
failure. Piston force is due to the hydrostatic pressure acting on the internal and external
shoulders of the casing string while thermal effects refer to the expansion or shortening of the
casing due to increase or decrease in temperature.
Cementing the Casing/Liner
The quality of the cementing operation is also critical in maintaining wellbore integrity. Besides
the selection of the proper cement systems, the placement of cement and the quality of the
cement job are critical elements in assuring the well's integrity. It is very important to
thoroughly circulate and clean out the well prior to cementing in order to prevent mud mixing
into the cement, causing cavities or channels, resulting in potential cement degradation and/or
creation of leakage pathways for the formation fluids.
Well deviation can also affect the quality and presence of the cement. Drilling mud is first
circulated in the hole to ensure that drill cuttings and borehole wall cavings have been removed
prior to running the casing. The mill varnish is also removed from the surface of the casing to
ensure that the cement will bond to the steel surface. Centralizers are used to ensure that the
casing is placed in the center of the borehole. For under-reamed or washed out holes, bow
108
-------
spring centralizers are used. After the cement slurry is pumped down-hole, a lighter drilling
mud follows. This results in the casing being under compression from a higher differential
pressure on the outside of the casing. Thus when the cement sets and drilling continues, the
casing will always have an elastic load on the cement-casing interface, which is essential for
maintenance of the casing-cement bond and to prevent channeling or micro-annulus effects in
the cemented annulus.
Many wells are subject to sustained casing pressures (SCP). The main cause is believed to be gas
flow through the cement matrix. The cementing problems that could result in SCP include: (1)
micro-annuli caused by casing contraction and/or expansion, (2) channels caused by improper
mud removal prior to and during cementing, (3) loss circulation of cement into fractured
formations during cementing, (4) flow after cementing by failure to maintain an overbalance
pressure, (5) mud cake leaks, and (6) tensile cracks in cement caused by temperature and
pressure cycles (Sweatman, 2006).
Mechanical Integrity Methods for Production/Injection Wells
In the United States, every production and/or injection well is required to demonstrate that it
has sound mechanical integrity prior to it being placed on production/injection. Statutes and
regulations have been implemented in every state to ensure that oil and natural gas operations
are conducted in a safe and environmentally responsible fashion and wellbore integrity is
maintained throughout their operating life-cycle. The regulatory requirements for injection
wells as codified under 40 Code of Federal Regulations (CFR) Parts 144 through 148 require that
the injection well demonstrate that it has both internal mechanical integrity (no leaks in
tubing/packer or casing) and external mechanical integrity (all injected fluids are exiting the
permitted injection interval and that there is no upward migration behind pipe due to
channeling or a bad cement job/micro-annulus etc.). Leakage out of the production/injection
zone into overlying USDWs could occur due to poorly cemented casing, casing failure,
improperly plugged and abandoned wells or other artificial conduits, and natural
fractures/faults etc. Cement that has properly set has very low permeability (approximately 10"2
m2) and no significant flow of formation fluids can occur unless the cement has degraded or has
not set properly. Casing failure could occur due to corrosion, erosion or improper design (Syed
etal, 2010)
Internal Mechanical Integrity
Throughout the life of a producing well and during fracturing operations, the well conditions
should be monitored on an ongoing basis to ensure integrity of the well and well equipment.
Maximum and minimum allowable annular surface pressures should be assigned to all annuli
(should be considered as "do not exceed" limits). Also, during initial drilling completion, positive
pressure tests of the casing, tubing and inner annulus (between tubing and casing above the
packer) are conducted. The required surface test pressure varies in each geologic area (but is
generally at least 0.25 psi/foot of vertical depth to the top of the packer and the inner casing
and may not exceed 70% of the minimum yield strength of the casing). A well has verified its
internal mechanical integrity if the total pressure loss within the test period is less than 10% of
109
-------
the initial test pressure and the pressure is stable (thermal stabilization effects). Thermal
stabilization can occur when liquids either expand or contract depending on temperature
differential, causing questionable test results. Pre-loading an annulus or using fluids that are
close to the same temperature as fluids in the well will help in mitigating this effect. The test
fluid is generally an inert non-corrosive fluid/water or in some instances it could be a 50-50 mix
of methanol/water, neat methanol or diesel (used in extremely cold environments for freeze
protection). Factors to consider when conducting such tests (also referred to as MITIA or SAPT -
Standard Annulus Pressure Test), is that when a liquid medium is used as the test fluid, the well
may pass the MITIA, but later when it is on gas injection, there may be slow annulus pressure
build-up (sustained casing pressure) that may not be easily detected over a long period of time.
Other factors to consider for a successful MITIA for wells include proper packer selection
(elastomers) and materials of construction for tubing and surface wellhead that will meet
production and/or injection service requirements.
External Mechanical Integrity
There are several techniques that can be utilized to verify that production fluids are contained
within the wellbore and that there is no upward flow behind the casing (due to
chanelling/micro-annulus etc.) that can impact overlying USDWs. Some of these techniques are
briefly discussed below (Syed et al, 2010).
Cement Evaluation
Acoustic cement logs are run to determine cement tops as well as the quality of the casing-
cement and cement-formation bonds. Acoustic bond logs do not measure hydraulic seal, but
instead measure the loss of acoustic energy as it propagates through casing. This loss of energy
M»«u»m»H position
Figure 18. Ultrasonic Imager (a) tool design
and (b) transducer position (Smolen, 1996)
is related to the fraction of the casing perimeter covered by cement. Two classes of sonic
110
-------
logging tools exist: (1) sonic (cement bond log/variable density log - CBL/VDL) or segmented
bond tool (SBT) and (2) ultrasonic (ultrasonic imaging tool - USIT) (Boyd et al, 2006).
The Ultrasonic Imaging Tool (USIT) is basically a continuously rotating pulse echo type tool, and
is an improvement over the Cement Evaluation Tool (GET) with nearly 100% coverage of the
casing wall. The processing of the echo is, however, quite different from the GET. The USIT is
shown schematically in Figure 18. The main working element is the rotating transducer
indicated as "sensor" on the bottom of the tool string. The transducer rotates, emitting and
receiving signals reflected back from the casing wall. The USIT tool is 3 3/8" in diameter and by
changing the rotating transducer subassemblies can operate in casing sizes from 4 V-i to 13
3/8". The rotating transducer is shown in Figure 18(b). In the measurement position it is aimed
toward the wall and in the fluid properties position it is aimed toward the target plate, with the
fluid properties measured when going in the hole. The USIT presentation uses highly
sophisticated computer processing and is color coded. It is very sensitive to the condition of the
borehole and is preferably run along with a CBL to provide best overall picture of well integrity.
An illustrative example of a USIT log is shown in Figure 19.
Acoustic impedance, Z, is defined as the product of the density (kg/m3) and acoustic
velocity (m/sec) of a medium and is expressed in MRayl (106 kg/m2 sec). A list of acoustic
impedance values for common down-hole materials is given in Table ITable 3.
Figure 19. Illustrative Example of USIT Log Run on Injection Well
ill
-------
Table 3. Acoustic Properties of Materials (Smolen, 1996)
Material
Air
Water
Drilling Fluids
Cement Slurries
Cement (Litefil)
Cement (Class G)
Limestone
Steel
Density
[Kg/nf]
1.3-130
1000
1000-2000
1000-2000
1400
1900
2700
7800
Acoustic Velocity
(m/sec)
330
1500
1300-1800
1800-1500
2200-2600
2700-3700
5500
5900
Acoustic Impedance
(MRayl]
0.0004-0.04
1.5
1.5-3.0
1.8-3.0
3.1-3.6
5.0-7.0
17
46
The Segmented Bond Tool (SBT) is a radial cement bond device, which measures the quality of
cement effectiveness, both vertically and laterally around the circumference of the casing. The
SBT is designed to quantitatively measure six segments, 60° each around the pipe periphery
and employs an array of high-frequency steered transducers which are mounted on six pads.
Each of six motorized arms positions a transmitter and receiver against the casing wall. The SBT
is usually run with a VDL (variable density log). A primary SBT presentation has (1) a correlation
trace and (2) two attenuation traces that are an average of the 6 segmented measurements
and a minimum attenuation trace representative of the 60° segment with the least attenuation.
A separation of the two attenuation curves indicates a cement void on one side of the casing
and a continuous wide separation over an extended depth interval infers the present of
channeling within the cement sheath. An example Segmented Bond Tool (SBT) log run on an
injection well is shown in Figure 20.
Figure 20. Example SBT Log
112
-------
Factors that Affect Cement Log Quality
There are many factors that affect the response of sonic logging tools. These
factors include: micro-annulus, logging tool centralization, fast formation arrivals,
use of lightweight cements and cement setting time (Boyd et a I, 2006).
Micro-annulus. A micro-annulus is defined as a very small (approximately 0.01 to
0.1 mm) annular gap between the casing and the cement sheath. A micro-annulus
can result in a misinterpretation of the CBL/VDL. Micro-annuli are caused by
temperature, mud-cake deposits, pipe coatings and constraining forces. A
common procedure is to pressure up the casing to approximately 1,000 to 1,500
psi and close the gap (if the cement job was good). Micro-annuli affect ultrasonic
tools much less than the CBL/VDL and SBT (pads) in the presence of liquid in the
gap with the opposite effect in the presence of gas.
Eccentralization. This may be an issue particularly in deviated and horizontal wells
with the absence of cement on the low side and the distance between the casing
Figure 21. and formation face is small.
MIT Tool
Logging Tool Centralization. It is mandatory that the USIT and the CBL/VDL tools
are well centralized. The SBT pads with their articulated arms are relatively unaffected by the
centralization issue, although the CBL/VDL part of the tools is affected. Tool centralization can
be checked in the log presentation.
Fast Formations. Formations with very high velocity and short transit time are called "fast
formations". Acoustic signals from anhydrites, low porosity limestone and dolomites often
reach the receiver ahead of the pipe signal. Fast formations affect the CBL/VDLs and SBT logs
but do not affect USIT interpretation.
Lightweight Cement. Cement evaluation relies on the contrast in the acoustic properties of the
cement and liquid. The acoustic properties of lightweight cement (commonly used in areas of
weak formation) are close to those of cement slurry making it difficult to distinguish between
the two.
Cement Setting Time. This is an important consideration in CBL interpretation. If the bond log is
run before the cement is fully set, a misinterpretation indicating poor bonding may result in an
unnecessary squeeze operation. The hardening time of cement slurries depend on their type
and formulation, the down-hole temperature profile and pressure conditions, and extent of
drilling mud contamination. The U.S.EPA recommends a 72 hour waiting on cement (WOC)
prior to logging UIC regulated wells, while the American Petroleum Institute (API) and the
Alberta Energy and Utilities Board suggest a 48 hour WOC time (for oil and gas related
production and injection wells). The ultrasonic cement analyzer (UCA) can be utilized to
determine when to log and has shortened the WOC time.
113
-------
To declare zonal annular isolation between two points behind casing, a minimum length of
continuous good quality cement should exist. A recommendation of 33 feet of continuous good
cement for the 7 inch casing and for 45 feet for 9 5/8 inch casing has been reported in a EPA
publication, while oil industry service company recommendations for continuous good quality
cement are 10 to 11 feet for 7 inch casing and 15 feet for 9 5/8 inch casing, to assure zonal
isolation (Boyd et a I, 2006).
Finally, it should be noted that even if cement quality logs indicate good bonding and zonal
isolation, there may be annular communication resulting from reactions between the rock,
cement and formation fluids in production wells.
Zone Isolation/Pressure Testing
Placement of the cement completely around the casing and at the proper height above the
bottom of the drilled hole (cement top) is one of the primary factors in achieving successful
zone isolation and integrity. It is good practice to pressure test the shoe after drilling out the
cement shoe on the surface and intermediate/longstring casing strings and confirm zonal
isolation at the shoe. This involves pressuring up inside the casing until the pressure at the shoe
exceeds the maximum hydrostatic pressure expected at that point during subsequent drilling
operations. Failure of cement around the shoe is usually due to contamination, either from the
original drilling mud or from the displacement fluid and usually results from poor cementing
techniques rather than poor quality cements since hard-set neat cement has sufficient strength
to withstand pressure tests.
Multi-finger Caliper Surveys
Multi-finger caliper logs (multi-finger imaging tools - MIT) are used to detect very small changes
to the internal surface condition of tubing from the impacts of corrosion and/or mechanical
damage. The tool may be run through tubing to log casing deeper in the well. They are available
in 24, 40 and 60 fingers or arms (tool diameters of 1.6875, 2.75 and up to 4.4 inches
respectively) to suit varying casing/tubing sizes. The number of fingers increases with the
diameter of the tool and when the tool is run in the hole, the fingers are closed to prevent
damage. Tool deployment can be via slick-line, e-line, coiled tubing or down-hole tractors. The
magnetic thickness tool (MTT) uses 12 miniature magnetic sensors, to investigate variations of
metal thickness within down-hole tubular. Data from the multi-finger imaging and magnetic
thickness tool can be combined to assess both the internal and external condition of the
tubular including maximum cross-sectional wall loss, maximum penetration (pitting etc.) and
reduction in wall thickness. A representative MIT and MTT tool is shown in Figure 21 and Figure
22, and an example multi-finger caliper survey run on an injection well is shown in Figure 23.
114
-------
Figure 22.
Magnetic
Thickness Tool
(MTT)
Figure 23. Example of Multi-Finger
Caliper Survey
Oxygen Activation/Water Flow Log/Hydrolog
Oxygen Activation logs also referred to as a Water Flow Logs (WFL) or Hydrologs are used to
detect water flow or channels behind casing in injection or production wells. The principle of
water detection using Oxygen Activation can be explained as follows - when the neutron burst
is generated by the tool, the oxygen
associated with the up-flowing water is
activated to an unstable nitrogen isotope
having a half-life of 7.35 seconds (oxygen
activation effect).When the nitrogen isotope
returns to its native oxygen, gamma rays are
emitted which may be detected by the near or
far background count measurement. The
times under consideration are long after the
inelastic or capture gamma rays have ceased.
The WFL is a dual burst TDT (thermal decay
time) with a modified pulse sequence. Unlike
a conventional TDT log, the OA/WFL needs to
be run centralized. The operation of a WFL is
shown in Figure 24. The neutron generator is
turned on for either 2 or 10 seconds, then
VOUJMfTH1C FLOW
RATE INCREASING
WTTM AREA
JO 30 40
TIME - SECONDS
Figure 24. WFL Measurement Technique
115
-------
turned off. If no water flow is present, then
the count rate decays as shown, reaching
background after about one minute. If
water flow is present, then the count rate
decays as before, until the activated water
moves adjacent to the detector. When that
occurs, excess counts are observed. After
the cloud of activated water passes, the
counts return to the background decay
curve. The data are recorded on three
detectors, typically the near (N), far (F), and
gamma ray (GR). Only one will be typically
optimized to provide good data. While each
burst and decay sequence takes about 1
minute, the data collected may be highly
statistical, and therefore the burst and
decay sequence will typically be repeated
up to about 10 to 15 times. Figure 25 shows
a WFL run on a well in Alaska.
Borax PNL Logs Figure 25. Example of WFL Log
Channel detection using temperature or noise logs is often ambiguous. In certain areas,
radioactive (RA) tracers cannot be used either due to safety, environmental, or political
reasons. As a result, a technique based on the higher capture cross section of boron has been
developed in Alaska to locate channels behind pipe. The borax compound generally used is
sodium tetra-borate penta-hydrate (Na2B407), due to its high capture cross section, low cost,
and ready availability. The mix rate used in Alaska is 7 pounds/barrel of warm seawater.
The Borax PNL technique involves comparing pulsed neutron log (PNL) passes run before and
after pumping a solution of borax dissolved in warm water as a tracer. A PNL indicates a
significant Sigma value when boron is present, so an overlay of log passes quickly indicates
those areas within and adjacent to the wellbore where boron accumulates due to injection of
the tracer. An illustrative example of a Borax-PNL log run in Alaska is shown in Figure 26.
Ultrasonic Leak Detection Logs
A new tool that has demonstrated success in the North Slope of Alaska in detecting leaks as
small as 0.0024 gallons per minute (gpm) is the ultrasonic leak detection logging tool run on
wire-line or on slick-line in memory mode (Julian et al, 2007). The tool is particularly useful
where rig workovers are expensive as in remote locations, offshore or in Arctic regions. It can
detect leaks through multiple strings because ultrasound is not significantly attenuated by gas,
liquid, or steel. Other advantages include: (1) it can be run in high pressure wells in which it is
difficult to maintain a pressure seal for the wireline, and (2) in memory mode a tandem multi-
finger caliper and a leak detection log can be obtained in one run. Many injection wells were
116
-------
previously producers and therefore have gas-lift mandrels. Ml gas consists of 35% methane,
20% each of ethane, propane, and carbon dioxide. Ml gas is an excellent solvent and easily
dissolves grease seals, o-rings, and elastomers. A schematic of the ultrasonic tool is shown in
Figure 27.
Figure 26. Example Borax-PNL Log
CJM« u 39 m 00^1 »
> ... !._
.'. . : .>'.' :- . .
I J25 m. OP* 1 «!>*
.,. . _ - i . . .
Knuato Jon (017m OO*1B° 3SOtt»l
(OH» OO-'W iCtai
WID »60m
'HHr 8'fctl
. :' Jl ',. . .-.
WLD 10 Mm, 00=169". 8.1 t»l
IT. on=' er i
HUM ofeM i
Figure 27. Ultrasonic Leak Detection Tool
Tree and Wellhead Integrity
The wellhead and tree are typically suitably engineered to withstand the normal operating
pressures. For normal operations and during hydraulic fracturing operations, if the annulus
between the production casing and the intermediate casing has not been cemented to the
surface, the pressure in the annular space should be monitored and controlled. The
intermediate casing annulus should be equipped with an appropriately sized and tested relief
valve. The relief valve should be set so that the pressure exerted on the casing does not exceed
the working pressure rating of the casing. Pressure exerted on equipment should not exceed
the working pressure rating of the weakest component.
Wellhead seal tests need to be conducted to test the integrity of the sealing elements
(including valve gates and seats) and confirm their ability to seal against well pressure. If
117
-------
abnormal annular pressures are noted, a re-pressure test of the wellhead system can help
determine whether it is a surface wellhead leak as opposed to a subsurface leak.
Horizontal Wells
In general, horizontal wells have had great success in high-permeability reservoir and
unconventional formations such as coal, chalk and shale. With the advancement of drilling and
completion technologies, horizontal wells have become the industry standard for
unconventional and tight formation gas reservoirs. Horizontal wells are commonly two to four
times more expensive to drill and complete than offset vertical wells, yet are theoretically
capable of up to three to five times the production. Environmental advantages with horizontal
wells include a smaller drilling footprint with a reduction of well locations.
Horizontal wells are typically drilled vertically to a "kick-off" point where the drill bit is gradually
turned from vertical to horizontal (see Figure 28). Horizontal wells use basically the same or
similar equipment as vertical wells such as safety valves, packers and seal assemblies, flow
control accessories, permanent down-hole gauges, artificial lift accessories etc. Tool
manipulation is hydraulic or with reciprocation, while rotationally actuated tools should be
used with caution. Intervention into the horizontal section requires coiled tubing, down-hole
tractors or workstring.
Horizontal wells are completed with various degrees of annular isolation. Un-cemented or
open- hole completions offer open access to fracture swarms, which may be plugged off or
inaccessible if annulus is cemented. With open- hole or barefoot completions the most
productive part of the interval has a better chance to be stimulated. Also, un-cemented
completions avoid perforation-related stress cages that can result in a large extraneous source
of treatment pressure drop. In this alternative, the producing portion of the well is the
horizontal portion of the hole and it is entirely in the producing formation. In some instances, a
short section of steel casing that runs up into the production casing, but not back to the
surface, is installed. Alternatively, a slotted or pre-perforated steel casing may be installed in
the open-hole section. These alternatives are generally called a "production liner" and are
typically not cemented in place. In the case of an open-hole completion, the tail cement should
extend above the top of the confining zone (the formation that limits the vertical growth of the
fracture).
Cased and cemented horizontal completions offer greater control over fracture treatment
placement and can be appropriate when dealing with relatively uniform rock. Where cemented
completions are warranted, sand jet perforating is preferred as it removes formation material
and thus avoids the stress cage related pressure drop.
Discontinuous multi-layer intervals such as stacked, fluvial-dominated sandstones are best
completed with vertical wells in multi-stage treatments.
118
-------
Figure 28. Example of a Horizontal and Vertical Well (API, 2009)
Conductor Casing
-Casifig
Intermediate Casing
Vertical Fractures in VerticaTwelts
Producing Formation
Hydraulic Fracturing
Hydraulic fracturing (HF) has been employed in the oil and gas industry since 1947 and allows
the production of hydrocarbons from low permeability (tight) reservoirs economically. The
process of hydraulic fracturing increases the exposed area of the producing formation, creating
a high conductivity path that extends from the wellbore through a targeted hydrocarbon
bearing formation for a significant distance, so that hydrocarbons and other fluids can flow
more easily from the formation rock, into the fracture, and ultimately into the wellbore.
During HF, fluid is pumped into the production casing, through the perforations (or open hole),
and into the targeted formation at high enough pressures to cause the rock to fracture; this is
known as "breaking down" the formation. As high pressure fluid injection continues, the
initiated fracture can continue to grow or propagate. The rate at which the fluid is pumped
must be fast enough that the pressure necessary to propagate the fracture is maintained. This
pressure is known as the propagation or extension pressure. As the fracture continues to
propagate, a proppant, such as sand, is added to the fluid. The proppant allows the fracture to
remain open when pumping is stopped (and the excess pressure is removed), allowing fluids to
flow more readily through this higher permeability fracture. During the HF process, some of the
fracturing fluid may leave the fracture and enter the untreated formation resulting in fluid leak-
off. The fluid flows into the micropores or pore spaces of the formation or may intersect
existing natural fractures in the formation.
119
-------
In order to carry out the HF process, a fluid must be pumped into the well's production casing
at high pressure. The production casing must be properly designed, installed and cemented so
that it is capable of withstanding the pressure that it will be subjected to during the HF process.
In some cases, a high pressure "frac string" may be used to pump the fluids, thereby not
exposing the production casing to the high treatment pressures. Once the HF process is
completed, the frac string is removed.
In the field, the HF process is called the "treatment" or "job" and consists of three stages:
Pad - The pad is the first stage of the job where the fracture is initiated and is
propagated in the formation. Another purpose of pad is to provide enough fluid volume
within the fracture to compensate for fluid leak-off into the formation.
Proppant Stages - Here proppants of varying concentrations are pumped. Most
common proppant is ordinary sand sieved to a particular size. Other proppants include
sintered bauxite and ceramic proppant.
Displacement - Here the previous sand laden stage is displaced to a depth just above
the perforations. This is done so that the proppant ends up within the fracture and not
within the pipe. Sometimes called the flush, the displacement stage is where the last
fluid is pumped into the well. The flush fluid could be plain water or the same fluid that
was pumped earlier.
In wells with long producing intervals (both vertical and horizontal), the HF process can be done
in a multi-stage process allowing for better control and monitoring of the HF process.
Post-Hydraulic Fracturing Monitoring
Prior to the HF treatment, the proppant, usually sand, may be "tagged" with a tracer. After the
proppant has been pumped into the formation, a cased-hole log, capable of detecting the
tracer, is run to confirm the proper placement of the proppant. A temperature survey in
conjunction with the tracer log can also be run. Since the HF fluid is typically at ambient
temperature at the surface and the formation temperature at the target depth is much higher,
the formation is cooled considerably during the HF treatment showing which perforations
accepted the fracturing fluid. The use of these techniques is declining with the advent of
sophisticated computer modeling techniques for mapping fracture growth and geometry.
Refracturing
Refracturing of oil and gas wells (also known as fracture re-stimulations) are becoming
increasing popular as this technique, under certain conditions, can restore or increase well
productivity and ultimate hydrocarbon recovery. Re-stimulations can by-pass near well-bore
damage and generate higher conductivity propped fractures resulting in more lateral extension
and deeper penetration of the fractures, with ultimate higher hydrocarbon recovery.
More than 30% of fracturing treatments are performed in older wells, therefore, mechanical
integrity of the tubular becomes critical in candidate selection for HF treatments. Surface casing
120
-------
vent flows must be checked and any indication of gas migration to the surface will result in the
elimination in the well as a candidate.
References
American Petroleum Institute, API Guidance Document HF1, "Hydraulic Fracturing Operations -
Well Construction and Integrity Guidelines", October 2009
Boyd, D., S.AI-Kubti, O.H.Khedr, N. Khan, K.AI-Nayadi, D.Degouy, A. Elkadi, Z.AI-Kindi, "Reliability
of Cement Bond Log Interpretations Compared to Physical Communication Tests
Between Formations", SPE 101420, Abu Dhabi, UAE, November 5-8, 2006
Juilan, J.Y., G.E.King, J.E. Johns, J.K. Sack, D.B. Robertson, "Detecting Ultra-small Leaks with
Ultrasonic Leak Detection - Case Histories from the North Slope, Alaska", SPE 108906,
Vera Cruz, Mexico, June 27-30, 2007
NORSOK, "Well Integrity in drilling and well operations", Standard D-010, August 2004
Smolen, J., "Cased Hole and Production Log Evaluation", Penn Well Books, Tulsa, OK, 1996
Sweatman, R., "Studies on Wellbore Integrity", Proceedings of the 2nd Wellbore Integrity
Network Meeting, Princeton, NJ, March 28-29, 2006
Syed, T., and Cutler.!., "Well Integrity Technical and Regulatory Considerations for C02
Injection Wells", SPE 125839, Rio de Janeiro, Brazil, April 12-14, 2010
121
-------
Case Study for Well Integrity over a Full Life Cycle
Lloyd H. Hetrick, PE, CSP
Newfield Exploration Company
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Abstract
This case study narrowly defines well integrity by one simple outcome: the prevention of
vertical migration of fluids in order to protect drinking water resources. This paper should not
be considered a stand alone document, rather an extension of the well design, construction,
and surveillance practices which have already been addressed in this Workshop. A generic shale
development well is presented, beginning with its basis of design, then construction, an
operational phase, and ultimately its plug and abandonment. This chronology is illustrated by a
series of well schematics, which are provided in Appendix A.
Regulations, industry standards, and best practices will be addressed, as will failure categories
and relative failure rates at each phase of the well's life cycle. This case study will also raise
relevant issues that may not have been fully discussed during this Workshop, such as the
difference between exploration and development phases, development well economics, the
potential for well integrity impacts from adjacent well activities, and a time line perspective.
Introduction
A brief process description for oil and gas projects might be helpful. Years before a well is
drilled, significant geological and geophysical "G&G" work is performed to identify prospective
areas. During this time, offset wells are studied to identify subsurface hazards that may be
present in order to avoid or mitigate them. Once a prospect is defined mineral leases are
acquired, additional G&G and reservoir analysis performed, and well design determined for
specific drilling locations. The first group of wells drilled are called "exploratory" and intended
to define the commercial value of the prospect. Exploratory wells require extra time to gather
data on the quality of the reservoir and are also used to identify well construction efficiencies
for the development phase. Once the project transitions from exploration to development,
each well has to pass an economic hurdle to be drilled.
Regardless of being exploratory or development, responsible oil and gas companies have a
strong business incentive to protect the environment, mineral reserves, and the well itself (1). It
is almost always more difficult and costly to re-enter and repair a well than to address design
deficiencies up front during construction.
This case study, although generic, is not unlike the Marcellus, Eagle Ford, and other
unconventional plays with multiple hydrocarbon zones. Even though only one reservoir is the
current development objective, additional reservoirs are candidates for future development.
122
-------
This case study will address technical issues but cannot explore very many technical details due
to a fifteen minute presentation limit. Accordingly, only the most relevant technical items such
as failure modes will be included and even then, will be greatly abbreviated. For example, if
corrosion is considered to be the primary failure category, the technical discussion will end
there with no deeper look into the true root cause failure mode such as galvanic corrosion,
sulfide stress cracking, etc.
Federal and State environmental laws protect underground sources of drinking water or
"USDWs". This paper will use USDW synonymously with the term "protected water" and refers
to an aquifer with less than 10,000 mg/l total dissolved solids or "TDS" (2).
State mineral law regulates the extraction and conservation of minerals unless on Federal BLM
or BIA land, then Federal mineral laws apply. In either case, the regulatory agency that oversees
mineral extraction is also the primary regulator for protecting USDWs during oil and gas
exploration and production activities (3) (4).
Protected water and hydrocarbons have natural separation (5) in most situations. There are
however, areas of the country where methane is routinely found to exist naturally in USDWs (6)
(7) and has been associated with bubbles in rivers as early as the mid 1800s (8). There are also
locations where methane vents to the surface via natural pathways having nothing to do with
oil and gas extraction activities (9) (10). It has been estimated from a review of Pennsylvania
regulatory records that over 95% of the complaints that oil and gas activities had contaminated
private water wells were actually due to preexisting or other land use activities (11). These
naturally occurring migrations are not limited to methane, as towns named Oil Springs, KY (12)
Oil Springs, Ontario (13) and historical sites such as Seneca Oil Spring, NY (14) and Brine Springs,
TX (15) all attest that oil and brine have been observed migrating to the surface dating back to
the 1600's.
Basis of Design
A development well is drilled only if there is confidence that the estimated recoverable
hydrocarbon reserves will provide an acceptable economic rate of return, given the cost to
construct and operate the well. For an unconventional gas play, development wells tend to
have generational designs where a group of wells will have a similar drilling, casing, cementing,
perforating, and hydraulic fracturing design. Over time as more wells are drilled, experience
provides opportunities to correct any design deficiencies, improve drilling efficiencies and well
performance, therefore subsequent generations of wells are seldom designed exactly the same.
Individual wells, regardless of their generational status, receive detailed engineering analysis
and planning which is communicated to the wellsite supervisor in the form of a written drilling
and completion procedure. These well specific procedures are a planned sequence of activities
which also incorporate regulatory compliance and industry best practices.
123
-------
Well Construction - Drilling
A typical onshore well is spud with a conductor pipe that is driven, drilled, or augered into the
ground by a construction crew or "spud rig" prior to the drilling rig's arrival. This conductor pipe
is a structural component that sometimes is not needed at all. Conductor pipe most often does
not reach the top and does not penetrate the base of protected water; therefore it is not
involved in protecting USDWs from vertical migration of fluids. Accordingly, failure categories
for the conductor pipe will not be discussed.
The surface hole is drilled to a prescribed depth below the base of protected water. This depth
is most often provided by the State Oil and Gas Regulator as in Oklahoma (16), or the State
Environmental Protection Regulator as in Texas (17), or not specifically provided other than to
protect all USDWs encountered as in Pennsylvania (18). In this latter situation, oil and gas
operators typically research a Pennsylvania Groundwater Information System "PaGWIS"
database and local water well driller's records to generate a hydro geological map in order to
determine depths of water that need to be protected.
The surface hole is not left open for more than a few hours while being drilled, cased, and then
cemented back to surface. Those zones left open during this brief period are all USDWs, so
vertical migration of fluids does not present a significant threat during surface hole drilling. The
surface hole on our case study well is drilled in a few hours on the first day of the drilling
operation.
The surface casing string is the primary barrier to prevent fluids from the wellbore from
entering protected water as the well is being drilled to the next casing setting depth. Unlike the
conductor pipe, surface casing is always required and is typically specified by regulation to be of
"suitable and sufficient" quality (19) or "suitable for all drilling and operating conditions such as
tension, burst, collapse" (20). For all casing strings, industry best practices provide extensive
guidance on the selection of proper casing size, grade, weight, connections, plus procedures for
field handling, inspection, and testing (21) (22) (23) (24) (25) (26) (27). For our case study well,
the surface casing is "run" or installed in a few hours during day #1 of the drilling operation.
Failure categories for the surface casing and all other casing strings can be divided into the
following five categories (28). It should be noted that two of these categories, mechanical and
corrosion, may be secondary to cement failures where a failed cement sheath can lead to
buckling or external corrosion that would not have otherwise occurred. Failure categories, their
respective failure modes, relative failure rates, and remedial options will be discussed briefly:
Materials - defects, tolerance busts, not getting the quality of pipe specified
Connections - wrong connection selected for the service, improper makeup
Wear and Handling - internal wear from drilling, external damage from handling
Mechanical -tensile, burst, collapse, buckling, cyclic loading
Corrosion - internal vs external; galvanic, C02, sulfide stress, hydrogen induced cracking
124
-------
Materials defects are supplier dependent and can be managed by inspections and other supply
chain quality control efforts. Connection problems are most often related to improper makeup
and can be minimized by onsite supervision. Wear for the surface casing string is seldom a
concern and occurs as a result of other problems encountered while drilling the well.
Mechanical problems with the surface casing are very few when compared to deeper casing
strings that are exposed to higher pressures and temperatures. External corrosion presents the
highest failure category for surface casing. Remedies may include external coatings, cement
squeezes, and cathodic protection systems.
The surface casing string's cement job provides the primary barrier against vertical migration of
fluids into protected water for the entire life of the well. In the context of USDW protection, the
importance of getting a good primary cement job on the surface casing string cannot be
overstated. Remedial cementing options do not provide high success rates for zonal isolation
and should be considered only for contingency purposes. Of all regulations for onshore wells,
the rules for surface casing cementing contain the most stringent requirements for hole size vs
casing size, centralization, cement quality, cement quantity, cement placement techniques, and
quality assurance than for any other casing string (29). Failure to properly cement the surface
casing string triggers both agency notification and corrective actions (30). The surface casing on
our case study well is cemented on day #2 of the drilling operation.
There is a significant body of information published on cement selection and cementing best
practices (31) (32) (33) (34). There is also a significant body of information available on
cementing failure rates (35) (36). This Well Integrity Case Study will focus on those conditions
which directly relate to zonal isolation for the protection of USDWs, briefly discussing three
failure categories, with their respective modes and relative failure rates, and remedial options:
Insufficient cement volume - underestimated annular volume, lost circulation
Low bond strength - poor slurry design, poor management of hydrostatic head pressure
Micro annulus, cracking, plastic deformation - thermal and pressure effects, cyclic loads
Cement failure rates are directly proportional to the ability to evaluate the top of and quality of
the cement sheath. Cement tops can be identified by a temperature log, relative cement bond
quality can be identified by a Cement Bond Log or CBL, while absolute cement bond quality
requires a combination of logging, testing, and engineering analysis (37).
For all three cement failure categories, remedial options are not optimum and include pumping
in from the top, spotting from the top via a small work string, or by perforating and squeezing.
It should be noted that two of these three remedies, pumping in from the top and perforating
and squeezing, might add new problems for zonal isolation if not properly executed.
There is a strong correlation between gas migration and uncemented or poorly cemented
casing strings. There is also a strong correlation between external casing corrosion and the
absence of a good cement sheath (35) (36).
125
-------
After the surface casing has been successfully tested, the float collar, float shoe, and
approximately 10' of new formation are drilled. Another integrity test is then performed, a
Formation Integrity Test or "FIT" which tests both the casing shoe and new formation together.
This is not a leak-off test and does not test the limits of the shoe and formation, rather the FIT
provides an assessment of the wellbore's ability to withstand additional pressure in case of an
influx of fluids and allows for safer drilling to the next casing point (38).
The next sections of well, which for this case study includes an intermediate and production
casing section, are essentially a repeat of the surface casing section described above except
that:
The design depth for intermediate and production casing strings are not as
comprehensively regulated (as for the surface casing depth) other than to provide safe
drilling operations and to prevent the waste of minerals.
The regulations concerning hole size vs casing size, centralization, cement quality,
cement quantity, cement placement techniques, and quality assurance for intermediate
and production casing strings are not as specific (as for the surface casing) other than to
provide safe drilling operations and prevent the waste of minerals.
Although this case study well has been drilled, cased, and cemented over a 30 day period, the
first two days are the most critical for zonal isolation of USDWs where the foundation for well
integrity is determined.
Well Construction - Completion
Well completion is the where the production casing is perforated, the formation is hydraulically
fractured, frac fluids are unloaded from the formation, and production operations commence.
This is basically the well's configuration for the rest of its life as it relates to protecting USDWs.
Prior to performing the hydraulic frac, the production casing is tested to anticipated frac
pressure plus a safety factor, as is the frac tree and all the surface pumping equipment and
lines. During the frac, all casing annuli are monitored, as is the injection rate, injection pressure,
and slurry properties. If during the frac job, significant pressure is found on the intermediate
casing annulus, or there is any indication of communication with the surface casing annulus, the
frac job is shut down and not resumed until corrective actions are made that only the intended
zone is subject to frac pressures.
Refracs are similar to original fracs as discussed above with the exception that a frac string or
wellhead saver might be used to protect older production casing strings and wellheads from
frac pressures. This is a case by case situation that requires additional testing and engineering
analysis in order to protect both the well and USDWs during refrac operations.
As the well is produced, reservoir pressures tend to drop and liquid rates tend to rise, therefore
devices for lifting liquids such as a tubing string with pumping or gas lift equipment becomes
126
-------
necessary. This internal configuration can have an impact on USDW protection and is addressed
during the operations phase.
Well Operations
Prudent operators monitor all casing annuli on a regular basis to be able to detect sustained
casing pressure or SCP. This condition could be caused by thermal expansion of annular fluids,
packer or liner leaks, leaks into the annulus from inner tubing or casing strings, or from annular
migration due to poor zonal isolation.
All states have rules for reporting and responding to the loss of well integrity which includes
releases, non-thermal SCP, and other abnormal situations (39) as does the BLM (40) and best
industry practices (41). The Commonwealth of Pennsylvania has new rules that require
quarterly mechanical integrity testing and annual reporting for all operating wells (42).
Adjacent well operations may have an impact on mechanical integrity of our case study well.
Hydraulic fracturing of a well near our case study well into a zone that is not protected, or not
adequately protected for the conditions imposed can lead to unwanted well to well
communication. This is currently a void where regulations and industry practices have not fully
recognized that well integrity can become a neighborhood issue.
Well Plug and Abandonment "P&A"
Similar to well construction regulations and industry practices, well P&A also has
comprehensive guidance to prevent vertical migration of fluids into USDWs. There is clear
guidance for plug location, cement quantity, quality, placement techniques, testing, and
reporting (43) (44) (45). Regulations may also specify that only approved cementing contractors
perform plugging, require independent onsite supervision, and require post cement job
certifications by both the operator and the cementing company.
There are also significant industry studies and best practices for well P&A (46) (47).
Failure studies have found that vertical migration issues in P&Aed wells are directly related to
the original primary cement job during well construction. Those wells with gas migration to the
surface prior to well P&A were likely to continue to have gas migration to the surface after P&A.
Additionally, those wells plugged with bridge plugs and dump bailed cement on top were found
to be more prone to leakage than wells plugged with cement that was circulated or squeezed in
place (35) (36).
Conclusions
Well integrity and well construction are inextricably linked, regardless of the completion
technique selected. Primary cementing is the critical step for preventing vertical migration of
fluids during the well's productive life, and afterwards.
State and federal regulations address casing and cementing with prescriptive rules and
reporting requirements, while industry employs a large body of technical studies and best
127
-------
practices. Five identified casing failure categories: materials, connections, wear / handling,
mechanical, and corrosion are not as problematic for zonal isolation as three identified
cementing failure categories: insufficient cement volume, low bond strength, and cement
sheath damage.
For hydraulically fractured completions, significant bodies of industry technical information and
best practices have been published. State and federal regulations address hydraulic fracturing
with rules and reporting requirements which are continuously adapting to keep pace with
technology advancements (48).
Adjacent wells and the potential for unwanted communication during hydraulic fracturing is a
concern. State and federal regulations are largely silent on this issue, as are industry studies
and best practices.
References
1. Interstate Oil and Gas Compact Commission "Hydraulic Fracturing, Texas Regulations
Protect Surface and Ground Water"
2. 40 CFR 144.3 Safe Drinking Water Act, definition for USDW
3. Texas Administrative Code Title 16, Part 1, Rule 3.7 and Title 16, Part 1, Rule 3.8 (b)
4. 43 CFR 3162.1 (a) Onshore Oil and Gas Order #2, Drilling Operations
5. U.S. Energy Information Administration "The Geology of Natural Resources" February 14,
2011
6. Pennsylvania Department of Environmental Protection "Methane Gas and Your Water Well"
November 2009
7. The National Academies "Management and Effects of Coalbed Methane Produced Water in
the United States" 2010
8. New York Times "Nature's Cheapest Fuel, The Natural Gas of the Delaware Valley Likely to
be Utilized" March 15, 1886
9. Western New York Trails.com "Eternal Flame Falls - Chestnut Ridge Park, Orchard Park,
New York" December 2010
10. American Geophysical Union Blogosphere "Natural Gas Seeps in Western NY" November
2010
11. Penn State University College of Agricultural Sciences "Gas Well Drilling and Your Private
Water Supply" March 2010
12. Wikipedia, Oil Springs, Kentucky
13. Wikipedia, Oil Springs, Ontario Canada
14. Wikipedia, Seneca Oil Spring, NY
15. Texas Landmarks and Vanished Communities, Salt Creek and Brine Springs, TX
16. Oklahoma Administrative Code 165:10-3-1
17. Texas Administrative Code Title 16, Part 1, Rule 3.13 (a)(2)(C)
18. Pennsylvania Code Title 25 Chapter 78.83
19. Oklahoma Administrative Code 165:10-3-4 (c)(l)
20. Pennsylvania Code Title 25 Chapter 78.84 (a)
21. API Spec 5CT "Specification for Casing and Tubing" Eighth Edition 2006
128
-------
22. NACE MR0175 / ISO 15156 "Materials for use in H2S Containing Environments in Oil and Gas
Production" Second Edition 2009
23. API RP 5C1 "Recommended Practice for Care and Use of Casing and Tubing" Eighteenth
Edition 2010
24. API RP 5C5 "Recommended Practice on Procedures for Testing Casing and Tubing
Connections" Third Edition 2003
25. API RP 5A5 "Field Inspection of New Casing, Tubing, and Plain-end Drill Pipe" Seventh
Edition December 2009
26. API RP 5B1 "Gaging and Inspection of Casing, Tubing, and Line Pipe Threads" Fifth Edition
2010
27. API RP 5A3 "Recommended Practice on Thread Compounds for Casing, Tubing and Line
Pipe" Third Edition 2009
28. Department of Energy, Sandia National Labs "Euler Buckling of Geothermal Well Casing"
Rechard and Schuler, April 1983
29. Texas Administrative Code Title 16, Part 1, Rule 3.13 (b)(2) (A) thru (G)
30. Texas Administrative Code Title 16, Part 1, Rule 3.13 (a)(l)
31. API RP 65 "Isolating Potential Flow Zones During Well Construction" Part 2, First Edition
May 2010
32. API Spec 10A "Specification for Cements and Materials for Well Cementing" December 2010
33. API RP 10B (1 thru 6) "Recommended Practice for Testing Well Cements"
34. API Spec 10D "Specification for Bow-Spring Centralizers" Sixth Edition 2002
35. SPE 106817 "Evaluation of the Potential for Gas and C02 Leakage Along Wellbores" Watson
and Bachu, March 2007
36. Alberta Energy and Utilities Workshop "Factors Affecting or Indicating Potential Wellbore
Leakage" Watson and Bachu, March 2007
37. API Technical Report 10TR1 "Cement Sheath Evaluation" Second Edition 2008
38. SPE 105193 "Improving Formation Strength Tests and Their Interpretation" van Oort and
Vargo, February 2007
39. Texas Administrative Code Title 16, Part 1, Rule 3.17 (a) & (b)
40. 43 CFR 3162.5-2 Onshore Oil and Gas Order #2, Drilling Operations
41. API RP 90 "Annular Pressure Management for Offshore Wells" First Edition, August 2006
42. Pennsylvania Code Title 25 Chapter 78.88(e)
43. Texas Administrative Code Title 16, Part 1, Rule 3.14
44. Oklahoma Administrative Code 165:10-11
45. Pennsylvania Code Title 25 Chapter 78.91 thru 78.98
46. API Bulletin E3 "Environmental Guidance Document: Well Abandonment and Inactive Well
Practices for U.S. Exploration and Production Operations" First Edition, June 2000
47. SPE 28349 "Issues and Techniques of Plugging and Abandonment of Oil and Gas Wells"
Calvert and Smith, September 1994
48. SPE 140482 "State and Federal Regulation of Hydraulic Fracturing: A Comparative Analysis"
Arthur, Hochheiser and Coughlin; January 2011
129
-------
Appendix A- PowerPoint Illustration for this Well Integrity Case Study-Slides 1 through 4
130
-------
Appendix A- PowerPoint Illustration for this Well Integrity Case Study- Slides 5 through 8
-1?
131
-------
Appendix A- PowerPoint Illustration for this Well Integrity Case Study-Slides 9 through 12
132
-------
Appendix A- PowerPoint Illustration for this Well Integrity Case Study-Slides 13 through 16
ia
-------
Appendix A- PowerPoint Illustration for this Well Integrity Case Study-Slides 17 through 20
t
-
II
^_3_" » > f x ii it B ^ » ii » ''1
II
134
-------
Appendix A- PowerPoint Illustration for this Well Integrity Case Study-Slides 21 through 22
-
C jiot/s r^i am on c«f uman art m
135
-------
Risks to Drinking Water from Oil and Gas Wellbore
Construction and Integrity: Case Studies and Lessons Learned
Briana Mordick
Natural Resources Defense Council
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA.
Introduction
Numerous cases of known or suspected drinking water contamination across the country have
been linked to oil and gas production. This paper will examine various published reports from
two such cases and discuss the potential roles of wellbore construction and integrity and
hydraulic fracturing in the resultant drinking water contamination.
Case Study #1: Bainbridge Township, Geauga County, Ohio
Incident Summary
On December 15th, 2007, an explosion was reported in the home at 17975 English Drive,
Bainbridge Township, Geauga County, Ohio. Early investigations determined that methane was
entering homes in the vicinity of the explosion through domestic water wells. The Ohio
Department of Natural Resources, Division of Mineral Resources Management (DMRM)
inspected local gas wells to identify the source of the gas. When inspectors arrived at the
English No. 1 gas well owned by Ohio Valley Energy Systems Corp (OVESC), representatives
from OVESC were on location examining the well and discussing remedial cementing
operations. OVESC proactively assumed responsibility for the incident without waiting for a
completion of the investigation by DMRM and initiated corrective action. In the weeks
following the explosion, DMRM initiated a monitoring program for methane in wells and homes
and to monitor the response of wells to corrective action at the English No. 1 well. DMRM
performed remedial work on affected water wells and provided in-home methane monitoring
systems and replacement sources of drinking water for affected homes (Ohio DNR DMRM,
2008).
Simplified Stratigraphy at the Location of the English No. 1 Well
The OVESC English No. 1 well was drilled to a total depth of 3,926'. The formations encountered
during drilling, listed in order of increasing depth, are as follows (Ohio DNR DMRM, 2008):
Unconsolidated glacial till. Less than 88' thick
Pennsylvanian and Mississippian aged interbedded sandstone and shale comprising the
drinking water aquifer: Sharon Conglomerate, Cuyahoga Formation, Berea Sandstone.
The Berea Sandstone has sometimes been noted to contain low-pressure natural gas.
Approximately 200' thick
Devonian aged Ohio Shale. Contains noncommercial quantities of low-pressure natural
gas. Approximately 1800 feet thick
136
-------
Devonian and Silurian aged "Big Lime"/Lockport Dolomite limestone and evaporate
deposits. Contains the Oriskany Sandstone and "Newburg" Dolomite members, which
are porous, permeable, brine-bearing zones which sometimes locally contain
noncommercial quantities of natural gas. Approximately 1600' thick
Thin interbedded shale and limestone partly comprising the seal for the gas-bearing
target reservoir. Contains the Packer Shell, a typically impermeable limestone but which
can be locally faulted or fractured near structural features. Approximately 100' thick
Low porosity and permeability Clinton Sandstone. Target formation containing
commercial quantities of natural gas. Approximately 200' thick
Sequence of Events Leading to Natural Gas Invasion into Drinking Water Aquifers
OVESC spud the English No. 1 well on October 18th, 2007. Conductor casing was installed to a
depth of 88 feet, through glacial till and into bedrock. The well was drilled through the
groundwater aquifers and surface casing was set at 263 feet and cemented to surface. Drilling
continued until the total depth of the well, 3,926 feet, was reached on October 26th. An open-
hole logging run was attempted but the logging tool bridged out at 3,658 feet, the depth of the
Packer Shell, due to an apparent filter cake build up. The logging tool could not be moved
below the bridge and open-hole logs were not obtained. OVESC proceeded to set 4-1/2"
production casing. Casing was run into the hole and became stuck at 3,659 feet, the depth of
the Packer Shell. The casing was washed down to 3,873 feet, became differentially hung, and
could not be lowered further. OVESC then proceeded to cement the production casing. Prior to
cementing, circulation of the wellbore was established but was subsequently lost during the
cementing operation and could not be re-established. The cementing operation was concluded
and, due to the lost circulation event, a cement bond log was run to establish the top of cement
(TOC). (Ohio DNR DMRM, 2008; Bair et al, 2010)
Based on the cement job design, TOC should have been 700-800 feet above the top of the
Clinton formation. The cement bond log revealed TOC to be at 3,640 feet, the depth of the
Packer Shell. This finding and the previous drilling, logging, and casing problems suggest the
Packer Shell thieved a large quantity of cement due to the presence of localized fracturing.
Despite the inadequate primary cement job, OVESC continued to complete the well. The well
was perforated from 3720-3740 feet, leaving only approximately 80 feet of cement covering
the Clinton between the top perf and the TOC/open annulus, and the planned hydraulic
fracture treatment proceeded on November 13th. The original frac design called for 105,000
gallons of water and 600 sacks of proppant. After pumping less than half the planned fluid and
proppant, fluid circulated out of the open valve on surface-production casing annulus. Pump
pressure and rate were reduced, 4000 gallons of fresh water was pumped to flush and recover
sand, and the frac job was discontinued. (Ohio DNR DMRM, 2008; Bair et al, 2010)
In the three days following the well completion, most of the frac fluid was recovered and
pressure on the surface-production casing was recorded. The pressure increased each day and
stabilized at 320 psi on the third day and gas was periodically blown off to reduce pressure.
Construction was completed and the well was shut in for the next 31 days. (Ohio DNR DMRM,
2008; Bair etal, 2010)
137
-------
While the well was shut in, gas from the Clinton, Newburg, and Ohio Shale formations migrated
into the uncemented annular space behind the production casing and caused the annulus to
become overpressured, reaching a maximum recorded pressure of 360 psi. This gas then
migrated from the high-pressure annulus, through fractures, into the shallow low-pressure
aquifer and subsequently into domestic water wells, culminating in the explosion on English
Drive. (Ohio DNR DMRM, 2008; Bair et al, 2010)
Remedial Action
OVESC performed two remedial cement jobs, one to seal the annulus from the current TOC to
above the Newburg formation and one to seal the remaining open annulus to surface. Small
amounts of gas were still detected in the annulus and a segmented bond log was run to
determine the source. The bond log showed channeling of the cement from 550 feet to surface,
which was allowing shallow Ohio Shale gas to enter the annulus. A good to excellent bond was
measured below that depth. (Ohio DNR DMRM, 2008; Bair et al, 2010)
Primary Causes of Gas Invasion into Drinking Water Aquifers
1. Poor Primary Cement Job: The poor primary cement job left the shallow Newburg
Dolomite and Ohio Shale gas-bearing zones open to the annulus behind the production
casing, allowing high-pressure gas to migrate into the annulus.
2. Decision to Hydraulically Fracture the Well Despite the Poor Cement Job: Circulation of
fluid and oil in the surface-production casing annulus during hydraulic fracturing
indicates that the fractures grew "out-of-zone" and allowed the frac to communicate
directly with the wellbore. The frac likely compromised the 80 feet of cement between
the top perf and the open annulus, causing a loss of cement bond between the
formation and production casing. This likely allowed Clinton gas to also migrate into the
annulus behind the production casing.
3. Shutting in the Well for 31 Days: The decision to shut in the surface-production casing
annulus for 31 days allowed the annulus to become over-pressured and gas to migrate
from the high-pressure annulus, through fractures, to the groundwater aquifer and
eventually into domestic water wells. (Ohio DNR DMRM, 2008; Bair et al, 2010)
Areas of Dispute
Subsequent to the well contamination incident, 42 property owners brought a suit against
OVESC and six other parties involved in the operations at the English No. 1 well. (Bair et al,
2010). As part of the suit, the attorneys for the plaintiffs contracted Eckstein & Associates
(E&A), a geological engineering firm, to review the causes of the incident. This subsequent
report differed from the DMRM assessment in several areas. Consequently, DMRM convened a
panel of experts to review the findings of Eckstein & Associates. The four main areas of dispute
are as follows:
1. Was the over-pressurization of the annulus of sufficient magnitude to induce fractures
in the geologic formations exposed in the uncemented annulus?
138
-------
a. The E&A report concluded that the pressures were indeed sufficient to create
fractures in the Ohio Shale and portions of the "Big Lime", providing migration
pathways for deep gas. (Eckstein, 2009)
b. The DMRM Expert Panel concluded that the pressures may have been sufficient
to create fractures in the Ohio Shale but that any fractures created would be
shallow, oriented horizontally, and of limited extent, and at most would
temporarily augment transport along natural fracture networks. (Bair et a I, 2010)
2. If the over-pressurization of the annulus did induce fractures, could they become
permanent migration pathways for deep gas to reach groundwater?
a. The E&A report concluded that the "deep- and far-reaching fractures" created by
the over-pressurization of the annulus will serve as long-term migration
pathways for methane to groundwater. Supporting evidence offered includes
data for wells in the affected area showing that methane concentrations have
remained high or increased over time. (Eckstein, 2009)
b. The DMRM Expert Panel report concluded that any induced fractures would be
shallow and of limited vertical, aerial, and temporal extent and consequently
would not create long-term migration pathways for gas to groundwater.
Supporting evidence offered includes data showing that the gas plume is
dissipating upward and gas pressures in affected wells are decreasing. (Bair et al,
2010)
3. Can methane concentrations in domestic water wells be used to delineate such fracture
networks?
a. The E&A report concluded that the presence of methane in water wells was
sufficient evidence for the presence of induced fractures, and therefore could be
used to map or delineate such fracture networks. (Eckstein, 2009)
b. The DMRM Expert Panel report concluded that the presence of methane alone,
in the absence of other corroborating evidence, was not sufficient to delineate
such fracture networks. They determined that other factors are in part
responsible for the patterns of methane concentrations measured in domestic
water wells over time. (Bair et al, 2010)
4. What is the nature and origin of the presence of black particulate matter in some
domestic water wells?
a. Following the English No. 1 well incident, some residential water wells began
yielding black particulate matter. Chemical analysis showed that the particles
consist of heavy metals, including lead and copper. The E&A report concluded
that the particulate matter was entrained in the gas leaking from the well, with
the likely source being the Ohio Shale. (Eckstein, 2009)
139
-------
b. The DMRM report concluded that the particulate matter was not widespread
and that it could not be determined whether it was created by the released
methane or by natural processes. (Bair et a I, 2010)
Case Study #2: Mamm Creek Field, Garfield County, Colorado
Incident Summary and Studies Considered for Review
In 2004, citizens notified the Colorado Oil and Gas Conservation Commission (COGCC) of the
presence of gas bubbling in the West Divide Creek, Garfield County, CO, near the Mamm Creek
Gas Field. Subsequent investigations identified the gas as thermogenic gas from the Williams
Fork (Mesaverde) Formation, which is the primary gas-bearing target in the Mamm Creek Field.
Water testing also detected the presence of BTEX compounds above regulated limits. It was
determined that the gas and other contaminants were leaking from a nearby wellbore which
had been improperly cemented, Encana's Schwartz #2-15B. Fines from this incident were used
to fund a study to determine the vulnerability of groundwater and surface water to impacts
from natural gas exploration and other human activities in Garfield County, CO near the Mamm
Creek Natural Gas Field.
The Phase I study, performed by URS Corporation, compiled and evaluated existing data on
water wells, gas wells, and water quality, and also included a limited amount of new field work
(URS, 2006). The Phase II Study, performed by S.S. Papadopulos and Associates, focused on two
field sampling tasks:
1. Water quality, gas composition, and methane stable isotope samples were obtained for
wells which previously had compounds of concern above regulated limits or had
sodium-chloride (Na-CI) concentrations which suggested mixing with deeper
brine/saline water.
2. Produced water and gas samples were taken from gas wells near the domestic water
wells which had water and/or gas chemistry which may have been influenced by deeper
formations, either by natural processes or through gas drilling activities (Papadopulos,
2008)
Subsequently, Dr. Geoffrey Thyne provided summaries and reviews of the Phase I and Phase II
studies (Thyne, 2008). Dr. Thyne's conclusions were in turn reviewed by S.S. Papdopulous and
Associates (Papadopulos, 2009), Bill Barrett Corporation (Donate et a I, 2009), and Dr. Anthony
Gorody of Universal Geoscience Consulting, Inc (Gorody, 2009).
Beginning in 2009 and completed in 2011, the United States Geological Survey (USGS), in
cooperation with the Colorado Department of Public Health, undertook a study to determine
the sources and sinks of nitrate and methane in domestic water wells screened in the shallow
Wasatch formation in Garfield County (McMahon et a I, 2011).
The following findings were generally consistent throughout all the studies considered:
140
-------
1. Some domestic water wells had increased concentrations of methane, relative to
background
a. Both biogenic and thermogenic methane were detected
2. Some domestic water wells had concentrations of fluoride, selenium, nitrate, and/or
arsenic which exceeded health-based standards
a. Fluoride and selenium concentrations do not appear to be related to oil and gas
activity
b. Nitrate concentrations are most likely related to agricultural activity, septic
system effluent, and/or animal waste
3. Some domestic water wells had concentrations of chloride, iron, manganese, and/or
total dissolved solids (IDS) which exceeded aesthetic-based standards
a. High chloride and IDS concentrations indicate the mixing or interaction of
shallow groundwater with deeper formation water. (URS, 2006; Papadopulos,
2008; Thyne, 2008; McMahon et al, 2011)
Several areas of dispute arose between the various studies, including:
1. Evidence for a temporal correlation of methane and chloride contamination and natural
gas activity
2. The nature and origin of methane in domestic water wells
3. The primary mechanism for deep Wasatch or Mesaverde formation water to mix with
shallow groundwater (URS, 2006; Papadopulos, 2008; Thyne, 2008; Donate et al, 2009;
Gorody, 2009; Papadopulos, 2009; McMahon et al, 2011)
Areas of Dispute
Evidence for a temporal correlation of methane and chloride contamination and natural gas
activity
In his review of the Phase I and II studies, Dr. Thyne observed that methane concentrations and
the number of wells with elevated chloride concentrations increased with time and were
correlated to the increasing number of gas wells with time. (Thyne, 2008) Papadopulos and
Associates, Bill Barrett Corporation, and Dr. Gorody disputed this claim and stated that there is
no statistically significant increase in methane or chloride concentrations with time (Donate et
al, 2009; Gorody, 2009; Papadopulos, 2009).
The nature and origin of methane in domestic water wells
The Phase I study found the presence of methane of biogenic, thermogenic, and unknown
origin in the water samples. Most samples that had elevated concentrations of methane
contained biogenic methane. The study indicates that biogenic methane can be formed by
various processes but does not offer a hypothesis for how the methane came to be present in
groundwater and domestic water wells. The implication, however, is that presence of biogenic
methane in domestic water wells is not related to oil and gas development. A smaller number
of samples contained thermogenic methane. In the area near the West Divide Creek seep, the
141
-------
origin of the methane is concluded to be from the leaking gas well which caused the seep.
Some of the highest methane concentrations were detected in the southeastern portion of the
study area. Although there had been little gas development activity in the area, there were
several old wellbores that records indicate may not have been properly plugged and
abandoned. The study concluded that the presence of thermogenic methane in water samples
could result from either migration along natural pathways, such as faults, or from natural gas
drilling, completion, or production activities or improperly abandoned wells. The researchers
concluded that more data would be necessary to conclusively determine which migration
pathway was responsible in each instance. The origin of the unknown methane types could not
be determined and may have resulted from mixing of different sources. (URS, 2006)
The Phase II study also found the presence of methane of both biogenic and thermogenic origin
in domestic water wells. Although most samples had isotopic compositions which indicated a
thermogenic origin, researchers determined that most samples were in fact biogenic in origin.
The conclusion was that the majority of samples which appeared to have a thermogenic
isotopic signature had undergone a "biogenic methane oxidation shift". This is a process by
which gas that is biogenic in origin undergoes oxidation, leaving the remaining fraction of
methane with an isotopic signature that appears to be thermogenic but is in fact biogenic. As
with Phase I, the researchers did not offer a hypothesis for how the biogenic methane came to
be present in domestic water wells. Again, the implication is that the presence of biogenic
methane in domestic water wells is not related to oil and gas development. A smaller number
of samples contained methane that the researchers believed to be truly thermogenic in origin.
Two hypotheses were offered to explain the nature and origin of these samples:
1. The samples may be derived from deeper gas-bearing formations, either tight sands gas
or coalbed methane gas
2. The samples may represent some mixture between biogenic and thermogenic gas
For those samples which the study determined to be truly thermogenic in origin, and not the
product of oxidation of biogenic methane, the researchers suggest that two mechanisms may
be responsible: migration along natural faults and fractures or gas exploration and production.
The study concluded that distinguishing between the two is not possible with the current data.
(Papadopulos, 2008)
In his review of the Phase I and Phase II studies, Dr. Thyne also agreed that the samples
contained methane which appeared to be of both biogenic and thermogenic origin. However,
unlike the previous researchers, Dr. Thyne concluded that the majority of samples were
thermogenic in origin. Dr. Thyne rejected the conclusion of the Phase II study that many of the
samples with thermogenic isotopic signatures were in fact biogenic methane which had been
oxidized. For those samples with isotopic values indicating biogenic origin, Dr. Thyne noted that
their origin was microbial C02 reduction, in which C02 is converted to methane by microbial
processes. Dr. Thyne concluded that the origin of this C02 was thermogenic C02 from the
Williams Fork (Mesaverde) Formation. Consequently, the methane produced by this C02 would
also be considered thermogenic in origin. Due to this finding that the majority of samples were
thermogenic in origin, Dr. Thyne concluded that gas development activities had impacted
142
-------
groundwater. (Thyne, 2008) Papadopulos and Associates disputed these conclusions and found
no basis to change the conclusions from their original report (Papadopulos, 2009).
The USGS study also sampled methane which appeared to be of both biogenic and thermogenic
origin. The USGS study used a more diverse geochemical data set than previous studies to
determine the nature and sources of the methane. Samples with the highest concentrations of
methane appeared to be biogenic in origin. These samples also contained high concentrations
of helium-4 and the co-occurrence implies that the methane was derived from a deep source
rather than being generated in-situ in domestic water wells. Researchers concluded that one
source for this deep biogenic methane could be the deep Wasatch Formation. Some samples
also contained methane which appeared to be thermogenic in origin. Researchers determined
that some of these samples may have contained biogenic methane which had undergone
oxidation while other samples contained methane which was truly thermogenic in origin. The
source of this thermogenic gas was most likely the Mesaverde (Williams Fork) Formation. The
study concluded that two migration pathways were possible for both the deep biogenic and
thermogenic gas: natural faults or fractures or the uncemented annular space in gas wells.
(McMahon, etal, 2011)
The primary mechanism for deep Wasatch or Mesaverde formation water to mix with shallow
groundwater
All four studies concluded that the geochemistry of some water samples may indicate mixing
between shallow groundwater and deeper water. All four studies also suggested that either
natural faults or fractures or gas wellbores could provide pathways for deep water to reach
shallow water, however there was some disagreement between the studies on which of these
pathways was most likely. (URS, 2006; Papadopulos, 2008; Thyne, 2008; McMahon et a I, 2011)
The URS study concluded that the cause of mixing could not be determined and could have
been the result of either natural pathways or gas development activities (URS, 2006). The
Papadopulos and Associates study also concluded that natural pathways, wellbores, or
hydraulic fractures may be possible migration pathways for deeper fluids but stated that the
samples with geochemical signatures indicating mixing were from wells in areas with only
modest gas development activity and therefore it was not possible to distinguish between
natural and manmade impacts (Papadopulos, 2008). In his review of the Phase I and Phase II
study, Dr. Thyne concluded that the number of domestic water wells with elevated chloride
concentrations was increasing over time and correlated to the number of gas wells drilled, and
the that source of the chloride was produced water (Thyne, 2008). The USGS study concluded
that both natural fractures and wellbores were likely migration pathways for deeper formation
water to reach shallow groundwater. They also determined that Mesaverde formation water
was an important source of chloride in some wells even when the actual fraction of Mesaverde
water in the sample was small (McMahon et a I, 2011).
Key Observations
Despite the areas of dispute discussed above, some key observations and conclusions emerged
from the studies. (URS, 2006; Papadopulos, 2008; Thyne, 2008; McMahon et al, 2011)
143
-------
Some domestic water samples contain methane and deep formation water which may
have migrated to water wells through either natural pathways or gas wellbores or both.
The study area is naturally faulted and fractured. Fault and fracture density increases
near structural features, such as the Divide Creek Anticline.
Regulations were updated in 2004 to require that all new wells have surface casing set
below the lowest USDW and cemented to surface and production casing cemented to
500' above the top of gas in the Mesaverde (Williams Fork) Formation. There is no
requirement to cement over the deep Wasatch Formation. Older wells may have been
constructed using different standards and may not have been properly abandoned.
Gas production wells with persistent or recurring elevated bradenhead pressures have
been identified near structural features.
Domestic water wells with elevated methane and chloride concentrations are often
coincident with structural features.
Natural fractures and faults may provide migration pathways for gas and fluids, both to
groundwater and to the uncemented annular space of wellbores. Fractures and faults
may also cause complications in well drilling, construction, and completion and result in
well integrity problems.
Challenges
Both these case studies and others around the country face challenges in determining causality
of water contamination. One of the most significant challenges is the fact that in many oil and
gas development fields, a systematic and comprehensive assessment of baseline water quality
predating oil and gas development does not exist. When water contamination related to oil and
gas development is suspected, investigators must piece together baseline water quality from
previous studies and reports or try to sample water which may be "outside" the influence of oil
and gas development.
Determining the extent and source of water contamination is also challenging. As noted by Dr.
Thyne, domestic water wells may not be ideally located to robustly determine the source of
contamination. (Thyne, 2008) As pollutants disperse from their source, they may undergo
chemical or physical changes, making it difficult to conclusively determine the source of
pollution. Pollutants and contaminants may also interact with any media between the source
and water well and result in the mobilization of naturally occurring contaminants. When such
naturally occurring contaminants are detected in groundwater, it may be difficult to distinguish
whether they migrated as a result of natural or anthropogenic causes or the potential link
between naturally occurring contaminants and human activities may not be investigated.
Selecting the proper set of test parameters to determine the source of water contamination is
also a challenge. As seen in the Garfield County example, many of the chemicals tested for in
the water samples could not be used to conclusively identify the source or method of transport
of contaminants because they were indicative of multiple sources and/or migration pathways.
144
-------
While it is unlikely that any water contamination investigation will test for all chemicals used or
released by oil and gas drilling, special care must be given to selecting proper indicator
chemicals. In these and other examples, investigators often assume that the presence of
biogenic gas in drinking water is not related to oil and gas activities and that thermogenic gas is
related to oil and gas activities. As shown in the USGS study, this is a poor assumption.
Investigators must take the next steps and determine both the source of methane in
groundwater and the mechanisms by which it could migrate from its source into groundwater.
One of the most significant concerns regarding the risk of hydraulic fracturing to contaminate
drinking water is that many of the chemicals used in hydraulic fracturing fluid are not known on
a well by well basis. In the Bainbridge, OH case, investigators tested for three chemicals which
were present in the hydraulic fracturing fluid used to frac the English No. 1 well (Ohio DNR
DMRM, 2008). However, the report did not state how many chemicals in total were used in the
hydraulic fracturing fluid or whether those they selected to test for represented the range of
mobility and/or toxicity of all the chemicals used. In the Garfield County example, none of the
studies tested the water for chemicals used in hydraulic fracturing. In their recommendations
for additional work, Papadopulos and Associates stated, "The effect on groundwater due to the
introduction of drilling or well completion/hydrofracturing fluids into the shallow aquifer was
not investigated for this study. A study evaluating possibly local effects of drilling or
hydrofracturing fluids on domestic groundwater should be considered." (Papadopulos, 2008)
Given that all studies found that deeper groundwater mixed with shallow water and that
natural fractures or wellbores could provide the pathways for this contamination, testing for
the presence of hydraulic fracturing chemicals and determining how induced fractures could
interact with natural fractures is an extremely important piece of additional research which
should be conducted.
Solutions and Lessons Learned
Detailed site characterization and planning and baseline testing prior to any oil and gas
development are crucial. An integral part of understanding how wellbore construction and
integrity and hydraulically induced fractures could create migration pathways to and potentially
contaminate groundwater is a thorough understanding of the current geologic and hydrologic
regimes. Site characterization and planning work may include but are not limited to:
Detailed study of regional and local geologic structure including faults, fractures, stress
regimes, rock mechanical properties, etc. through the use of 3D seismic surveys,
outcrop analog studies, collection of core and relevant analysis, well logs including
FMI/image logs, etc. As seen in Garfield County, the presence of natural faults and
fractures and areas of increased fracturing around structural features may be pathways
for gas, drilling fluids, hydraulic fracturing fluids or formation fluids to reach
groundwater or the uncemented annuli of hydrocarbon wells and may also compromise
wellbore integrity.
Detailed pre-drill maps of the extent and chemical composition of groundwater aquifers
Hydrologic flow and transport data collection and modeling
145
-------
Thorough identification of existing wellbores, determination of the integrity of those
wellbores (i.e. casing, cement, etc.), and mitigation where necessary
Hydrocarbon sampling and analysis to determine variations in chemical and isotopic
compositions of any hydrocarbons which may be encountered both vertically in a
wellbore and aerially throughout an oil or gas field
As development of an oil or gas field proceeds, these data sets must be continually updated as
new information becomes available, both temporally and aerially.
Wellbore construction and integrity are paramount in protecting drinking water. Wellbores
must be constructed so that any hydrocarbon or non-potable water bearing formations are
isolated. As seen in Garfield County and in other examples throughout the country, shallow gas-
bearing zones can be significant sources of methane in drinking water. Shallow brine or
formation water or its chemical constituents may also migrate into drinking water if not
isolated. Hydraulic fracturing must not occur if wellbore integrity is in question.
Wellbore maintenance is also crucial. Older wellbores which have degraded, been constructed
using less protective standards, or which have been improperly abandoned must be identified
and remediated. Such wellbores could provide migration pathways for contaminants to reach
groundwater and hydraulically induced fractures could provide new or enhanced migration
pathways for gas or fluids to reach these wellbores.
A water quality monitoring program should be developed and implemented throughout the life
of oil and gas exploration and production. The use of dedicated water quality monitoring wells
should be considered in order to help detect the presence of contaminants prior to their
reaching domestic water wells. Placement of such wells should be based on detailed hydrologic
flow models and the distribution and number of hydrocarbon wells.
Robust models and direct measurements of hydraulic fracture growth, including preferred
fracture orientation, frac half-length, and frac height growth, are also crucial. Techniques such
as microseismic monitoring, tiltmeters, and chemical and radioactive tracers should be
employed over the life of the field, especially as development progresses into new areas.
Equally critical is robust post-frac monitoring. This includes tracking injected volumes of frac
fluids as well as flowback volumes to better understand the potential for migration. In order to
effectively monitor where frac fluids go and whether they or the chemicals they contain
interact with groundwater, it is essential to know the exact chemical composition of all
constituents involved in the drilling and completion process, including but not limited to:
Drilling fluids/mud
Frac fluid
Connate water/produced water
Geochemistry of producing formations and formations which serve as potential barriers
between the producing formation and any aquifer
146
-------
Cumulative impacts must also be considered. The risks to groundwater may increase as
development progresses, as older wellbores are abandoned, and as drilling expands to new
areas. The impacts of increasingly more wellbores and increased fracture density due to
hydraulic fracturing and the potential impacts to drinking water must be examined.
References
Bair, E.S., Freeman, D.C., and Senko, J.M., 2010, Expert Panel Technical Report, Subsurface Gas
Invasion Bainbridge Township, Geauga County, Ohio: Prepared for the Ohio Department
of Natural Resources Division of Mineral Resources Management
Donate, S.A., Sterrett, R.J., and Hanna, B., 2009, Review of the Thyne Report's Analysis and
Conclusions of Mamm Creek Phase II Hydrogeologic Study: PowerPoint presentation
prepared for the Colorado Oil & Gas Conservation Commission, 10 p.,
http://cogcc.state.co.us/Library/Presentations/Glenwood Spgs HearingJuly 2009/(1 F)
ReviewOfTheThyneReport'sAnalysis Conclusionof%20MammCreekPh2.pdf
Eckstein, Y., 2009. Wellhead LEL records in Bainbridge residential water wells, PowerPoint
presentation to Bainbridge Gas Invasion Expert Panel, 43 p., www.ohiodnr.com/mineral
Gorody, Anthony, 2009, Summary Comments on the Report by Dr. G. Thyne: PowerPoint
presentation prepared for the Colorado Oil & Gas Conservation Commission, 26 p.,
http://cogcc.state.co.us/Library/Presentations/Glenwood Spgs HearingJuly 2009/(1 G
) Summary Comments Report.ppt
McMahon, P.B., Thomas, J.C., and Hunt, A.G., 2011, Use of diverse geochemical data sets to
determine sources and sinks of nitrate and methane in groundwater, Garfield County,
Colorado, 2009: U.S. Geological Survey Scientific Investigations Report 2010-5215, 40 p.
Ohio Department of Natural Resources, Division of Mineral Resources Management, "Report on
the Investigation of the Natural Gas Invasion of Aquifers in Bainbridge Township of
Geauga County, Ohio" September 1, 2008
Papadopulos & Associates, Inc., 2008, Phase II hydrogeologic characterization of the Mamm
Creek Field area, Garfield County, Colorado: Prepared for the Board of County
Commissioners, Garfield County, Colorado, 41 p.
Papadopulous & Associates, Inc., 2009, Evaluation of Thyne, 2008 "Review of Phase II
Hydrogeologic Study" Mamm Creek Field Area, Garfield County, Colorado: PowerPoint
presentation prepared for the Colorado Oil & Gas Conservation Commission, 29 p.,
http://cogcc.state.co.us/Library/Presentations/Glenwood Spgs HearingJuly 2009/(1 D
) Evaluation of Thyne Report vers2%20(dv).pdf
Thyne, Geoffrey, 2008, Review of Phase II Hydrogeologic Study prepared for Garfield County:
Prepared for Garfield County, 26 p.,
http://cogcc.state.co.us/Library/Presentations/Glenwood Spgs HearingJuly 2009/(1 A
) ReviewofPhase-ll-HvdrogeologicStudy.pdf
URS Corporation, 2006, Phase I hydrogeologic characterization of the Mamm Creek Field area in
Garfield County: Prepared for the Board of County Commissioners, Garfield County,
Colorado, 86 p.
147
-------
Revisiting the Major Discussion Points of the Technical
Presentation Sessions
The statements made during the workshop do not represent the views or opinions of EPA. The
claims made by participants have not been verified or endorsed by EPA. Any mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
The workshop lead and theme leads addressed the workshop participants and EPA at the
conclusion of presentations for each theme and at the end of the workshop. Leads summarized
the major discussion points and commented on research needs and data gaps.
Scott Anderson of Environmental Defense Fund, the workshop lead, concluded the well
integrity discussions by providing some context on Texas oil and gas regulation. In 1919, Texas
adopted well plugging and spacing rules. However, one week after the spacing rules were
adopted, the first spacing exceptions were granted, and the well plugging rules were not
enforced for some time, maybe decades. Mr. Anderson noted that some defensiveness on the
part of operators is understandable when it comes to these issues; companies commit very
large amounts of time and effort to identify and control risks, and the industry is justifiably
proud of its accomplishments. However, all stakeholders recognize that things can and do go
wrong. Mr. Anderson also stated his opinion that regulatory issues are relevant to the EPA
study, even though Congress did not ask for a review of regulations. According to Mr.
Anderson, when understanding risk, the state of the regulatory environment is just as relevant
as industry's performance history and best practices.
Bob Whiteside of Texas World Operations, the Well Construction theme lead, concluded the
discussions by describing how HF operations have changed in the past 40 years. Drilling now
requires more equipment, communications, and personnel. Wells are now highly engineered
and extensively examined.
Tim Beard of Chesapeake Energy Corporation, the Fracture Design and Stimulation theme lead,
summarized the main points of each of the Theme 2 presentations. Terry Engelder's
presentation described natural fractures above the Marcellus Shale and emphasized that there
is no leakage without a driver of pressure differences. The datasets described by John Williams
show that fresh water and gas can be in close proximity in some cases. Tim Beard's
presentation described how fracturing is highly engineered and depends on many variables and
that fractures tend to stay in low-stress zones and respond to barriers. David Cramer and Hal
Macartney's presentations addressed the conditions that encourage horizontal fracture growth
in shallow formations. Patrick Handren discussed how microseismic data can show that
fractures are generally well contained in the Barnett and that increased well density leads to
increased fracture complexity. Norm Warpinski described mineback and microseismic studies
showing that vertical fracture propagation across layers is inefficient and therefore limited due
to differing rock properties and stresses. Ahmed Abou-Sayed's presentation showed that
pressure transient analysis can be a useful tool. Daniel Soeder proposed a tracer test. Scott
148
-------
Cline explained the numerous reasons for low fluid return rates, none of which point to flow
into ground water. Mr. Beard noted that evidence shows that the risk to ground water from HF
is remote. However, wellbore integrity, construction, and maintenance practices, from drilling
to abandonment and beyond, are all crucial. Mr. Beard added that it is easy to lose sight of the
facts in favor of opinions and personal issues. He believes that the facts show HF itself is not
necessarily the cause of any incidents, well construction should continue to be addressed in the
future. Mr. Beard also thanked all of the workshop participants.
Jim Bolander of Southwestern Energy, the Well Integrity theme lead, summarized the main
points addressed during the Theme 3 presentations and discussions. The goal of Theme 3 was
to merge the previous two themes, well construction and fracture design. Mr. Bolander
emphasized the importance of casing, cementing and pressure management. He believes that
appropriate consideration of those three topics can prevent problems like the one in Bainbridge
Township, Ohio.
149
-------
Glossary of Terms
The sources of the definitions found in this glossary are noted at the end of each definition.
Sources include the following:
Abbreviated Source Full Source Name
SPE Society of Petroleum Engineers Exploration & Production Glossary
(http://www.spe.org/glossary/wiki/doku.php/)
Schlumberger Schlumberger Oilfield Glossary
(http://www.glossary.oilfield.slb.com/default.cfm)
ABBREVIATIONS
BHTP bottom hole treating pressure
BLM Bureau of Land Management
BOP blowout preventer
BTEX benzene, toluene, ethylbenzene, and xylene
CBL cement bond log
CET cement evaluation tool
COGCC Colorado Oil and Gas Conservation Commission
DMRM Ohio Department of Natural Resources, Division of Mineral Resources Management
E&A Eckstein & Associates
E&P exploration and production
ECP external casing packer
EMW equivalent mud weight
ESOGIS Empire State Oil and Gas Information System
FIT formation integrity test
GIS geographic information system
HHP hydraulic horsepower
HMX cyclotetramethylene trinitramine
HNS hexanitrosilbene
ISIP instantaneous shut-in pressure
LTC long thread and coupled casing connection
MASP maximum anticipated surface pressure
MATP maximum anticipated treating pressure
MIT multi-finger imaging tool
MTT magnetic thickness tool
NACE National Association of Corrosion Engineers
NORSOK Norsk Sokkels Konkuranseposisjon
NWIS National Water Information System
OD outer diameter
OVESC Ohio Valley Energy Systems Corporation
P&A plugging and abandonment (of a well)
PA DEP Pennsylvania Department of Environmental Protection
PNL pulsed neutron log
RDX cyclotrimethylene trinitramine
RRC Texas Railroad Commission
SAPT standard annulus pressure test
SBT segmented bond tool
150
-------
SCP sustained casing pressure
SRV stimulated reservoir volume
STC short thread and coupled casing connection
TCEQ Texas Commission on Environmental Quality
Tcf trillion cubic feet
TD total depth (of a well)
IDS total dissolved solids
TDT thermal decay time
TOC top of cement
UCA ultrasonic cement analyzer
USDW underground source of drinking water
USGS United States Geological Survey
USIT UltraSonic Imager Tool
WFL water flow log
WHIP wellhead treating pressure
WOC waiting on cement
GLOSSARY
API American Petroleum Institute
back pressure a pressure caused by a restriction or fluid head that exerts an opposing pressure to flow (SPE)
casing head a term that applies to the wellhead flange that forms the transition between pipe and the flange-
build tree. It may be attached by threads, welding, pressure forming or lock-ring/screw devices (SPE)
casing string a continuous string of casing, usually cemented over at least part of its length and usually extending
back to surface from the set point (SPE)
control head an extension of a retrievable tool that is used to set and release the tool (SPE)
displacement volume the volume of a wellbore occupied by fluid. When the swept volume varies from the
calculated displacement, part of the wellbore may not be actively swept (SPE)
equivalent mud weight The equivalent mud weight felt by the formation when circulating with a certain mud
weight and holding a backpressure. A 10 Ib/gal mud in a 10,000 ft well with 1000 psi backpressure would
generate an equivalent mud weight of about 11.9 Ib/gal. (SPE)
external casing packer a rubber bladder over a section of casing that is inflated, usually with cement, to give an
annular seal in open hole sections. Frequently used with liners and set at intervals along the open hole. (SPE)
fracture gradient the gradient needed to initiate a fracture (SPE)
gamma ray log A log of the total natural radioactivity, measured in API units. The measurement can be made in
both openhole and through casing. The depth of investigation is a few inches, so that the log normally measures
the flushed zone. Shales and clays are responsible for most natural radioactivity, so the gamma ray log often is
a good indicator of such rocks. However, other rocks are also radioactive, notably some carbonates and
feldspar-rich rocks. The log is also used for correlation between wells, for depth correlation between open and
cased hole, and for depth correlation between logging runs. The gamma ray log was the first nuclear well log
and was introduced in the late 1930s.
hydrostatic pressure pressure exerted by a column of fluid (SPE)
Instantaneous shut-in pressure Used to isolate the formation fracturing or injection effect from the friction
effects (SPE)
Intermediate Casing often a casing string or liner run to isolate a zone between the surface casing and the final
production casing (SPE)
interval the pay zone exposed to the wellbore. This may or may not be the entire pay. Also referred to as
completion interval. (SPE)
kick tolerance an estimate of the volume of gas influx at bottom hole condition that can be safely shut in and
circulate out of the well (SPE).
mesh a measurement of particle size based on the openings per inch in a screen (SPE)
151
-------
mud weight mud weight The mass per unit volume of a drilling fluid, synonymous with mud density. Weight is
reported in Ibm/gal (also known as ppg), kg/m3 or g/cm3 (also called specific gravity or SG), Ib/ft3 or in
hydrostatic gradient, Ib/in2/ft (psi/ft) or pptf (psi/1000 ft). Mud weight controls hydrostatic pressure in a
wellbore and prevents unwanted flow into the well. The weight of the mud also prevents collapse of casing and
the openhole. Excessive mud weight can cause lost circulation by propagating, and then filling, fractures in the
rock. Mud weight (density) test procedures using a mud balance have been standardized and published by the
American Petroleum Institute (API). (Schlumberger)
pay zone hydrocarbon producing interval (SPE)
plastic viscosity an absolute flow property indicating the flow resistance of certain types of fluids. A
measurement of shear stress. (SPE)
play a pay zone or set of pay zones with proven commercial reserves (SPE)
pressure out see screen out.
production casing the innermost casing string that straddles and isolates the producing interval (SPE)
PWS public water system
rathole Extra hole drilled at the bottom of the hole to leave expendable completion equipment, such as the
carriers for perforating gun charges (Schlumberger)
rheology the study of the deformation and flow of matter. Real fluids include non-elastic solids, non-Newtonian
fluids and viscoelastic substances. The added materials that provide viscosity range from clays to polymers to
complex surfactant chemistry (SPE)
riser pipe through which liquid travels upward (SPE)
screen out an early time frac failure when the frac width is too small and the fracture proppant bridges off on
the fracture. (SPE)
shoe the end of the casing, usually called a guide shoe, that helps insert the casing through the drilled hole (SPE)
slickwater a water base fluid with only a very small amount of a polymer added to give friction reduction benefit
(SPE)
sour service defined in NACE MR-0175/ISO 15156 as exposure to oilfield environments that contain H2S and can
cause cracking of materials by the mechanisms addressed by NACE MR-0175/ISO 15156 (SPE)
SPE Society of Petroleum Engineers
spud to begin drilling (SPE)
wireline Related to any aspect of logging that employs an electrical cable to lower tools into the borehole and to
transmit data. Wireline logging is distinct from measurements-while-drilling (MWD) and mud logging (SPE)
workover repairing a well. Usually implies opening the well and running in with a tubing string. May or may not
involve killing the well and may or may not involve a conventional rig. (SPE)
yield point the resistance to initial flow of a fluid or stress required to start fluid moving (SPE)
152
-------
Printed with vegetable-based ink on paper that
contains a minimum of 50% post-consumer fiber
content and processed chlorine free.
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGES FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
------- |