A Completions Guide Book to Shale-Play Development A Review of Successful Approaches Towards Shale

CSUG/SPE 133874

A Completions Guide Book to Shale-Play Development: A Review of Successful Approaches Towards Shale-Play Stimulation in the Last Two Decades

K.K. Chong, W.V. Grieser, A. Passman, C.H. Tamayo, N. Modeland, B. Burke, SPE, Halliburton

Copyright 2010, Society of Petroleum Engineers

This paper was prepared for presentation at the Canadian Unconventional Resources & International Petroleum Conference held in Calgary, Alberta, Canada, 19–21 October 2010.

This paper was selected for presentation by a CSUG/SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract

Almost three decades have passed since the early exploration of the north Texas, Barnett shale. The Barnett serves as an example study for the shale life cycle. Operators in North America have used the Barnett-shale development as a roadmap for the exploration of new shale plays like the Marcellus, Haynesville, and Eagle Ford, as well as others. Each new shale play is unique in nature with respect to geologic setting, lithology, and production mechanism. It is useful to have a defined strategy for the discovery, development, and decline phases of each individual shale play. The roadmap to shale well-completion designs should include the following key factors:

?Fracability: capability of the reservoir to be fracture stimulated effectively

?Producibility: capability of the completion plan to sustain commercial production

?Sustainability: capability of the field development to meet both economic and environmental constraints

This paper reviews the evolution and development of completion practices of the major US shale reservoirs in the last two decades and presents a roadmap for effective completion practices for shale stimulation. The completion roadmap uses the history of 16,000 shale frac stages in the Barnett, Woodford, Haynesville, Antrim, and Marcellus shales. Following the map through specific decision points will alter the path for individual shales. These decision points will be influenced by geologic, geochemical, and geomechanical information gathered along the way. The path toward a commercially viable shale play from the early asset-evaluation phase to the late asset maintenance-and-remediation phase evolves from a series of decision trees throughout the process.

Information presented in this paper provides a completion engineer with better understanding of the factors involved in shale-play stimulation and provides a methodical approach to select appropriate and optimum solutions that have evolved during the last two decades.

Introduction

In the mid 1800s, expanding uses for oil extracted from coal and shale began. Gas production from the Devonian shale in the US can be traced back to 1821. A review of more recent shale exploration and development can be found in SPE reprint series No. 45, “Production from Fractured Shales” (Lancaster et al. 1996). As the title infers, commercial production from nanodarcy shale was most likely from the existing natural fractures providing “transmissibility and economic permeability.” Later, when more cores became available, the natural fractures that existed in most gas-shale plays like the north Texas Barnett were found to be filled with calcite or quartz, not oil or gas (Lancaster et al. 1996).

However, filled natural fractures are thought to have dilated during the massive hydraulic fractures used to stimulate the Barnett from 1985 to 1991. This dilation of filled fractures created the large fracture networks, exposing large surface areas and sustainable production.

The north Texas Barnett serves as a study in shale-completion evolution. Table 1 illustrates the Barnett development history.

2 CSUG/SPE

133874 TABLE 1—NORTH TEXAS BARNETT-SHALE HISTORY

Life Cycle Phase Year

Cumulative

Number of Wells

Comments

Discovery 1979 5 Will gas flow from shale Wise Co. TX

1981 6 What stimulation fluid N2 , CO2

1984 17 30 to 50 lbm Crosslink 100,000 gal. 200,000 lbm

1985 49 30 to 50 lbm Crosslink 500,000 gal. 1,500,000 lbm 20% N2

1988 62 Move to Denton, CO

1991 76 First horizontal TP Sims B-1

1995 200 Reduce gel loading to horizontal wells

1997 300 Continue polymer reduction 1st slickwater frac 500,000 gal 100,000 lbm 1999 450 Added upper Barnett to vertical completions, first fracture tilt mapping 2001 750 Increase frac volumes 1to 2 MM gal, 500,000 lbm, first microseism mapping 2002 1,700 Drilling begins outside core area

Development 2003 2,600 Horizontal program begins 85 horizontal, 117 directional, 719 vertical 2004 3,500 150 horizontals drilled 2 to 4 stage fracs

2005 4,500 600 horizontals drilled reduce cycle time

2006 5,500

Wells completed without Viola frac barrier

Water encroachment from Ellenberger

2007 7,000 Equipment, material, and manpower shortage

2008 9,000

Reduce footprint

Increase efficiency

Lower environment impact; recycle frac water

Refrac methods tested

Decline 2009 13,000 Maintain production Lower cost Improve efficiency

The historical development of the Barnett illustrates the changes that occurred as the number of wells increased and as information of the geology, geochemistry, and geomechanical properties improved. The obvious historical production increase and overall increase of estimated ultimate recovery (EUR)/well brought about by changes in the drilling and completion program provided the momentum for the 30-year Barnett success story.

Shales represent source rock where the hydrocarbons are still trapped over geologic time. They contain huge quantities of natural gas and oil locked (adsorbed) within the organic particles found within the rock. This is similar to coalbed methane, where a larger amount of gas can be stored in the rock at low pressure as “adsorbed” gas, in contrast to a conventional rock that does not have adsorbed gas. Stimulation technology has allowed hydrocarbons contained within these source rocks to be produced at economic rates. Important factors that operators should consider before developing a plan for a well in a shale formation are ?Each shale play is unique, and therefore the completion design evolves during the reservoir’s life cycle

?Each shale play is heterogeneous, so a normal variation in reservoir quality and production outcome will exist even in a closely spaced area

?Hydrocarbon storage and production mechanisms are unconventional and not well-understood; therefore, the completion strategy and outcome are complicated

?An initial reservoir assessment and quantifiable quality is necessary to benchmark each wellbore so that intelligent decisions can be made during the discovery and development stage of the play

? A recompletion strategy is an important component of the overall development design, especially in the decline cycle.

?Economic viability should not be evaluated based solely on reserves, but on productivity as well

?Productivity-viability decisions are based on workflows consisting of formation evaluation, stimulation, and reservoir production

?The ultimate drilling program is to reach the optimum sweet spot with minimal nonproductive time

?Successfully addressing an individual challenge may not make a shale asset profitable—critical factors should be analyzed as a system

CSUG/SPE 133874 3 Need For a Shale Roadmap

As other emerging shale plays begin their life cycle in the discovery phase, the need to reduce the learning curve and increase the economic sustainability becomes a major factor in the business decision to go after the shale asset. The range in production outcome is illustrated in Fig. 1, which shows the 12-month cumulative gas production from 12,000 Barnett wells. The wells range from 50,000 to 1,000,000 Mcf in the first 12 months. Most of the range in production is probably a result of change in reservoir quality. However, some percent of this production range is caused by drilling and completion practices. The idea of using a shale roadmap is to reduce the occurrence of low-end producers and increase the production in the midrange region.

Fig. 1—Range in production outcome in Barnett shale.

Reasons for Completion Evolution

The most common reason to change or evolve is to survive. Economic sustainability of a shale-development project probably drives most of the evolution in the discovery and development stages of the shale life cycle. Changes in other factors listed below can also cause the operator to change the drilling and completion strategy:

?Geologic (geohazards, faults, karsts)

?Geochemical (lithology, hydrocarbon type, or mineralogy changes)

?Geomechanical (stress-field changes caused by tectonic stress, faults, or production)

?Petrophysical (porosity, saturation, permeability, or bottomhole pressure (BHP))

?Logistics (fluid, proppant, equipment, or manpower availability)

?Economics (well cost, product price, net present value (NPV), or return on investment (ROI))

?Environmental (state, bureau of land management (BLM), resource sustainability, safety, or cultural sensitivity)

In most shale plays, one of the most basic universal laws is that “change happens.” Most operators plan the early-discovery stage to include the highest degree of diversity in initial drilling-site selection to quickly discover the magnitude of change that can occur. Knowing what is being done correctly and what is being done incorrectly is usually measured in the time required to drill and complete the well, along with the associated cost. While this can satisfy the economic-sustainability need, it can lead to mediocre, unpredictable, or unexpected production outcomes. Meaningful production data usually lags the drilling and completion timetable by 3 to 6 months. A large-scale shale development can be 50 to 100 wells into the program before a need for an operational change or a change in the basic assumptions becomes apparent. This is why collaboration between all the stakeholders is important. Preparation of a roadmap and plan for contingencies should be encouraged, financed, and supported by the organization in advance of the first exploration well. Flexibility and authority to quickly change direction with changing geology through environmental factors can allow a sustainable harvesting of the shale asset.

4 CSUG/SPE

133874

Shale-Resource Workflow

This workflow (Fig. 2) serves as a basis for early shale development with the goal to shorten the learning curve in new shales. The

workflow is used as a baseline and then customized to fit the needs of the operator and shale formation. This workflow is designed

to

Evaluate specific behavior of key factors that impact the system

Identify potential locations of sweet spots

Evaluate geochemical and geomechanical parameters

Determine wellbore geometries

Evaluate completion and stimulation strategies

Predict and evaluate well performance

Optimize wellbore programs

Fig. 2—Shale-resource workflow.

CSUG/SPE 133874 5 Shale Well-Completion Roadmap Key Factors

Each new shale play is unique in nature with respect to geologic setting, lithology, and production mechanism. It is useful to have a defined strategy for the discovery, development, and decline phase of each individual shale play. The roadmap to shale well-completion designs should include the following key factors:

?Fracability (capability of the reservoir to be fracture stimulated effectively)

?Producibility (capability of the completion plan to sustain commercial production)

?Sustainability (capability of the field development to meet both economic and environmental constraints) Fracability

Coupling openhole wireline-log information with laboratory measurements of core or cutting samples provides a basis to calibrate the petrophysical model that describes essential geomechanical and geochemical characteristics of the shale. One of the characteristics is fracability.

Finding areas in the shale play that are fracable is important in the development of a fracture fairway large enough to connect the highest amount of the rock volume to the parent wellbore during the hydraulic-fracturing process; Fig. 3 depicts qualitatively the difference in the shale with low and high “fracability” fracture networks.

Fig. 3—Fracture fairway geometry created in low/high “fracability” shale.

Complex fracture geometry during hydraulic-fracturing treatments can lead to “pressure outs” and screenouts. One cause of the “pressure outs” is high process-zone stress (PZS). With high PZS, the chance for pressuring out is greater than screenout. Ramurthy et al. (2009) presented a study that clearly shows a relation between high PZS and poor production and low PZS and better production. A diagnostic fracture injection test (DFIT) can be used to identify high-PZS zones. Modification to injection rate increase, change in viscosity, and addition of acid may help reduce PZS effects.

Producibility

Identifying the producing potential of unconventional reservoirs requires knowledge of the geology, geochemistry, petrophysics, mineralogy, and rock mechanics. Identification of productive shale includes location of high organic content, thermally mature, porosity enhanced, high “fracability” rock. This can be accomplished with standard logging tools using advanced processing techniques calibrated with core-testing data. Ultimate production outcome is influenced by many factors, including ?Total organic content (TOC) and maturation (R o)

?Producable hydrocarbon content

?System permeability height (kh)

?Reservoir pressure

?Capability of the rock to create complex fracture networks “fracability” or brittleness

?Stimulated reservoir volume (SRV) and fracture-complexity index (FCI) (Grieser et al. 2007; Julia et al. 2006) Percent TOC is reported for most shale plays and is usually one of the major factors used when judging the productivity of the shale. However, TOC is simply the organic material remaining in the shale in its present form. While most productive shale has a minimum TOC value, the absolute magnitude has little to do with productivity. The amount of organic material converted has a much greater impact on production. Most midcontinent shales have converted ± 50% of the original organic material in place. This conversion is believed to be a contributor to the shale porosity and permeability. For example, it is estimated that the original organic content of the north Texas Barnett was 8%. As the organic material was converted to oil and gas, about 4% was used. So, today there is still about 4% TOC in the Barnett. The real number that affects rock properties is the percent of organic material converted. This is because conversion of organic material to hydrocarbon increases the bulk porosity of the rock. There is usually not much of a direct correlation between brittleness and TOC. There are many high-TOC shales that are ductile, immature, and nonproductive (Sylvan shale).

6 CSUG/SPE

133874

Sustainability

The main objective for a shale-fracture treatment is to obtain a large, highly fractured network that can produce from the ultralow-

permeability rock. This is usually achieved by creating a multiple-stage fracturing treatment in the long horizontal lateral. The

extent of the fracture fairway or complex network can be determined from microseismic-fracture mapping. Two terms commonly

used in the industry are stimulated reservoir volume (SRV) and fracture-complexity index (FCI), which have been described in

numerous papers published throughout the past few years (Warpinski et al. 2008; Mayerhofer et al. 2006; Cippolla et al. 2008;

Julia et al. 2006; Ramurthy et al. 2009; Grieser and Bray 2007).

FCI is defined by Cipola et al. (2008) as the ratio of total width of the microseismic cloud to the total length of the cloud, with

planar fractures having a relatively small FCI and network fractures having a larger FCI value. Fig. 4 shows a FCI plot.

Fig. 4—Complexity plot showing FCI values (Cipolla et al. 2008).

The SRV is defined as the product of gross stimulated area as measured by microseismic mapping and shale-pay thickness

(Mayerhofer et al. 2006). The case study showed how three years of cumulative gas production in the Barnett shale would be

affected by larger SRV and closer frac spacing to make a better well (Fig. 5). Fig. 6 shows a plot of 6 months of cumulative gas

versus SRV for the Barnett core area (Mayerhofer et al. 2006) and clearly indicates that the higher SRVs result in better well

performance.

CSUG/SPE 133874 7

Fig. 5—Observed and modeled production as a function of fracture spacing (Mayerhofer et al. 2008).

Fig. 6—SRV vs. well performance (Mayerhofer et al. 2006).

Increasing complexity can be accomplished by simultaneous fracturing or zipper fracturing. Simultaneous fracturing involves pumping down two or more wellbores at the same time into a common interval. Zipper fracturing is sequential fracturing of two or more wellbores in a common interval. Both the simultaneous- and sequential-fracture treatments create temporary stress fields that increase the FCI, thus increasing the SRV. Fig. 7 shows the impact of network size on gas production. Doubling the SRV almost doubled the well performance (Mayerhofer et al. 2006).

In the nanodarcy shale-gas rock, the SRV created is greatly impacted by fracture spacing, job volume, rate, viscosity, and proppant amount. Other factors include drilling, injecting, or producing of offset wells. Restimulation can add new SRV and increase FCI because the stress state might have changed since the last large treatment. The shale-gas production has typically steep production decline with the first few months; restimulation would boost the early production where the SRV and FCI were lower in the initial treatment. Valko (2009) published his work on over 10,000 wells with production history in the Barnett shale, demonstrating the successful restimulation on the group with 10 of the oldest wells in the Barnett shale.

8 CSUG/SPE

133874

Fig. 7—Impact of network size on cumulative gas production (Mayerhofer et al. 2006).

Shale Roadmap Components

The following is a list and description of the shale roadmap components. These components describe the following:

?Definitions of terms

?Critical data that needs to be collected

?Observations, tests, and diagnostic procedures used to stay on the path

?Tools, methods, and procedures used to shorten the path and stay on course

Geologic (Geohazards, Faults, Karsts)

In the initial phase, it is important to begin with a baseline understanding of the geologic interpretation of the area. In this phase, it

is imperative to

?Gather any available data, including maps, logs, cores, production reports, and interpretations, etc. into a single location

to compare and evaluate

?Gain an understanding of the current maturity of geologic models

The depositions of most shale formations are a series of thin laminations. These layers can alter the advancement of the

fracture and can blunt fracture growth or slip at the interface, causing a complex fracture path and possible horizontal fracture

extension. Shale reservoirs are not homogeneous deposits of organic shale. They usually consist of multiple thin layers of variable

lithology. These layers have a wide spread of mechanical properties and lack of adhesion strength. This can result in layer

slippage. The most common lithology combination is bulk shale, silt layers, laminated shales, and natural fractures, respectively.

Shales include the indurated, laminated, or fissile claystones and siltstones. Pressure of overlying sediments, dewatering during

diagenesis, and plastic flow produces anisotropy and secondary bedding. Shales are typically deposited in low-energy

environments and are often found in lake and lagoon deposits, river deltas, on flood plains, and offshore beach/sand-bed systems.

The productive shale plays are unique in that they are source rock, reservoir rock, and trap. In general, there is an agreement that

there are free and adsorbed components of hydrocarbon in most shale plays. A particular theory that has acceptance suggests that

the free gas is stored and produced from the microporosity in lamina and natural fractures, and the adsorbed gas is stored and

produced from the bulk-shale matrix. The production from shales is generally viewed as a three-component storage/flow unit.

?Laminar layers of siliceous or carbonaceous material that can have conventional porosity storage and flow (macro

porosity in thin laminations of chert, dolomite, calcite, and sand).

?Natural fractures can contribute when mineral deposition has not taken place (or nonhealed fractures).

?Black organic bulk shale that can feed both the laminated layers and the natural fractures and desorb gas through the

fracture-network surface area.

CSUG/SPE 133874 9 The adsorbed gas is believed to be attached to the organic and clay material in the bulk shale. Enhanced overall surface area of a shale will increase the desorption rate of gas from the shale. The amount of gas released should be directly proportional to the amount of surface area created. Gas production from hydraulically fractured shale is believed to come from desorption and diffusivity from the microporosity/fractures. The geologic study defines the regional extent of the play while mapping the structure, including tops, bottoms, faults, and karst features. Defining locations, size of the map structure, and map faults should be the first thing to evaluate on any shale play. The subsurface dimensions true vertical depth (TVD) of top and bottom barrier rock needs to be determined next. It has been found that seismic interpretation can provide valuable insight. To potentially identify high-quality areas for development, seismic analysis should be performed to

?Analyze available well logs with emphasis on full log suites; in particular, full waveform sonic and density logs

?Run models including fluid substitution to determine the response of pay in the shales

?Determine the applicability of amplitude variation offset (AVO) analysis in determining the location of “sweet spots”

?Compare the model to the seismic data

?Correlate any wells with pay to the 3D seismic data and perform an attribute analysis including (but not limited to) amplitude, phase, waveform classification, etc.

?With additional data, iterate the process to improve identification of pay

If seismic data is not available with the appropriate logging program, including Vp and Vs sonic logs, density, gamma ray, effective porosity, Sw, and Vshale, offset synthetic seismic data can be created and analyzed using the similar methods and processes as surface seismic data.

Geochemical Lithology, Hydrocarbon Type, or Mineralogy Changes)

An initial quantification of reservoir quality is necessary to benchmark each wellbore so that intelligent decisions can be made during the discovery and development stage of the play. This identifies where the pay is coming from in the formation to optimize completions and stimulation programs. It also provides evaluation of the reservoir fluids and rock to determine applicable programs. Geochemical analysis and screening provides an understanding of how the system will react to different programs. It is critical to have a strong understanding of the geochemistry, including analysis and evaluation of

?TOC, kerogen typing, and thermal maturity measurements to evaluate the formation for potential hydrocarbon generation, production, and hydrocarbon type. Table 2 illustrates kerogen typing and Table 3 touches on the concept of thermal

maturity

?Mineralogical analysis that use X-ray diffraction (XRD), glycolation testing used to profile the mineralogical traits, and propensity for clay swelling to determine viability and production capability

?Fluid screening to determine the compatibility of core samples with a variety of base fluids

?Acid solubility and capillary suction testing (CST) to understand swelling tendency and reactivity when formation samples are exposed to chosen base fluids

?Fracture-treatment screening to identify fluids and additives that reduce formation damage

?Rock mechanical properties-correlation back to acoustic logs

?Shale's ductility or brittleness by determining Young's modulus, Poisson's ratio, and Brinell hardness

?Lithology and potential reaction to the treatment fluid by conducting scanning electron microscopy (SEM)

?Correlation of geomechanics and geochemistry models

?Regional stress and fracturing

?Comparison of analog reservoirs

10 CSUG/SPE

133874

TABLE 2—KEROGEN TYPES AND CLASSIFICATION

Kerogen Type Depositional Environment Organic Precursors Hydrocarbon Product

I Lacustrine Algal bodies or structural debris of algal

origin

Very H-rich; very good for oil

II Marine, reducing conditions

Skins of spores and pollen, cuticle of

leaves and herbaceous plants

H-rich; precursors for oil and

gas

III Marine, oxidizing conditions Fibrous and woody-plant fragments and

structureless colloidal humic matter

H-poor; mainly precursors for gas

IV

Marine, oxidizing conditions Oxidized, recycled woody debris

Very H-poor; largely inert, but may

produce gas at very late stages of

maturation

TABLE 3—THERMAL MATURITY CLASSIFICATION

Ro%HC-Type

0 to 0.55 Onset of oil generation

0.55 to 0.9 Peak oil production

0.9 to 1.1 Wet gas

1.1 to1.4 Dry and wet gas

1.4 to

2.1 Dry gas only

> 2.1 CO2

Geomechanical (Stress-field Changes as a result of Tectonic Stress, Faults, and Production)

Geomechanical understanding is often overlooked as part of the planning and design process. Developing a geomechanical model can provide insight into predicting predrill pore pressures and improving wellbore stability. The model is created using all available analog, seismic, and geological information calibrated to direct measurements of pressure, in-situ stress, and observations of wellbore behavior. This model can be used to make trajectory-specific predictions of pore fluid, shear failure, and fracture pressures so that future wells plan can be designed to minimize formation-fluid influx, drilling-fluid loss, and wellbore instability. In addition, well trajectories can be designed to optimize fracture placement and direction, ultimately improving recovery (acceptable well trajectories should be determined if compromises between drilling efficiency and production enhancement are required). Geomechanical analysis should be such that it

?Correlates data, including drilling reports, logs, downhole measurements, and geologic and mud data

?Evaluates drilling records to relate any mechanical-instability incidents to depth, wellbore inclination and azimuth, mud density, and drilling activity

?Constructs 3D pore pressure, in-situ stress, and rock-properties model

?Develops a rock-strength model using log-based correlations and regional analog rock-strength data

?Characterizes stress-induced wellbore failure, if present

?Determines in-situ stress regime and the orientation of horizontal stresses

?Back-analyzes the borehole-failure data to constrain the magnitude of maximum horizontal stress consistent with borehole and pore pressure, rock strength, and drilling parameters

?Verifies and calibrates the strength and stress model against drilling experiences and borehole-failure data in offset wells

?Models wellbore stability to evaluate the well pressure at which excessive shear failure will occur at the borehole wall; the results will provide detailed stress magnitudes with depth consistent with the local in-situ conditions ?Defines the optimum mud weight and makes recommendations on casing-seat locations

CSUG/SPE 133874 11 Petrophysical (Porosity, Saturation, Permeability, and BHP)

The single most important step in shortening the shale learning curve is gathering the right data early in the life cycle. The right data does not refer to “as much data as possible” but instead to the necessary capital expenditure on data acquisition that provides the critical knowledge for success. Not only are core and cutting analysis important, but running the right logs can be a game changer. Wireline logging is an integral part of the formation-evaluation process. The logging program should consist of the following logging suite:

?Compensated spectral natural gamma ray for the identification of clay minerals and uranium-rich intervals that could be areas of high organic content

?Nuclear and acoustic porosity measurements for the determination of porosity and lithology

?Array resistivity to identify permeable zones indicated by invasion effects on the quint resistivity curves, as well as formation resistivity

?Wave sonic for rock-property evaluation by measuring both the fast and slow shear slowness

?Elemental analysis tool is used for precise evaluation of complex mineralogies and is useful in evaluating unconventional reservoirs

?Magnetic resonance logs for the determination of pore distribution and the identification of free porosity

?Extended range micro-imager for identifying natural and induced fracturing and sedimentological interpretation Design and Implementation.While well construction, including drilling fluid and casing design, is not a focus of this paper, it is important to mention that this phase of planning should not be neglected.

If the shale consists of discontinuous, stacked, multiple potential pay intervals over a long gross interval, the well will most likely be completed as a vertical wellbore with a multiple-stage fracture-stimulation treatment. If the shale is a single zone, then a horizontal wellbore might be the best application for an effective completion. When considering a horizontal wellbore, many options are possiblean openhole completion, cemented-casing completion, uncemented casing with perforated sections, or an uncemented casing with isolation packers and stimulation sleeves. If the well is to be a horizontal well in the exploratory phase, plan a vertical pilot wellbore so that formation samples and openhole logs can be obtained.

Shale Hydraulic-Fracturing Design.This design is based on SRV. Hydraulic fractures in shales usually do not take the form of a single biwing fracture. Instead, they are referred to as “fracture networks” consisting of many fractures of various lengths, heights, and widths. While a complex fracture-network model might be more appropriate for shale-fracture design, other methods using conventional multiple-fracture wings or adjustments to leakoff coefficients can approximate basic fracture geometry. Fracturing modeling using the appropriate 3D model is a critical component of the optimization process to understanding fracture penetration. Fracturing Fluid and Proppant Selection. Fracturing fluid-type selection is a critical decision point in the completion process. It is important to note that rock properties alone do not help to choose a shale fracturing fluid. For example, ductility alone does not help choose a shale fracturing fluid. If the shale is ductile, it will probably generate a more conventional biwing fracture. Also, when it closes, any unpropped area will probably not be conductive. So for ductile shales, a fully packed fracture is more desirable. Ductile laminated shales will probably require a fluid that can distribute proppant across the entire fracture height so that vertical permeability is established. Fig. 8 shows how selection of a fracturing fluid changes when moving from a high-permeability, more-ductile rock to a low-permeability, brittle rock.

133874 12 CSUG/SPE

Fig. 8—Relationship between fluid type and rock mechanical properties.

Like with conventional reservoirs, the operator looks for the most cost-effective fluid that is “compatible” with the formation

and reservoir conditions. Fines and sludges can be generated in all types of fracturing fluids. Any time formation fines, oil, and

fracturing polymers are together, there is potential for creation of emulsions and pore blockage. Maximum proppant concentration

in shale fracs is usually dependent on the frac width generated. Frac width can be varied with rate and viscosity. Because of the

low permeability of shale (0.0003 to 0.00001 md), the required proppant conductivity is also relatively low. For a high-

permeability shale (0.0003 md) and a long fracture half length (2000 ft), the required optimum fracture conductivity is ~10 to 100

md-ft. Softer shales, where proppant embedment would be more significant, will probably require higher lbm/ft2 proppant

concentrations to achieve the same fracture conductivity.

Until recently, the most common stimulation fluid for shale has been fresh water with a neutral pH. This fluid has the benefit

of being environmentally low-risk and the most economic system to pump. The consumption of large volumes of fresh water in

some areas, sometimes exceeding 100,000 bbl per wellbore, with as little as 20% frac fluid recovery led to consideration of the

effect of the large fresh-water volumes on the shale formations. The recent use of reactive fluids, such as those reported by Grieser

et al. (2006) (Fig. 9), has generated some excitement in reported production increases over conventional “water frac”

treatments. These early initial production results are difficult to verify, and a longer production history for multiple wells is

probably needed to see the full effect of these treatments. While the reported benefits of pumping reactive fluids have not been

fully investigated, there are some initial investigations underway to help understand the interaction between shale lithologies and

reactive fluids. It is important to note certain important points about usage of reactive fluids in shale stimulation treatments;

following are some of the key parameters to keep in mind with regard to shale-reactive fluids:

?Shale-reactive fluids expose micropores and increase surface area

?Most shales tend to exhibit some form of disruption when immersed in various fracturing fluids

?Reactive fluids tend to remove soluble minerals in the shale, exposing microfractures, and enhance etching and create

new surface

?Increased access to microporosity and/or natural fractures is believed to enhance shale production

?Wells treated with reactive fluids still need to be monitored to see long-term effects on production and scale precipitates

CSUG/SPE 133874 13

Fig. 9—Job plot from SE Oklahoma Woodford shale frac showing dramatic/unexpected pressure drop as reactive fluid hits the perfs (Grieser et al. 2006).

DFIT

A DFIT is a small-volume, low-rate injection followed by a shut-in period. There are four objectives of the diagnostic fracture injection test:

?Breakdown the formation

?Estimate closure pressure and pore pressure

?Estimate system permeability or (kg)(h)

?Determine fluid leakoff type

Fig. 10 shows a G-function plot signature curve that can be used to analyze the DFIT data.

Fig. 10—G-function plot analyzed based on DFIT.

14 CSUG/SPE

133874

The recommended injection volume for tight gas is small (10 to 24 bbl). The injection rate should approach the frac rate if

possible; however, injection rates of 5 to 20 bbl/min have been used successfully in some shales. Downhole quartz gauges are used

and should have an accuracy of ±0.01 psi. Shut-in time of a minimum of 24 hr is usually required to see closure. It is important

that the well remain static through the entire shut-in period. Bleeding pressure off backside or surface iron will distort the data. The

wellhead and any surface connections or gauges must be leak-tight.

Microseismic Fracture-Mapping Evaluation

Microseismic-fracture mapping has proven to be useful for determining important fracture parameters, including height, length,

and azimuth, as well as SRV and fracture complexity. It has been shown that SRV is directly related to production (Mayerhofer et

al. 2006; Mayerhofer et al. 2008; Warpinski et al. 2008; Warpinski 2009). A study was made in the early development of the

Barnett shale to determine the average SRV of water-fractured vertical wells. The average dimension of the SRV for a typical

1,000,000-gallon water frac is shown in Fig. 11.

Fig. 11—The average dimension of the SRV for a typical 1,000,000-gallon water frac.

Drainage area is believed to be directly related to fracture surface area. Thus, the basic fundamental of a stimulation design is

to expose as much surface area as possible at the least amount of cost. High-rate 50- to 80-bbl/min water fracs appear to meet this

requirement best.

For the Barnett, the general relationship between fracture half length (X f) and volume pumped (bbl) is

Fairway half length =X f ~ 44 (bbl) 1/3

Fairway width ~X f/2

Fairway thickness ~h

Feedback—Evolve and Optimize

Optimization uses the data from the initial pilot well(s) to evaluate what-if scenarios; each time a new well is drilled, completed,

and stimulated, the data are input into the feedback loop to quickly determine the ideal capital spending program and fine-tune the

development plan. A critical component of the optimization process is determining the viability of a horizontal program. Drilling

success in any reservoir is dependent on finding the most prospective areas, or the “sweet spots,” and aligning the wellbore for

maximum borehole exposure to these zones, whether it is vertical or horizontal. In shale reservoirs, this means placing the well in

the zones most conducive to fracturing. This requires a thorough understanding of the shale gas-reservoir characteristics from

testing and formation-evaluation analysis (aiming for the middle is rarely a successful strategy because shales can have significant

variance in thickness and composition). What-if scenarios are evaluated to determine which stimulation program ideally exploits

the resource. Then, by predicating whether the horizontal shale program increases production (whether IP or cumulative) and/or

increases reserves recovered, optimization economics determine if the additional capital expenditure is worthwhile.

CSUG/SPE 133874

15 Fig. 12—Shale roadmap path and feedback loop.

As the shale development passes through each department or silo, decisions are made based on the information available and the external conditions that exist at the time. Each decision affects the path to the next team member (Fig. 12). In early-discovery or development mode, success is usually judged by initial production. However, as more production history becomes available and a projected EUR can be estimated, the decision process can change and alter the roadmap to achieve a more-sustainable development.

US Shale-Completion Overview

Most US shale plays are now all horizontal-type completions. This drilling and completion type exposes the maximum amount of shale rock per wellbore. Various completion types have been investigated. These include

?Openhole, no pipe no isolation

?Open annulus, slotted or pre-perf liners

?Open annulus, compartmentalized with mechanical or swell packers

?Cemented, isolation

The most-favored completion method has some sort of isolation or compartmentalization aspect. Table 4 lists the most common or average stimulation data for various shale plays in the US.

TABLE 4—AVERAGE STIMULATION DATA FOR VARIOUS US SHALE PLAYS

Barnett Haynesville Marcellus Woodford Bakken Eagleford

TVD, ft 7,000 to

8,000

10,000 to

13,500

6,500 to 7,500 7,000 to 13,000 7,450 to 11,010 6,000 to 13,000

Horizontal length, ft. 3,000 to

5,000

4,000 to 7,600 4,000 to 5,500 3,000 to 5,000 4,000 to 10,000 3,500 to 4,500

Number of stages 4 to 6 10 to 18 6 to 19 6 to 12 5 to 37 7 to 17 bbl/Stage 17,100 10,600 10,000 17,000 1,800 12,500 sk/Stage 3,500 3,500 4,000 3,500 1,500 2,500

Rate, bbl/min 70 to 80 70 80 70 to 90 15 to 20 35 to 100

Avg. psi 3,000 to

5,000

10,500 to

14,000

6,500 to 8,700 5,000 to 13,000 2,800 to 8,000 9,000 to 12,500

Avg. lbm/gal 0.57 2.5 2.5 1.0 2.0 to 2.5 1 to 1.5

Fluid type FR-water

Linear gel

FR-water

Linear gel

Crosslink

FR-water

Linear gel

Crosslink

FR-water

Linear gel

Hybrid

Crosslink

FR-water

Linear gel

Crosslink

Proppant type

100-mesh

40/70 Sand

30/50 Sand

100-mesh

40/70 ISP

40/70 RCP

30/50 ISP

100-mesh

40/70 Sand

30/50 Sand

100-mesh

40/70 Sand

40/70 CRC

100-mesh

20/40 Sand

40/70 Sand

20/40 Ceramic

100-mesh

40/70 Sand

30/50 Sand

16 CSUG/SPE

133874

Miller et al. (2008) reported on their study for 423 laterals in 301 wells from Montana and North Dakota, Bakken play, and

indicated that wells completed using compartmental completion methods are among the best type of completion practice to date in

this play and were the best-producing group. The well with the highest normalized best period production (BPP) can be

characterized as follows:

?Lateral length: 4,000 to 7,000 ft

?Completion type: compartmental or preperforated liner

?Stimulation fluid: crosslinked

?Stimulation volume: greater then 150 gal per foot of lateral length

?Proppant type: 20/40-mesh or smaller natural sand or man-made proppant

?Proppant mass: greater then 300 lbm/ft of lateral length

?Average proppant concentration: greater than 2.5 lbm/gal

Figs. 13 and 14 show the normalized fluid and proppant volumes of the best producing group of Bakken wells with

compartmental completions and indicate more fluid and proppant increase production (or more fracturing stages; more than 20

with tendency to 30+).

Fig. 13—Production by ft, proppant per ft (based on Miller et al. 2008).

CSUG/SPE 133874 17

Fig. 14—Production by ft, frac fluid per ft (based on Miller et al. 2008).

With more-advanced technology and gained experience, the completion learning curve of the Haynesville shale has been seemingly more expedited than some of the earlier unconventional shale plays. However, the evolution is still in process, as operators continue to seek the most optimal solutions of producing the Haynesville shale reserves; theories and opinions differ between companies, but many of the major overlying paradigms of unconventional completions that impeded the progress of other plays have already been shattered. Beginning in 2009, as the majority of operators began their ventures into the Haynesville shale play, a wide variance could be seen in completions strategies, ranging from low-volume, high proppant-concentration completions to very large fluid volumes with low proppant concentrations. As the year progressed, the completion strategies became less variant as operators likely began unifying the completion strategies across the majority of their wells with what seemed to work best for them; thus, allowing an established baseline to which future changes to completion strategy could be quantified for effectiveness.

Analyzing the horizontal Haynesville shale completions trends beginning in 2009 (Fig. 15), it is clear that there has been a steady increase in the number of stimulation intervals being placed in the lateral. As the number of intervals per lateral have increased, the amount of proppant and fluid volume per interval also seemed to increase for the majority of 2009 (into the 4th quarter). However, in the fall of 2009 (likely after many operators had established a baseline productivity goal from their completion strategy), a trend of decreasing interval-fluid volume appears that is not mimicked by the interval proppant amount, which remained rather flat or decreased only slightly into the first half of 2010. Evaluating these trends can give a first glimpse as to where the Haynesville shale completion strategy is evolving. While speculation could suggest that the decrease in fluid-volume amount is caused by the increase in intervals over the lateral length (and thus smaller fracs can be placed as more fracs are being placed in the lateral), the reality is that the proppant amount remaining flat into 2010 does not support this idea. What could, however, be postulated is that as operators begin to make calculated step changes in attempts to optimize the completion design, beneficial results are being found in increasing the lateral coverage of the well (thus the continual increase in stages) and also by creating better conductivity in the fractures (which explains the steady proppant amounts with decreasing fluid volumes), both of which reiterate the evaluation that the rock properties of the Haynesville shale are less conducive to creating far-reaching complex fractures and are more subject to embedment and fracture rehealing than the Barnett shale and consequently require more conductive proppant packs.

133874 18 CSUG/SPE

Fig. 15—Haynesville shale completion roadmap.

The Marcellus play has also quickly closed the gap on uncertainty the past 3 years. With the Marcellus being located in one of

the more challenging areas of the country compared to its unconventional counterparts, operators have had to handle more than

just what is best for the reservoir. They have also had to determine what is economical, logistically feasible, and the optimized

treatment type to use. The early months of the Marcellus were to be more of a trial-and-error scenario, where operators were

testing their acreage to define the production fairways that proved to have the highest chance of success. Slickwater fracs were

predominantly used as the main fracturing fluid, with linear hybrids and crosslink hybrid tailins being used; as well, a few N2 assist

jobs were tried. Stage volumes varying from 8,300 to 10,900 bbl were being pumped. The actual stimulated treatment interval was

being explored and spacing of perforation clusters continued to be altered and tested using microseismic technology to identify

how efficiently each lateral was treating. At that time, the average number of stages ranged from four to eight, depending on the

length of the lateral. Sand volumes fluctuated slightly (Fig. 16) from 3,500 to 5,000 sk per stage, depending on operator.

The climb up the learning curve that has been seen over the past 3 years has changed only slightly but within a much shorter

time frame than its unconventional counterparts. Different types of proppant have been tested and eliminated because of economic

feasibility. Lateral lengths have tripled in length in some cases as the ROI is identified. Also contributing to the longer laterals are

difficult surface conditions, making pad drilling appear to be more appropriate and economically viable in order to efficiently

cover the entire acreage block. One thing to point out is that the average fluid volume pumped decreased slightly over the past 3

years (Fig. 17). This is misleading because of the fact that if a wider scope is viewed, such as the average number of stages

pumped and the overall well volumes, then the overall fluid volumes as well as overall sand volumes per well have been

consistently increasing because of laterals getting longer and spacing of stages getting tighter. With that being said, Marcellus

operators are still adjusting their frac designs according to what regions they are operating in. The northeast section and the

southwestern section of the state are completed differently, depending on what each operator has identified as being the primary

driver to better production, whether that is fluid volume, sand volume, sand type/mesh, or stage spacing.

Other treatment parameters, such as averarge treatment rate (Fig. 18) and treatment stages per wellbore (Fig. 19), trended

upward during the past three years.

CSUG/SPE 133874 19 Marcellus Completion Trend (2008 through Q1 2010)

Fig. 16—Marcellus shale completion trend; proppant (sk) per stage.

Fig. 17—Marcellus shale completion trend; treament volumes (bbl/stage).

20 CSUG/SPE

133874

Fig. 18—Marcellus shale completion trend; rate (bbl/min) per stage.

Fig. 19—Marcellus shale completion trend; number of stages per wellbore.

Woodford Completions Historical Perspective

The Oklahoma Woodford exploration and early development stage was plagued by completion problems that had to be overcome.

The historical completion history of the OK Woodford is provided in Fig. 20 and Table 5. Injection problems plagued the

completion program, resulting in less than 40% of stages pumped to completion in the horizontal lateral.