SPE 136183

Engineering a Successful Fracture-Stimulation Treatment in the Eagle Ford Shale N.A. Stegent, A.L. Wagner, J. Mullen, Halliburton; R.E. Borstmayer, GeoSouthern Energy

Copyright 2010, Society of Petroleum Engineers This paper was prepared for presentation at the SPE Tight Gas Completions Conference held in San Antonio, Texas, USA, 2�–3 November 2010. This paper was selected for presentation by an 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 The Eagle Ford shale in south Texas is the one of the most recent developments in unconventional reservoir exploration. Numerous existing completion methods have been applied in this nano-darcy formation with various degrees of success. The horizontal Eagle Ford wells in the northeast area of the current Eagle Ford play (DeWitt County) require a completion strategy that is reservoir specific. The production in this area has a high liquid/gas ratio and presents different challenges for commercial development than those in the typical �“high-rate water frac�” completions associated with dry-gas shale stimulation theory. Previous high-rate water frac completions in this area typically associated with the �“Barnett-style shale stimulation�” achieved poor results. Core analysis shows that a low Young�’s Modulus (YM) (soft rock), high clay content, and the potential for high liquid-hydrocarbon production require the need for a different completion strategy. Swelling formation clays and proppant embedment were formation issues to consider along with the multiphase hydrocarbon production. Higher conductivity fractures would be required, but various unknowns existed:

�• How many frac stages should be pumped? �• How much proppant should be pumped on each frac stage? �• What type of proppant should be used? �• What mesh proppant should be used? �• What perforation scheme was needed? �• What type of completion fluids should be used? �• What injection rate was needed? �• How would fracture-injection issues be handled?

This paper discusses how a collaborative, engineered approach was applied to the completion of the Eagle Ford shale to

deliver a commercial asset. To address the unknowns, the methodology included geologic and reservoir understanding applied to the stimulation design and execution. The stimulation resulted in hydrocarbon production that exceeded expectations. Comparative well results will be discussed. Introduction The first question typically asked regarding the stimulation completion in an unconventional reservoir is �“based on production results, what is the best way to complete this reservoir?�” In a relatively new play, this is a very difficult question to answer because little or no offset-well data exists. The next logical step in addressing this challenge is to use the experiences of others in similar reservoirs. Polling colleagues and asking for recommendations is a good place to start the climb up the learning curve, but it alone can�’t provide an adequate completion strategy. There will be enough subtle differences between the different shale reservoirs that a �“one-size-fits-all�” approach is probably not the best. In addition to conventional reservoir properties, such as porosity, resistivity, water saturation, permeability, pore pressure, and mechanical rock properties, shale characteristics such as the brittleness, ductility, complexity, mineralogy, laminations, etc., also need to be determined. A more complete reservoir characterization provides the asset team with knowledge to design the best completion for the well. Proppant volume, proppant type, fluid type, injection rate, number of frac stages, perforation scheme, and other completion decisions are made based on reservoir description and knowledge, not just experience in other areas. Petrophysical analysis is the key to understanding the reservoir. Unfortunately the data required for a good reservoir analysis does not always exist. When the data is available, the next step is to try to develop any correlations, with some degree of confidence, to aid in the

2 SPE 136183

completion design. This paper discusses the engineering methodology used to design a successful commercial completion in the Eagle Ford shale reservoir in south Texas. Understanding the Reservoir

The Challenge The operator was a participant in several horizontal Eagle Ford-shale wells in the DeWitt County area that had been completed using a Barnett shale-style water frac, and the results were disappointing. It was apparent that a different stimulation approach was needed to produce the commercial results necessary to continue development of their Eagle Ford Shale asset.

Whole-core data had been taken through the Eagle Ford during the drilling of a vertical offset well. A relatively complete core analysis had been performed, so the team had good rock and reservoir data from X-ray diffraction (XRD), geomechanical testing, embedment tests, total-organic-content (TOC) determination, desorption analysis, pressure/volume/temperature (PVT) analysis, and so on. On the well to be completed (known as the subject well), a vertical pilot hole was drilled and openhole electric logs were run across the zone of interest. The vertical section was plugged back, the lateral section drilled, and production casing cemented in place. The only data available in the lateral section was a mud log (gamma ray, rate of penetration, and total gas). To determine the number of stages to be pumped, the amount of proppant per stage, the perforation scheme, etc., it was necessary to obtain a good understanding of the reservoir.

Differences between the Barnett Shale and the Eagle Ford Shale

The current Eagle Ford development began as a shale-gas play in late 2008 in the middle/lower part of LaSalle County. The first fracture-stimulation designs were slick-water fracs, patterned after what had been done in the Barnett shale. The Barnett shale is predominantly a gas-bearing siltstone reservoir that is very brittle with a high YM of ~7×10E6 psi. The reservoir properties of the Eagle Ford shale are substantially different from those of the Barnett shale. The Eagle Ford shale produces both gas and high-gravity oil and is mainly a clay-rich limestone with very low quartz content. This tends to make it less brittle (more ductile) with a low YM of ~2×10E6 psi. Whole-core testing on the Eagle Ford shale indicates that because the rock is relatively soft (low YM), it is prone to proppant embedment (Fig. 1 and Fig. 2). Fig. 2 illustrates that the embedment in the Barnett shale at 5,000-psi closure stress will have about 0.20 grain diameters of embedment, while the embedment in the Eagle Ford shale at 10,000-psi closure stress can have an entire grain diameter of embedment. Embedment tests were conducted on Eagle Ford shale core to validate the simulation.

Fig. 1�—Brinell hardness number (from core tests) of various shale reservoirs in North America.

SPE 136183 3

Fig. 2�—Proppant-embedment simulation for various YM vs. closure stress (Cipolla et al. 2008). The Eagle Ford shale has a YM of around 2E6 psi while the Barnett shale has a YM around 7E6 psi. The Barnett shale is a very hard, brittle formation that contains many natural fractures, and has little, if any, stress

anisotropy (the difference between the maximum and minimum horizontal rock stress). This allows a complex network of fractures to be created rather easily by using a low-viscosity fluid. The high-YM rock tends to have little or no proppant embedment and mainly produces dry gas. Because there is very little difference between the maximum and minimum stresses in the rock, it is easy to open the natural fractures during the stimulation treatment. Under these reservoir conditions, a slick water frac with low volumes of very small-mesh proppant has been relatively effective. The reservoir conditions of the Barnett shale lend themselves to being stimulated with slick water, which can provide sustainable production results.

The Eagle Ford shale is significantly different. The Eagle Ford formation is a softer rock and could potentially have more stress anisotropy, which allows for more planar-type fractures. The low YM indicates that the rock is relatively soft and prone to proppant embedment; therefore, low concentrations of small-mesh proppant may not be as effective as in the Barnett formation. Higher concentrations of larger-mesh proppant placed using hybrid fluid systems provide sufficient conductivity to overcome embedment and multiphase flow. Cores of the Eagle Ford reservoir rock might not have a lot of visible natural fractures, but microfractures can be present (Fig. 3), which means a balance of net pressure may be required to maintain small fracture offsets along weak bedding planes and fissures during the stimulation treatment (Eqs. 1 and 2).

Fig. 3�—Normal and bedding-plane Fractures from Eagle Ford cores.

4 SPE 136183

A proper design of rate-and-fluid viscosity is required to create dominant fractures at the wellbore and establish the necessary fracture width while still remaining in the pay interval. Microseismic mapping can be used to verify how injection rate and fluid type will affect the fracture geometry.

.......................................................................................... (1) Eq. 1�—Viscosity (m) and pump rate (Q) influence on net pressure in a planar fracture (Rock Mechanics for Fracturing 2002).

.......................................................................................................................................(2) Eq. 2�—Nolte and Smith derived fissure opening pressure based on net pressure in the fracture (Cipolla, et al, 2008) Petrophysical Characteristics of the Eagle Ford Shale Reservoir The Eagle Ford formation lies beneath the Austin Chalk and extends laterally all the way across Texas from the southwest to the northeast part of the state (Fig. 4). Both formations were deposited during the cretaceous period and can be characterized as mudstones and chalks that have been enriched with organic-carbon material; thus, both might be hydrocarbon source rock (Fig. 5). The characteristics of the Eagle Ford formation change substantially across the SW-NE strike of the play. For example, thickness ranges from about 500 ft in the southwestern portion of the play, narrowing to around 45 ft in the Austin area, because of the San Marcos arch structural feature, and increasing to over 500 ft in the dark shales of the Eagle Ford formation that outcrop in Dallas County; depths range from ~2,500 to 13,000 ft TVD, pressure gradients from 0.4 to 0.8 psi/ft, TOC from 2 to 9%, and mineralogy varies significantly. To further complicate this technically challenging play, many of the aforementioned reservoir characteristics can also change significantly over relatively short distances; thus, understanding the local regional characteristics is critical to developing the asset commercially.

Corpus

SanAntonio

Austin

Houston

Oil

High Liquids

Gas

BeeDimmit

Frio Karnes

La SalleLive Oak

McMullen

Maverick Zavala Atascosa

DeWitt

Goliad

Wilson

GonzalesLavaca

Fayette

Colorado

Austin

Washington

Webb Duval

SubjectWell

Fig. 4�—Map of Eagle Ford play in Texas (for illustration purpose only).

SPE 136183 5

Fig. 5�—Update stratigraphic figure (source unknown). Petrophysical Analysis (Shale Log) Data from a variety of different sources, such as electrical logs, core data, and mud logs on an offset vertical well, were analyzed using a well-defined systematic approach (Rickman et al. 2008, Kundert and Mullen 2009, and Mullen et al. 2007) to develop an accurate petrophysical model to characterize the Eagle Ford shale in this area that could then be applied to other wells in the region to aid in developing a completion strategy based on the well�’s petrophysical characteristics.

The model used the electrical logs to first indentify the organic-rich shale using the spectral gamma ray, which was also used for clay typing. The other conventional logs were used for the mineral identification, which is critical for completion design, as the fluid effects on the formation must be considered (i.e., photoelectric absorption for pyrite volume, bulk density for limestone volume, neutron porosity for clay volume and porosity, and resistivity for water saturation). The sonic logs together with multiple other petrophysical relationships (where both mineralogy and shale type were considered) were used to calculate rock properties (Poisson�’s ratio and YM), and this was then calibrated to both dipole sonic data and core data. Image logs showed the Eagle Ford shale to be a finely laminated shale with bedding planes, all with dip angles of less than 10 degrees, indicative of a low-energy depositional environment, with the azimuth primarily SSE. Neither fractures, nor faults, nor any folds were discernable in this part of the Eagle Ford play.

Additionally, the shale-log model was used to calculate the TOC, kerogen content, free-gas volume, and rock-brittleness factor (Rickman et al. 2008), thereby grading the shale reservoir into potential completion intervals. The more ductile rock is indicated by a green color spectrum while the brittle rock is colored red in the brittleness factor track. The log-generated stress profile in conjunction with the rock properties was used for the completion design to simulate the sensitivity to frac rate and fluid type. The shale-brittleness factor, mineralogy, and the shale-classification analysis (Fig. 6) was used in the frac simulator to model the laminations and bedding planes. This was incorporated in the frac model to simulate any bounding layer properties that could affect fracture-height growth. The closure pressure, along with the embedment test data, was used to select the proppant type and mesh size. Finally, the mud log in the vertical pilot was reviewed to see the extent of any hydrocarbon shows.

6 SPE 136183

Fig. 6�—Offset well calibrated to core data.

Having evaluated the logs, the core data was then used to calibrate the model�— this is a critical component of the analysis

because without measured core data the log interpretation can be inaccurate. The entire petrophysical-analysis process loops around between logs and cores and back to the logs; thus, validating an accurate petrophysical model of this particular area. The mineralogy derived from the logs was calibrated to the X-ray analysis mineralogy, both of which were used to tailor the frac fluids used for the completion (Fig. 7 and Table 1).

Fig. 7�—(a) Average mineralogy and (b) clay-component percentages.

TABLE 1�—SUMMARY OF RESERVOIR PROPERTIES FROM CORE DATA ANALYSIS

min max

TOC,% 2 6

Porosity, % 8 18

Water saturation, % 7 31

Permeability, nanoDarcies 1 800

YM, psi 1.00E+06 2.00E+06

Poisson�’s ratio 0.25 0.27

SPE 136183 7

Core geochemistry identified the kerogen type and maturity to be in the mature-oil window, with some core sitting in the condensate wet-gas window with fair hydrocarbon potential. Thus, the completion design would need to focus on an oil-producing shale rather than dry gas, unlike the other early Eagle Ford wells. Fluid-sensitivity tests and proppant-embedment test were conducted on the core, and the findings were used in the stimulation design.

Thin sections of the core characterized the Eagle Ford shale visually as a planar laminated shale with numerous bedding-plane fractures, planktonic foraminifera, and organic- rich matrix. Occasional-bedding normal fractures and pressure-release fractures were also observed (Fig. 3). Whole core indicated that the overburden stress was normal. This information was used in the completion strategy because the potential existed for the swelling clays identified by XRD and the log to plug these fractures. No fractures were seen by the openhole image log, but this is not uncommon for shale reservoirs. The identification of natural fractures and natural-fracture orientation in the core, together with understanding the reservoir characteristics to tailor the stimulation design to the reservoir, are fundamental to developing commercially successful shale plays.

Having calibrated the offset well to its core data, the petrophysical model was then applied to the TVD vertical pilot hole (Fig. 8) of the subject well that was drilled, logged, and pluged back before drilling horizontally through the Eagle Ford shale. Cased-hole pulse-neutron log data in the lateral was preferable for a horizontal petrophysical analysis using CHI modeling techniques (Reed et al. 2007) to generate a synthetic openhole log which could then be used for petrophysical analysis. Because the cased-hole data was not acquired, it was decided to transpose the vertical petrophysical analysis to the hoizontal wellbore. The vertical petrophysical analysis data from the subject well was saved as an Excel VLOOKUP table and tied by TVD depth. Then the lateral-well mud-log TVD and MD data were taken and transposed via the aforementioned VLOOKUP table to the equivalent data from the offset vertical well to create a pseudo GR and pseudo VLOOKUP shale-log analysis of the 3,400 ft lateral well, which had been drilled within a 22-ft TVD window of the Eagle Ford shale.

Fig. 8�—(a) Shale log on vertical pilot hole TVD of offset well and (b) shale log on vertical pilot hole TVD of subject well. Using this transposition technique to aquire additional information for the completion employs several assumptions. The

first assumption is that the reservoir has no lateral variations and is homogenous. The second assumption is that the directional drilling data is accurate, and lastly that the dip of the reservoir can be correlated back to a level datum plane. Fig. 9 shows the pseudo VLOOKUP GR (black solid line) correlates well with the actual measured mud log GR (red line with green shading) when displayed in TVD, illustrating that the VLOOKUP technique in the deviated (>45°) wellbore matched the vertical pilot wellbore at the top of the Eagle Ford formation. Fig. 10 shows the entire transposed VLOOKUP pseudo shale-log analysis in the lateral wellbore. This provided valuable information (brittleness factor, stress, Poisson�’s Ratio, YM), which was used in conjunction with the mud-log hydrocarbon shows to determine the number of frac stages and where to place the perforation clusters in each stage. The VLOOKUP pseudo shale-log was used to identify the organic-rich shale, select the completion intervals by looking for rock with higher brittleness (red color), and place perforation clusters in completion intervals with similar stresses. Fig. 11 illustrates the actual lateral wellbore with reservoir dip and directional drilling data included.

It can be seen that in some places (Stages 2, 4, 9, and 11), the VLOOKUP GR matches the GR mud log very well and thus there was confidence in the analysis, while in other parts of the well there was not such a clear correlation. In Stages 2, 4, and 9, the brittleness factor shows as ductile rock (green). Note that in the more brittle rock (Stages 3, 5, 6, 7, 8, and 11), it was easier to execute the stimulation plan�—normal treating pressures were observed. In the more ductile rock (Stages 2, 4, and 9; displayed in green), higher treating pressures were observed and the stages were harder to complete.

8 SPE 136183

Fig. 9�—TVD�—measured GR vs. pseudo VLOOKUP GR.

GR (red curve)

VLOOUP GR (black curve)

More Brittle (red) More Ductile (green)

ROP Tgas

Stg 1Stg 2Stg 3Stg 4Stg 5Stg 6Stg 7Stg 8Stg 9Stg 10Stg 11Stg 12

Fig. 10�—Subject-well vertical-pilot shale log data transposed to lateral wellbore to generate pseudo shale-log brittleness using VLOOKUP.

SPE 136183 9

Fig. 11�—Subject-well visualization illustrating reservoir dip and directional-drilling data in the lateral wellbore. Completion Design The formation evaluation was a vital part of the completion design. Very little frac/production data was available in the area, so it was impossible to emulate other completions. Some of the data in the shale-log analysis was used qualitatively (mineralology, shale classification, TOC, brittleness index, etc.), while other parameters were used quantitatively (stress, Poisson�’s Ratio, YM, BHP, etc.) to design the treatment. Formation evaluation for a shale reservoir requires a different mindset because it is not a conventional porosity/permeability reservoir. The same thing applies to the completion design. The terms water frac, conventional frac, and a hybrid frac are generic descriptions for the types of treatments used to stimulate shale reservoirs. The issue that arises with using this nomenclature is that the definitions are not standardized in the industry. The systems are so variable that it�’s difficult to accurately define each; in this paper�’s context, they are described as the following:

Water Frac: A frac fluid that uses water with a very low viscosity to fracture the reservoir and transport proppant for the majority of the treatment. Friction-reducing chemicals or gelling agents are used only to reduce pipe friction, not for proppant transport. They are typically pumped at high injection rates and rely on fluid velocity, proppant banking, etc. to transport/distribute proppant in the fracture. They typically consist of alternate stage volumes of proppant followed by fluid �“sweeps�” throughout the treatment.

Conventional Frac: A frac fluid that uses a higher-viscosity fluid (from foams to crosslinked fluids) to create fracture geometry and transport proppant for the majority of the treatment. Injection rates can vary greatly depending on the reservoir properties and wellbore configuration. A typical conventional frac will begin with �“pad�” fluid, which is followed by increasing concentrations of proppant-laden fluid.

Hybrid Frac: Anything in-between a water frac and a conventional frac. Typically, a hybrid frac is a combination of the two. They tend to begin with a low-viscosity fluid (at a high rate), alternating proppant volumes with fluid �“sweeps,�” and then conclude (sometimes at a lower injection rate) with a continuously increasing proppant concentration in a high-viscosity or crosslinked fluid. An unorthodox conventional frac could also be considered a hybrid frac�—this is a frac fluid that is designed outside the standard bounds currently subscribed to by the industry, as was the case in this subject well. Frac Design The wellbore configuration was 4 1/2-in. casing, 15.1 lbm/ft, with a tie-back in the top of a liner at ~11,200 ft (TVD of the formation ~13,250 ft). The MD of the well was 16,500 ft with a lateral of ~3,800 ft. The wellbore was drilled in a NW-SE direction, so transverse fractures were anticipated (Soliman et al. 2004). A fracture simulator was populated with rock properties and reservoir data obtained from the detailed petrophysical analysis in the vertical pilot well. The �“hybrid frac�” pump schedule (Fig. 12) was input and sensitivities were run for pump rate, viscosity, and number of potential fractures generated in the horizontal section (based on number of perf clusters). Current frac simulators have limitations when applied in complex reservoirs (complex fractures, nonplanar fractures, bounding layer effects, etc.), but they can still be a useful tool when looking at design sensitivities. This exercise was also used to help determine how much lateral interval might be effectively stimulated with each frac stage. The initial stimulation design that was considered for this project was a hybrid, with a twist. The crosslinked portion of the frac design would be transporting proppant for approximately 60 minutes. The crosslinked fluid was designed to maintain a high viscosity for ~15 minutes (as per Model 50 fluid testing (Fig. 13)), or about 25% of the time proppant-laden fluid would be pumped. The intent was to create fracture geometry with viscosity, rather than rate, to effectively stimulate each frac interval and keep the fractures in the pay zone. The fluid would then be allowed to degrade, both with temperature and high-temperature breakers, to a low viscosity to maximize the potential stimulated reservoir volume (SRV).

10 SPE 136183

Fig. 12�—Hybrid frac pump schedule.

Fig. 13�—35-lbm borate-crosslinked fluid with 0.5 gal/1,000 gal breaker at 300°F.

The fracture simulations (Fig. 14) were based on a single stage with four perforation clusters in the stage. The model was set to equally distribute the fluid and proppant between the perf clusters; thus, creating four fractures. The goal was to achieve ~1 lbm/ft2 proppant concentration to handle the high liquid production that was expected. The results were as follows:

�• 1 Fracture: If only one perf cluster took fluid, then a single fracture, 450-ft long, was generated, which grew upward into the Austin Chalk formation above the Eagle Ford, and the conductivity was very low (<0.5 lbm/ft2).

�• 2 Fractures: If two perf clusters took fluid, then two fractures were generated and the results were a little better. The fractures also tended to grow out of zone. The effective propped length of each frac was only 150 ft (net of 300 ft), and the conductivity was again low (~0.5 lbm/ft2).

�• 3 Fractures: If three fractures were generated, the fracs stayed more contained in the Eagle Ford formation and achieved an effective propped frac length of ~200 ft (net of 600 ft). The conductivity was just over 1 lbm/ft2 and evenly distributed across the formation.

�• 4 Fractures: If four fractures were generated, then each frac generated only about 125 ft of effective propped frac length (net of 500 ft) with good conductivity (>1 lbm/ft2).

SPE 136183 11

Fig. 14�—Fracture-sensitivity simulations for a single frac stage. Perforation Design and Placement To create a transverse fracture with minimal tortuosity complications, the team relied on the work of El-Rabaa (1998). His work showed that (in the lab), as long as the perforated interval is less than two times the diameter of the wellbore, a single transverse fracture will be created. However, field experience has shown that going up to four times the diameter can work just as well. This is necessary to place 3 to 4 lbm/gal 20/40-mesh proppant in the fracture and avoid wellbore or near-wellbore screenouts.

Because it�’s impossible to determine how many perforations will actually breakdown and take fluid/proppant, it was decided to incorporate other methods that could enhance the chance of success.

The first method was based on limited-entry perforating. The concept of �“limited-entry perforating�” has been used for decades, and there are mixed opinions as to its success. It was decided that it could only enhance the treatment if it worked (Eq. 4). The injection rate of 45 bbl/min was selected partially because of wellhead-pressure limitations (10,000 psi wellhead). Subsequent treatments have incorporated a 15,000-psi wellhead to increase the safe surface-working pressure (typically 12,000 psi maximum treating pressure). A total of thirty-six 0.34-in. holes, 60° phased with a 45 bbl/min liquid rate, would generate ~500 psi perf friction, assuming that some of the holes would not take fluid (roughly 30%). Therefore, an injection rate of 30 bbl/min would still create ~500-psi perf friction if only 25 holes were open. The plan was to shoot four clusters, 2-ft each in length, with 6-spf (shots per foot or twelve holes), 0.34-in. diameter perforations. It was thought that it was unlikely that more than three clusters would actually initiate a fracture (Cheng, 2009).

An approximation of the perforation friction per perforation at various rates was calculated using the perforation equation:

.......................................................................................................... (3) The second method to enhance the effectiveness of the stimulation treatments was to use the petrophysical analysis from the vertical section, which was correlated to the horizontal section. The plan was to place the perforation clusters for each stage in �“like�” rock instead of just equally spacing the perforation clusters (Figs. 15 and 16). The intent was to place the perforation clusters where the reservoir rock near the wellbore was the same and should ideally have very similar breakdown pressures. The perforation scheme for each stage targeted the more brittle rock (red) and tried to avoid the ductile areas (green), if possible. This might seem like a very subtle detail and some might not think it would make any difference, but once

12 SPE 136183

the fracture initiates in a perf interval, it can be difficult to breakdown the other clusters. The objective is to make every effort to get all of the perf clusters to accept fluid and proppant to maximize the SRV. This is where limited entry (perforation friction) and fluid viscosity (crosslinked slugs) can help.

Fig. 15�—Perforate in the �“like�” rock.

Fig. 16�—Good VLOOKUP GR correlation for Stages 9 and 11. The solid red line on the bottom track (Track 1) is the openhole gamma ray and the dotted black line is the gamma ray transposed from the vertical pilot wellbore. Formation Compatibility�–Clay Control and Surfactants Core XRD analysis showed an average total clay content of ~20% (ranging from 5 to 45%) with about half (10%) of that being mixed-layer clay. The cores also indicated both normal and bedding-plane fractures (Fig. 3), which, if plugged with swollen clays, could greatly impede hydrocarbon production and ultimate recovery. It was decided to use 6% NaCl (~1.0 molar solution) in all of the completion fluids, based on work done by Gdanski and Schuchart (1997) (Fig. 17). Understanding that this could be more clay control than necessary, it was decided to err on the conservative side. Since liquid hydrocarbon production was expected, the team decided that fluid compatibility testing was necessary. It was determined that a non-emulsifying surfactant should be used in all of the treatment fluid based on samples obtained from an offset well.

Fig. 17�—Shows 7% KCl (or ~1-molar salt solution) is required to properly protect clay from swelling.

SPE 136183 13

Proppant Selection Proppant-embedment testing on Eagle Ford core samples shows embedment in the range of 300 to 700 microns in depth (0.30 to 0.75 mm) at 10,000 psi with high-strength proppant. Comparing this to the width of a grain of 40/70-mesh proppant (~300 microns wide) and the width of a grain of 20/40-mesh proppant (~700 microns wide), the indications are that low concentrations of small proppant can actually �“disappear�” into the formation face and propped-fracture conductivity can quickly approach zero. Based on this data, the decision was made to use larger-mesh proppants (30/50- and 20/40-mesh). This test data agrees with the simulation work (Fig. 2) done by Cipolla et al. (2008).

Because this well was expected to produce liquid hydrocarbon, higher concentrations of larger-mesh proppant were going to be needed to achieve the required fracture conductivity. Proppant embedment, proppant crushing at stresses, and formation-fines migration are major components that can cause the loss of fracture conductivity over time. The zone of interest was at 13,500 ft (TVD) and the bottomhole temperature (BHT) was calculated to be 327°F. Proppant diagenesis can also degrade the conductivity of a proppant pack over time and has a larger effect on man-made materials at elevated temperatures (Weaver et al. 2005). High-strength proppant was discussed because the formation closure stress was expected to be between 11,000 and 12,000 psi, based on data from the early offset well-completions history. The economics were reviewed, and based on the higher costs associated with premium proppants, it was decided to pump natural sand at the beginning of the job followed by premium resin-coated proppant (RCP). The split was basically 15% 30/50-mesh premium white sand, 35% 20/40-mesh premium white sand, and 50% 20/40-mesh premium RCP (white sand substrate). Based on the embedment testing, it was felt that the load distribution on the proppant grains would be more evenly distributed in the low-modulus formation than in a high-modulus rock, which would exert more of a point source stress (Fig. 18). The team knew the proppant would experience proppant crushing (especially at the grain-to-grain contact) and fines would be generated, so a surface-modifying agent (SMA) was recommended for all of the uncoated premium white sand pumped to minimize these conductivity-degradation effects (Nguyen et al. 1998). It was decided the SMA could be implemented on the next well, pending the production results on the subject well.

Fig. 18�—Illustration of proppant-embedment stress distribution. Job Execution and Onsite Redesign The first stage was perforated with a tubing-conveyed perforating gun. Some miscommunications resulted in a 10-ft long perf interval (6 spf) at the toe of the horizontal. Two 3-ft long clusters were added to that for a total of 96 holes across 16 feet, which was not the ideal configuration. The toe of a horizontal is routinely the hardest to frac (there are many theories as to why this occurs), and this well proved to fall into that category. The design was 45 bbl/min for the first frac stage, but the initial injection rate was less than 20 bbl/min at 9,800 psi (10,000 psi max WHP). Almost three hours were spent attempting to obtain sufficient fracturing injection rate without success (almost screened-out on 0.25 lbm/gal 100-mesh prop slug) (Fig. 19).

14 SPE 136183

Fig. 19�—Stage 1; first attempt could not establish sufficient injection rate.

It was decided to reperforate Stage 1. The injection rate was a little better (25 bbl/min at 9,900 psi). Various things were attempted to try to get a good breakdown (viscous gel slugs, 100-mesh proppant slugs, etc.). But the near-wellbore friction was too great to overcome. A step-down test (Hyden and Stegent 1996) indicated over 1,500 psi near-wellbore friction was present. The second attempt of Stage 1 had similar results (Fig. 20). The best rate that could be achieved was 25 bbl/min, and each time a 1,000-gal stage of 100-mesh proppant was pumped, the wellhead pressure increased. Crosslinked slugs displaced with linear gel seemed to have a much better impact at reducing the wellhead pressure than the proppant slugs. Because the maximum allowable wellhead-pressure limit was almost reached, it was decided to abandon this stage and move on to Stage 2.

Fig. 20�—Stage 1; second attempt still could not establish enough injection rate (25 bbl/min max).

A bridge plug was set and alterations to the original perforation scheme for Stage 2 were discussed. It was decided to adhere to the rules presented by El Rabaa (1998) and keep the perforation intervals to 1 foot in length. To ensure there were enough holes in the casing, a perforation scheme redesign of eight 1-ft clusters, 6 spf, with 60°phase was implemented. This would provide 48 holes for each frac stage.

SPE 136183 15

Stage 2 set the stage for the remainder of the frac-treatment stages. After all of the difficulty and lessons learned on the first stage, it was decided to discontinue using the 100-mesh and instead use the 30/50-mesh proppant for the initial slugs during the pad (Figs. 21 and Fig. 22). The bridge-plug ball was dropped in the first 20 bbl of acid (total of 120 bbl acid per stage) and displaced with linear gel. Once the ball landed and the injection rate was up to 30 bbl/min, crosslinked proppant slugs were pumped. This initial breakdown design has since been modified (Stegent et al. 2010). Again, it seemed like each time the proppant slug reached the formation, the pressure increased (initially) and then slowly started to decline. A stable injection rate of 30 bbl/min on the pad fluid was obtained with crosslinked gel (~35-lbm polymer load), so it was decided to start pumping the 30/50-mesh proppant. The concentrations were slowly increased from 0.25 lbm/gal to 1 lbm/gal, one wellbore volume at a time. When the 20/40-mesh proppant was started, the concentration was decreased back to 0.25 lbm/gal and slowly ramped up to 1 lbm/gal to make sure there were no wellbore-width issues and the proppant could be placed. This redesign on-the-fly process was continued until a 2.5 lbm/gal concentration was achieved. A total of 300,000 lbm of proppant was placed in 7,900 bbl of fluid.

Fig. 21�—Stage 2.

Fig. 22�—Stage 2 chemical plot. The rest of the stages (Stages 3 thru 11) were completed in a similar fashion. They all had the same perforation scheme,

and each stage started with a basic redesigned pump schedule (Fig. 23), which was adjusted on-the-fly based on how the interval was treating. The newly designed pump schedule consisted of more-linear and crosslinked gel than the original design because it was thought it would be easier to place the proppant in the formation than when low-viscosity fluids were pumped

16 SPE 136183

ahead. The redesigned schedule might first appear to look like a conventional frac design, but the fluid viscosity of the crosslinked fluids adheared to the original Model 50 fluid testing (Fig. 13) discussed earlier. It was also decided to maintain a constant injection rate throughout th majority of the job instead of using a variable rate (slowing down when switching to crosslinked fluid) that was originally designed.

The proppant concentration was only reduced by 0.5 lbm/gal (1.0 lbm/gal reduction was originally designed) when switching from 30/50-mesh to 20/40-mesh proppant once treatment experience (tribal knowledge) was gained. The stages that were easier to breakdown and establish an injection into early in the job used less fluid to place the design volume of proppant (Fig. 24) than the intervals that were harder to establish injection (Fig. 25). Some stages were difficult to pump (Stages 2, 4, and 9) while others (Stages 6, 8, and 11) were easier to place. Both Stages 9 (more ductile) and 11(more brittle) had a good correlation to the transposed horizontal shale log (Fig. 16) and were therefore used as examples to show the benefit of knowing the brittleness index of the rock in the lateral section of the wellbore and how it can be related to the job treatment. Figs. 24 and 25 illustrate how Stage 11 established injection without difficulty and the design rate was easily achieved (total pump time ~2 hours), whereas Stage 9 experienced difficulty getting injection established and had a difficult time getting to the designed pump rate (total pump time ~3 hours). Both job plots have the same time interval and axis scales for comparison purposes.

V o l um e (G a l s ) F l ui d Ty p e

R a te (B P M )

P ro p pa n t T y p e

P ro p Co nc

(l b / ga l )

P ro p V ol u m e p e r S ta ge

(lb s)

C u m ul a tiv e P r op

V ol u m e (lb s)1 0 0 0 2 0 # L in e a r 11 0 0 0 1 5 % H C L A c id 1 0

Dro p B a ll fo r B P 1 04 0 0 0 1 5 % H C L A c id 1 04 0 0 0 2 0 # Xlin k G e l 1 02 0 0 0 2 0 # L in e a r 1 0 `

L a n d B a ll o n B P 1 05 0 0 0 2 0 # L in e a r 3 5

1 0 00 0 3 5 # Xl in k 3 52 0 00 0 2 0 # L in e a r 4 02 0 0 0 3 5 # Xl in k 4 0 3 0 / 5 0 S a n d 0 .2 5 5 0 0 5 0 0

1 5 00 0 2 0 # L in e a r 4 5 0 5 0 02 0 0 0 3 5 # Xl in k 4 5 3 0 / 5 0 S a n d 0 . 3 6 0 0 1 1 00

1 0 00 0 2 0 # L in e a r 5 0 0 1 1 003 0 0 0 3 5 # Xl in k 5 0 3 0 / 5 0 S a n d 0 . 5 1 5 0 0 2 6 00

1 0 00 0 2 0 # L in e a r 5 0 0 2 6 003 0 0 0 3 5 # Xl in k 5 0 3 0 / 5 0 S a n d 0 .7 5 2 2 5 0 4 8 50

2 0 00 0 3 5 # Xl in k 5 0 0 4 8 501 0 00 0 3 5 # Xl in k 5 0 3 0 / 5 0 S a n d 0 .2 5 2 5 0 0 7 3 501 0 00 0 3 5 # Xl in k 5 0 3 0 / 5 0 S a n d 0 . 5 5 0 0 0 1 2 3 5 01 0 00 0 3 5 # Xl in k 5 0 3 0 / 5 0 S a n d 0 .7 5 7 5 0 0 1 9 8 5 01 0 00 0 3 5 # Xl in k 5 0 3 0 / 5 0 S a n d 1 1 0 0 0 0 2 9 8 5 01 0 00 0 3 5 # Xl in k 5 0 3 0 / 5 0 S a n d 1 .2 5 1 2 5 0 0 4 2 3 5 01 0 00 0 3 5 # Xl in k 5 0 3 0 / 5 0 S a n d 1 . 5 1 5 0 0 0 5 7 3 5 01 0 00 0 3 5 # Xl in k 5 0 2 0 / 4 0 S a n d 1 1 0 0 0 0 6 7 3 5 01 0 00 0 3 5 # Xl in k 5 0 2 0 / 4 0 S a n d 1 .2 5 1 2 5 0 0 7 9 8 5 01 0 00 0 3 5 # Xl in k 5 0 2 0 / 4 0 S a n d 1 . 5 1 5 0 0 0 9 4 8 5 01 5 00 0 3 5 # Xl in k 5 0 2 0 / 4 0 S a n d 1 .7 5 2 6 2 5 0 1 2 1 10 01 8 00 0 3 5 # Xl in k 5 0 2 0 / 4 0 S a n d 2 3 6 0 0 0 1 5 7 10 01 0 00 0 3 5 # Xl in k 5 0 2 0 / 4 0 R CP 2 2 0 0 0 0 1 7 7 10 01 3 00 0 3 5 # Xl in k 5 0 2 0 / 4 0 R CP 2 .2 5 2 9 2 5 0 2 0 6 35 01 5 00 0 3 5 # Xl in k 5 0 2 0 / 4 0 R CP 2 . 5 3 7 5 0 0 2 4 3 85 01 5 00 0 3 5 # Xl in k 5 0 2 0 / 4 0 R CP 2 .7 5 4 1 2 5 0 2 8 5 10 05 0 0 0 3 5 # Xl in k 5 0 2 0 / 4 0 R CP 3 1 5 0 0 0 3 0 0 10 0

1 0 00 0 1 0 # L in e a r F lu s h 5 0 3 0 3 0 0 0

Fig. 23�–Basic revised pump schedule.

SPE 136183 17

Fig. 24�—Stage 11.

Fig. 25�—Stage 9. Production Results Coiled tubing was brought in the day after the last frac stage was complete to drill out the bridge plugs and clean out the wellbore. The initial positive choke size was 8/64 in., which was increased to 9/64 in. after 24 hours. The flowback choke size was slowly increased by 1/64 in. until a 14/64-in. choke was installed about 2 weeks into the flowback. Sixty days later, the well was flowing on a 20/64-in. choke and had produced over 50,000 bbl of oil and 150 MMscf gas (~80,000 BOE). About 10% of the frac load water was recovered after 180 days, and the well currently only produces ~10 BWPD. The production on this engineered hybrid completion far exceeded the production of an offset slick-water completion (Fig. 26).

18 SPE 136183

Fig. 26�—Daily production: hybrid (subject well) vs. offset slickwater (SW) frac.

Production from the subject well was also compared to production from seven offset wells (Fig. 27). The offset wells were treated with either a slick-water frac or a hybrid-type treatment. The production data is from the public domain and different service companies performed the stimulation treatments; therefore, a broad approach was taken for the comparison. The gas was converted to a barrel of oil equivalent (BOE) using 170 bbl of oil per 1 MMCF. Because very little reservoir or treatment data is available through public sources, no attempt was made to further normalize the data. The wells included in the comparison all had at least 1 month of reported production, and the peak month was used in the bar graph (Fig. 28). Fig. 28 shows that the hybrid-type treatments seem to outperform the water frac treatments. The authors believe this is because of the higher fracture conductivity achieved with a hybrid treatment, allowing a more unrestricted flow path for the liquids. These are all early-time results and need to be investigated further after more production data becomes available (Khan and Callard 2010).

Fig. 27�—Location of subject well and offset Eagle Ford completions in Gonzales, DeWitt, and Karnes Counties, TX.

SPE 136183 19

Fig. 28�—Comparison of monthly peak production. Conclusions

�• Hybrid stimulation treatments out-performed slick-water frac treatments in the Eagle Ford shale in the area of high liquid production.

�• Formation evaluation combined with core data correlation to develop a good petrophysical model is critical in shale completions to provide the completion engineer with guidance toward the proper stimulation design.

�• Higher production appears to correlate to higher conductive fractures that use higher concentrations of larger-mesh proppant, especially in high liquid-production reservoirs.

�• Transposition of vertical petrophysical data to the lateral wellbore can be useful for the perforation strategy. �• Fracture conductivity is essential to sustain production when multiphase hydrocarbon flow is expected. �• Injection rates between 35 and 50 bbl/min were sufficient to effectively stimulate the Eagle Ford shale with this

hybrid fluid and engineered design.. �• 3 lbm/gal (or higher) 20/40-mesh proppant can be placed in the Eagle Ford shale with the proper perforation scheme

and a proper fluid design. �• Flexibility regarding onsite design change can be critical for successful job placement.

Acknowledgements The authors wish to thank the Houston Halliburton account team for all their support, the Halliburton frac crew and engineering staff in Caldwell, TX for handling all of the onsite design changes, Collins DeVilbiss for his operational guidance and input, and Ryan Bump for compiling data for the paper. Nomenclature

v = Poisons�’ ratio = Stress, M/LT2 c = Closure stress, M/LT2

hf = Fracture height, L Po= Fissure opening pressure, M/LT2

h = Difference between minimum and maximum horizontal stress, M/LT2 Propportional to

H = Fracture height, L = Porosity

KIc-app = Fracture-tip effect or fracture toughness p = Fracture pressure, psi Pnet = Net pressure, psi Q = Frac rate, bbl/min

20 SPE 136183

f = Fluid density, m/L3

xf = Hydraulic-fracture wing or half-length, L = Fluid viscosity, M/LT

E = Young's modulus E' = Plane strain modulus (= E/1- 2)

References Chen, Y. 2009. Boundary Element Analysis of the Stress Distribution around Multiple Fractures: Implications for the Spacing of Perforation

Clusters of Hydraulically Fractured Horizontal Wells. Paper SPE 125769 presented at the SPE Eastern Regional Meeting, Charleston, West Virginia, 23�–25 September. DOI: 10.2118/125769-MS.

Cipolla, C.L, Warpinski, N.R., Mayerhofer, M.J., Lolon, E.P. and Vincent, M.C. 2008. The Relationship Between Fracture Complexity, Reservoir Properties, and Fracture Treatment Design. Paper SPE 115769 presented at the SPE Annual Technical Conference and Exhibition, Denver, Colorado, 21�–24 September. DOI: 10.2118/115769-MS.

El Rabaa, W. 1998. Experimental Study of Hydraulic Fracture Geometry Initiated From Horizontal Wells. Paper SPE 19720 presented at the SPE Annual Technical Conference and Exhibition, San Antonio, Texas, 9�–11 October. DOI: 10.2118/19720-MS.

Gdanski, R. and Shuchart, C. 1997. Advanced Sandstone Acidizing Designs Using Improved Radial Models. Paper SPE 38597 presented at the SPE Annual Technical Conference and Exhibition, San Antonio, Texas, 5�–8 October. DOI: 10.2118/38597-MS.

Hyden, R.E. and Stegent, N.A. 1996. Pump-in/Shutdown Tests Key to Finding Near-Wellbore Restrictions. Paper SPE 35194 presented at the SPE Permian Basin Oil and Gas Recovery Conference, Midland, Texas, 27�–29 March. DOI: 10.2118/35194-MS.

Khan, M.U. and Callard, J.G. 2010. Reservoir Management in Unconventional Reservoirs. Paper SPE 130146 presented at the SPE Hydrocarbon Economics and Evaluation Symposium, Dallas, Texas, 8�–9 March. DOI: 10.2118/130146-MS.

Kundert, D. and Mullen, M. Proper Evaluation of Shale Gas Reservoirs Leads to a More Effective Hydraulic-Fracture Stimulation. 2009. Paper SPE 123586 presented at the SPE Rocky Mountain Petroleum Technology Conference, Denver, Colorado, 14�–16 April. DOI: 10.2118/123586-MS.

Mullen, M., Roundtree, R., Barree, B. and Turk, G. 2007. A Composite Determination of Mechanical Rock Properties for Stimulation Design. Paper SPE 108139 presented at the SPE Rocky Mountain Oil and Gas Technology Symposium, Denver, Colorado, 16�–18 April. DOI: 10.2118/108139-MS.

Nguyen, P.D., Weaver, J.D, Dewprashad, B.T. 1998. Surface-Modification System for Fracture-Conductivity Enhancement. Paper SPE 48897 presented at the SPE International Conference and Exhibition in Beijing, China, 2�–6 November. DOI: 10.2118/48897-MS.

Reed, S., Torne, J.P., Pacheco, E., Lara, A., Palacios, C., Riano, J.M., Corral, A., Casares, M., Simon, J.D. and Arteaga, R. 2007. Case History: Application of CHI Modeling using Pulsed Neutron to Create Pseudo Openhole Logs in Highly Deviated Wells Using Special Techniques for Logging and Perforating in Veracruz, Mexico. Paper SPE 107527 presented at the SPE Latin America Petroleum Conference, Buenos Aires, Argentina, 15�–18 April. DOI: 10.2118/107527-MS.

Rickman, R., Mullen, M., Petre, E., Grieser, B. and Kundert, D. 2008. A Practical Use of Shale Petrophysics for Stimulation Design Optimization: All Shale Plays are not Clones of the Barnett Shale. Paper SPE 115258 presented at the SPE Annual Technical Conference and Exhibition, Denver, Colorado, 21�–24 September. DOI: 10.2118/115258-MS.

Rock Mechanics for Fracturing. Frac Tips. (January 2002) 1�–2. Soliman, M.Y., East, L. and Adams, D. 2004. GeoMechanics Aspects of Multiple Fracturing of Horizontal and Vertical Wells. Paper SPE

86992 presented at the SPE International Thermal Operations and Heavy Oil Symposium and Western Regional Meeting, Bakersfield, California, 16�–18 March. DOI: 10.2118/86992-MS.

Stegent, N., Leotaud, L., Prospere, W. and C Veillette, C. 2010. Cement Technology Improves Fracture Initiation and Leads to Successful Treatments in the Eagle Ford Shale. Paper SPE 137441 presented at the SPE Tight Gas Completions Conference, San Antonio, Texas, 2�–3 November.

Weaver, J.D, Nguyen, P.D., Parker, M.A. and van Batenburg, D. 2005. Sustaining Fracture Conductivity. Paper SPE 94666 presented at the SPE European Formation Damage Conference, Scheveningen, The Netherlands, 25�–27 May. DOI: 10.2118/94666-MS.