hydraulic fracture optimization in unconventional reservoir

SPE 151128 Hydraulic Fracture Optimization in Unconventional Reservoirs Pedro Saldungaray, SPE, Terry T. Palisch, SPE, CARBO Ceramics Inc. Copyright 2012, Society of Petroleum Engineers This paper was prepared for presentation at the SPE Middle East Unconventional Gas Conference and Exhibition held in Abu Dhabi, UAE, 23–25 January 2012. 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 Hydraulic fracturing has become a critical component in the successful development of unconventional reservoirs. From tight gas, to oil and gas-producing shales and coal bed methane, resource plays rely on hydraulic fracturing for commercial viability. A primary goal in unconventional reservoirs is to contact as much rock as possible with a fracture or a fracture network of appropriate conductivity. This objective is typically accomplished by drilling horizontal wells and placing multiple transverse fracs along the lateral. Reservoir contact is optimized by defining the lateral length, the number of stages to be placed in the lateral, the fracture isolation technique and job size. Fracture conductivity is determined by the proppant type and size, fracturing fluid system as well as the placement technique. While most parameters are considered in great detail in the completion design, the fracture geometry and conductivity receives lesser attention. Some mistakenly anticipate that in extremely low permeability formations, hydraulic fractures act as “infinitely conductive” features. However, many factors that affect the realistic conductivity of the fracture are poorly understood or overlooked. This often leads to a less than optimal outcome with wells producing below the reservoir potential. This paper presents an approach to assess the realistic fracture conductivity at in-situ conditions and the economic implications on proppant selection. The effects of transverse fractures, low areal proppant concentration and flow dynamics, are considered among other variables. The theory behind this concept is presented and supported with case studies where it has been applied in the field to various unconventional reservoirs. Introduction Unconventional reservoir fracturing is unique in several aspects when compared to fracturing conventional wells. Very low to extremely low permeability, horizontal well geometries, multiple transverse fracs placed along a horizontal drain, and complex frac geometry - particularly in shales - all add to the complexity of designing and implementing fracture treatments. For the remainder of the paper we are assuming that horizontal wells with multiple fracs are utilized in unconventional reservoir developments. In order to optimize the stimulation treatment, the design process must attend to multiple parameters which can be grouped into four broad categories: • Wellbore placement and lateral length • Completion hardware and isolation techniques • Fracture spacing or number of fracs • Fracture geometry and conductivity Wellbore Placement and Lateral Length. These parameters are driven by geology, in situ stress regime, reserves to be developed per well, production rates to be handled by each individual well, future well intervention requirements, surface logistics and environmental impact. The trend within most unconventional plays through the years has seen an increase in the lateral length to maximize the reservoir contacted and reserves developed by each well (Figure 1). In most cases the main restriction to lateral length is the capability of both current and future intervention in the wellbore. This may include limitations on frac isolation equipment and perforating, as well as coiled tubing reach concerns. This trend of increasing the lateral length has favorably impacted the economics of field developments and leaseholds, and reduced the environmental impact of development. Lateral lengths ranging from 1,000 to in excess of 10,000 ft are common today [Rankin 2010] 2 SPE 151128 Completion Hardware and Isolation Techniques. The industry has developed a wide variety of completion hardware and isolation techniques for Horizontal Multi- Fractured wells (HMF). From barefoot openhole wells to uncemented or cemented liners, ball-activated sliding sleeves to pump-down plugs and perf guns, as well as hybrid systems, each technique strives to maximize operational efficiency by placing the maximum amount of stages in the minimum possible time. Current multistage sleeve systems are capable of placing dozens of stages in a continuous pumping operation, with the maximum limits being continually pushed. Plug and perf techniques are only limited by the ability to pump the plugs and guns down the lateral. In the Bakken, for example, operators are now routinely placing as many as 40 stages per lateral using combinations of sliding sleeve and plug and perf methodology [Rankin, 2010]. In fact, it is rumored that some are contemplating as many as 50 stages in the future. Fracture Spacing and Number of Fracs. This subject is largely dependent on the rock fabric and permeability of the formation. As spacing is reduced adjacent fracs start to interfere with each other affecting production while costs continue to increase due to the larger number of treatments. An economic evaluation dictates the optimal spacing where the benefit of adding fracs is balanced with the cost of the increased number of fracture stages [Rankin 2010, Norris, 1998]. Two parameters can affect this interference. First is the rock fabric, or the tendency of the hydraulic fracture treatment to generate complex fractures. The higher the tendency to generate a complex network of fractures, the greater the optimal spacing will be between fractures. In this case one will tend to measure the effectiveness of the fracture in terms of Stimulated Reservoir Volume (SRV). In some shales it has been demonstrated that increased SRV will yield higher production and EUR [Mayerhofer 2008]. The flow capacity or conductivity of the fracture network combined with the SRV provide an assessment method to predict well performance and hydrocarbon recovery. As the tendency of the formation to generate complex fractures decreases, the optimal spacing between fractures becomes more tightly correlated with reservoir permeability and resulting fluid mobility in the formation. In reservoirs more prone to conventional bi- wing planar fractures, a greater number of closely spaced stages are required to recover reserves in lower permeability reservoirs [Cipolla 2009]. In general where reservoir permeability is the determining factor, it is not uncommon to see long horizontal drains with dozens of fracs spaced 10s to 100s ft apart. In the Haynesville Shale (Figure 2) and Bakken (Figure 3), for example, increasing the number of stages has led to increased production. Figure 2– Initial Production as a function of Lateral Length and Total Number of Frac Stages for one operator in the Haynesville Shale [Pope 2009]. Figure 3– Increasing lateral length as well as decreasing the spacing between fracture stages has yielded positive impacts on production, EUR and well economics [Rankin 2010]. Figure 1 – Average horizontal lateral length in the Louisiana Haynesville Shale has shown a steady increase since development began in 2007, and a corresponding increase in average IP [Pope, 2010]. 0 2 4 6 8 10 12 14 16 18 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Aug07 Nov07 Feb08 Jun08 Sep08 Dec08 Mar09 Jul09 Oct09 Jan10 InitialProductionRate,MMCFD Distancebetweentopandbottomperf,ft "LateralLength"andIPvsTime <= Lateral Length, ft LouisianaWells ONLY IP=> 0 1 2 3 4 5 6 7 8 9 10 0 1020304050607080 Lateral Length x # Stages / 1000 Reported IP, MMCFD #2 #1 #4 #7 #5 #6 #3 SPE 151128 3 Fracture Geometry and Conductivity. The fracture geometry optimization involves defining the desired fracture half-length, width and conductivity for maximized production. While there are several optimization methods, all involve a relative comparison of the flow potential of the fracture to that of the reservoir, as described by the Dimensionless Fracture Conductivity (F CD ) parameter below: F CD = [k frac *w frac ]/ [k form *X frac ] ………………………………………………………………………………(1) For steady or pseudosteady state flow in oil wells, several authors [Prats 1961, Cinco-Ley 1981and McGuire & Sikora 1960] have developed correlations that allow the engineer to use F CD to predict the benefits of the fracture stimulation, yielding a method that balances fracture half length and drainage area with fracture conductivity for stimulation design. F CD is also used to optimize the design of the fracture in such methods as the Unified Fracture Design [Economides 2002]. While the F CD concept and various related fracture optimization methods are well understood, many in the industry fail to identify the correct fracture permeability to plug into the equation correctly estimated at realistic (downhole) flow conditions [Palisch 2007]. Given the critical nature as well as generally overlooked impact of fracture conductivity, the following sections will be devoted to describing the deficiencies of the laboratory procedures, as well as provide references to adjust reported conductivities to reflect more realistic conditions within an actual fracture in order to guide the proppant selection process. Fracture Conductivity The concept of fracture conductivity is often overlooked as an important stimulation design variable in unconventional reservoirs. For some, the presence of nano-Darcy rock does not intuitively lead to the need for high fracture conductivity. However, while the fracture conductivity required to economically produce a horizontal well in an unconventional play and to improve hydrocarbon recovery will vary in different reservoirs, many engineers fail to recognize the conductivity requirements to accommodate high velocity hydrocarbon flow in transverse fractures. The pack conductivity for a given proppant is a function of the proppant particle size, strength, proppant grain shape (roundness and sphericity), embedment into the frac faces, fracturing fluid residue, fines migration, effective stress on proppant and fluid flow effects (non-Darcy and multi-phase flow) which can be very pronounced in the limited intersection between a wellbore and a transverse fracture. When accounting for these effects, it is not uncommon for proppant pack reference conductivity to be reduced by two orders of magnitude [Palisch 2007 and Miskimins 2005]. In the following sections the authors will review the standard testing methodology and deficiencies in more detail. Conductivity Testing and its Limitations. In order to understand realistic conductivity, one must first understand how conductivity is measured and reported. The conductivity of the fracture represents the product of the permeability of the fracture and the fracture width, and can be represented by the following equation: Conductivity = k frac *w frac … …………………………………………………………………………………(2) In 1989 the American Petroleum Institute (API) issued the first standardized procedures under API-RP-61 for measuring the conductivity of proppants in the lab using the Cooke Conductivity Cell [API 1989]. The procedure was modified through the years to include longer flowing times, replacement of steel shims with sandstone cores and testing at elevated temperatures. In 2006 the International Organization for Standardization (ISO) set the current standard under number ISO-13503-5 [ISO 2006]. In 2008 the API adopted ISO-13503-5 under API-RP-19D, effectively replacing API-RP-61 [API 2008]. These standards set testing procedures for evaluating sand, ceramic media, resin coated proppants, gravel packing media, and other materials used for hydraulic fracturing and gravel-packing operations. The objective was to provide a consistent methodology for proppant conductivity testing and comparing proppant materials under comparable laboratory conditions. Recognizing the standard’s limitation given the differing conditions between lab and realistic downhole conditions, API-RP- 19D specifically states it “is not intended for use in obtaining absolute values of proppant pack conductivities under downhole reservoir conditions” [API 2008]. The current procedure consists of placing a representative sample of proppant at 2 lb/ft 2 in the test cell between two Ohio sandstone wafers with a Young’s Modulus (YM) of 5 million psi. The cell is heated to 150°F or 250°F (depending on proppant type) and stress is ramped at a prescribed rate to the first test point. After 50 hours a set of measurements is made and the process can then be repeated at each desired stress, holding for an additional 50 hours at each stress. Conductivity is calculated by applying Darcy’s Law from the pressure drop produced by a 2 ml/min 2% KCl flow stream through the proppant pack. Conductivities measured using this test are normally reported in service and proppant company published literature and may be denoted as “reference”, “laminar”, “baseline” or “long term” conductivities. The key testing conditions are summarized below: • 2% KCl fluid pumped at 2 ml/min • 2 lb/ft 2 proppant loading • Sample placed between Ohio Sandstone wafers with YM = 5.0 Mpsi 4 SPE 151128 • Single stress maintained for 50 hr • Temperature 150° F (for sand) or 250° F (for ceramics) Although these standard conditions allow for comparable testing between proppants, they rarely represent the realistic conditions in which proppant is placed in hydraulic fractures [Vincent 2009]. As such, these procedures ignore many parameters that affect the actual conductivity of the frac. Further complicating matters, different proppant types may be affected differentially by each parameter. A brief description of the key effects is given below. The interested reader can refer to SPE 106301 for a full description [Palisch 2007]. Non-Darcy and Multiphase Flow effects. The ISO/API test flow rate of 2 ml/min is not representative of actual flow rates in a proppant pack. This rate would equate to ~6 BPD in a fully perforated vertical oil well with a 50 ft tall bi-wing frac achieving 2 lb/ft 2 concentration, or ~15 MSCFD flowing at 1,500 psi and 250°F in a similar dry gas well. The fluid velocities resulting from more prolific wells will cause tremendous amounts of energy to be lost, which translate into additional pressure losses not described by Darcy’s Law. Forchheimer’s equation (below) includes the non-Darcy pressure drop ( 2 ) component for a single phase fluid and is dominated by the velocity-squared term [Forchheimer 1901]. Interpreting this extra pressure drop as a conductivity reduction typically shows a fracture conductivity impairment of 50 to 85% [Palisch 2007]. …………………………………………………………… …(3) Additionally, the fluid circulated in the ISO/API tests is a solution of silica-saturated, oxygen free 2% KCl water. In reality oil and gas wells rarely produce 100% water, or even a single phase fluid for that matter. Instead, two or three phases are typically present (oil, water and gas), yielding a much more complex flow regime than tested in the lab. Multiphase effects have been described in many ways by various researchers. Lab data consistently demonstrate that pressure losses in the fracture may increase significantly when both liquid and gas phases are mobile within the fracture. This is typically attributed to the highly inefficient flow regime that occurs when gas, oil and water molecules move through the proppant pack, each moving at different velocity. In fact, some tend to consider multiphase flow impacts as a multiplier to non- Darcy effects since the impacts are most pronounced at high velocity flow. Unfortunately, significant pressure losses are documented even when only small percentages of a second phase are mobile within the fracture (Figure 4). Proppant loading at 2 lb/ft 2 . It is generally accepted that in most slickwater or hybrid frac stimulations, the effective proppant loading achieved in the fracture is less than 1 lb/ft 2 . This means that the fracture is narrower than in the ISO/API test. In addition to directly impacting conductivity via the conductivity equation (fracture perm x fracture width), the much narrower width produced by the reduced concentration also increases the fluid velocity through the pack for a given flow rate. This in turn exacerbates the non-Darcy and multiphase flow effects in the fracture. If the fracture width is halved, and hydrocarbon velocity is doubled, then non-Darcy pressure losses are increased by a factor of 400% (2 squared). . Embedment and Spalling. The ISO/API test uses a sandstone core with a YM of 5 million psi. Many shale and unconventional reservoirs are significantly softer than these sandstone cores (e.g. the Eagle Ford Shale has a YM of 1-3 million psi). Softer rock leads to a loss of width and conductivity due to both proppant embedment and formation spalling. The reduced width has the double effect of diminishing conductivity (directly proportional), and increasing fluid flow velocity due to the smaller cross section of the resulting proppant pack. As a consequence non-Darcy pressure losses will also be increased. Temperature Effects. As noted earlier, the ISO/API conductivity test is performed at 150°F for sand proppant and 250°F for ceramic proppant. The reason for this difference is primarily due to the known detrimental impact of higher temperature on sand and sand-based proppants (i.e. Resin Coated Sand). Specifically, as temperatures exceed 200°F, sand based products can Figure 4– The impact of Multi-phase flow can be dramatic at very low fractional flow rates of liquid [Palisch 2007]. Increased Pressure Drop due to Mobile Liquid in Proppant Packs - 10 20 30 40 50 60 0% 5% 10% 15% Fractional Flow of Liquid Multiplier of Total Pressure Drop 0.75 MMCFD 0.25 MMCFD Trend Increased Pressure Drop due to Mobile Liquid in Proppant Packs - 10 20 30 40 50 60 0% 5% 10% 15% Fractional Flow of Liquid Multiplier of Total Pressure Drop 0.75 MMCFD 0.25 MMCFD Trend SPE 151128 5 experience a significant decrease in conductivity (Figure 5). For example, an uncoated sand, when exposed to 250°F at 6,000 psi stress will lose 40% of its conductivity when compared to 150°F, and this loss jumps to nearly 80% at 300°F and 8,000 psi. Coating the sand with a resin lessens the damage because the resin can encapsulate the crushed fines. However, even resin coated sand loses 30% of its conductivity at 8,000 psi and 300° F. Ceramic proppants are tested at 250° F due to their thermal stability. These proppants are sintered at ~2,700°F and are engineered for improved sphericity, strength and thermal resistance. Therefore, no correction is required when placing a ceramic proppant into higher temperature formations. Cumulative Conductivity Impact. When all of these effects are taken together, the overall impact of these damage mechanisms on the conductivity at actual bottom hole flowing conditions can be severe. In fact, it is not uncommon to see the overall loss of conductivity exceeding 90% (Figure 6). It should also be noted that while all proppants experience these several orders of magnitude reduction in conductivity, the individual damage mechanisms can have different impacts on the various proppant types [Schubarth 2006]. While the above conductivity damage is already severe, there are also other downhole realities that can exacerbate the damage, including long term conductivity degradation as well as gel/fluid residue damage and many other mechanisms [Palisch 2007, Barree 2003, Pearson 2001]. Regardless of the exact magnitude of these reductions, the bottom line is that the realistic conductivity in all hydraulic fractures is much less than measured in standard lab testing, and reported in industry literature. Further, if these reductions are not accounted for when designing hydraulic fractures and/or selecting the appropriate proppant, significant production may be deferred or in some cases not recovered in the existing completion [Blackwood 2011]. Proppant Selection in Unconventional Reservoirs The most common completion in unconventional plays consists of a horizontal wellbore with multiple proppant fracs placed along it. Despite the very low reservoir permeability driving F CD up, high conductivity proppant is still needed given the detrimental effects discussed in previous sections. Additional to the conductivity considerations, there are several other issues that must be addressed when selecting the appropriate proppant for use in these multi-stage fracs in horizontal wells. These include flow convergence in transverse fracs, proppant transport when low viscosity fluids are employed, and proppant crush at the typical low concentrations employed. Figure 5 – The effects of temperature on conductivity for Sand-based proppants [Pope 2009]. Figure 6 – The cumulative reduction in conductivity due to several damage mechanisms not accounted for in the standard ISO/API test [Palisch 2007]. 0 0.2 0.4 0.6 0.8 1 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 1000 0 150 deg F 200 degF 250 deg F 300 deg F 350 deg F 20/40 Premium White Sand C on d uc ti v it y C orrec ti on f rom 150 d eg F , f ac t or Closure Stress, psi 0 0.2 0.4 0.6 0.8 1 0 2000 4000 6000 8000 10000 12000 1400 0 150 de g F 200 de gF 250 de g F 300 de g F 350 de g F 20/40 P rem i um RC S an d Conductivity Correction from 150 deg F, factor 1540 5720 685 4310 225 1410 85 547 25 167 7 120 0 1000 2000 3000 4000 5000 6000 Effective Conductivity (md-ft) ISO 13503-5 Test (Base Case) "Inertial Flow" with Non-Darcy Effects Multiphase Flow Lower Achieved Width (1 lb/sq ft) Gel Damage Fines Migration / Cyclic Stress Jordan Sand Lightweight Ceramic Effective conductivities can be less than 2% of API test values 99% reduction 98% reduction Cumulative Conductivity Reductions Using PredictK Closure Stress, psi 6 SPE 151128 Flow Convergence in Transverse Fracs. Let’s reiterate that the goal in unconventional plays is to place numerous transverse fracs along a horizontal lateral, as opposed to conventional plays which may exploit a single frac in a vertical well. Production into a horizontal wellbore from an orthogonal fracture will exhibit linear flow in the far field as it travels down the fracture(s). However, as the fluids converge on the relatively small diameter wellbore (Figure 7), the fluid velocities in that near wellbore region increase dramatically. In fact, if one considers a single planar 100 ft tall vertical fracture, and places it fully connected in a vertical well and transversely in a horizontal 6 inch diameter wellbore, the fluid velocity in the near wellbore would be 127 times higher in the transverse fracture as compared to the vertical well. Further, recall that velocity is a squared term in the Forchheimer (see previous discussion) pressure drop calculation, therefore, the pressure drop in the transverse frac could be over 16,000 times greater than in a fully connected vertical well. This leads to the conclusion that it is practically impossible to place enough conductivity near the wellbore in a transverse/HZ well to be fully optimized. Completions in unconventional resources will benefit from more conductivity near-wellbore in transverse fracs [Besler 2007, Rankin 2010, Shah 2010, Vincent 2011, Economides 2000]. Proppant transport and placement via low viscosity fluids. Proppant placement is governed by a series of mechanisms involving the interaction between the fracturing fluid and proppant. A number of issues have been investigated through time that impact how proppant is transported into the frac and its final location in the created geometry. Proppant density and size have a determining impact on proppant settling, which in turn impacts where proppant will be placed in the frac. The simplest single-particle settling mechanism can be described by Stokes law, in which the velocity of a single particle falling through a stagnant liquid medium can be described as follows: …… …………………………………………………………………….………(4) Where v fall is the settling rate in ft/s, d prop is the average particle diameter in inches,  is the fluid viscosity in cp, and  prop and  fluid respectively are the specific gravity of the proppant and the fluid [Economides 2000]. The settling rate is directly proportional to the difference in density between the fluid and proppant, and inversely proportional to the fluid viscosity. This last condition makes settling an important consideration when pumping low viscosity Newtonian fluids as are typically used in HMF treatments conducted in dry gas shales. While Stokes Law does not fully describe the proppant transport in hydraulic fractures due to the many additional considerations for calculating settling rate, it shows two components are directly controlled by the proppant: proppant density and diameter. While much attention is typically given to proppant density, proppant diameter can actually be of greater importance in a fracturing treatment. As stated in Stokes law, settling velocity is proportional to d prop squared, thus having an exponentially larger effect on settling rate than fluid viscosity. As an example, despite being more dense, a smaller diameter 40/70 2.65 ASG LWC/RCS/Sand particle settles slower than a 20/40 1.75 ASG Proppant (Figure 8). Again it should be noted that while there are significant limitations of using Stokes Law to describe setting under dynamic conditions in a slurry situation it does serve the purpose to illustrate how smaller and lighter proppant aid easier placement. It is therefore no surprise that the most popular slickwater proppants are currently 40/80 LWC and 40/70 Sand/RCS. Extensive research and experimentation have been carried out to better describe and assess proppant placement and can be referenced outside this paper [Palisch 2008, Dayan 2008, Mobbs 2001]. Figure 7 – Fluids flowing within the hydraulic fracture in horizontal wells must converge into an extremely (relatively) small area as they cut transversely with the wellbore [Shah 2010]. Figure 8 – Relationship between proppant density and proppant diameter on settling rate in 2% KCl [Palisch 2008]. SPE 151128 7 Proppant crush at low concentrations. The typical low proppant concentrations pumped in waterfracs often designed for unconventional gas reservoirs can result in a low areal concentration being placed in the frac. Values between 0.25 and 0.50 lb/ft 2 are typical and much lower than the 4.00 lb/ft 2 load used in ISO 13503-2/API-RP-19-C crush test, or the 2 lb/ft 2 used in the standard ISO/API conductivity test. The impact of these low concentrations on proppant pack conductivity (due to the narrower width) were discussed previously in this paper. However, an additional (and often overlooked) result of these narrower fractures is the impact on proppant crush. When proppant grains are loaded into a crush cell, particles can be considered either interior or exterior grains. Grains in the interior of the pack are “protected” due to their contact with six to twelve neighboring grains, thus providing uniform stress distribution on the individual gains. However, exterior grains have fewer contact points leading to greater stress at the points of contact. For this reason, exterior grains experience greater damage in the crush and conductivity cells, and ultimately the fracture. Therefore, as proppant pack width (and proppant areal concentration) decreases the exterior grains comprise a larger percentage of the total grains in the pack, thereby leading to higher proppant crush [Palisch, 2009]. Some have also proposed partial monolayers as a means to boost conductivity, the idea being that voids between grains would provide open paths with infinite conductivity [Brannon 2004, Parker 2005]. Using conventional proppants (Sand/RCS/LWC), a partial monolayer will occur at concentrations of <0.20 lb/ft 2 , where less than a single layer of proppant should occur. While there is significant debate regarding whether partial monolayers can be reliably achieved over large portions of a created frac [Gidley 1989, Palisch 2008], even if they can be successfully placed many overlook the increased stress concentrated on individual proppant grains. This will lead to higher crush, higher embedment, and ultimately loss of fracture width and conductivity. Various specialty proppants have been introduced to exploit the advantages of partial monolayers, as well as purportedly promote their placement. Most of these new proppants are much lighter density (from 1.75 g/cc to nearly buoyant), formed from various substrates, including resin coated porous ceramic and/or walnut hulls, thermoplastics, nanocomposites, polymers and other resin or plastic components. In many cases these proppants do not “crush” as conventional rigid particles do, but instead deform, which is one reason why they are typically only considered useful at low stress. Caution should be exercised when employing these deformable proppants, however, as their usefulness is limited only to partial monolayers. Independent testing has shown that if these deformable proppant grains are actually placed in a traditional pack whereby they come in contact with each other, the grains tend to “squish” together and become a relatively impermeable plug [Stimlab 2009-2010]. In summary, increased crush, concentrated stress on individual proppant beads and embedment (the latter two in partial monolayers only) occur when low areal concentrations are placed. This phenomenon takes place regardless of whether the proppant is a Sand, Resin Coated Sand, Ceramic or specialty deformable proppant, constituting additional sources of frac width and conductivity loss one must consider for adequate proppant selection in unconventional plays. Proppant Selection Case Histories When one understands the realistic conditions within the proppant pack, and their impact on fracture conductivity, it becomes apparent that the fracture flow capacity is not optimized (i.e. the F CD is much lower than anticipated) in horizontal multistage fractures in unconventional reservoirs. Further, it means that in general, anything that can be done to increase the conductivity of the fracture should yield a corresponding increase in production. While there are many ways to increase the conductivity of a fracture, and all should be considered when designing fracture stimulations, one of the easiest and most common is to upgrade the proppant size and/or type. As one moves up the Proppant Conductivity pyramid, fracture conductivity (and production) improves (Figure 9). However, moving up the pyramid typically carries with it an increase in completion (proppant) cost. Therefore, the decision to increase conductivity must also involve an economic analysis, and ultimately will become an economic decision. So the process for selecting proppant (as well as any design changes) must involve four steps. 1. Calculate the conductivity of the fracture at realistic conditions 2. Predict the production performance achieved with each proppant 3. Evaluate the cost vs. benefit and select the proppant that maximizes the economics of the completion 4. Review the actual field production benefits to ensure validity to the previous evaluations The first two steps must typically be performed through the use of a fracture propagation model that is coupled to a reservoir simulator/model. The model must be able to account for the realistic conditions of the fracture and the corresponding impact Figure 9 – The Economic Conductivity pyramid showing the three Tiers of proppant. 99% of all proppant can be placed into one of three Tiers. As one moves up the Triangle, proppant p erformance ( conductivit y) im p roves [ Galla g her 2011 ] . 8 SPE 151128 of fracture conductivity. Step 3 can then be performed using the economic hurdles for the given situation; some production simulators automate this function. The last step is often the most overlooked step in the process, due to the significant activity level required of most engineers involved with exploiting unconventional reservoirs. The authors will present in the following sections, several case histories from unconventional reservoirs in which proppant was selected considering the realistic conductivity at bottomhole flowing conditions and the economic impact on the completion. These cases illustrate the robustness of the approach described above and demonstrate the production and economic benefits of placing enhanced conductivity in ultra-low permeability formations. Case History 1 – Barnett Shale. The successful development of the Barnett Shale is considered by most to be the driving force behind the success of today’s unconventional reservoir exploitation. Shortly after companies “cracked the code” for successful completion practices (multistage fracs along horizontal wellbores) in the early 2000’s, the Barnett Shale became one of the hottest plays in the United States. While depressed natural gas prices have tempered the activity in the play for the last several years, the Barnett still enjoys widespread activity. The Barnett Shale is a Mississippian-aged shale located between 6,500 and 8,500 ft TVD. It currently covers as many as 15 counties in the Fort Worth Basin of north Texas, and at its thickest point can be as much as 1,000 ft thick. The Barnett is a thermogenic reservoir and averages 4.5% Total Organic Content (TOC). As is the case with all shale gas plays, the primary challenge in the play is the ultra-low permeability of <0.0001 md. Numerous studies have been documented on the completions practices in this shale gas play. One such study illustrated the benefits of increasing the conductivity of the hydraulic fractures [Cipolla 2009]. An actual Barnett Shale completion was history matched and then sensitivities were performed to several parameters including fracture conductivity and fracture stage spacing. The well of interest utilized a Tier 3 uncoated sand and was stimulated using a slickwater fluid system. The history match indicated that the realistic fracture conductivity was ~2 md-ft, and showed that if the conductivity were increased from 2 md-ft to 20 md-ft, the well would see a 1 BCF increase in the 15 year cumulative production (Figure 10). In addition, the study also illustrated that the optimal staging between fractures (stage spacing) was highly dependent on the conductivity of the fracture (Figure 11). Namely, as the fracture conductivity was increased, the optimal spacing increased. While all scenarios modeled would need to be evaluated on a cost vs. benefit basis, the study illustrated the importance of accurately estimating the realistic fracture conductivity, as well as the overall value of increasing the conductivity. Figure 10 – Barnett well history match showing the potential impact of increasing the conductivity of the fracture. The blue line represents actual well production [Cipolla 2009]. Figure 11 – The relationship between main fracture conductivity and spacing [Cipolla 2009]. SPE 119366 Potential Target of 1 Bcf over 15 years Actual Well Production SPE 151128 9 Case History 2 – Haynesville Shale. Using the learnings from the Barnett, the industry opened up the shale gas frenzy when the multistage fracturing in horizontal wells was applied to the Haynesville Shale. Its development confirmed that this technology could be successful in other shale plays. The Haynesville Shale is a late-Jurassic age shale that is found between the Cotton Valley Group and the Smackover Limestone in East Texas and North Louisiana. It is a black, organic-rich shale with a TOC of 3-5% and 1.3- 2.4 vitronite reflectance (Ro). Although the shale is laterally extensive and variable, most of the development occurs at depths between 11,000-13,000 ft, and in areas where the total thickness is 150-400 ft. While porosity is moderate (6-12%), permeability is extremely low (5-800 nanoDarcies). The Haynesville Shale also has elevated temperature (>300° F) and reservoir pressure (0.84-0.88 psi/ft) [Pope 2009, 2010]. Since most operators in the HV Shale initially adapted their frac designs from their experience in the Barnett, the primary fluid system of choice was either a slickwater or “hybrid” design. Small mesh (40/70 or 40/80) proppant was primarily utilized given transport concerns with low viscosity fluids. At realistic conditions (Figure 12), Tier 1 proppants have two to twenty times the conductivity as Tier 2 and Tier 3 proppants, respectively [Pope 2009]. However, despite the high temperatures and stresses, and this conductivity disparity, operators in the Haynesville shale have used tremendous volumes of all three Tiers of proppants. This is primarily due to higher cost and limited availability of the Tier 1 proppants. However, this large diversity of proppant usage has allowed for an opportunity to evaluate actual field performance comparing proppant types. Once such study is comprised of a well “Set A” containing 56 wells operated by the same company within a 5 mile radius [Pope 2010]. Twenty of these wells were known to contain Tier 1 40/80 Lightweight Ceramic proppant, while 36 offset wells were known to contain primarily a Tier 2 40/70 Premium RCS. While there is significant debate over whether IP (initial production) is a good indicator of completion performance in shale gas plays, the authors illustrated that increasing the conductivity (from Tier 2 to Tier 1) of the fracture yielded nearly 50% increase in IP normalized for flowing pressure and lateral length (Figure 13). As stated previously, when comparing different completions, IP data alone can present concerns. IP is just a snapshot in time and is an imperfect measure of success. In addition, as a well is produced the stress on the proppant typically increases, therefore it is important to also look at the longer term performance of proppants. A comparison of production for the same operator shows the Tier 1 wells are producing, on average, 20% more normalized production after 6 months than the Tier 2 wells. In addition, a smaller subset of wells “Set B” (10 Tier 1 and 19 Tier 2) have at least 12 months of production. In these wells, the Tier 1 completions had produced 30% more normalized production than the Tier 2. The authors hypothesized that this increase may be indicative of the durability advantage of the Tier 1 proppant [Pope 2010]. In a follow up study, the above well sets were reanalyzed using an additional 12 months of data for each well set [Blackwood 2011]. After normalizing for lateral length, the Tier 1 wells in “Set A” have produced over 20% more gas Figure 12 – Realistic Conductivity Comparison at Haynesville/Bossier Shale conditions for several 40/70 proppants show that Tier 1 ceramic proppants (red and black lines) are far superior to Tier 2 and 3 proppants at 10,000 psi stress [Pope 2009]. Figure 13 – Cumulative frequency plots comparing Tier 1 (premium) proppant to Tier 2 (other) proppant, using IP data that has been normalized to both the flowing pressure and lateral length (right), show that increasing conductivity yields a 30-40% increase in production [Pope 2010]. Figure 14 – Per well projected average cumulative incremental present value (PV10) of Tier 1 performance versus Tier 2 performance, assuming $5/mcf gas price, $60/bbl condensate and 10% discount factor. Regardless of whether projecting from “6 month” or “12 month” wells, there is tremendous value to upgrading the conductivity and the incremental investment pays out in less than 2 months [Blackwood 2011]. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 CumFrequency PINormalizedtoLateralLength,MCFD/psi/ft "PI"NormtoLLCumFrequency OperatorA OtherProppant PremiumProppant 0.00 1.00 2.00 3.00 4.00 5.00 6.00 0 24 48 72 96 120 144 168 192 216 240 CumIncrementalValue,MM$ Months CumulativeIncrementalValue 12MonthWells 6MonthWells 10 SPE 151128 than the Tier 2. This represents an additional 340 MMCF of gas, or $2 million in value, per well. Similarly, the Tier 1 wells in “Set B” have produced an average of 39% more gas (or 500 MMCF gas) per well than Tier 2 wells. That represents nearly $3 million in incremental present value, and a tremendous return on the investment required to upgrade to Tier 1 proppant. The authors also performed hyperbolic decline curve analyses on the updated production, projecting recovery out to 20 years. For the two well sets, it is estimated that the Tier 1 wells will produce an average of 1.2-1.6 Bcf additional gas over the Tier 2 wells, and generate $4 – 5 million in incremental value over 20 years, and pay out the additional investment in proppant in less than 2 months (Figure 14). Case History 3 – Eagle Ford Shale. The Eagle Ford formation is an Upper Cretaceous deposit in the Gulf Coast region of South and Central Texas. It has been recognized for many years and considered the source rock for several of the producing reservoirs in South Texas. However, it wasn’t until completion practices used in the Haynesville Shale were employed in the Eagle Ford that it became economically viable to develop. Since that time development has accelerated significantly, with nearly 200 active rigs in late 2011. While it is typically referred to as a shale, the Eagle Ford is actually an organic rich calcareous mudstone which is prevalent across 6 million acres spanning 20 counties. The reservoir and geologic characteristics can vary significantly across the play, with TOC ranging from 1-7%, depths from 6,000-13,000 ft TVD and total thickness as much as 250 ft. In addition, the reservoir fluids range from primarily oil in the northwest, to liquids-rich condensate in the most active central portion, to dry gas in the deepest areas to the south. Since development of this play did not begin in earnest until the latter half of 2009 and into 2010, and given the geologic and reservoir variability across the play, completion strategies are still being developed and optimized in the Eagle Ford. However, one study was recently published that documents the impact of fracture conductivity in Eagle Ford completions [Bazan 2010]. Two wells were modeled and history matched in this study. They are located in the condensate and gassy areas of LaSalle County. The depth of the “gassy” well, Well A, is ~11,000 ft TVD, and of the condensate well, Well B, is ~8,500 ft TVD. Both wells were drilled with 4,000 ft laterals and contained 10 (Well A) and 12 (Well B) stages. Since these were some of the first wells completed for this operator, additional data were collected to assist in the history match, including radioactive (RA) proppant tracers and microseismic mapping. While significant work was performed and documented, two items of note will be discussed here. First, the fracture propagation and production match confirmed that the realistic fracture conductivity, despite using Tier 1 lightweight ceramic proppant, was much lower than measured in the standard conductivity test. In Well A, the matched conductivity was ~2 md-ft, while in Well B it was 1.75 md-ft. Keep in mind that these wells each produced in excess of 6 MMCFE per day, so the stimulations and completions were successful despite this low realistic conductivity. However, similar to the work presented earlier in the Case 1 Barnett study, when proppants are placed into downhole conditions, the conductivity can be damaged quite severely. Therefore, it is no surprise that when sensitivities were run in these models, the impact of conductivity was quite significant (Figure 15). A Tier 2 RCS would provide 75-100% greater cumulative production than a Tier 3 Uncoated Sand, and the Tier 1 LW Ceramic provides an additional 20-50% over the Tier 2, after just 3 years of production. The condensate well (B) enjoyed the largest increase, which is likely due to the impact of multiphase flow and the corresponding need for additional conductivity. Figure 15 – Upgrading from a Tier 3 Sand (blue) to a Tier 2 RCS (red) yields a 3 year increase in production of 75% (Well A-top) and 100% (Well B-bottom). These increases jump an additional 20% (Well A) and 50% (Well B) when upgrading to a Tier 1 LWC [Bazan, 2010]. [...]... September Cinco-Ley, H and Samaniego-V., F.: “Transient Pressure Analysis for Fractured Wells”, JPT (September, 1981) 1749 Cipolla, C.L.: “Modeling Production and Evaluating Fracture Performance in Unconventional Gas Reservoirs”, paper SPE-118536, SPE Distinguished Author Series 2009 Cipolla, C.L et al 2009: Fracture Design Considerations in Horizontal Wells Drilled in Unconventional Gas Reservoirs”,... count [Rankin 2010], but indicated a Time, Days tremendous increase in well productivity when comparing Tier 1 ceramic completions to Tier 3 wells However, this Figure 17 - Red wells contain Tier 1 ceramic and utilize plug and perf staging, while Blue wells contain Tier 3 sand and same operator revisited the well set, adding additional wells sliding sleeve staging [Vincent 2011] and data (time) [Vincent... Haynesville and Eagle Ford, are in short supply, and therefore many operators must use whatever is available, 100,000 including Tier 3 Sand One leading Bakken operator has published results showing the benefits of increasing the 50,000 conductivity while at the same time maximizing the number of stages in the well, utilizing a plug and perf methodology in 0 uncemented liners Initial publications were based... “Using Reservoir Modeling To Evaluate Stimulation Effectiveness in Multilayered “Tight” Gas Reservoirs: A Case History in the Pinedale Anticline Area”, paper SPE-100574 presented at the Gas Technology Symposium, Calgary, Alberta, Canada, 15-17 May StimLab Proppant Consortium Notes from 2009-2010 Vincent, M.C 2002: “Proving It – A Review of 80 Published Field Studies Demonstrating the Importance of Increased... “Improving Production in the Eagle Ford Shale with Fracture Modeling, Increased Conductivity and Optimized Stage and Cluster Spacing Along the Horizontal Wellbore”, paper SPE-138425, presented at the SPE Tight Gas Completions Conference, San Antonio, TX, 2-3 November Besler, M.R , et al 2007: “Improving Well Productivity and Profitability in the Bakken – A Summary of Our Experiences Drilling, Stimulating... completion can be employed In addition, selecting the appropriate proppant for use in multi-stage fracs in horizontal wells of unconventional reservoirs require accounting for several additional impacts, including flow convergence in transverse fracs, proppant transport if low viscosity fluids are employed, and proppant crush at the typical low concentrations achieved in many resource plays The proppant selection... conductivity [Vincent 2009] in a wide variety of formations and well types Four recent examples involving horizontal wells have been presented in this paper, including studies in the Barnett, Haynesville, Eagle Ford and Bakken These case histories support the robustness of the proppant selection process outlined above, based on realistic conductivity and impact on stimulation economics The benefits of placing... proppant increases, therefore, it is important to evaluate the increase in production and perform a cost-to-benefit analysis It is estimated that after 6 months, this production increase represents ~$1.4 million in incremental value generated by the Tier 1 conductivity, per well, which more than offsets the investment of upgrading from Tier 3 to Tier 1 Additional studies have also been documented in the... placing enhanced conductivity with high quality proppant in ultra low permeability formations is illustrated by the case studies Simply put, fractures rarely can be characterized as “infinitely conductive”, even in ultralow permeability reservoirs Fractures must be designed to accommodate realistic conductivity reductions as presented in the paper Unconventional resources are sometimes called “technology... milliliters per minute Million Standard Cubic Feet Million Standard Cubic Feet Equivalent Million Standard Cubic Feet per Day Million pounds per square inch pounds per square inch pounds per square inch per foot Present Value at 10% discount rate Radioactive Resin Coated Sand Stimulated Reservoir Volume Total Organic Content True Vertical Depth Velocity Settling rate Fracture Width Young’s Modulus Fracture