Summary Hydraulic fracturing treatments that use treated water and very low proppant concentrations (commonly referred to as water-fracturing treatments or "waterfracs") have been successful in stimulating low-permeability reservoirs. However, the mechanism by which these treatments provide sufficient conductivity is not well understood. To understand the effects of fracture properties on conductivity, a series of laboratory conductivity experiments was performed with fractured cores from the east Texas Cotton Valley sandstone formation. The results of this study demonstrate that fracture displacement is required for surface asperities to provide residual fracture width and sufficient conductivity in the absence of proppants. However, the conductivity may vary by at least two orders of magnitude, depending on formation properties such as the degree of fracture displacement, the size and distribution of asperities, and rock mechanical properties. In the presence of proppants, the conductivity can be proppant- or asperity-dominated, depending on the proppant concentration, proppant strength, and formation properties. Under asperity-dominated conditions, the conductivity varies significantly and is difficult to predict. Low concentrations of high-strength proppant overcome the uncertainty associated with formation properties and provide proppant-dominated conductivity. The implication of these results is that the success of a water-fracturing treatment is difficult to predict because it will depend significantly on formation properties. This dependence can be overcome by using high-strength proppants or proppants at conventional field concentrations. Introduction Although proppants are routinely used to achieve conductivity during hydraulic fracturing treatments, recent fracturing treatments using treated water and very low proppant concentrations (commonly referred to as water-fracturing treatments or "waterfracs") have been successful in low-permeability reservoirs.1–4 The mechanism by which these treatments provide sufficient conductivity is not well understood. The presence of residual fracture width caused, for example, by surface asperities and proppant bridging, and the lack of damage associated with the use of gels in conventional proppant treatments, are possible explanations.2,5 Residual fracture width has been observed during laboratory experiments6 and field tests7 and can be attributed to the combined effects of surface roughness and fracture displacement.8 The surface asperities are thought to withstand high formation-closure stresses and create sufficient conductivity for wells completed in very low-permeability formations. The low concentrations of proppant are added to supplement the asperities and improve overall fracture conductivity. Factors affecting fracture conductivity and proppant-pack permeability have been reported in the literature. The importance of parameters such as fracture displacement, fracture roughness, mechanical properties, and closure stress on fracture conductivity have been demonstrated in the absence of proppants.9–12 When proppants are present, parameters such as proppant strength, proppant concentration, and closure stress have been shown to be important.13,14 However, these studies were performed with fractured cores in the absence of proppants or with proppant and flat, parallel core faces. No study has addressed the effects of fracture properties on conductivity in the presence of low concentrations of proppant (i.e., conditions that may exist during water-fracturing treatments). This paper investigates the effects of fracture properties on conductivity for a variety of conditions ranging from fractured systems to water-fracturing conditions to conventional proppant fracturing conditions. A series of laboratory conductivity experiments was performed with fractured cores from the east Texas Cotton Valley sandstone formation. Jordan sand and sintered bauxite proppants were used at concentrations of 0, 0.1, and 1.0 lbm/ft2, and the conductivity was measured at effective closure stresses ranging from 1,000 to 7,000 psi. The work investigates the relative influence of proppants and asperities on conductivity and demonstrates the benefits of using proppants. Water-Fracturing Treatments. Water-fracturing treatments discussed in the literature use a water-based fluid containing friction reducer (usually a manmade synthetic polymer, but a low concentration of natural guar polymer is sometimes used as a substitute for the friction reducer), clay stabilizers, and surfactants as necessary. This fluid is intended to serve as the pad fluid and to provide proppant transport. Common water-fracturing treatments involve pumping a pad fluid for the first 50% of the job, followed by a proppant stage where the proppant concentration is held constant at 0.5 lbm/gal. At the end of the job (usually the last 5%, based on fluid volume), the proppant concentration is increased to 2 lbm/gal. The higher proppant concentration is intended to improve connection between the wellbore and the fracture. Some potential problems with water-fracturing treatments include low conductivity and poor proppant transport. In low-permeability formations, low fracture conductivity is not a major limitation to production, provided the conductivity is not too low. The poor proppant transport is caused by the low viscosities of the water-fracturing fluids and results in rapid settling of the proppant particles. This inability to carry proppant a significant distance away from the wellbore can severely limit the effective fracture length. Fracture length is the key variable for initial production potential and ultimate recovery from very low-permeability formations. Therefore, if the proppant does not get transported toward the tip of the fracture, the success of the water-fracturing treatment will depend entirely on the conductivity created by surface asperities or some other mechanism. Experimental Methods Fractured Cores. Sandstone cores from the east Texas Cotton Valley formation were used in this study. The cores (which were obtained from depths ranging from 8,500 to 10,000 ft) had porosities of approximately 12% and permeabilities of approximately 0.05 md. Rock mechanical properties were determined from static triaxial compressive strength tests conducted at ambient temperature and 4,500 psi confining pressure. Young's modulus ranged from 3.6×106 to 7.0×106 psi, and Poisson's ratio was approximately 0.32.

Summary The fracture-propagation process performed with polymer-based fracturing fluids is applied commonly to increase the productivity of producing wells, especially in tight gas formations. The fracture-cleanup process is complex and may suffer from the presence of a yield stress, non-Newtonian fluid in place, and both mechanical and hydraulic damage to the matrix near the fracture face. A previously published fast-and-robust single-well model was applied to study the important parameters involved in the fracture-cleanup process. This three-phase 2D model proved useful for assessing the significance of reservoir capillary pressure, broken-gel viscosity, yield stress, formation damage, and fracture conductivity on low-permeability-gas-reservoir production, with studied permeabilities ranging from 0.005 to 5 md. The observed trends may not carry over to nanodarcy reservoirs, such as the gas shales. The three phases included gas, water, and fracturing gel. Introduction Hydraulic fracturing has been used as a successful technology to increase productivity by means of significantly increased contact between the wellbore and the producing formation. To propagate an open fracture into a reservoir, fracturing fluids have been used to provide the two main functions of initiating and propagating the fracture and transporting propping agents along the fracture. Guar gum is the earliest example of an aqueous, viscous fluid used during the injection. The fracturing fluid must be viscous to allow the transport of the proppant during the injection, and it must have the ability to be broken easily after the injection to maintain high conductivity in the fracture during the production phase. To accomplish these tasks, crosslinkers (such as borates and zirconates) and delayed breakers (either oxidizers or enzymes) are added typically to the fluid (Economides and Nolte 2000). Injection of the viscous fracturing fluid results in fluid loss to the matrix and filter-cake formation. Filter cakes with high polymer concentration form on the faces of the fracture during the injection. Original fracturing fluid may remain in the fracture unless the fracture-face filter cake occupies the entire pore space of the propped fracture following closure (Ayoub et al. 2006). Varying exposure times to fracturing fluid (Seright 2002) cause local polymer-concentration changes along the fracture. Thus, breakers are seldom distributed uniformly, and the break of the concentrated fluid is seldom complete. At the end of a fracture treatment, there is normally a shut-in period to allow fracture closure during which fluid continues to leak off into the reservoir. Alternatively, and especially for tight gas reservoirs, the fracture can be forced to close by flowing back some of the fracturing fluid at controlled rates to prevent disturbing the proppant pack significantly. As a result, hydraulic fractures contain partially broken fracturing fluid, and residues remain after the breaker reacts with the polymer. It has been postulated that fracturing fluids need a minimum pressure gradient to begin the cleanup process in the proppant pack (May et al. 1997), and this has been verified experimentally (Ayoub et al. 2006). The fracturing process, depending upon reservoir-matrix permeability, can cause mechanical damage through various mechanisms including fluid invasion into the reservoir, polymer-solids deposition near the fracture face as filter cake forms, clay swelling in the case of incompatible fluids, broken-polymer/fines migration into the reservoir matrix, and chemical interactions between the fracturing fluid and the matrix such as pH alteration or polymer adsorption (Holditch 1979). In addition, hydraulic damage occurs from the increase in water saturation caused by leakoff. The hydraulic damage can include a reduction in gas relative permeability and relative permeability hysteresis in the matrix where fracturing fluid has leaked off as the water saturation is first increased during leakoff and then decreased during the production phase. A shift in the capillary pressure curve to higher values can also result from mechanical damage. The production process becomes even more complicated in tight gas formations with permeability less than 0.1 md when the combined effects of closure stress, non-Darcy flow, high capillary pressure in the matrix, and viscous fingering in the proppant pack cause additional issues and restrict the production rate. The objectives of this study were to develop a basic understanding of the major factors impacting the fracture-cleanup process in tight gas formations with permeability of 0.005 md or greater, including yield stress of the filter cake, capillary pressure changes, and formation damage, by use of available numerical models. A three-phase, 2D model reported in the literature (Friedel 2004) was used for this study.

A field study in east Texas showed that formation water production influences polymer recovery from hydraulic fractures. This study was conducted on 10 wells located in the Cotton Valley Taylor formation. A typical fracture stimulation design included between 9,000 to 14,000 bbl of zirconium crosslinked guar gel. Detailed chemical analysis of flowback samples was used to identify the effect of formation water on polymer recovery. Polymer and chloride concentrations were measured. Water produced during flowback averaged 52 5% of the fluid pumped on the jobs, while polymer recovery during flowback averaged 35 + 6% of the amount pumped. Polymer concentration in the flowback fluid from all the wells declined over time, as chloride concentration increased. This is attributed to the production of formation water. Flowback rate has a minimal effect on polymer recovery from these wells. The results of this study are compared with those previously obtained from another low permeability formation, specifically the Codell formation in Colorado. Flowback analysis is an important tool to determine how specific reservoir conditions influence fracture cleanup. Introduction Understanding how hydraulic fractures clean-up is essential for improving well stimulation. Residual gel can damage fracture conductivity, shorten effective fracture half- length, and limit the productivity of the well. The drive to develop fluids, additives, and procedures that minimize this damage continues to be a dominant theme in fracture fluid development programs. For example, reduced-polymer and viscoelastic surfactant (VES) fluid systems minimize damage by reducing or eliminating the polymer required to execute a treatment. Fluid additives have been developed to improve polymer recovery, including encapsulated breakers and aggregate dispersants. Aggressive procedures using fibrous material to maintain proppant flowback control have also been developed to maximize fluid recovery. Still, our understanding of the fundamental physical and chemical processes governing fluid recovery from hydraulic fractures is immature. Fracture cleanup is a complex problem, and many parameters - fluid system, job design, flowback procedure, and reservoir conditions - can influence polymer and fluid recovery efficiencies. Often specific products and methods that work well in one reservoir have little effect in other situations. Well productivity is the ultimate measure of treatment effectiveness. However, it does not provide direct information on fracture cleanup. A well with poor cleanup in a relatively rich pay zone may give superior performance compared to a well which had excellent cleanup but was in a relatively poorer pay zone. Additional information besides well productivity is needed to determine how the various reservoir and treatment parameters influence fracture cleanup. Systematic analysis of fluid and polymer returns after a treatment is completed is the only method to quantify fracture cleanup. In this paper this is referred to as flowback analysis. Water production and polymer returns are determined as a function of time and/or flowback volume. The concentration of other chemical species, such as chloride ions, is also determined to provide data on the mechanisms of fluid recovery and on fluid/formation interactions. Considering this is a basic approach, surprisingly little information is available in the published literature on actual flowback analysis field studies. Only load water recovery is commonly recorded after a fracturing treatment. However, as will be discussed below, load water provides insufficient information to quantify fracture cleanup, especially if the formation produces water. P. 531^

Hydraulic fracturing treatments using treated water and very low proppant concentrations (" waterfracs") have been successful in stimulating low-permeability reservoirs. However, the mechanism by which these treatments provide sufficient conductivity is not well understood. To understand the effects of hydraulic fractures on conductivity, a series of laboratory conductivity experiments were performed with hydraulically fractured cores from the East Texas Cotton Valley sandstone formation. Jordan sand and sintered bauxite proppants were used at concentrations of 0, 0.1 and 1.0 lb m/ft2, and the conductivity was measured at effective closure stresses ranging from 1,000 to 7,000 psi. The results of this study demonstrate that fracture displacement is required for surface asperities to provide residual fracture width and sufficient conductivity in the absence of proppants. However, the conductivity may vary by at least two orders of magnitude depending on formation properties such as the degree of fracture displacement, the size and distribution of asperities, and rock mechanical properties. In the presence of proppants, the conductivity can be proppant or asperity dominated, depending on the proppant concentration, proppant strength and formation properties. Under asperity dominated conditions, the conductivity varies significantly and is difficult to predict. Low concentrations of high-strength proppant reduce the effects of formation properties and provide proppant dominated conductivity. At conventional proppant concentrations, conductivity experiments performed with flat, parallel core faces tend to overestimate the conductivity observed with hydraulic fractures. Actual hydraulic fracture conductivity may be as much as an order of magnitude lower in the presence of low strength proppant. An important implication of this study is that the success of a " waterfrac" treatment is difficult to predict because it will depend significantly on formation properties. This dependence can be overcome by using high strength proppants or proppants at conventional field concentrations. Introduction Although proppants are routinely used to achieve conductivity during hydraulic fracturing treatments, recent fracturing treatments using treated water and very low proppant concentrations (" waterfracs") have been successful in low-permeability reservoirs1–4. The mechanism by which these treatments provide sufficient conductivity is not well understood. The presence of residual fracture width caused, for example, by surface asperities and proppant bridging and the lack of damage associated with the use of gels in conventional proppant treatments are possible explanations2,5. Residual fracture width has been observed during laboratory experiments6 and field tests7 and can be attributed to the combined effects of surface roughness and fracture displacement8. The surface asperities are thought to withstand high formation closure stresses and create sufficient conductivity for wells completed in very low-permeability formations. The low concentrations of proppant are added to supplement the asperities and improve overall fracture conductivity. Factors affecting the conductivity of hydraulic fractures and proppant packs have been reported in the literature. The importance of parameters such as fracture displacement, fracture roughness, mechanical properties, and closure stress on fracture conductivity have been demonstrated in the absence of proppants9–12. When proppants are present, parameters such as proppant strength, proppant concentration, and closure stress have been shown to be important13–14. However, these studies were performed with hydraulic fracture in the absence of proppants or with proppant and flat, parallel core faces. No study has addressed the effects of hydraulic fractures on conductivity in the presence of low concentrations of proppant (i.e., conditions that may exist during waterfrac treatments).

The ability to determine the effective half-length and conductivity of hydraulic fractures is important for estimating a well's long-term production performance. The determination of these properties, along with the formation effective permeability, can result in more accurate predictions of the ultimate hydrocarbon recovery. In very low permeability reservoirs, fracture half-length is the key to optimum reservoir development. Being able to quantify these properties also allows for improved understanding of the effects of treatment design changes. Post-treatment pressure buildup testing has been the most common method for determining the effective length of hydraulic fractures. One of the major drawbacks of the pressure transient test in very low permeability reservoirs is the extremely long shut-in time required to observe a sufficient amount of the fractured well transient behavior to properly characterize the formation and fracture properties. This long shut-in time is undesirable due to the fact that the well is not able to produce and generate revenue during this time. This paper reports on the most recent results of an ongoing study of the production performance of hydraulically fractured wells. The focus of the study is a comparison of the performance of conventionally fractured wells and those that have been completed with the treated water and low proppant concentration ("waterfrac") technique. A new evaluation technique for comparing the effectiveness of the treatments utilizing production data is introduced. The advantages and limitations of the production data analysis technique are discussed, as well as an improved understanding of the results of waterfrac treatments in low permeability gas reservoirs. The use of a comprehensive suite of analysis techniques for the production performance of fractured wells to obtain estimates of fracture half-length, fracture conductivity and formation effective permeability is detailed. Specialized diagnostics, performance history matching with analytic solutions and specialized type curve analyses have been used for several areas to estimate the fracture and formation properties from the bilinear, formation linear and pseudo-radial flow regimes. Introduction The practice of pumping waterfracs has spread to a wide range geographically. The results have been mixed in that the technique works better in some areas than in others. Waterfracs must be evaluated both on technical (scientific) and economical merit. For example, the best economic solution may not always be the best from a technical standpoint. The explanation of how the waterfrac treatments work has been limited to the theory of fracture conductivity created by surface asperities (mismatches in the geometry of the fracture faces upon closure). The small concentrations of proppant pumped could also concentrate at points of reduced fracture width creating, in effect, a wedge that is able to support and keep the fracture open to a certain degree depending on the magnitude of the in-situ stress and the properties of the proppant. Original publications related to the usage and results of waterfrac treatments pumped in the Cotton Valley sandstone formation in East Texas were published beginning in 19971–3. Since then methods to evaluate the success of the waterfrac treatments compared to conventional designs have been limited mainly to the direct comparison of well productivity. In many cases no adjustment was made for the varying conditions under which the wells were produced. Compensation for difference in flowing tubing pressure, initial reservoir pressure, differences in reservoir quality, etc. were generally not taken into account and left to the discretion of each individual looking at the comparison. The lack of accounting for these parameters is often due to the fact that the data is not readily available.