Handling flow through fractured media is critical for transient pressure and flow analysis in shale gas reservoirs, because gas production from such low-permeability formations relies on fractures, from hydraulic fractures and fracture network to various-scaled natural fractures, to provide flow channels for gas flow into producing wells. This study presents a numerical investigation of pressure and flow transient analysis of gas production from a horizontal, multi-staged well in shale gas reservoirs. A specialized three-dimensional, two-phase simulator is developed and used for this purpose, which incorporates known nonlinear flow behavior in shale gas reservoirs. First we discuss a multi-domain, multi-continuum concept for handling multi-scaled heterogeneity and fractures, i.e., using hybrid modeling approaches to describe different types and scales of fractures from explicit modeling of hydraulic fractures and fracture network in simulated reservoir volume (SRV) to distributed natural fractures, microfractures, and tight matrix. Then sensitivity studies of transient pressure responses and flow rates are presented with respect to hydraulic fractures geometry, stimulated reservoir volume (SRV), and natural fracture density. We will also compare the behaviors with two different interporosity flow assumptions, fully-transient and quasisteady state flow. Unlike conventional reservoirs, their difference cannot be ignored due to the extremely low shale matrix permeability and significant gas compressibility. Specifically, we will analyze a field example from Barnett shale to demonstrate the use of results and methodology of this study.

Although existing models for analysing single-phase flowback water production at the onset of flowback in tight oil and gas reservoirs provide estimates of fracture volume, they are not applicable to shale gas reservoirs. This is because flowback data from shale gas wells do not show this single phase region. Instead, they show a surprising trend of immediate gas breakthrough. This paper attempts to (1) understand the fundamental reasons for this early gas breakthrough, (2) develop a representative mathematical model that describes this behaviour and (3) estimate the effective fracture volume and equivalent fracture half-length by history matching the early-time two-phase flowback data. From the diagnostic plots generated from of rate/pressure data of 8 multi-fractured horizontal wells completed in the Muskwa Formation, the gas-water ratio (GWR) plots indicate the presence of initial free gas in the complex fracture network. This conclusion is backed by the imbibition experiments conducted on shale samples collected from the same formation showing the presence of gas-saturated natural fractures. The linear diffusivity equation is solved for early-time two-phase gas/water flow in the hydraulic fractures. The primary drive mechanism at the onset of flowback is initial free gas expansion within the fracture network. Secondary drive mechanisms considered include fracture water expansion and fracture closure. The driving forces are modeled by an effective compressibility term analogous to the total compressibility in conventional multiphase flow formulations. Also, two-phase water/gas flow is handled by an explicitly determined relative permeability function of time. Eventually a new pseudo-time function is defined to account for the changes in gas properties and relative permeability with time. Rate normalized pseudo-pressure versus pseudo-time plots give a straight line when applied to field data, thus the solution can be used to characterize hydraulic fractures in a manner similar to conventional well testing methods.

Hydraulic fracturing combined with horizontal drilling has been the technology that makes it possible to economically produce natural gas from unconventional shale gas reservoirs. For best performance of fracturing stimulation, hydraulic fracturing designing parameters, such as different proppants, fracturing liquids, and injection rate and pressure, should be evaluated before the oeration in a particular reservoir system. Traditional evaluation and optimization approaches are usually based on the simulated fracture properties along, such as fracture areas. In our opinion, the better methodologyshould be also related to production data from the stimulated wells afterwards, because the enhanced production is the ultimate goal. In this paper, we present a general fully-coupled numerical framework to simulate the hydraulic induced fracture propagation and post-fracture gas well performance. This three-dimensional, multi-phase simulator focuses on: (1) fracture width increase and fracture propagation that occurs as slurry is injected into the fracture, (2) erosion caused by fracture fluid and leakoff, (3) proppant subsidence and flowback, and (4) multi-phase fluid flow through various-scaled anisotropic natural and man-made fractures. Mathematical and numerical details on how to fully couple the fracture propagation part and the fluid flow part are discussed. Fracturing and production operation parameters, properties of the formation, and reservoir, fluid, and proppant properties are all taken into account in this model. The well may be horizontal, vertical, or deviated, as well as open-hole or cemented. This simulator is verified based on benchmarks the literature. We show its application by simulating the fracture network (hydraulic and natural fractures) propagation and production data history matching of field production data in China. We conduct a series of real-data modeling studies with different combinations of fracturing parameters and present the methodology to design fracturing operations with feedback of simulated production data. This unified model aids in the optimization of hydraulic fracturing design, operations, and production.

The Haynesville Shale is one of the largest unconventional gas fields in the US. It is also one of the deepest, with wells reaching more than 10,000 feet below ground. This uncommon depth and overpressure leads to initial reservoir pressures of up to 12,000 psi. These pressures are uniquely high among shale gas reservoirs, and require special treatment. We show that the methods developed by Patzek, et al., PNAS, 110, 19731-19736, can scale all individual wells' gas productions to fit upon a single, universal curve. Haynesville wells can take months or years for flowing tubing pressure to stabilize, so we modified the type curves to take this into account. Furthermore, we introduce a PVT solver in order to calculate gas properties at up to reservoir pressures. When we apply the scaling theory to 2,199 individual well productions in the Haynesville, we find 1,580 wells which have entered exponential decline due to pressure interference. We use a simple physical model to determine the time to interference for wells with geologic parameters typical of the Haynesville, and use this to determine a field-wide permeability from the wells where we have observed pressure interference. From this, we arrive at time to interference for the remainder of Haynesville wells, and use this to arrive at production forecasts for all individual wells.

Liquid-rich shale (LRS) gas reservoirs have gained increasing attention in recent years. The Eagleford Shale in the U.S. and the Duvernay Shale in Canada are examples of liquid-rich shale gas plays which are being exploited to produce more profitable liquid hydrocarbons with natural gas. These reservoirs may store liquid hydrocarbons in liquid or vapor state, as with gas condensate systems. Further, shale gas reservoirs may contain significant volumes of organic matter, which stores gas (and liquid hydrocarbons) in the adsorbed state. There have been several historical simulation studies investigating the impact of various reservoir and fluid properties on fluid production and recovery from LRS, however none have investigated the importance of desorption. Adsorption has previously been suggested to be an important storage mechanism in organic-rich shales, particularly for heavy hydrocarbon fractions. Matrix pore configuration and associated connectivity has also been previously demonstrated to be an important control on gas production from shale gas reservoirs, but the impact on condensate production has not been investigated for LRS. In this work, the impact of heavy hydrocarbon fraction desorption, pore configuration and connectivity, fluid composition and operating conditions (flowing bottomhole pressure) on LRS production is investigated using a commercial simulator. We use PVT data from previous studies to develop a compositional simulation model analog of an LRS reservoir. Hydrocarbon component adsorption amounts are modeled using the Langmuir isotherm. Different combinations of organic and inorganic matter and fracture porosity, and their connectivity, are also assigned in the simulation cases. Simulation sensitivities demonstrate that desorption can contribute significantly to condensate production, depending on fluid composition and pore connectivity. In cases where liquid-fraction adsorption contributes significantly to in-place volume, we recommend injection of light end gases to mitigate condensate blockage. This will be studied in detail in future work.

Unconventional shale reservoirs possess low matrix permeability and cannot produce economically without effective placement of horizontal laterals and effective hydraulic stimulation.It is critical to recognize brittle units amenable to hydraulic fracturing and rich in organic matter. Shale mineralogy affects not only the effectiveness of stimulation, but also the ratio of free to absorbed gas. Heterogeneity in shale reservoirs expressed by mineral composition, organic matter richness and brittleness significantly influences shale-gas production. Although providing the necessary information, core data and pulsed neutron spectroscopy (PNS) logs are limited by cost considerations to an insufficient number of wells in order to sufficiently characterize reservoir heterogeneity. We present an approach of petrophysical analysis and statistical reverse modeling to better determine mineral concentration, organic matter percentage, and volume of gas and water using conventional logs in the Marcellus Shale (Appalachian basin, USA). Core data and PNS logs are utilized to guide and demonstrate the results. The statistically based reverse model incorporates random errors of logging, and solves mineral composition through matching the or iginal logs with mineralogy models. The effectiveness of statistical reverse models is enhanced through removing the effects of minerals with significant petrophysical response (e.g., barite) and fluids. Barite strongly affects the mineralogical estimates in shale. The concentration of such minerals can be predicted by the difference between PE curve and a trend line. Fluid volumes are predicted by a modified shaly-sand model with uranium concentration through understanding the effect of organic matter. The methodology proposed provides an approach to predict mineral composition and organic richness and the corresponding mineralogical properties (e.g., geomechanical and free-to-absorbed gas ratio). The proposed approach aids in recognizing geologic and engineering targets, designing horizontal well trajectories, targeting fracturing strategies and understanding shale depositional environments and processes. The proposed approach can be extended to other organic -rich shale reservoirs.

The Eagle Ford Shale is recognized as the largest oil and gas development in the world, based on capital investment ( Wood Mackenzie 2012 ). Development typically consists of horizontal wells stimulated with multiple hydraulic fracture stages. Almost $30 billion will be spent developing the play in 2013, and optimizing the completion design and spacing of these wells can result in large rewards for the companies involved. This paper presents a pragmatic integrated workflow, used to optimize development and guide critical development decisions in the Black Hawk field, Eagle Ford Shale. Geoscientists, reservoir, and completion engineers worked collaboratively to identify the optimal completion designs and well spacings for development focus areas. Multiple simplistic simulation models were history matched to existing production wells. Wide uncertainty exists in many key reservoir and completion parameters. Using stochastic realizations from ranges of key properties, uncertainty was reduced using the history matching process. The resulting calibrated reservoir scenarios formed the basis of optimization studies for development drilling and down spacing. Completion design parameters, including fracture stage length, perforation clusters per stage and landing point for the lateral, were evaluated in hydraulic fracture models. The resulting fracture geometries were simulated and the optimum completion design, and well spacing determined for each area. The optimal development was shown to vary by region, due to changing reservoir, fluid and geomechanical properties. The use of multiple subsurface realizations, spanning an appropriate range of uncertainty, was critical to the success of this study. Economic analysis across a range of potential outcomes enabled robust development decisions to be made. As a result of this work, field trials to test proposed changes to the completion have been initiated, and development drilling plans updated to reflect the optimal well spacing for each lease.

The Eagle Ford shale is the shale with the greatest expectations in Mexico, given its abundant hydrocarbon liquids production and success in the US. Resource estimates were first published by the EIA and PEMEX in 2011 , and new production and geologic information has been obtained in the area since then. The objective of this paper is to provide an important updated assessment of oil and gas reserves, as well as contingent and prospective resources, in the Mexican Eagle Ford shale and to quantify the related uncertainties. Given the considerable uncertainties in shale reservoirs, a probabilistic decline-curve approach was chosen, as it allows performing a play-wide assessment with available production data, and it quantifies the uncertainty in the resource size. Analog areas in the US Eagle Ford shale were defined based on available geologic information in both the US and Mexico. The Duong model coupled with a Markov Chain Monte Carlo (MCMC) methodology was used to analyze and forecast production of wells located in the defined analog sectors in the US Eagle Ford shale. By combining the results of individual-well analyses, a type curve and estimated ultimate recovery (EUR) distribution was obtained for each of the defined analog sectors. These distributions were combined with well-spacing assumptions and sector areas to generate the prospective-resources estimates. Similar methodology was used to estimate the reserves and contingent resources associated with the few existing wells. As of March 2013, the total prospective resources (P90-P50-P10) for the Eagle Ford shale in Mexico are estimated to be 527-1,139-7,268 MMSTB of oil and 17-37-217 TSCF of gas. Estimated reserves associated with existing wells and corresponding offset well locations are 18-35-60 BSCF for gas and zero for oil. Estimated contingent resources are 14-64-228 MSTB of oil and 8.5-13-26 BSCF of gas. For comparison, total resources (including reserves) for the US side are estimated to be 2,950-8,250-23,530 MMSTB and 34-91-266 TSCF, as reported by Gong et al. (2013) . The results of the work presented here should provide a more reliable assessment of the resource size and uncertainty in the Mexican Eagle Ford shale for use in resource exploitation planning.

In order to understand the effect of injected stimulation fluids on nano-darcy permeability, naturally fractured shale reservoirs, an integrated study of spontaneous imbibition has been performed. In this study, oil recovery during spontaneous imbibition in naturally fractured shale samples was improved using different water formulations. Different water solutions were formulated by adding different amounts of HCl and NaOH to either distilled water or 2 wt% KCl-base brine solution. Eight water formulations, distilled water, 2% KCl brine, low pH HCl solutions (0.74-1.2), and high pH alkaline solutions (11.7-12.4) were examined to recover oil from shale rocks. Outcrop samples from the Barnett shale formation were used in this study. The samples were 2.54 to 3.81 cm in diameter and 0.762 to 5.08 cm in length. Firstly, we studied the average porosity of the used samples using CT Scanning. The average porosity was around 7%. Secondly, we studied the rock stability and spontaneous imbibtion of the different Barnett samples in distilled water, 2 wt% KCl, low pH solutions (0.74- 1.2), and high pH alkaline solutions (11.7-12.4). During the spontaneous imbibition, the maximum oil recovery was 24% using distilled water and low pH solution of pH0.74 (3 wt% HCl in 2 wt% KCl base brine solution). Clay swelling can contribute to the Barnett oil recovery improvement by creating along bedding cracks. Oil recoveries from the Barnett shale can be improved by 56% when used 2 wt% NaOH solution and 88% when used 3 wt% HCl when added to 2 wt% KCl simulation base fluid. High pH solutions did not improve oil recoveries from the Barnett samples when added to distilled water due to pore plugging resulted from the combination of clay swelling with mineral dissolution. Low pH solutions showed along bedding cracks similar to distilled water effect on the Barnett shale samples that helped recovery accelerations, while higher pH solutions showed sever samples damage with higher alkaline concentrations (2 wt% NaOH).

Since 2008, the Eagle Ford Shale has been one of the most active U.S. oil and gas plays. Regional variations of the frequency (cyclicity) and thickness of organic-rich marl and limestone interbeds influence well completion design, and these variations may be related to well production performance. To better understand the impacts of lithologic cyclicity and reservoir and fluid properties on well performance, we conducted an integrated, regional study using well logs, production data, and PVT reports. Regionally, the Eagle Ford Shale is composed of 3 units. The Lower Eagle Ford consists of cyclic, interbedded organic-rich marl and limestone. To analyze lithologic cyclicity of the Lower Eagle Ford Shale in South Texas, We evaluated gamma ray logs from more than 500 vertical wells. A Matlab script was developed to count the numbers of organic-rich marl and limestone interbeds. Average bed thickness was calculated at each well and was mapped regionally. The numbers of both limestone and organic-rich marl interbeds increase from less than 2 in the northwest to more than 20 on the southeast. Lithologic cyclicity is greatest in La Salle and Karnes Counties, which are the most productive gas and oil regions, respectively. Eagle Ford Shale reservoir and fluid properties from PVT reports were mapped to further evaluate regional variations of fluid types and well performance. Reservoir pressure, pressure gradient, oil gravity, and gas specific gravity were mapped. Reservoir pressure and pressure gradient are greatest on the northeast. Oil gravity and gas specific gravity maps suggest increasing thermal maturity from northwest to southeast, with increasing depth. Well production rates are strongly influenced by lithologic cyclicity, pressure, and fluid properties. Understanding these relations in the lower Eagle Ford Shale should assist with optimizing completion design and stimulation strategies. The results of this study provide a better understanding of Eagle Ford Shale reservoir characteristic and well performance, and the study approach may be used to assess productivity of other shale plays.

Transport properties and mechanisms as well as phase behavior under nanoscale confinement exhibit significant deviations from their bulk behavior. This is due to the significant effect of molecule-wall interactions as well as molecule-molecule interactions in shale formations which are mainly characterized by nanopores. Consequently, production from shale gas reservoirs is strongly influenced by pore sizes and their effects on phase behavior and transport properties. In this study, we focus on analyzing and determining the effect of phase behavior and transport properties change due to pore proximity on production from a shale gas condensate reservoir. Additionally, the effect of different connectivities between pore sizes on production is studied. The effect of pore size on phase behavior is considered by using modified critical properties for different pore sizes in the phase behavior calculations. A shale gas condensate reservoir with a ternary mixture of methane (80 mol%), n-butane (10 mol%), and n-octane (10 mol%) as the reservoir fluid is modeled. The reservoir pressure and temperature are 5000 psia and 180 °F, respectively. The dew point pressure is 3600 psia. Pore sizes change between 5-150nm. Based on Scanning Electron Microscopy (SEM) studies on shale reservoir rocks, the pore volume of the reservoir was divided into five regions: bulk (stimulated area and pore sizes more than 50nm (17% PV)), 20-50nm (4% of PV), 15-20nm (6% of PV), 10-15nm (45% of PV), and less than 10nm (28% of PV). Three different types of connectivities between pores were considered: 1- completely random distribution 2- pore sizes from smallest to largest connected to the SRV in series, and 3- pore sizes from largest to smallest connected to the SRV in series. Our study has shown that by decreasing the pore size, dew point pressures decrease between 5 to 17%. Also by decreasing pore size, two-phase region shrinks therefore condensate drop-out and near wellbore permeability impairment are reduced. After 10 years of production, condensate saturation around SRV is 6-10% less under confinement effects. Gas and condensate viscosities under confinement decrease 3-16% and 10-45% respectively. Considering effect of confinement did not affect gas production significantly but the liquid production increased significantly and doubled. The effect of different pore size connectivities caused a 20% change in liquid productions. The results of this study can have a significant impact on our understanding of gas condensation and transport in shale formations thereby enabling improved field planning, well placement, completions design and facilities management.

In order to understand the performance of waterflooding in nano-darcy permeability, naturally fractured shale reservoirs, an integrated study of spontaneous imbibition has been performed. In this study, oil recovery during spontaneous imbibition in naturally fractured shale samples was improved using different water formulations. Different water solutions were formulated by adding different amounts of HCl and NaOH to either distilled water or 2 wt% KCl-base brine solution. Eight water formulations, distilled water, 2% KCl brine, low pH HCl solutions (0.74-1.2), and high pH alkaline solutions (11.7-12.4) were examined to recover oil from shale rocks. Reservoir core samples from Eagle Ford shale were used in this study. The samples were 2.54 to 3.81 cm in diameter and 0.762 to 5.08 cm in length. Firstly, we studied the porosity of the used samples using CT Scanning. The average porosity was 1.6% for Eagle Ford shale samples. Secondly, we studied the rock stability and spontaneous imbibtion of the different Eagle Ford samples in distilled water, 2 wt% KCl, low pH solutions (0.74-1.2), and high pH alkaline solutions (11.7-12.4). During the spontaneous imbibition, the maximum oil recovery was 37% for Eagle Ford using pH 0.74 (3 wt% HCl in 2 wt% KCl) and 44% using pH 11.7 (0.1 wt% NaOH in distilled water). The oil recovery was improved by mineral dissolution using low and high pH solutions along with wettability alteration. The rock hardness was significantly affected by using both high and low pH solutions, which resulted in 93-98% loss of its initial value.

A large amount of hydrocarbon fluids in shale formations are stored within the organic matters where the pore sizes are in the order of nanometer scales. Inside these nanopores, the interactions between the fluid molecules and porous walls play such an important role that can change the phase behavior as well as transport mechanisms of the hydrocarbon fluids. For a shale gas reservoir, the natural gas in the reservoir is usually stored in two forms, free gas and adsorbed gas. The region where free gas is stored has negligible fluid-wall interactions while the region for adsorbed gas is under strong pore wall influence. The current available equations of state cannot capture the phase behavior of the adsorbed gas phase due to the ignorance of the fluid-wall interactions. This work focuses on modifying the Peng-Robinson equation of state (PR-EOS) using the Simplified Local-Density (SLD) theory. From the modified PR-EOS, the fluid density at any arbitrary position inside the pore can be calculated using the local density approximation. A density profile for any particular hydrocarbon fluids can be obtained by calculating the local densities of the fluids at each discretized interval along the pore. From the density profile one can distinguish the regions of adsorbed phase, transition phase and bulk phase of the fluids. The thickness and averaged fluid densities for each phase can also be obtained from the model. Once the thickness of the absorbed phase is known, it is possible to determine whether adsorption is a single layer or multilayer. Our preliminary results show that depending on fluid types, either a single layer or multilayer adsorption is presented in those nanometer pores near the pore wall. The pore size range we focused on was from 100 nm to 1 nm. Methane and n-Butane were considered as fluids. When the pore size gets smaller and smaller, the absorbed layers at opposite pore walls can be merged together and result in the absence of the bulk fluid phase in the center areas of the pores. In this case, all the fluids in the pore are under influence of the wall. Our results also indicate that the fluid-wall interactions can have a much larger impacts on light components (methane) rather than heavy components (n-butane). That is, the density of the adsorbed phase of methane is more than two times the free gas density of methane (bulk density), while the n-butane adsorbed density is only slightly higher than its bulk density. The model has also been validated with molecular simulations for accuracy approval.

Liquid-rich shale plays (LRS) in North America have recently gained a lot of attention. Commercialization of these plays is now possible due to new technology, such as multi-fractured horizontal wells (MFHW). Along with such developments, there is an increased requirement to develop consistent reservoir-engineering methods to analyze multi-phase production data in such reservoirs. Large drawdowns required to produce these very low permeability formations complicate production analysis due to condensate drop-out near the wellbore or fracture face. Hydraulically-fractured vertical and horizontal wells completed in tight formations typically exhibit long periods of transient linear flow. This paper discusses a novel production data analysis technique for constant flowing bottomhole pressure (p wf < p dew ) wells producing from fractured LRS reservoirs. Our focus in this work is on cases where the initial reservoir pressure is well above the dew point pressure, as occurs in highly-undersaturated portions of the Eagle Ford Formation. A theoretical basis is developed for analysis of the transient linear flow period for these cases, and the effect of initial pressure on well performance is studied. The governing flow equation is linearized using appropriately defined two- phase pseudopressure and pseudotime functions, where the liquid solution analogy ca n be applied. This approach provides an accurate estimation of the linear flow parameter (x f √k) in multi-phase flow situations. Fine-grid compositional and black oil numerical models are used to validate the results. This work provides a robust analytical framework for production analysis of liquid-rich shale reservoirs as well as practical guidelines for real world applications.

Although there have been several efforts to quantify storage in shale nanopores, these have largely been based on generalization of the formulations for conventional reservoirs. Additionally, there is a lack of data addressing the effects of pore proximity on multicomponent adsorption and storage at a diverse set of pressures. Because it is nearly impossible with the currently available technologies to assess storage at the nano-scale, our work relies on the use of Molecular Dynamic simulation (to be called as MDS henceforth) techniques as well as a modified version of the Peng-Robinson EOS appropriate for modeling fluid behavior under pore proximity effects. We first describe the modified PR-EOS and demonstrate applications of pore confined methane phase behavior for different pore size distributions. For these chosen pore size distributions that are representative of organic nanopores, we derive an effective pore size that reproduces the composite phase behavior of the distribution of pore sizes. An effective pore size is defined because of the need to employ only one EOS for compositional modeling. Current efforts at modeling pore-confined phase behavior are largely restricted to tubes of a specified radius and may necessitate several fit-for-purpose EOS to model fluid behavior in different subsets of the pore size distribution. We demonstrate the need for careful examination of phase behavior when the pore volume contribution from the smallest of pores (sub-2nm) is substantial. However, our results indicate that for internmediate sized nanopores, an effective pore size representing the entire porous media may be derived. We then extend our modeling work to multicomponent systems and focus on the storage characteristics and phase behavior under confinement of a mixture of methane and octane. These results also indicate that when a substantial percentage of the pore volume is contained in the smallest of pores, the search for an effective pore size can become challenging. We then demonstrate some of the issues associated with fluid storage in organic nanopores by employing the graphene slit pore model. We model a replica of a connected pore system and demonstrate that pore proximity effects can substantially alter our expectations of storativity of methane, especially in the adsorbed layer. Finally, we demonstrate the need for moving beyond monolayer Langmuir adsorption models for describing storage by highlighting observations of multilayer adsorption of methane in organic pores. The key findings from this paper are as follows: Firstly, because the properties of alkanes differ with pore size, this study is the first to demonstrate that with complex pore connectivities, a simple extension of analyses from a single pore to connected pore systems is somewhat inadequate. This has implications for generating adsorption curves for reservoir simulation, to quantify fluids-in-place and to understand vapor-liquid equilibrium under the influence of pore proximity. We finally demonstrate that careful consideration of pore proximity effects in connected pore systems is necessary for a more meaningful quantification of reserves and predictions of well performance.

In multistage fracturing of unconventional formations such as the Eagle Ford shale, wells are traditionally stimulated by fracturing several perforation clusters simultaneously. While the technique is operationally efficient, there is evidence from production logs, microseismic monitoring, and other measurements that several clusters produce below expectations or do not produce at all. This condition is to a degree because stages penetrate zones with stress heterogeneities, and consequently, fractures propagate unevenly from all of the clusters. Evenly fracturing all clusters in heterogeneous zones is challenging in long horizontal sections penetrating heterogeneous reservoirs. Furthermore, efforts to improve well economics result in reducing completion time by extending the length of each stage even further to decrease the number of interventions required for completing the well. To address this challenge, a new sequenced fracturing technique has been developed based on a novel composite fluid comprising of degradable fibers and multi-sized particles that diverts the remaining stimulation fluids to understimulated regions of the wellbore. The composite fluid is delivered downhole at high-concentration, to create temporary plugs in clusters already stimulated thereby creating diversion with a minor amount of material. The solids degrade completely after the fracturing treatment has been completed, leaving no residual formation or fracture conductivity damage. The channel fracturing technique ( Gillard et al. 2010 ) was chosen as the preferred fracturing method for use in tandem with the composite fluid. This technique has been reported to increase effective fracture length while reducing risk of screenout with respect to other conventional methods. The new composite fluid was used in a campaign where its effects were monitored with microseismic instruments. A case study presents field experiments where wells from the same pad are fractured in a similar fashion with and without diversion. In one application, with similar water and fluid volumes, the well treated with this technique produced more than 15% per stage than its conventionally treated offset well. The signature of the composite fluid, clearly visible on all measurement techniques, has proven consistent across stages of various lengths and wells having different characteristics.

Tight oil and shale gas resources produced 29% of total US crude oil and 40% of total US natural gas in 2012 (EIA). Technically recoverable quantities of shale gas and shale oil resources for the US are 665 trillion cubic feet and 58 billion barrels, respectively. Considering the impact of the "unconventional boom" on the economy, it is crucial to understand the production performance of wells to maximize the recovery from shale plays. The latest advances in Rate Transient Analysis (RTA) provide quick yet robust tools to assess the quality of the Stimulated Rock Volume (SRV) and long term performance of the wells by estimating EURs. The most common challenge in history-matching of production in shale gas/oil wells has been the non-uniqueness of the history-matched parameters. A lot of emphasis has been put on estimation of fracture half-length, which is believed to be a primary driver for the performance of shale gas/oil wells. Since linear flow is the main transient flow regime in the early life of a hydraulically-fractured shale gas/oil well, a Rate Normalized Pressure (RNP) versus Square Root of Time plot is the most commonly used diagnostic plot for the performance analysis of the wells. A*sqrt(k) or x f *sqrt(k) parameter groups are reported as a proxy for productivity in hydraulically fractured shale/gas oil wells. Besides having permeability as an unknown, the history-match is also sensitive to net hydraulic fracture height, which is one of the inputs to models that must be specified from other sources of information. This paper presents a novel approach for production performance analysis of shale gas/oil wells, which significantly reduces the non-uniqueness issues that one can have in comparison of performance. Twenty two Eagle Ford Shale wells were analyzed across the trend from lean gas to high-yield condensate to define a workflow that could be applied to other wells in different geologic areas, yet provide consistent comparison of long term performance (EURs).

A practical simulation model is developed and demonstrated with applications for accurate characterization of production rate and pressure behavior with time in shale-gas reservoirs featuring horizontal wells intersected with multi-stages of hydraulic fractures. The model determines the contribution of each fracture stage to the overall production and predicts the pressure changes occurring in the fracture and matrix zones by a compartmental simulation approach. Various flow regimes of drawdown test are investigated for fracture diagnosis with different fracture permeability values to determine the external boundary effect. Our practical mathematical modeling, coupling the wellbore and reservoir hydraulics, is solved numerically by an iterative method to determine the flow rates coming into the horizontal well from multiple hydraulic fractures like the commingled layers intersected by a vertical well. This approach provides a reasonable description of behavior of multiple stages in shale gas reservoirs by considering the alteration of the gas transport properties and the changing apparent permeability under the effect of pore proximity of shale. Various applications of our non-Darcy simulation model demonstrate the importance of the corrections relevant for shale-gas reservoirs compared with the conventional Darcy flow without consideration of such corrections. Commensurate with field observations, when the non-Darcy flow corrections are considered, the cumulative gas production is higher compared with the conventional Darcy flow calculations which impact significantly on fracture and matrix pressure responses and production forecasting.

Shale gas production has been gaining worldwide attention over the past several years. This is due to the economic gas reserves using two current advanced technologies that are horizontal drilling and multistage hydraulic fracturing. Shale has a high total organic content (TOC) that may adsorb significant amount of natural gas. In addition, laboratory and theoretical calculations indicate that organic-rich shales adsorb CO 2 preferentially over CH 4 . Hence, the extent of organic matter in shale plays an important role in determining the feasibility of CO 2 injection with potential benefit of enhanced gas recovery (EGR). The performance of CO 2 injection and CH 4 recovery in shale reservoirs is a complex function of several engineering parameters including fracture half-length, fracture conductivity, and fracture height, operating parameters such as injection volume and injection time, and geologic parameters including reservoir permeability, porosity, and thickness. Nevertheless, the effects of the above uncertain parameters on the process of CO 2 -EGR are not clearly understood and systematically studied. Therefore, it is absolutely critical to quantify uncertainties and investigate the most important influential parameters controlling this process. In this paper, we employ numerical reservoir simulation techniques to model multiple hydraulic fractures and multi-component Langmuir isotherms. Two scenarios for CO 2 injection are investigated when the primary gas production decreases to the economic limit: (1) CO 2 flooding in two horizontal wells, and (2) CO 2 huff-n-puff in a horizontal well. A series of reservoir simulations based on Design of Experiment (DOE) are performed on the best scenario to investigate the critical parameters that control this CO 2 -EGR process in the Barnett Shale. This work enables operators to plan ahead of time and optimize a tertiary shale gas production process by considering the different investigated influential parameters.

Our method to characterize a stimulated reservoir is a two-step process. In the first step, we match a microseismic (MS) pattern using a geomechanics model, which gives injection permeability and porosity. In the second step we match gas rate versus time using Production Data Analysis, which gives a much lower production permeability and a reduced size of the SRV (stimulated reservoir volume). Data from five wells in the Fayetteville shale have been analyzed using new software. Two perm-based diagnostics (injection and production perm) have been correlated to various fracture treatment parameters, to try to characterize and to potentially improve fracture stimulations. The injection permeability and porosity can provide information on average fracture spacing and aperture width during fracture stimulation. One way to retain more of the injection permeability, and the size of the SRV, is to tailor the proppant to the width and spacing of the induced fracture network. The spreading of the proppant depends on the spacing and aperture width of the network fractures, as well as the diameter and density of the proppant. All these factors have been included in a conceptual model for proppant spreading away from a horizontal well. Results and conclusions of the five-well analysis are: Potential horizontal fracture components in three wells may act to reduce breadth and height of the MS cloud (i.e., less outward and height spread of fracture fluid). Proppant volume is a very small fraction (<4%) of fracture network volume. Created fractures are further apart and open wider in deeper wells. This reflects the degree of compaction of the formation. This depth trend is consistent with fracture spacing from MS events in other fields. Average fracture width in the network increases with depth, which suggests that increasing the proportion of 40-70 sand with depth is plausible. In two of the studied wells, it may have even been possible to use 30-50 sand. A regression equation enables a prediction of the SRV productivity. When effective stress increases, SRV productivity decreases. This is expected: it is harder to sustain a fracture network in a formation with larger effective stress (i.e. deeper wells). When amount of proppant (40-70 or 100-mesh) increases, SRV productivity increases. This is expected if proppant is needed to stop fractures in the network from closing due to in-situ stress. But an increase in 100-mesh proppant will be more beneficial as compared with 40-70 proppant. The MS matching provides new information on spacing and aperture width of fractures in a network. Using a model for proppant spreading built around this information should help operators optimize proppant transport in shale gas and oil wells, to retain a larger SRV, and greater permeability enhancement within the SRV (SRV sizes from transient analysis of gas rates are much less than MS volumes). Our fracture widths are generally consistent with proppant sizes used in shales (100-mesh and 40-70), and this supports the geomechanics model which determines the fracture widths by matching microseismic data.