Unconventional Resources
Understanding Water Controls on Shale Gas Mobilization into Fractures Last Reviewed June 2017

ESD14085

Goal
The overall project goal is to understand and predict the dynamics of fracturing fluid interactions with gas and oil within stimulated unconventional shale reservoirs, in order to identify manipulations of fluid properties that will significantly increase reservoir production. Laboratory tests and modeling studies will be performed to track fracturing fluid imbibition and gas/oil counterflow in shales from unconventional reservoirs and in micromodels of fracture-matrix systems. Responding to the need to reduce water use in developing unconventional reservoirs, this research will also include investigating the effectiveness of non-water based fracturing fluids This updated knowledge will provide a scientific basis for designing more effective hydraulic fracturing strategies that may ultimately maximize gas/oil production while utilizing significantly less water.

Performer
Lawrence Berkeley National Laboratory

Background
The enabling technologies of horizontal drilling and hydraulic fracturing have proven effective in recovery of natural gas trapped within shales, facilitating rapid development of shale gas and tight gas reservoirs in the U.S. over the past decade. Although major advances have been made in modeling large-scale gas production from unconventional shale formations, understanding matrix processes and their interactions with fractures underlying reservoir productivity remains qualitative and limits confidence in long-term gas production predictions. The common problems of formation damage and water-blocking cannot be reliably predicted, with some wells being highly productive despite little water recovery while others have low productivity despite good fracture fluid recovery reservoir heterogeneities over many scales, and complex nanopore structures have been investigated to a considerable extent, yet quantitative linkages to reservoir responses are unclear. The presence of water, both native and injected (as the main component in hydraulic fracturing fluids), greatly complicates gas transport over all scales of interest, from nanopores to reservoir scales. Despite the controlling role of water in shale gas production, surprisingly few direct measurements of the multiphase behavior of water and gas in unconventional reservoir rocks are currently available. Because of the scarcity of experimental data indicating relationships among water saturation, capillary (disjoining) pressure, and permeabilities (to both gas and water) in unconventional reservoirs, models still rely heavily on scaling-based extrapolations from more permeable media. Spatial variability of material properties, complexity of pore networks, kerogen distributions, and fracture distributions have become well-recognized factors affecting the flow of gas in tight reservoirs. While further investigation into complex systems and processes such as these will certainly lead to better predictions concerning unconventional shale reservoirs, note that water is a key factor in these issues as well. Moreover, unlike the inherent highly variable structural complexity of rock over multiple scales, water is a potentially controllable parameter in well field operations. However, control of water in unconventional reservoirs will first require a better understanding of the relationships between fracturing fluid composition and the rock matrix.

In Phase I of this project, an integrated experimental/modeling effort was undertaken to understand key processes in a sequence that closely follows production from shale gas reservoirs. Laboratory experiments and numerical simulations were conducted in Phase I to develop an understanding of native pore waters in partially saturated shale and determine how pore water saturation and unsaturated flow depend on adsorption and capillarity. For direct, molecular- level measurements, instruments at the LBNL Molecular Foundry were used for atomic force microscopy and X-ray photoemission spectroscopy. The water saturation dependent permeability, gas sorption isotherms, and gas permeabilities were generated for Woodford shale samples in the characterization stage, along with supporting spectroscopic studies needed to understand magnitudes of gas sorption. Through experimental and computational simulation, LBNL investigated the processes of (1) finite volumes of hydraulic fracturing fluid migrating into shale cores, with the imbibing water temporarily blocking shale gas from flowing into the fracture; (2) water redistribution and saturation decrease in the shale matrix adjacent to the fracture boundary; (3) percolation of previously trapped gas through desaturation of the water block; and (4) flow to the fracture face, recovery of gas permeability, and dissipation/evaporation of water. The performance of the laboratory tests at reservoir conditions and close integration between laboratory and modeling studies allowed for a more quantitative understanding of the dynamic processes controlling gas flow from shale into fractures.

The project was extended for a Phase II research effort on October 1, 2016, with two general objectives. The first is to further expand understanding of the coupling between water imbibition and gas counterflow in shales in order to help identify approaches to improving production. Under this objective, LBNL seeks to understand consequences of immiscible fluid displacement at fracture-shale matrix interfaces, within shale matrix blocks, and at larger scales of matrix blocks transected by fracture networks. Understanding gained in this aspect of the research will help identify combinations of fluid and interfacial properties of water-based fracturing fluids that support optimal gas/oil recovery. The second general objective is to understand the influence of non-water fracturing fluids on shale gas/oil mobilization, and to identify the optimal formulas of fracturing fluids for specific oil types and reservoir wettability. Integrated experimental and modeling approaches will be undertaken in order to achieve these objectives.

Impact
The laboratory tests conducted at reservoir conditions and close integration between laboratory and modeling studies in Phase I enabled a more quantitative understanding of the dynamic processes controlling gas flow from shale into fractures. Anticipated key results of Phase I included (a) determining how water block dissipation times depend on shale properties and amounts of water imbibition, (b) understanding the shale surface chemistry of wetting and reduction of gas permeability, and (c) determining the role of initial native pore water saturations in shales and in water block formation and dissipation. Phase II work will continue to gain additional insights needed to improve predictions and performance of fracturing fluid imbibition into shale from fracture surfaces and gas counterflow across the water block, and compare performance of non-water fracturing fluids with water-based fluids. This updated knowledge will provide a scientific basis for designing more effective hydraulic fracturing strategies that may ultimately maximize gas/oil production while utilizing significantly less water.

Accomplishments
In Phase I, experimental work was completed to characterize Woodford Shale samples from five separate sample locations. The laboratory measurements conducted include total organic carbon (TOC), total inorganic carbon (TIC), X-ray diffraction (XRD), scanning electron microscopy (SEM), water adsorption/desorption, specific surface areas, porosity, and permeability. TOC values of the five samples range from 2.7 to 7.1% on a mass basis. One sample location (WH1) exhibited high TIC (3.6%) which is consistent with XRD results that indicate the minerals quartz and illite are common to all samples while calcite was only identified in the high TIC sample, WH1.

Water adsorption and desorption isotherms were completed on each sample using two grain sizes (500-800 μm and 250-500 μm) over a range of relative humidities (0 to ≥ 96%).Grain densities of the Woodford Shale samples were obtained by water pycnometry on crushed samples, and bulk densities were determined directly from oven dry masses of cores and their associated bulk volumes. The values of grain densities and porosities (ranging from 0.065 to 0.010) were used to calculate water saturations of shales equilibrated at different relative humidity levels during adsorption/desorption isotherm tests. Adsorption/desorption results indicate that grain size does not significantly impact the wetting and drying behavior of the shale (Figure 1). However, pores finer than 3nm likely remain water-filled. Furthermore, results indicate the calcite-rich sample absorbed significantly less water than the other shales. Adsorption measurements were completed at 30 and 50º C with significant agreement between measurements (Figure 2). Desorption isotherms were only completed at 50º C to reduce the excessive time required for duplicative results. Desorption isotherms are showing significantly higher water contents relative to levels obtained by adsorption at the same relative humidity, indicative of water blocking (Figure 1).

Adsorption and desorption isotherms for Woodford shales at 50 °C. Adsorption and desorption paths are indicated with blue and red arrows, respectively.
Figure 1: Adsorption and desorption isotherms for Woodford shales at 50 °C. Adsorption and desorption paths are indicated with blue and red arrows, respectively. The “a” and “b” sample designations denote 500-800 μm and 250-500 μm grain-size ranges, respectively.

Water vapor adsorption isotherms on Woodford Shale samples, conducted at 30 and 50 °.
Figure 2: Water vapor adsorption isotherms on Woodford Shale samples, conducted at 30 and 50 °. P/Po is the relative humidity, controlled using saturated salt solutions. The “a” and “b” sample designations denote 500-800 μm and 250-500 μm grain-size ranges, respectively.

Specific surface areas of the shales were anticipated to be relatively high, given the significant adsorption of water vapor (Figure 1). The specific surface areas of crushed shale samples were obtained using several methods. With the exception of the calcite-rich WH1 sample, all surface areas obtained with N2 or Kr BET analyses are less than 1 m2/g. Because such low specific surface areas are unlikely for gas shales, and inconsistent with significant water vapor adsorption, water vapor adsorption data was used to directly to obtain alternate estimates. 

Crushed rock permeameter measurements were completed to quantify the permeabilities of the shale samples. To measure low permeabilities, a transient gas pressure pulse method was applied to rock grains sieved to a narrow size fraction such that samples can be treated as collections of equivalent uniform porous spheres. The developed permeameter system consists of two small pressure vessels, one containing the crushed shale sample and the other serving as the pressurizing gas reservoir, a high-pressure syringe pump (Isco) for delivering the test gas into the gas reservoir, and valves (Figure 3). In the system, the reservoir pressure is raised while isolated from the sample chamber, followed by opening of the pressurization valve and monitoring decay of the transient pressure increase in the sample chamber. Main components are contained within an incubator that controls experimental temperatures.

Figure 3: Crushed rock permeameters. a. Schematic of modified experimental system. b. Photograph of system within temperature-controlled incubator. c. example rock chips (2.8 – 3.3 mm). d. opened sample chamber.
Figure 3: Crushed rock permeameters. a. Schematic of modified experimental system. b. Photograph of system within temperature-controlled incubator. c. example rock chips (2.8 – 3.3 mm). d. opened sample chamber.

Alternate permeability measurements were obtained from cores using a core probe permeameter method and water imbibition experiments. The three permeability measurements result in a large range of permeabilities in which the crushed rock permeameter reflect lower permeabilities, the g imbibition permeabilities yield intermediate values, and the gas probe permeameter results in the highest permeabilities due to the influence of microfractures.

In Phase I, a pore-scale model was developed based on the many-body dissipative particle dynamics (MDPD) method. After investigating several modeling approaches, MDPD was selected because it is more fundamental and easily applied relative to other approaches in terms of representing interactions with different minerals and organic matter in rock space, including but not limited to slip flow, wettability, and adsorption/desorption processes. The MDPD code has been tested and verified against existing model results.

(a) Initial model setup for the simulation, black representing centers of solid/wall particles and green representing centers of MDPD fluid particles, and (b) comparison of the MDPD results with the existing model results (Navier-Stokes solution) for steady flow under no-flow boundary conditions at the channel walls.
Figure 4: (a) Initial model setup for the simulation, black representing centers of solid/wall particles and green representing centers of MDPD fluid particles, and (b) comparison of the MDPD results with the existing model results (Navier-Stokes solution) for steady flow under no-flow boundary conditions at the channel walls.

Phase I MDPD simulations for single-phase flow in a capillary channel almost perfectly match the Navier-Stocks solution (Figure 4). Pore-scale simulations have been expanded to represent two-phase flow in nano-tubes and nonporous media to represent spontaneous imbition of a wetting fluid into a nonwetting fluid filled nanotube. Furthermore, simulations were conducted to assess forced drainage and imbibition of fluids (Figure 5). The resulting simulations were found to agree with published data and preliminary results suggest that water may move into nano-scale channels more rapidly than in large-scale capillary tubes. Additional testing is underway to determine if this phenomenon occurs in porous media with different grain sizes and initial water saturations. Pore-scale simulations have also been completed to understand imbibition of multiple fluids in nanoporous media. Pore-scale simulations were used to understand imbibition and fluid entrapment as well as to calculate macroscopic parameters, which can be compared with experimental data and used in the macroscopic-scale simulations. Finally, an algorithm was developed to extract image data into the MDPD pore-scale model to develop realistic model domains for additional simulations.   

Figure 5: A simple example of forced imbibition of a wetting fluid into a nano-porous medium.
Figure 5: A simple example of forced imbibition of a wetting fluid into a nano-porous medium.

In Phase I, a macroscopic-scale model was also developed based on the Maxwell-Stefan diffusion model to represent the diffusion and adsorption/desorption in crushed shale rock samples. The results from these simulations will be used to test the models and understand the transport processes in shale rock while comparing against experimental results. Based on the interpretations from these models, both the permeability measurements and the adsorption/desorption laboratory results help to better evaluate pore-scale distribution and connectivity in shale rock matrix. Eventually, the macroscopic scale models will be used to evaluate imbibition, redistribution and methane flow into fracture faces.

The Phase II research effort began on October 1, 2016. In Phase II, laboratory work has focused on water vapor adsorption isotherms and imbibition and drainage capillary pressure-saturation measurements on Woodford Shale samples as well as initial water imbibition measurements on Marcells Shale cores. Work has also been conducted to design and fabricate fracture-matrix micromodels and a high-pressure laboratory foam generating system for use in non-water based fracturing fluid experiments.

Phase II modeling work is underway to (1) apply the pore-scale model for multiphase processes in micromodel experiments, (2) develop a constitutive model for adsorption/desorption in the shale matrix, (3) simulate diffusion and adsorption of water molecules in shale rock blocks to delineate major diffusive processes, and (4) analyze core imbibition experiments using a modified inverse modeling approach.

Current Status (June 2017)
Laboratory and numerical simulations have been completed to increase our understanding of water imbibition in shale. Refined laboratory methods have been established for further shale characterization work. Laboratory characterization of the Woodford Shale samples is complete. A MBPD model has been developed for simulating fluid and gas dynamics in nanoporous media. The Phase I project ended on March 31, 2016, and the Phase I final report is available below.

The project has been funded for a Phase II research effort to expand upon Phase I work in order to understand and predict the dynamics of fracturing fluid interactions with gas and oil within stimulated unconventional shale reservoirs from additional shale formations. Furthermore, Phase II work will investigate the influence of non-water fracturing fluids on shale gas/oil mobilization, and to identify the optimal formulas of fracturing fluids for specific oil types and reservoir wettability. The Phase II research effort began on October 1, 2016. The laboratory systems and numerical models have been modified and prepared for experimentation. Laboratory experiments continue to make progress on adsorption/desorption and imbibition measurements. Fracture-matrix micromodels and a high-pressure laboratory foam generating system are currently being developed for use in non-water based fracturing fluid experiments. Numerical models have been developed and modified for simulating adsorption/desorption, imbibition, and micromodel experiments. Numerical models are currently being tested and verified against laboratory results.

Project Start: October 1, 2014
Project End: June 30, 2018

DOE Contribution: $968,000
Performer Contribution: $0

Contact Information:
NETL – Stephen Henry (stephen.henry@netl.doe.gov or 304-285-2083)
Lawrence Berkeley National Laboratory - Tetsu Tokunaga, (TKTokunaga@lbl.gov or 510-486-7176)

Additional Information:

Understanding Water Controls on Shale Gas Mobilization into Fractures (Aug 2017) [PDF]
Presented by Tetsu Tokunaga, Lawrence Berkeley National Laboratory, 2017 Carbon Storage and Oil and Natural Gas Technologies Review Meeting, Pittsburgh, PA

Water Interactions with Shales, and Impacts on Gas Mobilization into Fractures (Aug 2017) [PDF]
Presented by Tetsu Tokunaga, Lawrence Berkeley National Laboratory, 2017 Carbon Storage and Oil and Natural Gas Technologies Review Meeting, Pittsburgh, PA

Phase 1 Final Report (July 2016) [PDF]