Search

Argon gas and liquid (dodecane) permeability measurements in two mixed quality siliceous mudstone samples within the Wolfcamp formation demonstrate that it is possible to close multiple bedding parallel open artificial micro-fractures and obtain representative matrix permeability by applying two confining stress cycles at a constant pore pressure under effective stresses ranging from 6.9 MPa to 59.7 MPa. The clearly micro-fractured sample exhibited a three-order decrease in permeability from 4.4 × 10−17 m2 to 2.1 × 10−20 m2. In contrast, the relatively intact sample exhibited initial liquid permeability of 1.6 × 10−19 m2 that declined gradually with stress to 3.0 × 10−20 m2. In this study, we developed a new permeability testing protocol and analytical approaches to interpret the evolution of micro-fractures and estimate the matrix permeability based on initial pulse-decay gas permeability measurements. The insights gained are useful to identify pore-scale controls on fluid flow behavior using SEM petrography and pore-space characterization and validate fundamental fluid flow mechanisms examined using pore-scale numerical models.

We measured argon gas permeability in three intact and one partially fractured Eagle Ford Shale samples documenting the stress dependence of horizontal (bedding parallel) in situ permeability of intact samples which varies between 1 and 10 nD (1 nD = 0.9869233 × 10−21 m2), while the permeability of partially fractured sample varies between 18 and 37 nD. For all samples, permeability decreases by up to an order of magnitude while cycling the confining pressure (PC) between 27.7 and 55.2 MPa at a constant pore pressure (PP) of 14.4 MPa. Most of the permeability decrease is within the first loading and unloading cycle. During this first cycle, we also observe less than 2% decline in permeability over ~ 10 days when we held the PC constant at 51.6–55.2 MPa, respectively. This suggests that the ongoing creep plays a relatively minor role. The subsequent PC cycles result in a small decrease in permeability (~ 6 to 26% variation between the start and the end of each cycle). We interpret that the initial permeability loss is due to the closing of micro-fractures—which we infer are caused by stress relief and gas expansion during sample retrieval and/or preparation. We interpret that the higher permeability of the partially fractured sample is mainly due to incomplete closure of a preexisting fracture, which extends nearly two-third the sample length. We document this dual-permeability structure from the observation of a dual-timescale pressure response behavior during the experiments at lower PC–PP. We find permeability decreases with increasing PC–PP; stress dependency of permeability follows an exponential relationship with a stress-sensitive gradient of 0.019–0.040 MPa−1. A better understanding of permeability variation with stress will help to reliably estimate in situ permeability and to better understand production evolution from unconventional shale reservoirs.

We examined the permeability behavior of two Marcellus Shale samples and two Eagle Ford Shale samples using argon gas under multiple confining stresses (PC = 13.8–68.9 MPa) and pore pressures (PP = 1.6–28.2 MPa). We corrected our measured permeability values for the Klinkenberg effect and contoured permeability as a function of PP and PC to obtain the effective stress coefficient for permeability (nk), which scales the influence of PP on permeability relative to PC. We explored the underlying mechanisms controlling nk using two idealized models of a cylindrical capillary tube consisting of low and high compressibility material to model the unique pore structure present in our samples. Finally, we investigated the impact of PP reduction on permeability during production. We show that the permeability is stress-dependent and that nk is formation-dependent. Marcellus Shale samples, with predominantly organic matter hosted pores, have a Klinkenberg-corrected permeability of 11.0 to 1.9 nD (PC - PP = 5.3–33.8 MPa) and nk value of 1.11 and 1.60, respectively. In contrast, Eagle Ford Shale samples, with predominantly inter-particle pores, have Klinkenberg-corrected permeability of 33.6 to 2.0 nD (PC - PP = 13.3–40.8 MPa) and nk value of 0.96 and 0.84, respectively. We infer that the PP change produces considerable pore wall strain relative to the PC change in the Marcellus Shale organic matter pores resulting in nk > 1. In contrast, we infer a smaller pore wall strain for the same PP and PC changes in the Eagle Ford Shale inter-particle mineral pores giving in nk

This paper describes the results of triaxial compression tests carried out at effective cell pressures ranging from 12.5 to 100 kPa to investigate the influence of fabric structure on the yield and failure of intact Reigate silver sand. In some of the tests, a digital image-based technique was used to determine the instant of onset of strain localisation, and the distribution of strain localisations within the specimen as overall deformation progressed. Comparative tests on intact and reconstituted specimens showed that fabric structure in the intact material allows the mobilisation of stress ratios close to peak before the onset of dilation, and increases the shear modulus at a given effective cell pressure and strain. Localisation was found to start at or after the onset of dilation, with a tendency to delay at increasing effective cell pressure. More localised deformation was observed at low effective cell pressures. Consistency between the critical state strengths of intact and reconstituted specimens is demonstrated, provided that the effect of shear band geometry is taken into account in stress analysis.

Strain localization developing inside soft rock specimens is examined through experimental observation and numerical simulation. In the experimental study, soft rock specimens are sheared at different strain rates under plane strain conditions and deformation and strain localization characteristics are analysed. Transition of localization mode from highly localized mode for higher strain rate to distributed and diffused mode of strain localization for lower strain rates was observed. In the numerical study, simulations of plane strain compression tests are carried out at different strain rates by using an overstressed-type elasto-viscoplastic model in finite element computations. The role of strain rates on setting gradients of strain fields across shear band is clarified. The probable mechanism for transition of localization mode is discussed. Copyright © 2005 John Wiley & Sons, Ltd.