The evolution of the effective-pressure dependence of hydraulic properties of sandstone during deformation

The dependence of rock physical properties on pressure is commonly described using the concept of effective pressure, p_eff= p_c-α p_f. In this law, the effective-pressure coefficient α , which quantifies the relative influence of confining pressure p_c and fluid pressure p_f on the property may deviate from unity and take different values for different physical properties. Although some microphysical models describing the effective-pressure dependence of physical properties exist, in particular the effect of fracture damage on the dependence of properties such as permeability on effective pressure remains to be understood. We conducted experiments on samples of Wilkeson sandstone, a clay-bearing (c. 8 %) sandstone with a porosity of c. 9 %. Cylindrical samples were deformed in a conventional triaxial apparatus with pore-fluid factors λ = p_f/p_c of 0.05-0.55 and effective pressures of 60-120 MPa, thus covering the brittle and ductile deformation regimes. During deformation, hydraulic properties were continuously measured using the pore-pressure oscillation technique. For four samples, transport properties were additionally investigated at a range of hydrostatic pressures prior and subsequently to deformation. Experiments showed that strength and failure mechanism of the samples were purely controlled by the effective pressure. The effective pressure coefficient for permeability, κ , as well as the permeability modulus, K_k≡ -∂ Δ ( p_c-κ p_f ) / ∂ ln k |_{pf, obtained by fitting the experimental data with an exponential law, showed a remarkable dependence on the state of deformation. For intact samples, κ >1 is consistent with a linear poroelastic model for heterogeneous rocks. Furthermore, the permeability decreased with increasing pressure, indicative of the elastic closure of microfractures. Around peak stress, κ <1 and Kk<0, i.e. permeability increased with pressure. An extension of microphysical models can explain these observations, which have important implications, e.g., for geotechnical applications or for understanding the evolution of fluid flow in fault zones during the seismic cycle.}