We investigate the effects of chemical fluid-rock disequilibrium and true triaxial stress states on axial permeability at confining pressures of the order 35 MPa at temperatures below 80C. Two suites of experiments are described, the first involving anelastic compaction and eventual failure at a constant strain rate. The decay in inferred hydraulic diffusivity here is found to be exponential, with a rate constant which depends both on stress state and the fluid chemistry, and which may change during deformation. In particular strain hardening deformation is found to be associated with more rapid sealing above a critical strain of the order 1% or so. Predicted asymptotic diffusion lengths are of the order a few metres for such deformation. In a separate suite of experiments, designed to monitor the continuous flow of fluids through stressed rock, a similar exponential decay was observed using a synthetic brine. These changes were associated with detectable changes in pore fluid chemistry. Fine particles (clays and detrital fragments) present in the natural core sample produced rapid sealing with predicted asymptotic diffusion lengths of the order a few hundred m to a few km, depending on the initial fluid-rock equilibrium conditions.
Many of the practical applications of rock mechanics in engineering problems involve a transient, man-made disturbance to the subsurface in addition to the pre-existing tectonic stress state. Examples are dam impoundment, mining activity and hydrocarbon production, all of which are known to induce earthquakes on a variety of scales. For example even relatively small stress perturbations due to the drawdown of fluid pressure on the production of hydrocarbons, of the order 1 MPa, can initiate small earthquakes over a timescale of several years (Segall, 1989; Grasso & Feigner, 1990). This paper investigates the response of porous rocks to rapid perturbations in stress state and fluid composition during such transients. Two experimental procedures are described involving (a) rapid compaction and faulting under at a constant, servo-controlled strain rate of 10-5 s-1, and (b) constant flow tests where the sample is kept under a constant axial and confining pressure.
Before describing the results, it is important to summarise the relevant capabilities of the two new experimental rigs used to carry out the individual tests.
Most experiments carried out on rock samples which involve a differential stress use cylindrical core samples with an axisymmetric radial confining pressure and an increasing axial stress e.g. the classic 'Hoek' cell. In this 'triaxial' configuration, there are three principal stresses (s1, s2, s3) but two are equal (s1> s2=s3). In the Earth, however, there are in general three independent principle stresses (s1>s2>s3). We shall refer to this configuration as 'true triaxial' stressing. We use a deformation rig designed by Smart (1995) to apply a radially-asymmetric confining pressure by the use of 24 trapped tubes situated between the cylinder sample walls and a steel pressure vessel. By pressurising these tubes to different amounts, a 'true triaxial' stress field can be delivered to the sample.
