A cumulative damage model was formulated to quantify the stress-induced permeability reduction during the shear-enhanced compaction. Using this model, we are able to link the transient permeability anisotropy during cataclastic flow to the influence of stress state and loading path on the yield stress and microscopic stress heterogeneities. A general good agreement between the model predictions and experimental data implies that the cumulative damage model captures the key micromechanical processes operating during cataclastic flow. Because permeability and porosity are closely related, change of permeability as a function of porosity is desired. However, due to the variability from sample to sample, it is often difficult to formulate a robust porosity-stress relationship using conventional triaxial tests. Recently, we developed a new testing method to characterize the iso-porosity stress contours in the stress space. In a so-called modified undrained test, a constant pore pressure is maintained throughout the loading through simultaneous manipulation of the confining pressure and axial stress. The stress paths from such a test map out iso-porosity stress contours that would coincide with the yield caps, if the elastic strain is negligible. For highly porous sandstones, the yield caps correlate well with the iso-porosity contours obtained by this novel approach. Combination of the probabilistic damage model and the iso-porosity data provides new insight into permeability evolution as a function of porosity change in porous sandstones during compactive yielding.
Fluid percolation is important in virtually all crustal processes. Better understanding of permeability-porosity relationship in crustal rocks is essential for many geological problems. Laboratory studies under controlled pressure and temperature are useful in isolating different processes and in investigating interrelationships among pore structure, permeability and stress. Within the geological pressure and temperature range of the seismogenic zone (i.e., the upper 10-15 kilometers), rocks often fail by shear localization or by cataclastic flow. The latter is a special semibrittle failure mode in which the distributed macroscopic deformation is associated with pervasive microcracking . While the macroscopic manifestation of brittle faulting and cataclastic flow failure modes are drastically different (i.e., localized versus distributed deformation), the micromechanics of both regimes involves stress-induced microcracking at the grain scale . Recent laboratory studies have shown that permeability can be significantly modified under hydrostatic or non-hydrostatic stresses. While an increase in mean stress generally results in porosity reduction and permeability decrease, deviatoric stress causes the pore space to dilate or compact, and consequently its effect on permeability can be drastically different, dependent on rock type, initial porosity, confinement, and failure mode .
2. CUMULATIVE DAMAGE MODEL AND ISO-PERMEABILITY STRESS CONTOURS
In porous siliciclastic rocks, permeability evolution as a function of stress during cataclastic flow can be described in three different stages. Before the deviatoric stress reaches a critical stress state C*, which marks the onset of shear-enhanced compaction , permeability and porosity reduction are controlled mainly by the effective mean stress. At C*, the deviatoric stress exerts primary control over permeability and porosity evolution. The increase in deviatoric stress results in drastic permeability and porosity reduction and considerable permeability anisotropy.