Acoustic emissions and conventional strain measurements were used to follow the evolution of the damage surface and plastic potential in a limestone under triaxial compression. Confining pressures were chosen such that macroscopically, the limestone exhibited both brit- tle and ductile behavior. The parameters derived are useful for modeling the deformation of a pressure-dependent material and for computing when localization would occur.
A constitutive model for rock loaded to failure under general stress states must, at a minimum, be able to describe the pressure-dependence of failure and the dilatancy that most rocks exhibit. Pressure-dependence is included in constitutive laws by generalizing the pressure-independent yield surfaces used to describe metals to a form where the yield stress is pressure-dependent. Dilatancy is then automatically incorporated by assuming an associated flow role. However, an associated flow rule has not generally been found to work well in predicting the inelastic volume strain of rock under compressire loading, typically overpredicting the dilatant strain [Miller and Cheatham 1972, Maier and Hueckel 1979, Senseny, et al. 1983]. Assuming a non-associated flow rule is the usual solution to the problem. The additional degree of freedom allows the ratio of inelastic shear and volume strain increments to be adjusted and broadens the range of conditions for localization.
A consfitutive equation incorporating the possibility of non-normality and exploring the im- plications for localization has been developed in a series of papers by Rudnicki and Rice [Rud- nicki and Rice 1975, Rudnicki 1977, Rudnicki 1984]. There are three experimentally-determined functions in the model. The ratio between the increments of inelastic volume strain de?, and the inelastic shear strain de it is the dilatancy factor/3. Note that the superscript i is not a tensor index, but instead is used to indicate the inelastic component of the strain. A second parameter is E v, the hardening rate computed from the derivative of shear stress r with respect to a measure of the shear strain. Ep depends on the third parameter/z, the slope of the yield surface./z is frequently taken to be the same as the slope of the failure surface. One purpose of the present work is to show how the evolution of It can be measured using acoustic emissions.
