This paper describes, for the first time, a unique investigation of ASR using synthetic sandstone samples formed under applied polyaxial stresses. The strain relaxation in all tests invariably gave a good indication of the principal stress orientations. Measured strain relaxation ratios were approximately equal to corresponding applied stress ratios up 1.5, suggesting that, with a knowledge of one of the principal stresses, in regions of low stress anisotropy, stress can be calculated directly from strain. At stress ratios of 2 and above, corresponding relaxation strain ratios were variable, time- dependent and generally higher than the stress ratios. This suggests that shear stresses associated with stress anisotropy result in material damage that is partly responsible for the measured relaxation strain.

Correction of strains for "Poisson" effects reduced data scatter at the higher stress anisotropies but gave a poor correlation between applied stress and a) strain magnitudes, b) calculated stress, for a measured Poisson ratio (v) of 0.3. Lower v values gave better data fit, but it is suggested that the use of a Poisson correction of (he measured strain. Using low values of v, is unrealistic. The strain relaxation mode was assessed using the Kelvin-Voight model. The strain relaxation curves are not perfectly linear in normalised strain/time space. The material behaviour cannot therefore be fully reproduced using this visco-elastic model. Anisotropic samples especially seem to deviate from a linear trend, again suggesting that stress anisotropy and related fabric damage have an effect on the ASR response.

P. 161

This content is only available via PDF.
You can access this article if you purchase or spend a download.