Laboratory experiments on rock salt and phenomenological observations

We conducted an experimental campaign on salt samples from three different geographic origins. Short-term strain-controlled tests and constant mean stress tests, as well as long creep tests were performed. For reliable rock characterization and constitutive modeling, the experimental results should be analyzed considering heterogeneity. This is particularly important for the volumetric strains and the dilatancy onset (an important design criterion of underground facilities in salt formations). Heterogeneity can be quantified by the use of different measurement techniques, capturing the sample response at different scales. Material heterogeneity, due to the presence of impurities, can explain the differences observed between local and global measurements that cannot be justified by the heterogeneity induced by sample-piston friction. The volumetric strains measured during the isotropic phase are not negligible compared to those of the deviatoric phase, and the analysis of the isotropic phase can explain some of the differences observed among tests performed under equivalent conditions. To reduce bias, all samples should undergo the same preconditioning process. Long creep tests performed at constant stress deviator and short-term tests show that the mean pressure effect on the deviatoric response is of the same order of magnitude than natural variability. Reverse creep, a transient phenomenon that could be active when the stress deviator decreases, has a smaller amplitude than the associated elastic drop, and can be analyzed not only during creep tests, but also during unloading-reloading cycles, which are characterized by a progressive load drop, more representative of in situ conditions.