ABSTRACT
The design of underground facilities in rock salt requires a correct characterization of the rock as a first step, which is usually performed at the laboratory scale using samples of a given size, constrained by the drilling equipment in the field and by the size of the experimental devices in the laboratory. Rock salt often contains impurities of different nature, which may rise representativity problems if the size of the tested samples is smaller than the representative volume element of the rock mass. An improper characterization of the rock in the laboratory may lead to an improper design of underground facilities. These two aspects are illustrated through an experimental campaign and subsequent computations at the cavern scale. In order to evaluate heterogeneity, different measurement techniques should be used in laboratory tests, covering different spatial scales within the samples. The use of two design criteria at the cavern scale is also discussed and proves that design criteria embedded at the constitutive level overcome the limitations of external criteria applied during post-processing of the cavern simulation results. Hence, the use of constitutive laws that reproduce laboratory data and yield pertinent long-term predictions is of utmost importance.
