Numerical Investigation of the Applicability of Time-Dependent Models

Time-dependency manifests as several different mechanisms, namely creep, swelling, consolidation, stress relaxation, and dilatancy. Creep is an important mechanism to consider in the design of projects with ultra-long lifespans or for projects developed in rocks that readily creep, such as rock salt. Creep models are typically calibrated at the lab scale to a small range of applied constant stresses and the resulting strains are measured over time.

This paper critically examines the behaviour of a visco-elastic creep model, known as the Burgers model, the two-component power law, and a modified strain-hardening law, when modelled at the lab scale and tunnel scale at various times and induced stresses using rock salt as an analogue, calibrated at 10MPa. A sensitivity analysis using different applied stresses for the models is then examined. Additionally, the effect of boundary conditions for a circular tunnel numerical model is examined for time-dependent behaviour.

It is shown that each of the three creep models are adequate at capturing the strain behaviour of halite at the both the lab scale and tunnel scale at the calibrated stress. When stresses are markedly changed from the calibrated value, the three begin to differ drastically due to their strain rate-stress behaviour, which is significant for scenarios in which in-situ stresses are not the same as lab scale calibration stresses. Additionally, while the effect of boundary conditions is shown to be insignificant in a purely elastic analysis, the introduction of viscosity in the model leads to a significant effect of boundary conditions on the strain-rate behaviour of the models.