A Hydro-mechanical Approach to Model Swelling Tests of Clay-Sulfate Rocks

We developed a coupled hydro-mechanical (HM) model based on a semi-logarithmic swelling law to reproduce the outcomes of swelling tests of a clay-sulfate rock specimen collected from the Freudenstein tunnel, which was constructed in Triassic Grabfeld Formation (formerly Gipskeuper = Gypsum Keuper) in Southwest Germany in the period of 1987-1991. The swelling tests were conducted using an oedometer apparatus under constrained (no strain) or constant load conditions. We used the strain-time data obtained from the laboratory testing to calibrate the HM model. We then ran a sensitivity analysis to reveal the importance of influential parameters, namely the maximum swelling pressure s(max)(sw) , swelling parameter k, and diffusion coefficient D on the long-term swelling behaviour of clay-sulfate rocks under the oedometer conditions. The HM model is capable of predicting long-term swelling deformations, i.e., model results were found to agree reasonably well with the experimental data. The results also show that using only 12 months experimental strain-time data to calibrate the HM model leads to an underestimation of swelling strains at the equilibrium condition. The findings show that at least 24 months experimental data is required for the model calibration.

Automated summary

We developed a coupled hydro-mechanical model to mimic the swelling tests of a clay-sulfate rock specimen. The model was calibrated using the strain-time data obtained from laboratory testing, and a sensitivity analysis was conducted to determine the influential parameters. The model was able to accurately predict the long-term swelling behavior of clay-sulfate rocks, and it was found that at least 24 months of experimental data is needed for accurate calibration.