Suction approximation modelling gases in deep geological repositories

Accurately modelling gas behaviour in deep geological repositories is critical for assessing the long-term performance of engineered barrier systems, particularly those based on compacted bentonite. In the context of repository conditions, the non-ideal characteristics of gas mixtures recommend the implementation of robust equations of state, such as Peng–Robinson. However, the modelling of suction, a key variable in coupled thermo-hydro-mechanical analyses, involves an implicit dependency between gas pressure, vapour concentration, and suction itself, which significantly increases computational cost. This study presents two explicit strategies for estimating suction. The first assumes that the concentration of water vapour is negligible compared to that of the other gaseous species. The second considers the presence of water vapour and uses the gas pressure obtained in the previous approach to estimate the vapour concentration. Both methods were evaluated against an implicit reference solution under varying temperature, total gas concentration, and liquid pressure conditions representative of deep geological reservoirs. The results indicate that the proposed approximations yield low deviations, with errors of less than 0·15% in the first case, and practically negligible in the second. Therefore, due to its simplicity and efficiency, the second strategy is recommended for modelling gas transport and its interaction with bentonite barriers.