A prediction–evaluation approach to the full-scale emplacement experiment (FE) in Mont Terri

Abstract

AbstractA prediction–evaluation approach is developed to assess the propagation of parameter, conceptual and scenario uncertainties in the estimated near-field temperatures of the full-scale emplacement experiment at the Mont Terri rock laboratory. The uncertainty assessment is performed using a three-dimensional thermo-hydraulic numerical model of the full-scale emplacement experiment that represents the emplaced materials and surrounding Opalinus Clay and accounts for heat generation at the heaters. The propagation of parametric uncertainties is assessed using a first-order second-moment method supplemented by Monte Carlo simulations sampling the uncertain parameter space. The risk of uncertain parameters resulting in the failure of the maximum temperature criteria is evaluated with a first-order reliability method. Conceptual and scenario uncertainties are evaluated with deterministic simulation variants. After the conclusion of predictive modelling, a mid-term evaluation of the temperature predictions is performed through a comparison with measurements after 2.5 years of heating. The comparison indicates that the best estimates of temperature agree well with the measurements and that the 95% error bands assessed with parametric uncertainty envelope the measured values in almost all locations. Additional comparison with the measured degree of water saturation and the relative humidity is performed to assess the hydraulic behaviour and set the ground for the long-term evaluation, which will include predictions of the near-field pore pressures.