Thermal-hydraulic experiments with bentonite/crushed rock mixtures and estimation of effective parameters by inverse modeling

Abstract The permeability and thermal conductivity of bentonite/crushed rock mixtures used as backfill for a nuclear waste repository have an important impact on the maximum radioactive load of the waste canister that can be embedded. Our research pursues useful methods for estimating the permeability, thermal conductivity, and specific heat of various bentonite/crushed rock mixtures for the conditions expected to prevail at the Aspo Hard Rock Laboratory (ASPO HRL). We conducted laboratory experiments and employed inverse modeling techniques to estimate effective thermal and hydraulic parameters suitable for predictive modeling of non-isothermal flow and transport from a nuclear waste repository. Thermal parameters are often calculated based on empirical relationships developed for homogeneous clays, i.e., they are not necessarily valid for mixtures. The applicability of these methods to model thermal-hydraulic processes within the bentonite/crushed rock mixtures in a deep repository needs to be assessed. All experiments were conducted with mixtures containing sodium- (SPV Volclay) or calcium-bentonite (Calcigel) and we used water from Aspo. Hydraulic column experiments were carried out with a specially designed permeameter and Darcy's law was applied to determine the hydraulic conductivity, which followed a lognormal distribution with mean values of 1.64×10 −11 and 4.93×10 −9 m/s for the two bentonite/crushed rock mixtures studied. The thermal laboratory experiments were analyzed using inverse modeling techniques. The simulated temperature distribution matched the measured data very well at all locations along the column and for all times. The inversely estimated thermal conductivity ranged from 1.6 to 2.2 W/mK, and the specific heat from 810 to 1020 J/kg K, both consistent with the predictions of the empirical relationships. However, the calculation of the effective parameters was very sensitive to heat loss through the insulation. The newly developed experimental setup in combination with inverse modeling allows the identification of key parameters governing the hydraulic and thermal processes of bentonite/crushed rock mixtures under repository conditions.