OVER-VIEW OF IN-SITU THERMOMECHANICAL EXPERIMENTS IN CLAY - CONCEPT, RESULTS AND INTERPRETATION

The HADES project (High Activity Disposal Experimental Site) aims at d emonstrating the technical feasibility and the long-term safety of geo logical disposal of reprocessed HLW (High-Level Wastes) radioactive wa stes. This disposal could be realised in the Tertiary Boom clay format ion below the Mol/Dessel nuclear site. Previous studies in the 80's on the geomechanical behaviour of Boom clay, at host rock temperature (1 5 degrees C), have demonstrated the mining capabilities of this clay. European partners have collaborated to increase the number of in-situ tests to be developed and operated from the Underground Research Facil ity( URF). Integrated large-scale experiments have been developed duri ng the last four years, within the framework of the Commission of the European Communities (CEC) research contracts, in order to gain more i nsight into the thermal influence of heat-emitting wastes on the clay behaviour in the near field. This paper summarizes the knowledge gaine d from the in-situ hydro-thermo-mechanical experiments realised from t he URF. The objectives and main issues of the tests are described. Inf ormation regarding selection of parameters and representativeness of t heir measurements is discussed. Despite their different aims and desig ns, three hydro-thermo-mechanical experiments are compared in order to emphasize similarities in the general behaviour of the clay massif su bmitted to a thermal load. It is observed that its behaviour is qualit atively similar to the reaction upon a constant total stress increase. With the information now available, it is not obvious to draw conclus ions with respect to the irreversibility of some phenomena. Simple in- house analytical codes are able to simulate with a good accuracy the h ydro-thermo-mechanical behaviour of clay. Nevertheless, these codes ar e not sufficient to describe the complexity of the phenomena that are involved in hydro-thermo-mechanical processes as pointed out by held d ata. In order to validate more complex models, the possibilities to im prove the representativeness of the measurements are investigated.