A full-scale in situ heating test in Callovo-Oxfordian claystone: observations, analysis and interpretation

Abstract

The paper describes the performance, observations and numerical interpretation of a full-scale in situ heating test conducted on Callovo-Oxfordian (COx) claystone in the Meuse / Haute-Marne (MHM) Underground Research Laboratory (URL) simulating a heat-emitting, high-level radioactive waste disposal concept. In the experiment, five heaters 3 m-long are placed in the axis of a microtunnel excavated in COx claystone to simulate the heat production of radioactive waste. The test is fully instrumented, and attention is focused on the near-field region's thermo-hydro-mechanical (THM) behaviour consisting of the casing surrounding the heater and the host formation (COx claystone). The interpretation of the test is assisted by the performance of a numerical analysis based on a coupled formulation incorporating the relevant THM phenomena. The calculations have used a constitutive law especially developed for this type of material. Initial and boundary conditions for analysis as well as material parameters are determined from a comprehensive field and laboratory experimental programme. The paper presents and discusses the thermal, hydraulic and mechanical observations in COx claystone, casing and an annular air-filled gap between the host formation and the casing. Special attention is also paid to the mechanisms involved in the interface between the rock and the casing. Heating has been applied in two stages; a cooling stage (also applied in steps) completes the experiment. The numerical analysis performed has proved able to represent the progress of the experiment very satisfactorily. By performing the 3D numerical analysis, it has been possible to incorporate anisotropies of material parameters and of in situ stresses. The performance and analysis of the in situ test have significantly enhanced the understanding of a complex THM problem and have proved the capability of the numerical formulation to provide adequate modelling capacity.