A fully coupled thermo-hydro-mechanical (THM) analysis is applied to predicthe thermal, hydraulic and mechanical responses during an in situ heater experiment which simulates a nuclear waste deposition hole with waste over-pack and bentonitc buffer, surrounded by fractured rock. Predicted system response -including temperature, moisture content, fluid pressure, stress, strain and displacement -are compared to field observations at sensors located both in the clay buffer and the surrounding near-field rock. An overall good agreement between modeling and measurements indicates that most thermo-hydro-mechanical responses are a fairly well captured by the coupled analysis. Most challenging is modeling of the mechanical behavior of the fractured rock, which locally is strongly dependent on heterogeneifies in material properties.
INTRODUCTION
Most concepts for safe disposal of spent nuclear fuel are based on multiple barriers for protection of the environment against the radioactive nuclides. The spent fuel assemblies will be encapsulated in metal ,canisters, then placed in depositions holes deep in the bedrock, and the holes will be filled with expansive bentonitc clay to embed the canisters. A performance assessment of such system requires coupled thermo-hydro-mechanical modeling that simulates the coupled, highly non-linear behavior of each component (canister, rock, buffer, water) and their interactions.
As a part of the international co-operative project DECOVALEX (Jing et al. 1999), a number of computer codes have been applied to an in situ heater experiment conducted by the Japan Nuclear Cycle Development Institute (JNC) at the Kamaishi Mine in Japan. A full description of all the computer codes applied to the Kamaishi Mine heater test and intercomparison of the modeling results will be presented in a special issue to the International Journal of Rock Mechanics and Mining Sciences, during the spring of 2001 (Rutqvist et al. 2001a, and 200lb). This article describes the simulations of the Kamaishi Mine heater test using a coupled thermohydromechanical computer code ROCMAS. The analysis includes model prediction, comparison to observed field responses and recent improvements of the modeling.
