ATHLET-CD/COCOSYS Analyses of Severe Accidents in Fukushima Daiichi Units 2 and 3: German Contribution to the OECD/NEA BSAF Project, Phase 1

On behalf of the German Federal Ministry of Economics and Technology, Gesellschaft für Anlagen- und Reaktorsicherheit (GRS) participated in the Organisation for Economic Co-operation and Development/Nuclear Energy Agency (OECD/NEA) project titled Benchmark Study of the Accident at the Fukushima Daiichi Nuclear Power Plant (BSAF). Analysis of the severe accidents (SAs) that happened in the Fukushima Daiichi nuclear power plant (NPP) requires well-qualified methods and codes, e.g., ATHLET-CD and COCOSYS developed and applied at GRS. Coupled ATHLET-CD/ COCOSYS analyses for the SA progression during the first days for the similar Units 2 and 3 of Fukushima Daiichi have been provided as the German contribution to the OECD/NEA BSAF project, phase 1. ATHLET-CD is a detailed SA code based on the thermal-hydraulic code ATHLET of GRS to simulate the processes in the reactor circuit before and during core degradation. COCOSYS is focused on the simulation of design basis and SA progression in the containment and the surrounding buildings of the NPP.

The focus is on selected results of the SA analyses in the boiling water reactors at the Fukushima Daiichi site especially with regard to the conditions in the torus-shaped wetwell (WW) of the primary containment and specific modeling needs. The GRS results obtained in this OECD/NEA BSAF project, phase 1, are encouraging in terms of capturing essential SA signatures like reactor and containment pressure, reactor water level, and WW temperature history for the first days of the accident in the analyzed Units 2 and 3. A detailed plant model was built up especially with a detailed torus nodalization allowing modeling of relevant phenomena like thermal stratification in the torus water pool and consideration of plant-specific details with regard to local water/steam injections into the torus water pool through safety systems and valves. As a result, the calculated accident progression of the best-estimate analyses for both units follows the accident time line quite closely. This is a prerequisite for reasonable core degradation calculations, as the time window available for the onset of core degradation between known points in time when safety injection stops and mobile pump injection into the reactor starts is small. The analyses are useful to identify areas that require further attention, to define information needs to be gained from the decommissioning, and to define further research needs with regard to experiments and code improvement.