Obtaining a more realistic hydrogen distribution in the containment by coupling MELCOR with GASFLOW

The system code MELCOR provides an integral analysis capability for severe accidents in nuclear power plants. However, its Lumped Parameter model provides less accurate information about the thermal hydraulics in the containment during a loss of coolant accident. GASFLOW is a 3D CFD code that simulates the containment thermal hydraulics and the local hydrogen distribution more realistically. Currently, the common procedure is to use a source term from a previous MELCOR calculation in GASFLOW. Yet, the effect of the more realistic GASFLOW pressure to the mass flow through the leak cannot be taken into account in this approach. This inconsistency can be overcome by coupling both codes. the coupling interface existing in MELCOR 1.8.6 was tested by calculating a postulated accident in a simplified BWR using two coupled instances of MELCOR. The results agreed perfectly with the ones from a similar stand-alone calculation. Hence, MELCOR could be coupled to GASFLOW. LOW interface for an external, explicit, and asynchronous coupling to MELCOR was developed. It enabled to receive the source term from MELCOR and to send back the containment pressure during the run time. The correct functioning of this data exchange was verified for a representative blow-down problem. er, we tested the coupling by calculating the TH7 experiment in the THAI facility with the coupled code system and with GASFLOW in stand-alone mode for comparison. The calculation results agreed very well to one another. Accordingly, the coupling functioned correctly. The results from the coupled calculation also accorded well with the experimental data such as the pressure history. Hence, the thermal hydraulic phenomena that occurred in the experiment were predicted correctly. Consequently, the MELCOR–GASFLOW coupling can account for the influence of the more realistic containment pressure calculated by GASFLOW to the accident progression. pared the results obtained using the common procedure and using the coupling for a postulated LOCA in a generic PWR. We found significant differences in the containment pressures. These caused deviating leak flow rates and differences in the hydrogen distributions. The results from coupling were assessed as more realistic.