Problems of the RELAPS/MOD3 on the reflood thermal-hydraulics

FLECHT SEASET unblocked forced reflood tests are assessed using Apollo version of RELAP5/MOD3 7j. The main purpose of the study is to examine the code predictability under forced reflood conditions for various initial power levels and flooding rates. The assessment calculations are performed for 8 tests which have been selected because they have similar initial conditions but different initial peak rod power and/or flooding rates. The assessment results show that RELAP5/MOD3 predictions are generally in good agreement with experimental data. However, in order to obtain more reasonable code predictability, three deficiencies of RELAP5/MOD3 should be improved. These are lack of grid modelling, overprediction of droplet size in dispersed droplet flow mode and unphysical transition boiling heat transfer model. These deficiencies seriously affect the assessment results