Sensitivity Analysis of Ex-Vessel Corium Coolability Models in MAAP5 Code for the Prediction of Molten Corium–Concrete Interaction after a Severe Accident Scenario

Abstract

A postulated progressing severe accident scenario has been simulated using MAAP5 code with the focus on ex-vessel cooling of molten corium in the reactor cavity. Various parameters associated with the prediction of molten corium–concrete interaction (MCCI) are identified. Accordingly, a sensitivity analysis is performed to assess the impact of these parameters on the predicted cavity floor erosion depth during this MCCI postulated accident. The sensitivity index of each variable parameter is determined using the Cotter indices method and Sobol′ indices method. At the early stage of the accident, the predicted cavity floor erosion depth is found to be highly sensitive to the downward heat transfer coefficient parameter with Cotter and Sobol′ indices of 94% and 50%, respectively. At the late phase of the accident, however, the cavity floor erosion depth becomes sensitive to melt eruption (Cotter index of 40%), water ingression (Cotter index of 13%), and particulate bed (Cotter index of 15%) parameters alongside the downward heat transfer coefficient (Cotter index of 16%) with the melt eruption parameter becoming dominant. Thus, the sensitivity of the code′s predictions can be minimized by improving the physical models associated with these parameters. Moreover, the sensitivity indices of these parameters can be used by model developers to identify unimportant parameters in a bid to reduce the dimension of the problem with the aim of improving the current predictive capabilities to conduct MCCI-related safety analyses.