Prediction of in-vessel debris bed properties in BWR severe accident scenarios using MELCOR and neural networks

The goal of this work is to develop a computationally efficient surrogate model (SM) for prediction of main characteristics of corium debris in the vessel lower plenum of a Nordic BWR. The SM has been developed using artificial neural networks (ANNs). The networks were trained with a database of MELCOR solutions. The effect of the noisy data in the full model (FM) database was addressed by introducing scenario classification (grouping) according to the ranges of the output parameters. SMs using different number of scenario groups with/without weighting between predictions of different ANNs were compared. The obtained SM can be used for failure domain and failure probability analysis in the risk assessment framework for Nordic BWRs.