Probabilistic safety margin characterization of an integrated small modular reactor using MFM and adaptive polynomial chaos

Abstract

Characterization of the safety margins can provide essential information for both licensee and regulatory commission to support decision-making, especially for the next generation of nuclear power plants which implement several novel designs. However, the traditional monte-carlo based best estimate plus uncertainty method is not appropriate for large scale application because of its high computational cost. Thus, a novel hybrid method that combines multilevel flow modelling (MFM) and generalized polynomial chaos (gPC)is proposed in this paper. In the framework of the proposed method, the MFM is utilized to model the functions of collaborative working passive systems of an integrated small modular reactor (iSMR) and perform causal reasoning for selecting relevant uncertain parameters, then the gPC is implemented to construct an orthogonal polynomials-based surrogate model for the thermal hydraulic model to perform fast sensitivity analysis and safety margin characterization. An adaptive sampling algorithm is also implemented to improve the efficiency and accuracy of the safety margin characterization. The proposed method is applied to a test case to analyze the sensitivity of uncertain parameters during an SBO accident of iSMR. Subsequently, the adaptive method is implemented for probabilistic safety margin characterization after considering a delayed core makeup tank (CMT) injection and the conditional core damage frequency is predicted. The result shows great efficiency of the proposed MFM combining adaptive gPC method.