Sensitivity Analysis of Key Parameters for Dynamical Analysis of Reactor Coolant System

Dynamical analysis of reactor coolant system（RCS） under extreme accidents is a key technical approach for nuclear power plant safety assessment. Quantitative examination of the sensitivity of key RCS structural parameters against system dynamical responses is a crucial aspect for the reliable evaluation of RCS responses. This paper presents a sensitivity analysis of steam generator（SG） support stiffness against the RCS load distribution under seismic load by means of global sensitivity analysis and correlation analysis. It is shown that SG support stiffness is more influential to load distributions at local scale, namely, close to SG, and less influential to the load distribution of distant reactor pressure vessel（RPV）. Moreover, the input-output relationship that characterizes the mapping from key parameters to RCS load distributions is constructed and a regression model via artificial neural network（ANN） is built. The ANN model enables the fast and accurate estimation of RCS load distribution upon structural design modifications of SG supports.