Evaluation of the Parameter Importance Changes Over Time for Nuclear Reactor Transients

Abstract

Abstract The uncertainty analysis and sensitivity analysis are two important parts in the safety analysis of nuclear reactor accidents, and the main purpose of the sensitivity analysis is to quantify the importance of the input on the target output. In typical sensitivity analysis for the nuclear reactor transients, specific techniques such as the regression analysis or correlation coefficient analysis are utilized to calculate the sensitivity measures of the inputs on the target output at scalar points, for instance the peak cladding temperature (PCT). However, the phenomena and parameters vary significantly over time during the transients, thus the influence of each input also changes. In order to evaluate the importance changes of the inputs on the target output over time, a time-dependent global sensitivity analysis technique was developed base on the B-spline function system of the functional data analysis, the landmark registration strategy, and an efficient moment-independent sensitivity analysis method. The developed technique was adopted in the sensitivity analysis of a large break loss of coolant accident (LBLOCA) of the nuclear reactor, and the time-dependent sensitivity measures of the inputs on the maximum cladding temperature were acquired and analyzed. Results showed that the proposed technique is capable of evaluating the importance changes of the key parameters in nuclear reactor accident transients.