[2D13-15] Uncertainty Analysis Method

[2D13] Analysis of TI-SGTR initiated from ATWS in 3 Loop PWR Plants

(1)Development of Modeling Base Time Series Data Analysis System

*Akitoshi Hotta¹, Wataru Kikuchi¹, Tadashi Kamada² (1. Nuclear Regulation Authority, 2. FMIC R&D)

Keywords:severe accident, source term evaluation, model identification, ARMA method, multivariable time series analysis

The source term evaluation reflecting SA measures for domestic LWRs is underway using MELCOR. The realistic source term evaluation requires a statistical evaluation considering various uncertainties. We are developing FlexMelPos as a system for integrated processing of parameter sampling and analysis results. In this presentation, we will explain the overall functions of this system and focus on the Dynamic Analysis Package (DAP) , which is a time-series data analysis method obtained from accident progression analysis.

Statistical evaluation using MELCOR is hampered by the frequent abnormal termination of calculation. DAP performs model identification using the ARMA method being widely used in fields such as forecasting natural phenomena and financial trends, and visualizes causal relationships between variables through multivariate analysis. These methodologies enables to judge whether the terminated cases are significant with respect to the figure-of-merit. Furthermore, it is possible to visualize temporal changes in the probability structure of a time series based on the AIC and judge the influence of changing integration time steps.