Bayesian statistical analysis on chemical composition contributing to irradiation embrittlement at high fluence region

Takamizawa, Hisashi ; Nishiyama, Yutaka 

It has been accepted that neutron irradiation embrittlement of reactor pressure vessel is caused by irradiation-induced formation of solute clusters (SCs) and matrix damages (MDs). In the present study, to analyze the contribution of chemical composition contained in SCs to irradiation embrittlement at high fluence region, statistical analysis using the Bayesian nonparametric (BNP) method was performed for Japanese PWR surveillance data. The significance of P, Si and Mn contents, which are not necessarily included in embrittlement correlations unlike the Cu and Ni content, was evaluated. The BNP method can learn the complexity of the statistical model itself from the input data and infer the predicted data with individual probability distribution of predict condition. The result suggested that irradiation embrittlement was most affected by the Si content at high fluence region.