非破壊検査信号の確率論的取扱いによる配管減肉管理高度化方法論の構築

点击翻译按钮获取中文摘要

This study aims to develop a pipe reliability analysis framework that is able to incorporate the quantifications of detection uncertainty and sizing uncertainty. A hierarchical Bayesian network was built to represent the conditional relationship among the variables. Accordingly, a Bayesian model was formulated to incorporate the developed multi-variable POD model and the quantification model of estimation error to calibrate the prediction model of wall thinning with nondestructive evaluation data to achieve reasonable reliability analysis.