---

概要

Abstract This study proposes the innovative method for evaluating the failure distribution functions that are essentially necessary to optimize the replacement interval of plant equipment. It is not easy to precisely evaluate the parameters of failure dis tribution functions since the enough failure data can not be easily obtained in Japanese plants. Then it is important to evaluate the confidence interval of the parameters of distribution functions to consider the risks due to the equipment failures. It i s also not easy to evaluate the confidence interval by the ordinary static method. So this study proposes to use the Bayesian method that can easily evaluate the confidence distributions as posterior distribution and compares the evaluation results by maxi mum likelihood estimation method. I found that both estimation methods give almost the same results as for the point estimation and Bayesian method is preferable to evaluate the confidence distributions of the parameters of failure distribution functions.

---

全文掲載のPDFファイルダウンロード

• PDFファイルダウンロード