Bayesian analysis method on processing reliability data of high flux engineering test reactor. (July 2020)

• Record Type: Journal Article
• Authors: Bao, Hui; Guo, Yun; Zhang, Hang; Peng, Changhong; Lu, Jianchao
• Is Part Of: Reliability engineering & system safety. Volume 199(2020)
• Publication Date: 2020-07
• Subjects: Research reactor -- Reliability data -- Bayesian analysis; Reliability (Engineering) -- Periodicals; System safety -- Periodicals; Industrial safety -- Periodicals; Fiabilité -- Périodiques; Sécurité des systèmes -- Périodiques; Sécurité du travail -- Périodiques; 620.00452
• DOI: 10.1016/j.ress.2020.106912
• Languages: English
• ISSNs: 0951-8320

Abstract

Highlights: Introduction of Bayesian method calculation process. Particularization of the Bayesian analysis method on processing reliability data of HFETR. Comparison of mean (ME), error factor (EF) and probability distribution curve shape among different distributions. Abstract: Probabilistic safety analysis (PSA) is widely used for safety evaluation of research reactors at present. Reliability data analysis is a major foundation for conducting research reactor quantitative safety analysis. The components of research reactors usually have high reliability and reliability data recording of Chinese research reactors lacks standardization, which lead to the insufficiency of reliability data and the inapplicability of the classical method to reliability data evaluation. Bayesian analysis is a useful approach to compensate the shortage through comprehensive utilization of all kinds of information. In this paper, the Bayesian Analysis Method on Processing Reliability Data of High Flux Engineering Test Reactor (HFETR) is introduced. The existing reliability data of the similar research reactor are elected as the prior distribution, which are combined with the reliability data of HFETR to calculate the posterior distribution applying Bayesian Analysis Method. The analysis results show that Bayesian analysis is efficient for expanding reliability data sample space of HFETR. According to the results, it can be found that: (1) the mean ( ME ) of the posterior distribution is between the mean of prior data and the mean of sample data. (2) the error factor ( EF ) of posterior data is smaller than either the prior data or the sample data, which indicate a lower uncertainty of posterior data. (3) the information with smaller factor ( EF ) has more influence on the characteristics of posterior distribution