Application of continuous Markov-chain Monte-Carlo method to multi-unit risk evaluations considering interdependence of accident progression among multiple units. Issue 12 (2nd December 2021)

• Record Type: Journal Article
• Authors: Sawada, Kento; Yamamoto, Akio; Endo, Tomohiro; Sato, Chikahiro; Maeda, Keisuke; Jang, Sunghyon
• Is Part Of: Journal of nuclear science and technology. Volume 58:Issue 12(2021)
• Page Start: 1308
• Page End: 1317
• Publication Date: 2021-12-02
• Subjects: Probabilistic risk assessment -- multi-unit -- multi-source -- continuous Markov chain Monte-Carlo method -- inverse transform sampling method; Nuclear engineering -- Periodicals; Nuclear physics -- Periodicals; Nuclear energy -- Periodicals; 621.4805
• DOI: 10.1080/00223131.2021.1940341
• Languages: English
• ISSNs: 0022-3131

Abstract

ABSTRACT: The accident in Fukushima Dai-ichi nuclear power plants reconfirms the necessity of the safety assessment considering multiple nuclear reactor units. However, consideration of the interdependency among safety systems or events in multiple units as well as the time dependency of accident progressions is difficult in the conventional event tree method, which is widely used in probabilistic risk assessments. Recently, the continuous Markov-chain Monte-Carlo (CMMC) method coupled with a plant safety analysis or a severe accident analysis code has been paid attention to address these issues. In the present study, the CMMC coupling method is applied to the risk assessment of multiple units to clarify the benefits and issues to be resolved of this method. Since the CMMC coupling method requires many executions of an accident analysis code, a meta-model that simplifies systems and physical phenomena in accidents is used in this study to reduce computational cost. Furthermore, the inverse transform sampling method is newly adapted. Numerical analyses of BWR accident under station blackout with loss of cooling capability are carried out considering the correlation among the availabilities of mitigation systems. The results suggested that the CMMC coupling method can quantitatively treat the interdependency and time dependency among events in multiple units.