Reliability analysis of dynamic systems by translating temporal fault trees into Bayesian networks

Kabir, Sohag; Walker, Martin; Papadopoulos, Yiannis

Abstract

Classical combinatorial fault trees can be used to assess combinations of failures but are unable to capture sequences of faults, which are important in complex dynamic systems. A number of proposed techniques extend fault tree analysis for dynamic systems. One of such technique, Pandora, introduces temporal gates to capture the sequencing of events and allows qualitative analysis of temporal fault trees. Pandora can be easily integrated in model-based design and analysis techniques. It is, therefore, useful to explore the possible avenues for quantitative analysis of Pandora temporal fault trees, and we identify Bayesian Networks as a possible framework for such analysis. We describe how Pandora fault trees can be translated to Bayesian Networks for dynamic dependability analysis and demonstrate the process on a simplified fuel system model. The conversion facilitates predictive reliability analysis of Pandora fault trees, but also opens the way for post-hoc diagnostic analysis of failures.

Citation

Kabir, S., Walker, M., & Papadopoulos, Y. (2014). Reliability analysis of dynamic systems by translating temporal fault trees into Bayesian networks. In Model-Based Safety and Assessment; Lecture Notes in Computer Science (96-109). Springer Verlag. https://doi.org/10.1007/978-3-319-12214-4_8

• Publication Date: 2014
• Deposit Date: Apr 12, 2016
• Publicly Available Date: Apr 12, 2016
• Journal: Lecture notes in computer science
• Print ISSN: 0302-9743
• Electronic ISSN: 1611-3349
• Publisher: Springer Verlag
• Peer Reviewed: Peer Reviewed
• Volume: 8822
• Pages: 96-109
• Book Title: Model-Based Safety and Assessment; Lecture Notes in Computer Science
• ISBN: 9783319122137; 9783319122144
• DOI: https://doi.org/10.1007/978-3-319-12214-4_8
• Keywords: Bayesian networks; Fault tree; Input event; Conditional probability table; Fault tree analysis
• Public URL: https://hull-repository.worktribe.com/output/435722
• Publisher URL: http://link.springer.com/chapter/10.1007%2F978-3-319-12214-4_8
• Additional Information: This is the authors accepted manuscript of an article published in Lecture notes in computer science, v.8822. The final publication is available at Springer via http://dx.doi.org/10.1007/978-3-319-12214-4_8