A systematic fault tree analysis based on multi-level flow modeling

GOFUKU Akio¹ , and OHARA Ai²

1. Graduate School of Natural Science and Technology, Okayama University, 3-1-1 Tsushima-Naka, Kita-ku, Okayama 700-8530, Japan (fukuchan@sys.okayama-u.ac.jp)

2. Department of Systems Engineering, Okayama University, 3-1-1 Tsushima-Naka, Kita-ku, Okayama 700-8530, Japan

Abstract: The fault tree analysis (FTA) is widely applied for the safety evaluation of a large-scale and mission-critical system. Because the potential of the FTA, however, strongly depends on human skill of analyzers, problems are pointed out in (1) education and training, (2) unreliable quality, (3) necessity of expertise knowledge, and (4) update of FTA results after the reconstruction of a target system. To get rid of these problems, many techniques to systematize FTA activities by applying computer technologies have been proposed. However, these techniques only use structural information of a target system and do not use functional information that is one of important properties of an artifact. The principle of FTA is to trace comprehensively cause-effect relations from a top undesirable effect to anomaly causes. The tracing is similar to the causality estimation technique that the authors proposed to find plausible counter actions to prevent or to mitigate the undesirable behavior of plants based on the model by a functional modeling technique, Multilevel Flow Modeling (MFM). The authors have extended this systematic technique to construct a fault tree (FT). This paper presents an algorithm of systematic construction of FT based on MFM models and demonstrates the applicability of the extended technique by the FT construction result of a cooling plant of nitric acid.

Keyword: fault tree analysis, functional information, Multilevel Flow Modeling, chemical plant