Combining the bow-tie method and fuzzy logic using Mamdani inference model

Abstract

As a hybrid tool, Bowtie might be effective as a qualitative tool. However, when the quantitative contribution is considered, it relies on databases, and as such, the analysis is affected by the subjective judgment of the specialist, data variability, or lack of data which may lead to vagueness and imprecise information. Fuzzy logic can help to shed light on this matter as it is capable of resembling human reasoning. The approach offers various mechanisms to overcome these uncertainties. This paper proposes a combination of Bowtie and fuzzy logic, using a case study concerning the rupture of isobutane storage tank. The Mamdani inference method is investigated. Additionally, the Fussel-Vasely equation was used to calculate the critical importance for each basic event, which brings a complementary understanding of the system. Six tests were performed to analyse the frequencies and fussel-vasely numbers of the five basic events: “PI” (Pressure Indicator), “LIAH” (Level and Indicator Alarm High), “HE” (Human Error), “SV” (SAFE VALVE) and “DS” (Disposal System). The analysis showed that the basic event “SV” as the most critical event, followed by “LIAH” and “Human Error”.