Process system failure evaluation method based on a Noisy-OR gate intuitionistic fuzzy Bayesian network in an uncertain environment

Abstract

In the system reliability evaluation of the process industries, it is sometimes difficult to get precise and sufficient failure data of system components utilized to calculate the failure probability. In this study, a Noisy-OR gate Bayesian network method based on intuitionistic fuzzy theory is proposed in cases of imprecise and insufficient historical data. The main contributes of this method include: a set of triangular intuitionistic fuzzy numbers considering uncertainty and hesitation is defined based on the standards and industry practices, meanwhile, a corresponding probability conversion method is also proposed; an improved similarity aggregation method is employed for less uncertainty accumulation and reducing the deviation caused by individual differences during the aggregation; the uncertain causal relationship among the relevant nodes is determined by applying the Noisy-OR gate in the Bayesian network. Furthermore, a case study of the crude oil tank fire and explosion accident is performed to illustrate the applicability of proposed approach. The comparison between the obtained results and that from pre-existing methods shows that the proposed method can provide a more suitable result in an uncertain environment. The weak links of the crude oil tank system are identified through Bayesian reasoning and sensitivity analysis, which can aid decision-making and improve the security execution of the crude oil tank system.