Probabilistic risk analysis of process systems considering epistemic and aleatory uncertainties: a comparison study

The fault tree analysis (FTA) is one of the important probabilistic risk assessment tools that is extensively used in different industrial applications. However, the classical FTA has been widely criticized due to its ambiguity and vagueness in finding the probability of basic events (BEs), and accordingly, in computing the probability of top events (TEs). In this paper, we propose a new approach considering the integration of fuzzy set theory and evidence theory for handling both types of uncertainties, i.e., epistemic and aleatory. In addition, to estimate the probability of TEs alternatively, the normal distribution is applied. By performing a numerical example of risk assessment in a collapse case study for a common offshore facility platform, the effectiveness of the approach is demonstrated. Finally, in this study, the sensitivity analysis is also performed to represent the vulnerability of the causes and the interaction of different inputs, meaning that the proposed model is highly reliable and effective for real applications.