Hybrid Tw$T_w$‐Based Integrated FFT‐FBN Approach for System Reliability Analysis

ABSTRACT This research paper presents a novel system reliability analysis approach that integrates a fuzzy Bayesian network (FBN) with a fuzzy fault tree (FFT) and utilizes weakest t‐norm ()‐based fuzzy arithmetic operations. When quantitative failure data for system components is imprecise or unavailable, expert opinions expressed as qualitative linguistic terms are utilized. In this study, linguistic terms are quantified using triangular and trapezoidal fuzzy numbers. Since experts may have diverse opinions regarding the failure data or possibilities of system components, a ‐based improved similarity aggregation method (‐ISAM) is employed to consolidate their fuzzy assessments with reduced uncertainty spread. Additionally, ‐based fuzzy arithmetic operations are applied to minimize the propagation of uncertainty during fuzzy computations. This paper analyzes the impact of expert group selection and their opinions on the failure probability of the top event. To demonstrate the applicability of the proposed research study, we perform a case study of crankcase explosions in Marine diesel engines. Three distinct expert groups assess the occurrence possibilities of basic events(BEs). The results show that when experts agree, the group composition has minimal impact on top event (TE) probability. However, when experts' opinions differ, there is a significant change in the TE probability, showing that the results are sensitive to expert opinions. The Fussell‐Vesely importance measure is utilized to rank the BEs based on their contribution to the TE occurrence. For verification purposes, the obtained results are compared with some existing methods. The findings of this research study have significantly contributed to the theoretical understanding and practical advancements in maritime safety, environmental protection, and the global economy.