A weakest t-norm based fuzzy fault tree approach for leakage risk assessment of submarine pipeline

The leakage of oil and gas due to submarine pipeline damage will result in serious consequences while the reasons are diverse and complicated. The fault tree analysis (FTA) method provides an effective tool to systematically identify various root events and perform probabilistic risk assessments. However, crisp probability values of the basic events are requested for quantitative analysis due to the characteristics of the method itself. In this paper, a weakest t-norm (T-omega) based fuzzy fault tree approach is proposed to obtain a relative reliable probability value using domain experts' evaluations. The main contributes of this method include: a set of fuzzy numbers are defined based on the DNV-RP-F107, meanwhile, a converting method is also proposed to defuzzify the integrated fuzzy numbers; the weakest t-norm operators for trapezoidal fuzzy number are employed for less uncertainty accumulation during the aggregation process. Furthermore, a case study is presented for a detailed description of the proposed approach. A probabilistic risk assessment for leakage failure of submarine pipeline is conducted using both the proposed approach and traditional method. The results show good validity and applicability of the proposed method. The critical events are recognized after quantitative analysis and some improvement measures are put forward for engineering references.