Study on quantitative assessment of urban gas pipeline leakage risk using a TFN-BN hybrid model

• TFN–BN hybrid model quantifies urban gas pipeline leakage risk. • Failure causes are restructured, adding hidden accessory and install/trench defects. • Multi-tier Bow-tie links causes, barriers and three consequence modules. • Nine-scale expert scoring with TFNs reduces ambiguity in coupled failure probability. • FAHP-based consequence scoring and a Beijing case validate integrity-consistent risk. This paper presents a quantitative risk assessment model for urban gas pipelines to address the increasing risk of leaks due to the expansion of pipeline networks and the complexity of the installation environment. The model integrates trapezoidal fuzzy numbers (TFN) and Bayesian networks (BN) to enhance accuracy. By restructuring the failure cause system of gas pipelines, traditional failure causes are categorized into six groups: third-party damage, misoperation, pipeline corrosion, pipeline defect, safety accessory failure, and natural environmental risks. Additionally, this model introduces hidden safety accessory failures, pipeline trench construction defects, and pipeline installation defects. A multi-tier Bow-tie model is developed, comprising 51 intermediate events, 91 basic events, 6 autonomous safety barriers, and 3 accident consequence analysis modules. A framework is proposed for calculating the probability of coupling failure in gas pipelines, utilizing a nine-scale expert score–trapezoid fuzzy number system to mitigate the ambiguity in expert assessments. Additionally, a Bayesian dynamic inference model based on evidence specification is proposed to quantitatively evaluate the failure probabilities of risk factors in gas pipelines. Moreover, a multi–criteria quantification system for evaluating the consequences of leaks is devised by integrating fuzzy Analytic Hierarchy Process (FAHP). Using a typical pipeline segment in Beijing as a case study, the overall failure probability of the pipeline is 9.91×10 -4 (km -1 ·a -1 ). The risk assessments for fire, vapor cloud explosion, and safety discharge are 5.40×10 -5 , 2.26×10 -6 , and 1.50×10 -7 , respectively, falling within the low–risk category. The analysis conclusions are consistent with pipeline integrity assessment results in terms of risk classification. The empirical findings demonstrate that the proposed framework encompassing risk identification, failure probability computation, and consequence quantification offers an engineering-applicable theoretical basis for risk management assessment of urban gas pipelines under conditions of incomplete failure data.