Identification and analysis of failure risk of gas transmission pipeline girth welds based on Bayesian-bow-tie model

Author of the article:LIU Jiayue1,2, HUANG Jiawei1,2, DONG Shaohua1,2

Author's Workplace：1. China University of Petroleum (Beijing) Oil & Gas Pipeline Research Center, Beijing, 102200, China; 2. China University of Petroleum (Beijing) College of Safety and Ocean Engineering, Beijing, 102200, China

Key Words：Girth weld failure; Bayesian network; Conditional probability; Bow-tie model; Preventive measures

Abstract：In recent years, accidents caused by girth weld failure during operation of oil and gas pipelines have been increasing year by year. Therefore, to analyze the cause of girth weld failure and prevent future occurrence, in response to the 2018 CNPC China-Myanmar Gas Transmission Pipeline Incident, this paper established a Bayesian-bow-tie model to identify and analyze girth weld failure of gas transmission pipelines. The model covers the following scope:1)introduces fault tree, failure event tree, Bayesian network and bow-tie model, and calculates the conditional probability and importance of each basic event causing girth weld failure;2)analyses the possible consequences of the basic events with higher importance levels;3)proposes corresponding preventive measures for basic event which has the highest importance level. Preventive measures are proposed for basic event which has high importance level, and control measures to mitigate the consequences of the accident are also proposed. The model was verified in the girth weld cracking incident of SSGP gas transmission pipeline in Malaysia. The results show that the Bayesian-bow-tie model used to identify and analyze girth weld failure of gas transmission pipelines in this paper can deduce the cause of the accident more accurately, and can provide guidance for managing similar problems. It has important practical application significance for managing girth weld failure events.