Failure probability analysis of hydrogen doped pipelines based on the Bayesian network

Damage from hydrogen, such as hydrogen embrittlement in natural gas pipelines, is a possibility. There are currently lacking ways for calculating the failure of hydrogen doped pipelines quan-titatively; instead, research on the safety of them is mostly concentrated on hydrogen compati-bility tests and consequence analysis. A method for quantifying hydrogen doped pipelines losses using Bayesian network-based failure probability analysis is presented. The influence of hydrogen on pipeline failure is introduced using the expert scoring method based on the analysis of his-torical failure data for gas pipelines. The failure probabilities of various components of hydrogen doped pipelines are then quantified by combining fuzzy mathematical theory and hierarchical analysis, which is subsequently utilized as prior probabilities to input into the Bayesian network for quantitative calculation. The results show that the Bayesian network can more accurately quantify the failure probability of hydrogen doped pipelines. They also show that the Bayesian network can provide a practical approach for quantitative risk assessment and safety evaluation, as well as a decision basis for the prevention of hydrogen doping pipeline failure accidents.