Currently, the service life of buried gas pipelines in most Chinese cities has exceeded 20 a. Because of the combined effects of third-party damage, corrosion, and human error, the operational safety of these pipelines has decreased, leading to a high probability of accidents. Because gas pipelines are often buried underground, the transported medium is flammable and explosive, and they are located in densely populated areas, they possess characteristics of concealment and high risk. A leaking gas pipeline can easily trigger secondary fire and explosion incidents, causing severe casualties, property damage, and gas supply interruptions.

To prevent buried gas pipeline leakages and associated accidents in urban areas, a method for assessing the cascading risk evolution of pipeline leakages is proposed. Based on the operational characteristics of urban buried gas pipelines, risk factors, such as human factors, physical elements, environmental factors, and pipeline conditions are comprehensively integrated. These risk factors are further refined into a three-level indicator system. The logical relationships among the development process of leakage accidents and these risk factors are examined, and an integrated risk evolution and assessment model is developed through the Grey decision-making trial and evaluation laboratory (DEMATEL) method, interpretive structural modeling (ISM), and the dynamic Bayesian network (DBN) approach. The Grey-DEMATEL-ISM method is utilized to analyze causal hierarchical relationships among risk factors and identify key risk factors. The introduction of Grey numbers compensates for the subjective uncertainty of expert evaluations in the DEMATEL-ISM method, making the assessment results more aligned with actual conditions. Different types of risk factors are quantified through methods such as expert opinions, historical data, and probability distributions, and a pipeline leakage risk evolution DBN model is established based on the logical relationships among the risk factors. Furthermore, the probability of accident consequences is calculated and updated in real time to predict the potential development paths of accidents, thereby assisting in the safety operation and maintenance decision-making of buried gas pipelines.

The results indicated that the corrosion and human factors exhibited high levels of causality, centrality, and sensitivity, making them critical risk factors for preventing pipeline leakages. A dynamic risk analysis model was conducted the use of the gas pipeline explosion accident in Shiyan, Hubei Province, as a case study. The results showed that the posterior probabilities of corrosion-related nodes were significantly higher than those of other nodes. Based on the assessment results, targeted measures were proposed to prevent initial incidents and cut off the accident propagation paths throughout the entire life cycle during the design and manufacturing phase, laying phase, and operational phase of the buried gas pipeline. The case study verified the feasibility and effectiveness of the proposed method.

The Grey-DEMATEL-ISM-DBN model can comprehensively consider the characteristics and quantitative representation of risk factors relate to humans, machines, environment, and management, even in situations with limited data. Based on historical accident information, it establishes a multi-level hierarchical structure model for the causes of buried gas pipeline leakage accidents. The model quantifies the degree of influence and sensitivity among various risk factors and provides an intuitive display of accident evolution scenarios. The proposed method enables the dynamic analysis of the risk evolution process of buried gas pipeline leakages, providing support for decision-making in safety operations, maintenance, and accident investigations involving urban buried gas pipelines.