Answers are grounded in this paper and cite their sources.
As a critical infrastructure for urban energy supply, gas pipeline networks operate in complex and variable environments, being influenced by a multitude of intertwined factors. Consequently, accidents are prone to occur, and the risk characteristics are prominent. During the process of risk evolution, various risk factors exhibit nonlinear, feedback, and heterogeneous propagation characteristics. Traditional single propagation models struggle to accurately depict the true complexity of risk propagation in urban gas pipeline network leakage accidents (UGPNLAs). To fill this research gap, this study proposes a data-driven analytical framework for examining the multi-layer hierarchical action and heterogeneous propagation characteristics of risks in UGPNLAs. Based on accident reports, the study integrates text mining techniques, Fault Tree Analysis (FTA), an improved DEMATEL method, ISM, MICMAC, and complex network theory. The research steps are as follows: Firstly, utilizing the objectivity of text mining techniques and the logicality of FTA, risk factors for UGPNLAs are systematically identified. Secondly, by introducing association rule mining techniques and improving the DEMATEL method, an AR-DEMATEL-ISM-MICMAC model is constructed to deeply explore the complex multi-layer hierarchical action relationships. Finally, a topological network of risk factors is constructed based on complex network theory, and probabilistic and robust propagation models are established to analyze the accuracy of key risk factor identification and the sensitivity of risk propagation. The results indicate that: (1) The risk factors for UGPNLAs encompass 39 risk factors across 5 dimensions, with 208 strong association rules identified among them. (2) A seven-layer causative model for UGPNLAs is identified, revealing the "root-indirect-direct" hierarchy of risk factors. workers' non-compliant operations, failure of safety devices, metering instrument failure, representing key focuses for safety management and control of UGPNLAs. (3) During the process of risk propagation, there are hierarchical differences in node propagation intensity: betweenness centrality nodes are critical points of system vulnerability, an increase in load weakens the system's self-adaptability and intensifies the conduction of high-risk nodes.
No takes yet. Share an insight, caveat, or question.
Yang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: