With the rapid urbanization, urban underground utility tunnels have seen fast growth in China in the past few years. Urban utility tunnels can house various kinds of city ‘lifelines’ such as natural gas pipeline, heat pipeline, water supply system, sewer pipeline, electricity and telecommunication cables, which are of great significance to guarantee essential flows of energy, information and logistics for urban life. If a utility tunnel accident occurs, the consequences could be catastrophic. Risk assessment has been an important tool to examine the safety performance of industrial facilities and the effectiveness of safety measures. In this study, an integrated model based on dynamic hazard scenario identification (DHSI), Bayesian network (BN) modeling and risk analysis is proposed for risk assessment of urban utility tunnels. The worst-case scenario of urban utility tunnel accidents is identified by DHSI and modelled by BN. Meanwhile, risk analysis is conducted based on the results of BN considering casualties and economic losses. Finally, the integrated method is applied to evaluate the risk level of a real-world utility tunnel. The results indicate that the integrated quantitative risk assessment framework is an alternative and effective tool for safety assessment and land-use planning of urban utility tunnels.@en

With rapid urbanization in China, many underground utility tunnels have been established these years. This huge underground construction facilitates city life, but may introduce societal risks due to the installation of high-risk pipelines. Natural gas pipelines have the potential to cause catastrophic accident if a gas leakage and a subsequent explosion occurs. The potential hazards in the gas compartments of a utility tunnel are quite different from those in conventional directly buried gas pipelines. This study developed a dynamic quantitative risk analysis method for natural gas pipelines in a utility tunnel. First, potential accident scenarios of natural gas pipelines situated in a utility tunnel were identified and implemented in a Bow-tie diagram based on case studies of typical gas pipeline accidents and expert experience. Then, a Bayesian network was established from the Bow-tie diagram using a mapping algorithm. Based on a comprehensive analysis of the results of probability updating and sensitivity analysis, critical influencing factors were identified. The proposed framework provides a predictive analysis of the gas pipeline accident evolution process from causes to consequences and examines key challenges in gas pipeline risk management in utility tunnels.@en