| description abstract | Road collapse accidents are increasing in recent years together with the rapid utilization of underground spaces in cities. Due to the complexity of road collapse accidents, it is of practical significance to explore the correlation hidden behind the causal factors. Complex networks have shown the value to not only obtain the evolution of accidents but also reveal the relationships among causal factors. This paper uses a complex network to statistically investigate the relevance of causes in conjunction with association rules. A real-world data set is analyzed for accident cause analysis, including 1,345 urban road collapses resulting from 13 factors. The aim is to inspect the relationship among the complicated causes of accidents and further identify core factors, source factors, and purposive factors with the aid of centrality and visualization of the complex network. The results reveal that incomplete backfill, metro construction, and rain wash are core factors of road collapse accidents. Moreover, high road load is a significant source factor in road collapse accidents, likely leading to incomplete backfill and loose soil. The results provide possible strategies for effective accident prevention by focusing on the principal factors in safety management. With rapid urbanization and the intensified utilization of underground spaces, road collapse accidents have emerged as a pressing concern in numerous cities. Beyond merely identifying the main causes of road collapse accidents, it is also crucial to unravel the intricate roles of causal factors that contribute to these accidents. This research endeavors to identify and classify the principal causes of road collapse accidents into three distinct groups: core, source, and purposive factors. By examining the intricate correlations between these factors and tracing the accident chain, a deeper understanding of their individual roles can be gained. This differentiation of factor roles facilitates the formulation of targeted prevention strategies that are tailored to address the contributions of factors. Moreover, the method holds potential for various accident types, making it a valuable tool for gaining insights into the intricate causes of accidents, thereby contributing to the enhancement of safety measures and accident prevention strategies. | |
