Interactions among safety risks in metro deep foundation pit projects: An association rule mining-based modeling framework

Abstract

The deep foundation pit project (DFPP) in subway construction is characterized by a high accident rate. Insufficient examination of the interactions among relevant safety risks often leads to defective countermeasures. This study aims to develop a general modeling and analysis procedure for risk interactions based on association rule mining and the weighted network theory, and then take China as an example to investigate the interactions among DFPP safety risks. The Apriori algorithm is employed to mine the strong association rules among risks, and the risk interaction network is constructed by integrating the mining results and expert opinions. The analysis shows that this network confirms to both scale-free and small-word properties, implying that DFPP accidents are not random events, but the result of strong interactions among safety risks emerging from multiple stakeholders. The factors from the contractor, structure, and natural environment comprise the key risk groups, while factors from the owner, designer, and social environment act as the core sources of risk and tend to induce a domino effect. It is recommended that the contractor implement network risk management rather than separate risk responses. This study may facilitate the safety management of DFPP and the development of risk analysis methods.