Risk Coupling Analysis of Deep Foundation Pits Adjacent to Existing Underpass Tunnels Based on Dynamic Bayesian Network and N–K Model

Because deep foundation pits and tunnels are deformation-sensitive structures, the safety of these projects is generally affected by coupled risks. In deep foundation pit construction, if the existing tunnel structure adjacent to the deposit is damaged, it can produce a severe group disaster. It is necessary to identify an efficient risk analysis model to study the dynamic coupled risk of deep foundation pit projects adjacent to existing underpass tunnels and to analyze the risk evolution law to achieve effective real-time safety control. This study proposes a coupled risk analysis model using the N–K model and dynamic Bayesian network to construct deep foundation pits in adjacent existing underpass tunnels. The model is predicated on association rules to explore the interrelationship between risk factors to build a dynamic Bayesian network structure. In addition, the N–K model is utilized to quantify coupled risks under such complex working conditions and to optimize the dynamic Bayesian network structure. The developed model clarifies the risk coupling mechanism of deep foundation pit construction adjacent to an existing underpass tunnel, finds the critical points in the risk transfer process, and conducts dynamic risk prediction and accident causation diagnosis for the coupled risk to realize the dynamic control of the coupled risk in the adjacent existing underpass tunnel construction. Taking the Nanning underground comprehensive utilization project as an example, the validity and applicability of the proposed approach were tested. The results showed that the model is feasible and has application potential, providing effective decision support for safety control while constructing deep foundation pits adjacent to existing tunnels.