Shield Tunneling Parameters Optimization Based on Dynamic Bayesian Networks

To control the surface subsidence in shield tunnel construction, a DBN-based, parameters optimization method was proposed and applied to a tunnel project in Wuhan, China. Selected the construction parameters to optimize as network nodes and set discretizing rules to determine the nodes domain, then discretized monitoring data and applied them to parameter learning to obtain complete DBN model. This model was validated with engineering measured data , then applied it to determine the setting optimal operating-range of each parameters and within that optimized the parameters real-timely according to the monitoring data changing. The results show that this model can reflect the inner link between the surface subsidence and shield construction parameters and this DBN-based method can control the surface subsidence in shield tunnel construction effectively.