地铁隧道施工安全风险演化的BP-SD模型研究

为深入研究地铁隧道施工风险的动态演化过程，结合系统动力学(SD)、误差反向传播(BP)神经网络与平均影响值（MIV）算法，确定模型中影响因素间的因果关系和影响函数，构建地铁隧道施工风险演化的BP-SD模型。模型仿真结果表明:安全管理风险系统与施工人员风险系统的值的变化对地铁隧道施工风险值影响最大，施工环境系统的值的变化对地铁隧道施工风险值影响次之，机械设备系统的值的变化对地铁隧道施工风险值影响最小。因此，可通过降低安全管理风险值与施工人员风险值控制地铁隧道施工风险发生的概率。

Research on BP-SD model for evolution of construction safety risk of subway tunnel

In order to deeply study the dynamic evolution process of construction risk of subway tunnel, a BP-SD model for construction risk evolution of subway tunnel was constructed and the causal relationship and influencing functions of the influence factors in the model were determined by combining with the system dynamics (SD), error back propagation (BP) neural network and mean influence value (MIV) algorithm. The simulation results of the model showed that the change of safety management risk system value and construction personnel risk system value had the greatest impact on the construction risk value of subway tunnel, while the impact by the change of construction environment system value took the second place, and the change of the mechanical equipment system value had the least impact. Therefore, the occurrence probability of construction risk of subway tunnel can be controlled through reducing the safety management risk value and the construction personnel risk value.