基于DBN的盾构隧道施工参数优化方法研究

OPTIMIZATION OF SHIELD TUNNELING PARAMETERS BASED ON DYNAMIC BAYESIAN NETWORKS

为严格控制地表沉降,最大限度地减小隧道盾构施工给周围环境带来的不利影响,针对现有方法的不足,基于动态贝叶斯网络(DBN)理论提出一种盾构隧道施工参数优化方法,并将其应用于武汉某盾构隧道工程。首先选定待优化施工参数作为拟建网络结构的网络节点;然后设定离散化规则划分节点状态,以离散化工程实测数据完成参数学习,得到完整的DBN优化模型;该模型经验证用于工程施工参数优化,结果表明:该模型能够正确反映地表沉降与各施工参数之间的内在联系,具备一定的科学性;基于该模型进行反向诊断推理,能够确定各施工参数的最优设定区间;在该区间内对施工参数进行实时优化,能够降低地表沉降风险;该方法实时性强,具备一定的工程应用价值。

To control the surface subsidence and minimize the negative influence from shield tunneling to surrounding environment,a DBN-based,parameters optimization method in shield tunneling was proposed according to the deficiency of the existing method and was applied to a shield tunnel in Wuhan,China. Firstly the main construction parameters were selected to optimize as node variables and the network structure was built. Then discretizing rules were set to divide the nodes states and measured data were discredited for parameter learning to get the complete DBN optimization model. After model validation with engineering measured data applied this model to parameters optimization. The results show that this model can reflect the inner link between the surface subsidence and shield construction parameters. Based on this model the optimal setting-range of each construction parameters can be determined and within the range construction parameters can be adjusted and optimized real-timely,which is helpful to reduce surface subsidence;This method is valuable in practice