BP神经网络在隧道围岩力学参数反演中的应用

摘要

The aim of this research is to ensure the construction safety and optimize the design of tunnels usinginformation technology. With the construction of Zhuzang tunnel of Gucheng-Zhuxi highway as an engineering background, we predicted the final deformation by regression equation of exponential function deduced from the field displacement measurement data. Subsequently, on the basis of the predicted deformation, we carried out back a-nalysis on the mechanical parameters (deformation modulus E, cohesion C, internal friction angle ψ) of the tunnel ' s surrounding rock through BP neural network which has good nonlinear mapping ability. The surrounding rock type and material parameters can be obtained in time to provide parameters for the design and construction of the tunnel.%以谷城至竹溪高速公路珠藏洞隧道施工监测为工程依托,根据现场变形监测数据的指数函数回归方程,对最终变形量进行了预测,并基于其预测值,借助BP神经网络的超强非线性映射能力,对隧道围岩力学参数(变形模量E、黏聚力C、内摩擦角Ψ)进行反演,以及时掌握开挖围岩类型和材料特性参数,为隧道工程施工和设计提供参数依据,从而达到安全施工和优化设计的目的,以实现隧道的信息化施工与设计.