隔挡式背斜构造区隧道涌突水量的BP网络预测

Using BP Network to Predict Tunnel Water-inrush in Partiton Style Anticlinal Belt

川东隔挡式背斜区具有其特殊的岩溶地质构造，已修建的数十条隧道每每揭露背斜区，均发生较大地下水涌突水灾害，对这种特殊构造下涌突水灾害的研究具有重要现实意义。针对岩溶地下水系统具有强烈的非线性特征，建立合适的BP神经网络，评价某在建公路华蓥山隧道的涌突水灾害危险等级。结果显示，背斜两翼非可溶岩层等级为Ⅰ～Ⅲ级；核部可溶岩地层为Ⅲ～Ⅴ级，且越靠近核部危险性等级越高；西翼涌突水危险性等级高于东翼。评价结果与勘察阶段的研究相互印证。

In the eastern Sichuan, there are a series of a special geological structure--partiton style anticlinals. where have built dozens of tunnels, and occurred serious great water-inrush hazards. So research the tunnel water-inrush hazard in this area is practical significance. Groundwater system is with strong nonlinear characteristics in karst, to establish the appropriate BP neural network, assessment the hazard rating of Huayingshan tunnel which is been built. The results show that： the non-soluble rock in anticline wings is evaluated I -II level; the soluble rock in anticline core is evaluated III-V level, and the closer to core, the level of risk higher; risk level in west wing is higher than the east. The results is consistent with the research in investigation stage.