基于遗传神经网络算法的岩溶区隧道溶洞规模预测

PREDICTION OF KARST CAVE SCALE OF TUNNEL IN KARST AREA BASED ON GENETIC NEURAL NETWORK ALGORITHM

岩溶区溶洞发育受多种因素影响,具有高度复杂性和非线性特征。为准确预测岩溶区隧道溶洞规模,降低隧道施工过程中遭遇岩溶洞穴的风险,在参考相关文献的基础上,结合已经探明的溶洞信息与岩溶发育机理选取岩石可溶性、岩层厚度、地表汇水能力、单斜与褶皱构造、断裂构造、岩溶水垂直分带6个主要因素作为岩溶区隧道溶洞规模预测指标。针对岩溶发育的非线性特点,使用BP神经网络建立了溶洞规模预测模型,并使用遗传算法优化BP神经网络的权值与阈值。在工程应用上,对某隧道进行溶洞规模预测,预测结果与实际施工情况较一致。

The development of karst caves in karst areas is affected by many factors and has high complexity and nonlinear characteristics. In order to accurately predict the scale of tunnel caves in karst areas and reduce the risk of encountering karst caves during tunnel construction,based on reference to relevant literature,combined with the proven cave information and karst development mechanism,six main factors including rock solubility, rock stratum thickness, surface water catchment capacity, monocline and fold structure, fault structure and vertical zoning of karst water are selected as prediction indexes of tunnel karst cave scale in karst area. Aiming at the non-linear characteristics of karst development,a karst cave scale prediction model was established using BP neural network,and genetic algorithm was used to optimize the weights and thresholds of BP neural network. In engineering application,the scale of a certain tunnel is predicted,and the predicted result is more consistent with the actual construction situation.