论文部分内容阅读
大型水电站地下厂房多采用分层分区开挖方法,施工期围岩变形受施工程序的影响呈强烈的非线性特点。针对传统回归模型对此类地下厂房围岩变形预测精度较低的问题,考虑了影响围岩变形的主要因素,采用遗传算法优化BP神经网络,结合动态分析法建立了施工期围岩变形预测的GA-BP模型。GA-BP模型在向家坝地下厂房运用结果表明,与回归模型相比,GA-BP预测模型提高了预测结果的精度与稳定性,适合施工现场的监测分析与预测。
The underground powerhouse of large-scale hydropower station mostly adopts the method of stratified zone excavation. The deformation of surrounding rock during construction is strongly nonlinear by the construction procedure. Aiming at the low accuracy of the traditional regression model for the deformation prediction of the surrounding rock of this kind of underground powerhouse, the main factors affecting the deformation of the surrounding rock are considered. The BP neural network is optimized by genetic algorithm and the deformation prediction of the surrounding rock is established by dynamic analysis GA-BP model. The results of applying GA-BP model to the underground powerhouse of Xiangjiaba show that compared with the regression model, the GA-BP prediction model can improve the accuracy and stability of the prediction results and is suitable for the monitoring analysis and prediction of the construction site.