小净距2扩4隧道变形规律的BP小波神经预测  

BP Wavelet Neural Prediction of Deformation Law of Two-to-Four Lane Tunnels with Small Clear Interval

作　　者：林大炜 林从谋[1] 黄逸群[1] 黄清祥 孟希[1] 

机构地区：[1]华侨大学岩土工程研究所,福建泉州362021

出　　处：《华侨大学学报（自然科学版）》2014年第2期207-211,共5页Journal of Huaqiao University(Natural Science)

基　　金：国家自然科学基金资助项目(51278208);福建省交通科技发展基金资助项目(200910)

摘　　要：以泉厦高速扩建工程大帽山隧道为例,通过周边位移和拱顶沉降的监测数据对小净距扩挖隧道的围岩变形规律进行分析.研究表明:小净距2扩4隧道具有和其他隧道不同的变形规律.在此基础上将小波函数引入BP神经网络建立BP小波神经网络模型,对特大断面超小净距隧道2扩4时围岩变形进行预测,并将预测结果与BP神经网络的预测结果进行对比.结果表明:BP小波神经网络模型收敛快、精度高,优于BP神经网络模型,预测的精度达10%以内,满足工程精度要求.Taking the expansion engineering Damaoshan tunnels in the expressway from Quanzhou to Xiamen as an ex- ample, the deformation of surrounding rock of small interval expanded tunnels was analyzed by the monitoring data of pe- ripheral displacement and vault settlement. The results show： the deformation of two-to-four lane tunnels with small clear interval is different to the deformation of other tunnels. Introducing wavelet function to wavelet neural network, BP wavelet neural network model was established to predict the surrounding rock deformation of extra-large section and two- to-four lane tunnels with small clear interval. Comparing BP wavelet neural network results with BP neural network re- sults, it is indicated： BP wavelet neural network model has faster convergence and higher precision than BP neural net- work model. The accuracy of the forecast results is in 10%, which meets the engineering requirement.

关 键 词：隧道 围岩 变形规律 BP小波神经网络 预测方法 小净距 扩挖 

分 类 号：U456.3[建筑科学—桥梁与隧道工程]