摘要
将BP算法引入小波神经网络,自适应地调整小波系数和网络权重,同时利用自适应算法调节BP算法的学习率,提高收敛效率.以4车道隧道———前欧隧道的监测数据为基础,建立BP小波神经网络变形预测模型.预测结果表明:BP小波神经网络对地质条件相似,施工及初期支护方法相同的隧道断面变形进行预测,其预测结果满足工程精度要求,能较准确地预知该断面在施工过程中的变形值.
Introducing back propagation （BP） algorithm to wavelet neural network, and using adaptive algorithm for adjusting BP algorithm learning rate, under adaptive adjustment of wavelet coefficients and network weights, the efficiency of convergence was improved. Based on monitoring data of the four-lane Qian-ou tunnel, BP wavelet neural network prediction model is established. For the similar geological conditions and the same construction and initial support, BP wavelet neural network prediction of the tunnel-section deformation meets the engineering requirement of accuracy.
出处
《华侨大学学报（自然科学版）》
CAS
北大核心
2011年第6期680-683,共4页
Journal of Huaqiao University(Natural Science)
基金
福建省交通科技发展课题基金资助项目(200910)
关键词
BP小波神经网络
大断面隧道
变形预测
back propagation wavelet neural networks
large cross-section tunnel
deformation prediction
