摘要
地铁已经成为城市普遍的交通工具,为保障地铁运营的安全,需要及时掌握地铁隧道的结构变化情况。通过小波神经网络模型对地铁保护区进行预测,首先利用小波对原始数据进行分解、降噪,然后利用神经网络进行建模并预报。以南京某地铁保护区的监测项目为例,采用该模型的预测结果同神经网络模型的结果进行比较分析。结果表明:经过小波变换的神经网络的预测效果更好。
Metro has become the common vehicle in cities.In order to ensure the operating security,the structural deformation of metro tunnel needs to be obtained timely.A wavelet neural network model is proposed for the prediction of metro protected areas.The paper first uses the wavelet to decompose and reduce the original data,and then uses neural network for modeling and prediction.A monitoring project in Nanjing is taken as the case.The wavelet neural network model is compared with the time series neural network model.Results show that predictive effect of the neural network through wavelet transformation is better.
作者
王俊杰
徐东风
WANG Junjie XU Dongfeng(Nanjing Metro Resources Development Co. , Ltd. , Nanjing 210012, China)
出处
《黑龙江工程学院学报》
CAS
2017年第3期10-12,37,共4页
Journal of Heilongjiang Institute of Technology
关键词
小波变换
时间序列神经网络
地铁保护区
变形预测
wavelet transformation
time series neural network
metro protected areas
deformation monitoring
