摘要
在基坑施工过程中为保证地铁及其周边建筑物的安全,地铁基坑的变形预测变得越来越重要。ARMA模型作为一种时间序列分析模型,在地铁基坑监测序列中常常表现出较大的趋势项,降低了ARMA的预测精度。基于BP神经网络良好的拟合能力,提取基坑监测序列的趋势项,将剩余项建立ARMA模型,对基坑监测序列进行高精度的变形预测。改进ARMA模型提高了原有ARMA模型的预测精度,为地铁基坑的预测分析提供了较好的技术参考。
In order to ensure the safety of the subway and its surrounding buildings during the construction of the foundation pit,the deformation prediction of the subway foundation pit becomes more and more important.However,as the time series analysis model,ARMA model in the subway pit monitoring sequence often shows a great trend item,the prediction accuracy of ARMA is often lowered.Based on the good fitting capability of the BP neural network,and upon the basis of extracting the trend items of the pit monitoring sequence,an ARMA model for the remaining items is established in the present paper precisely to predict the deformation of the pit.The improved ARMA model presented in this paper improves the prediction accuracy of the former ARMA model and provides a good technical reference for the prediction analysis of subway foundation pits.
作者
刘畅
LIU Chang(The 4th Engineering Co.Ltd.of the 18th Bureau Group of China Railway,Tianjin 300000,China)
出处
《国防交通工程与技术》
2020年第1期25-27,共3页
Traffic Engineering and Technology for National Defence
关键词
地铁基坑
时间序列
BP神经网络
变形预测
subway foundation pit
time series
BP neural network
deformation prediction
