摘要
变形分析与预报是工程建(构)物在施工与运营期间的重要内容,目前应用较为广泛的是针对变形体各测点建立时间序列模型(ARMA),这种建模方法考虑的是各测点位移在时间变化上的关联性,而时空序列模型(STARMA)则同时考虑测点在时间以及空间上的相关性,从理论上来讲,能够更好的解释变形体的形变规律。本文以某地下管线沉降监测为研究对象,分别建立ARMA模型以及STARMA模型,通过计算各测点预测RSE、NMSE、RMSE、MAE四个误差指标值并进行比较,验证了STARMA模型在预测精度上好于ARMA模型,对于管线沉降监测具有一定的应用价值。
Analysis and prediction of deformation is an important part of engineering during construction and operation,and it is widely used to establish time series model for each measuring point of deformable body at present.This method considers that displacement of each point is related to the past,while space-time series model takes both the time correlation and spatial correlation into account,and it can reflect the deformation law better.In this paper,the settlement monitoring of an underground pipeline is taken as the research object,we both build ARMA model and STARMA model and have a comparison,according to prediction and comparison of four error Indices of RSE、NMSE、RMSE、MAE by calculating the measuring points,results show that STARMA model is better than ARMA model on prediction accuracy,and it has certain application value for pipeline settlement monitoring.
作者
柳新强
王涛
LIU Xinqiang;WANG Tao(Shaanxi Railway Institute,Weinan Shaanxi,714099 China;College of Geology Engineering and Geomatics,Changan University,Xi＇an Shaanxi710054,China)
出处
《北京测绘》
2018年第7期809-813,共5页
Beijing Surveying and Mapping
基金
陕西铁路工程职业技术学院科研基金项目(KY2016-47)
