摘要
为更加准确地反映出大坝变形数据的内部规律,及时对大坝稳定性和安全度做出评价并采取有效措施,选择广州市李溪拦河坝为研究对象,搜集了4#静力水准仪2015年9月—2019年3月共753组数据.选取了2015年9月—2017年12月共355组数据作为模型训练集建立模型,2018年1月—2018年12月328组数据作为模型验证集,用于评判模型性能;2019年1月—3月70组数据作为模型测试集,用于模型测试.分别采用逐步回归分析模型和时间序列模型进行建模及验证分析.样本测试集预测结果表明,逐步回归分析模型测试集均方差为0.022、决定系数为0.951;时间序列模型测试集均方差为0.007、决定系数为0.985.表明运用时间序列模型在拟合、预测和误差分析方面优于逐步回归模型.
In order to reflect the internal law of dam deformation data more accurately,the stability and safety of the dam are evaluated and effective measures are taken in time.The Lixi dam in Guangzhou was selected as the research object.753 sets of data of 4#hydrostatic level from September 2015 to March 2019 were collected,355 sets of data from September 2015 to December 2017 were selected as the training set to establish the model,and 328 sets of data from January 2018 to December 2018 were used as the model verification set to evaluate the performance of the model,From January 2019 to March 2019,70 sets of data were used as model test sets.Stepwise regression analysis model and time series model were used to model and verify the analysis.The forecasting results of sample test set show that the mean square error of stepwise regression analysis model test set is 0.022 and the determinant coefficient is 0.951;the mean square error of time series model test set is 0.007 and the determinant coefficient is 0.985.The results show that the time series model is better than the stepwise regression model in fitting,forecasting and error analysis.
作者
江显群
陈武奋
邵金龙
黄钲武
JIANG Xianqun;CHEN Wufen;SHAO Jinlong;HUANG Zhengwu(The Pearl River Hydraulic Research Institute,Pearl River Water Resources Commission,Guangzhou,Guangdong 510610,China)
出处
《排灌机械工程学报》
EI
CSCD
北大核心
2019年第10期870-874,920,共6页
Journal of Drainage and Irrigation Machinery Engineering
基金
广州市科技计划项目(201604020049)
关键词
大坝变形
逐步回归分析
时间序列模型
位移预测
dam deformation
stepwise regression analysis
time series model
displacement prediction
作者简介
第一作者:江显群(1976-),男,江西黎川人,高级工程师(scut_jiang@163.com),主要从事水利自动化与信息化研究;通信作者:陈武奋(1988-),男,广东陆丰人,工程师(chen.wufen@163.com),主要从事水利自动化与信息化、人工智能、深度学习研究.
