基于SCSO-BIGRU的大坝变形预测  

Dam deformation prediction based on SCSO-BIGRU

作　　者：谢逸丰 XIE Yifeng(School of Civil and Surveying&Mapping Engineering,Jiangxi University of Science and Technology,Ganzhou,Jiangxi 341000,China)

机构地区：[1]江西理工大学土木与测绘工程学院,江西赣州341000

出　　处：《北京测绘》2024年第10期1482-1486,共5页Beijing Surveying and Mapping

摘　　要：针对传统预测模型调参难度大及预测精度不够高的问题,本文将沙猫群算法(SCSO)和双向门控循环单元(BIGRU)引入到大坝变形分析中,通过SCSO对BIGRU模型的参数进行自动寻优,构建了一种基于SCSO-BIGRU的大坝变形预测模型。以某大坝垂直方向监测数据为例,对SCSO-BIGRU模型和对比模型进行分析,结果表明,SCSO-BIGRU模型预测精度更高,其得到的均方根误差(RMSE)sRMSE、平均绝对误差(MAE)sMAE、相关系数R2分别为0.1115、0.1422、0.9971,各项精度评价指标均优于BIGRU模型和门控循环单元(GRU)模型,可为大坝变形精准预测提供参考。In view of the difficulty of tuning the traditional prediction model parameters and the lack of high prediction accuracy,this paper introduced the sand cat swarm optimization(SCSO)and the bidirectional gated recurrent unit(BIGRU)into the dam deformation analysis and constructed a dam deformation prediction model based on SCSO-BIGRU by automatically optimizing the parameters of the BIGRU model through SCSO.By taking the vertical monitoring data of a dam as an example,the SCSO-BIGRU model and the comparison model were analyzed.The results show that the SCSO-BIGRU model has higher prediction accuracy,and the root mean square error(RMSE),average absolute error(MAE),and correlation coefficient R2 obtained are 0.1115,0.1422,and 0.9971,respectively.The accuracy evaluation indexes are better than those of the BIGRU model and the GRU model,which can provide a reference for the accurate prediction of dam deformation.

关 键 词：沙猫群优化算法 双向门控循环单元 大坝变形预测 预测精度 

分 类 号：P258[天文地球—测绘科学与技术] TV698.11[水利工程—水利水电工程]