基于EMD的重力坝变形监测统计模型及应用

Statistical Model of Deformation Monitoring of Gravity Dam Based on EMD and Its Application

监测时段的选取和统计模型各分量的构造形式会影响传统重力坝变形监测统计模型的水压分量、温度分量、时效分量的分离结果。采用经验模态分解(EMD)方法将原始监测效应量自适应地分解产生非周期性的时效分量,建立基于EMD的重力坝变形监测统计模型,将其应用于YL重力坝9#坝段位移测点。选取HTT、HST和HT_(pca)T统计模型作为传统统计模型,采用不同的评价指标对4种模型的精度和合理性进行对比。YL重力坝9#坝段位移监测结果表明,基于EMD的重力坝变形监测统计模型的预测精度更高,水压分量、温度分量和时效分量的分离更合理。

The selection of monitoring period and the structural form of each component of statistical model will affect the separation results of water pressure component,temperature component and aging component of traditional gravity dam deformation monitoring statistical model.The empirical mode decomposition(EMD)method was used to adaptively decompose the original monitoring effect into non⁃periodic time⁃effect components,and the statistical model of gravity dam deformation monitoring based on EMD was established,which was applied to the displacement measuring point of YL gravity dam No.9 dams⁃ection.HTT,HST and HT_(pca)T statistical models were selected as the traditional statistical models,and the accuracy and rationality of the four models were compared by different evaluation indexes.The displacement moni⁃toring results of No.9 dam⁃section of YL gravity dam shows that the prediction accuracy of statistical model based on EMD is higher,and the separation of water pressure component,temperature component and aging components is more reasonable.