基于VMD-KPCA-LSTM的桥梁监测应变数据预测  

Prediction of Bridge Monitoring Strain Data Based on VMD-KPCA-LSTM

作　　者：张希望 朱前坤[1,2] 王宪玉 杜永峰[1,2] ZHANG Xiwang;ZHU Qiankun;WANG Xianyu;DU Yongfeng(Institute of Earthquake Protection and Disaster Mitigation,Lanzhou University of Technology,Lanzhou 730050,China;Western Engineering Research Center of Disaster Mitigation in Civil Engineering of Ministry of Education,Lanzhou University of Technology,Lanzhou 730050,China;Gansu Provincial Transportation Planning Survey&Design Institute Co.,Ltd.,Lanzhou 730030,China)

机构地区：[1]兰州理工大学防震减灾研究所,甘肃兰州730050 [2]兰州理工大学西部土木工程防灾减灾教育部工程研究中心,甘肃兰州730050 [3]甘肃省交通规划勘察设计院股份有限公司,甘肃兰州730030

出　　处：《应用基础与工程科学学报》2025年第1期76-86,共11页Journal of Basic Science and Engineering

基　　金：国家自然科学基金项目(52168041);甘肃省重点研发计划-工业类资助项目(22YF11GA301)。

摘　　要：桥梁结构健康监测系统在采集数据时会受到各种干扰,数据异常时有发生,难以反应桥梁真实的健康状况.针对数据异常情况,提出了结合变分模态分解(Variational Mode Decomposition,VMD)、核主成分分析(Kernel Principal Component Analysis,KPCA)以及长短期记忆神经网络(Long Short-Term Memory neural network,LSTM)的异常数据处理方法,即VMD-KPCA-LSTM.首先,将采集到的数据通过小波降噪和3σ异常剔除进行简单的预处理;然后,利用VMD将数据分解为模态相对稳定的应变分量;再次使用KPCA进行非线性降维;最后,进行各分量的LSTM预测,整合得到总的应变重构时序.与BP模型、GRU模型、LSTM模型和VMD-PCA-LSTM模型相比,VMD-KPCA-LSTM模型的MAPE分别降低了19.948%、13.621%、11.724%、7.238%.因此,提出的VMD-KPCA-LSTM模型可以更好地用于斜拉桥应变异常数据的预测,为桥梁健康状况评估分析提供了坚实的数据基础.The bridge structural health monitoring system often encounters data anomalies due to various interferences during collection,making it hard to accurately reflect the bridge’s health.To address these anomalies,we propose a method called VMD-KPCA-LSTM,which combines Variational Mode Decomposition(VMD),Kernel Principal Component Analysis(KPCA),and Long Short-Term Memory neural networks(LSTM).First,the collected data undergoes wavelet denoising and 3σanomaly removal.Then,VMD decomposes the data into stable strain components,KPCA performs nonlinear dimensionality reduction,and LSTM predicts these components.The LSTM predictions are integrated to reconstruct the total strain time series.Compared to BP,GRU,LSTM,and VMD-PCA-LSTM models,the VMD-KPCA-LSTM model reduces MAPE by 19.948%,13.621%,11.724%,and 7.238%respectively.Thus,the VMD-KPCA-LSTM model is more effective for predicting strain anomaly data in cable-stayed bridges,providing a reliable data foundation for bridge health assessment and analysis.

关 键 词：桥梁工程 健康监测 变分模态分解 核主成分分析 长短期记忆神经网络 数据预测 

分 类 号：U446.2[建筑科学—桥梁与隧道工程]