摘要
针对桥梁监测数据受多重噪声干扰、影响结构真实响应获取的问题,提出了一种基于时变滤波经验模态分解(time-varying filtering empirical mode decomposition,简称TVFEMD)和本征模函数(intrinsic mode function,简称IMF)能量熵增量的桥梁监测数据降噪方法。首先,利用TVFEMD分解桥梁原始监测数据,得到多个子序列;其次,采用IMF能量熵增量确定多个子序列中的有效子序列;然后,划分子序列中的结构响应分量和噪声分量,对结构响应分量重组实现监测数据降噪;最后,利用平均绝对误差(mean absolute error,简称MAE)、均方根误差(root mean squared error,简称RMSE)和信噪比(signal-noise ratio,简称SNR)对不同方法的降噪效果进行评价。仿真算例和工程实例结果表明:TVFEMD相比经验模态分解(empirical mode decomposition,简称EMD),有效解决了模态混叠问题;TVFEMD结合IMF能量熵增量方法,有效抑制了多重噪声影响,对结果精度有较大提升;与EMD-IMF能量熵增量和Kalman滤波降噪法相比,TVFEMD-IMF能量熵增量法所得到降噪信号的MAE和RMSE值分别提升了23%和21%以上,降噪效果更好,信噪比提升38%以上,抗噪性能更佳。
Bridge monitoring data suffers from noise interference,which affects the acquisition of the true response.However,traditional empirical mode decomposition methods have limited de-noise effects.In order to enhance the de-noise effect of monitoring data,a bridge monitoring data de-noise method based on time-varying filtered empirical mode decomposition(TVFEMD)and intrinsic mode function(IMF)energy entropy increment is proposed.Firstly,the bridge monitoring data is decomposed using the TVFEMD to obtain a number of subseries.After that,the IMF energy entropy increment is used to determine the effective subsequence among several subseries.Then,the effective subsequences are recombined to achieve de-noise in the monitoring data.Finally,the effectiveness of the proposed method in terms of de-noise is evaluated using the mean absolute error(MAE),root mean square error(RMSE),and signal-noise ratio(SNR).The results of the simulation and engineering examples show that:The TVFEMD effectively solves the problem of mode mixing in the empirical mode decomposition(EMD);The combination of the TVFEMD with the IMF energy entropy increment method,which effectively suppresses the effects of multiple noise and provides significant improvements to the accuracy of the results;Compared with the EMD and the Kalman filtering methods,the MAE and the RMSE values are improved by more than 23%and 21%,respectively,which indicates that the proposed method is more effective in de-noise.The SNR value is improved by more than 38%,which demonstrates that the proposed method has superior noise immunity.
作者
李双江
辛景舟
蒋黎明
刘水康
巴建明
周建庭
LI Shuangjiang;XIN Jingzhou;JIANG Liming;LIU Shuikang;BA Jianming;ZHOU Jianting(State Key Laboratory of Mountain Bridge and Tunnel Engineering,Chongqing Jiaotong University Chongqing,400074,China;School of Civil Engineering,Chongqing Jiaotong University Chongqing,400074,China;China Gezhouba Group No.2 Engineering Co.,Ltd.Chengdu,610091,China)
出处
《振动．测试与诊断》
EI
CSCD
北大核心
2024年第1期178-185,206,共9页
Journal of Vibration,Measurement & Diagnosis
基金
国家自然科学基金资助项目(52278292)
重庆市杰出青年科学基金资助项目(CSTB2023NSCQ-JQX0029)
贵州省交通运输厅科技资助项目(2023-122-001)
重庆交通大学研究生科研创新资助项目(CYB23246)。
关键词
桥梁
健康监测
降噪
时变滤波经验模态分解
本征模函数能量熵增量
bridge
health monitoring
de-noise
time-varying filtered empirical mode decomposition
intrinsic mode function energy entropy increment
