基于SVD与改进EMD联合的泄流结构工作特性信息提取

Research on Characteristics Information Identification for Flood Discharge Structure Based on SVD and Improved EMD

针对泄流结构振动信号非平稳性和特征信息被强噪声淹没的实际问题,提出基于奇异值分解(SVD)和改进经验模态分解(EMD)联合的特征信息提取方法.首先,对一维泄流振动信号时程进行相空间重构,运用SVD分解技术提取振动信号的奇异值信息,并通过奇异熵增量定阶理论滤除泄流振动信号中的高频噪声,实现信号的初次滤波;其次,对初次滤波信号进行正交化EMD分解,运用频谱分析方法筛选包含主要结构振动信息的各IMF,滤除低频水流噪声,实现信号的二次滤波;最后,将包含结构振动信息的IMF分量重构,得到泄流结构的工作振动特征信息.通过数值信号仿真验证本文方法的正确性,可有效滤除高频和低频噪声,凸显结构振动特征信息.结合三峡5号坝段泄流振动实测数据,运用本文方法进行坝体特征信息提取,并与ERA辨识结果进行比较,进一步说明本方法在泄流结构振动信息分析中的优越性,可为泄流结构在线监测和安全运行提供依据.

In order to figure out the actual problem of the flood discharge structure non-stationary vibration signal whose feature information submerged by strong noise, a valid operating characteristics information identification method based on Singular Value Decomposition( SVD)and improved Empirical Mode Decomposition( EMD) is proposed. Firstly,make use of the original signal to get the appropriate features matrix with the method of phase space reconstruction,a part of white noises were filtered out by SVD and singular entropy increment theory,which can reduce the interference to EMD; Then decomposing the signal with improved EMD,obtained a series of intrinsic mode functions( IMF) which contains real physical meaning; Finally reconstructed the IMF of characteristic information to achieve the de-noised signal through spectrum analysis. Constructing the simulation signal, and comparing the filtering effect of this method with SVD and EMD,study shows that it is a superior de-noising method,which can filter the vibration noise accurately and retain the characteristic information.Taking the 5th overflow section of Three Gorges Dam as object of study,the results compared with Eigensystem Realization Algorithm( ERA) show that this method is effective and has highly correct recognition accuracy,which can provide the help for safely operation and on-line dynamic non-destruction monitoring of the flood discharge structure.