基于变分模态分解和复杂度分析的水电机组振动信号特征提取
|
摘要
目前大型水电机组通常安装有状态监测系统可记录机组的振动数据,而如何从海量的数据中提取出机组的故障特征是水电机组故障诊断的难点和热点。提出了一种基于变分模态分解和复杂度分析的振动信号特征提取方法,该方法首先对降噪后的振动信号进行变分模态分解,再结合复杂度算法求得各模态分量的复杂度值,得到以各模态分量复杂度值为元素的反映机组故障信息的特征向量,最后利用支持向量机对特征向量进行分类。试验结果表明:基于变分模态分解与复杂度分析的特征提取方法对水电机组不同运行状态具有较好的区分度,是一种有效的振动信号特征提取方法。
At present,large-scale hydropower units have installed condition monitoring systems,and how to extract the fault characteristics of the unit from the massive vibration data is the difficulty and hot spot of the fault diagnosis of hydropower units.This paper presents a method combining variational mode decomposition with complexity analysis for feature extraction. Firstly,it performs variational mode decomposition on denoised vibration signal. Then,the complexity value of each mode component through complexity algorithm and feature vectors consisting of each mode component complexity value characteristic elements are obtained. Finally,the feature vectors are classified by support vector machines. The results show that the feature extraction method based on the combination of variational mode decomposition and complexity has a good discrimination degree for different operating states of hydropower units,and it is a reasonable feature extraction method.
At present,large-scale hydropower units have installed condition monitoring systems,and how to extract the fault characteristics of the unit from the massive vibration data is the difficulty and hot spot of the fault diagnosis of hydropower units.This paper presents a method combining variational mode decomposition with complexity analysis for feature extraction. Firstly,it performs variational mode decomposition on denoised vibration signal. Then,the complexity value of each mode component through complexity algorithm and feature vectors consisting of each mode component complexity value characteristic elements are obtained. Finally,the feature vectors are classified by support vector machines. The results show that the feature extraction method based on the combination of variational mode decomposition and complexity has a good discrimination degree for different operating states of hydropower units,and it is a reasonable feature extraction method.
引文
[1]潘罗平.基于健康评估和劣化趋势预测的水电机组故障诊断系统研究[D].北京:中国水利水电科学研究院,2013.
[2]卢娜.基于多小波的水电机组振动特征提取及故障诊断方法研究[D].武汉:武汉大学,2014.
[3]付文龙.水电机组振动信号分析与智能故障诊断方法研究[D].武汉:华中科技大学,2016.
[4]李曦.基于小波包分析和支持向量机的齿轮箱故障诊断方法研究[D].太原:太原理工大学,2013.
[5] Huang N E,Shen Z,Long S R,et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis[J]. Proceedings of the Royal Society A Mathematical Physical&Engineering Sciences,1998,454(1971):903-995.
[6]杨永锋,吴亚锋.经验模态分解在振动分析中的应用[M].北京:国防工业出版社,2013.
[7] Dragomiretskiy K,Zosso D. Variational mode eecomposition[J].IEEE Transactions on Signal Processing,2014,62(3):531-544.
[8] An X,Yang J. Denoising of hydropower unit vibration signal based on variational mode decomposition and approximate entropy[J]. Transactions of the Institute of Measurement&Control,2016,38(3).
[9]马增强,李亚超,刘政,等.基于变分模态分解和Teager能量算子的滚动轴承故障特征提取[J].振动与冲击,2016,(13):134-139.
[10]随瑞斌.轻度认知功能障碍安静及认知状态脑电复杂度研究[D].天津:天津医科大学,2015.
[11]吕中亮.基于变分模态分解与优化多核支持向量机的旋转机械早期故障诊断方法研究[D].重庆:重庆大学,2016.
[2]卢娜.基于多小波的水电机组振动特征提取及故障诊断方法研究[D].武汉:武汉大学,2014.
[3]付文龙.水电机组振动信号分析与智能故障诊断方法研究[D].武汉:华中科技大学,2016.
[4]李曦.基于小波包分析和支持向量机的齿轮箱故障诊断方法研究[D].太原:太原理工大学,2013.
[5] Huang N E,Shen Z,Long S R,et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis[J]. Proceedings of the Royal Society A Mathematical Physical&Engineering Sciences,1998,454(1971):903-995.
[6]杨永锋,吴亚锋.经验模态分解在振动分析中的应用[M].北京:国防工业出版社,2013.
[7] Dragomiretskiy K,Zosso D. Variational mode eecomposition[J].IEEE Transactions on Signal Processing,2014,62(3):531-544.
[8] An X,Yang J. Denoising of hydropower unit vibration signal based on variational mode decomposition and approximate entropy[J]. Transactions of the Institute of Measurement&Control,2016,38(3).
[9]马增强,李亚超,刘政,等.基于变分模态分解和Teager能量算子的滚动轴承故障特征提取[J].振动与冲击,2016,(13):134-139.
[10]随瑞斌.轻度认知功能障碍安静及认知状态脑电复杂度研究[D].天津:天津医科大学,2015.
[11]吕中亮.基于变分模态分解与优化多核支持向量机的旋转机械早期故障诊断方法研究[D].重庆:重庆大学,2016.