基于VMD_IWT近似熵与PSO_SVM的转子故障诊断
|
摘要
针对转子故障振动信号的非平稳性、非线性特征,提出了基于变分模态分解_改进小波阈值(VMD_IWT)近似熵与粒子群优化支持向量机(PSO_SVM)的转子故障诊断方法。首先对振动信号进行VMD分解,对分解后的各分量进行改进小波阈值处理;然后,提取降噪后信号的近似熵作为特征值组成特征向量;最后,将特征向量输入PSO_SVM进行故障分类识别。将该方法用于实际转子实验数据,并通过对比分析,分别证明了VMD、IWT和PSO_SVM方法的有效性,且文中所提方法的故障诊断准确率高达95%,证明该方法具有一定的实际应用价值。
Aiming at non-stationary and non-linear characteristics of rotor fault vibration signals, a method was proposed using support vector machine(SVM) with particle swarm optimization(PSO) algorithm on the basis of variational mode decomposition and improved wavelet threshold(VMD_IWT) approximate entropy.Firstly, the signals was decomposed into several intrinsic mode functions(IMF) and each of them was processed with improved wavelet threshold method. Then, the approximate entropy of the signals which were reconstructed with IMFs after improved wavelet threshold method was extracted as feature vectors for pattern recognition. Finally, the feature vector was input to the PSO_ SVM for fault diagnosis of rotor system. The method is applied to the actual rotor experimental data, and the effectiveness of the VMD, IWT and PSO_SVM methods is proved by comparison analysis. The experimental results show that the proposed method can classify the fault type of rotor with high accuracy which is 95%, which proving that the proposed method has practical value in some degree.
Aiming at non-stationary and non-linear characteristics of rotor fault vibration signals, a method was proposed using support vector machine(SVM) with particle swarm optimization(PSO) algorithm on the basis of variational mode decomposition and improved wavelet threshold(VMD_IWT) approximate entropy.Firstly, the signals was decomposed into several intrinsic mode functions(IMF) and each of them was processed with improved wavelet threshold method. Then, the approximate entropy of the signals which were reconstructed with IMFs after improved wavelet threshold method was extracted as feature vectors for pattern recognition. Finally, the feature vector was input to the PSO_ SVM for fault diagnosis of rotor system. The method is applied to the actual rotor experimental data, and the effectiveness of the VMD, IWT and PSO_SVM methods is proved by comparison analysis. The experimental results show that the proposed method can classify the fault type of rotor with high accuracy which is 95%, which proving that the proposed method has practical value in some degree.
引文
[1] 左红艳,刘晓波,洪连环.用于转子故障诊断的改进小波聚类算法[J].振动、测试与诊断,2018,38(2):320-326,420.
[2] 向玲,张力佳.变分模态分解在转子故障诊断中的应用[J].振动、测试与诊断,2017,37(4):793-799,847.
[3] 郑近德,潘海洋,潘紫微,等.自适应无参经验小波变换及其在转子故障诊断中的应用[J].中国机械工程,2016,27(16):2218-2224.
[4] 崔心瀚,马立元,魏忠林,等.EEMD模糊熵和变量预测模型的转子故障诊断新方法[J].内燃机工程,2016,37(4):84-91.
[5] 刘尚坤,唐贵基,王晓龙.基于改进变分模态分解的旋转机械故障时频分析方法[J].振动工程学报,2016,29(6):1119-1126.
[6] Dragomiretskiy K,Zosso D.Variational mode decomposition[J].IEEE Tran on Signal Processing,2014,62(3):531-544.
[7] Ali A A.A new intelligent method based on combination of VMD and ELM for short term wind power forecasting[J].Neurocomputing,2016,203:111-120.
[8] 何巍,袁亮,章翔峰.改进小波去噪-Teager算子的齿轮微弱故障提取方法[J].振动、测试与诊断,2018,38(1):155-213.
[9] 李学军,何能胜,何宽芳,等.基于小波包近似熵和SVM的圆柱滚子轴承诊断[J].振动、测试与诊断,2015,35(6):1031-1036.
[10] 刘思思,谭建平,易子馗.基于MFCC 和SVM 的车窗电机异常噪声辨识方法研究[J].振动与冲击,2017,36(5):102-107.
[11] Lei Ni,Juncheng Jiang,Yong Pan.Leak location of pipelines based on transient model and PSO-SVM[J].Journal of Loss Prevention in the Process Industries,2013(26):1085-1093.
[12] 张小龙,张氢,秦仙蓉,等.基于ITD复杂度和PSO-SVM的滚动轴承故障诊断[J].振动与冲击,2016,35(24):102-107,138.
[2] 向玲,张力佳.变分模态分解在转子故障诊断中的应用[J].振动、测试与诊断,2017,37(4):793-799,847.
[3] 郑近德,潘海洋,潘紫微,等.自适应无参经验小波变换及其在转子故障诊断中的应用[J].中国机械工程,2016,27(16):2218-2224.
[4] 崔心瀚,马立元,魏忠林,等.EEMD模糊熵和变量预测模型的转子故障诊断新方法[J].内燃机工程,2016,37(4):84-91.
[5] 刘尚坤,唐贵基,王晓龙.基于改进变分模态分解的旋转机械故障时频分析方法[J].振动工程学报,2016,29(6):1119-1126.
[6] Dragomiretskiy K,Zosso D.Variational mode decomposition[J].IEEE Tran on Signal Processing,2014,62(3):531-544.
[7] Ali A A.A new intelligent method based on combination of VMD and ELM for short term wind power forecasting[J].Neurocomputing,2016,203:111-120.
[8] 何巍,袁亮,章翔峰.改进小波去噪-Teager算子的齿轮微弱故障提取方法[J].振动、测试与诊断,2018,38(1):155-213.
[9] 李学军,何能胜,何宽芳,等.基于小波包近似熵和SVM的圆柱滚子轴承诊断[J].振动、测试与诊断,2015,35(6):1031-1036.
[10] 刘思思,谭建平,易子馗.基于MFCC 和SVM 的车窗电机异常噪声辨识方法研究[J].振动与冲击,2017,36(5):102-107.
[11] Lei Ni,Juncheng Jiang,Yong Pan.Leak location of pipelines based on transient model and PSO-SVM[J].Journal of Loss Prevention in the Process Industries,2013(26):1085-1093.
[12] 张小龙,张氢,秦仙蓉,等.基于ITD复杂度和PSO-SVM的滚动轴承故障诊断[J].振动与冲击,2016,35(24):102-107,138.