自适应MCKD和CEEMDAN的滚动轴承微弱故障特征提取
|
摘要
针对滚动轴承振动信号受强噪声干扰,难以提取其微弱故障特征的问题,提出了自适应最大相关峭度解卷积(MCKD)和自适应噪声完全集合经验模态分解(CEEMDAN)的故障特征提取方法。由于MCKD方法的滤波效果受滤波器长度参数的影响,故采用变步长网格搜索法对滤波器长度进行寻优,自适应地实现MCKD降噪。首先以特征能量比(FER)作为目标函数利用变步长网格搜索法寻找最优滤波器长度,通过自适应MCKD算法对振动信号进行降噪;然后采用CEEMDAN方法分解降噪信号,并根据峭度准则选取故障信息丰富的敏感固有模态分量(IMF)进行信号重构;最后利用包络谱对重构信号进行分析,提取故障特征信息。经仿真与实验分析,该方法能够有效地提取出滚动轴承的微弱故障特征信息。
With the difficulty of extracting the weak fault features of vibration signal of rolling bearing which is strongly interfered by noise,the adaptive maximum correlated kurtosis deconvolution( MCKD) and complete ensemble empirical mode decomposition with adaptive noise( CEEMDAN) are proposed. Considering that the filtering effect of the MCKD is affected by the parameter of filter length,the variable step size grid search is applied for optimizing filter length. Firstly,with the help of variable step size grid search the feature energy ratio( FER) is used as the objective function to optimize filter length and to denoise by the adaptive MCKD. Then,after the decomposition of denoising signal incurred by CEEMDAN and option of intrinsic mode function( IMF) of rich fault signals based on kurtosis criterion,the signals are reconstructed. At last,with the use of envelope spectrum analysis,the reconstructed signals are analyzed to extract fault features. It is concluded that through simulation and experimental analysis,the aforementioned method shall effectively extract the weak fault information of rolling bearing.
With the difficulty of extracting the weak fault features of vibration signal of rolling bearing which is strongly interfered by noise,the adaptive maximum correlated kurtosis deconvolution( MCKD) and complete ensemble empirical mode decomposition with adaptive noise( CEEMDAN) are proposed. Considering that the filtering effect of the MCKD is affected by the parameter of filter length,the variable step size grid search is applied for optimizing filter length. Firstly,with the help of variable step size grid search the feature energy ratio( FER) is used as the objective function to optimize filter length and to denoise by the adaptive MCKD. Then,after the decomposition of denoising signal incurred by CEEMDAN and option of intrinsic mode function( IMF) of rich fault signals based on kurtosis criterion,the signals are reconstructed. At last,with the use of envelope spectrum analysis,the reconstructed signals are analyzed to extract fault features. It is concluded that through simulation and experimental analysis,the aforementioned method shall effectively extract the weak fault information of rolling bearing.
引文
[1]武哲,杨绍普,刘永强.基于多元经验模态分解的旋转机械早期故障诊断方法[J].仪器仪表学报,2016,37(2):241-248.WU ZH,YANG SH P,LIU Y Q.Rotating machinery early fault diagnosis method based on multivariate empirical mode decomposition[J].Chinese Journal of Scientific Instrument,2016,37(2):241-248.
[2]AHN J H,KWAK D H,KOH B H.Fault detection of a roller-bearing system through the EMD of a wavelet denoised signal[J].Sensors,2014,14(8):15022-15038.
[3]徐卓飞,张海燕,王丹,等.基于EMD与KPCA的滚动轴承故障特征提取及诊断方法研究[J].机械科学与技术,2014,33(10):1518-1524.XU ZH F,ZHANG H Y,WANG D,et al.Study on feature extraction and diagnosis method of rolling bearing faults based on EMD and KPCA[J].Mechanical Science and Technology for Aerospace Engineering,2014,33(10):1518-1524.
[4]WU Z H,HUANG N E.Ensemble empirical mode decomposition:A noise-assisted data analysis method[J].Advances in Adaptive Data Analysis,2009,1(1):1-41.
[5]马超,王少红,徐小力.基于EEMD的声阵列滚动轴承故障诊断[J].电子测量与仪器学报,2017,31(9):1379-1384.MA CH,WANG SH H,XU X L.Fault diagnosis for rolling bearing by using acoustic array based on EEMD[J].Journal of Electronic Measurement and Instrumentation,2017,31(9):1379-1384.
[6]TORRES M E,COLOMINAS M A,SCHLOTTHAUERG,et al.A complete ensemble empirical mode decomposition with adaptive noise[C].IEEEInternational Conference on Acoustics,Speech and Signal Processing,2011:4144-4147.
[7]MCDONALD G L,ZHAO Q,ZUO M J.Maximum correlated kurtosis deconvolution and application on gear tooth chip fault detection[J].Mechanical Systems&Signal Processing,2012,33(1):237-255.
[8]任学平,张玉皓,邢义通,等.基于角域级联最大相关峭度反褶积的滚动轴承早期故障诊断[J].仪器仪表学报,2015,36(9):2104-2111.REN X P,ZHANG Y H,XING Y T,et al.Rolling bearing early fault diagnosis based on angular domain cascade maximum correlation kurtosis deconvolution[J].Chinese Journal of Scientific Instrument,2015,36(9):2104-2111.
[9]钟先友,赵春华,田红亮,等.基于自适应最大相关峭度解卷积和频率切片小波变换的齿轮故障特征提取[J].中国机械工程,2014,25(21):2880-2885.ZHONG X Y,ZHAO CH H,TIAN H L,et al.Gear fault feature extraction method based on adaptive MCKD and FSWT[J].China Mechanical Engineering,2014,25(21):2880-2885.
[10]王朝阁,庞震,任学平,等.ELMD和MCKD在滚动轴承早期故障诊断中的应用[J].机械科学与技术,2017,36(11):1764-1770.WANG CH G,PANG ZH,REN X P,et al.Early fault diagnosis of roller bearing based on ensemble local mean decomposition and maximum correlated kurtosis deconvolution[J].Mechanical Science and Technology for Aerospace Engineering,2017,36(11):1764-1770.
[11]吕中亮,汤宝平,周忆,等.基于网格搜索法优化最大相关峭度反卷积的滚动轴承早期故障诊断方法[J].振动与冲击,2016,35(15):29-34.LV ZH L,TANG B P,ZHOU Y,et al.Rolling bearing early fault diagnosis based on maximum correlated kurtosis deconvolution optimized with grid search algorithm[J].Journal of Vibration and Shock,2016,35(15):29-34.
[12]王健,孙志礼,于震梁,等.基于支持向量机的机械零件剩余寿命区间估计[J].东北大学学报(自然科学版),2016,37(7):974-978.WANG J,SUN ZH L,YU ZH L,et al.Renaining useful life interval estination for machine parts based on SVM[J].Journal of Northeastern University(Natural Science),2016,37(7):974-978.
[13]HE W,ZI Y,CHEN B,et al.Automatic fault feature extraction of mechanical anomaly on induction motor bearing using ensemble super-wavelet transform[J].Mechanical Systems&Signal Processing,2015,54-55:457-480.
[14]张建伟,侯鸽,暴振磊,等.基于CEEMDAN与SVD的泄流结构振动信号降噪方法[J].振动与冲击,2017,36(22):138-143.ZHANG J W,HOU G,BAO ZH L,et al.A signal denoising method for vibration signals from flood discharge structures based on CEEMDAN and SVD[J].Journal of Vibration and Shock,2017,36(22):138-143.
[15]吴小涛,杨锰,袁晓辉,等.基于峭度准则EEMD及改进形态滤波方法的轴承故障诊断[J].振动与冲击,2015,34(2):38-44.WU X T,YANG M,YUAN X H,et al.Bearing fault diagnosis using EEMD and improved morphological filtering method based on kurtosis criterion[J].Journal of Vibration and Shock,2015,34(2):38-44.
[2]AHN J H,KWAK D H,KOH B H.Fault detection of a roller-bearing system through the EMD of a wavelet denoised signal[J].Sensors,2014,14(8):15022-15038.
[3]徐卓飞,张海燕,王丹,等.基于EMD与KPCA的滚动轴承故障特征提取及诊断方法研究[J].机械科学与技术,2014,33(10):1518-1524.XU ZH F,ZHANG H Y,WANG D,et al.Study on feature extraction and diagnosis method of rolling bearing faults based on EMD and KPCA[J].Mechanical Science and Technology for Aerospace Engineering,2014,33(10):1518-1524.
[4]WU Z H,HUANG N E.Ensemble empirical mode decomposition:A noise-assisted data analysis method[J].Advances in Adaptive Data Analysis,2009,1(1):1-41.
[5]马超,王少红,徐小力.基于EEMD的声阵列滚动轴承故障诊断[J].电子测量与仪器学报,2017,31(9):1379-1384.MA CH,WANG SH H,XU X L.Fault diagnosis for rolling bearing by using acoustic array based on EEMD[J].Journal of Electronic Measurement and Instrumentation,2017,31(9):1379-1384.
[6]TORRES M E,COLOMINAS M A,SCHLOTTHAUERG,et al.A complete ensemble empirical mode decomposition with adaptive noise[C].IEEEInternational Conference on Acoustics,Speech and Signal Processing,2011:4144-4147.
[7]MCDONALD G L,ZHAO Q,ZUO M J.Maximum correlated kurtosis deconvolution and application on gear tooth chip fault detection[J].Mechanical Systems&Signal Processing,2012,33(1):237-255.
[8]任学平,张玉皓,邢义通,等.基于角域级联最大相关峭度反褶积的滚动轴承早期故障诊断[J].仪器仪表学报,2015,36(9):2104-2111.REN X P,ZHANG Y H,XING Y T,et al.Rolling bearing early fault diagnosis based on angular domain cascade maximum correlation kurtosis deconvolution[J].Chinese Journal of Scientific Instrument,2015,36(9):2104-2111.
[9]钟先友,赵春华,田红亮,等.基于自适应最大相关峭度解卷积和频率切片小波变换的齿轮故障特征提取[J].中国机械工程,2014,25(21):2880-2885.ZHONG X Y,ZHAO CH H,TIAN H L,et al.Gear fault feature extraction method based on adaptive MCKD and FSWT[J].China Mechanical Engineering,2014,25(21):2880-2885.
[10]王朝阁,庞震,任学平,等.ELMD和MCKD在滚动轴承早期故障诊断中的应用[J].机械科学与技术,2017,36(11):1764-1770.WANG CH G,PANG ZH,REN X P,et al.Early fault diagnosis of roller bearing based on ensemble local mean decomposition and maximum correlated kurtosis deconvolution[J].Mechanical Science and Technology for Aerospace Engineering,2017,36(11):1764-1770.
[11]吕中亮,汤宝平,周忆,等.基于网格搜索法优化最大相关峭度反卷积的滚动轴承早期故障诊断方法[J].振动与冲击,2016,35(15):29-34.LV ZH L,TANG B P,ZHOU Y,et al.Rolling bearing early fault diagnosis based on maximum correlated kurtosis deconvolution optimized with grid search algorithm[J].Journal of Vibration and Shock,2016,35(15):29-34.
[12]王健,孙志礼,于震梁,等.基于支持向量机的机械零件剩余寿命区间估计[J].东北大学学报(自然科学版),2016,37(7):974-978.WANG J,SUN ZH L,YU ZH L,et al.Renaining useful life interval estination for machine parts based on SVM[J].Journal of Northeastern University(Natural Science),2016,37(7):974-978.
[13]HE W,ZI Y,CHEN B,et al.Automatic fault feature extraction of mechanical anomaly on induction motor bearing using ensemble super-wavelet transform[J].Mechanical Systems&Signal Processing,2015,54-55:457-480.
[14]张建伟,侯鸽,暴振磊,等.基于CEEMDAN与SVD的泄流结构振动信号降噪方法[J].振动与冲击,2017,36(22):138-143.ZHANG J W,HOU G,BAO ZH L,et al.A signal denoising method for vibration signals from flood discharge structures based on CEEMDAN and SVD[J].Journal of Vibration and Shock,2017,36(22):138-143.
[15]吴小涛,杨锰,袁晓辉,等.基于峭度准则EEMD及改进形态滤波方法的轴承故障诊断[J].振动与冲击,2015,34(2):38-44.WU X T,YANG M,YUAN X H,et al.Bearing fault diagnosis using EEMD and improved morphological filtering method based on kurtosis criterion[J].Journal of Vibration and Shock,2015,34(2):38-44.