基于LCD-Hilbert谱奇异值和QRVPMCD的滚动轴承故障诊断方法

PDF(1169 KB)

振动与冲击 ›› 2015, Vol. 34 ›› Issue (7) : 121-126.

论文

针对多变量预测模型的模式识别（Variable predictive model based class discriminate，简称VPMCD）方法在参数估计中存在的缺陷，采用分位数回归（Quantile Regression，简称QR）代替原方法中的最小二乘法进行参数估计，克服最小二乘回归中强假设、易受异常值影响等问题，以此提高模式识别的精度。因此，提出了基于分位数回归的多变量预测模型模式识别方法（Quantile Regression -Variable predictive mode based class discriminate ，简称QRVPMCD）。采用局部特征尺度分解（Local characteristic-scale decomposition，简称LCD）方法对滚动轴承振动信号进行分解得到若干个单分量信号，提取单分量信号的Hilbert谱奇异值组成故障特征向量，并以此作为QRVPMCD的输入进行滚动轴承故障诊断。对不同工作状态和故障类型下的滚动轴承振动信号进行了分析，结果表明了该方法的有效性。

作者信息 +

The rolling bearing fault diagnosis method based on the Hilbert spectrum singular value and QRVPMCD

Targeting the defects in the parameter estimation of VPMCD (Variable predictive model -based class discriminate), Quantile Regression (QR) is used for parameter estimation instead of least-squares method in the original method. The questions such as strong assumptions and easily affected by the outliers in the Ordinary Least-Square Regression could be overcome by QR so as to improve the accuracy of pattern recognition. Therefore, the Quantile Regression-Variable predictive mode based on class discriminate (QRVPMCD) was proposed in this paper. The Local characteristic-scale decomposition (LCD) is used to decompose the rolling bearing vibration signal into several mono-component signals, and then the Hilbert spectrum singular values were extracted from the mono-component signals and formed into fault feature vector, which can be used as input of QRVPMCD for rolling bearing fault diagnosis. The analysis results from different working conditions and failures of roller bearing demonstrate the effectiveness of the proposed method.

Author information +

文章历史 +

摘要

针对多变量预测模型的模式识别（Variable predictive model based class discriminate，简称VPMCD）方法在参数估计中存在的缺陷，采用分位数回归（Quantile Regression，简称QR）代替原方法中的最小二乘法进行参数估计，克服最小二乘回归中强假设、易受异常值影响等问题，以此提高模式识别的精度。因此，提出了基于分位数回归的多变量预测模型模式识别方法（Quantile Regression -Variable predictive mode based class discriminate ，简称QRVPMCD）。采用局部特征尺度分解（Local characteristic-scale decomposition，简称LCD）方法对滚动轴承振动信号进行分解得到若干个单分量信号，提取单分量信号的Hilbert谱奇异值组成故障特征向量，并以此作为QRVPMCD的输入进行滚动轴承故障诊断。对不同工作状态和故障类型下的滚动轴承振动信号进行了分析，结果表明了该方法的有效性。

Abstract

Targeting the defects in the parameter estimation of VPMCD (Variable predictive model -based class discriminate), Quantile Regression (QR) is used for parameter estimation instead of least-squares method in the original method. The questions such as strong assumptions and easily affected by the outliers in the Ordinary Least-Square Regression could be overcome by QR so as to improve the accuracy of pattern recognition. Therefore, the Quantile Regression-Variable predictive mode based on class discriminate (QRVPMCD) was proposed in this paper. The Local characteristic-scale decomposition (LCD) is used to decompose the rolling bearing vibration signal into several mono-component signals, and then the Hilbert spectrum singular values were extracted from the mono-component signals and formed into fault feature vector, which can be used as input of QRVPMCD for rolling bearing fault diagnosis. The analysis results from different working conditions and failures of roller bearing demonstrate the effectiveness of the proposed method.

导出引用

杨宇，何知义，潘海洋，程军圣. 基于LCD-Hilbert谱奇异值和QRVPMCD的滚动轴承故障诊断方法[J]. 振动与冲击, 2015, 34(7): 121-126

YNAG Yu HE Zhi-yi PAN Hai-yang CHENG Jun-sheng. The rolling bearing fault diagnosis method based on the Hilbert spectrum singular value and QRVPMCD[J]. Journal of Vibration and Shock, 2015, 34(7): 121-126

参考文献

[1] Chun Chieh Wang, Yuan Kang,Pingchen Shen,et al. Applications of fault diagnosis in rotating machinery by using time series analysis with neural network [J]. Expert Systems with Applications, 2010, 37(2):1696-1702.
[2] Wang Huaqing, Chen Peng, Intelligent diagnosis method for rolling element bearing faults using possibility theory and neural network [J]. Computers & Industrial Engineering, 2011, 60(4): 511-518.
[3] Breiman L. Heuristics of instability and stabilization in model selection [J].The Annals of Statistics, 1996, 24(4): 2350-2383.
[4] 杨宇，李杰，潘海洋等.VPMCD和改进ITD的联合智能诊断方法研究[J]. 振动工程学报.2013,26(4):608-616.
Yang Yu,Li Jie,Pan Haiyang,et al.Research on Combined intelligent diagnostic method based on VPMCD and improved ITD[J]. Chinese Journal of Mechanical Engineering, 2013.26(4):608-616
[5] Koenker R,Bassett G W.Regression quantiles [J].Econometrica,1978,46:33-50.
[6] 程军圣,郑近德,杨宇.一种新的非平稳信号分析方法—局部特征尺度分解[J].振动工程学报.2012, 25(2):215-220.
Cheng Junsheng, Zheng Jinde, Yang Yu. A nonstationary signal analysis approach—the local characteristic-scale decomposition method[J].Journal of Vibration Engineering. 2012, 25(2): 215-220.
[7] Cheng J S, Yu D J, Yang Y. Energy operator demodulating approach based on EMD and its application in mechanical fault diagnosis［J］.Chinese Journal of Mechanical Engineering, 2004,40( 8):115-118．
[8] Huang N E, Wu M-L C and Long S R. A confidence limit for the empirical mode decomposition and Hilbert spectral analysis [J]. Proceedings of the Royal Society of London. A, 2003, 459: 2317-2345.
[9] Mark G. Frei, Ivan Osorio. Intrinsic time-scale decomposition: time-frequency-energy analysis and real-time filtering of non-stationary signals [J]. Proceedings of the Royal Society. A, 2007, 463: 321-342.
[10] Koenker R, D’Orey V.A Remark on Computing Regression Quantiles[J].Applied Statistics, 1993,43:410-440.
[11] Yang Yu, Yu Dejie, Cheng Junsheng. A roller bearing fault diagnosis method based on EMD energy entropy and ANN[J]. Journal of Sound and Vibration, 2006, 294: 269-277.
[12] Mardia KV, Kent JT, Bibby JM. Multivariate analysis[M]. London: Academic Press. 1979. 322, 381.
[13] Yu-Dong Cai, Pong-Wong Ricardo, Chih-Hung Jen, et al. Application of SVM to predict membrane protein types [J]. Journal of Theoretical Biology, 2004,226: 373–376.