基于低秩和稀疏分解的滚动轴承故障特征提取方法对比研究

PDF(3092 KB)

振动与冲击 ›› 2023, Vol. 42 ›› Issue (21) : 182-191.

论文

王冉1，黄裕春1，张军武1，余亮2

作者信息 +

Contrastive study on fault feature extraction methods for rolling bearing based on low rank and sparse decomposition

WANG Ran1, HUANG Yuchun1, ZHANG Junwu1, YU Liang2

Author information +

文章历史 +

摘要

滚动轴承是机械设备中广泛使用的关键部件，其故障特征的准确提取对设备稳定运行至关重要。轴承的初始故障很微弱，容易被背景噪声掩盖，这使故障特征的提取较为困难，需要对轴承故障特征与噪声的特性进行准确刻画。针对上述问题，为了深入探究轴承故障特征及噪声在时频域中的低秩与稀疏特性及其内在关联，本文对轴承故障特征提取低秩稀疏分解框架下的两种代表性方法开展对比研究，以便充分利用故障特征与噪声成分的性质，为噪声干扰下的轴承故障提取方法选择提供一定的依据。利用周期性瞬态冲击信号在时频域中的稀疏与低秩特性建立矩阵分解模型，对比了Go分解(Go-Decomposition, Go-Dec)和非负矩阵分解(Non-negative matrix factorization, NMF)两种具有代表性的分解方法，并将其应用于时频域中滚动轴承的故障特征提取。首先，基于短时傅里叶变换(Short Time Fourier Transform, STFT)生成振动信号的时频矩阵，并揭示了轴承故障脉冲在时频域中具有的稀疏性和低秩性。利用Go-Dec和NMF两种矩阵分解方法，分解出表征故障特征的矩阵。最后，对分解的故障矩阵采用逆短时傅里叶变换重构瞬态脉冲信号，并对该信号取包络谱从而确定滚动轴承的故障类型和频率信息。仿真分析和实验对比了两种故障特征分解方法，结果表明Go-Dec可以更好地去除噪声干扰，有效提取出表征滚动轴承故障特征的稀疏分量。

Abstract

Rolling bearing is the critical component widely used in mechanical equipment and one of the main causes of equipment failure. The accurate extraction of its fault feature is crucial to the stable operation of the equipment. The initial faults of the bearings are pretty weak, and transient signals are easily masked by background noise, which makes the extraction of fault features more difficult, the characteristics of bearing fault features and noise are required to be accurately depicted. For the above problems, to deeply explore the low-rank and sparse characteristics of bearing fault features and noise in the time-frequency domain and their intrinsic correlation, the two representative methods in the framework of low-rank and sparse decomposition for bearing fault feature extraction are compared and studied in this paper, to make full use of the properties of fault features and noise components, which provides a certain basis for selecting bearing fault extraction methods under noise disturbance. To address the above problems, the matrix decomposition model is further developed by utilizing the sparse and low-rank characteristics of periodic transient signals. Two fault feature extraction techniques, Go-Decomposition (Go-Dec) and Non-negative matrix factorization (NMF), are compared in this paper. They are applied to fault features extracted from rolling bearings in the time-frequency domain. First, the time-frequency matrix of vibration signal is generated based on Short Time Fourier Transform (STFT), and the sparsity and low rank of fault impulses are revealed in the time-frequency domain. Then, two matrix decomposition methods, Go-Dec and NMF, are used to decompose the matrix characterizing the fault features. Finally, the transient signal is recovered by inverse short-time Fourier transform of the decomposed fault matrix. The envelope spectrum of the reconstructed time-domain signal is taken to determine the fault type and frequency information of the rolling bearing. The simulated analysis and experimental validation compare the two fault feature decomposition methods in which Go-Dec can better reduce the noise interference and extract the sparse components characterizing the rolling bearing fault features.

导出引用

王冉1，黄裕春1，张军武1，余亮2. 基于低秩和稀疏分解的滚动轴承故障特征提取方法对比研究[J]. 振动与冲击, 2023, 42(21): 182-191

WANG Ran1, HUANG Yuchun1, ZHANG Junwu1, YU Liang2. Contrastive study on fault feature extraction methods for rolling bearing based on low rank and sparse decomposition[J]. Journal of Vibration and Shock, 2023, 42(21): 182-191

参考文献

[1]. 陈是扦, 彭志科, 周鹏. 信号分解及其在机械故障诊断中的应用研究综述 [J]. 机械工程学报, 2020, 56(17): 17.
CHEN Shiqian, PENG Zhike, ZHOU Peng. Review of Signal Decomposition Theory and Its Applications in Machine Fault Diagnosis[J]. Journal of Mechanical Engineering, 2020, 56(17): 17.
[2]. 罗毅, 甄立敬. 基于小波包与倒频谱分析的风电机组齿轮箱齿轮裂纹诊断方法[J].振动与冲击,2015,34(03):210-214.
LUO Yi, ZHEN Lijing. Diagnosis method of turbine gearbox gearcrack based on wavelet packet and cepstrum analysis. [J]. Journal of Vibration and Shock. 2015,34(03):210-214.
[3]. 王宏超, 陈进, 霍柏琦. 强抗噪时频分析方法及其在滚动轴承故障诊断中的应用 [J]. 机械工程学报, 2015, 51(1): 7.
WANG Hongchao, CHEN Jin, HUO Baiqi. Noise-resistant Time-frequency Analysis Method and Its Application in Fault Diagnosis of Rolling Bearing, [J]. Journal of Mechanical Engineering, 2015, 51(1): 7.
[4]. 毋文峰, 陈小虎.基于小波变换模改进Perona-Malik模型的强噪声信号滤波算法[J].振动与冲击. 2018,37(17):277-282.
WU Wenfeng, CHEN Xiaohu. Strong noise signal filtering algorithm based on wavelet transform module and modified Perona-Malik model. [J]. Journal of Vibration and Shock. 2018,37(17):277-282.
[5]. 李恒, 张氢, 秦仙蓉, 等. 基于短时傅里叶变换和卷积神经网络的轴承故障诊断方法 [J].振动与冲击 2018, 37(19): 8.
LI Heng, ZHANG Qing, QIN Xianrong, et al. Fault diagnosis method for rolling bearings based on short-time Fourier transform and convolution neural network[J]. Journal of Vibration and Shock. 2018, 37(19): 8.
[6]. ZHANG ZiWei, HUANG Weiguo, Liao Yi, et al. Bearing fault diagnosis via generalized logarithm sparse regularization [J]. Mechanical Systems and Signal Processing, 2022, 167: 108576.
[7]. YAO Renhe, JIANG Hongkai, WU Zhenghong, et al. Periodicity-enhanced sparse representation for rolling bearing incipient fault detection [J]. ISA Transactions, 2021, 118: 219-37.
[8]. YANG Hongyu, MATHEW J, MA Li. Fault diagnosis of rolling element bearings using basis pursuit [J]. Mechanical Systems and Signal Processing, 2005, 19(2): 341-56.
[9]. WANG Ran, ZHANG Junwu, FANG Haitao, et al. Sparsity enforced time–frequency decomposition in the Bayesian framework for bearing fault feature extraction under time-varying conditions [J]. Mechanical Systems and Signal Processing, 2023, 185: 109755.
[10]. ZHANG Han, CHEN Xuefeng, ZHANG Xiaoli, et al. Aero-engine bearing fault detection: A clustering low-rank approach [J]. Mechanical Systems and Signal Processing, 2020, 138: 106529.
[11]. LEE D D, SEUNG H. Learning the parts of objects by non-negative matrix factorization [J]. Nature. 1999, 401: 788-91.
[12]. GAO Huizhong, LIANG Lin, CHEN Xiaoguang, et al. Feature Extraction and Recognition for Rolling Element Bearing Fault Utilizing Short-Time Fourier Transform and Non-negative Matrix Factorization [J]. Chinese Journal of Mechanical Engineering, 2014, 28: 96-105.
[13]. ZHOU Tianyi, TAO Dacheng. GoDec: Randomized Lowrank & Sparse Matrix Decomposition in Noisy Case; proceedings of the ICML, F, 2011 [C].
[14]. Guo Kailing, Liu Liu, Xu Xiangmin, et al. GoDec+: Fast and Robust Low-Rank Matrix Decomposition Based on Maximum Correntropy.[J] IEEE Transactions on Neural Networks and Learning Systems,2018, 29(6): 2323-2336.
[15]. YU Liang, DAI Wei, HUANG Shichun, et al. Sparse Time–Frequency Representation for the Transient Signal Based on Low-Rank and Sparse Decomposition [J]. Journal of Acoustics, 2019, 1(1): e190003.
[16]. WANG Ran, FANG Haitao, YU LONGJING, et al. Sparse and low-rank decomposition of the time–frequency representation for bearing fault diagnosis under variable speed conditions [J]. ISA Transactions, 2022, 128: 579-98.
[17]. LEE D D, SEUNG H S. Algorithms for Non-negative Matrix Factorization; proceedings of the International Conference on Neural Information Processing Systems, F, 2000 [C].