Multivariate and multiscale monitoring of large-size low-speed bearings using Ensemble Empirical Mode Decomposition method combined with Principal Component Analysis

With a view to detecting incipient failures in large-size low-speed rolling bearings and ensuring minimal effect of subjectivity on the process, a new data-driven multivariate and multiscale statistical monitoring method is proposed. The proposed method which combines the Principal Component Analysis (PCA) multivariate monitoring approach and the Ensemble Empirical Mode Decomposition (EEMD) method, which adaptively decomposes signals into various time scales, was called the EEMD-based multiscale PCA (EEMD–MSPCA). The method is very general in nature, which is why it could also be used in different areas and for various tasks. It can be used for controlling each time scale of decomposition or only the selected ones, for multivariate and multiscale filtering or for monitoring system operation on the basis of reconstructed i.e. filtered signals. The efficiency of the proposed EEMD–MSPCA method for the task of bearing condition monitoring and signal filtering was evaluated on simulated as well as on actual vibration and Acoustic Emission (AE) signals measured on a purpose built test stand. The fact that the proposed method is able to identify the local bearing defect of a very small size indicates that AE and vibration signals carry sufficient information on the bearing condition and that the proposed EEMD–MSPCA method ensures high-reliability bearing fault detection.     

基于流形学习的滚动轴承故障盲源分离方法

滚动轴承的故障信号是一种典型的非线性非平稳信号,其信号中常常混有噪声信号及其他干扰成分。提出了一种基于流形学习的滚动轴承故障盲源分离方法,首先,利用经验模态分解(empirical mode decomposition,简称EMD)对单通道模拟信号进行分解,对得到的多通道信号构造其协方差矩阵,计算矩阵的奇异值下降速比得到原始信号数目;其次,利用峭度等指标选择最优观测信号,利用核主成分分析(kernel principal components analysis,简称KPCA)提取信号的流形成分;最后,利用快速独立成分分析(fast independent component analysis,简称Fast ICA)还原得到源信号。该方法不但解决了故障信号的欠定盲源分离问题,还提出了最优观测信号的确定准则,并通过实例验证了方法的有效性。     

Subspace-based gearbox condition monitoring by kernel principal component analysis

Feature extraction is a key step for gearbox condition monitoring. The statistical features of the measured vibrations can be used to characterise gearbox conditions; however, their regularity and sensitivity in pattern space are different and may vary considerably under different operating conditions. This paper addresses the non-linear feature extraction scheme from the time-domain features with wavelet packet preprocessing and frequency-domain features of the vibration signals using the kernel principal component analysis (KPCA). Then two different KPCA-based subspace structures are constructed for representing and classifying the gearbox conditions. The proposed methods can extract the non-linear features of gearbox conditions using KPCA effectively, and perform conveniently with low computational complexity based on subspace methods. Experimental analysis with a fatigue test of an automobile transmission gearbox shows that the KPCA features outperform PCA features in terms of clustering capability, and both the two KPCA-based subspace methods can be effectively applied to gearbox condition monitoring.     

基于短时滑移模糊熵和LPP的轴承故障诊断

针对旋转机械设备的故障特征微弱和环境噪声强等问题,提出了一种基于短时滑移模糊熵和局部保留投影法(locality preserving projection,简称LPP)的故障特征提取方法。首先,通过对滑移截断短时序列的架构分析,引入多尺度复合模糊熵,获得信号在不同复合尺度下的特征信息和故障潜在特征,能准确反应信号复杂度和不确定性;其次,应用LPP流形降维并保留信号的局部数据特征,设计最优带通滤波器,对轴承振动信号进行故障冲击特征提取。仿真分析和实验数据结果验证了该方法在强背景噪声情况下降噪抑制方面的有效性,具有快速识别和提取滚动轴承的微弱冲击特征的能力。     

GFD和核主元分析的机械振动特征提取

针对旋转机械非线性特征提取的问题,提出了广义分形维数(generalized fractal dimension,简称GFD)和核函数主元分析(kernel principal component analysis,简称KPCA)的旋转机械振动特征提取方法。首先,通过广义分形维数进行初次特征提取,形成高维特征空间;其次,通过核主元分析方法对高维特征空间降维并进行第二次特征提取;最后,利用核主元分析方法和KN近邻(KNN)方法对转子和轴承不同状态下的特征进行了分类。研究表明,GFD-KPCA方法对旋转机械进行了有效的特征提取,对不同状态的数据有高精度的分类,对参数选取有较低的依赖性。轴承微弱振动特征提取结果显示,GFD-KPCA性能优于常规的KPCA特征提取算法,具有更好的精度和适用范围。     

Fault diagnosis of rolling element bearing using a new optimal scale morphology analysis method

Periodic transient impulses are key indicators of rolling element bearing defects. Efficient acquisition of impact impulses concerned with the defects is of much concern to the precise detection of bearing defects. However, transient features of rolling element bearing are generally immersed in stochastic noise and harmonic interference. Therefore, in this paper, a new optimal scale morphology analysis method, named adaptive multiscale combination morphological filter-hat transform (AMCMFH), is proposed for rolling element bearing fault diagnosis, which can both reduce stochastic noise and reserve signal details. In this method, firstly, an adaptive selection strategy based on the feature energy factor (FEF) is introduced to determine the optimal structuring element (SE) scale of multiscale combination morphological filter-hat transform (MCMFH). Subsequently, MCMFH containing the optimal SE scale is applied to obtain the impulse components from the bearing vibration signal. Finally, fault types of bearing are confirmed by extracting the defective frequency from envelope spectrum of the impulse components. The validity of the proposed method is verified through the simulated analysis and bearing vibration data derived from the laboratory bench. Results indicate that the proposed method has a good capability to recognize localized faults appeared on rolling element bearing from vibration signal. The study supplies a novel technique for the detection of faulty bearing.     

基于改进多元多尺度色散熵的齿轮箱多通道振动信号故障诊断

齿轮箱发生故障时,其振动信号具有不平稳和非线性等特征,而常用的齿轮箱故障诊断方法大多是建立在单通道振动信号分析基础上,容易造成故障信息丢失,故而在工业生产中实用性受限。为了克服此缺陷,将多元多尺度色散熵引入到齿轮箱故障诊断当中,并改进其粗粒化方式,提出了改进多元多尺度色散熵,用以提取齿轮箱多通道振动信号的故障信息。在此基础上,提出一种基于集合经验模态分解,改进多元多尺度色散熵和遗传算法优化支持向量机的齿轮箱故障诊断方法。通过实验数据分析,并与多元多尺度样本熵、多元多尺度模糊熵等现有方法相比较,证明该方法具有更高的准确率和稳定性,且在处理短时间序列时具有明显优势。     

基于关联维数的滚动轴承故障诊断的研究

针对滚动轴承系统产生的非线性振动信号的特点，提出用关联维数来描述轴承振动信号的工作状态，进而对其进行故障诊断的方法。同时详细讨论了关联维数的计算方法，并对由轴承系统产生的非线性振动信号进行了关联维数的定量计算。实验表明，滚动轴承不同工作状态由不同的动力学机理产生，其关联维数明显不同。以关联维数作为滚动轴承的工作状态监测的依据，可以为提高滚动轴承故障诊断的准确率提供了一种有效的新方法。     

A new rolling bearing fault diagnosis method based on multiscale permutation entropy and improved support vector machine based binary tree

A new bearing vibration feature extraction method based on multiscale permutation entropy (MPE) and improved support vector machine based binary tree (ISVM-BT) is put forward in this paper. Local mean decomposition (LMD), a new self-adaptive time–frequency analysis method, is utilized to decompose the roller bearing vibration signal into a set of product functions (PFs) and then MPE method is used to characterize the complexity of the principal PF component in different scales. After the feature extraction, a new pattern recognition approach called ISVM-BT is introduced to accomplish the fault identification automatically, which has the priority of high recognition accuracy compared with other classifiers. Besides, the Laplacian score (LS) is introduced to refine the fault feature by sorting the scale factors. Finally, the rolling bearing fault diagnosis method based on LMD, MPE, LS and ISVM-BT is proposed and the experimental results indicate the proposed method is effective in identifying the different categories of rolling bearings.     

基于粒子群优化的核主元分析特征的提取技术

针对核主元分析在参数设置上的盲目性,提出应用粒子群优化算法优化核函数参数.并将核主元分析应用于特征提取中.首先建立核函数参数优化的数学模型,然后应用加速度自适应粒子群优化算法对其寻优,并通过Iris数据集进行仿真研究,验证其提取特征的有效性.将优化的核主元分析方法应用于齿轮箱典型故障的特征提取中,结果表明:参数优化的核主元分析能有效降低齿轮箱特征向量的维数,较线性主元分析取得更好的故障识别效果.该方法在机械故障信号的非线性特征提取中具有优势.