基于线性局部切空间排列维数化简的故障诊断

为实现旋转机械故障诊断方法的自动化、高精度及通用性,提出基于线性局部切空间排列(Linear LocalTangent Space Alignment,LLTSA)维数化简的故障诊断模型。首先结合经验模式分解(Empirical Mode Decomposition,EMD)和自回归(Autoregression,AR)模型系数构造全面表征不同故障特性的混合域特征集,再利用LLTSA将高维混合域特征集化简为故障区分度更好的低维特征矢量,并输入到最近邻分类器(K-nearest Neighbors Classifier,KNNC)中进行故障模式识别。所提出的诊断模型充分融合混合域特征融合在故障特征的全面提取、LLTSA在信息的有效化简及KNNC在分类决策方面的优势,实现诊断方法的自动化、高识别率及较好的通用性。用深沟球轴承不同部位、不同程度故障诊断实例验证该模型的有效性。

Fault diagnosis model based on dimension reduction using linear local tangent space alignment

Abstract：Based on dimension reduction with Linear Local Tangent Space Alignment (LLTSA), a novel fault diagnosis model is proposed in this paper to achieve automation、high-precision and generality of fault diagnosis of rotating machinery. With this model, mixed-domain feature sets of training and test samples are first constructed to characterize the property of each fault comprehensively by the fusion of Empirical Mode Decomposition (EMD) and Autoregression(AR) model coefficients. After that, LLTSA is introduced to automatically compress the high-dimensional eigenvectors of training and test samples into the low-dimensional eigenvectors which have better discrimination. Finally, the low-dimensional eigenvectors of training and test samples are input into K-nearest neighbors classifier (KNNC) to carry out fault diagnosis. Compared to the existing approaches, the proposed diagnosis model combines the strengths of mixed-domain features fusion in extensive extraction of fault feature, LLTSA in effective compression of fault information and KNNC in classification decision-making, and realizes the automation、high-precision and generality of fault diagnosis method. The diagnosis example on different fault positions and severities of deep groove ball bearings validates the effectivity of proposed fault diagnosis model.