Research on rolling bearing fault diagnosis based on improved local linear embedding algorithm

Abstract

Manifold learning method has strong ability of complex data processing, is widely applied in the fault diagnosis of rolling bearing, in view of the manifold learning algorithm of the locally linear embedding (LLE) algorithm generalization ability is not strong, poor ability to deal with sparse data were problem, this paper proposes a local linear embedding algorithm based on improved method for rolling bearing fault diagnosis. In this study, the local linear embedding algorithm is improved by using Mahalanobis distance instead of Euclidean distance to construct the local neighborhood, and the distance between samples is homogenized to construct the feature mapping function, which is solved by Lagrange multiplier method. Comparing the dimensionality reduction effect of the traditional algorithm and the improved algorithm, the research results show that the improved algorithm in this study has better dimensionality reduction effect on data and can improve the accuracy of fault identification. Therefore, the proposed method can improve the data processing ability of the local linear embedding algorithm.