Rotating Machine Fault Diagnosis Based on Locality Preserving Projection and Back Propagation Neural Network–Support Vector Machine Model

Abstract

In order to effectively recognize the rotating machine fault, a new method based on locality preserving projection and back propagation neural network–support vector machine model is proposed. First, the gathered vibration signals are decomposed by the empirical mode decomposition, and the corresponding intrinsic mode functions are obtained. Then, Shannon entropies of the intrinsic mode functions are used as the original features. But the extracted features have the problems of high dimension and redundancy. So, the manifold learning algorithm locality preserving projection is introduced to extract the characteristic features and reduce the dimension. The characteristic features are inputted to the back propagation neural network–support vector machine model to train and construct the fault diagnosis model, and the rotating machine fault condition identification is realized. The running states of a normal inner race and several inner races with different degrees of fault were recognized; the results validate the effectiveness of the proposed algorithm.