A Fault Diagnosis Method Based on Support Vector Machine Optimized by Sparrow Search Algorithm

Early fault diagnosis of industrial systems can greatly eliminate potential safety hazards. The existing rolling bearing fault diagnosis algorithm relies on manual feature extraction, and the diagnosis effect is poor when the number of fault samples is limit. Aiming at these deficiencies, a new fault diagnosis method that combines stacked autoencoder with improved support vector machines is established. Firstly, introduce the sparse stack autoencoder (SSAE) including a softmax to extract the features of the bearing data, then use the sparrow search algorithm (SSA) to automatically optimize the key parameters of support vector machine (SVM), and then, the extracted features are classified by the optimized SVM. Finally, the effectiveness of the SSAE-SSA-SVM learning framework proposed in this paper is evaluated by simulation experiments based on bearing data from Western Reserve University (CWRU). In addition, compared with SSAE-PSO-SVM, SSAE-SVM, SSAE, and CNN, the results show that the method has better performance in classification accuracy.