Air Brake System Fault Diagnosis Method Based on Optimized SVM

This paper uses the 120-air brake as the research object to carry out its fault diagnosis research. In this paper, the 120-air brake is diagnosed with the help of optimized Support Vector Machine. And its diagnostic effect is compared with traditional models, including SVM, RBF and CNN. The confusion matrix is used to display the diagnostic results, and the performance of the four algorithms is compared and analyzed by F1-score and error indexes. In this study, the wind pressure data for five channels of the heavy-haul railway’s air brake system are extracted, selected and fused to form the optimal feature set required in this paper, and the optimal feature set is substituted into the diagnostic models to realize fault diagnosis. Through comparative analysis, this paper finally concludes that the SVM based on SSA performs best in the fault diagnosis of the air brake system in this paper.