Data Based Fault Diagnosis of Hot Axle for High-Speed Train

The axle temperature is an important performance index for high-speed train axle, and many types of high-precision detection device have been developed. However, because the existing hotbox detection system works by comparing the measured value with fixed and experiential temperature thresholds, without considering the dynamic influence of running environment, the rates of fault and failure alarm are still high. Regarding this problem, this paper is to realize high precision fault diagnosis. Firstly, the characteristics of the temperature rising rates of all axles, the axles on the same side of a carriage, the axles on the same side of a whole train are analyzed. Secondly, the support vector machine (SVM) is employed to establish a data-based fault diagnosis model, and the model is optimized using genetic algorithm. Finally, the proposed method is applied to an actual running data, the analysis results demonstrated the effectiveness and feasibility.