Fault type recognition of High-speed train bogie based on dual-channel integration of information entropy

Traditional High-speed train fault diagnosis methods only simply considers characters of a certain single domain to conduct information fusion. Aimed at improving that, a new approach to discriminate the fault type is brought forward, using dual-channel multiple domain information entropy to analysis the vibration signals of the high-speed train bogie. In the model based on dual-channel information fusion, singular spectrum entropy, power spectrum entropy, wavelet energy entropy as well as wavelet space feature entropy are extracted and figured, to conduct certain fault recognition, such as 8 locations of resist sinusoidal vibration damper fault, 4 positions of Lateral damper failure and 2 places of the air spring failure. Examples of diagnostic results show that the fault diagnosis model, combined with the data fusion theory and information entropy features, can effectively improve the accuracy of fault diagnosis, almost getting close to 100%.