Suspension structures in bogies are important components of railway trains. The mechanical performance of suspension structures has a significant influence on the running safety and reliability of railway trains. Therefore, online fault diagnoses on suspension structures in trains on basis of big data are very essential. In this work, a method is proposed to automatically extract high-level features from vibration signals and recognize the faults. The method is composed of a Convolutional Neural Network (CNN) on Fast Fourier Transform (FFT) of vibration signals. First, feature extraction is one of key steps in fault diagnosis for railway trains. The simulation data sets of vibration signals are acquired with considering the different faults of suspension components, and the vibration signals are preprocessed by FFT. Then, the FFT coefficient-vectors are used to train Convolutional Neural Networks layer by layer to recognize different faults. Finally, results show that this method is extensively applicable and can achieve very high diagnostic accuracy for different faults. It provides a new measure for fault diagnosis of the railway trains.

CICTP