Real-time intelligent fault diagnosis using deep convolutional neural networks and wavelet transform

Abstract

As an important part of the ship, the failure of the rotating machinery directly affects the safe navigation of the ship. Traditional fault diagnosis methods require manual feature extraction or selection which brings great limitations to the practical application. In our paper, a real-time intelligent fault diagnosis method is proposed based on wavelet transform algorithm and deep convolution neural network (DCNN) model. Firstly, original vibration signals of different kinds of fault is collected. Then, original signals are converted into time frequency image using wavelet transform method. Finally, these time-frequency images are fed into the DCNN to train the constructed model, and real-time signals are input into the trained model to achieve classification of the running state of the bearing. A real experiment is provided to estimate the effectiveness of the approach. The results demonstrate that the proposed method has high diagnostic accuracy and efficiency.