Gear faults diagnosis based on wavelet neural networks

Abstract

With the development of the industry, the machine system is becoming more and more complicated, and more and more difficult to detect the gear faults of such a large and complicated system. The wavelet neural network approach is developed for gear faults diagnosis. The wavelet neural work is trained by the gradient descent optimization algorithm in this paper. The wavelet neural network based on the gradient descent optimization algorithm is used to classify the gear crack faults in the early stage. The simulated result shows that the wavelet neural network approach is effective to distinguish the state of the gear and suitable to diagnose the gear crack faults in the early stage.