Fault Diagnosis Method of Rolling Bearing Based on Immune Neural Network and D-S Evidential Reasoning

Considering the features of vibration signals of rolling bearing and low diagnosis accuracy via traditional signal processing methods, a novel fault diagnosis method of the rolling bearing based on immune neural network and D-S evidential reasoning is presented. Firstly, immune neural network with time domain and wavelet packet analysis method is used in the local diagnosis step. And then, the results in the local diagnosis step are further calculated with the weighted D-S evidential reasoning. Finally, the different fault types of rolling bearing are obtained. In order to validate the presented method, the benchmark dataset of the four types of bearing fault, the inner race fault, the rolling body fault, the outer race fault and the normal state, is adopted in this paper. The experimental results show that, the accuracy and reliability of the presented method is much better than a single neural network diagnosis method.