A New Fault Diagnosis Model Based on AIR Scheme

A new model for fault diagnosis of mechanical facilities is proposed in this paper. The model synthesizes the structure of the neural network and the scheme of the artificial immune regulation (AIR). The training samples are clustered first by the immune algorithm based on AIR scheme. The centers of the clustering (memory B-cells) are saved as the nodes of the hidden layer in the model, therefore, the amount and positions of nodes in the hidden layer can be determined automatically. The weight matrix is determined by least squares (LS) algorithm. Finally the numerical experiments in fault diagnosis of a tapered roller bearing are performed. The results show that the method is simple and effective through comparing with NN