Online motor fault diagnosis using hybrid intelligence techniques

In this paper, a novel approach for online motor fault diagnosis is proposed based on artificial neural network (ANN) trained by immune clustering and genetic algorithm (IGA). The IGA is employed to adaptively optimize the structure of the radial basis function neural network (RBFNN). The clonal selection principle is responsible for how the centres will represent the training data set. The immune network theory is used to avoid individual redundancy. The selection probability based on density and fitness enhances the RBFNN training convergence performance. The simulative experimental results show that RBFNN trained by the proposed algorithm has a smaller number of parameters, a faster convergence speed and higher fault diagnosis accuracy.