An Improved PSO-Based Fuzzy Ensemble Classifier for Transformer Fault Diagnosis

To deal with the flaws of single neural network in process of transformer insulation fault diagnosis, for example, low diagnosis precision, long training time and bad generalized ability etc, in the paper we propose an ensemble fuzzy neural classifier based on improved particle swarm optimization (IPSO). The method fully utilizes the advantages of particle swarm optimization such as fast seeking speed and easy realization mode etc, the integrated time of the overall system therefore become very short. Thus, more neural networks are applied to diagnose transformer faults at the same time, and the final conclusion is identified based on all achieved results. Hence, the integrated neural network possesses higher diagnosis precision and reliability, and is an ideal pattern classifier. In the end, a practical application in transformer insulation fault diagnosis indicates that the proposed method is very effective