Application of adaptive neuro fuzzy inference system to the partial discharge pattern recognition

The application of adaptive neuro fuzzy inference system (ANFIS) to the partial discharge (PD) pattern recognition is presented in this paper. Four types of defect models are made according to the main reason of insulation failures in real power system. Experiments are carried out to acquire the sample data, from which eight statistical features are extracted to construct the ANFIS. Different characteristics of the proposed defect models are compared based on the extracted features. Then the ANFIS is trained by characteristic features. Testing samples are utilized to validate the performance of the recognition system. The result shows that ANFIS reaches a successful recognition rate in the application of PD pattern classification.