Fault Detection and Identification With Adaptive Neuro-Fuzzy Inference System

Abstract

An algorithm that is not influenced by the line parameters is an option to overcome a few drawbacks of conventional approach especially in fault identification. Such algorithm should be independence on the power network configuration to be more generalized. In artificial intelligence, artificial intelligent network seems efficient and practical when mathematical model of the system is not available. As for fuzzy logic, it possess a capability to copy the sensing, generalizing, operating, processing and learning capability of human operator. This research implemented fuzzy system in the framework of adaptive network and then combines with discrete wavelet transform to obtain a great performance. A novel approach to implementing Gustafson-Kessel clustering algorithm have performed significant results for a better fault analysis tool using artificial intelligence. Results show that the scheme has ability on justifying the type of fault and estimate the fault distance in a small range of error. The proposed method also has very less involving of calculation and almost no affected by the fault resistance.