ANFIS APPROACH FOR LOCATING FAULTS USING GPS COORDINATES IN A DISTRIBUTION SYSTEM

摘要

Power utilities are increasingly concerned about the issues on quality and reliability of power supply system to accomplish low Customer Average Interruption Duration Index (CAIDI). One of the important aspects to assure high quality and reliability of power supply is to handle faults in distribution networks. Faults need to be located accurately and as quickly as possible before appropriate action can be taken. This paper presents the development of an intelligent technique to locate faults in a power distribution system. The proposed technique implements the Adaptive Neuro-Fuzzy Inference System (ANFIS) for fault location in latitute and longitute as GPS coordinates. The output of the proposed method provides information such as type of fault and fault address from google map by using the post-fault three-phase current data. The results of ANFIS training show that the average error is almost zero for each types of fault, 0.004 for latitute coordinate and also almost zero for longitute coordinate. Simulations were carried out by using the commercial software package PSS-ADEPT to validate the proposed approach.