An ANFIS-based approach to locate arc fault in smart grid using phasor measurement units (PMU)

Smart grids experience hundreds of Arc faults every year. Due to the low current range of high impedance (arc) faults, it is difficult to detect and determine the location of this type of fault for protection equipment in distribution systems. In this paper, a method has been presented to determine the feeder and arc fault location based on ANFIS. In the introduced method, with the assumption of detecting the arc fault, the fault current signal was received from the equipped Phasor measurement units (PMU) at the beginning of each feeder, and using advanced signal processing methods (wavelet analysis), the important features of the fault current signal were extracted. The extracted features have been used as input to the Adaptive Neural Fuzzy Inference System (ANFIS). In this paper, Mayer and Cassie's combined equations are used to model the arc fault which is one of the most accurate models. No need to line parameters, increasing the accuracy, and reduction of complex calculations are the main features of the presented approach. The introduced scheme has been evaluated on a 10 kV distribution system and results show an acceptable performance of the presented method.