Fault Classification in Power Distribution Systems with Photovoltaic Distributed Generation Insertion

Abstract

Faults detection and classification considering distributed generation insertion is a crucial procedure for power distribution systems. The significant growth of distributed generation units in the utility grid results in impacts on the conventional protection systems’ operation. These impacts are mainly caused by changes in the short-circuit current’s characteristics, direction, and amplitude. In this sense, this study addresses the impact of photovoltaic penetration on short-circuit fault classification process. The fault conditions are available in a modified IEEE 34-bus test system and modeled in ATPDraw. The algorithm for shot-circuit fault detection and classification is developed using MATLAB programming environment. Multilayer Perceptron artificial neural network is used to classify short-circuits types. The accuracy rate of the fault classification algorithm reduces as long as the photovoltaic panels insertion increases, being 100% and 96.5% without photovoltaic insertion and with high photovoltaic insertion, respectively.