Grid Integration Issues of Photovoltaic Systems and Islanding Detection

This paper investigates the challenges in the grid caused by high penetration of photovoltaic (PV) systems and motivates them to achieve efficient islanding detection during grid abnormalities. To achieve this an islanding classification approach is developed by using a multi-layer perceptron neural network (MLPNN) to train with the possible islanding scenarios that can occur in a grid. The proposed approach aids the grid-connected system to isolate the healthy portion from the faulty portion of the system and ensure uninterrupted operations of the system even in the case of fault and grid maintenance. The performance of the proposed method was tested on the data obtained for different grid abnormalities simulated on a grid-connected photovoltaic system using a Typhoon-HIL environment. The obtained results depicted 95.6% training efficiency and 98.3% testing efficiency, and the trained approach can effectively detect an islanding scenario in 0.02 s with 1.7% misclassification error.