Intelligent Islanding Detection Scheme for Multiple DG Microgrids using Random Forest Classifier

Abstract

The detection of islanding is very important for the safe operation of distributed generators (DG) and microgrids (MG). A multiple DG microgrid based islanding detection scheme is proposed. Every DG’s harmonics and unbalanced voltage characteristics at the PCC are extracted using the discrete wavelet transform (DWT). The Random Forest (RF) classifier is used for classification. An IEEE-13 bus system with a solar PV array and a diesel generator as DGs modelled in SIMULINK is taken as the test system for the proposed method. The performance of the proposed method is tested by generating various scenarios like changing loads, introducing faults, and switching capacitors in the system. The results show the method is promising in terms of accuracy and speed.