Intelligent Passive Islanding Detection Scheme For Microgrids Through a State Observer with Artificial Intelligence

Abstract

Microgrids are modern power systems that have evolved because of the global distribution of renewable energy resources (RERs) close to ending users. However, due to the dynamic nature of these microgrids, islanding detection (ID) is a major concern. A novel passive islanding detection strategy for microgrids is introduced in this paper. Initially, the voltage signals are acquired at the point of common coupling (PCC). Then, an adaptive Kalman filter (AKF) is applied to the measured voltage signals as a state observer for noise-free state estimations of the non-fundamental harmonic features. In addition, the recurrent neural network (RNN) is deployed on the extracted harmonic features for the calculation of state observer-based intelligent harmonic factor (SOBIHF). Finally, the SOBIHF is compared with the threshold level to typify between islanding and non-islanding condition. The presented approach has been tested in MATLAB/Simulink® on the study microgrid system. The results depict that the presented scheme detects islanding events with 99.8% accuracy and reduces the non-detection zone (NDZ) in various cases.