Advanced Three-Phase Instantaneous Power Theory Feature Extraction for Microgrid Islanding and Synchronized Measurements

Real-time fault detection and classification of electrical measurements is a monitoring challenge in power systems because fault identification plays a vital role in the automatic islanding and resynchronization of Microgrids. Realtime recordings from synchronous measurement units bring new functional capabilities to the microgrid controllers among which real-time islanding detection and classification would be a critical task to be addressed. In this work, we initialize the idea of an instantaneous intelligent passive islanding detection (IIPID) strategy by interpreting the instantaneous power theories as a fast and reliable feature extraction framework for islanding event detection and recognition. We are going to use the advanced power theories as possible signal decomposition techniques to extract a set of distinctive features to model the behavior of a different type of faults in the smart microgrid systems. For this reason, we have used the synchronous reference frame method in addition to conservative power theory (also a combination of these techniques) to extract useful features from 3-phase electrical voltage signals under faulty situations. A dataset of 10 famous islanding event scenarios have been generated in Matlab/Simulink, and a variety of combined feature spaces were formed. Through comprehensive discussions over these case studies, it has been demonstrated that this technique will form a well-separable feature space which can be easily exploited by any nonlinear classifier to recognize the reason of the fault in almost real-time and with a high accuracy.