Islanding and Grid Disturbance Detection Based on Multi-SVMs Delay Validation in Synchronous Dominated Microgrids

Abstract

Listen

Translate article

Intrinsically active, passive, or a hybrid method of the two is the main methods for solving the problem of islanding detection in microgrids. For islanding and disturbance detection, most methods rely on the common coupling point of microgrid and large power grid. Based on multiple support vector machines, this paper proposes an islanding and interference detection method. First, analyze the on-site electrical quantities and form feature vectors before training the support vector machine on the islanding and interference conditions; second, train the support vector machine on the islanding and interference conditions; finally, through the complementarity of multiple classifiers, improve islanding detection accuracy. The simulation example considers interference in multiple scenarios to validate the effectiveness of the above method, and the results show that the in-situ detection method in this paper gains robustness.