An Intelligent Differential Protection Scheme for DC Microgrid

Abstract

This paper proposes a differential current based intelligent fault detection technique for DC microgrids. The currents of both ends of the faulted line are utilized for the computation of differential current. This differential current is then processed through the machine learning (ML) technique named as support vector machine (SVM) for fault detection in low voltage DC (LVDC) microgrids. The SVM based protection technique is tested on a LVDC microgrid integrated with different types of renewable energy sources (RES) modelled in the SIMULINK platform. The presented protection scheme is examined for pole-ground (PG) and pole-pole (PP) faults with varying fault resistance, fault location and RES penetration in grid-connected and islanding mode. The proposed methodology is verified for its capability to discriminate the transient no-fault condition from fault cases.