Fault Detection and Classification of Microgrid Based on Mode Decomposition and Extreme Learning Machine

Abstract

Accurate fault detection and classification is a measure issue in a microgrid (MG). The MG often experiences shunt faults inside or outside of it, the circuit breaker connected between the utility grid and MG must immediately respond and open the circuit. If the fault is not detected accurately, it hampers the system's reliability and load performances also increase faulty line restoration costs. This research proposes a robust fault detection and classification technique based on Empirical Mode Decomposition (EMD) and Extreme Learning Machine (ELM). Energy of Decomposed current signals are used for unbiased feature extraction in presence of noise. whereas ELM is used for accurate fault detection and classification. The proposed EMD-ELM technique is validated in standard test system and found to be performing better as compared to other competitive techniques.