EMD/HT-based local fault detection in DC microgrid clusters

DC faults can create serious damages if not detected and isolated in a short time. This paper proposes a fault detection technique for DC faults to enhance the protection of DC microgrid clusters. To detect such faults accurately and quickly, a DC fault detection scheme using empirical mode decomposition and Hilbert transform is proposed. Due to the strict time limits for fault interruption caused by fast high-rising fault currents in DC systems, DC microgrid clusters' protection remains a challenging task. Furthermore, high impedance faults (HIFs) in DC systems cause a small change in the current, which can damage the power electronic converters if not detected in time. Therefore, this paper proposes a local scheme for the fast detection of faults including HIFs in DC microgrid clusters. Both simulation and experimental results using a scaled DC microgrid cluster prototype and considering several scenarios (such as low impedance faults, HIFs, noise, overload, and bad calibration of sensors) demonstrate the successful and fast detection (less than 2 ms) of DC faults by the proposed method. Compared with other techniques, the proposed scheme presents its merits from the viewpoints of accuracy and speed.

Automated summary

This paper proposes a fault detection technique for DC faults in order to enhance the protection of DC microgrid clusters. The proposed method uses empirical mode decomposition and Hilbert transform for accurate and fast detection of DC faults. The local scheme proposed in this paper is able to quickly detect faults, including high impedance faults, in DC microgrid clusters.