Signal processing and machine learning techniques in DC microgrids: a review

Abstract

Low-voltage direct current (DC) microgrids have recently emerged as a promising and viable alternative to traditional alternating current (AC) microgrids, offering numerous advantages. Consequently, researchers are exploring the potential of DC microgrids across various configurations. However, despite the sustainability and accuracy offered by DC microgrids, they pose various challenges when integrated into modern power distribution systems. Among these challenges, fault diagnosis holds significant importance. Rapid fault detection in DC microgrids is essential to maintain stability and ensure an uninterrupted power supply to critical loads. A primary challenge is the lack of standards and guidelines for the protection and safety of DC microgrids, including fault detection, location, and clearing procedures for both grid-connected and islanded modes. In response, this study presents a brief overview of various approaches for protecting DC microgrids.