Protection Scheme for 4-Bus Grid Connected and Islanded DC Microgrid Using k-Nearest Neighbor Algorithm

The microgrids are smart power generation technology to supply power to near load centers. It is significant for rural and urban places due to its operation using nature-based distributed energy resources. Protection is still a challenging task in microgrid due to numerous issues, which are line and ground faults, variation fault current, and false tripping. Fault detection in the microgrid after its occurrence is a crucial prerequisite to quickly minimize major disconnection in the system. In proposed work, a kNN-based protection algorithm has been given to detect dissimilar pole-to-ground faults in the four-bus DC microgrid. The protection strategy is divided into three parts, where in the first step, signals are collected, then mode is identified, and further fault detection/classification and section identification tasks have been performed. Accuracy achieved after analysis reveals that the protection strategy is accurate and accomplished of analyzing faults in the microgrid.