A Novel Approach for Detection High Impedance Fault in DC Microgrid

The use of DC micro-grids enhances the efficiency of the power systems, and also reduces the complexity of the power network that is one of the major advantages for them. Furthermore, due to the abundance of DC generation resources, energy storage devices, and loads the use of DC micro-grid will also decrease the costs. However, detecting high impedance faults is one of the most important challenges for the protection of this type of network. For this reason, this paper presents a new scheme of high impedance fault detection based on signal prediction using GMDH neural network. The GMDH neural network has been presented as a multivariate analysis method for modeling and specifying complex systems. The proposed method in this paper has been tested using DIgSILENT and Matlab software on a microgrid with single-pole topology. The results illustrate that the proposed method is capable of detecting the faults with impedance value up to 700 ohms. In addition, this method detects the low-impedance faults as soon as they occur in the system, and also another prominent merit of the presented method is that the noisy conditions do not have any effect on its operation.