Impedance Ground Faults Detection and Classification Method for DC Microgrid

Aiming at the practical engineering problem, it is easy to be confused and difficult to detect, such as the single pole high impedance ground fault and load switching of the DC microgrid. This paper proposes a method, which the impedance ground faults detection and classification method. Based on the combination of improved complete ensemble empirical mode decomposition with adaptive noise and random forest. First, the dependence of original signal is reduced by complete ensemble empirical mode decomposition with adaptive noise. Secondly, by comparing the cumulative slope sum k with the threshold, the abnormal conditions can be distinguished from the normal condition, load switching and the ground faults can be further distinguished through the energy ratio R-ratio. Finally, the random forest is used to further classify the ground faults to achieve precise classification of impedance ground faults in the DC microgrid. The analysis of calculation examples shows the method, it is quickly and effectively to detect and classify impedance ground faults in DC microgrid quickly and effectively, without being affected by fault resistance, fault initiation time and fault location.