In modern electrical power distribution systems, the effective operation of inverter-based arc suppression devices relies on the accuracy of faulty phase selection. In the traditional methods of faulty phase selection for single-phase-to-ground faults (SPGs), power frequency-based amplitude and phase characteristics are used to identify the faulty phase. In the field, when a high-resistance SPG occurs in the system, traditional methods are difficult for accurately identifying the faulty phase because of the weak fault components and complicated process. A novel realizable and effective method of faulty phase selection based on transient current similarity measurements is presented when SPGs occur in resonantly grounded distribution systems in this paper. An optimized Hausdorff distance matrix (MOHD) is proposed and constructed by the transient currents of three phases’ similarity measurements within a certain time window of our method. This MOHD is used to select the sampling time window adaptively, which allows the proposed method to be applied to any scale of distribution systems. Firstly, when a SPG occurs, the expressions for the transient phase current mutation in the faulty and sound phases are analyzed. Then, the sampling process is segmented into several selection units (SUs) to form the MOHD-based faulty phase selection method. Additionally, the Hausdorff distance algorithm (HD) is used to calculate the waveform similarities of the transient phase current mutation among the three phases to form the HD-based faulty phase selection method. Finally, a practical resonant grounded distribution system is modeled in PSCAD/EMTDC, and the effectiveness and performance of the proposed method is compared and verified under different fault resistances, fault inception angles, system topologies, sampling time windows and rates of data missing.

No abstract

No abstract