Skip to Main Content
High-resistance grounding is implemented in the DC zonal electric distribution system (DCZEDS) to permit continuity of service during a ground fault. To prevent subsequent phase-to-ground faults on the other phase from causing a disruptive phase-to-phase fault, it is of great importance that fast and robust ground fault location methods be applied to the DCZEDS. Current fault locating techniques have shortcomings in their application to the DCZEDS with ldquocommon-railrdquo faults and with accommodating the high frequency noise caused by repetitive switching transients. In this paper, a new idea is presented which utilizes the background noise contained in a suitable system signal to extract information about the fault location. Multi-resolution wavelet analysis and fractal dimension analysis are used as examples of signal processing tools for recognition of characteristic features in the system signal. A notional, high-fidelity DCZEDS model in PSCAD/EMTDC is implemented and used to illustrate the effectiveness of the proposed method through simulation.