Detection of high impedance faults in PV systems using mathematical morphology

Use of Mathematical Morphology (MM) based filters to detect faults in power systems have been discussed in recent literature. These studies have been predominantly focused on AC systems and associated arc faults. However, fault detection applications based on MM filtering in DC systems have not been adequately discussed. This paper presents how MM based filtering techniques and its composite operations could be used to successfully detect arc faults in a PV system. Throughout the analysis, series and parallel arc fault signatures obtained from an actual PV system with maximum power point tracking (MPPT) control is used. Effect of MM parameters on detection sensitivity is explained.