Fault Classification in Power System with Inverter-Interfaced Renewable Energy Resources Using Machine Learning.

Fault classification is crucial in fault mitigation to maintain selectivity in tripping only the faulted phase or zone in power system networks. However, inverter-interfaced renewable energy sources' unique fault current profile poses challenges to classifiers designed for conventional systems, which are inadequate in the presence of renewable energy resources such as inverter-interfaced photovoltaic (PV) or wind turbine systems in the grid. The inverters have internal protection schemes that trip during unbalanced conditions; however, in grids with high penetration of renewable energy, the inverter must ride through the fault and let relays protect the system. Moreover, the different control strategies for inverters can make the fault current small enough to be unreliable to use as a parameter in fault classifications. This study proposes a reliable fault classification method that can accurately identify faults in power systems with high penetration of renewable energy sources. This paper discusses a machine learning (ML)-based classifier using phase current and voltage magnitude to classify faults. The performance of the proposed classifier is validated against different fault scenarios in power systems like the IEEE 9-bus system. The classifier discussed in this paper achieved a satisfactory accuracy of 99.78% with voltage measurements for test conditions within three-quarters of a cycle. The classifier can be used for any three-phase system to provide correct faulted phase information to other protection components. The same methodology is extended to identify evolving faults, achieving an accuracy of 99.6% in determining the evolving fault type. Thus, the proposed ML-based classifier provides a reliable and accurate method for fault classification in power systems with high penetration of renewable energy sources. [ABSTRACT FROM AUTHOR]