Comparison of Legacy Relay with Machine Learning Based Relay for Detecting Faults at Inverter Terminals in a Distribution System with Inverter Based Resources

With increasing penetration of inverter based resources (IBR) in distribution systems, the distinction between fault and load current is lost due to the current limiting feature of inverters. This may render the fault undetectable to traditional overcurrent relays. Voltage assisted overcurrent relays and machine learning-based relays have been shown to be promising in this regard. This paper compares the performance of the two relays in a distribution system with 100% IBR penetration using the voltage and current measurements obtained at inverter terminals.