Principal Investigator: Dr. Rob Cuzner
This project addresses transmission protective relay issues caused by the Inverter Based Resources (IBRs) at one or both ends on the location and isolation of transmission line faults. The project will demonstrate improved fault discrimination utilizing a machine learning classifier model that addresses the limitations of the conventional overcurrent relay coordination, including the variability of responses in practical systems, by accessing every possible data combination from all possible modes and configurations of transmission line. A digital twin will be implemented and used to train the protection relaying model. In actual systems, the digital twin would be implemented locally at IBRs, or as part of protective relaying hardware to make protection systems adaptable. The main advantages of this technique is improved accuracy and less computational overhead, compared to conventional approaches distance/impedance protection, given the limited dynamic overload/overcurrent ratings of the IBRs or variability of responses with and without IBRs. The transmission and distribution power grid models will be modeled on a Typhoon HiL 604 system. Controls for the IBRs and protective relaying will be implemented utilizing FPGA based NI devices. Real-Time simulation capability will be expanded through use of FlexRIO FPGA plug-in modules that communicate via peer-to-peer streaming, over a PCI Express (PXIe) backplane implemented in a National Instruments PXIe-1095 chassis.