PMU-Timescale Topology Identification of Sub-station Node-Breaker Models using Deep Learning

In an actual power system, the bus-branch model used in a state estimator is usually derived from the substation nodebreaker model. Therefore, incorrect topology information in the substation node-breaker model will cause the state estimation results to deteriorate. This paper presents a bus-branch topology identification framework for substation node-breaker models. The proposed method employs phasor measurement unit (PMU) data to identify the intra-substation connectivity at very high speeds. PMU data from Dominion Energy, a power utility in the U.S., is used to validate and test the proposed topology identification framework on the IEEE 30-bus system. The results demonstrate the applicability of the proposed framework for different cases in which varying levels of measurement noise are present under partial system observability provided by PMUs.