Systems and methods for correcting reactive power anomalies in electrical networks based on distribution measurements

Abstract

A non-transitory, processor-readable medium stores instructions that, when executed by a processor, cause the processor to receive electrical measurement data measured at a distribution network of an electrical grid and not a transmission network of the electrical grid. The electrical measurement data is provided as input to a machine learning model to detect a reactive power anomaly that is at the distribution network, and in response to detecting the reactive power anomaly, identify a capacitor bank from a plurality of capacitor banks and associated with the transmission network of the electrical grid. In response to identifying the capacitor bank, the instructions cause the processor to cause the capacitor bank to be one of electrically coupled to the transmission network or electrically decoupled from the transmission network, to mitigate the reactive power anomaly.