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

A non-transitory, processor-readable medium storing instructions that, when executed by a processor, cause the processor to:

• receive first electrical measurement data that characterizes a first portion of an electrical grid and not a second portion of the electrical grid, the first portion of the electrical grid having a first nominal voltage, and the second portion of the electrical grid having at least one second nominal voltage that is higher than the first nominal voltage;
• calculate a first reactive power value based on the first electrical measurement data, the first reactive power value being (1) associated with the first portion of the electrical grid and (2) outside of a predefined reactive power value range;
• provide the first reactive power value as input to a machine learning model to identify a contactor from a plurality of contactors, each contactor from the plurality of contactors being associated with (1) a capacitor bank different from remaining capacitor banks from a plurality of capacitor banks and (2) the second portion of the electrical grid;
• in response to identifying the contactor, cause a signal to be sent to the contactor to cause a capacitor bank (1) associated with that contactor and (2) from the plurality of capacitor banks to be one of coupled to the second portion of the electrical grid or decoupled from the second portion of the electrical grid;
• and
• in response to causing the signal to be sent to the contactor:
• receive second electrical measurement data associated with the first portion of the electrical grid, and
• calculate a second reactive power value based on the second electrical measurement data, the second reactive power value being one of (1) within the predefined reactive power value range or (2) closer to the predefined reactive power value range than the first reactive power value.

A non-transitory, processor-readable medium storing 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;
• provide the electrical measurement data 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;
• and
• in response to identifying the capacitor bank, 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.