Using LV Real-Time Data for Pseudo-Measurements Generation in MV Distribution Networks

Abstract

This paper presents a new method to generate pseudo-measurements of different electrical quantities for distribution secondary substations. The idea is to take advantage of the stronger correlation that exists between electrical variables in a given MV/LV substation and the correspondent downstream network. Autoencoders trained with historical data are integrated on an optimisation algorithm, together with real-time metering information of the LV network, for accurately estimate pseudo-measurements at the correspondent MV/LV substation. The proposed methodology is particular suitable either for MV/LV substations not being telemetered in real-time or in case of failures related to telemetry equipment existing in the substation.