Modelling and Optimization Based Control for Demand Response in Active Distribution Networks

Abstract: We explore how Demand Response (DR) can effectively provide electricity system services such as for the management of bi-directional power flows and the control of voltage deviations in active distribution networks, without compromising consumer comfort or adversely affecting grid operation in the transmission network. By translating intricate power system physics into straightforward control objectives, we design DR control algorithms that can operate within realistic computational time frames at scale. We conduct simulation-based experiments and find that minimizing the Euclidean-Norm of the total residual load at transformer sub-stations is an effective objective for harnessing DR for the dispatch of renewable electricity grids. We show that this control objective can be efficiently pursued in sequential order, without optimal power flow calculations or information about the topology of a grid. Additionally, we find that pursuing this objective reduces the sum of peak power flows along all lines in active distribution networks, and can therefore enable both lower transmission losses and lower voltage deviations.