Forecast-integrated hierarchical energy management of PV–BESS in residential distribution networks using COA–VSA metaheuristics

This paper presents a hierarchical energy management system (EMS) for photovoltaic (PV) and battery energy storage systems (BESSs) in residential distribution networks. Given that load and renewable generation vary over time, two-day-ahead predictions of PV generation and residential demands are derived from machine learning models. The EMS operates under a master-slave optimization scheme, where the coati optimization algorithm establishes the optimal sizing and siting of the PV–BESS units and the vortex search algorithm schedules operations to minimize costs and maximize reliability. The research is performed on a 38-bus radial network with some residential nodes equipped with the PV–BESS systems. Simulation results improve PV self-consumption >96%, reduce peak grid imports by >65%, and achieve voltage and thermal constraints. Therefore, the proposed approach is a reliable and scalable method for efficient PV–BESS facilitation in residential networks.