A hierarchical framework for optimal microgrid energy management with ESS, EV, PV and demand response considerations

The increasing integration of renewable energy sources and distributed energy resources (DERs) underscores the need for robust, coordinated microgrid management strategies. This paper presents a comprehensive hierarchical optimization framework that unifies local grid utilization and grid-level demand response, integrating advanced forecasting with dynamic control. Utilizing data-driven load forcasting to predict power consumption based on historical usage and weather data, the proposed approach employs both day-ahead planning and receding horizon control to address the inherent variability in energy generation and demand. The framework minimizes peak power and power range while maximizing battery state-of-charge (SOC), ensuring efficient energy utilization and enhanced grid stability. The framework models ESS as bidirectional storage with linear state-of-charge dynamics, EV chargers as unidirectional loads with dynamic connection windows, and PV systems as uncontrolled generation sources whose output is predicted via an analytical irradiance-to-power model combined with data-driven forecasting. Demand response is incorporated through two complementary soft-bound signals - one incentivizing increased consumption during grid surplus and one enforcing curtailment during peak demand - each embedded directly as constraints within the hierarchical optimization. Simulation results demonstrate a significant reduction in peak power and effective optimization of battery usage under realistic operating conditions. Preliminary demand response experiments indicating further potential for grid resilience. This work provides a novel, integrated solution for modern microgrid energy management, paving the way toward more adaptive and sustainable energy systems.