Learning-based hierarchical control for a net-zero commercial building with solar plus storage and high EV Chargers Penetration

The integration of renewable energy resources and electric vehicles in microgrids presents a significant challenge due to generation and demand uncertainty. However, our technical and market research in Solar Plus Storage microgrids, carried out as part of a US Department of Energy's Solar Prize project, demonstrates the efficacy of advanced microgrid controls in managing this uncertainty while saving billions of dollars. The control system must be equipped with advanced optimization tools and adaptive forecasting models that minimize planning errors for future optimal operation to achieve this. This study contributes to the literature by quantifying the impact of integrating adaptive learning-based forecasting tools for PV power and electrical load at the tertiary level of a hierarchical control system. Additionally, we use the developed learning forecasters and optimization algorithms of the tertiary level to solve intertwined optimal sizing and operation subproblems as a combined problem for a solar plus storage system. •Developed deep learning tools for forecasting PV and electrical demand.•Proposed a method to optimize the operational cost of a solar plus battery system.•Designed a hierarchical control system with learning-based forecasting tools.•Proposed method can be used to both size and operate the system optimally.•Performed extensive sensitivity analysis to tune the proposed algorithm.