Holistic Data-Driven Approach for Sizing and Energy Management of an Urban Islanded Microgrid

This chapter presents a data-driven approach for optimal sizing and operation of islanded microgrids within an urban context. The study employs a building-level urban islanded microgrid as a testbed for case studies. A randomized learning model is introduced for supply and demand forecasting, and based on these results, a data-driven approach is employed to generate uncertainty scenarios characterizing uncertain supply and demand characteristics. Optimal sizing is conducted considering operational constraints to determine the sizes of different components, particularly the energy storage system and distributed generators, with a focus on minimizing capital and operational costs. A two-stage coordinated energy management strategy is used to minimize operating costs while meeting various system constraints.