AI-assisted modeling and optimization of integrated distributed energy systems

Decarbonization of the energy sector and transition to cleaner alternatives are essential steps toward mitigating climate change. In pursuit of this goal, distributed energy systems (DES) play a vital role in the sustainable development of energy infrastructure. Well-designed DESs facilitate the integration of renewable resources, enhances energy security, strengthens grid resilience, and improves overall energy efficiency. The proposed DES in this study integrates multiple renewable and energy conversion technologies, including photovoltaic (PV) system, wind turbine (WT), solar thermal collector (STC), combined heat and power system (CHP), and energy storage systems. It is also connected to the power grid, and the energy storage system includes battery for electric energy storage and hot water tank for thermal energy storage. An artificial neural network (ANN) model trained using real data is utilized to model the PV system for the DES. The integration of various energy generation and storage technologies into a distributed energy system presents challenges spanning both design and operational domains. Therefore, this paper focuses on artificial intelligence (AI)-assisted modeling and design and operation optimization of an integrated distributed energy system using a multi-objective genetic algorithm. To demonstrate the viability of the proposed optimization framework, a detailed case study for a large hotel building in Colorado is carried out. The performance analysis of the optimal distributed energy system is presented and compared with that of the conventional system. The optimized DES design and operation strategy achieves annual savings of $53,476, and 332 tons/yr of reduction in carbon dioxide emission relative to the baseline.