Renewable Energy Community Sizing Based on Stochastic Optimization and Unsupervised Clustering

Renewable Energy Communities (RECs) are emerging as significant in the global paradigm shift towards a smart and sustainable energy environment. By empowering energy consumers to actively participate in local energy generation, and sharing, using renewable energy sources, energy storage, and flexible loads, REC participants can reduce costs, and also contribute to low-carbon objectives, providing the flexibility needed to address modern smart grid challenges. This article presents a mixed integer linear programming model for optimal sizing of the solar PVs and battery energy storage systems (BESS) of REC participants who engage in P2P energy exchange. The model is formulated using a two-stage stochastic optimization to address load and PV uncertainty, and unsupervised clustering to structure the data for the stochastic optimization process. The model enables sizing solar PVs for different rooftop geometries and the objective function includes comprehensively defined electricity, operational, and scaled investment costs for solar PV and BESS, where economic fairness constraints are analyzed and implemented. The model is validated on real solar and atmospheric measured data from Zagreb, Croatia, and publicly available household consumption data from Northern Germany. The article also analyzes how tariff models, and electricity prices affect PV and BESS sizes, cost reductions, and P2P energy exchange for different REC participants with varying consumption and production profiles.