Coordinated management under uncertainty and fair revenue distribution in a multi energy community

This paper investigates an energy community comprising residential users with controllable heating/cooling systems and distributed energy resources. A two-stage management is proposed with a day-ahead scheduling followed by near real-time correction of the controls to mitigate uncertainties. Decentralized model predictive control using an Alternating Direction Method of Multipliers (ADMM) algorithm is employed in both stages. Additionally, a monthly post-delivery reward system is suggested to evaluate each user's marginal contribution to community performance. The contributions are the real-time tuning of the models to limit the impact of equations inaccuracies, and the simulation of the operational phase while prior studies on communities focus on offline assessments with deterministic profiles. Also, the proposed methodology emphasizes technical objectives over economic considerations with i) the smoothing of the community power usage and ii) minimum deviations from committed profiles in real-time. Those objectives relate to the reduction of peak power generation and the need for energy reserve. Simulation results show a 55 % reduction in reserve requirements and a 35 % smoother community power profile, albeit with a 10 % increase in users' bills due to coordination. The post-delivery scheme evaluates each user's contribution to peak power reduction and energy deviation, facilitating fair revenue distribution based on energy and reserve prices.