A data-driven robust optimal day-ahead bidding strategy considering V2G operation for distribution-system-side virtual power plant

Virtual power plants (VPP) play a crucial role in electricity markets by optimizing energy consumption in a distributed environment. However, financial losses often arise due to increased forecast deviations caused by demand-side activities. This paper proposes a robust bidding strategy for a distribution-system-side VPP that integrates conventional loads, electric vehicles, and incentive-demand in the day-ahead market. A two-stage optimization model with vehicle-to-grid operations is developed to address uncertainties in electricity prices and grid loads. A scenario-based polyhedral uncertainty set, derived using data-driven methods, is employed to represent possible variations in these uncertain parameters. The optimization problem, formulated as a min–max-min model, is efficiently solved using strong duality theory and a column constraint generation algorithm. Case studies on a real-world campus in Shenzhen, China, demonstrates that the proposed approach increases VPP revenue by 17.9 %, 25.3 % and 2.4 % compared to stochastic programming, robust optimization and distributionally robust optimization, respectively.