A new machine learning based optimal bidding strategy for virtual power plants with solar power generators

Recently, virtual power plants (VPPs) are pioneering renewable based power systems that leverage progressive technologies to optimal operation of distributed energy resources (DERs) within an integrated platform. VPPs can manage and apply various energy resources such as solar panels, wind turbines, battery storage systems, and demand response programs. Then, increasing of the photovoltaic energy storage systems (PVESS) participation in the electricity market (EM) has increased considerably but the several limitations of the bidding mechanism for PVESS (like: the uncertainty of photovoltaic generation, the limitation of the bidding ability, and the single-revenue structure), will seriously affect its market revenue. This paper proposes an optimal strategic model for VPPs in balancing and day-ahead electricity markets, integrating solar power, energy storage, and conventional generators to mitigate solar fluctuations and imbalance charges. A hybrid machine learning forecasting model predicts solar generation and electricity prices on a day-ahead basis. Simulation results show that submitting a single optimized bid maximizes expected profits, demonstrating the framework’s potential to enhance availability and profitability of VPP operations in electricity markets.