Dynamic energy storage pricing for virtual power plants via bilevel game optimization under multiple uncertainties

Energy storage systems are crucial for enhancing Virtual Power Plant (VPP) profitability and operational flexibility. However, existing research predominantly focuses on external market bidding strategies, neglecting energy storage pricing as a core decision variable, which limits comprehensive characterization of price-driven internal VPP ecosystems. To address this gap, this study proposes a bilevel Stackelberg game model where the upper-level VPP operator establishes optimal charge and discharge prices as the leader, while lower-level users adjust their energy demand to minimize costs as followers. A hyper-parameter-tuned differential evolution algorithm is employed to solve this bilevel optimization problem. The approach is validated through Monte Carlo simulations across 200 distinct scenarios. Results demonstrate that the dynamic pricing model increases VPP's average daily revenue by 16.0%, of which physical energy storage arbitrage contributes 9.0% and the dynamic pricing strategy alone contributes an additional 6.4%. Value at Risk (VaR) analysis validates the strategy's robustness, achieving VaR 95 of 1254.3 RMB and VaR 99 of 1138.0 RMB that guarantee profitability even under extreme market conditions, demonstrating practical feasibility for real-world VPP deployments. • Bilevel Stackelberg game model treats storage pricing as core VPP decision variable. • Multi-uncertainty framework integrates renewable output and EV charging behaviors. • Dynamic pricing strategy increases VPP revenue by 16.0% with 6.4% from pricing alone.