CVaR-based bi-level bidding strategy for VPP to participate in energy-regulation-carbon markets

With the increasing integration of renewable energy into new power systems and the growing need for synergy among diverse flexible resources, virtual power plant (VPP) has become a crucial model for aggregating distributed resources to participate in multiple markets, thereby enhancing both economic and low-carbon benefits. This paper investigates the joint optimal bidding strategy for a VPP that aggregates photovoltaic (PV) systems, traditional energy storage, and electric vehicle (EV) across energy, regulation, and carbon markets. We propose a two-stage bi-level optimization model for VPP bidding in day-ahead and real-time markets, incorporating conditional value-at-risk (CVaR) to quantify uncertainties in renewable output and market prices. Risk aversion coefficients are dynamically determined using the Sharpe Ratio (SR) and its marginal change. Simulation results demonstrate that the proposed strategy effectively balances benefits and risks. Compared with participating solely in the energy market, the joint market strategy increases the total profit of the VPP by 65 %. Moreover, in contrast to a risk-neutral strategy that maximizes the SR, the proposed approach selects β = 0.7 based on the marginal change in SR. This reduces CVaR significantly by 24.8 % with only a 13.2 % decrease in expected profit, while maintaining a high SR of 1.174.