Two-stage distributionally robust bidding strategies for multi-energy virtual power plants in energy, reserve and carbon markets

The integration of multi-energy coupling and the involvement of multi-energy virtual power plants (MEVPPs) in the electricity market as a form of energy storage resource introduce escalating challenges for the bidding strategies of MEVPPs. A multi-market trading model not only provides market entities with greater choices, but also puts higher demands on their bidding strategies. This article proposes a distributionally robust optimization (DRO) model based on moment information for the market bidding of MEVPPs. To tackle the above issues, this article first designs a bidding framework and trading process for MEVPPs to participate in the energy, frequency regulation, and carbon markets simultaneously, and establishes a bidding optimization strategy for aggregator participation in the three markets, accounting for the inherent unpredictability in the generation of renewable energy sources and electricity market prices, in order to determine its comprehensive clearance strategy. Secondly, this paper analyzes the probability distribution of renewable energy output prediction errors and typical scenarios of electricity market prices, constructs a two-stage DRO operation model for MEVPPs, and searches the optimal day-ahead bidding and scheduling decisions under the worst scenarios. Finally, the efficacy of the suggested framework was confirmed through the application of a VPP case study. In that: 1) The newly developed strategy can reach the solution of multi-energy bids with execution times satisfying the above three market timelines. 2) Compared with the single market scenario, MEVPP can reduce the system cost by up to 15.6 % in the joint bidding of multiple markets.