Two-stage multi-objective optimal scheduling strategy for the virtual power plant considering flexible CCS and virtual hybrid energy storage mode

Promoting grid integration of renewable energy, exploring low-carbon retrofit, and aggregating flexibility resources at the same time are important means for virtual power plant (VPP) to balance low-carbon and economy under the "dual-carbon" goal. In this paper, a two-stage multi-objective optimal scheduling model of VPP considering flexible low-carbon retrofit and virtual hybrid energy storage expansion is designed. At the technical level, the carbon capture system (CCS) is first modified by decoupling and the flexible carbon storage mode is proposed. Then, in order to fully release the response potential of many heterogeneous resources, based on the electricity tariff elasticity relation (ETER) and the lumped equivalent thermal parameter (LETP) method, respectively, the virtual energy storage (VES) modeling method for electricity and thermal demand response (DR) is innovatively proposed, and Minkowski sum is introduced to represent the VES power range as an aggregated low-dimensional flexible region, and finally a virtual hybrid energy storage (V-HES) mode is designed by integrating the homogeneously modelled physical and virtual energy storage. At the scheduling level, a twostage distributed co-optimization strategy is proposed to construct the overall scheduling decision space, and the multi-objective optimization model is solved in equilibrium using a subject-objective weighting game model. The multi-scenario case analysis shows that: (1) The proposed flexible carbon storage mode improves the carbon market trading revenue by 31.24 %. (2) The proposed DR-oriented VES modeling method can improve the electric and thermal load response margins by 6.44 % and 7.50 %, respectively, and enable the full consumption of wind power. (3) Applying the two-stage multi-objective distributed co-optimization strategy improves the dayahead low carbon level by 10.12 %, improves the economy by 8.91 %, and reduces the intra-day system output deviation by 46.37 % compared to configuring each component individually. The feasibility and effectiveness of the methods and strategies proposed in this paper are fully verified.