Two-stage dynamic aggregation involving flexible resource composition and coordination based on submodular optimization

• Published in: Applied energy Vol. 360; p. 122829
• Main Authors: Ding, Zhetong, Li, Yaping, Zhang, Kaifeng, Peng, Jimmy Chih-Hsien
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 15.04.2024
• Subjects: Dynamic aggregation; Primary frequency regulation; Submodular optimization; Virtual power plants; Virtual power plants; Primary frequency regulation; Submodular optimization; Dynamic aggregation
• ISSN: 0306-2619; 1872-9118
• DOI: 10.1016/j.apenergy.2024.122829

Traditional virtual power plants (VPPs) with fixed resource composition and coordination strategies struggle to cost-effectively exploit the flexibility of large-scale resources for adapting variable regulation requirements and resources characteristics. To this end, this paper proposes a dynamic aggregation mechanism to flexibly select and coordinate individual resources for forming aggregators according to grids regulation requirements and resource characteristics. The proposed mechanism is operated through a two-stage dynamic aggregation model comprising resource selection and coordination. Considering the two-stage dynamic aggregation model is a combinational optimization problem with high computational complexity, the submodular optimization method is utilized to swiftly address this problem. First, the complementarity and submodularity of the dynamic aggregation process are formulated to elaborate how the aggregation regulation characteristics (ARCs) evolve with flexible resource composition and coordination. Next, a submodularity-based algorithm is developed to promptly solve dynamic aggregation model under three scenarios, where aggregation operators focus on the resources quantity, quality, and cost-effectiveness, respectively. The polynomial computational complexity of the proposed algorithm has also been evaluated. Simulations using the IEEE 39-bus (New England) system consists of 10,000 flexible resources were executed to assess the submodularity approach. The proposed algorithm demonstrates superior computing speed and better performance guaranteed results (90%, 97%, 90% in three scenarios) compared to other methods—making it more suitable for implementation in practice. •The complementarity and submodularity of dynamic aggregation process are revealed and proved.•A two-stage dynamic aggregation model consists of resources selection and coordination is established.•A dynamic aggregation algorithm is proposed to deal with different aggregation scenarios.•The approximate guarantees and polynomial computational complexity of algorithm are proved.