Virtual power plant frequency regulation package recommendation based on graph neural networks

With the continuous growth of demand in the power frequency regulation market, users have placed higher expectations on diversified frequency regulation package options. This trend presents both new opportunities and challenges for package design and user recommendation. To address this, a frequency regulation package recommendation method based on Graph Neural Networks is proposed in this paper. First, a variety of frequency regulation packages are designed by comprehensively considering factors such as regulation capacity, participation frequency, and settlement methods, enabling the satisfaction of different user capabilities and settlement preferences and supporting personalized selections. Second, a graph-structured model is built a mong users, packages, and regulation attributes. Leveraging the powerful modeling capabilities of Graph Neural Networks, the method effectively captures the complex associations between users and packages under limited information, dynamically updating node and edge features to reflect users’ real-time needs and preferences, thereby enabling accurate package recommendations. Results show that the designed packages align well with users’ regulation participation dimensions, and that the proposed method can accurately recommend the most suitable package using Graph Neural Networks.