电力系统自动化2025，Vol.49Issue(21)：74-84,11.DOI:10.7500/AEPS20241209002
基于图神经网络预测与主从博弈的虚拟电厂调频套餐设计
Frequency Regulation Package Design for Virtual Power Plants Based on Graph Neural Network Prediction and Stackelberg Game
摘要
Abstract
To improve the benefit and user satisfaction of virtual power plants in the frequency regulation auxiliary service market,a design method for the frequency regulation package based on graph neural network and Stackelberg game is proposed.Firstly,by comprehensively considering the frequency regulation capacity,participation frequency,and settlement methods,diversified packages that meet the frequency regulation participation characteristics of various users are designed to provide personalized choices for users.Secondly,a graph structure model of users,packages,and frequency regulation attributes is constructed by using the graph neural network to capture the complex relationships between users and packages.The model dynamically updates the features of nodes and edges to predict the package selection results of users.In response to the predicted user package selection results,a bi-level optimization model with multi-parameter Stackelberg game is further introduced.Taking the virtual power plant as the leader and the users as the followers,the optimization of package parameters is conducted with the objective of maximizing their respective benefits of the virtual power plant and users,ultimately achieving a balance and maximization of benefits for both parties.Experimental results show that packages with personalized customization and optimal design,not only enhances the flexibility of user choices,meets personalized needs,and enhances market adaptability,but also improves benefits of users and virtual power plant,achieving a win-win goal.关键词
新型电力系统/虚拟电厂/调频/套餐/图神经网络/预测/主从博弈Key words
new power system/virtual power plant/frequency regulation/package/graph neural network/prediction/Stackelberg game引用本文复制引用
江佳美,李嘉媚,艾芊,牛泽原,陈旻昱,杨璇..基于图神经网络预测与主从博弈的虚拟电厂调频套餐设计[J].电力系统自动化,2025,49(21):74-84,11.基金项目
云南省重大科技专项计划资助项目(202302AF080006). This work is supported by Major Special Science and Technology Project of Yunnan Province(No.202302AF080006). （202302AF080006）
