电力需求侧管理2025，Vol.27Issue(6)：51-57,7.DOI:10.3969/j.issn.1009-1831.2025.06.008
基于近端策略优化的虚拟电厂低碳经济调度
Low-carbon economic scheduling of virtual power plant based on proximal policy optimization
摘要
Abstract
With the deepening of"dual carbon"process,the transformation of power system to green and clean is imperative.The massive distributed resources in virtual power plant aggregate distribution network can realize the goal of power decarbonization and green transfor-mation through flexible load regulation.Based on this,a virtual power plant economic optimization model based on near-end strategy opti-mization is proposed.Firstly,based on the principle of proportional sharing,a carbon flow model is constructed to track the carbon density of each node in real time.Then,the low-carbon economic dispatching optimization objectives of virtual power plant are constructed,includ-ing load adjustment cost and carbon emission cost.Finally,the proposed model is solved using the proximal strategy optimization algo-rithm.The example analysis shows that the proposed low-carbon model can realize the carbon flow tracking in the whole time scale on the basis of guaranteeing the power flow safety of the distribution network.Moreover,the proposed near-end strategy optimization algorithm can realize the low carbon scheduling of internal resources in virtual power plant while ensuring economy.关键词
虚拟电厂/近端策略优化/深度强化学习/碳流网络Key words
virtual power plant/proximal strategy optimization/deep reinforcement learning/carbon flow network引用本文复制引用
杨娜,刘丽,宋梦,赵晨..基于近端策略优化的虚拟电厂低碳经济调度[J].电力需求侧管理,2025,27(6):51-57,7.基金项目
国网安徽省电力有限公司科技项目资助"安徽绿色低碳电力市场体系建设研究"(52120922000D) （52120922000D）
