Scheduling planning for virtual power plants based on an improved cost allocation method
The high proportion of electric vehicles(EVs)and distributed power sources would affect the power balance of the power system.Virtual power plants(VPPs)provide a new method to enhance the utilization rate of renewable energy and balance the load.Existing studies rarely address the cooperation between VPPs and EV charging stations(CSs)managed by different stakeholders.A cooperative operation framework is proposed for a multi-stakeholder VPP-EV charging station system.The conflict of interests of different stakeholders was resolved by theτvalue cost allocation method,while a feedback mechanism was established to collect stakeholders'opinions,continuously optimizing and improving the cost allocation scheme.A hierarchical reinforcement learning(HRL)algorithm was applied to the proposed model,to effectively overcome the challenges related to large state-action spaces and reward coefficient by decomposing the complex problem into multiple sub-problems and implementing control and optimization at different levels.Numerical cases were presented to demonstrate the effectiveness of the proposed method.
virtual power plantelectric vehicle charging stationmulti-stakeholderfeedback mechanismτ value cost allocation methodhierarchical reinforcement learning algorithmrenewable energy
国家科技期刊平台
NSTL
万方数据
维普
