Translated Title
Optimization of EV Charging Guidance Strategy Based on Multi-Agent Graph Reinforcement Learning
Translated Abstract
In order to solve a series of problems such as high charging cost caused by a large number of electric vehicles(EVs)charging in a complex power-transportation coupling system,inadequate utilization of charging station resources,and instability of the grid caused by a large number of EV loading connected to the grid,this paper takes into account the interests of EVs,charging stations and power grid,and proposes a multi-agent graph reinforcement learning(MAGRL)algorithm to optimize EV charging guidance strategy.First,under the power-transportation coupling system,a multi-objective EV fast charging demand model is constructed with EV charging time cost and economic cost,unbalance of charging station and average voltage deviation of power grid as multiple objectives.Then,a real-time online charging path guidance framework with dual timescale is constructed.On the slow timescale,second-order cone optimization is used to solve the optimal power flow of the distribution network to obtain the node marginal price,and on the fast timescale,a MAGRL algorithm is proposed to solve the charging strategy of electric vehicles.On this basis,in order to protect the privacy of EV users,a feature processing method is proposed of using independent information module to extract EV information.Finally,simulations are carried out on a regional 25-node traffic network and IEEE 33-node power system.The simulation results show that compared with the shortest path method,the proposed MAGRL algorithm can effectively reduce the time cost and economic cost of EV users,promote the balanced operation of charging stations and reduce voltage offset.In addition,it is verified that the proposed MAGRL algorithm can adapt to the scenarios with different charging needs,thus with good scalability.
Translated Keyword
charging guidance
electric vehicle
graph reinforcement learning
multi-agent
power-transportation coupling system
Access Number
WF:perioarticalhb-hbdl202403002
