Research On The Operation Mechanism Of Electric Vehicle Agents In The Electricity Market

In order to deal with global warming,China put forward to such solemn goals as "strive to reach the peak of carbon emissions by 2030 and strive to achieve carbon neutrality by 2060"at this year’s two sessions(NPC and CPPCC).“Do a good job of peak carbon dioxide emissions and carbon neutral" has become a hot word in China.As the backbone of this work,electric vehicles have been supported by policy and technology.When large-scale electric vehicles are disorderly connected to the grid,it will impact the power grid and increase the difficulty of power grid dispatching.The promotion of Chinese electric market reform provides new business opportunities for load side resources.To flexibly control large-scale electric vehicles,electric vehicle agents emerge as the times require.How to make profit by bidding according to the current electricity market rules and how to schedule the charging and discharging of electric vehicles are the difficulties in the promotion of agent form.Based on the current rules of the electricity market,this paper studies the following three aspects: the next day travel prediction of electric vehicles,the optimal charging bidding strategy of electric vehicle agents,and the bi-level optimal bidding strategy of the Bayesian game.They solved the problems of prediction difficulty,profit difficulty and malicious competition of electric vehicle agents.The main work of this paper includes:(1)This paper starts with the difficulty of electric vehicle agent scheduling-Travel prediction.According to the relationship between various travel data(time away from home,time at home and travel distance),the difficulty of prediction and the difference between weekdays and weekends,this paper proposes an electric vehicle travel prediction model which is based on multi-layer machine learning algorithm.Compared with other common probabilistic statistical models,this model fully considers the impact of historical travel data on the prediction.So,it has better prediction accuracy.(2)This paper first analyzes the optimal bidding model of charging at the initial stage.This research is based on the current electricity market rules.The model aims at maximizing the revenue of electric vehicle agents and uses interval optimization method to find the optimal competitive value of agents.The example shows that this method can avoid the impact of electricity price prediction error.(3)When the market on the load side matures,more and more electric vehicle agents will participate in the market for charging and discharging.In this paper,we propose a reasonable charging and discharging mechanism based on the current electricity market.Based on this mechanism,a bi-level optimal bidding model of Bayesian game is established.Among them,the upper-level agent scheduling model with the goal of maximizing the interests of agents and the simulation clearing model with the goal of minimizing the generation cost.According to the Nash equilibrium theory,the existence of Nash equilibrium is proved,and the Nash equilibrium solutions of agents bidding are obtained by ergodic method.