Research On V2G Scheduling Of Electric Vehicles Based On Dynamic Electricity Pricing In The Context Of Residential Areas

With the excessive consumption of fossil fuels,global environmental problems are becoming increasingly acute.Therefore,many countries are actively developing and utilizing new energy,and electric vehicle as an important part of the new energy use has received wide attention.With the rapid development of electric vehicle technology,its number has increased sharply.However,due to the daily travel patterns of electric vehicle users,there is a significant overlap between their charging time at private charging stations and the peak load of the power grid,which may lead to the load exceeding the rated value,and even threaten the safe operation of the power grid.In this paper,considering the uncertainty of electric vehicle charging under the background of a residential area,an electric vehicle V2 G dispatch strategy based on dynamic electricity price is designed,which can guide the charging behavior of electric vehicles reasonably,ensure the load balance of power grid and reduce the charging costs of users.The details of this paper are as follows.Firstly,this paper studies the influencing factors of electric vehicles participating in V2 G and establishes an electric vehicle behavior model through data analysis of electric vehicle travel behavior based on historical data of vehicle travel.This paper analyses the advantages of dynamic electricity price in solving the stochastic problem in EV dispatch compared with other electricity price modes and further presents a two-stage V2 G model for dynamic electricity price formulation by aggregators.This model combines the advantages of centralized and distributed,and achieves global optimum while fully considering the randomness of user charging,which provides a foundation for further research.Secondly,considering grid load fluctuations and user charging uncertainty during the V2 G scheduling process of electric vehicles,based on the demand side response of electricity price incentives,a peak-valley difference model of grid load,a charging and discharging cost model of electric vehicles,and their constraints are established;A two-stage reinforcement learning algorithm,Long Short-Term memory network and improved linear programming algorithm is proposed to solve the problems of peak to valley load difference and charging and discharging cost of electric vehicles.The first stage is the dynamic electricity pricing stage,which considers the fluctuation of power grid load and uses a Long Short-Term memory network algorithm to output dynamic electricity prices;In the second stage,based on the dynamic electricity price output from the first stage,considering the uncertainty of electric vehicle charging and discharging,an improved linear programming algorithm incorporating electricity price subsidies is used to conduct real-time scheduling of electric vehicle charging and discharging.Validate the advantages of the two-stage reinforcement learning algorithm,which can reduce the peak valley difference of power grid load and reduce user charging costs,while improving the robustness of the system.Finally,further considering that the actual power grid basic load is in a constantly changing state,a power grid basic load forecasting model is introduced into the above two-stage model,and a bidirectional Long Short-Term memory network-improved nonlinear programming algorithm is proposed.Considering the issue of user participation,a V2 G protocol that maintains user interests is added to the new two-stage model.At the same time,given the introduction of future predicted power grid load,a bidirectional Long Short-term memory network is adopted to replace the original Long Short-term memory network;In the objective function,considering the impact of electric vehicle charging and discharging on the battery,a battery degradation model is introduced.At this time,the objective function becomes nonlinear,and the linear programming algorithm in the original two-stage model is changed to a nonlinear programming algorithm.Validate the advantages of the bidirectional Long Short-term memory network compared to an ordinary Long Short-term memory network when introducing future values,the designed algorithm can achieve peak cuts in the power grid,and reduce user charging and discharging costs.