| As a green transportation tool,electric vehicles are very important in the process of achieving the goal of "carbon neutrality and peak carbon dioxide emissions".With the increase in the number of electric vehicles year by year,the charging of large-scale electric vehicle clusters will have a negative impact on the safe operation of the power grid.At the same time,electric vehicles can be used as a flexible load resource,which can participate in the demand side response of the power grid,cut peaks and fill valleys on the power grid,and ensure the safe operation of the power grid.Accurate characterization of electric vehicle travel behavior and accurate prediction of charging demand is an important basis for the power grid to optimally dispatch electric vehicle clusters.However,the randomness and uncertainty of electric vehicle travel behavior make it difficult to predict charging load demand.In recent years,big data technology has continued to develop,and data-driven methods are used to describe the travel behavior of electric vehicle users,overcoming the inaccurate problem of traditional methods to describe travel behavior.It makes the prediction result of charging demand load more accurate,and also provides a load basis for the optimal dispatch of electric vehicle clusters.Therefore,based on the data-driven method,this dissertation predicts and optimizes the dispatch of electric vehicle cluster load,and guides electric vehicles to participate in the regulation of the power grid.The main research contents are as follows:(1)Modeling the travel behavior of electric vehicle users based on data drivingFirstly,the real electric vehicle travel trajectory data is selected for cleaning and preprocessing,the abnormal data is deleted and the trajectory data points of each trip of the electric vehicle are sorted out in terms of the number of trips,and the time distribution and starting point distribution of each trip are obtained through data visualization.Then,the research area of this dissertation is selected,divided into spatial grids of the same size,the trajectory data points of electric vehicles are mapped on the spatial grid,the OD matrix of electric vehicle travel is obtained,and the functional areas of different spatial grids are divided by POI point of interest index.Finally,the travel time distribution of different types of electric vehicles and the probability distribution of travel time of electric vehicles in different functional areas are obtained through example simulation.(2)Charging demand load prediction of electric vehicles based on the coupling model of "transportation network-charging station-distribution network"The prediction of electric vehicle charging demand load is the premise of optimal scheduling,on the basis of modeling user travel behavior,the "traffic network-charging station-distribution network" model is first constructed,the transportation network considers the road resistance model,the charging station considers the dynamic queuing number and queuing time model,and the distribution network adopts the IEEE33 node standard model and uses the offset rate and network loss rate to evaluate the safety and economy of the power grid.Next,we will introduce the types of electric vehicles,select private cars and taxis for research,and introduce characteristics such as battery parameters,mileage consumption,and how to trigger fast or slow charging.Then,Monte Carlo simulation is used to predict the charging load demand of electric vehicles.Finally,the calculation simulation is used to obtain the charging demand load distribution of each functional area and each transportation node,and the impact on the power grid when large-scale electric vehicles are connected to the grid.(3)Optimal scheduling of electric vehicle clusters considering the adjustability of charging behavior characteristicsBased on user travel behavior and load forecasting of charging demand,large-scale centralized and fast charging electric vehicles charging demand load is selected to study.Not all electric vehicles meet the conditions of demand-side response participating in the power grid.Therefore,adjustability identification of electric vehicles must be carried out first.Firstly,the charging behavior characteristics of electric vehicles are analyzed.The charging time characteristics and charging preference characteristics of electric vehicles are deduced from the real travel trajectory data,and the charging preference characteristics of electric vehicles are depicted from four aspects:vehicle parameters,response willingness,charging cost and incentive mechanism.Then,build charging physical matrix and user willingness matrix,identify the adjustability of electric vehicles using Bi-GRU model,get the adjustable and non-adjustable clusters of electric vehicles,and calculate the dispatchable and non-dispatchable capacity.Then,for the dispatchable capacity of the adjustable electric vehicle,the objective is to maximize the profit of the electric vehicle.Finally,the example simulation shows that the optimal dispatching strategy reduces the load fluctuation to a certain extent and improves the profit of the electric vehicles load aggregator. |
