Research On Peak Shaving Control Strategy Of EV Clusters Based On V2G Technology

With the increasing energy shortage and environmental pressure in China,the gradual promotion of electric vehicles has become the only way to develop the automobile industry of China.Electric vehicles are not only clean and green,but also have special energy storage capabilities that can support the electrical system.If the storage capacity of a large number of electric vehicles is properly adjusted,the uncertainties existing in renewable energy in the power system can be alleviated.Thus,the permeability of new energy will increase effectively.In view of the peak cutting and valley filling requirements of electric vehicles,the grid-connected operation architecture of electric vehicle clusters was determined,and electric vehicles were regarded as mobile energy storage units.Starting from the grid-connected single electric vehicle,mathematical models were established for bidirectional DC/DC and AC/DC respectively and corresponding control strategies were proposed.On this basis,the V2 G simulation model was built.Simulation analysis was carried out in G2 V and V2 G operating modes,and it was verified that the model can realize the functions of V2 G and G2 V by user selection,and ensured bidirectional energy flow under the premise of meeting the requirements of grid-connection.Due to electric vehicles have the characteristics of large load,small capacity and scattered distribution,it is difficult to control the large-scale participation in peak load regulation of power grid.Classifying electric vehicles into centralized regulation and grid connection can effectively improve the speed of regulation and reduce the difficulty of regulation.The factors influencing the charging and discharging model of electric vehicles were analyzed,such as scale and type of electric vehicles,battery characteristics,user behavior characteristics,etc.Based on historical statistical data,a calculation model of electric vehicles charging and discharging power based on Monte Carlo method was established.The available charging and discharging power of electric vehicles in each time period was analyzed by calculation examples.In fact,considering the will of electric vehicle users and electric vehicles itself parameters,each electric vehicle involved in the power grid peak shaving time and available capacity is not the same.The convolutional neural network method was used to classify electric vehicle clusters,and then the available charging and discharging power at each time period was determined,which laid a foundation for electric vehicle clusters to participate in peak cutting and valley filling in the power grid.With the help of SPSS simulation platform,exponential smoothing method and RBF neural network were used to model and predict the power load in the next 24 hours.On this basis,the effective peak cutting and valley filling of electric vehicle clusters were analyzed and organized.With TOU power price as the background,aiming at reducing peak-valley difference of system load and reducing electricity cost of electric vehicle users,a multi-objective collaborative optimization scheduling model was constructed by comprehensively considering constraints such as battery capacity and charging-discharge power.The improved PSO algorithm was used to solve the problem.The simulation result showed that the optimal scheduling strategy achieved the goal of peak clipping and valley filling,reduced the load pressure of power grid and reduced the charging and discharging cost of electric vehicle users to a certain extent.