Research On Short-term Load Forecast And Scheduling For Charging Station Of New Energy Electric Vehicles

Electric vehicles have obvious advantages in terms of environmental protection and consumption of new energy.However,large-scale electric vehicle access has had a huge impact on distribution network load fluctuations,especially in the case of changing seasons,the charging load of electric vehicles fluctuates drastically.It has brought difficulties to the daily scheduling of the distribution network,so it is necessary to conduct research on the charging load of electric vehicles under sudden weather changes.This paper studies the load forecasting of large-scale electric vehicle access charging.Firstly,through a large amount of data research,the data characteristics of charging load of electric vehicle charging station are analyzed,and the main factors affecting the charging load of electric vehicle are determined.From the characteristics of charging battery of electric vehicle,charging mode and user charging mode,electric vehicle charging under different day types is analyzed.A load model that builds a collection of similar daily load samples in a catastrophic weather.Then consider the complex weather factors,use gray correlation analysis and data mining method to analyze the pre-correlation degree of load forecasting data,construct the correlation matrix of load and meteorology,and then establish a neural network prediction model with multiple inputs and single outputs.A three-stage electric vehicle charging load clustering integration prediction is proposed.Firstly,the genetic algorithm(GA)is used to optimize the parameters of the adaptive neuro-fuzzy system(ANFIS),then the ANFIS algorithm is used to predict the load,and the forward feedback neural network(FFNN)algorithm is used to comprehensively consider the meteorological correlation matrix.Optimize again to improve the accuracy of electric vehicle charging load prediction.this paper proposes a new combined forecasting model,which has the following improvements: 1)Considering the daily type factors and meteorological factors of the load,the grey relational analysis and data cluster mining are used to predict the data;2)Considering the accuracy and stability of the load forecasting process,the combined intelligent algorithm is used to perform multi-layer dynamic adjustment combined prediction results.Then,based on the historical load data of a charging station in Shanghai,the BP-ES,BP-GA and GA-ANFIS-FFNN algorithms are used to predict the load of the example and compare the results to verify the accuracy and superiority of the electric vehicle charging forecast in the season.Finally,aiming at the situation that the distribution network loss and the peak-to-valley difference become larger due to the disordered charging and discharging of electric vehicles,the research on the strategy of electric vehicle participating in the optimal dispatching of distribution network is proposed.The electric vehicle based on artificial fish swarm algorithm with minimum and peak-to-valley difference is involved in the optimization model of distribution network scheduling,and simulation scheduling optimization is carried out in IEEE33 node system and some examples in Shanghai to verify the effectiveness of electric vehicles participating in distribution network scheduling.