Study Of Charging Station Location Planning Based On Spatial Load Forecasting And Electric Vehicles Fast Charging Demand Forecasting

With the policy of developing and utilizing clean energy implementing, the development of electric vehicles has been strong supported by the power industry. Along with the rising number of electric vehicles, new opportunities and challenges have been brought to power system. With a large number of electric vehicles connected to power grid, it is bound to give an impact to the planning and operation of power grid. Electric vehicles charging behavior can be divided into fast charging, slow charging and battery replacement. Because of the high charging power, and the relatively concentrated charge location, fast charging will bring more influence to power grid compared with slow charging and battery replacement, it is needed to optimize and control the fast charging behaviour. Fast charging is concentrated completed in the charging stations, so the problem is equal to locating and sizing planning of charging stations. The economy and reliability of power grid can be improved and the impact of electric vehicles fast charging can be reduced through researches, and there are quite important theoretical value and practical significance.Based on the economy and reliability of power grid, the charging stations location planning problem is researched. First, distribution network geographic information system is introduced as the data source, its data features and layer information is analyzed in detail. Utilizing geographic distribution data of power load obtained from it, spatial load forecasting is employed to forecast the load of the selected area using support vector machine algorithm optimized by particle swarm algorithm, to obtain the spatial distribution classification of power load. Combined with geographic feature information of the area, the conclusion of charging stations for site is drawn. Then road network model based on Poisson arrival location model, queuing vehicles reaching charging station dynamic model based on traffic flow theory and electric vehicles fast charging load demand prediction model based on queuing theory are built, to calculate the traffic flow of off-peak traffic period and traffic congestion period, and deduce the number of electric vehicles in the region arriving in charging station, then predict the fast charging demand of charging stations at different times. Last, distribution network economy and reliability indexes considering electric vehicles charging load are constructed to quantify the impact of electric vehicles charging brought to the grid. Considering the electric vehicles charging stations investment costs, distribution net loss costs, reliability economic costs,the sale of electricity benefits and environmental benefits in the aspect of economy, the charging stations locating and sizing planning is carried out. The risk index of electric vehicle charging station bringing to the power system is calculated to give the charging stations reliability assessment in the aspect of reliability. As a case study, the IEEE 33-bus system is simulated for simulating the scenarios of electric vehicles charging connected to the grid, which shows optimization schemes and evaluation conclusions of the charging stations locating and sizing planning.