Research On Intelligent Location And Capacity Determination Method Of Electric Vehicle Charging Station

With the gradual growth of China’s economic strength and per capita income,the number of motor vehicles is increasing year by year,the serious shortage of oil in China and the serious aggravation of air pollution make electric cars popular.Electric vehicles have the characteristics of low carbon environmental protection and economic efficiency,but the lagging charging station planning and construction will shackle the development of the whole electric vehicle industry.Reasonable electric vehicle charging station planning is directly related to the prospect of electric vehicle industry.Therefore,this paper studies the intelligent location and capacity of electric vehicles.First of all,we discuss the service objects and influencing factors,and establish a multi-objective and multi-constraint optimal charging station location and sizing model that min imizes the charging station construction cost,operation and maintenance cost,charging cost for service users,user driving power consumption cost and waiting time cost.Then,a BP(back propagation)neural network pre-location evaluation model aiming at reasonable land rent of charging station is established.The model takes transportation convenience,land cost,candidate land area,accessibility,spatial characteristics and population density as inputs,rents as output indicators,and the hidden layer contains 10 nodes.130 groups of addresses of electric vehicle charging stations in Taiyuan were selected to train the BP neural network pre-location model,and another 20 groups of addresses were used to verify the availability and feasibility of this evalu ation model.Secondly,In order to make the search speed of the ant colony algorithm faster and the search direction clearer,the ant colony algorithm is improved.In the initial stage of search,in order to speed up the search,only the ants with strong s earching ability are allowed to release pheromones.In the later stage of search,in order to prevent running into local optimum,all ants are allowed to release pheromones to achieve global update.Taking the traveling salesman problem as the simulation object,the rate of convergence and astringency of ant colony algorithm before and after improvement are contrasted,and the advantages of the modified ant colony algorithm is demonstrated.Finally,the BP neural network pre-location model and improved ant colony algorithm are applied to the location and capacity of electric vehicle charging stations in Luzhou District of Changzhi City and Pingcheng District of Datong City.Firstly,the BP neural network is used to evaluate the alternative location of the el ectric vehicle charging station,and the addresses with better location rent are obtained.These better addresses are used as the initial address solutions of the improved ant colony algorithm,and the address of the charging station is further optimized by the improved ant colony algorithm.At the same time,the optimal capacity of the charging pile at this address is optimized,so that the optimal solution of location and fixed capacity is obtained.The study indicate show that the model and algorithm suggested in this paper is able to solve effectively the problem of site selection and capacity determination of electric car charging station.