Research On The Optimization Of The Scale Of The Electric Vehicle Charging Station Based On The Queuing Theory

Electric car with its advantages of green and low-power was being vigorously promoted, and there is a mutually incentive relationship between the charging infrastructure and the use of electric vehicles. This paper is about research on electric vehicle charging behavior and charging station scale optimization. The main work and conclusions are:(1)Comparative analysis on the advantages of electric cars and drawbacks of conventional cars were made, which revealed the long-term development foreground of electric cars and reflected the necessity and feasibility of related researches on electric vehicles. Research progress on the size and layout of electric vehicles, demand forecasting and some applications of queuing theory to optimize service facility configuration were reviewed, and it was concluded that research on the optimization of the service facility configuration combined with queuing theory is of great significance.(2)Analysis of the traditional car travel behavior characteristics based on the fourth and fifth traffic survey report of Shanghai City were made, from which the distribution law about the travel time and distance of electric vehicles were obtained. Methods, characteristics and applicability of the quantitative prediction were summarized. Considering the population, the scope of services, etc., the method combining exponential smoothing and linear regression was adopted to forecast pure electric car ownership for short and long term. Influence of electric vehicle mileage on charging demand was influenced, and a demand forecasting model of pure electric vehicle charging was established accordingly.(3)The derivation process of M (n)/M/c/N queuing model was introduced to describe electric vehicles queuing and charging behavior and modeling. Against features of objective function and constraints of the model, branch and bound method was adopted to solve. This algorithm and calculation steps were also introduced in details.(4)A numerical example was designed, correlation analysis was made and the pure electric vehicle charging demand was forecast. On the basis of numerical analysis on MATLAB, numerical simulation of JMT and microscopic simulation of AnyLogic, correctness and validity of M (n)/M/c/N queuing model were verified. Based on MATLAB software platform, the optimal solution of the empirical case using the branch and bound method was obtained. And the relationship among expectation of total cost of system in unit time and the service intensity, parameters reflecting arrival rate influence a, the charging station capacity N was discussed and analyzed. It can be concluded that when parameter a lies in the interval (0,1), the model can better describe the actual charging behavior; Under that condition, the closer the number of charger is to charging stations capacity, the smaller the extent of fluctuation is; when charging stations capacity is sufficient, the higher the service frequency is, the smaller model fluctuation is with parameter a in the interval(0,1).