Study On Location And Capacity Optimization Of Public Charging Stations Based On GPS Trajectory Data Of Electric Vehicles

For the purpose of energy saving and environmental protection,electric vehicles have been widely concerned in the world,showing a rapid growth trend.However,the expensive purchase cost prevents the sinking of electric cars with long range,users are often anxious about remaining range.The increasing charging demand and long charging time exacerbate congestion at public charging stations,which further reduces the user’s driving experience and restricts the development of the electric vehicle industry.At the same time,the current situation of charging service supply and demand does not match,resulting in low utilization rate of charger.Therefore,it is particularly urgent to systematically design the location and capacity of public charging stations based on current charging demand and future potential demand.In view of this,based on the mining of GPS trajectory data of electric vehicles,this paper discusses the locating and sizing optimization of public charging stations,and studies the multiperiod expansion and innovative service mode of public charging stations under dynamic demand.The specific work includes the following aspects:Firstly,the locating and sizing optimization of public charging stations were studied considering the perception of charging service response time of electric vehicle users.Based on the GPS trajectory data set of electric taxis,the journey was disassembled to identify the three seed tracks of seeking charging station,charging and free driving.Taking the charging service response time as the time threshold for electric vehicle users to obtain satisfactory charging service,based on existing charging facilities and budget constraints,and combined with queuing theory,an optimization model of public charging station locating and sizing optimization model was established.Then,based on the characteristics of the model,a bi-layer nested genetic algorithm was developed to solve the model,in order to find a balanced solution to the problem of charging facility finding difficulty and charging congestion.The performance of greedy algorithm and greedy replacement algorithm is improved in large-scale stochastic examples.In addition,based on the Shenzhen case study and the sensitivity analysis of key parameters of the model,suggestions are made for the optimal deployment of public charging stations.Secondly,in order to promote the market penetration of electric vehicles,the multi-period expansion of public charging stations under dynamic demand is studied.Considering that charging distance and waiting time will affect purchasing decisions of electric vehicles,this paper proposes a variation function of the market penetration based on public charging opportunities and waiting time satisfaction.Then,the charging demand distribution is predicted by simulating the change of electric vehicles market share and difference of charging demand among different vehicles.Furthermore,multi-period locating and sizing optimization model is constructed to adapt to the dynamically charging demand and observe the interaction between charging demand and facility supply.In order to solve large-scale network problems,the proposed model is divided into two independent stages: site selection and capacity determination.In this paper,GA-k-medoids heuristic algorithm is designed by grafting genetic algorithm and k-medoids clustering algorithm.In the example analysis part,this paper linearizes the locating model,and verifies the effectiveness of the proposed model and algorithm by comparing the results of heuristic algorithm with Gurobi’s.With the success of numerical experiments,the multi-period deployment strategy of public charging stations under dynamic demand scenario is discussed in a case study of Shenzhen.Finally,the two-level decision model is designed to study the deployment strategy of the charging/changing joint station with the consideration of path deviation and range anxiety.Under the dual influence of path deviation and range anxiety,electric vehicle drivers will choose the appropriate path to the destination.Based on this,this paper constructs the flow attenuation function of electric vehicles.Then,aiming at the mode of combining fast charging service and battery replacement service,a dual queuing system with priority is established,including a slow internal charging system for battery replacement and an external charging system for fast charging of electric vehicles that missed the opportunity of battery replacement.Thus,a two-level programming model is constructed.The upper model studies the location-path optimization of the combined stations.On this basis,the lower model combined with the dual queuing system and budget constraints to optimize the allocation of spare batteries and fast chargers,so as to reduce the average sojourn time of vehicles in the station.In terms of theory,this paper focuses on the impact of electric vehicles users’ time perception on demand coverage,and enriches the relevant theories of public charging station locating,sizing and route optimization to a certain extent by combining range anxiety and path deviation factors;This paper focuses on the development of effective spatial-temporal associated locating and sizing model that can cope with highdimensional inputs;The proposed bi-layer nested genetic algorithm and GA-k-medoids heuristic algorithm provide new ideas and directions for solving nonlinear models containing both 0-1 decision variables and integer decision variables;The concept of combined charging/switching station breaks the category boundaries of charging facilities,and its dual queuing system with priority further highlights the value of research problems from the theoretical level.In practice,analyzing the driving behavior and charging behavior of electric vehicles users based on GPS trajectory data is conducive to grasping the more realistic demand for fast charging demand and improving the utilization rate of public chargers;Based on the short-term and looking to the long-term development of electric vehicles,this paper focuses on solving the problem of charging network optimization,which will inspire enterprises to seek solutions to charging problems from the perspective of development;By including the optimization of the location and capacity of public charging stations into the study,it can not only alleviate the range anxiety of travelers due to the limited range of electric vehicles,but also help solve practical problems such as queue congestion caused by insufficient chargers.