Optimization Of Charging Infrastructure Siting For Electric Taxis In Beijing

With the environment problem and the energy shortages being more and more obvious in recent years,all of the countries are highly enthusiastic in the study of energy consumption restructure and environment-friendly technology.For the reason that traditional fuel vehicle is a major resource of environment pollution and energy consumption,the electric vehicle driven by clean electric power energy is concerned and favored widely,and a huge number of electric vehicles has been put into use at home and abroad as an alternative for traditional transportation.However,due to the limitation of electric vehicle technology such as a short endurance mileage and a relative long charging time,the construction mode and siting density of electric vehicle charging station cannot use the traditional ones for gas station,so the constructed charging stations are hardly to satisfy the highly increased charging demand of electric vehicle.It has been a critical barrier for further extension of electric vehicle's use and development.Aiming at charging infrastructure siting problem for electric taxis in Beijing,this thesis studies the travel behavior characteristics of both passenger and electric taxi,and applies them to Monte Carlo simulation to simulate the process of charging demand emergence.Then identifies the distribution of charging demand on time and space which can satisfies the normal travel activities of passengers under present situation.The service ranges and capacities of charging stations are divided by voronoi diagram,and they constitute the constraints of the charging station siting model,under the known charging demand distribution,while the construction and operation cost function constitutes the objective function of the siting model.The solution of the less constrained siting model is obtained with Basic Particle Swarm Optimization algorithm,and the validity of the less constrained siting model is proved by contrast with the solution of the P median siting model.Meanwhile,Tabu Particle Swarm Optimization algorithm is introduced to increase the solution accuracy.After the sensitivity analysis on the parameters which influences the results of the model,an optimal solution of the siting model has been found.The thesis used the probability distribution of travel distance,departure time and battery state of charge to make sure the travel events obeying the real events' distribution,so that the complexity of travel trace of electric taxi is avoided,the difficulty on data capturing is reduced,the parameter redundancy is decreased and the accuracy of charging demand prediction is improved.A less constrained siting model is established to decrease the modeling constraint conditions.It decreases the complexity of constraints and simplifies the siting model while keeping the accuracy of the solution.It also ensures very few adjusts on parameters facing with outer condition changes to increase the universality and ease of use.Tabu-PSO algorithm is applied to solve the siting model to increase the accuracy of solutions while keeping the speed of the algorithm and avoiding it getting into local optimum.It provides a method to solve the less constrained siting model with a satisfactory solution.