| Smog and traffic jams are still major concerns in a crowded center of a big city and the global weather issue is a significant challenge.In order to meet the requirements in energy saving and pollution reduction,electric vehicles(EVs)are being promoted worldwide.The amount of electric vehicles has been continuously increased.The charging manner of electric vehicles is one of the bottleneck problems in further EVs’ development due to restrictions in battery capacity and basic infrastructure.With the progress in wireless power transfer,dynamic wireless charging and quasi-dynamic wireless charging technologies can supply driving power to the EVs in driving states or in temporary stops(such as bus stops or at intersections during a driving process through a non-direct contact approach in a real-time manner.The quasi-dynamic wireless charging requires less battery capacity,smaller size,and lower power,significantly reducing complexities in applications and investment in vehicles and infrastructure,which is an effective solution for the bottleneck problem of EVs’ development to promote EVs.The quasi-dynamic wireless charging at intersections is not only consistent with the matured signal light control means for traffic flows but also makes closer coupling between transportation networks and electric power systems.Under the favorable policies and technologies,it is necessary to perform deeper researches in coordinated optimization for integrated urban traffic and power distribution systems based on the coupling scenario of using quasi-dynamic wireless charging.No research work in this area has been found yet by current literature searching.Based on quasi-dynamic wireless charging for electric vehicles at intersections,this thesis investigates the coordinated optimization for urban traffic and power distribution systems and adequacy evaluation.The main contributions are summarized as follows:(1)Since no research work has been conducted in the area of coordinated optimization for integrated urban traffic and power distribution systems based on the coupling scenario of using quasi-dynamic wireless charging,a framework for this topic is first established,in which the difficulty of no simple analytical expression for modeling the topological optimization of traffic system and the deficiency of only using an approximate macro-representation for traffic flow analysis is resolved.A two-hierarchy and two objective-function modeling method for the complex integrated traffic and power distribution systems is presented,in which control variables are assigned to high-level and low-level optimization models respectively,resulting in reduced scales of the optimization models.In the high-level model,signal light settings are used as decision variables to control driving routes of vehicles.In the low-level model,other control variables including departing times,driving speeds,and street states are dynamically optimized to impact routings of each vehicle.These factors together affect and optimize charging locations and timings of EVs.The NSGA-II genetic algorithm is used to successfully solve the established multiple objective optimization problem that is non-convex and embeds a great number of differential dynamic simulations of vehicles.The case studies verified the effectiveness of the proposed framework,model,and algorithm.The objectives in both urban traffic and power distribution systems are co-optimized and the resulting outcomes are much better than those obtained by using any random path or static shortest path in space.(2)The basic model mentioned above is then extended from the representation for one single period to that for multiple periods,from double objectives to triple objectives,from modeling EV’s quasi-dynamic wireless charging timings to modeling both timings and charging time lengths,from optimizing distribution network power losses caused by EV’s quasi-dynamic wireless charging to co-optimizing both power losses and energy losses.Three optimization strategies for multiple time periods are developed.Based on the features of the presented model,a new improved BO-NSGA-II algorithm,in which Bayesian Optimization(BO)is combined with NSGA-II,is developed.The results in the case studies indicated the effectiveness of the proposed multi-period model,the three optimization strategies,and the new algorithm.The extended multi-period co-optimization can be reached by using either a multi-period single-setting optimization strategy or a multi-period multi-setting optimization strategy.The proposed BO-NSGA-II algorithm can effectively reduce iterations and enhance global convergence.(3)The probability distribution modeling techniques are presented for the number of departing vehicles in a given period and the number of vehicles stopping at different moments within the setting length of red lights at an intersection so that the stochastic characteristics of these two variables are simulated in the proposed co-optimization model for urban traffic and power distribution system.The results in the case studies verified the importance and necessity of considering these two probability variables.The normal distribution is more suitable than the uniform distribution for portraying the probability characteristics of the number of departing vehicles in a given period.Compared with Weibull distribution to fit the waiting duration of vehicles in front of red lights,the kernel density estimation leads to the results more accurate and closer to reality.(4)Based on the co-optimization modeling of urban traffic system and power distribution system,the operation adequacy indices and their evaluation and optimization methods for the integrated systems are proposed.The results in the case studies demonstrated the necessity and effectiveness of the adequacy indices,assessment method,and co-optimization model.The adequacy co-optimization model can reduce the deterioration in operation adequacy of both urban traffic and power distribution systems during a rash hour period under both normal and outage road conditions.The presented framework,model and algorithm for the co-optimization of urban traffic and power distribution systems can provide effective information for day-ahead dispatch,operation decision and,planning scheme in a short or long term.The work in the thesis is a significant step in the research for interactive effects between urban traffic and power distribution systems in the EV’s quasi-dynamic wireless charging scenarios. |
