Scheduling Optimization Of Electric Bus Charging Station With PV And BESS Considering Uncertainty

With the development of the country and the improvement of environmental protection awareness,traditional energy vehicles have been criticized for their high emissions and high pollution.My country has started relevant research and will formulate a timetable for the production and sale of traditional energy vehicles.Not only the ownership and market occupancy rate of new energy vehicles have grown rapidly in recent years,but also new energy vehicles have a positive impact on energy conservation and emission reduction,and improve the utilization rate of urban electricity.It is a reasonable replacement for traditional energy vehicles.As a "pioneer" in the promotion of electric vehicles in the Chinese market,the electrification rate of buses is getting higher and higher,and the impact of electric bus charging on the safe and stable operation of the power grid can’t be ignored.Therefore,how to optimize bus operation scheduling,reduce charging costs and make reasonable adjustments electricity load has been explored by enterprises and scholars.In addition,setting up distributed photovoltaic and energy storage facilities in bus charging stations can improve energy utilization,reduce carbon dioxide emissions,and reduce the demand for grid electricity by buses.Therefore,in view of the uncertainty of photovoltaic power generation and the random charging demand of electric buses,thesis studies the operation law of the photovoltaic storage charging station for buses,and proposes a real-time optimal charging and discharging strategy based on a deep reinforcement learning algorithm.The specific work arrangement as follows:Firstly,combined with the key scientific and technological problems that need to be solved urgently in the national economy and social development,the application prospect is discussed,and the research literature on the operation optimization and energy coordination optimization of electric bus charging station is clarified.Secondly,the theory and algorithm of reinforcement learning for intelligent energy management are introduced.According to the system structure of the bus optical storage charging station,the operation model of the electric bus charging station is studied,and the optimization objective including the charging and discharging cost,power fluctuation and other penalty costs is established,and the clear safety operation constraints and operation requirements are considered as the algorithm.It is the basis for optimization.Thirdly,transform the charging and discharging optimization process of electric bus PV-BESS Charging station into a Markov decision process,and use the Proximal Policy Optimization algorithm in deep reinforcement learning to interact with the environment and adapt to the uncertainties of environment.In order to reduce the number of combinations of action spaces and make the algorithm easier to converge,the charging and discharging selection strategy based on decision time is proposed.In order to satisfy the charging demand and avoid the heuristic design of penalty cost,a charging and discharging control algorithm suitable for deep reinforcement learning model is proposed.Then,in a practical example,the feasibility and superiority of the Proximal Policy Optimization algorithm are proved by comparing with the other three deep reinforcement learning algorithms,and the advantages of the proposed algorithm and the Proximal Policy Optimization Algorithm are verified.The sensitivity of initial charge state and V2 G compensation coefficient is analyzed.Finally,thesis summarizes the innovations and shortcomings of the research on the scheduling optimization of electric bus PV-BESS Charging stations considering uncertainty,and points out the future work prospects.