Research On Energy And Computing Resource Scheduling Technology For Electric Vehicles In Park Microgrid

With the increasing global demand for sustainable energy,microgrids in industrial parks have become an effective way to achieve energy independence and reduce carbon emissions.In microgrids,electric vehicles(EVs)are energy-information cross-domain entities with both electrical energy and computing resources.They can serve as energy storage units to provide power scheduling services in the energy network domain,and as edge computing nodes to provide computation offloading services in the information network domain,which provides important support for the efficient operation and energy utilization of microgrids.However,during the integration process of EVs and microgrids,two main problems arise.Firstly,as EVs belonging to private users often have self-interest,their willingness to interact is relatively low,and they are unwilling to disclose their private information,which causes information asymmetry between the two parties.Secondly,after large-scale EVs are randomly connected to the microgrid,it is difficult to formulate accurate day-ahead scheduling plans to ensure the economic and safe operation of the microgrid due to the uncertain electricity demand of EV owners and renewable energy output.Therefore,this thesis conducts research on the above problems,and the specific research content and main contributions are summarized as follows:Firstly,to address the problem of information asymmetry between EVs and park microgrids in resource trading,a reverse selection incentive scheme based on contract theory is proposed.By evaluating and grading the service capabilities of EVs in different geographical locations during driving,an optimization problem for electricity and computing contracts is established with the goal of maximizing the total utility of EV aggregators.The optimal contract is obtained by solving it using the Lagrange multiplier method.Simulation results show that the optimal contract can overcome the information asymmetry between EVs and aggregators and that a contract scheme considering both electricity and computing resources is more beneficial than one considering only a single resource.Secondly,to address voltage fluctuations or current overloads in the grid caused by large-scale unordered access of EVs to park microgrids,a two-layered scheduling strategy for EV resources considering power flow constraints is proposed.The upper layer aims to minimize the cost of electricity purchases by optimizing the power of each electric vehicle charging station.Due to complex power flow constraints,this scheduling problem is modeled as a Constrained Markov Decision Process(CMDP)and solved using Lagrange multipliers and SAC(soft actor-critic)algorithm.The upper layer scheduling results affect the constraints of lower layer scheduling which aims to maximize computing services provided by EVs for smart terminals by optimizing their electricity and computing resources.Finally,based on these two research points on theoretical studies of EV resource management methods,an integrated scheduling system for EV electricity and computing resources was designed and implemented.This system provides an effective support platform for EV resource trading with good application value.