Power Grid Demand and Supply Interaction Oriented UAV-Assisted Mobile Edge Network

The dramatic integration of vast quantities of clean energy causes intense impact on power grid demand and supply balance. Due to the randomness and mass of demand and supply tasks along with the diversity of task requirements, existing data interaction methods have difficulty supporting the normal operation of power grid demand and supply. Aiming at solving existing four problems of movable load data collection, real-time generation of load classification data, security of load data privacy and limited network resource allocation, this article proposes the power grid demand and supply balance data interaction structure based on UAV-assisted mobile edge network and carries out related researches as follows. In preparation for UAV application physical basis, this article optimizes the deployment method of UAV groups. The UAV federated learning (FL) structure construction optimization method is proposed to execute computing tasks of real-time load classification data under the guarantee of privacy security. Finally, to meet the requirements of real-time performance and transmission accuracy of load control instructions, this article proposes the joint allocation optimization method of limited communication and computing resource. In corresponding simulation experiments, comparisons are performed among proposed and existing methods, such as simulated annealing algorithm and quality of service, with key performance metrics involved, such as convergence iteration time and value, in order to demonstrate the effectiveness and feasibility of proposed structure and methods.