Optimal scheduling method and fast-solving algorithm for large-scale virtual power plants communication networks

As the scale of virtual power plants (VPPs) continues to expand, the communication demands between VPPs and the management center are increasing. To maintain the communication of the entire system, VPPs operators must pay high costs, and then, how to reduce communication costs as much as possible while ensuring VPPs communication requirements has become an important and difficult issue. However, in the existing literature, there are few scheduling methods for large-scale VPPs communications. To this end, this paper proposes an optimal scheduling method based on software-defined wide area network (SD-WAN) to reduce communication costs. First, the communication network architecture of large-scale VPPs is analyzed in detail, and communication services are categorized according to delay requirements. Second, for the most expensive wide area network layer, a communication network control structure based on SD-WAN is designed, and an optimal scheduling model is established to minimize communication costs while ensuring communication service quality. This model is formulated as a mixed-integer nonlinear programming problem, and then linearized and constraint-relaxed to enable solved by the state-of-the-art solver (i.e., Gurobi). Third, to further overcome the challenge of solving large-scale problems, such as low computation efficiency and memory overflow, a two-stage fast-solving algorithm is proposed, which sorts and categorizes VPPs branch sites and optimizes the problem in two stages, enabling the expedited resolution of the problem. Numerical tests verify the effectiveness of the proposed method. Especially for large-scale VPPs, the proposed algorithm improves computation efficiency by a thousand times without perceivable degradation in performance, compared to the state-of-the-art solver.