面向多 FPGA 实时仿真器的资源优化配置方法

The integration of high proportion of renewable energy and power electronic equipment under“carbon emission peak and carbon neutrality”goals brings challenges to the real-time (RT) simulation of power systems. To realize an efficient and accurate RT simulation of the new power system with limited computing resources and RT constraints, the partitioning and resource allocation method for RT simulation computing tasks based on field programmable gate array (FPGA) is studied. First, from the two dimensions of computing resource demand and actual computing time, the solution units and functional modules under different simulation architectures of the RT simulator are finely modeled, and then the spatial-temporal demand model of the simulation task is established. On this basis, the optimal allocation model of the multiple FPGA RT simulator is established. With the goal of giving full play to the optimal solution performance of the RT simulator, the large-scale complex simulation tasks are reasonably divided, and the logical resources, memory and other computing resources are optimally configured. Finally, the correctness of the proposed spatial-temporal demand model and effectiveness of the optimal resource allocation method are verified using the IEEE 123-node test feeder and a practical distribution system in China.