Zhao Ying, Ye Hua
Yunnan Electric Power Grid Company, Kunming, 650011, Yunnan, China.
Sci Rep. 2023 Sep 5;13(1):14634. doi: 10.1038/s41598-023-41108-2.
As more and more intelligent devices are put into the field of power system, the number of connected nodes in the power network is increasing exponentially. Under the background of smart grid cooperation across power areas and voltage levels, how to effectively process the massive data generated by smart grid has become a difficult problem to ensure the stable operation of power system. In the complex calculation process of power system, the operation time of complex calculation can not be shortened to the greatest extent, and the execution efficiency can not be improved. Therefore, this paper proposes a two-phase heuristic algorithm based on edge computing. In solving the virtual machine sequence problem, for the main partition and the coordination partition, the critical path algorithm is used to sort the virtual machines to minimize the computing time. For other sub-partitions, the minimum cut algorithm is used to reduce the traffic interaction of each sub-partition. In the second stage of the virtual machine placement process, an improved best fit algorithm is used to avoid poor placement of virtual machines across physical machine configurations, resulting in increased computing time. Through the experiment on the test system, it is proved that the calculation efficiency is improved when the coordinated partition calculation belongs to the target partition. Because the edge computing is closer to the data source, it can save more data transmission time than cloud computing. This paper provides an effective algorithm for power system distributed computing in virtual machine configuration in edge computing, which can effectively reduce the computing time of power system and improve the efficiency of system resource utilization.
随着越来越多的智能设备被投入到电力系统领域，电网中连接节点的数量呈指数级增长。在跨电力区域和电压等级的智能电网协同背景下，如何有效处理智能电网产生的海量数据已成为确保电力系统稳定运行的难题。在电力系统复杂的计算过程中，复杂计算的运行时间无法最大程度缩短，执行效率难以提高。因此，本文提出一种基于边缘计算的两阶段启发式算法。在解决虚拟机序列问题时，对于主分区和协调分区，采用关键路径算法对虚拟机进行排序，以最小化计算时间。对于其他子分区，采用最小割算法减少各子分区的流量交互。在虚拟机放置过程的第二阶段，采用改进的最佳适配算法，避免虚拟机跨物理机配置放置不佳导致计算时间增加。通过在测试系统上的实验证明，当协调分区计算属于目标分区时，计算效率得到提高。由于边缘计算更靠近数据源，与云计算相比可节省更多数据传输时间。本文为边缘计算中虚拟机配置的电力系统分布式计算提供了一种有效算法，可有效减少电力系统的计算时间，提高系统资源利用效率。
