摘要
随着新型电力系统发展,接入海量分布式对象的配电网在信息感知、分析、控制能力等方面都面临挑战,将成为边缘计算重要的应用场景。为了更好地满足配电网业务的实时性要求,提出了基于Kubernetes(简称K8s)的配电边缘计算终端资源动态配置方法。首先,针对配电网计算资源弹性伸缩需求,建立了基于K8s技术的配电边缘计算终端模型;然后,基于配电网业务与Pod集合间的对应关系,提出了以Pod集合为伸缩单元的资源弹性伸缩算法;最后,在MATLAB上进行建模仿真,构建业务请求种类和数量各不相同的配电网业务场景,讨论了不同弹性伸缩机制和不同弹性伸缩系数下配电网业务处理结果。仿真结果表明,以Pod集合为伸缩单元且设置合理的弹性系数能够有效降低业务时延,验证了该方法的有效性。
With the development of new power systems,distribution network connected to massive distributed objects face challenges in terms of information sensing,analysis,and control ability,and will become an important application scenario for edge computing.In order to better meet the real-time requirements of distribution networkbusiness,a resource dynamic allocation method for K8s based distribution edge computing terminal is proposed in this paper.Firstly,a K8s based distribution edge computing terminal model is established considering the auto-scaling demand of computing resource in distribution.Then,the corresponding relationship between the services of distribution and Pod set is noted,and a resource auto-scaling algorithm taking Pod set as scaling unit is proposed.Finally,the modeling and simulation are carried out on MATLAB to construct the distribution service scenarios with different types and quantities of business requests.The microgrid service processing results under different auto-scaling mechanisms and different auto-scaling coefficients are discussed.The simulation results show that taking Pod set as the scaling unit and choosing reasonable auto-scaling coefficient can effectively reduce the delay,verifying the effectiveness of the proposed method.
作者
司徒友
苏俊妮
张鑫
苏忠阳
SITU You;SU Junni;ZHANG Xin;SU Zhongyang(Dongguan Power Supply Bureau,Guangdong Power Grid Co.,Ltd.,Dongguan 523000,China;Guangzhou Suihua Energy Technology Co.,Ltd.,Guangzhou 510663,China)
出处
《供用电》
2023年第8期49-57,共9页
Distribution & Utilization
基金
中国南方电网有限责任公司科技项目[031900KK52200012(GDKJXM20200441)]。
关键词
弹性伸缩
配电网
边缘计算终端
K8s
资源动态配置
auto-scaling
distribution network
edge computing terminal
K8s
resource dynamic allocation