| 引用本文: | 孙宇嫣,蔡泽祥,马国龙,郭采珊.电力物联网云主站计算负荷模型与资源优化配置[J].电力自动化设备,2021,41(4): |
| SUN Yuyan,CAI Zexiang,MA Guolong,GUO Caishan.Workload model and optimal resource allocation of cloud master station in Power Internet of Things[J].Electric Power Automation Equipment,2021,41(4): |
|
| 摘要: |
| 随着电力物联网建设不断发展，运行分析、电能交易、用户能效管理、电源管理等业务应用呈现指数级增长，给传统主站系统带来巨大的计算处理压力。为了解决上述问题，针对新一代“云主站”架构，建立了电力物联网的计算负荷模型，并基于此提出了云资源配置方法。从“云主站”内计算负荷动态变化的根源着手，利用马尔科夫链刻画电网运行状态的变化规律；根据业务处理流程规范，建立状态关联的计算负荷模型，包括计算负荷描述模型及状态-事件-应用关联模型，并将其建立为经验知识库，用于分析不同时段各类业务的计算负荷特征；基于上述计算负荷模型，以平均响应延时最短、“云主站”能耗最小为目标，建立云资源优化配置双目标优化模型，通过逼近理想解排序法从帕累托前沿解集中获得综合最优解。以改进的IEEE 33节点电网为例，仿真结果验证了所提方法能有效分析不同应用计算负荷的变化特征，进而合理配置云资源。 |
| 关键词: 电力物联网 云主站 计算负荷模型 云资源配置 马尔科夫链 经验知识库 逼近理想解排序法 |
| DOI： |
| 分类号: |
| 基金项目:广东省重点领域研发计划项目(2019B111109002) |
|
| Workload model and optimal resource allocation of cloud master station in Power Internet of Things |
|
SUN Yuyan, CAI Zexiang, MA Guolong, GUO Caishan
|
|
School of Electric Power, South China University of Technology, Guangzhou 510640, China
|
| Abstract: |
| With the continuous development of PIoT(Power Internet of Things),business applications such as operation analysis, energy trading, user energy efficiency management, power management, and so on, show exponential growth, which brings huge computing and processing pressure to the traditional master station system. In order to solve the above problems, the workload model of PIoT is established for the new genera-tion of cloud master station architecture, and based on this, a cloud resource allocation method is proposed. Starting from tracing the root of the dynamic change of workload in the cloud master station, the change rule of power grid operation state is depicted by using Markov chain. According to the specification of business processing flow, the state-related workload model is established, including the workload description model and the state-event-application correlation model, which is established as an experience knowledge base for analyzing the workload characteristics of various businesses in different periods. Based on the above workload model, a dual-objective optimization model for cloud resource allocation is established with the goals of shortest average response time delay and minimum energy consumption of cloud master station. The comprehensive optimal solution is obtained from the Pareto frontier solution set by using TOPSIS(Technique for Order Preference by Similarity to an Ideal Solution). Taking the modified IEEE 33-bus system as an example, simulative results verify that the proposed method can effectively analyze the variation characteristics of workload in different applications, and then reasonably allocate cloud resource. |
| Key words: PIoT cloud master station workload model cloud resource allocation Markov chain experience knowledge base TOPSIS |
