基于多目标粒子群的电力数据容灾技术研究  

Study on power data tolerance technology based on multi-target PSO

作　　者：韩霞[1] 郭易鑫 白建海[1] 王晖南 高强强 Han Xia;Guo Yixin;Bai Jianhai;Wang Huinan;Gao Qiangqiang(Shanxi Province Electric Power Corporation of State Grid,Taiyuan 030001,China;SGIT-GreatPower,Fuzhou 350003,China)

机构地区：[1]国网山西省电力公司,太原030001 [2]国网信通亿力科技有限责任公司,福州350003

出　　处：《国外电子测量技术》2018年第6期44-49,共6页Foreign Electronic Measurement Technology

摘　　要：当电力故障发生时,为了保障电力数据的安全性,提出了基于多目标粒子群算法的数据容灾技术。首先分析了数据容灾技术的架构,分析了HDFS存储技术,由于其具有高容错性,所以电力调度容灾系统采用HDFS存储技术,并对电力数据的远程备份和恢复技术进行了分析说明。然后对多目标粒子群算法的数学模型和实现步骤进行了总结,基于多目标粒子群算法,按照电力调度数据的备份成本和数据恢复时间最小的原则,构建了电力调度数据容灾技术的架构并给出了相应的数学模型。针对提出的多目标粒子群的电力调度数据容灾技术,采用了实验仿真,仿真结果与RTO-Cost方法做对比实验。实现结果显示,所提的方法综合考虑了成本和恢复时间最小的原则,在成本处理上性能优于RTO-Cost方法;在恢复时间上,当容量超出中心节点的容量时,恢复时间明显减少了很多,验证了所提方法的高效性和适用性。In order to ensure the safety of power data when power failure occurs,data tolerance technology based on multi-target particle swarm algorithm is proposed.Firstly,the architecture of data tolerance technology is analyzed and the HDFS storage technology is analyzed.Because of its high fault tolerance,the power scheduling system adopts HDFS storage technology and analyzes the remote backup and recovery technology of power data.Then the mathematical model and implementation steps of multi-target particle swarm algorithm are summarized.Based on the multi-target particle swarm algorithm,According to the principle of the minimum backup cost and data recovery time of power dispatching data,the architecture of power dispatching data tolerance technology is constructed and the corresponding mathematical model is given.The power scheduling data capacity of multi-target particle swarm is presented in this paper,and the simulation results are compared with the RTO-Cost method.The implementation results show that the proposed method considering the Cost and time to recover the smallest principle,on the cost of processing performance is superior to the RTO-Cost method,the recovery time,when the capacity beyond the capacity of center node,recovery time significantly reduced a lot,to verify the efficiency and applicability of the proposed method in this paper.

关 键 词：粒子群 电力数据 容灾 备份 

分 类 号：TP309.3[自动化与计算机技术—计算机系统结构] TN91[自动化与计算机技术—计算机科学与技术]