基于决策粗糙集模型的电网多源异构数据整合  

Multi⁃source heterogeneous data integration of power grid based on decision rough set model

作　　者：王尧 邵晶晶 宋云奎 WANG Yao;SHAO Jingjing;SONG Yunkui(Data Platform Business Division(Southnet Big Data Center),China Southern Power Grid Digital Power Grid Research Institute Co.,Ltd.,Guangzhou 510000,China)

机构地区：[1]南方电网数字电网研究院有限公司数据平台事业部(南网大数据中心),广东广州510000

出　　处：《电子设计工程》2024年第10期140-144,共5页Electronic Design Engineering

摘　　要：针对传统整合方法受干扰数据的影响,导致配电网多源异构数据整合时误差大,效率低等问题,提出基于决策粗糙集模型的电网多源异构数据整合方法。通过对电网数据属性、存储方式、读写方法等预处理,结合贝叶斯估计方法,剔除含有噪声不确定性数据;通过粗糙隶属函数甄别真值,构建决策粗糙集模型过滤数据,划分区域;按照贝叶斯决策过程,确定数据所在集合,过滤掉冗余数据,以防止误判;构建多源异构数据整合模型,利用人工神经网络的单层次分割算法,将结构化数据与非结构化数据融合,以此完成数据整合。实验结果表明,该方法获取的最佳数据整合点的标幺值为0.73,与实际整合点数据一致,可将全部数据整合到一起,表明所提方法具有良好整合效果。The traditional integration method is affected by the interference data,which leads to the problems of large error and low efficiency in the integration of multi⁃source heterogeneous data of distribution network,a method of power grid multi⁃source heterogeneous data integration based on decision rough set model is proposed.By preprocessing the power grid data attributes,storage methods,reading and writing methods,combined with Bayesian estimation method,the uncertain data containing noise is eliminated;The truth value is discriminated by rough membership function,and the decision rough set model is constructed to filter data and divide regions;Determine the set of data according to the Bayesian decision⁃making process,and filter out redundant data to prevent misjudgment;Build a model of multi⁃source heterogeneous data integration,and use the single level segmentation algorithm of artificial neural network to integrate structured data and unstructured data to complete data integration.The experimental results show that the per unit value of the best data integration point obtained by this method is 0.73,which is consistent with the actual integration point data and can integrate all the data together,indicating that the proposed method has a good integration effect.

关 键 词：决策粗糙集 电网多源 异构数据 整合 

分 类 号：TN102[电子电信—物理电子学]