基于主动学习的电力CPS虚假数据分层过滤方法  

Hierarchical filtering method of power CPS false data based on active learning

作　　者：常荣 李松霖 CHANG Rong;LI Songlin(Yuxi Power Supply Bureau of Yunnan Power Grid Co.,Ltd.,Yuxi 653100,China)

机构地区：[1]云南电网有限责任公司玉溪供电局,云南玉溪653100

出　　处：《电子设计工程》2025年第4期133-136,141,共5页Electronic Design Engineering

基　　金：云南电网有限责任公司科技项目(2021-YN-017)。

摘　　要：CPS虚假数据超量累积会造成电力网络响应速率降低,为避免上述情况的发生,实现对虚假数据对象的快速处理,设计基于主动学习的电力CPS虚假数据分层过滤方法。完善主动学习算法流程,并制定电力CPS虚假数据提取方案。定义电网CPS虚假数据分配原则,根据分层状态估算条件确定过滤参数取值范围,完成基于主动学习的电力CPS虚假数据分层过滤方法的设计。实验结果表明,采用该方法可将CPS虚假数据单位累积量控制在6.2×10^(13)Mb之下,解决了由虚假数据超量累积造成的电力网络响应速率过慢的问题。Accumulation of false data in CPS can lead to decrease response rates in power networks.To avoid such situations and achieve fast processing of false data objects,a hierarchical filtering method for power CPS false data based on active learning is designed.The algorithm flow of active learning is improved,and a power CPS false data extraction plan is formulated.The allocation principles for false data in the power grid CPS are defined,and the parameter value range for filtering is determined based on the estimated conditions of the hierarchical state.The design of the hierarchical filtering method for power CPS false data based on active learning is completed.Experimental results show that using this method can control the cumulative amount of CPS false data to below 6.2×10^(13) Mb,effectively solving the problem of slow response rates caused by excessive accumulation of false data.

关 键 词：主动学习 电力CPS虚假数据 分层过滤 分层状态 数据超量 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]