基于量测量突变检测与拓扑约束协同的不良数据检测
|
摘要
电力系统状态估计过程中,常常由于残差污染和残差淹没等问题使不良数据点变得模糊,导致不良数据辨识难度增大.充分利用量测量突变检测方法,将超过某一门槛值的数据列为可疑数据;引入了电气量在网络拓扑间的约束关系,对已检出的可疑数据进行检测.根据电力网络的拓扑约束特性来确定量测量中的不良数据,以避免系统中存在多个不良数据时的漏检和误检.以某4节点电力系统为例,通过与加权残差法和标准残差法检测结果的对比,验证了基于量测量突变检测与电力网络拓扑约束协同的不良数据检测方法的有效性和可行性.
In the process of power system state estimation,the defective data points become blurred due to residual pollution and residual submergence,which makes the identification of bad data more difficult. Aiming at this problem,the correlation measure is used to detect the data which can be detected beyond the certain threshold value,and introduce the relationship between the amount of electricity in the network topology to detect the suspicious data detected. Through the topology constraint of the power network Characteristics to determine the poor measurement data in the measurement,the system further reduces the number of bad numbers when the number of missed and false detection of the possibility. With a four node power system as an example to carry out the analysis,with weighted residuals and standard residuals method,the test results showthat the proposed method is effective and feasible for the identification of multiple bad data based on the combination of the abrupt change detection and the power grid correlation matrix.
In the process of power system state estimation,the defective data points become blurred due to residual pollution and residual submergence,which makes the identification of bad data more difficult. Aiming at this problem,the correlation measure is used to detect the data which can be detected beyond the certain threshold value,and introduce the relationship between the amount of electricity in the network topology to detect the suspicious data detected. Through the topology constraint of the power network Characteristics to determine the poor measurement data in the measurement,the system further reduces the number of bad numbers when the number of missed and false detection of the possibility. With a four node power system as an example to carry out the analysis,with weighted residuals and standard residuals method,the test results showthat the proposed method is effective and feasible for the identification of multiple bad data based on the combination of the abrupt change detection and the power grid correlation matrix.
引文
[1]于尔铿.电力系统状态估计[M].北京:水利电力出版社,1985:115-120.
[2]孟建良,刘德超.一种基于Spark和聚类分析的辨识电力系统不良数据新方法[J].继电器,2016(3):85-91.
[3]许勇.基于PMU/SCADA混合量测状态估计及不良数据检测方法[J].四川电力技术,2015(4):51-55.
[4]张艳军,刘海珊,周苏荃.新息图拓扑可观测性及不良数据可辨识性分析[J].电力系统自动化,2008(6):55-59.
[5]张艳军,葛延峰,高凯,等.基于广域测量系统的新息图辨识不良数据方法[J].电力系统自动化,2010(23):25-29.
[6]张兴民,毛玉华,朱剑峰,等.利用图论方法进行多不良数据检测与辨识[J].中国电机工程学报,1997(1):69-72.
[7]卢志刚,程慧琳,冯磊,等.基于证据融合理论的多不良数据辨识[J].电网技术,2012(1):123-128.
[8]周小宝,罗滇生,毛田,等.一种实用的高压配电网不良数据处理方法[J].计算机工程与应用,2014(8):255-259.
[9]董树锋,何广宇,梅生伟.电力系统状态估计中不良数据识别的新方法[C]//Intelligent Information Technology Application Association.Proceedings of the 2011 International Conference on Softw are Engineering and M ultimedia Communication(SEM C 2011 V3),2011:4.
[10]邱关源.电路[M].5版.北京:高等教育出版社,2006:99-102.
[2]孟建良,刘德超.一种基于Spark和聚类分析的辨识电力系统不良数据新方法[J].继电器,2016(3):85-91.
[3]许勇.基于PMU/SCADA混合量测状态估计及不良数据检测方法[J].四川电力技术,2015(4):51-55.
[4]张艳军,刘海珊,周苏荃.新息图拓扑可观测性及不良数据可辨识性分析[J].电力系统自动化,2008(6):55-59.
[5]张艳军,葛延峰,高凯,等.基于广域测量系统的新息图辨识不良数据方法[J].电力系统自动化,2010(23):25-29.
[6]张兴民,毛玉华,朱剑峰,等.利用图论方法进行多不良数据检测与辨识[J].中国电机工程学报,1997(1):69-72.
[7]卢志刚,程慧琳,冯磊,等.基于证据融合理论的多不良数据辨识[J].电网技术,2012(1):123-128.
[8]周小宝,罗滇生,毛田,等.一种实用的高压配电网不良数据处理方法[J].计算机工程与应用,2014(8):255-259.
[9]董树锋,何广宇,梅生伟.电力系统状态估计中不良数据识别的新方法[C]//Intelligent Information Technology Application Association.Proceedings of the 2011 International Conference on Softw are Engineering and M ultimedia Communication(SEM C 2011 V3),2011:4.
[10]邱关源.电路[M].5版.北京:高等教育出版社,2006:99-102.