基于数据特征的配电网同期线损异常辨识方法

Identification Method of Synchronous Line Loss Anomaly in Distribution Network Based on Data Characteristics

当前配电网同期线损异常辨识方法存在辨识结果不精准的问题,提出了基于数据特征的配电网同期线损异常辨识方法。确定配电网同期线损数据正“秩和”与负“秩和”近似相等特征关系,分析配电网同期线损异常情况,动态管控配电线路。构造密度指峰函数,避免出现相距较近的聚类中心,聚类数据特征向量。计算待辨识向量与实际向量之差,构建支线回路关联矩阵,判断数据是否为异常数据,实现配电网同期线损异常辨识。实验结果表明,所提方法的异常数据变化曲线与实际异常数据变化曲线一致,且在0~8、0~9支线上,只存在1个异常数据辨识误差,具有精准的辨识结果。

At present,the identification results of synchronous line loss anomaly identification methods in distribution network are inaccurate.A synchronous line loss anomaly identification method based on data characteristics is proposed.It determines the approximate equal characteristic relationship between positive“rank sum”and negative“rank sum”of distribution network synchronous line loss data,analyzes the abnormal situation of distribution network synchronous line loss,and dynamically controls the distribution lines.It constructs density index peak function to avoid clustering centers close to each other and cluster data feature vectors.It calculates the difference between the vector to be identified and the actual vector,constructs the branch circuit incidence matrix,judges whether the data is abnormal data,and realizes the abnormal identification of line loss in the same period of distribution network.The experimental results show that the abnormal data change curve of the proposed method is consistent with the actual abnormal data change curve,and there is only one abnormal data identification error on the branch lines 0-8 and 0-9,which has accurate identification results.