基于KPCA-SVM的同期线损数据治理研究

Research on the Data Management of the Time Line Loss Based on KPCA-SVM

针对智能电网中同期线损管理可以查询异常数据,从而找到异常挂载设备的问题,论文提出了基于核主成分分析(Kernel Principal Component Analysis,KPCA)关键因素提取的支持向量机(Support Vector Machine,SVM)电网线损预测模型,通过KPCA提取影响线损的特征向量,然后建立SVM线损预测模型,然后将预测线损电量与实际线损相比较,若差值大于设定阈值,则采用TF-IDF(term frequency–inverse document frequency)算法筛选异常挂载电力设备。最后,通过实验验证了KPCA-SVM的有效性以及TF-IDF方法筛选异常设备的准确性。

The problem of abnormal mounting equipment can be found by querying abnormal data for the management of the same time in smart grid.In this paper,the model of Support Vector Machine(SVM)grid line loss prediction based on Kernel Princi?pal Component Analysis(KPCA)is proposed.The characteristic vector of the affected line loss is extracted by KPCA and then the SVM line loss prediction model is established.Then,the forecast line loss is compared with the actual line loss.If the difference val?ue is greater than the set threshold,the TF-IDF(term frequency-inverse document frequency)algorithm is used to filter the abnor?mal mounted power equipment.Finally,the validity of KPCA-SVM and the accuracy of TF-IDF method are verified by experi?ments.