Classification and recognition of voltage sags based on KFCM — SVM

Voltage sag is a serious power quality problem which has a profound effect on the electrical equipment and the users. Reliable data platform has been provided for real-time monitoring and scientific management of voltage sags by construction and development of on-line monitoring system. The valuable information extraction from massive data is an important problem that needs to be solved urgently. Sags classification and recognition by data mining are the effective means. The KFCM — SVM method proposed in this paper has the advantages as following: Firstly, reasonable classification and optimization of historical data can be realized by KFCM. Secondly, effective recognition of the voltage sag events is executed by SVM. Thirdly, the typical features are selected with high separability. The model, which is more suitable for online systems, is simple with small calculation. The effectiveness of this proposed method is verified by historical data modeling.