Exploiting principal curves for power quality monitoring

This work exploits principal curves for power quality monitoring. The goal of the paper is to investigate how principal curves can be applied to the power quality monitoring problem in terms of event analysis, detection and classification. It is shown that this technique is an efficient tool for feature extraction, for power quality disturbance representation and, therefore, analysis. Successful applications on disturbance detection and classification using simulated and experimental data are presented.

► The paper exploits principal curves (PC) for power quality monitoring. ► PC shows to be efficient for feature extraction, event representation and analysis. ► Successful applications on disturbance detection and classification are presented. ► PC-based systems for power quality monitoring require reduced computational cost. ► The feature space of PC provides a good and compact representation for events.