Performance Tests of a Support Vector Machine used for Classification of Voltage Disturbances
Paper i proceeding, 2006

This paper proposes a novel method for classifying voltage disturbances in electric power systems by using the Support Vector Machine (SVM) method. The proposed SVM classifier is designed to classify five common types of voltage disturbances and experiments have been conducted on recorded disturbances with good classification results. The proposed SVM classifier is also shown to be robust in terms of using training data and testing data that originate from two different power networks.

power quality

statistical learning theory

support vector machine

event classification

Power distribution

Författare

Peter G.V. Axelberg

Chalmers, Signaler och system, Signalbehandling och medicinsk teknik, Signalbehandling

Irene Yu-Hua Gu

Chalmers, Signaler och system, Signalbehandling och medicinsk teknik, Signalbehandling

Math H.J. Bollen

in proc. of 12th International conf. on Harmonics and Quality of Power (ICHQP 2006), Cascais, Portugal, Oct.1-5, 2006

Ämneskategorier

Signalbehandling

Annan elektroteknik och elektronik