Digital signal processing techniques for improving the automatic classification of power quality events.

Abstract

The work presented in this thesis investigates the application of digital signal processing techniques in the power quality automatic classification field, and thus, proposes an optimized automatic monitoring system with an improved accuracy. The proposed monitoring system involves three main sections: detection of the power quality events, extraction of the distinctive features that characterise each event, and automatic classification of the similar events under pre-defined categories. The thesis proposes new power quality processing techniques for detection and feature extraction sections, including the Hilbert and Clarke transforms. The former proposed technique was used for analysing single phase signals, while the later technique was proposed for the simultaneous analysis of three signals.