Home > Research > Publications & Outputs > Optimal algorithms for wavelet-based denoising ...
View graph of relations

Optimal algorithms for wavelet-based denoising in PD measurement.

Research output: Contribution to conference - Without ISBN/ISSN Conference paper

Publication date08/2003
Number of pages4
<mark>Original language</mark>English
EventIEEE International Symposium on High Voltage Engineering - Delft, Holland
Duration: 23/08/200325/08/2003


ConferenceIEEE International Symposium on High Voltage Engineering
CityDelft, Holland


Recent research showed that the Wavelet-based approach is superior to the methods based in only time-domain or frequency-domain for PD measurement denoising, and that Wavelet Transform(WT) can potentially be used to extract PD pulses from strong noises irrespective of the form of noise present. However, the method is more complex than the Fourier Transform(FT), and it needs prior knowledge and experience to be applied to produce its best effect. The present paper presents the optimal algorithms which can be used to select the best wavelet to be used, the determination of the number of levels that should be analysed in the denoising process, and the algorithm for determination of the threshold values. Results show that with the proposed set of algorithms, automatic use of the WT for PD measurement denoising can be implemented without prior knowledge of the noises present.