Title :
Multi-resolution analysis based data compression for power transformer protection
Author :
Aktaibi, A. ; Rahman, Md Arifur
Author_Institution :
Fac. of Eng. & Appl. Sci., Memorial Univ. of Newfoundland, St. John´s, NL, Canada
Abstract :
In this paper, a multi-resolution hybrid technique based data compression for protection of power transformer is presented. This hybrid technique is established by combining the d-q axis transformation with the wavelet packet transform (WPT) to provide the proposed (dqWPT) hybrid protection technique for power transformer. In this technique, signal decomposition and reconstruction are carried out for the collected fault and inrush current signals for different types of mother wavelets at a certain number of levels. The number of levels has to be kept the minimum for optimal use. Minimum description length (MDL) method is used to select the best mother wavelet, and the optimal number of decomposition levels required for the power transformer protection using the proposed technique. For this analysis, the Daubechies db4 is selected at one level of resolution is utilized to implement the proposed technique. The proposed technique has shown good results with the selected mother wavelet for different disturbances at the power transformer.
Keywords :
data compression; fault diagnosis; power transformer protection; signal reconstruction; wavelet transforms; Daubechies db4; MDL method; d-q axis transformation; dqWPT hybrid protection technique; fault current signals; inrush current signals; minimum description length method; multiresolution hybrid technique based data compression; power transformer protection; signal decomposition; signal reconstruction; wavelet packet transform; Circuit faults; Current transformers; Power transformers; Surges; Transient analysis; Wavelet analysis; Wavelet transforms; d-q axis components; differential protection; minimum description length (MDL); multiresolution analyses; three-phase power transformers; wavelet packet transform (WPT);
Conference_Titel :
Electrical Machines (ICEM), 2014 International Conference on
Conference_Location :
Berlin
DOI :
10.1109/ICELMACH.2014.6960519