COMPRESSION AND RECONSTRUCTION FOR FAULTS SIGNALS OF ELECTRIC MACHINES BASED ON WAVELET PACKETS

A new method for compression and reconstruction of faults signals of electric machines based on the improved fast algorithms of wavelet packets (WP) are presented in this paper.The error limitation is selected , the best compression rate can be obtained by selecting decomposition orders and adjusting the threshold of every node. The compression indices changed with decomposition orders are analyzed under the different methods that the coefficients of scale space ( v 1 ) can be compressed and not compressed , and the whole decomposition and optimal decomposition based on “entropy”.The evident results show that the computing loads of the WP can be reduced greatly , and the effective compression and reconstruction without distortion can be obtained using the method proposed in this paper.