Abstract:In this paper, an adaptive redundant lifting wavelet denoising method is proposed to realize effective identification of weak fault information of roller bearing in mechanical equipments. According to the different features contained in scale coefficients that are to be decomposed, wavelets which can optimally match these features are adaptively selected based on the norm criteria. Meanwhile, algorithm is introduced here to guarantee sufficient information in each scale coefficients and wavelet coefficients after every decomposition. Then the variable scale threshold denoising algorithm is applied to process each wavelet coefficients, followed by reconstruction and envelope spectrum analysis to abstract fault features of roller bearing. Analytical results obtained by adopting the method presented above of both compound fault signal of experiment table and field practical signal show that, fault identification can be well realized and the method discussed in this paper is of effectiveness. Key words: Adaptive; Redundant lifting wavelet; Denoising; Bearing ; Fault identification