Research And Application Of The Redundent Lifting Wavelet Analysis In Fault Diagnosis

Wavelet analysis has been an area of concern for the past two decades; it has alsobecome one of important tools in the field of harmonic and signal analysis. The liftingwavelet as a unique algorithm structure, the fast computing power and low storagerequirements is suitable for adaptive, nonlinear, non oddity sampling and integer tointeger transform. This sampling will transform current methods and be advantageousin the field of information science. The paper has made some research of mechanicalfault around the lifting wavelet analysis algorithm.This paper firstly analyzes some basic theory of lifting wavelet such asdecomposition and reconstruction process of lifting wavelet analysis. Then throughthe analysis of lifting wavelet decomposition as each compartment after the secondsample so that each amount of data through a decomposition is reduced by half,therefore the less involved in terms of information analysis will inevitably lead todifferent degrees of distortion effects. The introduction of autoregressive spectralanalysis and the method of lifting wavelet signals, autoregressive spectral analysis ofshort data can be better than the results of Fourier transformation.In order to avoid lifting wavelet analysis of number is reduced by half after eachsampled by twice, the paper proposes a redundant lifting wavelet analysis algorithm.Redundant lifting wavelet analysis does not exist across two sampling steps, eachdecomposed approximate signal and detail signal and the amounts of data containedin the original signal such as length, data redundancy, and facilitate analysis of signalcharacteristics.Be conducive to match the characteristics of the signal and the extraction of weaksignal characteristics, this paper puts forward a new redundant lifting waveletanalysis based on adaptive algorithm by study the relevant literatures. In thetransformation of adaptive redundant lifting wavelet, adaptive selection of updater andpredictor to ensure that propose signals of more effective features according toanalysis the local information of signals.To verify the practicability of adaptive redundant lifting wavelet analysis, thispaper gives the application of algorithm in the analysis of vibration signals and thenextracts the early signal of weak fault. The results show that adaptive redundant liftingwavelet analysis can meet the requirements of extraction in early weak signal.