Design of mixture de-noising for detecting faulty bearing signals

Detecting the waveform of a noisy signal is a key problem in the detection of early bearing faults under actual plant conditions. Mixture de-noising is found to be a useful technique for identifying bearing signals and greatly improves the fault diagnostics of the bearings. The mixture de-noising technique consists of an adaptive noise-canceling filter and a wavelet-based de-noise estimator. The mixture de-noising technique can substantially improve the signal-to-noise ratio when the signal is contaminated by noise. The performance of mixture de-noising under different noise ratios, bearing failure sizes and shaft speeds are discussed in this paper. This paper shows that the diagnostic role of failure diagnosis and analysis techniques can be made more effective with the application of mixture de-noising.