An Improved Algorithm of Wavelet Denoising and Its Application In Bearing Fault Diagnosis

摘要

This paper combined advantages of traditional Donoho threshold denoising algorithm and modulus maxima reconstruction denoising algorithm, proposed an improved wavelet denoising algorithm, which not only can filter white noise, but also can filter the residual color noise in low frequency band at the same time, so that it can be precisely' denoised. Decomposing signal by wavelet transform, using the Donoho threshold alogrithm principle to denoise, based on the decomposed signal, according to the. .modulus maxima reconstruction algorithm principle, improve the selection of preset signal power value, then denoising and reconstruct signal. Through the algorithm improvemet, it can be accurately to signal denoising and can overcome the defect of large calculation amount by using modulus maxima algorithm denoising. Through analysis and process the simulation signals and the experimental acquisition vibration signals, the results shows that the improved algorithm denoising effect is better than traditional denoising algorithm.