Electric signals filtering of AC CMT welding based on wavelet analysis

An improved threshold de-noise algorithm based on the general de-noise threshold algorithm was proposed. A processing function and an adjustment coefficient were added to fill the gap within the threshold range. It is aimed at achieving the flexibility to adjust the degree of attenuation and smoothing and improving the continuity of the reconstructed signal and the degree of approximation of signals. Coif wavelet was used to decompose the electrical signal of stainless steel welding by AC CMT for 3-tier,then the decomposed wavelet coefficients were processed by improved threshold de-noise algorithm. The results showed that by adjusting the adjustment coefficient flexibly, high-frequency noise can be removed well as the useful mutation component and the rule pulsation signals were held through the processing of improved algorithm. The attenuation of the useful signals was lower than that the soft threshold method performed and improved the approximation to the real signals. The improved threshold de-noise algorithm is suitable for filtering of electrical signals of AC CMT.