De-noising in Electric Signals of Arc Welding Process via Wavelet Soft threshold

During the practical measuring of the electric signals of CO₂ welding, the noise always can't be avoided. De-noising is a key link in analyzing signals. Wavelets have good time-frequency characteristics. The signals can be decomposed into different frequency components with different wavelet scales.For continuous signals, wavelet transform coefficient will increase in direct ratio with the scale. For noise, the coefficient will decrease in inverse ratio with the scale. On the basis of the principle, noise may been removed from original signal. The de-noising via wavelet soft-threshold is emphatically analyzed in this paper. The de-noising results of three kinds of methods for the practically measured electric signals of arc welding process are given out. De-noising by the traditional low pass filter and wavelet crude filter, the signals will show serious distortion in the breaking position. De-noising via wavelet soft-threshold can keep the break position of signals out of distortion as well as eliminate the noise in the signals. So this method can improve the effect of extracting the characteristic information from the signals.