A denoising method with exponential decay threshold for square arc signals in submerged arc welding

A novel denoising method with exponential decay threshold is proposed in this paper,by using this method the original signals can be well recovered and the clear arc current and voltage waveforms are obtained,the high frequency components of electric signals are reserved perfectly,and the rising or falling edge of current or voltage waveforms crossing the zero point are evaluated precisely.Comparing with the traditional wavelet denoising method,this method has the characteristics of little error,good continuity,and random rank derivative at the threshold point.The experiment results show that the majority of noise in the electric signals is eliminated while the distortion-free status of transient components of the electric signals is maintained perfectly,the loss of high frequency detail components is reduced and the effect of the local character extraction from welding arc signals is improved.