磁控电弧焊缝跟踪传感信号的实时小波降噪方法

通过对磁控电弧传感信号的频谱分析,发现表征焊缝位置信息的特征信号与干扰信号处在不同的频段；针对焊缝跟踪系统的实时性要求,提出了一种基于滑动固定数据窗的电弧传感信号实时小波降噪方法.该方法动态地截取电弧传感信号的实时数据,并通过选用正交紧支撑小波与舍弃电弧传感信号的第一、二层细节信号的小波分解方法,大大降低了传感信号降噪算法的计算量,保证了算法的实时性；最后采用DSP技术,对实际的焊接电流信号按该方法进行了大量的在线实时降噪试验.结果表明,采用该方法对磁控电弧传感信号进行降噪处理,能够满足系统的去噪要求及实时性要求.

A real-time wavelet de-noising method of magnetism-control arc seam tracking signal

According to the frequency spectrum of magnetism- control transducing signal,it is found out that the frequency band of signal characterized the position of the seam is different from the interference signal.With regard to the realtime requirement of seam tracking system,a real-time wavelet denoising method was put forward based on sliding the fixed data window.In this way,a real-time data of arc transducing signal is taken dynamically.Through using decomposition method of orthogonal compactly supported wavelet and the 1^st and 2^nd layer''s detail signal of the discarded arc tranducing signal,the amount of calculation is reduced,therefore the real-time of this method is guaranteed.Then DSP is used in the experiment on the actual welding current signal to check the denoising effect.It is shown that this method used in de-noising satisfies the requirements of de-noising and real-time.