基于粒子滤波的焊接电流信号降噪处理

Welding Current Signals Interference Reduction Processing Based on Particle Filter

针对旋转电弧传感器获得的焊接电流信号不仅包含有焊缝坡口信息,同时受到了许多非线性非高斯噪声干扰,如在焊接过程中熔滴飞溅、短路过渡、焊接电源不稳定、送丝速度不均匀等干扰,提出了基于粒子滤波的焊接信号降噪方法。首先建立了不含噪的焊接信号的数学模型,将其作为粒子滤波的状态方程;然后提取背景噪声,将其和状态信号一起作为观测信号即观测方程,据此对原始真实信号进行估计,得到降噪后的信号,把降噪问题转化成在状态空间模型下的滤波问题,把粒子滤波降噪思想应用在焊接电流信号的降噪处理,获得了比较理想的效果。

The welding noise reduction method based on particle filter is proposed in view of the welding welding current signals obtained by rotating arc sensor contains not only the weld groove information, but also many nonlinear non-gaussian noise which includes melting point, short circuit transition, unstable welding power and wire feeding speed in the process of welding spatter..First mathematics model with no noise of welding signal is established, and make it as the state equation ofparticle filter. Extract the background noise and put it with state signal together as observation signal （observation equation）. Estimate the original real signal and get the signal by noise reduction, and make noise reducing problem into filter problem of state space model. The ideas of Particle filtering noise reduction is used in the welding current signal noise reduction processing, and obtained the quite ideal effect.