Weld quality assessment based on arc sensing for robotic welding

Abstract

Robotic Metal Insert Gas (MIG) and Tungsten Insert Gas (TIG) welding are responses for growing need from industry for welding process automation. With the well-established process knowledge, a weld monitoring system is developed in this study for manual or robotic TIG/MIG welding to allow the detection of process inconsistency and to assess the welding quality in real-time. A new method of welding process signals analysis in time and frequency domains is proposed to enable an easy feedback to the robot controller in real-time. The paper includes the experimental results and statistical analysis. In addition, this paper includes non-destructive testing (NDT) results allowing to correlate the results from the proposed method and the selected weld defects.