Human behaviour capturing in manual tungsten inert gas welding for intelligent automation

Abstract

Tungsten inert gas welding is extensively used in aerospace applications due to its unique ability to produce higher quality welds compared to other conventional arc welding processes. However, most tungsten inert gas welding is performed manually, and it has not achieved the required level of automation. This is mostly attributed to the lack of process knowledge and adaptability to complexities, such as mismatches due to part fit-up and thermal deformations associated with the tungsten inert gas welding process. This article presents a novel study on quantifying manual tungsten inert gas welding, which will ultimately help intelligent automation of tungsten inert gas welding. Through tungsten inert gas welding experimentation, the study identifies the key process variables, critical tasks and strategies adapted by manual welders. Controllability of welding process parameters and human actions in challenging welding situations were studied both qualitatively and quantitatively. Results show that welders with better process awareness can successfully adapt to variations in the geometry and the tungsten inert gas welding process variables. Critical decisions taken to achieve such adaptations are mostly based on visual observation of the weld pool. Results also reveal that skilled welders prioritise a small number of process parameters to simplify the dynamic nature of tungsten inert gas welding process so that part variation can be accommodated.