Proactive control of weld dimensions in robotised MAG welding

Other conference contribution, 2013

Keeping all welding dimension on target and within specification limits is a multiparameter optimisation problem that involves trade-offs between potentially contradicting requirements. In the fillet weld case study, the balancing act is to keep fillet weld bead thickness on target and simultaneously fulfil a minimum requirement of both weld toe radii and weld penetration of the base material. With a robustengineering approach, consisting of screening, modelling, multi-parameter optimisation and robust assessment, some interesting observations regarding the relationships between welding control parameters and resulting weld dimensions have been made. The parameters that directly influence welding productivity, such as beam power (U, I), speed and material feed, were found to be of secondary importance for controlling the toe radius, penetration dimensions and the weld bead thickness, compared to the influence of the welding geometry (torch and plate positions). With the multi-parameter model developed, it was possible to predict and verify a setting of the welding geometry that fulfilled all weld dimension specifications simultaneously and to estimate the width of a production window that allows for some control factor variability. This result opens up a new strategy to the development of both welding productivity and quality at the same time, even though it takes some offline response surface modelling (RSM) or on-line evolutionary process (EVOP) experimental procedures, at least for this particular welding scenario: a 5mm fillet welds with at least 2mm penetration and smooth (>1mm) weld toe transition.