Metal fusion using pulsed GasMetal Arc : Melt pool modellingand CFD simulation

Abstract: Pulsed gas metal arc is a highly efficient technique used in manufacturing processes like welding and additive manufacturing. It offers high productivity and cost benefits but it is also prone to defect formation when process parameters are not properly controlled and optimized. A deeper process understanding can support achieving improved process control and mitigate these potential drawbacks. Nevertheless, there are still several challenges. For instance, the correlation between the input and output process parameters is non-linear and complex due to the multi-physics nature of the process. In addition, the elevated temperature and the intense radiative emission from the arc, along with the smoke, and the non-transparency of metal, make in-process observation challenging. Modelling and simulation offer a complementary approach to gain a deeper process understanding. In this study, a thermo- and fluid dynamics model was developed, focusing on the melt pool and metal deposition, while simplifying the arc to boundary conditions (decoupled approach). This model incorporates various forces and phenomena such as thermocapillary and electromagnetic forces, melting and solidification, and tracking of surface deformation and droplet coalescence.In the first part of the thesis, the developed model was applied to investigate the effect of workpiece orientations on the melt pool dynamics and reinforced bead geometry in multi-layer gas metal arc welding of a V-groove joint. The comparison of the predicted fusion zone with macrographs obtained from the experiments showed good qualitative agreement. It was found that the force balance in the melt pool changes significantly when changing the workpiece orientation by as little as 20◦ relative to the flat position. This results in distinct melt flow patterns, melt pool shapes, bead geometries, and in some cases, defect formation such as humping, undercut, and insufficient fusion. It was concluded that to avoid these defects a lower angle range is necessary for multilayer welding with the uphill orientation and side inclination.The second part of the thesis focused on analyzing different variants of the model for the electromagnetic force with a decoupled approach. Three commonly used models were compared: (1) the analytical models proposed by Kou and Sun inintegral form, (2) by Tsao and Wu in algebraic form, and (3) the partial differential equations governing the electric and magnetic fields. The comparative investigation was supported by experimental tests that also provided estimates of unknown model parameters and validation data. It was found that the distinct assumptions on which these models rely are not all justified. They resulted inpredicting different melt flow patterns and amplitude of the free surface oscillations, as well as different melt pool shapes and bead geometries. Model (3) is recommended to advance to a predictive melt pool model and was subsequentlyused in the remaining work of the thesis.Furthermore, the literature shows that modeling the effect of pulsed arc on the melt pool using a decoupled approach involves various simplifications. Arc pulsation affects energy and force balance in the melt pool through arc heat flux, arc pressure, and electromagnetic force. A systematic investigation of model variants considering pulsing was conducted using previously documented experimental test cases. The results showed that the influence of arc pressure was insignificant in those cases. However, model variants simplifying arc pulsing to a time-averaged effect underestimated the amplitude of the Marangoni flow and downward flow compared to a more comprehensive approach that considered the time dependence of arc pulsation. Thus, it is recommended to use a meltpool model that accounts for the time-dependent arc pulsation, which was also subsequently utilized in the remaining work of the thesis.The electromagnetic force models discussed earlier assume a stationary free surface when computing the electromagnetic force. However, this force is often at leading order in the vicinity of the arc. In the same region, the metal drop transfer leads to a periodic deformation of the melt pool free surface. In the final part of the thesis, the model was extended to account for free surface deformation when computing the electromagnetic force. This extension was applied to experimental test cases, and a comparison was made with simulation results obtained using the stationary electromagnetic force model. Significant differences in the results were observed, particularly in predicting the experimentally observed fingertip-shaped fusion zone geometry. The proposed improvement in the electromagnetic force model provided better predictions in this regard.