Numerical modeling of defect formation in friction stir welding

This study presents an investigation of welding defect formation in joining AZ61 magnesium alloy with friction stir welding by means of a 3D finite element model. The numerical model was created and analyzed in the ABAQUS software. This numerical model uses coupled Eulerian Lagrangian formulation, modified Coulomb's friction law, Johnson-Cook material law, mass scaling technique, and temperature dependent friction coefficient values. The numerical model was validated with experimental results in terms of heat input, temperature dis-tribution, plastic deformation type-amount, and weld defect formation in the weld zone. In friction stir welding, the heat input with a certain tool simply changes directly proportional to the ratio of tool rotation speed to tool feed rate for constant tool compressive force. In the numerical model, as in the experimental study, when the ratio of tool rotational speed to tool feed rate is 2, low heat input results in insufficient plastic deformation, leading to the formation of weld defects in the form of void. When the ratio of tool rotational speed to tool feed rate is 3, however, it was observed that sufficient heat input is provided, and no welding defects occur. The generated numerical model enables the determination of the welding defects that occur in the form of voids in the friction stir welding in a very short processing time. Besides, weld seam geometry can be predicted quite accurately with this model.

Automated summary

This study investigates the formation of welding defects during the joining of AZ61 magnesium alloy using friction stir welding. A 3D finite element model is created and analyzed, and is validated with experimental results. The study finds that insufficient heat input leads to weld defects in the form of void when the ratio of tool rotational speed to tool feed rate is 2, while sufficient heat input prevents welding defects when the ratio is 3. The numerical model allows for the determination of welding defects and accurate prediction of weld seam geometry in a short processing time.