E R Watts, D G Staines, W M Thomas, and E D Nicholas
A dual-rotation FSW variant is being investigated at TWI, whereby, the probe and shoulder rotate separately. The dual-rotation friction stir welding variant provides for a differential in speed and/or direction between the independently rotating probe and the rotating surrounding shoulder as shown in fig 1.
The apparatus used for this investigation is shown sequentially in fig 2a & b.
Fig.2. Dual-rotation apparatus complete with white marks on the shoulder and probe to indicate relative rotational movement
The apparatus can enable a range of different rotational speeds to be pre-selected or varied automatically by in-process control to suit the desired welding conditions.
In conventional rotary FSW, the relative velocity of the tool increases from zero at the probe centre to maximum velocity at the outer diameter of the shoulder. The dual-rotation technique can significantly modify the velocity gradient between the probe centre and the shoulder diameter. This technique provides a differential in rotation speed and the option for rotation in opposite directions. For example the shoulder rotational speed can be varied from approximately 30% less than the probe rotational speed while rotating in the same direction or about 25% less than the probe rotational speed when the shoulder is rotated in the opposite direction.
This dual-rotation technique effectively allows for a high probe rotational speed with out a corresponding increase in shoulder peripheral velocity. This technique can provide for a more idealised rotational speed for both probe and shoulder.
Dependent on the material and process conditions used, over heating or melting along the 'near shoulder side' weld surface of certain friction stir welds can occur. Melting can lead to fusion related defects along the 'near shoulder side' weld surface. The dual-rotation technique can be used to reduce the shoulder rotational speed as appropriate and help reduce any tendency towards over heating or melting.
A double sided butt weld using non-optimised conditions was made to demonstrate that dual-rotation stir welding is practicable for certain applications. Figure 3 shows the macrostructural features provided by dual-rotation stir welding using a Flared-triflute TM type probe.
Guide bend testing demonstrated freedom from gross defects as shown in figure 4.
Figure 5 shows the appearance of the weld surface that is formed beneath the tool shoulder after dual-rotation stir welding.
Although further investigation needs to be carried out, the dual-rotation technique may prove to have the following potential advantages over conventional FSW.
- It may provide more idealised process conditions to suit dissimilar clad plate. Whereby the shoulder rotational speed is selected to suit the clad material and the probe rotational speed is selected to suit the substrate material.
- It should enable a reduction in the reactive torque necessary to clamp plates together when opposed rotation is used. This may also be useful for robotic applications. (In this regard it would provide a similar benefit to Com-stir TM ).
- It is capable of minimising any tendency towards over heating or melting associated with the 'shoulder near side' weld surface.
- It is speculated that the facility to reduce the shoulder rotational speed may help minimise preferential corrosion particularly associated with some friction stir welded 7000 series materials.
- It is believed that it could benefit material processing and subsequent plate bending where outer surface condition is particularly important.
Discussion and concluding remarks
The initial investigation of dual rotation stir welding has demonstrated the feasibility of the technique for butt welding 16 mm thick 5083-H111 condition aluminium alloy but further, work is required to fully characterise the performance and acceptability of the dual stir technique. Work will continue at TWI to investigate the use of dual-rotation on spot, butt, and lap welds. In addition trials will be undertaken to achieve improvements in traverse rate, and investigate tool tilt angle.
Further work will also be undertaken to study the use of the contra-rotation variant see fig 6.
The results from the dual-rotation show promise, but much more work will be required to develop and perfect the technique.
Acknowledgements are made for the support and contributions provided by C S Wiesner, I M Norris, P J Oakley, R E Dolby (retired) Roger Lilly (retired).