Subscribe to our newsletter to receive the latest news and events from TWI:

Subscribe >
Skip to content

Recent developments in friction stir welding of thick section aluminium alloys

   
Perrett J G (1) , Martin J (1) , Threadgill P L (2) and Ahmed M M Z (3)

(1) TWI Technology Centre Ltd, Rotherham
(2) TWI Ltd, Cambridge
(3) Department of Engineering Materials, The University of Sheffield

Paper presented at 6th World Congress, Aluminium Two Thousand, Florence, Italy, 13 - 17 March 2007.

Abstract

In the last 15 years, friction stir welding has developed from a laboratory curiosity to an important fabrication technique for many aluminium alloys. Industrial implementation so far has been limited almost exclusively to thin section materials, i.e. up to 20mm thick. Recent trials have shown that welds in excess of 20mm in a single pass can be made, opening up the technology to new applications.

This report describes work undertaken at TWI Technology Centre, the development programme, the process parameters, tooling design and gives an insight into the metallurgical and hardness properties of a series of welds in a range of20mm thick aluminium alloys.

Introduction

Friction stir welding (FSW) was invented and patented [1,2] in 1991 by TWI and has since been developed to a stage where it is being widely applied in production. Currently 163 organisations hold non-exclusive licences to use the process. Most of them are industrial companies and there have been more than 1700 patent applications filed related to FSW.

Applications in the shipbuilding, automotive and aerospace industries are well publicised and documented. [3,4,5] Using FSW for joining thicker section aluminium alloy sections is not as advanced and there is little information available in the public domain. It should be noted that FSW offers exciting possibilities for joining very thick aluminium alloys in a single pass, recent work at TWI Technology Centre includes producing quality welds in sections up to 75mm thick. This has previously only been possible with electron beam welding.

FSW is performed by a rotating non-consumable tool which is plunged into the material to be welded until the shoulder contacts the top surface of the plate; the tool is then traversed along the joint line to create a solid phase weld ( Fig.1). The material around the tool becomes softened and highly plasticised from the frictional heat generated during the process and is carried around the tool so that there is complete mixing of material from the two plates. In full penetration welding the tool probe extends almost through the thickness of the plates to be welded, but in partial penetration the probe can be much shorter. Because there is no melting or resolidification of material,there is usually lower distortion introduced by this welding process compared with other welding processes. The process can be used to make butt and lap joints in any orientation. Due to the high forces developed during the process, the plates must be rigidly clamped during welding. A tool tilt angle is usually used to aid material compaction behind the FSW tool.

Fig.1. Principle of the friction stir welding process invented at TWI
Fig.1. Principle of the friction stir welding process invented at TWI

This report provides base line data on welding parameters and joint characteristics when joining a range of 20mm thick aluminium alloys in both similar combinations and in a dissimilar combination.

Materials and equipment

The aluminium alloy grade, plate size and welding configuration are shown in Table 1 below.

Table 1 List of grades, dimensions and welding configuration for the materials used for this work.

Material GradeDimensionWeld Configuration
AA2017A-T451 20mm x 125mm x 500mm Butt-weld
AA5083-O 20mm x 125mm x 500mm Butt-weld
AA6082-T651 20mm x 125mm x 500mm Butt-weld
AA7075-T651 20mm x 125mm x 500mm Butt-weld
AA7020-T651 20mm x 125mm x 500mm Butt-weld
AA7020-T651 20mm x 125mm x 300mm Dissimilar butt-weld
AA7075-T651 20mm x 125mm x 300mm Dissimilar butt-weld

A common Triflat TM design FSW tool ( Fig.2) was used for all of the welds. The tool design was based on current best practice for welding AA7075, as this was believed to be the most challenging material within the trials. The general features of the tool include a 40mm diameter scrolled shoulder with a tapered, threaded Triflat TM 19.5mm long probe.

Fig.2. Illustration of FSW tool showing the scrolled shoulder and Triflat TM probe design used for the welding trials
Fig.2. Illustration of FSW tool showing the scrolled shoulder and Triflat TM probe design used for the welding trials

The machine used for this work was a precision spindle FSW machine ( Fig.3) manufactured for TWI by Transformation Technologies Inc in the USA. This machine was developed for friction stir welding of high temperature materials and incorporates a highly concentric spindle (spindle run-out<10µm). The machine is designed to withstand high forces with very low structural deflection, deflecting less than 0.8mm in the Z-axis at the full downforce of 100kN and deflecting less than 0.3mm in the X-axis at the full traverse force of 45kN. The machine is also very well instrumented to monitor force, torque, speeds and distances.

Fig.3. Photograph of the precision FSW machine that was used to perform all of the 20mm thick welds in this study
Fig.3. Photograph of the precision FSW machine that was used to perform all of the 20mm thick welds in this study

A steel jig with horizontal hydraulic and mechanical vertical clamping was used to hold the material during the welding cycle. A 2mm steel backing shim was placed underneath the plates during welding to prevent direct tool contact with the jig base.

Experimental approach

All of the welding trials in this study are of a butt-weld configuration with a tool tilt angle of 2° normal to the plate surface. The welds were performed in displacement control with the tool plunging into the material to apre-determined depth. A pilot hole 9mm in diameter and 18mm deep was used for all the welds to reduce the forces upon the FSW tool during the plunge cycle. The FSW machine input parameters employed to produce the welds during this study are shown in Table 2. After a brief parameter optimisation study, these parameters were found to produce defect-free welds.

Table 2 Weld parameters for the different alloys

Material gradeRotation speed, rev/minTraverse rate, mm/min
AA2017 300 120
AA5083 300 80
AA6082 500 350
AA7020 200 90
AA7075 200 80
AA7075 - AA7020 200 80

A dissimilar plate butt-weld was produced with the AA7020 and AA7075 materials. The plates were butted together with the AA7075 on the advancing side of the FSW tool and the AA7020 on the retreating side of the FSW tool. Each of the welds was sectioned approximately 15mm distance from the exit hole.

Results

The parameters resulting from the welding conditions are detailed in Table 3 showing the maximum torque and forces recorded during the welds and also the calculated heat input.

Table 3 Summary of maximum torque and forces and calculated heat input taken from each of the welding trials

Material gradeMax torque, NmMax downforce, kNMax traverse force, kNMax lateral force, kNHeat input, kJ/mm
AA2017 240 85 18 6 3.2
AA5083 220 75 11 4 4.4
AA6082 240 65 16 2 1.8
AA7020 355 55 3 2 4.2
AA7075 280 67 13 3 3.3
AA7075
AA7020
280 53 7 3 3.8

FSW butt-weld AA2017A-T451

Weld AA2017A-T451 was made using a rotation speed of 300rev/min and a traverse speed of 120mm/min. A macro photograph of the welding zone ( Fig.4) shows that a fully consolidated weld was produced.

The downforce required to weld this material was relatively high, 85kN, with a high torque of 240Nm. The recorded traverse force was 18kN and lateral force was 6kN. These values reduced as the weld approached the edge of the plate except for the lateral force, which showed a small increase. This could be due to heat reflecting at the plate edge. The high forces and torques generated during welding are due to the material properties as 2xxx alloys maintain significant strength at the welding temperature. With a parent material hardness of ~170Hv, the hardness in the heat affected zone reduced to ~120Hv ( Fig.5). The loss of hardness in the heat affected zone (HAZ) is typical of 2xxx welds made in thinner sections.

Fig.4. Macro photograph taken from weld AA2017-T451
Fig.4. Macro photograph taken from weld AA2017-T451
Fig.5. Hardness map of weld AA2017-T451 (Scale Hv5)
Fig.5. Hardness map of weld AA2017-T451 (Scale Hv5)

FSW butt-weld AA5083-O

Weld AA5083-O was produced at 300rev/min and 80mm/min and generated a downforce of 75kN, torque 220Nm, traverse force 11kN and lateral force 4kN. These values reduced as the weld approached the edge of the plate except the lateral force, which showed a small increase. The macro photograph ( Fig.6) shows that a fully consolidated, defect free weld was produced. The high heat input is due to 5xxx alloys retaining considerable strength at the welding temperature.

Fig.6. Macro photograph taken from weld AA5083-O
Fig.6. Macro photograph taken from weld AA5083-O

The hardness map ( Fig.7) shows that a uniform hardness is present across the weld zone and parent material. This is due to the material being in the fully annealed condition, the friction stir weld has had little effect upon hardness, and again this reflects behaviour of welds made in thinner section materials.

Fig.7. Hardness map of weld AA5083-O (Scale Hv5)
Fig.7. Hardness map of weld AA5083-O (Scale Hv5)

FSW butt-weld AA6082-T651

AA6082-T651 plates were welded together using a rotation speed of 500rev/min and a traverse rate of 350mm/min. The macro photograph ( Fig.8) shows that a fully consolidated weld was achieved. The maximum downforce recorded was 65kN, torque 240Nm, traverse force 16kN and lateral force 2kN. This material is known to be readily friction stir weldable, this is shown by the ability to weld at a relatively high traverse rate when compared with the other grades of aluminium. The energy required to make this weld was considerably less than for the other materials. This is because 6xxx alloys have very little strength at elevated temperatures, and this is part of the reason why they extrude so well.

Fig.8. Macro photograph taken from weld AA6082-T651
Fig.8. Macro photograph taken from weld AA6082-T651

The hardness map ( Fig.9) shows that the parent material hardness is in the region of 130Hv-140Hv. The welding zone hardness has dropped to ~90Hv. As for welds in thinner materials, there is a very significant loss of strength in the HAZ.

Fig.9. Hardness map of weld AA6082-T651 (Scale Hv5)
Fig.9. Hardness map of weld AA6082-T651 (Scale Hv5)

FSW butt-weld AA7020-T651

Weld AA7020-T651 was performed using a rotation speed of 200rev/min and a traverse speed of 90mm/min. The maximum torque recorded was 355Nm and downforce recorded was 55kN. The traverse force and lateral force are relatively low, traverse force 3kN and lateral force 2kN. The macro photograph ( Fig.10) shows that a fully consolidated weld was produced. Again the heat input reflects the strength of the alloy at the welding temperature.

Fig.10. Macro photograph taken from weld AA7020-T651
Fig.10. Macro photograph taken from weld AA7020-T651

The hardness map ( Fig.11) shows that a lower hardness was recorded at the edges of the welding zone with values between 110Hv and 100Hv. The hardness increases in the welding zone but is significantly lower than the parent values. The hardness appears to be reasonably uniform through the material thickness. Again, the properties are similar in principle to welds made in thinner section.

Fig.11. Hardness map of weld AA7020-T651 (Scale Hv5)
Fig.11. Hardness map of weld AA7020-T651 (Scale Hv5)

FSW butt-weld AA7075-T651

Weld AA7075-T651 was made at a tool rotation speed of 200rev/min and a traverse rate of 80mm/min. The maximum downforce recorded was 67kN, torque 280Nm, traverse force 13kN and lateral force 3kN. The macro photograph ( Fig.12) shows that a fully consolidated weld was achieved. The heat input required was 3.3kJ/mm, which is broadly similar to that required for other 2xxx/7xxx alloys.

Fig.12. Macro photograph taken from weld AA7075-T651
Fig.12. Macro photograph taken from weld AA7075-T651

The hardness map ( Fig.13) shows that a significant drop in hardness in the HAZ has occurred, approximately 60Hv, with a reasonably uniform hardness through the material thickness. Higher values are found at the centre of the welding zone close to the upper surface. As for the other welds, this is similar to the observations found in thinner materials.

Fig.13. Hardness map of weld AA7075-T651 (Scale Hv5)
Fig.13. Hardness map of weld AA7075-T651 (Scale Hv5)

FSW dissimilar butt-weld AA7075-T651 and AA7020-T651

A dissimilar weld between AA7075 and AA7020 was made with a rotation speed of 200rev/min and traverse rate of 80mm/min. The material was arranged with AA7075 on the advancing side of the weld, stirring into the AA7020. The maximum downforce recorded was 53kN, torque 280Nm, traverse force 7kN and lateral force 3kN. The macro photograph ( Fig.14) shows that the material is fully consolidated with two grades in the welding zone etching differently. The heat input was between the values found for the corresponding similar welds.

Fig.14. Macro photograph taken from the dissimilar butt-weld with the AA7075-T651 material on the advancing side and the AA7020-T651 material on the retreating side
Fig.14. Macro photograph taken from the dissimilar butt-weld with the AA7075-T651 material on the advancing side and the AA7020-T651 material on the retreating side

A drop in hardness from parent hardness ( Fig.15) is evident in the welding zone for both materials as expected. The hardness drop is more significant in the AA7075 material with this material having higher parent hardness than AA7020. Similar values are found through the thickness of the welding zone at the traverse positions.

Fig.15. Hardness survey of the dissimilar butt-weld between AA7075 and AA7020
Fig.15. Hardness survey of the dissimilar butt-weld between AA7075 and AA7020

Discussion

It should be stressed that no attempts have been made in this work to optimise parameters and tool design, and it is probable that improved weld properties could be obtained with further research. It is difficult to draw any detailed conclusions from this series of tests when looking across the series of alloys as the welding parameters were not constant. However, from Table 3 it is clear that AA6082 is the least challenging of the alloys. Traverse speeds of 350mm/min are readily achievable. The scrolled shoulder tool design was not optimum for this material as it overheated the surface ofthe component. Even at these high travel speeds, a relatively low downforce and torque were required. Welding of the higher strength alloys had to be performed at lower processing speeds. The 7xxx series alloys showed quite different weld records. AA7075 required a relatively high downforce and traverse force compared with AA7020, which needed high torque for a similar weld cycle. Both AA2017 and AA5083 were run at a higher rotation speed but still requiredrelatively high downforces and traverse. The lateral forces in all cases were relatively low. However this has demonstrated that no unusual effects have resulted from the welding of thick section materials, although detailed microstructural evaluations are still underway and will be reported at a future date. Similarly, TWI and The University of Sheffield are also investigating the welding of much thicker welds (~75mm single pass) and again the data will be obtained in due course.

One welding trial was performed on the dissimilar material grades due to availability of materials and time limitation. Although the welding parameters required to weld AA7075 and AA7020 were very similar, they did not readily transfer to welding the two materials together. The weld record and the visual appearance of the weld surface when welding AA7075 to AA7020 demonstrated that the mixing was inconsistent and unstable. Further work is required to establish the extent of the process window.

Conclusions

Work to produce friction stir butt-welds in a series of similar and dissimilar aluminium alloys has resulted in the following conclusions being drawn:

  • The design and manufacture of a FSW tool to produce 20mm thick friction stir welds was achieved. However, the use of a common tool design for use across a range of aluminium alloys may not be optimum.
  • Friction stir welding of 20mm thick AA2017, AA5083, AA6082, AA7020 and AA7075 alloys is readily achievable and repeatable.
  • Further work is required to establish the extent of the process window when friction stir welding dissimilar materials AA7075 and AA7020.

References

  1. Thomas W M, Nicholas E D, Needham J C, Murch M G, Temple-Smith P and Dawes C J (TWI), 'Improvements relating to friction welding'. European Patent Specification EP 0 615 480 B1.
  2. Midling O T, Morley E J, Sandvik A (Norsk Hydro, rights transferred to TWI), 'Friction stir welding'. European Patent Specification EP 0 752 926 B1.
  3. Kallee S W, 'Friction Stir Welders Provide Prefabricated Components and Panels'. Aluminium International Today July/August 2004.
  4. Kallee S W, Davenport J, 'Trends in the design and Fabrication of Rolling Stock. European Railway Review, Volume 13, Issue 1, 2007.
  5. Kallee S W, Kell J M, Thomas W M, Wiesner C S, 'Development and Implementation of Innovative Joining Processes in the Automotive Industry', DVS Annual Welding Conference 'Große Schweißtechnische Tagung', Essen, Germany, 12-14 September 2005.

For more information please email:


contactus@twi.co.uk