Faster and faster - welding speed increases with tool development - one of a series of steps

TWI Bulletin, July/August 2000

Christopher Dawes is a Chartered Engineer by the mature candidate route. He joined TWI, then BWRA, in 1964. He is known for his work in developing welding technologies with respect to micro, resistance, laser and friction welding.

A new friction stir welding (FSW) tool has been designed to raise welding speeds above those previously achieved in 6mm thick 5083 aluminium alloy plate. As Christopher Dawes reports, previous welding rates have been exceeded, higher tolerances to speed variation have been achieved and all welds have met the mechanical test requirements.

Friction stir welding which was invented and patented ^[1] by TWI in the autumn of 1991, has transpired to be a remarkable method of achieving cost effective, high quality, continuous butt and lap seam joints in aluminium alloys. Moreover the mechanical machine tool welding operation offers generous parameter tolerances and hence process repeatability. Although a relatively new process, numerous commercial organisations throughout the world have reported ^[2-4] the advantages of FSW when compared to arc welding techniques and the process is now used ^[5] for volume production of long welds with low distortion.

A friction stir weld, Fig.1, is formed by plunging a rotating shouldered pin tool, with a pin length slightly less than the weld depth required, into the faying workpiece faces until the tool shoulder is in intimate contact with the work surface and then moving the work against the pin, or vice versa. The rotating tool pin within the workpiece friction heats the metal and produces a plasticised tubular shaft of metal around the pin. As the pin is moved in the direction of welding the leading face of the pin, assisted by a special profile, crushes and forces plasticised material to the back of the pin whilst applying a substantial mechanical forging force which consolidates the weld metal.

The tool technology is one of the most important features of the FSW process. The tool shape determines the heating, shearing, crushing and forging action on the hot worked metal. Tool size determines the weld size and tool strength. The tool material determines the rate of friction heating, tool strength and working temperature, the last ultimately determining the range of materials to which FSW can be applied.

To date, most work on the process has been conducted on a confidential TWI Group Sponsored Project (GSP) which has developed a tool and schedules for welding various Al alloy sheets and plate of 1.5-12mm thickness. ^[6] (Other tool designs have also been published. ^[7-9] )

The 5xxx series alloys are widely used, especially where good strength and high corrosion resistance are required. The potential applications for FSW joining this series of alloys are many and range from road tankers to fast ferries, all of which generally require long, low distortion welds. For such welds there is much to be gained from manufacturing time savings if the welding speed can be increased, even by only 10%, whilst maintaining high weld quality.

It is unlikely that the rubbing velocity and hence tool rotation speed for welding 5xxx series alloys can be increased to anywhere near that for welding 6xxx alloys, without researching suitable low friction tool shoulder materials. Therefore the short term opportunity for increasing welding speed lies in trying to minimise the tool shoulder diameter, to increase tool rotation speed, and to develop tool shoulder and pin profiles so as to increase the effective forward movement of the tool per revolution. TWI has investigated these possibilities and this report describes the results. The work described is complementary to another TWI research report by Thomas et al ^[9] concerned with determining the effects, at certain generic tool geometries, on weld formation when welding 6.4mm aluminium alloy plate, and in using this knowledge to demonstrate the capability of FSW for welding 50 and 70mm plate.

Objective

• To increase welding speed, whilst maintaining good weld quality, when making FSW butt joints between 6mm thick aluminium alloy 5083-O plates.

The original tool concept

The GSP tool proved a concept based on a tool with a shoulder profile and a threaded pin. In principle the shoulder, as well as capturing the weld metal and heating and forging it on the trailing side of the tool pin, also provides compaction of a surface-proud moving cone of material into the upper thread forms on the 180° trailing face of the tool pin. The aim of this action is to try to ensure that the weld material within the thread is always over compressed by the rotating downward auger action of the thread, thus producing a continuous forged extrusion of weld metal off the trailing thread form flanks of the tool pin within the depth of the workpiece. This forged extrusion of weld metal is considered to be a major contributor in producing the very fine grain size found in the recrystallised weld microstructure of FSW joints when compared to that of the parent metal.

The scroll shoulder tool concept investigated

• Small active shoulder diameter.
• High shoulder surface heating area.
• Means of continually compacting the weld metal around the tool pin upper thread forms.
• Coarse tool pin thread pitch.

This forces an annular ring of material around the upper threads of the pin. Once the annulus reaches its maximum state of material compression, the surplus material within the scroll, which is continually friction heated and sheared from the plate surface, retains sufficient plasticity and radially inward mechanical force to maintain maximum material compression about the upper threads of the tool pin. Furthermore the small cross-section of the annulus of material about the upper portion of the tool pin provides easier plastic flow of this material across the plate surface. It also eliminates the tendency of the tool to climb the material cone on the trailing side of the pin at higher welding speeds.

An additional benefit is provided by the scroll entry profile at the tool shoulder periphery. This profile catches the small wave of material extruded from the work surface on the advancing side of the tool shoulder and returns most of this material back into the plate. Thus the tendency of material build up on the advancing side of the tool is almost eliminated.

The last design feature was to use a coarse FSW tool pin thread pitch of 1.5mm. This feature was taken to try and enhance the crushing and dispersion of any oxide within the weld metal.

Materials and preparation

The 5083-O condition alloy used for the tool evaluation trial was 6mm thick. The material tensile strength was 281 N/mm ² , with an average elongation of 20% and a 0.2% proof stress of 130 N/mm ² . The chemical composition supplied by the vendor was as shown in the Table.

Table: Chemical composition element wt% (supplied by the vendor)

The weld test coupons were 400mm long x 100mm wide, thus producing welded specimens 400mm x 200mm for testing. The faying surfaces were milled square to the plate surface and prior to welding were wiped clean with petroleum ether.

Equipment

Tool evaluation procedure

Butt welds were made using tool rotation speeds of 400, 500 and 600 rev/min. The welding speeds used were 100, 120, 140 and 160 mm/min at 400 rev/min tool speed and increased to 200 mm/min at 500 rev/min and 220 mm/min at 600 rev/min. For each combination of welding speed and tool rotation speed, the tool heel plunge depth was set to ~0.2mm below the tool plunge depth control wheels, which ran on the surface of the plate specimen. The tool was tilted 2.5° away from the direction of welding.

From each weld made at the above machine settings, three cross weld tensile test specimens, two bend test specimens and one macrosection specimen were taken. The positions within each welded pair of plates are shown in Fig.5. The tensile tests were conducted to BS 10002-1:1990 and the bend tests to BS EN 288: Part 4: 1992. Of the two bend test specimens taken from each butt welded plate, one was bent with the weld root in tension and one with the weld face in tension.

The specimens taken for macrosection examination were mounted in a cold setting epoxy resin, and then polished to a one micron finish and etched in Kellers' reagent. Macrophotographs were taken of all welds sectioned.

The criteria for weld quality acceptance was that the weld tensile strength would equal that of the parent material in the annealed condition. The welds could be bent through 180° without fracture and that the macrosections exhibited no signs of voids or cracks.

Results and discussion

Tensile tests

The TEST HOUSE

Certificate of Test

Fig.7. Weld tensile strength test results

Comments:   Extensometer number 135/100S, calibrated to BS EN 10002-4 grade 0.5, was used for these tests. Mark 3, 5, 6 failed in the weld region.
Note - The test results detailed above apply only to the sample(s) of material submitted to the laboratory.

Tests Performed by:  B Kersey Certificate Approved by:  J V Curzon Signed     Date

Witnessed by:

The Test House (Cambridge) Ltd, Abington Hall, Abington, Cambridge CB1 6AL.
  Tel: 01223 894252  Fax: 01223 894255   E-mail: testhouse@dial.pipex.com

Registered in England No. 2513984  Registered Office: Abington Hall, Cambridge.

The Test House is a trading name of The Test House (Cambridge) Ltd, a wholly owned subsidiary of TWI.

In the majority of cases the tensile specimens failed in the parent metal, well away from the weld region. Those that failed in the weld region exhibited classic 45° fine grained ductile shear failures and showed signs of propagating from the root of the joint and beneath the weld nugget. This indicates in these cases the tool pin depth may not have been enough. Examples of both failure zones are shown in Fig.8.

Bend tests

Macrosections

Figure 10a-e shows transverse sections through welds made at the four corners and at the position almost halfway along the longest boundary of the scroll shoulder tool envelope. In all five sections signs of the re-orientated, stretched and ruptured original joint interface can be traced; as can be found in most FSW sections. Also at the root of each section a very short, near vertical line of original interface can just be seen, which was not adequately encompassed by the action of the tip of the experimental FSW tool pin. In Fig.10c-e, signs of the original interface can be traced snaking up to the plate top surface just to the right of the centre of the weld. However, in the unetched specimen state these broken hair like marks cannot be found, thus indicating a material grain orientation etching effect.

The tool shoulder scroll, together with the tool pin changes, had a marked effect on the material flow pattern in the upper half of the weld. The macrosections show how the scroll shoulder almost eliminates the small wave of material normally extruded from the plate surface (see left hand side of macrosections) on the advancing side of the weld.

All the welds made within the scroll shoulder tool envelope studied proved to be very good in terms of bend and tensile strength properties, but more development work should be done with the tool pin profile to improve the lower region of the weld; especially in the region towards the weld root. In this respect the Mk II tool pin, with its 1.5mm pitch helix and flat crested thread form feature, has made no improvement.

The increase in welding speed

The scroll shoulder tool can maintain tensile and bend strength quality to a high welding speed. If a vertical line is struck across the envelope in Fig.6 at the 160 mm/min welding speed, a tolerance box is given for the scroll shoulder tool, which equates to a tool rotation speed of 500 ± 100 rev/min. However, since tool rotation speed can be accurately controlled and monitored, the wide tolerance on tool rotation speed variation is perhaps unnecessary. Consequently, in circumstances where the tool speed can be confidently maintained at 600 rev/min, a welding speed of 210 mm/min can be achieved in the 5083-O alloy in question, thus making possible a substantial increase in welding speed (90%) over that achieved previously with the GSP tool. ^[6] Even higher welding speeds may be possible, since acceptable welds were made at the highest speed studied.

The maximum forward tool movement per tool revolution for the scroll shoulder tool was 0.4mm at the welding speed of 160 mm/min and tool speed of 400 rev/min, and 0.35mm at 210 mm/min and 600 rev/min. The scroll shoulder profile and associated increased shoulder heating area has enabled a high tool forward movement per tool revolution. The scroll shoulder tool has an effective shoulder diameter of 20mm giving a peripheral velocity of 0.63 m/sec at 600 rev/min. At 0.8 m/sec this tool would be revolving at ~750rev/min. Therefore, at a forward movement per tool revolution of 0.35mm a theoretical welding speed of 260 mm/min is possible.

Conclusions

An FSW tool with a scroll shoulder has been designed and developed with a view to increasing the welding speed above that achieved previously under a confidential TWI GSP when using the GSP tool. The major feature of this new tool is a shoulder scroll profile designed to ensure continuous compaction of weld metal around the upper reaches of the tool pin.

The following conclusions are drawn from welds made in 6mm thick aluminium alloy 5083-O plate and assessed on the bases of weld tensile and 180° bend strength tests and from transverse sections of the welds.

• A welding speed of 210 mm/min was achieved with the scroll shoulder tool at a tool rotation speed of 600 rev/min.
• A high tolerance to variation in tool rotation speed was maintained at the higher welding speed with the scroll shoulder tool (500 ± 100 rev/min).
• Welds made at the above welding conditions met tensile and bend test acceptance criteria required by BS 10002-1 and BS EN 288.
• Even higher welding speeds may be possible with the scroll shoulder tool.
• This new tool has been proved on a 5xxx series alloy only and may not be suitable for all the other alloy series.

Recommendations

Those seeking to increase FSW speed, when butt welding 5xxx series alloys, should seriously examine the use of a scroll profile on the FSW tool shoulder.

Acknowledgements

The author would like to thank Philip Threadgill for his advice and encouragement and acknowledge the efforts of David Staines and Ted Spurgin who carried out the welding trials and weld testing.

References

The story today ...
The work reported here will shortly be superseded by a current follow-up Group Sponsored Project to develop a new generation of high speed FSW tools. For further information contact Christopher Dawes.