Hybrid Nd:YAG Laser-AC MIG welding of thin section automotive aluminium alloy

C M Allen

TWI Ltd., Granta Park, Gt. Abington, Cambridge CB1 6AL, United Kingdom.

Paper presented at Eurojoin 6, Santiago de Compostela, Spain, 28 - 30 June, 2006.

Abstract

Hybrid laser-arc welding processes, in which a laser and an arc are combined in the same process zone, can offer a number of benefits over autogenous laser welding, including increased productivity and a tolerance to fit up comparable with arc welding. This paper concentrates on the development of welding parameters for an Nd:YAG laser-AC MIG arc hybrid process for thin aluminium sheet, for potential application in the automotive industry. Melt runs, butt welding and lap welding trials have been carried out on 1.2mm thick 5251-H22 aluminium alloy, with characterisation by metallographic section and radiography. A hybrid condition has been developed which is stable to welding speeds in excess of 8m/min, representing a four fold increase in productivity compared to MIG welding or a two fold increase in productivity compared to laser welding. This hybrid condition has been applied with particular success to edge lap welds, where a gap tolerance of ten times that of autogenous laser welding has been demonstrated. Internal weld porosity is outside of BS EN ISO 13919-2:2001, but any application would require a case-by-case assessment of fitness for purpose, and/or more thorough material preparation to reduce porosity.

Introduction

Hybrid laser-arc welding, originally proposed in the late 1970s, ^[1,2] combines laser and arc welding processes in a single process zone. It offers the benefits of the separate processes, and overcomes some of their respective drawbacks. Reported benefits compared to laser welding ^[3,4,5] include:

• Increased tolerance to joint fit-up.
• Greater welding speed, leading to higher productivity.
• Increased penetration.
• Lower net heat input, leading to reduced distortion.
• Improved weld quality.
• The potential to replace some laser power, for a given depth of penetration, by some less expensive arc power, thereby increasing cost effectiveness.

Hybrid welding has received renewed interest in the past ten years, principally concentrating on laser-MAG welding of C-Mn steels, e.g. in shipbuilding panel lines. ^[6] An area of potential growth for the hybrid process is in the automotive sector for thin sheet aluminium joining. ^[7,8] Aside from hybrid welding operations, the automotive industry is making investments in laser technology, particularly in autogenous laser welding and/or laser brazing of thin sheet steels for car body construction, using robot mounted fibre optic delivered Nd:YAG laser beams. These fabrication methods are flexible, high productivity, and low distortion.

This paper addresses the further development of laser welding, through Nd:YAG laser-AC MIG arc hybrid welding of thin sheet aluminium alloy for future lightweight car body constructions.

AC MIG is a low heat input arc process, with demonstrated penetration control and good gap bridging ability when welding thin sheet aluminium. ^[9,10,11] In AC MIG, each current pulse passed through the consumable electrode consists of both an electrode positive (EP) part and an electrode negative (EN) part. The EN part melts the electrode in preference to the base material, increasing deposition rate and hence gap bridging ability, and also reducing heat input in to the base material. In summary, this low heat input, fine penetration control, high deposition rate technique is well suited to welding of thin sheet with demonstrated tolerance to joint gaps.

The AC MIG process has been hybridised with low power density Nd:YAG and diode lasers, ^[10,11] and welding speeds of up to 4m/min have been reported, compared to 3m/min using the AC MIG arc without laser, i.e. a productivity increase of 33%. In this work a Nd:YAG laser of higher power density and capable therefore of keyhole welding, is combined with AC MIG welding, with the aim of offering welding speeds and productivity better than that of existing autogenous keyhole laser welding processes, and with a gap bridging ability better than laser welding and more comparable to the AC MIG process. Introduction of such a hybrid process in to an existing Nd:YAG laser welding line would allow relaxation of fit up tolerances, and increases in productivity, with minimal additional capital expenditure being incurred.

Experimental Methods

Materials and preparation

5251-H22 (Al-2Mg) sheet aluminium alloy, 1.2mm in thickness, was used. Prior to welding the sheet edges were dry machined, and top and bottom faces of the sheets degreased with acetone. No further steps were taken to remove any hydrated aluminium oxides from the region of the weld, which might otherwise improve weld quality (e.g. reduce internal porosity). Such approaches would be unlikely in an automotive fabrication environment. Two filler wires were used, both 1.2mm in diameter: AWS ER5356 (Al-5Mg) wire and AWS ER5556 (Al-5Mg-1Mn) wire.

Equipment

A Trumpf continuous wave HL4006D Nd:YAG laser was used, operating at power levels at the workpiece of up to 3kW, whose beam was focussed by a robot mounted optic to a 0.6mm diameter spot on the sheet upper surface. A Daihen CPDACR 200 arc power source was used for AC MIG welding with an OTC-Daihen CMWH-147 wire feeding unit. The laser was used with a travel angle (dragging) of 10° off vertical, leading the AC MIG process by a separation of 2mm, with the MIG torch having a travel angle (pushing) of 15° off vertical. These parameters were chosen on the basis of previous TWI experience with hybrid welding of aluminium and reference to. ^[10,11] Shielding of the weld pool was provided by a flow of 20l/min of either He or Ar down the MIG torch. A slot had to be cut into the MIG gas shroud to avoiding clipping by the laser beam. In the case of full penetration welds, the underbead was shielded by a flow rate of 5l/min of Ar, supplied through an efflux channel measuring 10mm x 10mm, machined in to the welding jig.

Welding experiments

Given the number of process parameters and hence complexity of the hybrid process, welding parameters were optimised in turn for:

• Full penetration autogenous laser melt runs on sheet
• Full penetration autogenous laser butt welds
• Full penetration AC MIG melt runs on sheet
• Full penetration hybrid melt runs on sheet
• Partial penetration hybrid edge lap welds between two overlapping sheets. In selected cases, the gap between the two sheets was tapered from 0mm to 2mm, in order to assess the gap bridging ability of the hybrid process
• Full penetration hybrid butt welds.

The ranges of principal variables used in each set of experiments are summarised in Table 1.

Weld Examination

Radiography was performed to BS EN 1435:1997, to determine the presence of welding imperfections, eg extent of porosity. Transverse sections were prepared using standard metallographic techniques, with subsequent metallographic examination determining weld shape and depth of penetration.

In the absence of a broadly accepted standard for fitness for purpose of laser welds for automotive application, as a relative indicator of weld quality, a standard for laser weld workmanship was used, BS EN ISO 13919-2:2001. As this is solely a standard for workmanship, failure to meet a given class of weld in this standard does not imply failure to be fit for a given application or purpose. The latter should be assessed independently on a case by case basis.

Table 1. Ranges of principal experimental variables.

Expt. type	Laser power, kW	Travel speed, m/min	Top bead shielding	Arc current, amps	Voltage trim setting*	Penetration control setting**
Laser melt run	3.0	3.0-8.6	20l/min Ar or 20l/min He	n/a	n/a	n/a
Laser butt weld
AC MIG melt runs	n/a	1.0-2.0	20l/min Ar or 20l/min He	40-90	-2 to +3	-
Hybrid melt runs	0.5-3.0	7.6-10.0	20l/min Ar	80-130	0 to +5	-5 to +5
Hybrid butt welds
Hybrid edge lap welds

Notes:
- = not varied
* The Daihen AC MIG power source has an operation mode in which the mean arc current can be freely adjusted and set by the operator prior to welding. A feature of this operation mode, in common with other commercially available arc welding power sources, is that other welding parameters such as the arc length or voltage, and pulse characteristics of the current, are set synergically within the power source. The current set by the operator and the synergic program selected, determine the values of these other parameters. In all work a synergic program setting of '42' was used, corresponding to an AC pulsed MIG operation with a 1.2mm diameter Al-Mg wire. Slight adjustments or 'trims' by the operator to the arc voltage away from this synergic setting are possible, by selecting various positions on a potentiometer dial on the power source. In this manner, deviations of up to +/-5V from the synergic voltage can be selected. The details of the synergic program and trim settings are included as an aid to the practical user wishing to reproduce the results presented, although specific to the power source used.
** the EN ratio of the current setting can be adjusted by the operator away from the synergic setting, through different dial settings of a potentiometer, between arbitrarily denominated values of -5 to +5. The EN ratio has been documented to change the penetration characteristics when AC MIG welding. ^[10] Once again, these settings, when used, have been included as an aid to the practical user wishing to reproduce the results.

Results and discussion

Autogenous laser melt runs

With He top bead shielding of 20l/min, at travel speeds of 3 and 4m/min, melt-through occurred in a few positions ('pinholes') along the weld length. At a travel speed of 5m/min, a consistent top and underbead was achieved. At travel speeds of 6 and 7m/min penetration was lost.

Switching to 20l/min Ar top bead shielding, all welds produced visually exhibited a brighter top bead appearance, being more effectively shielded against oxidation due to the greater density of Ar compared to He. At a travel speed of 3m/min, pinholes occurred. At travel speeds of 4 and 5m/min consistent top beads and underbeads were achieved, with the reduced heat input at 5m/min being preferred. This condition was selected as optimum and repeated three times to check consistency. Fig.1 shows a cross-section through one of these melt runs, whose profile was acceptable to the highest class, class B (stringent), according to BS EN ISO 13919-2:2001. Radiography of these melt runs also indicated an internal quality acceptable to class B (stringent) of BS EN ISO 13919-2:2001. At higher travel speeds, of 6, 7 and 8m/min, penetration became intermittent.

Fig.1. Cross section through an autogenous laser melt run produced at 5m/min with a laser power of 3kW and with Ar shielding

Autogenous laser butt welds

The optimum laser melt run condition was transferred to a nominal zero gap butt weld between sheets with machined edges. At 5m/min a narrower top bead was observed than when performing a melt run, and some localised losses in penetration were observed. Reduction in travel speed by 10%, i.e. to 4.5m/min, increased penetration and top bead width. This condition was again repeated three times to check consistency. Fig.2 shows a cross-section through one of these butt welds. The weld profile was again acceptable to class B (stringent), as was the internal weld quality as determined by radiography.

Fig.2. Cross section through an autogenous laser butt weld produced at 4.5m/min with a laser power of 3kW and with Ar shielding

AC MIG melt runs

For AC MIG melt runs, the slotted torch shroud to be used for hybrid experiments was found unsuitable, therefore a conventional shroud improving both melt pool shielding and arc stability, was used. Similarly, arc stability was much improved using Ar as a shielding gas, with an electrode stick out and shroud stand off of 10mm, compared to using He as a shielding gas, and an electrode stick out and shroud stand off of 15mm. Varying both arc current setting in increments of 10A (without voltage trim being applied) and travel speed in increments of 0.25m/min, the most stable condition was found to be 50A at 1m/min. Penetration was less consistent than for the laser melt runs. Faster travel speeds resulted in loss of arc stability, and higher currents resulted in melt through.

In a second round of trials voltage trim was applied. The condition that gave a stable arc at the highest welding speed achieved was using an arc current set to 80A at 2m/min, with a voltage trim of +2 or +3. This indicates that the application of a positive trim stabilises the arc to both higher currents and welding speeds. Top bead and underbead photographs from an AC MIG arc melt run at 2m/min with an arc current set to 80A with a +3 trim arc are shown in Fig.3. Penetration was more consistent than that achieved when using a 50A arc current, untrimmed, at 1m/min, but still not fully consistent, as is shown. In terms of weld profile, and particularly internal quality (porosity), this melt run was not acceptable to BS EN ISO 13919-2:2001. Weldments can tolerate an appreciable amount of porosity, that amount being material and alloy dependant, without significantly affecting static mechanical properties such as yield strength, and tensile strength and elongation being reduced as cross-section area is reduced. ^[12,13] This philosophy, of an acceptance of a certain porosity level, is commonly taken in the automotive industry, where expensive, time-consuming, and difficult to apply pre-welding aluminium cleaning treatments to reduce weld porosity levels are not the norm. That said, reduced porosity levels in both AC MIG and hybrid laser-AC MIG welds can be achieved with more attention to material cleaning. ^[10,11]

Fig.3. Top bead and underbead of selected AC MIG along melt run condition at 2m/min travel speed, using a mean arc current set to 80A, with voltage trim set to +3
a) Top bead	b) Underbead

Hybrid melt runs

1. Basic conditions: Combining the arc and laser, with the AC MIG melt run conditions (initially, those developed without voltage trim) but at the optimum laser melt run travel speed, successfully resulted in a stable arc at much higher speeds than without the laser. However, the increase in heat input led to melt through. A variety of higher travel speeds and arc current settings (all without voltage trim) were tried to find a stable condition which led to consistent penetration without melt through, resulting in an optimum condition at 7.6m/min with a mean arc current set to 80~90A. This represented a ~50% increase in welding speed compared with autogenous laser welding, and just over three and a half times faster than the fastest condition established when using the AC MIG arc on its own.

2. Effect of penetration control trim: Aside from reducing travel speed down from 8m/min to 7.6m/min to achieve more consistent penetration, the penetration trim control was also adjusted on the arc power set. Penetration trim control settings were chosen between a minimum of arbitrary denomination '-5' to a setting of '+3' (maximum possible setting was '+5'). These adjustments change the EN ratio of the AC current, although the magnitude of these changes in EN ratio were not recorded. Penetration control trims were applied to a reference melt run condition of mean arc current of 80A (with zero voltage trim) at a travel speed of 8m/min. Unlike the reported effect of penetration control trim on AC MIG welds, ^[11] no such effect was seen in the hybrid melt runs. This was probably due to the dominance of the laser over the arc, in terms of achieving penetration, at the high speeds used in these experiments.

3. Effect of reducing laser power: Melt runs were performed at lower travel speeds with reduced laser powers, to determine the minimum laser power required to stabilise the arc, and up to what speed that arc would be stabilised. All experiments were performed with a mean arc current set to 80A with zero voltage trim. These experiments are summarised in Fig.4. As Fig.4 shows, 1kW of laser power stabilised the arc at 1.5m/min, representing a 50% increase on the welding speed of 1m/min achieved without the laser (shown as the 'arc alone' point in Fig.4). However, a trimmed 80A arc, as reported above, was stable to 2m/min. At these modest welding speeds, voltage trim appears to be a far more 'economic' means of stabilising the arc than adding a low power (1kW) focussed laser source.

Fig.4. Hybrid laser-arc melt run conditions, as a function of whether a stable arc was achieved, with laser powers <3kW, and with the two best hybrid conditions achieved with 3kW, and the best condition using the AC-MIG arc on its own also included. Dashed line indicates inferred delineation between region of stable and unstable arc conditions. All results without voltage trim

4. Effect of voltage trim: As with AC MIG melt runs, the application of voltage trim stabilises the arc to higher currents and welding speeds. With a welding speed of 8.2m/min, the arc was stabilised to a mean current setting of 110A with a voltage trim of +3. This represents a welding speed increase of up to 80% compared with laser welding, and a ~40% current increase compared to an untrimmed hybrid setting. Top bead and underbead photographs from this hybrid melt run are shown in Fig.5. This profile of this particular melt run was to class C (intermediate) in accordance with BS EN ISO 13919-2:2001. However, as with the AC MIG melt runs, internal porosity was not accepted to BS EN ISO 13919-2:2001, and a separate fitness for purpose assessment would be required.

Fig.5. Top bead and underbead of selected hybrid melt run condition at 8.2m/min, with mean arc current set to 110A, and with a voltage trim of +3:
a) Top bead	b) Underbead

Hybrid edge lap welds

1. Basic conditions: For edge lap welding, as a start point the same conditions (without voltage trim) and travel angles were used as for hybrid melt runs. Initially, a work angle of 20° off vertical was used but this led to penetration of the underlying sheet. The work angle was therefore increased to 40° off vertical, the maximum possible given the diameter of MIG shroud used, with the laser focussed on to the top surface of the lower sheet and the MIG wire aimed in to the corner of the joint. This also resulted in penetration of the underlying sheet. To avoid penetration a series of further experiments indicated that it was necessary to increase travel speed to 8.6m/min, reduce laser power to 2.8kW, and, to counteract the resulting loss in arc stability, increase the mean arc current set to 110A. Top bead and underbead photographs from a hybrid edge lap weld with these conditions are shown in Fig.6. Fig.7 shows a cross-section through this lap weld.

Fig.6. Top bead and underbead of hybrid lap weld welded at 8.6m/min, with a mean arc current set to 110A, without voltage trim, and using a laser power of 2.8kW
a) Top bead	b) Underbead

Fig.7. Cross section through a hybrid lap weld at a travel speed of 8.6m/min, with a mean arc current set to 110A, without voltage trim, and using a laser power of 2.8kW

2. Effect of voltage trim: Using the above conditions, applying a positive voltage trim of setting +3 further stabilised the arc, for example to current settings of up to 130A at a travel speed of 8.2m/min. It was found necessary to position the wire 2mm out of the joint line of the edge lap joint line on the underlying sheet to maintain a regular top bead appearance. This was probably due to preferential arcing along the shortest path, i.e on to the top corner of the upper sheet, which occurred when the wire was positioned pointing directly at the joint line. With these increased arc current conditions penetration of the lower sheet occurred once again. In an attempt to reduce penetration, the travel speed was increased to 8.6m/min, however, this destabilised the arc. More successful was to reduce the laser power to 2.9kW. The fact that small (<10%) changes in process parameters led to large differences in process stability and weld profile do indicate that the operating window of this process is relatively small. This condition was repeated three times to check consistency. Top bead and underbead photographs from one of these hybrid lap welds are shown in Fig.8. Fig.9 shows a cross-section through this lap weld. As with the hybrid melt runs reported above, the radiographs of these welds contained a number of fine pores, with a mean maximum diameter of ~0.3mm. As noted before, the presence of this porosity would necessitate a fitness for purpose assessment, rather than simple adherence to a standard of workmanship. It is anticipated that reduced levels of porosity could be achieved by more stringent parent material preparation prior to welding, but this was not considered to be representative of the preparation that would be carried out routinely in the automotive industry, to whom this work had been targeted.

Fig.8. Top bead and underbead of a hybrid lap weld at a travel speed of 8.2m/min, with a mean arc current set to 130A, with a voltage trim of +3, and using a laser power of 2.9kW
a) Top bead	b) Underbead

Fig.9. Cross sections through a hybrid lap weld at a travel speed of 8.2m/min, with a mean arc current set to 130A, with a voltage trim of +3, and using a laser power of 2.9kW

3. Gap bridging: Edge lap welds were made using the hybrid condition developed, with tapered gaps between the sheets both starting at zero and running to a nominal gap of 2mm, and from a nominal gap of 2mm running to zero gap. These welds were then compared with equivalent autogenous laser welds made at the same travel speed. Table 2 summarises the gap bridging results, with the actual gap sizes being determined by feeler gauge and/or cross-sectioning.

Table 2. Gap bridging results for edge lap welds made by the hybrid laser-AC MIG process and the autogenous laser process.

Weld type	Nominal gap	Gap bridging* until
Hybrid laser-AC MIG	0-2mm	~0.9mm
Hybrid laser-AC MIG	2-0mm	~1.1mm
Autogenous laser	0-2mm	~0.1mm
Autogenous laser	2-0mm	Determined by feeler gauge to be <0.1mm

Note:
*Gap bridging for the purposes of this work was defined as physical connection of the two sheets after welding, with the resulting weld bead of a profile not necessarily acceptable to ISO 13919-2.

As Table 2 shows, in the case of the hybrid welds, with a gap increasing from zero, bridging was maintained to a gap size of ~1mm, and with a gap tapering down to zero, gap bridging was first achieved at a gap size also of ~1mm. With larger gap sizes holes appeared in the top bead. In the autogenous laser welding process gap bridging was lost at a value ten times as small at ~0.1mm. The hybrid process is therefore far better in terms of gap bridging in the case of this joint geometry. This large difference arose from the supply of extra weld metal material from the MIG wire consumable. Laser with cold wire feed would be more tolerant than the autogenous process, but welding speeds would have to be reduced to allow the laser to not only melt the parent material but the wire as well. The hybrid process has the advantage that arc energy, not laser, is effectively used to melt the wire.

Hybrid butt welds

1. Basic conditions: For butt welding, as a start point the same conditions (without voltage trim) were again used as for hybrid melt runs. With these conditions penetration was heavier, and localised melt-through occurred in one position. Increasing welding speed to 8.4m/min still led to localised melt through, and higher speeds led to loss of penetration.

2. Effect of voltage trim, heat input, process separation and laser defocus position: Following on from earlier work, a +3 voltage trim was selected. However, this again resulted in localised melt-through. Reducing heat input by increasing speed and/or reducing laser power reduced, but could not entirely eliminate, the occurrence of these localised melt-throughs. Different laser-arc separations of 0mm and 4mm were tried, but without success, and in the case of a 4mm separation arc stability was lost. Different laser defocus positions of +2mm and +4mm were also tried, but led to loss of penetration. The origin of these localised melt throughs, or pinholes, may result from short time scale (<20ms) variations in any one of the following:

• Arc power: short term variations in the arc were indeed measured using high frequency monitoring equipment.
• Wire feed rate: consistent high speed feeding of soft aluminium wires during arc welding is a documented problem.
• Laser power arriving at and absorbed by the work: as opposed to variations in output power, these would more likely be fluctuations in power arriving at or absorbed by the work (e.g. due to fluctuations in the laser plume or keyhole).
• A 'random event' occurring in the weld pool (eg sudden localised build up of porosity).

Whatever the cause, these hole features were not seen at the low speeds used when using the arc on its own, where short time instabilities may be better accommodated due to the longer freezing time/slower solidification velocity. Nor were they seen in hybrid melt runs or non fully penetrating butt welds, where the weld pool may be slightly better supported, due to the absence of any abutting edges, or underlying non-melted material respectively.

Conclusions

The hybrid Nd:YAG laser-AC MIG welding process benefits, (and limitations), for joining 1.2mm thick sheets of 5251-H22 aluminium alloy have been quantified for both butt welding and edge lap welding. The main conclusions of this work are:

• The Nd:YAG laser and AC MIG welding processes can be successfully combined in a hybrid process suitable for high speed welding of thin sheet aluminium for automotive body construction.
• Hybrid welding at speeds of over 8m/min at 3kW laser power have been achieved, over four times faster than the AC MIG arc on its own, and up to 80% faster than autogenous laser welding.
• The application of a positive voltage trim stabilises the AC MIG arc to currents greater than 50% higher than without trim.
• When hybrid edge lap welding, gaps of up to 1mm between 1.2mm thickness sheets can be bridged, ten times those tolerated by equivalent autogenous laser welding process.
• An assessment of fitness for purpose of the welds made using this process would need to be made on a case by case basis. Attention to material cleanliness may be required to reduce the observed porosity levels, if indeed these prove unacceptable for a given application.
• This hybrid welding process appears sensitive to drop through in full penetration welds, e.g. butt welds, and is therefore better suited to partial penetration weld geometries, such as edge lap welds.

Acknowledgements

This work was funded by the Industrial Members of TWI as part of the TWI Core Research Program.

References

1. Steen W M and Eboo M: 'Arc Augmented Laser Welding'. Paper presented at the 4th International Conference on Advances in welding processes, Harrogate (UK), 9-11 May 1978. Part 1, paper 17, pp.257-265.
2. Steen W M and Eboo M: 'Arc Augmented Laser Welding'. Metal construction Vol.11, No.7, July 1979, pp.332-335.
3. Downs D L and Mulligan S J: 'Hybrid CO ₂ laser-MAG welding of carbon steel - a literature review and initial study' TWI Members Report 739, March 2002.
4. Shi S G and Hilton P A: 'A comparison of the gap-bridging capability of CO ₂ laser and hybrid CO ₂ laser MAG welding on 8mm thickness C-Mn steel plate'. TWI Members Report 792, February 2004.
5. Allen C M: 'Laser welding of aluminium alloys - principles and applications'. TWI Members Report 795, March 2004.
6. Roland F, Reinert T and Pethan G: 'Laser welding in shipbuilding - an overview of the activities at Meyer Werft'. Paper presented at the IIW International Conference on Advanced processes and technologies in welding and allied processes, Copenhagen (Denmark), 24-25 June 2002, pp.B-II.
7. Anon: AutoTechnology 6/2002, pp.36-39.
8. Graf T and Staufer F: 'Laser-hybrid welding drives VW improvements'. Welding Journal, January 2003, pp.43-48.
9. Mulligan S J: 'Pulsed MIG arc welding processes for joining of thin sheet aluminium'. TWI Members Report 771, July 2003.
10. Ueyama T, Tong H, Yazawa M, Hirami M, Nakata K, Kihara T and Ushio M: 'High speed welding of aluminium alloy sheets with laser assisted AC pulsed MIG process'. IIW Asian Pacific International Congress, Proceedings, Conference, Singapore, October 2002, Vol.3-4, Paper 33, pp.11.
11. Tomita N, Ueyama T, Hasegawa S, Yasufuku T and Ueda Y: 'Development of laser-arc hybrid welding robot system'. IIW-1791-2004.
12. Harris I D: 'A review of literature on porosity formation and recommendations on the avoidance of porosity in MIG welding'. TWI Members Report 386, December 1988.
13. Lawrence F V, Jr and Munse W H: 'Effects of porosity on the tensile properties of 5083 and 6061 aluminium alloy weldments', Welding Research Council Bulletin, No. 181.

Copyright ©2006 TWI Ltd