Static and Fatigue Behaviour of Spot Welded Aluminium Alloy

The limited effect of discontinuities in resistance spot welds on joint properties is demonstrated for automotive applications

(Originally published in Welding Journal, 1999, Vol. 78, No. 3, March, pp 80-s - 86-s
Publishers - American Welding Society - http://www.aws.org )

By A Gean*, S A Westgate**, J C Kucza*, and J C Ehrstrom*

* Pechiney CRV, Voreppe, France
** TWI, Abington, Cambridge, UK

Abstract

There is a strong interest in the use of aluminium alloy sheet for vehicle applications, particularly the body where resistance spot welding is the principal joining method. It is important that the particular discontinuities, which are often found in aluminium alloy spot welds, do not adversely affect the weld properties. The objectives of this work were to provide information about the effect of excessive porosity and surface indentation, and the effect of weld size on the fatigue performance of spot welds in aluminium alloy sheet.

Trials were conducted on 1.2mm thick 5182-0 aluminium alloy, in the mill finished condition. Static shear and fatigue tests were conducted on welds over a range of welding conditions to simulate severe weld discontinuities. The work indicated that nugget porosity, up to about 40% of the weld diameter, deep surface indentation and variation in weld size had no major impact on the fatigue properties of the welds.

Introduction

The need to reduce vehicle weight, to improve fuel economy and thus also to reduce exhaust emissions, has led to increased use of lightweight materials such as aluminium alloys. While the space frame concept ^[1] has been claimed as being a cost-effective way of achieving a high performance vehicle structure, it remains suited to low volume manufacture. Aluminium alloys have found applications in the more classical design of higher volume vehicles, competing with zinc coated steels for hoods, trunks and doors, for example.

Resistance welding remains the principal joining process in the vehicle industry. It is a rapid, reliable, cost effective and now highly automated joining process for the low carbon, high strength and coated steels currently used. Aluminium alloys can also be spot welded to commercial quality levels without the need for special cleaning or surface treatment of the material ^{[2 to 5]} . Typically the welding current is double that required for zinc coated steels but the weld time only one third.

Extensive work has been done to improve the suitability of spot welding of aluminium alloys for mass production industries, particularly on welding conditions, electrodes and power supply types, mainly to improve electrode life ^{[5 to 9]} . Fewer studies have been made on the quality of the spot welds themselves ^[5,9,10] . Discontinuities such as porosity, cracks and indentation are often found in commercial quality spot welds in aluminium alloys. It is important that such discontinuities do not adversely affect the fracture mode of welds during testing or the properties of welds, particularly fatigue ^[11,12] .

The objectives of this work were to provide information about the effect of excessive porosity and surface indentation, and the effect of weld size on the fatigue performance of spot welds in aluminium alloy sheet. Guidance from the French vehicle manufacturers has been taken for the choice of baseline welding conditions.

Experimental study

Materials

The material studied was the non-heat treatable aluminium alloy 5182-0 in a thickness of 1.2mm. This was supplied by Pechiney Rhenalu in the mill finished, as-received condition and was not cleaned prior to welding. The specified chemical composition and mechanical properties are shown in Table 1. Due to its high mechanical properties, good stamping performance and ease of weldability, it finds application in inner strengthening panels.

Limited comparison was made with 0.8mm zinc coated low carbon steel (XSG) typical of current automotive use. The details are also shown in Table 1.

Table 1. Typical chemical composition and mechanical characteristics of the 5182-0 aluminium alloy and the XSG steel.

Alloy	Condition	Thickness, mm	Chemical composition, weight %						Mechanical Properties
									R 0,2 (MPa)	Rm (MPa)	A% rupture
			Si	Fe	Cu	Mn	Mg	Cr
5182	0	1.2	0.10	0.31	0.025	0.34	4.32	0.025	140	275	28
			C	Mn	Si	P	S	Al
XSG	-	0.8	0.08	0.42	0.01	0.013	0.025	0.02	179	312	42

Equipment and test samples

The welds were made on a 315kVA pedestal machine having a 100kA current capacity and a 15+15kN force capacity with an in line tandem cylinder.

Static tests were conducted on a 200kN Avery Denison tensometer. The shear and cross-tension test samples were made according to the French standards A 87-001 and NF A 89-206, see Fig.1. The fatigue tests were conducted on a 20kN Amsler Vibrophore at a test frequency of approximately 100Hz and a load ratio of R = 0.1 (min/max load during the cycle). A two spot shear test sample was used in this case, as shown in Fig.2.

Experimental procedure

Test welds were made at conditions taken from the French Standard NF 87 001. These conditions were used for the baseline condition, referred to as 'the standard series'. The electrodes were Cu/Cr/Zr with an 11mm tip diameter with a 100mm face radius. A 4kN electrode force and a 3 cycle weld time were used, and current was nominally 26.5kA. This was adjusted to maintain a constant weld size of nominally 6.3mm (5.8 √, where t = sheet thickness in mm) on the cross-tension test samples.

Static shear and cross-tension tests were conducted and the failure mode, weld size and load to failure recorded. Some of the welds were radiographed to illustrate the extent of nugget discontinuities and metallographic sections taken. These were polished to 1µm finish and etched in Kellers reagent.

Fatigue tests were conducted and a log/linear plot produced of test load against endurance, over the range 10 ⁴ to 10 ⁷ cycles. A regression analysis of the 36 standard series tests gave a Wohler curve of the form Log S = a log N + b, including the 90% failure probability limits, according to the French standard A 03-405. (S is the load applied, N is the number of cycles to failure and the constants a and b are derived from the regression analysis).

Welding conditions were modified in the study of discontinuities and weld size, and the details are given in Table 2. The weld series are shown below.

Standard series H	- baseline condition
Series I	- excessive porosity (low electrode force)
Series HP	- nugget centre drilled out 2.5mm diameter (as baseline condition)
Series K	- excessive indentation (truncated cone electrodes)
Series L	- small weld size (low current)
Series M	- large weld size (high current)

Table 2. Welding conditions for the aluminium alloy series and comparison with steel.

	Standard series (H)	Porous series (I)	Drilled out series (HP)	High force series (P)
Electrode force, kN	4	1.5	4	6.5
Weld time, cycles*	3	3	3	3
Welding current, kA mean (standard deviation)	26.5 (0.3)	24.2 (0.4)	26.5 (0.3)	30.2 (0.6)
Electrodes**	Dome tip. 11mm diameter. 100mm radius dome.

	Heavy indent series (K)	Small welds series (L)	Large welds series (M)	Steel series (T)
Electrode force, kN	6	4	4	2.3
Weld time, cycles*	5	3	3	10
Welding current, kA mean (standard deviation)	33.1 (0.6)	21.6 (0.4)	34.9 (0.3)	9.6 (0.09)
Electrodes**	Truncated cone 120°. 6.5mm diameter. 100mm radius dome.	Dome tip. 11mm diameter. 100mm radius dome.		Cnomo (G) 30° 6mm diameter tip. 40mm radius dome.

* 1 cycle = 0.02 sec at 50Hz
** Electrode material - Cu/Cr/Zr
Hold time 30 cycles.

Weld testing was conducted in a similar way to the standard series and the results compared. In addition, welds were made in the steel for comparison with the standard series in the aluminium alloy, see Table 2.

Results

Effect of weld discontinuities

Standard series welds showed satisfactory nugget penetration with some lack of symmetry. There was substantial porosity in the nugget although the periphery was clear as shown in Fig.3a. Increased porosity was present in the low electrode force welds with a central pore usually 2 to 3mm diameter and often associated with cracks ( Fig.3b). The deeply indented spot welds (series K) had 40 to 50% indentation at the centre and greater than 1.2mm sheet separation 10mm from the weld edge. However, there was no porosity in the narrow weld nugget ( Fig.3c).

a) Standard series (H) - weld H13 (26.6kA)
b) Excessive porosity (I) - weld I44 (24.3kA)
c) Heavy indentation (K) - weld K77 (32.7kA)

The results of the static tests for the standard series (H) and series I, HP and K are given in Table 3. All the cross tension tested welds gave a button/plug failure, although the welds with excessive porosity or deep indentation were significantly weaker than the standard series. The shear test samples were of a similar strength except for the drilled out welds, and interface failures occurred in all but the deeply indented welds. In many cases the normal scatter of results exceeded the effect of the discontinuities or of welding conditions.

Table 3. Static test results in the study of discontinuities.

	Standard series (H)	Excessive porosity (I)	I/H	Drilled spots (HP)	HP/H	Deep indentation (K)	K/H	High Force (P)	P/H
CT failure mode	Button/plug	Button/plug		Button/plug		Button/plug		Button/plug
CT failure force, kN	2.30 (0.15)	1.82 (0.22)	79%	2.46 (0.06)	107%	1.52 (0.18)	66%	2.32 (0.07)	101%
Weld diameter, mm	6.35 (0.13)	6.18 (0.10)	97%	6.46 (0.15)	102%	6.21 (0.28)	98%	6.47 (0.31)	102%
d/ √t	5.8	5.6	97%	5.9	102%	5.7	98%	5.9	102%
TS failure mode	Interface	Interface	-	Interface	-	Button/plug	-	Interface
TS failure force, kN	3.93 (0.45)	4.15 (0.22)	106%	3.61 (0.22)	92%	4.17 (0.36)	106%	4.09 (0.46)	104%
Interface splash	sometimes slight	always all around	-	sometimes slight	-	always heavy	-	sometimes slight

CT - cross-tension
TS -tensile-shear
Failure forces shown as: mean (standard deviation)

The fatigue test results are shown in Figs 4 to 6 for comparison of the effect of discontinuities with the standard series. The welds with excessive porosity failed at a slightly lower load than the standard series (1.1kN compared to 1.3kN at 10 ⁶ cycles) but within the scatter band showing 90% chance of failure ( Fig.4). In addition, the failure mode was similar in each case, with a crescent shaped crack growing through the thickness from the edge of the nugget. The results for the drilled out samples gave the same fatigue performance as the standard series ( Fig.5).

As a means of checking the effect of force on fatigue properties, some untested low force welds were pressed cold between the welding electrodes at 4kN. This treatment improved the fatigue properties as the test results for these samples (e.g. 1.5kN at 10 ⁶ cycles) were within the scatter band for the standard series. Furthermore, additional welds made at an even higher force of 6.5kN, using domed electrodes, gave a higher load than the standard series at 10 ⁶ cycles of 1.5kN, on the upper limit of the scatter band ( Fig 4).

The deeply indented welds ( Fig 6) failed at a substantially higher load (1.9kN at 10 ⁶ cycles) than the standard series (1.3kN). In addition, the fatigue cracks started in the parent material 2 to 5 mm outside the notch at the interface.

Effect of Weld Size

Series L and M were welded at the standard conditions but with the welding current adjusted to give weld diameters in the target ranges of 4.5 to 5mm and 7.5 to 8mm.

The radiographs and metallographic sections ( Fig 7) showed that weld splash and porosity in the nugget increased with weld size. However, the periphery of the nugget was clear in each case and it was shown above that the porosity within the nugget had little effect on the mechanical test results. Thus, these tests gave a true comparison of the effect of weld size.

a) Small weld, (L) - weld L7, 21.3kA, 4.3mm diameter.
b) Standard series, (H) - weld H39, 26.7kA, 6.35mm diameter.
c) Large weld, (M) - weld M63, 34.8kA, 7.9mm diameter

Table 4 summarises the static tests in comparison with the standard series. As expected, the static strength was highly dependent on weld size. Increasing weld size from the approximately minimum acceptable 4.2 √t to the large 7.2 √t gave a 2.4 times increase in shear failure load and a 1.6 times increase in cross-tension failure load. The cross-tension tests failed by button/plug failure except for some of the small welds, which broke around the nugget rather than through the sheet thickness. In shear, the largest welds failed by forming a button/plug, whereas the standard and small welds failed across the interface.

Table 4. Static test results in the study of weld size and comparison with steel.

	Small welds (L)	Standard series (H)	Large welds (M)	Steel series (T)
CT failure mode	Button/plug	Button/plug	Button/plug	Button/plug
CT failure force, kN	1.69 (0.29)	2.30 (0.15)	2.68 (0.16)	3.31 (0.30)
Weld diameter, mm	4.63 (0.29)	6.35 (0.13)	7.93 (0.10)	5.21 (0.27)
d/ √t	4.2	5.8	7.2	5.8
TS failure mode	Interface	Interface	Button/plug	Button/plug
TS failure force, kN	2.33 (0.46)	3.93 (0.45)	5.55 (0.10)	4.32 (0.16)
Interface splash	none	sometimes slight	slight or heavy	none

CT - cross-tension
TS - tensile-shear
Failure forces shown as: mean (standard deviation)

The fatigue results showed less difference between the weld sizes than did the static results ( Fig 8 and 9). At low endurance, the small welds withstood a lower fatigue load than the standard series scatter band. At these conditions, the L-N curve for the large welds was slightly better than the standard series, but remained within the scatter band. In all except the small welds, which failed across the interface, classical failure occurred. This comprised a crescent shaped crack, starting from the edge of the weld or bonded zone and propagating through the material thickness. At high cycle/low load conditions, there was no significant difference between the weld sizes. All of these welds showed the classic failure mode, described above.

The results for the large welds were also very similar to the standard series with individual values lying within the scatter band over the whole test range. All of these welds showed the classic failure mode, with the crescent shaped crack starting from the edge of the weld or bonded zone and propagating through the thickness.

Comparison of steel and aluminium alloy

In order to provide a representative comparison between the aluminium alloy and steel, on the basis of sheets with similar stiffness, a 0.8mm, zinc coated steel was chosen. The 1.2mm aluminium alloy was still 50% lighter weight than the steel, despite being 50% thicker. The 5.2mm weld size in the steel was the same proportion of the sheet thickness as for the aluminium, (nominally 5.8 √t). Similar static and fatigue tests were conducted as for the aluminium alloy and the results are shown in Table 4 and Fig.10.

Although the steel sheet was double the weight of the aluminium sheet, the static shear strength of the steel spot welds was only about 10% higher than that of the aluminium spot welds.

The benefit of the aluminium alloy was less pronounced when considering fatigue performance. At 10 ⁶ cycles, the fatigue load for the two spot aluminium alloy samples was 1.3kN, compared to 2.2kN for the steel spot welds. However, load distribution, and joint stiffness can influence the fatigue properties of a joint. Thus, the actual load per spot is dependent on joint design and material thickness in a structural component.

Discussion

Significance of discontinuities

Two categories of discontinuity occur in resistance spot welding of aluminium. Those caused by incorrect choice of welding conditions or machine set-up include stuck welds, surface splash, deep surface indentation and sheet separation, but are easily avoided. Certain discontinuities are intrinsic in aluminium alloy spot welds, particularly shrinkage porosity and cracking in the weld nugget. If these become severe, surface cracking or a shrinkage pipe can form in the centre of the electrode indentation. Although not a discontinuity as such, weld splash is common. It was intended that the extreme conditions assessed in this work represented those beyond normal quality standards in production.

The excessive porosity and drilled out welds indicated that cavities up to 40% of the weld diameter had no significant effect on joint performance. The welds produced with excessive indentation of 40 to 50% showed an actual increase in fatigue performance compared to the standard welds. However, static cross tension strength was lower than standard welds as the parent material was thinned at the edge of the weld.

Influence of Electrode Force

Although, fatigue properties of welds were fairly insensitive to the discontinuities studied, electrode force appeared to play a more important role. It would appear that higher electrode force, or a post weld squeeze, improved the fatigue performance. This may be because the higher forces modify either the residual stress at the edge of the weld, where the fatigue cracks initiate, or reduce the sharpness and stress concentration at the notch. The radial residual stress in this zone, due to the cooling of the weld, was shown to be tensile and close to the material yield stress to aluminium ^[13] . Furthermore, mechanical treatment of welds ^[13] was shown to increase fatigue life tenfold, although thermal treatment gave no improvement. Post-weld compressive loading of spot welds in steel is a recognised means of improving fatigue performance by introducing compressive stresses at the edge of the weld.

The bonded zone is not normally considered as contributing strength to welds in aluminium alloys and the cross-tension tested welds showed failure at the edge of the fused zone. However, at low load fatigue conditions, there is sometimes sufficient strength to promote cracking from the outer edge of this zone.

Conclusions

The static and fatigue properties of resistance spot welds in 1.2mm 5182-0 aluminium alloy have been studied to establish the effect of weld discontinuities and weld size, and to compare with 0.8mm zinc coated low carbon steel. The following conclusions are drawn.

• Excessive porosity, up to about 40% of the nugget diameter did not affect the static or fatigue performance of the welds in shear when maintaining a constant 6.3mm weld diameter.
• The location of the fatigue cracks were not affected by the porosity or drilled hole.
• Increasing the weld diameter from 4.2 to 7.2mm, produced a large increase in the static properties of welds. However, in fatigue, the weld size had only a small positive effect for high load/low endurance conditions and no effect at all for low load/high endurance (10 ⁶ cycles).
• Electrode force had one of the most significant effects on fatigue strength. The fatigue load at 10 ⁶ cycles was 15% higher when electrode force was increased from 4.0kN to 6.5kN and 15% lower at 1.5kN electrode force. A post weld squeeze at 4kN largely restored the fatigue properties of the low force welds.
• Deep surface indentation increased the fatigue strength of welds of the standard size probably due to the higher electrode forces used which modified conditions at the edge of the nugget.
• When showing a 50% weight saving in comparison to steel, the fatigue strength of spot welds in aluminium was lower than the equivalent welds in steel.

References

1	Patrick, E. P. and Sharp, M. L. 1992.	Joining Aluminium auto body structure. SAE Paper 920282.
2	Kucza, J. C., Butruille, J. R. Hank, E. and Lancrenon, B. 1997.	Aluminium as rolled sheet for automotive applications - Effect of surface oxide on resistance spot welding and adhesive bonding behaviour. SAE Paper 970013.
3	Hoch, F. R. 1978.	Joining of Aluminium alloys 6009/6010. SAE Paper 780396.
4	Krause, A. R., Thornton, P. H. and Davies, R. G. 1994.	Effect of magnesium content on the fatigue of spot-welded aluminium alloys. Proceedings of the international symposium of recent developments in light metals, Toronto, Ontario, August 20-25, pp 31-49.
5	Auhl, J.R. and Patrick, E. P. 1994.	A fresh look at resistance spot welding of aluminium alloy components. SAE Paper 940160.
6	Rivett, R. M. and Westgate, S. A. 1980	Resistance welding of aluminium alloys in mass production. Metal construction, 12(10), pp 510-517.
7	Leone, G. L. and Altshuller, B. 1984.	Improvement on the resistance spot weldability of aluminium body sheet. SAE Paper 840292.
8	Pickering, E. R. and Hart, C. J. 1994.	Optimizing resistance spot welding on aluminium-alloy 6111 autobody sheet. SAE Paper 940662.
9	Thornton, P. H., Krause, A. R. and Davies, R.G. 1996.	The aluminium spot weld. Welding Research Supplement, 75(3), pp 101s-108s.
10	Watanabe, G. and Tachikawa, H. 1995.	Behaviour of cracking formed in aluminium alloy sheets on spot welding. Toyota Central R&D LABS, Inc, IIW Doc No. III-1041-95, Stockholm,
11	Nordmark, E. G. 1978.	Fatigue performance of aluminium joints for automotive applications. SAE Paper 780397.
12	Krause, A. R., Thornton, P. H. and Davies, R. G. 1993	The fatigue of spot-welded aluminium alloys. Light Metals Processing and Applications, pp 589-600', International Symposium Light Metals, 32nd Annual Conference Metallurgists, The Metallurgical Society of CIM, Quebec City, Canada.
13	Steffens, H. D. and Kern, H. 1985.	Influence of residual stresses and microstructure on the lifetime of resistance spot welded 2024 aluminium. The effects of Fabrication Related Stresses on Product Manufacture and performance. Proceedings, International Conference, Cambridge, UK. Paper 50 pp 269-283.