Welding Defects Occurrence and Their Effects on Weld ...

1883 © 2014 ISIJ

ISIJ International, Vol. 54 (2014), No. 8, pp. 1883–1889

Welding Defects Occurrence and Their Effects on Weld Quality in Resistance Spot Welding of AHSS Steel

Xiaodong WAN,1)* Yuanxun WANG1,2) and Cuixia FANG1)

1) Department of Mechanics, Huazhong University of Science and Technology, Wuhan 430074, China.2) Hubei Key Laboratory for Engineering Structural Analysis and Safety Assessment, Luoyu Road 1037, Wuhan 430074, China.

(Received on March 12, 2014; accepted on May 11, 2014)

An investigation was made on the weldability of DP600 steel. Influences of electrode force, weldingcurrent and welding time on welding defects occurrence were discussed. Metallographic analysis andtensile-shear tests were implemented for weld quality detections. Weld quality dependency on weldingdefects was studied. It was found that welding defects were easy to occur during spot welding processof DP600 steel, which could be explained by its rich chemistry. Lower electrode force, larger welding cur-rent and longer welding time all contributed to the easy occurrence of welding defects. HAZ softeningcaused by martensite tempering was observed under strong welding parameters. For interfacial failure,welding defects effects on weld quality should not be underestimated. As to pullout failure, the effect ofexpulsion on weld quality was not that significant as suggested by other researches. In the end, a methodwas introduced to predict peak load and maximum tensile-shear displacement for pullout failure.

KEY WORDS: dual phase steel; resistance spot welding; weld quality; expulsion; shrinkage void; solidificationcrack.

1. Introduction

For resistance spot welding, the electrode force is firstapplied on two or more overlapped workpieces, then jouleheat is generated by the passage of a large welding currentin a very short period of time, base metal is finally melt andjoints are formed after the cooling stage. Although manynew welding procedures are introduced, such as laser beamwelding and adhesive bonding, resistance spot weldingremains to be the most widely used sheet metal welding pro-cess in the automotive industry owing to its low cost, highefficiency and easy to operate.

The increasingly demands for vehicle safety and energyconservation are drawing more and more attentions on non-traditional materials like advanced high strength steels(AHSSs).1) Due to the good combination of strength andductility, dual phase steels including DP600 are more pre-ferred by vehicle manufacturing companies. However, thepoor weldability of dual phase steels has been an importantissue to restrict its wide application. Welding parameterslike electrode force, welding current and welding time allcan have significant effects on weld quality of dual phasesteels. High alloy elements level in DP600 makes expulsionand interfacial failure more frequent for spot welds. The tra-ditional failure criterion of does not work forDP600 resistance spot welding any more.2) However, limit-ed work has been done to investigate the effect of weldingdefects on weld quality of spot welded DP600 steels.

The presence of welding defects in weld nugget has beenreported in some investigations.3–5) According to these stud-ies, shrinkage void and solidification crack were two of themain reasons for interfacial failure of spot welds undertensile-shear and fatigue tests. Voids formed by expulsiongenerally decreased the load bearing capacity of spot weldsand interfacial failure occurred more often. Khan et al.6)

found that interfacial failure would occur for DP600 resis-tance spot welds although the nugget size exceeded the crit-ical value, which was attributed to shrinkage voids andsolidification cracks formation. Sun et al.7,8) studied themicrostructure of resistance spot welded magnesium alloyand suggested that a portion of welding cracks belonged tosolidification cracking. An incrementally coupled finiteelement procedure was then developed to simulate the spotwelding process of aluminum alloys. Tensile residual stressat the weld nugget was believed to promote through-thicknesscracking. Marya et al.9) showed that solidification crackscan adversely affect the overall bonded area in DP600 spotwelds. Pouranvari et al.2,10) suggested that the high tendencyto shrinkage void formation during AHSS resistance spotwelding was one of the main reasons for its high suscepti-bility to interfacial failure. An analytical model was thenproposed to predict the critical nugget size:

where P referred to the porosity factor induced by shrinkagevoids.

Considerable researches have been made on how toreduce welding defects occurrence. Marya et al.3) investigated

* Corresponding author: E-mail: [email protected]: http://dx.doi.org/10.2355/isijinternational.54.1883

D tcr = 4D

t

Pf

H

HcrPFL

FZ

=⎛

⎝⎜

⎞

⎠⎟

4

© 2014 ISIJ 1884

ISIJ International, Vol. 54 (2014), No. 8

the weldability of dual phase steels with different thickness.They found that shrinkage voids were less pronounced inthinner DP steels. Joaquin et al.4) and Ma et al.5) showed thatlonger holding time was helpful to reduce shrinkage voids.Lang et al.11) studied effects of welding parameters on cracksusceptibility. They tried to understand the crack formationmechanism and make improvements on weldability of mag-nesium alloys. Milititsky et al.12) conducted tensile-sheartests on DP600 spot welds and found that short holding timemay promote shrinkage voids and solidification cracks for-mation. Zhang et al.13) presented an weld expulsion estima-tion method in resistance spot welding of dual phase steelthrough changing the electrode force.

The objective of this study is to investigate the weldabil-ity of DP600 steel. Welding defects, such as expulsion,shrinkage void and solidification crack, were detected byexperiments. Influences of electrode force, welding currentas well as welding time on welding defects occurrence werediscussed. In the end, a preliminary analysis was made oncorrelation between welding defects and mechanical perfor-mance of spot welds.

2. Experimental Procedure

In this study, a 1.7 mm thick DP600 steel was used for resis-tance spot welding. Its chemical composition was measuredand given in Table 1. Detailed mechanical property of this steelsheet is shown in Table 2. Before welding, the steel sheets werecut into dimensions of 120 × 40 mm and all the specimens sur-faces were cleaned carefully by acetone. The schematicdescription of the overlapped steel is shown in Fig. 1.

Spot welding was conducted using a 120 kVA, PLC con-trolled 50 Hz AC spot welding machine (1 cycle = 0.02 s).A 450 truncated cone electrode made of copper alloy with aface diameter of 6 mm was used. Welding was repeated forsix times at each welding condition, including five speci-mens prepared for static tensile-shear test and one specimenprepared for metallographic investigation. During the weld-ing process, both the squeeze time and holding time were

kept constant at 10 cycles (0.2 s). To study the effect of elec-trode force, welding current and welding time on joint per-formance, specimens were divided into three groups. Ineach group, only one parameter was varied and all the otherparameters were kept constant. Electrode force was changedfrom 2.0 kN to 4.0 kN. Welding current, which referred tothe peak current value, increased from 6 kA to 12 kA. Weld-ing time was varied from 8 to 17 cycles at an interval of 3cycles. A detailed description of the welding schedules uti-lized in this study is shown in Table 3.

An Instron universal testing machine was used to conductthe tensile-shear tests at a constant crosshead speed of 10mm/min. Tabs were glued to the specimen ends to reducejoints deformation. Peak load was measured at the top posi-tion of the load-displacement curve. Failure energy absorp-tion was considered as the area under the load-displacementcurve up to failure:14)

Where F is the force, x is the displacement, the product ofF and x are summed for each data set (i, N) up to the peakload. By examining the failed specimen, failure modes wereidentified.

For metallographic analysis, specimens were cross-sectioned, ground and polished with carbide papers andetched using a 4% nital reagent until the martensite grainboundaries can be observed under the microscope. Figure 2

Table 1. The chemical composition of DP600 steel, wt-%.

C Si Mn P S Al N

0.0790 1.0000 1.5200 0.0150 0.0049 0.0230 0.0037

Table 2. Mechanical property of DP600 steel.

Steel Yield Strength(MPa)

Ultimate TensileStrength (MPa)

Elongation(%)

DP600 351 633 28

Table 3. Welding schedules in this study.

Squeeze Time (Cycle) 10Holding Time (Cycle) 10

Electrode Force (kN) 2.0–2.5–3.5–4.0Welding Current (kA) 6–8–10–12Welding Time (Cycle) 8–11–14–17

Fig. 2. A prepared metallographic specimen.

Q F i x i x ii

N

= − −=∑ ( )[ ( ) ( )]1

1

Fig. 1. Schematic illustration of the overlapped steels for resistance spot welding.

ISIJ International, Vol. 54 (2014), No. 8

1885 © 2014 ISIJ

shows a prepared metallographic specimen. The microstruc-ture of weld joints was detected by an Zeiss Axiovert 200MAT Inverted microscope. Nugget size was measured bythe Carl Zeiss AxioVision Software on cross-sectioned spotwelds. Microhardness testing was carried out using a HX-100 test machine (100 g load) at an interval of 0.5 mm.

3. Results and Discussion

3.1. Microstructure and MicrohardnessA typical microstructure distribution of resistance spot

welded DP600 steel is shown in Fig. 3. Microstructure inthe fusion zone shows directional and columnar grainsgrowing from the fusion boundary towards the weld center-line, as depicted in Fig. 3(a). In Fig. 3(b), the fusion zone ispredominantly composed of large lath martensite, whichcould be explained by the high cooling rate. Grain size isless coarser in the HAZ (Fig. 3(c) and left area of Fig.3(d))when comparing to that of the fusion zone, which is due toan incompletion of austenitizing.3) The base metal, as seenin Fig. 3(d), consists mainly of evenly distributed martensiteand ferrite.

Figure 4 shows the microhardness profiles across theweld nugget under two different welding conditions. As canbe seen, the fusion zone hardness is roughly around 360 HV,owing to the lath martensite formation. Fusion zone hardnessis approximately 1.5 times higher than that in the base metal.When welding time is 8 cycles, maximum hardness is foundin the heat affected zone (HAZ) closing to the weld nugget.A softening region in the HAZ can be observed under a largewelding time (11 cycles), which is attributed to martensitetempering. What’s more, mechanical performance of spotwelds is considered to be improved by HAZ softening.15)

3.2. Welding Defects3.2.1. Expulsion

Welding expulsion at the faying surface for a DP600 spot

weld is shown in Figs. 5 and 6. Expulsion, which refers tothe ejection of molten metal, is a common phenomenon dur-ing resistance spot welding process. It usually occurs ateither the electrode/workpiece interface or the faying sur-face owing to excessive heat input. Due to the presence ofexpulsion, excessive electrode indentation, shrinkage voidsand solidification cracks will be introduced into the weldnugget. Expulsion at the faying surface may severely affectthe weld quality.15)

Prediction and control of expulsion are important issuesin the manufacturing environment. Long welding time andlarge welding current are common choices for dual phasesteel spot welding with consideration of its poor weldability,which make expulsion even more popular. In general, plas-tic annulus formed by larger plastic strain around the weldnugget is helpful to prevent expulsion.16) What’s more,appropriate welding parameters, clean workpiece surfaceand elimination of electrode wear are all necessary to reduceexpulsion. The effect of expulsion on tensile-shear failureload and failure energy absorption of spot welds are dis-

Fig. 3. Microstructure variation across a DP600 spot weld.

Fig. 4. Two typical microhardness profiles across DP600 spot welds.

© 2014 ISIJ 1886

ISIJ International, Vol. 54 (2014), No. 8

cussed in the following sections.

3.2.2. Shrinkage VoidShrinkage voids are found under some welding condi-

tions, as shown in Figs. 7 and 8. Low magnification micro-scopic images are depicted in Figs. 7(a) and 8(a). High mag-nification microscopic images are shown in Figs. 7(b) and8(b). Heat input immediately reduces to zero on weldingcurrent is cut off. Temperature at the weld joint decreasesrapidly under the cooling effect of water cooled electrodes.Liquid metal starts to solidify when temperature drops to theliquidus temperature. Finally, cooling rate mismatch at dif-ferent locations results in the formation of shrinkage voids.

Effective area of weld nugget is reduced and local shearstress is increased under tensile-shear tests when shrinkagevoids form at the faying surface.17) Interfacial failure ten-dency, which means an unsatisfactory load bearing capacity,

is believed to increase by shrinkage voids and the criticalnugget size gets bigger. However, when pullout failureoccurs, the mechanical performance will not be affected byshrinkage voids. In this case, the weld nugget and heateffected zone are nearly stress free, which was confirmed byfinite element method and experimental observations.18,19)

Shrinkage voids formation and solidification cracks arestrongly dependent on carbon, alloy elements and theamount of S and P. A higher carbon equivalent results inmore shrinkage voids in spot welds.4) Shrinkage voids tend

Fig. 5. Expulsion occurrence at the faying surface.Fig. 6. Expulsion occurrence at the faying surface (electrode force =

3.5 kN, welding current =12 kA, welding time = 14 Cycles).

Fig. 7. Shrinkage void at weld nugget (electrode force = 3.5 kN, welding current =12 kA, welding time = 14 Cycles).

Fig. 8. Shrinkage void at weld nugget (electrode force = 3.5 kN, welding current =10 kA, welding time = 14 Cycles).

ISIJ International, Vol. 54 (2014), No. 8

1887 © 2014 ISIJ

to occur during resistance spot welding of dual phase steel,owing to the inherent rich chemistry as compared to lowcarbon steels.3) In addition, if an internal expulsion is gen-erated during the welding process, shrinkage voids are eas-ily formed at the weld nugget center for the reason of liquidmetal loss. To reduce shrinkage voids occurrence, greaterelectrode force can be adopted to while controlling otherwelding parameters to restrict the heat input rate. As can beseen, the electrode force is relative high in Figs. 7 and 8.However, shrinkage voids are generated due to the highwelding current and long welding time, leading to too muchheat input and expulsion is also observed.

3.2.3. Solidification CrackFigures 9 and 10 clearly show the solidification cracks.

A rich chemistry in dual phase steels is also contributed tosolidification crack occurrence. Internal solidificationcracks are easy to form and often companied with shrinkagevoids. Increased local shear stress and decreased weld qual-ity are results of solidification cracks. Most of the cracks areobserved to be perpendicular to the faying surface and willbe explained below.

During the spot welding process, thermal expansion andplastic deformation occur sequentially in weld joints as aresult of heat input by welding current passage. Inhomoge-neous deformation results in nonuniform stress and strainfields. Under the combined effects of electrode force andheated base metal, stress in the weld nugget is mostly com-pressive. However, temperature decreases rapidly during thecooling stage and stress in the weld nugget translates grad-ually to tensile state. Contraction rate is the largest in thedirection parallel to faying surface. What’s more, the mar-tensite grain boundary direction is mostly perpendicular tothe faying surface. That’s why solidification cracks is proneto occur in through-thickness direction and most of them aregrain boundary cracks.

3.3. Welding Defects OccurrenceExpulsion and shrinkage void (or solidification crack)

occurrence under various welding conditions are shown inFig. 11. Resistance spot welding under the same weldingparameters was repeated for five times. Expulsion waschecked after the completion of tensile-shear tests. Shrinkagevoid occurrence was indirectly determined by the significantdrop in maximum tensile-shear displacement. If no weldingdefects were found, the frequency is replaced by the valueof 2% for a better visualization. As it is shown in Figs.11(a)–11(c) and indicated in previous sections, expulsionand shrinkage void always occur simultaneously. A lowerelectrode force, larger welding current and longer weldingtime all make contribution to the easy occurrence of expul-sion and shrinkage void.

3.4. Effect of Expulsion on Weld QualityTensile-shear tests were conducted for each group with

the aforementioned five spot welded specimens under thesame welding condition. Both interfacial and pullout failuremode were observed during the tensile-shear tests. Partialinterfacial failure was categorised as interfacial failure,depending on its failure performance. After an eliminationof the welding parameters under which only interfacial fail-ure during tensile-shear testing was taken place, a detailedanalysis of the influence of expulsion on weld quality isdepicted in Fig. 12. Sample number is ordered by failuremode (from interfacial failure to pullout failure) in Fig. 12.For all spot welded specimens listed in Fig. 12, only thespecimen number 2–5 in Fig. 12(a) and specimen number 3–4 in Fig. 12(b) are expulsion free, all the rest experienced aslight or excessive expulsion. In addition, among the fourdifferent welding conditions, HAZ softening was notobserved only when welding time is 8 cycles. Peak load,maximum displacement were recorded and the ratiobetween them was also plotted.

Generally, plastic collapse is considered to be the failure

Fig. 9. Solidification crack at the transition region from HAZ tobase metal (electrode force = 3.5 kN, welding current =10kA, welding time = 17 Cycles).

Fig. 10. A magnified views of solidification crack at weld nugget(electrode force = 3.5 kN, welding current = 10 kA, weld-ing time = 14 Cycles).

© 2014 ISIJ 1888

ISIJ International, Vol. 54 (2014), No. 8

mechanism for pullout failure. In this case, failure initiatesaround the weld nugget from the location where hardness isminimum.20) Moreover, it seems that a proportional relation-ship exists between peak load (Fmax) and maximumdisplacement (Lmax) under pullout failure, as can be seen inFig. 12. This conclusion is further confirmed by an almosthorizontal green line on behalf of the Fmax/Lmax in Fig. 12.That is to say:

................................ (1)

In addition, different failure modes indicate different weldqualities at the same welding condition. Peak load and max-imum displacement at interfacial failure is typically lowerthan those of pullout failure. However, despite the expulsion

occurrence, experimental data varies slightly under the pull-out failure mode, as shown in Figs. 12(a)–12(d). Thus, it canbe inferred that the influence of expulsion occurrence on

Fig. 11. Expulsion and shrinkage void occurrence under variouswelding conditions.

F Lmax max∝ Fig. 12. Effect of expulsion on weld quality.

ISIJ International, Vol. 54 (2014), No. 8

1889 © 2014 ISIJ

weld quality is not that significant for pullout failure, whichis probably different from some researches.6,21)

In addition, Lmax reflects the tensile strength of the initialfailure location and nugget size d is also an key factoraffecting Fmax. The normalized Lmax/d can then be used torepresent the ductility of initial failure location. Moreover,the failure location hardness Hfl is inversely proportion to itsductility. Thus,

............................. (2)

can be established. Pullout failure usually initiates from theminimum hardness position, that is to say Hfl = Hmin.10) ωcan be calculated by the average experimental data obtained,as shown in Table 4. For occasions without HAZ softening,the average value of ω is 173.8 HV (No. A). When HAZsoftening occurs, ω is approximately equal to 144.3 HV(No. B–D). Once ω is determined, Lmax and Fmax under pull-out failure could then be predicted by Eqs. (1) and (2). Forfurther study, a critical nugget size criterion may be con-structed in combination with additional interfacial failureinvestigation.

4. Conclusions

Electrode force, welding current and welding time aremain factors affecting the weld qualities of DP600 spotwelds. Microstructure and mechanical performance of spotwelds with welding defects were investigated. Although thetest results here are limited, following conclusions can stillbe derived:

• Welding defects are easy to occur during resistancespot welding of DP600 steel sheets due to its richchemistry.

• Lower electrode force, larger welding current and lon-ger welding time all contributed to the easier occur-rence of welding defects.

• HAZ softening caused by martensite tempering isobserved under strong welding parameters.

• For interfacial failure, welding defects effects on weldquality should not be underestimated.

• As to pull out failure, the influence of expulsion occur-rence on tensile-shear mechanical performance is notthat significant.

• A parameter to is introduced to predict peak load andmaximum tensile-shear displacement for pullout fail-ure occasion.

AcknowledgementsWe would like to acknowledge the financial support from

the National Natural Science Foundation of China(11072083). The sponsorship of the Graduates InnovationFund of Huazhong University of Science & Technology(HF-11-14-2013) and the Fundamental Research Funds forthe Central Universities, HUST (CXY13Q047) are alsoappreciated.

REFERENCES

1) ULSAB-AVC Consortium: Technical Transfer Dispatch # 6 (BodyStructure Materials), American Iron and Steel Institute, Southfield,MI, (2001).

2) M. Pouranvari, S. P. H. Marashi and D. S. Safanama: Mater. Sci. Eng.A, 528 (2011), 8344.

3) M. Marya and X. Q. Gayden: Weld. J., 84 (2005), 172.4) A. Joaquin, A. N. A. Elliott and C. Jiang: Weld. J., 86 (2007), 24.5) C. Ma, D. L. Chen, S. D. Bhole, G. Boudreau, A. Lee and E. Biro:

Mater. Sci. Eng. A, 485 (2008), 334.6) M. I. Khan, M. L. Kuntz, P. Su, A. Gerlich, T. North and Y. Zhou:

Sci. Technol. Weld. Join., 12 (2007), 175.7) D. Q. Sun, B. Lang, D. X. Sun and J. B. Li: Mater. Sci. Eng. A, 460

(2007), 494.8) X. Sun and P. Dong: Weld. J., 79 (2000), 215.9) M. Marya, K. Wang, L. G. Hector and X. H. Gayden: J. Manuf. Sci.

Eng., 128 (2006), 287.10) M. Pouranvari and S. P. H. Marashi: Mater. Sci. Eng. A, 528 (2011),

8337.11) B. Lang, D. Q. Sun, Z. Z. Xuan and X. F. Qin: ISIJ Int., 48 (2008),

77.12) M. Milititsky, E. Pakalnins, C. Jiang and A. K. Thompson: On

Characteristics of DP600 Resistance Spot Welds, SAE Report 2003-01-0520, SEA International, Warrendale, PA, USA, (2003).

13) Y. Zhang, J. Shen and X. Lai: ISIJ Int., 52 (2012), No. 3, 493.14) M. I. Khan, M. L. Kuntz and Y. Zhou: Sci. Technol. Weld. Join., 13

(2008), 294.15) H. Zhang and J. Senkara: Resistance Welding: Fundamentals and

Applications, 2nd ed., Taylor & Francis, Boca Raton, FL, (2011).16) Z. Hou, I.-S. Kim, Y. Wang, C. Li and C. Chen: J. Mater. Process.

Technol., 185 (2007), 160.17) X. Sun, E. V. Stephens and M. A. Khaleel: Weld. J., 86 (2007), 18.18) X. Deng, W. Chen and G. Shi: Finite Elem. Anal. Des., 35 (2000), 17.19) M. Pouranvari and S. P. H. Marashi: Mater. Des., 31 (2010), 3647.20) Y. J. Chao: Sci. Technol. Weld. Join., 8 (2003), 133.21) M. Pouranvari, A. Abedi, P. Marashi and M. Goodarzi: Sci. Technol.

Weld. Join., 13 (2008), 39.

Table 4. Welding parameters and average experimental data forcalculating ω (Electrode force = 3.5 kN).

No.(–)

Weldingcurrent (kA)

Welding time(Cycle)

D(mm)

Lmax(mm)

Hfl (HV)

ω(HV)

A 10 8 6.42 4.65 240 173.8

B 10 14 7.08 5.57 185.1 145.6

C 12 14 7.37 6.24 170.5 144.4

D 10 17 6.71 6.42 149.2 142.8

L d H flmax / = −ω 1