Pruebas de cizalladura RSW

WELDING RESEARCH

APRIL 2008, VOL. 87-s96

ABSTRACT. Finite element modelingand fracture mechanics calculations wereused to predict the resistance spot weldfailure mode and loads in shear-tensiontests of advanced high-strength steels(AHSS). The results were compared tothose obtained for an interstitial-free (IF)steel. The results of the work confirmedthe existence of a competition betweentwo different types of failure modes,namely full button pullout and interfacialfracture. The force required to cause acomplete weld button pullout-type failurewas found to be proportional to the tensilestrength and the thickness of the base ma-terial as well as the diameter of the weld.The force to cause an interfacial weld frac-ture was related to the fracture toughnessof the weld, sheet thickness, and weld di-ameter. For high-strength steels, it was de-termined that there is a critical sheetthickness above which the expected fail-ure mode could transition from pullout tointerfacial fracture. In this analysis it wasshown that, as the strength of the steel in-creases, the fracture toughness of the weldrequired to avoid interfacial failure mustalso increase. Therefore, despite higherload-carrying capacity, due to their highhardness, the welds in high-strength steelsmay be prone to interfacial fractures. Ten-sile testing showed that the load-carryingcapacity of the samples that failed via in-terfacial fracture was found to be morethan 90% of the load associated with a fullbutton pullout. This indicates that theload-bearing capacity of the welds is notaffected by the fracture mode. Therefore,the mode of failure should not be the onlycriteria used to judge the quality of spotwelds. The load-bearing capacity of the

weld should be the primary focus in theevaluation of the shear-tension test resultsin AHSS.

Introduction

The use of advanced high-strengthsteels (AHSS), such as dual-phase andtransformation-induced plasticity (TRIP)steels, has been steadily increasing overthe past few years in automotive applica-tions (Refs. 1, 2). This is due to the advan-tages that AHSS grades offer, in terms ofhigher strength that enables the automak-ers to decrease the vehicle weight for im-proved fuel economy and improved crashenergy absorption for better occupantprotection. The two grades of AHSS thathave seen increased use in automobilesare the dual-phase and TRIP steels. Thesteel grades that are used commercially inautomotive bodies at present are thosewith minimum strength levels of 500, 590,and 780 MPa. Resistance welding is thepredominant mode of fabrication in auto-motive production with a typical vehicle inNorth America containing about 4000 to5000 welds. Therefore, good resistancespot welding behavior is one of the keycharacteristics of any steel grade to beconsidered for use in automobile bodyproduction.

Several tests are generally used to char-acterize the resistance spot welding be-havior of steels. These include the weldingcurrent range determination, metallo-graphic characterization of the weld and

heat-affected zone microstructures, mi-crohardness, and weld tensile tests (Ref.3). One type of weld tension test typicallydone is called the shear-tension test(sometimes referred to as lap-shear test).In this test, two sheet samples, 140 mmlong by 60 mm wide are overlapped by 45mm and joined with a single spot weld lo-cated at the center of the overlapped re-gion — Fig. 1. The sample is then pulledin tension. Due to the offset of the sheets,the application of tension creates a bend-ing moment that causes rotation of theweld nugget. This type of deformation isdemonstrated with a finite element simu-lation — Fig. 2. For clarity, only one halfof the model is shown. The combination ofbending and shear loading that resultsfrom this deformation causes a compli-cated stress pattern to develop in andaround the nugget.

There are two different failure modesthat are generally observed in shear-tension tests, namely, “interfacial frac-tures” and “full button pullout” — Fig. 3.In the interfacial fracture, the weld fails atthe interface of the two sheets, leaving halfof the weld nugget in one sheet and half inthe other. In the full button pullout, frac-ture occurs in the base metal or in the weldheat-affected zone at the perimeter of theweld. In this failure mode, the weld nuggetis completely torn from one of the sheetswith the weld remaining intact. It is alsopossible to get a combination of the twofailure modes in which a portion of thenugget is pulled out of one of the sheetsand the rest of the nugget shears at the interface.

A review of the literature showed thatconsiderable work has been done to un-derstand the behavior of spot welds undertensile and shear loading. Work done byDavidson and Imhof (Refs. 4–6) showedthat spot weld strength in the shear-tension test is related to the stiffness of thejoint. They found that, for stiffer test spec-imens the degree of rotation that the spot

Predicting Resistance Spot Weld FailureModes in Shear Tension Tests of Advanced

High-Strength Automotive Steels

To judge the quality of resistance spot welds in advanced high-strength steels, the load-bearing ability of the weld should be considered

more important than the fracture mode

BY D. J. RADAKOVIC AND M. TUMULURU

KEYWORDS

Dual-Phase SteelsTRIP SteelsFracture MechanicsFinite Element ModelingShear-Tension TestFailure Mode

D. J. RADAKOVIC and M. TUMULURU arewith the Research and Technology Center, UnitedStates Steel Corp., Munhall, Pa.Based on a paper presented at the 2006FABTECH International & AWS Welding Show,Atlanta, Ga.

Radakovic Suppl April 2008 corr:Layout 1 3/6/08 3:53 PM Page 96

WELDING RESEARCH

-s97WELDING JOURNAL

weld undergoes becomes smaller, which inturn, leads to increased joint strength.Davidson and Imhof correlated the jointstrength with weld nugget rotation but didnot offer a relationship for the stress in-tensity at the weld. Pook (Ref. 7) devel-oped a relationship for the stress intensityat the weld nugget based on analytical so-lutions. Pook showed that the weld diam-eter and sheet thickness have an effect onthe stress intensity at the nugget perime-ter. Chao (Ref. 8) studied the expectedfailure modes and failure loads for spotwelds based on assumed stress distribu-tions around the perimeter of the weld re-sulting from a combination of shear andtensile loading. Radaj and Zhang (Ref. 9)and Zhang (Ref. 10) performed detailedfinite element modeling of spot weldsunder shear tension load to predict thestress intensity around the weld. Thisanalysis suggested that correlations andtrends predicted in earlier work did notcorrelate well with the detailed computer

models. Radaj and Zhang developedsome simplified equations to predict thestress intensity of spot welds in varioustests, including the shear-tension test. Thecorrelations were consistent with the com-puter simulations. However, the simula-tions are computationally intensive andtime consuming.

In the evaluation of the shear-tensiontest results in spot welds, it is generally be-lieved in the automotive industry that aninterfacial shear failure is indicative ofpoor weld integrity. This has generallybeen true for low-strength steels (tensilestrength equal to or less than 300 MPa), inwhich interfacial failure is normally asso-ciated with insufficient fusion or some sortof a weld imperfection, such as grossporosity. However, it is not clear if inter-facial fractures in shear-tension tests indi-cate poor weld integrity in AHSS grades.With the increased use of these steels inautomotive bodies, it is important to studytheir fracture behavior in shear-tension

tests so that welds, otherwise sound, donot get rejected solely based on fractureappearance. Furthermore, an understand-ing of the fracture behavior may allow theautomotive companies to use these steelsand enable them to take advantage of thebenefits that these steel grades offer. Ad-ditionally, the tensile fracture behavior ofthe recently introduced advanced high-strength steels, such as dual-phase andTRIP steels, has not been reported previ-ously. Therefore, a study was undertakento examine and predict the fracture modespossible in shear tension tests in dual-phase steels with a minimum tensilestrength of 590 and 780 MPa and TRIPsteel with a minimum tensile strength of780 MPa. An attempt based on finite ele-ment modeling (FEM) and fracture me-chanics calculations on data collectedfrom actual shear-tension tests was madeto predict the resistance spot weld failuremodes in shear-tension tests.

Materials and ExperimentalProcedure

Dual-phase steel coils with a minimumultimate tensile strength of 590 and 780MPa, transformation-induced plasticity

Fig. 1 — Sketches showing the dimensions of a shear-tension test coupon. Thearrows in the bottom sketch show the direction of the application of tensile loadduring the test.

Fig. 2 — Sketches showing half sections of a shear-tension coupon before (left)and after (right) the application of load. Bending of the sample (shown in theright-side sketch) results from the offset of the sheets.

Table 1 — Welding Conditions

Welding Machine Manufacturer Taylor Winfield Corp.Welding Machine Type Pedestal-typeWelding Machine Transformer 100 kVAWelding Controller TrueAmp IVElectrode Face Diameter 6 mmElectrode Force For IF Steel: 3.1 kN (697 lbf)

For AHSS Grades: 4.2 kN (945 lbf) for 1 mm5.4 kN (1200 lbf) for 1.2and 1.6 mm

Squeeze Time 75 cyclesWeld Time 13 cycles (for 1-mm sheets)

14 cycles (for 1.2-mm sheets)18 cycles (for 1.6-mm sheets)

Hold Time 10 cyclesPreheating NonePostheating NoneElectrode Coolant Water Temperature 21°CTip Cooling 3.7 L/min (1 gal/min)

Table 2 — Material Properties Used

Steel Yield Tensile Elongation,Grade Strength, Strength, %

MPa MPa

IF 137 302 45590 Dual Phase 370 650 25780 Dual Phase 470 805 19780 TRIP 440 844 24

Note: Elastic Modulus of 207 GPa and Poisson’s Ratioof 0.29 used for each case.


WELDING RESEARCH

APRIL 2008, VOL. 87-s98

(TRIP) steel with a minimum tensilestrength of 780 MPa, and interstitial-free(IF) steel with a tensile strength of 300MPa were used in the study. The IF steelwas used because it is extensively used forbody panels in automobile bodies and theautomotive industry is quite familiar withits welding and fracture behaviors. There-fore, it would be a good candidate to pro-vide a comparison for the fracture behav-ior of AHSS grades.

All the coils were coated with 42/42g/m2 (42 g/m2 on each side) hot-dippedgalvanneal (HDGA) coating. Coatings aregenerally applied to steel sheets used inthe automotive industry for corrosion pro-tection. A galvanneal coating is obtainedby heating the zinc-coated steel at450°–590°C (840°–1100°F) immediatelyafter the steel exits the molten zinc bath.The nominal HDGA coating weights forthe coils used were 42/42 g/m2. These coat-ing weights are typical of current commer-cial automotive use. All the coils usedwere melted, hot and cold rolled at UnitedStates Steel Corp. — Gary Works andcoated subsequently at PRO-TEC Coat-ing Co. of Leipsic, Ohio. The coils rangedin thickness from 1 to 2 mm.

Nominally, the dual-phase and TRIPsteels contain about 0.09 to 0.14 wt-% car-bon and are generally alloyed with variousamounts of manganese, chromium, andmolybdenum to achieve the required ten-sile strength (Refs. 11, 12). In addition tothese alloying elements, TRIP steels typi-cally contain silicon or aluminum to effec-tively suppress the formation of cementiteby increasing the time required for its for-mation and lowering its thermodynamicstability (Ref. 12).

Tensile test samples were preparedfrom coils in the as-received conditionwithout any cleaning of the mill oil usedprior to shipping the coils. Weld shear-tension tests were conducted per Ref. 3.

The current required to produce a weldbutton size equal to or 90% of the face di-ameter of the electrode tip used was de-termined. This was done using the highestcurrent possible without causing expul-sion in the samples. The welding parame-ters used for making the test samples areshown in Table 1. Load to failure and theweld fracture morphology were noted ineach case. In each case, five tensile sam-ples were tested and the average values reported.

In order to better understand the sam-ple behavior and failure modes that occurin the shear-tension test, finite elementcomputer simulations of this test wereperformed. The modeling was done usingABAQUS Version 6.5 general-purpose fi-nite element modeling (FEM) software.The simulations were run on a SiliconGraphics Octane 2 workstation. Three-dimensional models of the test samplewere developed using eight-node brick el-

ements. Model strain predictions werefound to converge for a local element sizenear the weld that is less than 25% of thesheet thickness for each of the three thick-nesses evaluated. An example of a finiteelement model of the shear-tension testsample for a 1-mm sheet with a 6-mm but-ton is shown in Fig. 4. The lower picture inFig. 4 shows a close-up view of the mesh inthe vicinity of the spot weld. In order to re-duce computation time, symmetry condi-tions were applied so that only one half ofthe sample had to be represented. In thestudy, several models were developed torepresent different sheet thicknesses andweld diameters.

The stress-strain behaviors of the threetypes of steel grades used in the simula-tions are shown in Fig. 5. The computermodeling software allows for the defini-tion of elastic as well as plastic behavior ofmaterials and the appropriate data pointswere defined to describe the load dis-

Fig. 3 — A photograph of two shear-tension test samples showing the two differentweld fracture morphologies described in the text. The sample on the top showed in-terfacial fracture, and the sample on the bottom showed full button pullout. The spotwelds on both ends of each sample were made to attach shims for tensile testing.

Fig. 4 — Schematic showing a three-dimensional finite element modelof spot weld shear-tension test sample. The lower picture shows a close-up view of the mesh in the vicinity of the spot weld. Sheet thickness is1.0 mm and button size is 6 mm.

Table 3 — Results of Simulations for Interstitial-Free Steel Samples (Ultimate Tensile Strength =300 MPa)

Weld Diameter Nominal Sheet FEM Failure FEM Failure K-Factor(mm) Thickness (mm) Mode(a) Load (N) IF or PO

2.5 1.0 IF 1200 0.625.0 1.0 PO 3640 2.327.5 1.0 PO 5400 2.3010 1.0 PO 7070 2.26

2.5 1.5 IF 1200 0.615.0 1.5 IF 4720 0.607.5 1.5 PO 8140 2.3110 1.5 PO 10700 2.28

2.5 2.0 IF 1200 0.615.0 2.0 IF 4540 0.587.5 2.0 IF 10280 0.5810 2.0 PO 13920 2.2213 2.0 PO 16700 2.17

(a) IF – interfacial fracture; PO – full button pullout fracture.


WELDING RESEARCH

-s99WELDING JOURNAL

placement behavior shown in Fig. 5. Thematerial properties defined in the analysisare shown in Table 2.

As a starting point in the study, theproperties of the entire test specimenwere considered to be homogeneous. Itwas assumed that the properties of theweld and the heat-affected zone were thesame as those of the base metal. The ho-mogeneous model was used only to esti-mate loads required for pullout failure.This assumption was made to examine theeffect of button size and sheet thicknesson the behavior of the weld in this test.

To predict the failure loads associatedwith interfacial fracture, additional con-siderations were made. It was observedthat the interfacial failures initiated at thenotch created at the sheet interface at the

periphery of the weld. This suggested thatthe fracture toughness of the weld andheat-affected zone controlled the failuremode and load. Previous work (Ref. 10)provided a means to estimate the stress in-tensity at the root of the notch at the sheetinterface and is described in later sections.

In the analysis, sheet thicknesses of 1.0,1.5, and 2.0 mm were simulated. For eachsheet thickness, four or five different weldsizes were modeled ranging in diameterfrom 2.5 to 13 mm. For all models, thesample length and width were held con-stant. A total of 39 simulations were run,13 for each of the three steel types, namelyIF, dual-phase, and TRIP steels.

In all models, one end of the samplewas held fixed and an applied axial dis-placement was applied to the opposite

end. The reaction force was monitored asa function of the applied displacement.When the failure strain was reached in ei-ther the weld or the base metal outside theweld, failure was assumed to have oc-curred. The magnitude of the reactionforce at the point when failure strain wasreached was considered to be the load-carrying capacity of the sample.

Results and Discussion

The results of the analyses of the ho-mogeneous shear-tension test samples areshown in Tables 3–5 and in Figs. 6–11. Fig-ure 6 shows the model-predicted defor-mation of a sample at the onset of a pull-out failure. In this failure mode, thestrength of the material outside the weld

Fig. 5 — Stress-strain curves for the three steel grades used in the study. Fig. 6 — Model-predicted deformation around the weld nugget in the shear-tension test at the onset of a full button pullout failure. Sheet thickness is1.0 mm and button size is 6 mm.

Fig. 7 — Close-up view of the model-predicted plastic strain distribution that occurs during a full button pullout failure in the shear-tension test. Note the neck-ing of the base metal immediately outside the weld nugget.


WELDING RESEARCH

APRIL 2008, VOL. 87-s100

nugget is exceeded and necking begins tooccur at the location shown. After failureof the material at this location, the weldbutton is peeled out of the base metalaround the weld perimeter. A more de-tailed model of this mechanism is shown inFig. 7, which is a contour plot of the plas-tic strain in the homogeneous sample. Asthe figure shows, the maximum plasticstrain is concentrated outside the weldnugget, which leads to necking and failureat this location. The predicted deforma-tion that occurs during the interfacial fail-ure is shown in Fig. 8. Nearly all of thestrain is concentrated in the weld at the in-terface of the two sheets. Figure 9 shows

the plastic strain distribution associatedwith the interfacial failure. There is somestrain outside the weld nugget, but shearfailure of the weld occurs before the pull-out failure can initiate. These results areshown in graphical form in Fig. 10. Themaximum plastic strain in the weld at theinterface of the sheets and the maximumstrain in the base metal outside the weldnugget are plotted as a function of the ap-plied end displacement. For the case offull button pullout, the strain in the basemetal outside the weld nugget is greaterthan that developed at the weld interfaceand the opposite is true for the case of theweld interfacial failure. Based on these re-

sults, it is apparent that there is a compe-tition between the pullout and the interfa-cial failure modes

For the pullout failure, the results ofthe finite element simulations showed thatthere was a strong correlation betweenfailure load and the material strength,sheet thickness, and weld diameter. Theload to cause interfacial failure was foundto be more strongly dependent on the welddiameter and less on the sheet thickness.The predicted failure loads were found toadhere to the following correlations:

FPO = kPO · σUT · d · t (1)

FIF = kIF · σUT · d2 (2)

Where FPO is the failure load for pulloutfailure, FIF is the failure load for interfa-cial fracture, σUT is the tensile strength ofthe material, d is the weld diameter, and tis the sheet thickness. Equations 1 and 2were derived based on the fact that theforce required to cause failure is equal tothe product of the strength of the materialand the failed area of cross section. In thisanalysis, the material was assumed to behomogeneous. Therefore, the strength ofthe weld and the base metal are both equalto σUT. In Equations 1 and 2, kPO and kIFwere constants determined from the modeling.

The results of the analysis of the ho-mogeneous sample analysis are shown inTables 3–5. Each table corresponds to adifferent steel type (IF, DP, and TRIP) andshows the combinations of sheet thicknessand weld diameter that were simulated.As predicted in Equations 1 and 2, the fail-ure loads to cause pullout were found tobe proportional to the weld diameter andthe sheet thickness while those for inter-facial failure were proportional to thesquare of the weld diameters. The pre-dicted failure mode and failure load aswell as the constant kPO or kIF are shownin the tables. The values for kPO and kIFwere backcalculated from Equations 1and 2 by using the model-predicted failureloads, weld diameters, sheet thicknesses,and base metal strengths. The fact that thevalues of these two constants were fairlyconsistent for each case (kPO~2.2 andkIF~0.6) indicates that Equations 1 and 2are good estimates of the model-predictedfailure mode (for the homogeneous sam-ples that were simulated).

Examination of Figs. 6–9 showed that,due to offset of the sheets, the shear-ten-sion test specimens undergo rotationalong an axis perpendicular to the loadingdirection. The degree of rotation wasfound to be greater for the cases of fullbutton pullout failures (Figs. 6, 7) com-pared to those that failed interfacially(Figs. 8, 9). Davidson and Imhof (Ref. 6)

Table 4 — Results of Simulations for DP 590 Steel (Minimum Ultimate Tensile Strength = 590 MPa)


2.5 1.0 IF 2290 0.555.0 1.0 PO 7600 2.277.5 1.0 PO 11600 2.3110 1.0 PO 15060 2.25

2.5 1.5 IF 2260 0.545.0 1.5 IF 9660 0.587.5 1.5 PO 16860 2.2410 1.5 PO 22700 2.26

2.5 2.0 IF 2300 0.555.0 2.0 IF 9300 0.567.5 2.0 IF 22150 0.5910 2.0 PO 28700 2.1413 2.0 PO 35500 2.12


Fig. 8 — Mode-predicted deformation in the shear-tension test showing shear overload of the weld metalat the interface of the two sheets. Sheet thickness is 1.0 mm and button size is 4 mm.


WELDING RESEARCH

-s101WELDING JOURNAL

discussed the significance of specimen ro-tation during fatigue testing of spotwelded samples. In fatigue testing, shear-tension test coupons are subjected to re-peated tensile loads in the same manner asin static shear-tension testing. Davidsonand Imhof showed that the stiffness of testsamples affected the degree of rotationand that the weld failure mode was af-fected by the degree of rotation. They ob-served that stiffer samples (less rotation)failed interfacially, whereas, below a cer-tain critical stiffness, the samples failed inthe base metal and the weld remained intact.

Equations 1 and 2 are plotted graphi-cally in Fig. 11 for dual-phase 590 steel andshow the predicted interfacial and pulloutfailure load as a function of weld diame-ter. In this plot, the sheet thickness was as-sumed to be 1.5 mm and the sheet tensilestrength equal to 590 MPa. A sheet thick-ness of 1.5 mm was chosen because it rep-resented the mid-thickness of the normalrange of steels (1 to 2 mm) used in auto-motive bodies. The minimum allowabletensile strength of the steel (590 MPa) wasused for this case. According to the analy-sis, the lower of the two predicted failureloads will determine the mode of failurethat occurs. Thus, the point where the twocurves intersect indicates where the fail-ure mode changes. The figure shows that,for this sheet thickness and strength, in-terfacial failures were predicted to occurfor weld diameters below 6 mm and pull-out failures for diameters greater than 6mm. Figure 11B shows a similar plot withthe exception that the weld diameter wasassumed to be constant at 7 mm and thefailure loads are plotted as a function ofsheet thickness. These results indicatethat pullout failures are more likely tooccur on thinner sheet samples, and themode can change to interfacial when thesheet thickness reaches a critical value (forthis example ~ 1.7 mm). By setting Equa-tions 1 and 2 equal to each other, the ratioof the weld diameter to the sheet thicknessis determined to be roughly equal to 4.This suggests, for a homogeneous testsample, pullout failures will occur if theweld diameter is greater than four timesthe sheet thickness. Likewise, for a givenweld diameter, the sheet thickness has tobe less than 25% of the weld diameter forpullout to occur.

The homogeneous model predictionsagreed well with the test data for caseswhere pullout failures occurred. This isbecause the pullout failures most ofteninitiated in the base metal outside of thenotch at the perimeter of the weld. A com-parison of the actual failure loads andmodel-predicted failure loads for samplesthat failed via pullout is shown in Fig. 12.However, when comparing the homoge-

neous model results to actual test data, thepredicted failure loads for the interfacialfractures were not consistent — Fig. 12. Itwas theorized that there were two mainreasons for the differences. First, themodel assumed that the sample was ho-mogeneous. In reality, the mechanicalproperties of the weld, base metal, and dif-ferent parts of the heat-affected zone aresignificantly different. For example, it wasshown that the hardness of weld fusionzones and the heat-affected zones aremuch higher than those of the base mate-rial in dual-phase steels (Ref. 13). Further,the computer models did not account forthe stress intensity at the perimeter of theweld. The model considered that the weldwould fail when the strength of the weldmetal is exceeded at the interface of thetwo sheets. This would more closely rep-resent a shear overload of the weld metalrather than a fracture of the weld nuggetinitiating at the notch around the perime-ter of the nugget.

A fracture mechanics study by Zhang(Ref. 9) provided an estimate of the stressintensity at the perimeter of the spot weldspecifically for the shear-tension test.Fracture mechanics theory provides ameans to evaluate the load-bearing abilityof materials that have cracks or flaws. In

fracture theory, the fitness for service of astructural member that contains a crack isdetermined by comparing the predictedstress intensity at the crack tip to the frac-ture toughness of the material. Like mate-rial properties such as tensile strength andelongation, fracture toughness is a mater-ial property that is determined from test-ing. Fracture analysis of a cracked mem-ber (in which stress intensity is comparedto fracture toughness) is analogous to theanalysis of notch-free structural membersin which predicted stresses are comparedto the strength of the material. Zhang’sanalysis yielded the following relationshipfor the stress intensity at the perimeter ofthe weld in the shear-tension test.

In Equation 3, Ktseq is the equivalent stress

intensity factor at the spot weld, F is theapplied load, d is the weld diameter, and tis the sheet thickness. Some actual test re-sults are shown in Table 6. The table liststhe steel grades, weld dimensions, failuremode, failure load, and the calculated

F K d tIF C

= ⋅ ⋅ ⋅1.44 (4)

KF

d teqts = ⋅0.694 (3)

Table 6 — Actual Test Results and Predicted Stress Intensity at Failure

Material Nominal Sheet Weld Diameter Failure Mode Failure Load Stress IntensityThickness (mm) (mm) IF or PO (N) (N/mm3⁄2)

780 DP 1.2 7.0 IF 14230 1294590 DP 1.0 6.0 IF 9790 1126

780 TRIP 1.6 8.5 IF 21800 1385780 DP 1.2 7.0 PO 15520 1411780 DP 1.6 8.0 PO 22900 1581


Table 5 — Results of Simulations for Dual-Phase and TRIP 780 Steels (Minimum UltimateStrength = 780 MPa)


2.5 1.0 IF 3000 0.585.0 1.0 PO 8950 2.177.5 1.0 PO 13600 2.2010 1.0 PO 17780 2.15

2.5 1.5 IF 3030 0.595.0 1.5 IF 11900 0.587.5 1.5 PO 20500 2.2110 1.5 PO 26400 2.13

2.5 2.0 IF 2920 0.575.0 2.0 IF 11530 0.567.5 2.0 IF 26760 0.5810 2.0 PO 34760 2.1113 2.0 PO 43700 2.12



WELDING RESEARCH

APRIL 2008, VOL. 87-s102

stress intensity at failure. The stress inten-sities were calculated using Equation 3.The first three samples listed in the tablefailed via interfacial fracture. This sug-gests that the listed stress intensity at thetime of failure is comparable to the frac-ture toughness of the weld. The stress in-

tensities for the samples that failed viabutton pullout are also shown. For thesetwo cases, the fracture toughness of theweld was apparently high enough to avoidinterfacial fracture (i.e., the fracturetoughness was greater than the valueslisted for the stress intensity when the pull-

out failure occurred).Table 7 shows additional data compar-

ing the model predictions to measureddata. The table lists the steel grades, welddiameter, sheet thickness, actual failureload, and mode as well as the predictedfailure load had pullout failures occurred

Table 7 — Comparison of Actual Shear Tension Test Results and Model Predictions

Actual Test Data Model PredictionsMaterial Sheet Thickness Weld Diameter Failure Failure FEM Failure FEM Failure Percent of Max

(mm) (mm) Mode(a) Load (N) Mode Load (N) PO Load(b)

780 DP 1.2 7.5 PO 15500 PO 15820 —780 DP 1.6 8.3 PO 22900 PO 23250 —590 DP 1.0 7.0 IF 9900 PO 10340 96780 DP 1.2 7.5 IF 14680 PO 15820 93

780 TRIP 1.6 8.5 IF 22060 PO 24330 91

(a) IF — interfacial fracture; PO — full button pullout. (b) Ratio of the actual failure load to the model-predicted pullout failure load if the failure were to occur by the pullout mode (multiplied by 100).

Fig. 9 — Close-up view of the model-predicted plastic strain distribution that occurs during a shear overload of the weld in the shear-tension test.

Fig. 10 — Predicted maximum plastic strain in the weld and in the base metal immediately outside the weld as a function of applied displacement in the sheartension test. There is a competition between pullout (A) and interfacial (B) modes of failure. Sheet thickness is 1.0 mm for both cases. The weld button size is6 mm (A) and 4 mm (B).

A B


WELDING RESEARCH


in each case. The first two samples listedfailed via button pullout and the model-predicted failure loads were in goodagreement with the actual loads. The fol-lowing three samples in the table failed viainterfacial fracture. For these cases, themodel-predicted loads for pullout failureare listed. The interesting results shownhere are that, although interfacial fractureoccurred in the samples, the load-carryingcapacity of the weld was greater than 90%of the predicted failure load if pullout hadoccurred. This indicates that the load-bearing capacity of these welds was notsignificantly affected by the fracturemode. Therefore, the mode of failureshould not be the only criteria used tojudge the quality of spot welds. The load-

bearing capacity of the weld should be theprimary focus in the evaluation of theshear-tension test results in AHSS.

In order to get an estimate of the loadrequired for interfacial fracture in theshear-tension test, the stress intensity inEquation 3 was set equal to the fracturetoughness of the material and the load wassolved for. This is shown in Equation 4where FIF is the load to cause interfacialfracture of the weld, and KC is the fracturetoughness of the material. The fracturetoughness describes the ability of a mate-rial to carry load in the presence of a flawand is determined from testing. Generally,ductile materials tend to have high frac-ture toughness while the opposite is truefor brittle materials. Based on the work

done by Zhang (Ref. 9), Equation 4(rather than Equation 2) was thought tobetter represent the parameters govern-ing the interfacial fracture mode. The pre-viously determined load to cause pulloutfailure (Equation 1) was, however, consid-ered to be appropriate. The predicted re-lationship for pullout load likely agreedwell with the measured data because thisfailure initiates in the form of necking inthe base metal near the weld heat-affectedzone as opposed to at the notch radiusaround the perimeter of the weld nugget.

Equations 1 and 4 are plotted graphi-cally in Fig. 13. In this plot, the weld di-ameter was assumed to be constant at 8mm and the tensile strength assumed wasthe minimum allowable for this grade (780

Fig. 11 — Predicted failure load in the analysis of the homogeneous shear-tension test sample as a function of the following: A — Weld diameter; B — sheetthickness.

Fig. 13 — Predicted competition between failure modes in the shear-tensiontest.

Fig. 12 — Comparison of actual failure loads to model-predicted failure loadsin the shear-tension test for samples that failed via full-button pullout.

A B


WELDING RESEARCH

APRIL 2008, VOL. 87-s104

MPa), and the failure load is plotted as afunction of sheet thickness. The interfa-cial fracture load is plotted assuming afracture toughness of 1560 N/mm3⁄2, andthe pullout failure load is plotted for alow-strength steel (300 MPa) and ahigher-strength steel (780 MPa). For thecurve representing the low-strength steel,the predicted pullout failure load is lessthan the interfacial fracture load over thisentire range of sheet thickness. This indi-cates that pullout failure will occur inevery case. For the higher-strength sam-ple, the curve for pullout failure and in-terfacial fracture intersect at a sheet thick-

ness of 1.6 mm. This indicates that for thehigh-strength sheet sample, interfacialfracture will be the expected failure modefor sheet thicknesses greater than 1.6 mm.

The critical parameters that controlthe transition between failure modes canbe determined by setting Equation 4 equalto Equation 1. Some of these results areshown in the last four figures. Figure 14shows that critical sheet thickness abovewhich interfacial fracture will occur be-comes less as the strength of the sheet in-creases. Thus, for higher-strength sheet,interfacial fracture can become the ex-pected failure mode in thicker samples.

Figure 15 shows the fracture toughness ofthe weld required to maintain pullout fail-ures increases with the strength of thesheet indicating that interfacial fracturesare more likely to occur with high-strengthsteels. Figure 16 shows that the weldtoughness required to maintain pulloutfailures also increases with sheet thick-ness. Figure 17 shows the predicted pull-out and interfacial failure loads for a low-strength steel shear-tension test samplethat has a weld with poor load-carrying ca-pacity. The failure load required for pull-out is less than that for interfacial fractureover the entire range of sheet thickness

Fig. 14 — Sheet thickness above which interfacial fracture is predicted to occuras a function of sheet tensile strength and the fracture toughness of the weld.

Fig. 15 — Predicted weld toughness required to avoid interfacial fracture inthe shear-tension test as a function of sheet tensile strength for 1- and 2-mm-thick sheets.

Fig. 16 — Predicted weld toughness required to avoid interfacial fracture in theshear-tension test as a function of sheet thickness and tensile strength.

Fig. 17 — Predicted failure load as a function of sheet thickness for a low-strength steel having a weld with poor load-carrying capacity. For a low-strength sheet, a pullout failure is predicted even when the load-carrying ca-pacity of the weld is poor.


WELDING RESEARCH


shown in the plot. This indicates that forlow-strength steels, a full button pulloutfailure could be expected even for the casewhere the weld has poor load-carrying ca-pacity. For high-strength steels, however,this was not found to be the case. Interfa-cial fracture was predicted to occur forcases where the weld has superior fracturetoughness and high load-carrying capac-ity. These predictions shown in Figs. 15–17agree well with known characteristics ofspot welds.

Conclusions

An analysis was performed in which acombination of finite element modelingand fracture mechanics calculations wasused to predict the weld failure modes inthe shear-tension tests of resistance spotwelds in AHSS grades. In the finite ele-ment model, the base material and heat-affected zone and weld properties were as-sumed to be homogeneous. Thehomogeneous model predictions agreedwell with the test data for cases where pull-out failures occurred. This is because thepullout failures most often initiated in thebase metal outside of the notch at theperimeter of the weld. However, whencomparing the homogeneous model re-sults to actual test data, the predicted fail-ure loads for the interfacial fractures werenot consistent. For the case of the interfa-cial failure mode, a relationship was foundin the literature to estimate the stress in-tensity at the weld notch tip. This rela-tionship was used along with the equationdeveloped for the pullout failure to definevariables that affect the failure mode andload for the shear-tension test.

The results of the analyses showed thefollowing:

1. The mathematical equation derivedbased on the finite element modelingshowed that the force required to cause abutton pullout fracture was found to beproportional to the tensile strength of thesheet as well as the diameter of the weldand the thickness of the sheet.

2. The present analyses showed that,

for low-strength steels (tensile strengthless than or equal to 300 MPa), a full but-ton pullout could occur even when weldshave a poor load-carrying capacity. Forhigh-strength steels, however, this was notfound to be the case. Interfacial fracturewas predicted to occur in the shear tensiontest for cases where the weld has superiorfracture toughness and high load-carryingcapacity.

3. It was determined that there is a crit-ical sheet thickness above which the ex-pected failure mode could move frompullout to interfacial fracture. It wasshown that, as the strength of the sheet in-creases, the fracture toughness of the weldrequired to avoid interfacial fracturesmust also increase. In the higher-strength,less-ductile steels, this is not likely to occurand interfacial fracture could become theexpected failure mode.

4. The load-carrying capacity of thesamples that failed via interfacial fracturewas found to be more than 90% of themaximum load associated with the fullbutton pullout. This indicates that theload-bearing capacity of these welds is notsignificantly affected by the fracturemode. Thus, the mode of failure shouldnot be the only criteria used to judge theresults of the shear-tension test. The load-carrying capacity of the weld should beconsidered the most important parameterwhen evaluating the shear-tension test re-sults in AHSS.

References

1. Schultz, R. A. 2007. Metallic materialtrends for North American light vehicles. Paperpresented at the Great Designs in Steel Semi-nar, American Iron and Steel Institute, South-field, Mich.

2. Horvath, C. 2007. Material challengesfacing the automotive and steel industries fromglobalization. Paper presented at the Great De-signs in Steel Seminar, American Iron and SteelInstitute, Southfield, Mich.

3. D8.9M-2002, Recommended Practices forTest Methods for Evaluating the Resistance SpotWelding Behavior of Automotive Sheet Steel Ma-terials. Miami, Fla.: American Welding Society.

4. Davidson, J. A. 1983. Design-related

methodology to determine the fatigue life andrelated failure mode of spot-welded sheetsteels. International Conf. on Technology andApplications of High-Strength Low-AlloySteels, ASM International Technical Series8306-022, Metals Park, Ohio.

5. Davidson, J. A., Imhof, E. J. Jr., 1984. TheEffect of Tensile Strength on the Fatigure Life ofSpot-Welded Sheet Steels. SAE Technical PaperNo. 848110.

6. Davidson, J. A., and Imhof, E. J. 1983. AFracture Mechanics and System-Stiffness Ap-proach to Fatigue Performance of Spot-WeldedSheets. SAE Technical Paper No. 830034.

7. Pook, L. P. 1975. Fracture mechanicsanalysis of the fatigue behavior of spot welds.International Journal of Fracture Vol. 11, pp.173–176.

8. Chao, Y. J. 2003. Ultimate strength andfailure mechanism of resistance spot weld sub-jected to tensile, shear, or combinedtensile/shearloads. Journal of Engineering Mate-rials and Technology Vol. 125(4): 125–132.

9. Radaj, D., and Zhang, S. 1991. Simplifiedformulae for stress intensity factors of spotwelds. Engineering Fracture Mechanics Vol.40(1): 233–236.

10. Zhang, S. 1997. Stress intensities at spotwelds. International Journal of Fracture Vol. 88,pp. 167–185.

11. Rege, J. S., Inazumi, T., Nagataki, T.,Smith, G., Zudeima, B., and Denner, S. 2002.Development of HDGI/HDGA DP steel fam-ily at National Steel Corp. 44th MWSP Confer-ence Proceedings, ISS, Vol. XL.

12. Mahieu, J., Maki, J., Claessens, S., andDe Cooman, B. C. 2001. Hot dip galvanizing ofAl alloyed TRIP steels. 43rd MWSP Confer-ence Proceedings, ISS, Vol. XXXIX.

13. Tumuluru, M. 2006. An overview of theresistance spot welding of coated high-strengthdual-phase steels. Welding Journal 85(8): 31–37.

The material in this paper is intended forgeneral information only. Any use of this mate-rial in relation to any specific application shouldbe based on independent examination and ver-ification of its unrestricted availability for suchuse, and a determination of suitability for theapplication by professionally qualified person-nel. No license under any United States SteelCorporation patents or other proprietary inter-est is implied by the publication of this paper.Those making use of or relying upon the mate-rial assume all risks and liability arising fromsuch use or reliance.

Nominees Solicited for Prof. Koichi Masubuchi Award

November 3, 2008, is the deadlinefor submiting nominations for the2009 Prof. Koichi Masubuchi

Award, sponsored by the Dept. of OceanEngineering at Massacuusetts Institute ofTechnology. It is presented each year toone person who has made significant con-tributions to the advancement of materi-als joining through research and develop-

ment. The candidate must be 40 years oldor younger, may live anywhere in theworld, and need not be an AWS member.The nominations should be prepared bysomeone familar with the research back-ground of the candidate. Include a ré-sumé listing background, experience,publications, honors, awards, plus a leastthree letters of recommendation from

researchers.This award was established to recog-

nize Prof. Koichi Masubuchi for his nu-merous contributions to the advancementof the science and technoogy of welding,especially in the fields of fabricating ma-rine and outer space structures.

Submit your nominations to Prof. JohnDuPont at [email protected].

Radakovic Suppl April 2008 corr:Layout 1 3/6/08 4:00 PM 5

Pruebas de cizalladura RSW

Documents

Transcript of Pruebas de cizalladura RSW