weld bond

8/6/2019 Weld Bond

1/5

ABSTRACT. Using a three-dimensionalfinite element analysis (FEA) method, theeffect of elastic modulus and thickness ofadhesives on the stress distribution inweld-bonded joints was studied to ad-dress the role of adhesive layer. Normalstress and shear stress distributed at theedges of a spot weld and in the lap regionwere computed for weld-bonded joints

made with adhesives of different elasticmoduli or thicknesses. The resultsshowed great stress concentration at theedge of the spot weld in weld-bonded

joints when the adhesive layer was thickor had a low elastic modulus. Shear stressvalues in adhesive layers were low underthe same circumstances. Stress concen-tration around the spot weld was reducedand the shear stress in the adhesive layerwas increased by increasing the elasticmodulus or decreasing the thickness ofthe adhesive layer. An adhesive layerwith appropriate thickness and elasticmodulus is necessary to obtain reason-able distribution of stresses in the wholelap region of a weld-bonded joint. A thinadhesive layer of high elastic modulus isfavorable to the fatigue properties ofweld-bonded joints, and it is recom-mended on certain conditions.

Introduction

Weld bonding is an advanced hybridtechnology that has the advantages ofspot welding and adhesive bondingcombined (Refs. 15). The stress concen-tration at the periphery of spot welds is

reduced and the fatigue performance ofthe joints is significantly improved by theapplication of an adhesive. The corro-sion problem in the inner surface of the

joints lap region is successfully solved atthe same time. Compared with adhesive-bonded joints, the tearing strength ofweld-bonded joints is superior and jointreliability is favorable. At present, due tothe excellent mechanical properties ofweld-bonded joints, weld bonding hasbeen used widely in the aviation andspace-flight fields and on the production

lines of automobiles.In weld-bonded lap joints, both the

spot weld and the adhesive layer con-tribute to the joint strength. The load-bearing capability of the two con-stituents and the stress distribution inweld-bonded joints are determined bymany factors, such as the shape and sizeof the joints and the mechanical proper-ties of the adhesive and base metal.Many experimental results showed theproperties of adhesives used in weld-bonded technology have important ef-fects on fracture mode and load-bearingcapability of the joints. The experimen-tal results have been analyzed qualita-tively from the point of view of joint stiff-ness, but they have not been interpretedquantitatively. In the present investiga-tion, a three-dimensional elastoplasticfinite element method was used to studythe effect of the elastic modulus and thethickness of the adhesive on stress distri-bution in weld-bonded joints. The rela-tionship between the stress distribution

and the fracture mode, together with the joint strength, were considered here.The conclusions drawn have instruc-tional significance for designing weld-bonded joints, choosing the adhesivesand expanding weld-bonding technol-ogy.

Computational Model and

Properties of Materials

Generally, weld-bonded structuresand specimens for joint strength testing aremade in lap joint configuration. Hence, aweld-bonded lap joint with a single spotwas analyzed, as shown in Fig. 1. The ten-sile shear loads were applied at the twoends of the joint. Electrode indentationswere not taken into account in this finiteelement model. The joint was wellbonded, and no defects, such as cavitiesor inclusions, were located at the inter-face. In addition, i t was supposed the spotweld and the heat-affected zone (HAZ) ofthe joint had the same mechanical prop-erties. Since the specimen was symmetricabout the x- axis, only half of the specimenwas considered. The finite elementmeshes used in the computation areshown in Fig. 2. Three-dimensional brickelements were used. Both the upper andthe lower plates were divided into two lay-ers. The adhesive layer included two lay-ers of meshes. The zones near the edges ofthe lap region and the spot weld were di-vided into finer meshes because thestresses in those zones were the main con-cern. The minimum size of the mesh was

0.15 mm. Overall, 1808 elements and2471 nodes were included in the mesh.Mechanical properties of materials usedin computation are listed in Table 1. Thebase metal was 08Al steel (correspondingto AISI 1010, wi th chemical compositionsof 0.06% C, 0.1% Si, 0.23% Mn, 0.004%P, 0.016% S, 0.04% Al and 0.05% Cu),which is used in the manufacture of auto-mobi le structures. The ALGOR elastoplas-tic FEA program was used to compute thestresses in the weld-bonded joints, whichwere made using adhesives with different

WELDING RESEARCH SUPPLEMENT | 275-s

A Study on the Role of Adhesives inWeld-Bonded Joints

BY B. H. CHANG, Y. W. SHI AND S. J. DONG

Stresses in weld-bonded joints with adhesives of different elastic moduli andthicknesses are obtained by a three-dimensional finite element method

KEY WORDS

Weld BondingElastic ModulusFinite Element Method FEMStress AnalysisStress DistributionJoint StrengthAISI 1010Resistance WeldingAdhesive Bonding

B. H. CHANG is with the School of Mechani-

cal Engineering, Xian Jiaotong Uni versity,

Xi an, P.R. China. Y. W. SHI is a Professor,

School of Materials Science and Engineering,

Beij ing Polytechnic University, Beij ing, P.R.

China. S. J. DONG is an Associate Professor,

Materials Department, Hubei Automotive In-

dustries Institute, Shiyan, P.R. China.

8/6/2019 Weld Bond

2/5

elastic moduli and thick-nesses. The applied loadswere equal to 2.5 kN,with an average tensilestress of 125 MPa pre-dicted by materialsstrength theory.

The joint was onlylightly deformed. Onlymaterial nonlinearity aris-

ing from elastoplastic be-havior of materials wastaken into account in thecomputation, while thegeometrical nonlinearitycaused by a large dis-placement and strain wasneglected. All computa-tion was accomplished onan IBM microcomputer.

Results andDiscussions

Effect of Elastic Modulus

of Adhesive on Stress

Distribution

It was reported thatthe elastic modulus of theadhesive has more evi-dent effect on thestrength of weld-bonded

joints than does the ad-hesive toughness (Refs.69). The weld-bonded

joints, made with eithertoughened or nontough-ened epoxy adhesives,had almost the samestrengths. Great differ-ences were found be-tween the mechanicalproperties of weld-bonded joints made wi thtoughened epoxy adhe-sives and high modulusand joints made withacrylic adhesives andlow modulus. Whenweld-bonded structuresare tested under condi-tions of high or varyingtemperatures, the modu-

lus of the adhesive willchange. It is well known

that the stiffness of joints is mainly deter-mined by the elastic modulus if the sec-tion areas of the specimens do not vary.Therefore, the elastic modulus of the ad-hesive can be used to characterize thedeforming ability of the weld-bonded

joint; that is to say, weld-bonded jointswith a low modulus are easier to deformthan those with a high modulus.

Stress distribution curves were drawn

to make it clear how the elastic modul i ofthe adhesives affect the stresses in theweld-bonded joints. Generall y, the adhe-sive layers were considered to be frac-tured by the shear stress at the edges ofthe lap regions. Radial stress at the pe-riphery of the spot weld was the deter-mining factor for the fracture of a spotweld. Fatigue strength of weld-bonded

joints was mainly determined by themaximum radial stress at the edges of thespot weld and lap regions, so that x, thenormal stress in x direction, and zx, theshear stress at the edges of the spot weldand the lap region, were investigated.

The normal stress x and shear stresszx at the edges of the lap region areshown in Fig. 3. The stresses are distrib-uted in the direction of the plate width. Itcan be seen in Fig. 3A that the normalstress x decreased with the increase ofY for all five types of adhesives with arange of the elastic moduli from 50 to200,000 MPa. A high x existed in themiddle part of the specimen and a rela-tively low x at the two edges of the spec-imen. For joints with a high modulus, thex was lower than in joints with a lowmodulus. When the elastic modulus ofthe adhesives equalled 50 MPa, remark-able stress concentration was found atthe middle part of the weld-bonded

joints, which may have been a responseto the spot weld in the joint. Wi th an in-crease of the adhesives modulus, x inthe middle of the plate decreased, andthe stress concentration also decreased.

Shear stresses had uniform distribu-tion along the plate width for all fivemoduli, and a slight drop of shear stresswas observed at the plate edge. Thehigher the modulus, the higher the shearstresses in the adhesive layers.

Stress curves at the periphery of the

spot welds are il lustrated in Fig. 4, the ab-scissa of which is the angular degree of. The is defined in Fig. 5. It can befound from Fig. 5 that the angular rangeof 0 deg 180 deg corresponds to theupper half of the spot weld, the partwhere 0 deg 90 deg was the loadedside of the spot weld, and the part where90 deg 180 deg was on the load-freeside.

From Fig. 4, it was found when theelastic modulus of the adhesive wasgreater than 500 MPa, the normal stress

276-s | AUGUST 1999

RESEARCH/DEVELOPMENT/RESEARCH/DEV

ELOPMENT/RESEARCH/DEV


ELOPMENT

Fig. 1 Shape and dimensions of the specimen for finite ele-

ment analysis.

Fig. 2 Finite element mesh of a weld-bonded lap joint. A

Finite element meshes for the whole joint; B enlarged meshes

for spot weld and lap region.

A

B

8/6/2019 Weld Bond

3/5

x was low and had uniform distributionover the whole angular range. When themodulus decreased, the stresses aroundthe spot weld were no longer uniform.The tensile stressesx on the loaded sideincreased and the x on the free sidechanged from tensile to compressive.The maximums of both tensile stresses

and compressive stresses increased withthe decrease of the adhesive modulus.For the weld-bonded joints with an ad-hesive modulus of 50 MPa, distinct stressconcentration appeared at the edge ofthe spot weld.

Shear stresses at the edges of the spotweld were small for weld-bonded jointswith adhesives of high elastic moduli.When the elastic modulus of the adhe-sive was greater than 5000 MPa, themaximum shear stress was lower than 20MPa. The shear stresses increased when

the moduli of the adhesives decreased,which implied the spot weld bore moreof the shear load.

From the stress distribution laws men-tioned above, it can be concluded theelastic modulus of adhesive has an influ-ence on the stresses in weld-bonded

joints. When the elastic modulus is low,

shear stresses in adhesive layers weresmall and significant stress concentrationwas observed at the edge of the spotweld. With the increase of the modulus,stresses at the edge of the spot weld de-creased and shear stresses in the adhe-sive increased a little. This was becausewhen the moduli of the adhesives de-creased, joint stiffness became small andthe joints deformed more easily. The ap-plied loads were mainly borne by thespot weld, and stresses distributed in theadhesives were a small part of the total

loads, which resulted in high normalstress x and shear stress zx in the spotweld zone. Joint strength was consideredfrom the point of view of stress distribu-tion. Stress concentration occurred in theweld-bonded joints with low elasticmodulus, and high shear stresses werefound at the edge of the lap regions for

joints wi th adhesives of high elastic mod-ulus. When the adhesive modulus de-creases, the fracture mode of weld-bonded joints will transform fromfracture in the adhesive layer first to frac-ture in the spot weld (Ref. 10). The adhe-sive layer and the spot weld have differ-ent strengths. A reasonable stressdistribution can be obtained by using anadhesive with suitable modulus,whereby the adhesive layer and the spotweld reach the fracture stresses at thesame time. For this reason, i t is useful and


A

A

B

B

Fig. 3 Distribution of stresses at the edge of the lap region along the plate width for weld-bonded joints with different elastic moduli. A Dis-

tribution ofxalong the plate width; B distribution ofzxalong the plate width.

Fig. 4 Angular distribution of shear stresses at the periphery of spot weld in weld-bonded joints with different elastic moduli. A Angular dis-

tribution of normal stressx; B angular distribution of shear stresszx.

8/6/2019 Weld Bond

4/5

important to select adhe-sives with a specific elasticmodulus.

Effect of Thickness of

Adhesive Layer on

Stress Distribution

Under production condi-tions, the thickness of the

adhesive layer is influencedby such technological fac-tors as the assembling andfixing of adherents, thequantity of adhesives andmethods of applying adhe-sives. The thickness of theadhesive layer can vary in arange from 0.1 ~ 1.0 mm. Sofar, not many experimentsor computations have beendone to study the influenceof the adhesive thickness onthe joint strength. The 3-Delastoplastic FEA method

was used in this study togain insight into the rela-tionship between the stressdistribution and the adhe-sive thickness. A phenolicresin adhesive was used inthe computation, which hasa modulus of 2875 MPa.

Curves of normal stressx and shear stress zx dis-tributed along the platewidth are shown in Fig. 6. Itcan be seen from Fig. 6Athat x has the same distrib-ution tendency for five ad-hesive layers wi th differentthickness (0.1, 0.2, 0.3, 0.4and 0.5 mm). Thex distrib-uted uniformly in the mid-dle part of the width, whichwas greater than that at thetwo edges of the joints. Inthe whol e range of the platewidth, the x increased withthe increase of the adhesivethickness and the x in themiddle part had a greater in-crease than that at the twoedges. Figure 6B indicates

shear stresses distribute uni-formly along the plate widthand decrease near the plateedges. The shear stressesover the whole range ofplate width decreased withthe increase of the thicknessof the adhesive layer.

Stresses of x and zxaround the spot welds areplotted in Fig. 7 for fiveweld-bonded joints with dif-ferent thicknesses of adhe-

sives. The x distributed uniformly andthe stress differences between the loadedside and the free side of the spot weldwere small (about 8 MPa) when the ad-hesive layers were thin. For greater adhe-sive thickness, x of the loaded side in-creased while that of the free sidedecreased; then the differences becamegreater (about 80 MPa). It is shown in Fig.7B the shear stresses zx on the loaded

side of the spot weld were higher than thaton the free side. Near the zone of = 120deg, the shear stresses had the minimumvalues. The shear stress increased withthe increase of the adhesive thickness,and the distribution pattern did not vary.

Similar to the adhesive modulus, thethickness of the adhesive layer has a cleareffect on the stress distribution in weld-bonded joints. When the adhesive layerwas thick, stress concentration aroundthe spot weld was evident. When the ad-hesive layers became thinner, the normaland shear stresses around the spot welddecreased and the shear stresses in the

adhesive layers increased slightly. Thephenolic resin adhesive has an elasticmodulus of 2875 MPa, which is muchlower than 200,000 MPa of the modulusof the base metal 08Al . The restraint fromthe base metal to the adhesive layer de-creased wi th the increase in the thicknessand the stiffness of the joi nts decreased atthe same time. Then a larger deformationwill occur in the adhesive layers for thesame loads applied. More loads wereborne by the spot weld, and higher nor-mal and shear stresses were found there.Normal stressx at the edge of the lap re-gion increased with the adhesive layerthickness because of the larger deforma-tion of adhesive layer in the loading di-rection, while the shear stress zx theredecreased because of a smaller restrainton angular deformation of the thicker ad-hesive layer. For weld-bonded joints wi ththin adhesive layers, fracture was gener-ally initi ated at the edge of the lap regionof the joints. The fracture mode willtransform from fracture in the adhesivesto fracture in the spot weld with an in-crease in adhesive layer thickness (Ref.10). It is very important for improving the

joint fatigue strength and optimizing the

stress distribution to design and produceweld-bonded joints wi th a suitable thick-ness of adhesive.

The application of adhesives with thespot welding process greatly reduces thestress concentration around the spotweld and improves the fatigue strength ofthe joint, wi th the degree of the reductionand the improvement being dependenton the properties of the adhesive (Refs.1112). In this study, it was found stressconcentration around the spot weld wasreduced when the elastic modulus in-

278-s | AUGUST 1999

RESEARCH/DEVELOPMENT/RESEARCH/DEV



ELOPMENT

Fig. 5 Schematic of a spot weld in the lap region of a weld-

bonded joint.

Fig. 6 Distribution of stresses at the edge of the lap region

along the plate width for weld-bonded joints with different ad-

hesive thicknesses. A Distribution of x along the plate

width; B distribution ofzxalong the plate width.

A

B

8/6/2019 Weld Bond

5/5

creased or the thickness of the adhesive

layer decreased. The shear stress valueslocated at the edge of the lap region areless than 60 MPa, although greater shearstresses were found also. The dominantcontrolling factor for the joints fracturewas the normal stress x in this circum-stance; therefore, a thin adhesive layerwith a high modulus was favorable forthe fatigue property in weld-bonded

joints. Of course, when the adhesive lay-ers are too thin, the joints cannot bebonded well. Some technological prob-lems will be caused when the elasticmodulus of the adhesive is too high, suchas the difficulty with displacing the ad-

hesive layer beneath the welding elec-trodes and the insulating effect of the ad-hesive layer on the spot weld. Thesefactors must be taken into account inchoosing the type of adhesive and de-signing the thickness of its layers.

Conclusions

In the present study, a three-dimen-sional numerical analysis model was de-veloped for weld-bonded joints. Normaland shear stresses in the joints with dif-ferent adhesives were computed. Rela-tionships between stress distribution andelastic modulus and thickness of adhe-sives were discussed. From this work, thefollowing conclusions are drawn:

1) In the model developed, indenta-tions were ignored and the joint wasthought to be bonded well. In addition,the HAZ was considered to have the samemechanical properties as the spot weld.Results obtained in previous studiesshowed that these assumptions have anegligible impact on computation results.

2) Stresses around the spot weld in-

creased with the decrease in the elastic

modulus of the adhesive. Great stressconcentration was found when the elas-tic modulus of an adhesive was lowerthan 100 MPa. The stress concentrationaround the spot weld was reduced byusing adhesives with a high elastic mod-ulus, but greater shear stress was found atthe edge of the overlap region.

3) Stresses in weld-bonded jointswere influenced by the thickness of theadhesive layer. Greater normal stress andsmaller shear stress were found at the lapedge for weld-bonded joints wi th a thickadhesive layer, and magnitudes of bothnormal and shear stresses around the

spot weld increased with an increase inadhesive layer thickness.

4) It was found that an adhesive layerwith a high elastic modulus or smallthickness will cut down stress concen-tration in weld-bonded joints and im-prove the fatigue strength of the joint.Hence, a thin adhesive layer with highelastic modulus is recommended be-cause the joint can be bonded well andno technological problem is encoun-tered in its production.

References

1. Jones, T. B. 1995. Weld-bonding themechanism and properties of weld-bondedjoints. Sheet Metal Industries72(9): 2731.

2. Budde, L., and Hahn, O. 1992. Adhe-sive bonding in combination with spot weld-ing or clinching. Welding in the World30(1/2): 2032.

3. Pfau, P. 1975. Stress distribution inweld-bonded joints, calculation on the basisof a one dimensional model. Wiss. Zeit.(Th karl-Marx-Stadt) 17(1): 7184.

4. Pfau, P. 1977. Stress distribution inweld-bonded joints, calculation on the basis

of a two dimensional model. Wiss. Zeit. (Th

karl-Marx-Stadt) 19(3): 341364.5. Hi ll s, D. R., Parker, J. D., and Wil li ams,

N. T. 1995. The effect of increasing samplewidth on strength of weld-bonded lap joints.Key Engineering Materials99100: 119126.

6. Jones, T. B. 1978. A future for weld-bonding sheet steel. Welding and Metal Fab-rication46(7/8): 415420.

7. Jones, T. B., and Wil li ams, N. T. 1986.Fatigue in adhesive and weld-bonded steeljoints. SAE Journal94(7): 4853.

8. Wang, P. C., Chi sholm, S. K., Banas, G.,Lawrence, F. V. 1995. The role of fail ure mode,resistance spot weld and adhesive on the fa-tigue behavior of weld-bonded aluminum.Welding Journal74(2): 41-s to 47-s.

9. Ryazantsev, V. I., and Shavyrin, V. N.

1981. Strength characteristics of welded andbonded-welded joints. Welding ResearchAbroad27(6): 3537.

10. Hills, D. R., Parker, D. R., andWilliams, N. T. 1996. Effect of number ofwelds in spot welded and weld-bonded arrayson static failure load of individual spot weldsand adhesive bond. Ironmaking and Steel-making26(2): 150156.

11. Chang, B. H ., Shi, Y. W., and Dong, S.J. 1998. Comparative studies on stress in weld-bonded, spot-welded and adhesive-bondedjoints. Journal of Materials Processing Tech-nology87: 230236.

12. Chang, B. H. 1998. Studies on stressand strain fields and mechanical properties forweld-bonded joints. Ph.D. Dissertation. Xian,

Xian Jiaotong University, P. R. China.


A B

Fig. 7 Angular distribution of stresses at the periphery of the spot weld in weld-bonded joints with different adhesive thicknesses. A Angular

distribution of normal stressx; B angular distribution of shear stresszx.

weld bond

Documents