A Study of the Stress Concentration Factor on Spot Welds

Transcription

1 A Study f the Stress Cnentrt Ftr n Spt Welds Test results dite tht the full pplit f the stress nentrt ftr is nt reqred design BY J-H. KUANG AND A-H. LIU ABSTRACT. The stress nentrt effet f resistne spt weld between tw sheets f metl is shwn t be dependent upn the gemetri dimenss nd the plne periphery rk impsed by the terfes f the welded pltes. Due t the rtt f the spt weld nugget nd the stress nentrt effet f plne rk, the stte f stress f the spt weld under tensile sher ld is very mplited. A three-dimensl isprmetri fite element mdel is used t nlye the stress distribut rund the spt under stti tensile sher ldg. In rder t simplify the mplex nfigurt f the periphery rk, n elliptil blunt-nth ssumpt is pplied t pprximte the pex f the rk. The effet f the gemetri prmeters f the speimen, i.e., spt shpe, plte thikness nd width, re vestigted. A reltship between the mximum stress nentrt ftr, shpe nd the rientt f n elliptil spt is develped. Experimentl results re ls pplied t vlidte the lulted results qulittively. Intrdut Spt weldg hs been widely used fr the fbrit f sheet metl ssemblies se Elihu Thmsn disvered the resistne weldg phenmenn High speed nd dptbility fr utmt re the mjr dvntges f spt weldg. The press ls hs limitts, e.g., reltively lw tensile nd ftigue strengths when stressed tens. This is due t the shrp nth rund the periphery f the nugget between the sheets. Bsed upn nsiderble munt f experimentl dt prvided the 1940s nd 1950s n pl rbn steel, Heushkel (Ref, 1) fund tht the tensile sher strength f sgle spt is determed by the sheet thikness, bse metl ultimte strength, spt dimeter nd the mpsit f welded metls. Rudy, MCuley nd Green (Ref. 2) studied the str distribut rund spt welds tensile sher. Rudy dites tht str first ppers t the weld edge nd prgresses lng the terfe plne twrd the enter f the weld nugget. The differene ter mrphlgy is ntrlled by the mnner whih this mvement lng the terfe plne is blked. The ritil weld dimeter thikness rti nept ws first trdued by Rudy In 1977, n nlytil pprh ws prpsed by Vnden Bsshe (Ref. 3) t develp lult f the weld dimeter thikness rti. The strength f sgle spt sher is estimted by pplyg the Vn Mises distrt energy filure thery. Bsshe's pper is fused n the strength f the whle spt nugget, the stress distribut nd stress nentrt vrit re nt the primry bjetives. In 1980, Swhill (Ref. 4) mdified Heushkel's spt strength equts by trdug prprtlity ftr. Defurny (Ref. 5) ls develped the strength equt fr bth mild steel nd HSLA steel. An nlytil study but the ftigue resistne f sgle irulr spt weld hs been nduted by Kn (Ref. 6). In 1983, three-dimensl fite element mdel ws prvided by Lwrene (Ref. 7) t vestigte the stress distribu- Dy P *- _»» X w Dx I. C=200 mm V= 34 mm k-l W= 34 mm Tu =Tl = 1.5 mm Fig. 1 Dimenss f the A WS stndrd test speimen. Tu KEY WRDS Resistne Spt Weld Stress Cnentrt Isprmetri Mdel Fite Element Mdel Three-Dimensl Stti Tens Sher Tens Sher Ldg Impsed Plne Crk Elliptil Spt Weld Cirulr Spt Nugget l-h. KUANG is Prfessr nd A-H. LIU is Grdute Student the Deprtment i Mehnil Engeerg, Ntl Sun Yt-Sen University, Khsiung, Tiwn. Pper presented t the 68th Annul AWS Meetg, held Mrh 22-27, 1987, Chig, III. 468-s DECEMBER 1990

2 t ner irulr spt nugget. A urvefitted equt hs been prvided t estimte the stress nentrt ftr f irulr spt under tensile sher ld. All these reserhers hve mde imprtnt ntributs t the understndg f spt weld strength. This pper vestigtes the stress distribut rund sgle elliptil spt weld by usg the fite element pprh. An elliptil blunt nth is pplied t pprximte the nth effet rund the periphery f the nugget between the sheets. The strength f elliptil spt welded jts depends nt nly upn the rss-setl re f the nugget but ls shpe nd rientt. Therefre, mprmise must be mde between shpe nd the rientt f the elliptil spt weld rder t bt the mximum jt strength. Fite Element Mdel f Sgle Spt Weld A generl purpse fite element prgrm MSC/NASTRAN ws used fr the nlytil slut f spt welded pltes. As nted mny reprts, the re f stti nlysis, the element tehnlgy nd the nlysis lgrithm f MSC/NAS TRAN pper t be stbilig. In this pper n istrpi ler elsti mteril behvir f the spt nugget ws ssumed. Figure 1 shws tw pltes t be spt welded lp jt under rbitrry xil ldg. Se the nfigurt f the bdy, s well s the externl ndits re symmetril, it my be sfe t nsider nly hlf f the bdy fr fite element idelit. The symmetry ndits, hwever, hve t be rprted the slut predure. This is shwn Fig. 2, where nly hlf f the pltes with the spt nugget, hvg symmetry bth geme- Dx : Spt Dimeters X-Diret try nd ldg, is nsidered fr nlysis. The pltes nd spt nugget re subdivided t number f nnulr rg-type slid elements ner the nugget. T imprve ury, smller elements re used nes where rpid vrits stress re ntiipted, rther thn the ne f nstnt stress. Nte tht the elements with the rdil length f the spt nugget re mmn t elements bth pltes. The ttl meshes fr this hlf f welded pltes hve 1134 hexhedrn slid elements nd 1662 ndl pts with three degrees f freedm per nde. Unifrmly distributed lds re pplied t bth ends f the jed pltes. Due t the symmetry, there nnt be trnsverse displement lng the le f symmetry f the x- xis. The ndit tht nstred ndes the diret f the y-xis hs t be rprted while fdg the slut. The nugget enter the terfe plne hs ls been nstred withut ny trnslt t remve the pssible rigid bdy mt the slut. Figure 1 shws the stndrd speimen Ldg Diret Fig. 2 Fite element mesh ner irulr spt weld. suggested by AWS (Ref. 8). Tw 34 X 117- mm (1.3 X 4.6-.) pltes f unequl thikness T u nd T re spt welded tgether with 34-mm verlp. The spt nugget is lted t the enter f the verlpped prt. The jed shpe nd rss-setl re f the spt re pprximted t the shpe nd rss-setl re f the eletrde. A diffiulty enuntered preedg with the fite element mdel is the determt f the shpe f the nth rund the periphery f the nugget between the sheets. As nted, the urvture f the rk tip will severely ffet the stress distribut nd stress nentrt ftr. Mirspi exmt f the rss-set ner the nugget te revels tht the nfigurt f the rk pex is mplited nd vries signiftly. In rder t simplify the nfigurt, n elliptil blunt nth shwn Fig. 3 is ssumed t pprximte the pex f the periphery rk, where nd b re the lengths f the semimjr nd the semimr xes f the elliptil blunt nth, respetively. The mximum stress nen- Spt ugget ; : V Fr Cirulr Spt = 0.3 mm " Fig. 3 Elliptil blunt nth f the plne rk. i i i Dimeters Rti f Blunt Nth (lb) Fig. 4 - Effet f blunt-nth shpe n the mximum stress nentrt ftr f irulr spt. 40 WELDING RESEARCH SUPPLEMENT 1469-s

3 k Fr n Elliptil Spt Dx = 3.79 mm Dx/Dy = 1.0/2.5 = 0.3 mm 03 (D : C ' 1 1 ' Dimeters Rti f Blunt Nth (/b) Fig. 5 Effet f blunt-nth shpe n the mximum stress nentrt ftr f n elliptil spt. Rdil Distne frm Spt Center (p/r; Fig. 6 The stress distribut ner irulr spt weld. trt ftr fr sgle irulr spt with dimeter f 6 mm (0.24.) is shwn Fig. 4. A gemetri stress nentrt ftr K is defed s K = 7l where -\ = the mximum pripl stress; T vg = the nml stress (=P/A); P = the stti xil ld; A = the rss-setl re f nugget (=7rD 2 /4); nd D = the dimeter f eletrde. Results re nverged if n elliptil blunt nth with 0.3-mm (1-.) length f semimjr xis nd rti t semimjr xis is greter r equl t 15; i.e., = 0.3 mm nd /b 15. Similr results hve ls been slved fr sgle elliptil spt s shwn Fig. 5. Hene, n elliptil blunt nth with semimjr xis f 0.3 mm nd rti f 15 is used the fite element mdel. Results nd Disusss Sgle Cirulr Spt Weld Usg the fite element mdel desribed bve, the mid-plne stress distribut rund sgle irulr spt with dimeter f 6 mm ws mputed. Figure 6 shws the stress distribut rund the irulr spt nugget different diret. Where p is the xil distne frm the spt enter nd R is the rdius f the spt nugget, s, the rti p/r = 0.1 dites the enter nd periphery f the spt, respetively. The mximum pripl stress urs n the periphery the ldg diret. A mximum stress nentrt ftr f 7 is fund this se. There re number f ftrs tht shuld be nsidered when nlyg fr spt weldg; severl f these re width, dissimilr thikness f pltes nd spt dimeter. T unt fr the mst imprtnt effet f these ftrs, vriety f eh ftr is emplyed, eh f whih is tended t unt fr sgle effet. Figure 7 dites the stress nentrt effet rund irulr spt fr different speimen widths. It turns ut tht the stress nentrt ftr hs smller vlue fr lrger width. Results shw tht the stress nentrt effet is nt sensitive t the width vrit. The reltship between the stress nentrt nd the thikness f the welded pltes hs ls been vestigted. Tw ses re disussed here: ne is the stress nentrt effet fr the pltes with equl thiknesses, nd the ther is ssumed fr pltes with dissimilr thiknesses. Figure 8 is plt f the stress nentrt ftr fr irulr spt under the equl thikness ssumpt. Results dite tht smller thikness/dimeter rti my use mre severe stress nentrt prblem. Figure 9 shws the results fr welded pltes with dissimilr thiknesses. Fr this dissimilr thikness se, the mximum pripl stress lwys ltes n the spt periphery f the thner plte. Frm these results, it is lerned tht the thikness f welded pltes is very sensitive prmeter t the weld strength. The pssible resn fr this my be the eentriity f ldg. This eentriity f ldg will use the jt t rtte nd hnge the stress distribut llly. In prtie, het blne my ls present prblem when spt weldg tgether unequl thiknesses f the sme metl. A number f experiments fr the sgle irulr spt hve been nduted t vlidte the numeril results. The ll stress distribut rund the irulr spt is diffiult t mesure these experiments. Fr this resn, n pprximte wy is pplied. It is ssumed tht the se with 470-s DECEMBER 1990

4 " Width (W)/Dimeter (D) Fr Cirulr Spt T Fig. 7 Effet f speimen width n the mximum stress nentrt ftr f irulr spt weld. 01 C " 7.0 Fr Cirulr Spt D= 6 mm Tl = 1.5 mm (1 1.1 Thiknesses Rti (Tu/TI) Fig. 9 Effet f thikness rti n the stress nentrt irulr spt. r u. i 0.6" 0.5" /Qsuhy \y Fr Cirulr Spt D= 6 mm Tu = Tl = T 10 ftr f Thikness (T)/Dimeter (D) Fig Strength rtis fr different thiknesses f the welded pltes V Fr Cirulr Spt Tu = Tl = T Thikness (T)/Dimeter (D) 1 Fig. 8-Effet f thikness f the welded pltes n the mximum stress nentrt ftr n?- 1 ' 1 ' Fr Cirulr Spt fit - I ' r l Width (W)/Dimeter (D) Fig. 10 Strength rtis fr different speimen widths tu s _i -~^ Q u S Q. 0C 2 CL Q ~^ Z r S -i *». I t WELDING RESEARCH SUPPLEMENT 471-s

5 :s 0..j X e :E CL.J LJ x K : H Z E Q. I : t : Z LL! _i u ;/ 3 "N. E J / r Fig. 12-The distribut f pripl stress fr n elliptil spt with rti f dimeters I C C t u_ (J 3.0 i C IAI // 11 \ $://:!..8 = 85 II A 1 V v8/'' ' '/ 9-90 ih f\n Fr n Elliptil Spt Dx/Dy = 1.5/1.0 Dx = 7.35 mm e= 9=5 //', 9=15 ///,.6-30 ////, 9 = 45 ifl/fr 9-60,'///// 9 = Rdil Distne frm Spt Center (p/r) Fr n Elliptil Spt Dx/Dy =.0/1.5 Dx = 4.9C mm // / ' / / / ' /// / /! / / / / 5&C/'//" \ 'NNV,' f X^^' N^t^' \Jfll *MM+Jf /" 9= 0 9 = 15 6 = = 75 9 = Rdil Distne frm Spt Center (p/r) Fig. 13 The distribut f pripl stress fr n elliptil spt ^,_ r u_ * C " 7.0- I higher stress nentrt ftr my hve lwer weldg strength the tensile-sher test. Essentilly, the nlysis is qulittive mprisn nly. A high-rbn steel with speifit JIS 4051 S50C hs been emplyed fr this speimen. Bsed n the stndrd speimen, vrits width f 34, 39, 45, 51 nd 57 mm (1.3, 1.5, 1.8, 2.0 nd 2.2.), respetively, were used. Thirty speimens fr eh width hve been tested. In Fig. 10, the dt re pltted by usg strength rti where P ut is the ultimte ld f this spt welded speimen tensile-sher ldg test, nd the referene ld P ref is the prdut f the ultimte strength f the pltes nd the weld re f the spt. The spt weld will fil either by sher thrugh the nugget r by terg the nugget ut f the sheet under ultimte ld. The generlly resg vlues f the ultimte ld is qte pprent even thugh the test dt exhibit gret del f stter. Fr referene, the results with three times f the stndrd devit re ls pltted s dshed les this figure. Ardg t the thery f sttistil hypthesis, there is suffiient evidene t sy tht the wider speimen is strnger. The similr tensilesher tests hve ls been nduted t study the effet f the thikness vrit n the stress nentrt f spt welded pltes. Five different thiknesses f pltes, ///All J ^7 ///AHI v Fr n Elliptil Spt Dx/Dy - 1.0/2.5 Dx = 3.79 mm, 9_ Q If" 9=5 //'. 6= 15 jjff 9 = 30 ;'///,- 9 = 45 /////, 9 = 60 v lilll /;' 9 = 75 ^STT^/T,' 6 = 85 ^^TSgS^C"-" 9 = 90 \^~^, i i i i Rdil Distne frm Spt Center (p/r) Fig. 14 The distribut f pripl stress fr n elliptil spt with rti f dimeters D x/d y = 1.0/ s DECEMBER 1990

6 r Li- C 0) u Dimeters Rti f Elliptil Spt (Dx/Dy) Fig. 15 Lt vrit f the mximum stress nentrt fr different elliptil spt shpe. ftr Dimeters Rti f Elliptil Spt (Dx/Dy) Fig. 16 Effet f elliptil spt shpe n the mgnitude f the mximum stress nentrt ftr. 1.6, 1.8, 2.0, 2.6 nd 3.2 mm (0.62, 7, 8, 0.10 nd 0.13.), were emplyed. timte lds fr the spt welded speimens vried with thikness s shwn Fig. 11. Experimentl results dite tht the thner welded pltes hve lwer strengths; shwg gd greement with the numeril sluts bted by pplyg the fite element methd. Sgle Elliptil Spt Weld The effet f spt shpe n the weld strength is very terestg subjet. Here n elliptil spt is studied. Similrly, the sme elliptil peripherl blunt nth rund the spt nugget is emplyed the rrespndg fite element mdel. In rder t mpre the differene f the weld strength nd stress distribut between elliptil spt nd the irulr spt, the sme weldg re is ssumed the fllwg nlysis. The stress distribut f sgle elliptil spt with rti f dimeters D x/d y = 1.5/1.0 is shwn Fig. 12. Where D x is either the mjr r the mr xis the ldg diret, nd D y is the rrespndg xis the trnsverse diret. The diret ngle 8 is frmed by the psit vetr nd the ldg xis, where p/r is the dimensless rti f rdil distne t the vrible rdius f the elliptil spt the sme diret. Figure 13 shws the stress distribut ner the elliptil spt, whih is rtted t right ngle, i.e., D x/d y = 1.0/1.5. Under tensile ld f P = 780 N, the mximum pripl stresses re 218, 195 nd 175 N/ mm 2 fr the elliptil spt with the rti Dx/Dy = 1.5/1.0, 1.0/1.0 nd 1.0/1.5, respetively. Numeril results dite the rti f dimeters nd the rientts my nt nly hnge the mgnitude f the stress nentrt ftr, but ls my shift the lt f the mximum stress. Figure 14 shws the stress distribut f n elliptil spt with rti f dimeters Dx/Dy = 1.0/2.5. It is fund tht the mximum pripl stress is lted t the spt periphery with n ngle 8 = 60 deg frm the ldg diret. Figure 15 shws the stress vrit t the spt edge with 8 = 0, 45 nd 60 deg. Cmpred with the results Fig. 15, it is dited tht the mximum stress nentrt ftr is lted t the spt edge with n ngle 0 = 0 deg when the rti D x/d y is greter thn 0.5. And the lt f the mximum stress nentrt ftr will be shifted utwrd when D x/d y is deresed, e.g., the lt will be shifted t 8 = 60 deg when D x/d y 0.4. Figure 16 shws the vrit f the mximum stress nentrt ftr fr different rti f dimeters. It is evident tht the lwest stress nentrt ftr is bted fr D x/d y = 1.0/2.5. In ther wrds, it my hve the strngest weld strength by usg XD C TD C s E Dimeters Rti f Elliptil Spt (Dx/Dy) n elliptil spt with the rti f dirmeters D x/d y = 1.0/2.5. In rder t vlidte the numeril results, mprisn is nduted between lulted results nd the experimentl results. Weld strength fr different shpe f sgle elliptil spt is shwn Fig. 17. The strngest weld strength is fund fr n elliptil spt with rti f dimeters D x/d y = 1.0/ 2.5. Althugh the weld strength nd the stress nentrt ftr re defed differently, frm the results shwn Figs. 16 nd 17, the trend f vrit is very herent. The mprisn is mpletely qulittive. In this study, nly the gemetri stress nentrt ftr is f nern. As nted, the stress nentrt hs highly llied effet. The high stresses exist nly very smll reg the viity f the disntty. The tul mteril f whih the prt is mde des nt Fig. 17-Weld strength fr different shpe f n elliptil spt. WELDING RESEARCH SUPPLEMENT 1473-s

7 enter t ll t the prblem f determg the stress nentrt ftr. In the se f dutile metl, the sttement is usully mde tht the first ld pplied t the member uses the mteril t yield t the disntty nd relieves sme f the stress nentrt. Cnsequently, it is never neessry t pply the full vlue f the stress nentrt ftr tul design. Cnluss In this pper, the stress nentrt ftrs f sgle irulr spt nd sgle elliptil spt welds hve been vestigted. The fllwg nluss resulted frm this study: Sgle Cirulr Spt 1) The stress nentrt ftr f sgle spt weld sher is determed by the thikness/dimeter rti when equl thiknesses f the sme metl re emplyed. The use f lrger thikness/dimeter rti will trdue the lwer stress nentrt ftr. 2) Stress nentrt my present severe prblem when spt weldg tgether unequl thiknesses f the sme metl. The mximum pripl stress lwys urs the thner plte. Sgle Elliptil Spt 1) The shpe nd rientt f n elliptil spt weld my nt nly ffet the mgnitude f the stress nentrt ftr, but ls my shift the lt f the mximum stress nentrt. 2) An elliptil spt weld with rti f dimeters D x /D y = 1.0/2.5, whih the mjr xis is perpendiulr t the ldg diret nd 2.5 times length, will prvide the lwest stress nentrt ftr. Referenes 1. Heushkel, The express f spt weld prperties. Weldg jurnl 31(10): Rudy, j. F., MCuley, R. B., nd Green, R. S The behvir f spt welds under stress. Weldg jurnl 35(2): Vnden Bsshe, D. ] timte strength nd filure mde f spt welds high-strength steels, SAF pper. N Swhill, ). M., nd Bker, ). C, ]r Spt weldbility f high-strength sheet steels. Weldg jurnl 59(1): Defurny, J., D'Heyer, R., nd Bryrd, A Spt weldg f high-strength steel sheets fr deep drwg. IIW Dument III, pp Kn, Y. R Ftigue resistne f spt welds n nlytil study, Metls Engeerg, pp fwrene, F. V r., Wng, P. C, H, N. J., nd Crent, H. T Estimtg the ftigue resistne f tensile-sher spt welds. FCP Reprt 48, tu-enc We/dg Hndbk, 7th Ed., Vl. 3. Amer Weldg Siety, Mimi, Fl. WRC Bullet 340 Jnury 1989 Interpretive Reprt n the Mehnil Prperties f Bred Jts By M. M. Shwrt This reprt summries the mehnil prperties, brg predures nd testg f mterils, suh s lumum, beryllium, pper, steel, nikel, superllys, retive metls, ermis nd grphite. ver 120 referenes re luded this terpretive reprt. Publit f this reprt ws spnsred by the Interpretive Reprts Cmmittee f the Weldg Reserh Cunil. The prie f WRC Bullet 340 is $24.00 per py, plus $0 fr pstge nd hndlg. rders shuld be sent with pyment t the Weldg Reserh Cunil. 345 E. 47th St., Ste 1301, New Yrk, NY WRC Bullet 341 Februry 1989 A Prelimry Evlut f the Elevted Temperture Behvir f Blted Flnged Cnnet By J. H. Bikfrd, K. Hyshi, A. T. Chng nd J. R. Wter This Bullet nsists f fur Sets tht present prelimry evlut f the urrent knwledge f the elevted temperture behvir f blted flnged nnet. Set I Intrdut nd verview, by J. H. Bikfrd; Set II Histril Review f Prblem Het Exhnger, by J. R. Wter; Set III Develpment f Simple Fite Element Mdel f n Elevted Temperture Blted Flnged Jt, by K. Hyshi nd A. T. Chng; nd Set IV Disuss f the ABACUS Fite Element Anlysis Results Reltive t In-the-Field bservts nd Clssil Anlysis, by J. R. Wter. Publit f this reprt ws spnsred by the Submmittee n Blted Flnged Cnnets f the Pressure Vessel Reserh Cmmittee f the Weldg Reserh Cmmittee. The prie f WRC Bullet 341 is $20 per py, plus $0 fr pstge nd hndlg. rders shuld be sent with pyment t the Weldg Reserh Cunil, Ste 1301, 345 E. 47th St., New Yrk, NY s DECEMBER 1990