Stress Residual en Vigas Armadas

8/12/2019 Stress Residual en Vigas Armadas

1/13

R es i dua l S t resses inH eavy W e lded ShapesG e o m e t r y of p l a t e s and s h a p e s isan i m p o r t a n t v a r ia b l e a f f e c t i n gr e s i d u a l s t r e s s m a g n i t u d e and d i s t r i b u t i o n and i n i t i a l r e s i d u a l s t r e s s e s due to r o l l i n gc a n bea h i g h e r m a g n i tu d e t h a n t h o s e due to w e l d i n gB Y G. A. A L P S T E N AN D L T A L L

A B S T R A C T . R esidual s tresses can have asignificant influence on the load-ca rry ingbehavior of s t ruc tura l s tee l memb ers subjec ted to compress ive loads . Prev ious experimenta l re sea rch on residual s tressesan d the s t reng th of c o l u m n s was relatedto smal l and medium-s ize shapes . Intoday's large s tructures , increasinglyheavy shapes are employ ed . W hile heavycolumn shapes are being used extensively, very little information has been available on the residual s tresses and strengthof such members .

T h is paper p resen ts the results of thefirst phase of a major in vestigation intothe residual s tresses in, and the behav iorof, thick plates and heavy shapes used inc o mp re s s io n me mb e rs . The shapes considered in this initia l s tudy are a 15H290shape and a 2 3 H 6 8 1 s h a p e , as well astwo loose component p la tes , PL 16x2 andP L 2 4 x 3 V . , . For the smalle r shap e , comparative tests were carried out for differen t manufac tur ing condi t ions of the c o m ponent p la tes (un ive rsa l-mil l and flame-cut pla tes ), different weld type (pen etra t ion) and different yield s trengths of thema te r i a l .

T he resu l ts of residual s tress measurements ca rr ied out in this first phase ofthe s tudy indicate the fo l lowing forheavy fabrica ted me mb ers :1. All phases of the ma n u fa c tu re andfabrication procedure generally affect thefo rma t io n of residual s tresses .2. The weld type (pe ne tra t ion ) andthe yield s trength of the steel are notmajor fac to rs in the fo rma t io n of residual stresses.3. The g e o me t ry of the p la tes andshapes is one of the impo rtan t va r iab lesaffecting the residual s tress magnitudeand d is t r ibu t ion .4. The varia t ion of residual s tressac ross the th ickness of p la tes more than

G. A. A L P S T E N is A ssociate Director.Swedish Ins t i tu te tor S teel C onstructionan d was formerly with Lehigh University:L . T A L L is Director, Division for Fa t igueand Frac ture . Fr i tz Engineering Labora tory . Lehigh University, B ethlehem, Pa.

Paper presented at the AWS50th A nnualMeeting held in Philadelphia, Pa., duringA p r il 28 to May 2, 1969.

1 in.thickcan be cons ide rab le .5. The weldin g residual s tresses inport ions of the cross section other thanthe we ld a rea tend to decre ase w ithincreasing s ize of the mem ber, p rob ab ly

because the weld area, and consequent ly ,the heat input, is re la t ive ly smal le r inheavy p la tes and shapes as c o mp a re d tol igh t members .6. The initia l s tresses can be of ahigher magni tude than the welding residual s tresses .T he re la t ionsh ip be tween in i t ia l re s id ual s tresses in ccm ponen t p la tes andwelding residual s tresses implies thatefforts to l imit the ma g n i tu d e of residualstresses in heavy welded shapes should bedirec ted towards the ma n u fa c tu re of thec o mp o n e n t p l a t e s . T h u s , by using flame-cut plates in heavy welded shapes , the reis a prospec t for an increase in strengthwhen compared with l igh te r members atthe same s lenderness ra t io . T here iseven a possibility that such weldedshapes may be s t ronger than the ir ro l ledc o u n te rp a r t s .

IntroductionO n l y in the p a s t two d e c a d e s was

i t e x p e r i m e n t a l l y and t h e o r e t i c a l l yv e r i f i e d th a t r e s id u a l s t r e s s e s are ama jo r in f lu e n c e on the s t r e n g t h ofs t e e l m e m b e r s in c o m p r e s s i o n . It wass h o w n t h a t t h e y h a v e a s ign if ican teffect on c o l u m n s t r e n g t h - 2 and ont h e b u c k l i n g of p la t e s .3 - 4 It is a lsok n o w n th a t r e s id u a l s t r e s s e s can be ofg r e a t i m p o r t a n c e in fa t ig u e , b r i t t l ef r a c t u r e , an ds t r e s s c o r r o s i o n .

T h e p r e v i o u s e x p e r i m e n t a l r e s e a r c ho n re s id u a l s t r e s s and the s t r e n g t h ofw e ld e d s t e e l members wa s r e l a t e d tos m a l l and m e d i u m - s i z e shapesthatis , s h a p e s of c o m p o n e n t s w i t h a t h i c k n e s s e q u a l to or less than 1 in. ~rl Int o d a y ' s l a rg e s t ru c tu re s , i n c re a s in g lyh e a v y s h a p e s are used . Ver y l i t t lei n f o r m a t i o n has b e e n a v a i l a b le on ther e s id u a l s t r e s s e s and s t r e n g t h of h e a v yc o l u m n s , yet t h e y are u s e d e x te n s iv e lyi n N o r t h A m e r i c a and e l s e w h e r e . Ap p l i c a t i o n s i n c l u d e the l o w e r s t o r i e s of

mu l t i - s to r y b u i ld in g s , ma jo r b r id g e s ,a n d l a u n c h i n g g a n t r i e s for r o c k e t s ands p a c e v e h i c le s . S i m i l a r s t r u c t u r a l e l e m e n t s are u s e d in s h i p and s u b m a r i n ehul ls , and in vesse ls for a t o m i c r e a c to r s .

S o m e h e a v y c o l u m n s h a p e s u s e d ine x i s t i n g s t r u c t u r e s are s h o w n in Fig.1, w h ic h a l s o i l l u s t r a t e s the d iffe ren tw a y s of d e s i g n i n g a h e a v y c o l u m ns h a p e . R o l l e d s h a p e s can be u s e d anda re a v a i l a b le in s izes up to the1 4 W F 7 3 0 j u m b o s h ap e of Fig. la.If the s t r e n g t h of the a v a i l a b le ro l l e ds h a p e is insuff ic ien t for a p a r t i c u l a ra p p l i c a t i o n , p l a t e s can be w e l d e d to aro l l e d s h a p e as s h o w n in F i g s , l b andc . A h e a v y s h a p e ca n a l s o be b u i l t upf ro m w e ld in g to g e th e r th re e p l a t e s tof o r m an H - s h a p e , as i l l u s t r a t e d in Fig .l d . F i n a l l y , a b o x - s h a p e can be m a d ef r o m w e l d i n g t o g e t h e r f o u r p l a t e s F i g . l e . Th e p la t e s u s e d in a w e l d e ds h a p e can be m a n u f a c t u r e d e i t h e r asu n iv e r s a l -mi l l p l a t e s ( th a t is, ro l l e d toe x a c t w id th and u s e d w i th a s - ro l l e de d g e s ) or as flame-cut * p la t e s ( th a tis, flame-cut from a l a rg e r b a s ep l a t e ) .

A t p r es e n t , the d e s i g n of h e a v yc o l u m n s d o e s not d i f f er f ro m th a t ofs m a l l and m e d i u m - s i z e c o l u m n s . Thed e s ig n c r i t e r i a p re v io u s ly d e v e lo p e dfo r s ma l l and m e d i u m - s i z e r o l l e dc o l u m n s 2 h a v e b e e n e x t r a p o l a t e d toi n c l u d e h e a v y m e m b e r s . W h i l e e x p e r i e n c e has in d ic a te d t h a t t h i s l e a d s tos a fe d e s ig n , it may not be a c o m p l e t e l y r a t i o n a l d e s ig n m e t h o d .

A n e x t e n s i v e r e s e a r c h p r o g r a m isc u r r e n t l y u n d e r w a y at L e h i g h U n i v e r s ity to s tu d y re s id u a l s t r e s s e s in h e a v yw e ld e d p la t e s and s h a p e s . ( H e a v yp la t e s and s h a p e s are d e f in e d h e re asm e m b e r s w i t h a t h i c k n e s s e x c e e d i n g 1i n . ) . The s pe c i f i c o b je c t iv e s of the

*Or oxygen-cut (from oxygen cu tting asdefined by AWS in Terms and Definitions,A WS A 3 . 0-69 ).

W E L D I N G R E S E A R C H S U P P L E M E N T | 93 s


2/13

73 0 lb / ft(a ) 13 0 lb /ft(b ) 2 6 9 0 l b / f t(C)

1S c a l e ( i n c h e s ) :

10 2 0 3 01610 Ib/ff 1190 l b / f t(d) (e)

Fi g. 1Heavy co l u m n s h a p e s i n e x i s t i n g s t r u c t u r e s

X I8W40( a ) I I H 6 6(b) I 4 H 2 0 2(O1

1 1I5H 290( d )

2 3 H 6 8 I( e )

324D774( f ) 2 4 H I I 2 2(g)

S c a l e ( i n c h e s ) :i 1 1 i0 10 20 30

Fig.2Test c o l u m n sstudy are to determine the magnitudeand distribution of residual stresses inthick welded plates by both experimental and theoretica l means, and torelate this to the stability under compressive loads of s tructura l members .

Prior to the initiation of the currentstudy, critical information was lackingon the behavior of heavy columns.T he present status is illustrated in Fig.2 ( a -d ) which shows:a. T he largest test shape so far usedin a multi-story frame test.b. T he largest welded shape in abeam-column test .c . T he largest column shap e .d . T he largest s tub column shape .T he shapes in Figs . 2e through 2gare examples of the specimens in thecurrent research program . T hese shapescompare to the heavy column shapesused in construction (see Fig. 1) .

T he paper presents some exper imental results obtained in the firstphase of the investigation. While thespecimens in the overall program cover the complete range of dimensionswhich are practical in construct ion,that is, plates ranging in size from 9 xV 2 up to 24 x 6 in. and weldedH-shapes and a box-shape rangingfrom a 7H28 to a 24H1122 shape, thespecimens considered in this paper arethe 15H290 and 23H681 shapes (seeFigs. 2d and e) as well as two as-manufactured pla tes, 1 6 x 2 and 24 x 31/?in . F igure 3 shows a comparison between the 23H681 shape and the7H28 shape, corresponding to theheaviest and lightest welded specimenstested so far in the Lehigh program.T he results of the overall study willallow the prediction of residualstresses in any plate and any welded

shape used in constructio n. T he pla tesand shapes in the investigation willserve as reference data, and by knowing the dimensions of components andthe manufactur ing and fabr ication details of a practical column, the distribution of residual stress can be predic ted. T he column strength can thenbe determined, including the effect ofresidual stresses.T he study is of a fundamen talnature and of considerable importance

in many areas; however , the present

Fig. 3Largest ( 23H 681) and sm al l es t( 7 H 2 8 ) c o l u m n s h a p e s t e s t e d s o f a r i nt h e r e s e a r c h p r o g r a m

Fig. 4Fabrication of t h e w e l d e d H -s h a p e 2 3H 6 81 ( co u r t e s y B e t h l e h e mS t e e l C o r p . )

applicatio n is the developm ent of information which will be useful inpreparing design criteria for heavycolumn members as are used in construction. T he f indings obta ined forthe basic welded plates will be applicable also to other types of structuresfor instance, ship and su bma rinehulls, and atomic reactor vessels.F a b r i c a t i o n o f T e s t S p e c i m e n s

T he test specimens were fabr icatedby steel fabricators according to norma l p rac t ice s and procedures . A WSspecifications were followed in thefabr ication. Sub merged -arc weldingmethod was used for all specimens.Pertinent welding data have beensummar ized in T ab le 1.

T he first four specim ens were fabricated from universal-mill platesthatis, the component plates in the weldedshapes were used with as-rolled edges.T he rema ining specimens were fabr icated from flame-cut plates, obtainedby flame-cutting the component platesfrom larger base metal pla tes . Specimens of the 15H290 shape were fabric a te d in b o th A S T M A 3 6 an d A 4 4 1steel for comparative tests. In addition, both a fillet weld (partial penetrati on ) and a groove weld (full pen etra t ion) were used. For the heavyshape 2 3 H 6 8 1 , only one specimen ofA 36 steel with fillet welds was fabricated.

T he welds in the 15 H290 shapeswere deposited in a symmetrical pattern as indicated in T able 1, to minimize the distor t ion of the members .T wo passes were used for the 15H2 90shapes with fillet welds, but sevenpasses were required for the groovewelds .

94-s M A R C H 1 9 7 0


3/13

T a b l e 1 -

T e s ts p e c i m eno.

12

3456789

1011

- W e l d i n g D a t a fo r F a b

n S h a p ed e s i g n a t i o n15H 29015H 290

15H 29015H 29015H 29015H 29015H 29015H 290PL16 X 2PL24 X 3Vs2 3 H 6 8 1

M a t e r i aA36A36

A 4 4 1A441A36A36A441A441A36A36A36

r i c a t i o nS t a t i cy i e l ds t r e s s ,k s i33.736.0

46.745.235.935.944.642.032.53 2 . 132.2

o f T e s t S p e c i m e n s

M a n u f a c t u r eo f p l a t e s

U MU M

U MU MFCFCFCFCFCFCFC

S i z e o ff l a n g ep l a t e ,b i n .

14 X 2>A14 X 2/a

14 X 2 > /,14 X 21/.,14 X 2 7 ,14 X 2 7214 X 2' /. ,14 X 2 7 ,

2 4 X 3 7 ,

S i z e o fw e bp l a t e ,

i n .10 x 17210 x 172

10 X i y *10 x 17210 X 17-j10 x iys10 x 17210 x 17216 X 2

T y p es u b

m e r g e d -a r cw e l d

F i l l e tG r o o v e

F i l l e tG r o o v eF i l l e tG r o o v eF i l l e tG r o o v e

F i l l e t

D e t a i lofw e l d ' '

A -B

c f>/

Ft tSAW>lIIT S6 0 * ^

T he 23H 681 shape was weldedusing an automatic beam welding machine with two tandem electrodes .T hus, i t was possible to weld simultaneously on both the left and theright side of the shape, each weldbeing deposited in one pass from oned-c electrode and one a-c electrodespaced 41/2 in. apa rt. A fter the firstflange and the web were joined together , the T-shape was turned ov erand the second flange was welded tothe T to form the f inal H-shape . Figure 4 shows the 23H681 specimen andthe beam welding machine used forthe fabr ication. A mo re deta iled account of the fabrication of the23H681 shape can be found in thel i terature .

sIt was specified that no straightening operations in any form should beused after the welding. In practice,such ope ra t ions may become necessary to fulfill straightness requirements . T he comm on m e thods usedfor straightening heavy welded shapesthat is, gagging or local heating bya flamewill change the residualstress distribution locally at the straightened section. T he residual stresses inthe unstra ightened par ts of the column will remain unchanged.

Procedure for Measurement ofResidual StressT he residual s tress distr ibution in athick plate is generally three-di-

uAW>J

rl 8 , 4

6 . 2

1.5

3,7

mensional with stresses in the longitudinal as well as the transverse direct ions . While the transverse stresseswill affect the yielding behavior of thedifferent fibers of the cross section, thelongitudinal s tresses are of pr imaryinterest for column strength . T hus, thetheoretica l methods normally employed for the prediction of columnbehavior and maximum strength ofcolumns consider the longitudinal re-

16,9,8,2,25 24 17

27,22,19,14 11.6 3

4.5.I2A20,21,28

18,23,26,1.7 10 15

sidual stresses only.When only longitudinal stresses aretaken into account in the theories forcolumn strength predic tion, i t is infact more relevant in the residualstress measurement to consider theapparent longitudinal s tresses as obta ined directly f rom the measured released strain in the longitudinal direction, rather than to separate the influence of the stresses in all three direc-

SECTIQNING-Top S ur f ace R ead ing

S t ra igh t L ineApproximation

S LICING

S t r a i g h t L i n eApproxim at ionFrom Sectioning

- S h a d e d A re a E q u a l s F in a lR e s i d u a l S t r e s s D i s t r i b u t i o nFi g. 5Principle o f t h e s e c t i o n i n g m e t h o d f o r r e s i d u a l s t r e s s m e a s u r e m e n t s

Bot t om S ur f ace R ead ing

Addi t iona l S t ressFrom Slicing

WELDING R E S E A R C H S U P P L E M E N T 95-s


4/13

- 4 0- 2 0

KSI 02 04 06 0

k f\ rr*>^/ /\ t

\ \KSI

60 4 0 2 0 0 - 2 0 - 4 0KSI

6 0 4 0 2 0 0 - 2 0 - 4 0I i I i I ' I i I ' I

6 04 020

KSI 0- 2 0- 4 0

- \ \J \V^

N e a r S u r f a c eF a r S u r f a c e

Fig. 6Residual s t r e s s e s i n a we l d e d s h a p e 15H290univers a l -m i l l p l a t e s , A36 S t e e l , y , i n . f i l l e t w e l d s

Near Sur faceF a r S u r f a c e

Fi g. 7Residual s t r e s s e ss a l -m i l l p l a t e s , A36 S t e e l , i n a w e l d e d s h a p e 15H 2 90-"/ m i n . g r oo v e w e l d s - Uni ver -

t ions . 0 T his is because the effect oftransverse stresses on the measuredreleased strain in a sectioningprocedure is somewhat similar to thaton the yielding behavior of a columnunder load . T hus , the re a re two approximations, the error of which tendto cancel each other. W ith thisprocedure of trea ting the measurements as one-dimensional, and whenthe transverse stresses are reasonablysmallthat is, less tha n 5 0 % of thelongitudinal stressthe relative errorin the applied stress to cause yield in af iber is less than 2 0 % . Fo r the behavior of the complete cross section, therelative error will be considerably less.A nother imp ortant fea ture of theresidual stress distribution in thickcom pone nts is that the longitudinalstresses can be expected to vary significantly through the thickness of thecomponents . T he me thod for measurement of residual stresses musttake into account this var ia t ion.T he procedu re used for the measurements was a sectioning method,

involving longitudinal saw cuts bothacross the width and through thethickness of the com pone nts . T hemethod is basically similar to the sect ioning method used by Kalakoutsky

in 1888 for the measurement of residual stresses in steel cylinders. 10 T h etechnique of the sectioning method asapplied to heavy welded shapes wasdeveloped in the Lehigh research prog r a m . 1 1Gage points were first laid outaround the specimen (Fig . 5) , andreadings were obtained using a 10 in.Whi t temore mechanica l extensometer.T he specimen was then cut into elements conta ining one or more gagepoints on each surface ( sectioni ng ) ,and new measurements were made .T he re leased stra ins at both surfacesof the elements could be evaluatedfrom these measurements . I f the ra t ioof width to length and width to thickness of the elements is such thatbeam-type action will occur, thethrough-thickness var ia t ion of s tra insreleased in the sectioning (s(,,.t.) willapproximate a stra ight l ine goingthrough the data points obta ined onthe surfaces . M easurem ents have indicated that the straight-line assumptionis reasonable for the geometry of elements used, and this assumption is thebasis for the evaluation procedure .T he sectioning was then con tinuedto obtain the actual variation of residual s tress through thickness . A f ter the

first set of saw cuts were m ade ( sect ioning ) , addit ional gage points werelaid out along the sides of the elements . New readings were taken bythe extensometer, followed by sawingthe elements into strips across thethickness ( s l ic in g ) . Fro m extensometer readings before and after theslicing, additional strains, esli ., wereobta ined. T hese stra ins are super imposed upon the strains from the sectioning to furnish the total strain varia t ion (see Fig . 5) . A ssuming that a l lresidual strains have been released,the residual stress may be obtainedfrom the re la t ionship:

E **,


5/13

K S I60 40 20 0'

K S I

fl

u

6 0 -N e a r S u r f a c eF ar Sur f ace

4 02 0

K S I 0

F ig . 8Residual s t r e s s e s i n a w e l d e d s h a p e 15H290univers a l - m il l p l a t e s , A 44 1 S t e e l ,*/ , i n . f i l le t w e l d s

- 2 0- 4 0 -

N e a r S u r f a c eF a r S u r f a c e

F ig . 9Residual s t r e s s e s i n a w e ld e d s h a p e 15H290univers a l - m il l p l a t e s , A 441 S t e e l , U / M i n . g r oo v e w e l d s

the error sources in the measurementswas carried out in connection with theinvest igat ion .12T he exper imen tal work involved inthe method is enormous, both withrespect to the required number ofgage point readings and the necessarysawing operations . For example , thenumber of gage readings involved inthe measurements on a 24 x 31 / : ; in .plate (see Figs. 17 and 19) is morethan 5000; the sawing was done on a

band saw and required a net machinetime of the order of 100 hr.Test Re sults

T he residual s tress distr ibutions obtained from sectioning of the four15H29 0 specimens fabr icated f romuniversal-mill plates are given in Figs.6 through 9 . T he curves shown referto the results obtained on both surfaces of the shape components in thesectioningthat is, before slicing. T he

F ig . 10Two-dimensional v a r i a t i o n o fr e s i d u a l s t r e s s i n a w e l d e d s h a p e 15H 290 u n i v e r s a l - m i ll p l a t e s , A36 S t e e l ,Va i n .f i l l e t w e l d sW E L D I N G R E S E A R C H S U P P L E M E N T 97-s


6/13

- 4 0- 2 0

KSI 0 Xn n- rw

T T

4 4

KSI6 0 4 0 2 0 0 -2 0 -4 0I I I 'I I i

KSI6 0 4 0 2 0 20-40

Near SurfaceFar Sur f ace

F ig . 11Residual s t r e s s e s i n a w e l d e d s h a p e 1 5H 2 90 -c u t p l a t e s , A36 S t e e l ,7 i n . f i l l e t w e l d s - f l a m e -

Near SurfaceFar Sur f ace

- 4 0 LF ig . 12Residual s t r e s s e s i n a w e l d e d s h a p e 1 5H 2 9 0 fl a m e -c u t p l a t e s , A 3 6 S t e e l , "/, i n . g r oo v e w e l d s

shapes in Figs . 6 and 7 are of A 36steel, differing only in the type ofweld. Figures 8 and 9 show a similarcomp arison for A 441 steel .T he results of a complete sectioningand slicing procedure from one of thespecim ens is given in Fig. 10. T heresidual stress distribution is represented in the form of an iso-stressdiag ram , that is, con tou r lines forconstant s tress .For all four shapes of universal-millplates, the stresses at the flange tipsare in compression, and of a relativelyhigh magnitud e . T he average stress a tthe flange tip varies between 16 and24 ksi for the four shapes.Figures 11 through 14 show theresidual stresses as measured in thefour 15H 290 specimens fabr icatedfrom flame-cut plates. A gain, thecurves in the diagrams correspond tothe measurements obta ined on bothsurfaces of the components in the sect ioning test . Each specimen made offlame-cut plates in Figs. 11 through 14corresponds to a s imilar specimenmade of universal-mill plates (Figs. 6through 9) so the distr ibutions may becompared directly . Instead of re la t ively high compressive stresses as in the

universal-mill plates, there are veryhigh tensile stresses at the flange tipsof the shapes made of flame-cutplates.Figure 15 gives the results obtainedfrom sectioning and slicing of the15H290 specimen m ade of A 36 steel ,and with fillet welds. A s may be seenfrom the contour lines, there are steepgradients in the weld region and at theflame-cut edges.A s will be discussed further in thenext section, it is obv ious from theresults on the 15H290 specimens thatthe initial stresses existing in the component plates prior to welding are ofgreat impor tance . T he re fore , measurements on the component pla tes,before welding, were included in thestudy of the 23H681 shape . Figures16 and 17 give the average residualstress through the thickness for a 16 x2 in. plate, taken from the same basemetal as the web plate of the shape,and a 24 x 3 1 / . , in . pla te , correspon ding to the flange plates of the shape.T he stresses at the plate edges are intension with a maximum of approximately 50 ksi at the flame-cut surface. T he tensile stresses are balancedby compressive stresses in the center

of the plates.A comp lete pic ture of the actualdistribution of longitudinal residualstresses in the thick plates can beobtained only from a study of thetwo-dimensional var ia t ion of residualstresses over the cross section. Figures18 and 19 show iso-stress diagrams asobtained from sectioning and slicingof the 1 6 x 2 in. and the 24 x 31 / , , in.pla tes, respectively . T he var ia t ionthrough the thickness amounts to approx ima tely 12 ksi for the 16 x 2 in.plate and 15 ksi for the 24 x 3V 2 in.pla te . T hus, a l though the averagestress in the center part of the plates iscomp ressive (see Figs. 16 and 1 7) ,the actual stress in the interior istensile . Similar ly, while the maximumcompressive stress in the average diagram of the 24 x 3 1 / . , in. plate is 11 ksi , and true maximu m com pression a t any measured point throughthe cross section is actually 22 ksi .F igure 20 shows schematically theaddition of stresses during welding ofthe 23H6 81 shape . T he d is t ribu t ionsshown are those measured in separatespecimens of loose plates and shapes,so there is no true algebraic additionin the diagram s. T he addit ional residual stresses resulting from the welding98-s M A R C H 1 9 7 0


7/13


8/13

Fig. 15Two-dimensional v a r i a t i o n, o f r e s i d u a l s t r e s s i n a we l d e d s h a p e15H290flame-cut p l a t e s , A36 S t e e l , / , i n . f i l l e t w e l d sare of a smaller absolute and relativemagnitude in heavy pla tes and shapes .T his could also be predicted since theratio of weld area to that of basemetal is smaller for heavy practicalshapes . T hus, proportionate ly , there isa smaller heat input for shapes of

thick plates than there is for lightplates, and so heavy shapes would beexpected to conta in compressive welding residual stresses of a smaller magnitude .T he above finding also mea ns thatthe initial residual stresses existing in

KSI 0

KSI 0

PL 2 4 x 3 (FC)F ig .17Residual s t r e s s e s i n a 24 x 3 , i n . f la m e - c u t p l a t e

PLI6x2(FC)F ig .16Residual s t r e s s e s i n a 16 x 2 i n .f l a m e - c u t p l a t ethe component plates before weldingare of greater importance for heavyshapes . T his conclusion is obviousfrom the results obtained on the15H290 when comparing the residualstresses in the shapes made from universal-mill plates (Fig s. 6- 9) andfrom f lame-cut pla tes (Figs . 11-14) .T he only difference between the twosets of specimens is in the manufacturing procedure for the plates, so thecomparison reflects the effect of thisvariable only.A ll flanges of the 1 5H 290 s hapesmade of universal-mill plates show asimilar kind of distribution with highcompressive stresses at the flange tips.A s can be seen in Fig . 10, the compressive stresses are at the yield pointlevel in the flange tip co rners . T his isin accordance with theoretical predictions of residual stresses in universal-mill plates based upon the heat flow incooling. ,: t A ccording to the predictions, the residual stresses tend toincrease with increasing size of themember. For heavy plates subjectedto free cooling, the cooling residualstresses as predicted are in high compression at the plate edges and mayapproach the yield point for someplates.

Flame-cut plates show a reverseddistribution with high tensile stressesat the flame-cut edge, balanced bycompressive stresses in the center ofthe plate. T he stress at the flame-cutsurface is larger than the yield pointof the parent m ater ia l . T his is possiblebecause the mechanical properties ofthe material close to the flame-cutsurface have been increased from therapid cooling after cutt ing. A notherfactor which influences the conditionsis the three-dim ension al stress state .S ince the local stresses in the heatedmaterial normally are in high tension

100-s M A R C H 1 9 7 0


9/13

that is, of equal sign in all three directionsthe material can sustain ahigher stress than the yield point in auni-axial tension test.B asically, these conditio ns prevail

also in the weld regions in the weldedshapes; this can explain the high tension stresses invariably experienced inthe welds. For the weld, however, theweld area contains a mixture of elec

trode and base metal . T he m echanicalproperties of electrodes used for welding structura l carbo n steels norm allyhave a higher strength than the basemeta l. T hu s, there are three effects

o1

oCO

o o 60

KSI

2020

oO J

oO J

o

oO J1 ocooCM O*

oCD

/>/)


10/13

which explain the occurrance of highstresses in the weld areas:1. Increased s trength of ma terialdue to the electrode material .2. Increased strength due to thecooling rate .3. T hree-dimensional s t resses .T ension tes ts of specimens containing weld metal have indicated a yields trength of 50- 55 ksi for A 7 s teelthat is, abou t 50 % higher than theyield strength of the base metal . ' T heresidual stresses measured in the weldarea of the heavy shapes are all of thisorder of magnitude.T he weld type is not a majo r factorin the formation of residual stresses inth ick plates and shapes . A comp arisonof the distributions in Figs. 6 and 13.for fillet welds, and the correspondingdiagrams in Figs. 7, 9, 12, and 14 forgroove welds, indicates no significantdifference. T his is probably becausethe heat input in each weld pass is ofthe same order of magnitude for thefillet weld and the groove weld (seeT able 1) . S imilar resul ts were obtained previously for small and medi-

- 4 0 h -

um-size plates. 'T he residual stress distributions asmea sured in the shapes of A 36 steeland of A 441 steel are similar, indicating that the type of steel has no greatinfluence on residual stress. Since theeffect of residual stress on columnstrength is dependent on the ratio ofresidual stress to yield stress while atthe same t ime the magnitude of theresidual stresses are not much affectedby the type of material, it can beexpected that columns made of h igh-strength steels are stronger than thosemade of A 36 s teel , a lso when compared on a non-dimensional basis .T his has been confirmed previously byresidual s t ress measurements andcolumn tests on small and medium-size welded shapes. 1 4T he geometry of the p lates andshapes is one of the major variablesaffecting the residual stress distribution. A s discussed above, the contrib ution of residual stress due to welding isgreatly dependent on the size of thecomponentsthat is, whether theshape is light, medium-size, or heavy .

KSI6 0 4 0 2 0 2 0 - 4 0

N e a r S u r f a c eFar Surface

Fig. 21Residual s t r e s s e s i n a w e ld e d s h a p e 23H681flame-cut p l a t e s ,A 3 6 S t e e l , ' s i n . f i l l e t w e l d s

Figure 24 summarizes the results ofres idual s t ress measurements made ona number of welded shapes us ingflame-cut plates, ranging from the7H28 shape to the 23H681 shape .Generally, the residual stresses due towelding decrease with increasing sizeof the member.A noth er feature of residual stressesin thick plates is the variation throughthe th ickness of the compone nts . R esul ts of measurements including bothsectioning and slicing have shown thatthe variation in thick plates is consid-eiable . For ins tance, the nonsymmet-rical application of welds on a flangeplate results in high tensile stresses onthe welded side, whereas the stress atthe opposite side often is compressive.It is believed that the large variation of stress across the thickness forthe component p lates 24 x 31 / . , in.and 16 x 2 in. (F igs. 18 and 19)remains from the cooling after rollingof the base metal . T heoret ical predictions of cooling stresses in rolledplates of a similar size show a distribution of stress through thickness ofthe plate very close to that measuredin the center of these flame-cutplates. 1: i A lso , the variat ion across thewidth for the center portion of theplates appears to be influenced by theinitial stresses from the cooling afterrolling of the base metal plates.If there were no initial stresses before flame-cutting of the plates, thestress distribution in the center portions of the plates would approximatea plane, except for the heated andyielded regions at the flame-cut edges,where tensile stresses are introduced.T he measured dis tr ibut ions , however,show a variation across the platewidth of 10 ksi and 13 ksi for thecenter portions of the 16 x 2 in. and24 xV/. in . p lates , respect ively . T heactual type of distribution in the center of the plates, with higher compressive stresses towards the edges, is consistent with predictions of coolingstresses in similar rolled plates.V T h i smeans that the stresses in thick flame-cut plates result from both the coolingafter rolling of the base metal and thelocal hea t input in the flame-cuttingprocess .T hus , the residual stress distributionin a fabricated member of thick component plates generally is a complexsuperimposed pattern resulting fromthe different stages of ma nufa ctur eand fabricat ion proc edures . T his is tosay that cooling residual stresses afterrolling, residual stresses from flame-cutting as well as the welding residuals tress and any other procedure employed in the manufacture and fabrication will influence the final residualstress distribution in a heavy welded.shape. However, since the welding

102-s M A R C H 1 9 70


11/13

residual stresses in heavy weldedshapes are small, except in the weldregion, residual stresses in such shapesmay be predic ted by determining theinitial stresses in the component platesfrom existing data or from measurements and est imating the addit ionalresidual stresses due to weldin g. T hisis a further simplification of thefinding obtained previously for smallershapes that the residual stresses in awelded shape can be predicted from aknowledge of the residual stress distr ibution in the separate componentpla tes with simulated weld beads . 7T he re la t ionship between init ial residual stresses in component plates andwelding residual stresses implies thatefforts to limit residual stresses inheavy welded shapes should be directed towards the manufacture of thecompon ent p la te s . T hus , by us ingflame-cut plates in heavy weldedshapes, there is a prospect for anincrease in strength when comparedwith lighter members at the sameslenderness ra t io . On the other hand,the column strength of heavy columns

made of universal-mill plates with as-rolled edges may be lower, in particular for A 36 steel and at higher slenderness ra t ios . 16For heavy welded columns made offlame-cut plates, there is even a possibility that such welded shapes may bestronger than their rolled counterpar ts . F igure 25 shows the measureddistribution of residual stress in arolled shape 14W r426 . 17 T h e no mi nal strength characteristics of thisshape fall in between those of the twowelded shapes 15H290 and 23H681studied here . T he stresses are veryhigh, of a similar distribution andorder of magnitude as in the weldedshape 15H290 made of universal-millplates.T he favorable indications for thestrength of heavy welded columnsmade of flame-cut plates is in contrastwith previously published results forsmall and medium-size welded shapesof universal-mill plates which provedto be weaker than their rolled counterpar ts . 6 However, since the presentbasis for the implications on heavy

column strength is limited to the fewspecimens studied, the sta tements arepreliminary; the general conclusionswill be developed further when thecurrent research program is completed.Conclus ions

T he purpose of this s tudy was toinvestigate exper imentally the magnitude and distribution of residual stressin heavy column shapes built up bywelding plates with thickness in therange of 1V 2 to 3V2 in . T wo shapeswere investigateda 15H290 shapeand a 23H681 shape . For the smallershape comparative tests were conducted to study the influence of the manufacture of plates, weld type, and yieldstrength of the mater ia l . T he specimens referred to in this paper constitute the first phase of a major research program concerning residualstresses in thick welded plates.

B ased on the results of this firstphase of the program, the followingconclusions can be sta ted:

Fig.22Two-dimensional v a r i a t i o n o f r e s i d u a l s t r e s s in a w e l d e d s h a p e 23H681flame-cut p l a t e s , A 3 6 S t e e l ,Va i n . f i l l e t w e l d sW E L D I N G R E S E A R C H S U P P L E M E N T ' 103-s


12/13

20KSI 0 -

2 0 -

PLI0x /2 ( U M)

- 4 0-20'KSI 0

2 04 06 0

-^ ^ ^ ^ 5 ^ M* ^ D

- \ /

6 0 G ro o v e W e l d

- 4 0- 2 0

KSI 02 04 06 0

-

D o u b l e V W e l dF i g . 23Residual s t r e s s e s i n a 10 x 1/,i n . u n i v e r s a l - m i l l p l a t e

1. T he residual stress distribution ina fabricated member of thick component plates generally is a complexsuperimposed pattern resulting fromthe different stages of manufactureand fabrication proceduresthat is,cooling residual stresses from the rolling of the base plates, residualstresses from flame-cutting the plates,as well as the residual stressesfrom thewelding process, all influence the finalresidual stress distribution in a weldedshape.2 . T he welding residual stresses inportions of the cross section otherthan the weld area tend to decreasewith increasing size of the member,probably because the weld area, andconsequently, the heat input, is relatively smaller in heavy plates andshapes as compared to l ighter members .

3. T he resul ts of comparat ive meas-surements on 15H290 shapes indicatethat the weld type (penetration) isnot a major factor in the formation ofresidual stresses in heavy plates andshapes, as long as the heat input is notdrastically changed.4 . T he residual stress distributionsas measured in shapes of A S T M A 36steel and of A 441 steel are similar,indicating that the yield strength of

7 H 2 8

I 0 H 6 2

I 2 H 7 9

I 4H 202

I5H29023H68I

S t r e s s S c a l e ( k s i ) :60 40 20 0 -20 -40

23H 68I I5H290 I4H 202 I2H79 I0H62 7H28

F i g . 24Residual s t r e s s e s i n w e l d e d H - s h a p e s o f d i f f e r e n t g e o m e t r y . F l a m e - c u tp l a t e s , A 3 6 S t e e l , f i l l e t w e l d s . ( R e s u l t s f or 1 2 H 7 9 a n d 1 4 H 2 0 2 s h a p e s f r o mM c F a l l s a n d Tall1")

the steel has no great influence onresidual stresses in heavy shapes.5. T he magni tude and dis t r ibut ionof initial residual stresses dependsgreatly on the manufacturing procedurethat is, universal-mill platesor flame-cut plates.6. T he geom etry of the plates andshapes is one of the major variablesaffecting the residual stress distribution. More specifically, the residualstresses will depend upon whether theshape is l ight, medium-sized or heavy(effect of size of cross section) orwhether the plates are narrow, medium or wide (effect of width to thickness ra t io ) .7. R esidual stresses due to coolingafter rolling of a plate normally arecompressive at the edges, balanced bytension stresses in the center portion

of the plate. Measurements on fourwelded H-shapes , 15H290, containinguniversal-mill plates indicate that theinitial stresses in the flange plates arevery high. T his conform s to the observation made previously from the results of theoretical predictions that thecooling residual stresses in rolledplates tend to increase with increasingsize of the plate.8. R esidual stresses due to flame-cutting are in high tension at theflame-cut edge due to the local heatinput. For the flame-cut plates

studied, the stress at the burned surface is about 50% above the yieldpoint of the base metal . T he tensilestresses are balanced by compressionin the center part of the plate; the

distribution of residual stresses in thecenter of a heavy flame-cut plate isdependent also on the initial stressesdue to rolling.9. T he variation of residual stressacross the thickness of plates morethan 1 in. thick can be co nsidera ble.S uch variation will result in a platewith welds deposited on one side ofthe plate; also, a significant differencecan be expected between stresses inthe surface and the interior of heavyrolled plate components .

10. T he relat ionship between theinitial residual stresses in componentplates and the welding residual stressesimplies that the welding stresses inthick welded plates and shapes are ofless importance than the initial stressesexisting prior to welding. Efforts tolimit residual stresses in heavy weldedshapes should be directed towards themanufacture of the component plates .

11 . S ince the welding residualstresses in heavy welded shapes aresmall, except in the weld region, residual stresses in such shapes may bepredicted by determining the initialstresses in the component plates fromexisting data or from measurements,and estimating the additional residualstresses due to welding . T his is afurther simplification of the findingobtained previously for smaller shapesthat the residual stresses in a weldedshape can be predicted from a knowledge of the residual stress distributionin the separate component plates withsimulated weld beads .7

T h e e f f e c t o f r e s i d u a l s t re s s e s o n t h e

104-s M A R C H 1 9 7 0


13/13

load-carrying behavior and strength ofheavy columns will be discussed further in the future paper.16 B riefly, byusing flame-cut plates in heavy weldedshapes, there is a prospect for anincrease in strength when comparedwith lighter members at the sameslenderness ratio. Under certain circum stanc es, there is even a possibilitythat such welded shapes may be stronger than their rolled counterpar ts; thisis the opposite relationship as compared to the situation for small andmedium-sized shapes. On the otherhand, heavy welded columns of universal-mill plates and A 36 steel areexpected to show a lower columnstrength, especially for columns of amedium to high slenderness ra t io .

AcknowledgmentsT his pape r presents the results of an

experimental study of residual stressesin heavy welded shapes . T he investigation is the first phase of a majorresearch program designed to determine the residual stresses in thickwelded plates and shapes and to relatethis to the stability under load ofcompression members .T he investigation was conducted a tFr i tz Engineer ing Laboratory, LehighUniversi ty, B ethlehem, Pa . T he National Sc ience Foundation sponsorsthe current research program . A pilotstudy, included in this paper, was part

of a project sponsored jointly by thePennsylvania Department of Highways, the B ureau of Public R oads ofthe U . S . Depa r tment o f C omm erce ,the C o lumn R esearch Counc i l , theA merican Insti tute of S teel C onstruct ion, and the A merican Iron and S teelInsti tute . T he specimens were fabricated by the B ethlehem S teel C orp. ,and thanks are due to that corporation and its personnel who assisted inthe design and fabrication of the specimens. T he guidance of T ask G roup1 of the C olum n R esearch C ouncil ,under the chairm anship of John A .Gilligan, is gratefully ack now ledged.

Special thanks are due to Lynn S.B eedle, Director of Fr i tz Engineer ingLaboratory, for his advice and encouragement throughout the program.Fiorello R . Estuar and Will iam C .C ranston carr ied out the exper imentson the 15H2 90 shape and C har les R .Nordquist assisted in the measurements on the 23H681 shape and itscomp onent pla tes . T heir assistance issincerely appreciated.T hank s are a lso due to Kenneth R .

Harpel , laboratory foreman, and hisstaff for the preparation of the testspecimens, to Mrs . S haron B alogh forthe preparation of the drawings, and

l4W r426

.

V

KSI2 0 10 0 10 - 2 0 - 3 0T T

N e a r S u r f a c eFar Surface

Fig. 25Residual s t r e s s e s i n a h o t - r o l l e d s h a p e 14\A F426, A7 s t e e l

to Miss Joanne Mies and Mrs . LindaWelsch for their care in typing themanuscr ipt .References

1. H u b e r . A . W . . a n d B e e d l e , L . S . .R e s i d u a l S t r e s s e s a n d t h e C o m p r e s s i v eP r o p e r t i e s of S t e e l , W ELDING JOUR NA L, 33(12), R es e a r c h S u p p l . . 5 8 9- s to 6 1 4 -s , ( 1 9 5 4 ) .2. B e e d le . L . S . , a n d T a l l . L . . B a s i cC o l u m n S t r e n g t h , T r a n s . A S C E , V o l. 1 27,P a r t I I , pp . 138 to 179, 1962 .3. N i s h i n o , F . , U e d a . Y . , a n d T a l l . L . ,E x p e r i m e n t a l I n v e s t i g a t i o n of th e B u c k l i n gof P l a t e s w i t h R e s i d u a l S t r e s s e s , A S T MS T P N o . 4 19 . T e s t M e t h o d s f o r C o m p r e s s i o n T e s t i n g , A u g u s t 1 96 7.4. U e d a , Y . , a n d T a l l , L . , I n e l a s t i cB u c k l i n g o f P l a t e s w i t h R e s i d u a l S t r e s s e s ,P u b l i c a t i o n s , I n t e r n a t i o n a l A s s o c i a ti o n f o rB r i d g e a n d S t r u c t u r a l E n g i n e e r i n g , V o l.27. 1967.5. N a g a r a j a R a o . N . R . . a n d T a l l. L . .R e s i d u a l S t r e s s e s i n W e l d e d P l a t e s .W ELDING JOUR NA L, 4 0 ( 1 0 ) , R e s e a r c h S u p p l . ,468-s to 480-s (1961) .6 . E s t u a r , F . R . . a n d T a l l . L . . E x p e r i m e n t a l I n v e s t i g a t i o n o f Built-Up C o l um n s .

W E L D I N G J O U R N A L . 4 2 ( 4 ) , Ibid. 164-s to176-s (1963) .7. N a g a r a j a R a o , N . R . , E s t u a r , F . R . ,a n d T a l l , L . , R e s i d u a l S t r e s s e s i n W e l d e dS h a p e s . WELDING JOURNA L, 43 (7) , Ibid.,295-s to 396-s (1964).8. A l p s te n , G . A . , R e s i d u a l S t r e s s e s i na H e a v y W e l d e d S h a p e 2 3 H 6 8 1 , F r i t z

L a b o r a t o r y R e p o r t N o . 33 7. 9 , i n p r e p a r a t i o n .9 . A l p s t e n . G . A . . T h r e e - D i m e n s i o n a lR e s i d ua l S t r e s se s a n d C o l u m n S t r e n g t h ,F r i t z L a b o r a t o r y R e p o r t N o . 3 3 7. 2 0, i np r e p a r a t i o n .10. K a l a k o u t s k y , N . , T h e S t u d y o f I n t e r n a l S t r e s s e s i n C a s t I r o n a n d S t e e l ,Lo n d o n . 1 8 8 8 .11. E s t u a r , F . R . . W e l d i n g R e s i d u a lS t r e s s e s a n d t h e S t r e n g t h o f H e a v y C o l u m nS h a p e s , P h . D . D i s s e r t a t i o n . L e h i g h U n i v e r s i t y , A u g . 1 96 6 . ( U n i v e r s i t y M i c r o f i l m s ,Inc . . A n n A r b o r , M i c h i g a n ) .12. B r o z z e t t i . J . . a n d A l p s t e n . G . A . .A c c u r a c y o f t h e S e c t i o n i n g M e t h o d f o rR e s i du a l S t r e s s M e a s u r e m e n t s , F r i t z L a b o r a t o r y R e p o r t N o . 337 .11 . in p r e p a r a t i o n .13. A l p st e n , G . A . . T h e r m a l R e s i d u a lS t r e s s e s i n H o t - R o l l e d S t e e l M e m b e r s .F r i t z L a b o r a t o r y R e p o r t N o . 337 .3 , D e c e m ber , 1968 .14. K i s h i m a . Y . . A l p s t e n , G . A . , a n dT a l l , L . , C o l u m n S t r e n g t h o f WeldedS h a p e s U s i n g F l a m e - C u t P l a t e s . F r i t z L a b o r a t o r y R e p o r t N o . 3 21 . 4. in p r e p a r a t i o n .15. M c F a l l s , R . K . , a n d T a l l , L . , AS t u d v of W e l d e d C o l u m n s M a n u f a c t u r e df ro m F l a m e - C u t P l a t e s . WELDING JOURNA L,4 8 ( 4 ) , R es e a r c h S u p p l . , 1 41 -s to 151-s(1969) .16. A l p s t e n . G . A . , a n d T a l l , L . , C o l u m n S t r e n g t h o f H e a v y S h a p e s -A P r o g r e s s R e p o r t , F r i t z L a b o r a t o r y R e p o r t N o .3 3 7. 1 6. i n p r e p a r a t i o n .17. F u j i t a , Y . , T h e M a g n i t u d e a n d D i s t r i b u t i o n o f Residual S t r e s s e s , F r i t z L a b o r a t o r y R e p o r t N o . 2 2 0A . 2 0 , M a y , 1 95 5.

W E L D I N G R E S E A R C H S U P P L E M E N T | 105-s

Documents

Stress Residual en Vigas Armadas