Residual Stresses in Heavy Welded Shapes January 1969 (70-6)

7/28/2019 Residual Stresses in Heavy Welded Shapes January 1969 (70-6)

1/67

Lehigh University

Lehigh Preserve

Fritz Laboratory Reports Civil and Environmental Engineering

1-1-1969

Residual stresses in heavy welded shapes, January1969 (70-6)

G. AlpstenL. Tall

Follow this and additional works at: hp://preserve.lehigh.edu/engr-civil-environmental-fritz-lab-reports

is Technical Report is brought to you for free and open access by the Civil and Environmental Engineering at Lehigh Preserve. It has been accepted

for inclusion in Fritz Laboratory Reports by an authorized administrator of Lehigh Preserve. For more information, please contact

[email protected].

Recommended CitationAlpsten, G. and Tall, L., "Residual stresses in heavy welded shapes, January 1969 (70-6)" (1969).Fritz Laboratory Reports. Paper 334.hp://preserve.lehigh.edu/engr-civil-environmental-fritz-lab-reports/334
http://preserve.lehigh.edu/?utm_source=preserve.lehigh.edu%2Fengr-civil-environmental-fritz-lab-reports%2F334&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://preserve.lehigh.edu/engr-civil-environmental-fritz-lab-reports?utm_source=preserve.lehigh.edu%2Fengr-civil-environmental-fritz-lab-reports%2F334&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://preserve.lehigh.edu/engr-civil-environmental?utm_source=preserve.lehigh.edu%2Fengr-civil-environmental-fritz-lab-reports%2F334&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://preserve.lehigh.edu/engr-civil-environmental-fritz-lab-reports?utm_source=preserve.lehigh.edu%2Fengr-civil-environmental-fritz-lab-reports%2F334&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://preserve.lehigh.edu/engr-civil-environmental-fritz-lab-reports?utm_source=preserve.lehigh.edu%2Fengr-civil-environmental-fritz-lab-reports%2F334&utm_medium=PDF&utm_campaign=PDFCoverPagesmailto:[email protected]:[email protected]://preserve.lehigh.edu/engr-civil-environmental-fritz-lab-reports/334?utm_source=preserve.lehigh.edu%2Fengr-civil-environmental-fritz-lab-reports%2F334&utm_medium=PDF&utm_campaign=PDFCoverPagesmailto:[email protected]://preserve.lehigh.edu/engr-civil-environmental-fritz-lab-reports/334?utm_source=preserve.lehigh.edu%2Fengr-civil-environmental-fritz-lab-reports%2F334&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://preserve.lehigh.edu/engr-civil-environmental-fritz-lab-reports?utm_source=preserve.lehigh.edu%2Fengr-civil-environmental-fritz-lab-reports%2F334&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://preserve.lehigh.edu/engr-civil-environmental-fritz-lab-reports?utm_source=preserve.lehigh.edu%2Fengr-civil-environmental-fritz-lab-reports%2F334&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://preserve.lehigh.edu/engr-civil-environmental?utm_source=preserve.lehigh.edu%2Fengr-civil-environmental-fritz-lab-reports%2F334&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://preserve.lehigh.edu/engr-civil-environmental-fritz-lab-reports?utm_source=preserve.lehigh.edu%2Fengr-civil-environmental-fritz-lab-reports%2F334&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://preserve.lehigh.edu/?utm_source=preserve.lehigh.edu%2Fengr-civil-environmental-fritz-lab-reports%2F334&utm_medium=PDF&utm_campaign=PDFCoverPages


2/67

Residual Stresses in Thick Welded Plates

RESIDUAL STRESSES INHEAVY WELDED SHAPES

byGoran A. Alpsten

Lambert Tall

January, 1969

Fritz Engineering Laboratory Report No. 337.12


3/67

Residual Stresses in Thick Welded Plates

RESIDUAL STRESSES IN HEAVY WELDED SHAPES

by

Goran A. Alpsten and Lambert Tall

Fritz Engineering LaboratoryDepartment of Civil EngineeringLehigh UniversityBethlehem, Pennsylvania

January, 1969

Fritz Laboratory Report No. 337.12


4/67

337.12

TABLE OF CONTENTS

ABSTRACT 1

INTRODUCTION 4

FABRICATION OF TEST SPECIMENS 8

PROCEDURE FOR MEASUREMENT OF RESIDUAL STRESS 10

TEST RESULTS 14

DISCUSSION OF RESULTS 19

CONCLUSIONS 28

ACKNOWLEDGMENTS 33

TABLES AND FIGURES 35

REFERENCES 62


5/67

337.12 -1

ABSTRACT

Residual st resses can have a significant influenceon the load-carrying behavior of st ructural s teel members subjectedto compressive loads. Previous experimental research on residualstresses and the strength of columns was related to small andmedium-size shapes. In today's large st ructures, increasinglyheavy shapes are employed. While heavy column shapes arebeing used extensively, very l i t t l e information has beenavailable on the res idual stresses and strength of such members.

This paper presents the results of the f i r s t phaseof a major investigation into the residual stresses in , and thebehavior of, thick plates and heavy shapes used in compressionmembers. The shapes considered in this in i t i a l stUdy are a15H290 shape and a 23H681 shape, as well as two loosecomponentplates , PL16x2 and PL24x3t. Fo r the smaller shape, comparativet es t s were carried out for d ifferent manufacturing conditions ofthe component plates (universal-mill and flame-cut plates) ,different weld type (penetration) and different yield strengthsof the material.

The results of residual s tress measurementscarried out in this f i r s t phase of the stUdy indicate that ,for heavy fabricated members,


6/67

337.12

- a l l phases of the manufacture and fabricationprocedure generally affect the formation ofresidual s t resses ,

-2

- the weld type (penetration) and the yield strengthof the s teel are not major factors in the formationof res idual s tresses ,

- the geometry of the plates and shapes is one ofthe important variables affecting the residuals tress magnitude and dist r ibut ion,

- the varia t ion of res idual s tress across the thicknessof plates more than one inch thick can beconsiderable,

- the welding residual stresses in portions of thecross section other than the weld area tend todecrease with increasing size of the member,probably because the weld area, and consequently,the heat input, is relat ively smaller in heavyplates and shapes as compared to l ight members,

- the in i t ia l stresses can be of a higher magnitudethan the welding residual st resses.

The relationship between in i t ia l residual stresses incomponent plates and welding residual stresses implies thateffor ts to l imi t the magnitude of residual stresses in heavy


7/67

337.12 -3

welded shapes should be directed towards the manufactureof the component plates . Thus, by using flame-cut platesin heavy welded shapes, there is a prospect for an increasein strength when compared with l ighter members at thesame slenderness ra t io . There is even a possibi l i tythat such welded shapes may be stronger than the i r rol ledcounterparts .


8/67

337.12 -4

INTRODUCTION

Only in the past two decades was it experimentallyand theore t ica l ly v er i f ied tha t res idual stresses are amajor influence on the strength of s tee l members incompression. I t was shown tha t they have a signif icant

(1 2) (3 4)effect on column strength ' and on the buckling of pla tes . 'I t is also known tha t residual s tresses can be of greatimportance in fat igue, br i t t l e fracture, and st ress corrosion.

The previous experimental research on residual st ressand the strength of welded s t ee l members was related to smalland medium-size shapes, tha t i s , shapes of components with athickness equal . (5 6 7)to or less than one lnch. " In today'slarge s t ruc tures , increasingly heavy shapes are used. Veryl i t t l e information has been available on the residual stressesand strength of heavy columns, yet they are used extensivelyin North America and elsewhere. Applications include thelower stor ies of multi-story buildings, major bridges, andlaunching gantries for rockets and space vehicles. Similarst ructural elements are used in ship and submarine hul l s ,and in vessels for atomic reactors .

Some heavy column shapes used in existing structuresare shown in Fig. 1 , which also i l lus t ra tes the dif ferent ways


9/67

337.12-5of designing a heavy column shape. Rolled shapes can be usedand ar>e available in sizes up to th e l4WF730 "jumbo" shapeof Fig. lao I f the str>ength of the available r>olled shape i sinsuff icient for> a par>ticular> application, plates can bewelded to a r>olled shape as shown in Figs. lb and c . A heavyshape can also be bui l t up fr>om welding together> thr>ee platesto for>m an H-shape, as illustr>ated in Fig. ld . Finally, abox-shape can be made fr>om welding together four> plates ,Fig. leo The plates used in a welded shape can be manufactur>edeither> as univer>sal-mill plates , that i s , r>olled to exact widthand used with as-r>olled edges, or> as flame-cut plates , that i s ,flame-cut fr>om a lar>ger> par>ent plate.

At pr>esent, the design of heavy columns does notdiffer> fr>om tha t of small and medium-size columns. The designcr>iter>ia pr>eviously developed for small and medium-size r>olled

(2)columns have been extr>apolated to include heavy member>s.While exper>ience has indicated that this leads to safe design,i t may not be a completely r>ational design method.

An extensive r>esear>ch pr>ogr>am is cUr>r>ently underwaya t Lehigh Univer>sity to study residual str>esses in heavywelded plates and shapes. (Heavy plates and shapes ar>edefined her>e as member>s with a thickness exceeding one inch).The specific objectives of the study ar>e to deter>mine themagnitude and distr>ibution of residual str>esses in thick welded


10/67

337.12 -6

plates by both experimental and theoretical means, and to relatethis to the stabi l i ty under compressive loads of st ructuralmembers.

Prior to the in i t ia t ion of the current study,cr i t i ca l information was lacking on the behavior of heavycolumns. The present status is i l lustra ted in Fig. 2 whichshows (a) the largest tes t shape so far used in a multi-storyframe t es t , (b ) the largest welded shape in a beam-columntes t , (c) the largest column shape, and (d) the largest stubcolumn shape. The shapes in Figs. 2e through g are examplesof the specimens in the current research program. These shapescompare to the heavy column shapes used in construction (seeFig. 1).

The paper presents some experimental resul tsobtained in the f i r s t phase of the investigation. While thespecimens in the overall program cover the complete rangeof dimensions which are pract ical in construction, that is ,plates ranging in size from 9 x t up to 24 x 6 and weldedH-shapes and a Box-shape ranging from a 7H28 to a 24Hl122shape, the specimens considered in this paper are the l5H290and 23H681 shapes (see Figs. 2d and e) as well as two asmanufactured plates, 16 x 2 and 24x st. Figure 3 shows acomparison between the 23H68l shape and the 7H28 shape,corresponding to the heaviest and l ightest welded specimens


11/67

337.12

tested so far in the Lehigh program.The results of the overall study will allow the

prediction of residual stresses in any plate and any weldedshape used in construction. The plates and shapes in theinvestigation wil l serve as reference data, and by knowingthe dimensions of components and the manufacturing andfabrication detai ls of a pract ical column, the distributionof res idual s tress can be predicted. The column strengthcan then be determined, including the effect of res idualst resses.

The study is of a fundamental nature and ofconsiderable importance in many areas; h o w e v e r ~ the presentapplication is the development of information which wil l beuseful in preparing design cri ter ia for heavy column membersas are used in construction. The findings obtained. for thebasic welded plates wil l be applicable also to other typesof st ructures, for instance, ship and submarine hulls , andatomic reactor v e s s e l ~ .

-7


12/67

337.12

FABRICATION OF TEST SPECIMENS

The t e s t specimens were fabricated by s t ee lfabricators according to normal pract ices and procedures.

-8

The AWS Specif icat ions were followed in the fabrication. Thesubmerged arc welding method was used for a l l specimens.Per t inent welding data have been summarized in Table 1 .

The f i r s t four specimens were fabricated fromuniversal -mi l l plates , tha t i s , the component plates in thewelded shapes were used with as-rol led edges. The remainingspecimens were fabricated from flame-cut pla tes , obtained byflame-cutting the component plates from larger parent plates .Specimens of the 15H290 shape were fabricated in both ASTMA36 and A441 s t ee l for comparative tes ts . In addi t ion, botha f i l l e t weld (par t ia l penetrat ion) and a groove weld ( fu l lpenetrat ion) were used. For the heavy shape 23H681 only onespecimen of A36 s teel with f i l l e t welds was fabricated.

The welds in the 15H290 shapes were depositedin a symmetrical pat tern as indicated in Table 1 , to minimizethe dis tort ion of the members. Two passes were used for the15H290 shapes with f i l l e t welds, but seven passes were requiredfor the groove welds.

The 23H681 shape was welded using an automaticbeam welding machine with two tandem electrodes. Thus, i t was


13/67

337.12 -9

possible to weld simultaneously on both the l e f t and the~ i g h t side of the shape, each weld being deposited in onepass f ~ o m one DC e l e c t ~ o d e and one AC e l e c t ~ o d e spaced 4tinches a p a ~ t . A f t e ~ the f i ~ s t flange and the web w e ~ e joined

t o g e t h e ~ , the T-shape was t u ~ n e d o v e ~ and the second flangewas welded to the T to f o ~ m the f ina l H-shape. F i g u ~ e 4shows the 23H68l specimen and the beam welding machine used

the f a b ~ i c a t i o n . A m o ~ e detailed account of th e f a b ~ i c a t i o nof the 23H68l shape can be found in Ref. 8.

I t was specified that no s t ~ a i g h t e n i n g o p e ~ a t i o n sin any f o ~ m should be used a f t e ~ the welding. In p ~ a c t i c e ,such o p e ~ a t i o n s may become n e c e s s a ~ y to fu l f i l l s t ~ a i g h t n e s s

~ e q u i ~ e m e n t s . The common methods used s t ~ a i g h t e n i n gheavy welded shapes, tha t i s , gagging local heating by aflame, wil l change the ~ e s i d u a l s t ~ e s s d i s t ~ i b u t i o n l o c a l l ya t the s t ~ a i g h t e n e d section. The ~ e s i d u a l s t ~ e s s e s in the

u n s t ~ a i g h t e n e d p a ~ t s of the column wil l ~ e m a i n unchanged.


14/67

337.12 -10

PROCEDURE FOR MEASUREMENT OF RESIDUAL STRESS

The residual s tress distribution in a thickplate is generally three-dimensional with stresses in thelongitudinal as well as the transverse directions. Whilethe transverse st resses will affect the yielding behaviorof the different fibers of th e cross section, the longitudinalstresses are of primary interes t for column strength. Thus,the theoret ical methods normally employed for the predictionof column behavior and maximum strength of columns considerthe longitudinal residual stresses only.

When only longitudinal stresses are taken intoaccount in the theories for column strength prediction, i ti s in fact more relevant in the residual stresS" measurementto consider the apparent longitudinal stresses as obtaineddirect ly from the measured released s train in the longitudinaldirection, rather than to separate the influence of the stressesin a l l three directions.(9) This is because the effect oftransverse st resses on the measured released strain in asectioning procedure i s somewhat similar to that on theyielding behavior of a column under load. Thus, there aretwo approximations, the error of which tend to cancel each

(9)other. With this procedure of treating the measurementsas one-dimensional, and when the transverse st resses are


15/67

337.12 -11

reasonably small, that is , less than 50% of the longitudinals t ress , the re la t ive error in the applied stress to causeyield in a f iber i s less than 20%. For the behavior ofthe complete cross section, the relat ive error wil l beconsiderably less .

Another important feature of the residual stressdistribution in thick components is that the longitudinalstresses can be expected to vary significantly through thethickness of the components. The method for measurementof residual s t resses must take into account this variation.

The procedure used for the measurements was asectioning method, involving longitudinal saw cuts bothacross the width and through the thickness of the components.The method is basical ly similar to the sectioning method usedby Kalakoutsky in 1888 for the measurement of residualstresses in s teel c y l i n d e r s ~ l O ~ h e technique of the sectioningmethod as applied to heavy welded shapes was developed in the

. (11)L e h ~ g h research program.Gage points were f i r s t laid out around the

specimen (see Fig. 5), and readings were obtained using aten inch Whittemore mechanical extensometer. The specimenwas then cut into elements containing one or more gagepoints on each surface ("sectioning") and new measurementswere made. The released strains a t both surfaces of the


16/67

337.12 -12

elements could be evaluated from these measurements. I fthe rat io of width to length and width to thickness of theelements is such that beam-type action wil l occur, thethrough-thickness variation of s t rains released in thesectioning (e t) wil l approximate a s t raight l inesec going through the data points obtained on the surfaces.Measurements have indicated that the straight- l ine assumptioni s reasonable for the geometry of elements used, and thisassumption is the basis for the evaluation procedure.

The sectioning was then continued to obtainthe actual variation of residual stress through thickness.After the f i r s t set of saw cuts were made ("sectioning"),additional gage points were laid out along the sides of theelements. New readings were taken by the extensometer,followed by sawing the elements into str ips across the thickness("sl icing"). From extensometer readings before and afterthe sl icing, addit ional strains, l' , were obtained. Theses lC .strains are superimposed upon the strains from the sectioningto furnish the t o t a l strain variat ion. See Fig. 5. Assumingthat a l l residual strains have been released, th e residualstress may be obtained from the relationship

= -E ( sect . + (1)


17/67

337.12 -1 3

The residual s t resses released in the slicingprocedure must be in equilibrium for each sect ional element.Thus, there is no contribution to the average stress throughthickness from these stresses. Consequently, the average residualstress obtained as the mean value of readings from bothsurfaces in the sectioning procedure is equal to the actualaverage stress .through the thickness.

The accuracy in the stress measurement is of theorder ofl ksi . An extensive study of the accuracy and theerror sources in the measurements was carried out in

. . h h ' . . (12)connectlon Wlt t e lnvestlgatl0n.The experimental work involved in the method is

enormous, both with respect to the required numb.er of gagepoint readings and the necessary sawing operations. Forexample, the number of gage readings involved in themeasurements on a 24" x 3t" plate (see Figs. 17 and 19) ismore than 5000; the sawing was done on a band saw and requireda net machine time of the order of 100 hours.


18/67

337.12

TEST RESULTS

The res idual s tress dis t r ibut ions obtainedfrom sectioning of the four 15H290 specimens fabricated fromuniversal-mill plates are given in Figs. 6 through 9. Thecurves shown re fe r to the resu l ts obtained on both surfacesof the shape components in the sectioning, tha t i s , beforesl ic ing. The shapes in Figs. 6 and 7 are of A36 s tee l ,differ ing only in the type of weld. Figures 8 and 9 show asimilar comparison for A441 s teel .

The resul ts of a complete sectioning and sl ic ingprocedure from one of the specimens is given in Fig. 10. Theres idual s tress distr ibution i s represented in the form ofan iso-s tress diagram, tha t i s , contour l ines for constants t ress .

For a l l four shapes of universal -mi l l plates ,the stresses a t the flange t ips are in compression, and ofa relat ively high magnitude. The average s tress a t theflange t ip varies between -16 ksi and -24 ksi for the fourshapes.

-14

Figures 11 through 14 show the res idual stressesas measured in the four 15H290 specimens fabricated fromflame-cut pla tes . Again, the curves in the diagrams correspondto the measurements obtained on both surfaces of the components


19/67

337.12

in the sectioning t es t . Each specimen made of flame-cutplates in Figs. 11 t h ~ o u g h 14 c o ~ ~ e s p o n d s to a s i m i l a ~specimen made of u n i v e ~ s a l - m i l l plates, Figs. 6 t h ~ o u g h 9,so the d i s t ~ i b u t i o n s may be compared direct ly. Instead ofrelat ively high compressive stresses, as in the universalmil l plates , there are very high tensi le stresses a t theflange t ips of the shapes made of flame-cut plates .

Figure 15 gives the resu l t s obtained fromsectioning and s l ic ing of the 15H290 specimen mqde of A36stee l , and with f i l l e t welds. As may be seen from thecontour l ines, there are steep s t ress gradients in the weldregion and a t the flame-cut edges.

As wil l be discussed further in the nextsect ion, i t i s obvious from the resul ts on the 15H290specimens that the in i t i a l stresses existing in the .componentplates p ~ i o r to welding are of great importance. T h e ~ e f o ~ e ,measurements on the component plates, before welding,wereincluded in the study of the 23H681 shape. Figures 16 and17 give the average residual stress through the thicknessfor a.16" x 2" pla te , taken from the same parent plate as theweb plate of the shape, and a 24" x 3*" plate, correspondingto the flange plates of the shape. The stresses a t the plateedges are in tension with a maximum of approximately 50 ksia t the flame-cut surface. The tensi le s t resses are balanced

-15


20/67

337.12

by compressive stresses in the center of the plates .A complete picture of the actual distribution

of longitudinal residual stresses in the thick plates canbe obtained only from a study of the two-dimensionalvaria t ion of residual stresses over the cross section.Figures 18 and 19 show iso-stress diagrams as obtained fromsectioning and sl icing of the 16 x 2 and th e 24 x 3t plates ,respectively. The varia t ion through the thickness amountsto approximately 12 ksi for the 16 x 2 plate and 15 ksifor the 24 x ~ plate . Thus, although the average st ress inthe center part of the pia tes is compressive (see Figs. 16

-16

and 17), the actual stress in the interior is tensi le .Similarly, while the maximum compressive stress in the averagediagram the 24 x ~ plate i s - l l k s i , the t rue maximumcompression a t any measured point through the cross section isactually -2 2 ksi .

Figure 20 shows schematically the addition ofstresses during welding of the 23H68l shape. The distributionsshown are those measured in separate specimens of loose pla tesand shapes, so there is no true algebraic addition in thediagrams. The additional residual stresses result ing fromth e welding were obtained from gage readings on the platesbefore and after welding. These welding residual stresses


21/67

337.12 -17

add to the in i t i a l s tresses exist ing prior to welding. I tshould be noted tha t the welding stresses could be measuredonly in the regions which remain elas t ic throughout thewelding and cooling af ter welding. In the parts of the crosssection where plas t ic deformations occurred a t any stage ofthe process the measured strain wil l contain a plast iccomponent and cannot directly be converted to s t ress .

While the i n i t i a l stresses in the flangeplates vary between approximately -11 ksi and 50 ksi , theadditional stresses as measured in the elas t ic part of theflanges nowhere exceed 2 ksi (Fig. 20b). For the web thecomparison between in i t i a l stresses and welding stresses i ssimilar , although the welding stresses are of a somewhatlarger magnitude than encountered in the f langes, that i s ,up to -13 ksi . Measured welding stresses in most points werein compression. Naturally, the s tress in the weld regionmust be in high tens ion, but th is could not be measureddirect ly , as explained above.

Because of the small residual stresses due towelding of th is heavy shape, the dis tr ibution of the i n t i a ls tresses i s retained in the welded shape, only with somemodifications in the magnitude of s t ress . Figure 21 shows thes tresses as measured on both surfaces of the c:omponents of


22/67

337.12 -1 8

the welded shape 23H681. The d i a g ~ a m can be c o m p a ~ e d d i ~ e c t 1 ywith those in Figs. 6-9 and 11-14 the l5H290 shape.

A fu l l u n d e ~ s t a n d i n g of the d i s t ~ i b u t i o n of the~ e s i d u a l s t ~ e s s e s in the 23H68l shape can be gained only

f ~ o m the ~ e s u l t s of complete sectioning and sl ic ing, shownin Fig. 22. The d i a g ~ a m indicates that not only is the

v a ~ i a t i o n of s t ~ e s s a c ~ o s s the width ~ e t a i n e d in thewelded shape but also the v a ~ i a t i o n a c ~ o s s the thjckness is

s i m i l a ~ to tha t in the loose component plates ( C o m p a ~ ethe d i a g ~ a m in Fig. 22 with those in Figs. 18 and 19).M a j o ~ changes have o c c u ~ ~ e d only in the w e l d a ~ e a .


23/67

337.12 -1 9

DISCUSSION OF RESULTS

From the results obtained on thick platesand heavy shapes in the investigation it is clear tha t thewelding residual stresses in areas of the cross section awayfrom the weld are far smaller in heavy welded shapes thanf d Oh 1 0 0 0 ( 5 ,6 , 7 ) f 11oun In t e ear le r lnvestlgatlons or sma platesand shapes. Figure 23 shows a comparison between the in i t i a lresidual stresses in a universal-mill plate 10" x i" andthe residual stresses af ter applying welds in the center of

h 1 (5 ) Whol h 0 ht e pa t e . le t e maXlmum compresslve stress In t eas-rolled plate is -6 ksi , the maximum compressive stressin the welded plate is -26 ksi , indicating a contributionfrom the welding of -20 ksi a t th is point. Thus, for thispla te , which is representative of small and medium-size plates,the residual stressesfrom welding consti tute the major partof the residual stress dis t r ibut ion.

Comparing these resul ts with those summarizedin Fig. 20, i t is noted that the si tuat ion in the heavymaterial i s quite different ; the major portion of theresidual stresses in the 23H681 shape are those originatingfrom the in i t i a l plates. The maximum compressive weldingstress in the flange plate i s -2 ksi . The flange plate isbasically a center-welded plate and may therefore be


24/67

337.12

compared directly with the center-welded plate in Fig. 23.The comparison exemplifies the observation that the weldingresidual stresses are of a smaller absmlute and re la t ivemagnitude in heavy plates and shapes. This could also bepredicted since the ra t io of weld area to that of parentmaterial is smaller for heavy p!'actical shapes. Thus,proportionately, there is a smaller heat input for shapesof thick plates than there is for l ight plates, and soheavy shapes would be expected to contain compressive weldingresidual stresses .o f a smaller magnitude.

The above finding also means that the in i t i a lresidual stresses existing in the component plates befo!'ewelding a!'e of greater importance for heavy shapes. Thisconclusion is obvious from the resul ts obtained on the15H290 when comparing the residual stresses in the shapesmade from unive!'sal-mill plates (Figs. 6-9) and fromflame-cut plates (Figs. 11-14). The only difference betweenthe two sets of specimens is in th e manufacturing p!'ocedurefor the plates, so the comparison reflects the effect of th isvariable only.

-20

All flanges of the 15H290 shapes made of univeralmill plates show a similar kind of distribution with highcompressive stresses a t the flange t ips . As can be seenin Fig. 10, the compressive stresses are a t the yield point


25/67

337.12 -21

level in the flange t ip corners. This is in accordancewith theoret ica l predict ions of residual s t resses inuniversal -mi l l plates based upon the heat flow in cooling. (13 )According to the predict ions, the residual stresses tendto increase with increasing size of the member. For heavyplates subjected to free cooling, the cooling residual s t ressesas predicted are in high compression a t the plate edges andmay approach the yield point for some plates .

Flame-cut plates show a reversed dis tr ibutionwith high tensi le s tresses at the flame-cut edge, balancedby compressive s tresses in the center of the plate . Thes tress a t the flame-cut surface is larger than the yieldpoint of the parent material . This i s possible because themechanical propert ies of the material close to the flame-cutsurface have been increased from the rapid cooling af tercutting. Another factor which influences the condit ions isthe three-dimensional s tress s t a t e . Since the local stressesin the heated mater ial normally are in high tens ion, that i s .of equal sign in a l l three direct ions, the mater ial cansustain a higher stress than the yield point in a uni-axialtension t e s t .

Basical ly, these conditions prevai l also inthe weld regions in the welded shapes, which can explain the


26/67

337.12

high tension stresses invariably experienced in the welds.However, for the weld, the weld area contains a mixture ofelectrode and base material. The mechanical propertiesof electrodes used for welding s t ructural carbon steelsnormally have a higher strength than the parent material .Thus, there are three effects which explain the occurranceof high stresses in the weld areas: (1) increased strengthof material due to the electrode material , (2) increasedstrength due to the cooling ra te , and (3 ) three-dimensionals t resses . Tension t es t s of specimens containing weld metalhave indicated a yield strength of 50-55 ksi for A7

-22

s teel , that i s , about 50% higher than the yield strength of theparent material . (5 ) The residual s tresses measured in the weldarea of the heavy shapes are a l l of th is order ofmagnitude.

The weld type is not a major factor in theformation of residual s tresses in thick plates and shapes.A comparison of the distr ibut ions in Figs. 6 and 13, forf i l l e t welds, and the corresponding diagrams in Figs. 7, 9,12, and 14, for groove welds, indicates no significantdifference. This is probably because the heat input in eachweld pass is of th e same order of magnitude for the f i l l e tweld and the groove weld. (See Table 1). Similar resul tswere obtained previously for ~ m a l l and medium-size plates . (S)


27/67

337.12

The res idual st ress dis tr ibutions as measuredin the shapes of A36 s tee l and of A441 s teel are s imilar ,indicating tha t the type of s tee l has no great influence onresidual stress . Since the effect of res idual s tress oncolumn strength i s dependent on the ratio. of residualst ress to yield s t ress , while a t the same time the magnitUdeof the residual stresses are not much affected by the typeof material, it can be expected that columns made ofhigh-strength s teels are stronger than those made of A36stee l , also when compared on a non-dimensional basis. Thishas been confirmed previously by residual stress measurementsand column tes ts on small and medium-size welded shapes.(14)

The geometry of the plates and shapes is oneof the major variables affecting the residual s t ressdist r ibut ion. As discussed above, the contribution ofresidual st ress due to welding is greatly dependent on thesize of the components, tha t i s , whether the shape i s l ight ,medium-size, or heavy. Figure 24 summarizes the resul tsof residual s tress measurements made on a number of weldedshapes using flame-cut pla tes , ranging from the 7H28 shapeto the 23H681 shape. Generally, the res idual s tresses dueto welding decrease with increasing size of the member.

Another feature of residual stresses in thick

-23

plates i s the varia t ion through the thickness of the components.


28/67

337.12 -2 4

Results of measurements including both sectioning and sl ic inghave shown that the variation in thick plates is considerable.Fo r instance, the non-symmetrical application of welds on aflange plate resul ts in high tensi le stresses on the weldedside , whereas the stress a t the opposite side often i scompressive. I t i s believed that the large variation ofst ress across the thickness for the component plates 24" x 3t"and 16" x 2" (Figs. 18 and 19) remains from the coolingafter rol l ing of the parent plate . Theoretical predictionsof cooling st resses in rol led plates of a similar sizeshow a distribution of st ress through thickness of theplate very close to that measured in th e center of these

(13 )flame-cut plates . Also, the variation across thewidth for th e center portion of the plates appears to beinfluenced by the i n i t i a l stresses from the cooling afterrol l ing of the parent plates . I f there were no in i t i a l stressesbefore flame-cutting of the plates , the st ress dis t r ibut ionin the center portions of the plates would approximate aplane, except for the heated and yielded regions a t theflame-cut edges, where tensi le stresses are introduced. Themeasured dist r ibut ions, however, show a variation acrossthe plate width of 10 ksi and 13 ksi for the center port ionsof the 16" x 2" and 24" x ~ " plates, respectively. The


29/67

337.12 -25

actual type of distribution in the center of the p l a t e s ~ withhigher compressive stresses towards the e d g e s ~ is consistent' th d" f l' "'1 11 d 1 (13)l pre lct lons a coo lng stresses In Slml ar ro e pa t e s .

This means that the stresses in thick flame-cut platesresu l t from both the cooling af ter rol l ing of the parent plateand the local heat input in the flame-cutting process.

T h u s ~ the residual s t ress distribution in afabricated member of thick component plates generally is acomplex superimposed pattern resul t ing from the differentstages of manufacture and fabrication procedures. Thatcooling residual stresses af ter r o l l i n g ~ residual s tressesfrom flame-cutting as well as the welding residual stressand any other procedure employed in the manufacture andfabrication wil l influence the f inal residual stressdistribution in a heavy welded shape. H o w e v e r ~ since thewelding residual s t resses in heavy welded shapes are s m a l l ~except in the weld r e g i o n ~ residual s tresses in such shapesmay be predicted by determining the i n i t i a l stresses in thecomponent plates from existing data or from measurementsand estimating the addit ional residual s tresses due to welding.This is a further simplification of the finding obtainedpreviously for smaller shapes that the residual s tresses ina welded shape can be predicted from a knowledge of theresidual stress dis t r ibut ion in the separate component plates


30/67

337.12

with simulated weld beads.(7)The relationship between in i t i a l residual

stresses in component plates and welding residual stressesimplies that efforts to l imit residual stresses in heavywelded shapes should be directed towards the manufactureof the component plates . Thus, by using flame-cut platesin heavy welded shapes, there is a prospect for an increasein strength when compared with l ighter members a t the sameslenderness ra t io . On the other hand, the column strengthof heavy columns made of universal-mill plates with as-rol lededges may be lower, in par t icular for A36 s teel and a thigher slenderness rat ios. (16)

For heavy welded columns made of flame-cutplates, there i s even a possibi l i ty that such weldedshapes may be stronger than thei r rol led counterpart.s.Figure 25 shows the measured distribution of residual

. (17)st ress in a rolled shape 14WF426. The nominal e;trengthcharacter is t ics of this shape fa l l in between those ofthe two welded shapes 15H290 and 23H681 studied here. .Thestresses are very high, of a similar dist r ibut ion and orderof magnitude as in the welded shape 15H290 made qf universal-mill plates .

The favorable indications for the strength ofheavy welded columns made of flame-cut. plates is in contrast

-2 6


31/67

337.12

with previously published resul ts for small and medium-sizewelded shapes of universal-mill plates which proved to be

(6)weaker than their rol led counterparts. However, sincethe present basis for the implications on heavy columnstrength is l imited to the few specimens studied, thestatements are preliminary; the general conclusions wil lbe developed further when the current research program i scompleted.

-27


32/67

337.12

CONCLUSIONS

The purpose of this study was to investigateexperimentally the magnitude and distribution of residuals tress in heavy column shapes bui l t up by welding plateswith thickness in the range of I t to 3t inches. Twoshapes were inves t iga ted: a 15H290 shape and a 23H681shape. For the smaller shape comparative tests wereconducted to study the influence of the manufacture ofplates , weld type, and yield strength of the material. Thespecimens referred to in this paper constitute the f i rs t

-28

phase of a major research program concerning residual stressesin thick welded plates .

Based on the results of this f i r s t phase ofthe program, the following conclusions can be stated:

1. The residual s tress distribution in afabricated member of thick component plates generally is acomplex superimposed pattern resulting from the differentstages of manufacture and fabrication p r o c e d u r ~ s . That i s ,cooling residual st resses from the rol l ing of the parentplates , residual stresses from flame-cutting the plates ,as well as the residual stresses from the welding process,a ll influence the f inal residual s tress distribution in awelded shape.


33/67

337.12 -29

2. The welding residual stresses in portions of thecrOss section other than the weld area tend to decrease withincreasing size of the member, probably because the weldarea, and consequently, the heat input, is relat ively smallerin heavy plates and shapes as compared to l ight members.

3. The results of comparative measurements on l5H290shapes indicate that the weld type (penetration) is not amajor factor in the formation of residual stresses in heavyplates and shapes, as long as the heat input is notdrast ical ly changed.

4. The residual stress distributions as measuredin shapes of ASTM A36 stee l and of A44l s tee l are similar,indicating that the yield strength of the s tee l ha s no greatinfluence on residual stresses in heavy shapes.

5. The magnitude and distribution of in i t i a l residualstresses depends greatly on the manufacturing procedure, thati s , universal-mill plates or flame-cut plates.

6. The geometry of the plates and shapes i s one ofthe major variables affecting the residual stress distribution.More specif ical ly , the residual stresses wil l depend uponwhether the shape i s l ight , medium-sized or heavy (effectof size of cross section) or whether the plates are narrow,medium or wide (effect of width to thickness ra t io) .


34/67

337.12 -30

7. Residual stresses due to cooling af te r rol l ingof a plate normally are compressive a t the edges, balancedby tension stresses in the center portion of the plate .Measurements on four welded H-shapes, 15H290, containinguniversal-mill plates indicate that the in i t i a l stressesin the flange plates are very high. This conforms to theobservation made previously from th e results of theoret icalpredictions that the cooling residual stresses in rolled platestend to increase with increasing size of the plate .

8. Residual stresses due to flame-cutting arein high tension a t the flame-cut edge due to the localheat input. For the flame-cut plates studied, the s tressa t the burned surface is about 50% above the yield pointof the parent material. The tensi le stresses are balancedby compression in the center part of the plate; thedistribution of residual stresses in th e center of a heavyflame-cut plate i s dependent also on the in i t i a l stresses dueto rol l ing.

9. The variation of residual s tress across thethickness of plates more than one inch thick can be considerable.Such variation will resul t in a plate with welds depositedon one side of the plate; also, a significant differencecan be expected between stresses in the surface and the interiorof heavy rol led plate components.


35/67

337.12 -31

10 . The relationship between the in i t i a lresidual s tresses in component plates and the welding residualstresses implies that the welding stresses in thick weldedplates and shapes are of less importance than the in i t i a lst resses existing prior to welding. Efforts to l imitresidual stresses in heavy welded shapes should be directedtowards the manufacture of the component plates .

11. Since the welding residual stressesin heavy welded shapes are small , except in the weld region,residual stresses in such shapes may be predicted bydetermining the in i t i a l stresses in the component platesfrom existing data or from measurements, and estimating theadditional residual stresses due to welding. This is afurther simplification of the finding obtained previouslyfor smaller shapes that the residual stresses in a weldedshape can be predicted from a knowledge of the residual stressdistribution in the separate component pla tes with simulatedweld beads. (7)

The effect of residual stresses on the load-carrying behavior and strength of heavy columns wil l be

. (16)discussed further a future paper. Briefly, by using


36/67

337.12 -32

flame-cut plates in heavy welded shapes, there is a prospectfor an increase in strength wnen compared with l ighter membersa t the same slenderness ra t io . Under certain circumstances,there is even a possibi l i ty that such welded shapes may bestronger than thei r rol led counterparts, which is the oppositerelationship as compared to the situation for small andmedium-sized shapes. On the other hand, heavy welded columnsof universal-mill plates and A36 s teel are expected toshow a lower column strength, especially for columns of amedium to high slenderness rat io.


37/67

337.12

ACKNOWLEDGMENTS

This paper presents the resul ts of an experimentalstudy of residual stresses in heavy welded shapes. Theinvestigation i s the f i r s t phase of a major research programdesigned to determine the residual stresses in thick weldedplates and shapes and to re la te th is to the stabi l i ty underload of compression members.

The investigation was conducted a t Fri tz EngineeringLaboratory, Lehigh University, Bethlehem, Pennsylvania. TheNational Science Foundation sponsors the current researchprogram. A pi lot study, i n c l u d ~ d in th is paper, was part ofa project sponsored joint ly by the Pennsylvania Departmentof Highways, the Bureau of Public Roads of th e U. S.Department of Commerce, the Column Research Council, theAmerican Insti tute of Steel Construction, and theAmerican Iron and Steel Insti tute. The specimens werefabricated by the Bethlehem Steel Corporation, and thanks aredue to tha t corporation and i t s personnel who assisted inthe design and fabrication of the specimens. The guidanceof Task Group 1 of the Column Research Council, under thechairmanship of John A. Gilligan, is gratefully acknowledged.

Special thanks are due to Lynn S. Beedle, Directorof Fri tz Engineering Laboratory, for his advice and

-33


38/67

337.12 -34

encouragement throughout the program. Fiorello R. Estuarand William C. Cranston carried out th e experiments on thel5H290 shape and Charles R. Nordquist assisted in themeasurements on the 23H68l shape and i t s component plates .Their assistance is sincerely appreciated.

Thanks are also due to Kenneth R. Harpel,laboratory foreman, and his s taf f for the preparationof the tes t specimens, to Mrs. Sharon Balogh for thepreparation of the drawings, and to Miss Joanne Mies andMrs. Linda Welsch for thei r care in typing the manuscript.


39/67

337.12 -35

TABLES AND FIGURES


40/67

Test Shape Material Static Manufacture Size of Size of Welding Detail No. ofSpecimen Designation Yield of Plates Flange Web Method of PassesNo. Stressa FC/UM Plate Plate and Weld Weld for each(ks i ) (inch) ( inch) Type Weld

1 J . S H 2 ~ 0 A36 33.7 UM l 4 x 2 ~ l O x l ~ SAW,fi l let ~ 22 lSH290 A36 36.0 UM 14x2 t l O x l ~ SAW ,groovei f3 lSH290 A41.;.1 46.7 UM 1 4 x 2 ~ l O x l ~ SAW.fi l let SEE

4 lSH290 A441 4S.2 UM 14x2t l O x l ~ SAW ,groove SEE

S ISH290 A36 3S.9 FC 1 4 x 2 ~ l O x l ~ SAW,fillet SEE

6 15H290 A36 3S.9 FC 1 4 x 2 ~ l O x l ~ SAW,groove SEE

7 15H290 A441 44.6 FC 1 4 x 2 ~ l O x l ~ SAW,fi l let SEE

B lSH290 A441 42.0 FC l 4 x 2 ~ 10x1 t SAW,groove SEE

9 PL16x2 A36 32.S FC

10 P L 2 4 x ~ A36 32.1 FC

11 23H6Bl A36 32.2 FC 2 4 X ~ 16x2 SAW,fi l let5J 1a Weighted AveI'age

TABLE 1. DATA OF TEST SPECIMENS

Sequence Weldingof Weld Pass oltagePasses No. (Vol ts )

IJ.-4 33S-B 33

6,2 3,7

16,9,6,2, 4,5,12,13,25.24.17 20,21,28 1- 2 323- 4 3327,22,19,14.11,6.3 18,23,26,1,7,10,15 S-2B 33

TEST SPECIMEN NO.

TEST SPECIMEN NO.

TEST SPECIMEN NO.

TEST SPECIMEN NO.

TEST SPECIMEN NO.

TEST SPECIMEN NO.

st DC 26I 1. AC 31nd DC 261. AC 30I I

Datallr'rent

(Amps)

6S06S0

SOD6S06S0

1

2

1

2

1

2

700S30710S30

Speed( ipm)

2614

24IB16

15ISIBIB

I

I

I

I,II

II

IW(j )


41/67

730 IbItt(a)

1610 IbItt(d)

Fig. 1

1130 IbItt(b)

1190lbltt( e)

2690lb/ft(C )

Scale ( nches):o 10 20 30

Heavy Column Shapes in Existing StructuresIW' I


42/67

JC8W'40(O )

, .

23H681(e)

IIIH66(b)

Scale (inches):Io 10 20 30

I4H202(c)

240774(f )

Fig. 2 Test Columns

1:5H290(d)

24HII22( g)

IW(Xl


43/67

Fig. 3 Largest (23H68l) and smallest (7H28)column shapes tested so far in theresearch program

-39


44/67

Fig. 4 Fabrication of the Welded H-Shape23H68l (Courtesy Bethlehem SteelCorporation)

-40


45/67

\" } \. JV vSECTIONING SLICINGTop Surface Reading

Straight LineApproximation

-Additional StressFrom SI icingStraight LineApproximationFrom Sectioning

Bottom Surface Reading Shaded Area Equals FinalResidual Stress DistributionFig. 5 Principle of the Sectioning Method forResidual Stress Measurements

I-I="......


46/67

-40

60

I

I60

-40

I

l

KSI60 40 20 0

-0-0- Near Surface- Far Surface

Fig. 6 Residual Stresses in a Welded Shape 15H290.Universal-Mill Plates , A36 Steel , 1/2" Fil le t Welds.

-42


47/67

-40

KSI

60

Near Surface- Far SurfaceKSI

-20

-40

Fig. 7 Residual Stresses in a Welded Shape 15H290.Universal-Mill Plates, A36 Steel , 11/16"Groove Welds.

-43


48/67

-40

KSI

KSI

60 KSI

I I

I I60

40- 0 - - 0 - Nea r Surface- - - - Far Surface

Fig. 8 Residual Stresses in a Welded Shape l5H290Universal-Mill Plates , A44l Steel , 1/2"Fi l le t Welds.

-44


49/67

-40

KS I 0 ~ - - - - - - 1 J J ~ -

60

I

I

J

I

KSI60 40 20 0

-0 -0- Near Surface Far Surface

K S 0 - - - r I ~ - - \ : - -

-20-40

Fig. 9 Residual Stresses in a Welded Shape 15H290.Universal-Mill Plates , A441 Steel , 11/16"Groove Welds.

-45


50/67

15H290(UM)(KSI)

Fig. 10 Two-Dimensional Variation of Residual Stressin a Welded Shape 15H290. Universal-MillPlates, A36 Steel, 1/2" Fill.et Welds.

-46


51/67


52/67

-40-20

KS I 0 ~ ___ - - - + - - ~ t - +204060 KSI

60 40 20 0 -20 -40l I...

I I6040 a a Near Surface20 Far Surface

KS 0 1I\r--+--__._. ."", . . . . . . . ,

-20-40

Fig. 12 Residual Stresses in a Welded Shape 15H290.Flame-Cut Plates, A36 Steel , 11/16"Groove Welds.

-48


53/67

KSI

-40

-20

020

40

60 KSI60 40 20 0 -20 -40I I

I I60

40

20 ~ Near Surface Far Surface

- 2 0 - ---40

Fig. 13 Residual Stresses in a Welded Shape 15H290.Flame-Cut Plates, A44l Steel, 1/2"Fi l le t Welds.

-49


54/67

-40-20

K S 0 1 -+ - - - - + - - - 1 - - - - 4 , . . - -

204060

604020

KSI60 40 20 0 -20 -40

-...0-0- Near Surface Far Surface

K S I 0 I--t--............ - + - - - - - l ~

-20-40

Fig. 14 Residual Stresses in a Welded Shape 15H290.Flame-Cut Plates, A441 Steel , 11/16"Groove Welds.

-50


55/67


56/67

-52

20

40

60

PLI6x2(FC)

Fig. 16 Residual Stresses in a Flame-Cut Plate 16" x 2"


57/67

-53

-20

40

60

PL 24x 3V (Fe)

Fig. 17 Residual Stresses in a Flame-Cut Plate 24" x 3i"


58/67

Fig. 18 Two-Dimensional Variation of Residual Stressin a Flame-Cut Plate 16" x 2"

IU1-1=


59/67

m~

-55


60/67

-2 0 -20KSI 0 KSI 0''''" - -- . P _ __ _

20 2040 4060 60KSI KSI

KSI 0' ~ r= ,204060

L...---rn--;.::;: :::. ; ; ! . ~ ~ O ~ ' f < \ ; ~ o: : - - ~ ~ o I I 2,0 i I

604020

[

KSI 0 r\ .....- --", I-2 0

(a ) INITIAL RESIDUAL STRESSES

604020

I

KSIOt _ I ....__-2 0

I

(b) ADDITIONAL WELDING STRESSES

,1604020

KSI 0 I'. ....6 " '" i"-2 0

(e ) FINAL RESIDUAL STRESSES

Fig. 20 Formation of Residual Stresses in aHeavy Welded Shape 23H681Itn(J )


61/67

-57

-40

KSI o ~ - - ~ ~ - - - - ~ ~ ~

60 KSI

I I

I I60

-0--0- Near Surface- Far Surface

-40

Fig. 21 Residual Stresses in a Welded Shape 23H681.Flame-Cut Plates, A36 Steel , 1/2" Fil le t Welds.


62/67


63/67

-59

KSI20

PL IOX lt2 CUM)-40

4060

W600 Groove Weld

-40-20

KSI 0204060

ZDouble V,Weld

Fig. 23 Residual Stresses in a Universal-Mill Plate 10" x 1/2U..


64/67

An7H2.IOH62

12H79

t::::\ . r--" \I \ l ~ 14H202~ ~d \O J" 5H290 Stress Scale (ksO:i ' - . .At V \ 23H68 I , I Til I I I I I60 40 20 0 -20-4015H290 14H202 12H79 IOH62

Fig. 24 Residual Stresses in Welded H-Shapes ofDifferent Geometry. Flame-Cut Plates,A3 6 Steel , Fi l le t Welds. (Results fo rl2H79 and l4H202 Shapes from Ref. 15 )

7H28

I() )o


65/67

KSI

20

I14W"426

I20

KSI

-30

I

I

-0--0-- Near Surface Far Su rface

Fig. 25 Residual Stresses in a Hot-RolledShape 14WF425, A7 Steel

-51


66/67

337.12

REFERENCES

1. A. W. Huber, and L. S. BeedleRESIDUAL STRESSES AND THE COMPRESSIVEPROPERTIES OF STEEL, Welding Journal, Vol. 33,pp. 589-s to 6l4-s , 1954.

2. L. S. Beedle, and L. TallBASIC COLUMN STRENGTH, Trans. ASCE, Vol. 127,Part I I , pp. 138 to 179, 1962.3. F. Nishino, Y. Ueda, and L. Tall

EXPERIMENTAL INVESTIGATION OF THE BUCKLING

-62

OF PLATES WITH RESIDUAL STRESSES, ASTM STP No. 419,"Test Methods for Compression Testing," August 1967.4. Y. Ueda, and L. TallINELASTIC BUCKLING OF PLATES WITH

RESIDUAL STRESSES, Publications, InternationalAssociation for Bridge and Structural Engineering,Vol. 27 , 1967.5. N. R. NagarajaRao, and L. Tall

RESIDUAL STRESSES IN WELDED PLATES,Welding Journal, Vol. 40, pp. 468-s to 480-s, 1961.6 F. R. Estuar, and L. Tall

EXPERIMENTAL INVESTIGATION OF BUILT-UP COLUMNS,Welding Journal, Vol. 42(4), pp . l64-s to 176-s, 1963.7. N. R. NagarajaRao, F. R. Estuar, and L. Tall

RESIDUAL STRESSES IN WELDED SHAPES,Welding Journal, Vol. 43(7), pp. 295-s to 306-s, 1964.8. G. A. AlpstenRESIDUAL STRESSES IN A HEAVY WELDED SHAPE23H68l, Fritz Laboratory Report No. 337.9, InPreparation.9. G. A. AlpstenTHREE-DIMENSIONAL RESIDUAL STRESSES AND COLUMNSTRENGTH, Fritz Laboratory Report No. 337.20 ,In Preparation.


67/67

337.12

10.

11.

N. KalakoutskyTHE STUDY OF INTERNAL STRESSES IN CASTIRON AND STEEL, London, 1888.

F. R. EstuarWELDING RESIDUAL STRESSES AND THE STRENGTHOF HEAVY COLUMN SHAPES, Ph.D. Dissertation,Lehigh University, Aug. 1966. (University

~ i c r o f i l m s , Inc. , Ann Arbor, Michigan).12. J . Brozzetti , and G. A. Alpsten

ACCURACY OF THE SECTIONING METHOD FORRESIDUAL STRESS MEASUREMENTS, FritzLaboratory Report No. 337.11, inPreparation.

13 . G. A. AlpstenTHERMAL RESIDUAL STRESSES IN HOT-ROLLEDSTEEL MEMBERS, Fri tz Laboratory ReportNo. 337.3, December, 1968.

14. Y. Kishima, G. A. Alpsten, and L. TallCOLUMN STRENGTH OF WELDED SHAPES USINGFLAME-CUT PLATES, Fritz LaboratoryReport No. 321.4, In Preparation.

15. R. K. McFalls, and L. TallA STUDY OF WELDED COLUMNS MANUFACTURED FROMFLAME-CUT PLATES, Fritz Laboratory ReportNo. 321.1, June 1967.

16. G. A. A1psten, and L. TallCOLUMN STRENGTH OF HEAVY SHAPES - APROGRESS REPORT, Fritz Laboratory ReportNo. 337.16, In Preparation.

17. Y. FujitaTHE MAGNITUDE AND DISTRIBUTION OF RESIDUALSTRESSES, Fritz Laboratory Report No.220A.20, May, 1955.

-63

Documents

Residual Stresses in Heavy Welded Shapes January 1969 (70-6)