01 the fatigue and static properties of butt welds in structural - Ideals

........

~;2qA-

tf:8f CIVIL ENGINEERING STUDIES

STRUCTURAL RESEARCH SERIES NO. 81

~3 01

THE FATIGUE AND STATIC PROPERTIES OF BUTT WELDS

IN STRUCTURAL STEELS

By L. A. HARRIS

and

G. E. NORDMARK

Approved by

N. M. NEWMARK

and

W. H. MUNSE

UNIVERSITY OF ILLINOIS

URBANA, ILLINOIS



L. A. Harris

and

G. E. Nordmark

Approved by

N. M. Newmark

and

W.R. Munse

A Report of an Investigation 'Conducted by

THE DEPARTMENT OF CIVIL ENGDJEERING

UNIVERSITY OF ILLINOIS

In Cooperation With

THE OHIO RIVER DIVISION LABORATORIES

CORPS OF ENGINEERS

U. S • .ARMY

Contract No. DA-33-017-eng-221

UNIVERSITY OF TILINOIS

URBANA, ILLINO IS

JUNE 30, 1954

I.

CONTENTS

Introduction. . . • • • .

1.

2.

Object and Scope of Investigation ••

Acknowledgement . • . . . ~

Page No.

1

1

3

II~ Description of Test Procedures. • 5-

3· Description of Steel. • • • • • • • 5

4. Preparation o.f Longitudinal Butt Welds. · 5

5. Preparation of Transverse Butt Welds. 8

6. Test Procedures . • . • • • • • • • • 10

III. Tests of' Longitudinal Butt Welds in ASTM A 7 Steels. 13

7 .. Results of' Present Fa.tigue Tests of

Longitudinal Butt Welds .

8. Results of Present Static Tests of'

Longitudinal Butt Welds ~

9. Review of Previous Longitudinal Butt Weld

13

17

Fatigue Tests. • . • • • .• 18

rv. Tests of Transverse Butt Welds in ASTM A7 Steels.. 21

10. Results of Present Fatigue Tests on

Transverse Butt Welds. . .

11. Results of' Present Static Tests of'

Transverse Butt Welds

12. Scope of the Study of' Previous Fatigue Tests

2l

24

of Transverse Butt Welds. . . . • • • • • •• 24

CONTENTS (CON'T)

Page No.

13. Review of Past and Present Fatigue Tests

of Transverse Butt Welds with the Reinforcement

On: O-Tens ion Cycle. . • • • • • • • • • • 26

14. Review of Past Fatigue Tests of Transverse

Butt Welds With the Reinforcement On:

Tension = Compression Cycle • . •

15. Review of Past and Present Fatigue Tests of

Transverse Butt Welds with the Reinforcement

Removed. •

V. Review of Past Fatigue Tests of Transverse Butt

Welds in Silicon Steel and in Low Alloy High Tensile

steel. . . . . . . . . 16. Scope of Investigation.

17· Review of Past Fatigue Tests of Transverse

Butt Welds in Silicon Steel . . . . . . . . 18. Review of Past Fatigue Tests of Transverse

32

. 37

37

38

Butt Welds in Low Alloy High Tensile Steel. 39

VI. Summary and Conclusions •

20. Conclusions .•

Bibliography . • • •

41

41

43

4B

TABLES

1. Chemical Composition of Steel Plate

2. Physical Properties of Steel Plate

3. H'elding Procedures

4. Results of Longitudinal Butt Weld Fatigue Tests

50 Results of Plain Plate Fatigue Tests

6. Location of Initiation of Fatigue Fractures in

Longitudinal Butt Weld Specimens

7- Results of Static Tests of Longitudinal Butt Welds

8. Results of Past and Present Fatigue Tests of

Longitudinal Butt Welds

9. Properties of Plate Material Used in Past and Present

Longitudinal Butt Weld Fatigue Tests

10 . Results of Transverse Butt Weld Fatigue Tests

11. Results of Static Tests of Transverse Butt Welds

12. Properties of Plate Material Used in Past and Present

Transverse Butt Weld Fatigue Tests

13. Results of Past and Present Fatigue Tests of Transverse

Butt Welds in the As-Welded Condition: O-Tension Cycle

14. Average Results of Past and Present Fatigue Tests of

Transverse Treatments: O-Tension Cycle


Butt Welds with the Reinforcement On Prepared with Special

Treatments: O-Tension Cycle

TABLES (CON' T)


Butt Welds with the Reinforcement On:

Tension = Compression Cycle


Butt ltlelds with the Reinforcement Removed

18. Properties of Silicon Steel and Low Alloy High Tensile

Steels Used in Past Fatigue Tests of Transverse Butt Welds

19- Results of Past Fatigue Tests of Transverse Butt Welds in

Silicon Steel and in Low Alloy High Tensile Steels

FIGURES

1. Location of-Specimens in Parent Plate) Plate 10

2. Location of Specimens in Parent Plate) Plate 9

3. Location of' Specimens in Parent Plate) Plate 8

4. Location of Specimens in Parent Plate, Plate 7

5· Location of Specimens in Parent Plate) Plate 6

6. Details of Butt Weld Specimens

7· Welding Jig for Longitudinal Butt Weld Specimens

8. Details of Welding Sequence for Longitudinal Butt Welds

9. Details of Preparation of Transverse Butt Weld Specimens

10. Schematic Diagram of Wilson Fatigue Testing Machine

li. Results of Fatigue Tests of Longitudinal Butt Welds

12. Typical Fracture Surfaces of Plain Plate Fatigue Specimens

13· Typical Fatigue Fractures of Longitudinal Butt Welds with the

Reinforcement On.

14. Typical Fatigue Fractures of Longitudinal Butt Welds with the

Reinforcement Removed by Grinding

15. Results of Past and Present Fatigue Tests of Longitudinal Butt Welds

16. Results of Fatigue Tests of Transverse Butt Welds

17. Typical Fatigue Fractures of Transverse Butt Welds with the Reinforcement

On

18. Typical Fatigue Fractures of Transverse Butt Welds with the Reinforcement

Ground Off

190 Typical Macrographs of Plane Normal to Fatigue Fractures

20. Results of Past and Present Fatigue Tests of Steel Plate Used in

Transverse Butt Weld Investigations

FIGURES (CONTT)

21. Results of Past and Present Fatigue Tests of Transverse Butt

Welds in the As-Welded Condition Prepared with E6ol0 and E6ol2

~ectrodes: O-Tension Cycle

22. Results of Past Fatigue Tests of Transverse Butt Welds in the

As-Welded Condition Prepared with E6ol3 and E6030 Electrodes:

O-Tension Cycle

23· Results of Past Fatigue Tests of Transverse Butt Welds in the

As-Welded Condition Prepared by Automatic Procedures:

O-Tension Cycle

24. Summary of Results of All Past and Present Fatigue Tests of

Transverse Butt Welds in the As-Welded Condition: O-Tension Cycle

25. Results of Past Fatigue Tests of Transverse Butt Welds with the

Reinforcement On Welded with Special Procedures: O-Tension Cycle

26. Results of' Past and Present Fatigue Tests of Transverse Butt Welds

in The As-Welded Condition Prepared 1.;ith E6ol0, E6ol2, and E6ol3

Electrodes: Tension = Compression Cycle

27· Results of Past and Present Fatigue Tests of Transverse Butt Welds

in the As-Welded Condition Prepared with E6020 and E6030 Electrodes

and with Automatic Procedures: Tension = Compression Cycle

,28. Summary of Results of All Past and Present Fatigue Tests of Transverse

Butt Welds in the As-Welded Condition: Tension = Compression Cycle

29 . Results of Past Fatigue Tests of Transverse Butt Welds with the

Reinforcement On Welded with Special Procedures: Tension = Compression

Cycle

FIGURES (CON'T)

30. Results o~ Past and Present Fatigue Tests of Transverse Butt

Welds with the Reinforcement Off: O-Tension Cycle

31. Results of Past and Present Fatigue Tests of Transverse Butt

Welds with the Reinforcement Orf: Tension = Compression Cycle

32. Results of Past Tests of Transverse Butt Welds in Silicon Steel

and in Low Alloy High Tensile Steel



I. INTRODUCTION

1. Object and Scope of' Investigation

1.

It has been the purpose of most of' the previous fatigue tests

of' welded joints to provide sufficient experimental data from which

it would be possible to determine, quantitatively, the fatigue strength

-of' welded joints of different types, and thence, to determine allowable

stresses for use in the design of structures subjected to repeated

loads.. These data were accumulated both L."'1 the United Stat.es and abroad.

In the United States, the test programs were conducted under the guidance

of the Fatigue Committee of the Welding Research Council, most of the

experimental programS being conducted at the University of illinois

in f'atigue machines designed specifically for the purpose by Prof'essor

w. M .. Wilson.

With advances in welding techniques, recent interest has

been directed towards methods of improving the fatigue behaviour of

1ve1dments. During the last three years, the Ohio River Division

Laboratories of' the Corps of Engineers, U. S . .Army has sponsored a

research program at the University of Illinois which has had as its

chief objective a comparison of the properties of' welds produced

with the low hydrogen ~ype electrode or with the cellulose type

electrode, which has a high hydrogen content. The earlier work on

2.

this program was concerned with a comparison of the properties of bead-

on-plate welds produced with these two types of electrodes and subjected

1* to various treatments 0 On the basis of these tests, and related tests

performed elsewhere, it was felt that the use of the low hydrogen

electrode might result in a weld more resistant to repeated loadings

than a weld produced with other electrodes. Therefore, a program

of small scale, polished, all-weld-metal specimens was designed to

study this possibilityo The results of the tests of the all-weld

metal specimens2 indicated that the weld metal deposited with the

E7016 electrode had a fatigue strength about 25,000 psi greater than

the fatigue strength of the weld metal deposited with the E6ol0 electrode

when the weld pads were preheated and air-cooledo The greater fatigue

strength of the weld metal deposited with the low hydrogen electrode

was reduced about 8,000 psi if the weld pad was water quenched two

minutes after each pass had been depositedo Small scale, polished

specimens containing weld metal and base plate fractured at about the

same fatigue strength as the base plate, regardless of the electrode

used to prepare the weldo

On the basis of the encouraging results from the all-weld-

metal fatigue tests of the low hydrogen weld metal, it was decided

to inaugurate a series of fatigue tests to determine whether this

improvement would be reflected in tests of full scale welded jointso

*Numbers in superscript refer to the references listed in the Bibliography at the end of the report.

3·

The present investigation represents the first phase of this study

of fUll scale welded. joints: a study of the fatigue and static

properties of butt welds 0

It is the purpose of the tests reported herein to compare

the fatigue and static properties of butt welds produced with the

E60l0 and the E70l6 electrodes. I

Two types of butt weld specimens

were tested: the longitudinal butt welds were tested with the applied

stress parallel to the direction of welding, whereas the transverse

butt welds were tested with the applied stress perpendicular to the

direction of welding.. The butt welds have been tested either with

the reinforcement on or with the reinforcement removed by grinding.

The specimens were subjected to a stress cycle which varied .from a

low tension of about 2000 psi to a maximum tension which was of such

magnitude that failure generally occurred between 100,000 and

In order to more thoroughly understand the behaviour of

butt welds subjected to repeated stresses, an extensive review o.f

the fatigue tests of butt welds conducted previously at the University

of Illinois bas been included in the reporto

20 Acknowledgement

This report is part of a cooperative research project

conducted in the Engineering Experiment Station of the University

of illinois, Department of Civil Engineering, and sponsored by

4.

the Ohio River Division Laboratories, of the Corps o~ Engineers,

u. S . .A:rmy, under Contract DA-33-0l7-eng-22l.

The work constitutes a part o~ the structural research

program o~ the De:partment o~ Civil Engineering under the general

direction of N. M. Ne-wmark, Research Pro~essor of structural

Engineering, and W. R. Munse, Research Associate Professor of Civil

Engineering. The research was performed by L. A. Harris, Research

Associate in Civil. Engineering, who was the immediate :project

supervisor; R. B. Matthiesen, Research Assistant in Civil Engineering,

and G. E. NordInark, Research Assistant in Civil Engineering.

The metallurgical stUdies were performed in the laboratories

of the Department of Mining and Metallurgical Engineering under the

supervision of R. W .. Eohl, Assistant Professor of Metallurgical

Engineering. (

The authors wish to acknowledge the assistance of the

members of the Fatigue Committ'ee of the Welding Research Council,

who acted in an advisory capacity.

5·

IIo DESCRIPrION OF TEST PRCCEDURES

3. Description of' Steel

In order to more closely control the chemical content,

and hence the weldability, of' the steel plate used in these tests,

the steel was ordered to comply with the American Bureau of' Shipping

specification f'or TYPe B steelo The chemical analysis and the

physical properties of the steel are given in Tables 1 and 2,

respectively. The steel meets both the ABS specification and the

ASTM A7 specification for chemical composition, but the average

physical properties do not meet either specification for tensile

strength, the average tensile strength being only 57,400 psi. One

spec~en had a tensile strength as low as' 55,600 psi.

The steel was supplied as 3/4 in. thick plates which were

6 ft. wide x 12 ft. long 0 The locations of the test specimens in

the parent plates are shown in Figs 0 1 through 5.

4. Preparation of Longitudinal Butt Welds

The details of the completed butt weld specimens are shown

in Fig 0 6. As indicated in this figure, the longitudinal butt weld

specimen was prepared from three pieces, the central section containing

the test weld and the two outside pull heads. This procedure was

followed in order to conserve material, the pull heads being prepared

from ~reviously tested specimenso

6.

The central section o~ the longitudinal specimen, 12 in.

wide x 22 in. long, was flame-cut from the base plate as indicated

in the plate lS\Vout diagrams. The 12 ino wide sections were saw cut

along their centerline to a 6 in. width and the saw cut edges were

machined in a shaper for a double V weld with a 60 degree included

angle and a 1/16 ino root opening.

In making the longitudinal weld, the two halves of the

plate were ~irst securely bolted to the welding jig shown in Figo 7.

This jig could be rotated about a horizontal axis to permit

deposition o~ all passes in the flat positiono Fig. 7 shows the

plates in position for welding a pass on each side of the specimeno

After deposition of the first pass, the weld was back chipped until

no visible defects were present 0 Approximately 15 minutes of air

cooling was allowed between the first and the second passes, but

only five minutes was allowed between succeeding passes. After

completion o~ the weld, the specimen was air cooled for 10 minutes

and was then removed "from the welding jig 0

The double V joint was deposited in s:ix passes, each

pass being deposited in the flat positiono In order to assure

that- a change in electrode would occur in the test section, the

welding sequence, shown in Fig. 8, was used. This welding sequence

was such that adjacent passes were welded in opposite directions

and changes of electrode in adjacent passes did not occur one over

the other.

Two types o~ electrode were used to prepare the welds, one

corres~onding to the AWS E6ol0 des~tion and the other to the

AWS ElOl6 designation. .All o~ the electrodes were 3/16 ino in

diameter and were operated at the values given in Table 3. These

values were measured on portable meters connected as close to the

arc as possible.

After completion o~ the longitudinal butt weld, the ends

of the test section were prepared for welding and the previously

drilled pull heads were welded to the test sectiono The specimen

was then machined to the dimensions given in Fig. 6 and the edges

were drawf'iled"

A portable disk gJ:inder was employed to remove the weld

reinfoTcement and the mill scale ~or those specimens to be tested

with the reinforcement removed. T"ne grinding was started with a

36 grit wheel and then finished with a 120 grit wheel in such a

manner that the scratches were parallel to the direction of loading 0

The mill scale was ground o~~ from one pull head to the other to

prevent failure at irregularities out of the test sectiono Grinding

was continued to a depth sufficient to remove the mill scale and

the rolling defects on the surface of the plateso The transverse

butt welds connecting the test section to the pull heads were

ground flush for all of the specimens in order to reduce the

occurrence of failure at these weldso

8.

After preparation of' the specimen in the above manner, the

width was measured with a 4-5 ino micrometer and the thickness with a

1 ino micrometer. The area was computed f'rom the average of' several

readings and the reported stress was computed f'rom this area. The

spec~ens which were tested with the weld 'reinforcement on were

measured at several points in the test section away from the weld

and the average area, not including the area of' the weld reinforcement,

was used to compute the stress. In the latter case, the measured

thickrless includes the thickness of the mill scale 0 After they were

measured, the specimens were ready for testingo

50 Preparation of Transverse Butt Welds

The plate material to be used for the transverse butt

welds was first flame-cut to the shape shown in Fig. 9ao This plate

was then saw cut at the joint and the joint was prepared in a shaper

.for a double V butt weld having an included angle of 60 deg 0 and a

1/16 ino root opening 0 The weld specimen had an increased width at

the joint to enable the machining of the start and the end of the

weld from the specilnens before testing 0 As stated above J 3/16 in.

diameter electrodes conforming to the American Welding Society

designations E6010 and ET016 were used to prepare the welds 0 The

welding procedure employed 6 passes and the electrodes were operated

at· the values, given in Table 3, which were measured on portable

meters connected as close to the arc as possibleo The welds were

thoroughly back chipped after deposition of the f'irst passo

9·

In welding the specimens, the two halves of the specimen

were first firmly secured in a welding jig similar to that used in

the welding of the longitudinal butt yeld spec:imens. The jig was

so arranged that the specimen and its table mount could be rotated

180 degrees. about a horizontal axis; therefore, all welding was

done in the flat position.

The 7 in .. width of the weld specimen was too wide to

allow the completion of a single pass without changing electrodes.

In order to more closely control the duplication of specimens, the

changes in electrodes were made at the points specified in Fig. 9b.

By use of' this procedure, every pass of tpe weld was stopped and

restarted within the test section and adjacent passes were welded

from opposite directions.. The welds were allowed to air cool for

five minutes between passes and were air cooled to room temperature

after completion of the jointQ

After welding was completed, the specimen was machined

to the final dimensions, given in Figo 6, and the edges were

drawfiledo' Those specimens to be tested with the reinforcement

removed were ground in the same manner as the longitudinal butt

weld specimens, except that the specimens were ground only for a

distance of about- 2 ino from each edge of the weldo The specimens

were measured and the stresses computed in the same manner as the

longitudinal butt welds.

10.

6. Test Procedures

The fatigue tests were performed at room temperature in

two 200,000 lb. capacity Wilson lever type fatigue testing machines

which ran at a speed of approxilllately 200 cycles per minute. These

machines are capable of producing an alternating load cycle of

200,000 Ibs. compression to 200,000 Ibs. tensionj however, the only

stress cycle emploYed in the present tests was one which varied

from a low tension to a maximum tension. The low tension of about

6,000 Ibs. was used to keep the bearings seated properly throughout

the stress cycle.

The essential features of the fatigue testing machine,

shown in Fig. 10, are a variable throw eccentric which transmits

force through a dynamometer (for determining the load on the

specimen) to a lever which in turn transmits the force to the upper

pull head at a multiplication ratio of approximately 15 to 1. The

force that is exerted on the specimen originates in the double throw

eccentric, which is adjusted to give the desired range of stress

before the test is started. The maximum load is controlled by the

adjustable turnbuckle mounted between the eccentric and the

dynamometer 0 During operation of the machine, the test specimen is

subjected to a range of stress which is controlled by the preset,

constant deflection of the eccentric. The stress is determined by

an Ames dial which measures the vertical deflection across the

throat of the dynamometer to the nearest 0.001 in.

ll.

To calibrate the fatigue machine, SR4 strain gages were

mounted on a calibration blank which was bolted into the machine in

place of the test specimeno The readings of the Ames dial across the

throat of the dynamometer were then calibrated against the calibration

blank. Both fatigue machines had. a calibration constant of 3150 lbs.

per 0.001 in.

ID placing a specimen in the fatigue machine, a stress

equal to the maximum desired stress was first statically applied

. to the spec:imen after the bolts had been tightened 0 Plastic straining

occurred for those specimens to be tested above their yield pointSj

hence 7 it was necessar,y to allow sufficient time for the specimen to

strain statically under the maximum load before the load. could be

maintained under repeated loadings 0 Whenever plastic straining

occurred, the final reduced dimensions of the specimen were recorded,

but the reduction of area was negligible in all cases 0 After the specimen

had stopped yielding under the static application of the maximum stress,

the load. range was set on the eccentric and the machine was started 0

The fatigue machine could be run continuously because it was

equipped with a micro-switch which stopped the machine if the maximum

deflection of the specimen increased 0 The load was checked frequently

at the start of a test, but later only as often as neceSS&""""J to mah"1tain

the desired loado Failure is defined in these tests as the number of

cycles of loading resisted by the specimen before the micro-switch

stopped the fati~~e machineo In most cases the failure occurred when

the machine was not being watchedo

12.

One of the two fatigue machines was originally operated with

a control mechanism which continuously maintained the maximum load

during a test. The procedure of mounting a specimen in the grips

of the machine was the same as the procedure used when the dynamometer

was employed as the load measuring device, but once the machine was

started, the load was continuously, automatically maintained during

the test. The accuracy of the load measurement with the automatic

equipment was about the same as the accuracy using the manually

operated dynamometer. Use of the automatic equipment had to be

discontinued early in the program after some of the mechanical parts

of the system fractured in fatigueo

13.

III.. ~S OF LONGITUDINAL Burr WELDS IN ASTM A7 STEELS

7. Results of Present Fatigue Tests of Longitudinal Butt Welds

The results of the tests of the longitudinal butt welds and

of the corresponding plain plates are presented in Tables 4 and 5,

respectively, and in Figo 11.. The tabular results contain the test

data in terms of the maximum stress and the number of cycles to

failure. In addition to the test data, the fatigue strengths corresponding

to failure at 100,000 and at 2,000,000 cycles have been computed from

the formula* f = S (!)k, in which S is the stress corresponding to a n

test failure after N cycles,- f is the stress corresponding to failure

after the desired number of cycles, n)' and k is an experimentally

determined constante

Based on the result of previous tests, the values of k have

been set equal to 0013 for the weld tests and to 0018 for the plain

plate testso A variation in the value of k from 0013 to 0018 changes

the value of the computed fatigue strength corres:9onding to :failure

at 100,000 or at 2,000,000 cycles by approximately 9 percent if the

test specimen :failed after 500,000 cycleso The closer is the experimental

number of cycles to the desired number of cycles] the smaller will be

the difference in the computed fatigue strengths a The interpretation

o.f the test results in terms of a fatigue strength corresponding to

a predetermined number of cycles simplifies the comparison of the

test results 0

*See University of Illinois Engineering Experiment Station ~illetin 302, -p. Ill.

14.

The results of the plain plate tests are given at the top of

Fig. ll. All of the plain plate specimens were tested at stresses

greater than the yield strength of the plate. Above a stress of

40 ,000 psi, it was difficult to maintain the load on the test specimen

because it would strain plastically until the range of the adjustment

turnbuckle was exceeded. In Table 5, the plain plate fatigue strength

corresponding to failure at 2,000,000 cycles has been tabulated, and

these values, 34,600 psi for the specimens with the mill scale on and

35,300 psi for the specimens with the mill scale off, were greater

than the yield strength of the base plate, which was 33,100 psi. After

their removal from the fatigue machine, the fractured specimens were

pulled statically in order to examine the fracture surfaces 0 Typical

fracture surfaces of the plain plate fatigue specimens are shown in

Fig. 12.

The results of the tests of the longitudinal butt welds

are given in the lower half of Fig. li. In this figure, the o:pen

symbols represent the tests of the specimens with the reinforcement

on and the closed symbols the tests of the specimens with the

reinforcement ground off 0 From this figure, it is apparent that the

spec~ens tested with the reinforcement off generally have a higher

fati5~e strength than the specimens with the reinforcement on)

and that the welds produced with the E7016 electrode generally

are stronger in fatigue than the welds produced with the E6010

electrode.

15-

A closer comparison of the results of the welds tested with

the reinforcement on can be made by using the values in Table 4 for the

fatigue strengths correspqnding to failure at 100,000 or at 2,000,000

cycles. The values in this table show that the welds produced with the

E70l6 electrode are about 4,000 psi stronger than those produced with

the E60l0 electrode for failure at 100, 000 cycles.o For failure at

2,000~000 cycles, the E1016 welds are about 2,000 psi stronger than

the E6010 welds. Thus, for the longitudinal butt welds tested with

the reinforcement on, the use of the E7016 electrode results in a

greater improvement in the fatigue strength corresponding to failure

at a relatively low number of highly stressed cycles than at a

relatively large number of' less highly stressed cycleso

Removal of the weld reinforcement had a different effect

on the fatigue properties of the two types of weldso The E7016

longitudinal butt welds had fatigue strengths which were increased

about 7,000 psi for failure at 100,000 and about 4,000 psi for

failure at 2,000,000 cycleso On the other hand, the increase in the

fatigue strength of the E60l0 welds corresponding to failure at

2,0°°3000 cycles was about 5,000 psi, but the increase in the fatigue

strength corresponding to failure at 100 JOOO cycles was only about

2,000 psio Therefore, for failure occurring at 100,000 cycles, the

E70l6 longitudinal butt welds tested with the reinforcement off are

about 9,000 psi stronger than the E60l0 welds, whereas for failure

at 2,000,000 cycles, the two types of weld have about the same

fatigue strength.

In general, the results of the fatigue tests of the

longitudin~ butt welds, both with the weld reinforcement on and

with the weld reinforcement removed by grinding, indicate that

16.

the E7016 welds are likely to be more resistant to repeated loadings,

especially under conditions in which failure occurs after a relatively

low number of highly stressed cycles.

Examination of the fracture surfaces, as summarized in

Table 6, provides an interesting comparison of the behaviour of

the welds produced with the two types of electrodes. For the specimens

tested with the weld reinforcement on, all but one of the spec~ens

started to fracture at the surface of the weld, with most of the

fractures starting from a region of a surface pass in which the

electrode was changed during welding. Photographs of some tYJ?ical

fractures are shown in Fig. 13. The one specimen which did not fail

from the surface of the weld, specimen El9, an E6olO specimen,

started to fracture at an internal. defect, but the .fatigue strength

of this specimen was not significantly different from the fatigue

strength of the other E6ol0 specimens. All but one of the E6olO

spec~ens with the weld reinforcement on had a defect in the fracture

surf'ace , although the fracture originated from a defect in only one

spec~en. It appears that the severe condition existing at the

change of electrode is likely to cause the initiation of a fatig~e

failure, but the failure might not occur at a stress significantly

less than that of a specimen for which the fracture had initiated

away from a change of electrode.

17.

There was a more consistent difference in the initiation of

fracture for the specimens tested with the weld reinforcement ground

off'. As sUlIlIllarized in Table 6, all of the failures of the E6010

specimens with the reinforcement off initiated at an internal defect,

whereas all but one of the failures of the E7016 specimens initiated

from the surface of the weld. Typical photographs of the fracture

surfaces of the specimens tested with the reinforceme~t removed are

shown in Fig. 14.. The one E7016 specimen which fractured at a defect

in the weld metal, had a fatigue strength about the same as that of'

the other E7016 specimens.

In the discussion presented above, it was pointed out that

at the higher stress levels, removal of the weld reiriforcement did

not increase the fatigue strength of the E6010 welds nearly as much

as that of the E7016 welds. The smaller improvement in the fatigue

strength of the E6010 welds is probably caused by the defects, from

which f'racture originated in the weld metal. Thus, because the weld

metal contains fewer defects, use of the E7016 electrode for

longitudinal butt welds might prove advantageous for those applications

in which a weld with the reinforcement ground off is required to

withstand a relatively low number of highly stressed cycles.

8. Results of Present Static Tests of Longitudinal Butt Welds

Static tests were made of the longitudinal butt welds and

of the plain plates having the same dimensions as the fatigue specimens.

The results of these tests are presented in Table 7. All of the weld

18.

specimens, both those with the reinforcement on and with the reinforcement

off, had yield points of about 39,000 psi which was about 5,000 to 7,000

psi greater than the yield point of the plain plate specimens. Removal

of the ndll scale increased the yield point of the plain plate specimen,

but removal of the weld reinforcement had little effect on the yield

point of the weld speQ~ens.

The plain plate specimens all had maximum stresses which

were below about 54,000 psi. The weld specimens, on the other hand,

had maximum strengths greater than about 60,000 pSi" the increase

in strength compared to that of' the plain plate specimens probably

being a result of the greater strength of the weld metal. Removal of

the weld reinforcement considerably reduced the max~ strength of

the E6ol0 butt welds and slightly reduced the maximum strength of the

E7016 butt welds.

The elongation of the E70l6 specimens was slightly greater

than that of the E6010 specimens probably because of: the greater

ductility of the E70l6 weld metal.

9. Review of Previous LOngitudinal Butt Weld Fatigue Tests

The results of the past3 and the present longitudinal

butt weld fatigue tests are presented in Fig. 15 and Table 8. The

properties of the plain plate used in these tests are given in Table 9·

(a) Specimens Tested with the Weld Reinforcement on:

For the longitudinal butt welds tested in the as-welded

condition, the E60l0 welds of the present (1953-54) series of tests

have about the same average fatigue strength as the previous tests,

but the E70l6 welds have an average fatigue strength about 2,000 psi

. greater for failure at 2,000,000 cycles and about 3,000 to 4,000 psi

greater for failure at lOO,OOa cycles. The previous tests of specimens

with high temperature or mechanical stress relief treatments have a

greater fatigue strength than the E60IO welds without treatment, but

the fatigue strength of the E70l6 welds is about as great as the

specimens subjected to either of these two treatments. Only the E60l0

welds subjected to a low temperature stress relief treatment had a

substantially greater (about 10 to 20 percent) average fatigue strength

than the E7016 welds without treatment.

(b) Specimens Tested with the Weld Reinforcement Ground Orf:

Without subsequent treatment, the previously tested E6010

welds had a higher average fatigue strength for fracture after

100,000 cycles and a lower average fatigue strength ~or fracture at

2,000,000 cycles than the 1953-54 series of E6010 welds. For the

spec~ens with subsequent heat treatments, removal of the weld bead

increased the fatigue strength by 10 to 20 percent, except for the welds

subjected to a low temperature stress relief treatment, in which case

the fatigue strength corresponding to failure at 100,000 ~cles was

less with the reinforcement off than with the reinforcement on. For

the s~ecimens with the reinforcement removed, the present series of

E7016 welds performed as well as any of the tests of the longitudinal welds.

20.

In the previous tests, the specimens tested in the as-welded

condition but with the weld reinforcement off had a fatigue strength

at 100,000 cycles as great as, or greater than, the specimens with a

stress relief treatment, but all three stress relief treatments (high

temperature, mechanical, and low temperature) significantly increased

the fatigue strength corresponding to failure at 2,000,000 cycles.

2l.

IV. TESTS OF TRANSVERSE BUTr WELDS IN ASTM A 7 STEELS

10. Results of Present Fatigue Tests of Transverse Butt Welds

The results of the transverse butt weld tests are plotted on

an SN diagram in Fig. 16 and are summarized in Table 10. The results

of the plain plate tests, given in Table 5, are also included in

the SN diagram"

One of the most important features of the SN diagram is that

the test results for all of the specimens fall within a relatively

narrow scatter band, regardless of the variations in the welding procedure.

* The average fatigue strengths reported in Table 10, corresponding to

failure at 100,000 and 2,000,000 cycles, allow a closer examinat ion of

the relative fatigue strengths of welds produced with the E6010 and

with the E7016 electrodeso For the specimens tested in the as-welded

condition, both types of weld have about the same average fatigue

strength, about 24,000 psi, corresponding to failure at 2,000,000

cycles, whereas the E7016 welds have an average fatigue strength

corresponding to failure at 100,000 cycles which is about lO percent

* The values for fatigue strength were computed from the experimental

-data using the empirical relationship f = S (!!)k, in which the value n

of k has been assumed equal to O~13 for the weld specimens and

to 0.18 for the plain plate specimens.

22 ..

greater than that of the E6010 welds, which was about 35,000 ps i.

For the specimens tested with the weld reinforcement ground

off, the E7016 welds have an average fatigue strength about 10 percent

greater than the E6010 welds for failure at 100,000 cycles and about

30 percent greater for failure at 2,000,000 cycles. Removal of the

weld reinforcement has decreased the fatigue strength of the E6010

welds corresponding to failure at 100,000 and at 2,000,000 cycles.

The E70l6 welds with the reinforcement off also have a lower average

fatigue strength for failure at 100,000 cycles than the E70l6 welds

with the reinforcement on; however, the average fatigue strength

corresponding to failure at 2,000,000 cycles is greater by about

20 percent.

An important aspect of the tests was the investigation of

the fracture surfaces to determine the origin of fracture 0 Several

interesting features have been noted 0 For the E6010 welds with the

reinforcement on, the fractures originated more often in the weld

metal (five specimens) than at the edge of the weld (two specimens)o

This can be interpreted to mean that fracture is imminent at both

locations of the welded jOint, and that, in order to improve the

fatigue properties of the E60l0 transverse butt welds, both regions

of the joint would have to be strengthened 0 In contrast to the

E6ol0 welds, all of the E70l6 transverse butt welds tested with the

reinforcement on started to fail at the edge of the weld, indicating

that the fatigue properties of the complete joint are controlled by

the fatigue strength of the section at the edge of the weld. For the

23·

E70l6 welds, improvement of the fatigue resistance at the edge of the

weld would probably increase the fatigue strength of the entire joint.

Photographs of typical fracture surfaces are shown in Fig. 17.

For all but one of the transverse butt welds tested with

the reinforcement ground off, the fractures initiated at an internal

inclusion or gas pocketo The one exception was an E70l6 specimen which

failed in the base plate away from the weld at a higher fatigue strength

than the other specimens which failed at a defect in the weld metal.

It is interesting to note that the specimens with the weld reinforcement

removed generally had fatigue strengths slightly lower than those of the

specimens with the weld reinforcement on, with the exception of the

fatigue strength of the E7016 welds corresponding to failure at

2, 000 ~ 000 cycles. It can only be assumed that if' the weld metal were

free from defects, the fatigue strength of the transverse butt welds

with the reinforcement off would be greater. Typical photographs of

the fracture surfaces of the welds tested with the reinforcement

removed by grinding are shown in Figo 180

A metallurgical examination was made of some of the fractured

specimens along a plane normal to the fracture surfaces and through

any defect in the fracture surface. The plane was polished and etched

with 2 percent nital. The examination of the polished surfaces confirmed

the location of fracture determined from the appearance of the fracture

surface. Typical macrographs of the polished and etched sections are

shown in Fig. 190 The metallurgical examination was discontinued after

it had been determined that the point of initiation of fracture could

be satisfactorily determined from the appearance of the fatigue fracture

surface.

,... ..

24.

11. Results of Present Static Tests of Transverse Butt Welds

static tests were made of the transverse butt weld.specimens

in the as-welded condition and with the weld.reinforcement ground off

and of plain plate spec imens with the mill scale on and with the mill

scale ground off. The specimens were prepared and welded in the same

manner as the specimens used for the transverse butt weld fatigue

investigationo The plain plate spec:i:mens are the same as those reported

in Table 7.

A 600,000 lb. Riehle testing machine was used to test the

specimens to complete :failure at a strain rate of 0.10 in. per minute.

The results, in Table 11, are reported in terms of the yield point,

the maximum strength, the reduction of area, and the elongation

in 6 in. (measured along the edges) 0

All of the statically tested, transverse butt welds fractured

outside of the weld so that the static properties were those of the

plain plate. The ductility, yield strength, and the maximum strength

were all approximately the same as that of the plain plate specimens.

The maximum strengths were significantly below that required by the

ASTM specification A70

12. Scope of the study of Previous Fatigue Tests of Transverse Butt Welds

In order to determine what advantages might accompany the use

of various welding techniques, a re-evaluation has been made of the

fatigue tests of transverse butt welds conducted previously at the

25·

University of IllinOis),4,5,6,7. It was hoped that this investigation

would allow a more critical examjnation of the limits of the fatigue

strength of transverse butt welded joints and of the variables which

affect these limits.

The design of the s~ecimens tested previously was generally of

the type shown in Figo 6b, although some of the specimens were of

different thicknesses and others were of a reduced sizeG Most of the

previous tests were of butt welds in steels which were to meet the

AST.M specification A7~ A discussion of the previous tests of transverse

butt weld.s in other steels is contained in Chapter 5 0 The~hysica1

properties and the chemical analyses of the A7 steels used in the past

and the present butt weld fatigue tests are summarized in Table 120

Several of the steels do not meet the AST.M specification for yield

point and tensile strength; however, the tensile properties of the

different steels are similar 0 The chemical compositions are also quite

similar, the steels haVing a carbon content in the range from 0015 to

0.26 percent, values which are common for A7 steels 0

Comparison of' the fatigue strengths of the plate materials

indicates that, in the absence of stress raisers, the steels behave

in somewhat· the same manner when subjected to repeated stresses 0 This

similarity is fUrther confirmed in Fig. 20 which gives the results of

the past and present ~atigue tests of the plain plate specimens of

the steels used in the transverse butt weld programs. It is interesting

to note that only the steel used in the present (1953-54) program had a

fatigue strength corresponding to failure at 2,000,000 cycles which

exceeded the coupon yield strength 0

26.

The tabular results of the butt weld tests contain information

concerning the chemical analysis of the steel) the preparation of the

test spec±mens) the average fatigue strengths corresponding to failure

at 100,000 cycles and at 2,000,000 cycles) and the location of fracture.

The figures contain the SN diagrams. For those SN diagrams containing

the results from the welds prepared with individual electrodes, the

test data are labeled to indicate the location of failure.

The discussion of the transverse butt weld tests will be

divided into the following sections:

(1) Fatigue tests of transverse butt welds with the reinforcement

on: O-Tension Cycleo


on; Tension = Compression cycle.


off.

13. Review of Past and Present Fatigue Tests of Transverse Butt Welds

with the Reinforcement On; O-Tension Cycle

(a) Welds without special pre- or post-weld treatments:

The results of the tests of the butt welds produced with the

E6010 -electrode are given in Table 13 and in Fig. 21. Generally J the

specimens ,welded in the flat.· position had fatigue strengths which were

about 10 to 30 percent greater-·than those of the specimens welded in

the vertical or the horizontal positions, the average fatigue strengths

27·

corresponding to failure at 2,000,000 cycles varying fram 17,000 psi

to 24,000 psi. The tests of the ·specimens welded in the flat-overhead

. position had too much scatter and were too limited in number to make

a reasonable comparison, the results varying ~ram among the strongest

specimens to among the weakest. The effect of the position of welding

on the fatigue properties is indicated by comparing the open symbols,

which are the results of the specimens welded in the flat position,

with the closed symbols, which are the results of the specimens welded

in the other positions~

As indicated by the results in Table 13, E6010 specimens

prepared in commercial fabrication shops are likely to have a slightly

lower fatigue strength than those prepared in the laboratory,· although

it is possible that the commercial welds produced today might perform

somewhat better than those produced in 1941, when the previous specimsns

reported in Bulletin 344 were welded. In this regard, the laboratory

prepared specimens produced in the present, 1953-54 series of tests

were superior to any of the previous tests of E6010 butt welds in the

as-welded condition.

Tabulation of the locations of fracture of the E60l0 specimens

indicates that 18 spec:il:nens failed in the weld metal, 8 in the weld

metal and at the edge of the weld, 21 at the edge of the weld, 4 at the

edge of the weld and in the plate, and 4 specimens failed in the plate

away·~rom the weld. To further investigate the effect of the location

of failure, the exper:il:nental data in the SN diagram of Fig. 21 have

been labeled to indicate the location of failure. It is ap:parent in

28.

this diagram that the lower fatigue strengths are generally associated

with failure in the weld metal, and that if failure at this ~ortion of

the joint could be eliminated, the fatigue strength of the welds ~roduced

with the E6010 electrode might be somewhat greater.

The results of the tests of the welds de~osited with the

E6012 electrode, given in Table 13 and in Fig. 2l, again show the

scatter obtained from tests of butt welds ~roduced with a single

classification of electrodeo In general, the E6010 welds had about the

same fatigue strength as the E6010 welds. As with the E6010 welds,

the specimens welded in the flat position are stronger in fatigue

than those welded in the vertical position, and those welded in the

flat-overhead position ~roduce intermediate results. All of the welds

~roduced with the E6012 electrode were welded in commercial fabrication

shops. Tabulation showed that 9 failures occurred in the weld metal,

6 in the weld metal and at the edge of the weld, 12 at the edge of

the weld, 3 at the edge of the weld and in the plate, and 1 in the

plate away from the weldo The location of failure has been indicated

on the results in the SN diagram, but a correlation does not exist

between the location of fracture and the high or the low values of

fat~e strength, as was the case for the results of the tests of the

E6010 welds.

The results of the limited number of tests of welds deposited

with the 'E6013 electrode are ~resented in Table 13 and in Figo 220 The

tests are too limited to evaluate the effect of the position of welding.

The tabulated results show that 2 s~ecilnens failed in the weld metal

29·

and at the edge of the weld, 5 at the edge of the weld, 4 at the edge

of the weld-and in the plate, and 1 failed at all regions of the

spec:iJnen, but noting the location of fracture on the SN diagram does

not indicate a correlation with the fatigue behaviour of the E6013

welds. None of" the welds failed in the weld metal, and this might

account for the slightly greater fatigue strengths obtained :from the

tests of the welds produced with the E6013 electrode than from the

tests of' the welds prepared with the E6ol0 electrode.

As only three specimens prepared with the E6020 electrode

were tested, the limits of' the fatigue behaviour could not be

determined. The results are given in Table 13, but are not plotted

on an SN diagram.

The transverse butt welds prepared with the E6030 electrode

were all welded in the f'lat position and the results are given in

Table 13 and in Fig. 22. Neither the tabulation of' the location of

failure nor the notation of the location of failure on the SN diagram

indicates a correlation with the f'atigue behaviour of the individual

specimens 0

The tests of the E70l6 welds have been discussed in Section

1.0; in the discussion which follows, the results will be compared with

the tests of the welds produced with other electrodeso

The use of automatic welding procedures produced welds which

behaved no more consistently in fatigue than the manually deposited

welds. The results of the tests of the automatically prepared welds,

30.

given in Table 13 and in Fig. 23, do not indicate a correlation between

the location of the fracture and the fatigue behaviour.

Comparison of the welds produced with the different electrodes

is facilitated by use of the summary Table 14, which contains the average

values of the reported tests, and the summary Fig. 24, in which all of the

test results of the welds in the as-welded condition have been plotted on

an SN diagram. The summary SN diagram shows the very large amount of

scatter encountered in the tests of the butt welds prepared with the

various types of electrodes. For instance, for a stress cycle of zero

to 20,000 pSi, a transverse butt weld, produced with one of the standard

AWE specification electrodes under conditions as good as or better than

those which might prevail in a typical fabrication shop might fail after

as little as 200,000 cycles of loading or might withstand as many as

8,000,000 cycles without failing. In evaluating the test results in

terms of an average fatigue strength corresponding to failure at a

given number of cycles, this large amount of scatter is quite likely to

be overlooked. The average results which will be discussed in the next

paragraph are averages only, and the deviation from the average, clearly

indicated by the SN diagram, must be considered in designing a structure

to resist repeated loadings. It is likely that the fatigue strength of

an individual specimen will vary as much as 30 per cent from the averages

reported in the tables.

The results presented in the summary Table 14 clearly

illustrate the similarity in the fatigue behaviour of the welds produced

with the different electrodes. It is possible that welds produced

with the E7016 or with the E60l3 electrode will produce welds having

31.

a sligh~ly higher average fatigue strength than those produced with the

other electrodes_ For the welds deposited in the flat position, the

variation in the average fatigue strength of the welds prepared with

the- various electrodes is from 30,800 psi to 37,900 psi for failure at

100,,000 cye-les and from 18,000 psi to 23,800 psi for failure at

2,000,000 cycles. Individual specimens are likely to have a fatigue

strength corresponding to failure at 2,000,000 cycles, as low as

15,000. psi on a O-tension cycle. It is interesting to note that the

welds produced with the automatic procedures were among those having

the lower fatigue strengths.

(b) Welds with Special Pre- or Post-Weld Treatments:

The results of the specimens subjected to special treatments

are presented in Table 15 and in Fig. 25- Only 2 special treatments

have been tested; a high temperature stress relief treatment and a

series ·.of specimens in which each pass was peened. .All but one o:f the

specimens failed at the edge of the weld, the one exception failing

at the edge of the weld and in the plate. The results have been

retabulated in the summary Table 14, in which it is apparent that both

treatments produce welds having average fatigue strengths as great

as, but no greater than, most of the as-welded specimens; therefore,

the tests do not justify the use of either procedure solely to increase

the fatigue strength of the transverse butt welds.

14. Review of Past Fatigue Tests of Transverse Butt Welds

With the Reinforcement On; Tension = Compression Cycle

(a) Welds without special pre- or post-weld treatments:

32.

The results of the tests on a tension = compression cycle of

the butt welds in the as -welded condition are presented in Table 16

and in Figures 26 and 27. The SN diagrams indicate that the scatter in

the test results for the E6010 welds is greater than that for the E6010

welds tested on a O-tension cycle. The average fatigue strength

corresponcring to failure at 2,000,000 cycles on a tension = compression

cycle varied from 10,300 psi to 15,300 psi. With the exception of the

welds prepared with the E6010 electrode, the results are too l~ited

to determine any relationship between the location of fracture and the

fatigue strength. For the E6010 welds, there does not appear to be a

consistent correlation between the location of failure and the fatigue

behaviour of the welds, although many of the specimens having relatively

low fatigue strengths failed in the weld metal.

The results of the tests of the specimens produced with each

type of electrode and of those produced automatically have been

replotted in the summary SN diagram in Fig. 28. The most noteworthy

feature of this SN diagram is the large amount of scatter of the test

results. The lowest test result, an E6010 specimen, probably corresponds

to a fatigue strength of about 14,000 psi for failure at 100,000 cycles

and about 9,000 psi for failure at 2,000,000 cycles. With the extreme

variation in the test results, design by use of an average fatigue

strength corresponding to failure at a given number of cycles could

33·

well prove dangerous. The average fatigue strengths do, however, aJ.1ow

a c~arison of the results of tests of welds prepared with different

electrodes , although the comparison must be cons idered a crude one and

the scatter in the test results must not be overlooked.

The average fatigue strengths have been tabulated in the

sUlllII1.8XY Table 14 in order to compare, however crudely, the fatigue

strength of the welds produced with the different electrodes. Generally

speaking, the results are similar to those of the specimens tested on. a

O-tension cycle, with the tension = compression fatigue strengths being

approximately·60 to- 70 percent of the O-tension fatigue strengths. The

tests were too- limited to compare the ef''fect of the position of welding,

with the exception of' the E6010 welds, in which case it appears that

the specImens welded in the vertical position are again consistently

weaker in fatigue than the welds prepared in the flat positiono None

of the electrodes produced butt welds having an average fatigue strength

significantly different fram that of the welds produced with any of

the other electrodes. In view of the wide scatter and the limited

number of' tests, it can only be concluded that none of the electrodes

tested will significantly improve the average fatigue behaviour of

transverse butt welds tested on a tension = compression cycle,

although the scatter might be less for the welds produced from some

electrodes.

34.

(b) Welds with special pre- or post-weld treatments

The only specimens which were prepared with a special

procedure were a series of 7 specimens welded with the E6010 electrode

and subjected to a postweld stress relief treatment of 1200 deg. F.

Analysis of the data in Table 16 and in Fig. 29 indicates that the

results of the tests of the heat- treated welds are within, but on the

high side of, the scatter band of the tests of the E60l0 butt welds

without the post weld heat treatmento Additional tests might well

produce results in a scatter band overlapping that of the welds without

the stress relief treatment. The results have been included in the

summary in Table 14,.

150 Review of Past and Present Fatigue Tests of Transverse Butt Welds

With the Reinforcement Removedo

(a) Welds tested on a O-Tension cycle:

Only a limited number of tests were conducted of welds with the

reinforcement removed, and the results of these tests are summarized in

Table 17 and in Fig 0 30. .All of the welds, with the exception of the

E6010 welds of the 1953-54 series, had approximately the same average

fatigue strength corresponding to failure at 2,000,000 cycles. However,

there is a considerable variation in the average fatigue strengths

corresponding to failure at 100,000 cycleso Because the specimens were

tested at di:fferent times, it is possible that some testing technique

is inf'luencing the test results of those spec:ilJlens tested at stresses

which exceed the yield point of the material. The differences in the

behaviour of s:imilarly prepared joints is so extreme that no attempt

will be made to compare the results for failure at 100,000 cycles.

Additional closely controlled tests might explain the variation in the

test results ..

Tabulation of the data indicates that failure occurred 30 times

in the weld metal, 4 times at the edge of the weld, and. 9 times in the

plate. Notat ion of the location of' failure on the SN diagram indicates

that the fatigue strength is likely to be- greatE:r if failure in the

weld metal can be preventedo

A limited series of tests were conducted to determine the

effect of a high temperature stress relief treatment. Comparison of

the results of these tests, given in Table 17 and in Fig. 30, with those

of the specimens without stress relief shows that the stress relieved

specimens have a similar average fatigue strength, \ although the test

results were generally 'in the high range of the scatter band obtained

from the tests of the specimens without the stress relief heat treatment.

Tabulation showed that 9 S]?ecimens failed in the weld metal, 6 at the

edge of the weld, and 4 in the ~late; however, notation of the location

of failure on the SN diagram indicates that the location of failure did

not affect the fatigue strength.

Removal of the weld reinforcement has generally result~d in

an in the average fatigue strength; F0-thou---.gh some tests have

indicated a decrease in the fatigue strength. For the welds produced

with the E6010 and the E6012 electrodes, the O-tension average fatigue.

36.

strengths, given in Table 14, are increased fram 20 to 50 percent for

failure at 2,000,000 cycles, the average fatigue strength being about

26,500 psi for the welds produced with either of" these electrodes. For

failure at 100,000 cycles, the results vary widely, although improvement

generally results with removal of the weld reinforcemen~. In the limited

past tests, the improvement in the average fatigue strength for failure

at 100,000 cycles has varied from ° to 50 percent. It should be

remembered that some tests of E6010 welds indicated a decrease in the

fatigue strength of about 10 percent with removal of the weld reinforcement.

For the E70l6 welds, the fatigue strength is increased about 20 percent

for failure at 2,000,000 cycles, but is decreased about 6 percent for

fa;lure at 100,000 cyclese

(b) Welds tested on a Tension = Compression c~le:

Only a llmited number of fatigue tests on a tension = compress~cn

cycle were made of specimens with the weld reinforcement removed. Tbe

results of these tests, given in the SN diagram in Fig. 31, indicate

that the specimens subjected to a 1200 dego F. stress relie~-treatment

behaved about the same as the specimens without heat treatment, the

scatter bands overlapping for these two conditions 0 The tabulated

results in Table 17 indicate that the E60l0 welds might be slightly

stronger than the welds deposited automatically, but the difference is

probably not significant considering the amount of scatter likely to be

encountered if additional tests were performed. The average fatigue

strengths in Fig. 14 indicate that removal of the weld reinforcement

increases the tension = compression fatigue strength of the transverse

butt welds from 10 to 40 percent.

·v. REVIEW OF PAST FATIGUE TESTS OF TRANSVERSE BUTr WELDS IN SILICON

STEEL .AND m LOW ALLOY HIGH TENSILE STEEL

16. Scope of Investigation

Only a very limited number of fatigue tests have been made of I

transverse butt welds in structural steels other than plain carbonASTM A7

steels. Two series of fatigue tests have been conducted of transverse

butt welded joints in silicon steel and three series of tests of low

alloy high tensile steels 0 One additional series of tests has been made

of a transverse butt welded joint connecting an A7 steel to a low alloy

high tensile steel~

The chemical analyses and the physical properties of the steels

used in the past tests are presented in Table 180 The data in this table

indicate that the ductility and the tensile strength of the two types

of steel are about the same, but the yield point of the silicon steel

is about 50,000 psi compared to a yield point of from 55,000 psi to

60,000 psi for the low alloy high tensile steelso A direct comparison

cannot be made of the fatigue strength of the two types of steel because

the silicon steel was tested on a O-tension cycle, whereas the low alloy

steel was tested on a tension = compression cycle. If one assumes that

the tension = compression fatigue strength is 60 to 70 percent of the

* O-tension value , the fatigue strength corres:ponding to failure at

2,000,000 cycles for the low alloy high tensile steel would be

* See Section 14.

38.

38,000 psi to 44,000 psi on a O-tension cycle, compared to 36,000 psi

~or the silicon steel. The fatigue strength for failure at 2,000,000

cycles for the plain carbon ASTM A7 steel used in the past fatigue tests

varied from 30,000 psi to 55,000 psi. Thus, the silicon steel has about

the same fatigue strength as the A7 steels, whereas the low alloy high

tensile steel has a fatigue strength which may be about 10 percent

greater than that of the A7 steels.

17. Review of Past Fatigue Tests of Transverse Butt Welds in Silicon Steel

The results of the fatigue tests of the transverse butt welds

in silicon steel are given in Table 19 and in Fig. 320 Three specimens

were tested in the as-welded condition and three with the weld reinforcement

removed. The results of the tests indicate that the removal of the weld

reinforcement did not improve the fatigue behaviour of the joint, the

average fatigue strength corresponding to failure at 2,000,000 cycles

being about 25,000 psi for both series of testso The results of the

similarly prepared welds in plain carbon A7 steels, sunnnarized in Table

14, showed that the average O-tension fatigue strength corresponding to

failure at 2,000,000 cycles ranged from 17,000 psi to 24,000 psi in the

as-welded condition, and from 26,500 psi to 29,000 psi with the weld

reinforcement removed 0 Thus, the butt welds in silicon steel have

average fatigue strengths which are equal to or less than those of the

butt welds in plain carbon steel 0

The ratio of the yield point of the silicon steel to that

of the plain carbon A 7 steels was approximately 104 (50, 000 ps i/3 5 , 000 ps i) ,

whereas the ratio between their fatigue strengths was approximately 1.0 or

less. If the static design stress for a structure of silicon steel were

increased in the same ratio as the yield point, the structure would be

designed for an allowable stress of 28,000 psi, a value which exceeds the

fatigue strength for failure at 2,000,000 cycles on a O-tension cycle.

For silicon steel welded structures which must withstand repeated

loadings, the design stresses will have to be based on the ability of

the structure to withstand repeated loads. Because the fatigue strength

of the welded joints in silicon steel is about equal to that of similar

joints in plain carbon steel, it may be more economical to use the less

expensive plain carbon A7 steels' in structures subjected to fatigue

loadings.

18. Review of Past Fatigue Tests of Transverse Butt Welds in Low Alloy

High Tensile Steels

Only three series of .fatigue tests J all on a tension = compression

cycle, have been conducted of single-V butt welds in low alloy steels.

The results of these tests are presented in Table 19 and in Figo 32. For

these limited tests, the fatigue strength ranged from 23,000 psi to 26,400

psi for failure at 100,000 cycles, and from 15,400 psi to 18,000 psi for

failure at 2,000,000 cycles. Two of the series of specimens were tested

with the weld reinforcement on, whereas one of the series was tested

with the specimens ground flush on the root side only. Grinding on the

root side only did not significantly increase the fatigue strength.

40.

For the low alloy steels, the ratio of the yield point of the

low alloy steel to the yield point of the plain carbon steel was

approximately 1.65 (58,000 psi/35,OOO psi). In contrast, the ratio of

the fatigue strength of the butt welds in the same steels is only

about 1.2. If the design stress for a structure of low alloy steel were

increased in the-ratio of the yield points, the structure would be

designed for an allowable stress of 33,500 psi, a value which dangerously

exceeds the value of the average, tension = compression :fatigue strength

(approximately 17,000 psi) for the transverse butt welded joints. Thus,

as for the silicon steel, the design of a welded structure of low alloy

high tensile steel subjected to repeated loadings would have to be based

on the ability of the structure to withstand repeated loads.

One series of tests, reported in Table 19 and in Fig. 32, was

conducted of a transverse butt weld joining a plain carbon steel plate

to a low alloy high tensile steel plate. For this series of tests,

the fatigue strength for failure at 100,000 cyc~es was about the same

as that for the welds in the low alloy steels, whereas the fatigue

stre~h for failure at 2,000,000 cycles was about that of the welds

joining two plain carbon steel plates. Thus, as might be expected,

the fatigue strength of welds joining dissimilar steels is probably a

fun~tion of the fatigue strength of similarly prepared joints in the

weaker of the two steels.

41.

VI. SUMMARY AND CONCLUSIONS

19· Summary

It has been the purpose of the experimental program reported

herein to compare the fatigue and the static properties of butt welds

produced with the E6010 and the E70l6 electrodes. Two types of butt weld

specimens were tested~ the longitudinal butt weld was tested with the

applied stress parallel to the direction of welding, whereas the transverse

butt weld was tested with the a~plied stress perpendicular to the direction

of welding. The butt welds have been tested either with the reinforcement

on or with the reinforcement removed by grinding and were subjected to

a stress c,ycle which varied from a low tension of about 2,000 psi to a

maximum tension, which was ·of such magnitude that failure generally

occurred between 100,000 cycles and 2,000,000 cycles.

The results of the present tests are summarized in the

accomp~ing Table which gives the average fatigue strength for the E60l0

O-TENSION FATIGUE STRENGTH

E6010

fioo,OOO psi

Longitudinal Butt Welds

Reinforcement On ••••• o ••••••• 37,400 Reinforcement Ground Off •.••. 39,800

E7016 Reinforcement On •••••••..•••• 41,700 Reinforcement Ground Of'.f ••••• 48 ,300


E6010 Reinforcement On .••••••....•• 34,900 Reinforcement Ground Off .•..• 33,300

E7016 Reinforcement On •••• o ••• o •••• 37,900 Reinforcement Ground Off ••.•• 35,400

f 2,000,000 psi

24,500 29,600

26,300 30,200+

24,000 2l,800

23,800+ 29,lOO+

42.

and the E70l6 welds with the reinforcement on and with the reinforcement

removed by grinding. In interpreting the results, consideration must be

given to the scatter of the test results from individual specimens. In

many instances, the scatter bands from the tests of the two electrodes

overlapped each other.

In order to more thoroughly understand the behaviour of butt welds

subjected to repeated stresses, an extensive review of the fatigue tests

of longitudinal and transverse butt welds conducted previously at the

University of D.linois has been included in the report. These past

spec~ens were welded with different joint preparations, in different

positions, and with different electrodes. The tests reported were those

which were subjected either to a O-tension or a tension = compression

stress cycle, and the joints were tested either with the mill scale on

or with the mill scale off. Some specimens were subjected to special

pre- or post-weld treatments.

The average fatigue strengths corresponding to the various

conditions studied are included in Table 8 for the longitudinal butt

welds and in Table 14 for the transverse butt welds. In interpreting

these tables, it is of the utmost importance that the scatter of the

test results be considered. It is quite likely that individual

spectmens will have fatigue strengths varying as much as 30 percent

from the average fatigue strengths summarized in the tables.

20. Conclua ions

The t'ollowing conclusions are based on the results of the present

experimental. program.

(1) In the as-welded condition, the longitudinal butt welds

produced with the E7016 electrode generally had a greater O-tension

fatigue strength than the longitudinal butt welds prepared with the

E6010 electrode. For ~ailure at 100,000 cycles, the E7016 welds were

about 4,000 psi stronger than those produced with .the E6010 electrode,

but were only about 2,000 psi stronger for failure at 2,000,000 ~cles.

(2) Removal o:f the weld reinforcement significantly increased

the O-tension fatigue strength of the longitudinal butt welds produced

with both the E6010 and the E7016 electrodes. For f'ailure occurring at

100,000 cycles, the E7016 longitudinal butt welds tested with the

reinforcement ground off' were about 9,000 psi stronger than the E6010

welds, whereas for failure at 2,000,000 cycles, the two types of welds

had about the same fatigue strength.

(3) The results of all of the present transverse butt weld

tests fell within a relatively narrow scatter band. In the as-welded

condition, the welds produced with both the E6010 and the E7016 electrode

had the same fatigue strength corresponding to failure at 2,000,000

cycles, whereas the E7016 welds had a fatigue strength corresponding

to failure at 100 ,000 cycles which was about 10 percent greater than

that of the E60l0 welds.

44.

(4) With the weld reinforcement ground off, the transverse

E70l6 butt welds had a :fatigue strength which was about 6 percent

greater than that o:f the E6010 welds :for :failure at 100-,000 cycles and

about 20 percent greater for failure at 2,000,000 cycles. In general,

the removal o:f the weld reinforcement decreased the :fatigue strength

of the transverse butt welds by almost 10 percent: however, the

:fatigue strength o:f the E7016 welds :for :failure at 2,000,000 cycles was

increased by 20 percent with removal of the weld reinforcement.

The following conclusions are based on the results of the

re-evaluation o:f the past and the present :fatigue tests of transverse

butt welds in pl~in carbon ASTM A 7 steels.

(1) ~ large amount of' scatter was encountered in the past

fatigue tests of transverse butt welds; it is likely that the :fatigue

strength o:f an individual specimen will vary as much as 30 percent from

the average fatigue strength. For the welds deposited in the flat

position, the range in the average O-tension fatigue strength was from

30,800 psi to 37,900 psi for failure at 100,000 cycles and from

20,000 psi to 23,800 psi for failure at 2,000,000 cycles. Individual

specimens are likely to have a O-tension fatigue strength as low as

15,000 psi for failure at 2,000,000 cycles. For a ° to 20,000 psi

stress cycle, the transverse butt welds failed after as few as 200,000

cycles of loading or resisted as many as 8,000,000 cycles without

:failing.

r

(2) The scatter for the specimens tested on a tension =

compression cycle was even greater than that for a a-tension cycle.

Generally, the tension = compression fatigue strength was about 60 to 70

percent of the O-tension values. The lowest test result on a tension =

compression cycle corresponds to a fatigue strength of about 14,000 psi

for failure at 100,000 cycles and about 9,000 psi for failure at

2,000)000 cycles.

(3) The welds deposited in the flat position had a significantly

greater fatigue strength than those deposited in the horizontal or

vertical position.

(4) With the exception of the tests of the as-welded E6ol0

specimens tested on a O-tension cycle, there was no consistent correlation

between the location of failure and the fatigue strength. The lower

O-tension Iatigue strength of the E60l0, as-welded specimens were

generally associated with failure in the weld metal.

(5) Welds prepared with the E6ol3 or the E70l6 electrodes had

greater fatigue strengths than those produced with the other electrodes

tested. The welds produced with automatic procedures were among those

having the lower fatigue strengths.

(6) Neither peening of each pass nor heat treating the

completed weld increased the fatigue strength.

(7) Removal of the weld reinforcement generally, but not

alw~sJ resulted in an increase in the fatigue strength. The increase

in the fatigue strength varied considerably and was as great as 50 percent

in some instances 0

46.

The following conclusions are based on the results of the re

evaluation of the :past transverse butt weld fatigue tests in silicon

steel and in low alloy high tensile steel.

(1) The butt welds in silicon steel have average f'atigue

strengths which were equal to or less than those of the butt welds in

plain carbon steelo

(2) The average fatigue strengths of the butt welds in the

low alloy high tensile steels tested were about 20 percent greater than

those of the butt welds in plain carbon steels.

(3) For both the silicon steel and the low alloy high tensile

steels tested, an increase in the yield point and the maximum. strength

may result in a much smaller or no increase in the fatigue strength of

welded joints in the steels. As a consequence, the use of silicon steel

or of low alloy high tensile steel may prove dangerous in a structure

subjected to repeated loads if the allowable design stress is increased

in the ratio of the yield point of the steel to the yield point of the

plain carbon A 7 steels.

(4) A transverse butt weld joining a plain carbon steel to a

low alloy high tensile steel is likely to have a fatigue strength equal

to that of' a similarly prepared joint in the :plain carbon steel.

(5) In view of their increased use in welded structures, it is

imperative that additional fatigue tests o.f welded joints in the various

types of low alloy high tensile steels be conducted before extensive

application of' these steels is made in welded structures subjected to

repeated loadings. The very limited, past :fatigue tests of transverse

butt welds in low alloy high tensile steels indicate that the increase

in the fatigue strength, compared to that of simi] arly prepared trans

verse butt welds in plain carbon .ASTM A7 steels, varies :from 0 to 20

. percent, whereas there is at least a 50 percent increase in the yield

point ..

48.

BIBLIOGRAPHY

1. Harris, L. A., Matthiesen, R. B., and Sinnamon, G. K., "Effects of Hydrogen and. Related Variables on the Physical Properties of Welds on structural Steels,ft Report to the Ohio River Division Laboratories, Corps of' Engineers, U. S. Army, Contract DA-ee -017 -eng-96, Civil Engineering Studies, Structural Research Series No. 31, University of Illinois, Sept. 1952.

2. Matthiesen, R. B., and Harris, L. A., and Newmark, N. M., "Fatigue Properties o:f Weld Metal", Welding Journal, V. 32, N. 9, pp. 441s-453s, 1953·

3. Wilson, W. M., Munse, W. H., and Snyder, I. S., flFatigue Strength ' of" Various Types of" Butt Welds Connecting Steel Plates U

, University of" illinois Engineering Experiment Station, Bulletin Series No. 384, University of Illinois, March 1950.

4. Wilson, W. M. and Wilder, A. B., "Fatigue Tests of Butt Welds in structural Steel Plates n, University of illinois Engineering Experiment Station, Bulletin Series No. 310, University of illinOis, Jan. 1939.

5. Wilson, W. M., Bruckner, W. H., Coombe, J. V., and Wilde, R. A., ttFatigue Tests of Welded Joints in structural Steel Plates II, University of illinois Engineering Experiment Station, Bulletin Series No. 327, University of' illinois, Feb. 1941.

6. Wilson, w. M., Bruckner, W. H., McCrackin, Jr., T. H., and Beede, H. C., "Fatigue Tests of' Commercial Butt Welds in structural Steel Plates", University of' lllinois Engineering Experiment Station, Bulletin Series No. 344~ University of lllinois, Oct. 1953.

7. Wilson, W. M., Bruckner, W. H., Duberg, J. E., and Beede, H. C., "Fatigue Strength of Fillet-Weld and Plug-Weld Connections in Steel structural Members fl, University of illinois Engineering Experiment Station, Bulletin Series No. 350, University of illinois, March 1944.

TABLE l

CHEMICAL COMPOSITION OF STEEL PLATE , SPECIMEN CHEMICAL CONTENT TIl PERCENT

NUMBER C Mn P S Si

HT4 0.l8 0.68 0.016 0.039 0.03

HTl9 0.l6 0.68 O.Ol~ 0.039 0.04

HT34 0.l8 0.68 0.ol6 0.039 0.03

Average O.l7 0.68 o.ol6 0.039 0.03

TABLE 2

PHYSICAL PROPERTIES OF STEEL PLATE

STANDARD 8 IN. GAGE LENGTH TENSILE COUPON

UPPER TENSILE PERCENT PERCENT SPEC:IMEN YmLD STRENGTH ELONGATION REDUCTION

NUMBER STRENGTH IN 8 IN. OF AREA psi psi

PLATE 10

liT3 33,500 57,800 33 63 liT 5 32,900 57,900 33 64 BT13 32,700 57,800 33 61 HT15 31 z800 58z100 33 63

Average 32,700 57,900 33 63

PLATE 9

HT18 33,800 57,400 32 61 HT20 32,900 .56,200 31 63 HT23 31,800 57,300 32 63 HT25 33,500 57,000 32 62 -

Average 33,000 57,000 32 62

PLATE 8

HT32 32,300 56,700 31 62 HT36 31,100 55,500 33 60 HT40 32,400 57,000 34 63 HT43 32 z800 56z400 34 62

Average 32,100 56,400 33 62

PLATE 7 HT49 34,600 58,200 33 61 HT54 34z400 59 z000 32 62

Average 34,500 58,0-600 33 /-

b2

PLATE 6

HT59 36,800 58,300 33 58 m:62 35,700 57,600 34 61 -

Average 36,300 58,000 34 60

AVERAGE OF .ALL TESTS

33~300 .57·400 33 62

TABLE 3

WELDING PROCEDURES

OPEN CIRCUIT ELECTRODE DIAMETER CURRENT VOLTAGE VOLTAGE

in. amp volts volts


E6010 first pass 3/16 125 25 81 2-6 pass 3/16 160-170 26 83

E7016 first pass 3/16 145 20 82 2-6 pass 3/16 200 20 83

Lon~itudinal Butt Welds

E6010 first pass 3/16 135 24 85 2-6 pass 3/16 150-160 25 85

E7016 first pass 3/l6 l65 20 84 2-6 pass 3/l6 200 2l 86

'r

TABLE 4

RESULTS OF LONGITUDINAL BUTT WELD FATIGUE TESTS

CYCLES FATIGUE STRENGTH CYCLES . FATIGUE STRENGTrf SPEC. STRESS TO f 100,000 f SPEC. STRESS TO f 100,000 f

FA~URE 2,000,000

FAlLURE 2,000,000

psi 103 psi psi psi 103 psi psi

Specimens Tested With Reinforcement On E6010 E7016

H7 40,100 86.0 39,300 H9 40,000 184.0 43,400 Hl9 30,000 451·7 36,400 24,700 H74 29,500 751.1 26,000 H77 30,000 417.4 36,200 24,400 H88 30,000 447.3 36,700 24,800 H79 36,000 89.2 35,600 H90 39,500 177·3 42,600 H91 29,600 304.5 34,200 H98 30,000 1,223·7 28,200 H97 40,000 123.0 41,100 HlOO 40 ,000 207·7 44,000 Hl03 36,000 180.6 38,900

Average 37,400 24,500 Average 41,700 26,300

Specimens Tested With Reinforcement Off

Hll 40 ,000 124.2 41,100 IUS 40 ,000 503·7 49,400 H2O 40,000 3S.5 36,300 H76 30,000 4,514.7* 30,000+ H75 30,000 2,883.2 30,000 m8 . 29,500 1,244.0 27,800 H89 39,800 169.6 42,600 H92 39,900 361.5 47,100 H93 29,500 1,845.8 29,200 H96 40,400 400.1 48,400 32,800 H99 40,000 82·9 39,000 Hl02 30,000 3,189.1* 30,000+ Hl01 30,000 1,803.7 29,600

Average 39,800 29,600 Average 48,300 30,200+

* Specimens did not fail

~\. = 0.13

TABLE 5

RESULTS OF PLAIN PLATE FATIGUE TESTS

CYCLES FATIGUE STRENGTIf' SPEC. STRESS TO f'100,000 f'

FAILURE 2,000,000

psi 103 psi psi

Mill Scale On

HIO 35,500 1,508.0 35,200 H46 40,000 450.3 33,000 H49 37,000 1,078.3 34,100 H58 37,200 1,627.2 36,200

Average 34,600

Mill Scale Of'f'

H5l 39,000 810.9 34,700 H56 37,400 . 1,137·5 34,800 H61 36,400 3,454.1* 36,400+

Average 35,300+

~ = 0.18

* S~ecimen did not fail

~~ABLE 6

_ LOCATION OF INITIATION OF FATIGUE FRACTURES IN LONGITUDINAL BUTI' WELD SPECIMENS

WELDS DEPOSITED WITH E6010 ELECTRODE WELDS DEPOSITED WTIH E7016 ELECTRODE

\Velds tested l,.,ith the reinforcement on

HI

Hl9

H77

H79

Defect in fracture surface, but fracture probably started at surface of weld at change of electrode.

Fra~ture started at internal defect.

Fracture starT,e(l all I:JU:J.:l0 a.ce aI' '·leld. Defects in fracture surface.

Fracture started at surface of 1,veld. Defects in fracture surface.

h;jJ. Fracture started at. surface of w'eld at change of electrode. Defect in fracture surface.

H97 Fracture started at surface of 1-Teld at change of electrode. Defects in fracture surface.

Hl03 Fracture started at surface of weld at change of electrode. Defects in fracture surface.

H9 H74'

H88

H90

Fracture started at surface of weld.

Fracture started at surface of '\-leld at change of electrode ..

Fracture started at surface of weld at change of electrode.

Fracture started at surface of weld at change of electrode.

H98 Fracture started at surface of weld.

HIOO Fracture origin not determined. Specimen was not pulled statically because weld joining test section and pull head failed. Fracture occurred at a change of electrode.

Welds tested with the reinforcement removed

Hll Fracture started at internal defect. Bl8 Fracture started at surface of vleld. H2O Fracture started at internal defect. H76 Specimen did not fracture. H75 Fracture started at internal defect. H78 Fracture started at surface of weld. H89 Fracture started at internal defect. H92 Fracture started at surface of weld. H93 Fracture started at internal defect. H96 Fracture started at defect in weld at surface. H99 Fracture started at internal defect. HI02 Specimen did not fracture. mOl Fracture started at internal defect.

TABLE 7

RESULTS OF STATIC TESTS OF LONGITUD:rnAL :BUTr WELDS

PERCENT PERCENT YIELD MAXIMUM ELONGATION REDUCTION

SPECJMEN POnn' STRENGTH IN 6 IN.* OF AREA psi psi

Plain Plate Tests

H8 Mill Scale on 32,lOO 53,100 42 53 H52 Mill Scale on 31,600 52,900 4l 5l

H63 Mill Scale off 34,000 54,200 42 52

Tests of E60l0 Welds

Hl05 Rejnforcement on 39,300 66,400 31 40

H95 Reinforcement off 3S,9uU 59,700 33 37

Tests of E70l6 Welds

El04 Reinforcement on 39,700 6l,400 38 41

H94 Reinforcement off 38,400 60,100 39 47

*Gage lines along edges of specimen.

.~

TABLE 8

RESULTS OF PAST AND PRESENT FATIGUE TESTS OF LONGrTUDINAL BUTT WELDS LOCATION STEEL PREPARATION OF SPECIMEa-OF DATA AVERAGE LOCATION OF

" CH~ICAL FATIGUE STRENGTH FRACTURE C(}1POSITION IN PERCENT ...

z Z 10 lU lU..I '" '" lU

0 ffi O::lU<I) .... '" .... (I) (I) lU '" C) :z (I) <I) ;: Q.:II ... c " c

Z '" ;: W W Z 0: 0 2 0 12 0 ..I 0 ..I o 0: c ::> 0 0 0: C) CID Q. LIJ Q. - ::> (I) 2 Z Z 0 0: 0 ....... W 0 W W ~ 0 C 0 ... c 0: - 0 cto~ to 0-

.. U ..J W ... .... Lt.. 0- .., .., 000 '" -..I ;:r 0 CO U ::; ZOo. 0 0 - Z -0 .... (l)W .... (1) 0(1) '" c: c ..I ;: 0: Z -lU 0 lU (I)

o Q. -< 1.1.1..1 O:a.: O~ ~ 0 W W.... u.. :::> w c c 0 00:: 0: .... 0 w W 1.1.1 0:0 CD ... m~g~:to co (I) .... u ::E: -, Q. Q. LIJ a.. 0:: /))0:0: til; - (\.I 0 .... LL. Lt.. :w: =-=- LIJ -<:.:: -(A) SPECIMENS T!SlED IN THE AS~WELDED CONDITION

38'l GI in C 8.21 00 '+8 SING. V MANUAL E6010 FLAT ON NONE O-T 38.5 2'+.3 0.15 6

~-~ 1953coS\ i" C 00 17 0.68 00 03 OSL. V MANUAL E6010 FLAT ON NONE O-T 87.'+ 2'+.5 0.1 a 7 _ .. .., 1953-5~ In C 00 17 0.68 00 03 OSlo V MANUAL £1016 FLAT ON NONE I-T if1.7 26.8 0.18 6

38'+ Gil ~" C 00 21 0.'+8 SING. V MANUAL E6010 FLAT ON HIGH TEMPERATURE STRESS RELI!F O-T '+0.8 26.1 0.1'+ 6

38lt HIli !" C 0.21 D0 48 SING. V MANUAL E6010 FLAT ON Low TEMPERATURE STRESS RELIEF O-T 49.' 28.2 0.19 6

384 Gill !" C 6,,21 00 48 SING. V MANUAL E6010 FLAT ON MECHANICAL STRESS RELIEF O-T 39.9 27.7 0.13 7

(a) SPECIMENS TESTED WITH THE WELD REINFORCEMENT R8HOVED

38 .. GIV ~" C 00 21 0,,'+8 SING. V MANUAL E6016 FLAT OFF NONE (FILED) O .. T "9.3 25.3 0.23 6

--- 1953-5" in c 0 .. 17 0.68 0,,03 Dal. V MANUAL E6010 FLAT OFF NONE (GROUND OFF) O-T 99.8 29,6 8.13 7

....... 1953~54 in C 0 .. 17 00 68 00 03 OBl .. V MANUAL E7016 FLAT OFF NONE (GROUND OFF) O-T '+8.3 30.2+ 0.19 \ 2

384 GV ~" C 00 21 0."8 SING. V MANUAL E6010 FLAT OFF HIGH TEMPERATURE STRESS RELIEF O .. T "8.3 32.6 00 13 S

384 HIV ~" C 0.21 00 48 SING. V MANUAL E6010 FLAT OFF Low TEMPERATURE STRESS RELIEF O-T 46.3 32.7 e .11 7

384 GV I l" C 0.21 0.lt8 SING. V MANUAL E601. FLAT OFF MECHANIOAl STRESS RELIEF O-T "'l.S al.5 0.12 6

Yield STEEL Point

psi

Bulletin 384 35,800 ASTM A7

TABLE 9

PROPERTIES OF PLATE MATERIAL USED m PAST AND PRESENT LONGrruDINAL BUTr WELD FATIGUE TESTS

PHYSICAL PROPERTIES CHEMICAL CONTENT Tensile Percent Percent IN PERCENT Strength Elongation Reduction

psi in 8 in. of Area C Mn P S

58,600 31 53 0.21 0.48 0.013 0.037

FATIGUE STRENGTH f1OO,OOO f 2,OOO,OOO k

Si psi psi

49,100a a 34,700 0.10

1953-54 33,300 57,400 33 62 0.17 0.68 0.016 0.039 0.03 ------ 34 ,600ab 0.18

ASTM A7

aSpecimens tested with the Mill Scale on

bSpecimens tested with the Mill Scale off

------ 35,300+

·'·T

TABLE 10

RESULTS OF TRANSVERSE BUTr WELD FATIGUE TESTS

CYCLES FATIGUE STRENGTIf' LOCATION TO f 100,OOO f OF b

SPEC. STRESS FAILURE 2,000,000

FRACTURE

psi 103 psi psi

E60l0 z Reinforcement On

Hl2 29,400 53709 36,600 24,800 edge m4 28,000 539·4 34,900 23,600 edge R22 26,000 802.2 23,100 w.m. E26 35,000 95·3 34,800 w.m. R32 26,300 984.0 24,000 w.m. H48 36,000 53·2 33,200 w.m. H54 25,000 1,695.4 24,500 w.m.

Average 34,900 24,000

E70l6 z Reinforcement On

Hl3 29,100 354.8 34,300 edge Hl7 39,000 117.8 39,800 edge H21 26,000 1,097·0 24,000 edge H25 39,600 98,8 39,600 edge H31 24,200 1,243.8 21,800 edge H65 25,700 3,056.2 25,700+ edge

Average 37,900 23,800+

E60l0 z Reinforcement Off

H28 35,000 112·7 35,600 w.m. IDO 26,000 530.1 32,300 21,900 w.m. H40 23,400 771.0 20,700 w.m. H44 25,400 541.2 31,600 21,400 w.m .. H50 35,000 74.0 33,600 w.m. H60 23,000 2,810·9 23,000 w.m.

Average 33,300 21,800

E70l61. Reinforcement Orf

H27 32,000 179·1 34,600 w.m. H39 26,000 c 26,000+ 2,079·7 H43 35,200 177·1 37,900 w.m. H45 36,100 1,808.9 35,600 plate H57 30,500 533·7 38,000 25,700 ;'f.m. H64 30,500 118.8 31,200 w.m.

Average 35,400 29,lOO+

~ = 0.l3 b w.m. = weld metal; edge edge of weld; plate = plate away from weld =

cSpecimen did not fail

TABLE 11

RESULTS OF STATIC TESTS OF TRANSVERSE BUTT WELDS

PERCENT PERCENT YIELD MAXJMUM ELONGATION REDUCTION

SPECrnEN POINT STRESS IN 6 IN.* OF AREA psi psi

Plain Plate Tests

H8 Mill Scale on 33,000 53,100 42 53 H52 Mill Scale on 31,600 52,900 41 51 H63 Mill Scale off 34,000 54,200 42 52

Tests of E6010 Welds

H2 Reinforcement on 31,700 52,900 35 35 H62 Reinforcement o.ff 33,400 55,400 35 56

Tests of E7016 Welds

H29 Reinforcement off 33,200 54,900 33 47 H47 Reinforcement on 31,800 52,500 30 46

TABLE 12 PROPERTmS OF PLATE MATERIAL USED m PAST AND PRESENT TRANSVERSE BUTT WELD FATIGUE TESTS

PHYSICAL PROPERTIES CHEMICAL CONTENT FATIGUE STRENGTH* Yield Tensile Percent Percent IN PERCENT f 100,000 f 200 ,000 k

STEEL Point strength Elongation Reduction C MIl P S 8i psi psi in 8 in. of area psi psi

Bulletin 3l0c 31,000 51~, 500 34 61 0.15 0.50 0.024 0.022 0.06 30,300+ a 0.10

ASTM A7

Bulletin 327 31~ ,800 61,300 31 54 0.26 0.47 0.015 0.034 0.01 119,80~ 31,600+a

0.18 ASTM A7 59,600 0.18

Bulletin 31~4 ASTM A7 Series X 30,700 59,000 30 54 0.25 0.49 0.009 0.029 0.00 Series Y 29,500 62,300 29 52 0.24 0.47 0.017 0.029 0.03 Series Z 31,000 60,200 30 59 0.19 0.51 0.018 0.029 0.01 All but Above 33,600 63,300 30 51~ 0.26 0.54 0.010 0.034 0.01

Bulletin 350 ASTM A7 38,200 62,700 30 57 0.20 0.42 0.015 0.027 0.23

Bullet in 384 49,100a 34,700

a ASTM A7 36,100 58,900 31 53 0.21 o.~B 0.013 0 .• 037 ---- 0.10

1953-54 d 33,300 57,400 33 62 0.17 0.68 0.016 0.039 0.03 34,6ooa

b 0.18 ASTM A7 35,300+ 0.18

~~-Tension Stress Cycle

~einforcement on cTests reported in Bulletin 302

bReinforcement off dOrdered to American Bureau of Shipping Specification for Type B steel

Z

tW d ..J :!) al

~ a: w CI)

327 K~L

3IJIJ l

31+1+ P

31f1J S5

TABLE 13

RESULTS OF PAST AND PRESENT FATIGUE TESTS OF TRANSVERSE BUTI' WELDS

IN THE AS-WELDED CONDITION: O-TENSION CYCLE

CI) CI) w z ~ ()

:x: t-

STEEL

z o co 0: < o

w CI)

w z < (!J lZ

~

z o o .,J

en

z o .... < a:

~ : w

o 0:: -,0..

W 0: ::> o w () o 0:: a..

PREPARATION OF SPECIMEN

w o o 0:: to W

~

z o tCI)

o 0...

tZ

~ U 0:: o lL. Z

W a:

AVERAGE FATIGUE. STRENGTH

LOCATION ~ .. ·FRACTURE

10 W W..J ... I- en a:WCl) C2:)" 4; Z W a..:-I- CD O..J 00::

I Z «::) e UJ 0... ::> "'I-~ 0 ... " ... «C(J)f;: (J) 0- c- ~.., ~

o ~ 4; ~ ~ o""~ i ~ 0 0 W W ~ a1 w w a:o 0 ... ......Juu<..J 0.0::0:: 1->- - N jJ.IwoO ... ..JO .~ 01- en 0 _lL. ___ lL. ~ =-:11. W UJ .0- .cz

(A) SPEO~ENS PREPARED WITH THE E601m ELECTRODE

7/snc 09 26 O~1f7 00 01' SING a U MANUAL E6010 FLAT ON

7/8"C 00 19 00 51 00 01 DBL. V MANUAL E601Q FLAT ON

7/8nC 0.26 O~51+ 00 01 DBlo V MANUAL E6010

7/anc 00 26 0 0 51+ 00 01 DBl. V MANUAL E6010

FLAT ON

Ft.AT ON

NONE

NONE (ICOf-to1ERO I AL-SHOP)

NONE (CO,..,EROIAL-SHOP)

NONE (COt-f.1EROIAl-SHOP)

3a o ' 22 .. 5 0.13

30 .. 8 21Q2 0.13 5

al+ o l 21 .. 3 00 19

30,7 0.13 . 3

a 3

5

--- 195a~51+ jw C 00 17 0.68 oola OSlo V MANUAL E6010 FLAT ON NONE

O-T

O-T

O-T

O-T

O-T

O-T

o-T I-T

O-T

O-T

a~.9 2~.,O 0.13 5

32 .. 9 17.0 0 .. 19

18.5 e.19

17.2 0.13

27.a 19.9 0.13

2

3 5

2

31+1+ A 7/8"C 00 26 0.,5" oocn DBlo V MANUAL E6010 FLAT ON . OVERHDo

34IJ MH 7/a"c 00 26 00 54 00 01 OSlo V MANUAL E6010 HORIZ o ON

344 NH 7/8nC 00 26 00 54 00 .1 DOLo V MANUAL E6010 HORIZ. ON

341+ 05,06 7/8 11C 00 26 00 51+ 00 01 OBl. V MANUAL E601a VERTo ON

VERT. ON 344 MV

31+~ NV

7/a nc 0.26 0 .. 51+ 0.01 08l. V MANUAL E6010

7/8nC 00 26 00 51+ 8,11 OBl o V MANUAL E601~ VERT., ON

NONE (COMMEROIAL-SHOP)

,NONE (COMMEROIAL-FIElD TEMP 00-1+,,0 F.)

NONE (COMMERCIAL-FIELD)

NONE (COMMEROIAl-SHOP)

NONE (COr.1ERO I A L-F I ELD TEMP. c 00 - ~OOF)

NONE (COMMEROIAl-FIELD)

18.6 8.13

O-T 23.8 0.13 3

2

2 1 . 1

a

TABLE 13 (CON'T)

RESULTS OF PAST AND PRESENT FATIGUE TESTS OF TRANSVERSE BUTT WELDS

IN THE AS-WELDED CONDITION: O-TENSION CYCLE

lOCATION STEEL--- I"REmATToffor- SPEC IMEN AVERAGE [DUTTONOf' OF DATA FATIGUE STRENGTH FRACTURE

cAtf.1lcAL COMPOSITION IN PERCENT W !i '0 W t- (I)

Z W llJ...I C (!J < z W 0 ~ o:W(I) 0 a ...I o 0:

(I) W ;: W W W o..~E 0 0 LLJ 0.. - :::>

2 (I) (I) 0: 0 Z () a .. " ...I

W W Z < ::> 0 0 0: ...I .... (/) o co 0- .a ..:J W -l- (I) Z Z Z 0 0: a 0: 0 <(1):1: (l)W .. CoO 0(1) W (X: -< W W ~ 0 ..: 0 l-<O( W .... .... !J. -01- W..J o~ o ~ a a W W .... I.L.

..J 0:: 0 al " :; ZQ. () 0 C;; z o a.. < 0:0 0 0- ..I;..J""<...I ..J i 0: Z -w 0 ~ w w W tili ~ , N WWOO..l...lO

::> w < <C 00: 0: 0 0..0:0: U. !J. 111: :w =- LLJ. LLJ 0.. coC Z

co (I) t-> (.) :L en -, 0- 0- LLJ 0.. a:: en 0"-

(B) SPEOIMEN§ PREPARED WITH THE E6012 ELEOTRODE

344 X 7/8"C 0,,25 00 49 00 00 OBLo V MANUAL E6012 FlAT ON NONE (COMMEROiAlQSHOP) O .. T 3407 2101 0.13 5 2

34'4 XX 7/8"C 00 26 00 54 00 01 OBLo V MANUAL E6012 FLAT ON NONE (COMMEROIAL-SHOP) 0-1 33.8 23 0 l 00 13 2 2 2

S4~ B 7/a flC 00 26 0,,5'4 00 01 \

OeL Q V MANUAL E6012 Fur ON NONE (COMMEROIAl-SHOP) O-T 32 0 6 20.0 0.13 2 '*

2

344 E 7/8"C 0,,26 00 54 ooOt OBLo V MANUAL E6012 VERi. ON NONE (COMMERCIAL-SHOP) O .. T 29.8 18.1 0.13 2 4 1 1

(0) SPECIMENS PREPARED WBTH THE ,E6013 ELEOTROEE

344 C 7/8 ff C 00 26 00 5'4 0,,01 OSLo V MANUAL E6013 FLAT ON NONE (COMMEROIAL-SHOP) O-T 35 0 2 20.7 0.13 1 .. 3 1 OvERHD.

3 .. 4 F 7/S"C 00 26 0,,54 00 01 OSLo V MANUAL E6013 VERT. ON NONE (COMHE~OIAL~SHOP) O.,.T 3L6 21.3 00 13

(D) SPEOIMENS PREPARED WITH THE E6020 ELEOTRODE

34'4 T6 7/SnC 00 26 00 54 00 01 DBLo V MANUAL E6020 FLAT ON NONE (COr-t1ERCIAL"SHOP) O~T 32.0 ___ m

0.18 1

(E) SPEOIMENS PREPARED WITH THE E6030 ELECTRODE

a .... y 7/8 11 C 00 2" 00 "7 00 03 SING o U MANUAL E6030 FUT ON NONE (COMMEROIAL-SHOP) O-T 31.6 21.5 0.13 7

34'4 R 7/811 C 0,,26 0.5" 00 01 OBLo V MANUAL E6030 FLAT ON NONE (COMMEROIAL-SHOP) O-T 34.9 22.2 0.13 3 3

344 U4,U6 7/811 C 00 26 0,,54 00 01 OSL" V MANUAL £6030 FLAT ON NONE (COMMEROIAL-SHOP) O-T 2702 19.8 0.13 2

344 G5 7/S"C 0,,26 0.5" 0.01 SING" II MANUAL E6030 FLAT- ON NONE (COMMEROIAL-SHOP) O-T 28.7 19.4 0.13 2 E6010 OVERHEAD

TABLE 13 (CON'T)

RESULTS OF PAST AND PRESENT FATIGUE TESTS OF TRANSVERSE BtYIT WELDS

rn THE AS-WELDED CONDITION: O-TENSION CYCLE

roc-ATfON---~~'-U_-~~--STEE[- ---- _u_ PREPARATlONOF--SP£C1MEN AVERAGE LOCAT ION OF OF DATA FATIGUE STRENGTH FRACTURE CHEMIcAL -_ ..

z tW .J .J

~

CI) W

0:: W

CI)

(I) CI) w z )o! (.)

:z: t-

Z o Ql 0: <

U

---, 1953-54 in c 0.17

310 in C 0.15

944 K 7/8"C 0.2£)

3lf4 l 7/8"C 0.26

310 6 I" C 0.15

310 4 in C 0.15

COMPOSITION ~ I 0

IN PERCENT ~ ~ ~ iii CI) w - w W [il a.:' r CI) ... 0:: 0 Z (,) I ~ W Z C :J 0 0 0:: ..Jt-W z 0 IX: 0 0:: - 0 CCI)::I: C (,) t-< W r t- lI. -at-

~ ::i =: ~ g ~ ~ ~ ~o. ~ .~ _ 00: Q: ..J 0 W 0.0:0:: :I: CI) -,0. 0... Wo...o::: Cl)Ot-

(F) SPECIMENS PREPARED WITH THE E7016 ELECTR2~

00 68 0.03 DSL. V MANUAL E7016 FLAT ON NONE

(c) SPEOIMENS PREPARED BY AUTOMATIO PROOE~

00 50 CloD£) SINO. V AUTOMATIC-- FLAT ON NONE

00 54 00 01 ~ AUTO. CARBON- -- 50 TILT ON NONE (COMMERCIAL SHOP) ARO

O.51t 00 01 =X=AUTO. UNION- FLAT ON NONE (COr+t£RCIAL SHOP) MELT

(H) ELEGTRODE NOT SPEOIFIED

00 50 0.06 SINC. V MANUAL NOT GIVEN . FLAT ON NONE

0.50 0.06 SING. V MANUAL NOT GIVEN FLAT ON NONE (PURPOSELY POOR)

CI) (l)W W.J a:o );it;

0-1

O-T

O-T

o-T

O-T

0-1

0 0 (l) -

.. (I) O),C 0

~

37.9

28.9

32".6

0 0 0 .. 0-0(1) 0).: .. N

lI. ~

23.8+ 0.13

22.8 0.10

17.8 0.13

19.4 0.13

21.8 0.10

20.6 0.10

w W t- CI)

" j z w 0 o a: LU 0. - :;)

CI .J e(! ~ W

W 0: c oowwt- '-L ..J.J8"c.J WW ~.J..Jo =-:.W 0...<2

6

3

2 3

2 1

3

a

2

TABLE 14

AVERAGE RESULTS OF PAST AND PRESENT FATIGUE TESTS OF TRANSVERSE BUTT WELDS

FATIGUE STRENGTH: WELD REINFORCEMENT ON FATIGUE STRENGTH: \iELD REINFORCEMENT OFF ELECTRODE POSITION O-Tension Tension=Compression O-Tension Tension=Compression

OF flOO,ooo . f ° flOO,OOO f f 100 ,OOO f . flOO,OOO f

WELDmG 2,000, 00 2,000,000 2,000,000 2,000,000 psi psi psi psi psi psi psi psi

E6010 Flat 33,200 22,300 20,500 13,500 33,300- 26,400 28,000 ------54,200*

Flat-Overhead 32,900 17,000 ------ ------ ------ ------ ------ ------Horizontal ------ 17,800 ------ 12,200 ---_ ... - ------ ------ ------Vertical 25,800 19 z000 16z900 11,800 ------ ------ ------ ------

Average 31,300 18,000 19,900 12,900 --- ... -- 26,400 28,000 -----_.

E6012 Flat 34,200 22,100 . 22,600 12,700 38,100 26,500 ------ ------Flat-Overhead 32,600 20,000 ------ ------ ------ ------ ------ ------Vertical 29 z8oO 18z100 ------ ------ ------ ------. ------ ------

Average 32,700 20,600 . 22,600 12,700 38,100 26,500 ------ ------E6013 Flat-Overhead 35,200 20,700 ------ ------ ------ ------ ------ ------

Vertical 31,600 21,300 21,500 13,600 ------ ------ ------ ------

Average 33,400 21,000 21,500 13,600 ------ ------ ------ ------E6020 Flat 32,000 ------ 22,000 13,500 ------ ------ -_ ... _-- ------E6030 Flat 31,200 21,200 21,600 12,800 ... ----- ------ ------- ------

Flat-Overhead ------ ------ 20,000 13J800 -_ ... --- ------ - ... ---- ------Average 31,200 21 J 200 20,800 13,300 ------ -- ... --- ---_ .. - ------

E7016 Flat 37,900 23,800 -- ... --- ---- .. - 35,400 29,100 ------ ------Automatic Flat ____ ~ ___ _______30 ,~OO ____ _ g9-.l0Q~_ 21,000 12,800 ------ ------ 23,000 14 900 __ ~_ .. ~ _____________ ~ _____ .. _______ .____'-1._. __

* The range of the results was so large that an average 'Jas not computed.

-'i~

TABLE 14 (Con't)

AVERAGE RESULTS OF PAST AND PRESENT FATIGUE TESTS OF TRANSVERSE BUTr WELDS

FATIGUE STRENGTH: WELD REINFORCEMENT ON FATIGUE STRENGTH: WELD REJNFORCEMENT OFF POSITION 0 T' T' C . ELECTRODE OF - ens~on ens~on= ompress~on

WELDlNG flOO,ooo f 2 ,000,000 flOO,OOO f 2 ,000,000

O-Tension Tension=Compression

flOO,OOO f 2 ,000,000 flOO,ooo f 2 ,000,ooo psi psi psi psi psi psi psi psi

Specimens prepared with pre- and post- weld special treatments

Not Given Flat Each Pass Peened -- ... --- 22,300 ------ ------ ------ ------ ------ ------

E60l0 Flat High Temperature stress Relief 31,900 23,700 21,300 15,100 39,300- 28,000 24,500 16,600

53,300*

E6012 Flat ------ ------ ------ ------ 36,800 31,200 ------ ------High Temperature Stress Relief

Not Giyen Flat High Temperature Stress Relief 33,300 23,200 ------ ------ ------ ------ ------ ------

* The range of the results was so large that an average was not computed.

TABLE 15

RESULTS OF PAST AND PRESENT FATIGUE TESTS OF TRANSVERSE BUTT WELDS WITH THE REINFORCEMENT ON

PREPARED WITH SPECIAL TREATMENTS: O-TENSION CYCLE PROCEDURES

lOCATION STEEL -rREPAR~nmr~OfsprCTMt.ff ~- AVERAGE OF DATA

CHEMICAL FATIGUE STRENGTH

COMPOSITION IN PERCENT ... z 10

~ ffi lLI-I O::lLI(I) 0

CI) l&J ;: lLI lLI O-=r~ a z CJ) (.I) 0:: 0 Z 0 0 0 ;: w w z < ::> 0 0 0:: -It-~ 0

(I) Z Z Z 0 0:: 0 0:: ;: 0 <(.I) CJ) c - 0-lLI lLI )II: 0 < 0 !i.:~ W t- lJ... -0 (l)lLI o"~ ocn ...J

~ 0 co " ,g ,(.) - Z 00.. < W.J 0)11:

-I ;: 0:: Z ..I -lLI lLI CJ) W lLI lLI 0::0 0 .. ::> lLI -< ~ 00:: 0:: ...J 0 0..0::0:: t-)- .: N co (/) t- o (/) .., 0.. 0- W 0- 0:: (/)ot- (/)0 Il.

310 3 i" C O~15 0~50 0.06 SING o V MANUAL NOT FI~AT ON EAOH PASS PEENED O"T 22.9 GIVEN

327 IJ 7/8"e 0,,26 0.47 00 01 SING o U MANUAL t.6U', U FLAT ON 12000F STRESS RELIEF TREATMENT O"T 31.9 23.7

310 2 ,~Jt C 0.15 00 50 Ot~6 SING. V AUTO o CaN. MILO FLAT ON 12000F STRESS RELIEF TREATMENT O-T 33.3 'ARCI' STEEL

310 5 i" C 00 15 OQ50 00 06 SING. V MANUAL NOT FLAT ON 12000F STRESS RELIEF TREATMENT O-T 23.2 GIVEN

LOCATIoN OF FRACTURE

lLI lLI t- (I) CI < Z lLI 0 .J o a: W 0-

- ::l " -I e.O "'lLI~;

o 0 lLI W ... u.. rj-l"u<...J

lLIOO..l.J;i )II: :.:aUJUJO-cC'

0.10 3

0.,13 7

0.10 2

0.10 2 1

.. ~'.......,..-.,.~ , ; ·;tll .m

TABLE 16


WITH THE REINFORCEMENT ON TENSION = COMPRESSION CYCLE LOCAlTON STEEl PREPARATION Of SPECiMEN AVERAGE lOCATION OF OF D~rA FATIGUE STRENGTH FRACTURE

CHEMICAL COMPOSITION IN PERCENT

~ 10 W Z W W.J W t- eo a ffi g:~.~ 0 " 0< Z W

C4 W ;: W W a 0 .... 0 a:: 2 (I) (I) 0: 0 Z 0 12 0 0 l.LI 0..' c; ~ ;: W W Z < :l a 0 a:: .Jt-W 0 co. ....

(I) z z z 0 a:: 0 0:: ;: 0 !~~ (I) 0_ 0_ .,!J .,!J. W -W lIJ )t: a 0< 0 to< UJ ~ LL. (/)W ... (/) 0(4 W 0:: 0< -' ~ 0 III (j

~ .... PJ 0 0 - Z 00- < W.J ~w o .. ~ cow W t- U. ..J - a: Z 0 UJ (I) to UJ W Q:O .J.J(j (!I 0< ..J ::> w :r: < ~ ;n a 0:: 0:: ..I 0 n.a::o:: t->- ..: (\J ~ ~~ ~ci ~ 0 al CI) t- U -;) a. 0- UJ 0.. a:: Cl)Ot- CI) U LL. )I: z

(A) SPECIMENS PREPARED WITH THE E6010 ELECT RODEO AND TESTED iN THE AS-WELDED CONDITION

327 C,D 7/8"C 00 26 OalJ? 0.01 SINGo U MANUAL E6010 FLAT ON NONE TeC 22.3 l".lJ 0.13 6

3ltlt Z 7/B"C 0.19 0.51 00 01 DOLo V MANUAL E6010 FLAT ON NONE (COMMERCIAL-SHOP) T=C 20.2 13.0 6

34lJ 9 7/8"C 0026 O05lJ O()Ol DBLo V MANUAL E6010 FLAT ON NONE (COMMEROIAL-SHOP) T=C lB.B 11.B 0.13 3 2

344 54,56 7/B"C 0.26 0.54 00 01 OSLo U MANUAL E6010 FLAT ON NONE (COMMERCIAL-SHOP) T=C 21.lt 1lJ.O 3 3

950 19,25 ~II C 00 20 O.lJ2 0.23 SING o V MANUAL E6010 FLAT ON 19-25 PICKLED AND PAINTED T=C 19.9 15.8 3 lJ 40-42

34lJ MH 7/8"C 00 26 0.54 O.Dl DSl. V MANUAL E6010 HORIZ o ON NONE (COMMEROIAL FIELD T=C 13.1 3 TEMPo 00 <:' lJOOF) ,

34lJ NH 7/8"C 0026 O.5lJ 0.(;)1 DOLo V MANUAL E6010 HORIZo ON NONE (COMMEROIAL-FIELD) T=C 11.6 2 1

34lJ O~ 7/8"C 00 26 0.5'+ 00 11 OBLo V MANUAl E6010 YERT o ON NONE (COMMEROIAL-SHOP) T=C 16.9 1'04

344 MY 7/8 11 C 0 .. 26 0.54 0.01 OBL. V MANUAL E6010 V~RTo ON NONE (COMMEROIAL-~IELD T=C 13.8 3 TEMP. 10° - lJO F)

34lJ NV 7/8"C 00 26 0.5lJ 0.01 OBL. V MANUAL E6010 VERTo ON NONE (COMMERCIAL FIELD) T=C 10 0 3 2 2

(8) SPEOIMENS PREPARED WiTH THE E6012 ELECT~ODE AND TESTED IN THE AS-WELDED CONDITION

344 X 7/8"C 00 25 00 49 0.00 OBLo V MANUAL E6012 FLAT ON NONE (Cor+1EROIAL-SHOP) T=C 21.8 12.6 0.13 5

34'+ XX 7/8 t1 C 0.26 00 5'+ 0.01 OBL. V MANUAL E6012 FLAT ON NONE (COMMEROIAL-SHOP) T=C 23.5 12.B 0.13 ,. 2

TABLE 16 (CON'T)

HESULTS OF PAST AND PRESENT FATIGUE TESTS OF ~j{ANSVERSE Burr WELDS

wrrH THE REINFORCEMENT ON TENSION = COMPRESSION CYCLE

IOCAilOfr sTEEl. PITmRArnnrlWSPECTM~N -~VERAGE [OCATH>1f~Or--~

OF DATA FATIGUE STRENGTH FRACTURE CHEMICAL COMPOSITION IN PERCENT ... z to W z W W~ W t- c.o 0 ffi O:W(/) 0 (!J oC Z W

CI) W W W o..:-t- G a ~ 0 ~ z CI) II) t- o: 0 z 0 tz 0 co LLI 0.. ::l W W Z oC :;) 0 0 a: ~t-W 0 .. (!J ..J t- (/) Z Z Z 0 a: 0 a: 0 !~~ (/) 0- C!)- ~ ~ W W W ~ 0 oC 0 t-< W t- t- U. (/)w ... c.o 0(1) W 0: oC ~ 0 en (!J zo.. 0 0 z 00... oC W...J Ok .~ 0 o w W t- LL. ...J 0: 0: Z ...J -w 0 W (/) UJ -w 0:0 0 ... ~ .... (!J c.!J oC ~.

:;) W :c <C < 00: a: ~ 0 UJ 0.. a: 0: t-,... a:- N W W a a .... ~I~ co (I) ..... U :L (I) ..,0... 0.. OJ 0- 0: (1)0 ..... (l)U "" ~- =- :a OJ OJ 0...

(C:~ SPECIMENS PREPARED WITH THE E6013 ELEOTRODE AND TESTED IN THE AS-WELDED CONDITION

a~~ F ?/BIIC 00 26 0,,51+ 00 11 DSl. V MANUAL E6013 VERTo ON NONE (COM~EROIAL-SHOP) T=C 2105 13.6 .. 1 1

(D) SPEOIMENS PREPARED WITH THE E6020 ElEOTRODE AND TESTED IN THE AS-WELDED CONDITION

31+~ T~, T5 ?IBlle 00 26 0051~ 0.01 OSLo V MANUAL E6020 FLAT ON NONE (COM,"1ERO I Al-SHOP) T=C 22 0 0 13.5 3 1 1

(E) SPEOIMENS PREPARED WITH THE E6Q30 ELECTRODE AND TESTED IN THE AS-WELDED CONDITION

31J~ R ?IBIIC 0 .. 26 Oo51~ 0 0 01 DBLo V MANUAL E6030 FLAT ON NONE (COMMEROIAL SHOP) T=C 21.6 11.8 0.13 3 3

Sif4 U5 ?/BIIC 0.,26 • 0.51~ OoCl DBlo V MANUAL E6030 FLAY ON NONE (COMMEROIAL SHOP) T=C 13.7

34~ G~,G6 ?/Bnc 00 26 0.51~ 00 01 SING .. U MANUAL E6030 FLAT ON NONE (COMMEROIAL SHOP) T=C 20 0 0 13.8 2 3 E6010 OVERHDo

(F) SPEOIMENS PREPARED BY AUTOMATiC PROOEDIJRES AND TESTED IN THE AS-WELDED CONDITION

3~~ K ?/811C 00 26 O1)51~ 0 .. 01 =:t-r-r= AUTO.CBN -... SO TILT ON NONE (COMMEROIAl-SHOP) T=C 19.? 11.8 00 13 2 5 ARO

NONE (COMMEROIAL-SHOP) 0.13 31J~ L ?/Bnc 00 26 a ,,5'~ 0.01 ==x:== AUTOMAT I 0 -a FUT ON T=C 22 .. 2 19.7 3 8 UNIONMELT

(Q) ElEOTRODE NOT SPEOOFIED; TESTED IN THE AS-WELDED CONDITION'

82? C,E !tt C 00 26 Oo~'7 0 0 .1 SING o V MANUAL NOT GIVE:N

FLAT NONE NONE T=C 20 11 1 15.~ 0.10 4

(H) SPEOIMENS PREPARED WiTH THE E601D ELE[LRODE AND TESTED AFTER A 1200 DEG 2 f HEAT TREATMENT

327 A,B 7/8"C 00 26 0.47 00 01 SING. U MANUAL E6010 FLAT ON 12000F STRESS RELIEF TREATMENT T=C 21 0 315.1 0.13 5

• ·,1'" TABLE 17


WITH THE REmFORCEMENT REMOVED

mCAlioN SlEEt PRE~K[KTTON OF SPEclRtN AVERAGE (oeAT ION tF OF DATA

C!HOOCrc-FATIGUE STRENGTH FRACTURE

CI(X>1POS I T ION. liN PERCENT

~ 10 W Z

~ W..J W t- c.o

~ 0 a:wc.o 0 u ~ z W

(/) ILLI ;: W W 0..:''- D 0 ..J 0 a: z (/) I~ a: 0 at 0 • at G .. LIJ 0.. ' (j ::»

w ILLI 0< g 0 0 0: ...... W 0 .. ..J

t;j c.o z z :z 0 a: 0: ;: 0 !~~ (/) -- e- -.!I aiJ W :; W ~ 0 .: 0 t-.: W t- l&. Cl)W .. (/) .. c.o W 0:: ..J ~

0 a:I .c" :::; zo- o 0 - z o d-~ W..J o~ _lit o 0 W W'" L&.. ..J X 0: Z -w 0 W (/) W W D:U ., ... ~"'UU~..J ::l LLI < :~ 00: 0: ..J 0 LLI l};O:QI ... ,.. .... N o 0 ..J 0 Ol ' en t- o en -,d- d- LIJ d- o:: ot- enu l&. .. :..:: ~'LIJ LIJ t. -< .. z

(A) SPEOIMENS TESTED IN A O-TENSION CYOLE SPEOIMENS WITHOUT PRE- OR POST-VELD TREATMENTS 327 o,P ?/Bne 00 26 O;ti1 lfoOl SING o 0 RANUAL E6010 FLAT OFF NONE O"T ll8,,8 28." 0.18 4 2

327 Z 7/BIIC 0.26 Oolf7 O(,Gl SING o U MANUAl E6010 FLAT OFF NONE (GROUND OFF PERP~NDIOULAR TO t)oT 41+.5 26.3 0;18 3 lJ DIREOTION OF LOADING)

327 X 7/B"C.O,,2G 00lJ7 00 01 SING a U MANUAL E6010 FLAT OFF NONE (MAOHINED OFF AND POLISHED) O~l 5'+02 0.18 I

384 FI ~" C 00 21 00 '+8 SINOo V MANUAL E6010 FLAT OFF NONE (MAOH.NED AND FILED) O':"T 35.6 28.9 1.07 6

0="- 1953 ... 54 J" C 00 17 0 0 68 0.03 DBLo U MANUAL E6010 FLAT OFF NONE (GROUND PARALLEL TO DIREOTION O-T 33.3 21.8 0.13 6 OF'LOADING)

381+ HI ~" C 00 21 D.ll8 SINGo V MANUAl E6Q12 FLAT OFF NONE (MAOHINED OFF AND FILED) D .. T 38.1 26.5 0.12 6

~- ... 1953 ... 5lJ I" C 11017 10 68 80 03 DOL" U MANUAL E7016 FUT OFF NONE (GROUND PARALLEL TO DIREOTION D-T 35.4 29 0 1+ 0.13 4

in C OF LOADING)

310 7 00 15 0 0 50 00 16 SING o V MANUAL NOT FLAT OFF NONE (MACHINED OFF) O-T 27.9 0.18 3 GIVEN

(0) SPEOIMENS TESTED IN A TENSION = COMPRESSION CYOLE SPEC I HE~S WITHOUT SPEO I AL PRE- OR POST-WELD TREA TM'EN'TS

327 G~H 7/8f1 C 1026 eo .. 7 0,,11 SING .. U MANUAL E6118 FLAT OFF NONE (MAOHINED OFF) T=C 29 0 2 Do18 7

327 Y 7/8 11 C 00 26 00 '+7 00 11 SING. U MANUAL E6810 FLAT OFF NONE (GROUND OFF) T=C 26.8 .---- 1.18 2

344 K 7/8"C 0.26 0'054 00 .1 AUTO. CON. _ ... "" .... FLAT OFF NONE (GROUND OFF) T=C 23.0 llJ.9 2 ARC

SPEOIMENS WITH SPECIAL PA:E- OR POST-WELD TREATMENTS

327 E?F 7/8"C 0026 al o .. 7 0.01 SING .. U MANUAL E6010 FLAT OFF 12000F STJtESS RELI EF TREATMENT T=C 24.5 16.t; 1 0 18 1 3 5

TABLE 17 (CON'T)


WITH THE REINFORCEMENT REMOVED

-- . . AVERAGE lOCAl ION OF --..~~";;";"'--~I!Tt'1r:r-------~--~--~~~EPXR~nOrr-oF S~CIt"EN FAT IGUE STRENGTH . . FRACTURE .. LocATIoN STEEL \ OF DATA CHEMICil

z

ti ..J ..J :;)

CO

(I) LU

0: W (I)

(I) (I) LU Z ~ o :z: I-

z o m IX c: o

COMPOSITION IN PERCENT

&AI (A) LU Z c: U 2

~

z o ()

.... (I)

z o t..: 0::

t-..: ~fu 00:: -,0...

LU 0: :;) Q LU () o 0:

0...

SPEOIMENS WITH Af08T-WELD HEAT TREATMENT

327 M,N 7/8nC 00 26 0.\7 G.ll SING o U MANUAL

327 U 7/8 nG 1.26 Oo~7 0.01 SING. U MANUAL

3B~ FII I i" C 0.21 Oo~8

384 HII ~"C 0.21 0.~8

SING o V MANUAL

SING. V MANUAL

w o a 0: to LU ..J W

z o t(I)

o a..

tZ ill

ffi () IX o U. Z

W 0:::

E601. FLAT OFF

E6010 FLAT OFF

E6D1. FLAT OFF

E6012 FLAT OFF

, Q

~rn o...:.t-, z .ut-LU C:(I):J: -Oto o...\c: LU LU 0..0:0: (I) 0 ....

1200° STRESS RELIEF TREATMENT

12000 STRESS RELIEF TREATMENT

115&° STRESS RELIEF TREATMENT

11500 STRESS RELIEF TREATMENT

(I) ('oW W .... 0::0 t->(I) 0

D-T

O-T

O .. T

Ill-T

c .. o ID G .. -o e CI)

.. (I) • ~ «» ~ ca «) .. _ N "

U. u. ),t

.. 90 .. 27.8 B.18

53~3 0.18

W u a w e.G

a a

~'-~

39 0 3 29.1 l.tO ~

96.8 31.2 0.07 ~

w ... C ....

0...

e.G w W W tU U. C Q Q .... W W.o...

9

2

3

(A) Z LU o 0:

:;) C .... W 0: c

LL .... .... 0 ~Z

jl i!'

TABLE 18 PROPERTIES OF SILICON STEEL .AND LOW ALLOY HIGH TENSILE STEEL USED IN PAST FATIGUE TESTS

OF TRANSVERSE BUTI' WELDS

STEEL PHYSICAL PROPERTIES Tensile Percent Percent CHEMICAL CONTENT IN PERCENT Yield

Point psi

Strength Elongation Reduction--------------------------------------------------

Bullet in 310 Silicon Steel 49,800

Bulletin 327 Lmf Alloy Steel 59,100

Bulletin 350 Lmf Alloy Steel

Series 13-18 60,400 Series 46-48 55,300

*in 2 in.

psi in 8 ino of Area C Mn p S Si Cu

80,700 25 57 0·30 0.91 0.022 0.027 0.23

78,900 21 62 0.16 1.27 0.022 0.036 0.20 0.03

79,700 30* 68 0.18 0.96 0~019 0.026 0.15 0.20 76,800 23-

STEEL

Bulletin 310

61

Stress Cycle

0.15

Silicon Steel a-Tension

Bulletin 327 Low Alloy Steel Tension =

1.06 0.014 0.035 ---- 0.05

FATIGUE STRENGTH

flOOO,OOO f 2,000,OOO k psi psi

35,800a 0.10

35,300a 26,400a 0.10 Compression

~ill Scale On

Ni Cr Va

-- 0.02 0.12

0.04 --- 0.10 0·-37 --- 0.13

TABLE 19

RESULTS OF PAST FATIGUE TESTS OF TRANSVERSE Burr WELDS IN SILICON STEEL

AND IN LOW ALLOY HIGH TENSILE STEELS

LOCATI Off-~·- ~----SnEL pREPARATTON'OF~ SPtffifEN -~ ------·~VERAGE lOCATION Of OF DATA

(I)

Z (J) W

ti (I) z W ~

..J ~ 0

..J :l W x' en (I) t-'

311 e in 311 9 in

327 "'-'09 !" 350 13-18 ~" 350 46 .. 48 ~n

350 70 ... 75 ~"

'" (/) W z Z

0 -c: m Q 0= z <C <C u X

CHEMlCAL COMPOSITION IN PERCENT

Z 10 '0 ... el)

~ I Q Z W W.J 0 ffi ~wen

~ w 0..~!K I- 0 Z 0

~ :::» 0 0 ~ .J ... W a 0:: 0 ~~r:

~o: '" ... I- U. 0 0 Z uo..-c:

-w 0 W (I) '" W 00:: 0= ... 0 '" Q. 0:: 0:: """) 0.. 0.. L&J 0.. 0:: (I) 0 t-

(A) WELDS IN SILiCON STEEL

DOLo V MANUAL C-+1o ON NONE

DOL. V MANUAL C-Mo OFF NONE

(a) VELDS IN low ALLOY HIGH TENSlbE S~

S I NG o V MANUAL GRD. EA ON NONE

SI NGoV MANUAL E6010 FUT ON PLATES PICKLED AND PAINTED GROUND FLUSH ON ROOT-SIDE

SING. V MANUAL E6011 FLAT ON PLATES PIOKLED AND PAINTED

(0) WELDS JOINING CARBON STEEL TO Lo~ALLOY HIGH TENSIL~ STEEL

SING .. V MANUAL E6Dli FLAT ON NONE (CARBON STEEL TO LOW ALLOY 5;TEEL)

FATIGUE, STRENGTH - - "FRACTURE

e D • • .. ff:)"'-(I) o-

(l)W .. (I) .(1) W..J et~ .~ 0::0 .. 1->- - N (I) (.) u. U.

O-T 22.8

I) .. T 23.7

T=C 25.7 15.a.

r-C 26 0 4 18.1

T=C 23.'

T=C 25.8 la.7

¥

0.18

1.11

8.17

1.11J

0.14

1.17

iii

'" I- en t!J c Z W 0 .J 0 0::

L&J 0.. c; :3 eD ..:s W

"'0:: C o 0 W"'I- LL. iiJrn (St!J-c:-I

Q&}.J..J 0 =-=- UJ 0.. < ::z:

2

3

7

3

3 2

"!:;,-!~

3@ 4 1 - O· C-C II: 121- 0" --1

H 12

H 7

-I H I

H H ~ ~ 13 8

~

'- ... _/ H '0 H HT8 ~ '3 .1 14 H o 9 h

U ... . H 6 ,"/ H ~ 4 lEo ~ 15 HT9

~ 16

~-----HTIO

HI rl ~ ~ 17

FIG. I

.... DIRECTION OF ROLLING __

LOCATION OF SPECIMENS IN PARENT PLATE STEEL HJ PLATE 10, 3/4 IN. THICK

r

·0\ J co

_II "b I

(8) co.

3 crt 4'- d': 12'- o·

H H H 32 3~ 33A

H H 356 336

H ~ 31 H 51

fo I rJ C3> --~

~I r31 ~ - - - - - - - HI I ~ 281 HHI HH2 HH3 HH4 HH51

H H H

~71 38A 37A 38A

~ H H H

388 378 368

.. DIRECTION OF ROLLING ....

riG. 2 LOCATION OF SPECIMENS IN PARENT PLATE STEEL HI PLATE 9 ~ 3/4 IN. THICK

I H 21

H 22

H 23

H 24

H 25

H 26

I . ----"--

rl ~7

I I

: I 0 0 I

-to

:11 <==3 0 0 - I -

(2)

to

FIG.3

~ @ 41- OOI_~J 21-QQ~ ______ . ___ . ___

H451 H 6SAI H67',t\ H66A H131

-H68 B H 6713 H 668

-H -

H G9AI ~ 46 H 71 A H 70Jt\

-

~71 H 71 BI H 7013 H 698

I ~ -H7213 H72 A H 73

H H 77A ~I H74AI ~ 48 I

H 77 H748

H H7BA CD H75AI ~ 49 u...

I

H7B8 H758

fo1 H79A (J) H76 AI ~ u... :r:

H 79 H 768

DIRECTION OF ROLLING .. lOCATION OF SPECIMENS IN PARENT PLATE

STEEL H, PLA,TE 8, 3/4 IN. THICK

I H 39

H 40

H 41

H 42

H 43

H 44

3 e 41-001~ 121-00"

~I "

----.

~of H 82 A H 81 A H 80A ~ H 54

H828 H 818 H 80B -H ~ H 61 H85 A H 84A H83A 55

~ H85 B H848 H83B

gl ~ rl I ~ H 62 56

- I H87 H868 H86A w

gl ~/ H631 H9IAI~ H 88AI ~ H .I 57

- 1

H 91 B H 888 WI , " '~~/

H I H92A = H89A ~ H lJ... 64 I 58,·

H928 H898

H651 H93A C\J H90AI ~ H

lJ.. 59 I

H938 I H90B

DIRECTION OF ROLLING =C •

. FIG. 4 LOCATION OF SPECIMENS IN PARENT PLATE

STEEL H, PLATE 7, 3/4 IN. THICK

• H 120A1 H 110 A I HT 591 H 106AI H 100 ITIT I H94

HIIOB

H III A I H 120BI H 115 H 106 B

H 101 IT2TI H 95

~ H 121AI Hili B I HI07A

HI16

H 112 A I HI021T3TI H96

HI21B I H 107B

-0 HT621 HT611 HT60 0

I HI22A H 117 HI12 B H 108A HI03 TT H97 I <0 4 co

_II -'I -N ~ H 113 A @ HI22B H 108B ® c:o H 118 HI04 TT H98

<0

HII3 B 5

HI23A HI09A

HI14 A H 119 HI05 ITeTI H99

HI23BI HII4 B H 109B

~ C==2" ~I .. 2'-4" .. I~ 1'_4" 2'-2

11

:1 12'-0"

~ DIRECTION OF ROLLING ..

FIG.5 LOCATION OF SPECIMENS IN PARENT PLATE STEEL H, PLATE 6, 3/4 IN. THICK

911 R

H 0

-

]: 11- 8 11

~ II - 9-

4 1- a ..

a. LONGITUDINAL BUTT WELD

b. TRANSVERSE BUTT WELD

FIG.6 DETAILS OF BUTT WELD SPECIMENS.

O. JIG IN POSITION FOR FIRST PASS

t ," ~t:,.~~'· .,'

"

~~i~~:t;lii~i:~C;~~h~·

b. JIG IN POSITION FOR SECOND PASS

FIG.7 WELDING JIG FOR LONGITUDINAL BUTT WELD SPECIMENS

FIG.8 DETAILS OF WELDING SEQUENCE FOR LONGll·UDINAL BUTT WELD S'PECIMENS

'r

II

FLAME-CUT ~ OUTSIDE FINAL DIMENSIONS.

4 1 - 0"

SAW-CUT AND MACHINE FOR DOUBLE V BUTT WELD.

~I

o. PREPARATION OF SPECIMEN FOR WELDING

I , I

4 11 3

11

\

5

3 2" .5"

5" 2"

2 211

4 2"

6 3 11 4"

I 7"·~ ... TRANSVERSE WELD

NOTE:

X iNDiCATES CHANGE OF" ELECTRODE.

~IRROWS iNDiCATE DiRECTiON OF WELD ING.

b. WELDING SEQUENCE

FIG.9 DETAILS OF PREPARATION OF TRANSVERSE BUTT WELD SPECIMENS

, ______________________________________________________________ ~.r~~

o II 0 I LEVER '--

UPPER PULL HEAD DYNAMOMETER

TEST SPECIMEN TURNBUCKLE

VARIABLE ECCENTRIC MOTOR

LOWER PULL HEAD •• II I ... . . .

H • •

o II·· • • • •

FIG.IO SCHEMATIC DIAGRAM OF WILSON FATIGUE TESTING MACHINE

60

50 (I)

~

Z 40

f/J ~ 30 0::

---

t-(/)

2~0

60

;;; 50 x:: z 40

~ 30 w 0:: ... (/)

20 30

--

----,- ----- - --

50

- I--~

~ -

50

..., • oe

100 500 1000

CYCLES TO FAILURE I N THOUSANDS

O. PLAIN PLATE TESTS

I() E6010 REINFORCEMENT ON

K=0.13 ,l~ E7016 REINFORCEMENT ON -l. ~ - .6. ~ lI.

'-"'- ~ &.\ -----=~ ~ , 0 r--r--. r--~ ------l-.

~ I'JUl"" u --100 500 1000

CYCLES TO FAIILURE IN THOUSANDS

b. LONGITUDINAL BUTT WELD TESTS

JO MILL SCALE ON • MILL SCALE OfF

00 -- I

I

2000 . 5000 8000

• E6010 R[INFORCEMENT O~I A E70 18 RE INFORCEMENT OFF

I

I

-. -... .. ... .~ ------ - .. - 1-

-

2000 5000 8000

FIG. II RESULTS OF FATIGUE TESTS 'OF LONGITUDINAL BUTT WELDS

l •

HIO H 49

C. SPECIMENS TESTED WITH MILL SCALE ON

H51 H 56 .

b. SPECIMENS TESTED WITH MILL SCALE REMOVED BY GRINDING

FIG.I2 TYPICAL FRACTURE SURFACES OF PLAIN PLATE FATIGUE SPECIMENS

H 19 H 97

a. E6010 WELDS

H 74 H 88

b. E7016 WELDS

FIG.13 TYPICAL FA'TIGUE FRACTURES OF LONGITUDINAL BUTT WELDS WITH THE REINFORCEMENT ON ,

H 99

a. E 6010 WELDS

H 78 H 96

b. E 7016 WELDS

r1G.14 TYPICAL rATfGUE FRACTURES OF LONGITUDINAL· BUTT WELDS WITH THE REINFORCEMENT REMOVED BY GRINDING

CI)

~

60

~o

z 40

~ 30 w a: ... CI)

CI) ~

Z

20 30

60

50

40

~ 30 w 0: le/)

20 30

o MILL SCALE ON, 1953-54 • M ILL SCALE OFF, 19!>3-54

I~ 0

[] M ILL SCALE ON, 384 r-v

-- ~ O. 00 []

~ ~

[j. .... ~

\. ---------------------- --- --- -- , -.I..

50 100 500 1000 2000 5000

CYCLES TO rAILURE IN THOUSANDS

c. PLAIN PLATE TESTS

REINFORCE~ENT ON REINrORCE~ENT OFF o E6010 1953-~4 • E6010 1953-54 A E 7018 1953-54 A E 7016 1953-54 o E6010 384 G I • E6010 384 G IV V E6010 384 H III LOW TEMP. STRESS RELIEF • E 6010 384 H IV LOW TEMP. STRESS RELIEF <J E6010 384 Gil HIGH TEMP. STRESS RELIEF ~ E6010 384 GV HIGH TEMP. STRESS RELIEF I> E6010 364 G III MECHANICAL STRESS RELIEF .. E6010 384 GVI MECHANICAL STRESS RELIEF

• • .. -

50

~ t~ , ~ . ... an .- - ....... V~ •• ... .. r- .... £;<1 <l]Q

I" - - -~"

[]O I> <l •• .... . , ~ .tJ. In.n,..1 .. P'J. .... I> -~ A: ~ .. - ...

I""""LI .. []~ '" ... .. -- ---V ~ ~<l8

II--

100 500 1000 2000

C YC LES TO FAILURE IN THOUSANDS

b. LONGITUDINAL BUTT WELD TESTS

FIG.15 RESULTS OF PAST AND PRESENT FATIGUE TESTS OF LONGITUDINAL BlJTT WELDS

.&.. -

5000

---

8000

8000

60 o MILL SCALE ON

CI) 50 • MILL SCALE Off

x: z 40 1'\ - ..

O· 00 • ~ w 30 0= t-CI)

20 50 100 500 1000 2000 5000 8000

CYCLES TO fAILURE IN THOUSANDS

O. PLAIN PLATE TESTS

60 o E6010 REINfORCEMENT ON I! E 7016 REINfORCEMENT ON

50 • E6010 REINfORCEMENT Off • E 7016 REINFORCEMENT Off CI) ~

z 40

E, fAILURE AT EDGE Of WELD WM .. FAILURE IN WELD METAL

• CI) CI) w 30 0:: l-CI)

20 ---- -- -- --

50 500 lodo 8000. 100 2000 5000


b. BUTT WELD TESTS

FIG. 16 RESULTS OF FATIGUE TESTS OF' TRANSVERSE BUTT WELDS

HI4 H 26

a. E 6010 WELDS

H 17 H 21

b. E 7016 WELDS

FIG.17 TYPICAL FATIGUE FRACTURES OF TRANSVERSE BUTT WELDS WITH THE REINFORCEMENT ON

H 30 H 50

O. E6010 WELDS

H 27 H 57

b. E 7016 WELDS

FIG.18 TYPICAL· FATIGUE FRACTURES OF TRANSVERSE BUTT WELDS WITH THE REINFORCEMENT GROUND OFF

H26 HI3

a. SPECIMENS WITH THE WELD REINFORCEMENT ON

H 40 H31

b. SPECIMENS WITH THE WELD REINFORCEMENT GROUND OFF

f1G.19 TYP:ICAL MACROGRAPHS OF PLANE NORMAL TO FATIGUE FRACTURES

CI) ~

Z

Ul en w a:: .... CI)

60

50

40

30

20

15 40

0

• 1953-54 MILL SCALE ON 1953-54 MILL SCALE Off

I~ 6 M

100

II BULLETIN 327 MILl. SCALE: ON V BULLETIN 310 MILL SCALE ON

• BULLETIN 327 MILl. SCALE Off D BULLETIN 384 MILL SCALE ON

I

• • DIn.

~ O· 6 00 D

~ ..

V V ~o-r-)

-y --

500 1000 2000 5000 CYCLES TO FAILURE IN THOUSANDS

FIG.20 . RESULTS OF PAST AND P'RESENIT FATIGUE TESTS OF STEEL PLATE USED IN TRANSVERSE BUTT WELD INVESTIGATIONS

I

I

I

I

10000

60

50

U; 40 ~

z - 30 V)

~ a: t-el) 20

<J V 0

~ 40'

60

50

V) 40 ~

Z - 30 (/) (/) w a:

t; 20

1953-54 o 344-P ~ 344-05,06 • 344 -MH W = fAILURE IN WELD METAL 3.~7 -K)L A 344-55 .344-MV • 344 -NH E= fAILURE A.T EDGE OF WELD 3"~4-Z • 344-A .344-NV P= fAILURE I N PLATE

~, ~ ~P ~ E1' l~ .F ..J"V'\..J'\ ~~ WWW EP Ve: r-o

E - ~E pU} ~P ~ .... OO .... A • ~ ~ A 6.vE 0 EWW EW EW P E EW Er~~_ P

... ... ... _ ...

... -w WW EWE EWE EEEE"EP W'C.F

I __ 1_ _J~ _ __________ _ 100 500 1000 2000


a. SPECIMENS PREPARED WITH THE E 6010 ELECTRODE

J 0 344-XX .344-E W- FAILURE: IN WELD METAL o 344-X " 344- B E = fAILURE AT EDGE Of WELD

P = fAILURE IN PLATE

... .A ~~ WE J --wOE E www 'E~ vi ~\ • ~¥ 6 .~iv 0

WEWEW EWWW 1& a .A - -E- IF P E·P E:

----~ 15 40

'-----

100 500 1000 2000 CYCLES TO fAILURE IN THOUSANDS

b. SPECIMENS PREPARED WITH THE E6012 ELECTRODE

OPEN SYMBOLS SHOW SPECIMENS WELDED IN fLAT POSITION

5000 10000

OPEN SYMBOLS SHOW SPECIMENS WELDED IN fLAT POSITlON

-(P

-5000 10000

FIG .. 21 RES.ULTS OF PAST AND PRESENT FATIGUE TESTS OF TRANSVERSE BUTT WELDS IN THE AS-WELDED CONDITION PREPARED WITH

E6010 AND E6012 ELECTRODES.; 0- TENSION CYCLE

Ul

60

50

~ 40

Z

~ 30 w a: tf/)

~

20

1540

60

50

~ 40

z

~ 30 w a:: tCI)

20

I~

•

• 344-F Will FAILURE IN WELD METAL E= FAILURE AT EDGE OF WELD

• 344-C P= FAILURE I N PLATE

- - -P E-P E-P E I:'W

~~ - ." EW E

- - -E EP E

-

100 ~oo 1000 2000 5000 10000 CYCLES TO fAILURE I N THOUSANDS

o. SPECIMENS PREPARED WITH THE E6013 ELECTRODE

o 344-Y W=FAILURE IN WELD METAL OPEN SYMBOLS SHOW D 344-R E= FAILURE AT EDGE OF WELD SPECIMENS WELDED II 344-U4. U6 P= FAILURE 1 N PLAT E IN FLAT POSITION

U--+ ~ w ~ I

.... DN

r» All II ~~ , Dm WEE EP E E

Vf. ~ ..... - - .

t _________ I

100 500 1000 2000 5000 10000 CYCLES TO FAILURE I N THOUSANDS

b. SPECIMENS' PREPARED WITH THE E6030 ELECTRODE

FIG.22 RESULTS OF PAST FATIGUE TESTS OF TRA.NSVERSE BUTT WELDS IN THE AS-WELDED CONDITION PREPARED WITH E6013 AND EEi030 ELECTRODES; O-TENSION CYCLE

60

50

f/) ~ 40

z tf) 30 f/) w a: t-f/)

20

1540 100

o 310-1 PROCESS NOT SPECIFIED W= FAILU'RE IN WELD METAL Il 344-K (CARBON -ARC) E = FAILURE AT EDGE Of WELD 'V 344-L (UN ION MEL T) P=fAILURE IN PLATE

I

E

~~ E~E ~ EW, ~ VEW

A A An ... ........... ~ E-P EEP "E "

I 500 1000 2000 5000 10000

CYCLES TO fA I LURE I N THOUSANDS

FIG.23 RESUILTS OF PAST FATIGUE TESTS OF TRANSVERSE BUTT WELDS IN THE ASWELDED CONDITION PREPARED BY AUTOMATIC PROCEDURES; Q-TENSION CYCLE

60

50

4 (/)

~

Z 3

~ a::: .- 2 V)

0

0

0

~ '-40

• •• I .. 0 0

- ..." A •

~ n. JfV)(

" ....,

100

o E6010 o E6020 • E6012 ~ E6030 V E6013 • E7016

I • AUTOMATIC

.~ n.A JI rII:R. ~ ~--I\h ... 6.~ --(JV

• A r:e ~rL. on.", ~ ~ ~O o • O~J ~ ' ....... ~ ..... <: ~-

"" ,...,...,. - ... • .A __ n~ n .I..A A .... ....,...., - - - ., ~" - ~

I

I

500 1000 2000 5000 10000

CYCLES TO fAILUR:E IN THOUSANDS

FIG.24 SUMMARY ()F RESULTS OF AL.L PAST AND PRESENT FATIGUE TESTS OF TRANSVEF~SIE BUTT WELDS IN THE ~\S-WELDED CONDITION: O-TENSION CYCLE

z

V) CIl W 0= ICIl

60

~O

20

I~O

I E= fAILURE AT EDGE or WELD P=FAILURE IN PLATE

E ~ E. A " ""E

-y

E

100

OPE N SYMBOLS SHOW o 327-IJ E6010 J 1200·F. STRESS RELIEF SPECIMENS WELDED o 310-5 HIGH TEMP. 5 TRESS RELIEF I N flAT POSITION A 310-2 HIGH TEMP. STRESS RELIEf

V 310-3 EACH BEAD PEENED

H E EP-

lr E ~b E E ........

0 E ~7 0 0 E 0

E ... E:

L...-... ___ ~ L--.-L.........-___ --

500 1000 2000 5000 10000 CYCLES TO fAILURE IN THOUSANDS

FIG.25 RESULTS OF PAST FATIGUE TESTS OF' TRANSVERSE BUTT WELDS WITH THE REINFORCEMENT ON, WELDED WITH SPECIAL 'PROCEDURES: O-TENSION CYCLE

50

40

;;; 30

~

Z

CI) 20 CI) W 0::

~15

1025

50

40

(() 30 ~

z (() 20 (/} w Q:

~15

1~5

t

FIG.26

liO 327-C,D I A 344 - P . I <l 350-19-25, 40-42 1 1 344-MV I' 344-MH W=FAILURE IN WELD METAL o 344- Z V 344-54,56 • 344-04 A 344-NV ~ 344-NH E=FAILURE AT EDGE Of METAL

p. FAILURE IN PLATE

tt.E fl! ~ S;P ~~ 3; I ,

~u ~ ... y ... ~ OJ ~ ~-...

I Ef ~jr.v :I¥ I: 'WA t Ff.. l Wt 14'~1'p .t~P Jj~ b • ~ Y3rf EW ..... <l <l

.A

!

__ fYi- E-Ej-E->-EP- E >-W- E-EW-I- EP W EW ~

!

-

I

50 100 500 iOOO 2000 5000 10000 CYCLES TO FAILURE IN THOUSANDS

a. WELDS PREPARED WITH E6010 ELECTRODE

o 344-X,E6012 W- fAILURE IN WELD METAL !

A 344-XX,E 6012 E = FAILURE AT EDGE Of METAL • 344-F, E6013 P- FAILURE I N PLATE

I

~ VI. Y1. '!! Vi. Ey/' WWE ~WE -- T U E ~

W W W o ll. A T~~A!I W-E-

. fYi~E ----_ .. - ---- - -------

50 100 500 1000 2000 5000 10000 CYCLES TO FAILURE I N THOUSANDS

b. WELDS PREPARED WITH E6012 AND E6013 ELECTRODES

RESUL"rS OF PAST AND PRESENT FATIGUE TESTS OF TRANSVERSE BUTT WELDS IN THE AS-WELDED CONDITION PREPARED WITH E60tO, E60t2 AND

E6013 ELECTRODES; TENSION = COMPRESSION CYCLE

50

40

;; 30 x: z V) 20 U') w 0:: IU')15

1025

50 )

4

;;;3 ~

z

~2 w a: ICIlI

f' ....

0

0

5

I~

o 344- T 4 J T5 E6020 W= fAILURE IN WELD METAL o 344- R E 6030 E =FAfLURE AT EDGE OF WELD f!& 344-U5 E6030 P=FAILURE IN PLATE

w,..E ~ V.( WE "" wt.... ... r'"

~i 'g EP ~ ~6 0 I E EP

I

- -

50 100 500 1000 2000 5000 10000 CYCLES TO fAILURE IN THOUSANDS

a. WELDS PREPARED WITH E6020 AND E6030 ELECTRODES

I_ 0 344-K CARBON-ARC W=FAILURE IN WELD METAL A 344-L UNIONMELT E=FAILURE AT EDGE Of WELD

P=FAILURE IN PLATE

g' E.fP EP

at EP ~EP 11. f!&Wf-P i

~.9~JP if ~ EP ! f!& L..r

E"P EP

----- -

50 100 500 1000 2000 5000 10000 CYCLES TO fAILURE IN THOUSANDS

b. WELDS PREPARED BY AUTOMATIC WELDING PROCEDURES FIG.27 RESULTS OF PAST AND PRESENT fATIGUE TESTS OF TRANSV~RSE BUTT WELDS

IN THE AS-WELDED- CONDITION PREPARED WITH E6020 AND E6030 ELECTRODES AND WITH AUTOMATIC PROCEDURES: TENSION=COMPRESSJON CYCLE

50

40

C;; 30 ~

Z

(f) 20 ~ a: ~ 15

-()

1025

~.

.'-"

0

50

Q r< f:O .--. .... "~,ry'f1 "A~

0 Q( > . ~ 'P OO~

100

o E6010 o E6020 A E6012 A E6Q30 V E6013 • AUTOMATIC

...n. 1- -4l A A ,. ~ ~ --

~ .... ,

A • • 0

.. ~'Y~n: [7. r ...... .-. >J1

_ .... ..-~ ,...~ r.', M._ nI..I. J 4 >0 0 00 00 . 1-- ..... ....,

500 1000 2000 5000 10000 CYCLES TO FAILURE liN THOUSANDS

FIG.28 SUM~~ARY OF RESULTS OF ALL PAST AN[) PRESENT FATIGUE TESTS OF TRANSVERSE BUTT WELDS IN THE AS-WELDED CC)NDITION: TENSION = COMPRESSION CYCLE

50

40

CI) 30 ~

z

CI) 20 CI)

LLJ Q:

~ 15

1Cl.Z5

-0-0

50 100

o 327-A,B 120cf f. STRESS RELIEF

....

C ~ -1---1

500 1000 2000 5000 10000 CYCLES TO FAILURE IN THOUSANDS

FIG.29 RESULTS OF PAST FATIGUE TESTS OF TRANSVERSE BUTT WELDS WITH THE REINFORCEMENT ON WELDED WITH SPECIAL PROCEDURES; TENSION=COMPRESSION CYCLE

60

50

U) ~ 40

Z

~ 30 w 0:: .... U)

U)

20

1540

60

50

~ 40

Z

U) 30 ~ " t;

20

1\0

FIG. 30

o 327-0.P E6010 • 384-fl E6010 • 1953-54 E 7016 W=fAILURE IN WELD METAL I A 327- Z £6010 , 1953-54 E6010 o 310- 7~ELECTROD£' E· FAILURE AT EDGE Of WELD V 327-X £6010 A 384-HI E6012 NO SPEClrlED) P= fAll.URE IN PLATE

'(&W b ~~'qvp ... E , ~'q ~ y ~~ W E S ~

, • iAtMt~ w-w -~~~ A ~ Q-t ..

'II . WW W- rYp

~ W -~ ,-~ ...... j~

~

~ ~ -----

100 500 . 1000 2000 5000 10000 CYCLES TO fAILURE I N THOUSANDS

a. WELDS WITHOUT PRE- OR POST-WELD SPECIAL TREATMENTS

Waf AlLURE IN WELD METAL o 3~7 -M4' A 364-F

1200·F STRE5S RELIEF I V 384-H][ E6012t "50-F STRESS RELIEF 1I00of STRESS RELIEF 0 327-U HIGH EMP. STRESS RELIEF E=FAILURE AT EDGE OF METAL

P-fAILURE IN PLATE

.~~ r-S ~ E E.

W ~ YJ"!. p

~ ref 0

~w -""'ti-" .6

! w ~yy

100 500 1000 2000 5000 CYCLES TO FAILURE IN THOUSANDS

b. WELDS WITH PRE- OR POST-WELD SPECIAL· TREATMENTS

RESULTS OF PAST AND PRESENT FATIGUE TESTS OF TRANSVERSE BUT-r WELDS WITH THE REINFORCEMENT OFF; O-TENSION CYCLE

00 00

CI)

~

z

~ IJJ 0:: r V)

5 u

~o

30 EaE ~ - -

20

15

In 25

~ P P e. e f 1

.. ..., -~ . E P • 0

.. E

50 100

o 327-G.H • 327-EJFJ l20cfF. W=FAILURE IN WELD METAL A 327-Y STRESS RELIEF E = FAILURE AT EDGE OF WELD V 344-K P = FAILURE IN PLATE

P P ~ 0 0 p

0 J_ .. -P W WT'#J rp V v---W

i

~OO 1000 2000 5000 10000 CYCLES TO FAILURE IN THOUSANDS

FIG.31 RESULTS OF PAST AND PRESENT fATIGUE TESTS OF TRANSVERSE BUTT WELDS WITH THE REINFORCEMENT OFF; TENSION = COMPRESSION CYCLE

i· i/

60

CI) 50 ~

Z 40

en en W 30 0:: t-CI)

2~0

40

30 't----EW

CI)

~

Z - 20 (/') (/') w a: 15 .... CI)

I~O

A

1 0 310,8; REINFORCEMENT ON, C-MO. ELECTRODEJW- FAILURE IN WELD METAL • 310,9' REINFORCEMENT Of~ C-MO. ELECTRODE' E-FAILURE AT EDGE OF WELD

W • Ia

~ 0 • ~ E W ~ i -- --

50 100 500 1000 2000 5000 ·10000 CYCLES TO FAILURE IN THOUSANDS

a. RESULTS OF TESTS OF WELDS IN SILICON STEEL; O-TENSION CYCLE

o 327. 101-109; REINFORCEMENT ON, NAVY GRADE EAELECTROOC W= FAILURE IN WELD METAL A 350, 1~-18; GROUND FLUSH ON ROOT SIDE ONLY,. E6010 E= FAILURE AT ALLOY STEEL EDGE OF WELD V 350, 46-48; REINFORCEMENT ON, E6010 EC= FAILURE AT CARBON STEEL EDGE OF WELD • 350, 74-75; REINFORCEMENT ON. E6010, CARBON STEEL WELD- PA=FAILURE IN ALLOY STEEL PLATE

ED TO LOW ALLOY STEEL PC=FAILURE IN CARBON STEEL PLATE

E E PC PC A V E

• ·t '~tA. ~ I P 'A E E;,...J:-E EW A

\U 0 E~ -E A. --V ~ o •

LI

E I'C EC -EC

---- -- ________ ~L_. ____ ----

50 100 500 1000 2000 5000 10000 CYCLES TO FAILURE IN THOUSANDS

b. RESULTS OF TESTS OF WELDS IN LOW ALLOY HIGH TENSILE STEEL; TENSION=COMPRESSION CYCLE

FIG. 32 RESULTS OF PAS'- TESTS OF TRANSVERSE BUTT WELDS IN SILICON STEEL AND IN LOW ALLOY HIGH TENSILE STEEL

01 the fatigue and static properties of butt welds in structural - Ideals

Documents

Transcript of 01 the fatigue and static properties of butt welds in structural - Ideals