Embed Size (px)
Citation preview
10 r~qA =It;{54-
- CIVIL ENGINEERING STUDIES t"Of1 3 STRUCTURAL RESEARCH SERIES NO. 254
FATIGUE STRENGTH F TRANSVERSE BUTT-WELDED
JOINTS IN N-A-XTRA 100 STEEL
By G. R. SCOTT
J. E. STALLMEYER
and
W. H. MUNSE
A Report of an Investigation Conducted
by
THE CIVIL ENGINEERING DEPARTMENT
UNIVERSITY OF ILLINOIS
In Cooperation With
NATIONAL STEEL CORPORATION
UNIVERSITY OF ILLINOIS URBANA, ILLINOIS
FEBRUARY 1963
March 1963
ERRATA
SRS No. 254
PleL.se make the following changes and additions in the above noted report:
0'" po c.l.
p.,21
po 21
The paragraph at the top of this page should actually be at the bottom as the last paragraph of Section 6020
Line 7 - Il stress level in order" should read "stress cycle in ordero l1
Line 3 - Fig. 27( a) should be Fig. 27(b).
Line 8 - I1stress level" should be tlstress cycle. tI
Line 13 - Fig. 27(b) should be Fig. 27(a).
Change title of (a) to Specimen GLB-45. Add ne1VoJ subtitle belOi{ the bottom picture as follow's:
(b) Specimen GLB-57
FATIGUE BEHAVIOR OF BUTT=WELDED
JOINTS IN N=A=XTRA 100 STEEL
by
Go Ro Scott
Jo Eo Stallmeyer
and
Wo lIo Munse
A Report of an Conducted
THE CIVIL ENGINEERING DEPART~JEND:r
UNIVERSITY OF ILL1~OIS
In ((Jo«:rpe:raot,~cm With
National Steel ((Jo~:"'oolra)"tlCjn
UNIVERSITY OF ILLINOIS
URBAl\fAJ ILLI'NOIS
IIo
IIIo
TABLE OF CONTENTS
INTRODUCTION 0 0 0 0 " 0 o 0 0 0 o 000 000 0 0 0 0 0
Object and Scope of Program 0 0 0 0 0 0 0 0 0 0 0 0
Acknowledgment 0 0 0 0 0 0 0 0 0 0 0 0 g " 0 0
MATERIALS 0 0 gOO o " 0 0 0 " 0 000 000 o 0 00(1 0
WELD QUALIFICATION o 0 o 0 0 0 0 o 0 0 " 0 0 o 0 000 0
301 Tentative Welding Procedure 0 0 0 0 0 0 0 0 0 \) 0 \)
30101 First Tt"ial" " 0 0 0 0 0 0 " 0 0 0 0 0 0 0 0
30102 Second Trial 0 0 0 v- 0 0 0 0 0 0 0 0 0 \) 0 0
;02 Welding Procedure PIOQ"",llOl8c 0 0 0 0 0 0 0 0 0 0 0
303 Welding Procedure PIOO=llOl8B 0 0 0 0 0 0 0 0 0 0 " 304 Welding Procedure PIOO=llOlSD 0 0 0 0 \) 0 0 0 0 0 0
;05 Discussion 0 0 G 0 0 0 \) " 0 0 0 0 0 I) 0 0 0 0 0 0
SPECIMEN FABRICATION " 0 0 0 0 o 000 0 0 \) o 0 000 (I
General 0 0 0 0
Plain Plate 0 0
Welded Plates 0
o 0 0 000 0 0 0 0 0 0 0
o 0 0 0 0 0 0 0 0 0 0 0 0
o 0 000 0 0 0 0 000 0
00000
o 0 0 0 0
o 0 0
TEST PROCEDURE " 0 000 000 " 0 0 000 o 0 0 0 000
Static Butt=Welded Joints 0 0
Fatigue Tests 0 0 0 0 0 0 0 0
000 0 gOO 0 0 \) 0
o 0 0 0 0 0 0 0 000
VIo EXPERIMENTAL RESULTS AND DISCUSSION 00000 000 0 0
Evaluation of Fatigue strength 0 0 0 0 0 0 0 0 0 0
Static Tests of Butt=Welded Joints 0 0 0 0 0 0 0 0
Fatigue Tests of Plain As=Rolled Specimens 0 0 0 0
Fatigue Tests of Butt=Welded Joints 0 0 0 0 0 0 0 0
60401 Zero-to-Tension Stress Cycle 0 0 0 0 " 9 0 0
60402 Half=Tension=to=Tension Stress Cycle 0 0 0 0
60403 Completely Reversed stress Cycle 0 0 0 0 0 0
60404 One-Third Compression=to=Tension Stress a Q
60405 One~Quarter Tension=to=Tension Stress Cycleo Fretting Failures 0 0 0 a 0 0 0 0 a Q 0 ~ 0 0 Q a 0
VIla METALLURGICAL STUDIES o 000 0 0 000 0 0 0 0 0 000
VIII 0 SUMMARY o 000 000 0 0 000 Q Q 0 0 a 0 0 0 000 0
TABLES 0 000 0 000 0 0 0 a 0 0 0 0 0 0 0 0 000 0 0
FIGURES o 000 0 000 0 0 0 0 0 0 a 0 0 0 0 0 0 000
iii
1
1
"
4
5
5 5 6 6 7 8 9
11
11 12 12
14
14 14
16
16 17 18 19 19 19 20 20 21
21
23
25
26
37
.Table Noo
1
2
3
4
5
6
7
8
9
10
11
12
13
14
LIST OF TABLES
Chemical Composition of N=A=XTRA Plates
Physical Properties of N=A=XTRA Plates
Physical Characteristics of Base Metal
Physical Characteristics of Base Metal
Summary of Weld Qualification Tests Procedure PIOO=llOl8c
Summary of Weld Qualification. Tests Procedure PIOO=llOl8B
Summary of Weld Qualification Tests Procedure PlOO=llOlBD
Results of Static Tests of Transverse Butt=Welded Joints in the As=Welded Condition
Results of Fatigue Tests of Plain As=Rolled Plate Specimens (Axial Tension)
Results of Fatigue Tests of Transverse Butt Welds in the As=Welded Condition (Axial Tension)
Results of Fatigue Tests of Transverse Butt Welds in the As=Welded Condition
Results of Fati.gue Tests of Transverse Butt Welds in the As=Welded Condition (Complete Reversal, Axial Loading)
Results of Fatigue Tests of Transverse Butt Welds in the As=Welded Condition (Partial ReversalJ
Axial Loading)
Results of Fatigue Tests of Transverse Butt Welds in the As=Welded Condition (Pulsating TensionJ
Axial Loading)
iv
26
26
27
27
28
29
30
31
32
33
35
Figure Noo
1
LIST OF FIGURES
Location of Specimens in Parent Plate N-A=XTRA lOOJ Plate T-l j 3/4 ino Thick
Location of Specimens in Parent Plate N=A=XTRA 100» Plate T=2J 3/4 ino Thick
Location of Specimens in Parent Plate N=A=XTRA 100y Plate T-3J 3/4 inoThick
4 Location of ASTM 00505 Round Specimens
5 Tentative Welding Procedure (Transverse Butt Weld)
6 Results of Magna=Flux Inspection
7 Welding Procedure PIOO=llOl8C (Transverse Butt Welds)
8 Welding Procedure PlOO=11018B (Transverse Butt Welds)
9 Welding Procedure PIOO=llOl8D (Transverse Butt Welds)
10 Results of Face Bend Tests = Procedure PIOO=11018D
11 Results of 'Root Bend Tests = Procedure PIOO=llOl8D
12 Results of Free Bend Tests = Procedure P100=110lBD
v
13 Fractures of Tensile Test Specimens = Procedure P100=110lBD
14· Details of Test Specimens
15 Butt Welded Joint Before Fina,l Machining
16 Illinois u Fatigue Testing Machine as Used for Axial Loading of Welded Joints
11 ' Typical Failures of Static Tests on Butt=Welded Joint Specimens
18 Results of Fatigue Tests of Plain As=Rolled Specimens Axial Tension
19 Typical Failure of Plain Plate Fatigue Specimen
Figure Noo
20
21
22
23
24
25
26
27
28
29
30
31
vi
Results of Fatigue Tests of Transverse Butt Welds in the As-Welded Condition - Axial Tension
Fracture of Butt Welded Joint Specimens
Results of Fatigue Tests of Transverse Butt Welds in the As-Welded Condition = Pulsating Tensions Axial Loading
Fracture of Specimens GLB-25J GLB=35
Results of Fatigue Tests of Transverse Butt Welds in the As=Welded Condition = Complete ReversalJ Axj,al Loading
Results of Fatigue Tests of Transverse Butt Welds in the As=Welded Condition = Part,ial Reversaly Axial Loading
Results of Fatigue Tests of Transverse Butt Welds in the As=Welded Condition = ~llsating TensionJ Axial Loading
Typical Failures of Butt=Welded Fatigue Specimens
Diagram Showing the Effect of Welding and Weld Geometry on Fatigue Life = Axial Tension
Reduction in Fatigue strength of As=Rolled Plain Plates Due to a Transverse Butt Weld
Surn.rnary of Results = Fatigue Tests of Transverse Butt Welds in the As=Welded Condition = Axial Loading
MOdified Goodman Diagram for Transverse Butt Welds in N-A-XTRA 100 Steel,9 in -the As"'Welded Condition
SYNOPSIS
Axial load fatigue tests on butt-welded joints and on as-rolled
plate of N-A-Xtra 100 are reportedo All tests on as=rolled plate were
conducted on a zero-to-tension stress cycleo The majority of tests on
butt-welded joints were conducted on three different stress cycles~ complete
reversal, ze:ro-to-tension, and half tension-to-t,ensiono A few tests were
also carried out on two other stress cycles~ one-third compression-to
tension and one-quarter tension-to-tensiono Fatigue lives in the range
from 50,000 to 2,000,000 cycles were obtainedo The data are presented in
the form of S-N curves and are used to develop a MOdified-Goodman diagramo
All welded joints were subjected to x-ray examination and were
free from defects within the limits of 2 percent sensitivity on the x-rayso
Consistency of the welded joints is indicated by the resultso
The results show that N-A-Xtra 100 is a weldable steel and does
not require any extraordinary precautionso Properties of welded joints
subjected to fatigue loading areJ in general) in line with what one would
expect for a high strength heat-treated structural steel on the basis of
past experience with similar materials of the same thicknesso
FATIGUE BEHAVIOR OF BUTT~WELDED
JOINTS IN N-A~XT.RA 100 STEEL
I I) INTRODUCTION
1$1 Object and Scope of Program
Since the introduction several years ago of high strength steels
with good weldability there has been increasing interest in the use of these
materials for a wide variety of applicationso In many of these applications
the structures are subjected to Z'epeated loads varying in ma,gni tude and in
frequency of occurrence.. In order to provide such structures with an adequate
margin of safety it is necessary to determine allowable stresses from informa=
tion obtained in experiments on the fatigue strength of such materi.also
The role of stress concentrations in fatigue has been investigated
on a variety of specimen configurations and its effect is well known 0 Any
stress concentration can have a marked influence on the fatigue behavior
and may cause failures to initiate in materials subjected to repetitions
of nominal stresses considerably below their static ultimate strengtho In
fatigue failures are associated. with one or more of the
(a) stress concentrations,
(b) Repetitions stress,
( Large amplitudes and/or high mean stresseso
A weld or welded joint affects the fatigue behavior of the base
material in a threefold mannero First, the weld constitutes a mechanical
discontinuity and acts as a stress concentrationo The effect of this stress
concentration on the fatigue strength of structural components is well knowno
2
The second factor involves the change in properties which results
from the heat cycle imposed d-u.ring the welding operationo The metal in the
heat affected zone is subjected to) a thermal cycle which results i,n changes
in the grain structure in this regiano This change in grain structure is
usually accompanied by a change in the physical properties from those which
prevailed before thermal cyclingo
Residual stresses imparted to the weldment constitute the third
factor contributed by weld:i,ng 0 Since residual stresses :may change the local
cyclic conditionsJ there is reason to believe that they will also have an
influence on the fatigue behavior of a weldmento
The exact na tut"e of the three items mentioned above is not suf
ficie:rl'tly well defined to permit a direct evaluation for a:.ny particlllar
material () For the high strength steels the item of metallu:rgical st;ructure
is extremely complex because of the valC'iety of structures that occur in
heat affected zone o
Any study of the fatigue behavior of welded structures is furthe:~
complicated by the fact that it is virtually :impossible to
sound welds or to duplicate welds in all of their detailso There at
presentJ no way t,o evalua;t,e the effect of weld flawlQl on the fatigue behavior
of weldmentso
The work reported herein W8,lQl undertaken to provide basic information
on the fatigue resistance of N=A=XTR:.ti steelo This ma,terial is a quenched and
tempered steel available in plates up to 1 ino thick 0 The studies include
fatigue tests of plain as=X'olled plates a:nd butt=welded joints under a
variety of cyclic stresses 0 A sufficient number of tests have been carried
3
out to permit an evaluation of the fatigue relationship from 100»000 to
2J OOOJ OOO cycleso These data have then been used to construct a MOdified
Goodman Diagram for different fatigue liveso
Metallurgical examinations of typical specimens have been con=
ducted to determine the metallurgical changes imparted due to the welding
procedures usedo
102 Acknowledgments
The tests described in this report were ca~ried out during the
course of an investigation conducted in the Engineering Experiment Station
of the University of Illinoiso The program was c~ried out with ~lnds
provided by Great Lakes Steely a subdivision of National Steel Corporationo
The investigation constitutes a pal~t of the structural res~arch
program of' the Civil Engineering Deprurtment of which Dr 0 No Mo Ner'itJInaX'k is
the Heado The research ws carried out by Go Ro Scott» Research Assistant
ip, Civil Engineering under t;he supervision of J 0 Eo Stallmeyer,l' Professor
of Civil Engineeringo
The authors wish to express their appreciation to a number of
persons on. the UniversityU s staff who have assisted in the conduct of' the
investigationQ These include Wo To Becker» Research Assistant who conducted
the metall1l't"gical investigations» JQ Wo Rone who helped in a portion of the
testingj and the Civil Engineering Shop staff who prepared the test memberso
4
IIo MATERIALS
All of' the material used in the investigation was furnished by
Great Lakes Steel Corporation in the form of three 6 x 12=fto plates of
3/4-ino steel 0 Chemical analyses as furnished by Great Lakes Steel and
as obtained from check analysis are presented in Table 10 The physical
properties as reported by Great Lakes Steel Corporation a'iJ:"{e presented in
Table 2 0 In addition to the physical properties supplied by 'the mw:1ufact'UI"er»
a series of tests on flat plate coupon specimens» taken from one of the
original plates» was carried out 0 The resw;ts of' these tes'ts are presented
in Table 30
During the progress of the investigation a series of small coupon
specimens were prepared from the pull=heads of butt~welded specimens which
had been subjected to static testso These specimens were taken from the
pull-head material in order that they would be as free from the effects of
prior stress as possibleo In order to conform to this requirement ASTM
00505 round specimens were preparedo The results of tests c~ried out on
these specimens are presented in Table 40 It should be noted that the
digit of the specimeh designation represents the number of the full size
static specimen from which it was removedo The right hand digit is merely
an indication of the order of testingo
The location of all specimens in the original p~ent plate is
shown in Figso 1» 2 and 30 The locations of the 00505 rounds taken from
the pull-heads are shown in Figo 40
IIIo WELD QUALIFICATION
301 Tentative Welding Procedure
30101 First Trial
5
Two plates~ designated GLQl and GLQ2~ were cut off from the
original materialo These plates were bevelled for a 60 dego double=Vee
butt-weld and welded according to a tentative welding procedure» Figo 50
It can be noted that in this tentative procedure, and in those that are
to follow, the surface contour of the weld was obtained with one Single
pass (5 and 6 in this case)~ thus obtaining a weld which would probably
be more resistant to fatigue than one with a multiple pass contouro
On the second pass, while the weld was coolingv a definite crack
was observed visually and the welding operation was stopped immediatelyo
This weld along "With the adjoining portion of base metal; was removed and
constitutes specimen Q~lo The specimen was inspected the magnetic
particles principle (Figo 6) and J{:=rayedo The x=ray picture was tak.en to
indicate the actual length of the crack at those locations where it 'w,s
not visible to the naked eyeo The picture indicates that the pri~y
crack is approximately 8 ino long 0 There is? in addi tion,9 a second crack
which is approximately 1 1/2 ino longo
In order to maintain a gap of 1/8 ino between the two plates
being welded» a bead of filler metal had been deposited at one end of the
joint 0 It is possible that this procedure resulted, in large residual
s-tresses which caused or contributed to the formation of the cracko
Specimen Q-l was examined metallurgically and the results are discussed
in Section Vllo
6
30102 Second Trial
The remaining portions of plates GLQ~l and GLQ=2 were rebevelled
for the same type of weld and the same welding procedure was used in a second
trial 0
After the first passJl in the process of cooling,9 a definite crack
appeared and the welding operation was stopped 0 This time the two plates
were left under restraint in the clamping device for a longer periodo The
result was that the crack was more easily seen than in Specimen Q=lo
The weld and a small portion of' the base metal was removed from
this qualification specimeno This specimen was given the designation Q-2o
gave way and that an opening as wide as 1/4 ino in the central portion could
be seeno No supplementary inspection of the weld was madeo
It is apparent that the procedure used for the preparation of these
two test welds was not adequateo The manner in which the second crack opened
up confirmed the fact that high longitudinal residual stresses had been
created and the fact that the plates were tacked at the far end to keep the
gap constant probably contributed to the development of these high residual
stresses 0 One way to reduce these residual stresses is to put more heat
into the weldn This was taken into account in the next welding procedureo
302 Welding Procedure P100~110l8c
The designation used for this welding procedureJ and all of those
which follow} has the following meaning 0 The letter P stands for gSProcedure 9H;
the next three digits are an i'ndication of the approximate yield point of
the base metal in kips per sqo ino; the next five digits represent the AWS
7
designation for the electrode used and the letter serves to differentiate
between procedures used with the same base metal and electrodeo
The two plates} GLQ=l and GLQ=2y were once again prepared for a
60 dego double-Vee butt=weldo The details of the procedure are shown in
Figo 70 It should be noted that the end of the two plates to be joined
were not tack weldedo Instead, in this procedure a varying gap was used
to compensate for the effect of shrinkageo This welding ~ocedure was also
different from the one previously used in that the interpass temperature was
raised, the plates to be welded were preheatedJ more heat was furnished for
the first pass and the pass pattern was changed 0
After satisfactory completion of the qualification test weld the
joint was x=rayedo The test weld was then machined for qualification
specimens 0 Qualification specimens were subjected to guided bend tests
with a 1 1/2=ino diameter mandrel o The results of these qualification tests
are summarized in Table 50
After the tests had been completed the specimens were examined 0
The welds appear to be extremely cleano It appeared from the ~lided bend
tests that the specimens failed primarily because the mandrel was too small
in diameter and imposed conditions which were too severe for the high strength
weld metal o
)Q3 Welding Procedure PIOO=llOlBB
An additional weld qualification specimen was fabricated from plates
GLQ-3 and GLQ=4o The welding procedure used for this qualification weld is
shown in detail in Figo 80 With this procedure the only major change is
8
that the base material was prepared for a single=V butt weld» and this was
done to comply with the 1956 American Welding Society Standard Specification
For Welded Highway and Railway Bridges 0 The following is an excerpt from
this specification 0
D2030 Base Materials and its Preparation
(c) For groove welds where the thickness of the
material as specified in Article D205(a) exceeds 3/8 ino}
the preparation of the base material for welding shall
be for a single~V groove butt joint meeting the require
ments of Figo D=23, except that in the case of oxyacetylene
welding the joint preparation may be in accordance with
Figo D-25o
After completion of the qualification test weld the joint was
x=rayedo The completed weld joint was then machined for qualification
specimens 0 The results of the tests carried out are presented in Table 60
It should be noted that one of the face bend tests was carried out
with the 1 1/2-ino diameter mandrelo This one test was carried out in order
to verify the significance of mandrel size on the results of the testo All
other tests were carried out with a 2 1/2=ino diameter mandrelo
304 Welding Procedure PIOO=11018D
This procedure is essentially the same as P100=110l8c y except that;
a shorter weld was used and consequently the varying gap had to be reducedo
The welding procedure used is shown in detail in Figo 90
The test results u'om the specimens cut from the resulting plate
are presented in detail in Table 70 Free Bend Specimen FR=51 was bent more
9
than 90 dego to observe, if possibleJ the ductility left in the weldo It
was not until a bend angle of approximately 180 dego was obtained that a
small crack developed at one of the corners of the specimeno An elonga
tion of 31 percent between the gage marks on the weld was measured at that
time 0
Figures 10 through 13 show photographs of all specimens prepared
with this procedure after they had been subjected to the required testso
All butt~welded specimens used in the course of this investiga
tion were prepared by welding operators qualified with this procedure in
accordance with AWS requirementso
305 Discussion
There are several things which were brought out as a result of
the weld qualification tests conducted which warrant attentiono The first
attempts to produce sound welds failed before a joint could be completed
because of the presence of a tack weld which caused a most unfavorable
residual stress field to be developedo
After a procedure had been developed which produced a joint
which appeared to be satisfactory the individual bend specimens failed to
meet requirementso This result has been attributed to the fact that a
1 1/2=ino diameter mandrel was used for these tests in place of the 2 1/2=ino
diameter mandrel recommended by ASME for these high strength steelso It
should be noted} however, that these specimens were so prepared that the
welded joint was parallel to the direction of rollingo Consequently» in
the bend tests the direction of bending was transverse to the direction
of rollingo In this particular steel no indication of differences in
10
directional properties was detectedp howevery ih other steels with different
directional properties this could be a serious factoTo
In any event utilization of a proper welding procedure for this
material produced good results in both the bend tests and tensile tests of
the welded jointso All test results for the final welding procedure indiGated
excellent ductility across the welded joint and high static strengthQ
11
IV 0 SPECIMEN FABRICATION
401 General
All plates used in the fabrication of test specimens were
originally flame cut from plates T-lJ T-2 and T=3o All plates were
removed from the parent material in such a manner that the direction of
rolling was parallel to the applied stress 0 Each specimen has an overall
length of 4 u=ou with a section reduced to 4 ino in width at the centero
The straight portion in the center is 5 inc long with a transition radius
to full width of 9 ino The width of the central section is controlled
by the thickness of the materialy the capacity of the testing machine and
the cyclic stress to be appliedo The transition radius is provided to
avoid large stress concentrations which are usually more critical in notch
sensitive steels under fatigue loadingo
It is customary to remove part of the material to be machined by
removing a vee-shaped section on each side of the plate at the central
portion by sawing (Fig 0 15) 0 At no time is the metal removed near the
test section by flame cuttingo The remaining machine operations required
to bring the specimens to final dimensions were carried out on a milling
machine 0
After machining had been completed the edges gf the test section
and part of the transition radius were draw=filedo This was followed by
polishing with emery cloth to obtain smooth polished edges with all grinding
marks parallel to the direction of the applied loado
12
4,,2 Plain Plate
Eleven holes at each end are drilled in the flame cut plates
(48~t X lOftB) which are subsequently machined to their final dimensions~
(Figo l4a)o
403 Welded Plates
All the plates to be welded were bevelled for a double~Vee with
60 dego included angle at the edge opposite the designation number shown
on the cutting diagrams of the parent plateso The welded specimens con
sisted of two pieces numbered consecutively~ the odd number being the lower
of the twoo For futvre reference the odd number will be used to designate
a completed weld specimeno The two plates were secured in place in 8;c jig
which could be rotated 180 dego about its axis to permit successive deposits
of the different passes to be made in the dO'Wn=hand positiono Excessive
amounts of torsion by shrinkage while the plates cooled were prevented by
the restraint applied by the jigo
The weld bead was deposited in a 1~ino continuous pass~ Figo 150
Welding procedure PlOO"",llOl8D was usedJ and a constant gap of 1/8 in" was
used between the two pieces to be weldedo Special care was given to the
way in which a pass was terminated, since cutting the electrode arc too
fast without slightly building up the bead seemed to resultJ occasionally,
in a small crack at the end of the pass upon coolingo
Upon completion of the weld the specimens 'Were x=rayed 0 Results
of x-ray examination indicate that the 'Weld quality was excellent in all
but one case" No porosity, slag inclusions or lack of penetration were
13
detected with x-ray sensitivity gf at least two percento Specimen GLB=3
had one location of undercut which was indicated on the x=rayo
After x~rayexamination the specimens were machined to final
dimensions shown in Figo 14b and dxaw=:filedo
14
v 0 TEST PROCEDURE
501 static But,t=We1ded Joints
Four butt=we1ded specimens were tested statically in a screw
type 600,000 pOlxud testing machine at the University of I11inoiso The
four specimens are G18=l, GLS=3j GLS=5 and G18-7o
A load of 6,000 pounds was applied to the specimen prior to
tightening the bolts to insure that no slipping would occur while applying
the ultimate loado The speed of testing was not recorded for GLS=l and
G18-3 but was less than 004 in/mino recorded for the ~ther two specimens
tested 0 The latter speed is the maximum recommended for testing use when
dealing with loads of a magnitude of more than 200yOOO poundso
502 Fatigue Tests
All the fatigue tests reported herein were conducted at room
temperature in an ordinary non=corrosive environment using University of
Illinois 250,000 Ibo lever type fatigue machineso The speed of these
machines is approximately 100 cycles per minuteo
The essential features of the fatigue machines are shown sche=
mati cally in Fig 0 160 A variable throw eccentri.c transmits the force
through a dynamometer to a lever which» in turnJ transmits the force to
the upper pull~heat at a multiplication ratio of approximately 15 to 10
The force that is exerted on the specimen originat;es in the double throw
eccentric which is adjusted to give the desired range of load before the
test is started 0 The maximum load is controlled by the adjustable turn=
buckle mounted between the eccentric and the dynamometer 0 Dt:lx'i:ng operation
15
of the machine» the test specimen is subjected to a range of stress which
is controlled by the preset» constant deflection of the eccentrico The
stress is determined by an Ames dial which measures the vertical deflection
across the throat of the dynamometer to the nearest 00001 ino
The criterion for failure of the specimens» as far as feasible Jl
was taken as the number of cycles at which a:pproxima;t1ely half the cross""
sectional area of the specimens had be~n fract'Ul"ed 0 This could not be
followed in all cases because the automatic microswitches on the fatigue
machines did not always stop the machines before the members had f'r'actured
half way througho In cases where the complete specimen was fractured» no
attempt was made to correct the life to that for which approximately half
the cross-sectional area would have fracturedo There are two reasons for
this 0 First» there are not sufficient data available to determine the
rate of crack growth in welded members under repeated loadso And» second,
laboratory observa~ions have indicated that fatigue cracks usually take a
fairly small percentage of the fatigue life to propagate in the type
of members usedo Also» if one considers the i,nherent scatter associated
with fatigue testingy it does not seem, necessary to make small corrections
of the nature discussed aboveo
16
VIo EXPERIMENTAL RESULTS AND DISCUSSION
601 Evaluation of Fatigue Strength
To compare numerically the results of fatigue tests of specimens
tested at different stress levels, fatigue strengths corresponding to
failure at particular lives have been computed using the equation
(1)
where S is the stress at which the specimen failed after N cycles j n is the
number of cycles for which the fatigue strength F is desired and k is an
experimentally determined parametero
laboratory observations and is based on the assumption that the finite part
of the S-N relationship, when plotted to a logarithmic scale, can be repre-
sented by a straight lineo Laboratory investigations have revealed that k,
the slope of the S-N curve» is a function of the material properties, the
geometry of the specimen, and the cyclic loading to which it is subjectedo
As a result; the computed values of fatigue strength are only approximationso
Nevertheless, because of the logarithmic nature of the relationship, the
error associated with values of computed fatigue strengths resulting from
any error in the assumed value of k are generally relatively smallo
The results of all series of fatigue tests in this study have been
plotted on a logarithmic basis using nominal stresses and average curves
drawn through these datao A value of k was assumed initially and the fatigue
strengths for two lives computedo The average values of these fatigue strengths
were then used to determine a new value of k and the process repeated until
the assumed and computed values of k coincided 0
17
Typical failures of the static test specimens are shown in Figo 170
These photographs show very clearly the fact that the failure occurred, entirely
in the base metal at a fair distance from the butt=welded jointo Excellent
ductility is also apparent when one notes the reduction in width of the cross
section 0 A similar reduction in thickness was a.lso apparent 0
602 static Tests of Butti=Welded Joints
To determine the static strength of full-size butt=welded joints four
specimens (GLS=lJ GLS=3Jl GLB=5,~ GLS=7) were prepax'ed with 'welding procedure
PIOO=llOlSDo The plates for these specimens were originally cut from parent
plate T=2 and machined to the same final dimensions as the fatigue specimens,,,
In all four specimens failure Olccurred in the base metal at a distance
of approximately 2 ino from the tee of the butt="l.veldo The test results, pre=
sented in Table 8J irulicate that the large specimens had yield points and
ultimate strengths which were slightly below those obtained on the small coupon
speci,mens 0
The top portion of specimens GLS=l and GLS=3 were retained for further
examination and testso Tr!e bottom of each specimen was shipped to
National Steel Corporation fOlr add,i tional chemical and physical tests 0 Eight
ASTM 00505 ino diameter specimens were machined from the top portions retained
in the laboratoryo
The resu~ts obtained on the full=size butt=welded joints showed a
decrease of 3 percent and 5 percent respectively from the yield strength and
-ultimate strength obtai,ned from the 00505 ino diameter specimens taken from
the pull=heado These results do not appeat" to be significantly different from
those whi,ch were obtained on the flat plate coupon sp~cimens 0
18
603 Fatigue Tests of Plain As=Rolled Specimens
A total of six specimens were subjected to fatigue tests in which
the cyclic stresses varied from zero to a maximum tensiono The specimens
were tested in the as-rolled condition with no treatment to the plate surfaceo
Special precautions were taken on the sides of the specimens to insure that
all machining and polishing marks were in the direction of the applied stresso
The results of all of the tests are presented in Table 9 and are
presented graphically in Figo 180 Also presented are the locations of the
failures and the computed fatigue strengths from the individual test resultso
The S-N curve plotted in Figo 18 represents the average of the fatigue
strengths computed for the individual test results and corresponds to
k =: 001430
Some'difficulty was encountered in completing the tests for the
plain plate tests under fatigue cycling 0 At the long lives frettir~ fatigue
failures occurred at several locations in the vicinity of the pull=headso
These failures occurred at approximately 1J OOOJ OOO cycleso It was necessary'
to repair these specimens before the tests on the plain plate could be com=
pletedo It is for this reason that two of the specimens tested at low stress
levels show no failureo When the last fretting failure occurred after the
specimen had resisted 2»000JOOO cycles of stress there 'Was little information
to be obtained by continuing the test beyond this pointo
A typical failure contour and fractu.t"e s'llrface are shown in Fig 0 190
The transition from propagation as a fatigue failure to one which consists
primarily of static failure is shown very clearlyo
604 Fatigue Tests of Butt=Welded Joints
60401 Zero-to-Tension Stress Cycle
19
A total of nine specimens were used in this phase of the investi
gationo All specimens were tested continuously at stress levels such that
failur.es occurred f'r'om 50,000 to approximately 3J 000,9 000 cycles I) The
results of all tests conducted are presented in Table 10 and are plotted
in Figo 200 The fatigue strengths corresponding to failur.e at 100;-000 and
2»OOOjOOO cycles have been tabulated in Table 100 The average fatigue
stren~ths calculated were used in drawing the relationship shown in Figo 200
Several points are plotted in Figo 20 as open circles with a crosso
This method of plotting indicates that the data represented by this point
has been obtained from a specimen which had, previously been subjected to
fatigue stresses without failureo In this particular case the two results
obtained in this way were not used in evaluating the average fati~le strengtho
These two results were not used because in one case failure occurred by
fretting in the pull=headso In the other case the increase in applied stress
is not sufficiently large to guarantee that the coaxing effect would not be
significant 0
The fract'UX'e surface of GLB=7 and GLB=5 a:re shown in Figo 21 and
give a general indication of the type of failures which were obtained for
both high stress and low stress fatigue crack propagationo
60402 Half-Tension to Tension Stress Cycle
A series of seven specimens were subjected to this stress cycleo
One specimen was subjected to retest since it resisted more than 2J OOOJ OOO
cycles of stress without faiDlre at the original stress level to which it
20
was subjectedo The results of all tests are presented in Table 11 and shown
graphically in Fige 220 Average values of fatigue strengths have been used
to plot the relationship shown in Figo 220 Failure surfaces for GLB-25 and
GLB-35 are presented in Figo 230
60403 COmpletely Reversed Stress Cycle
Nine original specimens and three retests were subjected to this
stress level in order to obtain sufficient data for the determination of
the S=N relationshipo All three specimens used in the retest had resisted
at least 2j OOOjOOO cycles at the original test stresso
All failures occurred at the toe of the weld reinforcemento This
is true in practically all fatigu~ tests even though a great deal of care
was exercised to insure that as little reinforcement as possible was present
on all specimenso The results are extremely consistent and give a good
indication of the uniformity which can be obtained with good control on
welding oper.ations 0
The results of all tests are presented in Table 12 ~d are shown
graphically in Figo 240
60404 One-Third Compression to Tension Stress Cycle
A total of th,ree specimens were subjected to this stress cycle 0
The maximum stresses used in these tests were chosen at the expected strength
for a life of 100»000 cycleso These stresses were determined from the
results available for the other stress cycleso
The results of these tests are presented in Table 13 and F:Lgo 250
The BVk!i value used for the determination of the complete relationship is
the avel"age of the values obtained for the zero=to=tension stress cycle
21
and the full reversalo For this reason the value at a life of 2,OOOJOOO
cycles is only approximateo A photograph of a specimen with two fatigue
cracks is shown in Fig., 27(a)., The crack on the lower edge of the w€,ld
reinforcement is clearly visible.,
6.,4,,5 One=Quarter Tension to Tension stress Cycle
For purposes of a check on the exact location of the point for
100,000 cycles on the MOdified Goodman Diagram two additional tests were
conducted at this stress level., The results obtained are presented in
Table 14 and Figo 26" The uku value used in determining the S=N relation=
ship is the average obtained from zero=to-tension and half=tension to
tension stress cycles., A typical fracture surface showing multiple points
of crack initiation along the toe of the weld reinforcment is shown in
Fig., 27(b).,
6.,5 Fretting Failures
During the course of the tests carried out on this program a
number of fretting failures occurred in the pull=heads of both the plain
plate and the butt=welded joint specimens., Most of these failures initiated
near a bolt hole and~e probably influenced by the high pressure exerted by
the bolt as a result of the high clamping forceo In several cases failures
occurred at a point near the bottom of the pull plates used to install the
specimen in the fatigue machineo
The pull plates of the fatigue machine are quite massive and
undergo very small strains even at the maximum capacity of the machineo
The specimen pull-head an the other har!d undergo considerable strain in
.22
this region, thereby producing a differential moyement between the parts
in contacto It is this differential movement which.contributes to the
fretting conditiono
Such failures occurred in a total of nine specimens du.ring the ~'.
course of the program of tests reportedo Such failures occurred ata
range of fatigue lives from 70yOOO to 1»000»000 cycles and occixrred over
the entire range of stress cycles usedo It should be noted f'U.!1'ther that
the nominal stress cycle which occurs in the pull-head is 40 percent of
the nominal stress cycle being applied to the welded joint at the test
sectiono
23 VII 0 MN.rALLURGICAL STUDIES
Specimen Q~l was examined metallurgicallyo No oxide material was
present in the crack, indicating that the crack formed when the weld was coolo
It appears that some ferrite is present in the austenite grain boundaries,
and the initially transgranular crack propagated by branching intergranularlyo
The amount of ferrite is small, and it is suspected that the crack does not
propagate because of the ferriteo
After two unsuccessful trials, plates GLQ-l and GLQ-2 were finally
welded satisfactorily with welding procedure PlOO-llOl8co The resulting
specimen will be referred to as GLQ-lo Specimen GLQ~l was sectioned and
subjected to various bend tests 0 The results are presented in Table 50 This
specimen was welded with a gap of 3/32 ino at one end which was expanded to
5/32 ino at the other end) providing a change in cooling rate along the baro
Bend tests showed more ductility near the end with the wide gap? although
the opposite would have been expectedo The bend test specimens were sectioned
and polished in order to carry out microhardness surveys and to examine the
microstructure in an attempt to determine the cause of the change in bend
ductility 0
When attempts were made in the shop to section the material for
the bend tests» some specimens contained areas of sufficient hardness to
break teeth out of the saw blade 0 One specimen containing the hard zonej)
(Specimen A) and one specimen not containing this zone (Specimen B) . W8.S
polished for metallographic observationo The presence of nitride inclusions
in the base metal at the fusion line was the su~pected cause of the high
hardness 0
24
No microstructural differences were found between Specimen ~gA~f
and Specimen 6!B flY 0 However" in Specimen enA ee the hardness of the base metal
is 30-35 Knoop numbers higher than Specimen E» and the weld metal is 50=55
Knoop numbers higher 0 The base=metal structure is tempered m.artensi te; and
it is possible that the tempering treatment did not produce a reproducible
hardness levelo
No metallographic evidence of nitride inclusions in the heat=
affected zone was foundo
General familiarity of metallographic features of welded N~A=XT.RA
has been obtained through examination of several polished sectionso The
very fine base=metal structure of tempered martensite contains rectangular
shaped zirconium rich particles which appear white and at"e olxtli,ned by a
2 percent nital etcho They do not seem to stand in relief» nor do they
change size or shape when transversing the heat=affected zoneo Nearing the
heat~affected zone: the structure becomes considerably darker~ agglomeration
of carbides occursJl and pearlite appears 0 NextJ 80 regi,on of :m.artensi te is
reached where the prior austenite grain size was large 0 This martensite is
qui te coarse 0 Borderi,ng this zone is the colu1l1naJr' weldmetal structure which
shows a carbide networko
Since quite a difference was fotuld between the base=metal hardness
of Specimens ~~Au and u~Bu~J another series tests was cax'ried out on portions
of GLQ~3J welding procedure P100=110l8Bo The results of these studies showed
a slight decrease in hardness as compared to Specimen i~~~ and were still in
the order of 30 Kneol's higher than Specimen flBB ff~ 0
25
VIII 0 SUMMARY
The effect of the presence of a butt=welded joint on the fatigue
properties for a particular material can be shown in several wayso One
convenient way is to compare the S~N relationships for as-rolled material
and butt=welded joints obtained for the same stress cycleo Such a comparison
is presented in Figo 280 From the curves presented in this diagram it is
possible to determine the percentage reduction in fatigue strength as a
result of the presence of the butt = welded joint at different fatigue liveso
This procedure has been followed in the determination of values presented
in Figo 290
Various stress cycles can be compared on the basis of the pat"ticular
S=N relationshipso Such a procedure has been followed in preparing the
summary of all of the results obtained as presented in Figo 300 The most
useful form of presentation of the results is as a MOdified Goodman Diagramo
Su.ch a diagram is presented in Fig 0 31 for N ~A=XTRA 100 steel 0 The diagram
shows very clearly that the greatest advantage of the high strength steels
lies in the region of low lives or in the region of high stresses where the
amplitude of the variation is small 0 Examination of the Goodman Diagram in
Figo 31 reveals that the curve which represents a fatigue life of 100»000
cycles may have several changes in c'U.'!:"vature 0 This conclusion is not
definite because of the small number of tests which were carried out at
the intermediate stress levelso
26
TABLE 1
CHEMICAL COMPOSITION OF N -A -XTRA PLATES
Heat No. Analysis C Mn p S Si Cr Mo Zr Cu
5M-15629 Great Lakes 0.19 0.97 0.008 0.020 0.78 0.68 0.25 0.10
Check 0.21 0098 00010 0.017 0.75 0.68 0026 0.10 0.03
Heat No.
5M ... 15~29
'*
TABLE 2
* PHYSICAL PROPERTIES OF N-A-XTRA PLATES
Direction Yield Point Ultimate Strength
psi psi
Long. 104.9 040 116,350
Trans. 106,120 119,250
Infor.mation supplied by the Manufacturer.
Elongation in.
2 in. 4 in.
28 18
22 14
TABLE 3 27
PHYSICAL CHARACTERISTICS OF BASE METAL
ABTM Designa- Yield. Tensile Elongation Reduction Test tion Strength Strength in 8 in. in area
ksi ksi percent percent
-* 106 .. 2 1604 64.5 Coupon GLC1 9205
GLC2 9203 106 .. 1 15·0 63.6 GLe3 9301 10705 15·5 6405 GLC4 9502 108.2 1605 6308
average 93·3 10700 15·9 63.9
TABLE 4
PHYSICAL CHARACTERISTICS OF BASE METAL
ASTM Designa- Yield Tensile Elongation Reduction 'rest tion Strength Strength in 2 in .. in ar,ea
ksi ksi percent percent
505 GS-ll 95·5 11100 23,,0 6502 GS~12 9500 110,,0 2105 6504 as ... 31 9500 109·5 2200 6704 GS-32 9600 11100 2305 6404
*it 9706 2304 6505 as-13 11202
00-14** 97,,2 112~O 2309 6500 00-33** 9607 11207 23.8 6200
** 97'06 65·5. 00-34 112·3 2309 average 9603 11],·3 23.1 6501
* 3/4-in .. thick plate" lo5=ino i~ widtho
**The loading speed for these specimens was slightly faster than for the first four tested .. However, a.ll speeds used are well below the maximum permdtted by the ASTM Specification ..
TABLE 5
SUMMARY OF WELD QUALIFICATION TESTS . PROCEDURE PlOO-llOl8c
(Refer to Fig. 1)
Bend Tests
* Specimens Fabricated Test Results
2 Free Bend 2 Pass
2 Face Bend 1 Pass 1 Fail
2 Root Bend 2 Fail
** 2 Side Bend 1 Pass 1 Fail
Tensile Tests
28
Specimen Yield Load Ultimate Load Yield Stress Ultimate St.ress Elongation lbs o 1bso psi psi percent
.T-l
T-2
T ... 3
110J OOO:
113,200
114,000
126,600
128y oOO
128,800
* I 1/21 dia:meter mandrel used
101,100
102,300
**~ Not required according to AWS specifications
*** Elongation in 6 in.
Tensile test. failures in the base metal
113,~ 500
114.1300
l15~600
TABLE 6
SUMMARY OF WELD QUALIFICATION TESTS PROCEDURE PlOO-l1018B
Specimens Fabricated
Specimen
T-4
T-5
2 Free Bend
2 Face Bend
2 Root Bend
Ultimate Load lbs.
(Refer to Fig. 8)
Bend Tests
Tensile Tests
Ultimate Stress psi
110y 300
l09y 400
'* Tested with 1 1/2 in. diameter mandrel.
** Elongation in 6 in.
Test Results
2 Pass
1 Pass 1 Fail
2 Pass
Elongation percent
160)
*
29
Location of failure
base metal
base metal
2 1/2 in. diameter mandrel used where not indicated. (Recommended by ASME)
TABLE 7
SUMMARY OF WELD QUALI~ICATION TESTS PROCEDURE PlOO~'llOl8D
SJ;?ecimens Fabricated
2 Free Bend
2 Face Bend
2 Root. Bend
---,--"-----Spec:tmen
T·-61
Ultimate Load Ibs"
(Refer to Fig. 9)
Bend Tests
Tensile Tests
Ultimate stress psi
y. 2 1/2~in" diameter ma..l1drel used.,
** Elongation in 2 ina
* Test Results
2 Pass
2 Pass
2 Pass
** Elongation percent
30
Location of Fallure
basemetal
ba.semetal
TABLE 8
RESULTS OF STATIC TESTS OF TRANSVERSE BUTT-WELDED JOINTS IN THE AS-WELDED CONDITION
Specimen Yield Point*- Tensile Strength Elongation**- Reduction in Location of No. ksi kai per cent area.) :per cent fracture
GLS-l 91)600 104)700 25 .. 4 45 Base Metal
GLS-3 91,700 10),200 25 .. 2 44 Base Metal
GLS-5*** 93,800 105,700 22.8 45 Base Metal
GlS-7*** 94,400 107;;700 '23 .. 2 44 Base MetaJ.
average 92.9 105.8 24.1 44.5
* By drop of beam.
** In a 5-in. gage length.
*** Loading s:peed of 0.40 in., fmm ..
\.),l f-'
Speeime..""1
~lo"
GIA=4
GLA=5
GUl=l
GIJ.!.=6
GLA-2
GLA=3
* k ::;: 00143
TABLE 9 (Refer to Fig. 18)
RESULTS OF FATIGUE TESTS OF PLAIN AS~ROLL1ID PlATE SPECIMENS
(AXIAL TENSION)
__ ~~_~==_ ~· __ ~~~_.~~_.·~~~_O~=_· .. -=T~~-. __ ~.= .~~~~~~=~~, ~~~~- ~*~--~'
St;reS8 Cy<:le.; ksi
o t.o + 75
('. to + 76
o to + 60
o tc + 54·
o tCl + 52
o to + 52
Life
171;600
178)000
582;:100
2,.112,; 800+
2} 225,,· 700
20 394" 000+
Location of Fracture
at radius
at center
at~ radius
no failure
at :radius
no failure
Com,£uted F8~ti,gue S?:rengths ... " ksi
average
F F 100.9 000 2, 000 ,0 000
8100 82.,4-
7700
8001
5003 5400
5200 5200
52.1 ~~~-~~~-~.~~---~ .. ~.~~.~~~~~ .. =--=. =. ~
VJ I\)
TABLE 10 (Refer to Fig" 20)
R!SlllaTS OF FATIGUE TESTS OF TRANSVERSE BUTT WELDS IN THE AS-WELDED CONDITION
(AXIAL TENSION) .. Specimen stress Cycle, Life Location
** Computed Fatigue Strengths, ksi
Boo ksi of' Fra.cture
GLB-ll Oto+60 47,100 b
*** Oto+60 74,700 GLB- 1 Fretting Failure
GLB- 7 Oto+60 82,800 a.
GLB- 9 o to + 50 143,500 a
-.GLB-39 o to + 32 285,600 a
GLB- 3 o to + 40 452,600 a
GLB-47 O.to + 30 920,300 a
GLB-51 o to + 30, 1,162,800 a
*** o to + 34 GLB- 5 1,210,500 a
GLB- 1 o to + 34 2,768,400+ No Failure
GLB- 5 o to + 28 3,079,700+ No Failure
average
.. k = 00216
** (a): Failure lni tiated at edge of weld reinforcement 0
(b): Failure initiated at the center of the weld ..
F100,OOO
5101
51",06
54~_l
55,,4
54 .. 5
*** Specimen has been submitted previously to lower stress cycle.
F 2,000,000
29·1
2504
2508
34.0
28.0 -28·5
Voi \).I
TABLE 11 (Refer to Fig. 22)
RESULTS OF FATIGUE TESTS OF TRANSVERSE BUTI' WELDS IN THE AS-WELDED CONDITION
Specimen No"
GLB-35
37
33***
25
27
31
29
33
Stress. Cycle kei.
+41.5 to +83.0
+41 .. 5 to 83 .. 0
+30.0 to +60 .. 0
+30 .. 0 to +60.0
+26 .. 0 to +52 .. 0
+24 .. 0 to +48,,0~
+24 .. 0 to +48 .. 0
+22 .. 0 to 44 .. 0
(PUlSATING TENSION, AXIAL LOADING)
Life
169,700
208,600
·394,700
-589,900
1),144,000
1,187,800
1),583,400
2,758,400+
Loca.tion of Fracture**
a.
c
a.
c
a
a
a.
No failure
* Computed Fatigue Strengths, ksi
F 100,000 F 2,000,000
95.1
100 .. 3
85 .. 4 39.6
45.7
45 .. 0
42 .. 0
average 9386
45 .. 0
44 .. 0
43 .. 5
* Maximum pulsating strength k.:::O" 256
** (a) Failure initi.ated at edge of weld reinforcement .. (c) Two cra.cks initiated at different edges of weld reinforcement, one causing failure.
*** Specimen has been submitted previously to lower stress cycle ..
\.N ~
TA.BLE 12 (Refer to rige 24)
RE'SOI1rS OF FATIGUE TmrS OF TRANSVKRSE BUTT WELDS IN THE AS-WELDED CONDITION
Specimen :Stress Cycle No .. + ksi
*** 36 .. 0 GLB-53 GLB~43 *** 3600
GLB-15 36 .. 0
GLB-l1 36 .. 0 *** GLB-19 3600
GLB-13 3100
GLB-~-l 2000
GLB-23 2000
GLB-21 20 .. 0
GLB-43 16,,0
GLB-53 1600
GLB-19 2000
* k = 0 .. 271
-(COMPLETE REVERSAL, AXIAL LOADING)
Life
49,400
101,300
150,600
153,000
160;~OO
174,300
612,200
672,500
1,126,600
2,027,700+
2;535,800+
2,579,300+
Location ** of' Fracture
a
a
a
a
a
a
a
a
a
No Failure
No Failure
No Failure
* COmputed Fatigue Strengths,+ksi F F
100,000 2,000,000
2908
36.1
4002
4003
4008
3600
average 37 c 2.
1406
1409
17·4 16 .. 0
16 .. 0
20 .. 0 --16 .. 5
** (a): Failwr:e 1m tiated at edge of weld reinforcement.
*** Specimen has been submitted previously to lower stress cycleo
\)J \J1
TABLE 13 (Refer to Figo 25)
RESULTS OF FATIGUE TESTS OF TRANSVERSE BUTI' WELDS IN THE AS-WELDED CONDITION
(p ARTIAL REVERSAL, AXIAL LOADING)
Specimen Stress Cycle Life No. ksi
GLB-49 -17 to +51 83,600
GLB-55 -17 to +51 93,300
GLB-57 -17 to +51 111,000
* Maximum alternating stress k=O.244
(
Loc atlon*~f Computed Fatigue Fracture F 100,00
a 48.8
a 50·2
c 52.,
average 50.4
** (a) Failure initiated at ~dge of weld reinforcement.
Strengths, * ksi
F 2,000,000
24·3 ***
(c) Two cracks initiated at different edges of weld reinforcement, one causing failure
*** Value extrapolated from 100,000 cycles.
TABLE 14 (Refer to Fig. 26)
RESULTS OF FATIGUE TESTS OF TRANSVERSE BUTr WELDS IN THE AS-WELDED CONDITION
(PULSATING TENSION, AXIAL LOADING) ------------------------------~----.-.. ,,'>--Specimen
No. Stress Cycle
ksi Life Location of
*.* Fracture
~* Computed Fatigue Strengths, ksi
GLB-45 +18 to +72 75,700
GLB-59 +18 to +72 86,500
* Maximum pulsating strength k=O.236
F lOO,OOO
a 67.5
b 69·6 average 68·5
** (a) Failure initiated at edge of weld reinforcement. (b)' Failure initiated at center of weld.
*** Value extrapolated from 100,00 cycles.
F 2,000,000
33· 8***
09-&1~ GlB-59 eg·&1~ GlB-!57 gg= 9"'1 9 Gl~~5S
- C}' .." • , . I (..) ~ ~ ~ ....II
~ ~ (!) (!)
9-S1~ 9-S'19 l.S-81£) gg-81e £9-819 19-819
w ~ s U 'J:
CL .......
~ t;> -Bto
...... Z Z W
~ a: « rt) a.. z
.".
(f) I 'z I-w F? 9 ~
I.-S19 g-S1£) a9-8'£) 99-819 tog-En£) ~9= 81f)
~ w U t-W « a.. -.J (f) a..
GlB-74 ~l·819 GlB-72 tl.-Ene GlB-70 89-S'9
\
LL "" 0 0
0
~ P.' 8 N
z « 0 0:: ...... ...... « X U 8
0 <t i
....J Z
W U')
~ '" I all CD ..J ..J (!) (!)
i'
~
GlQ-S I.-m~ GLQ-S g-~ng ! I
"-1
o N
11-819
(:!·&1a
11-819 .,8119
t-S19
Gle-B7
01-&19
8·11e Z·119
1-819
91-&19 G18-15 .,H1I19
8·919 S-in9
l.ql19 ;-119 9-V1e
""S19 Z-S19
t c S19 i .. S1a £c'f19
Gle-13 GlB-IS \
w ~ ~ U
...J :r: a.. f--
Z
a: « a.. z .,.
c
'0 en N I
Q: Z I-W
~ ~ 01
U c W 01 ... a..
§;i-Y19 .,eY19 u U) Q) II..
0 u.. .,. 0 0
0
ti u 0 ...J
Z-Y19 1-'f19
:z i
:b> I
X -i ::0 1'>
0 0 .. 1) r :t:~
---t rn
I ()J ..
" G)
(JJ
r-0 0
~ 0 Z
0
" (J)
lJ m ()
'?
Z (f)
z 'U J> :::u rr1 Z -i
--1\'""0 r :r: 1> () --t ~ fT1
0
I~ 1;:iO
r
GlB·21 GlB·23 Gl&-25 Gll-l7 Gll-l9 GlS .. 31 Gll-33
, I GLIi~22 G18~24 GlS-26 GLO-28 GLB-30 G18 .. 32 . GlB-34
G,lB-l5 GLO-37 Gl8 0 39 GLe .. 41 Gl8-43 GL8-45 Gle .. 41
I
I I
-=~'"-,~ fl'",~,'~"'""",,, ___ n"~'" •
!
GLS .. 36 GLl a 38 Gle~&f,O GLB-42 GLB-44 GLB-46 GLS-48
,goa1,,1 c;;lli-!i~ I
6f1>-a1g GL.B-!)O ~jte~EJ,g _ GLS·54 GLII-20 I L -6. O'-IO'~- c =.5'-O:~--L!~~qj
~
Ii)
f\>_
o o
o Ii
::
o .-
0::
-I
(J)
--.J c.::>
-q--------------~~
0 60
5 5 5" 1"
4 2," 6"
6" sQ' 6"
2 ;3" 6"
3 7"
6 4" 2,"
l 2481
Transverse Weld
Root openino tin. Arrows indicate direction of welding
X indicates change Qf electrode
Pass Electrode size, in. Current, amps. Rate of travel, in.lmin.
5 1
I 32 1 130 5.0
2 J
I 140 8.0 ~ ii"
3 ~ I 230 8.0 ----- ---~ ..... ,-... ---.• ,.'------
4 J 220 7.0 Ii
5 :3 210 1.0 Ti
6 J 210 7.0 iT
Volfaoe : 21 Volts
Polarity : D.C. Reversed
Electrode : E 11018
Temperature o
200 F (Maximum)
Ail in flat position
Underside of pass I with air orc before pass 2.
FIG. 5 IVE ELDING PROCEDU E (Transverse Buti sid)
\
SOCk - goug ad with
and polished with
grinding wheel.
FIG. 6 RESU S
air arc,
F
0- 1
- F X ,i S PEe I 0
311
4
3 5 Root opening 3i to 32 in.
5 51" 6'·
:3 6" 6"
6" 6"
2 3" 6" 6"
1" 6" 6·' 4~--~---*--~--~--~--~--~-~
~ 6" ~ 6~~~~~~~~~~~~~~~~
I · 24" T ra nsyerse weld J Arrows indicate direction of welding
X indicates chon ge of electrode
Pass EI IS cf rode size, in. Current \I amps. Ra fe of travel I! in.lmin.
I 5 150 5,5 fi
--
2 5 140 7.5 fi
3 3 230 8,0 fi
4 3-
220 7.0 Ii
5 3 210 6.5 nr
6 3
210 5.5 Ti
Voltage 21 Volts
Polarity D. C. Reversed
Preheat 2000
F
Elecirode: E 11018
I n t e r pas s T e m per a t u r e' 250° F ( M a x i mum)
All welding in flat position
Undersi de of pass I back- with air arc before pass 2.
FIG. 7 ELDING P CEDURE P 100-110'8 C
(Transverse Butt Welds)
0 60
II 6" 6iB
," 10
6 2 6" 6"
9 6" flo I"
42'
I"
8 Sll 6 1
• 1"2 18
•
2- 6" 7 2
6 1°8
5 488 6'" S" '2,UI
4 6" Sl. 4M
:3 7 00 S··
Reet lOp ening ~ in. 2
e·' 6 1• 3"
688 s"
Arrows indicate direction of welding
L X Indicates change of electrede 2411 Transverse Weld ~l EI sRle, in. , amp$. Rote in.
----5 140 5.5 32
2 :5 230 6.5 16"
3,4 .l 220 6.0 16
5-1l :5 220 8.0 Ii" -----
Voltage 21 Volts
Polarity D.C. Reversed
0 Pr eneat 200 F
Electr E 11018
interpass perature 25 F (Maltoimum)
AU welding in
Unders! of with ClIfC
FI 8 NG PIOO-1I0IS 8 ( Transverse B Ids)
5 511 6" 411
:3 2'1 6" i'
311
'31i t
4" 2 611 .
4 611
6 411 6" 5
11
Of opening ~ to i in. Arrows in oir ection of wei dl in~
Pass EI e ct 1"0 de Size, in. Current D amps. Rote of trovel D in/min.
I ~ 150 5.5 fi
2 5. I 1.5 Sf
:3 3 230 8.0 fi
4 3 220 7.0 Ii
5*6 3 210 5.0 to 7.0 iT
Voltage Volts
POICHlty D. C. Rev
Preheat 2000
F
Elec1rode: E 11018
I nterpass Temperature 2 F (Maximum)
AU welding in flat position
Undersi de of pass I bock = with air arc before ~HJSI 2.
FIG. ELDIN p CEDU E P 100-110 o (Transverse Butt Welds)
FIG. 10 RESULTS OF FACE BEND TESTS PROCEDURE PIOO-HOI8D
G. II RESULTS ROOT BEND TESTS PROCEDURE PIOO-IIOIS D
FIG. 12 RESULTS OF FREE BEND TESTS PROCEDURE PIOQ-IIOI8D
FIG. 13 FRACTURES OF TENSILE TEST SPECIMENS PROCEDURE P 100-11018 D
IdB
0@G)(±)
--~~~~~----------------- ---~.~--.~
e008
i-/iIi Dia . Holes I _------------ ,------ ~;tll
::J ----r4
(0) Plain PI e - pe a
@(f)@0
_~~!lI+-mG
(±) G (£l e
(b) rans
FIG.
eee0 Q)G-.~
GGGG
See Fig. 1 for details
e Butt eld- pe b
F S CI s
I 8"
I'"'"II! 2'-0 lI!>1~ 21-0 ,go I
------ ------------7
11 Continuous Pass lO"
-------- ------ ---------- / ----Section Removed by Sawing I
.--------2 ; r NOTE; Holes Drilled After Welds Are In Place
FIG. 15 u ELDE JOINT BEFORE FINAL ACHI ING
pun Heads
o 0
o 0 Lever
Specimen
FIG. I ILLINOIS· FATIGUE TESTING MACHINE AS USED
AXIA LOADING OF 'WELDED JOINTS.
Dynamometer
( a ) Specimen GLS-I
(b) GLS-3
OF STATIC TESTS L..L..II.,L..1I..iI' JOINT SPECIMENS
o Q
I
I
I
I
I
1 {
I
I f I 1
o CO
I I I
J rt.J ~I !!/
I I
I
t I I Ii
o <D
I.
~
o <::t
o N
Q
I
I
I
t
J I I I I
---
§ <f)
c o o <::t
0 0 0 (\j
0 0 Q
0 0 CO
52 (£5
0 0 <:t
0 0 (\j
lX -
~
_0 tD
-
_0 <::t
-
0 C\.I
8
en en "'C <t c c en :z :::7 0 .c I- m c - W '" Q)
t... :::7 (J) ....J c <t
LL
0 I-
m (I)
0 ::>.
U
FIG. 19 TYPICAL FAILURE OF PLAIN PLATE FATIGUE SPECIMEN
100
80
60 .............. r---, J. I""" ..... - ~ f:- r--.
-------...., 40
lIP
20 t.J)
tn W Q.,. -(f.)
E ::J
E ~
Cl 10 ~
l-
:--:: ~-~ 8
6
I I I I I I R
20 40 60 80 100 5.
RESU F
S- EL o CON
~1(=O.216 --- ---- --0
200
Cycles To
E TESTS
ION.
------r--~ r-- l"'-
t-'
400 60
lure~ In Thousands
NSVE Sf
ENSION.
!
r-:t--. 0-..
~ :-----p..... ~
I o I Original Specimen Failure
-¢-I Retest Specimen Failure
0--1 Original Specimen-No Failure
+1 Retest Specimen-No Failure I - 6000
T L E
FIG. I ACTU E OF BUT ELDED JOiNT SPECI ENS
.i
"" qj) €I) w ...
en ~ c
-,;:: c
100
80
60 )
40
I 00 20 ::ll
Q"
E ::J
E )(
o ~ I
~
f-T-t T :: I- ..,q &
l-
I '20
.
FIG 2
S i I 40
RES s-
r--... r--.
~ LT I i I I
60 80 !)Q
OF c
----.... ~ 0
K=0256:7: ----r---) ~ "..,
1 r-:: r-- r-- r--- --.D 0 ___
~ ~ ---
0 Original Specimen Failure
-<?- Retest Specimen Failure
/
-- .- --..-. . 200 400 61 - 000
Cycles To Failure, In Thousands
IG E ESTS OF TRANSVERSE UTT WELDS THE IO.v. PULSATI G SION, XI L
FIG. 23 FRACTURE OF SPECIMENS GLB-25, GLB-35
(I)
~
+, (I) (I)
(V, a--CJ)
0\ tt:: -o c: b
ID +-
<l(
g E ~
o ~
100
so
60
40
20 I
I
F-
':-~ ,"F-
I 0
'20
FIG~
............ ~ r-...
-¢ r-.-. r-. D¢--1 ~
~~ ~
!---.. .............. r---... -..... """- - ..... ..... ~
....,
~ ~ -- -
=:- 0 Original Specimen Failure
f:1 -<>- Retest Specimen Failure 0-- Original Specimen-No Failure
I I I I I I
40 60 sb 100 200 400 6e -;::
000
Cycles To Failure, In Thousands
RESULTS OF FATIGUE ESTS OF TR ANSVERS E BUT T ELDS IN THE AS- OED CON ON. CO PLE R ERSAL, A L LOADING.
Ul! .x
fl
'&d) en ~ ".t-ID
100
80 I
60
40
I D' 20 b -101 C .... CD :; <!
E ~
E )(
10 ~
I
-
-T .1:: _ ~ 'lIP?>
--"
I
20
FIG.
~ 1
40
"'I--~ ~
""""" n. ----.........,
------r-- ...... --...,;
------1"-""",,-r--. L K =O.244 ........... ~
""-~ r...... .............
~ -kC-T i'"
~ iii>'
I I I I '. 60 8t> 1)0 200 400 6~- -
Cycles To Failure, In Thousands
SULTS OF FATIGUE TESTS OF SVERSE
RSAL,
T EL C
~ i"-o
000
S E
100
80 ...... ~ ~ ......cl..
60
iii oX
40 ff) w (I)
.::: en
0' s: -01 VI 20 ~
Q.
E .;:J
E ;( 0
~ 10 ....
8
6
:~T:: ~~ ~ 5 I I I I i i I '~O 40 60 80
FI u s-
------
1)0
IG
I
-~ .. ~
t-- ....... ........ ----.,.
r---~-r-- r--..... r--- r-- .....
200 400 60-
Cycles To Failure ll In Thousands
E F
SI
K = 0.236
-------.
~
,
~
LOS I
L LO
- .......
T E
G.
000
5
II FE
FIG. 27 TYPICAL FAILURES OF BUTT-WELDED SPECIMENS
.
"
J
I J
I !
f I
-tit ,
o 0 Q co
J
I ~I
I I
I I I
II ~
II I
I I
o <.0
. I
j
I
~ I I I I I I
I I
V
o ~
m
/ I /
o N
!
--0 en -o~ ID :0:0 ,0 ,01 I- ;-'-' .......
00° _N
0 LA:
C!!» ... 0 -C
0 IJi'j)
Q.. <!
iii Q 00
§ to
0 0 0 'It
0 0 0 N
0 0 2
0 0 co
Q US
0 0 v
0
° N
Q
It: -o -W _
~ -_0
<.0
0 b" _
• 0 _0 GJI ~
-Ih U) Q) \I,., -en
J 0 N
100
~
0
.J:. 1;:2 80 c: CD Q)3; ... .... en ..... ....
::J
~m 0» :;:0 0 ...... lL c: CD
::J Q
c: .~ CD .... +-(,) 0 ::J-"'Co.. CD O:::c
Q)~ ~Q... t7 ... II:: Q) (,)
'-CD
0...
60
40
20
I (
20
FIG. 29
45.3 %
~ ~
31.9 010 L---~ F""""'"
!---I----~ ~
I--~ I--.'--- -
Stress Cycle: Zero -To-Tension
I I I - -40 -60 80 100 200 400 ,...,....,.... "' ..... " I"',......" -- -- 00
Cycles To Failure, In Thousands
REDUC 10 IN IGUE STRENGTH OF AS-ROLLED PLAIN
ES U S E S
CJ)
~
U) CJ) Q> ~ -(J)
E ::l r-b::
y. L1
~
100
80
60
40
20 I
I
.... - : l-
I-
I
20
FIG. 30
r- r--... -........ r---..... ... I--- r=-------1"--,- -t--___..... --.... r---... "'D- r-- """'-"r--. r--r- --r- _____ ........... r-..... I"--- r-t----- r--. ........
--- I--- I""-- ............ --.......:;". --... :::--- --- -r--- -........
~ ----- ............ ~ ... r---. ... r---- r--. r--- .... 1 T-T i"'--. --- --~ --~ ------- 2
~ ---.. r-- I----r-- r--. r--r- ------~ ............ r....... r-----.... I
-------r-r------. ---I----- aT- T
~ F-- ____ O-T :---r--- r-----..... t-- ~C-T r--r- "'" r-----
~ I r------ C-T
~ :: I 1 I I I I
40 60 ab IX> 200 400 6t - ~
000
Cycles To Failure, In Thousands
SUMMARY OF RESULTS- FATIGUE ESTS OF TRANSVERSE BUT
WELDS I T E S- ELDED CONDITIO -- AXIA LOADING.
