3SS.

1
Low Cycle Fatigue Behavior
Of Joined Structures
FATIGUE CRACK
PROPAGATION IN
AS14 STEEL
by
R. WI Hertzberg
HR Nordberg
Fritz Engineering La boratory Report No, 358.7
i
1
Low-Cycle Fatigue
FATIGUE CRACK PROPAGATION IN
A514 STEEL
by
Richard Hertzberg
Hans Nordberg
This research was sponsored by the Office of Naval Research,
Department of Defense, under Contract N0014-68-A-0514;
NR064-509. Reproduction in whole or part is permitted for
any purpose of the united States Government.
Department of Metallurgy and Material Science
Lehigh University
Bethlehem, Pennsylvania
November, 1969
Fritz Laboratory Report No. 358.7
358.7
TABLE OF CONTENTS
Page
SYNOPSIS i
1. INTRODUCTION 1
2. EXPERIMENTAL PROCEDURES 7
1. Weld Preparation 7
2. Specimen Preparation 7
3. Experimental Testing 8
4. Fractographic Techniques 9
3. EXPERIMENTAL RESULTS AND DISCUSSION 10
1. Effect of Specimen Thickness on 10
Fatigue Crack Propagation
2. Effect of Mean Stress on Fatigue 11
Crack Propagation
3. Effect of Specimen Orientation 12
on Fatigue Crack Propagation
4. Effect of Weldments on Fatigue 14
Crack Propagation
5. Effect of Environment on Fatigue 19
Crack Propagation
6. Fractographic Observations 20
4. SUMMARY 21
5. NOMENCLATURE 23
358.7
6. ACKNOWLEDGEMENTS
7. TABLES AND FIGURES
8. REFERENCES
Page
24
26
35
358.7 -i
SYNOPSIS
This paper presents one phase of a major research
program designed to provide information on the behavior and
design of joined structures under low-cycle fatigue. This
paper reports on a detailed study of the fatigue crack
propagation process in the A5l4 steel base plate, heat
affected zone and weld metal. These data are evaluated in
terms of fracture mechanics concepts.
In the early studies of this phase, several
variables of potential importance in the fatigue process
were examined. It was observed that fatigue crack
propagation in A514 steel is strongly dependent upon the
stress intensity factor range and only moderately dependent
upon the maximum stress intensity level over a range of A =
. Krnax
1.0 - 4.0 where A = The growth rate is independent
of specimen thickness (up to one inch). Fatigue crack
propagation rates were determined in A5l4F and A5l4J steel
in the RW and WR orientations. No directionality in fatigue
was noted in A514F while A5l4J did exhibit some-
what faster crack growth in the WR direction as compared to
that observed in the RW direction. The above results are
consistent with data previously published for other materials.-
358.7 -ii
A particularly striking effect was noted in
crack propagation studies of A514F steel containing
weldments. Fatigue crack propagation rates in transverse
weld metal and the heat affected zone were substantially
less than in the base metal. For example, at a stress
intensity level of 40,000 psi in., the fatigue crack
propagation rate in the weld metal was almost an order of
magnitude less than that displayed by the base plate. Growth
rates in the three metallurgically distinct regions tended
to converge at large stress intensity levels. After a heat
treatment of 110QoF for one hour, the crack growth rates in
the HAZ and weld metal increased significantly and were then
comparable to the growth rate in the base plate. Stated
differently, the dependence of crack growth rate on the
stress intensity level in the weld metal and HAZ decreased
as a result of the heat treatment. Heat treatment of the
base plate did not affect fatigue behavior. Similar results
were found in fatigue studies of longitudinal welds though
the magnitude of growth rate differences was considerably
less than was observed in the transverse welds. Again, heat
treatment eliminated the major differences in crack growth
in the different weld regions. These results are discussed
with reference to residual stress.
358.7 -1
1. I N T ~ O D U C T I O N
When formulating the experimental program for a
study of the fatigue behavior of a material, it is
important at the outset to establish the relative importance
of the initiation and crack propagation stages in the fatigue
life of a component. It has been well established that in
many engineering applications where metallurgical and/or
manufacturing flaws are present, the crack propagation stage
is dominant in the fatigue process. This is particularly
true under conditions of high stress resulting in a relatively
small number of cycles to failure. Consequently, the main
thrust of the following remarks will deal with an evaluation
of the fatigue crack propagation behavior of ASTM type A514
steel and the effect of welding upon the fatigue crack
propagation process.
While most of the earlier studies of fatigue
crack propagation were performed with aluminum alloys,
thereby eliminating a direct comparison of these results
with the present investigation, these earlier results did
isolate the main variables affecting fatigue crack
propagation. Most of the earlier crack propagation laws
358.7 -2
described the crack growth rate as being related to some
function of applied stress and crack length. (1-7) . In 1963
Paris and Erdogan(8) presented a comprehensive compilation
of fatigue crack propagation data and showed that the
fatigue crack growth rate was strongly dependent upon
changes in the stress intensity factor range. Hence, the
concepts of fracture mechanics which have contributed much
to the understanding of static fracture can also be used in
evaluating the response of a material under cyclic loading.
Compiling the results of several investigators who looked
at the response of both ferrous and non-ferrous materials,
Paris and Erdogan(8) showed that the fatigue crack growth
rate could be described by the following relationship,
~ ~ = C ~ K n , where a is crack length, ~ K is stress intensity
factor range C, n are material constants. In their earlier
studies they found the exponent "n" to be equal to 4. It
is noted that the fatigue crack propagation rate is
primarily a function of the stress intensity factor range
1
with all other variables being combined in the constant
term. For a more accurate evaluation of the fatigue response
it is obviously necessary to incorporate in any given formula.
as many of the variables as possible. In attempting to
arrive at a more comprehensive relationship, Forman, Kierney
and Engle(9) h a ~ proposed a relationship of the following
form:
358.7
da AmliK
n
dn
=
-XAK K
c
where A
Kmax
==
LrK
Krnax = maximum stress intensity factor
K
=
fracture toughness
c
m,n
=
material constants
-3
It is to be noted that this relationship includes the stress
intensity factor range, 6K, a measure of the maximum stress
intensity factor,A= ~ ~ x , and an evaluation of the material
fracture toughness in the previous relationship incorporates
the role of metallurgical factors. At low growth rates
associated with low stress intensity conditions it has been
found that metallurgical factors exercise only a secondary
effect on fatigue crack propagation. This observation was
found by weber(lO) for the case of austenitic stainless
steel and 70-30 type brass, by Heiser (11) for the case of
banded steel and by Miller(12) for the case of 4340 steel.
However at high stress intensity factor values associated
with much higher growth rates, the metallurgical structure
plays an important role in the fatigue process as shown
by Miller(12) and Heiser. (11) It has been argued that
the increasing importance of metallurgical effects at higher
stress intensity factors is associated with the increasing
358.7 -4
amount of unstable crack extension caused by locallized
fracture of inclusions. It is to be noted that the rapid
increase in fatigue crack propagation rates at high stress
intensity levels can be predicted by the Forman et ale
relationship under conditions where the maximum applied
stress intensity factor approaches the material fracture
toughness level.
As mentioned above, the most important variable
in the fatigue crack propagation behavior of a material
has been shown to be the stress intensity factor range.
The influence of mean stress on fatigue crack has also been
investigated. Frost(13) and Yokabori et al. (14) found
that the effect of mean stress upon fatigue crack propagation
in steel alloys was of second order importance. However,
since these tests were performed at relatively low stress
levels, they do not assess the influence of high mean stress
levels which approach the fracture toughness level of the
material. It is to be expected that the mean stress level
should playa more important role under these conditions.
The effect of environment on fatigue crack growth
has been the subject of recent reviews by Achter,(15)
Johnson and Paris, (16) and Wei. (17) For the case of a
358.7
material sensitive to environment, stress corrosion
cracking can play an important role ~ n fatigue at
-5
intermediate stress levels where K is larger than
max
K and where the crack propagation time is sUfficiently
I
sec
long.
In a recent review article, McMillan and
Hertzberg(18) summarized the important role of electron
fractographic observations in fatigue studies. They showed
that fatigue striations found on fatigue fracture surfaces
are a source of considerable quantitative information.
While extensive use of fractography has been made in studies
of non-ferrous alloy fatigue behavior, it is unfortunately
far more difficult to develop similar data for ferrous
materials since fatigue striations are not nearly as easily
resolved on steel fracture surfaces as they are in aluminum
alloys.
The objective of this program is to evaluate the
fatigue response of welded structures. Since welded
components almost always contain metallurgical flaws
resulting from the welding process, and since the welded
structures of interest are subject to relatively few cycles
at high stress levels, the point of view taken in this
358.7
investigation will be that the fatigue crack propagation
stage is dominant.
-6
The fatigue crack propagation behavior of type
A514 steel will be examined as a function of specimen
thickness, mean stress at both low and high crack growth
rates, specimen orientation with respect to the rolling
direction, and in the presence of weldments. In the latter
case, fatigue crack propagation studies will be performed
in welded panels where the fatigue crack will be made to
grow both across the weld (longitudinal weld) and along
both of the weld metal and the heat affected zone,
(transverse welds).
358.7 -7
2. EXPERIMENTAL PROCEDURES
2.1 Weld Preparation
Four welded plates were prepared from 3/4 inch
thick plates of T-l steel. All weldments were made by a
Metal Inert Gas process using 60 degree joints. The filler
metal used was AIRCO AX-90. Longitudinal plates having
dimensions of 6" x 18" x 3/4" with the weld running
parallel to the long direction were welded with five passes
of the filler metal. The heat input for these passes
~ a n g e d from 35 to 42 Kj/inch. Transverse plates having
dimensions of 18" x 14" x 3/4" were welded in the long
direction in eight passes with heat inputs varying from
22 to 56 Kj/inch. During the welding process all weld
passes were allowed to cool in air approximately 1 1/2
hours before the next pass was made.
2.2 Specimen Preparation
The fatigue crack propagat,ion studies were performed
on single edge notch specimens 12
11
x 3
11
X 1/8" with 3/4"
loading pin holes, Fig. 1. Specimens .061, and .266 inches
thick were also prepared for determination of thickness effects.
358.7
-8
Notches were introduced to the specimen with a machine saw
cut followed by a jeweler's saw cut., The total length of
the starting crack was approximately 3/8".
Specimens containing weldments were prepared
from panels obtained by slicing the original plate in half
along the mid-thickness plane. For the studies involving
crack propagation entirely in one region (for example
along the weld metal on HAZ) specimens were prepared from
the transverse plates such that the weld was perpendicular
to the long direction of the specimen. For longitudinal
samples where a crack propagates across different metallurgical
structure, specimens were prepared from the longitudinal
plates with the weld parallel to the loading direction.
2.3 Experimental Testing
~ a t i g u e crack propagation tests were performed on
an MTS machine at a frequency of lOcps. With the aid of a
travelling microscope, the instantaneous crack length was
recorded with respect to the corresponding number of cycles
applied. These data along with ,applied loads and dimensional
data were subsequently analyzed with the aid of a CDC 6400
computer.
358.7 -9
2.4 Fractographic Techniques
Electron fractographic techniques were applied
to the fatigue fracture surfaces. Samples suitable for
viewing on the electron microscope were prepared by the
two stage replication technique. Fractured surfaces were
cleaned in acetone and then firmly covered with an acetone
softened strip of cellulose acetate. After drying, the
tape was stripped off, placed on a slide face up, and
shadowed in a vacuum using platinum carbon pellets as the
shadowing material. A layer of carbon, 400-500 Angstroms
thick, was then d e p o s i t ~ d upon the shadowed replica. These
samples were then cut into approximately 1/8" squares and
placed in an acetone bath where the cellulose acetate
replica was allowed to dissolve. The carbon films with the
shadowing material were mounted on 200 mesh grids for
examination in the electron microscope.
Viewing of the replicas was done with an RCA
EMU-3G electron microscope. Photomicrographs were taken
of various specimens at magnifications ranging from 5,400X
to 35,OOOX.
358.7 -10
3. EXPERIMENTAL RESULTS AND'DISCUSSION
In the initial studies of this phase of the
overall program, a pilot study was initiated to examine
several variables that might affect the fatigue crack
propagation behavior ofA514 steel with and without welds.
It will be shown in the following comments that while some
variables are of only mild importance other factors
contribute significantly to the fatigue behavior of the
material.
3.1 Effect of Specimen Thickness on Fatigue Crack propagation
To examine the effect of specimen thickness on
fatigue crack propagation in type A514J alloy steel, fatigue
studies were performed on panels of 0.061, 0.126, and 0.266
in. thickness. In all cases the crack was oriented normal
to the rolling direction of the panel. As shown in Fig. 2,
there was no observed effect of thickness on crack
propagation rate. One may conclude from this observation
that the crack propagation rate was independent of any
I '
metallurgical variation associated with the specimen
thickness and/or that the fatigue crack propagation rate
was independent of the relative size of the plastic zone
358.7
with respect to the sheet thickness for the range of
thickness examined.
with the crack oriented parallel to the rolling
direction of the panel, fatigue crack propagation rate in
1/8" thick type A514F steel was similar to that observed
by MCHenry(19) in 111 thick panels. Therefore, in a range
of .06 to 111 thick panels there was no observed effect of
thickness on fatigue crack propagation in A514 steel.
3.2 Effect of Mean Stress on Fatigue Crack Propagation
-11
To evaluate the effect of mean stress on fatigue
crack propagation in type A5l4J the fatigue crack propagation
rates were determined as a function of the A value, a
parameter defined as the ratio of the maximum stress
intensity factor to the stress intensity factor range. The
data given in Fig. 3 show the material response as a function
of A over a range of 1.07 to 4.3. It was found that the
data could best be portrayed by two straight line segments
in a log stress intensity factor range -- log growth rate
plot. It is not clear whether such a discrete discontinuity
should exist based upon either continuum or microscope
considerations. Nevertheless, it is possible to treat the
growth rate data from each linear segment as
358.7 -12
where C, ro, n = material constants.
In the linear segment at lower growth rates
the estimated values of m and n are 0.4 and 2.4, respectively,
while in the upper segment of the data the m and n values
are higher. These exponent values reveal that while fatigue
crack growth rates are very sensitive to 8K there is less
dependence on K
max
or K
mean
For example, a four-fold
increase in the stress intensity factor range can effect a
thirty-fold increase in the fatigue crack propagation rate
whereas a four-fold increase in the ~ value gives rise to
less than a factor of 2 increase in the fatigue crack
propagation rate. The greater dependence of K on growth
max
rate at high K levels, may be related to a greater degree
of local cracking of inclusions during the fatigue process.
Consequently the fatigue process at high stress intensity
levels may consist of a summation of a pure fatigue process
combined with local static fracture associated with broken
inclusion particles.
3.3 Effect of Specimen Orientation on Fatigue Crack
Propagation
To determine the extent of metallurgical
358.7
-13
anisotropy upon fatigue crack propagation, panels were
prepared such that fatigue. cracks were grown both parallel
and perpendicular to the sheet rolling direction. In Fig.
4 it is seen that no directionality effect was observed in
A5l4F material. The fatigue crack propagation rates were
identical with the crack traversing both parallel and
perpendicular to the rolling direction. In A514J material,
it was observed that fatigue crack propagation was faster
when the crack traversed parallel to the rolling direction.
This effect was negligible at low stress intensity levels
but did produce higher crack growth rates at the higher
stress intensity levels. For example, at a stress intensity
range of 50,000 psi in the fatigue crack propagated twice
as fast when running parallel to the rolling direction than
when it was oriented perpendicular to the rolling direction.
As mentioned in the previous section, local fracture of
oriented inclusions could have given rise to the faster
growth rate when the crack was oriented parallel to the
rolling direction. Since multiple tests were not conducted
to evaluate the role of crack orientation with respect to
the rolling direction, it is not possible to conclude that
the A5l4F materials was definitely superior in transverse
behavior to that in the A5l4J material. More testing would
be necessary to'validate this point. In any case, the factor
358.7
of two different in growth rate observed at high K levels
in this study is not considered to be a major effect and
suggests the metallurgical effect of crack plane to ~ e
of second order importance.
3.4 Effect of Weldments on Fatigue Crack Propagation
-14
Having determined the role of specimen thickness,
mean stress, and specimen orientation on fatigue ,crack
propagation of the base metal, it was possible to evaluate
the effect of weldments on fatigue crack propagation in
theA5l4F material. Tests were conducted with both
longitudinal and transverse welds.
Crack propagation rates in the heat-affected
zone and weld metal in transverse welds were considerably
different than that displayed by the base metal. In Fig.
5, it is seen that fatigue crack propagation in the weld
metal was substantially less than in the base metal. This
effect was most pronounced at low stress intensity levels;
for example, at a stress intensity level of 40,000 psi
in the fatigue crack propagation rate in the weld metal
was almost an order of magnitude less than that displayed
by the base metal. It was observed that the relative
advantage of the weld metal over the base metal tended to
358.7
be eliminated at the higher stress intensity levels.
Stated in another way, the power dependence of the growth
rate on stress intensity range was much higher in the weld
metal where n 4.6 than in the base metal where n 2.4.
-15
As shown in Fig. 6, a similar observation was
found with respect to fatigue crack propagation in the
heat-affected zone as compared to the base metal. Here
again, fatigue crack propagation rates were considerably
lower in the heat-affected zone than in the base metal,
especially at low stress intensity levels. It is possible
that the marked superiority of the heat-affected zone and
weld metal in the transverse specimens was due to some k ~ n d
of favorable residual stress pattern associated with the
welding process and the subsequent machining procedure to
produce the specimens. If, indeed, the marked superiority
of the heat-affected zone and weld metal at low crack growth
rates was due to a residual stress effect, it had to be of
a compressive nature. It is not clear at this time how
such a favorable residual stress pattern could have been
produced. It is important to note that an argument based
on residual stresses supports the observation that the
heat-affected zone and weld ,metal lose their superior fatigue
behavior at high stress intensity levels. This would be
358.7
associated with the residual stress pattern being swamped
by the much higher stress intensity level being applied.
-16
To test the supposition that a favorable residual
stress was present in the weld and heat-affected zones in
transverse specimens, duplicate samples were subjected to
a stress relief heat treatment of one hour at llOQoF. As
a control, a specimen containing only the base metal was
also subjected to a one hour 11000F stress relief heat-
treatment. As shown on Figs. 5 and 6, the heat treatment
did not alter the fatigue response of the base metal. On
the other hand, the marked low crack growth rate superiority
of the weld and heat-affected zones was almost eliminated
by the heat treatment. Since metallurgical changes
resulting from the heat treatment were not likely to alter
the fatigue response of the material to the extent observed,
it may be concluded that the heat treatment did remove a
favorable residual stress pattern in both the weld metal
and the heat-affected zones.
One additional parameter was evaluated with
respect to fatigue crack propagation in transverse welds.
Crack propagation studies were conducted with the crack
propagating both upstream and down stream with respect to
the solidification direction in the welding process. No
358.7
effect on fatigue crack propagation rate was observed for
-17
this variable. It is possible that the multi-pass welding
process eliminated or at least minimized the possibility
of anisotropic effects due to the solidification process.
Since the transverse weld fatigue test results
have led to some suprising conclusions concerning the nature
of. the residual stress patterns in the welded plate, additional
tests are planned to verify the reproductibility of these
findings. In addition, it has been suggested by Coffin(20)
that the observed difference in fatigue behavior between the
base metal HAZ and weld metal could be the result of strain
aging effects. This possibility will also be explored.
For longitudinal welds the response of the various
metallurgical structures in the welded specimen (that is,
the base metal, heat affected zone and weld metal) to cyclic
loading was evaluated by fatigue tests which were conducted
at a constant stress intensity factor range; therefore, as
the crack traversed the three metallurgically different
regions of the specimen it was possible to observe the crack
extension rate under the same stress intensity condition.
Recalling that one-eighth inch test panels were machined from
the top and bottom multiple pass one-inch plate, fatigue tests
358.7 -18
with longitudinal welds were performed with both top and
bottom plates. with a top plate speqimen, at a stress
intensity range of 70,000 psi in the growth rate in the
base metal was approximately 50 percent greater than that
of the weld metal. At 60,000 pis in, the growth rate in
the base metal was about twice that in the weld metal. In
both cases, the crack growth rate in the weld metal was
higher than that observed in transverse welds for comparable
stress intensity levels. As the crack traversed the heat
affected zone, considerable scatter was observed in the
fatigue crack propagation rate at both 60,000 psi in and
70,000 psi in. The results did reveal that fatigue crack
propagation in the heat-affected zone was roughly 50 to 100
percent higher than in the base metal.
To examine whether the slower crack growth rate
in the weld metal and the scatter in test results in the
HAZ were due mainly to a residual stress effect as suggested
in the transverse weld studies, an additional fatigue test
of heat treated material were conducted. At a c o n s t a n t . ~
stress intensity range of 60,000 psi in, the one hour
110QoF heat treatment increased the crack growth rate in
the weld metal from 1.25 - 1.50 ~ 10-
5
in/eye. to 2.50 x
10-
5
in/eye. in the base metal. In addition, the scatter
358.7 -19
of fatigue results in the heat affected zone was essentially
eliminated as a result of the heat Though only
one test of heat treated material with a longitudinal weld
conducted, the results suggest that residual stress
effects may 'have contributed to the fatigue crack
propagation process. For the case of the longitudinal
welds, the residual stresses parallel to the applied stress
direction appear to be of a compressive nature in the weld
metal and tensile in the heat affected zone.
Longitudinal fatigue. tests were also performed
in top and bottom plates at a constant stress intensity
range of 30,000 pis in. While the growth rate in the weld
metal was again lower than in the base metal, the crack
propagation rates in the three regions were observed to be
approximately five times larger than expected on the basis
of previous results (Figs. 5,6). At the present time, it
is not possible to explain this discrepancy. More tests
are being planned to explore this question.
3.5 Effect of Environment on Fatigue Crack Propagation
To determine the possible effect of water,
fatigue crack propagation studies were carried out in both
air and in water in A514J base plate and in AIRCO AX-90
weld metal. In both cases no effect was observed of water
358.7 -20
on fatigue crack propagation (Fig. 7). The relative humidity
in air for the AIRCO AX-gO study was 65 percent which suggest
that this level is equivalent to a saturated condition in
terms of moisture effects on fatigue crack propagation. (17)
Since all fatigue tests were conducted in a relative humidity
range of 35 - 80 percent, the preliminary results suggest
that humidity variations in testing conditions from day to
day probably plays a relatively minor role in the fatigue
crack propagation process.
3.6 Fractographic Observations
An evaluation of fractographic features of the
fatigue fracture surfaces was started. As expected,
relatively few areas of clearly defined fatigue studies
were observed. The spacing between these striations were
recorded and compared with the associated macroscopic
fatigue growth rate. As shown in Table I, there is general
agreement between the two growth rate measurements. More
fractographic studies are planned to establish this
correlation with greater confidence.
358.7 -21
4. SUMMARY
This paper reports on a detailed study of the
fatigue crack propagation process in A514 base plates,
heat-affected zone, and weld metal. Several variables of
potential importance in the fatigue process were examined.
It was observed that fatigue crack propagation in A514 steel
was strongly dependent upon the stress intensity factor
range and only dependent upon the mean stress
intensity level. In addition, the growth rate-was
of specimen thickness (up to one inch). The above results
are consistent with data previously published for other
materials.
A particularly striking effect was noted in crack
propagation studies of A514F steel containing weldments.
Fatigue crack propagation rates in transverse weld metal and
the heat affected zone were substantially less than in the
'base metal. For example, at a stress intensity level of
40,000 psi in., the fatigue crack propagation rate in the
weld metal was almost an order of magnitude less than that
displayed by the base plate. Growth rates in the three
metallurgically distinct regions tended to converge at large
stress intensity levels. After a heat treatment at llOOoF
358.7 -22
for one hour, the crack growth rates in t h ~ HAZ and
weld metal increased significantly and were then
comparable to the growth rate in the base plate. Heat
treatment of the base plate did not affect fatigue behavior.
It was concluded that a favorable residual stress was present
in th- stress direction in both HAZ and weld metal regions
of transverse weldments.
Residual stress effects were also indicated in
fatigue studies of longitudinal welds. The magnitude of
this effect was considerably less than was observed in the
transverse welds. Again, heat treatment eliminated the
major differences in crack growth rate in the different
weld regi9nso
358.7 -23
5 NOMENCLATURE
da
dn
K
AK
Kmax
Kc
C,m,n
crack growth rate, inches/cycle
stress intensity factor, psi in
stress intensity factor range (Kmax-Kmin) psi in
maximum stress intensity factor, psi in.
plane stress fracture toughness, psi in.
ratio of Kmax
L\K
material constants
358.7 -24
6. ACKNOWLEDGEMENTS
This paper presents the results of a detailed
study of the fatigue crack propagation process in A514
steel base plate, heat-affected zone and weld metal. The
investigation is one phase of a major research program
designed to provide information on the behavior and
design of joined structures under low-cycle fatigue.
The investigation was conducted in the Mechanical
Behavior Laboratory of the Materials Research Center and in
the Metallurgy and Materials Science Department, Lehigh.
University, Bethlehem, Pennsylvania. The authors
acknowledge the financial support of the Office of Naval
Research, Department of Defense under contract N 00014-68-
A-514; NR 064-509. The program manager for the overall
research project is Lambert Tall.
The guidance of, and the suggestions from, the
members of the Advisory Committee on Low-Cycle
Fatigue is gratefully acknowledged.
The authors are grateful to Michael Parry for his
-25
contributions to the preparation of this paper. The
secretarial services of Betty Brader and Louise Valkenburg
are sincerely appreciated. George P. Conrad II is
Chairman of the Department of Metallurgy and Materials
Science, and Joseph F. Libsch is Director of the Materials
Research Center, and Vice-President for Research, Lehigh
University.
358.7 -26
7. TABLES AND FIGURES
358.7
TABLE 1
Fatigue Crack Growth Rate Measurements
-27
Specimen L\K Macro Micro Structure
13 30KSi in
-6 -6
Weld Metal 7xlO in/eye 9.3xlO in/eye
30
7xlO-
6 -6
HAZ 9.3xlO
30
8xlO-
6 -6
Base Metal 6.8xlO
60
3xlO-
5 -5
Base Metal 4.3x10
16
19
28.5
38
30.8
39.6
-6
5.6xlO
-6
7.7xlO
9xlO-
6
1.2xlO-
5
Base Metal(Long)
Base Metal(Long)
Base Metal(Trans}
Base Metal(Trans)
'a 358.7
.....-...-----12.0" -------.....
-28
o
0-
-0-
- - , ,- -
o
-0
- 0-
3.0"
-----.........- 6)'-'---........
o
o
..-....-2D'---'~
.-..--- 32';------
2A"
Fig. 1 Goemetry of Single Edge Notch and Compact
Tension Fatigue Specimens. Dashed lines
represent relative position of weld metal.
358.7
, ,.
T I
-29
2 '-
x .265 IN

-
0 .126 II
0
A
,061 "
00
.fiO
10
A
8

0)(
A
-
0
y
(6)(

da
>(
dN

-
(lN/CY)

j
)

6
2

.;
-

.a
A

)(
X
x)(
)('1
10
5
...

-
'l
t'
j(
XA
>A.

1 I I
V
80 100 40 60
6K (KS1VfN)
Fig. 2
Effect of Specimen Thickness Upon Fatigue
Crack Propagation in AS14 Steel.
358.7
-30
5
~
dN
(IN/CY) 2
5
2
~ =2.14
20 50 100
l1K (KSI/iN)
Figa 3 Effect of A (Kmax/AK) Upon Fatigue Crack
Propagation on A514J steel.
358.7
10
3
A514F RW1WR
5
A514 J RW
A514J
2
1(j4
5
da
dN
(IN/CY) 2
10
5
2
-31
1
20 50 100
6K (KSI v'fN)
Fig. 4 Fatigue Crack Propagation in AS14 Steel
in RW and WR Orientations.
358.7
5
10
da
dN
(LN/CY) 2
2
A514F
AIRCO
AX-go
-32
20 50 100
6K (KSIIiN)
Fig. 5 Comparison of F ~ t i g u e Crack Propagation Behavior
Between Transverse Weld Metal (AX-gO) and Base Plate
(h514F) in Heat-treated and Unheat-treated Conditions.
358.7
-33
5
5
1
,
I
I
I
I
I
I
I
I I
2
1
H ~ Z
1100F/1h
10
6
tm.---------......---....----..
sii..
dN
(IN/CY) 2
20 50 100
6 K (KSI./iN)
Fig. 6 Comparison of Fatigue Crack Propagation Behavior
Between HAZ and Base Plate (A514F) in Heat-
treated and Unheat-treated Conditions.
358.7
-34
5
x AIR
o WATER
tJ
2
0
)(
't
AiReD
AX-90 x
)(
,;<
5
cP
0
da A514J WR )(
I
-----
dN
)(
)(
8
(IN/CY) 2
0
f
I
)(
tl
)(
16
5
~ f
0
I
)(
,..0
0
0
5
xl
,
0
R'Io
0
f
2
10
6
20 50 100
6K (KS I v'!N)
Fig. 7 Effect of Water on Fatigue Crack Propagation in
A514J and AIReo AX-90 Alloys.
35807 -35
8 0
Is Head, A. KG
THE GROWTH OF FATIGUE CRACKS,
MagG 44, (1953) 925.
2Q Head, A. K.
THE PROPAGATION OF FATIGUE CRACKS,
Trans ASME, Sera (1956) 407 Q
30 Frost, No and Dugdale, D. S.
THE .PROPAGATION OF FATIGUE CRACKS IN
SHEET SPECIMENS, J. Mech. Phys. Solids
6, (1958) 92.
4. McEvily, A. Jo and Illg, W.
THE RATE OF CRACK PROPAGATION IN TWO
ALUMINUM ALLOYS, NACA TC 4394, (1958)0
5. Frost, N. EG, et ale
EXPERIMENTAL STUDIES INTO THE BEHAVIOR
OF FATIGUE CRACKS, Prac. Crack Prop.
Symp. Cranfield (1961) 166.
6. Liu, Hit W.
FATIGUE CRACK PROPAGATION AND APPLIED
STRESS RANGE - AN ESSAY APPROACH, Trans.
ASME, Sere D 85 (1963) 116.
7fJ Weibull, w.
A THEORY OF FATIGUE CRACK PROPAGATION IN
SHEET SPECIMEN, Acta Meto 11 (1963) 745.
8Q Paris, Pe C. and Erdogan, Fo
A CRITICAL ANALYSIS OF CRACK PROPAGATION
LAWS, Trans. ASME, Sere D 85 (1963) 5280
Forman,Ro Go, Kearney, V. E. and Engle, R. Mo
NUMERICAL ANALYSIS OF CRACK' PROPAGATION
IN CYCLIC-LOADED STRUCTURE, Trans ASME,
Sere D 89 (1967) 459.
100 Weber, Jo H.
EFFECTS OF CRYSTALLOGRAPHS AND THERMO-
MECHANICAL TREATMENT ON FATIGUE CRACK
PROPAGATION, Ph.D. Dissertation, Lehigh
University, 1969.
358.7 -36
11. Heiser, F. A.
ANISOTROPY OF FATIGUE CRACK PROPAGATION IN
HOT ROLLED BONDED STEEL PLATE, Ph.D.
Dissertation, Lehigh University, 1969.
12. Miller, C. A.
THE DEPENDENCE OF FATIGUE CRACK GROWTH RATE
ON STRESS INTENSITY FACTORS AND THE MECHANICAL
PROPERTIES OF SOME HIGH STRENGTH STEELS, ASM
Transactions, Vol. 61, No.3, (1968) 442.
13. Frost, N. E.
EFFECT OF MEAN STRESS ON THE RATE OF GROWTH
OF FATIGUE CRACKS IN SHEET MATERIALS, J. Mech.
Eng. Sci. 4 (1962) 22.
14. Yokobori, T., et ale
FATIGUE CRACK PROPAGATION OF MILD STEEL AND
HIGH STRENGTH STEELS, Rep. Res. Inst. Strength,
Fracture of Materials, Tohoku Univ. 3 (1967) 39.
15. Achter, M. R.
EFFECTS OF ENVIRONMENT ON FATIGUE CRACKS,
ASTM STP, 415 p. 181.
16. Johnson, H. H. and Paris, P. c.
SUB-CRITICAL FLAW GROWTH, J. Eng.
Fract. Mech. Vol. 1, No.1, (1968) 3.
17. Wei, R. P.
SOME ASPECTS OF ENVIRONMENT-ENHANCED FATIGUE
CRACK GROWTH, to be published in Engineering
Fracture Mechanics
18. McMillan, J. C. and Hertzberg, R. W.
APPLICATION OF ELECTRON FRACTOGRAPHY
TO FATIGUE STUDIES, ASTM STP 436, p. 89, (1968).
19. McHenry, H., private communication.
20. Coffin, L., private commurnication.
Securaty Classification
DOCUMENT CONTROL DATA R &0
(SOt'urlty of Hilt', body of abstrltct tflld snnoJt'f/o'" nHlNt hi) tJtlteff1cJ wlul(I tI,O ovorall report Is dIJ!lBI/Jed)
1 ORIGINATING ACTIVITY (Corporate lJuthor) , 20. REPORT SECURITY CLASSIFICATION
3 REPORT TITLE
Lehigh University
Unrestricted
2b. GROUP
FATIGUE CRACK PROPAGATION IN A5l4 STEEL
4. DESC Rt P T I V E NOT E S (Typo of report and Inc/us/ve dates)
'0. AU THO RIS) (Firat neme, middle Inillal, laet name)
Richard Hertzberg and Hans Nordberg
6. REPORT DA TE
November 1969
Sa. CONTRACT OR GRANT NO.
N0014-68-A-0514; NR064-509
h. PROJECT NO. 358
c.
d.
10. DISTRIBUTION .STATEMENT
11, SUPPLEMENTARY NOTES
13. ABSTRACT
74. TOTAL OF PAGES rb. NO. OF
Ott. ORIGINATOR'S REPORT NUM8ER(S)
9b. OTHER REPORT NO(S. (Any other numbers that may ba lIsslgned
thl s report)
12. SPONSORING MILITARY ACTIVITY
This paper reports on a detailed study of the fatigue crack
propagation process in the A5l4 steel base plate, heat affected zone
and weld metal. These data are evaluated in terms of fracture
mechanics concepts.
It was observed that fatigue crack propagation in A5l4 steel is
strongly dependent upon the stress intensity factor range and only
moderately dependent upon the stress intensity level over a
range of A = 1.0 - 4.0 where A = The growth rate is independent
LU<
of specimen thickness (up to one inch). crack propagation
rates were determined in A514F and A514J steel in the RW. and WR
orientations. No directionality in fatigue behavior was noted in A514F
while A5l4J did exhibit somewhat faster crack growth in the WR,
direction as compared to that ovserved in the RW direction.
Fatigue crack propagation rates in transverse weld metal and the
heat affected zone were substantially less than in the base metal. For
example, at a stress intensity level of 40,000 psi in., the fatigue
crack propagation rate in the weld metal was almost an order of magni-
tude less than that displayed by the base plate. Growth rates in the
three metallurgically distinct regions tended to converge at large
stress intensity levels. After a heat treatment of 1100F for one hour,
Security Classification
Security CJRRsificl:ltion
, 4:
LINK A LIN K B LIN K C
KEV WORDS
...
the crack growth rates in the HAZ and weld
metal increased significantly and were
then comparable to the growth rate in the
base plate. stated differently, the
dependence of crack growth rate on the
stress intensity level in the weld metal
and HAZ decreased as a result of the heat
treatment. Heat treatment of the base
plate did not affect fatigue behavior.
ROLE WT ROLE WT ROLE WT
Security Cla8sification