Fatigue of A36 Steel

Laser Beam Weldments

Welds are as sound as base material
BY S. J. GILL, D. VV. MOON, E. A. METZBOWER AND T. W. CROOKER
ABSTRACT. Studies of fatigue-crack initia-
tion and propagation were conducted on
0.5 in. (12 mm) thick laser beam wel d-
ments in ASTM A36 structural steel.
Cycles to crack initiation and crack
growth rates were measured using frac-
ture mechanics test procedures. Compar-
ative fatigue characteristics were devel-
oped on five material conditions: as-
received base metal, weld metal and heat
affected zones, including both as-welded
and postweld heat treated conditions for
the latter t wo materials. Throughout the
investigation, the fatigue properties of
the laser weldments were found to be
equal or superior to those of the base
metal. Evidence of superior fatigue per-
formance of as-welded weld metal and
heat affected zone materials is attributed
to beneficial compressive residual
stresses.
Introduction
Laser beam welding procedures offer
the promise of producing structural steel
weldments both faster and wi th better
mechanical properties than can be
attained using conventional arc welding
procedures (Ref. 1). Previous studies of
laser beam welded ASTM A36 structural
steel have demonstrated the feasibility of
laser beam welding and have shown that
the tensile and fracture toughness prop-
erties were equal to or better than those
of the base metal (Refs. 2,3). Studies
reported in this paper were devoted to
extending evaluations of A36 laser beam
weldments to include fatigue-crack initia-
tion and propagation.
5. /. GILL, D. W. MOON, E A. METZBOWER,
and T W. CROOKER are with the Material
Science and Technology Division, Naval
Research Laboratory, Washington, D. C
Experimental Procedure
Materials
The base metal used in this study was
0.5 in. (12 mm) thick ASTM A36 steel
(maximum 0.25 wt-% C, 0.80-1.20 wt-%
Mn_ maximum 0.04 wt-% P, maximum
0.05 wt-% S) (Ref. 4). Typical tensile
properties for this steel are: yield strength
40 ksi (275 MPa), ultimate tensile strength
65 ksi (448 MPa), and elongation in 1 in.
(25 mm) 24% (Ref. 3).
Test specimens were machined from
weldments containing continuous laser
beam welds of approximately 12 in. (0.3
m) effective length of wel d.
Welding was performed using a 15 kW
continuous-wave carbon dioxide laser.
The power range used for welding was
between 12 and 13 kW, wi th the travel
speeds between 21 and 33 in./min (9 and
14 mm/s).
The plate surfaces to be welded were
machined for a square butt joint geome-
try and then grit blasted. Then the joint
was wi ped wi th acetone prior to weld-
ing. Helium was used as a shielding gas.
The gas shielding arrangement consisted
of four channels: 1) a leading shield t o
protect the front side of the molten weld
pool , 2) a trailing shield behind the molten
weld pool to protect the hot, newly
solidified wel d metal, 3) a bottom shield
to protect the bottom of the weld bead,
and 4) a plasma suppression shield. The
plasma suppressing gas was used to
remove the plasma which forms above
the molten wel d pool , thus allowing the
laser beam to interact directly with the
workpiece without being absorbed.
Half of the test specimens were given a
postweld heat treatment consisting of
1175F (635C) for one hour, fol l owed
by furnace cooling to remove residual
stresses. Optical microscopy showed
very little difference between the micro-
structure of the as-welded and the post-
weld heat treated specimens (Ref. 3).
Test Specimens
All fatigue tests were conducted using
modified wedge opening loaded
(WOL) type fracture mechanics speci-
mens of various planar sizes and notch
configurations (Ref. 5). The WOL speci-
men design was modified to allow the
use of pin loading in place of wedge
loading. Crack initiation tests were con-
ducted using blunt-notch 1-T specimens,
shown in Fig. 1. Most crack propagation
tests were conducted using sharp-notch
2-T specimens, shown in Fig. 2. However,
some crack propagation data were also
gathered on the smaller 1-T blunt-notch
specimens, as several crack propagation
tests on 2-T specimens were terminated
early due to out-of-plane cracking.
Base metal specimens were machined
directly from the nominal 0.5 in. (12 mm)
thick base metal. However, welded spec-
imens were machined to a reduced thick-
ness of 0.42 in. (10.5 mm) to remove all
undercut. After some crack propagation
specimens exhibited out-of-plane crack-
ing, the remainder were given side
grooves 5% deep on each side, with root
radii of 0.010 in. (0.25 mm).
In welded specimens, locating the
notch root to coincide exactly with the
heat affected zone (HAZ) proved to be
somewhat difficult to achieve, since the
HAZ was not always in the form of a
straight line perpendicular to the plate
surface. Mechanical polishing and etching
with a 1% nital solution were used to aid
in the location of the HAZ.
Test Procedures
All fatigue testing was conducted in an
ambient room air environment using a
sinusoidal wave form and a cyclic fre-
48-S | FEBRUARY 1986
quency of 5 Hz. A load ratio (R = P
m
j
n
/
Pmax) of 0.1 was used throughout.
Crack propagation tests were con-
ducted in accordance wi th ASTM E 647-
83 (Ref. 6). In most instances, crack
length measurements were made using a
crack-opening-displacement technique
(Ref. 7).
Blunt-notch fracture mechanics speci-
mens have been successfully utilized for
fatigue-crack initiation studies in recent
years (Refs. 8,9). Fatigue-crack initiation
was monitored by observing the region
of maximum local stress at the root of the
blunt-notch specimens, using a l ow-pow-
er optical microscope. The notch root
regions were honed in such a way that
the final machining marks were parallel to
the direction of the applied load. All
cracks reported under the category of
"initiation" were sufficiently small to be
within the confines of the notch root
wi dth, i.e., less than 0.25 in. (6 mm) in
length, wi th the average value in this
study being approximately 0.10 in. (2.5
mm). Actually, such crack sizes represent
a reasonable estimate of crack initiation
for engineering purposes because, in this
type of test specimen, shallow cracks
tend t o propagate very rapidly across the
face of the notch, once formed. Thus,
the number of cycles which elapse
between the formation of the first visible
microcrack and its growth to a size
reported as initiation in this study is rather
small (Ref. 9).
Results and Discussion
Fatigue-Crack Initiation
Fatigue-crack initiation in notched
members is controlled by local strains at
the notch root. A number of approaches
have been developed for calculating local
notch-tip stresses and strains from nomi-
nal stress and notch geometry considera-
tions. One such approach utilizes the
parameter AK/\/p, where AK is the frac-
ture mechanics stress-intensity range and
p is the notch root radius. The parameter
AK/ \ / p~ has been shown to correlate
wi th local notch-tip strain and provides a
means of normalizing cycles-to-initiation
data for various notch-tip geometries
(Ref. 9).
The relationship between the local
stress range at the notch-tip (Ao-max) and
A K / V^ is as follows (Ref. 10):
A -
2 ( AK
> r-i.
Affmax r== (1)
irp
For the WOL specimen, the nominal
stress range at the notch-tip (A<r
n
) is given
by (Ref. 11):
Ao-
n
=-
AP
B(W-a)| [-]
D (2 HOLES)
<t
<t
3>,
- W-
- W' -
W
W'
a
h
H
D
B
r
INCHES
2.55
3.20
0.95
1.24
2.48
0.50
VARIABLE
0.094
mm.
64.77
81.28
24.13
31.50
62.99
12.70
VARIABLE
2.39
Fig. I - 1-T blunt-notch WOL specimen
D (2 HOLES)
* B -
c,
L.
C
L.
< $ -
- L >
-w-
- w-
w
w
a
h
H
D
B
inches
5.10
6.40
variable
2.48
4.96
1.00
variable
mm.
129.54
162.56
variable
62.99
125.98
25.40
variable
s
Q.
O
_ i
UJ
>
O
tr
<
UJ
tn
a.
O
_ i
UJ
>
UJ
O
O
tr
<
UJ
CO
UJ
tr
z
ui
5
Q.
o
>
UJ
Q
X
o
tr
<
UJ
a.
O
X
o
ce
<
UJ
tn
UJ
tr
Q.
o
>
UJ
a
o
cr
<
UJ
tn
where AP is the load range, B is the Fig. 2 - 2-T sharp-notch WOL specimen
WELDING RESEARCH SUPPLEMENT 149-s
A 3 6 LASER WELDS
BASE
.'.-_. HAZ
AS
K
E L MD
a

STRESS
RELV'D


Fig. 3 Fatigue-crack initiation data
AK(MPo m"
2
)
20 50 100
A 3 6 LASER WELDS
SASE
WELD
HAZ
"AS
REC'D.
X
AS
WELDED
a
0
STRESS
RELV'D.

5 /
- /
X
4
BASE METAL
TREND LINE
J I l l l l l
I O"
4
10 2 0 50 IOO
STRESS-INTENSITY RANGE, AKl KSI a/Hi)
Fig. 4 Fatigue-crack propagation data
speci men thi ckness, W is t he speci men
wi dt h measur ed f r om t he l oad line, and a
is t he crack dept h measur ed f r om t he
l oad line.
The t heor et i cal elastic stress concen-
t r at i on f act or (K
t
) can be cal cul at ed by
t aki ng t he rat i o of <r
max
/o-
n
f or a gi ven
speci men geomet r y and l oad range. For
t he speci men used in this i nvest i gat i on, as
shown in Fig. 1, using p = 0. 094 i n. (2.4
mm) , t he f act or was K
t
= 3.75.
The f at i gue-crack i ni ti ati on dat a gener-
at ed in this st udy are pl ot t ed in Fig. 3,
whi ch is a semi l ogari t hmi c pl ot of AK/ -v/p*
versus cycl es-t o-i ni t i at i on (IM). The cur ve
of t he base met al t r end line is shown f or
r ef er ence purposes. Run-out poi nt s are
shown at 3 X 10
6
cycl es. Despi t e some
degr ee of scatter in t he dat a, whi c h s not
unusual in wel dment s, t he base met al and
t he post wel d heat t r eat ed wel d met al
and heat af f ect ed zone materi al s exhi bi t
similar f at i gue-crack i ni ti ati on behavi or . In
most i nstances, t he as- wel ded wel d met al
and HAZ materi al s exhi bi t somewhat
superi or f at i gue-crack i ni ti ati on per f or -
mance (l onger life). This di f f er ence
bet ween as- wel ded and post wel d heat
t r eat ed materi al s may be due t o benef i -
cial compr essi ve resi dual stresses f r om
t he wel di ng process.
Fatigue-Crack Propagation
Fati gue-crack pr opagat i on dat a are
shown in Fig. 4, whi ch is a l ogari t hmi c pl ot
of crack g r o wt h rat e ( da/ dN) versus
stress i ntensi ty range (AK). Agai n, as in t he
pr evi ous dat a pl ot , t he base met al t r end
line is shown f or r ef er ence pur poses. As
discussed earl i er, bot h speci men t ypes
wer e used. The bl unt - not ch speci men
was used f or all t he dat a in t he as- wel ded
condi t i on and f or some of t he ot her dat a.
The base met al dat a bet ween 23. 89 and
27.09 ksi x/in". (26.25 and 29. 77 MPa/ v/ m)
wer e gener at ed wi t h this speci men, whi l e
t he r emai nder of t he base met al dat a was
gener at ed wi t h t he shar p- not ch speci -
mens. The dat a f or HAZ not ched speci -
mens in t he PWHT condi t i on bet ween
32.97 and 42. 88 ksi \ / i n . (36.23 and 47. 12
MPa \ / m) and t he dat a f or wel d not ched
speci mens in t he PWHT condi t i on
bet ween 24. 12 and 34. 94 k s i \ / F- (26. 50
and 38. 39 MPav' r n) wer e also gener at ed
wi t h t he bl unt - not ch speci men.
The crack pr opagat i on dat a show t he
same or der i ng of f at i gue behavi or as was
seen in t he crack i ni ti ati on dat a, i.e.,
similar results f or t he base met al and
post wel d heat t r eat ed wel d met al and
HAZ materi al s, and superi or per f or mance
( l ower crack gr owt h rates) in t he as-
wel ded wel d met al and HAZ materi al s.
Such crack gr owt h r et ar dat i on in wel ds
has been obser ved el sewher e and is
at t r i but ed t o compr essi ve resi dual
stresses i nt r oduced by wei di ng (Ref.
12).
Conclusions
Based upon t he results of this st udy, it
appears t hat t he f at i gue pr oper t i es of
wel d met al and heat af f ect ed zones in
ASTM A36 steel wel dment s are equal or
superi or t o t hose of t he base met al .
However , a wo r d of caut i on is in or der
here. The f at i gue behavi or of st ruct ural
wel dment s is pri mari l y cont r ol l ed by f ac-
t ors whi ch lie beyond t he scope of this
i nvest i gat i on, namel y, not ches, di scont i -
nuities and resi dual stresses, whi ch can be
dependent on wel dment geomet r y and
wel di ng pr ocedur es (Ref. 13). However ,
t he results of this st udy pr ovi de assurance
that laser beam wel di ng pr oduces A36
steel wel dment s wi t h basic f at i gue pr op-
erties t hat are as sound as t hose of t he
base mat eri al .
References
1. Metzbower, E. A. 1981. Laser welding.
Naval Engineer /ournal 93(8):49-58.
2. Met zbower, E. A., and Moon, D. W.
1982. Laser beam welding of ASTM A36 steel.
Lasers in Metallurgy, pp. 255-262. Warren-
dale, Pa.: American Institute of Mining, Metal-
lurgical and Petroleum Engineers.
3. Met zbower, E. A., Moon, D. W. , and
Fraser, F. W. 1982. Mechanical behavior of
laser beam weldments. Trends in Welding
Research in the United States, pp. 563-577.
Metals Park, Ohi o: American Society for Met-
als.
4. ASTM A36-81a. 1982. Standard specifi-
cation for structural steel. 1982 Annual Book of
ASTM Standards, Part 4:1-4. Philadelphia, Pa.:
American Society for Testing and Materials. .
5. Saxena, A., and Hudak, S. J., Jr. 1978.
Review of extension of compliance informa-
tion for common crack growt h specimens.
International journal of Fracture 14(5):453-
468.
6. ASTM E 647-83. 1984. Standard test
method for constant-load-amplitude fatigue
crack growt h rates above 10
- 8
m/cycle. 1984
Annual Book of ASTM Standards 03.01:711-
731. Philadelphia, Pa.: American Society for
Testing and Materials.
7. Yoder, C. R., Cooley, L. A., and Crooker,
T. W. 1981. Procedures for precision measure-
ment of fatigue crack growt h rate using crack-
opening displacement techniques. Fatigue
Crack Growth Measurement and Data Analy-
sis, ASTM STP 738, eds. S. ). Hudak, Jr., and
R. |. Bucci, pp. 85-102. Philadelphia, Pa.: Amer-
ican Society for Testing and Materials.
8. Novak, S. R. 1983. Corrosion-fatigue
crack initiation behavior of four structural
steels. Corrosion Fatigue: Mechanics, Metallur-
gy, Electrochemistry and Engineering, ASTM
STP 801, eds. T. W. Crooker and B. N. Leis, pp.
26-63. Philadelphia, Pa.: American Society for
Testing and Materials.
9. Yoder, G. R., Cooley, L. A., and Crooker,
T. W. 1985. Observations on a fracture
mechanics approach to fatigue crack initiation
in Ti-6AI-4V. Fracture Mechanics: Sixteenth
Symposium, ASTM STP 868, eds. M. F. Kannin-
en and A. T. Hopper, pp. 392-405. Philadel-
phia, Pav American Society for Testing and
Materials.
10. Rolfe, S. T., and Barsom, J. M. 1977.
Fracture and Fatigue Control in Structures., p.
212. Englewood Cliffs, N.J.: Prentice-Hall.
11. Wessel, E. T Clark, W. C, ]r and
Wilson, W. K. 1966. Engineering methods for
the design and selection of materials against
fracture. DDC Report AD 801001, Defense
Documentation Center.
12. Kapadia, B. M. 1978. Influence of resid-
ual stresses on fatigue crack propagation in
electroslag welds. Fatigue Testing of Weld-
ment, ASTM STP 648, ed. D. W. Hoeppner,
pp. 244-260. Philadelphia, Pa.: American Soci-
ety for Testing and Materials.
13. Reemsnyder, H. S. 1978. Development
and application of data for structural steel
weldments. Ibid. 3- 21.
50-s | FEBRUARY 1986