Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Fatigue test data were analyzed to determine the relationship between defects
and fatigue performance of weld repaired aluminum castings
BY L. LI, Z. LIU, AND M. SNOW
ABSTRACT. The effect of porosity on the sary process in fabrication of cast alu- fects were smaller than a certain size.
fatigue life of weld repaired D357 alu- minum structures. Properly repaired alu- While ways to avoid weld cracking and
minum cast alloy was investigated. Invest- minum castings may have equivalent sta- porosity defects have been actively inves-
ment cast D357 aluminum alloy was re- tic strength, ductility, fracture toughness, tigated (Refs. 810), remedial methods to
paired with gas tungsten arc (GTA) and fatigue strength levels as the as-cast increase the weld fatigue strength are
welding using a matching filler metal. The structures. For cast structures designed to widely used. Shot peening and heat treat-
repaired alloy underwent a postweld heat be used for aerospace applications, fa- ing (Ref. 11) have been used to increase
treatment involving solutioning and artifi- tigue property is a critical performance the fatigue properties of welded alu-
cial aging. The number of cycles to failure parameter. Past experience and published minum joints.
of the welded specimens showed a much literature on this issue indicate that cast- The causes for the degraded fatigue
wider distribution band than those of the ing defects (cracking, porosity, inclusions) property in an industrial setting can be
as-cast specimens. Crack-like shrinkage and microstructure (coarse second many and complicated, ranging from base
porosity was identified to be the predom- phases) decrease the fatigue strength of materials composition, melting and cast-
inant factor leading to the inconsistent fa- aluminum castings (Refs. 14). Welding ing procedures, heat treatment, weld re-
tigue strength of the welded specimens. and weld repair may introduce further pair procedures, and quality control
The relationship between defects in the complications due to not only additional process. This study reports an attempt to
fusion zone and fatigue failure cycles was metallurgical defects, but also geometric analyze and interpret more than 200 sets
statistically analyzed. Crack-like shrink- discontinuity and residual stresses (Refs. of fatigue testing data, gathered over two
age porosity in the welds was found to 5, 6). Some researchers advised against re- years from as-cast and weld repaired D357
have the most significant effect in de- pair welding (Ref. 7), citing the possibility alloy, in order to find out the relationship
creasing the fatigue life of the welded of additional residual stresses and defects. between defects and the fatigue perfor-
specimens. Spherical pores also con- Some researchers suggest a tighter control mance of weld-repaired D357 aluminum
tributed to decreasing the fatigue life of of defect sizes. For instance, Nordmark et castings.
welded specimens, although not as signif- al. (Ref. 5) suggested that incomplete fu-
icant as the crack-like pores. sion and porosity in the repair welds did Experimental Procedure
not lower the fatigue strength if the de-
Introduction Specimens for fatigue tests were pre-
pared as part of the actual investment
Welding repair is occasionally a neces- KEYWORDS casting of D357 aluminum alloy, as speci-
C
( K m
Kthm )
Discussion tance path. The fatigue crack propagated
intergranularly along the grain/ ai (1)
The existence of crack-like porosity ex- interdendritic boundaries that are delin- where Nf represents fatigue life, C and m
posed to the stress-concentration hole, eated by silicon particles Fig. 15. The are material constants, K is the stress-
which was located near the center of the propagation connected the pores and by- intensity factor range, Kth is the thresh-
weld, might have provided the initiation passed or cut through the silicon particles, old stress intensity factor range, and ai and
sites for the fatigue failure. An example of which were dispersed in the interdendritic af are the initial and final crack lengths, re-
such accelerated formation of fatigue boundaries. The fracture surfaces re- spectively. The value of af is dependent on
fracture is shown in Fig. 12. Welded spec- vealed micromechanisms for the fatigue the fracture toughness of the material; ai
imens that showed the lowest fatigue cy- failure. The separation was mostly is dependent on the crack-initiating spher-
cles almost always contained significant through the dimple or micro microvoid ical or crack-like pores.
amount of crack-like porosity. The frac- coalescence along the grain/interdendritic Since the present study deals with fa-
ture surface of a typical crack-like pore is boundaries Fig. 16. Because of the ex- tigue testing of specimens with preexisting
shown in Fig. 13. Dendritic features and istence of defects, the specimens did not defects, the stress intensity factor range
shrinkage cavities indicated a typical so- fail by typical fatigue failure mechanism, K may be higher than the threshold Kth.
lidification hot cracking mechanism for but rather by a fast fracture mechanism. In Therefore, the fatigue propagation in the
the formation of the crack-like porosity. areas that are free of defects, small regions present study may have bypassed the Re-
Having a crack-like geometry, and a created by a typical fatigue failure mecha- gion I and directly entered the Region II
minute crack tip radius, these pores effec- nism could be seen. In Fig. 17, the fatigue (or steady-state) growth. While this study
tively acted as preformed fatigue cracks. striations and transgranular propagation is not aimed at fatigue crack growth rate
Consequently, there was an absence of path are clearly visible. It seems the fa- (da/dN), a discussion of the effect of de-
crack nucleation period, explaining the tigue specimens tested in this study in- fect size on K will help understand the
premature fatigue failure for the speci- volved a mixed fracture mechanism: fast role of defects on fatigue life.
mens containing crack-like porosity. fracture around defects and fatigue stria- The stress intensity factor KI for the
tions in pockets of defect-free regions. defect-containing fatigue specimens was
Fig. 12 The role of crack-like shrinkage poros- Fatigue crack growth rate has been estimated for both types of pores that were
ity in fatigue cracking initiation and propagation. characterized in fracture mechanics as a exposed to the surface of the stress-con-
( )
K lower than the critical value for crack
K eff = Y aeff aeff
suggested to be more effective than de-
(2) propagation; a specimen containing creasing the total amount of porosity,
where Y(aeff) was a geometrical parameter crack-like defects, however, would have a which in turn to be more effective than de-
defined as K greater than the critical value for crack creasing the average size of pores.
propagation. This analysis is qualitatively
supported by the fatigue data reported in References
Figs. 7 through 11. Although this discus-
a
sion only considered the peak load condi- 1. Evans, W. J., Lu, Z.-J., Spittle, J. A., De-
2 tan
W 2W
( )
Y aeff =
tion and the worst location and orienta- vlukia, J. 1997. Fatigue crack development from
a cos a
tion of the defects, it did provide a defects in a cast aluminum alloy. Proceedings of
2W
a
quantitative insight into the relative ef- the High Cycle Fatigue of Structural Materials, p.
0.75 + 2.02 W
fects of the shape and size of pores on fa- 445457, TMS Fall Meeting.
3
a
tigue crack initiation and propagation. 2.Yi, J. Z., Gao, Y. X., Lee, P. D., Flower, H.
+0.37 1 sin
2W
M., Lindley, T. C. 2003. Scatter in fatigue life
(3) Fig. 16 Magnified region A in Fig. 14. The frac- due to effects of porosity in cast A356-T6 alu-
in which the effective crack length was aeff ture mechanism appears to be microvoid coales- minum-silicon alloys. Metallurgical and Materi-
= a + ry , and the plastic zone size was es- cence (dimples) along the grain/inter-dendritic als Transactions A, 34(9): 18791890.
timated by boundaries. 3. Byczynski, G. E. 2004. A study of crack
2
1 KI
initiation sites in high cycle fatigue of 319 alu-
ry =
YS
Fig. 17 Magnified region B in Fig. 14. Fatigue minum alloy castings. JOM, 56(11): 188.
(4) striations are clearly seen. 4. Skallerud, B., Iveland, T., Harkegard, G.
The peak load during fatigue tests was 1993. Fatigue life assessment of aluminum al-
108 MPa, the measured yield strength Fig. 18 The model for stress intensity factor cal- loys with casting defects. Engineering Fracture
(YS) of the material was 290 MPa, plate culations. Due to symmetry, only half of the fa- Mechanics 44(6): 857874.
width (W) was 5.1 mm, crack length (a) for tigue specimen is considered in this model. 5. Nordmark, G. E., Herbein, W. C., Dick-
the average crack-like porosity was 0.4 erson, P. B., Montemarano, T. W. 1987. Effect of
mm, crack length (a) for the average pore Conclusions weld discontinuities on fatigue of aluminum
depth was 0.05 mm. These values were put butt joints. Welding Journal, 66(6): 162-s to 173-
into the above equations and Keffvalues The major factors that decreased fa- s.
were iteratively solved. The calculated tigue life for the welded specimens were 6. El-Soudani, S. M., Pelloux, R. M. 1975.
maximum Keff for the specimen with 0.05 the existence of welding defects. Crack- Anisotropy of fatigue crack propagation in alu-
mm pore exposed to the surface was 4.37 like shrinkage porosity in the welds had minum alloy butt welded joints. Welding Journal
WELDING JOURNAL 5 -s
54(5): 144-s to 152-s. propagation. This analysis is qualitatively September, 2003, p 1879-1890.
7. Shankar, K., Wu, W. 2002. Effect of weld- supported by the fatigue data reported in 3. Byczynski, Glenn E., A study of crack
ing and weld repair on crack propagation be- Figures 7 through 11. Although this dis- initiation sites in high cycle fatigue of 319
haviour in aluminium alloy 5083 plates. Materi- cussion only considered the peak load aluminum alloy castings, JOM, v 56, n 11,
als and Design. 23(2): 201208. condition and the worst location and ori- November, 2004, p 188.
8. Matsuda, F., Nakata, K., Shimokusu, entation of the defects, it did provide a 4. Skallerud, B.; Iveland, T.; Harkegard,
Y., Tsukamoto, K., Arai, K., Effect of additional quantitative insight into the relative ef- G., Fatigue life assessment of aluminum
element on weld solidification crack suscepti- fects of the shape and size of pores on fa- alloys with casting defects, Engineering
bility of Al-Zn-Mg alloy (report 1) Results of tigue crack initiation and propagation. Fracture Mechanics, v 44, n 6, Apr, 1993,
ring casting cracking test. Transactions of JWRI p 857-874.
(Japanese Welding Research Institute), 12(1): Conclusions 5. Nordmark, G. E.; Herbein, W. C.;
8187. The major factors for the decreased fa- Dickerson, P. B.; Montemarano, T. W.,
9. Kearney, A. L. 1974. Avoiding cracks in tigue life for the welded specimens were Effect of weld discontinuities on fatigue
weld repair of aluminum. Foundry 102(4): the existence of welding defects. Crack- of aluminum butt joints, Welding Jour-
6567. like shrinkage porosity in the welds had nal, v66, n, Jun, 1987, p 162s-173s
10. Ghosh, P. K., Ghosh, A. K. 2004. the most significant effect in decreasing 6. El-Soudani, S. M.; Pelloux, R. M,.
Control of residual stresses affecting fatigue life the fatigue life of the welded specimens. Anisotropy of fatigue crack propagation
of pulsed current gas-metal-arc weld of high- Spherical pores also contributed to de- in aluminum alloy butt welded joints,
strength aluminum alloy. Metallurgical and Ma- creasing the fatigue life of welded speci- Welding Journal, v 54, n 5, May, 1975, p
terials Transactions A, 35A(8): 24392444. mens, although not as significant as the 144s-152s
11. Bertini, L., Fontanari, V., Straffelini, G. crack-like pores. Fatigue cracking propa- 7. Shankar, K, Wu, W., Effect of welding
1998. Influence of post weld treatments on the gated intergranularly in the welded sam- and weld repair on crack propagation be-
fatigue behavior of Al-alloy welded joints. In- ple, connecting the pores along its path. haviour in aluminium alloy 5083 plates,
ternational Journal of Fatigue, 20(10): 749755. With increasing effects on fatigue life were Materials and Design, v 23, n 2, Apil, 2002,
12. Wang, Q. G., Apelian, D., Lados, D. A. mean size, area fraction, and aspect ratio p 201-208
2001. Fatigue behavior of A356-T6 aluminum of pores. Defects located near the fatigue 8. Matsuda, F.; Nakata, K.; Shimokusu,
cast alloys. Part I. Effect of casting defects. Jour- initiation sites were more effective in de- Y.; Tsukamoto, K., Arai, K., Effect of ad-
nal of Light Metals, 1(1): 7384. creasing the fatigue cycles. To improve fa- ditional element on weld solidification
13. Klesnil, M., Lukas, P. 1972. Influence tigue cycles, elimination of crack-like de- crack susceptibility of Al-Zn-Mg alloy (re-
of strength and stress history on growth and sta- fects (those with greater aspect ratios) was port 1) Results of ring casting cracking
bilization of fatigue cracks. Engineering Fracture suggested to be more effective than de- test, Transactions of JWRI (Japanese
Mechanics, Vol. 4, pp. 7792. creasing the total amount of porosity, Welding Research Institute), v 12, n 1,
14. Anderson, T. L. 2005. Fracture Mechan- which in turn to be more effective than de- Aug, 1983, p 81-87
ics Fundamentals and Applications, 3rd Edi- creasing the average size of pores. 9. Kearney, A. L., Avoiding cracks in
tion, p. 52, Taylor & Francis. weld repair of aluminum, Foundry, v 102,
th for fatigue crack propagation for cast References n 4, Apr, 1974, p 65-67
aluminum Alloy 357 is not available, but 1. Evans, W.J.; Lu, Z.-J.; Spittle, J.A.; De- 10.Ghosh, P.K., Ghosh, A. K., Control of
the threshold Kth for Alloy 356, which is vlukia, J., Fatigue crack development residual stresses affecting fatigue life of
similar to Alloy 357 in the present study, from defects in a cast aluminum alloy, pulsed current gas-metal-arc weld of high-
has been calculated by Yi, et al (Ref. 2]. Proceedings of the TMS Fall Meeting, strength aluminum alloy, Metallurgical
Their estimated threshold Kth value is High Cycle Fatigue of Structural Materi- and Materials Transactions A, v 35 A, n 8,
6.1 . From this, it can be suggested that a als, 1997, p 445-457. August, 2004, p 2439-2444
specimen with spherical pores would have 2. Yi, J.Z.; Gao, Y.X.; Lee, P.D.; Flower, 11.Bertini, L.; Fontanari, V.; Straffelini,
a K lower than the critical value for crack H.M.; Lindley, T.C., Scatter in fatigue life G., Influence of post weld treatments on
propagation; a specimen containing due to effects of porosity in cast A356-T6 the fatigue behaviour of Al-alloy welded
crack-like defects, however, would have a aluminum-silicon alloys, Metallurgical joints, International Journal of Fatigue, v
K greater than the critical value for crack and Materials Transactions A, v 34 A, n 9, 20, n 10, Nov, 1998, p 749-755
6 -s