Materials 10 01401 PDF
Materials 10 01401 PDF
Materials 10 01401 PDF
Article
Walker Parameter for Mean Stress Correction in
Fatigue Testing of Al-7%Si-Mg Alloy Castings
Hüseyin Özdeş † and Murat Tiryakioğlu *
School of Engineering, University of North Florida, 1 UNF Drive, Jacksonville, FL 32224, USA;
huseyinozdes@gmail.com
* Correspondence: m.tiryakioglu@unf.edu; Tel.: +1-904-620-1390
† Current Address: Auburn University, Auburn, AL 36849, USA.
Abstract: In this study, performance of two existing Walker parameter estimation models has been
investigated. Results show that those developed mainly for steel did not provide reasonable fits
to experimental Walker parameters for fatigue data for Al-7%Si-Mg alloy castings in the literature.
A strong relationship between the Walker parameter and the structural quality, quantified by the
quality index, QT , was observed and an empirical equation to estimate the Walker parameter for
these alloys was developed. These findings indicate that the Walker parameter is not an intrinsic
material property and the structural quality of the specimens must be taken into account for mean
stress correction in fatigue testing.
Keywords: mean stress effect; runout analysis; S–N curve; Basquin Law
1. Introduction
High cycle fatigue (HCF) performance of metals has been studied extensively in the literature.
The change in fatigue life, Nf , with stress amplitude, σa , is known to follow the Basquin law [1]:
where, σf0 is fatigue strength coefficient and b is the Basquin exponent. The Basquin law has been used
to express the high cycle fatigue (HCF) behavior of metals. Both Basquin parameters are strongly
affected by the material [2], specimen geometry [3] as well as the type of test conducted [4]. It is
well-known that mean stress has a significant effect on the HCF behavior [5]. When the mean stress is
compressive, fatigue life is longest for the same level of alternating stress. When mean stress is tensile,
fatigue life is lowest. For zero mean stress, fatigue life is between the two other S–N curves. In the
literature, numerous correction methods have been introduced to account for the mean stress effect
on fatigue test results. These models have been developed to modify the alternating stress and find
the equivalent stress amplitude, σeq , at zero mean stress. Among those, one of the most widely used
models is the one developed by Walker [6], in which the equivalent stress amplitude is found as:
γ
2
σeq = σa (2)
1−R
where R is the load ratio, (minimum to maximum stress ratio) and γ is the Walker parameter, which
is an empirical parameter that changes between 0 and 1, representing the maximum and minimum
sensitivity to the load ratio, respectively [7]. At zero mean stress, R = −1, and consequently, σeq = σa .
The Walker equation was reported [8,9] to provide best fits in metals. There have been several
studies [6,9–11] in which the Walker parameter was treated as a material property. To the authors’
knowledge, there has not been any effort to determine whether γ is affected by the structural quality
of components. This study is motivated to fill this gap.
2. Background
In aluminum castings, structural defects, such as pores and oxide films entrained from surface,
i.e., bifilms, have a strong, deleterious effect on the mechanical properties related to fracture, such as
elongation [12], fracture stress [13,14], fracture toughness and fatigue life [15,16]. Davidson et al. [17]
showed that there is a linear relationship between logarithms of fatigue life and largest defect size in
Al-7%Si-Mg alloy castings. Moreover, one of the authors [18,19] demonstrated that the fatigue life
distribution can be directly linked to the size distribution of largest defects in Al-7%Si-Mg alloy castings.
Iben Houria et al. [20] investigated the combined effect of largest defect size and mean stress
on the fatigue strength (at 107 cycles) of A356 alloy castings. At zero mean stress (R = −1), fatigue
strength was not affected by the largest pore sizes less than 400 µm. With increasing largest pore size
after 400 µm, however, there was a sharp decrease in fatigue strength. When only tensile stresses are
present (R ≥ 0), the sensitivity of fatigue strength to maximum pore size was found to start at almost
50 µm. At pores sizes, larger than 400 µm, fatigue strength was not affected by maximum pore size.
Hence the response of the aluminum casting with defects to alternating loads is completely opposite
of each other at the two different stress ratios. This difference may be explained by the effect of the
driving force of local stress concentrations at the tip of the crack during crack propagation, as observed
by Gall et al. [21], who conducted FEM analysis of fatigue behavior in the presence of pores and hard
inclusions embedded in a metal matrix. At R = −1 (fully-reversed), the contributions of pores and hard
inclusions to driving force for crack propagation were identical. The contribution of inclusions was
twice of that of pores when mean stress became positive. Therefore, there is evidence in the literature
that fatigue behavior in the presence of structural defects, such as pores, is determined by mean stress.
Consequently, the assumption made in correction models for mean stress that the contribution of
structural defects is the same regardless of R, is not necessarily valid for parts with different structural
defects and varying structural quality, such as castings.
Tensile testing is the most commonly used test to determine the mechanical properties of metals,
such as yield strength (σY ), tensile strength (ST ) and elongation (eF ). Tiryakioğlu and coworkers
collected hundreds of data points from the aerospace and premium castings literature for Al-7Si-Mg,
A206 and A201 alloy castings [22–24] and plotted elongation versus yield strength, which is minimally
affected by structural defects. They concluded that the highest elongation points form a linear trend,
which can be written as:
eF(max) = β0 − β1 σY (3)
where β0 and β1 are alloy-dependent coefficients. For cast Al-Si-Mg alloys, β0 and β1 are 36.0 and
0.064 (MPa−1 ), respectively. Equation (3) can be used to estimate the ductility potential of aluminum
and magnesium alloy castings [23–27]. The structural quality of aluminum alloy castings can then be
quantified by using a quality index, QT , which is expressed as:
eF eF
QT = = (4)
eF(max) β0 − β1 σY
The authors have found [28,29] a strong relationship between QT and fatigue life in cast aluminum
alloys, and demonstrated that QT can be used to predict fatigue strength [30,31] and Basquin coefficients
for Al-Si-Mg-(Cu) alloy castings [31].
Turning our attention to the Walker parameter, Dowling et al. [9] showed that the Walker
parameter, γ, changes linearly with tensile strength of steels:
γ = γ0 − a·ST (5)
Materials 2017, 10, 1401 3 of 9
Dowling et al. reported γ0 and a to be 0.883 and 2 × 10−4 MPa−1 , respectively, for steels. The same
authors stated that γ = 0.651 when ST is between 300 and 350 MPa, and γ = 0.473 for ST values between
450 and 600 MPa. Lv et al. [8] stated that γ is expected to be low for materials with low ductility and
proposed the following equation to determine the Walker parameter:
ST − σY
γ = 0.5 ± (6)
ST +σY
Lv et al. did not provide a physical reason for the ± sign but showed that there is agreement
between the γ values reported by Dowling et al. [9] and those estimated by Equation (6).
In the studies summarized above, γ has been stated as a material property [6,9,10]. However,
the results of Iben Houria et al. [20] and Gall et al. [21] on the fatigue behavior in aluminum castings
with defects suggest that the effect of the mean stress around the structural defect that leads to fatigue
failure is affected by the type of defect and the level of the mean stress. Consequently, it is hypothesized
that the Walker parameter is not an intrinsic material property and is affected by the extrinsic effects of
casting’s structure, i.e., casting quality. This hypothesis will be tested by using data from literature.
• Palmer [32] tested 356-T6 lost foam castings with as-cast and machined surfaces. Specimens with
as-cast surfaces were tested at seven R ratios ranging from −1 to 0.44, whereas eight R ratios
between −1 and 0.62 were used for machined specimens.
• Wang et al. [15] used Sr-modified A356 casting alloy in T6 heat treatment condition to study the
influence of casting defects on the room temperature fatigue performance by using unnotched
polished cylindrical specimens. R ratio of 0.1 and −1 have been included in the study.
• Koutiri et al. [33] studied the high-cycle fatigue behavior of cast hypo-eutectic Al-Si alloy to
investigate the fatigue damage mechanisms under complex loading conditions with two different
load ratios (R = 0.25 and 0.73).
• Oswalt [34] investigated unchilled test bars of 357-T6 alloy to determine the fatigue strength at
R ratios of 0.2 and −1 were used in the study.
• Mu et al. [35] used AS7G06 (A357) cast aluminum alloy in T6 heat treatment condition for two
different R ratios (0.1 and −1) to analyze types of defects at the origin of the failure.
• Munoz [36] focused on A357-T6 cast aluminum alloy to investigate the effect of the material
microstructural parameters that affect the small fatigue cracks. R ratios of 0.1 and −1 were used
in this study.
• Jana et al. [37] studied cast F357 (Be-free variant of A357) plates of ~3.3 mm thickness to investigate
the effect of friction stir processing (FSP) on fatigue life of sand castings. Fatigue tests were run at
stress ratio of R = 0 and R = −1 both before and after friction stir processing (FSP).
To quantify their structural quality levels, quality index values of each datasets were estimated by
using equation 3 based on their tensile properties. In addition, the analysis conducted in this study
incorporated a maximum likelihood (ML) method developed by Sarkani et al. [38] except for dataset
from Palmer [32] who published their Basquin and Walker parameters. Though it was not mentioned
above for the sake of clarity, it is known that varying stress ratios not only generate beneficial or
detrimental effects on fatigue results, but also do act significantly different at HCF or LCF regions.
Unsurprisingly, they have more noteworthy influence within the HCF region. In HCF region, damage
is driven by elastic deformation meaning that fatigue life is mostly spent on crack initiation. Since
there is not significant plastic deformation in this region, once crack initiated typically is followed
Materials 2017, 10, 1401 4 of 9
closely by sudden fracture. Tensile mean stresses then could be observed more clearly in this region.
Typically, due to time concerns in fatigue testing, tests are interrupted after certain cycles and the data
points referred to as runouts. In some cases, these runouts are discarded from the analyses or treated
as failure data which may cause misleading results. To this aspect of their significance, there has been
some attempt made by researchers with statistical basis. To the authors’ knowledge, such a study with
proper treatment of survival data had not been conducted before.
Table 1. Tensile data, quality index and estimated Basquin and Walker parameters for the nine datasets.
Note that QT values for each dataset were also provided, and ranged between 0.027 and 0.860,
representing almost the entire spectrum (between 0 and 1) of the quality index. The S–N diagrams for
three of the datasets before and after the Walker correction for mean stress are presented in Figure 1.
These datasets are specifically chosen to represent the minimum, maximum and intermediate level of
QT among the nine datasets investigated in this study. Regardless of the level of the structural quality
level, the Walker correction provides very respectable fits to the experimental data.
The hypothesis that the value of γ is affected by the structural quality of the castings has been
tested; QT versus γ plots is presented in Figure 2.
Note that there is strong evidence that the Walker parameter is affected by the structural quality
of aluminum castings. The best fit curve indicated in Figure 2 is expressed as;
with a coefficient of determination, R2 = 0.987. Hence with increasing quality, γ decreases and therefore
castings become more sensitive to the mean stress effects in fatigue testing. Moreover, the highest
quality Al-Si alloy casting (QT = 1.0) would have a Walker parameter of 0.358, which represents the
intrinsic value for these alloys. As stated above, the Walker parameter has been regarded as a material
property in the literature. However, the strong effect of QT on the Walker parameter found in this
study indicates that γ includes the quality of the material and therefore has an extrinsic component.
The performance of the equations proposed by Dowling et al. [9] and Lv et al. [8] performed with
the results summarized in Table 1 are presented in Figure 3.
Note that QT values for each dataset were also provided, and ranged between 0.027 and 0.860,
representing almost the entire spectrum (between 0 and 1) of the quality index. The S–N diagrams
for three of the datasets before and after the Walker correction for mean stress are presented in Figure 1.
These datasets are specifically chosen to represent the minimum, maximum and intermediate level
of QMaterials
T among the nine datasets investigated in this study. Regardless of the level of the structural
2017, 10, 1401 5 of 9
quality level, the Walker correction provides very respectable fits to the experimental data.
(a)
(c)
Figure
Figure1. The S–NS–N
1. The curves at Rat=R−1= after
curves Walker
−1 after mean
Walker meanstress correction
stress forfor
correction (a)(a)
sand
sand cast
castF357
F357bybyJana
et al.Jana
[37],et(b)
al. A357 by Mu
[37], (b) A357etby
al. Mu
[35],etand (c) friction
al. [35], and (c)stir processed
friction F357 by Jana
stir processed F357etbyal.Jana
[37]etThe
al. original
[37].
dataThefororiginal
R = 0 as data
wellfor
as their
R = 0 correction according
as well as their to Walker
correction accordingequation (Equation
to Walker equation(2)) are also (2))
(Equation indicated.
are
also indicated.
The hypothesis that the value of γ is affected by the structural quality of the castings has been
tested; QT versus γ plots is presented in Figure 2.
Figure 1. The S–N curves at R = −1 after Walker mean stress correction for (a) sand cast F357 by Jana
et al. [37], (b) A357 by Mu et al. [35], and (c) friction stir processed F357 by Jana et al. [37] The original
data for R = 0 as well as their correction according to Walker equation (Equation (2)) are also indicated.
The hypothesis
Materials that the value of γ is affected by the structural quality of the castings has
2017, 10, 1401 6 ofbeen
9
tested; QT versus γ plots is presented in Figure 2.
Note that there is strong evidence that the Walker parameter is affected by the structural quality
of aluminum castings. The best fit curve indicated in Figure 2 is expressed as;
.
γ = 0.588 − 0.230 ⋅ Q (7)
with a coefficient of determination, R2 = 0.987. Hence with increasing quality, γ decreases and
therefore castings become more sensitive to the mean stress effects in fatigue testing. Moreover, the
highest quality Al-Si alloy casting (QT = 1.0) would have a Walker parameter of 0.358, which
represents the intrinsic value for these alloys. As stated above, the Walker parameter has been
regarded as a material property in the literature. However, the strong effect of QT on the Walker
parameter found in this study indicates that γ includes the quality of the material and therefore has
an extrinsic component.
The performance of the equations proposed by Dowling et al. [9] and Lv et al. [8] performed
with the results summarized in Table 1 are presented in Figure 3.
(a)
(b)
Figure
Figure3.3.(a)(a)The
Therelationship
relationshipbetween
between γγ and
and SSTT,, as
as suggested
suggestedby
byDowling
Dowlingetetal.al.
[9],[9],
andand
(b)(b)
the the
comparison
comparison of Walker parameters for best fit and estimated by using the equation proposed by by
of Walker parameters for best fit and estimated by using the equation proposed
LvLv
et et
al.al.
[8].
[8].
However, the coefficient of determination is much lower and therefore the change in ST can
explain only 37.6% of the change in the Walker parameter. The equation proposed by Lv et al. was
found to perform very poorly on cast Al-7%Si-Mg alloys, as shown in Figure 3b, as indicated by the R2
being lower than zero.
5. Conclusions
In the light of observations made in this study, following conclusions could be drawn:
• Experimental Walker parameter has been observed to show a correlation with QT (derived from
tensile properties) meaning that it has an extrinsic component and should not be treated as
a material property without considering the casting’s quality level.
• Equations, developed mainly for steel, to estimate the Walker parameter do not provide accurate
results for Al-7%Si-Mg alloy castings.
• A new methodology for estimating the Walker parameter based on alloys quality level involving
proper runout analysis has been developed. The following empirical equation can be used to
estimate γ:
γ = 0.588 − 0.230 · QT 0.541
Acknowledgments: H.O. would like to acknowledge a graduate assistantship provided by Alotech Corporation.
Author Contributions: Hüseyin Özdeş and Murat Tiryakioğlu collaborated in all phases of this paper, namely
the data collection, statistical analysis and interpretation of results.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Basquin, O. The exponential law of endurance tests. Proc. ASTM 1910, 625–630.
2. Kun, F.; Carmona, H.A.; Andrade, J.S., Jr.; Herrmann, H.J. Universality behind basquin’s law of fatigue.
Phys. Rev. Lett. 2008, 100, 094301. [CrossRef] [PubMed]
3. Collins, J.A. Failure of Materials in Mechanical Design: Analysis, Prediction, Prevention; John Wiley & Sons:
Hoboken, NJ, USA, 1993.
4. Manson, S.S. Fatigue: A complex subject—Some simple approximations. Exp. Mech. 1965, 5, 193–226.
[CrossRef]
5. Ince, A.; Glinka, G. A modification of morrow and smith-watson-topper mean stress correction models.
Fatigue Fract. Eng. Mater. Struct. 2011, 34, 854–867. [CrossRef]
6. Walker, K. The effect of stress ratio during crack propagation and fatigue for 2024-T3 And 7075-T6 aluminum.
In Effects of Environment and Complex Load History on Fatigue Life; ASTM International: West Conshohocken,
PA, USA, 1970.
7. Duran, J.A.R.; Hernandez, C.T. Evaluation of three current methods for including the mean stress effect in
fatigue crack growth rate prediction. Fatigue Fract. Eng. Mater. Struct. 2015, 38, 410–419. [CrossRef]
8. Lv, Z.; Huang, H.-Z.; Wang, H.-K.; Gao, H.; Zuo, F.-J. Determining the walker exponent and developing
a modified smith-watson-topper parameter model. J. Mech. Sci. Technol. 2016, 30, 1129–1137. [CrossRef]
9. Dowling, N.E.; Calhoun, C.A.; Arcari, A. Mean stress effects in stress-life fatigue and the walker equation.
Fatigue Fract. Eng. Mater. Struct. 2009, 32, 163–179. [CrossRef]
10. Dowling, N.E. Mean stress effects in strain–life fatigue. Fatigue Fract. Eng. Mater. Struct. 2009, 32, 1004–1019.
[CrossRef]
11. Anon. Mil-hdbk-5j: Metallic Materials and Elements for Aerospace Vehicle Structures; United States Department
of Defense: Washington, DC, USA, 2003.
Materials 2017, 10, 1401 8 of 9
12. Tiryakioğlu, M. On estimating the fracture stress and elongation of Al–7%Si–0.3%Mg alloy castings with
single pores. Mater. Sci. Eng. A 2010, 527, 4546–4549. [CrossRef]
13. Tiryakioğlu, M.; Staley, J.T.; Campbell, J. The effect of structural integrity on the tensile deformation
characteristics of A206-T71 alloy castings. Mater. Sci. Eng. A 2008, 487, 383–387. [CrossRef]
14. Tiryakioǧlu, M.; Campbell, J.; Staley, J.T. The influence of structural integrity on the tensile deformation of
cast Al–7wt.%Si–0.6wt.%Mg alloys. Scr. Mater. 2003, 49, 873–878. [CrossRef]
15. Wang, Q.G.; Apelian, D.; Lados, D.A. Fatigue behavior of A356-T6 aluminum cast alloys. Part I. Effect of
casting defects. J. Light Met. 2001, 1, 73–84. [CrossRef]
16. Wang, Q.G.; Crepeau, P.N.; Davidson, C.J.; Griffiths, J.R. Oxide films, pores and the fatigue lives of cast
aluminum alloys. Metall. Mater. Trans. B 2006, 37, 887–895. [CrossRef]
17. Davidson, C.; Griffiths, J.; Machin, A. The effect of solution heat-treatment time on the fatigue properties of
an al-si-mg casting alloy. Fatigue Fract. Eng. Mater. Struct. 2002, 25, 223–230. [CrossRef]
18. Tiryakioğlu, M. On the relationship between statistical distributions of defect size and fatigue life in
7050-T7451 thick plate and A356-T6 castings. Mater. Sci. Eng. A 2009, 520, 114–120. [CrossRef]
19. Tiryakioğlu, M. Relationship between defect size and fatigue life distributions in Al-7 pct si-mg alloy castings.
Metall. Mater. Trans. A 2009, 40, 1623–1630. [CrossRef]
20. Iben Houria, M.; Nadot, Y.; Fathallah, R.; Roy, M.; Maijer, D.M. Influence of casting defect and sdas on the
multiaxial fatigue behaviour of A356-T6 alloy including mean stress effect. Int. J. Fatigue 2015, 80, 90–102.
[CrossRef]
21. Gall, K.; Horstemeyer, M.F.; Degner, B.W.; McDowell, D.L.; Fan, J. On the driving force for fatigue crack
formation from inclusions and voids in a cast A356 aluminum alloy. Int. J. Fracture 2001, 108, 207–233.
[CrossRef]
22. Tiryakioglu, M.; Campbell, J.; Alexopoulos, N.D. Quality indices for aluminum alloy castings: A critical
review. Metall. Mater. Trans. B 2009, 40, 802–811. [CrossRef]
23. Tiryakioğlu, M.; Campbell, J. Quality index for aluminum alloy castings. Int. J. Metalcast. 2014, 8, 39–42.
[CrossRef]
24. Tiryakioğlu, M.; Campbell, J.; Alexopoulos, N.D. On the ductility of cast Al-7 pct si-mg alloys. Metall. Mater.
Trans. A 2009, 40, 1000–1007. [CrossRef]
25. Tiryakioglu, M.; Campbell, J. Ductility, structural quality, and fracture toughness of Al–Cu–Mg–Ag (A201)
alloy castings. Mater. Sci. Eng. 2009, 25, 784–789. [CrossRef]
26. Űnal, O.; Tiryakioğlu, M. On the ductility potential and assessment of structural quality in mg alloy castings.
Mater. Sci. Eng. A 2015, 643, 51–54. [CrossRef]
27. Tiryakioğlu, M.; Campbell, J.; Alexopoulos, N.D. On the ductility potential of cast Al–Cu–Mg (206) alloys.
Mater. Sci. Eng. A 2009, 506, 23–26. [CrossRef]
28. Özdeş, H.; Tiryakioğlu, M. On the relationship between structural quality index and fatigue life distributions
in aluminum aerospace castings. Metals 2016, 6, 81. [CrossRef]
29. Tiryakioğlu, M. On the relationship between elongation and fatigue life in A206-T71 aluminum castings.
Mater. Sci. Eng. A 2014, 601, 116–122. [CrossRef]
30. Özdeş, H.; Tiryakioğlu, M. The effect of structural quality on fatigue life in 319 aluminum alloy castings.
J. Mater. Eng. Perform. 2017, 26, 736–743. [CrossRef]
31. Özdeş, H.; Tiryakioğlu, M. On estimating high-cycle fatigue life of cast al-si-mg-(cu) alloys from tensile test
results. Mater. Sci. Eng. A 2017, 688, 9–15. [CrossRef]
32. Palmer, D.E. Stress Ratio Effects in Fatigue of Lost Foam Cast Aluminum Alloy 356; University of
Wisconsin-Milwaukee: Milwaukee, WI, USA, 2014.
33. Koutiri, I.; Bellett, D.; Morel, F.; Augustins, L.; Adrien, J. High cycle fatigue damage mechanisms in cast
aluminium subject to complex loads. Int. J. Fatigue 2013, 47, 44–57. [CrossRef]
34. Oswalt, K. A Comparison of the Fatigue Behavior of High Strength cast Aluminum Alloys 201 and A357;
DTIC Document: Ft. Belvoir, VA, USA, 1971.
35. Mu, P.; Nadot, Y.; Nadot-Martin, C.; Chabod, A.; Serrano-Munoz, I.; Verdu, C. Influence of casting defects on
the fatigue behavior of cast aluminum AS7G06-T6. Int. J. Fatigue 2014, 63, 97–109. [CrossRef]
36. Serrano Munoz, I. Influence of Casting Defects on the Fatigue Behaviour of an A357-T6 Aerospace Alloy.
Ph.D. Dissertation, INSA, Lyon, France, 2014.
Materials 2017, 10, 1401 9 of 9
37. Jana, S.; Mishra, R.S.; Baumann, J.B.; Grant, G. Effect of stress ratio on the fatigue behavior of a friction stir
processed cast Al–Si–Mg alloy. Scr. Mater. 2009, 61, 992–995. [CrossRef]
38. Sarkani, S.; Mazzuchi, T.A.; Lewandowski, D.; Kihl, D.P. Runout analysis in fatigue investigation.
Eng. Fract. Mech. 2007, 74, 2971–2980. [CrossRef]
© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).