International Journal of Fatigue 158 (2022) 106562

Contents lists available at ScienceDirect

International Journal of Fatigue

journal homepage: www.elsevier.com/locate/ijfatigue

Influence of the grain size on the fatigue initiation life curve

M. Mlikota a, K. Dogahe a, b, *, S. Schmauder a, Ž. Božić c
a
University of Stuttgart, Institute for Materials Testing, Materials Science and Strength of Materials (IMWF), Pfaffenwaldring 32, 70569 Stuttgart, Germany
b
University of Stuttgart, Graduate School of Excellence Advanced Manufacturing Engineering (GSaME), Nobelstraße 12, Stuttgart, Germany
c
University of Zagreb, Faculty of Mechanical Engineering and Naval Architecture, I. Lučića 5, 10000 Zagreb, Croatia

A R T I C L E I N F O A B S T R A C T

Keywords: This study is dealing with the effect of the grain size on the fatigue lifetime of metallic components, which is
Fatigue life represented in the form of (S-N) Wöhler curves of two different grain sizes. The micro-models containing the
Finite element method microstructures of the two different grain sizes are handled by employing the Finite Element Method (FEM) and
Microstructure
the required number of cycles for crack initiation was calculated by using the physically-based Tanaka-Mura
Grain size
Physically-based modelling
model. By implementing the different grain sizes in the micro-model, the effect of the grain size on the fatigue life
and especially on the endurance limit, in the form of the S-Nini diagram, can be obtained.

that investigated the effect of grain size on S-N diagram experimentally.

Generally, the fatigue life of a component increases as the grain size
1. Introduction
decreases [9–11].
The effect of the grain size on the fatigue behavior of commercial
The phenomenon of fatigue is the consequence of cyclic loading,
carbon steel with fine grains (6.5 μm) and coarse grains (20 μm) by
which consists of the crack nucleation and subsequent crack propagation
means of the S-N diagram was investigated by Kanemaru et al. [12]
up to final fracture. In metallic components which are free of cracks and
experimentally. As it can be observed in Fig. 1-Left, samples with finer
other microstructural inconsistencies, the fatigue cracks initiate from
grains performed better under fatigue loading conditions in comparison
slip motion of the dislocations [1]. The resistance to the short cracks
to samples of the same material with coarse grains. Also in another work
propagation is remarkably influenced by the microstructural features of
by Sharifi et al. [13], the relationship between the ferrite grain size and
the material such as the grain size and orientation [2,3].
the endurance limit at 107 cycles for dual phase steels, with 25 % of
Whenever a metallic component experiences deformation, the grain
martensite volume fracture, has been investigated (Fig. 1-Right). Based
boundaries block the dislocation slip, and as a consequence, the dislo­
on the results, by decreasing the ferrite grain size from 74.3 μm to 9.3 μm
cations pile up leading to build up of stress concentrations in the
the endurance limit increased. Apart from that, it was also observable
component. Nevertheless, there are some other grain boundaries that
that the endurance limit varied linearly from 330 MPa to 265 MPa, with
allow dislocations to transmit across them [4]. The grain size of an alloy
respect to the ferrite grain size.
is a determining factor for its mechanical properties, which means that,
In the present study, the effect of the microstructure grain size on the
by reducing the grain size, the alloy shows better mechanical properties.
fatigue crack initiation and finally the fatigue life of the AISI 1141 steel
The reason behind this is that more boundaries prevent the movement of
in the form of the fatigue life (S-N) curve will be demonstrated for two
dislocations within the microstructure [5,6]. Grain refinement can be
microstructures with different grain sizes.
obtained by introducing some rare earth (REs) alloying elements into the
alloy such as Lanthanum and/or Cerium which may be added to the
2. Method and material
alloys during casting and cause the grain refinement of steels due to
heterogeneous nucleation on in-situ formed RE inclusions [7].
2.1. Method and reference to our previous work
An S-N diagram for a given material is a plot of the magnitude of
stresses versus the number of cycles up to failure. In this sense both stress
The S-N diagram for the AISI 1141 steel was numerically calculated
and number of cycles are displayed on logarithmic scales. Given the S-N
by Mlikota and Schmauder [14]. In order to develop such an S-N dia­
diagram and load-time history, one can determine the damage accu­
gram, the number of cycles required for crack initiation under the effect
mulation or fatigue life of a mechanical part [8]. There are many works

* Corresponding author.
E-mail address: [email protected] (K. Dogahe).

https://doi.org/10.1016/j.ijfatigue.2021.106562
Received 19 February 2021; Received in revised form 21 September 2021; Accepted 22 September 2021
Available online 25 September 2021
0142-1123/© 2021 Published by Elsevier Ltd.
M. Mlikota et al. International Journal of Fatigue 158 (2022) 106562

energy, ν for the Poisson’s ratio, d for the slip band length, Δτs is the
Nomenclature average shear stress range on the slip band, and finally, CRSS, is the
critical resolved shear stress value, necessary for a dislocation along the
CRSS critical resolved shear stress analyzed slip band to start to move. A more comprehensive explanation
C11, C12, C44 material elastic constants for cubic crystal about the application of the TM formulation in FEM modelling in order
d slip band length to simulate the crack initiation process is reported in the following
ds slip band segment length publications: Mlikota et al. [18,19], Mlikota and Schmauder [20–22],
E elastic modulus Božić et al. [22–25].
FEM finite element method
G shear modulus 2.2. Material
HCF high cycle fatigue
N number of loading cycles The material AISI 1141 steel is considered in this study, which its
Nf number of cycles to fracture a specimen or a component chemical composition is tabulated in Table 1.
Nini number of cycles to initiate the short crack Table 2 contains the mechanical properties of AISI 1141 steel,
Ns number of cycles to nucleate a crack segment in a single including the Young’s modulus E, shear modulus G, Poisson’s ratio ν, the
grain ultimate strength Rm, the 0.2% yield strength Rp0.2, the crack initiation
RVE representative volume element energy Wc, and CRSS value.
S applied stress level In order to define the material model property, elastic orthotropic
Wc crack initiation energy material properties are assigned to the micro-model. The elastic con­
ν Poisson’s ratio stants for the BCC AISI 1141 steel which are assigned to the micro-model
TM Tanaka-Mura are: C11 = C22 = C33 = E(1 − ν)/(1 − υ − 2υ2 ) = 255.682 GPa, C12 = C13
Δτs average shear stress range along the slip band segment = C23 = Eν/(1 − υ − 2υ2 ) = 99.432 GPa, C44 = C55 = C66 = G = 78.125
Rm ultimate strength GPa [27] For the macro-model the pure isotropic material behavior is
Rp0.2 substitutional 0.2% yield strength adopted.
The microstructure assigned to the sub-model is taken from the work
of Mirzazadeh and Plumtree [26], with 60 μm grain size. The model is
created as a 3D shell RVE, generated by the Voronoi tessellation tech­
of microstructure is calculated by employing the Tanaka-Mura (TM)
nique consisting of 253 grains. In order to examine the grain size effect
equation [15,16]. In this sense a three dimensional (3-D) deformable
on the fatigue crack initiation stage, the reference grain size of 60 μm is
shell micro-model, which includes microstructural characteristics of the
decreased in another microstructural model to 50 μm with 313 grains.
material, is developed by using the finite element method (FEM) and
Both generated microstructures are showed in Fig. 3.
embedded in the macro-model by using a sub-modelling technique in
ABAQUS as it is shown in Fig. 2. The red edges of the microstructural
2.3. Crack initiation analysis
model are those which are embedded within the macro-model. The
macro-model in this study is taken as the half of the test specimen which
Fig. 4 shows the microstructural models for the both investigated
was used in the work of Fatemi et al. [17]. The loading and boundary
grain sizes, placed in the notch region under the outer loading of 224
conditions are explained in the work of Mlikota et al. [14].
MPa, giving the von Mises stress distribution in both cases. As it is
As mentioned earlier, in order to calculate the number of cycles for
observable from the results, besides the reduction of the number of
micro crack nucleation inside a grain, the physical-based TM formula­
grains which favors crack nucleation, the crack coalescence decreased
tion for cyclic loading is employed based on the pile up of dislocations
significantly by reducing the grain size in the micro-model in accordance
inside a grain:
to the experimental results of Alzubi et al. and Taylor et al. [2].
8GWc Crack initiation provides the largest contribution to the whole fa­
Ng = (1)
π(1 − ν)ds (Δτs − 2CRSS)2 tigue life of a material, and by approaching towards the endurance limit,
the contribution of the crack initiation would be more pronounced. As
where G stands for the shear modulus, Wc, for the crack initiation mentioned above, the number of loading cycles Nini, required for fatigue

Fig. 1. (-Left) The effect of grain size on fatigue behavior of the commercial carbon steels. (-Right) The relationship between the ferrite grain size of dual phase steels
and the endurance limit at 10E7 cycles at R = − 1 [12].

2
M. Mlikota et al. International Journal of Fatigue 158 (2022) 106562

Fig. 2. 3D global model of the notched tensile specimen with a 3D deformable shell sub-model which is embedded in the notch area.

initiation life of the microstructures in the form of S-Nini curves, for the
Table 1 AISI 1141 steel with two different grain sizes. The fatigue endurance
Chemical composition of AISI 1141 steels.
limit can be obtained by incremental decrease of the stress magnitudes
Fe Mn C S P in microstructurally based crack initiation modelling, up to the points
200 78.125 0.28 875 564 where the extreme number of cycles cause just few or no cracks, inside
the micro-structural model, which are recognized as run-out results.
It is observable that by having a finer grain size the number of
Table 2 initiation cycles Nini will be higher and accordingly the initiation fatigue
Mechanical properties of AISI 1141 steels [17]. life will be longer. In consequence, the endurance limit, has been
increased from 76 MPa (for 60 μm grain size) to 78.5 MPa (for 50 μm
E (GPa) G (GPa) ν Rm Rp0.2 CRSS Wc (kJ/
(MPa) (MPa) (MPa) m2)
grain size). Furthermore, it is expected that the fatigue life curve in the
S-N-diagram shifts also to the right and upwards for smaller grain sizes.
200 78.125 0.3 875 564 117 19

3. Conclusions
crack initiation, is calculated by summing up the number of cycles spent
for the nucleation of the individual cracks by employing the TM equa­ The analyses performed in this work quantitatively agrees with the
tion (Eq. (1)). Namely, the grain size, or actually the slip band length, experimental observations. Aside from the experimental studies that
d is a part of the denominator of the TM equation, meaning that the show a decreasing trend of the fatigue resistance with increasing grain
estimated cycles Nini should increase by decreasing d if all other factors size, the expectations to obtain the same trend of the experimental ob­
of the Eq.1 are kept constant. Since the crack initiation is the major step servations in the numerical studies has been achieved by using the TM
in the fatigue life of the material, the total number of cycles which are equation. It has been observed that by decreasing the grain size, or the
required for a material to fail increases with decreasing grain size and slip band length d, the total number of cycles for the fatigue crack
the fatigue life curve shifts to the right. initiation, and the total fatigue life, is increased and the material per­
Based on the study by Mlikota et al. [14], the model that is used in formed higher fatigue resistance.
this work is in good agreement with experimental observations, e.g. of
Fatemi et al. [16], for the grain size of 60 μm.
Fig. 5 shows the results of the numerical study of the fatigue

Fig. 3. Microstructural models of AISI 1141 steel with: Left- 60 μm grain size and 253 grains; Right- 50 μm grain size and 313 grains in total.

3
M. Mlikota et al. International Journal of Fatigue 158 (2022) 106562

Fig. 4. Damage initiation in a microstructural models of AISI 1141 steel under the load of 224 MPa and loading ratio R = 0, Left - coarse grain (60 μm) micro­
structure under 9E+04 cycles, Right - fine grain (50 μm) microstructure under 15E+04.

Fig. 5. Numerical study of the influence of the grain size on S-Nini curves of AISI 1141 Steel.

Declaration of Competing Interest [5] Hall EO. The deformation and aging of mild steel: III Discussion results. Proc Phys
Soc London 1951;B64:747–53.
[6] Petch NJ. The cleavage strength of polycrystals. Iron Steel Inst 1953;174:25–8.
The authors declare that they have no known competing financial [7] Ji Y, Zhang M-X, Ren H. Roles of lanthanum and cerium in grain refinement of
interests or personal relationships that could have appeared to influence steels during solidification. J. Metals 2018;8(11):884. https://doi.org/10.3390/
the work reported in this paper. met8110884.
[8] Sangid MD. The physics of fatigue crack initiation. Int J Fatigue 2013;57:58–72.
[9] Hoeppner D. The effect of the grain size on fatigue crack propagation in copper,
Acknowledgement fatigue crack propagation in copper. ASTM 415, AM. SOC. testing mats; 1967. p.
486–504.
[10] Morrison DJ, Moosbrugger JC. Effects of grain size on cyclic plasticity and fatigue
This research was funded by German Research Foundation (Deutsche crack initiation in nickel. Int J Fatigue 1997;19(93):51–9.
Forschungsgemeinschaft, DFG), grant number SCHM 746/222-1, and [11] Liang F-L, Laird C. Control of intergranular fatigue cracking by slip homogeneity in
copper I: Effect of grain size. Mater Sci Eng, A 1989;117:95–102.
Ministry of Science, Research and the Arts of the State of Baden- [12] Kanemaru T, Kawagoishi N, Kondo E, Wang QY, Ohzono Y. influence of grain size
Wurttemberg (Ministerium für Wissenschaft, Forschung und Kunst on notch sensitivities in fatigue of carbon steel, in ‘advances in fracture and
Baden-Württemberg) within the sustainability support (Nachhaltig­ damage mechanics VII’. Trans Tech Publications; 2008. p. 197–200.
[13] Sharifi H, Salehi M, Saeri M. The effect of the ferrite grain size on the fatigue
keitsförderung) of the projects of the Exzellenzinitiative II. behavior of ferrite-martensite dual phase steels. Int J Iron Steel Soc Iran 2014;2:
11–6.
[14] Mlikota M, Schmauder S, Božić Ž. Calculation of the Wöhler (S-N) curve using a
References
two-scale model. Int J Fatigue 2018;114:289–97.
[15] Tanaka K, Mura T. A dislocation model for fatigue crack initiation. J Appl Mech
[1] Alzubi F, Timko M, Li YJ, Toal R, Tovalin K, Es-Said OS. Large versus small grain 1981;48:97–103.
sizes on fatigue life of Aluminum aircraft wheels. Defect Diffus Forum 2019;391: [16] Tanaka K, Mura T. A theory of fatigue crack initiation at inclusions. Metall Trans A
174–94. 1982;13(1):117–23.
[2] Taylor D, Knott JF. Fatigue crack propagation behaviour of short cracks; the effect [17] Fatemi A, Zeng Z, Plaseied A. Fatigue behavior and life prediction of notched
of microstructure. Fatigue Engng Muter Struct 1981;4(2):147–55. specimens made of QT and forged microalloyed steels. Int J Fatigue 2004;26:
[3] Akinawa Y, Tanaka K, Matsui E. Statistical characteristics of propagation of small 663–72.
cracks in smooth specimens of aluminium alloy 2024–T3. Muter Sci Engng 1988; [18] Mlikota M, Schmauder S. On the critical resolved shear stress and its importance in
104:105–15. the fatigue performance of steels and other metals with different crystallographic
[4] Abuzaid WZ, Sangid MD, Carroll JD, Sehitoglu H, Lambros J. Slip transfer and structures. Metals 2017;8(11):883. https://doi.org/10.3390/met8110883.
plastic strain accumulation across grain boundaries in Hastelloy X. J Mech Phys
Solids 2012;60(6):1201–20.

4
M. Mlikota et al. International Journal of Fatigue 158 (2022) 106562

[19] Mlikota M, Schmauder S. Virtual testing of plasticity effects on fatigue crack [23] Božić Ž, Schmauder S, Mlikota M, Hummel M. Multiscale fatigue crack growth
initiation. In: Zingoni A, editor. Advances in engineering materials, structures and modeling for welded stiffened panels. In: Schmauder S, Chen C-S, Chawla KK,
systems: innovations, mechanics and applications. London, GB: CRC Press; 2019. p. Chawla N, Chen W, Kagawa Y, editors. Handbook of Mechanics of Materials.
587–92. Singapore: Springer Singapore; 2018. p. 1–21. https://doi.org/10.1007/978-981-
[20] Mlikota M, Staib S, Schmauder S, Božić Ž. Numerical determination of Paris law 10-6855-3_73-1.
constants for carbon steel using a two-scale model. J Phys Conf Ser 2017;843: [24] Božić Ž, Mlikota M, Schmauder S. Application of the DK, DJ and DCTOD
012042. https://doi.org/10.1088/1742-6596/843/1/012042. parameters in fatigue crack growth modelling. Tech Gaz 2011;18:459–66.
[21] Mlikota M, Schmauder S, Božić Ž, Hummel M. Modelling of overload effects on [25] Božić Ž, Schmauder S, Mlikota M. Fatigue growth models for multiple long cracks
fatigue crack initiation in case of carbon steel. Fatigue Fract Eng Mater Struct 2017; in plates under cyclic tension based on DKI, DJ-integral and DCTOD parameter.
40:1182–90. Key Eng Mater 2011;488–489:525–8.
[22] Mlikota M, Schmauder S. Numerical determination of component Wöhler curve. [26] Mirzazadeh MM, Plumtree A. High cycle fatigue behavior of shotpeened steels.
DVM Bericht 2017;1684:111–24. Metall Mater Trans A 2012;43(8):2777–84.
[27] ABAQUS Analysis user’s guide, vol. 3.