International Journal of Fatigue 145 (2021) 106113

Contents lists available at ScienceDirect

International Journal of Fatigue

journal homepage: [Link]/locate/ijfatigue

On the effect of internal friction on torsional and axial cyclic loading

M. Mehdizadeh, Ali Haghshenas, M.M. Khonsari *
Louisiana State University Department of Mechanical and Industrial Engineering, 3283 Patrick Taylor Hall Baton Rouge, LA 70803, USA

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

Keywords: Failure of materials subjected to torsional and axial cyclic loading is studied by applying the principles of
Internal friction thermodynamics of dissipative processes. Extensive experimental tests are conducted using stainless steel (SS)
Fracture fatigue entropy 304 and SS316 materials to gain insight into both torsional and tension–compression fatigue life. It is shown that
Dissipated energy
the higher magnitude of dissipated energy in torsional loading, as reported in many studies, is due to the
Torsional loading
Fatigue life
anelastic deformation in the material and its associated internal friction. It is further shown that proper
assessment of the damaging dissipation energy requires one to exclude the non-damaging internal friction from
the total amount of energy dissipation. Accumulation of thermodynamic entropy generation as an index
parameter using the modified dissipated energy is estimated to predict the fatigue life. The experimental results
are found to be in good agreement with the proposed fatigue life prediction.

1. Introduction τa,max + kσn,max = constant (1)

where k is a constant and can be obtained by a set of experimental

Fatigue is a weakening or failure mechanism that occurs in all me­ data.
chanical components such as metallic parts, composites, concrete, etc., Brown and Miller [14] proposed a strain-based method based on the
subjected to cyclic stress [1–3]. It involves a very complex interaction of maximum shear strain (γmax ) and maximum normal strain (∊n ) on the
different types of loading (steady, variable, uniaxial, or multiaxial, critical plane (maximum shear strain plane), described by the following
including torsion, tension–compression, and bending), all affected by relationship.
environmental conditions (e.g., temperature and corrosion).
A common method to correlate uniaxial and multiaxial stresses is γ max + f (∊n ) = constant (2)
based on the equivalent stress or strain known as the von Mises criterion Yet another approach is based on implementing a scaling factor
[4–6]. However, the findings of many experimental tests associated with relating to normal and shear strain energies [6,15–17]. An example of
torsional fatigue reveal that results are not in good agreement with the this approach is due to Garud [6], who developed an energy-based
von Mises criterion [7–10]. For example, Blass and Zamrik [11] applied model in terms of the shear and normal energy weights. He proposed
the von Mises criterion to biaxial (tension–compression and torsion), Eq. (3) with an empirical weighting factor for shear energy as:
pure uniaxial, and pure torsion experiments with stainless steel 304 and ( )
showed that under the same equivalent stress, the pure torsional fatigue ΔεΔσ + CΔγΔτ = f Nf (3)
life was greater than that of pure uniaxial loading. They stated that the
where Nf represents the number of cycles to failure, C is the
difference was associated with the influence of the stress state on
weighting factor applied on shear plastic strain energy, Δε and Δσ are
ductility.
the axial strain and stress ranges, and Δγ and Δτ denote for torsional
Findley [12] used a stress-based linear relationship between the
strain and stress ranges, respectively. Here, the different contributions of
maximum shear stress, τa,max , and the maximum normal stress,σ n,max
the torsional and the uniaxial plastic strain energy on the fatigue
(based on the critical plane) to relate the uniaxial and torsional stresses
degradation must be quantified experimentally. For instance, having
(Eq. (1)). The critical plane refers to a single plane that experiences the
performed a set of experimental measurements, Garud [6] reported that
maximum damage defined by the criterion on which the initial crack
implementing the weighting factor of C = 0.5 yields good predictions of
initiates, propagates, and grows until the final fracture [13]. Accord­
fatigue life with cyclic multiaxial stress loading for 1% Cr Mo V steel.
ingly, Findley proposed the following equation.

* Corresponding author.
E-mail address: khonsari@[Link] (M.M. Khonsari).

[Link]
Received 11 November 2020; Received in revised form 18 December 2020; Accepted 21 December 2020
Available online 24 December 2020
0142-1123/© 2020 Published by Elsevier Ltd.
M. Mehdizadeh et al. International Journal of Fatigue 145 (2021) 106113

Other reported experimental results show that the empirical scaling the thermoelastic effect that happens due to the cyclic loading and is
factor varies significantly based on the materials tested. For example, proportional to the local principal stresses [28]. Research shows the
Glinka et al. [18] and Liu [19] reported that the shear and normal en­ thermoelastic effect (temperature fluctuation) increases with increasing
ergies have an equal effect on the material loading fatigue damage. the loading rate and frequency [29].
Experiment results of Tipton [20] on steel 1045 revealed that a According to Maquin and Pierron [30], there are different mecha­
weighting factorofC = 0.9 yielded the best correlation between the pure nisms responsible for heat dissipation in metals. Within the context of
torsional and pure axial fatigue results. Chu et al. [21] scaled the shear this study, it is important to relate temperature measurements to the
energy by a weighting factor of C = 2 to obtain an acceptable correlation mechanisms that are responsible for fatigue damage and eventual fail­
with fatigue data. Skibicki et al. [22] applied Garud model to CuZn37 ure. Based on fatigue damage theories, the irreversible part of micro­
and obtained C = 0.75, by least-squares fitting of all ten­ structure deformation caused by dislocations movement and the
sion–compression (T-C) and torsional data. Similar methods were used associated dissipated energy is of interest when modeling fatigue failure
by other researchers [23–26] for predicting the contribution of torsional [30–32]. Another form of the dissipation energy exists due to the so-
and axial energy in fatigue damage. called internal friction that is related to local defects in the lattice (i.
All the reports mentioned above utilize an experimental approach to e., interstitial atoms diffusion in the Snoek effect and substitution in the
obtain a scaling factor for predicting the fatigue life of components Zener effect) and the oscillation of dislocation, atoms, and grain
under different types of loading, such as tension–compression and tor­ boundary in their equilibrium positions [30,33,34]. The kinetic energy
sion. They experimentally found that the energy dissipation of the of atoms increases due to acoustic wave generation as soon as the atoms
torsional fatigue test is normally higher compared to the ten­ go back to their equilibrium position. This process generates heat
sion–compression one. However, the physical reason for this phenom­ dissipation that causes an increase in the specimen’s surface tempera­
enon was not provided. This approach also requires performing ture under fatigue loading [30]. Note that the latter is recognized to be
extensive experiments to obtain the appropriate scaling factor for each non-damaging, thus must be excluded from the total measure dissipated
material. In this paper, a new model is proposed to predict fatigue life energy to accurately assess the energy dissipation that responsible for
using the dissipated energy during cyclic loading based on excluding the fatigue damage.
non-damaging portion of dissipated energy, which is associated with the In low-cycle fatigue, internal friction is negligible since the micro­
internal friction, which is the main reason for higher energy dissipation plastic deformations and the dissipated energy due to them are dormant.
of the cyclic torsion loading in comparison with the ten­ However, in high cycle fatigue, the non-damaging portion becomes
sion–compression experiment. By excluding the dissipated energy due to significant due to the presence of anelastic oscillations.
the internal friction from total dissipated energy, the uniaxial and Conventionally in high-cycle fatigue when the operating stress is
torsional dissipated energies due to plastic deformation are calculated. below the material’s endurance limit, the specimen’s life is considered
To predict fatigue life, we apply the concept of fracture fatigue entropy to be indefinite. Strictly speaking, a true endurance limit should be
(FFE). associated with the stress levels below which no microstructural changes
The outline of the work is as follows. In Section 2, the theory behind are present. However, adopting the conventional standard, in this study,
the proposed model and formulas for the estimation of fracture fatigue endurance limit is taken as the stress below which failure does not occur
entropy (FFE) is presented. Section 3 provides the details of the exper­ up to 107 cycles [35–37]. The dissipated energy below the endurance
imental procedure as well as the utilized apparatus. The experimental limit is only due to the internal friction-related movements,QIF , and
results are presented, explained, and compared in Section 4. The po­ there is no significant dissipated energy due to plastic deformation, QP .
tential application of the study is discussed in Section 5, followed by Above the endurance limit, however, the total dissipated heat, Qt , results
concluding remarks. from both the microplastic deformations and the internal friction
mechanisms. Note that internal friction in metals varies in an approxi­
2. Theory mately linear fashion with the applied stress/strain [31,38–41]. There­
fore, the dissipated energy due to plastic deformation that is responsible
Research shows the surface temperature of metallic specimens rises for fatigue failure can be obtained as follow:
under cyclic loading and includes three different phases, as shown in {
Fig. 1. In phase I, metal experiences an initial rapid increase due to QP ≅ 0 σ < σ EL
(4)
QP ≅ Qt − QIF σ ≥ σ EL
dislocation density increase and multiplication. In phase II, the surface
temperature reaches a steady-state, and the specimen spends most of its
where σ and σ EL are applied stress and material’s endurance limit. To
fatigue life here. Phase III includes a rapid rise in temperature following
calculate the dissipated energy in cyclic loading, one can apply the first
the formation of cracks, and this continues until failure. According to
law of thermodynamics stated as follows.
previous research, the initial slope of temperature at the beginning of
phase I,Rθ |t=0 , can be used to estimate the dissipated energy under fa­ Ėin + Ėgen = U̇ + Ėout (5)
tigue [27]. Note that the temperature fluctuation in Fig. 1 is known as
where Ėgen is the internally generated inelastic energy and since this
energy appears due to work on the system, then Ėgen = Qt . The param­
eters Ėin and Ėout represent the incoming and outgoing energies to the
boundaries of the control volume, respectively. The outgoing energy is
due to convection, radiation, and heat transfer within the body is by
conduction. At the beginning of phase I, radiation is negligible and the
heat loss to the ambient is by convection only. Thus, assuming that the
lumped system applies, the convection heat transfer can be described by
the following equation [39,42].
A
Qconv = h (T − T∞ ) (6)
V

where T is the specimen surface temperature, h is the heat transfer co­

Fig. 1. The schematic of temperature evolution of a specimen under cy­ efficient, T∞ represents the ambient temperature and V and A are the
clic loading.

2
M. Mehdizadeh et al. International Journal of Fatigue 145 (2021) 106113

volume and the surface area of the specimen gauge section, respectively. described by Eq. (15), is a property of a material, independent of the
The variation of internal energy in Eq. (5) is denoted by U̇ and is defined specimen geometry, the frequency, the type, and the amplitude of cyclic
by Eq. (7). loading [62,63].

dT ∫ tf ( )
(7) Q̇p
U̇ = ρcp
dt FFE = dt (15)
0 T

where dT
dt is time–temperature gradient, cp and ρ represent the specific In general, a major portion of a specimen’s life is in steady-state
heat capacity and density of the material, respectively. Hence, Eq. (5) is where the plastic strain energy per cycle is nearly constant [64–67],
rewritten in the following form: and the specimen’s surface temperature remains approximately
invariant in this period. Therefore, Eq. (15) can be simplified to:
A dT
Qt − h (T − T∞ ) = ρcp (8)
V dt Nf Q p
FFE ≅ (16)
SincedT= Rθ|t=0 is the initial slope of temperature rise at the begin­ fTs
dt
ning of phase I (T = T∞ ), the average dissipated energy during the fa­
whereNf represents the fatigue life, Ts denotes for the steady-state
tigue life,Qt , is as in Eq. (9):
temperature, and f is the applied frequency.
Qt = QP +QIF = ρcp Rθ |t=0 (9)
3. Experimental
where QP and QIF denote the plastic and internal friction-related por­
tions of dissipated energy. The validity and details of the mentioned 3.1. Materials and specimen preparation
technique to obtain dissipated energy during fatigue life are given in our
previous works [43,44]. In another study, Jang and Khonsari showed A series of uniaxial and torsional fatigue tests are performed on the
that dissipation energy could be obtained analytically through one- specimens prepared from stainless steel 304 and 316. The specimens are
dimensional heat equations with the same initial slope of temperature prepared according to ASTM STP 566 and ASTM: E466-07 for torsional
[40]. As discussed in previous works [43,44], the initial slope of tem­ and axial fatigue experiments, respectively. Fig. 2 shows the dimensions
perature rise contains two portions of non-damaging and damaging as of the specimens used in torsional and uniaxial experiments. Tables 1
defined in Eq. (10). and 2 show the mechanical properties and chemical compositions of
stainless steel 304 and 316, respectively. To remove the effect of the
Rpθ |t=0 = Rθ |t=0 − RIfθ |t=0 (10)
surface roughness on fatigue life, all specimens’ gage sections are pol­
where RIfθ |t=0
is the initial temperature rise due to internal friction ished with different grit sizes sandpaper to reach a surface roughness
and Rpθ |t=0 is the irreversible or plastic portion. Consequently, the energy equal to Ra = 0.2μm.
dissipation due to the irreversible deformation (QP ) and the reversible
portion (QIF ) can be obtained as follow: 3.2. Equipment and test procedure
{
QIF ≅ Qt ≅ ρcp Rθ |t=0 σ < σEL
(11) Fig. 3 shows the schematic of the experimental setup (LFE-150) for
QIF ≅ ρcp RIfθ |t=0 σ ≥ σEL
cyclic torsional fatigue experiments (Fatigue Dynamics, Michigan, USA).
{ This test rig is a compact, bench-mounted machine that can apply fully-
QP ≅ 0 σ < σ EL reversed fatigue loading to the specimen with variable load and fre­
(12)
QP ≅ ρcp Rpθ |t=0 σ ≥ σ EL quency. The specimen is clamped on both sides, with one end remains

Note that many researchers have shown that by increasing the

temperature, the amount of internal friction of metals increases [45,46].
It is worth mentioning that Eqs. (11) and (12) are derived based on the
lumped system analysis, which applies when the Biot number is less than
0.1. In both cases of T-C and torsion specimens, the Biot number is below
0.1 [47,48].
The statement of the second law, according to the Clausius–Duhem
inequality is [49]:
1 1 gradT
Ṡi = σ : ε˙p − Ak V̇k − q. 2 ≥ 0 (13)
T T T

where Si represents the entropy generation rate, T is temperature, σ and

ε˙p stand for the plastic stress and strain rate, Ak denotes the thermody­
namic conjugate force corresponding to the internal state variable Vk
(such as damage and hardening), and q represents the heat flux.
Research shows that the contribution of the internal variable in the
entropy generation rate (Ak V˙k ) is about 5–10% and is often assumed to
be negligible [49–51]. Further, in the case of cyclic loading, the entropy
rate generated by conduction heat transfer (T12 [Link]) is negligibly
small [52]. Therefore, Eq. (13) simply reduces to [44,53,54]:

1 Q̇
Si = σ : ε˙p = p (14)
T T
According to recent papers [55–61], the accumulated entropy gen­
Fig. 2. Specimen design and dimensions for (a) Tension-compression fatigue
eration during a fatigue experiment up to failure (tf ), called FFE experiment (b) torsional fatigue experiment (units are millimeters).

3
M. Mehdizadeh et al. International Journal of Fatigue 145 (2021) 106113

Table 1
The properties of stainless steel 304 and 316 are used in the experiments.
Material Physical properties Mechanical properties Thermal properties

Density Elastic Modulus Shear Modulus Tensile strength, ultimate Tensile strength, yield Specific heat Capacity Thermal conductivity

Kg/m3 MPa GPa MPa MPa J/kg K W/m K

SS 304 8000 197 80 505 215 500 16.2

SS 316 8000 193 74 580 290 500 16.2

Table 2
The chemical composition of stainless steel 304 and 316 are used in the experiments.
Material C Cr Mn Mo Si S Ni N P Fe

SS 304 0.08 18–20 2 – 0.75 0.03 8–10.5 0.1 0.045 Balance

SS 316 0.08 16–18 2 2–3 0.75 0.03 10–14 0.1 0.045 Balance

of R = − 1. To record the temperature of the specimen surface during

fatigue tests, an IR-camera with a temperature range of 0 − 500◦ C,
reading accuracy of 2 percent, and resolution of 320 × 240 pixels is
used. The camera is connected to a computer with data acquisition
ability. The specimen’s test section is painted black to increase the
emissivity and reduce reflection during the recording of surface tem­
perature, Fig. 5 illustrates the thermal image of the torsional and uni­
axial SS 316 specimens under cyclic loading captured by an infrared
camera. Note that the photos are captured during the steady-state phase.
As these figures also show, the maximum temperature rise happens in
the middle of the gauge section in both cases of torsion and axial
specimens.

4. Results and discussions

Fig. 3. Schematic of torsional fatigue machine.

4.1. Plastic strain energy

stationary and the other oscillates back and forth by a torsional actuator.
The apparatus provides adjustable twist angles up to ±11 degrees at a
Research shows that crack initiation is associated with the localized
wide range of frequencies up to 50 Hz.
plastic deformation in persistent slip band for both low- and high-cycle
Fig. 4 presents the experimental setup for the uniaxial ten­
fatigue regimes [68–70]. The persistent slip bands and the maximum
sion–compression servo-hydraulic fatigue machine (TestResoures, Min­
shear strain tend to nearly align themselves in the same direction. Re­
nesota, USA). The machine can apply up to 50 kN axial loading up with
ported observation of fatigue cracks also reveals that they tend to
the user-defined frequency up to a maximum of 75 Hz. In this apparatus,
initiate on the maximum shear plane for both torsion and uniaxial fa­
the specimen vertically gripped. The top grip is fixed, and the bottom
tigue experiments [68–73]. Therefore, the crack initiation is generally
one oscillates vertically to apply the specified load and frequency.
controlled by the maximum shear strain mechanism. In the context of
An extensive set of experiments is conducted under the torsional and
reversible and irreversible deformations, plastic shear deformations
uniaxial fatigue load at the applied frequency of 10 Hz and the load ratio
occur if the existing shear strain surpasses the local critical resolved
shear stress (CRSS). Otherwise, if it stays below CRSS, it causes anelastic
deformations where its portion of dissipated energy is directly related to
If
Rθ |t=0 , as represented by Eq. (11).
Fig. 6 shows that internal friction is dominant when stress is below
If
the endurance limit. Above this limit, Rθ |t=0 increases linearly with
increasing the stress/strain amplitude. The experiments show the SS 304
and SS 316 specimens under cyclic loading for the stresses lower than
240–250 MPa for tension–compression experiment and shear strain
lower than 0.03 mm
mm for torsional experiments did not lead to the failure
for 3 × 106 cycles. Therefore, they can be considered as the endurance
limits for these materials [74,75]. Hence, the dissipated energy associ­
ated with internal friction exhibits the same characteristics as defined by
Eqs. (11) and (12). These results confirm the fact that the energy dissi­
pation in torsional loading is higher than that of the uniaxial
tension–compression.
Fig. 7 shows the temperature evolution of T-C and torsional fatigue in
SS 304. As the figure illustrates, for the same fatigue life, the steady-state
temperature of torsion is much higher than T-C loading. This leads to
higher total energy dissipation and an increase in the amount of internal
friction [45,46].
Fig. 4. Hydraulic fatigue machine used for tension–compression fatigue The dissipated energy per cycle, Qt , increases if the applied stress or
experiment with IR-camera.

4
M. Mehdizadeh et al. International Journal of Fatigue 145 (2021) 106113

Fig. 5. Captured thermal image of the specimen under cyclic loading via IR-camera on SS 316 at a testing frequency of 10 Hz (a) torsional test under of 370 MPa and
(b) uniaxial test under 340 MPa (temperature unit is Centigrade).

Fig. 6. The variation of initial slope of temperature as a function of applied stress (σ) and applied shear strain (γ) for different materials under T-C and
torsional loading.

strain is increased. Fig. 8 shows the dependence of the dissipated energy Further insight can be gained by recalling that, as discussed in Sec­
per cycle as a function of the number of cycles to failure,Nf , for SS 304 tion 2, the portion of the dissipated energy due to internal friction is non-
and SS 316. In these figures, Qt is calculated using Eq. (7) for both the damaging and must be excluded to obtain the damaging dissipated en­
pure torsion and tension–compression fatigue experiments. As these ergy for further calculations. Fig. 9 shows the variation of dissipated
figures show, all the results lay on two separate lines. The upper line for energy due to plastic deformations vs. the number of life cycles. It can be
both materials is associated with torsion, and the lower line pertains to seen that both the uniaxial and torsional dissipated energies due to
tension–compression. irreversible deformations, QP , using Eq. (12), now follow the same slope
The fact that the magnitude of Qt in torsional fatigue tests is higher for different materials. As this figure shows, for the same fatigue life, the
than that of the uniaxial tension–compression has also been documented energy dissipation of SS 316 is higher than SS 304 that leads to the
in many reports [6,20,70,73,76]. In this investigation, we show that this higher value of fracture fatigue entropy of SS 316 compared with SS 304
phenomenon is associated with internal friction, as described by Fig. 6. (based on Eq. (16)).
In Fig. 8, the separation of the line associated with the torsional dissi­
pated energy and the uniaxial dissipated energy with an increase in life
is due to the fact that internal friction becomes more significant with 4.2. Fatigue life prediction
decreasing the cyclic load amplitude or approaching the high-cycle fa­
tigue regime. Using the modified value of dissipated energy and using Eq. (16), the
fracture fatigue entropy is calculated for both SS 304 and SS 316 in the

5
M. Mehdizadeh et al. International Journal of Fatigue 145 (2021) 106113

Using Fig. 6 b, the value of the initial slope of temperature due to the
internal friction can be estimated and Rpθ |t=0 = 0.86 sC. Then, having the
◦

properties of SS 316, including FFE = 74.9 mMJ 3 K, ρ = 8000 and cp = 500

and also using Eqs. (12) and (16), the fatigue life of the mentioned
specimen under torsion load is:

Fig. 7. Temperature evolution of SS 304 under T-C and torsional fatigue

loading in an approximately the same fatigue life.

case of pure uniaxial and pure torsional fatigue experiments. Fig. 10

shows the FFE as a function of fatigue life. The mean values of FFE for SS
304 and SS 316 are 37.3 mMJ 3 K and 74.9 m3 K, respectively. Referring to
MJ

Table 1, the main difference between SS 304 and SS 316 is in their yield
and ultimate strength. Therefore, the higher fatigue life of SS 316 can be
associated with these properties. Similarly, the total entropy dissipation
up to failure is larger for SS 316, and its FFE value is greater than that of Fig. 9. Experimental results for dissipated energy due to the plastic deforma­
SS 304. In general, the higher the FFE, the longer is the fatigue life. tion as a function of the number of cycles to failure for SS 304 and SS 316.
Examination of the results of Fig. 10 further establishes that FFE is a
unique material constant.
Having the mean value of fracture fatigue entropy, we next proceed
to predict fatigue life. Results for pure torsion and pure ten­
sion–compression are shown in Fig. 11. As shown, the calculated fatigue
lives are in good agreement with the experimental results. The
maximum error in predicted fatigue life based on the proposed tech­
nique is less than ±%15 and ±%13 for T-C and torsion experiments,
respectively.

4.3. An illustrative example

A specimen made from SS 316 with mechanical and chemical

properties shown in Tables 1 and 2 undergoes torsional loading at room
temperature. The specimen is subjected to a stress level of 385 MPa at
the testing frequency of 10 Hz. At the onset of the experiment, the initial
slope of temperature and the steady-state temperature is measured
Fig. 10. Fracture fatigue entropy as a function of fatigue life for SS 304 and
(Rθ |t=0 = 1.06 sC and TS = 182 ◦ C.). We wish to predict the specimen’s
◦

SS 316.
fatigue life of this specimen.

Fig. 8. Dissipated energy as a function of the number of cycles to failure (a) SS304 (b) SS 316.

6
M. Mehdizadeh et al. International Journal of Fatigue 145 (2021) 106113

[3] Golewski GL, Sadowski T. A study of mode III fracture toughness in young and
mature concrete with fly ash additive. Solid State Phenomena. Trans Tech Publ;
2016.
[4] Ellison E, Andrews J. Biaxial cyclic high-strain fatigue of aluminum alloy RR58.
J Strain Anal 1973;8(3):209–19.
[5] Meurer H, Hanswillemenke H, Breitling H. Biaxial Fatigue Tests on Thin Walled
Tubes of NiCr23Col2Mo (Inconel 617) at 950◦ C. ICBMFF3. 1991.
[6] Garud Y. A new approach to the evaluation of fatigue under multiaxial loadings.
J Eng Mater Technol 1981;103(2):118–25.
[7] Parsons MW, Pascoe K. Development of a biaxial fatigue testing rig. J Strain Anal
1975;10(1):1–9.
[8] Manson, S. and G. Halford, Discussion:“Multiaxial Low-Cycle Fatigue of Type 304
Stainless Steel”(Blass, JJ, and Zamrik, SY, 1976 Winter Annual Meeting). journal of
Engineering Materials and Technology, 1977. 99(3): p. 283-285.
[9] Sakane M, Ohnami M, Sawada M. Fracture Modes and Low Cycle Biaxial Fatigue at
Elevated Temperature: Transactions of ASME. J Pressure Vessel Technol 1987;109:
236–43.
[10] Shatil G, Smith D, Ellison E. High strain biaxial fatigue of a structural steel. Fatigue
Fract Eng Mater Struct 1994;17(2):159–70.
[11] Blass J, Zamrik S. Multiaxial low-cycle fatigue of type 304 stainless steel. In: Creep-
Fatigue Interaction. ASME New York. 1976; 1976. p. 129–59.
Fig. 11. Predicted fatigue lives in comparison with the experimental results. [12] Findley WN. A theory for the effect of mean stress on fatigue of metals under
combined torsion and axial load or bending. J Eng Industry 1959;81(4):301–5.
MJ [13] Jirandehi AP, Chakherlou T. A fatigue crack initiation and growth life estimation
QP = Rpθ |t=0 ρcq = 0.86 × 8000 × 500 = 3.44 method in single-bolted connections. J Strain Anal Eng Des 2019;54(2):79–94.
m3 s [14] Brown MW, Miller K. A theory for fatigue failure under multiaxial stress-strain
conditions. Proc Inst Mech Eng 1973;187(1):745–55.
Nf Q p FFE fTs 74.9 × 10 × (182 + 273.15) [15] Jordan, E.H., M.W. Brown, and K.J. Miller, Fatigue under severe nonproportional
FFE ≅ →Nf ≅ = loading, in Multiaxial fatigue. 1985, ASTM International.
fTs Qp 3.44
[16] Ellyin F, Kujawski D. Plastic strain energy in fatigue failure. J Pressure Vessel
≅ 99, 100 cycles Technol 1984;106(4):342–7.
[17] Socie D. Multiaxial fatigue damage models. J Eng Mater Technol 1987;109(4):
293–8.
4.4. Experimental verification [18] Glinka G, Shen G, Plumtree A. A multiaxial fatigue strain energy density parameter
related to the critical fracture plane. Fatigue Fract Eng Mater Struct 1995;18(1):
37–46.
To validate the above prediction, we performed two additional fa­ [19] Liu K. A method based on virtual strain-energy parameters for multiaxial fatigue
tigue tests under the same condition (σ ≈ 385 MPa & f = 10 Hz). The life prediction, in Advances in multiaxial fatigue. ASTM International; 1993.
[20] Tipton SM. Fatigue behavior under multiaxial loading in the presence of a notch:
first specimen fractured after 106,700 and the second one after 111,200 methodologies for the prediction of life to crack initiation and life spent in crack
cycles. The related discrepancies between the experimental and pre­ propagation (shoulder fillet, plastic strain. prediction) 1986.
diction tests are 7.1% and 10.9% for the first and second specimens, [21] Chu C-C, Conle FA, Bonnen JJ. Multiaxial stress-strain modeling and fatigue life
prediction of SAE axle shafts, in Advances in multiaxial fatigue. ASTM
respectively. International; 1993.
[22] Skibicki D, Pejkowski Ł. Low-cycle multiaxial fatigue behaviour and fatigue life
5. Conclusions prediction for CuZn37 brass using the stress-strain models. Int J Fatigue 2017;102:
18–36.
[23] Lee B, Kim K, Nam K. Fatigue analysis under variable amplitude loading using an
In this study, a thermodynamic analysis of materials subjected to energy parameter. Int J Fatigue 2003;25(7):621–31.
cyclic loading is performed to investigate the effect of loading type, [24] Sugiura K, Chang K-C, Lee G. Evaluation of low-cycle fatigue strength of structural
metals. J Eng Mech 1991;117(10):2373–83.
including torsion and tension–compression. It is observed that for the [25] Shahrooi S, Metselaar I, Huda Z. Evaluating a strain energy fatigue method using
specimen undergoing torsion, the dissipated energy is higher than the cyclic plasticity models. Fatigue Fract Eng Mater Struct 2010;33(8):530–7.
tension–compression. It is found that the fatigue life can be made to be [26] Han C, Chen X, Kim K. Evaluation of multiaxial fatigue criteria under irregular
loading. Int J Fatigue 2002;24(9):913–22.
independent of the loading type by excluding the effect of internal
[27] Amiri M, Khonsari M. Rapid determination of fatigue failure based on temperature
friction that is non-damaging and the subsequent dissipated energy due evolution: Fully reversed bending load. Int J Fatigue 2010;32(2):382–9.
to anelastic deformations and internal friction mechanism. Then using [28] Cavaliere P, et al. Thermoelasticity for the evaluation of fatigue behavior of 7005/
the concept of fracture fatigue entropy, the FFE is used as an index Al2O3/10p metal matrix composite sheets joined by FSW. Int J Fatigue 2008;30
(1):198–206.
parameter to predict fatigue life independent of loading type. The mean [29] Lee H, Chen J, Wang J. Thermomechanical behaviour of metals in cyclic loading.
value of FFE are 37 ± 3 and 74±5 mMJ 3 K in the cases of SS304 and SS316
J Mater Sci 1993;28(20):5500–7.
[30] Maquin F, Pierron F. Heat dissipation measurements in low stress cyclic loading of
specimens, respectively. Finally, a set of T-C and torsion experiments are metallic materials: From internal friction to micro-plasticity. Mech Mater 2009;41
performed on specimens made of SS 304 and SS 316 to check the validity (8):928–42.
of the proposed methodology. The predicted fatigue lives utilizing this [31] Guo Q, Guo X. Research on high-cycle fatigue behavior of FV520B stainless steel
based on intrinsic dissipation. Mater Des 2016;90:248–55.
criterion compared with the experimental results in pure torsion (the
[32] Risitano A, Risitano G. Cumulative damage evaluation in multiple cycle fatigue
error is less than ±%13) and pure tension–compression (the error is less tests taking into account energy parameters. Int J Fatigue 2013;48:214–22.
than ±%15) show good agreement. [33] Caillard, D. and J.-L. Martin, Thermally activated mechanisms in crystal plasticity.
2003: Elsevier.
[34] Slimani, A., P. Fleischmann, and R. Fougères, Dislocation dynamic in aluminium
Declaration of Competing Interest polycrystals during cyclic plasticity studied by acoustic emission. Journal de
Physique. 3, 1992. 2(6): p. 933-945.
[35] Shiozawa K, Lu L, Ishihara S. S-N curve characteristics and subsurface crack
The authors declare that they have no known competing financial initiation behaviour in ultra-long life fatigue of a high carbon-chromium bearing
interests or personal relationships that could have appeared to influence steel. Fatigue Fract Eng Mater Struct 2001;24(12):781–90.
[36] Murakami Y, Yokoyama N, Nagata J. Mechanism of fatigue failure in ultralong life
the work reported in this paper.
regime. Fatigue Fract Eng Mater Struct 2002;25(8–9):735–46.
[37] Bathias C. There is no infinite fatigue life in metallic materials. Fatigue Fract Eng
References Mater Struct (Print) 1999;22(7):559–65.
[38] Guo Q, et al. An energy method for rapid evaluation of high-cycle fatigue
parameters based on intrinsic dissipation. Int J Fatigue 2015;80:136–44.
[1] Sadowski T, Golewski G. A failure analysis of concrete composites incorporating fly
[39] Mehdizadeh M, Khonsari M. On the role of internal friction in low-and high-cycle
ash during torsional loading. Compos Struct 2018;183:527–35.
fatigue. Int J Fatigue 2018.
[2] Naderi M, Khonsari M. Thermodynamic analysis of fatigue failure in a composite
laminate. Mech Mater 2012;46:113–22.

7
M. Mehdizadeh et al. International Journal of Fatigue 145 (2021) 106113

[40] Jang J, Khonsari M. On the evaluation of fracture fatigue entropy. Theor Appl Fract [59] Wang J, Yao Y. An Entropy Based Low-Cycle Fatigue Life Prediction Model for
Mech 2018;96:351–61. Solder Materials. Entropy 2017;19(10):503.
[41] Roebben G, et al. Fatigue resistant silicon nitride ceramics due to anelastic [60] Jang J, Mehdizadeh M, Khonsari M. Nondestructive estimation of remaining
deformation and energy dissipation. J Alloy Compd 2000;310(1–2):39–43. fatigue life without the loading history. Int J Damage Mech 2020;29(3):482–502.
[42] Hajshirmohammadi B, Khonsari M. A simple approach for predicting fatigue crack [61] Lijesh K, Mehdizadeh M, Khonsari MM. Online monitoring of metal fatigue life.
propagation rate based on thermography. Theor Appl Fract Mech 2020;107: Struct Health Monitor 2019. 1475921719871668.
102534. [62] Haghshenas A, Khonsari M. Non-destructive testing and fatigue life prediction at
[43] Mehdizadeh M, Khonsari M. On the role of internal friction in low-and high-cycle different environmental temperatures. Infrared Phys Technol 2019;96:291–7.
fatigue. Int J Fatigue 2018;114:159–66. [63] Jirandehi AP, Mehdizadeh M, Khonsari M. Temperature-induced buckling of
[44] Mehdizadeh M, Khonsari M. On the application of fracture fatigue entropy to ductile metals during cyclic loading and the subsequent early fracture. Int J Mech
variable frequency and loading amplitude. Theor Appl Fract Mech 2018;98:30–7. Sci 2020;176:105525.
[45] Forster F, Koster W. Modulus of Elasticity and Damping as a Function of the [64] Ellyin, F., Fatigue damage, crack growth and life prediction. 2012: Springer
Material State. Z. Metallk 1937;29:116. Science & Business Media.
[46] Zener C. Internal friction in solids. Proc Phys Soc 1940;52(1):152. [65] Meneghetti G. Analysis of the fatigue strength of a stainless steel based on the
[47] Mignard F, et al. Thermal shock behaviour of a coarse grain porous alumina. energy dissipation. Int J Fatigue 2007;29(1):81–94.
J Mater Sci 1996;31(8):2131–8. [66] Wang X, et al. Quantitative Thermographic Methodology for fatigue assessment
[48] Xu B, Li P-W, Chan CL. Extending the validity of lumped capacitance method for and stress measurement. Int J Fatigue 2010;32(12):1970–6.
large Biot number in thermal storage application. Sol Energy 2012;86(6):1709–24. [67] Glodež S, et al. The influence of PVD coating on the low cycle fatigue behaviour of
[49] Lemaitre J, Chaboche J. Mechanics of solid materials. UK: Cambridge University Cr-Mo-V steel at elevated temperatures. Surf Coat Technol 2017;321:358–65.
Press CrossRef Google Scholar; 1990. [68] Kanazawa K, Miller K, Brown M. Low-cycle fatigue under out-of-phase loading
[50] Halford G. The energy required for fatigue(Plastic strain hystersis energy required conditions. J Eng Mater Technol 1977;99(3):222–8.
for fatigue in ferrous and nonferrous metals). J Mater 1966;1:3–18. [69] Taira, S., T. Inoue, and M. Takahashi. Low cycle fatigue under multiaxial stresses
[51] Basaran C, Nie S. An irreversible thermodynamics theory for damage mechanics of (in the case of combined cyclic tension-compression and cyclic torsion in the same
solids. Int J Damage Mech 2004;13(3):205–23. phase at elevated temperature). in Proceedings of the 10th Japan Congress on
[52] Liakat M, Khonsari M. Rapid estimation of fatigue entropy and toughness in metals. Testing Materials. 1967.
Mater Des (1980–2015) 2014;62:149–57. [70] Fatemi A, Socie DF. A critical plane approach to multiaxial fatigue damage
[53] Hajshirmohammadi B, Khonsari M. On the entropy of fatigue crack propagation. including out-of-phase loading. Fatigue Fract Eng Mater Struct 1988;11(3):149–65.
Int J Fatigue 2020;133:105413. [71] Socie, D.F., L.A. Waill, and D.F. Dittmer, Biaxial fatigue of Inconel 718 including
[54] Bin Jamal MN, et al. Low Cycle Fatigue Life Prediction Using Unified Mechanics mean stress effects, in Multiaxial fatigue. 1985, ASTM International.
Theory in Ti-6Al-4V Alloys. Entropy 2020;22(1):24. [72] Pascoe K, De Villiers J. Low cycle fatigue of steels under biaxial straining. Journal
[55] Yao W, Basaran C. Computational damage mechanics of electromigration and of Strain analysis 1967;2(2):117–26.
thermomigration. J Appl Phys 2013;114(10):103708. [73] Brown M, Miller K. High temperature low cycle biaxial fatigue of two steels.
[56] Sun Y-J, Hu L-S. Assessment of low cycle fatigue life of steam turbine rotor based Fatigue Fract Eng Mater Struct 1979;1(2):217–29.
on a thermodynamic approach. J Eng Gas Turbines Power 2012;134(6):064504. [74] Zhang G, Takashima K, Higo Y. Fatigue strength of small-scale type 304 stainless
[57] Imanian A, Modarres M. A thermodynamic entropy approach to reliability steel thin films. Mater Sci Eng, A 2006;426(1–2):95–100.
assessment with applications to corrosion fatigue. Entropy 2015;17(10): [75] Strizak J, Mansur L. The effect of mean stress on the fatigue behavior of 316 LN
6995–7020. stainless steel in air and mercury. J Nucl Mater 2003;318:151–6.
[58] Sosnovskiy LA, Sherbakov SS. Mechanothermodynamic entropy and analysis of [76] Yokobori T, Yamanouchi H, Yamamoto S. Low cycle fatigue of thin-walled hollow
damage state of complex systems. Entropy 2016;18(7):268. cylindrical specimens of mild steel in uni-axial and torsional tests at constant strain
amplitude. Int J FractMech 1965;1(1):3–13.