International Journal of Fatigue 96 (2017) 102–113

Contents lists available at ScienceDirect

International Journal of Fatigue

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

Temperature dependent cyclic deformation and fatigue life

of EN-GJS-600 (ASTM 80-55-06) ductile cast iron
Benjamin Jost, Marcus Klein ⇑, Tilmann Beck, Dietmar Eifler
Institute of Materials Science and Engineering (WKK), University of Kaiserslautern, P.O. Box 3049, 67653 Kaiserslautern, Germany

a r t i c l e i n f o a b s t r a c t

Article history: The ductile cast iron EN-GJS-600 (ASTM 80-55-06) provides several beneficial properties for a wide spec-
Received 28 July 2016 trum of applications. Its near-net shaping capabilities combined with beneficial physical and mechanical
Received in revised form 4 November 2016 properties even at elevated temperatures result in a wide use for components of internal combustion
Accepted 8 November 2016
engines. Under service conditions, fatigue loading at different temperatures occurs. In this context, the
Available online 09 November 2016
present paper focuses on the isothermal high cycle fatigue behavior of ductile cast iron EN-GJS-600 in
the temperature range from ambient temperature up to 400 °C at a frequency of 5 Hz. At elevated tem-
Keywords:
peratures, the consideration of dynamic strain aging effects on cyclic deformation and lifetime behavior is
PhyBaLSIT
Temperature increase test
of major importance. It was proved that only one temperature increase test is necessary to determine the
Strain increase test temperature range of pronounced dynamic strain aging effects. The physically based fatigue lifetime cal-
Ductile cast iron culation approach ‘‘PhyBaLSIT” (SIT = strain increase test) was applied and validated for total strain con-
EN-GJS-600 trolled fatigue tests. To receive maximum information with minimum effort, both test procedures were
Dynamic strain aging combined enabling calculation of fatigue life in the temperature range from ambient temperature up to
Fatigue properties 400 °C based on only thirteen fatigue experiments.
Lifetime calculation Ó 2016 Elsevier Ltd. All rights reserved.
Cyclic deformation behavior

1. Introduction
High fatigue strength, good thermal conductivity and low cost
in mass production enable a wide spectrum of applications of duc-
tile cast iron in the automotive and commercial vehicle technology,
especially for components of internal combustion engines [1].
Application examples for the ductile cast iron EN-GJS-600 are cam-
shafts, bearing covers, crankcases, pistons and especially crank-
shafts [2]. Ongoing efforts to reduce vehicle weight and fuel
consumption lead to advanced design approaches of internal com-
bustion engines [3]. On the one hand, decreasing the number of
cylinders and the engine stroke volume while increasing the speci-
fic power output results in higher component stresses in a wide
temperature range up to 270 °C [4,5]. On the other hand, avoiding
overdimensioning to improve material utilization requires reliable
characterization of the fatigue behavior [6]. In this context, the
influence of dynamic strain aging (DSA) effects on the fatigue
behavior has to be considered to accomplish a weight-optimized
design of such components [7].
In the past years, several investigations on ductile cast iron
materials were performed, e.g. [8–13]. Most of the investigations
⇑ Corresponding author.
E-mail address: klein@mv.uni-kl.de (M. Klein).
on this topic in the past years deal with isothermal LCF
[11,12,14,15], TMF [14–17] and TMF/superimposed HCF
[14,18,19] behavior of ductile cast iron. In [14] the lifetime behav-
ior for isothermal LCF is described by the Smith-Watson-Topper
damage parameter (PSWT). This parameter proved to be too insen-
sitive to appropriately describe the effects of testing temperature.
In [16,17], a fracture mechanics based model assuming a direct
correlation between cyclic crack-tip opening displacement (CTOD)
and crack growth rate, including the effects of creep, relaxation,
Bauschinger-effect and tension-compression asymmetry of cast
irons, is presented and evaluated. With this model, finite-element
simulations to calculate the fatigue life of components under
thermo-mechanical loading are performed to reduce expensive
component tests. In [18,19] this model is extended to superim-
posed HCF-loadings.
The majority of the investigations on ductile cast iron under
total strain controlled fatigue found in literature were performed
at defined temperatures and frequencies. DSA effects are consid-
ered, but the systematic dependence of DSA on deformation rate
and temperature is not completely taken into account.
The overall target of the ongoing investigation is to characterize
the influence of deformation rate and temperature on dynamic
strain aging of ductile cast iron under isothermal testing condi-
tions. Therefore, different testing frequencies (0.005 Hz, 5 Hz,

http://dx.doi.org/10.1016/j.ijfatigue.2016.11.010
0142-1123/Ó 2016 Elsevier Ltd. All rights reserved.
 B. Jost et al. / International Journal of Fatigue 96 (2017) 102–113 103

Nomenclature

AT ambient temperature N number of cycles

BSE back scatter electron n0 M(HCF) cyclic hardening exponent HCF-regime
CAT constant amplitude test n0 M(LCF) cyclic hardening exponent LCF-regime
CATHCF constant amplitude test HCF-regime Nf number of cycles to failure
CATLCF constant amplitude test LCF-regime Ø diameter
DFG German Research Foundation R2 coefficients of determination
DSA dynamic strain aging Re strain ratio
ET elevated temperatures T temperature
exp. experimental values T(DSAmax) temperature of maximum dynamic strain aging
HCF high cycle fatigue DN step length
LCF low cycle fatigue DR change in electrical resistance
PhyBaL Physically Based Lifetime calculation method DT change in temperature
PSWT damage parameter Smith-Watson-Topper Dea,t total strain step
SIT strain increase test e strain
TC thermocouple e0 B fatigue ductility coefficient
TIT temperature increase test e0 B,M(HCF) (e-N) fatigue strength coefficient HCF-regime
TMF thermo mechanical fatigue e0 B,M(LCF) (e-N) fatigue ductility coefficient LCF-regime
b fatigue strength exponent ea,p plastic strain amplitude
bM generalized fatigue strength exponent ea,t total strain amplitude
c fatigue ductility exponent e-N-curve Wöhler curve
cM generalized fatigue ductility exponent Se endurance limit
CStSp cyclic total strain-plastic strain r stress
f frequency ra stress amplitude
HBW Brinell hardness r0 B fatigue strength coefficient
K0 M(HCF) cyclic hardening coefficient HCF-regime
K0 M(LCF) cyclic hardening coefficient LCF-regime

47 Hz and 92 Hz) are applied during total strain controlled fatigue processes. It should be noted that Wöhler (e-N) curves in most
tests in a wide range of elevated temperatures (ET). To enable total cases show a slope change between the LCF- and HCF-regime indi-
strain controlled fatigue including the determination of the plastic cating a change in damage mechanism from microplasticity (HCF)
strain amplitude at frequencies up to 92 Hz, an extensometer with to macroplasticity (LCF). Based on one SIT, the transition between
minimized mass and inertia was developed. LCF- and HCF-regime can be determined which allows to describe
In the present paper, a procedure to determine the temperature either the majority of the HCF- or the LCF-regime by one power
range of DSA and the temperature dependent calculation of Wöhler law.
(e-N) curves with minimal experimental effort is shown for the fre-
quency of 5 Hz with focus on the HCF-regime. Based on the pre-
2. Material
sented procedure, ongoing research deals with the investigation
of frequency influence on DSA.
The ductile cast iron EN-GJS-600 was provided as cylindrical
Light and scanning electron microscopy investigations were
bars with a diameter of 36 mm and a length of 300 mm by Daimler
performed to characterize the graphite-, pearlite- and ferrite frac-
AG, Germany. From each bar two fatigue specimens with the
tion of EN-GJS-600. To describe the fatigue behavior in constant
dimensions shown in Fig. 1 were machined. The gauge length
amplitude (CAT) as well as strain increase tests (SIT) in detail,
was mechanically polished to avoid crack initiation at machining
mechanical stress-strain (r-e) hysteresis loops, deformation
induced surface defects. Fig. 2 shows a light optical micrograph
induced changes in temperature (DT) and electrical resistance
of the investigated EN-GJS-600. The matrix is predominantly pear-
(DR) were measured. At ET, the change in temperature DT cannot
litic with ferrite arranged around the graphite spheroids. The Bri-
be measured because temperature is controlled and not a material
nell hardness as well as the graphite- and ferrite fraction are
reaction. The specific electrical resistance is highly temperature
given in Table 1. Hardness tests were performed at fifteen different
sensitive, therefore a temperature dependent correction would
bars. The graphite and ferrite fraction, respectively, were deter-
be necessary at elevated temperatures. In contrast to that, mea-
mined by image analysis of micrographs covering complete gauge
surement of the plastic strain amplitude ea,p by a mechanical
length sections of four different specimens taken from four differ-
extensometer is temperature independent as long as the exten-
ent bars.
someter temperature is kept sufficiently constant. Therefore, the
plastic strain amplitude ea,p is used as input data for the physically
based lifetime calculation approach ‘‘PhyBaL” at AT and ET. In [20–
22], the PhyBaL method is explained in detail for stress controlled
fatigue at AT. In [23] details of sample machining and the
microstructure of the investigated EN-GJS-600 are given. To deter-
mine the temperature range between onset and maximum of DSA,
only one temperature increase test (TIT) at constant total strain
amplitude is needed. The combination of PhyBaLSIT with one TIT
leads to significant acceleration of fatigue life characterization
and quantification of temperature dependent cyclic deformation Fig. 1. Fatigue specimen, dimensions in mm.
104 B. Jost et al. / International Journal of Fatigue 96 (2017) 102–113
Fig. 2. Light micrograph of the investigated EN-GJS-600.
Table 1
Microstructural parameters of the investigated EN-GJS-
600.
Brinell hardness [HBW 2.5/187.5] 223 ± 4
Graphite fraction [area-%] 9.6 ± 0.4
Ferrite fraction [area-%] 17.6 ± 2.6
3. Experimental setup
Total strain controlled constant amplitude tests, strain increase
tests and temperature increase tests were performed with a strain
ratio Re = 1, a frequency f = 5 Hz and triangular waveform.
According to the institute’s standard experimental setup, the plas-
tic strain amplitude ea,p, the change in temperature DT (only in
tests at ambient temperature (AT)) and the change in electrical
resistance DR were measured continuously. In the present work,
the plastic strain amplitude ea,p is used as input data for lifetime
calculation and characterization of DSA. Total strain amplitude ea,
t was measured by a capacitive extensometer with ceramic rods.
The plastic strain amplitude ea,p is taken from the measured
stress-strain hystereses by calculating half the total strain differ-
ence at mean stress and stored each 3 cycles. Heating was per-
formed by a resistance furnace. Fig. 3 shows the experimental
setup for AT (a) and ET (b). At ET, thermocouple 1 (TC1) was used
for closed-loop temperature control.
Fig. 3. Experimental setup for AT (a) and ET (b) temperatures.
4. PhyBaLSIT
The physically based fatigue life calculation PhyBaLSIT enables
the assessment of Wöhler (e-N) curves for total strain controlled
fatigue in the HCF- and LCF-regime based on fatigue data deter-
mined in strain increase and constant amplitude tests. According
to e.g. [8,24,25], a total strain Wöhler curve shows a slope change
between the HCF- and LCF-regime. While in the HCF-regime
microplastic deformations lead to failure, in the LCF-regime defor-
mation is predominantly macroplastic. The HCF- as well as the LCF-
regime can be approximated by one power law each. Conse-
quently, the macroplastic (microplastic) damage amounts can be
neglected in the majority of HCF- (LCF-) regime and therefore a
two-parameter-concept allows to calculate a large proportion of
the HCF- and LCF-regime, respectively. In the transition between
HCF- and LCF-regime, both, microplastic and macroplastic defor-
mation occurs in relevant amounts and, hence, a four-parameter
approach needs to be used.
To describe the HCF-regime, Basquin’s law [26] (Eq. (1)) was
modified for total strain control (Eq. (2)) while in the LCF-regime
the Manson-Coffin law [27,28] (Eq. (3)) was modified (Eq. (4)),
see Fig. 4c.
ra ¼ r0B ð2Nf Þb ð1Þ
ea;t ¼ e0B;MðHCFÞ ð2Nf ÞbM ð2Þ
ea;p ¼ e0B ð2Nf Þc ð3Þ
ea;t ¼ e0B;MðLCFÞ ð2Nf ÞcM ð4Þ
The fatigue strength coefficient r0 B, describing the position of
the Wöhler curve in the HCF-regime, is changed to e0 B,M(HCF) while
the fatigue strength exponent b, describing the slope of the
Wöhler-curve in the HCF-regime, is changed to bM. In the LCF-
regime, the fatigue ductility coefficient e0 B is changed to e0 B,M(LCF)
and the fatigue ductility exponent c to cM.
Based on the material’s response for each total strain step of the
SIT (Fig. 4a), the endurance limit Se can be estimated as the first
total strain amplitude resulting in significant microplastic
deformation.
By variation of the step length DN and the total strain step Dea,t,
the SIT can be adapted to investigations in the HCF- or LCF-regime.
For example, to investigate the HCF-regime in the present work, a
step length of DN = 9000 cycles and a strain step of Dea,t = 0.01%
were chosen to cover the majority of the HCF-regime. To character-
ize the deformation behavior under LCF conditions, the step length
DN must be shortened and the total strain step Dea,t increased.
Following the principles of Morrow with the modification to
plot the measured values M (e.g. plastic strain amplitude ea,p,
change in temperature DT, change in electrical resistance DR) ver-
sus the control parameter ea,t instead of the stress amplitude ra, a
cyclic total strain-plastic strain (CStSp-)curve can be plotted from
the SIT results (Fig. 4b). Modifying the Morrow equation [29] for
the HCF- (Eq. (5)) and LCF-regime (Eq. (6)), respectively, allows
the description of a CStSp-curve for both regimes. Note that differ-
ent slopes exist in the HCF- and the LCF-regime.
0
ea;t ¼ K0MðHCFÞ ðea;p ÞnMðHCFÞ ð5Þ
n0
ea;t ¼ e
K0MðLCFÞ ð a;p Þ MðLCFÞ ð6Þ
The cyclic hardening coefficients K0 M(HCF), K0 M(LCF) define the
position of the CStSp-curves while the cyclic hardening exponents
n0 M(HCF), n0 M(LCF) describe the slopes. Morrow showed that there is
a relation between the slope of a cyclic hardening curve for con-
stant amplitude loading and the slope of the Wöhler curve [29]
in both, HCF- (Eq. (7)) and LCF-regime (Eq. (8)). For all further work
it is therefore assumed that an ea,t-range of the CStSp-curve which
 B. Jost et al. / International Journal of Fatigue 96 (2017) 102–113 105

Fig. 4. PhyBaLSIT calculation for the LCF- and HCF-regime (schematically).

can be described by one power law corresponds with a portion of

the Wöhler-curve that can also be represented by a single power
law.

n0MðHCFÞ
bM ¼ ð7Þ
5n0MðHCFÞ þ 1
1
cM ¼ ð8Þ
5n0MðLCFÞ þ 1

In the HCF-regime, with the data of two constant amplitude

tests (CATHCF), the CStSp-curve is converted from SIT to constant
amplitude loading by linear interpolation of the mean value of
each strain step. In the LCF-regime ongoing research suggests that
there is, in contrast to the HCF-regime, no difference between the
slope of the CStSp-curve for SIT and CAT. This might be explained
by a fully developed cyclic hardening state even at the lowest total Fig. 5. Results of a strain increase test of EN-GJS-600 at ambient temperature with
strain amplitude in the LCF-regime. reduced step-length (a), cyclic total strain–plastic strain-curve with data deter-
Using Eqs. (7) and (8) for the HCF- and LCF-regime, respectively, mined from the SIT to confirm the estimation of LCF-/HCF-regime by different
the slope of the Wöhler curves can be calculated. The total strain power laws (b).
amplitude and the corresponding number of cycles to failure of
the CAT in the HCF- or LCF-regime (CATHCF, CATLCF, see Fig. 4b),
regime, the cyclic hardening exponent increases to 0.458 in the
allow to calculate the fatigue strength (ductility) coefficient e0 B,M
LCF-regime. Between both power laws a small transition area can
(HCF) (e B,M(LCF)). Thus, using data from three (two) fatigue experi-
0
be determined. This confirms that the CStSp-curve is well suited
ments, the HCF (LCF) Wöhler curve can be described by Eq. (2)
to calculate Wöhler (e-N) curves valid for either the HCF- or LCF-
(Eq. (4)).
regime by using the respective Morrow equation. In [30] this
approach to allocate total strain amplitudes either to the HCF or
5. Results and discussion the LCF regime is proven for a testing frequency of 0.005 Hz.

5.1. Determining the transition between LCF- and HCF-regime by one

strain increase test 5.2. Temperature increase test

Contrary to stress controlled fatigue, for total strain controlled
fatigue the ductile cast iron EN-GJS-600 shows a slope change in
the Wöhler (e-N) curve indicating the transition between the
microplasticity dominated HCF- and the macroplasticity domi-
nated LCF-regime [8,24,25]. Because PhyBaL follows a two-
parameter-concept (see Section 4), either the majority of LCF- or
the HCF-regime can be described mathematically by a correspond-
ing power law. To ensure the validity of the calculated Wöhler (e-
N) curve, the transition between the LCF and HCF-regime has to be
determined. Fig. 5a shows results of a strain increase test at AT,
with a step length of only DN = 50 cycles and total strain steps of
Dea,t = 0.02%, respectively, to reach sufficiently high total strain
amplitudes ea,t in the LCF-regime. The corresponding CStSp-curve
according to Morrow (Fig. 5b) shows a significant change in the
cyclic hardening exponent n0 M. Starting at n0 M = 0.257 in the HCF-
To characterize the temperature dependent cyclic deformation
behavior, several constant amplitude tests were performed at a
total strain amplitude of ea,t = 0.2%, a strain ratio of Re = 1 and a
frequency of 5 Hz. Fig. 6a shows the temperature dependent fati-
gue behavior in a 3-D-diagram. Note that two specimen batches
produced on different lathes were used, with different surface
residual stresses which influenced fatigue strength. Residual stres-
ses were measured by X-ray diffractometry on the {2 1 1} a-Fe lat-
tice planes using Cr-Ka-radiation and sin2 W analysis. Surface
residual stresses of 328 MPa ± 10 MPa were determined for batch
one and 245 MPa ± 9 MPa for batch two. Because of identical
chemical composition, the influence of dynamic strain aging on
the fatigue behavior is assumed to be unaffected by these differ-
ences. Therefore, batch one was used to characterize the influence
of DSA by constant amplitude and temperature increase tests
106 B. Jost et al. / International Journal of Fatigue 96 (2017) 102–113
Fig. 6. Cyclic deformation behavior and relative fatigue life minimum/maximum up to 400 °C (a) and determination of the relative fatigue life minimum/maximum with only
one temperature increase test (b).
(Fig. 6a and b) while the 2nd batch was used for delineation LCF
and HCF regime (Section 5.1), lifetime calculation and verification
(Section 5.3 ff).
Fatigue life decreases with increasing temperature until a rela-
tive fatigue life minimum is observed at 175 °C. Increasing temper-
ature up to 300 °C results in a relative fatigue life maximum and
most pronounced cyclic hardening due to dynamic strain aging.
A further temperature increase to 400 °C results in a significant
decrease of fatigue life. This temperature dependence is in accor-
dance with results of Lang [8] who measured a relative fatigue life
minimum in the range of 200–225 °C and a relative fatigue life
maximum in the range of 300–315 °C for two different total strain
amplitudes. With the relative fatigue life minimum at 175 °C and
the relative fatigue life maximum at 300 °C, the inflection points
of the DSA influence on fatigue life are known. To describe fatigue
life in the complete temperature range, it is then sufficient to esti-
mate the respective Wöhler (e-N) curves at AT, 175 °C, 300 °C and
400 °C. However, determining the DSA dependent inflection points
in lifetime with constant amplitude tests requires a large number
of experiments. As a more effective procedure, the temperature
just below DSA onset and the temperature of most pronounced
DSA are determined in the present study by one temperature
increase test (Fig. 6b).
To measure the influence of dynamic strain aging reliably, suf-
ficient dislocation movement has to be activated. Therefore, the
TIT was performed with a relatively high total strain amplitude
of ea,t = 0.2% at the same frequency as the CATs (f = 5 Hz). To pre-
vent full utilization of cyclic hardening potential before reaching
high temperatures, the step length was set to only 2500 cycles at
each temperature which was increased stepwise by 25 K. The TIT
shows that the temperature of the relative fatigue life minimum
(maximum) due to dynamic strain aging (cf. Fig. 6a) correlates
with the lowest (highest) cyclic hardening in the respective tem-
perature step. Contrary to the constant amplitude tests (Fig. 6a),
cyclic hardening in TIT (Fig. 6b) is less pronounced at 175 °C than
at AT. This is because during each temperature step the cyclic hard-
ening potential of the material is partially utilized resulting in less
pronounced hardening at the subsequent temperatures, including
175 °C compared to CAT. At 200 °C, dynamic strain aging signifi-
cantly activates cyclic hardening again, resulting in steeper slopes
of ra and ea,p vs. cycle number above 175 °C, proving that a tem-
perature increase test cannot completely resemble the cyclic defor-
mation behavior under constant amplitude loading at elevated
temperatures but is sufficiently sensitive to temperature depen-
dent, i.e. DSA induced cyclic hardening mechanisms. Therefore,
the onset of dynamic strain aging is concluded to occur in the tem-
perature range between 175 °C and 200 °C. At 300 °C, highest cyclic
hardening occurs consistent in constant amplitude tests and tem-
perature increase test representing the maximum DSA effect.
This confirms that a temperature increase test is well suited to
identify the temperatures of the relative fatigue life minimum/-
maximum of EN-GJS-600 and provides an efficient method to
determine the influence of DSA on cyclic deformation behavior.
5.3. Calculation of temperature dependent lifetime behavior by
PhyBaLSIT
A detailed explanation of PhyBaLSIT is given in Section 4 and the
application of this method to GJS-600 at ambient temperature is
shown in [23]. In the present study, calculations of Wöhler (e-N)
curves for AT, 175 °C, 300 °C and 400 °C, according to the results
of temperature increase tests (Fig. 6b), are shown.
The SITs were started at a total strain amplitude of 0.03%
(400 °C), 0.05% (300 °C; 175 °C) and 0.06% (AT). After each step
with a length of DN = 9000 cycles, the total strain amplitude was
raised by Dea,t = 0.01%. Based on the material’s response in each
step of the SIT, the endurance limit Se can be evaluated. Regarding
the changes in the slopes of the ea,p-N curves, the temperature
dependent endurance limits can be determined as shown in
Fig. 7a–d. The estimated endurance limits are summarized in
Table 2. Because the estimation of the endurance limit is based
on measured values representing the integral materials reaction
within the gauge length, the influence of exceptionally large
defects, e.g. graphite agglomerations or pores which might
increase fatigue life scatter at constant amplitude tests are not con-
sidered in this approach.
Compared to AT, a temperature increase to 175 °C and 300 °C
only slightly reduces the estimated endurance limit by 0.01% for
each temperature step while a further increase to 400 °C signifi-
cantly reduces the estimated endurance limit. Considering each
total strain step in the SIT, the cyclic deformation behavior at each
temperature can be characterized. Starting from AT (Fig. 7a), a
temperature increase to 175 °C (Fig. 7b) does not significantly
affect cyclic hardening while at 300 °C (Fig. 7c) maximum cyclic
hardening caused by DSA can be observed within all total strain
amplitude steps. At 400 °C (Fig. 7d), the weakest cyclic hardening
in the investigated temperature regime is observed.
In a second step, the total strain amplitudes of two constant
amplitude tests representing the boundaries of the actual lifetime
calculation are chosen up to two SIT steps above Se (lower CATHCF)
 B. Jost et al. / International Journal of Fatigue 96 (2017) 102–113 107

Fig. 7. Strain increase test of EN-GJS-600 at ambient temperature (a), 175 °C (b), 300 °C (c) and 400 °C (d) and cyclic total strain–plastic strain (CStSp-) curve, data are derived
from SIT (e).

Table 2
Estimation of the temperature dependent endurance limit Se by SIT. and up to two steps below the total strain amplitude where frac-
ture occurs in the SIT (upper CATHCF), see Fig. 4b. Table 3 summa-
AT 175 °C 300 °C 400 °C
rizes the temperature dependent total strain amplitudes of the
Se estimated [ea,t in %] 0.15 0.14 0.13 0.10 upper and lower CATHCF, respectively. The average ea,p values of
each SIT strain step between the two constant amplitude tests
are used as input parameters to describe the cyclic hardening
Table 3
Temperature dependent upper and lower total strain amplitude defining the behavior in the range of strain amplitudes to be covered by the
ea,t-range for PhyBaLSIT lifetime calculation. fatigue life calculation PhyBaLSIT. As mentioned in Section 4, the
SIT parameters chosen here (DN = 9000 cycles, Dea,t = 0.01%) cover
AT 175 °C 300 °C 400 °C
the HCF-regime. Because the HCF Wöhler curve is described by a
Upper CATHCF [ea,t in %] 0.22 0.19 0.21 0.13
power law (Eq. (2), see Fig. 4c), it is necessary to ensure that the
Lower CATHCF [ea,t in %] 0.16 0.16 0.15 0.11

Table 4
Temperature dependent cyclic hardening coefficient and exponent from SIT at AT, 175 °C, 300 °C and 400 °C, and R2 indicating the correlation of the fitted power laws with
experimental data in Fig. 7e.

AT 175 °C 300 °C 400 °C

Cyclic hardening coefficient K0 M(HCF) (SIT) 0.343 0.306 0.396 0.364
Cyclic hardening exponent n0 M(HCF) (SIT) 0.180 0.197 0.214 0.267
Correlation R2 (SIT) 0.998 1.000 0.995 0.999
108 B. Jost et al. / International Journal of Fatigue 96 (2017) 102–113

chosen ea,t-range (Table 3) is within the microplasticity dominated

HCF-regime.
Fig. 7e shows the corresponding CStSp-curves for AT, 175 °C,
300 °C and 400 °C. To verify that the chosen ea,t-range is within
the HCF-regime, the CStSp-curves were described by Eq. (5). Table 4
summarizes the temperature dependent parameters of the CStSp-
curves fitted to Eq. (5). The extremely high correlation coefficients
R2 prove that the ea,t-range for the PhyBaLSIT calculation is entirely
located in the HCF-regime, described by Basquin’s law of fatigue.
To convert the CStSp-curves of the SITs in CStSp-curves for con-
stant amplitude loading, two CATs were performed at each tem-
perature (ea,t from Table 3). With the ea,p values taken at 104
cycles the CStSp-curves of the SIT were converted to constant
amplitude loading, see Fig. 8a–d. Table 5 shows the cyclic harden-
ing exponent n0 M(HCF) for CATs compared to n0 M(HCF) directly
obtained from SIT data.
This conversion is essential, because the correlation between
the cyclic hardening exponent n0 M(HCF) (Morrow), describing the
slope of a CStSp-curve, in the HCF-regime and the fatigue strength
exponent bM (Basquin), describing the slope of the Wöhler-curve in
the HCF-regime (Eq. (7)) shows best accordance for data taken in a
sufficiently stabilized state of constant amplitude loading which is
considered to be reached at 104 cycles in the present case (see
Fig. 8a–d). A comparison of the cyclic hardening exponents derived
from the SIT results with the values after considering the CAT data
shows the importance of this re-calculation.
Note that after conversion to constant amplitude loading the
cyclic hardening exponent n0 M(HCF) for 300 °C takes a very high
value with n0 M(HCF) = 0.418 representing a strongly inclined Wöhler
(e-N) curve. Considering the position of the CStSp-curves, it is obvi-
ous that for T = AT to 300 °C the converted CStSp-curves are shifted
to lower ea,p. This can be explained by cyclic hardening which
occurs in each strain step of the SIT. Because the ea,p values are
taken at 104 cycles in CAT and the step length in SIT is just slightly
shorter (9  103 cycles), cyclic hardening during the SIT steps leads
to reduced ea,p values. The distance between both, CStSp-curve of
SIT and CStSp-curve of CAT, increases with cyclic hardening and,
consequently gets larger from AT to 175 °C to 300 °C. At 400 °C,
only weak cyclic hardening is observed in SIT (Fig. 7d). Therefore,
the CStSp-curve out of SIT is shifted to higher values than that
resulting from CAT. Comparing CStSp-curves from SIT and CAT
therefore enables direct conclusions regarding temperature depen-
dent cyclic deformation behavior.
Substituting the cyclic hardening exponent n0 M(HCF) (CAT) and
solving Eq. (7) leads to the following fatigue strength exponents
bM defining the slopes of the calculated Wöhler (e-N) curves
(Table 6).
After the fatigue strength exponent bM is calculated, only the
fatigue strength coefficient e0 B,M(HCF) has to be determined by solv-
ing Eq. (2) and substituting the total strain amplitude ea,t and the
number of cycles to failure Nf to describe the respective Wöhler
(e-N) curves (Eq. (9), Table 7). Due to generally smaller scatter in
lifetime at high loads and shorter test duration Nf exp. of the upper
CATsHCF are used.
ea;t
e0B;MðHCFÞ ¼ ð9Þ
ð2Nf ÞbM
With the determined fatigue strength coefficients e0 B,M(HCF) Fig. 8. Constant amplitude tests of EN-GJS-600 and converted cyclic total strain-
(Table 7) and the fatigue strength exponents bM (Table 6) for plastic strain (CStSp-) curves for constant amplitude loading at ambient temper-
ature (a), 175 °C (b), 300 °C (c) and 400 °C (d).
T = AT, 175 °C, 300 °C and 400 °C the respective Wöhler curves at
AT, 175 °C, 300 °C and 400 °C can be described via Eq. (2). The
determined Wöhler curves, drawn up to the total strain amplitude taken once per temperature, marked by the symbols in Fig. 9. In
where endurance limit was estimated by SIT (Table 2), are plotted [24,31,32] it is shown that the transition area between LCF- and
in Fig. 9. Note that the calculation was done using only fatigue data HCF-regime is located at Nf 6 104 cycles, which supports the anal-
of one strain increase and two constant amplitude tests for each ysis of the respective Morrow equation (Fig. 7e) given here allocat-
temperature while the numbers of cycles to failure Nf are only ing the calculated ea,t-ranges in the HCF-regime.
 B. Jost et al. / International Journal of Fatigue 96 (2017) 102–113 109

Table 5
Temperature dependent cyclic hardening exponent from SIT and converted from SIT to CAT at AT, 175 °C, 300 °C and 400 °C and R2 defined accuracy of the power laws.

AT 175 °C 300 °C 400 °C

0
Cyclic hardening exponent n M(HCF) (SIT) 0.180 0.197 0.214 0.267
Cyclic hardening exponent n0 M(HCF) (CAT) 0.201 0.291 0.418 0.252
Correlation R2 (CAT) 0.998 0.999 0.973 0.999

Table 6
Temperature dependent fatigue strength exponents bM defining the slopes of the PhyBaLSIT-calculated Wöhler (e-N) curves.

AT 175 °C 300 °C 400 °C

Fatigue strength exponent bM 0.100 0.118 0.135 0.112

Table 7
Temperature dependent fatigue strength coefficient e0 B,M(HCF).

AT 175 °C 300 °C 400 °C

Fatigue strength coefficient e0 B,M(HCF) 0.677 0.729 0.934 0.438

Fig. 9. PhyBaL calculated Wöhler (e-N) curves with fatigue data extracted from one Fig. 10. PhyBaL calculated Wöhler (e-N) curves in excellent accordance to Nf exp.
SIT and two CATs using only Nf exp. out of one CAT per temperature. from constant strain amplitude tests.

Starting from AT, a temperature increase to 175 °C leads to a
reduced fatigue strength and temperature affects fatigue life at
low- and high total strain amplitudes in a similar way, indicated
by the Wöhler (e-N) curve for 175 °C being shifted parallel to lower
fatigue strength with an almost identical slope as at AT. A further
temperature increase to 300 °C activates dynamic strain aging
effects which, in comparison to 175 °C, increases fatigue strength
in the major part of the considered ea,t-range. As a further conse-
quence of DSA, the slope of the e-N-curve increases from
bM = 0.118 (175 °C) to bM = 0.135 (300 °C), see Table 6. For high
total strain amplitudes a considerable benefit in fatigue strength
can be observed compared to 175 °C. This benefit declines with

Fig. 11. Back scatter electron micrographs and crack initiating defects of two specimens after fracture in constant amplitude tests with ea,t = 0.16% at 175 °C.
110 B. Jost et al. / International Journal of Fatigue 96 (2017) 102–113

decreasing total strain amplitudes until the numbers of cycles to

failure Nf for 175 °C and 300 °C get almost equal at ea,t = 0.14%. This
effect is also reflected by the estimated endurance limit Se deter-
mined from SIT (Table 2), i.e. at 175 °C (ea,t = 0.14%) only a slightly
higher endurance limit occurs than at 300 °C (ea,t = 0.13%).
Decreasing total strain amplitude leads to reduced dislocation den-
sity and -movement and therefore to a weaker influence of
dynamic strain aging until close to the endurance limit thermally
assisted dislocation movement, which would, in the absence of
other effects, decrease material strength, dominates the deforma-
tion behavior compared to DSA. This finding verifies that the total
strain amplitude of ea,t = 0.2% chosen for the TIT (Fig. 6b) is well
suited to figure out dynamic strain aging effects. At last, a further
temperature increase to 400 °C affects fatigue strength detrimen-
tally as also shown in [25]. While the slope of the 400 °C Wöhler
(e-N) curve is similar as for 175 °C and AT, the general fatigue
strength is significantly reduced.

5.4. Cyclic deformation behavior and lifetime verification

To verify the calculated Wöhler (e-N) curves and to evaluate

cyclic deformation behavior more accurately, constant amplitude
tests at AT, 175 °C, 300 °C and 400 °C were performed. Fig. 10
shows the calculated Wöhler (e-N) curves (Fig. 9) together with
Nf exp. determined in these additional CATs.
The endurance limits Se estimated from SITs were proved for AT
(0.15%) and 400 °C (0.1%). At 175 °C, CATs showed similar to AT an
endurance limit of 0.15% proving a conservative estimation by SIT
(ea,t:0.14%). After validating the endurance limit at 0.15%, the same
specimen (55UE) was tested at 0.16% reaching 4.26  105 cycles to
failure. In comparison to another CAT (42A) at 0.16%, which con-
firms the calculated e-N-Wöhler curve (Nf = 2.33  105), the num-
ber of cycles to failure is almost doubled. This can be explained by
a considerably smaller crack initiating defect size in sample 55UE.
Fig. 11 shows the corresponding back scatter electron (BSE) micro-
graphs as well as detailed views of the crack initiating defects. To
visualize the difference in defect size, the crack initiating graphite
agglomeration and pore are indicated by frames and shown in
higher magnification. Both specimens have a generally homoge-
nous graphite distribution. Specimen 55UE shows a relative small
graphite agglomeration near the surface while the crack in speci-
men 42A initiated at a considerably larger near-surface pore. Fur-
thermore, a lower graphite content was measured in the cross
section of the fracture for specimen 55UE (18.79%) compared to
specimen 42A (20.20%). Current investigations give a clear indica-
tion that fatigue cracks appear in the region of the gauge length
with the highest graphite content. Therefore, the graphite content
in the fracture surface might also explain the differences in fatigue
life of both investigated samples.
For all temperatures, the results of the constant amplitude tests
show an excellent accordance to the Wöhler (e-N) curves calcu-
lated by PhyBaL. Especially the calculated slope of the Wöhler
curve for 300 °C shows PhyBaLSIT being sensitive for the effect of
changed cyclic deformation mechanisms on both, the cyclic hard-
ening exponent n0 M(HCF) as well as the fatigue strength exponent
bM.
In general, the very low lifetime scatter does not meet the
expectations for cast iron. Slima et al. [33] showed, that lifetime Fig. 12. Selected cyclic deformation and cyclic stress curves for constant amplitude
scatter in cast irons can be caused by different defect sizes influ- tests at ambient temperature (a), 175 °C (b), 300 °C (c) and 400 °C (d).
encing crack nucleation. The specimens were classified as ‘‘with
defects” and ‘‘without defects”. Wöhler-curves determined sepa- surface pores while the defect size shows relatively small scatter
rately for both classes show a lower lifetime scatter than generally which might explain the very low lifetime scatter.
reported for cast iron, proving that lifetime scatter directly refers to Fig. 12a–d shows cyclic deformation curves ea,p-N and ra-N
the defect size in crack nucleation. Ongoing investigations on the determined in CATs at several ea,t and all investigated tempera-
EN-GJS-600-specimens used in the present work suggest that most tures. Generally cyclic hardening occurs which is more or less pro-
specimens fail at near surface graphite agglomerations and near nounced depending on temperature. For AT this agrees with data
 B. Jost et al. / International Journal of Fatigue 96 (2017) 102–113
presented in [8,34,35]. Note that cast iron with ferritic matrix was
investigated in [35]. Because plastic deformation processes mainly
occur in the ferritic areas, those results are nevertheless represen-
tative. The weakest cyclic hardening is observed at 400 °C, while
cyclic hardening is most pronounced at 300 °C and gets stronger
with increasing total strain amplitude ea,t. Šamec et al. [36] inves-
tigated the cyclic deformation and lifetime behavior of the ferritic-
pearlitic ductile cast iron EN-GJS-500 with a frequency of 2 Hz at
AT, 300 °C and 400 °C. At 300 °C, cyclic hardening is most pro-
nounced whereas at 400 °C cyclic hardening is significantly
reduced. Although not explicitly mentioned by the authors, it is
likely that the most pronounced cyclic hardening observed at
300 °C is related with dynamic strain aging.
The effect of total strain amplitude on dynamic strain aging is
confirmed by comparison of selected ea,p-N curves at 175 °C and
300 °C: For ea,t = 0.15%, at 175 °C no failure occurred until 2  106
cycles while ea,p approaches the value of 0.014  103 in satura-
tion. At 300 °C, fatigue failure occurs at 3.77  105 cycles when
ea,p saturates at a significantly increased value of 0.045  103.
For ea,t = 0.19%, at 300 °C a higher fatigue life of 5.8  104 cycles
occurs compared to 4.2  104 cycles at 175 °C. Thereby, the value
of ea,p at 104 cycles for 300 °C (0.062  103) is clearly lower than
for 175 °C (0.113  103). This comparison shows that for the duc-
tile cast iron EN-GJS-600 dynamic strain aging has a positive effect
on fatigue life due to reduced ea,p values for the majority of the
constant amplitude tests.
5.5. Comparison to other publications on ductile cast iron EN-GJS-600
Fig. 13 gives a comparison of the results presented in Fig. 10
with data published for EN-GJS-600 in literature. The four
Wöhler-curves determined in the present work are declared as Nf
calc. PhyBaLSIT and plotted as thick lines.
Leutwein [25] performed three total strain controlled fatigue
tests at total strain amplitudes of 0.3%, 0.4% and 0.5% at Re = 1.
The frequency is not specified. By determining the elastic, plastic
and total strain for each test based on the stress-strain-hysteresis
at Nf/2, the elastic and plastic strain amplitude vs. Nf were sepa-
rately described with good correlation by a power law. Combining
both power laws, the HCF- and LCF-regime is described by a
four-parameter-concept. Generally it is obvious that the fatigue
strength in the HCF-regime at AT calculated in the present work
is lower than that described in [25]. In analogy to the PhyBaL-
method, data taken in a characteristic fatigue state are assumed
in [25] to describe the cyclic deformation induced material reac-
tion. Based on this, the complete Wöhler curve is calculated from
results of only three fatigue tests. In contrast to that, the PhyBaL
method used in the present work requires at least four fatigue tests
Fig. 13. Wöhler curves for AT, 175 °C, 300 °C and 400 °C compared to results in literature.
111
(one strain increase test, two constant amplitude tests within the
HCF-regime and one constant amplitude test within the LCF-
regime) for lifetime calculation in both, the HCF- and the LCF-
regime (see Section 4) to consider the different cyclic deformation
behavior in both regimes.
Hübner et al. [32] performed fatigue tests at different total
strain amplitudes with a strain ratio of Re = 1 and a frequency
of 1 Hz. The results are described by the Manson-Coffin and Bas-
quin law. The reported fatigue strength is significantly reduced
compared to the results obtained in the present work. Even at
175 °C, the investigated material here shows a higher fatigue
strength than the results given for AT in [32]. Although it may be
expected that the lower frequency results in slightly lower fatigue
strength, it is obvious that the fatigue strength of the material
investigated in [32] is smaller than that of the GJS-600 tested in
the present study.
Lang [8] performed total strain controlled constant amplitude
tests at AT, a frequency of 5 Hz and a strain ratio Re = 1. In anal-
ogy to Leutwein, the fatigue strength is higher than the results of
the present work. Note that the slope change between HCF- and
LCF-regime occurs at higher number of cycles to failure compared
to the other results. Hence, comparing the Wöhler curves of Lang
and Hübner shows that a significant difference in fatigue strength
can be observed between material batches meeting the standards
for EN-GJS-600. On this foreground, the behavior of the EN-GJS-
600 investigated in the present work proves to be typical for this
material class, however with a tendency to higher fatigue strength
and small scatter. Furthermore, Nf exp. for 300 °C and 400 °C taken
from temperature-life curves given in [8] for the total strain ampli-
tudes 0.197% and 0.237% are plotted in Fig. 13. The data for 300 °C
are close to the LCF-HCF transition and approach the HCF Wöhler
curve for 300 °C determined in the present work while the
400 °C data taken from [8] are clearly located in the LCF regime.
In [37] the endurance limit for AT, 250 °C and 500 °C is given.
In the Materials Science Monographs 61 [38] (MSM 61) and 42E
[39] (MSM 42E), data of several publications on fatigue of GJS-600
are collected. For the LCF-regime at AT, the data fit perfectly to the
results of Leutwein while in the HCF regime the data fit quite well
to the Wöhler curves determined in the present study. At 250 °C,
the LCF fatigue strength is increased compared to AT while in the
transition range between LCF- and HCF the benefit in fatigue
strength decreases until in the HCF-regime the fatigue strength is
slightly reduced compared to AT. At 500 °C, fatigue strength is
clearly reduced, but higher compared to the 400 °C Wöhler curve
determined in the present study.
This comparison to literature data shows that, considering typ-
ical scatter, the results of the present work fit well into published
values on fatigue strength of EN-GJS-600.
112 B. Jost et al. / International Journal of Fatigue 96 (2017) 102–113
Fig. 14. Temperature, strain, fatigue life relation calculated via PhyBaLSIT with data
determined in 13 fatigue tests.
5.6. Temperature- and load-dependence on fatigue life
With one temperature increase test the temperature interval
between the onset and maximum of dynamic strain aging was esti-
mated for the frequency of 5 Hz (see Section 5.2). The respective
Wöhler (e-N) curves at ambient temperature, 175 °C, 300 °C and
400 °C were calculated by PhyBaLSIT using fatigue data determined
in one strain increase and two constant amplitude tests at each
temperature (see Section 5.3). Hence, one temperature increase
test, four strain increase tests and eight constant amplitude tests
were used to describe the strain amplitude and temperature
dependent fatigue behavior in detail. Fig. 14 shows the
temperature-strain-fatigue life behavior in the ea,t-range from
lower to upper CATHCF calculated from data determined in these
13 fatigue tests.
This comprehensive description of the temperature- and load-
dependence of fatigue life and endurance limit allows to consider
the effect of dynamic strain aging in component design by a man-
ageable and affordable number of fatigue tests. Ongoing research
in the frequency range between 0.005 Hz and 92 Hz aims to
describe the influence of deformation rate on the temperature
range of dynamic strain aging.
6. Summary and conclusions
The physically based fatigue lifetime calculation method
PhyBaLSIT was applied for lifetime calculation in isothermal total
strain controlled high cycle fatigue. One total strain increase test
allows to estimate the endurance limit Se at a given temperature
based on the development of the plastic strain amplitude ea,p.
The total strain amplitudes ea,t of the upper and lower constant
amplitude test should be chosen slightly above the endurance limit
Se and slightly below the total strain amplitude leading to fracture
in SIT. The cyclic deformation behavior can be considered for each
total strain amplitude step of the SIT. The average values of each
step are taken as input parameters for the calculation of cyclic total
strain–plastic strain (CStSp-)curves according to Morrow. Besides
the description of the CStSp-curves, the Morrow equation (modified
for total strain controlled fatigue) allows to determine the transi-
tion between LCF- and HCF-regime and therefore to ensure that
the calculated Wöhler (e-N) curve is entirely located in the HCF-
regime. Using the generalized Morrow and Basquin equations,
modified for total-strain controlled fatigue tests, together with
data taken at 104 cycles from two total strain controlled constant
amplitude tests allows the calculation of a Wöhler (e-N) curve in
the HCF-regime. It should be noted that (i) a Wöhler-curve
determined from only three fatigue tests does not consider the
lifetime scatter of the investigated material. To describe the scatter
in detail, several fatigue experiments at constant amplitude are
needed. (ii) PhyBaL uses integrally measured data representing
fatigue induced changes in the complete gauge length. While those
data determining the cyclic deformation behavior correlate well to
the slope of the Wöhler-curve, it is not possible to separate the
influence of each single mechanism as e.g. the models presented
in [16–19]. To estimate the effect of loading parameters such as
strain ratio and frequency the PhyBaL-procedure must be repeated
with changed parameters. Nevertheless, it is obvious that the pro-
posed assessment of HCF Wöhler-curves based on the Basquin law
can be used in a reliable and efficient way as soon as it is assured
that all tests used for lifetime calculations lie below the transition
range between LCF and HCF.
To describe the temperature dependence on fatigue behavior of
the ductile cast iron EN-GJS-600, dynamic strain ageing effects on
the endurance limit as well as on cyclic deformation behavior in
the temperature range from AT to 400 °C, must be considered. Con-
stant amplitude tests (ea,t = 0.2%) at different temperatures showed
a relative fatigue life minimum at 175 °C. At T = 300 °C a relative
fatigue life maximum and most pronounced cyclic hardening due
to dynamic strain aging is observed. Temperatures above 300 °C
lead to a significant reduction of fatigue life. To determine the tem-
perature range between the onset and the maximum of dynamic
strain aging, a temperature increase test at constant strain ampli-
tude was used. By a reasonable choice of the parameters in the
temperature increase test assuring sufficient cyclic plastic defor-
mation to activate DSA-effects, the onset and the temperature of
maximum dynamic strain aging were estimated with good accu-
racy. Thus an efficient estimation of the relevant temperatures
regarding the fatigue strength and endurance limit is possible.
The present work proves PhyBaLSIT based on the combination of
one temperature increase test providing an efficient method to cal-
culate fatigue life in the temperature range from AT to 400 °C at a
frequency of 5 Hz. Several constant amplitude tests verified the
excellent accordance between the PhyBaLSIT calculated Wöhler
(e-N) curves and Nf exp. determined in constant amplitude tests.
Because besides temperature, also the strain rate influences the
DSA-affected regime, research on total strain controlled fatigue
with frequencies from 0.005 Hz up to almost 100 Hz in the temper-
ature range AT to 400 °C is currently performed and will be pub-
lished soon, including a detailed discussion on fatigue lifetime
calculation in the macroplasticity dominated LCF regime.
Acknowledgements
The financial support of this work by the German Research
Foundation (DFG) is gratefully acknowledged. Special thanks to
the Daimler AG for providing the test material.
References
[1] O’Rourke R. Cast iron: the engineered metal. Adv Mater Process 2001; 159: 65–
8.
[2] Basshuysen R, Schäfer F. Handbuch Verbrennungsmotor, vol. 5. Vieweg +
Teubner; 2010.
[3] Karpuschewski B, Welzel F, Risse K, Schorgel M, Kreter S. Potentials for
improving efficiency of combustion engines due to cylinder liner surface
engineering. Proc CIRP 2016;46:258–65.
[4] Carey C et al. Extreme engine downsizing. In: I.o.M. Engineers, editor.
Innovations in fuel economy and sustainable road transport. Woodhead
Publishing; 2011. p. 135–47.
[5] Nozawa R, Morita Y, Shimizu M. Effects of engine downsizing on friction losses
and fuel economy. Tribol Int 1994;27(1):31–7.
[6] Thomas JJ, Verger L, Bignonnet A, Charkaluk E. Thermomechanical design in
the automotive industry. Fatigue Fract Eng Mater Struct 2004;27:887–95.
[7] Chaboche J-L, Gaubert A, Kanouté P, Longuet A, Azzouz F, Mazière M.
Viscoplastic constitutive equations of combustion chamber materials
including cyclic hardening and dynamic strain aging. Int J Plast 2013;46:1–22.
 B. Jost et al. / International Journal of Fatigue 96 (2017) 102–113
[8] Lang K-H. Das Ermüdungsverhalten von GGG-60, GGG-40, GGV-30 und GG-30
im Temperaturbereich 20 °C < T < 500°C. Dissertation. Universität Karlsruhe
(TH); 1985.
[9] Löhe D, Vöhringer O, Macherauch E. Dynamic strain aging of cast iron. In:
Strength of metals and alloys (ICSMA 6); 1982. p. 199–204.
[10] Löhe D, Vöhringer O, Macherauch E. The influence of graphite shape on
mechanical properties of ferritic cast iron. ICM 1979;3(3):179–88.
[11] Hallstein R. Lang K-H, Löhe D, Macherauch E. Isotherme
totaldehnungskontrollierte Wechselverformung von GGG-60 im
Temperaturbereich 20 °C 6 T 6 500 °C, Teil 1. Mat.-wissenschaften und
Werkstofftechtechnik, vol. 26; 1995. p. 5–13.
[12] Hallstein R, Lang K-H, Löhe D, Macherauch E. Isotherme
totaldehnungskontrollierte Wechselverformung von GGG-60 im
Temperaturbereich 20 °C 6 T 6 500 °C, Teil 2. Mat.-wissenschaften und
Werkstofftechtechnik, vol. 27; 1996. p. 53–9.
[13] Schoenborn S, Kaufmann H, Sonsino CM, Heim R. Variable amplitude fatigue of
high-strength cast iron alloys for automotive applications. Int J Fatigue
2016;91:445–58.
[14] Uihlein A, Lang K-H, Löhe D. Versagensverhalten von Gusseisenwerkstoffen für
Zylinderköpfe. Neue Brennverfahren Renningen: Expert-Verl; 2007. p. 98–107.
[15] Uihlein A, Lang K-H, Löhe D. Low cycle fatigue bevaviour of cast irons at
different temperatures. In: Sixth international conference on low cycle fatigue,
Berlin, Germany; 2008. p. 371–6.
[16] Seifert T, Riedel H. Mechanism-based thermomechanical fatigue life prediction
of cast iron. Part I: Models. Int J Fatigue 2010;32:1358–67.
[17] Seifert T, Maier G, Uihlein A, Lang KH, Riedel H. Mechanism-based
thermomechanical fatigue life prediction of cast iron. Part II: comparison of
model predictions with experiments. Int J Fatigue 2010;32:1368–77.
[18] Metzger M, Nieweg B, Schweizer C, Seifert T. Lifetime prediction of cast irons
under combined thermomechanical fatigue and high cycle loading using a
mechanism-based model. Int J Fatigue 2013;53:58–66.
[19] Metzger M, Seifert T. A mechanism-based model for LCF/HCF and TMF/HCF life
prediction: multiaxial formulation, finite-element implementation and
application to cast iron. Tech Mech 2012;32:435–45.
[20] Ebel-Wolf B, Walther F, Eifler D. Influence of elevated temperatures on the
cyclic deformation behaviour of the magnesium die-cast alloys AZ91D and
MRI 230D. Mater Sci Eng A 2008;486(1–2):634–40.
[21] Eifler D, Starke P, Walther F. ‘‘PHYBAL” a short-time procedure for a reliable
fatigue life calculation. Adv Eng Mater 2010;12.
[22] Germann H, Starke P, Kerscher E, Eifler D. Fatigue behaviour and lifetime
calculation of the cast irons EN-GJL-250, EN-GJS-600 and EN-GJV-400. Proc
Eng 2010;2(1):1087–94.
113
[23] Jost B, Klein M, Eifler D. PHYBALSIT – fatigue assessment and life time
calculation of the ductile cast iron EN-GJS-600 at ambient and elevated
temperatures. Charact Miner Met Mater 2015;2015:711–8.
[24] Korkmaz S. A methodology to predict fatigue life of cast iron uniform material
law for cast iron. J Iron Steel Res 2011;18(8):42–5.
[25] Leutwein H-U. Einfluss von Autofrettage auf die Schwingfestigkeit
innendruckbelasteter Bauteile aus Kugelgraphitguss; 2008.
[26] Basquin OH. The exponential law on endurance tests, vol. 10. ASTM Proc;
1910.
[27] Coffin Jr LF. A study of the effects of cyclic thermal stresses on a ductile metal.
Trans ASME 1954;76(6):931–50.
[28] Manson SS. Behavior of materials under conditions of thermal stress. National
Advisory Committee for Aeronautics, technical note 2933; 1953.
[29] Morrow JoDean. Cyclic plastic strain energy and fatigue of metals. ASTM
International, STP 378; 1964.
[30] Klein M, Jost B, Beck T, Eifler D. Talks of the 27th colloquium on mechanisms
on fatigue: determining the temp.dep. cyclic deformation behavior of EN-GJS-
600 grey cast Fe by temp. & strain increase tests 2016-03-22.
[31] Lagoda T. Energy models for fatigue life estimation under uniaxial random
loading. Part II: verification of the model. Int J Fatigue 2001;23(6):481–9.
[32] Hübner P, Pusch G, Krodel L. Fatigue behavior of cast iron with globular
graphite. Adv Eng Mater 2004;7(6):541–4.
[33] Slima K Ben, Penazzi L, Mabru C, Ronde-Oustau F. Fatigue analysis-based
numerical design of stamping tools made of cast iron. Int J Adv Manuf Technol
2013;67(5–8):1643–50.
[34] Meneghetti G, Ricotta M, Masaggia S, Atzori B. Comparison of the low-cycle
and medium-cycle fatigue behaviour of ferritic, pearlitic, isothermed and
austempered ductile irons. Fatigue Fract Eng Mater Struct, 2013.
[35] Petrenec M, Tesařová H, Beran P, Šmid M, Roupcová P. Comparison of low cycle
fatigue of ductile cast irons with different matrix alloyed with nickel. Proc Eng
2010;2.
[36] Šamec B, Potrč I, Šraml M. Low cycle fatigue of nodular cast iron used for
railway brake discs. Eng Fail Anal: Mater Struct Compon Reliab Des 2011;18
(6):1643–50.
[37] Lang KH, Löhe D. Untersuchungen zum ein-und wechselsinnigen
Verformungsverhalten verschiedener Gusseisensorten im Temperaturbereich
20 °C kleiner gleich T kleiner gleich 500 °C. Forschungsberichte
Verbrennungskraftmaschinen; 1982. Vorhaben Nr. 228.
[38] Bäumel Jr A. Seeger T. Materials data for cyclic loading. Supplement 1.
Materials Sciene Monographs; 1990. p. 61.
[39] Boller Chr, Seeger T. Materials data for cyclic loading. Part E: cast and welded
materials. Materials Science Monographs; 1987. p. 42E.