Abstract
In gas turbines, superalloys are exposed to thermal as well as
mechanical cyclic loadings during start-up and shut down processes,
which can accelerate the formation of fatigue failure mechanisms. In the
present study, low cycle fatigue behavior and fracture mechanism of a
directionally-solidified CM247 LC superalloy at two temperatures of 600
°C and 800 °C were investigated. For this purpose, strain-controlled low
cycle fatigue tests were carried out at 600 °C and 800 °C, and constant
total strain amplitudes of 0.4, 0.6, 0.8, and 1% were applied during
the totally reversed loading ratio (\(R\ =\ -1\)). The Coffin-Manson
model, based on plastic deformation and a model based on the hysteresis
energy criterion is used to predict fatigue life and evaluate the low
cycle fatigue behavior. SEM observations of the surface of the failed
specimen showed similar LCF failure mechanisms in all the strain
amplitudes and temperatures.
Keywords: Low cycle fatigue, Life prediction,
Superalloy, Fracture morphology
INTRODUCTION
In gas turbines, engineers use superalloys to achieve better efficiency
and to attempt to optimize performance and efficiency. Increased turbine
inlet temperatures (TIT) are aspired for reducing fuel consumption and
increasing thermal efficiency. To control the energy requirement, more
startup and shutdown operations are necessary. Throughout the start-up
operation, the outer surface of the rotary component heats up before the
inner surface. Then, compressive stress on the outer surface is created.
Throughout the shut-down operation, the outer surface cools down faster.
Then, tensile stress on the outer surface is created. Furthermore, the
microstructure of the superalloys unavoidably becomes degraded due to
operating at elevated temperatures and high rotation speeds1,2. Therefore, flexibility also becomes more
important 3. First-generation nickel-base superalloys
have been vastly improved over time; and finally, as a solution to
continual requisitions for higher turbine outputs, single-crystal and
directionally-solidified superalloys were developed 4.
Due to their notable thermo-mechanical resistance, single-crystal and
directionally-solidified Nickel-base superalloys are broadly used as a
material for components in the hot-section of land-based gas turbines
and aircraft engines 5. Currently, there are many
materials used at the same time with a selection made on the grounds of
the constraints of the specific engineering application6. The directionally-solidified CM247 LC is one of the
most commonly used superalloys for high-temperature applications, which
are vastly used in the blades of gas turbines 1,7,8.
The appropriate mechanical properties at high temperatures are a
consequence of the γ′ precipitates in the matrix of γ phase9,10. The γ and γ′ phases are generally solid-solution
strengthened. After accumulating damage during fatigue exploitations,
the parameters that are of high importance include the microstructural
ones such as volume fraction and coherency strains between γ-γ′ and so
on 11,12. Numerous damage mechanisms occur in gas
turbines, which can cause the failure of blades and nozzles, such as low
cycle and high cycle fatigue, creep, thermo-mechanical fatigue (TMF),
corrosion, foreign object damage, erosion, and oxidation13-17. Metal fatigue is one of the main mechanical
failure modes in real applications 18. Due to the
engine start-up and shut-down processes as well as cooling by the axial
air compressor, hot-section components of a gas turbine engine
experience significant thermal fluctuations, resulting in localized and
small plastic strains. Thus, these components are designed such that
they can be resistant to low cycle fatigue 11,19. To
guarantee adequate creep and fatigue resistance, the manufactures are
selected directionally-solidified nickel-based superalloy materials. The
low cycle fatigue failure can lead to trip the gas turbine engines and
economic loss markedly in power plants and gas transmission stations,
while this can also cause catastrophic damage to aero-engines20-22. Generally, stress-strain and temperature fields
are the basis of the models used for LCF calculation23.
Low cycle fatigue in equiaxed and directionally-solidified Ni-based
superalloys has been studied extensively over the past two decades. S.M.
Seo et al. 24 investigated low cycle fatigue and
fracture behavior of Ni-based super-alloy CM247LC at 760 °C. Antolovich
et al. 25 investigated the low cycle fatigue behavior
of René 80 at two temperatures of 871 and 982 °C. It was found that when
frequency decreased the fatigue life was increased. He et al.26 studied the effect of temperature on the low cycle
fatigue behavior of a directionally solidified Mar-M200 superalloy from
550 to 850 °C. It was found that the deformation of γ′ precipitated
phase becomes anomaly and some secondary precipitates can be observed in
the matrix by increases the temperature. He et al. 27investigated the effect of temperature and strain amplitude on the
dislocation structure of M963 superalloy during high-temperature low
cycle fatigue. It was found that at the temperature between 700 and
800°C, dislocation shearing γ′ precipitates by dislocation pairs and
stacking faults at medium and high strain amplitudes and at the
temperature between 900 and 950°C, dislocations by-passing γ′
precipitates was the main deformation mode. Deng et al.28 investigated the fatigue behavior and damage
mechanisms of Inconel 718 under TMF and isothermal LCF. It was found
that the S-N data can be fitted by the Coffin-Manson equation under the
LCF condition, and the model based on hysteresis loop energy can be
predicted lifetime under TMF condition. Kashinga et al.29 investigated the effect of dwell time on the low
cycle fatigue of a directionally-solidified Ni-based superalloy using
both experimental and computational methods. It was shown that the
misorientations between columnar grains resulted in heterogeneous
deformation and localized stress concentrations, which became more
severe when the loading direction was perpendicular to a solidification
direction, showing the shorter lifetime observed. Mukherjee et al.30 investigated LCF behavior of Haynes 282 and its
correlation with microstructure at temperature of 760 °C. the results
showed predict fatigue life and close agreement between calculated and
measured life indicates validity of these models for the present
material when plastic strain energy density based life prediction model
were used. Rao et al. 31 investigated low cycle
fatigue behavior of Inconel 617 alloy at two temperatures conditions of
room temperature and 750 °C. the results showed the cyclic hardening
was occurred and also, the fatigue life decreased with increase in
temperature from room temperature to 750 °C. Zhang et al.12 investigated low cycle fatigue behavior of a
nickel-based single-crystal superalloy with a focus on the effect of
crystal orientation and temperature. It was found depending on the
strain amplitude, crystallographic orientation, and temperature either
cyclic hardening or softening was observed during the low cycle fatigue.
Also, low cycle fatigue life was reduced significantly by changing
loading orientation from [001] to [111] or increasing
temperature to 825 °C.
Despite many works that have been done to investigate the fatigue
behavior and failure mechanism, the contact between mechanical loading,
temperature, and fatigue life prediction of directionally-solidified
Ni-based materials actual indistinct. The present study aimed to
investigate isothermal low cycle fatigue behavior in the
directionally-solidified superalloy CM247 LC in the two temperatures of
600 °C to 800 °C. In this regard, to investigate the LCF behavior the
fatigue life and the fatigue degradation mechanisms were investigated
and discussed in the study.
- EXPERIMENTAL DETAILS
- Material evaluations
The material selected for the current study was the
directionally-solidified CM247 LC, which is hafnium (Hf)
containing a Ni-based superalloy developed for grain turbine blades.
According to the analysis results, the chemical composition of the
material under study is provided in Table 1. The material was produced
through vacuum induction melting. The CM247 LC superalloy was
directionally-solidified into bars of \(\phi\ 22\times 120\ mm\). The
material under study went through a three-step heat-treatment process.
Initially, the material was solution heat-treated at 1250 °C for 2.5 h,
followed by cooling. Precipitation of a uniform fine γ/γ′ microstructure
has resulted from this step. As the second step, the material was
heat-treated at 1080 °C for 4.5 h, which resulted in an optimized
distribution size of γ′ precipitates. For aging heat-treatment, the
third step was applied at 860 °C for 24 h. The mechanical and physical
properties of the cast directionally-solidified CM247 LC are presented
in Table 2. The processing and development of the
directionally-solidified superalloy CM247 LC are reported in Ref.32. Generally, in the DS and SC Ni-based materials,
the mechanical characteristics are extremely dependent on the 𝛾′
precipitates’ specifications including size, volume fraction, and
distribution 33,34. For the examination of
microstructures, the specimens for metallographic examination were cut
from the casted material through the EDM process and then the standard
procedure of polishing was performed on the samples. The etched process
was carried out with a solution of lactic acid, nitric acid, and
Hydrogen fluoride regent for a time of 120 s and at a voltage of 7 V.
The OLYMPUS optical microscope and a TESCAN field emission scanning
electron microscope (FE-SEM) were used to examine the fractured
surfaces. For this purpose, a rod specimen with dimensions of 15 mm in
diameter and 10 mm in length was prepared. For statistical analysis of
the metallurgical data of the SEM images an image analysis software
(metallographical image processing (MIP) package) application was used.
The optical micrographs of the microstructures in the longitudinal and
traverse sections of the grain axis of the CM247 LC superalloy are shown
in Fig. 1(a) and (b), respectively. As can be observed, the
microstructure involves both primary and secondary dendrites as well as
inter-dendritic regions. In Fig. 1(b), the mean of primary dendrite arm
spacing size is \(250\ \mu m\) and the mean of secondary dendrite arm
spacing size is \(210\ \mu m\). Furthermore, some distributed
microstructural features like eutectic γ/γ’ are observed in the
white-colored. The SEM micrographs of the microstructure are shown in
Fig. 2. Four EDS points were taken from formed (HF/Ta) C, γ/γ’ eutectic
phase, γ matrix, and γ’, respectively. The EDS results are shown in Fig.
2. As can be seen, γ’ intermetallic phase is distributed in a matrix of
γ. The average γ’ precipitate phase fraction was obtained as 62% and
the average size of the γ’ precipitate was achieved as \(0.5\ \mu m\).
The grain boundaries consisted of the main percentage of primary γ/γ’
eutectic phase and only a small percentage within the grains.
Low cycle fatigue test procedure
Directionally-solidified CM247 LC was used after the heat treatment
procedures. The present study used the button-head specimen for the LCF
test. The LCF specimens were machined with an extensometer gauge section
of 5.75 mm, with 27 mm in length. The low cycle fatigue test specimen
and its schematic are shown in Fig. 3. All specimens used in this study
were designed according to ASTM standard E606, considering the fact that
the solidification direction is parallel to the loading axis (along
[001] crystal orientation with a misorientation less than ± 10°) and
they were machined from casted and heat-treated material. The ratio of
the gage section area to the gage section diameter is a vital factor in
ASTM E606 strain-controlled low cycle fatigue test standard, which is
obtained by the LCF specimens under study. All casted specimens had the
same conditions and the same casting parameters. Also, before the LCF
test, all specimens were polished using SiC sandpaper to
eliminate the surface machining defects and marks. Afterward, a visual
inspection procedure was carried out to ensure that the residual marks
were cleaned for each fatigue specimen. Finally, specimens were cleaned
by acetone and dried by air. In this study, the strain-controlled low
cycle fatigue tests were performed at the two temperature conditions of
600 °C and 800 °C. All low cycle fatigue tests were performed using a\(100\ kN\) servo-hydraulic testing machine (Instron Co.), where the
machine had a low-profile load cell with a resolution of\(\pm\ 0.2\ kN\). Two N-type thermocouples with a resolution of\(\pm\ 2\ C\) were welded to the gage section in order to monitor the
temperature gradient. A radio frequency induction system was used to
apply heat to the specimen, while a PID controller by using the
closed-loop feedback control was kept the LCF specimen temperature.
Approximately 2 minutes after the set-point temperature became stable,
the LCF tests were started. The RF induction system applied maximum
temperature on the gage section of LCF test specimens. The mechanical
strain amplitudes were 0.4, 0.6, 0.8, and 1%. Therefore, symmetrical
triangular waveform and a constant total strain amplitude were kept
throughout the totally-reversed loading conditions (\(R\ =\ -1\)).
In each LCF test, temperature and mechanical strain amplitude are
different. The strain rates were held constant during loading and
unloading slopes. Strain (mm/mm), Force (N), displacement (mm),
temperature (°C), and the values of the peak forces were monitored and
recorded for each cycle. Subsequently, to determine failure, the stress
history was periodically analyzed. Fig. 4 shows the mechanical strain
amplitudes and temperatures applied to the LCF tests.
- MODELING WORK
- Low cycle fatigue life
A strain-controlled fatigue curve includes elastic and plastic strain
data 35. In isothermal problems, the Basquin’s
equation (cyclic stress) and the Coffin–Manson equation (cyclic plastic
strain) use the elastic and plastic strain data, respectively36. The Basquin’s equation is more appropriate and it
is used for the high cycle fatigue (HCF) regime 37,38:
\(\frac{{\Delta\varepsilon}_{e}}{2}=\frac{{\sigma^{\prime}}_{f}}{E}\left(2N_{f}\right)^{b}\)(1)
On the other hand, the Coffin-Manson equation is used for the low cycle
fatigue (LCF) regime and defines the reliance of fatigue life under
isothermal loading on plastic strain amplitude 39,40:
\(\frac{{\varepsilon}_{p}}{2}=\varepsilon_{f}^{{}^{\prime}}\left(2N_{i}\right)^{c}\)(2)
Where εp is the plastic strain amplitudes, c is the
strain ductility exponent, \(\varepsilon_{f}^{{}^{\prime}}\ \)is the strain
ductility coefficient, and Ni is the cycle to crack
initiations. Nevertheless, sometimes in the LCF problems, the
Coffin-Manson law has an error in fatigue life prediction, which is
assumed because of the unsteady loading condition 41and unexpected failure modes 42,43. As a consequence,
the Coffin-Manson equation is not suitable for predicting the lifetime
for LCF, when a variable likewise temperature is introduced into the
problem 44. Furthermore, a suitable model for fatigue
life prediction has to accurately estimate the material’s fatigue life,
and it should also be evaluating fatigue property. Nevertheless, the
Basquin equation is suitable for calculating the material’s fatigue
behavior based on the cyclic stress amplitude, while the Coffin-Manson
equation is suitable for assessing the material’s fatigue property based
on the plastic strain. Therefore, neither of them would be suitable for
the evaluation of the integrated fatigue property. Thus, a new model is
identified and developed based on hysteresis energy, which is presented
below 45:
\(D_{i}=\left(\frac{W_{i}}{W_{0}}\right)^{\beta}\) (3)
\(D=\sum_{i=1}^{N_{f}}{D_{i}={\sum_{i=1}^{N_{f}}{(\frac{W_{i}}{W_{0}}})}^{\beta}=1}\)(4)
Where, Di and Wi are the damage
parameter and the hysteresis energy of the ith cycle,
respectively. W0 and b are material constants, and
W0 is the fundamental fatigue toughness. The relation
between plastic strain and hysteresis energy is shown in Fig. 5. As can
be observed, Wi is the stress-strain hysteresis loop
area for each cycle, as shown with the red zone, and the value of
Δεp·Δσ is the gray zone. The fatigue damage capacity of
the material is defined by W0. The b parameter is the
damage transition exponent. It is described as the ability to transform
mechanical work into operative damage to the materials which, in other
words, is described as the susceptibility that the damage reacts to the
diversity of loading conditions 45. Moreover, the
fatigue damage is denoted by D. In this model, when fatigue damage D
equals 1, the material fails. The simplified model is written as
follows:
\(D=\ \sum_{i=1}^{N_{f}}{D_{i}=\sum_{i=1}^{N_{f}}{\left(\frac{W_{s}}{W_{0}}\right)^{\beta}=1}}\)(5)
Where Ws is the half life cycle hysteresis energy, the
Eq. (5) can be converted into the Coffin-Manson equation as the
following:
\(W_{s}=W_{0}\times N_{f}^{\frac{-1}{\beta}}\) (6)
However, the Δσ/2, Δεp/2, and WS are
obtained from the fatigue data.
RESULTS AND DISCUSSIONS
The results and discussion of isothermal low cycle fatigue behavior in
the directionally-solidified CM247 LC superalloy is presented in this
section. Based on the LCF test procedure under two temperature
conditions of 600 °C and 800 °C, the low cycle fatigue behavior of the
directionally-solidified CM247 LC superalloy was investigated. The low
cycle fatigue data obtained by experimental tests as well as the
relationship between half-life stress range, cycles to failure under
constant strain amplitude, degree of softening, and the two temperature
conditions are presented in Table 3. The values of stress range
amplitude are obtained from the hysteresis curve of the half-life cycle.
Based on the obtained data of the Table 3, once the strain amplitude is
increased, the stress range is increased and the fatigue life (cycle to
failure) is decreased.
A set of strain-controlled experiments were performed to evaluate the
influence of temperature on the fatigue life under axial loading in the
direction along the solidification axis. In the low cycle fatigue, the
hysteresis loop (cyclic stress-strain behavior) indicates the
relationship between the strain range and the stress range for each
cycle. The stress-strain hysteresis curves can be achieved by using the
companion method which is connecting the vertices of the stable or
half-life cycle loops at different strain amplitudes. This method is
used to determine the stable cyclic stress-strain behavior of the
material throughout the low cycle fatigue test. The hysteresis loops for
LCF tests are plotted in Fig. 6 under two temperature conditions and the
mechanical strain amplitudes of 0.4, 0.6, 0.8, and 1 %. As shown in
Fig. 6, at the temperature condition of 600 °C, at the strain amplitude
of 0.4% the peak stress value decreased from 994 MPa at the first cycle
to 831 MPa at the half-life cycle, at the strain amplitude of 0.6% it
decreased from 1059 MPa at the first cycle to 880 MPa at the half-life
cycle, at the strain amplitude of 0.8% it decreased from 1144 MPa at
the first cycle to 1007 MPa at the half-life cycle, and the strain
amplitude of 1% it declined from 1205 MPa at the first cycle to 1058
MPa at the half-life cycle. At the temperature condition of 800 °C, at
the strain amplitude of 0.4%, the peak stress value decreased from 914
MPa to 655 MPa from the first cycle to the half-life cycle, at the
strain amplitude of 0.6% it decreased from 963 MPa to 748 MPa from the
first cycle to the half-life cycle, at the strain of 0.8% it decreased
from 1095 MPa to 836 MPa from the first cycle to the half-life cycle,
and at the strain amplitude of 1%, it declines from 1130 MPa to 825 MPa
from the first cycle to the half-life cycle. The hysteresis loops showed
the stress value is increased with increasing the strain amplitude, and
while the temperature is increased from 600 °C to 800 °C, of all strain
amplitudes the temperature-induced softening is observed and the
material CM247 LC shows a softening behavior. Therefore, the cyclic
stress-strain curves showed that the cyclic softening occurs at both
temperature conditions at all the strain amplitudes (Fig. 6). A
comparison of the experimental results, depicted in Fig. 6, for LCF
tests, reveals that the fatigue life is decreased when the temperature
shifts from 600 °C to 800 °C for all strain amplitudes. The occurrence
of the cyclic softening behavior in precipitation-hardened alloys
depends on some reasons which are the formation of dislocation networks
at the γ/γ’ interface, shearing of precipitates, dissolution of
precipitates, and coarsening of the precipitates11,46,47. The cyclic stress-strain hysteresis loops
obtained with the experimental test data for two temperature conditions
indicated that, at a lower temperature, the material had higher
stiffness when the load was aligned along the solidification direction
of the material. The plastic strain range can be obtained from the width
of the hysteresis loop (Fig. 5). For an assumed strain amplitude, the
plastic strain amplitude increased with transferred the temperature from
600 °C to 800 °C.
Based on the hysteresis loops obtained from the experimental test of
CM247 LC, the degree of softening is determined using the Eq. (7) as
follows 48:
\(\text{Degree\ of}\ \text{softening}=\ \frac{(\frac{\text{stress\ range}}{2}at\ half\ life\ cycle)\ -(\frac{\text{stress\ range}}{2}at\ the\ first\ cycle)}{\frac{\text{stress\ range}}{2}\text{at\ the\ first\ cycle}}\)(7)
The degrees of softening at the half-life cycle for all LCF tests are
presented in Table 3. The variations of the degree of softening vs
strain amplitudes are shown in Fig. 7. For temperature condition of 600
°C, the degree of softening at the strain amplitude of 0.4% to 1% is
increased from 18.2% to 23.1%, and for temperature condition of 800
°C, is increased from 20.6% to 39.8%. As can be observed, that through
increasing the strain amplitude the degree of softening increased for
all strain amplitudes and temperatures, and also, the degree of
softening is increased with increasing the temperature from 600 °C to
800 °C.
The cyclic stress amplitude versus the number of cycles at different
strain amplitudes is called the cyclic stress response that is a
significant parameter in low cycle fatigue, and describes the stress
amplitude at the final stress level, while also showing the deformation
throughout the cyclic loading of the material. Cyclic stress response
under two temperature conditions for individual strain amplitudes of the
directionally-solidified CM247 LC is shown in Fig. 8. Peak stress values
are used to show the cyclic stress response of the material. The cyclic
stress response showed at both temperature conditions and all the strain
amplitudes of the CM247 LC showed a normal cyclic response and by
increasing the strain amplitude the cyclic stress amplitude is
increased. Throughout the cyclic loading, the material reveals softening
behavior until the time when the overload step occurs. As shown in Fig.
8 at both temperatures and all the strain amplitudes, curves indicate
that by increasing the strain amplitude the cyclic stress amplitude is
also increased. For both temperature conditions, the cyclic softening
was seen throughout the test process and followed by a rapid drop in the
stress due to the initiation of microcracks and their growth leading to
final failure. Also, when the temperature transfers from 600 °C to 800
°C, the curves show that the stresses are steadily decreasing.
Furthermore, as shown in Fig. 8, as the softening increases the lifespan
would be less. The cyclic stress response hypothesis describes that
cyclic softening is expected in the materials with
SUTS/SYS less than 1.4 which is in line
with the hypothesis. When gradual strain softening is occurred at the
lower temperatures, rapid strain softening at higher temperatures is
observed and the degree of work hardening
(SUTS/SYS ratio) is higher at 800 °C
compared with that at 600 °C. In the precipitated hardening superalloys,
the dislocation–precipitate interaction mechanisms determine the cyclic
stress response behavior 12. While the principal
mechanism is precipitate shearing, the initial hardening is mostly
followed by a period of softening, or only softening can be observed.
In order to predict fatigue life and evaluate the low cycle fatigue
behavior, the Coffin-Manson model, based on plastic deformation, and a
model based on the hysteresis energy criterion is used. The values of
plastic strain amplitude throughout the cyclic deformation are shown in
Fig. 9 and the hysteresis energy during cyclic deformation for all LCF
tests are shown in Fig. 10. The values of plastic strain amplitude
throughout the cyclic deformation (Fig. 9) demonstrated that the plastic
strain amplitudes increase after a small cycling interval for all strain
amplitudes at 600 °C and 800 °C. The results obtained from Fig. 9 show a
monotonic increase in the plastic strain amplitude with the increasing
number of cycles, indicating the occurrence of cyclic softening in the
directionally-solidified CM247 LC. Based on the obtained results from
Fig. 10, it can be seen that hysteresis energy reveals cyclic stability
right after a small cycling interval for all strain amplitudes,
indicating a similar behavior to that of the plastic strain amplitude.
The fitted curves of the fatigue life prediction models based on the
Coffin-Manson equation for obtained results from Fig. 9 are plotted in
Fig. 11. Following the experimental LCF data, the fatigue lifetime
versus plastic strain amplitude is fitted following the Coffin-Manson
law (Fig. 11). The values of R2 are fitted based on
the Coffin-Manson equation and are 0.91 and 0.95 under 600 °C and 800 °C
LCF test conditions, respectively. A comparison of plastic strain
amplitude-life relationship at the two temperature conditions indicates
that the fatigue life decreased by transferring the temperature from 600
°C to 800 °C. The fitted curves of the fatigue life prediction models
based on the hysteresis energy model for the obtained results from Fig.
10 are plotted in Fig. 12. Following the experimental LCF data, the
fatigue lifetime versus hysteresis energy is fitted according to the
hysteresis energy model and plotted in Fig. 12. The values of
R2 are fitted based on the hysteresis energy model
equation and are obtained as 0.92 and 0.94 under 600 °C and 800 °C LCF
test conditions, respectively. This hysteresis energy model shows good
ability to calculate the fatigue life of low cycle fatigue mechanism at
high-temperature conditions. The plastic strain amplitude-life and
hysteresis energy-life relationship at both temperatures showed that the
fatigue life is decreased when the temperature increases.
After LCF testing and in order to investigate the failure mechanisms in
this study, fracture surfaces of the fractured low cycle fatigue test
specimens were separately examined and analyzed using field emission
scanning electron microscopy (FE-SEM). The fracture surface examination
using SEM for the low cycle fatigue specimens of the two temperature
test conditions are shown in Fig. 13. As shown in Fig. 13, the
inter-granular cracking was observed in the fractured surface of the
specimen for LCF test specimens. Cavities and inter-granular cracks were
observed during low cycle fatigue tests performed in all specimens.
According to the examination of all the LCF specimens, fracture surfaces
were generally perpendicular to the loading axis. Due to the loading of
LCF specimens along the solidification direction, the fracture surface
features, depicted in Fig. 13, were conceivably caused by inter-granular
cracking. Moreover, the fatigue marks in fractured LCF test specimens
and the final stage of the fatigue fracture region are shown in Fig. 14
and Fig. 15, respectively. As shown in Fig. 14 the fatigue striations,
as well as the cleavage facets with inter-granular cracks, appear on the
fractured surface. The striations are defined and uniformly spaced at
high strain amplitude and high temperature compared to those at low
strain amplitude and low temperature. At low temperatures, the
striations are not uniform and appear partially deformed or ductile. The
rapid decline in peak stress amplitude observed before the failure can
mostly be attributed to the fast propagation and ultimate coalescence of
cracks, resulting in a decrement in the load-bearing capacity.
Furthermore, once the temperature and stress amplitude are increased the
decrease in the area of fatigue fracture would be observed. The reason
for this phenomenon is the more inter-striation spacing as compared to
that at lower strain amplitudes and temperatures, which results from the
high cyclic stress response, and so, fewer cycles would fit in the
mentioned area. The rapid fracture surfaces of the LCF test specimens of
strain amplitudes of 0.6% and 1% at both temperature conditions are
shown in Fig.15. As demonstrated in these figures, some dimples and
plastic deformation marks are observed in the rapidly fractured region.
The fracture region was representative of the ductile fracture mechanism
at all strain amplitudes at both temperatures. Comparison of fracture
surfaces of two temperature conditions and all strain amplitudes showed
similar behaviors in the fracture characteristics.
CONCLUSIONS
Isothermal low cycle fatigue behavior of the directionally-solidified
CM247 LC Ni-based superalloy is studied under two temperature conditions
of 600 °C and 800 °C. The main conclusions obtained in the current study
can be written as follows:
- Comparison of low cycle fatigue under two temperature conditions of
600 °C and 800 °C showed that the fatigue life decreases with
increased temperature from 600 °C to 800 °C.
- The cyclic hysteresis loops obtained in the experimental tests for two
temperature conditions of LCF tests showed that the material had
higher stiffness in lower temperatures, and with changing temperature
from 600 °C to 800 °C, a cyclic softening was observed.
- The results of plastic strain amplitude tests showed a monotonic
increase in the plastic strain amplitude with the increasing number of
cycles, indicating that cyclic softening occurred in the LCF test
process.
- The failure modes that occurred on LCF specimens are inter-granular
fractures, and similar striation morphologies were observed in LCF
test specimens.
- The fatigue life of the directionally-solidified CM247 LC under the
isothermal LCF test condition can be predicted by the Coffin-Manson
and hysteresis energy model.
ACKNOWLEDGEMENTS
The first author would like to expresses his great appreciation to Mrs.
Z. Khoshkhou-Gilavaie from the University of Tehran for editorial helps
that greatly improved the manuscript.