Analysis of damage behavior based on EBSD method and the law of fracture life under creep–fatigue conditions for the polycrystalline nickel-base superalloy

Abstract

To clarify the experimental background of characteristics of fracture life of the investment-cast polycrystalline Nickel-base superalloy IN738LC (typical gas turbine blade’s materials), in-situ observational tests under creep–fatigue conditions were carried out, and the effects of both the cycle- and the time-dependent mechanisms on the fracture life were investigated. Creep ductility, stress holding time and temperature were combined with the relationship between the inverse value of fracture life and applied load frequency as a promoting factor of the time-dependent mechanism. Then a three-dimensional curved surface representation of fracture life under arbitrary creep–fatigue conditions has been proposed theoretically according to the non-equilibrium science. As the load frequency decreases or stress holding time and temperature increases, the time dependent mechanism begins to play a role in the manner of creep–fatigue interaction. In this region, the characteristics of the Relative Notch Opening Displacement (RNOD) and crack growth behavior change from a cycle-dependent mechanism caused by fatigue to a time-dependent mechanism caused by creep through an unstable equilibrium transition region. In this paper, comparison of the three-dimensional curved surface representation of fracture life which has unique characteristics and the actual damage behavior measured by the Electron Back Scattered Diffraction (EBSD) method was carried out. As a result, misorientation analysis based on the EBSD method has clarified that the unstable equilibrium transition region in the three-dimensional curved surface was the result of creep damage concentration in the vicinity of notches by a load frequency which has the effect of limiting the damage region.

Keywords

Nickel-base superalloy EBSD misorientation law of fracture life creep–fatigue conditions

1 Introduction

Heat-resistant materials need to be highly reliable and are used for important parts in high temperature components [1]. However, these materials are subject to complex thermal and mechanical histories during a typical cycle of operation [2,3]: consisting of startup, steady-state operation and shutdown. Gas turbines are also used under a wide range of conditions from base load to peak load depending on energy demand. In other words, heat-resistant materials, like nickel-base superalloys, are used under the complex conditions of high-temperature creep and fatigue interaction. Accordingly, the life-limiting factors of heat-resistant materials are predictable as they change from high-temperature creep to periodical thermal and mechanical fatigue. In maintaining operational safety and minimizing maintenance costs, it is necessary to clarify the characteristics of fracture life and improve life-assessment techniques on a unit-specific basis taking into account the effect of both creep and fatigue interaction on fracture life [4]. Meanwhile, intelligent scheduling of inspections also requires proper techniques for assessing the life expenditure of components. The correct choice of inspection schedules and methods can prevent accidents. From the viewpoint of performing reasonable inspections, it is important to predict both the damage site and crack initiation mechanism of the components [5].

The authors have proposed the method of separate estimation of the effect of creep and fatigue, namely whether the fracture life, $t_{f}$ , is dominated by the time-dependent mechanism or the cycle-dependent mechanism, using the experimental relationship between the inverse value of fracture life, $1 / t_{f}$ , and the load frequency, f [6–8]. There are some papers on fracture life [6,9] which investigated the effect of creep and fatigue interaction. However, it is difficult to comprehend the characteristics of load frequency of the fracture life under creep–fatigue interaction, since the characteristics indicate diversity of such factors as stress, stress holding time, material properties and temperatures. In many cases, this relationship occurs under complex interactive conditions due to the cycle-dependent and the time-dependent mechanisms. Against this backdrop, the authors have found that creep ductility, stress holding time, $t_{H}$ , and temperature, T, were unified as a promoting factor of the time-dependent mechanism: P, which was then combined with the relationship between the inverse value of fracture life, $1 / t_{f}$ , and applied load frequency, f, and they have proposed a three-dimensional curved surface representation of fracture life under arbitrary creep–fatigue conditions [10]. Furthermore, to perform the theoretical representation of this three-dimensional curved surface, it was analyzed according to the non-equilibrium science [11–15].

On the other hand, in contrast to the above theoretical analysis, the authors have investigated usability of the EBSD (Electron Back Scattered Diffraction) method [16–20] which is becoming established as a damage evaluation method of material in order to evaluate creep damage of Nickel-base superalloy [21,22] focusing on the effect of test temperatures, fatigue interaction [23,24] and geometrical influence at stress concentrated regions [25,26]. In this study, using the polycrystalline Nickel-base superalloy, microscopic analysis of damage behavior based on the EBSD method was carried out to clarify characteristics of creep–fatigue interaction represented in the three-dimensional curved surface of fracture life derived from experiments and theoretical analysis.

2 Representation of curved surface characteristic of load frequency of fatigue life based on the non-equilibrium science [10]

The schematic illustration of the separate estimation method for the cycle-dependent and the time-dependent mechanism is shown in Fig. 1 [7]. When the fracture life, $t_{f}$ , is dominated by the cycle-dependent mechanism, the relationship between log $1 / t_{f}$ and log f is linear with a gradient of 45 degrees [6–8]. In contrast, when the fracture life, $t_{f}$ , is dominated by the time-dependent mechanism, the relationship between log $1 / t_{f}$ and log f is linear and parallel to the log f axis, i.e. there is a horizontal linear relationship between them. Consequently, the experimental relationship between $1 / t_{f}$ and f was characterized whether the fracture life, $t_{f}$ , is dominated by the cycle-dependent mechanisms, the time-dependent mechanisms or by both these competitive mechanisms.

Fig. 1.

Separate estimation method for the cycle-dependent mechanism and the time-dependent mechanism.

Fig. 2.

Three-dimensional curved surface representation of load frequency characteristics for creep–fatigue crack growth life. (Colors are visible in the online version of the article; https://dx-doi-org.web.bisu.edu.cn/10.3233/SFC-150182.)

From the results mentioned in Section 1, creep ductility, stress-holding time and temperature have a similar effect which promotes the occurrence of the time-dependent mechanism of the fracture life, $t_{f}$ . Besides the axis of the inverse value of fracture life, $1 / t_{f}$ , and applied load frequency, f, a third axis was added as the time-dependant promoting factor, P, and the three-dimensional assessment space was proposed as shown in Fig. 2(a). And then, the characteristic curved surface shown in Fig. 2(b) was obtained by the experimental data. Using this curved surface and obtaining the relationship between log $1 / t_{f}$ and log f under a specified condition of creep ductility, $t_{H}$ and T, the relationship between log $1 / t_{f}$ and log f under the other conditions can be qualitatively predicted. For example, when the relationship is dominated by the cycle-dependent mechanism under the condition of $T = T_{1}$ , the relationship under the condition of $T ⩽ T_{1}$ is found (without conducting other experiments) to be dominated by the cycle-dependent mechanism. On the other hand, when the relationship is dominated by the time-dependent mechanism under the condition of $T = T_{2}$ , the relationship under the condition of $T ⩾ T_{2}$ is found to be dominated by the time-dependent mechanism. The relationship between log $1 / t_{f}$ and log f in the region of $T_{1} < T < T_{2}$ shows the interactive characteristics between creep and fatigue which shows non-linear characteristics.

Fig. 3.

A curved surface obtained by a catastrophic function.

The transition characteristic of fracture life from the cycle-dependent mechanism to the time-dependent mechanism shows a discontinuous characteristic with an inflexion point, which is similar to the transition characteristic of a catastrophic function, as shown in Fig. 3. The authors proposed a theoretical foundation to represent this curved surface characteristic of fracture life under the creep–fatigue conditions, where the function $F (z)$ was defined by Eq. (1). This equation is obtained from the catastrophic theory to fit to the experiment results: $F (z) = ω_{1} \frac{3 z^{3} - z}{z^{4 (1 - 1 / (2 exp (P)))}} - ω_{2} ln z,$ (1) where $ω_{1}$ , and $ω_{2}$ are given by Eqs (2) and (3), respectively: $\begin{array}{rclr} ω_{1} = 1 - (1 - \frac{P}{2}), & (2) \\ ω_{2} = 1 - \frac{P}{2}, & (3) \end{array}$ where z is given by Eq. (4), which is the non-dimensional value of load frequency; $f_{0}$ is the value of control frequency which indicates perfect time dependent characteristic ( $f_{0} = 10^{- 3} Hz$ in this study): $z = \frac{f}{f_{0}} .$ (4)P is a factor describing the influence of the time-dependent mechanism; it assumes a value ranging from 0 to 2 ( $0 ⩽ P ⩽ 2$ ). The first term of Eq. (1) represents the effect of creep (time-dependent effect) and the interaction of creep and fatigue effects by the value of P, ranging from 0 to 2, which coexists during the processes of loading and unloading. The second term of Eq. (1) represents the effect of fatigue (cycle-dependent effect). The values of the weight coefficients ( $ω_{1}$ , $ω_{2}$ ) change from $(0, 1)$ to $(1, 0)$ due to the change of the value of P, ranging from 0 to 2. These coefficients change the weight of the effects of creep–fatigue interaction on its fracture life. That is, the fact that $P = 0$ shows a perfect cycle-dependent mechanism of fracture life and the second term of Eq. (1) is significant. Similarly, the fact that $P = 2$ shows a perfect time-dependent mechanism of fracture life and the first term of Eq. (1) is also significant. The condition of $0 < P < 2$ shows the effects of creep–fatigue interaction (the first and second terms of Eq. (1)), which coexist during the processes of loading and unloading, as shown in Fig. 4. Using Eq. (1), Eq. (5) is derived as an equation which satisfies the boundary conditions of the cycle-dependent mechanism for $P = 0$ and of the time-dependent mechanism for $P = 2$ , $z < z_{crit}$ : $\frac{1}{t_{f}} \approx 1 + exp (- f (z)) .$ (5)

Using Eqs (1)–(5), the characteristics of fracture life under the creep–fatigue condition were derived and were compared with experimental characteristics. Based on this foundation, we also derived the lower limit of temperature at which the creep crack growth appears and the predicting law of fracture life under the creep–fatigue condition. In detail, the characteristic of load frequency of fracture life and the degree of involvement of the time-dependent mechanism can be quantified by the value of P given by Eq. (1). That is, the value of P is considered as a representative parameter of the degree of involvement of the time-dependent mechanism of fracture life. Therefore, in the case of $P = 0$ , fracture life is regarded as perfect domination of the cycle-dependent mechanism. In this region ( $P = 0$ ), even though the stress holding time, $t_{H}$ , increases, it is theoretically not feasible that the appearance of the time-dependent mechanism occurs. This fact then yields the theoretical threshold temperature at which the creep crack growth mechanism appears. Therefore, if the relationship between P and temperature, T, is obtained from experiment results, this proposed theory enables us to extrapolate the temperature at $P = 0$ , that is, the theoretical threshold temperature $T_{1}$ as shown in Fig. 5. It is not theoretically feasible for the cause of the time-dependent mechanism to occur when the temperature T is lower than $T_{1}$ . In this case, the critical temperature can be predicted as 678°C (951 K). IN738LC is considered to show a superior creep resistance lower than 678°C. In addition, it can be predicted that fatigue crack growth mechanisms will tend to appear more readily than creep crack growth mechanisms at temperatures less than 678°C.

Fig. 4.

Effects of time-dependent and cycle-dependent mechanisms on the stress wave form of load frequency.

Fig. 5.

Relationship between P parameter and temperature.

Initially, the above curved surface of fracture life was obtained by theoretical analysis to represent the experimental results of some materials, however, it was reported that many materials show similar characteristics [11–15,27,28]. Therefore it is considered that these characteristics of the curved surface represent the basic properties of the material on creep fatigue interaction. Furthermore, by the microscopic damage analysis using the EBSD method, the relationship between the actual creep–fatigue damage behavior and the characteristic shape of the curved surface will be clarified in this study.

3 Material, experimental procedure and EBSD measurement

The material used is the investment-cast polycrystalline nickel-base superalloy IN738LC, commonly used as a heat-resistant material of gas turbines. The chemical composition and the mechanical properties of IN738LC are shown in Tables 1 and 2, respectively. The specimen used is a double edge notched (DEN) specimen to observe crack growth behavior as shown in Fig. 6.

Creep-fatigue tests were carried out using two types of in-situ observational testing machines which were manufactured by SIMAZU SEISAKUSYO and YONEKURA SEISAKUSYO in Japan, the same concept as the machine proposed by Yokobori [8]. A schematic illustration of the machine manufactured by SIMAZU SEISAKUSYO is shown in Fig. 7 [11]. The notch opening value, crack initiation and crack growth behavior were observed continuously while running the tests without interruption, by using a high-temperature microscope through a peeping window. Specimens were heated with an infrared condensed lamp and temperatures were continuously monitored with a spot-welded thermocouple on the specimen surface. Temperatures were kept constant within ±1°C in inert gas (99.9999% Ar). The load waves were controlled as trapezoidal waves as shown in Fig. 8. The load frequencies of the load wave were changed from 0.05 Hz to 0.005 Hz by changing the stress holding time,

t_{H}

. The gross section stress,

σ_{g}

, corresponds to the maximum value of tensile stress,

σ_{max}

. The relative notch opening displacement (RNOD) [29] was calculated as follows Eq. (6) and was dealt with as well as creep strain:

RNOD = \frac{φ - φ_{0}}{φ_{0}} = \frac{Δ φ}{φ_{0}},

(6) where φ and

φ_{0}

is the notch opening value and its initial value, respectively. As mentioned above, the geometrical influence for creep damage evaluation has been investigated by using these different stress concentrated specimens. The experimental conditions and results are shown in Table 3.

Table 1

Chemical composition of IN738LC (Ni base, mass%)

C	Si	Mn	P	S	Cu	Cr	Mo
0.1	0.13	0.05	0.003	0.002	0.01	15.9	1.73

Co	W	Ti	Al	Nb	Zr	Fe	Ta
8.25	2.77	3.29	3.33	0.72	0.06	0.33	1.56

Table 2

Mechanical properties of IN738LC

Temperature (°C)	0.2% proof stress (MPa)	Tensile stress (MPa)	Elongation (%)	R.A. (%)
RT	840	984	3.0	6.5
700	758	1105	7.7	11.0
760	702	995	8.3	11.6
850	477	744	10.0	17.0

Fig. 6.

Geometry and size of a DEN (Double Edge Notched) specimen. (Colors are visible in the online version of the article; https://dx-doi-org.web.bisu.edu.cn/10.3233/SFC-150182.)

Fig. 7.

Schematic illustration of in-situ observational machine. (a) Whole structures, (b) surrounds of the furnace. (Colors are visible in the online version of the article; https://dx-doi-org.web.bisu.edu.cn/10.3233/SFC-150182.)

Fig. 8.

Applied stress waveform.

Crystal orientation measurements were made with an EBSD-SEM system, comprising a TSL EBSD system using an orientation imaging microscopy (OIM Version 5) interfaced to a JEOL JSM-7001F with a field emission electron gun. This system can obtain a crystal orientation map automatically by digitally scanning the electron probe over a rectangular area of the specimen surface which is tilted at 70° from the horizontal.

In the case of smooth specimens, samples were sectioned parallel to the stress axis and first polished to a mirror surface with 0.05 µm colloidal silica solution to remove the worked layer before EBSD measurements. Orientation imaging maps were taken over a rectangular area of about 4 mm square. The step size of the scanning was 5 µm at an accelerating voltage of 25 kV in the axial and transverse directions.

The kernel average misorientation (KAM) was employed as a misorientation function. The KAM is the local average misorientation of adjacent measuring points and shows damage in microscopic area.

Table 3

Test conditions, fracture lives and symbols

Temperature (°C)	$σ_{g}$ (MPa)	$t_{R} (s) = t_{D} (s)$	$t_{H}$ (s)	f (Hz)	$t_{f}$ (h)	Symbol
740	490	10	0	0.05	26	△
		10	180	0.005	36	□
		CREEP			79	○
830	294	10	0	0.05	120	▲
		10	180	0.005	56	■
		CREEP			110	●
830	230	CREEP			903	◆

Note: $t_{R}$ – stress rising time, $t_{H}$ – stress holding time, $t_{D}$ – stress descending time.

Fig. 9.

Characteristics of the RNOD at 740°C under the creep–fatigue condition.

Fig. 10.

Characteristics of the RNOD at 830°C under the creep–fatigue condition.

4 Result

The RNOD curves at 740°C and 830°C are shown in Figs 9 and 10, respectively. For either condition, the displacement in the case of triangular net fatigue conditions is small, however a displacement trend is equivalent to the creep in cases of creep–fatigue condition ( $t_{H} = 180 s$ ). Also, as seen from Fig. 10, the stress did not affect the amount of displacement [24].

In conjunction with past results [11], the relationship between the inverse value of the fracture life ( $1 / t_{f}$ ) and the load frequency (f) is shown in Fig. 11. The stress level differs for cases of 830°C or more but to infer the overall trend and to calculate the effect of stress levels on fracture life from the past experimental data, one-tenth of the fracture lives for cases of 830°C or more is plotted on the diagram.

Fig. 11.

Relationship between inverse value of fracture life and applied load frequency focused in the characteristics of creep–fatigue interaction (830°C) and the cycle dependent (740°C). (Colors are visible in the online version of the article; https://dx-doi-org.web.bisu.edu.cn/10.3233/SFC-150182.)

Fig. 12.

Crack growth behavior at $t / t_{f} = 0.99$ under 740°C. (a) 740°C × 490 MPa, $t_{H} = 0 s$ , (b) 740°C × 490 MPa, $t_{H} = 180 s$ , (c) 740°C × 490 MPa, Creep. (Colors are visible in the online version of the article; https://dx-doi-org.web.bisu.edu.cn/10.3233/SFC-150182.)

In the case of 740°C, load frequency shows the cycle-dependent characteristics in the range of 10⁻¹ to 10⁻³ Hz, however, as load frequency decreases at 830°C, transition behavior to the time-dependent characteristics with the maximum point was observed. It is considered that the creep–fatigue interaction region, in other words, the nonlinear interacted area means the competition of dislocation motion increased by repeated loading and vacancy diffusion due to constant stress. It is an interesting fact that the degree of interaction of dislocation motion and vacancy diffusion shows both short life and long life of crack growth.

Next, Fig. 12 shows the observation results of crack growth morphology of end-of-life by optical microscope at 740°C and in Fig. 13 the results at 830°C. In the case of 830°C, since it was not able to interrupt the test immediately before rupture under the 294 MPa creep condition, the observation result of the 230 MPa which was equivalent to that of 294 MPa was employed. (The result of the 230 MPa was also employed in Fig. 14(a) on the same reason.) The creep and creep–fatigue conditions at 830°C show multiple cracks at grain boundaries, whereas one main crack dominated the damage behavior in either condition in the case of 740°C. Therefore, it was confirmed that the influence of the stress wave on crack growth behavior varies depending on temperatures and elevating temperature promotes intergranular damage.

Fig. 13.

Crack growth behavior at $t / t_{f} = 0.99$ under 830°C. (a) 830°C × 294 MPa, $t_{H} = 0 s$ , (b) 830°C × 294 MPa, $t_{H} = 180 s$ , (c) 830°C × 230 MPa, Creep. (Colors are visible in the online version of the article; https://dx-doi-org.web.bisu.edu.cn/10.3233/SFC-150182.)

Fig. 14.

KAM maps at $t / t_{f} \approx 0.99$ . (a) 830°C × 230 MPa, Creep, (b) 830°C × 294 MPa, $t_{H} = 180 s$ , (c) 830°C × 294 MPa, $t_{H} = 0 s$ , (d) 740°C × 490 MPa, Creep, (e) 740°C × 490 MPa, $t_{H} = 180 s$ , (f) 740°C × 490 MPa, $t_{H} = 0 s$ . (Colors are visible in the online version of the article; https://dx-doi-org.web.bisu.edu.cn/10.3233/SFC-150182.)

The EBSD observations at the-end-of-life corresponding to the results of the load frequency characteristics as shown in Fig. 11 are shown in Fig. 14. In the case of 830°C where change occurred to the time-dependent mechanism from the cycle-dependent mechanism in a transition region due to a decrease in load frequency, the crack growth manifested changes to intergranular cracking from transgranular cracking perpendicular to the stress and expansion of the damaged area. On the other hand, in the case of 740°C, where the transition region did not manifest due to decreasing load frequency, the damaged area is limited relatively closely to notches, and intragranular damage due to stress concentration (slip deformation) is also observed in creep conditions.

In other words, the degree of creep damage due to the grain boundary diffusion of vacancies differs under the conditions of 740°C and 830°C, where the effect is of course greater at higher temperature conditions, and damage is promoted over a wide area. On the other hand, the cyclic effect probably has an effect of limiting the damaged area [8] and damage only occurred at the notch or crack tip under pure fatigue conditions. However, the effect is restricted to a limited area of intergranular damage under creep–fatigue conditions ( $t_{H} = 180 s$ ) and it is considered that the concentration of intergranular damage in a relatively small area led to fracture at an early stage (the appearance of the transition region) in the 830°C conditions. In contrast, for the 740°C conditions, the intergranular damage mechanism due to vacancy diffusion did not prominently appear for low temperatures, therefore the same creep–fatigue conditions ( $t_{H} = 180 s$ ) did not promote intergranular damage, nor did the transition region manifest due to the short life. However, because 740°C is higher than the perfect cycle-dependent mechanism temperature of the material, it is not the intention of the authors to deny the existence of a transition region at lower load frequencies.

Finally, using Fig. 15, the relationship between the three-dimensional curved surface of fracture life and environments where actual gas turbines are used is investigated. The back side of the curved surface represents a high-temperature region which is the time-dependent mechanism, the front side is a low-temperature region which is the cycle-dependent mechanism. In addition, the right side is a high-cycle region, the left side is a low cycle range, and the intermediate region between them depicts a three-dimensional curved surface characteristic resulting from creep–fatigue interaction. Stress holding time is short in the high cycle area on the right side and the cycle-dependent mechanism dominates crack growth behavior. Since the effect of vacancy diffusion in the grain boundaries is not significant, anticipating actual environments in which gas turbines are used, it is important to understand the four areas delineated by the dashed red line in the figure.

Fig. 15.

Mechanisms of creep–fatigue interaction regarding the three-dimensional curved surface representation of load frequency characteristic of IN738LC. (Colors are visible in the online version of the article; https://dx-doi-org.web.bisu.edu.cn/10.3233/SFC-150182.)

The upper right area (high temperature and low-cycle) is the environment corresponding to the high temperature section of the gas turbine to be used at middle to peak load. In this area, the influence of creep–fatigue interaction is significant. The transition region with a short life appears at a certain load frequency… It shows the cycle-dependent mechanism above a certain level of load frequency, and in contrast, shows the time-dependent mechanism below a certain level of load frequency. Therefore, further investigation is required to clarify which damage mechanism will dominate the vulnerable region of an actual component. As a result, while obtaining life characteristics, it is possible to discern the morphology of the actual damage whether that be fatigue cracks or creep voids and cracks.

The upper left area (high temperature and very low cycle) is the environment corresponding to the high temperature section of the gas turbine used at the base load. This region shows the time-dependent mechanism where the effect of stress concentration is relatively small and damage is widely distributed over a region of high, uniform stress.

The lower right area (low temperature and low cycle) is the environment corresponding to the low temperature portion of the gas turbine used at middle load to peak load. This region shows the cycle-dependent mechanism and in some cases deformation can be observed in the region of relatively high temperature and concentrated stress but basically the fatigue crack grows linearly in the direction perpendicular to the stress.

The lower left area (low temperature and very low cycle) is the environment corresponding to the low temperature section of the gas turbine used at the base load. In the low temperature range, vacancies were immovable and the time-dependent mechanism had difficulty manifesting. Instead, brittle fracture properties manifested due to the stress level of the stress concentrated region. There are cases where excessive stress leads to inelastic deformation within grains, however, it can also lead to fracture without visible damage or deformation in the case of low deformability material.

As described above, many materials exhibit a complex creep–fatigue damage behavior according to stress holding time or temperatures. Damage behavior analysis based on the EBSD method and three-dimensional curved surface characteristics of fracture life enabling the characterization of creep–fatigue interactions is highly significant for engineering.

5 Conclusion

Using polycrystalline Nickel-base superalloy, microscopic damage behavior was analyzed based on the EBSD method and the creep–fatigue properties expressed as a three-dimensional curved surface of fracture life derived from the theoretical analysis and experiment. This study has clarified the following:

(1) Depending on the load frequency or test temperature, when the cycle-dependent mechanism transforms to the time-dependent mechanism due to a decrease in load frequency, a shorter life transition region can be observed in some cases. However, it is concluded that based on damage behavior analysis using the EBSD method, intergranular damage due to vacancy diffusion which is a time-dependent mechanism and usually distributed over a broader region, became limited to notch areas due to the effect of load frequency, concentrating the intergranular damage in a relatively narrow range resulting in shorter life.

(2) The characteristics of the three-dimensional curved surface represent the basic properties of the material on creep fatigue interaction. The damage behavior analysis based on the EBSD method and three-dimensional curved surface characteristics of fracture life facilitate the characterization of creep–fatigue interaction. Damage morphology (creep crack and void or fatigue crack) at stress concentrated region can be effectively predicted depending on blade shape or operating conditions of actual gas turbines.

References

[1] Harada

and Yokokawa

, Materia Japan42 (2003), 621–625.

[2] Kobayashi

Kagiya

and Banno

, Thermal and Nuclear Power Generation Technology Association56(591) (2005), 1205–1209.

[3] Okada

Hisamatsu

and Fujioka

, in: Proc. Int. Gas Turbine Cong. 2007, Tokyo, 2007, TS-081.

[4] Viswanathan , Damage Mechanisms and Life Assessment of High-Temperature Components, ASM International, 1989, pp. 415–437.

[5] Kobayashi

Miyachi

Kagiya

and Chujo

, in: Proc. Int. Gas Turbine Cong. 2007, Tokyo, 2007, TS-093.

[6] Yokobori

A.T.

Jr. Yokobori

and Takasu

, in: Proc. of Am. Soc. Metals, 1986, pp. 1–9.

[7] Yokobori

A.T.

Jr. Kuriyama

Kaji

and Yokobori

, J. Jpn Soc. Strength and Fract. Mater.21(1) (1986), 25–32.

[8] Yokobori

A.T.

Jr. Kaji

and Kuriyama

, in: Advan. In Fract. Research, Proc. of ICF10 Ravi-chandar

et al., eds, Honolulu, Elsevier Science, 2001.

[9] Yokobori

Yokobori

A.T.

Jr. and Takasu

, Mechanical Behaviors of Materials – IV Carlsson

and Ohlson

N.G.

, eds, Pergamon Press, 1984, pp. 967–981.

10.

[10] Yokobori

A.T.

Jr. Akaishi

Sugiura

and Kobayashi

, Strength, Fracture and Complexity2 (2004), 35–45.

11.

[11] Kobayashi

Miyabe

Kagiya

Nagumo

Sugiura

Matsuzaki

and Yokobori

A.T.

Jr. , in: Proc. ASME PVP, Czech Republic, 26–30 July 2009, 2009, PVP2009-77289.

12.

[12] Kobayashi

Yokobori

A.T.

Jr. Sugiura

and Fuji

, Mater. Trans.51(12) (2010), 2201–2207.

13.

[13] Yokobori

A.T.

Jr. Sugiura

Takeuchi

Ozeki

Ihida

Kobayashi

and Hosono

, Strength, Fracture and Complexity8 (2013), 25–44.

14.

[14] Noguchi Catastrophe

, Science Pub. (1977) (in Japanese).

15.

[15] Yokobori

A.T.

Jr. Yokobori

Kosumi

and Takasu

, in: Proc. of Am. Soc. for Metal Goel

V.S.

, ed., 1986, pp. 1–9.

16.

[16] Mino

Fukuoka

and Yoshizawa

, J. of the Inst. of Metals, Japan64 (2000), 50–55.

17.

[17] Kamaya

Wilkinson

A.J.

and Titchmarsh

J.M.

, Acta Materialia54 (2006), 539–548.

18.

[18] Brewer

L.N.

Othon

M.A.

Young

L.M.

and Angeliu

T.M.

, Microscopy and Microanalysis12 (2006), 85–91.

19.

[19] Kamaya

, Trans. of the Jpn Soc. of Mech. Eng. Series A74(739) (2008), 315–322.

20.

[20] Yokomaku

and Yoda

, Inspection Technology12(1) (2007), 36–41.

21.

[21] Kobayashi

Miyabe

Kagiya

Nagumo

Sugiura

Matsuzaki

and Yokobori

A.T.

Jr. , J. of the Jpn Soc. for Strength and Fract. of Mater.45(1) (2011), 1–7.

22.

[22] Kobayashi

Miyabe

Kagiya

Nagumo

Sugiura

Matsuzaki

and Yokobori

A.T.

Jr. , Strength, Fracture and Complexity7(2) (2011), 157–167.

23.

[23] Kobayashi

Miyabe

Kagiya

Sugiura

and Yokobori

A.T.

, Metal. and Mater. Trans. A44(7) (2013), 3123–3135.

24.

[24] Kobayashi

Miyabe

Kagiya

Sugiura

and Yokobori

A.T.

, Mater. Sci. and Tech.30(1) (2014), 24–31.

25.

[25] Kobayashi

Miyabe

Kagiya

Nagumo

Sugiura

Matsuzaki

and Yokobori

A.T.

Jr. , Mater. at High Temp.29(4) (2012), 301–307.

26.

[26] Kobayashi

Miyabe

Achiwa

Omi

Sugiura

and Yokobori

A.T.

, in: Proc. of the ECCC2014 3rd Inter. ECCC Conf., USB, 2014, #17.

27.

[27] Yokobori

A.T.

Jr. Sugiura

Yoshino

Tabuchi

and Hasegawa

, in: Proc. of the ASME PVP07 Creep8, 2007, CD-ROM.

28.

[28] Yokobori

A.T.

Jr. Sugiura

Kimoto

Yoshino

and Matsuzaki

, in: Proc. of the ASME PVP09, 2009, PVP2009-PVP77989.

29.

[29] Yokobori

A.T.

Jr. Yokobori

Kuriyama

and Kako

, Adv. in Frac. Res, Proc. of ICF6 Valluri

S.R.

Taplin

D.M.R.

Rama Rao

and Knott

J.F.

, eds, Vol. 3, Pergamon Press, 1985, p. 2181.