Structural reliability of thermal barrier coatings(TBCs) used in advanced land-based gas turbine power generation systems, has recently been increased year by year. This energy system has a specific advantage that it is flexible to absorb fluctuation in power supply from current renewable energy systems. However, the TBCs are frequently subjected to thermo-mechanical fatigue damages, because thermal stress is significant there. Here, understandings on not only the steady state thermo-mechanical fatigue (SS-TMF) failures but those on the non-stationary TMF (NS-TMF) failure have been essential. In the latter a special consideration should be taken to non-steady or transient thermal response of TBCs. In this work the failure behavior of NS-TMF were studied, in comparison with those of NS-TMF. The experimental works clearly demonstrated that the TBCs were subjected to additional damages under the NS-TMF. In order to insight this phenomenon, a simple mechanical model by which the transient thermal is passively taken into account has been proposed. The numerical calculation showed that the additional thermal stress is developed in the NS-TMF test, resulting in the additional damages in NS-TMF. Here the thermal stress showed peak under a given frequency, and the peak frequency was dominated by the physical and mechanical properties of TBS system and the Biot-number. Some suggestions were also discussed to avoid this type of new subject.

Thermal Barrier Coating (TBC) protects structural components from high temperature environment of gas turbines. The TBC consists of ceramic top coat and metallic bond coat, deteriorate by the high temperature operating condition. Calcium-Magnesium-Aluminosilicate (CMAS) infiltration into the top coat is also one of degradation process for TBCs. Damage mechanism associated with the CMAS are significant in the novel turbine system, which operates at higher temperatures than the melting point of the CMAS. In this study, thermal deformation measurements between 300°C and 900°C on the CMAS infiltrated air plasma sprayed (APS) TBC topcoat specimens were conducted. Inflection point on a thermal strain curve for the CMAS infiltrated top coat specimen was observed around 700°C. The temperature corresponds to the glass transition temperature of CMAS grass. About 10 % decrease in linear expansion coefficient of the CMAS infiltrated top coat was also observed. These results indicated that the thermo-mechanical properties of CMAS infiltrated top coat were influenced by that of the CMAS, and such properties may play an important role to develop a thermal stress and failure mechanism of TBCs.

For evaluating Low temperature Melt Infiltration (LMI) process, adapted to amorphous SiC fiber with high toughness and low thermal stability, high-manufacturable and high-strength orthogonal 3-D woven amorphous SiC fiber / SiC / YSi₂-Si Matrix Composites was developed. Thermophysical properties of candidate molten Si alloys were measured by the electrostatic levitation furnace, Si-Y alloy was selected because of the low melting temperature and superior infiltration ability compared to reference material, Si-Ti alloy. Dense orthogonal 3-D woven amorphous SiC fiber / SiC / YSi₂-Si Matrix Composites of which porosity was less than 3vol.% was successfully fabricated by weaving an amorphous SiC fiber coated with CVD-BN and constructing SiC matrix via Chemical Vapor Infiltration (CVI) process and LMI. Process temperature of LMI was 50°C lower than general Melt Infiltration (MI) process with Si. It was demonstrated that the orthogonal 3-D woven amorphous SiC fiber / SiC / YSi₂-Si Matrix Composites has equal tensile strength until 1100°C and excellent fatigue properties at 1100°C over 150MPa against to crystalline SiC fiber/SiC-Si Matrix Composites.

High temperature components such as boiler tubes and jet engine turbine blades undergo multiaxial creep damage due to their complex shapes and combination of mechanical stress and inner pressure for a steam tube. Although multiaxial creep investigation is required for the safety of high temperature components designing, there are only a few commercial testing machines that can conduct multiaxial loading at high temperatures. The authors have developed a miniature cruciform specimen having a 5 mm square size plane stress condition gauge section and a biaxial tensile loading testing machine. This study discusses multiaxial creep rupture time and creep properties of type 304 stainless steels at 923 K using the miniature cruciform specimen. A specimen deformation during the creep tests could be observed clearly and the displacement of the specimen also could be measured using a developed non-contact measuring method. Creep rupture time under multiaxial stress condition was shorter than that of uniaxial stress loading.

The actual structures receive not only proportional loading (tension and compression loading) but also non-proportional loading in which principal directions of stress and strain are changed depending on time. In the researches for multiaxial non-proportional cyclic loading tests, it has been reported that number of crack initiation cycles is decreased and number of cracks is also significantly increased under multiaxial non-proportional loading compared with those in proportional loading. It also has been reported that an additional hardening occurs due to the non-proportional loading which leads to additional interaction between slip systems resulted from increase in active slip systems under non-proportional loading. In the present study, cyclic overloading tests were carried out on SUS316 and STPT410 steels, in which the cyclic overloads were applied after pre-crack fatigue tests under proportional and non-proportional loading. The effects of pre-crack loading under non-proportional loading on the deformation and fracture characteristics under cyclic overloading are discussed.

The fatigue lives of austenitic stainless steel SUS316 were evaluated under thermomechanical fatigue (TMF) loading condition with and without tension hold. When correlating the fatigue lives with the mechanical strain range, the fatigue lives under the in-phase condition were shorter than those under isothermal low cycle fatigue (LCF) condition at the maximum temperature of TMF tests, on the other hand, those under the out-of-phase condition was longer. The introduction of the hold dwell at the maximum tension strain extended the fatigue lives under the in-phase condition because of the decrease in the stress range due to cyclic softening augmented by the tension hold. The prediction method for TMF lives from the LCF data was proposed based on the small crack propagation law taking into consideration the temperature dependence of the tensile strength and the crack propagation path, with some assumptions; such as, the m values in Paris’s law are comparable in the intergranular and the transgranular crack propagation modes, the fatigue life is proportional to the crack propagation life, and the fatigue crack propagation morphology of the tested material is similar to that of the 316FR steel, etc. The TMF lives of austenitic stainless steel SUS316 could be predicted with higher accuracy by the proposed method.

The low cycle fatigue (LCF) properties of the TiAl alloy Ti-43Al-5V-4Nb (at.%) with triplex microstructure have been investigated. The LCF tests were conducted at room temperature, 600 °C and 760 °C. The result shows that test temperature has a small influence on the LCF life evaluated by the total strain range. On the contrast, the LCF life evaluated by stress range had a good correlation with ultimate tensile strength at each temperature. Though the elongation at 760 °C was relatively ductile, the material parameter of Coffin’s low fitted to experimental data means brittle fracture. In TiAl alloys, the brittle-to-ductile transition temperature which is around 700 °C is also affected by strain rate. Therefore the difference of strain rate between tensile tests and LCF tests may result in the brittle fracture. The fracture mode was brittle irrespectively of test temperature. The fatigue crack at room temperature was initiated in the globular γ phase. On the other hand, at the temperature of 600 °C, straight fatigue crack was initiated at β phase. This suggested that the oxidation of β phase decreased the LCF life in high temperature.

Creep damage preferentially progresses at stress concentration portion, where stress gradient exists under multiaxial stress state, in high temperature components. Therefore it is necessary to clarify creep damage evolution process and to establish a rupture time prediction method under multiaxial stress state. In this study, creep tests using plane specimens and four kinds of round notch bar specimens with different notch radius (notch tip radius 0.1mm(R0.1), 0.5mm(R0.5), 2.0mm(R2.0) and 4.0mm(R4.0)) on a 2.25Cr-1Mo steel were carried out. Rupture times of notch specimens were much longer than those of plane specimens under the same nominal stress indicating notch strengthening effect. Measured void number density on the notch root section indicated higher value at around notch surface for R0.1 and R0.5 specimens, while it was higher at around specimen center for R2.0 and R4.0. These distribution patterns corresponded to those of the maximum stress and the triaxiality factor of the notch specimens. Although, creep rupture times of notch specimens were not able to be predicted by using the stress or strain at the representative point on the notch root section, they were well predicted based on the area average damage concept, especially by considering selected damage area for R0.1 and R0.5 specimens.

The small punch (SP) creep testing technique was applied to five heats of Gr.91 steel, which had been actually used for boiler pipings in the different USC power plants for long periods of time. As a result, it was found that the difference in creep strength between the heats, that is, the heat-to-heat variation, was well reflected in the SP creep property. In addition, the conventional SP test was carried out at room temperature and 650°C to obtain the load/stress conversion coefficient (F/σ) which is necessary for the conversion of SP creep rupture data to the uniaxial ones. The F/σ for each heat was estimated by the displacement to maximum load (u_m) using the correlation of F/σ = u_m h₀/0.320. The experimental results revealed that the F/σ showed no significant change and it was 2.4-2.5 irrespective of heats. It was also found that the SP creep rupture data were relatively well converted to the uniaxial one with those coefficients.

A European 10Cr cast steel, GX12CrMoWVNbN10-1-1, which has been recently used in the steam turbine components of USC power plant, was subjected to thermal aging at 610, 630, and 650°C and creep test. The small punch (SP) test was carried out at the temperature range from -196 to 80°C using those aged and creep damaged steels to examine an applicability of the SP testing technique to the assessment of aging degradation. The experimental results revealed that the maximum load and the displacement to fracture changed depending on the aging and creep conditions, respectively. The temperature corresponding to the SP fracture energy of 1 J, T_1J, was employed for investigating the change in ductile-brittle transition property for convenience. The T_1J, which changed to the higher temperature side due to aging and creep at the beginning, decreased gradually with increasing Larson-Miller parameter (LMP), resulting in the recovery of embrittlement. Laves phase was clearly observed at the initiation site of quasi-cleavage cracking and the inside of dimple on the fracture surfaces. The T_1J was well correlated with the number density of Laves phase, and it increased with increasing density. These results indicated that the precipitation and coarsening of Laves phase was closely associated with the change in SP ductile-brittle transition behavior. Additionally, it was also found that the tensile strength predicted based on the SP test result was relatively in good agreement with the measured one, and it had a tendency to decrease with increasing LMP.

The change in hydrogen thermal desorption characteristic of reduced activation ferritic/martensitic (RAFM) steel for fusion reactor blanket due to creep was investigated to examine the applicability of thermal desorption analysis (TDA) to remaining-life assessment. The hydrogen charging into creep damaged specimens was conducted by means of cathodic electrolysis. The hydrogen-charged samples were subjected to the TDA for the measurement of hydrogen evolution curve. The experimental results revealed that the evolution curve was changed only by creep depending on the test conditions, and the amount of hydrogen desorbed during the analysis (C_H) increased with increasing the creep life consumption. A vacancy-type defect and dislocation, which were introduced by the electron beam irradiation and the cold rolling respectively, had a large influence on the desorption characteristic, but they were by no means causes of the change in desorption characteristic with creep. It was also found that the C_H was successfully arranged with the parameter derived based on the creep void growth’s law, irrespective of test conditions. The results obtained in this study indicated that the TDA could be a strong tool for the creep damage detection and remaining-life assessment of RAFM steel.

In order to avoid troubles caused by damage of steam piping of thermal power plants, it is strongly required to improve accuracy to predict Type IV creep damage of Mod.9Cr-1Mo steel. Although Type IV creep damage occurs at the heat affected zone, it is recognized that the creep strengths between the welded joint and the base metal have correlation. So, evaluation of remaining life of the base metal is very important. In this report, the method for estimating the change in hardness due to creep was applied to unused and serviced Mod.9Cr-1Mo steel, and from the results, a method for estimating the creep life equation for the serviced material was proposed. A thermal aging estimation equation that expresses the change in hardness due to thermal aging, and ΔG and ΔG_Serviced parameters to express the acceleration of softening of unused and serviced materials were formulated. Based on the ΔG_Serviced parameter and the creep life equation of unused materials, the creep life equation for the serviced materials was constructed, too. It was shown that this equation can accurately estimate the rupture time in short and long-term regions.

In power plants, Type IV creep damage and the softening structure initiate in FGHAZ (Fine Grain Heat Affected Zone) of high temperature steam piping made of high chromium steel. The evaluation methods of remaining life of the damaged piping are needed. Before the creep, the grain boundaries of FGHAZ are serrated and unclear. With increasing of creep damage, they become simpler lines and clearer. The grain boundaries are the fractal and their complexity are expressed with the fractal dimension. We studied on the method that measured the creep damage in FGHAZ with the fractal dimension. The creep tests were conducted with the notched specimens made of the simulated FGHAZ of Mod.9Cr-1Mo steel. Before the fracture they were observed with SEM and their fractal dimension were measured with the box-counting methods. The fractal dimension decreased with creep damage. The fractal dimension decreased rapidly when the creep strain was small and the stress triaxiality was large. This decrease in fractal dimension was saturated at some value. We defined the fractal dimension decreasing rate by dividing the decreased value of the fractal dimension by creep time. It correlated with the creep rate. Vickers hardness decreased and the average grain size increased with decreasing fractal dimension. Those observations suggested that the re-crystallized grain in FGHAZ grew during the creep. The saturation of the decrease in fractal dimension might represent dynamic re-crystallization that stopped the growth of re-crystallized grain. We made the assumption; the large stress triaxiality made the strain small and made the recovery rapid. It also made growth of re-crystallized grain rapid. Then the fractal dimension decreased rapidly. The sufficiently large re-crystallized structure was the softening structure. Therefore, the fractal dimension is probably useful for the prediction or the inspection of the softened structure.

Type IV creep damage is the phenomenon that many voids initiate, coalesce, become large cracks and break the weldment. It occurs in FGHAZ (Fine Grain Heat Affected Zone) of high temperature steam piping made of high chromium steel in power plants. The method to estimate the remaining life of damaged weldments is needed. The stress and the stress triaxiality affect TypeIV creep damage, but the quantitative influence of them are not clear. We conducted the creep tests on the three conditions of the different maximum principal stress and of the different stress triaxiality factors with simulated FGHAZ of the Mod.9Cr-1Mo steel and got the damaged samples. The test temperature was 650C. In previous research, we had proposed the simple method that estimates the initiation rate and the growth rate of voids by the distribution of radius of void. We applied this method to the samples and quantified the influence of the maximum principal stress and TF (stress Triaxiality Factor) on the initiation rate and on the growth rate of voids. The radii of voids were measured by optical microscope. The results were as follows; under the same TF of 2.9, the maximum principal stress of 79MPa gave the larger initiation rate and the larger growth rate than 73MPa.Under the same stress of 79MPa, TF of 2.9 gave the larger growth rate than TF of 3.5, but the initiation rates on both conditions were almost the same.

Type IV creep damage in fine grained heat affected zone of Mod.9Cr-1Mo steel has been concerned and extensively studied for steam piping of thermal power plants. However, the evaluation of Type IV damage by EBSD measurement is not enough. Serviced materials that have been used for a long time in an actual equipment are considered to effective materials for investigating microstructural changes in a long-term region. In this study, SEM observation and EBSD measurement were performed on the material of the Mod. 9Cr-1Mo steel FGHAZ used on the actual equipment and interrupted test specimen. As a result, as the creep progressed, strain concentration due to dislocation accumulation was observed in the portion near the voids, and recovery due to dislocation reduction was confirmed in the non-void portion, and the mechanism of Type IV damage caused by local strain concentration was confirmed. ΔKAM, which is obtained by subtracting KAM_local, which is the KAM value near the void portion, and KAM₀, which is the KAM value of the nonvoid portion, showed a monotonic increase with respect to the creep damage rate t/tr. The possibility that creep damage can be evaluated by this index was shown. In addition to examining the estimation formula expressing the change of the KAM value, it was found that the macroscopic and local void numbers could be evaluated from the KAM value.

In order to establish creep damage assessment method for Gr.91 steel longitudinal welds by the ultrasonic phased array (UPA) method, creep rupture tests with large test specimens were carried out under 650°C, 50-60 MPa. All creep specimens ruptured at heat affected zone (HAZ). The UPA method was performed using interrupted creep specimens and ruptured specimens. The UPA indication was recognized near HAZ of the root part at creep life ratio of about 0.7. After that, the UPA indication was locally propagated thorough the wall thickness direction with increasing creep life ratio. Next, the relationship between UPA indication and microstructure was investigated. Many creep voids were observed at UPA indication position. Also, creep voids were initiated at recovery area of microstructure near HAZ of the root part. On the other hands, there were hardly creep voids at non-indication positions. It is thought that UPA method can identify the area of dense creep voids. This result suggests that the application of UPA method for Gr.91 steel longitudinal welds is effective as creep damage assessment like creep voids. Finally, in order to clarify the reason why recovery of microstructure progressed at the root part, EBSD (Electron Back Scatter Diffraction) method and ECCI (Electron Channeling Contrast Imaging) method were carried out. As a result, it was found that recovery and recrystallization of microstructure is easy to occur preferentially at the root part by stress concentrations.

Creep rupture behavior of HR6W weldment using full thickness specimen cut from the circumferentially welded pipe was investigated for evaluating its applicability to advanced USC (A-USC) power plant. Creep tests were conducted at 750°C for durations up to 8,600 hours, and damage morphology of weldment during creep was characterized. The applicability of several nondestructive detection methods to the creep damage evaluation was discussed. It was found that full thickness specimen was broken at the base metal and main crack was inclined approximately at 45 degrees to the axial direction of the specimen. Based on stress at the ruptured location, times to creep rupture of full thickness specimen were comparable with those of the smooth bar specimen of base metal. Among several nondestructive detection methods, a small crack in base metal on the outer surface was first observed at life fraction of 35% by replication. PT can detect the crack up to 0.5mm in length in about half of the life. The crack whose length is longer than 3mm can be detected by UT in latter half of the life.

We studied the correlation between the evolutions of non-linear acoustic and magnetic properties and local misorientation on strain-induced martensitic transformation (SIMT) behavior in SUS 304 metastable austenitic stainless steel during tensile test. Two nonlinear acoustic characterizations; resonant frequency shift and high harmonic components with electromagnetic acoustic resonance (EMAR), parameters in magnetic hysteresis loops were measured. Furthermore, the evolution of microstructure was observed with electron backscatter diffraction (EBSD) was examined. EMAR is a contactless resonant method with an electromagnetic acoustic transducer (EMAT). This method enables not only to measure exact ultrasonic attenuation of measured sample but also to eliminate nonlinear acoustic effect between the sample and transducer. We used two type of axial SH wave EMATs, which transmit and receive axially polarized shear wave along a cylindrical surface of a circular rod, for ferromagnetic and non-ferromagnetic materials. The changes in non-linear acoustic and magnetic properties have good correlations with microstructural observations. They successfully capture SIMT behavior in a metastable austenitic stainless steel.