PSI - Issue 43

Vít Horník et al. / Procedia Structural Integrity 43 (2023) 136–141 Author name / Structural Integrity Procedia 00 (2022) 000 – 000

138

3

The fatigue tests were performed by resonant testing machine Amsler 10 HFP with 100 kN force range under a fully reverse load control regime. The frequency of sinusoidal cyclic loading was in the range of 107 ÷ 113 Hz. The specimens w ere heated to the test temperatures of 800, 900, and 950 °C by an electric resistance furnace on air controlled by two thermocouples with an accuracy of ± 1 °C in the middle of the gauge length of the specimen . Laboratory temperature was held at 23 ± 1 °C with a relative humidity of 50 ± 5 %. Cylindrical test specimens with a geometry shown in Fig. 2 were used for the purposes of this study. The gauge length of all specimens was mechanically ground. The fractographical analysis of fractured specimens was performed by a TESCAN Lyra3 XMU scanning electron microscope (SEM). The tests were terminated by reaching a number of cycles of 1 × 10 7 or by a failure of a specimen. The fatigue endurance limit was determined based on three unbroken specimens cycled on the same level of stress amplitude. 3. Results and discussion Experimentally determined dependence of the number of cycles to failure N f on the stress amplitude σ a at the temperatures of 800, 900, and 950 °C is shown in Fig. 3. Moreover, the obtained results are compared with the results of IN 713LC superalloy and MAR-M 247 superalloy, previously published by Šmíd et al. (2016). The polycrystalline IN 713LC superalloy was in “as cast” condition ( without HIP or heat treatment). Therefore, numerous defects with a size range from 300 to 800 μm were observed in the structure. The grain size was around 1.0 mm, and the volume fraction of γ’ precipitates was approx. 53 %. On the other hand, the polycrystalline MAR-M 247 superalloy investigated by Šmíd et al. (201 6) was processed by the HIP treatment reducing the casting defects size to approximately 400 μm , and the structure contained about 60 % volume fraction of γ’ phase. In both cases (IN 713LC and MAR-M 247), the fatigue limit was determined by stress amplitude when the number of cycles reached 2 × 10 7 . In Fig. 3, the IN 738LC and IN 713LC superalloys achieved a similar high-cycle fatigue behavior at 800 °C , while the MAR M-247 reached a higher fatigue performance, see Fig. 3 (a). Experimentally determined points exhibit a relatively high scatter of fatigue lifetime stemming from inevitable effect of casting defects. The fatigue limit of IN 738LC and IN 713LC superalloys is 200 MPa at 800 °C while for MAR -M 247 is 220 MPa. A significant difference in the fatigue performance at the temperature of 900 °C was observed in Fig. 3 (b). The lower number of cycles to failure was characteristic for IN 738LC in comparison with IN 713LC at all tested stress amplitudes. Even larger difference, up to 2 orders of magnitude in extreme cases (such as at σ a = 240 MPa), can be seen while comparing to MAR-M 247 alloy. . The fatigue limit of IN 738LC decreased to 170 MPa at 900 °C , and it is notably lower than the fatigue limit of IN 713LC (180 MPa) and MAR-M 247 (200 MPa) superalloys. High-cycle S-N curves of IN 738LC, IN 713LC, and MAR-M 247 obtained at 950 °C are shown in Fig. 3 (c). Despite the IN 738LC reached a nearly similar fatigue performance as IN 713LC alloy at higher stress amplitudes, the difference in performance significantly increased with decreasing stress amplitude. The IN 738LC superalloy exhibited the lowest fatigue limit of all tested alloys at 950 °C . The fracture surfaces of the IN 738LC superalloy loaded at two different stress amplitudes at 800 °C are shown in Fig. 4 (a) and (b). The crack initiation was affected by casting defects in all performed tests. The origin of the fatigue crack initiation was frequently from the surface at high-stress amplitudes. With decreasing stress amplitude, the crack initiation predominantly shifted into the specimen interior. The internal fatigue crack initiation (highlighted by the arrow in Fig. 4 (b)) was commonly followed by a fish-eye formation. The facets, a prominent sign of the crystallographical (stage I regime) crack propagation, were observed in the vicinity of crack initiation sites of all specimens cycled at 800 °C , including the fractured specimens with surface fatigue crack initiation, i.e. with active corrosion mechanisms. Outside of the fish eye area, the fatigue crack propagation was entirely non-crystallographic (stage II regime), perpendicular to the loading direction accompanied by numerous striations. The significant scatter of obtained fatigue life data can be directly attributed to the presence of casting defects. The specimen lifetime corresponds to the size of the casting defect responsible for the fatigue crack initiation ( Šmíd et al. (20 20)). Generally, the specimens with a smaller pore size at the crack initiation site reached a higher fatigue lifetime. Fig. 5 shows the fractured surfaces of specimens after fatigue tests at 900 °C in higher (Fig. 5 (a)) and lower (Fig. 5 (b)) stress amplitudes. The fatigue crack initiation position dependence on stress amplitude was similar