PSI - Issue 47

Jan Patrick Sippel et al. / Procedia Structural Integrity 47 (2023) 608–616 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

614

7

While still observing a fatigue fracture surface inside the fisheye (a to b), we also observe a fatigue like fracture surface in the beach mark outside the fisheye (c). The brighter areas are shown to correlate with brittle intergranular force fracture surfaces (e.g., d and e). However, most strikingly we observe an alternating pattern correlating to the bright and dark areas found on the fracture surface for the position from e to i. So, while observing brittle intergranular force fracture surface in the bright area marked with e, we observe a fatigue fracture surface with increasing distance in the darker area (marked with f). Under the assumption of a constant amplitude, the increasing crack length would result in an increased maximum SIF. Therefore, the alternation of brittle intergranular force fracture to fatigue fracture, observed from e to f and from g to h, is highly uncommon. Image i shows the final fracture surface with a brittle fracture surface implying final unstable crack propagation resulting in final failure. To gain further insight regarding the cause for the change in fracture surface morphology during the late crack propagation stages, the interaction between defect size, natural frequency, and resulting amplitude during the fatigue test is analyzed. For this purpose, the data acquired during testing regarding the power, the frequency and the amplitude of the US-fatigue testing setup are examined and compared for both materials in the following. 3.4. Comparison of power, natural frequency and amplitude during fatigue In order to identify the cause of the different fracture surface characteristics outside the fisheye described above, it is purposeful to compare the data recorded during the experiment. The power as well as the frequency and the amplitude of the US-fatigue testing setup are recorded by the logger in the generator with a data acquisition rate of 1000 Hz during the whole test. However, since the power pattern as well as the values of frequency and maximum amplitude show constant values within all pulses during the entire test period until the final fracture, Figure 7 a is an excerpt of the data only including several pulses before the final failure. The excerpt displays the power in red, the excitation frequency in green and the amplitude, correlating with the stress amplitude, in blue. It should be mentioned that for the sake of clarity, only the data during the pulse phases of the pulse pause mode testing is shown. The red rectangle marks the detailed diagram displayed in Figure 7 b and corresponds to the final failure of the specimen. While the pattern of the power is significantly altered during final failure, no direct correlation regarding the crack propagation can be drawn and therefore only the natural frequency as well as the amplitude are included in Figure 7 b. Herein, a sharp drop of the natural frequency (green) and amplitude (blue) are observed within one pulse. The maximum SIF at the border of the fisheye can be calculated using Equation 1 with C = 0.5 for internal defects and the red area displayed in Figure 7 c.

Fig. 7. (a) Excerpt of the logger data of an AISI 52100 specimen before final failure; (b) Close up of the natural frequency and amplitude data during failure; (c) Defect size of the fisheye responsible for the drop in natural frequency and amplitude. The resulting maximum SIF value of K max,FiE = 19.4 MPam 1/2 corresponds very well with the fracture toughness of K C,AISI 52100 = 18 – 19.5 MPam 1/2 . The observed fracture surface outside the fisheye can therefore be considered the result of unstable crack propagation and the fast increase in crack length results in the abrupt drop of the natural frequency as well as of the amplitude. Although this sharp drop of the amplitude is obvious, no influence on the unstable crack propagation is observed. Therefore, we assume that the increase in crack length due to the unstable crack propagation can overcompensate the amplitude decrease, resulting in the SIF still exceeding the fracture toughness for AISI 52100.