PSI - Issue 23

Sascha Gerbe et al. / Procedia Structural Integrity 23 (2019) 511–516 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

513

3

damage state by radiographic micro-computed tomography system Nikon XT H 160 (CT). A detailed description of the CT system and the chosen parameters are given in Tenkamp et al. (2019). In these investigations voxel sizes of around 11.5 µm were realized. The contrast between material and fatigue crack was improved by using contrast agent according to Mrzljak et al. (2018). To observe the evolution of fatigue damaging, specimens were scanned at initial state (prior testing) and at specific damage states when changes in resonance frequency or AC potential were observed during the tests. The quantitative fatigue crack examination for the CT scans were performed by 3D-image analysis software (VG Studio Max) determining quoted crack values such as encasing projected area, volume and shape. All uniaxial cyclic loading tests presented in this work were performed stress-controlled in fully reversed cycles (stress ratio R = -1), with constant load amplitude (constant amplitude test; CAT) and at room temperature. Furthermore, the HCF and VHCF specimen surfaces were ground and polished up to 1 µm diamond paste. For subsequent fractographic investigations selected specimens fromVHCF experiments were ruptured under tensile load. Fractography was carried out by use of a Zeiss Auriga high-resolution field emission scanning electron microscope.

Fig. 2. Experimental setups of (a) uniaxial ultrasonic resonance fatigue testing machine from Boku Vienna ( f = 20 kHz, cf. Gerbe et al. (2018)) with additional forced air-cooling and (b) resonance fatigue testing machine Rumul Testronic ( f = 70  5 Hz), equipped with video microscope, extensometer and Matelect ACPD crack length measurement system (cf. Tenkamp et al. (2018-2)).

3. Results and Discussion

The uniaxial cyclic loading experiments in the VHCF regime show a pronounce scatter of fatigue life data for the AlSi8Cu3 alloy. This and the influence of different SDAS values λ 2 on the fatigue behavior is shown in the S-N Woehler diagram with the respective Basquin analysis in Fig. 3a (cf. Basquin (1910)). Accordingly, equation 1 gives the bearable stress amplitude σ a in dependency of the number of load changes (expressed by 2 N f ; two load changes per cycle), with b being the fatigue strength exponent and σ´ f the fatigue strength coefficient. Using a combination of the cycle-dependent stress amplitude σ a (2 N f ) with the threshold stress intensity factor ranges ∆ K I,th (SIF) from crack propagation tests, carried out and shown in Gerbe et al. (2019), Eqns. 2 and 3 can be used to derive Kitagawa Takahashi threshold diagrams (cf. Kitagawa and Takahashi (1976)) modified according to El Haddad et al. (1979). Here, ∆ σ th is the bearable stress range, depending on the threshold SIF range ∆ K I,th , the defect length a and the intrinsic defect length a 0 (both assumed as equivalent pore diameter) , which is calculated by use of ∆ K I,th and ∆ σ f as fatigue strength range. The corresponding threshold diagram for the bearing seat, modified by multiple threshold curves for different fatigue-cycle threshold values N max is shown in Fig. 3b. Furthermore, data-points from VHCF experiments are added. The values for the defect length a (provided by fractography) are summarized for selected specimens in Table 1. Instead of the fatigue strength range, the respective stress amplitude σ a,th is used as the half of ∆ σ th .

   

(1)

(2 ) 2 b a f f N N 

f



, I th K a a

th   

(2)

(  



)

0