PSI - Issue 76

Vladimír Mára et al. / Procedia Structural Integrity 76 (2026) 123–130

128

domain (10 6 -10 7 cycle) is lower in comparison to noHT, the noHT configuration suffers from extensive scatter, see Table 3.

Table 3. Regression parameters of K&V function Specimen batch Heat treatment α

β

R 2

s

s s

B

C

logN

noHT

1 427 663

-0.938

9 087

24 882

0.42

0.781

25.16

1

T240

1 997 184

-0,880

32 904

106 283

0.96

0.116

6.69

T200

4 638 483

-0.912

70 152

177 574

0.91

0.283

10.28

2

T300

412 023

-0.698

58 361

300 996

0.97

0.134

6.32

T6

1 987

-0.216

28 373

1 893 742

0.98

0.141

4.55

3

T6 mod

516

-0.113

7 804

7 796 628

0.94

0.239

6.98

3.3. Fracture behavior and role of defects

Fractographic analysis was performed to evaluate the impact of microstructural features on fatigue crack growth and propagation, to identify the fatigue crack initiation and to quantify the overall combined role of defects and heat treatments. For each broken specimen, type and size of the killer defect were identified (see Fig. 5a). They were used for calculation of fatigue limit prediction using modified Murakami’s equation for Al alloys. From the fracture surfaces, maximum fatigue crack length before failure a m was measured to calculate maximum stress intensity factor (SIF) of the crack tip before the unstable propagation of fatigue crack K Imax (see Fig. 5b). Equations for modified Murakami and maximum SIF can be found in the work (Xu et al., 2020). Subsequently, the length of early crack growth a ECG , areas of crack growth region A CGR and residual fracture region A RFR were determined to evaluate the influence of defects and heat treatment on the fracture behavior (see Fig. 5b-c). In the A CGR , SEM micrographs of striations were captured and used for estimating fatigue crack growth rate for each heat treatment (see Fig. 5b).

Fig. 5. (a) SEM micrograph showing lack of fusion type of the killer defect on the fracture surface of T240 specimen; (b) Example of determination of fracture regions for noHT specimen with N f = 3.37×10 4 ; (c) Example of of maximum fatigue crack depth measurement for T240 specimen with N f = 7.46×10 4 ; (d) SEM micrograph illustrating determination of fatigue crack growth rate using striation count. Crack tunneling can be observed on most of the fracture surfaces of noHT and annealed specimens. The effect of tunneling leads to the undesirable faster growth of crack front across the load-bearing cross-section, while towards the edges, the crack propagates significantly slower (see Fig. 5c). For the rest of the annealed specimens and all T6 specimens, the crack growth front is uniform and propagates mostly evenly (see in Fig. 5b). Near the larger defects, parabolic striations can be often observed (see orange arrows in Fig. 5b). This is valid for nearly all testing specimens regardless of heat treatment type. Fractographic analysis revealed, that defects causing fatigue crack initiation in testing specimens are of two types – lack of fusions (LOFs) and pores. At T6/T6mod specimens the cracks initiated from small circular gas pores above all, while larger pores and LOF defects can be found in annealed specimens (see Fig.6a). The largest killer defects can be found in the T200 specimens (batch No. 2), while the smallest ones are in the T6 samples (batch No. 3) (see Fig.6b). The effect of the treatment is also evident in how fatigue cracks propagate. The a ECG is slowly increasing its length with gradual transformation of Si network (see Fig. 6b). This is confirmed by