PSI - Issue 79

Christoph Bleicher et al. / Procedia Structural Integrity 79 (2026) 239–247

244

The statistical evaluation was conducted based on the maximum likelihood method according to Spindel and Hai bach (1979) and Stoerzel (2021) to derive the S-N curve parameters for a probability of survival of P S = 10 %, 50 % and 90 %. Furthermore, the scatter band T σ and the slope of the S-N curve k was determined. The slope after the knee point N k was assumed to be k* = 45, which corresponds to a decrease in fatigue strength of 5 % per decade according to Sonsino (2007). 4. Discussion of results 4.1. Cyclic material behavior under artificial sea water and ambient air The evaluation of the fatigue test data was done for alternating and tensile loading for “Batch 1”, “Batch 2” and “Batch 7” for both ambient air and sea water and are summarized for R σ = -1 and R σ = 0 in Fig. 4 and Fig. 5. The results revealed for R σ = 0 a nominal stress amplitude of σ a,n = 111 MPa at N lim under ambient air. Here, a knee point N k at 4∙10 6 cycles and a slope k of 10.4 was derived. Under sea water the derived nominal stress amplitude σ a,n reached a value of only 70.4 MPa while the slope k got much steeper to 7.2. Additionally, a slight increase in the scatter band T σ from 1:1.13 to 1:1.28 under sea water was determined. This shows the significance of the influence of the corrosion effect under sea water. A large influence coming from different batches could not be shown.

Fig. 4. Comparison of fatigue strength under ambient air and sea water for R σ = 0.

Contrary to these results, for alternating loading (R σ = -1), no knee point could be identified under ambient air. The corresponding slope k showed with 10.9 more or less the same value as for tensile loading. This was the same case for the scatter band. At N lim the nominal stress amplitude evaluated to 174 MPa, if it is assumed that there is no additional knee point beyond N = 1∙10 7 cycles. This results in a mean stress sensitivity M air of 0.40 for the material under ambient air at N lim . Under artificial sea water, the slope k was under alternating loading with 5.5 even steeper than under tensile loading. The scatter band T σ was unaffected and comparable to the tests under ambient air. Due to the low value for k, the nominal stress amplitude at N lim reached 98.1 MPa with a mean stress sensitivity under sea water M corr = 0.39. So, the mean stress sensitivity was not affected by the corrosive effect of the medium. During the investigation of Dies (1966) for an inspected AlBz10Fe alloy, a nominal stress amplitude in the range of σ a,n = 147 to 235 MPa at 5·10 7 cycles for R σ = -1 was calculated, what is in good agreement to the present findings although results are touching the lower end of the values in Dies (1966). Also, the fatigue strength of AlBz10Fe alloy in artificial seawater was investigated for alternating loading and a nominal stress amplitude of σ a,n between 113 and 145 MPa at 5·10 7 cycles was reported depending on the manufacturing conditions without detecting a knee point. The investigated material in Fig. 5 reached a nominal stress amplitude of σ a,n = 112 MPa for the same number of cycles.