PSI - Issue 77

Bastian Roidl et al. / Procedia Structural Integrity 77 (2026) 119–126 Bastian Roidl / Structural Integrity Procedia 00 (2025) 000 – 000 ∆ = , − ,−1 ,−1

124

6

(3)

Table 3 - Equations of mean stress effect models examined.

, = 1− , = 1−( ) 2 , = √1− , =√( + ) ∗ , = ( − + ) , = 1− 2 ∗

Mean stress model Goodman [16] [17]:

Equitation

Gerber [18] [17]:

Dietmann [19] [20]:

SWT (Smith-Watson-Topper) [12]:

Smith [17] [21]:

Half-Slope [12]:

Table 4 - Accuracy of the mean stress effect models using the fatigue index error formula. Average fatigue index error Δ FI [%]

Standard deviation ΔFI [%]

Mean stress effect method

Geometry A

Geometry G

Geometry H

Geometry A

Geometry G

Geometry H

Goodman

17% -10% -3% 20% 54% -5%

13% -11% -4% 22% 45% -5%

-1%

11%

12%

14%

Gerber

-22% -16%

4% 5% 7%

5% 5% 6%

6% 7% 4%

Dietmann

SWT Smith

7%

27% -17%

24%

24%

26%

Half-Slope

5%

5%

6%

For the standard specimen (geometry A), the average fatigue index error shows that Dietmann model works best, followed by the Half-Slope and Gerber models. These last two models are best for specimens with G geometry, but with less accuracy and a higher standard deviation. SWT and Goodman models have the lowest average fatigue index error for specimens of geometry H. The average error of the fatigue index for specimens of geometry H is low when using the SWT model compared to the Goodman model. So only using the SWT model for this type of specimen makes sense. This evaluation shows similarly to Fig. 3, i.e., that the choice of the optimum mean stress effect model depends on the specimen geometry for additively manufactured parts from AlSi10Mg. The more similar the specimens are compared to the standard fatigue specimens of type A, the better the Half-Slope and Dietmann models works. For hollow specimens, the optimum solution relies on the Goodman or SWT models. 3. Discussion This study demonstrates that both geometry and applied mean stress significantly influence the fatigue performance of AlSi10Mg specimens produced by Laser Powder Bed Fusion (L-PBF). Some aspects raised in the study require a more in-depth discussion: The geometry of additively manufactured specimens affects not only fatigue strength but also sensitivity to mean stress. A glance at Fig. 2 shows H-type specimens behave in a completely different way than the other two specimen