PSI - Issue 38

Michaela Zeißig et al. / Procedia Structural Integrity 38 (2022) 60–69 Zeißig, Jablonski / Structural Integrity Procedia 00 (2021) 000–000

66

7

parameters as e.g. shown by Hatami et al. (2020). According to e.g. Radaj and Vormwald (2007), their influence on the fatigue behaviour may be treated in the same way as mean stresses. The residual stresses are included in the Weibull distribution in the following way, see Macherauch and Kloos (1982) and as applied e.g. in Jablonski (2001),







Mises

Mises RS





WV  

RS     

RS

M  

WV

1

2

3

(6) Here, the residual stresses � �� are considered as part of the material resistance and multiplied with the residual stress sensitivity M . For M , an approximative value of 0.14 is used, based on the assumption that its value is equal to the mean stress sensitivity as defined in (7) and using the data from Tab. 1, see Hück et al. (1983).

0.00035

0.1

M

R







(7)

m

The residual stress data used for the following calculations is approximated via FE simulations. A layer-wise building process which is able to reproduce the different residual stress distributions depending on the building direction is used. However, the effect of different scanning strategies within the layers is neglected. The sequentially coupled calculations are performed with the Abaqus Welding Interface (AWI), Abaqus Welding Interface (2017), an extension to Abaqus (Dassault Systemes Simulia Corp., 2016). The applied residual stress data is, therefore, merely a rough approximation to test the fatigue approach and highlight general effects.

Figure 2. Fatigue survival probability distributions for 316L. Effect of different exponents, left, considering residual stresses, right.

The survival probability distributions incorporating the effect of residual stresses for different building directions are shown in Fig. 2, right. A Weibull exponent of 51 is used. As one can see, the highest fatigue limits are predicted for the horizontally built specimens after the removal from the building platform and under consideration of the higher static properties based on the values in Tab. 1. The effect of the removal may also be seen as a shift of the curve of the horizontally built specimens towards higher fatigue limits. Thus, a reduction of residual stresses takes place. For the vertically built specimens, the effect is negligible, thus the curve is not shown. It should be noted that the chosen removal process during simulation might affect the results as well. The lowest fatigue limit is predicted for the vertically built specimens. This still holds for identical local fatigue limit values, thus is attributed to the residual stress distribution. Based on the local fatigue limit values, a reduction of a about 50% of this value may be attributed to the influence of residual stresses, based on the underlying FE results.