PSI - Issue 77

Georg Veile et al. / Procedia Structural Integrity 77 (2026) 348–356

351

4

Georg Veile et al./ Structural Integrity Procedia 00 (2026) 000–000

mm. Fatigue life assessment was conducted using a local approach in FEA, identical to [3]. All fatigue damage parameters (FDP(s)) used in this work are commonly used FDP, extended by shear- and strain gradients, see Eq. (1) and (2). The distance dx is not specified in any guideline. This work investigates the influence on the fatigue life prediction. ∗ = 1 � � (1) ∗ = 1 � � (2) Fig. 3 illustrates the stress gradient in x direction for demonstrative purpose, using the example of a weld geometry with idealized weld radii. The gradient is determined at notch with the highest load state, e.g. stress, strain or shear. Since austenitic stainless steels are known for differing significantly in stress [12,13], only strain- and shear-gradients are used in this work.

Fig. 3 Schematic evaluation of a representative gradient in FEA with idealised weld geometry.

One of these commonly FDP is the critical plane approach of Fatemi-Socie (FS), [14] resulting in FDP Rm and FDP FFS [15]: = � 1+ 1 ∗ � ∆ 2 � 1+ , � (3) = � 1+ 1 ∗ � ∆ 2 � 1+ 1+ 1 ∗ � , � (4) Alternatively, the well-known Smith-Watson-Topper (SWT) can be extended to FDP FGF : = � 1+ 1 ∗ � ∙ ∆ 2 + � 1+ 1 ∗ � ∆ 2 ∆ 2 (5) The aforementioned FDP are known to be less conservative, compared to commonly used FDP FS , FDP VM or FDP SWT . However, in [3,16,17] it was shown that especially FDP FFS can lead to non-conservative prediction. This study investigates the influence of gradient determination by means of dx on the fatigue life prediction. 3. Results & Discussion The FEA results were subject to a mesh convergence study with 1 % stress difference from node to node. Additionally, convergence of the gradients was checked and achieved. For more, reference is made to [3,8]. In Fig. 4 calculated and experimental fatigue life are opposed for different Δx.