PSI - Issue 79

Ibrahim T. Teke et al. / Procedia Structural Integrity 79 (2026) 17–25

21

o Fatigue life and failure mechanisms were compared with dissimilar Al – Mg joints incorporating Sn coated steel interlayers as studied by Sun et al. (2016). o These studies highlighted the role of interfacial reaction layers and bending-induced stress concentration in SLJ fatigue performance. • Tensile-Shear (TS) Joints: o Validation was performed using results from Teke et al. (2025b), who examined the fatigue and tensile performance of DIN1623 steel spot welds under optimized welding parameters. o Their study also provided microstructural data and failure mode maps which were used to verify the model's prediction accuracy. • Dissimilar Steel-to-Steel Joints: o Material asymmetry (CR210 steel – CR340 steel) and joint stack configuration from Asati et al. (2022) were integrated into the model to assess residual stress distribution and crack path sensitivity. • Parameter Matching and SWT Correlation: o All simulations adopted the same SWT-based fatigue life model calibrated with identical material constants and thermal pre-stress conditions. The residual stresses induced by welding were introduced via a thermal expansion simulation step based on the method described by Teke et al. (2025a). • Fatigue Prediction Scope: o Fatigue predictions were generated for TS, MTS (Ertas and Akbulut, 2021), and SLJ configurations using consistent mesh density, boundary conditions, and load ratios. o Experimental life values were compared against simulated SWT damage integrals, indicating strong agreement. This comprehensive cross- validation underscores the model’s effectiveness in representing real -world joint behavior across multiple configurations and material combinations, demonstrating that a single finite element model — combined with thermal-strain-based prestress estimation — can accurately predict fatigue life without recalibration for each geometry or material set. 3. Results and Discussion This section evaluates the fatigue life prediction capability of the proposed numerical model across multiple specimen configurations, with and without thermal pre-stress. Comparisons with experimental results demonstrate that integrating welding-induced residual stresses significantly improves prediction accuracy, especially for dissimilar joints. As shown in Table 1, the pre-stress-integrated Morrow model yielded life predictions much closer to the experimental results than those without prestress. For instance, under 2.7 kN loading, experimental life was 13,762 cycles, while the pre-stress prediction was 10,209 cycles — an acceptable error. In contrast, the non-prestressed model overestimated life by ~13% (15,608 cycles). This trend was consistent across the load range, confirming that residual stresses critically influence fatigue crack initiation in TS joints. Table 1. TS specimen life prediction capability with pre-stress. 3.1. Tensile-Shear (TS) Joints

Load (kN)

Without Prestress Prediction

Experimental

Prediction

Study

2.2 2.4 2.7

46575 25856 13762

29163 18564 10209

45258 28243 15608

Teke & Ertas (2025) Teke & Ertas (2025) Teke & Ertas (2025) Teke & Ertas (2025) Teke & Ertas (2025)

3

6226 4264

6204 2385

9111 3416

3.7