1. Assessment of the effect of the process-induced porosity defects on the fatigue properties of wire arc additive manufactured Al–Si–Mg alloy
- Author
-
Teng Zhan, Ke Xu, Zhipeng Fan, Hanlin Xiang, Congchang Xu, Tianjiao Mei, Yuanyuan Wei, Wentao Chen, and Luoxing Li
- Subjects
Porosity defects ,Wire arc additive manufacturing ,Al–Si7–Mg0.6 alloy ,Machine learning model ,Fatigue assessment ,Mining engineering. Metallurgy ,TN1-997 - Abstract
Process-induced porosity significantly affects the fatigue resistance of additive-manufactured components under cyclic loads. In this study, the influence of process-induced porosity defect characteristics on the fatigue properties of wire arc additive manufacturing (WAAM) Al–Si–Mg parts was quantitatively analyzed using the Kitagawa-Takahashi diagram and machine learning methods. Two specimens with different porosity and distributions were fabricated using two welding wires with different surface states. The cross-section porosity of the two groups was 0.42% and 2.06%, respectively. According to the test results, the static strength of the two groups was equivalent, but the elongation of the porosity group specimen was reduced by 58.7%, and the fatigue strength (139.8 MPa) was lower than that of the control group (160.2 MPa). Combining high-cycle fatigue post-mortem inspection and scanning electron microscopy analysis, the geometric features of the critical defects were obtained. The fatigue performance was assessed by combining extreme value statistics and the Kitagawa-Takahashi diagram according to critical defects, which showed good consistency but was still not conservative as some failure data were within the safe-life range. Therefore, four parameters of applied stress and the projected area, location, and morphology of the critical defects were trained using an extreme gradient boosting model (XGBoost) and random forest (RF). The XGBoost model is 95.7 % more accurate than the RF model in predicting fatigue life. The importance of four parameters in limiting fatigue life is ranked in the above order.
- Published
- 2025
- Full Text
- View/download PDF