This work proposes a data-driven design framework for high-pressure die-cast (HPDC) aluminum alloys that integrates robust data refinement, machine learning (ML) modeling, explainability, and inverse design. A total of 1237 tensile-test records from T5-aged HPDC alloys were aggregated into a curated dataset of 382 unique composition–heat-treatment combinations. Four regression models—Ridge regression, Random Forest (RF), XGBoost (XGB), and a multilayer perceptron (MLP)—were trained to predict yield strength (YS), ultimate tensile strength (UTS), and elongation (EL). Tree-based ensemble models (XGB and RF) achieved the highest accuracy and stability, capturing nonlinear interactions inherent to industrial HPDC data. In particular, the XGB model exhibited the best predictive performance, achieving test R2 values of 0.819 for UTS and 0.936 for EL, with corresponding RMSE values of 15.23 MPa and 1.112%, respectively. Feature-importance and SHapley Additive exPlanations (SHAP) analyses identified Mn, Si, Mg, Zn, and T5 aging temperature as the most influential variables, consistent with metallurgical considerations such as microstructural stabilization and precipitation strengthening. Finally, RF-based inverse design suggested new composition–process candidates satisfying UTS > 300 MPa and EL > 8%, a region scarcely represented in the experimental dataset. These results illustrate how interpretable ML can expand the feasible design space of HPDC aluminum alloys and support composition–process optimization in industrial applications.
13 November 2019 in Metals and Materials International
Two types of nanoclusters, Cluster (1) and Cluster (2), formed at around room temperature and 100 °C, respectively, affect the age-hardening behavior in Al-Mg-Si alloys. Formation of Cluster (1) during natural aging (NA) is more accelerated in the high-Mg (9M10S) alloy than in the low-Mg (3M10S) alloy. Hardness at the early stage of two-step aging at 170 °C is not increased for the natural aging samples. On the other hand, hardness is directly increased for the pre-aged (PA) specimens. Furthermore, the formation of Cluster (1) during natural aging is suppressed by the formation of Cluster (2) during pre-aging at 100 °C. To understand the effects of heat treatment histories and Mg contents on the microstructure, transmission electron microscopy (TEM) was utilized. All the images were obtained at (001) plane, and peak aged samples with different heat treatments were used. Lower number density of precipitates is confirmed for the natural aging samples compared with the single-aged and pre-aged specimens. A higher number density of precipitates is confirmed for 9M10S in comparison to 3M10S. Hardness results correspond well to the TEM images.
As the environmental pollution issue has recently become significant, environmental regulations in Europe and the United States are being strengthened. Thus, there is a demand for the quality improvement of emission after-treatment systems to satisfy the strengthened environmental regulations. Reducing the amount of welding heat distortion by optimization of the welding order of each part could be a solution for quality improvement since the emission after-treatment system consists of many parts and each assembly is produced by welding individual ones. In this research, a method to derive a welding sequence that effectively minimizes welding deformation was proposed. A two-stage simulation was performed to obtain the optimal welding sequence. In the first stage, the welding sequence was derived by analyzing the number of welding groups in each assembly of a structure. The derived welding sequence was verified by performing a thermal elasto-plastic analysis and comparing it with the experimental results.
9 January 2020 in Metals and Materials International