Fig. 1 | Scientific Reports

Fig. 1

From: Prediction and optimization of stretch flangeability of advanced high strength steels utilizing machine learning approaches

Fig. 1

Overview of the prediction and optimization framework for the stretch-flangeability of AHSS. (a) Data collection: gathering chemical compositions, microstructural characteristics and mechanical properties from 212 steel conditions. (b) Missing value imputation: employing MICE methods to address incomplete datasets, with missing value analysis and cross-validation for method selection. (c) Regression: utilizing an interpretable ML framework utilizing SVM, SR, or XGBoost algorithm, coupled with SHAP for feature importance interpretation. (d) MOO: implementing R-NSGA-III to simultaneously optimize HER, UTS, and TE while considering compositional and microstructural constraints.

Back to article page