Final published version
Research output: Contribution to Journal/Magazine › Journal article › peer-review
Research output: Contribution to Journal/Magazine › Journal article › peer-review
}
TY - JOUR
T1 - Machine learning for fatigue lifetime predictions in 3D-printed polylactic acid biomaterials based on interpretable extreme gradient boosting model
AU - Nasiri, Hamid
AU - Dadashi, Ali
AU - Azadi, Mohammad
PY - 2024/6/30
Y1 - 2024/6/30
N2 - Modeling of the fatigue lifetimes in 3D-printed biomaterials of Polylactic acid (PLA) is presented in this article based on machine learning (ML) techniques of interpretable extreme gradient boosting (XGBoost) and Shapley additive explanations. For this objective, standard testing samples were additive-manufactured from PLA under different 3D printing parameters. Then, the fatigue experiments were performed on specimens under various stress levels. Based on these data, three ML methods were utilized for modeling the PLA fatigue lifetimes, including XGBoost, random forest, and support vector regression, besides a common nonlinear regression analysis. The obtained results indicated that XGBoost had superior modeling results, compared to other ML techniques and the regression analysis. The coefficient of determination was 97.66 % with a scatter-band of ±1.3, which was a narrow scatter in fatigue modeling.
AB - Modeling of the fatigue lifetimes in 3D-printed biomaterials of Polylactic acid (PLA) is presented in this article based on machine learning (ML) techniques of interpretable extreme gradient boosting (XGBoost) and Shapley additive explanations. For this objective, standard testing samples were additive-manufactured from PLA under different 3D printing parameters. Then, the fatigue experiments were performed on specimens under various stress levels. Based on these data, three ML methods were utilized for modeling the PLA fatigue lifetimes, including XGBoost, random forest, and support vector regression, besides a common nonlinear regression analysis. The obtained results indicated that XGBoost had superior modeling results, compared to other ML techniques and the regression analysis. The coefficient of determination was 97.66 % with a scatter-band of ±1.3, which was a narrow scatter in fatigue modeling.
KW - 3D printing
KW - Biomaterials
KW - Extreme gradient boosting
KW - Fatigue lifetime
KW - Machine learning
U2 - 10.1016/j.mtcomm.2024.109054
DO - 10.1016/j.mtcomm.2024.109054
M3 - Journal article
AN - SCOPUS:85192179686
VL - 39
JO - MATERIALS TODAY COMMUNICATIONS
JF - MATERIALS TODAY COMMUNICATIONS
SN - 2352-4928
M1 - 109054
ER -