Tribological performance optimization of AZ91/TiB 2 magnesium matrix composites using stacked ensemble learning and Bayesian methods

Tribological performance optimization of AZ91/TiB 2 magnesium matrix composites using stacked ensemble learning and Bayesian methods Sathish Kamaraj Department of Mechanical Engineering, Vels Institute of Science Technology and Advanced Studies, Chennai, Tamil Nadu, India Baskar Sanjeevi Department of Automobile Engineering, Vels Institute of Science, Technology & Advanced Studies, Chennai, Tamil Nadu, India Hariharasakthisudhan Ponnarengan Department of Mechanical Engineering, Dr Mahalingam College of Engineering and Technology, Pollachi, Tamil Nadu, India https://orcid.org/0000-0002-9420-3196 Logesh Kamaraj Department of Mechanical Engineering, Vel Tech Rangarajan Dr Sagunthala R&D Institute of Science and Technology, Chennai, Tamil Nadu, India Sathickbasha Katharbasha Department of Mechanical Engineering, B.S. Abdur Rahman Crescent Institute of Science and Technology, Chennai, Tamil Nadu, India https://orcid.org/0000-0001-8974-154X

This work aims to develop and optimize lightweight, wear-resistant AZ91 magnesium matrix composites reinforced with titanium diboride (TiB 2 ) for demanding tribological applications. A hybrid ultrasonic-assisted mechanical stir casting technique was used to fabricate composites with 0–4 wt.% TiB 2 , ensuring uniform reinforcement dispersion. The tribological behavior was systematically evaluated under dry sliding conditions using a Taguchi L18 orthogonal design with sliding speed, normal load, and TiB 2 content as variables. To address the nonlinear interdependencies among these factors and predict key tribological responses, wear rate (WR) and coefficient of friction (COF), a stacked ensemble learning (SEL) model was developed, integrating random forest, gradient boosting, and a polynomially expanded meta-regressor. The SEL model achieved high prediction accuracy (R² = 0.948 for COF, 0.908 for WR). Subsequently, Bayesian Optimization (BO) based on Gaussian Process Regression was employed for multi-objective process parameter optimization. The optimal combination—13.77 N normal load, 1.94 m/s sliding speed, and 3.98 wt.% TiB 2 , yielded experimentally validated WR and COF values of 10.235 mg/km and 0.326, with <5% deviation from predictions. SEM analysis revealed mild abrasive wear, delamination and localized grain refinement as key mechanisms. This study conclusively demonstrates the efficacy of coupling advanced machine learning with experimental validation for intelligent design of Mg-based composites. The proposed SEL–BO framework reduces experimental workload while ensuring high-performance tribological outcomes, supporting its application in aerospace and automotive systems where both weight and surface durability are critical.
09 16 2025 13506501251377565 10.1177/13506501251377565 2 10.1177/sage-journals-update-policy journals.sagepub.com true https://journals.sagepub.com/page/policies/text-and-data-mining-license 10.1177/13506501251377565 https://journals.sagepub.com/doi/10.1177/13506501251377565 https://sage.cnpereading.com/paragraph/article/?doi=10.1177%2F13506501251377565 https://journals.sagepub.com/doi/pdf/10.1177/13506501251377565 https://journals.sagepub.com/doi/pdf/10.1177/13506501251377565 https://journals.sagepub.com/doi/full-xml/10.1177/13506501251377565 10.1016/j.matchar.2015.04.015 10.1177/09544089231167758 10.46793/adeletters.2024.3.4.5 10.18485/aeletters.2023.8.3.3 10.3390/ma17215157 Tribology of the al-si alloy based MMCs and their application in automotive industryEngineered metal matrix composites: forming methods, material properties and industrial applications Vencl A. 127 Vencl A. Tribology of the al-si alloy based MMCs and their application in automotive industry. In: Engineered metal matrix composites: forming methods, material properties and industrial applications. 2013, pp.127–166. https://hdl.handle.net/21.15107/rcub_machinery_1807. 10.1016/0043-1648(95)06657-8 Tribol Ind Vencl A 27 28 2006 Tribological properties of Al-Si alloy A356 reinforced with Al2O₃ particles Vencl A, Rac A, Bobić I, et al. Tribological properties of Al-Si alloy A356 reinforced with Al2O₃ particles. Tribol Ind 2006; 28: 27–31. 10.1177/13506501211004790 10.1177/13506501221113161 10.1007/s13369-023-08216-7 10.1080/02670836.2022.2156673 10.1007/s40962-021-00747-9 10.1007/s40430-024-05109-0 10.1007/s13369-022-07032-9 10.1080/02670836.2021.2006908 10.1088/2051-672X/ac2176 10.1155/2022/2158516 10.1007/s00170-023-12502-x 10.1016/j.jma.2023.07.021 Adv Mater Process Technol Singh S 4058 8 2022 Optimization of adhesive wear behaviour of B₄C/AZ91D-Mg composites Singh S, Chauhan NR. Optimization of adhesive wear behaviour of B₄C/AZ91D-Mg composites. Adv Mater Process Technol 2022; 8: 4058–4072. 10.1177/02670836241237681 10.46793/tribomat.2024.004 10.1108/ILT-07-2021-0262 10.3390/lubricants11010024 10.46793/tribomat.2024.016 10.1016/j.triboint.2018.07.003 10.1016/j.triboint.2019.02.029 10.1016/j.mtcomm.2024.109722 10.1007/s11837-023-05845-z 10.1007/s40962-020-00552-w 10.1016/j.triboint.2024.110189 10.1115/1.4062518 10.1016/j.mtcomm.2023.107507 10.1007/s10853-024-10233-2 10.1016/S1003-6326(20)65482-6 10.1016/j.triboint.2024.109696 10.1088/2051-672X/ac7065 10.1016/j.jmrt.2021.09.069 Int J Interact Des Manuf Veera Ajay C 1 2025 Optimization and prediction of the tribological parameters of biocompatible AZ31/Al₂O₃/Si₃N₄ metal matrix composites using CCD–RSM, MOORA and FNN models Veera Ajay C, Manisekar K, Andrews A, et al. Optimization and prediction of the tribological parameters of biocompatible AZ31/Al₂O₃/Si₃N₄ metal matrix composites using CCD–RSM, MOORA and FNN models. Int J Interact Des Manuf 2025: 1–22. 10.1177/09544089241300003 Adv Mater Process Technol Kumar S 1136 11 02 2024 Tribological study of magnesium-based AZ91/TiC/rGO hybrid composite prepared by powder metallurgy route: an experimental and analytical approach Kumar S, Gupta RN. Tribological study of magnesium-based AZ91/TiC/rGO hybrid composite prepared by powder metallurgy route: an experimental and analytical approach. Adv Mater Process Technol 2024; 11(02): 1136–1155. 10.26599/FRICT.2025.9440982 10.1115/1.4068056 10.1177/13506501241300905 10.1016/j.triboint.2023.108783 10.1016/j.jma.2024.01.005 10.1038/s41598-024-59100-9 10.1016/j.jmst.2025.02.059 10.3390/ma17184523 10.1007/s40735-023-00745-w 10.1088/2053-1591/ad4e0a 10.3103/S1067821218020116 10.1007/s11661-016-3447-1 10.3390/pr13061697 10.1177/13506501241274546 10.1016/j.heliyon.2024.e33967