Metalurgija, Vol. 65 No. 1, 2026.
Izvorni znanstveni članak
https://doi.org/10.64486/m.65.1.4
Prediction of mechanical properties of zinc alloy based on machine learning algorithm
Kairan Yang
; School of Software, Xinjiang University, Urumqi, Xinjiang 830046, China
Gulisitan Yisimayili
; School of Economics and Management, Xinjiang University, Urumqi, Xinjiang 830046, China
Junyu Yue
; School of Computing Science and Technology, University of Chinese Academy of Sciences, Beijing 101408, China
*
* Dopisni autor.
Sažetak
As a medical implant material, zinc alloys need to have high strength and suffi-cient hardness to support bone regeneration. Therefore, it is important to clearly define the design criteria for zinc alloys that meet the mechanical property re-quirements of degradable medical implants. In this work, mechanical property data of Zn-Mg-Mn alloys were obtained through experimental research and liter-ature collection. A performance-oriented machine learning (ML) model was used to predict the compressive yield strength and hardness of Zn-Mg-Mn alloys with different element types, contents and alloy preparation processes, and then the in-fluence of element types and contents on the microstructure and macroscopic me-chanical properties of the material was explored. Based on the existing dataset, the model compared six different ML algorithms and identified the optimal prediction algorithm. To further verify the accuracy of the model’s predictions, data outside the dataset were randomly selected for comparative analysis with the model re-sults. The results showed that the prediction errors of the algorithm designed in this work for compressive yield strength and hardness were less than 2 % and 2.4 %, respectively.
Ključne riječi
machine learning; zinc alloy; mechanical property prediction
Hrčak ID:
336511
URI
Datum izdavanja:
1.1.2026.
Posjeta: 72 *