Composition optimization design and high temperature mechanical properties of cast heat-resistant aluminum alloy via machine learning
Changmei Hao, Yudong Sui, Yanru Yuan, Pengfei Li, Haini Jin, Aojie Jiang
Abstract
• Machine learning optimize alloy composition design. • Predicts high tensile strength for optimal composition at 300 °C and 350 °C. • AdaBoost algorithm improves prediction accuracy. • R 2 of 0.94 for Ultimate Tensile Strength(UTS), with 7.75% deviation in validation. • The optimal UTS at 300 °C was 163.83 MPa, with the best EL of 9.19 %. Traditional trial-and-error methods for optimizing the composition of heat-resistant aluminum alloys often consume significant time and resources, making it difficult to achieve alloys with excellent mechanical properties. This study combines experimental and machine learning methods to predict the optimal alloy composition for maximum ultimate tensile strength(UTS) at 300 °C and 350 °C. The AdaBoost algorithm was chosen as the final model. Experimental results show that predictions of the machine learning model deviate by only 7.75 % from the actual results, with an R 2 of 0.94. Furthermore, the study found that Al 9 FeNi and Al 3 Ni play key roles in enhancing the high-temperature mechanical properties of cast heat-resistant aluminum alloys. This model accurately predicts the high-temperature mechanical performance of heat-resistant aluminum alloys, providing effective guidance for their composition design.