Prediction of the yield strength of as-cast alloys using the random forest algorithm

• Published in: Materials today communications Vol. 38; p. 108520
• Main Authors: Zhang, Wei, Li, Peiyou, Wang, Lin, Fu, Xiaoling, Wan, Fangyi, Wang, Yongshan, Shu, Linsen, Yong, Long-quan
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.03.2024
• Subjects: As-cast metals; Prediction accuracy; Random forest; Yield strength; Prediction accuracy; Yield strength; Random forest; As-cast metals
• ISSN: 2352-4928; 2352-4928
• DOI: 10.1016/j.mtcomm.2024.108520

Yield strength is an important indicator of material mechanical properties, and its prediction and evaluation are crucial for engineering design and material selection. Predicting yield strength can help optimize design, improve the strength of structural materials, and serve as an indicator for material quality control to ensure product quality and performance. This article uses a random forest model to predict the yield strength of 540 as-cast alloys, and selects four evaluation indicators to analyze the model. The yield strength of the alloy was divided into three ranges: low, medium, and high, and the relationship between prediction accuracy and yield-strength range was obtained. The yield strength of four different alloy systems, Ti-Fe-Sn-Nb, Ti-Mo-Sn, Ti-Cu-Co-Zr and Fe-Ni-Co-Cu-Ti, was predicted and experimentally verified. It was found that the random forest algorithm can effectively distinguish the yield strength exhibited by the different alloy systems. By analyzing the importance of characteristic parameters, it was found that four parameters and some main elements play an important role in the accuracy of yield strength prediction. Accurately predicting yield strength can save the cost and time of selecting engineering materials, therefore, this work is of great significance in materials science and engineering applications. [Display omitted] •Average prediction accuracy of yield strength is 82.16%.•The R2 fitting degree is good, and the average R2 value predicted 15 times is 0.8394.•Random forest algorithm can distinguish yield strength of the different alloy systems.•Four parameters and some main elements play an important role in yield strength prediction.