Optimization of stabilized annealing of Al-Mg alloys utilizing machine learning algorithms

Abstract

The corrosion properties of the alloy are influenced by the physical parameters involved in the preparation process. Experiments to explore the preparation process of Al-Mg alloys are very complex and time-consuming, and the amount of data is very limited. In this work, the analysis of the corrosion mechanism of Al-Mg alloy identified the alloy magnesium content, deformation, annealing temperature and time as important factors affecting the corrosion resistance of the alloy. Based on the existing experimental data, a machine learning framework that effectively promotes smart manufacturing is proposed. The results show that the machine learning framework constructed based on the existing experimental results can reliably predict the NAMLT values of the alloy. As more data is acquired, the method is expected to be used to adjust production processes for efficient and intelligent machining.