Efficient alloy design of Sr-modified A356 alloys driven by computational thermodynamics and machine learning

Abstract

• A reliable CALPHAD thermodynamic database of al-Si-Mg-Sr system was established. • The quantitative composition-microstructure-properties relation was constructed. • The optimal Sr additional amount in A356 alloys was designed as 0.005 wt.%. • The strengthening/toughening mechanisms of Sr-modified A356 alloys were analyzed. A356 alloys are widely used in industries due to their excellent comprehensive performance. Sr is usually added in A356 alloys to improve their mechanical properties. There have been various experimental reports on the optimal additional amount of Sr in A356 alloys, but their results are inevitably inconsistent. In this paper, a combination of computational thermodynamic and machine learning approaches was employed to determine the optimal Sr content in A356 alloys with the best mechanical properties. First, a self-consistent thermodynamic database of quaternary Al-Si-Mg-Sr system was established by means of the Calculation of PHAse Diagram technique supported by key experiments. Second, the fractions for solidified phase/structures of A356- x Sr alloys predicted by Scheil simulation, together with the measured mechanical properties were set as the input dataset in the machine learning model to train the relation of “composition-microstructure-properties”. The optimal addition of Sr in A356 alloy was designed as 0.005 wt.% and validated by key experiments. Furthermore, such a combinatorial approach can help to understand the strengthening/toughening mechanisms of Sr-modified A356 alloys. It is also anticipated that the present approach may provide a feasible means for efficient and accurate design of various casting alloys and understanding the alloy strengthening/toughening mechanisms.