About this Abstract |
Meeting |
2022 TMS Annual Meeting & Exhibition
|
Symposium
|
Aluminum Alloys, Processing and Characterization
|
Presentation Title |
Development of a Machine Learning Model to Predict Constitutive Behaviour of Macrosegregated A356 Alloy |
Author(s) |
Jun Ou, Shaul Avraham, Daan Maijer, Steve Cockcroft |
On-Site Speaker (Planned) |
Jun Ou |
Abstract Scope |
Macrosegregation during solidification results in spatial variation of microstructure and mechanical properties in a cast part. This study is aimed developing numerical methods to predict the local constitutive behaviour of A356 aluminum alloy following macrosegregation. Constitutive data was generated from a series of castings with compositional and microstructural variations in the as-cast and heat-treated conditions. Two numerical methods were developed to represent this data: 1) an equation fitting method based on a modified Hollomon equation, and 2) a machine learning technique based on a deep neural network (DNN). The parameters in both methods are trained using the experimentally measured constitutive data. The features (input) of the data includes the silicon content (composition), secondary dendrite arm space (SDAS) (microstructure) and heat-treated condition, and the labels (output) include the yield stress and flow stress as a function of plastic strain. Both methods are capable of making accurate predictions on the constitutive behaviour. |
Proceedings Inclusion? |
Planned: Light Metals |
Keywords |
Aluminum, Modeling and Simulation, Mechanical Properties |