| About this Abstract |
| Meeting |
MS&T23: Materials Science & Technology
|
| Symposium
|
Additive Manufacturing Modeling, Simulation, and Machine Learning: Microstructure, Mechanics, and Process
|
| Presentation Title |
A-2: Deep Learning Assisted Material Structure Property Linkage of 3D Printed AlSi10Mg Alloy |
| Author(s) |
Ibrahim Khalilullah, Constantin Virgil Solomon |
| On-Site Speaker (Planned) |
Ibrahim Khalilullah |
| Abstract Scope |
In this study, Laser Powder Bed Fusion-manufactured AlSi10Mg components were subjected to various post-processing techniques, including stress relief, hot isostatic pressing (HIP), quenching, solution heat treatment (SHT), and T6 heat treatment (T6 HT). The mechanical properties of these samples, such as tensile strength and hardness, along with microstructural images acquired using a scanning electron microscope and a light microscope, were used to generate a large microstructure-property benchmark dataset for additively manufactured (AM) AlSi10Mg parts with different post-processing. A deep artificial neural network (ANN) trained to classify images was modified and retrained with the newly developed dataset to predict material properties from the microstructure. Here we present a data-driven property determination technique for 3D-printed materials as an alternative to the experimental procedure. |