[T17] Composite/Hybrid Materials

[17D-T17-09] In-Situ/In-Operando Machine Learning Approach for Topology Optimization of Structural Components by Means of Thermal Imaging, Thermal Analysis, and CT Scan

Aniqa Lim¹, Puskar Pathak¹, Venkat Selvamanickam ¹, *Francisco Carlos Robles Hernandez¹ (1.University of Houston, USA)

キーワード：Laser Direct Energy Deposition , Thermal Imaging, Mechanical Properties, Non-traditional alloys, Machine Learning

Here we present a material-by-design approach to produce components with high performance architectures. We propose an in-situ/in-operando heat assisted transitions to achieve specific microstructures during the 3D printing process. The approach is synergistic including Machine Learning (ML), Topology Optimization (TO) based on thermal imaging, thermal analysis in operando and CT-scanning as post processing to evaluate the printed parts. The uniqueness of our work include in-operando identification of cold spots and other defects that deter the integrity of the printed components. This permits to use ML algorithms capable of predicting the presence of defects and the component's soundness in-operando. This ML approach is ideal for manufacturing to eliminate defective parts in-operando and before their completion.