| About this Abstract |
| Meeting |
MS&T23: Materials Science & Technology
|
| Symposium
|
Engineering Ceramics: Microstructure-Property-Performance Relations and Applications
|
| Presentation Title |
Strategy to Estimate Mechanical Properties of Engineering Ceramics by Using AI-determined Grain Information and Simulation |
| Author(s) |
Manabu Fukushima, Kiyoshi Hirao, Yuki Nakashima, Kimiya Aoki, Shingo Ozaki, Wataru Nakao |
| On-Site Speaker (Planned) |
Manabu Fukushima |
| Abstract Scope |
National project in Japan, entitled “Development of a Technology Base and Applied Technologies for the Manufacturing Processes of Next-Generation Advanced Ceramics” has been launched in 2022, by New Energy and Industrial Technology Development Organization (NEDO). We joined the project team dealing with fracture prediction technology. Mechanical properties of ceramic materials are substantially influenced by the surface condition and internal microstructures including porosities, grain boundary and their distribution. As a numerical conversion of their microstructures requires a great deal of manual procedures, we have developed the artificial intelligence (AI) to recognize those. In addition, the numerical data including grain/pore position coordinates and pore position coordinates, pore size and its distribution, which is utilized to estimate the mechanical propeties by using FEM based simulation. In this presentation, we will explain our strategy to achieve the automation by using deep neural network segmentation and making connection with a simulation based on microstructural data. |