Structural strength related technology working activity report (FRP WG/AI-ML technology application WG)
Events
| Date | Friday, March 2024, 3 1:13 - 00:14 | |
| Venue | Keio University Mita Campus South Building 2B41 https://www.keio.ac.jp/ja/maps/mita.html |
|
| Entry fee | Free | |
| plan | JSAE Structural Strength Committee | |
| Application deadline | 2024.2.28 (Wed) | |
| Where to apply | ||
| Inquiry | Technology Exchange Business Division tech@jsae.or.jp | |
| 13: 00 to 13: 05 | Opening remarks | Structural Strength Committee Chairperson Shigenobu Okazawa, Yamanashi University |
| 13: 05 to 13: 45 | Lecture 1 | [FRP working group activity report] "Development of collision energy absorption member using fiber reinforced composite material" -Study of simulation method for compressive fracture phenomenon of fiber-reinforced composite materials-" Mr. Atsushi Yokoyama, Kyoto Institute of Technology |
| For many years, structural FRPWG has shown that it is possible to create lightweight and highly energy-absorbing parts by using fiber-reinforced composite materials (FRP) for crash boxes, which are collision energy absorbing members. We are developing a numerical analysis model that is effective for expressing the crushing process during a collision in FRP collision energy absorbing members, and elucidating the fracture mechanism by experimentally verifying the compressive fracture behavior of FRP materials. . This year, we are working on an analysis method to simulate compressive fracture phenomena based on this knowledge. In this report, we will report on the results of the study of the compressive fracture phenomenon of FRP that has been clarified in this working, especially the characteristic fracture behavior due to the presence of reinforcing fibers, and introduce its numerical analysis model. | ||
| 13: 45 to 14: 25 | Lecture 2 | [AI-ML technology application working activity report] "Development of crash box evaluation surrogate construction method" Mr. Yoshitaka Wada, Kindai University |
| This working group will examine data generation and processing methods, learning methods, and prediction accuracy for constructing crash box surrogate models. Using these findings, we aim to develop surrogate model construction methods for physical problems. The absorbed energy of the crash box and the maximum reaction force generated were also predicted. We also evaluated the contribution of input parameters and the performance of machine learning algorithms. Prediction using machine learning is not a complete replacement for CAE, but is used for preliminary calculations and narrowing down parameter ranges. It is assumed that this method of use will be applied to reduce the load of CAE analysis. Through these results, we aim to share useful knowledge for social implementation. This report describes the working activities and results to date. | ||
| 14: 25 to 14: 30 | Closing remarks | Structural Strength Committee Chairperson Shigenobu Okazawa, Yamanashi University |