3D Generative AI for Next-Gen Design Exploration: Overcoming the Three Major Challenges of CAE via Inverse Shape Generation
3D生成AIによる次世代設計探査：性能目標からの形状生成で，CAEの『3つの壁』に挑む
- Delivery
- Available on this site
- Format
- Price
- Non-members (tax incl.)：￥1,100 Members (tax incl.)：￥880
- Publication code
- 20264073
- Paper/Info type
- Symposium Text
No.13-25
- Pages
- 17-21(Total 5 p)
- Date of publication
- Jan 2026
- Publisher
- JSAE
- Language
- Japanese
- Event
- JSAE Symposium 2025
Detailed Information
| Author(J) | 1) 西口 浩司 |
|---|---|
| Author(E) | 1) Koji Nishiguchi |
| Affiliation(J) | 1) 名古屋大学 |
| Affiliation(E) | 1) Nagoya University |
| Abstract(J) | 構造最適設計における「動的非線形」「製造制約」「設計空間探査」の3つの壁を克服するため，性能目標値から3D形状を直接生成するAI技術を提案する．DeepSDFを用いた本手法は，衝突性能を満たす形状や製造可能な薄板構造を瞬時に生成し，従来の試行錯誤を超えた広範な設計探査を可能にすることを実証した． Translation |
| Abstract(E) | In order to overcome the three barriers of "dynamic nonlinearity," "manufacturing constraints," and "design space exploration" in structural optimization design, we propose an AI technology that directly generates 3D shapes from performance target values. This method, using DeepSDF, has been demonstrated to instantly generate shapes that satisfy crash performance and manufacturable thin plate structures, enabling extensive design exploration that goes beyond conventional trial and error. |
