A method, apparatus, device, medium, and product for generating an enhanced structure
By training a forward graph neural network model in conjunction with a pre-defined generative network, an enhanced network graph structure for flexible support components is generated, solving the problems of long design time and lack of customization in existing designs, and realizing efficient and accurate flexible support product design.
CN122154356AInactive Publication Date: 2026-06-05SUZHOU POLLY NEW MATERIAL TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SUZHOU POLLY NEW MATERIAL TECH CO LTD
- Filing Date
- 2026-05-09
- Publication Date
- 2026-06-05
- Estimated Expiration
- Not applicable · inactive patent
Smart Images
Figure CN122154356A_ABST
Abstract
The application discloses a kind of enhanced structure generation method, device, equipment, medium and product.The enhanced structure generation method comprises: obtaining the geometric information of target surface and expected pressure difference distribution;Wherein, target surface is the support surface of target flexible support component, is discretized into multiple grid observation points, each grid observation point corresponds to an expected pressure difference value in expected pressure difference distribution and local geometric feature in geometric information;The expected pressure difference value and local geometric feature of each grid observation point are input into the preset generation network trained in advance, and the local candidate subgraph generated by the preset generation network for each grid observation point is obtained;Local candidate subgraph is fused and processed, and the enhanced network graph structure inside target flexible support component is obtained.The technical scheme of the embodiment of the application can efficiently customize the enhanced network graph structure inside flexible support component according to the expected pressure difference distribution.
Need to check novelty before this filing date? Find Prior Art