一种低气味多层复合胶带VOC释放预测方法及系统
By combining graph neural networks and physical information neural networks, the problems of accuracy and generalization ability in VOC release prediction of multilayer composite tapes were solved, achieving high-precision VOC release prediction and improving the interpretability and physical consistency of the model.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- 福建友谊胶粘带集团有限公司
- Filing Date
- 2026-04-29
- Publication Date
- 2026-07-17
AI Technical Summary
In existing technologies, it is difficult to accurately predict VOC release from multi-layer composite tapes due to low accuracy, high cost, and difficulty in covering multiple working conditions. Furthermore, traditional machine learning methods have insufficient generalization ability under small sample or new material conditions.
A method combining graph neural networks and physical information neural networks is adopted. By fusing molecular graph structure representation with macroscopic material properties, a neural network model constrained by diffusion equations is established. Combined with bidirectional recurrent neural networks and attention mechanisms, multi-source feature fusion prediction is performed.
It achieves high-precision and generalizable prediction of VOC release behavior of multilayer composite tapes, improves the interpretability and physical consistency of the model, and can accurately identify key time nodes and dynamic evolution patterns.
Smart Images
Figure CN122117137B_ABST