A physical model and data-driven hybrid prediction method towards quality assurance for composite components

Since composite components have been used in many fields with high-performance requirements, their quality is always of great concern. During production, improving temperature uniformity of the mold, which has close contact with the composites, is critical for reducing component deformation. However, the bottleneck is realizing the rapid and accurate prediction for the mold temperature distribution. Therefore, this paper designs a new hybrid modeling method for mold temperature prediction, which is driven by both physical and data models. The proposed method is applied in a case study of quality assurance for a plate component. Its advantages are also validated.