FC-Transformer multiple features fusion for predicting porosity and ILSS in CFRP autoclave processing

Curing pressure plays a critical role in controlling void evolution and improving the manufacturing quality of carbon fiber-reinforced polymer (CFRP) composites during autoclave processing. To gain deeper insight into this pressure–void relationship, a theoretical void-growth equation was employed as a reference framework to clarify the influence of curing pressure on void evolution. Based on this finding, 21 groups of autoclave curing experiments were designed using a nonlinear pressure gradient from 0 to 0.6 MPa. Void morphology and spatial distribution under different curing pressures were characterized by X-ray computed tomography, while the effects of void characteristics on mechanical behavior were evaluated through short-beam shear tests and scanning electron microscopy (SEM) observations. In addition, an FC-Transformer neural network model was developed to predict void morphology evolution and the corresponding mechanical properties during autoclave curing. The results show that increasing curing pressure significantly reduces porosity and transforms void morphology from elongated, clustered features to a more uniform near-spherical distribution. As porosity decreases, the Interlaminar Shear Strength (ILSS) of the composite increases markedly. These findings provide both experimental and data-driven support for understanding pressure-controlled void evolution and optimizing autoclave curing processes for CFRP composites.