An AI-based approach for flow front monitoring and prediction in liquid composite molding processes based on dielectric and visual data elaboration

Liquid composite molding processes involve the impregnation and saturation of a dry preform by a liquid reactive resin, driven by pressure gradients. The correct advance of the resin flow front during the infusion is crucial to achieve defect-free composite products. In the present work, monitoring architecture based on real-time acquisition of data provided by dielectric and visual sensors has been developed. A machine learning approach, based on the You Only Look Once (YOLO) algorithm, has been integrated with the visual monitoring system to detect and dynamically track the resin flow front, deriving relevant process parameters in real-time. The data obtained highlights the effectiveness of the combined monitoring strategy and proposes a sensing tool for the further study of impregnation and saturation phenomena in liquid composite processing.