A rapid geometry reconstruction method for automated dry fibre placement (ADFP) preforms using real-time manufacturing data

This paper introduces a rapid geometry reconstruction technique designed to capture variability in real time during Automated Dry Fibre Placement (ADFP). The approach supports enhanced quality control, facilitates downstream process and structural modelling, and enables the creation of a digital product passport for in-service monitoring and proactive maintenance. By integrating laser scanning and machine motion data collected during fibre deposition, the method generates detailed computer-aided design (CAD) models with solid representations of individual tapes. Demonstration panels are produced using a laboratory-scale ADFP machine and subsequently infused with liquid resin. The accuracy of the reconstructed geometries was validated through optical microscopy, revealing tape width measurement errors within 2.5 %. Positional and angular deviations of the deposited tapes were constrained to ±1 mm and ±2°, respectively. These deviations were predominantly concentrated within the initial ∼180 mm of each course, attributed to insufficient tape tension at the onset of deposition. In-plane shear testing reveals a direct correlation between the digitally identified gaps and overlaps and the localised strain variations measured via digital image correlation. Although these defects have minimal impact on the in-plane shear modulus, they are shown to significantly influence premature failure and reduce the ultimate shear strength of the composite. Notably, gaps of up to 5 mm in width were observed, corresponding to a decrease in ultimate strength of up to 13 %.