Coupled deformation responses during thermoset AFP recompaction: Process parameter analysis and defect mitigation viability

This study investigates transverse tow deformation in thermoset Automated Fibre Placement (AFP) through systematic process parameter variation, focusing on tow width change and transverse displacement under recompaction. This paper assessed whether compaction-induced width change could mitigate AFP gap defects in process. A full factorial experimental design evaluated the effects of compaction force (C), tooling temperature (T), layup speed (S), external heating (H), and lie time (L). Separate regression models and ANOVA were used to identify significant factors and interactions for each response. Distinct governing mechanisms were identified. Tow width change showed viscosity-dominated behaviour, with tooling temperature as the primary factor and compaction force significant through interaction effects. Strong multi-parameter coupling, particularly C × T × S, demonstrated that meaningful width expansion requires coordinated thermal, mechanical, and temporal control rather than single-factor optimisation. In contrast, transverse displacement was governed solely by compaction force, independent of thermal conditions, indicating a shear-driven mechanical mechanism. A focused follow-up experiment isolating compaction force at moderate temperature confirmed this asymmetry. Width change decreased even at maximum compaction, demonstrating the requirement for thermal activation, while transverse displacement remained fully active, confirming its purely mechanical nature. This coupling shows that compaction forces required for width change inevitably induce displacement, while thermal control cannot decouple the two. The study concludes that compaction-induced width change is ineffective for gap mitigation in thermoset AFP. Achievable width changes are far below gap tolerances and are accompanied by displacement that degrades placement accuracy, representing a fundamental material limitation rather than a parametric optimisation issue.