[home] [Personal Program] [Help]
tag
11:30
20 mins
Robotic Large Format Additive Manufacturing for On-Demand Tooling for Composites
Michael Cargill, Patrick Bailey, Aywan Das, Wout De Backer
Session: Session 5b: Advances in Manufacturing Automation
Session starts: Wednesday 15 April, 10:50
Presentation starts: 11:30
Room: Main


Michael Cargill (University of South Carolina)
Patrick Bailey (University of South Carolina)
Aywan Das (University of South Carolina)
Wout De Backer (University of South Carolina)


Abstract:
The increasing adoption of carbon fiber reinforced polymers (CFRPs) in aerospace manufacturing has amplified the demand for large, high-precision tooling capable of withstanding the temperatures and pressures required for composite manufacturing. Conventional metallic tools, often fabricated from alloys like Invar, offer excellent dimensional stability but incur prohibitive costs and extended lead times, constraining rapid prototyping and low-volume production. This work investigates the application of robotic large-format additive manufacturing (LFAM) for the on-demand fabrication of composite tooling using thermoplastic polymers. Employing a custom pellet-fed, single-screw polymer extruder mounted on a KUKA KR60 robotic arm, this study evaluates the feasibility, dimensional accuracy, and durability of printed polymer tooling under composite curing conditions. Experimental comparisons between printed and conventionally machined tools assess surface quality and composite part quality through measurements of surface roughness, fiber volume fraction, and void content. Key process parameters: including print speed, bead width, and infill orientation, are analyzed to determine their influence on mechanical and dimensional tool performance. The results aim to establish validated design and process guidelines for LFAM based composite tooling and demonstrate its potential to reduce tooling costs and lead times by an order of magnitude, thereby advancing automation and sustainability in composite manufacturing.