Real-time control for large scale additive manufacturing using thermal images

Abstract

Large scale additive manufacturing has process mechanisms similar to Fused Filament Fabrication (FFF) and can print parts with large sizes, such as car bodies. This capability provides large scale additive manufacturing great application potentials in a variety of industries, including aerospace, automotive manufacturing, and construction. To make large scale additive manufacturing a viable manufacturing solution, both production efficiency and product quality need to be considered. Specifically, the printing process is subject to constraints on print surface temperature to guarantee good product quality. In this paper, we propose a novel method for controlling layer time during the printing process by using thermal images. Specifically, several thin wall test components are printed by the Thermwood Large Scale Additive Manufacturing (LSAM™) machine, and the print surface temperature is monitored by a FLIR™ thermal camera. A regression model based on real-time thermal imaging data is built to predict the surface temperature. Then a layer time control method is proposed based on the temperature prediction model. By comparing the proposed method to the fixed layer time policy, the experiment results suggest that the control method by using real-time data can siginificantly reduce the layer time, and subsequently the total printing time, while satisfying the quality requirement.