Strategic layer reworking using hybrid additive manufacturing for defect-free ceramic parts

Abstract

Hybrid manufacturing combines additive and subtractive processes to create parts of high precision and density. However, extrusion-based processes are susceptible to stochastic defects such as voids, which lower yield and worsen material properties, leading to premature failure of components. This research demonstrates deep learning informed selective layer reworking for a ceramic hybrid additive manufacturing platform. We evaluate each layer in-situ for under and over extrusions using a vision-based monitoring system and a YOLOv8 model trained on a custom dataset. Through closed-loop control, a decision to repair defective layers via subtractive operations, prior to reprinting, was made using conditional gcode programming based on the results of the YOLOv8 model. The YOLOv8 model detected voids with a precision of 91 %, and a mean average precision of 83.5 % across both defect classes. Through CT analysis, it was determined that reworking achieved a 68 % reduction in void content compared to uncorrected parts, showcasing the potential of hybrid manufacturing in the creation of defect-free parts.