A Study of Optical Surface Analysis Methods for 3D Printed Continuous Fiber Composites

Abstract

Continuous fiber reinforced additive manufacturing has been growing in research and in commercial application. In this work, additively manufactured carbon fiber thermoset composites are being produced utilizing a liquid deposition modeling process with photo-curable resin. The quality and mechanical properties of additively manufactured composites is sensitive to the void concentration and distribution. The void percentage of additively manufactured composites had shown correlation with the surface quality of individual layers. Research has found that surface quality of individual layers could be controlled by process parameter adjustments such as material flow rate, speed and layer thickness. Regular monitoring of these variables becomes necessary for consistent and feasible production. To achieve this, the authors implemented a study utilizing a camera system and image processing algorithm for surface evaluation with a goal of print parameter adjustment on the fly. The resulting technology has been proven to accurately analyze each layer. This study aims to develop a surface feature extraction method for the purpose of training and deployment of a machine learning algorithm which will execute dynamic adjustment of process parameters.