Using Optical Imaging and Image Processing to Verify a Layer in a Laser Powder Bed Fusion Process

Abstract

Additive manufacturing (AM) allows for the creation of complex geometries that cannot be created with traditional manufacturing methods. AM is widely used in regulated industries such as medical and aerospace which require objective evidence of good manufacturing processes (GMP) for auditing purposes. Within AM, important powder layer characteristics must be met to ensure final part quality. Currently, no machine can provide objective evidence of a proper characterization of crucial powder layer properties with in-process monitoring equipment. Such properties are currently verified by unquantifiable means and can be classified within two categories of failure. This project investigates and analyzes possible sensor technologies that can provide in-process data to objectively quantify the characterization condition. Implementing in-process monitoring technologies will provide objective, quantitative evidence, prevent failed builds due to improper powder layer setups, and reduce the time it takes to set up an AM machine for a build. While the final solution for this project incorporates the use of both a 2D laser line sensor and an AM in-machine camera, this thesis will specifically focus on the in-machine camera. More specifically, this thesis will discuss camera repeatability tests that were conducted, the images taken during these tests, and the resulting pixel intensity values from these images. Analysis of the intensity values demonstrated that the in-machine camera could distinguish between different powder layer thickness values and that intensity values could be used as a quantitative metric to indicate whether certain powder layer characteristics are within specification.