US-based researchers have invented an Artificial Intelligence software for 3D printers that will allow for quality checks of printed parts without investing in characterization equipment. This software christened Peregrine is being developed at Oak Ridge National Laboratories and it supports advanced manufacturing digital thread. It will help in gathering and scrutinizing data throughout the production process from design, feedstock selection, print build, and material testing.
According to Vincent Paquit from ORNL’s Imaging, Signals, and Machine Learning Group, this collection of data will help in the creation of a digital clone for parts thus providing a valuable store of data for raw materials to printed products. From this data, it will be easier to have a prototype for future builds for various part geometries using multiple materials. The result will be better automation and product quality assurance.
A digital thread is an innovative way through which custom parts are developed using Computer Aided Design and printed using self -correcting 3D printers through an advanced communication network which is cost effective in terms of time, energy and raw materials when compared to conventional production. This invention needs process control methods that ensures printed parts are ready for installation in applications like automobiles, airplanes, and energy facilities.
The software is ideal for powder bed printers which allow for distribution of a fine layer of powder over a build plate. The material is then melted and fused using lase or electron beams. On the other hand, Binder jetting systems use a liquid binding agent to fuse powdered materials.
According to Luke Scime, Principal Investigator for Peregrine at ORNL, the major challenge with 3D printing is that you have to put into consideration everything that will occur at scales of tens of microns and happening at microseconds for days or even weeks depending on build time. Since a flaw can happen at any point, its usually a challenge to fully understand the process and to qualify a part.
This informed the development of this AI software at ORNL’s Manufacturing Demonstration Facility, the US Department of Energy user facility that collaborates with industry leaders to invent, test, and refine advanced manufacturing technologies.