The NextManufacturing Center 3D metal-printing roadmap includes several points of emphasis, including microstructure control and porosity modeling; machine learning and part qualification; topology optimization; and powder spreading and melt-pool geometry.
Beuth highlights the center’s focus areas in broad terms:
- Monitoring and feedback-based control systems
- Controlling build microstructure and material properties by location within a part
- Expanding the variable sets used to manufacture acceptable parts and eliminate or control internal porosity
- Expanding the acceptable range of powder properties, while developing new alloys to take advantage of the rapid solidification rates present during 3D metal printing.
Closed-Loop Process Control
“Closed-loop process control is coming,” Beuth says. “The ability to adapt for variability in the process is a big concern particularly in the aerospace industry. We have projects underway focused on how to manage the large amount of data emanating from the machines, and we have shown some positive results. For example, one of our Ph.D. students, Luke Scime, has developed a machine-learning algorithm that, for each image of the powder bed, analyzes the image to look for spreading failures. His algorithm can identify several types of spreading failures; the next step is to engineer the in-process adjustments needed to overcome those failures.”
“Heat is generated as the part builds,” Rollett says, “which causes lower cooling rates at the top of the part than at the bottom, and a corresponding difference in material grain size and properties.”
Another Ph.D. student, Brian Fisher, collaborating with the University of Texas El Paso, has demonstrated that with a dependable thermal-feedback control system, build parameters can be adjusted to control the cooling rates throughout the build and maintain consistent grain size.
Making Magic in the Lab
The NextManufacturing Center team also is hard at work developing process maps to allow aerospace manufacturers, and other adopters of metal-additive manufacturing, some leeway in how they operate their machines. As explained by the center’s executive director Sandra DeVincent Wolf, the machines—laser and electron-beam—can do much more than they currently are asked to do.
“The process-variable sets being used are very narrow,” Wolf explains. “We’re showing that it’s possible to work with a wider range of parameters—beam power and travel speed—and manipulate the process to achieve very specific outcomes. Within 5 yr., we expect that this will become an important skill set in industry.”
Two important aspects of bringing this work to life: controlling porosity during the build; and allowing manufacturers to work with a wider variety of powders—varying in particle size and shape—and still yield acceptable parts.
“We’ve identified the parameter windows for beam power and diameter, speed, layer thickness, path spacing and local part temperature that users need to stay within in order to control porosity,” says Rollett. “Research on porosity control has the potential to eliminate concerns with porosity across powder systems and alloys.”
Most recently, CMU mechanical-engineering Ph.D. student Sneha Narra explained, researchers have begun to understand how to vary process parameters at different locations in the build to customize material characteristics—more ductile in one area of a part, for example, and harder or more impact-resistant in another area. And, fellow mechanical-engineering Ph.D. student Colt Montgomery is looking closely at a fairly common concern regarding 3D metal printing: the challenges of printing overhanging features. A key variable here: being able to change beam-spot size during the build, which gives the user more flexibility with power and velocity. (The students note that not every 3D metal-printing machine offers this capability.)
New Powder Alloys on the Horizon
It’s also important to understand and prepare for the coming increase in the types of alloys available for 3D metal printing. For the most part, explains Wolf, manufacturers are printing with the same alloys that have been developed for other processes, such as casting and thermal-spray coating. However, that soon may change, she says, thanks in part to a new project, led by CMU assistant professor of materials science and engineering Bryan Webler, to develop new powder alloys designed specifically to take advantage of the high cooling rates of additive manufacturing.
An Opportunity to Guide 3D Metal-Printing Research The NextManufacturing Center at Carnegie Mellon University invites organizations to support its efforts to advance 3D-printing technology to enable widespread adoption and certification. Membership, at various levels, allows opportunities to guide the overall direction of research projects, invitations to attend member meeting, gain access to project progress reports and access to training programs. The star-studded list of NextManufacturing Center contacts and collaborators includes founding consortium members Alcoa (now Arconic), ANSYS Inc., Bechtel Marine Propulsion Corp. (Bettis Naval Nuclear Laboratory), Bosch, Carpenter Technology Corp., Federal Aviation Administration, General Electric Co., General Motors Co., Ingersoll Rand Inc., National Energy Technology Laboratory, SAE International and United States Steel Corp. Other center collaborators include Lockheed Martin, Oberg Industries, Eaton, Covestro, Kennametal, NIST, Sandia National Laboratories, NASA, Sciaky and Pratt & Whitney. Learn more by visiting www.engineering.cmu.edu/next.
“These new alloys will have unique material properties,” Wolf says, “combining strength, corrosion resistance, creep resistance, etc. Aerospace companies and others surely will recognize huge opportunities to improve part performance using these new alloys, which should become available within the next 10 yr.”
An Opportunity to Guide 3D Metal-Printing Research
The NextManufacturing Center at Carnegie Mellon University invites organizations to support its efforts to advance 3D-printing technology to enable widespread adoption and certification. Membership, at various levels, allows opportunities to guide the overall direction of research projects, invitations to attend member meeting, gain access to project progress reports and access to training programs.
The star-studded list of NextManufacturing Center contacts and collaborators includes founding consortium members Alcoa (now Arconic), ANSYS Inc., Bechtel Marine Propulsion Corp. (Bettis Naval Nuclear Laboratory), Bosch, Carpenter Technology Corp., Federal Aviation Administration, General Electric Co., General Motors Co., Ingersoll Rand Inc., National Energy Technology Laboratory, SAE International and United States Steel Corp. Other center collaborators include Lockheed Martin, Oberg Industries, Eaton, Covestro, Kennametal, NIST, Sandia National Laboratories, NASA, Sciaky and Pratt & Whitney.
Learn more by visiting www.engineering.cmu.edu/next.
Computed Tomography Paints Porosity Portraits
A large area of focus for the NextManufacturing Center is powder characterization and porosity, and defect analysis. Among the goals: understanding better how process variables impact part quality, and developing practical guidelines for setting limits on defect size and population so that users can then optimize their processes to work within those boundaries.
Walking me through the process of part-microstructure imaging and analysis in the lab were CMU materials science and engineering Ph.D. students Ross Cunningham and Samikshya Subedi. They’re using the latest in supercomputing tools to analyze data collected at the national synchrotron-radiation light-source research facility at Argonne National Laboratory in Chicago, IL.
“Part of our work here centers on powder characterization,” Cunningham explains. “We’re learning how the raw materials used impact part properties and defects. From the 3D images and models, we’re learning how powder porosity transfers to the parts, and have been able to share that knowledge with the powder suppliers. This work also supports the process-mapping projects underway at the NextManufacturing Center—looking at porosity results as the process changes.
“So, you might optimize the process for build rate or part tolerances,” Cunningham continues, “but the map might not be optimized for defect population. By better understanding this variable we can help users of the technology better balance their processes accounting for all of the build properties.”
Computer Vision Comes to Additive Manufacturing
Lastly, we sat with Elizabeth Holm, CMU professor of materials Science and engineering, to discuss work she’s leading for the NextManufacturing Center on automating the process of analyzing microstructural images. Capitalizing on concepts of computer vision, which enables computers to interpret visual images, this technology allows researchers to quickly and accurately analyze data-laden images and organize them by specific features—much more efficiently than can humans.
“The scans become fingerprints that represent the visual content of a material,” Holm explains. “Then, we can use that content and do many things that impact additive manufacturing. We can, for example, classify images by the material being processed, and what the processing conditions were that yielded the images. Another application involves characterizing the metal powders that we’re printing with. We then can ask, ‘Does this batch of powder match the previous batch?’ or ‘How has recycling changed this powder?’ In this way, we can train the system to recognize the powders used most often, quickly evaluate and classify a new batch of powder, and use that data to allow for timely adjustments to be made to the machine setup and processing parameters.”
Another application, of particular interest to the Federal Aviation Administration (FAA), is the ability to use microstructure images—taken, for example, of 3D-printed metal parts—to classify the success of a build.
“The FAA, Holm says, “would like to be able to scan 3D-printed structures and quickly prove-out a process based on this imagery and analysis. The industry needs to be able to efficiently classify successful build microstructures. Today this is done subjectively, by manually examining properties such as grain size and orientation. Our work in computer vision will allow this process to be automated, allowing the FAA and others to efficiently qualify builds, write specifications and establish visual tolerances.
“All of this,” concludes Holm, “opens up new avenues for technology development, and material scientists and designers will be working hand in hand to help 3D metal-printing processes evolve.” 3DMP
View Glossary of 3D Metal Printing Terms
Advanced DfAM Course September 28-30 in Youngstown, OH
Friday, September 16, 2022
ASTM Hosting Advanced DfAM Course at America Makes
Wednesday, July 27, 2022
New Extended-Reality Basic Training App for EOS M 290 Machin...
Wednesday, June 15, 2022
Pilot Apprenticeship Program the Focus of Meeting on AM Work...
Tuesday, November 23, 2021