Posted by Bharat Parihar
I recently attended the Additive Manufacturing in Defense Forum organized by Deloitte Consulting LLP, where participant Paul Boulware, an EWI applications engineer of drove home one particular point: “To make one-inch cube of metal, it takes five miles of weld and half a day. As you are building these parts, knowing the quality of the part is crucial. You cannot wait to know that there is a problem in layer five or six or seven, after the part is built” [and therefore fail your quality control tests].
Therein lies the challenges of quality control in AM processes: This layering of substrate upon substrate, or the systematic fusion layers of metallic powder, must be controlled and corrected to ensure that a part’s manufacturer meets quality control and quality assurance specifications.
But what is “quality”? As with many things in life, that depends…
For parts that have high quality requirements, advanced simulation and modeling tools are crucial for accelerating the quality assurance or quality control process. Moreover, industry collaboration is a viable way to expedite the creation of such simulation and modeling tools.
At this forum, Dr. Ade Makinde of General Electric Global Research, Dr. Wayne King of the Lawrence Livermore National Laboratory, and Paul Boulware of EWI delved into what constitutes quality–and the quantification, simulation, and realization of quality control–for additively manufactured components.
Quality control ranges from “low” to “medium” to “high,” depending on the threshold of acceptable quality for the manufactured part. At the lower end of the range, there are printers to make pancakes; at the higher end, aerospace and technology parts.
The intended purpose of a part therefore determines the quality assurance rigor (and the subsequent cost). Just as one does not need a jet to buy groceries, one must design one’s quality control/quality assurance requirements based upon one’s needs.
At the lower end, this is relatively easy–AM processes incorporate sensors that provide visible, audible, tactile, and even parameter driven feedback. That often proves sufficient, especially for rapid prototyping. On the higher end of the scale, it gets more complex, and this is where advanced physics, feed-forward (rather than feedback) mechanisms, advanced simulation technologies, and industry collaboration come into play.
Dr. Makinde made a case for physics-based tools for the rapid qualification1 of additive manufactured parts. A combination of physics-based tools and data driven models are needed to rapidly qualify a part. Ways to efficiently manage the large volume of data generated by data driven models are needed to enable feed forward and machine control.
Building upon the physics theme, Dr. King demonstrated the advanced simulation and modeling tools developed at the lab. These tools could accelerate the qualification of AM parts.
Indeed, such computations could be incorporated in AM devices as feed-forward mechanisms to anticipate (and therefore compensate) for the “melting, solidification, spots, and spatter” of metal powders during the AM process.
The cost and expertise of building these tools is significant, but as Dr. Makinde indicated that GE is seeing the time required to simulate a part level distortion build drop from several weeks to about two days, through their partnership with AmericaMakes (the national additive manufacturing innovation institute) that fosters industry collaboration.
To conclude, to realize high quality, additively manufactured parts advanced simulation and modeling tools are crucial for accelerating the quality assurance/quality control process. Secondly, industry collaboration is a viable way to expedite the creation of such simulation and modeling tools.
1Rapid qualification alludes to the process of verifying that a part meets its design specifications.