It was around this time last year that Eric Barnes shared his thoughts on additive manufacturing. He was in excellent company. With him were experts Alison Park, Youping Gao, Paul Gradl, and Dan Braley, all of whom rank among the aerospace industry’s most knowledgeable and enthusiastic supporters of 3D-printing technology. At the time, Barnes predicted that additive manufacturing (AM) was “about to enter a golden age,” and that we (i.e., the AM community), have survived the Gartner Hype Cycle’s notorious “Trough of Disillusionment” and are headed into the all-important next phase: the “Plateau of Productivity.”
He should know. Barnes is a Fellow of advanced and additive manufacturing with Falls Church, Va.-based Northrop Grumman Corp. He’s spent a large chunk of his life looking for ways to make the F/A-18, B-2, F-35, and countless other aircraft and flight systems faster, safer, and more cost-effective to build. He’s also watched “rapid prototyping machines” evolve from their stereolithographic roots to today’s multifaceted, production-grade, highly reliable metal and polymer 3D printers. And despite his view that AM is poised to become abundantly mainstream throughout the aerospace industry, Barnes will also tell you we still have a long way to go.
Manufacturing Engineering (ME): Let’s start with a fun one. Have you been involved in any cool projects recently?
Eric Barnes: To be honest, most of our work has non-disclosure agreements, so I can’t share too many details, but I can tell you that Northrop Grumman has transitioned four different AM materials into air vehicle production. These include laser-sintered Nylon 12 as well as a PAEK polymer with ESD [electrostatic discharge] capabilities, and we’ve signed off on ULTEM 9085 polymer on the B-2 program. In this case, we replaced 11 cockpit ducts, reducing the schedule by 16 months and eliminating $1.5 million in tooling. And on the ducts themselves, we saw a cost reduction of 55 percent. That was a great outcome. And lastly, we have implemented electron beam powder-bed fusion of Ti-6Al-4V [a titanium alloy] on one of our air programs along with several satellite programs.
ME: When we last spoke, you seemed a bit frustrated with the time needed to certify parts for aerospace use. Has anything changed?
Barnes: It’s getting better in many ways, although it still takes longer than I would like. Our customers are becoming a bit more specific about what they want in terms of AM parts testing and validation, and this helps streamline the process somewhat. But, from my perspective, we need to find more efficient ways of doing both. Granted, there’s significant work being done within the research and development arenas. We’re seeing better integration of modeling and simulation as well as probabilistic analysis methods, but still have work to do in developing a cohesive technology approach for rapid qualification. I think we’re still several years away from having a more digital approach to qualification but are making consistent progress in that direction.
ME: At the risk of asking a dumb question, what’s involved in the certification process?
Barnes: Just to be clear, you qualify a material and process, but you certify a component, structure, or design. So qualification requires the development of a B-basis [the value at which 10 out of 100 specimens will fail] allowables database based on a well-controlled and well-defined process parameter envelope, and includes 1,000-plus static tests at various temperatures and conditions. Various levels of fatigue, durability, damage tolerance, and other testing needs to also be performed. And then we typically validate those allowables through a representative structural test of a representative component. That allows us to analyze the part, know where it will fail and at what level, and demonstrate that our design allowables are sufficiently conservative.
ME: You’re clearly a proponent of digital simulation. If customers embraced it, could you minimize current testing methods, thus decrease development time and cost?
Barnes: Absolutely, although I don’t see that the industry will ever eliminate physical testing completely. That’s especially true when working with the government and the DoD [Department of Defense]. They’re conservative, and that’s understandable, but at the same time, I’m confident that incremental increases in modeling and simulation usage would lead to less and less testing. There’s also the potential for newer statistical approaches than those we use today that reduce the need for material specimen and coupon testing.
ME: With that in mind, are you satisfied with the current level of simulation technology?
Barnes: That’s a good question, and, unfortunately, the answer is no. I don’t think we’re there yet, and I certainly would not feel comfortable taking something to one of our customers and saying, “Hey, our simulations show that we’ve solved the issue; Let’s do it this way.” This is why simulation is still primarily used in the research and development phase. Part of my frustration with this, however, isn’t so much that there aren’t yet any robust solutions, but rather that we waited so long as an industry to consider simulation as an alternative to traditional qualification and certification methods. We‘ve been doing qualification and certification the same way for decades.
ME: Do you think bringing additional suppliers into the aerospace fold will also help reduce costs?
Barnes: A couple of things, actually. The materials and equipment have come a long way in terms of stability and repeatability, but there‘s certainly room for improvement. Here again, we’re not there yet. And I also think we’re not there yet on cost. Additive is not as cost-effective as we believe it should be. We rely a great deal on our supply base for providing AM components and it seems that parts should be getting more affordable, but we haven’t seen much movement in that direction. Automated pre- and post-processing have helped, as has automation of some 3D-printing equipment, but I have some hope that newer multi-laser systems will help bring down part costs even further.
Barnes: Perhaps, but to be blunt, we see far too many suppliers that advertise themselves as aerospace-qualified and, once we start the evaluation process, find just the opposite. They’re either unwilling or unable to check the necessary certification boxes, or they don’t have the capabilities to complete all of the manufacturing steps in-house. For example, many can do the 3D printing but then have to send parts out for final machining, heat treating, and NDI [non-destructive inspection]. This often means waiting several months for finished parts, not to mention paying more than we would if the shop was vertically integrated.
ME: We’ve also talked about the development of more advanced in-process monitoring systems, which ties back to what we were just discussing. Any exciting news there?
Barnes: It’s moving along, but we and others are still trying to figure out the right type and combination of sensors, what data is most important, and what you do with it all once it’s gathered. In a typical build, you’re potentially talking about terabytes of data, and while companies are working on reducing this by ML [machine learning], compression and other methods, we haven’t yet seen the silver bullet. Once you have it, though, I expect you’ll no longer have to NDI everything like we currently do, and that machines will become smart enough to use in-process data combined with ML to prevent defects.
ME: Last question. My understanding is that most of the focus for AM parts is maintenance and repair operations (MRO), which is referred to as sustainment. Is that changing?
Barnes: Absolutely. Although MRO is essential, it represents a minority of our efforts. We now have hundreds of production parts in autonomous vehicles and are transitioning to more and more crewed air vehicles. We recently received approval to implement specific additive processes up to durability critical, and even mission critical with review from a board. Above this are structural components. We’re on target to get our first true structural part—structural normal controls, as it’s called—placed on an air vehicle by the end of this year. We’re also heavily involved in a less-discussed type of additive manufacturing, continuous-fiber composites. In fact, we teamed up several years ago with machine builder Electroimpact (Mukilteo, Wash.), an effort that resulted in six patents. So despite what I’ve said earlier about all the challenges still facing us and the sometimes frustrating pace of change in the aerospace industry, it is still a very exciting time.
