TCT
Represented by one small word and taking less than a second to utter, it can be hard to believe it’s practically what makes the world go around.
It tells us who we are, shows us what we want, and proves what we need. It is the cause of much confusion, angst, stress. And although that is all true, it can still be hard to really understand.
These basic units of meaning are the focus of many a working professional, however. They take time to process, they dominate each and every day, and some are fed up with it.
“Data,” Boeing’s VP of Additive Manufacturing Melissa Orme, begins. “We spend a lot of time reformatting data. And I think it’s an industry standard that data scientists spend 80% of their time formatting data and 20% doing the work.”
This, obviously, is not particularly ideal for manufacturers who want to speed up and scale up, rather than be slowed down by monotonous tasks. But slowed down they often are because the various software tools required across the additive manufacturing (AM) workflow don’t work in sync like one might hope.
What manufacturers want is a digital thread of their workflow, from powder genealogy through the design of the part, build of the part, and post-processing of the part, with each tool along the way able to talk to the previous and the next. But what they get is disparate solutions failing to format data upstream and downstream.
“What we’re doing today simply isn’t scalable,” Erica Vlahinos, VP of AM at Authentise, says. “Importing/ exporting in and out of each tool – and it seems the list of necessary tools grows every day – is time consuming and leaves insular pockets of data with no ability to sift through it all.”
“Companies are not working in parallel, they're working in sequence,” Dominique Galmel, DELMIA Portfolio Director at Dassault Systèmes, adds. “They’re waiting because they know there will be change and change is difficult to manage.”
“We need a chain,” Vlahinos continues. “But no one wants to link up. We have information, but without integration we can’t decipher the whole picture.”
This frustrates the likes of Boeing because it prevents, or at least makes more difficult, the development of machine learning models that could provide essential information in the production of parts.
“Why do we care about machine learning models? It’s so that we can predict what will happen in our additive manufacturing factory,” Orme says. “We can create models that will tell us when to expect something to happen and it will give us the understanding of why it’s going to happen. And it will help us decide what we should do about it.”
Get your FREE print subscription to TCT Magazine.
Exhibit at the UK's definitive and most influential 3D printing and additive manufacturing event, TCT 3Sixty.
Boeing’s workflows are slightly different to most in that the aerospace and defence manufacturer has developed its own solutions for most of the AM workflow, but when it comes to steps of the process it doesn’t have an in-house solution for, the frustration is the same as any other AM user. What Boeing has developed upstream and downstream doesn’t always get along with the externally manufactured solution that fills the gap – like simulation, for example. It means progress is slowed down. And manufacturers don’t like being slowed down.
HURDLES
Digital threads are important, as Orme puts it, because they build in efficiency, predictability, and potentially scale. JEOL AM Applications Engineer Jon Buckley also adds traceability to that trifecta, telling TCT: “An untracked change in a CAD file can potentially continue down the next process steps and result in unintended changes to digital files in later manufacturing steps, potentially resulting in non-conformance.”
The digital storage of information is at once a great enabler for AM and a great hurdle. The reasons manufacturers are frustrated are plenty.
Galmel suggests that many manufacturers are selecting ‘best of breed’ applications for every step of their process, and while they are ‘good at what they do’ individually, “the digital thread is complex because you need to interface all those tools and thevexecution of these activities is taking a long time.” Allyce Jackman, Senior CFD Engineer at FLOW-3D, agrees, telling TCT that compromises currently have to be made by AM engineers when selecting software solutions that best address their problem. As you can imagine, manufacturers don’t like compromise either.
Roy Sterenthal, VP of Additive Industrial at Oqton, points out that compromise is needed but from original equipment manufacturers rather than the end users their business models rely on. “There’s a lack of standards,” he says. “It starts from the way that hardware manufacturers have interfaces that allow us to capture data.”
Sterenthal speaks here from the software providers’ point of view, but Orme says Boeing is on the same page: “To be able to extract the data that we want, some machine manufacturers are really good, and some don’t want you to at all. If that machine supplier says you absolutely cannot see the data, I’m not interested in that machine.”
The challenges, then, extend from the disconnect between software solutions to the openness of machine providers, with the capabilities of these tools not escaping constructive criticism either. Jackman notes how ‘a model that can predict part scale distortions and thermal stress cannot [necessarily] predict bead geometry and interlayer fusion due to the computational runtimes involved,’ while Orme has seen some of the best simulation tools on the market fall short when confronted with the most complicated part geometries.
SOLUTIONS
Oqton, Dassault Systèmes and FLOW-3D are all too aware that advancements are needed within the software platforms they are bringing to market. Oqton, whose manufacturing execution system software is supplemented by an umbrella that includes tools like the 3DXpert build preparation platform has demands from users to ‘close the loop’ with monitoring and inspection features for traceability. Dassault Systèmes, whose software solutions caters for part selection, design, engineering, and virtual printing is being encouraged to place a greater focus on print process analysis, while computational flow dynamics firm FLOW-3D is working to deliver better microstructure prediction capabilities through the analysis of heat transfer, fluid flow, and more.
Improvements in each of these areas will help to enhance the individual products, but what has to come next is openness, collaboration, and an understanding that a challenge like this has to be addressed by an entire industry. While the standards Sterenthal alluded to were based more on values and principles than benchmarks and criteria, there is an argument that the latter are needed just as much as the former. Between machine makers, software providers and end users, they all need to speak the same language to get the best out of their AM workflows. The phrase ‘a rising tide lifts all boats’ may be cliché, but when it comes to digital threads in AM it might just be true.
“Either through industry consensus or competitive pressure – whichever comes first – providers must move towards open platforms as a standard,” Vlahinos says. “The success of the digital thread is contingent on that openness. Once established, it will unlock the collaboration and insights necessary to propel the AM industry forward."
“A digital twin that seamlessly links together solutions across scales would significantly benefit those working with new materials or new applications who need to go through a qualification and process design stage,” offers Jackman. “This becomes important for mission critical applications such as in aerospace and defence industries where qualification and certification are still largely done manually via experiments.”
“It’s all about the big promise of additive,” Sterenthal finishes. “One of the first things that additive should enable is for you to manufacture when you need it, where you need it. Once you have the digital thread implementation in place – flexibility, reliability, repeatability – then you can fulfil the promise of additive.”
