Additive Manufacturing for “Real” Production: Where are we?

Here’s my take following Formnext 2018 and a year of evolving insight ...

There was one word that was used more than any other at this year’s edition of Formnext in Frankfurt, which took place this week. You guessed it: PRODUCTION. It was invariably accompanied by other words, usually “serial” or “high-volume” — all within the context of additive manufacturing (AM) processes. This is, and has been for a while, the ultimate application goal with AM. Indeed, now more than ever, it appears to be within reach, with some evidence, and plenty of hearsay, to suggest it is happening right now. 

For context, and following the changes of the industry as I see them, I am now differentiating between manufacturing and serial production applications of AM, as well as the well-established prototyping and increasing tooling applications.Specifically, I find myself categorizing applications as follows:

Product Development / Prototyping:using AM/3DP processes for developing new products – to prove concepts, improve design integrity through fast iterations and determine form, fit and function. Material selection and identification can also be included here – prototyping in the end use material, or close. 

Tooling:Using AM/3DP processes to produce rapid tooling and/or investment casting parts to speed up the production process for small to medium volume parts and to overcome many of the limitations of traditional tooling methodologies including but not limited to very long lead times and extortionate costs. 

Manufacturing:using AM/3DP for single use / very low volume products and parts. Examples from this category can be found right throughout the supply chain – from large OEMs down to individual “cottage industry” type companies. It also includes using AM for general and customized jigs and fixtures etc. 

Production– medium to high volumes (serial), high quality (to regulatory standards) repeatable production of products / parts. 

I know this is far from perfect, most notably how manufacturing and production can be used interchangeably and there is the potential for confusion here, but when offered with context and perspective, this is the best I’ve come up with, so far, at least. [Feel free to let me know any thoughts on this ….] 

Anyway, back to Formnext and the opinions that framed the conversations around production during the course of the week were many and varied — even vastly contrasting in some cases. 

For instance, one conversation was beyond bullish, and included the statement: “Serial production with AM is proliferating, everyone is doing it.” I raised a brow, and challenged that with my own thoughts that it is growing, but most serial production applications of AM are not visible; but it is far from “everyone.” 

Another conversation, was the converse, including the proposition that: “’REAL’ production with AM will take decades.” Guess what? I raised a brow, and challenged that with my own thoughts that it is growing, but most serial production applications of AM are not visible yet; but there is real progress, with evidence and I cited the usual. 

It was weird to find myself using virtually the same argument to counter two such contrary points of view. I have no idea if I left a lasting impression or tempered either of these views in any way, but you never know. To be fair, however, the majority of conversations were more in line with my position of growth and expansion around application development as well as AM being integrated into real production workflows at increasing volumes, albeit most still not in the public domain. The general consensus I picked up on this week is that serial production applications with AM processes offer the biggest opportunities for the growth of the sector. 

As you would absolutely expect, the AM Original Equipment Manufacturers (OEMs) are ‘bigging’ up the production narrative. The OEMs offering metal were leading the way, but there was also plenty of the same from the polymers side as well, notably Carbon, HP, Stratasys and 3D Systems. The latter currently offers both metal and polymer AM solutions while Stratasys and HP will, at some point in the next couple of years (maybe longer), also commercialize their metal systems that are currently in development to complement their industrial polymer AM solutions.  

The real / serial production narrative was everywhere in various forms, including but not limited to “increased productivity”, “increased / improved production workflows” and so on. So, while I do, generally, agree that this is the direction the industry is heading in terms of growth; it does not and should not in any way take away from the prolific and still proliferating prototyping, tooling and manufacturing applications with additive processes (as defined above). They are here to stay and will continue. The fact that this different narrative is generally understated — because it is now the norm for most companies that develop new products — is actually a good thing in many ways, but I am not convinced it should be quite so understated at an event like Formnext. 

The standard, well-documented production applications of AM are still highly visible and being cited as the “proof” cases. You know the ones – the GE LEAP engine bracket (metal) and the Carbon / Adidas FutureCraft shoes (polymers). It is hard to keep the frustrated tone at bay here, but this is where we are for one good reason — these are pretty much the only applications that have approval for publication and to be out in the public domain. Although, on that, there is another, albeit under-used example, that I was beautifully reminded of this week, by Lee-Bath Nelson, co-founder of LEO Lane. And that is Michelin. I can’t put it any better than Lee, so I am going to quote her: “Michelin opened the kimono on their use of metal AM production. They’re producing a million metal AM parts per year, which is incredible.” 

Michelin, however, had the means and the wherewithal to develop their own systems to achieve their goals. This was not an off-the-shelf solution, and there are not many other companies across the world with the financial standing to do this, or subsequently invest in a new company / partner to commercialize the resulting system (AddUp, which has now also acquired BeAM). 

But here is the rub, while these are the stories that continue to be told, there are actually many others that the OEMs can’t talk about in the public domain. It’s long been the case as far as ‘case studies’ are concerned, but this week I specifically asked OEMs about the “invisible applications” in generic terms – most, if not all agreed that compared with what they can talk about, there was >80% production applications, many serial production, not visible. A large majority also agreed that it was highly frustrating, particularly for the industry as a whole. 

Their inability to talk, invariably comes from their customers’ fears of losing a competitive advantage - probably rightly - if frustratingly, so. I’ve been clued in on a couple on trust, most wouldn’t go beyond talking about this generically. However, for me this does, up to a point, justify the increasing narrative and rhetoric around AM for production. There are many more happening than can be seen and the OEMs are scaling up to generate more based on what they already know but can’t say. 

Of course, there are still challenges across the entire production workflow – some specific to AM capabilities and some to do with integration. It is important to remember that not all AM processes (still, currently seven categories) have the capabilities to meet the requirements of serial production applications. And, even for the AM processes that can, they are rarely, if ever, going to provide a standalone solution, so integration into existing workflows can be a real bottleneck that has to be overcome. Costs also remain a major barrier to adoption — both capital and per-part. To date the production applications that are out / emerging utilize the capabilities of AM to justify the costs through added value that overcomes the cost either over the lifetime of the product / part (eg GE / Michelin) or through higher volumes (eg Carbon & Adidas). 

However, until costs come down, which I believe they will, wider adoption throughout the supply chain is unlikely to increase dramatically. Financing, too, can be brought in here, but I saw one chink of light on this issue, this week – from Siemens, believe it or not. Another barrier that remains is build speeds - EOS might have an alternative and interesting answer to this with its introduction of polymer LaserProFusion process this week, but it’s not an imminent solution for the market. On the metal side, GE and Betatype and others are working on solutions for speed. Other limitations are being overcome before our very eyes too, issues such as simulation, qualification and verification are key here. 

As ever – the reality of AM for production is a conundrum of positive progress and frustrating reservation. However, I believe the evolution is happening out there, right now.

This issue was so dominant at Formnext, I have chosen to write this post separately from the general overview of the show, which will include specific news, highlights and other themes touched on here. It will follow soon, and include more details on the companies, products and people I have referenced in this post, such as LEO Lane, LaserProFusion, GE, Siemens, Carbon etc.