Part 2 of our in-depth Q&A with UWaterloo’s Ehsan Toyserkani.

It’s no secret that the additive manufacturing (AM) industry is facing challenges. The enthusiasm that once greeted novel 3D printed parts has been replaced by skepticism about the viability of this technology beyond the niche use cases where it excels.

In the first part of our discussion with Ehsan Toyserkani, we discussed some of the most prominent 3D printing myths that often give rise to this sort of skepticism. In the latter half of the Q&A, we turn to the future of AM and the challenges that will need to be overcome to get there.

engineering.com: Based on your expertise in in-process monitoring and closed loop controls, what do you see as the major challenges for these areas in additive manufacturing?

Toyserkani: One of the major challenges is the lack of standards. In an effort to come up with a remedy, my group has been working with ASTM and ISO over the past three years. We’ve proposed a workflow for in-situ monitoring using a specific sensor, a photodiode in this case. This standard has been vetted by people from NASA, GE, Siemens, because we need to address all critics and clarify each point, one-by-one. It’s been a difficult three years, but now the standard is ready for publication and, hopefully, it will be out in the next few months.

The other challenge is collecting real-time data from the processes. It might not seem like it, but laser powder bed fusion is really high speed: the laser beam is moving four or five meters per second, so you’re collecting lots of data. In our lab, we have an L-PBF system that uses two sensing technologies — optical tomography and photodiode melt pool monitoring — and after one print that takes around 70 hours to complete, the data size is in the range of terabytes. With that volume of data and the way you have to collect it and synchronize the sensors, we see a lot of challenges in hardware. Then there’s the process variability that comes from the environment and other disturbances which introduces noise into the data.

I had a mentor, Dr. Jyoti Mazumder from the University of Michigan, and he was a renowned scholar in laser-based manufacturing. I was talking to him about the data we were collecting from laser powder bed fusion and he said, “Wow, these are very messy datasets!” [Laughs] But it’s the complexity of multiphysics phenomena that adds to this challenge. It’s not just a thermal process – it’s also a fluid process and it combines mechanical and chemical phenomena all together. This is what we call sensor fusion, and it is a big challenge.

Given these challenges and the state of the industry more broadly, are there any particular developments in materials or software that you’re watching closely?

Well, obviously I’m not going to endorse any particular companies, but there are a lot of activities at the material level: at the university, we’re developing process recipes for unweldable materials, super alloys and so forth. There are also significant advancements in the application of artificial intelligence and machine learning for the prediction and optimization of AM processes. We actually already have a patent pending in this area, though it hasn’t been granted yet.

Then there are activities by some companies for parameter optimizations for defect detection and quality assurance – very much the same as what we are doing – and also the use of AI for design for additive as well as optimization of supply chains. I couldn’t make it to RAPID this year but I’ll be going to Formnext and chatting with some of these companies to see where they stand.

AI is getting a lot of attention because dimensionality reduction is very important [in additive manufacturing]. So, because we’re dealing with not only large but multidimensional datasets, including images, we need to reduce the dimensionality without jeopardizing the integrity of the data, so this is a very important topic.

There seems to be a consensus that AM adoption has been slower than expected, or at least slower than the industry was hoping. Would you agree?

Absolutely, yes. I think one of the reasons for that is overpromising on capabilities which led to frustration in the manufacturing market. There are other issues related to the marketplace and disagreements over mergers. Multiple mergers could have happened for the benefit of the AM industry but, unfortunately, those didn’t happen and now the valuations of companies are coming down. Then there are higher valuations for some companies that undermine the AM industry as a whole. So, there’s less money in AM these days and that’s a challenge.

When I heard about the high valuations of some of these companies three or four years ago, I told them, “This is going to be a major loss, because they’re not going to maintain the expectations of their investors, and if there are unrealistic expectations, that puts pressure on the whole industry, not just one individual company.”

I also think a lack of understanding of the importance of design for additive was another challenge that’s caused issues for adoption. AM is not good for just any component, so you really need to do lifecycle assessments, and that can be done in collaboration with universities. We’re here to help de-risk the technology. But you had companies investing in additive by purchasing a few million-dollar machines and – because they didn’t do their homework – they ended up disappointed. That’s why you need to find the right niche. Finding the right application would help not only the company, but also the entire AM industry.

There are a lot of companies I’m working with that have found the right applications but I can’t disclose their names because it’s confidential, and that’s another problem. There are fantastic stories in the market about setting up serial production, but many companies want to keep that confidential because of competition. But there are at least three examples I know of in Canada where the companies are making fourteen or fifteen thousand parts per year at a good margin, but they don’t disclose it.

More often, you hear about companies that tried to adopt additive, they weren’t successful, and so they put the machine up for sale. The good news is that the hype is gone: people are more realistic because they have a better understanding of the merits of additive manufacturing.

On that optimistic note, if we’re looking ahead to the next decade, how do you envision AM evolving in industrial applications?

Obviously, I believe there are a lot of opportunities for the AM industry. With the hype gone and people being more realistic, that’s a great opportunity for further adoption of the technology across various industries, like medical, aerospace, tooling and energy.

I think there will be more advancements in materials, with materials designed specifically for additive to improve the quality and reliability of end-use components.

I think automation and the use of robotics goes hand-in-hand with 3D printing, and that will streamline the market so that the trial process is more of a complementary component to the main manufacturing stream, as well as obviously being more efficient and reducing the need for manual intervention.

Are you talking about pallet loading systems moving parts between printers and post-processing operations?

Yes, exactly. For post-processing specifically, having less human involvement is very important. One of the challenges right now is how to do de-powdering in an automated fashion. For components that are quite large, I think that’s coming soon, but in ten years’ time I think we’ll have more automated 3D printers as components of a larger manufacturing line that’s fully connected and fully automated. At that point, we could have more actual serial production and eventually reduced costs and lead times over conventional production. Sustainability may play a critical role in adoption of additive manufacturing as well, since we would have less waste and more local production.

You hear some people saying, “RIP to 3D printing” today but this is not correct. It’s still coming, but in a more realistic way.

Last question: Do you have any advice for young engineers that are looking to pursue careers in AM?

This is a very dynamic field. Obviously, there are a lot of challenges, but challenges bring opportunities and there are a lot of R&D opportunities here. I would recommend you go with hands-on experience: stay tuned with the AM industry and try to develop your CAD and design-for-manufacturing skills. I also think it’s very important to learn about quality assurance, AI and how these are relevant to AM.

Sustainability is also very important for the younger generations. Think about sustainability and how we would be able to actually minimize energy and waste when you are producing components.

I can tell you that we receive a lot of requests from people to join our lab, far beyond our capacity, but that’s why we’re trying to include more training in our curriculum for 3D printing. It’s all coming together and I think the next ten years are going to be fascinating.