As Head of Robotics at TRUMPF, Inc., Andreas Bunz has to see the entire process of fabricating. Robots come into play in many different places along the value chain, and in many functions—everything from tooling changes to machine tending to palletizing to transporting blanks and parts to their next destinations. Bunz is in a unique position to comment on the new world of fabricating, as he is right on the vanguard of it. In also managing the TruLaser Center 7030, he knows both integrated technology and purposely separate technology (like some of the robots!) and although he never mentions the word, he definitely sees the technological “habitat” of tomorrow.
Fifth Wave Manufacturing: Hello everyone. This is Dave Brambert, President of Fifth Wave Manufacturing. We’re here today in Hoffman Estates, Illinois at the location of one of the Smart Factories from TRUMPF. And with us today, Andreas Bunz, who does multiple duties here.
Andreas Bunz: That’s correct.
FWM: You take care of the 7030 and robotics. Today we have a chance to talk about robotics, and I thought we would take advantage of the available time. Thank you for having me here today, Andreas.
Bunz: Great to have you here, Dave, and as you said, I’m head of robotics here in Hoffman Estates, just outside Chicago, and the product manager of the TruLaser Center 7030. The 7030 is sort of the second role, and I’m doing that sort of in conjunction with my other duties right now.
FWM: And robotics really is found all over within TRUMPF. So is AI, and I’m sure both of those areas intersect for you. If my kids were a little bit younger, I might say, you should check out robots, or you should check out AI, but particularly robots. After all, a lot of AI function and a lot of vision function comes from robotic work.
Bunz: Yes, that’s correct.
FWM: Please talk about what TRUMPF is doing, where some of the biggest impact is from.
Bunz: TRUMPF is offering off-the-shelf solutions for sheet metal fabrication: lasers, press brakes, welding machines. The TruLaser Center also is one of the—if not the most—advanced laser cutting and sorting machine. However, not every challenge, not every goal that our customers have can be addressed with our off-the shelf-products.
So the idea was to do things—complementary to our standard portfolio—like custom robotic solutions that go well together with our products. And first and foremost of these is bending. Parts might be too large or they might be too heavy for our regular automated press brake portfolio, which we also see behind us in the Smart Factory. And secondly the sorting operation, the sorting of laser cut parts is a challenge that until 12 years ago, nobody had solved. That’s why TRUMPF successfully developed and established the TruLaser Center that that cuts and sorts autonomously like no other machine.
Now on the robotic side, it opens up more possibilities if we cut the parts on a very fast and efficient laser machine and then present it to a robot that does not even need further programming, but detects the parts and can sort them and stack them autonomously. It could really go well together with our established portfolio.
FWM: I was thinking a part could even be too small to really be safely handled by human hands. And here comes a robot for that too.
Bunz: Too small or, we have thin parts, right? They might be sharp, so that regularly employees get hurt. It’s a tiring operation if you only have to do two bends and it needs a couple of thousand parts a month. So there’s a lot of reasons why robots make sense to go with our products.
FWM: Yes. And the systems around that seem like they accommodate the robot very well. Even the tooling rack, every tool in that rack, whether it’s a punch or a die, has an RFID tag or some other kind of identifier so that you know where that is at any time.
Bunz: If we talk about robotic solutions for press brakes, of course we have the TruBend Cell 7000 and 5000 that also do tool changes. On a custom solution, we also want to incorporate the tool change to allow as much autonomy as is possible. However, if a customer has a goal, they want to do very flexible manufacturing from medium-sized to very large parts then we look into that and maybe the quantity for this project is not so high but we still qualify—find and qualify—and then build a custom solution for that customer, which maybe otherwise would have not gone with the TRUMPF portfolio.
FWM: Interesting. I didn’t think of the custom aspect of that. As we watch the whole industry develop over time, much was devoted to “our laser goes faster than your laser” or “our speed of the press brake until we hit that short approach speed is so fast you’ll never beat it.” But we shifted from that and said, “okay, we can make all these things go faster. The real time savings now is in between steps.” And we’re doing the things that humans do or that humans don’t do.
Bunz: I think there’s a natural advance, like laser power is increasing naturally so then the cutting speed increases. But everyone is realizing that it’s not about the cutting time as much. People look for a lot of green lights on the laser or on the cutting machine. But what happens after that? Where does the part need to go? How does it get there? Is it ready for the next step? Do I have to handle it? Maybe I have to break it out, have to sort it, stack it, count it, forward it, so we could talk about robotics or AGVs and logistics. How can they be integrated and what happens when they arrive at the final destination, which could be a press brake and suddenly we can connect them very flexibly, right? Yes. And yes, the indirect processes…studies have shown that 80% of time is not value added at the moment. That we address not only with robotics, but also with intra-logistics solutions.
FWM: And if lasers were living beings out there on the floor, they would have a good work life because they start off everything. If they mess up, the cost of a mistake isn’t anywhere near what it would be post-welding, for example. You can always get back to feeding another sheet. But having said that, that takes time too. And a lot of the robotic, and even non-robot automation within, say, a laser cutter is aimed at first pass is just as good a pass as a 10th pass. It’s the same with press brakes, first bend is a usable bend.
Bunz: First piece, good piece. Absolutely. And exactly like you said, further down in the production chain, it only gets more expensive if mistakes are discovered later. And it must not even be a real mistake. It can be just be one part missing and development cannot be finished.
FWM: Well, that’s it. Good point.
Bunz: I talked about sorting and then bending. All of that can be helped or solved with robotic solutions aided by vision and maybe AI. I think there’s a lot ahead of us. And then we approach welding and robots have been big in welding for quite a while—industrial robots or cobots. But how do the parts get there? How do they get into the fixture? Do we always need a perfect fixture? Is it multiple robots handling parts? I mean, there’s so much out there, and we didn’t even leave the production shop floor.
We can talk about kitting, for example, multiple parts from different primary operations. It’s a very broad portfolio. We are working on a project at the moment that involves sheets that need to be fed to our STOPA system. How do they get loaded? So there’s ideas and there’s concepts to load them with robots on the STOPA carts. That also eliminates error. It makes sure that every stack is perfectly square, that it’s in the right location, it’s the right count, it’s the right material, and that a laser center in this case can produce uninterruptedly, because nobody accidentally booked the wrong material code to that pallet.
We look at the whole value chain. Where we see a need or where our customers tell us that there is a need to say, “please TRUMPF, Can you think of something, even if not immediately, but we want to get better there.” Also, we have customers that still need to integrate that into our software ecosystem, for example. Connectivity and transparency is a big piece to make that work.
Also thinking about robotics on the shop floor: I believe at the moment we have in total seven different robots integrated as part of our machines. That number will only go up.
FWM: I go into many shops as I’m sure you do, and I ask people, where’s your bottleneck? And more than half the time they say welding or joining or whatever the case is. What if somebody invented a fixturing robot?
Bunz: We can invert the principle of the TruLaser Center. With the many pins that lift our parts, why not have them come into shape and the parts that sits on it and it adapts based on 3D geometry. I could see that happen. If you have a certain, a set geometry family of parts. It’s only a function of complexity to make that happen, I think.
FWM: I would think that’s where the most time might be gained. Now, TRUMPF is famous for having data for studying things. Do your studies and does your data help you say, we need to concentrate on this area of robotics right this minute? Does all that data help you do that?
Bunz: TRUMPF collects or has data available for that. I would phrase it in a way to say it doesn’t spark the field of activity where we look into robotics. But once we have ideas, once we hear requests then we try to double check, to get a solid data foundation underneath that supports the assumption that it makes sense, for example, to automate intralogistic workflow of a certain process within our fabrication environment. Or perhaps a certain bending process can be made better with a custom robot solution relative to our current portfolio. I am not saying that the current one is not good. TRUMPF makes very flexible equipment. But if somebody has a very specific requirement, there might be a more efficient solution—it could be a shorter cycle time, or it could be more space efficient. Or it could be higher autonomy so they can produce lights out in the night, and then we would look into the data, like which parts are they producing, and try to support our studies with that data.
FWM: Just in the welding piece of business here, you are employing both a cobot in the form of the UR (Universal Robots) and the TruArc Weld 1000, and then in the TruLaser Weld 5000, there’s a KUKA that’s the basis of that. And is the line about a robot being a little more precise true? Are you finding that to be the case or not?
Bunz: I think across all brands, cobots have the integrated safety which an industrial robot does not. So an industrial robot also has to have some surrounding equipment to help with the safety aspect in our environment. I think that’s the biggest difference. Technically, speaking on the arc weld, on the operational side, you can teach the cobot. In this case, it’s very simple, but with an industrial robot that is not so simple. But on the welding result, I believe that would not make a big difference. Now, on the bending side, it’s a different story. We want to be able to program robots offline to have, like we said earlier, first part, good part also on a fully automated bending operation. Only then can you bring in new parts online fast and still run it in a flexible way. And if you do offline programming, then it doesn’t matter. Is it a cobot, is it a robot? A robot always will move faster. It doesn’t need to account for the integrated safety. It has fencing around it, there are other peripheral components around that take care of that. So industrial robot will always be better in terms of time efficiency during a production run, as long as you don’t need the human interaction.
FWM: Yes. And in that case you want the robot, but for doing the coupon demo of the TruArc Weld. With the cobot, it’s not complicated. It’s something where somebody can say, okay, start here and end there.
Bunz: Exactly right. And we have the two-station operation in the TruArc Weld. The operator can be very close to the other station while it’s actually working on something. And with the industrial robot, that setup would certainly look different. We have an integrated industrial robot that is very versatile and very fast in feeding those little parts in a high quantity to the 7050 electric press brake. So we’re looking to have the right robotic solution for the right application. We have a 165 kilogram robot automating our panel vendor, for unloading and stacking the parts and overturns—everything that is needed.
FWM: How much is AI touching the robotics area?
Bunz: AI plays a big role. It’s surely not hype; it’s here to stay. I mean, we’re fast on the way to super intelligence, probably faster than we all really want, and few of us really understand what that means. But it does revolutionize also the way we as TRUMPF can approach challenges. Not always is the solution that we see in a classic development setting the best solution. AI tools are helping, especially in things like coding, especially in things like vision processing, and in predicting. We talked about data, and the data that we know can be fed into certain AI models to make them learn.
Our laser center extracts more than 40 parameters on every single part. It moves with the SortMaster Speed. Now, somebody might ask, what do you do with that data? At the moment, we monitor and we check statuses, and try and predict when, for example, suction cups might fail, but this will only get better once we couple that with AI and real prediction models and learning models. TRUMPF has been on that path now for a long time. In recent months this accelerated and put more emphasis on it as a focus for the company not only in R&D but for the global strategy.
FWM: I’m sure there were some smart people who could see that the main use of all this data is to improve everything through AI and provide that to customers. So they were doing themselves good by allowing that data capture to happen.
Bunz: Exactly. That is a dialogue that obviously has to happen. Not every customer wants to share all the data—which is totally fine. We want to have a dialogue. We want to be transparent what we would do with it and how a customer could benefit. And it might not be an individual benefit, but the more that participate, the higher the overall benefit will be. Improvements can be incorporated in future software versions, in updates, in bug fixes. We might get faster, we might get more tailored to what actually happens in the field. And I think that for me, it’s mostly about processing of data at the moment.
Our development cycles are relatively sophisticated, relatively long. We don’t put out a new, fully automated laser machine every five years. That is a difference relative to the past. Laser power scaling is relatively simple. Everyone can do it more or less successfully. But if it comes to bending, if it comes to welding, the new generation will be developed or we want to develop it in a way that is also here to stay and that it can then benefit from all those emerging parallel technologies like AI.
FWM: This is an industry wide thing. And you mentioned you, you can’t do a new 7030 every year, you know, here’s a 7031. And yet for other products, you can just feel it, you can sense that the cadence is increasing. And it used to be, “what are we showing at FABTECH every two years?” and now it’s every year, easily. It’s even kind of starting to go to like 10 months and eight months. Will robots help you with that?
Bunz: I’m personally convinced robots will help us. Not only because we just created the custom robotic solutions division at TRUMPF, which I’m heading. I have to be convinced. I’m excited for that. And FABTECH this year will be, will be a first—a very big milestone for us. We offer the adaptability of our existing solutions to ever-evolving requests or challenges or just ideas that our customers have. We might bring up different series of press brakes, for example, 1000, 3000, 5000, 8000, and like you said, not every series gets an update every year. We are carefully managing our portfolio on where we roll out, which updates, which features, where does it make sense, where do we introduce it? And we then carve that to fit our current market situation.
And this in the future is very exciting because let’s take a STOPA system, for example. It’s very large, the spine of all of our smart factory. It’s the central piece of material logistics in terms of from raw material to cut and and to bend parts at the moment. If somebody puts that in, it stays for 15, 20, 25 years. But our machines around, they evolve, they get changed. They need to adapt to the changing environment. And I think there is a, there is a gap at the moment, certainly that we can close with robotics with AI, with processing of relevant data and deliver better solutions for our customers.
FWM: I think that’s right. And I think AI is very related to software. And you can have the same machine do different things according to what software you put in it, even with only the tooling that it came with.
Bunz: You change a certain set of tools and suddenly you bend different parts with the different radius. We’re able to program that fully offline simulated, be sure that the first part is a good part, and predict the results. Then even during the bending process we have our integrated laser angle measurement. We feed that back to the robot, we correct the bend. So we make sure the quality is there. And that’s what customers expect from us.
FWM: Let’s say it’s 2030. Look around you. Look around you in the landscape of fabricating and what do you see that’s different?
Bunz: What do I see? I believe that the shop floor in the future will be more clean than it is today. Clean in terms of not so much waste lying around. There’s not so much material being handled manually. Transparency and knowing which piece is where, which order is where, which machine and which production job, which status—all of these things will play a big role because we need that information to automate the whole floor. And it will get more automated. We see it in logistics. Compare current logistics companies let alone the big ones like Amazon and the end-to-end solution providers, but compare them to 10 years ago or 15 years ago. The use of robots in mobile logistics solutions and so on, it’s growing exponentially.
A fabricating environment is more challenging because not every job is the same. Not every material that we get delivered is the same. You see it here. When we do production, sometimes the quality, the batch that we get, is not what we expect. And then you have to adapt to it. To find solutions that are able to do that will revolutionize the shop floor, in my opinion. And then having those flex configurations, let the machines do best what they can do best, for example, cutting, bending, welding, integrated in a flexible way that the operator will probably more become a logistics manager on the shop floor. That person will be an employee on the shop floor making sure the individual technologies are happy, but then the automation solutions, they will do the actual heavy lifting. We already see this, and the Midwest is a prime example. Good welders are insanely hard to get. And so many people don’t like to hammer out parts of a one-inch plate for 10 hours a day anymore. There are better solutions out there. And I think this will just continue and increase in momentum. I’m very happy to be part of that journey.
FWM: All I can say is I really appreciate your opinions, because I know how they’re driven. You’ve been in this business a while. I appreciate your time today, and for accommodating me.
Bunz: You’re welcome.
FWM: We’ll continue to do this as time moves on.
Bunz: That sounds very good, thank you, Dave.