The metal fabricator and the metaverse

ipopba / iStock / Getty Images Plus

Imagine it’s your first day at a metal fabricator that’s gone paperless and seems to be well on its way to connecting machines and becoming a truly data-driven shop. Throughout the shop you see strategically placed dashboards, each showing where work is headed; where it’s going; and whether it’s ahead, on, or behind schedule. Everything’s ticking along, smooth as can be.

Then an alarm goes off, and everyone hears a loud bang. The automated laser cutting system just crashed. The forks that pick up a cut nest jammed. Dirty slats and less-than-optimal cutting conditions trapped the still-to-be-evacuated molten metal from the kerf, ultimately welding the sheet to the cutting table. The scramble commences as unexpected machine downtime mounts.

Ideally, with the right sensors in the right place, and with data that’s used to develop helpful information and intelligence (not to mention the right maintenance regimen that requires frequent slat cleaning), such crashes should become a rarity. But the world isn’t ideal. Stuff happens, and knowing how organizations deal with an imperfect world matters. Here, modeling and simulation will play an increasingly important role.

“The 21st century will be the simulation century. The 20th century was all about the televised image. It was the first time in human history where, if we wanted to understand what was going on in the world, we could review footage of it. The 21st century will be about modeling and simulation. You can run a simulation to predict what could happen, then make decisions about what to do.”

That was Richard Boyd, founder and CEO of Tanjo (rhymes with “Bango”), a Chapel Hill, N.C., artificial intelligence (AI) firm that develops what it calls the “enterprise brain,” a concept he described during a talk at FABTECH 2019 in Chicago.

“A machine learning system can map everything in an organization,” he said at the 2019 show. “Then, the minute it detects someone is trying to do something, it says, ‘Oh, here is something similar we’ve done before. Let me pull that up for you, and let me connect you with people and the content that would help you do it.’”

He’s also founder of Ultisim, a startup that focuses on digital twin technology and the metaverse. Going beyond simulation, digital twins model a physical, real-world counterpart and use real-time data to show how that counterpart behaves in the messiness of the real world. Boyd returned to this year’s FABTECH, held in Atlanta Nov. 8-10, where he spoke on the potential of simulation and a rarely heard term in metal fabrication circles: the metaverse.

The Meta What?

The idea of a metaverse isn’t new. Author Neal Stephenson wrote about it in his science fiction novel “Snow Crash” in the early 1990s. Think of virtual and augmented reality, fueled by AI and machine learning, that simulates the world and all that’s in it.

From a manufacturer’s perspective, the metaverse could provide a predictive view of the entire enterprise and even its place in the broader market. The concept goes far beyond simulations that show what could happen under prescribed conditions. Most important, the metaverse adds the human element, incorporating customer relationships and how teams interact with each other and technology. In this way, the metaverse tackles how humans and technologies shape the world.

A key component is AI, a broad technology that makes machine learning possible. “At first, AI allowed humans to automate human work by telling machines what to do. The new AI is machine learning,where machines can now infer their own understanding from large datasets of examples and will let us do new things. It can also provide new insights into how humans and machines can cooperate [to achieve] goals.”

Much of his company’s work has focused on how humans and machines interact, as well as answering a fundamental question: What should humans do, and what should be left to the automation? What is the right balance to optimize outcomes?

After a stint at Lockheed Martin as director of its Virtual World Labs, Richard Boyd launched Tanjo, a company focused on artificial intelligence and machine learning, in 2013. In 2017, he co-founded Ultisim, a firm focused on digital twin technologies.

As AI has shown, automation can review and interpret a vast amount of data points in an instant, developing results no human ever could. Humans then can act on that data. Technology doesn’t diminish but instead amplifies the effectiveness of every human decision.

Boyd (whose Twitter handle is @metaversial) has been bullish about the metaverse’s potential for years. He’s worked with Hollywood studios and computer gaming companies, defense and aerospace manufacturers, various governments around the world, health care systems, oil and gas industry, universities, urban planners—the list goes on. His broad experience gives metal fabricators insight into the what’s happening in other sectors and how, just maybe, the technology one day could apply to the fab shop floor.

From Hospitals to War Zones

To illustrate, Boyd described digital twins developed for hospital systems. “In hospitals, you’ve got lots of equipment that’s configured on-site for different kinds of use cases. There’s not just one way. All kinds of problem-solving occurs in real time, and that’s where the human choreography comes into the picture. Here’s where the metaverse really becomes powerful, when you add the effects of human behavior into a system.

“Applying human behavior to the model is key,” Boyd continued. “Humans are notoriously unpredictable, but that’s where the challenge becomes interesting. At this point, we’ve done simulations for the Defense Department, and we’re simulating entire countries.”

Describing one effort, Boyd recalled a simulation that measured the effects of building a girls’ school in Helmand Province, Afghanistan. “Can you build a school for girls there? Sure you can, but how will people react? You can’t do this in a vacuum. Physically you can build it, but what are the second-, third-, or fourth-order effects? It’s about being able to model the values and traditions of all the people involved to avoid surprises.”

He added that such an approach could apply to engineering, purchasing, manufacturing, and supply chain management. What are the behavior elements that affect purchasing one material over another? How would this product design, using a component available only from certain suppliers, affect the supply chain and production resiliency?

“If you’re going to solve the problem, you need to solve the whole problem,” Boyd said. “And we’ve got the tools now to model and simulate that.”

Modeling Behavior

The metaverse’s comprehensiveness might make it still sound like science fiction. After all, you can model physical things; how that 16-ga. aluminum forms with this specific toolset; how these cutting variables affect this laser beam’s characteristics while cutting this profile in that material. But when talk of modeling people’s behaviors emerges, eyebrows start furrowing. That said, tracking and modeling behavior isn’t a new discipline. It’s helped build the big tech giants (especially the ones who rely on advertising) into what they are today.

“No matter how much you model them,” Boyd said, “humans are notoriously difficult to predict. We change behavioral states a lot. But we have disciplines around that,” adding that models have been used to build teams that go beyond just individual skills and delve into behavioral aspects—especially how humans interact with technology.

“We’re not just talking about ergonomics or a better HMI [human-machine interface], nor are we talking about adapting humans to technology. It’s a symbiotic thing, and it incorporates more than education, skills, and experience. If you’re a good leader, you’ll need to be paying attention to all of this.”

A New Perspective on Knowledge

From job shadowing to formal apprenticeship programs, manufacturing training has been rooted in some kind of show-and-tell training. The experienced teach the less experienced. Such an approach has a flaw, however.

“If a certain problem never occurs during the apprenticeship,” Boyd said, “then [the apprentice] never learns how to deal with it.”

Knowledge transfer in general can have an element of secrecy to it. Frontline personnel don’t document their processes, and managers don’t ask them to. Their results, they say, are what really matters. The same goes for other positions in the company, like a salesperson’s relationships with key clients.

“But that knowledge is critical,” Boyd said, “and we can encapsulate that in expert systems. It’s hard, and you get a lot of resistance, especially from people inside a company who have specialized knowledge. That’s job security. But if you’re a manufacturer, you can’t afford to have that person with that knowledge leave. They could retire. They could get hit by a truck. What then? We call this problem the silver tsunami. Ideally, systems are developed with their active cooperation, and they’re incentivized to [cooperate].”

Industry saw that resistance with the introduction of sales software, like Salesforce.com and other platforms. No salesperson wanted to log their activity, even though it might help the higher-ups make better decisions and forecasts. “The same applied years ago to CAD,” Boyd recalled. “When I was at an engineering firm, they tried to get draftspeople to stop using paper. And they fought it every step of the way, saying, ‘This is my specialized method.’”

Future generations might have a new perception of knowledge. Imagine a young machine operator entering the field decades from now. They learn from experienced workers, but augmenting that is a digital twin of the operation—not just of the machine in front of them but of the entire plant, incorporating the effects of human behavior. Someone doesn’t show up. The digital twin simulates how a cross-trained team would reorganize to keep production rolling.

Such simulated experiences run through problem after problem. The better information the digital twin has, the better and more powerful it becomes. Similar models could apply to engineering, purchasing, and even sales, analyzing behavioral patterns of the past to simulate possible outcomes.

As Boyd put it, “You no longer just ask, ‘Can I make it? Can I distribute it, and can I afford it? How is the product going to be used?’ Let’s envision it, understand it, and make sure we design human-machine systems so they function well together. Most people don’t think this way. Again, it can’t be about adapting humans to technology. It’s a symbiotic relationship. We need to ask how you integrate it all together, and you can’t discount the human part of it.”

Boyd described work done with school systems. “How do you shorten a path to mastery for kids in schools? We already know that they need good nutrition and financial and emotional support and stability at home to be a higher-performing student. You can’t solve that. But if you can detect it, you can come up with systems that try to augment it. We’ve got models now to simulate that and help train people.”

Those models reveal new ways for teachers to connect with students, for managers to connect with frontline employees, and for businesses seeking to integrate technology with teams in the office and on the floor. All this might shape people’s perception of knowledge and how it’s shared.

Sure, this might all sound as if it will never reach the typical fab shop. But as Boyd described, such modeling and simulation are real now, be it for health care, education, and, yes, certain areas of manufacturing.

One day, it might shape how even the smallest fabricator operates. Today, ask someone how they learned to perform a certain complex task—be it uncovering an innovative bend sequence on a press brake or an estimate for a big job—and they often say, “Through experience.”

Future generations might think differently as technology continues to quantify and predict reality. The better companies capture and use that knowledge through modeling and simulation—be it the details of a certain weld or the relationship between a salesperson and an important client—the more powerful each decision a person makes will become.