Who says you can't teach an old factory new tricks?
The term smart factory likely brings to mind a shiny new manufacturing building with end-to-end, modern, connected technology. However, an old factory that integrates modern technology with select old machinery can enable data-driven decisions about processes, capabilities that are at the heart of what is meant by smart. That's a central lesson to learn about creating a smart factory roadmap, one that's espoused by several experts who closely observe the trend of digital transformation, automation and information sharing in manufacturing.
To them, a factory doesn't require the newest, best or even the most technology to be smart. Instead, any level of digital connectedness is a smart step toward understanding precisely how manufacturing processes work, which, in turn, helps companies make the best decisions.
"It's all about connecting data to make smart decisions to improve manufacturing," said Thomas Hedberg, co-leader of National Institute of Standards and Technology's Smart Manufacturing Systems Test Bed and the project leader of NIST's Digital Thread for Smart Manufacturing project. Hedberg is suggesting even old factories can learn new digital tricks.
Smart factories vs. digital manufacturing vs. Industry 4.0
While smart factory is a relatively new term, it's tied to the concept of digital manufacturing, whose use can be pinpointed to at least 40 years ago when manufacturers digitized fabrication, according to Hedberg. "Because of Moore's Law, it's computing that's developing and enabling even more digital activities in manufacturing," he said. "Manufacturing has been digital for a long time, but now, there's just more computing power."
The terms smart factory, digital manufacturing and even Industry 4.0 (a synonym for the fourth industrial revolution) are interchangeable, according to Paul Miller, senior analyst at Forrester. An organization's preferred term typically reflects its marketing strategy, he added. In the end, however, the name itself doesn't matter, Miller said. What matters is that the factory technology connects.
Debbie Krupitzer of consulting firm Capgemini has so much conviction that digital connectedness will make this modern industrial revolution transformative for businesses that her confidence can be found in her job title: North America Industry 4.0 lead. With this fourth revolution, Krupitzer said it's not about achieving a concept that would be labeled a "smart factory," but rather it's simply a matter of connecting existing and new manufacturing technology, which can include AI, automated technologies or IoT.
Smart factory roadmap requires assessing your strengths
Regardless of the chosen technology, a manufacturer must first have a smart factory roadmap to see how it will all connect. When Krupitzer's clients say they want their roadmaps to include IoT, she stops the conversation and instead helps the company first assess its current manufacturing processes and technology. "We never first say, 'Plug in IoT,'" she said. "You have to look at the landscape and find the gaps. You have to see where they're collecting data from and apply technology as needed." The technology is just an enabler. In fact, Krupitzer said adding the newest technology to manufacturing for the sake of having new technology will only cause problems. People are still needed to create a proper business strategy for the technology, she said.
Krupitzer advises clients to "strengthen the muscles that are already there" by improving the existing data flow between technologies. "You might not be as connected as you think, but you might have small pockets of excellence," she said. "So, we draw a [smart factory] roadmap for that. For example, a company might have [programmable logic controllers] and data in the cloud. We want to know what's going on in those machines. We'll ask how the company knows when the machines are about to break down. Is there any predictive learning? To me, that's where you want to be. Predictive leads to prescriptive."
The level of investment in a smart factory depends on budget, obviously, and digital technology can be expensive, Krupitzer said, especially if a manufacturer has spent little or nothing on upgrades over the years and needs to connect its infrastructure. Once a manufacturer has streamlined existing data flows and gained insight on how to improve current manufacturing processes, then it can slowly make new technological investments that will further digitize operations, she said. "There can be incremental ways to make it happen on a budget," she said.
Smart factory roadmap can include IoT for old machines
Forrester's Miller similarly advises companies to not rush into getting smart, largely because many of them have old manufacturing equipment that wouldn't mesh well with IoT, 3D platforms and other new technologies. The reality of Industry 4.0 and digital manufacturing is making old technology work alongside new technology, he said.
Paul MillerSenior analyst, Forrester
"We have theoretical conversations about if everything was connected and digital. Well, in certain new factories, you can connect everything, but in most settings, the industrial equipment is 10, 20, 30 years old," Miller said. "A lot of these industrial machines have been putting out some sort of electronic signal, but connecting it with the modern analytical platforms, that's a tough job to do."
Indeed, sometimes, the only step manufacturers can take with old machinery is to indirectly connect it to monitoring software with a simple sensor. Miller said don't underestimate just how intelligent that step is. "A smart factory can simply be prototype sensors measuring things," he said. For example, Miller said that one relatively inexpensive investment is adding a sensor that gauges noise and vibration on a metal milling machine. This can detect whether the equipment needs maintenance.
Manufacturers can also take a page from Munich Airport, which found a workaround for measuring the performance of oil, gas and water machinery that had old dials that wouldn't connect to a digital platform. "They put small, cheap Huawei webcams in front of the meters, which sent a feed to a pattern recognition web service and that translated the dials into a digital meter reading that could be analyzed," he said.
Of course, not all manufacturers have to worry about meshing old machinery with new smart technology. Many big-name companies have the capital to build smart factories from scratch.
According to Capgemini, Audi invested $1.3 billion in a new smart plant in Mexico that offers centralized production control, smart logistics and electronic quality control processes, while GE spent more than $200 million on a flexible "brilliant factory" in India, enabling the company to produce diverse products for four different subsidiaries under one roof.
Krupitzer of Capgemini also pointed to Tesla for being "superautomated," Ford for being "very mature" in digital manufacturing and Cisco for having "super smart factories doing all of the right stuff."
Smart factory roadmap to Industry 4.0 rests on data
NIST works with manufacturers to solve real-world problems, from small issues such as how to repair a machine to large systematic ones like finding the right data to improve product design and quality. Eliminating the hurdles that prevent companies from creating smart factories is one such initiative, Hedberg said. For instance, NIST has advised manufacturers to not discard machines just because they have less computing power than a calculator. The data from, say, a decades-old, three-axis rate table that's used to test electro-optical systems can effectively detail which of its parts are wearing once it's connected to a diagnostics program, he said. While a manufacturer might be tempted to stop using the rate table because it appears to be slowing down, making it smart could show it still has about half of its life left.
So, while Audi attempts to optimize auto manufacturing with a $1.3 billion investment, manufacturers with a lot less to spend don't necessarily have to shoot for the moon just to glean insight to improve how they make products. Although digital manufacturing and smart factories have, as Hedberg said, actually been around longer than widely assumed, they are still maturing.
"No one has 100% of the full picture yet," Hedberg said. "You could even say there are things we don't know yet. We have a good handle on it but not completely. There are still significant hurdles, even for big companies, and that includes distributing all of their data." No matter how much money a company spends on a smart factory, it still has to coalesce data from databases all over the world, he said. "It's still hard to align those viewpoints."