After years of hype, the digital factory—the comprehensive integration of data from development, production and suppliers via new hardware and software meant to increase efficiency—is gradually becoming a reality. This is partly due to nimbler robotics and simulation, whose emergence has made the otherwise nebulous concept of the digital factory tangible for manufacturers, according to several practitioners.
Exactly what this eventual wave of digitization means for the manufacturing labor market is not yet clear. What is clear is that in the short term there will be growing pains and potentially big gains in efficiency and innovation for manufacturers that take the plunge into the digital factory.
“We believe the factory of the future is a factory in which you can sit in your home and look at a factory and control production, change products… do setups and get analytical data while getting your final product out,” said Ashish Patwardhan, a segment marketing manager at Schneider Electric.
That could all be done, he added, using a smartphone.
That future may be closer than some imagine, he said, noting a basic prerequisite for the digital factory—the ability of machines to talk to each other—is already possible through open standards like PackML for packaging.
“It’s always a step-by-step process,” Patwardhan said of the evolution to the digital factory. “I think it has more to do with capital investment and possibility rather than technology. That being said, it’s happening today with robotics and integration of software, and integration of ERP [enterprise resource planning] systems.”
Nevertheless, while the digital factory is for the most part technically feasible the barriers preventing many companies from realizing it are manmade, several analysts agreed.
Some manufacturers are deeply rooted in their analog ways, Douglas Ortiz, a process simulation expert at Siemens, said. “Some companies are going to find [solace] in that, that this is how they do it, this is how it works, regardless of whether it’s expensive and cumbersome.”
But this wall of resistance will gradually crumble as manufacturers realize that the price “to maintain the status quo becomes too costly to not take up the new technology,” he predicted.
Robotics, simulation will be drivers
While robots have been around for years, they are becoming more integral to manufacturing in ways that were hard to imagine just a few years ago. Now, collaborative robots, or cobots, can work alongside humans on the factory floor. At Automation in Chicago in April, one roamed the show floor, and another, with a soft outer “skin” inviting touch, stopped operating when it sensed it was coming into contact with a human.
James Stettler, an analyst at Barclays Capital, estimates that the global cobot market will explode to a value of $3 billion in 2020 from $100 million in 2015. In the meantime, last year was a record year for robotics orders and shipments in North America, according to the Robotic Industries Association. In 2016, 34,606 robots valued at about $1.9 billion were shipped, a 10 percent increase in units over the previous year, the association said.
Festo Corp., a pneumatic and electrical automation firm, is increasingly turning its sights on robotics, according to Nuzha Yakoob, the firm’s senior product manager for electric automation.
At Germany’s Hannover Messe trade show in April, Festo unveiled a new cobot product aimed at making assembly lines more “modular,” or capable of using small automation cells to perform specific tasks. Such modular production is one way to prepare for “mass customization,” in which manufacturers can produce custom-made goods at scale relatively cheaply, Yakoob said.
“Instead of investing heavily in one big machine or complex machine that is designed for a certain product, now you’re looking at building modular cells, or small automation cells, that do very specific tasks,” she said, adding, “they’re actually far less complex” than their predecessors.
Robots can also be a crucial component of the Industrial Internet of Things, said Khris Kammer, a commercial programs manager at Rockwell Automation. “They have a tremendous amount of data that can come out of them and that can be used for improvement efforts that, up until now, haven’t been done,” he said.
Rockwell has a close partnership with Japanese robotics giant FANUC Corp. that has produced high-tech gear for the automotive industry that combines FANUC’s robotics and Rockwell’s programmable automation controllers.
Robots “are some of the most data-rich things that are out there on the plant floor, so it’s a huge part of what we’re doing,” Kammer said.
Simulating the future
Another building block of the digital factory is simulation, including the ability to test-run a part through the manufacturing process without physically doing so. Like robotics, the concept has been around for decades but is now gaining traction, in part because of technological advances.
Siemens’s SIMIT software for factory simulation has existed in one form or another for over two decades, Ortiz said. What began as an internal tool to validate Siemens’s energy products grew into software that helps customers train their operators on an offline simulator.
Today, SIMIT has “libraries” that correspond with digital control system components, such as motors and valves, Ortiz said. An operator can use the software to simulate what a drop in pressure on a valve might look like and how the factory would react.
“You’re freeing up the resources to do more work, better work, for you, as opposed to just remedial checking and verification that the data was entered correctly in 30 different spots,” Ortiz said.
Schneider Electric’s Patwardhan echoed that point by saying that the digital factory will not necessarily mean fewer jobs for workers but different, higher-skilled jobs working with robots or implementing software. Just as the automotive industry replaced human welders with machines and found other jobs for people on the factory floor, the manufacturing labor market will stay robust while implementing robotics and other advances, he predicted.
Rockwell Automation’s manufacturing clients are warming to the “digital twin,” or using a digitized version of a physical factory asset to analyze its performance, Kammer said. “We’ve seen huge benefits coming out of applying this kind of technology at the operational phase.”
One such client, a container manufacturer in the food and beverage industry, has been able to produce at a high volume by using the digital twin to, for example, compare the time between repairs for machines, he said. “The digital twin concept has been huge for them because they’ve actually been able to do benchmarking of equipment versus equipment,” he added, declining to name the client.
Real-time maintenance to stay healthy
Predictive maintenance, or the ability to monitor the health of a machine to know the optimal time to take it offline for repairs, is also part and parcel to the digital factory.
Predictive maintenance is taking off in part because companies are exploring how to install processors to collect performance data over time, Yakoob said.
To make the most of more sensors coming online, factories can tap into cloud computing and detailed analytics, she said. That, in turn, could precipitate the next phase of smart maintenance—“prescriptive maintenance”—or taking proactive measures to reduce, or even eliminate, machine failures. But the industry doesn’t yet have the bandwidth for this, she added.
“At the moment, we don’t have this data,” Yakoob said. “We haven’t collected this much data in the past. You need to collect at all levels within the machine, from so many sensors, to come up with a pattern, and then from that pattern create these smart algorithms.”
While technical hurdles remain, the barriers along the path to the digital factory are mostly manmade.
Yakoob said that a lack of awareness of the current state of Industry 4.0, the broader concept of automated manufacturing that includes cloud computing, is preventing many manufacturers from taking advantage of the phenomenon’s benefits.
“I think there’s a lot of work to be done in that area; some people don’t even understand the concept” of Industry 4.0, she said. “For those who do understand, some have the perception that this is something in the future, that they’re not really ready to implement, which isn’t correct.”
While a handful of the top manufacturers have made waves in digitization, the great majority of shops still haven’t taken the prerequisite step of getting real-time data from the factory floor, Memex CEO David McPhail said.
“I think we’re a way’s off from SMEs [small- and medium-sized enterprises] being able to digitize their factory… given that the adoption curve is still pretty shallow,” he said. “Right now, I think what we need to do is make a call to action for manufacturers to at least get their assets connected.”
Smoothing the runway
Big automation companies are trying to make the leap to the digital factory for SMEs less daunting by making their technology easier to use.
Siemens, for example, has sought to make its simulation software more appealing to smaller shops through easier functionality, according to Ortiz. “We’re doing this to help out those smaller companies get by without those additional technological resources,” he said.
Rockwell Automation is trying to make the data streams that come with connected machines easier for factories to process, Kammer said.
Rockwell is exploring how to use machine learning and predictive analytics to reduce companies’ “reliance on having a data scientist on staff or a PhD that knows how to do regression analysis,” he added.
This means making analytics available to the user out of the box, which he said Rockwell does through the use of standard Ethernet.
Patwardhan made a similar point in saying that Schneider Electric’s products are built around machine-to-machine communication. Web servers are ubiquitous on the company’s programmable logic controllers and human-machine interfaces, he said, allowing for remote access to the gear. That connectivity lowers the threshold of human expertise needed to explore the digital factory.
The jump to the digital factory can be a puzzle for big multinational companies, too, Patwardhan pointed out.
An industrial giant like Coke or General Mills has numerous factories, and designing a strategy to digitize them all can be immensely time consuming for an industry short on time, he said.
“Do you change all of your machinery? Can you keep some of your machinery?” Patwardhan added, listing some of the strategic considerations. “Investigating all of that and actually implementing it takes time.”
Cybersecurity concerns addressed
Another up-front concern for companies big and small that are considering digitizing their assets is cybersecurity. The rise of high-profile hacking incidents that have ravaged the intellectual property of multinational companies has been a cautionary tale for all.
To assuage this concern, Kammer said Rockwell Automation worked with Cisco to draw up a “Converged Plantwide Ethernet Design and Implementation Guide” that helps companies securely connect their assets.
For Yakoob, cybersecurity concerns are one of the main hurdles to greater adoption of the digital factory. For all their benefits, connected devices can open up new security vulnerabilities to companies if not properly managed.
Despite these challenges, practitioners like Kammer and Patwardhan predicted a fast rate of factory digitization in the next few years.
“We’re seeing an inflection point now,” Kammer said. “The level of interest has increased by an order of magnitude, I would say, over the last couple of years.”
Referring to “the factory of the future,” in which managers can monitor factory assets and control production remotely, Patwardhan said, “I think we’re getting there and we’re getting there quicker and faster than we imagined.”
“Because of this connection of software, of robotics, it’s easier and easier to get faster production, scalable production and individual production,” he added. “And that is where we see the world moving towards.”