INDUSTRY 4.0: The Smart Factory Arrives

The plants that produce vehicles and their high-tech systems are increasingly employing intelligent systems, Big Data and advanced analytics to improve quality, safety and efficiency.

December 1, 2016

Ryan Gehm

Human-centered Industry 4.0 is utilized in the production of TraXon transmissions for commercial vehicles in ZF’s Friedrichshafen plant. The assembly process is supported step-by-step by software via an integrated user interface.

A new “data-driven” manufacturing facility that Faurecia opened in Columbus, IN, in October 2016 embodies perfectly the premise and promise of the Industrial Internet of Things (IIoT): a smart environment that gathers vast amounts of information, seamlessly intermingles automated robotic vehicles with humans and ultimately ushers in the end of “dirty jobs.” The 400,000-ft² (37,160-m²), $64 million state-of-the-art plant will employ 450 people and produce a new, advanced emissions-control product for the commercial vehicle industry.

Senior process engineer Tim Esham analyzes robot activity at Faurecia’s new $64 million Columbus South digital manufacturing plant in Indiana.

One of those employees will be a full-time, on-site mathematician whose sole purpose is to help mine the terabytes of data being generated daily at the factory, cull insights and forecast issues before they occur. This job illustrates the shifting nature of the manufacturing landscape, according to Dave DeGraaf, president of Faurecia Emissions Control Technologies North America — to one that is clean, technologically advanced and proactive, and aimed at attracting employees with different and advanced skillsets.

“This facility represents our entry into Industry 4.0, a revolutionary concept incorporating connectivity, automation, data processing and hardware to advance the manufacturing industry,” said Mike Galarno, plant manager of Columbus South. The new plant employs, among other technologies, self-learning autonomous intelligent vehicles (AIVs) to transport components to the assembly line.

Automotive supplier Magna develops advanced driver assistance systems (ADAS) for vehicles, and many of these same technologies can be applied in factories, according to Ian Simmons, vice president of business development, corporate engineering and R&D at Magna International.

“Radar, multiple sensors, sensor fusion in vehicles, some of these you’ll find in intelligent guided vehicles within the manufacturing space,” he said. “And even how to interface with people with different sensors and biometrics.”

Cloud communication is another shared experience between cars and factories: “It’s already a challenge for the automotive industry in terms of product,” Simmons said. “If manufacturing is going to start using the Cloud to send huge amounts of data for analytics, it’s going to be a concern. Is there capacity to handle this traffic?”

Self-learning autonomous intelligent vehicles, or “cobots,” transport components to the assembly line at Faurecia’s new plant.

Trends come and go, but the shift to an advanced manufacturing environment is no mere trend, according to Stu Johnson, director of product marketing, Plex Systems.

“The difference here is this level of connectedness,” he said. “Sensors communicating data about machines to the Cloud, the analytics, connecting people [who] are the orchestrators through mobility and wearables, and ideally connecting the entire supply chain. The Cloud’s really the only scalable model for this kind of data. This indeed is the next industrial revolution.”

Big Data, predictive analytics

There’s a popular saying, “You can’t improve what you don’t measure.” By collecting data and connecting plants to the Cloud, companies theoretically have more avenues to improve their manufacturing operations in areas like preventive maintenance, logistics, inventory operations and material handling.

The key is turning Big Data into “smart data” to support decision-making and offer new insights, according to Magna’s Simmons. “The more data you have, the more you can measure, the more you can improve the quality of your parts and processes — and you’ve got end-to-end transparency of the process.”

The problem, he added, is that 80% of all data today is currently unstructured and therefore unusable.

“Big data: this is probably where the real revolution will come,” Simmons said. “With a lot of legacy equipment out there, it becomes difficult to get and really understand the data necessary for decision-making.” Sensors are key to making this a reality, as technology accelerates and costs drop.

Magna’s smart factory concept includes a glove a person would wear when picking parts out of multiple bins—the color coding on the glove confirms the correct part is being picked.

“There’s the potential to add sensors to legacy equipment to start to get the data you need,” he noted.

On the flipside, collecting all of this data creates a huge Big Data and analytics problem, according to Plex Systems’ Johnson. Cloud computing is the solution.

“Taking these huge streams of data up into the Cloud, using analytic platforms to look for predictive analytics like machine tool wear or machine vibration, and delivering it back to the machine to make better decisions about production quality — in more and more cases, autonomous decisions. Something as simple as stopping a machine or stopping a line if an SPC (statistical process control) chart is trending out of tolerance. This is the state of play,” Johnson said.

Not only gathering data, but applying it in a proactive way — that’s the goal for many manufacturing companies, including American Axle & Manufacturing (AAM). “We have a lot of data [but] we’re using it in a reactive way,” said Nigel Francis, vice president of advanced engineering & electrification systems at AAM.

He said his company is moving toward actively applying predictive analytics. “It’s more than experimental but it’s not scale at this stage,” Francis noted. “The goal is to basically have adaptive machining driven by Big Data, so you have reliability, repeatability and good quality as a matter of fact.”

Experts agree that companies need to run their own value analysis to determine, first, if they even have enough data to be meaningfully measured and, second, can turn that data into value by improving process, throughput, and efficiency.

“It’s not going to be a panacea for every problem that exists within a manufacturing organization,” said Simmons.

Cybersecurity jitters

GE’s Brilliant Factory demonstrates how the “digital thread” can provide end-to-end connectivity for operational acceleration.

While careful analysis and some caution are recommended, GE Digital’s head of manufacturing industries Paul Boris wants to make manufacturers “a little more nervous” than they might already be, stressing the need to move “fast, small, light” for the digital factory.

“We’ve got to get comfortable with the fact that the lean use of technology drives the continuous improvement cycle in operations,” he told a recent conference of automotive professionals. OEMs need to be willing “to deploy something quickly, consume it, drive the process, get some value, and then ‘pivot’ — maybe we discard some things we built, maybe we add on to some of those things, but we’ve got to move [forward].”

Boris likened it to the consumer-goods manufacturing space, in which the Apple iPhone 7, for example, becomes available — and highly desired — even though there’s absolutely nothing wrong with the iPhone 6.

GE practices what it preaches. Its FastWorks program takes a software-development approach to developing new products, with quick deliverables and fast learning. The program was the answer to a poignant question: “How do we do a project that’s large enough to have a tangible impact on process, but small enough so that if it’s an utter disaster we can bury it in the backyard and we all get to keep our jobs?” Boris said, to audience laughter.

Another issue making manufacturers more than a little nervous is cybersecurity. As connectivity and autonomous advance, in factories as in vehicles, this issue becomes paramount, experts agree.

“It’s become very public in terms of cybersecurity impact on vehicles,” said Simmons. “But imagine if you’ve got a manufacturing plant with millions of dollars of capital equipment being automated and suddenly you’re hacked, possibly causing safety issues, perhaps crashing the equipment. So we have to apply the standards and concerns over cybersecurity to the ‘factory of the future’ as you do in automobiles.”

Toyota has a large team dedicated to cybersecurity issues from a vehicle perspective, and that mindset is migrating to security on the plant floor, according to Trever White, divisional information officer, manufacturing & engineering business systems, information systems, Toyota Motor North America (TMNA).

“We have a group that’s focused just on security and with this IoT program [in factories], we have a dedicated member from our security team that’s working with us,” said White. “We’ve created new network and security architecture that’s focused on IoT, because there’s a lot of opportunity on the plant floor now for exposing data that we haven’t had before.”

Enter the “cobot”

As manufacturing facilities become more connected and highly automated, integration of the worker with production is critical — perhaps even more so than with less-advanced systems. Because no matter how intelligent the factory floor becomes, humans will always be involved in the process, experts say.

“You’re never going to replace your human associates on the line,” said Magna’s Simmons. Robots will supplement what the workforce is doing and become more “user-friendly” toward manufacturing staff, he said.

Collaborating robots, or “cobots,” that support the workforce are integral in Faurecia’s new Columbus South plant. “People have been an important part of the equation,” said DeGraaf. “Ultimately, these advanced technologies, like the AIVs and cobots, will enable employees to work more efficiently, experience less physical stress and improve work-life balance.”

The interface between human and robotics is key. “We need to have agile, flexible, gentle and intelligent means of manufacturing things. If you look into the future, the possibility — or rather the probability — of an intelligent machine needing to work in the same workspace as a human being is very, very high,” said AAM’s Francis. “That brings about a completely different thought process. We’re already thinking about this and hope to be working on it within our own facility very soon.”

Research is being conducted that allows robots to sense humans and avoid them to prevent injuries.

Augmented reality — whether wearable, visual, or hearing sensor system — can be used to improve quality and inform decision-making. Volvo, for example, has been testing Microsoft HoloLens technology — essentially a self-contained, holographic computer — on the manufacturing floor to support training, problem solving, etc.

“But challenges remain,” said Simmons, “namely, cost, socialization of workers with the HMI [human-machine interface] and training.”

GE is exploring “smart” lighting that facilitates communication in the factory. “Those LEDs can talk on a unique frequency of light that can register them on a device, so when an operator is standing with an iPad, I can tell you within 10 cm where that operator is standing,” Boris said. “Clearpath AGVs can actually listen on a signal and go around and collect data from all of the material units on the floor.”

Don’t interrupt me

This is a sampling of the possibilities. But the question looms: How do you implement new technology without interrupting production?

“We have a sub-minute takt time and all of our plants are running 100% plus capacity right now, so getting new technologies implemented in a factory is a very challenging thing to do,” said Toyota’s White. TMNA typically focuses on new plants or a new line to introduce the latest technology.

“From an existing [plant] perspective, we try to do a lot of trial activities and parallel implementations, where we allow the team members to work in the traditional process and also use new technologies — the iPhone/iPad applications, for example — until they feel comfortable,” he explained.

Toyota began its Advanced IT for Manufacturing program about three years ago to help modernize its traditional approach to manufacturing. “We saw an explosion of digital opportunities on the plant floor, so we started to lean in,” White said.

ZF also tries to implement the latest technology when building new lines and plants. “But more interesting is the potential to bring in new technology in a brownfield environment,” said Dr. Jürgen Sturm, CIO at ZF Friedrichshafen AG. “This is easily possible at low cost. This includes bringing in sensing technologies, [and] we also started early to have Raspberry Pi technology, things like this.”