It’s not just a concept—Industry 4.0 offers solutions manufacturers can put to work today
Industry 4.0 often seems like a vague, faraway idea, but there are several practical and relatively easy steps that can be taken to jump into the “future of manufacturing” right now. From addressing skilled labor shortages to increasing machine utilization, a fancy title should not stop a shop owner from using what works.
Let’s start with a huge concern in manufacturing today: attracting and keeping enough qualified workers. As Jason Ray, co-founder and CEO of Boston-based Paperless Parts expressed it, “It’s great to think about taking 15 different types of CNC machines purchased from different manufacturers at different times and plugging them into MTConnect to harvest Big Data. But the biggest problem shops are facing is they can’t get the second shift up and running. They can’t go hire someone new right out of trade school to come in and run their old equipment.”
Meanwhile experienced workers are fast retiring. Automation, as wonderful as it is, won’t solve this problem. What’s also needed, said Ray, is a way to “enable your smart people to work more effectively for you” and a way to bridge the skills gap between the experienced workers and new hires.
Happily, much of the technology rolling out under the Industry 4.0 umbrella addresses these needs.
As Jeff Rizzie, director of digital machining sales for Sandvik Coromant, Fair Lawn, N.J., put it: “Today, when they prepare new jobs, most shops rely on tribal knowledge. They ask an engineer to remember the last time they made something similar, and with that knowledge develop a new process or program [from scratch].” With digital machining, Sandvik Coromant can build system knowledge and not rely so much on that tribal memory. Performance data from machine tools can be fed back into the system to deliver programs and cutting tool recommendations.
Michael Walton, industry solutions executive (manufacturing) at Microsoft Corp., Redmond, Wash., said he advises companies of all sizes to move their business from “running as tribal knowledge and artisans to something that’s more akin to scientific.” But this transformation “assumes you can communicate and have a solution that allows you to be able to digitally capture people’s knowledge.”
Inexpensive sensors allow operators to capture much of the hidden knowledge. “I can take a $49 sensor on a CNC machine and I can coach an employee faster than if I had an employee that was a 30-year CNC specialist. I can tell them what they’re doing wrong and right faster with that $49 sensor,” he said. If a tool is pushing too hard, if there is too much torque, or if the machine is laboring and how long were they in cycle can all be captured and used to give feedback to coach the employee.
Virtual Reality for Training
Chicago-based R&D center and manufacturing test-bed MxD (Manufacturing x Digital) USA has partnered with dozens of firms on just such practical applications. CEO Chandra Brown pointed to the example of using virtual reality (VR) to train technicians to work on a Rolls-Royce helicopter engine. “Take the best person working on that engine, have them do all the procedures, put everything in, and then train every other person on that,” said Brown. The trainees wear a VR headset and “get to climb in the engine” without the health and safety risks that exist in the real world. “We’ve extracted the knowledge of the experienced techs and also made it fun to train the next generation of workers on the exact procedures,” said Brown.
Tony Del Sesto, MxD’s vice president of projects and engineering, said the hardware to enable this kind of training is available and inexpensive. In this case the VR headset is the HTC Vive, which retails for about $600. Del Sesto explained that MxD used VisionThree LLC, Indianapolis, to help create the content, starting with CAD drawings of the engine. “In some VR creations there is a latency between moving your head and what you see in the headset, and this can be bothersome and even induce nausea,” explained Del Sesto. “VisionThree modifies the CAD model such that what you see is virtually instantaneous with your movements, for a realistic and helpful experience.”
Walton said automotive manufacturers have begun using VR and augmented reality (AR) in their hiring. He referenced a large manufacturer that used video to capture people doing their jobs and then created a simulation based on that knowledge. Its prospective hires are now tested several times in the virtual world to confirm that they’ll be able to achieve the necessary skills to do the job in the required takt time.
Watson also pointed to the benefits of this technology for remote assistance. “That’s a real big one,” he said, “[using] virtual reality or artificial augmented reality to get a little window so they can see either pictures or tasks to repair something.”
Instead of going to a job site with lots of manuals, not knowing which ones are needed, a tech can don an AR headset like the Microsoft HoloLens and get guides to help do the troubleshooting, or even get remote coaching. The time required to do maintenance is greatly reduced and the process is “far more accurate,” said Watson.
Gaming Technology for Manufacturing
In another case, MxD used an Xbox Kinect gaming platform purchased for $100 on eBay to capture the motions of an assembly worker. From there, explained Del Sesto, they programmed guidance for other operators, such as telling them in real time when they reached for the wrong bin.
“You can study every motion,” said Brown. “What they’re doing. Who’s your fastest person. Everything is tracked, with all the timing. People are always asking me: ‘How can you go paperless? It’s not possible.’ Actually, it is possible and it’s pretty low cost. And the amount of information you can get from it is really pretty amazing. So I like to tell people that a great place to start is with digital work instructions right at the workbench. It’s better for health, for safety, and for measuring. And the folks that do it really like it.”
Del Sesto said that although the Xbox Kinect has been discontinued, Microsoft now offers powerful Azure Kinect sensors for a few hundred dollars, and industrial applications are a prime market. There’s an entire ecosystem of developers and service bureaus creating solutions based on this technology. Or as Watson put it, “Microsoft is democratizing artificial intelligence. It’s our mission…That means great news for you because things like custom vision, using sound, using other types of machine learning algorithms, they’re published and they’re free to use, other than [obtaining] the data.”
Turning Experience into Guidelines
Not all tribal knowledge can be captured with a video camera or digital sensor, however, and the information that is not susceptible to such capture is often critical. Take the process of estimating and quoting, for example. You’re not going to learn anything by putting a camera over the estimator’s desk. And, said Ray from Paperless Parts, “You’re not going to sit a 55-year-old machinist in front of a computer and have him take a survey of 70 questions and expect to get quality information come out of that. The biggest challenge in extracting this tribal knowledge is that it has to be a human-to-human relationship.”
Paperless Parts creates customized, cloud-based quoting systems for clients that speed up, and to some extent automate, the proposal process. But it starts with one of its implementation specialists spending several hours with the client’s experienced estimator, learning how they assess jobs and establishing “guardrails” for future quotes, which are then refined as the system develops.
As Ray explained, there are things an experienced estimator will look for when quoting a job, and it varies from company to company. For example, every time they see a particularly tight tolerance they might insist on someone in the engineering team verifying the quote before it’s sent. So Paperless would build that rule into the system and account for its cost. Or an estimator might look at a part drawing and figure a 10-minute cycle time per hole on each of the 10 holes in the part, which would suggest a possible guardrail based on cycle time per hole.
Once the system is fielded, it continues to query the
experienced estimators every time they override a rule or adjust pricing. “For example,” explained Ray, “if they changed the programming time on a part from two to three hours, we ask why. Sometimes the answer is because if they’re making 500 parts it’s worth it to spend the extra programming time to cut the cycle time. So we put quantity rules in place. Or maybe it’s because this part has so many uniquely sized and shaped features that it will require 40 different tools.” The Paperless system captures these deviations and increasingly refines the rules.
Ray added that one reason an experienced estimator is so critical to a business is because they’ve learned from past mistakes. “They’ve quoted and they’ve lost money. That’s experience. That’s driving down the road and hitting all the potholes along the way. What we’re trying to do is get that experienced estimator to put a cone on every one of those potholes, so that when someone else comes down that road behind them they can avoid them.” In doing so, Ray said his company’s software typically delivers an ROI of at least 5× (per monthly period) and “measurably increases top-line revenue by 5 percent or more, whether that’s winning more work because of quoting faster or more accurately, or getting a higher price for work that wasn’t quoted accurately before.” Pricing out sheet metal fabrication is especially formulaic, he said. “We have a new customer who six weeks after the on-boarding process is sending million-dollar quotes out without any editing.”
Ray also observed that many shop owners have not even mapped out their organizational structure and the responsibilities of their key people in detail. He suggested that shop owners should periodically ask the question, “‘How would anyone else in the organization be able to do this? Where would someone go to look this up if I was out sick?’ Take those into consideration and start thinking about how you can systematize that information and pass it down to the next generation,” he said.
Future Jobs and Skills
A related challenge is defining the jobs and skills that will be required in the factory of the future. An MxD report estimated that 65 percent of the jobs that will be held by Generation Z (people born in 1996 and after) have not yet been invented. Within the manufacturing world, MxD sees today’s operators, technicians and programmers needing to transition to new roles, including 165 new jobs they’ve laid out in their Digital Manufacturing and Design Roles Taxonomy. Brown said she laughed when she read some of these job titles. “Whoever thought we would need a digital ethicist in the future?”
It may shock some people, but Rizzie said Sandvik Coromant’s studies show that globally the utilization rate for the average machine tool is less than 30 percent. But the bigger problem, he said, “is not that the utilization rates are so low, it’s that most machine shops don’t realize they’re as low as they are. We have to first understand what it is before we can even go to work on improving it.” Until recently, he added, the cost of accurately calculating utilization was fairly high. “People were sent out on the shop floor with clipboards and stopwatches to gather information, and the data they gathered wasn’t particularly good.”
Happily a good digital network and today’s Industry 4.0 technology makes it downright easy to gather that information.
Step one is networking the machines, which Rizzie pointed out has already occurred in some shops without them thinking of it in terms of Industry 4.0. That’s because many shops have installed a DNC (distributed numerical control) solution that connects their machines to a central server. If not, it’s easy to do for any CNC machine built within the last 10 years or so, and still doable for older machines with a bit more effort.
Step two is implementing a communication protocol, like MTConnect, that extracts useful data from the machines. Once that’s done, it is possible to implement machine monitoring and analysis tools, such as with Sandvik Coromant’s Machining Insights, a cloud-based service that connects securely with a central server, pulls data 24/7, and presents it on demand. It shows real-time dashboards of the status of each machine, such as whether or not it’s running, what program is running, and the utilization rate of each machine during a shift. Shops can organize their machines in cells and get a quick look at the overall equipment efficiency of each cell and see if a particular machine is holding things up.
Beyond that, Rizzie said the “real power” is in the reports, which present a wealth of well-organized information covering selected time frames or subsets of the operation. For example, shops can pick a particular machine and see its utilization rate, jobs, open tickets, alerts, and operator comments, all in one place. A parts metrics report provides the ability to study any part ever produced in the shop, and does so according to various criteria.
For example, it creates a report based on cycle time, utilization rate, or quality. “So if I have a certain part that I run in my shop on multiple machines,” explained Rizzie, “I can instantaneously compare how that part runs on machine number one, against machine number two, against machine number three. And I can compare that on first shift, second shift, third shift. So now I can understand how the operator potentially affects how parts run in my shop.”
Shops can also look at factors such as setup time or tool breakage and get rich data on what’s actually happening with their machines, according to Rizzie. There is even a report that counts the time lost during shift changes, and another that tallies all the time operators run the machines at less than the programmed feed rate.
Sandvik Coromant itself found that operators at its North Carolina plant frequently overrode feed rate settings. Now, the system automatically texts shop supervisors when an operator lowers the feed rate more than 20 percent for more than three minutes. With training, operators quickly eliminated 85 percent of the problems. Now when such an alert occurs, it’s an anomaly and the engineer knows he needs to go out to fix a problem, because the operator has probably reduced the feed rate for the right reason, such as tool chatter.
Addressing Security Concerns
We shouldn’t leave this topic without a word of caution. Every machine sensor and every network connection is a potential leak of valuable information or an opening to a cyberattack. In fact, said Brown, manufacturing is the economic sector most frequently targeted by cybercriminals. As a result, a key part of MxD’s mission is to be the national center for cybersecurity in manufacturing. Brown said all of MxD’s big partners, like Microsoft, Deere & Co., and Caterpillar, worry about their supply chains. “The small and medium-sized businesses [are] where the vulnerability is going to be,” she said. The key is making sure you work with reputable, knowledgeable firms when setting up your network or connecting to the cloud.
At the same time, said Rizzie, understand the distinction between what’s important to protect and what isn’t. “One example I’ve used is the fear people have about putting machine monitoring solutions on the cloud. But what happens if somebody actually gets to your data from your machine monitoring? What have they really gotten? Essentially, what they’ll know is how often you run your machines. And to some extent who really cares? That’s important to you internally, but not to somebody else. Nobody else is going to do anything with it.”
Ray of Paperless Parts offered another bit of practical advice for those daunted by an investment in cybersecurity. “The best place to start is just talking about it with your team. Many phishing attacks come in via email and they’ll use someone’s name in the email and say, ‘Hey, can you remind me what the password is to our WiFi?’ Or ‘Hey, you mind sending me over the name of that customer with drawings from X?’ These attackers will come at very sophisticated levels and try to access your team. Talking about cybersecurity is a priority. Invest 30 minutes!”
When you break it down, Industry 4.0 isn’t as daunting as it appears to be: Leverage technology to improve the productivity of your personnel. Systematize key tasks and the relevant knowledge. And connect machines and implement a monitoring solution that improves and optimizes machine utilization.