Get to know the engineers creating the next industrial revolution — and how they got their jobs.
In the world’s largest ketchup processing plant, a robot fires a continuous stream of freshly picked tomatoes across the factory floor using compressed air. A plethora of cameras make minute observations of every tomato as they fly by, checking for ripeness and damage. As soon as a defective tomato is identified, another robot fires a precision blast of air at it, unerringly knocking it out of the stream and into a separate hopper. At the other end of the factory, the finished bottles of ketchup are packaged up and placed on pallets by autonomous forklift trucks, 24 hours a day, seven days a week.
No humans are involved.
Every industry is affected by automation. It’s not just cars that are being made by robots now. Our power plants are becoming increasingly automated. Our food is grown and processed in automated farms, storage units, and factories. Buildings are prefabricated by machines. In the modern factory, smart machines talk to each other, and notify their human masters when they need attention.
In this industrial revolution, most of the human workforce is simply no longer necessary, replaced by increasingly sophisticated robots, more advanced sensors, and a more robust internet.
“What’s happening is that rules-based human activity is going away,” says Kevin Paveglio, who runs the mechatronics degree program at ECPI College of Technology in Virginia. “Low skilled jobs are disappearing. We don’t need people putting things in boxes any more. We’re taking out human elements because of safety & quality concerns. Cars now last longer, because they’re put together better. Twenty years ago, seventy thousand miles on a car was a lot, but now they’re just broken in at that age. The companies who automate are doing much better in the economy. Simply put, for manufacturing companies, it’s do or die.”
Faster, better, cheaper… automation offers all three. Machines are poised to take a huge proportion of the burden of manual labor from us. Whether that heralds a utopia or dystopia remains to be seen: are we ushering in an era of mass unemployment or increased leisure? But as the Luddites showed 200 years ago when they failed to destroy the machines that changed the textile industry, the technology itself is unstoppable. Society will have to adapt, not just in America, but worldwide.
The Day of the Machines is here.
Machines Are Just Plain Better
It says a lot about our powers of ingenuity that we invent machines which can outperform us in almost every way, from tomato screening to diagnosing cancer. The advantages to adopting automation are undeniable.
Prescient is a manufacturing and technology company that creates steel frame structures for hotels, dormitories and multi-family housing. They utilize robotics extensively to make the process more efficient, faster, and cheaper. “In the process of the design we incorporate standardization which allows us to design through proprietary components, like Lego sets. That leads us to a lot of opportunities to streamline the manufacturing because it’s all about repetition. We can easily incorporate elements like robotic welding,” explains Michael Lastowski, Founder and CTO.
Director of Technology, Rick Barrett, explains the process. “With this simplified product design we have panels, posts and trusses. Holes are drilled automatically, not by hand, then the panels are welded in a fully robotic system in a controlled environment. This gives us far more accuracy on the construction of the components, typically tolerance of less than a 16th inch on something like a hallway. We also get a far higher consistency of weld. The finished products still requires visual inspection, and some manual welding to correct errors, but at least 75% of the work is done by the machines.”
“Our fabrication plants are working 24 hours a day,” says Lastowski. “And we just need equipment operators, not a full staff. The commercial advantages are huge. Normally, it would take a month to construct the components for a 100,000 square foot 5-story hotel. We can do it in a week, for less money. And the faster that hotel is up, the sooner the owner is making revenue. Everybody wins.”
Paveglio echoes the improved quality that automation brings. “The Boeing 727 used to have around 2000lb of shims to keep it together when it was hand built. Now it has just 200lb of shims because there’s more precision.”
While fully robotic plants get all the attention, one underrated aspect of automation is men and machines working together. Professor Thomas Kurfess at Georgia Tech, an ASME fellow and former advisor to President Obama on manufacturing technologies, calls these systems “co-bots”. Robots are great for moving big things around, but they have their limitations. Often, they’re most effective as part of a man-machine team.
We know what you’re thinking now – will we ever see Alien-style exoskeletons? “Oh yes, full exoskeletons are happening,” Kurfess says. “There’s lots of work going on, but realistically, we’re not likely to see them in the near future. First we will get robotic arms or similar devices, that makes use of human perception, motor skills and decision making, combined with robot strength. If you’re putting an engine into a car, the human can align it while the robot carries the weight. In surgery, it works the other way round, normal size movements translate to micro movements.”
Cheaper, Safer, More Productive
However, working with robots has its own set of problems. Kurfess emphasizes the need to think about the manufacturing environment. “Most automated production facilities are mostly robotic welding & assembling. That’s not a healthy environment for humans, so robots and humans have to be physically separated: if a human gets in the robot area, everything shuts down. Allowing humans to work in proximity to robots can change the way we interact. Vision and other sensory systems track the human and put a virtual box around them to prevent robots getting in. Robot controllers are fast enough to sense if the robot is coming in contact with something, and react to that contact in real-time, so the robot can safely be around humans.”
But speed, cost and quality aren’t the only benefits that automation brings. Machines are making workplaces safer, operating in environments that are hazardous to humans, or detecting potential problems long before they become dangerous.
Mark Spindler, CTO at Lakeland Grain, develops and installs fully automated grain terminals. “As we move the grain, it’s dumped out of a truck into a pit, up a bucket elevator, onto a conveyor, and into the bin. This generates lots of grain dust, which is very explosive. When a belt gets out of alignment, they can heat up and cause an explosion. It’s a major problem for the industry. We still get 10-12 explosions a year across the US, some of which are fatal. Existing systems monitor the temperature every few minutes, and will sound an alarm to warn the human operator that there’s an issue. The problem is that a drifting belt can make it game over in mere seconds. Our technology detects when something is wrong and automatically shuts down the equipment instantly.”
This level of predictive maintenance saves lives and money. Operating on a “best practice” routine inevitably means unnecessary downtime while equipment is inspected for wear or damage, and expensive parts are often replaced before it’s strictly necessary. Factories need to keep stocks of components on hand, so that they’re always available, which locks up operating capital and storage space. And when a vital piece of equipment fails unexpectedly, it can delay production throughout the entire plant.
“It’s all about data collection,” says Kurfess. “Imagine a smart copy machine. When toner gets low, it sends a message to the company and they send new toner, so I never run out, it just shows up when I need it. Now apply that logic to machine tools. What if my tools can tell me which inserts are wearing out, so they can get replaced just in time, so we’re not spending money on inventory sitting on shelves? That improves our supply chains as well as productivity.”
Improved data collection also gives manufacturers much more flexibility in the way they operate. Once you know exactly what is going on within an individual machine or assembly plant, you can configure it in different ways. “You can choose to run your machines at high deterioration to maximize output, or you can opt to reduce wear to save costs,” explains Todd Walter, Chief Marketing Manager of Embedded Systems at National Instruments. “You can calculate exactly how much that will cost in in terms of additional maintenance and make a business decision about whether the additional revenue is worth it.”
The Industrial Internet of Things
The Internet of Things (IoT) may have been slow to catch on in the consumer world; not many of us have really seen the need for smart fridges or washing machines, but connectivity is changing everything in the industrial world. Traditional Internet may be adequate for homes and businesses, but for industry, the standards often need to be higher. “It’s taken a while for the Internet to really hit manufacturing, but for good reasons,” notes Industry Solutions Architect Paul Didier of Cisco. “If you have to wait 2 mins for your Amazon purchase, no problem, if you have to wait two minutes to shut down a power station, that can be catastrophic. They can’t move as fast as other businesses because of safety and regulations, but now technology is catching up with their needs.”
Bandwidth, latency, security and reliable connectivity can be major problems for many industries, whether they’re using ethernet or wireless. “We install systems in harsh environments such as oil & gas plants, transformer stations where there is a lot of metal and interference, or places where temperature or dust is an issue,” says Matt Nelson of AvaLAN Wireless, who pioneered the use of 900 mHz wireless in manufacturing. “When you’re putting in connections to and from transformer stations, unintentional emissions cause havoc. We have to keep channels narrow, power high, and connectivity on.”
However, once you solve those fundamental engineering problems, the transformation in industry is astonishing. “You can upgrade your connected equipment the same way you upgrade your cellphone,” explains Dan Sexton of GE. “You can simply send a firmware update directly from the manufacturer and change the way something works.”
For Prescient, connectivity offers even more flexibility. “We have plants round the world. If one of them is taken offline by a storm or something, we can simply re-route production to another plant with literally a press of a button,” says Barrett.
At Lakeland Grain, IoT is a vital part of improving grain storage. “The machines talk to each other, which cuts out human error,” explains Spindler. “You can’t accidentally put the wrong thing in the wrong bin, which is an expensive mistake. We also have wireless probes deep in the stored grain, looking for the hot spots that signify potential spoilage, and communicating with the fans that adjust temperature and humidity to minimize shrinkage if the corn dries out. They’re constantly reporting back to a control panel that operators can access from a browser on their phones, tablets, or whatever, and they’ll send out alarms via text or email if there’s an issue. Without this, you don’t know there’s a problem until it’s too late, and your grain has gone bad. This way, you can easily increase your revenues by $250,000 in a season with a $15,000 system.”
Engineering in the New Industrial Revolution
Developing for this new industrial world requires a different mindset to much traditional engineering. Designing machinery for a modern factory requires far more than mechanical engineering: it requires familiarity with robotics, information technology, and process engineering. In a very important sense, manufacturing engineers are verging into artificial intelligence, much more than just physical construction of objects.
“There’s a huge skills gap,” says Didier. “We need maybe 10 million engineers over the next decade who can work with this kind of equipment. The level of automation we are looking at is ever expanding: it’s not enough to have a robot that can do the task. Now you have to be able to automate fault detection and preventive maintenance. You have to ask, what are the scenarios that tell me something is about to fail and fix it before it’s a problem? Effectively, you’re automating optimization. The key for engineers is to take your understanding of the physical world and the product and apply that knowledge into building a system. Robots should be self-configuring, if it fails, I should have all the information I need.”
Walter pinpoints one key part of the problem. “There is a large divide between the IT groups and the operations groups, and that’s one of the areas we need to get sorted to create understanding for this to work. A lot of the issues are due to policies within companies. The IT groups need to understand operational issues, just as the ops guys need to understand the IT issues. What we really need is engineers who have practical operational experience.”
Steve Hechtmann of Inductive Automation specializes in integrating industrial control systems. “When I go in to a client, I’d ask to operate the process for a day to get how it actually works and what really matters to them. You’re surrounded by millwrights, hydraulics guys, electricians, chemists, and so on. You don’t need to do be able to do their job, but you do need to understand that, if you’re the computer guy, you need to know what they’re doing and why. It’s very interdisciplinary, and that’s a lot of fun.”
“We started the mechatronics program because traditional four-year engineering degrees don’t give engineers the practical skills they need,” notes Paveglio. “Even physical maintenance is a skilled job. You still need a guy with a wrench and a can of grease and micrometers, but he needs new skills. Machines need adjusting to thousandths of an inch. You need someone who can recognize whether a problem is electrical, mechanical, software or some combination of them all. And those skills aren’t easy to teach. You need lots of hands-on experience, and you often need to develop simulators for training on. After all, you don’t want students to wreck a million-dollar machine! But if you can program a 6-axis robot, you can be earning 30k more than a guy with a normal engineering degree.”
The technical challenge facing engineers is huge. “You’ll be in the workforce for fifty years, but in a few years, things will completely change. Are you willing to go back and retrain?” asks Kurfess. “You need to change the way you think. For example, these days, we’re no longer telling a robot to get an object based on its spatial position. Instead, it uses vision systems to find something of the right shape. When you’re setting up a production line, you have to ask, how do I make it easy for the robot? For a person, I’d put screws in a box, for robots I’d put them in holes. You need to understand how robots ‘think’ and change your thinking to match.”
Just keeping up with latest developments is a time-consuming task. “The integrators who put these systems together have to know what’s already available,” says Paveglio. “In most cases, you can use components off the shelf. Often they need to be customized. The rare projects are completely designed from scratch, but even they use a lot of ready-made components and assemblies. The key is someone who knows what’s out there and how to bring them together as a system. You can source parts from Germany, Switzerland, Italy, America, or anywhere, and they’re all designed to IEEE standards so they can interact. Almost anything you can think of, someone out there has something that can do it. You could spend hundreds of hours just training on what sensors are available. Do you need a camera that can go inside a 2000 degree oven? You can get it if you know where to look.”
Some companies, such as Siemens and BMW, have already taken the initiative to start training the next generation of engineering workforce themselves. “The robot manufacturers need to teach people how to work with their kit,” Paveglio stresses. “CEOs are worried about what’s going to happen if they can’t buy the technical skills they need.”
Other industry-led consortia such as AVNU are partnering with universities to try and get the necessary skills into the education system. “At Cisco, we’re establishing partnerships with companies like Rockwell to train IT guys to get familiar with operational parts of the business,” notes Didier. We have to see some blending of engineering capacities: for example, mechanical engineers need to understand the core IT technologies so as they develop new products, they can get them into the IoT world and get things done faster. We’ve established the Industrial IP Advantage consortium to deliver free IT/OT training to cross train engineers. And we’ve just announced the Cisco Certified Network Assistant Industrial program which will set the standard for engineers in this field.”
Walter encourages both engineers and companies to do their part to change the industry. “Getting involved in AVNU is a good way to come up to speed on what’s happening and find ways to participate. There are also forums in IEEE. We are looking for smart people to jump in and help to create change.”
“Getting the right training and experience is a real problem,” confirms Hechtmann. “Young engineers coming into the field are more or less clueless about how to run plants and factories, because they don’t teach that. We set up Inductive University so that engineers can become credentialled in using our systems. We work with universities, department heads, etc to get practical skills to students. With this you can get a job anywhere in the world with an integrator.”
Interspecies Communication is Key
However, mastering a wide range of massively different technologies is only the start of the skills engineers need. Engineers also need to develop a deep understanding of the business aspect of manufacturing: not just cost, but the needs of the business to adapt and adopt flexible procedures.
“Ultimately, all those machines need to communicate with the people who are running the plant,” explains Bob Giese, President of Versacall. “On any production floor, there is a whole spectrum of activity, so how does the management know where to direct their attention? The answer, as we’ve proven with companies like Harley Davidson and 3M, is that the machines themselves need to provide that information. When you have good feedback, you can see where the issues are and take decisions on what’s happening. We can eliminate tedious and inaccurate manual data capture, and that gives us better auditing. We can address health and safety issues, get better quality products, and increase productivity and control. We can identify whether we are hitting production targets, monitor the amount of waste, and monitor cycle time. We’re typically getting a 10% reduction in downtime, which gives our clients an increase of 3% or more in production. The key is the ability to direct information to the correct individual.”
But merely providing swathes of data isn’t enough. It has to be presented in a form that managers and line personnel can understand, and act on. “Understanding of the manufacturing process and presenting it in visual format, on anything from large screens to phones and tablets, is where technology is at,” stresses Giese. “Knowing that is what makes a difference from a job candidate standpoint.”
Walter concurs. “From a skill set perspective, the level of complexity that this type of system adds is substantial, and all of this needs to be abstracted from the end user. If you look under the covers of, say, a cell phone, think about what a good job the developers have done of making it easy for the users. That’s what we need to do in manufacturing. Operators want to optimize production of what they do, not be experts in the systems that hold it all together.”
The 21st Century Engineer
Barrett has reassuring words for engineers. “Faced with this level of automated manufacturing, the skilled labor workforce often gets nervous about job security. Perhaps surprisingly, it’s the opposite: we’ve grown faster and hired more people.”
There are a growing number of roles for engineers in this new world, and demand has never been higher for those with the right skills. At the root of the business, there’s a vast need for individual components, from sensors or servos to routers. The next layer of engineers creates standard assemblies or machines, from smart machine tools to highly sophisticated robots. Then come the integrators, who design the plants and factories, and finally, the on-site operators need the skills to maintain and run them.
Traditional engineering skills, including CAD/CAM, math, and all the specialist disciplines are still essential. Almost every branch of engineering is affected, from agricultural to architectural, aerospace and automotive. The challenge facing most engineers now is not merely whether we can build something, but how efficiently we can manufacture it. And to do that, the key, according to everyone we spoke to, is cross-disciplinary collaboration, adaptability, and creativity.
As Hechtmann says, your task is to breathe life into machines.
Just try to avoid creating Skynet by accident.
Featured Image Credit Dan Ruscoe, CC BY