Within the manufacturing industry, innovation in the sensor and control arena arises out of newfound abilities to implement features of existing sensors rather than astounding vibranium-based evolutions in sensor technology.
At this point, only a modest number of manufacturing companies utilize the advances that Industry 4.0 has delivered. Often, data is neither collected nor utilized, which means companies are not achieving optimum quality, nor maximizing profitability.
However, there’s industrywide respect for what’s happening at, quite literally, the sharp end of the spear. Sensory data flows into myriad channels: hyper local, factory net, and cloud. There, the dilemma of data emerges. Can it and should it be used? How should it be used? What does it mean?
Making things begins with raw commodities, and rendering materials into products typically involves a tight embrace in an age-old process relatively unchanged until the advent of numerical control. “When I first started, we went from simple numerical control to what CNC is today, with people looking to get more information quicker to predict the outcome of future production and manufacturing processes,” reminisced Larry Robbins, president, commercial division of Wheeling, Illinois-based SMW Autoblok Corp.
Meanwhile, he noted, sensors can provide instantaneous feedback and the ability to make accurate decisions in real time with precise parameter observance. “The fact you can go all the way from just jaw and part position in the chuck itself all the way to temperatures and individual clamp force, the limitation is in the customer’s imagination.”
With the ability to monitor what John Cosic, director of sales and marketing in North America for French supplier Digital Way Inc., relates as lower costs for data acquisition, the question becomes that data’s value. “We’re seeing more information being collected about the process and state of the part, vibrations occurring on the machine, or power being consumed, cutting forces, temperature—and lord knows what may be available on the machine tool bus. So, a very important question is: Is all this information useful? And, even if it is, who will analyze it?”
Until those two questions are answered, Cosic opined, the true benefits of Industry 4.0 technology will not be fully realized.
SMW Autoblok’s products have evolved into devices capable of measuring pressures at each contact point and Coriolis forces. These are not the three-jaw, connected-to-a-treadle-or-belt of yore, but non-contact devices able to pass up to 1.5 kW of energy, as well as data, while clamping with great force. According to Robbins, the company built, and continues to build, mechanical, hydraulic-pneumatic, and manual workholders—and now digital or mechatronic and totally electric workholders—that perform “non-contact transfer of power while simultaneously transmitting data back and forth.”
Over the life of a mold, castings or forgings previously might have varied by 0.007-0.008” (0.18-0.20 mm). “Now without any true mechanical intervention, I can (evaluate) my part, and, because each jaw has its own drive system, reset that part and take what would’ve been a bad part and make it good from the first operation,” Cosic said.
“Previously, a customer would make a part, inspect it and it would either pass or fail inspection, and depending on the complexity of that part and the operations performed, they could have more invested into the part in process (costs) than they did in material cost and tooling,” he continued. “So now I’ve made a part that’s $10,000 in value and it’s no good. With electronics, I can give you predictability and be flexible enough that I can offset what used to be a bad casting or forging and make a good part.”
SMW Autoblok uses similar non-contact technology to ensure that off-highway machinery couples to the correct tool (bucket, scraper, etc.) and even precisely measures temperatures for bagging just-baked bread—sliding into plastic sleeves without melting the sleeve.
Digital Way develops non-contact measuring products that provide tool-condition monitoring or machine-conditioning monitoring. “Beyond data collection, we process and make intelligent decisions about what that data is saying, on the machine or in the cell, based on the machine tool bus,” Cosic said, explaining how the company’s “Cyber Machine” aggregates data locally.
Data is available however customers choose, he added. “Our customers have discovered that you can also use these tools as process optimization tools; they’re windows into the process and that ends up being the really big bang for the buck. If they can shave seconds off cycle times or extend tool life and deliver improved part quality, there’s a lot of money to be saved.”
For Digital Way, the focus has been on sensor quality to establish process limits. “(Our sensors) are relatively easy to set up and maintain in the sense that good process looks vastly different from bad process,” Cosic said. With the increase in data, both low- and high-speed, to encompass a process, it becomes difficult to understand supplied data when hundreds of machines are connected to a factory network.
“We can collect data within a cell at very high speed and store it for weeks or months without bogging down the network,” Cosic remarked. “It remains available in an open-source format, and an external device can specify exactly what data it wants, filtered by type, time, or any specific machine. Our sense is that if you really want to fully characterize your data you have to commit to local storage, and there’s not a lot of people doing that.”
In fact, though changes to the factory floor are significant, much is happening at the edge—at on-machine nodes. Companies are discovering and innovating ways to pull data that previously had been available and put it to work. Such cutting-edge solutions can yield better production, improved quality, lessened part rejection, worker job security, and, ultimately, greater profitability.
“In the last couple of years we have been creating innovative software solutions to allow greater access to the native diagnostic capabilities of today’s smart instrumentation,” said Jack Roushey, manager of field product marketing for Siemens AG. His comments underscore the view that hardware has evolved minimally over the last decade, with instrumentation firmware/software and remote monitoring apps adding flexibility to information access regardless of location, and with user interfaces (UI) that no longer require any knowledge of machine language.
“Bandwidth and memory have expanded, (and) the speed of transferred information has grown, delivering data in real time,” Roushey, an avowed “instrument guy,” noted. “Formerly your control room might have concentrated on processes control, not the health of the instrumentation. But with this additional memory, this additional bandwidth, you have the wherewithal within the control room to start monitoring for predictive maintenance.”
Such information, which has often gone unused, requires an interface capable of acquisition and interpretation. Many, if not most facilities, lack capital for massive infrastructure change and pursue economical options to use their native data, Roushey said.
“With the development of Siemens SITRANS store IQ and serve IQ monitoring apps, we’re able to grab that information, translate it over IP-type networks or up into cloud-based networks, and give our customers the ability to either monitor what they’re doing or even to add some basic control capabilities that can fill the gap until resources for the addition of PLCs (programmable logic controller) become available.”
For example, a smartphone app was created that permitted monitoring and delivering alerts based on condition and parametric changes, ringing the phone for such user-defined critical changes.
Process instrumentation has come a long way since pressure transmitters gave way to force balance using chips and strain gauges. More recently, the biggest change has been the use of increased memory and diagnostics in instrument electronics, according to Roushey, that enable relatively simple predictive software and far more powerful A.I.-based tools.
Ultrasonic-level measurement, for example, is a great tool but operating temperatures could impact its performance metrics. Internal microprocessors could compensate and gather historical data that was seldom used. Knowing a sensor’s tolerance to heat could, predictively, suggest replacement based on the number of over-heat instances or their duration.
“In pursuit of greater productivity, manufacturing is being transformed through a new world of data-centric digital factories,” added Fiona Treacy, senior director, industrial automation marketing at Wilmington, Mass.-based Analog Devices Inc. “For us and our customers, ensuring data from the sensors deployed across the factory floor is transmitted reliably in a harsh environment and in real time is a major priority and design challenge, as is the balance between power and space within tiny sensor housings.”
Francis Richt, manager of Sandvik Coromant US’ digital machining business, has a similar view, noting that customers find digital solutions challenging.
“They are using digital solutions such as CAM (computer-aided manufacturing) and so forth. Customers are buying machines, but software solutions and sensorized cutting tools are something brand new.”
Manufacturers need to understand the value of digital solutions, and a new go-to-market approach for the hardware side, Richt stressed. “We need to do business in a different way than we’ve done in the past; how we transport electronics and batteries is very different from carbide cutting tools.”
The demand for process security and component quality also is affected by a growing skills gap. “Operators, technicians, and engineers who know the look, the feel, the smell of metal cutting, as well as chip formation, quality surfaces, surface finish—all those elements of manufacturing—are now difficult to find,” Richt said. In his opinion, conversations should avoid the term “automation,” which may be misinterpreted as terminations. He prefers discussions of how digital support decisions may assist the shop floor.
Sandvik provides internal turning tools that could be used in a bore, landing gear, or an engine shaft for a jet engine, possibly obscured by flowing coolant and inside a machine. “Sensors bring you valuable information, which should deliver process security (such as stability and a good surface finish).”
While machine-tool builders rule the data, Sandvik develops software to visualize that data and automate some actions. “Coroplus, which integrates software algorithms and the PLC, can alter machine operations to prevent damaging the component,” Richt said. “For example, a landing gear, an engine shaft in aerospace, all extremely expensive components. So by having this, what we call stop and retract functionality, we make sure our customers save the component and possibly save the machine from being underutilized, stopped, and serviced.”
A European Siemens data scientist and data engineer, Konstantin Schmidt, resides in the world of A.I.-driven data. Sandvik has created a “Lighthouse” initiative to help customers piece together various disconnected technologies including metrology, robotics, and digital tool management. According to Schmidt, whose expertise is in digitalization, A.I., and other data-driven topics, “Even small- and medium-sized companies are diving deeply into automation.”
Thus, all control or production steps are now connected not only to the process—that data is logged, yielding control over each product. Note that versus big data, there is also smart data as storage costs are also important.
Companies such as Siemens offer an array of PLCs, and customers chose products based on reliability, connectivity, data ownership, and cyber security. Early adopters can dig deeper into their data, build their own apps, and may have more flexibility in connectivity. In addition to horizontal communication within production lines, Schmidt said there is “a growing interest in vertical communications so that you can go onto private or even public cloud layers and make use of it there. If you want to boil it down to one word, it would be ‘interoperability.’”
Others emphasized data privacy and security, which must balance with interoperability among differing application programming interfaces (API) to prevent being locked into a single vendor, either in the cloud or even the shop floor. Though manufacturing tends to be conservative, digitalization’s competitive benefits are driving adoption.
Industry 4.0 is a continuance of the Digital Revolution delivered by early post-mainframe microcomputers running programming languages such as CP/M (control program for microcomputers). Modern data is now delivered via graphical interfaces, often running on tablets.
“The one thing you do need is the ability to tie back to the machine control, because you have to be able to predict something and then have an action that’s created by that prediction,” Autoblok’s Robbins explained. “So, the ability to interface into an existing machine control is our biggest limitation.
“Originally, when we developed this technology, it was only going to be able to be integrated into new machine tools,” he continued. “However, within our organization we said, ‘Wait a second, all we need is a chip from (the existing vendor’s) control and we can integrate into any age machine.’ As long as we have the right generation chip, we can feed back to that machine control in a language that it will understand, digest, and act upon.”
In Robbins’ view, Industry 4.0 reflects the Internet of Things and connectability between a machine’s control and any components it is using. Such systems may be cloud based or centrally located elsewhere with edge computing, the direction which he sees most of the machine tools and manufacturing processes moving toward. The goal is instantaneously taking advantage of the information available. “You don’t have to wait until it processes, post-processes offline, and then find out what improvement or negativity may be from that result. You can make that judgment on the fly and keep moving forward.”
ADI assists in sensor and actuator integration IO-Link connectivity, which Treacy said provides a lower barrier to deploying key sensing technology for digital factories. “From a connectivity perspective, one of the primary challenges in sensor integration is ensuring that the sensors are connected and communicating effectively with the rest of the factory’s systems, which requires careful consideration of the communication protocols, network infrastructure, and data storage requirements.”
Cosic added, “Heck, you might not even have a person standing in front of a line of machines on a regular basis. Without the machine or the control being able to assess part quality or machine quality in real time it’s a dangerous situation, you open yourself up to making a lot of bad parts—and not knowing until they exit the cell.”
Sandvik’s Richt concurred. “Industry 4.0 is about building an ecosystem between different partners, different solutions, and being able to share data is key. Machine-tool builders see this as an opportunity, but they also realize that they can’t do everything themselves as this type of technology development is expensive.”
A deep data dive can provide a better understanding of quality improvement and process stabilization. “We can optimize machine utilization so that we get the most of out of our high-end machinery, apply predictive maintenance, and turn unplanned down times into planned down times. In other words, process understanding and process monitoring or observation,” Schmidt commented.
Such vast amounts of data require strict attention to detail, which Schmidt suggested is best achieved via model building and A.I. algorithms. Teaming edge computing and the cloud enables a better, more comprehensive understanding of data and the mathematics of the algorithms that Sandvik can apply to its own software.
With modeling, Schmidt said his business unit was able to identify what test should be applied—and at what point and volume—for optimum quality and minimal part rejection.
Treacy suggested that, “Reducing downtime is a second key element. Through combining data from multiple sources including deployed sensors with embedded A.I. engines, individualized prediction models of production devices (or even entire facilities) can be built and leveraged to proactively identify deviations in performance data and other metrics. The goal is the convergence of all digital networks into one unified factory network. This enables the connection of data sources with state-of-the-art digital tools to deliver greater efficiencies and productivity through more flexible production systems, more optimized manufacturing processes, increased energy efficiency, and reduced raw material usage and waste, driving a more sustainable production process and potential increased profitability.”
Using data, whether at the edge, in the shop, facility, or cloud and evaluated by sophisticated models through A.I. seems inevitable, albeit expensive. And everyone throughout an organization needs to be on the same page.
“You have to have buy-in extending from the boardroom to the shop floor,” Robbins said, noting that line workers may be the most critical—otherwise they may fear their job is being replaced. “Making people understand that this is going to solidify their position and give them the ability to be more valuable to their organization is a necessity.”
People need to understand that technology is evolving on a daily basis, he added. This will help make the best use of advanced sensors and controllers, while driving continuous improvement throughout an organization.
Connect With Us
