Terms such as smart factory, Industry 4.0 and Industrial Internet of Things (IIoT) have become inescapable buzzwords, invoked by every developer of manufacturing-related equipment and software. Many discussions of smart factories begin with prognosticators sharing visions of how the global manufacturing industry will be transformed.
These descriptions often lump in burgeoning technologies such as additive manufacturing, 3D visual simulation and collaborative robotics. All this can seem out of reach for the average job shop—and far too expensive to even contemplate. Who really needs it?
That popular industrywide view of smart manufacturing is, unintentionally, misleading—akin to answering the question, “What is a motor car?” by describing the Interstate highway system. In both cases, the core idea is much simpler—and the answer to the question of who really needs smart manufacturing is, emphatically, the average job shop—and every manufacturer that wants to remain competitive in the future.
A useful way to understand the smart factory is to start with a comparatively simple technology: the inexpensive electronic sensor. In the pre-sensor world, the machine operator programs a CNC lathe or mill to cut a part and the machine does its best to follow the instructions and remove metal. The communication is all one way, from the operator to the machine.
Did the machine accurately follow orders? Barring an obvious disaster such as a crash, the manufacturer won’t know unless he inspects the part after the operation is completed—which takes time and costs money.
Enter the sensor. Modern sensors strategically placed in the machining center can allow two-way communication. The sensors allow the machine to record or communicate information about its own status and the status of the operation while it works. Vibration measurement, temperature variation, motor current analysis and other factors can be shared over a network for monitoring and analysis. The status of the machine, and how that status affects the operation, can be assessed quickly—even in real time.
Where the sensor meets the machine is where the rubber meets the road in the IIoT. New production equipment is generally equipped with appropriate sensors, but they can be installed in older equipment at relatively low cost.
Sensor-enabled machines have a lot to say—more than a user may know what to do with. But analytics software can crunch the data and use it to improve the operation, maintain the machine, improve the product design—even improve the design of the machining equipment itself.
This same two-way communication approach has been scaled up to make smart manufacturing cells, with automation and multiple machines. Zoom out farther and imagine an entire facility networked this way, with the manufacturing data accessible and usable across formerly separate silos—comparable to the way enterprise resource planning (ERP) data has become.
That enticing vision of a manufacturing facility in which the equipment can share data with people and other equipment to allow faster, higher-quality, continuously improving production of parts and products—is the smart factory.
But it can start with pulling sensor-derived data from a single machine and using it to improve its operation.
“Industry 4.0 is really about how you can use data in a way to better operate your business,” said Larry Megan. “And if you don’t have the data, then you don’t have a place to start.”
Megan is vice president of Advanced Manufacturing International (AMI), Clearwater, Fla., a not-for-profit with a mission of helping small to medium-sized manufacturers (SMMs) get started on this journey—the “digital transformation,” as it is known.
“The problems manufacturers are trying to solve are the same problems that people have been trying to solve for 100 years,” he pointed out: “‘How do I increase efficiency? How do I reduce energy usage? How do I maximize throughput? How do I address quality issues?’ What has changed is that we now have technology to quickly make visible the data that can supply the answers so it can be used as a jumping-off point for making better decisions.”
Megan and AMI’s work with SMMs often includes recommending the use of a device developed by LECS Energy called the LIMS Appliance.
LIMS, which stands for Low Investment Manufacturing System, is an unassuming little box consisting of a computer with proprietary Solution Engine software, an I/O (input/output) hub, and a 110-V power source that plugs into a standard outlet. When wired at the edge of a piece of production equipment, it becomes a simple solution for collecting and sharing complex sensor-derived data. (AMI will conduct live demonstrations of the LIMS system at HOUSTEX, EASTEC, SOUTHTEC and WESTEC in October and November. AMI is the exclusive distributor of the LIMS box.)
“For many manufacturers, especially in the discrete parts space—automotive, aerospace and others—we feel the LIMS solution is a great onboarding point,” Megan said. “It’s designed for manufacturers who are using a CNC or other machine but have no visibility around the productivity of that individual machine, or more broadly, what’s going on across the factory floor. The LIMS solution is a low-cost, easy way for folks to start to get data out of their processes and get it into a format that’s useful. And then they can start to make better decisions with it.”
In the words of the device’s primary developer, LECS Energy’s Nat Frampton, “LIMS was designed to try to help operators and engineers be able to touch their process, to be able to analyze and come up with a new understanding of their process—to be able to collect that data, look at it historically, to interact or move the data to databases, and finally to be able to get the results from that analysis and improve their process.”
According to Frampton, LIMS is a product of 20 years of development, originally designed around the manufacturing of explosives for the U.S. Army. A key goal from the start was for users without specialized programming skills to be able to install and operate it.
“If you understand your own operation, if you can work with your equipment, you already have all the computer skills you’re going to need in order to be able to configure this appliance,” Frampton said.
LIMS is capable of a dizzying array of tasks, but it can be simplified to a five-step application flow. The first step is connecting the system to a given piece of production equipment. How that connection works depends on the equipment being connected to and the shop’s networking setup—if it has one. LIMS is designed to connect in the most efficient way possible, depending on those variables.
“LIMS is based on open standards,” said Frampton. “So, if it’s connecting to a relatively new machine that uses a standard data protocol such as MT Connect, then it’s very easy for anyone to pick up the data off of the network and move it to a database. However, if the machine isn’t using standard data protocols, but the data is available on the local PLC [programmable logic controller] that’s running the machine, then typically, if it’s relatively new, we can get the data off of the PLC. We understand the PLC protocols and have drivers that are able to do that.”
The LIMS appliance “speaks” over 50 different industrial protocols, he pointed out. “Just know if you have any piece of equipment that ‘speaks,’ we probably can talk to it.”
And as mentioned, if none of those things work, sensors can be added to the production machine and wired to the LIMS device, said Megan.
“We’ve had cases where we’ve added sensors to enable certain types of analyses—temperature, humidity in the shop, those kinds of things. Those can be wired in as typical wired signals as would be going into a PLC or anything else,” he said.
Whichever way the connection is made, the benefit is that “now you’re going to land everything through one system so that you’ll timestamp everything right through,” Megan said. “Everything will be consistent in time. You’ll be able land it in a standard database and have the data well organized, in context and in one place that you can use.”
Once the connections are made and data can flow to the LIMS, the second step is filtering the data to reduce noise and strengthen signal. “The data coming in off the equipment isn’t always perfect,” Frampton said. “It may be noisy.” For that reason, the Solution Engine program includes a simple analytics engine that smooths it. For example, “maybe you have a current that’s coming in and you want to integrate it into kilowatt hours, so you know what your carbon footprint is for a particular part.” The filtering will enable a clearer, more accurate data set. This initial data filtering is different from the more complex analytics possible once the system has been in place to record and store data over a longer period of time.
The third step is collecting the data and saving it to the user’s database of choice, which may be local or on the cloud.
“No cloud is required,” Frampton asserted. The user could choose to set up an SQL database, for example, and “you could put that database in the cloud if you wanted to—but you don’t have to. You can just store it locally on the LIMS device,” he noted. The LIMS device has 16 gigabytes of storage standard.
Another point that distinguishes the system from others: “One of the things we don’t do is monitor the data usage and charge you for every time you store some data or give you tag limits or anything like that. The data is yours,” Frampton said.
The fourth step is report generation and analysis from the data. As most machine operators are not necessarily experts on databases, LIMS is designed to make data collection and storage as easy as a mouse click.
“If you’re not a database jockey or are uncomfortable with databases, don’t worry about it,” Frampton said. “You can just right-click on that data and put it out to MS Excel. I’m a mechanical engineer by training, so I’ve always made sure our tools go straight out to Excel.”
However, if the user does have Azure or Google Cloud or another cloud-based system, “we can take that data and move it directly out to tables in Azure so you can see it in Power BI” or other such programs or formats, he said. “We can also look at machine learning and get results coming back in.”
The fifth and final step is the reason the system is put in place to begin with—using the gathered information to improve the process.
“Once you’ve figured out how to improve the process, the most important thing is to be able to control it—to do something about it,” Frampton explained.
The desired action might be relatively simple to begin with, such as making a pertinent piece of information visible with the expedience of a red-yellow-green light stack. In time, however, the collected data may enable the user to do much more.
“We position LIMS as an entry point for digital transformation,” Megan said. “It typically starts with this idea of collecting the data and being able to just do simple visualization. Once you have that, over time you can start to do more sophisticated things with it. You can walk yourselves up the maturity curve—up from basic analytics ultimately to more advanced solutions like AI.
“But it’s best to start simple—and LIMS is a great starting point,” he concluded.
Connect With Us
