This article discusses the concept of industrial Internet, which is considered as a union of software and big machines. John Deere has created a system that combines embedded sensors, GPS antennas, and cellular communications technology to provide equipment operations data to anyone with an Internet connection. The system can even enable dealers to forecast impending maintenance issues and schedule repairs before there is a breakdown. Linking sensors to a network gives owners a way to gather data like vibration, voltage, sound, temperature, or electrical output from machines. This type of analog data, previously inaccessible or hard to measure, lets owners monitor machine health. When analog information is collected in the cloud, the supervisory control and data acquisition (SCADA)-type scale grows exponentially and so do its uses. Industrial Internet-type applications take a supervisory control and data acquisition system to the next level.
Most people have a romantic idea of farmers and their farm equipment. Farm tractors, for instance, are solidly built but a little old-fashioned, the kind a thing a farmer muscles into following straight lines and keeps running with elbow grease.
Anyone who believes that would have a hard time comprehending the kind of tractors John Deere & Co. sells today. Just as individual farms have grown to encompass entire square miles, farm equipment is now made up of large and complex pieces of industrial machinery. A single large tractor for sale by John Deere can sell for more than $400,000.
With so much money tied up in machines, farmers want them to work at peak efficiency. And so John Deere has created a system that combines embedded sensors, GPS antennas, and cellular communications technology to provide equipment operations data to anyone with an Internet connection. The system can even enable dealers to forecast impending maintenance issues and schedule repairs before there is a breakdown.
Linking sensors to a network gives owners a way to gather data like vibration, voltage, sound, temperature, or electrical output from machines. This type of analog data, previously inaccessible or hard to measure, lets owners monitor machine health.
John Deere is just one of several companies taking advantage of advances in data acquisition, transmission, and analysis to enhance their control of operations. They have created systems that link a network, sensors, and software. The result is a way to capture large amounts of analog information, organize it, and return it in bite-size chunks for business benefit.
There are different names for systems of this sort. National Instruments uses the term “Big Data Analog.” General Electric calls it the industrial Internet. The goal for all is the same: to protect the investment that customers have made in their large industrial machines.
For example, a system has been measuring turbine vibration at several Duke Energy power plants for nearly two years. The system, built by National Instruments, collects vibration information, digitizes it, and analyzes it to predict when a turbine will fail or need maintenance.
According to Tom Bradicich, research and development fellow at National Instruments, “That converts surprise outages into planned outages. If you know your equipment will fail next Wednesday, you can plan for it with backups.”
Jon Bruner, an editor at large for O’Reilly Radar, an analysis and research firm in Sebastopol, Calif., has published Industrial Internet, The Machines Are Talking, a 52-page report that looks at characteristics of this melding of technologies.
On the report’s home page, Bruner writes: “The big machines that define modern life—cars, airplanes, furnaces, and so forth—have become exquisitely efficient, safe, and responsive over the last century through constant mechanical refinement. But mechanical refinement has its limits, and there are enormous improvements to be wrung out of the way that big machines are operated: an efficient furnace is still wasteful if it heats a building that no one is using; a safe car is still dangerous in the hands of a bad driver.
“It is this challenge that the industrial Internet promises to address by layering smart software on top of machines.”
When analog information is collected in the cloud, the SCADA-type scale grows exponentially and so do its uses.
ON BEYOND SCADA
According to Bradicich at NI, the systems give industries a way to gather analog data such as vibration, voltage, sound, temperature, or electrical output from machines and other physical devices. This type of analog data had been previously inaccessible or hard to measure.
Industrial Internet-type applications, Bradicich says, take a supervisory control and data acquisition system to the next level. They move SCADA beyond a single manufacturing fl oor or a single use in the field. When analog information can be collected in a cloud system or a private network, the SCADA-type scale grows exponentially and so do its uses, he said.
John Magee, chief marketing offi cer at General Electric, says savings can be found by monitoring and analyzing data from long-lived assets like generators, turbines, and aircraft in order to find ways to increase productivity and save money by cutting fuel costs or storing energy.
One of GE's industrial Internet customers is First Wind, which owns 16 wind farms in America. First Wind has recently added more sensors to its turbines to measure temperature, wind speeds, and the location and pitch of blades. Software collects and analyzes that data, enabling the operation of each turbine to be tweaked moment-by-moment, in response to that data, Magee said. In the winter, the analysis might show that the turbine blades are icing up, and corrective measures, such as speeding the turbine up or changing the pitch of the blades, could be taken.
It's clear why networking systems are catching on, O'Reilly's Bruner says. “The cost of, say, a sensor is immaterial compared to the cost of the jet engine,” he said.
“The big driver is on the data end: broadband communication, faster embedded processors, and new software tools for handling really large amounts of data have made a big difference. I also think managers in heavy industry have observed the impact of Big Data analytics on other businesses, like advertising, and have seen the value of a similar approach applied to machine data.”
Bruner pointed out, for instance, that jet engines have always had sensors in them, but the information has usually just been registered on flight-deck readouts. Now, in addition to registering in the flight deck, the data from those sensors can be analyzed along with data from thousands of other jet engines to detect maintenance needs, tweak flight plans, and inform future engine design.
“Same idea with controls: microcontrollers in big machines aren’t new, but now they’re networked and accessible to software that can optimize entire systems of machines by adjusting their parameters in real time,” Bruner said.
Bradicich attributes the interest in harnessing analog information at least in part to the falling cost of sensors as well as to their increasing portability and capability to communicate wirelessly. Sensors can now be placed in large scale on airport runways or train rails to monitor them for problems, he said.
One example of this is a network of sensors deployed by the Union Pacific Railroad. The railroad has had acoustic and visual sensors in place on the underside of each rail car for decades, but according to a statement from the company, now that those sensors are networked, the railroad can call upon software to analyze sensor data to look for track and wheel problems that will likely to arise in the near future.
John Deere is pursuing new uses for smart machines and expanded analytical power.
The John Deere JDLink system—available on agriculture machines produced from 2011—currently enables farmers to track fuel use and machine performance. A controller collects machine information gathered by sensors and transmits it to a server.
The private information is available via the JDLink website, accessible from any Internet accessible location like a desktop computer or a smartphone, said Chris Batdorf, a mechanical engineer turned John Deere production marketing manager.
In addition to tracking and analyzing machine and fuel use, a dealer can—with a farmer’s permission—remotely diagnose upcoming machine issues, he added.
And John Deere itself is making use of the JDLink-returned information, in anonymous aggregate.
“Say we recognize an anomaly of higher temperatures on our machines in the Southeast; from that our engineers can pull that data apart and say that’s all related to the tolerance between the fan and the cooling package on that machine,” Batdorf said. “So maybe we could be proactive and make running changes to our products and to the next generation product.”
The technology can be retrofitted to machines produced as far back as 2009.
This spring, the company plans to enable customers to store crop production data like moisture and yield via wireless data transfer, Batdorf said. Farmers will be able to approve viewers who may access that information. So, based on the crop data, an agronomist could make recommendations about how much fertilizer a farmer should spread.
“When you add a network connection to a machine you undoubtedly create a new security risk, but that risk can be outweighed at a higher level by the security benefits of remote monitoring and control.”
Because harvest data and production data are proprietary, John Deere has built a robust, specific database and repository for crop data downloaded, Batdorf said. The information is made anonymous before John Deere aggregates it to inform decisions about future equipment design.
“We could make changes to our machines based on trends that are happening,” Batdorf said. “If our customers are doing 20-inch row spacing rather than 30-inch spacing we can offer more 20-inch planters for combines.”
Most information on the industrial Internet, including that collected by the John Deere program, sits in the so-called cloud: offsite servers on which a company rents or buys space. Given that the information exists on servers— even if a business uses its own private network—how secure is it?
According to Bruner, industrial security can be easier to address than general web security. Industrial networks tend to be highly structured systems with regular data-transmission patterns, which means security software is better at detecting intrusions and unusual activity.
“When you add a network connection to a machine you undoubtedly create a new security risk, but that risk can be outweighed at a higher level by the security benefits of remote monitoring and control,” Bruner said. “A gas valve located in an isolated area is safer with a network connection than without one because operators can see when there’s a problem.”
Bruner acknowledged that builders must always consider data security when constructing an industrial Internet analog system.
“That said, it’s usually more secure to provide the right channels for data to move through than it is to provide no channels at all,” he added. “Businesses need to move large amounts of data around, and dedicated systems for collecting and analyzing it can keep track of where it’s going and manage access much better than the sorts of ad-hoc systems, like e-mailing datasets as attachments, that would otherwise arise.”
“A gas valve located in an isolated area is safer with a network connection than without one because operators can see when there’s a problem.”
Bruner sums up the industrial Internet in his report: “The industrial Internet is this union of software and big machines—what you might think of as the enterprise Internet of Things, operating under the demanding requirements of systems that have lives and expensive equipment at stake. It promises to bring the key characteristics of the Web—modularity, abstraction, software above the level of a single device—to demanding physical settings, letting innovators break down big problems, solve them in small pieces, and then stitch together their solutions.”
While it’s not actually an Internet for machines, systems like the industrial Internet do give machines a voice. Heeding what they say can save a company money.