What shall we call it? Cognitive manufacturing? The Industrial Internet of Things? Factory Digitization? Industry 4.0? Emerson Automation Solutions calls it the Plantweb Digital Ecosystem, and Bob Karschnia, Emerson’s VP/GM Wireless, says that while each of these terms has a slightly different meaning, in general all of them are trying to address the same question: “How can I improve the performance of my plant by getting more information in, analyzing that, and making better decisions as a result?” We talked to Emerson Automation Solutions’ Bob Karschnia about the technology and the future.
It starts with sensing
The first part is monitoring, which has been greatly improved by the growth of wireless sensing. Wireless sensor networks cost much less to install than wired systems, can be installed where wired sensors could not, and are easy to relocate. The classic example is vibration monitoring. In the old days a pump, say, might be checked for vibration once every shift. But what if something goes wrong in the middle of a shift? A wireless sensor can be set to report once an hour or oftener, so fewer problems are missed. “So just by monitoring those pumps people are able to significantly reduce their downtime,” says Karschnia. “They can reduce their maintenance spend and run their plants at higher capacity because they don’t have any slowdowns as a result of failures.”
Analytics
The second part is analytics: the use of big data to allow the system not only to alert operators but to draw its own conclusions as to where things are going and what needs to be done. This is the more complicated part, and the most work to implement. It also promises great rewards. “On some levels,” says Karschnia, “I think, the objective is, how can I take a one-off problem like pumps and deploy that philosophy across my plant.”
A look at the future
What will the technology look like in a decade or so? Predictions are risky, Karschnia points out, because the pace of technological advancement keeps increasing: “Things that ten or 15 years ago we weren’t even thinking about are commonplace today.” But he does see that the application of analytics will have a profound effect on industry, and that those analytics will be pushed further and further down the chain. As an example, he cites a recent project that involved steam traps. While acoustic detectors can spot a leaking steam trap, the data tends to be noisy, says Karschnia, “and trying to determine what’s really going on is hard, so we applied big data analytics to it and came up with solutions that gave clear, insightful answer on what to do.”
What about the people?
The spread of analytics and artificial intelligence through industry will require a change in the skill sets of the people involved — basically because it will make things easier. Computer users once had to know DOS prompts and command lines. Now elementary school children learn programming. “There’s an old saying that, technology accelerates until it becomes invisible,” says Karschnia, “and it becomes invisible when it’s easy to use and you don’t have to think about it.”
And the ease of application extends down to the lowest levels. At one time planning a wireless installation required lots of calculations, but today all that’s needed is to remember that in a crowded environment the sensor/transmitters should be perhaps 100 meters apart, but can be up to 300 meters apart where things are more spread out.
Selling it to upper-level management
Any significant change in a plant requires approval — and money — from above. And while denizens of the C Suite may have heard of cognitive manufacturing, says Karschnia: their expectations may have to be tempered. We can be talking about a significant investment here, he explains, and CEOs tend to want quick results: “There is some risk in overinflating the expectations of this with senior management, to the point where they become disillusioned with it and turn their attention somewhere else.”
What about security?
Older plants may lack security. “A lot of systems in place today are pretty old, and security wasn’t really considered a problem 20 or 30 years ago, Karschnia explains. “It’s a two-part problem: How do you put in some sort of defenses against old systems like that, and how do you develop new systems that have better defenses?”
Future systems will be better protected, especially with encryption. “An encrypted piece of data flows from the sensor into the control system, where it’s decrypted, a control algorithm runs, then that data is encrypted and stuck in the database. Everything is encrypted, and there’s no way to get at that data, because only the controllers and the sensors handle the encryption.”
Bob Karschnia, vice president and general manager, wireless at Emerson Automation Solutions
