From Pervasive Sensing to the Cloud, Change Is Coming Quickly
by Paul Studebaker
 |
| "We will be using business parameters as the feedforward variables to run the plant, and the controls will provide feedback and constraints." Among the coming changes Walt Boyes discussed are shifts in the very control strategies we use across industry. |
It's an interesting time in our industry, and not an entirely happy one," said Walt Boyes, principal, Spitzer & Boyes, to open his keynote speech today at the Yokogawa Users Group Conference and Exhibition in Houston. "Changes are coming; changes are here, and all at once. Our traditional slow movement in adopting new technologies is breaking down."
The cloud, mobility, sensing, millennials, big data and analytics, the Internet of things (IoT), apps, social interaction, analysis, control technology, work practices and preferences are all coming at once. It's going to affect how plants are built and operated, both new plants and brownfield plants, "and it's not going to be unpainful," said Boyes.
The process industry workforce is going through its biggest change in history.
The current generation went from manual and single-loop control to distributed controls and advanced control. "We can walk through a plant and hear a problem, but that situational awareness is going away. The panel men who could look, see a problem and go fix it are going away," Boyes said. The current generation has never operated a plant in manual mode, and a modern plant can't be run in manual. They require readouts and alarms to tell what's wrong, but they don't know what to do about it. We can't give them enough training to know what to do in an upset, Boyes said. "To them, the plant is a magic box. They use the tools, but they don't understand the controls or the process."
Millennials think differently. They see a mouse, keyboard and display as limiting. They say, "Where's the tablet? Can't I just use my phone? You mean I have to type stuff in?" They want mobility, not to be trapped in a control room. "We had the opportunity to learn by doing; they want to learn by doing, too, to replace our situational awareness model with one of their own," Boyes said.
The Flattened Enterprise
Boyes predicts that the enterprise hierarchy will be flattened to two layers—a real-time layer of production and control, and a transactional layer of supply, enterprise and distribution. He said, "We will be using business parameters as the feed-forward variables to run the plant, and the controls will provide feedback and constraints."
Today, we operate the whole plant in open-loop control—the operator is an open loop. When the plant drops out of an operating state, you have an uncontrolled failover. We depend on the safety instrumented system [SIS ] to handle this. That's like relying on the emergency brake every time you bring a train into a station. We can make the control system smart enough to handle that 90% of the time."
Sensors and Analyzers Everywhere
Sensors are increasingly simple, less costly, wireless and able to measure many parameters, to perform control and to calibrate themselves. "We do single-loop control because Foxboro couldn't figure out how to multiplex a pneumatic controller, so they gave us only the most important parameter," Boyes said. "We don't have to do it that way. Now we can measure temperatures continuously all the way up and down a fractionating tower, and control on them."
Boyes predicts that there will be no analog outputs. "It will be all digital, and wireless will take over because wires cost too much, and wireless is just as reliable," he said. "Analyzers will be online, and their outputs used as control variables."
Big Data, the Cloud and IoT
Aggregation of data will revolutionize control. "Equipment will wake up, see if it has a problem, and if it can't fix it itself, write and send its own work order," Boyes said. "Maintenance and control people will work to optimize, not gather and analyze data."
The industrial Internet of Things (IoT) will do for sensory data what the PC did for control. A modern DCS can run on commercial, off-the-shelf (COTS ) PCs, and industrial control can run on COTS networks, Boyes said. "COTS networks will have to be more robust than industrial networks, because they're designed for housewives and kids."
New businesses will develop cloud databases of publicly available information, useful for refining operations and control systems—things like weather and injury statistics. "These knowledge bases are already here," Boyes said. "Google is trying to put together an oracle. Putting all these facts in one database allows you to start connecting the dots: Ask a question, get a fact, with a factor indicating its level of confidence."
The process industries kill hundreds and injure thousands of people every year, usually during start-ups, shutdowns and upsets. "It's not just the big-boom plants; it's also food and pharmaceutical facilities," Boyes said. Worker and asset safety must and will improve. "We need procedurecontrolled automation, smarter loops and better safety systems," he added. "With pervasive sensing and wireless networks, we'll know where everyone is so we can go and pull them out. We'll even know where we parked the fire truck."
Apps to Modularize Control and HMI
"Apps will fit together like Legos. You'll be able to pull the ones you need and the ones each employee needs," Boyes predicted. They'll be easy to program and upgrade, so you won't have to upgrade the whole system.
Human-machine interface (HMI) design will be decoupled from the control system. "Today, Southern Company has eight different control systems in its plants, each with its own HMI," Boyes said. "It's developing an HMI template overlay that gives any one of the systems the same look and feel as any other, so an operator trained on one system can work in any plant.
"We'll have people running around the plant in Google glasses with tablets in two to five years. We'll have to change how we classify hazardous areas to accommodate them, and it's not a problem, because with fugitive emission laws, most hazardous areas already aren't."
With more sensors, more data, much more complex models and information in real time, "We have virtual augmented reality control systems," Boyes said. "Operators can walk through the plant virtually within the control system, point at a pump and see what's going on with it. The cost of this technology is dropping every day.
"Yes, it's a security nightmare. We have to work out a security design for an entirely new paradigm. The rapid pace of change means only automation suppliers will be able to stay current on the technology, and they'll increasingly become the main automation contractors."
The proliferation of technology and productivity will put many people out of work, "just as horses could not be trained to become cars or tractors," Boyes concluded. But that's a subject for another story.