Why do we do single loop control? I keep asking this question when I speak at user groups, at ISA section meetings and just in general conversation. I’m going to keep asking the question because it has a bearing on where we go in the future for loop and process control. It is very simple, really. Back in the 1920s, when Edgar Bristol, Sr., was inventing the very first loop controller, he was working with pneumatic instrumentation, and it was very difficult to get more than one pneumatic signal to be used in the controller.
Both vendor and end-user companies have long tried to produce advanced control systems and strategies that would carry us beyond the simple PID single loop that we’re so familiar with. There have been many versions of designs for automatic loop tuning, process optimization, neural networks, artificial intelligence and so-called soft sensors.
Until fairly recently, all of them have been add-ons to the basic control system, and all of them have been subject to the standard performance degradation curve. At first, everything works fine, and the expected benefits accrue.
But then processes change. Variables change. Systems and equipment change, and the advanced control software is not updated to keep pace with the process changes. Performance starts to fall off, and the plant operators and engineers begin to lose faith in the system.
Finally, the system performance degrades to the point that the operators basically shut it off. That’s what has happened in the past, but in the crystal ball we’re gazing into, that’s not likely to happen for much longer.
Three trends are converging that will make our dependence on single-loop control part of history. The first is the movement of advanced control algorithms and strategies to the field controller. The second is the increasing availability of inexpensive sensors The third is the increasing use of fieldbus and wireless in the plant control networks.
All of the companies that make field controllers, whether part of traditional DCS companies, or programmable-controller-based control systems, are merging advanced control algorithms directly into the field controller, so that they work invisibly and are completely transparent to the operator. They will self-update, and since they will be connected to the plant control and asset management systems, they’ll be able to continue to optimize plant performance regardless of changes in the plant process or automation system.
These advanced-control field controllers will also be supplemented by new types of inexpensive sensors. We are at the dawn of the age of “lick ’n’ stick” sensors. These sensors will be inexpensive enough that we will be able to use many of them in a single installation, and take advantage of the fact that the precision and control accuracy of a many-sensor matrix is better and provides more information than a single high-accuracy sensor can.
There are even companies coming out with add-on wireless sensors that can take a mechanical pressure gauge and turn it into a wireless transmitter.
These sensor arrays will be paired with the advanced control algorithms that will depend on multiple variable inputs, and we’ll have true multivariable control for the first time.
The final piece of the puzzle that will shape the future of advanced control is the increasing ubiquity of fieldbus and wireless in plant sensor networks.
Whether connected wired or wirelessly, sensors will be able to report multiple variables and advanced diagnostics and receive updated control information and be remotely calibrated, all automatically from the advanced process control software.
The future of advanced control is to make the plant instrumentation even more transparent to the operator than ever before—so that the operator can concentrate on controlling and optimizing the process itself, not the process automation system.