Guest commentary: Low-cost, high-return process control improvements, part 2
In the second of a three-part essay, George Buckbee, P.E., vice president of marketing and product development for ExperTune, considers a number of very practical ways you optimize your plant control systems and improve profitability.
Last month we got warmed up by considering how much money can be saved by stopping compressed air leaks. Now that you have the concept in mind, let's look at more specifically control-related opportunities.
Idea 2: Reducing overuse of derivative action in PID loops
A PID controller is a very capable tool that is often misapplied. Most control loops do not require derivative action, but have it nonetheless. When derivative action is applied too broadly in your plant, it will cause process upsets, increase variability, drive valve wear, and waste compressed air.
So if most applications don't need it, how can you identify those that do? Derivative action should only be applied to loops when:
• There is little or no noise;
• Fast-responding control is critical;
• P and I control alone is not good enough; and
• There is some (small) filtering of the PV.
You can find candidates for derivative removal easily enough. Look for loops that have excessive valve travel (see graphic) and see where the D value is set. Valve travel is the total amount of valve movement, up and down, totalized over a one-day period. When you find such situations, the simple solution is to remove derivative action on these loops and then see if the valves settle down. The result will be a cost reduction in two immediate areas, including:
• Less compressed air consumption. Go back to the first part of this series to see how to quantify that.
• Reduced valve maintenance. Valves subject to constant cycling due to excessive derivative driven movement will wear out positioners, actuators, and seals. Eliminate the excessive action and you will save on valve repairs and replacements.
Cost Estimate : If you use software like Plant Triage to identify the problem loops, you can check 100 loops in an hour. The "repair" is simply changing the derivative setting, so the time involved is minimal.
Benefit : If you eliminate derivative action on just 5% of the 100 loops you check, that's 5 loops with a reduction of $500 per loop per year in air consumption, so that's $2,500 in the first year.
If in that group of five valves you eliminate even one valve repair during the next year, that can easily be another $2,000.
Idea 3: Add a filter
Some instrument signals are noisy. The cause may be related to the inherent characteristics of the sensor, its application, or its position in the process. Even if you don't have derivative action, this signal noise will be amplified by the proportional action of the controller, inducing more process upsets. Furthermore, the noise makes it difficult for operators to see what is happening at a given time.
The graphic shows a process variable with a high noise band. Note that the control output is hardly moving at all. You can identify these manually one at a time, or use Plant Triage to scan all the loops and identify the issues automatically.
Choose a filter for the loop based upon the dominant process dynamics. Too large a filter will hide the real process behavior, and too small a filter will result in no improvement. The filter should be at least 3X the sampling time, and no more than 1/3 of the process dead time. PlantTriage contains tools to develop a model of the process dynamics and recommend an appropriate filter.
Benefit : Excessive noise band problems often result in quality upsets and poor decision-making. This is especially true on a modern DCS, where PV's are displayed as rapidly-changing numbers on a screen. Consequently, it can be difficult to estimate a specific value for a given loop. However, from typical customer case studies, we have seen savings between $1,000 and $10,000 per affected loop in a year.
George Buckbee, P.E., is VP of marketing and product development for ExperTune . Reach him at firstname.lastname@example.org. His series Seven habits of highly successful control engineers received very high readership.
-Edited by Peter Welander, process industries editor, PWelander@cfemedia.com
Control Engineering Process & Advanced Control Monthly eNewsletter
Register here to select your choice of free eNewsletters .