Michael Brown answers your control loop questions
I like to read your column on process control in loop signature series, because I think it is suitable for practitioners without leaving behind the theoretical aspect of each subject (writes Harimurti from Indonesia) he continues: I am new to process control; I would like to ask a question about valve selection:
There are many types of valve and various actuating mechanisms; my question is when we should use a valve with electro-pneumatic actuator instead of a motorised valve?
Michael Brown responds:
It is a bit too involved to answer your query in full detail, but basically pneumatically actuated valves are used in process control in at least 95% of plants. This is probably due largely to cost, but electric actuated valves do have disadvantages, not to say that the pneumatic ones do not!
One of the main problems with the electric type is speed (relatively slow). They also suffer from the big disadvantage that electric motors have limited duty cycles and do not like being switched on and off and reversed at a high rate, which is what happens when you are running around setpoint, particularly with noisy process signals. Therefore one normally puts a dead band around setpoint, typically +2%, which gives you a control that probably has more variance than with pneumatic types.
Other problems with them can arise from the method in which the valve is actually positioned. There are several methods, but the most common is a valve position feedback signal that is sent to the controller. I have seen numerous problems with this in many plants, ranging from noisy signals from feedback potentiometers, short life of the potentiometer requiring frequent replacement. Also sometimes there is a large hysteresis in the mechanism driving the position feedback element.
Harimurti:
Your answer really helped me, most of the tutorials that I found on the Internet always use control valves with pneumatic actuators when describing process control. I also found a good description of control valves and actuators in Spirax Sarco website (www.spiraxsarco.com) in their learning centre link.
I have another question: Can I control two valves from single controller output?
For example, in a neutralisation process I would like to have neutral pH, then I have to add either basic or acid reagent. The controller receive the pH value from the pH probe and determine which reagent and how much to be added to. Since I only measure pH, I need only one controller but I have to control two valves, one for basic reagent and the other for acid reagent.
Michael Brown:
Yes, you can use one controller with an output to twp valves, one for the base, and the other for the acid. You use what is known as 'split-range control'. Typically depending on controller and valve actions you split the output range from the controller so that roughly the first half of the signal (say 4 to 12 mA) goes to the one valve, and the remaining half (12 to 20 mA) works the second half. Most valve positioners can be ranged for such signals.
Two words of warning:
The actual split must be done so that the process gain (controller input divided by controller output) is equal for both ranges. (See also my Loop Signature Series on process gain. This series is currently on my website, but will soon be available on a CDR.)
Secondly, pH control is very non-linear depending on where you are working on the titration curve. You may have to use Gain Scheduling (ie, different tuning values) at different pH values for optimum control.
© Technews Publishing (Pty) Ltd | All Rights Reserved
printer friendly version