Nonlinear controller for pH control in AB PLC?

  • Thread starter Richard Hendrickson
  • Start date
R

Thread Starter

Richard Hendrickson

I am trying to control the addition of caustic or acid based to a tank based on the pH value of the influent entering the tank. I have tried the PID block in the AB PLC 5 but since it is nonlinear the proportional, integral and derivative terms must have values that change with the measured process value. I am contemplating using a third degree polynominal function to control the addition. Has anyone found an accurate way to control pH in an Allen-Bradley PLC 5?
 
Could you supply more process informations ? Whatever, your approach is wrong -- you can not avoid using massflow loop. In other words, based on the input flow, the PH of the influent, the chemical constituents, temperature ...and laboratory or chemist analysis; then you can conclude the amount of chemicals to be added. Several years ago, Durco manufactured metering pumps of simple design but absolutly robust and resistant to very many chemicals. I redesign a cooling water system for one of the world largest system (108 ") c/w cooling towers. In 1940, Foxboro model 40 was adequate for this kind of PID. In 1990, AB (in particular) was unable. Reason: insufficient provision for both, derivative and integral. Looking at their numerical Derivative and Integral algorithms, I would tell in seconds, i.e: the algorithms resident in your system. Being highly specialised in numerical methods, I warn you about polynomial approximations (if you intend to look around about your loop). Titration curve exibits a non approximatable function via polynomials. The solution is of same nature as the phenomenon, i.e: exponential relationship. Going back to massflow, if everything is correct about the stoichometrics, then the PID will just seat and watch the loop, slighly retouching. Then probably Proportional will finish the job. [email protected]
 
J
The only thing I see that is non-linear here, is the logarithm used to report pH, that's easy to turn into concentration of acid or base in both the incoming solution and the chemical feed products. After that it's all algebra and chemistry, a PID loop won't help. You can calculate the required amount of extra acid or base needed per unit volume of incoming solution and adjust the feed rate accordingly. A PID controller would be useful here if another pH meter was located in the tank. Then you could trim the feed rate with that pH meter as your feedback.
 
C

Controls Engineer

I am going through the same application. I have had very good success with non-linear single loop controllers utilizing e^2 or 3-zone characterizers. The AB PLC does not have a non-linear PID nor do they have a basic PID algorithm that accepts an external error function. Our application is very difficult because we have numerous different chemicals to treat daily and the titration curves vary wildly during the day. We cannot simply mass balance based on any titration curve. Any ideas?
 
Top