• Not Answered

Tuning a pH control loop

Hey folks,

I have been trying to use DeltaV InSight to tune a pH control loop, so far i have not had a lot of success due to slow response.  Does anyone have a suggestion as to what is the best way to tune a slow response loop like pH?

5 Replies

  • Hey Chris,
    Are you using DeltaV InSight "On Demand Tuning" or "Adaptive Tuning"?

    Some pH loops have very fast dynamics and some have slow dynamics! If this is the pH of a stirred tank it will be slow. Is this your application? And, it could be an integrating or near integrating response. Do you have any trends where you have put the pH loop in manual and performed a step test to see if it is Self Regulating or Integrating? I'll await your answers and respond after I get more information!
    Good luck!
    James
  • When you say slow response, do you mean an integrating response?  If so, then try checking the Process Type: Integrator box and see if Insight can get a process model.
     
    Another possibility is that you have some nonlinearity in the process or the controls that is confusing Insight. Do you have a split-range with separate slave loops for acid and base addition?  If you have a single control valve, is it actually responding to the controller output?  Are you using strong reagents that can result in highly nonlinear response (steep titration curve)?
     
  • In reply to James Beall:

    Before adaptative tuning you can check with Insight Learning how your process is not linear.
    Typically pH can be split into 3 différents zone where process Gain is very différent ( inflexion point)
    In this case in the DV standard Library you can use PID Gainshed to adress this issue dur to non linear process.
    the appropriate term for pH Loop is not "slow" but long dead time ( depending of the process design)
  • In reply to James Beall:

    I am using "On Demand Tuning".

    In this particular case I am increasing the pH on a process stream before it reaches the next step.  Below is the results from this mornings attempt at using "On Demand Tuning", it's the first time I have had some success with it.  Usually the test times out.  

  • In reply to Mark Coughran:

    Mark,

    I will look and see if insight can get a process model. I did notice that Learning had not been not enabled on this loop, so I did go ahead and enable that.
    In this instance it is a single pH loop with a single valve for pH control. The control valve does respond, just not very well.
    Although in another part of my process I do a loop when I am adjusting the pH with separate valves, one for each reagent.
    I am using 32%HCl and 20%NaOH for pH adjustment throughout the process.