Reasonable Limit of Control Loop Time

I've been asked to control the output of an Ozone Generator based upon the Ozone Residual value measured after the ozone does it's work. Problem is that the time between generator setpoint change and resultant change in residual is in the twenty-minute ball-park. Compounding the problem is that the measured residual is not what one would call a "stable" signal; it tends to jump around a bit.

Any suggestions will be appreciated!

Are you trying to minimize residual? 20-30 minutes is not uncommon for ozone to make its way from the generator then out to the the destruct. I would put some kind of averaging on the residual signal and use that as a trim on the setpoint. But you would still be left with a lag of about 1/2 hour.

You've basically confirmed my suspicion that "trimming" is about all that can be expected. Some other method will be necessary to get a starting point.

And, yes, any ozone that is not absorbed in the process is wasted. So the idea is to minimize the waste.

Thanks for your response.

I would find some way to predict what is going to happen. I would make a model.