Three Variable PID control

Hey everyone, I am a student working on a control system senior project and could use some help.

I am working to maintain a dissolved oxygen setpoint with three possible ways to influence the value. The outputs are agitator speed(200-700rpm), air flow in(1slm-10slm), air/o2 concentration(30-100%).

I am doing this in Studio 5000 without PIDE(will need IT to add it to student package if its a better solution).

The process typically done is to set agitation to the 200 min, set air flow(30% o2 concentration) to 1 slm and then the 35% Dissolved oxygen set point. As the cells consume oxygen, the agitation can be increased to a max of 700 before bumping up the air flow.

If the DO setpoint can be achieved with just agitation control then airflow and concentration should not vary.

Sorry for the long background. My current idea is to only focus on one PID( DO vs speed)then when the max speed is reached bump the air flow .5 slm at a time and return to the PID. Expected result is a speed decrease. Air concentration would just be randomly changed if flow and agitation remain high.

I feel like i am not utilizing the PID to its potential. So in conclusion, how could I go about optimizing those 3 values while achieving desired setpoint.?

http://www.plctalk.net/qanda/showthread.php?t=113807

http://www.plctalk.net/qanda/showthread.php?t=98662

The above links may help you out.
Regards,

Skip the PID. Just use the floating control method. "Adjust wait and see" in other words.

I am curious about your system. You seem to be combining mechanical aeration and diffused aeration. This isn't common. Is this a field situation or a lab system?

You are correct. This is the standard way bench top fermenters are controlled. The system Im using for reference(Biostat B or DCU) has a cascade button but im not sure this is cascade since 3 outputs can be controlled.

Your method for controlling the aerator until it hit max and then controlling air flow should be OK. I'd use floating control on each, with time delays between actions as dictated by process system response.

Note: I don't know the purpose of your study, but be aware that field systems response dynamics are wildly different from lab tests.

Tom,



Is the graphic from your book?

What Tom suggests isn't complicated. I think there is more too this.
If I were a devious professor giving this problem to students I would be thinking,"who are the designers and who are the users?"
If there are three different ways to control something then which way is the best way?


Users will just make the system work.
Designers will optimize how the systems works.

Peter not a class problem, this problem is trying to replicate the $25k Reactor using PLC for around 7k. Just not sure how the bioreactor company does the control loops. My goal for now is to keep the set point without violent oscillation seen in research papers. A future goal is to optimize gas consumption since pure O2 will be the biggest cost. Also I plan to do lab testing in the next few weeks. Just seeing what ideas everyone has first.

danw said:

Tom,



Is the graphic from your book?

Click to expand...

Yes, it is Figure 9.8