I have a closed loop biological control system where i use a PID controller. Because the transfer function of the biological system is complex to derive, I used trial and error to tune the PID parameters. And the resulting response is nearly critical damped. What can I learn from the parameters value of PID values that I found?
Please attach any tech notes or papers. Thanks in advance.
Yes, iterative learning control could be a solution for closed loop controller tuning. However, if you need to know the more accurate system model, open loop identification will obviously provide more reliable results
Hi Rex,
First of all, because we understand that you managed to control your system, I guess that, by closing your loop with a proportional gain P, you already got a stable (maybe not too good, yet not unstable) system.
Then, playing with the D and I gains gave you better performance, although not necessarily as good as you may want.
Identification is a good idea, yet this requires knowing the number of parameters that you must identify, or, in other words, the exact order of your system, which I doubt is available if you talk about a totally unknown system.
However, now you can take a Signal Analyzer and find the frequency response of your stable closed-loop plant. This would give you the graphs of Amplitude and Phase and a first idea about the transfer function (TF) of your system.
From here, there are procedures (for example in MATLAB) that allow you to compute the open-loop TF, etc.
I hope it helps.
Have Fun!
Itzhak
- Badges
- Science topic
- Similar topics
- Psychology
- Mental Processes
- Learning