Hi!
Settling time, rising time and overshoot are parameters to be optimized or specified to get a desired performance.
What are other parameters we should consider when tuning PID controllers?
yep, stability. The higher the gain, the closer the control, and the greater tendency to oscillate.
You have the classic undesired parameters of under damped, over damped
then overshoot, undershoot
oscillations.
Usually tuning required a step response. I've never programmed auto-tuning.
In a software implementation, you have to program so that "reset windup" does not occur. There are a number of ways of doing that.
If p is an output of -100 to 100%, then making sure the Integral term is never bigger than 100% is one way, Limiting the error is another.
In one design I put a limit on dp/dt. You do want to make the units correct. e.g. repeats/min
There's all sorts of parameters that can be used. Limiting Pmin and Pmax. e.g. 0 to 90%.
For resistive heating, having access to I, V, Power and optionally T can make an interesting PID controller.
Back in the 80's my team added heat-up energy as a limit too as well as stability criteria and open heater detection. Heat-up energy allowed for shorted or badly positioned thermocouple detection too.
You can limit I, V or Power. Tungsten heating is different than resistive heating.
I used to do a lot of manual PID tuning using a strip chart recorder.
Hi!
Settling time, rising time and overshoot are parameters to be optimized or specified to get a desired performance.
What are other parameters we should consider when tuning PID controllers?
Many processes have a number of operating modes which may require a different parameter set. Recipe loading, Operator interaction, logic selection, etc.