0
$\begingroup$

I have four different plant systems which I want to control. They are similar in some aspects (equal inputs and outputs as well as disturbances, all are thermodynamical systems). Furthermore, none of the systems tends to oscillate. The major difference are the time constants. Additionally, each system defines specific states which should be kept in bounds (e.g. restrict the maximum temperature).

Is there a systematic way of how to design a controller for these systems? Can I classify the systems in a way that I can claim that for all systems the same control approach can be used? Is there any literature for designing a controller especially with focus on which control strategy to use (there are so many and I cannot find a guide on which system I should use a certain controller e.g. PID, MPC, H_inf, Fuzzy control, integrator backstepping, LQR)?

$\endgroup$
4
  • $\begingroup$ You'd want to describe the systems enough so one could classify their dynamics... Then to the question "Is there a systematic way of how to design a controller for these systems" the answer will probably be a strong yes $\endgroup$
    – Pete W
    May 18 at 13:31
  • $\begingroup$ @PeteW Can you recommend literature on how to classify the dynamics of a system? Especially, if the system is nonlinear? $\endgroup$ May 18 at 13:37
  • $\begingroup$ Can't really give you a recommendation for nonlinear systems, but if you write out the form of the systems maybe someone else would have an approach. If they the nonlinearities are mild, there is a possibility that linearization (with possible gain programming) will work. $\endgroup$
    – Pete W
    May 18 at 13:40
  • $\begingroup$ If this is all your criteria then throw bang bang control on all of them and you're done in 10 minutes. This is how all your home thermostats work, it's just fine if you don't have more exacting requirements. My "system" is to choose the simplest solution and start with that, if it won't work or is not acceptable for some reason, then move on to more complicated solutions. The next simplest would probably be PID. Evaluate how that will perform and consider whether you should go further, and so on. $\endgroup$
    – Drew
    May 20 at 3:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.