PID not being able to apply Ziegler Nichols method.

zazas321

Joined Nov 29, 2015
893
Hey guys. I have came across one problem trying to manually tune my system. I was trying to apply Ziegler Nichols method to find Critical gain by setting Kd and Ki gains to zero and increasing Kp gain till the system oscilates. But my system(see picture) is never reaching oscilation it is always stable.

Attachments

• 64 KB Views: 5

kjj

Joined Mar 30, 2018
31
First of all, your system is a type 1 system, so the integrator is not needed. https://image.slidesharecdn.com/con...p02/95/control-chap6-30-638.jpg?cb=1388519031

For Ziegler-Nichols, i qoute: "If the output does not exhibit sustained oscillations for whatever value KP may take, then this method does not apply" http://kurser.iha.dk/m/mtpri1/control/3_PID/f_ziegler/ziegler_nichols_method.html

Btw, tuning this system is very easy, first order transfer function and type 1 means a simple P-controller can be used. Add D to reduce overshoot. Example simulation i made now:

zazas321

Joined Nov 29, 2015
893
In my system, integrator is part of a plant I just did not put it a single block. It would have been the same if i used 1/Js^2 + Ds as a plant. So you are saying I am not able to apply ziegler nichols rule in my case? Any other methods that i could use to tune the system particulary using all three P I and D gains

kjj

Joined Mar 30, 2018
31
In my system, integrator is part of a plant I just did not put it a single block. It would have been the same if i used 1/Js^2 + Ds as a plant. So you are saying I am not able to apply ziegler nichols rule in my case? Any other methods that i could use to tune the system particulary using all three P I and D gains
I think you misunderstood, the integrator in the PID is not needed since you have type 1 system.

There are many methods for tuning. You should try some manual tuning to learn.

Simulink has built in PID tuner which uses a more advanced tuning method

zazas321

Joined Nov 29, 2015
893
I am reuploading the picture to show what the final plant looks like. I have tested this circuit with different values of gain and it is not reaching oscillations. I have also looked at cohen coon tuning method but could not really get it to work as well. I need to investigate atleast 1 tuning method that could be used to tune my system. Thank you though for the help

Attachments

• 64 KB Views: 1

kjj

Joined Mar 30, 2018
31
That picture is virtually the same as the previous one. And you dont even say the value of J and D!

And as i said, a P-controller or PD-controller is what you need. It is easy to tune manually. Have you even tried to tune it manually?

And when you say "i need to investigate", you mean like homework? You can also tune it analytically, or using the frequency response, or of course manually.