How to tune this Transfer Function using Zieger nichols or Cohen coon

because from the this

s=tf('s');

sys=(5809000*s + 28060)/(s^3 + 571300*s + 1266);

step(sys)

i only get this

and the other one

num = [5809000 28060];

den = [1 571300 1266];

Gp = tf(num, den);

p = pole(Gp);

step(Gp, 50)

i get this

how do i tune using the zieger nichols or cohen coon from this?

Answer 1

[Sam Chak]

@Isyraf Izuddin

There is a typo in the denominator:

s^3

I think that Ziegler-Nichols and Cohen-Coon methods are more suitable if the mathematical model of the plant is unavailable. Nevertheless, you can try some automated tuning algorithm available in FileExchange:

https://www.mathworks.com/matlabcentral/fileexchange/47836-automated-ziegler-nicholas-pid-tuning

I prefer human tuning (mathematically, or course) because the human operator can make the system to behave as desired. In this case, a compensator is used to make the system converge around 5 sec without signicant overshoot and undershoot.

s = tf('s');
Gp = (5809000*s + 28060)/(s^2 + 571300*s + 1266)                % plant
Gc = 0.09835/(s^4 + 2.205*s^3 + 3.511*s^2 + 2.817*s + 0.0135)   % compensator
Gsys = feedback(Gc*Gp, 1);
Gsys = minreal(Gsys)
step(Gsys, 15)

[Isyraf Izuddin]

thank you

Answer 2

[MRR]

Hi @Isyraf Izuddin,

You can try to obtain a FOPTD model (First Order Plus Time Delay). Then you can use several tuning rules, including Ziegler-Nichols, and Cohen-Coon. I link here one of our works related with the implementation of tuning rules by means of interactive tools developed in Matlab.

https://www.mdpi.com/2073-8994/10/11/569

Hope it helps.

Answer 3

[Kunal]

s = tf('s'); Gp = (5809000*s + 28060)/(s^2 + 571300*s + 1266) % plant Gc = 0.09835/(s^4 + 2.205*s^3 + 3.511*s^2 + 2.817*s + 0.0135) % compensator Gsys = feedback(Gc*Gp, 1); Gsys = minreal(Gsys) step(Gsys, 15)