Optimisation PID Using Sliding mode script

I want à script for optimise PID paramèters Using Sliding Mode Control

Answer 1

[Sam Chak]

Hi @Tarek

I don't think that SMC can optimize the PID gains. The structure of SMC is not an optimization algorithm.

Before optimizing the PID gains, have you designed a stabilizing PID controller?

If you want to optimize the PID gains, I think that the basic LQR algorithm works for a linear plant or a linearized system.

If you need a demonstration, please show the model of the system.

[Tarek]

(-0.014*s^2+3.436e-5 *s+1.214e-8) /(S^2+0.0004*s 5.834e-9) This is my model

Answer 2

[Sam Chak]

Hi @Tarek

Are you still working on this? I revisited this control problem and discovered that a PID controller can be tuned to eliminate the steady-state error in the step response.

s   = tf('s');

%% Plant
Gp  = (-0.014*s^2 + 3.436e-5*s + 1.214e-8)/(s^2 + 0.0004*s + 5.834e-9)

Gp =
 
  -0.014 s^2 + 3.436e-05 s + 1.214e-08
  ------------------------------------
       s^2 + 0.0004 s + 5.834e-09
 
Continuous-time transfer function.

%% Controller
Gc  = pidtune(Gp, 'PIDF')

Gc =
 
             1            s    
  Kp + Ki * --- + Kd * --------
             s          Tf*s+1 

  with Kp = 0.645, Ki = 1.98e-05, Kd = -1.57e+03, Tf = 1.73e+04
 
Continuous-time PIDF controller in parallel form.

%% Closed-loop system
Gcl = minreal(feedback(Gc*Gp, 1))

Gcl =
 
  -0.007826 s^4 + 1.84e-05 s^3 + 8.749e-09 s^2 + 7.389e-13 s + 1.402e-17
  ----------------------------------------------------------------------
      s^4 + 0.0004799 s^3 + 3.796e-08 s^2 + 1.079e-12 s + 1.402e-17
 
Continuous-time transfer function.

%% Plots
step(Gp ), hold on
step(Gcl), grid on
legend('Plant', 'Closed-loop', 'location', 'se')