Page 66 - 2024-Vol20-Issue2
P. 66

62 |                                                                                            Al-Najari, Hen, Paw, & Marhoon

      Fig. 7. Step response                                                      Fig. 9. Step response Goriginal

                                                                    TABLE II.
                                                                    ZIEGLER NICHOLS CLOSE LOOP PID PARAMETERS

                                                                    Variable              Relation       Value
                                                                      Kcr              From drawing  13.8114947
                                                                       Pcr             From drawing
                                                                       Kp                               47.009
                                                                       Ti                 0.6*Kcr    8.28689682
                                                                       Td                 0.5*Pcr
                                                                       Ki                0.125*Pcr      23.5045
                                                                       Kd                   Kp/Ti      5.876125
                                                                                           Kp*Td     0.352566395
                                                                                                     48.69484158

                    Fig. 8. Multi step response                     as a polynomial model using Pade approximation and hence
                                                                    control system is designed for it [16]. The first-order pade
1) Ziegler Nichols ZN                                               approximation is:
In this logic, the controller is set as a proportional controller,
and then the process is near the specific operating point of        e-? s  =  1-    ?  s                          (4)
marginal stability. The ‘Kp’ value will increase till the output                    2
shows oscillations that are studied. The value of ‘Kp’ cor-
responding to this is called critical gain ‘Kcr’. The period                  1  +  ?  s
of the oscillation ‘Pcr’ is called the critical period. The PID                     2
parameters are then calculated using Table II [15].
                                                                        Where ? is the delay time.
    Fig. 7 shows the step response of the original function             The delay time of the original function is ? =18.2. There-
using ZN tuning. Fig. 8 shows the Multi-Step Response of            fore, the equation (4) will be:
the original function using ZN tuning.
                                                                    e-18.2s   =  1-    18.2  s                    (5)
2) MATLAB Tuner                                                                  1+     2
The MATLAB Tuner method was used to improve the re-
sponse of the original function. Fig. 9 shows the step response                        18.2  s
original function using MATLAB Tuner. Fig. 10 shows the                                 2
Multi-Step Response of the original function using MATLAB
Tuner.                                                                  From equations (2) and (5), the approximated transfer
                                                                    function (Gapproximated) will be:
B. Pade Approximation
As shown in Fig. 9 and Fig. 10, the original function has a                                -0.8672s + 0.0953      (6)
delay time also the performance of the response is not good.        Gapproximated(s) = 58.24s2 + 15.5s + 1
The delay time affects control. To solve this problem, the Pade
approximation method was used to approximate the original               Fig. 11 shows the step response of approximated function
function (Goriginal). The FOPDT model of PH is approximated
                                                                    (Gapproximated) using MATLAB Tuner. Fig. 12 shows the
                                                                    Multi-Step Response of approximated (Gapproximated) func-
                                                                    tion using MATLAB Tuner.
   61   62   63   64   65   66   67   68   69   70   71