You can use the Matlab program that will help you obtaining the parameters of the PID. It's the simplest way !

I design this blocks still i dont know how can i generate step input to ramp input and how can i arrange to opening motor in 4 or 10 seconds.

Dear Uzcan Zenger,

In fact there are classic methods such as the Ziegler Nichols method to determine the gain of the PID controller, however the simplest way is the use of Matlab providing "pidtool" which automatically chooses the optimal Kp Ki Kd PID gains, which makes testing and error unnecessary. Also you can access the tuning algorithm directly using "pidtune" or through the graphical user using "pidtool".

For more details and information about this subject, i suggest you to see links and attached file on topic.

PID ... - Control Tutorials for MATLAB and Simulink - Introduction

ctms.engin.umich.edu/CTMS/index.php?example=Introduction...

Determination of PID controller parameters from step response ...

www.12000.org/my_notes/PID_ode/index.pdf

Best regards

Dear Zenger,

If I have well understood you main concern is the design of PID for you motor control. Here below you will find (a copy of the program is attached to the message) a Matlab program that will help you obtaining the parameters of the PID. Just copy and paste in your matlab editor then run it to get the PID parameters. Actually the program will give two different PID controllers and you can choose the one that fits your desired performances. The second PID provides faster response than the first one.

regards .

clc;

clear; % here is your system 5/(s^2+18s+1)

num=[0 5];%

den=[1 18 1];% poles: P1=-17.9443, P2=-0.0557

G=tf(num,den);% poles are negative that's your system is stable

[C_PID, info]=pidtune(G,'PIDF');

info

% the previous instruction gives the PID parameters and the chosen open-loop crossover

% frequency

%you can test your PID using the step response as follows:

T_PID=feedback(C_PID*G,1);

hold on

step(T_PID);

% the previous PID has crossover frequency of 0.0985 rad/s

% if you want a faster response, you have to change the open-loop crossover

% frequency, for example

[C_PID_fast, info]=pidtune(G,'PIDF',1.0);

TI_PD_fast=feedback(C_PID_fast*G,1);

hold on

step(T_PID_fast);

Hi Ozcan Zenger,

Covey's Habit #2: Begin With the End in Mind. This provides an alternative theoretical approach for the control design.

The plant's transfer function is given by Gp = 5/(s2 + 18*s + 1). The plant is already stable but very slow (due to 1

Hi Ozcan Zenger,

If the desired settling time is 10 sec, using the P-only controller will result in a larger steady-state error. Hence, a PI controller maybe considered. If so, then the closed-loop transfer function becomes

Gcl = 5*(Kp·s + Ki)/[s3 + 18·s2 + (5·Kp + 1)·s + 5·Ki]

Since the order of denominator is 3 and the coefficient of s2 cannot be manipulated, you can try to make the coefficients of the polynomial as "balanced" as possible. Balanced in this context means that the coefficients of s2 and s1 should be as close as possible and coefficient of s0 should be approximately one-third of the coefficient of s1, (because Gcl is a 3rd-order system).

One good candidate is s3 + 18·s2 + 18·s + 6. Solving the simultaneous equations yields Kp = 17/5 and Ki = 6/5. The following m-code plots the step response. Luckily, it turns out that the response settles around 10 sec and the overshoot is relatively acceptable.

s = tf('s');

Gp = 5/(s^2 + 18*s + 1)

kp = 17/5; ki = 6/5;

Gc = kp + ki/s

Gcl = Gp*Gc/(1 + Gp*Gc);

Gcl = minreal(Gcl)

step(Gcl, 20)

Since you have a known plant model i.e. the T.F, so you can tune the controller easly. However one easy way is to use auto tuning option from the PID block which gives you the parameters values, also you can use traditional or optimal methods for more accuracy...

You can use the Matlab program that will help you obtaining the parameters of the PID. It's the simplest way !

http://ctms.engin.umich.edu/CTMS/index.php?example=Introduction§ion=ControlPID

The PID controller block you use before the plant transfer function have auto tuning option to find the PID controller parameters. And you can manually set your desired response in PID tuner which changes the values of Kp, Ki and Kd accordingly.

Kp and Ki can be found by following methods

1. Trial and Error Method

• Set integral and derivative terms to zero first and then increase the proportional gain until the output of the control loop oscillates at a constant rate. This increase of proportional gain should be in such that response the system becomes faster provided it should not make the system unstable.
• Once the P-response is fast enough, set the integral term, so that the oscillations will be gradually reduced. Change this I-value until the steady-state error is reduced, but it may increase overshoot.
• Once P and I parameters have been set to the desired values with minimal steady-state error, increase the derivative gain until the system reacts quickly to its set point. Increasing the derivative term decreases the overshoot of the controller response.

2. Zeigler-Nichols Method

a)method 1

• Obtain a unit step response of the plant experimentally and it may look‘s’ shaped curve as shown in the figure below. This method applies, if obtained response exhibit s-shaped curve for unit step input otherwise it cannot be applied. This curve can also be obtained by dynamic simulation of the plant.

b)method 2

• it is very similar to the trial and error method where integral and derivative terms are set to the zero, i.e., making Ti infinity and Td zero.
• Increase the proportional gain such that the output exhibits sustained oscillations. If the system does not produce sustained oscillations then this method cannot be applied. The gain at which sustained oscillations produced is called as critical gain

please Ref the image for tuning Kp and Ki