We Used The Following System Through The Investigation Process

Name : Mustafa Mukhtar Elbashir Index : 174099
- System used in investigation:

We used the following system through the investigation process:
D(s)
- Steps of design:
i. Design a proportional controller:
C(s) = Kp
For the first three steps the disturbance is considered zero.
a- First, I entered the plant and sensor’s transfer functions in MATLAB as
follows:
>> s = tf('s');
>> p = 4/((2*s+1)*(0.5*s+1));
>> H = 1/(0.05*s+1);
Then I plotted the root locus of the system and used the rlocfind to find Kp*
which makes the system marginally stable as follows:
>> openloop_transfer_function = p*H;

>> rlocus(openloop_transfer_function);
>> rlocfind(openloop);
I found that Kp* = 14.0694
b- I used the command sgrid to find the value of the Kp that results in a
system with damping ratio of 𝜉 = 0.8 as follows:
>> sgrid
>> Kp = rlocfind(openloop_transfer_function);
I found that Kp = 0.3269
c- I plotted the step response for different values of Kp < Kp* and showed the
overshoot in each case as follows:
Kp_values = [3 5 7 8 10];
s = tf('s');
p = 4/((2*s+1)*(0.5*s+1));
H = 1/(0.05*s+1);
for i=1:5,
closed_loop = feedback(Kp_values(i)*p, H);
step(closed_loop);
hold on
end
set(findall(gcf, 'type','line'),'linewidth', 1); legend('Kp=3',
'Kp=5', 'Kp=7', 'Kp=8', 'Kp=10');
The results are shown below:
Then by trial and error I found the largest value of gain k̅ Ԓ which produce
about 20% overshoot and it is found to be:
k̅Ԓ = 1.255
And the response for this k̅Ԓ is shown below:
Then I calculated the gain margin and phase margin for this k̅Ԓ as follows:
>> closed_loop = feedback(1.255*p, H);

>> bode(closed_loop);
>> margin(closed_loop);
The results are shown below:

So, gain margin = infinity, phase margin = 106 deg (at 2.15 rad/s)
d- Finally, I used a unit step disturbance and visualized the response of the
system which is as follows:
ii. Proportional plus integral control design:

C(s) = 𝑲𝒑(S +𝑲𝒊) /s
In the first step the disturbance is zero.

a- I plotted the response for Kp changes from 0.3 to 1.7 with increments of 0.2
for the Ki = 0, 0.1, 0.25, 0.5, 1, 2 as follows:
Ki = x;
s = tf('s');
p=4/((2*s+1)*(0.5*s+1));
H = 1/(0.05*s+1);
for i=0.3:0.2:1.7,
c = (i*(s+Ki))/s ;
closed_loop = feedback(C*p, H);
step(closed_loop);
hold on
end
set(findall(gcf, 'type','line'),'linewidth',
1); legend('Kp=0.3', 'Kp=0.5', 'Kp=0.7', 'Kp=0.9',
'Kp=1.1', 'Kp=1.3', 'Kp=1.5', 'Kp=1.7')
x in the above code is replaced by the corresponding value of Ki. The

results are shown below:
Ki = 0:
Ki = 0.1:
Ki = 0.25:
Ki = 0.5:
Ki = 1:
Ki = 2:
The value of the Ki that results in settling time ts ≤ 6s is k̅ӏ = 0.5.
b- The value of the Kp that results in overshoot less than 10% is Kp = 0.5.
Then I plotted the response using these values of Ki and Kp when both reference
input and disturbance input is unit step. The results are shown below:
c- Comparing these results , proportional controller can reduce error by

increasing gain but can not eliminate it , however proportional plus integral
controller (PI) which is designed by connecting proportional controller in
parallel with integral controller can eliminate error to zero
d- I made the disturbance and reference inputs as ramp inputs and the response
is shown below:
iii. PID controller design:

C(S) = 𝑲𝒑 (KD * S^ 𝟐 + 𝒔 + 𝑲𝒊)/S
a- I plotted the response for Ki = 0.5, Kp changes from 1 to 10 with increments of 1,

and for the KD = 0, 0.1, 0.2, 0.3, 0.4, 0.5 as follows:
Ki = 0.5;
KD = x;
s = tf('s');
p =4/((2*s+1)*(0.5*s+1));
H = 1/(0.05*s+1);
for KP=1:10,
C = (KP*((KD*s^2)+s+Ki))/s;
closed_loop = feedback(C*p, H);
step(closed_loop);
hold on
end
set(findall(gcf, 'type','line'),'linewidth', 1);
legend('Kp=1', 'Kp=2', 'Kp=3', 'Kp=4', 'Kp=5',
'Kp=6', 'Kp=7', 'Kp=8', 'Kp=9', 'Kp=10')
x in the above code is replaced by the corresponding value of KD. The
results are shown below:
KD = 0:
KD= 0.1 :
KD = 0.2 :
KD = 0.3:
KD = 0.4:
KD = 0.5:
From above responses we see that the smallest KD that will allow overshoot and
rise time specifications is K*D = 0.4.
b- I plotted the root locus for the system with Ki = 0.5, KD = 0.4 to determine
the j 𝜔 crossing which corresponds to Kp value that make the system
marginally stable as follows:
>> KD = 0.4;
>> C = ((KD*s^2)+s+Ki)/s;
>> openloop_transfer_function = C*p*H;
>> rlocus(openloop_transfer_function);
The result is shown below:

As shown in figure the root locus does not cross the j𝜔 axis so we can not
Find a suitable value for KP.
c- From plot question (a) that corresponds to KD = 0.4 we can

see that the three specifications are satisfied when Kp = 0.7.
To officially verify this result I plotted the response with
these parameters and the result is shown below:
We see that the three specifications are satisfied when Kp = 7.
d- proportional controller reduces error by increasing gain ,

but can not eliminate it , however the proportional plus
integral controller can eliminate the error completely , the
derivative controller makes the system faster but affects on
stability and its gain is very low .
conclusion :
proportional controller can reduce error but will increase the
overshoot , reducing the settling time and the time constant ,
in the other hand the integral controller can eliminate error ,
increasing settling time and time constant , finally , the third
term controller or the derivative controller can highly
increase the overshoot but highly decrease the settling time
and time constant .

We Used The Following System Through The Investigation Process

Uploaded by

Copyright:

Available Formats

We Used The Following System Through The Investigation Process

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

We Used The Following System Through The Investigation Process

Uploaded by

Copyright:

Available Formats

Name : Mustafa Mukhtar Elbashir Index : 174099

- System used in investigation:

>> openloop_transfer_function = p*H;

I found that Kp = 0.3269

>> closed_loop = feedback(1.255*p, H);

The results are shown below:

ii. Proportional plus integral control design:

In the first step the disturbance is zero.

x in the above code is replaced by the corresponding value of Ki. The

The value of the Ki that results in settling time ts ≤ 6s is k̅ӏ = 0.5.

c- Comparing these results , proportional controller can reduce error by

iii. PID controller design:

a- I plotted the response for Ki = 0.5, Kp changes from 1 to 10 with increments of 1,

The result is shown below:

c- From plot question (a) that corresponds to KD = 0.4 we can

We see that the three specifications are satisfied when Kp = 7.

d- proportional controller reduces error by increasing gain ,

You might also like