Mr. Biruk Simaani (M.Sc.

)
Menschen für Menschen Foundation
Agro – Technical and Technology College
Electrical and Electronics Technology Department
 Chapter - Six
Control Systems Design

OBJECTIVES: The purpose of this chapter is to:

❖ Discuss the notion of controller design and the hardware elements
needed for its implementation.
❖ Identify the types of feedback controller which are available for
system control and examine their effect on the response of system.
❖ Design controller using Root Locus and Frequency Domain methods

Instructor: Mr. Biruk Simaani (M.Sc.) 1

 6.1. Introduction
Consider the generalize process shown in Figure below

It has manipulated variable 𝑚, Output variable 𝑦 and disturbance variable 𝑑

The feedback control action takes the following steps

❖ Measure the value of the output using the appropriate measuring device - Sensors
❖ Compare the measured value 𝑦𝑚 to the desired value 𝑦𝑠𝑝 (set point) of the output.
❖ Compute the deviation 𝝐 = 𝒚𝒔𝒑 - 𝒚𝒎
❖ The value of the deviation is supplied to the main controller
Instructor: Mr. Biruk Simaani (M.Sc.) 2
6.1.1. ON- OFF Controller: The output of the controller changes when the error
value changes from positive to negative or vice versa.
❖ The output signal from the controller 𝑚 𝑡 based on the actuating
error signal 𝜖 𝑡 may be either at a maximum or minimum value.

The block diagram of a two position ON–OFF

controller is shown in Figure below

Instructor: Mr. Biruk Simaani (M.Sc.) 3

Example of Two Position ON–OFF Controller
A. Electromagnetic Relay: It has two contacts: normally open (NO) and normally
closed (NC) contacts
The opening of the contact is controlled by the relay coil,
which when excited changes the contact from one position
to another (i.e., NO→NC or NC→NO).

Input Relay Output

Instructor: Mr. Biruk Simaani (M.Sc.) 4
 B. Electronic ON-OFF Controller - Comparator Circuit
❖The two-position ON–OFF controller implemented electronically
using a comparator.

Application of Two-Position ON–OFF Controller

The two position ON–OFF controller is used in
❖ Air conditioning/room heating system
❖ Refrigerator temperature control system
❖ Liquid bath temperature control
❖ Liquid-level control in tanks

Instructor: Mr. Biruk Simaani (M.Sc.) 5

6.2. Basic Control Modes: There are three basic types of feed back controller
I. Proportional Controller:
❖ It is simple to implement and the most widely used method of
control.
❖ The error signal is amplified to generate the control (output) signal.
❖ The output signal of the proportional controller 𝑚 𝑡 is
proportional to the error signal 𝑒(𝑡).

Where Kp is the proportional gain or constant.

The actuating output signal
𝑚(𝑡) = 𝐾𝑃 𝑒(𝑡) + 𝐶𝑠
Where 𝐶𝑠 is controller bias signal
Instructor: Mr. Biruk Simaani (M.Sc.) 6
 Taking Laplace transform, we obtain

Therefore, the transfer function of the P-controller is given by

The block diagram of the P-controller is shown in Figure below

Proportional Band: The gain in the proportional controller can be referred to the
proportional band and it is given as

Instructor: Mr. Biruk Simaani (M.Sc.) 7

 ❖ The proportional band characterizes the range over which the error must change in
order to drive the actuating signal of the controller over its full range.
Usually 𝟏 ≤ 𝑷𝑩 ≤ 𝟓𝟎
Note: The larger the gain 𝐾𝑃 , or equivalently, the smaller the proportional
band, the higher the sensitivity of the controller’s actuating signal to the
deviation 𝜖 𝑤𝑖𝑙𝑙 𝑏𝑒
Define the deviation 𝑚′ 𝑡 = 𝑚 𝑡 − 𝐶𝑠
and take 𝑚′ (𝑡) = 𝐾𝑃 𝑒 𝑡

Instructor: Mr. Biruk Simaani (M.Sc.) 8

 Example of Electronic P-Controller: It can be realized either by an inverting amplifier
followed by a sign changer or by a non-inverting amplifier with adjustable gain.

The transfer function of the system

is given by

But the transfer function of the P-controller is

Instructor: Mr. Biruk Simaani (M.Sc.) 9

 Advantages of P-controller
❖ It amplifies the error signal by the gain value Kp.
❖ It increases the loop gain by Kp.
❖ It improves the steady-state accuracy, disturbance signal rejection and relative
stability.
❖ The use of controller makes the system less sensitive to parameter variations.

Disadvantages of P-controller
❖ System becomes unstable if the gain of the controller increases by large value.
❖ P-controller leads to a constant steady-state error

Instructor: Mr. Biruk Simaani (M.Sc.) 10

II. Proportional Integral Controller
❖ Most commonly it is known as Proportional Plus Reset Controller.
❖ The output of the proportional integral controller (PI-controller) consists of
two terms:
❑ Proportional to error signal and
❑ Proportional to the integral of error signal.

Where Kp is the proportional gain and

Ki is the integral gain.
1
𝑚 𝑡 = 𝐾𝑃 𝑒 𝑡 + 𝐾𝑃 න 𝑒 𝑡 𝑑𝑡
𝜏𝑖
Instructor: Mr. Biruk Simaani (M.Sc.) 11
 Taking Laplace transform, we obtain

The transfer function of the PI-controller is given by

The block diagram of PI-controller is shown in Fig

Instructor: Mr. Biruk Simaani (M.Sc.) 12

The equivalent block diagram of the PI controller is shown in Figure below

Example of Electronic PI-Controller: The PI-controller can be realized by an op-amp

integrator with the gain followed by a sign changer.
The transfer function of the system is given by

But the transfer function of the PI-controller is

Instructor: Mr. Biruk Simaani (M.Sc.) 13

 Advantages of PI-controller
❖ Eliminates the offset present in the proportional controller
❖ Provides faster response than the integral controller due to the presence of
proportional controller.
❖ Fluctuation of the system around the set point is minimum
❖ Has zero steady state error
❖ Increases the loop gain

Disadvantages of PI-controller
❖ It has maximum overshoot.
❖ Settling time is more.

Instructor: Mr. Biruk Simaani (M.Sc.) 14

 III. Proportional Derivative Controller
❖ Used in the system to have faster response from the controller
❖ It combines the proportional and derivative controller in parallel
❖ The output of the proportional derivative controller consists of two terms:
❑Proportional to error signal and
❑Proportional to the derivative of error signal.

Where Kp is the proportional gain and

Kd is the derivative gain

Instructor: Mr. Biruk Simaani (M.Sc.) 15

 Taking Laplace transform, we obtain

Therefore, the transfer function of the PD-controller becomes

The block diagram of PD-controller is shown in Figure below

Instructor: Mr. Biruk Simaani (M.Sc.) 16

 The equivalent block diagram of PD controller is shown in Figure below

Example of Electronic PD-Controller: The PD-controller can be realized by an

op-amp differentiator with the gain followed by a sign changer

The transfer function of the system is

given by

But the transfer function of the PD-

controller is

Instructor: Mr. Biruk Simaani (M.Sc.) 17

 Advantages of PD-controller
❖ It has smaller maximum overshoot due to the faster derivative action.
❖ It eliminates excessive oscillations.
❖ Damping is increased.
❖ Rise time in the transient response of the system is lower

Disadvantages of PD-controller
❖ It does not eliminate the offset.
❖ It is used in slow systems

Instructor: Mr. Biruk Simaani (M.Sc.) 18

IV. Proportional Integral Derivative Controller
❖The PID controller combines the advantages of PI- and PD-controllers
❖It is the parallel combination of P, I and D-controllers
❖By tuning the parameters in the PID-controller, the control action for
specific process could be obtained
❖The output of the PID-controller consists of three terms:
❑Proportional to error signal
❑Proportional to the integral of error signal
❑Proportional to the derivative of error signal

Instructor: Mr. Biruk Simaani (M.Sc.) 19

Taking Laplace transform, we obtain

Therefore, the transfer function of the PID-controller is given by

The block diagram of PID-controller is shown in Figure below

Instructor: Mr. Biruk Simaani (M.Sc.) 20

The equivalent block diagram of PID-controller is shown in Figure below

Example of Electronic PID-Controller: The PID-controller can be realized by an

op-amp integrator and differentiator with the gain followed by a sign changer.
The transfer function of the system is given by

But the transfer function of the PID-controller is

Instructor: Mr. Biruk Simaani (M.Sc.) 21

 Advantages of PID-controller
❖ It reduces maximum overshoot
❖ Steady-state error is zero.
❖ It increases the stability of the system.
❖ It improves the transient response of the system.
❖ It is possible to tune the parameters in the controllers.

Disadvantages of PID-controller
❖ It is difficult to use in non-linear systems.
❖ It is difficult to implement in large industries where complex calculations are
required.

Instructor: Mr. Biruk Simaani (M.Sc.) 22

6.3. Controller Design via Root Locus
❖ A closed-loop system is usually an unstable system. Hence, because it is
unstable, there must be some kind of compensators(controller) that can
compensate the stability of a closed loop system.

❖ Compensators are used to alter the output response of a system in order

to accommodate to the set of desired criteria. This is achieved by
introducing additional Poles and/or Zeros to the system transfer
function.

❖ Sometimes, the word regulator or correction is used instead of the word

compensator.

Instructor: Mr. Biruk Simaani (M.Sc.) 23

❖ Consider the following diagram where we have a sample root locus of a transfer
function.

Instructor: Mr. Biruk Simaani (M.Sc.) 24

 ❖ In the previous diagram, we assume that the desired output that we want is at
point B. However, we note that currently, our output is at A. And we further note
that at both points A and B, we desire a system with a “nice” percent
overshoot and settling time.

❖ When we change the location of the root from point A to point B, the system
response is speed-up, and interestingly, the percent overshoot is not affected.

❖ When we want to change from point A to point B, we can’t simply adjust the gain
of the system; rather, we need to introduce additional poles and/or zeros.

Instructor: Mr. Biruk Simaani (M.Sc.) 25

 ❖ The introduction of additional zeros and/or poles will speed-up the response
of the system.

❖ When we introduce additional poles/zeros, we are actually improving the transient

response of the system, as well as reducing the steady-state errors.

❖ Additional poles will eventually improve the steady-state characteristics, while

additional zeros will improve the transient response.

❖ Recall that poles are also called integrators in s-domain while zeros are called
differentiators.

Instructor: Mr. Biruk Simaani (M.Sc.) 26

6.3.1. Compensation Configuration
❖ Two configuration of compensation are commonly used:
(a) Cascade Compensation:
The compensating network, G1(s),
is placed at the low-power end of
the forward path in cascade with the plant.

(b) Feedback Compensation:

The compensator, H1(s), is placed in the
feedback path.

❖Both methods change the open-loop poles and zeros,

thereby creating a new root locus that goes through the
desired closed-loop pole location.
Instructor: Mr. Biruk Simaani (M.Sc.) 27
6.3.2. Compensator Design
❖ There are two types of compensators:
(a) Pure Integration (1/s): Used for improving steady-state error
(b) Pure Differentiation (s): Used for improving transient response
❖ Both pure integration and differentiation compensators are ideal compensator
Example: Electromechanical ideal compensators, such as tachometers, are
often used to improve transient response
❖Other types of compensators include a lead and lag compensator system.

Instructor: Mr. Biruk Simaani (M.Sc.) 28

 6.3.3. Ways of Designing Compensators
❖ There are two ways of designing a compensator.
❖ The compensators will generally:
1) Improve the steady state error, by means of adding additional
poles or integrators.
2) Improve the transient response by means of adding additional
zeros or differentiators.

Instructor: Mr. Biruk Simaani (M.Sc.) 29

 1. Compensator Design: Improving the Steady-State Error
❖ This method improves the steady-state error.
❖ It uses:
a) Ideal Integral Compensation.
❑ Uses a pure integrator to place an open-loop, forward path
pole at origin.
❑ This will increase the system type and reduce the steady-state
error to zero.
b) Integral Compensation
❑ Does not use pure integrator.
❑ It place the poles near the origin. However, this won’t bring
the steady-state error to zero, but still, you get a considerable
reduction in error.

Instructor: Mr. Biruk Simaani (M.Sc.) 30

 Design of Ideal Integral Compensation
❖ Consider the following un-compensated system:

❖ Here, this system has a desirable transient

response which is generated by the closed-
loop poles at A.
❑ In this case, we see that the angular
contribution to this pole is 180°.

Instructor: Mr. Biruk Simaani (M.Sc.) 31

 ❖ If we add a pole to increase the system type, the
angular contribution to point A is no longer 180°,
and the root locus no longer goes through A.
❖ Adding a pole alone will alter the shape of the
root locus.
❖ And worst, the “new” root locus does not go
through point A. And this is what we do not want.
❖ Thus, to solve this problem, we add a
zero close to the pole at the origin.

Instructor: Mr. Biruk Simaani (M.Sc.) 32

 ❖ By adding extra zeros, the angular contribution of
the poles and zeros cancel out.
❖ And furthermore, point A is now on the root locus.
❖ And we note that the system type has been
increased.
❖ Therefore, the steady-state error is
reduced without appreciably affecting
the transient response.

Instructor: Mr. Biruk Simaani (M.Sc.) 33

 Design of Lag Compensation
❖ We note that the design of ideal integral compensation involves the use of
pure integrator.
❖ Now, this time, we will not use pure integrator.
❖ We will use poles and zeros that are close to the origin, but not necessarily
on the origin.
Let us consider a type 1 uncompensated system which is shown below:

Instructor: Mr. Biruk Simaani (M.Sc.) 34

 ❖ The above system can be improved by adding a compensator transfer function in
the feed-forward section of the loop

❖ The pole-zero plot of this compensator is shown below:

Instructor: Mr. Biruk Simaani (M.Sc.) 35

 ❖ How will this effect on the transient response of the system? And how will this
effect on the required gain, K?
Let us consider the root locus plot of the system:

Instructor: Mr. Biruk Simaani (M.Sc.) 36

 ❖ We establish here that:
❑ After the additional of compensated zeros and pole in Figure (b) of the
previous slide, the angular contributions are approximately zero, and
therefore, point P is still approximately on the location of the root locus.
❑ The gain K is still the same for both compensated and uncompensated
system.
❖ It can be concluded that in order to keep the transient response unchanged, the
compensator pole and zero must be close to each other; and the pair must also be
closed to the origin.
❖ This method, even though it is an approximate method, but it will still improve the
steady-state error.

Instructor: Mr. Biruk Simaani (M.Sc.) 37

 2. Compensator Design: Improving the Transient Response
❖ Recall that differentiators are the zeros in the s-domain.
❖ This technique will improve the transient response of the system.
❖ Usually, we wanted a desirable percent overshoot and shorter settling time.

❖ There are two techniques:

a) Design of ideal derivative compensation.
b) Design of lead compensation.

Instructor: Mr. Biruk Simaani (M.Sc.) 38

 Design of Ideal Derivative Compensation
❖ The ideal derivative is a pure differentiator.
❖ In designing an ideal compensation system, a pure differentiator is added
to the forward path of the feedback control system.
❖ The reason for this is to ensure that the closed-loop pole is placed on the
root locus.

❖ We have to remember that we wanted to design a system with better

transient response. Therefore, we need to have a dominant pole of the
characteristic equation.
❖ A dominant pole of the characteristic equation is usually a complex root.

Instructor: Mr. Biruk Simaani (M.Sc.) 39

 Recall – Dominant Poles
❖ A dominant pole is a complex root of characteristic equation.
❖ The general form of a dominant pole is given as follows, along with the
sketch of the dominant pole:

Instructor: Mr. Biruk Simaani (M.Sc.) 40

 Design of Ideal Derivative Compensation
❖Consider the following root locus
plot of an uncompensated system.

❖Pay close attention to the location of

the open-loop poles and the closed
loop poles.

Instructor: Mr. Biruk Simaani (M.Sc.) 41

❖ We can improve the transient response of this
system by adding a compensator zero at -2.

❖ Now, pay attention to the location of the open

loop pole and the closed loop poles.

❖ The closed-loop poles is now a dominant pole,

hence better transient response.

Instructor: Mr. Biruk Simaani (M.Sc.) 42

 ❖ We now add a compensator zero at -3.
Hence:

❖ And we now look at the location of the poles

(open-loop and the closed-loop).

Instructor: Mr. Biruk Simaani (M.Sc.) 43

 Let us compare the predicted characteristics of the system described in previous slides:

Instructor: Mr. Biruk Simaani (M.Sc.) 44

 The following observations can be made:
❖ Each of the compensated cases has dominant poles with the same damping
ratio as the uncompensated system. Therefore, we expect to have same
percent overshoot for all cases.
❑ The compensated, dominant, closed-loop poles with more
negative real parts will have the shortest settling time.
❑ The value of the imaginary part of the compensated system is
larger. Thus, all compensated system will have smaller peak time
compared to the uncompensated system.

Instructor: Mr. Biruk Simaani (M.Sc.) 45

 Time response of compensated and uncompensated systems:

Instructor: Mr. Biruk Simaani (M.Sc.) 46

Design of Lead Compensation
❖ In the design of lead compensation system, we make use of both compensator
pole and zero.
❖ We note here that if the pole is farther from the imaginary axis than the zero,
the angular contribution will still be positive, and thus approximate an
equivalent single zero.
❖ The following is the geometry of lead compensation.

Instructor: Mr. Biruk Simaani (M.Sc.) 47

 ❖ Rotating the compensator angle about the desired closed-loop pole location will
lead us to realize that an infinite number of lead compensators could be used to
meet the transient response requirement.

Instructor: Mr. Biruk Simaani (M.Sc.) 48

 Phase Compensators
❖ There are two types of phase compensators:
(a) Phase lag compensator.
(b) Phase lead compensator
(a) Phase lag compensator: It is use to reduce the steady-state errors.

Instructor: Mr. Biruk Simaani (M.Sc.) 49

 (b) Phase lead compensator: It is used to improve transient performance.

Instructor: Mr. Biruk Simaani (M.Sc.) 50

Example: A control system is described by the following block diagram:
The transfer function of the plant is:

Design a controller (compensator) so that the closed loop pole has an

overshoot of 5% and a settling time of 2 seconds.
Solution
According to the question, we desire:

From these given values, we can calculate the damping ratio and the natural
frequency.

Instructor: Mr. Biruk Simaani (M.Sc.) 51

 Next, we form a second order characteristic equation:

Solving the above characteristic (quadratic) equation yields a desired dominant poles:

Now, we look at the given plant transfer function.

The transfer function consists of integrators (having a pole at s=0 and s=-1).
Therefore, a suitable controller (compensator) would be a PD controller
because of the fact that PD will improve the transient response of the system.

Instructor: Mr. Biruk Simaani (M.Sc.) 52

 The general transfer function for a PD controller is described as follows:

❖ Recall that a PD controller is an example of a leading system in which we introduce

an extra zero that could provide a better transient response.
❖ And in a leading system, |𝑧| < |𝑝| and we can approximate as single zero .
❖ First, we sketch the pole-zero map of the plant transfer function

Instructor: Mr. Biruk Simaani (M.Sc.) 53

 Note:
❖We place the compensator zero anywhere on the s plane for as long as we meet the
condition that |z| <|p|. Or put it in another way, |p| > |z|. Therefore, in here we
see that the zero must be to the left of all the pole of the plant transfer function.
❖Point A is the location of the dominant poles which give the desired transient
response. In this case, at point A, s = -2 ± j2.

❖Let us calculate the angular contribution of θ1 and θ2

❖Note that in order for the dominant poles (A) to be located in the root locus, each
poles and zeros must contribute some angles to point A.
Instructor: Mr. Biruk Simaani (M.Sc.) 54
 Now, the contribution of the compensator angle can be calculated by using the
following formula:

Normalizing the angle gives:

Instructor: Mr. Biruk Simaani (M.Sc.) 55

 Now, once we know the angle, what do we do next? We refer to the pz-map in the
previous slide. What we did just now was to calculate the angular location of the
compensator zero. We now wish to determine the location of the compensator zero
on the pz-map.

Instructor: Mr. Biruk Simaani (M.Sc.) 56

 ❖ What we did so far was to calculate the location of the angular contribution of the
compensator zero.
❖ Once we know how to calculate the angular contribution, we can then use basic
and simple trigonometry to determine the exact (or approximate) location of the
compensator zero.
❖ Now, we know for a fact that zc = -3.
❖ We now proceed to calculate the gains Kp and Kd. To do this, we have to use the
magnitude condition

Instructor: Mr. Biruk Simaani (M.Sc.) 57

 We next do the following:

Instructor: Mr. Biruk Simaani (M.Sc.) 58

 ❖ We state that the computed values of Kp and Kd are 3 and 10 respectively.
❖ Since the gain of the plant transfer function is K=10, then we divide the values of
3 and 10 by 10, hence yielding

Therefore, the compensator transfer function is

Instructor: Mr. Biruk Simaani (M.Sc.) 59

 Assignment - II (10%)

1. Consider a unity feedback control system with

(Use Root Locus)

a) Design Gc = K such that the system operates at 10% overshoot.
Determine the settling time.
b) Design a compensator to decrease the settling time by a factor of 3
while operating at 10% overshoot.
c) Using MATLAB verify that the design specifications of part (b) are
satisfied.

Instructor: Mr. Biruk Simaani (M.Sc.) 60

 2. Consider a unity feedback control system with

(Use Open and closed loop Bode )

a) Design K such that the system operates at 5% overshoot. Determine

the settling time.
b) Design a lead compensator to decrease the settling time by a factor
of 2 while operating at 5% overshoot.
c) Using MATLAB simulation, compare the performance of the
compensated and uncompensated systems

Instructor: Mr. Biruk Simaani (M.Sc.) 60

 End of the Course
and
Thank you for your Attention

Instructor: Mr. Biruk Simaani (M.Sc.) 60