VICTORIA

UNIVERSITY
Controller Design
Using Root locus and
Robustness test
NEE3201 Introduction to Control
Systems

Submitted Via: VU
Collaborate
Student ID: 4462542
Submitted to: MS RUI LI
Lab Supervisor: MS RUI
LI
Date Submitted:
20/10/16

Table of Contents
1.0 Objective............................................................................ 3
2.0 Introduction........................................................................ 3
3.0 Design Specifications:........................................................3
4.0 Procedure........................................................................... 4
5.0 Analysis and Verification:.................................................10
5.1 Explanation:................................................................... 10
5.2 Controller and Compensator:.........................................10
5.3 Need for Pole zero Cancellation:....................................10
6.0 Controller design..............................................................11
6.1 Desired point coordinates calculation:...........................11
6.2 Angle of Deficiency:........................................................12
7.0 Lead Lag Compensator:...................................................12
7.1 Lead Compensator:........................................................12
7.2 Magnitude Condition to find the value of K:...................13
7.3 Lag Compensation:........................................................14
8.0 Robustness test:...............................................................15
9.0 Concluding Discussion.....................................................17
10.0 References:......................................................................18

Table of Figures
Figure 1 Block Diagram of the Field Controlled Servo Motor
Using Simulink........................................................................... 4
Figure 2 Root locus plot structure..............................................5
Figure 3 the graphical representation of root-locus design of
closed-loop system....................................................................5
Figure 4 Control & Estimation Tool Manager, Step Response.....6
Figure 5 Showing Change in Root Locus....................................6
Figure 6 Control and Estimation Design, with Integrator
response plot of Root Locus.......................................................7
Figure 7 the change of step response by changing the gain
constant in Gc(s) from 1 to 4......................................................7
Figure 8 Closed loop viewer and SISO design for estimation on
damping factor and closed loop poles.......................................8
Figure 9 Poles and Zeros of Lead Compensator for Angle
contribution............................................................................... 9
Figure 10 Poles and Zeros of Lag Compensator and change in
root locus..................................................................................9
Figure 11 Closed Loop Pole Viewer for Finalized Design...........10
Figure 12 System Output Response with Desired Specifications
................................................................................................ 10
Figure 13 Ideal Poles zeros from given specified Peak Overshoot
& Settling Time........................................................................ 11
Figure 14 Poles and Zeros s-plane angle deficiency calculations
................................................................................................ 12
Figure 15 Angle calculations for lead compensator Pole
placement...............................................................................13
Figure 16 Poles of Lead Compensator......................................13

Figure 17 Block Diagram of the FC servo motor with the

Controller design.....................................................................14
Figure 18 Output Response of the Final controller design with
the Plant.................................................................................. 14
Figure 19 Robust Root Locus....................................................16
Figure 20 Robust Root Locus Zoomed......................................16
Figure 21 Transient Responses.................................................16

1.0 Objective
The main purpose of this laboratory experiment is to use root-locus
method and design a unity feedback control system using MatLab. In
order to meet certain closed-loop system, time-domain performance
specifications and to analyse the behaviour of the obtained system in the
presence of uncertainty.
The goals are such that:

Use root-locus method and design a unity feedback control system

using MatLab.
Design a controller to closed loop performance specifications
including transient performance and steady error.
Use the angle to aid in the placement of a compensator poles and
zeros.
Robust analysis in the presence of 5% variation of the system
parameter values.

2.0 Introduction
Root-Locus method is very important method for designing closed-loop
control systems. Time-domain performance specifications are graphically
mapped in pole location regions in the complex plane, when all analysis
and design considerations are based on the well-accepted dominating
poles concept. Graphically it is related to the positions of closed-loop
poles to the positions of open-loop poles and open-loop zeros when a
tunable parameter varies. By taking advantage of the concept of
dominant closed loop-poles, root locus method allows transient response
of a closed-loop system to be shaped graphically and to design an
appropriate compensator in the s-plane.
Therefore, designing a controller, for meeting certain performance
specifications, by using the tools and the properties of the complex plane,
the construction of the so called gamma-regions widens the set of tools in
the plane, by giving new so called alternatives in studying plant
parametric uncertainties in designing controllers, classified in a category
tagged by the notion of robustness. [1]. In addition MATLAB script file is
used for the robustness analysis. The controller poles and zeros have been
manually calculated and are noted in this report.

3.0 Design Specifications:

In order to meet the design specifications for the given field-controlled
servomotor and its load

GP ( s )=

4500
( s+2 ) ( s +30 )

the following parameters has

to be achieved.
G P ( s )=

4500
( s+2 ) ( s +30 )

Such as the following specification to be satisfied:

1. The step response of the reference input has zero steady-state error.
2. Completely rejects step disturbances at steady state.
3. The velocity error constant (or coefficient) is

Kv

40.

4. The step response of the reference input has a 40 5% peak

overshoot.
5. The step response of the reference input has a 2% settling time of not
more than 3/4 of a second.
The reduced Block diagram for the plant is given as:

Figure 1 Block Diagram of the Field Controlled Servo Motor Using

Simulink

4.0 Procedure
In command window of MatLab, the transfer function of the plant can be
defined. MatLab uses sysg=tf(numg,deng) command for the purpose.
numg=4500;
deng=[1 32 60]

sysg=tf(numg,deng)
sysg=

4500
2
s +32 s + 60 s
2

Next step is to open the CAD tool, for such purpose ,the command rltool
at the MatLab prompt is used to invoke the CAD tool for root locus design,
given on Figure 1.

Figure 2 Root locus plot structure

Left click on the Control and Estimation Tools Manager in the task bar, in
the Control and Estimation tools Manager Window that appeared, we
selected Architecture at the top of the window with a left click, the left
click on System Data to replace the value of 1 in G, by sysg follows by
clicking OK. The graphical representation of root-locus is given on Figure
2.

Figure 3 the graphical representation of root-locus design of closedloop system

By selecting of Analysis Plots at the top of the Control and Estimation Tool
Manager Window and check the box for Closed Loop r to y then simply
pull down menu to select STEP for PLOT 1. This is the step response of the
closed- loop system defined by the system data in the specified
architecture. The plot is given on Figure 3.

Figure 4 Control & Estimation Tool Manager, Step Response

Right click in the plot area of the LTI Viewer window and then we selected
Grid by a left click in the menu that appears to check that the steady state
error is not zero, because of the settling time.
Left click on Graphic Tunning at the top of the Control And Estimation Tools
Manager window, then left click on the Plot 1 and then we used the right
most pull down menu to select root locus to able to see how a root locus
appears in the SISO Design Window. This is the root-locus of a unity
feedback control system, because H(s) =1, with plant transfer function
Gp(s) and a proportional controller Gc(s) in the forward path. There is a
small pink square located on the root locus. That is the position of the
closed loop-pole when K=1. We verify these by selecting Architecture
follows by the System data. With the LTI Viewer and SISO Design, which
are given on Figure 4 below, with left click on the pink area and drag it
with a small hand to a new position along the root-locus we were able to
observe the corresponding change of the closed-loop step response in the
LTI Viewer.

Figure 5 Showing Change in Root Locus

At the top of the Control and Estimation Tools Manager window, then left
click on Compensator Editor, given on Figure 5 below. Then right click on
the Dynamics area to add poles and zeros to the controller. Left click on
the Real pole option to change the position of the pole to be added from
-1 to 0, enter key must be pressed after to effect this change. Then the
controller has to be changed as well, from a proportional to integral
controller. Also the value of the constant term K in G c(s) to 1 has to
change. The desired Percentage Overshoot can be achieved at a value of
k=0.533

Figure 6 Control and Estimation Design, with Integrator response plot

of Root Locus

When the SISO Design has been opened we noticed that the system is
now a third order system instead of a second order system. Also, when
the LTI Viewer has been opened to observe the closed-loop step response,
it is noted that the required zero steady state error has been achieved.
By changing the value of the gain constant K in G c(s) from 1 to 0.53 we
noticed that the closed-loop step response in the LTI Viewer has been sped
up, given in Figure 6 below.

Figure 7 the change of step response by changing the gain constant in

Gc(s) from 1 to 4

The peak overshoot in the closed-loop step response has been notes that
is about 40% and the 2 settling time is about 4.62 sec, the damping
factor is 0.28 and the closed loop poles are at s=-0.839
considered idle closed loop poles are no longer

j2.9. the

s=5.33 j18.3 , because

the system is now 3rd order system. All results are given in Figure 7.

Figure 8 Closed loop viewer and SISO design for estimation on damping
factor and closed loop poles

At the top of the Control and Estimation Tools Manager window left click
on Compensator Editor the record of transfer function and the controller
has been made which is given on Figure 8.
The gain value at this stage:
k v =lim s
s 0

4500 k
4500 k
=
60
s ( s+2 )( s+30 )

k v =40 ;

4500 k 4500 k
2400
=
k=
=0.533
40 60 2400
4500

At the top of the Control and Estimation Tools Manager window left click
on the Compensator Editor. By Right clicking on the Dynamic area to add
poles and zeros to the controller GC(s), the left click on the Real Pole
option to change the position of the compensation pole to be added from
-1 to -29.426.

The previous step has been repeated, but left clicks the real zero option to
change the position of the compensation zero to be added from -1 to -2
shown on Figure 8.
The SISO Window has been open to observe the changes in the root-locus.

Figure 9 Poles and Zeros of Lead Compensator for Angle contribution

Again by adding poles and zeros for the lag compensator can meet the
desired requirements for velocity error constant and settling time, such
that a pole at 0.01 is placed and zero at 0.019 is placed as shown in
figures below, the desired P.O=40 5 and settling time 0.65 is achieved
by moving the pink square in root locus graph, the gain value is 3.6, the
given value for gain in the control and estimation tool manager is in time
constant form, however the calculated idle case value for K is 3.89

Figure 10 Poles and Zeros of Lag Compensator and change in root locus

The closed loop poles have been changed to

s=5.01 j19.

because of

adding the compensators, i.e. the root locus has been shifted. The
damping factor also been reduced to 0.255 because for Higher order
system the damping factordoes not follow the criteria for the Peak
overshoot.

Figure 11 Closed Loop Pole Viewer for Finalized Design

System response: The desired value of k is 3.6 the step response is

given in figure 11 below.

Figure 12 System Output Response with Desired Specifications

5.0 Analysis and Verification:

5.1 Explanation:
Using the root locus the controller design is in continuous domain and is
based on the approximation that the closed loop system has a complex
conjugate pole pair which dominates the system behaviour.

5.2 Controller and Compensator:

From previous studies and lab experiments different controllers such as PI
etc is generally used to improve the steady state performance where as
PD improve the relative stability or transient response. Similarly a lead
compensator improves the dynamic performance and lag compensator
improves the steady state response [2].

5.3 Need for Pole zero Cancellation:

Usually in designing controllers undesired poles or zeros of a plant transfer
function is cancelled by zeros and poles of controller, but the pole-zero

cancellation scheme does not always provide satisfactory solution.

Whereas new poles and zeros also has some advantageous locations. For
example if there is an undesired pole near j axis, inexact cancellation,
which is almost inevitable in practice, because it may lead to a marginally
stable or unstable closed loop system.

6.0 Controller design

The example considered with the following parameters:
K a=1 ;
K=50 ;

K t =1.5

T e=1 /30 ;
T m=1/2;

The goal is to design a controller giving closed-loop system performance

specified by overshoot 40 5 % and settling-time

3
4

2 %,.

From the parameter combination determines the following indirect

performance specifications in the complex plane.
i.e. Damping ratio = 0.279 for P.O=40; ,
Therefore, the idle root locus lies at point

n=18.3

n=5.33,

s=5.33 j18.3

6.1 Desired point coordinates calculation:

As the desired specification has not met yet, therefore, compensation is
needed for the achievement of the given specifications. Desired value, is
not possible by only altering the Evans-coefficient (the P controller gain) in
the right direction. It is necessary to design a controller that alters the root
locus, in order to ensure that its branches will contain the point specifying
the desired performance [1].
A possible solution may be found by integration. The additional integration
will ensure zero steady-state error to step input. The pole at origin of the
controllers will alter the current root locus, improve stability margins and
limit the system speed of response.

Figure 13 Ideal Poles zeros from given specified Peak Overshoot and
Settling Time

6.2 Angle of Deficiency:

The net angle contribution can be calculated using trigonometry from the
following figure 13.

Figure 14 Poles and Zeros s-plane angle deficiency calculations

1

1= tan

18.3
=36.6 ,
305.330
1

2=180 tan

3=180 tan

18.3
=100.3
5.332
18.3
=106.3
5.33

Angle of deficiency

d =180 (36.6 100.3 106.3 )

d =180 +243.2 =63.2

7.0 Lead Lag Compensator:

The compensator using root locus design can be given as

s+ z n
s+ p n ;

7.1 Lead Compensator:

For the additional angle contribution a lead compensator is required,
which will add pole and zero to the system, the location of these pole and
zero has an important effect on the system, therefore the location for
these pole and zero is very important. As the angle of deficiency is
calculated therefore the location of pole and zero can be found.
The zero characterizes the speed of response of the system. Its value is
determined by the performance specifying point, which the root locus
branches must contain. However in this case placing the pole at -2 will
cancel the pole and the root locus will shift.
Therefore, zero at -2 is chosen. Now the distance of the pole location is
required to determine from Point P. The calculated, angle of deficiency is
63.2 therefore; by simple geometry calculation of angles the desired
location can be found. Consider the given figure 14 below for the desired
pole location calculation.

Figure 15 Angle calculations for lead compensator Pole placement

p=100.3 63.2=37.1

Location of Pole from point P =

l p1=

18.3
=24.2
tan ( 37.1 )

Therefore the location of real axis from point P, 5.33 = 24.2 + 5.33= 29.5

Figure 16 Poles of Lead Compensator

Therefore

K ( s )=

k ( s+2 )
s ( s+29.5 )

7.2 Magnitude Condition to find the value of K:

s=5.33 1 8.3

Evaluate

at the current pole and zero location,

G(s ) K ( s)s=5.33+18.3 =1

4500 k (s +2)
s ( s+2 ) ( s+29.5 ) ( s +30 )

4500 k
=1
s ( s+29.5 ) (s +30) s=5.33 + j 18.3

=1

s=5.33+ j 18.3

||

4500 k
=1
5.33+ j18.3||5.33+ j18.3+29.5||5.33+ j 18.4 +30|

4500 k
=1
19.0630.2131.4
4500 k =16605.72 k =3. 925

Therefore,

k v =lim s
s 0

K ( s )=

( 3.696 4500 ) ( s+2 )

s (s+29.5)

17750.7( s+2)
17750.7
=
=20.057
s ( s+2 )( s+29.5 ) (s+30) 29.530

7.3 Lag Compensation:

The

required

velocity

compensation the

error

k v =20.057

constant

is

k v =

40,

with

the

lead

, therefore a lag compensator can be

introduced to put the poles in such a way that the required transient
specifications and velocity error constant specification can be achieved.
k required =40 ; k current =20.057
Therefore, the poles for the lag compensator can be placed as the ratio of
kv
kv;
the
required and current
i.e.
z2 k v
=
p2 k v

required

current

40
=1.9943 ;
20.057

z 2=1.9943p2
As Pole near origin is necessary at this stage [3], therefore, Lets choose a
z 2=1.99430.01=0.019943 0.02
pole at 0.01, therefore,
Finalised Controller block diagram with the Plant and step response is
given in figure 17 and figure 18 below.

Figure 17 Block Diagram of the Field Controlled servo motor with the
Controller design

Figure 18 Output Response of the Final controller design with the Plant

8.0 Robustness test:

The notion of robustness means insensitivity of the system in certain
degree to variations in the dynamic parameter values. The specification of
s= n j n
a -region [3], whose boundary is determined by equation

1 2
It sets a desired dynamic behaviour of the control system, and guarantees
robustness, in the case when all dominating poles, repositioned in the
plane due to plant parameter uncertainties, lie in it:
Plant parameter uncertainties form an uncertainty region in the complex
plane. The analysis of the relative location and overlapping of those
regions enables the interpretation of the robustness properties of control
systems. In the given sero mechanism motor, the uncertainty in the
control system is due to 5% variations of the system parameters, the
table of these variates values is given below [1].
Parameters
1.00
Ka

kt

te

tm
K

50

1.5

30

3.6

+5%=1.0
5
-5%=0.95
+5%=52.
5
-5%=47.5
+5%=1.5
75
5%=1.425
+5%=0.0
393
5%=0.027
3
+5%=0.5
25
5%=0.475
+5%=3.7

Peak
Overshoot

Settling Time

Rise Time

40

0.775

0.053

42.1

0.765

0.073

37.9

0.78

0.078

42.1

0.766

0.0729

37.9

0.78

0.078

42.1

0.765

0.0729

37.9

0.78

0.078

43.6

0.763

0.0774

35.9

0.701

0.0734

38.6

0.776

0.0775

41.2

0.762

0.0729

42.1

0.765

0.0729

8
-5%=3.42

37.9

0.78

0.078

Table 1, Parameter variation of 5%

The region of desired

lower bounds on the
performance in figure

damping ratio
evaluated to be n

performance gamma is specified by upper and

admissible deviations from the nominal system
16 below, The Lower and upper bounds on the
[0.255; 0.279] and the speed of response is

[4.48; 5.33]. As In the presence of 5% variation

of the system parameter values, the closed-loop characteristic equation

roots dominating the system transient response remain in the
performance specifying -region and the overall system possesses robust
properties in this sense.

Figure 19 Robust Root Locus

As we know a system is characterized by its poles and zeros in the sense

that they allow reconstruction of the root locus. The poles and zeros are
represented graphically by plotting their locations on the complex s-plane,
whose axes represent the real and imaginary parts of the complex
variable s. The location of the poles and zeros provide qualitative insights
into the response characteristic equation. The poles are represented by
cross x and zeros by o , provided in Figure 17 with the robust change
of s position on root locus.

Figure 20 Robust Root Locus Zoomed

In the time domain it is clearly showed that transient responses are

specified by admissible overshoot and settling-time for the entire range of
parameter deviations in figure 18.

Figure 21 Transient Responses

9.0 Concluding Discussion

This laboratory experiment was very useful in terms of understanding the
root-locus method, how to use Matlab prompt to see root-locus diagram to
record the poles and zeros. Also, by changing the location of poles and
zeros we were able to observe the step response of the closed loop
system, steady state error, settling time and peak overshot more than
40% and adjusted peak overshoot of 40 5 % with 2 3/4 sec settling
time. Also we had a closed-loop system with complex poles to observe the
step response of the closed loop system. The complex plane can be used
when evaluating the dynamic behaviour of control systems in the
presence of parametric uncertainties as an alternative to other common
approaches. It is relatively easy to form an uncertainty region, showing
the combinations of the parameter values which cause changes in system
characteristics. Root loci, and transient response plotted for different
dynamic parameter as Evans-gain show values of the respective

parameter, causing very low loss of stability. For the robustness test
analysis purpose, MatLab script is used for entering different parameters
to get the transient and root locus robustness outcome [1].
Other than that the system will tend to be stable with additional zeros.
However In many design exercises, zeros can be introduced to attract
closed-loop poles and alter the root locus location. It is also very useful to
applied stable pole-zero cancellation for improving system performance.
Therefore in conclusion, root locus method, in comparison with other
design methods, considerably facilitates the design of a controller
guaranteeing desired closed-loop system dynamic behaviour, since it
offers an adequate interpretation of the time-domain performance.

10.0 References:
[1 D. Kostov, V. Karlova and A. Todorov, Robust Root Locus Application In design
] and Analysis, Technical University, Sofia, 2008.
[2 B. Robert H and D. Richard C, Robust Control Systems, in Modern Control
] System, New Jersey, Printice Hall,, 2011, pp. 932-1956.
[3 N. Norman S, Design Via Root Locus, in Control Systems Engineerig,
] Pomona, Seventh Edition, 2015, pp. 449-485.
[4 R. H. B. Richard C. Dorf, Modern Control System, New Jersey: Printice Hall,
] 2011.
[5 N. S. Nise, Control System Engineering, Pamona: John Wiley & Son, 2008.
]
[6 E. Cheever, Linear Physical System Analysis, 2015. [Online]. Available:
] http://lpsa.swarthmore.edu/Root_Locus/RLocusExamples.html. [Accessed 10
10 2016].