Control Theory II Lecture by Prof. Dr. Yousif Al Mashhadany

University of Anbar Control Theory II
College of Engineering Prof. Dr. Yousif Al Mashhadany

Dept. of Electrical Engineering 2020 - 2021
Control Theory II
By
Prof. Dr.
Yousif Ismail Al Mashhadnay
2020 – 2021
i
Week
No Subject
(3 H/W)
Root Locus
 Root locus plot.
1.  General rules for constructing root loci. 1
 Special cases, conditionally stable system.
 Non-minimum phase systems.
Control system design by the root locus
2.  Effects of the addition of poles and zeroes.
3
 Lead compensator, lag compensator.
 Polar plots.
3. 2
 Nyquist stability criterion.
Bode diagram
4.  Bode plot (Logarithmic plot) 2
 Gain and phase margin.
Three term controller
 PD controller.
5. 2
 PI controller.
 PID controller.
 Introduction to sampled data systems and digital control
7. 2
systems
Analysis of control systems in state space
 Control system representation in s.s.
8.  Controllability, observability. 3
 Pole placement, design by pole placement.
 Ackermann's formula.
ii
References:
1) “ Linear Control System Analysis and Design with MATLAB ” by John
J. D’Azzo and Constantine H. Houpis , 2003.
2) “Classical Control Theory”, By Yousif Al Mashhadany, first Edition,

2014
3) “Modern Control Engineering” , By Katsuhiko Ogata, Fifth Edition,

2010
4) “Control Systems Engineering” , Sixth Edition, By Norman S. Nise, 2011
iii
Lecture No. Eight
Root Locus.
8.1. Introduction.
8.2. General Rules of Root Locus.
8.3. Examples.
4
8.1. Introduction.
To facilitate the application of the root-locus method, the following
rules are established for K > 0. These rules are based upon the
interpretation of the angle condition and an analysis of the characteristic
equation. These rules can be extended for the case where K < 0. The rules
for both K > 0 and K < 0 are listed in Sec. 7.16 for easy reference. The
rules presented aid in obtaining the root locus by expediting the plotting
of the locus. The root locus can also be obtained by using the MATLAB
program. These rules provide checkpoints to ensure that the computer
solution is correct. They also permit rapid sketching of the root locus,
which provides a qualitative idea of achievable closed-loop system
performance.
8.2. General Rules of Root Locus.

Rule 1: Number of Branches of the Locus:
The characteristic equation C.E.(s)=1+‫‏‬G(s)H(s)=0 is of degree n=m‫‏‬u;
therefore, there are n roots. As the open-loop sensitivity K is varied from
zero to infinity, each root traces a continuous curve. Since there are n
roots, there are the same numbers of curves or branches in the complete
root locus. Since the degree of the polynomial C.E.(s) is determined by the
poles of the open-loop transfer function, the number of branches of the
root locus is equal to the number of poles of the open-loop transfer
function.
Rule 2: Real-Axis Locus:
In Fig. 1 are shown a number of open-loop poles and zeros. If the angle
condition is applied to any search point such as s1 on the real axis, the
5
angular contribution of all the poles and zeros on the real axis to the left
of this point is zero. The angular contribution of the complex-conjugate
poles to this point is 360o. (This is also true for complex-conjugate zeros.)
Finally, the poles and zeros on the real axis to the right of this point each
contribute 180o (with the appropriate sign included). From Eq.(1) the
angle of G(s)H(s) to the point s1 is given by
- - - - - - (1)
Therefore, s1 is a point on a branch of the locus. Similarly, it can be

shown that the point s2 is not a point on the locus. The poles and zeros to
the left of
a point s on the real axis and the 360 o contributed by the complex-
conjugate
poles or zeros do not affect the odd-multiple-of-180o requirement. Thus, if
the total number of real poles and zeros to the right of a search point s on
the real
axis is odd, this point lies on the locus. In Fig.1 the root locus exists on the
real
axis from p0 to p1, z1 to p2, and p3 to z2. All points on the real axis
between z1 and p2 in Fig.1 satisfy the angle condition and are therefore
points on the root locus.
6
Fig.1. Determination of the real-axis locus.
However, there is no guarantee that this section of the real axis is part of
just one branch. Fig.2 a illustrate the situation where part of the real axis
between a pole and a zero is divided into three sections that are parts of
three different branches.
Fig.2.
7
Rule 3: Locus End Points:

The magnitude of the loop sensitivity that satisfies the magnitude
condition is given by Eq.(3) and has the general form in Eq.(4),
‫ ‏‏‏‬- - - - - (3)
‫ ‏‬- - - - - (4)
Since the numerator and denominator factors of Eq.(4) locate the poles
and zeros, respectively, of the open-loop transfer function, the following
conclusions can be drawn:
1) When s=pc (the open-loop poles), the loop sensitivity K is zero.
2) When s=zh (the open-loop zeros), the loop sensitivity K is infinite.
When the numerator of Eq.(4) is of higher order than the denominator,
then s=1 also makes K infinite, thus being equivalent in effect to a zero.
Thus, the locus starting points (K=0) are at the open-loop poles and the
locus ending points (K=1) are at the open-loop zeros (the point at infinity
being considered as an equivalent zero of multiplicity equal to the
quantity n - w).
8
Rule 4: Asymptotes of Locus as s Approaches Infinity.
Plotting of the locus is greatly facilitated by evaluating the asymptotes

approached by the various branches as s takes on large values. Taking the
limit of G(s)H(s) as s approaches infinity, based on Eqs.(5) and (6), yields
------- (5)
----------(6)
-------(7)
Remember that K in Eq.(7) is still a variable in the manner prescribed

previously, thus allowing the magnitude condition to be met. Therefore,
as s →∞, There are n-w asymptotes of the root locus, and their angles
are given by
----(8)
---------(9)
---(10)
9
Eq.(10) reveals that, no matter what magnitude s may have, after a

sufficiently large value has been reached, the argument (angle) of s on the
root locus remains constant. For a search point that has a sufficiently
large magnitude, the open-loop poles and zeros appear to it as if they had
collapsed into a single point. Therefore, the branches are asymptotic to
straight lines whose slopes and directions are given by Eq. (10) (see Fig.3).
These asymptotes usually do not go through the origin. The correct real-
axis intercept of the asymptotes is obtained from Rule 5.
s1=│s1│∟1‫‏‏‏‏‬
s2=│s2│∟2‫‏‏‬
.
.
sx=│sx│∟x‫‏‏‬
.
sx=∞∟ɤ
Fig. 3
Rule 5: Real-Axis Intercept of the Asymptotes
The real-axis crossing so of the asymptotes can be obtained by applying
the theory of equations. The result is
The asymptotes are not dividing lines, and a locus may cross its
asymptote. It may be valuable to know from which side the root locus
approaches its asymptote. The locus lies exactly along the asymptote if the
10
pole-zero pattern is symmetric about the asymptote line extended through

the point so. Rule 6: Breakaway Point
Rule 6: Breakaway Point on the Real Axis
The branches of the root locus start at the open-loop poles where K=0 and
end at the finite open-loop zeros or at s=1.When the root locus has
branches
Fig.4.
on the real axis between two poles, there must be a point at which the two
branches breakaway from the real axis and enter the complex region of
11
the s plane in order to approach zeros or the point at infinity. (Examples

are shown in Fig. 4.a-3: between p0 and p1, and in Fig. 4.b-2: between p2
and p3.) For two finite zeros (see Fig. 4.b-1) or one finite zero and one at
infinity (see Fig. 4.a-1) the branches are coming from the complex region
and enter the real axis. In Fig. 4.a-3 between two poles there is a point s a
for which the loop sensitivity K z is greater than for points on either side
of s a on the real axis.
In other words, since K starts with a value of zero at the poles and
increases in value as the locus moves away from the poles, there is a point
somewhere in between where the K’s for the two branches simultaneously
reach a maximum value. This point is called the breakaway point. Plots of
K vs. s utilizing Eq.(3) are shown in Fig.4 for the portions of the root locus
that exist on the real axis for K > 0. The point sb for which the value of K
is a minimum between two zeros is called the break-in point. The
breakaway and break-in points can easily be calculated for an open-loop
pole-zero combination for which the derivatives of W(s)=K is of the
second order. As an example, if
‫ ‏‏‏‏‏‏‬- - - - - - - - - - - - (11)
When s3+s2 +2s is a minimum, -K is a minimum and K is a maximum.

Thus, by taking the derivative of this function and setting it equal to zero,
the points can be determined:
12
Since the breakaway point’s s a for K > 0 must lie between s=0 and s=-1,
in order to satisfy the angle condition, the value is sa =-0:4257;
The other point, sb=-1.5743, is the break-in point on the root locus forK
<0.
Substituting sa= 0.4257 into Eq. (11) gives the value of K at the breakaway
Point for K > 0 as
When the derivative of W(s) is of higher order than 2,a digital-computer

program can be used to calculate the roots of the numerator polynomial
of dW(s)/ds; these roots locate the breakaway and break-in points. Note
that it is possible to have both a breakaway and a break-in point between
a pole and zero (finite or infinite) on the real axis, as shown in Figs.4a-1,
7.11a-2,
and 4.b-3. The plot of Kvs. s for a locus between a pole and zero falls into
one of the following categories:
1. The plot clearly indicates a peak and a dip, as illustrated between p1
and z1 in Fig. 4.b-3. The peak represents a ‘maximum’ value
of K that identifies a break-in point.
2. The plot contains an inflection point. This occurs when the breakaway
and break-in points coincide, as is the case between p2 and z1 in
Fig. 4a-2.
13
3. The plot does not indicate a dip-and-peak combination or an inflection

point. For this situation there are no break-in or breakaway points.
The next geometrical shortcut is the rapid determination of the direction
in which the locus leaves a complex pole or enters a complex zero.
Although in Fig.5.a complex pole is considered, the results also hold for a
complex zero.
In Fig.5.a, an area about p2 is chosen so that l2 is very much smaller than
l0, l1, l3, and (l )1. For illustrative purposes, this area has been enlarged
many times in Fig.5b. Under these conditions the angular contributions
from all the other poles and zeros, except p2, to a search point anywhere
in this area are approximately constant. They can be considered to have
values determined as if the search point were right at p2.Applying the
angle condition to this small area yields
Fig. 5
14
In a similar manner the approach angle to a complex zero can be

determined. For an open-loop transfer function having the pole-zero
arrangement shown in Fig.6, the approach angle c1 to the zero z1 is given
by
In other words, the direction of the locus as it leaves a pole or approaches
a zero can be determined by adding up, according to the angle condition,
all the
angles of all vectors from all the other poles and zeros to the pole or zero
in question. Subtracting this sum from (1+‫‏‬2h)180o gives the required
direction.
Fig. 6
Rule 7: Complex Pole (or Zero): Angle of Departure:
The next geometrical shortcut is the rapid determination of the direction

in which the locus leaves a complex pole or enters a complex zero.
Although inFig.7.a complex pole is considered, the results also hold for a
complex zero.
In Fig. 7.a, an area about p2 is chosen so that l2 is very much smaller than
l0, l1, l3, and (l) 1. For illustrative purposes, this area has been enlarged
many times in Fig. 7.b. Under these conditions the angular contributions
from all the other poles and zeros, except p2, to a search point anywhere
15
in this area are approximately constant. They can be considered to have

values determined as if the search point were right at p2.Applying the
angle condition to this small area yields. In a similar manner the
approach angle to a complex zero can be determined. For an open-loop
transfer function having the pole-zero
16
Arrangement shown in Fig.8, the approach angle c1 to the zero z1 is given

by
In other words, the direction of the locus as it leaves a pole or approaches

a zero can be determined by adding up, according to the angle condition,
all the angles of all vectors from all the other poles and zeros to the pole
or zero in question. Subtracting this sum from (1+‫‏‬2h)180 gives the
required direction.
Rule 8: Imaginary-Axis Crossing Point:

In cases where the locus crosses the imaginary axis into the right-half s
plane,
the crossover point can usually be determined by Routh's method or by
similar means. For example, if the closed-loop characteristic equation
D1D2‫‏‬+ N1N2=0 is of the form.
17
An undamped oscillation may exist if the s_1 row in the array equals zero.
For this condition the auxiliary equation obtained from the s 2 row is
- - - - - - - - - - - (12)
The loop sensitivity term K is determined by setting the s1 row to zero:

K =bc/d
For K > 0, Eq. (12) gives the natural frequency of the undamped
oscillation. This corresponds to the point on the imaginary axis where the
locus crosses over into the right-half s plane. The imaginary axis divides
the s plane into stable and unstable regions. Also, the value of K from Eq.
(K =bc/d) determines the value of the loop sensitivity at the crossover
point. For values of K < 0 the term in the s0 row is negative, thus
characterizing an unstable response. The limiting values for a stable
response are therefore
0 < K <bc/d
In like manner, the crossover point can be determined for higher-order

characteristic equations. For these higher-order systems care must be
exercised in analyzing all terms in the first column that contain the term
K in order to obtain the correct range of values of gain for stability.
Rule 9: Intersection or Non-intersection of Root-Locus Branches:

The theory of complex variables yields the following properties:
1. A value of s that satisfies the angle condition of Eq. (1) is a point on the
root locus. If dW(s)/ds ≠ 0 at this point, there is one and only one branch
of the root locus through the point.
2. If the first y_1 derivatives of W(s) vanish at a given point on the root
locus, there are y branches approaching and y branches leaving this
point; thus, there are root-locus intersections at this point. The angle
between two adjacent approaching branches is given by
18
Also, the angle between a branch leaving and an adjacent branch that is
approaching the same point is given by
Fig. 9. illustrates these angles at s=-3 ,with θy=45o and λy=90o.
19
Ex.(1).
20
21
22
23
Ex.(2). Plot Root Loci for system shown below:
Solution:
s=-2.366 s=-0.634
24
Ex.(3). Plot the root loci for the system has the following T.F:
Solution:
θ=-54.5o
θ=60o
s=-2.26
s=0.45
s=- 2/3
θ=54.5o
25
Solution:
26
Solution:
27
Solution:
28
Solution:
29
Chapter Ten
Frequency Response Analysis

(Polar Plot & Nyquist
Stability Criterion).
The outline of the chapter is as follows:
10.1. Review of Frequency Response
10.2. Sketch the Polar plot of Frequency Response
10.3. Polar plot for some standard system
10.4. General Shape of polar plot
10.5. Nyquist Stability Criterion.
10.6. Solved Problem
48
10.1. Review of Frequency Response
To specify the behavior of a system to a sinusoidal input at a particular angular frequency ω

i.e. the frequency response G(jω), we need both the magnitude |G(jω)| and phase φ of
G(jω).Recall that both the magnitude and phase are functions of the input frequency and
therefore vary with ω. As we have seen when calculating the magnitude and phase during
the section of our course on Bode Diagrams, G(jω) is a complex number and can be
represented in the form x+jy or the rectangular form, where x and y are the real and the
imaginary parts, respectively. An alternative to represent G(jω) (calculated at frequency ω)
is the use of a polar plot, as shown in Figure.10.1.
Figure.10.1 .Polar plot representation
In the above diagram, the output, for a unit amplitude sinusoidal input at frequency ω, has
magnitude |G(jω)| (i.e., the length of the vector ) and a phase angle φ.
10.2. Sketch the Polar plot of Frequency Response
To sketch the polar plot of G(jω) for the entire range of frequency ω, i.e., from 0 to infinity,
there are four key points that usually need to be known:
1) the start of plot where ω = 0,

2) the end of plot where ω = ∞,
49
3) where the plot crosses the real axis, i.e., Im(G(jω)) = 0, and
4) where the plot crosses the imaginary axis, i.e., Re(G(jω)) = 0.
Consider a first order system

1
G( s) 
1  s
Where: τ is the time constant

Representing G(s) in the frequency response form G(jω):
1
G ( j ) 
1  j
Multiplying both numerator and denominator by the conjugate of denominator, i.e.,

(1−jωτ) , one has
1  j 1  j
G ( j )    j
1  
2 2
1  
2 2
1   2 2
The magnitude of G(jω), i.e., G(jω), is

1  j 2 1
| G( j ) | ( )2  ( ) 
1   2 2
1  
2 2
1   2 2
And the phase of G(jω), denoted by φ, is

 t
(1   2 2 )
  tan ( 1
 tan 1 ( t / 1)
1
(1   2 2 )
Now that we have expressions for the magnitude and phase of the frequency response, we
can sketch the polar plot using the 4 key points.
Point 1: The start of plot where (ω=0)
1
| G ( j ) |  1,
1 0
  tan 1 (0 / 1)  0
Point 2: The end of plot where (ω=∞)
1
| G ( j ) |  0,
1 
50
  tan 1 ( / 1)  90
Point 3: Where the plot crosses the real axis, i.e., Im(G(jω)) = 0.
 j
 0    0 &  
1   2 2
These are the same as Points 1 and 2.

Point 4: Where the plot crosses the imaginary axis, Re(G(jω)) = 0.
1
0  
1   2 2
This is the same as Point 2.

As Point 3 coincides with Points 1 and 2 and Point 4 coincides with Point 2, we need more
values of G(jω) evaluated at different frequencies. Taking ω=1/τ, then one has
1 1
| G ( j ) |  ,
11 2
  tan 1 (1 / 1)  45

Sketching the polar plot for 1/(1+τs), one has
Figure.10.2. Sketching the polar plot for 1/(1+τs)
51
Example1: Sketch the polar plot for a third order system with a transfer function of
1
G( s) 
2 s  3s 2  3s  1
3
Solution:
Replacing s in the above transfer function with jω, one has the frequency response
1
G ( j ) 
2( j )  3( j ) 2  3 j  1
3
1 (1  3 2 )  j (3  2 3 )
G( j )  
(1  3 2 )  j (3  2 3 ) ((1  3 2 ))2  ((3  2 3 ))2
The magnitude and phase of G(jω) are then obtained as:
| G( j) | (Re l (G( j)))2  (Im g (G( j)))2
1
| G( j ) |
((1  3 2 ))2  (3  2 3 ) 2
Im g (G ( j )
  tan 1
Re l (G ( j )
 (3  2 3 )
  tan 1 ( )
1  3 2
H.W. Sketching the Polar plot?

In Matlab we can plot the Nquist by using the following instructions
>> num=1
num =
1
>> den=[2 3 1 1]
den =
2 3 1 1
>> sys=tf(num,den]
>> [re,im,w] = nyquist(sys)
52
>> nyquist(sys)
Nyquist Diagram
6
0 dB
2 dB -2 dB
2
Imaginary Axis
4 dB -4 dB
6 dB -6 dB
10 dB -10 dB
0
-2
-4
-6
-4 -3 -2 -1 0 1 2 3
Real Axis
Figure.10.3. Nyquist plot of the given example
10.3. Polar plot for some standard system.
The polar plot of sinusoidal transfer function G(jω) is a plot of the magnitude of G(jω)
versus the phase angle of G(jω) on polar coordinates as ω is varied from zero to infinity.
Therefore it is the locus of as ω is varied from zero to infinity.
Note that in polar plot a positive (negative) phase angle is measured counterclockwise
(clockwise) from the positive real axis. The polar plot is often called Nyquist plot. An
example of such a plot is shown in Figure.10.1 each point on the polar plot of G(jw)
represents the terminal point of a vector at a particular value of w. in polar plot, it is
important to show the frequency graduation of the locus. The projections of G(jw) on the
real and imaginary axes are its real and imaginary components. Both the magnitude |G(jw)|
and phase angle ˂G(jw) must be calculated directly for each frequency w in order to
construct polar plots. Since the logarithmic plot is easy to construct ,however ,the data
necessary for plotting the polar plot may be obtained directly from the logarithmic plot if the
latter is drawn first and decibels are converted into ordinary magnitude. Or, of course
53
Matlab may be used to obtain a polar plot G(jw) and ˂G(jw) accurately for various values
of w in the frequency range of interest. An advantage in using a polar plot is that its depicts
the frequency response characteristics of a system over the entire frequency range in a single
plot. One disadvantage is that the polar does not clearly indicate the contributions of each
individual factor of the open loop transfer function.
Figure.10.4. Nyquist plot

Integral and derivative factors ( j ) 1 .the polar plot of G(jw)=1/jw is the negative imaginary
axis since
1 1 1
G ( j )    j    90
j  
The polar plot of G(jw)=jw is the positive imaginary axis.
Figure.10.5. Nyquist plot of jw and 1/jw

First order factor (1  j ) .
1
The sinusoidal transfer function
54
1 1
G ( j )   tan 1 (T )
1  jT 1  T
2 2
The value of G ( j ) at   0 is
1 1
G ( j )   tan 1 (0)  10
1 0 1 0
The value of G ( j ) at   1/ T is
1 1 1
G ( j )   tan 1 (1)    45
11 11 2
Figure.10.6. a) Nyquist plot of (1/1+jwT); b) plot of G(jw) in X-Y plane
H.w. Prove the polar plot is a semicircle and is plot in Figure.10.6.b.

Quadratic factor [1  2 ( j / n )  ( j / n )2 ]1 : the low and high pass frequencies portion of the
polar plot of the following sinusoidal transfer function
1
G ( j )  , for   0
1  2 ( j / n )  ( j / n ) 2
Are given respectively by

lim0 G( j )  10 & lim G( j )  0  180
55
Figure.10.7. Polar plot of the quadratic factor

Example 2. Consider the following second order system
1
G( s) 
s(1  Ts )
Sketch the polar plot of this system

Solution:
Since the sinusoidal transfer function can be written as
1
G ( j ) 
j (1  jT )
T 1
G ( j )    j
(1   2T 2 )  (1   2T 2 )
The low frequency portion of the polar plot becomes

lim0 G( j )  T  j    90
The high frequency portion of the polar plot becomes

lim G( j )  0  j0  0  180
The general shape of the polar plot can be shown in Figure 10.8
56
Figure.10.8. Polar plot of 1/[jw(1+jwT)]

 jT
Transport lag factor G( j )  e can be written as
G ( j )  1 cosT  j sin T
The polar plot of the transport lag can be illustrated in Figure.10.9.
Im
1
0 Re
w =0
w
increase
Figure.10.9. Polar plot of transport lag
57
The following table represents the polar plot for some simple transfer functions:
Im Im
w =∞ w =∞
1
Re 0 Re
w w
1 ↓ 1 + 𝑗𝑤𝑇 ↓
𝑗𝑤 0 𝑗𝑤𝑇 0
Im Im
∞ ∞
↑ 1+jw ↑
jw w T w
w =0 w =0
0 Re 0 1 Re
Im
1 Im
𝑗𝑤𝑇
(𝑗𝑤𝑇)2
w =∞ 1 + 𝑗𝑤𝑇
w=∞
0← 0 Re
w 0 1 Re
w =0
Im Im
w= w=
1/ 1 ∞ 1
0a ∞ 0
Re w =0 Re
w =0
1 + 𝑗𝑤𝑇 1
1 + 𝑗𝑤 2 𝑇 (1 + 𝑗𝑤𝑇1 )(1 + 𝑗𝑤𝑇2 )(1 + 𝑗𝑤𝑇3 )
𝑤𝑛2 Im 1 + 𝑗𝑤𝑇1 Im
𝑗𝑤[(𝑗𝑤)2 + 2𝜁𝑤𝑛 (𝑗𝑤) + 𝑤𝑛2 ] w= 𝑗𝑤(1 + 𝑗𝑤𝑇2 )(1 + 𝑗𝑤𝑇3 ) w=
0 ∞ ∞
Re Re
0
w
w
Figure.10.10. Polar plot of some transfer functions
58
10.4. General Shape of polar plot
The polar plot of a transfer function of the form
K (1  jTa )(1  jTb )....... b0 ( j ) m  b1 ( j ) m1  b2 ( j ) m2 .....

G ( )  
( j )  (1  jT1 )(1  jT2 )..... a0 ( j ) n  a1 ( j ) n 1  a2 ( j ) n 2 ....
Where n>m or the degree of the denominator polynomial is greater than the degree
numerator polynomial, will have the following properties:
1. For λ=0 or type zero systems: the starting point of the polar plot which is at ω=0, is finite
and is on positive real axis. The tangent to the polar plot at ω=0 is perpendicular to the
real axis. The terminal point which is on ω=∞ is at the origin and the curve is tangent to
the one of the axis.
2. For λ=1 or type one systems: the jω term is in the denominator contributes -90 to the
total phase angle of G(jω) for 0    . At ω=0, the magnitude of G(jω) is infinity ,and
the phase angle becomes (-90).At low frequencies ,the polar plot is asymptotic to a line
parallel to the negative imaginary axis. At ω=∞, the magnitude of G(jω) is zero and the
curve converges to the origin and is tangent to the one of the axis.
3. For λ=2 or type two systems: the (jω)2 term is in the denominator contributes -180 to the
total phase angle of G(jω) for 0    . At ω=0, the magnitude of G(jω) is infinity ,and
the phase angle becomes (-180).At low frequencies ,the polar plot is asymptotic to a line
parallel to the negative imaginary axis. At ω=∞, the magnitude of G(jω) is zero and the
curve converges to the origin and is tangent to the one of the axis.
The general shape of polar plot for different types of the system can be shown in
Figure.10.11. There are two types of stability criterions by Nyquist concept Simplified
stability and General stability:
Simplified Stability criterion: It is depend on the path of Nyquist graph where if the Nyquist
graph passes to the right of point ( -1,0 ); [ i.e. The point lie to left of the plot as shown
Figure.10.12.a. below] then the system is stable. Otherwise the system unstable
(Figure.10.12.b. )
59
Figure.10.11. Polar plot of type 0, type 1 and type 2 systems
Figure.10.12. Nyquist stability criterion
General Stability Criterion:

The criterion can be expressed as
Z NP
Z is the zeros of 1+G(s)H(s) in the right half of the s-plane
N is the number of clockwise encirclements of -1+j0
P is the poles of G(s)H(s) in the right half of the s-plane.
If P is not zero, for stable system, we must have (Z=0 or N= -P) which means we must have
P counterclockwise encirclements of -1+j0 point. If G(s)H(s) does not have any poles in the
right half s plane ,then Z=N. thus, for stability there must be no encirclement of the -1+j0
60
point by the G(s)H(s) locus. In this case it is not necessary to consider the locus for the
entire jω axis ,only for the positive frequency portion. Therefore; the stability in this
criterion can be summarized as: If Z = 0
the system is stable and If Z > 0 the system is unstable. Note: If the encirclement in
clockwise N it has positive value and in counterclockwise encirclements N is negative. We
present only the essence of the Nyquist stability criterion and define the phase and gain
stability margins. The Nyquist method is used for studying the stability of linear systems
with pure time delay.
For a SISO feedback system the closed-loop transfer function is given by:
G( s)
M ( s) 
1  G( s) H ( s)
Where G(s): represents the system and H(s): is the feedback element.
Since the system poles are determined as those values at which its transfer function becomes
infinity, it follows that the closed-loop system poles are obtained by solving the following
equation:
1  G ( s ) H ( s )  0  ( s )
which, in fact, represents the system characteristic equation.
10. 5. Nyquist Stability Criterion
A stability test for time invariant linear systems can also be derived in the frequency
domain. It is known as Nyquist stability criterion. It is based on the complex analysis result
known as Cauchy’s principle of argument. Note that the system transfer function is a
complex function. By applying Cauchy’s principle of argument to the open-loop system
transfer function, we will get information about stability of the closed-loop system transfer
function and arrive at the Nyquist stability criterion (Nyquist, 1932).
The importance of Nyquist stability lies in the fact that it can also be used to determine the
relative degree of system stability by producing the so-called phase and gain stability
61
margins. These stability margins are needed for frequency domain controller design
techniques.
D( s )  1  G ( s ) H ( s )
whose zeros are the closed loop poles of the transfer function. In addition, it is easy to see
that the poles of D(s) are the zeros of M(s). At the same time the poles of D(s) are the open-
loop control system poles since they are contributed by the poles of H(s)G(s), which can be
considered as the open-loop control system transfer function obtained when the feedback
loop is open at some point. The Nyquist stability test is obtained by applying the Cauchy
principle of argument to the complex function D(s). First, we state Cauchy’s principle of
argument. Cauchy’s Principle of Argument, Let F(s) be an analytic function in a closed
region of the complex plane given in Fig. below except at a finite number of points (namely,
the poles of F(s ). It is also assumed that F(s ) is analytic at every point on the contour. Then,
as travels around the contour in the - plane in the clockwise direction, the function F(s )
encircles the origin in the(Re{F(s)},Im{F(s)})-plane in the same direction times (see Figure
below), with N given by:
N=Z-P
where Z and P stand for the number of zeros and poles (including their multiplicities) of the
function F(s) inside the contour. The above result can be also written as:
Arg{F(s)}=(Z-P)2π
which justifies the terminology used, “the principle of argument”.
Figure.10.13. Cauchy’s principle of argument

The Nyquist plot is a polar plot of the function D( s )  1  G ( s ) H ( s )
62
when travels around the contour given in Figure10.14:
Figure.10.14. Contour of the polar plot
Contour, the poles of on the imaginary axis must be encircled by infinitesimally small
semicircles. Nyquist Stability Criterion: It states that the number of unstable closed-loop
poles is equal to the number of unstable open-loop poles plus the number of encirclements
of the origin of the Nyquist plot of the complex function D(s). This can be easily justified by
applying Cauchy’s principle of argument to the function D(s) with the s-plane contour given
in Figure above. Note that Z and P represent the numbers of zeros and poles, respectively, of
D(s) in the unstable part of the complex plane. D(s)At the same time, the zeros of D(s) are
the closed-loop system poles, and the poles of D(s) are the open-loop system poles (closed
loop zeros).
The above criterion can be slightly simplified if instead of plotting the function
D(s)=1+G(s)H(s), we plot only the function G(s)H(s), and count encirclement of the Nyquist
plot of G(s)H(s) around the point (1,j0), so that the modified Nyquist criterion has the
following form.
The number of unstable closed-loop poles (Z) is equal to the number of unstable open-loop
poles (P) plus the number of encirclements (N) of the point (-1,j0) of the Nyquist plot of
G(s)H(s), that is:
Z=P+N
Phase and Gain Stability Margins: Two important notions can be derived from the Nyquist
diagram: phase and gain stability margins. The phase and gain stability margins are
presented in Figure 10.15:
63
Figure.10.15. Phase and gain stability margins
They give the degree of relative stability; in other words, they tell how far the given system
is from the instability region. Their formal definitions are given by:
Pm  180  arg{G( j ) H ( j )}
1
Gm [dB]  20 log
G ( j ) H ( j )
Example3: Consider a control system represented by:

1
G( s) H ( s) 
s( s  1)
Since this system has a pole at the origin, the contour in the -plane should encircle it with a
semicircle of an infinitesimally small radius. This contour has three parts (a), (b), and (c).
Mappings for each of them are considered below.
a) On this semicircle the complex variable (s) is represented in the polar form by
s  Re j with R   &(   / 2     / 2 ). Substituting s  Re j into G ( s ) H ( s ) ,we easily see
that G( s) H ( s)  0 .
Thus, the huge semicircle from the (s) plane maps into the origin in the G( s) H ( s) plane as in
the Figure.10.16.
64
Figure.10.16. Nyquist plot for the example above
b) On this semicircle the complex variable (s) is represented in the polar form by
j
s  re j with r  0 &(   / 2     / 2 ) so that we have G( s) H ( s)  1 / re   x arg(  ) . Since
ϕ changes from –π/2 at point A to π/2 at point B, arg{G(s)H(s)} will change from (–π/2 to
π/2). We conclude that the infinitesimally small semicircle at the origin in the s-plane is
mapped into a semicircle of infinite radius in the G(s)H(s)-plane.
c) On this part the contour s takes pure imaginary values, i.e, s  j with  changing from

  to  .due to symmetry, it is sufficient to study only mapping along o     we
can find the real imaginary parts of the function G( j ) H( j ),which are given by
1
Re{G ( j ) H ( j )} 
2  1
1
Im{G ( j ) H ( j )} 
 ( 2  1)
From these expressions we see that neither the real nor the imaginary parts can be made
zero , and hence the Nyquist plot has no points of intersection with the coordinate axis . for
  o  we are at point B and since the plot at   +  will end up at the origin ,the Nyquist
diagram corresponding to part (c) has the form as shown in Figure 10.17.
65
Note that the vertical asymptote of the Nyquist plot in Figure.10.17 is given by
Re{G( jo ) H ( jo )}  1 ,since at those point Im{G( jo ) H ( jo )}   ,from the Nyquist diagram
,we see that N=0,and since there is no open loop poles in the left hand side of the complex
plane, i.e (P=0),we have Z=0,so that the corresponding closed-loop system has no unstable
poles .
The Nyquist plot is drawn by using the MATLAB function Nyquist
num=1; den=[1 1 0];
nyquist(num,den);
axis([-1.5 0.5 -10 10]);
axis([-1.2 0.2 1 1]);
The MATLAB Nyquist plot is presented in Figure.10.17.
Nyquist Diagram
20
sys
15
10
Imaginary Axis
-5
-10
-15
-20
-1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0
Real Axis
Figure.10.17. MATLAB Nyquist plot for Example

It can be seen from two Figures previous, that 1 / Gm  0 , which implies that Gm   . Also,
from the same figures it follows that    .In order to find the phase margin and the
corresponding gain crossover frequency we use the MATLAB function margin as follows
[Gm,Pm,wcp,wcg]=margin(num,den)
producing, respectively, gain margin, phase margin, phase crossover frequency, and gain
crossover frequency. The required phase margin and gain crossover frequency are obtained
as, Gm = Inf, wcp=Inf, Pm= 51.8273, wcg=0.7862 rad/s.
66
10.6. Solved Problem
Prob.1. Consider now the following system, obtained from the one in the previous example
by adding a pole, that is:
1
G( s) H ( s) 
s( s  1)( s  2)
The contour in the -plane is the same as in the previous example. For cases (a) and (b) we
have the same analyses and conclusions. It remains to examine case (c). If we find the real
and imaginary
parts of G ( j ) H ( j ) ,we get
3 ;  (2   )
Re{G ( j ) H ( j )}  Im{G ( j ) H ( j )} 
9  ( 2   ) [9  ( 2   )]
It can be seen that an intersection with the real axis happens at   2 at the point
Re{ G ( j 2 ) H ( j 2 )} . The Nyquist plot is given in Figure.10.18. The corresponding Nyquist
plot obtained by using MATLAB is given in Fig. 10.18.
Figure.10.18. Nyquist plot for the prob.1.

Note that the vertical asymptotes is given by Re{G ( j 0) H ( j 0)}  3 / 4 ,thus we have N=0,P=0
and Z=0 so that the closed loop system is stable. The Matlab function margin is produce
Gm=6 dB,Pm=53.410,ωcg=0.445 rad/sec, ωcp=1.141rad/sec.
67
Prob.2. Check the stability of the following system

20
G( s) 
s(0.1s  1)(0.5s  1)
Ω Phase Magnitude
0 -90 ∞
∞ -270 0
20 20[0.6 2  j(1  0.05 2 )]

G( j )  
 0.6 2  j(1  0.05 2 ) 0.36 4   2 (1  0.05 2 )
Setting Im{G(jw)=0}
1  0.05 2  0 Thus , w = ± 4.47 rad / sec, G( j 4.47 ) = -1.667

The closed loop system is unstable. The characteristic equation has two roots in the right-
half s-plane. Nyquist Plot of G ( jw )
Figure.10.19. Nyquist plot of prob.2.

10
G( s) 
s(0.1s  1)(0.5s  1)
real axis by the G ( jw) plot is at -0.8333, and the corresponding w is 4.47 rad/sec. Thus, Z =
0. The closed-loop system is stable. Nyquist Plot of G ( jw ):
68

100( s  1)
G( s) 
s(0.1s  1)(0.2 s  1)(0.5s  1)
Ω Phase Magnitude
0 -90 ∞
∞ -270 0
100(1  j )
G ( j ) 
(0.01  0.8 2 )  j (1  0.17 2 )
4
100(1  j )[( 0.01 4  0.8 2 )  j (1  0.17 2 )]


(0.01 4  0.8 2 )  j (1  0.17 2 )
Setting Im G( jw) = 0 ,
0.01w4 -0 .8w2 -1+ 0.17w2 = 0
w4-63 w2 -100 = 0,
Thus, w 2 = 64. 55 & w = ±8.03 rad/sec.
G(j8.03)=-10
The closed-loop system is unstable
Nyquist Plot of G ( j w ):
69
Nyquist Diagram
500
Imaginary Axis
-500
-30 -25 -20 -15 -10 -5 0 5 10 15 20
Real Axis
Figure.10.21. Open loop Nyquist plot of prob.4.

10
G( s) 
s (0.2 s  1)(0.5s  1)
2
G( jw)  tan 1 0 / 10  180( for s 2 )  tan 1 0.2w  tan 1 0.5w

At w  0  G ( jw)  180
So the compete table will be
70
Ω Phase Magnitude
0 -180 ∞
∞ -360 0
10
G ( j ) 
(0.1 4   2 )  0.7 j 3
10[( 0.1 4   2  0.7 j 3 ]

(0.1 4   2 ) 2  0.49 6
Setting Im {G ( jw ) = 0}, w = ∞. The Nyquist plot of G ( jw) does not intersect the real axis
except at the origin where w = ∞.
The closed loop system is unstable. The characteristic equation has two roots in the right-
half s-plane.
Nyquist Plot of G ( jw ):
Nyquist Diagram
3
1
Imaginary Axis
-1
-2
-3
-2 -1.8 -1.6 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0
Real Axis
71
Nyquist Diagram
0.8
0.6
0.4
Imaginary Axis
0.2
-0.2
-0.4
-0.6
-0.8
-0.2 0 0.2 0.4 0.6 0.8 1 1.2
Real Axis
Figure.10.24.Closed loop Nyquist plot of prob.5.

0.1
G( s) 
s ( s  1)( s 2  s  1)
Ω Phase Magnitude
0 -90 ∞
∞ -360 0
0.1 0.1[( 4  2 2 )  j (1  2 2 )]
G ( j )   ;
( 4  2 2 )  j (1  2 2 ) ( 4  2 2 ) 2   2 (1  2 2 ) 2
Setting Im {G ( jw)} = 0, w = ∞ or w2 =0.5, w = ±0.707 rad / sec &

G( j 0.707 )= -0 .1333.
The closed loop system is stable.
72
Nyquist Diagram
5
Imaginary Axis
-5
-0.25 -0.2 -0.15 -0.1 -0.05 0
Real Axis
Prob. 7. Check the stability of the following system

10( s  10)
G( s) 
s( s  1)( s  100)
Ω Phase Magnitude
0 -90 ∞
∞ -180 0
10( j  10)
G ( j ) 
 101 2  j (100   2 )
10( j  10)[101 2  j (100   2 )]

10201 4   2 (100   2 ) 2
Setting Im {G ( jw)} = 0, w = 0 is the only solution. Thus the Nyquist plot does not
intersect the real axis except at the origin. The closed loop system is stable.
73
Nyquist Diagram
10
5
Imaginary Axis
-5
-10
-1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0
Real Axis
Figure.10.26. Matlab open loop Nyquist plot of prob.7.
Nyquist Diagram
1
0.5
Imaginary Axis
minimum stability margins

0
peak response
-0.5
-1
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Real Axis
Figure.10. 27. Matlab closed loop Nyquist plot of prob.7.
74
Prob. 8.
A. check the open loop and closed loop stability of the following system using Matlab?
B. Find gain margin, phase margin, phase crossover frequency and gain crossover
1
frequency? G ( s) 
s( s  1)
Solution :
The Nyquist plot is drawn by using the MATLAB function Nyquist
num=1; den=[1 1 0];
nyquist(num,den);
axis([-1 0.1 -20 20]);
Nyquist Diagram
20
15
10
Imaginary Axis
-5
-10
-15
-20
-1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1
Real Axis
Fig.10.28 .Nyquist plot of the prob.8.

In order to find the phase margin and the corresponding gain crossover frequency we use
the MATLAB function margin as follows:
[Gm,Pm,wcp,wcg]=margin(num,den)
Gm =Inf,Pm = 51.8273,
wcp = Inf, wcg = 0.7862
75
Lecture Nine:
Control system design by the root

locus method.
9.1. Introduction.
9.2. Effects of the addition of poles and zeroes.
9.3. Lag Compensators.
9.4. Lead Compensators.
9.5. Lag – Lead Compensators.
9.6. Examples
30
9. 1. Introduction.
The performance specifications for a system are usually based on a desired damping ratio
(ζ) , damped natural freq. (Wn) and study state error. The root locus ( R.L.) diagrams for the
system show that the desired performance can’t be achieved by gain adjustment alone. The
R. L. introducing dynamic compensators to meet the desired performance specifications.
9.2. Effects of the addition of poles and zeroes.

The selection of the dominant complex root can be based on a desired damping ratio (ζ)
(which determines the peak overshoot), the desired damped natural frequency wn (which is
the frequency of oscillation of the transient ), or the desired real part (σ) of the dominant root
(which determines the settling time ). Two additional factors must be included in the design
. First, the presence of additional poles will modify the peak overshoot. It may be necessary
to specify the value of the gain Km. For example, in a Type 1 system it must be necessary to
specify a minimum value of K1 in order to limit the magnitude of the steady state error e ss
with a ramp input , where ess (t) = Dr(t)/k1 .
The root–locus plots described in last lecture show the relationship between the gain of the
system and the time response . Depending on the specifications established for the system,
the gain that best achieves the desired performance is selected. The performance
specifications may be based on the desired damping ratio, undamped natural frequency, time
constant, or steady-state error.
The root locus may show that the desired performance cannot be achieved just by
adjustment of the gain. In fact, in some cases the system may be unstable for all values of
gain. The control-systems engineer must be then investigate the methods for reshaping the
root locus to meet the performance specifications. The purpose of reshaping the root locus
generally falls into one of the following categories:
31
1) A given system is stable and its transient response is satisfactory, but its steady-state
error is too large . Thus , the gain must be increased to reduce the steady-state error.
This increase must be accomplished without appreciably reducing the system stability.
2) A given system is stable , but its transient response is unsatisfactory. Thus, the root
locus must reshaped so that it is moved farther to the left, away from the imaginary axis.
3) A given system is stable, but both its transient response and its steady-state response are
unsatisfactory. The locus must be moved to the left and the gain must be increased.
4) A given system is unstable for all values of gain . Thus, the root locus must be reshaped
so that part of each branch falls in the left –half s –plane, thereby making the system
stable.
Compensation of a system by the introduction of poles and zeros is used to improve the
operating performance. However, each additional compensator pole increases the number of
roots of the closed-loop characteristic.
9.3. Lag Compensators

When the transient response of a F.B. control system is considered satisfactory but the
steady state error is to large , it is possible to eliminate this error by increased the system
type . This must be achieved with only a minimum of change in the dominant roots of the clc
equation. The system type can be increases by introducing Lag Compensator
Gc (s) = ;
When the compensator has been added to the system, the loop gain for the new system is
(compensated system):
K= =
32
Where K is the loop gain of uncompensated system. If the compensator is added to a type
(0) system to improve the steady state accuracy, then the value of the step error coefficient ks
before k after the addition of the compensator is :
♣ The following two points can be used as a guided to the design of phase retard
compensator:
1) If the angle contributed by the compensator at the original closed-loop dominant roots
is less them 5 then the new root-locus is only slightly displaced from the original. The ,
the pole (s = -1/αT), and the zero (s= -1/T) of the compensator are placed very close
together.
2)The ratio of the zero & the pole of the compensator, i.e. ,
at s=0 must approximately equal to the desired increase in gain.
♣The procedures for design lag compensator:
1. Plot the root locus for given system.
2. Plot the line with angle which depends on the value of damping ratio where this point can
be limit.
3. Calculate the gain at the dominant root.
4. Calculate the parameter of lag compensator (A,α,T).
5. Checking the required condition for compensator ordered.
33
Example: A unity feedback control system with forward path T.F.

G(s) =
a) Sketch the root locus and find k for limiting stability?

b) Design compensator to increase kv (velocity error coefficient) to at least 4 sec-1
without effecting system transient performance.
Solution:
G(s)= ; H(s)=1; G(s)H(s) =
Poles: S1=0; S2,3 = -1 j;

Center of asymptote = = -0.66 ; angle of asymptote = =60
Angle of departure=Øp=180-∑Øp -( ) - .
C.E.=s3+2s2+2s+k=0
Jw3 – 2w2 2jw k →jw3+2jw =0
W=0; or w2 2 → w2 = -2 ; w j √2
jw
1.5
160
After drawing the root locus;

For ζ . ; Draw a line of angle
Cos- 1 . .
34
Therefore the dominant roots are: S1,2 = - 0.5 + j 0.9 (from the drawing point A) static loop
sensitivity at that point :
K │s││s - j ││s j│ s= -0.5 + j 0.9
K │- . j . ││ . - j . │ │ . j . │ . ≈
Kv = = = velocity – error coff. Of the dominant roots
Now to increase Kv without changing the sy. transient response, a lag compensator is added
in cascade to the sgs.
K v′ Kv → where Kv′ (velocity –error coff. Of compensated sy).
α = =8
Let . (choosen) → = 0.0125
Gc (s) = A α
To know that the addition of this compensator will not effect on the actual root – locus of
the dominant roots s = -0.5 j 0.9, check the angle
Contribution of the compensator at s = -0.5 + j 0.9
Gc(s) = S + 0.1 – S+0.0125 S= -0.5+j0.9
= - 0.4 + j0.9 - - 0.375 + j0.9
= 113.96 – 118.4 = - 4.48 (i.e less than 5 ).
‫ ؞‬The compensator is acceptable .
35
9.4 Lead Compensator

When the transient response of a feedback system must be improved, it becomes necessary
to reshape the root –locus so that it is moved for there to the left of jw-axis. The introduction
of a zero can be obtained by using lead compensator with the following)
Gc (s) = A Where
By making a sufficiently small, the location of the pole at s = - is far to the left of the zero
at s = - and has therefor only a small effect on the important part of the root – locus .
♣ The design procedure is as follows :

1.Plot horizontal line from dominant root point parallel with real axis.
2. Plot the line from origin to dominant root.
3. Calculate the point (B) which represents the bisector of the angle (OPA).
4. Draw the line between dominant root and point (B) .
5. Construct two line from point (B) to point (D) and (C) by take the bisector angle (Ø/2) and
(Ø/2) where (Ø) angle condition for the cascade transfer function (Gc(s) G(s)).
6. The point (D) represents the zero of lead compensator while point (C) represents the pole
of lead compensator.
7. Checking the value of new gain by using the magnitude condition
At the desired location N of one of the dominant closed – loop. Poles calculate the angle
contributions of the open – loop. Poles & zero. The lead network must then contribute an
angle Ø such that the total some of the angles of the original & compensator is +180 (1+2) .
♣To find the value of T & α of the compensator, then:
1-Draw a horizontal line AP, P (location of one of dominant pole) .
2. Draw a line from P to the origin 0 .
3. bisect the included angle APO . Call the bisector Bp .
4. Construct two lines PC and PD which make angles + with the bisector .
36
5. The intersection of PC and PD with the -ve real axis give the required locations of the
pole and the zero of the lead compensator.
jw
∟oPA
ф/2
A
P
C B D o 
Pole= Zero=
Example: For the unity- feedback, C/L system, with:

G (s) = .
Design a compensator to satisfy the following conditions

a. Setting time (2% criterion) ≤ 4 sec .
b. Damping ratio ζ = 0.45
Solution:
G (s) = = =
Ts = (2% error criterion ) , Ts = (5% error)
4= → ζwn = 1 ‫ ؞‬wn = = = 2.2 rad
Wd = wn ζ 2 = 1.9
‫ ؞‬dominant roots required is S1,2 = - ζwn ± jwd

S1,2 = -1± j1.9
Angle deficiency for the desired dominant pole is-:

ф =1 -(⅀θp-⅀θz)|-1+j1.9
ф =1 -(117.75+25.407-0)=36.84
37
‫ ؞‬Gc(s) = =
α= = = 0.11
Find K1 from the magnitude condition → Gc(s)G(s) = 1
Gc(s) G(s) =
‫ ؞‬K1 = at S= -1+ j1.9
K1 = = 33.5
K1 = 5K → K = = 6.7
‫ ؞‬The compensated sy. is Gc (s)G(s) =
! H.W. sketch the root locus for this compensated system..
38
Root Locus
0.93 0.86 0.76 0.62 0.44 0.22
2.5
2
0.97
1.5
1
0.992
Imaginary Axis
0.5
6 5 4 3 2 1
0
-0.5
0.992
-1
-1.5
0.97
-2
0.93 0.86 0.76 0.62 0.44 0.22
-6 -5 -4 -3 -2 -1 0
Real Axis
Root Locus
2.5
1.5
1
Imaginary Axis
0.5
-0.5
-1
-1.5
-2
-5 -4 -3 -2 -1 0
Real Axis
39
Root Locus
0.2
0.15
0.1
0.05
Imaginary Axis
-0.05
-0.1
-0.15
-0.2
-0.25
-0.3
-0.45 -0.4 -0.35 -0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0
Real Axis
Root Locus
0.06
0.04
System: sys System: sys

Gain: Inf Gain: 0
0.02 Pole: -0.05 Pole: -0.005
Damping: 1 Damping: 1
Imaginary Axis
Overshoot (%): 0 Overshoot (%): 0

Frequency (rad/sec): 0.05 Frequency (rad/sec): 0.005
0
System: sys
Gain: 0
Pole: 0
-0.02 Damping: -1
Overshoot (%): 0
Frequency (rad/sec): 0
-0.04
-0.06
-0.1 -0.08 -0.06 -0.04 -0.02 0
Real Axis
40
Root Locus
1
Imaginary Axis
-1
-2
-3
-4
-16 -14 -12 -10 -8 -6 -4 -2 0

Real Axis
Root Locus
1
Imaginary Axis
0
Axis
-1
-2
-3
-5 -4 -3 -2 -1 0 1
Real Axis
41
9.5. LAG–LEAD COMPENSATION

Lead compensation basically speeds up the response and increases the stability of the system.
Lag compensation improves the steady-state accuracy of the system, but reduces the speed of
the response. If improvements in both transient response and steady-state response are
desired, then both a lead compensator and a lag compensator may be used simultaneously.
Rather than introducing both a lead compensator and a lag compensator as separate units,
however, it is economical to use a single lag–lead compensator. Lag–lead compensation
combines the advantages of lag and lead compensations. Since the lag–lead compensator
possesses two poles and two zeros, such compensation increase the order of the system by 2,
unless cancellation of pole(s) and zero(s) occurs in the compensated system.
An electronic lag–lead compensator using operational amplifiers presents by the following
Fig.
Lag-Lead Compensator (Electronic Form)

The transfer function for this compensator may be obtained as follows: The complex
impedance Z1 is given by:
42
If is required in Equation of compensator , then the preceding design procedure for the
lag–lead compensator may be modified as follows:
1) From the given performance specifications, determine the desired location for the
dominant closed-loop poles
43
44
Example: Consider the control system shown in Figure below. This system has closed-loop
poles at (s=-0.25±g1.9843; The damping ratio is 0.125, the undamped natural frequency is 2
rad/sec, and the static velocity error constant is 8 sec–1. It is desired to make the damping
ratio of the dominant closed-loop poles equal to 0.5 and to increase the undamped natural
frequency to 5 rad sec and the static velocity error constant to 80 sec–1. Design an
appropriate compensator to meet all the performance specifications?
Solution:
To design the phase-lead portion of the compensator, we first determine the location of the
zero and pole that will give 55° contribution. There are many possible choices, but we shall
here choose the zero at s=–0.5 so that this zero will cancel the pole at s=–0.5 of the plant.
Once the zero is chosen, the pole can be located such that the angle contribution is 55°. By
45
simple calculation or graphical analysis, the pole must be located at s=–5.02. Thus, the
phase-lead portion of the lag–lead compensator becomes:
46
H.W.
1) Plot the Root Locus for compensated system and uncompensated system and
compare between them?
2) By Using Malab plot the unit step response for compensated system and
uncompensated system and compare between them?
47
Chapter Eleven
Bode Diagram
11.1. Introduction
11.2. Definition of Bode Plot.
11.3. Bode plots rules:
11.4. Calculate Gain margin and phase margin
11.5. Examples
76
11.1. Introduction
put:
Output: magnitude and phase shift | |
Frequency response to is given by

Linear simulation result
2.5
2
1.5
1
Amplitude
0.5
0
-0.5
-1
-1.5
-2
0 2 4 6 8 10 12 14 16 18 20
Time (sec)
We know determines the frequency response .How to plot this

information?
● 1 independent variable:
● 2 dependent variables: and
𝐼𝑚 𝐺 𝑖𝜔
𝑅𝑒 𝐺 𝑖𝜔
An alternative is to plot polar variables

● 1 independent variable:
● 2 dependent variables: and | |
77
|𝐺 𝑖𝜔 |
𝐺 𝑖𝜔
● Advantage: all information corresponding to physical data. Can be

found directly using a frequency sweep.
If we only want a single plot we can use as a parameter.
Nyquist diagram
0.6
0.4
Imaginary axis
0.2
0
-0.2
-0.4
-0.6
-0.6 -0.4 -0.2 0 0.2 0.4 0.6

Real axis
A plot of vs. as a function of .

● Advantage: All information in a sign plot.
● AKA: NYquist plot.
11.2. Definition of Bode Plot.

The Bode plot is a pair of log-log and semi log plots:
1- Magnitude plot : | | vs.
2- Phase plot: vs.
| | Is a unit of Decibels (dB?)
● Used in power and circuits.
● | | in other fields.
78
Note that by log, we mean log base

● In mat lab, log means natural logarithm.
Valve frequency response
2 0
Amplitude
Amplitude
0
-1.3dB
Output amplitude ‒‒ dB
-2 90 -50
Phase lead ‒‒ ᵒ
Lag
-1.3dB Phase
Phase 90
-4 Lag
-6 -100
50HZ50HZ
-8
-10 -150
1 10 100
Solution frequency ‒‒ Hz
Let’s do a simple pole

We need
● Magnitude of
● Phase of
𝐼𝑚 𝑠
We can write:
| | | | 𝜔
| | 𝜔
| | | |
𝑅𝑒 𝑠
| |
| | √
How to plot | | ⁄√
| |
√
Three cases:
Case 1:
● Approximate
79
| |
Bode diagram
10
5
0
-5
Magnitude (dB)
-10
-15
-20
-25
-30
-35
Frequency (rad/sec)
Case 2:
| |
Case 3:
● Approximate
| |
Bode diagram
10
5
0
-5
Magnitude (dB)
-10
-15
-20
-25
-30
-35
Frequency (rad/sec)
But we use a log-log plot.

● -axis is ; -axis is | |
80
Conclusion: on the log-log plot, when

● Plot is linear i.e. slope is
Of course, we need to connect the dots.
Bode diagram
10
5
0
-5
Magnitude (dB)
-10
-15
-20
-25
-30
-35
Frequency (rad/sec)
Bode diagram
0
-10
Magnitude (dB)
-20
-30
-40
0
Phase (deg)
-45
-90
Frequency (rad/sec)
Now let’s do the phase. Recall:
∑ ∑
81
8 9 9 8
Graph of the tangent function

Again, 3 cases:
Case 1:
●
●
𝐼𝑚 𝑠
𝜔
𝑖𝜔
𝑅𝑒 𝑠
82
225
180
135
90
45
Phase (deg)
0
-45
-90
-135
-180
-225
Frequency (rad/sec)
Case 3: 𝜔
Case 2: 𝜔
● 𝐺 𝑖𝜔
● 𝐺 𝑖𝜔 ● 𝐺 𝑖𝜔 9
● 𝐺 𝑖𝜔 45 Fixed at 9 for large 𝜔
0
𝜔
𝜔
Phase (deg)
-45
𝜔
-90
Frequency (rad/sec)
We need to connect the dots somehow.
83
0
𝜔
Phase (deg)
𝜔
-45
𝜔
-45
Frequency (rad/sec)
So for, drawing Bode plots seems pretty intimidating.

● Solving ;
● dB and log-plots
● Lots of trig
The process can be greatly simplified:
● use a few simple rules.
Example: Suppose we have
Then
| | | || |
And
| | | | | |
11.3. Bode plots rules:

Rule # 1: magnitude plots add in log-space. For
| | | | | |
84
40 Zero
20 Bode zero
Magnitude (dB)
0 Sum
Bode sum
-20
Pole
-40
Bode pole
-60
1 10 100 1000 10000 100000
Frequency
Rule # 2:
Phase plots add. For
90 Zero
70
Bode zero
50
30 Sum
Phase (degrease)
10
Bode sum
-10
-30 Pole
-50
-70 Bode pole
-90
1 10 100 1000 10000 100000
Frequency
Rule # 3: The constant

Before rushing in, let’s make sure we don’t forget the constant term. If
85
Magnitude: | | | |
● | | | |
Bode diagram
10
5
0
-5 |𝑐|
Magnitude (dB)
-10
-15
-20
-25
-30
-35
Frequency 𝜔
Conclusion: Magnitude is constant for all
Phase: {
8
225
180
𝑐 0
135
90
45
Phase (deg)
0
-45 𝑐 0
-90
-135
-180
-225
Frequency (rad/sec)
Conclusion: Phase is if otherwise 8
86
Rule # 4: “pure” zero

Let’s start with a zero at the origin:
Magnitude:
● |𝐺 𝑖𝜔 | |𝑖𝜔| |𝜔|
● |𝐺 𝑖𝜔 | |𝜔|
Our x-axis is 𝜔
Phase: 𝐺 𝑠 𝑠
● Plot is linear for all 𝜔
● 𝐺 𝑖𝜔 𝑖𝜔 9
● Slope is +20 dB/decade
● Always 9
● Need a point: 𝜔
|𝐺 𝑖𝜔 |
𝜔
● Passes through 0dB at 𝜔
Bode diagram
20
15
Magnitude (dB)
10
5
0
-5
91
90.5
Phase (deg)
90
89.5
89
Frequency (rad/sec)
Rule # 5.A “pure” zero: multiple zeroes

What happens if there are multiple pure zeroes?
87
● Just what you would expect .
Bode diagram
10
5
0
-5
Magnitude (dB)
-10
𝑘
-15
𝑘 4
-20
-25
𝑘
-30
-35
𝑘 𝜔
Magnitude:
●| | | | | |
| | | | | |
● Slope is +20k dB/Decade
Need a point
● At
| | ; k pure zeroes added together.
● Still passes through 0dB at

A “pure” Zero: multiple zeroes k pure zeroes added together.
And phase for multiple pure zeroes?
Phase:
● 9 Always 9
88
405
𝑘 4
360
315
𝑘
270
225
Phase (deg)
𝑘
180
135
90 𝑘
45
0
-45
Frequency (rad/sec)
Rule # 6. Normal zeros

Compare this to the magnitude plot of
4 𝑎
6 𝑎
𝑎
4 𝑎
𝑀
8 𝑎
20
𝑎
6 𝑎 Slope = 6dB/octave
= 20dB/Decade
𝑎
𝑎 𝑎 𝑎 𝑎 𝑎
𝜔 𝑇 𝜔 𝑇
Frequency (rad/sec)
This is why we use the format

● We want 0dB (no gain) at low frequency.
89
11.4. Calculate Gain margin and phase margin
The phase and gain margins of a control system are a measure of the
closeness the polar plot to the (-1+j0) point. Therefore, these margins may
be used as design criteria. It should be noted that either the gain margin
alone or the phase margin alone doesn’t give a sufficient indication of the
relative stability. Both should be given in the determination of relative
stability. For a minimum-phase system, both the phase and gain margins
must be positive for the system to be stable. Negative margins indicate
instability. Proper phase and gain margins ensure us against variations in
the system components and are specified for definite positive values.
The two values bound behavior of the closed-loop system near the
resonant frequency. For satisfactory performance, the phase margin should
be between 30” and 6 , and the gain margin should be greater than 6dB.
Whit this value, a minimum-phase system has guaranteed stability, even if
the open-loop gain and time constant of the components vary to a certain
extent. Although the phase and gain margins give only rough estimates of
the effective damping ratio of the closed-loop system, the do offer a
convenient means for designing control system or adjusting the gain
constant of systems. For minimum-phase systems, the magnitude and phase
characteristics of the open-loop transfer function are definitely related.
The stability measured by phase and gin margins of stable and unstable
systems, by the Fig.11.1. (a) Bode diagrams; (b) polar plots; (c) log-
magnitude versus-phase plots.
90
Negative gain
Positive gain margin
margin
0 0
𝜔
|𝐺|in dB
𝜔
|𝐺|in dB
9 9
𝐺
8 8
𝐺
𝜔 𝜔
7 Positive phase 7
Negative phase
margin
margin
System stable (a) Unstable system
Positive gain Im Im
𝑘
margin𝑔 G plane
Negative
G plane
|𝐺 𝑗𝜔 | phase margin
𝑦 8 ∅
𝑘𝑔 𝑦
𝑦 Re Re
Ø
∅
Positive
phase margin 𝑘𝑔 Negative
gain margin
𝐺 𝑗𝜔 𝐺 𝑗𝜔
System stable (b) Unstable system
Positive
Phase
Margin
Negative Gain
Margin
|𝐺|In dB
|𝐺|In dB
0 0
Positive
Gain Negative
Margin Phase
Margin
7 8 9 7 8 9
𝐺 𝐺
(c)
System stable Unstable system
Fig. (a) Bode diagrams; (b) polar plots; (c) log-magnitude versus-phase plots.
91
The requirement that the phase margin be between and 6 means that
in a Bode diagram the slope of the log-magnitude curve at the gain
crossover frequency should be more gradual than -40dB/decade. In the
most practical cases, a slope of -20 dB/decade is desirable at the gain
crossover frequency for stability. If it is -40 dB/decade, the systems could
be either stable or unstable. (Even if the system is stable, however, the
phase margin is small). If the lope at the gain crossover frequency is -60
dB/decade or steeper, the system is most likely unstable. For non-minimum
phase system, the correct interpretation of stability margins requires careful
study.
The best way to determine the stability of non-minimum phase systems is
to use the NYquist diagram approach rather than Bode diagram approach.
The rules for phase and gain margin are:
Gain margin: | | = frequency at phase = 8

Phase margin: 8 | = frequency at gain 0dB
92
11.5. Examples
Ex. 1. obtain the phase and gain margins of the system shown in figure for
the two cases where k=10 and k=100.
R(s) + 𝑘 C(s)
_ 𝑠 𝑠 𝑠 5
Solution:
The phase and gain margins can easily be obtained from the Bode diagram.
A Bode diagram of the given open-loop transfer function with k=10 is
shown in figure (a). The phase and gain margins for k=10 are:
Phase margin = 21” , Gain margin = 8 dB
Therefore, the system gain may be increased by 8dB before the instability
occurs. Increasing the gain from k=10 to k=100 shifts the 0-dB axis down
by 20dB, as shown in figure (b). The phase and gain margins are:
Phase margin = , Gain margin = -12 dB
Thus the system is stable for k =10, but unstable for k=100.
30
K=10
20
10
|𝐺𝑚| 0 +8 dB (Gain margin)
(DB)-10
-20
-30
Phase9
(deg)
8 (Phase margin)
7
0.2 0.4 0.6 0.8 1 2 4 6 8 10
𝜔
Figure (a)
93
K=10
50
40
30 (Gain margin) -12 dB
20
10
|𝐺𝑚| 0
(DB)-10
(Phase margin)
Phase 9
(deg)
8 0.4 0.8 1 2 4 6 8
𝜔
7
0.2 0.6 10
Figure (b)
Notice that one of the very convenient aspects of the Bode diagram
approach is the easy with which the effects of gain changes can be
evaluated. Note that to obtain satisfactory performance, we must increase
the phase margin to 6 . This can be done by decrease the gain k.
Decreasing k is not desirable, however, since a small value of & will yield
a large error for the ramp-input. This suggests that reshaping of the open-
loop frequency-response curve by adding compensation may be necessary.
The mat lab solution for this example:

[ ]
[ 65 ]
Bode
[ ]
94
Bode diagram
80
60
Magnitude (dB)
40
20 GM
0
-20
-40
-60
-80
-100
-90
-135
Phase (deg)
-180 PM
-225
-270
Bode diagram
K=100:
Bode Diagram
100
Magnitude (dB)
50
GM
0
-50
-100
-90
Phase (deg)
-135
PM
-180
-225
-270
Frequency (rad/sec)
Ex. 2. Draw the Bode plots for the system with G(s) as follows:
[ ]
Solution:
We will make a Bode plot for the open-loop function
[ ]. The Bode plot is the sum of the Bode plots for each
95
first-order term. Thus, it is convenient to us the normalized plot for each of

these terms so that the low-frequency asymptote of each term, except the
pole at the origin, is at 0dB, making it easier to add the components of the
Bode plots. We rewrite G(s) showing each term normalized to a low-
frequency gain of unity. Hence
The tables described how plot the magnitude as:

Frequency (rad/sec)
0.1(start: 1(start: 2(start: 3(start:
Description Pole at 0) Pole at -1) Pole at -2) Zero at -3
Pole at 0 -20 -20 -20 -20
Pole at -1 0 -20 -20 -20
Pole at -2 0 0 -20 -20
Zero at -3 0 0 0 20
Total slope -20 -40 -60 -40
(dB/dec)
40
-20dB/Dec (-6dB/oct)
20 -40 dB/Dec (-12dB/oct)
0 -60dB/Dec (-18dB/oct)
𝑀
-40 dB/Dec (-12 dB/oct)

-20
-40
-60
-80
0.1 1 10 100
96
The table for phase as:

Frequency (rad/sec)
0.1(star 0.2(star 0.3(star 0(End: 20(End: 30(End:
Description t: t: t: Pole at Pole at Zero at
Pole at Pole at Pole at -1) -2) -3
-1) -2) -3)
Pole at -1 -45 -45 -45 0
Pole at -2 -45 -45 -45 0
Zero at -3 45 45 45 0
Total slope -45 -90 -45 0 45 0
(dB/dec)
-90
45 /Dec
-120
45 /Dec
-180
Phase (degrees)
9 /Dec
-225
45 /Dec
-270
-315
-360
0.1 1 10 100
Frequency (rad/sec)
H.W. Calculate the gain and phase margin and discuss the stability?
97
Bode Diagram G(s) = 1/s+1

0
-10
Magnitude (dB) -20
-30
-40
0
Phase (deg)
-45
-90
Frequency (rad/sec)
Bode Diagram
G(s) = 10/s+1
20
Magnitude (dB)
10
-10
-20
0
Phase (deg)
-45
-90
Frequency (rad/sec)
98
Bode Diagram G(s) = 1/(s+0.1)

20
Magnitude (dB)
10
-10
-20
0
Phase (deg)
-45
-90
Frequency (rad/sec)
Bode Diagram G(s) = 1000/s+1

60
50
Magnitude (dB)
40
30
20
0
Phase (deg)
-45
-90
Frequency (rad/sec)
99
G(s) = s/(s+1)
Bode Diagram
0
-10
Magnitude (dB)
-20
-30
-40
-50
90
Phase (deg)
45
Frequency (rad/sec)
Bode Diagram G(s) = 10/ (s2+2s+1)

50
Magnitude (dB)
-50
-100
-150
-90
Phase (deg)
-135
-180
-225
-270
Frequency (rad/sec)
100
Bode Diagram G(s) = 1/ s(s2+2s+1)

50
Magnitude (dB)
0
-50
-100
-150
-90
Phase (deg)
-135
-180
-225
-270
Frequency (rad/sec)
Bode Diagram
20
Magnitude (dB)
10
0
G(s) = 10(s+1)/(s2+2s+1)
-10
-20
0
Phase (deg)
-45
-90
Frequency (rad/sec)
101
Bode Diagram
G(s) = 512(s+3)/ s(s2+16s+256)
40
20
Magnitude (dB)
0
-20
-40
-60
-80
-45
Phase (deg)
-90
-135
-180
Frequency (rad/sec)
Bode Diagram
G(s) = s/(s3+3s2+2s+1)
0
-20
Magnitude (dB)
-40
-60
-80
90
Phase (deg)
-90
-180
Frequency (rad/sec)
102
Ex.3. Plot the Bode diagram for the G(s) = k(s+2)/s2(s+8)(s+10), then
check the stability and find the limit of gain?
Solution: at k = 1;
Bode Diagram
50
0
Magnitude (dB)
-50
-100
-150
-200
-135
Phase (deg)
-180
-225
-270
Frequency (rad/sec)
Bode Diagram
50 G(s) = k(s+2)/s2(s+8) (s+10), k=20
Magnitude (dB)
-50
-100
-150
-135
Phase (deg)
-180
-225
-270
Frequency (rad/sec)
103
Bode Diagram
G(s) = k(s+2)/s2(s+8) (s+10), k=40
50
System sys
0 Frequency (rad/s): 1.01 System sys
Phase (deg) Magnitude (dB) Magnitude (dB): 0.663 Frequency (rad/s): 2.58
-50 Magnitude (dB): -12.9
-100
-150
-135
System sys
-180 Frequency (rad/s): 6.87
Magnitude (dB): -182
-225
-270
Frequency (rad/sec)
5000
3000
1000
700
300
100
50
20
10
w
8
0
-268.2
-266.9
-260.9
-257.0
-240.4
-197.7
-170.4
-160.3
-166.3
-168.5
-137.8
-178.4
-180.0
Phase
Bode Diagram
G(s) =640(s+2)/s2(s+8)(s+10)
-100
-0
Magnitude (dB)
-50 Corner freq
𝜔1=1/0.5=2
50
Gm=-2;
Pm=-8 𝜔2=1/0.125=8
-135 𝜔3=1/0.1=10
Phase (deg)
-180
-225
The system unstable

-270
Frequency (rad/sec)
104
Ex.4. Plot the Bode diagram for the following system, then check the
stability and find the limit of gain?
Solution:
R(s) + 𝑠 C(s)
8 𝑒
_ 𝑠 𝑠 4 𝑠
G(s)
; he phase angle of the transport lag is
57
40
20
-20
0
-40
-20 9
|𝐺|
-40 8
-60
𝐺
-60 7
-80 6
-100 45
0.1 0.2 0.4 0.6 1 2 4 6 10 20 40 60 100
𝜔 In rad/sec
=4 (rad/sec) ; =10 (rad/sec); Gm=8 dB; Pm=45 ; Therefore; the

system is
105
University of Anbar Control Theory
Chapter Twelve
Design PID Controller
12.1. Introduction.
12.2. Choice of the controller type.
12.2.1. P Controller
12.2.2. PI Controller
12.2.3. PD Controller
12.2.4. PID Controller
12.3. PID controller parameter and topology identification
12.4. Parameter tuning according to Ziegler-Nichols

Recommendation.
12.5. PID controller using operational amplifiers.
12.6. Modifications of PID Control Schemes
106
12.1. Introduction.
Today, a number of different controllers are used in industry and in many
other fields. In quite general way those controllers can be divided into two
main groups:
● Conventional controllers
● Unconventional controllers
As conventional controllers we can count controllers known for years
now, such
As P, PI, PD, PID, Otto-Smith, all their different types and realizations,
and other controller types 2. It is a characteristic of all conventional
controllers that one has to know a mathematical model of the process in
order to design a controller. Unconventional controllers utilize a new
approaches to the controller design in which knowledge of a
mathematical model of a process generally is not required. Examples of
unconventional controller are a fuzzy controller and neuro or neuro-
fuzzy controllers.
Manny industrial processes are nonlinear and thus complicate to describe
mathematically. However, it is known that a good many nonlinear
processes can satisfactory controlled using PID controllers providing that
controller parameters are tuned well. Practical experience shows that
this type of control has a lot of
sense since it is simple and based on 3 basic behavior types: proportional
(P), integrative (I) and derivative (D). Instead of using a small number of
complex controllers, a larger number of simple PID controllers is used to
control simpler processes in an industrial assembly in order to automates
the certain more complex process. PID controller and its different types
such as P, PI and PD controllers are today a basic building blocks in
control of various processes.
107
In spite their simplicity; they can be used to solve even a very complex
control problems, especially when combined with different functional
blocks, filters (compensators or correction blocks), selectors etc. A
continuous development of new control algorithms insure that the time of
PID controller has not past and that this
Basic algorithm will h a v e its part to play in process control in
foreseeable future. It can be expected that it will be a backbone of
many complex control systems. PID controllers use a 3 basic behavior
types or modes: P - proportional, I - integrative and D - derivative. While
proportional and integrative modes are also used as single control
modes, a derivative mode is rarely used on its own in control systems.
Combinations such as PI and PD control are very often in practical systems.
It can be also s h o w n that PID controller is a natural generalization of
a simplest possible controller - On-off controller.
12.2. Choice of the controller type

Insofar were described proportional, integrative and derivative modes of
the controllers and a rational behind their use was explained. However,
excerpt for a few tips, a n a t t e n t i o n was not given to a question
w h e n to use different types of Controllers. The rest of this section will
give some answers on that particular topic.
12.2.1. P Controller
When P controller is used, large gain is needed to improve steady state
error. Stable system does not have a problem when large gain is used. Such
systems are systems with one energy storage (1st order capacitive systems).
If constant steady state error can be accepted with such processes, than P
controller can be used. Small steady state errors can be accepted if sensor
108
will give measured value with error or if importance of measured value is

not too great anyway. Example of such system is liquid level control in
tanks when exact approximate levels of liquid suffice for the proper plant
operation. Also, in cascade control sometime it is not important if there is
an error inside inner loop, so P controller can a good solution in such cases.
Derivative mode is not required if the process itself is fast or if the control
system as whole does not have to be fast in response. Processes of 1st
order react immediately on the reference signal change, so it is not
necessary to predict error (introduce D mode) or compensate for the steady
state error (introduce I mode) if it is possible to achieve satisfactory steady
state error using only P controller.
12.2.2. PD Controller
It is well known that thermal processes with good thermal insulation act
almost as Integrators. Since insulation is good and thermal losses are
small, the most significant part of the energy that is led to the system is
used temperature rise. Those processes allow for large gains so that
integral mode in the controller is not needed. These processes can be
described as different connections of thermal energy storages. Thermal
energy is shifted from one storage into another. In general, with such
processes there is present a process dynamics with large inertia. Since
dynamics is slow, derivative mode is required for control of such
processes. Integral mode would only already slow dynamics make more
slowly. The other reason for using PD controllers in such systems is that is
possible to measure temperature with low level of noise in the measured
signal. PD controller is often used in control of moving objects such are
flying and underwater vehicles, ships, rockets etc. One of the reason is in
stabilizing effect of PD controller on sudden changes in heading variable
109
y(t). Often a "rate gyro" for velocity measurement is used as sensor of

heading change of moving object.
12.2.3. PI Controller
PI controllers are the most often type used today in industry. A control
without D mode is used when:
a) fast response of the system is not required
b) large disturbances and noise are present during operation of the process
c) there is only one energy storage in process (capacitive or inductive)
d) there are large transport delays in the system
If there are large transport delays present in the controlled process, error
prediction is required. However, D mode cannot be used for prediction
because every information is delayed till the moment when a change in
controlled variable is recorded. In such cases it is better to predict the
output signal using mathematical model of the process in broader sense
(process + actuator).
12.2.4. PID Controller
Derivative mode improves stability of the system and enables increase in
gain K and decrease in integral time constant Ti, which increases speed of
the controller response. PID controller is used when dealing with
higher order capacitive processes (processes with more than one energy
storage) when their dynamic is not similar to the dynamics of an integrator
(like in many thermal processes). PID controller is often used in industry,
but also in the control of mobile objects (course and trajectory following
included) when stability and precise reference following are required.
Conventional autopilot is for the most part PID type controllers.
110
12.3. PID Controller parameter and topology identification

Problem of identification of PID controller arises when parameters of an
existing controller have to be tuned. Manufacturers usually don’t give
data about controller structure (serial or parallel), so its structure also has
to be determined. Controller parameters have to be manually tuned if they
are changed with time (that is often with hydraulic and pneumatic
controllers) or because process parameters have changed so the controller
does not perform satisfactory anymore. Not knowing the exact structure of
the controller is not critical if manual
Parameter tuning can be done in controlled environment using trial and
error method and if rules given in Table 12.1 are observed.
However, if parameters are not tuned manually or if it is not possible to
use trial and error method, then it is critical to know the controller
structure. Identification can be performed experimentally using a certain
type of reference signal on the controller input (signal y (t)) and
measuring response on the controller output (uPID(t)) - Fig. 1 2 . 1. From
the controller response it is possible to draw conclusions About
parameters and structure of the controller under test.
Table.12.1 Rules for tuning PID controller parameters

Speed of
Parameter Stability Accuracy
response
Increasing Increases Deteriorate Improves
Increasing Decreases Deteriorate Improves
Increasing Increases Improves No impact
111
Signal
generator
P
+ +
e Actuator
r I + UPID + y
+
+ + Process
Fig.12.1: Controller parameters identification – experiment setup
12.4. Parameter tuning according to Ziegler-Nichols recommendation
Ziegler-Nichols tuning is used for P, PI and PID controllers. It has to be

noted that controllers tuned using this procedure are tuned for control, not
tracking. Thus, controllers with parameters tuned according to Ziegler-
Nichols recommendation will perform well i n disturbance rejection
perform poor in tracking reference changes. First method.in the first
method .We obtained experimentally the response of the plant to a unit-
step input, as shown in Fig.12.2 the plant involves neither integrator(s).
Plant
u(t) c (t)
Fig.12. 2 Unit-step response of a plant.
Nor dominant complex-conjugate poles, then such a unit-step response

curve may look like an S-shaped curve, as shown in Fig.12.3 (if the
112
response does not exhibit an S-shaped curve , this method does not apply).
Such step response curve may be generated experimentally or form a
dynamic simulation of the plant.
The S-shaped curve may be characteristic by two constant, delay time L
and time constant T. The delay time and time constant are determined by
drawing and tangent line at the inflection point of the S-shaped and
determined the intersections of the tangent line with the axis and line
, as shown in Figure -3-.The transfer function ⁄ may
then be approximated by a first-order system with a transport lag as
follows:
Ziegler and Nichols suggested to set the values of and according to
the formula shown in table 12.2.
Notice the PID controller tuned by the first method of Ziegler- Nichols
rules gives
c(t)
k
0 t
L T
Fig 12.3 S shape for response
( )
( )
113
Table.12.2 Ziegler- Nichols tuning rule based on step response on plant

(first methods)
Type of
controller
P 0
PI 0
PID 2L 0.5L
Thus, the PID controller has a pole at the origin and double zeroes at
Second methods: In the second methods, we first set and ,

using the proportional control action only (see Fig. 12.5.a) increase
from 0 to a critical value where the output first exhibit sustained
oscillations.(If the output does not exhibit sustained oscillations for
whatever value may take, then this method does not apply).Thus, the
critical gain and the corresponding period are experimentally
determined (see Fig.12. 5.b). Ziegler and Nichols suggest that we set the
values of the parameter and according to the formula shown in
Table 12-2.
114
c(t)
r(t) + u(t) c(t)
𝑲𝑷 Plant
𝑷𝒄𝒓
(a) 0
(b)
Fig. 12.5. (a) Block diagram of control system . (b) parameter of second
method of Ziegler- Nichols
Table.12.2 Ziegler- Nichols tuning rule based on critical gain and
critical period (second method)
Type of
controller
P 0
PI 0
PID
Notice the PID controller tuned by the second methods of Ziegler- Nichols
rules gives: ( )
( )
( )
Thus the PID controller has a pole at the origin and double zeroes at
115
Example 1: By using second method of Ziegler- Nichols, calculate the

PID parameters and make tuning for this parameters to get accepting
response for the following system?
C(s)
R(s) + 𝟏
𝑮𝒄 𝒔 𝒔 𝒔 𝟏 𝒔 𝟓
PID controller
Solution:
Since the plant has an integrator, we use the second method of Ziegler-
Nichols tuning rules. By setting and , we obtain the closed-
loop transfer function as follows:
The value of that makes the system marginally stable so that sustained
oscillations occurs can be obtained by use of Routh’s stability criterion.
Since the characteristic equation for the closed-loop system
The Routh array becomes as follows:
Examining the coefficient of the first column of the Routh table, we find
the sustained oscillation will occur if the Thus, the critical gain
is
116
With gain set equal to the , the characteristic equation

become
From which we find the frequency of the sustained oscillation to be

√ .Hence, the period of sustained oscillation
√
Referring to table 10-2, we determine and as follows:
The transfer function of the PID controller thus as
( )
( )
R(s) + C(s)
𝟔 𝟑𝟑𝟐𝟑 𝒔 𝟏 𝟒𝟐𝟑𝟓 𝟐 𝟏
𝒔 𝒔 𝒔 𝟏 𝒔 𝟓
PID controller
The PID controller has a pole at the origin and double zeroes at
A block diagram of the control system with designed PID
controller is shown in Figure above.
117
Next, let examine the unit-step response of the system. The closed-loop
transfer function is given by
From the above response the max. Overshoot about this is

unaccepted in design of therefore, must be tuning the parameter s of PID to
get:
( )
Where the parameter values:
And the unit-step response can be shown below with final tuning
Can be getting the max. Overshoot about with the following response:
118
Example 2. Figure 12.6 shows an electrical circuit involving an

operational amplifier. Obtain the output let us define:
Noting that the current following into the amplifier is negligible, we have
Since we have
Taking the Laplace transformer of this last equation, assuming the zero
initial condition, we have
Which can written as
The op-amp circuit shown in Figure.1. Is a first-order lag circuit, (Several

other circuits involving op amps are shown in Table12.1.Together with
119
their transfer functions).

𝑖 C
𝑅 𝑅
𝑖
𝑒 +
𝑒𝑖
𝑒
Fig.12.6. First-order circuit using operational amplifier
I(s)
𝑍 𝑠
I(s)
𝑍 𝑠
𝐸 𝑠
+
𝐸𝑖 𝑠 𝐸 𝑠
Fig. 12.7.opreational amplifier
Impedance Approach to obtain transfer function. Consider the op amp

circuit shown in Fig. 12.7 .Similar to the case of electrical circuits we
discussed earlier, the impedance Approach can be applied to op-amp
circuits to obtain their transfer functions. For the circuit shown in Figure.2,
we have
Since we have :
120
Example 3. Referring to the op-amp circuit shown in Figure 12.8 obtains

the transfer function by use of the impedance Approach. The
complex impedances and for this circuit are.
The transfer function is, therefore, obtained as
This is, of course, the same as the obtained in Example1

𝐶
𝑍 𝐶 𝑍
𝑅
𝑖 𝐸 𝑠 𝑖
𝑅 +
𝐸𝑖 𝑠 𝐸 𝑠
(a)
𝐶
𝐶 𝑅
𝑅 𝑅
𝑅 + +
𝐸𝑖 𝑠 𝐸 𝑠 𝐸 𝑠
Load lag network Signer inverter

(b)
Fig.12.8.(a) operational amplifier circuit; (b) operational-amplifier circuit

used as lead or Lag compensator
121
Lead or lag networks using operational Amplifiers. Fig.12.6-(a) shows an

electronic circuit using an operational Amplifier. The transfer function for
this circuit can be obtained as follows. Define the input impedance and
feedback impedance as and , respectively. Then
Hence, referring to equation , we have
Notice that the transfer function in equation 12.3.contains a minus sign.

Thus, this circuit is sign inverting. If such a sign inversion is not
convenient in the actual application, a sign inverter may be connected to
either the input or the output of the circuit of Fig. 12.8-(a).An example is
shown in Figure 12.6-(b).The sign inverter has the transfer function of
The sign inverter has the gain of .Hence the network shown in Fig.
12.8-(b) has the following transfer function:
122
𝑅 𝐶
𝑍
𝑍 𝐶
𝑅
𝑅 + +
𝐸𝑖 𝑠 𝐸 𝑠 𝐸 𝑠
Fig.12.9. Electronic PID

Where :
Notice that
This network has dc gain of .

Notice this network is a lead network if or .It is a lag
network if
123
12.5. PID controller using operational amplifiers.

Figure.4 shown an electronic proportional-plus-integral-plus-derivative
controller (a PID controller) using operational amplifiers. The transfer
function is given by
Where
( )( )
Noting that
We have
( )
[ ]
Notice that the second operational-amplifier acts as a sign inverter as well

as again adjuster
When a PID controller is expressed as
( )
is called the proportional gain , is called the integral time , is

called the derivative time. From equation 12.8 can be obtain the
proportional gain , integral time , and derivative time to be
124
When a PID controller is expressed as
is called the proportional gain , is called the integral gain , is

called the derivative gain. For this control
Table is shows a list of operational-amplifier circuit that may be used as

controller or compensators.
125
Table 12.3. Operational amplifier circuit that may be used as a

compensators
Control
Operational amplifier circuits
Action
𝑅 𝑅
𝑅
_ 𝑅
_
+
𝑒𝑖 +
𝑒
1 P
𝐶 𝑅
𝑅
_ 𝑅
_
+
𝑒𝑖 +
𝑒
2 I
𝐶 𝑅 𝑅
_
𝑅 _
+ 𝑅
𝑒𝑖 +
𝑒
3 PD
126
𝑅 𝐶 𝑅
𝑅
_
_
4 PI + 𝑅
𝑒𝑖 +
𝑒
𝐶 𝑅 𝐶 𝑅
_
5 PID 𝑅 _
+ 𝑅
𝑒𝑖 +
𝑒
𝐶 𝐶
𝑅
_ 𝑅
Lead or 𝑅 _
6 + 𝑅
lag 𝑒𝑖 +
𝑒
𝑅 𝐶 𝑅 𝐶
𝑅
Leg- _ 𝑅
𝑅 _
7 𝑅
lead [ ] 𝑒𝑖
+
+
𝑒
[ ]
127
12.6. Modifications of PID Control Schemes

Consider the basic PID control system shown in Figure 12-10(a), where the
system is subjected to disturbances and noises. Figure 12-10(b) is a
modified block diagram of the same system. In the basic PID control
system such as the one shown in Figure 12-10(b), if the reference input is a
step function, then, because of the presence of the derivative term in the
control action, the manipulated variable u(t) will involve an impulse
function (delta function).
In an actual PID controller, instead of the pure derivative term Tds, we
employ:
where the value of ɤ is somewhere around 0.1.Therefore, when the

reference input is a step function, the manipulated variable u(t) will not
involve an impulse function, but will involve a sharp pulse function. Such a
phenomenon is called set-point kick.
Figure 12-10: (a) PID-controlled system; (b) equivalent block diagram.

128
 Set-Point Kick.
If the reference input is a step function, the manipulated variable will
involve an impulse function and such a phenomenon is called set-point
kick. To avoid set-point kick phenomenon derivative action is placed in
feedback path so that differentiation occurs only on the feedback signal and
not on the reference signal.
 PI-D Control.
To avoid the set-point kick phenomenon, we may wish to operate the
derivative action only in the feedback path so that differentiation occurs
only on the feedback signal and not on the reference signal. The control
scheme arranged in this way is called the PI-D control. Figure 12-11 shows
a PI-D-controlled system. From Figure 12-11, it can be seen that the
manipulated signal U(s) is given by
( ) ( )
Notice that in the absence of the disturbances and noises, the closed-loop
129
transfer function of the basic PID control system [shown in Figure 12-
10(b)] and the PI-D control system (shown in Figure 12-11) are given,
respectively, by
Figure 12-11 PI-D-controlled system.
( )
( )
And
( )
( )
It is important to point out that in the absence of the reference input and
noises, the closed-loop transfer function between the disturbance D(s) and
the output Y(s) in either case is the same and is given by
( )
 I-PD Control.
Consider the case where the reference input is a step function. Both PID
control and PI-D control involve a step function in the manipulated signal.
Such a step change in the manipulated signal may not be desirable in many
occasions. Therefore, it may be advantageous to move the proportional
130
action and derivative action to the feedback path so that these actions affect
the feedback signal only. Figure 12-12 shows such a control scheme. It is
called the I-PD control. The manipulated signal is given by
Figure 12-12 I-PD-controlled system.
( ) ( )
Notice that the reference input R(s) appears only in the integral control
part. Thus, in I-PD control, it is imperative to have the integral control
action for proper operation of the control system.
The closed-loop transfer function Y(s)/R(s) in the absence of the
disturbance input and noise input is given by
( )
( )
It is noted that in the absence of the reference input and noise signals, the
closed-loop transfer function between the disturbance input and the output
is given by
( )
This expression is the same as that for PID control or PI-D control.
131
Chapter Thirteen
Analysis of Sampled-Data
Control Systems
This chapter discusses the following topics :
13.1. Introduction.
13.2. Sampled data system.
13.3. Stability of Sampled-Data Control Systems.
13.4. Modified Routh Stability Criterion.
133
13. 1. Introduction
The analysis of sampled-data control systems is complicated by the

presence of both continuous and discontinuous (discrete) components in the
control loop. Just as with continuous systems, a block diagram of the
feedback control system can serve as the basis for analyzing the dynamic
behavior of a digital system. The main focus of this chapter is the
application of block diagram algebra to obtain the closed-loop transfer
function and the analysis of the resulting characteristic equation to evaluate
control system stability. However, care must be exercised in the
development of the closed-loop transfer function. In particular, the rules for
block multiplication used for continuous systems must be modified when
sampling operations are included. Several techniques based on analysis of
the characteristic equation of the closed-loop transfer function are resented
which can be used efficiently to determine loop stability. Again, as with
continuous systems, the stability of a sampled-data control system is a
necessary condition for operability.
13.2. Sampled data system.
This is an example of ideal impulse sampling of a continuous-time

signal to produce a discrete-time signal. Several different sampling periods
Ts will be used to study the effect of the sampling period on the ability of
the continuous-time signal to be reconstructed from the discrete-time
samples. With each sample period Ts, the values of the samples exactly
equal the values of the continuous-time signal at the corresponding sample
points in time. The continuous-time signal is the sum of two pure cosine
terms as shown by the equation(13.1) and illustrated in the Figure 1.
( ) ( ) ( ) (13.1)
134
The three sample periods which will be used are

and .The corresponding sampling frequencies are
. The
effect of sampling ( ) at each of these rates will be studied in turn.
Figure. 13.1. The continuous-time signal

● First, consider ( ) We will
call the sampled signal ( ) The frequencies that appear in the sampled
signal will be the frequencies in the original signal (
) the negatives of those values ( ) and
integer multiples of the sampling frequency added to those frequencies in
accordance with the following equation, where ( ) is the Fourier
Transform of x(t) and ( ) is the Fourier Transform of ( )
( ) ∑ [( )] ( 13. 2)
Note that both of the sinusoidal frequencies which make up ( )

and , are less than of the sampling frequency
( ). The first few frequencies which appear in
( ) when are:
135
Note that the original frequencies are there, and also note that there are no
frequencies between and other than the original
frequencies (and their negative values). Therefore, the original continuous-
time signal ( ) could be recovered from the samples using an ideal low-
pass filter with a bandwidth of . Even with practical filters, ( )
could be recovered fairly well since the next higher frequency is .
In the Figure. 2, note that the sample values are close enough together in
time to give an accurate picture of the continuous-time ( ).
Sample of ( ) with
Figure. 13.2. Sample of ( ) with

● Next, consider ( ). The
frequency content of the sampled signal ( ) is computed in the same
way as before. Both of the frequencies in ( ) are less than
136
, but now the higher frequency ( ) is closer to the top of

the primary strip ( ) than with the first sampling
period. The first few frequencies which appear in ( ) when
are:
Note that there are no frequencies between and except the

original frequencies, but now the filter used to recover ( ) from the
samples must be much closer to ideal than before since the next frequencies
outside the band of are at . In the Figure 13.3,
note that the samples exactly equal the ( ) values at the sample points, but
if you looked only at the sample values, you would not have as clear an
idea of what ( ) actually looks like as you would be the
sampling period. Sample of ( ) with
Figure. 13.3. Sample of ( ) with

137
● Now, consider ( ). Now the

higher frequency in ( ) ( ) is larger than .
Aliasing occurs in this situation. This means that there will be frequencies
in ( ) which have values between and which are not
in the original signal. "High" frequencies in the original signal ( in
this example) are appearing under the "alias" of low frequencies. For our
example, the frequency content in ( ) will have a term (
) in it, so the frequency will appear in ( )
(as well as ). The first few frequencis which appear in
( ) when are:
If an ideal low-pass filter with a cut-off frequency of

is used to process the discrete-time samples, the original
frequency of would be eliminated, and the frequency of
would be included. This produces a completely different signal than the
original ( ), even though the sample values correspond to ( ) exactly. If
we define the new signal by ( ).
( ) ( ) ( ) ( ) ( ) (12.3)
we can see that the samples with seconds exactly equal the
values of ( ) and ( ) at the sample points even though they are two
completely different signals see Figure 13.4(a,b). This illustrates the fact
that in order to recover the original continuous-time signal from discrete-
138
time samples, the sampling frequency must be at least twice as high as the
highest frequency in the continuous-time signal. The higher the sampling
frequency, the more reliable and easier will be the recovery -- compare the
sampled signals using . When the sampling frequency is less
than twice the highest frequency in the continuous-time signal, recovery
cannot be accomplished using only the sample values, even with ideal
filtering.
-a-
-b-
Figure 13.4(a,b). Sample of ( ) ( ) with
139
13.3. Stability of Sampled-Data Control Systems
For a satisfactory control system design, the closed-loop system must be

stable. The following definition of stability is normally used for sampled-
data systems:
Definition: A linear sampled-data system is stable if the output sequence

{ ( )}
Is bounded for any bounded input sequence { ( )} Otherwise the system
is said to be unstable.
Note that this definition contains no mention of the process response
during the inter sample periods. Specifically, it is theoretically possible to
have a system in which the continuous output is unbounded, even though a
sampled sequence of the output remains bounded. An example of such an
output is shown in Fig.1. The output hidden oscillation is rare in practice
and can be readily detected by changing the sampling period or by using
modified z-transforms. The necessary and sufficient condition for the
stability of a linear sample-data system can be expressed in two ways:
∑| ( |
2. ( ) has no poles on or outsides the unit circle in the z-plane.

In the above conditions ( ) is the pulse transfer function of the process
and { ( } is the set of the samples of the impulse response. The first
condition implies that the impulse response { ( } will be bounded for
the total time span ( ). Long division can be used to calculate
the value of ∑ | ( | but this may be time-consuming. It is preferable
to use the second condition, which is related to the partial fraction
expansion procedure.
140
For a pulse transfer function ( ) and input ( ), the output signal can be
calculated:
( ) ( ) ( ) ( ) (12.4)
( )( ) ( )
Then, using partial fraction expansion (distinct roots are assumed),
( )
[ ] (13.5)
Where, ri : partial fraction coefficient (residue). In a closed-loop system,
the terms shown in Equation (13.5) include the influence of the poles of the
closed-loop transfer function, plus the contribution of the set-point and/or
disturbance input (in brackets). The input terms appear separately from the
transfer function terms in the partial fraction expansion and are assumed to
be bounded. Hence, the stability of the loop can be determined by the
transfer function denominator terms. The sampled output response
corresponding to Equation (13.6) is
( ) ( ) ( ) ( ) [ ] (13.6)
If the absolute value of any the poles pi is greater than unity, ( )
as , thus violating the condition for stability. This is also true if the
root is a complex number (in polar form, if | | then
| | ).
0 𝑡 𝑡 𝑡 𝑡 𝑡 𝑡
Time
Figure 13.5. Continuous signal.
141
For both continuous-time and discrete-time system, the condition of

stability is determined by the roots of the characteristic equation (i.e., the
system poles). For continuous system the roots of the characteristic
equation must have negative real parts. Graphically, the roots must lie in
the left half of the plane, shown in Figure 13.6. We can map the stability
region in onto the complex plane. . Because the variable is a
complex number we can write The magnitude of is | |
.
Hence, if a root of continuous-time system lies in the left half of the

plane, then and the corresponding root in the plane is the less than
unity in magnitude, that is | | . Graphically, the region of stability in
the plane is the interior of the unit circle, since the left half of the plane
maps into the inside of the unity circle, of the plane.
Stability test: The roots of the characteristic equation can be computed
using a nonlinear equation solver to determine if a pulse transfer function is
stable. Alternatively, the frequency
Z plane Z plane
Imag
Imag
1 unstable
unstable Stable
Stable Real 1
Real -1
Unit circle
-1
Figure 13.6. Stability region in the s and z plane

142
Response of the pulse transfer function can be obtained. ( ),

allowing bode or Nyquist plots to be developed. Such methods can be used
for stability determination but do not readily lend themselves traditional
controller design (e.g., gain and phase margins).Either of these approaches
can be time-consuming for high-order process models. Hence alternative
techniques may be preferred for analyzing the stability of the characteristic
equation. These techniques include are:
1- Modified routh stability test.
2- Jury’s test.
3- Schur-cohn test.
13.4. Modified Routh Stability Criterion:

The Routh stability test was used to analysis the characteristic equation of a
continuous-time system for unstable roots, those roots that lie in the right
half of the complex plane. For sampled-data system, this test can also be
applied to determine whether any roots of the characteristic equation (in )
lie outside the unit circle. If a simple rational transfer-function from to
can be found that maps the interior of the unit circle into the left half of the
complex plane, then the Routh criterion can be used directly with -
transformrs.Such a transformation (or mapping) is provided by the bilinear
transformation.
(13.7)
This complex mapping or transformation does not corresponding to that for

the original -transform.It only approximates the original
transformation , hence the use of in place of s. However, the
boundary separating the stable and unstable regions does map exactly and
consequently the stability determination is exact. To apply this stability
143
test, first determine the characteristic equation of the sampled-data system

(written as positive power of ):
( ) (13.8)
Note that if the time delay is an integer multiple of , a polynomial in
will always result. Using the bilinear transformation, the characteristic
equation is transformed to a function of :
( ) ̅ ̅ ̅ ̅ (13.9)
Also yielding a polynomial in , where ̅ = real constant coefficient
( ). Note that the ̅ are not necessarily equal to the . The
Routh test can then be applied directly to Equation (13.9) to determine the
number of roots that lie in the unstable right half of the plane, or
equivalently, how many roots of ( ) lie outside the unit circle.
Example (1): Consider the characteristic equation
( )
Use the bilinear transformer to check stability
Solution
Substitution of the bilinear transformation
( ) ( ) ( ) ( )
After sum algebraic manipulation,
Examination of term shows that the necessary condition for stability is

satisfied because all coefficients are positive. The corresponding Routh
array is
144
Since no sign changes occur in column one, the system is stable. Equation
has actual roots of and , all of which lie inside the
unit circle.
Example (2): For the following system check the stability by modified
Routh stability?
( ) This is a third-order system ,
Solution
( ) ( ) ( )
or
( ) ( ) ( ) ( ) ( ) ( )
or
Since there is a negative sign in the characteristic equation in , the

equations has at least on root in the right-half plane.
Routh tabulation:
Since there are two sign changes in the first column of Rout
tabulation, ( ) has two roots outside the unit circle. The three roots at:
and .
145
Example: For the following system check the stability by modified Routh
stability?
( )
Solution
( ) ( ) ( ) ( )
Or
( ) ( ) ( ) ( )
Routh tabulation:
( )( )
Stability condition:
146
Chapter Fourteen
Analysis of control systems

in state space
This chapter discusses the following topics:
41.1. Introduction:
41.2. State Space Concepts (S.S)
41.3. Transfer Function (T.F):
14.4. Correlation between transfer functions and state-space equations.
14.5. Transfer Function from State-Variable Representation:
14.6. State Variable Representation from Transfer Function:
14.7. Properties of the State Transition Matrix:
14.8. Controllability, observability.
147
14.1. Introduction:
● Mechanical, electrical, thermal, hydraulic, economic, biological, etc, systems, may be

characterized by differential equations.
●The response of dynamic system to an input may be obtained if these differential equations
are solved.
●The differential equations can be obtained by utilizing physical laws governing a particular
system, for example, Newton's laws for mechanical systems, Kirchhoff's laws for electrical
systems, etc.
Mathematical models: The mathematical description of the dynamic characteristic of a
system. The first step in the analysis of dynamic system is to derive its model. Models may
assume different forms, depending on the particular system and the circumstances. In
obtaining a model, we must make a compromise between the simplicity of the model and the
accuracy of results of the analysis.
Transfer functions: The transfer function of a linear time-invariant system is define to be

the ratio of the Laplace transform ( z transform for sampled data systems) of the output to the
Laplace transform of the input (driving function), under the assumption that all initial
conditions are zero.
Example: Consider the linear time-invariant system
ao y ( n )  a1 y (n1)  ......  a n1 y .  a n y  bo x (m)  b1x (m1)  .....  bm1x.  bm x , n  m Taking the
Laplace transform of y(t)
(ao y (n) )  ao ( y (n) )  ao [S nY (S )  S n  1 y(0)  S n  2 y (0)  ......  y (n  1) (0)
(a1 y (n  1) )  a1( y (n  1) )  ao [ S n  1Y ( S )  S n  2 y (0)  S n  3 y (0)  ......  y (n  2) (0)
(a 2 y (n  2) )  a1( y (n  2) )  ao [ S n  2Y ( S )  S n  3 y(0)  S n  4 y (0)  ......  y (n  3) (0) .
148
(an y)  an ( y)  anY (s)
same thing is applied to obtain the L.T. of x(t).by substitute all initial condition to zero. The
transfer function of the system become.
bo S m  b1S (m1)  .....  bm1S  bm

Transfer function = G(s)  (14.1)
ao S n  a1S (n1)  .....  an1S  am
● Transfer function is not provide any information concerning the physical structure of the
system (the T.F. of many physically different system can be identical).
● The highest power of s in the denominator of T. F. is equal to the order of the highest
derivative term of the output. If the highest power of s is equal to n the system is called an
nth order system.
In general, variables that are functions of time are represented by lowercase letters. These are
sometimes indicated by the form x(t ), but more often this is written just as x. There are some
exceptions, because of established convention, in the use of certain symbols. To simplify the
writing of differential equations. The symbols D and1/D are defined by:
dy(t ) d 2 y(t )
Dy  , D2 y  (14.2)
dt dt 2
D1y 
1
y  0t y( )d 0  y( )d 0t y( )d Y0 (14.3)
D
where Yo represents the value of the integral at time t = 0, that is, the initial value of the
integral.
149
14.2. State Space Concepts:
Basic matrix properties are used to introduce the concept of state and the method of writing
and solving the state equations.
State: The state of a system is a mathematical structure containing a set of n variables x1(t ),
x2(t ), . . . , xi(t ), . . . , xn(t ), called the state variables, such that the initial values xi(to) of
this set and the system inputs uj(t ) are sufficient to describe uniquely the system’s future
response of t ≥ to. A minimum set of state variables is required to represent the system
accurately. The m inputs, u1(t ), u2(t ), . . . ,uj(t ), . . . ,um(t ), are deterministic; i.e., they
have specific values for all values of time t ≥ to.
Generally the initial starting time to is taken to be zero. The state variables need not be
physically observable and measurable quantities; they may be purely mathematical
quantities. The following additional definitions apply:
State Vector: The set of state variables xi(t) represents the elements or components of the n-
dimensional state vector x(t); that is,
 x1 (t )   x1 
 x (t )   x 
 2   2
X (t )   x3 (t )    x3   X (14.4)
   
.  . 
 xn (t )  xn 
   
The order of the system characteristic equation is n, and the state equation representation of
the system consists of n first-order differential equations. When all the inputs uj (t) to a
given system are specified for t> to, the resulting state vector uniquely determines the
system behavior for any t > to.
State Space: State space is defined as the n-dimensional space in which the components of
the state vector represent its coordinate axes.
150
State Trajectory: State trajectory is defined as the path produced in the state space by the
state vector x(t) as it changes with the passage of time. State space and state trajectory in the
two-dimensional case are referred to as the phase plane and phase trajectory, respectively.
The first step in applying these definitions to a physical system is the selection of the system
variables that are to represent the state of the system.
Note that there is no unique way of making this selection. The three common representations
for expressing the system state are the physical, phase, and canonical state variables.
The selection of the state variables for the physical-variable method is based upon the
energy-storage elements of the system. Table 1 lists some common energy-storage elements
that exist in physical systems and the corresponding energy equations. The physical variable
in the energy equation for each energy-storage element can be selected as a state variable of
the system. Only independent physical variables are chosen to be state variables.
Independent state variables are those state variables that cannot be expressed in terms of the
remaining assigned state variables. In some systems it may be necessary to identify more
state variables than just the energy-storage variables. This situation is illustrated in some of
the following examples, where velocity is a state variable. When position, the integral of this
state variable, is of interest, it must also be assigned as a state variable.
For the circuit of Series RLC Circuit (Fig.14.1). This circuit contains two energy-storage
elements, the inductor and capacitor. From Table 1, the two assigned state variables are
identified as x1=Vc (the voltage across the capacitor) and x2=i (the current in the inductor).
Thus two state equations are required.
Fig 14.1 is redrawn in Fig. 14.2 with node b as the reference node. The node equations for
node a and the loop equations are, respectively,
Cx .  x2
151
Lx2  Rx2  x1  u
.
Rearranging terms to the standard state equation format yields:
1
x1 
.
x2
C
(14.5)
1 R 1
x 2   x1  x 2  u
.
L L L
Equation (14.5) represents the state equations of the system containing two independent state
variables. Note that they are first-order linear differential equations and are n=2 in number.
They are the minimum number of state equations required to represent the system’s future
performance.
State equation. The state equations of a system are a set of n first-order differential
equations, where n is the number of independent states.
The state equation represented by Eq(14.5)is expressed in matrix notation as:
.
 X 1.  0 0 
1 / C   x1   
 .        1 u (14.6)
 X 2   1 / L  R / L   x2   L 
The standard form of the state space equation is:
X .  Ax  Bu
In this case ,the matrix A which is the system matrix will be :
0 1/ C 
A=   , nxn plant coefficient matrix
 1 / L  R / L 
Matrix B ,which is input matrix will be :
152
0 
B=  1  , nx1 control matrix
 
L
and, in this case, u=[u] is a one-dimensional control vector.
In ( X .  Ax  Bu ), matrix A and x are conformable. If the output quantity y(t) for the circuit
of Fig.8 is the voltage across the capacitor vC, then
y(t )  vc  x1
Thus the matrix system output equation for this example is:
 x1 
y (t )  Cx  Du  [1 0]   [0]u (14.7)
 x 2
Where
C is the output matrix with 1xn dimension for single input single output system (SISO),
D is the forward matrix =0.
For a multiple-input multiple-output (MIMO) system, with m inputs and l outputs, these
equations become:
X .  Ax  Bu
y  Cx  Du ; Where:
A=n×n plant or system matrix
B=n×m control or input matrix
C=l×n output matrix
153
D=l×m feed forward matrix
u = m- dimensional control vector
y = l- dimensional output vector
The block diagram of the state space representation can be shown in the Fig. 2.5.
Fig.14.3. Block diagram of the linear, continuous time control system represented in state
space.
14.3. Transfer Function (T.F):
If the system differential equation is linear, the ratio of the output variable to the input
variable, where the variables are expressed as functions of the D operator, is called the
transfer function.
Consider the system output vC=y in the RLC circuit of Fig.2.1 substituting i=CDvC into
Eq( LDi+Ri+1/CD i=e), yields: (LCD2+RCD+1)vc;(t)
The system transfer function is:
y (t ) vc (t ) 1
G ( D)    (14.8)
u (t ) e(t ) LCD  RCD  1
2
The notation G(D) is used to denote a transfer function when it is expressed in terms of the D
operator. It may also be written simply as G.
154
The block diagram representation of this system (Fig. 14.4) represents the mathematical
operation G(D)u(t)=y(t); that is, the transfer function times the input is equal to the output of
the block. The resulting equation is the differential equation of the system.
u(t) y(t)
G(D)
Fig.14.4. Block diagram representation.
Note: sometime used the symbol (s) instead of (D) and transfer function becomes writing as
(G(s)) and D≡s and D2≡s2 ……, and the equation (2.16) will write as:
1
G (14.9)
LCS 2  RCS  1
The program by using Matlab to change between the two forms for representation of control
system ( State Space and Transfer Function ) can be shown below;
% The change in form from SS to TF is:
Cy=1;Ly=10;Ry=100;
A=[ 0 /Cy -1/Ly -Ry/Ly ];
B=[0 1/Ly];
C=[1 0];
D=[0];
[num,den]=ss2tf(A,B,C,D)
% The change in form from TF to SS is:
numc=[0 0 1];
155
denc=[Ly*Cy Ry*Cy 1];
[AA,BB,CC,DD]=tf2ss(numc,denc)
In the Matlab/Simulink can be done as in the Fig. 14.5.
Fig.14.5. Simulink state space representation of control system
H.W. Implement the program and compare between two results?
14.4. Correlation between transfer functions and state-space equations
The following full derivation of transfer function of SISO system from the state-space
equations. Let us consider the system whose transfer function is given by:
Y ( s)
G( s)  (14.10)
U ( s)
This system may be represented in state space by the following equations:
X .  Ax  Bu (14.11)
y  Cx  Du
Where
x is the state vector ,u is the input and y is the output. The Laplace transform of the equation
2 is given by:
156
sx( s )  x(0)  Ax ( s )  Bu ( s ) 
 (14.12)
Y ( s )  Cx ( s)  Du ( s ) 
Since the transfer function is previously defined as Laplace transformation of the output to
the input with zero initial conditions, we assume that x(0)=0,then we have
sx( s )  Ax ( s )  Bu ( s ) or ( SI  A) X ( s )  BU ( s )
1
Multiplying ( SI  A) to both sides of the last equation we will obtain
X (s)  (SI  A) 1 BU (S ) (14.13)
Substitute (14.12) in (14.13) we get
Y (s)  [C (SI  A) 1 B  D]U (s) (14.14)
So that the transfer function of the system represented by state space will be:
G(s)  Y (s) / U (s)  [C (SI  A) 1 B  D] (14.15)
1
The right hand side of equation (14.15) involves ( SI  A) .Hence G(s) can be written as
G ( s )  Q( s ) / | ( SI  A) |
Where Q(s) is a polynomial in s .Therefore, | ( SI  A) | is equal to the characteristic

polynomial of G(s).in other words ,the Eigen values of A are identical to the poles of G(s).
157
Fig.14.6. Block diagram of state equation and output equation
From Fig.14.6. can be written the following states equations:
x1  y   0u
x2  y.   0u .  1u  x.1  1u
x3  y ..   0u ..  1u .  x. 2   2u
xn  y n1   0u n1  1u n2  .......... ..   n2u .   n1u  x. n1   n1u
Where  0 ,  1 ,  2 ,  n are determined from
 0  b0
1  b1  a1 0
 2  b2  a11  a2  0
 3  b3  a1 2  a2 1  a3  0
158
 n  bn  a1 n1  an2  n2  .....  an  0
With this choice of state variables, the existence and uniqueness of the solution of the state
equation is guaranteed. (Note that this is not the only choice of a set of state variables). With
the present choice of state variable, we obtain
x.1  x2  1u
x . 2  x3   2 u
x. n1  xn   n1u
x. n  an x1  an1 x2  ......  a1 xn   nu
In term of vector matrix equations, the output equation can be written as
 x .1  0 1 0 0 0   x1   1 
 .      
 x 2  0 0 1 0 0   x2    2 
 x . 3   0 0 0 1 0   x3     3  u
      
 x . n1  0 0 0 0 1   x . n1    n1 
 .  
 x n   a n  an1  a n  2  a n 3  a1   x . n    n 
 
 x1 
 
 x2 
 y   1 0 0 0 0 x3    0 u
 
 x . n 1 
 . 
x n 
Or
159
X .  AX  BU
Y  CX  DU
Where
 x1  0 1 0 0 0   1 
     
 x2  0 0 1 0 0   2 

X  x3  A  0 0 0 1 0  , B   3 
     
 x . n 1  0 0 0 0 1    n 1 
 .   a n  a n 1  a n  2  a n 3  a1   
x n   n 
C  [1 0 0 0 0]
D   0  b0
Note that the state space representation for the transfer function is
Y ( s ) b0 s n  b1 s n 1  b2 s n 2  bn
 n
U (s) s  a1 s n 1  a 2 s n 2  a n
Example (1): Obtain the state equations for the circuit of Fig.14.7. The output is the voltage
v1.The input or control variable is a current source i(t). The assigned state variables
are i1 , i2 , i3 , v1 , and v2 , ?
Fig.14.7. Circuit of Example 4.
160
Solution:
Three loop equations and two node equations are written:
v1  L1 Di1 or v1  L1
di1
dt
v2  L2 Di2  v1
v2  L3 Di3
i2  C1 Dv1  i1
i  i3  C2 Dv2  i2
L3i3  L2i2  L1i1  K
0 0  1 / C1  1 / C1  0 
0 0  L1 / C 2 L3  ( L2  L3 ) / C 2 L3  1 / C 
X 
. X   2
u
1 / L1 0 0 0  0 
   
  1 / L2 1 / L2 0 0  0 
y  1 0 0 0 X
14.5. Transfer Function from State-Variable Representation:
Having established the conditions for the equivalence of the state-variable representation
with that of the transfer-function, we are interested to find one representation from the other
by finding their relationship. Let us consider first the problem of determining the transfer
function of a system given the state variable representation
X . (t )  Ax(t )  Bu (t )
161
y (t )  Cx (t )
since the transfer-function representation is expressed in the frequency domain, we begin by

taking the Laplace transform of both equations, assuming as usual in transfer-function
determination that the initial conditions on x are all zero.
SX  AX ( s )  BU ( s ) (14.16)
Y ( s )  CX ( s ) (14.17)
Grouping the two X(s) terms in Equation (14.17) we have
(sI – A)X(s) = BU(s)
where the identity matrix has been introduced to allow the indicated multiplication
compatible. Now, pre-multiplying both sides of the above equation by (sI – A)-1, we get
X(s) = (sI – A)-1 BU(s)
We substitute this result in Equation (14.17) to obtain
Y(s) = C(sI – A)-1 BU(s)
Comparing this relation between Y(s) and U(s) with the Equation
(Y(s) = G(s)U(s))
we find that the transfer function matrix G(s) as:
G(s) = C(sI – A)-1 B
For the single input-single output case, this result reduces to
G(s) = c′(sI – A)-1 b
162
The matrix (sI – A)-1 is commonly referred to as the resolving matrix and is designated by
φ(s),
φ(s) = (sI – A)-1
In terms of this notation the Equations become
G(s) = C φ(s) B and G(s) =c′ φ(s) b (14.18)
Example (2): Consider the system represented by the equations
 0 1  0 
x . (t )    x(t )   u (t )
 10 7  1 
y(t )  1 0x(t )
The matrix (sI – A) in this example becomes
1 0  0 1  s 1 
s     
0 1  10  7  10 s  7 
Its inverse is found as
 s  7 1
 10 s 
adj( sI  A)
 ( s )  ( sI  A) 
1
 2 
det(sI  A) s  7 s  10
Hence, transfer function will be as
 s  7 1  0 
[1 0]
  10 s  1  1
G ( s )  c' ( s )b  ( sI  A) 
1
 2
s  7 s  10
2
s  7 s  10
In the above example, we observe that the determinant of the matrix (sI – A) is equal to the
denominator polynomial of G(s). This is always true for single input - single output systems.
163
Although Equation (14.18) provide a direct method for determining the transfer function of a
system from a state-variable representation of the system, it is generally not the most
efficient method.
14.6. State Variable Representation from Transfer Function:
In Section 14.5 we have shown how to get the transfer function model of a linear continuous
system when its state-variable form is available. We shall now take up the issue of getting
the state-variable model when the transfer function model is available. Since the state-
variable representation is not unique, there are, theoretically, an infinite number of ways of
writing the state equations. We shall present here one method for deriving a set of continuous
state variable representation from the transfer function. Analogous procedure may be
followed for writing the continuous state equation from pulse transfer function in S domain.
The transfer function of single-input-single-output system of the form:
a n 1 s n 1  a n 2 s n 2  ......  a1 s  a0
G(s) 
s n  bn 1 s n 1  bn 2 s n 2  .......  b1 s  b0
Can be written ,after introducing an auxiliary variable E(s) as
Y ( s) a s n 1  a s n 2  ......  a1 s  a0 E (s)
G( s)   n n 1 n 1 n2 n 2 *
U ( s ) s  bn 1 s  bn 2 s  .......  b1 s  b0 E ( s )
We let now
Y (s)  (an1s n1  an2 s n2  ......  a1s  a0 ) E(s)
U (s)  (s n  bn1s n1  bn2 s n2  .......  b1s  b0 ) E(s)
From Theorem of Laplace transform, we note the following relations between the variables
in the s domain and time domain with zero initial conditions
164
E ( s )  e(t )
sE(s)  e. (t )
s 2 E (s)  e.. (t )
Under this correspondence we define the state variables
x1 (t )  e(t )
x2 (t )  x1 (t )  e . (t )
.
x3 (t )  x2 (t )  e.. (t )
.
xn (t )  xn1 (t )  e n1 (t )
.
From above two Equations group we obtain the state equations
x .1 (t )  x2 (t )
x. 2 (t )  x3 (t )
x .3 (t )  x4 (t )
x. n (t )  xn (t )  b0 x1 (t )  b1 x2 (t )  b2 x3 (t )  bn1 xn (t )  u
In matrix notation this becomes

165
 x .1  0 1 0 0 0   x1  0
 .    
 x 2  0 0 1 0 0   x 2  0
 x . 3   0 0 0 1 0   x 3   0  u
      
 x . n 1  0 0 0 0 1   x . n 1  0
 .  
 x n   b0  b1  b2  b3  bn 1   x . n  1 
 
In compact form, it is written as :
x’(t) = A x(t) + Bu(t) ; The output equation is obtained from as
which may be written compactly a s y(t) = Cx(t)
Hence the last two Equations are a set of state equations for the continuous system described
by transfer function. Another convenient and useful representation of the continuous system
is the signal flow graph or the equivalent simulation diagram. These two forms can be
derived, after dividing both the numerator and denominator of first Equation by sn :
Y (s) a s n 1  a s n2  ......  a1 s  a0 E (s)

G(s)   n n 1 n 1 n 2 n2 *
U ( s ) s  bn 1 s  bn 2 s  .......  b1 s  b0 E ( s )
From this expression we can get two equations
Y (s)  (an1s n1  an2 s n2  ......  a1s  a0 ) E(s)
U (s)  (s n  bn1s n1  bn2 s n2  .......  b1s  b0 ) E(s)
The above equation can be rewritten as follows
E(s)  U (s)  bn1s 1 E(s)  bn2 s 2 E(s)  .......  b1s1n E(s)  b0 s n E(s)
Example (3): Let us consider a single input single output system of the last Example which
is reproduced below for quick reference:
166
 0
A
1

0 
,b    ,c  1 0, d  0 0
 10 7  1 
We are interested to find its solution with initial condition x′(t0) = x′(0) = [0 0] and unity
step input u(t) = us(t). The resolvent matrix φ(s) given by relation (φ(s) = (sI – A)-1 ) is
written as :
 s  7 1
 10 s 
adj( sI  A)
 ( s )  ( sI  A) 1   2 
det(sI  A) s  7 s  10
 s7 1 
 2  7 s  10 s  7 s  10 
2
 (s)   s 
  10 s 
 s 2  7 s  10 s 2  7 s  10 
1 5 2 1 1 1 
3 ( s  2  s  5) ( 
3 s2 s5
) 
 (s)   
10 ( 1  1 ) 1 5
( 
2 
)
 3 s  2 s  5 3 s  5 s  25 
1 2 5 1 2 
 3 (5e  2e ) (1e  1e 5 ) 
3
 (t )   
10 (1e  2  1e 5 ) 1
(5e  2e ) 
5 2
 3 3 
Substituting the value of x'(0)=[0 0] and unit step input in the equation we get
1 t 1
t

  (5e 
 2 ( t  )
 2e 5(t  ) )d (1e 2(t  )  1e 5(t  ) )d 
 0 
t
30 30
x(t )    (t   )bu( )d   t
10 1
t  1 
0
  (1e 2(t  )  1e 5(t  ) )d  (5e 5(t  )  2e  2 ( t  )
)d 
 3 0 30 
167
1 t  1 1
  (1e    e 2t  e 5t 
 2 ( t  )
 1e 5(t  ) )d 1
3
x(t )   t   10 6 15

0
10 1
t   1  2 t 1 5 t 
   (5e 5 ( t  )
 2e  2 ( t  )
)d   e  e 
 3 0 3 0   3 3
Therefore y(t) is computed as y(t)=cx(t)+du(t)
1 1  2 t 1 5t
y (t )   e  e ,t  0
10 6 15
14.7. Properties of the State Transition Matrix:
Some useful properties of the state transition matrix φ(t) are recorded below :
1.  (0)  e  I (identity matrix)

A0
2.  (t)  e  e (  At )   (t ) 1 or  (t )   (t )   (t )

At 1 1 1
A ( t t )
3.  (t1  t 2 )  e 1 2  e 1 .e 2   (t1 ) (t 2 )   (t 2 ) (t1 )
At At
4. ( (t ))   (nt)
n
5.  (t1  t 2 ) (t 2  t3 )   (t1  t3 ) for any t1, t2, t3
6.  (t )  A (t )
dt
14.8. Controllability and Observability
A system is said to be controllable at time t 0 if it is possible by means of an unconstrained

control vector to transfer the system from any initial state x(t0) to any other state in a finite
interval of time. A system is said to be observable at time t0 if, with the system in state x(t0),
it is possible to determine this state from the observation of the output over a finite time
interval. The concepts of controllability and observability were introduced by Kalman. They
168
play an important role in the design of control systems in state space. In fact, the conditions
of controllability and observability may govern the existence of a complete solution to the
control system design problem. The solution to this problem may not exist if the system
considered is not controllable. Although most physical systems are controllable and
observable, corresponding mathematical models may not possess the property of
controllability and observability. Then it is necessary to know the conditions under which a
system is controllable and observable.
169
170
171
172
H.W. Check the solution by fined G(s) by using the matrices of S.S. form?
173
174
Design By : Prof. Dr. Yousif Al Mashhadany
2020- 2021
Bode Diagram
100
80
60
40
20
Magnitude
0
(dB)
-20
-40
-60
-80
-100
180
150
120
90
Phase 60
30
(deg) 0
-30
-60
-90
-120
-150
-180
-210
-240
-270 -2 -1 0 1 2
10 10 10 10 10
Frequency (rad/s)

Control Theory II Lecture by Prof. Dr. Yousif Al Mashhadany

Uploaded by

Document Information

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

Control Theory II Lecture by Prof. Dr. Yousif Al Mashhadany

Uploaded by

Copyright:

Available Formats

University of Anbar Control Theory II

College of Engineering Prof. Dr. Yousif Al Mashhadany

2) “Classical Control Theory”, By Yousif Al Mashhadany, first Edition,

3) “Modern Control Engineering” , By Katsuhiko Ogata, Fifth Edition,

4) “Control Systems Engineering” , Sixth Edition, By Norman S. Nise, 2011

Lecture No. Eight

8.2. General Rules of Root Locus.

Therefore, s1 is a point on a branch of the locus. Similarly, it can be

Fig.1. Determination of the real-axis locus.

Rule 3: Locus End Points:

Rule 4: Asymptotes of Locus as s Approaches Infinity.

Plotting of the locus is greatly facilitated by evaluating the asymptotes

Remember that K in Eq.(7) is still a variable in the manner prescribed

Eq.(10) reveals that, no matter what magnitude s may have, after a

pole-zero pattern is symmetric about the asymptote line extended through

the s plane in order to approach zeros or the point at infinity. (Examples

When s3+s2 +2s is a minimum, -K is a minimum and K is a maximum.

When the derivative of W(s) is of higher order than 2,a digital-computer

3. The plot does not indicate a dip-and-peak combination or an inflection

In a similar manner the approach angle to a complex zero can be

Rule 7: Complex Pole (or Zero): Angle of Departure:

The next geometrical shortcut is the rapid determination of the direction

in this area are approximately constant. They can be considered to have

Arrangement shown in Fig.8, the approach angle c1 to the zero z1 is given

In other words, the direction of the locus as it leaves a pole or approaches

Rule 8: Imaginary-Axis Crossing Point:

The loop sensitivity term K is determined by setting the s1 row to zero:

In like manner, the crossover point can be determined for higher-order

Rule 9: Intersection or Non-intersection of Root-Locus Branches:

Fig. 9. illustrates these angles at s=-3 ,with θy=45o and λy=90o.

Ex.(2). Plot Root Loci for system shown below:

Ex.(4). Plot Root Loci for system shown below:

Ex.(5). Plot Root Loci for system shown below:

Frequency Response Analysis

The outline of the chapter is as follows:

10.1. Review of Frequency Response

10.2. Sketch the Polar plot of Frequency Response

10.3. Polar plot for some standard system

10.4. General Shape of polar plot

10.5. Nyquist Stability Criterion.

10.6. Solved Problem

10.1. Review of Frequency Response

To specify the behavior of a system to a sinusoidal input at a particular angular frequency ω

Figure.10.1 .Polar plot representation

10.2. Sketch the Polar plot of Frequency Response

1) the start of plot where ω = 0,

Consider a first order system

Where: τ is the time constant

Multiplying both numerator and denominator by the conjugate of denominator, i.e.,

The magnitude of G(jω), i.e., G(jω), is

And the phase of G(jω), denoted by φ, is

These are the same as Points 1 and 2.

This is the same as Point 2.

  tan 1 (1 / 1)  45

Figure.10.2. Sketching the polar plot for 1/(1+τs)

The magnitude and phase of G(jω) are then obtained as:

| G( j) | (Re l (G( j)))2  (Im g (G( j)))2

H.W. Sketching the Polar plot?

Figure.10.3. Nyquist plot of the given example

10.3. Polar plot for some standard system.

Figure.10.4. Nyquist plot

The polar plot of G(jw)=jw is the positive imaginary axis.

Figure.10.5. Nyquist plot of jw and 1/jw

The sinusoidal transfer function