Module 002 Block Diagram Algebra and Transfer Function: Frequency Domain Modeling

Download as pdf or txt
Download as pdf or txt
You are on page 1of 16

Feedback and Control Systems

1
Block Diagram Algebra and Transfer Function

Module 002 Block Diagram Algebra and


Transfer Function

At the end of this module you are expected to:


1. Represent a control system with block diagrams
2. Select controller parameters to meet design objectives
3. Understand the control theory for the analysis and design of a control
system

Frequency Domain Modeling


Signals are functions of time that represent the evolution of variables. There are two types
of signals, continuous-time signals and discrete-time signals. Continuous time signals are
functions of a continuous variable while discrete-time signals are functions of discrete
variable, that is, they are defined only for integer values of the independent variable (time
steps). It can be noted that the notation x(t) is often used to designate a continuous-time
signal and x(n) to designate a discrete-time signal.
A continuous-time signal x(t) is periodic if there exists a time T for which
Eq. 1 x(t) = x(t + T)
A discrete-time signal x(n) is periodic if there exists an integer N for which
Eq. 2 x(n) = x(n + N)
Exponential Signals
Exponential signals are important in signals and systems analysis since thay are
eigenfunctions of linear time-invariant system (they are invariant in form when
applied as an input to an LTI system.
Real Exponential Functions
Eq. 3 𝑥(𝑡) = 𝐶𝑒 𝑎𝑡 , 𝐶, 𝑎 𝑟𝑒𝑎𝑙
Case a = 0; The constant signal is taken x(t) = C
Case a> 0; The exponential tends to infinity as t →∞ (C > 0)
Complex Exponential Signals
𝑥(𝑡) = 𝐶𝑒 𝑎𝑡
C, a complex, 𝐶 = |𝐶|𝑒 𝑗𝜃
Course Module
Feedback and Control Systems
2
Block Diagram Algebra and Transfer Function

𝑡
Eq. 4 𝑥(𝑡) = 𝐶𝑒 𝑎𝑡 = |𝐶|𝑒 𝑗𝜃 𝑒 (𝑟+𝑗𝜔0 ) = |𝐶|𝑒 𝑟𝑡 𝑒 𝑗(𝜔0 𝑡+𝜃)
Using Euler’s relation
Eq. 5 𝑥(𝑡) = |𝐶|𝑒 𝑟𝑡 cos(𝜔0 𝑡 + 𝜃) + 𝑗|𝐶|𝑒 𝑟𝑡 sin(𝜔0 𝑡 + 𝜃)
2𝜋
For r = 0, a complex periodic signal or period 𝑇 = whose real and
𝜔0
imaginary parts are sinusoidal is obtained
Eq. 6 𝑥(𝑡) = |𝐶|𝑒 𝑗𝜃 𝑒 𝑗𝜔0𝑡 = |𝐶| cos(𝜔0 𝑡 + 𝜃) + 𝑗|𝐶|sin(𝜔0 𝑡 + 𝜃)
For r < 0, a complex periodic signal multiplied by a decaying exponential
whose real and imaginary parts are “damped sinusoids” is obtained while for
r > 0, a complex periodic signal multiplied by a growing exponential is
obtained.
In control system design the most common mathematical models of the behavior of interest
are, in time domain, linear ordinary differential equations with constant coefficients and in
the frequency or transform domain, transfer functions obtained from time domain
descriptions through Laplace transform.
Derivation of the Transfer Function
Transfer Functions
The transfer function of a linear time-invariant system is the ratio of the Laplace
transform of the output Y(x) to the Laplace transform of the corresponding input U(x)
with all initial conditions assumed to be zero.
From differential equations to transfer function: Let the equation
𝑑2 𝑦(𝑡) 𝑑𝑦(𝑡)
+ 𝑎1 + 𝑎0 𝑦(𝑡) = 𝑏0 𝑢(𝑡)
𝑑𝑡 2 𝑑𝑡

With some initial conditions


𝑑𝑦(𝑡) 𝑑𝑦(0)
𝑦(𝑡)|𝑡=0 = 𝑦(0) 𝑎𝑛𝑑 |= =0
𝑑𝑡 𝑑𝑡

Describe a process of interest. Taking Laplace transform of both sides, the following
is obtained
[𝑠 2 𝑦(𝑠) − 𝑠𝑦(0) − 𝑦(0)] + 𝑎1 [𝑠𝑌(𝑠) − 𝑦(0)] + 𝑎0 𝑌(𝑠) = 𝑏0 𝑈(𝑠)
Where Y(s) = L{y(t)} and U(s) = L{u(t)}. Combining terms and solving with respect to
Y(s) the following is obtained:
𝑏0 (𝑠+𝑎1 )𝑦(0)+𝑦(0)
𝑌(𝑠) = 𝑈(𝑠) +
𝑠 2 +𝑎1 𝑠+𝑎0 𝑠 2 +𝑎1 𝑠+𝑎0

Assuming the initial conditions are zero.


𝑌(𝑠) 𝑏0
Eq. 7 = 𝐺(𝑠) =
𝑈(𝑠) 𝑠 2 +𝑎1 𝑠+𝑎0

Course Module
Feedback and Control Systems
3
Block Diagram Algebra and Transfer Function

Where G(s) is the transfer function of the defined system.


The concern is the transfer functions G(s) which are rational functions, i.e., ratios of
polynomials s, G(s) = n(s)/d(s). The interest is in proper G(s) where lim 𝐺(𝑠) < ∞.
𝑎→∞
Proper G(s) have degree n(s) ≤ degree d(s).
In most cases, degree n(s) < degree d(s), which case G(s) is called strictly proper.
Considering the transfer function
𝑏𝑚 𝑠 𝑚 +𝑏𝑚−1 𝑠 𝑚−1 +⋯+𝑏1 𝑠+𝑏0
𝐺(𝑠) = 𝑤𝑖𝑡ℎ 𝑚 ≤ 𝑛
𝑠 𝑛 +𝑎𝑛−1 𝑠 𝑛−1 +⋯+𝑎1 𝑠+𝑎0

This is transfer function assuming that all initial conditions are zero, is the result of
taking the Laplace transform of both sides of the time domain equation.

Block Diagrams and Transfer Function


A system is a mathematical relationship between an input signal and an output signal. It can
be represented by an abstract operator equation y = Hx, where H is an operator that may be
represent anything from a simple gain to a complicated nonlinear differential equation.
System Block Diagrams
Systems may be interconnections of other systems as in Figure 1, while Figure 2 is a
cascade interconnection and parallel interconnection is shown in Figure 3, the
feedback interconnection is in Figure 4.

(a) Interconnected Systems

Course Module
Feedback and Control Systems
4
Block Diagram Algebra and Transfer Function

(b) Cascade Interconnection

(c) Parallel Interconnection

Course Module
Feedback and Control Systems
5
Block Diagram Algebra and Transfer Function

(d) Feedback Interconnection


Figure 1. Block Diagrams of Interconnected Systems

Source: Basic Dynamic and Control (4), by: Finn Haugen, 2014, Skein, Norway,
https://home.usn.no/finnh/publications/books/basic_dynamics_and_control/Basic_Dyn_Con_Textboo
k_free. Retrieved December, 2018

Manipulating Block Diagrams

A block diagram for a system can be manipulated into new forms. A block diagram is
developed for a system. This developed diagram is then simplified through a process
that is both graphical and algebraic. The following are equivalent blocks with its
algebraic proof.

r
G(D) c

+ -

H(D)

Is equal to

r 𝐺(𝐷) c
Course Module
1 + 𝐺(𝐷)𝐻(𝐷)
Feedback and Control Systems
6
Block Diagram Algebra and Transfer Function

Figure 2. A Negative Feedback Block Reduction

The algebraic proof would be:


𝑐 = 𝐺(𝐷)𝑒
𝑐 = 𝐺(𝐷)(𝑟 − 𝐻(𝐷)𝑐)
𝑐(1 + 𝐺(𝐷)𝐻(𝐷) = 𝐺(𝐷)𝑟
𝐺(𝐷)
𝑐 = [1+𝐺(𝐷)𝐻(𝐷)] 𝑟
𝐺(𝐷)
For a positive feedback block reduction, where the process would be 1−𝐺(𝐷)𝐻(𝐷) the
algebraic proof is
c = G(D) e
c = G(D)(r+H(D)c)
c (1-G(D)H(D) = G(D) r
𝐺(𝐷)
𝑐 = [1−𝐺(𝐷)𝐻(𝐷)] 𝑟

The blocks can be reversed as in Figure 3.

r c

Is equal to

c r
1
𝐺(𝐷)
Figure 3. Reversal of Function Blocks

The algebraic proof is


c = G(D) r

Course Module
Feedback and Control Systems
7
Block Diagram Algebra and Transfer Function

1
𝑟 = 𝑐 [𝐺(𝐷)]

Figure 4a shows the branches are moved before the blocks while in Figure 4b, the
branches are moved after the blocks.
r c
r c
G(D) Is equal to G(D)
) G(D)
c c
(a) Before blocks

r c c
Is equal to r
G(D) G(D)
1
r 𝐺(𝐷)

(b) After Blocks


r
Figure 4 Moving Branches

Figure 5 shows the combining sequential function blocks, Figure 6 is moving the
summation functions before blocks while shown in Figure 7 is moving summation
past blocks.

r x b
G(D) H(D)

Is equal to

r b
G(D)H(D)

Figure 5. Combining Sequential Blocks

The algebraic proof of combining sequential blocks is


x = G(D)r
b = H(D) x
b = H(D)G(D) r

Course Module
Feedback and Control Systems
8
Block Diagram Algebra and Transfer Function

r ± e r ± e
G(D) G(D)
Is equal to
± ±
b
1 b
𝐺(𝐷)
c

Figure 6. Moving Summation Functions Before Blocks

The algebraic proof of moving the summation functions before blocks is


e = ±G(D)r ± b
𝑏
𝑒 = 𝐺(𝐷) [±𝑟 ± ±𝐺(𝐷)]

r e
r e
G(D) G(D)

d Is equal to d

G(D)

Figure 7. Moving Summation Functions Past Block

And the algebraic proof of moving the summation functions past block is
e = G(D) (±R ±d)
e = ±R(G(D)) ± d(G(D))

The Bode Plot


It is convenient to use a logarithmic scale to plot the magnitude of a frequency response. The
reason is that frequency responses can have a wide dynamic range which covers many
orders of magnitude. Some frequencies may be amplified by a factor of 1000 while others
may be attenuated by a factor of 10-4. Another reason is that if log scale is used, the
magnitudes of the cascaded Fourier transforms can be added rather than multiplied making
easier graphically.

Course Module
Feedback and Control Systems
9
Block Diagram Algebra and Transfer Function

Eq. 8 log|𝑌(𝑗𝜔)| = log|𝐻(𝑗𝜔)| + log |𝑋(𝑗𝜔)|


The bel was defined as a power amplification of |H(jω)|2 = 10 for a system. Thus, for the
system with power gain of 10 at frequency ω,
𝑑𝐵
Eq. 9 10 𝑏𝑒𝑙 𝑥 log10 |𝐻(𝑗𝜔)|2 𝑏𝑒𝑙 = 10 log10 10𝑑𝐵 = 10𝑑𝐵
But to measure the actual linear gain of a system the following identity is used
Eq. 10 10 log10 |𝐻(𝑗𝜔)|2 𝑑𝐵 = 20 log10 |𝐻(𝑗𝜔)𝑑𝐵
Therefore, a magnitude plot of |H(jω)| is represented as 20log10|H(jω)| dB using a linear scale
in dB units. It is also convenient to use a log scale for the frequency as features of a frequency
response can be spread over a wide frequency band.
Bode Plot
A Bode plot is the combination of a magnitude and phase plots using log scales for the
magnitude and the frequency, and a linear scales (radians or degrees) for the phase.
Only positive frequencies are considered. The Bode plot is useful since the overall
frequency response of cascaded systems is the graphical addition of the Bode plots of
the individual systems. This property is used to hand sketch a Bode plot of a rational
transfer function in pole-zero form by considering each first-order factor
corresponding to a pole or a zero to be an individual its own Bode plot.
Example of first-order: Consider the first-order system with transfer function
1
𝐻(𝑠) = , 𝑅𝑒{𝑠} > −2
𝑠+2

Which has the frequency response


1
𝐻(𝑗𝜔) = 𝑗𝜔+2

It is written as the product of a gain and a first-order transfer function with unity gain
at dc:
1 1
𝐻(𝑗𝜔) = 2 𝑗𝜔
+1
2

The break frequency is 2 radians, the Bode magnitude plot is the graph of
1 1
20 log10 |𝐻(𝑗𝜔)| = 20 log10 |2| + 20 log10 |𝑗𝜔 | 𝑑𝐵
+1
2

𝑗𝜔
= −20 log10 2 − 20 log10 | + 1| 𝑑𝐵
2
𝑗𝜔
= −6𝑑𝐵 − 20 log10 | 2 + 1| 𝑑𝐵
For low frequencies, ω<<2
20 log10 |𝐻(𝑗𝜔)| ≈ −6𝑑𝐵 − 20 log10 |1|𝑑𝐵 = −6𝑑𝐵

Course Module
Feedback and Control Systems
10
Block Diagram Algebra and Transfer Function

For high frequencies, ω >>2


𝜔
20 log10 |𝐻(𝑗𝜔)| ≈ −6𝑑𝐵 − 20 log10 | 2 | 𝑑𝐵
= −6𝑑𝐵 − 20 log10 |𝜔|𝑑𝐵 + 20 log10 2 𝑑𝐵
= −20 log10 |𝜔|𝑑𝐵
For ω = 10, -20dB is the result, for ω = 100, -40dB, etc. The slope of this asymptote is
therefore -20dB/decade.
With the asymptotes meeting at the break frequency 2 radians/s, the magnitude Boe
plot can be sketched as follows (the dashed line is the actual magnitude).

Figure 8. Bode Magnitude Plot for First-Order System

Source: Basic Dynamic and Control (12), by: Finn Haugen, 2014, Skein, Norway,
https://home.usn.no/finnh/publications/books/basic_dynamics_and_control/Basic_Dyn_Con_Textboo
k_free. Retrieved December, 2018

The Bode phase plot is the graph of


−𝜔
1 1 1
∠𝐻(𝑗𝜔) = ∠ 𝑗𝜔 = ∠ 𝑗𝜔 = arctan ( 2 )
2 +1 +1 1
2 2

The phase is 0 at ω = 0 and –π/2 at ω = ∞. The linear (log frequency scale)


approximation to the phase that helps to sketch is given by
2
0 , 𝜔 ≤ 10
𝜋 𝜔 2
∠𝐻 (𝑗𝜔) = − 4 [log10 ( 2 ) + 1] , < 𝜔 < 20
10
𝜋
− 2 , 𝜔 ≥ 20
{
𝜔 𝜋 𝜋
The rule of the thumb is; the phase is 0 at 0 , − at ω0 and - at 10ω0 where ω0 is the
10 4 2
break frequency. The Bode plot is as shown in Figure 9.
Course Module
Feedback and Control Systems
11
Block Diagram Algebra and Transfer Function

Figure 9. Bode Phase Plot

Source: Basic Dynamic and Control (14), by: Finn Haugen, 2014, Skein, Norway,
https://home.usn.no/finnh/publications/books/basic_dynamics_and_control/Basic_Dyn_Con_Textboo
k_free. Retrieved December, 2018

The Bode magnitude plot of s + 2 (the inverse of H(s)) is the magnitude plot in
Figure 8, flipped around the frequency axis likewise its phase plot.
An example of the second-order system. Consider the second-order system
with transfer function
𝑠
𝑠+100 𝑠+100 +1
100
𝐻(𝑠) = = = 10 𝑠 , 𝑅𝑒{𝑠} > −1
(𝑠 2 +11𝑠+10) (𝑠+1)(𝑠+10) (𝑠+1)( +1)
10

Which has the frequency response


𝑗𝜔
+1
100
𝐻(𝑗𝜔) = 10 𝑗𝜔
(𝑗𝜔+1)( 10 +1)

The break frequencies are 1, 10 and 100 radians/s. The Bode magnitude plot
is the graph of
𝑗𝜔 1 1
20 log10 |𝐻(𝑗𝜔)| = 20 log10 |10 |+20 log10 | + 1| + 20 log10 |𝑗𝜔 | + 20 log10 | | 𝑑𝐵
100 +1 𝑗𝜔 +1
10

At low frequencies ω <<1,


20 log10 |𝐻(𝑗𝜔)| ≈ 20𝑑𝐵 + 20 log10 |1| − 20 log10 |1| − 20 log10 |1|𝑑𝐵 = 20𝑑𝐵
= 20 + 20 log10 |𝜔| − 40 − 20 log10 |𝜔| + 20 − 20 log10 |𝜔|𝑑𝐵
= −20 log |𝜔|𝑑𝐵
The asymptotes of each first order term can be plotted on the same magnitude
graph and then they are added together to obtain the Bode magnitude plot.

Course Module
Feedback and Control Systems
12
Block Diagram Algebra and Transfer Function

Figure 10. Bode Magnitude Plot

Source: Basic Dynamic and Control (15), by: Finn Haugen, 2014, Skein, Norway,
https://home.usn.no/finnh/publications/books/basic_dynamics_and_control/Basic_Dyn_Con_Textboo
k_free. Retrieved December, 2018

The same procedure is done to plot the Bode phase plot.

Figure 11. Bode Phase Plot

Source: Basic Dynamic and Control (15), by: Finn Haugen, 2014, Skein, Norway,
https://home.usn.no/finnh/publications/books/basic_dynamics_and_control/Basic_Dyn_Con_Textboo
k_free. Retrieved December, 2018

Course Module
Feedback and Control Systems
13
Block Diagram Algebra and Transfer Function

Laplace Transforms
Along with the transfer function is the use of Laplace transform and the Fourier Transform
for continuous time signals.
The Laplace transform of a signal x(t) is defined as
+∞
Eq. 11 𝑋(𝑠) = ∫−∞ 𝑥(𝑡)𝑒 −𝑠𝑡 𝑑𝑡
Where s is a complex variable. If x(t) is given, this integral may or may not converge,
depending on the value of 𝜎 (the real part of s).
Example is the Laplace transform of 𝑥(𝑡) = 𝑒 −𝑎𝑡 𝑢(𝑡), 𝑎 𝑟𝑒𝑎𝑙
+∞
𝑋(𝑠) = ∫−∞ 𝑒 −𝑎𝑡 𝑢(𝑡)𝑒 −𝑎𝑡 𝑑𝑡
+∞
= ∫0 𝑒 −(𝑠+𝑎)𝑡 𝑑𝑡
1
= , 𝑅𝑒{𝑠} > −𝑎
𝑠+𝑎

Inverse Laplace Transform


The inverse Laplace transform is generally given as
1 𝜎+𝑗𝜔
Eq. 12 𝑋(𝑡) = ∫ 𝑋(𝑠)𝑒 𝑠𝑡 𝑑𝑠
2𝜋𝑗 𝜎−𝑗𝜔

The partial fraction expansion technique is used to find the continuous time signal
corresponding to a Laplace transform.
An example of this is assuming no multiple order poles in the set of poles {𝑝𝑘 }𝑚
𝑘=1 of
the rational transform X(s), and assuming that the order of the denominator
polynomial is greater than the order of the numerator polynomial, X(s) can be
expanded in the form
𝐴
𝑋(𝑠) = ∑𝑚 𝑘
𝑘=1 𝑠−𝑝
𝑘

From the ROC of X(s), the ROC of each individual terms of the equation above can be
found and then the inverse transform of each of these terms can be determined from
the table of Laplace transform pairs. If the ROC is to the right of the pole at s = pi, then
the inverse transform of this term is 𝐴𝑖 𝑒 𝑝𝑖 𝑡 𝑢(𝑡), a right-sided signal. However, if the
ROC is to the left of the pole at s = pi, the inverse transform of this term is
−𝐴𝑖 𝑒 𝑝𝑖 𝑡 𝑢(−𝑡), a left-sided signal. Adding the inverse transforms of the individual
terms will give the inverse transform of X(s).
Convolution Property of the Laplace Transform
ℒ ℒ
If 𝑥1 (𝑡) ↔ 𝑋1 (𝑠), 𝑠 ∈ 𝑅𝑂𝐶1 and 𝑥2 (𝑡) ↔ 𝑋2 (𝑠), 𝑠 ∈ 𝑅𝑂𝐶2 then
+∞ ℒ
Eq. 13 ∫−∞ 𝑥1 (𝑇)𝑥2 (𝑡 − 𝑇)𝑑𝑇 ↔ 𝑋1 (𝑠)𝑋2 (𝑠), 𝑠 ∈ 𝑅𝑂𝐶 ⊇ 𝑅𝑂𝐶1 ⋂𝑅𝑂𝐶2

Course Module
Feedback and Control Systems
14
Block Diagram Algebra and Transfer Function

This is a very useful property for Linear Time Invariant (LTI) system analysis. It is
important to note that the resulting ROC includes the intersection of the two original
ROC’s, but it may be larger, example, is when a pole-zero cancellation occurs.
The time shifting property of the Laplace Transform can be demonstrated as follows:
ℒ ℒ
if 𝑥(𝑡) ↔ 𝑋(𝑠), 𝑠 ∈ 𝑅𝑂𝐶 , then 𝑥(𝑡 − 𝑡0 ) ↔ 𝑒 −𝑠𝑡0 𝑋(𝑠), 𝑠 ∈ 𝑅𝑂𝐶.
The Fourier Transform
The Fourier transform or spectrum of a signal x(t) is defined as:
+∞
Eq. 14 𝑋(𝑗𝜔) = ∫−∞ 𝑥(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡
The inverse Fourier transform is given by
1 +∞
Eq. 15 𝑥(𝑡) = ∫ 𝑋(𝑗𝜔)𝑒 𝑗𝜔𝑡 𝑑𝜔
2𝜋 −∞

The Fourier transform describes the frequency contents of a signal. If the ROC of the
Laplace transform X(s) contains the imaginary axis, then the Fourier transform of x(t)
exists and is given by 𝑋(𝑗𝜔) = 𝑋(𝑠)|s=jω
Transfer Functions of LTI Systems
The convolution property makes the Laplace transform useful in obtaining the
response of an LTI system to an arbitrary input. The Laplace transform of the output
of an LTI system with impulse response h(t) is given by
Eq. 16 𝑌(𝑠) = 𝐻(𝑠)𝑋(𝑠), 𝑅𝑂𝐶𝑌 ⊇ 𝑅𝑂𝐶𝐻 ⋂𝑅𝑂𝐶𝑋
Also, the frequency response of the system is H(jω) = H(s)|s=jω. The Laplace transform
H(s) of the impulse response of an LTI system is called the transfer function.
Stability
Bounded-input bounded-output (BIBO) stability of a continuous-time LTI
system is equivalent to its impulse response being integrable, in which case its
Fourier transform converges. Also, the stability of an LTI differential system is
equivalent to having all the zeros of its characteristic polynomial having a
negative real part. A causal system with rational transfer function H(s) is stable
if and only if all its poles are in the left-half of the s-plane, that is, all the poles
have negative real parts.
Transfer Functions of LTI Differential Systems
It was discussed that the transfer function of an LTI system is the Laplace transform
of its impulse response. For a differential LTI system, like the frequency response, the
transfer function of the differentiation operator s-1 represents the integration
operator.
Taking the general form of an LTI differential system

Course Module
Feedback and Control Systems
15
Block Diagram Algebra and Transfer Function

𝑑𝑘 𝑦(𝑡) 𝑑𝑘 𝑥(𝑡)
∑𝑁
𝑘=0 𝑎𝑘 = ∑𝑀
𝑘=0
𝑑𝑡 𝑘 𝑑𝑡 𝑘

Using the differentiation and linearity properties of the Laplace transform to obtain
the transfer function H(s) = Y(s)/X(s)
∑𝑁 𝑘 𝑀 𝑘
𝑘=0 𝑎𝑘 𝑠 𝑌(𝑠) = ∑𝑘=0 𝑏𝑘 𝑠 𝑋(𝑠)
𝑌(𝑠) ∑𝑀 𝑏 𝑠𝑘
Eq. 17 𝐻(𝑠) = = ∑𝑁𝑘=0 𝑎𝑘 𝑠𝑘
𝑋(𝑠) 𝑘=0 𝑘

The differential equation used can have many different impulse responses. If it is
known that the differential system is causal, then the ROC is the right half-plane to
the right of the rightmost pole in the s-plane.
Transfer Function of LTI State-Space Systems
Assume the general causal continuous-time state space system is initially at rest
ẋ = Ax + Bu
y = Cx + Du
Taking the Laplace transform on both sides, the following is obtained
sX(s) = AX(s) + BU(s)
Y(s) = CX(s) + DU(s)
Solving for the Laplace transform of the output,
Y(s) = [C(sIn –A)-1 B + D] U(s)
The transfer function of the system is therefore given by
Eq. 18 𝐻(𝑠) = 𝑈(𝑠) ↦ 𝑌(𝑠) = 𝐶(𝑠𝐼𝑛 − 𝐴)−1 𝐵 + 𝐷
Poles and Zeros of the Transfer Function
Let 𝑁(𝑠) = ∑𝑀 𝑘=0 𝑏𝑘 𝑠 be the numerator polynomial of the transfer function H(s) in
𝑘

Equation 17 and let 𝐷(𝑠) = ∑𝑁 𝑘=0 𝑎𝑘 𝑠 be its denominator polynomial. Then the poles
𝑘

of H(s) are the N roots of the characteristic equation D(s) = 0 (or the N zeros of the
characteristic polynomial). The zeros of H(s) are the M roots of the equation N(s)=0.
As an example, find the transfer function pole of the following causal differential LTI
system
𝑑2 𝑦(𝑡) 𝑑𝑦(𝑐𝑡) 𝑑2 𝑥(𝑡) 𝑑𝑥(𝑡)
+ 𝜔𝑐 √2 + 𝜔𝑐2 𝑦(𝑡) = + 𝜔𝑐 √2
𝑑𝑡 2 𝑑𝑡 𝑑𝑡 2 𝑑𝑡

Taking the Laplace transform on both sides, the result is


𝑠 2 𝑌(𝑠) + 𝜔𝑐 √2𝑠𝑌(𝑠) + 𝜔𝑐2 𝑌(𝑠) = 𝑠 2 𝑋(𝑠) + 𝜔𝑐 √2𝑠𝑋(𝑠)
Which will give

Course Module
Feedback and Control Systems
16
Block Diagram Algebra and Transfer Function

𝑌(𝑠) 𝑠 2 +𝜔𝑐 √2𝑠


𝐻(𝑠) = 𝑋(𝑠) = 𝑠2 +𝜔 𝑐 √2𝑠+𝜔𝑐2

The poles are the roots of 𝑠 2 + 𝜔𝑐 √2𝑠 + 𝜔𝑐2 = 0. Identifying the coefficients with the
standard second-order denominator 𝑠 2 + 2 𝜍𝜔𝑐 𝑠 + 𝜔𝑛2 , the damping ratio can be
√2
found to be 𝜍 = = 0.707 and the natural frequency ωn = ωc. the poles of H(s),are
2
then
3𝜋
√2 √2
𝑝1 = −𝜍 𝜔𝑛 + 𝜔𝑛 √𝜍 2 − 1 = − 𝜔𝑐 +𝑗 𝜔𝑐 = 𝜔𝑐 𝑒 𝑗 4
2 2
3𝜋
√2 √2
𝑝2 = −𝜍 𝜔𝑛 − 𝜔𝑛 √𝜍 2 − 1 = − 2
𝜔𝑐 −𝑗 2
𝜔𝑐 = 𝜔𝑐 𝑒 −𝑗 4
And its zeros are
𝑧1 = −𝜔𝑐 √2, 𝑧2 = 0
The poles and zeros of a transfer function are signified with the symbols ‘X’ and ‘O’
respectively on a pole-zero plot in the s-plane.

References and Supplementary Materials


Books and Journals
1. Kirsten Morris; 2015; Introduction to Feedback Control; San Diego, California;
Academic Press
2. Endra Joelianto and Darrell Williamson; Transient Response Improvement of
Feedback Control Systems using Hybrid Reference Control; International Journal of
Control; Volume 82; Issue 10; 2017
Online Supplementary Reading Materials
1. Introduction to Feedback Control Theory;
http://www.cds.caltech.edu/~murray/books/AM08/pdf/am06-
complete_16Sep06.pdf; December 2018
Online Instructional Videos
1. Laplace Transform Explained and Visualized in 3D;
https://www.youtube.com/watch?v=6MXMDrs6ZmA; January, 2019

Course Module

You might also like