EMS507 Control System

Analysis and Design

Week 2 – Transfer function and block
diagrams

Yunpeng Zhu
Transfer function and block diagrams
© Copyright 2023 Yunpeng Zhu. All Rights Reserved
Edition: v1.1
Table of Contents
1 Laplace Transform ............................................................................... - 1 -
1.1 Development of Laplace Transform ....................................................... - 1 -
1.2 Typical Laplace Transforms .................................................................... - 4 -
2 Transfer functions ................................................................................ - 6 -
2.1 The concept ........................................................................................... - 6 -
2.2 Zeros and Poles ...................................................................................... - 9 -
3 Block diagrams................................................................................... - 11 -
3.1 Concepts .............................................................................................. - 11 -
3.2 Block diagram connections .................................................................. - 13 -
3.2 Block diagram algebra .......................................................................... - 15 -

i
1 Laplace Transform
1.1 Development of Laplace Transform

We have tried to solve differential equation models in the previous lectures,

for example,

d 2 x(t ) dx(t )
  x(t )  0
dt 2 dt
with one possible solution of x(t )  A1e(   i)t  A2e(   i)t .

 Let’s watch a short video to recap the dynamics of an oscillator (1 min).

Consider the 4th case where the system is subject to an input:

What is the response looks like in this case?

Let’s guess:
The input loads can be seen as a set of impulse:

-1-
The system is linear, so the response under this input is the sum of responses
under those impulses.

The above illustration is not rigorous, but it shows the basic idea of using
transforms: To solve differential equations.

In 1744, Euler evaluated a transform X (a)  f [ x(t )]   x(t )eat dt to solve

differential equations. But he didn’t pursue this matter very far. Instead he
turned to study how things vibrate.
40 years later, until 1785 (when Euler had dead), Laplace
developed a new transform based on Euler’s form as:

X ( s )  L[ x(t )]   x(t )e  st dt , s    j
0

The lower bound start from  is called the bilateral Laplace

Transform, which is not considered in this lecture.

-2-
Both Euler’s transform and Laplace transform can solve differential equations
(In fact, there are many famous and less famous transforms to solve the same
problem), why the Laplace Transform is so important?
 Because the Laplace Transform L[ x(t )]  X ( s) has beautiful properties:

dx(t )
L[ ]  sL[ x(t )]  f (0)  sX ( s)  f (0)
dt
1. Differentiation Consider the initial condition f (0)  0 ,
d n x(t )
L[ n
]  s n X (s)
dt
lim x(t )  lim{sX ( s)}
t  s 0
2. Final Value
Theorem This result can be used to allow rapid
calculation of steady state gain.
lim x(t )  lim{sX ( s)}
t 0 s 
3. Initial Value
Theorem This result can be used to allow rapid
calculation of instantaneous gain.

4. Multiplication L[ Ax(t )]  AX ( s)
by a Constant e.g. If L[ x(t )]  1 s , L[5 x(t )]  5 s
L[ x1 (t )  x2 (t )]  X 1 ( s)  X 2 ( s)
5. Sum of terms If L[ x1 (t )]  1 s , L[ x2 (t )]  1 s 2 ,
e.g.
L[ x1 (t )  x2 (t )]  1 s  1 s 2

L[ x(t   )]  e  s X ( s )
A premultiplying exponential in the
6. Real time shift Laplace domain can be related to a time
shift (e.g., a signal delay) in the time
domain.
Proof of Laplace Transform properties:
See ‘EMS507 supplementary material - Proof of Laplace Transform properties’
on QM+.

-3-
1.2 Typical Laplace Transforms

There are some commonly used Laplace Transform pairs. Remember them:
Signal x(t ) X (s) Converge

Unit impulse function

1 -
 (t )

Unit step function 1

Re( s)  0
us (t ) or 1, t  0 s

Ramp function 1
Re( s)  0
t  us (t ) or t , t  0 s2

Exponential decay function 1

Re( s)  
 t
e us (t ) or e  t
,t 0 s 

Sine function 
Re( s)  0
sin(t )  us (t ) or sin(t ), t  0 s  2
2

Cosine function s
Re( s)  0
cos(t )  us (t ) or cos(t ), t  0 s  2
2

Further reading: More Laplace Transform pairs and their proofs:

https://en.wikipedia.org/wiki/List_of_Laplace_transforms
https://coertvonk.com/physics/lfz-transforms/laplace/laplace-transforms-14929

-4-
MATLAB command:
syms t syms s
y=exp(-t)*(2+cos(2*t+pi/4)) Y=1/(2*s+1)
laplace(y) ilaplace(Y)
Quiz 1.1: Solve the differential equations (assuming 0 initial conditions):
dx(t )
(a) 3  x(t )   (t ) ;
dt
d 2 x(t )
(b)  5 x(t )  us (t ) ;
dt 2
dx(t )
(c)  2 x(t )  et us (t )
dt

dx(t )
(a) 3  x(t )   (t )  3sX ( s)  X ( s)  1
dt
1 1 1 1 1t
Then, X ( s )    x(t )  e 3 , t  0
3s  1 3 s  1 3 3

d 2 x(t ) 1
(b) 2
 x(t )  us (t )  s 2 X ( s)  X ( s ) 
dt s
1 1 1 s
Then, X ( s)     x(t )   cos(t )  us (t ), t  0
s s2  1 s s2  1
dx(t ) 1
(c)  2 x(t )  et us (t )  sX ( s)  2 X ( s) 
dt s 1
1 1 1 1
Yields X ( s )     x(t )  e2t  e t , t  0
( s  2) ( s  1) s  2 s 1
Inverse Laplace Transform and partial fractions (Typical cases)
1
X (s) 
s 3 ( s  1)( s  2)( s  3) 2 ( s 2  4 s  5)
A B C D E F G Hs  I
  2 3     2
s s s s  1 s  2 s  3 ( s  3) 2
s  4s  5

-5-
Practice:
1 A B 0.5 0.5
X (s)  X (s)    
( s  1)( s  3) s 1 s  3 s 1 s  3

A B C
X (s)   
1 s  2 ( s  2) 2
s 1
X ( s) 
( s  2) 2 ( s  1) 1 1 1
  
s  2 ( s  2) 2
s 1

As  B C
X (s)  
X ( s) 
1 s  2s  2 s  1
2

( s 2  2s  2)( s  1) s 1 1
 2 
s  2s  2 s  1

Once you have all these partial fractions, look up the Laplace Transform table
to find the inverse results (time domain).

Further reading: Find Inverse Laplace Transform by using partial fractions:

https://ocw.mit.edu/courses/18-03sc-differential-equations-fall-
2011/pages/unit-iii-fourier-series-and-laplace-transform/partial-fractions-and-
inverse-laplace-transform/

2 Transfer functions
2.1 The concept

We have learned how to model a system using differential equations and how
to solve these differential equations using Laplace Transform. Consider an
example:
Recall the electrical circuit containing a resistor (R) and capacitor (C), the
model of supply voltage (Vi) and that measured over the capacitor (VC) is

-6-
 dVC
RC  VC  Vi
dt

This has been discussed in Week 1.

We now take the Laplace transform of both sides of this equation (assuming 0
initial conditions), giving
RCs  VC ( s )  VC ( s )  Vi ( s )
 (1  RCs )VC ( s )  Vi ( s)
1
 VC ( s )  Vi ( s )
1  RCs
1
It can be seen that, the part is not relevant to the input Vi , it is only
1  RCs
dependent to R and C , which are often fixed/unchanged if a circuit is desined.

This is what we called the Transfer Function: The rule is that the output equals
the input times what is in the block (Assuming 0 initial conditons), which we
call the TRANSFER FUNCTION.
In this case, the transfer function appears to be an “optional extra”, but it
proves to be VERY useful when we need to analyse a system of such blocks.
Such a requirement often arises in control system analysis where blocks often
follow each other “in cascade” and feedback is frequently employed.
 The idea of the transfer function means that it is straightforward to
determine the output from this circuit for ANY input whose Laplace
transform we can look up or otherwise determine.

-7-
In many control systems, this transfer function approach in the Laplace domain
is VERY POWERFUL since it allows us to readily understand the effects of
certain structures (feedback or feed forward loops). It also allows us, in many
situations, to simplify the structure of our control system (i.e. the arrangement
of the blocks) through a process called Block Diagram Reduction (We will
discuss this later in this lecture).
Quiz 2.1: What are the Transfer functions of the following systems?
Model:
dCA (t )
s  (1  kr s )CA (t )  CAi (t )
dt
 kr is the reaction rate constant
dependent on temperature,
Periodically operated Continuous   is the residence time dependent on
s
Stirred Tank Reactor (CSTR) the size of the CSTR
1
 s s  CA ( s)  (1  kr s )CA ( s)  CAi ( s)  CA ( s)  C ( s)
 s s  1  kr s Ai
1
Thus the transfer function is G ( s ) 
 s s  1  kr s
This is a First order Linear Time invariant system

Model:
d 2 x(t ) 3EI
m  3 x(t )  u (t )
dt 2 l
 I is the moment of inertia of the
beam’s cross section,
 E is the Young's modulus,
 m is the mass.
Building structure system
3EI 1
ms 2  X ( s )  X ( s )  U ( s )  X ( s )  U (s)
l3 ms 2  3EI l 3
1
Thus the transfer function is G ( s ) 
ms  3EI l 3
2

This is a Second order Linear Time invariant system

-8-
2.2 Zeros and Poles

i.e.

s2  zero s2

G( s)  
( s  1)( s  3) poles s  1, s  3
We can show zeros and poles by using a pole-zero diagram:

The pole-zero diagram is a plot on s-plane

which represents the locations of poles and
zeros of a Transfer function.
s-plane: s is a complex number represented
by s    j (We often use j or i to
represent the image part)
 Poles on s-plane is represented by X
 Zeros on s-plane is represented by O
Zeros and poles are often used to assess the stability of a system.
Quiz 2.2: What are the zeros and poles of the following systems?

s2  4
G (s) 
s2  4 s ( s  1)( s  3)
Y ( s)  U ( s)
s ( s 2  4s  3)  zero s  2i

poles s  0, s  1, s  3
s 1 s 1
G ( s)  
s  2s  1 ( s  1) 2
2

y (t )  2 y(t )  y(t )  u (t )  u (t )
 zero s  1

poles s  1 (2 repeated roots)

-9-
2.2 General form and standard forms

A transfer function can be written into a general form:

N ( s ) b0 s m  b1s m1   bm1s  bm

G(s)  
D( s ) a0 s n  a1s n1   an1s  an

where a0 , , an and b0 , , bm are real numbers.

 We consider m  n because in this case, the system is physically realizable

(or Strictly proper).
For example: the transfer function of
du (t )
y (t ) 
dt
is G( s)  s , which is improper.

This is because, consider u (t )  sin(t ) , then y(t )   cos(t ) . When    ,

no physical systems can reach such a high gain!
For the convenience of conducting various analyses, we often rewrite this
general form into two standard forms:
Pole-Zero Form (Head-1) and Time Constant Form (Tail-1)
(a) Pole-Zero Form (Head-1):
( s  z1 )( s  z2 ) ( s  zm )
G(s)  K
( s  p1 )( s  p2 ) ( s  pn )

where z1 , z2 , , zm are zeros, p1 , p2 , , pn are poles, K is the gain for root

locus (Will discuss this later in Root Locus – Week 3)
(b) Time Constant Form (Tail-1):
m1 m2

 ( k s  1) ( l 2 s 2  2 l s  1)
G (s)  K k 1
n1
l 1
n2

 (T s  1) (T
i 1
i
j 1
j s  2 T j s  1)
2 2

where  k , l , Ti , T j and  are real numbers, K is the gain for transfer function
(Will discuss this later in Bode Plot – Week 6)
- 10 -
The tail-1 form looks complex, but it aims to normalize the constant term in
both numerator and denominator.

3 Block diagrams
3.1 Concepts

Having determined model transfer functions for each of the blocks in a system,
we are often faced with the task of obtaining the overall transfer function for
a system consisting of several blocks. The basic elements of a block diagram
are a block, the summing point and the take-off point.
Let us consider the block diagram of a closed loop (Feedback) control system as
shown in the following figure to identify these elements.

The above block diagram consists of two blocks having transfer functions G(s)
and H(s). It is also having one summing point and one take-off point. Arrows
indicate the direction of the flow of signals.
Let us now discuss these elements one by one.
(1) Block
The transfer function of a component is represented by a block. Block has
single input and single output. The following figure shows a block having input
U(s), output Y(s) and the transfer function G(s).

Output of the block is obtained by multiplying transfer function of the block

with input.

- 11 -
(2) Summing point
The summing point is represented with a circle having cross (X) inside it. It has
two or more inputs and single output.
It produces the algebraic sum of the inputs. It also performs the summation or
subtraction or combination of summation and subtraction of the inputs based
on the polarity of the inputs. Let us see these three operations one by one.
The figure shows the summing point with two inputs (A,
B) and one output (Y). Here, the inputs A and B have a
positive sign. So, the summing point produces the
output, Y as sum of A and B.

The figure shows the summing point with two inputs (A,
B) and one output (Y). Here, the inputs A and B are
having opposite signs, i.e., A is having positive sign and B
is having negative sign. So, the summing point produces
the output Y as the difference of A and B.
The figure shows the summing point with three inputs
(A, B, C) and one output (Y). Here, the inputs A and B are
having positive signs and C is having a negative sign. So,
the summing point produces the output Y as
Y = A + B + (−C) = A + B − C

(3) Take-off point

The take-off point is a point from which the same input signal can be passed
through more than one branch. That means with the help of take-off point, we
can apply the same input to one or more blocks, summing points.

- 12 -
3.2 Block diagram connections

1. Cascade
Cascade connection is also called Series connection. In the following figure,
two blocks having transfer functions G1 ( s) and G2 ( s) are connected in series.

We can represent the cascade connection of two blocks with a single block.
The transfer function of this single block is the product of the transfer
functions of those two blocks.
Similarly, we can represent cascade connection of ‘n’ blocks with a single block.
The transfer function of this single block is the product of the transfer
functions of all those ‘n’ blocks.
2. Parallel
The blocks which are connected in parallel will have the same input. In the
following figure, two blocks having transfer functions G1 ( s) and G2 ( s) are
connected in parallel. The outputs of these two blocks are connected to the
summing point.

- 13 -
We can represent the parallel connection of two blocks with a single block.
The transfer function of this single block is the sum of the transfer
functions of those two blocks.
Similarly, we can represent parallel connection of ‘n’ blocks with a single block.
The transfer function of this single block is the algebraic sum of the transfer
functions of all those ‘n’ blocks.
3. Feedback loop
As we discussed in previous chapters, there are two types of feedback -
positive feedback and negative feedback. The following figure shows negative
feedback control system. Here, two blocks having transfer functions G ( s ) and
H ( s) form a closed loop.

Y ( s)  G( s)[U ( s)  Z ( s)]  G( s)[U ( s)  H ( s)Y ( s)]

G( s)
Y ( s )  G ( s ) H ( s )Y ( s )  G ( s )U ( s )  Y ( s )  U (s)
1  G( s) H (s)
We can represent the feedback connection of two blocks with a single block.
G( s)
The negative feedback closed loop transfer function is
1  G ( s) H ( s)

- 14 -
Similarly, you can represent the positive feedback connection of two blocks
with a single block. The transfer function of this single block is the closed loop
G( s)
transfer function of the positive feedback is
1  G ( s) H ( s)
Quiz 2.3: Evaluate the transfer funtion of the following system

2 2
s  1 1 s  1 1 2s 1 3s 2  s  2
G ( s)       
2 1 s s 2  s  2 s s 2  s  2 s s 3  s 2  2s
1
s 1 s s2  s
3.2 Block diagram algebra
As well as these rules for combining blocks, we can also move some junctions
or components to get parts of the block diagram into the forms needed by
these rules.
Rather than learning lots of these moves, it is better to get into the habit of
just checking the outputs from the group of blocks you are considering - if you
have got the compensating blocks right then the “before” and “after” outputs
will match up with each other.
1. Summing point shifting

- 15 -
2. Take-off point shifting

Quiz 2.4: Evaluate the transfer funtion of the following system

Solution:

- 16 -
- 17 -