Department of Electronic Engineering

Block Diagram
g and
Signal-Flow Graph Models

Control Systems Lecture 2

Song Wang
Song.Wang@latrobe.edu.au
1
ELE 3SPC / ELE5FPC– Signal Processing and Control
Outline
™ Block diagram representation
™ Block diagram
g reduction
™ Signal-flow graph
™ Mason’s ggain formula
™ Examples of block diagram & signal-flow
diagraph
g p methods
™ Introduction to Matlab functions used for system
modeling
2
ELE 3SPC / ELE5FPC– Signal Processing and Control
Block diagram representation
¾ The dynamic systems that comprise automatic control systems
are represented by a set of simultaneous differential equations.
¾ The Laplace transform converts linear differential equations into
algebraic equations in terms of variable s.
¾ The transfer function of a linear system is defined as the ratio of
the Laplace transform of the output variable to the Laplace
transform of the input variable, assuming zero initial conditions.

In control systems, controlled variables are related to

controlling variables by transfer function,
function which depicts the
input-output relationship. In order to show this input-output
relationship by graphical means, we commonly use block
diagram representation.
representation
3
ELE 3SPC / ELE5FPC– Signal Processing and Control
Block diagrams consist of unidirectional,
nidirectional operational blocks
that represent the transfer function of the variables of interest.

For example,
F l ffor a fi
field-controlled
ld t ll d DC motor
t andd load,
l d the
th
relationship between the displacement θ(s) and the input
voltage Vf(s) is shown by

In co
contrast
s too a pu
purely
e y abstract,
bs c , mathematical
e c representation,
ep ese o ,a
block diagram indicates more realistically the signal flows of
the actual system.

4
ELE 3SPC / ELE5FPC– Signal Processing and Control
A system may have several variables under control.
control For
example, a 2-input 2-output system is described by
Y1 ( s ) = G11 ( s ) R1 ( s ) + G12 ( s ) R2 ( s )
Y2 ( s ) = G21 ( s ) R1 ( s ) + G22 ( s ) R2 ( s )

We write this in matrix form

⎡Y1 ( s ) ⎤ ⎡G11 G12 ⎤ ⎡ R1 ( s ) ⎤
⎢Y ( s )⎥ = ⎢G ⎥⎢ ⎥
⎣ 2 ⎦ ⎣ 21 G22 ⎦ ⎣ R2 ( s )⎦

The block diagram

representation is

5
ELE 3SPC / ELE5FPC– Signal Processing and Control
Derivation of the closed-loop transfer function (CLTF)

When H(s)=1,
() it
is called unity
feedback.

When
G(s)H(s)>>1,
() ()
C(s)/R(s)=1/H(s)
6
ELE 3SPC / ELE5FPC– Signal Processing and Control
Consider again the following closed-loop system.

What is the transfer function E(s)/R(s)?

Closed-loop systems are used as they allow the system

performance to be dominated by the components in the
feedback path. This is an important consideration as some
components in the forward path are outside designers’ control.

7
ELE 3SPC / ELE5FPC– Signal Processing and Control
Block diagram reduction
• Cascading blocks:

• Moving a summing
point after a block
• Moving a pickoff point
ahead
h d off a bl
blockk
• Moving a pickoff
point after a block
• Moving a summing
point ahead of a block
• Feedback loop elimination
8
ELE 3SPC / ELE5FPC– Signal Processing and Control
Example – a multiple-loop feedback control system

Step 1. Move the pickoff point after G4

9
ELE 3SPC / ELE5FPC– Signal Processing and Control
Step 2. Eliminate loop G3G4H1

Loop L

Step 3. Eliminate loop L

Loopp M
Step 4. Eliminate loop M

10
ELE 3SPC / ELE5FPC– Signal Processing and Control
Signal-flow graph
™ Alternative method to block diagram representation,
developed by S.J.Mason.
™ Ad
Advantage: the
h availability
il bili off a flow
fl graphh gain
i formula,
f l
also called Mason’s gain formula.
™ A signal-flow
signal flow graph consists of a network in which nodes
are connected by directed branches.
™ It depicts the flow of signals from one point of a system to
another
h andd gives
i the
h relationships
l i hi among the h signals.
i l
Note that the signal flows in only one direction.

11
ELE 3SPC / ELE5FPC– Signal Processing and Control
Definitions
™ Node - a point representing a signal or variable.
™ Branch – unidirectional line segment joining two nodes.
™ Path – a branch or a continuous sequence of branches that
can be traversed from one node to another node.
™ Loop – a closed path that originates and terminates on the
same node and along the path no node is met twice.
™ Nontouching loops – two loops are said to be nontouching
if they do not have a common node.

12
ELE 3SPC / ELE5FPC– Signal Processing and Control
 Corresponding
Signal flow graph of
Signal-flow block diagram
interconnected system

The signal
g flow ggraphp is a pictorial
p method of writingg a set of
algebraic equations to indicate the interdependencies of variables.
For example,
a11 x1 + a12 x2 + r1 = x1 Signal-flow
Si l fl graphh
of two algebraic
a 21 x1 + a 22 x2 + r2 = x2
equations

13
ELE 3SPC / ELE5FPC– Signal Processing and Control
 Mason’s gain formula
The linear
Th li dependence
d d Tij between
b the
h independent
i d d variable
i bl xi
(also called the input variable) and a dependent variable xj is

14
ELE 3SPC / ELE5FPC– Signal Processing and Control
 ∑P Δk k

Write the gain formula in a simplified form: T= k

Δ

Systematic approach:
1. Calculate forward path transfer function Pk for each
forward path k
2. Calculate all loop transfer functions
3
3. C id nontouching
Consider t hi loopsl 2 att a time
ti
4. Consider nontouching loops 3 at a time
5
5. etc
6. Calculate Δ from steps 2,3,4 and 5
7. Calculate Δk as pportion of Δ not touchingg forward ppath k

15
ELE 3SPC / ELE5FPC– Signal Processing and Control
Consider a two-path signal-flow graph.
1. Calculate
C l l forward
f d pathh transfer
f function
f i Pk
for each forward path k.
P1 = G 1 G 2 G 3 G 4 (path 1) and P2 = G 5 G 6 G 7 G 8 (path 2)

2. Calculate all loop TF’s.

L1 = G2 H 2 , L2 = H 3G3 , L3 = G6 H 6 , L4 = G7 H 7
3. Consider nontouchingg loops
p 2 at a time.
Loops L1 and L2 do not touch Loops L3 and
L4.
4
4. Consider nontouching loops 3 at a time
time.
None.
5. Calculate Δ from steps 2,3,4.
Δ = 1 − (L1 + L2 + L3 + L4 ) + (L1 L3 + L1 L4 + L2 L3 + L2 L4 )
6. Calculate Δk as portion of Δ not touching
forward path k. Δ1 = 1 − (L3 + L4 ) and Δ 2 = 1 − (L1 + L2 )
The TF of the system is Y ( s ) P1Δ1 + P2 Δ 2
= T ( s) =
R( s) Δ 16
ELE 3SPC / ELE5FPC– Signal Processing and Control
 Examples of block diagram &
signal-flow graph methods
Consider the block diagram and signal-flow
signal flow models of an
armature-controlled DC motor.
There is one forward
path P1, which touches
the one loop L1.
1
P1 = G1 ( s )G2 ( s )
s
L1 = − K bG1 ( s )G2 ( s )

Δ = 1 − L1 , and Δ1 = 1

θ ( s) P1Δ1 P
= = 1
Va ( s ) Δ 1 − L1
(1 / s )G1 ( s )G2 ( s )
Using Mason’s gain formula, we obtain the TF =
1 + K b G1 ( s )G2 ( s )
17
ELE 3SPC / ELE5FPC– Signal Processing and Control
 Reconsider the multiple-loop feedback control system
Loop 2

Loop 3
L
Loop 1
Step 1. Find forward paths. P1 = G1G2 G3G4
Step 2.
2 Calculate loop gains.gains L = −G G G G H , L = −G G H , L = G G H
1 1 2 3 4 3 2 2 3 2 3 3 4 1

Step 3. Find nontouching loops 2 at a time. All 3 loops touch.

Step 4&5. Find nontouching loops 3 at a time. Ditto Step 3.
Step 6. Calculate Δ. Δ = 1 − (L1 + L2 + L3 )
Step 7. Calculate Δk. All 3 loops touch P1. Δ1 = 1
Applying Mason’s gain Y ( s) P1Δ1 G1G2G3G4
= =
formula yields same result R( s) Δ 1 + G1G2G3G4 H 3 + G2G3 H 2 − 18G3G4 H1
as before. ELE 3SPC / ELE5FPC– Signal Processing and Control
Block diagram
g & signal-flow
g graph
g p
summary
™ Both models provide the relationship between system
variables.
™ Bl k di
Block diagram simplification
i lifi i for f a complexl system can be b
cumbersome.
™ Mason’ss gain formula is especially useful in handling large
Mason
and complex systems without requiring step-by-step
reduction.
™ Mason’s’ gain
i formula
f l can be
b applied
li d systematically
i ll but
b
errors can be made and not be easily detected.

19
ELE 3SPC / ELE5FPC– Signal Processing and Control
Matlab provides a wide range of functions to aid
system modeling and system analysis.
™ To manipulate
T i l t polynomials,
l i l we have
h roots poly
roots, poly, conv
conv,
polyval.

20
ELE 3SPC / ELE5FPC– Signal Processing and Control
™ To compute
comp te TF and poles and zeros, se tf,
eros wee use tf pole
pole, zero
zero,
pzmap, minreal. Beware of multiple poles & zeros.

6s 2 + 1
G( s) = 3
s + 3s 2 + 3s + 1

21
ELE 3SPC / ELE5FPC– Signal Processing and Control
™ To perform block diagram transformations,
transformations wee have
ha e
series, parallel, feedback.

22
ELE 3SPC / ELE5FPC– Signal Processing and Control
Example

23
ELE 3SPC / ELE5FPC– Signal Processing and Control