EE 312/ Lecture no.

2/system modelling/ Signal flow graph approach

University of Benghazi
Faculty of Engineering
Electrical and Electronic Engineering Department
----------------------------------------------------------------------------------------------------------------------------
Control systems EE312 Fall 2011/2012

Course Format and Assessment

The course will be delivered in 56 hours, with 4 home works. Final exam will be in Feb. 2012.
Course material will include handouts.
----------------------------------------------------------------------------------------------------------------------------
Instructor:
Dr. Awad Shamekh
----------------------------------------------------------------------------------------------------------------------------

Lecture no. (2) Date: 4/12/2011

Learning Outcomes of this lecture

The student will understand:
• The signal flow graph.
• The concepts of characteristic equation, poles, and zeroes.

----------------------------------------------------------------------------------------------------------------------------
References:
- Modern control systems, R.C. Dorf
- Control systems Engineering, Norman Nise
- Linear control systems engineering, Morris Driels
----------------------------------------------------------------------------------------------------------------------------

Signal flow graph

Is another method of representing a linear system it is a graphical representation of the
system. This method was developed by Mason. It consists of many nodes with
interconnected directed branches, see fig (1) & (2).

Fig (1), an illustrative example of the signal flow method

1
 EE 312/ Lecture no. 2/system modelling/ Signal flow graph approach

Node a node is a point which is connected by branches. Each node represents a

system variable.
Incoming signals are added up and the resultant signal is transmitted along the out
going branches.
Source node will have only outgoing branches.
Sink node will have only incoming branches.
Mixed node will have both outgoing and incoming branches.
Path is the traversed of branches from one node to the other such that no node is
touched more than once.
Forward path is a path originating from a source node and terminating on a sink
node.
Feedback path is a path starting from one node and terminating on the same node.
Branch gain (or branch transmittance) is expressed by the gain between two nodes.
Path gain is the product of the branch gains forming the path.
Loop gain is the product of the branch gains forming the loop.

Fig (2) a schematic diagram of the signal flow modeling.

Example:
Draw the signal flow graph for the circuit shown in fig (3), where vi is the input and
v o is the output.

Solution
v − v2 v − vo
i1 = i i2 = 2
R1 R3
v 2 = (i1 − i2 )R2 v o = i 2 R4
Take Laplace transformation for the system variables
Li1 = I 1 ( s ) Li 2 = I 2 ( s ) Lvi = Vi (s ) Lv 2 = V2 ( s ) Lv o = Vo (s )

2
 EE 312/ Lecture no. 2/system modelling/ Signal flow graph approach

Fig (3)

Vi ( s ) V2 ( s )
I1 (s) = −
R1 R1
V 2 ( s ) = ( I 1 ( s ) R 2 − I 2 ( s ) R2 )
V 2 ( s ) Vo ( s )
I 2 (s) = −
R3 R3
Vo ( s ) = I 2 ( s ) R 4

From these variables, the signal flow graph can be constructed as in fig (4)

Fig (4), signal flow graph of the circuit shown above

3
 EE 312/ Lecture no. 2/system modelling/ Signal flow graph approach

Example
Represent the following simultaneous equations by
1. A block diagram method
2. Signal flow graph

1
X 1 ( s ) = R1 ( s ) + X 2 ( s)
s +1
1
X 2 ( s ) = −4 X 1 ( s ) + R2 ( s ) − 7 Y1 ( s) − X 2 (s)
s+4
s
Y1 ( s ) = X 2 (s)
s +3
2

Y2 ( s ) = 10 X 1 ( s ) − sY1 ( s )

Solution
1) The block diagram method

2) Signal flow graph

4
 EE 312/ Lecture no. 2/system modelling/ Signal flow graph approach

Mason's Gain Rule

Mason's gain rule is a formula for finding the transfer function of a single – input ,
single output system from its signal flow graph. It may be applied repeatedly to a
multiple-input, multiple-output system to obtain each of the transfer functions.
Transfer function is also called graph transmittance.
n

C (s) ∑P Δ k k
= k =1
(graph transmittance)
R( s) Δ
When n is the number of the forward paths.
Pk is the path gain of the K th forward path.
Δ is the determinant which can be calculated as:

The cofactor of a path is the determinant of the signal flow graph formed deleting all
the loops touching the path.
Δ k = 1 − ( sum of the loop gains for the loops nontouching with the K th forward path) +
( sum of products of the loop gains taken two at a time for nontouching loops
which are nontouching with K th forward path ) − (........................)
To illustrate the application of Mason's gain rule, consider the example of the signal
flow graph shown in the figure below.

From the definition, a path is any succession of branches from input to output in the
direction of arrows, which does not pass any node more than once. From the figure,
one can observe that the system has two paths. The path gain is the product of the
transmittance of the branches comprising the path.

5
 EE 312/ Lecture no. 2/system modelling/ Signal flow graph approach

1 1 1 1
P1 = ( )( 2 )(10)( )( )
s +1 s + s s s
1 8 1 1
P2 = ( )( )( )( )
s + 4 s +8 s s
2

A loop is any closed succession of branches in the direction of arrows, which does not
pass any node more than once. The loop gain is the product of the transmittances of
the branches comprising the loop.
Then

4 − 56
L1 = − L3 =
s +s2
s +8
−6
L2 = − s L4 =
s
1 1 3 s
L5 = (10)( )( )( )( )
s s s+3 s+2
Two loops are said to be touching if they have any node in common otherwise, they
are non-touching. Similarly, a loop and path are touching if they have any node in
common.
Δ = 1 − ( L1 + L2 + L3 + L4 + L5 ) + ( L1 L2 + L1 L3 + L1 L4 + L2 L3 + L2 L4 ) − ( L1 L2 L3 + L1 L2 L4 )
Δ 1 = 1 − L2 = 1 + s
4 4s
Δ 2 = 1 − ( L1 + L2 ) + ( L1 L2 ) = 1 + +s+ 2
s +s
2
s +s
P1 Δ1 + P2 Δ 2
T (s) = #
Δ
Example

6
 EE 312/ Lecture no. 2/system modelling/ Signal flow graph approach

Find the transfer function of the signal flow graph shown in the figure below using the
Mason's gain rule.

Solution
P1 = a b g h = (1)(6)(1) = 6
1 −4
P2 = a b c d g h = (1)(1)( )(−4)(1) = ( )
s +1 s +1
1 s 3 s
P3 = a b c d e f g h = (1)(1)( )(1)( )(3)(1) = ( )( )
s +1 s +1 s +1 s + 2
1 −3
L1 = (−3)( )=
s +1 s +1
s − 5s
L2 = (−5)( )=
s+2 s+2
3 5s 15s
Δ = 1 − ( L1 + L2 ) + ( L1 L2 ) = 1 + + +
s + 1 s + 2 ( s + 1) + ( s + 2)
3 5s 15s
Δ1 = 1 + + +
s + 1 s + 2 ( s + 1) + ( s + 2)
5s
Δ2 = 1+
s+2
Δ3 = 1
P1 Δ1 + P2 Δ 2 + P3 Δ 3
T (s) =
Δ
where
3 5s 15s
P1 Δ 1 = 6[1 + + + ]
s + 1 s + 2 ( s + 1)( s + 2)
⎛ − 4 ⎞⎛ 5s ⎞
P2 Δ 2 = ⎜ ⎟⎜1 + ⎟
⎝ s + 1 ⎠⎝ s + 2 ⎠
⎛ 3 ⎞⎛ 5 ⎞
P3 Δ 3 = ⎜ ⎟⎜ ⎟
⎝ s + 1 ⎠⎝ s + 2 ⎠
Example

7
 EE 312/ Lecture no. 2/system modelling/ Signal flow graph approach

A) Find the transfer function of the block diagram shown in the figure below.
Solution
C = ( RG1 + RG3 − CH 1 )G2
C = RG1G2 + RG2 G3 − CH 1G2
C (1 + H 1G2 ) = RG1G2 + RG2 G3
C G1G2 + G2 G3
= #
R 1 + H 1G 2

B) convert the previous system to a signal flow graph and find the transfer function
use Mason's gain rule.
Solution

Paths:
P1 = a b c d e = G1 G2
P2 = a b c d e = G3 G2
Loops:
L1 = c − d − c = − G2 H 1
Δ = 1 − ( −G 2 H 1 ) = 1 + G 2 H 1
Δ1 = 1
Δ2 = 1
P1 Δ1 + P2 Δ 2 G1G2 + G3G2
T (s) = =
Δ 1 + G2 H 1