Lect 2 - Modeling of Dynamic Systems

Modeling of Dynamic Systems
Ref: Control System Engineering

Norman Nise : Chapters 2 & 3
TM184521 – Modelling of Dynamic Systems
Review of the last lecture
1
Review of the last lecture

◼ Open-loop control: The controller does not use the
output signal y(t)
◼ Closed-loop control: The controller uses the output
signal y(t)
◼ Feedback: another word for closed-loop
◼ Model: a mathematical description (differential
equation) of a system
◼ Better performance can be achieved by closed-loop
control
Chapter objectives :
◼ Review the Laplace transform
◼ Learn how to find a mathematical model, called a transfer
function
◼ Learn how to find a mathematical model, called a state
variable representation
◼ How to convert between transfer function and state space
models
◼ How to linearize a non-linear system
2
Laplace in Frequency Domain
Laplace Transforms Review

1. Standard notation in dynamics and control (shorthand
notation)
2. Converts mathematics to algebraic operations
3. Advantageous for block diagram analysis
Laplace transforms can be used in process control for:

1. Solution of differential equations (linear)
2. Analysis of linear control systems (frequency response)
3. Prediction of transient response for different inputs
4. Control system design
5. Stability analysis
5
Definition
The Laplace transform of a function f(t) is defined as

F ( s ) = L  f (t )  =  f ( t ) e − st dt
0
where F(s) is the symbol for the Laplace transform, L is the

Laplace transform operator, and f(t) is some function of time, t.
Note: The L operator transforms a time domain function f(t) into an

s domain function, F(s). s is a complex variable:
s = a + bj, j = −1
3
Inverse Laplace Transform, L-1:

By definition, the inverse Laplace transform operator, L-1, converts an
s-domain function back to the corresponding time domain function:
f ( t ) = L−1  F ( s ) 
Important Properties:
Both L and L-1 are linear operators. Thus,
L  ax ( t ) + by ( t )  = aL  x ( t )  + bL  y ( t ) 
= aX ( s ) + bY ( s )
where:
x(t) and y(t) are arbitrary functions
a and b are constants
X ( s ) = L  x ( t ) and Y ( s ) = L  y ( t )
Similarly,
L−1  aX ( s ) + bY ( s )  = ax ( t ) + b y ( t )
4
Laplace Transforms of Common Functions

1. Constant Function
Let f(t) = a (a constant). Then from the definition of the Laplace
transform,

 − st a − st  a a
L ( a ) =  ae dt = − e = 0−−  =
0 s  s s
0
2. Step Function
0 for t  0
It is defined as: u (t ) = 
1 for t  0
Because the step function is a special case of a “constant”, it
follows from u(t) that
1
L u ( t )  =
s 9
3. Derivatives
This is a very important transform because derivatives appear in
the ODEs we wish to solve.
 df 
L   = sF ( s ) − f ( 0 )
 dt 
initial condition at t = 0
Similarly, for higher order derivatives:

dn f 
L  n  = s n F ( s ) − s n−1 f ( 0 ) − s n−2 f ( ) ( 0 ) −
1
 dt 
−... − sf ( ) 0 − f ( ) 0
n−2 n −1
( ) ( )
where: n is an arbitrary positive integer
10
10
5
Special Case: All Initial Conditions are Zeroes
dk f
f ( ) ( 0) = k
k
Suppose then
dt t =0
f ( 0 ) = f ( ) ( 0 ) = ... = f ( ) ( 0 ) .
1 n−1
dn f  n
L n  = s F (s)
 dt 
In process control problems, we usually assume zero initial

conditions or some initial steady-state.
11
11
4. Exponential Functions
Consider f ( t ) = e−bt where b > 0. Then,
  − b+ s t
L e−bt  =  e−bt e− st dt =  e ( ) dt
  0 0
1  −( b+ s )t   1
= −e =
b+s  0 s+b
5. Rectangular Pulse Function

It is defined by:
0 for t  0

f ( t ) = h for 0  t  t w
0 for t  t
 w
12
12
6
f (t )
The Laplace transform of the rectangular
h pulse is given by
F (s) =
h
s
(
1 − e −t w s )
t w Time, t
6. Impulse Function (or Dirac Delta Function)
The impulse function is obtained by taking the limit of the
rectangular pulse as its width tw going to zero, but holding the area
1
under the pulse constant at one. (i.e., let h = )
tw
 ( t ) = impulse function
Then, L  ( t )  = 1
13
13
Solution of ODEs by Laplace Transforms

Procedure:
1. Take the L of both sides of the ODE.
2. Rearrange the resulting algebraic equation in the s domain to
solve for the L of the output variable, e.g., Y(s).
3. Perform a partial fraction expansion.
4. Use the L-1 to find y(t) from the expression for Y(s).
14
14
7
Table Laplace Transforms for Various Time-Domain Functions

f(t) F(s)
15
15
f(t) F(s)
16
16
8
f(t) F(s)
17
17
Important Properties of Laplace Transforms

1. Final Value Theorem
It can be used to find the steady-state value of a closed loop
system (providing that a steady-state value exists).
Statement of FVT:
lim y ( t ) = lim  sY ( s )

t → s →0
providing that the limit exists (is finite) for all Re ( s )  0,

where Re (s) denotes the real part of complex variable, s.
18
18
9
Example: 5s + 2
Y (s) =
s ( 5s + 4 )
Suppose,
Then,
 5s + 2 
y (  ) = lim y ( t ) = lim   = 0.5
t → s →0  5s + 4 
2. Time Delay
Time delays occur due to fluid flow, time required to do an
analysis (e.g. gas chromatograph). The delayed signal can
be represented as
y ( t − θ ) ; θ = time delay
Also,
L  y ( t − θ )  = e−θsY ( s ) 19
19
Example 1
Solve the ODE,
dy
5 + 4y = 2 y (0) = 1 (3-26)
dt
First, take L of both sides 5 ( sY ( s ) − 1) + 4Y ( s ) =

2
s
5s + 2
Rearrange, Y (s) = (3-34)
s ( 5s + 4 )
 5s + 2 
Take L-1, y ( t ) = L−1  
 s ( 5s + 4 ) 
From LT Table, y ( t ) = 0.5 + 0.5e−0.8t (3-37)

20
20
10
Example 2
21
21
Partial fraction expansion

Basic idea: Expand a complex expression for Y(s) into
simpler terms, each of which appears in the Laplace
Transform table. Then you can take the L-1 of both sides of
the equation to obtain y(t).
1. Case 1. Roots of the denominator of F(s) are real and

distinct
2. Case 2. Roots of the denominator of F(s) are real and
repeated
3. Case 3. Roots of the denominator of F(s) are complex or
imaginary
22
22
11
s+5
Example 3 Y (s) =
( s + 1)( s + 4 )
Perform a partial fraction expansion (PFE)
s+5  
= 1 + 2
( s + 1)( s + 4 ) s + 1 s + 4
where coefficients 1 and  2 have to be determined.
To find 1 : Multiply both sides by s + 1 and let s = -1
s+5 4
1 = =
s+4 s =−1 3
Laplace - 23
23
To find  2 : Multiply both sides by s + 4 and let s = -4
s+5 1
2 = =−
s +1 s =−4 3
A General PFE
N (s) N (s)
Consider a general transfer function, Y ( s ) = =
D(s) n
 ( s + bi )
i =1
Here D(s) is an n-th order polynomial with the roots ( s = −bi ) all
being real and distinct so there are no repeated roots. N(s) is order
m with n > m
24
24
12
The PFE is:

N (s) n
i
Y (s) = =
( s + b1 )( s + b2 ) ...( s + bn ) i =1 s + bi
Note: D(s) is called the “characteristic polynomial”.
Special Situations:
Two other types of situations commonly occur when D(s) has:
i) Complex roots: e.g., bi = 3  4 j j = −1( )
ii) Repeated roots : e.g., b1 = b2 = −3
For these situations, the PFE has a different form.
25
25
Example 4
d3y d2y dy
3
+ 6 2
+ 11 + 6 y = 4
dt dt dt
y( 0 )= y( 0 )= y ( 0 )= 0 system at rest (s.s.)
To find transient response for u(t) = unit step at t > 0
1. Take Laplace Transform (L.T.)

2. Factor, use partial fraction decomposition
3. Take inverse L.T.
Step 1 Take L.T. (note zero initial conditions)
4
s3Y(s)+ 6s 2Y(s)+11sY(s) + 6Y (s ) =
s
26
26
13
4
Rearranging, Y(s)=
( s 3 + 6s 2 + 11s + 6) s
Step 2a. Factor denominator of Y(s)

s(s 3+6s 2+11s+6 )=s(s+1 )(s+2 )(s+3 )
Step 2b. Use partial fraction decomposition

4 α α α α
= 1+ 2 + 3 + 4
s(s +1 )(s + 2 )(s + 3 ) s s + 1 s + 2 s + 3
Multiply by s, set s = 0
4  α α α 
= α1 + s  2 + 3 + 4 
(s +1 )(s + 2 )(s + 3 ) s =0  s + 1 s + 2 s + 3  s =0
4 2
= α1 =
1 2  3 3
27
27
For a2, multiply by (s+1), set s = -1 (same procedure

for a3, a4) 2
α2 = −2, α3 = 2, α4 = −
3
2 2 2 2/3
Step 3. Take inverse of L.T. (Y(s)= − + − )
3s s +1 s + 2 s +3
2 2
y(t)= − 2e −t + 2e −2t − e −3t
3 3
2
t → y(t) → t = 0 y (0) = 0. (check original ODE)
3
You can use this method on any order of ODE, limited only by
factoring of denominator polynomial (characteristic equation)
Practice Matlab …..

28
28
14
Example 5
Example 6
For a system described by the differential equations with initial conditions,
c + 7c + 10c = r , c ( 0 ) = 1, c ( 0 ) = 4
Determine (i) the transfer function C/R, (ii) the transient response to the
initial condition, c(t) and its final output.
29
29
Solution to example no 5
Since input is a step input, we have R(s) = 1/s, therefore
5(s + 100) 5(s + 100)
Y (s ) = =
(
s s + 60s + 500
2
)
s (s + 10)(s + 50)
The partial fraction expansion of Y(s) is given by
Y (s ) = 1 +
K K2 K3
+
s s + 10 s + 50
5(s + 100)
K1 = =1
(s + 10)(s + 50) s =0
5(s + 100)
K2 = = −1.125
s(s + 50) s = −10
5(s + 100)
K3 = = 0.125
s (s + 10) s = −50
30
30
15
and hence Y (s ) = 1 − 1.125 + 0.125

s s + 10 s + 50
Using Laplace transform table, we find that :

y(t ) = 1 − 1.125e−10t + 0.125e−50t
The final value can be computed using the final value theorem :
 5(s + 100) 
lim y (t ) = lim s  2  =1
t → s →0 (
 s s + 60s + 500 )

You can also find final value of y(t) by inputting a large value of
t (or infinite) in the equation
31
31
Modeling the system

What is system?
The original meaning of the term is a combination of elements
intended to act together to accomplish an objective.
For example, a link in a bicycle chain is usually not considered

to be a system. However, when it is used with other links to
form a chain, it becomes part of a system. The objective for the
chain is to transmit force.
When the chain is combined with gears, wheels, crank,
handlebars, and other elements, it becomes part of a larger
system whose purpose is to transport a person.
32
32
16
Input and Output

Nowadays, for example, a factory manager will call a
meeting to seek “input,” meaning opinions or data, from the
employees, and the manager may refer to the products
manufactured in the factory as its “output.”
However, in the system dynamics meaning of the terms, an

input is a cause; an output is an effect due to the input.
Thus, one input to the bicycle is the force applied to the

pedal. One resulting output is the acceleration of the bike.
33
33
The behavior of a system element is speciﬁed by its input-output

relation, which is a description of how the output is affected by the
input.
The input-output relation expresses the cause-and-effect behavior
of the element.
Any causes acting on the system from the world external to this
boundary are considered to be system inputs. Similarly, a system’s
outputs are the outputs from any one or more of the system
elements that act on the world outside the system boundary.
34
34
17
Static and Dynamic Elements

When the present value of an element’s output depends only
on the present value of its input, we say the element is a static
element.
For example, the current ﬂowing through a resistor depends
only on the present value of the applied voltage.
If an element’s present output depends on past inputs, we say it

is a dynamic element.
For example, the present position of a bike depends on what its
velocity has been from the start.
35
35
Examples of dynamic systems
(a) one representation of an automobile, (b) a simplified representation
36
36
18
37
37
Need for Mathematical Model (or Model)

• Must have a quantitative mathematical model in order to
understand and control complex systems.
• Must have fundamental method for modeling many physical
systems:
❑ Mechanical
❑ Electrical
❑ Electromechanical
❑ Hydraulic
❑ Biological
❑ etc
38
38
19
What will the model be used for ?

o Solutions of the differential and algebraic equations allows
system response and performance to be analyzed and
designed.
o The Laplace transformation will be applied to the model to
allow convenient manipulation and dynamic analysis.
o Input-output relationships for systems and components will
be obtained.
o Controller models will be designed that can be
implemented in hardware.
39
39
Types of model
Models of dynamics system can be of many kinds, including
the following:
➢ Mental, intuitive or verbal models
➢ Graphs and tables
➢ Mathematical models
Constructing a model:
➢ Mathematical modeling or first principle modeling.
➢ Process or system identification.
Why does one perform system identification ?
40
40
20
Properties of the models obtained from system identification

➢ limited validity
➢ little physical insight → the parameters are used only as
tools to give a good description of the system's overall
behavior
➢ relatively easy to construct and use
Identification is not a foolproof methodology that can be used

without interaction from the user → reasons :
➢ An appropriate model structure must be found.
➢ There are certainly no “perfect” data in real life.
➢ The process may vary with time
➢ It may be difficult or impossible to measure some variables
signals
41
41
The procedure of deriving models :

• Understand the physics and interaction of the elements
• Construct a simplified diagrammatic rep of the system
• Apply element and interconnection laws
• Draw the FBD
• Identify or define the inputs, outputs and state variables
• Establish the system equations
• Obtain the desired form of the system model
• If the model is non-linear, determine the equilibrium
conditions and obtain a linearized model.
42
42
21
1. Mechanical Systems → translational & rotational
Translational mechanical systems

→ can have only horizontal or vertical motion
a. Variables
Variables of trans mechanical systems are :
➢ x, displacement in meters (m)
➢ v, velocity in meters per second (m/s)
➢ a, acceleration in meter per second square (m/s2)
➢ f, force in Newton (N)
➢ w, energy in Joule (J)
➢ p, power in watts (W)
all variables are functions of time

43
43
b. The element laws include in trans systems are mass, friction and
stiffness.
They relate the external force to the acceleration, velocity and
displacement associated with the element.
Mass
Friction → when two bodies slide over each other there is a frictional
force f between them that is a function of the relative velocity between the
sliding surface
44
44
22
45
45
Stiffness → when a mechanical element is subjected to a force f and goes

through a change in length ∆x, it can be characterized by a stiffness element.
46
46
23
c. Interconnection Laws
D’Alembert’s law → is developed from Newton’s law for translational system.
For a constant mass :
i ( f ext )i = M dt
dv
where the summation over the index i includes all the external forces (fext)i
acting on the body.
 ( f ) − M dt = 0
dv
ext i
i
The sum of the forces is zero provided -Mdv/dt is thought of as an additional

forces → this fictitious force is called inertial force or D’Alembert force.
f
i
i = 0 → D’Alembert’s law
47
47
48
48
24
The law of reaction force → accompanying any force of one element on

another, there is a reaction force on the first element of equal magnitude and
opposite direction (Newton’s third law : reaction forces)
49
49
The law for displacements → if the ends of two elements are connected, those
ends are forced to move with the same displacement and velocity.
50
50
25
d. Obtaining the system model → FBD
Example 7
51
51
Example 8
52
52
26
Example 9
Example 10
53
53
Series and parallel combination
k1k 2 b1b2
k eq = beq =
k1 + k 2 b1 + b2
keq = k1 + k2 beq = b1 + b2
54
54
27
Rotational mechanical systems

→ are modeled using the same techniques as those for
translational mechanical systems
a. Variables
Variables of rotational mechanical systems are :
✓ θ, angular displacement in radians (rad)
✓ ω, angular velocity in radians per second (rad/s)
✓ α, angular acceleration in radians per second squared (rad/s2)
✓ τ, torque in Newton-meters (N.m)
all variables are functions of time
55
55
b. The element laws include in rotational systems are moment of

inertia, friction, stiffness, levers and gears.
Moment of Inertia → J in kilogram-meters2 (kg.m2)
J =  r 2 dm where r is the distance from the axis of reference and

dm is the mass of the small element.
The net torque applied about the fixed axis of rotation is given by
=
d
(J ) where Jω is the angular momentum of a body.
dt
Parallel axis theorem states
J = J 0 + ma 2 where a is the distance between the parallel axes and J0
is the moment of inertia about the principal axis
56
56
28
Friction → A rotational friction element is one for which there is an algebraic

relationship between the torque and the relative angular velocity between two
surfaces.
Rotational viscous friction arises when two rotating bodies are separated by a
 = b = b(2 − 1 )
film of oil.
Rotational devices characterized by viscous friction

57
57
Stiffness → is usually associated with a torsional spring, such as the mainspring

of a clock, or with a relatively thin shaft.
For a linear torsional spring or flexible shaft,
 = k where k is the stiffness constant with units of newton-

meters (N.m) and ∆θ=θ2-θ1
(a) Rot stiffness el with one end fixed. (b) Rot stiffness el with ∆θ=θ2-θ1
Potential energy is stored in a twisted element and for a linear spring or shaft
is given by 1
W= k 2
2
58
58
29
The Lever → an ideal lever is assumed to be a rigid bar pivoted at a point

and having no mass, no friction, no momentum and no stored energy.
Let θ be the angular displacement of the lever from the horizontal position.
The lever
For small displacements

d2 d2 d2
x2 = x1 v2 = v1 f2 = f1
d1 d1 d1
59
59
Gears → an ideal gear is assumed to have no moment of inertia, no friction,

no stored energy, and a perfect meshing of the teeth.
N is a gear ratio
A pair of gears
r2 n2
N= = where r and n denote the radius and number of teeth
r1 n1
θ1 and θ2 are the angular displacements for the
gears
60
60
30
1 r2 1 r2
r11 = r2 2 ; = =N; = =N
 2 r1 2 r1
θ1 and θ2 are the angular
displacements for the gears
Ideal gears (a) ref position, (b) after rotation
 11 +  22 = 0
FBD for a pair of ideal gears

61
61
c. Interconnection Laws
D’Alembert’s law → for a body with constant moment of inertia
rotating about a fixed axis,
 ( ) − J = 0
i
ext i
where the summation over i includes all the torques acting on

the body, and the term –Jω can be considered an inertial
torque.
i
i = 0 → D’Alembert’s law
the torque –Jω is directed opposite to the positive sense of θ, ω

and α
62
62
31
The law of reaction force → for bodies that are rotating about the same axis,
any torque exerted by one element on another is accompanied by a reaction
torque of equal magnitude and opposite direction on the first element.
The law for angular displacements →
Rot sys to illustrate the law of ang disp
The reference marks on the rims are at the top of the two disks when no torque
is applied. The net angular displacement for the shaft K2 with respect to its
unstressed condition is θ2-θ1.
 ( )
i
i =0 around any closed path
63
63
d. Obtaining the system model → FBD
(a) Rotational system, (b) & (c) Its corresponding FBD

64
64
32
Example 11 Example 12
65
65
Computer simulation
Diagram for input-output equations
The steps :
1. Solve the given equation for the highest derivative of the
unknown output variable
2. Connect one or more integrator blocks in series to integrate the
derivatives
3. Use the result of step 1 to form the highest output derivative as
the output of a summer and a gain block
Example :
Mx + Bx + Kx = f a (t )
66
66
33
Model simulation → with ‘simulink’

Example 13
Create a simulation (using simulink) to calculate response of the

system below to an input force that increases from 0-8 N at t=1 s.
The parameter values are M=2 kg, K=16 N/m and B=4 Ns/m
67
67
Example 14
A person wearing a parachute jumps out of airplane. Assume there is no
wind and that the parachute provides viscous damping relative to a fixed
reference frame. The parachute has a relatively small mass Mp and a
large damping coefficient Bp. The jumper has a larger mass Mj and
smaller drag coefficient Bj. The cords attaching the parachute to the
jumper are called risers and are assume to be quite springy and the
elastic effects of the risers is represented by a spring Kr. The
deformation of the parachute itself can also be included in the value of
Kr. Write the modeling equation describing the motions of the jumper
and the parachute and then solve the equation using simulink. Take
positive direction to be downward. Use the following parameter values
Mp=10 kg, Mj=60 kg, Bp=100 Ns/m, Bj=10 Ns/m and Kr=400 N/m.
Initial velocity for jumper and parachute is 20 m/s
68
68
34

Lect 2 - Modeling of Dynamic Systems

Uploaded by

Copyright:

Available Formats

You might also like

Lect 2 - Modeling of Dynamic Systems

Uploaded by

Document Information

Original Description:

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

Lect 2 - Modeling of Dynamic Systems

Uploaded by

Copyright:

Available Formats

Modeling of Dynamic Systems

Ref: Control System Engineering

TM184521 – Modelling of Dynamic Systems

Review of the last lecture

Review of the last lecture

TM184521 – Modelling of Dynamic Systems

Laplace in Frequency Domain

Laplace Transforms Review

Laplace transforms can be used in process control for:

TM184521 – Modelling of Dynamic Systems

where F(s) is the symbol for the Laplace transform, L is the

Note: The L operator transforms a time domain function f(t) into an

Inverse Laplace Transform, L-1:

TM184521 – Modelling of Dynamic Systems

X ( s ) = L  x ( t ) and Y ( s ) = L  y ( t )

Laplace Transforms of Common Functions

TM184521 – Modelling of Dynamic Systems

Similarly, for higher order derivatives:

Special Case: All Initial Conditions are Zeroes

In process control problems, we usually assume zero initial

TM184521 – Modelling of Dynamic Systems

5. Rectangular Pulse Function

TM184521 – Modelling of Dynamic Systems

Solution of ODEs by Laplace Transforms

Table Laplace Transforms for Various Time-Domain Functions

TM184521 – Modelling of Dynamic Systems

TM184521 – Modelling of Dynamic Systems

Important Properties of Laplace Transforms

lim y ( t ) = lim  sY ( s )

providing that the limit exists (is finite) for all Re ( s )  0,

TM184521 – Modelling of Dynamic Systems

First, take L of both sides 5 ( sY ( s ) − 1) + 4Y ( s ) =

From LT Table, y ( t ) = 0.5 + 0.5e−0.8t (3-37)

TM184521 – Modelling of Dynamic Systems

Partial fraction expansion

1. Case 1. Roots of the denominator of F(s) are real and

To find 1 : Multiply both sides by s + 1 and let s = -1

TM184521 – Modelling of Dynamic Systems

To find  2 : Multiply both sides by s + 4 and let s = -4

The PFE is:

Note: D(s) is called the “characteristic polynomial”.

For these situations, the PFE has a different form.

TM184521 – Modelling of Dynamic Systems

To find transient response for u(t) = unit step at t > 0

1. Take Laplace Transform (L.T.)

Step 2a. Factor denominator of Y(s)

Step 2b. Use partial fraction decomposition

TM184521 – Modelling of Dynamic Systems

For a2, multiply by (s+1), set s = -1 (same procedure

Practice Matlab …..

TM184521 – Modelling of Dynamic Systems

and hence Y (s ) = 1 − 1.125 + 0.125

Using Laplace transform table, we find that :

TM184521 – Modelling of Dynamic Systems

Modeling the system

For example, a link in a bicycle chain is usually not considered

Input and Output

However, in the system dynamics meaning of the terms, an