ESE 499 โ Feedback Control Systems
SECTION 3: LAPLACE TRANSFORMS & TRANSFER FUNCTIONS
K. Webb ESE 499
This section of notes contains an introductionto Laplace transforms. This should mostly be areview of material covered in your differentialequations course.
Introduction โ Transforms2
K. Webb ESE 499
3
Transforms
What is a transform? A mapping of a mathematical function from one domain to
another A change in perspective not a change of the function
Why use transforms? Some mathematical problems are difficult to solve in their
natural domain Transform to and solve in a new domain, where the problem is
simplified Transform back to the original domain
Trade off the extra effort of transforming/inverse-transforming for simplification of the solution procedure
K. Webb ESE 499
4
Transform Example โ Slide Rules
Slide rules make use of a logarithmic transform
Multiplication/division of large numbers is difficult Transform the numbers to the logarithmic domain Add/subtract (easy) in the log domain to multiply/divide
(difficult) in the linear domain Apply the inverse transform to get back to the original
domain Extra effort is required, but the problem is simplified
K. Webb ESE 499
Laplace Transforms5
K. Webb ESE 499
6
Laplace Transforms
An integral transform mapping functions from the time domain to the Laplace domain or s-domain
๐๐ ๐ก๐ก โโ
๐บ๐บ ๐ ๐ Time-domain functions are functions of time, ๐ก๐ก
๐๐ ๐ก๐ก Laplace-domain functions are functions of ๐๐
๐บ๐บ ๐ ๐ ๐ ๐ is a complex variable
๐ ๐ = ๐๐ + ๐๐๐๐
K. Webb ESE 499
7
Laplace Transforms โ Motivation
Weโll use Laplace transforms to solve differential equations
Differential equations in the time domain difficult to solve
Apply the Laplace transform Transform to the s-domain
Differential equations become algebraic equations easy to solve
Transform the s-domain solution back to the time domain
Transforming back and forth requires extra effort, but the solution is greatly simplified
K. Webb ESE 499
8
Laplace Transform
Laplace Transform:
โ ๐๐ ๐ก๐ก = ๐บ๐บ ๐ ๐ = โซ0โ๐๐ ๐ก๐ก ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก (1)
Unilateral or one-sided transform Lower limit of integration is ๐ก๐ก = 0 Assumed that the time domain function is zero for all
negative time, i.e.
๐๐ ๐ก๐ก = 0, ๐ก๐ก < 0
K. Webb ESE 499
In the following subsection of notes, weโll derive a few important properties of the Laplace transform.
Laplace Transform Properties9
K. Webb ESE 499
10
Laplace Transform โ Linearity
Say we have two time-domain functions:๐๐1 ๐ก๐ก and ๐๐2 ๐ก๐ก
Applying the transform definition, (1)
โ ๐ผ๐ผ๐๐1 ๐ก๐ก + ๐ฝ๐ฝ๐๐2 ๐ก๐ก = ๏ฟฝ0
โ๐ผ๐ผ๐๐1 ๐ก๐ก + ๐ฝ๐ฝ๐๐2 ๐ก๐ก ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก
= ๏ฟฝ0
โ๐ผ๐ผ๐๐1 ๐ก๐ก ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก + ๏ฟฝ
0
โ๐ฝ๐ฝ๐๐2 ๐ก๐ก ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก
= ๐ผ๐ผ๏ฟฝ0
โ๐๐1 ๐ก๐ก ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก + ๐ฝ๐ฝ๏ฟฝ
0
โ๐๐2 ๐ก๐ก ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก
= ๐ผ๐ผ ๏ฟฝ โ ๐๐1 ๐ก๐ก + ๐ฝ๐ฝ ๏ฟฝ โ ๐๐2 ๐ก๐ก
โ ๐ผ๐ผ๐๐1 ๐ก๐ก + ๐ฝ๐ฝ๐๐2 ๐ก๐ก = ๐ผ๐ผ๐บ๐บ1 ๐ ๐ + ๐ฝ๐ฝ๐บ๐บ2 ๐ ๐ (2)
The Laplace transform is a linear operation
K. Webb ESE 499
11
Laplace Transform of a Derivative
Of particular interest, given that we want to use Laplace transform to solve differential equations
โ ๏ฟฝฬ๏ฟฝ๐ ๐ก๐ก = ๏ฟฝ0
โ๏ฟฝฬ๏ฟฝ๐ ๐ก๐ก ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก
Use integration by parts to evaluate
โซ ๐ข๐ข๐๐๐ข๐ข = ๐ข๐ข๐ข๐ข โ โซ ๐ข๐ข๐๐๐ข๐ข Let
๐ข๐ข = ๐๐โ๐ ๐ ๐ ๐ and ๐๐๐ข๐ข = ๏ฟฝฬ๏ฟฝ๐ ๐ก๐ก ๐๐๐ก๐กthen
๐๐๐ข๐ข = โ๐ ๐ ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก and ๐ข๐ข = ๐๐ ๐ก๐ก
K. Webb ESE 499
12
Laplace Transform of a Derivative
โ ๏ฟฝฬ๏ฟฝ๐ ๐ก๐ก = ๐๐โ๐ ๐ ๐ ๐ ๐๐ ๐ก๐ก ๏ฟฝ0
โโ ๏ฟฝ
0
โ๐๐ ๐ก๐ก โ๐ ๐ ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก
= 0 โ ๐๐ 0 + ๐ ๐ ๏ฟฝ0
โ๐๐ ๐ก๐ก ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก = โ๐๐ 0 + ๐ ๐ โ ๐๐ ๐ก๐ก
The Laplace transform of the derivative of a function is the Laplace transform of that function multiplied by ๐ ๐ minus the initial value of that function
โ ๏ฟฝฬ๏ฟฝ๐ ๐ก๐ก = ๐ ๐ ๐บ๐บ ๐ ๐ โ ๐๐(0) (3)
K. Webb ESE 499
13
Higher-Order Derivatives
The Laplace transform of a second derivative is
โ ๏ฟฝฬ๏ฟฝ๐ ๐ก๐ก = ๐ ๐ 2๐บ๐บ ๐ ๐ โ ๐ ๐ ๐๐ 0 โ ๏ฟฝฬ๏ฟฝ๐ 0 (4)
In general, the Laplace transform of the ๐๐๐๐๐๐ derivativeof a function is given by
โ ๐๐ ๐๐ = ๐ ๐ ๐๐๐บ๐บ ๐ ๐ โ ๐ ๐ ๐๐โ1๐๐ 0 โ ๐ ๐ ๐๐โ2๏ฟฝฬ๏ฟฝ๐ 0 โโฏโ ๐๐ ๐๐โ1 0 (5)
K. Webb ESE 499
14
Laplace Transform of an Integral
The Laplace Transform of a definite integral of a function is given by
โ โซ0๐ ๐ ๐๐ ๐๐ ๐๐๐๐ = 1
๐ ๐ ๐บ๐บ ๐ ๐ (6)
Differentiation in the time domain corresponds to multiplication by ๐๐ in the Laplace domain
Integration in the time domain corresponds to division by ๐๐ in the Laplace domain
K. Webb ESE 499
Next, weโll derive the Laplace transform of some common mathematical functions
Laplace Transforms of Common Functions
15
K. Webb ESE 499
16
Unit Step Function
A useful and common way of characterizing a linear system is with its step response The systemโs response (output) to a unit step input
The unit step function or Heaviside step function:
1 ๐ก๐ก = ๏ฟฝ0, ๐ก๐ก < 01, ๐ก๐ก โฅ 0
K. Webb ESE 499
17
Unit Step Function โ Laplace Transform
Using the definition of the Laplace transform
โ 1 ๐ก๐ก = ๏ฟฝ0
โ1 ๐ก๐ก ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก = ๏ฟฝ
0
โ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก
= โ1๐ ๐ ๐๐โ๐ ๐ ๐ ๐ ๏ฟฝ
0
โ= 0 โ โ
1๐ ๐
=1๐ ๐
The Laplace transform of the unit step
โ 1 ๐ก๐ก = 1๐ ๐
(7)
Note that the unilateral Laplace transform assumes that the signal being transformed is zero for ๐ก๐ก < 0 Equivalent to multiplying any signal by a unit step
K. Webb ESE 499
18
Unit Ramp Function
The unit ramp function is a useful input signal for evaluating how well a system tracks a constantly-increasing input
The unit ramp function:
๐๐ ๐ก๐ก = ๏ฟฝ0, ๐ก๐ก < 0๐ก๐ก, ๐ก๐ก โฅ 0
K. Webb ESE 499
19
Unit Ramp Function โ Laplace Transform
Could easily evaluate the transform integral Requires integration by parts
Alternatively, recognize the relationship between the unit ramp and the unit step Unit ramp is the integral of the unit step
Apply the integration property, (6)
โ ๐ก๐ก = โ ๏ฟฝ0
๐ ๐ 1 ๐๐ ๐๐๐๐ =
1๐ ๐ ๏ฟฝ
1๐ ๐
โ ๐ก๐ก = 1๐ ๐ 2
(8)
K. Webb ESE 499
20
Exponential โ Laplace Transform
๐๐ ๐ก๐ก = ๐๐โ๐๐๐ ๐
Exponentials are common components of the responses of dynamic systems
โ ๐๐โ๐๐๐ ๐ = ๏ฟฝ0
โ๐๐โ๐๐๐ ๐ ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก = ๏ฟฝ
0
โ๐๐โ(๐ ๐ +๐๐)๐ ๐ ๐๐๐ก๐ก
= โ๐๐โ ๐ ๐ +๐๐ ๐ ๐
๐ ๐ + ๐๐๏ฟฝ0
โ= 0 โ โ
1๐ ๐ + ๐๐
โ ๐๐โ๐๐๐ ๐ = 1๐ ๐ +๐๐
(9)
K. Webb ESE 499
21
Sinusoidal functions
Another class of commonly occurring signals, when dealing with dynamic systems, is sinusoidal signals โboth sin ๐๐๐ก๐ก and cos ๐๐๐ก๐ก
๐๐ ๐ก๐ก = sin ๐๐๐ก๐ก
Recall Eulerโs formula
๐๐๐๐๐๐๐ ๐ = cos ๐๐๐ก๐ก + ๐๐ sin ๐๐๐ก๐ก
From which it follows that
sin ๐๐๐ก๐ก =๐๐๐๐๐๐๐ ๐ โ ๐๐โ๐๐๐๐๐ ๐
2๐๐
K. Webb ESE 499
22
Sinusoidal functions
โ sin ๐๐๐ก๐ก =12๐๐๏ฟฝ0
โ๐๐๐๐๐๐๐ ๐ โ ๐๐โ๐๐๐๐๐ ๐ ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก
=12๐๐๏ฟฝ0
โ๐๐โ ๐ ๐ โ๐๐๐๐ ๐ ๐ โ ๐๐โ ๐ ๐ +๐๐๐๐ ๐ ๐ ๐๐๐ก๐ก
=12๐๐๏ฟฝ0
โ๐๐โ ๐ ๐ โ๐๐๐๐ ๐ ๐ ๐๐๐ก๐ก โ
12๐๐๏ฟฝ0
โ๐๐โ ๐ ๐ +๐๐๐๐ ๐ ๐ ๐๐๐ก๐ก
=12๐๐
๐๐โ ๐ ๐ โ๐๐๐๐ ๐ ๐
โ ๐ ๐ โ ๐๐๐๐ ๏ฟฝ0
โโ
12๐๐
๐๐โ ๐ ๐ +๐๐๐๐ ๐ ๐
โ ๐ ๐ + ๐๐๐๐ ๏ฟฝ0
โ
=12๐๐
0 +1
๐ ๐ โ ๐๐๐๐โ
12๐๐
0 +1
๐ ๐ + ๐๐๐๐=
12๐๐
2๐๐๐๐๐ ๐ 2 + ๐๐2
โ sin ๐๐๐ก๐ก = ๐๐๐ ๐ 2+๐๐2 (10)
K. Webb ESE 499
23
Sinusoidal functions
It can similarly be shown that
โ cos ๐๐๐ก๐ก = ๐ ๐ ๐ ๐ 2+๐๐2 (11)
Note that for neither sin(๐๐๐ก๐ก) nor cos ๐๐๐ก๐ก is the function equal to zero for ๐ก๐ก < 0 as the Laplace transform assumes
Really, what weโve derived is
โ 1 ๐ก๐ก ๏ฟฝ sin ๐๐๐ก๐ก and โ 1 ๐ก๐ก ๏ฟฝ cos ๐๐๐ก๐ก
K. Webb ESE 499
More Properties and Theorems24
K. Webb ESE 499
25
Multiplication by an Exponential, ๐๐โ๐๐๐ ๐
Weโve seen that โ ๐๐โ๐๐๐ ๐ = 1๐ ๐ +๐๐
What if another function is multiplied by the decaying exponential term?
โ ๐๐ ๐ก๐ก ๐๐โ๐๐๐ ๐ = ๏ฟฝ0
โ๐๐ ๐ก๐ก ๐๐โ๐๐๐ ๐ ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก =๏ฟฝ
0
โ๐๐ ๐ก๐ก ๐๐โ ๐ ๐ +๐๐ ๐ ๐ ๐๐๐ก๐ก
This is just the Laplace transform of ๐๐ ๐ก๐ก with ๐ ๐ replaced by ๐ ๐ + ๐๐
โ ๐๐ ๐ก๐ก ๐๐โ๐๐๐ ๐ = ๐บ๐บ ๐ ๐ + ๐๐ (12)
K. Webb ESE 499
26
Decaying Sinusoids
The Laplace transform of a sinusoid is
โ sin ๐๐๐ก๐ก =๐๐
๐ ๐ 2 + ๐๐2
And, multiplication by an decaying exponential, ๐๐โ๐๐๐ ๐ , results in a substitution of ๐ ๐ + ๐๐ for ๐ ๐ , so
โ ๐๐โ๐๐๐ ๐ sin ๐๐๐ก๐ก =๐๐
๐ ๐ + ๐๐ 2 + ๐๐2
and
โ ๐๐โ๐๐๐ ๐ cos ๐๐๐ก๐ก =๐ ๐ + ๐๐
๐ ๐ + ๐๐ 2 + ๐๐2
K. Webb ESE 499
27
Time Shifting
Consider a time-domain function, ๐๐ ๐ก๐ก
To Laplace transform ๐๐ ๐ก๐กweโve assumed ๐๐ ๐ก๐ก = 0 for ๐ก๐ก < 0, or equivalently multiplied by 1(๐ก๐ก)
To shift ๐๐ ๐ก๐ก by an amount, ๐๐, in time, we must also multiply by a shifted step function, 1 ๐ก๐ก โ ๐๐
K. Webb ESE 499
28
Time Shifting โ Laplace Transform
The transform of the shifted function is given by
โ ๐๐ ๐ก๐ก โ ๐๐ ๏ฟฝ 1 ๐ก๐ก โ ๐๐ = ๏ฟฝ๐๐
โ๐๐ ๐ก๐ก โ ๐๐ ๐๐โ๐ ๐ ๐ ๐ ๐๐๐ก๐ก
Performing a change of variables, let
๐๐ = ๐ก๐ก โ ๐๐ and ๐๐๐๐ = ๐๐๐ก๐ก
The transform becomes
โ ๐๐ ๐๐ ๏ฟฝ 1 ๐๐ = ๏ฟฝ0
โ๐๐ ๐๐ ๐๐โ๐ ๐ ๐๐+๐๐ ๐๐๐๐ = ๏ฟฝ
0
โ๐๐ ๐๐ ๐๐โ๐๐๐ ๐ ๐๐โ๐ ๐ ๐๐๐๐๐๐ = ๐๐โ๐๐๐ ๐ ๏ฟฝ
0
โ๐๐ ๐๐ ๐๐โ๐ ๐ ๐๐๐๐๐๐
A shift by ๐๐ in the time domain corresponds to multiplication by ๐๐โ๐๐๐ ๐ in the Laplace domain
โ ๐๐ ๐ก๐ก โ ๐๐ ๏ฟฝ 1 ๐ก๐ก โ ๐๐ = ๐๐โ๐๐๐ ๐ ๐บ๐บ ๐ ๐ (13)
K. Webb ESE 499
29
Multiplication by time, ๐ก๐ก
The Laplace transform of a function multiplied by time:
โ ๐ก๐ก โ ๐๐ ๐ก๐ก = โ ๐๐๐๐๐ ๐ ๐น๐น(๐ ๐ ) (14)
Consider a unit ramp function:
โ ๐ก๐ก = โ ๐ก๐ก โ 1 ๐ก๐ก = โ๐๐๐๐๐ ๐
1๐ ๐ =
1๐ ๐ 2
Or a parabola:โ ๐ก๐ก2 = โ ๐ก๐ก โ ๐ก๐ก = โ ๐๐
๐๐๐ ๐ 1๐ ๐ 2
= 2๐ ๐ 3
In general
โ ๐ก๐ก๐๐ =๐๐!๐ ๐ ๐๐+1
K. Webb ESE 499
30
Initial and Final Value Theorems
Initial Value Theorem Can determine the initial value of a time-domain signal or
function from its Laplace transform
(15)
Final Value Theorem Can determine the steady-state value of a time-domain
signal or function from its Laplace transform
(16)
๐๐ 0 = lim๐ ๐ โโ
๐ ๐ ๐บ๐บ ๐ ๐
๐๐ โ = lim๐ ๐ โ0
๐ ๐ ๐บ๐บ ๐ ๐
K. Webb ESE 499
31
Convolution
Convolution of two functions or signals is given by
๐๐ ๐ก๐ก โ ๐ฅ๐ฅ ๐ก๐ก = ๏ฟฝ0
๐ ๐ ๐๐ ๐๐ ๐ฅ๐ฅ ๐ก๐ก โ ๐๐ ๐๐๐๐
Result is a function of time ๐ฅ๐ฅ ๐๐ is flipped in time and shifted by ๐ก๐ก Multiply the flipped/shifted signal and the other signal Integrate the result from ๐๐ = 0 โฆ ๐ก๐ก
May seem like an odd, arbitrary function now, but weโll later see why it is very important
Convolution in the time domain corresponds to multiplication in the Laplace domain
โ ๐๐ ๐ก๐ก โ ๐ฅ๐ฅ ๐ก๐ก = ๐บ๐บ ๐ ๐ ๐๐ ๐ ๐ (17)
K. Webb ESE 499
32
Impulse Function
Another common way to describe a dynamic system is with its impulse response System output in response to an impulse function input
Impulse function defined by๐ฟ๐ฟ ๐ก๐ก = 0, ๐ก๐ก โ 0
๏ฟฝโโ
โ๐ฟ๐ฟ ๐ก๐ก ๐๐๐ก๐ก = 1
An infinitely tall, infinitely narrow pulse
K. Webb ESE 499
33
Impulse Function โ Laplace Transform
To derive โ ๐ฟ๐ฟ ๐ก๐ก , consider the following function
๐๐ ๐ก๐ก = ๏ฟฝ1๐ก๐ก0
, 0 โค ๐ก๐ก โค ๐ก๐ก0
0, ๐ก๐ก < 0 or ๐ก๐ก > ๐ก๐ก0
Can think of ๐๐ ๐ก๐ก as the sum of two step functions:
๐๐ ๐ก๐ก =1๐ก๐ก0
1 ๐ก๐ก โ1๐ก๐ก0
1 ๐ก๐ก โ ๐ก๐ก0
The transform of the first term is
โ1๐ก๐ก0
1 ๐ก๐ก =1๐ก๐ก0๐ ๐
Using the time-shifting property, the second term transforms to
โ โ1๐ก๐ก0
1 ๐ก๐ก โ ๐ก๐ก0 = โ๐๐โ๐ ๐ 0๐ ๐
๐ก๐ก0๐ ๐
K. Webb ESE 499
34
Impulse Function โ Laplace Transform
In the limit, as ๐ก๐ก0 โ 0, ๐๐ ๐ก๐ก โ ๐ฟ๐ฟ ๐ก๐ก , so
โ ๐ฟ๐ฟ ๐ก๐ก = lim๐ ๐ 0โ0
โ ๐๐ ๐ก๐ก
โ ๐ฟ๐ฟ ๐ก๐ก = lim๐ ๐ 0โ0
1 โ ๐๐โ๐ ๐ 0๐ ๐
๐ก๐ก0๐ ๐
Apply lโHรดpitalโs rule
โ ๐ฟ๐ฟ ๐ก๐ก = lim๐ ๐ 0โ0
๐๐๐๐๐ก๐ก0
1 โ ๐๐โ๐ ๐ 0๐ ๐
๐๐๐๐๐ก๐ก0
๐ก๐ก0๐ ๐ = lim
๐ ๐ 0โ0
๐ ๐ ๐๐โ๐ ๐ 0๐ ๐
๐ ๐ =๐ ๐ ๐ ๐
The Laplace transform of an impulse function is one
โ ๐ฟ๐ฟ ๐ก๐ก = 1 (18)
K. Webb ESE 499
35
Common Laplace Transforms
๐๐ ๐๐ ๐ฎ๐ฎ ๐๐
๐ฟ๐ฟ ๐ก๐ก 1
1 ๐ก๐ก 1๐ ๐
๐ก๐ก 1๐ ๐ 2
๐ก๐ก๐๐ ๐๐!๐ ๐ ๐๐+1
๐๐โ๐๐๐ ๐ 1๐ ๐ + ๐๐
๐ก๐ก๐๐โ๐๐๐ ๐ 1
๐ ๐ + ๐๐ 2
sin ๐๐๐ก๐ก๐๐
๐ ๐ 2 + ๐๐2
cos ๐๐๐ก๐ก๐ ๐
๐ ๐ 2 + ๐๐2
๐๐ ๐๐ ๐ฎ๐ฎ ๐๐
๐๐โ๐๐๐ ๐ sin ๐๐๐ก๐ก๐๐
๐ ๐ + ๐๐ 2 + ๐๐2
๐๐โ๐๐๐ ๐ cos ๐๐๐ก๐ก๐ ๐ + ๐๐
๐ ๐ + ๐๐ 2 + ๐๐2
๏ฟฝฬ๏ฟฝ๐(๐ก๐ก) ๐ ๐ ๐บ๐บ ๐ ๐ โ ๐๐(0)
๏ฟฝฬ๏ฟฝ๐ ๐ก๐ก ๐ ๐ 2๐บ๐บ ๐ ๐ โ ๐ ๐ ๐๐ 0 โ ๏ฟฝฬ๏ฟฝ๐ 0
๏ฟฝ0
๐ ๐ ๐๐ ๐๐ ๐๐๐๐
1๐ ๐ ๐บ๐บ ๐ ๐
๐๐โ๐๐๐ ๐ ๐๐ ๐ก๐ก ๐บ๐บ ๐ ๐ + ๐๐
๐๐ ๐ก๐ก โ ๐๐ ๏ฟฝ 1 ๐ก๐ก โ ๐๐ ๐๐โ๐๐๐ ๐ ๐บ๐บ ๐ ๐
๐ก๐ก ๏ฟฝ ๐๐ ๐ก๐ก โ๐๐๐๐๐ ๐ ๐บ๐บ ๐ ๐
K. Webb ESE 499
Weโve just seen how time-domain functions can betransformed to the Laplace domain. Next, weโll look athow we can solve differential equations in the Laplacedomain and transform back to the time domain.
Inverse Laplace Transform36
K. Webb ESE 499
37
Laplace Transforms โ Differential Equations
Consider the simple spring/mass/damper system from the previous section of notes
Applying Newtonโs 2nd law at the mass:
๐น๐น๐๐๐๐ ๐ก๐ก โ ๐๐๏ฟฝฬ๏ฟฝ๐ฅ โ ๐๐๐ฅ๐ฅ = ๐๐๏ฟฝฬ๏ฟฝ๐ฅ
Rearranging, gives the governing second-order differential equation
๏ฟฝฬ๏ฟฝ๐ฅ + ๐๐๐๐๏ฟฝฬ๏ฟฝ๐ฅ + ๐๐
๐๐๐ฅ๐ฅ = 1
๐๐๐น๐น๐๐๐๐ ๐ก๐ก (1)
This equations describes the dynamic response of the system to some applied input force, ๐น๐น๐๐๐๐ ๐ก๐ก
That input may take many forms, e.g.: Step Sinusoidal Impulse, etc.
K. Webb ESE 499
38
Laplace Transforms โ Differential Equations
Weโll now use Laplace transforms to determine the step response of the system
1N step force input
๐น๐น๐๐๐๐ ๐ก๐ก = 1๐๐ ๏ฟฝ 1 ๐ก๐ก = ๏ฟฝ0๐๐, ๐ก๐ก < 01๐๐, ๐ก๐ก โฅ 0 (2)
For the step response, we assume zero initial conditions
๐ฅ๐ฅ 0 = 0 and ๏ฟฝฬ๏ฟฝ๐ฅ 0 = 0 (3)
Using the derivative property of the Laplace transform, (1) becomes
๐ ๐ 2๐๐ ๐ ๐ โ ๐ ๐ ๐ฅ๐ฅ 0 โ ๏ฟฝฬ๏ฟฝ๐ฅ 0 +๐๐๐๐๐ ๐ ๐๐ ๐ ๐ โ
๐๐๐๐๐ฅ๐ฅ(0) +
๐๐๐๐๐๐ ๐ ๐ =
1๐๐๐น๐น๐๐๐๐ ๐ ๐
๐ ๐ 2๐๐ ๐ ๐ + ๐๐๐๐๐ ๐ ๐๐ ๐ ๐ + ๐๐
๐๐๐๐ ๐ ๐ = 1
๐๐๐น๐น๐๐๐๐ ๐ ๐ (4)
K. Webb ESE 499
39
Laplace Transforms โ Differential Equations
The input is a step, so (7) becomes
๐ ๐ 2๐๐ ๐ ๐ + ๐๐๐๐๐ ๐ ๐๐ ๐ ๐ + ๐๐
๐๐๐๐ ๐ ๐ = 1
๐๐1๐๐ 1
๐ ๐ (5)
Solving (5) for ๐๐ ๐ ๐
๐๐ ๐ ๐ ๐ ๐ 2 +๐๐๐๐ ๐ ๐ +
๐๐๐๐ =
1๐๐
1๐ ๐
๐๐ ๐ ๐ = 1/๐๐
๐ ๐ ๐ ๐ 2+๐๐๐๐๐ ๐ +
๐๐๐๐
(6)
Equation (6) is the solution to the differential equation of (1), given the step input and I.C.โs The system step response in the Laplace domain Next, we need to transform back to the time domain
K. Webb ESE 499
40
Laplace Transforms โ Differential Equations
๐๐ ๐ ๐ = 1/๐๐
๐ ๐ ๐ ๐ 2+๐๐๐๐๐ ๐ +
๐๐๐๐
(6)
The form of (6) is typical of Laplace transforms when dealing with linear systems
A rational polynomial in ๐ ๐ Here, the numerator is 0th-order
๐๐ ๐ ๐ =๐ต๐ต ๐ ๐ ๐ด๐ด ๐ ๐
Roots of the numerator polynomial, ๐ต๐ต ๐ ๐ , are called the zeros of the function
Roots of the denominator polynomial, ๐ด๐ด ๐ ๐ , are called the polesof the function
K. Webb ESE 499
41
Inverse Laplace Transforms
๐๐ ๐ ๐ = 1/๐๐
๐ ๐ ๐ ๐ 2+๐๐๐๐๐ ๐ +
๐๐๐๐
(6)
To get (6) back into the time domain, we need to perform an inverse Laplace transform An integral inverse transform exists, but we donโt use it Instead, we use partial fraction expansion
Partial fraction expansion Idea is to express the Laplace transform solution, (6), as a sum of
Laplace transform terms that appear in the table Procedure depends on the type of roots of the denominator polynomial
Real and distinct Repeated Complex
K. Webb ESE 499
42
Inverse Laplace Transforms โ Example 1
Consider the following system parameters๐๐ = 1๐๐๐๐
๐๐ =16๐๐๐๐
๐๐ = 10๐๐ ๏ฟฝ ๐ ๐ ๐๐
Laplace transform of the step response becomes
๐๐ ๐ ๐ = 1๐ ๐ ๐ ๐ 2+10๐ ๐ +16
(7)
Factoring the denominator
๐๐ ๐ ๐ = 1๐ ๐ ๐ ๐ +2 ๐ ๐ +8
(8)
In this case, the denominator polynomial has three real, distinct roots
๐ ๐ 1 = 0, ๐ ๐ 2 = โ2, ๐ ๐ 3 = โ8
K. Webb ESE 499
43
Inverse Laplace Transforms โ Example 1
Partial fraction expansion of (8) has the form
๐๐ ๐ ๐ = 1๐ ๐ ๐ ๐ +2 ๐ ๐ +8
= ๐๐1๐ ๐
+ ๐๐2๐ ๐ +2
+ ๐๐3๐ ๐ +8
(9)
The numerator coefficients, ๐๐1, ๐๐2, and ๐๐3, are called residues
Can already see the form of the time-domain function Sum of a constant and two decaying exponentials
To determine the residues, multiply both sides of (9) by the denominator of the left-hand side
1 = ๐๐1 ๐ ๐ + 2 ๐ ๐ + 8 + ๐๐2๐ ๐ ๐ ๐ + 8 + ๐๐3๐ ๐ ๐ ๐ + 2
1 = ๐๐1๐ ๐ 2 + 10๐๐1๐ ๐ + 16๐๐1 + ๐๐2๐ ๐ 2 + 8๐๐2๐ ๐ + ๐๐3๐ ๐ 2 + 2๐๐3๐ ๐
Collecting terms, we have
1 = ๐ ๐ 2 ๐๐1 + ๐๐2 + ๐๐3 + ๐ ๐ 10๐๐1 + 8๐๐2 + 2๐๐3 + 16๐๐1 (10)
K. Webb ESE 499
44
Inverse Laplace Transforms โ Example 1
Equating coefficients of powers of ๐ ๐ on both sides of (10) gives a system of three equations in three unknowns
๐ ๐ 2: ๐๐1 + ๐๐2 + ๐๐3 = 0๐ ๐ 1: 10๐๐1 + 8๐๐2 + 2๐๐3 = 0๐ ๐ 0: 16๐๐1 = 1
Solving for the residues gives๐๐1 = 0.0625๐๐2 = โ0.0833๐๐3 = 0.0208
The Laplace transform of the step response is
๐๐ ๐ ๐ = 0.0625๐ ๐
โ 0.0833๐ ๐ +2
+ 0.0208๐ ๐ +8
(11)
Equation (11) can now be transformed back to the time domain using the Laplace transform table
K. Webb ESE 499
45
Inverse Laplace Transforms โ Example 1
The time-domain step response of the system is the sum of a constant term and two decaying exponentials:
๐ฅ๐ฅ ๐ก๐ก = 0.0625 โ 0.0833๐๐โ2๐ ๐ + 0.0208๐๐โ8๐ ๐ (12)
Step response plotted in MATLAB
Characteristic of a signal having only real poles No overshoot/ringing
Steady-state displacement agrees with intuition 1๐๐ force applied to a 16๐๐/๐๐
spring
K. Webb ESE 499
46
Inverse Laplace Transforms โ Example 1
Go back to (7) and apply the initial value theorem
๐ฅ๐ฅ 0 = lim๐ ๐ โโ
๐ ๐ ๐๐ ๐ ๐ = lim๐ ๐ โโ
1๐ ๐ 2 + 10๐ ๐ + 16
= 0๐๐๐๐
Which is, in fact our assumed initial condition
Next, apply the final value theorem to the Laplace transform step response, (7)
๐ฅ๐ฅ โ = lim๐ ๐ โ0
๐ ๐ ๐๐ ๐ ๐ = lim๐ ๐ โ0
1๐ ๐ 2 + 10๐ ๐ + 16
๐ฅ๐ฅ โ = 116
= 0.0625๐๐ = 6.25๐๐๐๐
This final value agrees with both intuition and our numerical analysis
K. Webb ESE 499
47
Inverse Laplace Transforms โ Example 2
Reduce the damping and re-calculate the step response
๐๐ = 1๐๐๐๐
๐๐ =16๐๐๐๐
๐๐ = 8๐๐ ๏ฟฝ ๐ ๐ ๐๐
Laplace transform of the step response becomes
๐๐ ๐ ๐ = 1๐ ๐ ๐ ๐ 2+8๐ ๐ +16
(13)
Factoring the denominator
๐๐ ๐ ๐ = 1๐ ๐ ๐ ๐ +4 2 (14)
In this case, the denominator polynomial has three real roots, two of which are identical
๐ ๐ 1 = 0, ๐ ๐ 2 = โ4, ๐ ๐ 3 = โ4
K. Webb ESE 499
48
Inverse Laplace Transforms โ Example 2
Partial fraction expansion of (14) has the form
๐๐ ๐ ๐ = 1๐ ๐ ๐ ๐ +4 2 = ๐๐1
๐ ๐ + ๐๐2
๐ ๐ +4+ ๐๐3
๐ ๐ +4 2 (15)
Again, find residues by multiplying both sides of (15) by the left-hand side denominator
1 = ๐๐1 ๐ ๐ + 4 2 + ๐๐2๐ ๐ ๐ ๐ + 4 + ๐๐3๐ ๐
1 = ๐๐1๐ ๐ 2 + 8๐๐1๐ ๐ + 16๐๐1 + ๐๐2๐ ๐ 2 + 4๐๐2๐ ๐ + ๐๐3๐ ๐
Collecting terms, we have
1 = ๐ ๐ 2 ๐๐1 + ๐๐2 + ๐ ๐ 8๐๐1 + 4๐๐2 + ๐๐3 + 16๐๐1 (16)
K. Webb ESE 499
49
Inverse Laplace Transforms โ Example 2
Equating coefficients of powers of ๐ ๐ on both sides of (16) gives a system of three equations in three unknowns
๐ ๐ 2: ๐๐1 + ๐๐2 = 0๐ ๐ 1: 8๐๐1 + 4๐๐2 + ๐๐3 = 0๐ ๐ 0: 16๐๐1 = 1
Solving for the residues gives๐๐1 = 0.0625๐๐2 = โ0.0625๐๐3 = โ0.2500
The Laplace transform of the step response is
๐๐ ๐ ๐ = 0.0625๐ ๐
โ 0.0625๐ ๐ +4
โ 0.25๐ ๐ +4 2 (17)
Equation (17) can now be transformed back to the time domain using the Laplace transform table
K. Webb ESE 499
50
Inverse Laplace Transforms โ Example 2
The time-domain step response of the system is the sum of a constant, a decaying exponential, and a decaying exponential scaled by time:
๐ฅ๐ฅ ๐ก๐ก = 0.0625 โ 0.0625๐๐โ4๐ ๐ โ 0. 25๐ก๐ก๐๐โ4๐ ๐ (18)
Step response plotted in MATLAB
Again, characteristic of a signal having only real poles
Similar to the last case
A bit faster โ slow pole at ๐ ๐ = โ2was eliminated
K. Webb ESE 499
51
Inverse Laplace Transforms โ Example 3
Reduce the damping even further and go through the process once again
๐๐ = 1๐๐๐๐
๐๐ =16๐๐๐๐
๐๐ = 4๐๐ ๏ฟฝ ๐ ๐ ๐๐
Laplace transform of the step response becomes
๐๐ ๐ ๐ = 1๐ ๐ ๐ ๐ 2+4๐ ๐ +16
(19)
The second-order term in the denominator now has complex roots, so we wonโt factor any further
The denominator polynomial still has a root at zero and now has two roots which are a complex-conjugate pair
๐ ๐ 1 = 0, ๐ ๐ 2 = โ2 + ๐๐๐.464, ๐ ๐ 3 = โ2 โ ๐๐๐.464
K. Webb ESE 499
52
Inverse Laplace Transforms โ Example 3
Want to cast the partial fraction terms into forms that appear in the Laplace transform table
Second-order terms should be of the form
๐๐๐๐(๐ ๐ +๐๐)+๐๐๐๐+1๐๐๐ ๐ +๐๐ 2+๐๐2 (20)
This will transform into the sum of damped sine and cosine terms
โโ1 ๐๐๐๐๐ ๐ + ๐๐
๐ ๐ + ๐๐ 2 + ๐๐2 + ๐๐๐๐+1๐๐
๐ ๐ + ๐๐ 2 + ๐๐2 = ๐๐๐๐๐๐โ๐๐๐ ๐ cos ๐๐๐ก๐ก + ๐๐๐๐+1๐๐โ๐๐๐ ๐ sin ๐๐๐ก๐ก
To get the second-order term in the denominator of (19) into the form of (20), complete the square, to give the following partial fraction expansion
๐๐ ๐ ๐ = 1๐ ๐ ๐ ๐ 2+4๐ ๐ +16
= ๐๐1๐ ๐
+ ๐๐2 ๐ ๐ +2 +๐๐3 3.464๐ ๐ +2 2+ 3.464 2 (21)
K. Webb ESE 499
53
Inverse Laplace Transforms โ Example 3
Note that the ๐๐ and ๐๐ terms in (20) and (24) are the real and imaginary parts of the complex-conjugate denominator roots
๐ ๐ 2,3 = โ๐๐ ยฑ ๐๐๐๐ = โ2 ยฑ ๐๐๐.464
Multiplying both sides of (21) by the left-hand-side denominator, equate coefficients and solve for residues as before:
๐๐1 = 0.0625๐๐2 = โ0.0625๐๐3 = โ0.0361
Laplace transform of the step response is
๐๐ ๐ ๐ = 0.0625๐ ๐
โ 0.0625 ๐ ๐ +2๐ ๐ +2 2+ 3.464 2 โ
0.0361 3.464๐ ๐ +2 2+ 3.464 2 (22)
K. Webb ESE 499
54
Inverse Laplace Transforms โ Example 3
The time-domain step response of the system is the sum of a constant and two decaying sinusoids:
๐ฅ๐ฅ ๐ก๐ก = 0.0625 โ 0.0625๐๐โ2๐ ๐ cos 3.464๐ก๐ก โ 0.0361๐๐โ2๐ ๐ sin(3.464๐ก๐ก) (23)
Step response and individual components plotted in MATLAB
Characteristic of a signal having complex poles
Sinusoidal terms result in overshoot and (possibly) ringing
K. Webb ESE 499
55
Laplace-Domain Signals with Complex Poles
The Laplace transform of the step response in the last example had complex poles A complex-conjugate pair: ๐ ๐ = โ๐๐ ยฑ ๐๐๐๐
Results in sine and cosine terms in the time domain
๐ด๐ด๐๐โ๐๐๐ ๐ cos ๐๐๐ก๐ก + ๐ต๐ต๐๐โ๐๐๐ ๐ sin ๐๐๐ก๐ก
Imaginary part of the roots, ๐๐ Frequency of oscillation of sinusoidal
components of the signal
Real part of the roots, ๐๐, Rate of decay of the sinusoidal
components
Much more on this later
K. Webb ESE 499
Natural and Forced Responses56
K. Webb ESE 499
57
Natural and Forced Responses
So far, weโve used Laplace transforms to determine the response of a system to a step input, given zero initial conditions The driven response
Now, consider the response of the same system to non-zero initial conditions only The natural response
K. Webb ESE 499
58
Natural Response
๏ฟฝฬ๏ฟฝ๐ฅ + ๐๐๐๐๏ฟฝฬ๏ฟฝ๐ฅ + ๐๐
๐๐๐ฅ๐ฅ = 0 (1)
Use the derivative property to Laplace transform (1) Allow for non-zero initial-conditions
๐ ๐ 2๐๐ ๐ ๐ โ ๐ ๐ ๐ฅ๐ฅ 0 โ ๏ฟฝฬ๏ฟฝ๐ฅ 0 + ๐๐๐๐๐ ๐ ๐๐ ๐ ๐ โ ๐๐
๐๐๐ฅ๐ฅ 0 + ๐๐
๐๐๐๐ ๐ ๐ = 0 (2)
Same spring/mass/damper system Set the input to zero Second-order ODE for displacement
of the mass:
K. Webb ESE 499
59
Natural Response
Solving (2) for ๐๐ ๐ ๐ gives the Laplace transform of the output due solely to initial conditions
Laplace transform of the natural response
๐๐ ๐ ๐ =๐ ๐ ๐ฅ๐ฅ 0 + ๏ฟฝฬ๏ฟฝ๐ฅ 0 + ๐๐
๐๐๐ฅ๐ฅ 0
๐ ๐ 2+๐๐๐๐๐ ๐ +
๐๐๐๐
(3)
Consider the under-damped system with the following initial conditions
๐๐ = 1 ๐๐๐๐
๐๐ = 16 ๐๐๐๐
๐๐ = 4 ๐๐๏ฟฝ๐ ๐ ๐๐
๐ฅ๐ฅ 0 = 0.15 ๐๐
๏ฟฝฬ๏ฟฝ๐ฅ 0 = 0.1๐๐๐ ๐
K. Webb ESE 499
60
Natural Response
Substituting component parameters and initial conditions into (3)
๐๐ ๐ ๐ = 0.15๐ ๐ +0.7๐ ๐ 2+4๐ ๐ +16
(4)
Remember, itโs the roots of the denominator polynomial that dictate the form of the response Real roots โ decaying exponentials Complex roots โ decaying sinusoids
For the under-damped case, roots are complex Complete the square Partial fraction expansion has the form
๐๐ ๐ ๐ = 0.15๐ ๐ +0.7๐ ๐ 2+4๐ ๐ +16
= ๐๐1 ๐ ๐ +2 +๐๐2 3.464๐ ๐ +2 2+ 3.464 2 (5)
K. Webb ESE 499
61
Natural Response
๐๐ ๐ ๐ = 0.15๐ ๐ +0.7๐ ๐ 2+4๐ ๐ +16
= ๐๐1 ๐ ๐ +2 +๐๐2 3.464๐ ๐ +2 2+ 3.464 2 (5)
Multiply both sides of (5) by the denominator of the left-hand side
0.15๐ ๐ + 0.7 = ๐๐1๐ ๐ + 2๐๐1 + 3.464๐๐2
Equating coefficients and solving for ๐๐1 and ๐๐2 gives
๐๐1 = 0.15, ๐๐2 = 0.115
The Laplace transform of the natural response:
๐๐ ๐ ๐ = 0.15 ๐ ๐ +2๐ ๐ +2 2+ 3.464 2 + 0.115 3.464
๐ ๐ +2 2+ 3.464 2 (6)
K. Webb ESE 499
62
Natural Response
Inverse Laplace transform is the natural response๐ฅ๐ฅ ๐ก๐ก = 0.15๐๐โ2๐ ๐ cos 3.464 ๏ฟฝ ๐ก๐ก + 0.115๐๐โ2๐ ๐ sin 3.464 ๏ฟฝ ๐ก๐ก (7)
Under-damped response is the sum of decaying sine and cosine terms
K. Webb ESE 499
63
Now, Laplace transform, allowing for both non-zero input and initial conditions
๐ ๐ 2๐๐ ๐ ๐ โ ๐ ๐ ๐ฅ๐ฅ 0 โ ๏ฟฝฬ๏ฟฝ๐ฅ 0 + ๐๐๐๐๐ ๐ ๐๐ ๐ ๐ โ ๐๐
๐๐๐ฅ๐ฅ 0 + ๐๐
๐๐๐๐ ๐ ๐ = 1
๐๐๐น๐น๐๐๐๐ ๐ ๐
Solving for ๐๐ ๐ ๐ , gives the Laplace transform of the response to both the input and the initial conditions
๐๐ ๐ ๐ =๐ ๐ ๐ฅ๐ฅ 0 + ๏ฟฝฬ๏ฟฝ๐ฅ 0 + ๐๐
๐๐๐ฅ๐ฅ 0 +1๐๐๐น๐น๐๐๐๐ ๐ ๐
๐ ๐ 2+๐๐๐๐๐ ๐ +
๐๐๐๐
(8)
Driven Response with Non-Zero I.C.โs
๏ฟฝฬ๏ฟฝ๐ฅ +๐๐๐๐๏ฟฝฬ๏ฟฝ๐ฅ +
๐๐๐๐๐ฅ๐ฅ =
1๐๐๐น๐น๐๐๐๐ ๐ก๐ก
K. Webb ESE 499
64
Laplace transform of the response has two components
๐๐ ๐ ๐ =๐ ๐ ๐ฅ๐ฅ 0 + ๏ฟฝฬ๏ฟฝ๐ฅ 0 + ๐๐
๐๐๐ฅ๐ฅ 0
๐ ๐ 2+๐๐๐๐๐ ๐ +
๐๐๐๐
+1๐๐๐น๐น๐๐๐๐ ๐ ๐
๐ ๐ 2+๐๐๐๐๐ ๐ +
๐๐๐๐
Total response is a superposition of the initial condition response and the driven response
Both have the same denominator polynomial Same roots, same type of response
Over-, under-, critically-damped
Driven Response with Non-Zero I.C.โs
Natural response - initial conditions Driven response - input
K. Webb ESE 499
65
Driven Response with Non-Zero I.C.โs
Laplace transform of the total response
๐๐ ๐ ๐ =0.15๐ ๐ + 0.7 + 1
๐ ๐ ๐ ๐ 2 + 4๐ ๐ + 16
=0.15๐ ๐ 2 + 0.7๐ ๐ + 1๐ ๐ ๐ ๐ 2 + 4๐ ๐ + 16
Transform back to time domain via partial fraction expansion
๐๐ ๐ ๐ =๐๐1๐ ๐
+๐๐2 ๐ ๐ + 2
๐ ๐ + 2 2 + 3.464 2 +๐๐3 3.464
๐ ๐ + 2 2 + 3.464 2
Solving for the residues gives
๐๐1 = 0.0625, ๐๐2 = 0.0875, ๐๐3 = 0.0794
๐๐ = 1 ๐๐๐๐
๐๐ = 16 ๐๐๐๐
๐๐ = 4 ๐๐๏ฟฝ๐ ๐ ๐๐
๐ฅ๐ฅ 0 = 0.15 ๐๐
๏ฟฝฬ๏ฟฝ๐ฅ 0 = 0.1๐๐๐ ๐
๐น๐น๐๐๐๐ ๐ก๐ก = 1๐๐ ๏ฟฝ 1 ๐ก๐ก
K. Webb ESE 499
66
Driven Response with Non-Zero I.C.โs
Total response:๐ฅ๐ฅ ๐ก๐ก = 0.0625 + 0.0875๐๐โ2๐ ๐ cos 3.464 ๏ฟฝ ๐ก๐ก + 0.0794๐๐โ2๐ ๐ sin 3.464 ๏ฟฝ ๐ก๐ก
Superposition of two components Natural response
due to initial conditions
Driven response due to the input
K. Webb ESE 499
Transfer Functions67
K. Webb ESE 499
68
Transfer Functions
Consider a system block diagram in the Laplace domain ๐๐ ๐ ๐ is the Laplace-domain input ๐๐ ๐ ๐ is the Laplace-domain output ๐บ๐บ ๐ ๐ is a Laplace-domain model
for the plant To understand what ๐บ๐บ ๐ ๐ is, revisit the previous example
We saw that if we assume zero initial conditions, the Laplace-domain output is
๐๐ ๐ ๐ =๏ฟฝ1 ๐๐
๐ ๐ 2 + ๐๐๐๐ ๐ ๐ + ๐๐
๐๐๐๐ ๐ ๐
K. Webb ESE 499
69
Transfer Functions
๐๐ ๐ ๐ =1๐๐
๐ ๐ 2 + ๐๐๐๐ ๐ ๐ + ๐๐
๐๐๐๐ ๐ ๐
The Laplace transform of the output is the Laplace transform of the input multiplied by the stuff in square brackets
๐๐ ๐ ๐ = ๐บ๐บ ๐ ๐ โ ๐๐ ๐ ๐
This is the transfer function
๐บ๐บ ๐ ๐ =๐๐ ๐ ๐ ๐๐ ๐ ๐
The ratio of the systemโs output to input in the Laplace domain, assuming zero initial conditions
A mathematical model for a dynamic system This is the model we will use for control system design in this course
K. Webb ESE 499
70
Transfer Functions
The transfer function for our example system is
๐บ๐บ ๐ ๐ =๏ฟฝ1 ๐๐
๐ ๐ 2 + ๐๐๐๐ ๐ ๐ + ๐๐
๐๐
Note that, as was the case for the Laplace transform of the output, the transfer function is a rational polynomial in s
๐บ๐บ ๐ ๐ =๐ต๐ต ๐ ๐ ๐ด๐ด ๐ ๐
Roots of ๐ต๐ต ๐ ๐ are the system zeros Roots of ๐ด๐ด ๐ ๐ are the system poles These poles and zeros dictate the nature of the systemโs
response
K. Webb ESE 499
71
SISO vs. MIMO Systems
Note that we have assumed a system with one input and one output
A single-input single-output (SISO) system
Often, systems have multiple inputs and multiple outputs
A multiple-input multiple-output (MIMO) system
K. Webb ESE 499
72
MIMO Systems โ Transfer Matrix
For MIMO systems, the transfer function becomes a transfer matrix For an m-input p-output system, a pรm matrix of transfer
functions
๐ฎ๐ฎ ๐ ๐ =๐บ๐บ11 ๐ ๐ โฏ ๐บ๐บ1๐๐ ๐ ๐ โฎ โฑ โฎ
๐บ๐บ๐๐1 ๐ ๐ โฏ ๐บ๐บ๐๐๐๐ ๐ ๐
Transfer function ๐บ๐บ๐๐๐๐ ๐ ๐ relates the ๐๐๐ ๐ ๐ก output to the ๐๐๐ ๐ ๐ก input
๐บ๐บ๐๐๐๐ ๐ ๐ =๐๐๐๐ ๐ ๐ ๐๐๐๐ ๐ ๐
In this course, weโll restrict our focus entirely to SISO systems
m inputsp outputs
K. Webb ESE 499
73
Characteristic Polynomial
๐บ๐บ ๐ ๐ =๐ต๐ต ๐ ๐ ๐ด๐ด ๐ ๐
The denominator of the transfer function is the characteristic polynomial, ฮ ๐ ๐
๐บ๐บ ๐ ๐ =๐ต๐ต ๐ ๐ ฮ ๐ ๐
Poles of the transfer function are roots of ฮ ๐ ๐ System poles or eigenvalues Poles determine the terms in the partial-fraction-expansion of the
systemโs natural response Along with the input, system poles determine the nature of the time-
domain response
K. Webb ESE 499
Using the Transfer Function to Determine System Response74
K. Webb ESE 499
75
Using ๐บ๐บ ๐ ๐ to determine System Response
System output in the Laplace domain can be expressed in terms of the transfer function as
๐๐ ๐ ๐ = ๐๐ ๐ ๐ ๐บ๐บ ๐ ๐ (1)
Laplace-domain output is the product of the Laplace-domain input and the transfer function
Response to two specific types of inputs often used to characterize dynamic systems Impulse response Step response
Weโll use the approach of (1) to determine these responses
K. Webb ESE 499
76
Impulse response
Impulse function
๐ฟ๐ฟ ๐ก๐ก = 0, ๐ก๐ก โ 0
โซโโโ ๐ฟ๐ฟ ๐ก๐ก ๐๐๐ก๐ก = 1
Laplace transform of the impulse function is
โ ๐ฟ๐ฟ ๐ก๐ก = 1
Impulse response in the Laplace domain is
๐๐ ๐ ๐ = 1 ๏ฟฝ ๐บ๐บ ๐ ๐ = ๐บ๐บ ๐ ๐
The transfer function is the Laplace transform of the impulse response Impulse response in the time domain is the inverse transform of the
transfer function
๐ฆ๐ฆ ๐ก๐ก = ๐๐ ๐ก๐ก = โโ1 ๐บ๐บ ๐ ๐
K. Webb ESE 499
77
Step Response
Step function: 1 ๐ก๐ก = ๏ฟฝ0 ๐ก๐ก < 01 ๐ก๐ก โฅ 0
Laplace transform of the step function
โ 1 ๐ก๐ก = 1๐ ๐
Laplace-domain step response
๐๐ ๐ ๐ = 1๐ ๐ ๏ฟฝ ๐บ๐บ ๐ ๐
Time-domain step response
๐ฆ๐ฆ ๐ก๐ก = โโ1 1๐ ๐ ๏ฟฝ ๐บ๐บ ๐ ๐
Recall the integral property of the Laplace transform
โ โซ0๐ ๐ ๐๐ ๐๐ ๐๐๐๐ = 1
๐ ๐ ๏ฟฝ ๐บ๐บ ๐ ๐ , and โโ1 1
๐ ๐ ๏ฟฝ ๐บ๐บ ๐ ๐ = โซ0
๐ ๐ ๐๐ ๐๐ ๐๐๐๐
The step response is the integral of the impulse response
K. Webb ESE 499
78
Transfer Function and Dynamic Response
We just saw that the impulse response is๐๐ ๐ก๐ก = โโ1 ๐บ๐บ ๐ ๐
And the step response is
๐ฆ๐ฆ ๐ก๐ก = โโ11๐ ๐ ๏ฟฝ ๐บ๐บ ๐ ๐
Both are entirely determined by the system transfer function, ๐บ๐บ ๐ ๐ System poles and zeros determine the nature of these
responses ๐บ๐บ ๐ ๐ is a complete mathematical model for the system