Beruflich Dokumente
Kultur Dokumente
Tyrone Vincent
Lecture 4
Contents
1 Motivation 1
3 Quiz Yourself 5
3.1 Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.2 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1 Motivation
Motivation
• Differential equations model dynamic systems
• Control system design requires simple methods for solving these equations!
• Laplace Transforms allow us to
– systematically solve linear time invariant (LTI) equations for arbitrary inputs
– easily combine coupled differential equations
– use block diagrams to represent systems that are made up of smaller subsystems
Transforms
• A transform is a different but equivalent representation for a mathematical object
• Recall that complex numbers can be represented in two ways:
Rectangular Form Polar Form
s = a + jb s = r∠θ
• Given a complex number represented in rectangular form, we can transform it to polar form (and vice-versa)
(a, b) → (r, θ)
1
2 Laplace Transform Definition
Laplace Transforms
• We will introduce the Laplace Transform for functions defined for t > 0.
Let f (t) be a function of time defined for t > 0. The one-sided Laplace Transform of f (t) is defined to be
Z ∞
F (s) := L[f (t)] := f (t)e−st dt
0−
• 0− is shorthand for the lower limit approaching 0 from the left (always negative)
• The Laplace Transform exists if the integral converges for any value of s
– Region of convergence is not as important for “one-sided” Laplace transforms
f (t) = 3
2
– Second limit converges to 1 for all values of s
– First limit converges to zero if the real part of s is less than zero
3
F (s) = [0 − 1]
−s
3
=
s
• Note: the value of the function for negative time does not matter (other than what affects the limit as t → 0.)
• In order to identify a Laplace Transform with a single function, we will pick the function that is equal to zero
for negative time.
L 3
3u(t) ←→
s
where u(t) is the unit step function
1.5
1
Magnitude
(
1 t≥0 0.5
u(t) =
0 otherwise
0
−0.5
−2 −1 0 1 2
Time (s)
L A
Ae−at u(t) ←→
a+s
3
3 Laplace Transform Properties
The Laplace Transform has useful properties that can help us find the Laplace Transform of more complicated func-
tions.
A useful application of the linearity property is to find the Laplace Transform of cos(ωt)u(t) from the sum of
complex exponentials
Example 3. cos(ωt)u(t) can be represented as the sum of complex exponentials
1 jωt
e + e−jωt u(t)
cos(ωt)u(t) =
2
Using the linearity and scaling properties,
1 1
L{cos(ωt)u(t)} = L{ejωt u(t)} + L{e−jωt u(t)}.
2 2
We have already derived the Laplace Transform pair
L 1
e−at u(t) ←→
a+s
Using this, we have
1 1 1 1
L{cos(ωt)u(t)} = +
2 s − jωt 2 s + jωt
s
=
s + ω2
2
By subtracting two complex exponentials, the same method can used to find the Laplace Transform of sine
Ls
cos(ωt)u(t) ←→
s2 + ω 2
L ω
sin(ωt)u(t) ←→ 2
s + ω2
4
Time shift L {f (t − t0 )u(t − t0 )} = e−st0 F (s)
Frequency shift L {e−s0 t f (t)} = F (s + s0 )
Example 4. Use the frequency shift property to find the Laplace Transform of a decaying exponential function
Solution: Since
L s
cos(ωt)u(t) ←→
s2 + ω 2
The frequency shift property implies
−at s
L{e cos(ωt)u(t)} = 2
s + ω 2 s=s+a
s+a
=
(s + a)2 + ω 2
Remember, the notation f (0− ) indicates the limit as we approach 0 from the left, or
The proof of these theorems is actually pretty easy - see the appendix for the differentiation theorem. The differentia-
tion theorem will be used in the next lecture to solve differential equations.
Example 5. (using Integration theorem.) Find the Laplace Transform of the function f (t) = tu(t).
Solution: We can use the integration formula because a ramp is the integral of a step. That is,
Z t
u(τ )dτ = tu(t).
0−
Thus,
1
L {tu(t)} = L {u(t)} ,
s
11 1
= = 2.
ss s
5
4 The Impulse Function
A useful yet tricky signal that finds a lot of use in control systems is the impulse function. It is an approximation to a
signal which lasts for a short period of time - much shorter that the natural time scales of the system. It turns out that
the Laplace Transform of an impulse function is very simple. However, before looking at an impulse function, first
lets consider a pulse:
Example 6. Find the Laplace Transform of the pulse function
1
t0
Magnitude
0 t0
Time (s)
Note that the area of the pulse function is 1 for any value of t0 .
Solution: The pulse function can be represented as a superposition of steps
1 1
pt0 (t) = u(t) − u(t − t0 )
t0 t0
so the Laplace Transform of pt0 (t) is
1 1
L {pt0 (t)} = L u(t) − L u(t − t0 )
t0 t0
1 1 −st0
= − e
t0 s t0 s
1
1 − e−st0
=
t0 s
The impulse function is a pulse that gets shorter and shorter, but whose magnitude gets larger and larger, while
maintaining the same area under the curve.
6
1
t0
Magnitude
0 t0
Time (s)
The impulse function can also be defined via these two properties:
δ(t) = 0 t 6= 0
Z 0+
δ(t)dt = 1
0−
Any function which has these properties (including the limit of the pulse function) has Laplace Transform equal to one
Z ∞
L {δ(t)} = δ(t)e−st dt
0−
Z 0+
= δ(t)e−st dt
0−
Z 0+
−s0
=e δ(t)dt
0−
=1
7
f (t) F (s)
δ(t) 1
1
u(t) s
1
tu(t) s2
1
eat u(t) s+a
s
cos(ωt)u(t) s2 +ω 2
ω
sin(ωt)u(t) s2 +ω 2
−at s+a
e cos(ωt)u(t) (s+a)2 +ω 2
e−at sin(ωt)u(t) ω
(s+a)2 +ω 2
6 Appendix
6.1 Euler’s Formula
The definition of the Laplace Transform includes the function e−st , where s is a complex number. Let’s remember
how to integrate this function. The function ex is very important, as it appears in many places in engineering subjects.
The reason that it is so ubiquitous is that is satisfies the relationship
d x
e = ex
dx
and thus appears in situations when quantities increase in proportion to how much is already there. We want to extend
this function to complex numbers. In this case, we will need to calculate ez when z is complex.
• The key formula extending exponentials to complex numbers is Euler’s Formula:
d jθ
Let’s check that with this definition, ejθ satisfies dθ e = jejθ
d jθ d
e = (cos(θ) + j sin(θ))
dθ dθ
= − sin(θ) + j cos(θ)
= j (cos(θ) + j sin(θ)) [switch terms, pull out j and use j 2 = −1]
= jejθ
dest
Now, considering dt we can write s = a + jb, so that
dest de(a+jb)t
=
dt dt
d(eat ejbt )
=
dt
= aeat ejbt + jbeat ejbt
= (a + jb)eat ejbt
= sest
Thus, taking the derivative of est with respect to time simply brings down the factor s, just as would occur if s were
purely real. This also means that
Z b b
st 1 st
e dt = e
a −s a
8
6.2 Proof of Differentiation Theorem
Z ∞
d df (t) −st
L f (t) = e dt
dt 0 − dt
Choose u = e−st and dv = dfdt(t) dt. Then du = −se−st dt and v = f (t) so that
Z ∞
d ∞
L f (t) = f (t)e−st 0− + f (t)se−st dt
dt 0−
Z ∞
−st
0
− f (0− ) + s f (t)e−st dt
:
= lim f (t)e
t→∞ 0−
−
= sL {f (t)} − f (0 )
7 Quiz Yourself
7.1 Questions
1. Find the one-sided Laplace Transform for the following signals
(
1 0≤t≤1
(a) x(t) =
0 otherwise
(b) x(t) = (1 + e−t )
(c) x(t) = (1 + e−t )u(t)
(d) x(t) = t
2. Using the step function and the integration formula, find the Laplace Transform of the following function
x(t) = t2 u(t)
3. Using the frequency shift property, find the Laplace Transform of the following function
7.2 Solutions
1.
(a)
9
(b)
(c)
10
(d)
2.
11
3.
12