ACM95b/100b Lecture Notes

Caltech 2004

Laplace Transform
Z ∞
L{f (t)} ≡ F (s) ≡ e−st f (t) dt (1)
0
Examples
a) f (t) = 1, t ≥ 0
b
e−st 
Z ∞
−st 1
L{1} = e dt = lim  =
0 b→∞ −s  s
0

b) f (t) = eat , t ≥ 0
Z ∞ Z ∞
−st at 1
at
L{e } = e e dt = e−(s−a)t dt = , s>a
0 0 s−a

Existence
The Laplace transform exists for s > a if:

(1a) f is piecewise continuous on 0 ≤ t ≤ M

(2a) f is of exponential order as t → ∞

(i.e. |f (t)| ≤ Keat , t ≥ M , with K, a, M constant)

Writing Z Z Z
∞ M ∞
e−st f (t) dt = e−st f (t) dt + e−st f (t) dt, (2)
0 0 M

the first integral exists by (1a) and |e−st f (t)| ≤ Ke(a−s)t for t ≥ M so if s > a, the second integral
exists by (2a).

Linearity

L{c1 f (t) + c2 g(t)} = c1 L{f (t)} + c2 L{g(t)} (3)

Transforms of Derivatives
For existence of L{f 0 (t)} for s > a we require

(1b) f is continuous with f 0 piecewise continuous on 0 ≤ t ≤ M

(2b) f is of exponential order as t → ∞.

(i.e. |f (t)| ≤ Keat , t ≥ M , with K, a, M constant)
If the discontintuities in f 0 (t) are located at t1 , t2 , · · · tn then
Z ∞ Z t1 Z t2 Z ∞
0 −st 0
L{f (t)} = e f (t) dt = + +··· + e−st f 0 (t) dt (4)
0 0 t1 tn
| {z }
integrate by parts
( t 2 ∞ )
t 1
Z ∞
−st
 
= e f (t)  +  + · · · +
 
 +s e−st f (t) dt (5)
0 t1 tn 0
| {z }
telescope using (1b)
∞ Z ∞
−st
e−st f (t) dt

= e f (t) +s (6)
0 0
| {z } | {z }
use (2b) and s>a L{f (t)}
= −f (0) + sL{f (t)}, s>a (7)

L{f 0 (t)} = sL{f (t)} − f (0), s>a

For existence of L{f (n) (t)} for s > a we require

(1c) f (t), · · · , f (n−1) (t) is continuous with f (n) piecewise continuous on 0 ≤ t ≤ M

(2c) f (t), · · · , f (n−1) (t) are of exponential order as t → ∞.

(i.e. |f (t)|, · · · , |f (n−1) (t)| ≤ Keat , t ≥ M , with K, a, M constant)

and by induction we obtain

L{f (n) (t)} = sn L{f (t)} − sn−1 f (0) − · · · − sf (n−2) (0) − f (n−1) (0), s>a

Note that if f (t), · · · , f (n−1) (t) vanish at t = 0 then

L{f (n) (t)} = sn L{f (t)} (8)

so differentiation in t space corresponds to multiplication in s space.

Convolution Integrals
Suppose f (t), g(t), and h(t) have transforms F (s), G(s), and H(s) for s > a ≥ 0. If

H(s) = F (s)G(s) (9)

then h(t) is the convolution of f (t) and g(t)

Rt
h(t) = f ∗ g = 0 f (t − τ )g(τ )dτ,
Rt (10)
= g∗f = 0 f (τ )g(t − τ )dτ.

Hence, if H(s) can be expressed as a product of known transforms, h(t) can be expressed as a
convolution integral. To sketch the proof of (10) we start with the product of transforms (9) and
seek to identify the form of h(t).
Z ∞ Z ∞
H(s) = F (s)G(s) = e−sξ f (ξ) dξ · e−sτ g(τ )dτ (11)
0 0
Z ∞ Z ∞ 
= g(τ ) e−s(ξ+τ ) f (ξ) dξ dτ (12)
0 0
(let ξ=t−τ so dξ=dt)
Z ∞ Z ∞ 
= g(τ ) e−st f (t − τ ) dt dτ (13)
0 τ
(reverse order of integration over triangular region)
Z ∞ Z t 
−st
= e f (t − τ )g(τ )dτ dt (14)
0 0
| {z }
L−1 {H(s)} by definition
Z ∞
= e−st h(t) dt (15)
0

Reverse order of integration over triangular region