LECTURE 1: LAPLACE TRANSFORMS – ENGINEERING

PERSPECTIVE

LAPLACE SOLVE TO
REPRESENTATION Y(s) =
COMPLEX
DOMAIN
TIME
DIFFERENTIAL ? SOLUTION
DOMAIN
EQUATIONS Y(t) =

Transforming with Laplace is a means of simplifying the mathematics in solving

differential equations so that we are doing only algebra, not dealing with
differentiation. The “?” transition is often hard and sometimes impossible.

Using an electrical circuit as an example, where input voltage vi(t) generates output
voltage v0(t), we have:

Vi(s) = L{vi(t)} and Vo(s) = L {vo(t)}

The expression G(s) = Vo(s)/Vi(s) is the TRANSFER FUNCTION.

If we can specify G(s) (describe the system in the complex domain) and choose an
input Vi(s) we can find vo(t) as:

vo(t) = L -1{G(s)Vi(s)} where L -1{…} means INVERSE LAPLACE

TRANSFORM

c + j

f(t) = L {F(s)}
-1
= 1/(2πj) 
c − j
F(s)estds, t > 0

Finding the inverse is normally achieved using tables and comparing the form of
G(s).Vi(s) with standard forms.

EXAMPLE FROM LECTURE 0

If all the initial conditions are zero i.e. f(0) = f ’(0) = … = 0 then taking the transform
of a differential equation is like multiplying by s instead of differentiating. E.g.

K θi (t) = A  0 (t) + B  0 (t) + C θ0 (t) becomes

K θi(s) = (As2 + Bs + C) θ0(s) and the transfer function is

1
 𝜃𝑜 (𝑠) 𝐾
= 𝐺(𝑠) = 2
𝜃𝑖 (𝑠) 𝐴𝑠 + 𝐵𝑠 + 𝐶

PARTIAL FRACTIONS

Examples:

1. What is the unit step response from a system where the transfer function is:

G(s) = 5(1 + 0.4s)/[(s + 1)(0.2s + 1)] = (10s + 25)/[(s + 1)(s + 5)]

𝑉0 (𝑠)
𝐺(𝑠) =
𝑉𝑖 (𝑠)

The output is Vo(s) and the input is: Vi(s) = L {u(t)}=1/s

Vo = G(s) Vi(s) = (10s + 25)/[s(s + 1)(s + 5)]

= A/s + B/(s + 1) + C/(s + 5)

10s + 25 = A(s + 1)(s + 5) + Bs(s + 5) + Cs(s + 1)

Make s = 0 25 = 5A A=5

s = -1 15 = -4B B = -3.75

s = -5 -25 = 20C C = -1.25

Vo(s) = 5/s – 3.75/(s + 1) – 1.25/(s + 5)

v(t) = (5 – 3.75e-t – 1.25e-5t) u(t)

2. Calculate the response to a decaying exponential input vi(t)= e-t of a system

whose transfer function is

G(s) = 3(s2 + 9s + 18)/(s + 4)(s + 2)

We know Vi(s) = L {e-t} = 1/(s + 1)

Vo(s) = G(s)Vi(s)

Vo(s) = 3(s2 + 9s + 18)/[(s+1)(s + 4)(s + 2)]

Into partial fractions:

3(s2 + 9s + 18) = A(s + 4)(s + 2) + B(s + 1)(s + 2) + C(s + 1)(s + 4)

s = -1 30 = 3A A = 10
2
 s = -4 -6 = 6B B = -1

s = -2 12 = -2C C = -6

V(s) = 10/(s + 1) – 1/(s + 4) – 6/(s + 2)

vo(t) = (10e-t – e-4t – 6e-2t) u(t)

1
3. If 𝑌(𝑠) = 𝑠2 +2𝑠+2 , find y(t).

Take Case 22 from the table:

𝜔2 𝜔𝑛
𝑌(𝑠) = 𝑠2 +2𝜉𝜔𝑛 𝑠+𝜔2 , --> 𝑓(𝑡) = 𝑒 −𝜉𝜔𝑛𝑡 𝑠𝑖𝑛(𝜔𝑛 √1 − 𝜉 2 𝑡)
𝑛 𝑛 √1−𝜉 2

Make sure that all coefficients in Y(s) and F(s) match each other.
1 2
Numerator: 𝑌(𝑠) = 2 𝑠2 +2𝑠+2

1 1 2
Denominator: 𝜔𝑛 = √2 , 𝜉 = , therefore 𝑌(𝑠) = 2 1
√2 𝑠2 +2 √2𝑠+(√2)2
√2

1 𝜔𝑛
Now 𝑦(𝑡) = 2 𝑒 −𝜉𝜔𝑛𝑡 𝑠𝑖𝑛(𝜔𝑛 √1 − 𝜉 2 𝑡) = 𝑒 −𝑡 sin (𝑡)
√1−𝜉 2

IMPORTANT PARAMETERS OF TRANSFER FUNCTIONS

Take the general case of transfer function:

𝐾(𝑎𝑛 𝑠 𝑛 + 𝑎𝑛−1 𝑠 𝑛−1 ⋯ ⋯ + 𝑎0 )

𝐺(𝑠) = ℓ
𝑠 (𝑏𝑚 𝑠 𝑚 + 𝑏𝑚−1 𝑠 𝑚−1 ⋯ ⋯ + 𝑏0 )

The transfer function can be re-written it its

• Normalised form when bm=1, or
• Non-dimensional form when b0=1

From the normalised form we can find poles (roots of the denominator) and zeros
𝐴
(roots of the numerator). Take 𝐺(𝑠) = (𝑠+𝑎)(𝑠+𝑏) as an example:
Poles are –a and –b; this transfer function has no zeros.

From the non-dimensional form we can also deduce a few properties. Non-
𝐴⁄
𝑎𝑏
dimensionally 𝐺(𝑠) = 𝑠 𝑠 . From this form, we can pick at the following:
( +1)( +1)
𝑎 𝑏

𝐴⁄
𝑎𝑏 is the steady state gain (or DC gain) of the system (the ratio of the input and the
1 1
output as time approaches infinity). 𝑎 , 𝑏 are the time constants of the system (the

3
coefficients that will appear in the argument of the exponential terms in the time
domain answers).

To gauge the relevance of the steady state gain, let us use the FINAL VALUE
THEOREM to find this gain, i.e. when s→ 0 𝑜𝑟 𝑡 → ∞. We must specify what the
1
input is first, e.g. input = 𝑆 , a unit step.

1
lim 𝑢(𝑡) = lim 𝑠 𝑠 = 1 (input at infinity)
𝑡→∞ 𝑠→0
1 𝐴 1 𝐴
∴ lim 𝑠 𝐺(𝑠) ∙ 𝑠 = lim (𝑠 ∙ (𝑠+𝑎)(𝑠+𝑏) ∙ 𝑠 ) = 𝑎𝑏
𝑠→𝑜 𝑠→0
𝑄𝐸𝐷.