ENGIN 211, Engineering Mathpchen/engin211/Laplace Transforms.pdf · Laplace Transforms ENGIN 211,...

Post on 14-Sep-2020

8 views 1 download

Transcript of ENGIN 211, Engineering Mathpchen/engin211/Laplace Transforms.pdf · Laplace Transforms ENGIN 211,...

1

Laplace Transforms

ENGIN 211, Engineering Math

Why Laplace Transform ?

• Laplace transform converts a function in the time domain to its frequency domain.

• It is a powerful, systematic method in solving differential equations.

• It converts differential equations into algebraic equations.

• Initial or boundary conditions are automatically accounted for.

• In situations where other methods fail because the solution is discontinuous, it can succeed.

2

Definition

ℒ 𝑓 𝑡 = 𝑒−𝑠𝑡𝑓 𝑡 𝑑𝑡∞

0

= 𝐹 𝑠

The parameter 𝑠 known as the complex

frequency is assumed to have positive and large

enough real part to ensure that the integral

converges.

3

Useful Transforms

• 𝑓 𝑡 = 𝑎 → 𝐹 𝑠 =𝑎

𝑠

• 𝑓 𝑡 = 𝑒𝑎𝑡 → 𝐹 𝑠 =1

𝑠−𝑎

• 𝑓 𝑡 = sin 𝜔𝑡 → 𝐹 𝑠 =𝜔

𝑠2+𝜔2

• 𝑓 𝑡 = cos 𝜔𝑡 → 𝐹 𝑠 =𝑠

𝑠2+𝜔2

• 𝑓 𝑡 = 𝑡𝑛 → 𝐹 𝑠 =𝑛!

𝑠𝑛+1

• 𝑓 𝑡 = sinh 𝜔𝑡 → 𝐹 𝑠 =𝜔

𝑠2−𝜔2

• 𝑓 𝑡 = cosh 𝜔𝑡 → 𝐹 𝑠 =𝑠

𝑠2−𝜔2

4

The Unit Impulse δ(t)

5

The unit impulse function is defined as

𝛿(𝑡) =𝑑𝑢(𝑡)

𝑑𝑡

where 𝑢(𝑡) is the unit step function.

Total area underneath 𝛿(𝑡): 𝛿(𝑡 − 𝑡0)𝑑𝑡∞

−∞= 1

Sifting property of 𝛿(𝑡): 𝑓(𝑡)𝛿(𝑡 − 𝑡0)𝑑𝑡∞

−∞= 𝑓(𝑡0)

Laplace transform: ℒ 𝛿(𝑡) = 1, ℒ 𝛿(𝑡 − 𝑡0) = 𝑒−𝑠𝑡0

𝛿(𝑡 − 𝑡0)

Properties

Linearity

ℒ 𝑓 𝑡 ± 𝑔 𝑡 = ℒ 𝑓 𝑡 ± ℒ 𝑔 𝑡

ℒ 𝑘𝑓 𝑡 = 𝑘ℒ 𝑓 𝑡

But ℒ 𝑓 𝑡 𝑔 𝑡 ≠ ℒ 𝑓 𝑡 ℒ 𝑔 𝑡

Rather the convolution of two function,

ℒ 𝑓 𝑡 ∗ 𝑔 𝑡 = ℒ 𝑓 𝑡 ℒ 𝑔 𝑡

6

Examples

•ℒ 2 sin 3𝑡 + 4 sinh 3𝑡 = 2

3

𝑠2+9+ 4

3

𝑠2−9

=18 𝑠2+3

𝑠4−81

•ℒ 5e4𝑡 + cosh 2𝑡 =

5

𝑠−4+

𝑠

𝑠2−4

=6𝑠2−4𝑠−20

𝑠−4 𝑠4−4

•ℒ 𝑡3 + 2𝑡2 − 4𝑡 + 1 =

3!

𝑠4 + 22!

𝑠3 − 41

𝑠2 +1

𝑠

=1

𝑠4 𝑠3 − 4𝑠2 + 4𝑠 + 6

7

Main Theorems

Assume ℒ 𝑓 𝑡 = 𝐹 𝑠 :

• Similarity theorem ℒ 𝑓 𝑎𝑡 =1

𝑎𝐹

𝑠

𝑎

• Real Shifting theorem: for 𝑎 > 0

ℒ 𝑓 𝑡 − 𝑎 = 𝑒−𝑎𝑠𝐹 𝑠

• Complex Shifting theorem: for 𝑎 > 0

ℒ 𝑒−𝑎𝑡𝑓 𝑡 = 𝐹 𝑠 + 𝑎

8

Main Theorems (Cont’d)

• Derivative theorem:

ℒ𝑑𝑓 𝑡

𝑑𝑡= 𝑠𝐹 𝑠 − 𝑓 0 +

ℒ𝑑2𝑓 𝑡

𝑑𝑡2= 𝑠2𝐹 𝑠 − 𝑠𝑓 0 + − 𝑓′ 0 +

• Complex differentiation theorem:

ℒ 𝑡𝑛𝑓 𝑡 = −1 𝑛𝑑𝑛𝐹 𝑠

𝑑𝑠𝑛

9

Main Theorems (Cont’d)

• Real integral theorem:

ℒ 𝑓 𝜏𝑡

0

𝑑𝜏 =𝐹 𝑠

𝑠

• Complex integral theorem:

ℒ𝑓 𝑡

𝑡= 𝐹 𝜎

𝑠

𝑑𝜎 if lim𝑡→0

𝑓 𝑡

𝑡exists

• Convolution theorem: if 𝑓1 ∗ 𝑓2 = 𝑓1 𝜏𝑡

0𝑓2 𝑡 − 𝜏 𝑑𝜏

ℒ 𝑓1 ∗ 𝑓2 = 𝐹1 𝑠 ∗ 𝐹2 𝑠

10

Main Theorems (Cont’d)

Convolution theorem:

𝑓1 ∗ 𝑓2 = 𝑓1 𝜏𝑡

0

𝑓2 𝑡 − 𝜏 𝑑𝜏

ℒ 𝑓1 ∗ 𝑓2 = 𝐹1 𝑠 ∗ 𝐹2 𝑠

Initial theorem:

𝑓 0 + = lim𝑠→∞

𝑠𝐹 𝑠 if lim𝑡→0

𝑓 𝑡 exists

Final theorem:

𝑓 ∞ = lim𝑠→0

𝑠𝐹 𝑠 if lim𝑡→∞

𝑓 𝑡 exists

11

Example (Shifting theorem)

The real shifting theorem states that if ℒ 𝑓 𝑡 = 𝐹 𝑠 , then

ℒ 𝑒−𝑎𝑡𝑓 𝑡 = 𝐹 𝑠 + 𝑎

• Because ℒ sin 3𝑡 =3

𝑠2+9, then

ℒ 𝑒−2𝑡 sin 3𝑡 =3

𝑠 + 2 2 + 9=

3

𝑠2 + 4𝑠 + 13

• Because ℒ 𝑡3 =3!

𝑠4, then

ℒ 𝑡3𝑒−4𝑡 =3!

𝑠 + 4 4=

6

𝑠 + 4 4

12

Example (Complex Differentiation)

Complex differentiation theorem:

ℒ 𝑡𝑛𝑓 𝑡 = −1 𝑛𝑑𝑛𝐹 𝑠

𝑑𝑠𝑛

If ℒ cosh 3𝑡 =𝑠

𝑠2−9, then

ℒ 𝑡2 cosh 3𝑡 =𝑑2

𝑑𝑠2

𝑠

𝑠2 − 9=

2𝑠 𝑠2 + 27

𝑠2 − 9 3

13

Example (Complex Integral)

• ℒ𝑓 𝑡

𝑡= 𝐹 𝜎

𝑠𝑑𝜎 if lim

𝑡→0

𝑓 𝑡

𝑡exists

Example: find ℒ1−cos 2𝑡

𝑡=?

Since lim𝑡→0

1−cos 2𝑡

𝑡= lim

𝑡→0

2 sin2 𝑡

𝑡= 0 does exist, we obtain

ℒ1 − cos 2𝑡

𝑡=

1

𝜎−

𝜎

𝜎2 + 4𝑑𝜎

𝑠

= ln𝜎 −1

2ln 𝜎2 + 4

𝑠

= ln𝜎

𝜎2 + 4 𝑠

= 0 − ln𝑠

𝑠2 + 4= ln

𝑠2 + 4

𝑠

14

Inverse Laplace Transform

• Given Laplace transform 𝐹 𝑠 in 𝑠-domain, find its function

𝑓 𝑡 in 𝑡-domain.

ℒ−1 𝐹 𝑠 = 𝑓(𝑡)

• Useful pairs

15

𝐹 𝑠 =𝑎

𝑠→ 𝑓 𝑡 = 𝑎

𝐹 𝑠 =1

𝑠 − 𝑎→ 𝑓 𝑡 = 𝑒𝑎𝑡

𝐹 𝑠 =𝜔

𝑠2 + 𝜔2→ 𝑓 𝑡 = sin𝜔𝑡

𝐹 𝑠 =𝑠

𝑠2 + 𝜔2→ 𝑓 𝑡 = cos𝜔𝑡

𝐹 𝑠 =𝑛!

𝑠𝑛+1→ 𝑓 𝑡 = 𝑡𝑛

𝐹 𝑠 =𝜔

𝑠2 − 𝜔2→ 𝑓 𝑡 = sinh𝜔𝑡

𝐹 𝑠 =𝑠

𝑠2 − 𝜔2→ 𝑓 𝑡 = cosh𝜔𝑡

Example

16

Find ℒ−1 4𝑠2−5𝑠+6

𝑠+1 𝑠2+4=?

Solution: 4𝑠2−5𝑠+6

𝑠+1 𝑠2+4=

𝐴

𝑠+1+

𝐵𝑠+𝐶

𝑠2+4

4𝑠2 − 5𝑠 + 6 = 𝐴 𝑠2 + 4 + 𝐵𝑠 + 𝐶 𝑠 + 1

• 𝑠 + 1 = 0 → 𝐴 = 3

• Equating coefficients of various powers 𝐵 = 1, and 𝐶 = −6

ℒ−13

𝑠 + 1+

𝑠 − 6

𝑠2 + 4= ℒ−1

3

𝑠 + 1+

𝑠

𝑠2 + 4− 3

2

𝑠2 + 4

= 3𝑒−𝑡 + cos 2𝑡 − 3 sin 2𝑡

Partial Fractions

• Order in numerator less than denominator

• Factorize the denominator into prime factors

• A linear factor 𝑠 + 𝑎 gives 𝐴

𝑠+𝑎

• A repeated factor 𝑠 + 𝑎 2 gives 𝐴

𝑠+𝑎+

𝐵

𝑠+𝑎 2

• A quadratic factor 𝑠2 + 𝑝𝑠 + 𝑞 gives 𝑃𝑠+𝑄

𝑠2+𝑝𝑠+𝑞

17

Example (Partial Fractions) 𝑠2 − 9𝑠 + 7

𝑠2 + 𝑠 − 2 𝑠2 + 2𝑠 + 2=

𝐴

𝑠 + 2+

𝐵

𝑠 − 1+

𝐶𝑠 + 𝐷

𝑠2 + 2𝑠 + 2

• 𝐴 =𝑠2−9𝑠+7

𝑠−1 𝑠2+2𝑠+2 𝑠=−2

= −4 and 𝐵 =𝑠2−9𝑠+7

𝑠+2 𝑠2+2𝑠+2 𝑠=1

= −2

5

• 𝑠2 − 9𝑠 + 7 = 𝐴 𝑠 − 1 𝑠2 + 2𝑠 + 2 + 𝐵 𝑠 + 2 𝑠2 + 2𝑠 + 2 + 𝐶𝑠 + 𝐷 𝑠2 + 𝑠 − 2

• Equating highest-order 𝑠3 coefficient 𝐴 + 𝐵 + 𝐶 = 0 → 𝐶 =22

5

• Equating lowest-order 𝑠0 coefficient −2𝐴 + 4𝐵 − 2𝐷 = 0 → 𝐷 =16

5

ℒ−1𝐴

𝑠 + 2+

𝐵

𝑠 − 1+

𝐶𝑠 + 𝐷

𝑠2 + 2𝑠 + 2= ℒ−1

−4

𝑠 + 2−

2

5

1

𝑠 − 1+

2

5

11𝑠 + 8

𝑠 + 1 2 + 1

= ℒ−1−4

𝑠 + 2−

2

5

1

𝑠 − 1+

2

5

11 𝑠 + 1 − 3

𝑠 + 1 2 + 1

= ℒ−1−4

𝑠 + 2−

2

5

1

𝑠 − 1+

22

5

𝑠 + 1

𝑠 + 1 2 + 1−

6

5

1

𝑠 + 1 2 + 1

= −4𝑒−2𝑡 −2

5𝑒𝑡 +

22

5𝑒−𝑡 sin 𝑡 −

6

5𝑒−𝑡 cos 𝑡

18

Applying to Differential Equations

With the use of Derivative theorems:

ℒ𝑑𝑓 𝑡

𝑑𝑡= 𝑠𝐹 𝑠 − 𝑓 0 +

ℒ𝑑2𝑓 𝑡

𝑑𝑡2= 𝑠2𝐹 𝑠 − 𝑠𝑓 0 + − 𝑓′ 0 +

We can convert differential equations into algebraic equations

where all initial conditions are taken into account. Effectively, we

are converting problems in 𝑡-domain into 𝑠-domain. In the end,

we need to perform inverse transform to go back to 𝑡-domain.

19

1st Order Differential Equation

Example: solve 𝑑𝑥

𝑑𝑡+ 4𝑥 = 2𝑒−2𝑡 + 4𝑒−4𝑡 , at 𝑡 = 0, 𝑥 = 2.

Taking Laplace on both sides to s-domain:

𝑠𝑋 − 2 + 4𝑋 =2

𝑠 + 2+

4

𝑠 + 4

Then 𝑋 =2

𝑠+4+

2

𝑠+2 𝑠+4+

4

𝑠+4 2 =2

𝑠+4+

1

𝑠+2−

1

𝑠+4+

4

𝑠+4 2

𝑋 =1

𝑠 + 2+

1

𝑠 + 4+

4

𝑠 + 4 2

Inverse Laplace back to t-domain

𝑥 𝑡 = ℒ−1 𝑋 = 𝑒−2𝑡 + 𝑒−4𝑡 + 4𝑡𝑒−4𝑡

20

2nd Order Differential Equation Example: 𝒙" + 𝟓𝒙′ + 𝟔𝒙 = 𝟐𝟒 𝐬𝐢𝐧 𝟐𝒕 , 𝐚𝐭 𝒕 = 𝟎, 𝒙 = 𝟎 𝐚𝐧𝐝 𝒙′ = 𝟎

Taking Laplace transform on both sides:

𝑠2𝑋 − 𝑠𝑥 0 − 𝑥′ 0 + 5 𝑠𝑋 − 𝑥(0) + 6𝑋 = 242

𝑠2 + 4

Plug in initial conditions:

𝑠2 + 5𝑠 + 6 𝑋 =48

𝑠2 + 4

So 𝑋 =48

𝑠2+5𝑠+6 𝑠2+4=

𝐴

𝑠+2+

𝐵

𝑠+3+

𝐶𝑠+𝐷

𝑠2+4

where 𝐴 =48

(𝑠+3) 𝑠2+4 𝑠=−2

= 6, 𝐵 =48

(𝑠+2) 𝑠2+4 𝑠=−3

= −48

13

And

𝐴 𝑠 + 3 𝑠2 + 4 + 𝐵 𝑠 + 2 𝑠2 + 4 + 𝐶𝑠 + 𝐷 𝑠 + 2 𝑠 + 3 = 48

21

2nd Order Differential Equation (Cont’d)

Equating the coefficients for 𝑠3 and 𝑠0, we obtain

𝐴 + 𝐵 + 𝐶 = 0, and 12𝐴 + 8𝐵 + 6𝐷 = 48

Thus,

𝐶 = − 𝐴 + 𝐵 = −30

13, and 𝐷 =

24 − 6𝐴 − 4𝐵

3=

12

13

Taking inverse Laplace:

𝑥 𝑡 = ℒ−1 𝑋 = ℒ−1𝐴

𝑠 + 2+

𝐵

𝑠 + 3+

𝐶𝑠 + 𝐷

𝑠2 + 4

= 6ℒ−11

𝑠 + 2−

48

13ℒ−1

1

𝑠 + 3−

30

13ℒ−1

𝑠

𝑠2 + 4+

6

13 ℒ−1

2

𝑠2 + 4

= 6𝑒−2𝑡 −48

13𝑒−3𝑡 −

30

13cos 2𝑡 +

6

13sin 2𝑡

22

System of Differential Equations

𝑥" + 2𝑥 − 𝑦 = 0𝑦" + 2𝑦 − 𝑥 = 0

, with initial conditions 𝑥 0 = 4𝑦 0 = 2

and 𝑥′ 0 = 0

𝑦′ 0 = 0

Taking Laplace transforms using ℒ 𝑓" = 𝑠2𝐹 𝑠 − 𝑠𝑓 0 − 𝑓′ 0

𝑠2𝑋 − 𝑠𝑥 0 − 𝑥′ 0 + 2𝑋 − 𝑌 = 0

𝑠2𝑌 − 𝑠𝑦 0 − 𝑦′ 0 + 2𝑌 − 𝑋 = 0 or

𝑠2 + 2 𝑋 − 𝑌 = 4𝑠

−𝑋 + 𝑠2 + 2 𝑌 = 2𝑠

In matrix form

𝑠2 + 2 −1−1 𝑠2 + 2

𝑋𝑌

=4𝑠2𝑠

Using inverse matrix, we obtain

𝑋𝑌

=1

𝑠2 + 2 2 − 1𝑠2 + 2 1

1 𝑠2 + 24𝑠2𝑠

=2𝑠

𝑠2 + 2 2 − 12𝑠2 + 5𝑠2 + 4

Thus

𝑋 =2𝑠 2𝑠2+5

𝑠2+2 2−1=

2𝑠 2𝑠2+5

𝑠2+1 𝑠2+3, and 𝑌 =

2𝑠 𝑠2+4

𝑠2+2 2−1=

2𝑠 𝑠2+4

𝑠2+1 𝑠2+3

23

System of Differential Equations (Cont’d)

Now let’s find the inverse Laplace of 𝑋,

𝑋 =2𝑠 2𝑠2 + 5

𝑠2 + 1 𝑠2 + 3=

𝐴1𝑠 + 𝐵1

𝑠2 + 1+

𝐶1𝑠 + 𝐷1

𝑠2 + 3

(a) Multiply 𝑠2 + 1 and then let 𝑠2 = −1,

𝐴1𝑠 + 𝐵1 =2𝑠 2𝑠2 + 5

𝑠2 + 3 𝑠2=−1

= 3𝑠

Thus, 𝐴1 = 3, and 𝐵1 = 0

(b) Multiply 𝑠2 + 3 and then let 𝑠2 = −3,

𝐶1𝑠 + 𝐷1 =2𝑠 2𝑠2 + 5

𝑠2 + 1 𝑠2=−3

= 𝑠

Thus, 𝐶1 = 1, and 𝐷1 = 0

So we have 𝑋 =3𝑠

𝑠2+1+

𝑠

𝑠2+3

Inverse Laplace: 𝑥 𝑡 = ℒ−1 𝑋 = 3 cos 𝑡 + cos 3𝑡

24

System of Differential Equations (Cont’d)

Next let’s find the inverse Laplace of 𝑌,

𝑌 =2𝑠 𝑠2 + 4

𝑠2 + 1 𝑠2 + 3=

𝐴2𝑠 + 𝐵2

𝑠2 + 1+

𝐶2𝑠 + 𝐷2

𝑠2 + 3

(a) Multiply 𝑠2 + 1 and then let 𝑠2 = −1,

𝐴2𝑠 + 𝐵2 =2𝑠 𝑠2 + 4

𝑠2 + 3 𝑠2=−1

= 3𝑠

Thus, 𝐴2 = 3, and 𝐵2 = 0

(b) Multiply 𝑠2 + 3 and then let 𝑠2 = −3,

𝐶2𝑠 + 𝐷2 =2𝑠 𝑠2 + 4

𝑠2 + 1 𝑠2=−3

= −𝑠

Thus 𝐶2 = −1, and 𝐷2 = 0

We have 𝑌 =3𝑠

𝑠2+1−

𝑠

𝑠2+3

Inverse Laplace: 𝑦 𝑡 = ℒ−1 𝑌 = 3 cos 𝑡 − cos 3𝑡

25

Solution: 𝑥 = 3 cos 𝑡 + cos 3𝑡

𝑦 = 3 cos 𝑡 − cos 3𝑡

Impulse Response

26

Network

Network

Network

Network

Convolution and Laplace

27

If we define the impulse response as h(t), then the output y(t) is related to the input x(t) via the convolution integral:

Convolution in the time domain is multiplication in the frequency domain:

Convolution Theorem for Inverse Laplace

This theorem can be used to find inverse Laplace transform ℒ−1 1

𝑠2 𝑠−3=?

𝐹1 𝑠 =1

𝑠2, and 𝐹2 𝑠 =

1

𝑠 − 3

Their inverse: 𝑓1 𝑡 = 𝑡, and 𝑓2 𝑡 = 𝑒3𝑡

Using Convolution Theorem

ℒ−11

𝑠2 𝑠 − 3= 𝑥𝑒3(𝑡−𝑥)𝑑𝑥

𝑡

0

= 𝑒3𝑡 𝑥𝑒−3𝑥𝑑𝑥𝑡

0

= 𝑒3𝑡 𝑥𝑒−3𝑥𝑑𝑥𝑡

0

=1

9𝑒3𝑡 − 3𝑡 − 1

28

Summary

Key points:

Useful Laplace transform pairs

δ-function and its properties

Laplace transform of derivatives

Inverse Laplace using partial fractions

Solving differential equations with Laplace

Supplemental:

Convolution theorem

29