Você está na página 1de 45

MATH283 Advanced Engineering Mathematics

Dr Michelle Dunbar
SMART / School of Mathematics and Applied Statistics The University of Wollongong

May 8, 2013

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

1/22

Review

Last lecture we looked at two shift theorems for nding Laplace and inverse Laplace Transforms of dierent functions. These were helpful for nding the following Laplace transforms: L eat f (t) , L [f (t a)H (t a)]

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

2/22

Review

Last lecture we looked at two shift theorems for nding Laplace and inverse Laplace Transforms of dierent functions. These were helpful for nding the following Laplace transforms: L eat f (t) , L [f (t a)H (t a)]

In this lecture, we will introduce the concept of convolution and investigate some important derivative and products of transforms.

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

2/22

Review

Last lecture we looked at two shift theorems for nding Laplace and inverse Laplace Transforms of dierent functions. These were helpful for nding the following Laplace transforms: L eat f (t) , L [f (t a)H (t a)]

In this lecture, we will introduce the concept of convolution and investigate some important derivative and products of transforms. These tools will be very handy!

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

2/22

Products of Transforms
Motivation:

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

3/22

Products of Transforms
Motivation: Given F (s ) = s (s2 + 1)2 s 1 = 2 s + 1 s2 + 1

= L [sin t] L [cos t] , it is natural to ask the question:

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

3/22

Products of Transforms
Motivation: Given F (s ) = s (s2 + 1)2 s 1 = 2 s + 1 s2 + 1

= L [sin t] L [cos t] , it is natural to ask the question: - Is f (t) = L1 [F (s)] simply the product of sin t and cos t?

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

3/22

Products of Transforms
Motivation: Given F (s ) = s (s2 + 1)2 s 1 = 2 s + 1 s2 + 1

= L [sin t] L [cos t] , it is natural to ask the question: - Is f (t) = L1 [F (s)] simply the product of sin t and cos t? - i.e. Does it hold that: L [sin t cos t] = L [sin t] L [cos t] ....?

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

3/22

Products of Transforms
Motivation: Given F (s ) = s (s2 + 1)2 s 1 = 2 s + 1 s2 + 1

= L [sin t] L [cos t] , it is natural to ask the question: - Is f (t) = L1 [F (s)] simply the product of sin t and cos t? - i.e. Does it hold that: L [sin t cos t] = L [sin t] L [cos t] ....? Answer: No! This is not true. (Why?)
Michelle Dunbar, Week 9, Lecture 1 MATH283 Advanced Engineering Mathematics 3/22

Products of Transforms
Why is this not true? Since:

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

4/22

Products of Transforms
Why is this not true? Since: L [sin t cos t] = L 1 sin 2t 2 1 = 2 s +4 = L [sin t] L [cos t] = s (s2 + 1)2

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

4/22

Products of Transforms
Why is this not true? Since: L [sin t cos t] = L 1 sin 2t 2 1 = 2 s +4 = L [sin t] L [cos t] = Therefore, L [sin t cos t] = L [sin t] L [cos t] , that is, the Laplace Transform of the product is NOT the product of their transforms. s (s2 + 1)2

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

4/22

Products of Transforms
Why is this not true? Since: L [sin t cos t] = L 1 sin 2t 2 1 = 2 s +4 = L [sin t] L [cos t] = Therefore, L [sin t cos t] = L [sin t] L [cos t] , that is, the Laplace Transform of the product is NOT the product of their transforms. So, what is the connection? - The answer is the concept of convolution!. s (s2 + 1)2

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

4/22

Products of Transforms
Denition (Convolution)
If f and g are functions dened on [0, ), then the convolution, denoted (f g ) of f with g , is the function dened for t 0 as follows:
t

(f g )(t) =
0

f (t u)g (u) du ,

which satises the property: L [(f g )(t)] = F (s) G(s) .

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

5/22

Products of Transforms
Denition (Convolution)
If f and g are functions dened on [0, ), then the convolution, denoted (f g ) of f with g , is the function dened for t 0 as follows:
t

(f g )(t) =
0

f (t u)g (u) du ,

which satises the property: L [(f g )(t)] = F (s) G(s) . One can also show, via the substitution ( = t u) that the convolution is commutative, that is: (f g )(t) = (g f )(t).

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

5/22

Products of Transforms
Example (1)
Find the convolution of sin t and cos t. That is, nd (sin t cos t). Solution:
By denition: (f g )(t) =
0 t t

f (t u)g (u) du, cos(t u) sin u du.


0

so

(cos t sin t) =

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

6/22

Products of Transforms
Example (1)
Find the convolution of sin t and cos t. That is, nd (sin t cos t). Solution:
By denition: (f g )(t) =
0 t t

f (t u)g (u) du, cos(t u) sin u du.


0

so

(cos t sin t) =

Using the trigonometric identity: sin() cos( ) = 1 [sin( + ) + sin( )] , 2

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

6/22

Products of Transforms
Example (1)
Find the convolution of sin t and cos t. That is, nd (sin t cos t). Solution:
By denition: (f g )(t) =
0 t t

f (t u)g (u) du, cos(t u) sin u du.


0

so

(cos t sin t) =

Using the trigonometric identity: sin() cos( ) = 1 [sin( + ) + sin( )] , 2

with = u and = t u, we obtain:


t

(cos t sin t) =
0

1 [sin(t) + sin(2u t)] du 2


6/22

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

Products of Transforms
Example (1 contd)
t

(cos t sin t) =
0

1 [sin(t) + sin(2u t)] du 2


t

1 1 u sin t cos(2u t) 2 2 t sin t = 2 =

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

7/22

Products of Transforms
Example (1 contd)
t

(cos t sin t) =
0

1 [sin(t) + sin(2u t)] du 2


t

1 1 u sin t cos(2u t) 2 2 t sin t = 2 = That is, (sin t) (cos t) = 1 t sin t. 2

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

7/22

Products of Transforms
Example (1 contd)
t

(cos t sin t) =
0

1 [sin(t) + sin(2u t)] du 2


t

1 1 u sin t cos(2u t) 2 2 t sin t = 2 = That is, (sin t) (cos t) = 1 t sin t. 2

From the table of Laplace Transforms, we nd that L 1 s . t sin t = 2 2 (s + 1)2


MATH283 Advanced Engineering Mathematics 7/22

Michelle Dunbar, Week 9, Lecture 1

Products of Transforms

Example (1 contd)
Recall that we found earlier L [sin t] L [cos t] = s . (s2 + 1)2

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

8/22

Products of Transforms

Example (1 contd)
Recall that we found earlier L [sin t] L [cos t] = So indeed we have: L [(sin t) (cos t)] = L [sin t] L [cos t] , and the product of Laplace Transforms equals the transformation of their convolution. s . (s2 + 1)2

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

8/22

The Convolution Theorem

Theorem (The Convolution Theorem)


If the convolution, f (t) g (t) is dened, and F (s) = L [f (t)] and G(s) = L [g (t)], then: L [f (t) g (t)] = L [f (t)] L [g (t)] , and L1 [F (s) G(s)] = f (t) g (t) .

N.B. The inverse version of the convolution theorem provides a convenient alternative to using partial fractions for nding inverse transforms.

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

9/22

The Convolution Theorem


Example (2)
Compute L1
1 s(s4)2

. Solution:
1 s

We begin by writing F (s) =

and G(s) =

1 . (s4)2

Taking the inverse transform of F (s) and G(s) we obtain: f (t) = L 1 Hence, 1 =1 s and g (t) = L 1 1 (s 4)2 = te4t .

f (t) g (t) = 1 te4t


t

=
0

1 ue4u du ue4u 4
t

1 4

e4u du
0

1 4t 1 1 e + = te4t 4 16 16
Michelle Dunbar, Week 9, Lecture 1 MATH283 Advanced Engineering Mathematics 10/22

The Convolution Theorem

Example (2 contd)
Therefore: L 1 1 s(s 4)2 = L 1 1 1 s (s 4)2 (By the convolution theorem)

= f (t) g (t)

1 1 4t 1 = te4t e + . 4 16 16

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

11/22

The Convolution Theorem


Example (3)
Compute L1 Solution:
We begin by writing F (s) =
1 s2

1 s2 (s2 +9)

and G(s) =

1 . (s2 +9)

Taking the inverse transform of F (s) and G(s) we obtain: f (t) = L 1 Hence, f (t) g (t) = t 1 3 1 = 9 =
Michelle Dunbar, Week 9, Lecture 1

1 s2

=t

and

g (t) = L 1

1 1 = sin(3t). (s2 + 9) 3

1 sin(3t) 3
t

(t u) sin(3u) du
0

(integrate by parts)

1 sin(3t) . 3
MATH283 Advanced Engineering Mathematics 12/22

The Convolution Theorem

Example (3 contd)
Therefore: L 1 1 1 1 = L 1 2 2 s2 (s2 + 9) s (s + 9) = f (t) g (t) = 1 9 t (By the convolution theorem)

1 sin(3t) 3

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

13/22

The Convolution Theorem


Example (4)
Determine f (t) such that
t

f (t) = 2t2 +
0

f (t u)eu du.

Solution:

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

14/22

The Convolution Theorem


Example (4)
Determine f (t) such that
t

f (t) = 2t2 +
0

f (t u)eu du.

Solution:
We see that the integral on the right is the convolution f (t) et .

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

14/22

The Convolution Theorem


Example (4)
Determine f (t) such that
t

f (t) = 2t2 +
0

f (t u)eu du.

Solution:
We see that the integral on the right is the convolution f (t) et . Therefore the equation has the form: f (t) = 2 t2 + f (t) e t .

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

14/22

The Convolution Theorem


Example (4)
Determine f (t) such that
t

f (t) = 2t2 +
0

f (t u)eu du.

Solution:
We see that the integral on the right is the convolution f (t) et . Therefore the equation has the form: f (t) = 2 t2 + f (t) e t . Taking the Laplace transform of this equation and applying the convolution theorem yields: L [f (t)] = 2L t2 + L f (t) et F (s ) =
Michelle Dunbar, Week 9, Lecture 1

1 4 + F (s ) s3 s+1
14/22

MATH283 Advanced Engineering Mathematics

The Convolution Theorem


Example (4 contd)
Re-arranging we obtain: F (s ) 1 1 4 = 3. s+1 s

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

15/22

The Convolution Theorem


Example (4 contd)
Re-arranging we obtain: F (s ) 1 Solving this equation we get: F (s ) = 4 4 + 4. s3 s 1 4 = 3. s+1 s

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

15/22

The Convolution Theorem


Example (4 contd)
Re-arranging we obtain: F (s ) 1 Solving this equation we get: F (s ) = Therefore, f (t) = L1 [F (s)] = 2 t2 + 2 3 t . 3 4 4 + 4. s3 s 1 4 = 3. s+1 s

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

15/22

The Convolution Theorem


Example (4 contd)
Re-arranging we obtain: F (s ) 1 Solving this equation we get: F (s ) = Therefore, f (t) = L1 [F (s)] = 2 t2 + 2 3 t . 3 4 4 + 4. s3 s 1 4 = 3. s+1 s

From this example, we see that the convolution theorem is useful in solving equations involving integrals of the unknown functions.
Michelle Dunbar, Week 9, Lecture 1 MATH283 Advanced Engineering Mathematics 15/22

The Convolution Theorem


Example (5)
Solve the IVP y 5y + 6y = f (t), where f (t) is a function dened for t 0. Solution:
Apply the Laplace Transform, inserting the initial values, to obtain: L y = s2 Y (s) sy (0) y (0) = s 2 Y (s ) and L y = sY (s) y (0) = sY (s)

y (0) = y (0) = 0,

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

16/22

The Convolution Theorem


Example (5 contd)
So taking the transform of the DE: L y 5L y + 6L [y ] = L [f (t)] s2 Y (s) 5sY (s) + 6Y (s) = F (s) Solving this equation for Y (s): Y (s ) = Now, since we may write: 1 1 1 = , s2 5s + 6 s3 s2 we have L 1 1 1 1 = L 1 L 1 s2 5s + 6 s3 s2 = e 3t e 2t
Michelle Dunbar, Week 9, Lecture 1 MATH283 Advanced Engineering Mathematics 17/22

1 s2 5s + 6

F (s )

The Convolution Theorem

Example (5 contd)
Applying the convolution theorem, we have y ( t ) = ( e 3t e 2t ) f ( t ) = e 3t f ( t ) e 2t f ( t ) This is the solution for any f (t) having a convolution with e3t and e2t .

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

18/22

Dierentiation of Transforms

Theorem (Dierentiation of Transforms)


If F (s) = L [f (t)], then: L [tf (t)] = dF (s) . ds

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

19/22

Dierentiation of Transforms

Theorem (Dierentiation of Transforms)


If F (s) = L [f (t)], then: L [tf (t)] = Equivalently, dF (s) 1 f (t) = L 1 t ds . dF (s) . ds

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

19/22

Dierentiation of Transforms

Theorem (Dierentiation of Transforms)


If F (s) = L [f (t)], then: L [tf (t)] = Equivalently, dF (s) 1 f (t) = L 1 t ds Applying the rst result n-times gives: L [tn f (t)] = (1)n d n F (s ) , dsn n = 1, 2, 3, . . . . . dF (s) . ds

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

19/22

Dierentiation of Transforms
Proof : dF (s) d = ds ds

f (t)est dt
0

=
0

d st e f (t) dt ds test f (t) dt

=
0

=
0

[tf (t)] est dt (by defn).

= L [tf (t)]
( s) . That is, L [tf (t)] = dF ds

This property is useful for solving DEs with polynomial coecients!


Michelle Dunbar, Week 9, Lecture 1 MATH283 Advanced Engineering Mathematics 20/22

Dierentiation of Transforms
Example (6)
Find L te6t . Solution:
Let f (t) = e6t , use the table of Laplace Transforms to nd: F ( s ) = L e 6t = 1 s6

Then applying the property of the derivative of the Laplace Transform, we obtain: L [tf (t)] = F (s) = 1 d ds s 6 1 = (s 6)2
21/22

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

Dierentiation of Transforms
Example (7)
Find L [t sinh 3t]. Solution:
Let f (t) = sinh 3t, use the table of Laplace Transforms to nd: F (s) = L [sinh 3t] = 3 s2 9

Then applying the property of the derivative of the Laplace Transform we obtain: L [t sinh 3t] = F (s) = 3 d ds s2 9 2 =3 2 (s 9)2

Michelle Dunbar, Week 9, Lecture 1

MATH283 Advanced Engineering Mathematics

22/22

Você também pode gostar