College of Engineering and Computer Science

Mechanical Engineering Department

Mechanical Engineering 501B

Seminar in Engineering Analysis
Spring 2009 Class: 14443 Instructor: Larry Caretto

March 9 Homework Solutions

1.

Kreyszig, page 546, problem 2. Find u(x,t) on a string of length, L = 1 and c 2 = 1, when
the initial velocity is zero, and the initial deflection with small k (say 0.01) is k[sin x
(1/3) sin 3x].

From slide 456 of the lecture presentation for February 23, the solution to the wave equation for a
vibrating string with zero initial velocity is.

nct
n x
sin

L
L

u ( x , t ) Bn cos
n 1

[1]

The value for Bn is given by the following equation from slide 44 of the same lecture presentation.

2
m x
f ( x ) sin
dx

L0
L
L

Bm

[2]

Setting L = and applying the equation for Bm to the initial deflection in this case, f(x) = k(sin x
0,5 sin 2x) gives.

2
2
mx
Bm f ( x ) sin
dx k (sin x 0.5 sin 2 x ) sin mx dx
L0
0
L
L

2k
k
2k
k
sin x sin mxdx sin 2 x sin mxdx
m1

0
0
2
2 m2

[3]

In equation [3] we recognize that the initial conditions are orthogonal to the eigenfunctions.
Because of this, the Bm coefficients are nonzero only if m = 1 or m = 2. In these cases we use the

result that

sin
0

mxdx

for integer m.
2

Substituting c = 1, B1 = k and B2 = -k/2 into the general solution of equation [1] gives the solution
for this problem as

u ( x, t ) k cos t sin x

Engineering Building Room 1333

E-mail: lcaretto@csun.edu

k
cos 2t sin 2 x
2

Mail Code
8348

[4]

Phone: 818.677.6448
Fax: 818.677.7062

Solutions to March 9 homework

2.

ME501B, L. S. Caretto, Spring 2009

Page 2

Kreyszig, page 546, problem 4. Find u(x,t) on a string of length, L = 1 and c 2 = 1, when
the initial velocity is zero, and the initial deflection with small k (say 0.01) is kx(1 x 2).

This problem is done in exactly the same way as the previous problem; the solution is given by
equation [1] above with the Bn coefficients given by equation [2]. For given length, L = , and the
initial conditions of this problem, f(x) = 0.1x( 2 x2), the expression for Bm becomes.

2
2
mx
2
2
f ( x ) sin
dx 0.1x ( x ) sin mx dx

L0
L

0
L

Bm

[5]

The integration for these initial conditions may be expressed as a combination of two separate
integrals as shown below.

0.2 2

sin mx x cos mx
0 x sin mx dx 0.2 m 2 m

0.2 2 cos m
0.2 2 ( 1) m
[6]

m
m

0 .2 3
0.2 x 3 cos mx 3x 2 sin mx 6 sin mx 6 x cos mx

x
sin
mx
dx

0

m
m2
m4
m3

0.2 cos m 6 cos m

0.2 ( 1)
1.2( 1)

3

m
m
m
m3

[7]

We subtract the result in equation [7] from the result in equation [6] to get the value of B m.

2
0.2 2
Bm 0.1x ( 2 x 2 ) sin mx dx
0

0.2 3
0 x sin mx dx 0 x sin mx dx

0.2 2 ( 1) m 0.2 2 ( 1) m 1.2( 1) m

1.2( 1) m

m
m
m 3
m3

The MATLAB commands to perform this integration and the result are shown below.
>> B=0.2/pi*int('x*(pi^2-x^2)*sin(m*x)',0,pi)
B=
-4587328911378127/36028797018963968*(m^2*pi^2*sin(pi*m)-3*sin(pi*m)
+3*pi*m*cos(pi*m))/m^4
>> pretty(B)
2 2
4587328911378127 m pi sin(pi m) - 3 sin(pi m) + 3 pi m cos(pi m)
- ----------------- ------------------------------------------------36028797018963968
4
m
>>

[8]

Solutions to March 9 homework

ME501B, L. S. Caretto, Spring 2009

Page 3

Reducing the expression from MATLAB shows that this code gives the same result for B m as the
one found by direct integration. Substituting the expression for B m into equation [1] and setting L
= and c = 1 gives the desired solution.

( 1) n cos nt sin nx
n3
n 1

u ( x, t ) 1.2

3.

[9]

Kreyszig, page 547, problem 12. Find the deflection u(x,t) of a string of length, L =
and c2 = 1, for zero initial displacement and triangular initial velocity, u t(x,0) = 0.01x if
0 x /2 and ut(x,0) = 0.01( x) if /2 x .

Here we start with the general solution to the wave equation for the case of zero displacements.
This equation is derived from the general solution on slide 45 of the February 23 lecture
presentation by setting Bn = 0 for zero initial displacement. (See the equation for B n on slide 44 to
see that Bn = 0 for zero initial displacements.)

nct
nx
sin

L
L

u ( x, t ) An sin
n 1

[10]

The equation for the An coefficients is also taken from slide 44.

2
mx
Am
g ( x ) sin
dx

mc 0
L
L

[11]

Setting L = and c = 1 and substituting the given initial velocity in equation [11] gives.

2
2
mx
g ( x ) sin
dx

mc 0
m
L
L

Am

/2

(0.01x ) sin mxdx

0

2
m

0.01( x ) sin mxdx

/2

[12]
The first integral is found as follows.

2
m

/2

0.02 sin mx x cos mx

m m 2
m

(0.01x ) sin mxdx

0

m / 2 cos m / 2 0

/2

0.02 sin m / 2 0
m
m2

0.02
m m
m
m 3 sin 2 2 cos 2

[13]

The second integral is

2
m

0.01( x ) sin mxdx

/2

0.02 cos mx

m
m

/2

sin mx x cos mx

2
m
m

/2

0.02

m
m
m
m
m cos
cos m sin m sin
m cos mx
cos
3
m
2
2
2
2

[14]

Solutions to March 9 homework

ME501B, L. S. Caretto, Spring 2009

Page 4

0.02
m
m
m
cos m m m sin
m

cos
3

m
2
2
2

0.02 m
0.01 m 0.02 m
m
m
cos
sin
2 cos
3 sin

m 2
m
2
2
2 m
2

Adding the two integrals gives Am.

2
Am
m

/2

2
0 (0.01x) sin mxdx m

0.01( x ) sin mxdx

/2

0.02
0.04 m
m m
m 0.01 m 0.02 m
sin

cos
2 cos
3 sin
3 sin

m
m
m
2 2
2
2 m
2
2
[15]
The MATLAB commands and result for this problem are shown below.
>> syms m x
>>pretty(simplify( 0.02/(m*pi)*(int('x*sin(m*x)',0,pi/2)+int('(pi-x)*sin(m*x)',pi/2,pi))))
2 sin(1/2 pi m) - sin(pi m)
1/50 --------------------------3
pi m
The result for Am = (.004/m3)sin(m/2) will cycle with the sine term as the remainder when m is
divided by 4 (m mod 4) ranges from 0 to 3. The general result for this term is

0
0.04

Am

m mod 4 0
m3

0.04

m 3

m mod 4 1
m mod 4 2

[16]

m mod 4 3

We can write the result when this expression for Am is substituted into equation [1] (with Bm equal
to zero) by defining a new index, m = 2n + 1 which will cover all odd numbers as n ranges from
zero to infinity. A factor of (-1)n will give the correct sign. With this modification, the solution for
this problem becomes.

u( x, t )

4.

0.04 ( 1) n
sin(2n 1)t sin( 2n 1) x

n 0 ( 2n 1)3

[17]

Kreyszig, page 552, problem 5. Longitudinal Vibrations of an Elastic Bar or Rod. These
vibrations in the direction of the x axis are governed by the wave equation u tt = c2yxx, c2
= E/. If the rod is fastened at one end, x = 0, and free at the other, x = L, we have u(0,t)
= 0 and ux(L,t) = 0 (because the force at the free end is zero). Show that the motion

Solutions to March 9 homework

ME501B, L. S. Caretto, Spring 2009

Page 5

corresponding to initial displacement u(x,0) = 0 and initial velocity zero matches the
equation given in the text.
From slide 38 of the February 23 lecture notes, the general separation of variables solution for
the wave equation is.

u ( x, t ) A sin(ct ) B cos(ct ) C sin( x ) D cos(x )

[19]

In order to have u(0,t) = 0 we must have D = 0 to eliminate the cosine term. To evaluate the
gradient boundary condition at x = L, we can take the x derivative of equation [19] after setting D
= 0.

u
( x, t ) A sin( ct ) B cos(ct ) C cos( x )
x

[20]

To match the condition that u/x is zero at x = L, we must have cos(L) = 0. This will be true if L
is an odd multiple of /2. That is we must have

2n 1

[21]

where n is an integer greater than or equal to zero.

We can rewrite our general solution in equation [19] as the sum of all possible eigenfunctions,
combining constants AC and BC into two separate constants An and Bn.

u( x, t )

n 0

sin( n ct ) Bn cos(n ct ) sin( n x )

[22]

Taking the time derivative of this equation will give us the velocity.

u
( x, t )
t

A
n

n 0

cos(n ct ) Bn sin(n ct ) sin(n x )

[23]

At t = 0, the velocity is zero. This gives the following result.

u
( x ,0) 0
t

A
n 0

cos(0) Bn sin(0) sin(n x )

n 0

An sin( n x )

[24]

The only way we can satisfy this initial condition is by setting all values of An to zero. For the
initial displacement, we can use an eigenfunction expansion of equation [22] (with An set to zero).

u ( x,0) f ( x )

Bn cos(0) sin(n x )
n 0

sin(n x )

[25]

n 0

If we multiply this equation by sin(mx)dx and integrate from 0 to L we obtain the following result
from the orthogonality of the eigenfunctions.1

We have the usual result for the sin2 integral:

Solutions to March 9 homework

f ( x ) sin(

x )dx

ME501B, L. S. Caretto, Spring 2009

B
0 n 0

Page 6

sin( n x ) sin( m x )dx Bm sin 2 ( m x )dx

0

Bm L
[26]
2

This gives the result for Bm shown below.

L

Bm

2
f ( x ) sin(m x )dx
L 0

[27]

With An = 0, our solution in equation [22] becomes

u( x, t )

B
n 0

cos( n ct ) sin( n x )

[28]

Other than a difference in notation, this solution (along with equation [21] for and equation [27]
for Bm) match the solution given in the text.
5.

Show that Laplaces equation is elliptic, the heat equation is parabolic, the wave
equation is hyperbolic, and the Tricomi equation , yu xx +uyy = 0 is of mixed type (elliptic
in the upper half plane and hyperbolic in the lower half plane).

The classification of equations is based on the general equation shown below.

2u
2u
2u
u u
B
C
D x, y , u,
,
0
2
x
xy
y 2
x y

[18]

The equation is elliptic, parabolic, or hyperbolic if B 2 4AC is less than, equal to, or greater than
zero. Laplaces equation,

2 u 2u

0 , has the form of equation [18] with A = C = 1 and B =

x 2 y 2

D = 0. Thus, for this equation, B2 4AC = 02 (1)(1) = -1, which is less than zero. Thus
Laplaces equation is elliptic. The diffusion or heat equation,
equation [18] with D =

1 u 2u

0 , has the form of

t x 2

1 u
, A = 1, and B = C = 0. Thus, B2 4AC = 02 (1)(0) = 0,
t

indicating that this equation is parabolic. We can write the wave equation in the following form:
2
2u
2 u

c
0 . This has the general form of equation [18] with A = 1, B = D = 0, and C = -c 2.
2
2
t
x

Here, B2 4AC = 02 (1)(-c2) = c2, indicating that the wave equation is hyperbolic. The Tricomi
equation, y

2u 2u

0 , has the general form of equation [18] with A = y, B = D = 0, and C =

x 2 y 2

1. Here, B2 4AC = 02 (1)(y) = y. When y is positive, B2 4AC is negative indicating that the
L

x
(2n 1)x
L
2( 2n 1) x
sin
sin
dx

L
L

2 4(2n 1)
0

L
L
2(2n 1)L
sin
0

4(2n 1)
L
2

L 0

2 2

Solutions to March 9 homework

ME501B, L. S. Caretto, Spring 2009

Page 7

equation is elliptic. When y is negative, B2 4AC is negative indicating that the equation is
hyperbolic. Thus the Tricomi is hyperbolic in the lower half plane and elliptic in the upper half
plane. Along the horizontal (y = 0) axis, the equation is parabolic.