Sei sulla pagina 1di 4

### 18.02 Problem Set 11, Part I I Solutions

1.

We put the center of the sphere at the origin O as usual, and take the

“North Pole” N = (0, 0, a) as the ﬁxed p oint. Let P b e an arbitrary p oint on

the surface of the the sphere S , and D the straight-line distance from N to

P . Then D is the length of a side of the triangle Δ ONP. The other two sides

O N and O P b oth have length a and the angle b etween them is φ in spherical

co ordinates, so the Law of Cosines gives

D

=

a 2 (1 cos φ)

1

2

.

Then

¯

D

=

1

SA

� �

S

D dS .

We’ll use the formula SA = 4π a 2 for the surface area of

the sphere. The integral

� �

S

D dS =

2π

0

π

0

a

2 (1 cos φ)

1

• 2 a

• 2 sin φ dφ dθ

=

2π

2 a

3

π

0

(1 cos φ)

• 1 2 sin φ dφ

=

• 2 π

Dividing this by SA = 4π a 2 , we get D ¯

=

2 a

4a

• 3 .

3

2

3

(1 cos φ)

3

2

|

π

• 0 =

16π a 3

3

.

(As a check:

D ¯ clearly scales by a, i.e.

D ¯ = K a for some constant K .

D =

a when φ =

π

3

, or at 30 degrees North latitude. Since there are more p oints

on S b elow this latitude, we should have K > 1.

But D max = 2a (when P

is the South Pole), so we also must have K < 2.

So K = 4 is at least in the

3

correct range.

2.

Limits (in spherical co ordinates):

from 0 to 2π .

ρ from 0 to a,

φ from 0 to φ 0 ,

θ

If

dm

is lo cated at (x, y , z ) then the force due to

dm

is

dF = G x,

y , z dm = G x, y , z

ρ 3

ρ 3

δ dV

= G

x, y , z

ρ 3

dV .

Let the total force F = a, b, c. By symmetry a = b = 0.

We compute c =

� � �

D

G

z

3 dV

ρ

= G

0

2π

0

φ

0

0

a

ρ cos φ

ρ 3

ρ 2 sin φ dρ dφ dθ

= G

0

2π

0

φ 0

0

a

cos φ sin φ dρ dφ dθ ..

Inner integral: a cos φ sin φ.

Gπ a sin 2 φ 0 .

Middle integral: a

sin 2 φ 0 /2.

Outer integral:

F = 0, 0, Gπ a sin 2 φ 0 .

3.

a) T

On T :

= disk of radius 1 at height z = 1,

n = k.

F

·

n = 1

ﬂux =

� �

T

F

·

n dS =

� �

T

dS = area =

π .

b) See the picture.

Let D 1 b e the volume b ounded by S and T .

1

Let D 2 b e the volume b ounded by T and U .

Rememb er we are consistently using upward normals and upward ﬂux.

The divergence theorem gives

ﬂux through S - ﬂux through T =

� � �

D 1

div F dV =

� � �

D 1

dV = volume(D 1 ).

ﬂux through S = volume(D 1 ) + ﬂux through T = volume(D 1 ) + π .

Likewise, ﬂux through T - ﬂux through U = volume(D 1 ).

ﬂux through U = π - volume(D 2 ).

Computing volumes:

D 2 : Volume(D 2 )

1

• 3 base × height =

π
.
3

D 1 :

We do this at the end in two diﬀerent ways. The answer is volume(D 1 ) =

2π

2 2

2

5

6

=

4π 2

3

5π

3

.

Thus we have, ﬂux through S = volume(D 1 ) + π =

Flux through U = π - volume(D 2 ) =

.
3
4π √ 2
.
3
3

As promised we compute volume(D 1 ) two diﬀerent ways.

Metho d 1:

volume(D 1 ) =

0

2π

0

π /4

2

sec φ

ρ 2 sin φ dρ dφ dθ .

(The ρ limits are from z = 1

ρ = sec φ.)

Inner integral:

ρ

3

3

sin φ

2

sec φ

=

2

2

3

sin φ

cos

3

φ

3

sin φ.

Middle integral:

2

2

3

cos φ

cos

2

6

π /4

0

=

2

2

3

1

(1 − √

2

) +

1

6

(1 2) =

Outer integral: volume(D 1 ) = 2π

2

2

2

5

6

=

4π

2

3

5π

3

.

Metho d 2:

volume(D 1 ) = volume(D 1 + D 2 ) - volume(D 2 ).

2

2

3

5

.

6

Volume(D 1 + D 2 ) is an easier integral than in metho d 1.

Volume(D

1 +D 2 )

=

0

2π

0

π /4

0

2

ρ

2

sin φ dρ dφ dθ = 2π

2

2

3

(1 cos(π /4)) =

Using volume(D 2 ) = π /3 we get volume(D 1 ) =

4π 2

3

5π

3

(same as metho d 1)

4π

2

2

4π

3

.

2

• c) U is given by z =

n dS = �−z

x

, z

y

ﬂux =

� �

R

4.

∂ ρ

a)Use x

x 2 + y 2 = r .

F

, 1dx dy

·

n dS = z dx dy

r 2 dr dθ =

0

2π

0

1

3

.

x ρ 1 = ρ 3

x

etc.

z dx dy =

x

x
y
z
.
ρ 3 , ρ 3 , ρ 3

F = f =

= ρ , etc.

Now use

∂ x

3

= ρ 3

3x

2

ρ

5

(and similarly for y

and z ).

div F = (ρ 3 + 3x 2 ρ 5 ) + (ρ 3 + 3y 2 ρ 5 ) + (ρ 3 + 3z 2 ρ 5 ) = 3ρ 3 + 3ρ 2 · ρ 5 = 0.

Q

• b) The divergence theorem do es not apply b ecause F is not deﬁned at 0.

On S :

n = x, y , z ,

a

F =

�−x. y , z

a

• 3 .

F · n =

1

a

2

1

ﬂux = a 2 ·area =

1

2 4π a 2 =

a

−4π .

c)Let S b e any closed surface around 0.

Let S 1 b e a small sphere centered

at 0 and completely insided S . Use outward normals for b oth surfaces (and

b e careful with signs).

D is the volume b etween S and S 1 .

From part (a) we know div F = 0, so the divergence theorem gives

� �

S

F · n dS

� �

S 1

F · n dS =

� � �

D

div F dV = 0.

� �

S

F · n dS =

� �

S 1

F · n dS = 4π .

QED

• 5. Use the fact that f

is p erp endicular to the iso-surface f

= c, so that

dep ending on whether f p oints inward or outward, f · n = ± | �f |, where

n is the outward unit normal to S .

Then apply the Divergence Theorem

to get

� �

S

f � · n dS =

� � �

· ( f � ) dV , where G is the interior of S .

G

Substituting into the RHS integral then gives

± S

� �

| �f | dS = S f · n dS =

� �

� � 2 f dV .

G

3

MIT OpenCourseWare

18.02SC Multivariable Calculus

Fall 2010