next up previous
Next: About this document ...

18.02 Problem Set 10 (due Thursday, April 29, 1999)

Part I     (9 points)

Hand in the the underlined problems; the others are for practice.

Lecture 30 (Thurs. April 22): Surface integrals and flux.

Read: SN, vector calculus section 9.

Problems: SN p.9.7 nos. 1, 2, 3, 4, 6, 8 (S. 36,37).

Lecture 31 (Thurs. April 22): Applications of flux, divergence theorem.

Read: EP pp. 995-998, 1000-1001.

Lecture 32 (Tues. April 27): Divergence theorem.

Read: SN, Vector Calculus, section 10; EP pp. 1006-1008.

Problems: SN, p. 10.5 nos. 1a, 2, 3, 5; p. 10.6 6, 7i, 8 (S.38,39).

Part II     (20 points)

Directions: Try each problem alone for 25 minutes. If you subsequently collaborate, solutions must be written up independently. It is illegal to consult old problem sets.

1.
(Thurs. 8 pts: 2, 3, 3) Take the surface bounded below by the right-angled cone $ z=(x^2
+ y^2)^{1/2}$, and above by the unit sphere centered at the origin. The upper and lower surfaces intersect in a circle.

Let $ S$ be the disc having this circle as its boundary; $ T$ the spherical cap forming the upper surface, and $ U$ the cone forming the lower surface. Orient all three surfaces so that the normal vector points generally upward (i.e, has a positive $ k$-component).

Calculate the flux of the vector field $ F=z{\bf k}$ over each of these three surfaces. In each case, do the calculation directly from the definition of the surface integral for flux, as in the Notes and Part I problems.

Solution. Flux through $ S:$


$\displaystyle \iint_S \vec{F} \cdot \vec{n} dS$ $\displaystyle =$ $\displaystyle \iint_S z \, dS$  
  $\displaystyle =$ $\displaystyle \iint _{x^2+y^2 \leq 1/2} \, dx \, dy/ \sqrt{2} =
\frac{\pi}{2 \sqrt{2}} \, .$  

Since $ z= 1/ \sqrt{2}$ on the disk.

Flux through $ T:$

$\displaystyle \iint_T \vec{F} \cdot \vec{n} \, dS$ $\displaystyle =$ $\displaystyle \iint_T z^2 \sin \varphi \, d \varphi \, d \theta$  
  $\displaystyle =$ $\displaystyle \int^{2 \pi}_0 \int^{\pi /4}_0 \cos^2 \varphi \sin
\varphi \, d \varphi \, d \theta$  
  $\displaystyle =$ $\displaystyle \frac{2 \pi}{3} - \frac{\pi}{3 \sqrt{2}} \, .$  

Flux through $ U:$ On $ U$

\begin{multline*}
z= \sqrt{x^2 + y^2} = f(x,y) \, , {\bf\hat{n}} \, dS =
\left...
...t{2}}_0
r \cdot r \, dr \, d \theta = \frac{\pi}{3 \sqrt{2}}.
\end{multline*}

2.
(Thurs. 3 pts.) A pattern of heat generation and absorption produces the temperature distribution in space ($ u$ is the temperature at the point $ (x,y,z))$:

$\displaystyle u=e^{-x^2-y^2-z^2} \, .
$

Find the heat flow across a sphere of radius $ a$ centered at the origin (cf. EP  p. 1000; call the heat conductivity $ k$).

For what radius $ a$ will the heat flow across the sphere be greatest?

Solution: The heat flow across a surface is the flux of the vector field $ -k\nabla u$ where $ u$ is the temerature distribution. In this case,

$\displaystyle \nabla u=e^{-x^2-y^2-z^2}(-2x{\bf i}-2y{\bf j}-2z{\bf k}.$    

Thus the flux across the sphere of radius $ a$ (outwards) is

\begin{multline*}
\iint_S{\bf F}\cdot{\bf\hat n}dS\\
=(-k)\iint (-2)e^{-x^2-y^2...
...dot\frac1a(x{\bf i}+y{\bf j}+z{\bf k})dS\\
=8\pi a^3 ke^{-a^2}.
\end{multline*}

The maximum value for this will occur when the derivative vanishes, that is when $ 8\pi k a^2e^{-a^2}(3-2a^2)=0$ so $ a=\sqrt\frac32.$

3.
(Fri. 4 pts: 2,2) In problem 1 above, you calculated directly the flux of the vector field $ F=z{\bf k}$ over three surfaces shown in cross-section at the right. The unit with the spherical cap $ T$, the cone $ U$ given by $ z=r$, and the disc $ S$ bounded by the circle in which $ T$ and $ U$ intersect. In each case, the normal vector is the one with a positive $ k$-component.

The correct value for the flux over the disc $ S$ is $ \pi
\sqrt{2}/4$. Using this:

(a)
Use the divergence theorem (and the formula for the volume of a solid cone) to get the flux over the conical surface $ U$. (Watch for orientations!)

(b)
Similarly, use the divergence theorem to get the flux over $ T$. (get the volume of the whole ice cream cone and subtract the cone volume to get the volume of the spherical cap on top.)

Solution:

(a)
By the divergence theorem,

$\displaystyle \iint_{S}{\bf F}\cdot{\bf\hat n}dS+\iint_{U'}{\bf F}\cdot{\bf\hat n}dS=\iiint_{C}(\nabla\cdot{\bf F})dV=$Vol$\displaystyle (C)=\frac{\sqrt2\pi}{12}$    

where $ U'$ is $ U$ with the opposite normal and $ C$ is the solid cone. Therefore,

$\displaystyle \iint{U}{\bf F}\cdot{\bf\hat n}dS=-\iint_{U'}{\bf F}\cdot{\bf\hat n}dS =\frac{\sqrt2\pi}6.$    

(b)
Volume of solid is

$\displaystyle \int_0^{2\pi}\int_0^{\pi/4}\int_0^1\rho ^2\sin\varphi d\rho \,d\varphi \,d\theta =\frac{2\pi}3-\frac{\sqrt2\pi}3.$    

So the volume of the cap is $ \frac{2\pi}3-\frac{5\sqrt2\pi}{12}.$ By the divergence theorem

$\displaystyle \iint_{S'}{\bf F}\cdot{\bf\hat n}dS+\iint_{T}{\bf F}\cdot{\bf\hat n}dS=$Vol of cap    

so

$\displaystyle \iint_{T}{\bf F}\cdot{\bf\hat n}dS=\frac{2\pi}3-\frac{\sqrt2\pi}6.$    

4.
(Fri. 2 pts.) Let $ S$ be a smooth closed surface. Show that the field $ x{\bf i}+y{\bf j}+z{\bf k}$ cannot be tangent to $ S$ at every point $ (x,y,z) $ of $ S$.

Solution: By the divergence theorem $ \iint_S{\bf F}\cdot{\bf\hat
n}dS=\iiint_R\nabla\cdot{\bf F}dV.$ If $ {\bf F}=x{\bf i}+y{\bf j}+z{\bf k}$ then $ \nabla\cdot{\bf F}=3$ and the volume integral is $ 3$Vol$ (R)\not=0.$ Thus the flux of $ {\bf F}$ through $ S$ cannot be zero, so $ {\bf F}\cdot{\bf\hat n}=0$ at some point on $ S$ and therefore $ x{\bf i}+y{\bf j}+z{\bf k}$ cannot be tangent to the surface everywhere.

5.
(Fri. 3 pts.) Prove that if $ f(x,y,z)$ satisfies Laplace's equation (see Notes P), then the flux of its gradient field $ \triangledown f$ across any smooth closed surface is 0.

Solution: If $ {\bf F}=\nabla f=f_x{\bf i}+f_y{\bf j}+f_z{\bf k}$ then $ \nabla\cdot{\bf F}=f_{xx}+f_{yy}+f_{zz}=0,$ precisely the given condition that $ f$ satisfy Laplace's equation. The by the divergence theorem the flux of $ \nabla f$ through any closed surface is $ \iint_S{\nabla a}\cdot{\bf\hat n} dS=0.$




next up previous
Next: About this document ...
Richard B. Melrose
1999-04-29