2

I have a cylinder where the top is given by $ z=-2$ the bottom is $z=2$ and the "lateral surface" is $x^2+y^2=1$. I have to find$\iint_{s}(4x+3y+z^2) \ dS$. I know this involves using the divergence theorem and then plugging in the answer into the volume of a cylinder but I have no idea how to set up the limits on the triple integral once I have found the partials and split up the integral for the divergence theorem. Thanks.

Abigail
  • 21

2 Answers2

5

EDIT: The OP's desire to apply the divergence theorem is ambitious, but can be achieved. To apply the Divergence Theorem, we must have the surface integral in the form of $\iint_S \vec F\cdot\vec n dS$. So this problem gets sneaky. To this end:

Consider the vector field $\vec F_1=(4,3,(4x+3y)z/2)$. Since $\vec n = (x,y,0)$ on the lateral surface and $\vec n = (0,0,z/2)$ on the two disks, this accounts for $\iint_S (4x+3y)dS$ and gives us flux $\iiint_V \text{div}\vec F_1 dV =\iiint_V \frac12(4x+3y)dV = 0$, by symmetry. Now consider $\vec F_2 = z^2(x,y,z/2)$. By similar reasoning, its flux across $S$ is indeed $\iint_S z^2 dS$. So the flux of $\vec F_2$ is $$\iiint_V \text{div} F_2 dV = \iiint_V \frac72z^2\,dV =\frac72\int_0^{2\pi}\int_0^1\int_{-2}^2 z^2r dzdrd\theta = \frac{56}3\pi\,.$$
(Note that we do indeed get the same answer by the much simpler direct calculation.)

Ted Shifrin
  • 115,160
  • this is OP with official stackexchange accoun. I couldn't comment on Ted's solution bc of no reputation points but if anyone could edit this into a comment that would be great. Would you recommend using polar coordinates in which case the bounds on integrals would be easy or would you take a different approach? I understand how to apply the divergence theorem, I'm just having a little troubling setting up that end triple integral before finding an answer to pluginn. –  Feb 04 '14 at 06:12
  • Yes, Abigail, use cylindrical coordinates. This should be straightforward. (You definitely should practice setting up such triple integrals, but my answer suggested you can do this particular problem without setting up or evaluating any integrals!) – Ted Shifrin Feb 04 '14 at 13:41
  • First thank you for your answer. But I cannot understand why this problem is related with divergence. Can you explain ? I agree that we cannot put this problem into setting, related with divergence theorem. My solution is : $\int\int (4x+3y)\ dS=0$ and when we calculate $z^2$-term, by symmetry we are sufficient to consider $\int z^2\ dz$ – HK Lee Feb 08 '14 at 02:44
  • @HeeKwonLee: You're right that I was too hasty here. I imagined the flux of a vector field. I have corrected my answer. – Ted Shifrin Feb 08 '14 at 03:37
  • I see the construction of ${\bf F}_i$. Thank you for your answer. – HK Lee Feb 08 '14 at 04:07
-1

There is nothing wrong with this approach! The OP said "I know it involves using the divergence theorem.." but this does not mean that he is not interested in other approach.

Since the surface $S$ consists of three parts, then the integral can be evaluated as

$$ \iint_{S} = \iint_{S_1} + \iint_{S_2} + \iint_{S_3} \longrightarrow (1) $$

First, we parametrize the lateral surface as

$$ r(u,v)= \langle \cos u, \sin u, v \rangle, \quad -2 \leq v \leq 2,\, 0\leq u \leq 2\pi. $$

Then, we have

$$\iint_{S} (4x+3y+z^2) dS = \int_{-2}^{2}\int_{0}^{2\pi} (4\cos u + 3\sin u +v^2)|\textbf{r}_u \times \textbf{r}_v|dudv .$$

For the second and the third integrals in $(1)$, you do not need to parametrize, just use the technique. Now, you should be able to finish the problem.

Note: We used the following identity

$$ \iint_{S} f \,dS = \iint_{T} f(\mathbf{x}(u, v)) \left\|{\partial \mathbf{x} \over \partial u}\times {\partial \mathbf{x} \over \partial v}\right\| du\, dv $$

  • 5
    As usual, you completely ignore the question in the OP's posting. She specifically wants to apply the Duvergence Theorem. Moreover, you have parametrized only one-third of the surface. – Ted Shifrin Feb 04 '14 at 04:53
  • @TedShifrin: First of all, I did not ignore the question of the OP. Secondly, using the "Divergence theorem" is an approach you can use to solve the problem. You should read carefully the question the OP said "I know it involves..." which it does not mean that he really wants this approach. Even if he wants this approach, there is nothing wrong with this answer. So, just do not rush to judge things. So, it seems you understand things one way. – Mhenni Benghorbal Feb 04 '14 at 06:20
  • 3
    (I've not voted on this answer yet) Your parametrization seems to miss the top and bottom of the cylinder, no? –  Feb 04 '14 at 06:32
  • @T.Bongers: Thanks for the comment. Yes, I missed them. To many things were going on at the same time. – Mhenni Benghorbal Feb 04 '14 at 19:34
  • @downvoter: can you tell us what the downvote for? – Mhenni Benghorbal Feb 08 '14 at 04:53