Calculating flux using surface integrals.

Join the discussion
Registration is free. Ask a follow-up in this thread, or start your own.
9 replies · 4K views
thepatient
Messages
160
Reaction score
0
This isn't homework. I've been restudying vector calculus from the beginning to end on my free time and got stuck on this problem. I am not sure what I'm doing wrong, but it may be a calculation error since it has so much calculation involved.

Homework Statement


Evaluate the surface integral [itex]\int[/itex][itex]\int[/itex][itex]_{\Sigma}[/itex]f*d[itex]\sigma[/itex], where f(x, y, z) = x2i + xyj + zk and is the
part of the plane 6x + 3y + 2z = 6 with x ≥ 0, y ≥ 0, and z ≥ 0, with the outward unit
normal n pointing in the positive z direction.

Homework Equations


[itex]\int[/itex][itex]\int[/itex][itex]_{\Sigma}[/itex]f*d[itex]\sigma[/itex] = [itex]\int[/itex][itex]\int[/itex][itex]_{R}[/itex]f*n*d[itex]\sigma[/itex], where
d[itex]\sigma[/itex] = |[itex]dr/du[/itex] X [itex]dr/dv[/itex]|dudv, and r is a vector parametricized for the surface.

n = outward unit normal vector from the surface.

The Attempt at a Solution


First, I parametricized the surface as a vector in terms of u and v:
x = u
y = v
z = 2 -3x - 1.5y
[itex]\overline{r}[/itex](u,v) = <u, v, 2-3u - 1.5v>
R is the region on the u, v plane under the line v = 2 -2u, u≥0, v≥0
Took partial derivatives in terms of u and v:
d[itex]\overline{r}[/itex]/du = <1, 0, -3>
d[itex]\overline{r}[/itex]/dv = <0,1,-1.5>

Calculated |[itex]dr/du[/itex] X [itex]dr/dv[/itex]|:
i j k
1 0 -3
0 1 -1.5 =
|3i + 1.5j + k| = (9 + 2.25 + 1)^1/2 = 3.5

Calculated n:
Since the plane is 6x + 3y + 2z = 6, perpendicular to the plane would be the vector v = <6, 3, 2>, so n would be v/(|v|), the vector divided by its magnitude.
n = <6, 3, 2>/|7| = <6/7, 3/7, 2/7>

Calculating f(u,v) * n:
f(u,v) = <(u^2), (uv), (2 -3u - 1.5v)>* <6/7, 3/7, 2/7> =
1/7 (6u^2 +3uv + 4 - 6u - 3v)

Plugging into equation:
[itex]\int[/itex][itex]\int[/itex][itex]_{R}[/itex]f*n*d[itex]\sigma[/itex] R=[(u,v): 0<u<1, 0<v<2-2u]

[itex]\int[/itex][itex]\int[/itex][itex]_{R}[/itex] 1/7 (6u^2 +3uv + 4 - 6u - 3v) *(3.5) *dvdu =

1/2 [itex]\int[/itex][itex]\int[/itex][itex]_{R}[/itex]6u^2 + 3uv -6u - 3v + 4 dv du =
1/2 [itex]\int[/itex]6u^2v + 3uv^2 -3v^2/2 + 4v (evaluating v from 0 to 2-2u) dv

Using wolfram alpha to evaluate (got tired of doing it by hand like 6 times XD) :
1/2 [itex]\int[/itex]24u^3 + 30u^2 + 8u + 2 du from u = 0..1
1/2 * (24u^4/4 + 30u^3/3 + 8u^2/2 + 2u) from 0 to 1
1/2 (6 + 10+4+2) = 11

The answer should be 15/4 though. I'm not sure what I did wrong, maybe I missed something. Can anyone see what I did wrong in this problem? Thanks much.
 
Last edited:
Physics news on Phys.org
Oops. I redid it and got 13/2 with the -6u term. It came out to 13/2 though. :(
 
If I remember correctly, since the surface Σ is closed, the divergence theorem applies right?

∫∫Rf*n*dσ = ∫∫∫div(f) dV

x is from 0 to one
y is from 0 to the line -2x + 2
z is from 0 to the plane 3 - 3x -3y/2

f(x, y, z) = x2i + xyj + zk
div(f) = d/dx i + d/dy j + d/dz k * x^2i + xyj + zk
= 2x + x + 1 = 3x+1

∫∫∫div(f) dV = ∫∫∫3x+1 dzdydx
∫∫3xz+z | z = 0..3-3x-3y/2 dy dx
∫∫3x(3-3x-3y) + 3-3x-3y -[ 0] dydx
∫∫9x -9x^2 - 9xy +3 -3x - 3y dy dx
∫∫6x - 9x^2 -9xy +3 -3y dy dx
∫6xy - 9x^2y - 9xy^2/2 + 3y - 3y^2/2 | y = 0..-2x+2 dx
∫6x(-2x+2) - 9x^2(-2x+2) -9x(-2x+2)^2/2 +3(-2x+2) -3(-2x+2)^2/2 dx
∫-12x^2 + 12x +18x^3-18x^2-(9x/2) (4x^2 -8x + 4) -6x + 6 -3(4x^2 - 8x + 4) dx=
∫-12x^2 + 12x + 18x^3 -18x^2 -36x^3/2 + 72x^2/2 -36x/2 + 6 -12x^2 +24x - 12 dx=
-12x^3/3 + 12x^2/2 + 18x^4/4 -18x^3/3 -36x^4/8 + 72x^3/6 - 36x^2/4 +6x -12x^3/3 +12x^2 -12x from x =0..1
-12/3 + 12/2 + 18/4 -18/3 -36/8 +12 -9 +6 -4 + 12-12 =1


:\ I'm frustrated with this problem.
 
Hmm.. But the problem states "...part of the plane 6x + 3y + 2z = 6 with x ≥ 0, y ≥ 0, and z ≥ 0, with the outward unit normal n pointing in the positive z direction." That doesn't mean that the flux is the net flux of the closed surface, but the flux of only that portion of the plane?
 
thepatient said:
First, I parametricized the surface as a vector in terms of u and v:
x = u
y = v
z = 2 -3x - 1.5y
You solved for z incorrectly. It should be z = 3 - 3x - (3/2)y = 3 - 3u - (3/2)v.
[itex]\overline{r}[/itex](u,v) = <u, v, 2-3u - 1.5v>
R is the region on the u, v plane under the line v = 2 -2u, u≥0, v≥0
Took partial derivatives in terms of u and v:
d[itex]\overline{r}[/itex]/du = <1, 0, -3>
d[itex]\overline{r}[/itex]/dv = <0,1,-1.5>

Calculated |[itex]dr/du[/itex] X [itex]dr/dv[/itex]|:
i j k
1 0 -3
0 1 -1.5 =
|3i + 1.5j + k| = (9 + 2.25 + 1)^1/2 = 3.5

Calculated n:
Since the plane is 6x + 3y + 2z = 6, perpendicular to the plane would be the vector v = <6, 3, 2>, so n would be v/(|v|), the vector divided by its magnitude.
n = <6, 3, 2>/|7| = <6/7, 3/7, 2/7>
All this is fine.
Calculating f(u,v) * n:
f(u,v) = <(u^2), (uv), (2 -3u - 1.5v)>* <6/7, 3/7, 2/7> =
1/7 (6u^2 +3uv + 4 - 6u - 3v)
You propagated the error from above, but your method is okay. With the correct expression for z, you would get
[tex]\frac{1}{2}\int_0^1\int_0^{2(1-u)} (6u^2+3uv+6-6u-3v)\,dv\,du = \frac{7}{4}[/tex]
Plugging into equation:
[itex]\int[/itex][itex]\int[/itex][itex]_{R}[/itex]f*n*d[itex]\sigma[/itex] R=[(u,v): 0<u<1, 0<v<2-2u]

[itex]\int[/itex][itex]\int[/itex][itex]_{R}[/itex] 1/7 (6u^2 +3uv + 4 - 6u - 3v) *(3.5) *dvdu =

1/2 [itex]\int[/itex][itex]\int[/itex][itex]_{R}[/itex]6u^2 + 3uv -6u - 3v + 4 dv du =
1/2 [itex]\int[/itex]6u^2v + 3uv^2 -3v^2/2 + 4v (evaluating v from 0 to 2-2u) dv

Using wolfram alpha to evaluate (got tired of doing it by hand like 6 times XD) :
1/2 [itex]\int[/itex]24u^3 + 30u^2 + 8u + 2 du from u = 0..1
1/2 * (24u^4/4 + 30u^3/3 + 8u^2/2 + 2u) from 0 to 1
1/2 (6 + 10+4+2) = 11

The answer should be 15/4 though. I'm not sure what I did wrong, maybe I missed something. Can anyone see what I did wrong in this problem? Thanks much.
 
thepatient said:
Hmm.. But the problem states "...part of the plane 6x + 3y + 2z = 6 with x ≥ 0, y ≥ 0, and z ≥ 0, with the outward unit normal n pointing in the positive z direction." That doesn't mean that the flux is the net flux of the closed surface, but the flux of only that portion of the plane?
Yes, I think so, otherwise it wouldn't specify which direction the normal points in.EDIT: It turns out you get the same answer, 7/4, because the vector field vanishes on the other three faces, so there's no flux through them.
 
thepatient said:
If I remember correctly, since the surface Σ is closed, the divergence theorem applies right?

∫∫Rf*n*dσ = ∫∫∫div(f) dV

x is from 0 to one
y is from 0 to the line -2x + 2
z is from 0 to the plane 3 - 3x -3y/2

f(x, y, z) = x2i + xyj + zk
div(f) = d/dx i + d/dy j + d/dz k * x^2i + xyj + zk
= 2x + x + 1 = 3x+1

∫∫∫div(f) dV = ∫∫∫3x+1 dzdydx
∫∫3xz+z | z = 0..3-3x-3y/2 dy dx
∫∫3x(3-3x-3y) + 3-3x-3y -[ 0] dydx
You dropped the factor of 1/2 on the y term.

I would pull the (3x+1) out front since it's a constant for the y and z integrations: [tex]\int_0^1 (3x+1)\int_0^{2(1-x)} \int_0^{3-3x-3/2 y}dz\,dy\,dx[/tex]It'll simplify the algebra a bit. After you integrate over z, you'll have[tex]\int_0^1 (3x+1)\int_0^{2(1-x)} (3-3x-3/2 y)\,dy\,dx = \int_0^1 3(3x+1)\int_0^{2(1-x)} [(1-x)-y/2]\,dy\,dx[/tex]
If you keep things in terms of 1-x, I think the result of the y integral will be pretty simple. To do the x integration, you'll just have to bite the bullet and multiply it out, or try the substitution u=1-x.
∫∫9x -9x^2 - 9xy +3 -3x - 3y dy dx
∫∫6x - 9x^2 -9xy +3 -3y dy dx
∫6xy - 9x^2y - 9xy^2/2 + 3y - 3y^2/2 | y = 0..-2x+2 dx
∫6x(-2x+2) - 9x^2(-2x+2) -9x(-2x+2)^2/2 +3(-2x+2) -3(-2x+2)^2/2 dx
∫-12x^2 + 12x +18x^3-18x^2-(9x/2) (4x^2 -8x + 4) -6x + 6 -3(4x^2 - 8x + 4) dx=
∫-12x^2 + 12x + 18x^3 -18x^2 -36x^3/2 + 72x^2/2 -36x/2 + 6 -12x^2 +24x - 12 dx=
-12x^3/3 + 12x^2/2 + 18x^4/4 -18x^3/3 -36x^4/8 + 72x^3/6 - 36x^2/4 +6x -12x^3/3 +12x^2 -12x from x =0..1
-12/3 + 12/2 + 18/4 -18/3 -36/8 +12 -9 +6 -4 + 12-12 =1:\ I'm frustrated with this problem.
Either way, you should get 7/4 for the answer, so I wouldn't bother beating your head against the wall to try get 15/4. That answer is apparently wrong.
 
You're right. I should just try the problems on my other calculus book. This book was a free online book which I think didn't have much revisement. Thanks much for your help. :]