1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Homework Help: Stokes theorem problem

  1. Sep 29, 2007 #1


    User Avatar
    Gold Member

    1. The problem statement, all variables and given/known data
    Let C be the closed curve that goes along straight lines from (0,0,0) to (1,0,1) to (1,1,1) to (0,1,0) and back to (0,0,0). Let F be the vector field [tex]F = (x^2 + y^2 + z^2) i + (xy + xz + yz) j + (x + y + z)k. Find \int F \cdot dr[/tex]

    By Stokes Theorem, I know that I can transform this into a curl (instructions say don't evaluate the line integral directly).

    So I know I get a form of square that goes CCW. Projected on the xz-plane, it is simply a square (but the y-coordinate of the right end is 1).

    I think the plane is described by x = z, since there is no dependence on y. So x-z = 0

    Anyways, I take the curl of F to get

    [tex]F \cdot dr = (1-x-y) i + (1-2z) j + (y+z-2y) k.[/tex]

    I take the dot product of this with the plane z - x = 0 to get

    [tex] \nabla f = -i + k [/tex]

    and dot that with F to get
    [tex]F \cdot \nabla f = x + y -1 + y + z - 2y = x + z -1 [/tex]

    so then

    [tex]\int_0^1 \int_0^1 (x + z -1) dx dy.[/tex] Since x = z...

    [tex]\int_0^1 \int_0^1 dx dy = \int_0^1 x^2 - x dy = \frac{1}{3} - \frac{1}{2} = - \frac{1}{6}[/tex]

    am I doing this correctly?
  2. jcsd
  3. Sep 29, 2007 #2


    User Avatar
    Science Advisor

    Part of your LaTex didn't come through. How do you take the dot product of a vector with a plane?

    And what do you mean by "dot that with F"?

    You have [tex]F \cdot \nabla f [/tex]

    That confuses me greatly. You have a function "F" but say nothing about "f". If you meant [itex]F \cdot \nabla F[/itex], that is not at all what you want.
    Stokes theorem has
    [tex]\int\int \nabla \times \vec{F}\cdot \vec{n}dA[/tex]
  4. Oct 1, 2007 #3


    User Avatar
    Gold Member

    Okay, here you have the dot product of a vector with the gradient of the plane:
    (example 2). But this could be confusing, and there is an alternative way to express Stokes Theorem after the dot product has been taking.

    This is Stokes Theorem as a surface integral. (the gradient approach to the surface integral is at http://tutorial.math.lamar.edu/Classes/CalcIII/SurfIntVectorField.aspx )

    As we know,

    [tex]\int\int \nabla \times \vec{F}\cdot \vec{n}dA = \int\int F \cdot dS = \int\int -P \frac{\partial g}{\partial x} - Q \frac{\partial g}{\partial y} + R[/tex]

    [tex]\nabla \times F(x,y,z) = P i + Q j + R k[/tex] and [tex] g = f(x,y) = z[/tex]

    The equation of my plane should be z = f(x,y) = x, right? (the plane is f, and determined by "Let C be the closed curve that goes along straight lines from (0,0,0) to (1,0,1) to (1,1,1) to (0,1,0) and back to (0,0,0).")

    in that case, [tex]\frac{\partial g}{\partial x} = 1, \frac{\partial g}{\partial y} = 0[/tex]

    So correcting my arithmetic..

    [tex]F \cdot \nabla f = -(1 - x - y) + y + z - 2y dA[/tex]
    = [tex]\int_0^1 \int_0^1 -1 + x + z [/tex]

    We take z = x. Is this an appropriate choice? If so

    [tex]\int_0^1 \int_0^1 2x - 1 dx dy = \int_0^1 x^2 - x (1,0) = \int_0^1 0 ....[/tex]

    what am I doing wrong here?
    Last edited: Oct 1, 2007
  5. Aug 6, 2009 #4
    You set [tex]g = f(x,y) = z[/tex], so your D will be the projection of graph g at x-y plane.
    You can substitute [tex]z=x[/tex] at there. I got zero as well for this question.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook