next up previous
Next: Flux Through Cylinders Up: The flux Integral Previous: Introduction

Flux through Surfaces defined by a Function

Now we need to find $\vec{n}$ and dS. If the surface S is defined by a function, z=g(x,y) we can obtain a nice expression for $\vec{n}$ and dS. For the normal vector to S at the point (x0,y0,g(x0,y0)) we can use the normal vector to the tangent plane. Remember that the tangent plane to g(x,y) at the point (x0,y0) is given by:

\begin{displaymath}
z=g_{x}(x_{0},y_{0})\left( x-x_{0}\right) +g_{y}(x_{0},y_{0})\left(
y-y_{0}\right) +g(x_{0},y_{0}) 
\end{displaymath}

or

-gx(x0,y0)x-gy(x0,y0)y+z=c

where c=-gx(x0,y0)x0-gy(x0,y0)y0+g(x0,y0). A normal vector to the tangent plane is

\begin{displaymath}
\vec{n}=-g_{x}(x_{0},y_{0})\hat{\imath}-g_{y}(x_{0},y_{0})\hat{j}+\hat{k} 
\end{displaymath}

If we normalize $\vec{n}$ to get a unit normal vector we get

\begin{displaymath}
\hat{n}=\frac{-g_{x}(x_{0},y_{0})\hat{\imath}-g_{y}(x_{0},y_...
 ...t{k}}{\sqrt{g_{x}(x_{0},y_{0})^{2}+g_{y}(x_{0},y_{0})^{2}+1}} 
\end{displaymath}

So the normal vector as a function of $\left( x,y\right) $ is

\begin{displaymath}
\hat{n}=\frac{-g_{x}(x,y)\hat{\imath}-g_{y}(x,y)\hat{j}+\hat{k}}{\sqrt{g_{x}(x,y)^{2}+g_{y}(x,y)^{2}+1}} 
\end{displaymath}

The Area of the jth patch can be obtained by finding the vectors $\vec{u}$, $\vec{v}$ which define the edges of the patch (one parallel to the xz-plane and one parallel to the yz-plane????). The area of the patch is given by $\left\Vert \vec{u}\times \vec{v}\right\Vert $. If we employ the tangent plane approximation of g(x,y)we can see that $\vec{u}$ is the vector from (x0,y0,g(x0,y0)) to $(x_{0}+\Delta
x,y_{0},g(x_{0},y_{0})+g_{x}(x_{0},y_{0})\Delta x)$ or

\begin{displaymath}
\vec{u}=\Delta x\hat{\imath}+g_{x}(x_{0},y_{0})\Delta x\hat{k} 
\end{displaymath}

Similarly $\vec{v}$ is the vector that points from (x0,y0,g(x0,y0)) to the point $(x_{0},y_{0}+\Delta
y,g(x_{0},y_{0})+g_{y}(x_{0},y_{0})\Delta y)$ or

\begin{displaymath}
\vec{v}=\Delta y\hat{j}+g_{y}(x_{0},y_{0})\Delta y\hat{k} 
\end{displaymath}

Taking the cross product of $\vec{u}$ and $\vec{v}$ gives

\begin{displaymath}
\vec{u}\times \vec{v} =\left\vert 
\begin{array}
{lll}
\ha...
 ... \Delta y & g_{y}(x_{0},y_{0})\Delta y
\end{array}
\right\vert\end{displaymath}

\begin{displaymath}
=-g_{x}(x_{0},y_{0})\Delta x\Delta y\hat{\imath}-g_{y}(x_{0},y_{0})\Delta
x\Delta y\hat{j}+\Delta x\Delta y\hat{k}
\end{displaymath}

Finally, the area of the jth patch is

\begin{displaymath}
\Delta S_{j}=\left\Vert \vec{u}\times \vec{v}\right\Vert =\s...
 ...x}(x_{0},y_{0})^{2}+g_{y}(x_{0},y_{0})^{2}+1}\Delta x\Delta y 
\end{displaymath}

In the limit that $\Delta S_{j}\rightarrow dS$ this becomes

\begin{displaymath}
dS =\sqrt{g_{x}(x,y)^{2}+g_{y}(x,y)^{2}+1}dxdy\end{displaymath}

\begin{displaymath}
=\sqrt{g_{x}(x,y)^{2}+g_{y}(x,y)^{2}+1}d\end{displaymath}

When we put these together we get

\begin{displaymath}
d\vec{A} =\hat{n}dS=\frac{-g_{x}(x,y)\hat{\imath}-g_{y}(x,y)...
 ...^{2}+g_{y}(x,y)^{2}+1}}\sqrt{g_{x}(x,y)^{2}+g_{y}(x,y)^{2}+1}dA\end{displaymath}

 
 \begin{displaymath}
d\vec{A} =\left( -g_{x}(x,y)\hat{\imath}-g_{y}(x,y)\hat{j}+\hat{k}\right)
dA

\end{displaymath} (3)
So we've represented the integral over the surface in terms of an integral over the region R in the xy-plane over which S sits.

\begin{displaymath}
\int_{S}\vec{F}\cdot d\vec{A}=\int_{S}\vec{F}\cdot \hat{n}dS...
 ...x,y)\hat{\imath}-g_{y}(x,y)\hat{j}+\hat{k}\right)
\right] dA 
\end{displaymath}

Where R is the ''shadow'' of S on the xy-plane.

Notes:

Example

1.
Calculate the flux of the vector field $\vec{F}=-z\hat{\imath}+\frac{z^{3}}{y}\hat{j}+x\hat{k}$ through the upper hemisphere of x2+y2+z2=4 that lies above the rectangle $-1\leq x\leq 1$, $\frac{1}{2}\leq y\leq 1$, oriented upward. Since the surface consists of the upper hemisphere we can solve the equation of the sphere for $z=g(x,y)=\sqrt{4-x^{2}-y^{2}}$, so the surface is the graph of a function and we can use % latex2html id marker 1057
$\left( \ref{dA}\right) $ to find $d\vec{A}$.

\begin{displaymath}
d\vec{A}=\left( \frac{x}{\sqrt{4-x^{2}-y^{2}}}\hat{\imath}+\frac{y}{\sqrt{4-x^{2}-y^{2}}}\hat{j}+\hat{k}\right) dxdy 
\end{displaymath}

The flux integral is

\begin{displaymath}
\int_{S}\vec{F}\cdot d\vec{A}=\int_{R}\left( \frac{-xz}{\sqr...
 ...2}-y^{2}}}+\frac{yz^{3}}{y\sqrt{4-x^{2}-y^{2}}}+x\right) dydx 
\end{displaymath}

Where R is the rectangle $-1\leq x\leq 1$, $\frac{1}{2}\leq y\leq 1$. Now substitute $z=\sqrt{4-x^{2}-y^{2}}$ to eliminate the z dependence in the integral. The integral reduces to

\begin{displaymath}
\int_{-1}^{1}\int_{0.5}^{1}\left( -x+4-x^{2}-y^{2}+x\right) dydx\end{displaymath}

\begin{displaymath}
=\int_{-1}^{1}\int_{0.5}^{1}\left( 4-x^{2}-y^{2}\right) dydx=\frac{37}{12}
\end{displaymath}


next up previous
Next: Flux Through Cylinders Up: The flux Integral Previous: Introduction
Vector Calculus
8/21/1998