Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Math 21a
Spring, 2009
A surface integral is
ZZ
ZZ
f (x, y, z) dS =
D
z .................
..
...
.
...........
..
...
...
...
...
.
.
............................
.
.
.
.
.
.
.
.
.
.
.
.................
.................
.
.............
.
.
.
.
.
...............
........
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.. .............. ..
.. ............. ..
...................
...................
.
....................
.
.
.
.
.
..
.
........................................................
.
.
.
.
.
.........................................................
Suppose F is a continuous vector field on an oriented surface S with unit normal vector n. The surface
integral of F over S is
ZZ
ZZ
ZZ
F dS =
F n dS =
F (ru rv ) du dv
S
RR
Evaluate the surface integral S F dS, where F = yi xj + zk and S is the part of the sphere
x2 + y 2 + z 2 = 4 in the first octant with inward orientation.
RR
z = 4 x2 y 2 that lies above the unit square [0, 1] [0, 1] with the downward orientation.
RR
paraboloid z = 4 x2 y 2 that lies above the unit disk (centered at the origin) with upward
orientation.
RR
Evaluate the surface integral S F dS, where F = hz, x, yi and S is the full unit hemisphere
(including the base!) on and above the xy-plane (so x2 + y 2 + z 2 = 1 plus a disk) with the
outward orientation.
Here we use the parameterization r(x, y) = hx, y, 1 12 x 21 yi. From this we find that
i j k
rx ry = 1 0 12 = 12 , 12 , 1 .
0 1 1
2
Therefore
ZZ
ZZ
(1 + z) dS =
1
1
1+ 1 x y
dA.
2
2
The region D in our parameter space (the xy-plane) need to cover our surface is the triangle
D = {(x, y) : x 0, y 0, x + y 2} ,
so we can write our limits as follows:
Z
ZZ
(1 + z) dS =
0
Z
=
0
2x
Z
0
1
1
2 x y dy dx
2
2
1 2
x 2x + 3
4
dx
1
8
23 22 + 3 2 = .
12
3
r(, ) = a sin() cos(), a sin() cos(), a cos() .
One could then compute r r , but it is easier to just remember that weve done this before
and the answer is:
r r = a2 sin() sin() cos(), sin() cos(), cos() .
(See, for example, page 869 of the text.) The coefficient in front is simply the coefficient 2 sin()
from the spherical volume element (with = a) while the vector is simply the unit vector in the
direction hx, y, zi (the radial vector, which is perpendicular to the tangent plane to the sphere).
Notice that the orientation is specified to be the inward normal, so we actually want r r =
r r .
In any case, with this we proceed. In spherical coordinates (with = a = 2 in this case), our
vector field is
F = hy, x, zi = h2 sin() sin(), 2 sin() cos(), 2 cos()i .
Thus
F (r r )
= 4 sin() 2 sin() sin(), 2 sin() cos(), 2 cos() sin() cos(), sin() cos(), cos()
= 8 sin() cos2 ().
We integrate this over the domain {(, ) : 0 /2, 0 /2}, so we have
ZZ
Z /2 Z /2
8 sin() cos2 () d d
F dS =
0
Z
= 4
/2
sin() cos2 () d
/2
1
4
3
= 4 cos ()
= .
3
3
0
Another approach is to use the parameterization by x and y
p
r(x, y) = x, y, 4 x2 y 2
then integrating over the quarter circle in the xy-plane. Lets see how this goes:
i j
*
k
+
y
y
x
ry rx = 0 1 4x2 y2 = p
, p
, 1 .
4 x2 y 2
4 x2 y 2
1 0 x
2
2
4x y
p
4 x2 y 2 dy dx
F dS =
S
0
2
Z
=
0
=
4
as before.
4x2
p
4x2
2
y
4
x
y
4 x2 y 2 +
sin1
dx
2
2
4 x2 0
Z
0
16
4
4 x2 dx =
= ,
4 3
3
ZZ
=
2x2 2y 2 2z dx dy.
0
1
(8) dx dy = 8 Area(D) = 8.
=
0
One way to do this is to use the parameterization r(x, y) = hx, y, 4 x2 y 2 i, so rx = h1, 0, 2xi,
ry = h0, 1, 2yi, and so
i j
k
1
0
2x
rx ry =
= h2x, 2y, 1i.
0 1 2y
Thus we can write
F (rx ry ) = hx, y, 2x + 2yi h2x, 2y, 1i = 2x2 + 2y 2 + 2x + 2y.
Hence our flux is
ZZ
ZZ
2x2 + 2y 2 + 2x + 2y dx dy,
F dS =
S
where D is the unit disk. Thus we make the change to polar coordinates, where 2x2 + 2y 2 + 2x +
2y = 2r2 + 2r (cos() + sin()) and dx dy = r dr d. We get
Z 2 Z 1 h
ZZ
i
2r2 + 2r (cos() + sin()) r dr d
F dS =
0
0
2 Z
=
0
i
2r3 + 2r2 (cos() + sin()) dr d
0
2
1 2
=
+ (cos() + sin()) d
2 3
0
2
1
2
= .
= + (sin() cos())
2
3
0
Z
Another approach is to use polar coordinates to parameterize the surface from the very beginning.
That is, we could use the parameterization r(r, ) = hr cos(), r sin(), 4 r2 i, in which case
rr = hcos(), sin(), 2ri and r = hr sin(), r cos(), 0i, so
i
j
k
sin() 2r = h2r2 cos(), 2r2 sin(), ri.
rr r = cos()
r sin() r cos() 0
In these coordinates our vector field is F = hr cos(), r sin(), 2r(cos() + sin())i, and therefore
our flux is
ZZ
Z 2 Z 1
F dS =
hr cos(), r sin(), 2r(cos() + sin())i h2r2 cos(), 2r2 sin(), ri dr d
S
0
2
=
0
h
i
3
2
2r + 2r (cos() + sin()) dr d,
S1
where S1 is the hemisphere (with outward-pointing normal) and S2 is the unit disk in the xy-plane
(with downward-pointing normal).
The integral over S1 is very similar to the previous problem. Well use the first parameterization
of Problem 3, namely
r(, ) = sin() cos(), sin() sin(), cos() .
This gives us an outward-pointing normal
r r = sin() sin() cos(), sin() sin(), cos() ,
so
F (r r )
= cos(), sin() cos(), sin() sin() sin2 () cos(), sin2 () sin(), sin() cos()
= sin2 () cos() cos() + sin3 () sin() cos() + sin2 () cos() sin().
Thus
ZZ
F dS
S1
/2
=
0
sin2 () cos() cos() + sin3 () sin() cos() + sin2 () cos() sin() d d
2
Z /2
1 2
2
3
2
=
sin () cos() sin() sin () sin () sin () cos() cos()
d
2
0
=0
= 0.
The second integral is even simpler. Here S2 is the unit disk in the xy-plane, with the unit
normal n = k (straight down). Thus
ZZ
ZZ
ZZ
F dS =
z, x, y 0, 0, 1 dA =
y dA,
S2
where D is the unit disk in the xy-plane. We use polar coordinates to compute this integral:
ZZ
ZZ
Z 1 Z 2
F dS =
y dA =
r sin() r d dr
S2
Z
=
0
2
r cos() dr = 0.
2
F dS = 0 + 0 = 0.
F dS +
F dS =
S
ZZ
ZZ
S1
S2
Well see a simpler way of computing this integral on Wednesday when we learn about the
Divergence Theorem.