Lecture 3
Lecture 3
In the last lecture we had defined divergence of a vector field and had obtained an expression for the
divergence in the Cartesian coordinates. We also derived the divergence theorem which connects
the flux of a vector field with the volume integral of the divergence of the field.
Physically, the divergence, as the name suggests is a measure of the amount of spread that the field has
at a point.
For instance, in the figure above, the vector field shown to the left has a positive divergence while that
to the right has a negative divergence. In electrostatics, we will see that the field produced by a
positive charge has positive divergence while a negative charge produces an electrostatic field with
negative divergence.
Divergence, curl etc. were extensively used in fluid dynamics from which a lot of nomenclature have
arisen. Let us consider a fluid flowing through an elemental volume of dimension
with its sides oriented parallel to the Cartesian axes. In the figure below, we show only the ycomponent of the velocity of the fluid entering and leaving the elemental volume. Let the density of
the fluid at
be
) and the y-component of the velocity be . We define a vector
. The mass of the fluid flowing into the volume per unit time through
the left face which has an outward normal
is given by
velocity. We can write similar expressions for the flow in the x and z directions. The net increase in
mass per unit time is
element
defined above.
Another of of talking about the net increase in mass is to realize that since the volume is fixed, the
.
increase in mass is due to a change in the density alone. Thus the rate of increase of mass is
Equating these two expressions, we get what is known as the equation of continuity in fluid
dynamics,
To see what this equation implies, consider, for example a vector field given by
field has been plotted using Mathematica (see figure)
. The
The size of the arrow roughly represents the magnitude of the vector. The divergence is given by
. In
the first and the third quadrants (x,y both positive or both negative) divergence is positive. One can
see that in these quadrants, if you take any closed region the size of the arrows which are entering
the region are smaller than those leaving it. Thus the density decreases, divergence is positive.
Reverse is true in the even quadrants.
In the figure to the right, the force field
has zero divergence. If you take a closed region in this
figure, you find as many vectors are getting in as are going out. The field is a solenoidal (zero
divergence) field.
Divergence Theorem Examples :
Recall divergence theorem :
surface defining the enclosed volume. As an example consider the surface integral of the position
vector over the surface of a cylinder of radius a and height h.
Evaluating the surface integral by use of the divergence theorem is fairly simple. Divergence of position
. Thus the volum integral of the
divergence is simply three times the volume of the cylinder which gives
Direct calculation of the surface integral can be done as follows. For convenience, let the base of the
cylinder be in the x-y plane with its centre at the origin.
There are three surface of the cylinder, a top cap, a bottom cap and the curved surface. For the top cap,
the normal vector is along the direction, so that
. On the top surface z is constant and is
given by z=h. Thus
. For the bottom cap, is along the
direction so that
. However, the value of z on this surface being zero, the flux vanishes. We are now left
with the curved surface for which the outward normal is parallel to the x-y plane. The unit vector on
this surface is
. Thus
contribution from the top and the bottom face, the surface integral works out to
obtained from the divergence theorem
, as was
We have seen that the divergence of a vector field is a scalar field. For vector fields it is possible to define
an operator which acting on a vector field yields another vector field. The name curl comes from
circulation which measures how much does a vector field curls about a point.
Consider an open surface of the type shown something like an inverted pot with a rim. We wish to
calculate the surface integral of a vector field defined over this surface.
Let us divide the surface into a large number of small segments and calculate the line integral over the
boundary of such an elemental surface. One such elemental surface with boundary is shown. If we
define the bounding curve of this surface in the counterclockwise direction, the normal to such a
surface will be outward. If we consider each such segment and take the line integral over the
bounding curve
, the contribution to the line integral from adjacent regions will cancel because
the line integral from the boundary are traversed in opposite sense (look at the black and the red
arrows on the boundary of two segments). Considering all such segments, we will be only left with
the uncompensated line integral at the rim. This is illustrated in the following figure :
The line integral on the boundary for ABCD + line integral for
The boundary of CDEF is equal to the line integral for ABEF
because the integral is traversed along CD in the former and
along DC in the latter.
Thus we have,
(1)
The quantity in the parenthesis in the last expression is defined as the curl of the vector field when the
limit of the elemental surface goes to zero. Curl being a vector, its direction is specified as the
outgoing normal to the surface element.
(2)
.It may be noted that because the definition is valid in the limit of the surface area going to zero, it is a
point relationship. Using this definition, we can write the previous equation (1) as
This equation relates the surface integral of the curl of a vector field with the line integral of the vector
field and is known as Stokes Theorem.
Tutorial
1. The geometry of the cube along with the direction of surface normalsare shown in the figure.
Consider the base of the cube which is the plane z=0. On this face
Since the normal is
along the -z direction flux from this face is zero. Similarly, the flux from the other two faces
which meet at the origin are also zero. Consider the top face where z=1. On this face
. The normal is in the +z direction, so that the flux is
Likewise, the
flux from the other two faces are also each. The total flux, therefore, is 3/2. The divergence of
the field
is
the surface of the sphere is along the radial direction and is given by
, where R=1 is
The first term gives zero because of vanishing of the integral over . We are left with
3. Let the base of the cylinder be at z=0 and the top at z=h. The origin is at the centre of the base.
The cylinder has three surfaces. For the bottom surface, the direction of the normal is along
and on this surface z=0. The surface integral for this surface is
For the top
surface, the normal is along
For
the curved surface the direction of the normal is outward radial direction in the x-y plane which
is
4. Geometrically a torus is obtained by taking a circle, say in the x-z plane and rotating it about the
z-axis to obtain a solid of revolution. Let us define the mean radius of the torus to be
and the radius of the circle which is being revolved about the z-axis to be
The position
Consider the coordinate of an arbitrary point on the circle which is in the x-z plane. Let the
position of the point make an angle with the x-axis. The coordinate of this point is
8
above expressions. A surface element on the torus is then obtained by the area formed by an
The area
arc obtained by incrementing by d and the arc formed by incrementing by
element is therefore given by the cross product
surface is directed along the direction of the cross product.). The partial derivatives are given by
Thus
(the order of the cross product determines the outward normal). Thus , substituting
The second and the third term in the integral vanish, the remaining two terms give
, which
makes the total contribution to the surface integral as
.
However, the problem is straightforward if we apply the divergence theorem. The divergence of
the position vector is 3. Thus by divergence theorem, the surface integral is 3 times the volume
of the toroid. The volume of the toroid is rather easy to calculate if we note that if we cut it
along a section, the toroid becomes a cylinder of radius r and length
. Thus the volume of
the toroid is
. Thus 3 times the volume is
direct evaluation of the surface integral.
10
=
z
F = xi + yj + zk .
which gives
Parameterize
The flux through the slant surface is
. The
. Adding up all
the contributions, the total flux from the closed surface is zero. This is consistent with the
fact that the divergence of the field is zero.
2. There are six faces. For the face at x=0, since the surface is directed along
direction, the
The face at x=1 gives +1/2. The faces at y=0 and that
surface integral is
at z=0 gives zero because the field is proportional to y and z respectively. The contribution
to flux from y=1 is 1 and that from z=1 is 3/2. Adding, the flux is 5/2 units. This can also be
done by the divergence theorem. Divergence of the field is 2x+3y, so that the volume
integral is
3. The divergence of the given vector field is 3(x 2 + y 2 + z 2 ) =
3r 2 . Thus, by divergence
R
12 R 5
. We can show this result by direct
theorem, the flux is
3 r dV 3.4
r dr
=
=
5
0
2
+ jy
+ kz
ix
, so
R
x 4 + y4 + z 4
rdS
=
dS . This integral can be conveniently evaluated
R
11
z4
3.2
seen to be 3 dS =
(R cos )4R2 sin d . (Note that we decided to do the
R
R 0
integral involving z4 rather than x4 or y4 because the azimuthal integral gives 2 in this case.
The integral is easy to perform with the substitution = cos , d = sin d , which
1
12 R 5
gives the flux to be 6 R d =
.
5
1
5
4. The divergence of the field is 3. The flux, therefore, is 3 times the volume of the cone which
is (1 / 3) R 2h = / 3 . The flux is thus . The direct calculation of the flux involves two
surfaces, the slant surface and the cap, as shown in the figure. The cap is in the xy plane and
has an outward normal n = k . The flux
=
F
ndS
zdxdy= dxdy=
(because on
the cap z=1 and the cap is a disk of unit radius). Thus it remains to be shown that the flux
from the slanted surface vanishes. At any height z, the section parallel to the cap is a circle
of radius z. Since, the height and the radius of the cap are 1 each, the semi angle of the cone
is 450.
n
y
x
Thus the normal to the slanted surface has a component 1 / 2 along the z direction and
1 / 2 in the x-y plane. The component in the xy plane can be parameterized by the
cos i + sin j k
. The area element can
azimuthal angle and we can write n(z, ) =
2
be written as (zd )( 2dz ) , the factor
12
=
F ndS
Using x z=
cos , y z sin , this
d zdz x cos + y sin z .=
=
F
ndS
zdxdy= 2dxdy=
8 . Thus
the net flux is zero. (You can also try to get this result directly as done in problem 4, where
we showed that the flux from the curved surface is zero).
13