Scribd
Upload
249 views

Note On Vector Integration

1) The document discusses vector calculus concepts like line integrals, surface integrals, and their applications to work, flux, and vector fields. 2) It defines line integrals as the dot product of a vector field and infinitesimal displacement, integrated over a curve. Surface integrals are defined similarly as a double integral over a surface. 3) The document explains that line integrals can depend on path, but are path-independent when the vector field is conservative or curl-free. Surface integrals give flux across a surface.

Uploaded by

Rizwan 106
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
249 views

Note On Vector Integration

1) The document discusses vector calculus concepts like line integrals, surface integrals, and their applications to work, flux, and vector fields. 2) It defines line integrals as the dot product of a vector field and infinitesimal displacement, integrated over a curve. Surface integrals are defined similarly as a double integral over a surface. 3) The document explains that line integrals can depend on path, but are path-independent when the vector field is conservative or curl-free. Surface integrals give flux across a surface.

Uploaded by

Rizwan 106
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

College of Engineering Pune.

(MA-) Vector Calculus & Partial Differential Equations


(Academic Year: 2017-18)
Note on Vector Integral Calculus

It is known that work is done in moving a body from one point A to


another point B and that, in general, will depend on the path chosen to
move from A to B. How can this work done be calculated? Since force is
a vector quantity and intuitively we know that to find this we will have
to find the dot product of force with displacement and that if force is
changing with the position of the point along the path then some sort
of integration will only give us the answer.
Similarly if we want to find the flux across a surface S when a fluid
flow has velocity v then again intuitively we know that the flux will be
only in the direction of normal to S and hence the dot product with
unit normal vector will give the flux. And once again if the velocity is
changing at every point on the surface then we will find the flux over
a small elementary surface area around every point and then integrate
over the entire surface.
In short we need to know how to integrate vector valued functions.
Integral of a vector valued function can be defined as antiderivative of a
vector valued function exactly in the same way as we define the integral
of a scalar valued function. Similarly, one must also remember that the
integral of a vector valued function, when ever it exists, is once again,
in general, a vector valued function.
In our course we are interested in applications of vector integration
and hence we do not go into the details of the integral of a vector valued
function but straight away define the lineR integral of a vector valued
function to be a scalar quantity as follows: C F · dr, where the integral
is over the curve C (read as: line integral of F dotted with dr along the
curve C)
Depending on how the curve R C is given and what function F is given
we evaluate the integral as F1dx + F2dy + F3dz with limits of inte-
gration appropriately chosen as per the cartesian curve C,
1
or as F(t) · dr
R
dt dt with appropriate limits for parameter t as obtained
from the parametric curve C : r(t). (Note that the value of the integral
does not depend on the choice of equation chosen to represent the curve
C)
Note that this is a generalization of line integral of any scalar valued
function along the curve C with the scalar valued function considered
to be a vector with components in all except one direction to be zero,
which is itself a generalization of our usual definite integral where we
are integrating a function along an interval [a, b] on the x−axis.
Observe that line integral, in general, will depend on the integrand
as well as the line (curve) of integration. This can be shown by many
examples from the question bank. Before going any further we make an
assumption that all our paths will be piecewise smooth and the functions
continuously differentiable.
Line integral which does not depend on the curve C but depends
only on the initial point A and final point B on the curve is called path
independent. Since there are infinitely many paths between any two
given points, it is not possible to check if a given line integral is path
independent or not by using this definition. Hence we have theorems
(1,2,3 in 10.2 of text) which give us necessary and sufficient conditions
for a line integral to be path independent. These say that either F is
gradient of some scalar f which is equivalent to saying grad(f )·dr = df
and hence the line integral is, by Fundamental theorem of Calculus,
f (B) − f (A) which is equivalent to line integral around any closed
curve in the domain is zero. Moreover when grad(f ) · dr = df , we say
that F · dr is the exact differential of f and in this case the curl of F
is zero. Converse, however, is not true. That is there may be a vector
valued function whose curl is zero but it has no potential function. (See
solved example in the book) This happens when the domain of F is not
simply connected.
Thus we have that line integral of F is path independent iff (i) F = ∇f
or (ii) line integral around any closed curve is zero.
2
Observe that (ii) is again similar to definition and we cannot check for
every closed curve, there being infinitely many closed curves. But it is
very useful to find line integral around some closed curve, if (i) is checked
to be true.
In case curl is not zero then we know for sure that the line integral around
at least one closed path is not zero. But in addition to curl being zero if
the domain of F is simply connected then F has a potential and hence
the line integral is path independent.
Caution: This does not necessarily mean that if line integral of F is
path independent then the domain of F is simply connected. In other
words, if curl of F is zero and the domain of F is not simply connected
then line integral need not be path dependent.
In some cases it may either be very difficult or impossible to get the value
of a line integral around a closed curve in a closed form. For example if
the closed curve is made up of n pieces (as in a square or a n-gon) then
we will need to break up the line integral into n pieces and evaluate each
integral and add them up or in case the integrand function cannot be
integrated in a closed form using empirical methods. In such a situation
we can evaluate this line integral by converting it into a double integral
using the Stoke’s theorem or Green’s theorem.(Before moving ahead you
must solve examples 1,2,3 from QB3 and try to see if you can evaluate
all the line integrals in 12 and 23)
Line integral can be interpreted as the work done by a force F in
displacing a body from point A to point B along some path C or in
potential theory if a field has potential then the line integral is simply
the difference in the potentials. (Try problems 13, 14 now!)

Having seen that line integral of a vector valued function along a


curve C is defined as a scalar, we extend this to the surface integral of
a vector valued function over a surface S. For this we will first have to
understand surfaces in two and three dimensions and how they are best
represented. You must have realized (after solving all the above prob-

3
lems) that working with parametric curves is much simpler as compared
to cartesian curves (polar curves is still more difficult!) and hence to
2 2 2 2
understand the p surface of a sphere by x + y + z = c or that of a
cone by z = x2 + y 2 may look simpler (due to familiarity) but will be
much more difficult to work with as compared with their corresponding
parametric equations.
Parametric surfaces will clearly require two parameters since surfaces
are two dimensional and are characterized by having an area. Thus in
general, just as r(t) represents a curve, r(u, v) will represent a surface.
For example r(u, v) = [acosu, asinu, v] represents a cylinder with axis
as z-axis and base circle x2 + y 2 = a2.
∂r ∂r
It can be proved easily that if r(u, v) is a surface and ∂u × ∂v is non
zero then it represents a normal vector to S at any point P. Thus
n = | ∂r ×1 ∂r | ( ∂u
∂r ∂r
× ∂v ) is a unit vector normal to S at P. We have al-
∂u ∂v
ready proved that if S represents the level surface of some function f
then n = ∇f /|∇f |. More over the vectors ru and rv span a plane
which is precisely the tangent plane to the surface S at P. (Theorem 1
in 10.5).
Now we define the surface integral of a vector valued function as
Z Z Z Z
F · ndA = F(r(u, v) · N(u, v)dudv
S R
Here N = |N|n and |N| is the area of parallelogram with adjacent
sides ru and rv . Hence ndA = Ndudv is an elementary area of S. This
integral is also called as the flux integral since it gives the flux across S
when F = ρv. (Now do problems 5,6,7, 17,18)
You can solve by following simple steps like first find the parametric
representation of S, then identify the parameters u and v, then find
the partial derivatives to get the vectors ru and rv which will then give
you the normal vector N and then integrate using double integral with
appropriate limits for u and v.
Did you notice that we said ”a” unit normal to S? This is because −n
is also a unit normal to S. What we can do is associate an orientation
4
for the surface S according to the orientation of the boundary curve of
S and then we get an orientable surface which will have a unique normal
which is according to the right hand rule for the three vectors u, v, n.
Similar to evaluating a single integral over some interval can be gen-
eralized to evaluating a single integral over any curve in space, we can
generalize evaluating a double integral over some region in a plane to
evaluating a double integral over any surface in space i.e. we want to
integrate a scalar valued function over any surface. This is defined as
Z Z Z Z
G(r)dA = G(r(u, v))|N(u, v)|dudv
S R
where dA is an elementary area of S and there is no need to consider
any orientation of S. In particular if the scalar function G is 1 we will
get the area of S, if G is density then we get mass etc.(Do problem 19)
Coming back to our comment above that we can evaluate a line inte-
gral around a closed curve by using Stoke’s theorem, first let us under-
stand what it says.
When you have a closed curve it will enclose a surface S and then
you have Z Z Z
F · dr = (∇ × F) · ndA
C S
(See theorem 1 in 10.9 for a proper detailed statement of the theorem)
Do problems 9,10 to understand the theorem and its use. When a curve
is a closed curve in a plane then the surface it encloses will also be a
plane region and then the Stoke’s theorem that we get is nothing but
the Green’s theorem. (Try problems on Green’s theorem now)
Yet again we have a similar comment as before - it may not be possible
to evaluate any surface integral in a closed form. If we have a closed
surface over which we want to evaluate an integral then it could be done
by converting it to a triple integral by using the Divergence theorem.
(Theorem 1 in 10.7)
All this needs to be understood properly and to be taken with a pinch
of salt. By this we mean that there cannot be a blanket rule for solving
5
any vector integral. Sometimes one may be easier than the other while
at some other time it may be vice versa.

You might also like