Professional Documents
Culture Documents
Since in evaluating line integral we need to express the curve in parametric equation as
function of t, let recognize parametric representation first.
Bodies that move in space form paths that may be represented by curves C. This and other
applications show the need for parametric representations of C with parameter t, which may
denote time or something else (see Fig. 12.1). A typical parametric representation is given by
Here t is the parameter and x, y, z are Cartesian coordinates, that is, the usual rectangular
coordinates. To each value 𝑡 = 𝑡𝑜 , there corresponds a point of C with position vector 𝑟(𝑡𝑜 )
whose coordinates are 𝑥(𝑡𝑜 ), 𝑦(𝑡𝑜 ), 𝑧(𝑡𝑜 ).
When we give parametric equations and a parameter interval for a curve, we say that we have
parametrized the curve. The equations and interval together constitute a parametrization of
the curve. A given curve can be represented by different sets of parametric equations.
The advantages of using parametric representation are that, the coordinates x, y, z all play an
equal role, that is, all three coordinates are dependent variables. Moreover, the parametric
representation induces an orientation on C. This means that as we increase t, we travel along
1
the curve C in a certain direction. The sense of increasing t is called the positive sense on C.
The sense of decreasing t is then called the negative sense on C.
2
12.2 BASIC CONCEPTS
In a line integral, we shall integrate a given function, also called the integrand, along a curve C in
space or in the plane as shown in Figure 3.2. (Hence curve integral would be a better name but line
integral is standard.)
(1)
3
The curve C is called the path of integration. Look at Fig. 12.2a. The path of integration goes from A
to B. Thus A: r(a) is its initial point and B: r(b) is its terminal point. C is now oriented. The direction
from A to B, in which t increases is called the positive direction on C. We mark it by an arrow. The
points A and B may coincide, as it happens in Fig. 12.2b. Then C is called a closed path.
A line integral of a vector function F(r) over a curve C: r(t) is defined by
(2)
𝑑𝑟
Where 𝑟 ′ = 𝑑𝑡
and r(t) is the parametric representation of C as given in (2). Writing (2) in terms of
∫ 𝑭(𝒓) ⋅ 𝑑𝒓 = ∫ 𝑀 𝑑𝑥 + 𝑁 𝑑𝑦 + 𝑃 𝑑𝑧
𝐶 𝐶
𝑏
= ∫ 𝑀 𝑥′ + 𝑁 𝑦 ′ + 𝑃 𝑧 ′ 𝑑𝑡
𝑎
Note that the integrand in (2) is a scalar, not a vector, because we take the dot product. Indeed,
𝐹● 𝑟 ′ ⁄|𝑟 ′ | is the tangential component of F. Line integrals arises naturally in mechanics, where they
give the work done by a force F in a displacement along C. This will be explained in detail below. We
may thus call the line integral (2) the work integral.
4
EXAMPLE 12.2: Evaluation of a Line Integral in the Plane
Find the value of line integral when 𝐹(𝑟) = [−𝑦, −𝑥𝑦] = −𝑦𝒊 − 𝑥𝑦𝒋 and C is the circular arc
from A to B.
The evaluation of line integral in space is practically the same as it is in the plane. To see this,
find the value of line integral when 𝐹(𝑟) = [𝑧, 𝑥, 𝑦] = 𝑧𝒊 + 𝑥𝒋 + 𝑦𝒌 and C is the helix. Given
that
𝑟(𝑡) = [cos 𝑡, sin 𝑡, 3𝑡] = cos 𝑡𝒊 + sin 𝑡𝒋 + 3𝑡𝒌
5
EXAMPLE 12.4: Evaluation of a Line Integral along the Curve, C
Evaluate ∫c F ∙ dr, where F (x, y, z) = zi + xy j – y2k along the curve C given by 𝑟(𝑡) = 𝑡 2 𝒊 + 𝑡𝒋 +
√𝑡𝒌, 0 ≤ 𝑡 ≤ 1.
𝟏𝟕
𝑺𝑶𝑳 =
𝟐𝟎
Simple general properties of the line integral (2)
6
Figure 12.3 Formula (c)
where in (c) the path C is subdivided into two arcs and that have the same orientation as C (Fig. 12.3).
In (b) the orientation of C is the same in all three integrals. If the sense of integration along C is
reversed, the value of the integral is multiplied by -1.
Suppose that the vector field F = M(x, y, z)i + N(x, y, z)y + P(x, y, z)k represents a force throughout a
region in space (it might be the force of gravity or an electromagnetic force of some kind) and that
is a smooth curve in the region. For a curve C in space, we define the work done by a continuous
force field F to move an object along C from a point A to another point B as follows.
In other words, the work done by a force F is the line integral of the scalar component F.T over the
smooth curve from A to B as shown in Fig 12.4. The sign of the number we calculate with this integral
depends on the direction in which the curve is traversed. If we reverse the direction of motion, then
we reverse the direction of T in Figure 12.4 and change the sign of and its integral.
7
Figure 12.4: The work done by a force F is the line
EXAMPLE 12.5
8
EXAMPLE 12.6
Find the work done by the force fields F = xi + yj +zk in moving an object along the curve C
parameterized by 𝑟(𝑡) = cos(𝜋t) i + 𝑡 2 𝑗 + sin(𝜋𝑡) 𝑘, 0 ≤ 𝑡 ≤ 1
9
Path Dependence
Take, for instance, the straight segment 𝐶1 ∶ 𝑟1 (𝑡) = [𝑡, 𝑡, 0] and the parabola 𝐶2 ∶ 𝑟2 (𝑡) =
[𝑡, 𝑡 2 , 0] as shown in Figure 12.5 with 0 ≤ t ≤ 1 and integrate F = [0, 𝑥𝑦, 0]. Then
For curve 𝐶1 ∶ 𝑟1 ′ (𝑡) = [1, 1, 0], 𝐹 = [0, (𝑡)(𝑡), 0], therefore 𝐹(𝑟1 (𝑡))●𝑟1′ (𝑡) = 𝑡 2
For curve 𝐶2 ∶ 𝑟2 ′ (𝑡) = [1, 2𝑡, 0], 𝐹 = [0, (𝑡)(𝑡 2 ), 0], therefore 𝐹(𝑟2 (𝑡))●𝑟2′ (𝑡) = 2𝑡 4
It is obvious that the two integrands of the line integral are different even though the two
curves share the same endpoints A and B. Logically, this gives different values of 1/3 and 2/5
respectively.
As an additional information, if a vector function F is the gradient of a scalar function 𝐹 = ∇𝑓, this
vector function F is known as a conservative vector field, and its line integral is interestingly
independent on path, which means integrating F along C1 or C2 produces the same value for the same
10
EXAMPLE 12.7:
Evaluate the line integral with the vector function given as 𝐹(𝑥, 𝑦) = (𝑥 − 𝑦)𝒊 + 𝑥𝒋. The
curve C is a closed curve that forms a unit circle. A closed circular curve can be parameterized
as below, considering a range of t that forms the complete circular path:
Solution
𝑑𝑟
Therefore: 𝐹. 𝑑𝑡 = − sin 𝑡 cos 𝑡 + sin2 𝑡 + cos2 𝑡 = − sin 𝑡 cos 𝑡 + 1
2𝜋
2𝜋 (cos 𝑡)2
∫0 (− sin 𝑡)(cos 𝑡) + 1 𝑑𝑡 = [ 2
+ 𝑡]
0
1 1
= ( + 2𝜋) − ( + 0)
2 2
= 2𝜋
Figure 12.6: The vector field has a counter clockwise circulation of 2π around the unit circle.
EXAMPLE 12.8
Evaluate the closed-curve line integral
∮ −𝑦 2 𝑑𝑥 + 𝑥𝑦𝑑𝑦,
𝐶
Where C is the square cut from the first quadrant by the lines x = 1 and y = 1 (counter-clockwise)
11
Solution
𝑑𝑥 𝑑𝑦 𝑑𝑥 𝑑𝑦
∮𝐶 −𝑦 2 𝑑𝑥 + 𝑥𝑦𝑑𝑦 = ∮𝐶 (−𝑦 2 ( 𝑑𝑡 ) + 𝑥𝑦 ( 𝑑𝑡 )) 𝑑𝑡 + ⋯ + ∮𝐶 (−𝑦 2 ( 𝑑𝑡 ) + 𝑥𝑦 ( 𝑑𝑡 )) 𝑑𝑡 :
1 4
EXAMPLE 12.9
Compute ∮ 𝑥𝑦 𝑑𝑥 + 𝑥𝑦 𝑑𝑦, over the counter-clockwise rectangle with corners (1,1), (3,1), (3,2), (1,2).
Solution
𝑑𝑥 𝑑𝑦 𝑑𝑥 𝑑𝑦
∮𝐶 𝑥𝑦𝑑𝑥 + 𝑥𝑦𝑑𝑦 = ∮𝐶 (𝑥𝑦 ( 𝑑𝑡 ) + 𝑥𝑦 ( 𝑑𝑡 )) 𝑑𝑡 + ⋯ + ∮𝐶 (𝑥𝑦 ( 𝑑𝑡 ) + 𝑥𝑦 ( 𝑑𝑡 )) 𝑑𝑡
1 4
12
C3: y=2, 1≤x≤3 (from 3 to 1) Use x=t, y=2 1
∫3 ((𝑡)(2)(1) + (𝑡)(2)(0)) 𝑑𝑡
𝑑𝑥 𝑑𝑦 3
= 1, 𝑑𝑡 = 0 16
𝑑𝑡 = − ∫1 2𝑡𝑑𝑡 = − 2
The solutions of line integral involving closed curve can be tedious when the closed curve is defined
by multiple sections as in Examples 12.8 and 12.9. In some cases, it will be more convenient to relate
a closed-curve line integral with an integral involving the curl of the vector function F over a surface
that is bounded by the closed curve. This is known as the (circulation form of) Green’s theorem, which
is explained in the topic of Stokes’ theorem.
13
SURFACE INTEGRALS
WEEK 13: SURFACE INTEGRALS
13.1 SURFACES FOR SURFACE INTEGRAL
The idea that will lead to the concept of a surface integral is quite similar to that which led to a line
integral. We say briefly ‘surface’ also for a portion of a surface, just as we said ‘curve’ for an arc of a
curve, for simplicity. Representations of a surface S in xyz-space are
z = f (x , y ) or g(x , y , z ) = 0 (1)
a and centre 0.
Now for curves C in line integrals, it was more practical and gave greater flexibility to use a parametric
representation r = r(t) where a ≤ t ≤ b. This is a mapping of the interval a ≤ t ≤ b located on the t-axis,
onto the curve C in the xyz-space. It maps every t in that interval onto the point of C with position
vector r(t) (see Figure 13.1)
Similarly, for surfaces S in surface integrals, it will often be more practical to use a parametric
representation. Surfaces are two-dimensional. Hence we need two parameters, which we call u and
v. Thus a parametric representation of a surface S in space is of the form
when (u, v) varies in some region R of the uv-plane. This mapping (2) maps every point (u, v) in R onto
the point of S with position vector r(u, v) (see Figure 13.1(B)).
14
Parametric Representation of a Cylinder
The circular cylinder x + y = a , -1 ≤ z ≤ 1, has radius a, height 2, and the z-axis as axis. A parametric
2 2 2
representation is
The components of r are x = a cos u, y = a sin u, z = v. The parameters u, v vary in the rectangle R: 0 ≤
u ≤ 2π, -1 ≤ v ≤ 1 in the uv-plane. The curves u = const are vertical straight lines. The curves v = const
are parallel circles. The point P in Figure 13.2 corresponds to u = π / 3 = 60°, v = 0.7.
where the parameters u, v vary in the rectangle R in the uv-plane given by the inequalities 0 ≤ u ≤ 2π,
-π/2 ≤ v ≤ π/2. The components of r are
The curves u = const and v = const are the ‘meridian’ and ‘parallels’ on S (as shown in Figure 13.3). This
representation is used in geography for measuring the latitude and longitude of points on the globe.
Another parametric representation of the sphere also used in mathematics is
where 0 ≤ u ≤ 2π, 0 ≤ v ≤ π
15
Figure 13.3 Parametric representation of a sphere
The extension of the idea of an integral to line and double integrals are not the only generalization
that can be made. We can also extend the idea to integration over a general surface, S. Two types of
such integrals occur:
16
13.2.1 SURFACE INTEGRAL OF VECTOR FIELD
In general, the surface S can be described in terms of two parameters, u and v say, so that on S
where (u, v) varies over a region R in the uv-plane. We assume S to be piecewise smooth, so that S has
a normal vector
1
N = ru rv and unit normal vector n= N
N
at every point (except perhaps for some edges or cusps, as for a cube or cone). For a given vector
function F we can now define the surface integral over S by
Here N = N n by equation (4), and N = ru rv is the area of the parallelogram with sides ru and rv, by
the definition of cross product. Hence
n dS = nN du dv = N du dv (8*)
also F . n is the normal component of F. This integral arises naturally in flow problems, where it gives
the flux across S when F = ρv. The flux across S is the mass of fluid crossing S per unit time. Furthermore,
ρ is the density of the fluid and v the velocity vector of the flow, as illustrated by Example 13.1 below.
We may thus call the surface integral (8) the flux integral.
We can write (8) in components, using F = [F1,F2, F3], N = [N1,N2,N3], and n = [cos α, cos β, cos ϒ]. Here
α, β, ϒ are the angles between n and the coordinate axes; indeed, for the angle between n and i.
F . n dS = (F cos + F
S S
1 2 cos + F3 cos ) dS (9)
In (9) we can write cos α dS = dydz, cos β dS = dzdx, cosϒdS = dxdy. Then (9) becomes the following the
integral for the flux:
We can use this formula to evaluate surface integrals by converting them to double integrals over
regions in the coordinate planes of the xyz-coordinate system. But we must carefully take into account
the orientation of S (the choice of n), as described in Section 13.3.1.
17
The tangent vectors of all the curves on a surface S through a point P of S form a plane, called the
tangent plane of S at P (refer to Figure 13.4). Exceptions are points where S has an edge or a cusp (like
a cone), so that S cannot have a tangent plane at such a point. Furthermore, a vector perpendicular
to the tangent plane is called a normal vector of S at P.
Now since S can be given by r = r(u, v) in (2), the new idea is that we get a curve C on S by taking a pair
of differentiable functions
u = u(t), v = v(t)
whose derivatives u’ = du/dt and v’ = dv/dt are continuous. Then C has the position vector
~r (t ) = r(u(t ),v (t )) . By differentiation and the use of the chain rule, we obtain a tangent vector of C on
S
Hence the partial derivatives ru and rv at P are tangential to S at P. We assume that they are linearly
independent, which geometrically means that the curves u = const and v= const on S intersect at P at
a non-zero angle. Then ru and rv span the tangent plane of S at P. Hence their cross product gives a
normal vector N of S at P.
N = ru rv 0 (11)
1 1
n= N= ru rv (12)
N ru rv
1
n= grad g (12*)
grad g
18
Tangent Plane and Surface Normal
r r
If a surface S is given by (2) with continuous ru = and rv = at every point of S, then S has, at
u v
every point P, a unique tangent plane passing through P and spanned by ru and rv and a unique normal
whose direction depends continuously on the points of S. A normal vector is given by (11) and the
corresponding unit normal vector by (12) (see Figure 4.4).
Example 13.1:
velocity vector is v = F = 3z , 6 , 6 xz , speed being measured in meters/sec. (Generally, F = ρv, but
2
Solution:
S: r = u,u2 , v (0 u 2, 0 v 3)
On S, writing simply F(S) for F[r(u,v)], we have F(S) = [ 3v2, 6, 6uv]. Hence F(S).N = 6uv2 – 6. By
integration we thus get from (7) the flux
( )
3 2 3
( )
3
= 12v 2 − 12 dv = 4v 3 − 12v v =0 = 108 − 36 = 72 m 3 / sec
3
19
Or 72,000 liters/sec. Note that the y-component of F is positive (equal to 6), so that in figure above,
the flow goes from left to right.
Example 13.2:
Compute the flux through the surface S, of the cylinder 𝑥 2 + 𝑦 2 = 16 in the first octant between z =
0 and z = 5 if the velocity vector ∬𝑆 𝑭 . 𝑑𝑺 where 𝑭 = [𝑧, 𝑥, −3𝑦 2 𝑧].
Solution:
𝒊 𝒋 𝒌
𝜕𝑟 1 | = [ 𝑢 , 1,0]
𝒓𝒗 = = [0, 0,1] ∴ 𝑵 = 𝒓𝒗 × 𝒓𝒖 = |0 0
𝑢
𝜕𝑣 √16−𝑢2
1 − 0
√16−𝑢2
5 4
𝑢𝑣
∬ 𝑭. 𝑵𝑑𝑆 = ∫ ∫ + 𝑢 𝑑𝑢𝑑𝑣
0 0 √16 − 𝑢2
𝑆
5 4 4 5 4 4
𝑢 𝑣 −2𝑢
∴ ∫ (𝑣 [∫ 𝑑𝑢] + ∫ 𝑢 𝑑𝑢) = ∫ [− ∫ 𝑑𝑢 + ∫ 𝑢 𝑑𝑢] 𝑑𝑣
√16 − 𝑢2 2 √16 − 𝑢2
0 0 0 0 0 0
5 5 4
4 𝑢2
= ∫ −𝑣 [√16 − 𝑢2 ] 𝑑𝑣 + ∫ [ ] 𝑑𝑣
0 2 0
0 0
5 5
= ∫0 4𝑣 𝑑𝑣 + ∫0 8 𝑑𝑣 = 90
20
Comment: In this example, the solution shown uses a basic parameterization of x = u, 𝑦 = √16 − 𝑢2 ,
z = v, instead of the common cylindrical surface parameterization x = cos u, y = sin u, z = v as learned
earlier. Try to repeat this example using the cylindrical surface parameterization (Hint: integrals of the
same field and the same surface should result in the same answer).
When the field (function) to be integrated is a scalar field, the resulted integral is known as the surface
integral of a scalar field, or simply the scalar surface integral. So, ∬𝑆 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑆 has essentially the
same concept as vector surface integral, but with a scalar field instead. Just as the vector surface
integral, here, for a given surface S, suitable parameterization can be used: x = x(u, v), y = y(u, v), z =
z(u, v), which gives again the position vector r(u, v) = x(u, v)i + y(u, v)j + z(u, v)k.
Just as the vector surface integral, the surface S can be specified by a scalar point function C(r)= c,
where c is a constant. Curves may be drawn on that surface, and in particular if we fix the value of one
of the two parameters u and v then we obtain two families of curves. On one, 𝐶𝑢 (𝒓(𝑢, 𝑣0 )) the value
of u varies while v is fixed, and on the other, 𝐶𝑣 (𝒓(𝑢0 , 𝑣)), the values of v varies while u is fixed, as
shown in Figure 13.5. Then as indicated on Figure 13.5, the vector element of area dS is given by:
𝜕𝒓 𝜕𝒓 𝜕𝒓 𝜕𝒓
𝑑𝑺 = 𝑑𝑢 × 𝑑𝑣 = × 𝑑𝑢𝑑𝑣
𝜕𝑢 𝜕𝑣 𝜕𝑢 𝜕𝑣
𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑦 𝜕𝑧
=( , , ) × ( , , ) 𝑑𝑢 𝑑𝑣 = (𝐽1 𝒊 + 𝐽2 𝒋 + 𝐽3 𝒌)𝑑𝑢 𝑑𝑣
𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑣 𝜕𝑣 𝜕𝑣
where
𝜕𝑦 𝜕𝑧 𝜕𝑦 𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦
𝐽1 = 𝜕𝑢 𝜕𝑣 − 𝜕𝑣 𝜕𝑢, 𝐽2 = 𝜕𝑢 𝜕𝑣 − 𝜕𝑣 𝜕𝑢, 𝐽3 = 𝜕𝑢 𝜕𝑣 − 𝜕𝑣 𝜕𝑢
21
𝜕𝒓 𝜕𝒓
However, instead of 𝑑𝑺 = 𝜕𝑢 × 𝜕𝑣 𝑑𝑢𝑑𝑣, for scalar surface integral, we have:
𝜕𝒓 𝜕𝒓 𝜕𝒓 𝜕𝒓
𝑑𝑆 = |𝜕𝑢 × 𝜕𝑣| 𝑑𝑢𝑑𝑣, where the magnitude |𝜕𝑢 × 𝜕𝑣| = √(𝐽1 2 + 𝐽2 2 + 𝐽3 2 )
Once the entire integral can be described with only u and v, the integration can be carried out to
obtain the value of the integral.
Sometimes, a given surface can be very clearly described by z = z(x, y). A quick example is a vertical
cone with 𝑧 = √𝑥 2 + 𝑦 2 . In this situation, we can simply adopt x = u, y = v (and 𝑧 = √𝑢2 + 𝑣 2 ). In fact,
we can directly write the two parameters as x and y without introducing new symbols u and v. for
example, if z = z(x, y) describes a surface as in Figure 13.6, then
𝜕𝑧 𝜕𝑧
𝐽1 = − 𝜕𝑥 𝐽2 = − 𝜕𝑦 𝐽3 = 1
𝜕𝒓 𝜕𝒓 𝜕𝒓 𝜕𝒓
Since 𝑑𝑆 = | × | 𝑑𝑢𝑑𝑣 =| × | 𝑑𝑥𝑑𝑦,
𝜕𝑢 𝜕𝑣 𝜕𝑥 𝜕𝑦
𝜕𝑧 2 𝜕𝑧 2
∬𝑆 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑆 = ∬𝑆 𝑓(𝑥, 𝑦, 𝑧(𝑥, 𝑦))√(1 + (𝜕𝑥) + (𝜕𝑦) ) 𝑑𝑥𝑑𝑦 (14)
22
Example 13.3:
∬(𝑥 + 𝑦 + 𝑧) 𝑑𝑆
𝑆
where S is the portion of the sphere 𝑥 2 + 𝑦 2 +𝑧 2 = 1 that lies in the first quadrant.
Solution:
𝑧 = √1 − 𝑥 2 − 𝑦 2
we have
𝜕𝑧 −𝑥
=
𝜕𝑥 √1 − 𝑥 2 − 𝑦 2
𝜕𝑧 −𝑦
=
𝜕𝑦 √1 − 𝑥 2 − 𝑦 2
1
∴ ∬(𝑥 + 𝑦 + 𝑧)𝑑𝑆 = ∬ [𝑥 + 𝑦 + √1 − 𝑥 2 − 𝑦 2 ] 𝑑𝑥𝑑𝑦
𝑆 𝐴
√1 − 𝑥 2 − 𝑦 2
where A is the quadrant of a circle in the (x, y) plane illustrated in Figure (b) (refer to the above Figure).
Thus,
1 √1−𝑥 2
𝑥 𝑦
∬(𝑥 + 𝑦 + 𝑧) 𝑑𝑆 = ∫ ∫ [ + + 1] 𝑑𝑦𝑑𝑥
𝑆 0 0
√1 − 𝑥 2 − 𝑦 2 √1 − 𝑥 2 − 𝑦 2
23
1
𝑦 √1−𝑥 2
−1
= ∫ [𝑥 𝑠𝑖𝑛 ( ) − √1 − 𝑥2 − 𝑦2 + 𝑦] 𝑑𝑥
√1 − 𝑥 2 0
0
1
𝜋 𝜋 1 3
= ∫ [ 𝑥 + 2√1 − 𝑥 2 ] 𝑑𝑥 = [ 𝑥 2 + 𝑥√1 − 𝑥 2 + 𝑠𝑖𝑛−1 𝑥] = 𝜋
2 4 0 4
0
An alternative approach to evaluating the surface integral in this example is to evaluate it directly over
the surface of the sphere using spherical polar coordinates. As illustrated in the Figure (c), on the
surface of a sphere of radius a we have,
𝜋/2
1 1 1 3
= ∫ [ 𝜋𝑐𝑜𝑠𝜙 + 𝜋𝑠𝑖𝑛𝜙 + ] 𝑑𝜙 = 𝜋
4 4 2 4
0
Example 13.4:
Solution:
4 − 𝑥 − 2𝑦 𝑥 𝑦
𝑧= =1− −
4 4 2
24
𝜕𝑧 1 𝜕𝑧 1
𝜕𝑥
= −4 𝜕𝑦
= −2
𝜕𝑧 2 𝜕𝑧 2
∬ 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑆 = ∬ 𝑓(𝑥, 𝑦, 𝑧(𝑥, 𝑦))√1 + ( ) + ( ) 𝑑𝑥𝑑𝑦
𝜕𝑥 𝜕𝑦
𝑆 𝐴
𝑥 𝑦 1 2 1 2
∬ 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑆 = ∬ (𝑥 + 𝑦 + 1 − − ) √1 + (− ) + (− ) 𝑑𝑥𝑑𝑦
4 2 4 2
𝑆 𝐴
3𝑥 𝑦 √21
= ∬( + + 1) 𝑑𝑥𝑑𝑦
4 2 4
2 4−2𝑦 2 4−2𝑦
√21 3𝑥 𝑦 √21 3𝑥 2
= ∫ ∫ ( + + 1) 𝑑𝑥𝑑𝑦 = ∫( + 2𝑦𝑥 + 4𝑥) 𝑑𝑦
4 4 2 16 2 0
0 0 0
2
√21 2 √21
= ∫ (3(16 − 16𝑦 + 4𝑦 2 ) + 16𝑦 − 8𝑦 2 + 32 − 16𝑦) 𝑑𝑦 = ∫(80 − 48𝑦 + 4𝑦 2 ) 𝑑𝑦
32 0 32
0
2 2
√21 √21 𝑦3 √21 8 7√21
= ∫ 20 − 12𝑦 + 𝑦 2 𝑑𝑦 = [20𝑦 − 6𝑦 2 + ] = (40 − 24 + ) =
8 8 3 0 8 3 3
0
Scalar surface integral can also be used to find the area of a given surface S. This is when f(x, y, z) = 1
and the integral ∬𝑆 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑆 = ∬𝑆 𝑑𝑆 simply ‘sums-up’ all the elements’ small area to give the
area of the entire surface S. The following examples demonstrate this application.
Example 13.5:
25
Solution:
z = f(x, y)
2
f f
2
A(S ) = 1 + + dxdy
R* x y
f f f f
2 2
(
f ( x, y ) = x 2 + y 2 )1/ 2
f 1 2
= x + y2
x 2
( )
−1 / 2
2x =
x
x2 + y2
f 1 2
= x + y2
y 2
( )
−1 / 2
2y =
y
x2 + y 2
2
f f
2
x2 + y 2
1 + + = 1 + 2 = 2
x y x + y2
2
f f
2
But R is bounded by x + y = 1 , i.e. a circle, centre the origin and radius 1. area =
2 2
A(S ) = 2 dxdy = 2
R
Example 13.6:
Find the area of the surface S of the paraboloid z = x 2 + y 2 cut off by the cone z = 2 x 2 + y 2
26
Solution:
We can find the point of intersection A by considering the y-z plane i.e. x = 0. Then, coordinates of A
are A(2, 4).
x2 + y2 = 4
2
f f
2
A(S ) = 1 + + dxdy
R* x y
A(S ) = 1 + 4 x 2 + 4 y 2 dxdy
R*
x =2 y = 4 − x
2
but how do we carry out the actual integration? It becomes a lot easier if we use polar coordinates.
The same integral in polar coordinates is
=2 r =2
A(S ) = 1 + 4r 2 rdrd
=0 r =0
=2 r =2 1/2 2 2
(1 + 4r ) rdrd = (1 + 4r 2 ) d
1 3/2
A(S ) = 2
0 0
=0 r =0 12
2
12 0
27
13.3 SOME PROPERTIES OF SURFACE INTEGRAL
In this section, we discuss some properties of surface integral that are relevant to solving the integral.
In particular, we look into these two matters: (i) the orientation of a given surface, and (ii) the
continuity of a surface.
From (8) and (8*), we see that the value of the integral depends on the choice of the unit normal
vector n. We express this by saying that such an integral is an integral over an oriented surface S, that
is, over a surface S on which we have chosen one of the two possible unit normal vectors in a
continuous fashion.
If we change the orientation of S, as illustrated by Figure 13.7, this means that we replace n with –n.
Then each component of n is multiplied by -1. This gives a negative to the original value of the
surface integral.
The replacement of n by –n (hence of N by –N) corresponds to the multiplication of the integral in (8)
by -1
Example 13.7:
28
Note that this integrand is now the negative of that in Example 13.1. Consequently, it is obvious that
the integration also gives the negative of the value obtained in Example 13.1:
3 2 3
∬𝑠 𝑭. 𝑵 𝑑𝑢𝑑𝑣 = ∫0 ∫0 (6 − 6𝑢𝑣 2 )𝑑𝑢𝑑𝑣 = ∫0 (12 − 12𝑣 2 )𝑑𝑣 = −72
However, the orientation of a surface (the required direction of the unit normal vector) does not affect
𝜕𝒓 𝜕𝒓
a scalar surface integral. This is because ∬𝑠 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑆 = ∬𝑠 𝑓(𝑢, 𝑣) |𝜕𝑢 × 𝜕𝑣| 𝑑𝑢𝑑𝑣 , so even
though 𝒓𝑢 × 𝒓𝑣 = −(𝒓𝑣 × 𝒓𝑢 ), the magnitudes are the same. In fact, we do not speak of orientation
of the given surface when solving a scalar surface integral. Therefore, this is also sometimes known as
surface integral without regard to orientation.
This property is similar to that of a line integral. Its concept can be conveniently expressed by:
∬ 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑆 = ∬ 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑆1 + ∬ 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑆2 for ease of understanding. This means, when the
integration is to be carried out over a surface S which is a composite of two (or more) surfaces, e.g. S1
and S2, the surface integral can be obtained by summing all the component surface integrals. This is
true as long as the piecewise surface is continuous.
This property is usually applicable in situations involving closed surfaces, for example, as illustrated
below:
Example 13.8:
Evaluate the surface integral ∬𝑆 (𝑧 2 ) 𝑑𝑆, where S is the total area of the cone √𝑥 2 + 𝑦 2 ≤ 𝑧 ≤ 2
Solution:
𝑆1 = ∬ 𝑧 2 𝑑𝑆1 = √2 ∬ 𝑥 2 + 𝑦 2 𝑑𝑥𝑑𝑦
𝑆1
29
Change to polar coordinate
2𝜋 2
= √2 ∫ ∫ 𝑟 2 𝑟𝑑𝑟𝑑𝜃 = 8√2 𝜋
0 0
As a final remark, note that this property is applicable to both scalar surface integrals and vector
surface integrals.
30
STOKES’ THEOREM
WEEK 14: STOKES’ THEOREM
14.1 INTRODUCTION
From previous topic, we have learnt that double integrals over a region in the plane can be
transformed into line integrals over the boundary curve of the region and conversely. We shall now
see that more generally, surface integrals over a surface S with boundary curve C can be transformed
into line integrals over C and conversely.
Stokes’ theorem is the ‘curl analogue’ of the divergence theorem and related the integral of curl of a
vector field over an open surface S to the line integral of the vector field around the perimeter C
bounding the surface.
If F is a vector filed existing over surface S and around its boundary, closed curve c, then
This means that we can express a surface integral in terms of a line integral round the boundary curve.
Example 14.1:
Verify Stokes’ theorem, that curl F.dS = F.dr
s
c
𝑆: 𝑥 2 + 𝑦 2 + 𝑧 2 − 4 = 0
𝐹 = 2𝑦𝒊 − 𝑥𝒋 + 𝑥𝑧𝒌
c is the circle x 2 + 𝑦 2 = 4
𝟐𝝅
∮𝑭. 𝒅𝒓 = ∫ (𝟒 𝒔𝒊𝒏𝜽[−𝟐𝒔𝒊𝒏𝜽𝒅𝜽] − 𝟐𝒄𝒐𝒔𝜽𝟐𝒄𝒐𝒔𝜽𝒅𝜽)
𝒄 𝟎
31
= −4 (2sin2θ + cos2θ ) dθ
2π
0
2 2
= -4 (1 + sin2 ) d = - 2 (3 - cos2 ) d (1)
0 0
2
sin2
= -2 3 - = −12
2 0
i j k
curl F = = i(0 − 0 ) − j(z − 0 ) + k(− 1 − 2) = − zj − 3k
x y z
2y −x xz
S 2 xi + 2yj + 2zk xi + yj + zk
n= = =
S 4 x 2 + 4y 2 + 4 z2 2
Now
xi + yj + zk
Then curl F. n dS =
s
(− zj − 3k ).
s
2
dS
1
(− yz − 3z ) dS
2 s
=
1
curl F. n dS = (− 2 sin sin 2 cos − 6 cos )4 sindd
s
2
s
2 / 2
= −4 (2 sin cos sin + 3 sin cos )dd
2
0 0
2 /2
2 sin3 sin 3 sin2
= −4 + d
2 0
(2)
0
3
2
2 3
= −4 sin + dφ = -12
0
3 2
curl F.dS = F . dr
s c
32
Example 14.2:
Solution:
The surface and boundary involved are illustrated in the above figure. We are required to show that
∮ 𝑭. 𝑑𝑟 = ∬ 𝑐𝑢𝑟𝑙 𝑭. 𝑑𝑺
𝑐 𝑆
Since C is a circle of unit radius in the (x,y) plane, to evaluate ∮𝑐 𝑭. 𝑑𝑟, we take
𝑥 = cos 𝜙, 𝑦 = sin 𝜙,
so that
𝑟 = cos 𝜙𝒊 + 𝑠𝑖𝑛𝜙𝒋
Thus,
2𝜋
∮ 𝑭. 𝑑𝑟 = ∫ (2 𝑐𝑜𝑠𝜙 − 𝑠𝑖𝑛𝜙)𝒊 . (−𝑠𝑖𝑛𝜙 𝒊 + 𝑐𝑜𝑠𝜙 𝒋)𝑑𝜙
0
𝑐
2𝜋 2𝜋
2
1
= ∫ (−2𝑠𝑖𝑛𝜙 𝑐𝑜𝑠𝜙 + 𝑠𝑖𝑛 𝜙)𝑑𝜙 = ∫ [−𝑠𝑖𝑛2𝜙 + (1 − 𝑐𝑜𝑠2𝜙)] 𝑑𝜙 = 𝜋
2
0 0
𝒊 𝒋 𝒌
𝜕 𝜕 𝜕
𝑐𝑢𝑟𝑙 𝐹 = || |=𝒌
𝜕𝑥 𝜕𝑦 𝜕𝑧 |
2𝑥 − 𝑦 −𝑦𝑧 2 −𝑦 2 𝑧
33
The unit outward-drawn normal at a point (x,y,z) on the hemisphere is given by (xi+yj+zk) since 𝑥 2 +
𝑦 2 + 𝑧 2 = 1. Thus
When we were dealing with the divergence theorem, the normal vectors were drawn in a direction
outward from the enclosed region. With an open surface as we now have, there is in fact no inward
or outward direction. With any general surface, a normal vector can be drawn in either of two opposite
directions. To avoid confusion, a convention must therefore be agreed upon and the established rule
is as follows.
A unit normal n is drawn perpendicular to the surface S at any point in the direction indicated by
applying the right-handed screw sense to the direction of integration round the boundary c. This is
identical to right-hand grip rule. Having noted that point, we can now deal with the next example.
Example 14.3:
A surface consists of five sections formed by the planes x=0, x=1, y=0, y=3, z=2 in the first octant. If
the vector field F = yi+z2j+xyk exists over the surface and around its boundary, verify Stokes’ theorem.
34
If we progress round the boundary along c1, c2, c3, c4 in an anti-clockwise manner, the normal to the
surfaces will be as shown.
We have to verify that curl F.dS = F .dr
s
c
F .dr = (0 + 0 + 0 ) = 0
c3
F .dr = 0 + 0 − 3 + 0 = −3
c
c
F .dr = -3
(b) Now we have to find curl F.dS
s
F = yi + z 2 j + xyk
i j k
curl F = F =
x y z
y z2 xy
= i(x − 2z) − j(y − 0) + k(0 − 1) = (x − 2z)i − yj − k
Then for each section, we obtain curl F.dS = curl F. n
s
dS
(1) S1 (top) n = k
35
curl F. n dS =
S1
( x − 2 z)i − yj - k. (k) dS
S1
(−1)dS = −(area of S ) = −3
S1
1
Then, likewise
(2) S2 (right-hand end): n = j
curl F. n dS =
S2
( x − 2 z)i − yj - k. ( j) dS
S2
= (−y) dS
S2
curl F. n dS = (− 3) dS = (- 3)(2) = −6
S2 S2
(3) S3 (left-hand end): n = − j
curl F. n dS =
S3
( x − 2 z)i − yj - k. (− j) dS
S3
= (y) dS
S3
curl F. n dS = 0
S2
(4) S4 (front): n = i
curl F. n dS =
S4
( x − 2 z)i − yj - k. (i) dS
S4
= (x − 2z) dS
S4
3 2 3
curl F. n dS = (1 − 2z)dzdy = z − z 2 0dy
2
S4 0 0 0
3
= (− 2) dy = − 2y 30 = −6
0
(5) S5 (back): n = −i with x=0 over S5
36
Similar working to the above gives curl F. n dS = 12
S5
curl F. n dS = −3 − 6 + 0 − 6 + 12 = −3
S4
So, referring back to our result for section (a) we see that
Example 14.4:
A surface S consists of that part of the cylinder x 2 + y 2 = 9 between z=0 and z=4 for y≥0 and the two
semicircles of radius 3 in the planes z=0 and z=4. If F = zi + xyj + xzk , evaluate curl F.dS over the
s
surface.
curl F.dS
s
F = zi + xyj + xzk
Now we can work through this easily enough, taking c1, c2, c3, c4 in turn, and summing the results which
gives
37
F. dr = (0 + 0 + 0) = 0
c1 c1
4
− 3z 2
F . dr = c (0 + 0 − 3zdz ) = 2 = −24
c2 2 0
F . dr = (4dx + 0 + 0) = 4dx = 24
c3 c3 −3
0
3z 2
c F . dr = c (0 + 0 + 3 zdz ) = = −24
4 4
2 4
There is an alternative way of solving this example. We can consider a fictitious surface enclosed by
the rectangular curve C1-C2-C3-C4 (the vertical rectangular surface formed by the closed curve). This
fictitious surface shares the same closed curve as the hollow half-cylinder surface. Therefore, finding
∫𝑠 𝑐𝑢𝑟𝑙 𝑭. 𝑑𝑺 of this fictitious surface bounded by y = 0, -3 ≤ x ≤ 3, 0 ≤ z ≤ 4, which is more straight-
forward than finding the original surface integral, will also give the same answer:
𝒊 𝒋 𝒌
𝜕 𝜕 𝜕
𝑐𝑢𝑟𝑙 𝑭 = | | = (1 − 𝑧)𝒋 + 𝑦𝒌
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝑧 𝑥𝑦 𝑥𝑧
∫𝑠 𝑐𝑢𝑟𝑙 𝑭. 𝑑𝑺 = ∫𝑠 𝑐𝑢𝑟𝑙 𝑭. 𝒏 𝑑𝑆
3 4
= ∫−3 ∫0 (1 − 𝑧)𝑑𝑧𝑑𝑥
4
3 𝑧2
= ∫−3 [𝑧 − 2 0
] 𝑑𝑥
3
= ∫−3 −4 𝑑𝑥 = −24
This alternative solution demonstrates an interesting property related to Stokes’ theorem: if two or
more surfaces share the same closed curve, their surface integrals (of the Stokes’ theorem) give the
same value.
38
14.3 GREEN’S THEOREM (CIRCULATION FORM)
Green’s theorem enables an integral over a plane area to be expressed in terms of a line integral round
its boundary curve. Let P and Q be two functions of x and y that are, along with their first partial
derivatives, finite and continuous inside and on the boundary c of a region R in the x-y plane.
That is, a double integral over the plane region R can be transformed into a line integral over the
boundary c of the region – and the action is reversible.
Q P
(Pdx + Qdy ) = x − y dxdy
c S
where the line integral is taken round c in an anticlockwise manner. In vector terms, this becomes:
RHS: dS = dxdy
dS = n dS = k dx dy
i j k
Q P Q P
curl F = = i 0 − − j 0 − + k −
x y z z z x y
P Q 0
Q P Q P
But in the x-y plane, = = 0 . curl F = k −
z z x y
So curl F.dS = curl F. n dS and in the x-y plane, n = k
Q P Q P
curl F.dS = k − .(k )dS = − dxdy
S s
x y S
x y
39
Q P
curl F.dS = − dxdy (1)
S S
x y
LHS: in this case F .dr =
c
(Pi + Qj)(. idx + jdy + kdz )
c
= (Pdx + Qdy )
c
(2)
Q P
x − y dxdy = Pdx + Qdy
S c
Example 14.5:
(x
+ y 2 )dx + (x + 2y )dy taken round the boundary curve
2
Verify Green’s theorem for the integral
c
c defined by
y =0 0 x 2
x +y =4
2 2
0 x 2
x =0 0 y 2
Q P
Green’s theorem: x − y dx dy = (P dx + Q dy )
S c
( )
In this case x + y dx + ( x + 2y )dy = P dx + Q dy
2 2
P = x 2 + y 2 and Q = x + 2y
40
(1) c1: y=0; dy=0
2
2
x3 8
(P dx + Q dy ) = x dx = =
2
c1 0 3 0 3
x 2 + y 2 = 4 y 2 = 4 − x 2 y = (4 - x 2 )
1/2
(2) c2:
x + 2y = x + 2(4 − x 2 )
1/2
dy = (4 − x 2 ) (− 2 x )dx = − x 2 dx
1 −1 / 2
2 4−x
(
−x
(P dx + Q dy ) = 4 + x + 2 4 − x 2 )
dx
c2 4−x
2
c2
x2
= 4 − 2 x − dx
c2 4 − x2
Putting x = 2 sin θ 4 − x 2 = 2 cos dx = 2 cosd
Limits: x=2 = ; x=0 , θ=0
2
4 sin2
0
(P dx + Q dy ) = 4 − 4 sin − 2 cosd
c2 /2 2 cos
0
1 sin 2
= 4 2 sin − sin2 − −
2 2 / 2
= 4 − 2 − 1 − = − 4
4
Finally
0
(P dx + Q dy ) = 2y dy = y2 2 = −4
0
c3 2
8 16
(P dx + Q dy) = 3 + − 4 − 4 = − 3
c
Q P
That is one part done. Now we have to evaluate x − y dx dy
s
P
P = x2 + y2 = 2y
y
41
Q
Q = x + 2y =1
x
Q P
x − y dx dy = (1 − 2y ) dy dx
s S
/2 2
r 2 2r 3
=
0
2 − 3 sin d
0
/2 /2
16 16 16
=
0
2 − 3 sin d = 2 + 3 cos
0
= −
3
Q P
(P dx + Q dy ) = x − y dx dy
c S
42