Beruflich Dokumente
Kultur Dokumente
Double integrals over a plane region may be transformed into line integrals over the boundary of
the region and conversely. This transformation also known as Green’s (1793-1841) theorem is
of practical as well as theoretical interest. This concept is extended in a tool called planimeter
1
The region W can be represented in both of the form
a ≤ x ≤ b, u ( x ) ≤ y ≤ v ( x ) (1)
and
c ≤ y ≤ d, p ( y) ≤ x ≤ q ( y) (2)
Consider first
∂P ∂P b y = v ( x)
{ }
b v( x) b
∫∫ ∂y dxdy = ∫ ∫ ( ) dy dx = ∫ P ( x, y ) dx = ∫ P x, v ( x ) − P x, u ( x ) dx (3)
Ω
a u x ∂y a
y = u ( x ) a
Now
i∫ Γ
Pdx = ∫ Pdx + ∫ Pdx
C1 C2
(4)
b
∫C1 P ( x, y ) dx = ∫a P x, u ( x ) dx (5)
Similarly,
a b
∫ P ( x, y ) dx = ∫ P x, v ( x ) dx = − ∫ P x, v ( x ) dx (6)
C2 b a
Thus
{ }
b b b
i∫ Pdx = ∫ P x, u ( x ) dx − ∫ P x, v ( x ) dx = − ∫ P x, v ( x ) − P x, u ( x ) dx (7)
Γ a a a
∂P
∫∫ − ∂y dxdy = i∫
Ω
Γ
Pdx (8)
2
∂Q d q ( y ) ∂Q
{ }
d
∫∫Ω ∂x dxdy = ∫d ∫p( y ) ∂x ∫c Q q ( y ) , y − Q p ( y ) , y dy
dx dy = (9)
∂Q
∫∫ ∂x dxdy = i∫
Ω
Γ
Qdy (10)
Adding (8) and (10) completes the Green’s theorem which reads
∂Q ∂P
i∫ ( Pdx + Qdy ) = ∫∫ ∂x − ∂y dxdy
Γ
Ω
(11)
Consider the area of a plane region as a line integral over the boundary. In (11), let P=0 and
Q=x. Then
∫∫ dxdy = i∫
Ω
Γ
xdy
The left side of the above integral is the area A of W. Similarly, let P=-y and Q=0; then from
(11)
1
( xdy − ydx )
2 ∫Γ
A=
This interesting formula expresses the area of W in terms of a line integral over the boundary.
The triple integral is a generalization of the double integral introduced above. For defining this
integral, one considers a function f ( x, y, z ) defined in a bounded closed region T of space. One
subdivides T by planes parallel to the three coordinate planes. Then, one numbers the
3
parallelepipeds inside T from 1 to n. In each such parallelepiped, one chooses an arbitrary point,
n
J n = ∑ f ( xk , yk , zk ) ∆vk (1)
k =1
where ∆vk is the volume of the kth parallelepiped. One assumes that f ( x, y, z ) is continuous in
T, and T is bounded by finitely many smooth surfaces. It can be shown that Jn in (1) converges to
a limit which is independent of the choice of subdivisions and corresponding points (x,y,z) as n
approaches infinity. This limit is called the triple integral of f ( x, y, z ) over the region T, and is
denoted by
∫∫∫ f ( x, y, z ) dxdydz
T
or ∫∫∫ f ( x, y, z ) dv
T
(2)
It is intended to show here that the triple integral of the divergence of a continuously
differentiable vector function u over a region T in space can be transformed into a surface
integral of the normal component of u over the boundary surface S of T. This can be done by a
means of divergence theorem which is the three-dimensional analogue of Green’s theorem in the
Divergence theorem of Gauss - Let T be a closed bounded region in space whose boundary is a
and has continuous first partial derivatives in some domain containing T. Then
∫∫∫ div uˆ dV = ∫∫ u
T S
n dA (3)
4
is the component of û in the direction of then outer normal of S with respect to T, and n̂ is the
u = u1iˆˆ+ u2 j + u3 kˆ
(5)
n = cos α iˆˆ+ cos β j + cos γ kˆ
so that α , β , and γ are the angles between n̂ and the positive x, y, and z axes, respectively, the (3)
At this point, it may be useful to review a few elementary vector integral calculus formulas.
ds = 1 + f ′ ( x ) dx and s = ∫ 1 + f ′ ( x ) dx
2 2
(7)
l
Extending (7) (refer to the derivation, Grossman/Derrick, pg 620), the lateral surface area, s of
σ = ∫∫ dσ =∫∫ 1 + f x2 ( x, y ) + f y2 ( x, y ) dA (8)
S Ω
where ds an infinitesimal area in the surface domain S and dA is is an infinitesimal area in the
ˆ G = G iˆˆˆˆ
∇ ˆˆ
x + G y j + Gz k = − f x i − f y j + k (9)
and
5
ˆ G − f x iˆˆ− f y j + kˆ
∇
nˆ = = (10)
ˆG
∇ f x2 + f y2 + 1
Then n̂ is the outward normal vector and g is the acute angle between n̂ and the positive z-axis.
Hence, by definition
nˆ ⋅ kˆ 1
cos γ = = and sec γ = f x2 + f y2 + 1 (11)
nˆ kˆ f + f +1
x
2
y
2
and
∫∫ F ( x, y, z ) dσ = ∫∫ F x, y, f ( x, y ) sec γ ( x, y ) dxdy
S Ω
(12)
Rewriting (6)
∫∫ Fˆˆˆ⋅ nd
S
ˆ σ = ∫∫∫ div Fdxdydz = ∫∫∫ div Fdv
T T
(6)
Just as Green’s theorem equates a line integral to an ordinary double integral, the divergence
If
then
6
∂P ∂Q ∂R
∫∫ ( P cos α + Q cos β + R cos γ ) dσ = ∫∫∫ ∂x + ∂y + ∂z dxdydz
S T
(6)
where α , β , and γ are the angles that n̂ makes with the positive coordinate axes. From (6), it is
∂P
∫∫ P cos α dσ = ∫∫∫ ∂x dxdydz
S T
(14)
∂Q
∫∫ Q cos β dσ = ∫∫∫ ∂y dxdydz
S T
(15)
and
∂R
∫∫ R cos γ dσ = ∫∫∫ ∂z dxdydz
S T
(16)
The proofs of (14),(15), and (16) are virtually identical, hence only proof of (16) is illustrated.
That is, S consists of two surfaces: an upper surface S2: z = f 2 ( x, y ) and a lower surface S1:
z = f1 ( x, y ) . Then
From (12)
7
∫∫ R ( x, y, z ) cos γ dσ = ∫∫ R x, y, f ( x, y ) cos γ sec γ dxdy = ∫∫ R x, y, f ( x, y ) dxdy
S2 Ω
2
Ω
2 (18)
where Wis the projection of S2 into xy-plane. On S1 the situation is a bit different because the
outer normal n̂ on S1 points downward and it was assumed that the angle g is an acute angle
between the positive z-axis and n̂ . This situation can be readily rectified if one uses - n̂ in S1.
thus
(20)
f2 ( x , y ) ∂R ∂R
= ∫∫ ∫ dz dxdy = ∫∫∫ dzdxdy
f1 ( x , y ) ∂z ∂z
Ω T
Similarly (14) and (15) can be proved following exactly the same procedure outlined above.
Since (6) contains no reference to any special coordinate system, the result is true in all
mechanics).
Example 1
x 2 + y 2 = 4, 0 ≤ z ≤ 2 .
Solution
div Fˆ = 2 x + 2 + 8 z , so
8
2
∫∫ Fˆ ⋅ nˆ = ∫∫∫ ( 2 x + 2 + 8z ) dzdxdy = ∫∫ ∫ ( 2 x + 2 + 8 z ) dzdxdy
0
S T x2 + y 2 ≤ 4
z=2
= ∫∫ ( 2 xz + 2 z + 4 z ) z=0 dxdy = ∫∫ ( 4 x + 20 ) dxdy
2
x2 + y2 ≤ 4 x2 + y 2 ≤ 4
The differential area in the polar coordinate system becomes dA = dxdy = rdrdθ .)
2π 8
= 4∫ cos θ + 10 dθ = 80π
0
3
Example 2
where S is the closed surface consisting of the cylinder x + y = a ( 0 ≤ z ≤ b ) and the dicks
2 2 2
P = x 3 , Q = x2 y, R = x2 z
Hence, by taking advantage of the symmetry of the circle (evaluate only 1/4 of the circle), the
1 2
I = ∫∫∫ ( 3 x 2 + x 2 + x 2 ) dxdydz = 4 × 5∫ 3 ( a − y )
b a a2 −y2 b a 3/2
∫∫ x2 dxdydz = 20 ∫ ∫
2
0 0 0 0 0 dydz
T
(a − y2 ) = a2 ( 1 − sin2 θ ) = ( a2 cos2 θ )
3/2 3/2 3/2
2
= a3 cos3 θ
and
9
1 a 2 3/2
1 4 π /2 4 1 4 3 π π a4
3 ∫0
( a − y 2
) dy =
3 ∫0
a cos θ d θ = a =
3 8 2 16
b π a4 5
I = ∫ 20 × dz = π a 4b
0 16 4
Example 3
x y z
Fˆ = nˆ = cos ( x, n ) iˆˆˆˆ
+ cos ( y , n ) j + cos ( z , n ) kˆˆ= i + j + k
r r r
where
Solution
Assuming that r ≠ 0
∂P ∂ x x 2 + y 2 + z 2 − x 2 / x2 + y2 + z 2 r 2 − x 2
= = =
∂x ∂x x 2 + y 2 + z 2
x2 + y2 + z2 r3
Likewise,
∂Q r 2 − y 2 ∂R r 2 − z 2
= and =
∂y r3 ∂z r3
Hence,
3r 2 − r 2 2 2
div F =
r3
=
r
if r ≠ 0 and ∫∫∫ r dxdydz
T
Since the triple integral is discontinuous when r=0, care must be exercised. Returning back to
10
ˆ ⋅ nd x2 y2 z2
∫∫S σ = ∫∫S r 2 + r 2 + r 2 dσ = ∫∫S 1dσ = 4π a
2
F ˆ
The integral is simply the surface area of a sphere of radius a, thus avoiding an integral
The triple integral may be evaluated by transforming the Cartesian coordinates into the spherical
2 2π π a 1 2π π
∫∫∫ = ∫0 ∫0 ∫0 r φ φ θ = ∫0 ∫ sin φ dφ dθ
2 2
dxdydz 2 r sin drd d a
T
r 0
π 2π
= a 2 ( − cos φ ) dθ = 2π a2 ( 1 + 1) = 4π a2
0 ∫0
11