중앙대학교 건설환경플랜트공학과 교수
김 진 홍
- 8주차 강의 내용 -
9.6 Calculus Review: Functions of Several Variables.
Chain Rules Theorem 1
Let be continuous and be
functions that are continuous. Then the function is defined and
(1)
) , , (x y z f
w x x(u,v), y y(u,v), z z(u,v) ))
, ( ), , ( ), , (
(x u v y u v z u v f
w
v z z w v y y w v x x w v
w
u z z w u y y w u x x w u w
Special Cases of Practical Interest
- If w = f(x, y) and x = x(u, v), y = y(u, v), then (1) becomes
(2)
v y y w v
x x w v
w
u y y w u
x x w u
w
Chap. 9 Vector Differential Calculus. Grad, Div, Curl
If w = f(x, y, z) and x = x(t), y = y(t), z = z(t) then (1) gives
(3) dt
dz z w dt dy y w dt dx x w dt dw
If w = f(x, y) and x = x(t), y = y(t), then (3) reduces to (4) dwdt wx dxdt wy dydt
Finally, the simplest case w = f(x), x = x(t) gives
(5) dt
dx dx dw dw dt
Ex. 1) w x2 y2 and xrcos
, yrsin
, then
2 sin 2 cos 2 sin 2 cos 2 cos
2x y r 2 r 2 r
r
w
2 ( sin) 2 ( cos ) 2 cos sin 2 sin cos 2 sin2
2 2
2 r r
r r
y r
w x
Ex. 2) z exsin y, x st2, y s2t
) cos(
2 ) sin(
2 ) cos ( ) sin
(e y t2 e y st t2e 2 s2t ste 2 s2t s
y y z s x x z s
z x x st st
) cos(
) sin(
2 ) cos ( 2 ) sin
(e y st e y s2 ste 2 s2t s2e 2 s2t t
y y z t x x z t
z x x st st
Chap. 9 Vector Differential Calculus. Grad, Div, Curl
Mean Value Theorem Theorem 2
Let be continuous, and and be such that the straight line segment
P
0P
. Then(6)
the partial derivatives being evaluated at a suitable point of that segment.
) , , (x y z f
w P0:(x0,y0,z0) P:(x0h,y0k,z0l)
z l f y k f x h f z y x f l z k y h x
f
, , ) ( , , )
( 0 0 0 0 0 0
Special Cases
- For a function f(x, y) of two variables, (6) reduces to
(7) y
k f x h f y x f k y h x
f
, ) ( , )
( 0 0 0 0
and for a function f(x) of a single variable, (6) becomes
(8) dx
hdf x
f h x
f( 0 ) ( 0)
Chap. 9 Vector Differential Calculus. Grad, Div, Curl
Mean value theorem for a function of two variables
9.7 Gradient of a Scalar Field. Directional Derivative.
Definition Gradient
The gradient of a given scalar function f(x, y, z) is denoted by grad f or ∇f and is the vector function defined by
(1)
Here x, y, z are Cartesian coordinates in a domain in 3-space in which f is defined and differentiable.
z k j f y i f x f z f y f x f f
[ , , ]
- The differential operator ∇ is defined by
(1*) k
j z i y
x
x xz y
z y x
f( , , )2 34 3
*
] 4 , 6 , 3 4
[ z y2 x
f
grad
Chap. 9 Vector Differential Calculus. Grad, Div, Curl
The directional derivative or of a function f(x, y, z) in the direction of a vector b is defined by
(5)
If the direction is given by a vector a of any length (≠0), then (6)
Ex. 1) Find the directional derivative of at P:(2,1) in the direction of v = 2i+5j.
Sol) v 29
29 ) 32 40 8 29 ( ] 1
8 , 4 [ ] 5 , 2 29 [ ) 1
(P
f Da
y y x y x
f( , ) 2 34
Directional Derivative
f grad ds b
f df
Db
f grad a a
ds f df
Da 1
Chap. 9 Vector Differential Calculus. Grad, Div, Curl
ds f df Db
) 1 ( b
)], 4 3
( , 2
[ 3 2 2
xy x y f
grad gradf(P)[4,8]
Ex. 3) Find the directional derivative of at P:(1,3,0) in the direction of a = [1, 2, -1].
Sol) a 6
2 ) 6
3 6 ( ] 1
3 , 0 , 0 [ ] 1 , 2 , 1 6 [ ) 1
(P
f Da
The minus sign indicates that at
P
the function is decreasing in the direction of a.yz x z y x
f( , , ) sin
Chap. 9 Vector Differential Calculus. Grad, Div, Curl
Ex. 2) Find the directional derivative of at P:(2,1,3) in the direction of a = [1, 0, -2].
Sol) grad
f =
[4x, 6y, 2z], gradf
(P) = [8, 6, 6], a 5789 . 1 ) 12 0 8 5 ( ] 1
6 , 6 , 8 [ ] 2 , 0 , 1 5 [ ) 1
(P
f Da
2 2
2 3
2 ) , ,
(x y z x y z
f
], cos ,
cos ,
sin
[ yz xz yz xy yz f
grad gradf(P)[0,0,3]
Gradient Is a Vector. Maximum Increase
Theorem 1
Vector Character of Gradient. Maximum Increase
Let
f(P) = f(x, y, z)
be a scalar function. Then gradf
is a vector and its length and direction are independent of the Cartesian coordinates. If gradf(P) ≠
0 at some pointP
, it has the direction of maximum increase off
atP
.Theorem 2
Gradient as Surface Normal Vector
Let
f
be a differentiable scalar function. Letf(x, y, z)
= c represent a surfaceS
. Then if the gradient off
at a pointP
ofS
is not zero vector, it is a normal vector ofS
atP
.Gradient as surface normal vector
Chap. 9 Vector Differential Calculus. Grad, Div, Curl
Ex. 2) Find a unit normal vector n of the cone of revolution
at a point 4( )
2 2
2 x y
z
).
2 , 0 , 1 ( : P
Sol.) f(x, y, z)4(x2y2)z2
grad f [8x,8y,2z], gradf(P)[8,0,4] 5]
, 1 0 5 , [ 2 ) ) (
(
1
grad f P
P f n grad
n points downward since it has a negative z-component.
Theorem 3
Laplace's Equation
The Laplace's equation is defined by
(9) 2
2 2 2 2 2 2
z f y
f x
f f
- 2f is also denoted by The differential operatorf. (10) 2 22 22 22
z y
x
is called the Laplace operator.
Chap. 9 Vector Differential Calculus. Grad, Div, Curl
Cone and unit normal vector n
(1)
z v y v x v v
div
1 2 3
is called the divergence of v. For example,
v = [3xz,2xy,yz2] 3xzi2xyjyz2k, then div v3z2x2yz Another common notation for the divergence is
z v y
v x
v
k v j v i v z k
y j x i
v v z v
y v x
v div
3 2
1
3 2
1 3 2 1
] [
) (
] , , [ ] , , [
Note that ∇ v means the scalar div v, whereas ∇
f
means the vector gradf
defined in Sec. 9.7.
9.8 Divergence of a Vector Field
Let v(x, y, z) be a differentiable vector function, where x, y, z are Cartesian coordinates, and let v1 , v2 , v3 be the components of v.
Then the function
Chap. 9 Vector Differential Calculus. Grad, Div, Curl
Theorem 1
Invariance of the Divergence
The divergence
div v
is a scalar function, that is, its values depend only on the points in space (and, of course, on v) but not on the choice of the coordinates in (1), so that with respect to other Cartesian coordinatesand corresponding components of v,
(2) div v =
*
*
*, y , z
x v1*, v2*, v3*
*
* 3
*
* 2
*
* 1
z v y v x v
Let
f(x, y, z)
be a twice differentiable scalar function. Then its gradient exists,2 2 2 2 2 2
) (
] , , [
z f y
f x
f f grad div v div
z k j f y i f x f z f y f x f f
grad v
Hence, the divergence of the gradient is the Laplacian.
(3) div (grad f) 2f
Chap. 9 Vector Differential Calculus. Grad, Div, Curl
Ex. 2) Flow of a Compressible Fluid. Physical Meaning of the Divergence
We consider the fluid motion in a region R having no sources, or sinks, and consider the flow through a rectangular box B.
k v j v i v v v
v1, 2, 3] 1 2 3
[
(4) uv [u1,u2,u3] u1iu2ju3k and assume that u and v are continuously differentiable vector functions of x, y, z and t.
The mass of fluid entering through the left face ΔxΔz during a short time interval Δt is given by
where the subscript y indicates the left face.
t z x u t z x
v )y ( )y
( 2 2
The mass of fluid leaving the box B through the opposite face during the same time interval is approximately
where the subscript indicates the right face.
] ) ( ) (
[ 2 2 2
2
2 V t u u y y u y
y t u z x
u
t z x u )yy ( 2
y y The difference
is the approximate loss of mass.
Chap. 9 Vector Differential Calculus. Grad, Div, Curl
Physical interpretation of the divergence
Let v be the velocity vector of the motion. We set
If we equate both expressions, divide by , and let and approach zero,
Vt
x,y,z t v t
div u
div
( )
0 )
(
div v
t
or
(5) condition for the mass conservation, or continuity equationof a compressible fluid flow
If the flow is steady, then =0 and / t (6) div (v)0
If the density is constant, so that the fluid is incompressible, (7) div v = 0
* The divergence measures outflow minus inflow.
This loss of mass in B is caused by the time rate of change of the density and is equal to V t
t
t z V u y u x
u
)
( 1 2 3
where u1(u1)xx(u1)x and u3(u3)zz (u3)z
Two similar expressions are obtained by considering the other two faces. If we add these three expressions, the total mass of loss in B during the time interval Δt is approximately
Chap. 9 Vector Differential Calculus. Grad, Div, Curl
condition of the incompressibility
9.9 Curl of a Vector Field
Let
v(x, y, z) = [ v
1, v
2, v
3] = v
1i+v
2j+v
3k
be a differentiable vector function. Then the curl of the vector function is(1) k
y v x j v x v z i v z v y v v
v v
z y x
k j i v v
curl ( 3 2) ( 1 3) ( 2 1)
3 2 1
⇐ x, y, z are right-handed Ex. 1) Curl of a Vector Function
v = [yz, 3zx, z] = yzi + 3zxj +zk. Then
zk yj xi k
z z yj xi z
zx yz
z y x
k j i v
curl 3 (3 ) 3 2
3
Chap. 9 Vector Differential Calculus. Grad, Div, Curl
Theorem 1
Rotating Body and Curl
The curl of the velocity field of a rotating rigid body has the direction of the axis of the rotation, and its magnitude is twice the angular speed of the rotation.
Theorem 2 Grad, Div, Curl
Gradient fields are irrotational. If a continuously differentiable vector function is the gradient of a scalar function
f
, then its curl is zero vector,(2) curl (grad f ) = 0 and
(3) div (curl v) = 0
Theorem 3
Invariance of the Curl
Curl v is a vector. That is, it has a length and direction that are inde- pendent of the particular choice of a Cartesian coordinates system in space.