Você está na página 1de 8

Vector Differential Calculus

51

The element o f arc ds . on C is then defined from elementary calculus by d s 2 = dx 2 + d y 2 + d z 2. From (3.53) and the parametric equations, ds dt % * d tj , d ) A \dt J J di J \dt (3.53)

= { [ f '( 0 3 2 + E f/C O l* + [ i / ( i ) ] 2S1/2 = [f"(0


= |f '( 0 | H ence, the element of arc ds is ds = \ f '( t) | dt. T hus, the total length / is the integral of ds over C; i.e ., / f ds = (3.55) (3.54)

f*

|f.'(0 \dt = I b [ f'(f) f V ) T /2 dt. -'a

P R O B L E M 3.17

F in d the length of the curve represented by the vector function f ( 0 = (a cos t ) i + (a sin t) j, 0 < f < 2 jt . [3.41]

Solution:

The derivative of f(f) is f '(i) = - a sin t i + a cos t j, 0 < t < 2 jt , (3.56)

and its magnitude is |f '(f) | = (a2 s in 2 t + a 2 cos2 t f '2 = a. T hus, from (3.52), the length of the curve is
nZTT ' 277 f r-2n I

(3.57)

-I J0

V{ t ) \d t = \f'(t)\dt =\ Jq

a dt - 2na.

(3.58)

The arc length s (i) of a curve is a function of the scalar variable t from some fixed point a to t. It is the length I of C with the fixed upper lim it h in (3.52) replaced by t; i.e ., dt. * /a P R O B L E M 3.18 Show that the arc length s may serve as a parameter in the rep (3.59)

resentations of curves with no singular points.


So lu tion :

D ifferentiating the arc length, (3.59), of a curve, we obtain for a < t < b,

= |r(0 S > o.
dt Thus, s is an increasing and continuous function in a < i <

(3.60)
T his im plies that

s has a unique inverse t = q (s) for 0 < s < I. Hence, a smooth vector function g ( s ) | f [ q (s)] (3.61)

with parameter s can represent the same curve C that is represented by f ( i ) . By setting the position vector r = [x, y, z ] as r = g (s) = ^ ( s ) i + g2(s) j + g 3 (s )k , (3.62)

52

Vector A n a lysis

we obtain the parametric equations of C w ith parameter s; i.e., x = i(s ), y = g2( s ), z = g3(s). (3.63)

Instead of the vector functions f (f) or g (s ) to represent the space curve C, we often use r(f) or r(s ). T hus, f (i) = r(f) = x ( i ) i + y (t) j + z (t )k , g (s ) = r (s ) = x ( s ) i + y (s) j + z (s )k , a<t<b, 0 < s S I. (3.64) (3.65)

Although the same fu nctional forms, r (f) and r(s), are used, it should be noted that r(s ) is not obtained from r(f) by simply changing the parameter t to s. Rather r(s ) is obtained from r ( i) by c h an g ing the parameter t to q(s). Problem 3.19 illustra te s th is point.
P R O B L E M 3 .1 9

A curve C is denoted by the vector function f (f) = a cos t i + a s in t j, 0 < i < 2ff.

Represent C by a vector function (C f., Prob. 3.17.)


Solution:

w ith

the arc length s as the parameter.

From the d e fin itio n of arc length (3.59), and using (3.57), s = f |f '( t ) | dt = f a dt = at. (3.66)

Jo

Jo

Hence, t = s / a , and the desired vector fu nc tio n is g (s) = f a cos |


l +

a sin |

j,

0<s<277a.

(3.67)

P R O B L E M 3 .2 0 When a curve C is represented by a vector function g(s) with arc length s as the parameter, show that

ds

^ = g '( s ) = T,

(3.68)

where T is the unit tangent vector to C at any point.


So lution:

From (3.40), the tangent vector to the curve C represented by the

smooth vector function f ( f ) at any point is the vector f'(i)- Hence, for |f ' ( 0 | ^ 0, the unit tangent vector T to C at any point is expressed by ^
f'(0

no

KJ

(3.69)

Now since g (s) = f[q (s)] and t = q (s ), we have on differentiating and using the chain rule,

dg(s) - g '( s )
ds Hence, from (3.69-70),

= f'(f)

ds

ds dt

=J^L .
|f '(t) |

(3.70)

e'(s)=TiIrT P R O B L E M 3 .2 1

F in d the u n it tangent vector to the curve C represented by the

vector function

Vector Differential Calculus

53

f(f) = (a cos f ) i + (a sin i) j, at t =


ti

0 < t < 2n,

/ 2.

(C f., Prob. 3.13.) From (3.67), the curve C is also represented by

Solution:

S (s )

fs\ s \ j, a cos [ j i + a sin (

w.

W.

0<s<2;

[3.67]

From (3.68), the unit tangent vector T to C at any point is ( s\ (s \ T = g '(s) = sin i + cos j, \ a l. . \ a ). Now, from (3.66), we have at t = 77/ 2 , s = af|( or s = a n /2 , the unit tangent vector is 7 7 \ i + cos ( n\ j = sin / \2j. \2/. (3.72) 0 < s ii 2 n a . (3.71)

2 = \an. Hence, at point t = 77/2

since sin ( 77/ 2 ) = 1 and cos (n /2 ) = 0. A lternatively, the same result can be obtained from (3.69) by using (3.56) and (3.57). Thus, for 0 < t < 277, T = - = [(a sin t ) i + (a cos O il I I '( t ) | a -(sin t) i + (cos t) j. Hence, at t = 77/ 2 , =

(3.73)

( n\ i + sin . \2/ .

C OS

f 77\ J = - 1. V2 /

P R O B L E M 3 .2 2

Show that the vector defined by d T / d s is normal or perpendicu

lar to the unit tangent vector T at any point on the curve C. Solution: Since T is the unit tangent vector, the dot product yields T T = 1. D ifferentiating both sides with respect to s, dT ds T = 0, (3.75) (3.74)

which im plie s that d T / d s is normal to T. (See Prob. 3.7.)

3.5

Surfaces

In Sec. 3.4, vector equations of the type r ( i) = x ( i ) i + y ( 0 i + z(t)k (3.76)

in the sin gle parameter t describe space curves. The parametric representation of the space curves is x = x (f), y = y (t), z = z (t). (3.77)

Surfaces, in general, are described by the parametric equations of the type


X

(u, v),

y = y (11, v),

z = z

(u ,

v),

(3.78)

where u and v are parameters. If v is fixed, i.e ., v = c, a constant, then (3.78) becomes a one-parameter expression, which describes a space curve along which

54

Vector A nalysis

u varies. T his is the curve designated by v = c. Thus for each v, there exists a space curve. Sim ilarly, v varies along the curve u = k, a constant. The locus of all the curves v = c and u = k forms a surface S. The parameters u and v are called the curvilinear coordinates of the point s on the surface, and the u-curves and v-curves are calle d parametric curves. See F ig . 3.5. If the terminal point of the position vector r generates the surface S, then (3.78) can be rewritten as r (u, v) = x (u, v) i + y (u, v) j + z (u, v) k. (3.79)

P R O B L E M 3 .2 3

Show that the unit vector n, defined by ru x rv


m = --------------

F ig . 3 .5

A s u r f a c e in s p a c e .

(3.80)

where ru = ru x rv 4 0. Solution:

ou

rv =

dv

is normal to the surface S represented by r(u, v) if

According to (3.40), at any point P , dr du dx dy . dz x+ ~ J + k du du du Sim ilarly, dz ov (3.82) (3.81)

is a tangent vector to a constant curve v at P . dr ov dx .


l

dy . dv

dv

is a tangent vector to a constant curve u at P . Hence, it follow s that at any point P , the vector ru x r v is normal to the surface S at P . Since |ru x t v | is the magnitude of this vector, x rv
n =

is a unit normal vector to S at P .

A point (u, v) on a surface S is called a singular point if ru x r v . = 0; other wise, it is called a nonsingular point. If ru and rv are continuous, then the tan gent planes exist only at the nonsingular points. Geom etrically, the condition for ru x r v ^ 0 is that the curves u = k and v = c, where k and c are constant, are nonsingular and are not tangent to each other at their point of intersection.

P R O B L E M 3 .2 4

F ind a unit normal vector n for a surface S represented by x=x, y = y, z = z (x ,y ), (3.83)

where x and y are parameters. Solution: From (3.83), the surface S can be represented by r ( x ,y ) = x i + y j + z ( x ,y ) k . Hence, the partial derivatives are dr dx dy . dz . dz rx = = i + j + - k = 1 + -k, ox dx dx dx dx dr dx . dy . dz . dz . (3.85) , (3.84)

rJ' = T _ =T 1+ d j y J +s T _k = 3+T _kdy dy dy dy

(386)

Vector Differential Calculus

55

Now, the vector product of r x and r y is


i j k

dz !Tx dz d

dz .

dz .

(3.87)

0 T hus, the magnitude is

|Y + C 4 |+ 1
y H ence, from (3.80), the u n it normal vector is dz .
x r r v x rv

(3.88)

dz .

T xl ^
dz\2 dy

3 +k
(3.89) +1

/ d z x2

P R O B L E M 3 .2 5

F ind a unit normal vector n for a surface S that is represented

by <f>{x,y, z) = 0. (3.90)

Solution: If we regard (3.90) as im p lic itly defining z as a function of x and y, we can assume that <pz 4 - 0. Then from elementary c alc u lu s, (3.85), and (3.86), the p artial derivatives are d<fi

dz
r = i +

dx d$ dz d< f >

<bx < f)z


(3.91)

dx

k = i ------------k = i ---------k,

dz
ry = j +

dy d<f>

c f > 4>z
(3.92)

dy

k = j --- k = j - k.

H ence, the vector product is 1


r x rv

J 0
-

< t> z 0 1 4>y


-

JL_
:

k),

(3.93)

so that the unit normal vector is


rv x rv Irx x ry I <x i + <y j + c f> zk [(^>x) 2 + ( 0 y) 2 + ( , ) 1 *

d(b . dx

d(f) . dy dy)

dcf) dz \ <3z (3.94)

d<t>Y J d A ' / d c f ,

56

V e c t o r A n a l y s is

If (f> x 4 0 or c f> y4 = 0, a sim ilar com putation would y ie ld t h e same result.

The diffe rential elem ent o f surface area is a vector g iv e n by dS = du x d v = tu x r v d u d v . (3.95)

du

ov

From the definition of a u nit normal vector (3.80), we se e that dS is a vector normal to the surface represented by r(u , v) at any p o in t P ; its magnitude, d S = |d S | = |ru x rv |dudv, (3.96)

is approximately equal to the surface area S that is b o u n d e d by four curves on S. (See Fig. 3.6.) From (3.80) and (3.96), (3.95) can be rewritten as c?S = n d S . (3.97)

The surface area of S can then be obtained by in te g ra tin g (3.97) overS; i.e.,
F ig . 3 .6 A d i f f e r e n t i a l e l e m e n t of su rface area.

S= JJ'dS^JJn-dS,
where n is a unit normal vector o f S at any point P .

(3.98)

P R O B L E M 3 .2 6

F in d the surface area for a surface S t h a t is represented by y = a sin 0 sin < f> ,

x = a sin 6 cos cf> , where 0 <


Solution:
6 < n

z = a cos 6,

(3.99)

and 0

<

< 2 tt.

Here the two parameters 0 and c f> are used in s te a d of u and v. Hence,

dr

dx .

dy .

dz

T6 = J o = ~ d e 1 + J d i + d d

= a cos 6 cos < f) i + a cos 6 sin < f > j a sin 6 k,

< 9 r
r cb ~

dx .

dy .
~

dz
+

.
k

d(f>

dcf>

i +

d(f>

dcf>

= a sin 6 sin c f> i + a sin 6 cos c j . Thus, the vector product is i j cos
< f>

k sin
cf>

r# x

T cf, =

a cos

a cos

- a s in Q

a sin 6 sin (f>

a sin 6 cos (f>

= a 2 s in 2 6 cos < f > i + a 2 s in 2 6 sin < f > j + a 2 s in d cos 6 k, and its magnitude is |*d x r< | = a 2 sin 6 [sin 2 0 cos 2 < j) + s in 2 6 s in 2 9S + cos 2 0]^ = a 2 sin 0 [sin 2 6 (cos 2 c f> + s in 2 cf> ) + cos2 = a 2 sin 0. Hence, from (3.96), the m agnitude o f the differential elem ent of the surface area is dS = a 2 s in 6 dO dtp. Thus, on integrating, the surface area is (3.100)

V ecto r D iffe re n tia l C alcu lus

57

' -

r e

a 2 s in 6 dQd(f> = 277a 2

J"

sin 6 dO

= 277a2 (-cos 6)\ c l: \


= 2rra2{-[(-1) - l] i
= 47 7 a2.

It is noted that (3.99) are the param etric equations for a sphere x2 + y2 + z 2= a 2 . (3.101)

From F ig . 3.7, (3.100) can be ob ta ine d by observation. Thus, if we hold < fixed and vary 6 by d d , we ob ta in an arc o f length a dO. H olding 6 fixed and varying < f> by d< + > , we obtain an arc o f a c irc le of radius a sin 0 with the length, a sin 6 d<j>. Hence the elem ent o f d iffe re n tia l area dS can be expressed as dS = a d O -a sin 6 dcfr = a 2 sin 0 d d d f i.

PRO BLE M 3.27

Express the d iffe re n tia l element of surface area d S for a sur

face S that is represented by z = z (x , y ). Solution: From (3.88), (3.102)

l +(

(
\d y

dxj
Hence, using (3.96), dS = |rx x r y | dxdy

dxj

\dy

dxdy.

(3.103)

58

Vector A nalysis

3.S

Directional Derivative and Gradient

A scalar fie ld 0 (x, y, z) is the totality of scalars 0 (x, y, z ) assigned to each point (x, y, z ) of a region R in space. A vector fie ld f(x , y, z ) is the to ta lity of vectors f(x , y, z ) assigned to each point (x, y, z ) of a region R in space. L et P (x , y, z) be a nonsingular point on a curve C in space and Q (x + A x , y + Ay, z + A z), any other point on C. (See F ig . 3.8.) Then the position vec tors r and r + Ar of P and Q are r = x i + y j + zk, r + Ar = (x + Ax) i + (y + Ay) j + (z + A z) k. Let 0 ( x , y, z) be a continuous and differentiable scalar function in a region R that contains the arc o f C from P to Q. Then, the directional derivative of 0 ( x , y, z) at P in the direction of the unit tangent vector T to C at P is de fined as ds = lim (* + As >o y + As h^ ^ < ~ X y z S > (3.104)

Fig . 3.8

Th e directional d eriva tive .

where As is the arc length of C from P to Q. The gradient of the sc a la r function d > (x ,y ,z), written as grad defined by grad 0 s is a vector

d<f> .

ox

t + _ j + -T. fc. dy dz

dcf) .

dcf>

(3.105)

P R O B L E M 3.28 Show that the directional derivative of S(x, y, z) in the direc tion of the curve C can be expressed as dd> = grad 0 T, (3.106)

as

where T is the unit tangent vector to C. Solution: From (3.68), the unit tangent vector T to C at any point is _ dr dx . T = = 1 ds ds Now from calc u lu s, <90 ds d 0 dx ^ <90 dy dx ds dy ds <90 d z dz ds (3.108) dy .
+ - J

ds

dz . + k. ds

(3.107)

which is the form of a sc a la r product of grad 0 and T. Hence, d0 = grad 0 T.

By introducing the diffe re n tia l operator V (read nabla or del),

U V =
the gradient of 0 is w ritten as

3 d d l + j + k, dx dy dz

(3.109)