My Notes On Faber1983
My Notes On Faber1983
My Notes On Faber1983
Robert B. Scott,1,2
1
Institute for Geophysics, Jackson School of Geosciences, The University of Texas at Austin,
Austin, Texas, USA
2
Department of Physics, University of Brest,
Brest, France
1.1 Curves
Exercises I-1
p
ds = (x0 dt)2 + (y 0 dt)2 + (z 0 dt)2
= a2 + b2 dt (1.1)
Thus the length of the helix is 2 a2 + b2 .
3. (a) Note the typo in this question. It should read . . . and verify that
3
4
||0 (0)|| = V0 .
0 (t) = (V0 cos , V0 sin 32t),
0 (0) = (V0 cos , V0 sin ),
||0 (0)|| = V0 . (1.2)
3. (b)
The acceleration is:
00 (t) = (0, 32). (1.3)
The direction is in the negative ydirection and the magnitude is 32 perhaps
in [ft s2 ].
x2 + (y 1)2 = 1.
(x + 8)2 + (y 6)2 = 53 .
6. (b) Show that the graph y = f (x) has the given curvature.
6
Here
(x) = (x, f (x)).
The unit tangent vector is found from the chain rule:
d(s)
T=
ds
d(x)/dx
=
ds/dx
(1.6)
To find the relation between arc length and x we note that (Boas, 1983,
Chapter 9):
p
ds = (dx)2 + (f 0 dx)2 ,
p
= 1 + f 02 dx.
p
ds/dx = 1 + f 02 . (1.7)
So the unit tangent vector is given by:
d(x)/dx
T(x) = ,
ds/dx
(1, f 0 )
=p . (1.8)
1 + f 02
And we use the chain rule again to find the curvature:
dT(s)
k=
ds
,
dT(x)/dx
=
ds/dx
.
(1.9)
This gets a bit tedious, but is straightforward. Write T = (T x , T y ). The
x component is
dT x (1 + f 02 )1/2
= ,
dx dx
f 0 f 00
= . (1.10)
(1 + f 02 )3/2
Robs notes on Faber 7
7. Let
(t) = (a cos t, b sin t),
for 0 t 2, which defines an ellipse
x2 y 2
+ 2 = 1.
a2 b
Compute the curvature k(t) at t = 0 and t = /2 using the formula from
Exercise 6(a). Well need:
x(t) = a cos t,
x0 = a sin t,
x00 = a cos t.
and
y(t) = b sin t,
y 0 = b cos t,
y 00 = b sin t.
So
a
k(0) = , (1.18)
b2
b
k(/2) = 2 . (1.19)
a
Well need these later for solving Exercise I-2 problem 3.
1.3 Surfaces in E 3
Definition of surface involved a vector-valued function of two variables, X :
D E 3 , defined on an open subset D of <2 . Note that D is an open subset
of <2 . (D will be referred to in later sections without definition.)
Definition of regular on p. 22. In my words, amounts to being able to move
in 2 independent directions on the surface by changing the two parameters
(u, v). In Fabers words, p. 24, the regularity ensures that for each point of
D <2 there is an open neighbourhood about this point X is one-to-one
and there is a continuous inverse function X() . So each point on
X() M has a unique set of coordinates (u, v).
Defines a tangent vector and tangent plane, denoted Tp M , on p. 26.
Exercises I-3
X
X1 = = (f 0 (u) cos v, f 0 (u) sin v, g 0 (u))
u
X
X2 = = (f (u) sin v, f (u) cos v, 0) (1.20)
v
10
Hold v = v0 fixed.
X(u, v0 ) = (f (u) cos v0 , f (u) sin v0 , g(u))
In the x z plane we see the original curve,
(u) = (f (u), 0, g(u))
but squished in the xdirection by cos v0 . In the y z plane its the same
curve but squished by sin v0 . In short, its like the curve (u) has been
rotated by v0 from the x z plane. The coordinates in the x z plane are
those of the shadow one would see if a light were shown through the curve
toward the x z plane. To obtain the y coordinates one shines the light
through the curve toward the y z plane.
2 (a). Sketch profile curve in x z plane and surface in 3D, show X, and
give equation for the surface in the form g(x, y, z) = 0.
has magnitude
|X1 X2 |2 = u2 (1 + a2 )
so for u 6= 0, X is regular.
We can express the surface in the form g(x, y, z) = 0 by
a2 (x2 + y 2 ) z 2 = 0
Helicoid
X = (u cos v, u sin v, bv)
12
By inspection,
(v) = (cos v, sin v, 0)
so then
u(v) = (u cos v, u sin v, 0)
and we require
(v) = (0, 0, bv)
And thus we have expressed the helicoid surface in the form of a ruled surface:
X = (v) + u(v)
Hyperbolic parabola:
X(u, v) = (u v, u + v, 2uv)
By inspection,
(u) = (1, +1, 2u)
so then
v(u) = (v, +v, 2uv)
and we require
(u) = (u, u, 0)
And thus we have expressed the helicoid surface in the form of a ruled surface:
X = (u) + v(u)
X = (v) + u(v)
X1 = (1, 1, 2v)
X2 = (1, 1, 2u) (1.25)
We should also confirm that these equations are indeed the hyperbolic parabola
of Example 5, which was expressed as
1
z = (y 2 x2 )
2
Substituting from X we find z = 2uv and
1 2 1
(y x2 ) = [(u + v)2 (u v)2 ] = 2uv = z,
2 2
confirming it is the hyperbolic parabola.
X 1 X2 = 0
The surface will be regular when the tangent vectors are not parallel, which
means that X1 X2 6= 0.
The uparameter curves are given by (u), a general smooth function of
u producing a curve in E 3 .
The vparameter curves are given by (u0 )+v. These are straight lines
in the direction of the vector that pass through (u0 ).
14
7. Let (u) = (cos(u), sin(u), 0). Through each point of (u), pass a unit
line segment with midpoint (u) and direction vector
u u
(u) = sin (u) + cos (0, 0, 1)
2 2
The resulting surface,
1 1
X(u, v) = (u) + v(u), v
2 2
is called a Mobis strip!
Exercises I-4
1. Derive the formula ds2 + dr2 + r2 d2 for the differential of arc length
in polar coordinates by substituting
x = r cos y = r sin
into the formula ds2 = dx2 + dy 2 for the differential of arc length in Cartesian
coordinates.
x(r, ) x(r, )
dx = dr + d = cos dr r sin d
r
y(r, ) y(r, )
dy = dr + d = sin dr + r cos d (1.27)
r
Squaring and adding we find the cross terms cancel, leaving the desired result
ds2 + dr2 + r2 d2 .
2(a). L = 2(exp() exp())
2(b). L = 8a
3(a). Find (gij ), (g ij ), det (gij ), U, where U is the unit normal, for the
sphere of Example 7, see also Fig. I-16.
g11 = X1 X1 = R2 cos2 v
g12 = X 1 X2 = 0
g21 = g12 = 0
g22 = X 2 X2 = R 2 (1.29)
g ij = 1/gij
The determinant is
det(gij ) = R4 cos2 v
The unit normal vector U is
X1 X 2
U=
||X1 X2 ||
R2 (cos u cos2 v, sin u cos2 v, cos v sin v)
=
||R2 (cos u cos2 v, sin u cos2 v, cos v sin v)||
= (cos u cos v, sin u cos v, sin v) (1.30)
5. Compute the metric form (or first fundamental form) and the unit
normal vector U for the general surface of revolution (Exercise 1, Section 3).
X
X1 = = (f 0 (u) cos v, f 0 (u) sin v, g 0 (u))
u
X
X2 = = (f (u) sin v, f (u) cos v, 0) (1.31)
v
Robs notes on Faber 17
The metric form (or first fundamental form) (see Eqs. 6 & 4 and p. 35) can
be written:
ds2 = g11 du2 + 2g12 du dv + g22 dv 2
with
g11 = X1 X1 = f 02 + g 02
g12 = X1 X2 = 0
g22 = X2 X2 = f 2 (1.32)
X 1 X2 X1 X2
U= =p
||X1 X2 || det(gij )
(g f cos v, g f sin v, f 0 f )
0 0
= p
f f 02 + g 02
(g 0 cos v, g 0 sin v, f 0 )
= p (1.33)
f 02 + g 02
6(a). Show that the area A of the general surface of revolution (Exer-
cise 1, Section 3) between a u b and 0 v 2 is
Z b p
A = 2 f (u) f 0 (u)2 + g 0 (u)2 du
a
6(b). Show that the area A of the surface obtained by revolving the
graph y = f (x), a x b about the xaxis is
Z b p
A = 2 f (x) 1 + f 0 (x)2 dx
a
Exercises I-5
Most of these are straightforward computations. Exceptions are 1, 6, and
8. The first is quite easy. Here are my solutions for 6 and 8.
If one jumps in and starts computing the curvature of the surface etc.
without thinking through whats going on, one can quickly get lost in com-
plicated computations. (I know from personal experience!) On the other
hand if one understands Eqs. (16, 17, 18), then the answer if very obvious.
00 (s)
is the vector proportional to the curvature of the curve at point (s). It
00
can be projected onto the tangent plane of the surface giving tan and the
00
remainder projects onto the normal to the surface U giving nor , Eq. (14).
The curvature of the surface in the direction of the curve T(s) = 0 (s) is the
magnitude
00 0 0
||nor || = Lij ui uj
So Meusniers Theorem is essentially contained in Eq. (22)
0 0
kn (v) = Lij ui uj = 00 U = ||00 || cos
where v = 0 (s) is the direction of the curve at point s and is the angle
between 00 (s) and U, noting of course that U is a unit vector.
20
where we have used Eq. (10) on p. 35 relating the determinant of the matrix
g to the magnitude squared of X1 X2 . Here x is the unit vector in the
xdirection, and similarly for the other 2 directions in Euclidean space. Now
since Lij = Xij U (see Eq. (20) p. 44), its the xcomponent of Xij that
multiplies the x component of U. So we can simply replace the unit vectors
above with the corresponding components of Xij , giving the desired relation
xij yij yij
1
Lij = det x1 y1 z1
g
x2 y2 z2
Chapter 6
Extended Indices
d(t)
0 (t)
dt
The derivative vector or velocity of the smooth curve parameterized
by t, see p. 2.
00 d2 (t)
(t)
dt2
The acceleration vector of the smooth curve parameterized by t, see
p. 4.
D
Open subset of R2 , see p. 22.
E3
Three dimensional Euclidean space, see p. 1.
21
22
k(s) = ||T0 (s)||
Curvature of a smooth curve at point (s).
T0 (s)
N(s) =
||T0 (s)||
Principal Normal vector, see p. 6.
s
Arc length of a smooth curve, see pp. 2, 3. Use as a parameter of the
curve, see p. 5.
t
Often used as a real parameter for the curve (t), not necessarily time
but clearly meant to imply that in some contexts.
T(s) = 0 (s)
Unit tangent vector, see p. 5.
T0 (s) = 00 (s)
Curvature vector, see p. 5.
Tp M
Tangent plane at point P of surface M. See pp. 26, 47.
U
Unit normal vector, p. 33,
Robs notes on Faber 23
10. Theorem I-10: p. 68. Given a point P of M and a unit tangent vector
v at P, there exists a unique geodesic (s) such that (0) = P and
0 (0) = v.
Bibliography
25