Next: 3.4 Principal curvatures
Up: 3. Differential Geometry of
Previous: 3.2 First fundamental form
Contents Index
3.3 Second fundamental form II (curvature)
Figure 3.6:
Definition of normal curvature
In order to quantify the curvatures of a
surface
, we consider a curve
on
which passes through point
as shown in Fig. 3.6. The unit tangent vector
and the unit normal vector
of the curve
at point
are related by (2.20) as follows:
(3.22)
where
is the normal curvature vector and
is the geodesic curvature vector which are the components of the
curvature vector
of
in the surface normal direction and
in the direction perpendicular to
in the surface tangent
plane. Thus, the normal curvature vector
can be expressed as
(3.23)
where
is called the normal curvature of the surface at
in the direction
. In other words,
is the magnitude
of the projection of
onto the surface normal at
, with
a sign determined by the orientation of the surface normal at
.
By differentiating
along the curve
with respect to
we obtain
(3.24)
thus
(3.25)
(3.26)
where
(3.27)
Since
and
are perpendicular to
, we have
and
, and hence
we have an alternative expression for
,
and
(3.28)
Computation of curvatures at points where the surface representation is
degenerate (see Sect. 1.3.6) is given in [453].
The numerator of (3.26) is the second
fundamental form
, i.e.
(3.29)
and
,
,
are called second fundamental form coefficients.
Therefore the normal curvature is given by
(3.30)
where
is the direction of the tangent line
to
at
. We can observe that
at a given point
on
the surface depends only on
which leads to the following
theorem due to Meusnier.
Theorem 3.3.1.All curves lying on a surface
passing
through a given point
with the same tangent line have the
same normal curvature at this point.
Using this theorem we can say that the normal curvature is positive
when the center of the curvature of the normal section curve, which is
a curve through
cut out by a plane that contains
and
is on the same side of the surface normal (see Fig.
3.7 (a)). Sometimes the positive normal
curvature is defined in the opposite direction, i.e. the center of
curvature of the normal section curve is on the opposite side of the
surface normal as illustrated in Fig.
3.7 (b).
In such cases (3.23) (3.30) become
(3.31)
The latter convention is often used in the area of offset curves and
surfaces in the context of NC machining. Throughout this book we
refer to the first convention as convention (a) and to the second one
as convention (b). We have listed all the equations, which involve
changes due to this convention in the last page of this chapter.
Figure 3.7:
Definition of positive normal curvature: (a)
, (b)
Figure 3.8:
Geometrical illustration of the second fundamental form
Suppose
is a point on a surface and
is a point in the
neighborhood of
and
is the surface
containing
and
, as in Fig. 3.8. Now suppose
and
are the points
and
,
then Taylor's expansion gives
(3.32)
Therefore
(3.33)
Thus using (3.28), (3.29), the projection
of
onto
is
(3.34)
where the higher order terms are neglected
and since
, we get
(3.35)
Thus
is equal to twice the distance from
to the
tangent plane of the surface at
within second order terms.
We want to observe in which situation
is positive and negative or
in other words we want to examine
in which side of the tangent plane
lies.
When
, (3.35) becomes
, which can be considered as a quadratic
equation in terms of
or
.
If we solve for
, assuming
, we obtain
(3.36)
which leads us to the following four cases:
If
, there is no real root. This means
there is no intersection between the surface and its tangent plane
except at point
. Point
is called elliptic point
(Fig. 3.9(a)). For example, an ellipsoid consists entirely of
elliptic points.
If
and
, there are
double roots. The surface
intersects its tangent plane with one line
,
which passes through point
. Point
is called parabolic
point (Fig. 3.9(b)). For example, a circular cylinder
consists entirely of parabolic points.
If
, there are two roots. The surface
intersects its tangent plane with two lines
, which intersect at point
Point
is called hyperbolic point (Fig. 3.9(c)). For example, a
hyperboloid of revolution consists entirely of hyperbolic
points.
If
, the surface and the tangent plane have
a contact of higher order than in the preceding cases. Point
is
called a flat or planar point.
If
and
, we can solve for
instead of
.
If
and
, we have
, thus the iso-parametric lines
,
will be the two intersection lines.