RELATIVE EXTREMA FOR A FUNCTION

80

CHAP. 3

Suppose x differs only slightly from Qk. Then, cj << Ick for j
Specialize (a) for this case. Hint: Factor out Xk and C%.
(c) Use (b) to obtain an improved estimate for ,.
(b)

The exact result is

) = I

{1, -3}
x = ( 1 - 2)

3-12. Using Lagrange multipliers, determine the stationary values for the
following constrained functions:
2
f = X2 g = X2 + 2 = 0
(b) f = x 2 + x2 + x2

31

a=x[
x

k.

Differential Geometry
of a Member Element

(a)

gl = xl +

= 00

2 + X3-

+ 2 = 0
3-13. Consider the problem of finding the stationary values of f = xrax
xTaTx subject to the constraint condition, xx = 1. Using (3-36) we write
2 =

X1 -

X2 + 2X 3

H =f

(a)

g = x'ax -

(xx

1)

Show that the equations defining the stationary points of f are

xTx = 1
ax = )x

(b) Relate this problem to the characteristic value problem for a symmetri
cal matrix.
T
3-14. Supposef = XTx and g = 1 - xTax =0 where a = a. Show that
the Euler equations for Ii have the form
1
T
xax=
ax= x

The geometry of a member element is defined once the curve corresponding

to the reference axis and the properties of the normal cross section (such as
area, moments of inertia, etc.) are specified. In this chapter, we first discuss the
differential geometry of a space curve in considerable detail and then extend
the results to a member element. Our primary objective is to introduce the
concept of a local reference frame for a member.
4-1.

PARAMETRIC REPRESENTATION OF A SPACE CURVE

A curve is defined as the locus of points whose position vector* is a function

of a single parameter. We take an orthogonal cartesian reference frame having
directions XI, X 2, and X3 (see Fig. 4-1). Let ? be the position Vector to a point
X3

We see that the Lagrange multipliers are the reciprocals of the characteristic
values of a. How are the multipliers related to the stationary values of f ?
X3(y)

13

X2
,

1I

I/

/___

/ /

/' x 1 (0')

X2 (Y)
X1

Fig. 4-1. Cartesian reference frame with position vector f(y).

*The vector directed from the origin of a fixed reference frame to a point is called the position
vector. A knowledge of vectors is assumed. For a review, see Ref. 1.
81

82

DIFFERENTIAL GEOMETRY OF A MEMBER ELEMENT

CHAP. 4

I*
.
.1_
. I_ __. I r
.1
11 .I
on the curve having coordinates Xj
1J , 2, J5)
and let y e tne parameter. we
can represent the curve by
=(y)
(4-1)

*1

X2

. Fig.
.1 E4-1A

Since

xjij, an alternate representation is

j=1

xj = xj(y)

(j = , 2, 3)

(4-2)

Both forms are called the parametric representation of a space curve.

Example 4-1
Cnnider a circ.le in the XY.- YX,nne

(1

(Fi

.4-1A

We tke v s the nolnr anole and

Y.

let a = |r|. The coordinates are

X1 = a cos y
Y_ -

in

'12 -

x, 3

"'"' y
Y

r =acos yi + asin

and

l E4-1B

Y)2

(2) Consider the curve (Fig. E4-1B) defined by

X1 = a cos y

x2 = b sin y

(4-3)

X3 = cy

where a, b, c are constants. The projection on the X1 -X2 plane is an ellipse having semiaxes
n .sst
-nr h Th ntifln
c. Y. s- [

}.'votlull

x/ ort! f-r thiC pllrt

v,.,~,t

t*so

-U*

r = a cos
4-2.

thP f-rlM
tl'.,
t-,-/l l

. h-.
Ito

yl, + b sin y, +

cy'

X2

ARC LENGTH

Figure 4-2 shows two neighboring points, P and Q, corresponding to y and

y + Ay. The cartesian coordinates are xj and xj + Axj (j = 1, 2, 3) and the
, r 11-1
+X_ -1 -I r _ n
I
I
lengtn OI tne cnora rom r to Q is given y
y +Ay)

IP-Q2
As Ay - 0, the chord length

j=1

(AXj) 2

(a)

IPQI approaches the arc length, As.

In the limit,

ds2 =

dxJ

(b)

j=1

Noting that
dxj =

dx
dy
dv

(c)

we can express ds as
/(dx
[1 \2

UO-

LTi

(dX2'2

dx32]1

"y
LdLY
Y
It3

I
I,

Fig. 4-2. Differential segment of a curve.

83

MEMBER ELEMENT

nhcr-TPY nF A

,.,,,,,f,,,I

UltItNt"l-I

84

,V-

I IL

...

._.

..

CHAP. 4

4-3.

Finally, integrating (4-4) leads to

dy )
Y=dX2)2

s(y)

SEC. 4-3.

+ (k)

21

dy

(4-5)

It is customary to
We have defined ds such that s increases with increasing y.
call the sense of increasing s the positive sense of the curve.
To simplify the expressions, we let

We consider again the neighboring points, P(y) and Q(y + Ay), shown in
Figure 4-3. The corresponding position vectors are 7(y), (y + Ay), and

Note that
One can visualize a as a scale factor which converts dy into ds.
1.
+
x> 0. Also, if we take y = s, then a

t = lim
Ay-'

dr
ds

PQ
IPQI

=-

ds

is satisfied.
We suppose that b > a. One can always orient the axes such that this condition
asC2
a
express
we
Then,
2
2
2
2 12
a= (b2 + c ) [1 -k sin yll

1 dr

df dy
dy ds

da

is
Consider the curve defined by (4-3). Using (4-6), the scale factor
2 12
2
2
2
2
a = [a sin y + b cos y + c ] /

ady

drL1/2

-Ty dyy
Equation (4-10) reduces to (4-6) when
coordinates.

b2 - a

b +r c2
2

The arc length is given by

s=

x dy

(62

[1

c2)12

2
k sin

1
y] /2

r(y

dy

by E(k, y).
The integral for s is called an elliptic integral of the second kind and denoted

Then,

2 i 2

s = (b2 + C )

E(k, y)

b = a, the curve
Tables for E(k, y) as a function of k and y are contained in Ref. 3. When
to
reduce
is called a circularhelix and the relations
a = (a2
s = ay

+ C )11 =

(4-9)

Fig. 4-3. Unit tangent vector at P(y).

const.
* See Ref. 1,p. 401.

(4-10)

is expressed in terms'of cartesian

(4-8)

Since a > 0, t always points in the positive direction of the curve, that is, in the
direction of increasings (or y). It follows that d:/dy is also a tangent vector and

Example 4-2

k =

(a)

As Ay -- 0, PQ approaches the tangent to the curve at P. Then, the unit tangent

vector at P is given by*

(4-7)

a dy
Yo

where

f:(y)= AP

PQ = (y + Ay)-

Using the chain rule, we can express t as

Then, the previous equations reduce to

ds = a dy
s -=

85

UNIT TANGENT VECTOR

(4-6)

Tx)I 22 /2

UNIT TANGENT VECTOR

86

DIFFERENTIAL GEOMETRY OF A MEMBER ELEMENT

L'mr

.
u-

CHAP. 4

SEC. 4-4.

PRINCIPAL NORMAL AND BINORMAL VECTORS

11

We determine the tangent vector for the curve defined by (4-3). The position vector is
= a cos yt + b sin yi2 + cy
Differentiating

with respect to y,
Normal plane
-=
dy

-a sin yT. + b cos y2 + C 3

Rectify

and using (4-9) and (4-10), we obtain

a = + [a 2 sin2 v + b2 cos2 y + C2 ]112
t = - [-a sin y

+ b cos yT2 + c73 ]

When a = b, a = [a2 + c2] 1/2 = const, and the angle between the tangent and the X
3
direction is constant. A space curve having the property that the angle between the tangent
and a fixed direction (X3 direction for this example) is constant is called a helix.*

Osculating plane
t
Fig. 4-4. Definition of local planes.

4-4.

Example 4-4

PRINCIPAL NORMAL AND BINORMAL VECTORS

Differentiating t t = 1 with respect to y, we have

t -

dy

= O

We determine
and

(a)

It follows from (a) that d/dy is orthogonal to . The unit vector pointing in the
direction of dt/dy is called the principalnormal vector and is usually denoted by 1/.

I dl
d dy

and b for the circular helix. We have already found that

a = [a2 + d
C2

1/2

t= [-asill yT + a cos yi 2 + c73]

Differentiating

with respect to y, we obtain

dt

a
-

[cos Yl
1 + sin yL2J

Then,

where

(4 t11)
di
d 1 d7\
dydy \adyy

The binormal vector, b, is defined by

b= x

//=

..

(4-12)

We see that b is also a unit vector and the three vectors, i, fi, b comprise a right
handed mutually orthogonal system of unit vectors at a point on the curve.

* See Ref. 4, Chap. 1.

Id = -cos

t-

sill y2

The principal normal vector is parallel to the X,-X 2 plane and points in the inwardradial
direction. It follows that the rectifying plane is orthogonal to the X1 -X2 plane. We can
determine b using the expansion for the vector product.
th bxasf the~

122

11

Note that the vectors are uniquely defined once (y) is specified. The frame

associated with , b and ii is called the moving trihedronand the planes deter
mined by (, i), (i, ) and (b, ) are referred to as the osculating normal, and
rectifying planes (see Fig. 4-4).

1 dt

-asinyacosy
This reduces to

- cos v

b
1
The unit vectors are shown in Fig. E4-4.

F3

-sin y 0

- - COS Y 2

+ -a3

a(

87

88

DIFFERENTIAL GEOMETRY OF A MEMBER ELEMENT

CHAP. 4

SEC. 4-5.

K = [-sin 01i

and

7 I

IdsO

,1

+ Cos 02i]7 d
ds

-Id/dsl
i dO/ds

[-sin 07, + cos 0~2]

In the case of a space curve, the tangents at two consecutive points, say P
and Q,
are in the osculating plane at P, that is, the plane determined by and at P.
We can interpret as the radius of the osculating circle at P. It should be noted
that the osculating plane will generally vary along the curve.

x2
I

i
0
i

A1

4-5.

CURVATURE, TORSION, AND THE FRENET EQUATIONS

The derivative of the tangent vector with respect to arc length is called the
curvature vector, K.
di d2;
ds ds2
(4-13)

_:1
12

= id dy

=
Using (4-11), we can write

Y.

ds

Fig. 4-5. Radius of curvature for a plane curve.

(4-14)

Note that K points in the same direction as since we have taken K > 0. The
curvature has the dimension L - and is a measure of the variation of the tangent
vector with arc length.
We let R be the reciprocal of the curvature:
R = K-1

*See Ref. 4, p. 14, for a discussion of the terminology "three consecutive points."

The binormal vector is normal to both and fi and therefore is normal to the
osculating plane. A measure of the variation of the osculating plane is given
by db/ds. Since /bis a unit vector, d/ds isorthogonal to b. To determine whether
db/ds involves , we differentiate the orthogonality condition - b
= 0, with
respect to s.

(4-15)

In the case of a plane curve, R is the radius of the circle passing through three
consecutive points* on the curve, and K = ldO/dsl where 0 is the angle between
t and it. To show this, we express f in terms of 0 and then differentiate with
respect to s. From Fig. 4-5, we have
t = cos 0il + sin 0i 2

89

Then

Fig. E4-4

x3

CURVATURE, TORSION, AND THE FRENET EQUATIONS

db

ds
ds ds
But dt/ds = K and b n = 0, Then, db/ds is also orthogonal to and involves
only i. We express db/ds as
db
d =- -Tn
(4-16)
ds
where T is called the torsion and has the dimension, L-l

90

DIFFERENTIAL GEOMETRY OF A MEMBER ELEMENT

CHAP. 4

SEC. 4-6.

It remains to develop an expression for z. Now, b is defined by

GEOMETRICAL RELATIONS FOR A SPACE CURVE

To determine the component of di/ds in the

orthogonality relation, t- = 0.

b=t xi

dn
ds
It follows from (a) and (b) that

Differentiating with respect to s, we have

Ab
db a x - + t- x d
d
dsIds
ds
This reduces to

db
d

ds

-t

ds

The differentiation formulas for t, i, and

since ni x ii = 0. Finally, using (4-16), the torsion is given by

T

dh

1 . df

-- =

ds

dy

4-6.

(4-17)

Note that can be positive or negative whereas K is always positive, according

to our definition. The torsion is zero for a plane curve since the osculating plane
coincides with the plane of the curve and b is constant.
'-nmo
I...

'

dt
ds

-K

dii
d = - K + zb
ds

dn

= -n-:=

C
Ir
--

are called the Frenet equations.

SUMMARY OF THE GEOMETRICAL RELATIONS FOR

A SPACE CURVE

dr

I d

t = d

-=

tangent vector

1 di
n =ld l
= principalnormal vector

t = - [-a sin yvl + a cos vY2 + c]31

Ca

(4-19)

ii = -cos y'l - s

Yl

= 1 [c sin y -

Difeo
(a2 +

2)1/2

-a=

a_

a2

dy

DifferentiationFormulas (FrenetEquations)

Then,
= -d2

= binormal vector

c cos y12 + aT,]

where

)dy

(4-18)

OrthogonalUnit Vectors

The unit vectors for a circular helix are

K=-

(b)

We summarize the geometrical relations for a space curve:

-____

91

direction, we differentiate the

+ c2

const

dt

I di

ds

c dy

and
1

dii

dy

C2

= -b --- =

= 2
2

c
+

22
c

ds a dy

= const

(4-20)

We have developed expressions for the rate of change of the tangent and
binormal vectors. To complete the discussion, we consider the rate of change
of the principal normal vector with respect to arc length. Since ii is a unit
vector, d/ilds is orthogonal to ii. From (4-17),

dn

W..

f
(a)F

(a)

1 di
= curvature
-I =
b

dy

= torsion

We use the orthogonal unit vectors (, , b) to define the local reference frame
for a member element. This is discussed in the following sections. The Frenet

i
92

DIFFERENTIAL GEOMETRY OF A MEMBER ELEMENT

CHAP. 4

SEC. 4-7

equations are utilized to establish the governing differential equations for a

member element.

LOCAL REFERENCE FRAME FOR A MEMBER ELEMENT

93

are related by
t =t
t2 = cos bf + sin /b

4-7.

t3 = -sin

LOCAL REFERENCE FRAME FOR A MEMBER ELEMENT

Combining (4-21) and (4-24) and denoting the product of the two direction
cosine matrices by PJ,the relation between the unit vectors for the local and basic
frames takes the concise form

The reference frame associated with t, , and b at a point, say P, on a curve

is uniquely defined once the curve is specified, that is, it is a property of the
curve. We refer to this frame as the natural frame at P. The components
of the unit vectors (, ii, b) are actually the direction cosines for the natural
frame with respect to the basic cartesian frame which is defined by the ortho
gonal unit vectors (i, 1i, 13). We write the relations between the unit vectors as

t = pi

(4-25)

where

} =

t[ i
e2l422

==

f_31

3
423

33J

f32

1{
12

1= t2ecos + ft31sin
- 21 sin+e+ e31,,
cos

(4-21)

T3

22 sin

fljk = tj'

= 6jk

j, k =

1, 2, 3

32

sin
COS

t2

--

3 COS

23

sin

sin

33

33 cos 0

k = coS(Yj, Xk)

Since both frames are orthogonal, p =

jm k

+t32

Note that the elements of are the direction cosines for the local frame with
respect to the basic frame.

One can express* the direction cosines in terms of derivatives of the cartesian
coordinates (x1 , X2 , X3 ) by expanding (4-19). Since (t, 6i,b) are mutually or
thogonal unit vectors (as well as it, 12, 13) the direction cosines are related by ;
.

13

22 COS

f12
{

(4-24)

bhi+ cos

(4-26)

- . We will utilize (4-25) in the next

chapter to establish the transformation law for the components of a vector.

(4-22)

m=i

Equation (4-22) leads to the important result

[ Jk]T = [jk]

and we see that

[jk]

(4-23)

is an orthogonal matrix.t

The results presented above are applicable to an arbitrary continuous curve.

Now, we consider the curve to be the reference axis for a member element and
take the positive tangent direction and two orthogonal directions in the normal
plane as the directions for the local member frame. We denote the directions
ofthe local frame by (Y, Y2, Y3 )and the corresponding unit vectors by (t 1 , t 2, t ).
We will always take the positive tangent direction as the Y, direction (t 1 = t)
and we work only with right handed systems (t x t2 = t3). This notation
is shown in Fig. 4-6.
When the centroid of the normal cross-section coincides with the origin of
the local frame (point P in Fig. 4-6) at every point, the reference axis is called
the centroidal axis for the member. It is convenient, in this case, to take Y2,
Y3 as the principal inertia directions for the cross section.
In general, we can specify the orientation of the local frame with respect to
the natural frame in terms of the angle 4)between the principal normal direction
and the Y2 direction. The unit vectors defining the local and natural frames

Y1

X2

X1
* See Prob. 4-5.
t See Prob. 4-6.

Fig. 4-6. Definition of local reference frame for the normal cross section.

DIFFERENTIAL GEOMETRY OF A MEMBER ELEMENT

94

cl
- uurrr

r.XL111 iijl '-V

We determine

CHAP. 4

---

PIfor the circular helix.

a

sin y

-cos y

The natural frame is related to the basic frame by

a

- cos y
-sin y

[ljk]

_siny
-- cosy -JJ
CX
a

1 R
y(4-28)
1RI

--cos y

sin

+ -csiny cos

-sin y cos

- - cos y sinl
i'

sin y sin 4 - -C cos y cos ,

C

(4-2)

The differential arc length along the Yj curve is related to dy by

+ - sin y sin

95

The curve through point Q corresponding to increasing yl with Y2 and y3

held constant is called the parametric curve (or line) for y1. In general, there
are three parametric curves through a point. We define ij as the unit tangent
vector for the yj parametric curve through Q. By definition,

k}

Using (4-25)

cos

CURVILINEAR COORDINATES FOR A MEMBER ELEMENT

(X

SEC. 4-8.

--Cos
1

ds = I dyj = gj dyj
(4-29)
JOTdy
!~
l
y
This notation is illustrated in Fig. 4-8.
One can consider the vectors u
(or aR/Oyj) to define a local reference frame
at Q.
X3

4-8.

CURVILINEAR COORDINATES FOR A MEMBER ELEMENT

We take as curvilinear coordinates (yt, Y2, y,) for a point, say Q, the parameter
y, of the reference axis and the coordinates (Y2, y3) of Q with respect to the
orthogonal directions (Y2, Y3) in the normal cross section (see Fig. 4-7). Let
-- R(y 1 , Y2, Y3) be the position vector for Q(yl, Y2, Y3) and
= (y,) the
position vector for the reference axis. They are related by
R = r + Y 2 t2 + y 3 t 3

where

t2 - t(yI) = cos 4

+ sin 4ib

t3(Yl) = -sin 4,i + cos

We consider 4 to be a function of y,.
t3 =

(4-27)

axis

Y2

X1
Fig. 4-8. Vectors defining the curvilinear directions.
Y3

Y1

Operating on (4-27), the partial derivatives of R are

AR

Y2,

dF

aYi = dy

df 2

dT3

I
,./ '"-`

Reference axis

Fig. 4-7. Curvilinear coordinates for the cross section.

0y2

OR,

(a)

CHAP. 4

DIFFERENTIAL GEOMETRY OF A MEMBER ELEMENT

96

97

REFERENCES

system by taking

We see that
g2 = 1
g3 = 1

ii2 = t2
c23 = t3

(4-30)

dy
which requires

It remains to determine ti and g.

Now,
d:

= -ycr
o'

-=

(a)

at = t ,

Cos
d 3 _

dvl

b
-si(dy, -Y+ dy,
-sn0(

+ cos

-d

d3 = OK sin

dy

os

-(

c +

= cx(

- Ky)t,

91 =

d4v

dyl,

and b.

Then,

4) t3

(c)

(4-34)

U1 = l

(l - Ky2 )

In this case, the local frame at Q coincides with the frame at the centroid. One
should note that this simplification is practical only when cTcan be readily
integrated.
cIrl
Pr

rmn

a-

The parameters a and r are constant for a circular helix:

))2

= (a2 + C2) 1 /2

and finally,
OR

T+

(4-33)

and
(b)

We use the Frenet equations to expand the derivatives of

dt = -K

ay,

dO

k(

do ++ sin

dy

When (4-32) is satisfied,

dy l

Also,
+d

(4-32)

- Ky')T

aye

y2 = Y2 cos

4 - Y3

sin

(Y2+

- Y3 2 )

(4-31)

a2

Then,

4)

We see from Fig. 4-9 that y2 is the coordinate of the point with respect to the
principal normal direction.

and integrating (4-33), we obtain

5= -(Y

Y)

= -rs

Y
For this curve, q5varies linearly with y (or arc length). The parameter g1 follows from (4-34).
ds,
gc y

(1
a(

Ky)

REFERENCES
n
Fig. 4-9. Definition of y;.

Since OR/ay, (and therefore uft) involve t2 and t3 , the reference frame defined
by ii,, u2, U3 will not be orthogonal. However, we can reduce it to an orthogonal

1. THOMAS, G. B., JR.: Analytical Geometry and Calculus, Addison-Wesley Publishing

Co., Inc., Reading, Mass., 1953.
2. HAY, G. E.: Vector and Tensor Analysis, Dover Publications, New York, 1953.
3. JAHNKE, E., and F. EMDE: Tables of Functions,Dover Publications, New York, 1943.
4. STRUIK, D. J.: Differential Geometry, Addision-Wesley Publishing Co., Reading,
Mass., 1950.

98

DIFFERENTIAL GEOMETRY OF A MEMBER ELEMENT

CHAP. 4

PROBLEMS

4-4. The equations for an ellipse can be written as

PROBLEMS

4-1. Determine t, h, , , K, for the following curves:

(a) xl = 3 cos y
x2 = 3 sin y
x 3 = 5y
(b) xl = 3 cos y
x2 = 6 sinl y
x 3 = 5y
(c)
= YIl + y212 + 33
(d) xl = aeo Y cos y
x = ae y sin y
X3 = Cy
where a, /f, c are real constants.
4-2. If x,3 0, the curve lies in the X,-X2 plane. Then, zr 0 and b = i3.
The sign of b will depend on the relative orientation of hiwith respect to t.
Suppose the equation defining the curve is expressed in the form
X2 = f(x l)

x = acos y
2
_

1 dx,
lk =a

4a
g(xIb -

e2k

Deduce that a = sec 0. Express , hi, , and K in terms of 0.

(d) Specialize (c) for the case where 02 is negligible with respect to unity.
This approximation leads to
0

A curve is said to be shallow when 02 << 1.

Let K = 1/R and r = 1/R. Show that (see (4-20))
dy

[k1

Rt

dy

dv2

(df
Edy1

k22

e32- =
Let

el3

-k--Id (d2xk -

I da dx,
a dy dy
I da dXk2

1/2

xdy d

1 3

- Oe2J e22

ri

Then,

Efkl =
[ _jk] T -I

L tjk]

4-7. Determine J for Prob. 4-1a.

4-8. Determine 1 for Prob. 4-lb.
4--9. Specialize jI for the case where the reference axis is in the X - X2
plane. Note that b = 3 3 3 where 1g331 = 1. When the reference axis is a plane
curve and q0 = 0, we call the member a "planar" member.
4-10. We express the differentiation formulas for ti as
dt
-=
ds
(a)

at

Show that a is, in general, skewsymmetric for an orthogonal system

of unit vectors, i.e., tj tk Sjk Determine a.
(b) Suppose the reference axis is a plane curve but q- 0. The member
is not planar in this case. Determine a.

dy.
dfi

ds

Using (4-22), show that

cos 0 = It 1.

tan 0
1

ds

where a and b are constants. This is the equation for a parabola sym
metrical about x = b/2.
(c) Let 0 be the angle between and 7,.

sin 0
cos 0

dxk

dy
cady

dy

4-6.

(b)

d2 xk

(b) Apply the results of (a) to

X2 =

2
b2

Determine a, [, i for both parametric representations. Take xl as the parameter

for (b). Does y have any geometrical significance?
4-5. Show that

d2f
dx1x2 = f" etc.

df

dx=dx

(a)

x2

a2

(a)

X3 = 0

x2 = b sin y

or

Equation (a) corresponds to taking x, as the parameter for the curve.

(a) Determine the expressions for 7, h, b, a, and K corresponding to this
representation. Note that x t _= y and = xjt + /'(X l ) 2 + 073. Let

4-3.

99

R,