COORDINATE SYSTEM OF THE SURFACE

The parametric representation of the surface

can be given, as follows:
in vector form as r = r (,)
In scalar form as X = X (,); Y = Y (,) ;
Z = Z (,) ;
Therefore these parameters, & can be
called curvilinear coordinates of a given
surface
If the & coordinate lines are mutually
perpendicular at all points on a surface
(i.e., the angles between the tangents to
these lines are equal to 90 ), the curvilinear
coordinates are said to be orthogonal.

Prepared by: Prof. P G Patel

Introduction to Shell

The derivatives of the position vector r with respect to the

curvilinear coordinates & are given by the following:

r, & r, are the tangent vectors at any point of the surface to

the & coordinate lines, respectively (Fig. 11.1).
Indeed, since & are scalar quantities, the directions of
vectors r, & r, coincide with directions of the vector dr.
This vector dr points to the chords joining points M and N for r,
and M and N1 for r, .
Since the latter are assumed to be orthogonal, then r, . r, = 0
Prepared by: Prof. P G Patel

Introduction to Shell

PRINCIPAL DIRECTIONS AND LINES OF CURVATURE

Prepared by: Prof. P G Patel

Introduction to Shell

Consider one normal sections as a plane curve, as shown in

Fig. 11.4. The position of a point M on the curve is
determined by a single arc length coordinate s, measured
from a suitable datum point. Let M and M1 be two
neighboring points dening a short arc of length ds. At M
and M1 the normals OM and OM1 are drawn to the curve
(there is only one normal at each point, since the curve is
smooth) and they are inclined with respect to a suitable
datum direction at angles and +d, respectively.
Assume that the two normals intersect at a point O (the
center of curvature of the arc). The center of curvature is
dened strictly in terms of a limiting process in which ds
0. The length OM (i.e., the distance from the center of
curvature to point M) is called the radius of curvature, , of
the given curve at point M.

Prepared by: Prof. P G Patel

Introduction to Shell

It follows by simple trigonometry from the above description that

= ds/d:
In practice it is more convenient to use the curvature k as a
variable; this is defined as, k = 1/
It follows from the above that there are an infinite number of
possible radii of curvature and curvatures at a point M of the
surface because an infinite number of normal sections may be
drawn through that point. However, there are two orthogonal
normal sections at any point of a surface oriented such that one
radius of curvature is the maximum of all possible, whereas the
second radius of curvature is the minimum of all possible. These
normal sections are called principal normal sections or principal
directions. The curvatures of these sections and the
corresponding radii are referred to as principal curvatures
(denoted by k1 and k2)and principal radii (denoted by R1= 1/k1
and R2 = 1/k2) at a point.

Prepared by: Prof. P G Patel

Introduction to Shell

We also introduce lines of the principal curvature or simply

lines of curvature. A line of curvature is a curve on the
surface with the property that, at any point of the curve, it
has a common tangent with the principal directions. From
the above, it follows that the curvature takes on an extreme
value at that point of the lines of curvature. Hence, at any
point of a smooth surface there is at least one set of
principal directions and two orthogonal lines of curvature.
The principal directions and lines of curvature may coincide
or may not coincide, depending upon the surface geometry.
Figure 11.5a shows the surface of revolution. It is described
by rotating a plane curve (called a meridian) about the axis
of rotation (the Z axis in Fig. 11.5).

Prepared by: Prof. P G Patel

Introduction to Shell

Figure 11.5a shows the surface of

revolution. It is described by
rotating a plane curve (called a
meridian) about the axis of rotation
(the Z axis in Fig. 11.5).
Thus, meridians at any point of the
surface are one of the lines of
curvature and, at the same time, one
of the principal normal sections.
The radius of curvature of the
meridian equals MO1 and it is the
rst principal radius of curvature of
the surface of revolution, R1.
Prepared by: Prof. P G Patel

Introduction to Shell

The second principal normal section of the surface of

revolution is a curve, , which lies in the plane that is
tangent to the conical surface K at point M. It follows from
the above that the second principal section and parallel
circle have a common tangent at point M. Thus, the parallel
circles of the surface of revolution represent the second
lines of curvature but they are not the second principal
normal sections or principal directions. The generator
MO2 of the conical surface K is the second principal radius
of curvature of the surface of revolution, R2. This radius is
numerically equal to the distance along a normal to the
meridian curve from a point of interest of this curve to the
axis of rotation (the Z axis).
Prepared by: Prof. P G Patel

Introduction to Shell

From now on we assume that the coordinate lines & are

the lines of curvature. This system of the coordinate lines
has particularly simple properties, so that the equations of
the theory of shells acquire a relatively simple form in this
system. To be able to use the corresponding formulas of the
general theory of shells, knowledge of the lines of
curvature is required, and the determination of these
curves for a given surface is, in general, a fairly complicated
problem. However, for many of types of shells used in
practice, the geometry of the middle surface is of a simple
nature (e.g., surfaces of revolution, cylindrical surfaces,
etc.), so that the lines of curvature are already known.

Prepared by: Prof. P G Patel

Introduction to Shell

THE FIRST AND SECOND QUADRATIC FORMS OF SURFACES

dr = r, d + r, d
Denoting
a11 = r, . r, ;
a12= r, . r, ;
a22= r, . r,
ds2 = dr. dr = a11(d)2 + 2a12(d. d) + a22(d)2
The expression on the right-hand side of this equation is called the rst
quadratic form of the surface dened by the vector r( and )

Prepared by: Prof. P G Patel

Introduction to Shell

10

Prepared by: Prof. P G Patel

Introduction to Shell

11

SHELLS IN ENGINEERING STRUCTURES

Prepared by: Prof. P G Patel

Introduction to Shell

12

SHELLS IN ENGINEERING STRUCTURES

Prepared by: Prof. P G Patel

Introduction to Shell

13

Advantages of SHELL

Prepared by: Prof. P G Patel

Introduction to Shell

14

FUNDAMENTALS OF SHELLS

Prepared by: Prof. P G Patel

Introduction to Shell

15

FUNDAMENTALS OF SHELLS

Prepared by: Prof. P G Patel

Introduction to Shell

16

FUNDAMENTALS OF SHELLS
Owing to the curvature of the surface, shells are
more complicated than at plates because their
bending cannot, in general, be separated from
their stretching.
On the other hand, a plate may be considered as a
special limiting case of a shell that has no
curvature;
Consequently, shells are sometimes referred to as
curved plates. This is the basis for the adoption of
methods from the theory of plates, discussed in
Part I, into the theory of shells.
Prepared by: Prof. P G Patel

Introduction to Shell

17

CLASSES OF SHELL
Thin shells
A shell is called thin if the maximum value of the ratio h/R
(where R is the radius of curvature of the middle
surface) can be neglected in comparison with unity
For an engineering accuracy, a shell may be regarded as
thin if the following condition is satised:
Max h/R 20

Thick shells
Shells for which this inequality is violated are referred to
as thick shells.

Prepared by: Prof. P G Patel

Introduction to Shell

18

BRIEF OUTLINE OF THE LINEAR SHELL THEORIES

Love was the rst investigator to present a successful
approximation shell theory based on classical linear
elasticity. To simplify the straindisplacement relationships
and, consequently, the constitutive relations, Love applied,
to the shell theory, the Kirchhoff hypotheses developed
originally for the plate bending theory, together with the
small deection and thinness of the shell assumptions.
This set of assumptions is commonly called the Kirchhoff
Love assumptions.

Prepared by: Prof. P G Patel

Introduction to Shell

19

BRIEF OUTLINE OF THE LINEAR SHELL THEORIES

The Love theory of thin elastic shells is also referred to as the
rst-order approximation shell theory. In spite of its
popularity and common character, Loves theory was not
free from some deciencies, including its inconsistent
treatment of small terms, where some were retained and
others were rejected, although they were of the same order.
This meant that, for certain shells, Loves differential
operator matrix on the displacements, in the equations of
equilibrium, became unsymmetric. Obviously, this violated
Bettis theorem of reciprocity. Loves theory also contained
some other inconsistencies. The need for a mathematically
rigorous two-dimensional set of the shell equations
employing the KirchhoffLove assumptions led to different
versions of the rst-order approximation theories.
Prepared by: Prof. P G Patel

Introduction to Shell

20

BRIEF OUTLINE OF THE LINEAR SHELL THEORIES

E. Reissner developed the linear theory of thin shells (also
the rst-order approximation theory) where some
inadequacies of Loves theory were eliminated. He derived
equations of equilibrium, straindisplacement relations,
and stress resultants expressions for thin shells directly
from the three-dimensional theory of elasticity, by
applying the LoveKirchhoff hypotheses and neglecting
small terms of order z/Ri (where Ri = 1; 2 are the radii of
the curvature of the middle surface) compared with unity
in the corresponding expressions.

Prepared by: Prof. P G Patel

Introduction to Shell

21

BRIEF OUTLINE OF THE LINEAR SHELL THEORIES

Sanders also developed the rst-order-approximation shell
theory from the principle of virtual work and by applying
the KirchhoffLove assumptions. Sanders theory of thin
shells has removed successfully the inconsistencies of the
Love theory. A version that retains terms of magnitude
compatible with those retained by E. Reissner and resolves
the inconsistency in the expression for twist was developed
by Koiter .
Timoshenkos theory of thin shells was very close to the Love
theory. General relations and equations were obtained by
applying the KirchhoffLove hypotheses and neglecting
terms z/Ri in comparison with unity.
Naghdi analyzed the accuracy of the LoveKirchhoff theory
of by:
thin
shells. to Shell
Prepared
Prof. elastic
P G Patel Introduction
22

BRIEF OUTLINE OF THE LINEAR SHELL THEORIES

Lurye , Flugge , and Byrne independently developed the
second-order approximation theory of shells. The general
relations and equations of this theory are the direct result
of the application of the Kirchhoff hypotheses together
with the small-deection assumption to the corresponding
equations of the three-dimensional theory of elasticity. The
second-order approximation theory attempts a more
careful discard of terms z/Ri. It retains these terms in
comparison with unity in the straindisplacement relations
and stress resultant equations. Applications of this theory
have generally been restricted to circular cylindrical shells.
In addition, the general relations and equations of this
second-order approximation shell theory are found to be
cumbersome for application.
Prepared by: Prof. P G Patel Introduction to Shell
23

LOADING-CARRYING MECHANISM OF SHELLS

We can say that shells fall into a class of plates as arches relate
to straight beams under the action of transverse loading.
It is known that the efficiency of the arch form lies primarily
in resisting the transverse load with a thrust N, thus
minimizing the shear force V and bending moment M.
It is possible to specify the arch shape and the manner of its
loading in such a way that the arch does not experience
bending at all. In this case, the arch is in the so-called
moment-less state of stress. For example, for a parabolic
arch, bending will not be induced by a vertical load
uniformly distributed over its chord.
Thus, the ability of arches to support certain transverse loads
without bending is the reason for their structural
advantage over straight beams.
Prepared by: Prof. P G Patel

Introduction to Shell

24

LOADING-CARRYING MECHANISM OF SHELLS

A shell mainly balances an applied transverse load, much like an arch, by
means of tensile and compressive stresses, referred to as the membrane
or direct stresses. These stresses are uniformly distributed over the
shell thickness. Such a state of stress is called the momentless or
membrane state of stress. Although the shear force and bending and
twisting moments are still present in the general case of loading, the
efficiency of the shell form rests with the presence of the membrane
stresses, as the primary means of resistance with the bending stress
resultants and couples are minimized.
Thus, shells, like arches over beams, possess an analogous advantage over
plates; however, with the following essential difference
while an arch of a given form will support only one completely
determined load without bending, a shell of a given shape has,
provided its edges are suitably supported, as a rule, the same
property for a wide range of loads which satisfy only very general
requirements
Prepared by: Prof. P G Patel

Introduction to Shell

25

LOADING-CARRYING MECHANISM OF SHELLS

The membrane stress condition is an ideal state at which
a designer should aim. It should be noted that
structural materials are generally far more efcient in
an extensional rather in a exural mode because:
1. Strength properties of all materials can be used
completely in tension (or compression), since all bers
over the cross section are equally strained and loadcarrying capacity may simultaneously reach the limit
for the whole section of the component.
2. The membrane stresses are always less than the
corresponding bending stresses for thin shells under
the same loading conditions.
Prepared by: Prof. P G Patel

Introduction to Shell

26

LOADING-CARRYING MECHANISM OF SHELLS

The highest efficiency of a shell, as a structural member, is
associated with its curvature and thinness. Owing to the
shell curvature, the projections of the direct forces on the
normal to the middle surface develop an analog of an
elastic foundation under the shell. So, it can be said that a
shell resists an applied transverse loading as a at plate
resting on an elastic foundation. This phenomenon can
explain an essential increase in strength and stiffness of a
shell compared with a plate.
Thus, as a result of the curvature of the surface, a shell
acquires a spatial stiffness that gives to it a larger loadcarrying capacity and develops the direct stresses.
Owing to its thinness, a shell may balance an applied
transverse loading at the expense of the membrane stresses
mainly, with bending actions minimized.
Prepared by: Prof. P G Patel

Introduction to Shell

27

LOADING-CARRYING MECHANISM OF SHELLS

Note that the pure bending conditions have no advantages
and should be avoided because shells, in view of their small
thickness, possess a low strength for this deformation.
However, sometimes bending conditions cannot be avoided.
It turns out that strong severe bending conditions are
localized only in a small domain near some discontinuities
in loading and geometrical conditions, as well as near
supports, etc. As we move away from such a disturbance
zone, the bending stresses will diminish rapidly and a
considerable part of the shell will be in the moment less
stress condition.
Therewith, the thinner the shell the faster this decrease of
bending stresses.
Shell thinness demonstrates the high efficiency of shells. It is
associated with the shells low weight and simultaneously
its high strength.
Prepared by: Prof. P G Patel

Introduction to Shell

28

LOADING-CARRYING MECHANISM OF SHELLS

However, a shells thinness is, at the same time, a weak point
because all the advantages mentioned earlier hold for a tensile
state of stress. In this case, a shell material is stretched and its
strength properties are used completely.
On the other hand, a thinness of shells manifests itself in
compression. External forces, as before, are effectively
transformed in the constant membrane stresses over the shell
thickness.
However, the trouble is that the level of the critical stresses at
buckling be sufficiently low. This level is just determined by the
shell thickness. The thinner the shell, the lower is the level of the
critical stresses. The latter can be many factors smaller than the
proportional limit of the shell material.
In this case, the efficiency of thin shells can be reduced
considerably. To avoid the possibility of buckling, a shell
structure should be designed in such a way that a dominant part
of the structure is in tension.
Prepared by: Prof. P G Patel Introduction to Shell
29