Some New Thoughts On The Buckling of Thin Cylindrical Shells

XXI ICTAM, 15-21 August 2004, Warsaw, Poland
SOME NEW THOUGHTS ON THE BUCKLING OF THIN CYLINDRICAL SHELLS

*
Parthasarathi Mandal*
Manchester Centre for Civil & Construction Engineering,

UMIST, P.O. Box 88, Manchester, M60 1QD, U.K
Summary The classical theory of buckling of axially loaded thin cylindrical shells predicts that the buckling stress is
directly proportional to the ratio of thickness to diameter (t/R), other things being equal. But the empirical data show that the
buckling stress is proportional to (t/R)1.5, other things being equal. Also there is wide scatter in the buckling stress data. The
imperfection-sensitive, non-linear behaviour is thought to be the cause of the above. In this article, it is argued that the
boundary conditions, from the statical determinacy point of view could explain the shell-buckling paradoxes.
INTRODUCTION
The classical theory, developed in the beginning of the twentieth century, predicts that cylindrical shells will buckle
under uniform axial compressive stress cr , given by
cr =
3 1
t
t
R 0.6 E R

(1)
where E is the Youngs modulus of elasticity and the Poissons ratio of the isotropic material, t is the thickness and R
is the radius of the shell.
However, in the 1930s it was discovered that the experimental buckling loads of cylindrical shells are often small
compared to the predictions of the classical theory. Since then there have been extensive research to explain: (i) these
lower experimental buckling loads; (ii) the large scatter of the experimental data; and (iii) the catastrophic nature of the
failure.
The concepts of non-linearity and imperfection-sensitivity are widely accepted as explanations for the paradox
associated with shell-buckling. However, extensive studies based on precise measurements of geometrical imperfection
in physical shells, manufactured in various ways, have had limited success in predicting the buckling load. Therefore, it
is reasonable to claim that there must be some crucial factors involved in the mechanisms of shell buckling which are
missing from existing analytical studies. This article discusses some interesting features of experiments on the selfweight buckling of small-scale, thin, silicone rubber shells.
SELF-WEIGHT BUCKLING EXPERIMENTS
Consider a large-scale, open-topped cylindrical shell. There is a
tendency for the structure to collapse under its own weight; and
collapse may be precipitated by a disturbance such as an
earthquake or wind loading. In order to understand this
phenomenon, Calladine and Barber [1] conducted some simple
experiments by making long, open-topped, fixed-base thin
cylindrical shells in silicone rubber and cutting them down, little
by little, until they just stood upright on a flat table without
buckling under their own weight. These heights were regarded
as the buckling heights (Lcr) under self-weight. All of the
cylinders had the same outside diameter (nominally 172 mm)
and ranged in thickness from 0.23 mm to 1.12 mm. The results
are
shown in Fig. 1, as a log-log plot of dimensionless critical
Fig. 1 Log-log plot of experimental data of thin cylindrical shells.
The data referred as previous experiments have been taken from heights against R/t. The ordinate represents the self-weight stress
Brush & Almroth[4]. The heavy best-fit line has a slope of 1.5; at the base if the shell (=Lcr g, where g is the unit weight of
and broken lines at 1 and 2 standard deviations from the mean are
the silicone rubber material), normalized with respect to E.
also shown.
More recently, some more experiments have been conducted
with a larger mould, of diameter 241 mm: see Refs.[2] and [3]. These results are also plotted in Fig. 1, and they agree
very well with [1]. There are two distinguishing features of these results. First, in general the buckling clearly does not
take place according to the simple classical theory (Eqn 1). Second, there is very little scatter in the data, compared to
almost all buckling tests on thin cylinders that have been reported in the literature. To assess the situation, the data of
experimental buckling strength of cylindrical shells under axial compression that was collected and presented by Brush
and Almroth [4], are re-plotted here in Fig 1, on log-log scales of cr /E and R/t. It is immediately clear that the
experimental data lie on a band which is steeper than the classical line; and indeed the best-fitting line through the
data points has a slope of 1.5, to two significant figures. Specifically, we find
1.5
t
mean = 5 E
R
(2)
XXI ICTAM, 15-21 August 2004, Warsaw, Poland
Non-linear finite element analyses

In order to understand the paradoxical self-weight buckling, non-linear finite element analyses of several experimental
shells were conducted using ABAQUS [5]. The result of a typical computation is shown in Fig. 2. The figure shows that
the shell does not buckle until a load of 1.84g has been reached: this is near to the classical buckling load level for the
shell as revealed by a linear-eigenvalue buckling analysis. The load then falls to around 1.3g, with very little radial
displacement. After this, a dimple starts to appear at Node B, with significant inward displacement, and Node A at the
top starts to move outwards. The shell above the dimple deforms inextensionally henceforth, leading to a post-buckling
plateau corresponding to the experimental critical heights. The plateau remains virtually similar if the amplitudes of
the initial imperfection were increased by a factor of four.
xA
t
L
z
oB
q
2R
Imperfection used in the ABAQUS computation:

0.25 t ( z / L ) cos ( 7 ) e 6.9
Fig. 2 Plot of numerical gravity load factor against radial deflection at two points of an open-topped, fixed base shell (R/t 210) from the selfweight buckling experiments. An ordinate of value 1.0 corresponds to the physical shell under gravity at its measured critical height. The geometrical
imperfection was deliberately included, primarily as a requirement for the non-linear analysis to be performed smoothly; and the shape of the
imperfection was chosen so as to encourage buckling to begin at a desired location.
DISCUSSION AND CONCLUSIONS

The self-weight buckling experiments along with the non-linear finite-element analyses highlight that the critical stress
of cylindrical shells under axial compression is proportional to (t/R)1.5, compared to (t/R)1.0 as in classical theory. This
fits well with the large amount of previous experimental data. However, the experiments on self-weight buckling showed
negligible scatter compared to the earlier experiments in the literature. We suggest that the explanation lies in the fact
that the two sets of shells have a different status in terms of static determinacy: thus the open-topped shells in the selfweight buckling experiments are statically determinate, whereas almost all the shells used in standard buckling tests in
the past had closed ends, hence they are statically indeterminate from the membrane stresses point-of-view. The static
indeterminacy in a system may be held responsible for any locked-in stresses due to lack-of-fit resulting from the
manufacturing process. Such unaccounted, random locked-in-stresses may cause a shell to buckle at a premature load,
after reaching the critical stress in a localised area; or in general they may produce a big statistical variation in buckling
loads of notionally similar shells.
To verify the above hypothesis of static determinacy, another set of non-linear finite element analyses were conducted
by Vossos [6]. Three fixed-base cylindrical shells were considered. The boundary conditions at the top were: (i) open,
(ii) constrained to be both circular and plane, and (iii) attached to a ring beam. The self-weight buckling analyses of
these three shells were carried out in two steps. In the first step a preload was applied to mimic the residual stresses in
practice, and in the second step gravity load was applied by modified RIKS method in ABAQUS [5]. The preload was
applied in one of the following ways: (i) a concentrated load in the radial inward direction, applied near Node B (see
Fig. 2); (ii) a temperature increase at a small region near Node B; and (iii) uplift of the base at a small portion of the
perimeter of the shell. The comparative buckling load factors are shown in Table 1. The buckling load, of course,
depends on the amount of preload and their positions. Hence the values given in Table 1 should be read column-wise.
They do not have any meaning along a row.
It is clear that the statically determinate shells perform better in buckling than their statically indeterminate counterparts.
For example, an increase in temperature in a small region of a statically indeterminate shell will cause higher initial
stresses. As a result, when the shells are tested for buckling, the statically indeterminate shells will carry lower loads.
Shell 1
Shell 2
Shell 3
Boundary Conditions
Top
Base
Open
Fixed
Circular & Plane
Fixed
Ring beam
Fixed
Statical
Determinacy
Determinate
Indeterminate
Indeterminate
Table 1
Radial load
1.00
0.90
0.94
Buckling load factor

Temperature
1.00
0.66
0.57
Uplift
1.00
0.41
0.49
References
[1] Calladine C.R., Barber J.N.: Simple experiments on self-weight buckling of open cylindrical shells. J. App. Mech, Trans ASME: 1150-1151, 1970.
[2] Mandal P., Calladine C.R.: Buckling of thin cylindrical shells under axial compression. Int. J. Solids and Structures: 37:4509-4525, 2000.
[3] Zhu E., Mandal P., Calladine C.R.: Buckling of thin cylindrical shells: an attempt to resolve a paradox. Int. J. Mech. Sci.: 44:1583-1601, 2002.
[4] Brush D.O., Almroth B.O.: Buckling of bars, plates and shells. McGraw-Hill, NY 1975.
[5] Hibbit, Karlsson, Sorensen Inc: ABAQUS, Version 5.8, web site http://www.hks.com/home.html
[6] Vossos, P.: Buckling of thin cylindrical shells under axial compression. MSc Thesis, Civil Engineering, UMIST, UK, 2003.

Some New Thoughts On The Buckling of Thin Cylindrical Shells

Uploaded by

Copyright:

Available Formats

Some New Thoughts On The Buckling of Thin Cylindrical Shells

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Some New Thoughts On The Buckling of Thin Cylindrical Shells

Uploaded by

Copyright:

Available Formats

XXI ICTAM, 15-21 August 2004, Warsaw, Poland

SOME NEW THOUGHTS ON THE BUCKLING OF THIN CYLINDRICAL SHELLS

Manchester Centre for Civil & Construction Engineering,

XXI ICTAM, 15-21 August 2004, Warsaw, Poland

Non-linear finite element analyses

Imperfection used in the ABAQUS computation:

DISCUSSION AND CONCLUSIONS

Buckling load factor

You might also like