Journal of Constructional Steel Research 145 (2018) 289–299

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

Buckling behavior of cylindrical steel tanks with concavity of vertical

weld line imperfection
Mehdi Rastgar a,⁎, Hossein Showkati b
a
Department of Civil Engineering, Khoy Branch, Islamic Azad University, Khoy, Iran
b
Department of Civil Engineering, Urmia University, Urmia, Iran

a r t i c l e i n f o a b s t r a c t

Article history: Shell structures are built using a number of welded curved panel parts. Hence, some geometrical imperfections
Received 29 December 2016 emerge. These imperfections have a direct impact on structural behavior of shells during the external compres-
Received in revised form 6 February 2018 sive loading. In this research, a field study was accomplished on the implementation of the storage tanks in a
Accepted 22 February 2018
refinery site and then, the resulted imperfections were identified and categorized. The survey of imperfections
Available online xxxx
revealed that the imperfection in form of concavity of vertical weld line is the most prevalent type of imperfection
Keywords:
seen in the steel tanks. This imperfection experimentally modeled and the buckling behavior of these tanks
Field study was evaluated under uniform external pressure. Comparing obtained results of estimation, ASME code and
Cylindrical tank experimental research represented a considerable difference in the amount of buckling load. Results show that
Buckling behavior the imperfections due to concavity of vertical weld line are very important in buckling of the tanks under uniform
Vertical weld line imperfection external pressure. This imperfection decreases initial, full and post buckling capacity of the tanks under uniform
External pressure external pressure, significantly. Findings of this research show that for design of steel tanks under uniform exter-
nal pressure load, 65% of the buckling load obtained from the ASME Code should be used.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction and failure of these tanks is investigated under uniform external pres-
sure loading.
Shell structures have curved initial shapes with the thickness less
than other dimensions. In some states, radius-to-thickness ratio reaches
2. Stability in shell structures
3000. The force field generated in these structures includes membrane
and bending forces, which can change depending on shell thickness.
Evaluating the stability problems in thin-walled shells is important
Steel tanks designed in cylindrical form are commonly used shell struc-
for two reasons. First, the ratio of thickness to other dimensions is
tures in industrial facilities. In geometrical terms, these tanks have
very low in these structures, which highlights the instability problem.
a very small thickness compared to their other dimensions and are
Second, shell structures are exposed to compressive stresses and forces;
categorized as thin-walled structures for buckling failure caused by
however, shell buckling is caused by low thickness. One of the main
the influence of uniform external pressure loading on the tank wall [1].
properties of shells is having much higher membrane stiffness than
The buckling of tanks under uniform external pressure is usually
bending stiffness. Accordingly, a shell can absorb a large amount of
caused by operational problems during the discharge of the liquid
membrane energy without experiencing a great deformation, while
contents in such a way that partial vacuum is produced. These events
large deformations and rotation in the cross-section need to absorb
are usually classified as accidents, and occur in individual tanks in a
this energy through bending deformations [1].
tank farm, rather than affecting many tanks in the same event, such as
those produced by a natural disaster.
The objective of this paper is identification and categorization of 2.1. Buckling of shell structures
created imperfections in implementation of the storage tanks via a
field study. By identifying the prevalent imperfection, buckling behavior Buckling is considered as a nonlinear phenomenon, in which the
structure cannot take further load with the same geometry and changes
its shape in order to find alternative equilibrium configurations. Shell
⁎ Corresponding author.
buckling occurs as the structure response to the membrane forces.
E-mail addresses: m.rastgar@urmia.ac.ir (M. Rastgar), h.showkati@urmia.ac.ir Membrane forces act along the component axis and tangent to the mid-
(H. Showkati). dle surface of the shell and the buckling occurs when the structure

https://doi.org/10.1016/j.jcsr.2018.02.028
0143-974X/© 2017 Elsevier Ltd. All rights reserved.
290 M. Rastgar, H. Showkati / Journal of Constructional Steel Research 145 (2018) 289–299

converts membrane strain energy into bending strain energy without
any change in the applied external load [1].
There have been a number of studies addressing buckling behavior
and instability, see e.g. [2]. and references therein. Below we mention
several contributions relying on experimental approach.
Influence of primary boundary condition on the buckling of shallow
cylindrical shells was studied by Showkati and Ansourian experimen-
tally [3]. Wang and Koizumi investigated the buckling of cylindrical
shells with longitudinal joints under external pressure [4]. Buckling of
cylindrical shells with stepwise variable wall thickness under uniform
external pressure was considered by Chen et al. [5]. Aghajari et al.
conducted an experimental study on the buckling of thin cylindrical
shells with two-stepwise variable thickness under external pressure
[6]. Experimental and numerical investigation of composite conical
shells stability subjected to dynamic loading was investigated by Jalili
et al. [7]. Ghazijahani et al. studied longitudinally stiffened corrugated
cylindrical shells under uniform external pressure [8].
2.2. Geometric imperfections on shells buckling
Geometric imperfections include all the deviations in the form of a
structural component compared with its ideal geometric composition.
In construction of shell structures, due to the large dimensions, curved
plates or panels can be used. The seam between various surfaces of
the main source of deviation is in the real form, these deviations or
imperfections can be generated as a result of welding or appropriate
incompatibility of the plates with larger dimensions than other plates.
In contrast to various structures, the buckling strength of shells with
no imperfections is significantly different from the buckling strength
of the same shell with imperfections. In some states, geometric imper-
fections may strengthen the structure and increase its capacity. This
feature puts shells among the structures which are called sensitive to
imperfection. In Fig. 1, the buckling behavior of columns, flat plates
and cylindrical shells are schematically shown.
In these curves, black lines show the system with no geometric
imperfections or perfect system, while red lines represent the corre-
sponding behavior of the imperfect system. As can be seen, the flat
plates and column elements are not sensitive to imperfection, while
cylinders which are the examples of thin-walled structures are very
sensitive to imperfections [9].
There have been a number of studies addressing the influence of
imperfections on the buckling behavior of shell structures. Below we
mention several contributions relying on experimental approach.
Calladine studied the causes of imperfection-sensitivity in the
buckling of thin shells [10]. Teng et al. investigated the geometric imper-
fections in full-scale welded steel silos [11]. In other work, buckling
behavior of large steel cylinders with patterned welds was considered
by Hubner et al. [12]. Influence of imperfection on the buckling of thin
cylindrical shell under uniform external pressure was studied experi-
mentally by Lo Frano and Forasassi [13]. Maali et al. investigated the
Buckling behavior of conical shells under weld-induced imperfections
experimentally [14]. Yang et al. studied the buckling of cylindrical shells
with general axisymmetric thickness imperfections under external
pressure [15]. Fatemi et al. conducted experiments on imperfect
cylindrical shells under uniform external pressure and observed that
geometric imperfections have a more considerable impact on the
behavior of shells [16].
In other work, inelastic stability of liners of cylindrical conduits with
local imperfection under external pressure was studied by Khaled
El-Sawy [17]. Ghazijahani et al. conducted experiments on dented
cylindrical shells under peripheral pressure [18]. Thompson presented
advances in shell buckling. He studied on the buckling of axially
compressed cylindrical shells with arbitrary thickness imperfections
theoretically and experimentally [19]. Cao et al. studied buckling of
cylindrical shells with arbitrary thickness imperfections under axial
compression analytically [20]. Lee et al. investigated the geometric
role of precisely engineered imperfections on the critical buckling load
of spherical elastic shells. Their investigation combined precision exper-
iments, finite element modeling and numerical solutions of a reduced
shell theory, all of which were found to be in excellent quantitative
agreement [21].
Evkin et al. investigated the buckling of a spherical shell under
external pressure and inward concentrated load [22]. In other work,
Hutchinson and Thompson studied nonlinear buckling behavior of
spherical shells subject to external pressure. They found that the nonlin-
ear axisymmetric post-buckling behavior of perfect thin spherical shells
and their asymmetric bifurcations are characterized providing results
for a structure/loading combination with an exceptionally nonlinear
buckling response [23].
Due to the mentioned studies about imperfections in shell struc-
tures, field study and vertical weld line imperfection has not been
studied. In this research this imperfection has been introduced and
buckling behavior of cylindrical steel tanks has been investigated
under uniform external pressure.
3. Analytical equations for buckling of cylindrical shells
Donnell in 1933 obtained an equation for the buckling of cylindrical
shells as follow [24]:
!
Et ∂ w
4 2
∂ w
2
∂ w
2
∂ w
D∇8 w þ þ ∇4 Nx þ 2Nxy þ Nθ 2 ¼ 0 ð1Þ
R2 ∂x4 ∂x
2 ∂x ∂θ ∂θ
This equation can be used for various loading conditions. Let us
take the simple case of applied lateral pressure, where Nx = Nxy = 0
and Nθ = Pr − σcrt. Eq. (1) then becomes [24]:

4 2
!
8 Et ∂ w 4 ∂ w
D∇ w þ þ∇ Nθ ¼0 ð2Þ
R2 ∂x4 ∂θ
2

An expression for the deflection that satisfies the simply supported

boundary condition of a cylinder can be expressed as:

mπx nπy
w ¼ wmn sin sin ð3Þ
L πR

Substituting this expression into Eq. (2) gives the nontrivial

2 2
π KE
2 2 ðm þβ 2
Þ 12Z2
solution σ cr ¼ 12ð1−υ 2 Þ ðt=LÞ where K ¼
β2
þ 2, β ¼ πR
nL
,
π4 β2 ð1þβ2 =m2 Þ
Fig. 1. Load-axial displacement graph of columns, flat plates and cylindrical shells in pffiffiffiffiffiffiffiffiffiffiffiffiffi
L2
perfect and imperfect states [9]. Z ¼ Rt 1−υ 2
 M. Rastgar, H. Showkati / Journal of Constructional Steel Research 145 (2018) 289–299 291

The number of circumferential waves of cylindrical shell formed in

the body is as follows: [3]
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi sffiffiffiffiffiffiffiffiffiffiffiffi
u 2 pffiffiffiffiffiffiffiffiffiffiffiffiffi2 rffiffiffiffi
u 6π 1−ν R R
u
n¼t
4   ≈ 2:74 ð5Þ
 L 2 t L t
R R

3
So that D ¼ 12ð1−υ
Et
2 Þ, C ¼ ð1−υ2 Þ, m ¼
Et mπR
L
Based on the ASME, Boiler and Pressure Vessel Code Section VIII,
Rules for Construction for Pressure Vessels, the allowable external pres-
sure is calculated from: [24,26]

2:42E ðt=2RÞ2:5
P cr ¼ 3=4
h i ð6Þ
ð1−υ2 Þ L=2R−0:45ðt=2RÞ1=2
Fig. 2. A plot of K [24]

2 2
For structures with large (R/t) ratios, Eq. (6) can be simplified to:
Considering m = 1 the minimum value of K becomes: K ¼ ð1þβ
β2
Þ
þ [24,26]
12Z2
π4 β 2 ð1þβ 2 Þ
2
2:42E ðt=2RÞ2:5
P cr ¼ ð7Þ
A plot of K defined by above equations is shown in Fig. 2. ð1−υ2 Þ
3=4 ðL=2RÞ

3.1. Short cylinders

4. Field survey
For short cylinders, the curvature parameter Z can be set to zero and
then K = (β + 1/β)2 In order to store oil products in a refinery site, the construction of
numerous tanks has begun, which are being implemented as steel cylin-
3.2. Intermediate-length cylinders drical tanks. These tanks have the height of 12 m, diameter of 23 m,
and sheet thickness of 18 mm and 8 of them are under construction.
As the ratio L/R increases, the quantity (β + 1) can be approximated For the construction of these tanks, the steel sheets with the dimensions
2
by β. Hence K ¼ β2 þ π12Z
4 β6 of 6 × 1.5 m and thickness of 18 mm are used. The sheets are entered
into the rolling device and the curvature operations are performed
3.3. Long cylinders during several steps to reach the radius of the tank. Then, they are trans-
ferred to the relevant location by the special vehicles for transporting
For long cylinders, the buckling mode is similar to that of a circular the rolled sheets and their installation operations are carried out using
ring and is elliptic in shape. Hence, n = 2 and β = 2L/πR and K reduces the crane and human force. From this step on, numerous imperfections
to K = β2. caused by various implementation factors are made; some of them are
Considering simply supported conditions for the ends of the cylin- partly modified and some others are remained in the tank even after
drical shell and m = 1, and application of uniform external pressure to the implementation is complete.
the roof of cylindrical tanks that causes a uniform axial force to the Installation of sheets is shown in Fig. 3.
body of tanks, buckling load of these tanks under the simultaneous According to this Figure, the rolled sheets are put in their place
effect of uniform external pressure and axial load is obtained by the fol- by the crane and then temporarily welded next to each other (tack
lowing equation: [25] weld). When the first stage is completed, the next layers are imple-
mented similar to the first layer. The rolling, transportation, and instal-
 4  D   
lation of these sheets cause various imperfections in the implemented
m2 þ n2 þ m4 1−υ2 C layer. As shown in Fig. 3, the edges of the circumferential sheets do
Pcr ¼
1 R2 ð4Þ not match each other; these edges should get close to each other by a
 2  
R
m 2 þ n2 n2 þ 0:5m2 chain and a crane until the imperfection is modified. Basically, some of

Fig. 3. Circumferential sheets installation process in the body of tank and generation of different imperfections.
292 M. Rastgar, H. Showkati / Journal of Constructional Steel Research 145 (2018) 289–299

the created imperfections are removed from the body of the tanks and
some others are left permanently.
In this research, after attending the tank implementation location
and field monitoring of the installation of circumferential sheets, vari-
ous types of imperfections are observed which are classified as below:

a) Non-compliance to the edges of consecutive horizontal rows that

occurred outside or inside
b) Vertical seam not executed properly and sharp towards inside or
outside
c) Non parallel vertical edges
d) Nonaligned horizontal edges
e) Cause by unsuitable rolling
f) Vertical different layers in height
Fig. 5. Frequency distribution graph of the imperfection types in all tanks.
g) Effect of welding temperature which usually appear in circumferen-
tial form, and
h) Wind Effects causing troughs or bulge, in the peripheral of the last
layer of tank.
Accordingly, it can be seen that, among all the introduced imperfec-
These imperfections are shown in Fig. 4. tions, imperfection types b) and f) occur more frequently than others,
By continuing the field study on the implementation of these tanks respectively.
in a 6-month period and statistical inference from the imperfections In imperfection type b), the edges of the circumferential sheets do
introduced in all the 8 tanks being constructed in the refinery site, the not match each other. This Imperfection named that vertical weld line
frequency distribution graph of the imperfections is drawn, as shown imperfection includes concavity of vertical weld line towards inside
in Fig. 5. In this Figure, the horizontal axis shows the type of imperfec- at the location of two adjacent sheets. This imperfection is shown in
tion and the vertical axis is the frequency percentage of the correspond- Fig. 6. In this research buckling behavior of steel tanks with vertical
ing imperfection. weld line imperfection is studied.

a) b) c)

d) e) f)

g) h)
Fig. 4. Types of observed imperfections in the field study of tanks.
 M. Rastgar, H. Showkati / Journal of Constructional Steel Research 145 (2018) 289–299
Fig. 6. Imperfection type b) (vertical weld line imperfection)
5. Experimental program
5.1. Geometric specifications of experimental specimens
According to Fig. 5, imperfection type b) named as the vertical weld
line imperfection occurs more frequently than others. For the experi-
mental evaluation of the impact of this imperfection on the buckling
behavior of tanks, three experimental specimens are made with the
following characteristics:
The first specimen of Spec1 has continuous circumferential plate in
an integrated form without intended imperfection. Due to the limited
length of the used plates which is 3 m and providing for the 3.6 m
perimeter of this specimen, there are two vertical welding lines in the
body of tank. CO2 welding is used in this specimen and others. The
weld line is perfectly circular and has no dents or bumps.
Fig. 7. Experimental specimens.
Fig. 8. Measurement of intentional imperfection value.
293
The second specimen of Spec2 has entirely an intentional imperfec-
tion in terms of a vertical weld line with 15 mm in depth within tank
height. Similar to the previous specimen, this specimen has two vertical
weld lines on the body of the tank that the imperfection is located on
one of these weld lines and the tank is fully circular in another weld
line and has no dents or bumps.
The third specimen of Spec3 that is in an integrated form, has two
intentional imperfections of the type of vertical weld line with 15 mm
over the tank height and they are located on one of the tank's diameters
with 180 degrees angle between them. There are two vertical weld lines
in the body of this tank and the imperfections exist along these welded
lines. All three specimens have the diameter of 1.15 m, height of 0.6 m,
conical roof with the height of 0.2 m, body plate thickness of 1 mm and
floor and roof plate thickness of 2 mm. These specimens are shown in
Fig. 7. (See Fig. 8.)
In order to prevent the floor and roof buckling of these experimental
tanks influenced by external pressure, strengthening annular straps with
the diameter of 0.3, 0.6 and 0.9 m and section of 20 × 5 mm are used.
These straps could increase the bending stiffness in the roof and floor
of the tanks; so, under the uniform external pressure, these two parts
maintain their initial situation and hence, do not buckle; only the body
of the tanks starts to move and buckles. The size of the experimental
specimens is selected in accordance existing laboratory facilities. Accord-
ing to provided dimensions, it is observed that experimental-scale
models are on the scale of 1:20 compared to real tanks. Welding on
experimental models is quite different from real tanks. These specimens
are welded using CO2. The shielding gas reduces generated heat and
1 mm plates are subsequently welded without any significant imper-
fections. A great care is taken in construction of experimental speci-
mens and it is tried to produce only considered imperfection.
Despite high accuracy in producing the experimental specimens,
various local imperfections emerge in these specimens; the major one
294 M. Rastgar, H. Showkati / Journal of Constructional Steel Research 145 (2018) 289–299
(A)
(B)
Fig. 9. (A) The three samples (B) tensile test (C) stress-strain diagram of material.
which includes imperfections caused by welding. This fact is due to low
thickness of sheets used as troughs in the body, troughs in the floor and
roof of the specimens caused by transportation from the workshop to
laboratory, as well as movement at different places in the laboratory.
The local imperfections may also affect buckling behavior of tanks. As
previously noted, the experimental specimens have intentional imper-
fections in terms of a dent on vertical weld line on tank's body with
15 mm size.
In Fig. 6, the measurement of intentional imperfection value is
shown in specimens. A metal ruler is coordinated with circular body
of the tank to measure considered imperfection. In imperfection
location, the distance between the ruler and vertical weld line equals
15 mm, which is the same as considered imperfection. An intentional
imperfection in specimen 2 and two intentional imperfections in
specimen 3 are produced. But there is no intentional imperfection in
specimen 1 and vertical weld line is completely adapted to the ruler.
Therefore, this specimen does not have considered imperfection and is
indicated as the specimen with no imperfection.
By using the approximate Eqs. (4) and (6), the buckling load caused
by uniform external pressure along with axial load is obtained which is
equal to 23.5 and 23.3 kPa, respectively. Also number of circumferential
waves formed in the body of the tank is 13 waves based on Eq. (5),
In this study, the amount of the obtained buckling load is used for the
evaluation and comparison of the results taken from experimental
specimens.
5.2. Mechanical properties of the specimens
Mechanical properties of the specimens have been achieved from
tensile test. These properties including elasticity Modulus, yield stress,
failure phase stress and equivalent strain of this point. Tensile test was
carried out for 3 samples of the used sheets in experimental specimens.
The test temperature is the same as the temperature of the laboratory
and approximately 28 degrees Celsius. The material is carbon steel.
Table 1
Mechanical properties of the specimens.
Average modulus Average Average Average
of elasticity yield stress failure stress failure strain
200 GPa 194 MPa 325 MPa 30.24%
Dimensions of tensile samples are selected based on recommended
dimensions in the ASTM E8/E8M−15a code. In Fig. 9 the samples,
tensile test and stress-strain diagram of material are shown. Mechanical
properties of the specimens are given in Table 1. These values are the
average results of three samples.
6. Testing of specimens and evaluation of results
6.1. Uniform external pressure loading
In order to evaluate the buckling behavior of experimental speci-
mens influenced by uniform external pressure, the suction device is
utilized. It is an electro-pump device that discharges the air inside the
tank with the constant flow of 40 m3/h to the outside. So, by discharging
the air inside the tank, the atmospheric pressure is uniformly imported
into the external surfaces of this tank and loading is applied as uniform
external pressure to the tank. Since air discharge flow of the tank is too
high, it is necessary to embed a control valve to adjust the discharge
flow and thus control the external pressure.
In order to measure the internal pressure in the tank, a pressure
gauge is used. This device shows the moment-to-moment pressure
inside the tank, which is decreased due to the discharge by the suction
device. So, external pressure is obtained by subtracting atmospheric
pressure and the pressure inside the tank. There are three holes in the
tank floor: the first hole is connected to the suction device, through
which discharge operation is done (applying uniform external pressure).
The second hole is connected to the discharge valve and controls the flow
of tank discharge (loading control) and the third one is connected to the
pressure gauge to measure the internal pressure (loading measurement).
Fig. 10 depicts the laboratory equipment including the suction pump
for loading application, loading control valve, loading measurement
device, deformation measurement devices, location of test specimens,
the data logger, and the computer.
6.2. Measurement devices of deformation
In order to take the experimental specimens deformation, measure-
ment devices are installed in different parts of these specimens in
the middle of the height radially and at the top of the conical roof of
the specimens. These devices include LVDT, strain gauge and pressure
sensor. The LVDT is an acronym used for Linear Variable Differential
Transformer. It is a transformer which is used for measuring linear
 M. Rastgar, H. Showkati / Journal of Constructional Steel Research 145 (2018) 289–299
Fig. 10. A view of the laboratory equipment's and experimental specimen
displacement. LVDTs take the directional displacements vertically, hor-
izontally, or radially in the locations they are installed and then, send
them to the data logger. The strain gauges are electronic circuits that
are attached using a special adhesive in the desired location and they
measure the strain in their installed direction and send it to the com-
puter. In this research, in all the tests, five LVDTs and two horizontal
and vertical strain gauges are used to record the deformation of differ-
ent parts of the experimental specimens.
In Fig. 11, a sample of these measurement devices which is installed
on the experimental specimens is represented and Fig. 12 shows
schematic view of all installed sensors with location of imperfection
and failure.
6.3. Test implementation
Air discharge from Spec1 starts by turning on the suction pump de-
vice and the difference between the internal and external pressure of
the tank is gradually increased. As a result, external pressure is gradually
applied to the tank and the initial buckling of the tank body begins.
By increasing the external pressure, the number of the waves formed
in the tank body is also increased and the tank moves towards full
buckling. Then, after making 12 waves in the tank circumference, the
buckling of the body becomes complete and, finally, at the external
pressure of 21.6 kPa, the tank failure suddenly happens and the tank
becomes unstable. Fig. 13(A) shows full buckling and Fig. 13(B) demon-
strates instability and failure of the experimental Spec1.
In the Spec2, similar to Spec1, loading continues gradually and in a
controlled manner to the failure stage and full instability. After making
10 waves in the tank circumference, the buckling of the body becomes
Fig. 11. Installation of measurement devices on the experimental specimen.
295
Fig. 12. Schematic view of all installed sensors with location of imperfection and failure
complete and, finally at the external pressure of 16.5 kPa, failure
suddenly happens and the tank becomes unstable. Fig. 14(A) and (B)
shows full buckling and failure of the experimental Spec2 respectively.
The Spec 3, similar to previous specimens, continues gradually and
in a controlled manner to the failure stage and full instability. In this
specimen, by increasing the external pressure 10 waves are formed in
the tank circumference and the buckling of the body becomes complete
and, finally, at the external pressure of 19.5 kPa, failure suddenly hap-
pens and the tank becomes unstable. Fig. 15(A) shows full buckling
and Fig. 15(B) demonstrates instability and failure of the experimental
Spec3. According to these Figures, full buckling and unstable form of
all three specimens is similar and up to full buckling, regular circumfer-
ential waves are seen in the tanks body but unstable location of Spec2
and 3 is different from Spec1 and failure of these specimens occurs at
the location of imperfection.
7. Evaluation of experimental results
By taking the data from the measurement devices installed at differ-
ent parts of these tanks that shown in Fig. 12, radial displacement
graphs of these points are drawn relative to the applied uniform exter-
nal pressure. In Fig. 16 radial displacement diagrams of Spec1 at the
locations of installed LVDTs are shown.
Considering these diagrams it can be seen that radial displacement
at the location of LVDT3 is higher with respect to other portions.
Jumping in these diagrams indicates buckling wave created due to the
effect of external pressure applied on it. It is observed that the first
jump occurs at 5 kPa external pressure at LVDTs 3 and 8 locations
with 11 and 5 mm values, respectively. This stage is called the initial
296 M. Rastgar, H. Showkati / Journal of Constructional Steel Research 145 (2018) 289–299

(A) (B)

Fig. 13. (A) Full buckling of Spec1 and instability threshold (B) failure of the Spec1.

Fig. 14. (A) Full buckling of Spec2 and instability threshold (B) failure of the Spec2.

Fig. 15. (A) Full buckling of Spec3 and instability threshold (B) failure of the Spec3.

buckling of the specimen. By continuation of the loading, other buckling

Fig. 16. Radial displacement of Spec1 at the locations of installed LVDTs.
waves are created in the body of the tank where the jumps in the
diagrams exhibit this issue. Finally at pressure loading of 19 kPa the
buckling of the tank body is completed. This stage is called as the full
buckling of the tank.
By continuation of the pressure loading up to 22 kPa the specimen
reaches its ultimate strength and then it experiences failure and collapse
occurs suddenly. The stage between the full buckling and failure of
the specimen is called the post-buckling stage. With respect to the
diagrams, the post-buckling capacity of the specimen reached 3 kPa.
The specimen failed at LVDT3 location, and the corresponding diagrams
of this point are utilized in the consequent comparisons.
In Fig. 17 radial displacement diagrams of SPEC2 at the locations of
installed LVDTs are shown in terms of uniform external pressure.
These diagrams show that radial displacement in this specimen at the
location of LVDT3 differs with respect to other portions. In this specimen
 M. Rastgar, H. Showkati / Journal of Constructional Steel Research 145 (2018) 289–299
Fig. 17. Radial displacement of SPEC2 at the locations of installed LVDTs.
with the start of loading, the radial displacement of the tank body
begins at LVDT3 location. While at other LVDTs small displacement
about 1 mm is observed for pressure loading up to 10 kPa, the displace-
ment at LVDT3 is 6 mm. The reason for this issue is the imperfection in
the form of concavity of vertical weld line in the tank body which is in
accordance with the first mode shape of buckling. In this specimen,
there is no initial buckling stage and by continuation of the loading
other buckling waves occur in the tank body and the jumps in the
diagrams indicate this issue.
Finally at 16 kPa pressure, the buckling of different parts of the
tank body is completed. By continuation of the pressure loading up to
16.5 kPa the specimen reaches its ultimate strength where it experi-
ences damage and collapse occurs suddenly. With respect to the
diagrams, the post-buckling capacity for this specimen is 0.5 kPa. Failure
of this specimen occurred at LVDT3 location and the corresponding dia-
gram of this point is used in other comparisons.
In Fig. 18 radial displacement diagrams of SPEC3 at the locations of
installed LVDTs are shown in terms of uniform external pressure.
These diagrams show that radial displacement at the locations of
LVDT3 and LVDT8 differs with respect to other portions. In this speci-
men with the start of loading, the radial displacement of the tank
body begins at these LVDTs locations. While at other LVDTs small dis-
placement about 1 mm is observed for pressure loading up to 14 kPa,
the displacement at LVDT3 and LVDT8 is 20 and 17 mm respectively.
The reason for this issue is concavity of vertical weld line imperfection
in the tank body which is in accordance with the first mode shape of
buckling. In this specimen, there is no initial buckling stage and by con-
tinuation of the loading other buckling waves occur in the tank body
and the jumps in the diagrams indicate this issue.
Fig. 18. Radial displacement of the tanks body at the unstable location.
297
Fig. 19. Radial displacement in all specimens at the locations of installed LVDTs.
Finally at 19 kPa pressure, the buckling of different parts of the
tank body is completed. By continuation of the pressure loading up to
19.5 kPa the specimen reaches its ultimate strength where it experi-
ences failure and collapse occurs suddenly. With respect to the
diagrams, the post-buckling capacity for this specimen is 0.5 kPa. Failure
of this specimen occurred at LVDT3 location and the corresponding
diagram of this point is used in other comparisons.
In Fig. 19 radial displacement of these three specimens is compared
with each other.
As is seen in Spec1, Point A is the start of buckling, Point B is the
initial buckling, Point C is the full buckling, CD zone is the post-
buckling capacity and point D is the ultimate capacity and collapse. In
Spec2, Buckling starts from point O and continues to point E and here
buckling is completed. EF zone is the post-buckling capacity and point
F is the ultimate capacity and failure. In Spec3, Buckling starts from
point O and continues to point G and here buckling is completed. GH
zone is the post-buckling capacity and point H is the ultimate capacity
and failure.
In Table 2, numerical values of buckling loads at these points are
summarized which is indicative the buckling behavior of the specimens
Considering the diagrams in Fig. 19 and Table 2, it is observed that
the initial buckling and full buckling load of specimens SPEC2 and
SPEC3 are lower. Comparing these two specimens it is seen that
SPEC3, despite having two imperfections, is stronger than SPEC2
with an imperfection. Thus it could be concluded that the concavity
of vertical weld line imperfection causes reduced initial membrane
strength and reduction in the full buckling and ultimate load. What-
ever the amount of this imperfection is increased its impact on the
buckling strength increases. The reason for this issue is compatibility
between the form of imperfection and buckling mode shape of struc-
ture. So that by start of loading the location of imperfection moves
and the first buckling wave is formed at the location of imperfection
and ultimately damage takes place and collapse occurs at the imperfec-
tion location. It seems that increased number of concavity of vertical
weld line Imperfection causes slight increase in the tank membrane
strength.
Table 2
Buckling load of all experimental specimens.
Initial buckling Full buckling Failure Post-buckling
load load load capacity
Spec1 5 kPa 19 kPa 22 kPa 3 kPa
Without imperfection
Spec2 – kPa 16 kPa 16.5 kPa 0.5 kPa
An imperfection
Spec3 – kPa 19 kPa 19.5 kPa 0.5 kPa
Two imperfection
298 M. Rastgar, H. Showkati / Journal of Constructional Steel Research 145 (2018) 289–299

Table 3
Buckling behavior of all specimens based on theory, ASME code and experiment methods.

Buckling Load (Theory)a Buckling Load (ASME)b Initial Buckling Full Buckling Failure Post- Buckling
Eq. (4) Eq. (6) Load (EXP) Load (EXP) Load (EXP) Load (EXP)

Spec1 23.5 kPa 23.3 kPa 5 kPa 19 kPa 22 kPa 3 kPa

Without imperfection n = 13 –
Spec2 23.5 kPa 23.3 kPa – kPa 16 kPa 16.5 kPa 0.5 kPa
An imperfection n = 13 –
Spec3 23.5 kPa 23.3 kPa – kPa 19 kPa 19.5 kPa 0.5 kPa
An imperfection n = 13 –
a 4 4
2 4 2 2 4 2
1 ðm þ n Þ ðR2 Þ þ m ð1−υ ÞC 1 ð3 þ 13:13 Þ ð0:5752 Þ þ 3 ð1−0:3 Þ225274:72
2 D 2 0:0188
Pcr ¼ 2
¼ 2
¼ 23:5 kPa
R ðm2 þ n2 Þ ðn2 þ 0:5m2 Þ 0:575 ð32 þ 13:132 Þ ð13:132 þ 0:5  32 Þ

Et3 205  106  0:0013 Et 205  106  0:001 mπR 1  π  57:5

D¼ ¼ ¼ 0:0188; C ¼ ¼ ¼ 225274:72; m ¼ ¼ ¼3
12ð1−υ2 Þ 12ð1−0:32 Þ ð1−υ2 Þ ð1−0:32 Þ L 60
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi sffiffiffiffiffiffiffiffiffiffiffiffiffi
rffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
rffiffiffiffiffiffiffiffiffiffi
u 2 pffiffiffiffiffiffiffiffiffiffiffiffi2ffi
u
4 6π 1−ν R R 57:5 57:5
n¼t 2
≈ 2:74 ¼ 2:74 ¼ 13:13:
ðRL Þ ðRt Þ L t 60 0:1

b
2:42E ðt=2RÞ2:5 2:42  205  106 ð1=1150Þ2:5
P cr ¼ 3=4
¼ ¼ 23:3 kPa:
ð1−υ2 Þ ½L=2R−0:45ðt=2RÞ1=2  ð1−0:32 Þ
3=4
½600=1150−0:45ð1=1150Þ1=2 

In Table 3 the results of investigating the effect of imperfection due were identified and categorized. The survey of imperfections revealed
to concavity of vertical weld line in the experimental specimens are that the imperfection in form of concavity of vertical weld line is the
given and comparison is made between the theoretical relations and most prevalent type of imperfection seen in the steel tanks. Presence
the relations in the ASME Code. Regarding this table the following issues of this imperfection causes reduced created waves in the body of the
could be referred to: tanks; consequently, the lengths of these waves are increased, resulting
in a reduced buckling load of the tank.
1- The corresponding failure load of SPEC1 is lower with respect to that Results show that the imperfections due to concavity of vertical weld
of the ASME Code which is smaller about 8.5%. line are very important in buckling of the tanks under uniform external
2- The full buckling loads obtained in the experimental specimens pressure. This imperfection decreases initial, full and post buckling
SPEC2 and SPEC3 are about 10% lower with respect to that of the capacity of the tanks under uniform external pressure, significantly.
specimen without imperfection also it is about 27% lower with Findings of this research show that for design of steel tanks under uni-
respect to that of the ASME Code. Therefore the imperfection due form external pressure load, 65% of the buckling load obtained from
to concavity of vertical weld line is very important in buckling of the ASME Code should be used.
the tanks under uniform external pressure.
3- The corresponding failure loads obtained in the experimental speci-
mens SPEC2 and SPEC3 are lower with respect to the ASME Code and Nomenclature
are about 16% smaller. D bending stiffness of plate
4- The deformations created in the tank with two imperfections is E modulus of elasticity
4 times of those in the healthy tank and the tank with one imperfec- L height of cylindrical shell
tion is two times of those in the healthy tank. m number of waves in the z-direction
5- The post-buckling capacity of the tanks with imperfection due to n number of waves in the θ-direction
concavity of vertical weld line is negligible, and the full buckling is N number of circumferential waves in the buckling situation
accounted as the threshold of collapse and failure. While in the R radius of cylindrical shell
tanks without this type of imperfection, the post-buckling capacity t thickness of cylindrical shell
amounts to about 13% of the full buckling. w deflection in the z-direction
6- The number of buckling waves obtained from the theory is 13 waves Z curvature parameter
and the number of buckling waves obtained in the experiment is Nθ axial load in the θ-direction
10 waves. Pcr buckling load
7- As there is a high probability of imperfection due to concavity of υ Poisson's ratio
vertical weld line in the tanks, the effect of this imperfection should σcr critical stress
be accounted for when calculating the full buckling load in the steel
cylindrical tanks. It is recommended to multiply the full buckling
load obtained from the ASME Code by 0.65. In other words for the References
design of these tanks under uniform external pressure load it is
recommended that 65% of the buckling load obtained from the [1] M. Farshad, Stability of Structures, 2nd ed. Mc Grow Hill, New York, 1983.
[2] J.G. Teng, Buckling of thin shells: recent advances and trends, Appl. Mech. Rev. 49
ASME Code be used as the full buckling load i.e.:
(1996) 263–274.
[3] H. Showkati, H. Ansourian, Influence of primary boundary condition on the buckling
of shallow cylindrical shells, J. Constr. Steel Res. 36 (1) (1995) 53–75.
Pdesign ≤ 0:65 PASME [4] J.H. Wang, A. Koizumi, Buckling of cylindrical shells with longitudinal joints under
external pressure, J, Thin-Walled Struct. 48 (2010) 897–904.
[5] L. Chen, J.M. Rotter, C. Doerich, Buckling of cylindrical shells with stepwise
8. Conclusions variable wall thickness under uniform external pressure, J. Eng. Struct. 33 (2011)
3570–3578.
[6] S. Aghajari, H. Showkati, K. Abedi, Experimental investigation on the buckling of thin
In this research, a field study was accomplished on the implementa- cylindrical shells with two-stepwise variable thickness under external pressure, Int.
tion of the storage tanks in a refinery site and resulted imperfections J. Struct. Eng. Mech. 39 (6) (2011) 849–860.
 M. Rastgar, H. Showkati / Journal of Constructional Steel Research 145 (2018) 289–299
[7] S. Jalili, J. Zamani, M. Shariyat, N. Jalili, M.A.B. Ajdari, M. Jafari, Experimental and
numerical investigation of composite conical shells' stability subjected to dynamic
loading, J. Struct. Eng. Mech. 49 (5) (2014) 555–568.
[8] T.G. Ghazijahani, H. Jiao, D. Holloway, Longitudinally stiffened corrugated cylindrical
shells under uniform external pressure, J. Constr. Steel Res. 110 (2015) 191–199.
[9] J.R. Vinson, The behavior of thin-walled structures: beams, plates and shells,
Mechanics of Surface Structures, 8, Kluwer Academic Publisher, 1998.
[10] C.R. Calladine, Causes of imperfection-sensitivity in the buckling of thin shells,
J. Thin-Walled Struct. 23 (1995) 215–235.
[11] J.G. Teng, X. Lin, J.M. Rotter, X.L. Ding, Analysis of geometric imperfections in full-
scale welded steel silos, J. Eng. Struct. 27 (6) (2005) 938–950.
[12] A. Hubner, J.G. Teng, H. Saal, Buckling behavior of large steel cylinders with patterned
welds, Int. J. Press. Vessel. Pip. 83 (2006) 13–26.
[13] R. Lo Frano, G. Forasassi, Experimental evidence of imperfection influence on the
buckling of thin cylindrical shell under uniform external pressure, J. Nucl. Eng.
Des. 239 (2009) 193–200.
[14] M. Maali, H. Showkati, S.M. Fatemi, Investigation of the buckling behavior of conical
shells under weld-induced imperfections, J. Thin-Walled Struct. 57 (8) (2012)
13–24.
[15] L. Yang, Z. Chen, F. Chen, W. Guo, G. Cao, Buckling of cylindrical shells with general
axisymmetric thickness imperfections under external pressure, Eur. J. Mech. A/Solids
38 (2013) 90–99.
[16] S.M. Fatemi, H. Showkati, M. Maali, Experiments on imperfect cylindrical shells
under uniform external pressure, J. Thin-Walled Struct. 65 (2013) 14–25.
299
[17] M. Khaled El-Sawy, Inelastic stability of liners of cylindrical conduits with local imperfec-
tion under external pressure, J. Tunneling Undergr. Space Technol. 33 (2013) 98–110.
[18] T.G. Ghazijahani, H. Jiao, D. Holloway, Experiments on dented cylindrical shells
under peripheral pressure, J. Thin-Walled Struct. 84 (2014) 14–25.
[19] J.M.T. Thompson, Advances in shell buckling: theory and experiment, Int. J. Bifurcation
Chaos 25 (2015) 1530001.
[20] G. Cao, Z. Chen, L. Yang, H. Fan, F. Zhou, Analytical study on the buckling of cylindri-
cal shells with arbitrary thickness imperfections under axial compression, J. Press.
Vessel. Technol. 137 (2015) 1–9.
[21] A. Lee, F.L. Jimenez, J. Marthelot, J.W. Hutchinson, P.M. Reis, The geometric role of
precisely engineered imperfections on the critical buckling load of spherical elastic
shells, J. Appl. Mech. 83 (2016) 1–11.
[22] A. Evkin, M. Kolesnikov, D.A. Prikazchikov, Buckling of a spherical shell under exter-
nal pressure and inward concentrated load: asymptotic solution, J. Math. Mech.
Solids 1 (2016) 1–13.
[23] J.W. Hutchinson, J.M.T. Thompson, Nonlinear buckling behavior of spherical shells:
barriers and symmetry-breaking dimples, Philos. Trans. A Math. Phys. Eng. Sci.
375 (2093) (2017).
[24] M.H. Jawad, Theory and Design of Plate and Shell Structures, 1st edition Chapman &
Hall Inc, 1994.
[25] D.O. Brush, B.O. Almorth, Buckling of Bears Plates and Shells, Mc Grow Hill, New
York, 1975.
[26] American Society of Mechanical Engineers, Boiler and pressure vessel code section
VIII, rules for construction for pressure vessels, division 1. New York, ASME (2013).