JOURNAL OF CURRENT RESEARCH IN SCIENCE (ISSN 2322-5009)

CODEN (USA): JCRSDJ

2014, Vol. 2, No. 5, pp: 664-672
Available at www.jcrs010.com
ORIGINAL ARTICLE
MECHANICAL BEHAVIOR INVESTIGATION FOR AN ATMOSPHERIC STORAGE
TANK ACCORDING TO API 650 UNDER LOADS USING FEM
Alireza Naddaf Oskouei, Ehsan Nazari Naghani
Department of Mechanical Engineering, Imam Hossein Comprehensive University, Tehran, Iran
ABSTRACT: This paper analyzes the mechanical behavior of an atmospheric storage tank under loads
(dead load, weight of fluid, wind and earthquake) that has been designed according to standard analytical
equations of API 650. After defining analytical equations results, we study the tank behavior by numerical
analysis of finite element software ABAQUS. Finally, comparing these two results, after defining critical
displacement and buckling points we recommend some solution to increase the buckling strength of tank
cylindrical shell that will conclude to increase its shelf life. The results show that values of hoop
membrane stress in tank shell obtained by numerical method is proportional to those obtained by design
analytical equations of standard and they fluctuate in a parallel way. Moreover the values of hoop
membrane stress obtained by numerical method are 9% less than analytical method. The studies on
stability of the tank, represents that installing boxlike stiffening rings in tank shell causes an improvement
of 72% in buckling strength of tank shell.
KEYWORDS: Atmospheric storage tank, Cylindrical shells, Finite Element, Buckling strength, API 650
INTRODUCTION
In order to design structures safely, their
behavior analysis in a variety of loading is
considerably important. The first step to analyze
the behavior of structure faced with different
loads is the stress analysis that can be
performed in different methods of theoretical,
numerical and experimental.
Atmospheric storage tank is one of the most
important structures widely used in many
industries, especially in oil, gas and
petrochemical industries. The standard used in
the design of these tanks is API 650. The design
equations used in the standard is based on some
simplifier assumptions. They state in templates
of stress analysis method which can be used only
for classic loading and geometries. While the
results of the surveying geometric shapes and
the actual loading using numerical methods such
as finite element method can conclude to more
accurate, realistic and reliable result. Of course
in experimental stress analysis, there is not
numerical errors which is usual in there
numerical methods and structures are studied
under real conditions. Because of various factors
such as not take in account thorough parameters
affecting on the strength and stability of the
structure due to the different uncertainties in
design equations defined in the standard API
650 so these tanks face to failure and
deformation. for these reasons studying and
comparing their mechanical behavior using the
results of theoretical equations in standard API

650 and numerical analysis software gained via

software ABAQUS would be useful In offering
solutions to increase the shelf life of the tanks.
For the first time in 1916 in the United States of
America "tank Manufacturers Association"
which later renamed to the "Steel Tank
Institute" (STI) started standardizing and
regularizing equations for the large quantities of
petroleum storage with high degree of
flammability.
In1922,
the
Underwriter
Laboratories (UL) by developing their safety
standards published the first standard entitled
"Above Ground Steel Tanks for Flammable and
Combustible Liquid. American petroleum
Institute (API) was established in 1919 with the
publication of API 650 they started providing
standards criteria and requirements for
materials, design, fabrication and installation for
vertical tanks, cylindrical, above ground, open
roof, enclosed and welded steel tanks in a spread
variety of dimensions and capacity for different
storage (Long and Garner, 1961).
Housner, (1954) examined the effects of the
earthquake on atmospheric storage tanks.
McGrath, (1963) began to investigate the
stability of the tanks shell designed according to
the standard API 650 against the load of wind.
Novozhilov, (1964); Vlasov, (1964) and Dupuis
et al.,(1970) providing strain-displacement
equations studied on numerical analysis for
deep and shallow shells based on thin shell
theory in classical mechanics. Grafton and
Strome, (1963) were pioneer to obtain finite

Corresponding Author: Ehsan Nazari Naghani, Department of Mechanical Engineerng, Imam Hossein
Comprehensive University, Tehran, Iran. E-mail: [email protected]

Mechanical Behavior Investigation For An

element for axisymmetric shells with symmetric
loading that their finite elements composed of a
ring of frustum cone. Eslami and Alizadeh,
(1994) in the analysis of spherical shells under
the asymmetric loads employed a mixed finite
element model. Shakeri et al., (1993) using finite
element method analyzed cone shells elasticity
under the impact loading. Rahgozar et al., (2005)
examined the static analysis on rotating thin
shell based on finite element method. Irani and
Fathi, (2001) analyzed the free vibration of
shells via the finite element method, and in that
study they used a special Lagrangian curved
finite element.
In this paper, we compare the mechanical
behavior of a floating roof atmospheric storage
tank using the results of the theoretical
equations of standard API 650 and numerical
analysis gained from software ABAQUS. Finally,
we provide some recommendations to increase
the tank shell buckling strength and it shelf life.
THE TYPES OF TANKS WITH DIFFERENT
OPERATION PRESSURE
In viewpoint of operation pressure or internal
pressure, tanks are classified into three major
groups. First tanks, known as atmospheric
storage tanks are utilized in an operation
pressure or internal pressure of less than 18 kPa
and standard API 650 is used to design them
(API 650, 2010). The second group, is called the
low-pressure storage tanks have an internal
pressure of 18 kPa <P <100 kPa and standard
API 620 is used in their design (API 620, 2012).
The third category, known as pressurized
vessels, their operation pressure or internal
pressure is P> 100 kPa and the design standard
of the tanks is ASME Sec VIII (ASME Sec VIII,
2013).
TANK DESIGN
According to API 650 standard definitions,
process and geometric properties of the tank
assumed to be as follows:
3.1. Design Assumptions
The tank assumed to be an external floating roof
(EFR), Double deck and containing a
hydrocarbon condensate and is considered with
a density of 0.735 gr/cm3, having a diameter of
72 m, height 18.5 m and the design temperature
of 85 c. The wind speed at a height of 10 m
above ground level assumed 34.7 m/s, the
maximum design liquid level 16.67 m, corrosion
allowance for shell, floor and roof plates of the
tank 0 mm, the seismic zoom 4 and importance
factor for seismic use group considered 1.25.
The material specification of plates according to
mechanical properties in standard ASTM, type A

665
516-70 and A 283-C for the shell, A 283-C for the
floor, A 516-70 used for annular plates and A
283-C is chosen for the roof.
3.2. Design Of Bottom And Annular Plates
According to standard API 650, minimum
thickness tank floor plates without corrosion
allowance is 6 mm and a minimum nominal
width of rectangular plates and Sketch Plates
(floor plates that are connected to the annular
plate) is 1800 mm. A schematic image of the
plates layout is shown in Figure 1.

Figure 1: A schematic layout of the tank floor

plates.
In the standards, steels based on the increase in
toughness are classified in eight groups. If the
first Course of tank shell using allowance stress
of the material of five groups IV, IVA, V and VI is
selected, then the annular plates material of the
tank (Here A 516-70 of the Group IV ) will be the
same. Whenever allowable stress in hydrostatic
test condition (St) on the first course of shell is
less than 210 MPa and thickness of first course
was in the interval 32 <t 40 mm, then the
minimum thickness of the annular plates will be
11 mm. The minimum width of annular plates is
600 mm and should not be less than the value
obtained from Eq. (1):
=

215
()0.5

) 1(

Where tb is thickness of annular plates, H is

maximum design liquid level and G is specific
gravity of the stored fluid in terms of millimeter,
meter and gram per cubic centimeter
respectively. So the annular plates width is
considered 1400 mm.
3.3. Design Of Shell Plates
The minimum acceptable width in standard API
650 for tank shell plates is 1800 mm. If the tank
diameter is greater than 60 m, the minimum
shell plate thickness should be considered 10
mm. since by increasing height of the shell, the
hydrostatic pressure of the fluid decreases, thus
thickness of any tank shell courses must not be

666
less than its upper courses. Generally there are
two methods to design thickness of tank shell
plates. In the first that is called 1-foot method,
required thicknesses in the design point (0.3 m
higher than bottom of the each shell course) are
calculated. This method is applied for tanks with
diameters less than 60 m. The second method
called variable-design-point is used to design
tanks with diameters larger than 60 m. This is a
trial and error approach and the solution closes
to a converge number. Thickness obtained in
this approach is less than 1-foot method. This
method obtains lower thicknesses in different
courses, thereby reduces the weight of the tank.
The more notability of applying this method is
its capability to design tanks with larger
diameter when we are faced with limitation of
plate thickness (Carluccio, 2007). Using variabledesign-point method, the first shell thicknesses
t1d and t1t in the design condition and hydrostatic
test conditions is respectively obtained in terms
of millimeter from the following equations:

Naddaf Oskouei and Nazari Naghani, 2014

in terms of millimeters. In Eq. (6) the value of x
that is actually the distance between variable
design points from the bottom of each course
can be obtained by using the least value of Eq.
(7):
1 = 0.61( )0.5 + 320
2 = 1000
3 = 1.22( )0.5

Where C and K are:

0.5 ( 1)
=[
]
(1 + 0.5 )

0.0696 4.9
)(
) +

) 2(

1 = (1.06

0.0696 4.9
)(
)

) 3(

Where D is tank diameter, CA is corrosion

allowance in terms of meter and millimeter
respectively. Also Sd and St are maximum
allowable stress of shell plates used in the tank
for respectively design and hydrostatic test
conditions in terms of MPa. Thickness of courses
has to be greater than designed thickness for the
shell (td) plus corrosion allowance and shell
thickness in hydrostatic test mode (tt). To
calculate the shell thickness in upper courses for
design and hydrostatic test conditions, with
calculation the ratio of A in Eq. (4) for lower
course, thickness of upper course (t2) can be
calculated according to the Eq. (5) and Eq. (6):
=

1
(1 )0.5

) 4(

1.375 2 = 1
2.625 2 = 2 = ( , )
1.375 < < 2.625 2 =
1
2 + (1 2 )[2.1
]
1.25(1 )0.5

4.9 (

)
1000
+

4.9(
)
1000
=

) 5(

) 6(

Where h1 is lower course height, r is radius of

the tank and t1 is actual course thickness
without corrosion allowance that all of them are

) 8(

Where tL and tu are thickness of the lower and

upper courses in the joint location in terms of
millimeter respectively.
Regarding design
assumptions, size and material of tank shell
plates are obtained from Table 1.
Table 1: Size and material of tank shell plates.
No.

1 = (1.06

) 7(

1
2
3
4
5
6
7
8

THK
(mm)
36.64
32.97
27.81
21.79
15.98
11.43
10
10

Material
A 516-70
A 516-70
A 516-70
A 516-70
A 516-70
A 516-70
A 283-C
A 283-C

Height
(mm)
2433
2433
2433
2433
2433
2433
1901
1901

Course Weight
(kg)
158488.21
142613.43
120293.59
94253.76
69122.31
49441.05
33797.26
33797.26

3.4. Floating Roof

According to requirements of API 650 floating
roof of the tank is considered a double deck with
a thickness of 5 mm for plates of the top and
bottom deck, diameter 71.6 m, height 0.715 m in
the adjacent inner shells and it has seven rings
with the same length of the openings. Between
every two rings some compartment are located
to enclose the air for roof flotation. To control
deformation and to tighten the ring plates in
every compartment 5 radial trusses joint them
in the equal intervals. in each trusses two UNP
80 as longitudinal parts in its top and down is
used as well as two angle beams with
dimensions 50 mm 50 mm 5mm with equal
openings from each other as vertical parts are
applied. Considering above circumstances, mass
of the roof is estimated about 417647 kg.
3.5. Design For Resist Against Wind Load
The maximum height of the tank shell that does
not need to be reinforced can be calculated from
the following equation:
3 190 2
1 = 9.47( ) (
)

) 9(

Mechanical Behavior Investigation For An

The above equation is defined based on the
dynamic pressure of the wind at height z(m),
which the first time by McGrath, (1963) was
introduced and in standard ASCE1 7 has also
been discussed. According to him, the pressure
is expressed in Eq. (10):
= 0.00256 2 = 1.48 kPa

)10(

By adding a pressure of 0.24 kPa to exert drag

effects and vaccum (external pressure) inside
the tanks respectively closed and open top,
McGrath considered total wind dynamic
pressure to be 1.72 kPa. In the above equation
is velocity pressure exposure coefficient of
wind in height of z, topographic factor,
wind directionally factor, V design wind speed
(for a 3-sec gust) in height of 10 m of ground
level in terms of km per hour, I importance
factor and G is the wind impact factor.
Importance factor is calculated for buildings or
other structures, taking into account the degree
of risk to human life and damage to property.
After calculating H1, transformed shell height is
determined. This height is calculated using Eq.
(11) that actual width of each shell course
turned to transposed width with a thickness
equal to thickness of the top course.
= (

5
)

)11(

That Wtr and W respectively are transposed

width and actual width of each shell course,
tuniform thickness of the last course, tactual
thickness of the course that is set to transposed
its width and all are in millimeter. If
> 1 an intermediate wind girder is
necessary that should be located in middle of
transposed tank shell height. For the tank,
according to table 1, an intermediate wind
girder is necessary in a distance of 5.91 m from
the highest point of the shell that the minimum
required section modulus Z(cm3) is obtained
from below equation:

667
( + )
+ 0.4 <
2

Where Mw is overturning moment caused by

vertical and horizontal wind pressure, MPi is
internal pressure moment, MDL is dead load
moment and MF is fluid moment that all of them
are calculated in around of the joint between
shell and floor in terms of Newton-meter.
According to the design assumptions, these both
conditions were satisfied by the tank and it is
considered as a self-anchored tank.
3.6. Seismic Design Of Storage Tank
Based on the Housner theory, behavior of the
product liquid inside a vertical cylindrical tank
when the tank is disposed to an earthquake is
divided into two parts (Long and Garner, 1961).
That part of the liquid located in the lower part
of the tank as a solid object with the tank moves,
displaces and accelerates. It is called impulsive
section. The natural period of vibration
associated with this component is a function of
the size and stiffness of the tank itself and is
usually to be found in the 0.1 s to 0.4 s range.
The other part of the liquid located in the upper
part of the tank freely moves to form Slosh. This
part is known as convective section. This section
has a longer natural response rather than the
impulsive section and depending on the size of
the tank usually within 5 s to 10 s is formed. In
Housner model that is shown in Figure 2, the
impulsive component is rigidly linked to the
tank walls, while the convective component is
attached by weak springs.

)12(

Figure 2: Modeling of convective and impulsive

section with the weight of roof and shell.

To control the tank stability against overturning,

according to the standard requirements of API
650,whenever the tank satisfy two conditions of
Eq. (13) it does not need to use the anchor bolts
to anchorage it.

Accordance with API 650, Ti and Tc, the natural

period of the impulsive and convective sections
are respectively obtained from the following
equation:

2 2
(
)
17 190

0.6 + <

1.5

American Society of Civil Engineers

)13(

= 0.128
= 1.8
0.578
=
tanh (3.68 )

)14(

668
Then Wi and Wc which are weight of impulsive
and convective liquid (in Newton) for the short

tanks that 1.33 from the following equation

can be calculated:

tanh (0.866 )
]
= [

0.866

3.67
= 0.23 tanh(
)

= =

)15(

Where Wp is total weight of liquid storage. Xi and

Xc are distance between the bottom of shell and
the center of lateral seismic force respectively
for impulsive and convective sections that for
short tanks is obtained from follow equation in
terms of meter:
= 0.375
= [1

3.6
)1

]
3.67
3.67
sinh (
)

cosh (

)16(

Total design base shear V is also can be gained

by Eq. (17) in which Vi is design base shear
related to impulsive section due to weight of the
tank and its liquid storage. Vc is design base
shear related to convective section due to
sloshing weight.
= 2 + 2
= ( + + + )
=

)17(

Where is the weight of the fixed roof, is

the weight of the floor and both are in terms of
Newton. and are seismic factor for
impulsive and convective sections that for this
tank are respectively 0.33 and 0.02. Overturning
moment that is imposed to bottom of shell
surrounding is:
= [ ( + + )]2 + [ ( )]2

Naddaf Oskouei and Nazari Naghani, 2014

Total hoop stress in shell caused by seismic
movement for the short tanks is calculated via
this equation:

)18(

Where and are height of gravity center of

shell and roof from the floor in terms of meter.
The parameters defined in Eqs. (15) to (18) for
the tank is shown in Figure 3.

2 + 2 + ( )2

)19(

Where:
= 4.9( 0.3)

= 8.48 [ 0.5 ( ) ] tanh (0.866 )

3.68( )
1.85 2 cosh (
)

=
3.68
cosh (
)

)20(

That is product hydrostatic hoop stress in the

shell and is hoop stress in the shell due to
impulsive and convective forces of the stored
liquid and both are in terms of MPa. Nh is
product hydrostatic membrane force, Ni and Nc
are hydrodynamic hoop membrane forces in the
shell due to respectively convective and
impulsive sections and both are in terms of
Newton per millimeter. Y is the distance of fluid
surface to the analysis point and is vertical
earthquake acceleration coefficient defined as
0.23.
FINITE ELEMENTS TO ANALYZE SHELLS OF
REVOLUTION UNDER NON-AXISYMMETRICAL
LOADIND
Curved surfaces or shells that have practical and
industrial aspects are known as thin shells if
ratio of thickness to radius of curvature r is
equal or less than 1/20. For thin shells that are
important from a practical point of view, this
ratio may be 1/100 or smaller (Ugural,1981). In
many cases it appears that the shells of
revolution
are
located
under
Non
axisymmetrical loading. For example, wind or
earthquake
effects
on
chimneys
and
atmospheric storage tanks produce loads that do
not have axial symmetry. However any desired
loading on a revolving object can be decomposed
in two symmetric and asymmetric loading with
respect to the plane of symmetry which is
included axis of rotation (Plane through the axis
of symmetry). Symmetric loads analysis or
Fourier
decomposition
(Sokolnikoff
and
Redheffer, 1966) to m harmonic terms will
result in:

= cos
=0

= cos
=0

Figure 3: Seismic moment and force diagram.

= sin
=0

)21(

Mechanical Behavior Investigation For An

Where , and are function of only z and
j. when = 0 then = 0 and Eq. (21) will be in
axisymmetrical loading mode. Otherwise, it
means = 1,2, , Eq. (21) will be in Nonaxisymmetrical loading mode that respected to
the plane passing on z is symmetric.
Figs. 4-a, 4-b and 4-c are showing the first
harmonic loads term along the r, z and . If the
loading was anti-symmetric with respect to the
symmetry plane, position of cos and sin in
Eq.(21) has to be replaced.

669
For the analysis it is necessary to define global
displacement derivative of Eq. (22) with respect
to local coordinates. So:
,
,
,
3
,
, =
,
=1
,
,
[, ]

0 0 , 0

0
0 0

0 0
0

, 0 , 0

0
0
0

0 0 2

0 ,
0 ,
[ 2 ]
0

0
0
0
0 ]

,
0

0
0
0
0
0
[0

)24(

The following abbreviations are used in the

coefficients matrix:
= sin
= cos

= sin
= cos

)25(

By defining Jacobian matrix in Eq. (26) and using

its inverse, derivatives of global displacement
with respect to global coordinate is obtained as
below:
Figure 4: Loading without axial symmetry: (a)
1 cos (b) 1 cos (c) 1 sin
Response of a shell of revolution to the harmonic
load series is a harmonic series of global
displacement. For the shell finite element
(Figure 5-a), the displacements are defined as
follows:
cos
0
0

cos
0 ]
[ ] = [ 0

0
0
sin
=
3
3

sin 0
{ [ ] + [ cos 0] [ ] }
2

=1
0
1
=1

1 =

(1 )
2

2 = 1 2

3 =

(1 + )
2

)22(

)23(

Where is the chosen shape functions for the

shell finite element and is a small rotate axial
x(Figure 5-a). As before, if the loading was antisymmetric with respect to the symmetry plane,
the function position of cos and sin has to
be replaced.

,
= [,
0

,
,
0

,
[]=[
,

0
0]
1

)26(

] [ , ]
,

)27(

For loads without axial symmetry on revolution

finite element (Figure 5-a), 6 types of non-zero
displacements and strains are considered that
form equation of strain-displacement:

=
=

[ ]

,
,

1
( + , )

, + ,
,
, +

]
,
[

)28(

Using Eq. (27) and Eq. (28), i th part of the

matrix B, the matrix strain-displacement of finite
element for the j th harmonic response can be
written as Eq. (29):

( ) =

0

[

0

0

0
0

( )

0

0
1
1
( )

2
2
( )
0
1
1
( )
( )
2
2
1

( ) (
)
2

2 ]

)29(

Where = 1,2,3 and = 0,1,2, , and:

1
= 11
,

Figure 5: (a) Hoop finite element (b) Nodal

rotation

1
= 21
,

1 )
( + 12

2
1 )
( + 22

=
2

)30(

670
So the finite element stiffness matrix Kj can be
formulated as below:
1

= ||

)31(

1 1

wherein = 2 for
= 0 and = 1 for =
1,2, . In addition, equivalent nodal loads for
each harmonic series of asymmetric body forces
are in the form of following equation:
1

= ||

( = 0,1,2, , )

1 1

)32(

Where:
1 0 0 0 0
= [0 1 0 0 0]
0 0 1 0 0

( = 1,2,3)

)33(

And body forces (per unit of volume and

constant in thickness) is equal to:
= { , , }

Naddaf Oskouei and Nazari Naghani, 2014

COMPARING SOFTWARE NUMERICAL
RESULTS WITH THE STANDARD ANALYTICAL
RESULTS
The results obtained from the finite element
model and the standard API 650 is shown in
Table 2 and Figure 6. Because maximum
deflection and stress occur in the sixth shell
course (Figure 7 and 8), therefore, the numerical
results of software ABAQUS support the API 650
argument about selecting location for an
embedded stiffening ring.
Table 2: Results of API 650 Standard and
ABAQUS software for global hoop stresses, T.
Course
No.

(+)
(MPa)

()
(MPa)

1
2
3
4
5
6
7
8

159.43
170.98
175.88
188.48
205.33
211.03
150.03
76.45

73.46
77.03
72.77
70.97
69.49
62.78
36.77
11.76

S11(MPa)Mesh
Size=2
127.6
133.1
145.2
154.1
188.5
193.4
114.7
55.0

)34(

Furthermore, the equivalent nodal loads due to

initial strain can be expressed as Eq. (35):
1

0 = 0 ||

( = 0,1,2, , )

1 1

= {0 , 0 , 0 , 0 , 0 , 0 }

)35(
)36(

Finally, the stress values for each harmonic

response are as follows, that is the exact
equations of stress-strain:
= ( )

( = 0,1,2, , )

)37(

Figure 6: Comparison of stresses provided in

standard API 650 and software ABAQUS.

These stresses are along the local coordinates.

Finite elements for a revolution shell have
consistent mass matrix to vibrate with axial
symmetry and without axial symmetry
accordance with Eq. (38):
2

cos 2

( = 0,1,2, , )

)38(

By integrating the above equation with respect

to we have:
=

)39(

Figure 7: Contour of hoop stresses in the tank

shell (magnification 100).

Wherein
= 2 for = 0 and = 1 for =
0,1,2, , . Note that replacing cos with sin
in Eq. (38) cannot change the result.
Taking advantage of numerical analysis was due
to better understanding and interpreting, in
addition comparing the results obtained from
the standard theory and the computer software
used here is engineering software ABAQUS that
is based on finite element analysis.

Figure 8: Contour of horizontal displacement in

the tank shell (magnification 100).

Mechanical Behavior Investigation For An

Unlike API 650 prediction, the results of
numerical software (Figure 8) indicate the
occurrence of a buckling of 21.7 mm in the
height of 2.445 m from the shell bottom
(between the first and second courses).
Naturally, this buckling will seriously damage
the tank piping system. To avoid the buckling,
embedded stiffening rings on the tank shell in
several patterns can lead to a reduction in the
amount of displacement in mentioned location.
Installing a stiffening with T-section and
dimensions of 0.5 m 0.2 m 0.01 m (larger side
or web of T is perpendicular to the shell) at a
height of 2.445 m from the bottom of the tank
can result in reducing the amount of deflection
up to 19.4 mm (Figure 9).

Figure 9: Using stiffening with T-section

(magnification 100).
Using a band of 1m 0.04 m and boxlike
stiffening ring with dimensions 1 m 0.5 m
0.05 m (larger side is perpendicular to the shell)
at the mentioned height, can reduce the rate of
deflection respectively up to 13.15 mm and 6
mm that is shown in Figures 10 and 11.

Figure 10: Using band to reduce the deflection

(magnification 100).

Figure 11: Using boxlike stiffening ring to

reduce the deflection (magnification 100).

671
CONCLUSION
Comparing the results of the API 650 standard
analytical equations and numerical solution of
finite element software ABAQUS that is shown in
Table 2, shows that values of hoop membrane
stress in tank shell courses that gained via
numerical method is proportioned to those
gained via design analytical equation of standard
and they fluctuate in a parallel way (Figure 6). In
additional, the amount of hoop membrane stress
gained via numerical method is 9% less than
analytical way. The studies on stability of the
tank represents that installing a boxlike
stiffening ring in tank shell can conclude to an
improvement of 72% in buckling strength in the
tank shell.
REFERENCES
API STD 620. Recommended rules for design and
construction of large, welded, low-pressure
storage tanks. Washington, DC, USA:
American petroleum institute 2012.
API STD 650. Welded tanks for oil storage.
Washington, DC, USA: American petroleum
institute 2010.
ASME STD Sec VIII. Rules for construction of
pressure vessels. New York. American society
of mechanical engineers 2013.
Carluccio AD. Structural characterisation and
seismic evaluation of steel equipments in
industrial plants. University of
Naples
Federico II 2007;pp:106-110.
Dupuis G, Goel J. A curved finite element for thin
elastic shells. International journal of
structure 1970;6(11):1413-1428.
Eslami MR, Alizadeh, SH. Mixed galerkin finite
element analysis of non-axisymmetrically
loaded spherical shells. Scienta Iranica
1994;1(2):101-109.
Grafton PE, Strome DR. Analysis of
axisymmetrical shells by the direct stiffness
method. AIAA Journal 1963;1(10):23422347.
Housner GW. Earthquake pressure on fluid
containers. A report on research conducted
under contract with the office of naval
research. California institute of technology.
Earthquake research laboratory 1954.
Irani F, Fathi KA. Free vibration analysis of shells
via FEM. International journal of engineering
sciences, Iran University of Science and
Technology
2001;12(2):111-125.
[In
Persian]
Long B, Garner B. Guide to storage tank and
equipment.
Professional
engineering
publishing 1961;pp:263-274.
McGrath RV. Stability of API standard 650 tank
shells. Proceeding of the American Petroleum
Institute,Section III- Refining, American

672
Petroleum Institute, New York 1963;43:458469.
Novozhilov VV. Thin shell theory. Noordhoff Ltd.
Netherlands 1964.
Rahgozar R, Sohi SG, Javanmardi M. Static
analysis of thin shells of revolution via FEM.
Second conference on thin shell structures.
Kerman,
Shahid
Bahonar
Univercity
2005;pp:227-242. [In Persian]
Shakeri M, Eslami MR, Babayi R. Elastic analysis
of conical shells under impact load.
International congress on computation
methods in engineering, Shiraz University
1993;pp:27-33.
Sokolnikoff IS, Redheffer RM. Mathematics of
physics and modern engineering. McGrawHill, New York, 1966.
Ugural AC. Stresses in plates and shells. McGrawHill, New York 1981;pp:199.
Vlasov VZ. General theory of shells and its
applications in engineering. NASA TTF-99
1964.

Naddaf Oskouei and Nazari Naghani, 2014