See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/342894794

Effect of Supporting System on Dynamic Buckling of Elevated Water Tanks: A

Case Study

Article in International Journal of Safety and Security Engineering · June 2020

DOI: 10.18280/ijsse.100304

CITATIONS READS

5 1,013

3 authors:

Nasser Dine Hadj Djelloul Mohammed Djermane

University of Bechar University of Bechar
18 PUBLICATIONS 29 CITATIONS 56 PUBLICATIONS 259 CITATIONS

SEE PROFILE SEE PROFILE

Noor Sharari
University of Technology Sydney
10 PUBLICATIONS 35 CITATIONS

SEE PROFILE

All content following this page was uploaded by Nasser Dine Hadj Djelloul on 14 July 2020.

The user has requested enhancement of the downloaded file.

 International Journal of Safety and Security Engineering
Vol. 10, No. 3, June, 2020, pp. 333-342
Journal homepage: http://iieta.org/journals/ijsse

Effect of Supporting System on Dynamic Buckling of Elevated Water Tanks: A Case Study
Nasser Dine Hadj Djelloul1*, Mohamed Djermane1, Noor Sharari2
1
FIMAS Laboratory, University of Tahri Mohammed, Bechar B.P 417, Bechar 08000, Algeria
2
Department of Civil Engineering, University of Technology Sydney, Sydney, NSW 2007, Australia

Corresponding Author Email: [email protected]

https://doi.org/10.18280/ijsse.100304 ABSTRACT

Received: 27 January 2020 The elevated water tanks are high importance structures for the humanity lifelines. These
Accepted: 6 June 2020 elevated tanks are considered as very sensitive structures for seismic movement
conditions. Among the reasons for the damage and failure of elevated tanks is the design
Keywords: of its support systems. For this reason, several theoretical and experimental researchers
tank staging, nonlinear analysis, finite studied the performance of this type of structure under seismic loading. The present study
elements, seismic, dynamic buckling, fluid– aims to demonstrate the supporting system effect on dynamic buckling of the elevated
structure interaction water tank, using three dimensional finite element technique the seismic response of two
elevated water tanks was established taking into account the following factors; the fluid-
structure Interaction (FSI), the wall flexibility, different nonlinear time histories analysis,
and the material and geometric nonlinearity. Indeed, the application of three different
instability criteria for the critical PGA estimate using two seismic excitations, namely El
Centro and San Fernando earthquake. The numerical values are compared and no
significant effect is found of the supporting system for convective fundamental frequency;
however, strongly disturbed impulsive fundamental frequency. In addition, the effect of
supporting system and the frequency content of the earthquake on PGAcr are clearly
shown. A percent increase of PGAcr can reach up to 37.48%.

1. INTRODUCTION was assumed to be non-viscous and incompressible. Analyses

indicated that the difference between the results obtained from
Storage tanks are important means used mainly to store intze tank and the equivalent cylindrical tank were negligible.
water (for daily life use), and hydrocarbon (for industrial use). Dutta et al. made a comparison between raised tanks support
These tanks constructed in a different shape mainly systems and proposed approximate empirical equations for
rectangular and cylindrical. Cylindrical tanks are the most evaluating the lateral, horizontal and torsional stiffnesses for
commonly used because they have a simple design and more different support systems [4-6]. Shrimali and Jangid
functional when it comes to the load’s resistance. These investigated two types of isolated tank models in which the
elevated tanks are supported on steel or reinforced concrete bearings are placed at the top and base of the steel tower [7].
systems to provide self-pumping from the force of gravity. In Livaoglur and Dogangun [8] studied the effects of soil-
addition, the elevated tanks ought to function in the seismic structure interaction on the seismic behaviour of the elevated
location after earthquakes to maintain the water supply during tanks with a structural frame supporting system. It was
and after the earthquake. Environmental pollution can be result concluded that the soil-structure interaction effect has
if the nuclear power plants and oil industry tanks are leakage, significant impacts on the shear stress, overturning moments,
which created catastrophic results when they were damaged axial forces and lateral displacement.
and totally fell down. Moslemi et al. [9] studied the performance of elevated tanks
For example, the earthquakes in Alaska 1964, Niigata 1964, under seismic excitations. In this study, the analysis of free
Parkfield 1966, San Fernando 1971, Miyagi prefecture 1978, vibrations is carried out on models. The response of rigid and
Imperial County 1979, Coalinga 1983, Northridge 1994, flexible models with different basic conditions which are fixed
Asnam 1980 and Koaceli 1999 demonstrated the same and articulated under the two horizontal and vertical
noticeable damages [1]. components of the earthquake are obtained using the direct
In terms of seismic analysis and modelling, many integration method. It was concluded that there is a very good
researchers have studied the dynamic behavior of elevated agreement between current practice and the finite element
tanks. Housner [2] allowed the engineers to perform the method in seismic analysis.
analysis of the seismic responses of the elevated rigid tanks by Leonard et al. [10] evaluated the buckling capacity using
using the two-mass analytical method. The liquid was assumed finite element technique. In this paper the analysis used
to be incompressible and inviscid; this analytical model has provide powerful tools for estimating buckling modes and
been adopted in many codes for it is simplicity and accuracy. evaluating the effects of stiffener arrangements.
The dynamic behavior of rigid Intze tanks was investigated Malhothra et al. [11] proposed simplified theoretical
by Joshi [3]. The results of the model were evaluated and calculation tables for seismic design of ground-supported
compared to those of equivalent cylindrical tanks. The fluid cylindrical tanks. The procedure takes into account both

333
impulsive and convective components.
Algreane et al. [12] studied a method to add the impulsive
mass to the walls of tanks alternative to Westergaard approach.
They simulated six models approach distributed masses to
determine the fundamental period. The results showed that the
distribution did not significantly affect the dynamic analysis
of elevated tanks.
Shakib et al. [13] investigated the filling level of the tank
supported by a moment-resisting frame by using FE analysis
subjected to horizontal seismic excitations. Their findings
indicated that the maximum response does not always occur in
the full tank case.
Sweedan [14] investigated the effect of vertical earthquake
excitation on the dynamic behaviour of elevated tank and
propose a simplified equivalent mechanical model. It has been Figure 1. (a) Collapse elevated tank 1980 El-Asnam
shown that the vertical excitation of these tanks can cause a earthquake [23], Algeria; (b) Elephant foot buckling [24]
significant increase in compression stresses generated in the
walls of the tanks. Furthermore, Chaduvula et al. [15]
investigated an experimental model of cylindrical steel 2. THEORETICAL SIMPLIFIED MODEL AND CODE
elevated water tank subjected to combined horizontal, vertical PROVISIONS
and rocking motions. The main conclusions they found; there
is not much rocking effect from the base on sloshing by visual In order to understand what is behind the various
observation and pressure on the tank walls due to vertical international codes, a brief review of codes is presented in this
excitation increases with increasing excitation. section and they are used to validate the finite element model
The Damages caused by past earthquakes showed the in the following sections.
importance of the supporting system for the elevated tanks. Many current codes such as Eurocode-8 [25] employed
Jabar and Patel [16] made a comparison between basic staging Veletsos and Yang’s model [26] for determining the dynamic
patterns, staging with radial bracing and staging with cross behaviours for the rigid circular tanks and Veletsos [27];
bracing. They studied the effect of this supporting staging on Haroun and Housner [28] for the flexible circular tanks with
the base shear, bending moment and displacement of the the approach proposed by Malhotra et al. (2000) Table 1. The
elevated concrete tank under different earthquake time American Concrete Institute ACI 350.3-06 [29], the American
histories using Sap2000 software. In the same subject, Water Work Association AWWA [30], and the American
Vyankatesh and Varsha [17] considered the effect of Petroleum Institute API 650 [11] codes employed the
supporting staging on the dynamic behaviour of elevated Housner’s mechanical model [2] with some modifications for
concrete tanks with various fluid levels in different seismic determining the dynamic behaviours of the liquid in a
zones. In a recent study published by Pole and Khedikar [18] container. According to Euro code-8 (2003); the parameter
three different types of supporting systems were analysed with should be calculated using the following equations [Eq. (1) to
different capacities of the tank, and they made a comparison Eq. (5)]:
between displacement and base shear of each supporting
system. 𝑔 1.84 ℎ
𝑘𝑐 = 𝑚 𝑐 1.84 tanh ( ) (1)
Shell Buckling mode is one of the most common forms of 𝑅 𝑅
damage in steel tanks generally classified as elastic buckling ‘’
diamond shape ‘’ and elasto-plastic buckling ‘’elephant foot’’, R 1.84 h
mc = me 0.318 tanh ( ) (2)
as shown in Figure 1. This instability appears usually around h R
the bottom of tanks. For this reason, many several theoretical
and experimental research studies were performed such as Liu cosh(1.84 h⁄R) − 1
and Lam, Nagashima et al., Virella et al. and Djermane et al. hc = [1 − ]h (3)
1.84 h⁄R sinh (1.84 h⁄R)
[19-22].
On the basis of previously-mentioned research, we can say
tanh (1.74 R⁄h)
that it is found that nearly all the published literature discussed mi = me (4)
the dynamic buckling of the tank and relationship between the (1.74 R⁄h)
dynamic behavior of the elevated water tank with the
supporting system. However, there were no studies for the 3
impact of the supporting system on the dynamic buckling of hi = h (5)
8
elevated water tanks.
The aim of this study is to investigate the effect of where, h, R, kc, mc, mi, me et hc are the fluid height, radius tank,
supporting system and the frequency content of the earthquake convective rigidity, convective mass, impulsive mass, total
on dynamic buckling of elevated water tank, which will lead mass and convective Hight.
to a clarification and understanding of the dynamic behavior Then, a more simplified approach was elaborated by
of elevated water tank under seismic excitation. This work is Malhotra et al. This spring-mass model is used in Eurocode 8
also motivated by the need for the enhancement of Algerian (2003) (Table 1) [11].
seismic code (RPA99/2003) which does not contain any
provisions for liquid storage tanks design yet.

334
Table 1. Recommended design values for the first impulsive 𝑰 = 𝜋 𝑅3 𝑒 (9)
and convective modes of vibration [11]
33
𝑃′ = 𝑃 + 𝑝𝑙 (10)
HL/R 𝐂𝒄 mi/mw mc/mw hi/HL hc/HL 140
0.3 2.09 0.176 0.824 0.400 0.521
0.5 1.74 0.300 0.700 0.400 0.543
0.7 1.60 0.414 0.586 0.401 0.571
1 1.52 0.548 0.452 0.419 0.616
1.5 1.48 0.686 0.314 0.439 0.690
2 1.48 0.763 0.237 0.448 0.751
2.5 1.48 0.810 0.190 0.452 0.794
3 1.48 0.842 0.158 0.453 0.825

The natural periods of the convective (Tc) responses is:

𝑇𝑐 = 𝐶𝑐 √R (6)

where, HL, R, Cc, mi, mc, hi, hc and mw are Total liquid height, Figure 3. Console (constant section)
radius tank. The coefficients of convective period, impulsive
mass, convective mass, impulsive height, convective height where, ki, R, E, I, l, P and p are rigidity of a tower, tower
and total liquid mass. radius, Young’s modulus, the inertia of the cross section, the
height of the tower, the weight of the concentrated mass and
linear weight of the tower respectively.

3. NUMERICAL MODEL

Figure 2. Model 2ddl [9]

h, R, kc, mc, mi, me and hc are the fluid height, radius tank,
convective rigidity, convective mass, impulsive mass, total
mass and convective height, as shown in Figure 2.
The rigidity of support can be calculated by using the finite
element method or according to:
The rigidity of a frame support can be given by [4-6]:

12 𝐸𝑐1 𝐼𝑐1 𝑁𝑐1

𝑘𝑖 = 3
ℎ𝑐1

1 (7)
𝐸𝑐1 𝐼𝑐1
2 𝐼𝑐1 𝑁𝑝 (4 𝑁𝑝2 − 1) ℎ
+ 𝑁𝑝 + 2(𝑁𝑝 − 1) 𝐸 𝑐1
𝐴𝑐 𝑅𝑠2 𝑏 𝐼𝑏
[ 𝐿 ]

where, Ecl, hcl, Icl and Ncl are Young’s modulus of the column
material, the height, the moment of inertia and the number of
columns, respectively. Eb, L and Ib are Young’s modulus of
the beam material, span and moment of inertia of the beam,
respectively. Np is the number of panels and Rs is the staging
radius.
By the method of Rayleigh, the rigidity of a tower with a
constant section (Figure 3) is given by [31]:

𝑃 3𝐸𝐼
𝑘𝑖 = (8)
𝑃′ 𝑙 3 Figure 4. (a) Shell 63, (b) Shell 181, (c) Beam 188

335
 In this study, two elevated tanks with different supporting P = pressure
systems are modeled by Ansys software using the finite γ = shear strain
element technique. The wall and roof are modeled in this study S = K×10−9 (arbitrarily small number to give element some
by using Shell63 for modal analysis and shell181 “plastic shear stability)
capability’’ for transient analysis. The two elements have six τ = shear stress
degrees of freedom at each node: translations in the nodal x, y, Ri = rotation about axis i
and z directions and rotations about the nodal x, y, and z-axes. B=K×10−9 (arbitrarily small number to give element some
For the shaft support the beam188 with 2-node and six degrees rotational stability)
of freedom at each node is used [32], as shown in Figure 4. Mi=twisting force about axis i
A damping matrix is also developed based on:
3.1 Fluid structure interaction
𝜀𝐵𝑢𝑙𝑘
̇ 0 0 0 0 0 0 0 𝑃
The effect of the fluid-structure interaction is taken into 𝛾𝑥𝑦̇ 0 1/𝜂 0 0 0 0 0 𝜏𝑥𝑦
account by properly coupling the nodes that lies in the 𝛾𝑦𝑧̇ 1/𝜂 0 0 0 0 𝜏𝑦𝑧
common faces of these two domains [33], as shown in Figure 0 0
𝛾𝑥𝑧̇ = 0 0 0 1/𝜂 0 0 0 = 𝜏𝑥𝑧
5.
𝑅𝑥̇ 0 0 0 0 1/𝐶 0 0 𝑀𝑥
𝑅𝑦̇ 0 0 0 0 0 1/𝐶 0 𝑀𝑦

{ 𝑅𝑧̇ } [0 0 0 0 0 0 1/𝐶] { 𝑀𝑧 }

where, η= viscosity and C = 0.00001η.

Figure 5. Fluid–structure models Figure 6. Fluid 80

3.2 Fluid domain 3.3 Free vibration analysis

FLUID80 is used to model fluids contained within vessels Modal analysis is used to determine the vibration
having no net flow rate (Figure 6). This fluid element is characteristics of this model. The important parameters in the
particularly suitable for calculating hydrostatic pressures and design of a structure under dynamic loading conditions are the
fluid/solid interactions, acceleration effects, such as in natural frequencies and mode shapes. The free system
sloshing problems. The element is defined by eight nodes vibration equation is given by:
having three degrees of freedom at each node: translation in
the nodal x, y, and z directions. The stress-strain relationships [M]{ü } + [K]{u} = 0 (11)
used to develop the stiffness matrix and thermal load vector
are as follows [32]: where, [M]=structural mass matrix, [K]=structural stiffness
matrix, {ü } = nodal acceleration vector, and {u} nodal
1 displacement vector. For a linear system, free vibration will be
0 0 0 0 0 0 expressed as:
𝐾
1
0 0 0 0 0 0 u = ϕi coswi t (12)
𝜀𝐵𝑢𝑙𝑘 𝑆 𝑃
1
𝛾𝑥𝑦 0 0 0 0 𝜏𝑥𝑦 where, ϕi : eigenvector representing the mode shape of the ith
𝛾𝑦𝑧 𝑆 𝜏𝑦𝑧 natural frequency, wi: ith natural circular frequency in radians
𝛾𝑥𝑧 = 1
0 0 0 0 0 0 = 𝜏𝑥𝑧 per unit time, t = time in seconds. Substitution of Eq. (12) in
𝑅𝑥 𝑆 𝑀𝑥 Eq. (11) gives [32]:
0 0 1
𝑅𝑦 0 0 0 0 0 0 𝑀𝑦
{ 𝑅𝑧 } 0 0 𝐵 { 𝑀𝑧 } 〈[K] − [M]〉ϕi = 0 (13)
1
0 0 0 0 0 0
𝐵 3.4 Transient analysis
1
0 0 0 0
[ 𝐵] For reasons of large displacements and relatively large
deformations, the material and geometrical nonlinearities are
where, considered. Plasticity was included using the software offers
∂u ∂v ∂w several options, among which we selected the simplest
εBulk = + + = Bulk strain
∂x ∂y ∂z consisting of a bilinear kinematic hardening curve, taking into
K = fluid elastic (bulk) modulus account the Beauschinger effect.

336
 The transient dynamic analysis solves the basic equation: 4. FREE VIBRATION ANALYSIS RESULT

of motion: In this section, free vibration analysis is carried out on the

two elevated tanks to validate the proposed numerical model
[M]{ü } + [C]{u̇ } + [K]{u} = {F(t)} (14) using the proposed FE technique (Figure 9). The values for the
convective and impulsive responses were obtained and
where, [C] = damping matrix, {u̇ } =nodal velocity vector and compared with the current practice values (which is based on
{F(t)} = load vector [16]. the Housner’s formula). The results of modal analysis are
summarized in Table 2 and 3, and Figures 10-11.
3.5 Stability criteria

The most used stability criteria such as that of Budiansky-

Ruth, the phase plane of energy is very time consuming
(Figure 7-8). In this work, a prediction of the critical level is
obtained by the Pseudo-Dynamic criterion. This prediction is
then checked and improved using the conventional criteria
[34-36].

Figure 9. Elevated tank geometries

(a) Criteria of Budiansky and Roth

(b) Criteria of Ari Gur and Simonetta Figure 10. Mode participation mass ratio (Type 1)
Figure 7. Critical load Table 2. Frequency and effective mass fraction (Type 1)

Finite elements Eurocode 8

Effective
Type Order Frequency mass Frequency
fraction
1* 0.156 0.41 0.155
Convective 2 0.286 0.013 /
3 0.35 0.0035 /
1* 0.77 0.51 0.95
Impulsive 2 6.01 0.035 /
3 16.57 0.011 /

Figure 8. Phase plane diagram before and after the Pcr Figure 11. Mode participation mass ratio (Type 2)

337
 Table 3. Frequency and effective mass fraction (Type 2) From Tables 2 and 3, we can realize that the convective
period remains the same for the two types of the supporting
Finite elements Eurocode 8 systems, which means that the convective component is
Effective independent of the support type. On the contrary, a significant
Type Order Frequency mass Frequency difference is observed between the impulsive periods of the
fraction two types of elevated tanks.
1* 0.156 0.54 0.157
Convective 2 0.29 0.016 /
3 0.35 0.00013 /
1* 2.87 0.37 3.5
5. RESULTS OF DYNAMIC BUCKLING
Impulsive 2 9.99 0.033 /
3 15.37 0.027 / The Budiansky and Roth criterion, which was used in the
literature to determine the dynamic buckling load, is employed
The obtained mode shows that the shapes of fundamental in this study. Different analyses of the two elevated tanks for
mode involve sloshing of the contained liquid without any rising PGA (peek ground acceleration) levels are done, and
participation of the shell walls (given by Veletsos) [27], as the two types of support are subjected to the horizontal
shown in Figures 12-13. The fundamental impulsive mode is excitations of the San Fernando 1971 and El Centro1940
a column mode type. It can be observed that the calculated FE earthquakes (Figure 14).
results are in reasonable agreement with current practice
values. The two fundamentals (impulsive and convective)
modes can achieve modal mass participation for more than
90% of the total mass of the system which means that the two
modes are the most dominant. The tank undergoes one of the
large deformations. This is due to higher stiffness of the
support in comparison with the tank part [9]. These results
indicate the validity of the proposed FE model.

Figure 14. Accelerograms (a) San Fernando, PGA =12.02 m/

s²,(b) El centro PGA =3.41 m/ s²San Fernando 1971
earthquakes

5.1 San Fernando earthquake

Shows the pseudo equilibrium path for the two type of

elevated tanks. The discontinuity on curve indicates that the
(PGAcr) occurs at 1.5 for type 2 and 1.2 g for type 1. It shows
several history curves corresponding to different levels of
excitation. Figures 15-16 show clearly the difficulty of using
Figure 12. Modal shape (Type 1) the Budiansky-Ruth criterion for determining the (PGA)cr
which requires, in fact, a lot of experience and attention. At
level 1.2 and 1.5, a disproportionate increase in displacements
is distinguished.

Figure 15. Pseudo-Equilibrium path for elevated tank type-2-

Figure 13. Modal shape (Type2)

338
 1.5 This increase does not correspond to a monotonic jump for
the above-mentioned reasons. The phase planes criterion
illustrated in Figures 19 and 20 shows more easily in this case
the instability in the vicinity of the (PGA). The difficulty for
1
using this criterion is, in some cases, the same as that reported
PGAcr (g)

for Budiansky and Ruth, but the use of the two criteria
simultaneously can be more illustrative.
0.5

0
0 0.5 1 1.5

Displacement (m)

Figure 16. Pseudo-equilibrium path for elevated tank type-1-

Figures 17-18, show several history curves corresponding

to different levels of excitation. These Figures show clearly the
difficulty of using the Budiansky-Ruth criterion for
determining the (PGA)cr which requires, in fact, a lot of
experience and attention.
Figure 19. Phase plane before and after PGAcr for
elevated tank type-2-

Figure 17. Time history curves before and after PGAcr for Figure 20. Phase plane before and after PGAcr for elevated
elevated tank type-2- tank type-1-

Figure 21. Dynamic buckling of elevated tanks under San

Fernando earthquake

5.2 El Centro earthquake

Figures 22-27 give the PGAcr for this excitation. Using an

estimation given by the pseudo dynamic path, the Budiansky–
Figure 18. Time history curves before PGAcr for elevated Ruth and phase plane criteria are then used to confirm the
tank type-1- obtained value.

339
 2

1.5
PGAcr (g)

0.5

0
0 0.5 1 1.5

Displacement (m)

Figure 22. Pseudo-equilibrium path for the elevated tank.

Type -1-
Figure 25. Time history curves before PGAcr for elevated
1.2
tank type-2-
1

0.8
PGAcr (g)

0.6

0.4

0.2

0
0 0.5 1 1.5 2

Displacement (m)

Figure 23. Pseudo-equilibrium path for the elevated Figure 26. Phase plane before PGAcr for elevated tank type-
tankType -2- 1-

Using an estimation given by the pseudo dynamic path, the

Budiansky–Ruth and phase plane criteria are then used to
confirm the obtained value

Figure 27. Phase plane before PGAcr for elevated tank type-
2-

5.3 Discussion of the results

The comparison between the pseudo dynamic path curves

for the two elevated tanks clearly shows the big difference
between the displacements that corresponds to each PGA for
Figure 24. Time history curves before PGAcr for elevated the two elevated tanks. Comparing the PGAcrt obtained for
tank type-1- two models, it can be seen that the PGAcr of the elevated tank
with tower support is 1.5 g m / s2 and 1.2g m / s2 for a raised

340
tank with frame support; with 20% difference for San The transient study of the elevated tanks under the
Fernando earthquake. 1.5345 g m / s2 for elevated tank with a earthquakes of San Fernando1971 and El Centro 1940 shows
frame support and 0.93775 g m / s2 for elevated tank with a that:
tower support; and 38.88% difference for El Centro Comparing the PGAcrt obtained for two models, it can be
earthquake. seen that the type of support has a great influence on the
The curve of the type 1 model indicates that the critical dynamic behavior of the elevated tanks.
value PGA (PGAcr) occurs at 1.5345 g (El Centro excitation), The PGAcr of the structure does not only depend on the
and it is increased by 21.79% compared to the San Fernando characteristics of the structure, but also the frequency content
earthquake result (Figure 28). and characteristics of the earthquake.
In the curve of the type 2- model, the PGAcr occurs at level Most of the deformations in the support tank interface
0.93775 g (El Centro excitation), and it is decreases by 37.48% resulted from the sudden change of rigidity and geometry of
compared to the San Fernando earthquake result (Figure 28). the interface (higher stiffness in the support of the tank).
From Figure 21, it is concluded that the maximum stresses The criterion of dynamic buckling response is studied
are located along the support-tank interface, and this is due to through testing the deformations of the excitation critical level.
the change of rigidity and geometry at this interface. The results of the deformations of tanks indicated a high level
of explosiveness in plastic buckling type.
Finally, based on the comparison made between two
elevated tanks results, it can be concluded that the highest
value of PGAcr does not always occur in the more rigid
supporting system, and PGAcr does depend on characteristics
of the earthquake excitation These results indicate the need
for consideration of the support rigidity and frequency content
of earthquake excitation design in current practice.

ACKNOWLEDGMENT

This work was supported by FIMAS Laboratory, University

of Tahri Mohammed Bechar, Algeria.

Figure 28. Comparison between the values of PGAcr REFERENCES

[1] Hadj-Djelloul, N.D., Djermane, M. (2020). Effect of

6. CONCLUSION geometric imperfection on the dynamic of elevated water
tanks. Civil Engineering Journal, 6(1): 85-97.
In high seismicity areas and high population density, a tank https://doi.org/10.28991/cej-2020-03091455
rupture following an earthquake can cause enormous damage [2] Housner, G.W. (1963). The dynamic behavior of water
and loss of property and human life. In addition, the use of tanks. Bulletin of the seismological society of America,
tanks in the nuclear industry and the storage of chemical or 53(2): 381-387.
toxic materials make this type of structure very sensitive to [3] Joshi, S. (2000). Equivalent mechanical model for
their safety with regard to seismic effect. In the present work, horizontal vibration of rigid Intze tanks. ISET Journal of
an efficient 3D finite element method analysis was used to Earthquake Technology, 37(1-3): 39-47.
analysis the dynamic behavior of two elevated tanks, [4] Dutta, S.C., Jain, S.K., Murty, C.V.R. (2000). Assessing
considering nonlinear temporal analysis, wall flexibility, the the seismic torsional vulnerability of elevated tanks with
material and geometric non-linearities, and fluid-structure RC frame-type staging. Soil Dynamics and Earthquake
interactions, using ANSYS software. First, a modal analysis Engineering, 19(3): 183-197.
was carried out and compared with the Eurocode Code 8 to https://doi.org/10.1016/S0267-7261(00)00003-8
confirm the numerical models, and to view the effect of the [5] Dutta, S.C., Jain, S.K., Murty, C.V.R. (2000). Alternate
support system on the frequencies of the two structures. tank staging configurations with reduced torsional
Secondly, the dynamic buckling analysis of the two elevated vulnerability. Soil Dynamics and Earthquake
tanks was performed to review also the effect of the support Engineering, 19(3): 199-215.
system on the dynamic behavior of these two elevated tanks, https://doi.org/10.1016/S0267-7261(00)00004-X
the obtained results showed that: [6] Dutta, S.C., Jain, S.K., Murty, C.V.R. (2001). Inelastic
The frequency of the convective component is independent seismic torsional behavior of elevated tanks. Journal of
of the rigidity of the structure, and the frequency of the Sound and Vibration, 242(1): 151-167.
impulsive component depends on the rigidity of the structure https://doi.org/10.1006/jsvi.2000.3343
which confirms Housner’s theory. The impulsive mode [7] Shrimali, M.K., Jangid, R.S. (2003). Earthquake
frequency of elevated tank with tower support is higher than response of isolated elevated liquid storage steel tanks.
the raised tank with frame type which means that tower type is Journal of Constructional Steel Research, 59(10): 1267-
more rigid than frame type. The two fundamental modes 1288. https://doi.org/10.1016/S0143-974X(03)00066-X
(impulsive and convective) reach modal mass participation of [8] Livaoğlu, R., Doğangün, A. (2006). Simplified seismic
more than 90% of the total mass of the system, which means analysis procedures for elevated tanks considering fluid–
that the two fundamental modes are dominant. structure–soil interaction. Journal of Fluids and

341
 Structures, 22(3): 421-439. earthquake excitation. Journal of Constructional Steel
https://doi.org/10.1016/j.jfluidstructs.2005.12.004 Research, 62(6): 521-531.
[9] Moslemi, M., Ghaemmaghami, A.R., Kianoush, M.R. https://doi.org/10.1016/j.jcsr.2005.10.001
(2016). Parametric based study for design of liquid-filled [22] Djermane, M., Zaoui, D., Labbaci, B., Hammadi, F.
elevated tanks. Canadian Journal of Civil Engineering, (2014). Dynamic buckling of steel tanks under seismic
43(7): 619-630. https://doi.org/10.1139/cjce-2015-0218 excitation: Numerical evaluation of code provisions.
[10] Allen, L.R., Hutchinson, G.L., Stevens, L.K. (1990). Engineering Structures, 70: 181-196.
Buckling considerations in the design of elevated steel https://doi.org/10.1016/j.engstruct.2014.03.037
water tanks. Thin-Walled Structures, 9(1-4): 389-406. [23] Soroushnia, S., Tafreshi, S.T., Omidinasab, F.,
https://doi.org/10.1016/0263-8231(90)90054-3 Beheshtian, N., Soroushnia, S. (2011). Seismic
[11] Malhotra, P.K., Wenk, T., Wieland, M. (2000). Simple performance of RC elevated water tanks with frame
procedure for seismic analysis of liquid-storage tanks. staging and exhibition damage pattern. Procedia
Structural Engineering International, 10(3): 197-201. Engineering, 14: 3076-3087.
https://doi.org/10.2749/101686600780481509 https://doi.org/10.1016/j.proeng.2011.07.387
[12] Algreane, G.A., Osman, S.A., Karim, O.K., Kasa, A. [24] Steinbrugge, K.V. (1970). Earthquake damage and
(2011). Behavior of elevated concrete water tank structural performance in the United States. pp. 183-226.
subjected to artificial ground motion. EJGE, 16: 387-406. [25] European Committee for Standardization (CEN). (2006).
[13] Omidinasab, F., Shakib, H. (2012). Seismic response Eurocode 8: Design of structures for earthquake
evaluation of the RC elevated water tank with fluid- resistance—Part 4: Silos, tanks and pipelines.
structure interaction and earthquake ensemble. KSCE J [26] Veletsos, A.S., Yang, J.Y. (1977). Earthquake response
Civ Eng, 16: 366-376. https://doi.org/10.1007/s12205- of liquid storage thanks. Proc. 2nd Adv. Civil. Eng.
011-1104-1 through Eng. Mech. Conf., ASCE, North Carolina, pp. 1-
[14] Sweedan, A.M.I. (2009). Equivalent mechanical model 24.
for seismic forces in combined tanks subjected to vertical [27] Veletsos, A. S. (1984). Seismic response and design of
earthquake excitation. Journal of Thin Walled Structures, liquid storage tanks. Guidelines for the Seismic Design
47: 942-952. https://doi.org/10.1016/j.tws.2009.02.001 of Oil and Gas Pipeline Systems, pp. 255-370.
[15] Chaduvula, U., Patel, D., Gopalakrishnan, N. (2013). [28] Haroun, M.A., Housner, G.W. (1981). Earthquake
Fluid-structure-soil interaction effects on seismic response of deformable liquid storage tanks. Journal of
behaviour of elevated water tanks. Procedia Engineering, Applied Mechanics, 48: 411-418.
51: 84-91. https://doi.org/10.1016/j.proeng.2013.01.014 https://doi.org/10.1115/1.3157631
[16] Jabar, A.M., Patel, H.S. (2012). Seismic Behavior of RC [29] ACI Committee 350. (2006). Seismic Design of Liquid-
Elevated Water Tank under different staging pattern and containing Concrete Structures and Commentary (ACI
earthquake characteristics. International Journal of 350.3-06): An ACI Standard. American Concrete
Advanced Engineering Research and Studies, IJAERS, 1. Institute.
[17] Mor Vyankatesh, K., More Varsha, T. (2017). [30] American Water Works Association (AWWA), 1984 and
Comparative Study on Dynamic Analysis of Elevated 2005, Welded steel tanks for water storage. AWWA
Water Tank Frame Staging and Concrete Shaft D100, CO.
Supported. IOSR Journal of Mechanical and Civil [31] Capra, A., Davidovici, V. (1980). Calcul dynamique des
Engineering (IOSR-JMCE), 14(1): 38-46. structures en zone sismique (No. BOOK). Eyrolles.
https://doi.org/10.9790/1684-1401013846 [32] The ANSYS Structural Software System. ANSYS INC,
[18] Pole, S.M., Khedikar, A. (2017). Seismic behaviour of Vol. 12.
elevated water tank under different staging pattern. [33] Cheng, X., Jing, W., Yin, C., Li, C. (2018). Stability
International Journal of Engineering and Management parameter analysis of a composite foundation of an oil
Research (IJEMR), 7(3): 199-205. storage tank in a loess area treated with compaction piles.
[19] Liu, W.K., Lam, D. (1983). Nonlinear analysis of liquid- Soils and Foundations, 58(2): 306-318.
filled tank. Journal of Engineering Mechanics, 109(6): https://doi.org/10.1016/j.sandf.2018.02.004
1344-1357. https://doi.org/10.1061/(ASCE)0733- [34] Budiansky, B. (1962). Axisymmetric dynamic buckling
9399(1983)109:6(1344) of clamped shallow spherical shells. NASA TN, 1510:
[20] Nagashima, H., Kokubo, K., Takayanagi, M., Saitoh, K., 597-606.
Imaoka, T. (1987). Experimental study on the dynamic [35] Djermane, M., Chelghoum, A. (2007). Nonlinear
buckling of cylindrical tanks: Comparison between static dynamic and dynamic buckling of thin shells using a
buckling and dynamic buckling: Solid-mechanics, drilling finite element. International Journal of Advanced
strength of materials. JSME International Journal, Manufacturing Technology, 1(2): 175-84.
30(263): 737-746. [36] Ari-Gur, J., Simonetta, S.R. (1997). Dynamic pulse
https://doi.org/10.1299/jsme1987.30.737 buckling of rectangular composite plates. Composites
[21] Virella, J.C., Godoy, L.A., Suárez, L.E. (2006). Dynamic Part B: Engineering, 28(3): 301-308.
buckling of anchored steel tanks subjected to horizontal https://doi.org/10.1016/S1359-8368(96)00028-5

342

View publication stats