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

net/publication/338234187

Seismic performance assessment of FPS isolated liquid storage tanks at

various intensity levels

Conference Paper · September 2017

CITATIONS READS

3 177

2 authors, including:

Hossein Hayati Raad

University of Tabriz
5 PUBLICATIONS 27 CITATIONS

SEE PROFILE

All content following this page was uploaded by Hossein Hayati Raad on 30 December 2019.

The user has requested enhancement of the downloaded file.

Seismic performance assessment of FPS isolated liquid storage
tanks at various intensity levels

Saman Bagheri1, Hossein Hayati Raad2

1
Associate Professor, Faculty of Civil Eng., University of Tabriz, Tabriz, Iran
2
Graduate Student, , Faculty of Civil Eng., University of Tabriz, Tabriz, Iran

ABSTRACT: On-grade steel tanks are widely used to store liquids, such as oil or other
petroleum products. The inappropriate behavior of these industrial facilities which frequently
resulted in serious structural or nonstructural damages, have been reported in some recent
earthquakes. Seismic base isolation is one of the most efficient techniques to mitigate
earthquake damage in both new and existing liquid storage tanks. This paper deals with the
seismic performance and damage assessment of on-grade liquid storage tanks isolated by the
friction pendulum system (FPS). As a sliding type isolation system, FPS introduces nonlinearity
into the dynamic analysis of the model, leading to likely different performances at various levels
of input excitation. To estimate the seismic demands for different levels of intensity,
incremental dynamic analysis (IDA) under a suit of strong ground motion records is carried out.
The effects of tank aspect ratio, isolation period and friction coefficient on the seismic response
of isolated tanks are investigated and the obtained results are compared with those from fixed
ones.
KEYWORDS: Storage tank, earthquake, base isolation, incremental dynamic analysis.

1 INTRODUCTION
Liquid storage tanks are one of the most important industrial structures that their dynamic
behavior is different from other structures such as buildings and bridges because of the
hydrodynamic forces of the contained liquid. The observations of some typical damages to on-
grade steel tanks during past earthquakes have repeatedly demonstrated their inappropriate
performances (Cooper, 1997; Hamdan, 2000). Some techniques to reduce the damages in these
structures have been examined by several researchers. One of these techniques is to use base
isolators. As a first attempt to study this subject experimentally, Chalhoub and Kelly (1990)
conducted shake table tests of fixed-base and base isolated cylindrical steel tanks. They
observed that as a result of isolation, the dynamic pressures were reduced considerably with a
slight increase in sloshing motion of the liquid.
Malhotra (1997) proposed a new method for seismic base isolation of on-grade cylindrical
liquid storage tanks by disconnecting the wall of the tank from the base plate and showed that
the overturning moments and axial compressive stresses in the tank wall could be reduced
significantly. The feasibility of using FPS bearings for seismic base isolation of rigid cylindrical
storage tanks was explored by Wang et al. (2001). The effectiveness of this sliding-type
isolation system was verified through numerical simulations under the 1940 El Centro
earthquake. A comparative study of performance of various isolation systems for on-grade
storage tanks by Shrimali and Jangid (2002) also confirmed that the sliding-type isolation
systems were more effective in controlling the seismic response of the tanks in comparison with
the elastomeric bearings. Moreover, the seismic behavior of elevated liquid storage tanks
isolated by high damping rubber bearings and friction pendulum isolators was analyzed and
compared by Paolacci (2015). Bagheri and Farajian (2016) investigated the nonlinear dynamic
behavior of FPS isolated liquid storage tanks under ground motion with different characteristics
and observed that the FPS was more effective in reducing the seismic responses under far-field
ground motions compared to near-fault ground motions.
As a sliding-type isolation system with velocity-dependent friction law, FPS introduces
nonlinearity into the dynamics of the system which can result in different performances at
various levels of input excitation. To account for this, incremental dynamic analysis (IDA), that
was initially developed for frame structures (Vamvatsikos and Cornell, 2002), can be extended
to our case. In this procedure, a series of nonlinear dynamic analyses of a structural model under
an ensemble of ground motion records, each scaled to several levels of intensity, is performed.
In the present work, the seismic responses of the FPS isolated liquid storage tanks are obtained
from the IDA procedure and compared with those of fixed-base ones at various levels of input
intensity. Also, the influences of tank aspect ratio, isolation period and friction coefficient on
the performance of the isolated tanks are studied at various intensity levels.

2 MODELING

2.1 Mathematical model of the isolated liquid storage tank

Housner (1963) proposed a simple approximate mechanical model to estimate the dynamic
effects of liquid in a ground supported tank with rigid wall and incompressible fluid under the
horizontal seismic excitation. The model was then improved by Haroun and Housner (1981) to
take into account the flexibility of the tank wall. In such mechanical models the liquid mass is
divided into two parts: the impulsive component that moves in unison with the tank wall near
the base of the tank, and the convective component that experiences sloshing motion near the
free-surface. Later on, Malhotra et al. (2000) proposed a simple, yet accurate, and more
generally applicable equivalent mechanical model of the tank-liquid system by combining the
higher impulsive modal mass with the first impulsive mode and the higher convective modal
mass with the first convective mode.
The mathematical model used in the present study is the latter which is resting on a base
isolation system as shown in Figure 1. The geometrical parameters of the tank are the liquid
height (H), radius of the tank (R), and the equivalent uniform thickness of the tank wall (t). The
convective and impulsive masses (mc and mi) are connected to the tank wall by springs having
stiffnesses of kc and ki, respectively. The damping coefficients of the convective and impulsive
masses are denoted as cc and ci, respectively. The calculation method of these parameters has
been described in detail by Malhotra et al. (2000).

2.2 Governing equations of motion

A three-degree of freedom system of a base isolated liquid storage tank is shown in Figure 2.
The equations of motion can be written as:
Figure 1. Mathematical model of a base isolated liquid storage tank used in this study

Figure 2. Three-degree of freedom system of a base isolated liquid storage tank

mc u&&c + cc (u& c − u&b ) + k c (u c − u b ) = −mc u&&g (1)

mi u&&i + ci (u& i − u&b ) + k i (u i − u b ) = −mi u&&g (2)

mbu&&b − kc (uc − ub ) − ki (ui − ub ) − cc (u&c − u&b ) − ci (u&i − u&b ) + F = −mbu&&g (3)

where uc, ui, and ub are displacements relative to the ground for the convective, impulsive, and
base masses respectively; üg is the earthquake ground acceleration; mb is the base mass; and F is
the horizontal force exerted by FPS. This force and the isolation period, Tb, are given as:
W
F= u b + µW sgn (u& b ) (4)
R
R
Tb = 2π (5)
g
where W is the vertical load on the isolator produced by the total weight of the system, i.e.
(mc+mi+mb)g, R is the radius of curvature of the sliding surface, and µ is the velocity-dependent
coefficient of friction governed by Eq. (6). In the aforementioned force of Eq. (4), the first term
is the linear elastic spring force with its stiffness based on the curvature of the spherical dish and
the second term is the friction force. In the latter term a continuous hysteretic variable ranging
between -1 and 1 governed by a modified Bouc-Wen model can be replaced the sign function to
avoid difficulties in the numerical solution of the equations (Fenz and Constantinou, 2008). The
velocity-dependent coefficient of friction is given as:
µ = µ max − ( µ max − µ min ) exp(−a u&b ) (6)

where ůb is the sliding velocity, µmax and µmin are the sliding coefficients of friction at large
velocity and nearly zero sliding velocity, respectively, and a is a rate parameter that controls the
variation of friction coefficient. a=100 s/m is often used for the type of material commonly used
in FPS bearings (Fenz and Constantinou, 2008).
The governing equations of motion are solved using the state-space method by a provided
MATLAB routine. The numerical results will be presented in terms of the overturning moment
(M) and vertical displacement of the liquid surface due to the sloshing motion (dv) according to
Eqs. (7) and (8), respectively (Malhotra, 1997). The overturning moment directly indicate the
seismic demand on the tank, which may lead to the buckling of the tank wall, possible damage
to the bottom plate, and rupture of connections. On the other hand, the free surface displacement
reflects the need for a sufficient freeboard and can be considered in relation to damage of the
roof or top wall and spill of the contents over the tank.
M = −{mc hc (u&&c + u&&g ) + mi hi (u&&i + u&&g )} (7)

ω c2 (u c − u b )
d v = 0.837 R (8)
g

2.3 Tank properties

Two cylindrical steel tanks with different aspect ratios, one broad and one slender, have been
used for numerical studies. The main properties of these tanks are summarized in Table 1, in
which ρ indicates the mass density of the contained liquid (water) and E is the modulus of
elasticity of the tank material (steel). The damping ratios for convective and impulsive masses
are taken as 0.5% and 2%, respectively, as suggested previously for on-grade steel storage tanks
(Haroun and Housner, 1981; Malhotra, 1997; Malhotra et al, 2000).

Table 1. Properties of the considered broad and slender tanks

Tank type H (m) R (m) H/R t (m) E (GPa) ρ (kg/m3)
Broad 14.6 24.4 0.6 0.0203 200 1000
Slender 11.3 6.1 1.85 0.0058 200 1000

3 SEISMIC ASSESSMENT PROCEDURE

The twenty-two ground motion records in the far-field record set of FEMA P695 (2009) are
adopted herein for the incremental dynamic analysis (IDA). A series of nonlinear dynamic
analyses of the tank models under the selected ground motion records is performed by scaling
each record to several levels of intensity. In this study, peak ground acceleration (PGA) is
chosen as the earthquake intensity measure in the IDA procedure.
In order to assess the performance of the seismic isolation system at various levels of intensity,
a non-dimensional performance index is used according to Eq. (9) for each response parameter.
Positive values of the performance index mean reduction in the seismic response due to the
isolation while negative values indicate the increase in the response.
fixed base tank response − isolated tank response
Performance Index = (9)
fixed base tank response

4 RESULTS AND DISCUSSION

4.1 Effect of base isolation

The seismic performance of the isolation system at various intensities for the broad and slender
tanks with µmax = 0.06, µmin = 0.03, and Tb = 2 s is shown in Figure 3. Results illustrated herein
are median values of the selected earthquake ground motions.
It is observed that due to the isolation the overturning moment is considerably reduced in both
the broad and slender tanks. As the earthquake intensity increases, performance index for the
overturning moment increases until it reaches an almost constant value of about 80% for the
broad tank or a little more for the slender tank. This happens when the peak ground acceleration
as the seismic intensity measure reaches about 0.3 g.
It is also seen that unlike overturning moment, the vertical displacement of the liquid surface not
only does not decrease significantly by isolation, but also increases slightly at some levels of
earthquake intensity. Indeed, because of the long period nature of sloshing motions, aseismic
base isolation, which increases the period of the system, does not have much effect on this
response parameter.

Broad tank Slender tank

dv

Figure 3. Seismic performance of the isolated tanks at various intensity levels

4.2 Effect of isolation period
In order to investigate the effect of isolation period on the seismic performance of the isolated
tanks at various intensity levels, seven different isolation periods (Tb = 1, 1.5, 2, 2.5, 3, 4, 5 s)
are considered for both broad and slender tanks. The friction coefficients are the same as in the
previous subsection. The numerical results in the form of median IDA curves are shown in
Figure 4.
The results indicate that reductions in the overturning moment due to base isolation of the tanks
are almost identical for all assumed isolators until the PGA reaches about 0.2 g. After this stage,
increase of the isolation period can increase the seismic performance of the system, but for
isolation periods of more than about 3 s, the effect of this parameter is not significant. The
figure also shows that different isolators can lead to a slight increase or decrease in the sloshing
displacement at various intensity levels except those with high periods that decrease
considerably this response parameter, especially at the high intensity levels.

Broad tank Slender tank

dv

Figure 4. Effect of isolation period on the seismic performance of the isolated tanks at various intensity
levels

4.3 Effect of friction coefficient

To investigate the effect of friction coefficient on the seismic performance of the isolated tanks
at various intensity levels, µmax has been varied from 0 to 0.12 and the value of µmin has been
assigned half of µmax. The isolation period is the same as in the first subsection of this section.
The obtained numerical results for the broad and slender tanks at various intensity levels are
shown in Figure 5.
As is evident, under the frictionless condition, i.e. µ = 0, the nonlinear behavior of the isolation
system vanishes and the constant performance results for all input intensities. This condition has
a considerable negative impact on the performance of FPS in reducing sloshing motions, but the
effectiveness of the isolation system in reducing overturning moments remains as well. As the
friction coefficient increases, the overturning moment increases in both broad and slender tanks
in low intensity levels, which results in low performance of the isolator in reducing this
response parameter. However, at high intensity levels, reductions in the overturning moment
due to base isolation are almost identical for all assumed nonzero friction coefficients. On the
other hand, in terms of vertical displacement of liquid surface, an increase in the friction
coefficient can decrease and even eliminate the negative effect of isolation system, especially at
high intensity levels.

Broad tank Slender tank

dv

Figure 5. Effect of friction coefficient on the seismic performance of the isolated tanks at various intensity
levels

5 CONCLUSIONS
This research was focused on the nonlinear behavior of FPS isolated storage tanks at various
intensity levels of input ground motions. Following the estimation of seismic demands through
incremental dynamic analysis, seismic performance assessment of isolated broad and slender
tanks was performed at various intensity levels. It was observed that the Friction Pendulum
System could be a quite effective isolation system in order to reduce the overturning moment as
an indicator of structural damage in different levels of input excitation. However, the vertical
displacement of the liquid surface due to the sloshing motion not only does not decrease
significantly by the isolation system, but also increases at some levels of earthquake intensity. It
was also seen that the isolation system had better performance in reducing overturning moments
at high intensity levels. A number of different isolation periods and friction coefficients were
finally examined at various intensity levels, and the associated effects on the seismic
performance were discussed.

6 REFERENCES
Bagheri, S. and Farajian, M. 2016. The effects of input earthquake characteristics on the nonlinear
dynamic behavior of FPS isolated liquid storage tanks, Journal of Vibration and Control, doi:
10.1177/1077546316655914 (published online before print).
Chalhoub, MS, and Kelly, JM. 1990. Shake table test of cylindrical water tanks in base-isolated
structures, Journal of Engineering Mechanics, 116: 1451-1472.
 Cooper, TW. 1997. A study of the performance of petroleum storage tanks during earthquakes 1933-
1995, Report for US Department of Commerce: NIST GCR 97-720.
FEMA P695. 2009. Quantification of building seismic performance factors, Federal Emergency
Management Agency, Washington DC.
Fenz, DM, and Constantinou, MC. 2008. Development, implementation and verification of dynamic
analysis models for multi-spherical sliding bearings, Technical Report MCEER-08-0018,
Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo, State University
of New York.
Hamdan, FH. 2000. Seismic behaviour of cylindrical steel liquid storage tanks, Journal of Constructional
Steel Research, 53: 307-333.
Haroun, MA, and Housner, GW. 1981. Seismic design of liquid storage tanks, Journal of the Technical
Councils of ASCE, 107: 191-207.
Housner, GW. 1963. The dynamic behavior of water tanks, Bulletin of the Seismological Society of
America, 53: 381-387.
Malhotra, PK. 1997. New method for seismic isolation of liquid storage tanks, Earthquake Engineering
and Structural Dynamics, 26(8): 839-847.
Malhotra, PK, Wenk, T, and Wieland, M. 2000. Simple procedure for seismic analysis of liquid-storage
tanks, Structural Engineering International, 10: 197-201.
Paolacci, F. 2015. On the effectiveness of two isolation systems for the seismic protection of elevated
tanks, Journal of Pressure Vessel Technology, 137(3): 031801/1-8.
Shrimali, MK, and Jangid, RS. 2002. A comparative study of performance of various isolation systems
for liquid storage tanks, International Journal of Structural Stability and Dynamics, 2(4): 573-591.
Vamvatsikos, D, and Cornell, CA. 2002. Incremental dynamic analysis, Earthquake Engineering and
Structural Dynamics, 31(3): 491-514.
Wang, YP, Teng, MC, and Chung, KW. 2001. Seismic isolation of rigid cylindrical tanks using friction
pendulum bearings. Earthquake Engineering and Structural Dynamics, 30(7): 1083-1099.

View publication stats