Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 00 (2017) 000000
www.elsevier.com/locate/procedia

X International Conference on Structural Dynamics, EURODYN 2017

Influence of base plate bending stiffness on the seismic performance

of liquid storage tanks
Diego Hernandez-Hernandez *, Tam Larkin and Nawawi Chouw
Department of Civil and Environmental Engineering, The University of Auckland, Auckland Mail Centre, Private Bag, 92019, Auckland 1142,
New Zealand

Abstract

In the past, a number of investigations have been performed on the seismic behaviour of storage tanks. At the University of
Auckland, a number of shake table experiments on tanks with different aspect ratios have been recently performed. The earthquake
ground excitation was simulated based on a design spectrum from the New Zealand code NZS 1170.5, classification D for soft
soils. The results show that the base plate stiffness of storage tanks may play a significant role in the magnitude of hoop stresses
caused by the passage of the earthquake. To incorporate a low-damage seismic design philosophy, the tank is allowed to separate
partially at the base. This temporary separation initiates rigid body like motion of the tank and thus reduce the wall local
deformations. Consequently, the stress development along the wall can be reduced from that occurring with full base plate contact.
This research focuses on the effect of the base plate thickness on the hoop stress development along the height of the tank wall for
five radial directions.

2017 The Authors. Published by Elsevier Ltd.

Peer-review under responsibility of the organizing committee of EURODYN 2017.

Keywords: Storage tanks; base plate stiffness; seismic analysis; hoop stress.

1. Introduction

Liquid storage tanks are lifeline structures used in industrial and urban areas. Past experiences of seismically induced
damage of storage tanks have demonstrated their susceptibility to earthquakes. The damage can manifest as elephant
footing, rupturing of pipeline connections, overturning of the tank or simple collapse. In the case of failure during a
strong earthquake, they may produce environmental damage, explosions and, in an extreme case, they could pose a

*
Corresponding author. Tel.:+64 027-394-6841;
E-mail address: dher840@aucklanduni.ac.nz

1877-7058 2017 The Authors. Published by Elsevier Ltd.

Peer-review under responsibility of the organizing committee of EURODYN 2017.
2 Diego Hernandez-Hernandez et al./ Procedia Engineering 00 (2017) 000000

threat to human life. Post-earthquake recovery necessitates a reliable water supply for fire-fighting and drinking water.
For all these reasons tanks are a critical lifeline facility for communities recovering from the effects of strong ground
motion.

One of the principal causes of failures of storage tanks is the high stresses developed in the tank wall. The partial
separation of the base from the foundation, known as uplift, significantly affects the magnitude and distribution of
those stresses and, for unanchored tanks, uplift is the phenomenon that governs seismic design [1]. Theoretical and
experimental investigations have been made to predict and improve the seismic performance of liquid storage tanks.
Haroun and Badawi [2] and Haroun and Al-Zeiny [3] reported that uplifting involves inherent nonlinearities due to
the plastic hinges developed in the base plate. Due to the complexity and time demands for developing a meaningful
solution [4-5], Ishida and Kobayashi [6] and Malhotra and Veletsos [7] developed simplified methods of analysis
considering the behaviour of the base plate as a beam. More recently, Kobayashi et al. [8] presented a computational
model based on the finite element method that considers uplift and cross-sectional deformations of the tank wall.

Experimental research on seismic uplift has been carried out by Clough [9] and Niwa [10], who reported on
experimental studies for broad and tall storage tanks respectively. They presented the static and dynamic axial and
hoop stress distribution and the cross-sectional deformation at mid-height, as a function of time. Both studies
considered constant the base plate thickness. Recently, Taniguchi [11] performed experimental tests and evaluated the
influence of the mass of liquid that contributes to uplift but the contribution of both the tank wall and base plate
thickness were ignored. Ormeo et al. [12] experimentally assessed the effect of uplift on the tank wall stresses. Even
though most codes and guidelines state that uplift should be avoided, in that study, it was demonstrated that, for certain
aspects ratios (ratio of the tank radius to liquid height: /), uplift may reduce both axial and hoop stresses, even
though higher displacements and accelerations occur. The favourable effect of uplift has also been reported in the
seismic behaviour of other structures [13-16].

The base plate stiffness directly affects the uplift resistance, which is evident in the results of some of the studies above
[3,5,7]. However, more work is required to clarify the influence of the base plate stiffness on the hoop stress
development along the tank height. To the authors knowledge, experimental tests varying the base plate stiffness have
not been reported. Hence, the objective of this work is to experimentally evaluate the influence the base plate stiffness
has on the distribution of hoop stresses circumferentially and along the height of the tank wall.

2. Experimental methodology

2.1. Tank model

A low-density polyethylene (LDPE) container (750mm height, 450mm diameter) is used to simulate a prototype steel
tank. Three aspect ratios (1, 2 and 3) were considered. The tank is unanchored hence uplift can occur. The properties
of the prototype and model are described in Table 1. The fundamental period of the prototype was estimated according
to the New Zealand Society for Earthquake Engineering (NZSEE): Recommendations for the design of liquid storage
tanks [17].

Veletsos and Tang [18] demonstrated that convective effects (oscillations at the fluid free surface) can be evaluated
independently of the impulsive effects (inertia forces on the tank wall). For this reason, and for practical purposes,
storage tanks are usually analysed as a single degree of freedom (SDOF) system considering only the impulsive
effects. Hence, the similitude conditions between the prototype and the model were defined using the Buckingham
theorem [19] and applying the Cauchy number, defined in [20], for SDOF systems. The similitude conditions are
shown in Table 1 whereas the scale factors are shown in Table 2.

2.2. Setup

Five vertical columns of strain gauges were attached to the external face of the container to measure hoop stresses, as
shown in Fig. 1. The columns are referred to by the azimuth with respect to the line of motion of the shake table. In
 Diego Hernandez-Hernandez et al./ Procedia Engineering 00 (2017) 000000 3

order to measure uplift, linear variable differential transformers (LVDTs) were used to record the vertical movement
of a plastic block attached to the container 200mm above the tank base. The default base plate thickness of the
container was 4mm. Three more round LDPE sheets, each of 3mm thickness, were added to increase the base plate
stiffness, giving a total of 13mm thickness for Test BP4. The thickness ratio (ratio of tank wall thickness to base plate
thickness / ) and the four cases studied are presented in Table 3.

Table 1. Properties of the model and prototype

Model Prototype
Material LDPE Steel
Elastic Modulus () 0.3 x 109 2.068 x 1011
Radius () 0.225 4.500
Maximum Liquid Height () 0.675 13.500
Wall thickness () 0.004 0.010
Maximum liquid Mass () 107 858833
Fundamental period () 0.066 0.143

Table 2. Scale factors Table 3. Four cases of base plate thickness

Dimension Scale factor Base Plate Base plate thickness
Thickness ratio
Length 20 (BP) (mm)
Mass 8000 1 4 1.00
Time 2.16 2 7 1.75
Acceleration 4.31 3 10 2.50
4 13 3.25

(a) (b) (c)

Fig. 1. (a) Radial distribution of strain gauges; (b) Top view of the container; strain gauges columns arrangement; (c) LDPE container

2.3. Shake table input

Based on the NZSEE procedure to simulate earthquakes, a ground motion for a soft soil site, classification D, was
generated. The target spectrum was estimated based on NZS 1170.5 [21] along with the NZSEE recommendations for
the design of liquid storage tanks [17]. The ground motion is shown in Fig. 2 whereas the target spectrum and the
4 Diego Hernandez-Hernandez et al./ Procedia Engineering 00 (2017) 000000

response spectrum are plotted in Fig.3. According to Ormeo et al. [22], the NZSEE 1170.5 code establishes a
restriction of the minimum period of the range of interest and, consequently, the full range of possible tank frequencies
is no accounted for. As a result, it may lead to an underestimation of stresses in the tank wall. In order to avoid that
underestimation, that restriction is ignored is this study. Thus, the lower limit is 0.4 = 0.057 () and the upper limit
is 1.3 = 0.186 (), see Figure 3.

1
Acceleration (g)

0.5

-0.5

-1
0 2 4 6 8 10 12 14 16 18 20
Time (s)
Fig. 2. Simulated ground motion for soft soils generated by NZSEE procedure

10
Spectral Acceleration (g)

Target spectrum

1
Ti=0.143

Range of interest
0.057 0.186
0.1
0.01 0.1 Period (s) 1 10
Fig. 3. Target spectrum and response spectrum

3. Results

The maximum hoop stresses of the four cases, for the three aspect ratios, (H=225mm, 450mm, 675mm), are shown in
Fig. 4. Considering the Fig. 4(a) and comparing the thinnest (4mm) and the thickest (13mm) base plate, the maximum
hoop stresses increase 29% for the azimuth of 180 degrees and decrease 33% for the azimuth of 0. For the remaining
radial directions, the variation is minimal. A similar tendency is presented in Fig. 4(b). The hoop stresses rise 63% for
an azimuth of 180, whereas a decrease of 22% occurs for zero degrees. Fig. 4(c) shows for the azimuth of 180, the
hoop stresses increase in more than 100%, when the BP1 (4mm) case and the BP3 (10mm) case are compared.
However, the hoop stresses fall 55% in the thickest base plate (13mm). This stress reduction may be caused by uplift
of the base that only occurs for an aspect ratio of three and for the thickest base plate.

Ormeo et al. [12] demonstrated that if uplift occurs, the axial stress distribution may be reduced, depending on the
aspect ratio. Those results, however, do not provide enough information about the distribution of hoop stresses in the
tank wall. In the present investigation, the container undergoes a maximum uplift of 2.53mm. Even though this is a
relatively small uplift, it affects significantly the hoop stress distribution in the direction of the excitation and confirms
what Ormeo et al. found (see Figure 5a). Nevertheless, for the direction perpendicular to the excitation (azimuth of
90), when uplift occurs, an increase of the maximum hoop stresses is noted, Fig. 5b. The magnitude of the hoop stress
in the azimuth of 90 in one-third greater than the hoop stress in the azimuth of 180. For an azimuth of 180, the
maximum hoop stress occurred 75mm above the tank base. For an azimuth of 90, however, it occurred 100mm from
the tank base regardless of the thickness of the base plate. In Fig. 4, the height () is normalized with respect to the
water level ().
 Diego Hernandez-Hernandez et al./ Procedia Engineering 00 (2017) 000000 5

0.8 180 135 90 45 0

0.62
Maximum hoop

0.6 0.54
stress (kPa)

0.49
0.42 0.39 0.39 0.43 0.42
0.4

0.2

0
BP1 = 1 BP2 = 1.75 BP3 = 2.5 BP4 = 3.25
(a) / = 1

2
1.68
Maximum hoop

1.5 1.46
stress (kPa)

1.24
1.03
1

0.5

0
BP1 = 1 BP2 = 1.75 BP3 = 2.5 BP4 = 3.25
(b) / = 2

3.5 3.17
3
Maximum hoop
stress (kPa)

2.5
2.06
2 1.80
1.52 1.52 1.50 1.52 1.41
1.5
1
0.5
0
BP1 = 1 BP2 = 1.75 BP3 = 2.5 BP4 = 3.25
(c) / = 3
Fig. 4. Influence of slenderness on the maximum hoop stress

1 1
H/R = 3 (180) H/R = 3 (90)
0.8 BP1 0.8 BP1
BP2 BP2
0.6 0.6
BP3 BP3
d/H
d/H

0.4 BP4 0.4 BP4

0.2 0.2
(a) (b)
0 0
0 1 2 3 4 0 1 2 3 4
Hoop stress (kPa) Hoop stress (kPa)
Fig. 5. Hoop stress distribution for / = 3, (a) 180 and (b) 90
6 Diego Hernandez-Hernandez et al./ Procedia Engineering 00 (2017) 000000

4. Conclusions

A low-density polyethylene container representing an unanchored steel liquid storage tank has been tested on a shake
table using simulated earthquake ground motion. The main aim was to assess the influence of the base plate thickness
on the distribution of hoop stresses in the wall. Four base plate thicknesses, i.e. 4 mm, 7 mm, 10 mm and 13 mm, were
considered. The experiments were performed in triplet for each case.

The results revealed that the base plate thickness plays a significant role in reducing hoop stress development in the
case of an aspect ratio of 3 as the result of transient uplift. The maximum hoop stress was reduced up to 50%. In
contrast, considering the wall cross-section perpendicular to the excitation direction, the maximum hoop stress
increased up to 25%. In the case of aspect ratios of 1 and 2 the effect is marginal as the result continuous full contact
between base plate and support.

Acknowledgements

The authors wish to thank the Mexican Government for awarding the first author the doctoral scholarship
CONACyT-SENER Hidrocarburos for his research at the University of Auckland.

References

[1] R. Peek and P. C. Jennings, Simplified analysis of unanchored tanks, Earthq. Eng. Struct. Dyn., 16, (1988) 10731085.
[2] M. A. Haroun and H. S. Badawi, Seismic behavior of unanchored ground-based cylindrical tanks, Ninth World Conference on Earthquake
Engineering (1988) 643648.
[3] M. A. Haroun and A. A. Al-Zeiny, Nonlinear transient response of unanchored liquid storage tanks, Pressure Vessels and Piping
Conference, (1995) 3541.
[4] S. Natsiavas, An analytical model for unanchored fluid-filled tanks under base excitation, Trans. ASME, 55, (1988) 648653.
[5] R. Peek, Analysis of Unanchored Liquid Storage Tanks Under Lateral Loads, Earthq. Eng. Struct. Dyn., 16, (1988) 10871100.
[6] K. Ishida and N. Kobayashi, An effective method of analyzing rocking motion for unanchored cylindrical tanks including uplift, J. Press.
Vessel Technol., 110, (1988) 7687.
[7] P. K. Malhotra and A. S. Veletsos, Uplifting response of unanchored liquid-storage tanks, J. Struct. Eng., 120 (1994) 35253547.
[8] N. Kobayashi, T. Tashita, S. Takizawa, and T. Taniguchi, Simplified rocking model of unanchored cylindrical tank including baseplate
uplift and cross sectional deformation of tank shell due to seismic load, Pressure Vessels and Piping Conference (2015) 19.
[9] D. P. Clough, Experimental evaluation of seismic design methods for broad cylindrical tanks, Berkeley, California, 1977.
[10] A. Niwa, Seismic behavior of tall liquid storage tanks, Berkeley, California, 1978.
[11] T. Taniguchi, Rocking response on unanchored flat-bottom cylindrical shell tanks subjected to horizontal base excitation. Part II: Flexible
bottom plate, Pressure Vessels and Piping Conference (2003)18.
[12] M. Ormeo, T. Larkin, and N. Chouw, The effect of seismic uplift on the shell stresses of liquid-storage tanks, Earthq. Eng. Struct. Dyn.,
44(12) (2015) 19791996.
[13] Y. Chen, C. Kun, T. Larkin, and N. Chouw, Impact of Vertical Ground Excitation on a Bridge with Footing Uplift, J. Earthq. Eng., 21
(2016) 119.
[14] Y. Chen, T. Larkin, and N. Chouw, Experimental assessment of contact forces on a rigid base following footing uplift, Earthq. Eng.
Struct. Dyn., 2017.
[15] W. Y. Loo, C. Kun, P. Quenneville, and N. Chouw, Experimental testing of a rocking timber shear wall with slip-friction connectors, Int.
Assoc. Earthq. Eng., 43, (2014) 16211639.
[16] M. Ali, R. Briet, and N. Chouw, Dynamic response of mortar-free interlocking structures, Constr. Build. Mater., 42 (2013) 168189.
[17] NZSEE, Seismic design of storage tanks, 2009: Recommendations of a study group of the New Zealand Society for Earthquake
Engineering. Wellington: New Zealand Society for Earthquake Engineering, 2009.
[18] A. S. Veletsos and Y. Tang, Soil-Structure Interaction effects for laterally excited liquid storage tanks, Earthq. Eng. Struct. Dyn., 19
(1990) 473496.
[19] E. Buckingham, On physically similar systems; Illustrations of the use of dimensional equations, Phys. Rev., 4(4), (1914) 345376.
[20] X. Qin, Y. Chen, and N. Chouw, Effect of Uplift and Soil Nonlinearity on Plastic Hinge Development and Induced Vibrations in
Structures, Adv. Struct. Eng., 16 (1) (2013) 135148.
[21] NZS 1170.5, Structural design actions, Part 5: earthquake actionsNew Zealand. Code and supplement. Standards New Zealand, vol. 5.
2004.
[22] M. Ormeo, T. Larkin, and N. Chouw, Evaluation of seismic ground motion scaling procedures for linear time-history analysis of liquid
storage tanks, Eng. Struct., 102 (2015) 266277.