Sunteți pe pagina 1din 15

KSCE Journal of Civil Engineering (2011) 15(6):1041-1055 DOI 10.

1007/s12205-011-0945-y

Structural Engineering

www.springer.com/12205

Seismic Response of Liquid Storage Steel Tanks with Variable Frequency Pendulum Isolator
V. R. Panchal* and R. S. Jangid**
Received August 4, 2009/Revised June 8, 2010/Accepted October 27, 2010

Abstract
The seismic response of liquid storage steel tanks (slender and broad) isolated with Variable Frequency Pendulum Isolators (VFPIs) is investigated under normal component of six near-fault ground motions. The continuous liquid mass is lumped as convective mass, impulsive mass and rigid mass. The corresponding stiffness associated with these lumped masses is worked out depending upon the properties of the tank wall and liquid mass. The frictional forces mobilized at the interface of the VFPI are assumed to be velocity independent. The governing equations of motion of liquid storage steel tanks isolated with the VFPIs are derived and solved in the incremental form using Newmarks step-by-step method assuming linear variation of acceleration over small time interval. For comparative study, the seismic response of liquid storage steel tanks with the VFPIs is compared with that of the same liquid storage steel tanks isolated using the Friction Pendulum Systems (FPSs). In order to measure the effectiveness of isolation system, the seismic response of isolated steel tanks is compared with that of the non-isolated steel tanks. Further, a parametric study has been carried out to critically examine the behaviour of liquid storage steel tanks isolated with VFPIs. The important parameters considered are the friction coefficient of the VFPI, the Frequency Variation Factor (FVF) of the VFPI and the tank aspect ratio. The difference between the liquid storage tanks isolated with the VFPI and the FPS isolators subjected to the harmonic and far-field ground motions was also investigated in this study. Effect of vertical component of ground motions on the behaviour of the VFPI-isolated liquid storage tanks is investigated under triaxial ground excitations by considering the interaction of forces in two orthogonal directions. From these investigations, it is concluded that seismic response, viz. the base shear, the sloshing displacement and the impulsive displacement, of liquid storage steel tanks during near-fault ground motions can be controlled within a desirable range with the installation of the VFPI. Under strong harmonic excitations, the base shear of the VFPI reduces than that of the FPS whereas the sloshing displacement, the impulsive displacement and the isolator displacement of the VFPI exceeds than that of the FPS. It is also found that the isolation by the FPS and VFPI isolators has almost the same effect in the tank to the far-field ground motions. The triaxial ground motions have noticeable effect on the response of the VFPI-isolated liquid storage tanks relative to unilateral ground motion and if ignored, the sliding displacement and base shear will be underestimated. Keywords: base isolation, liquid storage steel tank, vibration control, near-fault ground motion, variable frequency pendulum isolator, far-field ground motion, harmonic support motion

1. Introduction
Liquid storage steel tanks are lifeline structures and strategically very important, since they have vital use in industries and nuclear power plants. Unlike most structures (such as buildings or bridges), the weight of storage tanks varies in time because of variable liquid storage level, and they may contain low-temperature (e.g., Liquefied Natural Gas (LNG)) or corrosive substances. Typical damage at tanks during past earthquakes, 1989 Loma Prieta earthquake, 1994 Northridge earthquake, Ji-Ji Taiwan and 1999 Kocaeli earthquqke, were in the form of cracking at the corner of the bottom plate and compression buckling of tank wall (elephant foot buckling) due to uplift, sliding of the base, anchorage failure, sloshing damage around the roof, failure of piping

systems and plastic deformation of base plate (1988, 1990). Failures of storage tanks not only instantly disrupts essential infrastructure but can also cause fires or environmental contamination when flammable materials or hazardous chemicals leak. Consequently, protection of liquid storage steel tanks against severe seismic events has become crucial. For over three decades, seismic isolation technology has been recognized as one of the promising alternatives for protecting liquid storage steel tanks against severe earthquakes. The main concept in isolation is to increase the fundamental period of structural vibration beyond the energy containing period of earthquake ground motions. The other purpose of an isolation system is to provide an additional means of energy dissipation, thereby reducing the transmitted acceleration into the superstructure. The innovative design ap-

*Professor, Dept. of Civil Engineering, Sardar Vallabhbhai Patel Institute of Technology, Vasad 388 306, Gujarat, India (E-mail: vijay_svit@yahoo.co.in) **Professor, Dept. of Civil Engineering, Indian Institute of Technology Bombay, Powai, Mumbai 400 076, Maharashtra, India (Corresponding Author, Email: rsjangid@civil.iitb.ac.in) 1041

V. R. Panchal and R. S. Jangid

proach aims mainly at the isolation of a structure from the supporting ground, generally in the horizontal direction, in order to reduce the transmission of the earthquake motion to the structure. A number of authors have investigated intensively the seismic response of the liquid storage tanks. Housner (1957, 1963) developed an analytical model. The model assumes that the tank walls are rigid, the fluid is incompressible, and fluid displacements are small. In this model, ground-supported liquid storage tank was idealised by a system with two-degrees-of-freedom by concentrating the mass of the tank at two points. The masses and stiffnesses are functions of the geometry of the tank and fluid elevation. This model has been widely used to investigate the seismic response of the liquid storage tanks. Haroun and Housner (1981) have developed a three-mass model of ground-supported tanks that takes tank-wall flexibility into account. Later, Haroun (1983) developed designed charts to estimate impulsive, convective, and rigid masses, assuming the liquid in the tank as incompressible with irrotational flow. Malhotra (1997a) demonstrated the effectiveness of the isolation in reducing significantly the overturning base moments transferred to the foundation and the axial compressive stresses generated in the tank wall. Malhotra (1997b) investigated the seismic response of base-isolated steel tanks and found that isolation was effective in reducing the response of the tanks over traditional fixed base tank without any significant change in sloshing displacement. Shenton III and Hampton (1999) investigated the seismic response of isolated elevated tanks and found that seismic isolation is effective in reducing the tower drift, base shear, overturning moment, and tank wall pressure for the full range of tank capacities. Wang et al. (2001) investigated the response of liquid storage tanks isolated by the FPS and observed that the isolation was effective in reducing the response of the tanks. Shrimali and Jangid (2002) investigated the seismic response of liquid storage steel tanks isolated by lead-rubber bearings under bi-directional earthquake excitation and observed that the seismic response of isolated tanks is insensitive to interaction effect of the bearing forces. Shrimali and Jangid (2003) investigated earthquake response of elevated liquid storage steel tanks isolated by the linear elastomeric bearings under real earthquake ground motion. Shrimali and Jangid (2004) presented the earthquake analysis of baseisolated liquid storage steel tanks using linear theory of base isolation. Jadhav and Jangid (2006) investigated the seismic response of liquid storage steel tanks isolated by elastomeric bearings and sliding systems under near-fault ground motions and observed that both elastomeric and sliding systems were effective in reducing the earthquake forces of the liquid storage tanks. Akkse (2007) investigated the effectiveness of the Double Concave Friction Pendulum (DCFP) bearings for liquid storage tanks subjected to near-fault and far-field ground motions. Panchal and Jangid (2008) proposed a new friction base isolator named Variable Friction Pendulum System (VFPS) for seismic isolation of liquid storage steel tanks under near-fault ground motions. Abali and Ukan (2010) made a parametric study of liquid storage

tanks isolated by curved surface sliding bearings to compute the sensitivity of critical response parameters such as, period of isolation, tank aspect ratio and the coefficient of friction of sliding bearings to various ground motions. Yet, there are only a few works available in the literature where the seismic response of base-isolated liquid storage steel tanks under near-fault ground motions has been studied. Consequently, the effects of these motions on liquid storage steel tanks are not yet understood fully. Near-fault ground motion can introduce more devastating response to isolated structures than an equal or larger (higher peak ground acceleration) far-field ground motion (Loh et al., 2002). This concern profoundly influenced seismic isolation design requirements in the 1997 Uniform Building Code (ICBO, 1997). In the earlier code there were no near-fault effects but in the recent code, near-fault effects viz. source type and distance dependent near-fault factors to the customary design spectrum have been introduced. However, it is believed that these factors are not sufficient to solve the problem consistently, because they pay little attention to the physical characteristics of near-fault ground motions. In view of the above, it is necessary to conduct the reliable numerical studies on the behaviour of base-isolated liquid storage steel tanks under near-fault ground motions in order to provide assistance to current research and engineering practice. In this paper, the seismic response of liquid storage steel tanks (slender and broad) isolated by the VFPIs is investigated under near-fault, harmonic and far-field ground motions. The specific objectives of the present study may be summarized as: 1. to investigate the performance of the VFPI for seismic isolation of liquid storage steel tanks under near-fault ground motions, 2. to compare the seismic response of liquid storage steel tanks isolated with the VFPI and the FPS in order to measure the effectiveness of the VFPI, 3. to investigate the influence of important parameters on the response of liquid storage steel tanks isolated with the VFPI through a parametric study. The important parameters considered are the friction coefficient of the VFPI, the Frequency Variation Factor (FVF) of the VFPI and the tank aspect ratio, 4. to study the dynamic behavior of the VFPI-isolated tanks under harmonic and far-field ground motions, and 5. Effect of vertical component of ground motions on the behaviour of the VFPI-isolated liquid storage tanks is investigated under triaxial ground excitations by considering the interaction of forces in two orthogonal directions.

2. Near-fault Ground Motions


Seismologists have identified the forward directivity and flingstep as the primary characteristics of near-fault ground motions. These characteristics make near-fault earthquakes unique compared to far field ground motions. The fling-step usually induces only limited inertial demands on structures due to the longperiod nature of the static displacement. On the other hand,
KSCE Journal of Civil Engineering

1042

Seismic Response of Liquid Storage Steel Tanks with Variable Frequency Pendulum Isolator

ground motions that are influenced by forward-directivity effects can be very damaging to structures. Forward-directivity effects are seen when the rupture direction is aligned with the direction of slip, and the rupture front moves towards a given site. These conditions occur readily in strike-slip earthquakes when the rupture propagates horizontally towards a given site. Forward directivity can also occur for dip-slip faulting; although the conditions required are met less readily than for strike-slip faulting. During the last two decades, an ever increasing database of recorded ground motions have demonstrated that the kinematic characteristics of the ground motion near the faults of major earthquakes contain large displacement pulses, say one or two pulses from 0.5 m to more than 1.5 m with peak velocity of 0.5 m/sec or higher. Their period is usually in the range of 1 to 3 sec, but it can be as long as 6 sec. Such distinct pulses do not exist in ground motions recorded at locations away from the near-fault region, i.e., El Centro 1940 earthquake ground motion. In some cases, the coherent pulse is distinguishable not only in the displacement and velocity histories, but also in the acceleration history. In other cases, acceleration histories recorded near the source contain high frequency spikes and resemble the traditional random-like signal; however, their velocity and displacement histories uncover a coherent longperiod pulse with some overriding high-frequency fluctuations. Six ground records are selected as near-fault ground motions for the present study. These near-fault ground motions are taken from Makris and Chang (2000), Jangid and Kelly (2001) and Alavi and Krawinkler (2004). The reason for the selection of only six ground motions is that these ground motions have a variety of Peak Ground Acceleration (PGA), Peak Ground Velocity (PGV) and Peak Ground Displacement (PGD). Moreover, these ground motions cover equivalent velocity pulse period range from 0.75 sec to 5 sec and velocity pulse amplitude range from 0.6 m/sec to 1.75 m/sec. Some characteristics of the selected

ground motions, such as duration, PGA, PGV and PGD are summarized in Table 1 whereas the pulse characteristics of selected near-fault ground motions are summarized in Table 2.

Fig. 1. Fourier Spectrums of Normal Component of Near-fault Ground Motions

Fig. 2. Acceleration and Displacement Spectra of the Six Nearfault Ground Motions for 5% Damping

Table 1. Some Characteristics of Normal Component of Near-fault Ground Motions Used in the Study Near-fault earthquake ground motions (Normal component) October 15, 1979 Imperial Valley, California October 15, 1979 Imperial Valley, California January 17, 1994 Northridge, California June 28, 1992 Landers, California January 17, 1994 Northridge, California January 17, 1994 Northridge, California Recording station El Centro Array #5 El Centro Array #7 Newhall Lucerne Valley Rinaldi Sylmar Duration (sec) 39.420 36.900 60.000 49.284 14.950 60.000 PGD (m) 0.765 0.491 0.381 2.300 0.391 0.311 PGV (m/sec) 0.98 1.13 1.19 1.36 1.75 1.22 PGA (g) 0.37 0.46 0.72 0.71 0.89 0.73

Table 2. Equivalent Pulses for Recorded Normal Component of Near-fault Ground Motions Near-fault ground motions Recording station Magnitude (Mw) 6.4 6.4 6.7 7.3 6.7 6.7
1043

Velocity Pulse Type Cycloidal type-B Cycloidal type-B Cycloidal type-C1 Cycloidal type-B Cycloidal type-A Cycloidal type-C2

Velocity Velocity Pulse period (sec) Pulse amplitude (m/sec) 3.2 3.8 0.75 5.0 0.8 2.3 0.7 1.13 0.9 1.0 1.75 0.6

October 15, 1979 Imperial Valley, California El Centro Array #5 October 15, 1979 Imperial Valley, California El Centro Array #7 January 17, 1994 Northridge, California June 28, 1992 Landers, California January 17, 1994 Northridge, California January 17, 1994 Northridge, California
Vol. 15, No. 6 / July 2011

Newhall Lucerne Valley Rinaldi Sylmar

V. R. Panchal and R. S. Jangid

These near-fault ground motions cover various forms of frequency range as shown in Fig. 1. In order to give further insight into the near-fault ground motions, the acceleration and displacement spectra of the six ground motions for 5% damping are shown in Fig. 2. The spectra of these ground motions indicate that the ground motions are recorded at a firm soil or rock site.

3. Modeling and Idealization


The 3D Finite Element Model (FEM) of a base isolated liquid storage tank is extremely complicated. Therefore simplified models are used for preliminary analysis of tanks. In the present study, the selected liquid storage tank is idealized by a mechanical model. The reason to select mechanical analog model is that recently, Christovasilis and Whittaker (2008) compared the results of a FE and a mechanical analog. They concluded that the Housners (1963) mechanical analog could be used with confidence for the preliminary analysis and design of conventional and isolated tanks. The model considered for the base-isolated liquid storage steel tanks is shown in Fig. 3. The VFPI is installed between the base and the foundation of the tank. The contained liquid is considered as incompressible, inviscid and has irrotational flow. During the base excitation, the entire tank liquid mass vibrates in three distinct patterns: sloshing or convective mass (i.e., top liquid mass which changes the free liquid surface), impulsive mass (i.e., intermediate liquid mass vibrating along with tank wall) and rigid mass (i.e., the lower liquid mass which rigidly moves with the tank wall). There are various modes in which sloshing and impulsive masses vibrates but the response can be predicted by considering first sloshing mode and first impulsive mode as observed experimentally by Kim and Lee (1995) and numerically by Malhotra (1997b). Therefore, the continuous liquid is modeled by lumped masses as suggested by Haroun (1983) with flexible tank. The sloshing, impulsive, and rigid lumped masses are denoted by mc , mi, and mr, respectively. Thus, the base-isolated tank system has three degrees of freedom under uni-directional earthquake excitation. These degrees of freedom

are denoted by uc, ui, and mb which denote the absolute displacement of sloshing, impulsive, and rigid masses, respectively. The various assumptions made for the system under consideration are as follows: 1. The sloshing and impulsive masses are connected to the tank wall by corresponding equivalent spring having stiffness kc and ki , respectively, and are worked out depending upon the properties of the tank wall and liquid mass. 2. The self-weight of the steel tank is neglected since it is very small (less than 5% of the effective weight of the tank). Such assumption is valid because the effective weight of the tank is considerably greater than the dead load of the tank liquid system. 3. The damping constant associated with the movement of sloshing and impulsive masses is expressed by the assumed damping ratio. 4. The friction coefficient of the VFPI is assumed to be independent of the relative velocity at the sliding interface. This is based on the findings that such effects do not have noticeable effects on the peak response of the isolated structural system (Fan et al., 1990). 5. The restoring force provided by the VFPI is considered as non-linear (i.e., not proportional to the relative displacement). 6. The system is excited by normal component of near-fault ground motion only and the contribution of parallel component is ignored. This is based on the findings that the resultant maximum isolator displacement is mainly due to the normal component of the near-fault ground motions (Jangid and Kelly, 2001; Rao and Jangid, 2001; Jadhav and Jangid, 2006). The sloshing, impulsive and rigid masses in terms of liquid mass, m are expressed as: m c = Yc m m i = Yi m m r = Yr m m = R Hw
2

(1a) (1b) (1c) (1d)

where w is the mass density of the tank liquid; Yc , Yi and Yr are the mass ratios which are function of average thickness of tank wall, th and aspect ratio of the tank, S=H/R; H is liquid height; and R is radius of the tank. For th/R = 0.004, the various mass ratios are expressed (Haroun, 1983) as:
Yc = 1.01327 0.87578 S + 0.35708 S 0.06692 S + 0.00439 S
2 3 2 3 4

(2a)
Yi = 0.15467 + 1.21716 S 0.62839 S + 0.14434 S 0.0125 S Yr = 0.01599 + 0.86356 S 0.30941 S + 0.04083S
2 3 4

(2b) (2c) The fundamental frequency of impulsive mass, i and of sloshing mass, c are given by the following expressions as:
Fig. 3. Model of Liquid Storage Steel Tank Isolated with the VFPI
1044

P E - ---i = --H s

(3a)
KSCE Journal of Civil Engineering

Seismic Response of Liquid Storage Steel Tanks with Variable Frequency Pendulum Isolator

g -- tanh ( 1.84 S ) c = 1.84 R

(3b)

where E and s are the modulus of elasticity and density of tank wall, respectively; g is the acceleration due to gravity; and P is a dimensionless parameter expressed as: P = 0.037085 + 0.084302S 0.05088S2 + 0.012523S 3 0.0012S 4 (4)

b = x b = 0 ), if the frictional force the non-sliding phase ( x mobilized at the interface of the VFPI is less than the limiting frictional force (i.e., | Fx| < F s). However, the system starts sliding b 0 ) as soon as the frictional force attains the limiting fricb x (x tional force (i.e., |Fx| = Fs). Due to the highly non-linear behavior of the system, a very small time steps of the order of 0.0001 sec is selected such that it is at least one-hundredth of the impulsive period of the liquid storage tank. To stop the Newmark iterations in each time step, the following convergence criterion is selected. ( x ) j + 1 ( x ) j - Relative error = -----------------------------------( x ) j (8)

The Eqs. (2), (3) and (4) for modeling of liquid storage steel tank are taken from Shrimali and Jangid (2004) derived by curve fitting of the charts given by Haroun (1983) for th/R = 0.004. The similar equations can also be worked out for other th/R ratios. The equivalent stiffness and damping of the sloshing and impulsive masses are expressed as:
2 k c = m c c

where x is the incremental relative displacement; j is the iteration number; and is a small threshold parameter. The convergence parameter is taken as 10-5 in each time step.

(5a) (5b) (5c) (5d)

k i = m i

4. Description of VFPI
A new isolator called the VFPI (Pranesh and Sinha, 2000) incorporates the advantages of both the friction pendulum system (FPS) and Pure-Friction (P-F) isolators (refer Figs. 4a and 4b). In this isolator, the shape of the sliding surface is non-spherical. To be more specific, its geometry has been derived from the basic equation of an ellipse, with its semi-major axis being a linear function of sliding displacement. This is equivalent to an infinite number of ellipses continuously transforming into one another such that the semi-major axis is larger for larger sliding displacement. The performance of the VFPI is found to be very effective for a variety of excitation and structural characteristics. The VFPI is relatively flatter than FPS, which results in smaller vertical displacement for similar displacements. This is an additional advantage of the VFPI compared to the FPS since flatter sliding surface will result in the generation of smaller overturning forces in the structure. The most important properties of this system are: (i) its time period of oscillation depends on the sliding displacement and (ii) its restoring force exhibits softening behaviour. The isolator geometry is such that its frequency decreases with an increase in the sliding displacement and asymptotically approaches zero at very large displacement. As a

2 i

cc = 2 c mc c ci = 2 i mi i

where c and i are damping ratio of sloshing and impulsive masses, respectively. The governing equations of motion of isolated liquid storage tank subjected to near-fault ground motion are expressed in the matrix form as: } + [ C ] { x } + [ K ] { x } + { F } = [ M ] { r } x g [ M ]{x (6)

where {x}={xc, xi, xb}T and {F}={0, 0, Fx}T are the relative displacement and frictional force vectors, respectively; xc = uc ub is the displacement of the sloshing mass relative to bearing displacement; xi = ui ub is the displacement of the impulsive mass relative to bearing displacement; xb = ub ug is the displacement of the bearing relative to ground; [M], [C], and [K] are the mass, damping and stiffness matrices of the system, respectively; {r}={0, 0, 1} is the influence coefficient vector; Fx g is the earthis the frictional force mobilized in the isolator; x quake ground acceleration; T denotes the transpose; and overdots indicate derivative with respect to time. The matrices [M], [C] and [K] are expressed as: mc 0 mc kc 0 0 cc 0 0 [ M ] = 0 mi mi ; [ C ] = 0 ci 0 ; and [ K ] = 0 ki 0 mc mi Mt 0 0 kb 0 00

(7)

where Mt = mc + mi + mr is the total effective mass of the isolated liquid storage tank. The coupled governing equations of motion of the baseisolated liquid storage steel tanks cannot be solved using the classical modal superposition technique due to non-linear forcedeformation behaviour of the VFPI. As a result, the governing equations of motion are solved in the incremental form using Newmarks step-by-step method assuming linear variation of acceleration over small time interval, t. The system remains in
Vol. 15, No. 6 / July 2011 1045

Fig. 4. Details, Hysteresis Loop and Modeling of the VFPI

V. R. Panchal and R. S. Jangid

result, the dominant frequency of excitation and the isolator frequency are not likely to tune. The response of structure with the FPS increases for higher time periods, whereas the response of the VFPI is almost independent of the structural time period. The instantaneous stiffness of the VFPI (Pranesh and Sinha, 2000) can be written as:
2( x ) kb ( xb ) = M b b

, based on the parameter M (refer Eqs. (13) and (14)). The


practical range of FVF recommended by Pranesh and Sinha (2000) is from 3 to 10 per m (i.e., d = 0.10 0.33m). In the present study, the selected value of the d is 5 per m (i.e., d = 0.2 m). For all investigations, the parameters of VFPI are selected as b = 0.04 m and d = 0.2 m (FVF 5 per m) so that it has initial time period of 2.0 sec. For comparison, an example with FPS isolators is also taken with an isolation period of 2.0 sec. For both isolators, the coefficient of sliding friction of 0.05 is selected. The tank parameters such as damping ratio of convective mass, c and the impulsive mass, i are taken as 0.5 and 2%, respectively. The tank wall considered is made of steel with a modulus of elasticity of E = 200 GPa and mass density, s = 7900 kg/m3. The properties of the broad and slender tanks taken from Shrimali and Jangid (2002) are: (i) aspect ratio, S for slender and broad tanks is 1.85 and 0.6, respectively; (ii) the height, H, of water filled in the slender and broad tanks is 11.3 and 14.6 m, respectively; (iii) the natural frequencies of convective mass, c and impulsive mass, i for the broad and slender tank are 0.123, 3.944 Hz and 0.273, 5.963 Hz, respectively; and (iv) the ratio of tank wall thickness to its radius, th/R is taken as 0.004 for both the tanks. Note that the same value of th/R = 0.004 is used in deriving Eqs. (2a)-(2c) and (4). The response quantities of interest of liquid storage tank are the base shear, Fb and the relative displacements of sloshing mass, xc, the impulsive mass, xi, and the isolation system, xb. The base shear is a measure of the hydrodynamic forces generated in the tank which is directly proportional to earthquake forces exerted in the liquid storage tank. Also, the base shear is directly proportional to the axial compressive forces induced in the cylindrical tank wall which causes the buckling (Malhotra, 1997b). The impulsive component controls the hydrodynamic pressures, thus the base shears and overturning moments, while the convective component controls the vertical displacements of the free-surface, hence the freeboard requirement. On the other hand, the relative isolator displacement is crucial from the design point of view of the isolator. 5.1 Response of Liquid Storage Tanks with VFPI under Near-fault Ground Motions The time variation of the base shear, the sloshing displacement, the impulsive displacement and the isolator displacement of slender tank under the Imperial Valley, 1979 near-fault ground motion (recorded at El Centro Array #5 station) is shown in Fig. 5. As earlier mentioned, two types of friction base isolators, i.e., FPS (Tb = 2.0 sec and = 0.05) and VFPI (Ti = 2.0 sec and = 0.05), are considered for comparison of the seismic response. The horizontal patches in the time history of base shear (isolator force) indicate that the base shear, which consists of the restoring force and frictional force, remains more or less constant. This is due to the fact that the isolator frequency of the VFPI is a function of the isolator displacement. In other words, the isolator stiffness (restoring force) decreases as the isolator displacement increases. The contribution of restoring force to the isolator force is negligible beyond particular value of the isolator displacement.
KSCE Journal of Civil Engineering

(9)
2 b

2 b ( xb ) = -------------------------------------2
(1 + r) rb sgn ( xb ) r = --------------------d 2 = gb ----b 2 d
2

(1 + 2r)

(10) (11) (12)

dTi = 2 ----(13) gb where M = mc + mi + mr is the total effective mass of the isolated liquid storage tank; b and d are semi-minor axis and initial value of the semi-major axis (which is greater than zero) of sliding surface; sgn(xb) is incorporated to maintain the symmetry of the sliding surface about the central vertical axis. The signum function has a value of +1 for positive value of sliding displacement and -1 for negative value of sliding displacement; r is the nondimensional parameter for the sliding surface; b is the instantaneous frequency of VFPI which depends on the geometry of the sliding surface; i is the initial frequency of VFPI at zero isolator displacement; and Ti is the initial time period of the VFPI. It can be noticed that the ratio b/d 2 governs the initial frequency of the isolator. Similarly, the value of 1/d determines the rate of variation of isolator frequency, and this factor has been defined as frequency variation factor (FVF) (Pranesh and Sinha, 2000). It can also be seen from Eq. (10) that the rate of decrease of isolator frequency is directly proportional to the FVF for a given initial frequency. The limiting value of the frictional force, Fs, to which the sliding system can be subjected in a particular direction, is expressed as: Fs = W (14) where is the friction coefficient of the sliding system; and W = Mg is the weight supported by the isolator. Thus, the modeling of VFPI (refer Fig. 4c) is required the specific value of the two parameters, namely initial time period, Ti, and friction coefficient, .

5. Numerical Study
In the study, the seismic response of liquid storage steel tanks isolated with the VFPI is investigated under normal component of six near-fault ground motions. For comparative and detailed parametric study two different types of tanks, namely the broad and slender tanks, are considered. The VFPI isolator is designed to provide the specific values of two parameters, namely Ti and

1046

Seismic Response of Liquid Storage Steel Tanks with Variable Frequency Pendulum Isolator

Fig. 5. Time Variation of the Base Shear, the Sloshing Displacement, the Impulsive Displacement and the Isolator Displacement of Slender Tank Isolated with the FPS (Tb = 2.0 sec and = 0.05) and VFPI (Ti = 2.0 sec and = 0.05) under Imperial Valley, 1979 (El Centro Array #5) Near-fault Ground Motion

Fig. 6. Time Variation of the Base Shear, the Sloshing Displacement, the Impulsive Displacement and the Isolator Displacement of Slender Tank Isolated with the FPS (Tb = 2.0 sec and = 0.05) and the VFPI (Ti = 2.0 sec and = 0.05) under the Imperial Valley, 1979 (El Centro Array #7) Nearfault Ground Motion

Hence, the isolator force is more or less equal to frictional force which is the product of coefficient of friction and gravitational force. This means that isolator force remains more or less constant due to constant frictional force. On the other hand, similar types of horizontal patches are not observed in the FPS, because the restoring force in the FPS varies linearly with sliding displacement, so that the isolator force increases with sliding displacement. The peak values of the base shear, the sloshing displacement, the impulsive displacement and the isolator displacement of slender tanks isolated with the FPS and the VFPI are 0.2776 and 0.0818 W, 181.1422 and 94.3859 cm, 0.2580 and 0.1620 cm and 22.6237 and 94.0743 cm, respectively. This implies that with the installation of the VFPI in liquid storage slender tanks, seismic response of the slender tanks excluding the isolator displacement can be controlled within desirable range during the near-fault ground motions. The figure also reveals that the performance of the VFPI is quite effective for controlling the base shear, the sloshing displacement and the impulsive displacement than that of the FPS. However, the isolator displacement of the VFPI exceeds that of the FPS. In other words, besides optimum characteristic variable curvature sliding isolators may give higher isolator displacement without any reduction in isolator force. This is
Vol. 15, No. 6 / July 2011

expected as the horizontal stiffness of the VFPI is lower than that of the FPS. Such large isolator displacement of the VFPI will lead to the requirement of very large isolators, costly flexible connections for utilities and an extensive and expensive loss of space for a seismic gap. Under near-fault ground motions, this feature of the VFPI reduces its effectiveness in comparison to the FPS. Similar differences in the response of slender tanks isolated with FPS and VFPI are also depicted in the Table 3 and Fig. 6 for the Imperial Valley, 1979 near-fault ground motion (recorded at El Centro Array #7 station). The response of the corresponding non-isolated slender tanks is also shown in Figs. 5 and 6 in order to show the effectiveness of both isolators. It is demonstrated from figures that there is significant reduction in the base shear, the sloshing displacement and the impulsive displacement of base-isolated slender tanks in comparison to non-isolated slender tanks, implying that the base isolation is effective in reducing earthquake response of the liquid storage slender tanks. Figure 7 illustrates the time variation of the base shear, the sloshing displacement, the impulsive displacement and the isolator displacement of broad tanks under the Imperial Valley, 1979 near-fault ground motion (recorded at El Centro Array #5 station). The peak values of the base shear, the sloshing displacement, the impulsive displacement and the isolator displacement

1047

V. R. Panchal and R. S. Jangid

Table 3. Peak Seismic Response of Non-isolated and Isolated Liquid Storage Slender Tanks under Near-fault Ground Motions Near-fault ground motions (Normal component) Imperial Valley, 1979 (El Centro Array #5) Tank condition Non-isolated Isolated (FPS) Isolated (VFPI) Non-isolated Isolated (FPS) Isolated (VFPI) Non-isolated Isolated (FPS) Isolated (VFPI) Non-isolated Isolated (FPS) Isolated (VFPI) Non-isolated Northridge,1994 (Rinaldi) Isolated (FPS) Isolated (VFPI) Non-isolated Northridge,1994 (Sylmar) Isolated (FPS) Isolated (VFPI) Peak response quantities Fb/W 0.6303 0.2776 0.0818 0.4445 0.2991 0.0818 1.1248 0.3096 0.0818 0.9058 0.318 0.0818 1.0454 0.5857 0.0818 0.8345 0.4988 0.0818 xc (cm) 151.0800 181.1422 94.3859 154.8600 186.1282 88.2184 86.3520 95.3830 76.7994 154.6100 179.3184 91.3185 64.01400 111.7960 52.4487 84.9740 144.2996 55.5689 xi (cm) 0.7584 0.2580 0.1620 0.4165 0.2733 0.1511 1.3269 0.2872 0.1733 1.3110 0.2859 0.1850 1.1862 0.5161 0.1694 1.0070 0.4408 0.1462 xb (cm) 22.6237 94.0743 24.7575 72.4209 25.8002 35.1287 26.6429 166.8001 53.2483 45.8927 44.6096 38.5011

Imperial Valley,1979 (El Centro Array #7)

Northridge,1994 (Newhall)

Landers,1992 (Lucerne Valley)

of broad tanks isolated with the FPS and the VFPI are 0.1659 and 0.0818 W, 125.0670 and 137.6852 cm, 0.5842 and 0.4082 cm and 11.5202 and 22.0765 cm, respectively. On the other hand, the peak values of the base shear, the sloshing displacement and the impulsive displacement of non-isolated broad tanks are 0.3154 W, 124.5 cm and 1.1095 cm, respectively. This indicates that the performance of the VFPI in broad tanks is also quite effective for controlling the base shear and the impulsive displacement. However, the sloshing displacement is nearly identical for non-isolated broad tanks and broad tanks isolated with the FPS and the VFPI; this implies that the VFPI is more effective in slender tanks in comparison to broad tanks. Fig. 8 and Table 4 show similar phenomenon those observed in Fig. 7. Figure 9 shows hysteresis loops of the FPS and the VFPI for different near-fault ground motions. Similar trend those observed in earlier figures are also observed in this figure. The base shear of base-isolated liquid storage steel tanks with the VFPI is not as large as that of liquid storage steel tanks with the FPS because the horizontal stiffness of the VFPI is softened with an increase of the isolator displacements. From this figure, it is also observed that the number of force reversal cycles in slender tanks appears to be less as compared to the broad tanks. Fig. 10 shows variation of the sloshing displacement, the impulsive displacement and the isolator displacement against the coefficient of friction, under different near-fault ground motions. The friction coefficient of the VFPI is varied in the range of 0.01 to 0.2. The responses are plotted for slender and broad tanks. From this figure, it is observed that the sloshing displacement and the impulsive displacement increase with an increase in the friction coefficient whereas the sliding displacement decreases with an increase in

Fig. 7. Time Variation of the Base Shear, the Sloshing Displacement, thE Impulsive Displacement and the Isolator Displacement of Broad Tank Isolated with the FPS (Tb = 2.0 sec and = 0.05) and the VFPI (Ti = 2.0 sec and = 0.05) under the Imperial Valley, 1979 (El Centro Array #5) Nearfault Ground Motion
KSCE Journal of Civil Engineering

1048

Seismic Response of Liquid Storage Steel Tanks with Variable Frequency Pendulum Isolator

the friction coefficient. This is due to the fact that for higher values of the friction coefficient the isolator becomes relatively stiff; as a result, the isolator displacements are reduced. Figures 11 and 12 demonstrate the effects of the FVF on the base shear, the sloshing displacement, the impulsive displacement and the isolator displacement of slender and broad tanks, respectively. The FVF is varied in the range of 0.01 to 100 m-1. From these figures, it is found that the base shear, the sloshing displacement and the impulsive displacement of both tanks decre-

Fig. 8. Time Variation of the Base Shear, the Sloshing Displacement, the Impulsive Displacement and the Isolator Displacement of Broad Tank Isolated with the FPS (Tb = 2.0 sec and = 0.05) and the VFPI (Ti = 2.0 sec and = 0.05) under the Imperial Valley, 1979 (El Centro Array #7) Nearfault Ground Motion

Fig. 9. Force-deformation Loops of the FPS (Tb = 2.0 sec and = 0.05) and the VFPI (Ti = 2.0 sec and = 0.05) under Different Near-fault Ground Motions

Table 4. Peak Seismic Response of Non-isolated and Isolated Liquid Storage Broad Tanks under Near-fault Ground Motions Near-fault ground motions (Normal component) Imperial Valley, 1979 ( El Centro Array #5) Imperial Valley, 1979 (El Centro Array #7) Tank condition Non-isolated Isolated (FPS) Isolated (VFPI) Non-isolated Isolated (FPS) Isolated (VFPI) Non-isolated Northridge, 1994 (Newhall) Isolated (FPS) Isolated (VFPI) Landers, 1992 (Lucerne Valley) Non-isolated Isolated (FPS) Isolated (VFPI) Non-isolated Northridge, 1994 (Rinaldi) Isolated (FPS) Isolated (VFPI) Northridge, 1994 (Sylmar) Non-isolated Isolated (FPS) Isolated (VFPI) Peak response quantities Fb/W 0.3154 0.1659 0.0818 0.3570 0.2192 0.0818 0.9724 0.3651 0.0818 0.5966 0.177 0.0818 0.7745 0.5633 0.0818 0.6728 0.3806 0.0818
1049

xc (cm) 124.5000 125.0670 137.6852 105.9300 106.5683 105.3314 54.2060 60.5593 48.6036 201.6200 196.8790 192.2722 49.7840 78.9332 61.1497 49.3650 73.8820 59.5372

xi (cm) 1.1095 0.5842 0.4082 1.3103 0.7668 0.4424 5.0369 1.0732 0.6099 2.7021 0.6704 0.602 3.8143 1.6012 0.6341 2.6484 1.1859 0.7511

xb (cm) 11.5202 22.0765 16.8203 34.5351 31.3150 16.7932 12.6207 26.2812 51.0161 36.2166 32.8582 37.9159

Vol. 15, No. 6 / July 2011

V. R. Panchal and R. S. Jangid

Fig. 10. Variation of the Peak Sloshing Displacement, the Impulsive Displacement and the Isolator Displacement of Slender and Broad Tanks against the Friction Coefficient of the VFPI

Fig. 12. Variation of the Peak Base Shear, the Sloshing Displacement, the Impulsive Displacement and the Isolator Displacement of Broad Tanks Against the FVF of the VFPI

ase with an increase in the FVF whereas the isolator displacement increases with an increase in the FVF. Further, the effects of the FVF on the impulsive displacement and the isolator displacement of the broad tanks are less than that of the slender tanks. Figure 13 illustrates the effects of the aspect ratio on the sloshing displacement, the impulsive displacement and the isolator displacement. The response of tank is obtained for different aspect ratio (i.e., 0.5 to 4). From this figure, it is found that the sloshing displacement decreases with an increase in the aspect ratio whereas the isolator displacement initially increases up to 1.5 and beyond that there is no significant change in the isolator displacement. On the other hand, the impulsive displacement first decreases, attains a minimum value and then increases with an increase in the aspect ratio. This implies that the sloshing displacement and impulsive displacement in slender tanks isolated with the VFPI are less than that of the broad tanks isolated with the VFPI whereas the isolator displacement in the slender tanks

is more than that of the broad tanks. The variation of base shear against aspect ratio is not shown in the figure. This is due to the fact that the variation of base shear against aspect ratio is negligible. This can be observed from Tables 1 and 2 of slender and broad tanks, respectively wherein the values of peak base shear for various near-fault ground motions are constant. 5.2 Response of Liquid Storage Tanks with VFPI under Harmonic Ground Motions Presented in this section is the response of a liquid storage tank isolated with the VFPI to harmonic support motion having amplitude in the range of 0.075 - 0.45 g and excitation frequency of 2 Hz. The peak values of the base shear, the sloshing displacement, the impulsive displacement and the isolator displacement of slender and broad tanks isolated with the FPS and the VFPI are shown in Table 5. This table shows that the base shear, the sloshing displacement, the impulsive displacement and the isolator displacement of liquid storage tank isolated with the VFPI and the FPS are nearly identical for low level of excitation and the difference between the above response quantities of liquid storage tanks with the VFPI and the FPS increases with an increase in the PGA. Figs. 14 and 15 show the time histories of base shear, sloshing displacement, impulsive displacement and isolator displacement of the liquid storage slender and broad tanks isolated with VFPI subjected to sinusoidal excitations of 0.45 g amplitude. These figures indicate that under strong excitations, the base shear of the VFPI reduces than that of the FPS whereas the sloshing displacement, the impulsive displacement and the isolator displacement of the VFPI exceeds than that of the FPS. These figures also reveal that the maximum isolator displacement occurs in the transient part of the response and requires long time to reach steady state phase of response. Fig. 16 shows the comparison of the hysteresis loops of the VFPI and FPS for sinusoidal excitation of 0.45 g amplitude. From this figure, it is observed that for high amplitude excitation, the shape and pattern of hysteresis loops of both the isolators are different.
KSCE Journal of Civil Engineering

Fig. 11. Variation of the Peak Base Shear, the Sloshing Displacement, the Impulsive Displacement and the Isolator Displacement of Slender Tanks Against the FVF of the VFPI

1050

Seismic Response of Liquid Storage Steel Tanks with Variable Frequency Pendulum Isolator

Fig. 13. Variation of the Peak Sloshing Displacement, the Impulsive Displacement and the Isolator Displacement of Tank Isolated with the VFPI Against the Aspect Ratio Table 5. Peak Seismic Response of Isolated Liquid Storage Tanks to Various PGA PGA (g) 0.075 0.15 0.225 0.3 0.375 0.45 Tank condition Isolated (FPS) Isolated (VFPI) Isolated (FPS) Isolated (VFPI) Isolated (FPS) Isolated (VFPI) Isolated (FPS) Isolated (VFPI) Isolated (FPS) Isolated (VFPI) Isolated (FPS) Isolated (VFPI) Fb (W) Slender 0.0526 0.0525 0.0683 0.0643 0.0888 0.0736 0.1105 0.0788 0.1328 0.0813 0.1553 0.0818 Broad 0.0503 0.0503 0.0597 0.0585 0.0758 0.0685 0.0945 0.0755 0.1141 0.0795 0.1342 0.0814 3.9368 3.9324 8.1079 7.9809 12.7276 12.8476 17.6486 18.6002 22.7806 24.2829 28.1082 30.9512 xc (cm) Slender Broad 7.9716 7.9627 15.5767 15.5367 23.2989 23.2949 31.0645 31.252 38.4541 39.0447 45.6848 47.993 0.0805 0.0805 0.1167 0.1200 0.1447 0.1452 0.1717 0.1522 0.1705 0.1808 0.1746 0.1932 xi (cm) Slender Broad 0.1961 0.1962 0.4658 0.4624 0.5811 0.5907 0.5481 0.5936 0.5342 0.5817 0.5559 0.5519 0.2550 0.2547 1.8169 1.8429 3.8517 3.9857 6.0143 6.3409 8.2280 9.1445 10.4687 13.2882 xb (cm) Slender Broad 0.0299 0.0301 0.9603 0.9690 2.5601 2.6455 4.4233 4.6835 6.3758 6.9036 8.3719 9.2498

On the other hand, the shape and pattern of hysteresis loops of both the isolators are nearly identical for low level of the excitation (not shown). Figs. 5 and 14, it is observed that the isolation by the VFPI isolators is more effective than the isolation by the FPS isolators in the liquid storage tank subjected to the near-fault ground motion. 5.3 REsponse of Liquid Storage Tanks with VFPI under Far-field Ground Motions The difference between the liquid storage tanks with the VFPI and the FPS isolators subjected to the far-field ground motion was also investigated. The far-field ground motion selected for the study is North-South component of May 18, 1940, Imperial Valley earthquake (recorded at a site in El Centro, California). The PGA, PGV and PGD values of this ground motion are 0.319 g, 0.362 m/sec and 0.214 m, respectively. Figs. 17 and 18 show the time variation of the base shear, the sloshing displacement, the impulsive displacement and the isolator displacement of the
Vol. 15, No. 6 / July 2011

slender and broad tanks isolated with the VFPI and the FPS under El Centro, 1940 far-field ground motion, respectively. The comparison between hysteresis loops of FPS and VFPI isolators are shown in Fig. 19. It is shown from the figures that the base shear, the sloshing displacement, the impulsive displacement and the isolator displacement of structure isolated with the VFPI and the FPS are nearly identical. For far-field ground motions, the hysteretic loop is similar to friction pendulum system and clearly shows bi-linear behavior. Because the intensity of the El Centro, 1940 ground motion is not very severe, there is no significant difference on dynamic behavior between the VFPI and the FPS under the El Centro, 1940 earthquake. In other words, the isolation by the FPS and the VFPI isolators has almost the same effect in the tank to the far-field ground motion. From the Figs. 5 and 17, it is observed that the isolation by the VFPI isolators is more effective than the isolation by the FPS isolators in the water storage tank subjected to the near-fault ground motion.

1051

V. R. Panchal and R. S. Jangid

Fig. 14. Time Variation of the Base Shear, the Sloshing Displacement, the Impulsive Displacement and the Isolator Displacement of Slender Tank Isolated with the FPS (Tb = 2.0 sec and = 0.05) and VFPI (Ti = 2.0 sec and = 0.05) Subjected to Sinusoidal Excitation (PGA = 0.45 g)

Fig. 15. Time Variation of the Base Shear, the Sloshing Displacement, the Impulsive Displacement and the Isolator Displacement of Broad Tank Isolated with the FPS (Tb = 2.0 sec and = 0.05) and VFPI (Ti = 2.0 sec and = 0.05) Subjected to Sinusoidal Excitation (PGA = 0.45 g).

Fig. 16. Hysteretic Behavior of the FPS (Tb = 2.0 sec and = 0.05) and the VFPI (Ti = 2.0 sec and = 0.05) under Sinusoidal Excitation (PGA = 0.45 g)

5.4 Effect of the Vertical Acceleration on the Response of the VFPI-Isolated Liquid Storage Tank The necessity of including vertical component of ground excitation in the design of sliding structure was noticed by Liaw et al. (1988). They observed that as the frictional stress is a function of vertical reaction on the foundation, both vertical and frictional forces vary when there is vertical motion on the sliding system. Lin and Tadjbakhsh (1986) studied the effect of vertical ground motion on the horizontal response of a two-dimensional P-F

system and indicated that the effect of vertical motion is significant for harmonically excited foundations only. Hwang and Hsu (2000) performed experimental study of base-isolated building under triaxial ground excitation and showed that the vertical component of ground motion increases the peak story acceleration significantly, compared to those obtained without considering vertical component. Responses of a rigid structure with a frictional base isolation system subjected to random horizontalvertical earthquake excitations were studied by Su and Ahmedi (1988). They observed that vertical excitation amplifies the horizontal slip of the structure to an extent. Almazan et al. (1998) developed different models of the structure isolated by FPS to study the effect of vertical ground motion. They concluded that if isolator deformations larger than 15 cm are expected, exact structural model with vertical flexibility should be used for dynamic analysis including vertical ground motion component and large isolator deformations. Shakib and Fuladgar (2003) evaluated the effect of vertical ground motion on the response of a threedimensional single-story structure resting on the P-F system and stated that the peak absolute acceleration and peak base displacement of the low period superstructure may be underestimated if the vertical component of earthquake is not considered. In spite of the thorough work already done to understand the effect of vertical component of the on the dynamic behaviour of
KSCE Journal of Civil Engineering

1052

Seismic Response of Liquid Storage Steel Tanks with Variable Frequency Pendulum Isolator

Fig. 17. Time Variation of the Base Shear, the Sloshing Displacement, the Impulsive Displacement and the Isolator Displacement of Slender Tank Isolated with the FPS (Tb = 2.0 sec and = 0.05) and VFPI (Ti = 2.0 sec and = 0.05) under the El Centro, 1940 Far-field Ground Motion

Fig. 18. Time Variation of the Base Shear, the Sloshing Displacement, the Impulsive Displacement and the Isolator Displacement of Broad Tank Isolated with the FPS (Tb = 2.0 sec and = 0.05) and VFPI (Ti = 2.0 sec and = 0.05) under the El Centro, 1940 Far-field Ground Motion

stant but rather varies with changes in the vertical acceleration. For vertically rigid structures, the normal reaction on VFPI is given by: gz u - N = W 1 + ---- g (15)

Fig. 19. Hysteretic Behavior of the FPS (Tb = 2.0 sec and = 0.05) and the VFPI (Ti = 2.0 sec and = 0.05) under the El Centro, 1940 Far-field Ground Motion

sliding structures, especially buildings, there are still important aspects of the behaviour of liquid storage tanks isolated by the VFPI under triaxial ground excitations that need further investigation. The normal reaction at the sliding surface does not remain con-

gz is the z-component of the earthquake ground accelerwhere u ation; and N is the normal reaction at the sliding surface. The effect of vertical component of earthquake on the horizontal response of the system is studied by comparing the behaviour of the system subjected to unilateral, bilateral and triaxial ground motions. The response of the VFPS-isolated liquid storage tank is obtained for two earthquake ground motions listed in Table 6. Force displacement curves of the slender and broad tanks isolated with VFPI under the unilateral, bilateral and triaxial ground motions are shown in Fig. 20. It is observed from the figure that the triaxial ground motion has noticeable effect on the response of the VFPI-isolated liquid storage tanks relative to unilateral ground motion and if ignored, the sliding displacement
Applied in y-direction Component 090 140 PGA (g) 0.583 0.376 Applied in z-direction PGA (g) 0.548 1.695

Table 6. Triaxial Ground Motions used in the Study Earthquake 1994 Northridge (New Hall Fire Station) 1979 Imperial Valley (El Centro Array #6)
Vol. 15, No. 6 / July 2011

Applied in x-direction Component 360 230 PGA (g) 0.589 0.436


1053

V. R. Panchal and R. S. Jangid

Fig. 20. Force Displacement Curves of the Slender and Broad Tanks Isolated with VFPI (Ti = 2.0 sec and = 0.05) under the Unilateral, Bilateral and Triaxial Ground Motions

and base shear will be underestimated. Similar trends those observed in the Fig. 20 are also observed in the Table 7 for slender and broad tanks.

6. Conclusions
The seismic response of liquid storage steel tanks isolated with Variable Frequency Pendulum Isolator (VFPI) is investigated under six recorded near-fault ground motions. The normal component of six near-fault ground motions is utilized as input to study the variation of the base shear, the sloshing displacement, the impulsive displacement and the isolator displacement. The comparison of the seismic response of the liquid storage steel tanks isolated with the VFPI and the FPS is made in order to verify the effectiveness of the VFPI. Further, a parametric study has been carried out to critically examine the behaviour of the liquid storage steel tanks isolated with the VFPI. The important

parameters considered are the friction coefficient of the VFPI, the Frequency Variation Factor (FVF) of the VFPI and the tank aspect ratio. The difference between the liquid storage tanks isolated with the VFPI and the FPS isolators subjected to the harmonic and far-field ground motions was also investigated in this study. Effect of vertical component of ground motions on the behaviour of the VFPI-isolated liquid storage tanks is investigated under triaxial ground excitations by considering the interaction of forces in two orthogonal directions. From the trends of the numerical results of the present study, the following conclusions may be drawn: 1. With the installation of the VFPI in liquid storage tanks, the base shear, the sloshing displacement and the impulsive displacement during near-fault ground motions can be controlled within a desirable range. On the other hand, significant increase in the isolator displacement is observed. 2. The performance of the VFPI for seismic isolation of the liquid storage tanks is quite effective in reducing the base shear and impulsive displacement of the liquid storage tanks in comparison to the FPS. 3. The VFPI is found to be more effective for slender tanks in comparison to the broad tanks as the sloshing displacement of the broad tanks is not much influenced due to isolation of tanks by the VFPI. 4. The base shear, sloshing displacement and impulsive displacement of both tanks decrease with an increase in the FVF whereas the isolator displacement increases with an increase in the FVF. 5. The sloshing displacement and impulsive displacement increase with an increase in the friction coefficient whereas the isolator displacement decreases with an increase in the friction coefficient. 6. The sloshing displacement and impulsive displacement in slender tanks isolated with the VFPI are less than that of the broad tanks isolated with the VFPI whereas the isolator displacement in the slender tanks is more than that of the broad tanks. 7. Under strong harmonic excitations, the base shear of the VFPI

Table 7. Peak Response Quantities of Liquid Storage Tanks Isolated with VFPI (Ti = 2.0 sec and = 0.05) under Different Components of Earthquakes used in the Study Earthquake Components Unilateral (Slender) Bilateral (Slender) 1994 Northridge (Newhall Fire Station) Triaxial (Slender) Unilateral (Broad) Bilateral (Broad) Triaxial (Broad) Unilateral (Slender) Bilateral (Slender) 1979 Imperial Valley (El Centro Array #6) Triaxial (Slender) Unilateral (Broad) Bilateral (Broad) Triaxial (Broad) Fbx (W) 0.082 0.081 0.114 0.082 0.080 0.116 0.082 0.080 0.123 0.082 0.074 0.122 Fby (W) 0.082 0.079 0.083 0.082 0.078 0.088 0.082 0.059 0.106 0.082 0.066 0.084 xc (cm) 78.2140 52.3480 49.8790 55.8060 56.5000 55.0210 106.5400 87.9140 88.5620 111.5400 112.5300 112.5500 yc (cm) 27.9270 30.9700 31.0560 31.0260 32.8310 32.9330 72.7590 46.1440 45.8430 55.2190 55.1790 54.8460 xi (cm) 0.1625 0.1003 0.1505 0.4531 0.3117 0.4739 0.1299 0.1258 0.2095 0.4382 0.4268 0.4840 yi (cm) 0.1900 0.1043 0.1296 0.6075 0.3833 0.3552 0.1960 0.1725 0.1084 0.4908 0.2562 0.3157 xb (cm) 34.2560 40.1470 41.1520 22.6860 24.8150 24.4230 102.8400 84.6550 84.7960 24.3140 35.7490 35.8710 yb (cm) 16.335 15.3470 15.0850 10.2440 10.9810 11.1290 35.6110 34.9740 35.3030 12.3920 26.1640 26.0190

1054

KSCE Journal of Civil Engineering

Seismic Response of Liquid Storage Steel Tanks with Variable Frequency Pendulum Isolator

reduces than that of the FPS whereas the sloshing displacement, the impulsive displacement and the isolator displacement of the VFPI exceeds than that of the FPS. The maximum isolator displacement occurs in the transient part of the response and requires long time to reach steady state phase of response. 8. The isolation by the VFPI and FPS isolators has almost the same effect in the tank to the far-field ground motions. 9. The triaxial ground motions have noticeable effect on the response of the VFPI-isolated liquid storage tanks relative to unilateral ground motion and if ignored, the sliding displacement and base shear will be underestimated.

References
Abali, E. and Ukan, E. (2010). Parametric analysis of liquid storage tanks base isolated by curved surface sliding bearings. Soil Dynamics and Earthquake Engineering, Vol. 30, Nos. 1-2, pp. 21-31. Akkse, M., Ates , S ., Bayraktar, A., and Dumanolu, A. A. (2007). The seismic response of liquid storage tanks isolated by double concave friction pendulum bearings to near-fault ground motions. 10th World Conference on Seismic Isolation, Energy Dissipation and Active Vibrations Control of Structures, Istanbul, Turkey. Alavi, B. and Krawinkler, H. (2004). Behaviour of moment-resisting frame structures subjected to near-fault ground motions. Earthquake Engineering and Structural Dynamics, Vol. 33, No. 6, pp. 687-706. Almazan, J. L., De la llera, J. C., and Inaudi, J. A. (1998). Modeling aspects of structures isolated with the frictional pendulum system. Earthquake Engineering and Structural Dynamics, Vol. 27, No. 8, pp. 845-867. Chalhoub, M. S. and Kelly, J. (1988). Theoretical and experimental studies of cylindrical water tanks in base isolated structures. A Report No. UCB/EERC-88/07, USA, Berkeley. Chalhoub, M. S. and Kelly, J. (1990). Shake table test of cylindrical water tanks in base isolated structures. Journal of Engineering Mechanics, ASCE., Vol. 116, No. 7, pp. 1451-1472. Christovasilis, I. P. and Whittaker, A. S. (2008). Seismic analysis of conventional and isolated LNG tanks using mechanical analogs. Earthquake Spectra, Vol. 24, No. 3, pp. 599-616. Fan, F.-G., Ahmadi, G., and Tadjbakhsh, I. G. (1990). Multi-story baseisolated buildings under a harmonic ground motion-Part II: Sensitivity analysis. Nuclear Engineering and Design, Vol. 123, No. 1, pp. 17-26. Haroun, M. A. (1983). Vibration studies and test of liquid storage tanks. Earthquake Engineering and Structural Dynamics, Vol. 11, No. 2, pp. 179-206. Haroun, M. A. and Housner, G. W. (1981). Seismic design of liquid storage tanks. Journal of Technical Councils, ASCE, Vol. 107, No. TC1, pp. 191-207. Housner, G. W. (1957). Dynamic pressures on accelerated fluid containers. Bulletin of Seismological Society of America, Vol. 1, No. 4, pp. 15-37. Housner, G. W. (1963). Dynamic behaviour of water tanks. Bulletin of Seismological Society of America, Vol. 53, No. 1, pp. 381-387. Hwang, J. S. and Hsu, T. Y. (2000). Experimental study of isolated building under triaxial ground excitations. Journal of Structural Engineering, ASCE, Vol. 126, No. 8, pp. 879-886. International Conference of Building Officials. (1997). Uniform building code, earthquake regulations for seismic isolated structures, Appendix Chapter 16, Whittier, CA.
Vol. 15, No. 6 / July 2011

Jadhav, M. B. and Jangid, R. S. (2006). Response of base-isolated liquid storage tanks to near-fault motions. Structural Engineering and Mechanics, Vol. 23, No. 6, pp. 615-634. Jangid, R. S. and Kelly, J. M. (2001). Base isolation for near-fault motions. Earthquake Engineering and Structural Dynamics, Vol. 30, No. 5, pp. 691-707. Kim, N. S. and Lee, D. G. (1995). Pseudodynamic test for evaluation of seismic performance of base-isolated liquid storage tanks. Engineering Structures, Vol. 17, No. 3, pp. 198-208. Liaw, T. C., Tian, Q. L., and Cheung, Y. K. (1988). Structures on sliding base subjected to horizontal and vertical motions. Journal of Structural Engineering, ASCE, Vol. 114, No. 9, pp. 2119-2129. Lin, B. C. and Tadjbakhsh, I. G. (1986). Effect of vertical motion on friction driven systems. Earthquake Engineering and Structural Dynamics, Vol. 14, No. 4, pp. 609-622. Loh, C.-H., Liao, W.-I., and Chai, J.-F. (2002). Effect of near-fault earthquake on bridges: Lessons learned from Chi-Chi earthquake. Proceedings of 3rd National Seismic Conference and Workshop on Bridges and Highways, FHWA Western Resource Center and MCEER, Portland, Ore, pp. 211-222. Makris, N. and Chang, S.-P. (2000). Effect of viscous, viscoplastic and friction damping on the response of seismic isolated structures. Earthquake Engineering and Structural Dynamics, Vol. 29, No. 1, pp. 85-107. Malhotra, P. K. (1997a). Method for seismic base isolation of liquidstorage tanks. Journal of Structural Engineering, ASCE, Vol. 123, No. 1, pp. 113-116. Malhotra, P. K. (1997b). New methods for seismic isolation of liquidstorage tanks. Earthquake Engineering and Structural Dynamics, Vol. 26, No. 8, pp. 839-847. Panchal, V. R. and Jangid, R. S. (2008). Variable friction pendulum system for seismic isolation of liquid storage tanks. Nuclear Engineering and Design, Vol. 238, No. 6, pp. 1304-1315. Pranesh, M. and Sinha, R. (2000). VFPI: An isolation device for aseismic design. Earthquake Engineering and Structural Dynamics, Vol. 29, No. 5, pp. 603-627. Rao, P. B. and Jangid, R. S. (2001). Performance of sliding systems under near-fault motions. Nuclear Engineering and Design, Vol. 203, Nos. 2-3, pp. 259-272. Shakib, H. and Fuladgar, A. (2003). Response of pure-friction sliding structures to three components of earthquake excitation. Computers and Structures, Vol. 81, No. 4, pp. 189-196. Shenton III, H. W. and Hampton, F .P. (1999). Seismic response of isolated elevated water tanks. Journal of Structural Engineering, ASCE, Vol. 125, No. 9, pp. 965-976. Shrimali, M. K. and Jangid, R. S. (2002). Non-linear seismic response of base-isolated liquid storage tanks to bi-directional excitation. Nuclear Engineering and Design, Vol. 217, Nos. 1-2, pp. 1-20. Shrimali, M. K. and Jangid, R. S. (2003). Earthquake response of isolated elevated liquid storage steel tanks. Journal of Constructional Steel Research, Vol. 59, No. 10, pp. 1267-1288. Shrimali, M. K. and Jangid, R. S. (2004). Seismic analysis of baseisolated liquid storage tank. Journal of Sound Vibration, Vol. 275, Nos. 1-2, pp. 59-75. Su, L. and Ahmedi, G. (1988). Response of frictional base isolation systems to horizontal-vertical random earthquake excitations. Probabilistic Engineering Mechanics, Vol. 3, No. 1, pp. 12-21. Wang, Y.-P., Teng, M.-C., and Chung, K.-W. (2001). Seismic isolation of rigid cylindrical tanks using friction pendulum bearings. Earthquake Engineering and Structural Dynamics, Vol. 30, No. 7, pp. 1083-1099.

1055

S-ar putea să vă placă și