Liquid Storage Tank

Review of Code Provisions on Design
Seismic Forces for Liquid Storage Tanks

by
Dr. O. R. Jaiswal
Department of Applied Mechanics
Visvesvaraya National Institute of Technology
Nagpur
Dr. Durgesh C Rai

Dr. Sudhir K Jain
Department of Civil Engineering
Indian Institute of Technology Kanpur
Kanpur
Review of Design Seismic Forces for Liquid Storage Tanks
Review of Code Provisions on

Design Seismic Forces for Liquid Storage Tanks
Abstract
It is well recognized that liquid storage tanks possess low ductility and energy
absorbing capacity as compared to the conventional buildings. Accordingly, various
design codes provide higher level of design seismic forces for tanks. In this article,
provisions of IBC 2000, ACI, AWWA, API, Eurocode 8 and NZSEE guidelines are
reviewed, to assess the severity of design seismic forces for tanks vis--vis those for
buildings. It is seen that, depending on the type of tank, design seismic force for tanks
can be 3 to 7 times higher than that for buildings. Based on the comparison of provisions
in these documents, various similarities, discrepancies and limitations in their
provisions are brought out. At the end a brief description of Indian code is given along
with a few suggestions to remove the inadequacies in Indian code.
1. INTRODUCTION
Seismic safety of liquid storage tanks is of considerable importance.
Water storage tanks should remain functional in the post earthquake period to
ensure potable water supply to earthquake-affected regions and to cater the
need for fire fighting. Industrial liquid containing tanks may contain highly
toxic and inflammable liquids and these tanks should not loose their contents
during the earthquake. Liquid storage tanks are mainly of two types: ground
supported tanks and elevated tanks. Elevated tanks are mainly used for water
supply schemes and they could be supported on RCC shaft, RCC or steel
frame, or masonry pedestal.
Failure of tanks during Chilean earthquake of 1960 and Alaska
earthquake of 1964 led to beginning of many investigations on seismic analysis
of liquid storage tanks. Following two aspects came to forefront:
(a) Due consideration should be given to sloshing effects of liquid and
flexibility of container wall while evaluating the seismic forces on tanks.
(b) It is recognized that tanks are less ductile and have low energy absorbing
capacity and redundancy compared to the conventional building systems.
Studies focused on the first aspect resulted in the development of
mechanical models of tank by Housner (1963) and Veletsos (1974), which
represented tank-fluid system in a more realistic fashion. Many investigations
IITK-GSDMA-EQ01-V1.0
followed along this line to further refine these mechanical models to include
effects of flexibility of soil (Hori (1990), Veletsos et. al. (1992)) and base
uplifting of unanchored tanks (Malhotra (1997)). Further studies have provided
more simplifications to these mechanical models (Malhotra (2000)). Most of the
design codes use these mechanical models to represent dynamics of tank-fluid
system, which are applicable to ground supported as well as elevated tanks.
The second aspect which is related to low ductility and redundancy in
tanks as compared to the conventional buildings, has been dealt with in a
rather empirical manner. Lateral seismic coefficient for tanks is generally taken
higher than for the buildings. Wozniak and Mitchell (1978) state the high
value of lateral seismic coefficient for tanks in comparison with buildings is appropriate
because of the low damping inherent for storage tanks, the lack of nonstructural load
bearing elements, and lack of ductility of the tank shell in longitudinal compression.
Most of the design codes do follow this approach and assign higher design
seismic action for tanks as compared to buildings. How high this design action
should be, is perhaps decided on ad-hoc basis or based on past experiences,
however, it is influenced by type of tank, supporting subgrade, type of
anchorage to tank etc. Basically it depends on how good ductility and energy
absorbing capacity a particular type of tank can provide. For elevated tanks,
ductility, redundancy and energy absorbing capacity is mainly governed by the
supporting structure, which could be in the form of a RCC shaft, RCC frame,
Steel frame or even masonry pedestal.
This article presents an assessment of design seismic force for tanks vis-vis design seismic force for buildings as mentioned in the following
documents:
(a) IBC 2000
(b) ACI Standards ACI 371 (1998) and ACI 350.3 (2001)
(c) AWWA D-100 (1996), AWWA D-103 (1997), AWWA D-110 (1995) and
AWWA D-115 (1995)
(d) API 650 (1998)
(e) Eurocode 8 (1998)
(f) NZSEE guidelines and NZS 4203:1992

It may be noted here that IBC 2000, ACI, AWWA and API standards are
from USA. The quantification of design seismic action in ACI, AWWA and API
standards is in a different fashion than IBC 2000. However, FEMA 368 (NEHRP
2000) has provided modifications to these quantifications to bring them in
conformity with provisions of IBC 2000. In the present article, provisions of
ACI, AWWA and API standards will be discussed along with the modifications
of FEMA368. Similarly, in New Zealand, the NZSEE recommendations
(Priestly et. al., 1986) on seismic design of tanks, is being presently revised by a
study group to bring it in line with New Zealand loading code NZS 4203:1992.
The outline of the procedure proposed by this study group is given by
Whittaker and Jury (2000). In the present article, procedure described by
Whittaker and Jury is considered along with NZS 4203:1992.
The assessment of design seismic force for tanks is presented in terms of
design response spectra. This assessment is done with respect to corresponding
design seismic force for buildings. Such a comparative assessment helps in
knowing how severe design seismic action for tank is, as compared to that for a
building under similar seismic exposures. First, provisions on design seismic
action for tanks described in the above-mentioned documents are discussed,
followed by a comparison of design seismic actions from various codes. At the
end a brief description of Indian Standard, IS 1893:1984 is given. Inadequacies
of IS 1893:1984 in quantifying suitable seismic design forces for tanks are
brought out and a few modifications are proposed to remove these limitations.
2. IBC 2000
International Building Code (IBC) 2000 does provide provisions for
certain types of non-building structures which include tanks. For buildings, the
seismic base shear is given by V = Cs W, where, W is the effective seismic
weight. Seismic response coefficient or base shear coefficient, Cs should be
minimum of the following two values Cs =
SDS
SD1
or Cs =
, where SDS and
R/I
( R / I )T
SD1 are the design spectral response accelerations at short periods and 1 second
period, respectively; I is importance factor; R is response modification factor

and T is the fundamental time period of building. The minimum value of Cs
should not be less than 0.044 SDS I. IBC suggests a value of R = 8.0 for buildings
with ductile frames. For most of the buildings, importance factor, I = 1.0. Figure
1 shows the variation of base shear coefficient, Cs = (V/W) with time period.
The values of SDS and SD1 are taken as 1.0 and 0.6 respectively, which
correspond to SD = 1.5, Fa = 1.0, S1 = 0.6 and Fv = 1.5 with site class D.
For tanks, due to low ductility and redundancy, low values of R are
specified. Table 1 gives details of various types of tanks mentioned in IBC
along with their R values. Four values of R are specified, i.e. R = 1.5, 2.0, 2.5 and
3.0. For most of the tanks the value of importance factor I will be I =1.25.
However, for tanks containing highly toxic materials, importance factor could
be I =1.5. The expression for base shear of tank is same as that for building
with suitable values of R and I. For tanks, the minimum value of Cs should not
be less than 0.14 SDSI as against 0.044 SDSI for buildings. For ground-supported
tanks (i.e., at-grade tanks), IBC suggests to include the effects of sloshing.
Similarly, for elevated tanks (i.e., above-grade tanks), IBC states that when
sloshing mode period of the stored liquid is within 70 percent to 150 percent of
the fundamental period of the supporting structure, the effects of sloshing shall
be included in the design of tank and supporting structure. However, IBC 2000
does not provide any particular details on evaluation of sloshing or convective
mode forces. Thus, the values of R specified for tanks can be considered only
for impulsive modes. The variation of base shear coefficient (BSC) for tanks,
with time period is also shown in Figure 1. It is seen from this figure that
depending on the type of tank, base shear coefficient is 3 to 7 times higher than
that of a ductile building. The ratio of base shear coefficient of tank and
building, (BSCtank/BSCbldg), plotted in Figure 2, directly indicates how severe
design base shear for tank is with respect to a ductile building (R = 8.0). The
effect of response reduction factor of tank is seen up to 2 sec. For time period
greater than 3 sec, all types of tank have same base shear coefficient, which is
about four times that for a ductile building.
3. ACI STANDARDS
ACI 371 and ACI 350.3 describe provisions for seismic design of liquid
storage concrete tanks. ACI 371 deals with pedestal supported elevated RCC
tanks only; on the other hand, ACI 350.3 deals with ground supported as well
as elevated tanks. Further, ACI 371 describes consideration of impulsive mode
only, whereas, ACI 350.3 has provisions for impulsive as well as convective
modes. The quantification of design seismic action in these ACI standards is in
a manner different from IBC 2000. In order to bring these quantifications in
conformity with IBC 2000, FEMA 368 has suggested modifications to the base
shear coefficient expressions of ACI standards. Prior to study of various
provisions of ACI standards, it will be appropriate to review the modifications
suggested by FEMA 368. Table 2 gives the details of base shear coefficient
expressions of ACI 371 and ACI 350.3 along with the modified expressions of
FEMA 368.
From Table 2 it is seen that as per ACI standards, in velocity-critical
range of spectra, the impulsive mode base shear coefficient, Cs decreases as a
function of 1/T2/3. However, in FEMA 368, impulsive base shear coefficient, Cs
decreases as 1/T. For convective mode base shear also similar difference can be
noted. To have a better understanding of modifications proposed by FEMA, a
comparison of base shear coefficient obtained from ACI 371 expression and one
obtained from the modified expression of FEMA368 is shown in Figure 3. The
base shear coefficient values shown in Figure 3, correspond to the most sever
zone of ACI 371 and equivalent seismic conditions of FEMA 368. It is seen that
in short period range ( T < 0.6 s), base shear coefficient values of FEMA368 are
about 12% higher than one obtained from ACI 371. In the long period range,
values obtained from both the expressions match well. It may be noted that in
the ACI 371, the importance factor does not appear in the expression for base
shear coefficient, whereas, FEMA368 modification has introduced an
importance factor I =1.25.
Similarly, comparison of base shear coefficient of impulsive mode
obtained from the expression of ACI 350.3 and the one modified by FEMA 368
is shown in Figure 4a. These results are for the most severe zone of ACI 350.3
and FEMA 368. It may be noted that in ACI 350.3 value of response
modification factor, Rw is in the range of 2.0 to 4.75, whereas in FEMA 368, R
varies in the range of 1.5 to 3.0. In Figure 4a, base shear coefficients are shown
for the lowest and the highest values of response modification factor. It is seen
that for the highest value of response modification factor (i.e. Rw = 4.75 and R =
3.0), base shear coefficient from both the expressions match well. For the lowest
value of response modification factor (i.e. Rw = 2.0 and R = 1.5), results of
FEMA 368 are on lower side by 15%. In Figure 4b, base shear coefficient
corresponding to convective mode is compared. For T > 2.4 s, ACI 350.3 and
FEMA 368 expressions give same values of convective base shear coefficient.
3.1 ACI 371 (1998)

This ACI standard provides recommendations for evaluating design
seismic forces on concrete pedestal supported elevated tanks. With FEMA
modifications, the base shear is given by V = Cs .W, where W is the summation
of weight of water, container and support structure above the base. The base
shear coefficient Cs is given by
Cs =
SD1
( R / I)T
Cs =
4S D1
( R / I)T 2
Cs
S DS
( R / I)
for Ts < T < 4 s

for T 4 s
and C s 0.2S DS
(1)
(2)
(3)
The quantities, SDS, SD1, R, I and T are same as defined in IBC 2000 and Ts =
SD1/SDS. FEMA 368 states that except for the above stated modifications,
concrete pedestal supported elevated tanks should be designed as per
provisions of ACI 371. In this ACI standard, a load factor of 1.1 is given for
strength design method, as opposed to unity in IBC 2000. A closer look at the
base shear formula mentioned above reveals that, this is same as one given in
IBC 2000 except for the minimum values of Cs = 0.2 SDS (which is 0.14 SDS I in
IBC 2000) and a load factor of 1.1 for strength design. FEMA 368 specifies R = 2
for pedestal supported elevated tanks. The ratio of base shear coefficient of
tank obtained from ACI 371 (i.e. from Eqs. 1 to 3) and base shear coefficient of
building (obtained as per IBC 2000 with R = 8) i.e. BSCtank / BSCbldg is plotted in
Figure 5. The ratio of tank to building base shear coefficient as obtained from
IBC 2000 is also shown in Figure 5. It is seen that values of base shear
coefficient obtained from ACI 371 are higher than the one obtained from IBC
2000. The higher value of base shear coefficient as per ACI 371 is due to a load
factor of 1.1 and higher value of lower bound limit. ACI 371 does not give
details about convective mass component, and recommends its consideration if
water weight is less than 80 percent of the total gravity load of tank.
3.2 ACI 350.3 (2001)

ACI 350.3 gives procedure for seismic design of liquid containing
concrete tanks, with detailed description of impulsive and convective
components. It mainly deals with ground supported concrete tanks and limited
information on pedestal supported elevated tanks is also provided.
Considering the modifications suggested by FEMA 368, the base shear
coefficient for impulsive mode, is given by
(0.6
(Cs ) i =
SDs
Ti + 0.4SDS )
T0
1.4(R / I)
for 0 < Ti <T0
(4)
S DS
1.4(R / I)
for T0 Ti < Ts
(5)
S D1
1.4(R / I)Ti
for Ti Ts
(6)
T0 = 0.2SDS/SD1 and Ti is time period of impulsive mode. All other parameters

are as defined in earlier sections. Similarly the base shear corresponding to
convective component is given by Vc = (Cs)c Wc, where Wc is convective mass
and base shear coefficient, (Cs)c is given by
(Cs)c =
6S D1 I
Tc2
for all values of Tc
(7)
Here, Tc is the time period of convective mode. In convective mode, damping is

taken as 0.5%. It is relevant to note that response modification factor, R does
not appear in the expression for convective mode base shear coefficient. This
implies that value of R is taken as unity i.e. convective base shear corresponds
to pure elastic response.
Variation of base shear coefficient for impulsive and convective modes
with time period is shown in Figure 6. Since ACI 350.3 forces correspond to
working stress design, they were multiplied by 1.4 to bring the forces to the
level of strength design or ultimate load design. It may be noted that FEMA 368
suggests importance factor, I =1.25 for tanks and there is no lower bound limit
on base shear coefficient in eqs. 4 to 6. In Figure 7, ratio of base shear coefficient
of impulsive mode and base shear coefficient of a building i.e. BSCtank / BSCbldg
is plotted. The base shear coefficient of building is obtained as per IBC 2000
with R = 8. It is seen up to T = 2 s, for a tank with R = 1.5, design base shear is 7
times higher than that of a building. However, for longer time period (i.e. T > 2
s), severity of tank shear as compared to that of a building reduces. This is due
to the fact that for tanks no lower bound on design seismic force has been
defined in ACI 350.3, whereas IBC 2000 specifies a lower bound on design
seismic force for buildings.
For convective mode, ACI 350.3 and FEMA 368, specify 0.5% damping
and for impulsive mode damping is 5%. Convective mode period is usually
grater than 2.0 sec and hence from eq. (6) and (7) it is seen that convective
mode spectrum (0.5% damping) is 8.4R/T times higher than impulsive mode
spectrum (5% damping).
4. AWWA STANDARDS
American Water Works Association (AWWA) standards provide
guidelines for design and manufacturing of different types of water storage
tanks. AWWA D-100 (1996) deals with welded steel tanks, AWWA D-103
(1997) is for factory-coated bolted steel tanks. Similarly, AWWA D-110 (1995)
deals with wire- and strand- wound, prestressed concrete water tanks and
AWWA D-115 (1995) is for prestressed concrete water tanks with

circumferential tendons. All these AWWA standards deal with circular tanks
only and quantification of seismic loads provided in them is in a fashion
different from IBC 2000. To bring these quantifications in line with IBC 2000,
FEMA 368 and its commentary (FEMA 369), has provided modifications to the
expressions for base shear coefficients of these AWWA standards. The
provisions of these AWWA standards will be discussed along with the
modifications of FEMA 368.
4.1 AWWA D-100 (1996) and D-103 (1997)
AWWA D-100 and AWWA D-103 deal respectively with welded steel
tanks and factory-coated bolted steel tanks. AWWA D-100 has provisions for
ground supported as well as elevated water tanks, whereas AWWA D-103 has
provisions for ground-supported tanks only. Provisions of AWWA D-100 and
D-103 for seismic design of ground-supported tanks are identical. First the
provisions for ground-supported tanks will be discussed.
4.1.1 Ground supported tanks
In Table 3, expressions for base shear coefficients for ground-supported
tanks as given in AWWA D-100 and D-103 are given along with the modified
expressions of FEMA 368. It is seen that for ground supported steel tanks, both
impulsive and convective modes are considered. For impulsive base shear a
constant value independent of time period is given. Since ground supported
steel tanks will mostly have time period in the constant-acceleration range of
spectra, hence it suffices to mention the constant value. It may be noted here
that AWWA D-100 specifies the base shear coefficient for working stress design
and hence the modified expression given in FEMA 368 contains a factor of 1.4
in the denominator. In AWWA standards, importance factor I is taken as I =
1.25. Further in AWWA D-100, the response reduction factor, Rw varies in the
range from 3.5 to 4.5, whereas in FEMA 368, it varies from 2.5 to 3.0. To have
better understanding of modifications suggested by FEMA 368, it will be
appropriate to compare the numerical values of base shear coefficients
10
obtained from AWWA D-100 and FEMA 368. In the most severe zone of
AWWA D-100, for the lowest and highest value of response reduction factor
(i.e. Rw = 3.5 and 4.5) the impulsive base shear coefficient turns out to be 0.36
and 0.28 respectively. The corresponding values of base shear coefficient as per
the modified expression of FEMA 368, are 0.357 and 0.298 respectively. The
parameters of most sever zone of AWWA D-100 are: Z = 0.4, S = 1.5 and soil
type C. The equivalent parameters of FEMA 368 are: Ss = 1.5, S1 = 0.6, Fs = 1.0,
Fv = 1.5, SDS = 1.0, SD1 = 0.6, Ts = 0.6, site class D.
Regarding the convective mode base shear coefficient it is interesting to
observe that in AWWA D-100, convective mode base shear coefficient depends
on the response reduction factor. It may be noted that in the provisions of ACI
350.3, the convective base shear is independent of response reduction factor. A
comparison of convective base shear coefficient obtained from AWWA D-100
and FEMA 368 for the lowest and highest values of response reduction factor is
shown in Figure 8. It is seen that the values obtained from AWWA D-100
expressions are on much higher side. It appears that due to modified
expressions of FEMA 368, the convective base shear values have reduced. It is
relevant to note here that, as per FEMA 368, the modified expression for
convective base shear contains a factor of 1.4 in the denominator to bring the
base shear values to working stress level. In this context, one may observe that
while suggesting modified expression for convective base shear of ACI 350.3,
which is also based on working stress method, FEMA 368 did not use a factor
of 1.4 (Table 2). Thus, while modifying the convective base shear expression of
AWWA D-100, if a factor of 1.4 is not used in FEMA 368, then the results of
AWWA D-100 and FEMA 368 will match well.
4.1.2 Elevated Tanks
For elevated tanks, AWWA D-100 does not consider sloshing effect, and
only the impulsive mode is considered. The impulsive base shear is given by V
= Cs.W, where, W is the summation of weight of water, container and support
structure above the base. The expression for base shear coefficient of elevated
tank is given in Table 3, along with the modified expression from FEMA 368. It
11
is seen that as per AWWA D-100, response reduction factor for elevated tank is
in the range of 3.0 to 4.0, whereas in FEMA 368 it varies from 2.0 to 3.0. A
comparison of the base shear coefficient obtained from AWWA D-100 and
FEMA 368 is shown in Figure 9. These results correspond to the most sever
zone and are for the lowest and highest values of response reduction factors. In
the short period range (i.e., T < 0.6 s) base shear coefficient corresponding to
the highest response factor (i.e., Rw = 4 and R = 3) match well. However, for
the lowest value of response reduction factor, the FEMA 368 results are on
lower side by 15%. In the long period range (i.e., T > 0.6 sec), FEMA 368 results
are on lower side.
The modified expression for base shear given in FEMA 368, is in the
same fashion as that of IBC 2000. In figure 10, the base shear coefficient of
elevated tank obtained from the modified expression of FEMA 368, is
compared with the base shear coefficient of a ductile building (with R = 8, I =
1.0 as per IBC 2000). In this figure, the ratio of BSCtank/BSCbldg is plotted for
two values of response reduction factor of tanks, i.e., R = 2 and 3.0. Since IBC
2000 provisions correspond to strength design, the FEMA 368 values of base
shear coefficient are multiplied by 1.4 to bring them to strength design level. It
is seen that base shear of elevated tanks is about 3 to 5 times higher than that of
a building up to T = 1.5 sec. In the long time period range, i.e. T > 1.5 sec, the
ratio of tank to building base shear reduces which is due to no lower bound
limit on tank base shear in the long period range.
4.2 AWWA D-110 (1995)

AWWA D-110 and AWWA D-115 deal with ground supported
prestressed concrete tanks, but since their provisions for seismic design are
quite different, these two standards will be discussed separately. First,
provisions of AWWA D-110 are discussed, and the next section will describe
provisions of D-115.
Expressions for impulsive and convective base shear coefficients, as per
AWWA D-110 are given in Table 4, along with the modified expressions of
12
FEMA 368. The base shear coefficients given in AWWA D-110 correspond to
working stress method. A comparison of the impulsive base shear coefficient
from AWWA D-110 expressions and modified expression of FEMA 368, is
shown in Figure 11a. This comparison is shown for the highest and lowest
values of response reduction factors. As per AWWA D-110, the values of
response reduction factor are in the range of 2.0 to 4.5, whereas in FEMA 368
these values are in the range of 1.5 to 3.0. It is seen that for the highest value of
response reduction factor (i.e., Rw = 4.5 and R = 3.0), base shear coefficient
values from AWWA D-110 and FEMA 368 match well in the short period range
(i.e., T < 0.6 sec). However, for the lowest value of response reduction factor
(i.e., Rw = 2.0 and R = 1.5), base shear coefficient values from FEMA 368 is 15%
less than that of AWWA D-110. In the long period range (i.e., T > 0.6 sec), base
shear coefficient values obtained from FEMA 368 are on lower side for the
highest as well as lowest response reduction factor.
Regarding the modified convective base shear expressions of FEMA 368,
there are certain inconsistencies. As per AWWA D-110, convective base shear
coefficient does not depend on response reduction factor (i.e., Rc = 1.0 for all
types of tanks). In section 14.7.3.7.3 of FEMA 368, on page no. 311 it is
mentioned that the modified convective base shear coefficient shall be taken as
(C s ) c =
6SD1I
Tc
. However, on page no. 312 it is stated that the modified
6SD1I
convective base shear is given by Vc =
Wc . Clearly there is
2
1.4RTc
inconsistency in the two expressions, second one considers a factor of 1.4R in

the denominator and the first one does not. Further, in FEMA 369 (pp 366),
which is commentary to FEMA 368, the modified base shear expression is also
given as (Cs ) c =
6SD1I
Tc
Since in AWWA D-110 the convective base shear
coefficient does not depend on response reduction factor, in Table 4, the

modified base shear coefficient is taken as (Cs ) c =
6SD1I
Tc
. A comparison of
13
convective base shear coefficient obtained from AWWA D-110 and FEMA 368
is shown in Figure 11b. It is seen that base shear coefficient values obtained
from FEMA 368 about 1.5 times higher than those obtained from AWWA D110.
Impulsive base shear coefficients given by the modified expressions of
FEMA 368 are in line with those given by IBC 2000 for buildings. Hence these
expressions of FEMA 368 can be used to compare the base shear coefficient of
tank with that of a ductile building (R = 8 and I =1 as per IBC 2000). Such a
comparison is shown in Figure 12, wherein, ratio of BSCtank/BSCbldg is plotted
for three different values of response reduction factor of tanks. While plotting
these results, the base shear coefficient of tank is multiplied by 1.4 to bring
them to strength design level. Again, due to absence of a lower bound limit on
base shear coefficient of tanks, the values of BSCtank/BSCbldg decreases in long
period range, i.e., for T > 2.0 sec.
4.3 AWWA D-115 (1995)

This AWWA standard deals with circular prestressed concrete tanks
with circumferential tendons. Expressions for impulsive and convective base
shear coefficients are given in Table 5 along with the modified expressions of
FEMA 368. It will be appropriate to note the differences in the base shear
coefficient expressions of AWWA D-110 and AWWA D-115. In AWWA D-115,
the response reduction factor varies from 1.0 to 3.0 whereas in AWWA D-110 it
varies from 2.0 to 4.75.
A comparison of impulsive base shear coefficient obtained from AWWA
D-115 and modified expression of FEMA 368 is shown in Figure 13, for extreme
values of response reduction factors. It is surprising to observe a large
difference in the base shear coefficient values obtained from these two
expressions. The values obtained from FEMA 368 expressions are on lower side
and there appears to be some inconsistency in the expressions of AWWA D-115
and the modified expression of FEM 368. In this context, it is to be noted that
FEMA 368 suggests same modified expression for impulsive base shear
14
coefficient of AWWA D-110 as well as AWWA D-115. Thus as per FEMA 368,
the values of response reduction factor are identical for AWWA D-110 as well
as AWWA D-115. However, as pointed out earlier, response reduction factor
values in AWWA D-110 and AWWA D-115 are quite different. Thus, it appears
that while suggesting the modified base shear expressions, FEMA 368 has not
considered this difference in the range of values of response reduction factors
of AWWA D-110 and D-115.
Another major inconsistency noted is that in AWWA D-115 convective
base shear coefficient depends on response reduction factor, whereas in FEMA
368 the corresponding modified expression does not have any dependence on
response reduction factor.
Due to this, it is not possible to compare the
convective base shear coefficient values obtained from AWWA D-115 and
FEMA 368. Again it may be noted that the FEMA 368 expression for convective
base shear coefficient is same for AWWA D-110 and D-115, and in AWWA D110 also convective base shear does not depend on response reduction factor.
It is quite clear that due to the inconsistencies mentioned above, no
comparison can be done between the impulsive base shear coefficient of tank
obtained from modified expression from FEMA 368 and the base shear
coefficient of building.
Thus, a review of provisions of AWWA standards and their
corresponding modifications in FEMA 368 reveals quite a few inconsistencies
and contradictions. These are summarized below:
(a) In AWWA D-100 and D-115 convective base shear coefficient depends on
response reduction factor whereas, in AWWA D-110 this coefficient does
not depend on response reduction factor.
(b) Though AWWA D-110 and D-115 deal with prestressed concrete tanks, they
use quite different values of response reduction factors. Moreover, in the
modified expression of FEMA 368, only one range of value of response
reduction factor is suggested for AWWA D-110 and D-115. This led to a
large difference in the values of base shear coefficients obtained from
AWWA D-115 and FEMA 368.
15
(c) In AWWA D-115, convective base shear depends on response reduction

factor, however, the base shear given by the corresponding modified
expression of FEMA 368 does not depend on response reduction factor.
(d) All AWWA codes are based on working stress method. While modifying
the convective base shear coefficient of AWWA D-100, the FEMA 368 has
considered a factor of 1.4 in the expression; however this factor is not
present in the modifications suggested for AWWA D-110 and D-115. In this
context it may be noted that ACI 350.3 is also based on working stress
method, and while modifying its convective base shear expression, FEMA
368 has not considered a factor of 1.4 in the expression (Table 2).
5. API 650 (1998)

American Petroleum Institute (API) has two standards namely, API 650
and API 620 which provide provisions for design and construction of
petroleum steel tanks. These API standards deal with ground supported
circular steel tanks only. The seismic design provisions in these API standards
are identical and hence provisions of only API 650 are described here. API 650
considers impulsive as well as convective component in the seismic analysis.
The base shear corresponding to impulsive and convective components is
respectively given by Vi = Ci Wi and Vc = Cc Wc. Here, Wi and Wc are impulsive
and convective masses respectively and likewise, Ci and Cc are base shear
coefficients corresponding to these components. In API 650, quantification of
base shear coefficients is in a manner different than IBC 2000. Hence, FEMA
368 has suggested modifications to base shear coefficients of API 650, to make
them consistent with IBC 2000. Table 6 provides details of base shear
coefficients given in API 650 along with the modified expressions given by
FEM 368
For impulsive mode, the base shear coefficient is defined as a constant
value which does not depend on the time period. It may be recognized here
that ground supported steel tanks will generally have time period in the
constant-acceleration range of spectra and hence it suffices to specify the
16
magnitude of the spectra in constant-acceleration range. It may be recalled here

that AWWA D-100 has also specified a constant value of base shear coefficient
for ground supported steel tanks. In API 650 and FEMA 368 the value of
importance factor is taken as 1.25. API 650 prescribes seismic forces at working
stress level, and they are to be multiplied by a factor of 1.4, to bring them
strength design level. Thus, as per API 650, Ci = 0.84 ZI and as per FEMA 368,
at strength design level, Ci = 0.336SDSI, which when compared with IBC 2000
expression (i.e. SDSI/R), implies that for ground supported steel tanks, R =
2.976 3.
For the most severe zone of API 650, Z = 0.4 and for S3 type site, S = 1.5.
For equivalent seismic condition, parameters from FEMA 368 will be SD = 1.5,
Fa = 1.0, S1 = 0.6 and Fv = 1.5, SDS = 1.0, with site class D. For these parameters,
the value of Ci, at strength design level, will be 0.42 from API 650 as well as
FEMA 368. This value of base shear coefficient is about 3.4 times higher than
the value of base shear coefficient of a building (obtained as per IBC 2000 with
R = 8). Comparison of convective mode base shear coefficient, at strength
design level, obtained from API 650 and FEMA 368 is shown in Figure 14. It is
seen that FEMA 368 results are about 20% less than those of API 650.
6. EUROCODE 8 (1998)
Five parts of Eurocode 8 (1998) cover provisions for seismic design of
various types of civil engineering structures. Part-4 of Eurocode 8 deals with
tanks, silos and pipelines. This code describes in detail about dynamic
modeling of convective and impulsive components, and also discusses other
aspects like base uplifting of unanchored tanks, soil-structure interaction etc.
The seismic action is specified in terms of response spectrum. In this
code, behavior factor, q, accounts for energy dissipation capacity of the
structure, mainly through ductile behavior and other mechanisms. For elastic
structures, q = 1.0 and for structures with good energy dissipation capacity, q =
5.0. Eurocode 8 specifies two types of response spectrum, first one is elastic
spectrum, Se(T), (for q =1.0) and second is spectrum for linear analysis, Sd(T), (for
17
q > 1.0). For tanks on ground, elastic spectrum is to be used, i.e., behavior factor,
q = 1.0. For buildings with ductile frames, behavior factor q can be as high as q
= 5.0, and spectrum for linear analysis is used. Seismic base shear for building is
given by V = I Sd(T) W, where, I is the importance factor and Sd(T) is the
spectrum acceleration at time period T. Expression for base shear of tank is
same as that of building, except that instead of spectrum for linear analysis Sd(T),
elastic spectrum Se(T) is to be used. The expressions for elastic spectrum and
spectrum for linear analysis are given in Table 7. A closer look at these
expressions reveals that elastic spectrum depends on damping correction factor
, whereas, spectrum for linear analysis does not depend on . Further, it can be
seen that there is a lower bound limit on values of spectrum for linear analysis in
long period range. There is no such lower bound limit on elastic spectrum.
It may be noted that Eurocode 8 provides indicative values of behavior
factor, q and various parameters defining shape of spectrum. However, there is
no indication on maximum value of design ground acceleration, ag, which
corresponds to a reference return period of 475 years. It mentions that National
Authorities of the member countries should arrive at suitable values of ag for
various seismic zones. In the present study, value of ag = 0.3g (i.e., = 0.3) is
assumed.
Figure 15 shows variation of base shear coefficient with time period for
ductile building (q= 5.0) and impulsive mode of tank. These results correspond
to S =1.0, = 2.5, = 1.0, K1 =1.0, K2 =2.0, Kd1 =2/3, Kd2 =5/3, TB =0.15, TC =0.6
TD = 3.0 and = 0.3. For buildings, I =1.0 and I =1.2 for tanks with high risk to
life, and large environmental, economic and social consequences. Sub soil class
B is considered which is similar to site class D of IBC 2000. In Figure 15, base
shear coefficient for convective mode with 0.5% damping is also shown. It is to
be noted that for buildings, i.e., when spectrum for linear analysis is used, there is
a lower bound limit on spectrum values, however for elastic spectrum no such
limit is specified. Further, it is important to note that shapes of Se(T) and Sd(T)
are also different beyond T = TB. Elastic spectrum, Se(T) reduces much faster with
time period than spectrum for linear analysis, Sd(T). Hence, for higher time
18
periods (T>0.6 s), the difference between base shear of tank and building is
reduced considerably.
This is clearly seen from Figure 16, wherein ratio
BSCtank / BSCbldg is plotted. Eurocode 8 specifies only one value of q (= 1.0) for
ground-supported tanks. Neither it mentions about different types of groundsupported tanks nor does it give specific values of q for different types of
ground-supported tanks. However, it states that if adequately demonstrated,
inelastic response (i.e., q > 1) can be considered.
For elevated tanks also, Eurocode 8 does not give very specific values of
q. It mentions that supporting structure may be designed to respond beyond
the yield level, thereby allowing energy dissipation in it. Elevated tanks with
simple support and which have little risk to life, negligible environmental and
social consequences due to failure, will have the value of q = 2.0. For elevated
tanks under higher risk category, the selected value of q should be properly
substantiated and proper ductility be provided through ductile design of
supporting structure.
For convective mode, in all types of tanks, the value of behavior factor, q
=1.0 and damping value, = 0.5% is suggested. For this value of
, the
damping correction factor is 1.673. As a result of this, elastic spectrum

corresponding to convective mode (0.5% damping) turns out to be 1.673 times
higher than that for impulsive mode (5% damping).
7. NZSEE GUIDELINES AND NZS 4203 :1992

In New Zealand, seismic design of liquid storage tanks follow NZSEEs
document (Priestley, et. al., 1986). A study group of NZSEE is presently
revising this document to bring it in line with the NZS 4203:1992, the New
Zealand loading code for buildings. Details of the proposed seismic loading are
given by Whittaker and Jury (2000). This section discusses the proposed
seismic loading along with the seismic loading provisions for building in NZS
4302:1992.
As per NZS 4203:1992, seismic base shear for building is given by V = CW,
where, lateral force coefficient or base shear coefficient, C is given by
19
C = Ch(T, ) Sp R Z Lu
(7)
0.03
where
Ch(T, ) = Basic seismic hazard coefficient which accounts for different soil
conditions, ductility factor and natural period T of building,
Sp
= Structural performance factor. For buildings Sp = 0.67
= Risk factor. For ordinary buildings R =1.0
= Zone factor as per seismic map
Lu
= Limit state factor for ultimate limit state. Lu = 1.0
NZS 4203:1992 has provided tabulated values of Ch(T, ) for three different
soil conditions, eight different values of ( from = 1.0 to 10.0) and values of
time period T from T = 0.0 to 4.0 sec. For buildings with ductile frames the
value of could vary from = 6 to10.
As per Whittaker and Jury (2000), seismic base shear for tanks is V=CW,
where
C = Ch(T, 1) Sp R Z Lu ) Cf(, )
(8)
Eq.(8) differs from Eq.(7) in two ways: First, basic seismic hazard coefficient,
Ch(T,1) corresponds to =1.0, i.e., purely elastic spectrum is used and secondly,
an additional factor, Cf(,) termed as correction factor is included. This
correction factor accounts for ductility and level of damping. For tanks,
performance factor Sp=1.0 is recommended as opposed to 0.67 for buildings.
Value of risk factor, R for tanks is arrived at by considering four aspects,
namely, risk to number of persons, risk to environment, community
significance of the tank and value of adjacent property. Value of R can vary
from 0.5 to 1.6 depending on the risk associated with the tank and a tank with
serious risk has R = 1.3. The value of damping depends on material of tank
shell and supporting soil. Whittaker and Jury (2000) have provided values of
ductility factor, for different types of tanks, which are shown in Table 8.
Similarly, damping level, for different types of concrete and steel tanks are
also provided. Further, provisions are made for increasing the damping values
of tank depending on flexibility of supporting soil, i.e., to consider radiation
20
damping in soil. For different values of ductility factor, and damping level, ,
values of correction factor Cf (, ) are provided as shown in Table 9.
For three different values of Cf, namely, Cf=0.72, 0.54 and 0.38, variation
of base shear coefficient with time period is plotted in Figure 17. These results
are for the most severe zone (Z =1.2) and site subsoil category (C), i.e., flexible
and deep soil condition, which is similar to site class D of IBC 2000. Also
plotted in this figure is the variation of base shear coefficient for a building with
= 6.0. It would be appropriate to note that variation of Ch(T,6) and Ch(T,1)

with time period is of different nature. This gets reflected in the variation of
base shear coefficient of tanks and building with time period. For T > 0.6s,
reduction in Ch(T,1) with time period is slower than corresponding reduction
in Ch(T,6). This implies that elastic spectrum (Ch(T,1)) is more flatter than
inelastic spectrum (Ch(T,6)). Moreover values of Ch(T, 1) and Ch(T,6) at various
time period do not have a constant ratio.
From the ratio of BSCtank / BSCbldg shown in Figure 18, it is seen that the
ratio of base shear coefficients increases at around T= 0.6s, then remains almost
constant up to T=3.0s and for T>3.0s it starts decreasing. The increase is due to
the fact that values of Ch(T, 6) (i.e. for building) and Ch(T, 1 ) (i.e., for tank) are
not in same proportion for values of T from 0.6-3.0s. The decrease in ratio of
base shear for higher value of T (i.e., T > 3.0s) is due to the fact that there is a
lower bound on value of C (i.e., C cant be less than 0.03) for buildings, but for
tanks (i.e., elastic case of =1.0) there is no such lower bound limit.
For elevated tanks, Whittaker and Jury (2000) do not provide specific
information on ductility factor, . However, it mentions that for elevated tanks,
ductility factor as appropriate for support structure should be considered. This
may imply that if supporting structure is quite ductile then value of can be as
high as for buildings (i.e., = 6 to 10). At the same time it should be noted that
in Table 9, values of response modification factor, Cf is given only for
maximum value of = 4.0, i.e., for values of greater than 4.0, response
modification factor is not available.
21
Whittaker and Jury (2000) have specified 0.5% damping for convective
mode. Thus, in Table 9, values of Cf(,) corresponding to = 0.5, can be used
for convective mode. These values of Cf(,) change with ductility factor, ,
implying that convective mode base shear will vary with ductility of tank. It
may be noted that in AWWA D-100 also convective base shear depends on
ductility of tank i.e. on response reduction factor. However, in ACI 350.3 and
Eurocode 8, convective mode shear does not depend on ductility of tank and it
corresponds to pure elastic case, i.e., R = 1, and q =1.0, respectively.
8. COMPARISON OF DESIGN FORCES FROM VARIOUS CODES

In the previous sections provisions related to base shear coefficients for
tanks as given in various codes, standards and guidelines are described. It will
be interesting to compare base shear coefficients obtained from these
documents. Before comparing the results for tanks, a comparison of base shear
coefficient for a ductile building is shown in Figure 19. In this figure, base shear
coefficients of building (BSCbldg), obtained from IBC 2000, Eurocode 8 and NZS
4203:1992 are shown. These results correspond to the most severe zone of each
code. It is seen that in the short period range (i.e., T=0.1-0.6s), results from
Eurocode 8 and NZS 4203 match well. In this short period range, IBC 2000
results are on lower side by about 15%. Further, all the three codes have
different shape of spectra in constant-velocity range (i.e., T>0.6s). Moreover,
magnitude of the lower bound limit on spectra is also seen to be different in
these codes. To obtain similar comparison for tanks, first of all, for a particular
type of tank, all the relevant parameters (such as R, q, Cf) from different codes
will have to be identified. It is seen that most of the codes consider ground
supported unanchored concrete water tank as a low ductility tank or a tank
with low energy absorbing capacity. For such a tank the relevant parameters
will be as shown in Table 10. In Figure 20, comparison of base shear coefficient
for this tank (BSCtank) obtained from different codes is shown. It may be noted
here that FEMA 368 has modified the base shear expressions of ACI 350.3 and
AWWA D-110 and brought them in line with IBC 2000. In view of these
22
modifications, parameters from ACI 350.3, AWWA D-110 and IBC 2000 are
same. From Figure 20 it is seen that in the short period range (T<0.6s),
Eurocode 8 results are 10% higher and NZSEE results are 35% higher than the
one obtained from IBC2000. Further, it can also be seen that except for IBC
2000, no other code has lower bound limit on base shear coefficient in long
period range. Comparison of ratio of base shear coefficient of tank and building
(BSCtank/BSCbldg) is shown in Figure 21. Here, base shear coefficient of tank
from a particular code is divided by corresponding base shear coefficient of a
ductile building. It is seen that from T=0.1-0.6s, this ratio is constant for all the
codes. This constant value is 6 for Eurocode 8 and for IBC and NZSEE it is 6.7
and 7.3 respectively. The decrease in the value of this ratio for T>0.6s for the
case of Eurocode 8, is due to difference in shapes of spectrum used for tank and
building. Another factor contributing to this decrease, particularly in higher
period range, is absence of lower bound limit on spectral values for tanks. The
decrease in the value of this ratio in long period range, for NZSEE, ACI 350.3
and AWWA D-110 is also attributed to similar reasons. For the case of IBC
2000, due to lower bound limit on spectral values for tanks, the ratio of tank to
building shear does not fall below the value of 4, even in long period range.
Results similar to one presented in Figure 21, can be obtained for a high
ductility tank, i.e., a tank with high energy absorbing capacity. For such a tank,
various parameters of different codes are given in Table 11. These parameters
can as well be applicable to some of the elevated tanks. For Eurocode 8, value
of q = 2 is considered, which is suggested for a low risk category elevated tank
with simple type of supporting structure. Results on ratio of base shear
coefficient of tank to building, (BSCtank / BSCbldg), are shown in Figure 22. It is
seen that maximum value of this ratio is about 3 to 4 in all the codes, as against
a value of 6 to 7 for low ductility tanks. This implies that design base shear for
a low ductility tank is double that of a high ductility tank. Variation in the ratio
of base shear of tank and building, in the higher time period range is seen in
Figure 22 also, which is due to reasons discussed earlier.
23
On similar lines, comparison of convective base shear coefficients

obtained from various codes is shown in Figure 23. Before looking into this
results it will be appropriate to recognize that as per AWWA D-100, AWWA D115 and NZSEE (i.e., Whittaker and Jury (2000)), the convective base shear
coefficient depends on response reduction factor or ductility factor. However,
as per ACI 350.3, AWWA D-110, API 650 and Eurocode 8 the convective base
shear coefficient does not depend on response reduction factor. The
comparison shown in Figure 23 considers only those documents in which
convective base shear coefficient does not depend on response reduction factor.
Hence in Figure 23, results of ACI 350.3, AWWA D-110, API 650 and Eurocode
8 are shown. It can be seen that convective base shear coefficient obtained from
Eurocode 8 and API 650 match well. However, ACI 350.3 gives very high
values of convective base shear coefficient. For time period greater than 3.0s,
the results of ACI 350.3 are about 2.5 times higher than that of API 650.
9. PROVISIONS OF INDIAN CODE

Indian Standard IS:1893-1984 provides guidelines for earthquake
resistant design of several types of structures including liquid storage tanks.
This standard is under revision and in the revised form it has been divided into
five parts. First part, IS 1893 (Part 1):2002, which deals with general guidelines
and provisions for buildings has already been published. Second part, yet to be
published, will deal with the provisions for liquid storage tanks. In this section,
provisions of IS:1893-1984 for buildings and tanks are reviewed briefly
followed by an outline of the changes made in IS 1893 (Part 1):2002.
In IS: 1893-1984, base shear for building is given by V = CsW, where, Cs
is the base shear coefficient given by
Cs = K C I o
(9)
Here, K = Performance factor depending on the structural framing system and

brittleness or ductility of construction; C = Coefficient defining flexibility of
structure depending on natural period T; = Coefficient depending upon the
soil-foundation system; I = Importance factor; o = Basic seismic coefficient
24
depending on zone. For buildings with moment resisting frames, K = 1.0.

Importance factor, for buildings is usually I = 1.0.
IS:1893-1984, does not have any provision for ground supported tanks. It
has provisions for elevated tanks, for which it does not consider convective
mode. Base shear for elevated tank is given by V = Cs W, where, base shear
coefficient, Cs is given by
Cs = I Fo (Sa/g)
(10)
Here, Sa/g = average acceleration coefficient corresponding to the time period

of the tank, obtained from acceleration spectra given in the code; Fo = seismic
zone factor; W = weight of container along with its content and one third
weight of supporting structure. For elevated tanks, Importance factor I = 1.5. It
may be noted that in the expression for base shear coefficient of tank, the
performance factor K does not appear, i.e. K = 1 is considered, which is same as
that for a building with ductile frame. This implies that in IS:1893-1984, there is
no provision to account for lower ductility and energy absorbing capacity of
elevated tanks. Thus, as per IS:1893-1984, base shear coefficient for tank will be
only 1.5 times higher than that for a building, which is due to higher value of
importance factor. This is in contrast to other codes, reviewed in earlier
sections, wherein tank base shear coefficient is seen to be 3 to 7 times higher
than buildings. This lacunae needs to be corrected in the next revision of the
code.
As mentioned earlier, IS 1893 is under revision and first part, of the
revised code, IS 1893 (Part 1): 2002, has already been published. In this revised
code, base shear for building is given by V = Cs W, and base shear coefficient Cs
is given by
Cs =
ZI
(Sa / g)
2R
(12)
where Z = zone factor, I = importance factor, R = response reduction factor and

Sa/g = average response acceleration coefficient, obtained from acceleration
spectra given in the code. For buildings with ductile frames value of R is 5. In
Figure 24, a comparison of base shear coefficients for building obtained from
25
IS:1893-1984 and IS 1893 (Part I):2002 is shown, along with the base shear
coefficient from IBC 2000. Since IS:1893-1984, does not specify specific value of
load factors for strength design, the results in Figure 24 are presented for
working stress level. It is seen that base shear coefficient from IS:1893-1984 is
lower than one from IS 1893 (Part I):2002. Further, unlike IBC 2000, there is no
lower bound in IS 1893 (Part I):2002 and IS:1893-1984.
Subsequent parts of IS 1893:2002, will be using acceleration spectra
given in Part I, and will be based on same design philosophy. Thus, for liquid
storage tanks, base shear coefficient will be given by Eq.(12), in which suitable
values of R will have to be used for different types of tanks. From the review
presented in earlier sections, it is seen that low and high ductility tanks have
design base shear 3 to 7 times higher than ductile buildings. In Figure 25, base
shear coefficients for low and high ductility tanks, from IBC 2000 (i.e., tanks
with R=1.5 and R=3.0) are shown. To achieve this level of base shear
coefficients the value of R in IS 1893 (Part 1):2002 should be 1.1 and 2.25 as can
be seen from Figure 25. Also shown in this figure is the base shear coefficient
for tank obtained from IS:1893-1984, which is on much lower side. Based on the
comparison shown in Figure 25, proposed values of R which can be used in IS
1893 (Part 2):2002 for different types of tanks, are given in Table 12.
10. DISCUSSION
Due to low ductility and energy absorbing capacity, liquid storage tanks
are generally designed for higher seismic forces as compared to conventional
buildings. In this article, provisions of various codes on design seismic forces
for tanks are reviewed.
It is found that there is considerable variation in the types of tanks
described in various codes. For example, ground supported tanks described in
IBC 2000, ACI 350.3, AWWA D-110 and NZSEE guidelines are having different
types of base conditions. Eurocode 8 does not provide any details about base
supports of ground-supported tanks. Provisions on elevated tanks are
described in IBC 200, ACI 371, ACI 30.3, AWWA D-100, Eurocode 8 and
26
NZSEE guidelines. Less information is available on energy absorbing capacity

of different types of supporting towers of elevated tanks. Less frequent use of
elevated tanks in these countries may be one reason for low emphasis on
elevated tanks in these codes.
All the codes, consider impulsive and convective modes of vibration in
the seismic analysis of ground-supported tanks. The level of design seismic
force for a particular tank obviously depends on its ductility and energy
absorbing capacity. It is observed that for a tank with low ductility, impulsive
base shear coefficient (ratio of lateral force to weight) is 6 to 7 times higher than
the base shear coefficient of a ductile building; and for a high ductility tank this
value is 3 to 4 in all the codes (Figures 21 and 22). However, this is so only for
tanks with short time period (i.e., T<0.6s). Beyond this short period range, there
is considerable difference in the values of BSCtank/BSCbldg. For example, at
T=1.5s, for a tank with low ductility, the value of BSCtank/BSCbldg, as per
NZSEE guideline is 8.2 and as per IBC 2000 and Eurocode 8 this value is 6.7
and 4.4, respectively. Thus, Eurocode 8 results are on lower side by almost
50%. In fact, beyond T=0.6s, as per Eurocode 8, the value of BSCtank/BSCbldg,
decreases continuously, which is due to two reasons. First, in Eurocode 8,
elastic spectrum (used for tank) has much faster reduction with time period than
spectrum for linear analysis (used for buildings) in the constant-velocity range.
Secondly, unlike for buildings, there is no lower bound limit on spectrum used
for tanks.
In NZSEE guidelines [i.e., NZS 4230:1992 and Whittaker and Jury (2000)]
the elastic spectrum (used for tanks) reduction with time period is slower than
inelastic spectrum (used for buildings). Due to this reason, the NZSEE results in
Figure 21, show a slight increase in the value of BSCtank/BSCbldg at T=0.6s. In
IBC 2000, spectra used for tank and building, have same shape in constantvelocity range. Further, IBC 2000 specifies, lower bound limits on spectral
values for buildings as well as tanks. Hence, as per IBC 2000, values of
BSCtank/BSCbldg , do not fall below four even in the long period range. In case of
ACI 350.3 and AWWA D-110 also, there is no lower bound specified on
27
spectral values for tanks. This, as explained earlier, leads to loss of severity of
tanks shear in long period range as compared to that of buildings (Figures 21
and 22).
The reason for providing a lower value of response reduction factor for
tanks is their low ductility and low energy absorbing capacity as compared to
buildings. However, it is seen that in many codes, due to non-availability of
lower bound limit on spectral values of tanks, the ratio of tank to building base
shear reduces in long period range. ACI 371, which deals with shaft supported
elevated tanks, is a good example on provision of lower bound limit on spectral
values. Due to availability of suitable lower bound limit in ACI 371, the value
of BSCtank /BSCbldg does not reduce even in long period range (Figure 5). Thus,
there is a need to provide suitable lower bound limit on spectral values of tanks
in other codes. Absence of such a lower bound limit can lead to nonconservative estimate of base shear for tanks with longer time period,
particularly the elevated tanks on flexible supports.
For the case of ground supported steel tanks, API 650 and AWWA D-100
specify a constant value of base shear coefficient, which does not depend on
time period. For steel tanks, it is quite likely that impulsive time period will be
in the constant-acceleration range of spectra, and hence it suffices to specify
this constant value of base shear coefficient. However, in IBC 2000, the base
shear coefficient for ground supported steel tanks is not defined as a constant
value. Further, as per AWWA D-100 and IBC 2000, response reduction factor
for ground supported steel tanks changes with type of base support provided.
However, in API 650, tanks with different types of base supports are not
described. As per API 650, the base shear coefficient for ground supported steel
tank is about 3.4 times higher than that of a ductile building; and in AWWA D100 and IBC 2000, depending on type of base support, the base shear coefficient
of ground supported steel tank is 3 to 3.7 times higher.
While considering the convective base shear, all the codes suggest a
damping value of 0.5%. However, in the evaluation of convective base shear
coefficient, considerable differences are seen in the provisions of various codes.
28
Firstly, as per ACI 350.3, API 650, AWWA D-110 and Eurocode 8, convective
base shear coefficient does not depend on response reduction factor. However,
as per AWWA D-100, AWWA D-115 and NZSEE guidelines (i.e. Whittaker and
Jury (2000)), convective base shear coefficient depends on response reduction
factor. From the comparison presented in Figure 23, it is seen that base shear
coefficient evaluated as per ACI 350.3 is about two and half times greater than
the one obtained from AWWA D-110 and Eurocode 8.
For the elevated tanks, AWWA D-100 does not recommend
consideration of convective mode of vibration. However, Eurocode 8 and
NZSEE guidelines recommend consideration of convective mode. At the same
time, IBC 2000 and ACI 371 suggest that convective mode need not be
considered if certain conditions on weight of water and time period of
convective mode are met with.
As far as AWWA Standards are considered, there appears to be quite a
few inconsistencies in them and also in the modifications suggested for them in
FEM 368. These have been described in detail in section 4. There is a need to
properly address these inconsistencies.
In the context of Indian codes it is noted that design seismic forces for
buildings, as per revised Indian code (i.e., IS 1893 (Part 1):2002), compare well
with those specified in IBC 2000 (Figure 24). However, Indian code does not
have a lower bound limit on spectral values for buildings, which otherwise is
present in all the other codes. As far as liquid storage tanks are concerned,
Indian scenario is bit different. In India, elevated tanks are quite commonly
used in public water distribution systems and a large number of them are in
use. These tanks have various types of support structures, like, RC braced
frame, steel frame, RC shaft, and even masonry pedestal. Ground supported
tanks are used mainly by petroleum and other industrial installations. For
different types of elevated and base supports for ground-supported tanks,
values of response modification factor, R, to be used in Indian code are
proposed (Table 12). However, it is felt that for elevated tanks with different
types of supporting structures, a detailed investigation is needed to ascertain
29
their energy absorbing capacity and ductility characteristics. Similarly, suitable

values of lower bound limits on spectral values for buildings as well as other
types of structures, including tanks, needs to be arrived at.
11. CONCLUSIONS
Following conclusions are drawn from the comparative assessment of
provisions of different codes on seismic design of liquid storage tanks:
1) There is no uniformity in types of tanks described in various documents.
Most of the codes put emphasis on ground-supported tanks and very
limited information is available on elevated tanks.
2) All the documents suggest consideration of convective and impulsive
components in seismic analysis of tanks.
3) For a particular type of tank with short period (less than 0.6s), ratio of base
shear of tank and building is almost same in all the codes. This ratio is 6 to 7
for low ductility tanks and 3 to 4 for high ductility tanks. However, for
tanks with time period greater than 0.6s, there is a large variation in the
values of this ratio obtained from different codes. For example, at time
period of 1.5 sec, value of this ratio from Eurocode 8 is almost 50% less than
the one obtained from NZSEE guidelines. This is attributed to the use of
spectra of different shapes for buildings and tanks.
4) Unlike for buildings, most of the documents do not provide lower bound
limit on spectral values for tanks. This results in decrease in the ratio of base
shear of tank and building, in long period range. This effectively results in
reduction in severity of tank base shear as compared to building base shear.
5) Suitable provisions for lower bound limit on spectral values for tanks are
necessary. Only ACI 371, which deals with elevated tanks, and IBC 2000
have provisions for lower bound limit on spectral values of tanks.
6) Convective mode base shear values obtained from API 650 and Eurocode 8
match well, however one obtained as per ACI 350.3 is 2.5 times higher than
that of ACI 350.3.
30
7) There are quite a few inconsistencies among different AWWA standards.

These need to be resolved. Further modifications suggested by NEHRP
recommendations (FEMA 368) to provisions of AWWA codes are also
having certain inconsistencies.
8) Indian code needs to include provisions on lower bound limit on spectral
values of buildings and tanks. Further, provisions for inclusion of
convective mode of vibration in the seismic analysis of tanks also need to be
included. Based on the review of various international codes presented in
this paper, it is recommended that IS 1893 should have values of response
reduction factor in the range of 1.1 to 2.25 for different types of tanks.
Acknowledgements
This work has been supported through a project entitled Review of
Building Codes and Preparation of Commentary and Handbooks awarded to IIT
Kanpur by the Gujarat State Disaster Management Authority (GSDMA),
Gandhinagar through World Bank finances. The views and opinions expressed
therein are those of the authors and not necessarily those of the GSDMA or the
World Bank.
References
ACI 350.3, 2001, Seismic design of liquid containing concrete structures, An
American Concrete Institute Standard
ACI 350, (2001), Code requirements for environmental engineering concrete
structures, An American Concrete Institute Standard
ACI 371, 1998, Guide for the analysis, design and construction of concrete
pedestal water towers, An American Concrete Institute Standard
API 650, 1998, Welded storage tanks for oil storage, American Petroleum
Institute Standard, Washington, DC
AWWA D-100, 1996, Welded steel tanks for water storage, American Water
Works Association, Colorado
AWWA D-103, 1997, Factory-coated bolted steel tanks for water storage,
American Water Works Association, Colorado
AWWA D-110, 1995, Wire- and strand-wound circular, prestressed concrete
water tanks, American Water Works Association, Colorado
AWWA D-115, 1995, Circular prestressed concrete water tanks with
circumferential tendons, American Water Works Association, Colorado
31
Eurocode 8, 1998, Design provisions for earthquake resistance of structures,

Part 1- General rules and Part 4 Silos, tanks and pipelines, European
committee for Standardization, Brussels
FEMA 368, 2000, NEHRP recommended provisions for seismic regulations for
new buildings and other structures Part I: Provisions, Building seismic
safety council, National Institute of Building Sciences
FEMA 369, 2000, NEHRP recommended provisions for seismic regulations for
new buildings and other structures Part II: Commentary, Building
seismic safety council, National Institute of Building Sciences.
Hori,N., 1990, Effect of soil on dynamic response of liquid-tank systems,
Journal of Pressure Vessel Technology, ASME, 112, 118-123.
Housner,G.W., 1963, The dynamic behavior of water tanks, Bulletin of the
Seismological Society of America, 53 (2), 381-387.
IBC 2000, International Building Code International Code Council, 2000, Falls
Church, Virginia.
IS 1893 (Part 1):2002, Criteria for earthquake resistant design of structures,
Bureau of Indian Standards, New Delhi.
Malhotra,P.K., 1997, Seismic response of soil-supported unanchored liquidstorage tanks, Journal of Structural Engineering, ASCE, 123 (4), 440-450.
Malhotra,P.K., 2000, Simple procedure for seismic analysis of liquid-storage
tanks, Structural Engineering International, 3, 197-201.
NZS 4203:1992, Code of practice for general structural design and design
loading for buildings, New Zeeland.
Priestley,M.J.N., et. al., 1986, Seismic design of storage tanks,
Recommendations of a study group of the New Zealand National Society
for Earthquake Engineering
Veletsos,A.S., 1974, Seismic effects in flexible liquid storage tanks,
Proceedings of Fifth World Conference on Earthquake Engineering, Rome
Italy, 1, 630-639
Veletsos,A.S., Tang,Y., and Tang,H.T., 1992, Dynamic response of flexibly
supported liquid-storage tanks, Journal of Structural Engineering, ASCE,
118 (1), 264-283.
Whittaker,D. and Jury,D., 2000, Seismic design loads for storage tanks, Paper
No. 2376, 12th World Conference on Earthquake Engineering, New Zealand
Wozniak,R.S. and Mitchell,W.W., 1978, Basis of seismic design provisions for
welded steel oil storage tanks, American Petroleum Institute 43rd midyear
meeting, session on Advances in Storage Tank Design, Toronto, Canada.
32
Table 1: Various types of tanks and their R values as mentioned in IBC 2000
Type of tank
Ground supported tanks
Flat bottomed welded or bolted steel tanks Anchored
Flat bottomed welded or bolted steel tanks Unanchored
Reinforced or prestressed concrete tanks
with anchored flexible base
Reinforced or prestressed concrete tanks
with reinforced nonsliding base
Tanks with unanchored and unconstrained flexible base
Elevated tanks
On braced legs
On unbraced legs
On irregular braced legs single pedestal or skirt supported
Tanks supported on structural towers similar to buildings
R
3.0
2.5
3.0
2.0
1.5
3.0
3.0
2.0
3.0
33
Table 2: Expressions for base shear coefficients from ACI and FEMA 368
ACI 371
FEMA 368
S
1.2C V
D1
Cs =
for Ts < T < 4 s
Cs =
2/3
( R / I )T
RT
2.5Ca
4S D1
=
for T 4 s
R
( R / I)T 2
0.5 Ca
S DS
Values of coefficients, CV and Ca depend

( R / I)
on seismic zones; R is response
0.2 SDS
modification factor with maximum
R = 2.0 for pedestal supported elevated
value of 2.0
tanks
ACI 350.3
FEMA 368
Impulsive mode
Impulsive mode
S
2.75ZI
(Cs)i =
for Ti 0.31 s
(0.6 Ds Ti + 0.4S DS )
RW
T0
(Cs)i =
for 0 < Ti <T0
2.75ZI
1.25ZIS
1.4(R / I)
=
<
for Ti > 0.31 s
RW
R W Ti2 / 3
S DS
=
for T0 Ti < Ts
1.4(R / I)
S D1
=
for Ti Ts
1.4(R / I)Ti
Convective mode
Convective mode
1.875ZIS
6S D1 I
(Cs)c =
for Tc < 2.4 sec
for all values of Tc.
(Cs)c =
2/3
Tc
Tc2
< 2.75 ZI
Values of R are in the range of 1.5 to 3.0
6ZIS
for different types of tanks
=
for Tc 2.4 sec
2
Tc
Z is zone factor; S is soil profile
coefficient; I is importance factor and RW
is response modification factor. Values
of RW are in the range of 4.75 to 2.0 for
different types of tanks.
34
Table 3: Expressions for base shear coefficients from AWWA D-100, D-103 and
FEMA 368
AWWA D 100 and D103
(Ground supported tanks)
Impulsive mode
2.52 ZI
(C s ) i =
Rw
3ZIS
for Tc 4.5 s
R w Tc
13.5ZIS
R w Tc 2
Impulsive mode
(C s ) i =
Convective mode
(C s ) c =
FEMA 368
for Tc > 4.5 s
Z is zone factor; I is importance factor;

S is soil factor; Tc is time period of
convective mode; Rw is response
reduction factor with values in the
range of 3.5 to 4.5
AWWA D 100
(Elevated tanks)
1.25ZIS
( Cs ) i =
R wT2 / 3
0.75ZI
Rw
2.75ZI
Rw
Rw is in the range of 2.0 to 3.0
S DS I
1.4R
Convective mode
1.5S D1 I
(C s ) c =
1.4RTc
=
6S D1 I
1.4RTc 2
for Ts < Tc 4 s
for Tc > 4 s
I is importance facor; R is response

reduction factor with values in the
range of 2.5 to 3.0 for steel tanks.
FEMA 368
SDSI
for T < Ts
1.4R
S I
= D1
for Ts < Tc 4 s
1.4RT
4SD1I
for Tc > 4 s
=
1.4RT 2
R is in the range of 2.0 to 3.0
(C s ) i =
35
Table 4: Expressions for base shear coefficients from AWWA D-110 and
FEMA368
AWWA D110
Impulsive mode
1.25ZIS
(C s ) i =
2/3
R i Ti
2.75ZI
Ri
FEMA 368
Impulsive mode
S
(0.6 Ds Ti + 0.4SDS )
T0
for 0 < Ti
(Cs)i =
1.4(R / I)
<T0
SDS
=
for T0 Ti
1.4(R / I)
< Ts
SD1
=
for Ti Ts
1.4(R / I)Ti
Convective mode
Convective mode
6S I
4 ZIS
(Cs ) c = D21
(Cs ) c =
2
Tc
R cTc
Ri is response reduction factors for Values of R are in the range of 1.5 to
impulsive mode with values in the 3.0 for different types of concrete tanks
range of 2 to 2.75 and Rc is response
reduction factor for convective
mode; Rc = 1.0 for all types of tanks
36
Table 5: Expressions for base shear coefficients from AWWA D-115 and
FEMA368
AWWA D 115-95
Impulsive mode
1.25ZIS
(Cs ) i =
2/3
R w Ti
2.75ZI
Ri
Convective mode
ZIS
(C s ) c =
R w Tc
Rw is response reduction
factor with values in the
range of 1.0-3.0.
FEMA 368
Impulsive mode
S
(0.6 Ds Ti + 0.4SDS )
T0
(Cs)i =
for 0 < Ti <T0
1.4(R / I)
SDS
=
for T0 Ti < Ts
1.4(R / I)
SD1
=
for Ti Ts
1.4(R / I)Ti
Convective mode
6S I
(Cs ) c = D21
Tc
Values of R are in the range of 1.5 to 3.0 for
different types of concrete tanks
Table 6: Expressions for base shear coefficients from API 650 and FEMA 368
API 650
Impulsive mode
Ci = 0.6 Z I
Convective mode
0.75ZSI
Cc =
for Tc 4.5 s
Tc
3.375ZSI
for Tc > 4.5 s
=
Tc2
Z is zone factor; I is importance
factor; S is site coefficient
FEMA 368
Impulsive mode
Ci = 0.24 SDS I
Convective mode
0.6S DS I
for Ts < Tc 4.0 s
Cc =
Tc
2.396S DS I
=
for Tc > 4.0 s
Tc2
37
Table 7: Elastic spectrum Se(T)

Eurocode 8
T
Se(T) = S1 + (0 1)
TB
0 T < TB
= S0
TB T < TC
T
= S0 C
T
and Spectrum for linear analysis Sd (T)
of
T
Sd (T) = S1 + ( 0 1)
TB q
0 T < TB
= S 0
q
TB T < TC
K1
T
= S 0 C
q T
( 0.2 ) TC T < TD
Kd1
TC T < TD
K1
T T
= S 0 C D
TD T
TD T
K2
= S 0
q
TC

TD
Kd1
Kd 2
TD
T

( 0.2 ) TD T
where
=ag/g; ag=design ground acceleration ; 0=Spectral acceleration
amplification factor for 5% viscous damping; S = Soil parameter;
=Damping correction factor. = 1.0 for 5% damping. For any other
damping , the value of = {7/(2+)}0.5 ; K1, K2, Kd1, Kd2 = Exponents
which influence the shape of spectrum. Values of these exponents depend
on soil condition. Their values for different soil conditions are given in
Tables 4.1 and 4.2 of Eurocode 8 Part 4; TB, TC = Limits of the constant
spectral acceleration branch; TD = Value defining beginning of the constant
displacement range of the spectrum
38
Table 8: Different types of tanks with their ductility factor, (Whittaker and
Jury (2000))
Type of Tank
Steel Tanks on Grade
Elastically supported
Unanchored tank designed for uplift (elephant foot shell
buckling may occur under seismic overload)
Unanchored tank designed for uplift and elastic (diamond
shaped) shell buckling mode
Anchored with non-ductile holding down bolts
Anchored with ductile tension yielding holding down bolts
Ductile skirt pedestal
On concrete base pad designed for rocking
Concrete Tanks on Grade
Reinforced Concrete
Prestressed Concrete
Tanks of other materials on Grade
Timber
Non-ductile materials (eg. Fiberglass)
Ductile materials and failure mechanisms
Elevated Tanks
1.25
2.001
1.25
1.25
3.002
3.002
2.002
1.25
1.00
1.00
1.00
3.00
As
appropriate
for support
structure 3
Notes
1. Check that elastic buckling does not occur before elephant foot
2. Capacity design check required to protect against other forms of failure
3. Capacity design approach shall be used to protect elevated tanks against failure while
yielding occurs in the chosen support system
Table 9: Correction factor, Cf (Whittaker and Jury (2000))

Ductility
factor,
1.0
1.25
1.5
2.0
2.5
3.0
4.0
0.5
1.75
0.92
0.75
0.58
0.49
0.43
0.36
Damping level, (%)

1.0
2.0
5.0 10.0 15.0
1.57 1.33 1.00 0.80 0.71
0.88 0.83 0.72 0.62 0.58
0.72 0.68 0.61 0.54 0.51
0.56 0.54 0.48 0.44 0.42
0.48 0.46 0.42 0.38 0.36
0.43 0.41 0.38 0.35 0.33
0.36 0.35 0.33 0.30 0.29
20.0
0.67
0.55
0.48
0.40
0.35
0.32
0.28
39
Table 10: Parameters for a low ductility tank

Code
IBC 2000 and ACI 350.3
Eurocode 8
NZSEE guidelines
Parameters
R = 1.5, I =1.25
q = 1.0, I = 1.2
Sp = 1.0, = 1.25, = 5%, Cf = 0.72
Table 11: Parameters for a high ductility tank

Code
IBC 2000 and ACI 350.3
Eurocode 8
NZSEE guidelines
Parameters
R = 3.0, I =1.25
q = 2.0, I = 1.2
Sp = 1.0, = 3.0, = 5%, Cf = 0.38
Table 12: Proposed values of Response reduction factor, R for IS 1893:2002

Tank
Ground supported tanks
Unanchored steel tank
Anchored steel tank
Concrete tank with unconstrained flexible base
Concrete tank with non-sliding base
Elevated tanks
Supported on RCC shaft
Supported on RCC frame staging
Supported on steel frame staging
Supported on masonry shaft
Proposed value of R
1.80
2.25
1.10
1.50
1.50
2.25
2.25
1.10
40
R = 8.0
For building
R = 1.5
Base shear coefficient (BSC)
0.8
R = 2.0
For tanks
R = 2.5
0.6
R = 3.0
0.4
0.2
0
0
Time Period (S)
Figure 1: Variation of base shear coefficient with natural period;

IBC 2000 (SD =1.5, S1 = 0.6, Fa =1.0, Fv = 1.5, Class D site)
R = 1.5
7.5
R =2.0
For tanks
R = 2.5
BSCtank / BSCbldg
R = 3.0
4.5
(SS = 1.5, S1 = 0.6, Fa = 1.0, Fv = 1.5, site class D)

(I = 1 and R = 8 for building; I = 1.25 for tanks)
1.5
0
0
Time Period (S)
Figure 2: Ratio of tank and building base shear coefficient (IBC 2000)
41
0.8
As per modified equation of FEMA368

(SS = 1.5, S1 = 0.6, Fa = 1.0, Fv = 1.5 ,
I =1.25, R = 2, site class D)
0.6
0.4
0.2
As per equation of ACI 371

(Ca = 0.44, Cv = 0.64, R =2.0, soil type D)
0
0
Time period (S)
Figure 3: Comparison of base shear coefficient from ACI 371 and FEMA 368
Base shear coefficient(BSC)
0.8
(a) Impulsive mode
ACI 350.3; Rw = 2.0

FEMA 368; R=1.5
0.6
ACI 350.3; Rw = 4.75

FEMA 368; R = 3.0
0.4
0.2
0
0
Time period (S)
1.2
(b) Convective mode
ACI 350.3
FEMA
0.8
0.4
0
1
Time period (S)
Figure 4: Comparison of base shear coefficient from ACI 350.3 and FEMA 368
(For ACI 350.3: Z = 0.4, S = 1.5, I = 1.25, soil type C;
For FEMA 368: SS = 1.5, S1 =0.6, Fa =1.0, Fv = 1.5, I = 1.25, site class D)
42
BSCtank / BSCbldg
ACI 371
IBC 2000
2

(I = 1 and R = 8 for building; I = 1.25, R = 2 for tanks)
0
0
Time Period (S)
Figure 5: Ratio of tank and building base shear coefficient (ACI 371)
R = 1.5
1.5
R = 2.0
R = 2.5
R = 3.0
x Convective mode
1.0
No lower bound
0.5
0.0
0
Time Period (S)
Figure 6: Variation of base shear coefficient with time period (ACI 350.3)
(SD =1.5, S1 = 0.6, Fa =1.0, Fv = 1.5, site class D)
43
R = 1.5
R = 2.0
R = 2.5
R = 3.0
BSCtank / BSCbldg
5
4
3
2

1
0
0
Time Period (S)
Figure 7: Ratio of base shear coefficient of tank (impulsive mode) and building (ACI 350.3)
0.4
AWWA D-100; R = 3.5
Base shear Coefficient
AWWA D-100; R = 4.5

FEMA 368; R = 2.5
0.3
FEMA 368; R = 3.0

0.2
0.1
0
1
Time (S)
Figure 8: Comparison of convective base shear coefficient for ground-supported tank from
AWWA D-100 and FEMA 368 (For AWWA D-100 Z = 0.4, S = 1.5, I = 1.25, soil type C;
For FEMA 368: SS = 1.5, S1 =0.6, Fa =1.0, Fv = 1.5, I = 1.25, site class D)
44
0.5
AWWA D-100; Rw = 3
AWWA D-100; Rw = 4
0.4
FEMA 368; R = 2
FEMA 368; R = 3
0.3
0.2
0.1
0
0
0.5
1.5 (S)
Time
2.5
3.5
Figure 9: Comparison of impulsive base shear coefficient for elevated tank from AWWA
D-100 and FEMA 368 (For AWWA D-100 Z = 0.4, S = 1.5, I = 1.25, soil type C; For
FEMA 368: SS = 1.5, S1 =0.6, Fa =1.0, Fv = 1.5, I = 1.25, site class D)
R=2
R=3
For tanks
BSCtank/BSCbldg

0
0
0.5
1.5
2
Time (S)
2.5
3.5
Figure 10: Ratio of elevated tank and building base shear coefficient (AWWA D-100)
45
AWWA D-110; Rw = 2.0
0.8
AWWA D-110; Rw = 4.5

FEMA 368; R = 1.5
0.6
FEMA 368; R = 3.0
0.4
0.2
0
0
Time (S)
(a) Impulsive base shear

1.5
AWWA D-110
FEMA 368
0.5
0
1
Time (S)
(b) Convective base shear

Figure 11: Comparison of base shear coefficient from AWWA D-110 and FEMA 368
(For AWWA D-110 Z = 0.4, S = 1.5, I = 1.25, soil type C; For FEMA 368: SS = 1.5, S1 =0.6,
Fa =1.0, Fv = 1.5, I = 1.25, site class D)
46
R = 1.5
R = 2.0
R = 3.0
5
BSCtank/BSCbldg
For tanks
4
3
2
(SS = 1.5, S1 = 0.6, Fa = 1.0, Fv = 1.5 site class D)

(I = 1.0 and R = 8 for building; I = 1.25 for tanks)
1
0
0
Time (S)
Figure 12: Ratio of tank and building base shear coefficient (AWWA D-110)
AWWA D-115; Rw = 1.0
1.5
AWWA D-115; Rw = 3.0
FEMA 368; R = 1.5

FEMA 368; R = 3.0
0.5
0
0
Time (S)
Figure 13: Comparison of impulsive base shear coefficient from AWWA D-115 and
FEMA368
(For AWWA D-115: Z = 0.4, S = 1.5, I = 1.25, soil type C; For FEMA 368: SS = 1.5, S1
=0.6, Fa =1.0, Fv = 1.5, I = 1.25, site class D)
47
Base shear coefficient
0.8
API 650 (Z = 0.4, S = 1.5, I = 1.25, site type S3)

0.6
0.4
0.2
FEMA 368
(Ss = 1.5, S1 = 0.6, Fa = 1,
Fv = 1.5, site class D)
0
0
3
Time period (S)
Figure 14: Convective base shear coefficient from API 650 and FEMA 368
1.0
Building ( q = 5)
Tank (q = 1)
0.8
x Convective mode
0.6
No Lower bound
0.4
Lower bound
0.2
0.0
0
Time Period (S)
Figure 15: Variation of base shear coefficient with time period as per Eurocode 8
(S =1.0, = 2.5, = 1.0, K1 =1.0, K2 =2.0, Kd1 =2/3, Kd2 =5/3, TB =0.1, TC =0.4 TD = 3.0, I
=1.0 for building and I =1.3 for tank)
48
Tank (q = 1)
6
BSCtank / BSCbldg
5
4
3
2
1
0
0
Time Period (S)
Figure 16: Ratio of base shear coefficient of tank impulsive mode and building (Eurocode
8)
Building
1.2
Tank Cf = 0.72
Tank; Cf = 0.54
Tank; Cf = 0.38
0.9
No Lower bound
0.6
Lower bound
0.3
0
0
Time Period (S)
Figure 17: Variation of base shear coefficient with time period (NZSEE Guidelines)
( Sp = 0.67 , R = 1.0 for building, Sp = 1.0 R =1.3 for tank, Z =1.2, Lu =1.0)
49
Tank Cf = 0.72
Tank; Cf = 0.54
BSCtank / BSCbldg
Tank; Cf = 0.38
0
0
Time Period (S)
Figure 18: Ratio of base shear coefficient of tank and building (NZSEE Guidelines)
Eurocode 8 ( = 0.3,S = 1, = 2.5, = 1, Kd1 = 2/3, Kd2 = 5/3,

TA = 0.15, TB = 0.6, q = 5, sub soil class B)
Base shear coefficient (BSCbldg)
0.3
IBC 2000 ( SS = 1.5, S1 = 0.6, Fa = 1.0,

Fv = 1.5, R = 8, I = 1, site class D)
0.2
NZSEE ( Z = 1.2, Sp = 0.67, R = 1,

Lu = 1, Site category C)
0.1
0
0
Time (S)
Figure 19: Comparison of base shear coefficient for ductile building obtained from various
codes. Most severe zone in each code is considered
50
Eurocode 8 ( = 0.3, S = 1, = 2.5, = 1, = 1.2,

K1 = 1, K2 = 2, TA = 0.15, TB = 0.6,
q = 1, sub soil class B)
Base shear coefficient (BSCtank )
1.2
0.8
NZSEE ( Z = 1.2, Sp = 1.0, R = 1.3,

Cf = 0.72, Lu = 1, site category C)
ACI 350.3 and

AWWA D-110
0.4
IBC 2000 ( SS = 1.5, S1 = 0.6,

Fa = 1.0, Fv = 1.5, R = 1.5, I
= 1.25, site class D)
0
0
2
Time (S)
Figure 20: Comparison of base shear coefficient for ground supported unanchored
concrete water tank obtained from various codes. Most severe zone in each code is
considered.
NZSEE ( Z = 1.2, Sp = 1.0,
R = 1.3, Cf = 0.72, Lu = 1,
site category C)
10
IBC 2000 ( SS = 1.5, S1 = 0.6,

Fa = 1.0, Fv = 1.5, R = 1.5,
I = 1.25, site class D)
BSCtank / BSCbldg
ACI 350.3 and

AWWA D-110
6
4
Eurocode 8 ( = 0.3,S = 1, = 2.5,

= 1, = 1.2, K1 = 1, K2 = 2,TA = 0.15,
TB = 0.6, q = 1, sub soil class B)
2
0
0
Time (S)
Figure 21: Comparison of ratio of base shear coefficients of tank and building from
various codes (Low ductility tank).
51
BSCtank / BSCbldg
IBC 2000 ( SS = 1.5, S1 = 0.6,

Fa = 1.0, Fv = 1.5, R = 1.5,
NZSEE ( Z = 1.2, Sp = 1.0,

R = 1.3, Cf = 0.38, Lu = 1,
site category C)
ACI 350.3 and

AWWA D-110
2
Eurocode 8 ( = 0.3,S = 1, = 2.5,

= 1, = 1.2, K1 = 1, K2 = 2,
TA = 0.15, TB = 0.6, q = 2, Sub soil class B)
0
0
Time (S)
Figure 22: Comparison of ratio of base shear coefficient of tank and building from various
codes (High ductility tank).
ACI 350.3 and AWWA D-110 ( SS = 1.5, S1 = 0.6,

Fa = 1.0, Fv = 1.5, I = 1.25, site class D)
Eurocode 8 (S = 1, = 2.5, = 1.673,
1.5
= 1.2, K1 = 1, K2 = 2, TA = 0.15,
TB = 0.6, q = 1, sub soil class B)
API 650 (SS = 1.5, S1 = 0.6,

Fa = 1.0, Fv = 1.5,
0.5
0
1.5
2.5
3.5
4.5
5.5
Time (S)
Figure 23: Comparison of base shear coefficient for convective mode
52
IBC 2000 ( SS = 1.5, S1 = 0.6, Fa = 1.0,

Fv = 1.5, R = 8, I = 1, site class D)
0.1
0.075
IS 1893 (Part I):2002 (Z = 0.36,
R = 5, I = 1, soft soil)
0.05
0.025
IS 1893-1984
(K=1, = 1, 0 = 0.08, soft soil)
0
0
0.5
1.5
2.5
Time Period (S)
Figure 24: Comparison of base shear coefficient for building from IBC 2000, IS 1893:1984,
IS 1893(Part 1):2002. IBC values are divided by 1.4 to bring them to working stress level.
0.8
IBC 2000 ( Lowest value of R = 1.5)
IS1893: 2002 (R = 1.1)
0.6
IBC 2000 ( Highest value of R = 3)
IS1893: 2002 (R = 2.25)
0.4
IS1893: 1984
0.2
0
0
0.5
1.5
2.5
Time period (S)
Figure 25: Base shear coefficients for tanks from IBC 2000, IS 1893:1984 and IS 1893(Part
1):2002. IBC values are divided by 1.4 to bring them to working stress level.
53

Liquid Storage Tank

Uploaded by

Copyright:

Available Formats

Liquid Storage Tank

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Liquid Storage Tank

Uploaded by

Copyright:

Available Formats

Review of Code Provisions on Design

Seismic Forces for Liquid Storage Tanks

Dr. Durgesh C Rai

Review of Design Seismic Forces for Liquid Storage Tanks

Review of Code Provisions on

Review of Design Seismic Forces for Liquid Storage Tanks

Review of Design Seismic Forces for Liquid Storage Tanks

(f) NZSEE guidelines and NZS 4203:1992

Review of Design Seismic Forces for Liquid Storage Tanks

period, respectively; I is importance factor; R is response modification factor

Review of Design Seismic Forces for Liquid Storage Tanks

Review of Design Seismic Forces for Liquid Storage Tanks

3.1 ACI 371 (1998)

for Ts < T < 4 s

Review of Design Seismic Forces for Liquid Storage Tanks

3.2 ACI 350.3 (2001)

for 0 < Ti <T0

T0 = 0.2SDS/SD1 and Ti is time period of impulsive mode. All other parameters

for all values of Tc

Review of Design Seismic Forces for Liquid Storage Tanks

Here, Tc is the time period of convective mode. In convective mode, damping is

Review of Design Seismic Forces for Liquid Storage Tanks

AWWA D-115 (1995) is for prestressed concrete water tanks with

Review of Design Seismic Forces for Liquid Storage Tanks

Review of Design Seismic Forces for Liquid Storage Tanks

4.2 AWWA D-110 (1995)

Review of Design Seismic Forces for Liquid Storage Tanks

. However, on page no. 312 it is stated that the modified

inconsistency in the two expressions, second one considers a factor of 1.4R in

Since in AWWA D-110 the convective base shear

coefficient does not depend on response reduction factor, in Table 4, the

Review of Design Seismic Forces for Liquid Storage Tanks

4.3 AWWA D-115 (1995)

Review of Design Seismic Forces for Liquid Storage Tanks

Due to this, it is not possible to compare the

Review of Design Seismic Forces for Liquid Storage Tanks

(c) In AWWA D-115, convective base shear depends on response reduction

5. API 650 (1998)

Review of Design Seismic Forces for Liquid Storage Tanks

magnitude of the spectra in constant-acceleration range. It may be recalled here

Review of Design Seismic Forces for Liquid Storage Tanks

Review of Design Seismic Forces for Liquid Storage Tanks

This is clearly seen from Figure 16, wherein ratio

damping correction factor is 1.673. As a result of this, elastic spectrum

7. NZSEE GUIDELINES AND NZS 4203 :1992

Review of Design Seismic Forces for Liquid Storage Tanks

= Structural performance factor. For buildings Sp = 0.67

= Risk factor. For ordinary buildings R =1.0

= Zone factor as per seismic map

= Limit state factor for ultimate limit state. Lu = 1.0

Review of Design Seismic Forces for Liquid Storage Tanks

= 6.0. It would be appropriate to note that variation of Ch(T,6) and Ch(T,1)

Review of Design Seismic Forces for Liquid Storage Tanks

8. COMPARISON OF DESIGN FORCES FROM VARIOUS CODES

Review of Design Seismic Forces for Liquid Storage Tanks

Review of Design Seismic Forces for Liquid Storage Tanks

On similar lines, comparison of convective base shear coefficients

9. PROVISIONS OF INDIAN CODE

Here, K = Performance factor depending on the structural framing system and