Is 1893 - Tanks & Water Retaining

Lecture 1
January 17, 2006

In this lecture
Types of tanks
IS codes on tanks
Modeling of liquid
© Sudhir K. Jain, IIT Kanpur E-Course on Seismic Design of Tanks/ January 2006 Lecture 1 / Slide 2
Types of tanks
Two categories
Ground supported tanks
Also called at-grade tanks; Ground Service Reservoirs (GSR)
Elevated tanks
Also called overhead tanks; Elevated Service Reservoirs (ESR)
Types of tanks
Ground supported tanks

Shape: Circular or Rectangular
Material : RC, Prestressed Concrete, Steel
These are ground supported vertical tanks
Horizontal tanks are not considered in this course
Types of tanks
Elevated tanks
Two parts:
Container
Staging (Supporting tower)
Types of tanks
Elevated tanks
Container:
Material: RC, Steel, Polymer
Shape : Circular, Rectangular, Intze, Funnel, etc.
Staging:
RC or Steel frame
RC shaft
Brick or masonry shafts
Railways often use elevated tanks with steel frame staging
Now-a-days, tanks on brick or stone masonry shafts are not
constructed
Use of tanks
Water distribution systems use ground
supported and elevated tanks of RC & steel
Petrochemical industries use ground supported

steel tanks
Indian Codes on Tanks
IS 3370:1965/1967 (Parts I to IV)
For concrete (reinforced and prestressed) tanks
Gives design forces for container due to
hydrostatic loads
Based on working stress design
BIS is considering its revision
IS 11682:1985
For RC staging of overhead tanks
Gives guidelines for layout & analysis of staging
More about this code later
IS 803:1976
For circular steel oil storage tanks
IS 1893:1984
Gives seismic design provisions
Covers elevated tanks only
Is under revision
More about other limitations, later
IS 1893 (Part 1):2002 is for buildings only

Can not be used for tanks
Hydrodynamic Pressure
Under static condition, liquid applies pressure on
container.
This is hydrostatic pressure
During base excitation, liquid exerts additional
pressure on wall and base.
This is hydrodynamic pressure
This is in additional to the hydrostatic pressure
Hydrodynamic pressure
Hydrostatic pressure
Varies linearly with depth of liquid
Acts normal to the surface of the container
At depth h from liquid top, hydrostatic pressure =
γh
γh
Hydrostatic pressure
Has curvilinear variation along wall height
Its direction is opposite to base motion
Base motion
Summation of pressure along entire wall surface

gives total force caused by liquid pressure
Net hydrostatic force on container wall is zero
Net hydrodynamic force is not zero
Circular tanks (Plan View)

Hydrostatic pressure Hydrodynamic pressure
Base motion
Net resultant force = zero Net resultant force ≠ zero
Note:- Hydrostatic pressure is axisymmetric; hydrodynamic is asymmetric

Rectangular tanks (Plan View)
Hydrostatic pressure Hydrodynamic pressure
Base motion
Net resultant force = zero Net resultant force ≠ zero
Static design: Hydrostatic pressure is considered

Hydrostatic pressure induces hoop forces and
bending moments in wall
IS 3370 gives design forces for circular and
rectangular tanks
Net hydrostatic force is zero on container wall
Hence, causes no overturning moment on
foundation or staging
Thus, hydrostatic pressure affects container
design only and not the staging or the
foundation
Seismic design: Hydrodynamic pressure is

considered
Net hydrodynamic force on the container is not
zero
Affects design of container, staging and
foundation
Procedure for hydrodynamic pressure & force:

Very simple and elegant
Based on classical work of Housner (1963a)
Housner, G. W., 1963a, “Dynamic analysis of fluids in
containers subjected to acceleration”, Nuclear Reactors and
Earthquakes, Report No. TID 7024, U. S. Atomic Energy
Commission, Washington D.C.
We need not go in all the details
Only basics and procedural aspects are
explained in next few slides
Modeling of liquid
Liquid in bottom portion of the container moves with
wall
This is called impulsive liquid
Liquid in top portion undergoes sloshing and moves
relative to wall
This is called convective liquid or sloshing liquid
Convective liquid
(moves relative to tank wall)
Impulsive liquid
(moves with tank wall)
Modeling of liquid
Impulsive liquid
Moves with wall; rigidly attached
Has same acceleration as wall
Convective liquid
Also called sloshing liquid
Moves relative to wall
Has different acceleration than wall
Impulsive & convective liquid exert pressure on
wall
Nature of pressure is different
See next slide
Modeling of liquid
Impulsive Convective
Base motion Base motion
Modeling of liquid
At this point, we will not go into details of

hydrodynamic pressure distribution
Rather, we will first find hydrodynamic forces
Impulsive force is summation of impulsive
pressure on entire wall surface
Similarly, convective force is summation of
convective pressure on entire wall surface
Modeling of liquid
Total liquid mass, m, gets divided into two

parts:
Impulsive liquid mass, mi
Convective liquid mass, mc
Impulsive force = mi x acceleration

Convective force = mc x acceleration
mi & mc experience different accelerations
Value of accelerations will be discussed later
First we will find mi and mc
Modeling of liquid
Housner suggested graphs for mi and mc

mi and mc depend on aspect ratio of tanks
Such graphs are available for circular &
rectangular tanks
See Fig. 2a and 3a of Guidelines
Also see next slide
For taller tanks (h/D or h/L higher), mi as fraction
of m is more
For short tanks, mc as fraction of m is more
Modeling of liquid
1 1
mi/m mi/m
0.5 0.5
mc /m
mc /m
0 0
0 0.5 h/L
1 1.5 2
0 0.5 1 1.5 2
h/D
For circular tanks For rectangular tanks
See next slide for definition of h, D, and L
Modeling of liquid
D
h
Plan of Circular tank

Elevation
Base motion
L L
Base motion
Plan of Rectangular tank
Modeling of liquid
Example 1:
A circular tank with internal diameter of 8 m, stores 3 m height
of water. Find impulsive and convective water mass.
Solution:
Total volume of liquid = π/4 x 82 x 3 = 150.8 m3
∴ Total liquid mass, m = 150.8 x 1.0 = 150.8 t
Note:- mass density of water is 1000 kg/m3; weight density of water is

9.81 x 1000 = 9810 N/m3.
D = 8 m, h = 3 m
∴ h/D = 3/8 = 0.375.
1
mi/m
0.5
mc /m
0
0 0.5 1 1.5 2
h/D
Modeling of liquid
From graph, for h/D = 0.375

mi/m = 0.42 and mc/m = 0.56
mi = 0.42 x 150.8 = 63.3 t and

mc = 0.56 x 150.8 = 84.5 t
Modeling of liquid
Impulsive liquid is rigidly attached to wall

Convective liquid moves relative to wall
As if, attached to wall with springs
Kc/2 Kc/2 Convective liquid

mc (moves relative to wall)
Rigid m Impulsive liquid

i (moves with wall)
Modeling of liquid
Stiffness associated with convective mass, Kc

Kc depends on aspect ratio of tank
Can be obtained from graph
Refer Fig. 2a, 3a of guidelines
See next slide
Modeling of liquid
1
Kch/mg
mi/m
0.5
mc/m
0
0 0.5 1 1.5 2
h/D
Modeling of liquid
Example 2:
A circular tank with internal diameter of 8 m, stores 3 m height
of water. Find Kc.
Solution:
Total liquid mass, m = 150.8 t (from Example 1)
= 150.8 x 1000 = 150800 kg
g = acceleration due to gravity = 9.81 m/sec2
D = 8 m, h = 3m
∴ h/D = 3/8 = 0.375. From graph, for h/D = 0.375;
Kc h/mg = 0.65
Modeling of liquid
1
Kch/mg
mi/m
0.5
mc/m
0
0 0.5 1 1.5 2
h/D
Kc = 0.65 mg/h
∴ Kc = 0.65 x150800 x 9.81/3.0 = 320,525.4 N/m
Note: - Unit of m is kg, hence unit of Kc is N/m. If we take m in ton, then

unit of Kc will be kN/m.
Modeling of liquid
Now, we know liquid masses mi and mc

Next, we need to know where these are
attached with the wall
Like floor mass in building acts at centre of
gravity (or mass center) of floor
Location of mi and mc is needed to obtain
overturning effects
Impulsive mass acts at centroid of impulsive
pressure diagram
Similarly, convective mass
Modeling of liquid
Impulsive mass acts at centroid of impulsive

pressure diagram
Location of centroid:
Obtained by dividing the moment due to
pressure distribution by the magnitude of
impulsive force
Similarly, location of convective mass is obtained
See next slide
Modeling of liquid
Resultant of impulsive Resultant of convective
pressure on wall pressure on wall
hc
hi
hi, hc can be obtained from graphs

They also depend on aspect ratio, h/D or h/L
Refer Fig. 2b, 3b of guidelines
See next slide
Modeling of liquid
1 1
0.8 0.8
hc/h hc/h
0.6 0.6
hi/h hi/h
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0.8 1 h/D 1.2 1.4 1.6 1.8 2 0 0.5 1 h/L 1.5 2
For circular tanks For rectangular tanks
Modeling of liquid
Example 3:
A circular tank with internal diameter of 8 m, stores 3 m
height of water. Find hi and hc.
Solution:
D = 8 m, h = 3m
∴ h/D = 3/8 = 0.375.
1
0.8
hc/h
0.6
hi/h
0.4
0.2
0
0 0.5 1 h/D 1.5 2
Modeling of liquid
From graph, for h/D = 0.375;

hi/h = 0.375
hi = 0.375 x 3 = 1.125 m
and hc/h = 0.55
hc = 0.55 x 3 = 1.65 m
Note :- Since convective pressure is more in top portion, hc > hi.
Modeling of liquid
Hydrodynamic pressure also acts on base

Under static condition, base is subjected to
uniformly distributed pressure
Due to base motion, liquid exerts nonuniform
pressure on base
This is in addition to the hydrostatic pressure on the base
See next slide
Modeling of liquid
Base motion
Hydrostatic pressure on base Hydrodynamic pressure on base
Modeling of liquid
Impulsive as well as convective liquid cause

nonuniform pressure on base
Nonuniform pressure on base causes overturning
effect
This will be in addition to overturning effect of
hydrodynamic pressure on wall
See next slide
Modeling of liquid
hi
Overturning effect Overturning effect

due to wall pressure due to base pressure
Note:- Both the overturning effects are in the same direction
Modeling of liquid
Total overturning effect of wall and base

pressure is obtained by applying resultant of
wall pressure at height, hi* and hc*.
• In place of hi and hc discussed earlier
For overturning effect due to wall pressure alone,
resultant was applied at hi

For hi and hi , see next slide
*
Modeling of liquid
h*i
hi
Location of resultant of wall Location of Resultant of wall

pressure when effect of base pressure when effect of base
pressure is not included pressure is also included
Modeling of liquid
Similarly, hc and hc* are defined
h*c
hc
Location of resultant of wall

pressure when effect of base
Location of Resultant of wall
pressure is not included
pressure when effect of base
pressure is also included
Modeling of liquid
hi and hi* are such that

Moment due to impulsive pressure on walls only = Impulsive force x hi
Moment due to impulsive pressure on walls and base = Impulsive force x hi*
hc and hc* are such that

Moment due to convective pressure on walls only = Convective force x hc
Moment due to convective pressure on walls and base = Convective force x hc*
Modeling of liquid
hi* is greater than hi

hc* is greater than hc
Refer Fig. C-1 of the Guidelines
hi* & hc* depend on aspect ratio
Graphs to obtain hi, hc, hi*, hc* are provided
Refer Fig. 2b & 3b of guidelines
Also see next slide
Please note, hi* and hc* can be greater than h
Modeling of liquid
2.5
hc*/h
2
1.5
hc/h hi*/h
0.5
hi/h
0
0 0.5 1 1.5 2
h/D
Modeling of liquid
Example 4:
A circular tank with internal diameter of 8 m, stores 3 m height of
water. Find hi* and hc*.
Solution:
D = 8 m, h = 3m
∴ h/D = 3/8 = 0.375. From graph, for h/D = 0.375;
Modeling of liquid
2.5
hc*/h
2
1.5
hc/h hi */h
0.5
hi /h
0
0 0.5 1 1.5 2
h/D
hi*/h = 1.1
Hence hi* = 1.1 x 3 = 3.3 m
Similarly, hc*/h = 1.0
Hence, hc* = 1.0 x 3 = 3.0 m
Modeling of liquid
This completes modeling of liquid

Liquid is replaced by two masses, mi & mc
This is called mechanical analogue or spring
mass model for tank
See next slide
Modeling of liquid
mi = Impulsive liquid mass

Kc/2 Kc/2 mc = Convective liquid mass
mc Kc = Convective spring stiffness
hi = Location of impulsive mass
hc (without considering overturning
Rigid m (hc*) caused by base pressure)
i
hi hc = Location of convective mass
(hi*) (without considering overturning
caused by base pressure)
Mechanical analogue hi* = Location of impulsive mass
or (including base pressure effect on
spring mass model of tank overturning)
hc* = Location of convective mass
(including base pressure effect on
overturning)
Modeling of liquid
mi, mc, Kc, hi, hc, hi* and hc* can also be obtained
from mathematical expressions:
These are given in Table C 1 of Guidelines
These are reproduced in next two slides
Modeling of liquid
For circular tanks
⎛ h⎞
⎛ D⎞ tanh ⎜ 3.68 ⎟
tanh ⎜ 0 . 866 ⎟ mc
= 0.23 ⎝ D⎠
mi
= ⎝ h ⎠ m h
m D D
0 . 866
h
⎛ h⎞
cosh ⎜ 3.68 ⎟ − 1.0
hi hc
= 1− ⎝ D⎠
= 0 . 375 for h / D ≤ 0 . 75 h h ⎛ h⎞
h 3.68 sinh ⎜ 3.68 ⎟
= 0 .5 −
0 . 09375
for h / D > 0 . 75 D ⎝ D⎠
h / D
D ⎛ h⎞
0.866 cosh⎜ 3.68 ⎟ − 2.01
hi *
= h hc * ⎝ D⎠
h ⎛ D⎞
- 0.125 for h / D ≤ 1.33 =1−
2 tanh⎜ 0.866 ⎟ h h ⎛ h⎞
⎝ h⎠ 3.68 sinh⎜ 3.68 ⎟
D ⎝ D⎠
= 0.45 for h / D > 1.33

mg ⎛ h⎞
K c = 0.836 tanh 2 ⎜ 3.68 ⎟
h ⎝ D⎠
Modeling of liquid
For rectangular tanks
⎛ h⎞
⎛ L⎞ tanh⎜ 3.16 ⎟
tanh⎜ 0.866 ⎟ mc
= 0.264 ⎝ L⎠
mi
= ⎝ h⎠ m h
m L L
0.866
h
⎛ h⎞
cosh ⎜ 3.16 ⎟ − 1.0
hi hc
= 1− ⎝ L⎠
= 0 . 375 for h / L ≤ 0 . 75 h h ⎛ h⎞
h 3.16 sinh ⎜ 3.16 ⎟
L ⎝ L⎠
for h / L > 0 . 75
0 . 09375
= 0 .5 −
h/L
⎛ h⎞
cosh⎜ 3.16 ⎟ − 2.01
L hc * ⎝ L⎠
hi *
0.866 = 1−
= h − 0.125 h h ⎛ h⎞
h ⎛ L⎞ for h / L ≤ 1.33 3.16 sinh⎜ 3.16 ⎟
2 tanh⎜ 0.866 ⎟ L ⎝ L⎠
⎝ h⎠
= 0 .45 for h / L > 1.33 mg ⎛ h⎞

K c = 0.833 tanh 2 ⎜ 3.16 ⎟
h ⎝ L⎠
Modeling of liquid
Note, in Table C-1 of the Guideline, there are two

typographical errors in these expressions
For circular tank, first expression for hi/h shall have
limit as “for h/D ≤ 0.75”
For circular tank, in the expression for hi*/h, there
shall be minus sign before 0.125
These two errors have been corrected in the
expressions given in previous two slides
Modeling of liquid
mi and mc are needed to find impulsive and

convective forces
Impulsive force, Vi = mi x acceleration
Convective force, Vc = mc x acceleration
Kc/2 Kc/2
mc Vc
Rigid m Vi
i
Modeling of liquid
Vi and Vc will cause

Bending Moment (BM) in wall
Modeling of liquid
BM at bottom of wall
BM due to Vi = Vi x hi
BM due to Vc = Vc x hc
Total BM is not necessarily Vi X hi+ Vc X hc
More about this, later
Kc/2 Kc/2
mc Vc
hc Rigid
mi Vi
hi
Modeling of liquid
Overturning of the container is due to pressure

on wall and base
Pressure on base does not cause BM in wall
Overturning Moment (OM) at tank bottom
OM is at bottom of base slab
Hence, includes effect of pressure on base
Note the difference between bottom of wall and
bottom of base slab
Modeling of liquid
OM at bottom of base slab

OM due to Vi = Vi x hi*
BM due to Vc = Vc x hc*
Kc/2 Kc/2
mc Vc
hc* Rigid
mi Vi
hi*
Modeling of liquid
mi and mc will have different accelerations

We yet do not know these accelerations
ai = acceleration of mi
ac = acceleration of mc
Procedure to find acceleration, later
Use of mi, mc, hi, hc, hi* and hc* in next example
Acceleration values are assumed
Modeling of liquid
Example 5:
A circular tank with internal diameter of 8 m, stores 3 m height of
water. Assuming impulsive mass acceleration of 0.3g and
convective mass acceleration of 0.1g, find seismic forces on tank.
Solution:
Geometry of tank is same as in previous examples.
D = 8 m, h = 3m
From previous examples:
mi = 63.3 t mc = 84.5 t
hi = 1.125 m hc = 1.65 m
hi* = 3.3 m hc* = 3.0 m
Impulsive acceleration, ai = 0.3g = 0.3 x 9.81 = 2.94 m/sec2
Convective acceleration, ac = 0.1g = 0.1 x 9.81 = 0.98 m/sec2
Modeling of liquid
Example 5 (Contd..)
Impulsive force, Vi = mi x ai = 63.3 x 2.94 = 186.1 kN

Convective force, Vc = mc x ac = 84.5 x 0.98 = 82.8 kN
Bending moment at bottom of wall due to Vi = Vi x hi

= 186.1 x 1.125 = 209.4 kN-m
Bending moment at bottom of wall due to Vc = Vc x hc
= 82.8 x 1.65 = 136.6 kN-m
Overturning moment at bottom of base due to Vi = Vi x hi*
= 186.1 x 3.3 = 614.1 kN-m
Overturning moment at bottom of base due to Vc = Vc x hc*
= 82.8 x 3.0 = 248.4 kN-m
At the end of Lecture 1
In seismic design, mechanical analogue of tanks
are used, wherein, liquid is replaced by
impulsive & convective masses
These masses and their points of application
depend on aspect ratio
Graphs and expressions are available to find all
these quantities
These are based on work of Housner (1963a)
E-Course (through Distance Learning Mode) on
Seismic Design of Liquid Storage Tanks

January 16 - February 6, 2006
Response to Questions and Comments on Lecture 1

Question from Mr. S. K. Kukreja (guduja@rediffmail.com)
1. clarify following
1. fig 2a,3a of guidelines as pointed in lecture1/slide25
2. table C-1 of guidelines as pointed in lecture1/slide57
3. Housner graphs to find h* are not given for rectangular tanks
In Figure 2a of the Guideline, mi/m, mc/m and Kch/mg are plotted as a

function of h/D. This is for circular tanks and corresponding figure for
rectangular tanks is given in Figure 3a. Expressions for these quantities are
given in Table C-1 of the Guideline. Basically, graphs plotted in these
figures are based on these expressions only. Depending on convenience,
one can find these quantities either from graphs or from expressions. In
the Lecture we have given graphs of hi* and hc* only for circular tanks
(Slide 54), which is Figure 2b of the Guideline. Corresponding figure for
rectangular tank is Figure 3b of the Guideline, which has not been
reproduced in the Lecture. Expressions for hi* and hc* for circular as well as
rectangular tanks are also given in Table C-1.
Question from Mr. Umesh H. Patil (uhp9999@yahoo.co.in)
1. pl.something about hydrostatic design
Under static conditions, liquids apply a pressure on the bodies in contact.

This pressure is a product of unit weight of liquid and depth of the liquid at
the point under consideration. Direction of this pressure is always normal
(perpendicular) to the surface of the body. This is known as hydrostatic
pressure, and it varies linearly with height.
Hydrostatic pressure induces hoop forces and bending of the wall. IS

3370:1967 gives these forces for circular and straight walls. Refer standard
textbooks for more information.
2. show to calculate hydrodynamic pr.parameters for tapering sections of

wall/container
Lecture 3 has covered information on tanks of other shapes such as Intze,
conical, etc. If cross-section of wall is tapered, i.e., its thickness varies along
the height, then there will not be any change in hydrodynamic pressure.
Recall, hydrodynamic pressure is obtained considering wall as rigid.
3. effect of top covered slab on hydrodynmic pressure
Top slab will impart more stiffness to wall and make it more rigid. This is
not going to affect hydrodynamic pressure distribution. However, if
sloshing liquid touches roof slab, then hydrodynamic pressure distribution
will be affected. Please refer Lecture 6 and Commentary to Clause 4.11 of
the Guideline.
4. there is difference in answers from graph & equations
Reading the value from graph is likely to induce some approximation and
hence, expressions are also given. The difference, however, may not be
significant from engineering view point.
Question from Mr. H. I. Abdul Gani (gani@igcar.ernet.in)
The following are my questions on Lecture 1.

1) What exactly does Horizontal Tank and Vertical Tank mean
A circular tank kept horizontally on ground, i.e., longitudinal axis of tank is

horizontal is termed as horizontal tank.
In petrol pumps, steel tanks used for storing petrol/diesel are usually
horizontal tanks. Railway tankers carrying petroleum products are also
horizontal tanks. Another good example is heating greasers in bathrooms,
some are of horizontal type and some are of vertical type !!. A vertical tank is
the one, in which longitudinal axis of tank is vertical. Classification as
horizontal and vertical tank is more appropriate for circular tanks only.
2) It was shown that the hydrodynamic pressure is curvilinear in shape. Does

the pressure profile depend on the stiffness of the tank walls?
3) How the stiffness of the wall affects i) Convective mass and ii) impulsive
mass
These have been addressed in Lecture 3
4) In example 1, why the sum of impulsive and convective masses (147=2E8kg)

is less than the total mass of water (150.8 kg)?
This also has been addressed in Lecture 3

5) In the hydrodynamic pressure profile in the base, it is not clear how pressure
can develop in the upward direction? Can you please explain
Under static condition, liquid exerts uniform pressure on base. During

lateral excitation, the liquid increases vertical pressure on some portion of
slab and reduces pressure on other portion. This increase or decrease in
pressure is the hydrodynamic effect. The decrease in the pressure is like
an upward pressure.
6) How to find height of sloshing liquid? Does it depend on acceleration
This is addressed in Lecture 6
7) For tanks with arbitrary shapes like triangular, hexagonal shapes etc, how to
find impulsive and convective masses? Any simplified methods available
Please refer Lecture 3
8) During vertical acceleration, how to find impulsive and convective mass?
Information on effect of vertical excitation is covered in Lecture 6. Also read

Clause 4.10 of the Guideline. Vertical excitation will cause change in
weight density of liquid, which in turn will change hydrostatic pressure on
wall.
9) Why effect of base pressure is considered along with both convective mass
and impulsive mass instead of convective mass only?
In our modeling, liquid is divided into two parts, one that vibrates along with
wall (impulsive mass) and other which moves relative to wall (convective
mass). During lateral excitation, both impulsive and convective liquids exert
non-uniform pressure on base.
10) How to check adequacy of freeboard in a water tank during dynamic

excitation?
This issue is addressed in Lecture 6
11) For lumping mass in mathematical model for impulsive and convective mass
which height is to be used, h or h*?
If bending moment at the bottom of wall is to be obtained, then, masses shall

be lumped at h. If overturning moment at the bottom of base slab is to
obtained, then masses shall be lumped at h*. For design of staging and
foundation also, mass must be lumped at h*.
12) For load application in a mathematical model, only C.G. (h) of the impulsive/
convective force is known. How to find the area and distribution of load (or
mass) with respect to height of the tank?
Details about distribution of impulsive and convective pressure along wall

are covered in Lecture 6.
Question from (cshekhar66@rediffmail.com)

I didnt understood the following
pl explain mi/m=3D tanh(.866*12/5)/.866*12/5
now in this expression mi/m=3Dtanh 2.078/2.078
mi/m=3Dtan 5*2.078/2.078 is it finding tan 10.39 and then dividing it by 2.078 all
the other thing i understood and also since i have referred charts from the
guidlines i also got the answers correct
Abut only i didnt understood the formulae part how the value arrived
thanking you
chandrshekhar
The expression which you have given above contains 3D term, which is not
present in the expressions given in Lectures and in Table C-1 of the
Guideline. You are also interested in knowing how this formula is arrived at.
The derivation of this formula is given in Housner’s paper (1963a) and you
may go through the same. It comes from potential equation of liquid or
Laplace equation. Since it is a bit mathematical, we have not included this in
the course contents.
Question from Mr. R. Murugan (rmurugan@igcar.ernet.in)
1. How convective pressure is distributed and upto what depth ?
We need to recognize that we are dividing liquid in two parts. That does not
mean that physically liquid gets divided in two parts and liquid only up to
certain depth participates in convective mode and liquid below it participates
in impulsive mode. It is an idealization, and convective and impulsive
pressures are present over the entire height of liquid. This can also be
observed in the pressure distributions shown in Lecture 6.
2. In slide 54 How hi* is higher than the tank height?
hi* is that height where mi should act so that moment due to impulsive
pressure on wall and moment due to impulsive pressure on base is equal to
moment due to mi (actually, mi × g) . Since, effect of moment due to base
pressure is included, this height can become larger than h. On the other
hand hi, which includes effect of impulsive pressure only on the walls, is
always less than h.
3. The formulae derived for the above tanks are applicable for open tank or
closed tank ? (i.e. tank with cover slab or without cover slab)
Actually these are derived for open tanks, wherein, question of liquid
touching roof slab does not arises. However, due to presence of roof slab,
pressure distribution on wall is not going to change, provided sloshing
liquid does not touch roof slab. If it touches roof slab then some additional
issues will come up. These issues have been addressed in Lecture 6 and
in Clause 4.11 of the Guideline.
4. The formulas and graph derived to find hi, hc, mi, mc etc is applicable for RC
tank or steel tank? The formulae are irrespective of materials or the materials
proportion will change the co-efficient?
This issue has been addressed in Lecture 3. Effect of flexibility of wall on

pressure distribution has been studied by Veletsos (1984). However, for
design purpose we use formulae derived for rigid walls. Please refer
Lecture 3 and commentary of Clause 4.2.1.2
5. What is mean by squat tank, slender tank?
Slender tank is one in which h/D or h/L is quite high. In squat tank h/D or
h/L is less. There is no fixed limit on these values at which this
demarcation begins. Generally, these terminologies are used for impulsive
and convective masses. For example, we say that in slender tanks,
convective mass is less and in squat tanks, impulsive mass is less. In this
context, from Figure 2a of the Guideline, we can see that for h/D = 0.2,
about 80% mass contributes to convective mode and for h/D = 1.0 about
80% mass contributes to impulsive mode. Thus, one can say that a tank
with h/D less than 0.2 is squat tank and a tank with h/D greater than 1.0 is
slender tank.
6. How Question 1.2 is false in Assignment i.e.., in short tank mi is less only See
Example-1 in Lecture -1.
This was error in our solution, and we have sent addendum to it.
7. So, far we learnt about impulsive mass, convective mass and it is centeroid.
When we do the modeling in computer. How impulsive mass is distributed
i.e how applied and upto what depth?
Distribution of impulsive and convective pressure along wall height is

covered in Lecture 6.
8. Why mi + mc is not coming to m ? In solution 1, mi/m = 0.466; mc/m = 0.503;

when both are added the value is 0.969; Nearly 3% of the liquid mass is
unaccounted. Why is it so? Is it not correct to include the missing mass in
impulsive liquid component?
This has been discussed in Lecture 3.
9. Table C1-Expression for parameters of spring mass model, hi/h = 0.375 for
h/D less than or equal to 0.75 is to be given, while the actual one furnished
reads, h/d greater than 0.75
This is a typographical error in Table C-1, which we have pointed out in

the Lecture 1. There is another typographical error in the expression for hi*
for circular tanks. This also has been pointed out in Lecture 1 (Slide no.
60).
Question from Mr. Abdul Gani (gani@igcar.ernet.in)
It is explained in lecture 1, impulsive pressure distribution as below:
Impulsive
Let us assume mi/m =x:
If the pressure distribution is as shown in the above figure, hi should come less
than or equal to 0.5*x* h, i.e., hi/h should always be lower than 0.5x.
Table below compares the hi expected versus actual computed

hi/h
Problem no. mi/m hi/h expected
computed
Less than or equal to
1.1 Circular tank 0.466 0.375
0.233
1.2 Rectangular Less than or equal to
tank in longer 0.54 0.54/2 0.375
direction 0.27
1.2 Rectangular Less than or equal to
tank in shorter 0.72 0.72/2 0.4
direction 0.36
From above computations, it appears that the pressure distribution for impulsive
loading explained in lecture 1 appears to be incorrect.
Kindly explain.
This question arose because of the feeling that physically liquid got
divided into two parts. You are thinking that liquid below ( x ) × (h ) height is
impulsive and liquid above this height is convective. As has been
explained in the answer to earlier question, there is no physical line, which
demarcates impulsive and convective portions of liquid. Impulsive liquid is
present throughout the height; however, more liquid from bottom portion
participates in impulsive mode. Similarly, convective liquid is also
distributed along the entire height, but more liquid from upper portion
participates in convective mode. Since more liquid from bottom portion is
participating in impulsive mode, hi/h is always less than 0.5h. Similarly,
hc/h is always greater than 0.5. Please refer Figure 2b and 3b of the
Guideline.
Lecture 2
January 19, 2006

In this lecture
Seismic force evaluation
Procedure in codes
Limitations of IS 1893:1984
During base excitation
Structure is subjected to acceleration
From Newton’s second law
Force = mass x acceleration
Hence, seismic force acting on structure
= Mass x acceleration
For design, we need maximum seismic force

Hence, maximum acceleration is required
This refers to maximum acceleration of structure
This is different from maximum acceleration of
ground
Maximum ground acceleration is termed as
peak ground acceleration, PGA
Maximum acceleration of rigid structure is same
as PGA.
Seismic force = mass x maximum acceleration

Can be written as:
Force = (maximum acceleration/g) x (mass x g)
= (maximum acceleration/g) x W
W is weight of the structure
g is acceleration due to gravity
Typically, codes express design seismic force as:
V = (Ah) x (W)
V is design seismic force, also called design base
shear
Ah is base shear coefficient
Maximum acceleration of structure depends on

Severity of ground motion
Soil conditions
Structural characteristics
These include time period and damping
More about time period, later
Obviously, base shear coefficient, Ah, will also
depend on these parameters
Seismic design philosophy is such that, design

seismic forces are much lower than actual
seismic forces acting on the structure during
severe ground shaking
Base shear coefficient has to ensure this
reduction in forces
Hence, base shear coefficient would also have a
parameter associated with design philosophy
Thus, base shear coefficient depends on:

Severity of ground motion
Soil condition
Design philosophy
Let us examine how following codes have

included these parameters in base shear
coefficient
IS 1893 (Part 1): 2002
IS 1893:1984
International Building code (IBC) 2003 from USA
Study of IBC provisions will help us understand
the present international practice
IS 1893 (Part 1):2002
Ah = (Z/2). (I/R). (Sa/g)
Z is zone factor
I is importance factor
R is response reduction factor
Sa/g is spectral acceleration coefficient
IS 1893 (Part 1):2002
Zone factor, Z
Depends on severity of ground motion
India is divided into four seismic zones (II to V)
Refer Table 2 of IS 1893(part1):2002
Z = 0.1 for zone II and Z = 0.36 for zone V
IS 1893 (Part 1):2002
Importance factor, I
Ensures higher design seismic force for more
important structures
Values for buildings are given in Table 6 of
IS :1893
Values for other structures will be given in respective parts
For tanks, values will be given in Part 2
IS 1893 (Part 1):2002
Response reduction factor, R

Earthquake resistant structures are designed for
much smaller seismic forces than actual seismic
forces that may act on them. This depends on
Ductility
Redundancy
Overstrength
See next slide
IS 1893 (Part 1):2002
Δ
Total
Horizontal
Maximum force Load
if structure remains elastic Fel
Due to
Linear Elastic
Total Horizontal Load
Ductility
Response
Non linear
Maximum Response
Load Capacity Fy First Due to
Load at Significant Redundancy
First Yield Fs Yield
Due to
Overstrength
Design force Fdes Figure: Courtesy
Dr. C V R Murty
0 Δw Δy Δmax
Roof Displacement (Δ)
Maximum Elastic Force (Fel)

Response Reduction Factor =
Design Force (Fdes)
IS 1893 (Part 1):2002
Response reduction factor (contd..)

A structure with good ductility, redundancy and
overstrength is designed for smaller seismic force
and has higher value of R
For example, building with SMRF has good ductility and has
R = 5.0 as against R = 1.5 for unreinforced masonry
building which does not have good ductility
Table 7 gives R values for buildings
For tanks, R values will be given in IS:1893 (Part 2)
IS 1893 (Part 1):2002
Spectral acceleration coefficient, Sa/g

Depends on structural characteristics and soil
condition
Structural characteristics include time period and damping
Refer Fig. 2 and Table 3 of IS:1893
See next slide
IS 1893 (Part 1):2002
For 5% damping
IS 1893 (Part 1):2002
For other damping, Sa/g values are to be

multiplied by a factor given in Table 3 of
IS:1893
Table 3 is reproduced below
% 0 2 5 7 10 15 20 25 30
damping
Factor 3.20 1.40 1.00 0.90 0.80 0.70 0.60 0.55 0.50
For higher damping, multiplying factor is less

Hence, for higher damping, Sa/g is less
IS 1893:1984
Let us now look at the provision of IS

1893:1984
IS 1893:1984 suggests two methods for
calculating seismic forces
Seismic coefficient method (SCM)
Response spectrum method (RSM)
These have different expressions for base shear
coefficient
IS 1893:1984
Ah= KCβIαo Seismic Coefficient Method (SCM)

= KβIFoSa/g Response Spectrum Method (RSM)
K is performance factor
C is a coefficient which depends on time period
β is soil-foundation system coefficient
I is importance factor
αo is seismic coefficient
Fo is zone factor
Sa/g is average acceleration coefficient
IS 1893:1984
Seismic coefficient, αo
Used in seismic coefficient method
Zone factor, Fo
Used in response spectrum method
IS 1893:1984
IS 1893:1984
β is soil foundation coefficient

Depends on type of soil and foundation
In IS 1893:2002, type of foundation does not have any
influence on base shear coefficient
IS 1893:1984
Ensures higher design seismic force for more
important structures
IS 1893 (Part 1):2002, gives values only for buildings
IS 1893:1984
Performance factor, K
Depends on ductility of structure
Similar to response reduction factor of IS1893(Part 1):2002
K is in numerator whereas, R is in denominator
For buildings with good ductility, K = 1.0
For ordinary buildings, K = 1.6
Thus, a building with good ductility will have
lower value of base shear coefficient than
ordinary building
IS 1893:1984
IS 1893:1984
Coefficient, C
Depends on time period
see next slide
Spectral acceleration, Sa/g
Depends on time period and damping
See next slide
IS 1893:1984
Graphs for C and Sa/g from IS 1893:1984
Natural Period (Sec) Natural Period (Sec)
IS 1893:1984
IS 1893:1984 has provisions for elevated tanks

only
Ground supported tanks are not considered
For elevated tanks, it suggests
Ah = βIFoSa/g
Performance factor, K is not present
Implies, K = 1.0 for all types of elevated tanks
Unlike buildings, different types of tanks do not have
different values of K
This is one of the major limitation of IS1893:1984
More about it, later
IBC 2003
International Building Code (IBC) 2003
In IITK-GSDMA guidelines IBC 2000 is referred
This is now upgraded to IBC 2003
In USA codes are regularly upgraded every three year
There is no change in the base shear coefficient
expression from IBC 2000 to IBC 2003
IBC 2003
Base shear coefficient

Ah = SD1 I/(R T)
≤ SDS I/R
Ah shall not be less than 0.044 SDSI for buildings
and not less than 0.14 SDSI for tanks
This is a lower limit on Ah
It ensures minimum design seismic force
This lower limit is higher for tanks than for buildings
Variation with time period is directly given in
base shear coefficient
Hence, no need to have response spectrum
separately
IBC 2003
T is time period in seconds

SDS and SD1 are design spectral accelerations in
short period and at 1 sec. respectively
SDS and SD1 depend on seismic zone and soil
I is importance factor and R is response
modification factor
IBC suggests I = 1.0, 1.25 and 1.5 for different
types of structures
Values of R will be discussed later
IBC 2003
More about SDS and SD1

SDS = 2/3 Fa SS and SD1 = 2/3 Fv S1
SS is mapped spectral acceleration for short
period
S1 is mapped spectral acceleration for 1-second
period
SS and S1 are obtained from seismic map
This is similar to zone map of our code
It is given in contour form
Fa and Fv are site coefficients
Their values for are given for different soil types
IBC 2003
Response modification factor, R

IS 1893(Part 1):2002 calls it response reduction
factor
Values of R for some selected structures are
given in next slide
IBC 2003
Type of structure R
Building with special reinforced concrete moment resisting concrete frames 8.0
Building with intermediate reinforced concrete moment resisting concrete frames 5.0
Building with ordinary reinforced concrete moment resisting concrete frames 3.0
Building with special steel concentrically braced frames 8.0

Elevated tanks supported on braced/unbraced legs 3.0
Elevated tanks supported on single pedestal 2.0
Tanks supported on structural towers similar to buildings 3.0
Flat bottom ground supported anchored steel tanks 3.0
Flat bottom ground supported unanchored steel tanks 2.5
Ground supported reinforced or prestressed concrete tanks with reinforced
2.0
nonsliding base
In summary,
Base shear coefficient from these three codes are:
IS 1893 (Part 1): 2002 IS 1893: 1984 IBC2003
Ah = (Z/2).(I/R).(Sa/g) SCM: Ah = KCβIαo Ah = SD1 I/(R T)

RSM: Ah = KβIFoSa/g ≤ SDS I/R
> 0.044 SDS I for buildings
For tanks: > 0.14 SDS I for tanks
Ah = βIFoSa/g
Important to note that:

IS codes specify base shear coefficient at
working stress level
For limit state design, these are to be multiplied
by load factors to get factored loads
IBC specifies base shear coefficient at ultimate
load level
For working stress design, seismic forces are
divided by a factor of 1.4
Once, base shear coefficient is known, seismic

force on the structure can be obtained
Recall, seismic force, V = Ah. W
This is same as force = mass x acceleration
Let us compare base shear coefficient values

from these codes
Comparison will be done at working stress level
IBC values are divided by 1.4 to bring them to working
stress level
This shall be done for similar seismic zone or
seismic hazard level of each code
This comparison is first done for buildings
Comparison for buildings

Following parameters are chosen
IS 1893 (Part 1): 2002 IS 1893: 1984 IBC2003
Z = 0.36; Zone V αo = 0.08; Fo = 0.4; SDs = 1.0; SD1 = 0.6
I = 1.0; R = 5.0 Zone V I = 1.0; R = 8.0
Soft soil β = 1.0 Soil type D, equivalent to
5% damping K = 1.0; I = 1.0 soft soil of IS codes
Soft soil, raft foundation 5% damping
5% damping
They represent similar seismic hazard level
Building with good ductility is chosen

Say, buildings with SMRF
In IBC, for buildings with SMRF, R = 8.0
Refer Table shown earlier
For building with T = 0.3 sec

IS 1893(Part 1):2002
Sa/g =2.5
Ah = Z/2.I/R.Sa/g = (0.36/2 )x (1.0/5.0) x 2.5 = 0.09
IS 1893:1984
C = 1.0 and Sa/g = 0.2
SCM: Ah = KCβIαo = 1.0 x 1.0 x1.0 x1.0x0.08 = 0.08
RSM: Ah = KβIFoSa/g = 1.0 x 1.0 x1.0x 0.4x0.2 = 0.08
IBC 2003
Ah = SDSI/(1.4xR) = 1.0 x1.0/(1.4 x 8.0) = 0.089
For building with T = 1 sec

IS 1893(Part 1):2002
Sa/g = 1.67
Ah = Z/2.I/R.Sa/g = (0.36/2 )x(1.0/5.0)x1.67 = 0.06
IS 1893:1984
C = 0.53 and Sa/g = 0.11
SCM: Ah = KCβIαo = 1.0 x0.53x1.0 x1.0x0.08 = 0.042
RSM: Ah = KβIFoSa/g = 1.0x1.0x1.0x0.4x0.11 = 0.044
IBC 2003
Ah = SD1I/(1.4xRxT) = 0.6x1.0/(1.4 x 8.0x1.0) = 0.054
For building with T = 1.5 sec

IS 1893(Part 1):2002
Sa/g = 1.11
Ah = Z/2.I/R.Sa/g = (0.36/2 )x(1.0/5.0)x1.11 = 0.040
IS 1893:1984
C = 0.4 and Sa/g = 0.078
SCM: Ah = KCβIαo = 1.0 x0.4x1.0 x1.0x0.08 = 0.032
RSM: Ah = KβIFoSa/g =1.0x1.0x1.0x0.4x0.078 = 0.031
IBC 2003
Ah = SD1I/(1.4RT) = 0.6x1.0/(1.4 x 8.0x1.5) = 0.036
Base shear coefficients for four time periods
T IS 1893 IS 1893: 1984 IBC2003

(Sec) (Part 1):
2002 SCM RSM
0.3 0.09 0.08 0.08 0.089

1.0 0.06 0.042 0.044 0.054
1.5 0.040 0.032 0.031 0.036
2.0 0.03 0.024 0.024 0.0314*

* Due to lower bound, this value is higher
Graphical comparison on next slide
Comparison of base shear coefficient (Buildings)

0.1
IS 1893(Part 1):2002
IBC 2003
0.075
IS 1893:1984; RSM Note the lower

IS 1893:1984; SCM bound of IBC
0.05
0.025
0
0 0.5 1 1.5 2 2.5 3
Time Period (S)
We have seen that:

Codes follow similar strategy to obtain design
base shear coefficient
In similar seismic zones, base shear coefficient for
buildings matches reasonably well from these
three codes
Similarly, let us compare design base shear

coefficients for tanks
From IS1893:1984 and IBC 2003
IS 1893(Part 1):2002 is only for buildings
Hence, can’t be used for tanks
Only elevated tanks will be considered
IS 1893:1984 has provisions for elevated tanks only
Zone and soil parameters will remain same as
those considered for buildings
Importance factor for tanks are different than
those for buildings
In IBC
I = 1.25 for tanks
R = 3.0 for tanks on frame staging (braced legs)
R = 2.0 for tanks on shaft or pedestal
In 1893:1984
I = 1.5 for tanks
K is not present in the expression for base shear
coefficient (implies k=1.0). Hence, base shear
coefficient will be same for all types of elevated
tanks
For tank with T = 0.3 sec

IS 1893:1984
I = 1.5, Sa/g = 0.2
Ah = βIFoSa/g = 1.0 x 1.5 x 0.4x0.2 = 0.12
This value is common for frame and shaft staging
IBC 2003
For frame staging, I = 1.25, R = 3.0
Ah = SDSI/(1.4xR) = 1.0 x1.25/(1.4 x 3.0) = 0.298
For shaft staging, I = 1.25, R = 2.0
Ah = SDSI/(1.4xR) = 1.0 x1.25/(1.4 x 2.0) = 0.446
For tank with T = 1 sec

IS 1893:1984
I = 1.5, Sa/g = 0.11
Ah = βIFoSa/g =1.0x1.5x0.4x0.11 = 0.066
IBC 2003
For frame staging, I = 1.25, R = 3.0
Ah = SD1I/(1.4xRxT) = 0.6x1.25/(1.4 x 3.0x1.0) = 0.178
For shaft staging, I = 1.25, R = 2.0
Ah = SD1I/(1.4xRxT) = 0.6x1.25/(1.4 x 2.0x1.0) = 0.268
Base shear coefficients for tanks
T IS 1893:1984* IBC 2003

(Sec)
Frame staging Shaft staging
0.3 0.12 0.298 0.446
1.0 0.066 0.178 0.268

* Base shear coefficient values are common for frame and shaft staging
Graphical comparison on next slide
Comparison of base shear coefficient (Tanks)

0.5
IBC 2003; Tanks on shaft staging
0.4
IBC 2003; Tanks on frame staging

0.3
IS 1893:1984; All types of
elevated tanks
0.2
0.1
0
0 0.5 1 1.5 2 2.5 3
Time period (S)
Base shear coefficient for elevated tanks from

IS1893:1984 is on much lower side than IBC
2003
IBC value is about 2.5 times for frame staging
and 3.5 times for shaft staging than that from
IS1893:1984
Recall, for buildings, IS 1893:1984 and IBC have
much better comparison
Reason for lower values in IS 1893:1984

IBC uses R = 2.0 and R = 3.0 for tanks as against R
= 8.0 for buildings with good ductility
IS 1893:1984 uses K = 1.0 for tanks. Same as for
buildings with good ductility.
Clearly ,elevated tanks do not have same
ductility, redundancy and overstrength as
buildings.
This is a major limitation of IS 1893:1984
Another limitation of IS 1893:1984

In Lecture 1, we have seen, liquid mass gets
divided into impulsive and convective masses
IS 1893:1984, does not consider convective mass
It assumes entire liquid mass will act as impulsive
mass, rigidly attached to wall
In IITK-GSDMA guidelines, these limitations
have been removed
Let us now, get back to seismic force evaluation

for tanks
Design base shear coefficient is to be expressed
in terms of parameters of IS 1893(Part 1):2002
Ah = (Z/2). (I/R). Sa/g
Z will be governed by seismic zone map of Part 1
I and R for tanks will be different from those for
buildings
R depends on ductility, redundancy and overstrength
Sa/g will depend on time period
Impulsive and convective masses will have

different time periods
Hence, will have different Sa/g values
Procedure for finding time period in next lecture
Seismic force = (Ah) X (W)
Base shear coefficient, Ah, depends on
Seismic Zone
Soil type
Ductility, Redundancy and overstrength
IS 1893:1984 has some serious limitations in
design seismic force for tanks
Sesimic Design of Liquid Storage Tanks


Question from Mr. Abdul Gani (gani@igcar.ernet.in)
The following are my questions on Lecture 2.
1. What exactly is meant by ‘Overstrength’?
Overstrength means additional strength which structure possess over and

above the design strength. This additional strength comes from factors on
gravity loads, live loads and earthquake loads. For limit state design of RC
structures, we use partial safety factor of 1.5 on gravity loads. Partial
safety factors on materials also provide overstrength. Moreover, materials
may have strength greater than their characteristic strength, Non-
structural elements and special ductile detailing also constitute another
source of overstrength. Please note, in design we do not include strength
provided by these sources, however, actually it is present in the structure.
2. As per fig C-4a of the Guidelines, the base shear coefficients from IS 1893
(Part 1):2003 is much less than that from IBC 2000 for structures with time
period less than 0.1s and greater than 1.7s. How to account for this effect
while designing a structure which falls in this time period zone?
Figure C-4a gives comparison of base shear coefficient for buildings from
IS 1893(Part 1):2002 and IBC 2000. In IBC 2000, spectrum is flat (or
horizontal) from T = 0.0 sec itself, whereas, in IS 1893, there is a rising
portion from T = 0.0 up to T = 0.1 sec.
A very stiff or rigid structure will have very low time period, which may fall
in this range. For such rigid structures, Sa/g and base shear coefficient will
be less. In such rigid structures, reduction in seismic forces on account of
ductility is not allowed. Thus, strictly speaking, use of R for these
structures is not appropriate. In order to address this issue, codes make
spectrum flat in short period range and allow the use of R values. IS
1893(Part 1) should in fact remove this rising portion. More discussion on
this aspect is available in the following document:
Proposed draft provisions and commentary on IS 1893(Part 1):2002
(www.iitk.ac.in/nicee/IITK-GSDMA/EQ05.pdf)
Please note, for tanks, this rising portion has been removed in the IITK-
GSDMA Guidelines.
For time period greater than 1.7 sec also there is difference between IBC
and IS 1893(Part 1). This is due to the fact that IBC has given a lower limit
on its base shear coefficient, which means to say that, beyond certain time
period, design seismic force will remain same. This is done to ensure
certain minimum strength against lateral loads. Moreover, if time period is
incorrectly estimated on higher side, then also this lower limit acts as a
safeguard. IS 1893 does not have such lower limit and for tanks also in
the IITK-GSDMA Guideline, as of now there is no such lower limit.
3. Why response reduction factor is different for convective and impulsive

masses?
In IITK-GSDMA Guideline, response reduction factor is same for impulsive

and convective modes.
However, in some international codes (like, ACI 350.3 and AWWA
D-115), response reduction factors are different for impulsive and
convective modes. These codes argue that convective mode is low
frequency mode (large time period) and hence, convective forces would
vary slowly with time. That means, convective forces are like static forces
and hence, reduction due to ductility shall not be applicable to them. ACI
350.3 and AWWA D-115 use R = 1.0 for convective mode, i.e. they do not
allow any reduction in convective seismic forces. However, it is important
to note that, these codes use different response spectrum for convective
and impulsive modes. For convective mode, which is usually having very
large time period, response spectrum has 1/Tc2 variation in long period
range. On the other hand, IS 1893 spectrum has 1/T variation for all time
periods.
At present, in IITK-GSDMA Guideline, R is kept same for impulsive
and convective modes and same spectrum is used for impulsive and
convective modes. This makes calculations simple and does not cause
significant error.
4. How to find base shear coefficient based on IS 1893:2003 for two identical
buildings designed for 50 and 100 years respectively.
At present, as per IS 1893:2002, base shear coefficient will be same for

both the buildings. Please note, if design forces have to depend on
expected life of building, then zone map or zone factor or PGA will have to
be arrived at using probabilistic seismic hazard analysis (Refer
“Geotechnical Earthquake Engineering” by Kramer S L; Pearson
Education, first Indian reprint, 2003). In probabilistic analysis, PGA is
specified with certain probability of exceedance in given number of years.
For example, IBC specifies maximum considered earthquake with 2%
probability of being exceeded in 50 years (2500 year return period). Indian
codes have not yet adopted such an approach for quantifying seismic
hazards.
5. Why IS code specifies base shear coefficients at working stress level, while
IBC specifies at Ultimate stress level?
In India, working stress method is still quite common. Hence, IS 1893 has
given design spectrum at working stress level. However, it does not really
matter. We use load factors to arrive at loads to be used for limit state
design. Users of IBC use a factor of 0.7 to arrive at forces for working
stress design.
6. In some of the international codes (eg. New Zealand and French codes), the
spectral amplification factor for hard rock is higher followed by medium and
soft soils. Why is it so? But in IS 1893:2003, the spectral amplifications are
same for both rock and soil sites. Can please explain why?
In hard soil, high frequency (low time period) waves get amplified and in
soft soil, low frequency (long time period) waves get amplified. Thus, in
low time period range, spectral amplification factor will be higher for hard
soil. Similarly, in long time period range, spectral amplification for soft soil
will be higher. The spectral amplification factors are obtained from
recorded data of ground motion in various soil conditions. As of now, IS
1893 (Part 1) has kept same spectral amplification in short period range
for all types of soils. With the availability of more reliable data, spectra in
short period range may be modified for different soil conditions.
7. In some of the standards (eg. TecDoc 1347) provision is made to account for
local site effects (Site amplification etc). In IS 1893:2002 why no such provision
is made?
Site effects in IS 1893 are included by specifying different design spectra

for rock, medium and soft soils, especially accounting for the higher
ground velocities observed in soft soils. Other topographic effects, such as
hill slopes, ridges valleys etc. are currently not included in the design force
calculations, as available information is not sufficient for reliable estimation
of their effects.
8. In IBC, what is the time period at which SDS is specified?
SDS is specified in short period range and there is no specific value of time
period at which it is specified. One may assume that this period will be
less than approximately 0.1 sec. Unlike this, SD1 is specified at 1.0 sec
time period.
Lecture 3
January 23, 2006

In this lecture
Modeling of tanks
Time period of tanks
Modeling of tanks
As seen in Lecture 1 liquid may be replaced by
impulsive and convective mass for calculation of
hydrodynamic forces
See next slide for a quick review
Modeling of tanks
mi = Impulsive liquid mass
mc = Convective liquid mass
Kc/2 Kc/2
Kc = Convective spring stiffness
mc
hi = Location of impulsive mass
(without considering overturnig
hc caused by base pressure)
Rigid m (hc*)
i
hc = Location of convective mass
hi
(without considering overturning
(hi*) caused by base pressure)
Mechanical analogue hi* = Location of impulsive mass

or overturning)
spring mass model of tank hc* = Location of convective mass
overturning)
Graphs and expression for these parameters are given in lecture 1.
Approximation in modeling
Sometimes, summation of mi and mc may not be
equal to total liquid mass, m
This difference may be about 2 to 3 %
Difference arises due to approximations in the
derivation of these expressions
More about it, later
If this difference is of concern, then
First, obtain mc from the graph or expression
Obtain mi = m – mc
Tanks of other shapes
For tank shapes such as Intze, funnel, etc. :
Consider equivalent circular tank of same
volume, with diameter equal to diameter at the
top level of liquid
Example:
An Intze container has volume of 1000 m3. Diameter of
container at top level of liquid is 16 m. Find dimensions of
equivalent circular container for computation of
hydrodynamic forces.
Equivalent circular container will have diameter of 16 m

and volume of 1000 m3. Height of liquid, h can be obtained
as :
π/4 x 162 x h = 1000
∴ h = 1000 x 4/(π x 162) = 4.97 m
Thus, for equivalent circular container,

h/D = 4.97/16 = 0.31
All the parameters (such as mi, mc etc.) are to be
obtained using h/D = 0.31
16 m
16 m 4.97 m
Equivalent circular container

Intze container
volume = 1000 m3
Effect of obstructions inside tank
Container may have structural elements inside
For example: central shaft, columns supporting
the roof slab, and baffle walls
These elements cause obstruction to lateral
motion of liquid
This will affect impulsive and convective masses
Effect of these obstructions on impulsive and

convective mass is not well studied
A good research topic !
It is clear that these elements will reduce
convective (or sloshing) mass
More liquid will act as impulsive mass
In the absence of detailed analysis, following

approximation may be adopted:
Consider a circular or a rectangular container of
same height and without any internal elements
Equate the volume of this container to net
volume of original container
This will give diameter or lateral dimensions of container
Use this container to obtain h/D or h/L
Example: A circular cylindrical container has internal

diameter of 12 m and liquid height of 4 m. At the center
of the tank there is a circular shaft of outer diameter of
2 m. Find the dimensions of equivalent circular
cylindrical tank.
12 m
4m
12 m
Hollow shaft of Plan
Elevation 2 m diameter
Solution:
Net volume of container = π/4x(122 –22)x4 = 439.8 m3
Equivalent cylinder will have liquid height of 4 m and its
volume has to be 439.8 m3.
Let D be the diameter of equivalent circular cylinder, then
π/4xD2x4 = 439.8 m3
∴ D = 11.83 m
Thus, for equivalent circular tank, h = 4 m, D = 11.83m
and h/D = 4/11.83 = 0.34.
This h/D shall be used to find parameters of mechanical
model of tank
Effect of wall flexibility
Parameters mi, mc etc. are obtained assuming

tank wall to be rigid
An assumption in the original work of Housner
(1963a)
Housner, G. W., 1963a, “Dynamic analysis of fluids in
containers subjected to acceleration”, Nuclear Reactors and
Earthquakes, Report No. TID 7024, U. S. Atomic Energy
Commission, Washington D.C.
RC tank walls are quite rigid
Steel tank walls may be flexible
Particularly, in case of tall steel tanks
Wall flexibility affects impulsive pressure distribution

It does not substantially affect convective pressure
distribution
Refer Veletsos, Haroun and Housner (1984)
Veletsos, A. S., 1984, “Seismic response and design of liquid
storage tanks”, Guidelines for the seismic design of oil and gas
pipeline systems, Technical Council on Lifeline Earthquake
1Engineering, ASCE, N.Y., 255-370, 443-461.
Haroun, M. A. and Housner, G. W., 1984, “Seismic design of liquid
storage tanks”, Journal of Technical Councils of ASCE, Vol. 107,
TC1, 191-207.
Effect of wall flexibility on impulsive pressure depends
on
Aspect ratio of tank
Ratio of wall thickness to diameter
See next slide
Effect of wall flexibility on impulsive pressure distribution

h/D = 0.5
tw / D = 0.0005
tw / D = 0.0005
tw is wall thickness
z Rigid
h tank
Impulsive pressure on wall From Veletsos (1984)
If wall flexibility is included, then mechanical

model of tank becomes more complicated
Moreover, its inclusion does not change seismic
forces appreciably
Thus, mechanical model based on rigid wall
assumption is considered adequate for design.
All international codes use rigid wall model for

RC as well as steel tanks
Only exception is NZSEE recommendation
(Priestley et al., 1986)
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.
American Petroleum Institute (API) standards,
which are exclusively for steel tanks, also use
mechanical model based on rigid wall
assumption
API 650, 1998, “Welded storage tanks for oil storage”,
American Petroleum Institute, Washington D. C., USA.
Effect of higher modes
mi and mc described in Lecture 1, correspond
to first impulsive and convective modes
For most tanks ( 0.15 < h/D < 1.5) the first
impulsive and convective modes together
account for 85 to 98% of total liquid mass
Hence, higher modes are not included
This is also one of the reasons for summation of
mi and mc being not equal to total liquid mass
For more information refer Veletsos (1984) and
Malhotra (2000)
Malhotra, P. K., Wenk, T. and Wieland, M., 2000, “Simple
procedure for seismic analysis of liquid-storage tanks”,
Structural Engineering International, 197-201.
Modeling of ground supported tanks
Step 1:
Obtain various parameters of mechanical model
These include, mi, mc, Kc, hi, hc, hi* and hc*
Step 2:
Calculate mass of tank wall (mw), mass of roof
(mt) and mass of base slab (mb)of container
This completes modeling of ground supported
tanks
Modeling of elevated tanks
Elevated tank consists of container and staging
Roof slab
Wall Container
Floor slab
Staging
Elevated tank
Liquid is replaced by impulsive and convective

masses, mi and mc
All other parameters such as hi, hc, etc, shall be
obtained as described earlier
Lateral stiffness, Ks, of staging must be
considered
This makes it a two-degree-of-freedom model
Also called two mass idealization
mc
Kc/2 Kc/2
mc Kc
mi hc
hi
mi + ms
hs
Ks
Spring mass model Two degree of freedom system

OR
Two mass idealization of elevated tanks
ms is structural mass, which comprises of :

Mass of container, and
One-third mass of staging
Mass of container includes
Mass of roof slab
Mass of wall
Mass of floor slab and beams
Two Degree of Freedom System
2-DoF system requires solution of a 2 × 2 eigen

value problem to obtain
Two natural time periods
Corresponding mode shapes
See any standard text book on structural
dynamics on how to solve 2-DoF system
For most elevated tanks, the two natural time
periods (T1 and T2) are well separated.
T1 generally may exceed 2.5 times T2.
Two Degree of Freedom System
Hence the 2-DoF system can be treated as two

uncoupled single degree of freedom systems
One representing mi +ms and Ks
Second representing mc and Kc
mc
Kc mi + ms
mi + ms
Ks mc
Ks Kc
Two uncoupled
Two degree of freedom system single degree of freedom systems
when T1 ≥ 2.5 T2
Priestley et al. (1986) suggested that this

approximation is reasonable if ratio of two time
periods exceeds 2.5
Important to note that this approximation is
done only for the purpose of calculating time
periods
This significantly simplifies time period calculation
Otherwise, one can obtain time periods of 2-DoF
system as per procedure of structural dynamics.
Steps in modeling of elevated tanks

Step 1:
Obtain parameters of mechanical analogue
These include mi, mc, Kc, hi, hc, hi* and hc*
Other tank shapes and obstructions inside the container shall
be handled as described earlier
Step 2:
Calculate mass of container and mass of staging
Step 3:
Obtain stiffness of staging
Recall, in IS 1893:1984, convective mass is not

considered
It assumes entire liquid will act as impulsive mass
Hence, elevated tank is modeled as single
degree of freedom ( SDoF) system
As against this, now, elevated tank is modeled
as 2-DoF system
This 2-DoF system can be treated as two
uncoupled SDoF systems
Models of elevated tanks
m +ms m = Total liquid mass

mi + ms
mc Ks
Ks
Kc
As per the Guideline As per IS 1893:1984
Example: An elevated tank with circular cylindrical

container has internal diameter of 11.3 m and water
height is 3 m. Container mass is 180 t and staging mass
is 100 t. Lateral stiffness of staging is 20,000 kN/m.
Model the tank using the Guideline and IS 1893:1984
Solution:
Internal diameter, D = 11.3 m, Water height, h = 3 m.
Container is circular cylinder,
∴ Volume of water = π/4 x D2 x h
= π /4 x 11.32 x 3 = 300.9 m3.
∴ mass of water, m = 300.9 t.
h/D = 3/11.3 = 0.265

From Figure 2 of the Guideline, for h/D = 0.265:
mi/m = 0.31, mc/m = 0.65 and Kch/mg = 0.47
mi = 0.31 x m = 0.31 x 300.9 = 93.3 t

mc = 0.65 x m = 0.65 x 300.9 = 195.6 t
Kc = 0.47 x mg/h
= 0.47 x 300.9 x 9.81/3 = 462.5 kN/m
Mass of container = 180 t

Mass of staging = 100 t
Structural mass of tank, ms

= mass of container +1/3rd mass of staging
= 180 +1/3 x 100
= 213.3 t
Lateral stiffness of staging, Ks = 20,000 kN/m
mi + ms m + ms
Ks mc Ks
Kc
mi = 93.3 t, ms = 213.3 t, mc = 195.6 t, m = 300.9 t, ms = 213.3 t,

Ks = 20,000 kN/m, Kc = 462.5 kN/m Ks = 20,000 kN/m
Model of tank as per the Guideline Model of tank as per IS 1893:1984
Time period
What is time period ?
For a single degree of freedom system, time
period (T ) is given by
M
T = 2π
K
M is mass and K is stiffness

T is in seconds
M should be in kg; K should be in Newton per
meter (N/m)
Else, M can be in Tonnes and K in kN/m
Time period
Mathematical model of tank comprises of

impulsive and convective components
Hence, time periods of impulsive and convective
mode are to be obtained
Time period of impulsive mode
Procedure to obtain time period of impulsive
mode (Ti) will be described for following three
cases:
Ground supported circular tanks
Ground supported rectangular tanks
Elevated tanks
Ti for ground-supported circular tanks
Ground supported circular tanks

Time period of impulsive mode, Ti is given by:
h ρ ⎛ 1 ⎞
Ti = Ci Ci = ⎜ ⎟
t/D E ⎝ (
⎜ h/D 0.46 − 0.3h/D + 0.067(h/D)2 ) ⎟
⎠
ρ = Mass density of liquid

E = Young’s modulus of tank material
t = Wall thickness
h = Height of liquid
D = Diameter of tank
Ci can also be obtained from Figure 5 of the

Guidelines
10
6
C
C
i
4
C
c
0
0 0.5 h/D 1 1.5 2
This formula is taken from Eurocode 8

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.
If wall thickness varies with height, then
thickness at 1/3rd height from bottom shall be
used
Some steel tanks may have step variation of wall
thickness with height
This formula is derived based on assumption

that wall mass is quite small compared to liquid
mass
More information on time period of circular
tanks may be seen in Veletsos (1984) and
Nachtigall et al. (2003)
Nachtigall, I., Gebbeken, N. and Urrutia-Galicia, J. L., 2003, “On the
analysis of vertical circular cylindrical tanks under earthquake
excitation at its base”, Engineering Structures, Vol. 25, 201-213.
It is important to note that wall flexibility is

considered in this formula
For tanks with rigid wall, time period will be zero
This should not be confused with rigid wall
assumption in the derivation of mi and mc
Wall flexibility is neglected only in the evaluation
of impulsive and convective masses
However, wall flexibility is included while
calculating time period
This formula is applicable to tanks with fixed

base condition
i.e., tank wall is rigidly connected or fixed to the
base slab
In some circular tanks, wall and base have
flexible connections
Ground supported tanks with flexible base are

described in ACI 350.3 and AWWA D-110
ACI 350.3, 2001, “Seismic design of liquid containing
concrete structures”, American Concrete Institute,
Farmington Hill, MI, USA.
AWWA D-110, 1995, “Wire- and strand-wound circular,
prestressed concrete water tanks”, American Water Works
Association, Colorado, USA.
In these tanks, there is a flexible pad between
wall and base
Refer Figure 6 of the Guideline
Types of connections between tank wall and base slab
Such tanks are perhaps not used in India
Impulsive mode time period of ground

supported tanks with fixed base is generally
very low
These tanks are quite rigid
Ti will usually be less than 0.4 seconds
In this short period range, spectral acceleration,
Sa/g has constant value
See next slide
Impulsive mode time period of ground supported

tanks likely to remain in this range
Sa/g
Example: A ground supported steel tank has water

height, h = 25 m, internal diameter, D = 15 m and
wall thickness, t=15 mm. Find time period of
impulsive mode.
Solution: h = 25 m, D = 15 m, t = 15 mm.
For water, mass density, ρ = 1 t/m3.
For steel, Young’s modulus, E = 2x108 kN/m2.
h/D = 25/15 = 1.67. From Figure 5, Ci = 5.3
h ρ
Time period of impulsive mode, Ti = Ci
t/D E
25 1.0
Ti = 5.3
0.015/15 2x10 8
= 0.30 sec
Important to note that, even for such a slender tank of

steel, time period is low.
For RC tanks and other short tanks, time period will be
further less.
In view of this, no point in putting too much

emphasis on evaluation of impulsive mode time
period for ground supported tanks
Recognizing this point, API standards have
suggested a constant value of spectral
acceleration for ground supported circular steel
tanks
Thus, users of API standards need not find
impulsive time period of ground supported tanks
Ti for ground-supported rectangular tanks
Ti for ground-supported rectangular tanks

Procedure to find time period of impulsive mode
is described in Clause no. 4.3.1.2 of the
Guidelines
This will not be repeated here
Time period is likely to be very low and Sa/g will
remain constant
As described earlier
Hence, not much emphasis on time period
evaluation
Ti for Elevated tanks
For elevated tanks, flexibility of staging is important
Time period of impulsive mode, Ti is given by:
mi + ms Δ
Ti = 2π OR T = 2π
Ks g
mi = Impulsive mass of liquid
ms = Mass of container and one-third mass of staging
Ks = Lateral stiffness of staging
Δ= Horizontal deflection of center of gravity of tank when a
horizontal force equal to (mi + ms)g is applied at the
center of gravity of tank
These two formulae are one and the same

Expressed in terms of different quantities
Center of gravity of tank refers to combined
mass center of empty container plus impulsive
mass of liquid
Example: An elevated tank stores 250 t of water. Ratio of

water height to internal diameter of container is 0.5.
Container mass is 150 t and staging mass is 90 t.
Lateral stiffness of staging is 20,000 kN/m. Find time
period of impulsive mode
Solution: h/D = 0.5, Hence from Figure 2a of the

Guideline, mi/m = 0.54;
∴ mi = 0.54 x 250 = 135 t
Structural mass of tank, ms
= mass of container + 1/3rd mass of staging
= 150 + 90/3 = 180 t
mi + ms
Time period of impulsive mode Ti = 2π
Ks
135 + 180
Ti = 2π
20,000
= 0.79 sec.
Lateral stiffness of staging, Ks
Lateral stiffness of staging, Ks is force required
to be applied at CG of tank to cause a
corresponding unit horizontal deflection
CG
P δ Ks = P/ δ
For frame type staging, lateral stiffness shall be

obtained by suitably modeling columns and
braces
More information can be seen in Sameer and
Jain (1992, 1994)
Sameer, S. U., and Jain, S. K., 1992, “Approximate
methods for determination of time period of water tank
staging”, The Indian Concrete Journal, Vol. 66, No. 12,
691-698.
Sameer, S. U., and Jain, S. K., 1994, “Lateral load analysis
of frame staging for elevated water tanks”, Journal of
Structural Engineering, ASCE, Vol.120, No.5, 1375-1393.
Some commonly used frame type staging
configurations are shown in next slide
Plan view of frame staging configurations
4 columns 6 columns 8 columns
9 columns 12 columns
24 columns 52 columns
Explanatory handbook, SP:22 has considered

braces as rigid beams
SP:22 – 1982, Explanatory Handbook on Codes for
Earthquake Engineering, Bureau of Indian Standards, New
Delhi
This is unrealistic modeling
Leads to lower time period
Hence, higher base shear coefficient
This is another limitation of IS 1893:1984
Using a standard structural analysis software,
staging can be modeled and analyzed to
estimate lateral stiffness
Shaft type staging can be treated as a vertical

cantilever fixed at base and free at top
If flexural behavior is dominant, then
Its stiffness will be Ks = 3EI/L3
This will be a good approximation if height to
diameter ratio is greater than two
Otherwise, shear deformations of shaft would
affect the stiffness and should be included.
Time period of convective mode
Convective mass is mc and stiffness is Kc
Time period of convective mode is:
m
T = 2π
c
c
K c
mc and Kc for circular and rectangular tanks can

be obtained from graphs or expressions
These are described in Lecture 1
Refer Figures 2 and 3 of the Guidelines
For further simplification, expressions for mc

and Kc are substituted in the formula for Tc
Then one gets,
For circular tanks:
2π
Cc =
Tc = Cc D/g 3.68 tanh (3.68h / D)
For rectangular tanks:

2π
Tc = Cc L/g Cc =
3.16 tanh (3.16(h / L))
Graphs for obtaining Cc are given in Figures 5

and 7 of the Guidelines
These are reproduced in next two slides
Convective mass and stiffness are not affected
by flexibility of base or staging
Hence, convective time period expressions are
common for ground supported as well as
elevated tanks
Convective mode time periods are usually very
large
Their values can be as high as 10 seconds
10
C
C
4
C
c
0
0 0.5 h/D 1 1.5 2
Fig. 5 For circular tanks
10
6
Cc
0 0.5 1 1.5 2
h/L
Fig. 7 For rectangular tanks
Example: For a circular tank of internal diameter, 12 m

and liquid height of 4 m. Calculate time period of
convective mode.
Solution: h = 4 m, D = 12 m,
∴ h/D = 4/12 = 0.33
From Figure 5 of the Guidelines, Cc = 3.6
Time period of convective mode, Tc = Cc D/g

Tc = 3.6 12/9.81
= 3.98 sec
Based on mechanical models, time period for
impulsive and convective modes can be
obtained for ground supported and elevated
tanks
For ground supported tanks, impulsive mode
time period is likely to be very less
Convective mode time period can be very large


Question from Mr. Murugan.R (rmurugan@igcar.ernet.in )
• In Slide-16: Please correct the tw/D ratio. Both dotted line and thick line are
showing the same ratio of 0.0005.
You are right. Graph with solid line is for tw/D = 0.005
• In Slide-26: We are treating 2-DOF system as two uncoupled SDOF system.

How to combine them?
Response from these two uncoupled SDOF systems is combined using

SRSS rule. Please note, total base shear is SRSS combination of base
shear in impulsive mode (i.e., first SDOF) and convective mode (i.e.,
second SDOF).
• Kindly furnish the formula for time period of a ground supported tanks with
finned base?
Finned base? You are perhaps asking formula for tanks with flexible
base. This has been answered in a subsequent question.
• In Slide-62: Kindly provide the formula for stiffness of staging based on shear
deformation.
This has been discussed in Lecture 7

Question from Mr. H. I. Abdul Gani (gani@igcar.ernet.in )
The following are my doubts regarding lecture 3.
1. How to find the time period of ground supported tanks with flexible
connection?
Tanks with flexible connections are those wherein, between wall and base
there is a flexible pad. These types of tanks are described in ACI 350.3 and
are reproduced in Figure 6 of IITK-GSDMA Guideline. ACI 350.3 suggests
following formula for impulsive mode time period of circular tanks on flexible
base
8π (Ww + Wr + Wi )
Ti =
gDKa
Where, Ww, Wr, and Wi are weights of wall, roof and impulsive liquid
respectively. D is diameter of tank and Ka is stiffness of base pad in the
units of force/m2.
Flexible base does not affect convective mode time.
2. Even though the time period of impulsive mode of ground supported tanks
with flexible base is low, the sa/g values as per IS 1893 is low in this region. How
to account this reduced sa/g values in design?
For tanks, Sa/g graph has been made horizontal in this low period range,
hence, this question will not arise. Please refer Clause 4.5.2 of the
Guideline.
3. How to find the damping if impulsive and convective time periods are not
well separated?
If two time periods are not well separated, then we have to solve 2-DoF
system, wherein, damping is different for both the DoF. This will be non-
classically damped system. This has been pointed out in the Commentary to
clause 4.2.2.4. Time period, mode shapes and modal superposition
methods for non-classically damped systems have been developed and this
information is available in some advanced textbooks on Structural
Dynamics.
4. How the flexibility of pad affects the impulsive mode time period?
This question is answered in your first question

5. In the case of ground supported rectangular tanks impulsive and convective
time periods will be different in both directions. Which value is to be adopted for
the design?
In rectangular tanks, for wall design, total lateral force and bending moment
acting on that wall is required. Hence, for the design of a particular wall,
forces arising due to seismic load in the perpendicular direction shall be
considered.
6. How we can incorporate the effect of soil flexibility to find the impulsive mode
time period?
Usually codes suggest to refer specialized literature for soil structure

interaction. Some discussion on this topic is included in Lecture 7.
7. What is the criteria for decoupling mi+ms,ks and mc,kc (apart from Priestly's
method). Can ASCE4-98 guideline for decoupling equipment masses be used
here?
For elevated tanks used in practice, there is a range of various parameters

(mi, mc, Ti and Tc). For most of the tanks, the criterion suggested by
Preistley is OK. However, there may be some extreme cases wherein,
Priestley’s criterion may not work well. You need to be cautious while using
decoupling criterion of ASCE4-98, which is for secondary systems like
equipments.
Interestingly, one can derive this criterion. What you need to do is to solve
2-DoF system using standard modal analysis (undamped case) and find its
time periods and compare them with time periods of decoupled systems.
You will notice that time period of 2-DoF systems depends on ratio of two
masses and ratio of two uncoupled time periods.
8. While decoupling, only 1/3rd of the mass of staging is considered. How the
remaining mass is accounted?
This has been explained in the commentary to Clause 4.2.2.3. Staging acts
like a lateral spring. You consider a SDoF system and include mass of the
spring also. Assuming that spring deflection varies linearly along its length,
one finds that in the time period only one-third mass of spring contributes.
9. In slide 16,
a) On the graph, what does 'z' represent?
b) The curves are presented for h/D = 0.5. For other values are curves
available?
c) tw/D = 0.0005 is represented by both bold and dotted curves. Please correct.
d) For rigid tank, what will be the value of tw/D?
z represents the height at which pressure is to be obtained. Note h is total

depth of liquid and z/h varies from 0 to 1.0. This figure is taken from
Veletsos (1984). More studies on effect of wall flexibility are available in
the literature. Refer paper Haroun, M. A. and Housner, G. W., 1984,
“Seismic design of liquid storage tanks”, Journal of Technical Councils of
ASCE, Vol. 107, TC1, 191-207.You can also try your hands on literature
on pressure vessels. They deal with flexible walls.
Question from Mr. Rushikesh Trivedi (rushikeshtrivedi@yahoo.com)
The formula specified by the code to consider lateral stiffness of shaft type
staging considers Ks = 3*EI /L^3. (Lecture 3, Slide 62, E- Course on Tanks)
As in the case of Tanks supported on framed stagings, here also we should

consider stiffness through a force applied at the CG of container. Hence, the
symbol "L" being referred to in the above formula should be identified as
distance of CG of container from top of Footing.
If we consider "L" to be the distance from top of footing up to the bottom of

container (as considered in Proposed Draft (Example 3, Page 72)), it will be
grossly on conservative side.
Kindly opine on the same.
Some discussion on this issue is given in Lecture 7. It is to be recognized

that we need to model the stiffness of staging properly.
In the simple formula (k = 3 E I /Lcg^3) for lateral stiffness of
staging, one should consider Lcg being the height of cg of container from
the base (with or without water, as the case may be). Of course, this only
considers flexural deformation of the staging and the approximation is
generally acceptable if the relative stiffness of the tank container is within
50% of the relative stiffness of the tank staging.
If the tank container is believed to be more stiff (or rigid), then one
should determine the lateral stiffness from the deflection of the entire
structure acting as a cantilever beam of length Lcg subjected to a
concentrated load at Lcg from the base for which the portion (say L1)
above the tank staging is considered as rigid which undergoes only rigid
body rotation (Refer Figure below).
In frame staging base beams, which are comparatively more rigid,
are also modeled and diaphragm effect is included to account for in-plane
rigidity of slab. The force at CG is applied by putting a rigid link from
staging top to CG. This rigid link does not undergo much higher deflection
than staging top.
Lcg
L1 W
∆
θ
θ
Δ
T = 2π
g
Lecture 4
January 30, 2006

In this lecture
Z, I, Sa/g and R values for tanks
Seismic force V = (Ah) x (W)
Ah is base shear coefficient
⎛Z⎞ ⎛ I ⎞ Sa
Ah = ⎜ ⎟ ⎜ ⎟
⎝ 2 ⎠ ⎝ R⎠ g
Zone Design
philosophy Depends on time period
Depends on severity
and damping
of ground motion
Tanks have two modes

Impulsive
Convective
Seismic force
In impulsive mode, Vi = (Ah)i x impulsive weight
In convective mode, Vc = (Ah)c x convective weight
(Ah)i and (Ah)c are base shear coefficient in

impulsive and convective modes, respectively
Impulsive base shear coefficient

(Ah)i = (Z/2) x (I/R) x (Sa/g)i
Convective base shear coefficient
(Ah)c = (Z/2) x (I/R) x (Sa/g)c
Note, R has been used in (Ah)i as well as (Ah)c
Zone factor, Z
As per Table 2 of IS 1893(Part1):2002
I, R, (Sa/g)i and (Sa/g)c will be discussed here
First, (Sa/g)i and (Sa/g)c
(Sa/g)i and (Sa/g)c
(Sa/g)i is average response acceleration for
impulsive mode
Depends on time period and damping of
impulsive mode
(Sa/g)c is average response acceleration for
convective mode
Depends on time period and damping of
convective mode
(Sa/g)i and (Sa/g)c
Sa/g is obtained from design spectra

Figure 2 of IS 1893(Part 1):2002
These spectra are slightly modified for tanks
See next slide
(Sa/g)i and (Sa/g)c
Modifications are:
The rising portion in short period range from (0 to
0.1 sec) has been made constant
Very stiff structures have time period less than 0.1 sec
There may be modeling errors; actual time period may be
slightly higher
As the structure gets slightly damaged, its natural period
elongates
Ductility does not help in reducing response of very stiff
structures
Hence, rising portion in the range 0 to 0.1 sec is usually
disallowed by the codes.
Spectra is extended beyond 4 sec
Since convective time period may be greater than 4 sec.
Beyond 4 sec, 1/T variation is retained
(Sa/g)i and (Sa/g)c
Sa/g
Sa/g
Sa/g
Modified spectra
Spectra of IS 1893 (Part 1):2002
For 5% damping
(Sa/g)i and (Sa/g)c
Expressions for design spectra at 5% damping

Expressions for spectra of Expressions for spectra for tanks
IS 1893(Part 1):2003
For hard soil sites For hard soil sites
Sa/g = 1 + 15 T 0.00 ≤ T < 0.10 Sa/g = 2.50 T < 0.40
= 2.50 0.10 ≤ T < 0.40 = 1.0 / T T ≥ 0.40
= 1.00 / T 0.40 ≤ T ≤ 4.0
For medium soil sites For medium soil sites
Sa/g = 1 + 15 T 0.00 ≤ T < 0.10 Sa/g = 2.50 T < 0.55
= 2.50 0.10 ≤ T < 0.55 = 1.36 / T T ≥ 0.55
= 1.36 / T 0.55 ≤ T ≤ 4.0
For soft soil sites For soft soil sites
Sa/g = 1 + 15 T 0.00 ≤ T < 0.10 Sa/g = 2.5 T< 0.67
= 2.50 0.10 ≤ T < 0.67 = 1.67 / T T ≥ 0.67
= 1.67 / T 0.67 ≤ T ≤ 4.0
(Sa/g)i and (Sa/g)c
Sa/g values also depend on damping

Multiplying factors for different damping are given in Table
3 of IS 1893(Part 1)
Recall from Lecture 2, higher damping reduces base shear

coefficient or design seismic forces
Multiplying factor =1.4, for 2% damping
Multiplying factor = 1.0 for 5% damping
Multiplying factor = 0.8 for 10% damping
This multiplier is not used for PGA
Damping
Damping for impulsive mode
5% of critical for RC tanks
2% of critical for steel tanks
These are kept in line with IS 1893(Part 1)
Clause 7.8.2.1 of IS 1893(Part 1) suggests 5% damping for
RC and 2% damping for steel buildings
However, IBC 2003 suggests 5% damping for all
tanks
It suggests 5% damping for all types of buildings also
Damping
Damping depends on material and level of

vibration
Higher damping for stronger shaking
Means that during the same earthquake,
damping will increase as the level of shaking
increases
We are performing a simple linear analysis, while
the real behavior is non-linear
Hence, one fixed value of damping is used in our
analysis
Damping
IS 1893(Part 1), needs to have a re-look at the

damping values
Accordingly, damping values for tanks can also
be modified
Damping
Damping for convective mode

0.5% of critical for all types of tanks
Convective mode damping does not depend
on material of tank or type of liquid stored
In Table 3 of IS 1893(Part 1):2002
Multiplying factor for 0.5% damping is not given
Values are given for 0% and 2% damping
Linear interpolation shall not be done
Multiplying factor = 1.75, for 0.5% damping
In Eurocode 8 this multiplying factor is 1.673
In ACI 350.3, this factor is 1.5
Importance factor, I for tanks is given in Table 1
of the Guideline
This Table is reproduced here
Type of liquid storage tank I

Tanks used for storing drinking water, non-volatile
material, low inflammable petrochemicals etc. and 1.5
intended for emergency services such as fire fighting
services. Tanks of post earthquake importance.
All other tanks with no risk to life and with negligible 1.0
consequences to environment, society and economy.
NOTE: Values of importance factor, I given in IS 1893 (Part 4)

may be used where appropriate
I = 1.5, is consistent with IS 1893(Part 1)

IS 1893(Part 1):2002 suggests, I = 1.5 for
Hospital buildings
Schools
Fire station buildings, etc.
Tanks are kept at same importance level
Footnote below this Table is given to avoid

conflict with I values of IS1893(Part 4)
IS 1893(Part 4) will deal with industrial structures
Not yet published
Some industries assign very high importance
factor to tanks storing hazardous materials
Depending on their own requirements
For such tanks, Importance factor (I) will be as
per part 4
R values for tanks are given in Table 2 of the
Guideline
This is reproduced in next two slides
Elevated tank R
Tank supported on masonry shafts
a) Masonry shaft reinforced with horizontal bands * 1.3
b) Masonry shaft reinforced with horizontal bands and vertical bars at corners 1.5
and jambs of openings
Tank supported on RC shaft
RC shaft with two curtains of reinforcement, each having horizontal and vertical 1.8
reinforcement
Tank supported on RC frame#
a) Frame not conforming to ductile detailing, i.e., ordinary moment resisting 1.8
frame (OMRF)
b) Frame conforming to ductile detailing, i.e., special moment resisting frame 2.5
(SMRF)
Tank supported on steel frame# 2.5

#These R values are meant for liquid retaining tanks on frame type staging which are inverted pendulum type
structures. These R values shall not be misunderstood for those given in other parts of IS 1893 for building
and industrial frames.
* These tanks are not allowed in Zone IV and V
Ground supported tank R
Masonry tank
a) Masonry wall reinforced with horizontal bands* 1.3
b) Masonry wall reinforced with horizontal bands and vertical bars at 1.5
corners and jambs of openings
RC / prestressed tank
a) Fixed or hinged/pinned base tank (Figures 6a, 6b, 6c) 2.0
b) Anchored flexible base tank (Figure 6d) 2.5
c) Unanchored contained or uncontained tank (Figures 6e, 6f) 1.5
Steel tank
a) Unanchored base 2.0
b) Anchored base 2.5
Underground RC and steel tank+ 4.0
+ For partially buried tanks, values of R can be interpolated between ground supported and underground
tanks based on depth of embedment.
R values for tanks are smaller than buildings

This is in line with other international codes
As discussed earlier, R depends on
Ductility
Redundancy
Overstrength
Tanks possess low ductility, redundancy and
overstrength
First let us consider, elevated tanks on frame

type staging
Staging frames are different than building
frames
Hence, following footnote to Table 2
These R values are meant for liquid retaining tanks on frame
type staging which are inverted pendulum type structures.
These R values shall not be misunderstood for those given
in other parts of IS 1893 for building and industrial frames.
Staging frames are non-building frames and are
different than building frames
There are critical differences between building

frames and non-building frames
International codes clearly differentiate between
these two types of frames
Building frames have rigid diaphragms at floor
levels
Frames of staging do not have rigid diaphragms
In buildings, seismic weight is distributed along
the height at each floor level
In elevated tanks, almost entire seismic weight is
concentrated at the top
These are inverted pendulum type structures
Moreover in buildings, non-structural elements,

such as infill walls, contribute significantly to
overstrength
Staging are bare frames
In view of this, for staging with SMRF, R = 2.5
as against R = 5.0 for buildings with SMRF
With R = 2.5, base shear coefficient for elevated
tanks on frame staging matches well with other
international codes
See next slide
Comparison for frame staging

Zone and soil parameters are same used in Lecture 2
0.5 IBC 2003;
Frame staging, R = 3.0
0.4
Guideline;
Frame staging, R = 2.5
0.3
IS 1893:1984;
0.2 All types of staging, K = 1.0
0.1
0
0 0.5 1 1.5 2 2.5 3
Time period (sec)
Let us now consider, elevated tanks on RC shaft

They possess less redundancy and have single
load path
RC shafts are usually thin shell and possess low
ductility
There are analytical and experimental studies on
ductility of hollow circular sections used in RC
shafts
Some references are given on next slide
Studies on ductility of shaft

Zanh F A, Park R, and Priestley, M J N, 1990, “Flexural
strength and ductility of circular hollow reinforced concrete
columns without reinforcement on inside face”, ACI Journal
87 (2), 156-166.
Rai D C, 2002, “Retrofitting of shaft type staging for elevated
tanks”, Earthquake Spectra, EERI, Vol. 18 No. 4, 745-
760.
Rai D C and Yennamsetti S, 2002, “Inelastic seismic demand
on circular shaft type staging for elevated tanks”, 7th
National Conf. on Earthquake Engrg, Boston, USA,
Paper No. 91.
Rao M L N, 2000, “Effect of confinement on ductility of RC
hollow circular columns”, a Master’s thesis submitted to
Dept. of Earthquake Engineering, Univ. of Roorkee, India.
These studies have revealed that ductility of

shaft depends on
Thickness of wall (ratio of outer to inner diameter)
Axial force on shaft
Longitudinal and transverse reinforcement
Some results from these studies on ductility of
RC shafts are discussed in next few slides
Effect of Axial Load on Ductility
Figure from Rai (2002)
Ast/Ag = ratio longitudinal

reinforcement to concrete
area.
P = axial load on shaft
fc’ = characteristic strength of

concrete
Ag = gross area of concrete
Hollow circular section
In this figure, curvature ductility is plotted as a

function of longitudinal reinforcement
These results are for inner (Di) to outer (Do)
diameter ratio of 0.94.
If ratio of axial load (P) to ultimate load (fck.Ag) is
0.1 then, curvature ductility is about 9 for Ast/Ag =
0.02
This value reduces to 3 for P/ (f’c.Ag) of 0.25
Now, let us see some results on effect of shaft
thickness
Effect of Shell Thickness on Ductility
Effect of ratio of inner to outer diameter (Di/Do) is shown

This result corresponds to P/(f’c.Ag) = 0.05
Very low axial force ratio
For thin shaft with Di/Do = 0.95, curvature

ductility is 12
For longitudinal steel ratio Ast/Ag = 0.02
This value increases to about 25 for thick shaft
with Di/Do = 0.8
Thus, thickness has significant effect on ductility
A thick shaft has reasonably good ductility
These analytical studies clearly indicate that

thin RC hollow sections possess very low
ductility
Issues connected with poor ductility of shaft,
inadequate provisions of IS 1893:1984, and
their correlation to behavior during recent
earthquakes is discussed in following paper:
Rai D C, 2002, “Review of code design forces for shaft
supported elevated water tanks”, Proc.of 13th Symposium
on Earthquake Engineering , Roorkee, Ed. D K Paul et al.,
pp 1407 -1418.
(http://www.nicee.org/ecourse/12_symp_tanks.pdf)
Based on all these considerations, R = 1.8 for

shaft supported tanks
With this value of R, base shear coefficient for
shaft supported tanks matches well with
international codes
Comparison with IBC 2003 on next slide
Comparison for shaft staging

Zone and soil parameters are same as used in Lecture 2
0.5 IBC 2003;

Shaft staging, R = 2.0
0.4
Guideline;
0.3 Shaft staging, R = 1.8
IS 1893:1984;
0.2 All types of staging, K = 1.0
0.1
0
0 0.5 1 1.5 2 2.5 3
Time period (Sec)
Some useful information on RC shaft is given in

ACI 371-98
ACI 371-98 , 1998, “ Guide for the analysis, design , and
construction of concrete-pedestal water Towers”, American
Concrete Institute, Farmington Hill, MI, USA.
It exclusively deals with tanks on RC shaft
It suggests same design forces as IBC 2003
It gives information on:
minimum steel
construction tolerances
safety against buckling
shear design etc.
We have seen comparison with IBC 2003

Comparison with other international codes is
available in following documents:
Jaiswal, O. R. Rai, D. C. and Jain, S.K., 2004a, “Codal
provisions on design seismic forces for liquid storage tanks:
a review”, Report No. IITK-GSDMA-EQ-01-V1.0, Indian
Institute of Technology Kanpur, Kanpur.
(www.iitk.ac.in/nicee/IITK-GSDMA/EQ01.pdf )
Jaiswal, O. R., Rai, D. C. and Jain, S.K., 2004b, “Codal
provisions on seismic analysis of liquid storage tanks: a
review” Report No. IITK-GSDMA-EQ-04-V1.0, Indian
Institute of Technology Kanpur, Kanpur.
(www.iitk.ac.in/nicee/IITK-GSDMA/EQ04.pdf )
In the above two documents, following

international codes are reviewed and
compared:
IBC 2000 (now, IBC 2003)
ACI 350.3
ACI 371
AWWA D-110 and AWWA D-115
AWWA D-100 and AWWA D-103
API 650 and API 620
Eurocode 8
NZSEE recommendations (From New Zealand)
Priestley et al. (1986)
Now we know Z, I, R and Sa/g for tanks

One can now obtain base shear coefficient for
impulsive and convective modes
An example follows.
Example
Example: An elevated water tank has RC frame
staging detailed for ductility as per IS: 13920 and is
located in seismic zone IV. Site of the tank has soft
soil. Impulsive and convective time periods are 1.2
sec and 4.0 sec, respectively. Obtain base shear
coefficient for impulsive and convective mode.
Solution:
Zone: IV
∴ Z = 0.24 From Table 2 of IS 1893 (PART I):2002,
I = 1.5 From Table 1 of the Guideline
R = 2.5 for RC frame with good ductility (SMRF)
From Table 2 of the Guideline
Example on (Ah)i and (Ah)c
Impulsive time period, Ti = 1.2 sec, and soil is soft,

Damping = 5% (RC Frame)
∴ (Sa/g)i = 1.67/Ti = 1.67/1.2 = 1.392
(Clause 4.5.3 of the Guideline)
Convective mode time period, Tc = 4.0 sec and soil is soft

Damping = 0.5% (Clause 4.4 of the Guideline)
Factor 1.75 is to be used for scaling up (Sa/g) for 0.5%

damping (Clause 4.5.4 of the Guideline)
∴ (Sa/g)c = (1.67/Tc) x 1.75 = 1.67/4.0 x 1.75 = 0.731
Example on (Ah)i and (Ah)c
Base shear coefficient for impulsive mode

(Ah)i= (Z/2) x (I/R) x (Sa/g)i
= 0.24/2 x 1.5/2.5 x 1.392
= 0.10
Base shear coefficient for convective mode
(Ah)c = (Z/2) x (I/R) x (Sa/g)c
= 0.24/2 x 1.5/2.5 x 0.731
= 0.053
R values for tanks are less than those for
buildings.The basis for this is
Analytical studies
Provisions of international codes, and
Observed behavior of tanks
For tanks, slight modifications are recommended
for design spectrum of IS 1893(Part1)
Damping for convective mode may be taken as
0.5% for all types of tanks
Lecture 5
January 31, 2006

In this Lecture
Impulsive and convective base shear
Critical direction of seismic loading
© Sudhir K. Jain, IIT Kanpur E-Course on Seismic Design of Tanks/ January 2006 Lecture 5/ Slide 2
Base shear
Previous lectures have covered
Procedure to find impulsive and convective
liquid masses
This was done through a mechanical analog model
Procedure to obtain base shear coefficients in
This requires time period, damping, zone factor, importance
factor and response reduction factor
Now, we proceed with seismic force or base
shear calculations
Base shear
Seismic force in impulsive mode (impulsive base

shear)
Vi = (Ah)i x impulsive weight
Seismic force in convective mode (convective
base shear)
Vc = (Ah)c x convective weight
(Ah)i = impulsive base shear coefficient
(Ah)c = convective base shear coefficient
These are described in earlier lectures
Base shear
Now, we evaluate impulsive and convective

weights
Or, impulsive and convective masses
Earlier we have obtained impulsive and
convective liquid mass
Now, we consider structural mass also
Base shear : Ground supported tanks
Impulsive liquid mass is rigidly attached to

container wall
Hence, wall, roof and impulsive liquid vibrate
together
In ground supported tanks, total impulsive mass
comprises of
Mass of impulsive liquid
Mass of wall
Mass of roof
Hence, base shear in impulsive mode
Vi = ( A h )i (mi + mw + mt ) g
mi = mass of impulsive liquid

mw = mass of container wall
mt = mass of container roof
g = acceleration due to gravity
This is base shear at the bottom of wall

Base shear at the bottom of base slab is :
Vi’ = Vi + (Ah)i x mb
mb is mass of base slab
Base shear at the bottom of base slab may be
required to check safety against sliding
Base shear in convective mode
Vc = ( A h )c m c g
mc = mass of convective liquid
Total base shear, V is obtained as:
V = V i 2 + V c2
Impulsive and convective base shear are

combined using Square Root of Sum of Square
(SRSS) rule
Except Eurocode 8, all international codes use
SRSS rule
Eurocode 8 uses absolute summation rule
i.e, V = Vi + Vc
In the latest NEHRP recommendations (FEMA

450), SRSS rule is suggested
Earlier version of NEHRP recommendations (FEMA
368) was using absolute summation rule
FEMA 450, 2003, “NEHRP recommended provisions for
seismic regulations for new buildings and other structures”,
Building Seismic Safety Council, National Institute of Building
Sciences,, USA.
FEMA 368, 2000, “NEHRP recommended provisions for
seismic regulations for new buildings and other structures”,
Building Seismic Safety Council, National Institute of Building
Sciences,, USA.
Bending moment:Ground supported tanks
Next, we evaluate bending or overturning

effects due to base shear
Impulsive base shear comprises of three parts
(Ah)i x mig
(Ah)i x mwg
(Ah)i x mtg
mw acts at CG of wall
mt acts at CG of roof
mi acts at height hi from bottom of wall
If base pressure effect is not included
mi acts at hi*
If base pressure effect is included
Recall hi and hi* from Lecture 1
Bending moment at the bottom of wall

Due to impulsive base shear
M i = ( A h )i (mi hi + mw hw + mt ht ) g
Due to convective base shear

M c = ( A h )c (mc hc ) g
hi = location of mi from bottom of wall

hc = location of mc from bottom of wall
hw = height of CG of wall
ht = height of CG of roof
For bending moment at the bottom of wall,

effect of base pressure is not included
Hence, hi and hc are used
Bending moment at the bottom of wall

(Ah)imt
(Ah)cmc
(Ah)imw
hc
ht
hi (Ah)imi hw
Ground level
Mi = ( Ah )i (mi hi + mw hw + mt ht ) g
M c = ( A h )c (mc hc ) g
Total bending moment at the bottom of wall
M = M i2 + M c2
SRSS rule used to combine impulsive and
convective responses
Overturning moment:Ground supported tanks
Overturning moment
This is at the bottom of base slab
Hence, must include effect of base pressure
hi* and hc* will be used
Ground level
Overturning moment:Ground supported tanks
Overturning moment in impulsive mode
⎡ mi ( hi + tb ) + mw (hw + tb ) +⎤
*
Mi = ( Ah )i ⎢
*
⎥g
⎢⎣mt (ht + tb ) + mb tb / 2 ⎥⎦
Overturning moment in convective mode
= ( Ah ) c mc ( hc + tb ) g
* *
M c
tb = thickness of base slab
Overturning moment is at the bottom of base

slab
Hence, lever arm is from bottom of base slab
Hence, base slab thickness, tb is added to
heights measured from top of the base slab
Total overturning moment
*2 *2
M = M *
i +M c
Example
Example: A ground-supported circular tank is shown
below along with some relevant data. Find base shear
and bending moment at the bottom of wall. Also find
base shear and overturning moment at the bottom of
base slab.
Roof slab 150 mm thick mi = 141.4 t; mc = 163.4 t

10 m mw = 65.3 t, mt =33.1 t,
4m Wall 200 mb = 55.2 t,

mm thick
hi =1.5 m, hi* = 3.95 m,
hc = 2.3 m, hc* = 3.63 m
Base slab 250 mm thick
(Ah)i = 0.225, (Ah)c = 0.08
Example
Solution:
Impulsive base shear at the bottom of wall is
Vi = (Ah)i (mi + mw + mt) g
= 0.225 x (141.4 + 65.3 + 33.1) x 9.81
= 529.3 kN
Convective base shear at the bottom of wall is

Vc = (Ah)c mc g
= 0.08 x 163.4 x 9.81
= 128.2 kN
Total base shear at the bottom of wall is

V = Vi2 + Vc2 = (529.3 ) + (128.2 )
2 2
= 544.6kN
Example
For obtaining bending moment, we need height of CG of roof slab

from bottom of wall, ht.
ht = 4.0 + 0.075 = 4.075 m
Impulsive bending moment at the bottom of wall is
Mi = (Ah)i (mihi + mwhw + mtht) g
= 0.225 x (141.4 x 1.5 + 65.3 x 2.0 + 33.1 x 4.075) x 9.81
= 1054 kN-m
Convective bending moment at the bottom of wall is

Mc = (Ah)c mc hc g
= 0.08 x 163.4 x 2.3 x 9.81
= 295 kN-m
Total bending moment at bottom of wall is
M = Mi2 + M2c = (1054 )2 + (295 )2 = 1095kN - m
Example
Now, we obtain base shear at the bottom of base slab
Impulsive base shear at the bottom of base slab is

Vi = (Ah)i (mi + mw + mt + mb) g
= 0.225 x (141.4 + 65.3 + 33.1 + 55.2) x 9.81
= 651.1 kN
Convective base shear at the bottom of base slab is

Vc = (Ah)c mc g
= 0.08 x 163.4 x 9.81
= 128.2 kN
Total base shear at the bottom of base slab is
V = Vi2 + Vc2 = (651.1 ) + (128.2 )

2 2
= 663.6kN
Example
Impulsive overturning moment at the bottom of base slab

Mi* = (Ah)i [mi (hi* + tb) + mw(hw + tb) + mt(ht +tb) + mb tb/2]g
= 0.225 x [141.4(3.95 + 0.25) + 65.3(2.0 + 0.25) +
33.1(4.075 + 0.25) + 55.2 x 0.25/2] x 9.81
= 1966 kN-m
Convective overturning moment at the bottom of base slab
Mc* = (Ah)c mc (hc* + tb) g
= 0.08 x 163.4 x (3.63 + 0.25) x 9.81
= 498 kN-m
Total overturning moment at bottom of base slab
*
M = (M ) + (M )
* 2
i
* 2
c = (1966 ) 2
+ (498 ) = 2028 kN - m
2
Notice that this value is substantially larger that the value at the
bottom of wall (85%)
Base shear : Elevated tanks
In elevated tanks, base shear at the bottom of
staging is of interest
Ms is structural mass
Base shear in impulsive mode
Vi = ( A h )i (mi + ms ) g

V c = (A h )c m c g
Total base shear
V = V i 2 + V c2
Bending moment:Elevated tanks
Bending moment at the bottom of staging

Bottom of staging refers to footing top
Impulsive base shear comprises of two parts
(Ah)i x mig
Ah)i x msg
Convective base shear has only one part
(Ah)c x mcg
mi acts at hi*
mc acts at hc*
Bending moment at bottom of staging is being
obtained
Hence, effect of base pressure included and hi*
and hc* are used
Structural mass, ms comprises of mass of empty

container and 1/3rd mass of staging
ms is assumed to act at CG of empty container
CG of empty container shall be obtained by
considering roof, wall, floor slab and floor beams
[ (
M i = ( Ah )i mi hi + hs + ms hcg
* *
) ]g
M c = ( Ah )c m c hc + hs g
*
( *
)
hs = staging height
Measured from top of footing to bottom of wall
hcg = distance of CG of empty container from
bottom of staging

(Ah)i mig (Ah)c mcg
hi* hc*
(Ah)i msg
hcg
hs hs
Top of footing
[ (
M = ( Ah )i mi h + hs + ms hcg g
i
*
i
*
) ] *
(
M c = ( Ah )c mc hc + hs g
*
)
Total bending moment
*2 *2
M = *
M i +M c
For shaft supported tanks, M* will be the design

moment for shaft
For analysis of frame staging, two approaches

are possible
Approach 1: Perform analysis in two steps
Step 1:
Analyze frame for (Ah)imig + (Ah)imsg
Obtain forces in columns and braces
Step 2:
Analyze the frame for (Ah)cmcg
Obtain forces in columns and braces
Use SRSS rule to combine the member forces
obtained in Step 1 and Step 2
Approach 2:
Apply horizontal force V at height h1 such that
V x h1 = M*
V and M* are obtained using SRSS rule as
described in slide nos. 26 and 32
In this approach, analysis is done in single step
Simpler and faster than Approach 1
Example
Example: An elevated tank on frame staging is shown

below along with some relevant data. Find base shear
and bending moment at the bottom of staging.
A is CG of empty container
A
2.8 m
mi = 100t; mc = 180 t
Mass of container = 160 t
Mass of staging = 120 t
hi* = 3 m, hc* = 4.2 m
hs = 15 m
(Ah)i = 0.08, (Ah)c = 0.04
GL
Example
Structural mass, ms = mass of container + 1/3rd mass of staging

= 160 + 1/3 x 120 = 200 t
Base shear in impulsive mode
Vi = ( A h )i (mi + ms ) g
= 0 .08 x (100 + 200 ) x 9 . 81
= 78.5 + 157 = 235.5 kN

Vc = (A h ) c m c g
= 0 . 04 x 180 x 9 . 81 = 70 . 6 kN
Example
Total base shear V = V i 2 + V c2
= 235 .5 2 + 70.6 2 = 245.8kN
Now, we proceed to obtain bending moment at the bottom staging
Distance of CG of empty container from bottom of staging,

hcg = 2.8 + 15 = 17.8 m
Example
Base moment in impulsive mode
M i
*
= ( A h )i [m i (h i *
)
+ h s + m s h cg ] g
= 0 . 08 [100 x (3 . 0 + 15 )+ 200 x 17 . 8 ]x 9 . 81
= 78.5 x 18 + 157 x 17.8
= 4207 kNm
Note: 78.5 kN of force will act at 18.0m and 157 kN of force will act at
17.8 m from top of footing.
Example
Base moment in convective mode
M c = ( Ah )c mc hc + hs g
*
( *
)
= 0 . 04 x180 x (4 . 2 + 15 )x 9 . 81
= 70.6 x 19.2
= 1356 kNm
Note: 70.6 kN of force will act at 19.2 m from top of footing.
Total base moment

* 2 *2
M *
= M i + M c
= 4207 2 + 1356 2
= 4420 kNm
Example
Now, for staging analysis, seismic forces are to be

applied at suitable heights
There are two approaches
Refer slide no 33
Approach 1:
Step 1: Apply force of 78.5 kN at 18 m and 157 kN at 17.8
m from top of footing and analyze the frame
Step 2: Apply 70.6 kN at 19.2 m from top of footing and
analyze the frame
Member forces (i.e., BM, SF etc. in columns and braces) of
Steps 1 and 2 shall be combined using SRSS
Example
Approach 2:
Total base shear, V = 245.8 kN will be applied at height h1,
such that
V x h1 = M*
245.8 x h1 = 4420
∴ h1 = 17.98 m
Thus, apply force of 245.8 kN at 17.98 m from top of footing
and get member forces (i.e., BM, SF in columns and
braces).
Elevated tanks:Empty condition
Elevated tanks shall be analysed for tank full as
well as tank empty conditions
Design shall be done for the critical condition
In empty condition, no convective liquid mass
Hence, tank will be modeled using single
degree of freedom system
Mass of empty container and 1/3rd staging mass
shall be considered
Elevated tanks:Empty condition
Lateral stiffness of staging, Ks will remain same

in full and empty conditions
In full condition, mass is more
In empty condition mass is less
Hence, time period of empty tank will be less
Recall, T = 2 Π MK
Hence, Sa/g will be more
Usually, tank full condition is critical
However, for tanks of low capacity, empty
condition may become critical
Direction of seismic force
Let us a consider a vertical cantilever with
rectangular cross section
Horizontal load P is applied
First in X-Direction
Then in Y-direction (see Figure below)
More deflection, when force in Y-direction
Hence, direction of lateral loading is important !!
Y
X P
On the other hand, if cantilever is circular

Direction is not of concern
Same deflection for any direction of loading
Hence, it is important to ascertain the most critical
direction of lateral seismic force
Direction of force, which will produce maximum
response is the most critical direction
In the rectangular cantilever problem, Y-direction is the most
critical direction for deflection
For frame stagings consisting of columns and braces, IS

11682:1985 suggests that horizontal seismic loads shall be
applied in the critical direction
IS 11682:1985, “Criteria for Design of RCC Staging for Overhead
Water Tanks”, Bureau of Indian Standards, New Delhi
Clause 7.1.1.2 Horizontal forces – Actual forces and moments resulting

from horizontal forces may be calculated for critical direction and used
in the design of the structures. Analysis may be done by any of the
accepted methods including considering as space frame.
Clause 7.2.2 Bending moments in horizontal braces due to horizontal

loads shall be calculated when horizontal forces act in a critical
direction. The moments in braces shall be the sum of moments in the
upper and lower columns at the joint resolved in the direction of
horizontal braces.
Section 4.8 of IITK-GSDMA Guidelines contains

provisions on critical direction of seismic force for
tanks
Ground-supported circular tanks need to be

analyzed for only one direction of seismic loads
These are axisymmetric
Hence, analysis in any one direction is sufficient
Ground-supported rectangular tanks shall be

analyzed for two directions
Parallel to length of the tank
Parallel to width of the tank
Stresses in a particular wall shall be obtained
for seismic loads perpendicular to that wall
RC circular shafts of elevated tanks are also

axisymmetric
Hence, analysis in one direction is sufficient
If circular shaft supports rectangular container

Then, analysis in two directions will be necessary
For elevated tanks on frame staging

Critical direction of seismic loading for columns
and braces shall be properly ascertained
Braces and columns may have different critical
directions of loading
For example, in a 4 - column staging
Seismic loading along the length of the brace is
critical for braces
Seismic loading in diagonal direction gives
maximum axial force in columns
See next slide
Critical directions for 4 - column staging
Bending Axis
Critical direction for Critical direction for

shear force in brace axial force in column
For 6 – column and 8 – column staging, critical

directions are given in Figure C-6 of the
Guideline
See next two slides
More information available in Sameer and Jain
(1994)
(http://www.nicee.org/ecourse/Tank_ASCE.pdf)
Bending Axis
Critical direction for shear Critical direction for shear force and
force and bending moment bending moment in braces and axial
in columns force in columns
Bending Axis
Critical direction for Critical direction for shear force,

shear force and bending bending moment and axial force
moment in braces in columns
As an alternative to analysis in the critical

directions, following two load combinations can
be used
100 % + 30% rule
Also used in IS 1893(Part 1) for buildings
SRSS rule
100%+30% rule implies following combinations

ELx + 0.3 ELY
ELY + 0.3 ELx
ELx is response quantity when seismic loads are
applied in X-direction
ELY is response quantity when seismic loads are
applied in Y-direction
100%+30% rule requires

Analyze tank with seismic force in X-direction;
obtain response quantity, ELX
Response quantity means BM in column, SF in brace, etc.
Analyze tank with seismic force in Y-direction;
obtain response quantity, ELY
Combine response quantity as per 100%+30%
rule
Combination is on response quantity and not
on seismic loads
Important to note that the earthquake

directions are reversible
Hence, in 100%+30% rule, there are total eight
load combinations
SRSS rule implies following combination
EL2x + EL2y
Note:
ELx is response quantity when seismic loads are
applied in X-direction
ELY is response quantity when seismic loads are
applied in Y-direction
Hence, analyze tank in two directions and use
SRSS combination of response quantity
This completes seismic force evaluation on tanks
There are two main steps
Evaluation of impulsive and convective masses
Evaluation of base shear coefficients for
SRSS rule is used to combine impulsive and
convective responses
Critical direction of seismic loading shall be
properly ascertained
Else, 100%+30% or SRSS rule be used

Questions and Comments on Lectures

Question from Mr. H. I. Abdul Gani (gani@igcar.ernet.in )
Lecture 4
1. Why damping in convective mode is independant of type of liquid stored? For

highly viscous liquids, does not viscosity of the liquid affect damping?
This is an interesting question. The formulation used for obtaining

hydrodynamic pressure, considers liquid to be non-viscous, homogeneous
and incompressible. The potential equation is written with these
assumptions on liquid properties. Now, if liquid is highly viscous, not only
the damping, but the convective mass will also change. One can guess, that
in highly viscous liquid, less mass will undergo sloshing motion, thereby
implying that more liquid will participate in impulsive mode. However, this
needs to be substantiated with in-depth study.
2. How the multiplicative factor of 1.75 was arrived at for 0.5% damping? Was
logarithmic interpolation used?
Yes, logarithmic interpolation is used. In the commentary to Clause 4.5.4,

we have referred Newmark and Hall (1982), wherein, this issue is
discussed.
3. For multistoreyed buildings with "soft storey" what is the importance factor to
be adopted?
This is a question on buildings. Importance factor is not related to soft

storey. If a building has soft storey, then it would be an irregular building.
Please refer Clause 7.10 of IS 1893(Part 1):2002
Question from Mr. N. Devanarayanan (devanarayanan@indianoil.co.in)
Dear Sir,
I have been attending the ecourse on the design of liquid storage
structures. Since we deal mostly with petroleum tank storages I am more
interested in learning about the same.There were a few points I would like
clarification on.
In certain areas due to the soil characteristics the storage tanks are some
times rested on RCC pile foundations
In certain cases the tanks are resten on sandpad foundations, the
underlying starta below the sand pad foundations are strngthened with stone
columns.
In such cases do we have to take the effect of seismic forces and design
the tanks using the two degree of freedom, and will the piles/stone columns
act as a staging thus making the whole structure behave like an elevated
tank.
In the first case, there will be a RC ring beam (or circular pile cap), which is
supported on piles. Tank wall and some portion of base plate rest on this
ring beam. Inside the ring beam there will be sand filling, which supports
rest of the base plate. This tank will be considered as ground-supported
tank and not as elevated tank. In the second case, the entire base plate
rests on sand fill, which is supported on compacted stone columns. For this
case also, tank will be treated as ground supported tank. The two degree
of freedom model of elevated tank is not applicable to these tanks. For
circular steel tanks, first find the impulsive time period using formula given
in Lecture 3 (also refer Clause 4.3.1.1 of the Guideline). Then, find the
convective mode time period and proceed with the evaluation of base
shear coefficient.
Lecture 5
1. Why Vi and Vc (also Moi&Moc) are combined using SRSS rule? Wouldn’t
absolute sum be more accurate as the effects of impulsive and convective mass
are simultaneously present?
Though impulsive and convective masses (or modes) may be present

simultaneously, these modal responses may not reach their maximum
values simultaneously. Hence, absolute summation is not used. Till
sometime back, NEHRP recommendation of 2000 (FEMA 368) had
suggested absolute summation. However, some recent numerical
simulation studies have indicated that absolute summation rule
overestimates the response. Hence, the latest NEHRP recommendation of
2003 (FEMA 450) has used SRSS rule.
Lecture 6
February 7, 2006
In this Lecture
Hydrodynamic pressure on wall and base
Effect of vertical ground acceleration
Sloshing wave height
Anchorage requirements for ground supported
tanks
So far, emphasis was on lateral forces on tanks
These lateral forces comprised of
Impulsive component
Convective component
Impulsive component has two parts
One due to impulsive liquid mass
Second due to structural mass of tank
Convective component has one part
Due to convective liquid mass
Recall from Lecture 1

Impulsive force is summation of impulsive
Convective force is summation convective
Now, we will see procedure to obtain
Impulsive and convective pressure distribution on
wall and base
Pressure distribution on wall is needed to obtain

Hoop forces, and
Bending moment in wall
IS 3370 (Part IV):1967 gives hoop forces and
bending moments due to hydrostatic pressure
Hydrostatic pressure has linear distribution along
wall height
IS 3370(Part IV):1967, “Code of Practice for Concrete
Structures for the storage of Liquids”, Bureau of Indian
Standards, New Delhi
Impulsive pressure:Circular tanks
Impulsive pressure on wall of circular tanks is
given by
p iw = Q iw ( y ) ( A h ) i ρ g h cos φ
⎡ ⎛ y⎞ ⎤
2
⎛ D⎞
Qiw ( y ) = 0 .866 ⎢1 − ⎜ ⎟ ⎥ tanh ⎜ 0 .866 ⎟
⎢⎣ ⎝ h ⎠ ⎥⎦ ⎝ h⎠
Qiw = Coefficient of impulsive pressure on wall
(Ah)i = Impulsive base shear coefficient
φ = Circumferential angle (see next slide)
y = Vertical distance of a point on wall from the bottom of wall
Note that impulsive pressure varies with

circumferential angle, φ
At φ = 00, pressure will be maximum
At φ = 900, pressure will be zero
Recall following figure from Lecture 1
zero pressure at φ = 900
φ
maximum pressure at φ = 00
Qiw(y) can also be read from Figure 9(a) of the

Guideline
This Figure is reproduced here
1
0.8
0.6
Y/h
0.4
1.0 0.5 0.25
1.5
h/D =2
or h/L
0.2
0
0 0.2 0.4 Q 0.6 0.8 1
iw
Impulsive pressure on base of circular tanks

changes in radial as well as circumferential
direction
A strip of length l’ is considered
Refer plan of circular tank shown in Figure 8(a) of the
Guideline l’
x
Direction of D/2
seismic force φ
Plan
Impulsive pressure on this strip is given by
⎛ x⎞
sinh ⎜ 0 . 866 ⎟
p ib = 0 . 866 ( Ah )i ρ g h ⎝ h⎠
⎛ l' ⎞
cosh ⎜⎜ 0 . 866 ⎟⎟
⎝ h ⎠
x = horizontal distance in the direction of seismic force, of a point

on base slab, from the reference axis at the center of tank
Example: A circular tank has internal diameter

of 12 m, and water height of 5 m. Impulsive
base shear coefficient is 0.23. Find the impulsive
pressure distribution on wall
Solution:
D = 12 m, h = 5 m, (Ah)i = 0.23
Impulsive pressure on wall is given by,
piw = Qiw (y) (Ah)i ρ g h cos φ
Where,
Qiw (y)= 0.866 [ 1- (y /h)2] tanh(0.866 D/h)
Maximum pressure will occur at φ = 0
At base of wall, y = 0;
Qiw = 0.866 tanh(0.866 D/h)
= 0.866 tanh(0.866 x 12/5)
= 0.84
Now, at φ = 0,
piw = Qiw (y) (Ah)i ρ g h
Mass density of water, ρ = 1000 kg/m3
∴ piw (y =0) = 0.84 x 0.23 x 1000 x 9.81 x 5
= 9476.5 N/m2 = 9.48 kN/m2
For different values of y, piw is similarly calculated. Some values are
tabulated, and a distribution plotted on the next slide.
y (m) 0 1.25 3.75 5

Qiw (y) 0.84 0.79 0.37 0
piw (y) (kN/m2) 9.48 8.91 4.17 0
h = 5m
At φ = o
9.48 kN/m2
Impulsive pressure distribution on wall
Impulsive pressure:Rectangular tanks
Impulsive pressure on walls of rectangular tanks
is given by
p iw = Q iw ( y ) ( A h ) i ρ g h
⎡ ⎛ ⎞ ⎤
y
2
⎛ L ⎞
Q iw ( y ) = 0 . 866 ⎢1 − ⎜ ⎟ ⎥ tanh ⎜ 0 . 866 ⎟
⎣⎢ ⎝ ⎠ ⎦⎥
h ⎝ h ⎠
Except following, this expression is same as that
for circular tanks
D/h is replaced by L/h
Angle φ is not present
Qiw(y) can be read from Figure 9(a) of the
Guideline
This pressure is on walls perpendicular to

direction of seismic force
Pressure remains same along the length of
these perpendicular walls
However, it changes with wall height
Walls perpendicular to
direction of seismic force
B L

Impulsive pressure on base of rectangular tanks

Impulsive pressure will be defined for a point at
a distance x from central axis
As shown in Figure below
Refer plan of rectangular tank in Figure 8 (b) of the
Guideline L
Direction of
Seismic Force
Plan
Impulsive pressure is given by

pib = Qib ( x ) ( Ah )i ρ g h
⎛ x ⎞
sinh ⎜ 0 . 866 ⎟
Q ( x ) = ⎝ h ⎠
⎛ ⎞
ib
L
cosh ⎜ 0 . 866 ⎟
⎝ h ⎠
Qib(x) is coefficient of impulsive pressure on base

It can be read from Figure 9(b) of the Guideline
See next slide
Coefficient of impulsive pressure on base Qib(x)

1.2
0.25
0.5
0.8 Qib(x)
1.0
1.5
0.4 2.0
Qib
0
-0 .4 -0 .2 0 0.2 0.4
x/L
2.0 -0.4
1.5
1.0
-0.8
0.5
0.25 = h/L
-1.2
Figure 9(b) of the Guideline
Convective pressure:Circular tanks
Convective pressure on wall of circular tanks is
given by
⎡ 1 2 ⎤
pcw = Qcw( y )( Ah ) c ρ g D⎢1 - cos φ ⎥ cosφ
⎣ 3 ⎦
⎛ y⎞
cosh ⎜ 3.674 ⎟
Qcw ( y ) = 0.5625 ⎝ D⎠
⎛ h⎞
cosh ⎜ 3.674 ⎟
⎝ D⎠
(Ah)c = Convective base shear coefficient

Qcw = Coefficient of convective pressure on wall
φ = Circumferential angle
y = Vertical distance of a point on wall from the bottom of wall
Qcw can be obtained from Figure 10(a) of the

Guideline
See next slide
Coefficient of convective pressure on wall, Qcw

1
0.8
2.0
Y/h 1.5
1.0
0.6
0.5
0.4
h/D=0.25
0.2
0
0 0.1 0.2 Q 0.3 0.4 0.5 0.6
cw
Figure 10(a) of the Guideline

Convective pressure on base of circular tanks is

given by
p cb = Q cb ( x ) ( Ah )c ρ g D
⎡ x 4 ⎛ x ⎞3 ⎤ ⎛ h⎞
Qcb ( x ) = 1.125⎢ − ⎜ ⎟ ⎥ sech⎜ 3.674 ⎟
⎢⎣ D 3 ⎝ D ⎠ ⎥⎦ ⎝ D⎠
Qcb can also be obtained from Figure 10(b) of

the Guideline
See next slide
Coefficient of convective pressure on base, Qcb

0.3
h/D=0.25
0.2
Qcb
0.5
0.1
b
c 0.75
Q
1.0
0
1.0 x/D
-0.4 -0.2 0 0.2 0.4
0.75
-0.1
0.5
-0.2
h/D=0.25
-0.3
Figure 10(b) of the Guideline

Example: A circular tank has internal diameter

of 12 m, and water height of 5 m. Convective
base shear coefficient is 0.07. Find the
convective pressure distribution on wall
Solution:
D = 12 m, h = 5 m, (Ah)c = 0.07
Convective pressure on wall is given by,
pcw = Qcw (y) (Ah)c ρ g D [1- 1/3 cos 2φ] cos φ
Where,
Qcw (y)= 0.5625 cosh(3.674 y/D)/cosh(3.674 h/D)
Maximum pressure will occur at φ = 0
At base of wall, y = 0;
Qcw (y =0) = 0.5625/cosh(3.674 h/D)
= 0.5625/cosh(3.674 x 5/12) = 0.23
Now, at φ = 0,
pcw = Qcw (y) (Ah)c ρ g D [1- 1/3 ]
Mass density of water, ρ = 1000 kg/m3
∴ pcw (y =0) = 0.23 x 0.07 x 1000 x 9.81 x 12 x [1- 1/3]
= 1263.5 N/m2 = 1.26 kN/m2
For different values of y, pcw is similarly calculated. Pressure
distribution with wall height is shown in the next slide.
y (m) 0 1.25 3.75 5

Qcw (y) 0.23 0.25 0.40 0.5625
pcw (y) (kN/m2) 1.26 1.37 2.20 3.09
3.09 kN/m2
h = 5m
At φ = o
1.26 kN/m2
Convective pressure distribution on wall
Convective pressure:Rectangular tanks
Convective pressure on wall of rectangular tank
is given by
p cw = Q cw ( y )( Ah )c ρ g L
⎛ y⎞
cosh ⎜ 3 .162 ⎟
Qcw ( y ) = 0 .4165 ⎝ L⎠
⎛ h⎞
cosh ⎜ 3 .162 ⎟
⎝ L⎠
Qcw can also be obtained from Figure 11(a) of
the Guideline
See next slide
Coefficient of convective pressure on wall, Qcw

1
0.8
Y/h
2.0
1.5
0.6
1.0
0.5
0.4
h/L=0.2
0.2
0
0 0.1 0.2 Q 0.3 0.4 0.5
cw
Figure 11(a) of the Guideline

Convective pressure on base of rectangular

tanks is given by
pcb = Qcb( x )( Ah )c ρ g L
⎡ x 4 ⎛ x ⎞3 ⎤ ⎛ h⎞
Qcb ( x ) = 1 .25 ⎢ − ⎜ ⎟ ⎥ sec h ⎜ 3 .162 ⎟
⎣⎢ L 3 ⎝ L ⎠ ⎦⎥ ⎝ L⎠
Qcb can be obtained from Figure 11(b) of the

Guideline
However, Figure 11 (b) of the Guideline is
incorrect
Correct Figure is shown in next slide
Coefficient of convective pressures on base, Qcb

0.4
0.3
h/L = 0.25
0.2
0.50
Qcb
0.1 0.75
1.0
0
1.0
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
0.75 -0.1 x/L
0.50 -0.2
-0.3
h/L = 0.25
-0.4
Figure 11(b) of the Guideline (corrected)

Linearised Pressure Distribution
Impulsive and convective pressures on wall
have curvilinear distribution
Hoop forces and bending moment in wall due to
such curvilinear pressure distributions are not
readily available
Hoop forces and BM are available in codes only
for linearly varying pressure
As mentioned earlier, refer IS 3370 (Part IV):1967
Hence, equivalent linearised pressure

distribution is suggested
For impulsive as well as convective pressure
This linearisation is such that, base shear and

bending moment at the bottom of wall remains
same
Equivalent linear distribution for impulsive

pressure will be such that
Total base shear and bending moment due to
linear distribution will be same as that due to
curvilinear distribution
Similarly, for convective pressure
Equivalent linear impulsive pressure
bi
qi qi
h ˜
hi hi
Actual Impulsive ai
pressure distribution Equivalent linear
pressure distribution
The ordinates of linear distribution are
a i = 2 (4 h − 6 hi ) b i = 2 (6 h i − 2 h )
qi qi
h h
Both for circular and rectangular tanks
Both for qi is different for circular and rectangular
tanks
( A h )i m i
For circular tanks qi = g
π D / 2
( A h )i m i
For rectangular tanks qi = g
2B
Here
mi = impulsive liquid mass
hi = height of impulsive mass
h = Total liquid height
(Ah)i = Impulsive base shear coefficient
All these quantities are known

Procedure to obtain these quantities is covered
in previous lectures
Similarly, equivalent linear convective pressure

b c = 2 (6 h c − 2 h )
bc qc
qc qc h
a c = 2 (4 h − 6 h c )
qc
hc
h
˜ hc h
For circular tanks

ac
Actual convective ( Ah )c mc
pressure distribution
Equivalent linear qc = g
pressure distribution π D/2
For rectangular tanks
( A h )c m c
qc = g
2B
Linearised distribution can be treated as

summation of two parts
Uniform pressure
Triangular pressure
bi
= +
ai bi ai - bi
Hoop forces and BM in wall due to uniform and

triangular pressure can be obtained from IS
3370(Part IV):1967
Linearised pressure distributions discussed here

are taken from ACI 350.3
ACI 350.3, 2001, “Seismic design of liquid containing
concrete structures”, American Concrete Institute,
Farmington Hill, MI, USA.
Example: A circular tank has internal diameter of 12 m,

and water height of 5 m. Impulsive and convective base
shear coefficients are 0.23 and 0.07, respectively. Find
linearised impulsive and convective pressure distribution
on wall.
Solution:
D = 12 m, h = 5 m, (Ah)i = 0.23, (Ah)c = 0.07
Capacity of tank = π x 122 /4 x 5 = 565.5 m3
∴ Mass of water = 565.5 x 1000 = 5,65,500 kg
For h/D = 5/12 = 0.42, from Figure 2a of the Guideline:

mi/m = 0.47, ∴ mi = 0.47 x 565500 = 2,65,800 kg
mc/m = 0.5, mc = 0.50 x 565500 = 2,82,800 kg

hi /h = 0.375, hi = 0.375 x 5 = 1.875 m
hc /h = 0.58, hc = 0.58 x 5 = 2.9 m
a) Impulsive mode
( Ah ) i mi 0 .23 × 265 ,800 × 9 .81

qi = g= = 31 .81kN/m
π D /2 π × 12 / 2
Pressure at bottom & top is given by
a i = 2 (4 h − 6 hi ) = (4 × 5 − 6 × 1 .875 ) = 11 .13 kN/m

qi 31 . 81 2
2
h 5
bi = 2 (6hi − 2h ) = 2 (6 × 1.875 − 2 × 5) = 1.59kN / m 2

qi 31.81
h 5
Note:- For the same tank, in earlier example, we had obtained actual
distribution of impulsive pressure. Linear and actual pressure distributions
are shown below:
bi = 1.59 kN/m2
h = 5m
ai = 11.13 kN/m2 9.48 kN/m2

Linear pressure distribution Actual pressure distribution
( Obtained in earlier example)
b) Convective mode
( Ah ) c m c 0 .07 × 282 ,800 × 9 .81

qc = g= = 10 .3kN/m
π D/2 π × 12 / 2
ac =
qc
2
(4 h − 6 h c ) =
10 . 3
2
(4 × 5 − 6 × 2 . 9 ) = 1 . 07 kN/m
h 5
b c = 2 (6 h c − 2 h ) = (6 × 2 . 9 − 2 × 5 ) = 3 . 05 kN/m
qc 10 . 3 2
2
h 5
Note:- For the same tank, in earlier example we had obtained actual
distribution of convective pressure. Linear and actual pressure
distributions are shown below:
3.09 kN/m2
bi = 3.05 kN/m2
h = 5m
ai = 1.07 kN/m2 1.26 kN/m2

Linear pressure distribution Actual convective pressure distribution
(Obtained in earlier example)
Circumferential distribution
In circular tanks, pressure varies along the
circumference of the wall
Recall there is cosφ term in the expression for
pressure on wall
Hoop forces and BM in wall for such varying

pressure are also not readily available
Hence, for convenience in stress analysis
Hydrodynamic pressure may be approximated
by an outward axisymmetric pressure distribution
Intensity of this axisymmetric distribution is equal
to that of maximum pressure
See next slide
Simplified distribution in circumferential direction
pmax pmax
Actual pressure distribution Simplified pressure distribution
Effect of vertical acceleration
Due to vertical acceleration, weight of liquid and
tank will increase or decrease depending on
direction of acceleration
If acceleration is downward, weight will decrease
If acceleration is upward, weight will increase
Increase in weight = vertical acceleration x mass
Increase in liquid weight would lead to increase

in hydrostatic pressure exerted by liquid on wall
Decrease in weight will reduce hydrostatic
pressure, and this need not be considered for
design purpose
Hydrostatic pressure varies linearly with height
Pressure due vertical acceleration would also
vary linearly with height
Pressure, pv due to vertical acceleration
pv = ( Av ) ρ g h (1 − y h)
2 ⎛Z I Sa ⎞
Av = ⎜⎜ × × ⎟⎟
3 ⎝ 2 R g ⎠
ρ = Mass density of liquid pv

h = Total liquid height
g = Acceleration due to gravity
y
Z = Zone factor
I = Importance factor
R = Response reduction factor
Design acceleration spectrum in vertical direction

is taken as 2/3rd of design acceleration spectrum
in horizontal direction
This is consistent with IS 1893 (Part 1):2002
For finding the value of Sa/g
One needs to know time period of tank in vertical
mode
Usually tanks are quite rigid in vertical direction

Hence, time period in vertical mode may be
taken as as 0.3 sec for all tanks
This implies, value of Sa/g will be 2.5
Vertical mode of vibration for circular tanks is
called breathing mode
More information on time period of breathing
mode is available in Eurocode 8 and ACI 350.3
Pressure due to wall inertia
Mass of wall is distributed along its length and
height
Hence, during lateral excitation, wall is
subjected to lateral pressure
Pressure due to wall inertia is given by
p ww = ( Ah )i t ρ m g
ρm = mass density of tank wall

pww
t = thickness of tank wall
g = acceleration due to gravity
(Ah)i = impulsive base shear coefficient
Pressure due to wall inertia
Pressure due to wall inertia may not be

significant in steel tanks
Steel tank walls have less mass
However for RC tanks, pressure due to wall
inertia may be significant
Total pressure on wall
Total pressure on wall comprises of
Impulsive hydrodynamic pressure, piw
Convective hydrodynamic pressure, pcw
Pressure due to vertical acceleration, pv
Pressure due to wall inertia, pww
Total pressure on wall is obtained as
p = ( piw + pww ) + p + pv 2 2 2
cw
Example
A ground supported circular RC water tank has internal
diameter of 12 m and liquid height is 5 m. It is located in
zone IV and has fixed base. Wall thickness is 200 mm.
Find maximum pressure due to vertical excitation and
wall inertia.
Solution:
h = 5 m, D = 12 m, (Ah)i = 0.23,
Wall thickness = 200 mm
Zone IV; Z= 0.24, I = 1.5,
It is ground supported RC tank with fixed base, R = 2.0
Example
For vertical mode, time period is taken as, T = 0.3 sec for all tanks
∴ Sa/g = 2.5
Av = 2/3 . Z/2. I/R . Sa/g
= 2/3 x 0.24/2 x 1.5/2.0 x 2.5
= 0.15
Maximum pressure due to vertical acceleration will occur at

y = 0, i.e. at the base of wall,
∴ pv = (Av) ρ g h (1- y/h)
= 0.15 x 1000 x 9.81 x 5 x (1- 0/5)
= 7.36 kN/m2
Example
Maximum pressure due to wall inertia,

pww = (Ah)i t ρm g
= 0.23 x 0.2 x 25
= 1.15 kN/m2
Example
For this tank, from the previous examples, at the bottom of wall (y =
0) we have
Piw = 9.48 kN/m2; Pcw = 1.26 kN/m2;
Pv = 7.36 kN/m2; Pww = 1.15 kN/m2;
Hence, total pressure
p = ( piw + pww)2 + pcw + pv

2 2
= ( 9.48 + 1.15 ) + 1.26 + 7.36
2 2 2
= 13 kN/m2
Total hydrostatic pressure at base = ρ g h = 49 kN/m2.
Thus, total hydrodynamic pressure of 13 kN/m2 is 26% of hydrostatic
pressure.
Convective liquid undergoes sloshing motion
i.e., liquid undergoes vertical motion
dmax
Maximum sloshing wave height is given by
D
d max = ( A h )c R For circular tanks
2
L
d max = ( A h )c R For rectangular tanks
2
(Ah)c = convective base shear coefficient

R = response reduction factor
D = diameter of circular tank
L = Length of rectangular tank in the direction of seismic force
Multiplication with R is due to the fact that

design forces are reduced by factor R
Recall, (Ah)c = (Z/2) (I/R) (Sa/g)c
Displacements will be R times higher than
displacements due to design forces
Expression for sloshing wave height is taken
from ACI 350.3
Sloshing wave height is useful in determining

the free board to be provided
If loss of liquid needs to be prevented
If free board is less than sloshing wave height,
liquid will exert pressure on roof
In the process, sloshing liquid mass will also
change
Refer Malhotra (2005) for more information
Malhotra P K (2005), “Sloshing loads in liquid-storage tanks
with insufficient free board”, Earthquake Spectra, Vol. 21,
No. 4, 1185-1192
Anchorage requirement
Ground supported tanks have tendency to
overturn during lateral base excitation
Particularly tall steel tanks
Anchorages are needed to ensure safety against
overturning
Tank will overturn, if overturning moment due to
lateral force is more than stabilizing moment
As a quick estimate, one can say, overturning

will occur if
h 1
>
D (A h )i
h = height of liquid
D = diameter of tank
(Ah)i= impulsive base shear coefficient
For rectangular tanks, diameter D is to be

replaced by L
It is an approximate estimate for circular tanks

It assumes that entire tank mass and liquid mass is
acting at the center of liquid height
Read Clause 4.12 of the Guideline
Impulsive and convective pressure have
curvilinear distribution along the wall height
Equivalent linear distribution may be used
For obtaining hoop forces and BM in wall
Wall is also subjected to pressure due to vertical
acceleration and wall inertia
Sloshing wave height can be obtained using
simple expression
Lecture 7
February 16, 2006

In this Lecture
Examples 2 & 3 of IITK-GSDMA Guideline
These examples have been sent separately
Here, we will discuss only important points
These tanks will also be analysed using IS
1893:1984
Comparison of forces obtained from the
Guideline and IS 1893:1984 will be presented
Example 2
Elevated tank considered in Example 2 of the
Guidelines has
RC Intze container of 250 m3 capacity
RC Frame staging on six columns
Staging is assumed to have been designed for good ductility
Tank is located in seismic zone IV on hard soil
Please refer the examples sent to you
In this example, base shear and base moment
at the base of staging are obtained
Example 2
Analysis involves following calculations

Step 1: Weight of various parts
Step 2: Center of gravity of empty container
Step 3: Parameters of spring-mass model
From h/D of equivalent circular container, all the
parameters of spring-mass model are obtained
Step 4: Lateral stiffness of staging
Using standard structural analysis software,
stiffness is obtained
More about this later
Example 2
Step 5: Impulsive convective mode time period

Step 6: Design horizontal seismic coefficient
(Ah)i and (Ah)c are obtained using Z, I, R, and
damping values
Step 7: Base shear and base moment
Step 8: Analysis of empty tank
Example 2
Mass of water = 255.6 t

Mass of empty container = 160.7 t
Mass of staging = 105.6 t
Structural mass, ms = 160.7 + 105.6/3 = 196 t
ms = empty container + 1/3rd mass of staging
Total mass = 255.6 + 196 = 451.6 t
Example 2
Parameters of spring-mass model

Here, container is of Intze type
Equivalent circular container of same volume
and diameter is obtained
Equivalent circular container has D = 8.6 m and
h = 4.4 m; hence, h/D = 0.51
This h/D is used to find mi, mc, hi, hc etc.
mi = 140.6 t; about 55% of total water mass
mc =110 t; about 43% of total water mass
mi + mc is about 2% less than total water mass.
This was discussed in Lecture 3.
Example 2
hi = 1.65 m, hi* = 3.43 m

hc = 2.68 m, hc* = 3.43 m
Incidentally, hi* and hc* are same in this case.
hi* is twice that of hi
However, hc* is only 1.3 times hc
Indicates that impulsive pressure at the base is
more significant than the convective pressure at
the base. Hence,
hi* is significantly larger than hi
hc* is not much larger than hc
Example 2: Lateral stiffness
Lateral stiffness of staging
Staging is modeled using structural analysis
software
Force is to be applied at the CG of tank
CG of tank is combined CG of container and
impulsive liquid mass, mi.
In this example, CG of empty container is
treated as CG of tank
A reasonable approximation for design purpose
CG of empty container is at a distance of 2.88 m

from top of circular ring beam
In the model, center-line dimensions are used
Hence, force is applied at 2.88 + 0.3 = 3.18 m
from center-line of circular ring beam
0.3 m is half-depth of ring beam
Since only staging is modeled, a rigid link of
length 3.18 m is provided to apply force at that
height
Model is shown in next slide
10.0
Rigid link
Structural model of staging used for analysis
A force of 10 kN is applied at CG
Deflection of CG = 5.616 x10-4 m
Hence, lateral Stiffness of Staging
Ks = 10/(5.616x10-4) = 17,800 kN/m
Here, computer software is used for analysis
Traditionally, designers use manual methods
Present frame is a polygon frame
For finding stiffness of such frames, general

practice in India has been to use method
suggested by SP 22
SP:22 – 1982, Explanatory Handbook on Codes for
Earthquake Engineering, Bureau of Indian Standards, New
Delhi
In this method, braces are considered as rigid
beams
If braces are considered as rigid beams, then

No rotation is allowed at joints of the columns
Hence, stiffness of a column in a panel or
between two brace levels is
K C = 12EI
L3
Where L is length of column in a panel
Example 2: Analysis as per IS 1893:1984
For the frame in the present example

Stiffness of each column in a panel, Kc
K C =12EI =12 ×22360 ×10 33×8.76 ×10 9 =36,730 kN/m

L3 4.0 ⎛
⎜
⎞
⎟
⎜ ⎟
⎜ ⎟
⎝ ⎠
In each panel, there are 6 columns

Hence, stiffness of each panel
∑K C = 6 × 36730 = 220,400 kN/m
Stiffness of one panel = 220400 kN/m

There are four such panels
Each having same stiffness
They act as springs in series
Hence, lateral stiffness of staging, Ks
Ks =
∑ K C
=
220400
= 55,090 kN/m
4 4
Thus, stiffness of frame

Ks = 17,800 kN/m from computer analysis
Ks = 55,090 kN/m from method of SP22
Thus, SP 22 grossly overestimates the stiffness
Clearly, braces should not be treated as rigid
beams
Their flexibility should be included in the analysis
Jain and Sameer (1992) have developed a

simple method for obtaining stiffness of
polygon frames
This method considers flexibility of braces
Sameer, S. U., and Jain, S. K., 1992, “Approximate methods for
determination of time period of water tank staging”, The Indian
Concrete Journal, Vol. 66, No. 12, 691-698.
(http://www.nicee.org/ecourse/Tank_ICJ.pdf)
As per Jain and Sameer (1992)
1 1 1
= +
K s K flexure K axial
Np Top most
∑K
1 1
= panel
K flexure i =1 panel
Kpanel = stiffness of each panel

Np = number of panels Bottom most
panel
12 EI c N c ⎡ Ib L ⎤
K = ⎢ ⎥ For intermediate panels
+
panel
h3 ⎣ b
I L 2 I c h ⎦
12 EI c N c ⎡ I cbr L ⎤ For top most panel

K = ⎢ ⎥
+
panel
h3 ⎣ cbr
I L I c h ⎦
12 EI c N c ⎡ Ib L ⎤ For bottom most panel

K panel = ⎢ ⎥
h3 ⎣ b
I L + I c h ⎦
Np
1 2
K axial
=
N c Ac E R 2
∑
i =1
2
Hi h
E = Modulus of elasticity
Ic = Moment of inertia of column
Nc = Number of columns
h = Height of each panel
L = Length of each brace
Ib = Moment of inertia of brace beam
Icbr = Moment of inertia of circular ring beam
Hi = Distance of lateral load from inflection point of ith panel
Hi = Distance of lateral load from inflection point

of ith panel
For intermediate panels, point of inflection is at mid
height of panel
For the top most panel, point of inflection is at a
distance y from circular ring beam
⎛ 3 I b I c ⎞
⎜ + ⎟
y = ⎝ L h ⎠
h
⎛ 6 I b ⎞
⎜ ⎟
⎝ L ⎠
For the bottom most panel, distance y is measured

from fixed end
For 6–column staging of this Example

E = 22,360 N/mm2,
Ac = π x (650)2/ 4 = 3,31,800 mm2,
Nc = 6,
L = 3,141 mm
R = 3,141 mm
Ic = π x (650)4/ 64 = 8.76 x 109 mm4
Ib = 300 x (600)3 / 12 = 5.4 x 109 mm4
Icbr = 500 x (600)3 / 12 = 9 x 109 mm4
h = 4,000 mm
H1 = 6030 mm; H2 = 9,180 mm; H3 = 13,180 mm; H4 = 16,330 mm
Note:- Load is applied at CG of container, and distances Hi are from

the load.
Using this method, we get

Stiffness of staging, Ks = 16,440 kN/m
From computer analysis, we have
Ks = 17,800 kN/m
Difference is about 7%. Which is very reasonable
Recall, as per SP 22, Ks = 55,090 kN/m
Which is on much higher side
Note, usually in practice, method of SP 22 is

used
Also this method has been used in a number of
text books in India.
For further calculations here, Ks = 17,800 kN/m
is used
Example 2: Time period
Tank full condition
Impulsive mode time period, Ti
mi + ms
Ti = 2π
Ks
= 0.86 Sec
Convective time period, Tc
Tc = Cc D/g
For h/D = 0.51, from Figure 5 of the Guideline,
Cc = 3.35
Hence, Tc = 3.14 sec
Convective time period is 3.65 times impulsive

time period
Hence, two DoF model can be treated as two
uncoupled SDoF systems
Example 2: Base shear
Design horizontal seismic coefficient
Zone IV, hence, Z = 0.24
I = 1.5
Frame has ductile detailing, hence, R = 2.5
For Ti = 0.86 sec and hard soil; (Sa/g)i = 1.16
From Clause 4.5.3 of the Guideline
For Tc = 3.14; (Sa/g)c = 0.318 x 1.75 = 0.56
Multiplication by 1.75 is for 0.5% damping
Design Horizontal Coefficient

(Ah)i = Z/2 . I/R . (Sa/g)i = 0.084
(Ah)c = Z/2 . I/R . (Sa/g)c = 0.040
(Ah)c is less than (Ah)i
Convective mass is subjected to much smaller
acceleration
Example 2:Base shear
Base Shear, V
Impulsive mode
Vi = (Ah)i (mi + ms)g
Vi = 0.084 x (140.6 + 196) x 9.81
= 116 +161 = 277 kN
Convective mode
Vc = (Ah)c mc g = 0.04 x (110) x 9.81 = 43 kN
Convective forces are less than impulsive forces
Convective mass derives much smaller forces
than impulsive mass
2 2
Total Base shear, V = Vi +Vc = 280 kN
Base shear is about 6.3 % of total seismic
weight of 4429 kN (tank is located in seismic
zone IV)
Now, to obtain member forces, impulsive and
convective base shear is to be applied at
corresponding locations
Recall from Lecture 5, there are two
approaches to apply these forces
Approach 1:
Apply impulsive forces at their respective
locations
Impulsive forces have two parts
Part1: 116 kN at hi* + hs
Part 2: 161 kN at hcg
Apply convective forces at its location
43 kN at hc* + hs
Location of forces as per Approach 1
116 kN 43 kN
161 kN hi*= 3.43 m hc*=3.43 m
hcg = 19.18 m hs = 16.3 m hs = 16.3 m
Top of footing
Impulsive forces Convective forces
Approach 2:
Apply base shear, V at height h1 such that
V x h1 = M*
V = 280 kN and M* = 5448 kN-m
Hence, h1 = 5448/280 = 19.46 m
Thus, a force of 280 kN to be applied at 19.46 m
Location of forces as per Approach 2
280 kN
h1 =19.46
Top of footing
Analysis for tank empty Condition

No convective mode
Impulsive mass will be only structural mass, ms
V = (Ah)i ms g = 211 kN
This will act at CG of empty container
Tank empty condition

Moment = 211 x 19.18 = 4053 kNm
211 kN
hcg =19.18
Top of footing
Base shear and moment are more in tank full

condition
Hence, design will be governed by tank full
condition
These forces are to be applied in suitable
directions to get critical design forces in columns
and braces
Recall, the issue of critical direction for frame
staging
Analysis as per IS 1893:1984
Now, this tank will be analysed as per IS
1893:1984
This will help us compare design forces as per
the Guideline and IS 1893:1984
Let us recall two points
In IS 1893:1984 entire liquid mass is lumped as
impulsive mass
Hence, no convective mode
While finding the stiffness of staging, braces are
treated as infinitely rigid beams
This was general practice as suggested by solved example
in SP22.
Even though not suggested by IS:1893-1984.
Zone and soil parameters from IS 1893:1984

β = 1.0
I = 1.5
Fo= 0.25 (Zone IV)
Time Period, T
M
T = 2π
K
M is total Mass
This is sum of structural mass and total liquid mass
K is Stiffness of Staging
Total Mass, M = Structural mass + liquid mass

= 196 +255.6 = 451.6 t
Lateral stiffness Ks
Braces are treated as as rigid beams as per
prevailing practice
Hence, Ks = 55,090 kN/m
This we have seen earlier
Time period
M 451.6
T = 2π = 2π = 0.57 Sec
K 55090
This time period is less than the impulsive time

period of 0.86 sec calculated earlier
Because, stiffness of braces is taken infinite
For 5% damping & T = 0.57 Sec
Sa/g = 0.155
From Figure 2 of IS 1893 : 1984
Example 2: Analysis as per 1893:1984
Design horizontal seismic coefficient, αh

αh = βIFo(Sa/g) = 1.0 x 1.5 x 0.25 x 0.155
= 0.058
Base Shear, V = αh x W
= 0.058 x (451.6 x 9.81)
= 257 kN
Base shear is 5.8 % of seismic weight of 4429 kN
This will act at the CG of container
CG of container is at 19.18 m from footing top
Hence base moment, M = 257 x 19.18 = 4,929 kNm
Analysis for empty condition

Structural mass, M =196 t
M 196
T = 2π = 2π = 0.375 Sec
K 55090
For 5% damping & T = 0.375 Sec
Sa/g = 0.2
As per Figure 2 of IS 1893:1984
αh = βIFo(Sa/g) = 1.0 x 1.5 x 0.25 x 0.2 = 0.075
= 0.075 x (196 x 9.81) = 144 kN
Moment at the bottom of staging, M
= 144 x 19.18 = 2762 kNm
Next, we compare seismic forces obtained from
the Guideline and IS 1893:1984
Example 2:Comparison of forces
Comparison of results from the Guideline and IS 1893:1984
Guideline IS 1893:1984
Lateral stiffness of staging 17,800 kN/m 55,089 kN/m
Time period
Impulsive mode,
Tank empty ( Ti ) 0.66 sec 0.375 sec
Tank full ( Ti) 0.86 sec 0.57 sec
Convective mode,
Tank full ( Tc) 3.14 sec -----
Design seismic
horizontal coefficient
Impulsive mode
Tank empty ( Ah)i 0.11 0.075
Tank full ( Ah)i 0.084 0.058
Convective mode,
Tank full ( Ah)c 0.040 -----

Guideline IS 1893:1984
Total base shear (V)
Tank empty 211 kN 144 kN
Tank full 280 kN 257 kN
Bending Moment
Tank empty 4053 kN 2762 kNm
Tank full 5448 kN 4929 kNm
In tank full condition, base shear and moment

from the Guideline are about 10% more than
those obtained using IS 1893:1984
Two points may be noted

IS 1893:1984 implied K = 1.0, which leads to
smaller seismic forces
However, the practice was to overestimate
stiffness,which increases seismic coefficient.
These two effects compensate each other
Hence, there is not much difference in design
forces from the Guideline and IS 1893:1984 for
elevated tanks on frame type staging.
Example 2: Member forces
To obtain design forces in columns and braces
Staging frame is to be analysed for lateral seismic force
One can use standard structural analysis programs.
Alternatively, one may use approximate analysis methods.
Plane frame analysis methods are well known
For example, moment distribution method and
Kani’s method
In this example, staging has polygon frame
It is not a plane frame
Strictly speaking, it is a space frame
Some text books describe methods for analysis

of such polygon frames
Jain and Jai Krishna (1980) and Dayaratnam
(1986)
Jain O P and Krishna J, 1980, “Plain and reinforced
Concrete”, Vol. 2, Eighth Edition, Nem Chand Brothers,
Roorkee
Dayaratnam P, 1986, “Design of Reinforced Concrete
Structures”, Oxford & IBH Publishing Co., New Delhi
These methods make simplifying assumptions
Accuracy of these methods depends on validity
of these assumptions
Major assumptions in these methods are:

1) Axial force in column due to lateral force is
proportional to its distance from bending axis
2)Point of inflection in columns and braces are at
mid-span
3)Dayaratnam’s book assumes that lateral shear is
equally distributed to all columns, OR
Jain and Jai Krishna distribute lateral shear similar
to that in a cross-section of a cantilever column
Sameer and Jain (1994) have critically

examined these assumptions
(http://www.nicee.org/ecourse/Tank_ASCE.pdf)
They pointed out that assumption regarding
shear distribution is not necessary
That is, assumption 3) in the previous slide.
Column shear is obtained by moment
equilibrium in the plane of bending
Point of inflection for the top and bottom panels

is not at mid-span
Sameer and Jain obtained point of inflection for
these panels more accurately.
Based on these observations, Jain and Sameer
(1994) have developed a new approach to
obtain design forces in columns and braces
For the present example, design forces for

columns and braces are obtained
Total lateral force = 280 kN acting at 19.46 m
from footing top (i.e., fixed end)
Design forces in columns and braces from
following methods are compared in next slide
Computer analysis
Jain and Sameer
Jain and Jai Krishna
Dayaratnam
Comparison of design forces in columns and braces
Computer Jain & Jain & Dayaratnam

analysis Sameer JaiKrishna
Columns
Axial force (kN) 492 494 519 519
Bending moment (kNm) 150 160 187 94
Shear force (kN) 66 70 93 47
Braces
Bending moment (kNm) 175 187 280 187
Shear force (kN) 112 119 178 119
Jain & Jai Krishna’s approach overestimates

bending moments and shear forces in columns
and braces
Dayaratnam’s approach underestimates bending
moments and shear forces in columns
The point here is that designer should properly
assess the limitations of a particular method
Example 3:Elevated Tank on Shaft
Example 3 of the Guideline

Elevated tank on RC shaft is considered
Container is same as in Example 2
Thickness of shaft = 0.15 m
Centre line diameter of shaft = 6.28 m
Height of shaft above GL = 15 m
Hard soil; Zone IV
Example 3:Elevated tank on Shaft
Weight of staging, i.e., shaft

= π x 6.28 x 0.15 x 16.4 x 25 = 1,213 kN
Mass of staging = 123.7 t
Hence, structural mass ms
ms = 160.7 + 123.7/3 = 201.9 t
Container is same as in Example 2, hence,
mi = 140.6 t
mc =110 t
Example 3:Lateral stiffness
Lateral stiffness of staging

Shaft is considered as a cantilever of length
16.4 m
This is the height of shaft from top of footing
upto bottom of circular ring beam
Lateral stiffness is given by
3EI
KS = 3
L
Where,
E = Modulus of elasticity = 22.36 x 106 kN/m2
I = Moment of inertia of shaft cross section

= π x (6.434 – 6.134)/64 = 14.59 m4
L = Height of shaft = 16.4 m
Thus,
3 × 22.36 ×10 6 ×14.59
KS = = 2.22 ×10 5
kN/m
16.43
Height of shaft is taken up to ring beam only

and not up to CG of empty container
Recall, in frame staging, stiffness was obtained
by applying force at the CG of empty container
In frame staging, a rigid link was used from
staging top to apply force at CG
The effect of larger rigidity of container portion is thus
included
Even if load is applied at top of staging rather
than CG, there is not much difference in stiffness
Rigid link portion does not have much higher deflection than
deflection at staging top
In shaft, if we wish to consider height up to CG,

then, cantilever will have two different cross-
sections
One up to staging top
Second from staging top to CG
This will represent container portion up to CG
Due to larger dimensions in the second portion,
It will undergo only rigid body rotation
Deflection at CG will not be much higher than
that at staging top
Hence, staging height taken up to ring beam is

adequate for design
This will be on the conservative side
Also recall time period varies as square root of
stiffness
A 15% error in stiffness will give about 7% error in
time period
There is another approximation in stiffness of
shaft
Shear deformations
Effect of shear deformation

Consider a cantilever subjected to tip load , P
Tip deflection due to shear deformation
PL
δs =
κ ' AG
G = Shear modulus
A = Cross-sectional area
κ’ = Shape factor
Depends on shape of cross-section
Thus, total deflection of cantilever

PL3 PL
δ = + '
3EI κ AG
First part is due to flexural deformation

Second part is due to shear deformation
Hence, stiffness will be

1
Ks =
L3 L
+ '
3EI κ AG
For hollow circular cross-section

κ’ = 0.5
Hence, lateral stiffness of shaft staging, including
shear deformations, will be
1
Ks =
L3 L
+
3EI 0.5 AG
In the present example, if shear deformations

are included, then
Ks = 1.77 x 105 kN/m
Recall, without shear deformation, Ks = 2.22 x 105

kN/m
By including shear deformations, stiffness has
reduced by 20%.
For further calculations, Ks=2.22 x 105 kN/m is
used
This value is also used in the Guideline
However, it should have been 1.77 x 105 kN/m
Analysis for tank full condition
Impulsive mode time period, Ti
mi + ms
Ti = 2π
Ks
= 0.25 Sec
Shafts are quite rigid, hence, low time period

Recall, for frame staging, Ti = 0.86 sec
Convective time period, Tc = 3.14 sec

Same as in Example 2
Ratio of Tc and Ti is about 12.0
Impulsive and convective time periods are quite well separated
Hence, the assumption of two uncoupled SDoF
can be used
Design horizontal seismic coefficient
Zone IV, hence, Z = 0.24
I = 1.5
R = 1.8
For Ti = 0.25 sec and hard soil; (Sa/g)i = 2.5
From Clause 4.5.3 of the Guideline
Note:
If we include the effect of shear deformation,
then time period Ti would have been 0.28 sec
And, (Sa/g)i will remain 2.5
For Tc = 3.14; (Sa/g)c = 0.318 x 1.75 = 0.56

Multiplication by 1.75 is for 0.5% damping
Design Horizontal Coefficient, Ah

(Ah)i = Z/2 . I/R . (Sa/g)i = 0.25
(Ah)c = Z/2 . I/R . (Sa/g)c = 0.06
(Ah)c is much less than (Ah)i
Convective mass is subjected to much smaller
acceleration
Base Shear, V
Impulsive mode
Vi = (Ah)i (mi + ms)g
Vi = 0.25 x (140.6 + 201.8) x 9.81
= 345 + 495 = 840 kN
Convective mode
Vc = (Ah)c mc g = 0.06 x (110) x 9.81 = 65 kN
Convective forces are much smaller than
impulsive forces
2 2
Total Base shear, V = Vi +Vc = 843 kN
Base shear is about 19 % of total seismic
weight of 4,488 kN (tank is located in seismic
zone IV)
Impulsive forces have two parts

Part1: 345 kN at hi* + hs
Part 2: 495 kN at hcg
Apply convective forces at its location
65 kN at hc* + hs
Base moment in impulsive mode

Mi* = 345 x (3.43 + 17) + 495 x 19.88
= 169,00 kNm
Base moment in convective mode
Mc* = 65 x (3.43 + 17) = 1322 kNm
Total base moment
* * 2 * 2
M = M i +M c
= 16,940 kNm
Tank Empty Condition

Time Period Ti = 0.19 Sec
Base Shear, V = 495 kN
Base Moment, M* = 9,842 kN-m
Seismic base shear and moment are higher in
tank full condition
Hence, tank full condition will govern the design
Example 3:Analysis using IS 1893:1984
This tank is also analysed using IS 1893:1984

Entire liquid mass is lumped as impulsive mass
No convective mode
For tank full condition
Mass, M = Mass of container + Mass of water
+ 1/3rd Mass of staging
= 160.7 + 255.6 + 1/3 x 123.7 = 457.5 t
Example 3: Analysis using IS 1893:1984
Stiffness of staging, Ks = 2.22 x 105 kNm

Time period
M 457.5
T = 2π = 2π = 0.285 Sec
K 2.22 ×10 5

αh = βIFo(Sa/g) = 1.0 x 1.5 x 0.25 x 0.2
= 0.075
= 0.075 x 4488
= 336.6 kN
Base Moment, M = V x hcg
= 336.6 x 19.88
= 6,692 kN-m
Analysis for Tank Empty

Structural mass, M
M = Mass of container + 1/3rd Mass of staging
= 160.7 + 1/3 x 123.7 = 201.9 T
M 201.9
T = 2π = 2π = 0.19 Sec
K 2.22 ×10 5

αh = βIFo(Sa/g) = 1.0 x 1.5 x 0.25 x 0.2 = 0.075
= 0.075 x 1980 = 148.5 kN
Base Moment, M = V x hcg
= 148.5 x 19.88 = 2952 kN-m
Let us compare results from the Guideline and
IS 1893:1984
See next slide
Example 3: Comparison of results
Idealization of Tank Guideline IS 1893:1984

Lateral stiffness of staging 2.22 x 105 kN/m 2.22 x 105 kN/m
Time period
Impulsive mode,
Tank empty ( Ti ) 0.19 sec 0.19 sec
Tank full ( Ti) 0.25 sec 0.285 sec
Convective mode,
Tank full ( Tc) 3.14 sec -----
Design seismic horizontal coefficient
Impulsive mode
Tank empty ( Ah)i 0.25 0.075
Tank full ( Ah)i 0.25 0.075
Convective mode,
Tank full ( Ah)c 0.060 -----
Example 3:Comparison of results
Idealization of Tank Guideline IS 1893:1984

Total base shear (V)
Tank empty 495 kN 148.5 kN
Tank full 843 kN 336.6 kN
Total base Moment (M)
Tank empty 9,842 kN-m 2,952 kN-m
Tank full 16,940 kN-m 6,692 kN-m
For shaft supported tanks, seismic forces from

the Guideline are on much higher side
Base moment is 150% higher in full condition
Example 3:Comparison of results
Comparison of forces from the Guideline and IS

1893:1984 reveals:
For frame staging
Design forces increase marginally
However, this increase will vary for different tanks
For shaft staging
Design forces increase significantly
In these tanks, increase in base moment is of
particular interest
Analysis using IS 1893:1984
Main reason for this increase is:

IS 1893:1984 has K = 1.0 for all types of tanks
Thereby implying that all elevated tanks have same ductility
or energy absorbing capacity as that of a building with good
ductility
IITK-GSDMA Guideline has specified different
values of response reduction factors for different
types of staging
Depending on their ductility, redundancy and overstrength
In frame staging, earlier practice was to

overestimate lateral stiffness
Thereby increases design seismic forces
To some extent, this compensates the lower values
of design forces due to K = 1.0
Hence, in frame staging net increase in forces is not
very significant
For shaft staging, increase in design seismic

forces will not necessarily cause proportionate
increase in cost of the tank
Since lateral forces are higher, one will have to
suitably choose the dimensions of shaft
A wider shaft will help in achieving more economical
design
To explain this point, cost of shaft and foundation is
obtained for one tank
See next example
Issue of cost
Consider a 1000 m3 elevated tank on RC shaft
with following data
Height of shaft = 20 m (from foundation top)
Shaft thickness = 200 mm
Zone V, Medium soil, SBC = 25 t/m2
Solid raft foundation is used
Issue of cost
In the first instance, shaft of 7 m diameter is

considered
Then, shaft of 12 m diameter is considered
Comparison of quantity of concrete and steel
required in both the cases is given in next slide
Quantity for shaft and foundation are given
Issue of cost
Material quantities required

7 m dia. Shaft 12 m dia. shaft
IS 1893:1984 Guideline IS 1893:1984 Guideline
Shaft
Concrete 90 m3 * 150 m3 150 m3
Steel 6t * 7t 9t
Foundation
Concrete 200 m3 540 m3 135 m3 340 m3
Steel 8t 21 t 9t 12 t
* Design of 7 m shaft is not possible for forces as per the Guideline
Issue of cost
Thus, by increasing shaft diameter from 7 m to

12 m
About 200 m3 of concrete and 9 t of steel in
foundation could be saved
This example is shown to emphasize the

importance of proportioning of dimensions in
achieving economy in design
Issue of cost
Some more economy in foundation can perhaps

be achieved by using different type of
foundation system
Some options are:
Solid mat foundation
Annular mat foundation
Strip foundation on piles
Shell type foundation
Issue of cost
Note, design of container does not get much

affected, since hydrodynamic pressure is about
30 to 40% of hydrostatic pressure
In working stress design, permissible stresses are
increased by 33.3%
This would usually accommodate the
hydrodynamic pressure
Issue of cost
It is to be recognized that, so far, in India,

seismic design forces for tanks were rather low
Hence, simplistic design approaches prevailed
And above discussed options were not needed
and were not attempted
Moreover, in many elevated tanks, wind forces
used to govern the design
Issue of cost
In view of new design forces

There is a need to try some of these options
To achieve better economy, and
Good seismic performances
In Part B of the Guideline, six solved examples
are given
These will be of help to the users of the Guideline
We are now almost at the end of this course
Next Lecture, which is the last one, will discuss
some miscellaneous issues.
Lecture 8
February 21, 2006

In this Lecture
Following issues will be discussed
Soil structure interaction
Buried tanks
P-Delta effect
Flexibility of piping
Buckling of shell
IS 11682:1985
Soil Structure Interaction
Soil condition is important in seismic analysis
Type of soil affects the ground motion and
structure’s response
Effect on ground motion is reflected in design
spectrum
Recall, there are different response spectra for
hard, medium and soft soil
Spectrum for soft soil is higher than hard soil
Except in short period range
Effect of soil on structure’s response is included

in soil structure interaction
Soil-structure interaction is a complex problem
Soil properties are frequency dependent
Based on research, simplified methods have
been proposed
We will briefly discuss method suggested by

Veletsos (1984)
This has been used in Eurocode 8 and NZSEE
Recommendations (Priestley, 1986)
Veletsos, A. S., 1984, “Seismic response and design of liquid
storage tanks”, Guidelines for the seismic design of oil and
gas pipeline systems, Technical Council on Lifeline
Earthquake Engineering, ASCE, N.Y., 255-370, 443-461.
From Veletsos (1984):

Soil structure interaction has two effects
It changes time period
It changes total damping
Effect on time period
If structure is on hard soil, it is treated as fixed at
base
If soil is soft, it imparts flexibility and increases time
period
Soil flexibility affects impulsive mode time period
It does not affect convective mode time period
Effect on damping
Soil medium has higher damping
Damping in soil comes from hysteric action and
radial waves
Also called radial damping in soil
Damping of soil increases total damping of soil-
structure system
Veletsos (1984) suggested following expression

for impulsive time period of tank on soft soil
− K ⎡ K x h 2f ⎤
Ti = Ti 1 + ⎢1 +
f
⎥
Kx ⎢⎣ Kθ ⎥⎦
Kx = Horizontal translational stiffness of soil

Kθ = Rocking stiffness of soil
Kf = Stiffness of impulsive mode
hf = Distance from the tank base to the point of application of
impulsive mass, mi
Ti = Impulsive mode time period of tank with fixed base
8
KX = Gs R0
2 −νs
8
Kθ =
3
G s R0
3( 1 − ν s )
Gs = Shear modulus of elasticity of soil

νs = Poisson’s ratio of soil
R0 = Radius of the foundation
For elevated tanks

Stiffness of impulsive mode, Kf is lateral stiffness of
staging, Ks
For ground supported tanks
We have considered circular and rectangular
tanks
For both types of tanks, time period of impulsive
mode, Ti is known
Recall expression for Ti for circular tanks from Lecture 3
Also refer Clause 4.3.1.1 and 4.3.1.2 of the Guideline
Hence, stiffness of impulsive mode, Kf for

ground supported tanks can be obtained from
mi
Ti = 2π
Kf
mi being impulsive mass is also known
hf is distance of impulsive mass from base of

tank
In ground supported tanks
hf = hi + tb
In elevated tanks
Total impulsive mass = mi + ms
Hence, hf = hcg
Recall, we assume that entire mass is lumped at
CG of empty container
Veletsos (1984) included effect of soil damping

as follows:
Effective damping ratio of structure is
−
ξ
ξ = ξS + 3
⎛ − ⎞
⎜ Ti / T i ⎟
⎝ ⎠
−
ξ = Effective damping ratio of structure including soil effect
ξs = Damping ratio for soil
ξ = Damping ratio of structure
Ti = Impulsive mode time period with fixed base
−
Ti = Impulsive mode time period including soil effect
It is clear that soft soil would have lower values

of lateral stiffness Kx and rocking stiffness, Kθ.
Hence, more effect on impulsive time period
Similarly, damping in soft soil is more compared
to hard soil
Hence, damping of structure will be more
Thus, soil structure interaction would be
important for structures on soft soil
Increase in time period and damping would

reduce seismic forces
This is also called beneficial effects of soil
Before using these beneficial effects of soil, one
needs to properly ascertain the soil properties
Note, soil properties are frequency dependent
Buried tanks
In buried tanks, effect of soil pressure on wall
shall be included
During lateral excitation, soil induces dynamic
pressure on wall
Strictly speaking, a detailed soil-structure
interaction study is needed to find earthquake
induced soil pressure
Specialised literature shall be referred to find
dynamic earth pressure and its distribution
Buried tanks
NZSEE has suggested following simplified

distribution for earthquake induced soil pressure
Tank wall
σx = 1.5 (Ah)i γs He Ground surface
PE
Embedment
depth, He
Idealised pressure
distribution
Tank base
σx = 0.5 (Ah)i γs He
Buried tanks
Total force per unit length due to this pressure is

given by
PE = (Ah)i γsHe2
This is in addition to soil pressure in at-rest condition

This earth pressure shall be used to check wall
design
This dynamic earth pressure shall not be relied
upon to reduce dynamic effects due to liquid
Buried tanks
Seismic analysis procedure for buried tanks is

same as that for ground supported tanks
i.e., procedure to find mi, mc etc. remains same
Time period of impulsive mode will be very less,
hence, (Sa/g)i = 2.5
Buried tanks are quite rigid
Convective time period will be same as that for
ground supported tanks
Buried tanks
For buried tanks, response reduction factor, R =

4.0
This is for fully buried tank
Refer Table 2 of the Guideline
If tank is partially buried, then, R value shall be
linearly interpolated
Depending on depth of embedment
Buried tanks
For a RC tank with fixed base

R = 2.0 for ground supported tank
R = 4.0 for fully buried tank
If tank is half buried, then,
R = 2.0 + (4.0 – 2.0) x 0.5 = 3.0
For 2/3rd buried tank,
R = 2.0+(4.0 – 2.0)x 0.67 = 3.33
For a steel tank with anchored base
R = 2.5 for ground supported tank
If tank is half buried, then,
R = 2.5 + (4.0 – 2.5) x 0.5 = 3.25
Buried tanks
Dynamic soil pressure would also act on shaft

staging of elevated tanks
For large tanks, shaft diameter could be more than 10
m
Foundation for shaft is usually 2 to 3 m below ground
Hence, soil pressure will act over buried portion of shaft
Buried tanks
Dynamic soil pressure acts opposite to direction

of seismic loading
Seismic force
Soil
pressure
Buried tanks
Soil pressure would induce moment on the

foundation
This moment would be opposite to moment due
to seismic forces
Thus, net moment on foundation would reduce
Codes do not allow this reduction for design
purpose
Due to uncertainties in compactness of soil etc.
Buried tanks
In frame staging, column sizes are quite small

Compared to shaft diameter
Members with such small width may not be able
to mobilize dynamic soil pressure
Hence, no dynamic soil pressure on embedded length
of columns
P-Delta effect
Consider a vertical cantilever subjected to lateral
force Q and vertical force P
Bending moment at the base, M = Q x L
P-Delta effect
Due to force Q, let deflection be Δ

Point of application of P shifts
And, additional moment p x Δ will act on the
cantilever
P
Now, total moment at the base =
Q Δ QxL+PxΔ
P-Delta effect
Thus, vertical loads interact with lateral

displacements
And, induce additional forces on structure
This is termed as P-Delta effect
Due to additional moment of P x Δ, lateral
deflection would further increase
Which, in turn would cause more moment
P-Delta effect
P-Delta effect can be minimised by limiting the

lateral deflection
For buildings, IS 1893(Part 1) requires that storey
drift shall be less than 0.4%
i.e., total lateral deflection will be always less than
h/250.
h is total height of the building
P-Delta effect
Compared to buildings, P-Delta effect will be

more severe in elevated tanks
In buildings, vertical loads are distributed over
the entire height
In elevated tanks, almost entire vertical load is at
the top
P-Delta effect
Thus, to minimise P-Delta effect in elevated tanks

Permissible maximum lateral deflection shall be
less than that for buildings
h/500 may be a reasonable limit
Recall, in building limit is h/250
i.e., total lateral deflection shall be less than 0.2%
of height
In past earthquakes, many tanks have suffered
damages at the junctions of tank and piping
These junctions are critical locations
Severe stress concentration occurs at junctions
Hence, piping shall have sufficient flexibility to
accommodate seismic movement
Without causing damage to tank shell or base
All the international codes emphasize on flexible

piping system
Flexibility in piping can be imparted by
Using flexible elbows
Special coupling devices
These flexible devices need not be used at the
junction of piping and tanks
They can be used at nearby joints also
Buckling of shell
Ground supported circular steel tanks are
sensitive to buckling
These are thin shell structures
In the past, many steel tanks have suffered
buckling failure during earthquakes
API 650, AWWA D-100 are exclusively for ground
supported steel tanks
These codes give provisions for safety against
buckling
Buckling of shell
During lateral base excitation, tank wall is

subjected to excessive compressive force
This may trigger buckling
Buckling loads are very sensitive to initial
imperfections in geometry
Particularly, for thin shells
NZSEE recommendations (Priestley et al., 1984)
have also given information on buckling of steel
tanks
Buckling of shell
For safety against buckling

Codes restrict maximum allowable axial stresses in
steel tanks
RC shaft staging may also be thin shell structure
They are not as thin as steel tanks
Nevertheless, safety against buckling should be
ensured
Buckling of shell
Not much significant information is available on

buckling of RC shaft staging
ACI 371 (1998) has specified a limit on maximum
axial force on shaft
This limit depends on slenderness coefficient or
ratio of thickness to diameter
ACI 371 , (1998) “ Guide for the analysis, design , and
construction of concrete-pedestal water Towers”, American
Concrete Institute, Farmington Hill, MI, USA.
Buckling of shell
As per ACI 371, in limit state design, for shafts

under pure compression
P < 0.7 x Cw x fck x βwxAc
P = Axial load on shaft

Cw = Wall strength coefficient = 0.55
fck = Specified compressive strength of concrete
Ac = Gross area of concrete
βw = Wall slenderness coefficient
Buckling of shell
Wall slenderness coefficient, βw is obtained from

classical elastic buckling strength of cylinder and
is given by
βw = 80 t/dw
where, t = Wall thickness and
dw= Center-line diameter of shaft
βw shall be less than 1.0
For wider shafts with less thickness, βw will be

less than 1.0
Which will reduce the permissible axial load
Buckling of shell
Geometrical imperfections are invariably present

in shaft staging
Stringent construction tolerances should be
followed to minimize these imperfections
Recall, buckling load is sensitive to imperfections
Effect of these imperfections on static response,
dynamic response and buckling should be studied
A good research topic !!
Buckling of shell
ACI 371 specifies limits on construction

tolerances; Some of these are
Variation in wall thickness:
-3% and + 5% of wall thickness
Variation in plumb of shaft:
Less than 10 mm in any 1.6 m of height
Less than 40 mm in any 16 m of height
Maximum of 75 mm in total height
Variation in diameter of shaft:
0.4% of diameter and shall not exceed 75 mm
IS 11682:1985
In India, we have IS 11682:1985 exclusively for
RC staging
However, it does not have any provisions on
construction tolerances for shaft staging
These should be included in IS 11682
In fact, there are many other limitations in IS
11682:1985
Some of these will be discussed briefly
IS 11682:1985
Clause 3.2 of IS 11682: “…. Weight of water may be

considered as live load for members directly containing the
same….”
This implies that for container design, water can be
treated as live load
In seismic analysis of buildings, one applies
reductions in live load.
This confuses the design
Hence, this clause should be corrected
IS 11682:1985
Clause 5.4: Increase in permissible stresses for column staging

shall be as per IS 456-1978
Clause 5.4.1: “The increase in permissible stresses need not be
allowed in the design of braces for forces as wind or
earthquake, which are primary forces in them
Increase in permissible stresses is allowed for
columns and not for braces
i.e., braces will be stronger than columns
Whereas, in seismic design we need strong
column-weak beam combination
IS 11682:1985
Clause 7.3.3. “Loss of contact in the soil under footing should

not be allowed. In locations where the soil bearing capacity is
high, loss of contact may be allowed provided it is safe against
overturning and such other conditions that are to be fulfilled.”
This clause is for column footings
It needs to be more specific regarding how much
loss of contact can be allowed or how much
tension in soil may be permitted
Seismic loadings are momentary and like increase
in SBC, some tension may also be permitted
These should also be permitted for raft foundations.
IS 11682:1985
Clause 7.2.4. “For staging in seismic zones where seismic

coefficient exceeds 0.05 twin diagonal vertical bracing of steel
or RCC in addition to horizontal bracing may be provided….”
This clause is for frame staging
Limit of 0.05 on seismic coefficient may be
removed
Suggestions on use of shear walls in addition to
vertical diagonal bracing may be included
IS 11682:1985
Clause 8.2 provides provisions on minimum thickness of RC

shaft, reinforcement and detailing near openings.
These provisions should include suggestions on
proportioning of shaft and container diameters
Similarly, different types of foundation systems for
shaft should be suggested
This has been done in ACI 371
IS 11682:1985
Clause 8.5.2 suggests formulae for stresses in shaft

These are based on working stress design method
As per IS 456:2000, limit state design is compulsory
for members subjected to combined direct load
and flexural
Clause B-4.3 of IS 456:2000
Hence, provisions on limit state design of shaft
are needed in IS 11682
IS 11682:1985
There are many more such modifications needed

in IS 11682
An urgent revision of this code is needed.
This would help designers in analysis and design of
staging
And finally…..
With this Lecture, we conclude this E-course
Please spare some time to go through the
Lectures again
We have received some very good questions
Discussion and answers to questions on Lectures
1, 2 and 3 have already been sent to you
Please send your remaining questions
Thank You………
Assignment 1
Date of assignment: 18 January 2006
E-course on Seismic Design of Tanks
1
What is expected?
Part I: Reading assignments are to be
completed as soon as possible
Best if you finish the reading assignments on the
same day
It will help you appreciate the next lecture
© Sudhir K. Jain, IIT Kanpur E-Course on Seismic Design of Tanks / January 2006 Assignment 1 / slide 2
What is expected?
Parts II and III

To be completed by 19 January; before we send
out the solutions of these to you.
Please do not send the solutions to us.
Compare your answers with the solutions we will
send.
Feel free to go back and review the lecture
while answering the questions.
Some questions may not have very specific
answers
This is in line with our engineering profession wherein
certain questions have no specific answers.
Part I: Reading Assignment 1
Read carefully Preface and Clauses 0., 1., 2.,
4.1, 4.2, & 4.2.1 of the Guidelines
Mark whatever you find difficult to follow
Also mark portions that you find interesting
It will help you review the materials later.
Part II: True / False
Identify the following statements as True or False
1.1)Hydrodynamic pressure varies linearly with liquid height.
1.2)Impulsive mass is less than convective mass for short
tanks.
1.3)Net hydrodynamic force on the container wall is zero.
1.4)Hydrodynamic pressure on base causes overturning
moment on tank.
1.5)Convective mass acts at lower height than impulsive
mass.
1.6)With the inclusion of base pressure effect, overturning
moment on tank reduces.
1.7)Impulsive and convective masses depend only on height
of liquid.
Part III: Questions
1.1)A circular tank has internal diameter of 12 m
and water height of 5 m. Obtain mi, mc, Kc, hi,
hc, hi* and hc*.
1.2) A rectangular tank has internal plan
dimensions of 10 m x 16 m. Water height is 8
m. Obtain mi, mc, hi, hc, hi* and hc* in both the
directions.
Solution 1

Date of solution: 20 January 2006
1
1.1)Hydrodynamic pressure varies linearly with liquid height.
False
Hydrodynamic pressure has curvilinear variation along the
height. In fact, Hydrostatic pressure varies linearly with
height.
1.2)Impulsive mass is less than convective mass for short

tanks.
False
In short tanks, convective liquid or liquid undergoing
sloshing motion is more. In tall tanks, impulsive liquid mass
is more.
© Sudhir K. Jain, IIT Kanpur E-Course on Seismic Design of Tanks / January 2006 Solution 1 / slide 2
1.3)Net hydrodynamic force on the container wall is zero.

False
Rather, net hydrostatic force on wall is zero
1.4)Hydrodynamic pressure on base causes overturning

moment on tank.
True
Refer slide no. 46 of Lecture 1
1.5)Convective mass acts at lower height than impulsive
mass.
False
Liquid in upper portion, undergoes convective or sloshing
motion, hence, convective mass is always located at
higher height than impulsive mass. Note the graphs for hi
and hc in figure 2b and 3b of guidelines
1.6)With the inclusion of base pressure effect, overturning

moment on tank reduces.
False
Hydrodynamic pressure on the base produces overturning
moment in the same direction as that due to
hydrodynamic pressure on wall. Hence, with the inclusion
of base pressure effect, overturning moment increases.
This can be seen from the fact that hi* is always greater
than hi and hc* is always greater than hc.
1.7)Impulsive and convective masses depend only on height

of liquid.
False
Impulsive and convective masses depend on aspect ratio
(h/D or h/L) rather than only on height, h.
Part III: Solutions
1.1)A circular tank has internal diameter of 12 m and water
height of 5 m. Obtain mi, mc, Kc, hi, hc, hi* and hc*.
Solution:
Total volume of water = π/4 x 122 x 5 = 565.5 m3
∴ Total water mass, m = 565.5 x 1.0 = 565.5 t
D = 12 m, h = 5 m
∴ h/D = 5/12 = 0.417
For this value of h/D, from Figures 2a and 2b of the Guideline,

values of various parameters can be read. One can also use
formulae given in Table C-1 of the Guideline. Here, these formulae
will be used to obtain various parameters.
Part III: Solutions
⎛ D⎞ ⎛ 12 ⎞
tanh ⎜ 0 . 866 ⎟ tanh ⎜ 0 . 866 ⎟
mi ⎝ h⎠ = ⎝ 5 ⎠
= 0.466
= 12
m D 0 . 866
0 . 866
h 5
⎛ 5⎞
⎛ h⎞ tanh⎜ 3.68 ⎟
tanh⎜ 3.68 ⎟ ⎝ 12 ⎠
mc
= 0.23 ⎝ D⎠ = 0.23 = 0.503
m h 5
D 12
Since, h/D is less than 0.75, hi/h = 0.375
⎛ h⎞ ⎛ 5⎞
cosh⎜ 3.68 ⎟ − 1.0 cosh⎜ 3.68 ⎟ − 1.0
hc ⎝ D⎠ ⎝ 12 ⎠
= 1−
⎛ h⎞
= 1− = 0.579
h h
3.68 sinh⎜ 3.68 ⎟
5 ⎛ 5⎞
3.68 sinh⎜ 3.68 ⎟
D ⎝ D⎠ 12 ⎝ 12 ⎠
Part III: Solutions
D 12
0.866 0.866
hi * 5
= h - 0.125 = - 0.125 = 0.947
h ⎛ D⎞ ⎛ 12 ⎞
2 tanh⎜ 0.866 ⎟ 2 tanh⎜ 0.866 ⎟
⎝ h⎠ ⎝ 5⎠
⎛ h⎞ ⎛ 5⎞
cosh⎜ 3.68 ⎟ − 2.01 cosh⎜ 3.68 ⎟ − 2.01
hc * ⎝ D⎠ ⎝ 12 ⎠
= 1− = 1− = 0.878
h h ⎛ h⎞ 5 ⎛ 5⎞
3.68 sinh⎜ 3.68 ⎟ 3.68 sinh⎜ 3.68 ⎟
D ⎝ D⎠ 12 ⎝ 12 ⎠
mg ⎛ h⎞
K c = 0.836 tanh 2 ⎜ 3.68 ⎟
h ⎝ D⎠
∴ ⎛ 5⎞
K c h / mg = 0.836 tanh 2 ⎜ 3.68 ⎟ = 0.694
⎝ 12 ⎠
Part III: Solutions
Thus, we get,
mi/m = 0.466 ∴mi = 0.466 x 565.5 = 263.5 t
mc/m = 0.503 ∴mc = 0.503 x 565.5 = 284.5 t
hi/h = 0.375 ∴hi = 0.375 x 5 = 1.88 m
hc/h = 0.579 ∴hc = 0.579 x 5 = 2.90 m
hi*/h = 0.947 ∴hi* = 0.947 x 5 = 4.735 m
hc*/h = 0.878 ∴hc* = 0.878 x 5 = 4.39 m
Kch/mg = 0.694
Kc = 0.694mg/h = 0.694 x 565.5 x 9.81/5.0 = 770 kN/m
Part III: Solutions
1.2) A rectangular tank has internal plan dimension of

10 m x 16 m. Water height is 8 m. Obtain mi, mc, hi, hc,
hi* and hc* in both the directions.
Solution:
Y
16 m
10 m X
Plan view
Total volume of water = 10 x16 x 8 = 1280 m3

Total mass of water, m = 1280 x 1.0 = 1280 t
Part III: Solutions
First, consider base motion in X-direction:

L = 16 m, h = 8 m
∴ h/L = 8/16 = 0.5
From Figure 3a, 3b of guidelines, for h/L = 0.5:

mi/m = 0.54 ∴mi = 0.54 x 1280 = 691.2 t
mc/m = 0.48 ∴mi = 0.48 x 1280 = 614.4 t
hi/h = 0.375 ∴hi = 0.375 x 8 = 3.0 m
hc/h = 0.57 ∴hc = 0.57 x 8 = 4.6 m
hi*/h = 0.80 ∴hi* = 0.8 x 8 = 6.4 m
hc*/h = 0.87 ∴hc* = 0.87 x 8 = 7.0 m
Part III: Solutions
Now, consider base motion in Y-direction:

L = 10 m, h = 8 m
∴ h/L = 8/10 = 0.8
From Figure 3a, 3b of guidelines, for h/L = 0.8:

mi/m = 0.72 ∴mi = 0.72 x 1280 = 921.6 t
mc/m = 0.32 ∴mi = 0.32 x 1280 = 409.6 t
hi/h = 0.40 ∴hi = 0.40 x 8 = 3.2 m
hc/h = 0.65 ∴hc = 0.65 x 8 = 5.2 m
hi*/h = 0.58 ∴hi* = 0.58 x 8 = 4.64 m
hc*/h = 0.73 ∴hc* = 0.73 x 8 = 5.84 m
Part III: Solutions
Thus, in x-direction, wherein, h/L = 0.5, impulsive and convective

masses are almost same. Whereas, in Y-direction, for which h/L = 0.8,
more liquid contributes to impulsive mass.
Also note the difference in hi and hi* values ( or hc and hc*) for two
cases. For shallow tanks, inclusion of base pressure significantly changes
the height .
Additional Slide for Solution 1
Insert after slide 1 of Solution 1
1
Correction in Part II True/False - 1.2
Notice that there is a correction in Part II

True/False – 1.2
1.2)Impulsive mass is less than convective mass for

short tanks.
It is printed as
False
It should be printed as
TRUE
Assignment 2
Date of assignment: 20 January
1
Read the following:
Clause 4.5 of the Guidelines
Clause 6.4 of IS 1893(Part 1):2002
Clause 3.4, 4.2.1.1, and 5.2.6 of IS 1893:1984
© Sudhir K. Jain, IIT Kanpur E-Course on Seismic Design of Tanks / January 2006 Assignment 2/ slide 2
2.1) Seismic force depends on mass of the structure.
2.2) In IS 1893(Part 1):2002, there are five seismic zones.
2.3) In IBC 2003, response modification factor for buildings
with good ductility is lower than that for elevated tanks on
frame type staging.
2.4) In IS 1893:1984, performance factor, K for buildings with
good ductility is more than that for elevated tanks.
2.5) Importance factor for tanks is higher than for buildings.
2.6)Damping of structure does not affect base shear
coefficient.
2.7) In IS 1893(Part 1):2002, base shear coefficient depends
on type of foundation.
Part III: Questions
2.1)As per IS 1893:1984, obtain base shear coefficient for
following structures using response spectrum method
(Fo = 0.4, β = 1.0, consider 5% damping)
1) A building with good ductility and time period of 1.1 sec.
2) A building with low ductility and time period of 1.1 sec.
3) An elevated tank with time period of 1.1 sec.
2.2) For following elevated tanks, obtain base shear
coefficient as per IS 1893:1984 (Fo = 0.4, β = 1.0)
1) Time period = 0.5 sec; damping of 0%, 2% and 5%
Solution 2

Date of solution: 23 January
E-Course on seismic design of tanks

1
Part II: True/ False
2.1) Seismic force depends on mass of structure.
True
2.2) In IS 1893(Part 1):2002, there are five seismic zones.
False
There are four zones, numbered from II to V.
2.3) In IBC 2003,response modification factor for buildings
with good ductility is lower than that for elevated tanks on
frame type staging
False
For a building with good ductility, response modification
factor, R = 8.0 as against, R = 3.0 for elevated tanks on
frame staging
© Sudhir K. Jain, IIT Kanpur E-Course on Seismic Design of Tanks / January 2006 Solution 2 / Slide 2
Part II: True/ False (contd…)
2.4) In IS 1893:1984,performance factor, K for buildings with

Good ductility is more than that for elevated tanks
False
Performance factor, K is same (K = 1.0), for a building with
good ductility and elevated tanks.
2.5) Importance factor for tanks is higher than for buildings.

True
2.6) Damping of structure does not affect base shear

coefficient.
False
See Table 3 of IS 1893(Part 1):2002 and Fig. 2 of IS 1893:1984
Part II: True/ False (contd…)
2.7) In IS 1893(Part 1):2002, base shear coefficient depends

on type of foundation.
False
In IS 1893(Part 1):2002, base shear coefficient does not
depend on type of foundation. Rather, in IS 1893:1984, it
use to depend on type of foundation.
Part III: Solutions
2.1)As per IS 1893:1984, obtain base shear coefficient for
following structures using response spectrum method
(Fo = 0.4, β = 1.0; consider 5% damping)
1) A building with good ductility and time period of 1.1 sec.
2) A building with low ductility and time period of 1.1 sec.
3) An elevated tank with time period of 1.1 sec.
Solution: In response spectrum method, base shear coefficient is

expressed as:
Ah = KβIFoSa/g for buildings
Ah = βIFoSa/g for elevated tanks
We have, Fo = 0.4, β = 1.0
Part III: Solutions
All three structures have time period of 1.1 sec and damping of 5%,
and hence, from Figure 2 of IS 1893:1984, Sa/g = 0.1
For building with good ductility:

K = 1.0, I = 1.0
Ah = KβIFoSa/g = 1.0 x 1.0 x 1.0 x 0.4 x 0.1 = 0.04
For building with low ductility:
K = 1.6, I = 1.0
Ah = KβIFoSa/g = 1.6 x 1.0 x 1.0 x 0.4 x 0.1 = 0.064
For elevated tank:
I = 1.5
Ah = βIFoSa/g = 1.0 x 1.5 x 0.4 x 0.1 = 0.06
Part III: Solutions
2.2) For following elevated tanks, obtain base shear

coefficient as per IS 1893:1984 (Fo = 0.4, β = 1.0)
Solution: Base shear coefficient for tank is:

Ah = βIFoSa/g
We have, Fo = 0.4, β = 1.0 and Importance factor, I = 1.5
Values of Sa/g are to be obtained from Fig. 2 of IS 1893:1984
Part III: Solutions
Time period Damping Sa/g Ah = βIFoSa/g

(Sec) (%)
0 0.50 1.0 x 1.5 x 0.4 x 0.5 = 0.300
0.5 2 0.25 1.0 x 1.5 x 0.4 x 0.25 = 0.150
5 0.16 1.0 x 1.5 x 0.4 x 0.16 = 0.096
0 0.15 1.0 x 1.5 x 0.4 x 0.15 = 0.090
1.4 2 0.105 1.0 x 1.5 x 0.4 x 0.105 = 0.063
5 0.08 1.0 x 1.5 x 0.4 x 0.08 = 0.048
Note the effect of damping on base shear coefficient
Assignment 3
1
Read Clause 4.2.2, 4.2.3 and 4.3 of the
Guidelines
3.1)Time period does not depend on mass of the structure.
3.2)Time period increases with stiffness.
3.3)Impulsive mode time period of ground supported tanks is
usually less than 0.4 seconds.
3.4)Convective mode time period depends only on aspect
ratio of the container.
3.5)In elevated tanks, convective mode time period
depends on lateral stiffness of staging.
3.6)While finding lateral stiffness of frame staging, braces shall
be treated as rigid beams
Part III: Questions
3.1) A ground-supported rectangular tank has length of 12
m and width of 6 m. Liquid height is 4m. Find time
period of convective mode in the two horizontal
directions.
3.2) An elevated tank with circular cylindrical container has
internal diameter of 12 m and water height is 4.5 m.
Weight of container is 3,000 kN and weight of staging
is 1,700 kN. Lateral stiffness of staging is 50,000
kN/m. Obtain two mass model and find time period of
impulsive and convective modes.
Solution 3

Date of solution: 27 January
1
3.1)Time period does not depend on mass of the structure.
False
3.2)Time period increases with stiffness.

False
Time period decreases with stiffness and is inversely
proportional to square root of stiffness. T = 2π M/K
© Sudhir K. Jain, IIT Kanpur E-Course on Seismic Design of Tanks / January 2006 Solution 3/ slide 2
3.3)Impulsive mode time period of ground supported tanks is

usually less than 0.4 seconds.
True
3.4)Convective mode time period depends only on aspect

ratio of the container.
False
It depends not only on aspect ratio but also on diameter
or lateral dimension of container. See formula for Tc in slide
no. 65.
3.5)In elevated tanks, convective mode time period

depends on lateral stiffness of staging.
False
3.6)While finding lateral stiffness of frame staging, braces shall

be treated as rigid beams.
False
For proper evaluation of lateral stiffness, it is necessary that
flexibility of brace is included in the model.
Part III: Solutions
3.1) A ground-supported rectangular tank has length of 12 m and
width of 6 m. Liquid height is 4m. Find time period of convective
mode in the two horizontal directions.
Solution:
Y
12 m
6m X
Plan view
Part III: Solutions
1) Analysis in X-direction:
Length, L = 12.0 m, Height of liquid, h = 4.0 m
∴ h/L = 4/12 = 0.33
From Figure 7 of guideline, for h/L = 0.33, one gets Cc = 4.0
Time period of convective mode, TC = CC Lg
TC = 4.0 (12.0 9.81)
= 4.42 Sec.
2) Analysis in Y-direction:
Length, L = 6.0 m, h = 4.0 m, ∴ h/L = 4/6 = 0.67
From Figure 7 of guideline, for h/L = 0.67, one gets, Cc = 3.65
Time period of convective mode, T C = C C Lg
Tc = 3.65 (6.0 9.81)
= 2.85 Sec.
Part III: Solutions
3.2) An elevated tank with circular cylindrical container has internal

diameter of 12 m and water height is 4.5 m. Weight of container
is 3,000 kN and weight of staging is 1,700 kN. Lateral stiffness of
staging is 50,000 kN/m. Obtain two mass model and find time
period of impulsive and convective modes.
Solution:
Internal diameter, D = 12 m, Water height, h = 4.5 m.
Container is circular cylinder,
∴ Volume of water = π/4 x D2 x h = π /4 x 122 x 4.5 = 509 m3.
mass of water, m = 509 t.
h/D = 4.5/12 = 0.375
From Figure 2 of guideline, for h/D = 0.375:
mi/m = 0.43, mc/m = 0.545 and Kch/mg = 0.665
Part III: Solutions
mi = 0.43 x m = 0.43 x 509 = 218.9 t

mc = 0.545 x m = 0.545 x 509 = 277.4 t
Kc = 0.665 x m g/h = 0.665 x 509 x 9.81/4.5 = 737.9 kN/m
Weight of container = 3,000 kN

∴ Mass of container = 3000/9.81 = 305.8 t
Weight of staging = 1,700 kN

∴ Mass of staging = 1700/9.81 = 173.3 t
Structural mass of tank, ms = mass of container +1/3rd mass of staging

= 305.8 + 1/3 x 173.3
= 363.6 t
Lateral stiffness of staging, Ks = 50,000 kN/m
Part III: Solutions
Thus, we have:
mi =218.9 t, ms = 363.6 t, mc = 277.4 t,
Kc = 737.9 kN/m and Ks = 50000 kN/m
Two mass model:
mc 277.4 t
Kc 737.9 kN/m
218.9t +
mi + ms
363.6t
Ks 50000 kN/m
Part III: Solutions
For evaluating time period, this two mass model will be treated
as two uncoupled single degree of freedom system:
277.4 t
737.9 kN/m
218.9t +
582.5 t 363.6t
277.4 t
50000 kN/m 50000 kN/m
737.9 kN/m
(a) Two mass idealization (b) Uncoupled system
Part III: Solutions
mi + ms
Now, Impulsive Time Period, Ti = 2π
Ks
218.9 + 363.6
Ti = 2π
50,000
= 0.678 Sec.
mc
Convective Time Period, Tc = 2π
Kc
277.4
Tc = 2π
737.9
= 3.85 Sec.
Assignment 4
1
Read Clause 4.4, 4.5 of the Guideline
4.1) R values for elevated tanks with SMRF staging are lower
than for buildings with SMRF.
4.2)Frame staging has less redundancy than shaft staging.
4.3)Damping in convective mode depends on material of
tank.
4.4)Elevated tanks are inverted pendulum type structures.
4.5)Thin shafts have more ductility than thick shafts.
4.6)Response reduction factor for tanks depends on
importance of tank.
4.7)In elevated tanks, damping in impulsive mode depends
on material of container.
Part III: Questions
4.1) A ground-supported rectangular RC water tank with
fixed base is located in zone III on medium soil. Along
the length direction, Ti = 0.2 sec and Tc = 5.0 sec.
Along the width direction, Ti = 0.4 sec and Tc = 3.0
sec. Find impulsive and convective base shear
coefficients in both the directions.
Part III: Questions
4.2) An elevated water tank on ductile RC frame staging is
located in zone V. Ti = 1.5 sec and Tc = 4.0 sec. Find
the base shear coefficient in impulsive and convective
mode for following soil conditions.
i) Hard ii) Medium iii) Soft.
Solution 4

Date of solution: 01 February 2006
1
4.1) R values for elevated tanks with SMRF staging are lower
than for buildings with SMRF
True
For elevated tanks with SMRF staging R = 2.5, whereas for
buildings with SMRF, R = 5.0
4.2)Frame staging has less redundancy than shaft staging
False
4.3)Damping in convective mode depends on material of

tank
False
Convective mode has 0.5% damping for all types of tank
materials.
4.4)Elevated tanks are inverted pendulum type structures

True
4.5)Thin shafts have more ductility than thick shafts

False
4.6)Response reduction factor for tanks depends on

importance of tank.
False
Response reduction factor depends on ductility,
redundancy and overstrength. Importance of tank
decides the value of importance factor, I.
4.7)In elevated tanks, damping in impulsive mode depends

on material of container
False
In elevated tanks, damping in impulsive mode depends on
material of staging and not of container.
Part III: Solutions
4.1) A ground supported rectangular RC water tank with fixed base is
located in zone III on medium soil. Along the length direction, Ti
= 0.2 sec and Tc = 5.0 sec. Along the width direction, Ti = 0.4
sec and Tc = 3.0 sec. Find impulsive and convective base shear
coefficients in both the directions.
Solution:
Width Direction
Length Direction
Plan of tank
Part III: Solutions
Zone III, Z = 0.16 (As per Table 2 of IS 1893 (Part I): 2002)
Tank is being used for water storage, hence
I = 1.5 (As per Table 1 of the Guideline)
For RC ground supported tank with fixed base,
R = 2.0 (As per Table 2 of the Guideline)
Part III: Solutions
1) Analysis in Length Direction:

Impulsive mode time period, Ti = 0.2 sec,
Damping in impulsive mode = 5% (RC tank)
∴ (Sa/g)i = 2.5 (Clause 4.5.3 of the Guideline)
Impulsive base shear coefficient, (Ah)i = Z/2 x I/R x (Sa/g)i

= 0.16/2 x 1.5/2.0 x 2.5 = 0.15
Convective mode time period, Tc = 5.0 sec

Damping in convective mode = 0.5% (Clause 4.4 of the Guideline)
∴ (Sa/g)c = 1.75 x 1.36/5.0 (Clause 4.5.3 of the Guideline)
= 0.476
Note, factor 1.75 is for 0.5% damping (Clause 4.5.4 of the
Guideline)
Part III: Solutions
Convective base shear coefficient (Ah)c= Z/2 x I/R x (Sa/g)c

= 0.16/2 x 1.5/2.0 x 0.476
= 0.029
Hence, in length direction, (Ah)i = 0.15 and (Ah)c = 0.029
2) Analysis in Width Direction:

Impulsive mode time period, Ti = 0.4 Sec,
Damping in impulsive mode = 5% (RC tank)
∴ (Sa/g)i = 2.5 (Clause 4.5.3 of the Guideline)
Impulsive base shear coefficient, (Ah)i = Z/2 x I/R x (Sa/g)i
= 0.16/2 x 1.5/2.0 x 2.5 = 0.15
Part III: Solutions
Convective mode time period, Tc = 3.0 sec

Damping in convective mode = 0.5% (Clause 4.4 of the Guideline)
∴ (Sa/g)c = 1.75 x 1.36/3.0 (Clause 4.5.3 of the Guideline)
= 0.793
Convective base shear coefficient (Ah)c= Z/2 x I/R x (Sa/g)c
= 0.16/2 x 1.5/2.0 x 0.793
= 0.048
Hence, in width direction, (Ah)i = 0.15 and (Ah)c = 0.048
Part III: Solutions
4.2) An elevated water tank on ductile RC frame staging is located in

zone V. Ti = 1.5 sec and Tc = 4.0 sec. Find the base shear
coefficient in impulsive and convective mode for following soil
conditions.
i) Hard ii) Medium iii) Soft.
Solution:
Zone V; Z = 0.36 (As per Table 2 of IS 1893 (Part I): 2002)
Tank stores water, hence, I = 1.5 (As per Table 1 of the Guideline)
Staging is ductile RC frame, hence, R = 2.5
(As per Table 2 of the Guideline)
Part III: Solutions
Time period of impulsive mode, Ti = 1.5 sec

Damping is 5% (RC staging)
Values of (Sa/g)i and impulsive base shear coefficient , (Ah)i for
different types of soil are given below:
Soil type (Sa/g)i (Ah)i = Z/2 x I/R x (Sa/g)i

Hard 1/Ti = 0.67 = 0.36/2 x 1.5/2.5 x 0.67 = 0.072
Medium 1.36/Ti = 0.91 = 0.36/2 x 1.5/2.5 x 0.91 = 0.098
Soft 1.67/Ti = 1.11 = 0.36/2 x 1.5/2.5 x 1.11 = 0.120
Part III: Solutions
Time period of convective mode, Tc = 4.0 sec

Damping is 0.5%
Values of (Sa/g)c and convective base shear coefficient , (Ah)c for
different types of soil are given below:
Soil type (Sa/g)c (Ah)c = Z/2 x I/R x (Sa/g)c

Hard 1/Tc x1.75 = 0.36/2 x 1.5/2.5 x 0.438 = 0.047
= 0.438
Medium 1.36/Tc x 1.75 = 0.36/2 x 1.5/2.5 x 0.595 = 0.064
= 0.595
Soft 1.67/Tc x 1.75 = 0.36/2 x 1.5/2.5 x 0.731 = 0.079
= 0.731
Assignment 5
Date of assignment: 3 February

Date of solution: 5 February
1
Read Clause 4.6, 4.7 and 4.8 of the Guidelines
5.1)In ground supported tanks, base shear at the bottom of
wall is same as base shear at the bottom of base slab.
5.2)In elevated tanks, base moment at the top of footing
includes effect of base pressure.
5.3)For calculating bending moment at the bottom of wall,
effect of base pressure is included.
5.4)In elevated tanks, impulsive mass of liquid, mi, acts at CG
of empty container.
5.5)In IITK-GSDMA Guideline, impulsive and convective base
shear are combined using absolute summation rule.
5.6)Critical direction of seismic loading is always same for
braces and columns of frame staging.
5.7)In elevated tanks, base shear at the bottom of container
wall, will not depend on mass of staging.
Part III:
5.1) Plan and elevation of a rectangular RC tank are shown below along
with some relevant data. Thickness of wall is 250 mm, roof slab is
120 mm thick and base slab is 250 mm thick. Obtain base shear,
bending moment at the bottom of wall and overturning moment in
X and Y directions.
0.1 m projection
12 m 6m along the
4.5 m periphery
Y 12 m
X Plan
Elevation
Earthquake mi hi hi * mc hc hc* (Ah)i (Ah)c
Direction (t) (m) (m) (t) (m) (m)
X 138 1.69 4.73 189 2.48 5.07 0.15 0.024
Y 230 1.69 2.61 112 2.93 3.29 0.15 0.037
Part III:
5.2) An elevated water tank has CG of empty container
at 3.0 m from top of floor slab. Height of staging
from footing top to top of floor slab is 12 m.
Impulsive and convective mode parameters are: mi
= 140 t, hi* = 3.43, mc = 110 t, hc* = 3.43 m, ms =
225 t. Base shear coefficient of impulsive mode is
0.15 and of convective mode is 0.06. Find base
shear and base moment at the bottom of staging.
Solution 5

1
5.1)In ground supported tanks, base shear at the bottom of

wall is same as base shear at the bottom of base slab.
False
Base shear at the bottom of base slab is sum of base
shear at the bottom of wall and shear due to mass of
base slab.
5.2)In elevated tanks, base moment at the top of the footing
includes effect of hydrodynamic pressure at the base.
True
Since staging is below the base slab (or floor slab), base
moment in staging includes the bending effect caused by
hydrodynamic pressure on base slab.
5.3)For calculating bending moment at the bottom of wall,

effect of base pressure is included.
False
Base pressure does not affect bending moment in wall.
Rather, base pressure affects overturning moment, which
is obtained at the bottom of base slab.
5.4)In elevated tanks, impulsive mass of liquid, mi, acts at CG

of empty container.
False
In elevated tanks, impulsive mass of liquid, mi, is
considered to act at hi*. Structural mass of tank, ms, acts
at CG of empty container.
5.5)In IITK-GSDMA Guideline, impulsive and convective base

shear are combined using absolute summation rule.
False
They are combined using Square Root of Sum of Squares
(SRSS) rule.
5.6)Critical direction of seismic loading is always same for braces
and columns of frame staging.
False
Columns and braces of frame staging can have different
critical directions.
5.7)In elevated tanks, base shear at the bottom of container

wall, will not depend on mass of staging.
True and False
True: Base shear at any section depends on all the horizontal

forces acting above that section. Since staging is below
the container wall, it is not going to affect base shear at
the bottom of container. Base shear at the bottom of
staging depends on mass of staging.
False: Mass of staging affects the natural period of the tank

which decides the seismic coefficient (Ah)i .
Part III: Solutions
5.1) Plan and elevation of a rectangular RC tank are shown below

along with some relevant data. Thickness of wall is 250 mm, roof
slab is 120 mm thick and base slab is 250 mm thick. Obtain base
shear, bending moment at the bottom of wall and overturning
moment in X and Y directions.
0.1 m projection
along the
12 m 6 m 4.5 m periphery
Y
12 m
X Plan Elevation
Earthquake mi hi hi * mc hc hc* (Ah)i (Ah)c
Direction (t) (m) (m) (t) (m) (m)
X 138 1.69 4.73 189 2.48 5.07 0.15 0.024
Y 230 1.69 2.61 112 2.93 3.29 0.15 0.037
Part III: Solutions
Solution:
First we calculate mass of roof slab, wall, and base slab.
Weight density of concrete = 25 kN/m3
Thickness of roof slab= 120 mm

Plan dimensions of roof slab = 12.5m x 6.5m
∴ Weight of roof slab = 12.5 x 6.5 x 0.12 x 25 = 243.8 kN
and mass of roof slab, mt = 243.8/9.81 = 24.9 t
Thickness of wall = 250 mm

Internal dimensions are 12m x 6m and height is 4.5 m.
∴ Mass of wall, mw = 2 x (12.25 + 6.25) x 0.25 x 4.5 x 25/9.81
= 106.1 t
Part III: Solutions
Thickness of base slab, tb = 250 mm

Plan dimensions of base slab are 12.7 m x 6.7 m
∴ Mass of base slab = 12.7 x 6.7 x 0.25 x 25/9.81 = 54.2 t
Since tank is rectangular in plan, it will have different seismic forces in

X- and Y-direction.
Analysis in X- direction
In X-direction, we have,
mi = 138 t, mc = 189 t, hi = 1.69 m, hi*= 4.73 m,
hc = 2.48 m, hc*= 5.07 m, (Ah)i = 0.15, (Ah)c = 0.024
Part III: Solutions

= 0.15 x (138 + 106.1 + 24.9) x 9.81
= 395.8 kN
Vc = (Ah)c mc g
= 0.024 x 189 x 9.81
= 44.5 kN
Total base shear at bottom of wall is
V = Vi2 + Vc2 = (395.8 ) + (44.5 )

2 2
= 398.3kN
Part III: Solutions

= 0.15 x (138 x 1.69 + 106.1 x 2.25 + 24.9 x 4.56) x 9.81
= 861.5 kN-m
Mc = (Ah)c mc hc g
= 0.024 x 189 x 2.48 x 9.81
= 110.4 kN-m
M = Mi2 + M2c = (861.5 ) 2

+ (110.4 ) = 868.5kN - m
2
Part III: Solutions
This is total bending moment and will be shared by two walls which
are perpendicular to the direction of seismic force. Hence, bending
moment at the bottom of one wall = 868.5/2 = 434.25 kNm.
Part III: Solutions
Impulsive overturning moment at the bottom of base slab is

= 0.15 x [138(4.73 + 0.25) + 106.1(2.25 + 0.25) +
24.9(4.56 + 0.25) + 54.2 x 0.25/2] x 9.81
= 1588 kN-m
Convective overturning moment at the bottom of base slab is
= 0.024 x 189 x (5.07+ 0.25) x 9.81
= 236.7 kN-m
Total overturning moment at bottom of base slab is
M =*
(M ) + (M )
* 2
i
* 2
c = (1588 )2 + (236.7 )2 = 1605kN - m
Part III: Solutions
Analysis in Y- direction
In Y – direction we have,
mi = 230 t, mc = 112 t, hi = 1.69 m, hi*= 2.61 m,
hc = 2.93 m, hc*= 3.29 m, (Ah)i = 0.15, (Ah)c = 0.037
= 0.15 x (230 + 106.1 + 24.9) x 9.81 = 531.2 kN

Vc = (Ah)c mc g
= 0.037 x 112 x 9.81 = 40.7 kN
Total base shear at bottom of wall is
V = Vi2 + Vc2 = (531.2 ) + (40.7 )

2 2
= 532.8kN
Part III: Solutions

= 0.15 x (230 x 1.69 + 106.1 x 2.25 + 24.9 x 4.56) x 9.81
= 1090 kN-m
Mc = (Ah)c mc hc g
= 0.037 x 112 x 2.93 x 9.81
= 119.1 kN-m
M = Mi2 + M2c = (1090.3 )2 + (119.1 )2 = 1097kN - m
Part III: Solutions
This is total bending moment and will be shared by two walls which
are perpendicular to the direction of seismic force. Hence, bending
moment at the bottom of one wall = 1097/2 = 548.5 kN-m.
Part III: Solutions
Impulsive overturning moment at the bottom of base slab is

= 0.15 x [230(2.61 + 0.25) + 106.1(2.25 + 0.25) +
24.9(4.56 + 0.25) + 54.2 x 0.25/2] x 9.81
= 1544 kN-m
Convective overturning moment at the bottom of base slab is
= 0.037 x 112 x (3.29+ 0.25) x 9.81
= 143.9 kN-m
Total overturning moment at bottom of base slab is
M =*
(M ) + (M )
* 2
i
* 2
c = (1544 )2 + (143.9 )2 = 1551kN - m
Part III: Solutions
5.2) An elevated water tank has CG of empty container at 3.0 m

from top of floor slab. Height of staging from footing top to
top of floor slab is 12 m. Impulsive and convective mode
parameters are: mi = 140 t, hi* = 3.43, mc = 110 t, hc* = 3.43
m, ms = 225 t. Base shear coefficient of impulsive mode is
0.15 and of convective mode is 0.06. Find base shear and
base moment at the bottom of staging.
Solution:
Impulsive base shear at the bottom of staging is
Vi = (Ah)i (mi + ms) g
= 0.15 x (140 + 225) x 9.81 = 537.1 kN
Convective base shear at the bottom of staging is
Vc = (Ah)c mc g
= 0.06 x 110 x 9.81 = 64.7 kN
Total base shear at bottom of staging is
V = Vi2 + Vc2 = (537.1) + (64.7 )
2 2
= 541.0kN
Part III: Solutions
Distance of CG of container from bottom of staging,

hcg = 3.0 + 12.0 = 15.0 m
Impulsive base moment at the bottom of staging is
Mi* = (Ah)i [mi (hi* + hs) + ms hcg] g
= 0.15 x [140 (3.43 + 12) + 225 x 15.0] x 9.81
= 8145 kN-m
Convective base moment at the bottom of staging is
Mc* = (Ah)c mc (hc* + hs) g
= 0.06 x 110 x (3.43 + 12) x 9.81 = 999.0 kN-m
Total base moment at bottom of staging is
M* = (M ) + (M )
* 2
i
* 2
c = (8145 )2 + (999.0 )2 = 8206kN - m
Thus, base shear = 541.0 kN & Base moment = 8206 kN-m
Assignment 6
Date of assignment: 8 February 2006
1
Read Clause 4.9, 4.10, 4.11, 4.12 of the
Guideline
6.1)Convective pressure varies linearly along wall height.
6.2)In circular tanks, impulsive pressure on wall varies in
circumferential direction.
6.3)In rectangular tanks, convective pressure on a particular
wall varies along the length of that wall.
6.4)Due to vertical ground acceleration, lateral pressure
exerted by the liquid on the wall changes.
6.5)In rectangular tanks, hydrodynamic pressure acts on
walls parallel to the direction of seismic force.
6.6)Impulsive pressure is maximum at the bottom of wall.
6.7)Convective hydrodynamic force is more significant for
tall tanks.
Part III: Questions
6.1) A ground supported rectangular RC water tank has plan
dimensions of 12 x 6 m and water height of 4.5 m. Tank is fixed
at base. Wall has uniform thickness of 200 mm. Tank is located on
hard soil in seismic zone III. In X – direction: (Ah)i = 0.15, (Ah)c =
0.024. In Y – direction: (Ah)i = 0.15, (Ah)c =0.037.
Find in X- and Y- directions:

12 m 6m 1) Linearised impulsive and convective
pressure distribution on the walls
Y
2) Pressure due to vertical excitation
X Plan 3) Pressure due to wall inertia
4) Total pressure at the bottom of wall
Solution 6
Date of assignment: 8 February 2006

Date of solution: 10 February 2006
1
6.1)Convective pressure varies linearly along wall height.
False
Convective pressure has curvilinear distribution along wall
height.
6.2)In circular tanks, impulsive pressure on wall varies in
circumferential direction.
True
Impulsive pressure has cosφ variation in circumferential
direction.
6.3)In rectangular tanks, convective pressure on a particular
wall varies along the length of that wall.
False
In rectangular tanks, convective pressure on a wall
remains constant along the length of that wall. This
assumes that the direction of shaking is perpendicular to
the wall.
6.4)Due to vertical ground acceleration, lateral pressure
exerted by the liquid on the wall changes.
True
when subjected to vertical excitation, weight density of
liquid increases or decreases depending on direction of
vertical acceleration. Hence, hydrostatic pressure which
depends on weight density of liquid, will also increase or
decrease. This additional pressure is called pressure
due to vertical excitation.
6.5)In rectangular tanks, hydrodynamic pressure acts on

walls parallel to the direction of seismic force.
False
In rectangular tanks, hydrodynamic pressure acts on
walls perpendicular to the direction seismic force.
6.6)Impulsive pressure is maximum at the bottom of wall.

True
6.7)Convective hydrodynamic force is more significant for tall

tanks.
False
Convective pressure in tall tanks (high value of h/D or h/L) is
not as significant as for short tanks (low h/D or h/L ratio).
Refer Figures 10(a) and 11(a) of Guidelines
Part III: Solution
6.1) A ground supported rectangular RC water tank has plan
dimensions of 12 x 6 m and water height of 4.5 m. Tank is fixed
at base. Wall has uniform thickness of 200 mm. Tank is located on
hard soil in zone III. In X – direction: (Ah)i = 0.15, (Ah)c = 0.024,
In Y – direction: (Ah)i = 0.15, (Ah)c =0.037.
Find in X- and Y- directions:

12 m 6m 1) Linearised impulsive and convective
pressure distribution on wall
Y
2) Pressure due to vertical excitation
X Plan 3) Pressure due to wall inertia
4) Total pressure at the bottom of wall
Part III: Solution
Solution:
Capacity of tank = 12 x 6 x 4.5 = 324 m3
∴ Mass of water = 12 x 6 x 4.5 x 1000 = 324,000 kg
Thickness of wall, tw = 200 mm
Weight density of concrete = 25 kN/m3
Analysis in X- direction:
(Ah)i = 0.15, (Ah)c = 0.024
h = 4.5 m, L = 12 m, B = 6 m, h/L = 4.5/12 = 0.375.
For h/L = 0.375, from Figure 3 of the Guideline, we have:
mi/m = 0.425, mi = 0.425 x 324,000 = 137,700 kg
mc/m = 0. 584, mc = 0. 584 x 324,000 = 189,200 kg
hi/h = 0.375, hi = 0.375 x 4.5 = 1.69 m
hc/h =0.551, hc = 0.551 x 4.5 = 2.48 m
Part III: Solution
a) Impulsive mode
( A h )i m i 0 . 15 × 137 ,700 × 9 . 81
qi = g = = 16 . 90 kN/m
2B 2×6
ai =
qi
2
(4 h − 6 h i ) =
16 . 90
2
(4 × 4 .5 − 6 × 1 .69 ) = 6 .60 kN/m 2
h 4 .5
bi = 2 (6hi − 2h ) = ( )
qi 16.90
6 × 1 . 69 − 2 × 4 . 5 = 0 . 95kN/m 2
h 4.52 b = 0.95 kN/m2

i
Linear impulsive pressure distribution ai = 6.60 kN/m2

Part III: Solution
b) Convective mode
( Ah ) c m c 0 . 024 × 189 , 200 × 9 . 81

qc = g = = 3 . 71 kN/m
2B 2×6
a c = 2 (4 h − 6 hc ) = ( )
qc 3 .71
4 × 4 . 5 − 6 × 2 . 48 = 0 . 57 kN/m 2
h 4 .52
bc = 2 (6hc − 2h ) = ( )
qc 3.71
6 × 2 . 48 − 2 × 4 . 5 = 1 . 08 kN/m 2
h 4.52 2
bc = 1.08 kN/m
Linear convective pressure distribution ac = 0.57 kN/m2
Part III: Solution
c) Pressure due to vertical acceleration

Zone III, Z = 0.16
I = 1.5
It is ground supported RC tank with fixed base, hence, R = 2.0
For vertical mode, T = 0.3 Sec, ∴ Sa/g = 2.5
Av = 2/3 . Z/2. I/R . Sa/g = 2/3 x 0.16/2 x 1.5/2.0 x 2.5 = 0.10
Maximum pressure due to vertical acceleration occurs at y=0, i.e., at
the base of wall,
∴ pv = (Av) ρ g h (1- y/h)
= 0.10 x 1000 x 9.81 x 4.5 x (1- 0/4.5)
= 4.42 kN/m2
d) Pressure due to wall inertia,
pww = (Ah)i t ρm g
= 0.15 x 0.2 x 25 = 0.75 kN/m2
Part III: Solution
Total pressure,
p= (p iw + p ww )
2
+ p cw + p v
2 2
= (6 . 60
+ 0 . 75 ) + 0 . 57 2 + 4 . 42 2
2
= 8.59 kN/m 2
Analysis in Y- direction:
(Ah)i = 0.15, (Ah)c = 0.037
h = 4.5 m, L = 6 m, B = 12 m, h/L = 4.5/6 = 0. 75.
For h/L = 0.375, from Figure 3 of the Guideline, we have:
mi/m = 0.71, mi = 0.71 x 324,000 = 230,000 Kg
mc/m = 0. 346, mc = 0. 346 x 324,000 = 112,100 Kg
hi/h = 0.375, hi = 0.375 x 4.5 = 1.69 m
hc/h =0.551, hc = 0.65 x 4.5 = 2.93 m
Part III: Solution
a) Impulsive mode
( A h )i m i 0 . 15 × 230 ,040 × 9 . 81
qi = g = = 14 . 10 kN/m
2B 2 × 12
a i = 2 (4 h − 6 hi ) = (4 × 4 .5 − 6 × 1 .69 ) = 5 .47 kN/m
qi 14 . 10 2
2
h 4 .5
bi = 2 (6hi − 2h ) = ( )
qi 14.10
6 × 1 . 69 − 2 × 4 . 5 = 0 . 79 kN/m 2
h 4.52 b = 0.79 kN/m2

i
Linear impulsive pressure distribution ai = 5.47 kN/m2

Part III: Solution
b) Convective mode
( A h )c m c 0 . 037 × 112 ,104 × 9 . 81

qc = g = = 1 . 70 kN/m
2B 2 × 12
a c = 2 (4 h − 6 h c ) = (4 × 4 .5 − 6 × 2 .93 ) = 0 .035 kN/m
qc 1 . 70 2
2
h 4 .5
bc = 2 (6hc − 2h ) = ( )
qc 1.70
6 × 2 . 93 − 2 × 4 . 5 = 0 . 72 kN/m 2
h 4.52 2
bc = 0.72 kN/m
Linear convective pressure distribution ac = 0.035 kN/m2
Part III: Solution
c) Pressure due to vertical acceleration

Zone III, Z = 0.16
As it stores water, I = 1.5
It is ground supported RC tank with fixed base, R = 2.0
For vertical mode, T = 0.3 Sec, ∴ Sa/g = 2.5
Av = 2/3 . Z/2. I/R . Sa/g = 2/3 x 0.16/2 x 1.5/2.0 x 2.5 = 0.10
Maximum pressure due to vertical acceleration is occur at y=0, i.e., at
the base of wall,
∴ pv = (Av) ρ g h (1- y/h)
= 0.10 x 1000 x 9.81 x 4.5 x (1- 0/4.5)
= 4.42 kN/m2
d) Pressure due to wall inertia,
pww = (Ah)i t ρm g
= 0.15 x 0.2 x 25 = 0.75 kN/m2
Part III: Solution
Total pressure
p= (p iw + p ww )
2
+ p cw + p v
2 2
= (5 .47 2
+ 0 . 75 )2
+ 0 . 035 2
+ 4 . 42 2
= 7.62 kN/m 2
At the base of wall pressure in kN/m2 are:

piw pcw pv pww Total pressure
X-direction 6.6 0.57 4.42 1.15 8.59

Y-direction 5.47 0.035 4.42 1.15 7.62
Assignment 7
1
Read Examples 2 and 3 of the Guideline
7.1) If braces are treated as rigid beams, then seismic forces
are overestimated .
7.2) Inclusion of shear deformation, reduces time period of
shaft staging.
7.3) Effect of shear deformation on lateral stiffness is more
important in shafts with large height-to-diameter ratio.
7.4) If stiffness decreases by 1.44 times then, time period will
increase by 1.2 times.
7.5) Empty tank does not have convective mode of vibration.
Part III: Questions
7.1) A eight column RC frame staging has four brace levels. All
columns have 400 mm diameter; braces are 300 mm wide and
400 mm deep. Grade of concrete is M 25. Top ring beam is 400
mm wide and 750 mm deep. CG of tank is at a distance of 3m
from top ring beam (Refer Figures below). Find the lateral
stiffness by (a) computer analysis of a 3-dimensional model, (b)
approximate analysis considering braces as rigid beams as per
SP:22 approach, and (c) approximate analysis using the approach
of Sameer and Jain (1992). Compare and discuss your results.
3m
4m
3m
4m
4m
4m
4m
Plan Elevation
Part III: Questions
7.2) A RC shaft has inner diameter of 6m and wall thickness of 200
mm. Grade of concrete is M 25. Find the lateral stiffness of shaft
with and without considering shear deformation for following
three staging heights: L = 15 m, 20 m and 25 m. Take Poisson’s
ratio as 0.2.
Solution 7

1
7.1) If braces are treated as rigid beams, then seismic forces
are overestimated .
True
If braces are treated as rigid, they do not allow any
rotation at joints of columns and braces. This leads to
significant overestimation of stiffness. This reduces time
period and hence, higher seismic forces.
7.2) Inclusion of shear deformation reduces time period of

shaft staging.
False
With the inclusion of shear deformation, lateral stiffness
decreases and time period which is inversely proportional
to square root of stiffness, increases.
7.3) Effect of shear deformation on lateral stiffness is more
important in shafts with large height-to-diameter ratio.
False
The lateral stiffness is given by
1
Ks =
L3 L
+
3 EI 0 . 5 AG
Here, L3/3EI is flexural contribution and L/0.5AG is shear contribution.

Flexural contribution varies as L3, whereas, shear contribution varies
linearly with length, L. Assuming same cross section, if height, i.e., L is
more, contribution of flexural deformation increase as L3 whereas
contribution of shear deformation increase linearly. Hence, in shafts
with more height, flexural deformation governs the stiffness and
effect of shear deformation is low. This is also explained in the
solution of problem no. 7.2 of part III of this assignment.
7.4) If stiffness decreases by 1.44 times then, time period will
increase by 1.2 times.
True
Time period T = 2 π (M/K)0.5. If K decreases to K/1.44, then
T will increase by 1.2 times. Note: (1.44)0.5 = 1.2
7.5) Empty tank does not have convective mode of vibration.
True
In empty tank, no water, hence no convective mode.
Part III: Solutions
7.1) A eight column RC frame staging has four brace levels. All
columns have 400 mm diameter; braces are 300 mm wide and
400 mm deep. Grade of concrete is M 25. Top ring beam is 400
mm wide and 750 mm deep. CG of tank is at a distance of 3m
from top ring beam (Refer Figures below). Find the lateral
stiffness by (a) computer analysis of a 3-dimensional model, (b)
approximate analysis considering braces as rigid beams as per
SP:22 approach, and (c) approximate analysis using the approach
of Sameer and Jain (1992). Compare and discuss your results.
3m
4m
3m
4m
4m
4m
4m
Plan Elevation
Part III: Solutions
Solution:
a) Stiffness from computer analysis
Staging is modeled using SAP 2000. Frame elements are used to
model braces, columns and circular ring beams. Columns are
considered to be fixed at the base.
Lateral load is to be applied at a height of 3.0 m from circular
ring beam. For this purpose, a rigid link of 3.0 m length is
required at the center of staging. In order to have a node at the
center, eight radial beams of same size as circular beams are also
modeled. From the central node, a vertical rigid link of 3.0 m
length is put. On this rigid link, a lateral force of 10 kN is applied.
Computer model is shown on next slide.
Part III: Solutions
10 kN
Rigid link
Radial beam
Column
Brace
Part III: Solutions
For lateral force of 10 kN, deflection = 0.00267 m
∴ Lateral stiffness of staging, Ks = 10/0.00267

= 3,745 kN/m
Deflection of staging
Part III: Solutions
b) Stiffness by considering braces as rigid beams
If braces are treated as rigid beams, then stiffness of column

between two brace levels or stiffness of column in a panel is
given by
KC = 12EI
L3
E = Modulus of elasticity, I = Moment of inertia and L = Length

of panel.
Part III: Solutions
In the present example, Grade of concrete is M 25

Hence, E = 5000 (25)0.5 = 25000 N/mm2 = 25 x 106 kN/m2
Column diameter = 400 mm = 0.4 m

∴ Moment of inertia, I = π (0.4)4/64 = 1.26 x 10-3 m4
Hence,
K C = 12EI = 12 × 25 × 10 6 × 1 .2 6 × 10 - 3 = 5 906 .3 kN/m

L3 ⎛ 4.0 ⎞
⎜ ⎟
3
⎝ ⎠
In each panel, there are 8 columns. These eight columns act

like eight springs in parallel. Hence, stiffness of each panel
Kp = ∑ K C = 8 × 5906.3 = 47250 kN/m
Stiffness of one panel = 47,250 kN/m
Part III: Solutions
Staging comprises of five such panels. All panels have same stiffness
and they act like five springs in series. Hence, Stiffness of staging Ks is
given by
1 1 1 1 1 1
= + + + +
Ks Kp Kp Kp Kp Kp
i.e., Stiffness of staging, Ks = Kp/5 = 47,250/5 = 9,450 kN/m
Part III: Solutions
c) Stiffness as per approach of Sameer and Jain (1992)

Stiffness of staging is given by
1 1 1
= +
K s K flexure K axial
Where,
1
Np
1 ⎡ ⎤
12EI c N c ⎢ Ib L
K flexure
= ∑K
i =1
, K panel =
h 3
⎢
I L +
−
α I h
⎥
⎥
panel ⎣ b c ⎦
α = 1.0 for end panels and α = 2.0 for intermediate panels
1 2 Np
= ∑H
2
i h
K axial N c A cE R 2 i =1
Part III: Solutions
E = Modulus of elasticity
Ac = Area of one column
Nc = Number of columns
Np = Number of panels
L = Length of each brace beam
R = Radius of staging system
Ib = Moment of inertia of brace beam
h = Height of each panel
Hi = Distance of point of inflection of ith panel from the load
Part III: Solutions
For intermediate panels, point of inflection is at mid-height of panel.

For end panels, point of inflection is at a distance, y from the
restrained end.
3I
b + Ic Top most panel Panel 1
y = L h h
6I
b Panel 2
L
Intermediate panels Panel 3
For the top most panel,restrained
end is at top ring beam, hence, y
Panel 4
is measured from top end.
Bottom most panel Panel 5
For the bottom most panel, y is
measured from bottom end.
Part III: Solutions
For this Example, we have:

E = Modulus of elasticity = 5000 (25)0.5 = 25,000 N/mm2
Ac = Area of one column
= π x (400)2/ 4 = 125,664 mm2
Nc = Number of columns = 8
Np = Number of panels = 5
L = Length of each brace beam
= 3000 x sin (22.50) x 2 = 2,296 mm
R = Radius of staging system = 3,000 mm
= π x (400)4/ 64 = 1.26 x 109 mm4
Part III: Solutions
Ib = Moment of inertia of brace beam = 300 x(400)3/12 = 1.6x109 mm4

= 400 x (750)3 / 12 = 1.41 x 1010 mm4
h = Height of each panel = 4,000 mm
3I
b + Ic 3 × 1.6 × 10 9 1.26 × 10 9
+
y = L h h = 2296 4000 × 4000 = 2301 mm
6I 6 × 1.6 × 10 9
b
L 2296
Distance of point of inflection of each panel from the load

H1 = 3000 + 2301 = 5,301 mm
H2 = 3000 + 4000 + 2000 = 9,000 mm
H3 = 3000+ 8000 + 2000 = 13,000 mm
H4 = 3000 + 12000 + 2000 = 17,000 mm
H5 = 3000 + 20000 – 2301 = 20,699 mm
Part III: Solutions
Now, Kaxial is given by
1 2 N
= 2 i =1 i h
H
p 2
∑
K axial Nc AcE R
2 2 − 18
= = 8 .84 x10
Nc A c E R 2 8 x125664 x 25000 x (3000) 2
Np
2
∑ Hi h = ((5301) 2 + (9000) 2 + (13000) 2 + (17000) 2 + (20699) 2 ) x 4000
i =1
12
= 3 . 98 x 10
1
∴ = 8.84 x10 −18 x 3.98 x1012 = 3.52 x10 − 5 mm / N
K axial
Part III: Solutions
Now, let us find Kpanel

For the upper most panel, ( α = 1.0 )
12 EI c N c ⎡ I cbr L ⎤
K panel = ⎢ ⎥ = 4.5 x 104 N/mm.
⎢⎣ I cbr L + α I c h ⎥⎦
3
h
Similarly, for bottom most panel, ( α = 1.0 )
K panel = ⎢ ⎥ = 3.25 x 104 N/mm.
⎢⎣ I b L + α I c h ⎥⎦
3
h
Similarly, for intermediate 3 panels, ( α = 2.0 )

K panel = ⎢ ⎥ = 2.48 x 104 N/mm.
h3 ⎢⎣ I b L + α I c h ⎥⎦
Part III: Solutions
Np
1 1
K flexure
= ∑K
i =1 panel
1 1 1 2 2 2
= + + + +
K flexure 4.5 ×10 4 3.25 ×10 4 2.48 ×10 4 2.48 ×10 4 2.48 ×10 4
= 2.95 x 10-4 mm/N
Now, Ks is obtained as
1 1 1
= + = 2.95 × 10 − 4 + 3.52 × 10 −5 = 3.302 x10 − 4 mm / N
Ks K K
flexure axial
∴ Ks = 3,029 N/mm = 3,029 kN/m.
Part III: Solutions
Thus, stiffness of staging using these three approaches is

a) From computer analysis of 3-D model
Ks = 3,745 kN/m
b) From approximate analysis considering braces as rigid beams
Ks = 9,450 kN/m
c) From approximate analysis of Sameer and Jain (1992)
Ks = 3,029 kN/m
Part III: Solutions
Thus, if braces are treated as rigid beams, as suggested by SP 22,

stiffness is almost 2.5 times higher than that obtained computer
analysis of 3-D model of staging.
The stiffness obtained by approach suggested by Sameer and Jain is

about 19% lower than that from computer analysis.
In computer analysis, we have put radial beams, which impart

rotational rigidity to columns at top level, Hence, some of the difference
in Sameer and Jain method is due to model error of computer
Analysis.
Part III: Solutions
7.2) A RC shaft has inner diameter of 6m and wall thickness of 200

mm. Grade of concrete is M 25. Find the lateral stiffness of shaft
with and without considering shear deformation for following
three staging heights: L = 15 m, 20 m and 25 m. Take Poisson’s
ratio as 0.2.
Solution:
Lateral stiffness of RC shaft without considering shear deformation
3EI
Ks =
L3
If shear deformations are considered, then

1
Ks =
L3 L
+
3EI κ' AG
Part III: Solutions
Grade of concrete is M 25.

Hence, Modulus of elasticity, E = 5000 x (25)0.5 = 25,000 N/mm2.
= 25000 x 103 kN/m2
G = Shear modulus = E/(2 x (1 + ν))
where, ν = Poisson’s ratio = 0.2
∴ G = E/(2 x (1 + ν)) = 25000 x 103 /(2 x (1 + 0.2)) = 10,417 x103 kN/m2
I = Moment of inertia of shaft cross section

= π x (6.44 – 6.04)/64 = 18.74 m4
A = Area of cross section of shaft

= π x (6.42 – 6.02)/4 = 3.90 m2
κ’ = 0.5 (for hollow circular cross section)
Part III: Solutions
Now, for shaft with L = 15 m:

Ks without considering shear deformation
3EI
Ks =
L3
3x25000x10 3 x18.74 3
= 3
= 416 .4 x10 kN / m
15
Ks with considering shear deformation

1
Ks =
L3 L
+
3EI κ' AG
1
= 3
= 318 .5 x10 3 kN / m
15 15
+
3x25000x10 3 x18 .74 0.5x3.9x10 417x10 3
Part III: Solutions
Thus, with the inclusion of shear deformation, Ks reduces by 23.5%.
Similarly, lateral stiffness is obtained for L = 20 m and 25 m and

results are given in Table below
Height Height-to- Lateral stiffness, Ks (kN/m)
of shaft diameter
without shear with shear % Reduction
ratio
deformation deformation
15 m 2.5 416.4x103 318.5x103 23.5 %
20 m 3.3 175. 7x103 149.7x103 14.8 %
25 m 4.2 90x103 81x103 10 %
Note: Thus, for shafts with large height-to-diameter ratio, shear

deformation has less effect on the lateral stiffness.

Is 1893 - Tanks & Water Retaining

Transféré par

Informations du document

Description originale:

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Is 1893 - Tanks & Water Retaining

Transféré par

Droits d'auteur :

Formats disponibles

Lecture 1

January 17, 2006

 Ground supported tanks

 Petrochemical industries use ground supported

 IS 1893 (Part 1):2002 is for buildings only

 Summation of pressure along entire wall surface

Circular tanks (Plan View)

Note:- Hydrostatic pressure is axisymmetric; hydrodynamic is asymmetric

Rectangular tanks (Plan View)

Hydrostatic pressure Hydrodynamic pressure

Net resultant force = zero Net resultant force ≠ zero

 Static design: Hydrostatic pressure is considered

 Seismic design: Hydrodynamic pressure is

 Procedure for hydrodynamic pressure & force:

 At this point, we will not go into details of

 Total liquid mass, m, gets divided into two

 Impulsive force = mi x acceleration

 Housner suggested graphs for mi and mc

For circular tanks For rectangular tanks

 See next slide for definition of h, D, and L

Plan of Circular tank

Note:- mass density of water is 1000 kg/m3; weight density of water is

From graph, for h/D = 0.375

mi = 0.42 x 150.8 = 63.3 t and

 Impulsive liquid is rigidly attached to wall

Kc/2 Kc/2 Convective liquid

Rigid m Impulsive liquid

 Stiffness associated with convective mass, Kc

Note: - Unit of m is kg, hence unit of Kc is N/m. If we take m in ton, then

 Now, we know liquid masses mi and mc

 Impulsive mass acts at centroid of impulsive

 hi, hc can be obtained from graphs

For circular tanks For rectangular tanks

From graph, for h/D = 0.375;

Note :- Since convective pressure is more in top portion, hc > hi.

 Hydrodynamic pressure also acts on base

Hydrostatic pressure on base Hydrodynamic pressure on base

 Impulsive as well as convective liquid cause

Overturning effect Overturning effect

Note:- Both the overturning effects are in the same direction

 Total overturning effect of wall and base

resultant was applied at hi

Location of resultant of wall Location of Resultant of wall

 Similarly, hc and hc* are defined

Location of resultant of wall

 hi and hi* are such that

 hc and hc* are such that

 hi* is greater than hi

 This completes modeling of liquid

mi = Impulsive liquid mass

= 0.45 for h / D > 1.33

= 0 .45 for h / L > 1.33 mg ⎛ h⎞

 Note, in Table C-1 of the Guideline, there are two

 mi and mc are needed to find impulsive and

 Vi and Vc will cause

 Overturning of the container is due to pressure

 OM at bottom of base slab

 mi and mc will have different accelerations

Impulsive force, Vi = mi x ai = 63.3 x 2.94 = 186.1 kN

Bending moment at bottom of wall due to Vi = Vi x hi

Seismic Design of Liquid Storage Tanks

Response to Questions and Comments on Lecture 1

Ground supported tanks

Petrochemical industries use ground supported

IS 1893 (Part 1):2002 is for buildings only

Summation of pressure along entire wall surface

Static design: Hydrostatic pressure is considered

Seismic design: Hydrodynamic pressure is

Procedure for hydrodynamic pressure & force:

At this point, we will not go into details of

Total liquid mass, m, gets divided into two

Impulsive force = mi x acceleration

Housner suggested graphs for mi and mc

See next slide for definition of h, D, and L

Impulsive liquid is rigidly attached to wall

Stiffness associated with convective mass, Kc

Now, we know liquid masses mi and mc

Impulsive mass acts at centroid of impulsive

hi, hc can be obtained from graphs

Hydrodynamic pressure also acts on base

Impulsive as well as convective liquid cause

Total overturning effect of wall and base

Similarly, hc and hc* are defined

hi and hi* are such that

hc and hc* are such that

hi* is greater than hi

This completes modeling of liquid

Note, in Table C-1 of the Guideline, there are two

mi and mc are needed to find impulsive and

Vi and Vc will cause

Overturning of the container is due to pressure

OM at bottom of base slab

mi and mc will have different accelerations

For design, we need maximum seismic force

Seismic force = mass x maximum acceleration

Maximum acceleration of structure depends on

Seismic design philosophy is such that, design

Thus, base shear coefficient depends on:

Let us examine how following codes have

Response reduction factor, R

Response reduction factor (contd..)

Spectral acceleration coefficient, Sa/g

For other damping, Sa/g values are to be

For higher damping, multiplying factor is less

Let us now look at the provision of IS

Ah= KCβIαo Seismic Coefficient Method (SCM)

β is soil foundation coefficient

Graphs for C and Sa/g from IS 1893:1984

IS 1893:1984 has provisions for elevated tanks

Base shear coefficient

T is time period in seconds

More about SDS and SD1

Response modification factor, R

Important to note that:

Once, base shear coefficient is known, seismic

Let us compare base shear coefficient values

Comparison for buildings

They represent similar seismic hazard level

Building with good ductility is chosen