0% found this document useful (0 votes)
22 views14 pages4 Vol2 B4-11
Uploaded by
Pranav DevaleCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
22 views14 pages4 Vol2 B4-11
Uploaded by
Pranav DevaleCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
/ 14
STUDY ON THE FARTHQUAKE PROOF DESIGN OF
ELEVATED WATER TANK
by: ¥. Sonobel!) and T. Nishikawal!!)
ABSTRACT
Elevated water tank has different vibratory characteristics compared
with ordinary structures, because "water! affects the vibration behavior.
Several papers on this analysis have been published, but it is very com
plicated to analyze precisely its vibration behavior,
‘This report deseribes the possible method of replacing the elevated
water tank by the appropriate equivalent vibratory system by experimental
‘Studies and of analyzing the dynamic response of the structures supporting
the water tank and fluid in the tani,
‘This method may be applicable to the dynamic analysis of the same type
of structares, such a8 petroleum tank, silo ete.
It was confirmed that measured responses of the supporting structur
and tho fluid by vibration tests of small models for the actually designed
elevated wator tanks wore explained by the analysis by replacing the
structures to the modified spring-mase system,
Two models were made: one is a cylindrical tank model, wnich has
60 em diameter and the ratio of the model to the prototype of 1/10 ~ 1/204
another is a spherical model, of same size, Each tank was supported by @
frame which has several kind of rigidity.
By free vibration tosts, experimental equations of the fundamental
portods of the wator (Tw), corresponding to depth of the water and the most
exciting periods of the frames (TP), were obtained. Alter that, foreed
stationary vibration tests were carried out by the electro-magnetic vibration
fable. By those tosts, rosponse curves, periods of the elevated tank includ~
ing higher order, damping coefficients, participation functions were obtained.
Furthermore, forced vibration tests were carried out, using the input of
pseudo El Centro NS 1940 record (time scale 1/2.5, max. acc. 80.2 gal), to
the eylinderical tank, From its response, maxisum displacement and accel
ation of the framo and maximum response height of the water in tho tank were
found out.
Considering the test results, simplified two degree, of freedom system,
which consisted of the weight of fame plus dead water assumed to be fastened
to the tank rigidly, and the weight of the free water assumed to be fastened to
the tank by springs, was analyzed by analog computer 'SERAC.!
From this response analysis, the good agreement between the experi-
‘ment and tho analysis was confirmed. On this basis a possible method of the
‘actual dynamic design of clevated tanks were proposed.
() Dr. Asst, Prof., Chiba Institute of Technology
(i) Assistant of Tokyo Metropolitan University
By a 1z
STUY ON THF RARTHQUAKE PROOF DRSIGN OF
ELEVATED WATER TANK
(i)
Yasuhisa Sonobe") and Takao Nishikawa'
SYNOPSIS
Vibration tests were carried out on 1/10 ~ 1/20 scale models of eylindi-
cal and spherical elevated water tanks,
The test comprises free and forced vibration tests under sinusoidal and
random seismic wave.
In addition, a response analysis was effected on an analog computer to
compare the results with those obtained in the tests.
‘The present research enables the authors to clarify that vibration of
elevated tanks may be expressed simply as vibration of two degrees of freedom
system.
1, Introduction
Rigorous treatment of dynamic analysis of an elevated water tank is very
complicated in comparison with other ordinary structures, because the fluid in
the vessel takes a role as a part of vibrating body. Recently, however, an
analytical method of responses due to arbitrary ground motion has been devel
oped in the range of elasto-plasticity, Therefore, if the clovated water tank
‘could be replaced by an appropriate vibratory system, it is able to predict the
behaviors of the structure supporting the tani and fluid contained in the tanle
during earthquake.
‘The dynamic analysis of these structures is felt keenly necessary by the
fact that great fire of petroleum company in 1964, Niigata earthquake, was
caused mainly by the vibration of fluid in the tank.
Such a study was done in Prof. Muto!s Laboratory at the University of
‘Tokyo in 1957 by one of the authors, In this study the water in the tank was
divided into two parts from the vibrational behavior, one was the dead water
which was fastened to tho tank rigidly and the olher the free water fastened to
the tank by spring. The ratio of water of each part was determined quantative-
ly by experiments (ref. 3.2.(3)), and further it was proposed that the external
force to be taken in the design of the elevated water tank was obtained by
multiplying the sum of the equivalent weights of structure and dead water by
the seismic costficient. After that, the study concerning with the higher
‘modes of vibration of fluid in the tank was done, but the purpose of this study
was concentrated in the problem of the dead waler.
Now in the current study, the above experiments have been more devel
oped and, aiming at a dynamic ‘design taking the coupling effect of structure and
fluid inte’ account, small type models were made to measure the vibration
() Dr. Asst. Prof., Chiba Institute of Technology
(M1) Assistant of Tokyo Metropolitan University
By behavior of structure and fluid contained in the tank under the arbitrary ground
motion besides the stationary vibration experiments on the free water and dead
From these experimental results, possibility was anticipated of replacing
the structure simply with a two freedom vibratory system consisting of the
weight of free water and of the dead load plus dead water. Confirming the
above anticipation, the authors have attempted to make a guiding principle of
the earthquake-proof design of structures of this kind.
2. Outline of experiment
2.1 Models
‘Two prototypes as shown in Fig. 1 wore taken as the, subject of study,
that ia, an actually designed cylindvical elevated water tank of 12 = 40.0 in
height and 250~ 400 tons in volume and'a spherical ono of 12~ 20 m in height
and 40~ 150 tons in volume, As.a principle of model making, neglecting the
viscosity of water and the weight ratio of each part being equal, the reduced
Seale for period may bo taken a5 VS, where S 1s the reduced scale for length
inthe model, because the vibration period of water in the tank is proportional
to the squaro-root of tts diameter.
* Since the vibration behavior had not been measured on th
water tank, the natural period was estimated following the pre
Studies and the result was applied in the design of models to be used.
‘The models were constructed in such a way that, as shown in Fig, 2, the
water tank is supported on a steel frame of 4-pillar type. Three kinds of
frames of different lengths are prepared so as to make the ratio of period of
frame and water be in the range of that in the actual tank. The distance be=
tween the fixed points of frame as plate spring are 39 cm, 64.5 em. 80 em
which are common for the both models. The tank is made of aluminium plates
1mm thick and its diameter is 60 cm in the cylindrical tank and 50 em in the
spherical tank. The bottom shape of both models has a similar shape to the
actual tank and the reduced scale of length corresponds to 1/17 ~ 1/20 in the
cylindrical model and 1/10 1/14 in the spherical model.
2.2 Contents of experiment
‘Static loading tests: The experiment was conducted to calibrat
Between the strain and deformation of the frame.
the relation
Pree vibration tests: Free vibration of the frame was caused by releasing
Suddonly a slight forced deformation, Periods and damping cosificients of the
frames were obtained. The vibration of water was caused by moving the tank
by hand and when the fundamental vibration became stationary, the period and
damping coefficient were obtained.
Stationary forced vibration testa: An electromagnetic vibration table was used
nrtnis Torbed vibration teste, Keoping the total amplitude of the table at cor
tain definite values, the frequency of the sinusoidal wave was varied continu~
ously. a4
Forced vibration teste due to actual earthquake motion: El Centro NS 1940
fartiquake record in which the time avis Is reduced to 1/2.5 was taken as an
nput, and dynamic strain-meter was used for the measurement of response as
in the case of free and forced vibration tests mentioned above. Further, in
order to examine the surface wave of water, uniformly spaced 18 floats were
Uniformly placed in the diameter line in the direction of Vibration, and their
yertical movements were photographed on a movie film. This experiment was,
carried out for the cylindrical type water tank model only.
Results of tests
3.1 Static loading tests
According to the test results, the relationship of the displacement and the
strain of the frame are kept linear to the storey shearing farce. But when the
displacement becomes largo, the rigidity ts slightly lowered and a trivial
hysteresis characteristic 1s observed. Table 1 shows the spring constant ac—
cording to the secant modulus corresponding to the displacements of 4, 6 and
Bmm. ‘The theoretical values in the table are the values calculated on the as~
sumption that the pillars are fixed at the position of bolts. Table 2 shows ex-
perimental and theoretical values of period obtained on the models without
water, From these data, it is considered that the statical and dynamic spring
constants are nearly coincident,
3.2, Free vibration tests
(1) Periods and damping coefficients
Cylindrical model: Fig. 3 shows the relation between the period of the
fundamental vibration of water Tw, the period of the most exciting vibration of
the frame Ty and the depth of water H. H means the water depth from the posi~
tion at the inter section of the bottom surface and the side surface
In each caso, the 1st period of water Tw is extremely large when the
water depth is low, but it is decreased with the increase of the water depth,
Tinally converging to a certain definite value (0.11/D ~0.13/D sec D; d=
ameter in om). Tnis characteristics may be also observable in the cases of
other type of water tanks to be described in 3.2 (3), Also Tw increases as the
rigidity of frame becomes higher when the water depth is kept constant.
Spherical model: Pig. 4 shows the relation between the period Tw, Te
and the water quantity in the case of spherical model. Ty approaches the
Yalue of 0.1/@ as the water depth is higher. As for Tp» a similar relation
ship is observed as in the case of the cylindrical model,
Regarding to the damping in the case of cylindrical model, any damping
characteristics is hardly found for the Tet vibration of water, "The amount of
damping of the most exciting vibration of frame can not obtained precisely’,
because the wave shape of vibration becomes irregular due to tho effect of
water, especially its tendency is enhanced when the rigidity of frame 1s
lowered. ‘It seems, however, that as a gonoral tondeney tho lower the rigidity
of frame the larger the damping, and there Is no relation to the water quantity,
‘The damping coetficient is estimated approximately at | ~ 28 for h = 39cm,
2-2 for b= 64,5 om, and 3 ~ 6% for h= 80cm,
BY (2) Dead water and free water
From the most
ia assumed having t
formal
sxciting period of frame, quantity of doad water Wo which
ixing effect to the frame is caloulated by the following
w[bf-t]m
where
Jp * Period of main vibration of frame
Tia! Period of frame without water
Wat Weight of frame
Fig. 5 shows the relation between the ratio of the dead water quantity Wo of
the cylindrical model and the Jepth of water expressed by taking the tank diam-
ster as unity, togetior with the experimental formula, ‘The free water quantity
tends to converge to a definite value as the water quantity le increased. Ite
Value is estimated at I= 0,30 In caso of h~39 om and 64.5 cm, and 1 = 0.33D
In the case of h= 80 om, taking the depth from the water surface. Pig. 6
shows the relation between We/ily of the spherical model and the water quan=
{ity Vy together with the experimental Formula. ‘The ratio of dead water is
Increased with the increase of water quantity. ‘The values of We/y, colncide
well in case of f= 39 cm and 64.5 om, but in case of 80 em, the free water
‘Wantity is larger than the above two, which shows a simila® tendeney ao in the
case of cylindrical model.
(3) Period of free vibration, quantitios of dead water and free water in the
elevated water tanks of othor types
Now the results of experiments carried out in Prot. Muto's Laboratory
using the following 4 types of elevated water tanks are described.
Quadratic prism tank 200 (width) x 400 (depth)
Cylindrical tank with flat base 300 ¢
Gylindrical tank with the half spherical base 300
‘Thuncated conical tank slope of generator 3.3/1
‘The experimental methods were almost the same as the experiments deseribed
previously. Fig. 7 shows the relation between the free vibration period and
the water depth. The most exciting period of frame Tr is increased with the
water dopth in all tanks, whereas tho poriod of 1st mode vibration of water Tw
convorges to certain definite value in each tank, Denoting the tank diameter
by D, we obtain
Tw =0.11/D for prismatic tank
Tw =0.10/5 for eylindrical tank
Tw 0.1 for conical tank (Ty in see, D in om)
ow as Fig. 8 shows the relation of the quantity of dead water and free water. The
quantity of free water approaches a definite value as the quantity of water is
larger, Denoting the free water quantity by the depth Hpi
we have
H_'*0.3D for prismatic and cylindrical tank
Hp
