313

Proceedings: Fourth International Conference on Case Histories in Geotechnical Engineering, St. Louis, Missouri,
March 9-12, 1998.
A CASE STUDY ON SETTLEMENT OF OIL STORAGE TANK FOUNDATIONS
G.Ramasamy
University of Roorkee
Roorkee- 247 667
India
ABSTRACT
G. Kalaiseh'an
Bharat petroleum Corporation Ltd.
Noida-201301
India
1. 33
Foundations of8 Steel Oil storage tanks and two fire \Vater tanks were proportioned limiting total settlement to 100 mm. The soil at
the site consists of alternating layers of cohesive and cohcsionless soils. Settlement estimates were based on currently available
methods with suitable modifications to the situation met with. The tanks were load tested (Hydrotcst) and settlements observed at
nine locations along the periphery on tank shell base. These observed settlements arc compared with the estimated values.
KEYWORDS
OIL TANKS. LAYERED SOIL, SETTLEMENT. HYDROTEST. CASE STUDY
INTRODUCTION
Cylindrical storage tanks form a familiar part of petroleum
refineries. chemical plants and many other manufacturing
units. They hold large volumes of hazardous products.
Failure of such tanks can lead to severe environmental
damage, loss of human life and big financial losses. Literature
suggests that differential settlement has been a major cause of
distress in such tanks. Therefore. reliable estimation of
settlements constitutes an important step in design of
foundations of oil tanks.
Available methods on estimation of settlement are many. The
estimates vary quite significantly dqxnding on the method
adopted. This necessitates evaluation of prediction methods
through comparison of the estimated and observed values.
The present article deals \\-ith one such exercise carried out
with reference to foundations of a number of oil storo:1ge tanks
constructed in a tank fann at a oil depot in the Gangetic
plains of India.
DETAILS OF TANKS AND SITE CONDlTIONS
A tank farm (Fig. I) consisting of eight oil tanks form a part
of the oil depot. The tanks are of different diameters varying
from 9.0 m. To 17.0 m. and of height from 13.5 m. To 15.0
m. These tanks arc intended to store High Speed diesal
(HSD), superior kerosene oil (SKO) and motor sprit.
A detailed soil investigation was planned and executed to
map the soil strata. The investigations consisted of three
. . . '
. .. - ..
BORE HOLE -18m
o BORE HOLE- 9 m
f:1 OCPT
-12m
, ..
'' ~ - - - . . . : . . . - . : . .,_;- ~ - .:..:..; :....:.. ,.:.. . ,;:_:-. . : : . : ~ - ...:,._" ..
' . ,....,_ . : ...
SERVICE ROAD ' ' -.
, '
'. ,
'- .
1
5
0
()11
-'
160
812
,'
' '
,'
0
' .
17
013
' .
. .
' ..
. '
1s@
0
1
4
..
: I
' . . ..
Fig. I Tank Farm
Fourth International Conference on Case Histories in Geotechnical Engineering
Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu
boreholes and one dynamic cone penetrate test at each of the
tank locations. At the location of the largest tank (T -I). the
boreholes were made upto a depth of 18.0m. and at other
locations, they were made upto a depth of 9. Om. In each of
the boreholes, standard penetration test (SJYT) was conducted
at 1.5m depth intervals. Representative sample collected
through the SJYT sampler were used for classification tests.
Undislnrbcd samples collected in clay layers through thin-
walled samplers were used for shear and consolidation tests.
Based on the field and laboratory test data, bore logs were
prepared. It was observed that the subsoil conditions at the
site are more or less identical at all bore hole locations.
Accordingly. an average representative soil profile as shown
in Fig.2 was obtained for the site. It can be seen that the
subsoil consists alternating layers of clay and non plastic silt
of varying thickness upto 18.0m, the maximum depth of
exploration.
ELEVATION
(m) G. L
o ~ ~ ~
CL
-3-75 -i--W-l
ML- NONPLASTIC; N =12
-6-00-------------
CL; WL=25'1,;Wp=17'1,; N=9
-8-25---------------------
ML- NONPLASTIC N=18
-11 '50--------------
CL WL=30'1,;Wp=18'/,;N=19
-14-0o-------------------
M L /5 M ; N = 24
Jo/g. 2 Soil Profile
PROPORTIONING OF FOUNDATION
The oil tank foundations are generally proportioned based on
limiting total and/or differential settlement Marr et.al. ( 1982)
describe detrimental settlement patterns that a tank
foundation may develop and suggest that variable soil
thickness and I or compressibility over the plan area of the
tank foundation is the major of cause of foundation failures.
D'Orazio and Duncan ( 1987) have studied data pertaining to
26 oil tanks and suggest that a differential settlement of 0.5%
to 2.5% of tank diameter can be tolerated depending upon the
settlement pattern of the bottom plate. Chen eta! (1987)
compiled the work of a number of investigators and suggested
314
a set of criteria on limiting shell and bottom plate settlements.
Though these works suggest a much higher tolerable
settlement, the foundations in the present case are
proportioned for a tolerable total settlement of I OOmm of the
bottom plate.
ESTIMATION OF SETTLEMENT
The soil at the site is stratified consisting of layers of cohesive
and cohesionless soils. In such cases, the settlement is
calculated separately for each layer and then summed up to
get the total settlement.
Settlement of Cohesionless Soil Layers
A number of methods are available for the estimation of
settlement of cohesionless soil deposits of which. the methods
proposed by Peck ct.al. (1974). Burland and Burbidge (1985)
which are based on SJYT data and those proposed by De-Beer
and Martens (1957) and Schmcrtmann et.al (1978) which are
based on SCPT (Static Cone Penetration Test) are widely
accepted. These have been developed for homogeneous
deposits. In the present case, the method proposed by Peck
et. al ( 197 4) has been adopted with suitable modifications for
the layered system as below
The settlement of each layer is obtained using the equaiton,
q H
S ~ ------------------ X
where.
H-
D ~
0.044 X N X Cw D
settlement of the layer considered in mm
average corrected N value for the layer considered
load intensity at the tank base_ tlm
2
water table correction factor. taken as 0.5 for
submerged layers
thickness oflaycrs considered
Diameter of the tank
The cohesinless soil layers coming within a depth equal to
the diameter of the tank are considered in the computation.
Settlement of Cohesive Sojl Lavers
In the case of cohesive soil, a small part of total settlement
occurs upon application of the load and the major part
consists of the primary consolidation settlement. The
settlement. when estimated using c-log p cnrve, includes both
the immediate and consolidation settlement. In the present
case. the settlement of the cohesive layers is estimated using
the c-log p curves obtained from consolidation tests conducted
on undisturbed soil samples.
The total settlement computed as above exceeds 100 mm for
the anticipated load intensity of 15 tlm
2
. The topmost
Fourth International Conference on Case Histories in Geotechnical Engineering
Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu
cohesive layer contributes a major part of the lola! settlement.
Therefore, replacement of the top cohesive soil by a well
compacted gmnular material upto 2.0 m below ground level
for tanks of diameter more than 14.0m and, l.5m in the case
of smaller tanks, as shown in Fig.3 was proposed to limit the
bottom plate selllemcnt at the centre of the tank to I 00 mm.
It is known from elastic theory that in the case of a flexible
circular foundation resting on a clastic material. the
settlement at the edge of the foundation is equal to 70 percent
of that at the centre. Accordingly, it may be stated that the
foundations are proportioned in the present case limiting the
settlement of the shell base to 70 mm.
k::_ TANK
1-0m
cTANK BASE i:..:t GROUND
-'7;- ./ EMBANKMENT [LEVEL

'"' ""'
'"?"! STONE SAND MIXTURE
J7Sm
CLAY LAVER
NON PLASTIC SILT /SAND
Fig. 3 Proportioned tank foundation
OBSERVED SETTLEMENT
After the construction of the tanks. the performance of the
foundations were tested for full water load (llydrotcst). The
tanks were filled in stages. 114. 1/1, 112, 213. 3/4 and full
capacities and at each stage, the settlements were observed at
nine locations along the periphery on the tank shell base.
The settlements were observed 24 hours after loading at each
loading stage. Thus, it may be stated that the observed
settlements represent mainly the immediate (or elastic)
settlement and only a very little part of the consolidation
settlement. It is observed that settlements were more or less
the same at all locations at each of the loading stages. These
observed settlements are plotted in the form of load -
settlement curve in Fig.4 for the tanks of different diameters.
COMPARISON OF PREDICTED AND OBSERVED
SETTLEMENTS
As the settlements observed during hydrotest
represent the immediate (or elastic) settlement of the soil
strata, the same is compared with the corresponding
estimated values. T n the present case. the settlement of the
cohesionless layer constitute most of the immediate
settlement. At the design stage. this was estimated based on
SPT data using the Peck ct.al (1974) procedure. However. as
an exercise of back - analysis, the settlement of the
cohcsionless layers is also computed using De-Beer and
315
Martens ( 1955) method which is based on SCPT data. To
enable settlement computation by De-Beer and Martens
( 1955) method. the SPT values arc converted into equivalent
SCPT values using the correlation proposed by Peck et.al
(1974). These computed and the corresponding observed
settlements for various tanks are shown in Table I.
A comparison of the observed and estimated settlements show
that whereas Peck et.al ( 1974) procedure, as adopted here for
the layered SYStem underestimates settlement, De-Beer and
Martens (1955) method provides overestimation of
settlement. The possible immediate settlement of clay layers,
which is left out in the calculations, when estimated based on
a procedure suggested by D'orazio and Duncan (1987) on a
conservative basis, works out to less than 5 mm. Thus, there
is a significant difference in the estimated and observed
values. However, in view of the fact that the ground situations
for which the adopted settlement computation procedures are
valid and those actually met with, are not the same, the
observed and the estimated values can be considered to agree
satisfactorily.
E 10
E
.
)- 20
z
w
::;:
W30
-'
)-
)-
wL.o
lf)
LOAD IN TONNES
1000 1500 2000 2500 3000
Fi.g 4 Observed load- settlement tank foundations
CONCLUDING REMARKS
Foundations for a number of oil tanks resting on a deposit
consisting of alternating layers of cohesive and cohesionless
soils were proportioned limiting the total settlement to 100
mm. The tanks were load tested (Hydro test) and the observed
settlements arc compared with the estimated values. The
exercise suggests that estimation of settlement of foundations
in a real situation as the present one involves some logical
modifications to currently available methods and judicial
selection of soil parameters. The back analysis and
comparison of observed and estimated values have provided
Fourth International Conference on Case Histories in Geotechnical Engineering
Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu
valuable data base for judicious design decisions in
foundation work of similar nalurc in the area.
REFERENCES
1. Burland, J.B. and M.C. Burbidge. [19851. of'
fOundations on sand and graver, Proc. of Institution of
Engineers, London. Part I. 78, pp. 1325-1381.
2. Chen. H.M. Pan. CC. and S.T. Chung 11987].
"Settlement cnteria for large oil storage tanks", 9ll'
Southest Asian Geotechnical Conference, Bangkok,
Thailand. 4-41 to 4-52.
3. De Beer, E. and A Martens [1957]. "Method of'
computation an upper limit fhr the iJ?fluence qf
heterogenity C?{sand layers in the settlement of bridges".
Proc. 4"' Int. Coni on SMFE. London, Vol. I.
316
4. D'Orazio, T.B. and TB D'Orazio, [1984]. "Stability of
oil storage tanks", J. Geotech. Engg. Div .. ASCE, 1139,
(9), 967-983.
5. Marr. W.A. Romas J.A. and J.W. Lambe [1982].
"Criteria fOr seillement f?fianks", J. Gcotech. Engg. Div.
ASCE. 108 (8), 1017-1039.
6. Peck, R.B., Hansen, W.E.. and T.H. Thornburn, [1974].
"Foundation Engineering", John Wiley and Sons, Inc.,
New York.
7. Schmertmann. J.H., Hartmann, J.P. and P.R. Brown,
[1978]. "Improved strain it?fluence factor diagrams",
Journal of Gcolech. Engg .. ASCE, vol. 104, No. GE 8.
Table I - Comparison of Observed and Estimated Settlements
Estimated Settlement
TankDia. Load intcnsily, Observed Peck eta! De-Beer and
M tlm
2
settlement mm method Martens method
17.0 15.0
.;o
19.7 81
14.0 15.0 39 18.3 66
12.6 15.0 40 20.0 58
9.0 13.5 26 12.8 40
Fourth International Conference on Case Histories in Geotechnical Engineering
Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu