The Behavior of Existing Pile Foundation Against Buckling Instability in Liquefiable Soil

(A Case Study : Coal Fired Power Station Tanjung Jati B, Jepara, Central Java, Indonesia)

Rini Kusumawardani1, Muhammad Hamzah Fansuri2, Priya Purwanta3

1Head of Soil Mechanics Laboratory, Universitas Negeri Semarang, Indonesia

2Planning Engineer, Mitsui Engineering and Shipbuilding. Co. Ltd., Indonesia
3Planning Manager, Mitsui Engineering and Shipbuilding. Co. Ltd., Indonesia

Abstract
The liquefaction phenomenon normally occurs on the saturated non-cohesive soil (silt to sand). In
this condition, the effective soil stress reduces significantly due to dynamic load, such as earthquake
loading or pillar driving, subjected on it for interval time duration. Some evidences of liquefaction
phenomen are found in coastal area of Jepara regency in which it is built a Coal Fired Power Station
of Tanjung Jati B unit 5 and 6.
Foundation pillar which is buried in soil depth has a great potential of destruction when a dynamic
loading is subjected. Thus, it could provoke a failure construction in the study area due to pile
bending or pile buckling. Unstability of pillar buckling should be considered as it could be treated as
a column loaded with lateral loads. An identification method of pile structure in liquified soil is
achieved by calculating the safety factor of structure.
By using high-risk earthquake on soft soil, the analysis results of liquefaction safety factor showed
that the critical depth of the pillar foundation is less than axial loading at the potential liquefaction
depth.
The liquefaction safety {g(x)} with high-risk earthquake on non-cohesive soil is 4.87, so the g(x) is
more than 0. Coal Fired Power Station of Tanjung Jati unit 5 and 6 can be indicated safe in which
the critical depth of the foundation pillar is lower than the one on the axial burden in depth
potentially causing liquefaction.
Keywords: Buckling, Bending Pile, Liquefaction

A. Introduction earthquake at 9.22o -109.58 o east with a depth of 55

Indonesia lies on the earthquake – prone km on August 19th, 2004 and 5.5 SR earthquake with a
area because its location is situated at the confluence of depth of 33 km on July 19th, 2005.
four major tectonic plates of the earth which The earthquake potential in Jepara regency occurred
are Eurasian, Indo-Australian, Pacific and Philippine. on December 25th, 1821, when the earthquake in
Central Java Province as one of the areas adjacent Jepara reached the intensity of the earthquake on the
to the plate collision zone is prone to tectonic scale of MMI ( Modified Mercalli Intensity ) VI-VII.
earthquakes. Some earthquakes with a magnitude of Therefore, every development planning in Indonesia
scale 5 SR occurred in the past 25 years in various needs to take into account the risks caused by the
regions of Central Java, although there were no earthquake. Such risks not only include the risk of
casualties and significant damage. Some of these failure on the building structure but also the risk
occurrences were an earthquake with a strength of 6 SR of a failure that could occur on the supporting structure
at 7.20 o south - 09.30 o east with the depth of 33 km on of the building above it.
March 14th; 1981, 6.6 SR earthquake with a depth of The collapse of the supporting structures of each
106 km on June 9th, 1992; 6.2 SR earthquake at pole is observed in potentially liquefied soil after the
8.62 o south / 110.11 o east on May 25th, 2001, 6.3 SR occurrence of the earthquake.
For example the collapsed building in Figure 1 (a)
Contact of Author: Rini Kusumawardani, Soil Mechanics and 1 (b). In both cases, the building must be
Laboratory, Universitas Negeri Semarang destroyed.
Phone: +62818270087 The pattern of the overall failure of this building may
e-mail:[email protected] seem similar in which both are highly skewed. The
building in Figure 1 (a) is located in a lateral spread
where the dock wall near the building with a lateral performance analysis during the 1987 Edgecumbe
distance of more than 1 meter. earthquake in New Zealand. By detail, the movement
pile phenomenon due to of earthquake loading is
illustrated in Figure 2 (JRA, 1996).

(a)

(b) Fig. 2 Understanding the incidence of pile failure (Finn

and Thavaraj, 2001).

JRA (1996) recommends that the design of

foundation structures as illustrated in Figure 3. Figure
3 suggests that designing bending failure resistance can
be achieved by assuming non-liquefaction soil
produces passive earth pressure on the pile foundation
and soil liquefaction with the amount of 30% of the
overburden pressure. In addition, it is assumed that
the bending failure does not take into account inertia
load, the effect of lateral spread and inertia
(Ishihara, 1997; CEN, 2004; BSSC, 2000).

Fig. 1 (a) The collapse of a pile foundation caused by

the Kobe earthquake; (b) Failure of pile foundation
at surface soil level in 1995 caused by Kobe
earthquake (Tokimatsu).

The condition of foundation failure

as shown in Figure 1 was suspected as a result of
soil liquefaction phenomena. The soil
liquefaction phenomenon was caused when the soil
tends to lose its shear strength (kusumawardani, 2016).
Fig. 3 The ideal pile design at JRA, 1996 (qNL denoted
As a consequence, the elements of the soil which lied earth passive pressure and 30% overburden
on top experienced movement that could push the pile pressure).
position which then caused non-liquefaction
movement of the crust above on its own. This Collapsing occurrence is the instance of lateral
occurrence was the cause of the bending failure on the instability under axial pressure. This phenomenon
pile structure (Hadama & O'Rourke, 1992; Finn and occurs as a result of the final carrying capacity of the
Fujita, 2002 , Hamada, 2000 ). It is also corroborated pile foundation which passes through the solid soil to
by Berril et al. (2001) who performed bridge runway the medium solid layer and the soil around the pile is
experiencing liquefaction when
the earthquake occurs (Hattacharya el al, 2005, The actual load failure (P failure ) has several factors ,
Ψ (Ψ <1) . Based on equation (3) above, the instability
Knappett and Madabhushi, 2006). The stress on the
can be estimated at axial loads of 0.35 Pcr. Therefore,
pile is initially elastic and bending length is equal to the the value of Ψ is 0.35. This factor is affected
entire length of the pile in the liquefied ground. The by axial load ( P dynamic ), failure , and residual
resulting lateral is caused by deflection movement, stress on pole stake caused because
inertia or the decrease in bending capacity which of the piling process . It can be seen on figure 4.
shortly increases the lateral deflection, and leads to the
increasing plasticity of the pile and the decrease of the
collapsing load thereby increasing the failure. This
lateral load, however, is a secondary requirement for
the pile foundation in the liquefied soil and should be
checked against Euler's collapse. In other words, part
of the pile in the liquefied soil needs to be assumed as
non- clamped structural columns.
This article discusses the collapsing behaviour on
piles that experience liquefaction. As a unity of
structural construction, the pile foundation should not
be collapsed even if the soil experiences liquefaction.
Fig. 4 the buckling amplification factors to normal
B. Literature Review axial loads.
The inclination of the building in Figure 1 makes
the authors review for the analysis of collapsing factors The critical length (HC) can be achieved by looking at
due to liquefied soil. The thing to note is the increase in the condition of failure of the pile
pore water stress when an earthquake occurs by assuming P dynamic = P failure, as well as Leff = HC. The
(Battachary aet al., 2005). When earthquake vibration critical length can be calculated using this following
begins to occur, then the pore water stress gradually equation (4),
increases on the soil layer on which liquefaction is 3.45𝐸𝐼
HC = √ (4)
indicated. This increase in pore water stress causes the 𝑃𝑑𝑦𝑛𝑎𝑚𝑖𝑐
soil to experience zero effective stress. Therefore, the For the pile stability factor with working axial load can
soil does not have the ability to support the load. As a be calculated using equation (5) which is a function of
result of these events, the frictional force between the the relationship between liquefaction depth (DL) and
soil and the pile decreases drastically and this is what critical depth (HC).
causes the collapsing of the pole. g(X) = HC – DL (5)
Some of the parameters that need to be considered Information;
in the analysis of these collapsing factors HC = Critical depth
include static axial load, collapsing load of the pile, DL = Unbraced length, calculation of liquefied
and the damage load on the pole and limit function to a depth added by the depth of foundation.
stable state. In the analysis of static axial load (P static), X = The vector represents the set of variables
it is assumed that each pile is burdened with the same that governs the stability of the pile
load. This assumption ignores the load eccentricity for foundation at the site of the designed
loads in a static condition. The dynamic load equation earthquake.
(1) is obtained by multiplying dynamic axial load X variables can be classified into 4 categories with
factor (Pdynamic) respect to earthquake characteristic factor, pile
Pdynamic = (1 + α) x Pstatic (1) characteristics, soil data and surrounding
For analysis of buckling pole ( Pcr ), equation (2) is conditions, type of structure. Function limit when a
used (2) pile is considered as unstable or fail is g(X)<0 and
𝐸𝐼 g (X)> 0 is the indication of a safe condition.
Pcr = 𝜋 2 𝐿2 𝑒𝑓𝑓 (2) Parameter g(X) = indication of limit condition
Where L eff is the Euler's buckling length by assuming
a clamped pile at both ends ( Bhattacharya
et. a l ., 2004, 2005). In addition, pile failure at one end
caused by the liquefaction phenomena can be
calculated by the following equation (3)
Pfailure = Ψ x Pcr (3)
C. Case Study with radius 100 km from project location at Tanjung
The location of Tanjung Jati Coal – Fired Power Jati B Unit 5 & 6 Jepara.
Plant B building units 5 & 6 is administratively located Further study about the potential of liquefaction
in Tubanan Village, Kembang District, Jepara with calculation method that have been implemented
Regency, Central Java Province, ± 32 km in North earlier researchers is required. Test results and sample
Jepara. soil showed that there were 9 locations that had the
The results of the identification and classification of liquefaction potential. The axial load (Pstatic) of the
the potential liquefaction showed that the dominant soil building was 1700 kN. Plan foundation can be seen in
was that alluvial at the river and the old river. The figure 6.
coastal sediment was brought to this area through small
rivers flowing around the project site. The top deposit
is comprised of a very soft sublayer to soft clay. The
results of soil sampling can be seen in figure 5.

(a)

Fig. 6 Plan foundation Boiler #6 Tanjung Jati B Unit 5

& 6.

The building design can be seen in Table 1 and the

pile table design can be seen Table 2.

Tabel 1. Building Data Design

Building Height 26 m above Ground
Level
Building Dimensions 94.5 m x 79.5 m
Foundation Type MAT Foundation
Axial Load on Each 1700 kN
Pile
(b) Dead Load of the 1899300 kN
Fig. 5 (a), (b) the results of N-SPT sampling data. Building

The results of the liquefaction analysis based on Tabel 2. Pile Design Data
soil testing at 12 boreholes and 16 CPTU test identified Length 22 m
that the site is susceptible to liquefaction. Based on the External Diamater 600 mm
soil data susceptible to liquefaction, the normal Internal Diameter 400 mm
classification based on uniform gradation and fine sand Material Prestressed Concrete
with the average diameter size (D50) is 0.1 to 0.2mm.
Young’s Modulus (E) 33892181 kN/m2
The other factor which triggers mechanism of the
Bending stiffness of 173022.57 kN.m2
liquefaction in soil is a strong earthquake with a
the pile (EI)
minimum magnitude M = 7. The norm for disaster risk
liquefaction can be calculated with magnitude 7.5.
USGS Map of earthquake showed that there were 1. Dynamic Axial Load Factor Analysis
1.1 The initial of the earthquake before liquefaction
no earthquake with M> 6 in Central Java area from the
The occurrence of earthquake caused by the shear
last 100 year with period from year 1900 until 2012,
force in buildings is usually estimated by lateral force
calculation procedure (BSSC, 2000; BIS 2000; CEN, µk = the modulus subgrade reaction, (MN / m 3 )
2004) according to Table 3.
VB = Cs x W (6) Tabel 3. Value of Subgrade Modulus Reaction
when; µk Value (MN/m3)
Relative
Z = Maximum earthquake zone consideration Sand under Sand on the
Density
factor water water
I = Structural factors 40% 8 13
R = response factor reduction 60% 24 42
Sa/g = Average of acceleration response coefficient 80% 40 75
The rigidity value was 1.3T. The average value of
The approach of the natural period of vibration (Ta) in N-SPT was 4.51. Therefore, based on the value, the
seconds is formulated in this following equation (7) soil can be classified as soft soil (Terzagi and Peck
(BIS, 2002) 1967).
Ta = 0.09h/√𝑑 (7) Relative density value (RD) can be calculated using
the equation (12) below. The calculation of empirical
h is the height of the building, in meters and d is correlation between N-SPT and effective vertical stress
the thickness of the basic dimension of the building, in and % relative density, Mayerhof (1957);
meters.
𝑁
The results of the calculation showed that period RD = 21 √𝜎′ 𝑣+0.7 (12)
(Ta) was equal to 1. 25 seconds. Then the spectral
acceleration for the period difference can be calculated when;
in this following equation (8); N = N -SPT value
1+15Ta ; 0.0 ≤ Ta ≤ 0.10 (8) 𝜎’v = effective vertical stress, kgf/cm2
Sa/g ; 0.10 ≤ Ta ≤ 0.67
1.67/Ta ; 0.67 ≤ Ta ≤ 4.0 The calculation showed that the relative density
was 43.05% and subgrade reaction modulus was 42
Ta is the period of the building. The calculation MN/m3 (above water level). The stiffness factor of the
resulted in an earthquake coefficient (Cs) with this above equation resulted in a value of 1.32 m.
following equation (9); The depth of the mat foundation was 3.5 m, then the
Cs = (Sa x Ie)/R (9) moment rocking value (MR) can be calculated by the
when; equation (13) as follows;
Sa = Spectrum response design MR = VB x Fi (13)
Ie = The primacy factor
R = The coefficient of response modification when;
Fi = The depth of the foundation (meter)
The calculation resulted in the value of Cs of 0.10
(Tokimatsu and Asaka, 1998). Horizontal earthquake Therefore, the value of MR was 686563.12 kN.m
force (VB) was obtained using this following equation Half of the total supporting pile (n) was
(10); 650 piles. The compression axial load value acting on
VB = Cs x W (10) each pile was obtained according to the following
when equation (14);
W = the combination of reduced dead load and Pinc = MR / (S x n) (14)
live load (kN).
VB value of 196160.89 kN was obtained through when;
the calculation. Davisson and Robisson’s (1965) S = the distance between piles (m)
method approach for lateral pile load analysis was N = 0.5 . pile total
used. In the procedure, the lateral load was assumed as Pinc = 686563.12 / (1.8 x 650)
the anchorage point in the soil and the depth depended = 586.806 kN.
on relative stiffness between soil and pile. This method
was used to calculate the stiffness factor (T) for the Therefore, the calculation showed that the increase
combination of pile and soil, with equation (11) as of the axial load value was 35%.
follows;
1.2 After Liquefaction
5 𝐸𝐼
T= √ (11) It has been discussed earlier, in full liquefaction, the
𝜇𝑘
base shear decreased because of the earthquake
excitation. Estimated reduction of the base shear was
when, required to predict the period vibration on the building
EI = Pile stiffness for a small vibration when the soil around the pile has
liquefied. Considering the foundation depth of 3.5m
below the water level and the depth of the pile (L) was 3. The Buckling Load Safety Factor
19 m below ground level. The rigidity of the The length of the unbraced pile (DL) was 12m,
building (k) for small vibration in the building with the corresponding to the equation DL < HC. It can be
condition of the full liquefaction was calculated with concluded that the building was safe from the failure of
equation (15) as follows; resisting the buckling instability. The buckling load
K = 38 x (12EI / L3) (15) safety factor can be formulated with the ratio of HC and
= 11.502 MN/m DL as follows;
The buckling load safety factor corresponds to this
This assumption showed the stiffness of the initial following equation (18),
pile contribution for the total stiffness on the pile SF = HC/DL (18)
system. This assumption was confirmed by the elastic = 1.4
stiffness of soil liquefaction which was calculated by
the earthquake magnitude which was less than the The mathematical equation resulted in the stability
stiffness of the concrete pile. of the pile with this following function g(X) =
The time period of the building during the full HC - DL > 0 which can be indicated as
liquefaction can be calculated by the equation (16) as safe. The calculation above showed that the pile was
follows; strong against shear and safe from liquefaction danger.
𝑀
Ta. Liq = 2𝜋 𝑘 (16) 4. Conclusion
= 25.77 s The calculation method of the pile design based on
the moment mechanism in which the lateral load was
when, caused by inertia and slope displacement causing
M = the weight of the building (kN) moment damage on pile was performed in this study.
K = the value of building rigidity The calculation which has been performed indicated
that there was no indication of buckling moment
The above calculation shows the time period of resistance at the time of liquefaction.
vibration increased approximately 27 times. The A simpler method used to evaluate in this study was
selection of the response modification coefficient (R) the safety factor equation (18). This method checked
with seismic force resisting system of the moment – for the foundation stability due to the resistance
bearing concrete frame resulted in the factor value of 8. instability in case of liquefaction, when the stiffness of
Therefore, the base shear (VB') could decrease with the soil around the pile was weakened, there were two
value of 24520.1 kN. parameters, the critical depth (HC) and the unbraced
The increasing the axial load value of each pile pile length caused by the calculated liquefaction (D L ).
could be calculated by the following equation (17); This case study has considered illustrating the
Pinc = (VB’ x L)/(S x n) (17) application of the methodological objectives.
= 398.19 kN
5. References
This increased the axial load value by 23% and Berrill, J.B., Christensen, S.A., Keenan, R.P.,
the base shear decreased by 8 times. From the above Okada, W. and Pettinga, J.R. (2001). “Case Studies of
calculation, the axial load may approximately increase Lateral Spreading Forces on a Piled Foundation”,
by 40% during the liquefaction stage. Then, the α value Géotechnique, Vol. 51, No. 6, pp. 501-517.
of 0.2 can be used. BSSC (2000). “The 2000 NEHRP Recommended
Provisions for New Buildings and Other Structures,
2. Critical pile depth estimation Part 2: Commentary (FEMA 369)”, Building Seismic
Table 4 below illustrated about critical pile depth Safety Council, National Institute of Building
estimation. Sciences, Washington, DC, U.S.A.
CEN (2004). “Eurocode 8: Design of Structures
Table 4. Critical pile depth estimation for structure for Earthquake Resistance – Part 5: Foundations,
Retaining Structures and Geotechnical Aspects”, EN
Parameter Value Description
1998-5:2004, Comité Européen de Normalisation,
Pstatic 1700 kN Following the Project of
Brussels, Belgium.
Tanjung Jati B Power Unit
Davisson, M.T. and Robinson, K.E. (1965).
5&6
“Bending and Buckling of Partially Embedded Piles”,
Pdynamic 2098 kN Following the dynamic
Proceedings of the Sixth International Conference on
axial load factor Soil Mechanics and Foundation Engineering,
equation on the full Montreal, Canada, Vol. 2, pp. 243-246.
liquefaction estimated at
Finn, W.D.L. and Fujita, N. (2002). “Piles in
0.2
Liquefiable Soils: Seismic Analysis and Design
HC 16.87m Following the equation (4)
Issues”, Soil Dynamics and Earthquake Engineering,
Vol. 22, No. 9, pp. 731-742.
Hamada, M. (2000). “Performances of
Foundations against Liquefaction-Induced Permanent
Ground Displacements”, Proceedings of the 12th
World Conference on Earthquake Engineering,
Auckland, New Zealand, Paper No. 1754 (on CD).
Hamada, M. and O'Rourke, T.D. (editors) (1992).
“Case Studies of Liquefaction and Lifeline
Ishihara, K. (1997). “Geotechnical Aspects of the
1995 Kobe Earthquake”, Proceedings of the 14th
International Conference on Soil Mechanics and
Foundation Engineering, Hamburg, Germany, pp.
2047-2073.
JRA (1996). “Design Specifications of Highway
Bridges. Part V: Seismic Design”, Japan Road
Association, Tokyo, Japan.
K. Kertapati Engkong & Eka Adi Saputra,
Sukahar. Seminars and intellectual property
right. Indonesia Construction Progress. Seismotectonic
Study of Muria Region for Eligibility and the Safety of
the Planned Electricity Center of Data Reactor of
PLRD-Ujung Lemah Abang-Ula-Jepara-Central Java.
Kusumawardani, R., Suryolelono, K.B., Suhendro,
B. And Rifa’i, A. 2016. The dynamic response of
unsaturated cleand sand at very low frequency.
International Journal of Technology, Vol. 1.
Performance during Past Earthquakes, Volume 1,
Japanese Case Studies”, Technical Report
NCEER92-0001, State University of New York at
Buffalo, Buffalo, U.S.A.
Tokimatsu, K. and Asaka, Y. (1998). “Effects of
Liquefaction-Induced Ground Displacements on Pile
Performance in the 1995 Hyogoken-Nambu
Earthquake”, Soils and Foundations, Special Issue on
Geotechnical Aspects of the January 17, 1995
Hyogoken-Nambu Earthquake, No. 2, pp. 163-177.