Settlement and load sharing of a piled raft with ground improvement on soft ground

Tassement et partage de charges d’une fondation mixte radier-pieux dans un sol meuble amélioré

K. Yamashita & T. Yamada

R&D Institute, Takenaka Corporation, Chiba, Japan

ABSTRACT
This paper offers a case history of design and performance of a seven-story building of a piled raft on loose sand underlain by soft silt
layers. In order to reduce consolidation settlement of the soft silt and to cope with the liquefiable loose sand, a piled raft combined
with grid-form ground improvement was adopted. The design of the foundation was found appropriate by field measurements which
concerned settlement, axial forces of the piles and earth pressures and pore-water pressure beneath the raft from the beginning of
construction to the time about four years after the end of construction.
RÉSUMÉ
Cet article présente la conception et le comportement d’une fondation mixte radier-pieux d’un bâtiment à sept niveaux. Cette
fondation est construite dans une couche de sable lâche reposant sur des couches de limon. Ce type de fondation a été adopté en
combinaison avec l’amélioration, en forme de quadrillage, d’une partie du sol meuble et ce, afin de réduire le tassement de la couche
de limon et contenir les effets induits par la couche de sable lâche à fort potentiel de liquéfaction. L’article discute le comportement de
cette fondation sur la base d’une série de mesures sur site, qui se poursuivent pour la quatrième année depuis le début de la
construction du bâtiment. Le dimensionnement de cette fondation est estimé convenable considérant les mesures de tassement, des
forces axiales sur pieux, des contraintes sur le sol et de la pression de l’eau interstitielle sous le radier.

Keywords : piled raft, ground improvement, settlement, load sharing, monitoring

1 INTRODUCTION pore-water pressure beneath the raft from the beginning of

construction to the time about four years after the end of
In recent years there has been an increasing recognition that the construction.
use of piles to reduce raft settlements can lead to considerable
economy without compromising the safety and performance of
the foundation (Poulos, 2001). In Japan piled rafts have been
used for more than 80 buildings since it was first applied to the
four-story building in Urawa (Yamashita & Kakurai, 1991;
Yamashita et al., 2008). However there exist not so many case
histories on monitoring load sharing between raft and piles as
well as settlement. Thus accumulation of field evidences by
monitoring full-scale structures is required to develop more
reliable design method (Mandolini et al., 2005).
This paper offers a case history of design and performance of
a piled raft on loose sand underlain by soft silt layers. In order
to reduce consolidation settlement of the soft silt and to cope
with the liquefiable loose sand, a piled raft combined with grid- Photo 1. Seven-story office building in Minamisuna.
form ground improvement was adopted. To confirm validity of
the foundation design, field measurements were performed on
the settlement, axial forces of the piles, earth pressures and
9 CXGXGNQEKV[
526 7 PFTCKPGFUJGCT % QPUQNKFCVKQP[KGNF 2 Y CXG8 2 O U
0 8 CNWG UVTGPIVJSW M2 C UVTGUU M2 C 5 Y CXG8 5 O U
㨙

    ‫ ޓ‬  㨙

).r (KNN
5KNV
5CPF & KPKPIJCNN
 ) TKFHQTO UQKNEGO GPVY CNNU
Upper
5KNV 㨙 㨙
 㨙 㨙
Lower
 㨙
5KNV 㨙
& GRVJ O

85 82 㨙
㨙 2*%2KNGU

5CPF[
5KNV
'HHGEVKXGQXGTDWTFGP 㨙 㨙
 5CPF RTGUUWTG‫ޓ‬ ’
‫ޓ‬ 5VTCKPICWIGU

 5KNV
5CPF O
5CPFCPF 5GVVNGOGPVICWIGU
 ITCXGN
Figure 1. Schematic view of the building and foundation with soil profile.

Proceedings of the 17th International Conference on Soil Mechanics and Geotechnical Engineering 1236
M. Hamza et al. (Eds.)
© 2009 IOS Press.
doi:10.3233/978-1-60750-031-5-1236
 K. Yamashita and T. Yamada / Settlement and Load Sharing of a Piled Raft 1237

2 GPGVTCVKQP 6 KRVTGCVO GPV 9 KVJFTCY KPI ǾO O

+O RTQXGFUQKN
EQNWO PU O O
࡮& KUEJCTIGQH
UQKNFKHKGFO CVGTKCN
࡮5VKTTKPICPFO KZKPI % QWRNGQHUQKNEQNWO PUNCRRGFCPF
 EQPVKPWQWUN[CTTCPIGF


.KSWGHKCDNGNC[GT

+O RTQXGFUQKN
EQNWO PU

(a) Construction procedure (b) Grid-form soil-cement walls

Figure 2. Grid-form ground improvement. Photo 2. Grid-form soil-cement walls at the foundation level.

Grid-form soil-cement walls Office building Dining hall

٨ ٧٧ ٨ ٨ ٤ ٤ ٨ ٨ ٤ ٤ ٨ ٨ ٤ ٤ ٨ ٨ ٧ ٧ Pile‫ޓ‬diameter
٨
٧ d=0.9m ٤ d=0.8m
٤٤

٤
٤ ٨ ٤ ٤ ٤ ٤ ٤ ٤ ٤ ٨ ٨ ٤
٤
٤ d=0.7m ٨ d=0.6m
Tributary area
‫ڏ‬
٤٤

٤
٤ ٨ ٨ ٤ ٤ ٤ ٤
٤ ٤ ٤ ٨ ٨ ٤
Monitoring devices
E1 E2 ٤
W ٨ ٤ ٤ Instrumented pile
٤
٤
٤ ٨ ٨ ٤ ٤ ٨ ٨ ٤ ٤ ٨ ٨ ٤ ٤ ٨ ٨ ٧ ٧ ٨
‫غ‬ Earth pressure cell
ٌ Piezometer
‫ ڏ‬Settlement gauges
0

Figure 3. Layout of piles and grid-form soil-cement walls with locations of monitoring devices.

2 BUILDING AND SOIL CONDITION than 200 mm of consolidation settlement was predicted in the
soft silt layers down to a depth of 18 m. In order to reduce the
The seven-story office building of 29.4 m in height above the consolidation settlement and to ensure the differential
ground surface with a flat dining hall is located in Minamisuna, settlements being below a tolerable amount, a piled raft was
Tokyo (Photo 1). The building is a steel-frame structure and a proposed. The piled raft was designed based on a design
schematic view of the building and foundation with soil profile philosophy of “creep piling”, i.e. sufficient piles are included to
is shown in Fig. 1. The foundation levels are at depths of 2.2 m reduce the net contact pressure between raft and soil to below
in central part and 1.6 m in both ends below the ground surface. the consolidation yield stress of the clay (Hansbo, 1984;
The subsoil consists of an alluvial stratum to a depth of 46 m, Jendeby, 1986). The allowable bearing capacity of each pile at a
underlain by a diluvial sandy layer of SPT N-values of 50 or working condition was determined to be sufficiently larger than
more. The soil profile down to a depth of 11 m is made of soft the load which should be carried by the pile based on the “creep
silt and loose sand. Between depths of 11 m to 42 m below the piling”. At the same time the design load of each pile should be
ground surface, there lies a thick soft to medium silt stratum. smaller than the load at which significant creep starts to occur,
The upper silt layer between depths of 11 m to 18 m is slightly at about 70 % of the ultimate bearing capacity. The piles were
overconsolidated with an overconsolidation ratio (OCR) of embedded in the lower medium silt layer enough to ensure the
about 1.1. The lower silt layer between depths of 18 m to 42 m frictional resistance. Consequently a piled raft with the grid-
is overconsolidated with an OCR of 1.8 or more. form ground improvement was adopted which has a total
number of 70 piles of 30 m in length, 0.6 m to 0.9 m in diameter.
The piles are pretensioned spun high-strength concrete (PHC)
3 FOUNDATION DESIGN piles and were constructed by inserting a set of 15 m-long PHC
piles into a pre-augered borehole filled with mixed-in-place soil
3.1 Ground improvement cement to avoid noise and vibration. Figure 3 shows a layout of
the piles and the grid-form soil-cement walls.
It appeared that the loose sand between depths of 6m to 11 m
has a potential of liquefaction during earthquakes of the
maximum horizontal ground-surface acceleration over 200 Gal, 4 INSTRUMENTATION
according to the simplified method (Tokimatsu and Yoshimi,
1983). To cope with the liquefiable sand layer, a grid-form The locations of the monitoring devices are shown in Fig.3.
ground improvement was introduced. The grid-form soil- Two piles, 7A and 7B, were installed with a couple of LVDT-
cement walls are constructed by deep mixing method as type strain gauges at pile head (at a depth of 4.3m). The pile 7B
illustrated in Fig. 2 and the high-modulus soil-cement walls was installed with the other couples of strain gauges at depths of
confine loose sand so as not to cause excessive shear 15.3 m and 31.3 m. Earth pressure cells were installed at depths
deformation in the loose sand during earthquakes (Photo 2). The of 2.2 m (E1) and 1.6 m (E2) and a piezometer (W) was
effectiveness of the grid-form ground improvement has become installed at a depth of 2.2 m beneath the raft. The vertical
evident at the 1995 Kobe earthquake (Tokimatsu et al., 1996). ground displacements below the raft were measured by
settlement gauges. The settlements of the foundation were
3.2 Pile specification measured by an optical level. The reference point was set to an
existing nearby building founded on end-bearing piles. The
A total load in structural design is 378 MN which corresponds measurement of the axial forces of the piles, earth pressures and
to the sum of dead load and live load of the building. The pore-water pressure beneath the raft and the vertical ground
average contact pressure over the raft is 100 kPa with the local displacements started early in December 2003, at the time just
maximum of 142 kPa. If a raft foundation alone was used, more before constructing steel reinforcement of the foundation slab.
1238 K. Yamashita and T. Yamada / Settlement and Load Sharing of a Piled Raft
GD\
         
ٕ ٕ ٕ
  ( QWPFCVKQPUNCD
& KURNCEGO GPV㧔O O 㧕
[GCTUCHVGTGQE 9HUW LFDOJURXQGGLVSODFHPHQW 웉PP웊
'PFQHEQPUVTWEVKQP
     
     
 㨙
& GRVJ㧦O

O
  㨙
O

'HSW K웉P웊
 㨙




Figure 4. Measured vertical ground displacements. & GE
& GE
& GE
 & GE
The measurement of the foundation settlements began at the 㨙 & GE
& GE
time just after concrete casting of the foundation slab. Figure 5. Measured
 vertical ground displacements with depth.


Ԙ 
ԙ 
Ԛ 
ԛ 
Ԝ 
ԝ 
Ԟ 
ԟ 
Ԡ 
ԡ 
Ԣ 
ԣ
5 OBSERVATIONS 

5GVVNGO GPV O O

5.1 Settlement of the foundation


The building completed in mid-November 2004 and started in 

operation late in November. In this paper the measurement in 

Dec.2, 2004 are referred to as those “at the end of construction”. 6WUHHW 웢웛
6WUHHW 웥웛
'HF
'HF
2FW
2FW
'HF
'HF
'HF
'HF
Figure 4 shows the measured vertical ground displacements
below the raft at depths of 3.0 m, 11.6 m and 20.4 m relative to (a) Along the streets A and D
a reference point at a depth of 46 m from the ground surface. 
Ԙ 
ԙ 
Ԛ 
ԛ 
Ԝ 
ԝ 
Ԟ 
ԟ 
Ԡ 
ԡ 
Ԣ 
ԣP
The ground displacement at a depth of 3.0 m amounted to 15.6 

mm at the end of construction. After that the displacement 5 GVVNGO GPV O O


slightly increased and reached 19.4 mm in Dec.13, 2008, at the

time about four years after the end of construction. Figure 5
shows the distributions of the measured vertical ground 

displacements with depth. At the time just before casting of the 

foundation slab (Dec.26, 2003), heaving of the ground due to 6WUHHW 웣웛
6WUHHW 웤웛
'HF
'HF
2FW
2FW
'HF
'HF
'HF
'HF
the excavation amounted to 2.6 mm at a depth of 3.0 m.
(b) Along the streets B and C
Considering the heaving the vertical ground displacement at a
depth of 3.0 m amounted to 18.2 mm at the end of construction Figure 6. Measured longitudinal settlement profiles.
and reached 22.0 mm in Dec.13, 2008. Figure 6 shows the
longitudinal settlement profile of the foundation measured by an (QWPFCVKQPUNCD 'PFQHEQPUVTWEVKQP [GCTUCHVGTGQE
optical level. The measured settlements were 14-24 mm at the 
ٕ ٕ ٕ

end of construction. The settlements slightly increased to 17-31 & GRVJ㧦

O
QTEG㧔/ 0 㧕


mm in Dec.13, 2008 and the maximum inclination angle of the 
foundation amounted to 1/1200 radian between the columns 
11B and 12B which is less than the serviceability limit of
# ZKCN㨒


1/1000 radian in structural design. 

    
5.2 Pile load, earth pressures and pore-water pressure  
         
GD\
Figure 7 shows the measured axial forces of the piles 7A and (a) Pile 7A
7B. The measured values at pile head increased after the end of (QWPFCVKQPUNCD 'PFQHEQPUVTWEVKQP [GCTUCHVGTGQE
construction and reached a state of equilibrium on the pile 7A, 
ٕ ٕ ٕ

but still slightly increased on the pile 7B at the time about four & GRVJ㧦
O
# ZKCNHQTEG㧔/ 0 㧕


years after the end of construction. Figure 8 shows the 
O

distributions of the measured axial forces on the pile 7B. At the 
end of construction, the average skin friction between depths of 
15.3 m to 31.3 m through the layers of soft to medium silt was  O
90 kPa whereas the value between depths of 4.3 m to 15.3 m 
through the layers of soft silt and loose sand was 23 kPa. The  
    

average skin friction in the lower part of the pile is consistent          
with the average undrained shear strength of the silt (81 kPa) (b) Pile 7B
GD\

$[LDOIRUFH 01 Figure 7. Measured axial forces of the piles 7A and 7B.

      

㨙 (QWPFCVKQPUNCD 'PFQHEQPUVTWEVKQP [GCTUCHVGTGQE
ٕ ٕ ٕ

2 TGUUWTG㧔M2C㧕

 
'HSW KᄼP

㨙  '

 '


& GE 9
& GE 
& GE     
 㨙 & GE  
& GE          
GD\
Figure 8. Measured axial force distribution on the pile 7B. Figure 9. Measured earth pressures and pore-water pressure.
 K. Yamashita and T. Yamada / Settlement and Load Sharing of a Piled Raft 1239

between depths of 15.3 m to 31.3 m. After that the skin friction

in the lower part considerably increased and reached 123 kPa in
Dec.13 2008, which is 52% larger than the average undrained
shear strength. The skin friction in the upper part slightly
increased after the end of construction. Figure 9 shows
development of the measured earth pressures and pore-water
pressure beneath the raft. The earth pressures seemed to reach
constant values in early stage of construction despite of the
successive increase in construction loading.

5.3 Load sharing between raft and piles

For piled rafts the equilibrium equation is expressed as follows:

Figure 10. Time-dependent load sharing between raft and piles on the
W = W’+ Uw = Pp + Pr’+ Uw (1) tributary area.

where W is a total building load, W’ is a net building load, Uw Table 1. Load sharing among piles, soil and soil-cement walls.
At the end of
is buoyancy acted on the raft, Pp is pile resistance and Pr’ is net construction
4 years after the
end of construction
raft resistance. The net raft resistance consists of the net 㧔 Dec.2,
㧔Dec.13, 2008㧕
2004㧕
resistance of soil and that of the grid-form soil-cement walls.
Ratio of load carried by piles
On the tributary area shown in Fig.3, Eq.(1) can be expressed as to net building load
0.54 0.69
follows: (0.50)* (0.64)*
Įp’=Ppt / (Wt - uw (At- Ap))
Ratio of net load carried by soil
Wt = Ppt + ps’(At-Ap-Ag) + pg’Ag + uw(At- Ap) (2) to net building load 0.21 0.18
Įs’=ps’(At-Ap-Ag) / (Wt - uw (At- Ap))
where Wt is a total building load and Ppt is the sum of the pile- Ratio of net load carried by soil-cement
head load on the tributary area, ps’ is a net contact pressure walls to net building load 0.25 0.13
between the raft and soil, pg’ is a net contact pressure between Įg’=pg’ Ag / (Wt - uw (At- Ap))
the raft and the grid-form soil-cement walls, uw is a pore-water *Values in parentheses are load-sharing ratios to total building load
pressure beneath the raft, At is a tributary area, Ag is a plane
area of the grid-form soil-cement walls and Ap is the sum of the which is less than the serviceability limit of 1/1000 radian. The
cross-sectional area of the pile. Assuming that the total load ratio of the load carried by the piles to the net building load on
applied to the foundation is constant after the end of the tributary area was estimated to be 0.69. This means that the
construction and Wt is equal to the total load in structural design average pressure acted on the raft corresponds to 85% of the
on the tributary area, the load sharing between raft and piles can effective overburden pressure before excavation. Consequently
be estimated by Eq.(2) with the measured values. validity of the foundation design based on the “creep piling” is
Figure 10 shows the time-dependent load sharing among the generally confirmed.
piles, soil, the soil-cement walls and the buoyancy on the
tributary area. Table 1 shows the load-sharing ratios to the net
building load at the end of construction and those at the time ACKNOWLEDGEMENT
about four years after the end of construction. The ratio of the
load carried by the piles considerably increased after the end of The authors are grateful to Messrs. Nobuo Nakayama and
construction. On the other hand the ratio of the net load carried Akihiro Miyashita, of Takenaka Corporation, for their
by the grid-form soil-cement walls to the net buiding load contribution to the structural design of the piled raft, Dr. Junji
considerably decreased after the end of construction while the Hamada and Mr. Tomohiro Tanikawa for their assistance in the
ratio of the net load carried by the soil slightly decreased after field measurement.
the end of construction. As for the raft-soil-pile interaction
behaviour on the above, the following mechanism might be
REFERENCES
presumed: consolidation settlement occurred in the normally
consolidated silt layer down to a depth of 6 m due to loading by Hansbo, S. 1984. Foudations on friction creep piles in soft clays, Proc.
the raft over the effective overburden pressure of the silt. The 1st Int. Conference on Case Histories in Geotechnical Engineering,
excessive load transferred to the soil-cement walls and this Vol.2, pp.914-915.
caused consolidation settlement in the upper silt layer with an Jendeby, L. 1986. Friction piled foundations in soft clay -A study of
OCR of about 1.1 just below the bases of the soil-cement walls. load transfer and settlements, Thesis, Chalmers University,
As a result a part of the load carried by the soil-cement walls Gothenburg, Sweden.
gradually transferred to the piles. Mandolini, A., Russo, G. & Viggiani, C. 2005. Pile foundations:
Experimental investigations, analysis and design, Proc. 16th
Based on the results of monitoring on load sharing at the time
ICSMGE, Vol.1, pp.177-213.
about four years after the end of construction, it is known that Poulos, H. G. 2001. Piled raft foundations: design and applications,
the net load carried by the raft per unit area corresponds to 85% Geotechnique 51, No2, pp.95-113.
of the effective overburden pressure before excavation, i.e. the Tokimatsu, K. & Yoshimi, Y. 1983. Empirical correlation of soil
average pressure acted on the raft is close to the effective liquefaction based on SPT N-value fines content, Soils &
overburden pressure. Foundations, Vol.23, No.4, pp.56-74.
Tokimatsu, K., Mizuno, H. and Kakurai, M. 1996. Building damage
associated with geotechnical problems, Special Issue of Soils &
6 CONCLUSIONS Foundations, pp.219-234.
Yamashita, K. and Kakurai, M. 1991. Settlement behavior of the raft
foundation with friction piles, Proc, 4th Int. Conf. on Piling and
Long-term monitoring on settlement and load sharing of the Deep Foundations, pp.461-466.
piled raft with the grid-form ground improvement have been Yamashita, K., Yamada, T. and Hamada, J. 2008. Recent case histories
performed. At the time about four years after the end of on monitoring settlement and load sharing of piled rafts in Japan,
construction, the measured settlements were 17-31 mm and the Proc. of the 5th Int. symposium on Deep Foundations on Bored and
maximum inclination angle of the foundation was 1/1200 radian, Auger Piles, pp.181-193.