See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/280320816

Applicability of Simple Method to Piled Raft Analysis in Comparison with

Field Measurements

Article · June 2015

CITATIONS READS

4 456

3 authors, including:

Kiyoshi Yamashita Junji Hamada

Takenaka Research & Development Institute Takenaka Research & Development Institute
45 PUBLICATIONS   319 CITATIONS    36 PUBLICATIONS   206 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

studying about piled rafts with deep mixing walls View project

Settlement and load sharing of piled raft foundations View project

All content following this page was uploaded by Kiyoshi Yamashita on 24 July 2015.

The user has requested enhancement of the downloaded file.

 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 46 No.2 June 2015 ISSN 0046-5828

Applicability of Simple Method to Piled Raft Analysis in Comparison

with Field Measurements
K. Yamashita1, T. Tanikawa2 and J. Hamada3
1,2,3
Research and Development Institute, Takenaka Corporation, Chiba, Japan
E-mail: [email protected]

ABSTRACT: An applicability of the simple method of the combined pile group and raft method proposed by Clancy and Randolph (1996)
to piled raft analysis was examined through comparisons with the field monitoring results from four case histories in Japan. To deal with
multi-layered soils with finite depth, the simple method was modified using the Steinbrenner’s solution. The shear modulus of soil used in
the analysis was determined by degrading the shear modulus at small strains using a reduction factor, where a set of reduction factors were
employed in Case 2 while a single reduction factor was used in Case 1. Consequently, it was found that the presented approach based on the
method of Clancy and Randolph gave an approximate average settlement and load sharing between the pile group and the raft, when the
reduction factor of shear modulus was 0.4 in Case 1 and 0.3 in Case 2.

KEYWORDS: Piled raft foundation, Design method, Analysis, Settlement, Load sharing

1. INTRODUCTION 2. COMBINED PILE GROUP AND RAFT METHOD

In recent years, there has been an increasing recognition that the use 2.1 Stiffness of piles rafts and load sharing
of piles to reduce raft settlement can lead to considerable economic
savings without compromising the safety and the performance of the Randolph (1983) has proposed the approximate pile group-raft
foundations (Poulos, 2001). The effectiveness of piled rafts in interaction approach which employs a flexibility matrix to combine
reducing average and differential settements has been confirmed not the individual stiffness of the pile group and the raft. The overall
only on favorable ground conditions as shown by Katzenbach et al. stiffness of piled raft, kpr, and the ratio of load carried by pile group
(2000) and Mandolini et al. (2005), but also on unfavorable ground to structure load may be estimated by the following equations (1)
conditions with ground improvement techniques (Yamashita et al., and (2):
2011a; Yamashita et al., 2011b; Yamashita et al., 2013a). k p  kr (1  2αrp ) (1)
The design process of piled rafts generally involves three key kpr  2
1  (kr /kp )αrp
stages, i.e., preliminary design, main design and detailed design
(Poulos, 2001). At the main and detailed design stages, the three- Pp k p αrp k r (2)
dimensional finite element method or the hybrid finite element- ＝
Pr  Pp k p  k r (1  2αrp )
elastic continuum method have been used to consider the complex
interaction of the raft-soil-pile system carefully. For the preliminary where
design stage, it is required to develop more practical and simple Pp: load carried by pile group
methods for estimating the settlement and the load sharing between Pr: load carried by raft
pile group and raft. kp: overall stiffness of a pile group
The combined pile group and raft method, which combines the kr: overall stiffness of a raft,
equivalent pier method for pile groups (Poulos and Davis, 1980) and αrp : interaction factor
the flexibility matrix method for piled rafts (Randolph, 1983), has
been proposed by Clancy and Randolph (1996). Horikoshi and Clancy and Randolph (1993) showed that the value of αrp tended
Randolph (1999) has extended their work to deal with piled rafts in towards about 0.8 as the number of piles increases, independent of
a non-homogeneous soil where the soil modulus increases linearly pile spacing or pile slenderness ratio. This leads to the following
with depth, allowing for the low aspect ratio of the equivalent pier expressions:
and the finite depth of soil. The calculated stiffness of piled rafts
was compared with those obtained by more rigorous numerical 1  0.6( k r /k p ) (3)
k pr＝ kp
analysis such as the hybrid method of analysis. In addition, Clancy 1  0.64（k r /k p）
and Randolph (1996) presented two case studies of small-scale piled
rafts, in which the results obtained by the combined pile group and Pp 1  0.8 (k r /kp ) (4)
＝
raft method were compared with the field measurements. However, Pr  Pp 1  0.6 (kr /k p )
there exist not so many case studies where the applicability of such
simplified methods has been examined through real behavior of 2.2 Stiffness of pile group in multi-layered soils
piled rafts obtained from field measurements. Poulos and Davis (1980) have proposed the equivalent pier method
In this paper, an applicability of the combined pile group and for estimating the load-settlement response of a pile group, where
raft method is examined through comparisons of the calculated the pile group is replaced by an equivalent pier. For a pile group of
settlements and load sharing between pile group and raft with the plan area, Ag, the diameter of the equivalent pier deq is approximated
field monitoring results from four case histories in Japan. The shear by Eq. (5):
modulus of soil used in the analysis is determined by degrading the
shear modulus at small strains using a reduction factor. Finally, an d eq  1.13 Ag (5)
optimum reduction factor of shear modulus, which gives the
minimum deviation from the measured settlements, is discussed. where Ag is plan area of a pile group.

43
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 46 No.2 June 2015 ISSN 0046-5828

Young’s modulus of the equivalent pier, Eeq, is calculated by  I ( H ) n I ( H )  I s ( H k－1 )  (9)

Eq. (6): we  2(1  s ) qB  s 1   s k 
 G1 k 2 Gk 
Where
Ap (6)
Eeq  Es  ( Ep  Es )
Ag 1  2 s
I s ( H k )  F1k  F2k (9.1)
where 1  s
Es: average Young’s modulus of soil penetrated by piles
Ep: Young’s modulus of piles
Ap: total cross-sectional area of piles  （1  m 2  1） m 2  nk
2
( m  m 2  1) 1  n k
2 
1 
F 1k  m  ln  ln
 2
m(1  m 2  nk  1)
2
m  m 2  nk  1 
Figure 1 shows a schematic of the combined pile group and raft  
approach for piled rafts in multi-layered soils with finite depth. The
(9.1.1)
overall stiffness of the equivalent pier in a non-homogeneous soil nk m
shown in Figure 1(b) may be estimated using a load-transfer F2 k  tan -1
2 2
nk m 2  n k  1 (9.1.2)
approach of a single pile in a non-homogeneous soil proposed by
Randolph and Wroth (1978):
q: rectangular uniform pressure on the soil surface
we: elastic settlement at the raft center
2 2 tanh (Lp ) Lp Gk: shear modulus of soil of the k-th layer

Pｔ (1   s )ξ ζ Lp dp (7) L: length of raft

wt d pGL p 8 tanh (L ) L B: breadth of raft
1 p p

(1   s ) Lp dp Hk: depth of bottom of the k-th layer

Is(Hk): shape factor for finite depth of Hk
where
m =L/B
Pt: pile-head load
nk=2Hk/B
wt: pile-head settlement
Lp: pile length
dp: pile diameter Overall stiffness of the raft may be estimated assuming that the
Gave: average shear modulus of soil along pile length average settlement is equal to the settlement at the raft center
GLp: shear modulus of soil at a depth of pile length multiplied by π/4.
Gb: shear modulus of soil below the level of pile base
νs: Poisson’s ratio of soil 2.4 Differential settlements
H: finite depth of soils To evaluate differential settlement as a proportion of the average
rm: maximum radius of influence of pile settlement of the raft, the raft-soil stiffness ratio for rectangular rafts,
ξ= GLp/Gb Krs, was proposed by Horikoshi and Randolph (1997) as follows:
ρ= Gave/GLp
λ= Ep/GLp (= Eeq/GLp) 0.5 3
Er 1  s2  B   t ｒ  (10)
ζ= ln(2rm/dp) K rs  5.57    
Es 1  r2  L   L 
μLp= 2 2/ζ λ (Lp/dp)
where
In the original paper, shear modulus of soil below the level of pile Er: Young’s modulus of raft
base, Gb, is obtained using the Boussinesq solution for a rigid punch νr: Poisson’s ratio of raft
acting on a homogeneous elastic half space. In the present approach, tr: Raft thickness
to deal with multi-layered soils below the pile base, equivalent shear
modulus, with which the overall stiffness of raft on the multilayered According to their study, rafts may behave substantially flexible
soils is equal to that on a homogeneous elastic half space, is when the value of Krs is less than about 0.05. If a raft can be
employed as Gb in Eq. (7). The equivalent shear modulus may be regarded as flexible, settlements of a piled raft may be estimated
obtained using the Steinbrenner’s solution given by Eq. (9) using the method proposed by Clancy and Randolph (1996) to
(Terzaghi, 1943). Randolph (1994) suggested a modification to the account for the overall reduction in settlements due to the presence
value of the maximum radius of influence of pile, rm, to improve the of the piles in the piled raft. Namely, the settlements of a piled raft
load-transfer approach for stubby piles (Lp/dp < 2.5). For the stubby can be calculated by multiplying the stiffness ratio, kr/kpr, to
piles in a non-homogeneous soil where the soil modulus increases settlements of raft alone.
linearly with depth, the following equation has been proposed by
Horikoshi and Randolph (1999): 3. PROFILES OF SOIL MODULUS
In evaluating shear modulus of soil for static loading, Tatsuoka et al.

ζ  ln A  0.25  2.5  1   s   0.25  2 Lp / d P  (1991) have recommended to modify the shear modulus at small
strains depending on the magnitude of shear strain. To estimate the
(A=5, for stubby piles) (8) overall stiffness of an equivalent pier and a raft, the shear modulus
of soil, G, was determined by degrading the shear modulus at small
2.3 Stiffness of raft in multi-layered soils strains using a reduction factor of shear modulus as follows:

Figure 1(c) shows a schematic of a raft on multi-layered soils with G = RGG0 (11)
finite depth. The elastic settlement of a fully flexible rectangular raft where
at the raft center may be expressed using the formula of G0: shear modulus of soil at small strains
Steinbrenner as follows (Terzaghi, 1943): RG: reduction factor of shear modulus

44
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 46 No.2 June 2015 ISSN 0046-5828

q
P GaveGLpGb Pt G

G1 H1 G1
H2
Lp / 2 Eeq Gave Hk
G2 G2
Hk+1
deq
Gk H Lp Gk n=H
Gk+1 Gk+1
Gb

Gn Gn

Depth Depth
(a) Piled raft (b) Equivalent pier (c) Raft

Figure 1 Combined pile group and raft approach in multi-layered soils

Figure 2 shows the profiles of soil shear modulus assumed in the 1991; Yamashita et al., 1994; Yamashita et al., 2011a). Figure 3
presented approach. As for the reduction factors for equivalent pier shows schematics of the four structures with the soil profiles.
analysis and raft analysis, the following two cases were assumed. In These piled rafts have common characteristics as follows:
Case 1 shown in Figure 2(a), a single reduction factor, RG, was used ・Larger pile group than 3x3 pile group is installed beneath the full
for both the equivalent pier analysis and the raft analysis. The area of the raft. The pile length is equal in the pile group and the
constant A in Eq. (8) was assumed to be 5 according to the proposal plan shape of the raft is substantially rectangle with B/L>0.5.
by Randolph (1994) and Horikoshi & Randolph (1999). In Case 2 Under these conditions, the interaction factor (αrp=0.8) proposed
shown in Figure 2(b), for the raft analysis, RG was used for the soil by Clancy and Randolph (1993) may be applicable.
layers above the level of the pier base. On the other hand, the ・The bearing capacity of the raft alone is adequate since the raft is
reduction factor for those below the level of the pier base was fixed founded on relatively stiff clay or dense sand.
at a larger value, since theoretical solutions for the vertical strain Table 2 shows the measured maximum settlement of the
beneath the centre of a uniformly loaded rectangular area indicates foundation and the ratio of the load carried by the pile group to the
that the vertical strain level is markedly higher within a depth of the effective structure load, αp’, estimated from the measurement results.
breadth of the rectangular area, i.e., the strain level below the depth The values of αp’ were 0.49 to 0.93 and relatively wide range, while
is relatively lower than that above the depth (Poulos, 1993). So, the maximum settlements were 11 to 29 mm. The average
considering that pile length is usually roughly close to breadth of the settlements and the values of αp’ used in comparison with the
raft, the reduction factor for the soil layers below the pier base was analysis, i.e., the modified measured values, were shown in Table 2.
assumed to be 0.8 empirically. For the pier analysis, RG is used for The average settlements were obtained as follows: For the 4-story
the whole soil layers. The constant A for the pier analysis was building, the measured maximum settlement multiplied by π/4; For
reduced to 1.5 to compensate the relative increase in raft stiffness to the 5-story building, the average value of the measured settlements;
pier stiffness based on the preliminary analysis (Yamashita et al., For the 11-story building, the average value of the settlements
2013b). measured between the center and the corner of the raft; For the
47-story building, the vertical ground displacement measured
4. COMPARISONS WITH FIELD MEASUREMENT between the center and the edge of the raft. The modified values of
RESULTS αp’ for the 4-story and 5-story buildings were described in Sections
4.1 Case histories 4.3 and 4.4, respectively.
To obtain the shear modulus of soil at small strains, shear-wave
Table 1 shows case histories of the four piled raft foundations velocity measurements of soil were carried out using down-hole
supporting 14 to 162 m high buildings in Japan, where field technique in the sites of the buildings except for the 4-story building.
measurements were performed on the foundation settlements and the For the 4-story building, the values of G0 were determined based on
load sharing between the piles and the raft (Yamashita and Kakurai, those for the 5-story building, since the 4-story building is located
very closely to the 5-story building.

q q
Pt Pt
G01 G01
Eeq RG, G0ave Eeq RG, G0ave
G02 RG G02
A=5 A=1.5
deq deq
 RG G0k   G0k 
G0k
G0k+1 G0k+1
RG, G0b RG, G0b RG=0.8
G0n G0n

(a) Case 1 (b) Case 2

Figure 2 Profiles of soil shear modulus in analytical model

45
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 46 No.2 June 2015 ISSN 0046-5828

Table 1 Case histories of piled raft foundations

Maximum Total Depth of Piles
Structure Construction height pressure foundation References
period Length Diameter Number Type
(m) (kPa) (m) (m) (m)
0.25×0.25-
4-story office building 1986-87 14.1 61 2.1 15.1 0.40×0.40 16 Steel-H pile Yamashita & Kakurai (1991)
0.30×0.30-
5-story office building 1992-93 17.1 84 2.4 14.6-15.8 0.41×0.41 20 Steel-H pile Yamashita et al. (1994)

11-story office building 2004-05 60.8 145 3.0，3.6 27.5，26.9 1.1-1.5 40 Cast-in-place Yamashita et al. (2011a)
concrete pile

47-story residential tower 2006-09 161.9 600 4.3 50.2 1.5-1.9 36 Cast-in-place Yamashita et al. (2011a)
concrete pile

SPT 17.1m SPT

14.1m N-value
N-Value N-Value
00 50
50 0 50
GL±0 2.1m GL±0 2.4m
Kanto loam Kanto loam
Medium sandy clay
Loose sand
10 15.1m
Medium clay 10 15.8m
Loose to medium sand
Medium sand
Stiff silt Stiff silty clay
Medium to dense sand
Medium to dense sand
Depth (m)

20
Depth (m)

20 Stiff silt
Stiff silty clay
Stiff silt
Stiff silty sand Medium silty sand
Stiff clayey silt
30 30
Hard silt Hard clay

161.9m
40 40 Hard silt
Hard clay
Dense sand and gravel
Dense clay
Dense sand Dense sand
50 50

(a) 4-story office building (b) 5-story office building

SPT
N-Value
0 50 Laminated rubber bearings
GL±0
60.8ｍ

Silt
Sand and gravel 5.3m 5.8m 4.3m
10
Sand

20
18m
SPT Silt
Depth (m)

N-Value
Laminated rubber bearings 30 Sand
0 50 GL 32m
GL±0 Sand and gravel
3.6m 3.0m
3.9m 40 Sand
Sand 41m
10
Sandy silt Silt
Sand and 50 52.4m
Depth (m)

20 gravel
Clay Sand and gravel 54.5m
Sand 60
28m
30 Sand and 30.5m
gravel
70m Settlement gauges
Sandy silt
40
Sand and
gravel
Sand
50

(c) 11-story office building (d) 47-story residential tower

Figure 3 Schematics of four structures with soil profiles

Table 2 Measured settlement and load sharing between piles and raft
Measured values Modified measured values
Structure Max. settlement Ratio of piles to Ave. settlement Ratio of piles to
(mm) structure load, αp’ wmeas (mm) structure load, αmeas
4-story office building 10.5*1 0.56*2 8 0.59
5-story office building 18.5*1 0.49*1 14 0.57 *1 End of construction (E.O.C.)
*2 Average in 3-12 months after E.O.C.
11-story office building 10*3 0.65*4 9 0.65 *3 2.5 months before E.O.C.
47-story residential tower 29*5 0.93*5, 0.87*5 24 0.90 *4 Average in 22-60 months after E.O.C.
*5 17 months after E.O.C.

46
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 46 No.2 June 2015 ISSN 0046-5828

4.2 Parameters used for analyses

Table 3 shows parameters used for the presented method of analysis.
The slenderness ratios of an equivalent pier, Lp/deq, are much less
than 2.5, which may correspond to extremely stubby piles. The
variation of soil modulus with depth, ρ= Gave/GLp, was determined
on the assumption that average shear modulus of soil along pile
length at small strains, G0ave, is equal to the average shear modulus
at small strains along pile length, taking account of the thickness of
each layer. Poisson’s ratio of soil, νs, was fixed at 0.3 since the
effects of νs on the accuracy of the method are small (Horikoshi et
al., 1999). The finite depth, H, was assumed to be (Lp+B) which
might be the minimum depth from the point of design practice.
However, for the 47-story building, the finite depth was set to be 70
m from the ground surface, which corresponds to the depth of the
reference point of the settlement gauge.
The reduction factor of shear modulus, with which the calculated Figure 4 Foundation plan with layout of piles
average settlement of the piled raft matched the measured one, is (Yamashita et al., 1991)
defined as an equivalent reduction factor (denoted as RGeq). In order
to back-calculate the equivalent reduction factor, the reduction
factor of shear modulus was chosen as a variable, i.e., the value of
6.0
RG was varied from 0.20 to 0.60 in Case 1 and 0.15 to 0.50 in Case

Sum of pile head load (MN)

2, in increments of 0.05.

4.3 Four-story office building 4.0

The 4-story reinforced concrete building, shown in Figure 3(a), is
located in Saitama, suburb of Tokyo. Figure 4 shows the foundation
plan with a layout of the piles. Figure 5 shows the relationship 2.0
between the sum of the measured pile-head load and the structure
load in the tributary area during and after the construction. It can be
seen that the sum of the measured pile-head loads was nearly zero 0
after the casting of the raft since almost entire load of the raft was 0 2.0 4.0 6.0 8.0
directly transferred to the soil. Thereafter, the load carried by the
piles increased in proportion to the increase in the structure load. In Estimated building load on tributary area (MN)
case of low-rise buildings where the ratio of the self-weight of the
raft to the total load of the structure is not negligible, the ratio of the Figure 5 Relationship between sum of pile-head load and structure
load carried by the piles to the effective structure load, αp’, was re- load in the tributary area (Yamashita et al., 1991)
estimated by applying a linear regression to the relationship between
the sum of the pile-head load and the structure load. The modified Figure 8(a) shows the calculated average settlement, denoted as
value of αp’, αmeas, was shown in Table 2. wcalc, versus the reduction factor of shear modulus, together with the
Figure 6 shows the analytical models of Cases 1 and 2 with the measured average settlement, denoted as wmeas, shown in Table 2.
profile of shear modulus of soil at small strains. For the raft analysis, The values of wcalc increased with decrease in the value of RG, as
the plan of the raft was modelled as an equivalent rectangle as expected. The value of the equivalent reduction factor, RGeq, was
shown in Table 3. The applied pressure was set to 54 kPa, which interpolated as 0.36 in Case 1 and 0.29 in Case 2. Figure 8(b) shows
corresponded to the total pressure (61 kPa) minus the pressure of the the calculated value of αp’, denoted as αcalc, versus the reduction
0.3-m thick raft (7 kPa), considering that the measurements of the factor, together with the modified measured value of αp’, denoted as
foundation settlement began after constructing the raft. Figure 7 αmeas. The value of αcalc with RGeq=0.36 in Case 1 and that with
shows the calculated overall stiffness of piled raft, pile group and RGeq=0.29 in Case 2 agreed well with the measured one, while the
raft. For the same value of RG, the stiffnesses in Case 2 were slightly values of αcalc were only slightly larger than the measured one in
larger than those in Case 1 as expected. Incidentally, the stiffness of both cases. Figure 8 indicates that the value of αcalc, which is
piled raft was close to the stiffness of pile group in both cases. This equivalent to the value of kp/kr, in Case 1 was affected clearly by
may be expected from Eq. (3) as pointed out by Randolph (1994). change in the value of RG than that in Case 2.

Table 3 Parameters used for analyses

Raft Equivalent pier Soil
Structure Breadth Length Lp Ag deq Ap Ep H ρ νs
Lp/deq
B (m) L (m) (m) (m2) (m) (m2) (MPa) （m）
4-story office building 20.0 24.6 15.1 492 25.0 0.24 205800 0.60 Lp+B 0.50 0.3
5-story office building 24.0 25.0 15.8 600 27.6 0.38 205800 0.57 Lp+B 053 0.3
11-story office building 45.0 80.0 27.5 3600 67.7 52.77 20580 0.41 Lp+B 0.55 0.3
47-story residential tower 30.5 47.0 50.0 1434 42.7 82.18 29400 1.17 70 0.50 0.3

47
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 46 No.2 June 2015 ISSN 0046-5828

G0 (MPa) q=54kPa (61kPa)

0 200 400 Pier Raft Pier Raft
0 2.1m
G =RGG 0ave G =RGG 0ave
G k =RGG 0k
10 deq=25.0m Lp =15.1m G 0ave=90MPa G 0ave=90MPa
RG=0.15-0.5
RG=0.2-0.6 RG=0.15-0.5
G k =RGG 0k
Depth (m)

20 G b=RGG 0b
RG =0.2-0.6
G b=RGG 0b
B =20.0m G 0b =117MPa G 0b =117MPa G k =0.8G 0k
30 RG =0.2-0.6 RG =0.15-0.5
37.2m
40
(a) Case 1 (b) Case 2
50
Figure 6 Analytical models for 4-story building

Case 1 and 0.24 in Case 2. Figure 13(b) shows the calculated value
of αp’, αcalc, versus the reduction factor of shear modulus. The values
of αcalc with the RGeq in both cases were slightly larger than, but
agreed well with the measured one.

kpr kpr
kp kp
kr kr

(a) Case 1 (b) Case 2

Figure 7 Calculated stiffness of piled raft, pile group and raft

Figure 9 Foundation plan with layout of piles (Yamashita et al.,

1994)
0.29 0.36 0.29 0.36

(a) Average settlement (b) Ratio of pile group to

structure load

Figure 8 Comparison of calculations with measurements

4.4 Five-story office building

The 5-story reinforced concrete building, shown in Figure 3(b), is
located close to the 4-story building. Figure 9 shows the foundation
plan with a layout of the piles. Figure 10 shows the relationship
between the sum of the measured pile-head load and the structure Figure 10 Relationship between sum of pile-head load and structure
load in the tributary area. In the same way as the 4-story building, load in the tributary area (Yamashita et al., 1994)
the ratio of the load carried by the piles to the structure load was re-
estimated by applying a linear regression to the relationship between 4.5 Eleven-story office building
the sum of the pile-head load and the structure load. The modified The 11-story steel-frame building, shown in Figure 3(c), is located
value of αp’ was shown in Table 2. in Aichi Prefecture. Figure 14 shows the foundation plan with a
Figure 11 shows the analytical models of Cases 1 and 2 with the layout of the piles. Figure 15 shows the analytical models of Cases 1
profile of shear modulus of soil at small strains. The applied and 2 with the profile of soil shear modulus at small strains. The
pressure was set to 77 kPa, which corresponded to the total pressure shear wave velocity below a depth of 50 m was assumed to be the
(84 kPa) minus the pressure of the 0.3-m thick raft (7 kPa), as in the same value as that above the depth. The applied pressure was set to
4-story building. Figure 12 shows the calculated overall stiffnesses 106.5 kPa while the total pressure was 145 kPa, considering that the
of piled raft, pile group and raft. The stiffnesses were slightly larger measurements of the foundation settlement began after constructing
than, but similar to those in the 4-story building. the 0.8-m thick raft and ended 2.5 months before the end of the
Figure 13(a) shows the calculated average settlement, wcalc, versus construction. Figure 16 shows the calculated overall stiffnesses of
the reduction factor of shear modulus, together with the average piled raft, pile group and raft. The stiffnesses were much larger than
settlement, wmeas. The value of RGeq was interpolated as 0.30 in those in Figures 7 and 12.
48
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 46 No.2 June 2015 ISSN 0046-5828

G0 (MPa) q=77kPa (84kPa)

0 200 400 Pier Raft Pier Raft
0 2.4m
G =RGG 0ave G =RGG 0ave
Lp =15.8m G k =RGG 0k
10 deq=27.6m G 0ave=101MPa G 0ave=101MPa
RG=0.15-0.5
RG=0.2-0.6 RG=0.15-0.5
G k =RGG 0k
Depth (m)

20
RG=0.2-0.6 G b=RGG 0b
G b=RGG 0b
B =24.0m G 0b =107MPa G 0b =107MPa G k =0.8G 0k
30
RG=0.2-0.6 RG＝0.15-0.5

40 42.2m

50 (a) Case 1 (b) Case 2

Figure 11 Analytical models for 5-story building

kpr kpr
kp kp
kr kr 0.24 0.30 0.24 0.30

(a) Case 1 (b) Case 2 (a) Average settlement (b) Ratio of pile group to
structure load
Figure 12 Calculated stiffness of piled raft, pile group and raft Figure 13 Comparison of calculations with measurements

Pile diameter Monitoring devices

1.5m－1.8m 1.3m－1.6m Instrumented pile
1.4m－1.8m 1.1m－1.4m Earth pressure cell

Y4
16.0m

GL-3.0m

Y3
E2 Tributary
43.5m
11.5m

area
Y2
P1
16.0m

GL-3.6m
E1
Y1

X1 X2 X3 X4 X5 X6 X7 X8 X9 X10
7.5m 9.6m 9.6m 9.6m 9.6m 9.6m 9.6m 8.6m 6.3m
80.0m

Figure 14 Foundation plan with layout of piles (Yamashita et al., 2011a)

G0 (MPa) q = 106.5kPa (145kPa)

0 400 800 Pier Raft Pier Raft
0 3.0m
10 G =RGG 0ave G =RGG 0ave G k =RGG 0k
deq=67.7m Lp=27.5m G 0ave=203MPa G 0ave=203MPa R =0.15-0.5
20 RG =0.2-0.6 RG=0.15-0.5
G
Depth (m)

30 G k =RGG 0k
RG =0.2-0.6
40 G b=RGG 0b G b=RGG 0b
G 0b =577MPa G 0b =577MPa G k =0.8G 0k
50 B =45.0m RG =0.2-0.6 RG＝0.15-0.5
60
70
75.5m
(a) Case 1 (b) Case 2
Figure 15 Analytical models for 11-story building
49
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 46 No.2 June 2015 ISSN 0046-5828

Calculated (Y1 & Y4)

kpr kpr
kp kp Calculated (Y2 & Y3)
kr kr

(a) Case 1 (b) Case 2 Figure 18 Calculated and measured settlement profiles for 11-story
building
Figure 16 Calculated stiffness of piled raft, pile group and raft
4.6 Forty-seven-story residential tower
Figure 17(a) shows the calculated average settlement, wcalc,
versus the reduction factor of shear modulus, together with the The 47-story reinforced concrete residential tower, shown in Fig.
measured average settlement. The value of RGeq was interpolated as 3(d), is located in Aichi Prefecture. Figure 19 shows the foundation
0.50 in Case 1 and 0.40 in Case 2. Figure 17(b) shows the calculated plan with a layout of the piles. Figure 20 shows the analytical
value of αp’, αcalc, versus the reduction factor of shear modulus. The models of Cases 1 and 2 with the profile of soil shear modulus at
value of αcalc with RGeq=0.50 was somewhat less than the measured small strains. The applied pressure was set to 580 kPa, which
value, while the value of αcalc with RGeq=0.40 in Case 2 was in good corresponded to the effective pressure, namely total pressure of 600
agreement with the measured one. Figure 18 shows the comparison kPa minus the measured pore-water pressure beneath the raft of 20
of the measured longitudinal settlement profile of the raft with the kPa. Figure 21 shows the calculated overall stiffnesses of piled raft,
calculated settlement profile under the condition that the calculated pile group and raft. The stiffness of piled raft was similar to that in
average settlement matched with the measured one. The settlements the 47-story tower. In both cases, the stiffness of piled raft was
of the piled raft were calculated by multiplying the stiffness ratio of almost identical to that of pile group while the stiffness of raft was
kr/kpr to the settlements of the raft alone. While the calculated considerably smaller than that of piled raft.
settlements in both cases are almost identical, those in Case 2 are
shown in Figure 18, where the value of kr/kpr was 0.82 as shown in
Figure 16(b). The calculated settlements roughly agreed with the
measured ones. This is likely that the raft-soil stiffness ratio, Krs, is
0.002, much less than 0.05.

5D 7D

Tributary area

0.40 0.50 0.40 0.50

(a) Average settlement (b) Ratio of pile group to

structure load

Figure 17 Comparison of calculations with measurements Figure 19 Foundation plan with layout of piles
(Yamashita et al., 2011a)
G0 (MPa) q =580kPa (600kPa)
0 400 800
0 Pier Raft Pier Raft
4.3m
10

20 G =RGG 0ave G =RG1G 0ave

G k =RG1G 0k
G 0ave=159MPa G k =RGG 0k G 0ave=159MPa
Depth (m)

deq=42.7m Lp=50.2m
30 RG=0.2-0.6 RG1=0.15-0.5
RG1=0.15-0.5
RG=0.2-0.6
40

50
G b=RGG 0b G b=RG1G 0b
60 G 0b =688MPa G 0b =688MPa G k =0.8G 0k
RG=0.2-0.6 RG1＝0.15-0.5
70 70.0m

(a) Case 1 (b) Case 2

Figure 20 Analytical models for 47-story tower

50
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 46 No.2 June 2015 ISSN 0046-5828

Figure 24 shows the ratio of the calculated average settlement to the

measured one, wcalc/wmeas, versus the reduction factor of shear
modulus for the four case histories. Figure 25 shows the ratio of the
calculated value of αp’ to the measured value of αp’, αcalc/αmeas,
versus the reduction factor of shear modulus. The curves in
Figures 24 and 25 were obtained on the assumption that the
kpr kpr calculated values were interpolated approximately by the 6-order
kp kp polynomials.
kr kr

(a) Case 1 (b) Case 2

Figure 21 Calculated stiffness of piled raft, pile group and raft

Figure 22(a) shows the calculated average settlement, wcalc, versus

the reduction factor of shear modulus, together with the average
settlement. The value of RGeq was interpolated as 0.38 in Case 1 and
0.30 in Case 2. Figure 22(b) shows the calculated value of αp’, αcalc,
versus the reduction factor of shear modulus. The values of αcalc
with the RGeq in both cases were only slightly larger than, but in 0.36 0.29
RGeq=0.30 0.38 0.50 RGeq=0.24 0.30 0.40
good agreement with the measured one. Figure 23 shows the
comparison of the measured longitudinal settlement profile of the
raft with the calculated settlement profile (Case 2) under the
condition that the calculated average settlement matched with the (a) Case 1 (b) Case 2
measured one. The settlements of the piled raft were calculated by
multiplying the stiffness ratio to the settlements of the raft alone, Figure 24 wcalc/wmeas vs. reduction factor with of RGeq
where the value of kr/kpr was 0.19 as shown in Figure 21(b). The
calculated settlements generally agreed with the measured ones,
where the raft-soil stiffness ratio was relatively small (Krs=0.06).

0.30 0.38 0.30 0.38

0.36 0.29
RGeq=0.30 0.38 0.50 RGeq=0.24 0.30 0.40

(a) Average settlement (b) Ratio of pile group

to structure load
(a) Case 1 (b) Case 2
Figure 22 Comparison of calculations with measurements
Figure 25 αcalc/αmeas vs. reduction factor with RGeq

The values of the equivalent reduction factor of shear modulus, RGeq,

are shown by open circles in Figures 24 and 25. The values of
αcalc/αmeas with those of RGeq were summarized in Table 4. The
values of RGeq in both cases were within a limited range, i.e., 0.30 to
Calculated (A & E) 0.50 in Case 1 and 0.24 to 0.40 in Case 2. The values of αcalc/αmeas
were 0.83 to 1.13 in Case 1 and 0.95 to 1.09 in Case 2. Therefore,
the calculated values with the RGeq in both cases were generally
consistent with the measured ones, while the variation in the values
Calculated (C)
of αcalc/αmeas in Case 2 was less than that in Case 1. These results
suggest that an appropriate reduction factor of shear modulus could
be chosen in design practice.
In order to find an optimum reduction factor of shear modulus,
Figure 23 Calculated and measured settlement profiles for which gives the minimum deviation of the calculated and measured
47-story tower values, the sum of the difference between the calculated and
measured settlements in the four case studies, Δw, was evaluated as
5. DISCUSSIONS follows:
4
To discuss an accuracy of the calculated values in Cases 1 and 2, the
ratios of the calculated values to the measured ones are examined. Δw =  {(wcalc/wmeas)i -1}2 (12)
i1

51
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 46 No.2 June 2015 ISSN 0046-5828

Figure 26(a) shows the Δw versus RG relationship. The optimum

reduction factor of shear modulus, which give the minimum value of
Δw, denoted as RG*, was 0.39 in Case 1 and 0.31 in Case 2. The
minimum values of Δw in both cases were almost identical. The sum
of the difference between the calculated and measured values of αp’
in the four case studies, Δα, was also evaluated using the following
the equation.
4
Δα =  {(αcalc/αmeas)i -1}2 (13)
i1

Figure 26(b) shows the values of Δα corresponding to the RG*. The

values of Δα in both cases were almost identical. Figure 27 shows the
values of wcalc/wmeas with the RG* on the wcalc/wmeas versus RG curve
shown in Figure 24. Figure 28 shows the values of αcalc/αmeas with RG*=0.39 RG*=0.31
the RG* on the αcalc/αmeas versus RG curve shown in Figure 25.
Table 5 summarizes the values of wcalc/wmeas and αcalc/αmeas with the
RG*. Since the values of wcalc/wmeas with the RG* were 0.79 to 1.22 in
both cases, the calculated settlements with the optimum reduction (a) Case 1 (b) Case 2
factors were generally consistent with the measured ones. Moreover,
the values of αcalc/αmeas with the RG* were 0.96 to 1.09 in Case 2 Figure 27 wcalc/wmeas vs. reduction factor with values of RG*
while 0.91 to 1.07 in Case 1. Therefore, the calculated values of αp’
with the optimum reduction factors were in good agreement with the
measured ones in both cases, while the difference between the
maximum and minimum values of αcalc/αmeas in Case 2 was slightly
less than that in Case 1.

Case1 Case1
Case2 Case2

RG*=0.31 0.39 RG*=0.31 0.39

RG*=0.39 RG*=0.31

(a) Deviation from measured (b) Deviation from measured

settlements, Δw values of αp’, Δα (a) Case 1 (b) Case 2

Figure 26 Deviation of calculations and measurements with Figure 28 αcalc/αmeas vs. reduction factor with values of RG*
optimum reduction factor

Table 4 Values of αcalc/αmeas with equivalent reduction factor RGeq

4-story building 5-story building 11-story building 47-story tower
Case A
RGeq αcalc/αmeas RGeq αcalc/αmeas RGeq αcalc/αmeas RGeq αcalc/αmeas
1 5 0.36 1.05 0.30 1.13 0.50 0.83 0.38 1.05
2 1.5 0.29 1.09 0.24 1.08 0.40 0.95 0.30 1.06

Table 5 Values of wcalc/wmeas and αcalc/αmeas with optimum reduction factor RG*

wcalc/wmeas αcalc/αmeas
Case A RG*
4-story building 5-story building 11-story building 47-story tower 4-story building 5-story building 11-story building 47-story tower

1 5 0.39 0.92 0.79 1.22 0.98 1.02 1.07 0.91 1.05

2 1.5 0.31 0.92 0.79 1.22 0.97 1.09 1.07 0.96 1.06

52
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 46 No.2 June 2015 ISSN 0046-5828

6. CONCLUSIONS Poulos, H.G., 1993. Settlement prediction for bored pile groups,
Proc. of the 2nd International Seminar on Deep Foundations
The applicability of the combined pile group and raft method on Bored and Auger Piles, 103-117.
proposed by Clancy and Randolph (1996) to piled raft analysis was Poulos, H.G., 2001. Piled raft foundations: design and applications,
examined through comparisons with the monitoring results from the Geotechnique 51, No.2, 95-113.
four case histories. To deal with multi-layered soils with finite depth, Poulos, H.G. and Davis, E. H., 1980. Pile Foundation Analysis and
the combined pile group and raft method was modified using the Design, John Wiley & Sons.
Steinbrenner’s solution. The shear modulus of soil is determined by Randolph, M.F., 1983. Design of piled raft foundations, Proc. of the
degrading the shear modulus at small strains using a reduction factor Int. Symp. on Recent Developments in Laboratory and Field
of shear modulus, where a set of reduction factors were employed in Tests and Analysis of Geotechnical Problems, Bangkok, 525-
Case 2 while a single reduction factor was used in Case 1. Through 537.
the examination, the following conclusions can be drawn: Randolph, M.F., 1994. Design methods for pile group and piled
1. The equivalent reduction factors of shear modulus, RGeq, with rafts, Proc. of the 13th Int. Conf. on SMFE, Vol.5, 61-82.
which the calculated settlements matched the measured ones, Randolph, M.F. and Wroth, C.P., 1978. Analysis of deformation of
were back-calculated to be 0.30 to 0.50 in Case 1 and 0.24 to vertically loaded piles, J. Geotech. Engrg. Division, ASCE,
0.40 in Case 2. The values of RGeq in both cases were found to Vol. 104, GT12, 1465-1488.
be within a limited range. The calculated values of αp’ with the Terzaghi, K., 1943. Theoretical Soil Mechanics, John Wiley &
RGeq were found to be in good agreement with the measured Sons.
ones. Tatsuoka, F. and Shibuya, S., 1991. Deformation characteristics of
2. The calculated settlements by a raft-alone analysis with the soils and rocks from field and laboratory tests, Proc. of the
equivalent reduction factor of shear modulus, which were 9th Asian Regional Conf. on SMFE, 53-114.
factored by the stiffness ratio of kr/kpr, were found to be Yamashita, K. and Kakurai, M., 1991. Settlement behavior of the
generally consistent with the measured ones, where the ratios of raft foundation with friction piles, Proc. 4th Int. Conf. on
the raft-soil stiffness, Krs, were relatively small. Piling and Deep Foundations, 461-466.
3. The optimum reduction factor of shear modulus, RG*, which Yamashita, K., Kakurai, M. and Yamada, T., 1994. Investigation of
gave the minimum deviation from the measured settlements, a piled raft foundation on stiff clay, Proc. 13th Int. Conf. on
was calculated to be 0.39 in Case 1 and 0.31 in Case 2. The SMFE, Vol.2, 543-546.
ratios of the calculated settlement with the RG* to the measured Yamashita, K., Yamada, T. and Hamada, J., 2011a. Investigation of
settlement were 0.79 to 1.22 in both cases. The ratios of the settlement and load sharing on piled rafts by monitoring full-
calculated value of αp’ with the RG* to the measured one in Case scale structures, Soils & Foundations, Vol.51, No.3, 513-532.
2 were 0.96 to 1.09, while those in Case 1 were 0.91 to 1.07. As Yamashita, K., Hamada, J. and Yamada, T., 2011b. Field
a result, it was found that the presented approach based on the measurements on piled rafts with grid-form deep mixing
combined pile group and raft method gave an approximate walls on soft ground, Geotechnical Engineering Journal of the
average settlement and load sharing between the pile group and SEAGS & AGSSEA, Vol.42, No.2, 1-10.
the raft, when the reduction factor of shear modulus was about Yamashita, K., Wakai, S. and Hamada, J., 2013a. Large-scale piled
0.4 in Case 1 and about 0.3 in Case 2. raft with grid-form deep mixing walls on soft ground, Proc. of
4. It appeared that the analysis in Case 2 may give more the 18th ICSMGE, 2637-2640.
satisfactorily accurate estimation of load sharing between the Yamashita, K., Hamada, J. and Tanikawa, T., 2013b. Applicability
piles and the raft. However, further study on the examination of simplified method to piled raft analysis for estimating
through comparisons of the calculations with the measurements settlement and load sharing, Proc. of the 2nd Int. Conf.
would be required. Geotec Hanoi 2013, 53-61.

7. ACKNOWLEDGMENTS
The authors are grateful to Prof. Tsutomu Tsuchiya of Muroran
Institute of Technology for his useful discussions and constructive
comments. The authors are also grateful to Messrs. Y. Soga, K.
Shimono, M. Yamada, T. Yokonami of Takenaka Corporation for
their contribution to the foundation design and the field
measurements.

8. REFERENCES
Clancy, P. and Randolph, M. F., 1993. An approximate analysis
procedure for piled raft foundations, Int. J. Num. & Anal.
Methods in Geomechanics, Vol. 17, 849-869.
Clancy, P. and Randolph, M. F., 1996. Simple design tools for piled
raft foundations, Geotechnique 46, No.2, 313-328.
Horikoshi, K. and Randolph, M. F., 1997. On the definition of raft-
soil stiffness ratio, Geotechnique 47, No.4, 741-752.
Horikoshi, K. and Randolph, M. F., 1999. Estimation of overall
settlement of piled rafts, Soils & Foundations, Vol.39, No.2,
59-68.
Katzenbach, R., Arslan, U. and Moormann, C., 2000. Piled raft
foundation projects in Germany, Design applications of raft
foundations, Hemsley J.A. Editor, Thomas Telford, 323-392.
Mandolini, A., Russo, G. and Viggiani, C., 2005. Pile foundations:
Experimental investigations, analysis and design, Proc. of the
16th ICSMGE, Vol. 1, 177-213.

53

View publication stats