HOSTED BY Available online at www.sciencedirect.

com

ScienceDirect
Soils and Foundations 62 (2022) 101179
www.elsevier.com/locate/sandf

Technical Paper

Experimental investigation of bearing mechanism of closed-

and open-ended piles supported by thin bearing layer using
X-ray micro CT
Ryunosuke Kido a,⇑, Riki Suezawa b, Yasuo Sawamura c, Makoto Kimura d
a
Department of Civil and Earth Resources Engineering, Kyoto University, Japan
b
Japan Ministry of Economy, Trade and Industry, Japan
c
Department of Urban Management, Kyoto University, Japan
d
Department of Civil and Earth Resources Engineering, Kyoto University, Japan

Received 6 October 2021; received in revised form 16 May 2022; accepted 1 June 2022
Available online 28 June 2022

Abstract

In order to clarify the bearing mechanism of closed-ended and open-ended piles supported by a thin bearing layer, pile-loading tests
are conducted on model grounds with different bearing layer thicknesses, and the soil deformation characteristics around the pile tips are
observed by X-ray micro CT. In the case of open-ended piles supported by a thin bearing layer, the soil in the pile greatly displaces fol-
lowing the downward displacement of the soil located more deeply than the pile tip, and the soil density in the pile becomes lower than
when the bearing layer thickness is sufficiently large. These characteristics probably cause lower inner friction and lower base resistance,
resulting in a lower bearing capacity. When the bearing layer thickness is more than three times the pile diameter, the bearing capacity is
much higher than when the bearing layer thickness is the same as the pile diameter. In addition, soil deformation which occurs is almost
entirely in the bearing layer, and the changes in bearing capacity are hardly affected by the soft layer below the bearing layer. The exper-
imental findings obtained in the present study support the idea that the criterion for the bearing layer thickness, where the influence of a
thin bearing layer on the bearing capacity can be ignored, is three times the pile diameter, regardless of whether the pile tip is open or
closed.
Ó 2022 Production and hosting by Elsevier B.V. on behalf of The Japanese Geotechnical Society. This is an open access article under the CC BY-
NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Keywords: Thin bearing layer; Open-ended pile; Closed-ended pile; Pile-loading test; X-ray micro CT

1. Introduction piles), and are widely applied for both onshore and offshore
structures. Open-ended piles can be more easily penetrated
A pile foundation is used to support a structure that is than closed-ended piles, and steel pipe piles, which show
constructed on soft ground that cannot be supported by high horizontal resistance and rigidity, are suitable for deep
a shallow foundation. Bearing piles, which are designed construction works. In the design and construction of piles,
to provide a high bearing capacity, are deeply driven into it is important to accurately predict the ultimate bearing
the ground. They are classified as either closed-ended piles capacity. For this purpose, the bearing capacity character-
(e.g., cast-in-place piles) or open-ended piles (e.g., steel pipe istics of piles must be understood.
The bearing capacity characteristics of closed-ended
piles (e.g., Kim et al., 2009) and open-ended piles (e.g.,
Peer review under responsibility of The Japanese Geotechnical Society. Doherty and Gavin, 2011) have been investigated by in-
⇑ Corresponding author.
E-mail address: [email protected] (R. Kido). situ loading tests and laboratory tests. Especially in Japan,

https://doi.org/10.1016/j.sandf.2022.101179
0038-0806/Ó 2022 Production and hosting by Elsevier B.V. on behalf of The Japanese Geotechnical Society.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 R. Kido et al. Soils and Foundations 62 (2022) 101179

the focus of many studies is on evaluating he ultimate bear- bearing layer, and enable the establishment of a theory for
ing capacity of piles (e.g., Okahara et al., 1990), the charac- the ultimate bearing capacity. This will provide further
teristics of the bearing capacity and the settlement of open- rationalization of a practical design for pile foundations.
ended piles (e.g., Goto and Katsumi, 1967; Katsumi, 1984). The aim of the present study is to clarify the bearing
Some researchers have reported that the evolution tenden- mechanism of piles supported by a thin bearing layer.
cies of the frictional resistance and the base resistance The thin bearing layer discussed herein is defined as a bear-
between these two types of piles differ (Paik et al., 2003; ing layer with the same thickness as the diameter of the
Han et al., 2019). The main reason is that open-ended piles pile. For simplicity, the model grounds prepared in the pre-
exhibit the plugging effect, a unique feature that closed- sent study comprise three layers, all of which have different
ended piles do not exhibit. This phenomenon contributes bearing layer thicknesses. So that the soil deformation
to penetration resistance and bearing capacity (Paik and using X-ray micro CT which qualitatively simulates the
Salgado, 2003). Therefore, the mechanism of the plugging behavior of field-scale piles can be included in the discus-
phenomenon and its influence on the bearing capacity have sion, a stiff bearing layer and a soft non-bearing layer are
been discussed by conducting numerous pile-loading tests included, with poorly-graded sands of different relative
(e.g., Ko and Jeong, 2015; Kurashina et al., 2016). density values, Dr. Pile-loading tests are conducted under
Meyerhof (1951) and Vesic (1963) have proposed bear- isotropic confining pressure in order to demonstrate the
ing mechanisms for closed-ended piles. These mechanisms stress state in a deep foundation. The changes in soil den-
are often used to estimate the ultimate bearing capacity sity during the pile loading are visualized using X-ray
because they are more convenient and are more practical micro CT, and then the CT images are analyzed to quantify
than performing field-scale tests on piles. Their validity the range in soil displacement and the strain distributions
was clearly confirmed by Otani (2003), who applied X- in the model ground. The influence of the bearing layer
ray computed tomography (CT) to pile-loading tests. More thickness on the changes in bearing capacity and the differ-
recently, X-ray CT has been applied to soil-structure inter- ences in bearing mechanisms between closed-ended piles
action problems (e.g., Kikuchi and Morikawa, 2008; and open-ended piles are then determined.
Doreau-Malioche et al., 2018). The plugging mechanism
of open-ended piles has been investigated from a micro-
scopic viewpoint in laboratory experiments using transpar- 2. Experimental methods
ent soil and a PIV analysis (Zheng et al., 2020) and the
discrete element method (Liu et al., 2019; Li et al., 2021). 2.1. Pile-loading device with X-ray micro CT
Based on previous studies, it is important to understand
the soil deformation/failure characteristics under pile load- Fig. 1 shows the test apparatus. A part of the pullout
ing and to link them to the changes in bearing capacity in test apparatus for reinforcements (Kido et al., 2021) is
order to clarify the bearing mechanism of piles. modified in order to perform the pile-loading tests under
Piles are generally supported by a thick and stiff bearing any level of isotropic pressure. The upper part of the appa-
layer. However, it is expected that piles supported by a thin ratus consists of an acrylic triaxial cell, a load cell, a motor,
bearing layer will shorten the construction period and and a displacement gauge, while the lower part consists of a
improve the economic efficiency of the foundation works. base plate and pedestal. The cell is made of lucid acrylic,
Matsui and Oda (1991) reported that one of the most and uses 2 MPa in pressure resistance. This allows the X-
important characteristics of piles supported by a thin bear- ray attenuation to be reduced to a maximum degree, and
ing layer is that the bearing layer thickness should be at also ensures the axial load can be supported without steel
least three times the diameter of the pile as the criterion pillars (e.g., Otani et al., 2002). The motor controls the
for negligible influence of the thin bearing layer on the loading rate of the model pile from 0.1 mm/min to
bearing capacity. Previous centrifuge model tests (Horii 1.0 mm/min. The load cell is connected to a steel bar
and Nagao, 2018; Martinez et al., 2020) and laboratory 20 mm in diameter and 200 mm in length, and the tip of
tests using X-ray micro CT (Suezawa et al., 2020) have con- the steel bar is placed into a hollow part at the model pile
firmed that a lower bearing capacity was apparently head, allowing for the measurement of the pile head load.
observed when the bearing layer thickness was the same The X-ray micro CT facility used in the present study is
as the pile diameter. However, neither the mechanism for KYOTO-GEOlXCT (Higo et al., 2011). The X-ray
the reduction of the bearing capacity in a thin bearing absorption coefficient depends on the material density,
layer, nor the reason for deciding the criterion as three which is converted to a CT value. It is empirically known
times the pile diameter have been sufficiently clarified. Pre- that there is an almost linear relationship between the CT
vious studies on piles supported by a thin bearing layer value and the material density (Desrues et al., 1996). Each
have been conducted only on closed-ended piles. The bear- CT slice acquired in the present study consists of 10242
ing mechanism of open-ended piles has not been reported. voxels with each voxel presenting a gray value, i.e., a CT
An investigation of soil deformation characteristics will value with a 16-bit signed integer. Details of the X-ray
confirm the bearing mechanism of piles supported by a thin CT facility are given in Kido et al. (2020).

2
 R. Kido et al. Soils and Foundations 62 (2022) 101179

Fig. 1. Experimental device: (a) photograph and (b) schematic illustration.

Fig. 2. Schematic illustration of model pile: (a) close-ended pile and (b) open-ended pile.

2.2. Model piles the open-ended pile models used in previous laboratory
tests (e.g., Paik and Salgado, 2003; Doherty and Gavin,
Fig. 2 presents illustrations of the two model piles used 2011; Ko and Jeong, 2015; Kumara et al., 2015;
in the present study. For the convenience of X-ray CT Kurashina et al., 2016), 0.44–0.85. The wall thickness t is
scanning, each model pile is fabricated using ordinary 1 mm, and was designed to be the smallest thickness that
acrylic. The open-ended pile model is made of a combina- would prevent the acrylic model pile from buckling during
tion of hollow and solid parts to ensure that the inside of loading so that the value R as close as possible to that of
the model ground is exposed to atmospheric pressure, while actual piles.
the outside of the model ground is subjected to various
levels of confining pressure. 2.3. Materials
The degree of the opening of the open-ended pile, R, is
expressed by the following equation: The actual ground is comprised of various kinds of
pd 2  2 materials, including gravel, sandy soil, and soft clay. These
d materials are often in a saturated condition below the
R ¼ pD4 2 ¼ ð1Þ
4
D groundwater level. In our study, two kinds of uniform
poorly-graded sands (Toyoura sand and silica sand #5)
where d is the inner diameter and D is the outer diameter. were used as the ground materials under an air-dried con-
The open-ended pile model used in the present study has a dition. The present study focuses on visualizing and analyz-
D of 10 mm and d of 8 mm, resulting in an R of 0.64. ing the soil deformation characteristics around the pile tips
Although the value of R is smaller than that of typical using X-ray micro CT. It should be noted that it is possible
full-scale steel pipe piles, 0.84–0.97, it is similar to that of that the real soil-plug effect may not be obtainable ideal
3
 R. Kido et al. Soils and Foundations 62 (2022) 101179

Table 1
Physical properties of sand materials.
Material Silica sand #5 Toyoura sand
Specific gravity Gs 2.64 2.64
Average particle diameter D50 [mm] 0.438 0.20
Internal friction angle (peak) [deg] 41.4 38.9
Internal friction angle (residual stress) [deg] 33.0 33.4
Internal friction angle (sand – acrylic) [deg] – 20.3
Viscosity c [kpa] 0.0 0.0
Maximum void ratio emax 1.04 0.975
Minimum void ratio emin 0.705 0.585

140 mm. For the closed-ended pile, there are three cases
with different bearing layer thicknesses: seven times the pile
diameter (Case C-7D), three times the pile diameter (Case
C-3 D), and one times the pile diameter (Case C-1 D). In
Case C-7 D, the entire ground below the pile tip is the bear-
ing layer. The initial position of the pile tip is aligned with
the surface of the bearing layer for all three cases. For the
open-ended pile, there are three cases with different bearing
layer thicknesses: eight times the pile diameter (Case O-8
D), three times (Case O-3 D) the pile diameter, and one
times the pile diameter (Case O-1 D). The initial position
of the pile tip is 20 mm higher than the surface of the bear-
ing layer.
To experimentally compare the soil deformation charac-
teristics during pile loading when a bearing layer thickness
Fig. 3. Grain size distribution curve.
is different, the model ground must have a clear difference
in bearing capacity. Kurashina et al. (2016) confirmed that
samples and small pile models are used. The physical prop- higher bearing capacity was obtained during the loading of
erties of the samples are listed in Table 1, and the grain size open-ended piles into air-dried sandy ground with a higher
distribution curves are shown in Fig. 3. value of Dr. Based on this result, the best method for
preparing model grounds in the present study was to ensure
2.4. Model grounds the bearing layer was as dense as possible, and the non-
bearing layer was as loose as possible. It was found from
Fig. 4 shows schematic diagrams of the test cases. The preliminary investigations that the maximum value of a
model grounds had a diameter of 70 mm and a height of Dr of 85% and the minimum value of a Dr of 30% were fea-

Fig. 4. Experimental cases: (a) closed-ended pile and (b) open-ended pile.
4
 R. Kido et al. Soils and Foundations 62 (2022) 101179

sible for both the silica sand and Toyoura sand when below reveal that no soil deformation reached the mem-
preparing model grounds. Therefore, the former value brane boundary of each model ground during pile loading.
was used for the bearing layer and the latter value was used This is evidence that the pile loading tests conducted in the
for the non-bearing layer, respectively. present study is a model test that is not affected by the scale
Two preliminary loading tests were conducted for the effect.
closed-ended pile to determine which samples should be The small scale of the model does not allow the quanti-
used for the bearing and non-bearing layers, respectively. tative simulation of actual pile behaviors in terms of the
Two model grounds (one for the Toyoura sand and the bearing capacity level and soil-plug phenomenon. The pre-
other for the silica sand #5) were prepared with a Dr of sent study focuses on the qualitative interpretation of the
30%. The loading rate was 0.5 mm/min and the confining behaviors of both the pile and soil by observing the differ-
pressure was 100 kPa. When the confining pressure in the ence between the bearing capacity and soil deformation
triaxial cell is higher than 100 kPa, some of the air pressure when the bearing-layer thickness is varied. The test results
will leak from the cell, even after it has been sealed by a top described below confirm that this objective has been
cover (Fig. 1b). Considering that bearing layers are located achieved.
several meters to several tens of meters below the ground
level, pile loading tests should be conducted under as large 2.6. Preliminary tests for open-ended pile
a level of confining pressure as possible in order to demon-
strate the stress state around the bearing layer in the field. Considering the actual construction of piles, an open-
In the present study, therefore, a confining pressure of ended pile penetrated from the surface layer will pass
100 kPa was applied. The load–displacement curves through the non-bearing layer before reaching the bearing
(Fig. 5) indicate that the silica sand has a higher load than layer deeper underground. Therefore, a certain amount of
the Toyoura sand, so the silica sand with a Dr of 85% soil will enter into the pile during penetration of the pile,
should exhibit a higher bearing capacity and greater stiff- and then a soil plug might appear. The reproduction of this
ness than the Toyoura sand with a Dr of 30%. It is certain, event makes it easier to compare the open-ended pile and
therefore, that the silica sand is used for the bearing layer the closed-ended pile: that is, the soil behavior at locations
with and the Toyoura sand is used for the non-bearing deeper than the surface of the bearing layer and its influ-
layer, and that the bearing layer is sufficiently stiff com- ence on the bearing capacity can be more clearly compared.
pared to the non-bearing layer. Therefore, a preliminary loading test on the open-ended
pile was performed with a model ground using Toyoura
2.5. Scale effect of models on experiment sand with Dr of 30% to investigate the relationship between
the displacement level and the degree of plugging. The pile
Schnaid and Houlsby (1991) reported that the ratio of penetration of 30 mm was applied under 100 kPa as the
the diameter of a model ground to the diameter of a model confining pressure and 0.5 mm/min as the loading rate.
pile should be higher than 50. In this study, the ratio is 7. The degree of plugging for an open-ended pile is evalu-
This ratio was included in the design to ensure that the ated by the incremental filling ratio (IFR), i.e., the ratio of
model ground size would be as large as possible, with the incremental soil length in the pipe, DL, to the incremen-
two factors taken into account: the weight limitation of tal displacement, Dd (e.g., Paikowsky et al., 1989). The
the X-ray turn table and sufficient X-ray transmission fully plugged state is indicated by an IFR of 0, while the
energy to obtain clear images. Due to the small scale of partially plugged state is indicated by an IFR of 0 to 1.
the model ground, the boundary conditions could have The IFR is calculated by measuring the soil length in the
affected the test results. The X-ray CT images described pipe from the CT images. The open-ended pile used in
the present study is usually settled by 3 mm (l0) when
installed on the soft layer at the initial state. Therefore,
at each degree of indentation (d/D), the net soil length in
the pipe (L) is firstly calculated by subtracting l0 from the
soil length in the pipe (l), and then the ratio of the incre-
mental net soil length (DL) to the incremental indentation
(Dd) between two degrees of indentation is calculated as
the IFR.
Fig. 6 shows schematic diagrams of how to calculate the
IFR and the changes in IFR obtained in the preliminary
test. The IFR becomes less than 1 with the indentation dis-
placement of 20 mm, indicating the partially plugged state.
It seems desirable to have an indentation displacement of
10 mm, equal to the pile diameter D, from the surface of
the bearing layer in order to compare the results between
Fig. 5. Load-displacement curve. the open- and closed-ended piles. Therefore, the initial
5
 R. Kido et al. Soils and Foundations 62 (2022) 101179

Fig. 6. Increment of soil height inside open-ended pile calculated by CT image.

position of the tip of the open-ended pile was determined to 2. Pour the Toyoura sand and the silica sand #5 into the
be 20 mm higher than the surface of the bearing layer. mold by the air-pluviation technique using a funnel
(Fig. 7a). The former should be poured slowly from a
2.7. Test method low position without any vibration to create as loose a
layer as possible. The latter should be poured rapidly
Fig. 7 shows a schematic diagram of the experimental from as high a position as possible to create a dense
procedure. Details of the procedure are as follows: layer. Each material is to be poured into 10 equal layers,
and the positions where each material is poured must be
1. Fasten a 0.3-mm-thick membrane onto the pedestal determined by calibration works such that Dr for the
placed on the base plate. Once the mold is attached, bearing layer becomes 85% and that for the non-
apply negative pressure to tightly fix the membrane onto bearing layer becomes 30%.
the mold. 3. Attach a jig to the top of the mold and place the model
pile vertically on the surface of the bearing layer
(Fig. 7b). At this time, pass four rubber sheets,
0.3 mm in thickness, through the model pile to improve
the airtightness of the model ground. After that, create
the upper non-bearing layer with the Toyoura sand
according to the method given in Step 2).
4. Remove the jig and place the four rubber sheets on the
surface of the non-bearing layer. Then, attach an alu-
minum cap to the non-bearing layer to push the model
pile vertically (Fig. 7c).
5. Wrap the membrane around the aluminum cap and fix it
with an O-ring. After that, remove the mold by turning
off the negative pressure acting on the mold. Then, apply
negative pressure to the model ground to make it stand
vertically (Fig. 7d).
6. Cover the acrylic cell with the base plate and adjust the
position of the stainless-steel bar such that the steel bar
fits into the hollow part of the model pile.
7. Seal the cell by fixing it to the base plate using screws
(Fig. 7e). Turn off the negative pressure being applied
to hold the model ground. Maintain a difference of
100 kPa between the cell pressure and the pressure in
the model ground.

The loading rate is 0.5 mm/min. The following parame-

Fig. 7. Procedure of model ground preparation. ters are measured: pile head load, indentation displace-
6
 R. Kido et al. Soils and Foundations 62 (2022) 101179

ment, cell pressure, and model ground pressure, each of almost consistent. This trend is consistent with the results
which is measured every second. The total indentation dis- of previous studies (e.g., Horii and Nagao, 2018;
placement is 10 mm for the closed-ended pile and 30 mm Martinez et al., 2020; Matsui and Oda, 1991), indicating
for the open-ended pile. that the results obtained in the present study represent sim-
ilar properties to the behavior of actual piles.
3. Results Fig. 9c confirms that the bearing capacity of the open-
ended pile varies in three stages: (I) the stage in which
3.1. Load-displacement curves the soil continues to penetrate into the inner pipe (Stage
1), (II) the stage in which the plugging phenomenon contin-
Figs. 8 and 9 show the load–displacement curves for the ues to occur until the pile tip reaches the surface of the
closed- and open-ended piles, respectively. One loading test bearing layer (Stage 2), and (III) the stage in which the pile
with CT scanning is performed for each case. In each of tip penetrates the bearing layer (Stage 3). The bearing
these, load relaxation is observed at some degrees of inden- capacity tends to be similar or slightly lower in Case O-
tation because pile loading is temporarily suspended during 1D than in Cases O-3D and O-8D, starting in Stage 2. In
the CT scan. These figures confirm the high reproducibility Stage 3, the bearing capacity is very low in Case O-1D.
of the load–displacement relationship, including the test These results indicate that the bearing capacity is higher
with CT scanning. Fig. 8 also includes a plot of the incre- when the bearing layer thickness is larger, regardless of
mental load per displacement of 1 mm, calculated by divid- the open/closed conditions at the pile tip, and that the cri-
ing the incremental load by the incremental indentation terion for sufficient bearing layer thickness is at least three
displacement during one second. The convergent displace- times the pile diameter.
ment and convergent load shown in Fig. 8a are defined as
the indentation displacement and the load at which the 3.2. CT images
incremental load reaches zero, respectively.
Fig. 8 shows that, for any case, the load increases lin- Fig. 10 shows the CT scan volume. The voxel size is
early until the d/D reaches 0.05. The convergent load is 48.12  53.0 lm3. Fig. 11 shows the vertical cross-
smaller in the order of Cases C-7D, C-3 D, and C-1 D. sectional slices at the initial and last states of the pile-
The levels of load between Cases C-7D and C-3D are loading tests. In the CT images, the high-density regions

Fig. 8. Load-displacement relationships of closed-ended pile: (a) Case C-7D, (b) Case C-3D and (c) Case C-1D.

Fig. 9. Load-displacement relationships of open-ended pile: (a) Case O-8D, (b) Case O-3D and (c) Case O-1D.
7
 R. Kido et al. Soils and Foundations 62 (2022) 101179

Fig. 10. Scan volume in model ground for (a) close-ended model and (b) open-ended pile.

are indicated in white, while the low-density regions are Case O-1D is reduced beyond the d/D of 2.30 (Fig. 9),
indicated in black. Fig. 12a and 12b describe the gray value which is inconsistent with the above assumption.
histograms for the bearing and non-bearing layers, respec-
tively. The pixels for each layer in the CT images at the ini-
tial state are extracted, and then the number of pixels
showing each gray value is normalized by the total number
of pixels for the probability density. Note that each gray
value has a positive correlation with the material density;
hence, Fig. 12 expresses the density distributions in both
layers. It is found that the density distribution in each layer
is similar to those of the other cases, which is evidence of
the uniformity of the density of each model ground.
At the last state for Cases C-7D and 3D (Fig. 11), a
wedge-shaped higher-density region develops downward
from the base of the pile, and the surrounding region
appears to be of lower density. The same trend is observed
in Case C-1D; however, the soil structure is disturbed in
both the bearing and non-bearing layers. At the last state
for the open-ended pile, the soil around the pile tip
becomes dense as in the case of the closed-ended pile. It
is probable, therefore, that the bearing capacity even for
the open-ended pile is enhanced by the reaction force from
the densified ground around the pile tip, and is independent
of the bearing layer thickness.
Fig. 13 presents the variations in the IFR and net soil
length L in the open-ended pile. The IFR decreases rapidly
to less than 1 beyond the d/D of 1.70, where the pile tip
begins to penetrate the bearing layer. The IFR for each
case tends to decrease even further as the d/D increases,
with Case O-1D showing the lowest IFR beyond the d/D
of 2.10. In other words, the amount of soil penetration in
the pile is the smallest. According to these results, it can
be assumed that the plugging effect contributes to an
increase in both inner frictional resistance and base resis-
tance even when the bearing layer thickness is the same Fig. 11. Vertical cross section of CT images: (a) close-ended pile and (b)
as the pile diameter. Nevertheless, the bearing capacity of open-ended pile.
8
 R. Kido et al. Soils and Foundations 62 (2022) 101179

Fig. 12. Gray value histogram: (a) for bearing layer and (b) for non-bearing layer. The symbols ‘‘L” and ‘‘U” indicate lower non-bearing layer and upper
non-bearing layer, respectively.

the pile indentation due to the high compressibility of the

lower non-bearing layer. At the d/D of 3.00, the soil density
for Case O-1D decreases over a wide area of the upper non-
bearing layer in the pile (depth of 2.0–2.5 in Fig. 14b). It is
probable, therefore, that when an open-ended pile is sup-
ported by a thin bearing layer, the contribution to the inner
frictional resistance and base resistance is small because of
the lower soil density in the pile and the large downward
settlement. To clarify this mechanism in more detail, it
would be effective to simulate the particle-scale behavior
of the plugging phenomenon for model grounds with differ-
ent bearing layer thicknesses using, for example, the dis-
crete element method.
Fig. 13. Changes in IFR and net length L in open-ended pile.

3.3. Soil deformation quantified by image processing

The reason for this inconsistency is considered in terms 3.3.1. Digital image correlation
of the gray value distributions in each pile, which are The deformation behavior in the model ground is ana-
shown in Fig. 14. A search region, 7.36 mm in width and lyzed by digital image correlation (DIC) using CT images.
3.23 mm in height (153  61 voxels), is applied to the soil To improve the accuracy of the calculated displacement
area of the pile in the 2-dimensional CT slice. The median fields, subvoxel refinement is also applied. Details of the
of the gray value distributions for all the voxels occupying algorithms used in this study can be found in Higo et al.
the search region is evaluated. 560 voxels are finally ana- (2013) and Kido (2019), respectively. In the present study,
lyzed to evaluate the vertical distribution of the soil density the displacement in the pile region was set to be zero so that
in the open-ended pile by moving the search region by one the displacement field at the ground with that at the pile
voxel from the top to the bottom of the soil in the pile. The can be identified. For this purpose, voxels for the pile were
reason for using the median of the gray value rather than extracted using a region-growing method (e.g., Kido et al.,
the mean value is to avoid overestimating the soil density 2020) and then a gray value which was sufficiently larger
due to the presence of very high-density material, such as than the gray values for the ground, was added to the
the iron particles contained in Toyoura sand and silica extracted voxels prior to DIC.
sand.
At the initial state, the gray value of the bearing layer in 3.3.2. Displacement vectors
both cases is higher than that of the upper non-bearing Fig. 15 shows distributions of the incremental displace-
layer. When the d/D reaches 2.50, the gray value near the ment vectors. The white line in the figure indicates the
pile tip (depth of 2.50 in Fig. 14) is the highest. As the d/ boundary between the bearing and the non-bearing layers.
D reaches 3.00, the position of the highest gray value For both types of piles, the displacement region extends
moves downward in both cases; however, the degree of bulbously downward from the pile base, and the displace-
the movement in Case O-1D is much larger (black arrows ment near the pile base is the largest. Fig. 15a clearly indi-
in Fig. 14). This is probably because besides the soil in cates that the displacement region in Case C-1D extends
the pile, the soil deeper than the pile tip also sinks following downward more than that in the other two cases, penetrat-
9
 R. Kido et al. Soils and Foundations 62 (2022) 101179

Fig. 14. Gray value distributions of soil inside open-ended pile: (a) Case O-8D and (b) Case O-1D.

ing into the non-bearing layer. In the case of Case C-3D, downward is that the sand particles along the pile shaft and
the displacement region is situated almost entirely in the those inside the pile follow the displacement of the pile in a
bearing layer. vertical direction due to the partially plugged condition.
Fig. 15b shows that the sand in the open-ended pile Fig. 15b also indicates that the displacement level of soils
tends to be displaced downward, indicating that the sand near the pile tip and those in the pile for Case O-1D are sig-
particles themselves do not penetrate into the pile. In cases nificantly larger than that for Cases O-8D and 3D. This
where the IFR is less than 1 between the displacement from tendency corresponds to the results where the soil level in
2.10D to 2.30D, the reason the sand in the pile is displaced the pile for Case O-1D is the lowest (section 3.2). It also

10
 R. Kido et al. Soils and Foundations 62 (2022) 101179

Fig. 15. Displacement vectors calculated by DIC: (a) close-ended pile and (b) open-ended pile.

supports the idea that not only the soil in the pile, but also the edge of the pile base and the shear deformation extends
the soil deeper than the pile tip, sinks following the pile to a lower portion, without a clear wedge region. This ten-
indentation into the bearing layer when the bearing layer dency for Case C-1D probably corresponds to the punch-
is thin. ing failure phenomenon confirmed by Matsui and Oda
(1991). The volumetric strain shows expansion in the
3.3.3. Shear strain and volumetric strain region of the large shear strain and compression at the pile
Figs. 16 and 17 show the distributions of shear strain base (Fig. 17).
and volumetric strain, respectively. The strain is calculated Fig. 16 indicates that large shear strain develops near the
using the B-matrix for the 8-node finite element, and the side of the open-ended pile. This may be attributed to the
second invariant of the deviatoric-strain tensor is defined soil deformation related to the frictional resistance, or
as the shear strain. The volumetric strain is positive for may be due to the low accuracy of the analysis around
compression and negative for expansion. Because the the soil-pile interface due to the spatial resolution of CT
results of Case C-7D and Case O-8D are similar to those images. For this reason, focus is mainly placed on the soil
of Case C-3D and Case O-3D, those of the former two deformation deeper than the pile tip. At the d/D between
cases are not shown in Fig. 17. In Fig. 16, there is no shear 1.70 and 1.90, large shear strain develops downward from
strain around the left side of the model pile for Cases O-8D the pile tip, where the expansive strain occurs. For Case O-
and O-3D. This is probably because the strain is calculated 3D in Fig. 17, a blue line (expansive strain region) exists in
based on zero displacements at the pile and the stationary the ground between the two pile tips just above the white
ground. Since outer skin friction should occur during pile line (the boundary between the upper non-bearing and
loading, zero-shear strain seems odd. The accuracy of the bearing layers). A similar tendency for the strain distribu-
image correction (Section 3.1.1) probably affects the tions was confirmed by Kikuchi and Morikawa (2008).
results. Even so, these results are less significant for the Therefore, the region around the pile tip is in a plugged
main findings and conclusions in the present study. state. On the other hand, as the pile indentation increases,
In Cases C-7D and C-3D, shear strain develops diago- the expansive region disappears and the soil in the bearing
nally from both ends of the pile base, and a wedge- layer probably begins to penetrate the pile, resulting in the
shaped shear-strain region is finally generated. The wedge unplugged state. It is probable, therefore, that the open-
region corresponds to the high-density region of the ended pile exerts bearing capacity due to the alternating
ground (Fig. 11a) at the d/D of 0.80. In contrast, Case C- repetition of the plugged and unplugged states in the
1D shows large shear strains vertically downward from ground near the pile tip, regardless of the bearing layer
11
 R. Kido et al. Soils and Foundations 62 (2022) 101179

Fig. 16. Distributions of shear strain calculated by DIC for: (a) closed-ended pile and (b) open-ended pile.

Fig. 17. Distributions of volumetric strain calculated by DIC for: (a) closed-ended pile and (b) open-ended pile.

thickness. The shear strain near the pile tip at the d/D 4. Discussions
between 2.10 and 2.30 indicates that the amounts of large
shear strain generated from the two pile tips are related 4.1. Bearing mechanism of open-ended pile supported by thin
to each other in all cases, and their tendency is the most bearing layer
pronounced in Case O-1D. In addition, expansive strain
develops in the ground where the large shear strain occurs Fig. 18 shows a schematic diagram of characteristics of
(Fig. 17), indicating positive dilatancy due to shear defor- bearing capacity and soil deformation for an open-ended
mation. These tendencies and the displacement field shown pile supported by a thin bearing layer. Each bracketed
in Fig. 15b strongly support the idea that Case O-1D shows number shown in Fig. 18c corresponds to the numbers
more pronounced shear deformation in both the bearing shown in Fig. 18a and b.
and the lower non-bearing layers than Case O-8D or Case When the open-ended pile begins to penetrate the
3D. ground, the external friction and annulus resistance of
12
 R. Kido et al. Soils and Foundations 62 (2022) 101179

Fig. 18. Distributions of volumetric strain calculated by DIC for: (a) closed-ended pile and (b) open-ended pile.

the pile are exerted first. As the pile indentation is On the other hand, the bearing layer thickness affects the
increased, the level of the soil in the pile increases. The sur- range and magnitude of the soil deformation, causing the
face area of the soil in contact with the inner pile surface is difference in bearing capacity. When the pile tip approaches
increased, which likely generates larger frictional resis- the surface of the bearing layer (Stage 2 in Fig. 18c), the
tance. This idea is supported by the results obtained by level of the soil in the pile is lower when the bearing layer
Ko and Jeong (2015). As the pile tip approaches the surface thickness is the same as the pile diameter than when the
of the bearing layer, the ground near the pile tip becomes bearing layer thickness is more than three times the pile
denser and the plugging near the pile tip progresses, result- diameter. In that case, Case O-1D already shows a lower
ing in increases in the internal friction and base resistance. bearing capacity than the other cases (Fig. 18c). That is,
This enhances the overall bearing capacity. After the pile the inner friction seems less significant due to the lower
tip reaches the bearing layer, the soil in the bearing layer level of the soil in the pile when a soft layer exists below
begins to penetrate the pipe, and then plugging and unplug- a thin bearing layer. When the pile tip penetrates the bear-
ging alternately occur. Shear deformation progresses dur- ing layer (Stage 3 in Fig. 18c), the ground near the pile tip
ing this process, forming a wedge-shaped region from the can be densified even for the case of a thin bearing layer.
pile tip to the deeper ground (Fig. 18b), where positive dila- However, the pile is displaced significantly in a downward
tancy occurs and pushes out the surrounding soil. The direction, causing the plastic deformation to reach the soft
above tendency probably indicates that the inner friction layer below the thin bearing layer, which results in lower
and base resistance related to the ground reaction force base resistance. Since the soil density in the pile is lower
make a particular contribution to the changes in bearing during the pile penetration for the thin bearing layer, the
capacity. Moreover, based on the results which show that inner friction is also lower. For these reasons, an open-
the trend in load variation is similar for all three cases ended pile supported by a thin bearing layer has a much
(Fig. 18c), the bearing mechanism of the open-ended pile lower bearing capacity than that supported by a sufficiently
is basically the same for all cases, although the bearing thick bearing layer (Fig. 18c).
layer thickness differs among them.
13
 R. Kido et al. Soils and Foundations 62 (2022) 101179

4.2. Comparison of bearing mechanisms between closed- and can be obtained is narrower for the open-ended pile than
open-ended piles for the closed-ended pile, resulting in a lower contribution
to the bearing capacity. This tendency is independent of the
Fig. 19 shows a comparison of the characteristic DIC bearing layer thickness.
results for the closed- and open-ended piles. The results Secondly, focus is placed on the region where shear
of the DIC given here represent the soil deformation during strain develops. When the bearing layer is sufficiently thick,
pile indentation from the surface of the bearing layer to the the closed-end pile exhibits a clear wedge-shaped region
d/D of 0.200. related to the large shear deformation, while the open-
Firstly, focus is placed on the range in soil displacement. ended pile does not show such a region. While the range
The tendency of the soil displacement to expand bulbously in which the base resistance develops for the open-ended
from the pile tip is the same for both closed- and open- pile is not as extensive as that for the closed-ended pile,
ended piles, whereas the soil displacement for the closed- the open-ended pile exerts not only internal friction during
ended pile expands more widely in the horizontal direction pile indentation, but also base resistance related to soil den-
than the open-ended pile. This is probably because the sification. When the bearing layer is thin, large shear strain
closed-ended pile pushes out a greater amount of soil due develops downward at the edge of the closed-ended pile
to its larger cross-sectional area than the open-ended pile, without such a wedge. That is, punching failure occurs.
and the soil tends to penetrate the open-ended pile more In the case of the open-ended pile, a clear wedge is shown.
deeply. Namely, the range in which the base resistance In the case of an open-ended pile supported by a thin bear-

Fig. 19. Comparison of soil deformation characteristics of closed- and open-ended piles.
14
 R. Kido et al. Soils and Foundations 62 (2022) 101179

ing layer, the soil below the pile tip and that in the pile lar- Then, the soil density in the pile becomes lower than in
gely sink downward, causing decreases in both inner fric- the case for which the bearing layer is sufficiently thick,
tion and base resistance. However, Figs. 8 and 9 indicate which probably causes lower inner resistance and base
that the load levels for the closed- and open-ended piles resistance, resulting in a low bearing capacity.
are almost equivalent to each other when compared with 3. Regardless of the bearing layer thickness, the closed-end
the same bearing layer thickness. Considering that the pile tends to yield the base resistance in a wider range of
open-ended pile shows a narrower range for obtaining the ground and exhibits a higher bearing capacity than
the base resistance compared to the closed-ended pile, it the open-ended pile.
is possible that the inner friction of the open-ended pile 4. When the bearing layer thickness is more than three
contributes to the bearing capacity more effectively than times the pile diameter, the bearing capacity is much
the base resistance. It is also possible, however, that the dif- higher than when the bearing layer thickness is the same
ference in bearing capacity between the closed- and open- as the pile diameter for both the closed- and open-ended
ended piles is not generated because of the small dimension piles. In addition, the soil deformation occurs almost
of the model piles used in the present study. In future work, entirely in the bearing layer and the changes in bearing
it is important to conduct various experiments with model capacity are hardly affected by the soft layer below the
piles of different scales and to compare the results acquired bearing layer. These experimental findings support the
from them. idea that the criterion for the bearing layer thickness,
The results obtained in the present study indicate that where the influence of a thin bearing layer on the bear-
both closed- and open-ended piles can be considered ing capacity can be ignored, is three times the pile diam-
well-supported piles as long as the bearing layer thickness eter, regardless of whether the pile tip is open or closed.
is more than three times the pile diameter. The reason is
that soil deformation occurs almost entirely in the bearing
layer and that the change in bearing capacity is hardly Declaration of Competing Interest
affected by the soft layer below the bearing layer.
The authors declare that they have no known competing
5. Conclusions financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
The purpose of the present study was to clarify the bear-
ing mechanism of piles supported by a thin bearing layer.
For this purpose, model grounds with different bearing References
layer thicknesses were prepared for pile loading tests using
closed- and open-ended model piles, and the changes in soil Desrues, J., Chambon, R., Mokni, M., Mazerolle, F., 1996. Void ratio
evolution inside shear bands in triaxial sand specimens studied by
density during the pile loading were visualized using X-ray
computed tomography. Géotechnique 46 (3), 529–546.
micro CT. Pile-loading tests were conducted under isotro- Doreau-Malioche, J., Combe, G., Viggiani, G., Toni, J.B., 2018. Shaft
pic confining pressure to demonstrate the stress state in a friction changes for cyclically loaded displacement pile: An X-ray
deep foundation. Although the model grounds were pre- investigation. Géotech. Lett. 8 (1), 66–72.
pared using air-dried poorly graded sands, and the sizes Doherty, P., Gavin, K., 2011. Shaft capacity of open-ended piles in clay. J.
Geotech. Geoenviron. Eng. 137 (11), 1090–1102.
of both the grounds and the piles were much smaller than
Goto, H., Katsumi, T., 1967. Fundamental studies on settlements of large
those of actual conditions, the results obtained in the pre- diameter steel pipe piles. Trans. Jpn. Soc. Civil Engineers 1967 (138),
sent study were qualitatively consistent with those obtained 1–10.
in previous studies on field-scale piles supported by a thin Han, F., Ganju, E., Salgado, R., Prezzi, M., 2019. Comparison of the load
bearing layer. That is, the soil deformation characteristics response of closed-ended and open-ended pipe piles driven in gravelly
sand. Acta Geotech. 14 (6), 1785–1803.
newly investigated in the present study confirm the bearing
Higo, Y., Oka, F., Kimoto, S., Sanagawa, T., Matsushima, Y., 2011.
capacity mechanism of piles supported by a thin bearing Study of strain localization and microstructural changes in partially
layer, despite the small-scale of the model tests. The main saturated sand during triaxial tests using microfocus X-ray CT. Soils
findings of this study are listed below. Found. 51 (1), 95–111.
Higo, Y., Oka, F., Sato, T., Matsushima, Y., Kimoto, S., 2013.
Investigation of localized deformation in partially saturated sand
1. In the case of a thin bearing layer, the high compressibil-
under triaxial compression using microfocus X-ray CT with digital
ity of the soft layer below the bearing layer causes a image correlation. Soils Found. 53 (2), 181–198.
large downward displacement of the soil near the pile Horii, Y., Nagao, T., 2018. Centrifuge modelling of non-displacement
tip. The difference between closed- and open-ended piles piles on a thin bearing layer overlying a clay layer. Phys. Modell.
is that the former exhibits punching failure, while the Geotech., 1371–1376
Katsumi, T., 1984. Studies on stresses and deformations of soils in pipe
latter exhibits a clear wedge related to large shear
piles by finite element analysis. J. Jpn. Soc. Soil Mech. Found. Eng. 24
deformation. (3), 151–158, in Japanese.
2. When an open-ended pile is penetrated into a thin- Kido, R., 2019. Microscopic characteristics of partially saturated soil and
bearing layer, the soil in the pile is displaced downward their link to macroscopic responses. Doctoral dissertation, Kyoto
following the soil displacement deeper than the pile tip. University, 2019.

15
 R. Kido et al. Soils and Foundations 62 (2022) 101179

Kido, R., Higo, Y., Takamura, F., Morishita, R., Khaddour, G., Salager, Matsui, T., Oda, K., 1991. End bearing mechanism of bored pile on thin
S., 2020. Morphological transitions for pore water and pore air during layer. In: Proc. of the 9th Asian Regional Conference on Soil
drying and wetting processes in partially saturated sand. Acta Mechanics and Foundation Engineering, 1, 251–254.
Geotech. 15 (7), 1745–1761. Meyerhof, G.G., 1951. The ultimate bearing capacity of foundations.
Kido, R., Sawamura, Y., Kimura, K., Kimura, M., 2021. Investigation of Geotechnique 2 (4), 301–332.
soil deformation characteristics during pullout of a ribbed reinforce- Okahara, M., Nakatani, S., Taguchi, K., Matsui, K., 1990. A study on
ment using X-ray micro CT. Soils Found. 61 (3), 642–657. vertical bearing characteristics of piles. J. Jpn. Soc. Civil Eng. 1990
Kikuchi, Y., Morikawa, Y., 2008. Plugging mechanism in a vertically (418), 257–266.
loaded open-ended pile. In: Proc. or the Second BGA International Otani, J., 2003. State of the art report on geotechnical X-ray CT research
Conference on Foundations. 1, 169–180. at Kumamoto University. In: Proc. of the International Workshop on
Kim, D., Bica, A.V., Salgado, R., Prezzi, M., Lee, W., 2009. Load testing X-ray CT for Geomaterials, GeoX2003, 43–77.
of a closed-ended pipe pile driven in multilayered soil. J. Geotech. Otani, J., Mukunoki, T., Obara, Y., 2002. Characterization of failure in
Geoenviron. Eng. 135 (4), 463–473. sand under triaxial compression using an industrial X-ray CT scanner.
Ko, J., Jeong, S., 2015. Plugging effect of open-ended piles in sandy soil. Int. J. Phys. Modell. Geotech. 2 (1), 15–22.
Can. Geotech. J. 52 (5), 535–547. Paik, K., Salgado, R., 2003. Determination of bearing capacity of open-
Kumara, J.J., Kurashina, T., Yajima, T., Kikuchi, Y., 2015. Undestand- ended piles in sand. J. Geotech. Geoenviron. Eng. 129 (1), 46–57.
ing inner friction mechanism of open-ended piles - an experimental Paik, K., Salgado, R., Lee, J., Kim, B., 2003. Behavior of open- and
study. Jpn Geotech Soc Spec Publ 2 (37), 1333–1338. closed-ended piles driven into sands. J. Geotech. Geoenviron. Eng. 129
Kurashina, T., Kikuchi, Y., Hyodo, T., Janaka J. Kumara, Nakano, A., (4), 296–306.
Hara, H., 2016. Effect of the relative density to the mechanism of the Paikowsky, S.G., Whitman, R.V., Baligh, M.M., 1989. A new look at the
end bearing capacity development of the open ended piles. J. Jpn. Soc. phenomenon of offshore pile plugging. Mar. Geotechnol. 8 (3), 213–
Civil Engineers (Coastal Eng.) 72 (2), I_390–I_395. 230.
Li, L., Wu, W., Liu, H., Lehane, B., 2021. DEM analysis of the plugging Schnaid, F., Houlsby, G.T., 1991. An assessment of chamber size effects in
effect of open-ended pile during the installation process. Ocean Eng. the calibration of in situ tests in sand. Géotechnique 41 (3), 437–445.
220, 1–15. Suezawa, R., Kido, R., Sawamura, Y., Kimura, M., 2020. Influence of
Liu, J., Duan, N., Cui, L., Zhu, N., 2019. DEM investigation of bearing layer thickness on vertical bearing capacity of pile and soil
installation responses of jacked open-ended piles. Acta Geotech. 14 (6), deformation behavior. J. Jpn. Soc. Civil Eng. Ser. C (Geosphere Eng.)
1805–1819. 76 (4), 429–440, in Japanese.
Martinez, I.G., Kido, R., Sawamura, Y., Kimura, M., 2020. Centrifuge Vesic, A.B., 1963. Bearing capacity of deep foundations in sand. Highway
model tests on bearing characteristics of a pile supported by a thin Res. Rec. 39, 112–153.
sand layer. Proc. of the Japan Society of Civil Engineers 2020 Annual Zheng, J.-H., Qi, C.-G., Zhao, X., Wang, X.-Q., Shan, Y.-L., 2020.
Meeting, p. CS2-07.. Experimental simulation on open-ended pipe pile penetration using
transparent granule. KSCE J. Civ. Eng. 24 (8), 2281–2292.

16