Applied Acoustics 139 (2018) 189–202

Contents lists available at ScienceDirect

Applied Acoustics
journal homepage: www.elsevier.com/locate/apacoust

Feasibility study of using acoustic emission signals for investigation of pile T

spacing effect on group pile behavior
⁎
Wuwei Maoa,b, , Shogo Aoyamac, Ikuo Towhatad
a
Department of Geotechnical Engineering, College of Civil Engineering, Tongji University, Shanghai 200092, China
b
Key Laboratory of Geotechnical and Underground Engineering, Ministry of Education, Tongji University, Shanghai 200092, China
c
Kiso – Jiban Consultants Co., Ltd., Tokyo, Japan
d
Department of Civil Engineering, University of Tokyo, Tokyo, Japan

A R T I C LE I N FO A B S T R A C T

Keywords: Understanding the subsoil behavior subjected to group pile loading is of foundation important to clarify the
Acoustic emission group pile behavior. The acoustic emission (AE) signals produced by the stressed subsoil subjected to group pile
Sand loading are expected to contain useful information about the behavior of group pile. In this study, measurements
Group pile of AE signals generated during group pile loading were performed. To evaluate the effect of pile spacing, group
Pile spacing
pile loading tests were performed with two different pile spacing to represent strong and weak group effects.
Frequency
Results showed that the group pile in case of narrow spacing might significantly reduce the apparent stiffness of
the ground, which was evidenced by lower values of ground secant modulus and the released AE energy. In
addition, center pile carried more load and emitted higher level of AE compared with other pile positions when
the pile spacing was narrow. However, it is observed that such group pile effect was affected by the loading
history as well. Considering the high consistency between the results of conventional load-settlement mea-
surement and the AE monitoring, it is therefore suggested that the AE monitoring can be potentially used as an
alternative method for evaluation of the group pile behavior.

1. Introduction increasing demand on further understanding of pile bearing behavior,

Group pile has been a common design in modern infrastructure
constructions. It is well acknowledged that the capacity of group pile
depends on the individual pile capacities, and is influenced by the
spacing between the piles [1–3]. There exists a certain value of pile
spacing at which the mechanism of group failure changes [4–7]. For
those spacing larger than this value, the failure of group pile is analo-
gous with local penetration of individual piles; for those spacing smaller
than this value, the failure of group pile occurs together as a block.
Therefore, it is suggested that the pile spacing should be as large as
practicable to avoid significant pile interaction. However, close pile
spacing is practically unavoidable and accordingly the group pile effect
has to be considered during pile foundation design. Although the im-
portance of clarifying the group pile behavior has been well recognized,
there is a general lack of scientific understanding about the effect of pile
spacing on the behavior of group piles. Overall, the empiricism remain
plays an important role in the current practice of group pile design [8].
Conventional pile studies focus on load-settlement measurement.
Such measurement, however, does not provide enough information for
more refined investigation of the stressed subsoil behavior. With the
some recent studies have made attempts to investigate the subsoil be-
havior near the pile end [9–14]. In particular, some novel approaches
such as the X-ray CT and the particle image velocimetry (PIV) enabled
direct observations of stress and strain distributions of subsoil subjected
to group pile loading [15,16], which provided close inspections into the
subsoil behavior in a nondestructive way. However, it should be noted
that the X-ray CT or PIV method is only possible in a laboratory setting
condition.
The objective of this study is to evaluate the feasibility of using the
signatures of Acoustic Emission (AE) signals to investigate the group
pile behavior. The AE monitoring method has been introduced into a
wide variety of applications with different materials including metal,
rock, concrete and composites [17–20]. Although the AE method does
not receive widespread popularity for monitoring of porous granular
soils, it has brought many insightful results in a variety of successful
applications such as triaxial compression of soils [21], soil seepage and
erosion measurement [22,23], and slope stability measurement
[24].The basic concept behind AE monitoring during pile loading is that
when soils below the pile tip are stressed, AE signals are generated due
to sand particle movements, surface grinding or individual particle

⁎
Corresponding author at: Department of Geotechnical Engineering, College of Civil Engineering, Tongji University, Shanghai 200092, China.
E-mail address: maowuwei@tongji.edu.cn (W. Mao).

https://doi.org/10.1016/j.apacoust.2018.05.001
Received 10 January 2018; Received in revised form 27 April 2018; Accepted 1 May 2018
Available online 08 May 2018
0003-682X/ © 2018 Elsevier Ltd. All rights reserved.
W. Mao et al.
Fig. 1. Schematic diagram of the group pile apparatus: (a) front view; (b) top view of pile locations and air bags; (c) overall picture of the experimental setup.
Fig. 2. Definition of pile positions among the group and the selected piles for
AE monitoring.
breakage. In case of group pile situation, previous studies generally
found that narrower pile spacing had more significant effect on group
pile behavior [5–7,16]. When pile spacing changes, the interlinked
local soil properties will also change during the process of group pile
loading. These changes would be eventually featured by different AE
characteristics. To investigate the acoustic phenomena for different pile
spacing conditions, two types of pile spacing, 2.5D (2.5 times of pile
diameter) and 5D, are tested in the current study to represent strong
and weak group effects. The AE signals detected from these two cases
are analyzed in terms of different parameters and the feasibility of the
applicability is accordingly evaluated.
190
Applied Acoustics 139 (2018) 189–202
Fig. 3. Illustration of group loading and individual loading processes and
loading steps.
2. Experimental details
2.1. Setup of apparatus
The schematic diagram and the general layout of the testing appa-
ratus are shown in Fig. 1. The soil tank has an inner dimension of
1600 mm (width) × 1600 mm (length) × 1680 mm (height). The group
pile model used in this study is constituted of 9 piles (3 × 3), and is
located in the center of the model box. For axially loaded piles, previous
studies [10], [13] suggested that the intense stress and strain
W. Mao et al.
Fig. 4. Load-settlement behavior of the three piles monitored with AE (5D case):(a) 1st group loading;(b) 2nd group loading;(c) 4th group loading;(d) 5th group
loading.
concentrations emanate from the pile tip, and the range of significant
stress and strain variations is mainly restricted within 5D. In the current
experimental setup, the horizontal distance to the sidewall is 15D for
the side pile in case of 5D group setup. It is therefore expected that the
boundary effects are negligible if not totally eliminated.
The model ground is made of silica sand with mean particle size of
0.52 mm, and is compacted to a relative density around 90% during
each test. Eight air bags, which is capable of applying up to 200 kPa
surcharge to the model ground, are placed on the surface of the model
ground surrounding the model piles, as is shown in Fig. 1(b). The
purpose of the surcharge applied herein was used to simulate a deeper
pile embedment. Due to the limited space, there is no confining stress
applied in between the piles. The electrical motor with a maximum
loading capacity of 500kN is placed on two perpendicular trails, and is
191
Applied Acoustics 139 (2018) 189–202
movable in a 2D plane so that piles at different locations can be
reached. The pile used in the test has an outer diameter of 40 mm, with
the pile tip flat and closed-ended. Calibrated strain gauges are attached
inside the pile (50 mm above the tip) to measure the axial shaft force.
The height of the model ground is 1200 mm. The piles are put in po-
sition with soil compacted around and the initial embedment depth is
550 mm. A rigid steel pile cap associated with hoop accessories is ap-
plied to connect the piles in group. An overview of the experimental
setup is shown in Fig. 1(c).
Regarding a 3 × 3 pile group, the positions of piles can be divided
into three categories, which are the center pile, the middle pile and the
corner pile, as shown in Fig. 2 Therefore, three piles representing three
different positions corresponding to the center, middle and corner pile
respectively are selected for AE monitoring.
W. Mao et al.
Fig. 5. Load-settlement behavior of the three piles monitored with AE (2.5D case):(a) 1st group loading;(b) 2nd group loading;(c) 4th group loading;(d) 5th group
loading.
The piezo-ceramic type AE sensor, manufactured by Fuji Ceramics
Corporation, Model R-cast M304A, is used for detection of AE signals.
The AE sensor is mounted on the surface of pile shaft and fixed with
tape. The captured signals are further amplified by the pre-amplifier,
manufactured by Fuji Ceramics Corporation, Model A1201, and digi-
talized with a time interval of 0.5 µs. The AE energy, calculated by
summarizing the area under the rectified signal envelop, is used to
characterize the AE activity.
2.2. Loading procedure
It is known that when pile spacing is close, the bearing load dis-
tribution within the pile group is not uniform among different pile
positions. For example, previous experimental data mentioned that the
192
Applied Acoustics 139 (2018) 189–202
center pile could carry more load than the corner pile in sandy ground
[25]. However, a recent study showed that such location effect existed
only when the piles were loaded together as a group [16]. Once the
group loading was disturbed by the intermission of individual loadings,
the bearing load of the center pile showed no significant difference
compared with the other piles. This observation implies that the in-
teraction within group pile is influenced not only by the pile spacing,
but also by the loading history. Therefore, the current study investigates
the above mentioned behavior through specific loading procedures.
For each test, the pile loading consists of 5 steps. Each step of
loading includes 20 mm pile penetration, and is followed by unloading.
The loading is performed in a displacement control manner and the rate
of displacement is 1 mm/min. It should be noted that the connection
between the loading unit and the pile group cap (or individual piles) is
W. Mao et al.
Fig. 6. Load-settlement behavior during individual pile loading: (a) 5D case; (b) 2.5D case.
not fixed (free touch style). During unloading stages, negative dis-
placement is developed due to the rebounding of ground, and the ul-
timate rebounding value can be reached after the loading unit dis-
connected with the pile group cap (or individual piles).
Two types of loading processes are performed, which is termed as
the group loading and the individual loading respectively. In case of
group pile loading, the piles were connected to the pile cap with hoop
accessories. An additional loading head with large contact area was
placed on the pile cap in order to realize a more uniform load dis-
tribution. While for individual loading, the hoop is removed, and there
is no connection between the pile and the pile cap .The axial force is
applied on individual piles. The purpose of individual loading is to
examine the effect of individual loading on the group pile behavior. The
process of the group loading and individual loading, as well as the 5 pile
loading steps are schematically illustrated in Fig. 3.
3. Test results
3.1. Load-settlement behavior
Fig. 4 shows the load-settlement relations of the three piles mon-
itored with AE during 5D group pile test. The load here refers to the pile
tip resistance measured by the strain gauges, which were located at
50 mm above from the pile tip. It can be seen that the load-settlement
curves during the 1st loading showed notable difference compared with
the reloading steps, especially at the very beginning of penetration. A
clear transition point indicating the ground yielding can be identified
from the reloading curves. It can be generally found that, except for the
1st loading, the load-settlement curves of the three piles were close to
each other before ground yielding. In addition, the rebounding dis-
placement is found to be around 1–2 mm during unloading stages.
Fig. 5 shows the results of load-settlement relation in case of 2.5D
group loading. The overall tendency of load evolution is similar with
that of 5D case. However, it is clearly noted that the center pile carried
more load compared with middle and corner piles. The non-uniform
distributions of load and AE among different pile locations will be
discussed in detail later via the secant modulus and the correlated AE
activity.
Regarding individual loading stage, the load-settlement relation-
ships are summarized in Fig. 6. On the whole, it is seen that the bearing
load in case of 5D increased more rapidly at the beginning of each
193
Applied Acoustics 139 (2018) 189–202
loading compared with that of 2.5D case, demonstrating that the
ground is stiffer. The bearing loads at the end of the 3rd load stage are
in the range of 10.5kN ∼ 12.5kN for 5D case, and 9.5kN ∼ 11.5kN for
2.5D case. For both 5D and 2.5D conditions, the center pile did not
show notable difference compared with other pile positions.
3.2. AE activity during group pile loading
Fig. 7 shows the AE evolution during 5D group pile loading. Similar
tendencies can be observed compared with the load-settlement curves
as shown in Fig. 4. Higher tip resistance of piles yielded higher level of
AE. Note that both soils below pile tip and pile-soil interface sur-
rounding the pile shaft could generated AE signals. An estimation of the
AE generated by pile-soil interface friction is performed by pulling out
the pile from the ground. It is found that the contribution is relatively
small, approximately one-tenth of the push down conditions. For sim-
plification, the AE discussed in this study is assumed to be generated
only at the pile tip. However, it should be noted that the AE generated
at the pile-soil interface could still be used for monitoring of soil be-
havior in specific application scenarios. For example, Berg et al. [26]
used AE generated from pile-soil interface for monitoring of slope de-
formation process and established satisfying correlations between AE
rate and slope deforming velocity.
In Fig. 7, the center and the corner pile showed no obvious differ-
ence during reloading processes. However, the middle pile appeared
much lower AE level during all loading processes. This might be caused
by the initial non-uniform ground conditions since the soil below the
middle pile was subjected to a smaller load.
For many materials such as rock, metal or composites, the AE re-
sponse is different upon reloading compared to the original loading due
to the Kaiser Effect. It is expected that no AE can be observed if the load
does not exceed the previous maximum load. This feature of Kaiser
Effect allows it to be used for assessment of soil behaviors, such as pre-
stress determination or stability monitoring [27,28]. In the current
study, the AE activities during the beginning stages of the 2nd group
loading and the 5th group loading were significantly lower than those
of the 1st and the 4th group loading (e.g. Fig. 7), demonstrating the
influence of Kaiser Effect. However, despite the AE activity became less
significant, it did occur even when the stress was smaller than the
previous maximum stress. This is because that the sources of AE are
generated by numerous individual sand particles. Even a very small
W. Mao et al.
Fig. 7. AE activity evolution during group pile penetration (5D case):(a) 1st group loading;(b) 2nd group loading;(c) 4th group loading;(d) 5th group loading.
load may lead to the rearrangement of particle positions, resulting in
the releasing of AE signals. Therefore, this issue deserve more attention
when the Kaiser Effect is used for investigation of soil behavior.
Particle breakage is an important observation during model pile
tests [29]. According to the previous study [30], the high frequency
component (> 100 kHz) of AE signals are correlated to particle
breakage mechanism, while the low frequency components may be
originated from particle sliding. For stressed soils, it is expected that
higher stress level will lead to larger amount of particle breakage [31].
Accordingly, the intensity of high frequency AE signals can be used as
the representative parameter to evaluate the subsoil stress condition.
Based on the above ideas, the detected AE signals were further treated
with a high-pass filter to eliminate the low frequency components
(< 100 kHz). Fig. 8 shows the results of high-pass AE activity during 5D
194
Applied Acoustics 139 (2018) 189–202
group pile loading. The overall tendencies were similar with those of
the total energy cases, but the magnitude is decreased due to the re-
moval of low frequency components. The ratio of the high-pass AE
energy to the total AE energy is shown in Fig. 9. It is noteworthy that all
curves initiated from a considerably small value and rose rapidly to a
relatively stable value, suggesting that the particle breakage was not
significant at the beginning of pile loading, and it increased rapidly
until the ground is yielded.
Regarding different pile locations among the pile group, the center
pile and the corner pile shared similar ratio levels during all group
loading steps. The middle pile, during the 1st and 2nd group loading,
showed notable difference from the center and corner piles. However,
such difference became insignificant during the 4th and 5th group
loading. It is therefore indicated that when the effect of the initial
W. Mao et al.
Fig. 8. AE activity after high-pass filter (100 kHz) (5D case):(a) 1st group loading;(b) 2nd group loading;(c) 4th group loading;(d) 5th group loading.
ground inhomogeneity was eliminated after certain depth of loading,
the piles among a 5D group demonstrated similar behavior. In other
words, the pile-pile interaction among the 5D group pile is not sig-
nificant.
Fig. 10 shows the AE evolution during 2.5D group pile loading.
Again, similar tendencies can be observed compared with the load-
settlement curves as shown in Fig. 5. A notable feature from Fig. 10 is
that the AE originated from the center pile was much higher than those
from middle and corner piles. On the other hand, the AE detected from
middle and corner piles had similar level of AE activity. The distinct
difference between the center pile and the other piles suggested that the
pile interaction in case of 2.5D group pile was more significant than
that of 5D group pile condition. In addition, the results of the high-pass
AE activity also showed a similar tendency with the total AE activity, as
195
Applied Acoustics 139 (2018) 189–202
shown in Fig. 11. In view of the high-pass AE ratio, as shown in Fig. 12,
the center pile exhibited much higher value compared with middle and
corner piles, indicating that there was more sand particle breakage
occurring below the center pile. This was also consistent with the ob-
servation that the center pile carried higher level of bearing load.
3.3. AE activity during individual pile loading
The AE activities during individual pile loading stage, including the
total AE energy, the high-pass AE energy as well as the ratio of high-
pass AE energy, are summarized in Fig. 13, Fig. 14 and Fig. 15 re-
spectively. Results of pile B3 and C3 are not presented due to un-
expected noise during the test. The overall tendencies are similar with
those of the group loading conditions. However, the magnitudes of the
W. Mao et al.
Fig. 9. Ratio of high-pass AE energy to the total AE energy (5D case):(a) 1st group loading;(b) 2nd group loading;(c) 4th group loading;(d) 5th group loading.
AE energy are significantly larger. This may be attributed to the reason
that during group loading, all piles are connect to the footing plate so
that part of the AE energy is dissipated when the AE signals propagate
from the pile shaft to the footing plate.
Considering different pile positions, the center pile in case of 2.5D
group pile condition demonstrated higher value of AE activity either in
view of AE energy or high-pass energy ratio, especially at the very
beginning of each penetration. It seems that the ground below the
center pile in case of 2.5D group pile condition is densified due to group
loading. While for 5D group pile condition, the center pile did not show
any significant difference from the other piles.
3.4. Observation of ground deformation
Additional tests were performed to investigate the ground de-
formation characteristics under different pile spacing conditions.
196
Applied Acoustics 139 (2018) 189–202
Several layers of color sand were placed below the pile tip to mark the
ground deformation, which could be observed through post-test ground
excavations. As can be seen from Fig. 16, in case of 5D group pile,
individual concave shaped ground deformations were observed below
each pile tip. However, such deformation is mainly restricted below
each pile tip and does not extend any further. The color sand layers
between the piles does not deform notably as well. This mode of ground
deformation suggested no significant effect of group pile interactions in
5D group pile condition. In contrast, in case of 2.5D group pile, a large
concave shape of ground deformation can be observed, suggesting that
the ground below the pile tip deforms as a block. This tendency is also
observed for color sand layers between the piles, demonstrating sig-
nificant group pile interactions. Therefore, it is suggested that for
narrow pile spacing, the group pile is tended to fail as a block.
W. Mao et al.
Fig. 10. AE activity evolution during group pile penetration (2.5D case):(a) 1st group loading;(b) 2nd group loading;(c) 4th group loading;(d) 5th group loading.
4. Discussions
For functioning piles in the field, the allowable ground deformation
is usually small and the ground should be in the state before yielding.
Traditionally, the linear part of the load-settlement curve is considered
as elastic deformation and the secant modulus, defined by the ratio of
pile tip resistance to the pile settlement within this range, is used as an
important parameter to evaluate the pile bearing behavior [16]. It is
worth noting that due to strong pile interactions in case of narrow pile
spacing, the magnitude of settlement experienced by each pile is a
combination of the settlement due to the pile’s “own load” and the
settlement experienced from influence of the adjacent piles. Thus, the
secant modulus essentially represents the apparent stiffness of the
ground.
197
Applied Acoustics 139 (2018) 189–202
Fig. 17 exhibits the features of the secant modulus of the AE mon-
itored piles during 5D and 2.5D group pile tests. The secant modulus
was calculated between 0.5 mm and 1.0 mm settlement, where the
load-settlement curve was almost linear. It can be seen that, for both 5D
and 2.5D cases, the value of secant modulus increased from the 1st
loading to the 2nd loading, while decreased during the 4th reloading
due to individual loading, and increased again during the 5th reloading.
It is therefore demonstrated that the individual loading could sub-
stantially affect the apparent ground stiffness. In general, the values of
secant modulus in case of 2.5D were smaller than those of 5D condi-
tions during the same loading step. From a practical point of view, it is
suggested that stronger group pile effect could result in greater ground
settlement.
It is also noted that the secant modulus in case of 2.5D reduced
W. Mao et al.
Fig. 11. AE activity after high-pass filter (100 kHz) (2.5D case):(a) 1st group loading;(b) 2nd group loading;(c) 4th group loading;(d) 5th group loading.
significantly during individual loading stage, while in case of 5D, it
continued increasing compared with the 2nd group loading stage. On
the other hand, during the 4th group loading, the secant modulus in
case of 5D reduced dramatically, while slight increasing is observed for
2.5D case. The variations of secant modulus caused by individual
loading is difficult to explain because the piles are disconnected and
reconnected to the footing plate in order to perform individual loading,
which may cause unexpected disturbances to the piles.
In view of different pile positions, for 5D case, the center pile
showed no notable difference in the bearing capacity after group
loading except during the 2nd-group loading, which demonstrated that
the group effect in case of 5D pile spacing was not significant. While for
2.5D case, the modulus of the center pile was notably greater than those
of other piles after group loading (2nd and 5th loading). However, for
the 4th group loading which was performed after individual loading,
198
Applied Acoustics 139 (2018) 189–202
the center pile showed lower secant modulus than the corner pile. In
other words, the pile location effect is significant only when the group
loading is following a previous group loading. This observation suggests
that the interaction among group piles was influenced not only by the
current loading condition but also by the history of loading.
To evaluate the correlations between the pile bearing behavior and
the AE activities, the accumulated AE energy was also calculated during
the same pile penetration period (0.5–1 mm), as shown in Fig. 18.
Compared with the secant modulus shown in Fig. 17, the overall dis-
tribution of AE energy demonstrated remarkable similarity. It rose from
the 1st loading to the 2nd loading, while decreased during the 4th re-
loading due to individual loading, and increased again during the 5th
reloading. Also, the AE from the center pile showed no obvious dif-
ference compared with the other piles for 5D case. As for 2.5D case, the
AE from the center pile was substantially higher than those of the other
W. Mao et al.
Fig. 12. Ratio of high-pass AE energy to the total AE energy (2.5D case):(a) 1st group loading;(b) 2nd group loading;(c) 4th group loading;(d) 5th group loading.
piles during all loading steps. In particular, although the AE from the
center pile after individual loading remained higher than the other
piles, the gap became much smaller. Also, the AE from middle and
corner piles showed notable difference after individual loading (4th
loading), while they were almost identical during 2nd and 5th group
loading. The AE characteristics revealed from 2.5D case suggested that
strong group pile effect existed when pile spacing was narrow.
The above experimental results and discussions manifested high
similarity in the evolution tendencies between load measurement and
AE monitoring concerning group pile behaviors. For practical applica-
tions, it is always more difficult to measure the load, especially for long
time monitoring. The AE monitoring provides an alternative non-de-
structive measurement to investigate the ground behavior. However, it
199
Applied Acoustics 139 (2018) 189–202
is worth noting that quantitative correlations between load and AE
remains difficult to be established due to the nature of the AE method.
Because such relationships can be influenced by many factors, which
reduces its practical significance. For example, in the current study, the
AE energy detected during individual loading is generally larger than
that in group loading. Because part of the AE energy is dissipated due to
the footing plate that connects all piles together. This may result in
different AE-load relationships. In addition, it is difficult to quantify the
contribution of the Kaiser Effect regarding granular soils in the current
study. If the ground is subjected to historical stresses, or different stress
paths, the AE-load relationship may be changed as well. For field ap-
plications, the above issue is further affected by many other factors,
such as the location of the sensor, local ground condition, as so on.
W. Mao et al. Applied Acoustics 139 (2018) 189–202

Fig. 13. AE activity evolution during individual pile penetration: (a) 5D case; (b) 2.5D case.

Fig. 14. AE activity after high-pass filter (100 kHz) during individual pile penetration: (a) 5D case; (b) 2.5D case.

5. Conclusions
This study presents the application of AE monitoring method for
evaluation of pile spacing effect on group pile behavior. Three piles
representing the center, middle and corner pile locations in a 3 × 3 pile
group were monitored with AE. Based on AE monitoring results, the
following conclusions can be drawn:
(1) The evolution tendencies of AE activity showed high similarity
compared with the load-settlement curves of piles at different lo-
cations. It is therefore suggested that the AE monitoring could be
potentially used as an alternative tool for evaluation of soil stress
state.
(2) Two types of pile spacing, 2.5D and 5D, were tested to represent
strong and weak group effects. It is generally found that the group
pile in case of narrow spacing may significantly reduce the apparent
stiffness of the ground, evidenced by lower values of secant mod-
ulus and AE energy rate in case of 2.5D group pile. From a practical
point of view, it is suggested that stronger group pile effect could
result in greater ground settlement.
(3) High frequency AE component was interpreted to be correlated
with particle breakage mechanism. The evolving characteristics of
high-pass AE energy suggested that the particle breakage was not
significant at the beginning of pile loading, and it increased rapidly
until the ground was yielded. For 5D case, ratio of the high-pass AE
energy turned to be similar among different piles after the 5th
loading step. While for 2.5D case, the center pile always showed
larger values than the other piles, demonstrating the group pile

200
W. Mao et al. Applied Acoustics 139 (2018) 189–202

Fig. 15. Ratio of high-pass AE energy to the total AE energy during individual pile penetration: (a) 5D case; (b) 2.5D case.

Fig. 16. Observation of ground deformation after group pile loading: (a) 5D; (b) 2.5D.

Fig. 17. Secant modulus of the piles monitored with AE.

Fig. 18. AE energy increasing rate during 0.5–1 mm group pile penetration.

201
W. Mao et al.
effect in case of narrow spacing.
(4) The group pile effect was affected by the pile spacing as well as the
loading history. Center pile carried more load and emitted higher
level of AE when the pile spacing was narrow. However, such group
pile effect was significant only when the group loading was fol-
lowing a previous group loading, and became less significant if the
group loading was interrupted by an individual loading.
Acknowledgement
This study is sponsored by the Shanghai Sailing Program (Grant No.
18YF1424000), the Fundamental Research Funds for the Central
Universities of China (Grant No. 22120170118), and Shanghai
Education Commission (Peak Discipline Construction, Grant No.
0200121005/052).
References
[1] Poulos H-G. Modified calculation of pile-group settlement interaction. J Geotech
Eng 1988;114(6):697–706.
[2] Lee C-Y. Settlement of pile groups-practical approach. J Geotech Eng
1993;119(9):1449–61.
[3] Zhang H-H, Small J-C. Analysis of capped pile groups subjected to horizontal and
vertical loads. Comput Geotech 2000;26(1):1–21.
[4] Whitaker T. Experiments with model piles in groups. Geotechnique
1957;7(4):147–67.
[5] Rollins K-M, Lane J-D, Gerber T-M. Measured and computed lateral response of a
pile group in sand. J Geotech Geoenviron 2005;131(1):103–14.
[6] Lee S-H, Chung C-K. An experimental study of the interaction of vertically loaded
pile groups in sand. Can Geotech J 2005;42(5):1485–93.
[7] Leung Y-F, Klar A, Soga K. Theoretical study on pile length optimization of pile
groups and piled rafts. J Geotech Geoenviron 2009;136(2):319–30.
[8] Randolph M-F. Science and empiricism in pile foundation design. Geotechnique
2003;53(10):847–76.
[9] Jardine R-J, Zhu B-T, Foray P, Yang Z-X. Measurement of stresses around closed
ended settlement piles in sand. Geotechnique 2013;63(1):1–17.
[10] Yang Z-X, Jardine R-J, Zhu B-T, Rimoy S. Stresses developed around settlement
piles penetration in sand. J Geotech Geoenviron 2014;140(3):04013027.
[11] Kuwajima K, Hyodo M, Hyde A-F. Pile bearing capacity factors and soil crushability.
J Geotech Geoenviron 2009;135(7):901–13.
[12] Zhang C, Yang Z-X, Nguyen G-D, Jardine R-J, Einav I. Theoretical breakage me-
chanics and experimental assessment of stresses surrounding piles penetrating into
202
Applied Acoustics 139 (2018) 189–202
dense silica sand. Geotech Lett 2014;4:11–6.
[13] White D-J, Bolton M-D. Displacement and strain paths during plan-strain model pile
installation in sand. Geotechnique 2004;54(6):375–97.
[14] Mao W, Aoyama S, Goto S, Towhata I. Acoustic emission characteristics of subsoil
subjected to vertical pile loading in sand. J Appl Geophys 2015;119:119–27.
[15] Ekisar T, Otani J, Hironaka J. Visualization of soil arching on reinforced embank-
ment with rigid pile foundation using X-ray CT. Geotext Geomembr 2012;32:44–54.
[16] Aoyama S, Mao W, Goto S, Towhata I. Application of advanced procedures to model
tests on the subsoil behavior under vertical loading of group pile in sand. Indian
Geotech J 2016;46(1):64–76.
[17] Chai M, Zhang J, Zhang Z, Duan Q, Cheng G. Acoustic emission studies for char-
acterization of fatigue crack growth in 316LN stainless steel and welds. Appl Acoust
2017;126:101–13.
[18] Maillet E, Baker C, Morscher G-N, Pujar V-V, Lemanski J-R. Feasibility and lim-
itations of damage identification in composite materials using acoustic emission.
Compos Pt A-Appl Sci Manuf 2015;75:77–83.
[19] Prem P-R, Murthy A-R. Acoustic emission monitoring of reinforced concrete beams
subjected to four-point-bending. Appl Acoust 2017;117:28–38.
[20] Cai M, Kaiser P-K, Martin C-D. Quantification of rock mass damage in underground
excavations from microseismic event monitoring. Int J Rock Mech Min Sci
2001;38(8):1135–45.
[21] Wang Y-H, Ma C-H, Yan W-M. Characterizing bond breakages in cemented sands
using a mems accelerometer. Geotech Test J 2009;32(2):186–95.
[22] Hung M-H, Lauchle G-C, Wang M-C. Seepage-induced acoustic emission in granular
soils. J Geotech Geoenviron 2009;135(4):566–72.
[23] Lu Z, Wilson G-V. Acoustic measurements of soil pipe flow and internal erosion. Soil
Sci Soc Am J 2012;76(3):853–66.
[24] Dixon N, Spriggs M-P, Smith A, Meldrum P, Haslam E. Quantification of reactivated
landslide behaviour using acoustic emission monitoring. Landslides
2015;12(3):549–60.
[25] Vesic A-S. Experiments with instrumented pile groups in sand. Performance of deep
foundations. ASTM International: West Conshohocken, PA; 1969. p. 177–222.
[26] Berg N, Smith A, Russell S, Dixon N, Proudfoot D, Take W-A. Correlation of acoustic
emissions with patterns of movement in an extremely slow moving landslide at
Peace River, Alberta. Canada. Can Geotech J 2018. http://dx.doi.org/10.1139/cgj-
2016-0668.
[27] Koerner R-M, Lord Jr A-E, Deutsch W-L. Determination of prestress in granular soils
using AE. J Geotech Eng 1984;110(3):346–58.
[28] Smith A, Dixon N. Quantification of landslide velocity from active waveguide–-
generated acoustic emission. Can Geotech J 2014;52(4):413–25.
[29] Yang Z-X, Jardine R-J, Zhu B-T, Foray P, Tsuha C-H-C. Sand grain crushing and
interface shearing during displacement pile installation in sand. Geotechnique
2010;60(6):469–82.
[30] Mao W, Towhata I. Monitoring of single-particle fragmentation process under static
loading using acoustic emission. Appl Acoust 2015;94:39–45.
[31] Yu F. Particle breakage and the drained shear behavior of sands. Int J Geomech
2017;17(8):04017041.