Acoustic emission characteristics of subsoil subjected to vertical pile loading in sand

Journal of Applied Geophysics 119 (2015) 119–127

Contents lists available at ScienceDirect

Journal of Applied Geophysics journal homepage: www.elsevier.com/locate/jappgeo

Acoustic emission characteristics of subsoil subjected to vertical pile loading in sand Wuwei Mao ⁎, Shogo Aoyama, Shigeru Goto, Ikuo Towhata Department of Civil Engineering, The University of Tokyo, Tokyo 113-8656, Japan

a r t i c l e

i n f o

Article history: Received 6 January 2015 Received in revised form 23 April 2015 Accepted 21 May 2015 Available online 27 May 2015 Keywords: Acoustic emission Pile Sand Subsoil Yielding

a b s t r a c t The response of the subsoil subjected to pile loading is crucial to clarify the bearing mechanism of pile foundations. This study presents a novel acoustic emission (AE) method to monitor the subsoil behavior in a model pile testing system. The AE testing aims to capture the “micro-noises” released from sand grain dislocation and crushing around the pile shaft during penetration. The correlations between the pile settlement and the AE characteristics including count, amplitude and energy are revealed and discussed, highlighting that the ground density and the shear zone formed during pile penetration mainly affect the AE behavior. The results also suggest that the yielding of ground can be determined based on the development of the AE activity. The technique shows promise as an insitu methodology for monitoring of subsoil behavior during the process of pile loading. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Pile foundation is extensively used in infrastructural constructions in order to provide a better performance in ground bearing capacity or displacement behavior (Poulos and Davis, 1980). The design of pile foundation basically relies on the determination of the bearing capacity of the pile shaft or pile groups. In general, current design guidelines estimate the ground bearing capacity from either in-situ tests or theoretical calculations based on laboratory testing, combined with empirical modifications by considering local ground conditions and foundation types. However, it is still a subject of great uncertainty in foundation design to determine the bearing capacity of piles in sandy ground (Foray et al., 1998; Zhang and Chen, 2012; Burlon et al., 2014). Despite various modification approaches were established in order to correlate the testing parameters (e.g., the cone resistance qc) to the end-bearing resistance (qb) of a pile. Nevertheless, the bearing mechanism of the subsoil exhibits complicated cluster of situations which are not well understood, thus the proposed correlations have inevitably been confronted with unreliability, and some of the criteria are even not consistent with the physical processes involved (Randolph et al., 1994). Therefore, there is an ongoing demand to clarify the mechanism of subsoil reactions and the inner interactions between the pile and the surrounding soils (Randolph, 2003). The stress and strain paths developed around the pile tip are most relevant to the pile-end bearing capacity. In some recent studies, several attempts have been made to clarify the subsoil behavior adjacent to the pile tip, such as evaluation of stress development ⁎ Corresponding author. E-mail address: [email protected] (W. Mao).

http://dx.doi.org/10.1016/j.jappgeo.2015.05.017 0926-9851/© 2015 Elsevier B.V. All rights reserved.

around the pile tip (Jardine et al., 2013; Yang et al., 2014), soil crushability and the corresponding pile bearing mechanism (Kuwajima et al., 2009; Zhang et al., 2014). It is generally found that the stress around a pile is radially developed and the high stress zone is restrained near to the pile shaft. Furthermore, particle crushing is a common observation close to the pile tip. Considering relatively few information on soil–pile interaction provided from load–settlement analysis, some novel approaches were developed to achieve more direct observations. Ekisar et al. (2012) used the X-ray CT to visualize the soil arching on reinforced embankment with rigid pile foundation. The effect of load transfer inside the ground was able to be recognized, however, the tested model was restricted to a relatively small size. White and Bolton (2004) applied the particle image velocimetry (PIV) method to investigate the ground settlement and strain path development in a pile penetration test. The PIV results provided a direct vision of sand particle movement through a transparent window. However, it was limited to be a 2-dimensional observation and difficult for field application as well. With the objective of better understanding the subsoil behavior during pile penetration in sand, a more flexible and effective method is preferred. Acoustic emission (AE), or stated as micro-seismic sometimes, refers to the elastic waves generated by the rapid release of energy in stressed materials (Swindleh, 1973). As a non-destructive testing technique, AE testing has been frequently used on materials such as metal, rock, concrete, wooden beams and composites materials. The applications have been concerned with either damage estimation (Ohtsu and Watanabe, 2001), fracture procedure (Takano et al., 2013), flaw localization (Jomdecha et al., 2007), or failure mechanism (Labuz et al., 2001), which have led to the discovery of many new features of the

120

W. Mao et al. / Journal of Applied Geophysics 119 (2015) 119–127

tested materials. Nonetheless, AE testing in soils was less reported in the literature due to the propagation complexity and high attenuation of elastic waves in porous materials. Pioneer works on laboratory and field aspects of AE in geotechnical engineering had been carried out by Koerner and Lord's group in the 1970s, with special focus on the correlation of soil stability and AE activity (Koerner and Lord, 1974; Lord and Koerner, 1975; Koerner et al., 1981). Since AE is capable of reflecting the inner stress state of a material, more recent studies have connected the AE characteristics to some new aspects of soil mechanism, e.g., stress measurement (Kurlenya et al., 1997), soil seepage (Hung et al., 2009), and soil erosion (Lu and Wilson, 2012). It is found that AE has obvious advantages in monitoring the mechanical state of stressed soils. In case of pile installation, the soils immediately below the pile tip and around the pile shaft are stressed and the energy is released due to particle sliding or fracture in form of elastic waves. Information carried within these AE signals may provide direct insights to the subsoil response during pile penetration. Such information can be beneficial to understand the bearing mechanism of the ground. In this study, an AE monitoring system was developed to investigate the subsoil response under vertical loading of pile in sand. The AE signals were analyzed in terms of various parameters: e.g., count, magnitude and energy. The features of AE activity associated with the development of ground resistance and the subsoil behavior were revealed and discussed. 2. Test system details 2.1. Loading apparatus and test procedure Fig. 1 shows the layout of the pile loading apparatus equipped with the AE monitoring system. The sand sample was filled in a rigid soil tank. The internal dimension of the tank was 600 mm long, 600 mm wide and 500 mm high. The closed-ended pile was made of aluminum with diameter of 40 mm. A vertical loading unit was installed on top

of the frame with displacement control assembly. The applied load was recorded by a load cell connected with the loading unit and settlement of the pile was recorded by a linear variable differential transformer (LVDT) transducer. Four air balloons were fixed below the top cover of the model box to apply the confining pressure to the ground, and were set to be at 10 kPa in this study. Five load–unload cycles were conducted for both dense and loose ground cases. The implementation of sequential loading enables the comparison of subsoil behavior due to loading history. In addition, not only ground behavior of pre-yielding but also post-yielding are concerned in the current study. Therefore, the pile is loaded until fully yielded and the penetration depth was set around 20 mm for each sequential cycle at a speed of 1 mm/min. Unloading tests were conducted at the same rate before the next loading sequence.

2.2. Material Silica sand No. 5 was used in this study. Silica sand is extensively used as construction materials in inland engineering, and has been widely used in laboratory testing as well. The physical and mechanical properties of the sand used are shown in Table 1, and the grain size distribution of the sand is shown in Fig. 2. The air-dried sand was prepared inside the testing tank in layers, and compacted after every 50 mm thickness. The total height of the sample was 400 mm. Two types of ground conditions were tested, with relative density (Dr) of 93% for dense case and 67% for loose case.

2.3. Acoustic emission and parameters Fig. 3 shows the schematic view of the AE monitoring system in the test. In order to enhance the signal strength captured by AE transducers, a metal waveguide is usually employed when investigating AE behaviors in soils, especially in large scale field studies (Dixon et al., 2003). The pile herein was made of aluminum which functioned similar to a wave guide. The piezo-ceramics (PZT) type sensor was directly attached to the pile surface so that the excessive attenuation of elastic wave within the sand could be avoided. The AE sensor (effective working range 10 kHz–5 MHz, resonant frequency at 300 kHz) was connected to a preamplifier to obtain high sensitivity with low noise levels. Due to the huge amount of data flow, most of the previous studies recorded the AE data in trigger mode with a preset threshold to avoid the buffer overflow. The main drawback of the trigger mode recording is that signals with amplitude less than the threshold will not be saved, and the data loss could be significant when the AE activity is high. Therefore, a high performance data logger was used in this study and continuous recording of AE signals was carried out during pile loading process with a sampling rate of 2 M/s. Fig. 4 shows a typical AE signal received by the sensor and the AE parameters defined based on it. The AE amplitude represents the peak voltage of the signal wave. The counting numbers are the events with maximum amplitude that exceeds the preset threshold. While the AE energy (EMARSE) is measured by summarizing the area under rectified signal envelop (ASM Handbook, 1992), it is not associated with the threshold. Table 1 Basic properties of the sand.

Fig. 1. Schematic view of the pile test system.

Property

Value

Specific gravity, Gs Maximum void ratio, emax Minimum void ratio, emin D50, mm Coefficient of uniformity, Cu Coefficient of curvature, Cc

2.651 1.09 0.66 0.557 1.692 1.051

W. Mao et al. / Journal of Applied Geophysics 119 (2015) 119–127

121

Percentage finner by weight

100 80 60 40 20 0 0.01

0.1

1

10

Grain size (mm) Fig. 2. Grain size distribution of the sand.

Fig. 4. Typical AE signal and relevant parameters defined.

3. Results and discussions 3.1. Load–settlement behavior Fig. 5 illustrates the relationship between the load and settlement during five sequential loadings. It shows that the ground resistance of dense case measured by the load cell was almost twice as large as loose case at the same penetration depth. The settlements in the unloading stages illustrated that the rebound displacement for dense ground was more prominent than loose ground. Fig. 6 shows the secant modulus in each loading step in case of loose and dense ground conditions, calculated by the following equation: Esec ¼

F Sd

ð1Þ

where Esec is the secant modulus, S is the sectional area of pile end, F is the pile bearing load increment between 0.5 to 1.0 mm settlements of each loading sequence and d (equals to 0.5 mm) is the corresponding settlement. Generally, dense ground showed larger secant modulus than that of loose ground, and the moduli tended to increase with loading steps due to ground compaction. It suggested that the stiffness of the ground was closely associated with soil densification. The dependence of the AE characteristics on ground conditions will be discussed in the next section. 3.2. AE activities

the activity of AE was low at the beginning of loading, increased rapidly to high values, and finally maintained at relatively constant level. This tendency exhibited more obviously during the reloading tests. Particularly, it is worth noting that the tendency of the AE development was in good agreement with the bearing load development as shown in Fig. 5. This correlation confirmed the validity of AE technique in pile penetration monitoring. Furthermore, the AE counts for dense case (Fig. 7(a)) were almost twice as much as those of loose case (Fig. 7(b)) at the end of each loading. It suggests that the dense ground tended to be more emissive than the loose ground. Generally speaking, the subsoil subjected to sequential loading underwent a process of densification (dilatancy may also exist around the shear zone). Accordingly, reloading stages were expected to generate more AEs. The results from loose case confirmed this assumption well where significant gaps between 1st and the latter loading steps can be observed. For the dense case, however, the AE counting showed no obvious difference between 1st and 2nd loadings and decreased slightly during the 3rd–5th loadings, suggesting that there are also other factors affecting the AE behavior, which will be discussed later. Besides the AE count, the amplitude of the AE signal is another important parameter indicating the strength of the released energy. Fig. 8 shows the evolutions of normalized average AE amplitude A / Amax, which were calculated by averaging all detected AE amplitudes within 10 s. It can be seen that the AE amplitude tendency appeared in the same manner with AE count. The average amplitude rose with increased pile penetration, and turned to be relatively constant after a certain penetration depth. Dense case generally showed higher AE amplitude than loose case, but with the absolute magnitudes decreased

3.2.1. AE count and amplitude The results of AE count rate evolutions during all loading processes for both dense and loose cases are summarized in Fig. 7. It shows that

0

Settlement (mm)

20 Dense

Loose

40 60 80 100

0

2000

4000

6000

8000

Load (N) Fig. 3. Schematic view of the AE monitoring system.

Fig. 5. Load–settlement relationship in case of loose and dense ground conditions.

W. Mao et al. / Journal of Applied Geophysics 119 (2015) 119–127

Secant modulus of resistance (kPa/m)

122

2.0E6

rectified signal envelop (MARSE) (ASM Handbook, 1992). Since the AE signals were recorded in discrete numerical values, the following equation is used for calculation:

1.6E6 EMARSE ¼

1.2E6

8.0E5

Dense case Loose case

4.0E5

0.0

1st

2nd

3rd

4th

5th

Loading step Fig. 6. Secant modulus of ground resistance during sequential loadings.

during the latter loading steps. By contrast, loose case exhibited a significant increasing of amplitude between 1st and 2nd loadings and maintained at the relatively constant level during the latter loadings. Such AE characteristics are closely related to the micro-scope mechanics of subsoil response, which will be discussed next. The distribution of AE events by peak amplitudes are shown in Fig. 9. For both cases, the emission occurred over a wide range of peak amplitudes. The majority of AE events, however, were dominated by relatively low peak amplitudes (less than 1 V), which accounted for around 80%–85% for dense case and over 90% for loose case. In view of sequential loadings, for dense case, it is found that the share of higher peak amplitude (N 4 V) showed a continuous decreasing tendency. While for loose case, it exhibited more fluctuated trends. On the whole, it appeared to be increasing from 1st to 2nd loading and then deceased in the later loading steps. It suggested that after ground densification caused by previous loading steps, the loose ground behaved more like the dense ground. 3.2.2. AE energy AE energy is considered as a more preferable parameter in the analysis because it is sensitive to both signal amplitude as well as signal duration, which are the indices of signal strength. As mentioned above, the released AE energy is calculated from the measured area under the

t2 t2 1X 1X V þ ðt ÞΔt þ jV − ðt ÞjΔt: 2 t1 2 t1

ð2Þ

Fig. 10 shows the AE energy development during pile penetration in case of (a) dense ground and (b) loose ground. In general, the released AE energy was relatively low at the beginning of penetration, followed by a rapid increase period and eventually reached certain stable value. Similar with the AE count and amplitude features, the overall shape and tendency between the load–settlement and AE–settlement curves showed high similarity during each loading step, suggesting that the AE characteristics could be a good indicator of the ground condition. Compared with the reloading processes, the AE in the initial loading (1st) showed significant difference, where the AE increasing period was longer and the increasing rate was much lower. It was particularly obvious in case of loose ground as illustrated in Fig. 10(b). By contrast, AE during reloading stages exhibited relatively consistent tendency. The relationships of load–settlement and AE energy rate (EMARSE) during initial loading and 2nd-loading were shown in Fig. 11 for dense case and Fig. 12 for loose case in the semi-log scale. It is clearly illustrated that the transition part from initial rising to final stabilizing during 1st loading was more gradual than the 2nd-loading. The AE activities in the 3rd– 5th reloading cycles were found to be similar with the 2nd-loading so that the detailed description was not presented in this paper. Fig. 13 shows the AE energy evolutions during the unloading stages. The unloading was initiated about 5 s after the loading stopped at the same speed (1 mm/min). It can be seen that the AE energy decreased rapidly to a negligible level after the unloading started, then it rose slightly until the loading unit disconnected with the pile, and finally decreased to zero after a certain period of stabilization. It is also seen that the maximum AE energy (indicated with arrows in Fig. 13(a)) reached at the end of each unloading increased slightly compared with the previous unloading sequences, and the absolute values in the dense cases were generally higher than the loose case. In addition, the elapsed time for AE stabilization became longer during the latter loading steps in both dense and loose cases. Such properties may be potentially used for assessment of the subsoil condition in some dynamic process, e.g., monitoring of AE during the stabilization process of the subsoil after dynamic pounding of the pile, although more investigations are still needed to deepen the knowledge on this issue.

Fig. 7. Evolution of AE count under pile loading in case of (a) dense case and (b) loose case.

W. Mao et al. / Journal of Applied Geophysics 119 (2015) 119–127

123

Fig. 8. Evolution of normalized average AE amplitude under pile loading in case of (a) dense case and (b) loose case.

The intersection of the bisector of the two tangent lines and the curve was defined as the yielding point. Although the yielding point depicted herein is a bit scale sensitive, the method provides an overall impression of ground status. The semi-log scale provides a better observation of the load–settlement/AE–settlement relationships when the pile settlement is small. By applying the same method to the AE–settlement curve, another “yield point” can be obtained. A summary of the yielding settlement is shown in Fig. 14. The yield settlements determined by AE turned to be a bit higher than those determined from load measurements, still it can be stated that the yielding of the ground is marked by the AE characteristics. The high consistency between AE and bearing load behavior make it possible to substitute one for the other at certain circumstances, e.g., AE monitoring during displacement pile installation.

It is also worth noting that the pile embed depth was relatively small in the current study and therefore the contribution of AE from the side of the pile was limited. An estimation of AE generated from the side of pile was carried out by pulling the pile out from the soil. It was found that the maximum AE energy rate during pile pulling reached only about one-tenth of the penetration situations. It suggests that the recorded AE during the penetration test was mainly originated from the pile tip. 3.2.3. AE and ground yielding The “yielding” is a basic concept in soil mechanics and is usually determined from stress–strain relationships. In plasticity theory, it is of great importance to describe the yielding behavior of soils properly in constitutive models. After yielding, the soils will be dominated by plastic deformation and small amount of load increment could result in large deformation. The yielding of soils is associated with irrecoverable plastic deformation and releases energy in form of thermal energy or stress waves. Tanimoto and Tanaka (1986) stated that such dissipated energy could be monitored by means of AE and correlated the onset of soil yielding with AE characteristics through a series of triaxial tests. According to the above revealed AE features, it is found that the yielding of ground was also highly distinguished by AE characteristics in case of pile penetration. The arrows in Figs. 11 and 12 indicate the yielding points, which were derived from the two tangent lines drawn at the beginning and final linear part of the load–settlement curves.

3.2.4. Relation between AE and subsoil behavior According to the above results, the AE activity was observed to be rising before ground yielding, while turned to be relatively constant after yielding. Such tendency was closely related with the ground stress status during pile penetration. Before the ground yielding, the sand below the pile tip was densified and the bearing capacity was built gradually. It may be stated that the ground stayed at the elastic status before the stabilization of AE rate, which was also evidenced by the linear load– settlement curve within this period. From view of the micro-level, the increased stress among sand particles caused significant sliding and crushing, resulting in increased releasing of elastic waves. However,

100

90

Event number (%)

80 70 60 50 40 30 20

90

1st 2nd 3rd 4th 5th

4

5

6

7

8

9

10

Event number (%)

1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0

80

0.8

70

0.7

1st 2nd 3rd 4th 5th

0.6

60

0.5

50

0.4

40

0.3 0.2

30

0.1

20

0.0

10

4

5

6

7

8

9

4

5

6

7

8

9

0

0 1

2

3

4

5

6

7

8

9

1

2

3

Peak amplitude (V)

Peak amplitude (V)

(a)

(b)

Fig. 9. Distribution of event numbers by peak amplitude for (a) dense case and (b) loose case. The insets of the figure show the enlarged view for amplitudes higher than 4 V.

124

W. Mao et al. / Journal of Applied Geophysics 119 (2015) 119–127

Fig. 10. AE energy rate during loading processes in case of (a) dense ground and (b) loose ground.

after the ground yielding, due to continuous pile penetration, the failure surface was well formed and the bearing capacity of the ground failed to increase notably. As a result, AE sources from the particle sliding and crushing turned to be constant as well due to the mobilized failure surface. Relative constant bearing load resulted in constant particle contacts after yielding, which also demonstrates that AE is consistent with the ground stress status. Fig. 15 shows a typical failure pattern below a pile tip. The failure surface was restrained within a limited region and the shape was analogous to half-sphere. Such failure shape was well observed by other researchers as well (Yasufuku and Hyde, 1995; Kuwajima et al., 2009). A closer observation made by Yang et al. (2010) also showed that sand crushing was most significant along this failure surface. The shear zone developed along the failure surface may result in the observed AE behavior mentioned above, where up boundaries of the AE activities were detected. The block of soil below the pile tip moved steadily and progressively downwards along the failure surface, resulting in constant AE rate after the ground yielding. As is mentioned above, the majority of the AEs detected in the current study were originated from the pile tip. Considering the significance of sand crushing below the pile tip, it might be possible to characterize the extent of crushing by means of AE. The sources of AE from pile penetration tests are mainly originated from

sand dislocation, crushing and pile–sand interface friction. The spectrum characteristics of AE signal may provide useful insights to distinguish different AE source mechanisms. For example, AE signals corresponding to sand crushing may contain very high frequency components (Mao and Towhata, 2015), while friction of sand particles may exhibit varied frequency characteristics relating with the stress level (Dagois-Bohy et al., 2010; Michlmayr and Or, 2014). Further investigations toward a better understanding of different AE source mechanisms should be worth trying. In view of sequential loading, it may be generally stated that the overall density of the subsoil was gradually increased, as was revealed by the increased secant modulus shown in Fig. 6. Meanwhile, dense ground was more emissive than loose ground as revealed in Fig. 10. However, the AE observed in sequential loading steps showed different tendency (Fig. 16). During the pre-yielding periods (Fig. 16(a)), the average AE energy rates (calculated during 1–3 mm pile settlement in Fig. 10) of both dense and loose cases were increased from the 1st loading to the 2nd loading, and then deceased to a relatively stable level during 3rd–5th loading steps. In view of the absolute AE energy value, the dense case exhibited higher level than the loose case during the first two loading steps and turned to be almost identical during the 3rd– 5th loading steps. During the post-yielding periods (Fig. 16(b)), in the

Fig. 11. Load–settlement and AE characteristics in case of dense ground for (a) initial loading and (b) 2nd-loading.

W. Mao et al. / Journal of Applied Geophysics 119 (2015) 119–127

125

Fig. 12. Load–settlement and AE characteristics in case of loose ground for (a) initial loading and (b) 2nd-loading.

Fig. 13. AE energy rate during unloading processes in case of (a) dense ground and (b) loose ground.

0.20

by Load-Settlement by AE-Settlement

0.16 0.12 0.08 0.04 0.00

1st

2nd

3rd

4th

5th

Normalized Settlement (S/D)

Normalized Settlement (S/D)

0.20

by Load-Settlement by AE-Settlement

0.16 0.12 0.08 0.04 0.00

1st

2nd

3rd

Load step

Load step

(a)

(b)

Fig. 14. Yield settlement normalized by pile diameter in case of (a) dense ground and (b) loose ground.

4th

5th

126

W. Mao et al. / Journal of Applied Geophysics 119 (2015) 119–127

loading step, it seemed that the ground started to behave like dense ground, since the increment was not obvious. This demonstrates the effect of initial ground condition on the AE characteristics. 4. Conclusion Sandy ground is highly stressed during pile penetration that causes substantial micro-movements of sand grains (e.g., sand slide and sand crushing), releasing energy in form of elastic waves. Based on AE monitoring results, the following findings can be drawn:

Fig. 15. Typical failure pattern below a pile tip (the dash line indicates the failure surface).

dense case, the average AE energy in the constant period (calculated during the 15–20 mm pile settlement in Fig. 10) exhibited a decreasing trend. While in the loose case, AE increased notably from the 1st loading to 2nd loading, and turned almost constant during the 2nd–5th loading cycles. In addition, the distribution of the peak amplitude also showed similar difference between the dense and loose cases (Fig. 9). A comparison of the two testing cases showed that the loose ground tended to behave like the dense ground after the sequential loading. In view of the different loading steps, the AE evolution was not always showing the increasing trend with the sequential loading. It seems that apart from soil density, there are other aspects affecting the AE behavior. One reason could be the failure surface formed in the previous loading steps. The sand particles within failure zone underwent severest delocalization, accompanying particle sliding and crushing which initiated substantial AE waves. After the failure zone was formed in the 1st-loading step, the AE contribution from this zone may decrease notably during the reloading stages. Because the particles within this zone had better arrangement and directivity during reloading, which caused less sliding compared with randomly distributed condition in case of 1st-loading. In dense case, the AE increment from ground densification was not as significant as in loose case, and consequently, the overall AE was dropping. Furthermore, the dilatancy of soil around the shear zone may create a zone of soil less dense, which could be more evident in case of dense ground and attributed to the decreased AE observed in Fig. 16(b). In loose case, the densification effect might be still dominant, and therefore the overall AE increased significantly in the 2nd-loading step despite of the pre-existed failure surface. However, from the 3rd-

(1) The process of pile penetration was highly distinguished by AE activities. The evolution tendency of the AE count, AE amplitude and AE energy showed high similarity with load–settlement curves. The AE at the beginning of the pile penetrating was relatively low, followed by a rapid rising period and eventually became relatively constant. (2) The test results showed that the yield settlements obtained from both load and AE data were close. Considering that the yielding of the soil is closely related to the release of irrecoverable energy, therefore, the use of AE energy measurement for yielding determination seems reasonable. This suggests a new method of studying the yielding of soil other than the traditional stress– strain-based method. (3) In general, dense ground was more emissive than loose ground. However, under sequential loading conditions, despite the subsoil generally became denser, the AE energy dropped after each loading step in dense case. By contrast, loose ground showed significant increment from the 1st to 2nd-loading, and turned to be relatively constant from the 3rd–5th loading cycles. It suggested that the loose ground turned to behave like dense ground after being densified by the previous loading steps. (4) Apart from the density effect, it is believed that the dislocations of sand particles in the shear zone resulted in better arrangement and directivity, which decreased the AE activity within this zone during the latter sequential loadings. The effect of dilatancy may also attribute to the decreased AE activity in dense case. (5) The AE features revealed herein provide new insights into the energy dissipation of subsoil subjected to pile load. The features of energy dissipation are closely related with the ground bearing capacity development. Potential application of the technique to field pile monitoring is promising. However, various in-situ aspects of the effects on AE characteristics including anisotropy, water content effect and wave attenuation would deserve for further investigation in order to obtain optimal results.

Fig. 16. Average AE energy rate during (a) 1–3 mm and (b) 15–20 mm pile settlement.

W. Mao et al. / Journal of Applied Geophysics 119 (2015) 119–127

Acknowledgment The financial assistance provided by Japanese Ministry of Education, Culture, Sports, Science & Technology (MEXT, Grant No. 123113) for PhD studies of the first author is gratefully acknowledged. The authors are grateful to the advice and expertise provided by Professor Junichi Koseki, Professor Reiko Kuwano and Associate Professor Taro Uchimura of the University of Tokyo. We also thank Dr. Sadao Yabuuchi, President of Japan Pile Corporation for his invaluable support.

References ASM Handbook, 1992. 9th ed. Nondestructive Evaluation and Quality Control vol. 17. ASM International, pp. 593–682. Burlon, S., Frank, R., Baguelin, F., Habert, J., Legrand, S., 2014. Model factor for the bearing capacity of piles from pressure meter test results—Eurocode 7 approach. Geotechnique 64 (7), 513–525. Dagois-Bohy, S., Ngo, S., Pont, S.C.D., Douady, S., 2010. Laboratory singing sand avalanches. Ultrasonics 50 (2), 127–132. Dixon, N., Hill, R., Kavanagh, J., 2003. Acoustic emission monitoring of slope instability: development of an active waveguide system. Geotech. Eng. 156 (2), 83–95. Ekisar, T., Otani, J., Hironaka, J., 2012. Visualization of soil arching on reinforced embankment with rigid pile foundation using X-ray CT. Geotext. Geomembr. 32, 44–54. Foray, P., Balachowski, L., Colliat, J.L., 1998. Bearing capacity of model piles driven into dense overconsolidated sands. Can. Geotech. J. 35 (2), 374–385. Hung, M.H., Lauchle, G.C., Wang, M.C., 2009. Seepage-induced acoustic emission in granular soils. J. Geotech. Geoenviron. 135 (4), 566–572. Jardine, R.J., Zhu, B.T., Foray, P., Yang, Z.X., 2013. Measurement of stresses around closedended settlement piles in sand. Geotechnique 63 (1), 1–17. Jomdecha, C., Prateepasen, A., Kaewtrakulpong, P., 2007. Study on source location using an acoustic emission system for various corrosion types. NDT E Int. 40 (8), 584–593. Koerner, R.M., Lord Jr., A.E., 1974. Acoustic emissions in stressed soil samples. J. Acous. Soc. Am. 56 (6), 1924–1927. Koerner, R.M., McCabe, W.M., Lord, A.E., 1981. Acoustic emission behavior and monitoring of soils. Acoustic Emissions in Geotechnical Engineering Practice, a symposium 750. ASTM STP, pp. 93–141. Kurlenya, M.V., Petrov, V.E., Popov, S.N., Tkach, K.B., 1997. Applicability of acoustic waveguides for stress measurement in soils. J. Min. Sci. 33 (1), 88–93.

127

Kuwajima, K., Hyodo, M., Hyde, A.F., 2009. Pile bearing capacity factors and soil crushability. J. Geotech. Geoenviron. 135 (7), 901–913. Labuz, J.F., Cattaneo, S., Chen, L.H., 2001. Acoustic emission at failure in quasi-brittle materials. Constr. Build. Mater. 15 (5), 225–233. Lord Jr., A.E., Koerner, R.M., 1975. Acoustic emissions in soils and their use in assessing earth dam stability. J. Acous. Soc. Am. 57 (2), 516–519. Lu, Z., Wilson, G.V., 2012. Acoustic measurements of soil pipe flow and internal erosion. Soil Sci. Soc. Am. J. 76 (3), 853–866. Mao, W., Towhata, I., 2015. Monitoring of single-particle fragmentation process under static loading using acoustic emission. Appl. Acoust. 94, 39–45. Michlmayr, G., Or, D., 2014. Mechanisms for acoustic emissions generation during granular shearing. Granul. Matter 16 (5), 627–640. Ohtsu, M., Watanabe, H., 2001. Quantitative damage estimation of concrete by acoustic emission. Constr. Build. Mater. 15, 217–224. Poulos, H.G., Davis, E.H., 1980. Pile Foundation Analysis and Design (No. Monograph). Wiley, New York. Randolph, M.F., 2003. Science and empiricism in pile foundation design. Geotechnique 53 (10), 847–875. Randolph, M.F., Dolwin, R., Beck, R., 1994. Design of driven piles in sand. Geotechnique 44 (3), 427–448. Swindleh, W., 1973. Acoustic emission. 1. Introduction. Nondestruct. Test. 6 (3), 152–158. Takano, T., Maeda, T., Miki, Y., Akatsuka, S., Hattori, K., Nishihashi, M., Hirano, T., 2013. Detection of microwave emission due to rock fracture as a new tool for geophysics: a field test at a volcano in Miyake Island, Japan. J. Appl. Geophys. 94, 1–14. Tanimoto, K., Tanaka, Y., 1986. Yielding of soil as determined by acoustic emission. Soils Found. 26 (3), 69–80. White, D.J., Bolton, M.D., 2004. Displacement and strain paths during plan-strain model pile installation in sand. Geotechnique 54 (6), 375–397. Yang, Z.X., Jardine, R.J., Zhu, B.T., Foray, P., Tsuha, C.H.C., 2010. Sand grain crushing and interface shearing during displacement pile installation in sand. Geotechnique 60 (6), 469–482. Yang, Z.X., Jardine, R.J., Zhu, B.T., Rimoy, S., 2014. Stresses developed around settlement piles penetration in sand. J. Geotech. Geoenviron. 140 (3), 04013027. Yasufuku, N., Hyde, A.F.L., 1995. Pile end-bearing capacity in crushable sands. Geotechnique 45 (4), 663–676. Zhang, L., Chen, J.J., 2012. Effect of spatial correlation of standard penetration test (SPT) data on bearing capacity of driven piles in sand. Can. Geotech. J. 49 (4), 394–402. Zhang, C., Yang, Z.X., Nguyen, G.D., Jardine, R.J., Einav, I., 2014. Theoretical breakage mechanics and experimental assessment of stresses surrounding piles penetrating into dense silica sand. Géotech. Lett. 4, 11–16 (January–March).