You are on page 1of 11

Measurement 153 (2020) 107365

Contents lists available at ScienceDirect

Measurement
journal homepage: www.elsevier.com/locate/measurement

A new real-time monitoring technique in calculation of the p-y curve of


single thin steel piles considering the influence of driven energy and
using strain gauge sensors
Hossein Moayedi a,b, Ramli Nazir c, Mesut Gör d, Khairul Anuar Kassim c, Loke Kok Foong e,⇑
a
Department for Management of Science and Technology Development, Ton Duc Thang University, Ho Chi Minh City, Viet Nam
b
Faculty of Civil Engineering, Ton Duc Thang University, Ho Chi Minh City, Viet Nam
c
Centre of Tropical Geoengineering (Geotropik), School of Civil Engineering, Faculty of Engineering, Universiti Teknologi Malaysia, 81310 Johor Bahru, Johor, Malaysia
d
Department of Civil Engineering, Division of Geotechnical Engineering, Firat University, Turkey
e
Institute of Research and Development, Duy Tan University, Da Nang 550000, Viet Nam

a r t i c l e i n f o a b s t r a c t

Article history: This paper presents the modelling of a single thin steel pile to simulate the relationship between soil
Received 26 July 2019 pressure reaction (p) and lateral deflection (y). The presented simulation helps to produce a more reliable
Received in revised form 9 November 2019 p-y curve graph using skin-based strain sensors. Therefore, the main objective of this study is to evaluate
Accepted 3 December 2019
the p-y curve obtained from physical modelling in order to calculate the subgrade reaction modulus (Ks)
Available online 10 December 2019
and maximum bending moment in both loose and dense soil conditions using differentiation and integra-
tion methods. In this regard, extensive instrumentation and unique data acquisition system was used
Keywords:
around the single thin steel pile, where the results provided the most reliable graph of the moment
Strain gauge sensors
Pile monitoring
(M) versus depth (z) from the soil surface. In addition, the effect of pile driving energy with respect to
Driven energy the different relative densities of soil is obtained. The results indicated that as the density of sand
Thin piles increases, the subgrade modulus and maximum bending moment of the pile increases. The increases
Lateral loading in passive soil pressure could cause significant decreases in pile lateral deflection. Furthermore, the p-
y behaviour of the single thin pile is hardly affected by the installation of derivation energy.
Ó 2019 Elsevier Ltd. All rights reserved.

1. Introduction The pile-soil interaction is important in civil engineering design


(e.g., designing deep foundations) and most structural health mon-
Due to stress-dependent, nonlinear, anisotropic and heteroge- itoring subjects [7–11]. The top part of the pile is the critical por-
neous nature of the soil, it has very complex mechanical behaviour tion that need to be assessed during the structural life time of
[1–4]. Therefore, in most engineering design projects, instead of the piles. This is because higher displacement and the maximum
modelling the subsoil (e.g., with all its complexity) the subgrade bending moment occur at the top of the pile [12–15]. For this rea-
soil is replaced by a simpler system called a subgrade reaction son, it is imperative to study the maximum bending moment fur-
model [5,6]. In this regard, several researchers proposed models ther in pile design. The maximum bending moment is the most
in order to simulate subgrade modulus of loose and dense sand. important design criterion; even more important than lateral dis-
Several researchers proposed models that assume a linear relation- placement. However, both of these factors are important in several
ship between vertical settlement (y) and the associated soil force cases. Fig. 1 shows the schematic of a laterally-loaded pile response
reaction (p) at any given point. This is given by the coefficient of (i.e., the horizontal load applied at the top of the pile). It shows that
subgrade reaction, Ks (Eq. (1)). the top layer is the critical part especially with regard to the
motion because sometimes the soil does not have full contact with
P
Ks ¼ ð1Þ the pile in this region.
y
According to various studies, several factors need to be consid-
ered in lateral design response such as (i) load transfer approach,
⇑ Corresponding author at: Institute of Research and Development, Duy Tan (ii) soil-pile interaction and (iii) non-linear p-y curves (e.g., Khari
University, Da Nang 550000, Viet Nam (L. Kok Foong). et al. [16], Suryasentana and Lehane [17], Kim et al. [13] and Guo
E-mail addresses: hossein.moayedi@tdtu.edu.vn (H. Moayedi), lokekokfoong@ and Lehane [18–21]. The non-linear p-y curve includes the lateral
duytan.edu.vn (L. Kok Foong).

https://doi.org/10.1016/j.measurement.2019.107365
0263-2241/Ó 2019 Elsevier Ltd. All rights reserved.
2 H. Moayedi et al. / Measurement 153 (2020) 107365

Fig. 1. Schematic of a Laterally Loaded Pile Response and data acquisition system.

forces and the lateral deflection. Higher lateral resistance enables more, Rovithis et al. [28] used back-calculated p-y loops
the width of the pile to be decreased. In a centrifuge test study constructed from sample data on pile bending moments to evalu-
on lateral pile response and p-y curves presented by Barton et al. ate the seismic soil-pile interaction. Fundamental properties of p-y
[22], the stiffness of the p-y curves varies as the square root of loops are used to derive distributed springs and dashpots [28].
depth rather than linearly, and the ultimate resistance is underes- McGann et al. [29] presented the evaluation of the response of
timated near the pile head and overestimated at depth by the API the soil-pile system to lateral spreading in two alternative lateral
curves. The test used is the cyclic lateral load test on model pipe load cases using FE modules. These modules were used to present
piles driven into inundated and saturated sand. Khari et al. [16] a kinematic analysis of a single pile embedded in the laterally
investigated on development of p-y curves of laterally loaded in spreading layered soil profile. Bouzid et al. [30] used the principles
piles in cohesionless soil and found that the soil reaction for vari- of mobilized strength design (MSD) to construct a p-y curve. Two
ous relative density (from 30% to 75%) increases in range from 40 of the coefficients used were Nc (for stress) and Mc (for strain).
to 90% for smooth pile at small displacement and 90% at a large The p-y curve represented the discrete non-linear springs which
displacement. characterize the soil-pile interaction at various depths.
Naggar et al. [23] developed a simple 2D analysis to correctly Several studies, such as Kaloop and Li [32], Moschas and Stiros
model the pile response (e.g., lateral pile deflection to the applied [33], Vazquez et al. [34], and Guzman-Acevedo et al. [35], stated
horizontal load on pile cap) to dynamic loads. Energy dissipation that the structural health monitoring refers to the process of char-
and wave propagation were also calculated for, along with discon- acterization strategy for engineering structures and implementing
tinuity conditions at the soil-pile interface [23]. Simpson and a damage detection. Here the damage is defined as geometric prop-
Brown [6] stated that the criteria used to develop p-y curves erties of a structural system and/or any changes to the material,
should result from in situ testing and/or laboratory data. Mirza including changes to the boundary conditions. These concerns will
[24] evaluated the behaviour of a pile group in uniform and layered adversely influence the system’s performance [36]. Real-time mon-
soil (sand and/or clay). This study was performed based on the itoring of Infrastructure plays a vital role in public safety in regards
model approach called strain wedge that was proposed to analyse to both post extreme event scenarios and most importantly long-
the response of a long (e.g., also considered as flexible pile) under term damage accumulation. There are few studies, such as Choi
lateral stresses [24]. On the other hand, Dewaikar and Patil [25] et al. [37], Zhang [38], Moayedi and Armaghani [39] and Moayedi
used a simple method for the construction of hyperbolic p-y curves et al. [1], that covered the assessment of soil-pile interaction
for a soft clay. The hyperbolic p-y curve was proposed for static parameters, based on the laboratory experimental results, includ-
loading conditions. Large magnitude lateral loads such as wind ing the vertical or lateral load-settlement behaviours of the bored
and ocean waves were considered. Dash et al. [26] stated that in piles. The previous solutions are hard to use for non-trained per-
the Beam on Nonlinear Winkler Foundation (BNWF) model, the soil sonnel. Besides, potential hybrid mathematical solutions requires
is represented by a set of independent springs, situated at discrete large amount of laboratory works to be done before to be used in
locations along the pile. Generally, the lateral soil spring refers to a such a complex engineering problem [57–61]. This is because of
p-y spring. The shape, strength and stiffness parameter affect the assigning various non-critical parameters and conditions through
response of the pile subjected to lateral seismic loading according laboratory modeling as well as complicated equations obtained
to Dash et al. [26]. Kim et al. [27] used the p-y curves for major from the various procedure. Most previous studies focused on a
parameters relating to the flexibility of the pile (pile diameter, pile particular soil type such as clay (e.g., Prasad and Rao [40], Geor-
length, pile bending stiffness and modulus of subgrade reaction) in giadis et al. [41], Das and Suman [42], Wang and Liu [43]) or sand
numerical analysis. Different bending moments and lateral dis- (e.g., Bouzid [44], Lee et al. [45], Khari et al. [46], Qin and Guo [15],
placements resulted in flexible piles and rigid piles [27]. Further- Reddy and Ayothiraman [47] and Park et al. [48]) which have not
H. Moayedi et al. / Measurement 153 (2020) 107365 3

discussed about using different values of the driving energy. The Table 1
solutions that are provided in most of these studies did not provide Samples of soil used for the experimental program.

a simple design formula for use in resolving the geotechnical and Soil density The density of sand (DR) The height of Sample
geomechanically practical problems. It is appealed that the exist- conditions (cm)
ing models do not show uniform and reliable outputs under differ- A 75% (Dense Sand) 65
ent project conditions and this issue cause uncertainty in the B 30% (Loose Sand) 65
prediction of the lateral bearing capacity of the bored pile. In brief, C 30% dense + 75% loose 15; 50
(Combined Sand)
the currently available solutions do not offer a simple formula and/
or required features such as forecasting the stress-strain behaviour
of laterally loaded bored piles. The authors are therefore motivated tion constant. The result obtained from the lateral deflection of
to propose new modelling for the problem. In this regard, physical the pile at the top (i.e. using LVDTs) was then used to compare with
modelling was conducted to produce the p-y curve of a single pile the deflection calculated by second integration of the observed
under the cyclic loading. Hence, the main objectives of this current moment curves.
study were to find a reliable and simple computation model to esti- There are numerous studies on the scale effects on pile founda-
mate the modulus of subgrade reaction and also to assess the max- tion behavior [49–51]. Regarding the scale effect during the exper-
imum bending moment on loose and dense sand using the p-y iment works, we scaled down the real problem of the pile
curve. foundation based on limitations that we had in the geotechnical
laboratory. One important issue that should be taken into account
2. Materials and methods for lateral or axial vertical loading was the ratio between the pile
diameter and tank diameter. The maximum ratio we found was
A schematic view of the laboratory measuring equipment and 12 pile diameter for horizontal loading and a 5 pile size for ver-
steps that were taken in the geotechnical laboratory are shown tical loading. In our example of pile horizontal loading, we
in Fig. 2. The samples of soil used for the experimental program employed a size larger than 12 of the pile diameter to avoid
is tabulated in Table 1. A total of 12 steps were followed to record any side effects on the testing piles. One more restriction was the
the data for the linear variable differential transformer (LVDT) and ratio between the pile diameter and gran size as it should be prop-
static tests. Note that the LVDTs were only used to measure the erly selected, avoiding the scale effect. In this regard also we have
deflection at the ground surface (e.g., at the top of the pile). How- selected the uniform gran size distribution which was smaller than
ever, the strain gages (see Fig. 3) were used to collect the lateral 0.1 of the pile diameter as suggested in several other studies
deformation of the whole embedded pile’s length. The strain values [52,53].
observed at each strain gauge station were immediately converted The arrangement of the strain gauges is shown in Fig. 3. As sta-
to moments by multiplying the strain by the appropriate calibra- ted, these strain gauges were placed along the pile’s length. The

Fig. 2. Summary of sensors and data collection systems, (a) LVDT to record horizontal displacement in pile’s head, (b) strain gauge sensors, (c) 32 channels data logger.
4 H. Moayedi et al. / Measurement 153 (2020) 107365

Fig. 3. Arrangement of strain gauge sensors (a) Schematic view (b) Actual photograph of sensors installation.

main idea of considering several strain gauges along the pile was to 50 cm. The difference in the height of the free fall produced differ-
take a lateral deflection of the piles along the embedded pile and ent values of driving energy, i.e., 2 J and 4 J respectively. The value
use it through the back-calculation process. The arrangement of of the driving energy was calculated using Eq. (2), and Table 2
the strain gauge on single pile was 2, 6, 10, 15, 20, 26 and 35 cm shows the test number set for each sample with each value of driv-
from the ground surface for the G1–G8 strain gauges, respectively ing energy.
(Fig. 3). Note that the used pile through the experimental program
had 5 cm free length which means, for each unit of a strain gauge, Ea ¼ 9:81ðWhÞðHf Þ ð2Þ
the calculated distance was taken 5.0 cm from the bottom of the Wh is the weight of the hammer (kg), and Hf is the height of free
pile cap. fall (m)
Two values of the height from the bottom of the hammer to the The mobile pluviator developed by Khari et al. [54] was used to
top of the pile cap were used in the free-fall method, i.e., 25 cm and fill the box with sand (Fig. 4a). Same technique is used by Khari

Table 2
Sample details and the driving energy used.

Soil density conditions A (dense) B (loose) C (combined)


Test number 1 2 3 4 5 6
The value of driving energy (Joule) 2 4 2 4 2 4

Fig. 4. laboratory equipment used in this study, (a) The mobile pluviator used to fill the test box (b) experimental set up for single pile driving.
H. Moayedi et al. / Measurement 153 (2020) 107365 5

et al. [31]. By using this apparatus, the density of the sand was 90
accurately determined, and a good soil configuration obtained.
80
Three different relative soil densities were used, i.e., 75%, 30%

Horizontal applied load (N)


and combined (75% and 30%). Two of the sets of soil samples had 70
to be prepared before running the test. This is because samples 60
A, B and C were tested twice, using different values of the driving
energy, (i.e., 2 J and 4 J). Therefore, the total number of soil samples 50
prepared was six. Before the test was conducted, the steel U chan- 40
nel was connected to the box at the mid-point of its length using
the G-clip. This U channel was used to support the driving pile 30
stand and two LVDT units during the testing process. Before start- 20
ing the process of driving the pile into the soil sample, the driving
10
pile stand was connected to the U channel. Subsequently, a ham-
mer with a weight of 825 g was used to drive the pile into the soil 0
samples using the free-fall method (Fig. 4b). 1 2 3 4 5 6 7 8
During testing, the driving pole process was stopped when the Loading step
height between the pile cap and the surface of the sand was 5 cm.
Fig. 7. The loads used in the experiments (for each test).
The driving pole process was conducted carefully to ensure that
the pole was in a vertical position (at 90° to the sand surface).
The test will produce inaccurate results for a pole that is not in a
vertical position, and the data could not then be used for further 90
analysis. Therefore, in such a case, the pole and the sand sample 80
must be removed, the sample preparation repeated and the test

Applied horizontal load (N)


70
re-done. After the pole achieved the pile cap level, the driven pile
stand was removed, and two LVDTs were connected at the U chan- 60
nel using magnetic clips. All LVDTs were positioned touching the 50
pile cap, as shown in Fig. 5. This is because the LVDTs were used
40
30
20
10
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Lateral deflection of pile's head (cm)
Test 1 Test 2 Test 3
Test 4 Test 5 Test 6

Fig. 8. Graph of load (N) against lateral deflection (mm) for six tests.

to measure the horizontal deflection at the pile cap. After the LVDT
setup is completed, a second apparatus must be set up for the sta-
tic test. For the static test, a stand with a pulley was placed beside
Fig. 5. Two LVDT sensors on pile setup.
the sample with the completed driving pile setup and LVDT. Then,

Fig. 6. Applied load in the static test setup; (a) setting the load direction, (b) the static load for a particular stage.
6 H. Moayedi et al. / Measurement 153 (2020) 107365

a rope was tied to the middle of the pile cap, placed on the pulley To conduct the test, various values of the applied horizontal
and connected to a load holder (Fig. 6). load must first be applied to the test apparatus in the so-called sta-
After preparing all apparatus, i.e., after placing the strain gauge tic condition. The various lateral loads were applied to the pile at
at the pile, driving the pile vertically, placing the LVDT at the pile the surface of the ground through a pulley arrangement. Fig. 7 pre-
cap and setting up the static test without applying any loads, all sents the horizontal loads applied as dead weights, which were
the strain wires (7 wires) and all the LVDT wires (2 wires) were gradually increased during the static test.
connected to the data logger. After the wires were connected, the
coefficient of strain, coefficient of LVDT and gauge factor were
input to the employed data logger, as follows: 2.1. p-y curve and horizontal modulus of subgrade reaction (kh)

The coefficient of strain gauge 0.93 Two types of outputs were produced in the experimental study,
Coefficient of LVDT 0.005 i.e., two items of data from the LVDT in mm and seven items of data
Gauge factor 2.13 ± 1% from the strain gauge in le. These results were used to obtain the

Lateral deflecon of pile (cm) Lateral deflecon of pile (cm) Lateral deflecon of pile (cm)
0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 0 0

5 5 5

10 10
10

15 15
15
Depth (cm)

Depth (cm)
Depth (cm)

20 20
20
25
25 25
30
30 30
35
35
35
40
40 H = 4.905 N H = 9.81 N
40 H = 4.905 N H = 9.81 N
H = 4.905 N H = 9.81 N H = 14.715 N H = 19.62 N H = 14.715 N H = 19.62 N
H = 14.715 N H = 19.62 N H = 29.43 N H = 39.24 N H = 29.43 N H = 39.24 N
H = 29.43 N H = 39.24 N H = 49.05 N H = 58.86 N H = 49.05 N H = 58.86 N
H = 49.05 N H = 58.86 N H = 78.48 N H = 78.48 N
H = 78.48 N
(a) (b) (c)

Lateral deflecon of pile (cm) Lateral deflecon of pile (cm) Lateral deflecon of pile (cm)
0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2
0 0 0
5 5 5

10 10 10

15 15 15
Depth (cm)

Depth (cm)
Depth (cm)

20 20 20

25 25 25

30 30
30
35 35
35
40 40
40
H = 4.905 N H = 9.81 N H = 4.905 N H = 9.81 N
H = 4.905 N H = 9.81 N
H = 14.715 N H = 19.62 N H = 14.715 N H = 19.62 N
H = 14.715 N H = 19.62 N
H = 29.43 N H = 39.24 N H = 29.43 N H = 39.24 N
H = 29.43 N H = 39.24 N
H = 49.05 N H = 58.86 N H = 49.05 N H = 58.86 N
H = 49.05 N H = 58.86 N
H = 78.48 N H = 78.48 N
H = 78.48 N
(d) (e) (f)
Fig. 9. Lateral deflection of pile versus depth after 9 cycles of loading in: (a) Test 1, (b) Test 2, (c) Test 3, (d) Test 4, (e) Test 5 and (f) Test 6.
H. Moayedi et al. / Measurement 153 (2020) 107365 7

graph of soil reaction and pile deflection (p-y curve). The p-y curve versus depth for an applied horizontal force of 78.5 N for each exper-
was used to determine the spring constant or modulus of subgrade iment. The highest lateral deflection is found for test #6. As stated
reaction. The value of the spring constant was used in the numerical earlier, test #6 is the combined soil (75% and 30% relative density)
analysis for pile deflection. The model was developed using several with 2 J driving energy. On the other hand, test #2 recorded the low-
different specifications. These experiments were carried out to ana- est lateral deflection. Test #2 is the loose sand (30% relative density)
lyze the relationship between soil reaction, p, and pile deflection, y. with 4 J driving energy (Fig. 9b).
The data from the experiments were further used to develop the
graph of the p-y curve for each test. For this purpose, there were
4. Discussion
three types of instrumentation sets used for each experiment: (i)
strain gauges, (ii) LVDTs and (iii) static weights.
4.1. Moment analysis

3. Results For the second analysis, the moment value must first be
obtained. To obtain the moment value, all the strain values must
For the first analysis, the graph of load (N) against lateral deflec- be multiplied by their own gauge coefficient. The strain value is
tion (mm) is provided. To produce this graph, all loads were con- the output of the experiment. It is different for each load, and each
verted into units of Newtons (N) and also the average value of the load produces seven strain values. Eq. (3) was used to calculate the
LVDT was calculated (e.g., using the coefficient of LVDT = 0.005) to first moment,
find the lateral deflection values. Fig. 8 shows the graph of load
(N) against lateral deflection for six tests obtained from the installed Mn ¼ kn:Gn ð3Þ
LVDTs. In addition, Fig. 9(a–f) illustrates the results of the lateral where Mn is the moment value for each gauge (N.m), kn is each
deflection of the thin wall pile used versus depth in test number gauge coefficient (refer to Table 3) and Gn is the value of strain
1–6 (also see Table 2), respectively. From Fig. 8 it can be seen that for each gauge.
the tests number 3 and 4 (i.e., pile embedded in loose sand) pre- After calculating all the moments for each gauge for the various
sented the highest lateral deflection for a particular horizontal force. loads of each sample, the graph of moment M (N.cm) against depth
On the other hand, it can be said that the lowest applied horizontal from the soil surface z (cm) is produced to obtain the best fitting of
stresses are obtained when (i) the pile derived with low driving the s-curve for the moment. Furthermore, the new value of the
energy (Fig. 9a, c, e) and (ii) the ground covered with the loose sand gauge coefficient for a particular depth from the soil surface, z
(Fig. 9c, d). The lateral bearing capacity of the driven pile in dense (cm), was produced from the best fitting of the s-curve moment.
sand (Fig. 9a, b) was found to be up to 2.2 times higher than when The new values of z are 2.0, 7.6, 13.2, 18.8, 24.4, 30.0, 35.6, 41.2,
it is embedded in the loose sand (Fig. 9c–f). Note that, the results 46.8, 52.4 and 58.0 cm. Finally, the value of the moment based
using the p-y curved developed in other studies may be different on the depth from the best-fitting curve was calculated.
from the results obtained in this study. For instance, according to Fig. 11(a, b) shows the effect of relative density on the maxi-
Khari et al. [54], the soil-pile reaction for various relative densities mum bending moment. The graph shows the lateral load against
(from 30% to 75%) was increased by 40–95% for a smooth pile at the maximum bending moment for two different driving energies.
small displacements and by 90% at large displacements. This could These two graphs represent the two different driving energies of 2 J
be due to the roughness of the pile that has been stated in their stud- and 4 J respectively. Fig. 10.a presents the graph of lateral load (N)
ies. Fig. 10 shows a comparison of the lateral deflection of the pile against the maximum bending moment (Ncm) for driving energy
of 2 J. The highest maximum bending moment occurs for the com-
bined relative density (75% and 30%) soil. The highest maximum
Lateral deflecon of pile (cm)
0.0 0.2 0.4 0.6 0.8 1.0 1.2 bending moment is 1224.9 Ncm, and the highest lateral load is
0 78.48 N. Fig. 10.b shows the graph of the lateral load (N) against
the maximum bending moment (Ncm) for driving energy of 4 J.
The loose sand (Dr 30%) recorded the highest maximum bending
5 moment. The highest maximum bending moment is 1124.9 Ncm
at lateral load 78.48 N. From a combination of these three graphs;
10 it is observed that the pile deflection is lower in dense sand and
higher in loose sand.
15 Fig. 12a and b show the graphs for the lateral load against the
deflection at the ground surface. These two graphs also represent
Depth (cm)

different driving energies, as the graph shown in Fig. 12a shows


20 the lateral load against deflection at ground surface for driving
energy of 2 J, whereas Fig. 12b shows the graph for driving energy
25 of 4 J. The highest deflection occurs for a pile in loose sand. As
expected, the highest lateral load results in the largest pile deflec-
30 tion in loose sand. The value for the deflection for the driving pile
energy of 2 J was 0.3786 (Fig. 12a). Less deflection occurs at the pile
in dense sand. The deflection at the ground surface is shown for
35
dense sand, loose sand and combined (Dr 75% and Dr 30%) The
highest deflection recorded in dense sand was 2.2 cm, which was
40 much lower than that recorded in loose sand (Fig. 12b).
Test 1 Test 2 Test 3
4.2. Development of the p-y curve
Test 4 Test 5 Test 6

Fig. 10. A comparison of lateral deflection of pile versus depth for an applied The development of the p-y curve depends on the soil reaction
horizontal force of 78.5 N for 6 test models. and the deflection of the pile. According to Juirnarongrit and Ash-
8 H. Moayedi et al. / Measurement 153 (2020) 107365

Table 3
Gauge coefficient.

G1 G2 G3 G4 G5 G6 G7
k 1.1374 1.1397 1.0851 1.1572 1.1198 0.9490 0.9958

Fig. 11. Graph of lateral load (N) against maximum bending moment (N.cm) for driving energy: (a) 2 J, (b) 4.

2
ford [55], p-y curves for soils are classified into two different types, d M ð zÞ
i.e., curves for cohesive soil and curves for cohesionless soil. For p¼ ð4Þ
dz2
this study, cohesionless soil (sand) was used to develop the p-y
curve. The p-value is the soil reaction (kN/m), and the ‘y’ value is ZZ
the deflection of the pile. The graph of soil reaction against the M ðzÞdz
y¼ ð5Þ
deflection was plotted as shown in Fig. 12. The values of the soil EI
reaction p (kN/m), and the deflection, y (m) were calculated using
Eqs. (4) and (5) and, respectively, where EI = 91,000 Nmm2
H. Moayedi et al. / Measurement 153 (2020) 107365 9

Fig. 12. Graph of lateral load (N) against deflection (mm) at the ground surface for driving energy: (a) 2 J, (b) 4 J.

Fig. 13a illustrates the p-y curve for the pile where the driving were continuously recorded using a data logger. The data logger
energy used is 2 J. There were three tests for the soil reaction gave the results for the strain values. The moment was calculated
against deflection. The relative densities of the soil samples used by multiplying the gauge coefficient with the strain value. The
were 75%, 30% and combined (75% and 30%). For the graph of soil result from the strain test shows that higher deflection occurred
reaction against deflection, the pile in the dense sand shows less in loose sand and lower deflection occurred at the pile in dense
deflection compared with the other two samples. However, the pile sand. According to Ranjan and Rao [56], the percentage of relative
in the loose sand records the highest deflection. Fig. 13b shows the density for dense soil is between 65% and 85%. For the dense, or
soil reaction against the deflection for 4 J driving pile energy. The high relative density, granular soil, the shear strength found to
highest deflection occurs for the pile in loose sand, i.e., 0.27 mm be high. Besides, the obtained laboratory data can be a basis for
at 0.35 kN/m. The sand with a relative density of 75% shows the many mathematical-based solutions [62–67].
least deflection although the soil reaction is higher than test #4
(Dr 30%) and test 6 (Dr 75% and Dr 30%). 5. Conclusion
There were six tests conducted for three different relative den-
sities of sand, i.e., 75% (dense sand), 30% (loose sand) and 30% plus The p-y curve was developed from the experimental work, and
75% (combined). These three samples were tested using two values the results were generated by the data logger. The data logger
of the driving energy, i.e., 2 J and 4 J. Two items of data from the produced two items of LVDT data and seven items of data from
LVDTs were analyzed to determine the displacement of the pile the strain gauge. The experiments were only conducted on one
cap under load. From the results, the graph of load (N) against lat- type of soil. This was due to the fact that a limited number of soil
eral deflection (mm) was plotted. The highest lateral deflection types were similar to the actual condition of the soils at the site.
was 15.25 mm with 0.078 kN load. For the strain gauges, the data Two different driving energies were used (2 J and 4 J); each driv-
10 H. Moayedi et al. / Measurement 153 (2020) 107365

Fig. 13. The p-y curve for different relative densities of soil with (a) 2 J and (b) 4 J driving pile energy.

ing energy was used for three samples with different relative Declaration of Competing Interest
densities. The following conclusions are obtained from the cur-
rent study. The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared
 The subgrade modulus and maximum bending moment of loose to influence the work reported in this paper.
and dense sand were obtained using the p-y curve.
 The highest deflection of 0.27 mm occurs for the pile in loose
Acknowledgments
sand.
 The sand with relative density 75% shows the least deflection
This research was performed in the geotechnical laboratory of
although the contact pressure (p) is higher than for the test with
Universiti Teknologi Malaysia (UTM). We like to thank Dr. Mahdy
Dr 30% and/or test #6 with relative densities of 75% and 30%
Khari for his immense help during the experimental program.
combined.
 The p-y curve can be determined by using differentiation and
integration methods. The p-y behaviours of the pile were hardly References
influenced by the installation of derivation energy.
[1] J. Fan, D. Jiang, W. Liu, F. Wu, J. Chen, J. Daemen, Discontinuous fatigue of salt
 The highest deflection occurred at the pile in the loose sand rock with low-stress intervals, Int. J. Rock Mech. Min. Sci. 115 (2019) 77–86,
although the least deflection was in the dense sand. From the https://doi.org/10.1016/j.ijrmms.2019.01.013.
p-y curve, the average value for the subgrade reaction was [2] Z. Zhang, D. Jiang, W. Liu, J. Chen, E. Li, J. Fan, K. Xie, Study on the mechanism of
roof collapse and leakage of horizontal cavern in thinly bedded salt rocks,
found to be 12,823 kN/m3. Environ. Earth Sci. 78 (2019) 292, https://doi.org/10.1007/s12665-019-8292-2.
H. Moayedi et al. / Measurement 153 (2020) 107365 11

[3] A. Marto, N. Latifi, M. Janbaz, M. Kholghifard, M. Khari, P. Alimohammadi, A.D. Felix, GPS, accelerometer, and smartphone fused smart sensor for SHM on
Banadaki, Foundation size effect on modulus of subgrade reaction on sandy real-scale bridges, Adv. Civ. Eng. 2019 (2019).
soils, Electron. J. Geotech. Eng. (EJGE) 17 (2012) 2523–2530. [36] J.R. Gaxiola-Camacho, J.A. Quintana-Rodríguez, G.E. Vazquez-Becerra, F.J.
[4] P. Alimohammadi, M. Khari, K.A. Kassim, A. Adnan, A.D. Banadaki, N. Latifi, The Carrion-Viramontes, J.R. Vazquez-Ontiveros, F.J. Lopez-Varelas, Structural
influence of plasticity index on the dynamic soil behavior, Electron. J. Geotech. Health Monitoring of Bridges in Mexico–Case Studies.
Eng. 18 (2013) 613–622. [37] H.Y. Choi, S.R. Lee, H.I. Park, D.H. Kim, Evaluation of lateral load capacity of
[5] I. Alkroosh, H. Nikraz, Simulating pile load-settlement behavior from CPT data bored piles in weathered granite soil, J. Geotech. Geoenviron. Eng. 139 (2013)
using intelligent computing, Open Eng. 1 (2011) 295–305. 1477–1489.
[6] M. Simpson, D.A. Brown, Development of PY curves for piedmont, Residual [38] L. Zhang, Nonlinear analysis of laterally loaded rigid piles in cohesionless soil,
Soils (2003). Comput. Geotech. 36 (2009) 718–724.
[7] V. Raman, M. Drissi-Habti, P. Limje, A. Khadour, Finer SHM-coverage of inter- [39] H. Moayedi, D.J. Armaghani, Optimizing an ANN model with ICA for estimating
plies and bondings in smart composite by dual sinusoidal placed distributed bearing capacity of driven pile in cohesionless soil, Eng. Comput. (2017) 1–10.
optical fiber sensors, Sensors 19 (2019). [40] Y. Prasad, S.N. Rao, Lateral capacity of helical piles in clays, J. Geotech. Eng.-
[8] K.H. Teng, P. Kot, M. Muradov, A. Shaw, K. Hashim, M. Gkantou, A. Al- ASCE 122 (1996) 938–941.
Shamma’a, Embedded smart antenna for non-destructive testing and [41] K. Georgiadis, M. Georgiadis, C. Anagnostopoulos, Lateral bearing capacity of
evaluation (NDT&E) of moisture content and deterioration in concrete, rigid piles near clay slopes, Soils Found. 53 (2013) 144–154.
Sensors 19 (2019). [42] S.K. Das, S. Suman, Prediction of lateral load capacity of pile in clay using
[9] B.S. Vien, W.K. Chiu, M. Russ, M. Fitzgerald, A quantitative approach for the multivariate adaptive regression spline and functional network, Arab. J. Sci.
bone-implant osseointegration assessment based on ultrasonic elastic guided Eng. 40 (2015) 1565–1578.
waves, Sensors 19 (2019). [43] T. Wang, W. Liu, Development of cyclic py curves for laterally loaded pile
[10] B. Xie, J. Li, X. Zhao, Research on damage detection of a 3D steel frame model based on T-bar penetration tests in clay, Can. Geotech. J. (2016).
using smartphones, Sensors 19 (2019). [44] D.A. Bouzid, Finite element analysis of a piled footing under horizontal loading,
[11] X. Zhu, M. Cao, W. Ostachowicz, W. Xu, Damage identification in bridges by Geomech. Eng. 3 (2011) 29–43.
processing dynamic responses to moving loads: features and evaluation, [45] J. Lee, M. Prezzi, R. Salgado, Experimental investigation of the combined load
Sensors 19 (2019). response of model piles driven in sand, Geotech. Test. J. 34 (2011) 653–667.
[12] D.S. Liyanapathirana, S.D. Ekanayake, Application of EPS geofoam in [46] M. Khari, A.K. Kassim, A. Adnan, An experimental study on pile spacing effects
attenuating ground vibrations during vibratory pile driving, Geotext. under lateral loading in sand, ScientificWorld J. 2013 (2013) 8.
Geomembr. 44 (2016) 59–69. [47] K.M. Reddy, R. Ayothiraman, Experimental studies on behavior of single pile
[13] G. Kim, D. Kyung, D. Park, J. Lee, CPT-based p-y analysis for mono-piles in sands under combined uplift and lateral loading, J. Geotech. Geoenviron. Eng. 141
under static and cyclic loading conditions, Geomech. Eng. 9 (2015) 313–328. (2015) 10.
[14] L. Tang, X. Ling, Response of a RC pile group in liquefiable soil: a shake-table [48] D. Park, D. Park, J. Lee, Analyzing load response and load sharing behavior of
investigation, Soil Dyn. Earthq. Eng. 67 (2014) 301–315. piled rafts installed with driven piles in sands, Comput. Geotech. 78 (2016)
[15] H.Y. Qin, W.D. Guo, Nonlinear response of laterally loaded rigid piles in sand, 62–71.
Geomech. Eng. 7 (2014) 679–703. [49] B.M. Lehane, C. Gaudin, J.A. Schneider, Scale effects on tension capacity for
[16] M. Khari, A.K. Kassim, A. Adnan, Development of p-y curves of laterally loaded rough piles buried in dense sand, Geotechnique 55 (2005) 709–719.
piles in cohesionless soil, Sci. World J. 2014 (2014) 8. [50] G.G. Meyerhof, Scale effects of ultimate pile capacity, J. Geotech. Eng. 109
[17] S.K. Suryasentana, B.M. Lehane, Numerical derivation of CPT-based p-y curves (1983) 797–806.
for piles in sand, Geotechnique 64 (2014) 186–194. [51] M. Gui, M. Bolton, Geometry and scale effects in CPT and pile design, Geotech.
[18] F. Guo, B.M. Lehane, Lateral response of piles in weak calcareous sandstone, Charact. 2 (1998) 1063–1068.
Can. Geotech. J. 53 (2016) 1424–1434. [52] P. Foray, Scale and boundary effects on calibration chamber pile tests,
[19] A. Bassil, X. Wang, X. Chapeleau, E. Niederleithinger, O. Abraham, D. Leduc, Proceedings of the First International Symposium on Calibration Chamber
Distributed fiber optics sensing and coda wave interferometry techniques for Testing/ISOCCTl, Potsdam, New York, 1991, pp. 147-160
damage monitoring in concrete structures, Sensors 19 (2019). [53] Z. Yang, R.J. Jardine, B. Zhu, P. Foray, C.d.H.C. Tsuha, Sand grain crushing and
[20] K. Itakura, I. Kamakura, F. Hosoi, Three-dimensional monitoring of plant interface shearing during displacement pile installation in sand, Geotechnique,
structural parameters and chlorophyll distribution, Sensors 19 (2019). 60 (2010) 469
[21] Z. Li, J. He, D. Liu, N. Liu, Z. Long, J. Teng, Influence of uniaxial stress on the [54] M. Khari, K.A. Kassim, A. Adnan, Sand samples’ preparation using mobile
shear-wave spectrum propagating in steel members, Sensors 19 (2019). pluviator, Arab. J. Sci. Eng. 39 (2014) 6825–6834.
[22] Y.O. Barton, W.D.L. Finn, R.H.G. Parry, I. Towhata, Lateral pile response and py [55] T. Juirnarongrit, S.A. Ashford, Effect of pile diameter on the modulus of sub-
curves from centrifuge tests, Offshore Technology Conference, Offshore grade reaction, SSRP, 22, 2001.
Technology Conference, 1983. [56] G. Ranjan, A.S.R. Rao, Basic and applied soil mechanics, New Age International,
[23] M. Naggar, Hesham El, K. Bentley, Dynamic analysis for laterally loaded piles 2007.
and dynamic p-y curves, Can. Geotech. J. 37 (2000) 1166–1183. [57] W. Qiao, K. Huang, M. Azimi, S. Han, A novel hybrid prediction model for
[24] K. Mirza, Seismic Structural Health Monitoring of Bridges, University of British hourly gas consumption in supply side based on improved whale
Columbia, 2006. optimization algorithm and relevance vector machine, IEEE Access 7 (2019)
[25] D.M. Dewaikar, P.A. Patil, Analysis of a laterally loaded pile in cohesion-less 88218–88230.
soil under static and cyclic loading, Indian Geotech. J 36 (2006). [58] W. Qiao, W. Tian, Y. Tian, Q. Yang, Y. Wang, J. Zhang, The forecasting of PM2.5
[26] S.R. Dash, S. Bhattacharya, A. Blakeborough, M. Hyodo, PY curve to model using a hybrid model based on wavelet transform and an improved deep
lateral response of pile foundations in liquefied soils, (2008) 04-01. learning algorithm, IEEE Access 7 (2019) 142814–142825.
[27] Y. Kim, S. Jeong, J. Won, Effect of lateral rigidity of offshore piles using proposed [59] W. Qiao, Z. Yang, Forecast the electricity price of U.S. using a wavelet
py curves in marine clay, Mar. Georesour. Geotechnol. 27 (2009) 53–77. transform-based hybrid model, Energy (2019) 116704.
[28] E. Rovithis, E. Kirtas, K. Pitilakis, Experimental py loops for estimating seismic [60] W. Qiao, Z. Yang, Solving large-scale function optimization problem by using a
soil-pile interaction, Bull. Earthq. Eng. 7 (2009) 719–736. new metaheuristic algorithm based on quantum dolphin swarm algorithm,
[29] C.R. McGann, P. Arduino, P. Mackenzie, Applicability of conventional py IEEE Access 7 (2019) 138972–138989.
relations to the analysis of piles in laterally spreading soil, J. Geotech. [61] J. Chen, D. Lu, W. Liu, J. Fan, D. Jiang, L. Yi, Y. Kang, Stability study and
Geoenviron. Eng. 137 (2010) 557–567. optimization design of small-spacing two-well (SSTW) salt caverns for natural
[30] D.J. Bouzid, S. Bhattacharya, S.R. Dash, Winkler Springs (py curves) for pile gas storages, J. Energy Storage 27 (2020) 101131.
design from stress-strain of soils: FE assessment of scaling coefficients using [62] W. Qiao, H. Lu, G. Zhou, M. Azimi, Q. Yang, W. Tian, A hybrid algorithm for
the Mobilized Strength Design concept, Geomech. Eng. 5 (2013) 379–399. carbon dioxide emissions forecasting based on improved lion swarm
[31] M. Khari, K.A. Kassim, P. Alimohammadi, Response of Single and Grouped Pile optimizer, J. Cleaner Prod. 244 (2020) 118612.
Subjected to Lateral Load in Cohesionless Soil, Applied Mechanics and [63] W. Qiao, Z. Yang, An improved dolphin swarm algorithm based on kernel fuzzy
Materials, Trans Tech Publ, 2015, pp. 1397–1401. C-means in the application of solving the optimal problems of large-scale
[32] M.R. Kaloop, H. Li, Sensitivity and analysis GPS signals based bridge damage function, IEEE Access (2020).
using GPS observations and wavelet transform, Measurement 44 (2011) 927– [64] W. Liu, Z. Zhang, J. Chen, J. Fan, D. Jiang, D. Jjk, Y. Li, Physical simulation of
937. construction and control of two butted-well horizontal cavern energy storage
[33] F. Moschas, S. Stiros, Measurement of the dynamic displacements and of the using large molded rock salt specimens, Energy 185 (2019) 682–694.
modal frequencies of a short-span pedestrian bridge using GPS and an [65] W. Qiao, Z. Yang, Z. Kang, Z. Pan, Short-term natural gas consumption
accelerometer, Eng. Struct. 33 (2011) 10–17. prediction based on Volterra adaptive filter and improved whale optimization
[34] G.E. Vazquez, B.J.R. Gaxiola-Camacho, R. Bennett, G.M. Guzman-Acevedo, I.E. algorithm, Eng. Appl. Artif. Intell. 87 (2020) 103323.
Gaxiola-Camacho, Structural evaluation of dynamic and semi-static [66] W. Qiao, Z. Yang, Modified dolphin swarm algorithm based on chaotic maps
displacements of the Juarez Bridge using GPS technology, Measurement 110 for solving high-dimensional function optimization problems, IEEE Access 7
(2017) 146–153. (2019) 110472–110486.
[35] G.M. Guzman-Acevedo, G.E. Vazquez-Becerra, J.R. Millan-Almaraz, H.E. [67] W. Gao, J.L.G. Guirao, B. Basavanagoud, J. Wu, Partial multi-dividing ontology
Rodriguez-Lozoya, A. Reyes-Salazar, J.R. Gaxiola-Camacho, C.A. Martinez- learning algorithm, Inf. Sci. 467 (2018) 35–58.

You might also like