Magazine of Civil Engineering. 2024. 17(3). Article No.

12704

Magazine of Civil Engineering ISSN

2712-8172

journal homepage: http://engstroy.spbstu.ru/

Research article
UDC 624
DOI: 10.34910/MCE.127.4

Performance of geosynthetics-strengthened unconnected

piled raft foundations under seismic loading
A.G. Jihad1,2, M. Karkush1✉
1Civil Engineering Department, University of Baghdad, Baghdad, Iraq
2Road and Transport Engineering Department, Al-Qadisiyah University, AL-Qadisiyah, Iraq
 mahdi_karkush@yahoo.com
Keywords: unconnected piled raft, geosynthetics, foundations, seismic loading
Abstract. Over the past 20 years, geosynthetic reinforcement has been widely used in earthen
constructions. For constrained building projects and challenging ground conditions, geosynthetic-
reinforced-pile-supported (GRPS) embankments are regarded as dependable options. There has been
significant research into ways to improve the stability of pile foundations, and there has also been some
research into the viability of using unconnected piled raft under seismic load. This study investigates the
use of geosynthetics to improve the performance of unconnected pile foundations under seismic load. A
series of reduced-scale physical model experiments conducted on a shaking table was used to assess the
seismic soil-structure interaction behaviors of geosynthetics-reinforced pile foundation systems. The study
makes use of disconnected, closed-end aluminum piles with different thicknesses of cushion layers of sand
soil and different spacing between piles are used in the study. The performance of disconnected footing
was examined using a shaking table to determine the impact of the geosynthetics reinforced. These model
findings are used to identify and thoroughly describe the contradictions in the present design
methodologies. This study shows the settlement of the foundation under seismic loading decreased by
reinforcing the cushion layer with one and two layers of geogrid materials, increasing the thickness of the
cushion layer led to increasing the settlement of the foundation under seismic loading, where the settlement
decreases with decreasing the value of spacing between piles. However, with further increases in either
the cushion thickness or cushion stiffness or both, the effect of improvement would decrease.
Citation: Jihad, A.G., Karkush, M. Performance of geosynthetics-strengthened unconnected piled raft
foundations under seismic loading. Magazine of Civil Engineering. 2024. 17(3). Article no. 12704.
DOI: 10.34910/MCE.127.4

1. Introduction
One of the key functions of geotechnical engineering is the design of the foundation system. A safe,
effective, and cost-effective project will undoubtedly result from the proper selection and design of the
foundation system. A foundation system can use piles for two main reasons: to increase bearing capacity
and/or to limit predicted settlements to acceptable levels. The general assumption made by traditional pile
foundation design techniques is that the piles would support the entire weight of the superstructure. Over
the past few decades, a more modern strategy has been used, one that uses piles as settlement reducers
in what is known as a piled raft foundation. Davis and Poulos [1] introduced the concept of piled raft
foundation in 1972. After that, many researchers have been involved in the design and analysis of such
type of foundation. Among those [2–8] and others, the foundation of a piled raft should only contain the
minimum number of piles necessary to make the settlement at a reasonable level. A portion of the loads
that are transmitted from the superstructure to them via the raft is also carried by these piles. In other words,
the loads are shared between the raft and piles. High axial stress concentration at the pile head may occur
[9, 10].

© Jihad, A.G., Karkush, M. 2024. Published by Peter the Great St. Petersburg Polytechnic University.
 Magazine of Civil Engineering, 17(3), 2024

In addition, if the foundation is planned as a structurally connected piled raft in seismically active
zones or places prone to high winds, large shear forces in addition to overturning moments may arise in
the pile’s head due to the lateral dynamic load. For all of that, the ultimate capacity may be governed by
the structural capacity of the pile rather than the geotechnical capacity. In these situations, more likely than
soil failure is the structural failure of a foundation. Therefore, to avoid structural damage to the foundation,
the dimensions of piles should be increased and the result is an uneconomical design of the foundation.
The use of disconnected piles and the use of an intervening compacted fill layer known as a cushion layer
between the raft and the piles have recently been suggested as alternatives to the traditional method for
dealing with constraint responses. The piles in such a foundation system are considered soil reinforcement
members rather than structural components [11–13]. In unconnected piles, the carrying loads could be
much higher than that carried by structurally connected piles, and still, the economic benefits have been
gained. Moreover, there should be no possible damage to structural connections. In addition, because of
the mobilized adhesion force along the soil-raft interface, the horizontal loads can be effectively transmitted
as shown in Figure 1 [11,14,15].
In this study, the propagation and degradation of the sandy soil in the cushion layer that separates
the raft and unconnected piles were investigated in detail by conducting several tests using a shaking table.
The cushion layer used in this study was reinforced by several layers of geosynthetic materials at a different
cushion layer thickness and change the space between piles, all the raft footing was tested under seismic
load.

Figure 1. Transfer mechanism for horizontal loads: a) piles structurally connected to raft
and b) piles structurally unconnected from raft [11].

2. Materials and Methods

2.1. Shaking Table
Using a shaking table device, the uniaxial sinusoidal motion was adopted with various frequencies
as shown in Fig. 2. The shaking table was produced in the local market which has a dimension of 1×1 m
size with 1.25 tons loading capability. It was supported by a heavy-duty steel frame that had 4 rollers that
could move smoothly in one direction. All parts of the shaking plate were fixed to a rigid steel frame by 8
bolts to ensure rigidity and eliminate any motion, vibration, or sliding of the shaking table during the
controlled seismic loading. In addition, the AC driver is contacted by the computer to control the time and
frequency of the shaking. To adjust the exact frequency magnitude a calibration process must be done,
this process was executed as follow: for a specific period, the number of cycles was recorded compared to
the AC frequency, and using the accelerometer in the base of the shaking table the results were recorded,
the result is the frequency in Hz by change the velocity of the motion a new value of frequency would appear
as shown in Fig. 3. Thus, for a specific value of frequency, it is easy to select the correct number in the AC
drive to determine the speed of the motor to get the required acceleration in the base of the shaking table.
 Magazine of Civil Engineering, 17(3), 2024

Figure 2. Testing model scheme.

700

600

500
Motor Speed (rpm)

400

300

200

100

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Acceleration (g)

Figure 3. The calibration of the motor speed (rpm) and the acceleration.
2.2. Steel Container
A steel container had a depth of 850 mm and 800×700 mm cross-section and was made of a 5 mm
plate thickness. The steel container was set up on a mobile frame base. In every test, the container’s side
dimension was greater than three times that of the raft side. The piles’ end tips were about 25D or more
above the container’s base. It was made out of a steel frame of L-shape steel angles of 75x75 mm in size.
Three sides of the steel container were covered with a steel plate 5 mm thick, and the fourth side was
covered with strengthened transparency glass which help to watch the behavior of the foundation during
the test. One of the improvements applied was the flexible rubber joints at the box corners which permit
motion within a specific limit, this movement was limited to 20 mm and could occur at the same time
alternately for each opposite side. The purpose of this flexible joint was to lessen the effect of the fixed joint
on the behavior of the soil sample during shaking, which resulted in unanticipated behavior by wave
reversals, damping, and other factors.
2.3. Pile Model
The pile model used in the current study was made of aluminum alloy of hollow square section of
14 mm in outside dimension and 11 mm inside dimension. The embedded pile length of 350 mm was used
in all experimental model tests. The slenderness ratio (length to diameter ratio (L/d)) was 22.15.
2.4. Raft Model
The square steel plate with side dimensions 210×210 mm and a thickness of 12 mm was used as a
model raft. The modulus of elasticity (Er) and Poisson’s ratio (𝜈𝜈𝑟𝑟) of the steel plate were 210×106 kPa and
0.30, respectively. The total weight of the suggested footing was 18 kg which is equivalent to the ordinary
building weight.
2.5. Geosynthetics
This study looked into how soil degradation was affected by geosynthetic reinforcement. A non-
woven geotextile sheet was laid on top of a biaxial geogrid in the reinforced fill tests. The geosynthetic
characteristics and forms and its location are presented in Table 1 and Fig. 4, respectively. Given that the
fill material was sand, the non-woven geotextile was used over the geogrid to both transmit the load from
the foundation to the geogrid and prevent sand from flowing through the geogrid apertures.
 Magazine of Civil Engineering, 17(3), 2024

Table 1. Physical characteristics and geometry of geosynthetic materials.

Geosynthetic Properties Units Values
Aperture dimensions mm 25×33
Minimum rib thickness mm 0.76
Biaxial geogrid Tensile strength @ 2% strain kN/m 6.6
Tensile strength @ 5% strain kN/m 13.4
Ultimate tensile strength kN/m 19
Mass/unit area gm/m2 96
Thickness mm 1.04
Non-woven Wide-width tensile strength kN/m 7.23
geotextile Wide-width tensile elongation % 54
Puncture strength (CBR) N 1218
Opening size mm 0.15
Note: geosynthetic properties are based on the manufacturer-provided datasheet.

Biaxial geogrid Non-woven geotextile Locations of geosynthetic layers

Figure 4. Geosynthetic materials used in the experimental work.
2.6. Instruments
A 25 mm accurate LVDT (Linear Variable Differential Transformer) was placed on the center of the
foundation system to measure the vertical settlements of the foundation model. Through metal rods fixed
to the horizontal beam at the top of the box. The system was supported by metal rods and fixed to the base
of the shaking table. The horizontal movement for the foundation and the shaking table was measured by
LVDTs of lengths (400 and 500 mm) respectively shown in Fig. 5.
The acceleration of the shaking table, inside the soil, and on the plate and the foundation was
measured by using an analog accelerometer sensor ADXL335, which can measure three directions of
acceleration excitation as shown in Fig. 5. Five sensors were used; one on the shaking table, three within
the soil column at different depths (each 20 cm depth apart), and the last one was fixed on the foundation.
The selected sensor ADXL335X was a simple breakout board that allows easy, direct, and rapid evaluation
of the performance of the ADXL335 accelerometer. The used sensor was a 3-axis analog-output
accelerometer with a ±3 measurement range. The small size (25.4×25.4 mm) of the accelerometer made
it easy to maneuver and fix on the different systems without needing additional hardware. All the
accelerometers were connected to the data acquisition system channel.

a b
Figure 5. a) Accelerometer and b) LVDTs used in the tests.
 Magazine of Civil Engineering, 17(3), 2024

Data acquisition with LabVIEW Application was used to record and save all the readings of LVDTs
and accelerometer. LabVIEW (Laboratory Virtual Instrumentation Engineering Workbench) is a platform
and development environment for a visual programming language from National Instruments. It is
commonly used for data acquisition, instrument control, and industrial automation on a variety of platforms
including Microsoft Windows, various flavors of UNIX, Linux, and Mac OS X. The data logger is supplied
with 32 channels for connecting the transducers for all the instruments where 5 channels are used for
accelerometers and 3 channels are used for LVDTs, as well 4 channels are used for AC drive control and
internal connecter. The problem was represented by disturbance of the instrument’s signals by the
magnetic radiations from the motors of the shaker leading to incorrect readings. To solve this problem, two
actions were adopted as described in the following points. First using ground connections for all the
equipment and connecting them to a copper rod that has been previously installed in the ground soil to
ensure the removal of any charges generated that may cause noise in the reading, the second use of a
transformer for electricity equipped for all equipment to ensure the elimination of any noise caused by the
electric current on the readings.
2.7. Soil Box Model and Bedding Soil
A relatively uniform dry sand of grain size between 0.35 and 1.18 mm was utilized in this investigation
as foundation soil. The maximum and minimum dry unit weight of the tested sand were found 17.44 kN/m3
and 15.17 kN/m3, respectively. Table 2 shows the physical properties of the used soil. According to the
Unified Soil Classification System (USCS), the soil is classified as poorly graded sand (SP).
Table 2. Physical Properties of the tested soil.
Property Value Standard Specification
Specific gravity 2.64 [16]
Sand (0.35–1.18 mm), % 100
Coefficient of curvature 1.05
Coefficient of uniformity 1.52
D10, mm 0.5 [17,18]
D30, mm 0.6327
D60, mm 0.7597
USCS soil type SP
Maximum dry unit weight, kN/m3 17.44 [19]
Minimum dry unit weight, kN/m3 15.17 [20]
Used dry unit weight, kN/m3 16.627
Relative density, % 50

To reconstitute or prepare the sand samples for experimental model tests, several ways were used
such as raining, vibration, or tamping. In general, as it was reported by [21–25], to provide relatively
homogenous sand samples and to get the required relative density, the raining procedure is considered
the best. The adopted raining technique consisted of a steel hopper of 35 cm height with a valve at the
bottom to control sand raining manually. A sieve of 4.75 mm opening size was attached to the bottom via
a funnel connected to a hopper to distribute the sand uniformly. The hopper was hanging by a pulley to
adjust the height of the rain. Many trials were performed to draw a relationship between the falling height
and the corresponding dry density as shown in Fig. 6. From the curve, it is clear that the falling height
should be about 200 mm to get the required dry density of 16.627 kN/m3.
The inside walls of the steel container were divided into 8 intervals 10 cm each to ensure that the
desired dry density was obtained properly. To produce a homogenous sample overall in the container, the
required quantity of sand for each layer was weighted such that the designed dry density could be obtained.
After the completion of the raining of the soil for each layer, the sand layer was equalized by a steel plate
of section area (175×200 mm) to an average of one blow per unit area for grain fixation and insurance of
the required layer height. However, the target relative density of the sand sample was 50% ± 2%.
 Magazine of Civil Engineering, 17(3), 2024

17.5

17

Unit weight (kN/m3)

16.5

16

15.5

15
0 100 200 300 400 500 600 700 800
Distance from sieve to sand soil surface (mm)
Figure 6. The relationship between the height of falling and dry density of sandy soil.

2.8. Seismic loading

Seismic loading, which is known as an application of seismic oscillation to a structure, can be
considered one of the most essential basics of earthquake engineering. It depends on seismic hazard,
geotechnical site parameters, and structural natural frequency which occurs at the surface of the structure
and in contact with the ground or an adjacent structure. In the years 2017 and 2018, seismic activities were
recorded with different strengths in Iraq. During the last two months of the year 2017, more than seventy
earthquakes hit Iraq. As shown in Table 3, the magnitude of these earthquakes ranged from 4.0 to 7.3 on
the Richter scale, with depths ranging from 6.21 to 42.32 km.
Table 3. History of some earthquakes in Iraq.
Magnitude (Richter
Year Location
scale)
1960 6.0 to 6.7 Halabja City which locates at northeast of Baghdad
1967 6.1 100 km south of Halabja City
2013 5.6 and 5.8 Two earthquakes hit about 60 km south of Halabja City, northeast of Baghdad
2017 7.3 Series of earthquakes 30 km south of Halabja City, northeast of Baghdad, Iraq
2018 4.0 to 4.5 A series of earthquakes hit northeast of Baghdad

Halabja City is located northeast of Baghdad (Iraq) and near the international border between Iraq
and Iran. Halabja earthquake was used in this study. The acceleration of this earthquake was 0.24 g, and
seismic loading conditions have been applied to the model as acceleration is shown in Fig. 7.

Figure 7. The acceleration time history of the applied Halabja earthquake.

2.9. Testing procedure
The test procedure followed in this investigation consisted of many steps such as:
1. The preparation of each model test was commenced by weighing the required quantity of sand to
get the target density, then it was poured into the container by rainfall method into 8 layers each of
100 mm thickness.
 Magazine of Civil Engineering, 17(3), 2024

2. Sand raining was paused when reaching the level above the pile tip by 50 mm. The pile was
inserted 50 mm into the soil bed. This level was calculated previously to ensure proper seating of
piles into the sand. The reached sand surface was leveled with a straight edge, and the steel mesh
of square spaces of 70 mm was used to support the piles and ensure their verticality as shown in
Fig. 8.

a b c d
Figure 8. a) Piles fixation by steel mesh, and b) horizontal leveling of
upper head of piles, c) vertical leveling of piles, and d) final level of soil and unconnected piles.
3. Continuing the raining of soil to reach the middle of the pile level, stop the raining work and remove
the steel mesh because the pile was fixed by the soil, through that the top level and side of the
unconnected piles checked carefully by water bubble in longitudinal, transverse, and diagonal
directions as shown in Fig. 9. Although the pile installation method utilized in this study is not the
same as in practice, it was chosen instead of the pushing or driving method because using these
methods may affect the unit weight of the nearby soil of the piles and consequently, the modulus
of elasticity and the wires of the strain gauges in the pile affected the soil in the pushing method.
4. Through the last steps, the instruments fixed in their location. The accelerometer number one in
the base plate of the shaking table and the numbers two, three, and four was put at three levels in
the body of soil every 200 mm apart as shown in Fig. 9.
5. Sand surface should be leveled carefully such that it should be at the same level as the top pile's
head in order to ensure that the applied cushion was definitely at one level with both the top head
of the piles and the sand surface as shown in Fig. 9.
6. The cushion layer is raining in a special mold, the special mold used in this step depends on the
dimension side and the thickness of the cushion. After that, the soil around the cushion prepared
by raining to the level required and then carefully lifting the plate mold to ensure the density did not
change.
7. In the reinforcement cushion layer tests, the geosynthetic (biaxial geogrid and non-woven
geotextile) was added to the cushion layer in single or double layers. The first layer was at the head
piles level and in case of double reinforcement the second layer was in the medium of the cushion
layer.
8. The raft was placed on the cushion and the special loading was seated on it; the accelerometer
number 5 was fixed in the raft previously.
9. The LVDTs were placed in the designated location, LVDT number one was fixed on the plate of the
shaking table to determine the horizontal movement of the shaking plate, while LVDTs number two
and three are fixed on the raft to measure the horizontal and vertical movement of the raft as shown
in Fig. 9. To prevent the inclination in the LVDT2 that may be happened because of the vertical
displacement of the raft, therefore sliding channel was used to allow vertical and horizontal
movement at the same time.
10. After that, the instruments were connected to the data acquisition including all the instruments such
as accelerometers and LVDTs. The wires extend through the piles and within the soil at the least
possible effect on the test.
11. After assurance that all the measurement devices were connected correctly to the data acquisition
system and connected to the computer in order to record and save all the readings automatically.
12. Starting the test by operating the computer and starting the shaking table work, the computer
recorded and save all the readings automatically.
13. Removing the raft model and all the instruments, discharging sand from the model box, and keeping
it in containers (water withdrawal and air drying the soil in case of saturated test).
14. Repeating the procedure mentioned above points to make another test.
 Magazine of Civil Engineering, 17(3), 2024

Figure 9. Schematic diagram of model test and instrument’s location.

3. Results and Discussion

In this study, the settlement of the unconnected pile was experimentally measured by changing the
thickness of the cushion layer, spacing between the piles, and reinforcement of the cushion layer by
geosynthetics (biaxial geogrid and non-woven geotextile) using one and two layers. The thickness of the
cushion used in the study was 2d, 3d, and 4d where d is the pile diameter. The spacing between piles used
was 3d, 4.5d, and 6d. The settlement was determined at the end of the test after 30 seconds after starting
the seismic load. The notations of conducted tests in this study are shown in Table 4.
Table 4. Notations and description of conducted shaking table tests.
Test no Cushion thickness Spacing between piles Reinforcement
T1 2d
T2 3d 3d
T3 4d
T4 2d
Without
T5 3d 4.5d
reinforcement
T6 4d
T7 2d
T8 3d 6d
T9 4d
T10 2d
T11 3d 3d One layer consists
of
T12 4d
biaxial geogrid and
T13 2d
non-woven
T14 3d 4.5d geotextile
T15 4d
T16 2d
T17 3d 3d Two layers, each
consisting of
T18 4d
biaxial geogrid and
T19 2d
non-woven
T20 3d 4.5d geotextile
T21 4d

The results of experimental tests showed that the interposed cushion layer between the raft and piles
had caused a downward relative displacement of the soil-pile system. This relative displacement is
maximum at the pile heads and extends to a certain depth below the pile heads. The relative displacement
becomes zero at a certain depth which is known as the depth of the neutral plane. Initially, the surrounding
soil and the raft settle greater than the piles which produce negative skin friction along the upper part of the
piles (above the neutral plane). Thus, the piles loaded through their heads as well as through the negative
skin friction. Because of the load on the pile head and the negative skin friction, the pile is settling and a
mobilized positive skin friction occurring at the lower part of the pile shaft, as well as resistance at the lower
tip. It is found that this mechanism was governed by the thickness and stiffness of the cushion layer.
 Magazine of Civil Engineering, 17(3), 2024

Fig. 10 shows the variation of settlement of the tested soil models without reinforcement for several
spacings between piles 3d, 4.5d, and 6d where d is the equivalent pile diameter. Also, the thickness of the
cushion layer varies from 2d to 4d. The settlement increases with increasing the thickness of the cushion
layer for a small spacing ratio of 3d but increasing the spacing between piles to 4.5d reflects a different
trend of behavior. For this ratio of spacing (4.5d), increasing the thickness of the cushion layer to 3d reduces
settlement more than 2d, but still, 4d gives higher settlement than the foundation with cushion layers of 2d
and 3d. In case of spacing between piles is 6d, the settlement is decreased with increasing the thickness
of the cushion layer, but still cushion layer of thickness 4d gives settlement higher than 3d. Generally,
reducing the spacing between piles increases the stresses in the foundation and consequently increases
the settlement, therefore the thickness of the cushion layer must be matched with spacing between piles to
get the optimum benefit of cushion layer thickness. The difference in measured settlement due to changing
the thickness of the cushion layer ranged between 0.19 to 9.75%.
To study the effects of changing the spacing between piles on the variation of settlement with time
three ratios (3d, 4.5d, and 6d) have been studied as shown in Fig. 11. In each case, the thickness of the
cushion layer is assumed constant. The settlement increased with increasing the spacing between piles,
but the variation in the final value of settlement after 30 seconds of shaking ranged from 12.75 to 14.56%.
Generally, the trend of settlement variation with time is almost the same for all tested cases. It can be seen
that most of the settlement 90% occurred during the first 20 seconds of dynamic loading, then a slight
increase noticed in the value of the settlement continues to the end of the loading period.
The stiffness of the cushion layer has an important role in increasing the efficiency of the
unconnected piled raft foundation in reducing the settlement. However, if the stiffness of the cushion layer
is not high enough, the performance of the unconnected piled raft in reducing the settlement is lower but
it’s better than the unconnected piled raft without reinforcement. Increasing the thickness of the cushion
layer without reinforcement led to increasing the value of settlement because the seismic load decreases
the angle of internal friction of soil used in the cushion layer and that makes the cushion layer creeps out
from the cushion space, but this case is reversed in case of using reinforcement in the cushion layer whether
using one or two layers of geogrid, so the settlement decreases with increasing the thickness of reinforced
cushion layer. The geosynthetic materials used in reinforcing the cushion layer helps in the redistribution
of the load between the pile and increases the stiffness of the cushion layer. Fig. 12 shows the effect of
using reinforcement in the cushion layer on the variation of settlement with time for tested soil models. The
efficiency of using reinforcement appears significantly when the thickness of the cushion layer is 3d and
more and for any spacing between piles. Also, using one layer of reinforcement reduces the settlement of
footing better than the cushion layer without reinforcement, and using two layers of reinforcement in the
cushion layer reduces the settlement better than using one layer of reinforcement as shown in Fig. 12.

Figure 10. Variations of the settlement with time for soil samples without reinforcement
of cushion and change cushion thickness.
 Magazine of Civil Engineering, 17(3), 2024

Figure 11. Variations of the settlement with time for soil samples
without reinforcement of cushion and change spacing between piles.
The mechanisms of the geosynthetic material in soil work as a tensioned membrane and help to
stabilize the soil. The interlocking of geosynthetic-soil particles applied additional horizontal stress to the
soil particles under loading and in turn increased the soil shear strength thus creating more stable soil. The
vertical component of the tensioned membrane reduced the pressures on the soil between the piles and
subsequently increased the pressures on the stationary supports (piles), so the tensioned membrane effect
was more effective at the displacement. Fig. 13 and 14 show the effects of using one and two layers of
geogrid on the settlement of soil model subjected to dynamic loading with constant spacing between piles.
The results presented in these figures exhibited decreasing the settlement with increasing the number of
reinforcement layers and this reduction was associated with increasing the thickness of the cushion layer.

Figure 12. Variations of the settlement with time for soil samples
with changing the number of reinforcement layers.
 Magazine of Civil Engineering, 17(3), 2024

Figure 13. Variations of the settlement with time for soil samples
of cushion layer reinforced with one layer of geogrid.

Figure 14. Variations of the settlement with time for soil samples
of cushion layer reinforced with two layers of geogrid.

4. Conclusions
This work is devoted to investigating the effects of improvement of the cushion layer in unconnected
piles subjected to seismic loading using one and two layers of geogrid. The lateral restraint and the
tensioned membrane effect of geosynthetic materials helped transfer the more vertical load onto the
supports. The following points can be concluded from the results of the experimental work conducted in
this study:
• The settlement of the foundation under the seismic loading decreased by reinforcing the
cushion layer with one and two layers of geogrid materials. The improvement in settlement
by reinforcement of the cushion layer was between 8 and 27%. There is no significant
difference between using one and two layers of reinforcing the cushion layer but using two
layers of reinforcement is better than using one layer.
• The reinforcement and the thickness of the cushion layer govern the amount of settlement.
Increasing the thickness of the cushion layer led to increasing the settlement of the
foundation under seismic loading in the case of using a cushion layer without reinforcement,
the settlement increased between 4 and 10% with an increase in the thickness of the cushion
from 2d to 3d and 4d. Especially in the case of the small thickness of the cushion layer and
spacing between piles.
• The variation of the spacing between piles led to a change in the settlement, where the
settlement had decreased with decreasing the value of spacing, the value of the settlement
had decreased between 7 and 14% when decreasing the spacing from 6d to 4.5d and 3d.
Reduction of the distance between piles makes the foundation system works as a block and
that reduces the settlement.
• For a constant thickness of the cushion layer using geosynthetic materials in the
reinforcement of the cushion layer reduces the settlement of the unconnected piled raft
system.
• The results of tests indicated that there had been a considerable role for the thickness and
the stiffness of the cushion layer on the loading sharing ratio between soil, raft, and piles.
However, with further increases in either the cushion thickness or cushion stiffness or both
the effect of adjustment will decrease.
• It was observed the variation of the spacing between piles led to a change in the settlement,
where the settlement had decreased with decreasing the spacing between piles which had
made the foundation system work as a block consequently leading to reduce the settlement.
 Magazine of Civil Engineering, 17(3), 2024

References
1. Davis, E.H., Poulos, H.G. The analysis of Piled Raft Systems. Australia Geotechnique Journal. 1972. 2(1). Pp. 21–27.
2. Cooke, R.W. Piled raft foundations on stiff clays – a contribution to design philosophy. Géotechnique. 1986. 36(2). Pp. 169–203.
DOI: 10.1680/geot.1986.36.2.169
3. Chow, Y.K. Iterative analysis of pile-soil-pile interaction. Géotechnique. 1987. 37(3). Pp. 321–333. DOI:
10.1680/geot.1987.37.3.321
4. Randolph, M.F. Design methods for pile groups and piled rafts. Thirteenth International Conference on Soil Mechanics and
Foundation Engineering. 1994. 5. Pp. 61–82.
5. Karkush, M.O., Aljorany, A.N. Analytical and Numerical Analysis of Piled-raft Foundation of Storage Tank. Construction in
Geotechnical Engineering. Lecture Notes in Civil Engineering. 2020. Vol. 84. Pp. 373–384.
6. Karkush, M.O., Ala, N.A. Numerical Evaluation of Foundation of Digester Tank of Sewage Treatment Plant. Civil Engineering
Journal. 2019. 5(5). Pp. 996–1006. DOI: 10.28991/cej-2019-03091306
7. Ta, L.D, Small, J.C. Analysis of piled raft systems in layered soil. International Journal for Numerical and Analytical Methods in
Geomechanics. 1996. 20(1). Pp. 57–72. DOI: 10.1002/(SICI)1096-9853(199601)20:1<57::AID-NAG807>3.0.CO;2-0
8. Horikoshi, K., Randolph, M.F. Centrifuge modelling of piled raft foundations on clay. Géotechnique. 1996. 46(4). Pp. 741–752.
DOI: 10.1680/geot.1996.46.4.741
9. Poulos, H.G. 16. Practical design procedures for piled raft foundations. Design applications of raft foundations. 2000. Pp. 425–
467. DOI: 10.1680/daorf.27657.0016
10. Kim, K.N., Lee, S.H., Kim, K.S,, Chung, C.K., Kim, M.M., Lee, H.S. Optimal pile arrangement for minimizing differential
settlements in piled raft foundations. Computers and Geotechnics. 2001. 28(4). Pp. 235–253. DOI: 10.1016/S0266-
352X(01)00002-7
11. Wong, I.H., Chang, M.F., Cao, X.D. 17. Raft foundations with disconnected settlement-reducing piles. Design applications of raft
foundations. 2000. Pp. 469–486. DOI: 10.1680/daorf.27657.0017
12. Burghignoli, A., Jamiolkowski, M., Viggiani, C. Geotechnics for the preservation of historic cities and monuments: components of
a multidisciplinary approach. Proceedings of the 14th European Conference on Soil Mechanics and Geotechnical Engineering.
2007. Pp. 24–27.
13. Alhassani, A.M.J., Aljorany, A.N. Parametric Study on Unconnected Piled Raft Foundation Using Numerical Modelling. Journal
of Engineering. 2020. 26(5). Pp. 156–171. DOI: 10.31026/j.eng.2020.05.11
14. Al-Saidi, A. Correction Factor for Methods of Installation of Piles Group in Sandy Iraqi Soils. Journal of Engineering. 2016. 22(9).
Pp. 172–181. DOI: 10.31026/j.eng.2016.09.11
15. Jebur, M.M., Ahmed, M.D., Karkush, M.O. Numerical Analysis of Under-Reamed Pile Subjected to Dynamic Loading in Sandy
Soil. IOP Conference Series: Materials Science and Engineering. 2020. 671(1). Article no 012084. DOI: 10.1088/1757-
899X/671/1/012084
16. American Society of Testing and Materials (ASTM). Standard test method for specific gravity of soil solids by water pycnometer.
ASTM D854. 2006.
17. American Society of Testing and Materials (ASTM). Standard test method for classification of soils for engineering purposes
(Unified Soil Classification System). ASTM D2487–06. 2006.
18. American Society of Testing and Materials (ASTM). Standard test method for particle size-analysis of soils. ASTM D422–63
(2002). 2006.
19. American Society of Testing and Materials (ASTM). Standard test method for direct shear test of soils under consolidated drained
conditions. ASTM D3040–04. 2006.
20. American Society of Testing and Materials (ASTM). Standard test method for minimum index density and unit weight of soils and
calculation of relative density. ASTM D4254–00 (2006). 2006.
21. Karkush, M.O., Ali, S.D., Saidik, N.M., Al-Delfee, A.N. Numerical Modeling of Sheet Pile Quay Wall Performance Subjected to
Earthquake. Geotechnical Engineering and Sustainable Construction. 2022. Pp. 355–365. DOI: 10.1007/978-981-16-6277-5_28
22. Karkush, M.O., Jafar, G.S. Effects of Surcharge on the Behavior of Passive Piles in Sandy Soil. International Journal of Scientific
& Engineering Research. 2015. 6(10). Pp. 392–397.
23. Karkush, M.O., Resol, D.A. Remediation of Sandy Soil Contaminated with Industrial Wastewater. International Journal of Civil
Engineering. 2017. 15(3). Pp. 441–449. DOI:10.1007/s40999-017-0195-6
24. Jihad, A., Karkush, M. Behavior of Geosynthetics-Strengthened Saturated Cushion Layer in Disconnected Piled-Raft System
under Seismic Loading. Indian Geotechnical Journal. 2024. DOI: 10.1007/s40098-024-00877-x
25. AlSaadi, K.A., Almurshedi, A.D., Karkush, M. Effect of Geosynthetics-Reinforced Cushion on the Behavior of Partial Connected
Piled Raft Foundation in Dry and Saturated Sandy Soil Using Shaking Table. Indian Geotechnical Journal. 2024. DOI:
10.1007/s40098-024-00890-0

Information about authors:

Mahdi Karkush, Doctor of Technical Sciences
ORCID: https://orcid.org/0000-0003-1304-0303
E-mail: mahdi_karkush@yahoo.com

Amer G. Jihad,
E-mail: amer.ghani@qu.edu.iq

Received: 12.02.2023. Approved after reviewing: 14.04.2023. Accepted: 29.11.2023.