DOI: https://doi.org/10.4038/engineer.v57i1.

7638

ENGINEER
ENGINEER- -Vol.
Vol.LVII,
LVII,No.
No.01,01,
pp.pp.
[57-70],
[page2024
range], 2024
©
© The
TheInstitution
Institutionof
ofEngineers,
Engineers,Sri
SriLanka
Lanka

Failure of Gravel Compaction Pile (GCP) Improved

Embankment – Case Study
T. Nekasiny, N.H. Priyankara, N.B.G. Imali and M.S. Nilawfer

Abstract: Due to the scarcity of suitable land, it becomes necessary to utilize marshy lands
consisting of soft soil for infrastructure development. Soft soil has a problematic nature due to its high
moisture content, high compressibility, high void ratio and very low shear strength. Hence, it is a
responsibility of the geotechnical engineers to overcome these issues by adopting suitable ground
improvement techniques. Gravel Compaction Pile (GCP) is one of the most popular soft ground
improvement techniques used in the field. This technique has been successfully applied during the
construction of the Colombo-Katunayake Expressway project and the Outer Circular Highway project.
However, GCP technique has failed in the Southern Expressway Extension Project from Matara to
Beliatta section. Therefore, in this study, the causes of the failure of the GCP improved embankment
section were studied based on the field records. The slope stability of the embankment during
construction was analysed using Matsuo and Kawamura's method. Back analysis revealed that once
shear failure occurred in the subsurface, the shear strength of the soft soil reduces to its residual value
and it takes longer time to regain its original strength. Further, it was noted that when the soft soil
thickness is greater than 10 - 12 m, it becomes extremely challenging to improve the soft ground
without any reinforcement. Therefore, it is strongly recommended to accurately interpret the
subsurface characteristics in order to select the most suitable ground improvement technique.

Keywords: Gravel Compaction Pile (GCP), Matsuo and Kawamura method, Slope stability
analysis, Very soft peaty clay

1. Introduction Further, Karunawardena et al. [11] stated that

peaty clay found in OCH has a compression
The rapid development and population growth
index of 1.95 and undrained shear strength of
of a country would increase the demand for
substantial number of infrastructures by 7.2–19.0 kPa. As such, peaty soil does not
provide favourable conditions for construction.
making land very scarce. The scarcity of
suitable lands for construction encourages the Therefore, geotechnical engineers face many
challenges because of peaty soil which cannot
widespread use of areas underlain by weak soil
support heavy loads, hence construction can
deposits that are considered either marginal or
result in excessive settlement. In order to
inappropriate. Due to the scarcity of suitable
overcome these challenges, there are two ways,
land, a considerable percentage of the number
of infrastructure development projects in Sri namely, transfer the structural load to an
underlying hard stratum through piles or
Lanka, such as the Colombo – Katunayake
Expressway (CKE), Southern Expressway, improve the engineering properties of soft soil.
Outer Circular Highway (OCH) and Central Eng. T. Nekasiny, AMIE(SL), B.Sc. Eng. (Hons) (Ruhuna),
Expressway was constructed on flood plains Department of Civil and Environmental Engineering,
University of Ruhuna, Sri Lanka.
and marshy terrain made up of extremely soft Email:tnekasiny@gmail.com
peat, organic soils, and clay [3][11-15]. https://orcid.org/0009-0003-6891-8379
Eng. (Dr.) N.H. Priyankara, CEng, MIE(SL), B.Sc. E ng .
(Hons). (Moratuwa), M.Eng. (AIT), PhD (Tohoku),
Road embankment construction over peat MSLGS, Senior Lecturer, Department of Civil and
deposits is quite challenging, because of the Environmental Engineering, University of Ruhuna, Sri
inherent properties of peat such as high Lanka.
moisture content, high compressibility, high Email:nadeej@cee.ruh.ac.lk
https://orcid.org/0000-0002-8775-7246
void ratio and very low shear strength [1][3] Eng. N.B.G. Imali, CEng, MIE (SL), B.Sc. E ng . ( Ho ns ).
[12-14]. The primary consolidation of peat is (Peradeniya), M.Eng. (Ruhuna), Road Development
very high with significant secondary Authority, Sri Lanka.
compression [11-13]. Madhusanka and Email:gimali69@yahoo.com
Eng. M.S. Nilawfer, CEng., MIE(SL), B.Sc. Eng. ( Hons) .
Kulathilaka [14] reported that Sri Lankan (Moratuwa), Dip in Arbitration, PG Dip in Structural Eng.
peaty soil has a high moisture content of about (Moratuwa), Additional Director, Expressway Operation,
300 %, low shear strength of about 0.99 kN/m2 , Maintenance and Management Division, Road Development
and a compression index ( ) of 1.51. Authority, Sri Lanka.
Email:nilawfershareef@yahoo.com

1
57 ENGINEER
ENGINEER
However, transferring of load through piles to 2. Details of GCP Trial
a hard stratum is not an economical solution for Embankment
roads occupying a large plan area and To expand the expressway network in Sri
moderately loaded buildings. Improving Lanka, the Government of Sri Lanka decided to
engineering properties of soft soil is the most extend the existing Southern expressway from
economical option and it is a responsibility of Matara to Mattala (Figure 2). Section 1 of the
the geotechnical engineers to find appropriate Southern Expressway Extension Project from
ground improvement technique/s based on the Matara to Beliatta, is mainly going through the
subsurface soil conditions. Nilwala flood plain which consists of 10.1 km
viaducts, 0.6 km bridges/ underpass/drainage
Gravel Compaction Pile (GCP) is one of the
box culverts, 15.3 km non-treated area and
most popular soft ground improvement
technique that aims to increase load-bearing 4.0 km soft ground treated area [18]. By
considering the subsurface soil profile, initially
capacity and reduce settlement by densification
GCP technique had been proposed as the soft
of subsoil [1] [11]. The installation of GCP
ground treatment method. In order to examine
consists of a sequence of routine work as shown
the performance of the GCP improved ground,
in Figure 1. In this method, a 40 cm diameter
a trial embankment section was done at
casing is forced into the ground, down to the
chainage (Ch.) 7+405 to 7+475. The typical cross
required depth, under vibration at a frequency
of 10 Hz [16]. Then, casing is retracted stepwise section of the GCP improved embankment is
shown in Figure 3.
while supplying granular material into the
casing and compacted the granular material by
It can be noted that Existing Ground Level
casing tip under vertical vibration. As this
(EGL) is about 0.250 m MSL whereas design
process is repeated, a well compacted 70 cm
Road Finished Level (RFL) is about
diameter GCP is created.
9.760 m MSL as shown in Figure 3. The side
slope of the embankment was planned to
maintain as 1:1.5 (1 vertical to 1.5 horizontal)
above the berm section and 1:1.8 below the
berm. Width of the carriageway is 24.40 m.

2.1 Subsurface Soil Profile

The subsurface soil profile in the trial area was
investigated by advancing five boreholes as
illustrated in Figure 4. Two boreholes, BH-21
and BH-22, were done in the initial stage of the
project at the centre line at Ch. 7+405 and
7+500, respectively, and three boreholes, BH-04,
BH-05 and BH-06, were done in the detail
Figure 1 - GCP Construction Procedure design stage at Ch. 7+452 covering the entire
cross section.
Even though GCP technique has been
According to the borehole logs at Ch. 7+405
successfully utilized in the CKE project and
(BH-21) and Ch. 7+500 (BH-22), soft soil
OCH project [11][12] to improve soft soils, this thickness was identified as 17.0 m and 13.0 m,
technique failed when it was applied to the soft
respectively. The subsurface soil profile across
soil improvement in Southern Expressway Ch. 7+452 can be idealized as shown in Figure
Extension Project from Matara to Beliatta
5. It can be seen that soft soil thickness
(Section – 1), causing huge financial loss to the
significantly varied from left to right, where
contractor. As such, this paper presents a
soft soil thickness on the Left-Hand Side (LHS)
detailed analysis of the causes of GCP
is about 4.0 m whereas that on the Right-Hand
improved embankment failure and lessons that Side (RHS) is about 16.0 m.
can be learnt from the incident.

ENGINEER
ENGINEER 2
58
 Section 1

Figure 2 - Project location – Southern Expressway Extension Project (Section 1)

Figure 3 - Typical Embankment at Ch. 7+405 – 7+475 (Not to Scale)

2.2 Geotechnical Parameters of compression index were found as 0.8 m2 /year

Subsurface Soil and 0.1, respectively. For a conservative design,
The physical and engineering properties of the coefficient of consolidation in horizontal
subsurface soil based on field and lab test data direction ( ) was assumed as 0.8 m2 /year.
are summarized in Table 1.
2.3 Gravel Compaction Pile (GCP)
Based on the data presented in Table 1, it can be Installation
noted that subsurface mainly consists of very Before installation of GCP, 1.5 m high soil fill
soft peaty clay followed by completely had been placed on the existing ground as the
weathered rock layer. The peaty clay has very working platform for the movement of GCP
high moisture content with very high void installation machine. Based on the geotechnical
ratio. Both laboratory triaxial tests and insitu parameters presented in Table 1, GCPs were
vane shear test results indicated that the peaty installed at 1.3 m spacing on the Right-Hand
clay has very low shear strength. Furthermore, Side (RHS) and at 1.6 m spacing on the Left-
laboratory oedometer test results indicated that Hand Side (LHS) in square pattern as there is a
the peaty clay has high compression index with considerable variation of the soft soil thickness
very high modified compression index of about from LHS to RHS as shown in Figure 5. The
0.3. The average coefficient of consolidation in spacing of the GCP was taken as 1.6 m when
vertical direction ( ) and modified secondary the soft soil thickness is less than 10 m.

3
59 ENGINEER
ENGINEER
 Table 1 – Properties of Subsurface Soil

Property Value
Blackish
grey very
Soil type
soft peaty
clay
Natural moisture content (%) 107 – 150
Unit weight (kN/m3 ) 11.38 – 12.85
Coefficient of consolidation ( )
0.8 – 1.5
(m2 /year)
Compression Index ( 0.997 – 1.140
Initial void ratio ( ) 2.40 – 2.62
Modified Compression Index
0.28 – 0.33

Modified Secondary
Compression Index 0.08 – 0.1

Undrained shear strength

5.75
(kPa)
Undrained friction angle (u) (o) 0
Liquid Limit (%) 109 - 159
Plastic Limit (%) 51 - 72

Figure 4 - Borehole Locations within the GCP

Trial Embankment Section

Figure 5 - Variation of Soft Soil Thickness

Across Ch. 7+452
With the help of vertical vibrating excitation of
the vibro-hammer, casing was penetrated
through the soft soil to the hard stratum. When
the applied current (generally 35A) of the GCP
machine to penetrate the casing was Figure 6 – GCP Termination Criteria Based on
significantly increased (up to about 80A), the Electric Current Variation at Ch. 7+470
driving of the casing was terminated as
Further, based on the GCP installation records,
depicted in Figure 6. When the required current
variation of soft soil thickness along a particular
to penetrate the casing is significantly
cross section can be identified as shown in
increased, that implies casing has reached the
Figure 7. By combining borehole logs and GCP
hard stratum.
installation records, variation of soft soil
thickness within the trial embankment section
can be idealized as shown in Figure 8.

ENGINEER
ENGINEER 4
60
Based on this idealization, it can be seen that
soft soil thickness in RHS varied from 12-18 m
whereas in the LHS, it is about 4-10 m.
Particularly between Ch. 7+405 and 7+420, the
thickness of soft soil varied from 16 to 18 m
even on the LHS. This clearly illustrates the
highly fluctuating nature of the subsurface soil
profile.

2.4 Quality Assurance of GCP

The quality of the GCP was investigated at
thirteen (13) GCP locations by conducting
Dynamic Cone Penetration Tests (DCPT) at the
center of the GCP [4][5][17]. Then, the DCPT-
Measured (DCPT–M) values were converted
into SPT–N values using the correlation as
shown in Equation 1 [4].

… (1)

The variation of SPT-N values over depth at

different GCP locations are presented in
Figure 9. Figure 8 - Soft Soil Thickness Contour Map
within the Trial Embankment Section
It can be clearly seen that the SPT–N values
gradually increased with depth due to Then friction angle of the compacted GCP
overburden pressure. Even though SPT-N materials (s ) was estimated using Equation 2
values increased with depth, when the [6], where N is the field measured SPT-N value.
overburden correction was applied to estimate By taking the average SPT-N value of 16 as
the SPT-N value, the corrected SPT-N values indicated in Figure 9, it can be calculated that
are around 40–50 irrespective of the depth. This the friction angle of the GCP material (s ) is
implies GCP was properly compacted. about 32 0 .

…(2)

Figure 7 - Typical GCP Installation Records at Ch. 7+460

5
61 ENGINEER
ENGINEER
Figure 9 - SPT N Value over Depth at GCP
Locations
Figure 10 – Instrumentations in the GCP Trial
2.5 Embankment Construction and Field Section
Instrumentation
Soon after the GCP installation, a 0.5 m thick 3. Field Observations
gravel mat was placed as a drainage layer. The embankment was constructed in four
After that, a geotextile was laid as a separator phases, namely, Embankment construction-
and embankment filling commenced. The Stage 1, Waiting period, Berm construction and
performance of the embankment during Embankment construction–Stage 2. The
construction was evaluated based on the field summary of the trial embankment construction
instrumentation and monitoring data. During is illustrated in Table 2. The graphical
the construction of the embankment, the field presentation of the four phases of the
behaviour was monitored using 16 settlement embankment construction is shown in
plates, 5 vibrating wire piezometers, 2 Figure 11.
inclinometers, and 38 surface stakes. The field
instrumentation arrangement is shown in It can be noted that, due to upheaving of the
Figure 10. ground during GCP installation, the working
platform elevation before start of the
Settlement plates were used to measure the embankment filling was about 2.0 m MSL. The
settlement of the soft soil below the settlement of the subsurface soft ground due to
embankment, while piezometers were used to embankment construction is presented in
measure the variation of the pore water Figure 12. It can be seen that settlement on the
pressure. Surface stakes were installed near the RHS is much higher than that of LHS.
toe of the embankment to measure the lateral Settlement after 277 days at RHS and LHS were
movement of the subsoil, close to the ground about 1.35 m and 0.3 m, respectively. The
surface (not varying with depth) and to check variation of lateral displacement in RHS of the
the stability during construction. Lateral embankment over depth at different time
displacement measurements obtained from intervals is illustrated in Figure 13. The
inclinometers indicated the continuous variation of rate of lateral displacement over
horizontal movement of the subsoils with time at RHS of the embankment at the critical
depth under the embankment. depths is shown in Figure 14.

ENGINEER
ENGINEER 6
62
Table 2 – Summary of Trial Embankment of embankment construction built to a height of
Construction 4.0 m MSL elevation.
Duration
Phase Details
(Days) According to the data obtained, the variation of
pore water pressure over time was drawn as
Platform construction and
0 - shown in Figure 15 for Ch. 7+450 and Ch. 7+435
GCP installation
at depths of 3.75 m and 8.45 m, respectively. It
Embankment construction is clearly seen that pore water pressure is
1 100
– Stage 1 gradually increasing with time during all the
2 Waiting period 89 phases. Even during the waiting period,
without embankment filling, the excess pore
3 Berm construction 40
water pressure was not dissipated.
Embankment construction
4 48
– Stage 2 4. Analysis
4.1 Slope Stability Based on Matsuo and
Kawamura Method
It has been widely recognized that the failure of
soft ground is closely related to the magnitude
and history of the deformation which had taken
place before final failure. It uses information
from practical measurements in the field to
control embankment construction to be safe
and efficient. When the soft ground is under
loading, in addition to the consolidation, there
is a possibility for horizontal soil flow (shear
deformation). This fact makes it difficult to
theoretically distinguish the displacement and
the failure of soft ground. It is obvious that
Figure 11 – Variation of Embankment failure occurs when the progress of shear
Elevation Over Time at Ch. 7+455 deformation is faster than the consolidation
settlement. Therefore, the graphical method
proposed by Matsuo and Kawamura [7] is
commonly used to estimate the stability of the
embankments constructed on soft ground
based on the field monitoring data as shown in
Figure 16.

Figure 16 indicates a relationship between

settlement and ratio of lateral displacement to
settlement, and each curve corresponds to a
different Factor of Safety (FOS) value, varying
from 1.0 to 1.67. Then, based on the obtained
field monitoring data, embankment settlement
versus ratio of lateral displacement to
Figure 12 – Settlement Over Time at Ch. 7+455 settlement was plotted in Matsuo and
Kawamura’s diagram as shown in Figure 17.
To monitor the excess pore water pressure This clearly indicates the variation of FOS at
dissipation, it is necessary to install the different depths of the soft soil against slope
piezometers before starting the embankment failure of embankment at different phases of
construction. According to the information construction.
provided, three (3) piezometers were installed
soon after the GCP installation. However, none It can be observed that FOS gradually decreases
of the piezometers were under working with the increment of embankment height at
condition during the embankment construction Stage 1 (Figure 17) due to huge lateral
- Stage 1. As such, two additional piezometers displacement of about 200 mm on RHS of the
were installed about 60 days after the initiation embankment as shown in Figure 13.

7
63 ENGINEER
ENGINEER
 Figure 13 – Variation of Lateral Displacement Over Depth at RHS of the Trial Embankment

According to the field practice norms, if the rate

of lateral displacement is between 4 to 5
mm/day, frequent monitoring is required.
Further, FOS at the end of Stage 1 of
embankment construction can be estimated as
1.24. Therefore, the embankment filling was
stopped as indicated in Figure 11 (waiting
period) expecting an improvement in FOS.

The continuous increase of pore water pressure

(Figure 15) during embankment construction –
Stage 1 clearly indicated the unsafe nature of
the embankment at that phase. As such,
embankment construction was stopped for a
period of 89 days (waiting period) expecting
the dissipation of excess pore water pressure to
improve the stability.
Figure 14 – Variation of Rate of Lateral
Displacement Over Time for RHS of Although the excess pore water pressure was
Embankment at Ch. 7+455 not dissipated during the waiting period, the
rate of increment of pore water pressure
The significant outward lateral movement can significantly decreased at shallow depth as
be observed down to a depth of 12.0 m at the illustrated in Figure 15(a). However, at greater
end of Stage 1 of embankment construction. depth [Figure 15(b)], pore water pressure has
The rate of lateral displacement per day has increased even during the waiting period. This
increased up to 4-5 mm/day as indicated in behaviour clearly indicates that during Stage 1
Figure 14. of the embankment construction, the existing

ENGINEER
ENGINEER 8
64
 path in the GCP has been significantly reduced
due to smear effect and/ or higher horizontal
permeability of soft soil has been diminished
due to disturbance during GCP installation. As
a result, excess pore water pressure has not
dissipated during the embankment
construction – Stage 1 and even during the
waiting period. Moreover, it can be noted that if
length of the GCP was greater, the vertical
drainage through the GCP due to capillary
action may not have been effective. Even
though rate of lateral displacement has been
significantly reduced during the waiting
period, additional 70 mm outward lateral
movement can be observed during the waiting
period. As such, this behaviour further reduces
the FOS against embankment slope failure
down to 1.20 even during the waiting period.

Then, a berm has been introduced up to the

elevation of 5.33 m MSL in both RHS and LHS
of the embankment to reduce the lateral
displacement. The newly introduced berm and
revised design cross section is shown as a
dotted line in Figure 7. It is well known that a
berm provides an additional lateral support
against outward movement of the
embankment. Berm construction process and
settlement due to berm construction are
depicted in Figure 11 and Figure 12,
respectively.
Figure 15 – Variation of Pore Water Pressure
with Time
It can be noted that settlement due to berm
construction in RHS is about 0.4 m whereas in
LHS is about 0.05 m. This clearly indicates the
effects of variation of soft soil thickness across
the trial embankment area. By providing the
berm, embankment moved about 10 mm in the
inward direction as illustrated in Figure 13.
Further, rate of lateral displacement has
changed to 1-2 mm/day in the inward direction
as depicted in Figure 14. However, there is no
reduction in pore water pressure due to berm
construction. It can be clearly seen that
settlement versus lateral displacement to
settlement ratio graphs in the Matsuo and
Figure 16 - Matsuo and Kawamura’s Diagram Kawamura’s [7] plot have been moved to the
[7] left, indicating a slight improvement in
stability. This behaviour is clearly illustrated in
GCPs were ineffective in helping to transfer the the enlarge view of the variation of FOS at the
pore water at greater depth to the ground depth of 3.0 m at different phases of
surface [10]. However, during the design stage, embankment construction in the same figure.
generally it is assumed that pore water in the
soft ground flow towards the GCP due to Once the FOS has increased and indicated a
higher horizontal permeability of the soft soil slight improvement in stability, embankment
and collected pore water transfer to the ground construction – Stage 2 has been started as
surface through the GCP. Based on field shown in Figure 11.
monitoring data, it seems that vertical drainage

9
65 ENGINEER
ENGINEER
Figure 17 - Variation of FOS in RHS of the Embankment using Matsuo and Kawamura Method [7]
at Ch. 7+455
When the embankment elevation increased to ground surface. According to the field
7.0 m MSL, embankment has again started to observations, only three GCPs were installed
move in the outward direction as shown in beyond the toe of the embankment. In addition,
Figure 13. Even though the rate of lateral 200/200 geogrids (tension in both transverse
displacement is about 2.0 mm/day, the FOS and longitudinal direction is 200 kN/m2 ) were
curves in the Matsuo and Kawamura’s [7] plot placed in the embankment to enhance the slope
moved in the vertical direction, indicating stability.
further reduction in FOS. Reduced FOS values
are about 1.13 – 1.17 at depths 3 m to 7 m which 10

is less than the required minimum FOS of 1.20

3 4

200/200 Geogrids
5 Embankment fill
for short term stability [2]. Therefore, it is Berm 13

3
2 1 5 14

4
Berm

clearly indicated that the embankment is

16 12 1 6 15 23
0 7 8
9 2 10
11

Unimproved
5

unstable at this stage. ground GCP improved composite ground

Elevation (m)

17
18 22
-5
6 Unimproved ground
7

4.2 Determination of Shear Strength -10

21

Parameters Based on Back Analysis -15 20 19 24

The above embankment conditions were 8

Completely weathered rock

numerically modelled using GEOSLOPE -20

SLOPE/W software as shown in Figure 18. 26

-25
25

Using the back-analysis technique, shear 0 10 20 30 40 50 60 70 80 90 100

Distance (m)
strength parameters of GCP improved
Figure 18 – Numerical Model used for the
composite ground at different phases of the
Back Analysis
embankment construction were estimated.
Spencer’s method was used as the constitutive
The geotechnical parameters of other materials
model while “Entry and Exit” method was used
(except GCP improved composite ground) used
to generate the slip surfaces. Since the
for the analysis are shown in Table 3. Effective
occurrence of soft soil is critical on RHS of the
shear strength parameters (c’ and ’) were used
embankment, only RHS of the trial
for embankment fill material and completely
embankment section was considered for the
weathered rock while undrained shear strength
slope stability analysis. In the model, it was
parameters (cu and u) were used for
assumed that only the soft soil portion below
unimproved ground for the analysis. Based on
the embankment was improved with GCP and
the GCP spacing and diameter, area
the water table was maintained at the existing

ENGINEER
ENGINEER 10
66
replacement ratio (as ) was estimated as 0.228. [9], undrained shear strength gain depends on
Further, it was assumed that unit weight of the the applied external load and degree of
GCP improved composite ground ( ) does consolidation of the soft soil. Phase 1 to phase 3
not vary with the embankment fill height, and of the embankment constructions is under the
average shear strength method was used to same applied external vertical load
compute the unit weight of GCP improved (embankment height = 6.0 m). Even though, 229
composite ground [1] as presented in days have been spent for the construction from
Equation 3. By taking unit weight of GCP beginning to phase 3, there is no improvement
material as 22 kN/m2 , the average unit weight in shear strength gain during this period due to
of GCP improved composite ground can be less dissipation of excess pore water pressure.
estimated as 14.28 kN/m3 . Poor dissipation of excess pore water pressure
may result in the very slow primary
… (3) consolidation. Further, introduction of a berm
has not had any influence on the shear strength
In this analysis, assuming that FOS obtained gain even though it reduced the outward
through Matsuo and Kawamura method [7] is movement of the embankment.
correct, shear strength parameters of the GCP
improved composite ground were estimated As such, it is very clear that there is no shear
from trial and error method. Since friction angle strength gain of soft soil due to GCP installation
of the GCP improved composite ground is and stage construction of the embankment. The
governed mainly by the GCP material, it was continuous reduction of FOS and outward
assumed that friction angle of GCP improved lateral movement of embankment led to
composite ground is not varying with the abandoning of the GCP ground improvement
embankment height. As such, the average technique in the project and huge financial loss
friction angle (avg ) of GCP improved composite has occurred to the contractor.
ground is estimated using Equation 4 [1], where Table 4 – Back Analysis Results
is the stress ratio in GCP. By taking stress
Shear
concentration ratio (n) as 3, was estimated to
FOS strength
be 2.06. parameters

… (4) Construction

Cohesion cu
Matsuo and
Kawamura

angle  u (o)
SLOPE/W
phase

Friction
(kPa)
The shear strength parameters obtained
through back analysis are shown in Table 4 and
critical failure surfaces at different stages of
embankment construction are illustrated in Stage 1 1.240 1.240 17.3 16.4
Figure 19. Based on above explanation, the Waiting period 1.200 1.203 15.7 16.4
friction angle of the GCP improved composite
After berm
ground can be computed as 16.4 0 . 1.205 1.209 12.0 16.4
construction
Stage 2 1.130 1.131 12.0 16.4
Table 3 – Material Parameters used for
SLOPE/W Analysis
5. Conclusions and

Material Recommendations
(kN/m3 ) (kPa) ( )
Based on this research study, the following
Embankment Fill 20 5 32
conclusions can be drawn:
Completely 1. According to limited borehole investigations,
Weathered Rock 21 10 38
soft soil thickness in the RHS was identified
Unimproved as 15.6 m and GCP spacing was decided by
ground 12 5.75 0 taking the average soft soil thickness as
GCP improved 13.7 m. However, during GCP installation, it
ground 14.28 - - was realized that average thickness of soft
soil on the RHS is about 16 m and, at some
locations, it may be around 18 m. As such, it
Based on the results presented in Table 4, it can can be concluded that GCP spacing was
be noted that undrained cohesion (cu) gradually decided by wrong interpretation of the soft
decreases with the increment of embankment soil thickness. This clearly indicates the
height. According to Skempton and Bjerrum importance of proper site investigation prior

11
67 ENGINEER
ENGINEER
 10

1.240

5 200/200 Geogrids Embankment fill

Unimproved ground GCP improved composite ground

-5
Elevation (m)

-10

Unimproved ground
-15

Completely weathred rock

-20

-25
0 10 20 30 40 50 60 70 80 90 100

Distance (m)

(a) At the end of embankment construction – Stage 1

10

1.203

200/200 Geogrids
5
Embankment fill

Unimproved ground
GCP improved composite ground
-5
Elevation (m)

-10

Unimproved ground
-15

-20 Completely weathered rock

-25
0 10 20 30 40 50 60 70 80 90 100

Distance (m)

(b) At the end of waiting period

10

1.209

200/200 Geogrids
5
Embankment fill
Berm
Berm

Unimproved ground
GCP improved composite ground
-5
Elevation

-10

Unimproved ground
-15

-20 Completely weathered rock

-25
0 10 20 30 40 50 60 70 80 90 100

Distance

(c) At the end of berm construction

10
1.131

200/200 Geogrids
5
Embankment fill
Berm Berm

Unimproved ground
GCP improved composite ground
Elevation (m)

-5

-10

Unimproved ground
-15

-20 Completely weathered rock

-25
0 10 20 30 40 50 60 70 80 90 100

Distance (m)

(d) At the end of embankment construction – Stage 2

Figure 19 – Critical Failure Surfaces at Different Stages of Embankment Construction

ENGINEER
ENGINEER 12
68
 to start of the GCP installation. 7. The slow rate of dissipation of excess pore
water pressure after embankment loading is
2. As per depth measurements carried out
due to the shear failure of the soft soil. Due
during the GCP installation, soft soil
to disturbance of the soft soil during GCP
thickness was higher than the value
installation and other field activities,
considered for the GCP design. However, no
horizontal permeability of soft soil reduces
precautions have been taken to revise the
and the smear zone surrounding GCP
original design or no countermeasures have
increases. As a result, the vertical drainage
been implemented before start of the
path within the GCP reduces, causing
embankment filling. Hence steps should
reduction in dissipation of pore water
have been taken to revise the original design
pressure. This behaviour clearly indicates
if the field records indicated a somewhat
the importance of proper maintenance of
different soil profile than the soil profile
field activities without disturbing the
used for the original design, and, if required,
subsurface during GCP ground
install additional GCPs in between already
improvement.
installed GCPs.
8. Once shear failure has occurred, the strength
3. It was observed that only three GCPs were
of the soft soil reduces to its residual value
installed beyond the toe of the embankment
and it takes a longer time to regain its
at a particular row. However, it was realized
original strength [16]. After the shear failure,
that three GCPs are insufficient to provide
even constructing a berm to provide lateral
toe support to the embankment due to high
support will not be successful.
thickness of the soft soil layer. If the
thickness of the soft soil is greater, adequate
stability could have been achieved against Acknowledgment
slope failure by installing precast concrete
piles as toe support. The authors wish to acknowledge the Road
Development Authority (RDA) for granting
4. The construction of the trial embankment permission to use data related to Southern
started immediately after the GCP Expressway Extension Project from Matara to
installation. However, it is a well-known fact Beliatta section.
that the strength and stiffness of the
surrounding soft soil are reduced as a result References
of disturbance during GCP installation.
Hence, it is necessary to allow adequate time 1. Bergado, D.T., Anderson, L.R., Miura, N. &
to stabilize the surrounding soft ground to Balasubramaniam, A.S. (1996) “Soft Ground
recover its strength. The waiting period can Improvement: in Low land and Other
be decided based on the pore pressure Environments”, American Society of Civil
Engineers, New York, 427 p.
measurements using piezometers.
5. It can be noted that, in the trial embankment 2. ICTAD Publication No. – SCA/5, (2009)
section, piezometers were installed after “Standard Specifications for Construction and
starting the embankment construction. This Maintenance of Roads and Bridges”.
clearly indicates that embankment filling was
3. Priyankara, N. H., Wijesooriya, R.M.S.D.,
started without proper prior monitoring of
Wickramasinghe, W.R.M.B.E. & Yapa, S.T.A.
pore water pressure. Hence, it is
(2009) “Suitability of Quarry dust in
recommended to establish a proper field Geotechnical Applications to Improve
monitoring system, prior to starting Engineering Properties”, Engineer -Journal of
embankment construction. Institution of Engineers Sri Lanka, Vol. 42, No. 3,
pp. 7-13.
6. Based on the field observations and data
analysis, it can be concluded that, when the 4. Spagnoli, G. (2007) “An Empirical Correlation
soft soil thickness is more than 10 – 12 m, it between Different Dynamic Penetrometers”,
is really difficult to improve the soft ground Electronic Journal of Geotechnical Engineering
without any reinforcement. Hence it is (EJGE), Vol. 13, pp. 1-11.
strongly recommended to select the most
suitable ground improvement technique by 5. Marawan, M.S. & Ahmed, F. (2013) “Estimation
accurately interpreting the soft soil of Deformation Modulus of Gravelly Soils using
Dynamic Cone Penetration Tests”, Ain Shams
thickness.
Engineering Journal, Vol. 4, pp. 633-640.

13
69 ENGINEER
ENGINEER
6. Shioi, Y. & Fukui, J. (1982) “Application of N- 17. Salagoda, R. & Yoon, S. (2003) “Dynamic Cone
value to Design of Foundation in Japan”, 2nd Penetration Test (DCPT) for Subgrade
ESOPT, Vol. 1, pp. 40-93. Assessment Final report,” FHWA/IN/JTRP-
2002/30.
7. Matsuo, M. and Kawamura, K. (1977) “Diagram 18. http://rda.gov.lk/supported/project-progress-
for Construction Control of Embankment on pmu/esep/esep.html, Visited, 13/07/2023.
Soft Ground”, Japanese Geotechnical Society, Vol.
17, pp 37–52, ISSN 0038-0806.

8. Priyamali, M.W.S. & Priyankara, N.H. (2023)

“The Influence of Preconsolidation Pressure on
Undrained Shear Strength Characteristics of
Peaty Clay in Sri Lanka”, Engineer - Journal of
Institution of Engineers Sri Lanka, (Accepted).

9. Skempton, A.W. & Bjerrum, L. (1957) “A

Contribution to the Settlement Analysis of
Foundations on Clay”, Geotechnique, Vol. 7, pp.
168-178.

10. Dangwal, K.K. & Aggarwal, M. (2019)

“Irrigation in Hilly Areas by Capillary Lift”,
International Journal of Innovative Technology and
Exploring Engineering, Vol. 8(12).
doi:10.35940/ijitee.8.12., pp. 3227-3230.

11. Karunawardena, A., Ooka, A. & Nithiwana, W.

(2017) “Improvement of Sri Lankan Peaty Clay
using the Gravel Compaction Pile Method for
the Construction of a Highway Embankment.”,
Proceedings of the 19th International Conference on
Soil Mechanics and Geotechnical Engineering, Seoul
2017, pp. 3003-3006.

12. Karunawardena, A. & Toki, M. (2013) “Design

and Performance of Highway Embankments
Constructed over Sri Lankan Peaty Soils”,
Proceedings of the 18th International Conference on
Soil Mechanics and Geotechnical Engineering, Paris
2013, pp. 2949-2952.

13. Kulathilaka, S.A.S. & Fernando, M. (2015)

“Effect of Preload on Secondary Consolidation
of Peaty Clay”, International Conference on
Geotechnical Engineering (ICGE – 2015), pp. 371-
374.

14. Madhusanka, K.A.C. & Kulathilaka, S.A.S.

(2015). “Possible use of Paddy Husk Ash in
Improvement of Engineering Characteristics of
Peaty Clay”, International Conference on
Geotechnical Engineering (ICGE – 2015), pp. 407–
410.

15. Saputhanthri, D.R. & Kulathilaka, S.A.S. (2011)

“Enhancement of Engineering Characteristics o f
Peaty Clay due to Mixing with Cement”, Annual
Transactions of the Institute of Engineers Sri Lanka,
pp. 118-126.

16. Nozu, M. & Ohbayashi, J. (1998) “Application of

the Static Sand Compaction Pile Method to
Loose Sandy Soil”, Proceeding of International
Symposium on Problematic soils, IS – TOHOKU,
Vol. 2, pp. 751-755.

ENGINEER
ENGINEER 14
70