Geotextiles and Geomembranes xxx (xxxx) xxx

Contents lists available at ScienceDirect

Geotextiles and Geomembranes

journal homepage: www.elsevier.com/locate/geotexmem

Use of geosynthetic clay liner as a remedial measure of claystone

degradation in Lam Ta Khong hydropower plant
Suttisak Soralump a, *, Avishek Shrestha b, Apiniti Jotisankasa a, Chinoros Thongthamchart b,
Rattatam Isaroran b
a
Geotechnical Engineering Division, Department of Civil Engineering, Kasetsart University, Bangkok, 10900, Thailand
b
Geotechnical Engineering Research and Development Center, Kasetsart University, Bangkok, 10900, Thailand

A R T I C L E I N F O A B S T R A C T

Keywords: In this study, the downstream slope of a dam impounding the upper reservoir of Lam Ta Khong (LTK) hydro­
Geosynthetic clay liner electric energy storage in Thailand was found to slide at a higher pace during the rainy season. After a thorough
Ruzi grass site investigation, laboratory tests, and numerical modeling to identify the main cause of the movement, it was
Claystone degradation
found that as rainfall infiltrated the upper soil layer, the claystone of the downstream slope deteriorated when in
Lam Ta Khong
contact with the water. As a remedial measure, 174,750 m2 of geosynthetic clay liner (GCL) was used to cover the
entire downstream slope of the dam, and proved to be an effective and economical solution for reducing the
ongoing movement. The GCL included a textured high-density polyethylene (HDPE) layer for improving the
resistance and minimizing the slippage at the interface between the GCL and underlying rock. Before the dam’s
remediation, the settlement point demonstrated a movement of nearly 0.1 m/year. In contrast, after the
placement of the GCL, almost all settlement points moved less than 0.1 m for a recorded period of more than 4
years.

1. Introduction construction of the reservoir, which were consequently repaired.

Moreover, high-density polyethylene (HDPE) sheets were overlaid on
A pumped hydropower electric storage (PHES) system is different the surface of the pond; after this process, in 2004, the reservoir
from conventional hydroelectricity plants, and is generally designed to impounding began. During the operation phase, the settlement points
meet the peak power demand. A PHES generally consists of two reser­ showed movements, indicating that the downstream slope of the Upper-
voirs, in which energy stored in the form of the potential energy of water pond was settling; horizontal displacements were also observed. It was
is pumped from a lower reservoir to a higher reservoir. During periods of indicated that the degradation of the claystone caused the settlement of
low power demand, low-cost electric power is used to pump water from the dam body, which became more pronounced during rainfall. The
the lower reservoir to the higher reservoir. When power is in high de­ degrading nature of claystone in wet-dry processes has been reported
mand, the stored water is then freed through the turbines, generating (Rodvinij and Wattanachai, 2017; Sadisun et al., 2005) and many
electricity (Rehman et al., 2015). To cope with the high peak power studies characterizing its properties have been conducted (Chiarelli
demands in Thailand, the first PHES system was constructed in the et al., 2003; Monfared et al., 2011; Tian et al., 2014; Zhang et al., 2012).
Nakhon Ratchasima province, and was designed to connect the existing Similarly, problems associated with using claystone as a construction
lower reservoir of Lam Ta Khong to the upper reservoir (hereafter called material in dams have also been reported (Deschamps et al., 1999). As a
“Upper-pond”). It was situated 350 m above an underground power­ means to stop the degradation of claystone and prolong the reservoir’s
house consisting of 250 MW reversible Francis-type generating units. life, the use of a geosynthetic clay liner (GCL) with textured HDPE has
The Upper-pond is rectangular in plan and impounds 9.9 million been proposed for mitigating the downstream slope of a dam.
cubic meter of water, with a surface area of 0.34 sq. km. Between 2000 GCLs consist of a thin layer of sodium or calcium bentonite bonded to
and 2002, cracks were observed on the floor of the pond during the a layer or layers of geosynthetic materials, i.e., geotextiles or

* Corresponding author.
E-mail addresses: [email protected] (S. Soralump), [email protected] (A. Shrestha), [email protected] (A. Jotisankasa), [email protected]
(C. Thongthamchart), [email protected] (R. Isaroran).

https://doi.org/10.1016/j.geotexmem.2021.04.001
Received 25 August 2020; Received in revised form 8 March 2021; Accepted 5 April 2021
0266-1144/© 2021 Elsevier Ltd. All rights reserved.

Please cite this article as: Suttisak Soralump, Geotextiles and Geomembranes, https://doi.org/10.1016/j.geotexmem.2021.04.001
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Daniel, 1996; Ishimori and Katsumi, 2012; Kul and Ören, 2018; Mazzieri
and Pasqualini, 2000; Petrov et al., 1997; Shan, 1991). It is fascinating to
many researchers how this cost-effective material with low permeability
can be used and adopted by design engineers and environmental
agencies in various fields, especially for controlling seepage (Bhowmik
et al., 2018; Gilson-Beck, 2019).
This study presents a success story on the use of GCL with textured
HDPE to inhibit the settlement and displacement of the Upper-pond of
the Lam Ta Khong PHES systems. Various laboratory and field tests,
along with numerical modeling of the dam, indicated that the primary
cause of the settlement was the degradation of the claystone when in
contact with water. From the monitoring data of approximately 4 years
after GCL placement, it was found that the rate of movement decreased
significantly, with most of the settlement points demonstrating less than
0.1 m movement, whereas before the remediation, the rate was
Fig. 1. Upper-pond of Lam ta Khong pumped hydropower electric storage approximately 0.1 m/year.
(PHES); Survey lines (A–H) for measuring electrical resistivity in the study area.
2. Upper-pond construction
geomembranes (Bouazza, 2002). Many studies have been conducted on
the hydraulic conductivity of GCLs with a variety of permeants, and for The body of the Lam Ta Khong (LTK) Upper-pond (Fig. 1) is a rock-
different states of the GCLs (AbdelRazek and Rowe, 2019; Boardman and filled dam located in the northeastern part of Thailand, between the

Fig. 2. Cross-section of the dam.

Fig. 3. Test pits (TP1–TP4), settlement points (S1–S17), and observation wells (W1–W13) installed on the dam body.

2
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Fig. 4. Dam movement behavior (a) Plot of settlement along with time and rainfall (b) Plot of horizontal displacement along with time and rainfall.

Sikhio and Pak Chong districts of the Nakhon Ratchasima Province. The (Brachiaria ruziziensis) was planted in the topsoil layer. In Thailand, this
dam was constructed using local materials predominant at the site. A species can be grown in both highlands and lowlands, and under various
cut-and-fill method was adopted. The rocks excavated from the foun­ soil conditions. Owing to its relatively fast growth, this grass is
dation were used as fills in the construction of the dam body, which commonly used for erosion control and shallow soil reinforcement
primarily consisted of sandstone and claystone. To maximize the use of (Teerawattanasuk et al., 2014).
the materials obtained from the excavation and owing to the difficulty in Overall, the present structure design and the norms followed for the
transporting materials uphill, claystone was allowed in the fill at the construction of the dam body were considered as adequate.
downstream face, provided that a gentle vertical-to-horizontal
embankment slope of 1:2.5 was maintained. The cross-section at the 3. Instrumentations and dam movement behavior
greatest height (approximately 50 m) of the dam is shown in Fig. 2, and
it maintained a freeboard of 2 m throughout the total length of 2170 m. 3.1. Instruments installed
The asphaltic concrete facing was originally provided as an impervious
surface, although later cracks were observed, and consequently, the dam During the construction of the dam, instrumentation was placed
leaked. In response to this, the basin area was then overlaid with HDPE within the dam body and its crest to monitor the dam’s safety and
sheets, which could effectively prevent water from seeping out of the identify any movement patterns. Settlement points S1–S12 were
basin. installed on the downstream roadway around the dam ridge, whereas
There was also a topsoil layer approximately 30 cm thick, provi­ S13–S17 were installed on the dam’s northern and western slopes, as
sioned for planting grass by means of hydro-seeding. Ruzi grass shown in Fig. 3. The groundwater levels were monitored at points

3
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Fig. 5. Longitudinal cracks (a) Location of the crack, (b) Observed on the dam crest, (c) Close up view of longitudinal crack.

Fig. 6. Difference between water level and dam foundation.

W1–W12 at the foot of the dam. The locations of the test pits TP1 – TP4 through settlement points S1–S12 was studied beginning on July 2nd,
were dug to obtain the soil samples for the laboratory tests, as well as for 2002, whereas the settlement readings from S13–S17 on the slope of the
the in situ investigations. Settlement plates ME1–ME5 (Fig. 2), were 3C zone were recorded starting from February 7th, 2003.
installed at the 3A zone of the dam body at five different locations, and
piezometers P1–P5 were installed at the foundation of the highest sec­
tion of the dam to measure the pore water pressure and detect seepage. A 3.2. Behavior of the dam movement
rain gauge was also installed at this site. The behavior of the dam
Although a dam, in general, is not allowed to have a settlement of

4
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Fig. 7. Resistivity survey results (a) Line E (b) Line F.

more than 1% of its height (USBR, 2012), the Upper-pond demonstrated The plot of settlement, along with the time and rainfall for S1–S17, is
a vertical settlement of 93.4 cm for 50 m at the settlement point S15. The shown in Fig. 4(a). The rainfall (in mm/day) is shown as a bar chart.
reported vertical settlement and horizontal displacement were Even after more than five years of operation, it can be seen that the
measured through the movement of the settlement points along the y settlement points S5–S9, S11, and S13–S17 demonstrate a continuous
and x directions, respectively. increase in settlement compared to the others. Even among them, the

5
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Fig. 8. Ruzi grass (a) On the downstream slope of the dam (b) Ruzi grass forming a hay-like surface that allowed considerable infiltration.

settlements of S13 to S17, which lie on the slope of the 3C zone, are 4.1. Resistivity survey
greater, as this zone consists of claystone. For S1–S12, because they lie
on the crest of the Upper-pond, some of them are supported by the 3B A resistivity survey is a non-invasive method that can be used to map
zone, which comprises sandstone and is less degradable than claystone. a subsurface, and has been used to determine water leakages (Manataki
However, for all of the settlement points, the settlement rate seems to et al., 2014) including from embankments and dams (Al-Fares, 2014;
accelerate each time after rainfall. Panthi and Soralump, 2020; Sjödahl et al., 2010). The resistivity survey
Furthermore, it was also observed that the dam not only settled in a in Lam Ta Khong was performed along eight lines (A-H) (Fig. 1) to
vertical direction, but also moved in the horizontal direction. A plot of evaluate the wetting area within the dam. The survey was conducted in
the horizontal displacement of the settlement points with time and two periods: during the dry season in March 2009, and during the rainy
rainfall is shown in Fig. 4(b). If the settlement was occurring primarily season in August 2010. The results from the resistivity test were inter­
owing to consolidation, then it would have mostly been vertical; how­ preted based on the inverse model, which is a technique employed in
ever, in this case, the dam had a pronounced lateral displacement, many studies (Abu-Shariah, 2009; Šumanovac and Weisser, 2001). From
indicating that the slope was sliding. Points S5 and S7 (installed near the the results of the resistivity survey conducted at different lines, it was
dam crest) showed the maximum lateral displacement, and longitudinal found that there were portions with relatively low electrical resistance
cracks were also observed between them, as shown in Fig. 5. As observed (6–96 Ω) spreading along the reservoir. Some typical results along lines
through the settlement points, the magnitudes of the movement in the E and F are shown in Fig. 7.
vertical direction and displacement in the horizontal direction are Line E is located at the maximum dam height section corner, and a
nearly the same. This indicates the sliding of the downstream slope, low-resistivity material is dispersed therein. The low resistivity at the
rather than the action of consolidation alone. middle portion suggests that there might be a presence of a water pocket
One of the possible hypotheses for this rainfall-induced slope as well. Line F runs perpendicular to Line E, and it can be seen that the
movement was that the water level rose above the dam foundation, spot where the low resistivity was demonstrated occurred in both survey
subsequently weakening it. However, based on the observations from periods, and was quite stable. This observation of the low-resistivity area
W1–W12, the groundwater remained well below the foundation level in Line F also confirms the results for Line E. The presence of the low
for at least 5 m throughout the entire length, as shown in Fig. 6. It can be resistivity indicates the moisture content. The source of the moisture
inferred that the groundwater played a minimal role in deteriorating the content cannot be identified through the leakage of the dam, as the in­
foundation and causing dam settlement. However, the dam body side of the dam is paved and overlaid by HDPE, and there are no signs of
continued to show the settlement at an unflagging rate, which was un­ leakage. Thus, the only possible source could be rainwater infiltrating
usual; hence, investigations to inhibit the dam settlement and avoid its the dam body.
failure were necessary.

4.2. Effect of Ruzi grass on pore water pressure measurement

4. Site investigation and laboratory tests

As discussed in Section 2, Ruzi grass (Brachiaria ruziziensis) was

A site investigation and several laboratory tests, in-situ tests, were
planted on the dam surface by hydroseeding. During the site investiga­
performed to determine the cause of the unusual dam settlement. Four
tion, it was found that the Ruzi grass was fully grown. During the
test pits were excavated to explore the characteristics of the 3C material.
summer, the grass dried off, with residues accumulating on top of each
A resistivity survey was performed in situ to observe the soil profile.
other to form a hay-like surface, as shown in Fig. 8. This hay-like surface
Samples of materials were also collected from the site, and various
provided a mulching role, thereby reducing runoff and erosion by
laboratory tests were conducted, such as large direct shear tests, triaxial
improving the infiltration. Accordingly, the moisture was preserved
tests, and slake durability tests. Furthermore, a field permeability test
beneath the surface. Despite its benefits in reducing erosion, it was
was conducted on site to determine the coefficient of permeability of the
suspected that the Ruzi grass may have had an adverse role in retaining
cover soil and the 3C zone.
the infiltrated water, which could soften and degrade the underlying
claystone. Therefore, miniature tensiometers (Mairaing et al., 2012;
Shrestha et al., 2019) capable of measuring both positive and negative

6
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Fig. 9. Pore water pressure measurement with and without Ruzi grass on top surface.

Fig. 10. Observed phenomena with digging of test pits a) Accumulation of water after leaving for a night b) Intact rocks being decomposed.

Table 1
Basic soil properties and soil classification according to Unified Soil Classification System (USCS).
Soil from Depth Soil Specific Atterberg’s Limit Test Sieve Analysis Test (% Finer)
Test Pit (m) Type Gravity
Liquid Plastic Plasticity 3/ 3/8" #4 #8 #10 #40 #80 #100 #200
no. (Gs)
Limit (LL) Limit Index (PI) % 4"
% (PL)%

TP-1 5 CL 2.63 39.39 18.21 21.18 100 96.52 90.36 86.95 86.51 84.59 83.58 83.37 81.95
TP-2 0.3 CL 2.55 34.63 19.77 14.86 100 100 93.3 86.08 84.93 81.33 77.51 75.64 69.15
3 CL 2.69 42.18 19.51 22.67 100 100 93.96 84.07 81.32 72.49 69.05 68.28 65.81
5 CL 2.66 41.08 24.31 16.76 100 100 77.2 64.26 62.39 56.77 55.04 54.52 51.35
4
TP-3 2.5 CL 2.63 39.36 20.39 18.97 100 97.99 97.20 96.26 95.97 94.82 94.71 94.62 93.75
TP-4 2.5 CL 2.7 37.85 20.38 17.46 100 100 98.35 97.74 97.59 96.91 96.86 96.67 95.08

pore water pressures within a range of 100–600 kPa were used for increase in suction (Fredlund et al., 1993, 1996; Kayadelen et al., 2007).
monitoring the 3C zone with Ruzi grass on the slope surface. After Accordingly, this indicated that a certain strength of the soil could be
approximately a week, Ruzi grass was removed from the top surface, and recovered.
the pore water pressure was measured again at the same location. From these measurements, it can be deduced that the Ruzi grass’s
Fig. 9 shows higher readings for the pore-water pressure in the area hay-like surface entrapped water and increased the soil moisture con­
with Ruzi grass, and once the Ruzi grass was removed, there was an tent. The increase in moisture content led to a reduction in the soil
increase in suction in the same location at all depths. At a depth of 0.2 m, strength. These revelations increased the confidence that the Upper-
positive readings of pore water pressure were also noted for the area pond slope movement could be stopped if the infiltrated water was
with Ruzi grass, indicating saturation of the soil. In contrast, after prevented from entering the underlying claystone.
removing the Ruzi grass from the top surface, a significant suction was
developed. It is well-established that a soil’s strength increases with an

7
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Fig. 11. Grain size distribution for the soils from test pits (3C zone).

revealed that the claystone had already decomposed and become soil-
Table 2
like, as shown in Fig. 10(b).
Jar Index (JI) values indicating the deterioration of rock upon soaking.
During the excavation of the test pits, a thin-wall miniature sampler
No. Rock Location Jar Classification (AHTD (63 mm in diameter and approximately 30 mm high) was used to collect
Sample Index TM, 2010)
(JI)
undisturbed soil samples for use in the direct shear test, triaxial test, and
soil water characteristic curve test. Furthermore, box samples (15 × 15
1 Claystone Dam Foundation 1 Soft
cm2 and 30 × 30 cm2) representative of the foundation material were
2 Claystone Rock fill material (Dam 1 Soft
3 Siltstone body), Random Zone 3 Soft collected for a large direct shear test. The disturbed soil sample collected
4 Sandstone 6 Hard was used for basic property tests, a sieve analysis, specific gravity tests,
and durability tests.
The basic soil properties of the samples collected at different depths
Table 3 from the test pits are listed in Table 1. The results from the sieve analysis
Slake index of the rocks tested. of the soil samples collected at the varying depths of the different test
pits comprising zone 3C are shown in Fig. 11. The grain size distribution
Material Test conditions Slake Durability
Classification
curves clearly indicate that the soils at a shallow depth are finer than
those at greater depths.
Soaked Natural (Gamble, 1971)

Id1, Id2, Id1, Id2, Soaked Natural

% % % % 4.4. Durability test

Claystone from 0.87 0 16.49 0.18 Very Very

foundation Low Low
Jar slake and slake durability tests were performed to determine the
Rock fill material 62.77 40.33 82.47 54.31 Low Low durability of the materials used in the dam construction. The jar slake
(Claystone + test was conducted based on procedures described by the Kentucky
Siltstone + Transportation Research Program (Hopkins and Gilpin, 1981), where
Sandstone)
Jar Index (JI) values from 1 to 6 could be assigned. A JI value of 1
corresponds to the softest nature of the rock, demonstrating a complete
breakdown into mud or flakes, whereas a JI value of 6 indicates that the
Table 4 rock remains intact. The JI values of the rock samples collected from the
Permeability test results at different locations. site are listed in Table 2.
Test pit Location of the test Discharge (Q, cm3/ Permeability (k, cm/ The slake durability test (Franklin and Chandra, 1972), as outlined in
no. s) s) American Society for Testing and Materials (ASTM) D4644 (ASTM,
TP1 1.50 m below soil 56.6 × 10− 2
6.47 × 10− 5 2008), was conducted to determine the weathering resistance of the
surface rocks used in the construction of the dam (claystone, siltstone, and
TP1 at the soil surface 49.08 × 10− 1 6.16 × 10− 4
sandstone). The results from the slake-durability tests are given in
TP1 at the soil surface 27.3 × 10− 1 4.69 × 10− 4

TP2 1.50 m below soil 25.95 × 10− 1 2.91 × 10− 4 Table 3. From the results, it can be inferred that the durability of the rock
surface materials is very low for claystone. For the mixture of claystone, silt­
TP2 at the soil surface 89.96 × 10− 2 1.21 × 10− 4
stone, and sandstone obtained from the dam body, the durability is
TP3 at the soil surface 27.84 × 10− 1 3.54 × 10− 4
comparatively higher, but when classified on the standard scale, it re­
TP4 at the soil surface 16.9 × 10− 1 2.09 × 10− 4
mains low.

4.3. Test pits and basic soil properties 4.5. Soil permeability test

Four circular test pits (TP-1, TP-2, TP-3, and TP-4) with a 1-m The soil permeability test was conducted using a constant head
diameter and 5-m depth were excavated to examine the ground condi­ borehole test method, as described in the USBR Earth Manual 2 (USBR,
tions. During the exploration of TP-1, a wet portion was detected at a 1990). The tests were conducted in situ at seven different locations. The
depth of 1.10 m, consistent with the findings of the resistivity survey. results of the permeability test are tabulated in Table 4, where the
When left for a night, the circular pit received a significant amount of permeability ranges from 6.47 × 10− 5 cm/s at a depth of 1.5 m (TP1) to
water, as shown in Fig. 10(a). At a depth of 3 m in TP-1 and TP-3, it was 2.09 × 10− 4 cm/s at the surface of TP4.

8
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Fig. 12. Soil water characteristic curve during drying and wetting processes. Fig. 13. Permeability function of the soil samples obtained from the test pits at
different depths.
This result demonstrates that the claystone might have decomposed
into clay, as the permeability readings are high. However, even for clay, (top, middle, and bottom, as shown in Fig. 13). The weight of the sample
the permeability readings are still considered high (Horpibulsuk et al., was monitored continuously using a data-logging balance. The volu­
2011; Tavenas et al., 1983; Wilkinson and Shipley, 1972), most of which metric water content (θ) of the soil was then calculated using the
are on the order of 10− 4 cm/s. Accordingly, it can be suggested that the soil-phase relationship, as follows:
clay derived from the decomposed claystone has greater macrovoids
between the clay aggregates, resulting in the higher permeability. θ=
Gs.w
(1)
1+e
In the above, Gs is the specific gravity of the soil, w is the water
4.6. Soil water retention behavior and permeability function content measured through the change in the weight of the sample, and e
is the void ratio of the sample.
The soil water retention curve (SWRC) and permeability function of Furthermore, the suction values obtained at the three locations were
the soil, two of the major parameters required for the seepage analysis, used to calculate the hydraulic gradient i, as follows:
were determined in the laboratory from an undisturbed sample collected ( )
using a thin-wall miniature sampler, based on a tensiometer suction d z − γsw
measurement and the instantaneous profile method explained by i= (2)
Mairaing et al. (2012) and Jotisankasa and Sirirattanachat (2017). dz
The retention capacity for water during the drying and wetting Here, z is the elevation head as measured from the base of the sample
processes is depicted in Fig. 12. The test involved soaking the sample in of each tensiometer, s is the measured matric suction, and γ w is the unit
water for approximately a week before the drying instantaneous profile weight of the water. Likewise, the discharge velocity, v, at any particular
(IP) test, where the soil suction was increased from zero to approxi­ time is obtained as follows (Jotisankasa and Sirirattanachat, 2017;
mately 100 kPa by the end of the test. The drying IP test was performed Mairaing et al., 2012):
to determine the drying portion of the SWRC and the unsaturated
permeability function. The top surface of the soil sample was left dVw
v= (3)
exposed, whereas the base of the sample was underlain by an imper­ A.dt
meable steel base plate. The suction readings were continuously moni­ In the above, dVw is the change in the volume of the soil mass (as
tored using miniature tensiometers (Jotisankasa and Vathananukij, obtained through the change in the soil mass), A is the cross-sectional
2008; Shrestha et al., 2019) at three locations on the sides of the sample

9
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Fig. 14. Soil water retention curve (SWRC) and permeability function used in the seepage analysis.

Table 5 Table 8
Results from triaxial unconsolidated undrained (UU) test. Parameters used in the numerical modeling.
Test Method Degree of Saturation Undrained shear strength, Su (kPa) Parameters Values used in the model

Target Before After Inside the dam body Foundation of dam

UU test 80 78.5 78.5 203 Zone Zone Zone Sandstone Claystone Sound
100 99.42 99.42 37 2A, 3B 3C Rock
80 81.95 81.95 202 2B, 3A
80 78.12 78.12 204
Cohesion, c 0 10 5 0 100 0
100 100 100 54
(kPa)
100 99.62 99.62 68
Friction 40 30 29 40 24 45
Angle
(Degrees)
Total Unit 18.5 18.5 21 23 20 24
Table 6 Weight,
Results from large direct shear test. (kN/m3)
Young’s 20000 20000 20000 500,000 400,000 700,000
Test Method Test Sample Size Cohesion, c Friction angle
Modulus,
condition (cm2) (kPa) (degrees)
E (kPa)
Large direct Natural 15 × 15 80 20 Poisson’s 0.25 0.25 0.25 0.2 0.2 0.2
shear test Natural 30 × 30 100 24 Ratio, μ
Soaked 15 × 15 7 24
Soaked 15 × 15 6 23
Soaked 30 × 30 11 24 different moisture contents. After each suction measurement, water was
added to the sample to reach an approximately 1–2% moisture content,
based on using a very fine spray to obtain suctions at different moisture
Table 7 contents. After changing the moisture content to a desirable amount, the
Residual strength parameters obtained by cut-plane method. sample was sealed and curated for 3–4 days to ensure suction equilib­
rium throughout the sample. The suction was then measured for the
Test method Cohesion, c (kPa) Friction angle (degrees)
corresponding moisture content, and the process was repeated until the
Residual strength test 0 31.56 sample reached a near-saturation level. The obtained SWRC points and
2.035 31.4
hydraulic conductivity for the 3C zone were used as data point
0 37.33
3.805 27.74 functions.
0.033 32.98 Because the soil samples from zones 3A and 3 B could not be
13.212 28.73 collected, the volumetric water content – suction relationship was esti­
0 37.94
mated based on the soil grain size distribution, dry density, and specific
5.708 29.35
1.691 26.89
gravity, following a physio-empirical model (Arya and Paris, 1981)
10 18.16 calibrated based on residual soils in Thailand (Jotisankasa et al., 2009).
Similarly, the hydraulic conductivity function was estimated according
to (Jackson, 1972). The reliability of the estimated SWRC and hydraulic
area, and t is the elapsed time. Linear regression was used to calculate conductivity were confirmed by the good agreement between the esti­
the slope of the water volume change with time. Using Darcy’s law, the mations and experimental results from the laboratory for the 3C zone.
permeability k at any depth z for any average elapsed time dt can be Hence, the estimated SWRC and k-function were considered reliable for
calculated by dividing the water flow rate by the corresponding hy­ the other zones. The SWRC and k-function used in the seepage analyses
draulic head gradient (Ng and Leung, 2012), as follows: for zones 3A and 3B B are shown in Fig. 14.
v From the SWRC and k-function plot, it can be deduced that even at a
k= (4)
i higher suction, the water-retaining capacity of the 3C material is rela­
The wetting curve was obtained by a point-wise measurement tively high compared to the materials from the 3A and 3B zones. The
method, incorporating incremental soil suction measurements at permeability function of the 3C material also suggests that it can carry a

10
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Fig. 15. Comparison between the predicted and measured values of settlement from (a) ME1, (b) ME2 (c) ME3 (d) ME4 (e) ME5.

fair amount of water, even at higher suction values. of the soil, γ s is the unit weight of the soil, and γw is the unit weight of the
water. The sample was then air-sealed inside a moisture chamber for 3–4
days to ascertain the equal distribution of moisture throughout the
4.7. Triaxial unconsolidated undrained (UU) test, large direct shear test sample.
and residual strength test The results indicate a sharp decrease in the undrained shear strength
when the degree of saturation increases from 80% to 100%, as indicated
The triaxial unconsolidated undrained (UU) test (ASTM, 2015) is one in Table 5. Notably, the cohesion decreases by approximately 66%–82%
of the recommended tests for evaluating stability at the as the degree of saturation increases from 80% to 100%, as observed in
end-of-construction loading in dams (USBR, 2011). Because the down­ samples obtained from the TP-3 and TP-4 locations. It can be inferred
stream slope of the dam is not fully saturated at the end of construction, that the strength of these decomposed claystone from TP-3 and TP-4 is
the resulting change in the undrained shear strength (Su) owing to the very sensitive to changes in water content, which reflects the deterio­
variation in moisture content was more likely from the change in degree rating nature of the claystone upon contact with water.
of saturation from 80% to 100%. The test was performed on undisturbed Similar results were also obtained from a large direct shear test based
soil samples retrieved from TP-3 and TP-4. on ASTM D4554 (ASTM, 2012), which was conducted on claystone
The initial moisture content (w) of each sample was determined. The samples collected from the foundation area. The test was performed on
target degree of saturation was then achieved by spraying water using a sample sizes of 15 × 15 cm2 and 30 × 30 cm2, both under natural and
very fine spray, and subsequently measuring the weight of the sample soaked test conditions. Here, “natural condition” refers to a test in which
with the added water. The degree of saturation (Sr) of the sample ob­ the moisture content of the undisturbed sample was in situ and in an
tained through the change in moisture content (w) was then calculated undisturbed state as much as possible. The “soaked condition” indicates
using equations (5) and (6). that the sample was immersed in water for 24 h prior to and throughout
(
Gs (1 + w)
) the test. The soaking procedure did not strictly ensure the complete
e= .γw − 1 (5) saturation of the sample, but rather simulated a real field scenario in
γs
which a zero-suction condition could be ascertained. A series of tests on
w.Gs four sets of the block sample were performed to determine the shear
Sr = (6) strength parameters, cohesion c, and angle of friction φ. The results
e
obtained from the large direct shear test are shown in Table 6, and
In the above, e is the void ratio of the sample, Gs is the specific gravity

11
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Fig. 16. Deformation analysis upon filling with water (a) X-displacement (b) Y-displacement.

Fig. 17. Intensity of rainfall during August–October 2007 used in the analysis.

suggest that the cohesion of the foundation claystone is drastically 5. Deformation and seepage modeling
decreased once in contact with the water.
A residual shear strength test was also conducted using the cut-plane After a thorough site investigation and laboratory tests, the Upper-
method (Head, 1994), where a pre-shear plane was defined prior to pond was modeled using Geo-studio suite software for the deforma­
testing. In the residual state, the soil undergoes a large strain, the tion and seepage analysis. The modeling was intended to simulate the
interparticle bonds are broken, and clay particles are oriented parallel to observed behavior, so as to understand and predict the various scenarios
the slip plane. The results of the residual strength tests are given in that might occur in the real field. Sigma/W and Seep/W of Geostudio
Table 7. From these results, it is observed that some samples demon­ were used for the deformation and seepage analysis, respectively.
strate a cohesion intercept, which is thought to be owing to various The model was given a fixed support at the base of the ground, and
factors such as suction, bonding, and the size effect of the shear box. A roller supports were provided at its side ends. This ensured the vertical
cohesion value of 10 kPa and friction angle of 18.2◦ at a 68.1% degree of settlement of the ground owing to consolidation and dam erection, but
saturation were chosen as the reduced strength parameters for the nu­ prevented the ground from moving sideways.
merical modeling. The validity of this chosen operational shear strength
was subsequently demonstrated based on the agreement between the
5.1. Material properties
analysis and field observations.
The material properties used in the numerical modeling are

12
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Fig. 18. Pore water pressure contour considering continuous rainfall.

summarized in Table 8. Most of the parameters were determined 5.2. Modes of analysis
through laboratory and in situ testing. For the soil parameters that could
not be tested, references were made to the design reports. The cross-section chosen for the numerical modeling of the dam is
For instance, the Young’s modulus was obtained from Equation (7), shown in Fig. 2, and represents the section with the maximum height. It
where the settlement (δs) was measured through the settlement plates consists of Zones 2A, 2B, 3A, 3B, and 3C, together with sandstone,
(ME1–ME5) installed in the 3A zone of the dam body. claystone, and sound rock at its foundation. The analysis was performed
considering three scenarios: (a) normal conditions, (b) filled with water,
Erc = γH d1/δs (7)
and (c) rainfall infiltration into the dam.
here, Erc is the modulus during construction (in MPa); γ is the unit
weight of the rockfill (in kN/m3); H is the height of the fill above that 5.2.1. Analysis under normal condition
layer (in m); d1 is the thickness of the layer of the dam (in m); and δs is The first step in numerical modeling is to determine the initial
the settlement of layer (in mm). effective stresses acting on the foundation and dam body. The dam was
The materials used in the Upper-pond were similar to those used in constructed in stages, meaning that the filling was conducted in several
the Vajiralongkorn Dam in Thailand. The material properties of the steps. Therefore, the analysis chosen was a couple-consolidation type.
Vajiralongkorn dam were obtained from the Electricity Generating Au­ The settlement values from the model were then compared with the
thority of Thailand (EGAT, 1984), as well as other published literature settlement values obtained from the settlement plates (ME1–ME5). The
(Davis et al., 2011; Hunter and Fell, 2003). The Young’s moduli for left axis consisted of the elevation in meters above sea level and
zones 2A, 2 B, 3A, 3 B, and foundation material were estimated based on demonstrated the various filling steps involved in the construction of the
these reviews. For the 3C material, triaxial consolidated undrained (CU) dam with time, whereas the right axis showed the settlement in meters,
tests (ASTM, 2011) were performed, along with consideration of the and represented the settlement occurring at the corresponding settle­
settlement plate readings (ME1–ME5) as in Equation (7), to obtain a ment plates (ME1–ME5) with time. The values simulated in the model
Young’s Modulus of 20 MPa. Similarly, the friction and cohesion values were found to be in line with the measured values (depicted in Fig. 15),
for zones 2A, 2B, 3A, and 3B were chosen based on the design report, and therefore justified the model as representative of the field.
whereas for the 3C zone, the values obtained from the triaxial CU test
were employed. 5.2.2. Analysis with the filled water
Skempton’s pore water coefficients A and B (Skempton, 1954) are As the water is filled in the dam, the load and deformation mode of
required to predict the change in a pore water pressure with a change in the analysis is chosen so as to acquire the resulting deformation from the
the total stresses. In this study, the coefficients were determined through infill of water in the Upper-pond. The displacements in both the x and y
triaxial CU tests. The value of the coefficient B was 0.98, and the coef­ directions from the numerical modeling are shown in Fig. 16 (a) and
ficient A, which is a function of the deviator stress, was dictated by Fig. 16 (b), respectively, these are not in line with the actual behaviors as
Skempton’s equation (Equation (8)), as follows: demonstrated by the measuring instruments in the field, i.e., the
movement in the downstream slope (3C zone). This implies that the
Δu = B[Δσ1+A(Δσ1-Δσ3)] (8) ongoing movement was not caused by the filling of water into the dam.
Here, Δu is the pore pressure change, Δσ1–Δσ3 is the change in
5.2.3. Analysis under rainfall infiltration
deviator stress, and A and B are Skempton’s coefficients.
Rainfall infiltration into the dam body is expected to affect the water
pressure changes inside the dam body. The water seepage into the soil

13
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Fig. 19. Infinite slope stability results (a) Pore water pressure distribution in the dam (b) Pore water pressure along the line C2 (c) Factor of Safety (FS) using strength
parameters from direct shear test (d) FS using reduced shear strength.

depends on many soil characteristics and properties, especially its 0.002, 0.02, 0.08, and 0.04 m/day, respectively.
permeability under unsaturated conditions, and the soil-water retention From the seepage analysis, it is found that most of the time there is a
curve. The governing differential equation (Ng and Shi, 1998) for the bulb of water (October 11, 2007) in zone 3C at the level of +652 to
water flow in unsaturated soils is presented in Equation (9), as follows: +659 m or at a depth of 1.5–8.5 m from the surface, consistent with
( ) ( ) ( ) findings from the resistivity survey (where pockets of water were
∂ ∂h ∂ ∂h ∂uw
kx + ky + Q = mw (9) observed). As shown in Fig. 18, the pore water pressure change is more
∂x ∂x ∂y ∂y ∂t
noticeable in the upper slope section (C1) than at the lower slope (C2).
In the above, kx and ky are the hydraulic conductivity of soil in the x
and y directions, respectively, h is the total hydraulic head, Q is the 6. Slope stability analysis
applied boundary flux, and mw is the slope of the water retention curve.
From Equation (9), we can see that the change in pore water pressure A slope stability analysis was performed to determine whether the
results in a change in effective stress and is mainly dependent on Q, the downstream slope would fail under the maximum rainfall intensity. A
seepage of rainwater into or out of the slope, and k/mw, i.e., the ratio of period covering the maximum intensity of rainfall (from August 26 to
the hydraulic conductivity to the slope of the SWRC curve. Therefore, 29, 2007) was chosen. Before performing slope stability analysis, a
soils with high values of k and low values of mw exhibit a rapid rate of seepage analysis was conducted with this intensity of rainfall. It revealed
infiltration and seepage. that the downstream slope would be saturated, resulting in a perched
The seepage analysis was performed under transient conditions, as it water table, as shown in Fig. 19(a). Shallow landslides are associated
was the most likely scenario in the field. It involved changing the in­ with the advancement of a wetting front in an unsaturated zone
tensity of the rain with time, thereby potentially affecting the water resulting from the infiltration of rainfall, which increases the water
pressure inside the dam. The year 2007 received the highest daily content and reduces the matric suction of the soil (Chae et al., 2015). In
rainfall (0.124 m/day); therefore, this year was chosen for the analysis, this situation, there is a possibility of the occurrence of a shallow depth
as shown in Fig. 17. Continuous rainfall data from October 9 to 12, 2007 failure parallel to the slope of the dam; therefore, an infinite slope model
were chosen. The intensities of the rainfall during this period were was used to evaluate the stability, as follows:

14
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Fig. 20. Deformation analysis using residual strength (a) X-displacement (b) Y-displacement (c) Deformation mesh.

c + (γ*zcos2 β)*tanϕ − uw *tanϕb

′ ′
was modeled using the reduced strength, the displacement pattern was
F= (10) similar to that occurring in the field (Fig. 20). If the strength of the soil
γ*zsinβ*cosβ
reached the reduced strength, the slope might have failed and caused a
here, c’ is the effective cohesion, γ is the unit weight of the soil, β rep­ disaster. Thus, an appropriate remedy was urgently needed to stop the
resents the angle of the slope, i.e., 20◦ , uw is the pore water pressure, and ongoing movement in the 3C zone.
ϕ and ϕb are the angles of shearing resistance with respect to the
′

effective stress and suction, respectively, such that when uw > 0, ϕb = 7. Use of geosynthetic clay liner (GCL)
ϕ , z is the depth of the failure surface.
′

As a remedial measure, a GCL consisting of bentonite sandwiched

It was found that the soil at a depth of 3 m was saturated, resulting
between a geotextile and textured HDPE was overlaid on the dam’s
from the perched water table. On the day of the highest rainfall (Aug.
downstream slope. The material specifications of the GCL are listed in
27), the factor of safety (FS) decreased significantly but the dam was still
Table 9; it provides an effective hydraulic seal, and excellent resistance
considered safe, as shown in Fig. 19(c); however, when using the
to physical or chemical breakdowns. To remedy the Upper-pond, the
reduced strength of the soil, the FS fell below 1 (Fig. 19(d)).
road section where the crack appeared was dismantled, and the topsoil
This analysis illustrates that the dam was in a critical condition. The
and grasses of the downstream slope were cut off. The downstream slope
strength of the soil during the investigation was somewhere between the
surface was then adjusted by topsoil excavation for laying the GCL with
peak and reduced strength. When the displacement in the X–Y direction

15
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Table 9 and the drainage pipe were then covered by grit and small particles of
Specification of the geosynthetic clay liner (GCL) used. stones and sand, which were then overlaid by a sand drainage with a 0.3-
Properties Test Method Test Typical cm thickness. These arrangements would release any likely pore water
Frequency Value pressure on the downstream side. The top surface was then overlaid by a
Bentonite Properties 30-cm thick soil, and grass planting was employed to prevent erosion. In
Bentonite Swell Index American Society for 1/50,000 ≥24 ml/2 g this way, a total area of 174,750 m2 of GCL was placed over the
Testing and Materials kg downstream slope of the Upper-pond, and the rainwater was prohibited
(ASTM) D 5890 from coming into contact with the claystone of the 3C zone.
Bentonite Fluid Loss ASTM D 5891 1/50,000 ≤18 ml
kg
Finished GCL Properties 8. Post-remediation measurements
Bentonite Mass ASTM D 5993 1/4000 m2 ≥2800 g/sq.
m
The settlement points S1–S12 were monitored from Jan. 10th, 2014
Grab Strength ASTM D 6768 1/4000 m2 ≥7000 N/m
Permeability ASTM D 5887 1/10,000 ≤1.0*10− 8 to Sept. 25th, 2019, to identify if there was a continual movement of the
m2 cm/s downstream slope of the Upper-pond. Settlement points S14–S17, were
Finished Geomembrane Properties removed during the remediation process. Fig. 23 shows the movement of
Textured High-Density ASTM D 5199 1/20,000 ≥0.45 mm the dam as recorded during the monitored period.
Polyethylene (HDPE) m2
Thickness
There was a significant reduction in the settlement and horizontal
Sheet Density ASTM D 1505 1/20,000 ≥0.92 g/ displacement of the downstream slope of the dam. Before remediation, it
m2 cm3 took only approximately a year on average for nearly all the settlement
points to move by 0.1 m. However, after laying the GCL on the down­
stream slope of the dam, all of the settlement points were comparatively
held intact. Settlement point S8 exhibited the largest movement, which
was approximately 0.1 m in both the horizontal and vertical directions
at the end of 4 years, whereas the rest of the settlement points demon­
strated values of less than 0.1.

9. Conclusion

This study presents a case study of the use of a GCL with a textured
HDPE layer as a remedial measure for downstream slope movement in
the Lam Ta Khong pumped hydroelectric storage in Thailand. The use of
GCL with textured HDPE geomembranes has been very rarely used to
stop claystone degradation in dams. This study provides insights on how
its use can prevent the failure of such large structures. Extensive site
investigations, laboratory tests, instrumentation, and numerical
modeling were carefully performed to understand the causes of the
movement, and to arrive at the appropriate remedial approach.
One of the settlement points (S15) installed on the dam’s body
indicated a vertical settlement of 93.5 cm within 8 years. Similarly, most
of the settlement points continued to demonstrate movement in the
Fig. 21. Laying of geosynthetic clay liner (GCL) in downstream slope of the horizontal and vertical directions at an average rate of 0.1 m/year.
Upper-pond. Through field investigation, it was found that the Ruzi grass
implanted on the top surface of the dam was responsible for water
the textured HDPE (Fig. 21). An anchorage trench was also dug to infiltration, which decreased the suction beneath it. The permeability
improve the GCL stability, and it was ensured that the GCL had a proper measured at the top surface and at 1.5 m in the field was found to be in
overlap between them while covering the entire downstream slope. the range of 10− 5 to 10− 4 m/s, which is very high, considering that the
A typical section of the remediated downstream slope is shown in 3C zone consists of claystone. Furthermore, the resistivity test results
Fig. 22. A horizontal polyethylene drainage pipe with a diameter of 150 showed the presence of water pockets at several locations.
mm was installed as appropriate for a proper drainage system. The GCL Disturbed and undisturbed soil samples were collected from the

Fig. 22. Cross section - a part of the remediated downstream slope of Upper-pond.

16
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

Fig. 23. Dam movement behavior after remediation (a) Plot of settlement along with time and rainfall (b) Plot of displacement along with time and rainfall.

dam’s body and foundation to investigate their properties. A jar slake laboratory tests) suggested that the claystone degraded upon contact
test and slake durability test indicated that the claystone used in the 3C with rainwater, causing movement in the dam’s downstream slope.
zone had very low durability. Similarly, triaxial UU tests on the undis­ A transient seepage analysis was performed using Seep/W. The
turbed samples indicated a decrease in the undrained shear strength by continuous rainfall data of October 9–12, 2007 (0.002, 0.02, 0.08, and
approximately 65–80% when the degree of saturation increased from 0.04 m/day, respectively) were chosen to study the effect of rain infil­
approximately 80%–100%. These field observations (and several trating the dam. A water bulb was formed in the 3C zone, similar to the

17
S. Soralump et al. Geotextiles and Geomembranes xxx (xxxx) xxx

observation from the resistivity test. Furthermore, a period covering the Horpibulsuk, S., Yangsukkaseam, N., Chinkulkijniwat, A., Du, Y.J., 2011. Compressibility
and permeability of Bangkok clay compared with kaolinite and bentonite. Appl. Clay
maximum intensity of rainfall (August 26–29, 2007) was chosen for its
Sci. 52, 150–159.
considerable effect on the stability analysis of the slope. When using the Hunter, G., Fell, R., 2003. Rockfill modulus and settlement of concrete face rockfill dams.
reduced strength of the dam material along with the highest rainfall of J. Geotech. Geoenviron. Eng. 129, 909–917.
August 27 (0.125 m/day), the FS fell below 1, indicating that the dam Ishimori, H., Katsumi, T., 2012. Temperature effects on the swelling capacity and barrier
performance of geosynthetic clay liners permeated with sodium chloride solutions.
was in a critical state. Geotext. Geomembranes 33, 25–33.
In total, 174,750 m2 of GCL was laid on the downstream slope to Jackson, R.D., 1972. On the calculation of hydraulic conductivity. Soil Sci. Soc. Am. J.
avoid the infiltration of rainfall reaching the claystone material. The 36, 380–382.
Jotisankasa, A., Sirirattanachat, T., 2017. Effects of grass roots on soil-water retention
GCL with textured HDPE provides resistance against slipping along the curve and permeability function. Can. Geotech. J. 54, 1612–1622.
slope, and provides the best leak protection. Post-measurements indi­ Jotisankasa, A., Vathananukij, H., 2008. Investigation of soil moisture characteristics of
cated a considerable decrease in the movement of the downstream slope. landslide-prone slopes in Thailand. In: International Conference on Management of
Landslide Hazard in the Asia-Pacific Region 11th-15th November, p. 12.
All of the settlement points except for S8 demonstrated a settlement of Jotisankasa, A., Vathananukij, H., Coop, M., 2009. Soil-water retention curves of some
less than 0.1 m in more than 4 years of monitoring, whereas S8 silty soils and their relations to fabrics. In: Proc. The 4th Asia Pacific Conference on
demonstrated a movement of slightly over 0.1. Unsaturated Soils, pp. 263–268.
Kayadelen, C., Tekinsoy, M., Taşkıran, T., 2007. Influence of matric suction on shear
strength behavior of a residual clayey soil. Environ. Geol. 53, 891.
Acknowledgement Kul, T.Ö., Ören, A.H., 2018. Liquid limit based assessment of geosynthetic clay liners
subject to hydration and hydraulic conductivity testings. Geotext. Geomembranes
46, 436–447.
The author would like to thank the Electricity Generating Authority
Mairaing, W., Jotisankasa, A., Soralump, S., 2012. Some applications of unsaturated soil
of Thailand (EGAT) for providing research funding (grant no. mechanics in Thailand: an appropriate technology approach. Geotechn. Eng. J.
4120005794) and for the continuous monitoring of the dam movement, SEAGS & AGSSEA 43, 1–11.
which proved to be valuable data in the research. In addition, the Manataki, M., Papadopoulos, N., Sarris, A., Agapiou, A., Themistocleous, K.,
Hadjimitsis, D.G., 2014. Monitoring of water leakages from pipes through
Geotechnical Innovation Lab (GIL) of Kasetsart University deserves geophysical imaging methods. In: Integrated Use of Space, Geophysical and
special gratitude for their support in conducting the various laboratory Hyperspectral Technologies Intended for Monitoring Water Leakages in Water
tests. Supply Networks. InTech.
Mazzieri, F., Pasqualini, E., 2000. Permeability of damaged geosynthetic clay liners.
Geosynth. Int. 7, 101–118.
References Monfared, M., Sulem, J., Delage, P., Mohajerani, M., 2011. A laboratory investigation on
thermal properties of the Opalinus claystone. Rock Mech. Rock Eng. 44, 735.
AbdelRazek, A., Rowe, R.K., 2019. Interface transmissivity of conventional and Ng, C.W.W., Leung, A.K., 2012. In-situ and laboratory investigations of stress-dependent
multicomponent GCLs for three permeants. Geotext. Geomembranes 47, 60–74. permeability function and SDSWCC from an unsaturated soil slope. Geotech. Eng.
Abu-Shariah, M.I., 2009. Determination of cave geometry by using a geoelectrical 43, 26–39.
resistivity inverse model. Eng. Geol. 105, 239–244. Ng, C.W.W., Shi, Q., 1998. A numerical investigation of the stability of unsaturated soil
Al-Fares, W., 2014. Application of electrical resistivity tomography technique for slopes subjected to transient seepage. Comput. Geotech. 22, 1–28.
characterizing leakage problem in Abu Baara earth dam, Syria. Int. J. Geophys. Panthi, K., Soralump, S., 2020. Estimation of failure surface of Pa bon dam from site
2014. investigation. Jordan J. Civil Eng. 14.
Arya, L.M., Paris, J.F., 1981. A physicoempirical model to predict the soil moisture Petrov, R.J., Rowe, R.K., Quigley, R.M., 1997. Selected factors influencing GCL hydraulic
characteristic from particle-size distribution and bulk density data 1. Soil Sci. Soc. conductivity. J. Geotech. Geoenviron. Eng. 123, 683–695.
Am. J. 45, 1023–1030. Rehman, S., Al-Hadhrami, L.M., Alam, M.M., 2015. Pumped hydro energy storage
ASTM D 2850-15, 2015. Standard Test Method for Unconsolidated-Undrained Triaxial system: a technological review. Renew. Sustain. Energy Rev. 44, 586–598.
Compression Test on Cohesive Soils. American Society for Testing and Materials, Rodvinij, P., Wattanachai, P., 2017. Influence of claystone deterioration on shear
West Conshohocken, PA. strength of backfill. Int. J. Geomate 13, 54–59.
ASTM, D 4644, 2008. Standard Test Method for Slake Durability of Shales and Similar Sadisun, I.A., Shimada, H., Ichinose, M., Matsui, K., 2005. Study on the physical
Weak Rocks. ASTM International, West Conshohocken, PA, USA. disintegration characteristics of Subang claystone subjected to a modified slaking
ASTM D 4767-11, 2011. Standard Test Method for Consolidated Undrained Triaxial index test. Geotech. Geol. Eng. 23, 199–218.
Compression Test for Cohesive Soils. ASTM International, West Conshohocken, PA. Shan, H.-Y., 1991. Results of laboratory tests on a geotextile/bentonite liner material.
Bhowmik, R., Shahu, J., Datta, M., 2018. Failure analysis of a geomembrane lined Geosynthetics 91 2, 517–535.
reservoir embankment. Geotext. Geomembranes 46, 52–65. Shrestha, A., Jotisankasa, A., Chaiprakaikeow, S., Pramusandi, S., Soralump, S.,
Boardman, B.T., Daniel, D.E., 1996. Hydraulic conductivity of desiccated geosynthetic Nishimura, S., 2019. Determining shrinkage cracks based on the small-strain shear
clay liners. J. Geotech. Eng. 122, 204–208. modulus–suction relationship. Geosciences 9, 362.
Bouazza, A., 2002. Geosynthetic clay liners. Geotext. Geomembranes 20, 3–17. Sjödahl, P., Dahlin, T., Johansson, S., 2010. Using the resistivity method for leakage
Chae, B.-G., Lee, J.-H., Park, H.-J., Choi, J., 2015. A method for predicting the factor of detection in a blind test at the Røssvatn embankment dam test facility in Norway.
safety of an infinite slope based on the depth ratio of the wetting front induced by Bull. Eng. Geol. Environ. 69, 643–658.
rainfall infiltration. Nat. Hazards Earth Syst. Sci. 15. Skempton, A., 1954. The pore-pressure coefficients A and B. Geotechnique 4, 143–147.
Chiarelli, A.-S., Shao, J.-F., Hoteit, N., 2003. Modeling of elastoplastic damage behavior Šumanovac, F., Weisser, M., 2001. Evaluation of resistivity and seismic methods for
of a claystone. Int. J. Plast. 19, 23–45. hydrogeological mapping in karst terrains. J. Appl. Geophys. 47, 13–28.
Davis, G.H., Reynolds, S.J., Kluth, C.F., 2011. Structural Geology of Rocks and Regions. Tavenas, F., Leblond, P., Jean, P., Leroueil, S., 1983. The permeability of natural soft
John Wiley & Sons. clays. Part I: methods of laboratory measurement. Can. Geotech. J. 20, 629–644.
Deschamps, R., Hynes, C., Wigh, R., 1999. Analysis and stabilization of failing earth dam Teerawattanasuk, C., Maneecharoen, J., Bergado, D.T., Voottipruex, P., Le, G.L., 2014.
founded on claystone. J. Perform. Constr. Facil. 13, 143–151. Root strength measurements of vetiver and ruzi grasses. Lowland Technol. Int. 16,
Franklin, J., Chandra, R., 1972. The slake-durability test. In: International Journal of 71–80.
Rock Mechanics and Mining Sciences & Geomechanics Abstracts. Elsevier, Tian, H., Ziegler, M., Kempka, T., 2014. Physical and mechanical behavior of claystone
pp. 325–328. exposed to temperatures up to 1000 C. Int. J. Rock Mech. Min. Sci. 70, 144–153.
Fredlund, D.G., Rahardjo, H., Rahardjo, H., 1993. Soil Mechanics for Unsaturated Soils. USBR, 1990. Earth Manual, Part 2, third ed. Materials Engineering Branch Research and
John Wiley & Sons. 614 Laboratory Services Division Denver.
Fredlund, D., Xing, A., Fredlund, M., Barbour, S., 1996. The relationship of the USBR, 2011. Design Standard No.13, Embankment Dams, Chapter 4: Static Stability
unsaturated soil shear strength to the soil-water characteristic curve. Can. Geotech. Analysis. Denver, Colorado.
J. 33, 440–448. USBR, 2012. Reclamation Managing Water in the West. Denver, Colorado.
Gilson-Beck, A., 2019. Controlling leakage through installed geomembranes using Wilkinson, W., Shipley, E., 1972. Vertical and Horizontal Laboratory Permeability
electrical leak location. Geotext. Geomembranes 47, 697–710. Measurements in Clay Soils, Developments in Soil Science. Elsevier, pp. 285–298.
Head, K.H., 1994. Manual of soil laboratory testing. In: Permeability, Shear Strength and Zhang, F., Xie, S., Hu, D., Shao, J.-F., Gatmiri, B., 2012. Effect of water content and
Compressibility Tests, vol. 2. Pentech Press Ltd. structural anisotropy on mechanical property of claystone. Appl. Clay Sci. 69, 79–86.
Hopkins, T.C., Gilpin, B.C., 1981. Identification of Kentucky Shales.

18