You are on page 1of 14

Indian Geotech J (April 2020) 50(2):238–251

https://doi.org/10.1007/s40098-019-00396-0

S.I. : DEVELOPMENTS IN TRANSPORTATION GEOTECHNICS

Geogrid Reinforced Soil Walls with Marginal Backfills Subjected


to Rainfall: Numerical Study
Midhula Jayanandan1 • B. V. S. Viswanadham1

Received: 4 July 2019 / Accepted: 5 November 2019 / Published online: 12 December 2019
Ó Indian Geotechnical Society 2019

Abstract Altering natural landforms without impacting Keywords Geosynthetics  Reinforced soil walls 
stability has become quintessential due to unprecedented Backfill  Rainfall  Pore water pressure  Drainage
infrastructure development. Geosynthetic reinforced soil
wall (GRSW) as a retaining structure is a congenial option
List of Symbols
addressing land scarcity and right of way challenges.
c0 Cohesion of soil (Pa)
However, with dearth of good-quality backfill soil and
f Percentage open area (dimensionless)
prohibitive transportation costs, engineers are being forced
h Height of the slope (m)
to use native low-permeable marginal soils for constructing
I Average hourly rainfall intensity (m/s)
geosynthetic reinforced soil walls. Adverse climatic chan-
mw Coefficient of volumetric water change with
ges leading to erratic rainfall have triggered numerous
respect to change in negative pore pressure
failures in such soils due to their inefficiency to dissipate
(Pa- 1)
excess pore water pressure. The objective of the present
k Coefficient of permeability (m/s)
paper is to evaluate applicability of different backfill soils
ks Saturated hydraulic conductivity of soil (m/s)
for GRSW. Seepage and stability analyses were conducted
q Applied rainfall flux (m/s)
to appraise the behaviour of a geogrid reinforced soil wall
Tult Ultimate tensile load (kN/m)
with rigid facing of height 9.6 m having 6° batter during
u/ch Normalised pore water pressure
rainfall. The effects of backfill soil types with fines content
(dimensionless)
of 0%, 20% and 40% on seepage behaviour and stability
umid/ch Normalised pore pressure measured at the base,
were assimilated under moderate rainfall. It was ascer-
at a point vertically below the midlength of the
tained that failure occurred faster when the backfill soil
slope crest (measured at a distance 4.8 m from
with 20% fines was considered. For backfill soil with 40%
the upstream end of the wall) (dimensionless)
fines, a significant portion of rainfall was observed to
utoe/ch Normalised pore pressure measured on the base
simply disappear as runoff. In the section with sand
at the toe (measured at a distance 9.6 m from
backfill, failure did not occur as water drained off faster
the upstream end of the wall) (dimensionless)
than rainfall infiltration, restraining mounting of water
Dupwp/ch Variation in normalised pore water pressure
table.
from upstream end to downstream end of the
wall is measured along the base of the GRE
wall from t = 0 days to t = 40 days
& Midhula Jayanandan (dimensionless)
midhula.jayanand@iitb.ac.in cdmax Maximum dry unit weight of soil (kN/m3)
B. V. S. Viswanadham eult Ultimate tensile strain (dimensionless)
viswam@civil.iitb.ac.in hs Saturated volumetric water content of soil
(m3/m3)
1
Department of Civil Engineering, Indian Institute of hw Volumetric water content of soil (m3/m3)
Technology Bombay, Powai, Mumbai 400076, India

123
Indian Geotech J (April 2020) 50(2):238–251 239

sb Bond skin friction (Pa) less than 15% of fines passing 0.075-mm sieve and with
/0 Angle of internal friction of soil (°) plasticity index less than or equal to six. A recent report by
/b Angle indicating the rate of increase in shear National Concrete Masonry Association [8] permits the
strength relative to the matric suction (°) usage of soils having less than 35% fines, whereas
sf Shear strength of soil (Pa) MORTH (Ministry of Road Transport and Highway, Govt.
r Total stress on the failure plane (Pa) of India) restricts the fines to 15%. Casagrande has defined
ua Pore air pressure (Pa) that the value of 10-6 m/s for saturated coefficient of
uw Pore water pressure (Pa) permeability is the boundary of poor drainage material
qw Density of the soil domain (in g/m3) (Holtz and Kovacs [9]).
Numerical investigation carried out by Yoo [10, 11] on a
reinforced earth wall with fine-grained backfill soil having
segmental concrete block facing, studied the effect of long-
duration rainfall on the performance of reinforced earth
Introduction wall. It was observed that the factor of safety significantly
reduced which can be attributed to the reduction of shear
Geosynthetic reinforced walls owing to their innate flexi- strength due to the reduction in matric suction. Kim and
bility, expediency in construction, cost-effectiveness and Borden [12] performed transient seepage analyses and
tolerance to deformations without structural distress, have stress deformation analyses on a 5.4-m-high MSE wall
gained immense popularity. Stringent guidelines have been with marginal backfills under surface infiltration and
imposed on the selection of backfill soil. Good-quality observed that the soil yielded with the reduced shear
granular soil is the preferred backfill for construction of strength and large wall-face deflection and vertical settle-
geosynthetic reinforced soil walls. But the scarcity and ment. As reported by Yoo and Jang [13], geotextile rein-
exorbitant cost of good-quality soils have compelled forced soil wall with a poorly grade non-plastic sand
engineers to use locally available low-permeable soil as backfill subjected to cycles of wetting and drying, devel-
backfills. The usage of such low-quality low-permeable oped maximum pore water pressure of 4 kPa only
soils (with high fines) directly results in effective use of throughout the rainfall period. The pore water pressure
waste materials and reduction in the overall cost of con- developing within the low-permeable soils should be
struction by approximately 30%. However, the typical effectively dissipated either using external drainage means
characteristics of such marginal soils such as low peak or incorporating internal drainage measures within the soil.
shear resistance, loss of shear strength on wetting, poor Groundwater drainage must be intercepted by a back drain
drainage characteristics and volume changes and creep (chimney drains) between the reinforced and retained soil
potential are aggravated when exposed to rainfall condi- zones given as a continuation of base drain which should
tions. In fact, a steady increase in rainfall-triggered natural be provided throughout the reinforced zone length. The use
disasters has been witnessed recently. Heavy rains even for of sand and geocomposite chimney drains has been proved
a few days or incessantly happening for several days can be effective in arresting the development of positive pore
detrimental as they cause saturation of overburden, induce water pressures in the reinforced zone of low-permeable
positive pore water pressure which reduces matric suction marginal soils to a great extent (Bui Van et al. [14],
and result in low shear strength and subsequent failure Viswanadham et al. [15], Mamaghanian et al. [16, 17]).
(Koerner and Koerner [1], Christopher et al. [2], Christo- Mamaghanian et al. [16] noted that the use of a geotextile
pher and Stuglis [3]). After evaluating 320 failures of as a chimney drain could not prevent the failure of the wall
geosynthetic reinforced mechanically stabilised earth on inducing seepage; however, it can manage to delay the
walls, Koerner and Koerner [4] concluded that 73% used time of failure initiation. Another viable solution is
silty and clayey backfill soils and 94% were geogrid rein- encapsulating geogrid layers with sand cushion (Abdi et al.
forced. The selection of backfill material from a wide [18], Raisinghani and Viswanadham [19], Abdi and Arjo-
variety of reinforced backfill materials available today is mand [20], Balakrishnan and Viswanadham [21]). But the
contingent on factors such as strength and deformation difficulty in placing sand and marginal fill soils in alternate
requirements, availability, interaction with the reinforce- layers and the possible intermixing of the two different soil
ment, grain size, plasticity, permeability, shear strength and layers make this method unfeasible. The use of geocom-
compaction characteristics of the backfill material. posite layers (sometimes referred as hybrid geosynthetics)
There are several guidelines for selection of the backfill as internal drainage measures which offers dual function-
soil. While Koerner et al. [5] recommend a granular ality of reinforcement and drainage has also proved to be a
backfill material which is devoid of fines, AASHTO [6] major breakthrough in enhancing behaviour of soil walls
and FHWA (Berg et al. [7]) allow backfill materials with and slopes with such low-permeable marginal backfills

123
240 Indian Geotech J (April 2020) 50(2):238–251

(Raisinghani and Viswanadham [22], Bhattacharjee and soils, failures are caused primarily by long-duration–
Viswanadham [23, 24], Razeghi et al. [25]). moderate-intensity rainfall events (Wieczorek [28]).
Viswanadham et al. [15] conducted seepage and stabil- In the present paper, it is intended to appraise the
ity analysis to determine the performance of geogrid rein- behaviour of a GRSW with different backfill soil condi-
forced soil walls with panel facing using marginal backfill tions subjected to rainfall by performing numerical analy-
(a blend of sand and kaolin in 4:1 proportion by dry ses. Three soil types with fines content of 0%, 20% and
weight) with and without chimney sand drain subjected to 40% were used as backfill materials. The finite-element-
seepage. On comparison, a soil wall reinforced with stiff based software SEEP/W and the limit equilibrium-based
geogrid layers was found to perform effectively at the onset software SLOPE/W were employed, respectively, for
of seepage than the wall using low stiffness geogrid and seepage and stability analyses of this wall section (GeoS-
without any chimney drain. The normalised pore water tudio [29]). A 9.6-m-high wall having 6° batter with ver-
pressure values were considerably reduced at the midheight tical provided with rigid precast concrete panel facing was
and wall toe after provision of chimney sand drain resulting subjected to rainfall. A panel facing of thickness 185 mm
in enhanced wall performance under the effect of seepage was adopted according to Viswanadham et al. [15]. The
forces. Also, the effect of increase in thickness of sand rainfall of intensity 10 mm/h was simulated numerically
layer in chimney drain was studied and it was noticed that for 25-day duration, and seepage and stability analyses
there was an improvement in the discharge values and were performed at the onset of rainfall, during rainfall and
factor of safety against piping near the toe region. Yang up to a total duration of 40 days. The numerical results
et al. [26] numerically assessed the combined effect of were compared for the three backfill conditions to com-
different backfill–reinforcement–drainage conditions on prehend the hydraulic response as well as static global
the infiltration response and global stability of a reinforced stability of the GRSW with an emphasis on the pore water
soil slope (RSS). The investigation deduced that RSS with pressure developed within the soil and factor of safety of
sand backfill independent of reinforcement and drainage reinforced wall.
provision was more competent under heavy rainfall con-
ditions owing to its high permeability and rapid pore water
dissipation. On the other hand, for geotextile RSS with silty Details of the Reinforced Wall Section
and clayey soils, sandwiching the reinforcement within a
thin layer of sand was proven to be effective in improving A wall with 9.6 m height having 6° batter with vertical
interaction properties and shear strength preventing built- provided with rigid precast concrete panel facing of
up of pore water pressures and capillary barrier effect. In thickness 185 mm subjected to rainfall was considered for
high-permeability soils, high-intensity–small-duration numerical analyses. The wall was founded on 2.4-m-thick
rainfall is considered to cause failures with antecedent foundation soil with levelling pad at an embedment depth
rainfall having little influence on landslide occurrence of 1 m from wall base. The initial water table was assumed
(Corominas [27]). On the contrary, in low-permeability at a height of 2.4 m from the base. Three different backfill
soils with fines content 0% (Soil A), 20% (Soil B) and 40%

Fig. 1 Schematic diagram of


the model RE wall used in the
present analysis

123
Indian Geotech J (April 2020) 50(2):238–251 241

analysis was carried out using SLOPE/W software. A


rainfall with intensity 10 mm/h was applied on the wall
crest for a period of 25 days. The change in pore water
pressures and loss in matric suction with progress of
infiltration were recorded. The analysis was continued for a
total duration of 40 days to monitor the recovery of matric
suction in the soil. The corresponding change in factor of
safety values against rainfall progression was also noted
throughout the analysis period. The procedure was repeated
for all the three backfill soil conditions, and the results
were compared.

Validation of the Present Study

The present study is a numeral analysis which appraises the


seepage and stability performance of a geogrid reinforced
soil wall with different backfill conditions subjected to
Fig. 2 Particle size distribution curves of different backfill soils used rainfall. However, as the numerical model is not validated
in the present study
by any experimental or real data, the present study was
validated with one of the published literature (Rahardjo
(Soil C) were considered in the present analysis. A geogrid
et al. [31]) by performing a finite-element analysis using
with ultimate tensile strength of 40 kN/m and ultimate
GeoStudio [29]. Rahardjo et al. [31] conducted a para-
tensile strain of 25% in machine direction (as per ASTM
metric study to examine the effect of groundwater
D6637-10 [30]) was used as the reinforcement in the
table position in affecting slope stability. A 15-m-high, 27°
analysis. The percentage hopen area fiof the geogrid used
unreinforced slope was considered with three different
was 97.23% (f defined as ðal aþb
l at
l Þ ðat þbt Þ
where al, at are grid
groundwater table positions corresponding to the wettest,
opening sizes and bl, bt are rib widths and l, t are longi-
average and driest periods in Singapore. In Singapore,
tudinal and transverse directions, respectively). Six layers
majority of slope failures occurred within a shallow depth.
of geogrid reinforcements were attached to the wall facing
Typical soil properties of a residual soil obtained from the
with a uniform centre-to-centre spacing of 1.6 m. The
Bukit Timah (BT) Granite zone in Singapore were incor-
length of reinforcements was taken as 8 m (* 0.85H) to
porated into the numerical analyses. To simulate a maxi-
prevent chances of any rupture failure. The geogrid and its
mum total amount of precipitation of 533.2 mm reported in
aforementioned properties used for the analysis were
a day, a rainfall intensity of 22 mm/h was applied contin-
selected according to Balakrishnan and Viswanadham [21].
uously on the ground surface for 24 h. The analysis con-
Figure 1 depicts the schematic diagram of the geogrid
tinued for another 24 h to evaluate the behaviour of BT
reinforced wall used in the present analysis.
slope in terms of matric suction recovery. The configura-
In the present study, Goa sand with zero per cent fines is
tion of model slope adopted for validation is presented in
used as Soil A. The soil types Soil B and Soil C are
Fig. 3.
obtained by blending locally available Powai soil and
Two-dimensional transient seepage analyses were per-
uniformly graded sand passing through 4.25-mm sieve in
formed in this study using the finite-element software
different proportions to achieve 20% and 40% fines con-
SEEP/W (GeoStudio [29]). SWCC and permeability
tent, respectively. The grain size distribution curves of the
function adopted for the analyses are considered according
three different backfill soils are depicted in Fig. 2. The
to Rahardjo et al. [31]. Non-ponding boundary condition
properties of the three backfill soils are summarised in
was selected to prevent any excessive accumulation of
Table 1.
rainfall on the slope surface. The flux boundary q equal to
the desired rainfall intensity and duration was applied to
the surface of the slope. Further, global stability analysis
Numerical Modelling of Geogrid Reinforced Soil
was carried out using SLOPE/W and the factor of safety for
Wall Subjected to Rainfall
different time intervals was plotted corresponding to the
three different ground water table positions. The results of
The geogrid reinforced wall section was numerically
the present study were found to be in close agreement with
modelled using GeoStudio [29] suite, wherein seepage
the results of Rahardjo et al. [31] as illustrated in Fig. 4.
analysis was carried out using SEEP/W and slope stability

123
242 Indian Geotech J (April 2020) 50(2):238–251

Table 1 Properties of soils used in the present analysis


Properties Units Soil A Soil B Soil C

Fine content % 0 20 40
Specific gravity, Gs –a 2.654 2.697 2.737
Particle size distribution
Gravel (G) % 0 0 0
Sand (S) % 100 79.44 59.965
Silt (M) % 0 19.96 38.035
Clay (C) % 0 0.6 2
Effective particle size, D10 mm 0.101 0.045 0.015
Average particle size, D50 mm 0.191 0.165 0.085
a
USCS classification – SP SM SM
Plasticity characteristics (Atterberg limits)
Liquid limit (%) % –a 27.7 35.02
Plastic limit (%) % –a Non-plastic Non-plastic
Plasticity index (%) % –a –a –a
Compaction characteristics (standard proctor)
Optimum moisture content, OMC % –a 15.25 16.76
3
Maximum dry density, cd kN/m –a 17.65 19.87
Permeability characteristics
c
k m/s 1.54 9 10-4 d
2.5 9 10-7 d
6 9 10-9
Shear strength parametersb
Cohesion, c kPa 0 0.2 2
Angle of internal friction, / degrees 32 29.63 25
a
Not relevant/ reported
b
From direct shear test on sample moist compacted at cdmax and OMC (saturated for 24 h at a normal stress 100 kPa)
c
Constant head permeability test (at 40% relative density)
d
Falling head permeability test conducted on sample moist compacted at cdmax and OMC

The observed deviation is attributed to the difference in employed. The governing equation representing the tran-
mesh properties and tolerances in time step increments. sient seepage flow through an unsaturated, heterogeneous
Factor of safety of the BT slope reached a minimum and anisotropic soil in SEEP/W is:
value when rainfall stopped regardless of the position of      
o oh o oh o oh oh
groundwater table. The changes in factor of safety during k þ k þ k ¼ m2w qw g ð1Þ
ox ox oy oy oz oz ot
rainfall were not affected significantly by the groundwater
table near the ground surface due to the relatively small where mw is the coefficient of volumetric water change
changes in matric suction during rainfall. Factor of safety with respect to change in negative pore pressure (in Pa-1)
for the BT slope in the driest period 2.16 and at the average and is the slope of soil water characteristic curve (SWCC),
condition 2.04 decreased drastically and reached a mini- h is the hydraulic head (in m), k is the permeability of soil
mum value at 1.7. After the rainfall ceased, the factor of (in m/s), qw is the density of the soil domain (in g/m3) and
safety for the BT slope at the driest period and at the g is gravitational acceleration (in m/s2).
average condition increased rapidly at the same recovery The rainfall in SEEP/W was simulated by providing a
rate. The factor of safety for the BT slope in the wettest unit flux boundary condition with a positive value on a
period decreased gradually from 1.7 to 1.45, and then, it potential seepage face. Non-ponding condition was acti-
increased slowly after the rainfall stopped. vated to prevent accumulation of excess water along the
soil surface. The initial position of water table was pro-
Seepage Analysis of the RE Wall Section Subjected vided along the wall base. No flow condition (Q = 0 m3/s)
to Rainfall Using SEEP/W was provided at the base and along the side boundaries. A
toe drain was simulated within the sand layer towards the
For conducting transient seepage analysis, finite-element left of the base layer (pore pressure head, H = 0 m) to
analysis-based SEEP/W software of GeoStudio [29] was drain out all the excess water along with runoff. Since the

123
Indian Geotech J (April 2020) 50(2):238–251 243

Fig. 3 Model slope adopted for validation of Rahardjo et al. [31]

Fig. 4 Comparison of variation


of FOS with time for a rainfall
of intensity 22 mm/h with
different GWT positions of the
present study and published
literature work

wall facing is constituted by discrete precast concrete medium in SEEP/W software by assigning it a very low
panels, potential seepage face was provided along the wall value of hydraulic conductivity. The value of hydraulic
facing to account for water seeping out through gaps conductivity was selected such that it encompasses the
between the facing panels. A stiff panel facing in a rein- effect of gaps between panels Mamaghanian et al. [16, 17].
forced soil wall has the function of reducing the wall Geogrid inclusions are generally considered to exhibit only
deformation significantly and retaining the backfill soil and reinforcement function. Further, geogrids rarely showcase
hence minimising the transferred loads to reinforcement drainage function capabilities and accordingly they are not
layers. The facing is modelled as an impervious soil modelled in seepage analysis. This geogrid behaviour was

123
244 Indian Geotech J (April 2020) 50(2):238–251

previously established by Bhattacherjee and Viswanadham within the infiltration zone. The seepage analysis results for
[23]. Saturated/unsaturated constitutive material properties geogrid reinforced walls with various backfill soils are
were adopted for both backfill soil and foundation base depicted using pore pressure contours for 25 days (end of
layer. Foundation soil was assumed to have similar prop- rainfall) and 40 days (end of analysis duration) as given in
erties as that of backfill soil with both compacted at OMC Figs. 6a–c and 7a–c, respectively.
and cdmax. The hydraulic conductivity function (HCF) was It was identified that the ratio of saturated hydraulic
obtained by considering the saturated soil properties in conductivity of the soil (ks) to the rainfall flux (I) has a
conjunction with the SWCC based on Iryo and Rowe [32] predominant role in determining the infiltration rate of
database. The SWCC and hydraulic conductivity function rainfall into the soil medium. When ks/I [[ 1, the rainfall
(HCF) curves for the three backfill soils thus obtained are flux is far lower than the saturated hydraulic conductivity
illustrated in Fig. 5a and b. of the soil, resulting in almost all the rain being percolated
The global element size for the model mesh was taken into the soil. This results in very high infiltration rate, and
as 0.25 m. According to Tsaparas et al. [33], the pore water water drains rapidly through the soil. On the other hand,
pressure change during a rainfall is predominantly confined when ks/I \\ 1, the rainfall flux is much higher than the
to a depth of 2 m from the ground surface, when the water permeation capacity of the soil. As a result, the
groundwater table is at a significant depth from the surface. infiltration is observed to be very low and majority of the
Hence, a finer mesh size of 0.05 m was considered up to a rainwater disappears as runoff. Hence, it can be inferred
depth of 2 m from the surface to get more precise results that soils with very low permeability when subjected to
heavy intensity rainfall usually fail due to surface runoff
causing erosion.
The pore pressure profiles clearly indicate that Soil A
with 0% fines (ks/I * 55) has the least pore water pressure
(no measurable positive pore water pressures observed
throughout the wall section) at the end of rainfall period.
Soil A had persistent matric suction of 5 kPa and negligible
rise in phreatic level at the end of rainfall which indicates
that the wall section is safe from failure. This is attributed
to its high permeability and high water drainage capacity.
In case of Soil B with 20% fines (ks/I = 0.9), the pore water
pressure developed substantially with progress of rainfall,
resulting in mounting of water table. The infiltration rate
was adequate for the soil to develop positive pore water
pressures, but its inefficiency to dissipate this built-up
pressure was ascribed to its low hydraulic conductivity.
Only a small part of the applied rainfall entered the wall
section with backfill Soil C with 40% fines (ks/I = 0.002),
which resulted in less pore water pressure build up by the
end of 25 days of continuous rainfall. However, it can be
deduced that a prolonged duration rainfall can result in
high pore pressures in Soil C. Once the rainfall was stop-
ped, apart from Soil C, both Soil A and Soil B demon-
strated recovery of matric suction. The recovery rate was
very fast in Soil A, whereas Soil B exhibited a slow rate of
matric suction recovery. Soil C, owing to its extremely low
permeability, was inefficient in the revival of matric
suction.
The normalised pore water pressure u/ch (defined as a
ratio of pore water pressure to the product of bulk unit
weight of the soil and height of the wall) was measured at
two points, at the toe and midpoint, along the wall base of
the RE wall. The dimensionless umid/ch values are nor-
Fig. 5 Hydraulic properties of soils used in seepage analysis of the malised pore water pressure values measured at the base, at
present study. a SWCC. b Hydraulic conductivity function a point vertically below the midlength of the slope crest

123
Indian Geotech J (April 2020) 50(2):238–251 245

Fig. 6 Pore pressure contours


of geogrid reinforced soil wall
subjected to rainfall intensity of
10 mm/hour after 25 days for
different backfills (pore water
pressure in kPa). a Soil A with
0% fines. b Soil B with 20%
fines. c Soil C with 40% fines

123
246 Indian Geotech J (April 2020) 50(2):238–251

Fig. 7 Pore pressure contours


of geogrid reinforced soil wall
subjected to rainfall intensity of
10 mm/hour after 40 days for
different backfills (pore water
pressure in kPa). a Soil A with
0% fines. b Soil B with 20%
fines. c Soil C with 40% fines

123
Indian Geotech J (April 2020) 50(2):238–251 247

which is measured at a distance 4.8 m from the upstream It was also observed that in all three backfill cases, u/ch
end of the wall. Further utoe/ch values are the normalised values increased with the distance from the toe/toe drain,
pore pressure values measured on the base at the toe (9.6 m i.e. umid/ch values are greater than utoe/ch values. This is
from the upstream end of the wall). The variation of nor- corroborated using Fig. 9a and b, which shows the varia-
malised pore water pressures with time elapsed during tion of u/ch along the base of the RE wall with various
rainfall for different wall sections with various backfills is backfills. It can also be clearly stated that for all the backfill
illustrated in Fig. 8a and b. As indicated in Fig. 8a and b, soils, the pore water pressure values decreased as the
the maximum u/ch is attained at the end of 25 days (rainfall rainfall stopped. In case of Soil A due to its high draining
stopped) which then decreased for the rest of the analysis capability, the rise in phreatic surface above the base layer
period in both Soil A and Soil B. At the end of rainfall, it was inconsiderable and hence the pore water pressure built-
can be observed that Soil A does not develop any positive up was also considered trivial. Also, it is evident that Soil B
pore water pressure even after 25 days of continuous displayed highest u/ch values under the considered rainfall
rainfall indicating very high drainage capacity of such throughout the duration of analysis. Even after the end of
good-quality soils. The final phreatic surface was formed rainfall, the previously infiltrated rainwater gradually per-
well within the base layer and did not rise to the wall colated through Soil C resulting in continuous increase in
section. Soil B exhibited maximum umid/ch and utoe/ch u/ch values because of the low draining capability and low
values of 0.269 and 0.096, respectively (Fig. 8a, b permeability of such soils. Similarly in the present study,
considered). Dupwp/ch which is the variation in normalised pore water

Fig. 8 Variation of normalised pore water pressures with time Fig. 9 Variation of normalised pore water pressures along the base of
elapsed during rainfall of intensity 10 mm/hour for wall sections with geogrid reinforced wall section for different backfills a After 25 days.
different backfill soils. a At toe. b At midpoint b After 40 days

123
248 Indian Geotech J (April 2020) 50(2):238–251

pressure from upstream end to downstream end is mea-


sured along the base of the geogrid reinforced soil wall.
Dupwp/ch is an index which shows the relative change in
pore water pressure from the initial ground water
table condition (at t = 0 days) to the final water table con-
dition (at t = 40 days). Figure 10 illustrates the variation of
pore water pressure (Dupwp/ch) with the horizontal distance
along the base of the wall. Since Dupwp/ch values are
positive irrespective of the backfill conditions, it can be
clearly ascertained that there is an increase in pore water
pressures with time due to the rainfall. It can also be
concluded that Dupwp/ch values decreased from upstream
end to downstream end (toe) of the wall. Soil A, as
expected, exhibited the least Dupwp/ch values as there was
no significant mounting of water table and hence very less
increase in pore water pressures. Soil B showed the max-
imum Dupwp/ch values compared to Soil A and Soil C.
Figure 11 represents the u/ch values measured along the
height of RE wall sections with different backfill condi-
tions for a period of 25 days and 40 days at a distance
4.8 m from the right boundary indicating midcross section
of the wall. Major portion of RE wall section modelled
with Soil C backfills exhibited matric suction contrary to
the belief that the soil with such high fines content (i.e.
40%) is more susceptible to develop positive pore pres-
sures. This is again attributed to the fact that Soil C being
less permeable did not allow the rainwater to percolate into
the wall section; therefore, phreatic surface rise was very
slow resulting in positive pore pressures developing only to
a height 1.6 m and 2.2 m from the wall base after 25 days
and 40 days, respectively (Fig. 11a, b). As revealed in
Fig. 11a and b, the entire wall cross section with 0% fines
developed no positive pore water pressures throughout the
40-day analysis period. The geogrid reinforced soil wall Fig. 11 Variation of normalised pore water pressures along the
height of geogrid reinforced wall section for different backfill soil
cross section with 20% fines backfill showed positive pore
types. a After 25 days. b After 40 days

pressures throughout the wall section by the end of rainfall.


However, with the stoppage of rainfall, the soil recovered
matric suction up to a depth of 2.6 m from the crest.

Stability Analysis of RE Wall Section Subjected


to Rainfall Using SLOPE/W

The static stability analysis of the reinforced wall section


was carried out using limit equilibrium-based software
SLOPE/W. The stability analysis of the soil wall under
rainfall condition was performed by incorporating the
phreatic surfaces captured from SEEP/W transient seepage
analyses during various stages of rainfall into SLOPE/W.
The variation of negative pore pressures was incorporated
in the slope stability analysis, by utilising the modified
Fig. 10 Variation of Dupwp/ch with the horizontal distance measured
from upstream end to downstream end of the GRE wall

123
Indian Geotech J (April 2020) 50(2):238–251 249

Mohr–Coulomb failure criterion presented in Eq. (2), as of safety in all the wall sections till the end of rainfall. RE
proposed by Fredlund et al. [34]: wall section with Soil A showcased superior performance
sf ¼ c0 þ ðr  ua Þ tan /0 þ ðua  uw Þ tan /b ð2Þ with a very small reduction in FOS during the rainfall and
showed fast recovery rate poststoppage of rainfall. This, as
where sf is the shear strength (in Pa); r is the total stress on expected, is due to its high-permeability characteristics and
the failure plane (in Pa); (r - ua) is the net normal stress drainage capability. The above observation is supported by
on the failure plane (in Pa); (ua - uw) is the matric suction Yang et al. [26, 35]. Their study revealed that the loss of
(in Pa); and /b is the angle indicating the rate of increase in matric suction induced by rainfall infiltration influenced the
shear strength relative to the matric suction (in °). stability of the reinforced sand slope minimally. In case of
Bishop’s simplified method of analysis was chosen for Soil B, there was an abrupt reduction in FOS after the 19th
conducting the stability analyses. The entry and exit day which is assumed to be caused by the complete loss of
method was used for identifying the potential slip surface. matric suction throughout the wall section. The lowest FOS
For achieving convergence, thirty slices were considered in was obtained after 25 days of rainfall which gradually
each iteration for a maximum of 2000 iterations with a increased after the rainfall stopped with failure occurring
tolerance of 0.01 in factor of safety values. The geogrid on the 22th day. It is evident that the recovery of matric
reinforcements which exhibit tensile strength have been suction after rainfall resulting in increase in FOS values is
modelled as fabric-type reinforcement and the ultimate very gradual in low-permeability soils. It was observed that
fabric capacity (or ultimate tensile load) of 40 kN/m which FOS values continuously reduced in Soil C backfill because
was selected in accordance with Balakrishnan and of incompetency to drain off infiltrated water. In the pre-
Viswanadham [21]. The ultimate tensile load adopted was sent study, since the stability analysis is carried out purely
considered to be distributed evenly along the length of the considering the rainfall infiltration into the wall section
reinforcement with a bond resistance (sb) of 10.07 kPa. (evaporation losses or runoff is neglected), it can be con-
The facing was modelled in SLOPE/W equivalent to the cluded that a rainfall with prolonged duration and moderate
soil with an unit weight of concrete (M30 grade). The intensity is required to trigger failure in soil walls with
material properties for the soil considered in the stability backfills having very high fines content (Soil C).
analysis are summarised in Table 1. The stability analysis
was carried out for a total duration of 40 days with a
rainfall of intensity 10 mm/h applied for 25 days. The Conclusions
variation in factor of safety with time during rainfall for
geogrid reinforced soil wall with different backfill soils The behaviour of a geogrid reinforced soil wall having
was obtained and is depicted in Fig. 12. rigid facing with marginal backfill subjected to a rainfall of
From Fig. 12, it can be inferred that irrespective of 10 mm/hour was evaluated based on a series of stability
backfill soil type, factor of safety values decreased with the analyses coupled with transient seepage analyses. The
progress of rainfall. There was a steady reduction in factor rainfall was simulated for 25 days, and seepage and sta-
bility analyses were performed at the onset of rainfall,
during rainfall and up to 40-day completion of rainfall. The
applicability of different backfill soils varying in fines
content (0%, 20% and 40%) was inspected numerically.
Based on the analysis and interpretation of the results, the
following conclusions can be drawn:
(1) Soil A with 0% fines exhibited a matric suction of
5 kPa, and umid/ch values were found to be negative
even at the end 25 days of rainfall. The matric suction
increased to 20 kPa by the end of 40 days after the
rainfall was stopped. Also, the rise in phreatic level
was observed to be insignificant. Similarly, the entire
cross section of wall with Soil A developed no
significant positive pore water pressure built-up
throughout the 40-day analysis period. This is
attributed to the extremely high-permeable nature of
Soil A which allows rapid pore water dissipation and
Fig. 12 Variation of factor of safety with time elapsed during rainfall rapid percolation of rainfall. Thus, it can be noted that
of intensity 10 mm/h

123
250 Indian Geotech J (April 2020) 50(2):238–251

the wall section with backfill soil containing no fines at the upstream end. The increase in pore water
content considered in the present study is safe from pressure was negligible in Soil A with 0% fines and
failure. hence had the least Dupwp/ch values. In comparison
(2) There was a considerable increase in the normalised with Soil A and Soil C, Soil B which showed
pore water pressure values in Soil B with 20% fines maximum increase in pore water pressures also
when compared to Soil A and Soil C. A maximum observed maximum Dupwp/ch values by the end of
normalised pore water pressure value (u/ch) of 0.33 40 days.
was observed by the end of 25 days of rainfall at the
In the present study, deformation analysis of the various
farthest point from toe along the base of the wall
soil wall cross sections was not carried out to study the
which reduced to 0.28 with stopping of rainwater
effect of different backfill soils and rainfall on a geogrid
entry into the soil. The phreatic surface had risen to
reinforced soil wall. Further studies are also warranted to
the crest of the wall by the end of rainfall which
understand the seepage and stability behaviour of the
gradually lowered by the restoration of matric suction
reinforced walls subjected to different rainfall intensities,
after the rainfall was stopped.
durations and with different reinforcement types and lay-
(3) In Soil C with 40% fines, the mounting of water
out. Also, numerical studies shall be further extended by
table was found to be very slow, but at the same time
adopting different reinforcement layouts to achieve the
the phreatic surface continued to rise with no recovery
same initial factor of safety for walls with the three dif-
in matric suction even after the end of the rainfall.
ferent backfill soils considered, thereby providing an
Major portion of rainwater is disappearing as runoff,
insight to the effectiveness of spacing, length and tensile
and the rainwater infiltrating into the wall section is
load of reinforcement. Further, it is also warranted to val-
very less and hence phreatic surface had risen to a
idate the above by performing appropriate physical model
height of 1.6 m only.
studies on geogrid reinforced soil walls subjected to
(4) With application of rainfall, the FOS values decreased
rainfall.
steadily in all backfill soils. However, for the
considered rainfall characteristics, Soil A demon- Acknowledgements The authors would like to acknowledge editor
strated high competence than marginal soils (Soil B and reviewers for their constructive comments in improving the
and Soil C) with high fines content as the factor of quality of the manuscript.
safety values never reduced below 1.38 throughout
the stability analysis. There was a gradual recovery in
the FOS values in Soil B with 20% fines after the References
rainfall; however, in Soil C with 40% fines content,
1. Koerner RM, Koerner GR (2011) The importance of drainage
FOS values continued to decrease even after the control for geosynthetic reinforced mechanically stabilized earth
rainfall was stopped. This indicates soil walls with wall. J GeoEng 6(1):3–13
very high fines content fail mostly due to soil erosion 2. Christopher BR, Zornberg JG, Mitchell JK (1998) Design guid-
or due to long-duration rainfall. Hence, it can be ance for reinforced soil structures with marginal soil backfills.
Proc Sixth Int Conf Geosynth 2(6):797–804
inferred that granular soil with no fines content is the 3. Christopher BR, Stuglis RS (2005) Low permeable backfill soils
ideal backfill soil for RE wall section. However, it in geosynthetic reinforced soil wall: state of the practice in North
appears that a higher quantity of fines could be America, In: Proceedings of North American geosynthetics
allowed in the reinforced fill, provided properties of conference (NAGS), GRI-19, cooperative conference proceed-
ings, Las Vegas, Nevada, USA, pp 16
materials are well defined and controls are well 4. Koerner RM, Koerner GR (2018) An extended data base and
established. recommendations regarding 320 failed geosynthetic reinforced
(5) The values of Dupwp/ch indicate the change in mechanically stabilized earth (MSE) walls. Geotext Geomembr
normalised pore water pressures of final ground water 46(10):904–912
5. Koerner J, Soong TY, Koerner RM (1998) Earth retaining wall
table condition from the initial ground water table con- costs in the USA. Geosynthetic Institute, Folsom
dition. Irrespective of the backfill soil conditions, 6. AASHTO (2009) Standard specifications for highway bridges.
Dupwp/ch values were observed to be positive which American Association of State Highway and Transportation
substantiated the built-up of pore water pressures in Officials, Washington D.C
7. Berg RR, Christopher BR, Samtani NC (2009) Design and con-
soil with time due to the rainwater infiltration. The struction of mechanically stabilized earth walls and reinforced
variation of Dupwp/ch values with the horizontal soil slopes. FHWANHI-10-024 Report. National Highway Insti-
distance measured along the base of the wall showed tute, Washington, DC, vol 1, pp 306
a decreasing trend from upstream end to downstream 8. National Concrete Masonry Association (1997) Design manual
for segmental retaining walls, Second edition, Second Printing, J
end. This indicated that the change in pore water Collin (Editor), Herndon, Virginia
pressure from t = 0 days to t = 40 days is maximum

123
Indian Geotech J (April 2020) 50(2):238–251 251

9. Holtz RD, Kovacs WD (1981) An introduction to geotechnical 22. Raisinghani DV, Viswanadham BVS (2011) Centrifuge model
engineering. Prentice-Hall, Englewood Cliffs. ISBN 978-0-13- study on low permeable slope reinforced by hybrid geosynthetics.
484394-0 Geotext Geomembr 29(6):567–580
10. Yoo C, Jung HY (2006) Case history of geosynthetic reinforced 23. Bhattacherjee D, Viswanadham BVS (2015) Numerical studies
segmental retaining wall failure. J Geotech Geoenviron Eng on the performance of hybrid-geosynthetic-reinforced soil slopes
ASCE 132(12):1538–1548 subjected to rainfall. Geosynth Intl 22(6):411–427
11. Yoo C (2011) Effect of rainfall on performance of reinforced 24. Bhattacherjee D, Viswanadham BVS (2018) Effect of geocom-
earth wall. In: Han J, Alzamora DE (eds) Geo-frontiers congress posite layers on slope stability under rainfall condition. Indian
2011. ASCE, GSP 211, vol 3, pp 1852–1861. Geotech J 48(2):316–326
https://doi.org/10.1061/41165(397)189. ISBN: 9780784411650 25. Razeghi HR, Viswanadham BVS, Mamaghanian J (2019) Cen-
12. Kim WS, Borden RH (2013) Numerical Simulation of MSE Wall trifuge and numerical model studies on the behavior of geogrid
Behaviour Induced by Surface-Water Infiltration. J Geotech reinforced soil walls with marginal backfills with and without
Geoenviron Eng ASCE 139(12):2110–2124 geocomposite layers. Geotext Geomembr.
13. Yoo C, Jang DW (2013) Geosynthetic reinforced soil wall per- https://doi.org/10.1016/j.geotexmem.2019.10347. (in Press)
formance under heavy rainfall. In: Delage P, Desrues J, Frank R, 26. Yang KH, Thuo JN, Huynh VDA, Nguyen TS, Portelinha FHM
Puech A, Schlosser F (eds) Proceedings of the 18th international (2018) Numerical evaluation of reinforced slopes with various
conference on soil mechanics and geotechnical engineering: backfill reinforcement- drainage systems subject to rainfall
challenges and innovations in geotechnics, Paris, vol 1, infiltration. Comput Geotech 96:25–39.
pp 2131–2134 https://doi.org/10.1016/j.compgeo.2017.10.012
14. Van Bui D, Chinkulkijniwat A, Horpibulsuk S, Yubonchit S, 27. Corominas J (2001) Landslides and climate. Keynote Lect 8th
Limrat I, Arulrajah A, Jothityangkoon C (2017) Steady flow in Intl Symp Landslides. 4:1–33
mechanically stabilised earth walls using marginal soils with 28. Wieczorek GF (1987) Effect of rainfall intensity and duration on
geocomposites. Geosynth Intl 24(6):590–606 debris flows in central Santa Cruz Mountains, California. Geolog
15. Viswanadham BVS, Razeghi HR, Mamaghanian J, Manikumar Soc Am Rev Eng Geol 7:93–104
CHSG (2017) Centrifuge model study on geogrid reinforced soil 29. GeoStudio (2012) SEEP/W and SLOPE/W. Ver. 7.15. User’s
walls with marginal backfills with and without chimney sand Guide, GeoSlope International Ltd, Calgary, Canada
drain. Geotext Geomembr 45(5):430–446 30. ASTM D6637 (2010) Standard test method for determining ten-
16. Mamaghanian J, Razeghi HR, Viswanadham BVS, Manikumar sile properties of geogrids by the single or multi-rib tensile
CHSG (2018) Behaviour of geogrid reinforced soil walls with method. American Society for Testing and Materials, West
marginal backfills with and without chimney drain in a Conshohocken
geotechnical centrifuge. In: McNamara A, Divall S, Goodey R, 31. Rahardjo H, Nio AS, Leong EC, Song NY (2010) Effects of
Taylor N, Stallebrass S, Panchal J (Eds). Proceedings of the 9th groundwater table position and soil properties on stability of
international conference on physical modelling in geotechnics. slope during rainfall. J Geotech Geoenviron Eng ASCE
17–20 July 2018, University of London. Taylor & Francis Group 136(11):1555–1564
(Pubs.), London, ISBN 978-1-138-34422-8. 1:1211–1216 32. Iryo T, Rowe RK (2005) Hydraulic behaviour of soil–geocom-
17. Mamaghanian J, Viswanadham BVS, Razeghi HR (2019) Cen- posite layers in slopes. Geosynth Intl 12(3):145–155
trifuge model studies on geocomposite reinforced soil walls 33. Tsaparas I, Rahardjo H, Toll DG, Leong EC (2002) Controlling
subjected to seepage. Geosynth Intl 26(4):371–387 parameters for rainfall-induced landslides. Comput Geotech
18. Abdi MR, Sadrnejad A, Arjomand MA (2009) Strength 29(1):1–27
enhancement of clay by encapsulating geogrids in thin layers of 34. Fredlund DG, Morgenstern NR, Widger RA (1978) Shear
sand. Geotext Geomembr 27(6):447–455 strength of unsaturated soils. Can Geotech J 15(3):313–321
19. Raisinghani DV, Viswanadham BVS (2010) Evaluation of per- 35. Yang KH, Thuo JN, Chen JW, Liu CN (2018) Failure investi-
meability characteristics of a geosynthetic-reinforced soil through gation of a geosynthetic-reinforced soil slope subjected to rain-
laboratory tests. Geotext Geomembr 28(6):579–588 fall. Geosynth Int 26(1):1–24.
20. Abdi MR, Arjomand MA (2011) Pullout tests conducted on clay https://doi.org/10.1680/jgein.18.00035
reinforced with geogrid encapsulated in thin layers of sand.
Geotext Geomembr 29(6):588–595 Publisher’s Note Springer Nature remains neutral with regard to
21. Balakrishnan S, Viswanadham BVS (2019) Centrifuge model jurisdictional claims in published maps and institutional affiliations.
studies on the performance of soil walls reinforced with sand-
cushioned geogrid layers. Geotext Geomembr.
https://doi.org/10.1016/j.geotexmem.2019.103496) (in Press)

123

You might also like