Document Actions

IJEGE-11_BS-Kuo-et-alii

background image
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
303
DOI: 10.4408/IJEGE.2011-03.B-035
EXPERIMENT ON THE STABILITY OF GRANULAR SOIL SLOPES
BY RAINFALL INFILTRATION
R
onG
-H
eR
CHEN
(*)
, k
wo
J
ane
KUO
(**)
& C
HiH
-m
inG
CHANG
(***)
(*)
Professor, Department of Civil Engineering, National Taiwan University, Email: rongherchen@ntu.edu.tw, 886-2-23629851
(**)
PhD student at National Taiwan University, E-mail:clay.kuo@msa.hinet.net
(**)
Former master’s student at National Taiwan University
INTRODUCTION
Many slope failures were caused by infiltration
of water that induced increase in pore-water pressure
and seepage force in soil (l
umP
, 1975; b
Rand
, 1982;
i
veRson
& m
aJoR
, 1986; C
Hen
et alii, 1999; k
im
et
alii, 2004; C
Hen
et alii, 2004; C
Hen
et alii, 2009). The
stability of an unsaturated soil slope can be dramati-
cally altered as the soil becomes gradually saturated,
even resulting in a slope failure. The collapsed soil
may turn into a flow if there is plenty of water in the
soil along with favourable geological condition.
Using model tests to study rainfall induced slope
failures have been conducted by a few research. w
anG
& s
assa
(2003) performed tests on silica sand of dif-
ferent initial dry densities and fines contents to inves-
tigate the effects of grain size on pore-water pressure
generation and the failure behaviour of rainfall-in-
duced landslides. m
oRiwaki
et alii (2004) conducted
a full-scale experiment to clarify the failure process of
landslide triggered by rainfall. The soil was loose sand
with a relative density of 35 % and an initial water
content of 8 %. The flume was subjected to rainfall
of an intensity of 100 mm/hr. The sliding process of
the slope was recorded and evaluated by measuring
the displacement of slope surface, and the piezometric
level and pore-water pressure within the soil. o
Rense
et alii (2004) conducted model sandy slopes to inves-
tigate slope failure by water percolation from side
upslope, with a 80 cm constant height, and by artifi-
cial rain falling on the slope with rainfall intensities
ABSTRACT
Slope failure is commonly caused by rainfall in-
filtration due to increase in pore-water pressure within
the slope. In order to understand the failure mechanism
of granular soil slopes, a model slope subjected to rain-
fall infiltration was developed and employed to per-
form experiment on sandy soils. Different geological
conditions, fines contents of sand, and rainfall intensi-
ties were considered as variables. Infiltration of rainfall
was simulated by an overland flow infiltrating uniform-
ly into a platform on the top of the slope. During the
experiment, the variations in pore water pressure and
volumetric water content in the soil were measured.
The characteristics of the failure mechanism and the
responses of pore pressure and water content in four
model slopes were observed and discussed. According
to the observation, the failure in permeable sand was a
sliding mode. Initial failure was noticed as a piping oc-
curred at the toe of slope, and it then gradually propa-
gated upward as a retrogressive failure. However, the
failure in less permeable silty sand was initiated by
erosion at shallow depth, the rill later expanded and
turned into either a flow or a complex mode of flow
and slide, depending on the rainfall intensity applied.
Moreover, sandy soil had a marked increase in pore-
water pressure when approaching failure; but this phe-
nomenon was less obvious in silty sand.
K
ey
words
: model test, slope, stability, granular soil, infiltration
background image
R.-H. CHEN, k. J. kUO & C.-M. CHANG
304
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
MODEL TEST
Two kinds of samples were tested; sample A was
composed of 100% sand and sample B was a mixture
of 90 % sand and 10 % silt. According to the Unified
Soil Classification System, samples A and B are classi-
fied as SP and SM, respectively. Therefore, they are ex-
pected to behave differently under rainfall infiltration.
SIMILARITY ANALYSIS
In the studies on slope failure, model test is one
of the efficient methods. This is because model test
has the merits that it can be performed in well control-
led conditions, e.g., boundary and loading conditions;
more importantly, it can also simulate how a slope
failure is initiated and also the whole process of fail-
ure can be observed in the laboratory.
In a model test, the model and the prototype
should be related through satisfying the geometric
similarity, the kinematic similarity, and the dynamic
similarity so that the feasibility and reliability of test
results are ensured. For instance, the scaling factor of
length between the prototype and the model can be
chosen as λ (= L
p
/L
m
) and the materials used in the
prototype-slope and the model-slope can have the
same unit weight. From similarity analysis (R
oCHa
,
1957; R
osCoe
, 1968), the relationship between the
physical quantities of prototype- and model- slopes
are obtained as shown in Table 1.
EXPERIMENT SETUP
The experimental setup is shown in Figures 1a
and 1b for different geological conditions. The device
includes a sand tank supported by a steel frame; the
of 42~262 mm/hr. The material was silty sand with
9% of fines content. The soil samples were obtained
from Omigawa landslide site in Japan. The maximum
rainfall intensity was 42 mm/hr in that area.
A set of laboratory-scale experiment for study-
ing rainfall induced slope failure was performed by
t
oHaRi
et alii (2007). The soils were sand and re-
sidual granite soil of various relative densities and
initial water contents. They employed two types of
water supply in the study; one was a rainfall simulator
designed to produce an effective rainfall intensity of
approximately 100 mm/hr, and the other was a water
tank which could vary the rising rate of water level
within the slope. H
uanG
et alii (2008) also conduct-
ed model tests on sandy soil slopes to look into the
rainfall-induced failure process in the slope. In addi-
tion, they observed and reported the characteristics of
cumulative solid discharge versus time.
C
Hen
et alii (2004) investigated the triggering
mechanism for a hazardous mudflow caused by a ty-
phoon which delivered approximately 300 mm/day
of precipitation. The destructive 50-m long mudflow
rushed down a hill without any forewarning from a
man-made platform of 30 m x 80 m right at the top of
the mudflow. This funnel-shaped mudflow of a volume
of 2,000 m
3
damaged three houses below and resulted
in five deaths. According to the test results of the sam-
ples taken from the erosion gully, most of soils belong
to silty and clayey sand. The bedding, which appears
in the erosion gully, has a dip angle ranging from 30°
to 60° toward the toe of the slope. Besides, the cause of
the disaster was found likely due to the malfunction of
the drainage system around the platform because more
overland flow on the platform could infiltrate into the
slope. This disaster was deeply concerned for the dis-
aster area occurred in an urbanized territory.
For a better understanding about relevant slopes
as the case study described above, model test was em-
ployed to study the failure mechanism of soil slopes
composed of sand and silty sand, respectively. Differ-
ent geological conditions, fines contents of sand and
rainfall intensities were considered as variables to
test the stabilities of model slopes. The infiltration of
rainfall was simulated by an overland flow infiltrating
uniformly from the top of the slope. The character-
istics of the failure mechanism and the responses of
pore pressure and water content for four model slopes
were observed and are discussed below.
Tab. 1 - Dimensional analysis of the physical quantities in
model test
background image
EXPERIMENT ON THE STABILITY OF GRANULAR SOIL SLOPES BY RAINFALL INFILTRATION
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
305
accuracy is ± 0.5 %, and the allowable pressure
is 20 kPa. The piezometers are installed either
as shown in Figure 1a for a homogeneous soil
slope, or as shown in Figure 1b for a slope with
an impervious dip stratum of 10° underneath.
Moreover, three CCD cameras are aimed
at different angles, normal to the side of the
tank, the lower slope and the deposition area,
to record the process of the test. These cameras
have a resolution of 640×480 pixels and the
maximum frame rate is 30 fps.
TEST MATERIAL
The sand used for testing is uniform, sub-
angular, and contains 99.8% silica. It is classified
as poorly-graded sand (SP) as per the Unified Soil
Classification System, with a coefficient of uniform-
ity of C
u
= 1.566, and a coefficient of curvature of
C
c
= 0.946. The physical properties of the sand are:
maximum unit weight γ
d,max
= 16.9 kN/m
3
, minimum
unit weight γ
d,min
= 14.6 kN/m
3
, median diameter d
50
= 0.2 mm, and effective grain size d
10
= 0.13 mm.
The friction angles of the sand are obtained
by triaxial consolidated-undrained test at rela-
tive densities of 55 % and confining pressures of
50, 100, and 200 kPa, respectively. The friction
angles of samples A and B are 36.6° and 35.1°,
respectively. Furthermore, samples A and B have hy-
draulic conductivities of 3×10
-4
m/s and 4×10
-5
m/s,
respectively. The saturated volumetric water content
of soil is obtained from the soil-water characteristics
curve using pressure plate test; the values are 41.9%
and 47.1% for samples A and B, respectively.
TEST PROGRAM
The test program is shown in Table 2. The param-
eters considered include geological condition, silt content
in sand, and rainfall intensity. Two geological conditions
are the homogeneous slope (Figure 1a) and the slope with
an impervious dip stratum (Figure 1b). The latter was to
wall of the device is made of 10-mm thick reinforced
glass. The profile of the slope has two sections: a 40°
upper slope and a 20° lower slope, with a platform at
the top of the slope. This geometric configuration was
arranged in order to study the failure mechanism of
concave steep slopes. It is also hoped that the result
from this experiment may provide a reference for fu-
ture remedial work of similar slopes. The flow chute
is 0.3 m wide, while that of the deposition area is 0.6
m wide and 0.4 m long (Figure 1c).
A rainfall device is set directly above the plat-
form. Water can also be supplied from either the
bottom or the rear of the tank to simulate the change
of ground water level.
Several moisture sensors, Delta-T SM200, are
used to measure the volumetric water content of
the soil. The accuracy of the sensor is ±3 %, with
an operating range between 0-60°. Moisture sen-
sors W1~W7 are installed on one side of the tank
as shown in Figures 1a and 1b. Pore-water pressures
are recorded by piezometers, Kyowa PGM-G. The
Tab. 2 - Test program
Fig. 1 - Experimental setup
background image
R.-H. CHEN, k. J. kUO & C.-M. CHANG
306
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
reached the top of the slope, the valve was closed.
The slope was then left 24 hours for soaking in
order to reach as much saturation as possible.
5. Then the valve was opened again to drain out wa-
ter until the water table was lowered to the eleva-
tion at the toe of the slope. It usually took about
two hours for water drainage, e.g., at a rate of 0.02
m
3
/min, so that soil settlement was insignificant.
6. The monitoring system was turned on; it inclu-
ded three CCDs, pore-water pressure transdu-
cers and piezometers.
7. The infiltration test was started by turning on
the rainfall device to generate rainfall at a spe-
cified intensity.
8. During the test, the volumetric water content and
pore-water pressure were recorded at frequencies
of 5/s and 1/s, respectively.
9. When there was no significant soil movement ob-
served, the test was stopped.
Note that soaking soil before the test was not only
to simulate the effect of precedent rainfall on site but
also to expedite the test.
TEST RESULTS AND DISCUSSION
Before discussing test results, the definition of
symbols for several important points during a test are
given, i.e., t
p
= the time to occur initial failure when pip-
ing was observed, t
f
= the time when an obvious mass
movement was initiated, and t
e
= the time at the end of
significant mass movement. The failure mode and im-
portant time points of each test are tabulated in Table 3.
FAILURE PROCESS
Based on the observation, the failure mode of
sample A belongs to a sliding mode; a typical exam-
ple is displayed in Figure 2 and the whole process is
presented in Figure 3. However, sample B failed more
like a flow. The reason could be attributed to their
hydraulic conductivities and fines content. Sample A
was composed mainly of relatively permeable sand
and the seepage with high velocity could be expected
research the effect of impervious dip stratum. In addition,
due to relatively low hydraulic conductivity of sample B,
only one homogeneous slope was studied for sandy soil.
The rainfall intensity chosen was based not only
on the hydraulic conductivity of the soil, but also on
the site. The higher value of rainfall intensity, 286 mm/
hr or 8 x10
-5
m/s, was calculated from the empirical
equation for Taipei City (C
Hen
et alii, 2004); the lower
value, 140 mm/hr or 3.8 x10
-5
m/s, was chosen to be
equal to the hydraulic conductivity of sample B. Based
on our previous studies, these values exceeded the
threshold value of 80 mm/hr for slope failure to occur.
Furthermore, because of the pervious nature of sample
A, it was tested only under high rainfall intensity.
Strictly speaking, according to Table 1, the rain-
fall intensity and the hydraulic conductivity of the soil
should be modified when conducting a model test.
However, the hydraulic conductivity of the soil is so
small that the modified value seems to have little ef-
fect on the test results. Moreover, the scaling factors
for these two quantities are the same. Accordingly, the
rainfall intensity was not modified either.
TEST PROCEDURE
The test procedures are briefly described as fol-
lowing.
1. For constructing the homogeneous slope, a layer
of sand of 0.1 m thick was placed at the bottom of
the tank at first. When constructing the non-ho-
mogeneous slope, a 10° stratum made of plywood
was placed in the beginning. After that, the gap
between this stratum and the wall of the tank was
sealed with silicone sealant for waterproof.
2. The slope was formed with the help of a shaping
plate to retain soil in place. The soil was then plu-
viated by layers to obtain a uniform sand deposit,
each being 0.02 m thick, from the bottom to the
top. The plate was lifted every time a layer of soil
had been placed. This procedure was repeated till
a specified full height was reached.
3. After the model slope had been constructed, a bar
working as a spillway gate was placed at the frin-
ge of the platform. This bar was 0.3 m long and
0.018m high. The purpose of using it was to retain
surface water to simulate an overland flow.
4. The water valve at the bottom of the tank was
opened to allow water to flow slowly into the tank
at a rate of 0.01 m
3
/min. After the water table had
Tab. 3 - Important time points of each test
background image
EXPERIMENT ON THE STABILITY OF GRANULAR SOIL SLOPES BY RAINFALL INFILTRATION
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
307
to induce piping at the toe of a slope. On the other
hand, owing to water infiltrating more slowly in a silty
sand, fines particles were washed out by water from
the interstices at shallow depths and became fluidized.
More details about each test are described as follows.
• Test T1 - As aforementioned, the wetting of slope
started from the toe where piping occurred. This
small zone of failure subsequently extended quic-
kly upward to about the interface between lower
and upper slopes, forming a shallow sliding. Later
on the failure surface continued to deepen and ex-
tended beyond the crest to produce a deep failure.
The extension of the collapsed area was about one
half of the platform.
• Test T2 - The wetting of slope also began from
the toe, in the same way as T1. However, due to
an impervious dip stratum, initiation of piping and
developing a shallow sliding were faster than T1
test. Even so the collapsed area on the platform
was also about one half of the original area.
• Test T3 - In this test, the wetting of slope began
from the interface between upper and lower slopes,
but there was no obvious piping occurred in this
silty sand as previously described. Instead, a crack
on the platform was noticed first, which then deve-
loped quickly from a rill into a small gully. Figure
4 shows how the rill was eroded and turned into a
gully. The final dimension of the eroded hole near
the crest was about 0.13 m wide and 0.14 m deep.
• Test T4 - The applied rainfall intensity in this
test was only one half of that in T3. Although
initial failure starting from the toe of the slope
was more obvious as compared to T3, it was less
significant than T2. The failure mode tended to
be a complex mode of flow and slide. An eroded
gully was also developed on the surface of upper
slope and extended upward to the crest. Never-
theless, this gully was only 0.067 m wide, about
one half of that in T3.
Fig. 2 - Failure surfaces for T2 at different times
Fig. 3 - The retrogressive failure surfaces in a sandy soil slope (T2)
background image
R.-H. CHEN, k. J. kUO & C.-M. CHANG
308
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
viously, the phenomena were owing to the less perme-
able characteristic of sample B as well as the rainfall
of higher intensity applied to T3.
For W5 located at the interface of lower and up-
per slopes, all curves in Figures 5b had high initial
values because the passage for flow became narrower
adjacent to this area. In Figure 5b1, the curve of T2
was higher than that of T1, resulting from the narrow
passage on the impervious stratum. In Figure 5b2, it is
also not surprising to see the curve of T3 subjected to
higher rainfall intensity was above that of T4. Another
characteristic of the curves at this interface was none
of them declined after t
f
.
PORE-wATER PRESSURE
The variation of pore-water pressure in soil was re-
corded using six piezometers shown in Figures 1a and b.
Piezometers P1 and P4 were below the right side of the
platform, P2 and P5 were below the crest, and P3 and P6
were close to the toe of the slope. Their coordinates are
given in Figure 1b. For convenience to compare, total
head instead of piezometric head was adopted, taking
the bottom of the model as the datum. The total head
versus time are displayed in Figures 6a and b.
As illustrated in Figure 6a, the total head for sample
A generally increased markedly from t
p
to t
f
. The period
between t
p
and t
f
was shorter for T2 than that of T1, in-
dicating failure occurred sooner in T2. After t
f
, the total
head either kept steady or even increased due to water
was continuously supplied into the soil. On the other
hand, the increase in total head from t
p
to t
f
for sample
From Table 3 it can be seen that both t
p
and t
f
for
T2 occurred sooner than T1; apparently, it was result-
ed from the dip stratum. With regard to samples A and
B, both t
p
and t
f
of test T3 took place slower than test
T2. The reason could be explained by the difference
in the fines content and hydraulic conductivity of the
sample. As to the effect of rainfall intensity on T3 and
T4, it seems to be only more significant on t
p
. Finally,
the time to end mass movement, t
e
, expectedly took
longer for sample B than sample A.
VOLUMETRIC wATER CONTENT
The volumetric water content of soil was measured
by seven sensors, and their coordinates with respect to
the toe of the slope are given in Figure 1. The initial vol-
umetric water content and the specific water contents
corresponding to t
p
, t
f
, and t
e
are denoted as θ
0
, θ
p
, θ
f
,
θ
e
, respectively. These values are tabulated in Table 4.
Figure 5 shows the time histories of the volumet-
ric water contents of W2 and W5; W2 was at the crest
and W5 was at the interface of upper and lower slopes.
It can be seen from Figure 5a that the curves of two
samples display different trend, even though the initial
values of the two samples were about the same. Since
sample A is relatively permeable, hence the curves in
Figure 5a1 did not reach full saturation. In addition to
that, the curve of T2 declined significantly after tf, as
opposed to that of T1. On the contrary, in Figure 5a2,
the curves of sample B attained high values, and the
curve of T3 could even reach complete saturation. Ob-
Fig. 4 - Top view of the failure process for T3
Tab. 4 - Summary of volumetric moisture contents
background image
EXPERIMENT ON THE STABILITY OF GRANULAR SOIL SLOPES BY RAINFALL INFILTRATION
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
309
exception, shown in Figure 6b2, is that a peak value
was noticed at t
f
. Apparently, this was related to the lo-
cation of P5 from where the cross-sectional area gradu-
ally became smaller down to the toe of the slope, and
B was less obvious. As can be seen from Figure 6b, t
f
for both T3 and T4 were only slightly different; this has
been indicated in Table 3. Furthermore, the total head
after t
f
essentially did not have a marked change. One
Fig. 5 - Volumetric water content versus time obtained from moisture sensors w2 and w5
Fig. 6 - Total head versus time for tests on different samples
background image
R.-H. CHEN, k. J. kUO & C.-M. CHANG
310
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
therefore the pore-water pressure was easily to build
up. Comparing respectively the dotted curves of P4, P5,
and P6 in Figures 6a and b, the peak total heads in sam-
ple A were 0.05 ~ 0.08 m higher than those in sample
B, as a result of more permeable nature of sample A.
In addition, the profiles of total head at different
times for the four tests are illustrated in Figure 7. For ho-
mogeneous slope, T1, the profiles are concave, whereas
those for T2-T4 are convex. Besides, the initial profiles
of T2-T4 were approximately parallel to the impervious
stratum. From this result, it is easy to know that the up-
per part of the slope was not saturated during the test.
But most of the lower slope, especially near the toe, was
saturated when slope failure occurred. As illustrated in
Figure 7a2, the final total head at P6 approximated the
final surface. This observation suggests a mode of retro-
gressive failure started from the toe of the slope.
The final surfaces of T1 and T2 are also plotted
in Figures 7a1 and a2. Comparing these figures, T2
had a deeper failure and somewhat farther run-out
distance. Nevertheless, the final surfaces for T3 and
T4 are not plotted because there was no well defined
failure surface observed.
CONCLUSIONS
This experiment employed rainfall induced over-
land flows to infiltrate into sandy soil slopes to study
their stability. The characteristics of the failure mech-
anism and the responses of pore pressure and water
contents in two samples composed of sand and silty
sand, respectively, were observed.
The failure in permeable sand was a sliding mode.
Initial failure was noticed as a piping occurred at the
toe of slope, and it then gradually propagated upward
as a retrogressive failure. Apart from this, the failure
occurred in the slope with a dip stratum was fast (t
p
=
20 s and t
f
= 40 s), compared to that in the homoge-
neous slope (t
p
= 30 s and t
f
= 120 s). However, the
failure in less permeable silty sand was initiated by
erosion at shallow depth, the rill later expanded and
turned into either a flow or a complex mode of flow
and slide, depending on the rainfall intensity applied.
Moreover, sandy soil had a marked increase in pore-
water pressure when approaching failure; neverthe-
less, this phenomenon was less obvious for silty sand.
The initial profiles of total head in the slopes with
a dip stratum were approximately parallel to the im-
pervious stratum. Hence most of the lower slope, es-
pecially near the toe, was saturated when slope failure
occurred. This observation also confirmed the obser-
vation that the mode of retrogressive failure started
from the toe of sandy slopes.
The findings from this experiment are helpful to
clarify the failure mechanisms of granular soil slopes
such as the case study described in the introduction.
It is also hoped that the observation from this experi-
ment may provide a reference for numerical analyses
as well as future remedial work for similar soil slopes.
ACKNOWLEDGEMENTS
This research was financially supported by the
National Science Council, ROC.
Fig. 7 - Profile of total head at different times in tests T1 ~ T4
background image
EXPERIMENT ON THE STABILITY OF GRANULAR SOIL SLOPES BY RAINFALL INFILTRATION
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
311
REFERENCES
b
Rand
e.w. (1982) - Analysis and design in residual soil, Proceedings of the ASCE Geotechnical Engineering Division,
Specialty Conference- Engineering and Construction in Tropical and Residual Soils, Honolulu, Hawaii, 89-141.
C
Hen
H., C
Hen
R.H. & l
in
m.l. (1999) - Initiation of the Tungmen debris flow, eastern Taiwan, Environmental and Engineering
Geosciences, 5 (4): 459-473.
C
Hen
H., C
Hen
R.H., y
u
f.C., C
Hen
w.s. & H
unG
J.J. (2004) - The Inspection of the triggering mechanism for a hazardous
mudflow in an urbanized territory, Environmental Geology, 45 (7): 899-906.
C
Hen
R.H., C
Hen
H.P. & C
Hen
k.s. (2009) - Simulation of a slope failure induced by rainfall infiltration, Environmental Geology,
58 (5): 943-952.
H
uanG
C.C., l
o
C.l., J
anG
J.s. & H
wu
l.k. (2008) - Internal soil moisture response to rainfall-induced slope failures and debris
discharge, Engineering Geology, 101 (3): 134-145.
i
veRson
R.m. & m
aJoR
J.J. (1986) - Ground water seepage vectors and the potential for hillslope failure and debris flow
mobilization, Water Resources Research, 22 (11): 1543-1548.
k
im
J., J
eonG
s., P
aRk
s. & s
HaRma
J. (2004) - Influence of rainfall-induced wetting on the stability of slopes in weathered soils,
Engineering Geology, 75 (3): 251-262.
l
umb
P. (1975) - Slope failure in Hong kong. Quarterly Journal of Engineering Geology, 8: 31-65.
m
oRiwaki
H., i
nokuCHi
t., H
attanJi
t., s
assa
k., o
CHiai
H. & w
anG
G. (2004) - Failure processes in a full-scale landslide
experiment using a rainfall simulator, Landslides, 1 (4): 277-288.
o
Rense
R.P., s
Himoma
s., m
aeda
k. & t
owHata
i. (2004) - Instumented model slope failure due to water seepage, Journal of
Natural Disaster Science, 26 (1), 15-26.
R
oCHa
M. (1957) - The possibility of solving soil mechanics problems by the use of models - 4th International Conference on Soil
Mechanics and Foundation Engineering, London, U.K., 1: 183-188.
R
osCoe
K. (1968) - Soils and model tests - Journal of Strain Analysis, 3: 57-64.
t
oHaRi
a., n
isHiGaki
m. & k
omatsu
m. (2007) - Laboratory rainfall-induced slope failure with moisture content measurement,
Journal of Geotechnical and Geoenvironmental Engineering, 133 (5): 575-587.
w
anG
, G. & s
assa
k. (2003) - Pore-water pressure generation and movement of rainfall-induced landslides: effects of grain size
and fine-particle content, Engineering Geology, 69 (2): 109-205.
Statistics