Content uploaded by Pratik Chaturvedi
Author content
All content in this area was uploaded by Pratik Chaturvedi on Mar 03, 2019
Content may be subject to copyright.
Training of Sensors for Early Warning
System of Rainfall-Induced Landslides
Naresh Mali, Pratik Chaturvedi, Varun Dutt, and Venkata Uday Kala
Abstract
Landslides have been a major issue in the Himalayan
region where slopes are cut and reformed for construction
practices for infrastructure development, deforestation,
and many other human activities. In lieu of the mitigation
measure for rainfall-induced landslides to improve the
factor of safety against failure, several warning techniques
have been suggested. However, they are quite expensive,
resulting in an only limited application for infinite slopes.
In lieu of the existing conditions, early warning systems
(EWS) for detecting slope failure using the sensors have
been found to be handy to control the fatality of the
disaster. But, the various sensors have been used for these
warning systems are not unique. Hence, they need to be
trained for each type of soil and other favorable
conditions. For the proposed study, Micro-Electro-
Mechanical Systems (MEMS) based sensors have been
used to predict the slope failures under rainfall conditions
at controlled laboratory scale prototype and to perform a
series of flume tests in order to develop the threshold for
moisture levels and movement that can trigger the slope
failure.
Keywords
Slope-instability Flume test Sensors
Early warning system
1 Introduction
In Himalayan region of India; slope failures predominantly
occur during or immediately after rainfall [1–4], which leads
to increase in piezometer levels such as rainwater infiltration,
thereby, triggering slope failures. Several mitigation tech-
niques have been proposed in order to decrease the effect
due to landslides. However, where the slopes are steep and
extend to greater heights, most of these may be neither
applicable nor economical. In such situations, the early
warning systems for detecting slope failure using the sensors
have been found to be handy to control the fatality of the
disaster. But these warning systems employing various
sensors are not unique. Hence, efforts have been made to
train the sensors for each type of soil and other conditions to
retrieve the data from the field. During the course of the
study, the multi-disciplinary involvement for bringing out
the sensors, assembly, applications, calibration, testing,
placing, data retrieving and model-based predictions were
developed [1,3,4].
In the proposed study, Micro-Electro-Mechanical Sys-
tems (MEMS) based sensors were used to predict the slope
failures under rainfall conditions at controlled laboratory
scale prototype and to perform flume tests in order to
develop the threshold for moisture levels and movement that
can trigger a slope failure.
2 Materials and Methodology
Most of the slope failures in and around Mandi region of
Himachal Pradesh are shallow failures. The index properties
of soil (Table 1) were determined (IS-2720). The present
methodology developed by performing flume tests (Fig. 1.)
considering various conditions such as dry state and also by
increasing the amount of moisture levels [1,3,4] within the
slope. However, while performing the tests, the amount of
moisture content, displacement, acceleration and velocity
N. Mali (&)V. U. Kala
School of Engineering, Indian Institute of Technology Mandi,
Kamand Campus, Mandi, 175005, Himachal Pradesh, India
e-mail: malinareshmudhiraj@gmail.com
P. Chaturvedi
Scientist ‘D’with DTRL, DRDO, New Delhi, India
V. Dutt
School of Computing and Electrical Engineering, Indian Institute
of Technology Mandi, Mandi, Himachal Pradesh 175005, India
©Springer Nature Switzerland AG 2019
A. Kallel et al. (eds.), Recent Advances in Geo-Environmental Engineering, Geomechanics and Geotechnics,
and Geohazards, Advances in Science, Technology & Innovation, https://doi.org/10.1007/978-3-030-01665-4_104
449
were monitored using sensors and identifying the response
of soil slope failure. Based on the experimental threshold
sensors data, the SMS alert will be generated.
3 Results and Discussion
Tri-axial accelerometer: It is capable of measuring accel-
eration forces (static and dynamic) by providing simulta-
neous measurements in three orthogonal directions x, y and
z. Thus, the sensor could be used for the analysis of dif-
ferent vibrations experiencedbyastructure.Thissensor
expends two capacitors formed by a moveable plate held
between two fixed plates. Under zero net force, the two
capacitors are equal but a change in force causes the
moveable plate to shift closer to one of the fixed plates,
increasing the capacitance.
Soil-moisture sensor: It is usually used to detect the soil
humidity. When the soil is wet, the output voltage decreases
but it increases when the soil is dry.
Force sensor: It is a piezo-resistive conductive polymer,
which changes resistance in a predictable manner following
application of force to its surface. Like all resistive sensors,
this requires a relatively simple interface and can operate
satisfactorily in moderately hostile environments.
Tilt sensor: Tilt sensors are devices that produce an
electrical signal, which varies with angular movement.
These sensors are used to measure slope and tilt within a
limited range of motion. They are usually made by a cavity
and a conductive free mass inside, such as a blob of mercury
or rolling ball. One end of the cavity has two conductive
elements (poles). When the sensor is oriented so that its end
is downwards, the mass rolls onto the poles and shorts them,
acting as a switch.
3.1 Soil Moisture Sensor
For sensing the soil-moisture content in percentage, we used
YL-69 module (Fig. 2a). Analog readings for soil-moisture
sensors in dry and completely wet states were 395 (0%
moisture) and 1022 (100% moisture), respectively. To rep-
resent the analog values from the soil-moisture sensor in
percentage, and hence the moisture percentage was calcu-
lated by the following equation.
Moisture Percentage ¼1022 Analog value
1022 395 100 ð1Þ
The results depicted that, based on the performing of
flume tests, initially the tests were conducted for more than
Table 1 Physical properties of
sand Symbol Description Value
G
s
Specific gravity 2.62
%Gravel size 4
%Sand size 49
%Silt size 44
%Clay size 3
USCS Soil Classification SP (poorly graded sand)
C (kPa) Cohesion 5
UAngle of internal friction 28°
Fig. 1 Line diagram of
experimental setup (Not to scale)
450 N. Mali et al.
the critical angle the failure is observed at around 12%
moisture contents whereas, the tests conducted for less than
the 17% moisture levels. According to [1,3,4], moisture
content after reaching the threshold value causes the failure
of the slope leading to debris flow.
The failure of slope also depends upon several factors
such as soil type, angle of inclination, ramp surface texture
and others.
3.2 Accelerometer
For the determination of acceleration, GY-61 accelerometer
module was employed. It measures the tri-axial accelerations
in three orthogonal axes (Fig. 2b). This accelerometer is
capable of measuring acceleration in the range of ±3g
(where, g = 9.81 m/s
2
), in each of the three orthogonal axes.
For the accelerometer placed at the top and base of the pipe,
a
x
,a
y
, and a
z
refers to the axis perpendicular to the base of
the ramp (pointing upwards), along the width of the ramp
(from right to left), and sloping at an angle with the soil
away from the ramp (his the angle of the ramp with hori-
zontal), respectively. Based on the predefined threshold
values, the modules outputs LOW, otherwise, it outputs
HIGH. The threshold value for the digital signal can be
adjusted using the built-in potentiometer. The integration of
different sensors to a microcontroller and then the micro-
controller’s connection to an Internet cloud was carried out.
The microcontroller receives data from sensors and it is
connected to a GSM modem for transmitting the sensors’
readings to the web server. The values from sensors were
logged onto an in-house developed web server http://www.
landslidemonitoring.esy.es/ with the help of GSM Module.
After the successful reception of the data from soil
moisture sensors and accelerometer sensors, the landslide
probability was computed with the weights assigned to
individual sensor values. When the probability value crosses
a prefixed threshold (in this study, 85), an alert would be
triggered to the registered users.
3.3 Experiences During the Training Sensors
While Performing the Tests
1. Soil moisture sensor needs to be calibrated frequently for
calculating the soil-moisture percentage.
2. The probes of the soil moisture sensor are at a distance of
37 mm apart, hence the resistance will be created at the
probe, but not in the gap.
3. Calibration should be achieved before placing these
sensors in the soil at the time of testing.
4. Calibration values may not be the same for all the sensors
5. Flex sensors should be supported to avoid getting
detached from the one end of the node.
6. Any change in voltage would affect the tilt sensor.
4 Conclusion
Early Warning System (EWS) architecture is in place for the
people to be alerted about the oncoming disaster. The
threshold values are evaluated from the analysis of logged
0 5 10 15 20 25 30 35 40 45 50 55 6 0 65 70 75 80 85 90 95 100
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Moisture Content (%)
Time (min)
Tests condcucted for less than critical angle
Tests condcuted for more than critical angle
01020304050
-6
-4
-2
0
2
4
6
8
10
12
14
16
18
20
22
Accleration (m/s2)
Timer (min)
ax2
ay2
az2
Plot showing the moisture levels with
respect to time period
Plot showing acceleration (m/s2) with respect
(a) (b)
to time period
Fig. 2 aPlot showing the moisture levels with respect to time period. bPlot showing acceleration (m/s
2
) with respect to time period
Training of Sensors for Early Warning System of …451
soil-moisture and soil-movement values, by performing the
set of flume tests on the ramp. Once the soil-moisture or
soil-movement thresholds are reached, the above-mentioned
alert generation unit generates landslide alerts. As soon as
any activity in the area under surveillance of sensors crosses
a pre-determined limit, an alert would be triggered to inform
the concerned people to take necessary steps. Hence, as soon
as any value of the database crosses the threshold level, an
alert is sent via an SMS.
Acknowledgements The authors would like to thank the State
Council for Science, Technology & Environment, Himachal
Pradesh, India, for providing the financial support to pursue this study.
References
1. Huang, C.-C., Lo, C.-L., Jangand, J.-S., Hwu, L.-K.: Internal soil
moisture response to rainfall-induced slope failures and debris
discharge. Eng. Geol. 101, 134–145 (2008)
2. Fell, R.: Landslide risk assessment and acceptable risk. Can.
Geotech. J. 31, 261–272 (1994)
3. Lourenco, S.D.N., Sassa, K., Fukuoka, H.: Failure process and
hydrologic response of a two layer physical model: implications for
rainfall-induced landslides. Geomorphology 73, 115 –130 (2006)
4. Wu, L.Z., Huang, R.Q., Xu, Q., Zhang, L.M., Li, H.L.: Analysis of
physical testing of rainfall-induced soil slope failures. J. Environ.
Earth Sci. 73(12), 8519–8531 (2014)
452 N. Mali et al.
