Survey Review

ISSN: 0039-6265 (Print) 1752-2706 (Online) Journal homepage: http://www.tandfonline.com/loi/ysre20

STUDY OF THE DEFORMATION OF KOYNA DAM

USING THE GLOBAL POSITIONING SYSTEM

Ankur Manake & Madhav N. Kulkarni

To cite this article: Ankur Manake & Madhav N. Kulkarni (2002) STUDY OF THE DEFORMATION
OF KOYNA DAM USING THE GLOBAL POSITIONING SYSTEM, Survey Review, 36:285, 497-507

To link to this article: http://dx.doi.org/10.1179/sre.2002.36.285.497

Published online: 19 Jul 2013.

Submit your article to this journal

Article views: 8

View related articles

Citing articles: 2 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=ysre20

Download by: [University of Georgia] Date: 04 April 2016, At: 05:54

 Survey Review, 36, 285 (July 2002)

STUDY OF THE DEFORMATION OF KOYNA DAM USING THE

GLOBAL POSITIONING SYSTEM

Ankur Manake1 and Madhav N. Kulkarni2

1DeEartment of Mechanical Engineering

Department of Civil Engineering
Indian Institute of Technology, Bombay
Mumbai, India

ABSTRACT

Large engineering structures such as dams, though essential for development, are also a potential
hazard, causing loss of human lives and property, in case of their failure under excessive deformations.
Downloaded by [University of Georgia] at 05:54 04 April 2016

Hence, monitoring dam deformations has assumed great importance worldwide. Classical geodetic
techniques, and more recently, the revolutionary space geodetic technique: Global Positioning System
(GPS), are valuable tools for generating important input data to monitor such deformations. A project
for crustal dynamic studies in the Koyna Region in Maharashtra, India, using GPS, has been taken up,
which also includes the deformation studies for Koyna Dam. A GPS network has been established for
this purpose, and observed in two sessions viz. December 2000 and May 2001. The study results
indicate a definite correlation between the change in reservoir water level and the deformation of the
dam. The results obtained using the conventional methods are found in close agreement with the GPS-
derived results, thus proving GPS to be an effective toolfor dam deformation studies.

INTRODUCTION

Large engineering structures are a symbol of socio-economic development of a

region. However, these structures, such as dams, bridges, high-rise buildings, etc.,
undergo constant deformations due to various loading factors. As a result, they pose a
potential hazard, causing loss of lives and property in the event of their failure under
excessive deformations. Monitoring the deformations of such structures, particularly
dams, \vhose failure can cause havoc in the surrounding region, has assumed great
importance worldwide. These deformations are monitored using various techniques,
such as mechanical sensors, geodetic methods, and more recently, the revolutionary
space geodetic technique: the Global Positioning System (GPS).
A major factor contributing to the deformations of dams is a result of varying
stresses and strains due to the effects of the local crustal movements. Crustal
movements cause shifts in the relative locations of points situated on the dam body and
in the vicinity of the dam, thus subjecting the structure and its surroundings to various
stresses. In addition, the short-term elastic deformations can also be attributed to other
factors such as varying reservoir water level, temperature, seepage, erosion, concrete
deterioration, liquefaction etc. All these factors, when put together, can result in a
critical situation for the dam under peak load, which may lead to its failure. During the
last few decades, about fifty dam failures have been reported, out of which five have
been in India [4]. Hence, it becomes extremely important to study the deformations in
a dam structure due to changing water level, crustal motions, and other loading factors.
The study of deformations of Koyna Dam, in Maharashtra, India, has been carried
out in the past using various geodetic techniques, with which the second author was
associated [4, 5]. In order to carry out dam and crustal deformation studies using GPS
in this seismically active region, the Department of Science and Technology,
Government of India (DST), sanctioned a project in Sept. 2000 to the Indian Institute
of Technology, Bombay, with the second author as the principal investigator. The

497
 THE DEFORMATION OF KOYNA DAM

observations taken on this dam structure and the results obtained thereof have been
presented here, along with a comparison with the results obtained from the
conventional methods.
The implementation of a high precision GPS-based structural deformation
monitoring system at the few millimetre level accuracy requires careful consideration
of error mitigation, system design and operational efficiency. GPS provides three
dimensional positions with reference to a fixed global coordinate system,
simultaneously at all sites, with automated data collection, and no line-of-sight
requirement between the points. This is the primary advantage of GPS over
conventional surveying techniques. Another great advantage is the capability of GPS
to measure deformations in real time mode. Due to the various advantages that GPS
offers for structural deformation studies, many structures worldwide are being
monitored using this technique.
The Pacoima Dam in California, USA, which is a 113 metre tall concrete arch
structure, is being monitored continuously, using GPS [2]. One receiver is placed at the
Downloaded by [University of Georgia] at 05:54 04 April 2016

abutment of the dam and the other is placed at the center of the arch of the dam. A
reference station is mounted several kilometre away on bedrock, which is expected to
be stable. The measurements using dual frequency P-code receivers have revealed
displacement-dependence in the E-W direction on factors like temperature and
reservoir water level.
The Calgary tower in Canada, which is a 160 metre tall structure, was monitored
using the differential GPS carrier-phase approach [1]. Two GPS receivers were
installed on the tower (one to act as a back up, in case of failure of the other), with the
third receiver on a fixed station, to act as a reference station. The N-S amplitude of the
tower deformation is reported to be ±15 mm, with E-W amplitude of±5 mm. Since the
standard deviation of the E-W direction movement was estimated as 5mm, it was not
considered significant. The deformation was attributed primarily to wind and
temperature.

CASE STUDY: KOYNA DAM, MAHARASHTRA, INDIA

Koyna Dam is an 85 metre high rock-filled structure, with 892 sq. km. of catchment
area, situated in the Indian Peninsular area. It is a rubble concrete dam, which is a
mechanised, and modem version of old cyclopean/masonry or rubble concrete. The
Indian Peninsular Shield, one of the oldest continental blocks on the earth's surface,
was traditionally assumed to be a stable block; hence, the 1967 earthquake, with
magnitude 7.5 on the Richter scale, came as a surprise to earth scientists. The dam
withstood this significant event without any damage. However, this evoked interest on
the part of geologists, geodesists, dam experts and engineers, leading to various studies
of the stability of the dam structure and the surrounding region. The dam has been well
instrumented for observations of the structural behaviour, right from the design stage.
These observations assumed significant importance because of the continuing tremors
in the zone, after the 1967 event. This instrumentation consists of various gauges
embedded in the dam structure or installed around the dam body, including uplift
pressure cells, coordimetres, thermometres, dial gauges, piezometers, and stress and
strain meters. The data from these instruments has been continuously collected and
analysed by the engineers concerned [3]. Based upon this huge amount of data,
collected over several years, a mathematical model for predicting the dam deflection
has been formulated.

498
 A. MANAKE AND M. N. KULKARNI

Mathematical Model to predict deflection of Koyna Dam

A mathematical model for forecasting the displacement of dam structure provides a

tool for its monitoring, as abnormalities in deflection serve as a warning to initiate a
close examination of the dam. The mathematical model developed by Electricite' de
France, (EDF) Grenoble, for forecasting the dam displacements, assumes a multiple
linear regression model, which comprises three components [3]:

1. An irreversible time-depe.ndent component f](t), expressed as:

fi(!) = b Ie + b2e-1 (1)

2. A component 6(Z) due to water level, expressed as:

Downloaded by [University of Georgia] at 05:54 04 April 2016

(2)

where Z is the ratio of deviation of reservoir level from the normal level to the height
of the dam, and b 1. b 2 . .. etc. are constants to be estimated.

3. A cyclical component f3(S), expressed as:

13(S) = b7Cos S + b8 Sin S + b9 Sin2 S + b 10 Sin S Cos S (3)

where S is the angel assigned to the day to which displacement of the dam
corresponds, the angle from a
to 360 being assigned serially to the successive days of
0

a calendar year.
When the collected data is fitted into the function, the resulting expression is:

Y = -70329.6035 - 0.1865 ADSFT + 2.0657 TE - 0.5601 Cos S

- 0.8090 Sin S+ 7.5573W (4)

where
Y is deflection of Koyna dam in mm
ADSFT is the average downstream face temperature in deg F
TE is the time elapsed since January 1, 1969 with a decade as a unit of time
S is the angle assigned to the day, as explained in equation 3 above,
W is the dam reservoir level in feet

The results from the GPS data, collected in this project, can contribute significantly
to the enhancement/augmentation of the mathematical model explained above.
However, several years of GPS observations are required for this task.

GPS NETWORK FOR KOYNA DAM REGION

In order to carry out structural and crustal deformation studies using GPS in this
region, a DST-funded research project has been taken up since September, 2000. An
extensive GPS network comprising 35 stations has been established in the area. Of
these, 12 stations are established on the dam body, and the rest on the ground
surrounding the dam and the reservoir. Out of the 12 stations established on the dam, 6
are on the dam axis, 3 on dam piers (top of the buttresses) and 3 on the base of the

499
 THE DEFORMATION OF KOYNA DAM

dam. A total of 9 old geodetic network stations [4] are included in this GPS network,
with one geodetic station of the Great Trigonometrical survey (GT), which acts as the
basic geodetic framework for the Indian subcontinent. Proper monumentation has been
done for all the stations. The targets, which were embedded in the down-stream face of
the dam wall for the conventional geodetic monitoring (by observing them as
intersected points by theodolite), could not be occupied, as it is not possible to set up
GPS antenna on these points. Hence, new points on the dam body have been
established, so as to cover the entire dam structure, from the top to the base of the dam.
This distribution of points is expected to reveal the deformation behaviour of the dam
structure at various points. This GPS network is shown in Figure 1.

LEGEND
D DEFORMATION
Downloaded by [University of Georgia] at 05:54 04 April 2016

POINTS [~]
fJ SAFE1Y
POlNTS[&]
RESERVOIR • POINIS C«
DAM AXIS 16]

• POINI'SON
DAMPIER [3]

• POINI'S 00-
DAM:BASB(3)

.• FAULT
POINTS [8J

o BACKSlGHT
POINTS(3)

• GTPOINT[l]
!.ANGALI JAIG.AD

• lIG POllITS [2]

• :BASE STATION

Figure 1: GPS Network at Koyna Dam Region

In order to study the dam deformation, the GPS network comprising the points on the
dam body, and the points in the deformation zone (which is in the vicinity of the dam,
on the down-stream side of wall), i.e. Dl, D2, D3, D4 and D5, is considered here. The
points in the deformation zone are monitored, in order to establish a relationship
between the crustal motion and the dam deformation. The changes in baseline lengths
between these points can reveal the deformation pattern of the dam structure, as well as
the surface deformations of the earth's crust. The observations have been carried out
using four Trimble 4000SS1 dual frequency geodetic GPS receivers, with choke-ring
antennas. The salient features of the GPS data collection are:

- Period of Observation : 6 hours

- Sampling rate : 15 seconds
- Minimum Number of Satellites : 4
- Satellite Elevation Mask: 150

A base station was set up on the top-most point of the dam, and was run continuously
throughout the period of observations. This base station has been used as a reference

500
 A. MANAKE AND M. N. KULKARNI

station for relative positioning in the post-processing mode. The GT station, Jangli
Jaigad H. S., was occupied for a longer period of 24 hours, in order to achieve higher
accuracy in the estimation of the longer baseline length (about 7 km), to the base
station. The change in baseline length between GT point and the base station gives us a
good estimate of the deformation of the reference station on the dam body, from the
stable GT point.
Till-date, this GPS network has been observed in two sessions: first time in
December 2000 (Epoch S1), and second time in May 2001 (Epoch S2), during the last
6 months. Over this period, the changes in the reservoir water level and changes in the
atmospheric temperature and the temperature of the dam body have been recorded.
Generally, these factors contribute significantly to the deformation of the structure, in
addition to the effect of the seismic activity in the region.

ANALYSIS OF THE RESULTS

Downloaded by [University of Georgia] at 05:54 04 April 2016

The GPS data has been processed using Trimble GPSurvey software version 2.3, and
the network adjustment has been done using Columbus 3.0 software. A detailed
analysis of the results obtained for the displacement of the base station, and
deformations within the dam body are presented here.

Variations in Base Station Co-ordinates

In order to analyse the variations in the co-ordinates of the reference station, data
was collected in different patterns in both the sessions. In S1, the data was collected for
about 36 hours on hourly basis, in order to analyse the hourly variations of the absolute
position of this point. This variation can be attributed mainly to the changes in the
temperature over the day, as well as the inherent GPS signal fluctuations. A somewhat
cyclic trend can be observed in the latitude, longitude variations plotted in Figure 2.

VI
"U 1.5
c
o
o
4J
(I)

~c 0.5
ro
4J
~
4J
£
E
~
-1
4J
C)
c
~ -1.5
u

-2

Hour of the Day

I Second = JO m appro:\:.
-+- Longitude -. -.- .. Latitude

Figure 2: Hourly Variations of Base Station Co-ordinates

501
 THE DEFORMATION OF KOYNA DAM

Displacement of the Base Station

The base station displacement has been studied with reference to the Jangli Jaigad
GT point, which we are assuming as un-disturbed, being sufficiently away from the
deformation zone of Koyna (this will be verified subsequently, by connecting to the
GT network). In session SI, the base station was occupied and its absolute co-ordinates
in WGS84 datum were determined. In session S2, the co-ordinates of GT point were
held fixed, in order to estimate the relative deformation of the base station, using the
relative GPS technique in post-processing mode. The changes in the co-ordinates of
the base station are tabulated in Table 1.

Table 1: Changes in Base Station Co-ordinates

LATITUDE LONGITUDE
(Seconds Only) (Seconds Only)
Downloaded by [University of Georgia] at 05:54 04 April 2016

December 2000 06.45012 07.76487

May 2001 06.45493 07.73760
Change 0".00481 = 0.144 m - 0".02727= - 0.818 m

The results given in Table 1 indicate that there has been a displacement of the Base
Station in the N-W direction. The change in the baseline length between the Base
Station and the GT point is tabulated in Table 2.

Table 2: Change in Baseline Length over the two epochs

Baseline Length Estimated Std. Deviation

(m) (mm)
Dec. 2000 (Epoch S1) 7909.643 0.9
May 2001 (Epoch S2) 7909.647 0.5
Change (S2 - Sl) 0.004

The change in the base line length indicates an expansion of the baseline by 4 mm,
which is in agreement with the estimated change in the co-ordinates of the base station.
The high precision of the results obtained here is clearly evident from the sub-
millimetre level standard deviation, as shown above.

Deformation of/he Dam Structure

A total of 6 GPS' stations have been established on the dam axis. The reference
station for estimating the deformations of these points is the Base station set up on the
dam, hence the deformations estimated are only relative. The estimated changes in the
baseline lenbrt:hsbetween these points, along with the standard deviations, are shown in
Table 3. These deformations in the baselines indicate the strain accumulating at
various points on the dam. The estimated deformations show a decrease in both
latitude and longitude, indicating a shift of the dam axis with reference to the base
station, in the S-W direction~ however, more repeat observations are required to
confirm this trend. A detailed analysis of deformations relative to GT stations, in NS
and EW directions, is in hand.

502
 A. MANAKE AND M. N. KULKARNI

Table 3: Changes in Baselines between points on the Dam Axis

S. Base Line DECEMBER 2000 MAY 2001 Change in

No From To Baseline Std. Baseline Std. Baseline
Length Dev. Length Dev. (mm)
(m) (mm) (m) (mm)
1 DAMA1 DAMA4 515.964 0.5 515.966 0.6 2.0
2 DAM4 DAMA6 187.503 0.4 187.494 0.4 -9.0
3 DAMA6 DAMA1 703.463 0.8 703.455 0.9 -8.0
4 DAMA2 DAMA3 184.165 0.1 184.156 0.2 -9.0
5 DAMA3 DAMA5 349.234 0.2 349.240 0.3 6.0
6 DAMA2 DAMA5 533.389 0.2 533.386 0.4 -3.0
Downloaded by [University of Georgia] at 05:54 04 April 2016

In a similar manner, three GPS points have been established on the Dam Piers, one on
the downstream, and the other two on the upstream side. These points were observed
with the base station as the reference station. Estimated displacements of these points
also indicate a shift in the S-W direction. In addition, three more GPS stations have
been established on the base of the dam. These are located such that they cover
maximum area at the bottom and also near the overflow section, since the load of the
water reservoir is maximum at the base of the dam. The estimated displacetnents of
these points indicates a shift towards the N-W direction, which is anomalous to the
trend we have seen so far, which is the S-W direction. This needs further
investigations.

Deformations in the Deformation Zone

In order to look at the relationship between the deformation of the dam and of the
previously identified deformation zone, located in the vicinity of the dam on the down-
stream side, the changes in the baseline lengths between the 5 deformation pillars on
the downstream side of the dam are considered. Out of these, three are on the left bank
and two on the right bank of the river. The deformation vectors for these pillars are in a
direction away from the Koyna river channel down-stream of the dam, i.e. the pillars
on the left bank indicate movement in the S-W direction and the pillars on the right
bank, in the S-E direction (see Fig.4). This can be attributed to the sideward force
exerted by the downstream flow, on the channel banks.

Integrated Analysis of the Deformation Pattern

An integrated analysis of the deformation pattern in the region, taking into

consideration the structural deformations of the dam body, and the crustal
deformations in the deformation zone and safety zone, has been taken up. The
deformation vectors for the GPS stations, over the period between the two epochs of
observations, are plotted in Figure 4. As discussed above, the deformation vectors
indicate a uniform trend, with the points on the dam body showing displacements away
from the reservoir, and the deformations pillars following a trend of movement away
from the downstream water channel. This deformation pattern obtained from the GPS
results is compared with the pattern indicated by the instrumentation data, in order to
study the correlation between the results from these two independent techniques.

503
 THE DEFORMATION OF KOYNA DAM

CORRELATION WITH RESULTS OBTAINED FROM INSTRUMENTATION DATA

The dam deformations estimated from the data from instrumentation are reported in
the annual reports by the Irrigation Department, Konkan region, for Koyna Hydro-
Electric project [3]. In order to analyse the structural behaviour of the dam, the
deflection of the dam perpendicular to the dam axis is monitored. This deflection has
been attributed to various factors, which include reservoir water level and temperature.
The data is being collected since 1961 to study the stability of the structure, and these
studies were intensified after the 1967 earthquake [5].

From the instrumentation data obtained from the Dam Safety Organization, Nashik,
Maharashtra, for deflections measured by coordimeter at M25 (monolith number 25)
for the year 2000, the curve shown in Figure 3 has been plotted. This deflection is
plotted against the reservoir water level, since it is the highest contributor to the dam
deformation. It is observed that over the years there is a cyclic pattern in the
Downloaded by [University of Georgia] at 05:54 04 April 2016

deformation of the dam. The coordimeter is located at a height of 583.41m, which is

close to the Dam base points, established for GPS measurements.

t160

~
650 ~
w
~
!;

640 ~
+VF. DER.EcnON TOWARDS DIS SlOE
.VE DEflECTION TOW ARDS U/S SIDE

o 4 7

DEFLECTION PERPENDICULAR TO AXIS IN MM

Figure 3: Variations of Deflection perpendicular to the Dam with Reservoir Water Level

As evident from Figure 3, the deformation observed using this method shows a
negative deformation, i.e. a deformation in the upstream direction. The deformation
using GPS on a level similar to the level of the coordimeter, i.e. at the Dam base
points, also indicates a change in the N-W or the upstream direction. The order of
magnitude of the deformation in between these two epochs, estimated from these two
methods is also the same, which is 3 mm using the coordimeter, and 3 mm to 8 mm at
different locations on the base of the dam, using GPS (see Figure 4).

504
 A. MANAKE AND M. N. KULKARNI

I:lAMA! I' .,
Downloaded by [University of Georgia] at 05:54 04 April 2016

Figure 4: Deformation Vectors of various points on and near the dam. \V.r.t. the Base Station

Interpretation of the Results

The instrumentation data archived since 1966 show deflections ranging from 2.7 mm
to 13 mm at different locations. The upstream deformation of the base of the dam can
be attributed to the fall in water level. The maximum load due to water level is at the
bottom of the reservoir and of the dam structure. Thus, as the water level increases the
dam is pushed southwards or towards the downstream. In between the two epochs
observed, there was no rainfall, hence the water level had decreased. This reduction of
water level was more than 15 metre, thus reducing the load on the bottom of the datu,
resulting in elastic movement of the structure northwards or in the upstream direction.
This deflection in the upstream direction at the bottom also induces an axis of rotation,
because of the rigidity of the structure. This axis would lie above the base level
observed, and hence cause the top of the dam to shift in the opposite direction i.e.
southwards or in the downstream direction. This phenomenon is also supported by the
shift in the coordinates of the axis and pier points, as inferred from GPS data, which
shows a deflection towards the downstream. Thus, the deformation patterns obtained
from the results of these two independent techniques indicate a very high correlation.

505
 THE DEFORMATION OF KOYNA DAM

CONCLUSIONS

An extensive network has been set up at the Koyna Dam and surrounding region to
carry out deformation studies using the revolutionary space geodetic technique: GPS.
GPS proves to be a very effective tool to carry out the deformation studies, due to its
high precision and accuracy. From the results and inferences presented here, it is
evident that GPS is a very economic, efficient and effective technique for monitoring
the structural and crustal deformations. The following conclusions can be drawn:

GPS is an effective method for round-the-clock static deformation monitoring.

It works under any weather conditions, without compromising on the accuracy.
The hourly data variation shows a definite correlation between the change in
the day temperature and the changes in the co-ordinates of the GPS stations on
the dam body.
Downloaded by [University of Georgia] at 05:54 04 April 2016

The results from the two sessions of GPS observations indicate that the
deformations estimated at the various points on the dam structure follow a
uniform pattern. A direct correlation between fall in the reservoir water level
and the dam deformation is also established. As the water level of the reservoir
falls, the base of the dam tends to shift in the upstream direction and as an
inverse action, the points on the axis tend to shift towards downstream. These
phenomena observed from the GPS observations have also been confirmed by
the results from the conventional instrumentation methods used to monitor the
dam structure. However, more repeat observations, and comparison with
conventional data are required for conclusive analysis.

The Mathematical model developed for predicting dam deflection, which is

based upon the results of the conventional instrumentation methods, can be
enhanced/augmented/modified, using GPS-derived results. This would help in
making a better forecast for the dam deflection. However, this will only be
possible after repeat GPS observations of the network over several years.

The GPS-derived deformation pattern is in close agreement with the results

from the conventional instrumentation data. Hence, we can conclude that these
two independent techniques are complimentary to each other.

In order to carry out detailed statistical analysis to determine the deformation

characteristics, data over a longer period of time is required. At least about 10
sessions of the data needs to be collected in order to formulate a statistical
model using GPS·for determination of deformations of the dam structure. Such
a model would probably be applicable, at least partially, to other large
engineering structures, designed to last a few decades, which undergo
deformations at a very slow rate.

ACKNOWLEDGMENTS

This work forms a part of a research project of the Department of Science and
Technology, Government of India. The financial & administrative support provided by
DST, and Indian Institute of Technology Bombay, is duly acknowledged. The
fieldwork for the project was carried out with the support of research fellows: V.S.
Tomar and Praveen Pillai. This study also forms a part of the M. Tech. dissertation

506
 A. MANAKE AND M. N. KULKARNI

submitted by the first author, under the guidance of the second author, at 1. 1. T.
Bombay [6].

References

1. Celebi, M. et. AI., 1998. GPS Monitoring of Structures: Recent Advances, proc.
30th Joint Meeting of US-Japan Co-operative Programme in Natural Resources,
August.
2. Kenneth, W. et. AI., 1998. Continuous GPS Monitoring of Structural Deformation
at Pacoima Dam, California, Seismological Research Letters, Vol. 69, No.4.
3. Irrigation Department, Konkan Region, 1982, Data Compilation Report on
Structural Behaviour of Koyna Dam, Koyna Hydro-electric project, Govt. of
Maharashtra, Koyna Nagar, India.
4. Kulkarni, M.N., 1986, Geodetic Surveys for Dam Deformation Studies, Indian
Downloaded by [University of Georgia] at 05:54 04 April 2016

Surveyor, January, Journal of the Institution of Surveyors (India): 23-25.

5. Kulkarni, M. N. and A. Manake, 2000. Dam Deformation Studies using GPS &
Geodetic techniques, proc. Indian Geotechnical Conference-2000, Civil Eng.
Dept., I. I. T. Bombay, India: 253-256.
6. Manake, A., 2001, Study of Deformation of Large Engineering Structures Using
Global Positioning System, M. Tech. dissertation, Mechanical Eng. Department, I.
I. T. Bombay, India.

507