Long-term Deformation Patterns of Earth-fill Dams based on Geodetic Monitoring Data: the Pournari I Dam Case Study

Abstract

Dam safety is crucial taking into account that the vast majority of these structures are expected to exceed their design lifespan by 2020 alongside with the increasing magnitude of extreme flood incidents. Earth-fill dams deform significantly during the critical phase of the first reservoir filling while the rate of deformations is decreased in the long-term. Systematic monitoring and data analysis enables the evaluation of the ongoing performance of a dam, the validation of laboratory models and the assessment of different engineering designs. Relationships describing the long-term evolution of dam deformations are currently not well documented due to lack of available monitoring data. This study investigates the long-term deformation patterns of earth-fill dams using as a case study the Pournari I dam. A 29-year long dataset was analysed which consisted of geodetic measurements of vertical and horizontal deformations of the dam crest and the reservoir level fluctuations. A comparison was then carried out with the results obtained from other case studies, e.g. the Kremasta dam. The analysis indicates that the crest settlements of Pournari I dam remained within normal limits for the time period examined. The rate of deformations was also stabilised almost seven years after the completion of the dam, which is longer than the period suggested in the international literature.

Full text

3rd Joint International Symposium on Deformation Monitoring Vienna, 30th March–1st April 2016
Long-term Deformation Patterns of Earth-fill
Dams based on Geodetic Monitoring Data: the
Pournari I Dam Case Study
P. Michalis, S.I. Pytharouli
Department of Civil and Environmental Engineering,
University of Strathclyde, James Weir Building, 75 Montrose Street, G1 1XJ, Glasgow, UK
Email: p.michalis@strath.ac.uk
S. Raftopoulos
Public Power Corporation S.A., Agisilaou 56-58, 104 36, Athens, Greece
Abstract. Dam safety is crucial taking into account
that the vast majority of these structures are expected
to exceed their design lifespan by 2020 alongside
with the increasing magnitude of extreme flood
incidents. Earth-fill dams deform significantly
during the critical phase of the first reservoir filling
while the rate of deformations is decreased in the
long-term. Systematic monitoring and data analysis
enables the evaluation of the on-going performance
of a dam, the validation of laboratory models and the
assessment of different engineering designs.
Relationships describing the long-term evolution of
dam deformations are currently not well documented
due to lack of available monitoring data.
This study investigates the long-term deformation
patterns of earth-fill dams using as a case study the
Pournari I dam. A 29-year long dataset was analysed
which consisted of geodetic measurements of
vertical and horizontal deformations of the dam crest
and the reservoir level fluctuations. A comparison
was then carried out with the results obtained from
other case studies, e.g. the Kremasta dam.
The analysis indicates that the crest settlements of
Pournari I dam remained within normal limits for the
time period examined. The rate of deformations was
also stabilised almost seven years after the
completion of the dam, which is longer than the
period suggested in the international literature.
Keywords. dam deformations, dam safety, geodetic
monitoring, long-term, embankment.
1 Introduction
The last decades, the increasing demand for water,
flood control projects and power supply has led to a
corresponding increase in the construction of dam
structures. However, the vast majority of the existing
dam infrastructure is expected to exceed its lifespan
by 2020. Heavy rainfall events and more frequent
flooding incidents are considered major contributing
factors to the failure of earth-fill dams. The rate of
failure during the lifespan of a dam is estimated to be
1/10,000 per dam/year. During the first five years of
operation this value becomes 1/1,000 per dam/year
(Baecher and Christian, 2000) making this period the
most critical in the operational life of a dam.
Deformations of earth-fill dams start to take place
during the construction phase due to increase of
effective stresses among the various zones of the dam
and the creep mechanism of the earth material. Dams
continue to deform significantly during the first
filling of the reservoir, which is considered the most
critical phase as seepage resistance, foundation,
abutments and reservoir rim are tested for the first
time (USSD, 2008). The rate of deformations is
decreased in long-term and eventually drops to a
small constant rate per year. Factors influencing the
deformation rate are: the reservoir level fluctuations,
internal erosion, slope instability, creep of the
shoulder fill, the effect of secondary consolidation
and seismic activity (Tedd et al., 1997).
Systematic monitoring and analysis of dams is a
very important tool that provides early warning signs
of an impending failure and a better understanding of
the on-going performance of the structure. Numerous
studies have focused on the analysis of the behaviour
and the safety of dams based on monitoring data, e.g.
Heck (1984), Dascal (1987), Tedd et al. (1997),
Boros-Meinike and Jankowski (2003), Popovici et al.
(2004), Alonso et al. (2005), Kyrou et al. (2005),
Guler et al. (2006), Szostak-Chrzanowski and

3rd Joint International Symposium on Deformation Monitoring Vienna, 30th March–1st April 2016
2
Massiera (2006), Pytharouli and Stiros (2009),
Kalkan et al. (2010).
This paper presents the post-construction
behaviour of one of the largest earth-fill dams in
Greece based on geodetic monitoring data. We
evaluate the deformations of the control points
located on the crest of the dam based on geodetic
observations without any geotechnical constraints.
The investigated control points correspond to
locations where maximum deformations are
anticipated.
2 The Pournari I Dam
The Pournari I dam is located in Western Greece and
construction was completed in 1980 while the first
impoundment of the reservoir started in 1981. The
dam is owned by the Public Power Corporation of
Greece (PPC S.A.) and is rated 5th in energy
production amongst the hydroelectric power stations
in Greece. Pournari I dam, shown in Figure 1(a), is
the 5th larger dam in Greece with material volume of
9×106 m3 and one of the largest dams in Europe.
Fig. 1 (a) Downstream shoulder of Pournari I dam with the
concrete spillway (left) and part of the hydroelectric station
(right). (b) Cross section of the Pournari I dam at the location
where maximum height is obtained.
Pournari I is an earth-fill dam composed of sand
and gravel material with a central clay core, shown
in Figure 1(b), while the maximum capacity of its
reservoir is 865×106 m3. The maximum and
minimum water levels for hydroelectric power
production are at an altitude of 120 m and 100 m
above mean sea level (AMSL), respectively, with the
critical water level for the safety of the dam at 126 m
AMSL (PPC S.A., 1981).
The central clay core is covered with sandy gravel
filters. Rock fill shoulders of varied gradient were
constructed at each side of the dam with a step at
elevation of 80 m and 65 m at the upstream and the
downstream sides respectively, as shown in Figure
1(b). The width at the base of the dam is 453 m, its
maximum height is 107 m (from the foundation
level), while the crest length is 580 m (PPC S.A.,
1981).
3 Available Data
The Pournari I dam has been monitored since 1981.
The geodetic monitoring system consists of 79
control stations in total located on the crest, the
downstream and upstream shoulders, right and left
abutments, the spillway right bank and the power
station area. The vertical and horizontal deformations
are measured using reference stations that are located
on stable ground close to the right and left abutments.
The geodetic monitoring record was provided by
the PPC S.A. and covers a period of 29 years, from
February 1981 to April 2010. The record consists of
(1) the horizontal deflections and (2) the vertical
deformations of all control points on the dam crest
and shoulders and (3) daily values of reservoir level
fluctuations. During the period of the first filling of
the reservoir (between 1981 and 1984) the sampling
rate was every 20 days, while from 1984 onwards this
rate was reduced to once per year.
This study is focused on the analysis of the control
points S1-S14 located on the crest (see Figure 2) as
this part of the dam is the most susceptible to
deformations.
Fig. 2 Location of control points located on the crest of
Pournari I dam (the dataset from S4 was not available).
(a)
(b)

3rd Joint International Symposium on Deformation Monitoring Vienna, 30th March–1st April 2016
3
4 Data Analysis and Results
4.1 Crest Settlement
Figure 3 presents the crest settlements of Pournari I
dam and the reservoir level fluctuations for the
period of 29 years. The reservoir level remained
within the operational limits set for hydroelectric
power production (i.e. between 100 m and 120 m,
PPC S.A., 1981) reaching maximum and minimum
values of 119.9 m and 100.6 m on August 2008 and
January 1987, respectively.
Fig. 3 Evolution of the settlements of all crest control points
of Pournari I dam and reservoir level fluctuations. The
upstream side of the crest appears to have the maximum
observed settlements. The rate of settlements became more
gradual after 1988 when the reservoir drawdown level
remained relatively higher compared to the period 1981-1987.
The settlements of the crest were developed at a
higher rate during the first years of the dam operation
which was anticipated. Significant settlements that
reached maximum values were also recorded on the
upstream side of the crest compared to the
downstream crest control points. In particular, the
maximum settlement was recorded at the control
point S9, which is located at the cross section with
the maximum dam height, on April 2010 with a
cumulative value of 609 mm (see Figure 3). This
value corresponds to 0.57% of the dam height.
The crest settlements were also evaluated using
the settlement index (SI) given by equation 1
(Charles, 1986):
(1)
where s (mm) is the crest settlement between two
different measurement periods t1 and t2 at each
control point located at a height H (m) from the
foundation level. Tedd et al. (1997) suggested that if
the value of the dimensionless parameter SI is greater
than 0.02, the crest settlement is attributed to
mechanisms other than creep requiring further
investigation to be conducted. Figure 4 presents the
SI obtained for all the crest control points of Pournari
I dam for the period of 29 years. From the
construction of the dam (1981) until the last available
measurement (2010), the settlement index SI has
never exceeded the critical value of 0.02 which
indicates that the deformations at the body of the dam
can be attributed to normal anticipated mechanisms
(i.e. creep).
Fig. 4 The settlement index of the crest control points of
Pournari I dam has never exceeded the threshold value of 0.02
from 1981 until 2010.
The annual rate of settlement (Sa) was also used to
assess the deformations of the crest with equation 2:
(2)
where Sii and Si are consecutive yearly settlement
measurements and H is the height from the
foundation level at each crest control point. Any Sa
settlements equal or less than 0.02% of the height of
the dam were regarded normal and the dam stabilised










1
2
log1000 t
t
H
s
SI
100






H
SS
Siii
a

3rd Joint International Symposium on Deformation Monitoring Vienna, 30th March–1st April 2016
4
(Dascal, 1987). As presented in Figure 5, the annual
rate of settlement for all the crest control points of
Pournari I dam is stabilised below the value of 0.02%
of the dam height almost 8 years after the first
impoundment (1981).
Fig. 5 The annual rate of settlement of the crest control points
from 1981 until 2010 was stabilised below the value of 0.02%
in 1989, almost 8 years after the first reservoir filling.
4.2 Crest Horizontal Deformations
The horizontal deformations of the crest control
points of Pournari I dam is presented in Figure 6.
Fig. 6 Horizontal deformations of the crest control points of
Pournari I dam from 1981 until 2010. Positive values indicate
deformations towards the downstream direction.
Both the upstream and downstream crest control
points were displaced in the horizontal direction
towards the upstream side of the dam. The rate of
deformations reached maximum values during the
first years of the dam operation which corresponds
to the first reservoir filling and it is gradually
decreased after the year 1986. The maximum
horizontal deformation was recorded for the control
point S10 with a value of 139 mm (see Figure 6). The
S10 is located at the cross section where maximum
dam height is recorded.
5 Discussion and Conclusions
This study presented the post-construction analysis of
the behaviour of one of the largest earth fill dams in
Greece based on long-term monitoring data. The 29
year long dataset consisted of geodetic measurements
of vertical and horizontal deformations from the crest
of Pournari I dam and the reservoir level fluctuations.
The maximum crest settlement (609 mm) was
recorded for the control point located at the cross
section with the maximum dam height. Similar
behaviour was recorded for the Kremasta dam as the
maximum vertical deformation (764 mm) was
recorded at the maximum height of the dam
(Pytharouli and Stiros, 2009).
The settlement index (SI) was also used to assess
the crest deformations of Pournari I dam. The SI has
never exceeded the critical value of 0.02 from the
construction of the dam (1981) until the last available
measurement (2010), which indicates that the dam
deformations can be attributed to normal anticipated
mechanisms (i.e. creep, secondary consolidation of
the clay core). The same does not apply for the SI of
Kremasta dam which has exceeded the value of 0.02
four periods of time reflecting effects other normal
creep (Pytharouli and Stiros, 2009).
The annual rate of settlement of the crest of
Pournari I dam was stabilized below the value of
0.02% almost 8 years after the first impoundment of
the reservoir. This is not in agreement with the
estimation proposed by Dascal (1987) that suggested
that the maximum period for the stabilisation of crest
settlements varies from 24 to 30 months.
The analysis of the dataset also revealed that the
crest deformed towards the upstream direction for the
examined period of time which is considered a rather
unusual behaviour. Horizontal deformations towards
the upstream direction are common for earth-fill
dams for the time period after construction and
during the first stages of the first impoundment of the
reservoir. For the case of Pournari I dam, the
direction of the horizontal deformations of the crest
for the majority of the control points did not change

3rd Joint International Symposium on Deformation Monitoring Vienna, 30th March–1st April 2016
5
over time. It should be noted that this is not the case
for the horizontal deformations of the control points
located on the downstream face of the dam, below
the crest level. The upstream movement of the crest
is potentially originated by the hydrostatic pressure
from the reservoir causing the submerged part of the
dam to deform towards the downstream direction but
at the same time resulting in an upstream movement
of the crest.
The results of this study indicate that the long term
behaviour of the crest of Pournari I dam is normal
and within anticipated limits. However, it also
highlights that the stabilization period of the rate of
deformations is slightly different compared to the
expected from the published literature. This does not
imply that the dam’s structural integrity is at risk. It
rather highlights the complexity of the deformation
mechanisms for earth-fill dams and the fact that
dams of similar size and type can behave differently
but still operate within safety limits.
Future research entails the investigation of the
deformations of the crest and the body of the dam for
the period that corresponds to the first reservoir
filling. This will enable to quantify the factors
affecting the deformations of the dam and will
provide an insight of its behaviour.
Acknowledgments: Authors would like to thank the
Public Power Corporation (PPC) of Greece S.A., A.
Kountouris and T. Sioutis (PPC S.A.) for providing
the monitoring record and their assistance. Special
thanks to the Management and personnel of the
Hydroelectric Station of Pournari I dam.
References
Alonso, E.E., S. Olivella and N.M. Pinyol (2005). A review
of Beliche Dam. Geotechnique, 55(4), pp. 267-285.
Baecher, G.B. and J.T. Christian (2000). The practice of risk
analysis and the safety of dams. Available at:
http://oldpm.umd.edu/files/public/water_library/2000/Con
ference_The%20Practice%20of%20Risk%20Analysis%2
0and%20Safety%20of%20Dams_Cairo_2000_Baecher.do
c, [Accessed 12 January 2016].
Boros-Meinike, D. and W. Jankowski (2003). A monitoring-
based assessment of the ageing process of selected polish
dams. In: Proc. of 21st Congress on Large Dams, Montreal,
ICOLD, Q.82-R.
Charles, J.A. (1986). The significance of problems and
remedial works at British earth dams. In: Proc. of
BNCOLD-IWES Conference, Reservoirs, pp. 123-141
Dascal, O. (1987). Post-construction deformation of rockfill
dams. Journal of Geotechnical Engineering, 113(1), pp. 46-
59.
Guler, G., H. Kilic, G. Hosbas and K. Ozaydin (2006).
Evaluation of the Movements of the Dam Embankments by
means of Geodetic and Geotechnical Methods. Journal of
Surveying Engineering, 132(1), pp. 31-39.
Heck, B. (1984). Monitoring dam deformations by means of
geodetic control networks. In: Proc. of the International
Conference on Safety of Dams, Coimbra, Portugal, 23-28
April, pp. 455-466.
Kalkan, Y., M.R. Alkan and S. Bilgi (2010). Deformation
Monitoring Studies at Atatürk Dam. In: FIG Congress 2010,
Facing the Challenges – Building the Capacity, Sydney,
Australia, 11-16 April 2010. Available at:
http://www.fig.net/pub/fig2010/papers/fs01d\fs01d_kalkan
_alkan_et_al_4466.pdf, [Accessed 20 May 2010].
Kyrou, K., A. Penman and C. Artemis (2005). The first 30
years of Lefkara Dam. In: Proc. of the ICE - Geotechnical
Engineering, 158(2), pp. 113 –122.
Public Power Corporation (PPC) S.A. (1981). Brochure:
Hydroelectric project Pournari. Available at:
http://www.itia.ntua.gr/~nikos/arx_int/CDfrag/scanarismen
a/POURNARI/Untitled.pdf, [Accessed 12 January 2016].
Popovici, M., A.M. Abdulamit, C.I. Ilinca and M.M. Malai
(2004). Practical strategies for Surveillance at Romanian
Dams. Hydro Review Worldwide, pp. 32 – 36.
Pytharouli, S. and S. Stiros (2009). Investigation of the
parameters controlling the crest settlement of a major
earthfill dam based on the threshold correlation analysis.
Journal of Applied Geodesy, 3(1), pp. 55-62.
Szostak-Chrzanowski, A. and M. Massiera (2006). Relation
between monitoring and design aspects of large earth dams.
In: Proc. of 3rd IAG/12th FIG Int. Symposium on
Deformation Mesurements, Baden, 22-24 May. Available
at:http://www.fig.net/commission6/baden_2006/PDF/MD
A/Szostak-Chrzanowski.pdf, [Accessed 20 May 2010].
Tedd, P., J.A. Charles, I.R. Holton and A.C. Robertshaw
(1997). The effect of reservoir drawdown and long-term
consolidation on the deformation of old embankment dams.
Geotechnique, 47(1), pp. 33-48.
United States Society on Dams (USSD) (2008). Why Include
Instrumentation in Dam Monitoring Programs?. Available
at:http://www.ussdams.org/instrumentation.PDF,
[Accessed 12 January 2016].