KSCE Journal of Civil Engineering (2011) 15(2):317-325 Structural Engineering

DOI 10.1007/s12205-011-0918-1
www.springer.com/12205

A Study On Dam Instrumentation Retrofitting

A. Melih Yanmaz* and Onur Ari**
Received July 1, 2009/Accepted May 11, 2010

···································································································································································································································

Abstract

Continuous monitoring and inspection of large dams would provide means for proper rehabilitative actions to be taken on time
which intend to maintain the desired safety throughout the physical life of dams. Comprehensive inspections can be carried out using
a number of instruments. An ideal dam instrumentation system should provide time-dependent information about critical parameters
so that possible future behaviour of the structure can be predicted. Numerous of the existing dams do not have adequate instruments
or current instrumentation systems may not be in good condition. By implementing the modern equipment to existing dams, the
uncertainty associated with the impacts of aging or unexpected severe external events will be reduced and possible remedial
measures can be taken accordingly. This paper addresses the concept of instrument retrofitting to an unmonitored structure,
Gokcekaya Dam, 159 m-high arch dam, Turkey, with reference to investigation of the current condition of the structure. The
deficiencies observed are listed and the corresponding rehabilitative measures are suggested. Finally, different alternatives of a new
instrumentation system are introduced and compared in monetary terms.
Keywords: dams, safety, instrumentation, rehabilitation, monitoring
···································································································································································································································

1. Introduction use of new types of instruments possible, which have greater

Many dams were designed and constructed during the last
century using conventional procedures (DeMichele et al., 2005).
Therefore, the adequacy of such dams with respect to current
conditions needs to be checked. On the contrary, new dams are
designed using sophisticated approaches, which are normally
based on some simplifying assumptions. Validity of design as-
sumptions and instantaneous status of safety of new dams can be
assessed using some instruments installed in and around dams.
Unexpected behaviours may be observed throughout the physi-
cal life of the structure due to the inability to model uncertainties
associated with extreme events, foundation conditions, material
properties as well as deficiencies in construction practices. With
sufficient data and experience gained from the behaviour of the
existing structure, challenging future designs can then be per-
formed.
Early dam examination practices consider only the visual in-
spection, done by walking on dam body and checking the struc-
tural integrity visually from inside of the galleries. Although this
method with an experienced examiner gives valuable informa-
tion about the physical condition of dam body, more complex
items, such as seepage, movement, and state of internal stresses
cannot be determined without using proper equipment. With the
advancement in dam monitoring, a number of instruments have
been developed and used. Recent electronic technology made the
accuracy, and longer life with smaller dimensions. Moreover,
new instruments are equipped with their automated data loggers,
which can lead to computer assisted evaluation of long-term be-
haviour of the structure. Possible deficiencies should be moni-
tored either by direct or indirect measuring techniques. Charac-
teristics of items to be monitored should then be analyzed for
determining the remedial actions to be taken. The extent and
degree of monitoring is decided during the design phase of a dam
with reference to the critical items that are likely to cause a hazard.
The type of dam, site characteristics, and the purpose of the
project should also be considered. Time-dependent probabilistic
safety analyses should be carried out in order to ensure the desir-
ed strength throughout the lifetime of dams. The critical items
should then be monitored constantly using the most appropriate
equipment in view of the performance of monitoring and economy.
Although the required monitoring equipment would normally
differ from dam to dam, pore water pressure and seepage, body
and foundation movement, mass temperature, and seismic effects
are mostly the parameters to be measured. Automated systems
offset human-induced errors in measurements and monitor the
equipment continuously. They collect relevant inputs precisely
and trigger possible warning messages.
As a dam gets aged, several deficiencies may occur i.e., con-
crete body deteriorates, drains may be clogged, grouts and cut-
off walls lose their effectiveness, contraction joints separate due

*Professor, Civil Engineering Department, Middle East Technical University, Ankara 06531, Turkey (Corresponding Author, E-mail: myanz@metu.edu.tr)
**Research Assistant, Civil Engineering Department, Middle East Technical University, Ankara 06531, Turkey (E-mail: oari@metu.edu.tr)

− 317 −
 A. Melih Yanmaz and Onur Ari
to cyclic loading, and old instrumentation system may become
out of order. These adverse effects of aging bring the dam to a
more vulnerable condition against failure. In order to protect the
downstream, old dams should be rehabilitated. As a first step, the
current condition of the dam should be assessed properly. Fur-
thermore, a number of instruments can be retrofitted to measure
the critical parameters. Although retrofitted instruments cannot
give information about the events that have occurred before the
installations, they provide a basis for assessing possible future
events. In the days of the design period of an existing dam,
structural and geotechnical design techniques might have been
incapable of precise modelling. In order to ensure the validity of
assumptions made and the adequacy of the design, relevant data
may be obtained from the retrofitted instruments. The aim of this
paper is to illustrate the application of dam instrumentation re-
trofitting to an existing dam, Gokcekaya Dam, Turkey. Various
alternatives have been generated in terms of monetary units.
2. Literature Survey
This section introduces some case studies on dam instrument-
ation retrofitting practices. The facing of the Morávka Dam in
Czech Republic has been rehabilitated, drainage system has been
improved and the monitoring system has been upgraded by auto-
matic monitoring and data transfer (Kratochvíl and Glac, 2006).
Surface monitoring system and pendulum have been automated
for Talvacchia Dam in Italy and additional equipment, monitor-
ing dynamic and seismic responses, have been installed (ICOLD,
1992). Dam leakage and surface movement of Baitone Dam
(Italy) have been monitored periodically. The existing monitor-
ing system has been supplemented with one inverted pendulum,
three piezometers, and an automated data acquisition system
(ICOLD, 1992). Marunuma Dam in Japan experienced severe
cracking due to freezing and thawing. Increased leakage has
been observed from the joints. The dam has been equipped with
strain and joint monitoring instruments. Based on detailed
inspections, the deck has been covered with a new concrete slab.
New stress, strain and joint movement monitoring equipment
have been implemented (ICOLD, 1992). Kohmyo-Ike Dam
(Japan) was subject to extensive seepage during the period of 60
years. After intensive analyses, the upstream face has been
covered with silty clay and new piezometers and weirs have been
installed. The measurements verified decrease in seepage after
this rehabilitation (ICOLD, 1992). Upper Huia Dam in New
Zealand had uplifting problems. New piezometers were installed
to constantly monitor the pore water pressures beneath the dam.
By using the automated system, the measured pore water pres-
sure data gave invaluable information about the effectiveness of
the cut-off wall and the necessary remedial actions were then
implemented on that basis (Ahmed-Zeki et al., 2000). A new
monitoring system has been designed for Compuerto Dam (Spain)
with some rehabilitative actions. Grout curtains have been
extended and all existing drains have been cleaned and bored.
Weirs have been installed to monitor the seepage rate. One of the
− 318 −
ten weirs has also been equipped with an automatic gauge. Each
monolith has been instrumented with two piezometers. Three
direct and three inverted pendulums have been installed. Two
extensometers have also been introduced. In order to monitor the
rotation of the monoliths, seven clinometer bases and one clin-
ometer have been added to the system. For monitoring contrac-
tion and expansion of the joints, 14 thermometers have been
installed to one of the monoliths. Finally, an automatic gauge has
been installed for reservoir level monitoring (ICOLD, 1992).
The initial instrumentation of Chandreja Dam (Spain) was set for
measurement of drainage flow and uplift monitoring. The new
system has been proposed for further monitoring of uplift, sur-
face movement, and joint movement. An array of piezometers has
been installed under the buttresses. Angular collimation targets
have been introduced for horizontal movement monitoring and
an array of jointmeters has been implemented for monitoring the
relative movements of joints (ICOLD, 1992). Seeuferegg Dam
(Switzerland) is located in the Alps. Existing monitoring system
consisting of a direct pendulum and seepage rate monitoring
weirs was supplemented with additional direct and inverted
pendulums. A new geodetic network has been provided for sur-
face monitoring. A number of standpipe piezometers have also
been installed for uplift pressure monitoring (ICOLD, 1992).
After 1971 San Fernando Earthquake, vertical contraction joints
of Pacoima Dam (USA) have been opened resulting in horizontal
cracks. Due to the movement of plates, a slight rotation of the
dam body has also been noticed. No instrumentation had been
provided for the dam initially. After rehabilitation, extensometers
have been installed in abutments. A total of 20 piezometers were
provided to monitor the uplift pressures. An inclinometer, an array
of accelerometers, and a number of thermometers have also been
installed at various locations on dam body (ICOLD, 1992).
3. Case Study
3.1. General
Gokcekaya Dam, located on Sakarya River, 60 km northeast of
Eskisehir province, Turkey, generates electricity. It is an arch
dam, which was constructed during the period 1967-1972. A
descriptive sketch of it is shown in Fig. 1. The dam is 159 m high
from the foundation, 115 m high from the thalweg and it has a
crest length of 479.7 m (including thrust blocks). Its crest width
is 8 m. A concrete volume of 650,000 m3 was used in the
construction of the body. The reservoir lake has a surface area of
20 km2 with maximum volume of 910 million m3. The hydro-
electric power plant has three units with an installed capacity of
278.4 MW. The spillway of Gokcekaya Dam, 62 m long, has
been constructed in a neighbouring valley as a separate structure
on the right side of the dam. Three radial gates are supported by
two intermediary piers. The energy dissipater of the spillway is
of a deflector bucket type. Detailed information on basin hydro-
logy and site characteristics can be found in Ari (2008). There
were two important fault lines at the dam site, which directly
affected the body construction. Treatments of these faults were
KSCE Journal of Civil Engineering
 A Study On Dam Instrumentation Retrofitting
Fig. 1. Characteristic Sections of Gokcekaya Dam (Not to scale)
completed by removing the debris up to a predefined level and
then filling the fault by concrete (DSI, 1974).
The authors visited Gokcekaya Dam site to be informed on the
current conditions. Several deficiencies observed throughout the
site visit are summarized below.
a) Traces of leaked water from the contraction joints, some
calcium deposits, and some diagonal cracks and openings were
noticed on the walls and ceilings of the galleries. Leakage from
contraction joints can be evaluated as a sign of joint movement.
b) A relatively large leakage has been noticed along one of the
contraction joints on the downstream face of the dam at the same
axis as one of the faults.
c) At the entrance of the gallery from the power plant access, a
crack was observed on the base of the gallery, perpendicular to
the dam axis, which may result from the differential settlement
due to foundation movement.
d) On the right abutment around the right thrust block, eleven
drain wells exist, which have been used as standpipe piezo-
meters. One of them is out of order due to a subsidence, whereas
the other three had no water, possibly due to clogging.
e) The chute and the deflector bucket sections of the spillway
were heavily covered by plants, which would significantly decre-
ase the capacity of the spillway during operation. The local
authorities stated that the gates have not been opened since 1982.
It is, now, suspicious that whether or not the gates can operate
properly during a possible subsequent flood.
3.2. Rehabilitative Recommendations
Some remedial actions for dam rehabilitation are introduced in
Vol. 15, No. 2 / February 2011 − 319 −
this section. The remaining economical life of Gokcekaya Dam
is roughly estimated to be around 40 years (Ari, 2008). Due to
lack of routine inspection and necessary repair works, the
expected safety performance of Gokcekaya Dam was assumed
to be reduced. So it is essential to perform some rehabilitative
works. To this end, monitoring the performance of the structure
by modern retrofitted instruments is of utmost importance. In
order to be consistent on rehabilitative and instrumental recom-
mendations, not only the present situation, but also some impor-
tant past events should be taken into account. The dam has
experienced a major problem of leakage of some mud flow from
the sluiceway drainage system in August 1983. It clouded almost
one-third of the downstream reach to brownish colour. One week
after this event, a new mud flow had been started from various
drain holes in the floor of the main gallery. The situation of the
dam was inspected by a team who observed mud flow coming
from some drain holes, mud residue on the stairs of the inspec-
tion gallery, and opening of some contraction joints around drain
holes. The drain canals were observed to be clogged and no
measurement for seepage was available. Furthermore, they also
noticed that stress-strain monitoring equipment were out of order
(Tanriverdi et al., 1983). The mud observed in galleries and in
drain holes was of clay origin, containing some organic materials
with light-brown colour. Physical and chemical studies on both
the mud found in drain holes and the material taken from the
reservoir bottom showed identical properties. So the origin of the
mud was assumed to belong to the reservoir bottom. Because of
lack of maintenance, the residues of the mud are still observable
in drain holes in the gallery, which may lead to pore water pres-
sure built-up by increasing the uplift pressure at the foundation
level.
The dam has 17 weirs which are located in a sectioned manner
in order to get an idea in case of an abnormal behaviour. However,
most of them are inoperable. Therefore, they should be renewed.
A number of standpipe piezometers mounted around the right
thrust block are assumed to monitor seepage due to the right fault
line. There exist eleven piezometers in that group. In addition to
those on the right thrust block, five standpipe piezometers also
exist around the spillway. After detailed studies on the past
readings of these standpipe piezometers, four wells at the left
thrust block have been selected to be retrofitted with reference to
their location and condition. In Gokcekaya Dam, many joints
experienced significant leakage due to the relative movement
between blocks. The condition of joints should then be inspected
and repaired urgently. Jointmeters had been embedded to the
concrete mass during construction phase but almost all of them
are out of order. Main problems observed at the dam site and the
corresponding rehabilitative actions to be taken have been
summarized in Table 1.
3.3. Instrumentation Retrofitting
Instrument application design of a dam requires extensive
structural, hydraulic and geotechnical analysis, prediction of
weak zones, determination of best-suited instruments, cost
 A. Melih Yanmaz and Onur Ari

Table 1. Rehabilitation Program Table 2. Unit Prices of Instruments (IIC, 2007)

Problems Unit price per item
Rehabilitative action required Equipment Type
identified at (€ )
• Cleaning channels and weirs VW Piezometer 400.00

Drain holes
• Re-drilling of drain holes, if necessary V-Notch Weir and Automatic VW Readout Unit 2,145.00
and weirs • Possible additional foundation grouting based on Automatic VW Weir Readout Unit 1,915.00
chemical analysis of the materials removed from
drain holes Parshall Flume and Automatic VW Readout Unit 2,350.00

• Cleaning the cracks and expanded joints Surface-mount VW Jointmeter 550.00

• Filling cracks and joints with urethane sealant for Accelerometer 11,253.00
Cracks and
watertightness Surface-mount Clinometer (Tiltmeter) 895.00
joints
• Filling cracks and joints with epoxy resin injection Extensometer 980.00
for structural bonding
Instrument Cable 4.90 /m
• Removing the accumulated debris and plants from
spillway face Cable Carrier Tray 26.00 /m
Spillway and
gate mechanism
• Repairing concrete damages Multiplexer 4,425.00
• Inspecting and repairing spillway gates and operat- Data Logger 15,104.00
ing mechanism
Software 16,300.00
Common Triggering Unit and Data Recorder for
23,788.00
optimization, and experience. For existing dams, however, some accelerometers
of the relevant information may be missing as in the case of Handheld Data Readout 5,627.00
Gokcekaya Dam. It has been believed that after some time, the
items, such as concrete temperature, are stable in terms of
measured values and further monitoring may not be required. achieved by considering several items, such as types of instru-
Recent studies showed that aging of dams results in more ments, installation locations, the cable lengths to data logger
damage than expected. In order to monitor and predict faulty from the individual instrument, required cable sag, and the length
elements and conduct repair works, it is necessary to install of the tray carrying cable. The multiplexers and data loggers are
modern and precise equipment to an old dam. In the following located at the entrance of the gallery from the hydroelectric
sections, the alternatives of instrument retrofitting will be power plant for access ease. Furthermore, all length calculations
discussed for Gokcekaya Dam. The proposed basic instruments have been completed accordingly. The required sag of the cables
are piezometers, automated weirs, surface jointmeters, surface has been chosen as 15% of the total length of cable (USBR
clinometers, borehole extensometers and strong motion accelero- 1987). The complete price list of all proposed instruments is
meters. All equipment is to be planned considering a central given in European currency unit (EUR) with reference to IIC
automated data acquisition system in order to obtain precise (2007) in Table 2.
results, free of human induced errors. The calculations have been After computing the details and the quantity of the individual

Table 3. Cost Computations

Monitoring item Option Ci (€ ) Cc (€ ) Cca (€ ) Ct (€ )
1 14,000.00 32,848.11 13,925.60 60,773.71
Piezometer 2 12,400.00 24,821.61 11,390.60 48,612.21
3 6,400.00 13,081.65 11,390.60 30,872.25
1 25,850.00 8,739.32 16,265.60 50,854.92
Seepage
2 21,065.00 8,739.32 16,265.60 46,069.92
1 35,750.00 49,595.33 23,821.20 109,166.53
Joint movement
2 18,700.00 27,363.00 12,430.60 58,493.60
1 17,005.00 18,629.31 17,758.00 53,392.31
Rotation (tilt)
2 2,685.00 2,163.84 7,397.00 12,245.84
Foundation movement - 13,720.00 12,069.61 5,570.24 31,359.85
1 57,547.00 4,194.13 11,499.80 73,240.93
Earthquake acceleration
2 35,041.00 1,052.62 4,856.80 40,950.42

− 320 − KSCE Journal of Civil Engineering

 A Study On Dam Instrumentation Retrofitting
instrument items, cabling, cable tray and instrument costs have
been determined and total costs of each option are attained by
summing them up as shown in Table 3 in which Ci, Cc, Cca, and
Ct stand for cost of instrument, cost of cable, cost of carrier, and
total cost, respectively.
3.3.1. Piezometer Options
In Gokcekaya Dam, three options for piezometer applications
are proposed. Although uplift is less critical in arch dams,
Gokcekaya Dam has been observed to have some problems
related with the seepage and uplift reducing facilities. In order to
monitor possible deficiencies, installation of a new pore water
pressure monitoring system would be beneficial. A definition
sketch is presented in Fig. 2 to demonstrate the instrumentation
configuration for piezometers. In Option 1, the total number of
required vibrating wire piezometers are determined as 35. On the
right thrust block, TR-2 and TR-3 (See Fig. 2) are selected and
one of their drain holes is considered to be utilized for piezo-
meter installation. Similarly, TL-2 to TL-5, on the left thrust
block, are chosen to be monitored. Since this option concentrates
on the amount of data gathered, all cantilevers from 0 to 24 are
equipped with piezometers to monitor the change of pore water
pressure between the sequential blocks. In this option, the exist-
ing standpipe piezometer holes on the right abutment are also
evaluated.
By considering their condition and location, wells 149, 147,
145 and 143 are added to the monitoring system by installing
piezometers. In the computations, the average depth, where
piezometers are embedded, is taken as 10 m. The possible costs
of drilling and backfilling are ignored. In Option 2, the existing
right abutment drain holes are excluded because of the consistent
results of manual well reading up to date and the closeness of the
piezometers on blocks, TR-2, TR-3, and cantilever 0. The other
installations in the dam body are taken as the same as Option 1.
The total number of instruments in this option is 31. The Option
3 is suggested as a lower-cost alternative to Options 1 and 2. In
Fig. 2. Definition Sketch for Piezometer Arrangement
Vol. 15, No. 2 / February 2011 − 321 −
Option 3, the piezometers on the existing right drain holes are
discarded as in Option 2. Moreover, unlike the previous options,
only one piezometer for two cantilevers is utilized. So the canti-
levers with even numbers are equipped with piezometers. These
cost-cutting actions reduce the total number of instruments to 16.
The costs of three piezometer options in Euros are compared
according to their total costs in Table 3.
3.3.2. Seepage Monitoring Options
After evaluating past reading logs for seepage measurements,
it is clear that neither the monitoring has been properly scheduled
nor the readings had the required precision. By evaluating the
condition of the existing weirs, those that need to be instru-
mented have been determined. For seepage measurement, two
options are suggested. In Option 1, the existing weirs are proposed
to be replaced by Parshall flumes because of less maintenance
and smaller headlosses. The selected Parshall flumes in this
option should also be automated by vibrating wire weir gauges in
order to fully adapt to automated data logging system. They are
planned to be installed to the place of the existing weirs. So the
weirs should be cut from the base of the canal to install Parshall
flumes. Like the piezometers, cables have been routed to the
ceiling of the gallery and then carried via cable carrier trays to
the multiplexer units. Option 2 suggests repairing and cleaning
of the existing weirs. Afterwards, by using the vibrating wire
weir gauges, the existing weirs can be equipped and monitored
continuously by data logger device. For cable routing, similar
approach as that of Option 1 is applied. Table 3 gives the costs of
two seepage monitoring options.
3.3.3. Joint Movement Monitoring Options
Gokcekaya Dam consists of 25 cantilevers and 10 thrust
blocks. During the construction, several embedded jointmeters
had been installed along the contraction joints. However, almost
all of them are out of order and past readouts are not available.
Furthermore, the readout ports have been out of order due to
 A. Melih Yanmaz and Onur Ari
heavy corrosion. In addition to those, a horizontal joint opening
and leakage has been observed on cantilever 1 and several
leakages on vertical joints inside the gallery have been observed.
After extensive studies on dam projects, some repair works and
subsequent instrument retrofitting are recommended. Two options
for instrument retrofitting are suggested. In Option 1, every joint
along the inspection gallery is equipped with two surface-mount
jointmeters. One jointmeter is proposed for the downstream side,
whereas the other one is for the upstream side of the same joint
as seen in Fig. 3.
The reason behind this layout is that the hydraulic loading on
an arch dam creates either tension or compression on the
different zones of the shell and two jointmeters on both faces
help to monitor the movement precisely. Moreover, after the
repair works on the leaking joint on cantilever 1, the downstream
face of the horizontal joint is suggested to be instrumented with
surface-mount jointmeters in order to verify the performance of
the repair works done. It is computed that 65 surface-mount
jointmeters are required in this option. Cables are conveyed
along the ceiling of the gallery by cable trays and a total length of
10.12 km of cable is required. Option 2 suggests a more econo-
mical layout than Option 1 by proper placement of 34 surface-
mount jointmeters. In this option, a contraction joint is equipped
with only one surface-mount jointmeter on either downstream or
upstream face. If a joint is monitored on its downstream face, the
neighbouring joints are instrumented on the upstream face as
shown in Fig. 3. This layout has created an alternating plan and
thus reduced the required number of instruments. As a result of
having fewer instruments, the total length of cables is reduced to
5.59 km. Table 3 presents the costs of two joint movement
monitoring options.
3.3.4. Rotation (Tilt) Monitoring Options
In Gokcekaya Dam, no pendulums had been installed during
the construction. Furthermore, no vertical shafts had been pro-
vided. Installation of surface-mount clinometers is planned for
rotation monitoring. Two options of surface-mount clinometers
Fig. 3. Definition Sketch for Jointmeter Arrangement (A part of the
gallery)
− 322 −
are proposed for this dam with reference to the gallery layout.
Option 1 recommends installing surface-mount clinometers over
the three sections inside the gallery and onto the downstream
face. The dam body is first divided into two and the location of
the first group of clinometers is determined as cantilever 11. In
this central section, 9 clinometers are proposed with a vertical
spacing of 20 m. The other sections are selected as cantilevers 4
and 21. Each section is equipped with 5 clinometers. Option 2
suggests installing only three clinometers inside the galleries of
above sections in order to reduce the cabling and equipment
costs. It is obvious that the data gathered from this option are
limited. Table 3 gives costs of two rotation monitoring options.
3.3.5. Earthquake Acceleration Monitoring Options
Gokcekaya Dam is placed in the third earthquake zone on the
Ankara - Eskisehir province border. However, the dam site is
only 30 km far from the first earthquake zone and 80 km from
Duzce, the epicentre of the 12 November 1999 Earthquake with
a magnitude of 7.2. Gupta (2002) investigated possible sources
of earthquakes in dam environments. His results covering more
than 90 earthquakes indicated that these earthquakes were
triggered by reservoirs with the recorded largest event at Koyna
Dam reservoir, India, in 1967 having a magnitude of 6.3. Gupta
(2002) also stated that the depth and the volume of the reservoir
are the most important factors in reservoir induced seismicity.
One of the recent studies implied that reservoir induced earthquakes
have been occurring over 44 years in Koyna region (Gupta et al.,
2007). Considering the left and right faults at the foundation and
closeness to the North Anatolian Fault Line, Gokcekaya Dam
site can be considered as a seismically active region. Moreover,
the depth and the volume of the reservoir dictate that this dam is
prone to reservoir induced seismicity.
Earthquake performance and behaviour under seismic loading
can only be evaluated with proper measurement of earthquake
acceleration. That is why Gokcekaya Dam needs to be equipped
with strong motion accelerometers. Strong motion accelerometers
have not been monitored by automated data loggers like the
other instruments. So they require a common triggering unit and
a recorder. Common triggering unit activates the strong motion
accelerometer at a time, in case it senses any earthquake (Sezgin,
2008). In Gokcekaya Dam, two options are suggested for installing
strong motion accelerometers. In Option 1, three accelerometers
are recommended. Two accelerometers are proposed to be placed
on the dam body and the remaining one to the left abutment as a
free-field accelerometer. A typical distance for free-field accelero-
meter in dam applications is defined as twice the height of
concrete dams and half of this value if the modulus of elasticity
of foundation is equal to or greater than the modulus of elasticity
of the dam concrete (Darbre, 1995). The first accelerometer is
placed to the dam-foundation interface at cantilever 10, whereas
the second one is placed on its crest. The third accelerometer is
installed as a free-field accelerometer, 300 m from the left
abutment. Option 2 is suggested as consisting one accelerometer
which is placed on the crest of cantilever 10. No matter which
KSCE Journal of Civil Engineering
 A Study On Dam Instrumentation Retrofitting

Table 4. Total Costs of Alternatives

Alternative Option Code ni nm Cic (€ ) Ca (€ ) Ct (€ )
1 P1 A1 W1 J1 T1 E 147 18 378,788.24 111,054.00 504,842.24
2 P1 A1 W1 J1 T2 E 131 16 337,641.77 102,204.00 454,845.77
3 P1 A1 W1 J2 T1 E 116 15 328,115.31 97,779.00 440,894.31
4 P1 A1 W1 J2 T2 E 100 13 286,968.84 88,929.00 390,897.84
5 P1 A1 W2 J1 T1 E 147 18 374,003.24 111,054.00 500,057.24
6 P1 A1 W2 J1 T2 E 131 16 332,856.77 102,204.00 450,060.77
7 P1 A1 W2 J2 T1 E 116 15 323,330.31 97,779.00 436,109.31
8 P1 A1 W2 J2 T2 E 100 13 282,183.84 88,929.00 386,112.84
9 P1 A2 W1 J1 T1 E 145 18 346,497.73 111,054.00 472,551.73
10 P1 A2 W1 J1 T2 E 129 16 305,351.26 102,204.00 422,555.26
11 P1 A2 W1 J2 T1 E 114 15 295,824.80 97,779.00 408,603.80
12 P1 A2 W1 J2 T2 E 98 13 254,678.33 88,929.00 358,607.33
13 P1 A2 W2 J1 T1 E 145 18 341,712.73 111,054.00 467,766.73
14 P1 A2 W2 J1 T2 E 129 16 300,566.26 102,204.00 417,770.26
15 P1 A2 W2 J2 T1 E 114 15 291,039.80 97,779.00 403,818.80
16 P1 A2 W2 J2 T2 E 98 13 249,893.33 88,929.00 353,822.33
17 P2 A1 W1 J1 T1 E 143 18 366,626.75 111,054.00 492,680.75
18 P2 A1 W1 J1 T2 E 127 16 325,480.28 102,204.00 442,684.28
19 P2 A1 W1 J2 T1 E 112 14 315,953.82 93,354.00 424,307.82
20 P2 A1 W1 J2 T2 E 96 12 274,807.35 84,504.00 374,311.35
21 P2 A1 W2 J1 T1 E 143 18 361,841.75 111,054.00 487,895.75
22 P2 A1 W2 J1 T2 E 127 16 320,695.28 102,204.00 437,899.28
23 P2 A1 W2 J2 T1 E 112 14 311,168.82 93,354.00 419,522.82
24 P2 A1 W2 J2 T2 E 96 12 270,022.35 84,504.00 369,526.35
25 P2 A2 W1 J1 T1 E 141 18 334,336.23 111,054.00 460,390.23
26 P2 A2 W1 J1 T2 E 125 16 293,189.76 102,204.00 410,393.76
27 P2 A2 W1 J2 T1 E 110 14 283,663.30 93,354.00 392,017.30
28 P2 A2 W1 J2 T2 E 94 12 242,516.83 84,504.00 342,020.83
29 P2 A2 W2 J1 T1 E 141 18 329,551.23 111,054.00 455,605.23
30 P2 A2 W2 J1 T2 E 125 16 288,404.76 102,204.00 405,608.76
31 P2 A2 W2 J2 T1 E 110 14 278,878.30 93,354.00 387,232.30
32 P2 A2 W2 J2 T2 E 94 12 237,731.83 84,504.00 337,235.83
33 P3 A1 W1 J1 T1 E 128 16 348,886.79 102,204.00 466,090.79
34 P3 A1 W1 J1 T2 E 112 14 307,740.32 93,354.00 416,094.32
35 P3 A1 W1 J2 T1 E 97 12 298,213.86 84,504.00 397,717.86
36 P3 A1 W1 J2 T2 E 81 10 257,067.39 75,654.00 347,721.39
37 P3 A1 W2 J1 T1 E 128 16 344,101.79 102,204.00 461,305.79
38 P3 A1 W2 J1 T2 E 112 14 302,955.32 93,354.00 411,309.32
39 P3 A1 W2 J2 T1 E 97 12 293,428.86 84,504.00 392,932.86
40 P3 A1 W2 J2 T2 E 81 10 252,282.39 75,654.00 342,936.39
41 P3 A2 W1 J1 T1 E 126 16 316,596.27 102,204.00 433,800.27
42 P3 A2 W1 J1 T2 E 110 14 275,449.80 93,354.00 383,803.80
43 P3 A2 W1 J2 T1 E 95 12 265,923.35 84,504.00 365,427.35
44 P3 A2 W1 J2 T2 E 79 10 224,776.88 75,654.00 315,430.88
45 P3 A2 W2 J1 T1 E 126 16 311,811.27 102,204.00 429,015.27
46 P3 A2 W2 J1 T2 E 110 14 270,664.80 93,354.00 379,018.80
47 P3 A2 W2 J2 T1 E 95 12 261,138.35 84,504.00 360,642.35
48 P3 A2 W2 J2 T2 E 79 10 219,991.88 75,654.00 310,645.88

Vol. 15, No. 2 / February 2011 − 323 −

 A. Melih Yanmaz and Onur Ari
option is selected, the required number of common triggering
units and data recorder is only one. In Table 3, the costs of two
earthquake acceleration monitoring options are presented.
3.3.6. Foundation Movement Monitoring
Foundation of Gokcekaya Dam had been equipped with Carlson
type foundation deformation meters at several places during the
construction. Despite of being a sound system for foundation
monitoring in the past, foundation deformation meters are inad-
equate compared to vibrating-wire extensometers. To this end,
replacement of existing foundation deformation meters with
borehole extensometers is proposed. In order to be consistent
with the dam design and to reduce the retrofitting costs by
elimination of extensive foundation drilling, existing boreholes
used for foundation deformation meters are planned to be
equipped with extensometers.
3.4. Evaluation of Alternatives
Table 4 presents possible 48 combinations with three options
for pore water pressure monitoring; two options for seepage
monitoring, joint movement monitoring, rotation monitoring,
earthquake acceleration monitoring, and finally one option for
foundation movement monitoring. In Table 4, ni and nm stand for
number of instruments and number of multiplexers, respectively,
whereas Cic and Ca are instrument and cable cost and automation
cost, respectively. Options are designated by two-character
codes. The first character defines the instrument group as P for
piezometers, W for seepage monitoring, J for joint movement
monitoring, T for rotation (tilt) monitoring, A for earthquake
acceleration monitoring, and E for foundation monitoring. The
second character identifies the aforementioned option numbers.
The total numbers of instruments in alternatives are also pre-
sented in Table 4. The highest and lowest numbers of instru-
ments are 147 and 79, respectively. Total number of instruments
is used to determine the required number of multiplexers, which
are used to increase the number of instruments to be monitored
by automated data acquisition system. For every 8 instruments
including temperature sensors, one multiplexer should be added
to the automated data acquisition system.
The aforementioned calculations made the computation of the
automation cost possible, which consists of the required number
of multiplexers, a data logger device, and software. The costs of
these items are also given in Table 1. After determination of
instrument and automation cost, total costs of alternatives are
computed and presented considering the combinations of
different options. A constant cost of 15,000 € has been added to
the alternatives in order to represent the cost of portable data
logger which should be available in all automated systems for
determining the possible faults and the cost of some additional
consumables, such as cable splicing kits. The highest total cost of
the proposed system is then determined as 504,842.24 € and the
most economical alternative has a cost of 310,645.88 € . Total
costs of all alternatives are presented in Table 4. Cost computa-
tions are carried out considering initial costs only. As the opera-
− 324 −
tion, maintenance, installation, and workmanship costs are almost
the same for all alternatives, these cost items are omitted.
A proper decision-making for the selection of a suitable instru-
mentation system is based on consideration of the achievement
to be expected in reduction of future risks as a result of imple-
mentation of the proposed instrument retrofitting and rehabilita-
tive actions to be taken. Therefore, the net benefit of such a
retrofitting project would be the difference between the failure
cost of the dam and the total cost of instrumentation and rehabili-
tative actions. The cost of failure of Gokcekaya Dam cannot be
determined since all relevant information covering the dam site
and downstream failure possibilities of all types of properties is
not available. However, it is quite clear that this cost is much
greater than the initial investment cost of such a dam, which was
52.9 million U.S. Dollars as of 1967 prices (EMO, 1967). This
cost is converted to 2008 prices (FTF, 2008) to obtain $288.71
million. The instrumentation cost is about 1% of the total project
cost (USBR, 1987), which comprises approximately $2.89 million.
When compared to the highest-cost alternative, this value is
more than 4 times the cost of retrofitted instruments to be
proposed for Gokcekaya Dam. Since instrumentation retrofitting
cost is much smaller than the initial investment cost and dam
failure cost, selection of a reasonable alternative may be
governed by the performance expectation from the monitoring
system. The alternative having the highest ability to monitor dam
performance may then be selected. To this end, it is recommend-
ed to choose either alternative 1 or 5 since they have maximum
number of instruments with the desired performance for moni-
toring the critical parameters.
3.5. Further Instrumentation
Recent developments in electronic technology lead to the
production of high accuracy GPS devices, especially in elevation
calculations. These advanced GPS units can be used on dams for
surface monitoring. Together with the body and foundation
movement monitoring, surface monitoring will give the behaviour
of the dam body precisely. In Gokcekaya Dam, the surface
monuments for conventional geodetic measurements, such as
triangulation, are in good condition but the lack of routine
measuring make these monuments obsolete. The initial cost of
installing such a GPS system for surface monitoring would be
high but the automatic data collection and processing abilities of
the system would remove the necessity for skilled operators.
Temperature monitoring is vital, especially during concrete
pouring. After some time, the variation of concrete temperature
stabilizes. Joint movements in concrete dams can also be correl-
ated with temperature changes. Therefore, the mass concrete
should be monitored throughout the life of the dam. In Gokcekaya
Dam, thermistors had been placed during construction in order to
monitor dehydration temperature. However, the present condi-
tions of these sensors are not known and a comprehensive survey
is needed to check their operability. Almost all types of modern
instruments are equipped with temperature sensors. However,
additional ones can be added to the proposed system. As a final
KSCE Journal of Civil Engineering
 A Study On Dam Instrumentation Retrofitting
recommendation, accurate reservoir water level measurement is
recommended with a vibrating-wire water level gauge.
4. Conclusions
Instrument retrofitting requires installation of modern equip-
ment to old dams in order to assess the destructive effects of
aging and thus take the necessary corrective actions on time. In
this paper, Gokcekaya Dam, Turkey, has been considered in a
case study to demonstrate instrument retrofitting. First, numerous
deficiencies on the dam body and appurtenant structures have
been identified. Then, these evidences were interpreted. After
evaluation of results, a number of rehabilitative repair works and
a modern dam monitoring system have been proposed with cost
analysis of different alternatives, which are formed by the com-
bination of various options for different types of instruments.
Cabling costs were found to be of the order of instrument costs.
The highest-cost belongs to Alternative 1 (504,842.24 €), whereas
Alternative 48 has the lowest cost (310,645.88 € ). It should be
noted that these alternatives consist only of the initial instrument-
ation costs. The costs of rehabilitative measures as well as repair,
maintenance, installation, and workmanship costs should also be
added to these values to determine the overall costs. Since the
cost of instrumentation retrofitting is normally much smaller
than the failure cost of a dam, the reduction in future risks as a
result of implementation of a retrofitting system would highlight
the benefits to be gained.
References
Ahmed-Zeki, A. S., Logan, N., McQuarrie, W., and Jaduram, R. (2000).
“Monitoring the cut-off wall performance of Upper Huia Dam.”
Massachusetts: Geo-Institute Specialty Conference, Amherst, pp.
212-227.
Ari, O. (2008). A study on dam instrumentation retrofitting: Gokcekaya
Dam, MSc Thesis, Middle East Technical University, Ankara,
Vol. 15, No. 2 / February 2011 − 325 −
Turkey.
Darbre, G. R. (1995). “Strong-motion instrumentation of dams.”
Earthquake Eng. and Structural Dynamics, Vol. 24, No. 8, pp. 1101-
1111.
De Michele, C., Salvadori, G., Canossi, M., Petaccia, A., and Rosso, R.
(2005). “Bivariate statistical approach to check adequacy of dam
spillway.” J. Hydr. Engrg., ASCE, Vol. 10, No. 1, pp. 50-57.
DSI (1974). Gokcekaya Dam with photos, State Hydraulic Works,
Ankara (in Turkish).
EMO (Chamber of Electrical Engineers) (1967). Electrical Engi-
neering, EMO, Vol. 127, Ankara, Turkey (in Turkish).
FTF (Financial Trend Forecaster) (2008). Inflation Rate Calculator.
From http://inflationdata.com/Inflation/Inflation_Calculators/Inflation_
Rate_Calculator.asp
Gupta, H. K. (2002). “A review of recent studies of triggered
earthquakes by artificial water reservoirs with special emphasis on
earthquakes in Koyna, India.” Earth-Science Reviews, Vol. 58, No.
3, pp. 279-310.
Gupta, H., Shashidhar, D., Pereira, M., Mandal, P., Purhachandra Rao,
N., Kousalya, M., Satyanarayana, H. V. S., and Dimri, V. P. (2007).
“Earthquake forecast appears feasible at Koyna, India.” National
Geophysical Research Institute, No. 93, pp. 843-848.
ICOLD (International Commission on Large Dams) (1992). Improve-
ment of existing dam monitoring, Bulletin 87, ICOLD, Paris.
IIC (IIC-GrupoEP) (2007). Unit price list for material supply, from
http://www.iic-grupoep.com.
Kratochvíl, D. and Glac, F. (2006). “Prevention of seepage and mod-
ernization of monitoring system on the Morávka Dam.” Twenty-
Second International Congress on Large Dams, Barcelona, pp. 135-
144.
Sezgin, Ö. I· . (2008). An evaluation study on instrumentation system of
Cindere Dam, MSc Thesis, Middle East Technical University,
Ankara, Turkey.
Tanriverdi, I· ., Karadayl, N., Solakoglu, Ö., and Kucur, T. (1983).
Investigation of sediment-laden flow from drain holes of Gokcekaya
Dam, Unpublished Report, Turkish State Hydraulic Works, Ankara
(in Turkish).
USBR (1987). Embankment dam instrumentation manual, USBR, U.S.,
Goverment Printing Office Washington, D.C.