ЗДРУЖЕНИЕ МАКЕДОНСКИ КОМИТЕТ ЗА ГОЛЕМИ БРАНИ

MACEDONIAN COMMIITTЕE ON LARGE DAMS

Зборник на трудови
Proceedings

ТРЕТ КОНГРЕС ЗА БРАНИ

THIRD CONGRESS ON DAMS

13÷14 јуни, 2013 Струга, Република Македонија

13th÷14th Јune, 2013 Struga, Republic of Macedonia
 TOPIC NO. 2
CONSTRUCTION OF DAMS AND TAILINGS DAMS
14 SPECIAL WORKING TECHNOLOGIES DEVELOPED DURING DAM
CONSTRUCTION IN ROMANIA
Dan Stematiu, Adrian Popovici, Radu Sârghiuţă, Altan Abdulamit,
Catalin Popescu, Daniel Gaftoi 145
15 CONSTRUCTION OF THE DAM “SV. PETKA DAM”
Mile Slavevski, Zlatko Nikolovski, Rade Mihajlovski, Petar Jankov,
Grozde Aleksovski, Stanislav Milovanovik 157
16 CONSTRUCTION TECHNOLOGY OF CONCRETE DAM SVETA
PETKA
Mile Slavevski, Marjan Levajkovski, Dragan Janev, Nikola Krajchevski 173
17 METHODOLOGY OF GROUTING FOR CONSOLIDATION AT
LOCATION FOR АRCH DAM “SVETA PETKA”
Milorad Jovanovski, Nikola Gjoncevski, Mile Slavevski, Igor Peshevski 187
18 GROUTING CURTAIN OF DAM ST. PETKA
Mile Slavevski, Jordan Petkov, Nikola Gjonchevski, Ljubisha Jovanovski 199
919 CONSTRUCTION AND GROUTING OF RADIAL JOINTS OF DAM SV.
PETKA
Jordan Petkov, Marjan Levajkovski, Joncho Kalchev 213
20 REALIZATION OF THE DESIGN ON PREPARATION AND FIRST
IMPOUNDING OF RESERVOIR SVETA PETKA
Ilija Rusev, Zvonko Hribernik, Zlatko Nikolovski, Stanislav Milovanovich,
Grozde Aleksovski, Aleksandar Jovanovski, Dragan Janev 223
21 HYDRAULIC CAPACITY OF THE RIVER BED OF RIVER TRESKA
DOWNSTREAM OF HPP SV. PETKA
Ljupcho Petkovski, Mile Slavevski, Zlatko Nikolovski, Ljupco
Gjurcinovski 233
22 DESIGNING, PRODUCTION, INSTALLATION AND TESTING OF
HYDRO MECHANIC EQUIPMENT OF DAM KNEZHEVO
Dijana Likar 245

TOPIC NO. 3
MONITORING AND RESTORATION OF DAMS AND TAILINGS DAMS
23 ASPECTS CONCERNING SURVEILLANCE ACTIVITY
OF THE GURA RÂULUI DAM - ROMANIA
Dumitru Bobocu, Cornel Ilinca, Adrian Popovici 259
24 MODERNIZATION OF MONITORING SYSTEM ON NERETVA POWER
PLANTS
Armin Hadrovic, Ismet Hajduk, Suad Zalihic, Dani Rahimic 271
25 ADVANCED MONITORING OF ROMANIAN DAM PORTFOLIO
Dan Stematiu, Adrian Popovici, Radu Sârghiuţă, Altan Abdulamit,
Catalin Popescu, Daniel Gaftoi 279
26 MONITORING OF SV. PETKA DAM
Zlatko Nikolovski, Aleksandar Jovanovski, Vlatko Mantev 287
 ADVANCED MONITORING OF ROMANIAN DAM PORTFOLIO
Dan Stematiu1, Adrian Popovici1, Radu Sârghiuţă1, Altan Abdulamit1,
Catalin Popescu1, Daniel Gaftoi1
Summary
At present, in Romania there are registered 427 dams included in importance classes A
(exceptional) or B (special) that, according to the law, require a special monitoring. The dam
monitoring system well established nowadays is the result of an evolution that in Romania
lasted for half a century. The tracing of this evolution can be done by analysing the changes
recorded in the course of time both in dam instrumentation with measuring devices and in the
data processing and normal behaviour assessment as well. In the recent years new advanced
monitoring approaches were promoted in order to enhance the monitoring activity. The paper
presents in details two such techniques, namely the infrared image techniques employed in
defining the seepage phenomena in embankment dams and systematic in situ dynamic
measurements of vibration modes of concrete dams that may reveal the aging phenomena by
changes in the dynamic characteristics.
Key words: monitoring, seepage, aging, infrared, natural vibration

НАПРЕДЕН МОНИТОРИНГ НА БРАНИТЕ ВО РОМАНИЈА

Дан Стематиу1, Адриан Поповичи1, Раду Саргита1, Алтан Абдуламит1,
Каталин Попеску1, Даниел Гафтои1
Резиме
Моментално, во Романија се регистрирани 427 брани вброени во важни класи: A
(исклучителни) или В (специјални), кои според законот бараат специфичен
мониторинг. Добро воспоставениот систем за мониторинг на браните е резултат на
еволуцијата која во Романија трае веќе половина век. Патот на оваа еволуција е преку
анализирањето на промените во тек на време во вградените инструменти во браната и
уредите за мерење од една страна и во интерпретацијата на податоците и оцената за
однесувањето на браната од друга страна. Во последните неколку години се
импементирани нови напредни пристапи со цел да се подобри мониторинг процесот.
Во овој реферат детално се презентирани два пристапа, едниот е пристапот на
инфрацрвена слика применет за дефинирање на филтрационите феномени кај насипни
брани и “in sity” динамичко мерење на вибрациони тонови кај бетонски брани, кои
може да го откријат феноменот на стареење преку промената на динамичките
параметри.
Клучни зборови: мониторинг, филтрација, стареење, инфрацрвено, природни
вибрации.

1,2,3,4,5,6
Technical University of Civil Engineering Bucharest, Bv. Lacul Tei, 124, Sect. 2, Bucharest
stematiu@utcb.ro
1. INTRODUCTION

The dams and their associated reservoirs undoubtedly bring advantages and direct and indirect
benefits to a large part of population. But, at the same time, the dams induce an additional risk
to the population situated downstream of them that could be affected by the uncontrolled
spillage of water from the reservoir in the case of failure or technical accidents.
The recent inventory of dams has identified 2366 dams on Romania’s territory that
correspond to the definitions stipulated by the Water Law and the Dam Safety Law. Among
them, 427 dams are included in importance classes A (exceptional) or B (special) that,
according to the law, require a special monitoring [11].
The role of dam monitoring within the safety management, well established nowadays, is
the result of an evolution that in Romania lasted for half a century. The tracing of this
evolution can be done by analyzing the changes recorded in the course of time both in dam
instrumentation with measuring devices and in the data processing and normal behaviour
assessment as well. The dam monitoring is made by observations and measurements. Usually,
dam monitoring systems employ conventional instruments based on electrical, hydraulic or
pneumatic principles, which yield important information on the changes of different physical
quantities such as pressure, stress, strain, displacement or temperature. Nevertheless, the
values measured by these instruments refer only to their location. Between the positions of the
instruments the distribution of the measured physical parameters has to be estimated. It is
unavoidable that the use of these data for the assessment of the condition of the overall
structure contains uncertainties, especially in the case of lateral dams or dikes that have a
length development.
Effective monitoring of seepage within embankment dams is essential in regards to
management of dam safety and prevention of failure. Effective, early detection of seepage in
embankment dams has been difficult as it originates and develops in the subsurface. Infrared
Thermal Imaging is an appropriate technique that is non-contact, non-intrusive, simple and
flexible.
On the other hand, in the case of concrete dams the aging phenomena are usually present
but can not be directly identified by regular instrumentation. The decrease of concrete strength,
the fissure development, the foundation weathering has as the end result the change of the
overall stiffness of the dam structure. This change has to be early detected in order to be a
starting point for further investigation concerning the actual dam safety.
The dam structure stiffness is directly reflected by its dynamic properties since the natural
vibration properties are depending on the mass (assumed constant) and stiffness. The in situ
measurements of the natural vibration modes (shape and periods) are a very convenient
approach of the dam aging investigation.
These new dam monitoring techniques - infrared thermal imaging and in situ dynamic
measurements - are presented in the followings based on two case studies.

2. INFRARED INVESTIGATION TO DEFINE SEEPAGE CONTROL FOR

OSTROVUL MIC DAM

2.1 THE PROBLEM

Uncontrolled seepage can progressively erode soil from the embankment or its foundation in
an upstream direction towards the reservoir and develop a flow conduit (pipe) to the reservoir.
If the seepage forces are large enough, soil will be eroded from the foundation and be
Трет конгрес за брани, Струга, Јуни, 2013 260
deposited in the shape of a cone around the outlet. If these "boils" appear, professional advice
should be sought immediately. Seepage flow which is muddy and carrying soil particles may be
evidence of "piping" and complete failure of the dam could occur. If a seepage problem has
already been identified investigation is required in order to determine the probable cause of the
seepage and the remedial action needed.
The normal approach is to define an adequate numerical model and to examine the actual
causes by defining several scenarios. The correctness of the investigation is dependent on the
accuracy of the model that in its turn depends on the model calibration [3], [5]. The current
field data available are the piezometric measurements. However, for long dikes the extensive
seepage may occur in zones not covered by readings. One of the most convenient methods in
terms of costs and duration that can furnish the needed information is seepage detection using
infrared thermal imaging [6], [8], [10].
The present paper section deals with the seepage problems encountered at the Ostrovul
Mic hydropower plant dikes.

2.2 OSROVUL MIC DAM

Ostrovul Mic development consists in a gated dam and a hydropower station closing the river
bed area and dikes (lateral dams) that create the reservoir contour (figure 1). The power
station has an installed discharge of 90 m3/s and a head of 20 m the power being 15.9 MW.

Figure 1. Plan and panoramic view of Ostrovul Mic reservoire and dam

The dikes are made of ballast (sand and gravel). The watertightening system provides
concrete slabs (4.0 x 5.0 m) on the upstream face and a cutoff wall 5 to 6 m deep in the
foundation (Figure 2). The cutoff wall was performed in an open cut excavation of trapezoidal
261 Third Congress on Dams, Struga, June, 2013
shape 4 m in the upper part and 1m at the basis. To allow the closure of the cut-off wall into
the marle bed rock the upper layer of the river aluvium (3 to 5 m) was removed in front of the
dike upstream toe.

Figure 2. Dike cross section

The investigations conducted during the low level reservoir operation have revealed
several faults in the actual condition of foundation watertightening: - the cutoff wall is not fully
closed into bedrock; - there are gaps beteen the concrete face support beam and the cut-off
wall; several windows were created in the cut-off wall by oversized blocks in the foundation
ground.

2.3 INFRARED SEEPAGE INVESTIGATION AND MATHEMATICAL MODEL

Large leakage through the dikes and esspecial through the left bank one was noticed since the
reservoir commissioning in 1986. This was the cause for restricted water level operating mode
as a prime measure. Over the time, various remedial works have been performed, but they
failed to effectively control the seepage flow. The total seepage was up to 900 l/s. The largest
defective seal behaviour was found between km 1 + 230 and 1 + 370 at the left bank dike. In
the last years, in spite of a significant lower level in the reservoir (5 to 6 m) the wet area on the
downstream slope has extended, the number of springs has raised and the level of emergence
on the slope has also raised.
The main causes of the seepages were investigated based on a calibrated mathematical
model. The 2D finite element model was calibration based on the seepage line pozition within
the dike body. Since the existing piezometric prophiles were far from the seepage affected zone
it was decided to have as calibration data the level of seepage emergence on the downstream
face. A simple visual evaluation was considered very inaccurate due to slope vegetation and
the superficial soil layer. Consequently, the infrared image techniques was emploid.
Infrared light or thermography is the use of an infrared imaging and measurement camera
to "see" and "measure" thermal energy emitted from an object. Seepage flows affect the
temperature field within the dams and their foundations. Leakage detection by means of
infrared images allows detecting the presence and movement of water by evaluating the
thermal field. In the Romanian dam monitoring practice the FLIR cameras are used [4] (FLIR
took its name from the acronym for Forward-Looking InfraRed imaging systems). The analysis
draws on the temperature behaviour and the heat capacity of materials within the body of the
dam and consequently allows the user to identify and isolate temperature variations along the
surface of interest.
Infrared images of the downstream face of the dike were taken. From the same location
digital photos were also taken (figure 3). The images were taken in summer. The body of dike
was warmer than the water in the outer slope of dike, therefore it is shown with red tone in the
Трет конгрес за брани, Струга, Јуни, 2013 262
picture in compression with the yellow one above the saturated zone. By a simple calculation
the elevation of the border between the saturated and unsaturated zones was determined as
455.90 mASL for a reservoir water level of 462.50 mASL.

Figure 3. Model calibration: a) infrared and digital images; b) flow net after calibration.

Starting from the permeability values determined in laboratory and having as calibration
target the emergence of the seepage line on the downstream face, the actual permeability
coefficients were defined by an iterative process. The calibration was performed for two
distinctive cases, one considering the cut-off wall closed into the bedrock, the other one
considering a 0.5 m gap between the cut-off wall deep end and and the bedrock. The values are
listed in table 1.

Table 1. Permeability coefficients

kx (m/s) ky (m/s)
Dike body 2 x 10-3 2 x 10-2
Alluvium
4 x 10-3 4 x 10-3
foundation
Bedrock 1 x 10-6 1 x 10-6
Deteriorated
2 x 10-4 2 x 10-4
concrete face
Cut-off wall 1 x 10-6 1 x 10-6
Rehabilitated
1 x 10-5 1 x 10-5
concrete face
New cut-off wall 1 x 10-6 1 x 10-6
Geomembrane 1 x 10-7 1 x 10-7

The emergence level on the downstream face shown by the flow net in figure 3
corresponds to the measured one. Simmilar values were obtained in both cases – closed or
263 Third Congress on Dams, Struga, June, 2013
flowting cut-off – thus revealing that the main cause of the recorded seepage phenomena is not
the deficiency of the performed foundation sealing. The simulations pointed out that the most
significant contribution to seepage and to the high level of seepage line is done by the
deteriorated slabs on the lower part of the concrete face and especially by the damaged joint
between the concrete face and the cut-off wall. The excavations in front of the upsream toe,
required to implement the cut-off wall in an open trench, have exposed the most pervious layer
above the bedrock to direct infiltration.
Three rehabilitation alternatives were proposed – jet grouting diaphragm, additional cut-
off wall and PVC geomembrane on upsream face. In order to support the final decision three
comparison criteria were selected: the seepage flow, the condition of the downstream face and
the maximum seepage gradient.

3. INVESTIGATION OF PALTINU DAM AGING BASED ON IN SITU

DYNAMIC MESUREMENTS

3.1 THE PROBLEM

The challenge of managing dams aging became a principal focus of dam engineering
throughout the world. In many practical situations damage can be reasonably interpreted and
quantified as a stiffness reduction, with the significant advantage of treating a state of damage
in a linear elastic context. The concrete strength reduction and the extended fissures lead to
Young modulus reduction that is regarded as a damage indicator. A variation in the elastic
modulus induces a change in the dynamic response of the structure.
The main tool available to reveal the changes in the dam dynamic response is the
recording of the free and forced vibrations of the dam and the processing of recordings in
order to identify the vibration periods and shapes and the critical damping eventually.
The dynamic tests consist in recording the vibration field in the dam induced by ambient
noise excitation (free vibrations) or by the hydromechanical equipment vibrations (forced
vibrations). The spectral power density and Fourier spectrum of the recordings can render
evident the first vibration modes, mainly the natural frequencies.
When evaluating the dynamic behaviour of concrete dams, it is reasonable in most cases
to assume that the response is linear for low or moderate-intensity dynamic loading. Such an
assumption of linear response simplifies both the formulation of the mathematical model used
to represent the dam, reservoir water and foundation rock system and also the procedures used
to calculate the response.
The present paper section deals with the field testing of dynamic response of Paltinu arch
dam and the corresponding calibrated model used to asses the dam safety or its deterioration
due to aging phenomena.

3.2 PALTINU DAM

Located on Doftana River, Paltinu arch dam was constructed between 1960 and 1971. The
dam height is of 108 m and the crest length is of 330 m. Storage has been purposed mainly to
provide the discharge for drinking and industrial water supply.
The cross profile of the location is characterised by a pronounced asymmetry caused by a
terrace located on the left bank having a width of about 100 m along the valley. The dam
foundation consists of Carpathian flish and includes detrial sedimentary rocks: sandstone,
micro conglomerates and shistous clays.
Трет конгрес за брани, Струга, Јуни, 2013 264
 The foundation rock presents a marked bedded arrangement. The rock mass is affected by
a dense cracking system, more evident in sandstone. The dam area is crossed by several faults
generally located transverse on the valley. The permeability of the foundation rock was quite
large (2…8 lugeons) and an extensive grout curtain was provided.
Dam structure is made up of central symmetrical double curvature arch body that rests on
a pulvino by means of a peripherical joint and a parabolic wing extending over the left
abutment terrace (figure 1).
During the first 3 years after commissioning the reservoir was operated at lower levels. At
the new cycle of reservoir filing in 1974 when the water levels have exceeded the previous
recorded ones, an abnormal behaviour of the dam was noticed: dam displacements larger then
predicted, movements at the foundation level, joint openings, cracks at rock surface and
significant increase of seepage (from 10 l/s to 150 l/s).
Remedial works had in view the causes of the unusual loading scheme. They consisted
mainly of additional rock watertightening upstream the fault, drainage downstream and
additional mass on the left abutment. To reduce the seepage forces within the rock mass the
upstream face of the rock was covered with concrete and a new grout curtain was performed
along the left abutment. To prevent possible water pressure effects a new drainage system was
provided by means of two drainage galleries located 10 m downstream the grout curtain. To
increase the abutment stability a concrete cover was performed on the left bank downstream
area that allowed also rock mass consolidation by grouting and controlled drainage (fig. 4, 5).

Figure 4. Paltinu arch dam

265 Third Congress on Dams, Struga, June, 2013
Figure 5. Downstream view of Paltinu dam

3.3 FIELD DYNAMIC MEASUREMENTS

The dynamic response of Paltinu dam was achieved by means of highly sensitive
KINEMETRICS equipment. Consequently, a chain of SS-1 RANGER seismometers in
simultaneous configurations, were used in order to record the response (figure 6).
The SS-1 Ranger is a structural dynamics instrument, used in the determination of multi-
modes of vibration under low-level excitation. The positioning scheme for the transducers on
the dams was adopted in accordance with the expected response.

Figure 6. Positioning of equipment on dam crest

The layouts of the sensors were performed in order to point out the maximum
displacements at the crest level, the characteristics of the fundamental eigenmode of vibration
Трет конгрес за брани, Струга, Јуни, 2013 266
of the dam, together with the frequency content of the recorded signals. The location of
seismometers used in the 2006 campaign for the dam response recording is shown in figure 7.

Figure 7. Location of the sensors

Seven sensors were simultaneously positioned at the crest level, on a horizontal radial
direction, on a horizontal tangential direction and on a vertical direction. The records have
been carried out taking into account vibration sources as micro tremors, together with
vibrations produced by hydromechanics equipment operating in the body of the dam.
The processing of the data was carried out digitally, which allowed the determination of
mean values, peak-peak values, Fourier spectra and response power spectral densities,
relationships between input and output, auto correlation functions, etc. Following the process
and interpretation of the instrumental data, the natural frequencies/ periods of vibration of the
Paltinu dam have been established. The first – fundamental period of vibration identified by
means of the measurement was T1 = 0.44 sec (f1 = 2.26 Hz).

3.4 DATA INTERPRETATION BASED ON THE MATHEMATICAL MODEL

A 3 D finite element model of the dam – foundation system was developed using the ANSYS
computer program (see figure 8).
The dam-foundation interaction effects were represented by a "standard" massless
foundation model, in which only flexibility of the foundation rock is considered but its inertia
and damping are ignored. The rock foundation was modelled with 2780 solid elements
connected in 3510 nodes. The arch dam was modelled as a monolithic structure represented by
574 of finite elements of appropriate types.
The dam-water interaction effect was represented by the added hydrodynamic mass
models according to Westergaard. The added water mass is realized by means of 173 lumped
mass elements at the upstream face of the dam.
The sensitivity analysis leads to a concrete elasticity modulus of 25106 kPa
corresponding to the measured fundamental frequency
The computed eigenfrequency of the system with the impounded reservoir and the added
mass approach for structure reservoir interaction are presented in table 2. From the frequency
response analysis it can be seen, that the first 5 eigenmodes are sufficient to represent the
structural response.
267 Third Congress on Dams, Struga, June, 2013
Figure 8. The finite element model

Table 2. Vibration periods

Frequency Period
Eigenmode
[Hz] [sec]
1 2.25 0.44
2 2.45 0.41
3 2.86 0.35
4 3.23 0.31
5 3.69 0.27
In order to acquire advance knowledge of the dynamic behaviour of the dam and also to
examine the accuracy of the results, the computed natural modes of vibration were presented in
the form of deflected shapes. The first two modes are presented in figure 9.

Figure 9. The first two vibration modes of the FE model

Трет конгрес за брани, Струга, Јуни, 2013 268
 Performing several modal analyses on the basis of a finite element model of the dam,
each of them for a different reservoir water level one can establish a graphical correlation
between the fundamental period and reservoir condition. The analysis results are presented in
table 3. In figure 10 the actual variation of the fundamental vibration period in terms of the
water level is presented.

Table 3. Modal frequencies in terms of the water level

Mod 1 Mod 2 Mod 3 Mod 4 Mod 5
Wate r le vel
f1 [Hz] f2 [Hz] f3 [Hz] f4 [Hz] f5 [Hz]
640 2.258 2.480 3.542 3.854 4.545
635 2.283 2.540 3.583 3.865 4.573
630 2.313 2.618 3.628 3.886 4.610
625 2.322 2.647 3.643 3.897 4.621
620 2.340 2.704 3.670 3.919 4.643
610 2.354 2.757 3.694 3.939 4.658
600 2.361 2.787 3.709 3.950 4.664
590 2.364 2.802 3.717 3.956 4.667
Lac gol 2.365 2.813 3.725 3.963 4.670

650
(mASL)

640
630
[mdM]

620
in laclevel

610
600
water

590
580 T1
W Nivel

570
Reservoir

560
550
540
0.42 0.425 0.43 0.435 0.44 0.445
T [s]

Figure 10. First vibration period in terms of water level

New acquired field measurements, in special annual campaigns, are used for assessing the
aging effects. If the new values does not differ significantly from the previous value one can
conclude that the dam has preserved its structural properties. If this is not the case, further
investigation are required in order to evaluate the causes and significance of the changes which
can be due to either ageing or fatigue or structural local degradations.

REFERENCE
[1] Abdulamit, A., Stematiu, D., Toma, I. (1995) Identification of Dynamic Elastic Material Properties Using
Hybrid Models, Proceedings New Advances in Modal Synthesis of Large Structures. Non-linear, Damped
and Non-deterministic Cases, Lyon, France, October.
[2] Allen, R. (2011) Safety Evaluations of Hidden Dam – Seepage Models Meet Reality. 2011 AEG Annual
Meeting.
[3] CRWMS M&O (2001). Seepage Calibration Model and Seepage Testing Data. 153045MDL-NBS-HS
000004 REV 01. Las Vegas, Nevada.
[4] FLIR SYSTEMS.(2006). ThermaCAM TM B4. User manual.
[5] Guardo, M., Rohrer, K, P. (2000) Calibration of a steady-state seepage model from transient recovery of
field data† JAWRA Journal of the American Water Resources Association. Volume 36, Issue 1, pages 87–
94.
269 Third Congress on Dams, Struga, June, 2013
[6] Moel, M. (2010). Seepage Detection within Embankment Dams Using Infrared Thermal Imaging.
Proceedings of the 2010 ANCOLD Conference.
[7] Sarghiuta, R., Abdulamit, A., Bugnariu, T. (2007) Safety Assessment of Arch Dams using Global Elastic
Modulus Concept. Proceedings of International Symposium Thirty Years from the Romania Earthquake
of March 4, 1977 Bucharest, Romania
[8] Shutko I and col. (2009). New Technologies in Monitoring and Emergency Mapping of Water Seepage
and Dangerously High Groundwater. Research report.
[9] Stematiu, D., Bugnariu, T., Toma, I. (1995) Global Elastic Modula for Arch Dams, Proceedings of the
Conference on Research and Development in the Field of Dams, Crans–Montana, Switzerland September.
[10] Stematiu, D. (2008). Use of infrared images to revealed infiltrations from tailings dams (in Romanian).
Hidrotehnica, Vol. 53, Nr. 12.
[11] *** Dam Safety Law No 466 (2001) M.Of. Nr. 428/31 iul. 2001 (in Romanian).

Трет конгрес за брани, Струга, Јуни, 2013 270