FACULTY OF ENGINEERING AND THE BUILT ENVIRONMENT

DEPARTMENT OF CIVIL ENGINEERING

n
w
To
e
ap

INVESTIGATING THE OPERATIONAL BEHAVIOUR OF A

C

DOUBLE CURVATURE ARCH DAM

of
ity

By
rs
ve

Zac James Prins

ni

Submitted in partial fulfillment of the requirements for the degree

U

M.Eng.

Civil Infrastructure Management and Maintenance

Supervisor: Professor Pilate Moyo

 n
The copyright of this thesis vests in the author. No

w
quotation from it or information derived from it is to be

To
published without full acknowledgement of the source.
e
The thesis is to be used for private study or non-
ap
commercial research purposes only.
C
of

Published by the University of Cape Town (UCT) in terms

of the non-exclusive license granted to UCT by the author.
ity
rs
ve
ni
U
 PLAGIARISM DECLARATION

I know the meaning of plagiarism and declare that all the work in the document, save
for that which is properly acknowledged, is my own. This thesis/dissertation has
been submitted to the Turnitin module (or equivalent similarity and originality
checking software) and I confirm that my supervisor has seen my report and any
concerns revealed by such have been resolved with my supervisor.

Signature Removed
Signature: ______________________

Date: 12 May 2017

i
 DEDICATION

To my loving family for their unwavering support.

ii
 ABSTRACT

The safety of dams is crucial in ensuring the continual availability of water, safety of
the surrounding communities and infrastructure. Surveillance systems are
implemented to monitor the structural integrity of certain dams which have a safety
risk. The components and extent of the surveillance systems adopted depends on
many factors, which include the type of dam wall structure used to impound the
reservoir, geotechnical and environmental conditions.

The case study used for this thesis is Kouga Dam located in the Eastern Cape
Province of South Africa. It is a double curvature, concrete arch dam which supplies
water for domestic, irrigation, and industrial use to the Gamtoos River Valley and
Nelson Mandela Metropolitan. During construction the stability of the right flank was
questioned and subsequently remedial measures were taken in order to increase the
shear resistance of this flank. Previous dam safety evaluations also noted the
possibility of Alkali Silica Reaction (ASR) occurring within the structure which
resulted in concrete swelling and loss of strength. Due to these factors and the large
hazard potential rating associated with this dam, an intensive surveillance system
has been used to monitor the dam's behaviour during operation.

In this thesis the results of the surveillance system is analysed. A strong linear
relationship exists between the temperature loading and displacement response of
the dam wall. Changes in temperature initiate the response of the structure almost
instantaneously. A more complex relationship exists between hydrostatic loading and
the displacement response of the structure. A phase lag of approximately one to
three months is evident between these two variables. Since construction the
displacement and strain rates in the upstream (y) and upward (z) directions are
0.3mm/annum and 8.6µε/annum respectively. However, since 1989 there has been a
reduction in the average displacement and strain rates in all directions by
approximately 70%. This may suggest that the ASR has stabilized. The vertical
construction joints, especially the central and upper joints, are relatively open during
low water levels. The structure is found to transfer the imposed loading mainly to the
central foundation via dominant cantilever action. As a result the reaction forces on
the upper foundation have been found to be relatively low, lowering the risk of
potential shear failure of the right foundation. Small foundation movements of less
than 0.3mm have been observed within the foundation downstream of the dam wall
on the right flank. These movements are between 10 and 40m within the
foundations.

iii
 ACKNOWLEDGEMENT

Firstly I would like to thank Professor Pilate Moyo for his supervision and guidance
throughout this thesis. Dr. Chris Oosthuizen for urging me to pursue postgraduate
studies and stoking my interest in dam engineering. As well as Patrick Bukenya for
his help in the field and during the write-up of the document.

I would also like to acknowledge the Department of Water and Sanitation, for
providing the means for further studies, allowing me to utilize invaluable information
and also my fellow colleagues who assisted with the research.

Lastly, I would like to thank my family for their continuous support.

iv
 TABLE OF CONTENTS

Plagiarism Declaration.............................................................................................. i

Dedication ................................................................................................................. ii

Abstract.................................................................................................................... iii

Acknowledgement................................................................................................... iv

Table of Contents ..................................................................................................... v

List of Tables ......................................................................................................... viii

List of Figures ......................................................................................................... ix

List of Abbreviations............................................................................................... xi

Nomenclature ......................................................................................................... xii

Chapter 1................................................................................................................... 1

1 Introduction.................................................................................................... 1
1.1 Background ................................................................................................ 1
1.2 Problem Statement ..................................................................................... 2
1.3 Objective..................................................................................................... 2
1.4 Limitations and Scope of Research ............................................................ 2
1.5 Thesis Structure ......................................................................................... 2

Chapter 2................................................................................................................... 4

2 Design, Construction, and Management of Arch Dams: A Review ........... 4

2.1 Design of arch dams ................................................................................... 4
2.1.1 Brief historical review ................................................................... 4
2.1.2 Site considerations....................................................................... 4
2.1.3 Loading ........................................................................................ 5
2.1.4 Analysis ....................................................................................... 6
2.2 Construction of arch dams .......................................................................... 9
2.3 Management of arch dams ....................................................................... 10
2.3.1 Purpose (Statutory requirements) .............................................. 10
2.3.2 Structural response .................................................................... 11
2.3.3 Failure modes ............................................................................ 15
2.3.4 Alkali Silica Reaction.................................................................. 20

v
 2.3.5 Instrumentation .......................................................................... 23
2.4 Chapter Summary .................................................................................... 24

Chapter 3................................................................................................................. 25

3 Methodology ................................................................................................ 25
3.1 Introduction ............................................................................................... 25
3.2 Kouga Dam............................................................................................... 25
3.2.1 Locality....................................................................................... 25
3.2.2 Dam Characteristics................................................................... 26
3.2.3 Background ................................................................................ 27
3.2.4 Geology ..................................................................................... 28
3.2.5 Right flank foundation conditions ............................................... 29
3.2.6 Loading on right flank foundation ............................................... 31
3.3 Monitoring System .................................................................................... 33
3.3.1 Overview .................................................................................... 33
3.3.2 Water Levels .............................................................................. 33
3.3.3 Temperature .............................................................................. 33
3.3.4 Displacements ........................................................................... 34
3.3.5 Ambient Vibration Measurements .............................................. 38
3.4 Chapter Summary .................................................................................... 42

Chapter 4................................................................................................................. 43

4 Results and Discussion .............................................................................. 43

4.1 Introduction ............................................................................................... 43
4.2 Instrumentation Data Results ................................................................... 43
4.2.1 Water Levels .............................................................................. 43
4.2.2 Temperature .............................................................................. 44
4.2.3 Geodetic Data ............................................................................ 45
4.2.4 GNSS/GPS Data........................................................................ 51
4.2.5 GNSS/GPS and Water Level ..................................................... 52
4.2.6 GNSS/GPS and Temperature .................................................... 54
4.2.6 Trivec System ............................................................................ 55
4.2.7 AVM and Water Level ................................................................ 57
4.2.8 Crack widths .............................................................................. 59
4.3 Comparative Study ................................................................................... 61
4.4 Dam Forces on Foundation ...................................................................... 62
4.5 Chapter Summary .................................................................................... 63

vi
Chapter 5................................................................................................................. 64

5 Conclusion and Recommendations ........................................................... 64

5.1 Summary .................................................................................................. 64
5.2 Conclusion ................................................................................................ 64
5.3 Recommendations .................................................................................... 65

References .............................................................................................................. 66

Appendices ............................................................................................................. 70

A. Trivec results ............................................................................................... 70

vii
 LIST OF TABLES

Table 1: Typical static and dynamic loads on dams ................................................... 6

Table 2: Size classification ....................................................................................... 10
Table 3: Hazard potential classification .................................................................... 10
Table 4: Category classification of dams with a safety risk ...................................... 10
Table 5: Damping in concrete arch dams (USACE, 1994: 7-11) .............................. 13
Table 6: Operational Failure Modes of Dams ........................................................... 16
Table 7: Affect of ASR on concrete at 28 days (Oberholster, 2009: 211) ................. 23
Table 8: Investigated parameters ............................................................................. 33
Table 9: Installation levels and height of Trivec systems.......................................... 38
Table 10: Summary of displacement and strain rates .............................................. 50
Table 11: Trivec data summary ................................................................................ 57
Table 12: Crack width measurement periods ........................................................... 59
Table 13: Comparative dam statistics ...................................................................... 61

viii
 LIST OF FIGURES

Figure 1: Schematic topography of various dam sites (USACE, 1994:1-3) ................ 5

Figure 2: Deflected arch (left) and cantilever (right) units (USACE, 1993:2-4) ........... 7
Figure 3: Generalised Westergaard added mass (Federal Energy Regulatory
Commission [FERC], 1999: 11-122)........................................................................... 8
Figure 4: Flaming Gorge dam construction in 1962 (USBR, 2016) ............................ 9
Figure 5: Schematic of force transfer within an arch dam (Shaw, 2015: 324) .......... 11
Figure 6: Typical stress distributions for thin arch and arch/gravity dams (Shaw,
2015: 325, 327) ........................................................................................................ 12
Figure 7: Dynamic system configuration .................................................................. 13
Figure 8: Unstable blocks during seismic loading (Ibid.) .......................................... 17
Figure 9: Abutment instability causing Malpasset dam failure (Ibid.: 42) .................. 19
Figure 10: Classic ASR induced map cracking (Poraver, 2016) ............................... 21
Figure 11: Cored sample of ASR deteriorated concrete (Concrete Microscope
Library, 2016) ........................................................................................................... 22
Figure 12: Discolouration arround ASR related cracks (Ibid.: 207) .......................... 22
Figure 13: Locality of Kouga Dam (Google Earth, 2013) .......................................... 25
Figure 14: Kouga Dam ............................................................................................ 27
Figure 15: Close-up of right flank abutment (Hobbs et. al., 1967: 255) .................... 29
Figure 16: Right flank prior to completion and stress pad remedial measures (Hobbs
et. al., 1967: 654) ..................................................................................................... 29
Figure 17: Capping beam and stressing slab on right flank (Hobbs et. al., 1967:
664,650) ................................................................................................................... 30
Figure 18: Arch and cantilever reactions on the rock foundation in relation to the
stratification dip and strike directions (Forbes Dick, 1989: 14) ................................. 31
Figure 19: Resultant forces on the right foundation (Forbes Dick, 1989: 16) ........... 32
Figure 20: Top view showing locations of thermometers on upstream face ............. 34
Figure 21: Downstream general elevation of monitoring points ................................ 35
Figure 22: Kouga Dam GNSS/GPS Network (Pretorius & du Toit, 2014: 8) ............. 35
Figure 23: GNSS/GPS System at Kouga Dam (Ibid.: 4)........................................... 36
Figure 24: Trivec system components (Naude, 2002: 4) .......................................... 37
Figure 25: Locations of Trivec systems .................................................................... 38
Figure 26: Tri-axial accelerometer sensor (GeoSIG, 2010: 1) .................................. 39
Figure 27: Data Acquisition System ......................................................................... 40
Figure 28: The computer used for data acquisition and processing ......................... 40
Figure 29: Location of accelerometers within gallery ............................................... 41
Figure 30: Roving and permanent accelerometers .................................................. 41
Figure 31: Water level record ................................................................................... 44
Figure 32: Trajectory of the Sun across the dam wall .............................................. 44
Figure 33: Average temperature for the left and right flanks .................................... 45
Figure 34: Water record and geodetic data
................................................................................................................................. 46
Figure 35: Relative upstream/downstream displacement (116) ............................... 47

ix
Figure 36: Relative right flank/left flank displacement of right flank quarter point
(111) ......................................................................................................................... 48
Figure 37: Relative right flank/left flank displacement of left flank quarter point (122)
................................................................................................................................. 48
Figure 38: Relative vertical displacement at centre (117)......................................... 49
Figure 39: Schematic of permanent x-y displacement of dam wall .......................... 50
Figure 40: Displacement data for left flank (KG12) .................................................. 51
Figure 41: Displacement data for right flank (KG13) ................................................ 52
Figure 42: Displacement and WL data for left flank .................................................. 52
Figure 43: Displacement and WL data for right flank ............................................... 53
Figure 44: Displacement VS Water Level................................................................. 53
Figure 45: Displacement and Temperature data for left flank ................................... 54
Figure 46: Displacement and Temperature data for right flank ................................ 54
Figure 47: Displacement VS Temperature ............................................................... 55
Figure 48: Trivec Kogr1 during high water levels ..................................................... 56
Figure 49: Natural frequencies of the first four modes of vibration at varying water
levels ........................................................................................................................ 58
Figure 50: Natural frequencies for varying modes of vibration at FSL ...................... 59
Figure 51: Period A - Falling water level................................................................... 60
Figure 52: Period B - Rising water level ................................................................... 60
Figure 53: Natural frequencies for Kouga and Roode Elsberg Dams ....................... 61
Figure 54: Force transfer through arch during FSL hydrostatic loading.................. 62

x
 LIST OF ABBREVIATIONS

APP Approved Professional Person

ASCE American Society of Civil Engineers
ASR Alkali Silica Reaction
AVM Ambient Vibration Measurements
dB Decibels
DSO Dam Safety Office
DWA Department of Water Affairs
DWAF Department of Water Affairs and Forestry
DWS Department of Water and Sanitation
EFDD Enhanced Frequency Domain Decomposition
FEA Finite Element Analysis
FERC Federal Energy Regulatory Commission
GNSS Global Navigation Satellite System
GPS Global Positioning System
GRGWS Gamtoos River Government Water Scheme
Ha Hectares
Hz Hertz
MDOF Multiple Degrees of Freedom
NWA National Water Act
OMA Operational Modal Analysis
RL Reduced Level
SDOF Single Degree of Freedom
TMG Table Mountain Group
USA United States of America
USACE United States Army Corps of Engineers
USBR United States Bureau of Reclamation

xi
 NOMENCLATURE

Latin Upper Case

°C Degrees Celcius
K Canyon Shape Factor
Pn Westergaard hydro-dynamic pressure

Latin Lower Case

c Damping
f0 Amplitude of excitation force
g Acceleration due to gravity
k Stiffness
ke Equivalent stiffness
m Mass or meters
mm Millimeters
3
m Cubic meters
me Equivalent Mass
t Time
ütn Total normal acceleration
Displacement
̇ Velocity
̈ Acceleration

Greek Lower Case

α Westergaard pressure coefficent
ζ Damping Ratio
φ Mode shapes
ρw Density of water
µε Micro-strain
µm Micron/Micrometer
ω Excitation Frequency
ωn Natural Frequency

xii
 CHAPTER 1

1 INTRODUCTION

1.1 Background
South Africa is deemed a water scarce country. In 2014 it averaged less than
500mm of rainfall annually compared to the global average of approximately
1160mm (The World Bank, 2016). The country is significantly reliant on water, not
only for domestic purposes, but in order to support major contributing sectors to the
country's economy. These include the mining and agricultural sectors. In addition to
these water demands, domestic consumption has increased over the years due to
population growth and urbanization.

In light of the water demands, the South African government has taken steps to
protect this resource1. Water impounding infrastructure within the country is subject
to legislation which ensures their safety. The National Water Act (NWA) mandates
dam safety regulations (NWA, No. 36 of 1998, 1998: s123) which include routine
inspections to evaluate the safety of the dams. Monitoring systems are one of the
tools used by engineers during these evaluations. These tools assist in
understanding the structural behaviour of the dam during operation and when under
extreme loading.

The monitoring system records loading and response data. An analysis of this data
identifies trends which give insights into the operational behaviour of the structure
and assist in highlighting potential modes of failure. These insights allow engineers
to make informed decisions regarding dam safety.

Kouga Dam, which is the thesis case study, is categorised as a double curvature,
concrete arch dam. These dam types are generally chosen in canyon sites which
have high strength abutments. This is because the dam is curved in section and in
plan views, resisting the loading not only by self-weight, but also by arch action.
Therefore the dam wall is thinner in cross-section compared to gravity dams, which
resist the loading through sheer mass.

The relatively smaller cross-sectional area of double curvature arch dams is not
large enough to resist the imposed loading by mass alone. Consequently, significant
loss in the arch action, or foundation instability in the direction of arch thrusts may
lead to potentially catastrophic failure of the structure.

1
For further reading, see: (National Water Act [NWA], No. 36 of 1998, 1998: chap12); (NWA, No. 36
of 1998. Regulation, 2012); (South African National Congress on Large Dams [SANCOLD], 1991)

1
1.2 Problem Statement
Double curvature arch dams transfer the imposed loading on the structure via a
combination of arch and cantilever action. This unique way in which arch dams
transfer forces depend on many factors, which include the environmental conditions,
site topography, design assumptions, and construction techniques.

In evaluating the safety of arch dams the operational behaviour of the structure
needs to be understood. Especially in light of factors such as; variable material
properties of foundations, material deterioration, aging, changes in environmental
conditions and loading, which may change the way the structure behaves over time.
Thorough data analysis of the monitoring systems installed at dams assist in
providing greater insight into the structural response. These insights include general
behaviour trends and highlighting potential failure modes. This allows informed
decisions to be made with regards to ensuring the safety of the structure.

1.3 Objective
The purpose of this research is to analyse the instrumentation data recorded at
Kouga, a double curvature arch dam located in the Eastern Cape Province of South
Africa. The objective of this analysis is to gain insight into the operational behaviour
of the dam, namely the effects that different measured parameters have on the
structure's response, to track deterioration over time, and to identify potential modes
of failure.

1.4 Limitations and Scope of Research

The scope of the thesis is limited to Kouga Dam. The parameters considered, as
recorded by the monitoring system, include the following;

 Water levels
 Temperature
 Displacements
 Natural frequencies

1.5 Thesis Structure

Chapter 1 - Introduction

This chapter considers the need for dams and the associated governmental dam
safety regulations. The objective of the thesis is presented, followed by the scope of
the study.

2
Chapter 2 - Design, Construction, and Management of Arch Dams: A Review

This chapter presents literature relating to arch dams. These works focus on design
and analysis, general construction practices, and the management of arch dams.

Chapter 3 - Methodology

This chapter provides background information on the thesis case study and its
monitoring system which records the loading, response, and integrity of the structure
over time.

Chapter 4 - Results and Discussion

This chapter presents the recorded data of the instrumentation and the identified
behavioural trends are discussed. In addition a comparative study is done with a
dam of similar size, comparing the determined structural characteristics.

Chapter 5 - Conclusion and Recommendations

This chapter presents conclusions based on the identified behavioural trends and
provides recommendations for future studies related to the topic of this thesis.

Appendices

This section consists of the Trivec displacement data recorded at Kouga Dam.

3
 CHAPTER 2
2 DESIGN, CONSTRUCTION, AND MANAGEMENT OF ARCH DAMS:
A REVIEW

2.1 Design of arch dams

2.1.1 Brief historical review

The design of arch dams has developed significantly over the past century and is
documented in manuals and guidelines 2 . The Trial Load Method (United States
Bureau of Reclamation [USBR], 1977a: chap4b) was adopted in the United States of
America (USA) before the advent of computers. This method estimates the stresses
by assuming a dual structural system consisting of horizontal arch and vertical
cantilever units. The method's contribution to analysis of arch dams is further
elaborated in Section 2.1.4. In Europe the membrane method was used in defining
the arch shape (da Silva & Júlio, 1995: 534). This method involved water being
loaded on a rubber membrane which is fixed to a frame. The boundary of the frame
would be the shape of the valley in which the dam is situated. The rubber membrane
deforms into a funicular shape off which the arch is based. Shaw (2015: 321) notes
that in South Africa a similar method was developed by Myburgh (1960), known as
the mesh model. In this model perpendicular wires were opted for instead of the
rubber membrane, and on the intersections of the wires, weights were applied to
simulate the hydraulic loading on the dam wall.

All of these methods proved effective in defining the arch shape during the early
period of arch design. However, following the introduction and rapid development of
computers, finite element analyses (FEA) has become the primary tool in arch dam
design and analysis. This has allowed for the design and construction of efficient
structures which require less material and are more economical. In all of the above
mentioned methods one of the critical factors in the design of arch dams is the site
topography. This is discussed in the section below.

2.1.2 Site considerations

Arch dams function by transferring imposed loads on the structure to the supporting
abutments and foundation. The chosen characteristics of the arch structure and how
it transfers the forces to the supports is largely dependent on the specific site and
foundation conditions.

The site topography may broadly be categorized into four types, namely narrow-V,
wide-V, narrow-U, or wide-U (See Figure 1). Canyon shape factors (K) are often
used to describe the topography and is defined as the total developed length of

2
For further reading, see: (United States Bureau of Reclamation [USBR], 1977a); (USBR, 1977b);
(United States Army Corps of Engineers [USACE], 1994).

4
foundation surface divided by the maximum dam height (Shaw, 2015: 323). Arch
dams are most effective in narrow valleys and particularly V-shaped with low K
values. These topographies allow the dam wall to be relatively thin in section, whilst
transferring the majority of the imposed loads to the abutments via arch action. The
thin, flexible cantilevers allow the forces to transfer horizontally, which results in a
much more distributed load throughout the foundation. Conversely, the wider the
canyon and higher the K value, the thicker the arch and the more the dam relies on
cantilever action to resist the loading. This in turn requires the structure to increase
in wall thickness the closer to the riverbed in order to accommodate for the reliance
on cantilever action.

Figure 1: Schematic topography of various dam sites (USACE, 1994:1-3)

2.1.3 Loading
Dams are subjected to loading which are categorised as either static or dynamic.
Static loads can be further divided into sustained and cyclic loads. The sustained
loading does not change with time and includes loads such as the dead load of the
dam wall and appurtenant structures. The magnitude of cyclic loads change with
time, ranging from hourly to seasonally. These loads include changing temperature
and water level. The response of arch dams to static loading is predominantly
governed by the stiffness of the structure.

Dynamic loading implies a load which results in inertial forces. The frequency of this
loading is much higher compared to that of cyclic loads. Examples of dynamic
loading include micro-seismic vibrations caused by wind loading and oceanic wave
motion, as well as macro-seismic activity such as earthquakes. Depending on the
ratio of excitation to natural frequency, the structural response to the dynamic
excitation is governed by inertia (mass), damping, and stiffness. The dynamic
response of structures will be expanded in more detail in Section 2.3.2. Typical loads
on dams are summarized in Table 1.

5
Table 1: Typical static and dynamic loads on dams
 Gravity loads (Own weight)
 Silt loads
Static: Sustained
 Ice loads
 Concentrated loads (Appurtenant structures)
 Hydrostatic loads
i. Water level
ii. Pore pressure
Static: Cyclic
iii. Uplift

 Thermal loads (Temperature)

 Waves
 Wind
Micro-seismicity  Atmospheric phenomena
(Ambient loads)  Human activities
i. Traffic
ii. Machinery

Dynamic
Macro-seismicity Earthquakes
 Impact loads
i. Explosions ("Acts of war")
ii. Lightning
Other iii. Collisions
 Landslides
i. Direct contact with dam wall
ii. Resultant wave action

2.1.4 Analysis

Trial Load Method

One of the earliest analysis procedures adopted for the design of arch dams is
known as the Trial Load Method. The method estimates the stresses developed
within the structure in a simplistic and acceptably accurate (for the purposes of early
design) manner. The underlying assumption is that arch dams resist imposed loads
by two systems of structural members, namely horizontal arch and vertical cantilever
units (USBR, 1977a: chap4b). Arch units are bound by horizontal planes and transfer
the forces as hoop stresses throughout the arch. Cantilever units are bound by
vertical planes and transfer the forces as cantilever stresses. This is often referred to
as arch and cantilever action respectively. It is assumed that the corresponding arch
and cantilever deflections at any point throughout the dam are equal (Shown
significantly exaggerated in Figure 2). After grouting the structure is assumed to act
as a monolith, resisting imposed loads by a combination of arch and cantilever
action.

6
 Figure 2: Deflected arch (left) and cantilever (right) units (USACE, 1993:2-4)

Finite Element Analysis

In recent times, FEA has been used in order to calculate stresses at a much greater
accuracy compared to previous techniques. This method divides a structure into
inter-connected finite elements, and solves for the resultant displacements with the
use of equilibrium equations, incorporating the stiffness matrices and load vectors of
the system. The stresses for each element can then be determined from the stress-
displacement relationship (Bathe, 1996: chap1).

Dynamic Analysis: Generalised Westergaard Method

Westergaard (1933) first developed the notion of using an added mass to represent
the interaction between the dam wall and reservoir during the dynamic analysis.
Initially the method assumed a concrete gravity dam with vertical upstream face,
however this was later extended to a more general form to accommodate other dam
types. The generalised Westergaard method (Kuo, 1982) can be applied to arch
dams as it accounts, to some extent, for the upstream curvature and three
dimensional transfer of forces. For this case the same added mass profile is used,
however the hydro-dynamic forces act normal to the curved dam wall surface. See
Figure 3 for the graphical representation of this method.

7
 Figure 3: Generalised Westergaard added mass (Federal Energy Regulatory Commission
[FERC], 1999: 11-122)

The hydro-dynamic forces exerted on the upstream face are given by the following
equation;
̈ (2.1)

with,

Pn: Hydro-dynamic pressure

α: Westergaard pressure coefficent
ütn: Total normal acceleration

The pressure coefficient is described by;

(2.2)

with,

ρw: Density of water

H: Height of water level
Z: Distance to point of interest

8
2.2 Construction of arch dams
The USACE (1994: chap13), notes the following construction considerations for arch
dams.

i. Diversion
ii. Foundation excavation
iii. Consolidation and curtain grouting
iv. Concrete operations
v. Monolith joints
vi. Galleries and adits
vii. Drains
viii. Appurtenant structures

Careful consideration and quality assurance of each of these aspects are required in
order for the design assumptions to be realised.

Arch dams require competent abutments for the arch to thrust against, therefore the
selection and preparation of the foundations are crucial. Weaker regions of the
foundations are identified and removed prior to consolidation grouting (Ibid.: 13-5).
Inadequate grouting or concrete placement may result in high pore pressures and
possibly significant uplift forces in saturated voids within the foundation. Curtain
grouting is done in order to control seepage underneath the dam wall.

Arch dams are constructed as independent cantilevers with leader and follower
sections as shown in Figure 4. The interfaces between the blocks are vertical
contraction joints. The design may incorporate shear keys at these joints in order to
increase the shear resistance between adjacent blocks. After all the concrete lifts
have been constructed the joints are grouted in order to allow the structure to act as
a monolithic system.

Figure 4: Flaming Gorge dam construction in 1962 (USBR, 2016)

9
2.3 Management of arch dams

2.3.1 Purpose (Statutory requirements)

Chapter 12 of the National Water Act (1998) relates to the safety of dams and aims
at "... improving the safety of new and existing dams with a safety risk so as to
reduce the potential for harm to the public, damage to property or to resource
quality.". Dams with a safety risk are defined as structures "... which can contain,
store or dam more than 50 000 cubic metres of water..., and which has a wall of a
vertical height of more than five metres...". It is then further classified into categories
based on size and hazard potential (NWA, No. 36 of 1998. Reg. 139, 2012). See
Tables 2 to 4.
Table 2: Size classification

Size class Maximum wall height in metres (m)

Small Less than 12 m
Medium Equal to or more than 12 m but less than 30 m
Large Equal to or more than 30 m

Table 3: Hazard potential classification

Potential adverse
Hazard potential Potential loss of Potential economic
impact on resource
rating life loss
quality
Low None Minimal Low
Significant Not more than ten Significant Significant
High More than 10 Great Severe

Table 4: Category classification of dams with a safety risk

Hazard potential rating

Size class
Low Significant High
Small Category I Category II Category II
Medium Category II Category II Category III
Large Category III Category III Category III

The NWA further mandates that an approved professional person (APP) should
conduct dam safety evaluations at dams with a safety risk. An APP is registered in
terms of the Engineering Profession Act (No. 46 of 2000, 2000) and is approved by
the Minister of the Department of Water and Sanitation (DWS).

The Regulations Regarding the Safety of Dams (NWA, No. 36 of 1998. Reg. 139,
2012) requires that Category I dams be evaluated on their safety at intervals of

10
between five and ten years, while Category II and III dams should be evaluated
every five years with a possible extension to a maximum of ten years.

2.3.2 Structural response

Structural functioning of arch dams

In order to effectively manage arch dams, the operational behaviour and response to
the imposed loading needs to be understood. The idealistic force transfer throughout
an arch dam is shown visually in Figure 5. As shown in the image compressive
stresses are located at the upstream face, close to the crest of the dam, and
subsequently the forces are then transferred towards the downstream abutment
contact. The more flexible the cantilevers units of the arch dam wall, the greater the
arch action. The degree of double curvature (curved in plan and section views) also
increases the horizontal force transfer to the abutments, whilst pushing the dominant
compressive forces lower on the upstream face, away from the crest (in elevation).
For a well designed arch which exhibits dominant arch action to distribute the
loading, the maximum toe stress may be found as high as a third of the dam height
during operational loading, as opposed to the maximum toe stress being located at
the riverbed. In the case where the cantilevers are stiffer, the horizontal arching zone
becomes much shallower, while the forces tend to transfer more to the central
foundation via cantilever action (Shaw, 2015: 325).

Figure 5: Schematic of force transfer within an arch dam (Shaw, 2015: 324)

11
The examples in Figure 6 show the minimum principal stress (compression) vectors
for a thin double curvature arch which exhibits effective arching (left), and an
arch/gravity dam (right) which utilizes larger cantilever forces to resist the loading.

Figure 6: Typical stress distributions for thin arch and arch/gravity dams (Shaw, 2015: 325,
327)

The thin arch displays significant arch action throughout its full height along the
upstream face, which peaks at approximately two thirds of the height of the dam.
The thrust then transfers toward the downstream face, resulting in maximum
compression forces at approximately one third of the dam height along the abutment.

The arching on the upstream face for the arch/gravity dam is considerably less
compared to the thin arch. Little to no arching occurs close to the heel of the dam,
with these forces predominantly resisted via cantilever action. The maximum
compression forces on the downstream are located within the riverbed, as opposed
to higher up along the abutment flanks.

Dynamic response
Dynamic systems are defined as structures which are acted on by time varying
forces. The structures response to this dynamic loading is oscillating displacements
or vibrations. Once a system becomes dynamic certain parameters start having
more significant effects on the system's response. These parameters include mass
(m), stiffness (k), and damping (c) (Moyo, 2013: 9). Figure 7 describes the dynamic
system configuration. The mass, stiffness, damping, and natural frequency (ωn)
collectively constitute the modal parameters of a system. These parameters control
the system's behaviour. For a given input or excitation imposed on the system a
corresponding response or output will result.

12
 Figure 7: Dynamic system configuration

Damping corresponds to the energy dissipation within a given structure or system

and is summarized for concrete arch dams in Table 5.

Table 5: Damping in concrete arch dams (USACE, 1994: 7-11)

Location of Damping Type of Damping Description

Internal friction within:

Concrete Arch Friction  Concrete material
 Construction joints

Propagation of elastic waves away from

Radiation
Foundation Rock the dam wall
Hysteretic Sliding along cracks/fissures
Absorption of the energy by:
Reservoir Refraction  Reservoir bottom
 Upstream water body

Dynamic systems can be simplified as single or multiple degree of freedom

(SDOF/MDOF) systems depending on the complexity of the system. A degree of
freedom (DOF) represents the number of independent coordinates in space required
to describe the motion of the system (rotations and displacements in particular
directions) (Moyo, 2013: 7).

The equation below represents the basic equation of motion of a damped, SDOF
system subject to an external dynamic force.

̈ ̇ (2.3)

me: Equivalent mass

c: Damping
ke: Equivalent stiffness
f0: Amplitude of excitation force
ω: Excitation frequency
t: Time

As the excitation frequency approaches the natural frequency, the system

approaches resonance and the dynamic response is controlled primarily by the
damping. At resonance the displacements of the structure in response to the
dynamic loading becomes significantly amplified. Therefore it is the excitation

13
frequency and not the amplitude of the excitation force which is the significant factor
when concerned with the magnitude of the response of a dynamic system (Moyo,
2013: 56).

Dams in Southern Africa are rarely exposed to large dynamic excitation, often
caused as a result of earthquakes, compared to world standards (Brandt, 2011: 6),
but are continually subject to micro-seismic loading during operation. Therefore the
structural system can be simplified as a damped, free vibration response system with
the input or excitation force being the ambient vibrations recorded on the structure
during operational conditions (Moyo, 2014: 2). During large dynamic excitation the
excitation frequency may approach the natural frequency of the structure and cause
resonance. Coupled with large excitation forces the response of the structure may
become exaggerated. In this case it would be more appropriate to represent the
system as a damped, forced vibration response system (Moyo, 2013: 53).

Structural Health Monitoring

Comprehensive structural health monitoring (SHM) systems have been intensified in
recent years especially within the dam industry. This is due to the aging
infrastructure, loss of institutional knowledge, and significant economic and social
implications associated with failure, especially for large dams with high hazard
potential. The objective of SHM systems is to track the structure's integrity over time,
and to assess any damage (Chang, et al., 2003: 257) or change in structural
parameters as a result of remedial works. The process involves response data
acquisition through relevant instrumentation, analysis, interpretation of this data, and
storing the results for future reference and comparison.

Ambient vibration monitoring (AVM) is a form of SHM which aims to determine the
dynamic properties of a structure. The ambient vibrations recorded on a dam
structure have been shown to exhibit similar response signatures when compared to
the larger, macro excitations caused by earthquakes and forced vibration tests
(Okuma et. al., 2008: 8). The significance of this type of testing is that it gives insight
into the dynamic system of the actual structure in operation. Therefore, AVM is often
referred to as Operational Modal Analysis (OMA) (Moyo, 2014: 2). This method is
opposed to forced vibration testing which stimulates the structure artificially. For
large, stiff civil structures such as dams, it is difficult and expensive to artificially
stimulate the structure. Therefore, the benefits associated with AVM are that it is
economical, convenient, and records data which is representative of the as-built
structure during operation (Ibid.: 32).

The dynamic system consists of dynamic modal parameters and structural

parameters. The dynamic parameters include natural frequency, mode shapes (φ),
and damping ratio (ζ), which are extracted from the data through data processing
techniques. The natural frequency is related to the structural parameters of stiffness
and mass as defined in the equation below (Moyo, 2013: 21).

14
 √ (2.4)

Once the system has been defined for a period of time the structural performance
can be assessed, tracked over time, and used to develop calibrated finite element
models. Shifts in the natural frequencies and in the mode shapes may be indicative
of changes to the structural parameters, namely the stiffness of the structure.
However, modal parameter shifts are also sensitive to changes in temperature,
moisture, and other environmental factors (Chang et al., 2003, 258). Therefore, the
interpretation of the recorded data requires significant analysis by experienced
engineers who have a good understanding of the structure. This will enable the
tracking of trends in the behaviour which may potentially be linked to deterioration
and impending failure.

Theoretically the ambient vibrations result in broadband excitation causing all modes
of vibration to be excited. However in reality the input has a certain spectral
distribution due to differences in weighting of specific modes (Moyo, 2014: 6). All of
the structure's modes may not be excited at any given moment and this emphasizes
the importance of accumulating AVM records over many periods throughout the
operational life of the structure.

2.3.3 Failure modes

Dam safety is defined as a dam's "adequacy against an uncontrolled release of
reservoir water" (FERC, 1999:11-2). The result of uncontrolled release of impounded
water is downstream flooding which can have catastrophic effects on life, property,
and infrastructure.

Operational failure modes occur during normal operation of the dam. The causes
range from human error to mechanical malfunction. These failure modes are
summarised in Table 6.

Dams may also fail due to the integrity of the structure or the founding conditions not
being able to resist the imposed loads acting on the structure. The major structural
failure modes associated with concrete arch dams are discussed in the following
sections.

15
Table 6: Operational Failure Modes of Dams

Cause Mode of failure

Negligence in terms of outflow Over-topping and/or large out-
Human Error
release and water level within dam flow
Lack of lubrication of moving Faulty operation of gates and
Maintenance mechanical parts outlet works leading to
Fatigue fracture and loose bolts in overtopping and/or large out-
gates flow
Gates fail to open Overtopping
Mechanical Gates open inadvertently Large outflow
Malfunction Outlet works such as valves fail to
Overtopping
operate
No warning downstream in
case of large outflow
Technological
Communication breakdown No warning of large inflow from
Malfunction
upstream leading to
overtopping
Reduced storage Overtopping and/or large out-
Siltation
capacity flow
Debris plugs spillway (log boom
Reduced release
failure)
capacity Overtopping
Ice attack on outlet works
Loss of access to operate gates
Over pumping of upper reservoir in
pumped storage project Overtopping and/or large out-
Other
Loss of power supply resulting in flow
inoperable gates

Concrete Over-Stressing
Concrete will fail in either compression (crushing) or tension (cracking) when the
loading causes stresses which exceed the capacity of the concrete. As the load on
the structure is increased to capacity, internal micro-cracks develop into stable larger
cracks and the stress strain behaviour varies from linear to non-linear. This results in
permanent deformation and strain of the concrete. When the concrete is stressed
beyond its capacity the cracks become unstable, fracture, and ultimately leads to
material failure (Gillian et. al., 2011: 175).

Vertical construction joints exhibit little to no tensile strength and in the event of
significant seismic excitation these joints become vulnerable. Earthquakes impose
seismic tensile arch stresses at high frequencies on the dam wall. This rapidly opens
and closes the construction joints when the limited tensile capacity of the joints is
exceeded. These stresses also tend to increase the tensile cantilever stresses
throughout the dam wall. When the tensile capacity is exceeded by this tensile
cantilever stress a horizontal joint forms and is most likely to occur at the relatively
weaker construction lift joints (cold joints). The heel of the dam wall is an area that is
vulnerable to cracking failure due to the maximum tensile stresses which exist during
the operational hydraulic loading. The combination of the open vertical joints and

16
newly formed horizontal joints may lead to the formation of unstable blocks (See
Figure 8). In the case where a block fails, the stabilizing compression (arch) stress
within the dam can no longer be effectively transferred to the foundations and the
remaining structure is compromised and at risk to failure. Slip planes or plastic
displacement of the foundations are similar to the abovementioned case. In these
cases instead of the seismic tensile arch stresses, the displacement may potentially
cause the vertical construction joints to open (Ghanaat, 2004: 4).

Figure 8: Unstable blocks during seismic loading (Ibid.)

Alkali Silica Reaction (ASR) causes swelling of the dam wall predominantly on the
upstream face (See Section 2.3.4). Depending on the severity of the deterioration
there is a reduction in stiffness, strength, and modulus of elasticity at the affected
area (Oberholster, 2009: 211). In effect this reduces the cross sectional area of the
dam. This could have significant structural implications especially if the structure is a
thin arch. The swelling also results in a re-orientation of the forces within the arch.
Instead of acting perpendicular to the wall-foundation interface on the downstream
face, the principal stresses re-orientate into a more parallel orientation (Oosthuizen,
2004). As a result the dam resists loading in a manner it was not designed for and
causes the arch to transfer the forces in a way which may initiate sliding in areas
where the foundation or dam wall is more vulnerable. Once the resistance of the
compromised structure is overcome by the loading, cracks and hinges will develop
and could lead to failure (Ibid.).

Sliding Instability
The material characteristics of the concrete dam wall and foundation are notably
different. This interface acts as a discontinuity with a potential for sliding failure. Arch
dams transfer a significant portion of the imposed loads to the abutments as hoop
stresses, therefore the structure is essentially 'wedged' into the foundations (Ibid.).
This lowers the likelihood of central foundation failure especially in narrow-V
topographies. However certain aspects do make arch dams more susceptible to this
mode of failure.

17
Arch dams built within wider topographies tend to develop dominant cantilever action
in resisting the imposed loads (Shaw, 2015: 324). As a result the central foundation
becomes much more stressed as the maximum compression forces develop at the
toe of the dam wall. In the case where the dam wall-foundation bond is inadequate to
resist the developed stress, sliding failure may occur.

The degree of slenderness of a dam influences the structures susceptibility to

cracking at the dam-foundation interface. Lombardi and Fanelli (1992) propose that
the slenderness and height of a dam relates to the total shear resistance of the dam-
foundation interface. The greater the slenderness ratio (as defined by Lombardi) and
height of the arch dam, the lower the shear capacity of the structure.

Pore pressure and uplift forces are as a result of the interaction between the dam
wall, impounded water, tail water, and foundation. As concrete is a permeable
material, water permeates through the voids following the phreatic profile. The water
then pressurizes within these voids resulting in a tensile stress gradient within the
concrete. Uplift forces develop when this pressure acts normal to a discontinuity
(Oosthuizen, 2014). High pore pressures are often caused as a result of inadequate
drainage. The discontinuities may be in the form of;

i. Sub-vertical cracks at the heel of the dam

ii. Lift joints within the dam wall
iii. De-bonded dam wall-foundation interfaces
iv. Joints within the rock foundation

Uplift forces reduce the effective normal stress at the dam foundation interface and
results in reduced frictional resistance. This lowers the stability of the structure and
makes it more susceptible to sliding failure (Ibid.).

Foundation instability may lead to sliding failure in the case where the movement of
certain supporting material would jeopardise the integrity of the structure. Un-
fractured rock exhibits high intact strength. This inherent strength is significantly
compromised by the presence of fractures and discontinuities especially when it is
located in an area where significant forces are present. These discontinuities may be
in the form of joints, faults, bedding planes, or any other geological change within the
foundation (Gillian et. al., 2011: 179). Discontinuities, which may be viewed as
potential failure planes, lower the shear capacity of material and increases the risk of
shear failure. In the event of one potential failure plane, the failure is described as
plane sliding (FERC, 1999: 11-2). When two or more of these planes intersect they
may form wedges and could trigger wedge sliding (Ibid.). High pore pressures and
uplift forces within the foundation will reduce the shear capacity of discontinuities
increasing the likelihood of foundation failure (Ibid.: 11-10).

Certain conditions are required for foundation failure to occur. Firstly, the resultant
shear stress imposed on the joint should be greater than the shear capacity of the

18
joint. The failure plane should be orientated in such a way that the foundation is
capable of moving in a direction which could cause failure (Gillian et. al., 2011: 179).
Therefore the direction of sliding must intersect a free surface in the downstream
direction. Due to the high strength of intact rock, failure becomes more probable the
lower the amount of intact rock required to be sheared in order to induce sliding
failure. Two or more potential failure planes may result in wedges of various sizes
depending on the individual joint orientations and their connectivity with each other.
Once again a free surface is required for abutment instability to be a potential failure
mode. The Malpasset dam failure of 1959 (FERC, 2014: 35) was as a result of
abutment instability (See Figure 9).

Figure 9: Abutment instability causing Malpasset dam failure (Ibid.: 42)

As previously stated the orientation of a discontinuity is important in assessing the

potential of sliding failure as faults naturally have a reduced shear resistance
compared to the intact rock foundation. The permeability of the discontinuity is
important because it could provide a passage for seepage through the dam wall, or
result in abnormally high uplift forces at certain locations in the foundation and this
will decrease the shear stability (Oosthuizen, 2014).

Downstream plunge pools are natural dissipating structures for arch dams. In some
cases where no concrete apron is provided the erosion of the foundation
accompanying the dissipation of energy could impact on the sliding stability of the
structure (Gillian et. al., 2011: 181).

19
2.3.4 Alkali Silica Reaction

Overview of reaction
ASR is a chemical reaction occurring within the pores of the concrete matrix between
alkali and reactive silica minerals to form an expansive alkali-silica gel within and on
the surface of the aggregate. As the silica component originates from certain
aggregates within the concrete the reaction is also referred to as alkali-aggregate
reaction. The gel product absorbs water and swells within the concrete matrix and
cracking will occur once the expansive pressure exceeds the tensile capacity of the
concrete (Oberholster, 2009: 189).

Factors influencing the reaction

In order for the reaction to take place, all of the following factors should be present.
Sufficient alkalinity within the pore solution, sufficient reactive minerals within the
aggregate, and conducive environmental conditions (Ibid.: 190).

Usually the main source of alkalinity contributing to the pore solution originates from
the cement within the concrete. The highly soluble neutral sulphates dissociate in the
solution and increase the concentrations of Na+, K+ and pH levels (Ibid.). Alkalinity
may also be increased due to the products of the hydration process. See Equations
2.5 and 2.6 (Addis, 1998: 73). In the equations below H represents water.

(2.5)

(2.6)

Alkali's are produced from the reaction of calcium hydroxides (Portlandite) with alkali
minerals within the aggregate, which then dissociate in the pore solution. The
calcium hydroxide (and consequent increase in pH levels) also stimulates the
development of a "passivating" layer which protects reinforcement from rusting within
the concrete matrix (Ballim et. al., 2009: 215). In the case of ASR, these alkali's have
a much more detrimental effect as it increases the risk of the swelling reaction.
Alkali's may also be introduced into the system externally by sea water, vapour, and
salt containing ad-mixtures (Oberholster, 2009: 190).

The fine and course aggregate should be dense and contain sufficient amount of
alkali-reactive minerals. The rate of the reaction depends on the amount of reactive
minerals, the reactivity, and the surface area of exposed reactive mineral.
Aggregate which has been identified as potentially alkali reactive within the Eastern
Cape of South Africa are the following;

i. Cape Super Group – Ortho-quartzite of the Table Mountain Group

ii. Enon Formation – Quartzite pebbles
iii. Quaternary Period – Quartzite pebbles (Ibid.: 192)

20
The rate of swelling approximately doubles for every 10°C increase in mean annual
ambient temperature (Ibid.). The reaction also requires the concrete to be moist and
exhibit an internal relative humidity of 75-85% or higher (Ibid.). Daily and seasonal
fluctuations in temperature and moisture increase the number and width of cracks
associated with ASR.

Indicative features of ASR

The swelling effect associated with ASR causes internal stress which results in
cracking if the tensile capacity of the concrete is exceeded. The cracks are relatively
large (compared to shrinkage cracks), with the pattern for unrestrained concrete
being random. This is often referred to as "map" cracking (Ibid.: 205) and is shown in
Figure 10.

Figure 10: Classic ASR induced map cracking (Poraver, 2016)

Restrained or load bearing concrete structural elements significantly influence the

crack pattern. The macro-cracking tends to be parallel to the direction of constraint.
For example, columns are unrestrained laterally therefore the concrete swells in this
direction and causes vertical cracks or increases existing structural cracks
(Oberholster, 2009: 205). In concrete arch dams the concrete expansion caused by
ASR also results in the closure of joints, warping of concrete and misalignment of
certain effected elements (Ibid.).

The ASR product or gel may be visible on the surface cracks of the concrete and
often form white stains on the perimeter of the aggregate (See Figure 11). Even
though the cracks are caused by the reaction, the voids are sealed due to the
leaching of the gel product. This limits the ingress of other substances through the
cracks.

21
 Figure 11: Cored sample of ASR deteriorated concrete (Concrete Microscope Library, 2016)

Another indicative feature of ASR on the concrete surface can be dark

discolouration, which may be due to dampness, with lighter shades immediately
surrounding the cracks (Oberholster, 2009: 206), as shown in Figure 12.

Figure 12: Discolouration arround ASR related cracks (Ibid.: 207)

Structural effects
ASR may in some cases increase the severity of existing structural cracks and as a
result degrade the structural integrity even further. The extent, rate, and location of
ASR affected areas vary throughout the concrete. For this reason the structural
effects on a given element or structure is case specific.

Table 7 shows the effects of ASR on the mechanical properties of concrete. In

practice these properties should be larger than the figures in the table due to
restraint from adjacent material and bi/tri-axial stress states (Ibid.: 211). It is clear
that the properties do decrease the more the concrete is affected by ASR. The

22
findings also show that tensile strength and stiffness are most significantly reduced
due to the deterioration.
Table 7: Affect of ASR on concrete at 28 days (Oberholster, 2009: 211)

2.3.5 Instrumentation
Surveillance is defined as close observation of a structure to prevent or detect
problems and it encompasses monitoring and instrumentation (Oosthuizen, 2014).
Monitoring uses equipment or instrumentation to gather information such as loading
and response of the structure, to assess performance, and to continue assurance of
the safety of the dam throughout its service life. Therefore surveillance is a process
of systematic review. Instrumentation monitoring programs for dams compliments
dam safety inspections and surveillance programs.

The objective of instrumentation monitoring programs is to diagnose the state of the

structure, predict its behaviour, and identify potential failure modes. The
instrumentation implemented should answer specific questions that take into account
the dam structure type, site specific geotechnical conditions, and environment
conditions (Ibid.). The diagnostic aspect is to verify design parameters and to identify
the causes of adverse events. Observations from monitoring systems can shed light
on unknown variables and assumptions made in the design phase. Therefore it
allows for refinement of the analysis of the existing structure and improves future
designs. Instrumentation also allows the cause of adverse events to be identified
through the recorded data and remedial work effectiveness can also be tracked with
this data (Moyo, 2014: 2).

Parameters which may need to be monitored in concrete dams include:

i. Displacements
ii. Cracks
iii. Settlement
iv. Seepage
v. Stress and strains
vi. Uplift pressures within foundation
vii. Integrity shifts (Natural frequencies)

23
2.4 Chapter Summary
This chapter reviews arch dams from the design stage to construction, and the
management of the structure over time. Each section gives insight into aspects
integral in understanding how the structure was designed to resist the imposed
loading, and the reasons for deviations from this behaviour.

The design of arch dams is largely dependent on the specific site conditions.
Competent flank foundations are required as the structure resists the loading to
some extent via arch action. Therefore the strength of the concrete is important as
the loading is resisted structurally and to a lesser extent by mass. FEA have been
used in recent times in the design of new arch dams as well as in the analyses of
existing structures. By mathematically modelling the structure and loading conditions
accurately enough, a realistic response of the structure can be simulated (Shaw,
2015: 321). It is also a particularly useful tool for assessing remedial alternatives
during rehabilitation of existing dams.

The correct construction practice is important in order to materialise the design. The
main construction considerations are highlighted such as foundation preparation and
grouting. It is essential that the arch is grouted adequately to ensure that the
structure acts as a monolith as opposed to individual cantilevers.

Responsible management of large dams is a statutory requirement and calls for a

deep understanding of the structural system. Dam safety engineers need to identify
potential failure modes, understand the structural response, and consider the effects
of deterioration such as ASR. Monitoring systems are crucial tools which give insight
into the safety evaluation of dams.

By exploring the various stages associated with arch dams a better understanding of
the structure and its response to changing loading over time can be achieved.

24
 CHAPTER 3
3 METHODOLOGY

3.1 Introduction
This chapter focuses on Kouga Dam as the case study of the thesis. The chapter
begins with a brief background of the structure followed by a review of previous dam
safety evaluations where the important findings and foundation conditions are noted.
Lastly the chapter describes in detail the monitoring system installed at the dam.

3.2 Kouga Dam

3.2.1 Locality
The locality of Kouga Dam is shown in Figure 13.

Figure 13: Locality of Kouga Dam (Google Earth, 2013)

25
3.2.2 Dam Characteristics
Name of dam Kouga Dam3
Locality number: L820-01
River: Kouga River
Nearest town: Patensie
Distance to nearest town: 27km West
Province: Eastern Cape
Latitude: 33°44’30”
Longitude: 24°35’15”
Wall type: Double curvature concrete arch
Crest length: 317m
Wall height (LFL to NOC): 69m
Wall thickness (FSL): 6.93m
Wall thickness (NOC): 6.43m
Level of UNOC: RL169.164m
Level of NOC: RL156.972m
Level of FSL: RL149.352m
Level of gauge plate zero: RL 54.950m
Storage capacity: 128.490x106 m3
FSL area: 554.5 ha
Purpose: Irrigation, domestic, and industrial use
Owner: DWS
Designer: DWS
Contractor: DWS
Completed: 1969
(Dam Safety Office [DSO], 2011)

3
Formerly known as Tweerivieren Dam and Paul Sauer Dam

26
3.2.3 Background

General
Kouga Dam is located approximately 27km West of Patensie, on the Kouga River in
the Eastern Cape. The dam forms part of the Gamtoos River Government Water
Scheme (GRGWS). It was designed, constructed, and owned by the then
Department of Water Affairs (DWA), now known as the Department of Water and
Sanitation. Construction on the dam was completed in 1969 (DSO, 2011).

The dam is a double curvature, concrete arch dam (See Figure 14) and was the first
of its kind to be built in South Africa. It supplies water primarily for irrigation purposes
to the Gamtoos River Valley and also supplements the water supply to the Nelson
Mandela Metropolitan. The initial design made provision for the possibility of future
raising in order to increase the capacity and provide additional water supply to Port
Elizabeth. The dam has an approximate height of 69m and crest length of 317m. The
spillway system comprises of a centrally located, uncontrolled ogee spillway and a
chute spillway on the left flank which is controlled by two radial gates. The outlet
works releases into a concrete channel which is situated on the right flank and closer
to the left flank a scour outlet can be found on the concrete lined apron. A post-
tensioned concrete stress pad has been provided on the right flank and provides
additional stiffness and restraint against sliding failure. The dam was constructed on
a pulvino pad which is a thicker layer of concrete which assists in distributing the
forces from the dam wall to the foundations. Kouga Dam is classified as a Category
III dam due to its high hazard potential rating and size class.

Figure 14: Kouga Dam

27
Alkali Silica Reaction
Dam safety evaluations and laboratory tests have identified the presence of ASR at
Kouga Dam. The dam structure and the surroundings exhibit the three main factors
which contribute to the ASR. These factors include high levels of alkalinity within the
cement used, reactive minerals associated with the aggregate, and conducive
environmental conditions.

The cement used during construction (1960's and 1970's) originated from Port
Elizabeth and exhibited a Na2O equivalent of greater than 0.8% (Elges et. al., 1995:
2). Oberholster (2009: 191) states that a Na2O equivalent of greater than 0.6% is
regarded as a high alkali cement which can contribute to the presence of ASR. In
addition the entire catchment of the dam also consists of quartzitic sandstone which
are identified as potentially alkali reactive (See Section 2.3.4). The alluvial deposits,
which consist of the same quartzitic sandstone, in the Kouga River were processed
and used as course and fine aggregate for the concrete during construction. The
orientation and conditions on the structure are also conducive to ASR. The ASR is
likely to occur on the upstream face of the dam wall because this is the interface
which has constant exposure to moisture. The orientation of the upstream face is
north facing and as a result is exposed to direct sunlight for most of the day. This
translates into high temperatures that fluctuate daily and which increase the rate of
the reaction.

After 1981 tests were conducted in order to verify ASR on the structure. The tests
were conducted on concrete cores drilled from the structure and included
petrographic examinations, X-ray diffraction analysis, scanning electron-micrograph
and standardless Energy Dispersive X-ray (EDX) analysis, drying shrinkage, and
wetting expansion tests (Elges et. al., 1995: 4). ASR was confirmed from the tested
samples and to date no further examinations were done to determine if the reaction
has stabilized.

3.2.4 Geology
The rock which the dam is founded on is characterized by stratified and jointed
quartzitic sandstone of the Table Mountain Group (TMG). The bedding planes dip
approximately 50° downstream and slightly towards the left flank (See Figure 15).
Between this competent rock there are inter-bedded layers of weaker, consolidated
phyllite which generally vary in thickness, up to a maximum of approximately 0.3m
(Hobbs et. al., 1967: 257). This phyllite material deteriorates rapidly when saturated
and exposed to air. Hobbs (1967: 257) also noted that boreholes drilled in the outlet
tunnel showed that the quality of the foundation improved considerably with depth.

28
 Figure 15: Close-up of right flank abutment (Hobbs et. al., 1967: 255)

3.2.5 Right flank foundation conditions

General
Forbes Dick (1989) concludes that the left and central foundations of Kouga are
stable due to the favourable orientation of the strata with relation to the arch thrusts.
The arch forces on the left flank act almost perpendicular to the strike of the bedding
planes and this keys the arch into the abutment.

The right flank of the arch abuts into a spur on the mountain (See Figure 16).
Upstream of the spur is a gully and on the downstream is a deep embayment. The
embayment results in less support against the arch thrust. During excavation of the
right flank foundation phyllite seams were encountered which rapidly deteriorated.
This raised further concerns regarding the integrity of the foundations.

Figure 16: Right flank prior to completion and stress pad remedial measures (Hobbs et. al.,
1967: 654)

29
Remedial measures
Unfavourable conditions were identified on the right flank foundation during
construction, and as a result remedial measures were undertaken to improve the
foundation conditions and to ensure the stability of the right flank. These remedial
measures were also undertaken in light of the Malpasset Dam failure noted in
Section 2.3.3. Hobbs (1967: 660-665) notes the following remedial measures
undertaken at the dam;

i. Treatment of phyllite seams

ii. Provision of extensive concrete apron
iii. Foundation grouting
iv. Increasing the radii of the arch rings
v. Reinforced concrete thrust pad
vi. Capping beam with slab on surface of right flank
vii. Drainage and grout curtains

The primary function of the capping beam (See Figure 17) is to act as a thrust block
and to redistribute the resultant forces from the upper arch into the right abutment as
parallel to the strike of the strata as possible. This allows more bedding planes to
contribute in resisting the thrust. Foliation or lifting of the strata is reduced the more
parallel the thrust is relative to the strata. This lowers the force perpendicular to the
bedding planes.

The stressing slab is anchored into the right abutment with 62 high tensile steel
cables. These cables are anchored to a depth of 40m and post-tensioned to 245
tonnes. The anchoring confines the layers of strata together into a monolith and as a
result deepens the load carrying rock mass. This further reduces the tendency of
buckling failure of the strata under the arch thrust load.

Figure 17: Capping beam and stressing slab on right flank (Hobbs et. al., 1967: 664,650)

30
3.2.6 Loading on right flank foundation

Potential failure mode

Figure 18 shows the stratification of the foundation and arch thrusts of the structure.
Forbes Dick (1989: 13) notes that the most probable potential failure mode in the
direction of thrusting on the right flank is shear across the bedding planes in
combination with tensile foliation of the phyllite layers. The bedding layers act as
cantilever slabs that are stacked on each other, with the fixed end within the
abutment and the free end closest to the dam. Movements within the foundation on
the right flank may be initiated by deteriorating phyllite layers and produce near
perpendicular forces relative to the bedding planes, acting in a downstream direction.
This would result in excessive bending moments on the bedding layers, cause the
layers to buckle under the load, and initiate a shear failure downstream.

Figure 18: Arch and cantilever reactions on the rock foundation in relation to the stratification
dip and strike directions (Forbes Dick, 1989: 14)

Arch thrusts
Due to the valley shape and arch design the structure transfers the imposed loading
to a larger extent via cantilever action to the central and lower flank foundations (The
transfer of forces to the foundation is discussed in Section 4.4). The magnitude and
dip of the resultant forces decrease from the centre of the arch to the abutments
(See Figure 19) with the orientation tending towards the right abutments.

31
The forces which could initiate shear failure occur adjacent to the spillway due to its
large resultant and downstream components. The thrust has a similar dip to the
bedding planes and is directed towards the right abutment well below the power
station. The thrusts closer to the abutment are less favourable in terms of the
orientation as the resultant forces are near horizontal. The resultant and downward
components are significantly lower compared to that closer to the spillway. The
upper arch thrusts on the capping beam and assists in redirecting the forces along
the strata below the stress pad.

Figure 19: Resultant forces on the right foundation (Forbes Dick, 1989: 16)

The arch thrusts on the right flank and adjacent to the spillway seem favourable due
to the strata which keys in the arch. Closer to the abutment the capping beam and
anchored stress slab reduce the risk of sliding failure due the redirection of the
thrusts to displace a much larger rock mass in order to induce failure. At present the
embayment is not a concern with regards to a lack of abutting mass, due to the
direction of the resultant forces of the wall being well into the foundation in a mostly
downward direction. The effect of ASR on the re-orientation of arch thrusts may be a
cause for concern and should be investigated in order to assess the safety of the
right foundation.

32
3.3 Monitoring System

3.3.1 Overview
The previous dam safety evaluations at Kouga dam have raised concerns regarding
the potential instability of the right abutment and the aforementioned ASR. As a
result the monitoring and surveillance of the dam has intensified. The following
parameters were investigated and are summarized in Table 8.
Table 8: Investigated parameters

Type Parameters Description Instrumentation

Hydraulic Water level Gauge plates
Load Temperature Ambient, water, concrete Thermometers
Dynamic AVM Accelerometers
Geodetic surveys
Global Navigation Satellite
Dam wall face and crest System (GNSS)/Global
Response Displacements
Positioning System (GPS)
Crack width gauges
Dam wall and foundation Trivec systems
Dynamic Natural frequencies and mode
Integrity Processing of AV data
response shapes

3.3.2 Water Levels

A gauge plate installed at the dam is used to record the water level on a daily basis.
The gauge plate level refers to the height at which the water surface is above the
lowest inlet level. This gauge reading can be converted to the corresponding
reduced level above mean sea level and volume of water within the dam.

3.3.3 Temperature
A total of four thermometers have been installed on the upstream face of the dam
wall. Two closer to the left flank and two closer to the right flank. Figure 20 shows the
approximate locations (red dots) of the thermometers. The thermometers are
estimated to have been installed at the gallery level. Temperature readings are
recorded 4 times per day (6-hourly).

33
 Figure 20: Top view showing locations of thermometers on upstream face

3.3.4 Displacements

Geodetic Network
The geodetic system uses a triangulation network of beacons and targets on the
downstream face of the dam wall and is complimented by precise traverse
measurements within the gallery. The final results are three dimensional
displacements of the monitoring points. The accuracy in the x and y directions are
approximately 0.3-0.5mm while the z (vertical) measurements are 0.1-0.2mm
(Pretorius et. al., 2001: 210).

Figure 21 shows the location of measurement targets for geodetic surveys. The
geodetic surveys are carried out twice a year, once during summer (March) and
again during winter (September). The points investigated are all positioned at
RL144.780m (gallery level) and measure displacements in three directions.

Three positions were investigated in detail and these are located at two quarter
points and the mid-point of the dam wall. The right flank quarter point is represented
by traverse station 111, the midpoint stations are 116 and 117, and the left flank
quarter points are 122 and 123. The recorded displacements are referenced to a
baseline date (15/12/1972) for comparison purposes.

34
 Figure 21: Downstream general elevation of monitoring points

GNSS/GPS
The GNSS/GPS is an automatic monitoring system that operates continuously to
provide deformation information in the x, y, and z directions. It records displacements
at a logging rate of 1 hour and allows the behaviour of the structure to be observed
in much more detail compared to the geodetic survey data which only allows for data
points every six months.

The installation layout includes two primary base sensors in the surrounding area of
the dam structure. These include two secondary sensors on the dam structure and a
server located in the outlet house. The base stations are used as reference stations
for the relative positioning during post-processing. The two secondary sensors are
positioned adjacent to the overspill section on the dam wall. The layout is shown
graphically in Figure 22.

Figure 22: Kouga Dam GNSS/GPS Network (Pretorius & du Toit, 2014: 8)

35
The four Trimble Net R5 receivers and Trimble Zephyr Geodetic antennae are used
to make the observations (See Figure 23). The features of the GNSS/GPS system
are summarised below;

 Period of Observation: 24 hours

 Sampling rate: 10 seconds
 Tracking Interval: 1 Hz
 Continuous Logging Rate: 1 hour
 Minimum Number of Satellites: 5
 Satellite Elevation Mask: 5° (Ibid.: 6)

Figure 23: GNSS/GPS System at Kouga Dam (Ibid.: 4)

Crack width gauges

The then Department of Water affairs and Forestry (DWAF) developed in-house
crack width measuring devices in the 1990's which were implemented throughout the
country (Oosthuizen et.al, 2003: 600). The three dimensional crack width gauges are
still used at Kouga and many other dams in South Africa despite a national upgrade
to automatic crack width gauge systems. The device allows the crack widths
(translations) and relative rotations (tilt) to be measured across cracks or joints.
Sinco 3-D crack gauges from the USA have been installed within the gallery of
Kouga dam, while the DWAF94 gauges are used to record the crack and joint
displacements throughout the rest of the dam. The readings are manually recorded
on a monthly basis with electronic digital indicator gauges. The accuracy of both
crack width gauge types in the x (across the crack or tangential direction), y
(vertical), and z (radial) directions is 0.02mm for all translations

For the purpose of this study the three dimensional crack width gauges located on
the downstream face construction joints and gallery of the dam wall were
investigated.

36
Trivec
The Trivec system measures the relative displacements in three directions (x, y, and
z) at 1m intervals along a vertical borehole. Unplasticized polyvinyl chloride (uPVC)
tubes which are connected with stainless steel measuring couplings are installed
within the borehole and grouted into position. A measuring probe is lowered into the
casings and records the displacements in all three directions at the various
measuring points. The first set of readings acts as the baseline data set to which the
future readings are compared. The measurements are done bi-annually, once during
summer and another during winter. The equipment used for the system is shown in
Figure 24.

Figure 24: Trivec system components (Naude, 2002: 4)

At first four probes were first installed at the dam (Kogr1-4) with the baseline
readings taken between 2001 and 2002. In 2013 two more probes were installed.
The majority of the Trivec systems are located on the right flank due to the concerns
about the foundation stability. The locations of the Trivec systems are shown in
Figure 25.

37
 Figure 25: Locations of Trivec systems

Kogr1 is the only probe which has been installed through the dam wall and it extends
approximately 25m into the foundation. Kogr2 and Kogr5 are located on the right
flank and Kogr4 is on the left flank. Kogr3 and Kogr6 were installed in the lowest
main inspection tunnel on the right flank. The installation level (highest measuring
point on Trivec) and height of the boreholes are given in Table 9.
Table 9: Installation levels and height of Trivec systems

Trivec System RL (m) Height (m)

Kogr1 156.9 87.2
Kogr2 144.8 47.6
Kogr3 103.6 33.5
Kogr4 140.2 54.0
Kogr5 99.1 47.0
Kogr6 103.6 49.0

3.3.5 Ambient Vibration Measurements

Operational modal analysis (OMA) is used to extract the dynamic modal parameters
during ambient conditions. The modal parameters include natural frequencies, mode
shapes, and damping ratios. This method is often used on large civil structures
because of the relatively low cost and the recorded measurements are comparable
with that of forced vibration tests. The ambient loads that act on the dam during
operation include, but are not limited to the following; wave action, wind, atmospheric
pressure, and human activities.

38
Instrumentation
Kouga dam has an extensive AVM system in place to record the ambient vibrations.
Accelerometers that are permanently positioned within the gallery measure the
vibrations which are then recorded by a data acquisition system. Post-processing is
done in order to extract the modal parameters from the recorded data. The
equipment used at Kouga dam to record the ambient vibrations are listed below.

 Accelerometers

The permanent AVM system comprised of AC-2x tri-axial (Figure 26) accelerometer
sensors manufactured by GeoSIG is used to measure the ambient vibrations acting
on the structure. Some of the specifications associated with this accelerometer
includes; bandwidth of 0.1-100Hz, dynamic range of >125 dB, and an accuracy of
±0.4 dB. The roving system used Q-Flex QA-700 accelerometers manufactured by
Honeywell. Some of the specifications associated with this accelerometer includes;
bandwidth of >300Hz, dynamic input range of approximately 30g, and resolution of
<1µg (GeoSIG, 2010: 1, 2).

Figure 26: Tri-axial accelerometer sensor (GeoSIG, 2010: 1)

 Data Acquisition System

The CR-5P monitoring system (Figure 27) is also manufactured by GeoSIG and was
used to monitor and record the accelerations of the structure. The dynamic seismic
data logger digitized the data to a 24 bit resolution and at a sampling rate of up to
250 samples per second. The raw data is then downloaded to a computer (Figure
28) for processing and modal extraction.

39
 Figure 27: Data Acquisition System

Figure 28: The computer used for data acquisition and processing

Test Layout and Procedure

The dynamic measurements were performed using two set-ups. One set-up was a
permanent system which had a total of 21 tri-axial accelerometers with a fixed
reference location. The other set-up used only six accelerometers and the travelling
accelerometer method in order to cover all the blocks. The two systems were used in
order to verify the recorded data. The accelerometers were placed at the centre of
blocks 106 to 126 within the gallery of the dam (See Figure 29). The permanent
accelerometers were housed inside survey brackets which were mounted on the

40
upstream side of the gallery wall while the roving accelerometers were placed on top
of the brackets (See Figure 30).

Figure 29: Location of accelerometers within gallery

Figure 30: Roving and permanent accelerometers

The measurements were performed over three periods with the first in August 2011,
the second in February 2014, and the third in March 2014. Trial measurements were
done in order to test the system and to verify whether the data recorded was a
realistic sample. For the roving set-up the data was recorded for periods which
varied from 20-60 minutes. The permanent set-up recorded data continuously and
only required the data to be downloaded from the data acquisition system in order to
perform the data processing. A sampling frequency of 1000Hz was used for the
measurements.

41
Data Processing
The dynamic modal parameters were extracted from the recorded acceleration data
using ARTeMIS Extractor Pro 2010. The Enhanced Frequency Domain
Decomposition (EFDD) method was used in order to determine the associated
natural frequencies of the structure.

3.4 Chapter Summary

This chapter introduces Kouga dam as the case study of the thesis and provides a
summary of the foundation conditions, loading on the structure, and ASR identified at
the dam. In addition previous dam safety evaluations which have highlighted the
instability of the right abutment as a potential failure mode are explored in greater
detail.

The chapter outlines the monitoring system at Kouga dam and highlights how it has
intensified over time as a result of the aspects identified in past dam safety
evaluations. This chapter then discusses the layout of the instrumentation and
measurement procedure associated with each instrument. The monitoring system
records the imposed loading, structural response, and integrity of the structure. The
parameters measured at Kouga dam and considered in this thesis include water
level, temperature, displacement, and ambient vibrations.

42
 CHAPTER 4
4 RESULTS AND DISCUSSION

4.1 Introduction
Monitoring systems are integral to dam safety surveillance. The chosen
instrumentation is dependent on the type of dam, site specific conditions, and an
understanding of which parameters are important for defining the structural
behaviour and potential modes of failure. The specific parameters investigated at
Kouga dam in order to deduce behavioural trends of the dam wall structure include;
water levels, temperature, displacement data, and AVM.

This chapter presents the instrumentation results recorded at Kouga dam. A brief
study is done with a dam of similar type and size to compare the natural frequencies
and stiffness of the structures. Lastly an analysis is done to determine how Kouga
transfers the forces on the structure to the foundations. The effects of the
topographic conditions on the force transfer of the structure are discussed and a
FEA is presented.

4.2 Instrumentation Data Results

4.2.1 Water Levels

The Eastern Cape experiences a cold and wet winter with minimum temperatures in
August/September and a dry warm summer with maximum temperatures in
February/March. There is cyclic behaviour with regards to the temperatures, rainfall,
and corresponding water levels within the dam. However there are dry periods which
occur for extended periods of time and these continual low water levels have had an
effect on the behaviour of the dam. Figure 31 shows the water record for Kouga dam
from 1972-2015.

43
 Water Level FSL
150 55.598

145 50.598

Gauge Plate Reading (m)

140
45.598
Water Level (m)

135
40.598
130
35.598
125
30.598
120

115 25.598

110 20.598
1972/03/01

1977/03/01

1982/03/01

1987/03/01

1992/03/01

1997/03/01

2007/03/01

2012/03/01

2017/03/01
2002/03/01
Date
Figure 31: Water level record

4.2.2 Temperature
The dam spans approximately from East to West between a narrow valley along the
Kouga River. The upstream side of the dam wall is North-facing and as a result the
un-wetted portion of the upstream wall is exposed to constant and direct sunlight
during the day. Conversely the downstream side of the wall is in constant shade.
Figure 32 shows the position of the dam wall within the valley and the trajectory of
the Sun.

Figure 32: Trajectory of the Sun across the dam wall

44
The recorded data of the thermometers on the left and right flanks were averaged in
order to make the data more manageable. Figure 33 shows the recorded
temperature data.
Right flank average temperature Left flank average temperature
27
26
25
24
23
Temperature (Degrees Celcius)

22
21
20
19
18
17
16
15
14
13
12
2011/03/01

2011/08/31

2012/03/01

2012/08/31

2013/03/02

2013/09/01

2014/03/03

2014/09/02
Date
Figure 33: Average temperature for the left and right flanks

From this figure it is shown that the temperature of the upstream face is cyclic and
changes seasonally. The recorded temperatures reach a maximum during summer
(February/March) and are at its lowest during the winter (August/September)
months. It is also evident that the right flank experiences slightly higher temperatures
compared to the left flank and the average temperature difference is approximately
2.75°C. This could be attributed to the trajectory of the sun and the possibility of a
morning shadow cast on the left flank due to the mountain.

The trends demonstrate that the maximum temperatures have remained constant
(26°C and 23.5°C for the right and left flanks respectively) while the minimum
temperatures have increased over the 3.5 year period. The minimum temperatures
for both data sets have increased by approximately 1.5°C.

4.2.3 Geodetic Data

The relative displacements in the x, y and z directions are shown in Figure 34. The
recorded data is presented with the water record (six monthly moving average) and
sign convention used.

45
 Relative Vertical Displacement (mm) Relative RF/LF Displacement (mm) Relative US/DS Displacement (mm) Water Level (m)

-5
0
5
10
15
25
30
35

20
-50
-40
-30
-20
-10
0
10
20
30
110
115
120
125
130
140
145
150

135

-20
-15
-10
-5
0
5
10
15
20
72/03/01
73/03/01
74/03/01
75/03/01
76/03/01
77/03/01
78/03/01
79/03/01
80/03/01
81/03/01
82/03/01
83/03/01 Block 111

Block 111
84/03/01

Block 111
85/03/01
86/03/01
87/03/01
88/03/01
89/03/01
90/03/01
91/03/01
92/03/01

46
93/03/01
94/03/01

Upwards
Water Level

95/03/01

Downwards
Left Flank

Figure 34: Water record and geodetic data

Up-Stream
Down-Stream

96/03/01

Right Flank
Block 116

Block 116

Block 117
97/03/01
98/03/01
99/03/01
00/03/01
(X) Relative US/DS Displacement

01/03/01
(Y) Relative RF/LF Displacement

(Z) Relative Vertical Displacement

02/03/01
03/03/01
04/03/01

111
05/03/01
06/03/01
07/03/01
08/03/01
Block 122

X
Block 122

116
Block 123
09/03/01
10/03/01
11/03/01

Sign Convention
12/03/01

122

Y
13/03/01
14/03/01
15/03/01
16/03/01
At the end of 1989 a severe drought occurred and reduced the volume of water
within the dam to 10% of full capacity (13 million m3). The following sections
compare displacement trends before and after this drought.

Relative upstream/downstream Displacement (X): Centre point

As expected the upstream and downstream displacements are dependent on the
water level. The severe droughts can be identified in the displacement graphs as the
low extremities and this indicates upstream displacement (negative direction). The
displacement for both quarter and centre points follow a similar signature. However
since 1985 the right flank exhibits 10mm larger seasonal displacement fluctuations
compared to the left flank. The upstream and downstream displacements for the
centre of the dam are shown in Figure 35. In the figure the red line indicates the
drought that occurred in 1989.
30
25
Downstream
20
Relative US/DS Displacement (mm)

15
10
5
0
-5
-10
-15
-20
-25
-30
-35 Upstream
-40
80/03/02

86/03/02

09/03/02

15/03/02
72/03/02
73/03/02
74/03/02
75/03/02
76/03/02
77/03/02
78/03/02
79/03/02

81/03/02
82/03/02
83/03/02
84/03/02
85/03/02

87/03/02
88/03/02
89/03/02
90/03/02
91/03/02
92/03/02
93/03/02
94/03/02
95/03/02
96/03/02
97/03/02
98/03/02
99/03/02
00/03/02
01/03/02
02/03/02
03/03/02
04/03/02
05/03/02
06/03/02
07/03/02
08/03/02

10/03/02
11/03/02
12/03/02
13/03/02
14/03/02

Date
Figure 35: Relative upstream/downstream displacement (116)

The general radial displacement trend has been steadily moving towards the
upstream. This rate has decreased following the drought of 1989.

Relative right flank/left flank Displacement (Y): Right flank quarter point
The right flank and left flank displacement of the centre of the dam wall fluctuates
around 0mm, whereas the right and left quarter points diverge towards its opposite
flank. The tangential displacement of the right flank quarter point is shown in Figure
36.

47
 Relative RF/LF Displacement (mm) Relative RF/LF Displacement (mm)

-8
-6
-4
-2
0
2
4

-16
-14
-12
-10
-8
-6
-4
-2
0
2
4
6
8
10
12
14
16
18
20

72/03/02 72/03/02
73/03/02 73/03/02
74/03/02 74/03/02
75/03/02 75/03/02
76/03/02 76/03/02
77/03/02 77/03/02
78/03/02 78/03/02
79/03/02 79/03/02
80/03/02 80/03/02
81/03/02 81/03/02
82/03/02 82/03/02
83/03/02 83/03/02
84/03/02 84/03/02
85/03/02 85/03/02
86/03/02 86/03/02
87/03/02 87/03/02
88/03/02 88/03/02
89/03/02 89/03/02
90/03/02 90/03/02
91/03/02 91/03/02

decreased following the drought of 1989.

92/03/02 92/03/02

48
93/03/02 93/03/02

Date

Date
Left Flank

94/03/02 94/03/02
Right Flank

95/03/02 95/03/02

Left Flank

Right Flank
96/03/02 96/03/02
97/03/02 97/03/02
98/03/02 98/03/02
99/03/02 99/03/02
00/03/02 00/03/02
01/03/02 01/03/02
02/03/02 02/03/02
03/03/02 03/03/02
04/03/02 04/03/02
05/03/02 05/03/02
06/03/02 06/03/02
07/03/02 07/03/02
08/03/02 08/03/02
09/03/02 09/03/02
10/03/02 10/03/02
11/03/02 11/03/02
12/03/02 12/03/02
Relative right flank/left flank Displacement (Y): Left flank quarter point

13/03/02 13/03/02
14/03/02 14/03/02
The tangential displacement of the left flank quarter point is shown in Figure 37.

15/03/02 15/03/02

Figure 37: Relative right flank/left flank displacement of left flank quarter point (122)
Figure 36: Relative right flank/left flank displacement of right flank quarter point (111)

16/03/02 16/03/02
Block 111 on the right flank tends to displace towards the left flank. This rate has
Block 122 on the left flank tends to displace towards the right flank. This rate has
decreased following the drought of 1989.

The strain rate for the entire period (1973-2014) is larger on the right flank compared
to the left flank (See Table 10). The right flank also exhibits larger seasonal
fluctuations of 6-9mm compared to 4-6mm on the left flank. In part this may be
attributed to the higher average temperature and longer length of the right flank
section of the wall.

Relative Vertical Displacement (Z): Centre point

The displacement for both quarter and centre points follow a similar signature. The
centre of the dam wall exhibits the greatest relative vertical displacement and has a
larger strain rate compared to the flanks. The vertical displacements for the centre of
the dam are shown in Figure 38.

35
Upwards
30
Relative Vertical Displacement (mm)

25

20

15

10

5
Downwards
0

-5
72/03/01
73/03/01
74/03/01
75/03/01
76/03/01
77/03/01
78/03/01
79/03/01
80/03/01
81/03/01
82/03/01
83/03/01
84/03/01
85/03/01
86/03/01
87/03/01
88/03/01
89/03/01
90/03/01

92/03/01
93/03/01
94/03/01
95/03/01
96/03/01
97/03/01
98/03/01
99/03/01
00/03/01
01/03/01
02/03/01
03/03/01
04/03/01
05/03/01
06/03/01
07/03/01
08/03/01
09/03/01
10/03/01

12/03/01
13/03/01
14/03/01
15/03/01
16/03/01
91/03/01

11/03/01

Date
Figure 38: Relative vertical displacement at centre (117)

The seasonal fluctuations have remained constant at approximately 5mm. The dam
tends to displace upwards. The general trend shows that the rate of vertical
displacement has decreased over time. The strain rates in the vertical direction are
considerably larger than that in the tangential direction (See Table 10).

49
Summary
The geodetic targets at the gallery level indicate that the dam is displacing in upward
and upstream directions and the quarter points tending towards opposite flanks.
Table 10 summarizes the results.
Table 10: Summary of displacement and strain rates

Displacement
Rate Strain Rate (µε/annum)
(mm/annum)
Period
US/DS RF/LF RF/LF Vertical
(RF quarter (LF quarter
(Centre) (Centre)
point) point)
1973-1988 -0.8 2.5 -2.2 17.2
1989-2014 -0.2 0.7 -0.9 5.0
1973-2014 -0.3 1.6 -1.3 8.6
Reduction in
rate since 1989 75 72 59 71
(%)

Curtis (2000) notes that the response of ASR-affected arch dams are vertical and
upstream displacements over. Kouga is swelling to a lesser extent in the constrained
direction towards the flanks and considerably more along the unrestrained
boundaries in the upward and upstream directions. The average permanent
displacements (in the x-y plane) exhibited by the dam are shown in the schematic
below (Figure 39).

Figure 39: Schematic of permanent x-y displacement of dam wall

The whole structure has moved upstream with the right flank undergoing significantly
larger relative displacement. The additional stiffness on the left flank could be
attributed to the favourable orientation of the foundation bedding planes, parking
area, and spillway chute. It appears that due to the increased axial load associated
with the swelling the structure may be buckling in an attempt to relieve the swelling
stresses. This is due to the foundations being more rigid than the "softer" dam wall
structure and this provides a fixed restraint against the swelling load that occurs

50
within the concrete. The dam wall is not symmetrical and the buckling is occurring at
the slightly longer section on the right portion of the wall.

Previously it has been assumed that until 1976 the swelling may have closed up the
vertical contraction joints of the dam (Elges et. al., 1995: 3), therefore during this
early period (1972-1977) the dam does not show any clear swelling evidence in the
vertical direction. The strain rates in all directions have steadily decreased over the
years and are expected to plateau in the near future. This indicates that the ASR has
appeared to stabilize. The fact that this decrease in strain rate occurred after the
major drought indicates that the dam may have permanently displaced into a position
which manages the stresses in a more effective manner.

The geodetic data provides evidence of general seasonal variations and long term
trends. However, the small temperature sample does not provide enough data to
make confident conclusions on its effect on the structure's response.

4.2.4 GNSS/GPS Data

The GNSS/GPS has a higher resolution when compared to the geodetic data and
this enables the behaviour of the structure to be observed in more detail. As the GPS
monitoring points were only installed in 2011/08/09 the data sample is much smaller
compared to the geodetic sample space. Over this small period both flanks exhibit
similar behaviour in terms of seasonal fluctuations and amplitudes. See Figures 40
and 41.

10

Down-Stream
5
Plan View
Relative US/DS Displacement (mm)

-5

-10

-15

Up-Stream
-20
11/03/01

11/08/30

12/02/29

12/08/29

13/02/28

13/08/30

14/02/28

14/08/30

15/03/01

15/08/30

Date
Figure 40: Displacement data for left flank (KG12)

51
 5

Down-Stream
0
Plan View
Relative US/DS Displacement (mm)

-5

-10

-15

-20

Up-Stream
-25
11/03/01

11/08/30

12/02/29

12/08/29

13/02/28

13/08/30

14/02/28

14/08/30

15/03/01

15/08/30
Date
Figure 41: Displacement data for right flank (KG13)

4.2.5 GNSS/GPS and Water Level

Both the displacements and water levels follow cyclic patterns. Intuition would
suggest that the larger the hydraulic loading the greater the downstream deflection.
However it is clear from Figure 42 and 43 that the structure's response is not initiated
by changes in the reservoir level.
KG12 dx Water Level
10 152

Down-Stream
151
5
Relative US/DS Displacement (mm)

Plan View
150

0
Water Level (m)

149

-5 148

147
-10

146

-15
145
Up-Stream
-20 144
11/08/30

12/02/29

12/08/29

13/02/28

13/08/30

14/02/28

14/08/30

15/03/01

15/08/30
11/03/01

Date
Figure 42: Displacement and WL data for left flank

52
 KG13 dx Water Level
5 151

Down-Stream
150
0
Plan View
Relative US/DS Displacement (mm)

149
-5

Water Level (m)

148
-10
147

-15
146

-20
145
Up-Stream
-25 144
11/08/30

12/02/29

12/08/29

13/02/28

13/08/30

14/02/28

14/08/30

15/03/01

15/08/30
11/03/01

Date
Figure 43: Displacement and WL data for right flank

The graphs show that the structure changes the direction of displacement before the
water level changes. The dam moves downstream between one to three months
before the water level rises and vice versa. This phase lag may be due to another
loading which has a more immediate effect on the dam's behaviour.

Figure 44 shows the relationship between displacement and water level. The
relationship between the two variables is complex and no clear or obvious correlation
is evident.
5

Down-Stream

0
Plan View
Relative US/DS Displacement (mm)

-5

-10

-15

-20

Up-Stream
-25
141 142 143 144 145 146 147 148 149 150 151

Water Level (m)

Figure 44: Displacement VS Water Level

53
4.2.6 GNSS/GPS and Temperature
The displacements of the structure are approximately 180 degrees out of sync with
the temperature at any given point in time. See Figures 45 and 46 below.

KG12 dx Temperature
10 26

Down-Stream
24
5
Plan View
Relative US/DS Displacement (mm)

Temperature (Degrees Celcius)

22

0
20

-5 18

16
-10

14

-15
12
Up-Stream
-20 10
11/03/01

11/08/30

12/02/29

13/02/28

13/08/30

14/02/28

14/08/30

15/03/01

15/08/30
12/08/29

Date
Figure 45: Displacement and Temperature data for left flank

KG13 dx Temperature
5 30

Down-Stream
28
0
Relative US/DS Displacement (mm)

Plan View 26
Temperature (Degrees Celcius)

-5 24

22
-10
20

-15 18

16
-20
14
Up-Stream
-25 12
13/02/28

13/08/30

14/02/28

14/08/30

15/03/01

15/08/30
11/03/01

11/08/30

12/02/29

12/08/29

Date
Figure 46: Displacement and Temperature data for right flank

54
Figure 47 shows the relationship between displacement and temperature. The
relationship is linear with an increase in temperature resulting in an upstream
displacement and vice versa.

Down-Stream
0
Plan View
Relative US/DS Displacement (mm)

-5

-10

-15

-20

Up-Stream

-25
10 15 20 25 30
Temperature (Degrees Celcius)
Figure 47: Displacement VS Temperature

4.2.6 Trivec System

For the Trivec analysis the recorded data during high water levels were compared to
that of low water levels. Figure 48 shows Kogr1 during high water levels. The
warmer colours represent the data that was recorded during summer while the
cooler colours represent the data that was recorded during winter.

The full set of displacement graphs are shown in Appendix A. A summary of the
results for each Trivec is given in Table 11.

55
 Water level
60
Gauge plate level
50
40
30
20
10
0
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15

07/13 03/08 02/13 03/14 07/07 07/08 01/07 09/14

Legend:
Gallery (11) and Foundation (60) Base date

Cumulative R Displacement Cumulative T Displacement Cumulative Z Displacement

(micron) (micron) (micron)
0 0 0
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)
10 10 10

20 20 20

30 30 30
Downstream

Right Flank

Downwards
Left Flank
Upstream

Upwards
40 40 40

50 50 50

60 60 60

70 70 70

80 80 80

90 90 90
0

1 500
-4 500

-3 000

-1 500
0
-5 000

5 000
-10 000

20 000
10 000

15 000

25 000

3 000
1 500

4 500

6 000
-3 000

-1 500

Figure 48: Trivec Kogr1 during high water levels

Summary
The displacement trend of Kogr1 for the top third of the dam is consistent with the
geodetic measurements. The quarter points of the dam wall at the gallery level and
above tends towards the opposite flank. However below this level the Trivec
measurements of Kogr1 show that the lower two thirds of the wall is displacing
towards the same flank, namely the right flank. This may be because the dam is
stiffer below the FSL compared to the section between FSL and NOC. The missing
concrete section across the spillway and NOC levels result in the observed
behaviour of the top third of the dam wall tending to the opposite direction compared
to the lower stiffer sections.

The displacements throughout the foundations (right and left) fluctuate within a
bandwidth of approximately 0.2mm. The largest tangential displacement towards the
right flank occurs in winter during the high water levels. This increases the forces on
the foundation and can be seen to have an effect on Kogr2 and Kogr5 which are
positioned in the right flank foundation downstream of the dam. The displacements

56
may be an indication of weaker phyllite material inter-bedded between the highly
folded and compressed sandstone and quartzites.

Table 11: Trivec data summary

Trivec
High Water Level Low Water Level Common trend
System
 Largest downstream  Largest upstream
and downward and upward
displacement during displacement during
winter. summer.  Below foundation
 Below 2/3 of height,  Below 1/2 of height, level the movements
dam tends towards dam tends towards are stable (<0.2mm)
Kogr1 the right flank. Above, the left flank. Above, in all directions.
tends slightly towards tends slightly  Clear temperature
left flank (Observed in towards right flank. effect on behaviour
geodetics). Only observed for of dam wall
Exaggerated during summer readings.
winter.

 Clear displacement  No clear  Generally stable

trends changing in displacement trends with movements
direction from left to <0.2mm throughout
right flanks throughout height.
Kogr2 the height of the  Erratic movements
borehole. in top 5m.
 No clear
temperature effect
on behaviour
 Similar behaviour regardless of water level or season.
Kogr3
 Greatest radial and tangential displacement range in lower half of borehole.
 Similar behaviour regardless of water level or season.
 Generally stable with movements <0.2mm throughout height.
Kogr4  Erratic movements in top 10m.
 Slight tendency within the top 5m to displace downstream and upwards
during summer, and vice versa during winter.
 Spike in  No clear  Generally stable
displacements for all displacement trends with movements
Kogr5 directions (including <0.2mm throughout
vertical) 10m into height.
borehole (<0.3mm.)
Kogr6  Generally stable with movements <0.2mm throughout height.

4.2.7 AVM and Water Level

Clear changes occur in the recorded natural frequency as a result of changes in
water level. The natural frequencies (ωn) for the first four modes of vibration (φi), at
varying water levels are shown in Figure 49.

57
 Mode 1 Mode 2
3.42 3.66
y = -0.005x2 + 1.691x - 118.7 y = -0.007x2 + 2.279x - 162.1
3.40 3.64 R² = 0.927
R² = 0.963
3.38 3.62
Natural Frequency (Hz)

Natural Frequency (Hz)

3.36
3.60
3.34
3.58
3.32
3.56
3.30
3.54
3.28
3.26 3.52

3.24 3.50
3.22 3.48
146 147 147 148 148 149 149 150 150 151 146 147 147 148 148 149 149 150 150 151
Water Level (m) Water Level (m)

Mode 3 Mode 4
4.72 6.18
y = -0.013x2 + 3.953x - 287.0 y = -0.014x2 + 4.140x - 296.0
R² = 0.651 R² = 0.810
4.70
6.16
Natural Frequency (Hz)

Natural Frequency (Hz)

4.68
6.14
4.66
6.12
4.64

6.10
4.62

4.60 6.08
146 147 147 148 148 149 149 150 150 151 147 147 148 148 149
Water Level (m) Water Level (m)

Figure 49: Natural frequencies of the first four modes of vibration at varying water levels

As the water level increases the impounded water has an added mass effect on the
dam wall structure. The relationship between ωn, stiffness (k), and mass (m) has
previously been discussed (See Equation 2.4 in section 2.3.2), and is repeated here
for convenience.

As the water level rises (and added mass increases) the ωn decreases. The first,
second, and fourth modes strongly fit a second order polynomial regression line.
These strong modes are easily identifiable during the data processing of the AVM.
Mode three is often referred to as a weak mode and displays a low coefficient of
determination for a similar regression model.

The natural frequencies for the higher modes increase sequentially given that the
water level is constant. This is shown in Figure 50.

58
 7

6
Natural Frequency (Hz)

3
0 1 2 3 4 5

Modes
Figure 50: Natural frequencies for varying modes of vibration at FSL

The AVM signature (ωn) for a structure will change if either the mass or stiffness
changes. Therefore if the signature of the structure at a certain water level (constant
mass) is compared at various time periods the changes may be attributed to the
structural stiffness. A downward shift of the ωn's for the respective modes would thus
indicate a decrease in the stiffness of the structure. This demonstrates the
importance of AVM's over long periods of time in order to identify possible structural
deterioration.

4.2.8 Crack widths

The three dimensional crack width gauges were installed at five levels of the
downstream face of the dam wall (See Figures 51 and 52). Two gauges on each of
the joints on the Ultimate Non Overspill Crest (UNOC), NOC, and gallery levels. Only
single gauges were installed at the upper and lower walkways.

Two periods were investigated in order to observe the impacts of rising and falling
water level on the vertical construction joints of the structure. The periods are
summarized in Table 12. The first date acts as the baseline reading for the specific
period.
Table 12: Crack width measurement periods

Period Date Water Level (m) Volume (million m3)

2009/01/27 RL 144.03 99
A: Falling Water Level
2010/12/20 RL 127.51 39
2010/12/20 RL 127.51 39
B: Rising Water Level
2011/07/28 RL 149.97 130

59
Figures 51 and 52 show the crack width gauge results for the displacement across
the joints (tangential). A positive value indicates an opening (green) of the joint while
a negative value indicates a closure (red/orange) of the joint. The yellow values
indicate little change in joint movement.
25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
-0.08 0.31 0.27 0.62 0.21

0.13 -0.27 0.06 UNOC 0.62 0.19

-0.15 -0.30 0.04 0.11 0.44 0.19 0.17

0.37 0.09 0.19 0.00 NOC 0.41 0.21 0.12

FSL
0.04 0.14 0.22 -0.04 0.16 -0.01 0.43 0.87 0.54 0.63 0.30 0.53 0.30 0.06 0.54 0.38 0.40 0.18 0.10 0.24 -0.08
Gallery -0.27 0.12 0.37 0.25 0.39 0.64 -0.11 0.77 0.26 0.51 0.48 0.29 0.44 0.67 0.04 0.21 -0.07 0.04 0.14 0.02 0.23

0.28
Upper Walkway 0.25 0.26 0.09 0.13 0.21 0.18 0.28 0.34 0.46 0.44 0.42 0.38 0.37 0.00 0.25 0.15 0.29

Lower Walkway 0.03 -0.02 -0.08 0.16 0.05 0.08 0.17 0.19 0.08 0.09 0.32 0.17 0.08

Figure 51: Period A - Falling water level

25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
0.15 0.75 1.07 0.38 -0.04

0.13 0.77 1.07 UNOC 0.38 -0.02

0.42 0.05 0.63 0.37 0.11 0.12 0.13

-0.10 0.12 0.07 0.36 NOC 0.19 0.17 0.14

FSL
-0.10 0.06 -0.23 -0.15 -0.18 0.02 -0.48 -0.55 -0.41 -0.31 -0.43 -0.42 -0.33 -0.36 -0.25 -0.21 -0.26 -0.03 -0.01 0.13 0.08
Gallery 0.20 -0.03 -0.39 -0.28 -0.07 -0.34 0.35 -1.01 -0.45 -0.51 -0.43 -0.42 -0.34 -0.34 -0.25 -0.20 -0.08 0.00 0.02 0.16 0.10

-0.14
Upper Walkway -0.29 -0.19 -0.10 -0.06 -0.16 -0.19 -0.25 -0.22 -0.45 -0.32 -0.28 -0.32 -0.30 0.00 -0.21 -0.09 -0.31

Lower Walkway -0.07 0.03 -0.01 -0.04 0.00 0.00 -0.81 -0.72 0.00 0.03 -0.17 -0.09 -0.01

Figure 52: Period B - Rising water level

For period A the water level falls from just below the FSL to approximately the upper
walkway. The joints open up with the maximum being approximately 1mm.

For period B the water level increases from the upper walkway to above the FSL.
The joints are more closed compared to period A and have a maximum closure of
approximately 1mm. The closure of the joints seems to only occur below the level of
the water surface (FSL). It is also evident that the largest joint movements for both
periods occur in the middle of the structure.

The data indicates that the joints of the structure are more open during low water
levels. Thus the hydrostatic loading is not transferred to the foundations
predominantly via hoop stresses because the dam wall does not act completely as a
monolith. When the water level is high the joints are more closed and the dam wall is

60
in a more compressed state. The joints above the water level are however still open
despite the high water level. This may also indicate that the swelling effects
associated with ASR may be predominantly located in the lower sections of the dam
wall.

4.3 Comparative Study

Roode Elsberg Dam is a thin, double curvature, arch dam located in the Western
Cape Province of South Africa. Kouga and Roode Elseberg dams are similar in
terms of dam type and size, and a comparison was done on the respective recorded
modal parameters and stiffness of the dams. The fact that Kouga has been exposed
to ASR may allude to the dam having a high stiffness as a result of the added
restraint caused by the swelling of the concrete. Table 13 compares statistics of
Kouga to Roode Elsberg dam.

Table 13: Comparative dam statistics

Parameter Roode Elsberg Dam Kouga Dam

Height (m) 72 69
Crest Width (m) 2.6 6.3 (10m at pulvino)
Crest Length (m) 274 317
Spillway Length (m) 76 127

Although the dams exhibit similar heights Kouga spans a greater length and is much
thicker in comparison. As a result Kouga has a greater mass and presumably lower
natural frequencies for the various modes at a given water level. Figure 53 compares
the natural frequencies of both dams.

Kouga Roode Elsberg

7
Natural Frequency (Hz)

3
0 1 2 3 4 5
Modes
Figure 53: Natural frequencies for Kouga and Roode Elsberg Dams

61
Figure 53 indicates that Kouga exhibits greater natural frequencies for the respective
modes of vibration, despite it having the greater mass. This confirms that Kouga dam
is a stiff structure for its size.

4.4 Dam Forces on Foundation

The valley in which Kouga dam was constructed exhibits a canyon shape factor (k)
of approximately 5.3 (See section 2.1.2), and this relates to a wide U-shaped valley.
With a base width and height ratio of approximately 0.15 Kouga may be described as
a thin arch dam. Due to the combination of topography and design it can be argued
that the dam wall would naturally tend to transfer the imposed loading via dominant
cantilever action to the central foundations. Carvalho et. al. (2014: VI17) observed
that the upper construction joints of the dam opened during the low water levels. This
observation corresponds with the crack width results (See section 4.2.8). The
structure therefore does not act as a monolith during low water levels and the forces
are transferred to the foundation predominantly via cantilever action.

The simplified finite element model of Kouga dam with a hydrostatic load at full
supply level is shown below in Figure 54. This model illustrates how the structure
transfers stresses to the foundations.

Figure 54: Force transfer through arch during FSL hydrostatic loading

The upstream face displays significant arch action with maximum compressive
stress occurring at approximately two thirds of the height of the dam. The thrusts
then drop rapidly on the downstream face and result in the largest compressive
stresses which are located at the wide river bed. It is evident that the structure
transfers the imposed loading predominantly via cantilever action to the central
foundation. This results in relatively large tensile stresses at the heel of the dam. In
light of the aforementioned ASR identified at the dam, which would most likely be

62
located on the wetted upstream face, the reduction in tensile capacity associated
with the deterioration together with the high tensile stresses increases the possibility
of tensile cracking at the heel.

4.5 Chapter Summary

This chapter presents the results of the instrumentation data recorded by the
monitoring system at Kouga Dam. An analysis of the individual data sets is
discussed and some of the results are overlaid in order to deduce the relationship
between certain measured parameters. In addition this chapter presents a
comparative study between Kouga and Roode Elsberg Dam. These two dams are
similar in terms of dam type and size, and a comparison was done on the respective
recorded modal parameters and stiffness of the dams. In conclusion the chapter
describes how a simplified finite element model was used to determine the
magnitude and nature of the forces within the concrete arch structure during
operation.

63
 CHAPTER 5
5 CONCLUSION AND RECOMMENDATIONS

5.1 Summary
The NWA mandates that dams with a safety risk be routinely evaluated on their
safety. Monitoring systems installed at the dams assist the engineers with this task
and gives insights into the structure's operational behaviour. Following the data
analysis instrumentation results identify the structural response to changes in
loading, and long term trends may indicate deterioration of the structure over time.
Once the behaviour is understood a more informed argument can be made for the
potential failure modes of the structure and the likelihood thereof.

The purpose of this research is to analyse the instrumentation data recorded at a

double curvature arch dam in order to identify trends and better understand the
operational behaviour of the structure. The case study of this thesis is Kouga Dam in
the Eastern Cape Province of South Africa. Kouga was completed in 1969, and after
the catastrophic failure of Malpasset Dam in 1959, a decision was made to stabilise
the right abutment due to suspected instability. In 1981 ASR was positively identified
in the dam wall. Due to the aforementioned aspects and catastrophic consequences
associated with a dam failure an extensive monitoring system has been installed at
Kouga Dam.

In the sections below the conclusions of the analysis of the monitoring system and
recommendations for future work relating to this thesis are presented.

5.2 Conclusion
The individual instrumentation results from the comprehensive monitoring system
have created a more complete picture of the structure's behaviour. The data analysis
has assisted in identifying trends and has provided significant insights into the
operational behaviour of the structure. The way in which the data was organised and
displayed proved most telling when identifying how the structure responds to specific
loads and long term tendencies.

The following conclusions were drawn from the results and discussion of this thesis:

i. A linear relationship exists between temperature loading and displacement

response of the dam wall. Temperature initiates the response of the
structure almost instantaneously. An increase in temperature results in an
upstream displacement and vice versa. The most extreme upstream
displacement occurs during low water levels and high temperature. The
former allowing a larger portion of un-wetted concrete (upstream face)
exposed to direct sunlight.

64
 ii. A phase lag of approximately one to three months is evident between
hydrostatic loading and the displacement response of the structure. The
relationship between these two variables is complex. Hydrostatic loading
however exhibits a more significant effect on natural frequencies, with the
relationship for most of the lower modes of vibration fitting a second order
polynomial regression line. An increase in water level results in a decrease
in natural frequency for the respective modes and vice versa. The recorded
natural frequencies, and in light of the ASR, has also indicated that Kouga
dam is a relatively stiff structure when compared to a dam of similar type and
size.

iii. Since construction the dam wall has tended towards the upstream (y),
upwards (z), and quarter points of the structure tending towards opposite
flanks (gallery level). The displacement and strain rates in the y and z
directions are 0.3mm/annum and 8.6µε/annum respectively. However, since
1989 there has been a reduction in the average displacement and strain
rates in all directions by approximately 70%. This may suggest that the ASR
has stabilized.

iv. The vertical construction joints, especially the central and upper joints, are
relatively open during low water levels. Therefore the dam wall does not act
completely as a monolith, and transfers the imposed loading mainly to the
central foundation via dominant cantilever action. The reaction forces on the
upper foundations have been found to be relatively low when compared to
the central foundation, lowering the risk of potential shear failure of the right
foundation.

v. Small foundation movements of less than 0.3mm have been observed within
the foundation downstream of the dam wall on the right flank. These
movements which are between 10 and 40m within the foundations are most
notable during winter at high water levels. This is the period when the dam is
in its most downstream position.

5.3 Recommendations
The following recommendations are given for future work relating to this thesis:

i. The extent and activity of the ASR should be determined. This information
would assist in the development of a calibrated finite element model.

ii. A finite element model should be developed and should incorporate the
swelling effects associated with ASR, temperature, and hydraulic loading.
This model should be calibrated using the data obtained from the monitoring
system.

65
 REFERENCES

Addis, B. 1998. Fundamentals of Concrete. G. Owens, Ed. Midrand, South Africa:

Cement and Concrete Institute.

Ballim, Y., Alexander, M., Beushausen, H. 2009. Chapter 9: Durability of Concrete.

In Fulton's Concrete Technology. Owens, G. Ed. 9th ed. Midrand, South Africa:
Cement & Concrete Institute. 155-188.

Bathe, K. J. 1996. Finite Element Procedures. Upper Saddle River, New Jersey:
Prentice Hall.

Brandt, M. 2011. Seismic Hazard in South Africa. (Report Number: 2011-0061).

Pretoria, South Africa. Council for Geoscience.

Carvalho, E., Valentim, N. & Oosthuizen, C., 2014. On-line dynamic monitoring of
Cahora Bassa Dam.... the next level. Proceedings of the International Symposium on
Dams in a Global Environmental Challenges. 1-6 June 2014. Bali, Indonesia. VI-11
to VI-20.

Chang. P. C., Flatau, A., Liu, S. C. 2003. Review Paper: Health Monitoring of Civil
Infrastructure. Sage. 257-267.

Concrete Microscope Library. 2016. Alkali-Silica Reaction. Available:

http://publish.illinois.edu/concretemicroscopylibrary/alkali-silica-reaction/ [2016,
November, 26].

da Silva, V. D. & Júlio, E. N. B. S. 1995. Design and Analysis of Arch Dams by the
Membrane Method. Proceedings of the International Conference on Education,
Practice and Promotion of Computational Methods in Engineering Using Small
Computers. 1-4 August 1995. Macau, China. 533-538.

Dam Safety Office. 2011. Registration Details of a Dam Registered in terms of Dam
Safety Legislation of Chapter 12 of the National Water Act (Act No. 36 of 1998).
Pretoria, South Africa: Directorate: Dam Safety Office, Department of Water and
Sanitation.

Elges, H., Geertsema, A., Lecocq, P., Oosthuizen, C. 1995. Detection, monitoring
and modelling of alkali-aggregate reaction in Kouga Dam (South Africa). (Report
CONF-9510182--TRN: 95:008180-0012). Denver, CO, United States. Committee on
Large Dams.

Engineering Profession Act, No. 46 of 2000. 2000. Government Gazette.

426(1821). 1 December. Government notice no. 1304. Cape Town: Government
printer.

66
Federal Energy Regulatory Commission, 1999. Chapter 11 - Arch Dams.
Engineering Guidelines for the Evaluation of Hydropower Projects. Washington, DC.

Federal Energy Regulatory Commission, 2014. Chapter R5 - Concrete Dam

Analysis: Arch, Gravity Buttress Dams. Engineering Guidelines Risk-Informed
Decision Making. Washington, DC.

Forbes Dick. 1989. Engineering Geological Report on the Safety of Paul Sauer Dam.
(Report Number: GSO/1/89/5). Forbes Dick and Associates (Consulting Engineering
Geologists). Claremont.

GeoSIG. 2010. AC-23/AC-22/AC-21 Accelerometer [Leaflet].

Ghanaat Y. 2004. Failure Modes Approach to Safety Evaluation of Dams.

Proceedings of the 13th World Conference on Earthquake Engineering. 1-6 August
2004. Vancouver, Canada. Paper no. 1115.

Gillian, C., Lund, G., Weldon, J. 2011. Three Predominate Failure Modes of Thin
Arch Dams. Proceedings of the 31st Annual USSD Conference: 21st Century Dam
Design - Advances and Adaptations. 11-15 April 2011. San Diego, California. 173-
187.

Google Earth 7.1.2.2041. 2013. Baviaanskloof Nature Reserve, Eastern Cape, South
Africa [Map, July 10]. 33° 44' 26.00"S, 24° 35' 16.00"E. Borders and labels; places
layers. NOAA, US Department of State Geographer 2016. [2016, November 27].

Hobbs, L.D., du Plessis, J.G., Kriel, J.P. 1967. Stabilising highly stratified rock in the
right abutment of Tweerivieren Dam. Proceedings of the Commission Internationale
Des Grands Barrages. Istanbul, Turkey. 647-667.

Shaw, Q.H.W. 2015. The Structural Function of Different Arch Dam Types.
Proceedings of the SANCOLD Annual Conference 2015. 1-3 September 2015.
Cape Town, South Africa. 320-332.

Kuo, J. S. 1982. Fluid-structure Interactions: Added Mass Computations for

Incompressible Fluid. National Science Foundation.

Lombardi, G. & Fanelli, M. 1992. Practice and Theory of Arch Dams. Proceedings of
the International Symposium on Arch Dams. 17-20 October 1992. Nanjing, China. 1-
4.

Moyo, P. 2013. Dynamics: Definitions and concepts. [CIV5113Z PowerPoint

presentation]. Department of Civil Engineering, University of Cape Town.

Moyo, P. 2014. Operational Modal Analysis - Ambient Vibration Testing. [CIV5119Z

PowerPoint presentation]. Department of Civil Engineering, University of Cape Town.

67
National Water Act, No. 36 of 1998. 1998. Government Gazette. 398(19182). 26
August. Government notice no. 1091. Pretoria: Government printer.

National Water Act, No. 36 of 1998. Regulation. 2012. Government Gazette.

560(35062). 24 February. Government notice no. R139. Pretoria: Government
Printer.

Naude, P. 2002. Kouga Dam: Dam Safety Installation Report on Trivec System.
(Report Number: L820/01/DX02). Pretoria, South Africa. Directorate: Civil Design,
Sub-Directorate: Dam Safety, Department of Water Affairs and Forestry.

Oberholster, B. 2009. Chapter 10: Alkali-Silica Reaction. In Fulton's Concrete

Technology. Owens, G. Ed. 9th ed. Midrand, South Africa: Cement & Concrete
Institute. 189-218.

Okuma, N., Etou, Y., Kanazawa, K., Hirata, K. 2008. Dynamic properties of a large
arch dam after forty-four years of completion. Proceedings of the 14th World
Conference on Earthquake Engineering 2008. 12-17 October 2008. Beijing, China.

Oosthuizen, C., 2014. Safety of Special Structures. [CIV5118Z: Guest lecture]

Department of Civil Engineering, University of Cape Town.

Oosthuizen, C., Naude, P.A., Dorfling, C.J. 2003. A simple 3-Dimensional crack-
width-tilt gauge out of Africa. (ISBN 90 5809 602 5). Field Measurements in
Geomechanics. Proceedings of the 6th International Symposium. 23-26 September
2003. Oslo, Norway. 599-604.

Poraver. 2016. Influence of expanded glass granulate on the alkali-silica-reaction in

cementitious systems. Available: http://www.poraver.com/us/influence-expanded-
glass-granulate-alkali-silica-reaction-cementitious-systems-2/ [2016, November, 26]

Pretorius, C.J. & du Toit, I. 2014. The Department of Water Affairs in South Africa
expands Dam monitoring to GPS/GNSS Deformation Monitoring [Paper]. Pretoria,
South Africa. Directorate: Strategic Asset Management, Sub-Directorate: Dam
Safety Surveillance, Department of Water and Sanitation. (Unpublished)

Pretorius, C.J., Schmidt, W.F., van Staden, C.S., Egger, K. 2001. The Extensive
Geodetic System used for the Monitoring of a 185 metre High Arch Dam in Southern
Africa. Proceedings of the 10th International Symposium on Deformation
Measurements. 19-22 March 2001. California, United States of America. 203-213.

Shaw, Q.H.W. 2015. The Structural Function of Different Arch Dam Types.
Proceedings of the SANCOLD Annual Conference 2015. 1-3 September 2015.
Cape Town, South Africa. 320-332.

South African National Committee on Large Dams. 1991. Guidelines on Safety in

Relation to Floods. (Report no. 4). Pretoria, South Africa: Safety evaluation of dams,
SANCOLD.

68
United States Army Corps of Engineers. 1993. Theoretical Manual for Analysis of
Arch Dams. (Technical Report no. ITL-93-1). Washington, DC.

United States Army Corps of Engineers. 1994. Arch Dam Design. (Engineering
Manual no. 1110-2-2201). Washington, DC.

United States Bureau of Reclamation. 1977a. Design of Arch Dams: Design Manual
for Concrete Arch Dams. Denver, Colorado: A Water Resources Technical
Publication.

United States Bureau of Reclamation. 1977b. Guide for Preliminary Design of Arch
Dams. Denver, Colorado: A Water Resources Technical Publication.

United States Bureau of Reclamation. 2016. Flaming Gorge Unit. Available:

http://www.usbr.gov/uc/rm/crsp/fg/ [2016, November 25]

Westergaard, H. M. 1933. Water Pressures on Dams During Earthquakes.

Transactions. American Society of Civil Engineers. Volume 92. 418-433.

World Bank. 2016. Average precipitation in depth (mm per year). Available:
http://data.worldbank.org/indicator/AG.LND.PRCP.MM?end=2014&locations=1W&st
art=2012&view=map [2016, November, 22].

69
 APPENDICES
A. TRIVEC RESULTS

70
Data obtained,
Position 1: Highprocessed
water Leveland graphs prepared by DWAF Pretoria. Version 2016/08/29 3:33 PM

Water level
60

Gauge plate level

50
40
30
20
10
0
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15

- 07/13 03/08 02/13 03/14 - 07/07 07/08 01/07 09/14

Legend:
Gallery (11) and Foundation (60) Base date

Cumulative R Displacement Cumulative T Displacement Cumulative Z Displacement

(micron) (micron) (micron)
0 0 0

10 10 10
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

20 20 20

30 30 30

40 40 40

50 50 50

60 60 60

70 70 70

80 80 80

90 90 90
0

0
20 000

25 000
5 000

10 000

15 000

1 500
-10 000

-5 000

-4 500

-3 000

-1 500

1 500

3 000

4 500

6 000
-3 000

-1 500

Different Scale

Relative Rad Displacement Relative Tan Displacement Relative Z Displacement

(micron) (micron) (micron)
0 0 0

10 10 10
Trivec measuring point number (numbered from the top)
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

20 20 20

30 30 30

40 40 40

50 50 50

60 60 60

70 70 70

80 80 80

90 90 90
0

200

400
-200
200

400
0
-200
800
200

400

600
0
-400

-200

Different Scale

Positive directions: R=Downstream, T= Left Flank and Z= Downwards

Figure 1 Kouga Dam: Kogr1 Displacements obtained with the TRIVEC at position 1.
Data obtained,
Position 1: Lowprocessed
water Leveland graphs prepared by DWAF Pretoria. Version 2016/08/29 3:35 PM

Water level
60

Gauge plate level

50
40
30
20
10
0
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15

- 07/04 02/06 02/10 02/11 - - - 02/05 07/05

Legend:
Gallery (11) and Foundation (60) Base date

Cumulative R Displacement Cumulative T Displacement Cumulative Z Displacement

(micron) (micron) (micron)
0 0 0

10 10 10
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

20 20 20

30 30 30

40 40 40

50 50 50

60 60 60

70 70 70

80 80 80

90 90 90
0

0
20 000

25 000
5 000

10 000

15 000

1 500
-10 000

-5 000

-4 500

-3 000

-1 500

1 500

3 000

4 500

6 000
-3 000

-1 500

Different Scale

Relative Rad Displacement Relative Tan Displacement Relative Z Displacement

(micron) (micron) (micron)
0 0 0

10 10 10
Trivec measuring point number (numbered from the top)
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

20 20 20

30 30 30

40 40 40

50 50 50

60 60 60

70 70 70

80 80 80

90 90 90
0

200

400
-200
200

400
0
-200
800
200

400

600
0
-400

-200

Different Scale

Positive directions: R=Downstream, T= Left Flank and Z= Downwards

Figure 1 Kouga Dam: Kogr1 Displacements obtained with the TRIVEC at position 1.
Data obtained,
Position 2: Highprocessed and graphs prepared by DWAF Pretoria.
water Level Version 2016/08/25 12:44 AM

Water level
60

Gauge plate level 50

40
30
20
10
0
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15

- 07/13 03/08 02/13 03/14 - 07/07 07/08 01/07 09/14

Legend:
Rock discontinuities Base date

Cumulative R Displacement CumulativeT Displacement Cumulative Z Displacement

(micron) (micron) (micron)
0 0 0

5 5 5
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

15 15 15

20 20 20

25 25 25

30 30 30

35 35 35

40 40 40

45 45 45

50 50 50
0

500

1 000
-500
0

500

1 500
1 000
-500
-1 000

500

1 000
-500

Different Scale

Relative R Displacement Relative T Displacement Relative Z Displacement

(micron) (micron) (micron)
0 0 0

5 5 5
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

15 15 15

20 20 20

25 25 25

30 30 30

35 35 35

40 40 40

45 45 45

50 50 50
0

0
100

100

200

100
-300

-200

-100

-200

-100

-100

Positive directions: R =Downstream, T= Left Flank and Z= Vertical Downwards

Figure 2 Kouga Dam: Kogr2 Displacements obtained with the TRIVEC at position 2.
Data obtained,
Position 2: Lowprocessed
water Leveland graphs prepared by DWAF Pretoria. Version 2016/08/25 12:33 AM

Water level
60

Gauge plate level 50

40
30
20
10
0
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15

- 07/04 02/06 02/10 02/11 - - - 02/05 07/05

Legend:
Rock discontinuities Base date

Cumulative R Displacement CumulativeT Displacement Cumulative Z Displacement

(micron) (micron) (micron)
0 0 0

5 5 5
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

15 15 15

20 20 20

25 25 25

30 30 30

35 35 35

40 40 40

45 45 45

50 50 50
0

500

1 000
-500
0

500

1 500
1 000
-500
-1 000

500

1 000
-500

Different Scale

Relative R Displacement Relative T Displacement Relative Z Displacement

(micron) (micron) (micron)
0 0 0

5 5 5
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

15 15 15

20 20 20

25 25 25

30 30 30

35 35 35

40 40 40

45 45 45

50 50 50
0

0
100

100

200

100
-300

-200

-100

-200

-100

-100

Positive directions: R =Downstream, T= Left Flank and Z= Vertical Downwards

Figure 2 Kouga Dam: Kogr2 Displacements obtained with the TRIVEC at position 2.
Data obtained,
Position 3: Highprocessed and graphs prepared by DWAF Pretoria.
water Level Version 2016/08/25 1:42 AM

Water level
60

Gauge plate level 50

40
30
20
10
0
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16

- 07/13 03/08 02/13 03/14 - 07/07 07/08 01/07 09/14

Legend:
Rock discontinuities Base date

Cumulative R Displacement Cumulative T Displacement Cumulative Z Displacement

(micron) (micron) (micron)
0 0 0

5 5 5
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

15 15 15

20 20 20

25 25 25

30 30 30

35 35 35
0

0
1 000

1 000
-4 000

-3 000

-2 000

-1 000

-3 000

-2 000

-1 000

500
-500
-1 000

Different Scale

Relative Rad Displacement Relative Tan Displacement Relative Z Displacement

(micron) (micron) (micron)
0 0 0

5 5 5
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

15 15 15

20 20 20

25 25 25

30 30 30

35 35 35
0

0
200
100

100

200

100
-200
-300

-100

-100
-300

-200

-100

Different Scale

Positive directions: R=Downstream, T= Left Flank and Z= Vertical Downwards

Figure 3 Kouga Dam: Kog3 Adjusted displacements obtained with the TRIVEC at position 3.
Data obtained,
Position 3: Lowprocessed
water Leveland graphs prepared by DWAF Pretoria. Version 2016/08/25 1:43 AM

Water level
60

Gauge plate level 50

40
30
20
10
0
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16

- 07/04 02/06 02/10 02/11 - - - 02/05 07/05

Legend:
Rock discontinuities Base date

Cumulative R Displacement Cumulative T Displacement Cumulative Z Displacement

(micron) (micron) (micron)
0 0 0

5 5 5
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

15 15 15

20 20 20

25 25 25

30 30 30

35 35 35
0

0
1 000

1 000
-4 000

-3 000

-2 000

-1 000

-3 000

-2 000

-1 000

500
-500
-1 000

Different Scale

Relative Rad Displacement Relative Tan Displacement Relative Z Displacement

(micron) (micron) (micron)
0 0 0

5 5 5
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

15 15 15

20 20 20

25 25 25

30 30 30

35 35 35
0

0
200
100

100

200

100
-200
-300

-100

-100
-300

-200

-100

Different Scale

Positive directions: R=Downstream, T= Left Flank and Z= Vertical Downwards

Figure 3 Kouga Dam: Kog3 Adjusted displacements obtained with the TRIVEC at position 3.
 Data obtained,
Position 4: Highprocessed and graphs prepared by DWAF Pretoria.
water Level Version 2016/08/25 2:43 AM

Water level
60

Gauge plate level

50
40
30
20
10
0
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15

- 07/13 03/08 02/13 03/14 - 07/07 07/08 01/07 09/14

Legend:
Rock discontinuities Base date

Cumulative R Displacement Cumulative T Displacement Cumulative Z Displacement

(micron) (micron) (micron)
0 0 0
Trivec measuring point number (numbered from the top)
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

20 20 20

30 30 30

40 40 40

50 50 50

60 60 60
500
-500

1 000
-1 000

0
0

3 000
1 000

2 000
-1 000

500

1 000

1 500
-500
-1 000

Different Scale

Relative Rad Displacement Relative Tan Displacement Relative Z Displacement

(micron) (micron) (micron)
0 0 0
Trivec measuring point number (numbered from the top)
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

20 20 20

30 30 30

40 40 40

50 50 50

60 60 60
-600

-400

-200

200

400

600
0

0
400
200

100

200
-400

-200

-300

-200

-100

PositiveDam:
Figure 4 Kouga directions:
Kogr4R=Downstream, T= Left Flank and Z= Vertical Downwards
Adjusted displacements obtained with the TRIVEC at position 4.
 Data obtained,
Position 4: Lowprocessed
water Leveland graphs prepared by DWAF Pretoria. Version 2016/08/25 2:43 AM

Water level
60

Gauge plate level

50
40
30
20
10
0
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15

- 07/05 02/06 02/10 02/11 - - 07/04 02/05 -

Legend:
Rock discontinuities Base date

Cumulative R Displacement Cumulative T Displacement Cumulative Z Displacement

(micron) (micron) (micron)
0 0 0
Trivec measuring point number (numbered from the top)
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

20 20 20

30 30 30

40 40 40

50 50 50

60 60 60
500
-500

1 000
-1 000

0
0

3 000
1 000

2 000
-1 000

500

1 000

1 500
-500
-1 000

Different Scale

Relative Rad Displacement Relative Tan Displacement Relative Z Displacement

(micron) (micron) (micron)
0 0 0
Trivec measuring point number (numbered from the top)
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

20 20 20

30 30 30

40 40 40

50 50 50

60 60 60
-600

-400

-200

200

400

600
0

0
400
200

100

200
-400

-200

-300

-200

-100

PositiveDam:
Figure 4 Kouga directions:
Kogr4R=Downstream, T= Left Flank and Z= Vertical Downwards
Adjusted displacements obtained with the TRIVEC at position 4.
Data obtained,
Position 5: Fullprocessed
Record and graphs prepared by DWAF Pretoria. Version 2016/08/25 2:04 AM

Water level
60

Gauge plate level

50
40
30
20
10
0
13 14 15 16

- 09/14 - 04/14 03/15 - - 09/13 - 07/15

Legend:
Rock discontinuities Base date

Cumulative R Displacement CumulativeT Displacement Cumulative Z Displacement

(micron) (micron) (micron)
0 0 0

5 5 5
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

15 15 15

20 20 20

25 25 25

30 30 30

35 35 35

40 40 40

45 45 45

50 50 50
0

1 000
-3 000

-2 000

-1 000
0

1 000
-2 000

-1 000

1 000

2 000
-1 000

Different Scale

Relative R Displacement Relative T Displacement Relative Z Displacement

(micron) (micron) (micron)
0 0 0

5 5 5
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

15 15 15

20 20 20

25 25 25

30 30 30

35 35 35

40 40 40

45 45 45

50 50 50
100
0
-300

-200

-100

100

200

100

200

300
0

-100

0
-300

-200

-100

Positive directions: R =Downstream, T= Left Flank and Z= Vertical Downwards

Figure 5.1 Kouga Dam: Kogr5 Displacements obtained with the TRIVEC at position Kogr5.
Data obtained,
Position 6: Fullprocessed
Record and graphs prepared by DWAF Pretoria. Version 2016/08/25 2:17 AM

Water level
60

Gauge plate level

50
40
30
20
10
0
14 15

- 07/15 - 04/14 03/15 - 06/14 09/14 - -

Legend:
Rock discontinuities Base date

Cumulative R Displacement CumulativeT Displacement Cumulative Z Displacement

(micron) (micron) (micron)
0 0 0

5 5 5
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

15 15 15

20 20 20

25 25 25

30 30 30

35 35 35

40 40 40

45 45 45

50 50 50
0

1 000
-2 000

-1 000
500
0
-500
-1 000

250

500
0
-500

-250

Different Scale

Relative R Displacement Relative T Displacement Relative Z Displacement

(micron) (micron) (micron)
0 0 0

5 5 5
Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

Trivec measuring point number (numbered from the top)

10 10 10

15 15 15

20 20 20

25 25 25

30 30 30

35 35 35

40 40 40

45 45 45

50 50 50
100
0
-200

-100

100

100
0

-100

0
-200

-100

Positive directions: R =Downstream, T= Left Flank and Z= Vertical Downwards

Figure 6.1 Kouga Dam: Kogr6 Displacements obtained with the TRIVEC at position Kogr6.