Delft University of Technology

CIE4061-09 Multidiscliplinary Project

Durban Dig-Out Port Research

Part 2 Report - Breakwater Engineering

Supervisors WSP:
Andre van Tonder Pr Eng
Team Project Durban:
Rob Leach
Rik Gijsman
Supervisors TU Delft:
Frans de Haan
ir. Bas Wijdeven
Rick de Koning
dr. ir. Phil Vardon
Tuan Le
Client TNPA:
Sander Steeneken
David John McGillewie
Selvan Pillay

January 29, 2015

Preface

The second objective of our Multidisciplinary Project is presented in this report. A breakwater
design is investigated for the proposed port layout for the Durban Dig-Out Port. As a project
team we are satisfied with the results and eager to continue with the third objective, which is the
recommendation for the Transnet National Ports Authority regarding the expected design vessels.
Our stay in South Africa has been successful and pleasant from the start. In this, the WSP and
Transnet staff have played a significant role and we would like to show our gratitude towards them.

Stellenbosch,
January 2015,

i
Contents

1 Introduction vii
1.1 Multidisciplinary Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
1.2 Durban Dig-Out Port Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
1.3 Research Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x
1.4 Context of the Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x
1.5 Report Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi

2 Length and Orientation 1

2.1 Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2.2 Model setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3 Cross Section 11
3.1 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.1.1 Type of Breakwater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.1.2 Type of Armour Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1.3 Probability of Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2.1 Design Water Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2.2 Wave Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.3 Probabilistic Determination of the Armour Units . . . . . . . . . . . . . . . . . . . 14
3.3.1 Limit State Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.3.2 Parameters and Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.3.3 Armour Unit Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.3.4 Probabilistic Calculation Files . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.4 Cross Section Dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4.1 Crest Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4.2 Toe Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.4.3 Under- and Filter Layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.5 Cross section Drawings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

4 Settlement and Stability Analysis 27

4.1 Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.2 Model Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

ii
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

5 Concluding on Breakwater Design 45

6 Sand Bypass System 47

6.1 Situation without Sand Bypass System . . . . . . . . . . . . . . . . . . . . . . . . . 48
6.2 Sand Bypass Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
6.3 Beach Nourishment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.4 Sand Trap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.5 Overview Sand Bypass System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

7 Care for the Mangroves 55

7.1 Proposed Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
7.2 Concluding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

iii
List of Figures

1.1 Location of the project shown at two scales (Source: Google Maps) . . . . . . . . . ix

2.1 Computational grid and splines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2.2 Bathymetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.3 Storm significant wave height originating from the east . . . . . . . . . . . . . . . . 6
2.4 Storm significant wave height originating from the south . . . . . . . . . . . . . . . 6
2.5 Storm significant wave height penetration . . . . . . . . . . . . . . . . . . . . . . . . 7
2.6 1% exceedance significant wave height penetration . . . . . . . . . . . . . . . . . . . 8

4.1 Cross section in Plaxis for the -25m CD breakwater . . . . . . . . . . . . . . . . . . 28

4.2 Cross section northern breakwater -15 m CD . . . . . . . . . . . . . . . . . . . . . . 31
4.3 Cross section northern breakwater -25 m CD . . . . . . . . . . . . . . . . . . . . . . 31
4.4 Cross section in Plaxis for the -25m CD breakwater with water level of +1.10 m CD 36
4.5 Breakwater cross section scenarios for northern breakwater analysis in Plaxis . . . . 37
4.6 Settlement of the -25 m CD northern breakwater after 50 years, ’Average/Reference‘
parameter set. Water level +1.1 m CD (Lee and Sea side) . . . . . . . . . . . . . . 39
4.7 Settlement of the -25 m CD northern breakwater, ’Average/Reference‘ parameter
set. Water level + 3.7 m CD (Lee Side) and + 6.85 m CD (Sea Side) . . . . . . . . 41
4.8 Horizontal displacement at Xbloc layer -25 m CD northern breakwater, ’Lower
Limit‘ parameter set. Water level + 3.7 m CD (Lee Side) and + 3.7 m CD with
3.15 m amplitude (Sea Side) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

7.1 Summary of measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

iv
List of Tables

2.1 Delft3D-Wave - Imposed wave conditions . . . . . . . . . . . . . . . . . . . . . . . . 4

2.2 Delft3D-Wave - Imposed obstacles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

3.1 Determination of the design water level . . . . . . . . . . . . . . . . . . . . . . . . . 13

3.2 Design wave heights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.3 Parameters and distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.4 Parameters and distributions at -25 m CD . . . . . . . . . . . . . . . . . . . . . . . 16
3.5 Designed Xbloc for tip of northern breakwater . . . . . . . . . . . . . . . . . . . . . 16
3.6 Parameters and distributions at -15 m CD . . . . . . . . . . . . . . . . . . . . . . . 17
3.7 Designed Xbloc for lee side of northern breakwater . . . . . . . . . . . . . . . . . . 17
3.8 Parameters -25 m CD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.9 Parameters -15 m CD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

4.1 Overview of the parameters necessary for the different material models . . . . . . . 32
4.2 Input parameters for the materials that are used in the Linear Elastic model . . . . 33
4.3 Input parameters for the materials that are used in the Mohr-Coulomb model . . . 33
4.4 Parameters rock fill . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.5 Input parameters for Xbloc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.6 Settlements compared to required calculation time for different mesh options . . . . 35
4.7 Results of the consolidation analysis using different parameter sets for the subsoil . 38
4.8 Results of the long term settlement analysis for different parameter sets . . . . . . . 38
4.9 Results of static calculation (Lee side: 3.7 m CD and Sea side: 6.85 m CD [3.7 +
3.15]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.10 Results of static calculation (Lee side: 0.0 m CD and Sea side: 0.0 m CD) . . . . . 40
4.11 Results of harmonic calculation (Lee side: 3.7 m CD and Sea side: 3.7 + 3.15 m
CD amplitude) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

v
Chapter 1

Introduction

In this chapter the research and the project are introduced. Afterwards the context and the outline
of this report are presented.

1.1 Multidisciplinary Project

Delft University of Technology (DUT) offers their master students Civil Engineering the elective
course CIE4061-09: Multidisciplinary Project. The purpose of the course is to solve an engineering
problem in practise, with a group of students specialising in different areas. The course is rewarded
with 10 European Credits (ECTS) per person, equivalent to a work load of 280 hours. The five
members of Project Durban have elected the course and arranged a period of 12 weeks attached
to WSP Group Africa in Stellenbosch, South Africa. In this period research is conducted for
Transnet National Ports Authority (TNPA), towards the Dig-Out Port Project in Durban (DDOP),
South Africa. The lessons learned at DUT are applied and during the research the team faces
discrepancies between the Dutch theoretical lessons learned in university and the practical South
African solutions based on experience.
Below, the team members and their specialisation area are introduced:
• Rik Gijsman - Coastal Engineering;
• Frans de Haan - Hydraulic Structures;
• Rick de Koning - Hydraulic Structures;
• Tuan Le - Ports Engineering;
• Sander Steeneken - Geo-Engineering.
This research is performed for Transnet National Ports Authority (TNPA). Transnet SOC Ltd. is
a state owned company with different operational divisions;
• National Ports Authority (TNPA);
• Port Terminal (TPT);
• Freight Rail (TFR);
• Rail Engineering (TRE);
• Pipelines (TPL);
• Supporting, including Transnet Capital Projects.

vii
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Transnet has approximately 50,000 employees, a revenue of R49 billion per year (± e3.5 billion) and
assets worth R150 billion (± e11,- billion). TNPA assets contain more than one third (± R56 bil-
lion) of the total, since it owns the eight commercial ports in South Africa [TNPA, 02-2014].
During the project, the team is supervised from Delft University of Technology side by ir. Bas
Wijdeven and dr. ir. Phil Vardon. From WSP Group Africa side the team is supervised by
Andre van Tonder Pr Eng and Rob Leach. Other involved parties are sponsors Boskalis, Van
Oord and Delft Infrastructure & Mobility Initiative.

1.2 Durban Dig-Out Port Project

The port of Durban is South Africa’s premier port and hub of the whole region, especially for the
Johannesburg (Gauteng Province) area. The expectations are that the port is not able to cope with
the vessel and cargo increase in the coming decades. For the next 30 years, a container demand
growth of 4% per year is expected [TNPA, 05-2014]. For the economic situation of South Africa,
it is of utmost importance that the port expands accordingly in order to support the economic
growth of the country.
Since the current port is located within an environmentally sensitive area and surrounded by
developed urban properties, sufficient expansion of the port at its current location is not feasible
[A. Mather and K. Reddy, 2008]. The focus is therefore changed from expanding the port to the
development of a new port in the direct surroundings of Durban. The old international airport,
located along the coastline, is a suitable new location, see Figure 1.1. The property, located about
11 km south-eastwards from the city centre, is not in use since the FIFA world cup of soccer in
2010. Durban has built a new international airport north of the city (King Shaka International
Airport). Due to the growing demand of throughput for the containers, liquid bulk and vehicles,
the current port of Durban will eventually not be sufficient to handle the capacity demand anymore
[TNPA, 05-2014]. A part of the current port will therefore be expanded. However, this expansion
is not sufficient to handle all cargo for the coming 30 years. The design of the Durban Dig-Out
Port needs to cover the lack of space for capacity increase of the current port. A container terminal
with a 10,000,000 TEU throughput per year, a liquid bulk terminal with 5,000,000 kL throughput
per year and a vehicle terminal of 300,000 of units per year need to be constructed, in order to
meet the demand forecasts.

viii
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

(a) Durban on the map of South Africa

(b) The location of the potential new location (left) and the current port of Durban (right)

Figure 1.1: Location of the project shown at two scales (Source: Google Maps)

ix
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

1.3 Research Objectives

The project team has three distinguishable objectives which are summarised hereafter. The focus
lies with supplying TNPA with a full solution of the port expansion at the Durban Dig-Out Port
(DDOP). These research objectives are also defined in the Task Brief, which is agreed on November
the 17th 2014 by David John McGillewie and Selvan Pillay from TNPA.

Part 1: Design Global Port Layout Create a global port layout for the complete Durban
Dig-Out Port. The design is based on capacity demand, planning, boundary conditions and the
life-time of the port. This global port layout will include:
• Breakwaters;
• Sand by-pass system;
• Channels;
• Basins;
• Turning circles;
• Quay walls;
• Road and rail connections;
• Port Control and Port Authority Admin Buildings.

Part 2: Preliminary Design Create a preliminary design for specified aspects of the designed
DDOP lay-out, which fit within the various disciplines of the team (coastal-, geo- and hydraulic
engineering). To give the reader an idea, possible aspects are engineering the breakwater, entrance
channel, quay-wall and modelling coastal- and geo-engineering aspects.

Part 3: Recommendations on Vessel Size Comment on the TNPA’s choice of container

design vessel size for the new port in the Durban area over the next 50 years.

1.4 Context of the Report

This report presents the final outcome of Part 2 of the Project Durban research regarding the
Durban Dig-Out Port. Before this report, Project Durban has written an analysis report and a
report Part 1. In those reports the area was investigated and a possible port lay-out is given. The
breakwater, which is included in this lay-out, is worked out in a preliminary state in this report.
Project Durban continues their research after Part 2 with the objective to give a recommendation
to the port authority on the assumed vessel size. This will be presented by the end of January
2015.

x
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

1.5 Report Outline

The outline of this report is as follows. In Chapter 2 the length and orientation of the breakwater
are simulated in a wave model program. After that the cross section is determined for the two
breakwaters on different locations. With the dimensions of the cross sections known, a simulation
of the stability is done in Chapter 4. To do this, the program Plaxis 2D is used. In Part 1 the
sand bypass system en the mangroves are shortly mentioned. In Chapter 6 and 7 those two topics
are explained more indepth and the aim is to deliver a concept solution.

xi
Chapter 2

Length and Orientation

The length and the orientation of the breakwaters in the proposed port layout are based on
nautical rules of thumb. The length of the breakwater is determined by the length of the entrance
channel and the orientation is based on the present wave conditions. In the proposed layout, two
different configurations for the breakwaters are distinguished in terms of orientation and length.
The proposed longer breakwaters are based on a relatively high maximum vessel entrance speed
(10 kn) and the proposed shorter breakwaters on a lower maximum allowed entrance speed (6 kn).
A shorter breakwater might be preferred since the depth increases rapidly in front of the project
locations coast. This results in a rapid increase in construction costs for the longer breakwaters.
The implementation of shorter breakwaters however causes more wave action inside the port, which
possibly results in much downtime at the terminals. In this chapter an investigation is performed
with the numerical Delft3D-Wave software, the wave action in the port is determined for the two
different configurations. Especially the wave action at the liquid bulk terminal is investigated since
it is located next to the entrance channel of the port.
This chapter has the following structure. First the approach for the research is introduced, in
which the reason for the selection of the Delft3D-Wave software is explained. Afterwards the
model setup is presented. After the results of the simulations, the limitations of the methodology
and the software are discussed.

2.1 Approach
In order to come up with a conclusion and recommendations about the breakwater length and
orientation with regard to the hydraulic conditions, several steps are taken. In this section these
are elaborated. First, the methodology is discussed. Afterwards, a brief introduction into the used
Delft3D-Wave software is given.
Methodology
Since there are several processes which affect the propagation of waves, it is a complicated process
to determine analytically. Next to that, there are no measurements available because it is a new
port. Therefore it is decided to investigate the wave action in the port with the numerical Delft3D-
Wave software. This software simulates the evolution of random, short-crested, wind-generated
waves in estuaries, tidal inlets, and lakes.

1
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

A model grid has been developed to represent the proposed port layout. Afterwards the bathymetry
of the project location is implemented in the grid. The boundary conditions are based on the
apparent wave condition at the project site. Since there are two dominant wave directions, two
simulations are performed. For both wave directions the wave action in the turning circle and
at the liquid bulk terminal are determined. A distinction is made between storm conditions and
normal conditions, to investigate the behaviour of the system during extreme and during average
conditions. The different configurations of the breakwaters are imposed in the model to compare
the results for both wave conditions and directions.
Obviously the methodology and the numerical model have their limitations. In the model setup
section the different physical and numerical processes which are included or excluded from the
simulations are discussed. In the discussion the limitations of the methodology and the software
are elaborated.
Delft3D - Wave
In this subsection the used Delft3D-Wave model is briefly introduced. The software is part of
the Delft3D software package from Deltares. The model provides a user friendly interface which
interacts with the third generation 2D model SWAN (Simulating WAves Nearshore). This 2D
model is often used to simulate the evolution of random, short-crested, wind-generated waves in
estuaries, tidal inlets and lakes.
Based on the discrete spectral action balance equation the SWAN model accommodates random
wave fields from various orientations simultaneously. SWAN computes the evolution of random,
short-crested waves in coastal regions with deep, intermediate and shallow water and ambient
currents. The SWAN model accounts for (refractive) propagation due to current and depth and
represents the processes of wave generation by wind, dissipation due to whitecapping, bottom
friction, depth-induced wave breaking and non-linear wave-wave interactions (both quadruplets
and triads) [Deltares, 2014].

2.2 Model setup

In this section the setup of the simulation is presented. The Graphical User Interface of Delft3d-
Wave is followed and the different options are discussed. For the not discussed processes and
parameters, the Delft3D-Wave default values and settings are used.
Hydrodynamics
In the simulations there is no coupling made with the hydrodynamics of the system. Tidal el-
evations and currents are not taken into account. The average water level is found equal to
1.1m CD and the design water level is taken 3.7m CD based on research done by AECOM
[AECOM, 2012].
Grid
For the proposed port layout a curvilinear grid is developed with the Delft3D module RGFGRID.
So-called splines (the dotted lines in Figure 2.1) are drawn to represent both the wave directions
and the land boundaries of the proposed port layout. The area of interest, which is the entrance
channel of the port, has a higher density of splines. The other areas have a lower density of splines
to save computational time. The computational grid is created by RGFGRID in between the
splines. Since following the land boundary and the wave directions are contradictory goals, some

2
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

inaccuracies are found in the grid. At some location the land boundary is not exactly followed,
which causes staircases in the grid. At other locations the splines are not exactly orthogonal, which
results is computational errors. The grid has the following properties:
• Grid M-Direction (Horizontal grid lines): 325
• Grid N-Direction (Vertical grid lines): 486
• Grid cells: 114,021

Figure 2.1: Computational grid and splines

Bathymetry
The bathymetry of the study area is based on topographic charts. The depths are manually
implemented with the Delft3D module QUICKIN. The 20 meter and 30 meter depth contours as
indicated in the proposed layout are followed. The bathymetry which is implemented in the model
is presented in Figure 2.2.
Timeframe
Because there are no changes in the boundary conditions over time, the result reaches an equilib-
rium state. Therefore the results of the simulations do not include time series.
Boundary Conditions
The boundary conditions represent the wave climate before the coast of the Durban Bluff. Near
shore swell waves are generally directed from the sector S to SE with periods longer than 10 s.
There are however also waves directed from the sector SW to ENE near the shore, these waves
are commonly wind waves with periods between 4 s and 10 s. These two wave directions are
represented and imposed in the simulation as shown in Table 2.1.
The storm wave characteristics represent the conditions during extreme situations. Also the effect
of normal wave conditions is researched and is imposed in the model as shown in last two rows of
2.1. The storm significant wave height is found from AECOM [AECOM, 2012], where also the 1%
exceedance wave heights are found.

3
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Figure 2.2: Bathymetry

Table 2.1: Delft3D-Wave - Imposed wave conditions

Significant Peak wave pe- Direction [de- Directional

wave height riod Tp [s] grees] standard devi-
Hs [m] ation [degrees]
Storm south 10.1 12 0 10
Storm east 6.3 6 270 10
1% Exceedance south 3.2 12 0 10
1% Exceedance east 3.2 6 270 10

Obstacles
The breakwaters are implemented in the simulation by means of obstacles. The heights of the
dams and the reflection coefficients are based on the breakwater cross section design presented
in Chapter 3. The transmission of the breakwater is calculated with the default alpha and beta
values (see Table 2.2).

Table 2.2: Delft3D-Wave - Imposed obstacles

Height [m +CD] Reflection coefficient [-] Alpha [-] Beta [-]

Southern Breakwater 18.7 0.25 2.6 0.15
Northern Breakwater 11.7 0.25 2.6 0.15

4
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Physical Processes and Parameters

The input wave conditions are determined near shore by AECOM [AECOM, 2012]. The trans-
formation of waves from deep to shallow water is therefore not included in this simulation. The
computational domain is also relatively small. Because of that, the influence of wind is not in-
cluded. Depth induced breaking and bottom friction are included since the water is relatively
shallow. The model does calculate the radiation stress of the waves and takes whitecapping into
account according to the theory of Komen et al (1984). The processes concerning wave propaga-
tion in the spectral space taken into account are refraction and frequency shift. The wave set-up is
chosen to be left out of the model, due to inaccuracies of the SWAN model, see Subsection 2.4 for
more information. Diffraction around obstacles is disregarded due to iteration flaws in the model.
Also non-linear triad interactions (LTA) are disregarded.
The physical constants implemented are a gravitational acceleration of 9.81 m/s2 and a water
density of 1030 kg/m3 .
Output
The simulation results shown in the next subsection are presented as map files. Next to that, two
horizontal grid lines are presented with their relative significant wave height to represent the wave
penetration in the turning circle and in the liquid bulk terminal. The specific M lines followed are
133 for the liquid bulk terminal and 150 for the turning circle.

2.3 Results
In this section the results of the wave simulations are presented. First the wave penetration
during storm conditions is presented. This scenario is chosen to represent the extreme conditions
at the project location. For these simulations the imposed boundary conditions are the storm
significant wave heights. Afterwards the simulations are presented for significant wave heights
with a probability of exceedance of 1% to represent the normal conditions at the project location.
A comparison is made between the two configurations of the breakwaters.
In Figure 2.3 and 2.4 the penetration of waves under storm conditions originating from respectively
the east and the from the south is displayed. Please note that the scales for the different directions
are not the same. Next to that, the scale is bounded. Higher values than the imposed scale are
not represented in the figures.

5
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

(a) Long breakwater (b) Short breakwater

Figure 2.3: Storm significant wave height originating from the east

For waves originating from the east there is more wave action in the entrance channel with the
longer breakwaters configuration. Probably this is caused by the reflection of the long southern
breakwater. The shorter configuration of the breakwaters result in lower wave action and therefore
a more sheltered entrance channel.

(a) Long breakwater (b) Short breakwater

Figure 2.4: Storm significant wave height originating from the south

The southern approaching waves show the contrary result. The longer breakwaters create a more
sheltered entrance channel for the vessels. Especially at the liquid bulk terminal severe wave
heights are apparent during storm conditions.

6
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

As explained in the model setup, two grid lines are followed to investigate the wave heights in the
entrance channel and in the liquid bulk terminal. The location of liquid bulk terminal is about
1500 to 2000 m offshore from the start of the turning circle. Figure 2.5 presents the wave heights
along the indicated grid lines.

(a) Liquid bulk terminal (b) Turning circle

Figure 2.5: Storm significant wave height penetration

The simulation shows that there is a small difference between the different breakwaters options
regarding the waves approaching from the east. In both cases the wave heights inside the liquid
bulk terminal are less than 0.5 m and also inside the turning circle less than 0.5 m. Concerning
the storm waves approaching from the south, there is a difference. The longer breakwaters shelter
the liquid bulk terminal better, only allowing approximately 1 m waves. The shorter breakwaters
cause wave heights of approximately 2.75 m inside the liquid bulk terminal, which will result
in downtime. According to H. Ligteringen [H. Ligteringen and H. Velsink, 2012], mooring cannot
take place when wave heights exceed 1.5 m.
It has to be remarked however that the results are presented for storm conditions. Since these
conditions are only expected 6 times per year, this downtime could be accepted. To get a better
idea of the waves inside the entrance channel of the port under normal conditions, the 1% exceeding
significant wave height is imposed in the model. The results of this simulation are presented in
Figure 2.6. From this figure can be read that the 1% exceeding significant wave height at the liquid
bulk terminal becomes less than 1 meter. In these conditions the terminal can be operational. The
wave heights in the turning circle which are only exceeded in 1% of the time are lower than 0.5
m.

7
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

(a) Liquid bulk terminal (b) Turning circle

Figure 2.6: 1% exceedance significant wave height penetration

2.4 Discussion
As this is a computer simulation model, the results do not necessarily represent reality. Next
to that, there are no data available to validate the model results. Therefore it is important
to understand the shortcomings and simplifications of this method while interpreting the re-
sults. In this section the limitations of the followed methodology are discussed. Also the lim-
itations of the used software are discussed, these are extracted from the SWAN user manual
[The SWAN team, 2006].
Grid
Inaccuracies in the grid can be ascribed to the fact that the real world is modelled by a set of
grid cells of a certain area. This causes for example averaging of bathymetry and velocity over
the cells in horizontal, and in vertical direction as well when running the model. When processes
have a length scale smaller than the grid cell, these processes disappear and have to be taken into
account by introducing new parameters and formulas for the estimation of those lost processes.
Due to the orthogonality and the size of grid cells the grid will not perfectly follow the real land
boundary.
As the land boundaries are not perfectly followed, errors are introduced. By decreasing the size
of the grid cells, land boundaries are better represented. Decreasing the size of the grid cells also
decreases the amount of averaging that is done, which will lead to a better simulation of the actual
wave patterns. However, decreasing the size of grid cells means that more grid cells are needed for
the area. This leads to more computational time that was not present in this study. For a better
representation of the waves in the area a denser grid is needed.
Bathymetry
The bathymetry inserted in the DDOP model is derived from geographical contour lines and only
gives an indication of the bottom depth. In between the contour lines, the computer model Delft3D
interpolates the bottom depth by taking into account the inserted contour lines. The result is a
smooth bottom profile. This bottom profile might differ from the real bottom profile and can cause
inaccuracies with regard to the wave height and wave orientation.

8
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Wave Conditions
Two wave design conditions and two average wave conditions are applied. The determined wave
heights, period, orientation and spreading represent simplifications from measurements [AECOM, 2012]
in the area. Therefore, the actual wave conditions might differ from the simulated conditions due
to both simplifications and the measurement errors.
Currents
The currents along the coast might affect the wave heights and orientation. These effects are not
included in the model, which will create small inaccuracies.
Wind
Waves generated by wind are not taken into account in the input of the DDOP model. In other
words, the model does not calculate the wave height with the wind input. These waves are already
taken into account in the wave conditions, and were calculated by AECOM [AECOM, 2012]. The
normative wind wave direction is from the north-east. As can be read in item “Wave conditions”,
the modelled wind waves are inaccurate due to simplifications and measurement errors.
Time Step
To solve differential equations with a computer, these equations need to be discretized, in both
time and space. This will lead to errors, as the differential equations cover a continuous process.
These errors can therefore not be avoided using a numerical model. However, decreasing the time
step also decreases these numerical errors. On the other hand, more computational time will then
be needed.
Refraction
When depth varies considerably over one spatial grid step (which is the case in this project), this
results in inaccurate refraction computations in such grid steps. The inaccurately computed effects
may radiate far into the computational area.
Diffraction
The computation of diffraction in arbitrary geophysical conditions is rather complicated and re-
quires considerable computing effort. To avoid this, a phase-decoupled approach is employed so
that same qualitative behaviour of spatial redistribution and changes in wave direction is obtained.
Diffraction is not taken into account in the DDOP model due to computational iteration divergence
problems during the simulation runs.
Wave Induced Currents
SWAN does not calculate wave-induced currents. If relevant, such currents should be provided
as input to SWAN, e.g. from a circulation model which can be driven by waves from SWAN in
an iteration procedure. This process is time consuming and is not relevant for the wave height
determination in the DDOP.
Wave Induced Setup
As an option SWAN computes wave induced setup. In one-dimensional cases the computations
are based on exact equations. In 2D cases, the computations are based on approximate equations
(the effects of wave-induced currents are ignored). The wave induced set-up option is not used in
the DDOP model, due to the inaccuracy of the program.
Triad Wave-Wave Interactions
The approximation for the triad wave-wave interactions depends on the width of the directional
distribution of the wave spectrum. The present tuning in SWAN seems to work reasonably well in

9
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

many cases but it has been obtained from observations in a narrow wave flume and might be not be
accurate. The obtained results in wave height therefore need to be interpreted with caution.
Quadruplet Wave-Wave Interactions
The approximation for the quadruplet wave-wave interactions depends on the width of the direc-
tional distribution of the wave spectrum. It seems to work reasonably in many cases but it is a
poor approximation for long-crested waves (narrow directional distribution). It also depends on
the frequency resolution. It seems to work reasonably in many cases but it is a poor approximation
for frequency resolutions with ratios very different from 10 %. This is a fundamental problem that
SWAN shares with other third-generation wave models.
Model Validation
It is not known whether this model represents the real situation, because the new port has not been
constructed and therefore no practical reference data is available. Furthermore, measurements at
location are costly and time consuming and therefore cannot be done with regard to finances and
the amount of time in the project. As a result it is recommended to interpret the results with
care.

10
Chapter 3

Cross Section

This chapter covers the design of the cross sections of the two breakwaters designed for the port.
The different stages of the design process are explained and the involved calculations are elaborated.
At the end of the chapter the outcome of the design process is presented in technical drawings.

3.1 Design Considerations

The starting point for the design involves several important choices that affect the whole of the
process. Following subsections cover these choices.

3.1.1 Type of Breakwater

The breakwaters are included in the proposed port layout. In this chapter a preliminary design of
the cross section of the breakwaters at different depths is presented. It is important to remark that
the proposed breakwater extends to a relatively large depth. For the design of the breakwater,
the first decision is to design a rubble mound breakwater, a caisson breakwater or a composite
breakwater. The design of a caisson at the DDOP is found problematic since the window in
which the elements can be placed is small. Therefore it is decided to design a rubble mound
breakwater.

11
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

3.1.2 Type of Armour Units

The armour layer is designed with Xblocs. The advantages of these concrete armour elements are
summed up hereafter:
• Xbloc is a single layer armour unit. This reduces the number of necessary units compared to
the double layer armour layers. For the large breakwater necessary in the DDOP this saves
a significant amount of units.
• Xbloc has a high stability coefficient and a low packing density which results in a lower
concrete demand.
• The units are stacked in multiple layers to minimize the stacking area for construction.
• The shape is such that it can be placed in a straightforward way, with the right density. This
results in high level of freedom in placing the Xblocs.
With regard to the mentioned advantages and after discussions with experienced engineers it is
decided to use Xblocs for the design of the the armour layer.

3.1.3 Probability of Failure

The size of the armour layer elements is determined probabilistic since the used parameters are
uncertain. Therefore an accepted probability of failure is determined for the design storm. The
design life of the breakwater is 50 years. The probability of failure during the lifetime of the
structure is an economic decision and a probability of failure between 5% and 20% is recommended
[H.J. Verhagen, K. d’Angremond, F. van Roode, 2009]. For this design, the accepted probability
of failure of the structure is 10% in 50 years. To translate this failure probability to a design storm
frequency per year, the Poisson distribution is used:

1
f =− · ln(1 − p) (3.1)
tl

1
Hence the storm frequency per year becomes: f = − · ln(1 − 0.10) = 2.11 · 10−3 .
50
From offshore NCEP data the average number of storms per year is determined. The data consists
of a 15 years 3 hourly significant wave height record. Assuming a storm threshold, the storms
per year are determined. The 15 years average amount of annual storms is 6. The allowed failure
probability per storm event is therefore 3.51 · 10−4 . This value will be the upper limit of the
probability of failure during the probabilistic calculations.

3.2 Boundary Conditions

The hydraulic boundary conditions are discussed in this section. The cross section of the breakwa-
ters is designed at different depths and with different wave attack. For the different cross sections
these are the variable parameters. The design water level is applicable to both breakwaters.

12
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

3.2.1 Design Water Level

The lowest astronomical tide level is equal to 0 m CD. The design water level is determined with
the mean high water level, the water level surge and the expected sea level rise. The values which
are extracted from the analysis report [Project Durban, 2014] are summed up below in Table 3.2.
The height of the breakwaters is based on the design water level and the design wave. The height
of the toe is based on the lowest apparent water level and the design wave.

Table 3.1: Determination of the design water level

Tide MHWS 2.01 [m]

Water level surge (1:100 year) 0.72 [m]
Sea level rise 1.00 [m]
Cumulative 3.73 [m]

The design water level for the crest of the breakwater is set at + 3.7 m CD. As explained, there
is designed for a storm frequency with a frequency of 2.11e-3 per year. Since there is limited
amount of data available, there is decided to include a storm water level surge with return period
of 100 years and also a significant storm wave height with a return period of 100 years. This does
not exactly match the design storm frequency, and therefore a probabilistic calculation method is
performed.

3.2.2 Wave Conditions

The significant storm wave height is extracted from [AECOM, 2012] The storm wave heights
are calculated for different directions with a return period of 100 years. The wave heights are
calculated with wave transformation modelling. A water level rise of 3.7 meter is included in the
simulations, which makes the output conservative. As explained the design storm frequency is not
exactly matched, but these data are the most representative and the uncertainty is included in the
probabilistic calculation of the armour units. The northern breakwater is designed for the waves
approaching from the east, the southern breakwater is designed for waves approaching from the
south. It is assumed that the wave heights along the breakwater are constant.

Table 3.2: Design wave heights

Hss [m]
Southern storms 10.1
Eastern storms 6.3

13
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

3.3 Probabilistic Determination of the Armour Units

Due to the uncertainties in wave conditions a probabilistic method is used to determine the armour
layer units. The probabilistic Monte Carlo method is used, which calculates the stability of the
units 4e6 times. The input parameters all have a probabilistic distribution. If the limit state
function becomes lower than zero, it means that the unit has failed. The failure probability is the
amount of failures divided by the total amount of calculations. In this section the probabilistic
calculations are elaborated.

3.3.1 Limit State Function

To determine the the size of the armour layer probabilistic, a limit state function is developed.
The limit state function is constructed such that the unit fails if the result becomes smaller than
zero. The unit does not fail if the result is larger than zero. The limit state function is based on
the Xbloc manual [Delta Marine Consultants, 2011], which gives the stability formula as presented
in Equation 3.2.

 3
Hs
V = (3.2)
2.77 · 4
in which,
• V = Volume Xbloc [m3 ]
• Hs = Significant wave height [m]
• 4 = Relative density [-]
With this stability formula, the limit state function is determined, see Equation 3.3.

 3
Hs
Z=V − (3.3)
2.77 · 4

As presented in Equation 3.3 the volume of the required units is an input in the equation. Different
unit volumes of the XBloc are implemented. The designed unit is the smallest unit for which the
probability of failure is smaller than the calculated probability of failure per event.

14
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

3.3.2 Parameters and Distributions

The input parameters in Equation 3.3 have a probabilistic distribution. In this subsection the used
distributions for the calculations are presented. The more uncertain a parameter is, the higher the
deviation from the mean value is chosen.
Next to that, the significant wave height depends on the location along the breakwater. The stone
and water density are constant on every location. The input parameters and their distributions
are summed up in Table 3.3.

Table 3.3: Parameters and distributions

Parameter Distribution Mean value [µ] Standard Deviation [σ]

ρc Gaussian 2450 [kg/m3 ] 75 [kg/m3 ]
ρw Gaussian 1030 [kg/m3 ] 10 [kg/m3 ]
Hs Gaussian
V Triangular

Most of the parameters are Gaussian (normal) distributed. This distribution is characterised with
a mean value and a standard deviation. The standard deviation is a measure for the uncertainty
of the mean value. ρc and ρw are assumed equal for every cross section. The value of the two
parameters is explained first. The value of the wave height is specified for every cross section
afterwards.
ρc
The Xbloc is made in-situ, therefore the stones do not have exactly the same density. For instance a
change in weather conditions influences the quality of the concrete. The concrete density is assumed
Gaussian distributed with a standard deviation of 75 kg/m3 (3% of the mean value).
ρw
The density of water is related to the pressure, the temperature and the salinity of the water. Small
changes in the water density are possible, the average over a larger area stays however relatively
constant. The density of water is assumed Gaussian distributed with a standard deviation of 10
kg/m3 (1% of the mean value).
Hs
Also the significant storm wave height is Gaussian distributed. The data to determine the signifi-
cant storm wave height are limited. Also the return period of the obtained storm wave height does
not exactly match the design storm frequency. Therefore a relatively large standard deviation is
imposed for the design wave height. The uncertainty is in this way included in the determination
of the armour units.
V
The armour unit volume is assumed triangular distributed. The triangular distribution has a
defined upper and lower boundary. The volume of the armour units is bounded since these are
constructed on site. Large deviations from the mean value are not expected.

15
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

3.3.3 Armour Unit Size

The most important input parameter in the design of the armour layer is the significant storm
wave height. The design process started with the southern breakwater, see remark at the end of
this subsection. Afterwards, the northern breakwater is designed at two different locations. This
is presented below. As explained, the probability of failure per event needs to be smaller than
3.5 · 10−4
Tip of northern breakwater [-25 m CD]
The tip of the northern breakwater reaches to a depth of 25 m below CD. In the area around
the tip of the breakwater the same stone sizes are used. Both on the sea and on the lee side the
same armour units are applied. The length of the southern breakwater is designed such that the
incoming waves from the SSE cannot reach the northern breakwater. The governing wave height
is coming from the east with a significant storm wave height of 6 m. A 10% standard deviation is
is applied. In Table 3.4 the parameters are given with their distribution and values.

Table 3.4: Parameters and distributions at -25 m CD

Parameter Distribution Mean value [µ] Standard Deviation [σ]

ρc Gaussian 2450 [kg/m3 ] 75 [kg/m3 ]
ρw Gaussian 1030 [kg/m3 ] 10 [kg/m3 ]
Hs Gaussian 6.0 [m] 0.6 [m]
V Triangular 12 [m] Boundaries: 11.9 - 12.1 [m]

The calculated probability of failure for an Xbloc with a volume of 12 m3 is 2.1 · 10−4 . In Table
3.5 the characteristics of this Xbloc are given.

Table 3.5: Designed Xbloc for tip of northern breakwater

Xbloc characteristics
Volume 12 [m3 ]
Weight 29 [t]
Height 3.3 [m]

16
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Northern breakwater [-15 m CD, sea side]

The governing waves in a depth of -15 m CD at sea side of the breakwater are the waves from
the East. The waves arrive at the breakwater under a small angel, which is not included in the
calculation. The weight of the armour units at sea side is equal to the weight as calculated for the
tip of the breakwater.
Northern breakwater [-15 m CD, lee side]
On the protected lee side of the breakwater it is not necessary to design the armour layer with
similar Xbloc sizes. Application of smaller sizes is preferred for economic reasons. Therefore the
significant wave height at the lee side at a depth of -15 m is determined. This information is
extracted from the wave model, see Chapter 2. With this model a first estimation is made about
the expected wave height within the breakwaters. A 3 m high wave is used. Due to the high level
of uncertainty, a deviation of 17% is imposed. In Table 3.6 the parameters are given with their
distribution and values.

Table 3.6: Parameters and distributions at -15 m CD

Parameter Distribution Mean value [µ] Standard Deviation [σ]

ρc Gaussian 2450 [kg/m3 ] 75 [kg/m3 ]
ρw Gaussian 1030 [kg/m3 ] 10 [kg/m3 ]
Hs Gaussian 3.0 [m] 0.5 [m]
V Triangular 2.5 [m] Boundaries: 2.4 - 2.6 [m]

The calculated probability of failure for an Xbloc with a volume of 2.5 m3 is 9.75 · 10−5 . In Table
3.7 the characteristics of this Xbloc are given.

Table 3.7: Designed Xbloc for lee side of northern breakwater

Xbloc characteristics
Volume 2.5 [m3 ]
Weight 6 [t]
Height 1.96 [m]

Remark:
A misinterpretation of the wave heights has lead to the wrong dimensions of the southern break-
water. Unfortunately this mistake was discovered when the design was already finished. Due to
the limited time available Project Durban has decided, in agreement with WSP staff, to go on with
the research to the northern breakwater with correct wave heights. The results of the Southern
breakwater are included in the report at the end of this chapter. It has to be reminded that the
breakwater is designed for too low wave heights. The design process was equal to the process of the
northern breakwater design. The cross section of the Southern breakwater are shown at the end of
this chapter.

17
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

3.3.4 Probabilistic Calculation Files

In the paragraphs below, the used Matlab M-files are given. With this code the probability of
failure of the armour layer is calculated. The first file displays the input parameters, which are
used in the limit state function given in the second file.
Input of parameters and distributions
1 %% d e f i n e t h e s t o c h a s t s
2 % c r e a t e a s t r u c t u r e with f i e l d s ’Name ’ , ’ D i s t r ’ , ’ Params ’ and ’
propertyName ’
3 stochast = struct ( . . .
4 ’Name ’ , { % d e f i n e th e s t o c h a s t i c v a r i a b l e names :
5 ’ RhoS ’ . . . % [ kg /m3 ] RhoS Density concrete
6 ’RhoW ’ . . . % [ kg /m3 ] RhoW D e n s i t y water
7 ’ Hs ’ . . . % [m] Hs S i g n i f i c a n t Wave
Height
8 ’V ’ . . . % [ m3 ] V Xbloc volume
9 } ,...
10 ’ D i s t r ’ , { % d e f i n e th e p r o b a b i l i t y d i s t r i b u t i o n f u n c t i o n s
11 @norm inv . . . % [ kg /m3 ] RhoS Density concrete
12 @norm inv . . . % [ kg /m3 ] RhoW D e n s i t y water
13 @norm inv . . . % [m] Hs S i g n i f i c a n t Wave
Height
14 @trian inv . . . % [ m3 ] V Xbloc volume
15 } ,...
16 ’ Params ’ , { % d e f i n e th e p a r a m e t e r s o f th e p r o b a b i l i t y
distribution functions
17 {2500 1 0 0 } . . . % [ kg /m3 ] RhoS Density concrete
18 {1030 1 0 } . . . % [ kg /m3 ] RhoW D e n s i t y water
19 {6 0 . 6 } . . . % [m] Hs S i g n i f i c a n t Wave
Height
20 {11.9 12.1 1 2 } . . . % [ m3 ] V Xbloc volume
21 } ,...
22 ’ propertyName ’ , { % s p e c i f y h e r e t o c a l l t h e z−f u n c t i o n with
propertyname−p r o p e r t y v a l u e p a i r s
23 true . . . % [ kg /m3 ] RhoS Density concrete
24 true . . . % [ kg /m3 ] RhoW D e n s i t y water
25 true . . . % [m] D Nominal s t o n e
diameter
26 true . . . % [ m3 ] V Xbloc volume
27 } ...
28 );
29

30 %% main matter : running th e c a l c u l a t i o n

31 % run t h e c a l c u l a t i o n u s i n g Monte Carlo
32 resultMC = MC( . . .
33 ’ stochast ’ , stochast , . . .

18
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

34 ’ NrSamples ’ , 4 e6 , . . .
35 ’ x2zFunction ’ , @prob Durban x2z ) ;

Limit State Function

1 %% c r e a t e samples−s t r u c t u r e based on i n p u t arguments
2 samples = s t r u c t ( . . .
3 ’ RhoS ’ , [ ] , . . . % [ kg /m3 ] RhoS density concrete
4 ’RhoW ’ , [ ] , . . . % [ kg /m3 ] RhoW d e n s i t y water
5 ’ Hs ’ , [ ] , . . . % [m] Hs s i g n i f i c a n t waveheight
6 ’V ’ , [ ] ) ; % [ m3 ] V Xbloc volume
7

8 samples = s e t p r o p e r t y ( samples , v a r a r g i n { : } ) ;
9

10 %% c a l c u l a t e z−v a l u e s
11 % pre−a l l o c a t e z
12 z = nan ( s i z e ( samples . RhoS ) ) ;
13 % l o o p through a l l samples and d e r i v e z−v a l u e s
14 f o r i = 1 : l e n g t h ( samples . RhoS )
15 D e l t a = ( samples . RhoS ( i ) − samples .RhoW( i ) ) / samples .RhoW( i ) ;
% [ −] r e l a t i v e d e n s i t y
16 z ( i , : ) = samples .V( i ) −( samples . Hs ( i ) / ( 2 . 7 7 ∗ D e l t a ) ) ˆ 3 ;
% Z−f u n c t i o n
17 end

19
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

3.4 Cross Section Dimensions

In Section 3.3 the size of the armour units is determined. The other elements of the breakwater
are based on the dimensions of the armour units. In this section these elements are determined.
The following aspects are included in the preliminary design:
• Crest height determination
• Toe design
• Under- and filter layer design
The final cross sections are presented at the end of this section. First the two lay-outs of the
northern breakwater and after that the southern breakwater cross sections are given.

3.4.1 Crest Design

The governing process for the determination of the crest height of the breakwater is overtopping.
Wave run-up, reflection and transmission are disregarded in the determination of the crest height.
The allowable overtopping determines what the freeboard of the breakwater becomes. This is
calculated with Equation 3.4 [H.J. Verhagen, K. d’Angremond, F. van Roode, 2009].
 
q Rc
p = 0.2 · exp −2.3 · (3.4)
g · Hs3 Hs · γf · γb

In which,
• γf = roughness coefficient: 0.45 [-];
• γb = angle of attack: 1 [-];
• g = gravitational acceleration: 9.81 [m/s2 ];
• Hs = significant wave height [m];
Depends on the location.
• q = allowable overtopping quantity: 10 [l/s/m];
With this condition only vehicles driving slowly are able to access the breakwater during storm
condition.
• Rc = Crest height above design water level [m].
Depends on the location.
The freeboard is calculated with the Equation 3.4. Afterwards the relative freeboard is determined.
This is the ratio between the freeboard and the significant wave height. When this ratio is in
between 1.2 and 1.5, the minimum crest width in front of the crown wall is recommended to
be equal to the diameter of one Xbloc [Delta Marine Consultants, 2011]. A crown wall width
of 12m is applied to ensure enough space for maintenance, inspection and construction of the
breakwater.

20
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

3.4.2 Toe Design

Design rules for the toe design are given in the Xbloc manual [Delta Marine Consultants, 2011].
The required rock size depends on the water depth and the wave height. The required rock mass
is derived by the approach of Van de Meer, see Equation 3.5.

Hs
Dn50 =  2.7 ! (3.5)
ht 0.15
2 + 6.2 · Nod ·4
h

In which,
• Dn50 = nominal stone diameter [m];
• Hs = significant wave height [m];
Depends on the location
• ht = depth above toe [m];
According to the Xbloc guidelines this value is in between 1 - 1.5 ·Hs . The cross section is
designed with a value of 1.25 ·Hs .
• h = water depth in front of toe [m];
This value is depending on the thickness of the toe and therefore also on the stone diameter.
For this reason the formula becomes a iterative formula, because the output is also an input
value.
• Nod = damage value number of displaced units: 0.5 [-];
This value is recommended by the guidelines. Larger values lead to settlements of the armour
layer.
• 4 = relative stone density: 1.52 [-].
(ρs − ρw )
with ρs = 2600 [kg/m3 ] and ρw = 1030 [kg/m3 ]
ρw
For the dimensions of the toe also rules of thumb are given. The width of the toe is calculated as
6 · Dn 50. The height of the toe becomes 3 · Dn 50.

3.4.3 Under- and Filter Layers

The rock grading of the first under layer is found according to the Xbloc guidelines
[Delta Marine Consultants, 2011]. The grading is based on the size of the armour layer units. The
general filter layer rules are applied, which recommend the weight ratio between two consecutive
layers to be between 1/10 and 1/25 [H.J. Verhagen, K. d’Angremond, F. van Roode, 2009]. The
core material fills up the area between the layers and the bottom. Under the toe at the sea side,
large pressure gradients can build up. Filters are constructed there to prevent the washing out of
material.

21
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

3.5 Cross section Drawings

Cross section Northern Breakwater Tip at -25 m CD
As mentioned, the tip of the breakwater has a symmetrical cross section. The lee side is lowered
with one meter to save material and to allow for proper drainage of overtopping water. The Xbloc
guidelines indicate that the Xbloc armour layer needs to be placed on a slope from 2:3 to 3:4. For
the 29 tons Xbloc it is decided to use a 2:3 slope. The slope of the breakwater below toe level is
1:2. For construction it is not convenient to have different stone classes in one layer. The average
stone sizes are therefore constant per layer. As mentioned, the first under layer is determined with
the Xbloc guidelines and simplified filter rules.
In Table 3.8 the specific values for the northern breakwater at a depth of -25 m CD are given. In
the first drawing the cross section is presented.

Table 3.8: Parameters -25 m CD

Hs 6 [m]
Rc 8 [m]
ht 7.5 [m]

Cross section Northern Breakwater at -15 m CD

At a depth of -15 m the breakwater is not symmetrical. Smaller armour units are used at the lee
side. The sea side of the breakwater is designed with the same units as on the sea side at -25 m
depth. To reduce the needed volume of material, the slope on the lee side of the breakwater with
the smaller Xblocs is 3:4.
In Table 3.9 the specific values for the northern breakwater at a depth of -15 m CD are given. In
the second drawing this cross section is presented.

Table 3.9: Parameters -15 m CD

Hs 3 [m]
Rc 8 [m] (equal to sea side value)
ht 3.75 [m]

Cross sections Southern Breakwater

The third and fourth drawing show the cross sections of the southern breakwater. A remark about
these drawings is made on page 17.

22
 LEE SIDE SEA SIDE

Xbloc 29ton
1000-3000kg
3000-6000kg
60-300kg
10-60kg

NORTHERN BREAKWATER TIP CROSS SECTION

LEGEND

3000-6000kg

1000-3000kg
60-300kg

10-60kg
PROJECT DURBAN

BREAKWATER DESIGN
TOE DETAIL CREST DETAIL
 LEE SIDE SEA SIDE

Xbloc 6ton

Xbloc 29ton
300-1000kg
1000-3000kg
1000-3000kg
60-300kg 3000-6000kg
10-60kg 60-300kg

NORTHERN BREAKWATER CROSS SECTION

LEGEND

1000-3000kg
3000-6000kg
300-1000kg
60-300kg 1000-3000kg
10-60kg 60-300kg
10-60kg
PROJECT DURBAN

BREAKWATER DESIGN
TOE LEE SIDE DETAIL TOE SEA SIDE DETAIL
 SEA SIDE LEE SIDE

Xbloc 29ton
1000-3000kg
3000-6000kg
60-300kg
10-60kg

SOUTHERN BREAKWATER TIP CROSS SECTION

LEGEND

3000-6000kg

1000-3000kg
60-300kg

10-60kg
PROJECT DURBAN

BREAKWATER DESIGN
TOE DETAIL CREST DETAIL
 SEA SIDE LEE SIDE

Xbloc 29ton
1000-3000kg
3000-6000kg
60-300kg
10-60kg

SOUTHERN BREAKWATER CROSS SECTION

LEGEND

3000-6000kg

1000-3000kg
60-300kg
10-60kg
PROJECT DURBAN

BREAKWATER DESIGN
TOE LEE SIDE DETAIL CREST DETAIL
Chapter 4

Settlement and Stability Analysis

In this section the breakwater cross sections are modelled with the finite element software Plaxis2D.
This is a sophisticated software program used by professional geo-engineers around the world. First
the approach of this research is elaborated. Then, the model setup is described. After that,the
results are presented. This section concludes with a discussion and recommendations for further
research.

4.1 Approach
In this section the approach of the numerical modelling is elaborated. First, the goal of the
analysis is explained. Thereafter, the software Plaxis2D is briefly introduced. Then, the breakwater
implementation in Plaxis from the drawings is discussed. Finally, an overview is given of the
different material models in Plaxis2D and which material models can be used for this problem.
Goal of Plaxis Modelling
The goal of this analysis is to calculate the settlements and the overall stability of the structure.
Also, a flow function is defined for the waves to model the wave action on the breakwater. This
flow function is not as sophisticated modelled as in Delft3D, see Chapter 2, but it gives an idea of
the impact on the breakwater.
Plaxis 2D Software
Plaxis2D is a two-dimensional finite element program, developed for the analysis of deformation,
stability and groundwater flow in geotechnical engineering [R. Brinkgreve et. al, 2014]. Plaxis is
a geotechnical simulation tool and its models can be regarded as a qualitative representation of
soil behaviour whereas the model parameters are used to quantify the soil characteristics. The
simulation remains an approximation and involves some inevitable numerical errors and modelling
errors.
Breakwater Implementation in Plaxis
The northern breakwater is modelled at full scale in Plaxis. As mentioned in the southern breakwa-
ter is designed with incorrect wave heights. For this reason southern breakwater is not modelled in
Plaxis. Full scale modelling has several advantages and disadvantages that need to be mentioned.
The advantages are that a full scale model comes closer to reality than a scaled model. Also, the

27
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

settlements and displacements of the breakwater are better readable. The disadvantages are that
the model becomes quite big and therefore the calculation time is longer.
From the cross section drawings the coordinates of the intersections are determined. The bottom
left part of the breakwater has an x-coordinate of 0 m CD and an y-coordinate of -15 m CD and -25
m CD, respectively. Figure 4.1 shows the cross section in Plaxis for the -25 m CD breakwater.
The -15 m CD breakwater is also modelled, and the results of both simulations are presented in
Section 4.3.

Figure 4.1: Cross section in Plaxis for the -25m CD breakwater

Material Models
When a new material is created, a material model needs to be set to this material. Different
material models are designed to cope with different materials, applications and accuracy of the
modelling. A short overview of the different material models are given below:
• Linear Elastic (LE);
• Mohr-Coulomb (MC);
• Hardening Soil (HS);
• Hardening Soil with small-strain stiffness (HSsmall);
• Soft Soil (SS);
• Soft Soil Creep (SSC);
• Jointed Rock (JR);
• Modified Cam-Clay (MMC);
• NGI-ADP;
• Hoek-Brown (HB).
To keep the report concise, only the relevant material models for this application are considered.
The model consists of a layer of bedrock (Siltstone) and a covering sand layer, both situated below
the breakwater. For the bedrock the Linear-Elastic model can be used
[Plaxis2D Reference Manual, 2014]. The best material model to model the sand layer is the Hard-
ening Soil small model, but this requires highly specialized soil parameters, which are unlikely to
obtain. Therefore, depending on available parameters, the Mohr-Coulomb model or the Hardening
Soil model should be used. Regarding the rock of which the breakwater is constructed, the best
material model to use is the Hoek-Brown model. Here again, it depends on the available data if
this model can be used.

28
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Otherwise, the Mohr-Coulomb model is used. For the concrete crown, the Linear Elastic model is
used. The Xblocs are modelled using the Mohr-Coulomb model.
Hereafter a short elaboration of the Linear Elastic-, Mohr-Coulomb-, Hardening Soil- and the
Hoek-Brown model is given:
• Linear Elastic (LE): isotropic elasticity model, E’ and ν’ (nu) are the most important pa-
rameters. It is not suitable to model soils, but can be used to model stiff volumes in soils
(like concrete walls and intact rock formations).
• Mohr-Coulomb (MC): linear elastic perfectly plastic model. The parameters are E’ and
ν’ (nu) for elasticity, φ’ and c’ for plasticity and ψ for dilatancy. It represents first-order
approximation of the soil/rock behaviour. This model is used as a first analysis of the
problem, to obtain first estimates of deformations.
• Hardening Soil model (HS): advanced model for soil simulation. The parameters are φ’ and
c’ for plasticity and ψ for dilatancy. Stiffness is described more accurately by using three
different input stiffnesses; triaxial loading stiffness (E50 ), triaxial unloading stiffness (Eur )
and oedometer loading stiffness (Eoed ). All stiffness increases with increasing pressure. Initial
soil conditions (pre-consolidation e.g.) play a significant role in most soil deformations.
• Hoek-Brown model (HB): The HB model is an isotropic elastic perfectly-plastic model for
weathered rock based on the Hoek-Brown failure criterion. This non-linear stress-dependent
criterion describes shear failure and tensile failure by a continuous function. Besides the
elastic parameters (E and ν’ (nu)), the model involves practical rock parameters as uni-axial
compressive strength of the intact rock (σci ), the Geological Strength Index (GSI) and the
disturbance factor (D).
These different material models all have several limitations. These limitations are summed up
hereafter, because it is important to realize to what extent the different material models can give
accurate results:
• Linear Elastic (LE): Soil behaviour is highly non-linear and irreversible. The linear model is
not sufficiently capable to capture essential features of the soil. Be careful with this model
for materials that are loaded up to their material strength.
• Mohr-Coulomb (MC): This first order model includes only a limited number of features
that soil behaviour shows in reality. The model does neither include stress-dependency nor
stress-path dependency nor strain dependency of stiffness or anisotropic stiffness.
• Hardening Soil (HS): It is a hardening model that does not account for softening due to
soil dilatancy and de-bonding effects. In fact, it is an isotropic hardening model so that
it models neither hysteretic and cyclic loading nor cyclic mobility. The model does not
distinguish between large stiffness at small strain and reduced stiffness at engineering strain
levels. The user has to select the stiffness parameter in accordance with the dominant strain
levels in the application. The calculation time is longer than the Mohr-Coulomb model, since
the material stiffness matrix is formed and decomposed in each calculation step.
• Hoek-Brown (HB): The model is not suitable for stratified or jointed rock sections with a
significant anisotropic stiffness or with one or more dominant sliding directions.

29
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

4.2 Model Setup

In this section the model setup is elaborated. First, the breakwater material parameters are
investigated. These are necessary for accurate model results. Thereafter, the meshing options
and the groundwater flow boundary conditions are described. This subsection closes with the
elaboration of different test scenarios for the consolidation and stability analysis of the northern
breakwater.
Breakwater Material Parameters
Accurate parameters are of great importance for accurate model results. The amount of avail-
able parameters and the material models that can be used are related to each other. The more
parameters available, the more advanced material models can be used.
TNPA carried out a geotechnical investigation of the area of the DDOP. In this investigation
boreholes are conducted onshore [Project Durban, 2014]. Also, eight boreholes are conducted near
the shore. These boreholes are used to discover the type of bedrock and the overlying layer. The
bedrock consists of SANDY SILTSTONE, Mzamba Formation (Cretaceous). The overlying layer
consists of dense, fine grained sand. In the two cross sections both the bedrock and the sand layer
are assumed to be rectangular. This is of course simplified, in reality the transition from bedrock
to sand layer is not that smooth.
It is assumed that the bedrock and overlying layer are representative for the location of the break-
water. The overlying sand layer should be of sufficient quality for the breakwater to be constructed
on. Additional geotechnical investigations should be done to determine the quality of the sand and
the conditions of the bedrock at the breakwater location.
As seen in Figure 4.2 and 4.3, the breakwater consists of eight different materials:
• Core (10 – 60 kg);
• Filter layer (60 – 300 kg);
• Under layer (300 – 1000 kg);
• Under layer (1000 – 3000 kg);
• Toe (3000 – 6000 kg);
• Xbloc (6 ton);
• Xbloc (29 ton);
• Concrete crown.

30
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Figure 4.2: Cross section northern breakwater -15 m CD

Figure 4.3: Cross section northern breakwater -25 m CD

Including the bedrock and the overlying layer this results in a total of ten different materials
that need to be modelled. The amount of data that is available is limited. Therefore, it is not
possible to use the advanced models (Hardening Soil and Hoek-Brown model) for this numerical
analysis. The Linear Elastic model is used to model the bedrock and the concrete crown. The
Mohr-Coulomb model is used to model the sand layer, the rock fill material and the Xblocs. Table
4.1 gives an overview of the different parameters that are needed for the different material models
concerned.

31
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Table 4.1: Overview of the parameters necessary for the different material models

Parameter Name Linear Elastic model Mohr-Coulomb model Unit

General
Drainage Type Type Drained Drained -
Weight above p.l. γunsat X X kN/m3
Weight below p.l. γsat X X kN/m3
Parameters
Young’s Modulus E0 X X kN/m2
Poisson’s ratio ν0 X X -
Cohesion c0 X kN/m2
Friction angle φ0 X ◦
◦
Dilatancy angle ψ X
Flow
Horizontal permeability kx X X m/s
Vertical permeability ky X X m/s

The total amount of parameters that need to be determined are 84 (12 for the Linear Elastic model
and 72 for the Mohr-Coulomb model). This determination has to be done with limited information
and limited time. This has consequences for the accuracy of the model results. In consultation
with professionals at WSP Group Africa assumptions are made.
First, the stiffness and strength parameters of the rock fill material can be considered the same
when simple macro stability analysis is done. Also, the unsaturated and saturated unit weights
are estimated as a function of free volume between the rocks. The permeability for the different
layers is adopted using the diameter range of the rock fill.
The rock fill considered consists of granite, since the breakwater is a coastal structure
[A. Waltham, 2009]. Information about the stiffness parameters of rock fill is limited. The values
for plastic and cyclic loading largely depend on the crushing potential of the grains. Quartz sand
and excellent quality rock hardly crushes in the range of effective stresses relevant for hydraulic
engineering structures, like breakwaters. In this case, excellent rock is assumed to be used for the
breakwater. If this will actually be used, it is highly dependent on the availability and/or economic
factors. The best way to find good estimates of the parameters of the rock fill is to perform large
scale oedometer tests and shear tests. When excellent rock quality is assumed, the values of
the stiffness parameters are in the same range as quartz sand [CUR & CETMEF, 2007]. Also,
it is assumed that the quarried rock fill consists of round particles and is more or less uniformly
graded.
Three different input parameter sets are used to create a range of possible solutions. The first
set is based on the lower limit of values presented in the literature and is called ‘Lower Limit‘.
The second set is called ‘Average/Reference‘ and is based on average values from literature and
values from a reference project discussed with the engineers at WSP Group Africa. The third and
last set is based on the highest values presented in the literature and is called ‘Upper Limit‘. No
variation in the parameters of the concrete crown is taken into account.

32
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Table 4.2 and 4.3 give an overview of the different input parameters for the different ranges
considered. The unit weight and the permeability are not constant for the different types of rock
fill used in the breakwater, see Table 4.4. These values are used in the three different parameter
sets considered. Therefore, the parameter sets focus on the stiffness and the strength parameters,
which have the most influence on the settlements and stability.

Table 4.2: Input parameters for the materials that are used in the Linear Elastic model

Parameter Siltstone (bedrock) Concrete crown Unit

General
Parameter set ‘Lower Limit’ ‘Ave/Ref’ ‘Upper Limit’ - -
Drainage Type Drained Drained Drained Drained -
γunsat 21 23 25 25 kN/m3
γsat 21 23 25 25 kN/m3
Parameters
E0 5.0E+04 1.0E+05 1.0E+06 35.0E+06 kN/m2
ν0 0.25 0.3 0.35 0.15 -
Flow
kx 0.01E-06 0.01E-06 0.01E-06 0,1000E-12 m/s
ky 0.01E-06 0.01E-06 0.01E-06 0,1000E-12 m/s

Table 4.3: Input parameters for the materials that are used in the Mohr-Coulomb model

Parameter Sand Rock fill (breakwater) Unit

General
Parameter set ‘Low Lim’ ‘Ave/Ref’ ‘Up Lim’ ‘Low Lim’ ‘Ave/Ref’ ‘Up Lim’ -
Drainage Type Drained Drained Drained Drained Drained Drained -
γunsat 20 20 20 * * * kN/m3
γsat 20 20 20 * * * kN/m3
Parameter
E0 3.5E+04 5.0E+04 6.9E+04 1.0E+04 5.0E+04 1.0E+05 kN/m2
ν0 0.25 0.3 0.35 0.25 0.3 0.35 -
c0 0 0 0 0 0 1 kN/m2
φ0 30 32 35 40 45 55 ◦
◦
ψ 0 2 5 10 15 25
Flow
kx 8.25E-05 8.25E-05 8.25E-05 * * * m/s
ky 8.25E-05 8.25E-05 8.25E-05 * * * m/s

33
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Table 4.4: Parameters rock fill

Parameters rock fill Constant for all parameter sets Unit

Unit weight parameters γunsat γsat -
Core (60 – 100kg) 16.5 20.5 kN/m3
60 – 300kg 16.5 20.5 kN/m3
300 – 1000kg 16.5 20.5 kN/m3
1000 – 3000kg 16 20 kN/m3
Toe (3000 – 6000kg) 16 20 kN/m3
Flow Parameters kx ky -
Core (60 – 100kg) 0.3 0.3 m/s
60 – 300kg 0.5 0.5 m/s
300 – 1000kg 0.8 0.8 m/s
1000 – 3000kg 1 1 m/s
Toe (3000 – 6000kg) 1 1 m/s

The input parameters for the two types of Xblocs are considered the same. This is a simplification,
but considered the limited amount of data available it is not possible to make a difference between
the two Xbloc types. An overview is given in Table 4.5.

Table 4.5: Input parameters for Xbloc

Parameter Xbloc (6t & 29t) Unit

General
Parameter set ‘Low Lim’ ‘Ave/Ref’ ‘Up Lim’ -
Drainage Type Drained Drained Drained -
γunsat 16 16 16 kN/m3
γsat 20 20 20 kN/m3
Parameter
E0 5.0E+04 1.0E+05 1.0E+06 kN/m2
ν0 0 0 0 -
c0 0 0 0 kN/m2
φ0 45 50 55 ◦
◦
ψ 0 0 0
Flow
kx , ky 1 1 1 m/s

It must be noted that the output of the model is highly dependent on the input parameters. Since
the amount of data to conduct this numerical analysis is limited, a range of input parameters for
the different materials is considered. This results in a range of possible answers, which can then
be compared. In no way it is possible to give a final and single answer concerning the settlements
and the stability of the breakwater. This might be possible when the input parameters are very
accurate and the person in charge has a lot of experience with modelling breakwaters in finite
element software. Also, a geo-engineering expert of WSP Group Africa mentioned that numerical
modelling of breakwaters is a very difficult task and there is no perfect system to model these kind
of structures.

34
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Meshing Options
Finite element software uses elements and nodes at which the calculations are carried out. One
element consists either of 6 or 15 nodes, depending on the properties the user selects. It is possible
to select different mesh options, from very coarse to very fine. The finer the mesh, the more
elements and nodes are created. Increasing the number of elements increases the accuracy of the
model, but also increases the calculation time. A balance has to be found between accuracy and
calculation time.
In modelling the breakwater a concise investigation is done to find this balance. For the -25 meter
CD breakwater the settlements for different mesh options is compared to the required calculation
time, see Table 4.6. The parameter sets used is ‘Average/Reference’ for all mat.

Table 4.6: Settlements compared to required calculation time for different mesh options

Mesh option Settlement breakwater [m] Calculation time [s]

Very Coarse -0.2578 3
Coarse -0.2578 5
Medium -0.2574 7
Fine -0.2574 9
Very Fine -0.2574 10

It shows that the there is no further increase in accuracy after the medium mesh option. Therefore,
the medium mesh option is selected for the calculation.
Groundwater Flow Boundary Conditions
To make sure that the boundaries have no influence on the model results, the boundaries should
be at a sufficient distance from the breakwater. The width of the -25 m CD breakwater is 170
m. The bottom left part of the breakwater has x-coordinate 0 m and the bottom right part has
x-coordinate 170 m. Xmin is -170 m and Xmax is 340 m, which results in one breakwater width at
each side of the breakwater, which is sufficient. Ymin is set to -60 m CD, which is 35 m CD below
the breakwater, and Ymax is set +15 m CD.
Within Plaxis, groundwater flow boundary conditions can be set. The Xmin , Xmax and Ymax are
set to ‘Open’, which indicates that groundwater can freely flow through the boundaries. The Ymin
is set to ‘Closed’, which indicate that groundwater flow through the Ymin boundary is not possible.
These settings are commonly used for modelling dams and breakwaters.
Test Scenarios
Now, the test scenarios are elaborated. The different scenarios are discussed with engineers from
WSP Group Africa. In total five different test scenarios have been devised. The five different tests
that are conducted are:
• Subsoil consolidation analysis;
• Long term settlements, see Figure 4.4;
• Stability analysis, see Figure 4.5;
– Static; Lee side: design high water level (+3.7 m CD), Sea side: highest wave (+6.85
m CD),
– Static; Lee side: 0 m CD, Sea side: 0 m CD, which is design low water level,
– Harmonic; Lee side: design water level (+3,7 m CD), Sea side: wave modelled with a
harmonic function.

35
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

These tests are done using the three different parameter sets, for both the cross sections of the
northern breakwater. For the harmonic test scenario an exception is made, only the -25 meter
CD breakwater is considered due to very long calculation times. For all the four test scenarios, a
‘Safety‘ analysis is conducted to determine the safety factor of the breakwater.
Subsoil consolidation analysis
This test is conducted to get an idea of how significant the influence of the subsoil layers is on the
total settlement of the breakwater. A soft subsoil will obviously lead to bigger settlements than
a stiff subsoil. The three different parameter sets are used for the subsoil and the breakwater is
modelled using the ’Average/Reference’ parameter set.
Long Term Settlement
The long term settlement is modelled with the ‘Consolidation’ calculation type. In this calculation
the time interval is set to 50 years, in which it is assumed that no significant consolidation occurs
after this time. A water level of +1.1 m CD is set to be the normative water level, since this is
the average water level. The settlement calculated is the final settlement of the whole structure,
excluding the construction stages of the breakwater. In reality, the different construction stages
induce settlements as well.

Figure 4.4: Cross section in Plaxis for the -25m CD breakwater with water level of +1.10 m CD

Stability Analysis
For the stability the ‘Fully coupled flow-deformation’ calculation option is selected. Three different
stability analysis are conducted, in which two static and one harmonic. The first static calculation
is done to check whether the breakwater can cope with the (constant) horizontal external force
from the water. The second static calculation is done to check whether the model can cope with
the lowest water level possible.
Finally, the harmonic calculation is done to see how the breakwater reacts on a harmonic induced
waves. Actually, this is a changing horizontal force, which can lead to horizontal displacements in
the breakwater. The time considered is 10 wave periods, which is 1 minute. This amount of wave
periods is small, since the average time of a storm can be estimated to be between 6 and 12 hours.
The choice to decrease the amount of periods is due to the long calculation time. Ten wave periods
already takes more than 30 minutes to calculate. Related to this, the decision is made that only
the -25 meter CD northern breakwater cross section is modelled.

36
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

(a) Lee side waterlevel of +3.70 m CD and sea side waterlevel of +6.85 m CD

(b) Waterlevel of +0 m CD

(c) Waterlevel of +3.70 m CD and amplitude of 3.15 m

Figure 4.5: Breakwater cross section scenarios for northern breakwater analysis in Plaxis

37
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

4.3 Results
In this section the results of the Plaxis numerical modelling of the breakwater are elaborated.
Subsoil Consolidation Analysis
The results of the consolidation analysis using different parameter sets for the subsoil is shown in
Table 4.7. As expected, the material sets with the lowest stiffness and strength give the biggest
settlement. In further analysis the Average/Reference parameter sets are used for the subsoil of
the breakwater.

Table 4.7: Results of the consolidation analysis using different parameter sets for the subsoil

Parameter set Settlement [m] Safety Factor [-]

Lower Limit 0.388 1.47
Average/Reference 0.257 1.46
Upper Limit 0.135 1.45

Long term settlements

It can be concluded from Table 4.8 that the model behaves consistently. When lower stiffness
and strength parameters are applied, the settlement increases. Evidently, the safety factor de-
creases.

Table 4.8: Results of the long term settlement analysis for different parameter sets

Parameter set ux,max [m] ux,min [m] uy,min [m] Safety Factor [-]
Breakwater -15 m CD
Lower limit 0.087 -0.106 -0.504 1.13
Average/ Reference 0.025 -0.021 -0.188 1.41
Upper Limit* - - -0.039 1.45
Breakwater -25 m CD
Lower Limit 0.094 -0.092 -0.822 1.24
Average/ Reference 0.035 -0.035 -0.257 1.46
Upper Limit 0.033 -0.033 -0.186 1.92

* Calculated with Upper Limit parameter set for the sub soil (sand and bedrock layer), otherwise
the soil collapses. The settlement and safety factor are not comparable with the other results in the
table. This implicates that the subsoil plays a significant role in the possible stiffness and strength
of the breakwater.
From Figure 4.6 it is seen that the settlement occurs at the top of the breakwater and gradually
decreases within the rock fill, which is expected. The midpoint of the settlement is not exactly in
the median of the breakwater. This is due to the fact that the breakwater is not symmetrical in
the y-axis, but the mediam lies slighty right of the middle. The same type of settlements occur
considering the other parameter sets and the -15 meter CD northern breakwater.

38
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Figure 4.6: Settlement of the -25 m CD northern breakwater after 50 years, ’Average/Reference‘
parameter set. Water level +1.1 m CD (Lee and Sea side)

Stability
The stability is divided in two parts, the static and the harmonic part. First the static results are
given and elaborated and thereafter the harmonic results.
Static calculation results
Both the two static test scenarios show the same consistent behaviour as the long term settlement
analysis. The results are summarized in Table 4.9 and 4.10.

Table 4.9: Results of static calculation (Lee side: 3.7 m CD and Sea side: 6.85 m CD [3.7 + 3.15])

Parameter set ux,max [m] ux,min [m] uy,min [m] Safety factor [-]
Breakwater -15 m CD
Lower limit 0.019 -0.128 -0.539 1.11
Average/ Reference 0.011 -0.028 -0.258 1.27
Upper Limit 0.012 -0.030 -0.219 1.37
Breakwater -25 m CD
Lower Limit 0.045 -0.103 -0.804 1.29
Average/ Reference 0.020 -0.038 -0.287 1.46
Upper Limit 0.021 -0.037 -0.223 2.22

39
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Table 4.10: Results of static calculation (Lee side: 0.0 m CD and Sea side: 0.0 m CD)

Parameter set ux,max [m] ux,min [m] uy,min [m] Safety factor [-]
Breakwater -15 m CD
Lower limit 0.081 -0.105 -0.592 1.17
Average/ Reference 0.021 -0.021 -0.275 1.40
Upper Limit 0.022 -0.022 -0.232 1.53
Breakwater -25 m CD
Lower Limit 0.083 -0.082 -0.873 1.30
Average/ Reference 0.030 -0.030 -0.307 1.48
Upper Limit 0.031 -0.031 -0.236 2.34

The difference between the parameter sets are not equally distributed. For example, the difference
between the ‘Lower Limit‘ and ‘Average/Reference‘ parameter sets with respect to uy,min is signif-
icantly bigger than the difference between the ‘Average/Reference’ and ‘Upper Limit’ parameter
sets. It is possible that one of the stiffness and strength parameters has a more dominant influence
on the results. Discussion with an expert of WSP Africa Group revealed that the sensitivity with
respect to the friction angle is particularly present within breakwater modelling.
Also, the settlements are different when different water levels within the breakwater are compared.
If the water level is set to +0 meter CD, the settlements are greater than when the water level is
set on +3.7 meter CD. A reason could be that the ‘extra‘ water in the +3.7 meter CD case induces
an upward (buoyancy) force on the rock fill, reducing the settlements in this scenario. The effect
has a greater effect on the lower stiffness and strength parameter sets, since lower stiffness and
strength material tends to settle more initially.
The height of the breakwater has only limited influence on the total settlement of the breakwater.
Therefore, the ratio ’breakwater height‘ versus ’total settlement‘ is higher for the -15 meter CD
than for the -25 meter CD breakwater. Whether this has to do with the subsoil or with the
rock fill material of the breakwater should be investigated further and is out of the scope of this
investigation.
Figure 4.7 shows the vertical settlement of the -25 meter CD breakwater with ‘Average/Reference‘
parameter set. Also here it showns that the midpoint of the settlement lies not exactly in the
median of the breakwater.

40
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Figure 4.7: Settlement of the -25 m CD northern breakwater, ’Average/Reference‘ parameter set.
Water level + 3.7 m CD (Lee Side) and + 6.85 m CD (Sea Side)

Harmonic Calculation Results

The results for the harmonic calculation are presented in Table 4.11. The ux,min , uy,max and uy,min
can be neglected, since the values approach zero. The ‘Lower Limit’ is calculated with 1 wave
period (1T) and with 10 wave periods (10T). The difference is roughly a factor 1.5 in horizontal
displacements. The ‘Average/Refernce’ and the ‘Upper Limit’ parameter sets give approximately
the same results for ux,max .

Table 4.11: Results of harmonic calculation (Lee side: 3.7 m CD and Sea side: 3.7 + 3.15 m CD
amplitude)

Parameter set ux,max [m] Safety factor [-]

Breakwater -25 m CD
Lower Limit (1T) 0.098 1.28
Lower Limit (10T) 0.167 1.20
Average/Reference (1T) 0.018 1.46
Average/Reference (10T) 0.018 1.68
Upper Limit (1T) 0.015 1.96
Upper Limit (10T) 0.015 1.98

41
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Figure 4.8: Horizontal displacement at Xbloc layer -25 m CD northern breakwater, ’Lower Limit‘
parameter set. Water level + 3.7 m CD (Lee Side) and + 3.7 m CD with 3.15 m amplitude (Sea
Side)

The horizontal displacements occur at the Xblocs, see Figure 4.8. In the ‘Lower Limit’ parameter
set the Xbloc have a lower stiffness and strength, which results in higher displacements. Again,
as also seen in the results of the settlements, the differences between ‘Lower Limit’ and ‘Av-
erage/Refence’ are significantly larger than between the ‘Average/Reference‘ and ‘Upper Limit‘
parameter sets.
It must be noted that the modelling period of the harmonic wave is very small. In reality storm
can have a duration between 6 and 12 hours. In this analysis it was not possible to model a full
strom of 6 to 12 hours, unfortunately. The calculation time for modelling 1 minute was already
more than 30 minutes.
The neglectable differences in ux,max for the ‘Average/Reference’ and ‘Upper Limit’ parameter sets
need further investigation. A very general assumption is made considering the stiffness and strength
parameter of the Xblocs. Therefore, it is likely that the results are inaccurate/incorrect.

42
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

4.4 Discussion
The model is consistent, for both the -15 m CD cross section and the -25 m CD cross section of
the northern breakwater:
• Higher stiffness and strength parameters result in lower settlements and higher factors of
safety;
• For higher water levels within the breakwater the settlements are lower than with lower water
levels in the breakwater. This is due to the buoyancy forces that will uplift the rock fill layer;
• Higher stiffness and strength parameters lead to lower horizontal displacements and higher
factors of safety when harmonic wave functions are applied onto the breakwater;
• If the settlements from Table 4.8 are compared with Tables 4.9 and 4.10 it can be seen that
for the settlements the results are approximately the same. This indicates that higher water
levels at the ’Sea Side’ do not have significant influence on the vertical settlements of the
breakwater.
The differences in displacements between the different parameter sets are not equally distributed,
the differences between ‘Low Limit‘ and ‘Average/Reference‘ results and ‘Average/Reference‘ and
‘Upper Limit‘ results are not equal. This could indicate that the influence of the individual stiffness
and strength parameters on the results is not the same. As mentioned earlier, the friction angle is
likely to have a great influence on the results.
Important notes have to be made concerning the modelling of the breakwater in Plaxis:
• Plaxis is highly dependent on correct soil/rock material parameters. During parameter de-
termination it is seen that the model consists of many different material sets and therefore
different parameters. A lot of these parameters are assumptions in this model, which have a
negative effect on the accuracy of the results;
• ‘Simple’ material models are used, due to the fact that the availability of parameters is
limited. More advanced models would lead to more accurate results. This was not possible
in this analysis;
• The outcome is highly dependent on the expertise of the person who makes the model. Since,
the experience is limited, it can be expected that some mistakes are made that will lead to
inaccurate results;
• Modelling breakwaters is a difficult task and there is no perfect system. The model is very
sensitive to especially the friction angle.

43
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Recommendations for further research

For more accurate modelling the breakwaters of the DDOP, the following topics should be ad-
dressed.
Parameters subsoil
Accurate soil parameters are of vital importance for accurate model results. Therefore, the stiffness
and strength parameters of the bedrock and overlying sand layer should be known. This can be
accomplished by an extensive geotechnical investigation (offshore) by the use of boreholes and
Cone Penetration Tests (CPT). Also samples of the siltstone should be taken and tested in the
laboratory. The more parameters are known, the more advanced models can be used to model
the bedrock and overlying sand layer. When the sand layer is not of sufficient quality to function
as foundation of the breakwater, this layer needs to be dredged, and material of a higher quality
should be laid on the bottom. This is extremely expensive at the depths of this breakwater.
Parameters rock fill
As mentioned earlier, the determination of the parameters of the rock fill is very difficult and little
is found about the parameters in the literature. But, literature states that extensive oedometer
and shear tests should be applied to the rock fill material to achieve the needed accuracy to model
this problem [CUR & CETMEF, 2007]. Also expert judgment and reference project can be very
useful to determine accurate parameters for the rock fill. The same goes for the Xblocs.
Xblocs
In this model the Xblocs are modelled as soil volumes. Research should be done whether it is
possible to model Xblocs as a ‘Structure’ in Plaxis. This could have a positive effect on the
results.
Groundwater flow
In this analysis no attention has been paid to the groundwater flow other than the permeability of
the different rocks within the rock fill. In this investigation there was too little time to take this
into account.

44
Chapter 5

Concluding on Breakwater Design

In the proposed layout two different configurations for the breakwaters are distinguished in terms of
orientation and length. The proposed longer breakwaters are based on a relatively high maximum
vessel entrance speed (10 kn) and the proposed shorter breakwaters on a lower maximum allowed
entrance speed (6 kn). In this chapter the research is concluded.
From the research with the simplified wave model it is concluded that the shorter breakwaters
cause more wave action in the port. The conditions are such that the liquid bulk terminal cannot
be operational during storm conditions from the south. Storm conditions from the north are on
the contrary better mitigated by the shorter breakwaters because less waves are reflected by the
southern breakwater. For both configurations this however does not result in downtime. During
normal conditions it is found that the liquid bulk terminal can be operational for more than 99%
of the time with the shorter configuration of the breakwaters.
In the design of the breakwater cross section the necessary dimensions are determined. There is a
distinction made between the cross section at the largest depth, for the longer configuration, and
at a lower depth for the shorter configuration. Comparing these two gives insight in the increase of
dimensions at larger depth and therefore the increase in costs of the breakwaters. In the research
it has become obvious that the breakwater at larger depth is significantly larger. In combination
with the result of the wave model it is therefore concluded that the shorter configuration of the
breakwaters is preferred. This leads to significantly lower construction costs and only leads to
downtime of the liquid bulk terminal during storm conditions.
By the use of numerical analysis the stability of the cross section of the northern breakwater is
calculated for several test scenarios. The different input parameter sets have given consistent
results. The stiffer and stronger the material, the smaller the settlement. Evidently, with stiffer
and stronger materials, the safety factor increases. For the stability analysis the same conclusion
is given. For both the cross sections (-15 m CD and -25 m CD) the models are consistent. The
settlement to construction height ratio is higher at the -15 m CD breakwater, thus the relative
settlements are bigger compared to the -25 m CD breakwater. The ‘Lower Limit’ parameter set
gives a settlement in the order of 0.8 m, the ‘Upper Limit’ parameter set gives a settlement in
the order of 0.2 m. It can be concluded that the input parameters of the model have a significant
influence on the results.

45
Chapter 6

Sand Bypass System

Durban’s coastal stretch is interrupted by the breakwaters of the DDOP. The current net alongshore
sediment transport is approximately 500,000 m3 /year northwards [WSP, 2008] and sediment will
start accumulating at the south side of the southern breakwater. Next to that, sediment will also
accumulate at the north side of the northern breakwater. The wave climate in combination with
secondary current patterns due to the sheltering zone of the breakwater also causes southward
directed sediment transport.
The alongshore sediment transport at the Durban coast is however mainly induced by the obliquely
incident waves from the south. Downstream of the breakwaters (north side), where the waves reach
the coast again, the alongshore sediment will start transporting sediment from the existing beaches.
Erosion of the bluff, which is a steep and high sand dune, has to be controlled and prevented. The
upstream accumulating sediment is used to prevent erosion of the bluff by means of a sand bypass
system.
First the effect of accretion and erosion in the situation with breakwaters is quantitatively estimated
in this chapter. After that, various sand bypass options are discussed. The beach nourishment
and sand trap are discussed in the following subsections. Lastly an overview of the preferred sand
bypass system is illustrated.
The reader has to keep in mind that the sand bypass design is highly conceptual. Further research
has to be done before the proposed system can be implemented. However, this concept might
provide the reader more insight in the order of magnitude of the coastal processes and the measures
which could be taken in order to prevent these coastal processes from happening.

47
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

6.1 Situation without Sand Bypass System

The majority of the alongshore sediment transport occurs in between a cross shore distance (mea-
sured from the shoreline) of 50 m and 400 m, representing 98% of the total sediment transport
[TNPA, 06-2006]. The length of the breakwaters is longer than the zone in which alongshore sed-
iment transport occurs, which implies that most of the sediment transport will be blocked, under
the assumption that the sediment is not transported cross shore (via rip-currents).
As a consequence of the sediment blocking, accretion of sediment occurs at the southern side of the
breakwaters and coastline erosion occurs to the north of the breakwaters, which might have fatal
consequences for the bluff. The amount of erosion and accretion is quantitatively approximated
with an analytical single line model.
Approach single line model
Pelnard-Considere (1956) derived an analytical solution for accretion of sediment near breakwaters
[J. Bosboom and M.J.F. Stive, 2013]. The approach is based on the principle of volume balance,
which in other words implies that the amount of sediment that comes into a control volume
is equal to the amount that comes out added with local accretion or erosion. Therefore, the
erosion at the northern side of the breakwater has a similar profile as the accretion profile at the
southern side, where second order circulation patterns and differences in wave height are neglected.
Furthermore, the amount of sand transported is dependent on the angle of wave incidence relative
to the coastline.
Model results and conclusions
With the single line theory first the cross shore length of sand accumulation at the breakwater
L(t) is calculated for a yearly nett transport rate of 500,000 m3 :
r
0 4at
L(t) = φ (6.1)
π

In which φ0 represent the angle of wave incidence relative to the coastline (36 degrees, equal to
0.63 radians). Furthermore, t represents the time in years, which is taken to be equal to one.
Lastly, a is equal to S/φ0 d, where S is the amount of nett sediment transport equal to 500,000
and d is the profile depth taken to be equal to 8 m. This results in an accretion length of 224
m. As discussed before, the erosion profile is more or less similar to the accretion profile which
implies a yearly setback of the coastline of about 220 m at the northern part of the breakwaters.
This is a huge setback and therefore measures need to be taken in order to prevent the bluff from
eroding.
The affected width of accretion and erosion along the shore can be estimated with the following
relation:
√ L
B = 2.5 π (6.2)
φ

This results in a width of 1.6 km. It can be concluded that approximately 1.6 km north and 1.6
km south of the port the coast will adapt to the existence of breakwaters, see the figure at the end
of this chapter.

48
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

Model limitations and assumptions

The single line theory is a simple analytical solution and therefore simplifies and neglects a lot of
processes. More advanced models (Delft3D) could be used to determine the accretion and erosion
patterns with better accuracy. However, due to the complexity of these models, the amount of
time to set up and to run the models is longer than the amount of time available in the project.
The processes not taken into account are summed up below.
• The single line theory assumes that the shape of the cross shore profile does not change in
time, which implies an equilibrium profile at arbitrary depth
• Tidal influences, variations in wave height, direction and sediment characteristics are not
taken into account in this model
• Non-eroding obstacles such as the liquid bulk terminal are left out of the model.
• The shoreline evolution is calculated under the assumption of steady wave conditions (only
one wave condition)
• Outbreaking alongshore currents together with its sediment transport are not taken into
account in the model
• The cross shore exchange of sediment is not taken into account in the model
As discussed before, the accretion profile is assumed to be more or less the same as the erosion
profile if second order circulation patterns and differences in wave height are neglected. In reality,
the erosion pattern can be different from the accretion pattern. This can be explained by the
following reason: The incoming waves are sheltered at the lee side of the breakwater. A difference
in wave height then occurs at the point where the waves are not sheltered anymore, inducing
circulation patterns towards the breakwater. This effect reduces the setback of the coastline.
Furthermore, the effect of cross shore currents was not taken into account in the model. In reality
strong rip-currents exist near the breakwater, transporting sand offshore and past the breakwater
in the port entrance channel and to the northern side of the port. Therefore the sand transport
will not be blocked completely, but will continue to a certain extend.

49
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

6.2 Sand Bypass Options

From Subsection 6.1 can be concluded that a sand bypass system is needed to prevent the coast from
eroding at the northern side of the port entrance. In this subsection various sand bypass systems
are discussed briefly, together with the advantages and disadvantages. Lastly, a recommendation
is given based on the advantages and disadvantages of the systems.
Mobile maintenance
Mobile maintenance can be established with a trailer suction hopper dredger. This vessel dredges
the sand from the sand trap, brings it into port where it can be filtered and can be dumped at
the lee side of the northern breakwater. Taking into account the waves, the dredger cannot come
close to the port. For that reason, a pipeline is deployed to pump the sand onto the beach.
The advantages of mobile maintenance are the flexibility, manoeuvrability and repair costs. The
dredger is flexible in usage because it can be deployed in various places in the port. Furthermore
the vessel is manoeuvrable; it can sail to the furthest corners of the port to dredge up or to dump
the sand. Lastly, repair of the vessel can be done onshore, which is efficient, reliable and less costly
than in the water.
The disadvantage is the discontinuous dredging activity. The vessel only dredges a large bulk of
material limited times per year.
Continuous maintenance
Continuous maintenance can be done by means of a fixed jet pump system deployed on a jetty. The
pumps are place below the sea bottom and create a pit with its jets where sand can be trapped.
The trapped sand is transported via pipelines to the spot where beach is nourished.
An advantage is the continuous beach nourishment. However, also disadvantages exist: This
system comes with high maintenance costs. The dug up soil material is not only sand, it also
contains tyres, cables etc. which are often stuck in the pipelines. Next to that, the sandpits
created by the pumps will not collect all longshore transport due to the fact that the pumps are
not working 24 hours per day, 365 days per year and the fact that the jetty and trap might not
cover the whole surf zone. Therefore, often an extra dredging vessel is needed to maintain the
depth of the entrance channel, which obviously comes with extra costs.
Recommendation
A continuous system needs a fixed pump system and also a dredging vessel to dredge the remainders
of the sand in the entrance channel. Furthermore, maintenance of the pipelines is costly and has
to be executed with care.

50
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

6.3 Beach Nourishment

The beach nourishment is based on the mobile sand bypass option. As calculated in Section 6.1
the coastline setback will be approximately 200 m per year. This setback is fatal for the bluff and
therefore the beach needs to be nourished. Two options of beach nourishment are proposed: two
month nourishment and one year nourishment. The client can choose one of the two options based
on the available crew, dredging vessels, aesthetics and planning schedule. For both options more
research needs to be done to come to the most effective location of nourishment and the shape of
the sand heap.
Two month nourishment
Calculating with the same model as in Section 6.1 the setback of the coast for two months is about
90 m. The volume of sand needed to nourish this setback is approximately 80,000 m3 . Since the
modern dredging vessels can carry a volume of about 30,000 m3 , the vessel needs to sail three
times back and forth.
It is recommended to dump the amount of sand just after the construction of the breakwater, in
order to make sure that the existing land will not be eroded. The sand can be either be extracted
from the constructed sand trap or the beach at the entrance channel, depending on the sand
quality.
One year nourishment
The beach setback per year is 200 m with an amount of eroded sand equal to 500,000 m3 , indicating
17 times sailing back and forth. This amount of sand can be dumped at the northern side of
the breakwater, forming a large heap of sand similar to the Sand Engine at the coast of The
Netherlands.

6.4 Sand Trap

The sand trap has to be located at a strategic position where most sand is collected and where
dredgers can do their work effectively. It is chosen to locate the sand trap close to the breakwater
at a certain distance from the coast, because most of the sand will go past this area via a combi-
nation of alongshore sediment transport and rip-currents. Furthermore dredging vessels have good
nautical working environment, in which enough distance is kept from both the shore as well as
breakwater keeping in mind the wave climate.
Depending on the type of nourishment, the sand trap needs to have at least a capacity of 80,000
m3 or 500,000 m3 . In both cases it is recommended to make the capacities larger, keeping in mind
the gross sediment transport and the possible change in dredging schedule.

51
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

6.5 Overview Sand Bypass System

The recommended mobile sand bypass system is illustrated in the figure presented below. The
natural accretion and erosion pattern are given according to the single line theory. The erosion
pattern is affected by wave sheltering and has a slightly different shape than the accretion pattern.
The process involving the sand bypass is as follows: the vessel dredges the sand, sails into port
and dumps the sand at the sediment offloading area. Subsequently the sediment is filtered and
lastly the sand is transported via pipelines to the spot where the beach needs to be nourished and
is dumped at the right place.

52
 EROSION ZONE

INLAND BREAKWATER EXTENSION

SEDIMENT OFFLOADING
AREA

ANGLE OF
WAVE INCIDENCE

ACCRETION
SAND TRAP
ZONE
Chapter 7

Care for the Mangroves

A major advantage of the proposed layout for the Durban Dig-Out Port is that the area available for
the Isipingo estuary and its mangrove trees is maintained. Even so, the area can be expanded in this
design. However only maintaining the area will not cure its current bad condition. Environmental
study to the estuary states that “at this point the natural functioning of the system has effectively
ceased” [SSI and MER, 2011]. With the DDOP construction new opportunities arise to revitalize
the estuary. A solution for the estuary will have to tackle problems involving the fresh water
supply, water pollution, the estuary mouth and managing flood risk.
Also in this chapter the reader has to keep in mind that the proposed solutions for the mangrove
system are highly conceptual. Further research has to be done before the proposed system can
be implemented. However, this concept might provide the reader more insight into the processes
involved in estuary upkeep. All statements in this chapter are based on the report “Isipingo
Estuary Management Plan: Situation Assessment” by SSI Engineers [SSI and MER, 2011].

7.1 Proposed Measures

Fresh Water Supply
An ideal situation for the mangrove estuary at Isipingo would be the natural exchange between
fresh water from the river and salt water from the ocean. The flora and fauna would be healthy,
but occasional flooding of the Isipingo River would occur. To reduce flood risk it was decided to
divert the natural course of the Isipingo River to the new Umlazi storm channel when the airport
was constructed, reducing the fresh water supply to the estuary drastically.
The result of the cut-off in fresh water supply is that the estuary lost its ability to ‘breathe’ and
the sediment processes on the sea side closed the entrance to the estuary. Artificially keeping the
mouth of the estuary open with a pipeline has failed and a sustainable solution will include getting
more fresh water into the estuary.
Looking at the course of nearby rivers it shows that the Isipingo River also has a branch running
into the Mbokodweni River. Without taking into account river dynamics, an option is to close off
the Isipingo run-off into the Mbokodweni River, increasing the fresh water supply into the Isipingo
estuary. To be able to steer the estuary into a natural balance, a discharge regulating structure

55
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

could be applied instead of closing the branch off completely.

Water Pollution
In addition to the reduction of fresh water supply human activity in the area has increased the
levels of pollution in the water. The problems with water pollution are that people are unaware
of the damage they do to the environment by using open water as a dump site. And also that the
consequences for people and businesses polluting the water are negligible. A sustainable solution
for the Isipingo estuary will include increasing awareness of the people and businesses around the
river about how they can limit their polluting behaviour. On the other side, the government can
put greater effort in discouraging pollution of the environment, for example by putting hefty fines
on offenders.
The Estuary Mouth
In an effort to save the estuary from dying a pipeline from the sea into the area was constructed to
let seawater flow in and out. The interaction between salt and fresh water is vital for an estuary’s
survival, but the health of the estuary kept on declining, even with the pipeline in place. Oxygen
levels in the water have declined and minimal flushing of the area gave way to extensive fine
sediment settling.
The cut-off in fresh water supply disturbed the natural equilibrium and caused the intense sediment
processes in the ocean to close the mouth of the estuary. Simply dredging away the sand on a fixed
interval would not fix the problem, in fact the ocean would just close the entrance again.
A sustainable solution for the estuary issue is one where human involvement and maintenance are
as small as possible and where nature runs its course. With the construction of the breakwaters for
the proposed DDOP that might be possible. The southern breakwater will prevent the longshore
sediment transport to reach the estuary entrance and will greatly reduce wave action at the estuary
mouth.
The environmental study to the Isipingo estuary concludes that an open mouth estuary, thus
without pipelines, is ideal for it to function. A thorough look into sedimentation processes at the
new sheltered situation of the estuary mouth is needed, but it is imaginable that with an increase
in fresh water supply and limited sea action a new equilibrium for the estuary mouth can be
reached.
Flood Risk
With a large number of people living and working in the Isipingo area the risk of flooding has to
be acceptably low. In recent times heavy rainfall has led to flooding of a large part of the estuary.
The situation with the DDOP in place will have a positive impact on flood risk on a number of
ways. At the moment the estuary is best described as a bowl that can be filled with water from the
north because the berm of the area is 4.0 meters high and especially the pipeline is a bottleneck
in releasing flood water because of its limited discharge capacity.
The assessment on the state of the estuary concluded that the consequences of a flood event can
be reduced greatly when the berm is lowered or, even better, the mouth of the estuary becomes
open and the pipeline is removed. In addition the construction of the entrance channel next to the
estuary raises the opportunity to construct an extra outlet that can be used to discharge water in
case of flooding.

56
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

The last flooding of the Isipingo area occurred in March 2008, but according to the same assessment
the consequences of similar weather conditions would be almost negligible if an open-mouth system
and an extra culvert in the entrance channel were used. Therefore the entrance channel bank will
be constructed so that it protects the estuary and that it is able to discharge excess water in case
of heavy rainfall.

7.2 Concluding
Constructing the DDOP as proposed will give possibilities to bring the Isipingo mangrove estuary
back to life. Reaching a new natural equilibrium for the mouth of the estuary will make the
trees and animals in the area flourish again and it will also prevent the area from severe floods.
The new entrance channel will also help prevent flooding by constructing culverts. The proposed
solutions for the estuary use passive systems like the open mouth and the culverts to keep human
involvement to a minimum and make it a truly sustainable solution. The proposed measures are
summed up in Figure 7.1.

Figure 7.1: Summary of measures

57
Bibliography

[AECOM, 2012] AECOM SA (Pty) Ltd., (2012). Non-public information. Document title: Hydro-
dynamic data extract from AECOM report FEL 1 report.
[R. Brinkgreve et. al, 2014] Brinkgreve R.B.J., Engin, E., Swolfs, W.M., (2014).
2DAnniversaryEdition-0-Gen-Info.pdf (SECURED), part of PLAXIS 2D AE Manual.
[CUR & CETMEF, 2007] CIRIA, CUR, CETMEF, (2007). The Rock Manual; The use of rock
in hydraulic engineering (2nd edition). Publisher: C683, CIRIA, London, ISBN 978-0-86017-
683-1
[Deltares, 2014] Deltares (2014). Delft3D-Wave, User Manual, Version:3.05.34160.
[Delta Marine Consultants, 2011] Delta Marine Consultants (DMC) (2011). Guidelines for Xbloc
Concept Designs.
[H. Ligteringen and H. Velsink, 2012] Ligteringen, H. and Velsink, H., (2012). Ports and Termi-
nals. Published by VSSD. ISBN: 978-90-6562-288-4.
[J. Bosboom and M.J.F. Stive, 2013] Bosboom, J., Stive, M.J.F. (2013). Coastal Dynamics 1.
Published by VSSD. ISBN: 978-90-6562-286-0.
[A. Mather and K. Reddy, 2008] Mather, A.A. and Reddy, K., (2008). Expansion plans for the
port of Durban: What are the issues for the city of Durban? COPEDEC VII, Paper No:
M-05.
[Plaxis2D Reference Manual, 2014] Plaxis B.V., (2014). 2DAnniversaryEdition-2-Reference.pdf
(SECURED), part of the PLAXIS 2D AE Manual.
[Project Durban, 2014] Gijsman, R., de Haan, F.S., de Koning, R.J., Le, T.N, Steeneken, S.P.,
(2014). Analysis report Durban Dig-Out Port. TU Delft.
[SSI and MER, 2011] SSI Engineers and Environmental Consultants (SSI) and Marine and Estu-
arine Research (MER), 2011. Isipingo Estuary Management Plan, Part II.
[The SWAN team, 2006] The SWAN Team (2006). SWAN User Manual, SWAN Cycle III version
40.51. Published by Delft University of Technology.
[TNPA, 06-2006] Transnet National Port Authority, (June 2006). Non-public information. Docu-
ment title: 385-J-105 - Sand Bypass Sediment Dynamics Rev00 June 2006.
[TNPA, 02-2014] Transnet National Ports Authority, (Feb 2014). Non-public PowerPoint presen-
tation, Document title: Port of Durban Presentation Feb 2014.ppt.

59
 TU Delft - Durban Dig-Out Port Research - Part 2: Port Engineering

[TNPA, 05-2014] Transnet National Ports Authority, (May 2014). Non-public PowerPoint presen-
tation, Document title: Port of Durban - PDFP May 2014.ppt.
[H.J. Verhagen, K. d’Angremond, F. van Roode, 2009] Verhagen, H.J., d’Angremond, K., van
Roode, F. (2009). Breakwaters and Closure Dams. Published by VSSD. ISBN: 978-90-6562-
173-3.
[A. Waltham, 2009] Waltham, A.C., (2009). Foundation of Engineering Geology, third edition.
Published by: Taylor and Francis, ISBN 10 0-415-46959-7.
[WSP, 2008] WSP Group Africa (Pty) Ltd., (2008). Non-public information. Document title: Re-
view of Sand Bypass Systems for Port of Durban 01.

60