Master Thesis 0304

ISSN 1651-7660
Skrift 2003-7

APARTMENT BARGES –
A COMFORT AND SAFETY ANALYSIS

M Sc Thesis
by

Sebastian Brunes

Aeronautical and Vehicle Engineering

Stockholm Division of Naval Systems
2003 ROYAL INSTITUTE OF TECHNOLOGY
NAVAL ARCHITECTURE April 2003
DEPARTMENT OF VEHICLE ENGINEERING

APARTMENT BARGES –
A COMFORT AND SAFETY ANALYSIS

M Sc Thesis

by

Sebastian Brunes

Performed at Saltech Consultants AB

Kungliga Tekniska Högskolan

Aeronautical and Vehicle Engineering
Division of Naval Systems
Royal Institute of Technology (KTH)
SE-100 44 Stockholm
Sweden

Master Thesis 0304

ISSN 1651-7660
Skrift 2003-7
Acknowledgement
I would like to thank Thomas Milchert and the other consultants at Saltech Consultants AB
for all their help and interest in my Master Thesis and for making this report possible. It is
always nice to find companies with a true interest in ship research and maritime projects.
I would also like to thank all the people who have helped me with information, data and feed-
back. Among these people, I thank my girlfriend Paula for all her help with the report as well
as for being a patient “discussion companion”.

1
Summary
This report presents the results of a study of the comfort and safety concerning apartment
barges.
There are mainly two types of houses on barges: the one family house and the apartment
block. Most of the people who choose to live in a one family house barge like the sea and
have indulgence with the possible discomfort. However, in a floating apartment block a new
group of people will be moving in, with less tolerance of the sea. The most common
discomforts are due to vertical and lateral accelerations and high roll angles. Other aspects of
comfort are good design of gangways, well designed heating, ventilation etc.
Comfort criteria and theoretical background
For passenger ships there are comfort criteria regarding vertical accelerations, lateral
accelerations and roll angles that have to be met. In this report, the criteria used when
designing cruise liners are applied.
All floating objects will move due to the effects of wind, waves and changes in water levels.
This report is concentrating on the movements induced by wind-generated waves, since
surges never last long enough to make a large barge roll in resonance.
Potential locations for apartment barges are all in the archipelago or in protected water areas,
where the wind is reduced. The wind and waves in protected areas are described, in order to
give a theoretical background and an explanation to how a barge will move under different
conditions.
A working procedure
A working procedure for planning apartment barges is proposed. The planning should be
carried out through an iterative process, until a functional design is achieved. The different
steps of the planning process are illustrated in the figure below.

Wind and Barge

wave data design

Damping measures
• Bilge keels • Adjust dimensions
• Mooring damping • Tanks
• Break waters • Etc.

Strip calculations

Evaluation of comfort

Are the
movements and Choice of precautions to improve the
accelerations
No behaviour of the barge
below the comfort
criterias?

Yes

Continue
projecting

2
The location of interest is studied regarding fetch length, wind data and water depth. An
initial design is chosen, based on the basic demands of the customer. The dimensions,
draught, distance from keel to center of gravity and radius of gyrations are calculated, along
with other barge characteristics. These will, together with wind and wave data, be needed in
the strip calculations performed in order to predict the barge’s behaviour.
Strip theory is the common way of solving 3-dimensional problems using a 2-dimensional
technique. The strip calculations are used to calculate movements and accelerations for
different sea states.
With the results from the strip calculations, it is possible to evaluate the expected comfort. If
the barge’s movements exceed the chosen criteria, appropriate precautions could be chosen
from the ones listed below.
• Bilge keels
• Anti-roll tanks
• Mooring damping Seaflex
• Pontoons as mobile breakwaters
• Changing the beam
• Changing the centre of gravity KG
Risk analysis
A rough risk analysis for apartment barges is performed. The risk is defined as the probability
that an accident occurs, multiplied by its consequences. If a certain risk is high and it is
impossible to eliminate or reduce the probabilities of occurrence, precautions to minimise the
consequences should be taken. The risks identified for apartment barges are:
1. Fall accidents caused by slippery footpaths
2. Health problems due to high moisture and wind
3. Taking in water by inlets, outlets or scuttles
4. Corrosion
5. Mould/water damage
6. Collision
7. Broken mooring arrangements
8. Damages to the hull from ice-pressure
9. Sewer leakage
10. Fire on board
Fall accidents, Taking in water and Fire on board are identified as the highest risks.
However, these risks can be reduced to acceptable levels certain proposed precautions.
Example projects
The iterative planning process described briefly above is used on the two example projects:
Bällstaviken and Pampas Marina.
Bällstaviken
The behavior of three barges with different dimensions was studied. Strip calculations were
made for bare hull as well as for barges equipped with anti-roll tanks and bilge keels.
Calculations were done for two different wave inclination angles. The best result was
achieved for the narrowest barge, equipped with bilge keels.

3
Pampas
For Pampas Marina strip calculations were made for the Generation III barge, equipped with
and without different damping measures. For this barge an anti-roll tank improved the
behaviour significantly.
Conclusions
The conclusions drawn in ship literature and according to common beliefs, are that the wider
the barge the smaller the movements. This assumption is however contradicted by the results
of the strip calculations made in this study. The low stability of the narrower barges gives
high roll angles, as expected, but the accelerations are low. By equipping the barge with for
example bilge keels, the roll is easily reduced.
The damping measures studied are all working in different ways with their own advantages
and disadvantages.
Damping
device Advantages Disadvantages
Bilge keels A simple and efficient way to reduce the roll Sometimes gives an increase in
angles. accelerations.
Anti-roll tank Reduces the accelerations efficiently, if the Reduces the stability and takes
barge is not too stable. valuable space inside the barge.
Breakwaters Reduce both accelerations and roll angles, Expensive and not always possible.
since they reduce the waves.
Seaflex Reduces the roll angles and keeps the A probably increase in accelerations
barge in position. and very careful installation is
needed.

In protected areas, barges wide enough not to produce any responses at all must be possible to
build. However, full-scale tests are probably needed in order to estimate the beam required.
If measures are taken to reduce the potential high risks involved for apartment barges, living
like this is assumed to be safe.

4
Table of contents

1 INTRODUCTION............................................................................................................ 7
1.1 OUTLINE...................................................................................................................... 7
2 COMFORT ....................................................................................................................... 8
2.1 COMFORT CRITERIA AND MOTION SICKNESS ................................................................ 8
2.2 OTHER COMFORT DEMANDS ...................................................................................... 10
3 WIND AND WAVES ..................................................................................................... 11
3.1 WAVE GENERATION .................................................................................................. 11
3.2 WIND WAVES IN PROTECTED BAYS ............................................................................ 12
3.3 ENERGY..................................................................................................................... 17
4 THE MOVEMENT OF BARGES ................................................................................ 18
4.1 ROLLING ................................................................................................................... 18
4.2 STABILITY ................................................................................................................. 20
4.3 NATURAL FREQUENCY .............................................................................................. 22
4.4 STRIP THEORY – A WAY TO PREDICT MOVEMENTS AND ACCELERATIONS .................. 23
5 METHODS OF LIMITING ACCELERATIONS AND ROLL ................................ 27
5.1 BILGE KEELS ............................................................................................................. 27
5.2 ACTIVE FINS .............................................................................................................. 30
5.3 ANTI-ROLL TANKS .................................................................................................... 30
5.4 MOORING DAMPING .................................................................................................. 35
5.5 PONTOONS AS MOBILE BREAKWATERS...................................................................... 37
5.6 THE SAIL-EFFECT FROM THE WALLS .......................................................................... 37
5.7 CHOOSING APPROPRIATE DIMENSIONS OF THE BARGE ............................................... 38
6 RULES AND REGULATIONS .................................................................................... 42
6.1 FIRE SAFETY RULES ................................................................................................... 42
6.2 STABILITY AND LEAK STABILITY............................................................................... 42
7 SAFETY AND SYSTEM ANALYSIS.......................................................................... 43
7.1 RISK ANALYSIS .......................................................................................................... 43
8 PLANNING APARTMENT BARGES - A WORKING PROCEDURE .................. 52
8.1 STUDIES OF THE LOCATION........................................................................................ 52
8.2 BARGE DESIGN .......................................................................................................... 53
8.3 STRIP CALCULATIONS ................................................................................................ 53
8.4 EVALUATION OF COMFORT FROM WIND DATA ........................................................... 53
8.5 DAMPING MEASURES ................................................................................................ 54
8.6 COMMENTS TO THE PLANNING .................................................................................. 55
9 EXAMPLE PROJECTS ................................................................................................ 56
9.1 BÄLLSTAVIKEN ......................................................................................................... 56
9.2 PAMPAS MARINA ...................................................................................................... 71
9.3 MARIEBERGSVIKEN IN KARLSTAD ............................................................................ 80
10 CONCLUSIONS......................................................................................................... 81

5
11 REFERENCES ........................................................................................................... 82

APPENDIX A ROLL RESPONSES FOR DIFFERENT B/L RATIOS

APPENDIX B FETCH CALCULATIONS FOR THREE EXAMPLES

6
1 Introduction
People have for thousand’s of years been living on various kinds of ships: sailing yachts,
small cargo ships and tugs. All people living like this liked the sea and had no problems
accepting the minor disadvantages that came along with the fact that their “house” was
moving. All floating objects will move, due to the effects of wind, waves and changes in
water levels
There are mainly two types of houses on barges: the one family house and the apartment
block. Most of the people who choose to live in a one family house barge like the sea and
have indulgence with the possible discomfort. However, in a floating apartment block a new
group of people will be moving in, with less tolerance of the sea. The most common
discomforts are due to accelerations and high roll angles.
With the intention of providing a home like anywhere else, the comfort of the barges must be
good and this might demand improvements in today’s building practices. That is why this
investigation concerning comfort and safety is conducted; to estimate roll angles,
accelerations, overall safety and comfort of a barge in a specified location. Methods to
improve the barge’s behaviour, and thus eliminating or reducing the discomfort, will also be
studied.

1.1 Outline
This report consists of two parts. The first one (chapter 3-9) gives a theoretical background.
The wind and wave generation in protected areas are described, in order to show how to
estimate the expected movements. Methods to reduce the roll angles and accelerations are
accounted for in order to find out how changes in the barge design effect the movements. A
rough risk analysis is also done (chapter 9) in order to identify the highest risks. Possible
precautions reducing the probabilities and consequences are presented.
In the second part of the report, (chapter 10-12) a working procedure is proposed and applied
on two example projects in different locations and with different characteristics.

Used notations

B = Breadth [m] Hs = Significant Wave height [m]

L = Length [m] Ts = Significant Wave period [s]
T = Draught [m] T = Wind duration [s]
U = Speed [m/s] Fe = Effective fetch length [m]
2
I44 = Mass moment of inertia [m kg] g = Gravity [m/s2]
A44 = Added mass of water [m2kg] ςa = The wave amplitude [m]

∆ = Displacement [kg] ω = Angular velocity [rad/s]

ωN = Natural angular velocity [rad/s]

7
2 Comfort
Making the barge a comfortable place to live must be the number one priority for the design
of apartment blocks on barges, besides the usual economic considerations. Good comfort of a
barge consists of low accelerations and roll angles as well as good design of gangways, well
designed heating, ventilation etc.
The advantages of living at the sea are many; the view and the accessibility to water are some.
However, when choosing to live on a barge instead of on land, some potential discomfort, like
the risk of being seasick due to the movements, must be accepted. In this chapter, some
comfort criteria that are used to minimise this risk will be presented.

2.1 Comfort criteria and motion sickness

When constructing passenger ships, it is common to calculate the vertical and the lateral
accelerations as well as the roll angles in order to predict how many percent of the passengers
that will suffer from seasickness. Based upon these predictions, criteria can be set, that have
to be met by the floating construction under certain conditions. The most common criteria
used are so called RMS values1 for Roll, Vertical acceleration and Lateral acceleration.
The vertical and the lateral accelerations are products of different kinds of motions, e.g.
heave, pitch and roll. The accelerations can be summarized according to the following
definition:

∑a
2
a RMS = i i = heave, pitch, roll…
i

The RMS values of accelerations are calculated in the centre of gravity or in any chosen point
on the barge. As an example, the vertical acceleration and movement will be calculated for a
point P of interest; with x=4 m and y=5 m. The barge is moving in roll and in pitch. Then the
Vertical velocity equals the Roll velocity times 5 plus the Pitch velocity times 4 according to
the formula above. The vertical acceleration is the time derivative of the vertical velocity.
z

Fig 2.1 The co-ordinate system for a barge

1
The RMS value, or the standard deviation value, is a statistic term describing the expected distribution and not
under any circumstances the highest expected value. It is known from statistic analysis that the highest expected
value depends on the number of cycles. For a long time distribution four times the RMS can be expected, and for
extremely many cycles five times the RMS. In this case 3-4 times the RMS value seems to be an appropriate
assumption. The average value = 1.25 * RMS and other characteristics of the movements can be derived from
the RMS value. [8]

8
The results of such calculations allow for proper precautions to be taken, in order to reduce
the accelerations and roll to satisfactory levels. There are several potential ways of improving
the behaviour of a ship or a barge. Some of them will be thoroughly described in chapter six.
How are comfort criteria set for apartment barges? Naturally, the tolerance for discomfort will
be lower when considering a permanent home, than for ships, where transport is the main
issue, or the journey is limited in time. Applying the stricter criteria for accelerations and roll
angles used for cruise liners could be a good aim and will be used in the various evaluations
made in this report. For cruise liners these criteria are [5]:
RMS value
Roll 2°
Vertical acceleration 0.02 g or 0.2 m/s2
Lateral acceleration 0.03 g or 0.3 m/s2

Accelerations and roll below these RMS values are unlikely to produce vomiting.
The comfort limits for cruise liners are very strict and can be used for predicting how often it
will be uncomfortable to stay on the barge. When constructing apartment barges it may be
acceptable with higher accelerations during extreme weather conditions, but then there will be
a risk that people will feel uncomfortable.
Individuals have different sensibility to accelerations. Experiments have also shown that
young children and women are more susceptible to seasickness than men. [19]
The motion sickness incidence for a group of people is reduced over time. The diagram in Fig
2.2 shows an example of how the occurrence of seasickness decreases over time in a group of
people, for a specific acceleration. The diagram should be interpreted as follows: If 22 percent
of the passengers on the ship were seasick on the first day, that share had decreased to 4
percent after five days. This implicates that the ones mostly affected by the accelerations of a
apartment barge, are probably visitors unaccustomed to the sea or the inhabitants themselves
when returning home after a long absence.

Motion Sickness
Incidence

25%

20%

15%

10%

5%

0%
Day 1 Day 2 Day 3 Day 4 Day 5

Fig 2.2 Motion sickness incidence: effect of acclimatization. [12]

9
When designing a barge for apartment blocks, one must choose appropriate maximum levels
of acceleration, which will assure satisfactory comfort most of the time. Only a few
unpleasant occasions per year can be tolerated. However, during heavy storms it must be
accepted that some people get seasick, but there must be no, or at least only minor, damages
to the barge. A criterion that can be defined is for example that the barge should cope with
waves at least as bad as the waves generated by the highest measured wind speed during the
last fifty years on that specific location.
To achieve a satisfactory level of comfort, choosing a suitable location and appropriate
dimensions of the barge is important. The dimensions chosen must be based on a proper
analysis of the possible sea states. It is essential to be familiar with the wind and wave
characteristics, when deciding on a suitable location for an apartment barge. In the next
chapter wave generation will be described. It will also be shown how to predict the most
important parameters of sea states.

2.2 Other comfort demands

In order to get a high total comfort, there are some important qualities that should be
investigated, which differs from the construction of regular land based housing.
Properly designed heating and ventilation systems. Apartment barges are exposed to
higher wind speed, moisture from water and waves and this must be taken into account
when designing the systems.
Protection of gangways and actions taken to avoid ice formation on gangways
wintertime.

10
3 Wind and Waves
Waves are induced either by the wind or by ships passing by, making speed through the
water. The wind-generated waves have a characteristic frequency and their pattern is repeated,
but for the ship-generated surge only a few waves, spreading away from the ship, are
produced. A surge never lasts long enough to make a barge roll in resonance, and will only be
a potential problem for small one family barges. For that reason, this report will be
concentrating on the movements induced by wind-generated waves.
The wind pressure itself, acting on the apartment barges can also be a problem, giving high
static roll angles, if the wall area is big and/or the stability of the barge is limited.

3.1 Wave generation

In order to predict the waves, the characteristic parameters of the wind must be known. The
three parameters that are decisive for the wave generation are the fetch length, the duration of
wind and of course the wind speed. This seems logical – the wind must have blown for a
certain minimum time, so that one particular wave has had the time to travel all the way from
the opposite shore.
The waves are growing and the period gets longer as the wave travels.

Wind direction

H1 H2

λ1
λ2

Fig 3.1 The waves at different distances from the shore

Wave characteristics
The two most important parameters characterising a wind-generated sea state are:
• Hs - the significant wave height
• Ts - the mean period

Fig 3.2 Definition of water wave characteristics

11
The significant wave height is defined as the mean height of the largest third of all waves,
which also corresponds to the wave height normally estimated by a trained observer. The
wave characteristics are often expressed in terms of wavelength (m), wave period (s), wave
celerity (m/s), wave number, angular velocity (rad/sec) and frequency (Hz). The relations
between these essential wave parameters are presented in the table below. The formulas are
valid for regular waves in deep water.

Table 3.1 A matrix of the relations between wave parameters

λ T C k ω f

Wave g ⋅T 2 2π ⋅ C 2 2π 2π ⋅ g g
λ
length λ 2π g k ω 2 2 ⋅ f 2 ⋅π

2π ⋅ λ 2 ⋅π ⋅ C 2π 2π 1
Wave T
period g T g g ⋅k ω f

λ⋅g g ⋅T g g g
Wave C
celerity 2π 2π C k ω 2 ⋅ f ⋅π

Wave 2π 4π 2 g ω2 4 ⋅π 2 ⋅ f 2

k
number λ gT 2 C2 k g g

2π ⋅ g 2π g 2 ⋅π ⋅ f
Angular ω g ⋅k
velocity λ T C ω
2
Wave 1 g g ⋅k ω
f 2π ⋅ g f
frequency 2⋅ T 2 ⋅ C ⋅π 2 ⋅π 2 ⋅π
λ

For most people, it is natural to characterise a wave by its frequency or wavelength. However,
in marine literature, the angular velocity is the most frequently used parameters. The reason
for this is that it makes the analysis easier, as will be shown later. For example, it is easy to
understand that for wavelengths going towards infinity, the angular velocity goes towards
zero. This implicates that for such a wave a floating object’s response goes towards one,
which means that the object’s response is the same as the wave’s amplitude, since the object
will follow the slow change in water level.
The wave will initiate large movements of a barge if the dominant wave frequency is close to
the natural frequency of the barge, and if the waves are high enough. In general, short waves
give rise to no or very small movements of the barge. This is because waves with a short
wave length have high frequencies (see table above) and these frequencies never coincide
with the natural frequency of a barge.

3.2 Wind waves in protected bays

According to studies that have been made, the waves in protected areas are reduced for two
reasons, a lower wind speed and a shorter Fetch length. The Fetch length is the distance over
water from the shore in the upwind direction to the actual location studied.

12
The wind speed over the sea is decreasing when the wind is approaching land because of the
increase in surface roughness. A usual value of the wind speed in a typical archipelago is
approximately 70 % of what is measured out at sea [2]. This reduction is made if the wind
comes from the direction of open sea. If local wind data is available it is of course better than
approximations of data from another location.
The boundary layer is defined as the layer close to the ground where the speed is reduced by
the surface. Boundary layers are typically around 100-1000 meters thick. [23] The occurrence
of a boundary layer is due to the no slip condition, which means that the speed of an air
particle on the surface is always equal to zero. An increase in roughness of the ground makes
the boundary layer thicker. The change in boundary layer when approaching land is illustrated
in Fig 3.3.

Fig 3.3 The change in Boundary layer when approaching land

Effective fetch length

For a wide water area, the fetch length corresponds to the distance from the upwind shore to
the point of interest. However, if the water area is narrow compared to the length, or if islands
shelter the location in question, the Effective fetch length must be calculated.
In the picture below, the effective fetch length for a wind blowing from WSW is illustrated.
Most important is of course how far it is to the shore in the actual wind direction (illustrated
by the thick line) but the longer distance 18° south of the wind direction will also help
building up the waves. The fetch length will also be reduced because by the shelter from the
island.

13
 6°

6°

6°

18 °

Fig 3.4 Fetch lengths used when calculating the Effective Fetch length

To calculate the effective fetch length (Fe) several distances are added and weighted
according to the formula below. This formula is often used to estimate which fetch length to
use when predicting the waves in a specific location. [2]
n

∑ Length (α i ) ⋅ cos 2 (α i )
Effecive fetch length Fe = i =1
n

∑ cos( α
i =1
i )

Where α i = every measured distance

There are other formulas proposed in the literature, but according to full-scale tests this one is
very simple and gives a result that is accurate enough; only a little conservative. [2]
The effective fetch length is usually rather short in the archipelago because islands reduce the
free water surface. Even if there is a long narrow water surface between the location of
interest and where the wind is coming from, one must consider the distances for the angles +/-
6°, +/-12°… +/- ~ 45° to get the effective length. It may seem strange, that a very long
distance 45° away from the actual wind direction should affect the result significantly.
However, since the factor (cos (45°)) 2 = 0.25 in the formula above, it only contributes to a
small part of the total sum.
In narrow water areas where B/L is small, the wave generation is more limited and a section
within the angles +/- 30 ° is a more accurate assumption for calculating Fe. [2]
It is also common to take account of the effects induced by “high shore” and hence reduce the
effective fetch length. The figure below illustrates to what extent the fetch length can be
reduced as a function of the slope of the opposite shore. Especially for locations characterized
by a short fetch length and high shores, the lee effects near the shore will influence the result.
The maximal expected wave height and the maximal wave length can be reduced.
In the figure 3.5 below, the Fetch reduction as a function of the slope of the shore can be
seen. The wind varies between shoreline a and line c, where full wind velocity is reached. The
same wave energy is obtained for this varying wind between a and c as for full wind velocity

14
between b and c. Thus the Fetch reduction is b-a if full wind velocity is used for wave
calculations. [2]

Fig 3.5 Fetch reduction behind high shores [2]

Wave generation in protected areas

On open sea, the time needed to achieve Full Arisen Sea (FAS)2 is very long, sometimes as
much as days. However, in protected areas, the time needed is much shorter. This means that
the dimensioning factor, together with the wind speed, for waves in protected areas is the
fetch length, not the duration of wind. Depending on the fetch length and wind speed the time
to achieve FAS can be as low as 10-20 minutes.
Out at sea, many different kinds of combinations of wavelengths and mean periods can occur,
but in protected areas the fetch length is short and thus dimensioning the wave characteristics.
Since the fetch length is almost set, FAS will soon be a steady state where a specific wind
speed generates a sea state defined by HS and TS.
Calculating the effective fetch length and knowing the wind speed, Hs and Ts can be calculated
using the following empirical equations. [22] The relationship between the wave height (Hs)
and the wind speed (U) is given by:
g ⋅ Hs   g ⋅ Fe  
0 , 42

= 0,283 ⋅ 0,0125 ⋅   
U2 
2
 U  

The correlation between the wave period (T) and wind speed (U) is given by:

g ⋅ Ts   g ⋅ Fe  
0 , 25

= 1,20 ⋅ tanh 0,077 ⋅  2  

2π ⋅ U   U  

g ⋅T
= 6,5882 ⋅ e EX
U
Where

2
FAS is the steady state, where the wind speed has been constant long enough so that all waves have
experienced the same wind speed during their entire journey from shore to shore. This gives that time is no
longer a parameter decisive for the waves.

15
 0,5
  g ⋅ Fe 
2
 g ⋅ Fe    g⋅F 

EX = 0,0161 ⋅  ln  − 0,3692 ⋅ ln   + 2, 2024  + 0,8798 ⋅ ln 2 e 
  U 2
  U 2
   U 
 

Where Hs = Significant Wave height [m]

Ts = Significant Wave period [sec]
Fe= Effective Fetch length [m]
U= Wind speed 10 m above water level [m/s]
T= Wind duration [sec]
In chapter 11, these formulas have been used to calculate the effective fetch length, the mean
period and the significant wave height for some locations in the Stockholm area; Bällstaviken
and Pampas Marina, and a location in Karlstad in the lake Vänern.
If a barge is moored in shallow water, the waves affecting it will lose energy as a result of
dissipation3 and friction to the seabed. A substantial amount of energy is lost in water depths
of 10-50 meters. [11] Further investigations of the effect of shallow water are needed. More
knowledge regarding both dissipation for waves and roll damping are vital, to find out to what
extent the shallow water effect limits the movement of barges. As the case is now no extra
reductions are assumed because of the shallow water effects.

Wave spectra
Even if the mean period TS and thus the angular velocity (ω) are constant, there will be a
spread in angular velocity within every specific sea state. Wave spectra are used to illustrate
the distribution of energy, as a function of angular velocity (ω), at any given location and
wind state.
The wave spectra in protected bays are narrower in frequencies than in open water areas,
since the variation in distance and time that each wave can travel is limited. Typical wave
spectra for two different water areas are shown in Fig 3.6.

Fig. 3.6 Wave spectrum [1]

3
The Dissipation consists of the bottom Dissipation and the Dissipation in the water column. The dissipation
loss in energy of the wave is because internal friction and eddies. [27]

16
3.3 Energy
The wind transfers energy to the wave if the speed of the wind is higher than the wave speed
and if the wave’s direction coincides at least partially with the wind direction. The wind will
help building up waves not only in the actual wind direction, but approximately +/- 45° from
both sides of that direction [2].
The wave energy consists of potential and kinetic energy. The mean energy (E) per square
meter of water area is given by:
ζa= the wave amplitude
1
E = ⋅ ρ ⋅ g ⋅ς a
2
ρ= the density of water 1000 kg/m3
2
g = the gravity 9.82 m/s2

A linear increase in wave height corresponds to a quadratic increase in energy.

17
4 The movement of barges
Floating objects will move in six degrees of freedom and with six types of motions; Roll,
Pitch, Yaw, Surge, Sway, Sway, and Heave. The convention for translatory and angular
displacement is illustrated in Fig 4.1. [13] Since the roll will be the far most important motion
causing acceleration, and thereby discomfort, for apartment barges, the roll will be described
more in detail below.

Fig 4.1 The six degrees of freedom From [13]

4.1 Rolling
Rolling occurs when the centre of buoyancy is moved transversally away from the centre of
gravity. This happens when a wave contribute with lift to one side of the barge when in the
same moment the lift on the other side is decreased. See Fig 4.2 below. As soon as the barge
has passed the point of equilibrium it slows down, and eventually it starts rolling back. After
the turn, the barge’s roll amplitude slowly decreases in the next roll cycle, if no further
excitation.

Fig 4.2 Roll induced by wave. Long waves give high moments, short waves
give small moments.

Since barges often are wider than normal ships, they typically have a high initial stability,
which will lead to:
High retarding
force. High accelerations and
Small roll angels uncomfortable behaviour
A short natural
period.

18
Calculation of responses in waves
The movements of a barge in irregular waves can be calculated from the wave spectrum
function and the transfer function of the barge’s hull in a particular loading condition. The
transfer functions define the steady state responses of the hull per unit wave amplitude as a
function of effective angular frequency and heading. Each response, such as a roll angle, has
its own transfer function.
By means of a simple multiplication for every angular frequency, a response spectrum is
obtained defining the response per unit amplitude for every angular frequency in the wave
spectrum.
Transfer functions are usually calculated by means of strip theory computer programs,
compare section 4.4. For detailed description of the mathematical procedures and definitions,
compare ship hydromechanics textbooks.

Magnification Transfer function Response Response spectrum

10 30

8 25

20
6
15
x 4 =
10
2 5

0 0
Angula Angular
w1 w2 w3 w4 w5 w6 w7 w8 velocit w1 w2 w3 w4 w5 w6 w7 w8 velocity

Fig 4.3 Simplified definitions: wave spectrum x transfer function = response

spectrum

Instead of transfer function, also the term RAO (Response Amplitude Operator) is used in the
literature and in this report. This term is used with varying definitions.
Here, the RAO is defined similar to the transfer function and varies with wave frequency.

RAO = Response / Excitation

For instances

RAORoll = Roll angle/Wave slope

19
Roll Damping
The roll is the movement hardest to predict. Damping in roll is very small and it is measured
as a damping ratio, which indicates how much the roll amplitude is decreasing between two
roll cycles. It is typical for roll motions to have an effective non-dimensional damping ratio of
considerably less than 5 % for a bare hulled ship. [1] That is why waves with an Encounter
frequency4 near the barge’s resonance peak in roll, are so critical, and should be avoided. If
that is impossible, extra damping is needed. The usual roll damping components are: [20]

• Lift damping The damping energy corresponds to lift effects at

forward speed due to potential flow effects. The lift
damping is zero at zero forward speed. I.e.
apartment barges have no lift damping.
• Wave making damping The damping energy corresponds to the energy in
the Radiated waves.
• Frictional damping The damping energy corresponds to body-fluid
friction due to viscous effects. This component is
very small at zero speed and thus negligible for
apartment barges.
• Vortex shedding damping The damping energy corresponds to vortex shedding
at sharp corners and bilge keels also due to viscous
effects. The contribution from this component is
large at zero speed.

The first two damping components are considered to be linear in roll velocity. The other two
are non linear, and basically quadratic to roll velocity, due to viscous effects. [4]
In cases where the water depth is limited, an increase in damping could be expected since the
water will be compressed between the seabed and the bottom of the barge, when the barge is
rolling. However, no test results concerning this have been found in the literature.

4.2 Stability
The ability of a floating object to regain its equilibrium after a disturbance is often referred to
as stability.
The stability describes the expected behaviour of the barge. The stability consists of an initial
stability, corresponding to the righting moment trying to retain equilibrium when exposed to a
small disturbance, and the stability width, which is the maximum static roll angle that the
barge can have, and still try to retain balance.

4
The encounter frequency is the frequency of the waves that the barge will comprehend, with speed and course
taken into account. The encounter frequency is given as ω = ω − ω ⋅ U ⋅ cos( µ ) but for moored barges U=0 and
2

e
g
thereby ωe = ω .

20
 Fig 4.4: Cross section of the barge. [32]

A plain barge has high initial stability, often GM ≅B. For a simple rectangular shaped barge,
like in Fig 4.5, the GM can be calculated as [1]:

GM = KB + BM − KG

Where KB = the distance from keel to the centre of buoyancy

L ⋅ B3
BM = ∆ = The displacement
12 ⋅ ∆

KG = The distance from keel to centre of gravity

The GM can be used to calculate the GM − curve and here the behaviour of the barge can be
seen, the initial stability as well as maximum static roll angle.

21
 Fig 4.5 A typical GM-Curve

A wide barge will have a high initial stability, GM , and a high stability width. The high
stability width is always a good thing from a safety point of view, since it reduces the risk for
capsizing caused by icing or collision. However, high GM means stiff roll movements and
high vertical acceleration.

4.3 Natural Frequency

In engineering it is always important to predict the natural frequency of the object engineered,
in order to avoid coincidence with any expected disturbances. If that happens, it will give a
peak in the transfer function.
The natural period of a ship or a barge can be calculated as:

2 ⋅π ⋅ k4
Tn =
g ⋅ GM

I 44 + A44
Where k 4 = the radius of inertia
∆
I 44 = Inertia (m2kg)
A44 = The added mass of water (m2kg)
∆ = The displacement (kg)
In a simplified version of the formula, Tn is estimated to:
B where K is empirically determined to 2,27 for ships with a normal
Tn = K ⋅
g ⋅ GM mass distribution, which can be used for the apartment barges.

22
This typically gives a natural period Tn of around 2.5 - 4 sec, e.g. a natural angular velocity of
1.6-3 rad/s, depending on beam and GM , for the barges studied in this report.
The dream would be to avoid that the natural frequency of the barge is close to any wave
frequency that could occur at the actual location. Unfortunately this is impossible and
therefore extra damping is needed or other roll damping actions should be considered.
One conceivable method is to use a wide barge that should be less sensitive to the shorter
waves and thereby avoid most wave responses. On the other hand the higher stability of a
wide barge will be a problem.
A third way is trying to lift the centre of gravity and thereby decrease the stability and get a
longer natural period. The loss in stability will decrease the accelerations and create more
comfortable movements for small waves, but worse behaviour in heavy sea. A higher KG will
also decrease the stability width.

4.4 Strip theory – a way to predict movements and accelerations

Strip theory is the common way of solving 3-dimensional problems using a 2-dimensional
technique. By modelling the ship as consisting of many 2-D strips, the equations for heave,
sway, roll and the coupling between these motions are solved much easier. [5]

Fig 4.6 The ship divided into 2-D Strips [14]

23
Strip calculations will be used in an example below. The effects of different fetch lengths on
the movements of barges with different shapes are calculated.
The program used is Wolfson Ship Motion program [8], a strip program developed at the
University of Southampton. The program is used in Applied Marine Technology. The sea
spectra used in all calculations made in this report are the I.T.T.C Two Parameter spectrum,
where the parameters characterizing the waves are Hs and Ts .
The following assumptions are made when using strip theory [7]:
1. The ship is slender. (The length is much greater than the beam, or the draught and the
beam is much less than the wavelength.)
2. The hull is rigid so that no flexure of the structure occurs
3. The speed is moderate
4. The motions are small
5. The ships hull sections are wall-sided
6. The water depth is much greater than the wave length so that the deep water
approximation may be applied
7. The presence of the hull has no effect on the waves

Assumptions number 1 and 6 are not completely fulfilled for barges. These inaccuracies will
be dealt with as follow:
Number 1 The influence of different barge lengths has been studied in a test using the
Wolfson Motion Program. The results show that no differences appear for
B/L ratio of 1, 2 or 5 if the waves are coming right from the side. See
appendix A. For other inclination angles the results must be taken with “a
stitch of salt”. The results from the strip calculations are more reliable for
inclinations of 90°, since the ratio B/L does not have any effect of the result.
Number 6 The fact that the water depth is not sufficiently deep near the shore, where
barges are usually moored, is assumed to contribute to a conservative
estimation of the roll damping. In a physical sense it is clear that the damping
in roll must be increased on shallow water. Ikeda has also confirmed this as
logic assumption, but no tests on these effects have been found. [Mail]

According to strip theory the motion damping arises because the oscillating hull radiates
waves away from the ship. In all motions but roll, this is a correct assumption, but in roll there
are other effects that are more important. It is the eddy making and skin friction in
combination with any existing appendages that contribute to most of the total damping. [7]
Waves with a frequency near roll resonance will induce high roll angles and a high vertical
acceleration. In resonance, strong nonlinearities in the hydrodynamic roll damping make any
analysis concerning roll difficult, especially using strip calculations. [1] Since the non-linear
viscous effects are significant in the estimation of roll at resonance, the results should not be
given too much importance. Where the responses are high and it can be assumed that the
barge is rolling in resonance, it is common to “cut the peaks” or take the response value
besides the peak with the purpose of adjusting the calculated responses to match reality. [32]
This implies that a linear first-order roll motion does not exist at resonance and thus causing

24
 inaccuracies in the prediction of second-order hydrodynamic forces. [18] In order to make the
strip results more accurate, a function in the strip program can be used, enabling the use of
results from full-scale test on roll damping, in order to calibrate the predictions of
accelerations and roll.

The effect of fetch lengths on roll and accelerations

As an example of how important the fetch length is for the movements of barges, the
accelerations and roll angle have been calculated for two different wind speeds, 15 m/s and 20
m/s. The barge used in the calculations has the following dimensions:
• B=14
• L= 30
• T= 1.7 m
The point of measure on the barge is amidships, 0.5 m from the side at height 2 m above keel.
This point is chosen because the accelerations are worse closest to the edge.
The fetch lengths studied were 500-2500 m. For every wind speed and fetch length, the wave
height and the dominating angular velocity are calculated.

Table 4.1 The different sea states as function of the fetch length

Wind speed 15 m/s Wind speed 20 m/s

Fetch length (m) Hs (m) ωN Hs (m) ωN

500 0,2 2 0,27 1,7
1000 0,26 1,7 0,36 1,4
1500 0,31 1,5 0,43 1,3
2000 0,35 1,4 0,49 1,2
2500 0.38 1.4 0.54 1.2

The calculations have been made in Beam Sea, where the waves are coming in right from the
side of the barge. The results are illustrated in Fig 4.7-4.8 below. In each diagram, the levels
of the comfort criteria are indicated.
m/s2 (RMS)
8
15 m/s
20 m/s
Comfort
6

0
0m 500m 1000m 1500m 2000m 2500m

Fig 4.7 The roll for different fetch lengths 90°

25
 2
m/s (RMS)
3
15 m/s
20 m/s
Comfort

0
0m 500m 1000m 1500m 2000m 2500m

Fig 4.8 The vertical accelerations for 14m x 30 m barges and different Fetch
length 90°

The importance of a short fetch length is seen above, where the accelerations are well above
the maximum acceptable.

26
5 Methods of limiting Accelerations and Roll
It is mostly high accelerations and not the roll angles that will cause seasickness. In general,
wider barges with high stability give high accelerations and small roll angles, and vice versa
goes for narrow barges with low stability. A good example of this is an old fishing boat,
which has low stability that gives rise to high roll angles, and low accelerations, but the boat
is still comfortable.
To get a comfortable apartment barge where the furniture is not moving around and the
crystal glasses are not thrown to the floor at every second wave, the designer wants to limit
the accelerations without getting too high roll angles.
In this chapter, seven different potential approaches to limiting the roll angles and
accelerations will be described.
• Bilge keels
• Active fins
• Anti-roll tanks
• Mooring damping
• Pontoons as mobile breakwaters
• The sail damping effect from the building
• Changing the beam
• Changing the centre of gravity KG
Additional methods of limiting the rolls and accelerations, not considered suitable for
apartment barges are, among others, Paravanes, Twin Keels and Pneumatic breakwaters.
Further information concerning them can be found in marine literature.

5.1 Bilge keels

Bilge keels are the cheapest and most common way to add damping in roll. The bilge keels
consist of standing plates along the ship/barge; often placed in the corners, see the Figure
below. For ships, it is important to have the bilge keels along the Streamlines 5, in order to
minimize the extra resistance. For apartment barges the resistance is of minor interest since
they are moored most of the time. Therefore bilge keels can be placed across the direction of
the imaginary streamlines. Transverse keels can be needed when the barge is almost quadratic
and therefore shows almost the same behaviour in pitch as in roll.

CL

Bilge keel

Fig 5.1 Typical Bilge keels

5
Streamlines are the trajectories that the water particles follow when the ship make speed through the water.
From the equation of Bernoulli, it can be seen that the pressure is always constant along each specific
Streamline.

27
Ships are made of steel and it is very simple to put on bilge keels, by simply welding them
onto the hull. Apartment barges are made of either steel or concrete. For the latter, attaching
keels by bolts into the concrete is possible, but it is easier if thought of prior to manufacturing.
Bilge keels increase the viscous damping and are efficient even in severe sea. [1] The roll
damping coefficient is often increased with a factor of 4 for waves close to the ships natural
frequency even with moderate sized keels. A typical reduction in accelerations and roll are
35%-50%. [15]
Some recent experiments show that the roll damping from bilge keels is decreased with
shallow draught. This is often the case for barges, but the loss in damping is expected to be
small. [6]

Fig 5.2 The change in RAO when using bilge keels. [1]

28
The extra damping with bilge keels works by vortex shedding or as an increase in added
water mass6 to the barge. The damping force (L) can be expressed as [5]:
1
L≈ ⋅ ρ ⋅ Α ⋅V 2 ⋅ CD
2

U max ⋅ T
C D = 8,0 ⋅
2⋅b
Where T = Period time
b = Breadth of bilge plate
Since the damping corresponds to the square of the velocity, the bilge keels will give
maximum contribution to the damping when the velocity in roll is maximal. This occurs when
the barge has a roll angle equal to zero, but is rolling.
In experiments, the positioning of the bilge keels has been shown to have great importance.
The longer the distance between the keel and the centre of roll, the higher the velocity and
consequently an increased damping is gained. The flow passing around the bilges is impeded,
which has a damping effect as well. [6] A long distance between the keels and the centre of
roll also increases the lever of the retarding force and thereby the damping moment.

Fig 5.3 The damping for different positioning of bilge keels[1]

6
The added water mass follows the ship as it moves and can be seen as an extra mass added to the mass of the
hull. According to the laws of Newton, an increase in mass causes a reduced acceleration as a result of a specific
force.

29
5.2 Active fins
Active fins work in a way similar to the wings of an airplane, but underneath the water. A
control unit adjusts the fin; thereby give a restoring moment counteracting the roll. By
moving the fin, the angle of attack is changed; sometimes increasing the lift force and
sometime giving a negative lift, depending on where in the cycle of roll the ship is at each
moment. The active fins also contribute to the damping by increasing the vortex shedding in
the same way as bilge keels. However the small area of the fins only gives a limited effect.

Roll moment
due to waves - Ship

Roll moment Roll

due to fins

Active fins

Fig 5.4 The Open loop control system for active fins

In order to be efficient, fins generally demand a high forward speed. In the right speed
conditions, active fins are the best anti rolling devices for ships. However, producers of active
fins have recently expanded the use also to low and zero speed. The active fins are most
common on yachts where the rolling causes great discomfort, when anchoring or drifting. By
moving the fins in a proper way, the damping is increased. This is an expensive and complex
system, not very well designed for permanent use on barges. The use on yachts is more
suitable, because of their small displacement and the yacht owner’s willingness to pay.

5.3 Anti-roll Tanks

Anti-Roll Tanks, also called Sloshing Tanks, is a rather common and sometimes an efficient
method of limiting the roll angles. In successful tests, they have reduced the amplitudes and
the accelerations in the range of 50%-60%. [15] Anti-roll tanks can never eliminate the rolling
entirely, since it needs a roll angle in order to function. The tanks work by moving the centre
of gravity of the water in the tank and thus give rise to a moment that counteracts the roll. The
optimal behaviour of an anti rolling tank is shown below.

Fig 5.5 Anti-roll tank behaviour. [1]

In roll, the RAO, the Response Amplitude Operator, can be higher than 10 for a bare hull.
Using a properly designed tank, the roll is reduced. The diagram in figure 5.6 illustrates the
typical influence that anti-roll tanks can have on the RAO of a ship.

30
 Fig 5.6 Comparison of the computed roll response of a typical ship with and
without an anti-roll tank. [1]

The important parameters, when considering using anti-roll tanks, are positioning, size, duct
area, natural frequency of the tank and the quota dGM/GM, i.e. the decrease in metacentric
height as a result of adding the ballast tank, divided by the original GM. The dGM/GM should
vary between 0.15 and 0.3. [5] In experiments, it is generally found that a satisfactory degree
of stabilization for ships is achieved if the weight of the liquid in the tank is in the order of 2-5
% of the ship’s displacement. [8] According to Lloyd 1-5 % is enough. [7]
It is always best to install the tank as high as possible in the ship, with the duct close to the
centre of gravity. One disadvantage with tanks is that they have to be optimised for a certain
combination of speed, heading, GM and sea state. An advantage when applying this technique
on apartment barges is that their speed is always zero. On most locations one can also predict
from which direction the worst waves will come, and thereby optimise the tanks for those
conditions.
Because of the free water surface in the tank the barge will lose some of its metacentric
height. This will give a slight loss of stability, which contributes to the decrease in
accelerations, but also to an increase of the natural period in roll.
There are active and passive tanks. The active ones are complex and expensive systems with
valves or pumps. If the wavelength is varying a lot then it might be justified to install active
tanks since they can adjust their behaviour according to the wavelength. They are of course
better than passive ones, but to this purpose they are seldom motivated for barges.

31
The different types of anti-roll tanks are
• Free surface tank
• U-Tube tank
• External tank

Fig 5.7 Different types of passive anti roll tanks [1]

32
Free surface tanks
When designing passive free surface tanks, the natural period of the tank should be chosen to
be equal or up to 10 % higher than the barge’s natural period in order to make the tank as
efficient as possible. The natural period for a free surface tank can be estimated to be:

2 ⋅π 1 4 ⋅π ⋅ b h = the water depth in the tank

ωN = = ≈
T N  g ⋅π π ⋅h
1/ 2
(g ⋅ h )1 / 2 b = the breadth of the tank
 ⋅ tanh 
 b b  b is normally >> than h.

U-tube tanks
The angular velocity ωN corresponding to the resonance frequency of the U-tube tank should
be chosen equal to the angular velocity corresponding to the resonance frequency of the
barge. The formula below [16] can be used when choosing the dimensions of the u-tube tank:

c ττ 2 g ⋅ hd
ω t = =
a ττ W rW + 2 h r ⋅ h d

For all passive tanks, the maximum moment due to the motion of the fluid occurs when the
hydraulic jump is at its midpoint, i.e. when the roll velocity has its maximum. [5] In order to
keep down the aggregate accelerations for the tank and the barge, it is important that the
maximums do not coincide in time.
For a barge with the dimensions 18x23 m and a draught of 1,7 m, the formula above has been
used to design a U-Tank with ωt = 1.9 rad/sec. The tank dimensions in the figure are chosen
through an iterative process.

2m

3m
2.6 m

4m

Fig 5.8 The chosen dimensions of the U-Tank used in a test

33
Strip calculations were done, comparing the RAO for the barge with and without an anti-roll
u-tank. The length of the tank, and hence the water volume, was augmented in small steps
from 1 m until a significant reduction in RAO was achieved. This occurred at a length of 2 m.
The total mass of the water in the tank is 21 ton, which is 9 % of the displacement of the
barge. This is more than expected – according to earlier findings concerning anti-roll tank size
for ships, a water volume around 5% of the ship’s displacement should be more than enough.
The reason for this difference is probably that the high initial stability limits the dGM/GM.
The roll response is decreasing, but not as much as expected. One of the reasons for this is the
low initial RAO response for this wide bare hull. Even more important, in the case with a
wide barge, the high GM gives that the optimal dGM/GM-relation (0.15-0.35) is not
achieved, unless a very large tank is used. The typical appearance of the RAO-curve for a
barge with a tank is shown in Fig 5.6. The barge’s behaviour in resonance is improved, but for
angular velocities above and below resonance it is deteriorated.

In the diagram 6.6 above the RAO for roll is showed, with following definitions:
Roll angle of the Ba rge
RAO =
Slope of the wave

External tanks
External tanks work by providing extra viscous damping when the water pours in and out
through holes and by transforming kinetic energy from the rolling to potential energy, that is
needed to lift the water in the tank. This might be a very good solution for apartment barges.
The extra resistance caused by external tanks make them a bad solution for ships, but for
moored barges the only disadvantages are the extra cost and need of maintenance of the tank.
A clever solution can be to integrate the external tank in gangways to land, or to a breakwater.

Fig 5.9 External tank for reduction of roll as part of the gangway

34
5.4 Mooring damping
In standard mooring, iron chains are used. A long and heavy chain, partly lying on the seabed,
will slow the movements down and prevents the chain from getting tight and causing
uncomfortable jerks. However, for movements caused by strong wind, current or big waves,
there might be uncomfortable wrenches in the anchor chains, despite their considerable mass
and length. This problem gets worse by the fact that apartment barges are moored in shallow
waters, thus limiting the possible length, and thereby weight, of the chain.
In roll, the lifting and stretching of the chain is meant to increase the damping by absorbing
energy, but in experiments this has been shown to have only small effects, since the mass of
the chain lifted is much less than the living mass of the barge and therefore this effect is
negligible. [21]
There are other products for mooring, for example Seaflex [28], consisting of elastic rubber
“ropes” with high elongation and an increased force per unit stretched. What is special with
Seaflex is that there is no pulling back force.

Fig 5.10 The Seaflex is available with of up to eight parallel rubber stripes
and can be used both for vertical mooring and for “cross mooring”

The Seaflex can be installed both in a cross and vertically, with a pretension counteracting the
roll as well as taking care of positioning the barge. Careful calculations must be made to
ensure that the Seaflex has the correct length and strength, giving the best behaviour possible
in high water levels and rough waves. The maximum elongation is 120 %, so the Seaflex must
be long enough to ensure that this percentage is never exceeded. The installation of the
Seaflex must be carefully done, since a correct pretension is a prerequisite for the method to
work as planned. The diagram below shows the force per stretched percentage for one
Seaflex.

35
 Fig 5.11 Force per elongation for Seaflex

In view of the fact that the elongation of the Seaflex gives an increase in force per unit of
length, the retarding force will increase and have its maximum when the barge starts to roll
back. This fact gives a peak in negative acceleration in roll, because the normal roll
acceleration and the acceleration provided by the Seaflex coincide in time as well as in
direction. The peak in acceleration must be considered, since it is potentially increasing the
incidence of seasickness.
The Seaflex can limit the roll angles, but further testing on the increase in accelerations must
be done.

36
5.5 Pontoons as mobile Breakwaters
Moored pontoons are commonly used in order to protect small harbours or other sensitive
marine installations. This method can be suitable also for apartment barges. Besides the wave
damping effect, the pontoons can be used as gangways between the barge and land or as a part
of a marina enclosing the apartment barges. By using pontoons enclosing the apartment
barges, an extra evacuation possibility gets available, which improves the total safety.
The incoming energy of a wave can be divided into three parts;
The transmitted part
The reflected part - Depending on the pontoon’s dimensions and the wave
characteristics, a part of the wave’s energy will be reflected out from the breakwaters.
The dissipated part, which is energy losses because of the disturbed wave motion.
The transmission coefficient, KT, is a measure of the efficiency of the breakwaters. It is
defined as the transmitted wave height divided by the incoming wave height. KT often has a
value around 0.3-0.6 depending on the wavelengths and the breakwaters dimension.
The apartment barges themselves also work as breakwaters for adjacent units. This means that
mooring barges beside each other might be beneficial. The first barge will be exposed to the
undisturbed wave system, but the barges behind will experience a damped wave system and
the wave height will be reduced for every barge the waves passing. The KT coefficient for a
floating pontoon can be estimated as [17]:
1 W = the weight of the pontoon per unit length of wave crest
KT =
 π ⋅W 
2 intercepted
1+  
 w⋅ L⋅d  w = the specific weight of the water

However, this formula is valid only when the pontoon penetrates almost to the seabed, but the
KT increases when the ratio draught/depth increases.
For a barge with dimensions 14x30 with draught 1.7m, this will give a KT of 0.57, which will
reduce the wave height with approximately 50 % and thereby the energy with almost 70 %.

5.6 The Sail-effect from the walls

On old fishing boats sails are often used to increase the damping in roll during storms. By
moving back and forth, the area of the sail gives rise to a force counteracting the roll. The
walls of the building will also work as sail. They will act like bilge keels, but in the air instead
of in the water. However, there is a big difference in density between water and air. The
relation between the media’s densities is:
ρ air 1 .2
= = 0 .012
ρ water 1000
This means that a “sail” above the water must be much larger than a bilge keel under water in
order to achieve the same effect. As an example, the sail effect will be evaluated for a 10 x 42
m barge. These dimensions are chosen since a narrow barge has a large wall area and thereby
the wall damping should be more significant than for wide barges. The height of the walls
above the water level is approximately 7 m for a two-store apartment barge.
To find out whether the sail-effect is worth taking into account, strip calculations were
performed. To enable the calculations, the wall must be translated to a bilge keel with the

37
same damping effect. The uniform pressure by the wind is replaced by a force acting in one
point of the wall. The lever from the point force to the centre of roll must be equal to the lever
from the imaginary bilge keel. See figure 5.12 below. The wall area is in this case 42 x 7 m =
294 m2. Translated to an area below water, this will be equivalent to 0.35 m2, a small area
even for a bilge keel.
As a comparison, the area of the bilge keels in example 5.1above was chosen to 20 m2 for the
same barge size. This leads to the conclusion that the damping effects from the walls are
negligible, which is confirmed by the strip calculations.

Fig 5.12 The Sail damping effect translated to a bilge keel with the same
effect

5.7 Choosing appropriate dimensions of the barge

Changing the dimensions of the barge might be seen as a more fundamental change than just
small improvements and damping devices. However, since the barge’s dimensions influence
its behaviour, some different modifications, all able to give the same useful living area will be
presented in this section. The “methods” studied are:
• Changing the beam
• Changing the centre of gravity

Changing the beam

A test was performed where the displacement and deck area was kept constant, while the ratio
B/L was changing. The Wolfson Strip program’s input data was L, B, KG and radius of
inertia7. These tests were made using the wave data from Bällstaviken, one of the example
locations studied later in this report.
The following sea states for Bällstaviken are used in the calculations:
Table 5.1 the different sea states for Bällstaviken
Sea State Wind speed (at Sea) Wave Height Mean Period Wave Length
B1 5 m/s Hs=0.06 Tm=0.9s L = 1.3 m
B2 10 m/s Hs=0.13 Tm=1.3s L = 2.8 m
B3 15 m/s Hs=0.21 Tm=1.6s L = 4.2 m
B4 20 m/s Hs=0.29 Tm=1.9s L = 5.6 m
B5 25 m/s Hs=0.37 Tm=2.1s L = 7.0 m
B6 30 m/s Hs=0.46 Tm=2.3s L = 8.5 m

7
The radius of inertia was estimated to be: Kxx=0.35*B, Kyy=0.25*L, Kzz=0.25*L

38
In sea state 6, the wavelength λ was 8.5 m and the mean period was 2.3 sec. According to
PNA small movements are to be expected if B is larger than 2⋅λ (for B=18 m) but according
to the strip calculations made for different barge dimensions, see diagram below, roll angles
of 9° (RMS) were to be expected. With these RMS-values, roll angles of 20-25° can be
expected, maybe not producing seasickness symptoms, but definitely a total mess in the
apartments.
The diagram below shows the roll angles resulting from waves coming from the side, which is
of course the wave inclination angle giving the highest rolls. If the barges can be moored so
that the incoming waves are not coming in right from the side, a substantial reduction will be
achieved. As an example, the roll responses for the same sea states as above are calculated for
waves coming in from an angle of 60°. The result was a significantly decreased RMS values.

Roll angles
Wave inclination angle 60 ° Wave inclination angle 90 °

Degrees (RMS) Degrees (RMS)

10 10
Barge 10x42 Barge 10x42
Barge 14x30 Barge 14x30
Barge 18x23 Barge 18x23
Comfort Comfort
8 8

6 6

4 4

2 2

0 0
B1 B2 B3 B4 B5 B6 B1 B2 B3 B4 B5 B6

Fig 5.13 Roll response for different B/L and constant displacement, with
different wave inclination angles.

39
When taking the wave inclination angle into account, the roll angles (RMS) are always below
2°, except for the short and wide barge. For a complete analysis of roll responses and
accelerations in the Bällstaviken project, see chapter 9 below.
The results from the strip calculations do not agree with PNA, which predict small
movements and accelerations for barges this wide. This contradiction will be discussed later.

Changing the center of gravity

The distance from keel to the centre of gravity, KG, of an apartment barge is possible to
adjust within an interval. If benefits are gained, the designer should work with that in mind
when choosing dimensions. Strip calculations have been used to evaluate the differences in a
barge’s behaviour when KG is allowed to vary one meter in both directions, between 1.7 and
3.7 m, from the estimated 2.7 m.

Roll angles
Degrees (RMS)
16
KG=1.7m
KG=2.2m
KG=3.2m
12 KG=3.7m
Comfort

0
B1 B2 B3 B4 B5 B6

Fig 5.14 Roll response for different sea states and different KG (Barge 10x42
m and wave inclination of 90°

With a high KG there will be a decrease in metacentric height GM. For the sea states 1-4 that
gives a better behaviour, with small roll angles. On the other hand, the loss in stability in
heavy weather causes extremely high roll angles. If the actual location is well protected and
there are no risks for waves like in sea state B5 and B6, a higher KG will give a more
comfortable behaviour.

40
Vertical accelerations
The same pattern is shown in the case of vertical accelerations, as can be seen in the diagram
below.
2
m/s (RMS)
3
KG=1.7m
KG=2.2m
KG=3.2m
KG=3.7m
Comfort
2

0
B1 B2 B3 B4 B5 B6

Fig 5.14 Vertical accelerations for different sea states and different KG
(10x42m)

41
6 Rules and Regulations
The Swedish Maritime Organization considers apartment barges not as ships, but as ordinary
houses/buildings. The construction rules from the Swedish National Board of Housing are to
be used and for every specific construction, the Community Administration must grant
permission. Some of the construction rules are difficult to fulfil, since ships/barges are
normally designed to satisfy the rules of the Maritime Organization, where safety is the main
issue. There are for example problems with designing emergency exits suitable for the
disabled. Since the barges are considered as ordinary buildings, it is possible to have the barge
as registered address.

6.1 Fire safety rules

The rules to follow concerning fire safety are the common rules from the Swedish National
Board of Housing, concerning all housing. According to their rules: “Further fire protection
measures may be required in addition to the requirements set out in this section (Section 5) if
action by the rescue service cannot be expected within normal attendance time and if such
action is essential in order that
• …escape from the building may be possible as intended, and
• …the spread of fire to nearby buildings may be limited.” [30]
Since the barges are situated on the water, the accessibility from land is restricted. Even
though the barge is not accessible from all sides, the fire department should be able to fight
the fire, within normal time. [29]

6.2 Stability and Leak stability

The standard rules from the Swedish Maritime Organization are not mandatory but could be
used as guidelines. The use of separate watertight sections of the barge will reduce the risk of
sinking, and should be considered, at least for large barges with apartment block.

42
7 Safety and System Analysis
In order to assure the safety for the people living on board and their guests, a thorough
analysis of the system must be performed, including possible risks and risk limiting
precautions.
The barge is connected to land in different ways. The connections are part of the system and
therefore they should be included in the safety analysis. The connections between the barge
and land are:
• Electricity for use onboard
• Freshwater for use onboard
• Sewer from the barge to land
• Gangway for entering the barge
• Mooring arrangement
All these connections must be flexible and able to follow the movements of the barge. The
barge is normally moving in the xy-plane because of changes in wind and waves and also
along the z-axis depending on shifting water levels.
Restrictions in the barge’s movements in the horizontal plane can be achieved by using a pole
driven down into the seabed and attached in a ring on the barge. Depending on the water
current and the wind pressure between two and four poles are used. However, most common
is the use of anchor chains, since it is a simple solution giving damped movements in the xy-
plane.

7.1 Risk analysis

The risk is usually defined as the probability that an accident occurs, multiplied by the
consequence if it occurs. [10]

Often it is the impossible, the only

thing not thought of, that in the end
will occur and thereby causing a
disaster.

Risks can be separated into three groups: the risk for humans to be injured or getting killed,
the risk for economic damage including loss of income during repair, and risk for damaging
the environment. [9]

RISK

Environmental
Personal injury Economical loss
damage

Fig 7.1 Three different types of risks

43
A good way to start is to list all possible risks along with their consequences and probability
of occurrence. When all risks are identified, precautions can be taken in order to minimise the
probability that each event occurs. If the probability is still not considered negligible, efforts
should be made to reduce the consequences.
Classes of probabilities and consequences are sometimes used when estimating the different
risks, and probability values are calculated. [24] Since it is often impossible to make accurate
estimations, the use of probability classes is often a better way, since they are less sensitive to
inaccuracies in the assumptions made. The system used in this report is presented in the tables
below. [24]
Table 7.1 Probability classes

Probability class Name Frequency

A Common 1-10 times/year
B Very likely 1/1-10 years
C Likely 1/10-100 years
D Possible 1/100-1000 years
E Unlikely < 1/1000 years

The consequences can be divided in classes in a similar way. In the table below, the system
used for consequences classes is presented, with examples of event from the apartment barge
case.
Table 7.2 Consequence classes

Consequence class Description Example of event

Bilge alarm mal
1 Near-accident At least one level of safety remaining
functioning
Property damages approx.<0.5% of Crashed scuttle above
2 Minor accident
total value water level
Personal injury and/or Property
3 Accident Fall accident
damages approx. 0.5-5% of total value
Severe Personal injury with up to a few
4 Major accident dead. Property damages approx. 5-30% Sinking during daytime
of total value
Many dead people. Property damage Quick sinking during
5 Catastrophe
>30 % of total value night

For a more sophisticated analysis a possibility is to make a consequence table for each risk
type.
In order to make it more obvious which are the worst risks, each risk can be put in a risk
matrix, see figure 7.3 below.

44
 Probability

Consequence

Fig 7.3 The Risk Matrix

The risks assigned a high probability and severe consequences, will be placed in the upper
right corner. The risk matrix is a good tool when choosing which risks one should concentrate
on reducing. [9] The risks in the upper-right corner must be eliminated, by reducing the
consequence or the probability (or both). An efficient way of decreasing the probability of an
essential function to fail is to insert redundancy in the system.

Risks for apartment barges

In order for apartment barges to fulfil safety demands and to become comfortable and
popular, actions to avoid the potential high risks recognised, must be taken. In the cases where
the risk is high and it is impossible to eliminate or reduce the probability of occurrence,
precautions to minimise the consequences will have to be made. Risks identified for
apartment barges are presented in the table below.
Table 7.3 The identified risks
Injury/discomfort Property Environmental
to humans damages damage
1 Fall accidents caused by slippery footpaths X
2 Health problems due to high moisture and wind X
3 Taking in water by inlets, outlets or scuttles X X X
4 Corrosion X
5 Mould/water damage X X
6 Collision X X X
7 Broken mooring arrangements X X X
8 Damages to the hull from ice-pressure X X X
9 Sewer leakage X
10 Fire on board X X X

Even though the list of possible risks above is probably not complete, it will serve as a
foundation for an example of a brief safety analysis. To find out all possible risks, “focus
groups” can be used, consisting of people with different experience related to the activity.
Depending on the barge’s location, the potential risks can vary a lot and must be studied

45
carefully for every specific project. For each of the risks listed above, probabilities and
consequences will be discussed and classes will be estimated.
At first, the probabilities and the consequences for each risk in the table above will be
estimated, if possible. After that, possible precautions will be described, and new probabilities
and consequences will be estimated. The consequences and probabilities are very hard to
predict, but they are hopefully good enough to serve as examples when comparing the
different risks.
In order to estimate the probabilities, statistics of failure is essential. Records are often kept,
for example of fire accidents. If data like this is available, it is a very important source of
information.

1. Fall accidents caused by slippery footpaths

Without any precautions what so ever, the risk is assumed to be very high and is estimated to
be in probability class A. The consequences are estimated to class 2-4 since the risk for lethal
accidents are obvious in the case with drowning and potential fractures in fall.
Equipping the gangways with anti-slip devices and perhaps heating of the gangways during
wintertime reduce the probabilities of fall accidents to class B.
By trying to avoid sharp corners and edges on the barge and gangways, it is possible to reduce
the consequences of a fall. Other precautions reducing the consequences is proper railing,
preventing people from falling into the water and the installation of ladders down to the water,
where people can climb up, if they have fallen. This will reduce the drowning risk. The
consequence is then reduced to 1-3.
1. Fall accidents Probability Consequences
Initially A 2-4
With precautions B 1-3

2. Health problems due to high moisture and wind

There are no signs of any increase in pneumonia or mould causing unhealthiness, due to
increased humidity. On the contrary, higher humidity will increase the comfort, especially
during winter.

3. Taking in water by inlets, outlets or scuttles

Without any extra precautions, the probability for sinking caused by freezing of inlets/outlets
or leaking scuttles is pretty high and is given the value B in probability. The consequences are
in this case 3-5 and will affect property damages as well as potential death of the habitants
and environmental damages.
The first measures to be taken, should be to minimise the number of all inlets/outlets and
those necessary, should be installed above the water surface where they do not cause water
leakage if not tight. The scuttles close to the water surface should be properly sealed and
impossible to open. They should be located on a proper distance above the water level and
well protected from being broken from the outside by external violence from i.e. boats.
A good idea is to install a timer on the fresh water feeder, to avoid that a broken pipe inside
the barge will lead to that the barge is filled with water. This method is used in the house
barges at Pampas Marina, one of the example locations studied in this report. The water is
automatically shut down after 30 minutes of use.

46
As an extra safety precaution, bilge pumps should be installed with capacity enough to
evacuate the water even from potential fire extinction. The consequences of water penetration
can be reduced if it is possible to separate the barge in watertight sections. This will improve
the chances of survival dramatically. The consequences can also be reduced by alarms,
waking people asleep.
The most dangerous chemicals/substances stored in ordinary housed are the diesel oil for
heating, and this must be the case even for the apartment barges. Environmental damages
might be caused by a discharge of chemicals and diesel oil for heating. Most efficient is of
course to avoid diesel oil completely, but if that is impossible, efforts to reduce the
environmental consequences should be taken in order to minimise the quantity, discharged.
Removing all inlets and outlets below the water surface and not having any leakable scuttles
close to the water line can reduce the probability to class C-D. By redundancy in the bilge
system and alarms for high water level inside the barge, the probability for any consequence
at all, can be reduced one class, getting D-E. The reason that it is impossible to always have
sufficient bilge capacity is that during potential sinking scenarios the electricity is often put
out.
3. Taking in water Probability Consequences
Initially B 3-5
With precautions D-E 3-5

One barge design, not mentioned in this report, uses expanded styrene in the space below
water. This makes taking in water impossible, but it also rends impossible any use of that area
below water. This design is most favourable for smaller one family barges where the losses in
useful areas below water are smaller.

4. Corrosion
There are two main manufacturing materials for barges; steel or reinforced concrete. For steel
barges painting is essential for protection against corrosion and the barge must be sandblasted
and repainted every 5-20 years. For the steel reinforced concrete pontoon there should be no
need for maintenance of the hull and the estimated lifetime is 50-100 years. It is certain that
the steel pontoon will be a victim of corrosion and corrosion is possible also for the concrete
pontoon if the reinforcements get in contact with water.
Corrosion is expected, and in order to avoid it, regular maintenance is needed. This includes
painting of the steel barge and well-exerted concrete casting. Corrosion will therefore not be
considered as a risk and will not be analysed further.
The consequence of corrosion is primarily economic, although if maintenance measures are
not taken, severe corrosion might, in the long term, lead to damages of the hull with leakage
as a result.

5. Mould/water damage
Neither the barges made of steel nor the ones made of concrete are damaged by moisture and
water, if well protected and properly ventilated. The interior fittings will be damaged only if
water gets inside and then the precautions and risks are as described in risk “Taking in water”
above.

47
6. Collision
The probability of damages to the hull, caused by collision, is very hard to predict. The
probabilities depend on the strength of the barge hull, but most important are the
characteristics of the ship traffic passing the barge.
For both the steel barges and the concrete barges, the hull is dimensioned to withstand forces
by water pressure, ice pressure and the weight of the apartments. When dimensioning the hull,
sometimes the maximum pressure used is assumed to be the same as the dimensioning ice
pressure used for bridge pylons8. Using this value, gives a conservative estimation of the
needed strength for the hull, since the ice pressure can hardly push as hard on a floating object
as on a pylon for a bridge.
In collisions with small pleasure boats, made of plastic or wood, there will be no or only
minor damages to an apartment barge, independently of the boat’s speed. In order for passing
ship to pose a threat for the barge, two conditions must be satisfied:
1. The impact energy, which equals the kinetic energy of the colliding vessel9, must
exceed the maximum absorbable energy of the two objects in the impact. [30]
2. The penetration of the barge must be in the water line or below.
All merchant vessels made of steel will constitute a threat to the barges. In order to estimate
the probability of collision, a study of the passing ship traffic needs to be done. The
probability cannot be assumed to be negligible and in the general case the probability is
estimated to class D. Serious collisions fulfilling the two conditions above will cause severe
damages, with consequences in class 4-5.
If the probability of collision is considered too high, a solution reducing the probability can be
the use of breakwaters or sandbanks surrounding and protecting the barges. This is estimated
to reduce the probability to class E.
To reduce the consequences of collision, dividing the barge in watertight sections is a good
idea that will reduce the seriousness of an accident, since the probability of sinking is
reduced. Even the use of fenders will provide some protection for the barge, if the colliding
ship is not too big. However the consequences will still be severe if a collision occurs.
6. Collision Probability Consequences
Initially D 4-5
With precautions E 4-5

7. Broken mooring arrangements

The mooring arrangements can include anchor chains, Seaflex or vertical poles driven down
into the seabed. Independently of the method used, dimensioning of the mooring devices must
be based on calculations and include a certain safety factor. If the actual location is windy and
open for waves, the possibility of the mooring system to break must be anticipated, here
estimated to be in probability class C-D. The consequences are varying depending on the

8
The values of the ice pressure used for dimensioning concrete pylons for bridges, are in the south of Sweden
250 kN/m, and in the north of Sweden 500 kN/m. [25]
9 1
The kinetic energy is ⋅ m ⋅ V 2
2

48
adjacent shores, ship traffic and wind conditions to that extent that an adequate consequence
class is hard to predict, but in the general case it is estimated to class 3-4.
To reduce the probability of drifting away, redundancy is efficient. The use of extra anchor
chains, which are not to be used under normal circumstances, can significantly reduce the
probability. See figure 7.4 below for an example of redundancy. The probability class is
hereby reduced to D-E.

Fig 7.4 Redundancy in the mooring arrangement

7. Broken mooring arr. Probability Consequences

Initially C-D 3-4
With precautions D-E 3-4

8. Damages to the hull from ice pressure

The hull is made to withstand the pressure expected from ice, no matter what the construction
material is; steel or concrete. The values of the ice pressure used for dimensioning concrete
pylons for bridges dimensioning are as mentioned above, 250 kN/m in the south of Sweden
and 500 kN/m in the north of Sweden. For steel barges, the ice will give an extra tear on the
painting, which might increase the maintenance needs, but this should be included in the
expected repairs. However, for a badly manufactured concrete barge, the probability for
explosive effects from water freezing in cracks is assumed to be pretty high. In areas next to
scuttles and bolts, the probability is higher. Without detailed knowledge of the actual quality,
the probability class is estimated to be B-C.
A potential accident is not assumed to cause any acute consequences and the material
damages are assumed to be small and easy to repair. Moreover, there is no danger for the
inhabitants. The consequence class is therefore estimated to 1-2.
Since the risk is assumed to be low, no extra precautions are made.
8. Damages due to ice press. Probability Consequences
Initially B-C 1-2

9. Sewer leakage
As mentioned earlier, the sewer must be able to move along with the barge’s translation
movements. Since it is unavoidable that the barge will move, the probability of leakage is
higher for sewers to barges than for sewers to houses on land. If the mooring arrangement is
broken the sewer will of course also be broken. The probability class is estimated to be B-C.
The environmental consequences of a leakage are assumed to be limited and no personnel
damage is expected. The consequence class is therefore 2.

49
The total risk is tolerable, since the high probability is compensated for by the low
consequences.
The consequences can be even further reduced, by an alarm for low pressure in the sewer.
This will limit the time until discovery and thereby the total dump, giving a consequence class
1.
9. Sewer leakage Probability Consequences
Initially B-C 2
With precautions B-C 1

10. Fire onboard

The risk for fire is divided in two; 10.1 for the one family barge and 10.2 for the apartment
barge. Fire accidents are the most common reason for death in houses on land. There is no
reason to assume that the probability of fire should be lower on barges, since smoking and
forgotten candles are the most common fire-starters according to statistical data [33]. The
probability class is estimated to be D for one family barges and class C for apartment
barges10.
The consequences of a fire on a barge are the same as on land, plus the extra consequence of
sinking or capsizing during fire extinguishing. The free mass of the water used in the fire
extinction can eliminate the stability and make the barge capsize. The stability must be taken
into account, when estimating the probability for capsizing. There are obviously risks for
severe consequences, with people dead or injured, thus giving a consequence class of 4-5.
The building material, concrete or steel, in the apartment barges has a potential of reducing
the possibility of the fire to spread. The surrounding water limits the risk that the fire spreads
to adjacent barges even more. By good design and this in mind, it should be possible to
reduce the probabilities one step, for the one family barge to E and to the apartment barge to
D.
The need for early warning for the inhabitants should be emphasised: in only 25% of the
accidents with a mortal consequence, a working fire alarm was present.
If proper fire alarms and bilge pumps with sufficient capacity to handle even water from fire
extinguishing are installed, the consequences could be decreased to class 3-4.
10. Fire on board Probability 10.1 Probability 10.1 Consequences
Initially D C 4-5
With precautions E D 3-4

10
In Sweden the total number of households in owned houses are 1.6 millions (SCB 1990) and in apartments 2.1
millions (SCB 1990). The numbers of fires were approximately 3 000 each. (Räddningsverket). This giving the
probability of approximately 0.2 % for the one family barge and for a block of 20 apartments, the probability
will be approximately 3 % and class C.

50
Risk matrix – a summary
The risks with and without precautions are summarized in the table below and are then
illustrated in a risk matrix.
Table 7.4 The risks before and after precautions
Probability class Consequence class
With pre- With pre-
Risk Initially cautions. Initially cautions
1 Fall accidents caused by slippery footpaths A B 2-4 1-3
2 Health problems due to high moisture and wind - - - -
3 Taking in water by inlets, outlets or scuttles B D-E 3-5 3-5
4 Corrosion - - - -
5 Mould/water damages See 3 See 3 See 3 See 3
6 Collision D E 4-5 4-5
7 Broken mooring arrangement C-D D-E 3-4 3-4
8 Damages to the hull from ice pressure B-C B-C 1-2 1-2
9 Sewer leakage B-C B-C 2 1
10.1 Fire on board (one family barge) D E 4-5 3-4
10.2 Fire on board (apartment blocks) C D 4-5 3-4

Probability

HIGH
A R1
RISK

B R1 R3

R9 R9

C R
10.2

R7
R
D 10.2 R
10.1 R6

R7 R3
LOW
E R
R6
RISK 10.1

Consequence
1 2 3 4 5
Fig 7.5 The Risk matrix with the identified risks. The round symbol
indicates the risk after precautions.

51
8 Planning Apartment Barges - a Working Procedure
Planning the construction of a successful apartment barge or a cluster of barges can include
the following steps and considerations, and should be carried out through an iterative process
as described below.

Wind and Barge

wave data design

Damping measures
• Bilge keels • Adjust dimensions
• Mooring damping • Tanks
• Break waters • Etc.

Strip calculations

Evaluation of comfort

Are the
movements and Choice of precautions to improve the
accelerations behaviour of the barge
below the comfort No
criterias?

Yes

Continue
projecting

Fig 8.1 The iterative working procedure

Each step of the procedure is described more in detail below.

8.1 Studies of the location

First, the location of interest is studied regarding effective fetch length, wind data and water
depth. If wind shelter effects have an influence upon the wind at the location, reductions of
the wind speeds can be made.
By studying wind data at a location as close as possible to the mooring place in question, and
for a sufficient period of time, a good picture of the wind spectrum at that location can be
obtained.

52
8.2 Barge design
An initial design is chosen, based on the basic demands of the customer. The dimensions,
draught, distance from keel to center of gravity and radius of gyrations are calculated, along
with other barge characteristics. These will, together with wind and wave data, be needed in
the strip calculations performed in order to predict the barge’s behaviour.

8.3 Strip calculations

By performing strip calculations, the movements and accelerations can be calculated for the
chosen design of barge and different sea states. The strip programs can calculate many
different movements and accelerations, but the most important for barges are roll and the
accelerations caused by roll, pitch, and the other movements. The purpose of the strip
calculations is to find out if the expected motions and accelerations are acceptable, by
comparing them with comfort criteria. The strip programs can often automatically calculate
predictions of the Body Sickness Index (BSI), which is a measure of “how many percentage
of the passengers that will get seasick” as a function of sea state. The BSI indices can be used
as an alternative or a complement to comparing accelerations with comfort criteria.

BSI Index
The Body Sickness Index is derived using the methods described in British Standard and are
based upon measurements and data from mixed groups of passengers on ferries. [8] It takes
no account of the acclimatisation, which means that the calculated BSI value will be higher
than the value that the inhabitants experience after a couple of days on the barge. The BSI
value can be calculated at any specific position on the ship, but the value is valid only for that
spot. The position chosen in the calculated examples below is always the position with the
highest accelerations on the barge, and thus the highest BSI. Since the barge is moving both in
roll and pitch, the largest movements are expected to be found as far away as possible from
the roll and pitch centre, but sometimes amidships has had the highest accelerations.
When choosing an appropriate maximum value for BSI, the wind data and the target group
must be considered. For apartment barges, the acclimatisation should be taken into account.
In the examples in this report, the BSI is calculated for different barge designs and damping
measures, in order to compare them. However, no maximum limits have been set, since this
demands a more thorough analysis.

8.4 Evaluation of comfort from wind data

When the results from the strip calculations are evaluated, it is possible to evaluate the
expected comfort, by comparing the available wind data with the wind giving the maximum
acceptable movements and accelerations.
A level of acceptable discomfort should be selected. Then, the expected number of occasions
when the comfort limits are exceeded can be estimated. (An appropriate maximum value of
the number of uncomfortable occasions can be for example 5 times a year.)
For a more thorough analysis a statistical method can be used where the number of expected
moments of discomfort/year is estimated.
Using the fact that the wind follows the Rayleigh distribution and the accelerations from the
strip calculations, statistical analysis can be used to compare the result with the comfort
criteria in order to get the expected number of uncomfortable sea states.

53
According to statistical analysis, the probability that a value X is exceeded, if the values
follow the Rayleigh distribution, is:

If RH is 72 and X = which means

 X2 
14 m/s has been  14 2 
that in 6,6 % of
− 
 shown, in strip − 
 the time, the
 RH  72
P( x ≥ X ) = e 
calculations, to give P=e 
= 6,6% comfort limits
the maximum will be
tolerated exceeded
accelerations, then:

If the Rayleigh parameter (RH) is known, the formula for estimating how often there will be a
sea state, causing uncomfortable accelerations, can be used. The Rayleigh parameter is
defined as 8*σ2, where σ is the standard deviation of the recorded wind data. The periods of
time with considerable accelerations, but below the comfort criteria are also interesting and
need to be analysed.
It is also possible to calculate the highest expected wind speed, during a period of for example
twenty years. Since every sea state lasts for three hours, 20 years are composed of 58 400 sea
states. According to the formula below, the highest wind speed is expected to be 28 m/s.

U 1 / N = RH ⋅ ln N = 72 ⋅ ln(58400) = 28 m/s Where N = 58 400

In the next chapter, Example projects, these different ways of evaluating a barge’s behaviour
at defined locations will be illustrated.
If the result of the evaluation turns out to be satisfactory, the planning procedure can continue
with the chosen dimensions. Otherwise, if high roll angles or/and accelerations are expected
too often, damping measures will have to be considered in order to improve the behaviour of
the barge.

8.5 Damping Measures

Proper actions in order to reduce the roll angles and accelerations are, as mentioned earlier in
the report, bilge keels and tanks. As described in the chapters describing these damping
methods, they will not necessarily decrease the accelerations. In some cases, the
uncomfortable movements of the barge will even increase as a result of the extra device,
depending of the dimensions and characteristics of the barge.
As an initial measure, extra damping through bilge keels and tanks should be considered. If
the motions are still too large after new calculations and evaluations, further improvements of
the barge or adjustments to the mooring location must be made. Recommended methods are
mooring damping, wave breakers or changing the barge’s centre of gravity.
Also when the barge responses are expected to stay below the acceptable levels, damping
measures should be considered, especially cheap and easily accomplished ones. The over all
comfort of the barge will augment and thus present a better solution to the inhabitants.

54
By protecting the mooring location with breakwaters, both the roll angles and the
accelerations of the barge are always reduced. This is because it is always more efficient to
attack the source; the waves, instead of the symptoms, the movements caused by the waves.

8.6 Comments to the planning

After stepping through the iterative process described above, the designer has hopefully found
a functional design of the barge and can continue with the planning process. However, if a
suitable solution is not found, the reason might be that the fetch length is too great and the
chosen location is too windy. The location might not be unsuitable for comfortable apartment
barges. Breakwaters, a relatively expensive solution, might be needed in order to obtain a
sufficient reduction of the waves. If an acceptable technical and economically feasible
solution is not possible to find, another location will have to be chosen.

55
9 Example Projects
Three example projects are presented in this chapter. They are locations where there are either
existing house barges or plans to build apartment barges.
Bällstaviken. This location will be the most thoroughly investigated one. The influence
of the barge’s dimensions, mooring angle and different damping devices will be
investigated. The statistical approach will be used, as well as a more “common sense”
estimation, based on the BSI-output from the strip calculations together with wind data.
Pampas Marina. This location is investigated in a similar way. The number of
occasions when the comfort limits are exceeded will be estimated for one barge design,
with and without damping devices.
Karlstad. Since no wind data is available for this location, only a summary estimation
will be done, in order to check for the need of further investigations.

9.1 Bällstaviken
Bällstaviken is a location that has been studied for apartment barges since a couple of years. It
is an attractive site in the Stockholm area and permission to moor barges there is already
granted for several years ahead. The area is relatively well protected from waves and there is
no traffic with ships big enough to create swells that will provoke discomfort on the barges.

Fig 9.1 Map of Bällstaviken. Published with permission from the Swedish
Maritime Organisation “© Sjöfartsverket tillstånd nr 03-01370”

56
Effective fetch length - Bällstaviken
The largest waves hitting the barges will arise in heading 135 ° from the barges (North= 0 °),
through the relatively narrow water area. The effective fetch length is 572 m. For calculations
see Appendix B.

Wind analysis - Bällstaviken

The wind data used in this case has been collected out at sea, at Almagrund. The direction and
speed of the wind has been measured every third hour for two years. In this analysis, each
wind state will be assumed to last for three hours. The total number of observations is 5 840
and they follow a Rayleigh distribution, illustrated in Fig 11.2 below.

15%

10%

5%

0%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Wind speed (m/s)

Fig 9.2 Wind data out at sea for all directions – proportion of observations
for each wind speed.

In Bällstaviken, the wind directions generating waves are 135° +/- 45°. Therefore, the
observations in this interval, from 90°-180°, are extracted from the rest of the observations.
The highest waves are generated when the wind blows from 135°. This will give an
estimation of the waves that is higher than the real value, because most of the time the wind is
not blowing from exactly 135°, and therefore gives smaller waves. According to the
observation data, the probability for a wind direction of 90-180° is 21%. The distribution of
these observations is shown in Fig 9.3 below.

57
 15%

10%

5%

0%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Wind speed (m/s)

Fig 9.3 Wind data for directions 90-180 ° – proportion of observations for
each wind speed.

Another possible way of estimating the proportion of time with wind coming from 90°-180°,
is to use a wind rose. In this example, a rose measured at Bromma Airport during 1961-1998
was used, giving a value of 20% for wind directions of 90°-180°. This confirms the value
calculated above.

Fig 9.4 Wind rose from Bromma Airport

58
Since the observations almost follow the Rayleigh distribution, the Rayleigh parameter can be
used if the standard deviation σ is known. [14]. The standard deviation is calculated for the
wind data from Bällstaviken which is 70 % of the values out at sea. The standard deviationσ,
of all wind measurements, is calculated and is found to be 2.1.
RH = 8 ⋅ M 0 35 where M0 = the variance σ 2 =4.4
In these calculations it has been assumed that one wind state lasts for tree hours and this will
give 58 400 wind states during 20 years. Using statistic analysis and the Rayleigh parameter,
the highest expected wind speed during 20 years is estimated to be:
U 1 / N = RH ⋅ ln N = 35 ⋅ ln(58400) = 20 m/s where N = 58 400 and RH = 35
The Rayleigh parameter will also be used later in this example, in order to calculate the
proportion of time when the wind speed is within certain intervals and thus give rise to
different sea states.

Expected waves - Bällstaviken

When the fetch length is known, the generation of the waves is depending only on the wind
speed. The wind in the bay supposedly has decreased to 70% of the wind speed out at sea,
because of the increase in roughness, consisting of islands. The relation between wave heights
and mean periods for Bällstaviken is shown in the table below. The formulas used to calculate
the wave characteristics as a function of wind speed and fetch length are the ones described in
Ch 3. [22] The table below shows the correlation between wind speed, wave height and
angular velocity and this is founded on the assumption of effective fetch length of 570 m.
Table 9.1 Sea states for Bällstaviken; winds up to 33 m/s and effective fetch length 570 m.

Wind speed out at The Reduced Mean angular

sea wind speed Wave length Wave height velocity (rad/s)
3 m/s 2.1 m/s 0.75 m 0.03 m 9 rad/s
6 m/s 4.2 m/s 1.6 m 0.07 m 6.3 rad/s
9 m/s 6.3 m/s 2.5 m 0.11 m 5.0 rad/s
12 m/s 8.4 m/s 3.3 m 0.16 m 4.3 rad/s
15 m/s 10.5 m/s 4.2 m 0.21m 3.9 rad/s
18 m/s 12.6 m/s 5.0 m 0.26 m 3.5 rad/s
21 m/s 14.7 m/s 5.9m 0.31 m 3.2 rad/s
24 m/s 16.8 m/s 6.8 m 0.36 m 3.0 rad/s
27 m/s 18.9 m/s 7.6 m 0.41 m 2.8 rad/s
30 m/s 21.0 m/s 8.5 m 0.46 m 2.7 rad/s
33 m/s 23.0 m/s 9.3 m 0.52 m 2.6 rad/s

The highest wind speed taken into consideration is 33 m/s, since this is the highest recorded
one in the Stockholm area during the last 50 years, measured out at sea. This means that no
higher significant wave height than 0.52 m is likely to arise and this should be taken into
account when dimensioning the barges.

59
The barges – Bällstaviken
Three different barges bas been studied, all providing the same usable living area.
Barge 1 10x42 m
Barge 2 14x30 m
Barge 3 18x23 m
They have the same displacement, 715m3, draught = 1.7m and KG = 2.7 m, since this makes
it easier to compare their behaviour. The barges are planned to have three stores with
apartments, all with an own entrance from an external gallery. Each barge are planned to
consist of between 20-30 apartments.

Strip calculations – bare hull

Strip calculations have been done with the, for Bällstaviken, possible sea states. The roll
angles, vertical accelerations, and lateral accelerations have been calculated. The BSI-index
has also been calculated.
When nothing else is indicated, the measure point on each barge is amidships, two meters
above the keel and close to the side. When a measurement amidships is not generating the
highest accelerations, the x coordinate is changed in order to obtain the highest accelerations
onboard. The reason for choosing a measure point close to the side of the barge, is that the
accelerations will always be highest there, regardless the area and shape of the barge. The
reason for this is that the wider the barge, the longer the distance from the centreline. A longer
distance from the measure point to the centre will give accelerations, for an equal roll angle.
The sea states in the table below will be used in the strip calculations. Each sea state is a
function of the fetch length and a wind speed. All wind results are measured out at sea and
therefore reduced to 70%.
In the last column, the proportion of time when the wind speed is within each interval is
calculated. The sea state B1, with wind speeds up to 3.5 m/s, will for example occur in 30 %
of the time according to the statistical analysis:

 X2   3.52 
− 
 − 

 RH   35 
1 − P( x ≥ X ) = 1 − e = 1 − P( x ≥ 3.5) = 1 − e = 30%

Since the wind only generates “interesting” waves 21 % of the time, the proportion in the
table below will be approximately one fifth of the calculated value above.
Table 9.2 The different sea states with probabilities

Sea Local wind speed

state (reduced to 70%) Sign Hs Tm L Estimated proportion of time
B1 3.5 m/s 0.06 m 0.9 s 1.3 m P(0>x>3.5)= 6.3%
B2 7.0 m/s 0.13 m 1.3s 2.8 m P(3.5>x>7.0)= 9.7%
B3 11,5 m/s 0.21 m 1.6s 4.2 m P(7>x>11.5)= 4.7%
B4 14.0 m/s 0.29 m 1.9s 5.6 m P(11.5>x>14)= 0.4%
B5 17.5 m/s 0.37 m 2.1s 7.0 m P(14>x>17.5)=0.074%
B6 21.0 m/s 0.46 m 2.3s 8.5 m P(17.5>x>)=0.003%

60
In Bällstaviken it is possible to moor the barges so that the waves are coming from the
quarter, with an inclination angle of 60°. To study the differences in the barge’s movements
when moored so that the waves are coming from the side, calculations have been made for an
inclination angle of 90° as well.

Roll angles
Wave inclination angle 60° Wave inclination angle 90°
Degrees (RMS) Degrees (RMS)
10 10
Barge 10x42 Barge 10x42
Barge 14x30 Barge 14x30
8 Barge 18x23 8 Barge 18x23
Comfort Comfort

6 6

4 4

2 2

0 0
B1 B2 B3 B4 B5 B6 B1 B2 B3 B4 B5 B6

Fig.9.5 Calculated roll angles for three barges in Bällstaviken

60 deg: The roll angles are reaching the limit of comfort in B6, but might be accepted
since this sea state very seldom occurs. However it should be noted that a RMS
value of 2 ° means that 5-6 ° can be expected, creating potential damages to the
interior of the barge.
90 deg: With the waves coming right from the side, the roll angles get very high and
cannot be accepted.

61
Vertical accelerations
Wave inclination angle 60 ° Wave inclination angle 90 °
2 2
m/s (RMS) m/s (RMS)
3 3
Barge 10x42 Barge 10x42
Barge 14x30 Barge 14x30
Barge18x23 Barge 18x23
Comfort Comfort
2 2

1 1

0 0
B1 B2 B3 B4 B5 B6 B1 B2 B3 B4 B5 B6

Fig 9.6 Calculated vertical accelerations for three barges in Bällstaviken

60 deg: The comfort criteria for vertical accelerations are exceeded in B4-B6. The fact that
the widest barge has the highest vertical accelerations can be partially explained by
the longer distance from the measure point to the centreline and that the velocity
component induced by the pitch is increasing, when the length of the barge gets
shorter.
90 deg: For this wave inclination, the accelerations are even higher and cannot be accepted
during any circumstances. Even a wind as low as 12 m/s will cause too high
accelerations.

The fact that barges this wide will move at all is surprising. The results of the strip
calculations contradict the opinion of PNA, which claims that small accelerations and
movement are to be expected if the barge is wider than two times the wavelength. This
confirms the difficulties in predicting roll and accelerations, without full-scale tests.

62
Lateral accelerations
Wave inclination angle 60 ° Wave inclination angle 90 °
m/s2 (RMS) m/s2 (RMS)
0,5 0,5
Barge 10x42 Barge 10x42
Barge 14x30 Barge 14x30
0,4 Barge18x23 0,4 Barge 18x23
Comfort Comfort

0,3 0,3

0,2 0,2

0,1 0,1

0 0
B1 B2 B3 B4 B5 B6 B1 B2 B3 B4 B5 B6

Fig 9.7 Calculated lateral accelerations for three barges in Bällstaviken

60 deg: The lateral accelerations are moderate and below our comfort criteria for all sea
states and barges.
90 deg: The comfort criterion is exceeded for B6.

BSI
Wave inclination angle 60 ° Wave inclination angle 90 °
BSI BSI
80% 80%
Barge 10x42 (x=20) Barge 10x42
Barge 14x30 (x=15) Barge 14x30
Barge 18x23 (x=11.5) Barge 18x23
60% 60%

40% 40%

20% 20%

0% 0%
B1 B2 B3 B4 B5 B6 B1 B2 B3 B4 B5 B6

Fig 9.8 Calculated BSI for three barges in Bällstaviken

63
60 deg: The calculated BSI value is quite high, between 15 and 25% for B6, but since the
acclimatisation is not taken into account it might be acceptable.
90 deg: The BSI-index is as high as 20 % at B3 for the wider barges, which must be
regarded as too high.

When the results from the strip calculations above are analysed, it is clear that it is the levels
of vertical accelerations that will be limiting. The vertical accelerations exceed the comfort
criterion 0.2 m/s2 at different sea states for the different barges. In the table below, statistic
analysis is used in order to estimate how often uncomfortable sea states will occur for the
different barges.
Table 9.3 Statistic analysis of the different barges in Bällstaviken
Wave Sea states when Proportion of time
inclination comfort criterion is when comfort criterion Expected number of
Barge size angle exceeded is exceeded occasions per year*
60 B5, B6 0.08% 2.3
10x42 m
90 B3, B4, B5, B6 5.2% 151
60 B4, B5, B6 0.5% 14
14x30 m
90 B3, B4, B5, B6 5.2% >151
60 B3, B4, B5, B6 5.2% 151
18x23 m
90 B3, B4, B5, B6 5.2% >151
* One year consists of 2 920 occasions and each occasion corresponds to one three-hour period.

As shown in the table, it is very seldom that the narrowest barge will experience comfort
problems (if moored with a 60 inclination angle), and the behaviour of that barge without
additional damping measures might be acceptable. However, the other two barges seem to
need extra damping, especially when moored with a 90 inclination angle.
Bilge keels are usually the easiest and cheapest method and they will be evaluated next.
Calculations have been made for the same barges, with bilge keels.

64
Strip calculations – bilge keels
The same strip-calculations have been done for the same model barges but equipped with
bilge keels with width 0.5 m.

Roll angles
Wave inclination angle 60 ° Wave inclination angle 90 °
Degrees (RMS) Degrees (RMS)
3 3
Barge 10x42 Barge 10x42
Barge 14x30 Barge 14x30
Barge 18x23 Barge 18x23
Comfort Comfort
2 2

1 1

0 0
B1 B2 B3 B4 B5 B6 B1 B2 B3 B4 B5 B6

Fig 9.9 Calculated roll angles for three barges with bilge keels (0.5m) in
Bällstaviken

60 deg: The roll angles have decreased to 15-30 % of what they were without bilge keels
and the roll is well below our comfort criteria.
90 deg: The roll angles in case with waves coming from 90 ° are still much higher, but
they are now below the comfort limit for all sea states.

65
Vertical accelerations
Wave inclination angle 60 ° Wave inclination angle 90 °
2
m/s (RMS) m/s2 (RMS)
0,8 0,8
Barge 10x42 Barge 10x42
(x=20)
Barge 14x30 Barge 14x30
Barge18x23
0,6 0,6 Barge 18x23
Comfort
C f

0,4 0,4

0,2 0,2

0 0
B1 B2 B3 B4 B5 B6 B1 B2 B3 B4 B5 B6

Fig 9.10 Calculated vertical accelerations for three barges with bilge keels
(0.5m) in Bällstaviken

60 deg: The vertical accelerations have decreased and are now below our strict comfort
criteria for all sea states, but for the widest barge, vertical accelerations are close to
the limit for discomfort in B6.
90 deg: Even with bilge keels, it is only for the narrow barge that the vertical accelerations
are below the tolerable. The fact that the two other barges get almost the same
accelerations is strange, but that was the result from the strip calculations.

66
Lateral accelerations
Wave inclination angle 60 ° Wave inclination angle 90 °
2 2
m/s (RMS) m/s (RMS)
0,5 0,5
Barge 10x42 Barge 10x42
Barge 14x30 Barge 14x30
0,4 Barge18x23 0,4 Barge 18x23
Comfort Comfort

0,3 0,3

0,2 0,2

0,1 0,1

0 0
B1 B2 B3 B4 B5 B6 B1 B2 B3 B4 B5 B6

Fig 9.11 Calculated lateral accelerations for three barges with bilge keels
(0.5m) in Bällstaviken

60 deg: According to the strip calculations, only a small improvement in lateral

accelerations due to the bilge keels can be seen, but the lateral acceleration is low
for all the barges in all sea states.
90 deg: The lateral accelerations have been reduced and are now below the comfort
criterion for all sea states and all barges. For sea state B6 the RMS value has
almost been halved.

67
BSI
Wave inclination angle 60 ° Wave inclination angle 90 °
BSI BSI
20% 20%
Barge 10x42 (x=20) Barge 10x42 (x=20)
Barge 14x30 (x=15) Barge 14x30
Barge 18x23 Barge 18x23
15% 15%

10% 10%

5% 5%

0% 0%
B1 B2 B3 B4 B5 B6 B1 B2 B3 B4 B5 B6

Fig 9.12 Calculated BSI Index for three barges with bilge keels (0.5m) in
Bällstaviken

60 deg: From the calculated BSI value it is seen that the probability of the inhabitants
getting seasick is very low, especially if the acclimatisation is taken into account.
90 deg: The narrow barge is showing a low BSI value, but the wider barges have high
predictions, even from B3 and up.

Equipping the barges with such a simple device as bilge keels, the accelerations and roll
angles can be decreased enough to achieve a comfortable behaviour in all sea states and for all
of the investigated barges when mooring them with a wave inclination angle of 60 °. There
will be no problem fulfilling the comfort demands. However, the wider barges moored with a
90° inclination angle still have problems with the vertical accelerations. Statistic analysis is
again used in order to estimate how often uncomfortable sea states will occur. The result is
presented in the table below and should be compared with table 9.3 for an evaluation of the
effect of bilge keels.

68
Table 9.4 Statistic analysis for expected uncomfortable occasions
Wave Sea states when Proportion of time
inclination comfort criterion is when comfort criterion Expected number of
Barge size angle exceeded is exceeded occasions per year*
60 --- 0 0
10x42 m
90 --- 0 0
60 --- 0 0
14x30 m
90 B3, B4, B5, B6 5.2% 151
60 --- 0 0
18x23 m
90 B3, B4, B5, B6 5.2% 151
* One year consists of 2 920 occasions and each occasion corresponds to one three-hour period.
If the barges are moored 60° towards the waves the qualities of barges equipped with bilge
keels seem to be superior all the time and such barges will probably be a more comfortable
and popular place to live on, even though the extra damping is not needed to get below the
comfort criteria.
If the 90° mooring will be used, the wider barges will require further damping measures to be
taken. Therefore, strip calculations will next be performed for a barge of size 14x30m
equipped with both tanks and bilge keels.

Strip calculations – bilge keels and tanks

The result from the strip calculations below shows the effectiveness of bilge keels and passive
tanks, both separately and in combination, and to what extent they limit the roll angles and the
accelerations. The wave inclination angle is 90°.
Since the lateral accelerations are below the comfort criteria, for all barges except the bare
hull in sea state 6, only the roll angles and the vertical accelerations have been displayed in
the diagrams below.
Roll angles Vertical acceleration
Degrees (RMS) m/s2 (RMS)
6 3
Bare hull Bare hull
Bilge Bilge
Tank Tank
Bilge+Tank Bilge+Tank
Comfort Comfort
4 2

2 1

0 0
B1 B2 B3 B4 B5 B6 B1 B2 B3 B4 B5 B6

Fig 9.13 Roll and vertical acceleration for four different versions of the 14x30
m barge, with a wave inclination of 90 ° in Bällstaviken.

69
Only tanks will reduce both the roll angle and the vertical acceleration, but to a less
significant extent than bilge keels. However, a combination of both devices seems to be a
good solution. Unfortunately not good enough, since the comfort criterion for vertical
acceleration is still exceeded for sea states B5 and B6.
The high stability of the barge reduces the efficiency of the tank and the difficulties in
designing the parameter of tanks, might contribute to their scanty result.

Summary – Bällstaviken
Without additional damping, the barge’s low roll damping will give high roll angles and the
high stability gives high accelerations, even in a protected area as Bällstaviken.
However, with bilge keels, the roll motions are effectively reduced and so are the
accelerations. An even better behaviour of the barge is obtained with a combination of a
passive tank and bilge keels.
The great importance of wave inclination is also shown in the example. The movements and
accelerations were often twice as high for 90° as for 60°, implying that the mooring angle is
also very important to consider.
If an unfavourable mooring angle is the only possible choice, wave breakers should be
considered.
As a conclusion, mooring the barges 60° and using bilge keels must be considered very
comfortable and well-designed solution.

70
9.2 Pampas Marina
Pampas Marina is situated in the centre of Stockholm near Ulvsunda in the lake Mälaren. In
combination with a regular yachting marina, it is also the home for 8-10 house barges. The
Marina has produced most of these barges, but some of them are constructed by the people
living there. All necessary installations and mooring arrangements are taken care of by the
Marina. The barges produced by the Marina are sold as a tenant-ownership. The inhabitants
pay a fee as in an ordinary tenant-owner’s association. The barges are 6x14 m, approximately
130m2 living area, with a villa feeling. As a result of the smaller size, and hence smaller beam,
of these one family barges, the stability will be much lower than for larger apartment barges.
The roll angles and accelerations will be calculated below.

Fig 9.14 Map of the Ulvsunda Lake and Pampas Marina Published with
permission from the Swedish Maritime Organisation“©
Sjöfartsverket tillstånd nr 03-01370”

Effective fetch length - Pampas

As can be seen on the map of the area, the wind direction generating the highest waves is
approximately 290°. Since it is a wide water area, winds coming from 290°+/-48° are
assumed to contribute to the wave generation. The distances are measured on the map for
every 6°. There is always an uncertainty for every particular distance, for example of how big
the reduction should be for possible shelter effects. The effective fetch length for Pampas
Marina is estimated to 985m. See Appendix B for calculations.
The effective fetch length is longer in Pampas Marina, than in Bällstaviken. This will give a
little higher significant wave heights and longer waves.

71
Wind analysis – Pampas
Since it has been shown that the worst wind direction from a wave making point of view is
290°, WNW, these observations are extracted from the wind data measured out at sea. In 25%
of the time, the wind will blow from a direction 290°+/-45°. The wind distribution is shown in
the diagram below. During the two-year measuring period, the highest recorded wind speed
was 18 m/s. The reason that it is lower than in Bällstaviken is probably the higher roughness
over land in the direction from W. Since the wind coming from West travels over land before
it is measured, the wind speeds measured out at sea are only reduced with 10 % in this case.

15%

10%

5%

0%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Wind speed (m/s)

Fig 9.15 The wind distribution extracted for wind directions 245-335 °

The observations almost follow the Rayleigh distribution and the Rayleigh parameter is
calculated below. The standard deviation σ is known and is found to be 2.55.
RH = 8 ⋅ M 0 = 52 where M0 = the variance σ 2 =6.5
In these calculations it has been assumed that one wind state lasts for tree hours and this will
give 58 400 wind states during 20 years. Using statistic analysis and the Rayleigh parameter,
the highest expected wind speed during 20 years is estimated to be:
U 1 / N = RH ⋅ ln N 52 ⋅ ln(58400) = 24 m/s where N = 58 400 and RH = 52

Expected waves – Pampas

When the fetch length is known, the generation of the waves is depending only on the wind
speed. The wind in the bay supposedly has decreased to 90% of the wind speed out at sea.
The relation between wave heights and mean periods for Pampas is shown in the table 11.5
below. The formulas used to calculate the wave characteristics as a function of wind speed
and fetch length are the ones described in Ch 3.

The Barge – Pampas

The barge used in this example is the Aqua Villa Generation III with the dimensions 6.1 x 14
m and three storeys high. The total living area for this Villa is 132 m2. The lower floor is used
for bedrooms and bathrooms and is situated mostly below the water surface, but with the

72
scuttles close to the ceiling. On the first floor, the kitchen and living room is situated as well
as the entrance. The second floor is intended for social activities or as master’s bedroom.
There are three different models of the villas but they are all built on the similar pontoons.

Fig. 9.16 The Aqua villa Futura

The barge characteristics are:

B x L x T = 6,3 x 14,1 x 1,77
Displacement 151 m3
KG 2,04 m
In order to get improvement of the passive tank it has to be installed high up on the barge, this
makes it less suitable for barges, this small.

Strip calculations
Roll angles, vertical acceleration and lateral acceleration will be calculated with the Wolfson
strip program.
The sea states in the table below will be used in the strip calculations. Each sea state is a
function of the fetch length and a wind speed. All wind results are measured out at sea and
therefore reduced to 70%.
In the last column, the proportion of time when the wind speed is within each interval is
calculated. The sea state P1, with wind speeds up to 3.5 m/s, will for example occur in 21 %
of the time, according to the statistic analysis:
Since the wind only generates “interesting” waves 25 % of the time, the proportion in the
table below will be approximately one fourth of the calculated value above.

73
Table 9.5 The different sea states with probabilities

Sea Local wind speed

state (reduced to 90%) Sign Hs Tm L Expected proportion of time
P1 3.5 m/s 0.07 m 1.0 s 1.7 m P(0>x>3.5) = 5.2 %
P2 7 m/s 0.16 m 1.5 s 3.6 m P(3.5>x>7.0) = 10.0 %
P3 8.8 m/s 0.21 m 1.7 s 4.5 m P(7>x>8.75) = 4.0 %
P4 10.5 m/s 0.26 m 1.9 s 5.4 m P(8.75>x>10.5) = 2.7 %
P5 12.3 m/s 0.31 m 2.0 s 6.4 m P(10.5>x>12.25)=1.6 %
P6 14 m/s 0.36 m 2.2 s 7.3 m P(12.25>x>14) = 0.8 %
P7 15.8 m/s 0.42 m 2.3 s 8.3 m P(14>x>15.75) = 0.4 %
P8 17.5 m/s 0.47 m 2.4 s 9.2 m P(15.75>x>17.5) = 0.14%
P9 21 m/s 0.58 m 2.7 s 11.1 m P(17.5>x) = 0.07%

Calculations have been performed for four barges, with the same dimensions but with
different damping devices. In the table below, the different strip calculation scenarios with
barge data and damping measures are summarised.
Table 9.6 The damping devices on the barge Generation III in Pampas Marina
Damping device characteristics KG
Bare hull No additional damping 2.04 m
Bilge keels Two bilge keels
2.04 m
L=10 m and H =0.5 m
Tank One passive U-tank
Dimensions: B=3.5m, H=1m, b=1m, h=0.5m
2.24 m
Volume: 3.75 m3
Centre of gravity: z=4 m
Tank + Bilge keels See above 2.24 m

In Pampas, the barges are moored on different sides of a jetty, which means that the wave
inclination angle will also differ. The results from strip calculations made for 30°and 90° will
be presented in the diagrams below.

74
Roll angles
Wave inclination angle 30 ° Wave inclination angle 90 °
Degrees (RMS) Degrees (RMS)
3 3
Bare hull Bare hull
Bilge keels Bilge keels
Tank Tank
Bilge+Tank Bilge+Tank
Comfort Comfort
2 2

1 1

0 0
P1 P2 P3 P4 P5 P6 P7 P8 P9 P1 P2 P3 P4 P5 P6 P7 P8 P9

Fig 9.17 Calculated roll angles for four barges in Pampas Marina

30 deg: The roll angles are very small for all barges
90 deg: The roll angles are well below the comfort criteria even for a bare hull, but the fact
that the simple precaution of installing bilge keels or a tank reduces significantly.
Please note that the largest expected roll response will be 3-4 * RMS and are
therefore not negligible.

75
Vertical accelerations
Wave inclination angle 30 ° Wave inclination angle 90 °
2 2
m/s (RMS) m/s (RMS)
0,6 0,6
Bare hull (x=6) Bare hull (x=6)
Bilge keels Bilge keels (x=6)
Tank Tank
Bilge+Tank (x=6) Bilge+Tank
Comfort Comfort
0,4 0,4

0,2 0,2

0 0
P1 P2 P3 P4 P5 P6 P7 P8 P9 P1 P2 P3 P4 P5 P6 P7 P8 P9

Fig 9.18 Calculated vertical accelerations for four barges in Pampas Marina

30 deg: The vertical accelerations are lowest for the barge with a tank. For the barges with
bilge keels, wind speeds over 16 m/s will give rise to uncomfortable vertical
accelerations.
90 deg: The tank reduces the vertical acceleration efficiently and so do bilge keels.

76
Lateral accelerations
Wave inclination angle 30 ° Wave inclination angle 90 °
2 2
m/s (RMS) m/s (RMS)
0,5 0,5
Bare hull (x=6) Bare hull (x=6)
Bilge keels (x=6) Bilge keels
Tank Tank
0,4 0,4
Bilge+Tank (x=6) Bilge+Tank (x=6)
Comfort Comfort

0,3 0,3

0,2 0,2

0,1 0,1

0 0
P1 P2 P3 P4 P5 P6 P7 P8 P9 P1 P2 P3 P4 P5 P6 P7 P8 P9

Fig 9.19 Calculated lateral accelerations for four barges in Pampas Marina

30 deg: The lateral accelerations are acceptable for all sea states
90 deg: For the barges with bilge keels and a tank, the lateral accelerations are acceptable
for all sea states and for the bare hull the comfort criterion is exceeded for P8 or
more exactly wind speeds above 17 m/s.

77
BSI
Wave inclination angle 30 ° Wave inclination angle 90 °
BSI BSI
20% 20%
Bare hull (x=6) Bare hull
Bilge keels Bilge keels
Tank Tank
Bilge+Tank Bilge+Tank
15% 15%

10% 10%

5% 5%

0% 0%
P1 P2 P3 P4 P5 P6 P7 P8 P9 P1 P2 P3 P4 P5 P6 P7 P8 P9

Fig 9.20 Calculated BSI for four barges in Pampas Marina, measure point: (x,
y, z) = (6,3,4.5)

The highest BSI was found for waves coming from 30°. The reason is probably the relatively
high pitch motion compared with the roll motions when the wave inclination angle is 30°.
The increase in BSI value for the barges equipped with anti rolling devices is a result of the
amplified accelerations caused by the additional damping. Despite this, the total comfort of
barges equipped with anti-rolling devices is probably better thanks to the reduced roll angles.
By comparing the wind speed giving the highest acceptable accelerations and roll angles, with
the expected wind speeds, it is possible to estimate how often these will be exceeded. A total
time in percentage, when the barge has higher accelerations than the comfort criterion is
summarized and the expected number of times/year is calculated.

78
Table 9.6 Statistic analysis of expected occasions with discomfort
Wave Sea states when Proportion of time
Barge inclination comfort criterion is when comfort criterion Expected number of
version angle exceeded is exceeded occasions per year*
30 P8-P9 0.21% 6
Bare hull
90 P8-P9 0.21% 6
30 P7-P9 0.61% 18
Bilge keels
90 P9 0.07% 2
30 P9 0.07% 2
Tank
90 P9 0.07% 2
30 P7-P9 0.61% 18
Tank + Bilge
90 - 0% 0
* One year consists of 2 920 occasions and each occasion corresponds to one three-hour period.

The expected value is an estimation giving a hint of the likely behaviour and cannot be used
as a guarantee.

Summary Pampas
In view of the fact that the one family barges at Pampas Marina are relatively small, their
behaviour is quite good. The low stability gives low accelerations and hence improves the
barges’ qualities. However, a low stability enhances the static roll angles.
It should be remembered that the greatest expected value is up to 3-4 times the RMS value.
This means that roll angles up to 4° can be expected for the bare hull. According to people
living on bare hull barges at Pampas Marina, it happens, on extremely windy occasions that
books fall out of the shelves, which is likely to occur for roll angles of 5-6°. This fact
confirms the result of the strip calculations and adds to making them credible.
There have been problems mostly with high static roll angles, when the barge gets a list by
the wind pressure. The people interviewed did not get seasick from the roll and the
accelerations. The reason is probably that they like and are used to the sea-induced
movements and that they have had time to acclimatise. Another reason might be that the
barges are giving shelter to each other and thereby the approaching wind and waves are
reduced.
Strip calculations show that simple damping devices like bilge keels reduce the roll angles but
increase the accelerations. As seen in the diagrams, tanks give a very efficient damping for
small barges but there are problems finding a place for the tank and the tank will reduce the
stability even more.
Since the barges at Pampas are so small, the mooring angle has not the same clear-cut
influence on the movements as for the larger barges in the Bällstaviken example. The results
from the strip calculations are more reliable for inclinations of 90°, since the B/L ratio does
not have any effect of the result.

79
9.3 Mariebergsviken in Karlstad
Plans for housing on water in Karlstad were evaluated by three students in their M. of Sc.
Thesis. The idea was to get an environmentally friendly way of living, using the water areas
in the centre of town. Their idea is to connect nine pontoons and build several small houses on
it, see picture below.
The basis for their work is more of a civil engineer character than a naval architecture
approach. They assume that wind-generated waves can be neglected, despite a fetch length of
approximately 1 000 m. Consequently, there has been no analysis of the wave induced
movements, which seems risky, since waves of considerable wave height and wavelength will
probably affect the barge.

Fig 9.21 The planned barge with houses

The water depth on the chosen location is limited, which will lead to a considerable risk of the
barge hitting the seabed, when windy and low water levels. Of course this problem can be
solved by dredging on the location and thereby increase the water depth.
In this report, no analysis of the accelerations and roll angles for the Karlstad project has been
performed, but the wave heights will be estimated: With the assumption made in the report of
A fetch length of approximately 1000 m gives a wave height Hs = 0,36 m for a wind speed of
14 m/s.
This is likely to happen several times every year, which shows that there is a need for a more
thorough analysis of both the location and barge characteristics.

80
10 Conclusions
The conclusions drawn in ship literature and according to common beliefs, is that the wider
the barge the smaller the movements.
The main conclusion of the strip calculations in this report is that narrow barges are more
comfortable than wider barges. The narrow barges have lower accelerations but higher roll
angles than the wider ones, but equipping the barge with, for example bilge keels easily
reduce the roll. External tanks have not been able to test in the strip calculations, but if they
can be integrated in the gangway system, a good technical solution is achieved.
The damping measures studied are all working in different ways with their own advantages
and disadvantages.
Damping
device Advantages Disadvantages
Bilge keels A simple and efficient way to reduce the roll Sometimes gives an increase in
angles. accelerations.
Anti-roll tank Reduces the accelerations efficiently, if the Reduces the stability and takes
barge is not too stable. valuable space inside the barge.
Breakwaters Reduce both accelerations and roll angles, Expensive and not always possible.
since they reduce the waves.
Seaflex Reduce the roll angles and keeps the barge A probably increase the accelerations
in position in the same time. and very careful installation is
needed.

Building barges wide enough not to produce any responses at all, in protected areas, must be
possible. Full-scale tests are probably needed in order to estimate the beam required.
If measures are taken to reduce the potential high risks involved for apartment barges, living
like this can be assumed to be safe. Depending on location the highest risks use to be frozen
inlets or collision since their consequence is severe.

81
11 References
1. Principles of Naval Architecture
The Society of Naval Architects and Marine Engineers
1989-Second Revision
2. Per Vallander
Vågor i skärgårdsvatten
1977
3 Beach Erosion Board, Dept of the Army, Corps of Engineering.
1962
4 Madelaine Nordqvist
Study of non-linear models of ship roll damping- evaluation from roll decay
tests, KTH
1998
5 O.M. Faltinsen, Cambridge ocean technology
Sea loads on ships and offshore structures
1995
6 Yoshiho Ikeda, Osaka Prefecture University
Prediction methods of roll damping of ships and their application to determine
optimum stabilization device
2002
7 A.R.J.M Lloyd
Sea keeping: Ship Behaviour in Rough Weather
1989
8 Wolfson Unit, For Marine Technology & Industrial Aerodynamics
University of Southampton
2000
9 Elisabeth Ahlenius, Väg och Vattenbyggaren
Om lönsam och effektiv riskhantering
1.1999
10 H SystSäk, Försvarsmaktens handbok för Systemsäkerhet
1996
11 Ingemar Carlen, FOI-R-0109-SE
Våglaster på fundament till vindkraftverksystemdynamik och utmattning
2001
12 J.D Walters
Motion Sickness, Medicine and Travel. Documenta Geigy
1964
13 Anca Daniella Hansen
PH.D Thesis at DTU
1996
14 Anders Rosen
Föreläsningsanteckningar I sjöegenskaper, KTH
2001
15 Michael Kasten
Roll Attenuation Strategies
2002
16 S.A.Ragab, A.H.Nayfeh et Al
Roll Stabilization by Passive Anti-Roll Tanks
1998

82
17 R. Silvester
Coastal Engineering 1
1974
18 O.M. Faltinsen and R. Zhao, Journal of Ship Research No2 pp 93-106
Slow-Drift Motions of a Moored Two-Dimensional Body in Irregular Waves
1989
19 J.H Nieuwenhuilsen
Experimental investigations on Seasickness
1958
20 Mikael Palmqvist
Some model test analysis and comparative strip theory calculations, KTH
1996
21 Kristoffer Reinius
Förbättringar av standardpontoners vågdämpande funktion samt förankring av
pontonerna, KTH
1978
22 Bretschneider
Prediction of waves and currents. Look Lab Hawaii,
1973
23 H. Bergström
Wind energy meteorology, Uppsala University
2002
24 K. Magnusson SSPA
Riskbedömning B/S Sjöbussen
2002
25 L.Bergdal
Islaster mot vindkraftverk till havs
2002
26 Bali’c och Moberg
Bostäder på vattnet, Karlstad universitet
2002
27 Ian R. Young
Source Term Balance for Finite depth Wind Waves, Adelaide University
2000
28 Seaflex , www.anchro.se
29 Karlsson
Housing on water, Karlstad Universitet
2002
30 FS1998:38
31 Avances in Marine Technology, Proceedings Vol 2
1979
32 M.Huss
Vågor och gensvar KTH
1983
33 SCB Statistik over brandorsaker in bostadshus

83
 APPENDIX A

ROLL RESPONSES FOR DIFFERENT b/l RATIOS

Roll responses for two barges with the same waves and displacement/meter. The first barge
has dimension 14 m x 14 m and the other is 14 m x 70 m. There is no difference in roll
response when the waves are coming from 90°.

Barge 14m x 14m

1
 APPENDIX A

Barge 14m x 70m

2
 APPENDIX B

FETCH CALCULATIONS FOR THREE EXAMPLES

The effective fetch length calculated first as an example on a hypothetic location from
Vallander. [2] and after that for the two examples studied in this report.

1
 APPENDIX B

Bällstaviken

Angle Cos (α) Length (m) Length* (cos(α))2

30 0,87 250 188
24 0,91 290 242
18 0,95 325 294
12 0,98 463 443
6 0,99 563 557
0 1 2520 2520
6 0,99 650 643
12 0,98 400 383
18 0,95 400 362
24 0,91 210 175
30 0,87 200 150

∑ cos(α n ) = 10.4 and ∑ [Length(α n ]

) ⋅ cos(α n ) 2 = 5956
α n αn

∑ Length (α i ) ⋅ cos 2 (α i )
Effecive fetch length Fe = i =1
n
= 572 m
∑ cos( α
i =1
i )

2
 APPENDIX B

Pampas Marina
Table 11.6 The effective fetch length in Pampas Marina

Angle cos (α) Length (m) Length* (cos(α))2

48° 0.67* 0 0
42° 0.74* 0 0
36° 0.81* 0 0
30° 0.87 1492 1119
24° 0.91 1616 1349
18° 0.95 2113 1911
12° 0.98 1119 1070
6° 0.99 994 984
0° 1.00 808 808
6° 0.99 746 738
12° 0.98 746 714
18° 0.95 808 731
24° 0.91 932 778
30° 0.87 994 746
36° 0.81 1119 733
42° 0.74 870 481
48° 0.67 622 279

* These values are not used in the calculations, since the land area renders wave generation impossible in these
directions.

∑ cos(α n ) = 12.6 and ∑ [Length(α n ]

) ⋅ cos(α n ) 2 = 12 440
αn αn

∑ Length (α i ) ⋅ cos 2 (α i )
Effecive fetch length Fe = i =1
n
= 985 m
∑ cos( α
i =1
i )