Dynamic Positioning Simulator

Dynamic Positioning Simulator
Interim Report
Jalitha Wills
1104551
j.d.wills@student.tudelft.nl
Delft - Rotterdam, 2007
CONTENTS i
Contents
Preface vi
Used symbols and units vii
1 Introduction 1
2 Basics about Ships 3
2.1 Moves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 Types of Ships . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.3 Thrusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.3.1 Main propellers . . . . . . . . . . . . . . . . . . . . . . . . 6
2.3.2 Tunnel thrusters . . . . . . . . . . . . . . . . . . . . . . . 6
2.3.3 Azimuth thrusters . . . . . . . . . . . . . . . . . . . . . . 7
3 Dynamic Positioning 8
3.1 Denitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.2 Why Dynamic Positioning? . . . . . . . . . . . . . . . . . . . . . 8
3.2.1 Dynamic Positioning training . . . . . . . . . . . . . . . . 9
3.2.2 Comparison of Dynamic Positioning to other systems . . 9
3.2.3 Operating with DP . . . . . . . . . . . . . . . . . . . . . . 10
3.2.4 Applications where DP can be used . . . . . . . . . . . . 11
3.3 Dynamic Positioning System . . . . . . . . . . . . . . . . . . . . 12
3.3.1 Mathematical model . . . . . . . . . . . . . . . . . . . . . 13
3.3.2 Position Reference Systems . . . . . . . . . . . . . . . . . 13
4 Overview of Ship simulator 15
J.D.Wills j.d.wills@student.tudelft.nl
CONTENTS ii
5 Hydromechanics 25
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.2 Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.3 Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.4 Ship movements . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
5.5 Thrusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
5.5.1 Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6 Modeling 37
6.1 Power Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
6.1.1 Without DP . . . . . . . . . . . . . . . . . . . . . . . . . 37
6.1.2 With DP . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
6.2 Programming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
6.2.1 A rst model . . . . . . . . . . . . . . . . . . . . . . . . . 39
6.2.2 Method to nd the new conguration . . . . . . . . . . . 40
6.2.3 Minimization methods . . . . . . . . . . . . . . . . . . . . 43
7 Results 47
8 Evaluation and Future Goals 52
8.1 Evaluation of rst three months . . . . . . . . . . . . . . . . . . . 52
8.2 Future Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
A Figures 55
B Testproblem 60
LIST OF FIGURES iii
List of Figures
2.1 Coordinate system of a ship . . . . . . . . . . . . . . . . . . . . . 4
3.1 Scheme of Dynamic Positioning System . . . . . . . . . . . . . . 12
4.1 Ship editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.2 Harbour patrol boat . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.3 Hovercraft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.4 Ship approximated by a block. . . . . . . . . . . . . . . . . . . . 18
4.5 The eect of the mass of a ship. . . . . . . . . . . . . . . . . . . . 19
4.6 Curve of resistance. . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.7 Curve of thrust . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.8 Curve of thrust response. . . . . . . . . . . . . . . . . . . . . . . 21
4.9 Curve of liftforce of yawing ship. . . . . . . . . . . . . . . . . . . 21
4.10 Curve of sideforce. . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.11 Ship in editor sailing forward . . . . . . . . . . . . . . . . . . . . 23
4.12 Ship in editor sailing sideward . . . . . . . . . . . . . . . . . . . . 24
5.1 Drag and lift force . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.2 Transverse and Longitudinal projection. . . . . . . . . . . . . . . 30
5.3 Coordinate systems . . . . . . . . . . . . . . . . . . . . . . . . . . 32
5.4 Wave encounter . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
5.5 Thrust and Torque of a propeller. . . . . . . . . . . . . . . . . . . 34
6.1 Coordinate system . . . . . . . . . . . . . . . . . . . . . . . . . . 39
7.1 Conguration with 3 thrusters . . . . . . . . . . . . . . . . . . . 48
LIST OF FIGURES iv
A.1 Conning display . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
A.4 Ship terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
A.5 Importance of the dierent external inuences in specic sailing
conditions and applications. . . . . . . . . . . . . . . . . . . . . . 58
A.6 Example of Mathematical Model . . . . . . . . . . . . . . . . . . 59
B.1 Optimal solution two thrusters . . . . . . . . . . . . . . . . . . . 62
LIST OF TABLES v
List of Tables
1 Introduction on (Physical) Quantities and their units: Basic Units. vii
2 Introduction on (Physical) Quantities and their units: Special
Quantities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
3 Introduction on (Physical) Quantities and their units: Other Units. viii
7.1 Number of thrusters vs. minimal energy. . . . . . . . . . . . . . . 48
vi
Preface
This report is written as a result of the literature study for my Masters Thesis
project for Applied Mathematics of Delft University of Technology, carried out
at VSTEP BV. Since March 2007 I have been working on Dynamic Positioning
of vessels, especially for use in a training simulator.
Jalitha Wills
Rotterdam, June 2007
vii
Used symbols and units
Quantities
In this section a few tables can be found with the denition of the quantities
used.
1
Quantity Unit symbol
Length m
Mass kg
Time s
Table 1: Introduction on (Physical) Quantities and their units: Basic Units.
Physical Quantity Unit symbol
Plane angle rad
Force N = kg m/s
2
Work J = N m
Power W = J/s
Frequency Hz = s
1
Pressure Pa = N/m
2
Table 2: Introduction on (Physical) Quantities and their units: Special Quanti-
ties.
1
From [9, 10].
viii
Physical Quantity Unit symbol
Area m
2
Volume m
3
Density kg/m
3
Velocity m/s
Acceleration m/s
2
Angular Velocity rad/s
Angular Acceleration rad/s
2
Dynamic Viscosity Ns/m
2
Kinematic Viscosity m
2
/s
Table 3: Introduction on (Physical) Quantities and their units: Other Units.
ix
Symbols
In this paper several symbols are used and dened. In this section an overview
on them.
Symbols from Chapter 5
Roman Symbols
A
L
: cross section of the wind area. (longitudinal projection)
A
LS
: cross section of the submerged area. (longitudinal projection)
a
r
: resistance augmentation factor.
A
T
: section of the wind area. (transverse projection)
A
TS
: section of the submerged area. (transverse projection)
c: wave velocity.
C
D
: dimensionless drag coecient.
C
L
: dimensionless lift coecient.
C
c
:
c
dependent current load coecient.
C
w
:
rw
dependent wind load coecient.
C
wd
:
wd
and f
wd
dependent wave load coecient.
C
Q
0
: torque coecient of the propeller in open water.
C
Q
b
: torque coecient of the propeller working in the wake of the ship.
C
T
: thrust coecient.
D: cylinder diameter.
D
p
: diameter propeller.
F
D
: drag force.
F
l
: lift force.
F
: buoyant force.
f
v
: vortex shedding frequency.
f
wd
: regular wave frequency.
G: center of gravity
g: gravitational constant.
h: distance below the water surface.
H
1/3
: signicant wave height.
J: advance number from propeller.
k: wave number.
L: length of the ship.
m: mass of the ship.
n: number of revolutions per minute of propeller.
N
c
: current moment.
N
w
: wind moment.
N
wd
: wave moment.
(nabla): volume of the submerged part of the ship.

O: origin of earth-bound coordinate system.
p: hydrostatic pressure.
P
E
: eective power of the propulsion.
P
T
: power from the thrust.
Re: Reynolds number.
Q
p
: torque from propeller.
R
T
: calm-water resistance of the ship.
x
S: origin of steady translating coordinate system.
T: wave period.
t: time.
t
d
: thrust deduction factor.
T
e
: wave elevation period.
T
p
: thrust from propeller.
U: undisturbed ow velocity.
u: velocity of the ow in x-direction.
V : velocity of the steady translating coordinate system.
v: velocity of the ow in y-direction.
V
A
: average inow speed to the propeller.
V
c
: relative velocity of the current.
V
rw
: relative wind velocity (relative to ship velocity).
V
s
: velocity of the ship. V
w
: wind velocity.
w: velocity of the ow in z-direction.
w
r
: wake fraction.
x: steadily translating coordinate.
x
a
: surge coecient.
0
: earth-bound coordinate.
2
b
: body-bound coordinate.
X
c
: longitudinal current force.
X
w
: longitudinal wind force.
X
wd
: longitudinal wave drift force.
y: steadily translating coordinate.
y
a
: sway coecient.
Y
c
: transverse current force.
Y
w
: transverse wind force.
Y
wd
: transverse wave drift force.
z: steadily translating coordinate.
z
a
: heave coecient.
Greek Symbols
c
: current direction.
rw
: relative wind direction.
wd
: wave drift direction.
F
l
: phase shift of lift force.
x
: phase angle for surge movement.
y
: phase angle for sway movement.
z
: phase angle for heave movement.
: phase angle for roll movement.
: phase angle for pitch movement.
: phase angle for yaw movement.
a
: wave amplitude.
0
: open water eciency.
H
: hull eciency.
r
: relative rotative eciency.
a
: pitch coecient.
: wave length.
2
with {x, y, z}.
xi
: rotation angle of coordinate system.
: kinematic viscosity of the uid.
: density of water.
air
: density of air.
: potential function.
a
: roll coecient.
a
: yaw coecient.
: circular wave frequency.
e
: wave frequency of encounter.
Symbols from Chapter 6
Roman Symbols
c
1
(i):part of transverse force of (n 1)
th
thruster (as a coecient of (n-1)
longitudinal forces).
c
2
(i):part of transverse force of (n 1)
th
thruster (as a coecient of (n-2)
transverse forces).
F
: thruster conguration (which satises the forces constraints) with the least
energy.
F
max
: vector of maximum forces of thrusters.
F
x
: total of external forces in longitudinal direction.
F
xcurrent
: current force in longitudinal direction.
(F
x
)
i
: see (F
x
)
thruster
i
.
(F
x
)
now
: total longitudinal force produced by present conguration of thrusters.
(F
x
)
thruster
i
: longitudinal force produced by i
th
thruster.
F
xwanted
: longitudinal force demanded by skipper.
F
xwave
: wave force in longitudinal direction.
F
xwind
: wind force in longitudinal direction.
F
y
: total of external forces in transverse direction.
F
y
current
: current force in transverse direction.
(F
y
)
i
: see (F
y
)
thruster
i
.
(F
y
)
now
: total transverse force produced by present conguration of thrusters.
(F
y
)
thruster
i
: longitudinal force produced by i
th
thruster.
F
y
wanted
: longitudinal force demanded by skipper.
F
y
wave
: wave force in longitudinal direction.
F
y
wind
: wind force in longitudinal direction.
F
z
: total of external forces in z-direction.
g ((F
x
)
1
, . . . , (F
x
)
n1
, (F
y
)
1
, . . . , (F
y
)
n2
): energy function of the thruster
forces.
M: total of external moments.
M
current
: current moment.
M
now
: total moment produced by present conguration of thrusters.
M
wanted
: moment demanded by skipper.
M
wave
: wave moment.
M
wind
: wind moment.
s: part of transverse force of (n 1)
th
thruster.
x
i
: x-position of i
th
thruster.
y
i
: y-position of i
th
thruster.
xii
Greek symbols
1
: error tolerance for longitudinal force thrusters.
2
: error tolerance for transverse force thrusters.
1
Chapter 1
Introduction
Dynamic Positioning may be dened as a system which automatically controls
a vessel to maintain her position and heading exclusively by means of active
thrust.
1
This denition speaks of two special modes which are available for Dy-
namic Positioned ships. More special modes are available and almost every
Dynamic Positioned ship has 2 or 3 dierent modes installed. The main
advantage of Dynamic Positioning is that the system corrects for the external
forces working on the ship, namely wind, wave and current forces. A small
setback is that the system has high startup costs, but there will be a save of
money through lower fuel costs and no expenditures on tugboats.
Ships that are equipped with Dynamic Positioning often have azimuth thrusters,
but not necessarily so. These rotating propellers allow for greater manoeuvring
exibility than standard xed propeller-rudder arrangements. Under standard
sailing conditions, these are steered with rotating engine handles. But for
precise manoeuvring at low speeds, the skipper can often switch over to joystick
control. With joystick control, the ship can be moved and rotated by moving
the joystick in the desired direction. The ships control system has built in
logic to bring over the joystick steering to the correct rotation and rpm of the
engines.
A problem with the Dynamic Positioning system is that training is still very
expensive. The training is very elaborate, because the skipper should learn
to work with the system and should learn what to do in case of failure or
erroneous input from the sensors on the ship.
A Dynamic Positioning Simulator should be able to train the skipper in
dierent weather situations and in dierent operational situations.
Of importance in creating such a simulator are the hydromechanics involved
in sailing a ship and the mathematical model behind Dynamic Positioning.
Knowing all the forces acting on the ship, it is necessary to determine the
conguration of the thrusters in such a way that minimal power is needed.
Algorithms like Steepest Descent with Linesearch and BFGS method are
implemented in Matlab to achieve this.
1
From: [3]
2
VSTEP is interested in creating a Dynamic Positioning Simulator and
already has a lot of expertise in developing Training Simulators, since they have
created a Driving Simulator, whereby people learn to steer a car while learning
the dierent trac regulations, as well as a (Professional) Ship Simulator,
whereby people learn to steer various ships in increasingly dicult conditions.
This Ship Simulator game draws much attention also from the professional
nautical schools and related companies.
VSTEP is a company with locations in Rotterdam (headquarters) and Oxford
(UK). VSTEP means Virtual Safety Training and Education Platform. They
work in 3 elds: Scenario Training, Procedure Training and Training Simulators.
This report will start with an introduction of ships in Chapter 2. After
this in Chapter 3 an introduction about Dynamic Positioning will be given. The
necessity of training will be discussed as well. Then in Chapter 4 the programs
that are used to create the Ship Simulator game will be briey introduced.
In Chapter 5 the hydromechanics involved will be explicated. After this, in
Chapter 6, the dierent models are described. In Chapter 7 the results are
discussed. Finally, in Chapter 8 an overview of the rst three months and the
future goals will be given.
3
Chapter 2
Basics about Ships
From 1800 until after the WWII trade and passenger travel with the Dutch
colonies in the East and the West and the stream of emigrants who moved to
mostly North America led to a growing demand for developments in shipbuild-
ing. The technology in this area endured a slow, but radical development. This
development was characterized by:
transition from wooden ships to iron and later on steel ships,
transition from sail ships to steam ships and later on motor vessels,
emergence of new types of ships as tankers and refrigerated and freezer
vessels,
a partially increase in the speed of the vessels, the size of the vessels and
their safety.
After the WWII the world trade and with that the shipping changed very fast.
Everlasting more and bigger ships were needed. Transport of passengers, oil,
containers, heavy loads, animals and (dry) bulk with one and the same ship was
normal until about 1970. Most of the modern multi-purpose ships are not able
to transport so much dierent types of load.
The transport of passengers on long distant trips is now almost completely el-
bowed out by air trac. But the number of passengerships that are specialized
in vacation trips (cruises) has grown strongly.
1
In the next sections a general introduction will be given about ships, the move-
ments, and the type of ships in the Ship Simulator 2007 game. In Appendix
A gure A.4 a nice picture of a ship can be found with the terms used for the
dierent parts of the ship.
1
This introduction to the chapter is based on Chapter 1 from [5], p.50-51
2.1 Moves 4
2.1 Moves
The coordinate system of the ship is given in gure 2.1.
Figure 2.1: The coordinate system of a ship. The origin is in the ships center
of gravity.
There are six standard movements a ship can make and those are divided in
movements and rotations.
The three movements:
Surge: a horizontal oscillating movement in x-direction, i.e. a transient
motion in a fore and aft direction;
Sway: a horizontal oscillating movement in y-direction, i.e. a motion from
side to side;
Heave: a vertical oscillating movement in z-direction, i.e. a transient up-
and-down motion;
The three rotations:
Yaw: an oscillating rotation around z-axis, i.e. the bow yaws from side to
side;
Pitch: an oscillating rotation around y-axis, i.e. the bow pitches up and
down;
Roll: an oscillating rotation around x-axis, i.e. the vessel rotates from side
to side, about the fore-aft axis;
2.2 Types of Ships
There are dierent kind of ships in the simulator game. Some of them will be
described in this chapter. Oshore structures like oil rigs will be in the game, and
with this supply vessels like the Fairmount Sherpa and Fjell ships. Cruiseships,
push boats, rescue boats, tankers, life rafts, containerships and tugboats can
also be found in the game.
2.2 Types of Ships 5
Containerships
Since 1960 the transport of containers is increased and it still is increasing.
Containers can be transported from customer to customer and not only from
harbour to harbour. On land the transport is with trains or trucks, so a great
part of this transportation goes by other means than ships. There are large,
intercontinental containerships that only pass the large harbours and smaller or
mediocre containerships, that transport containers from the smaller to bigger
harbours, and provide services that are not protable for larger ships. The big
containerships from the game should be able to load and unload containers with
a crane.
2
Cruiseships
Nowadays the passengerships are exclusively meant for making very luxurious
trips to tropical countries and harbours. These ships mostly have swimming
pools, cinemas, bars, casinos and more on board.
3
There are stabilisation ns
to make sure the wiggles of the ship are not too extreme. The Titanic
4
, which
is simulated in the game, was a luxurious ship that was launched May 31, 1911.
On the rst trip of this ship (April 10, 1912), it crashed into an iceberg and
it sank within 3 hours. The vessel was 46,328 tons, had a length of 882.6 feet
and a breadth of 92.6 feet. The boat deck and bridge were 70 feet above the
waterline.
5
Oshore structures
Drilling for oil and gas on the seas is nowadays done in more than 40 countries,
100 of kilometres from the coasts, with increasing depths. There are dierent
types of oshore units build. In this game a semi-submersible drilling unit is
available. This can be used for drilling in waters with waterdepth between
150 and 2500 metres. Those platforms have Dynamic Positioning Systems (see
Chapter 3) in almost all cases.
Such units also need supply ships and oiltankers can come to retrieve the oil.
6
Tankers
There are dierent kind of tankers, namely gas tankers, (crude) oil tankers,
chemicals tankers and tankers with other substances, such as nished oil prod-
ucts like motor fuel or diesel oil, but also other uids like drinking water and
wine. In the game there is a crude oil tanker from the type VLCC (Very Large
Crude Carrier) which can contain about 200,000-300,000 DWT (dead weight
2
From: [5] p. 52-54
3
From: [5] p. 58
4
Text about the Titanic is from website: http://www.encyclopedia-titanica.org/index.php
5
http://www.titanicinquiry.org/
6
From: [5] p. 64-73
2.3 Thrusters 6
tonnage). Oil tankers can be load and unload through a exible hose or a pipe
arm.
7
Tugboats
A characteristic of tugboats is the low stern. This is necessary to give the wire
enough freedom of movement. The point were the force of the wire engages to
the ship, should be near midship in such a way that this force does not inuences
the steering of the ship. The towing winch, which should translate the complete
towing force from the propeller to the wire, is very important. Tugboats are
used for salvage, towing, anchoring on oshore applications and assistence to
ships in the state of emergency.
8
2.3 Thrusters
There are three types of thrusters that make up the majority of units found in
DP vessels, which are: main propellers, tunnel thrusters and azimuth thrusters.
9
Combinations of these thrusters can be found on dierent kind of ships. A semi-
submersible drillship for instance can come with six azimuth thrusters.
2.3.1 Main propellers
Main propellers, sometimes in combination with rudder systems, provide for
the main propulsion. In some vessels these rudders are also part of the DP
system. If the rudders are not part of the DP system, then it is important
that the rudders are amidships, because else they can lead to sway or yaw
movements when the propellers are switched on.
2.3.2 Tunnel thrusters
The eective thrust of a tunnel thrusters depends on the depth of immersion of
the propeller, the length of the tunnel and the speed of the vessel. The tunnel
creates an increase in thrust in most of the cases. The ow of the water through
the tunnel gives the water inside the tunnel a higher velocity than the water
outside the tunnel and this results in a pressure dierence, which results in an
increase in thrust force. The noise from the propellers is also reduced.
If the propeller is too near the surface then it is less ecient, because there is
air resistance, due to the air on the surface, and also because this can lead to
cavitation, which is a result from air bubbles -due to pressure dierences- that
implode and lead to damage of the propeller-blade.
The eective thrust decreases also when the speed of the vessel is too high (this
bound depends on the kind of ship). The higher the speed is, the higher the
7
From: [5] p. 55-56
8
From: [5] p. 59-60
9
Information for this chapter from [3] chapter 4, [5] chapter 12 and [15]
2.3 Thrusters 7
frictional resistance to the tunnel is, and if this increase in friction is bigger
than the improve in thrust, then there is a reduction in eciency.
A long tunnel causes increased friction losses, and a too short tunnel causes
losses due to turbulence.
2.3.3 Azimuth thrusters
An azimuth thruster is, as the name says, a unit able to generate thrust in any
direction. These units may be tted to provide the desired manoeuvrability, or
sometimes as a substitute for the main propellers. The most important advan-
tage of azimuth thruster is an optimal thrust in every direction.
In some congurations these thrusters are positioned right underneath the hull
of the vessel, in which case the azimuth thrusters are vulnerable to grounding
damage and one should take precautions this will not occur. The thrusters can
be placed higher on the ship too, but in this case they are closer to the surface
and then they are less eective.
There are also azimuth thrusters that are retractable, horizontally in the hull
or vertically. The rst one takes up less space. When the thruster is retracted
it cannot be used.
Some azimuth thrusters are removable. This might be interesting for vessels
where retraction of the thrusters is not possible (when there is not enough
space for this for instance) or when it is too expensive to install. This is nice
for ships that spend most of the time in deep water; when travelling to shallow
water one can remove simply the thrusters.
An azimuth thruster can work in the ahead and the astern direction. In the
astern direction the amount of thrust available falls to only 60%, but since sail-
ing in reverse is faster than rotating the unit through 180
and since sometimes

there is not enough space for rotation, this still is an interesting possibility. But
although this is true, some DP systems will rotate the vessel when the eectivity
is that low, and other DP systems force the thruster to operate in the ahead
mode, so it is important to keep this in mind.
In congurations where the azimuth thrusters are located near each other, it
is necessary to ensure that they do not intervene with each other, i.e. that the
stream of one will not aect the other thruster, because the other thruster can
then overspeed or become less eective. This is also true if the other thruster
is a main propeller or tunnel thruster. To cope with this, the thrusters can be
barred, which means the range of movement is restricted. This can also be done
when it is undesirable to have a thrust in a certain direction, for instance when
there are divers in that region. Such a possibility is called a thruster inhibit. If
this is not available, then it might be necessary to shut down the thruster.
8
Chapter 3
Dynamic Positioning
Dynamic positioning was rst discovered as a way of positioning ships in the
early 1960s. It rapidly growed and now dynamic positioning is established as a
reliable way of maintaining a vessels location or heading over a long period of
time. Dynamic positioning systems perform well in deep as well as shallow water
and the costs do not increase signicantly with water depth. Other advantages
are better station-keeping accuracy and exibility.
1
3.1 Denitions
There are many denitions for Dynamic Positioning. From [17]:
A means of holding a vessel in relatively xed position with respect to the
ocean oor, without using anchors accomplished by two or more propulsive
devices controlled by inputs from sonic instruments on the sea bottom and on
the vessel, by gyrocompass, by satellite navigation or by other means.
Another nice denition, from [3], is:
Dynamic Positioning may be dened as a system which automatically controls a
vessel to maintain her position and heading exclusively by means of active thrust.
And nally one other denition, from [5]:
With Dynamic Positioning the ship will be held in position within prescribed,
in advance given tolerances, in spite of current and wind inuences.
3.2 Why Dynamic Positioning?
This is an interesting question. Dynamic Positioning has some advantages when
comparing with other systems, but also disadvantages. But there are a lot of
1
From [3] p.1-4
3.2 Why Dynamic Positioning? 9
possibilities with this system. First a short introduction on Dynamic Positioning
training to understand the importance of training.
3.2.1 Dynamic Positioning training
With a Dynamic Positioning system the skipper, when steering, does not need
to correct for the wind, current and wave inuences. The system will make such
a correction. So why is training necessary one can ask then? First of all the
skipper should learn to work with the system, what are the dierent modes
possible (see section 3.2.3), and he should learn how to steer the ship. Most DP
systems work with joystick steering, he should learn how to handle this. This is
especially important in applications that need a lot of precision, like anchoring
to an oil rig or in a harbour.
Furthermore it is important for the skipper to learn what to do in case of an
emergency. What if a part of the systems falls out, or what if the entire system
fails? And of course he need to see when the wind or current sensors return an
unusual value that might be wrong. He might need to overrule the system then.
Also extreme weather situations should be practised.
Target cases will be dened with several conditions, which are important to
learn.
3.2.2 Comparison of Dynamic Positioning to other sys-
tems
To see why Dynamic Positioning (from now on referred to as DP) is so interesting
for position control besides the obvious advantages, this system is compared to
other systems.
2
Jack-up barge: in this system there is no need for power, thrusters, extra
generators or controllers. This means that the system is immune to power
blackouts or system failures. Another advantage is that there will be no
underwater hazards from thrusters, so this is nice for boats with divers.
It is also true that once on location, the vessel will not need position
references anymore.
But of course this system also has some disadvantages. There is in the
rst place a limitation to the water depth this system is applicable to.
Some books speak of limited water depts of up to about 60m, others
speak of 150m, but both are very limited depths. When spudded in, the
manoeuvrability is gone too. A last disadvantage is that for rigmoves
3
one
needs tugs.
Spread mooring to anchor pattern (Anchoring): again we have no need
for complex systems with thrusters, extra generators and controllers, so
the system is immune to power blackouts or system failures. Also for this
system it holds that when on location there are no position references nec-
essary anymore. Without thrusters we have again no underwater hazards.
2
From [3] p.5-6
3
This is a move of a rig, for instance an oil rig.
The advantages are the same.
The disadvantages are pretty much the same. Only for this system there is
limited manoeuvrability, when moored, instead of none. The waterdepth
is also less limited, up to 600m. For rigmoves and for laying of mooring
tugs are necessary. And, this is also an important disadvantage, excessive
time is required for setting up the anchor pattern.
DP: this system has lots of advantages, but also disadvantages. One very
important advantage is that tugs are not necessary anymore. The set-up
on location is quick, and during operation the vessel is very manoeuvrable.
If the weather changes, or there are other changes in the situation on the
water, the system can react on this very fast. If the operation requirements
change, the system can adapt to this also easily. Task can be completed
fast too, so this just means that it is cheaper. The water depth does not
matter anymore, the system can work in any water depth, although in
shallow water there might be problems. This has also less risk for the en-
vironment, since there are no mooring lines on the seabed.
The main disadvantages are vulnerability to power failures and shortages,
thruster failure and electronics failures (since DP is a very complex sys-
tem). Since the system needs a lot of power, the fuel costs are higher too.
The initial costs are also high. Diving can be dangerous with the thrusters
under the ship. Position reference is needed continuously. And last but not
least, there is need for highly educated sta members, thus training and
experience are needed now.
From this we see that DP has many advantages as well as disadvantages, but
important is that DP can be used on locations where the other systems can not
be applied. Furthermore the manoeuvrability at any time plays an important
role in the need for DP.
DP also has many advantages in comparison to conventional steering with turn-
ing gas handles. Dynamic Positioning can save power since the system calculates
the forces needed to overcome the external forces, while a skipper can make mis-
takes in estimating these forces. No tugboats are necessary anymore, where as
it is to hard in many situations for a skipper to sail himself with the ordinary
gas handles. A last interesting advantage is that DP can be used in situations
where a lot of precision is needed. A supply vessel that should anchor to an oil
rig can do this more easily with DP for instance.
3.2.3 Operating with DP
DP has many applications. There are several modes of operation possible,
namely:
4
manual/joystick mode: the operator has full control over the vessel;
auto-heading mode: the system maintains required heading automatically;
auto-position mode: the system maintains required position automatically;
4
From: [1] Chapter 10 p.15-16
auto area position mode: the system maintains automatically within a
specied area, while using minimum power;
auto track mode: the vessel follows a specied track described by a set of
way-points;
autopilot mode: the vessel steers automatically along a pre-dened course;
follow target mode: the vessel follows a constantly moving target such as
an ROV (remotely operated vehicle).
3.2.4 Applications where DP can be used
Most DP systems are equipped with two or three of such applications. With
this and the advantages, we can see directly a few examples of applications of
DP systems:
5
Cable- or pipelay: in this case it is needed that the vessel will sail in a
straight line or course;
Exploration/production drilling: the water can be deep and the ship should
stay stable during this kind of operation;
Oceanographical research: the vessel should be able to move a lot of times,
and lay still on other times, so anchoring every time is not an option. This
is also better for the environment, since the vessel will not be anchored into
the seabed. And with this system the vessel is able to follow a predened
course;
Wreck survey, salvage and removal: a ship equipped with DP can search
the area of the wreck to nd it (if the exact position is not available) and
then stay stable there while the wreck is recovered;
Cruiseliners: the ship does not need tugboats to moor in tight harbours.
Dredging: DP is used here for track keeping.
Search for sea mines: DP is used here for track keeping. The ship should
follow a predened course in search of the mines.
In the oshore industry, which is of most interest to VSTEP at the moment,
the following applications can be found:
Supply vessel: for supplying oshore platforms.
Anchor handling: an oshore rig should be anchored as fast as possible
and this should be done very accurately as well.
Floating crane operations: the crane vessel should keep the correct heading
and position.
5
From: [3] p.3
3.3 Dynamic Positioning System 12
3.3 Dynamic Positioning System
Vessels that need to be dynamically positioned need to have Dynamic Position-
ing systems that compensate for wind forces, waves and current, each of which
can have a dierent direction and force onto the vessel. Worst-case combinations
can occur when wind, current and waves come from dierent directions.
There is a need for sensors on the ship that measure the velocities and di-
rection of wind, waves and current. A mathematical model will then calculate
the power and direction of the thrusters to stay on course. An example of a
schematic drawing of this system is given in gure 3.1.
Ships that are equipped with Dynamic Positioning equipment often have az-
imuth thrusters, but not necessarily so. A combination of twin screws at the
back with conventional rudder blades plus a bow and/or stern thrusters also
allows for straight sideways movements.
There are two Dynamic Positioning systems available: systems based on PID
regulator, and systems based on model control. The rst type can only correct
after the ship gets o course, the other type can predict deviation and react on
this on beforehand.
6
Figure 3.1: This is a schematic drawing of a dynamic positioning system.
7
6
more in OE 5663
7
source: http://content.answers.com/main/content/wp/en/f/fc/Controller.jpg
3.3.1 Mathematical model
The model behind this DP system should calculate the resulting thruster cong-
uration, using the inputsignals from the sensors on the ship. The model should
describe the reality as accurate as possible, but one should not forget that there
are already errors made when measuring the forces of wind, waves and current.
From all the forces the direction and velocity is needed. For waves one also
needs the wavelength. From this data all the resulting forces will be estimated.
The thruster force is a function of the rotations per minute, the blade area,
the interactions between the thruster and the hull, and between thrusters. The
model should then calculate the force needed by the thrusters to move the ship
in the correct direction (or to hold the ship stable on one location) and mini-
mize the energy produced by these thrusters. So this model holds a combination
of various mathematical and physical methods including hydromechanics, wave
theory, systems theory and optimization. For the calculations numerical meth-
ods can be used to approximate reality.
In Appendix A, gure A.6, an example of such a mathematical model is given.
3.3.2 Position Reference Systems
To know the position of the ship there are several reference systems possible.
Some of them are:
Taut-wire/rise inclinometers: these systems are particularly useful when
the vessel may spend long periods on a static location, where the water
depth is limited. Tautwire-based reference may also be used when the
vessel need to maintain a location relative to a moving vehicle.
There are a number of congurations possible for a taut-wire system, but
the most commonly used system lowers a wire with a weight on the end
of the wire to the seabed. Then when in operation the angle of the wire
is measured. The length of the wire together with the angle dene the
position of the ship. Strong currents degrade the accuracy of the system,
especially in deep water, and in shallow water the movement is limited,
because the angle of the wire can not become too large.
8
Hydro-acoustic Position Reference systems: these systems need a
transponder positioned on the seabed and a transponder on the ship. An
acoustic signal is send from the ship to the seabed and the time needed
for the signal to return to the ship is measured. With that time one can
measure the position of the ship relative to the transponder on the seabed.
It is important though, not to forget that there should be a compensation
for the roll and pitch moves of the ship.
There are also systems, called Long Baseline acoustic position reference
systems, that have multiple transponders on the seabed (not all are used,
some are for redundancy). This will improve accuracy. No angle measures
are required because of the multiple signals that are received. Dierent
kinds of systems like this are of course possible.
8
From [3] p.89-94
(Dierential) Global Positioning Systems: this system works with data
that is received from dierent satellites. From these the system calculates
the position of the vessel. With dierential GPS this data is combined
with data from a reference station which position is know exactely.
15
Chapter 4
Overview of Ship simulator
In this chapter the ship editor of VSTEP will be introduced.
Figure 4.1: Screenshot from the ship editor.
16
Two examples of ships can be found in gure (4.2) and gure (4.3).
Figure 4.2: Screenshot from the ship editor. This is a harbour patrol boat.
Figure 4.3: Screenshot from the ship editor. This is a hovercraft.
17
A ship has a complex shape. To calculate all the forces working on the ship,
one should take its shape into account. Since this is a complex process, one can
also try to approximate the shape. In this ship editor the ship is approximated
by a block. This block can take any size necessary and can be rotated. In this
way it is possible for the block to t the ship very well. This can be seen in
gure (4.4(a)) and (4.4(b)).
As is known several forces work on a ship in the water. Even when the ship lies
still in the water the buoyancy force and gravity work on the ship. The gravity
works in this editor on the center of mass, which can be seen in for instance
picture (4.4(b)). This can be seen in pictures (4.5(a)), (4.5(b)) and (4.5(c)),
where the mass of the ship is increased and the ship starts to sink, because the
buoyancy force is not big enough to overcome the gravity.
1
A lot of properties of the ship and thrusters are not calculated, but read by
the program from curves. Other properties are calculated, but since the shape
of the ship is approximated, these properties are not correct, because the shape
of the ship is more complex. Now some of those properties can be corrected
then by a curve to take this complex shape into account. Some of these curves
can be found in gures (4.6), (4.7), (4.8), (4.9) and (4.10). These curves show
the dierent kind of possibilities in this editor. Although the forces, resistances
and thrust are not calculated exactly, they are modied to make it more realistic.
1
This is explained in Chapter 5.
18
(a) Ship approximated by a block.
(b) The block can be rotated.
Figure 4.4: In the upper screenshot one can see that the ship is approximated
by a block. In the lower screenshot one can see that it is possible to rotate and
move this block. From the last picture the orientation and origin of the axis
are also clear. In this editor the axis have dierent names than dened in this
paper.
19
(a) Harbour patrol boat with normal mass.
(b) Harbour patrol boat with higher mass.
(c) Harbour patrol boat with even higher mass
Figure 4.5: The eect of the mass of a ship. In the upper graph the harbour
patrol boat has a normal mass. In the center picture the harbour patrol has a
higher mass. The ship lies lower in the water, but still buoyancy is higher than
gravity. In the lower graph this is not the case, the gravity has won and the
harbour patrol boat can not be seen anymore, only the antenna is above water.
20
Figure 4.6: Curve of the correction of the resistance of dierent parts of the
ship. The resistance is calculated by the ship editor, but since this is calculated
on a simple block, the result should be scaled to take the complex shape into
account. The scaling can be found from the graphs, where on the x-axis the
speed in knots can be found and on the y-axis the scaling factor.
Figure 4.7: Curve of the thrust at dierent speeds. The thrust is the force the
propeller produces.
21
Figure 4.8: At lower velocities the thrusters can adjust to a higher speed easier
than at higher speeds. This is illustrated in this curve.
Figure 4.9: Curve of the liftforce that is generated by yawing movements of the
ship.
22
Figure 4.10: When the ships starts to accelerate, it banks over a bit and this is
caused by a sideward directed force, also called the torque of the propeller. This
force is not calculated in the ship editor, but read from this curve.
23
The results of all these curves can be seen in a few screenshots from the editor,
where the ship is sailing. In gure (4.12) the ship is sailing forward and one can
see the white, long arrow indicating the thrust in forward direction. In gure
(4.11) the ship is sailing sidewards and there is a white arrow in that direction
as well.
Figure 4.11: Screenshot of the ship sailing forward. The white arrows at the
thrusters show the thruster forces working on the ship.
24
Figure 4.12: Screenshot of the ship sailing sideward. The white arrows at the
thrusters are showing the side forces working on the ship. There is also a force
working on the ship near the center of gravity which is due to the ships sideward
movement through the water.
25
Chapter 5
Hydromechanics
In this chapter the basics of hydromechanics will be described.
As an introduction a few laws and basic equations will be mentioned before
getting into details.
5.1 Introduction
At rst there is Archimedes law:
m =
, (5.1)
where m is the mass of the ship, is the density of the water and

is the
volume of the submerged part of the ship. This leads to the law of buoyancy:
F
= g
, (5.2)
where F
is the buoyant force working on the ship, is the density of the water,
g is the gravitational constant and

is the volume of the submerged part of
the ship. This law states basically that when a solid is immersed in a liquid, it
experiences an upthrust equal to the weight of the uid displaced.
The hydrostatic pressure follows from this equation:
p = gh, (5.3)
where p is the hydrostatic pressure, is the density of water, g is the gravita-
tional constant and h is the distance below the water surface.
5.2 Flow
The continuity equation:
t
+
(u)
x
+
(v)
y
+
(w)
z
= 0, (5.4)
5.2 Flow 26
where is u the velocity of the ow in the x-direction, v the velocity of the ow
in the y-direction and w the velocity of the ow in z-direction.
If the uid is treated as incompressible, then is constant and this leads to the
simplied continuity equation:
u
x
+
v
y
+
w
z
= 0. (5.5)
Denition 5.2.1 (Potential Function). A potential function, , associated with
a potential ow eld is a mathematical expression having the convenient prop-
erty that at any point in the ow, the velocity component in any chosen direction
is simply the derivative of this potential function in that chosen direction:
u =

x
v =

y
w =

z
.
From this denition the new continuity equation for incompressible ows be-
comes the Laplace equation:
x
2
+

2
y
2
+

2
z
2
= 0. (5.6)
Very important in the uid dynamics are the Navier-Stokes equations (here
for incompressible ows):
(
u
t
+u
u
x
+v
u
y
+w
u
z
) = f
1

p
x
+(
2
u
x
2
+

2
v
y
2
+

2
w
z
2
),
(
v
t
+u
v
x
+v
v
y
+w
v
z
) = f
2

p
y
+(
2
u
x
2
+

2
v
y
2
+

2
w
z
2
), (5.7)
(
w
t
+u
w
x
+v
w
y
+w
w
z
) = f
3

p
z
+(
2
u
x
2
+

2
v
y
2
+

2
w
z
2
).
These equations can be simplied by neglecting the volumetric forces (f
i
) and
assuming that the ow is non-viscous ( = 0). This leads then to the Euler
equations:
u
t
+u
u
x
+v
u
y
+w
u
z
=
1
p
x
,
v
t
+u
v
x
+v
v
y
+w
v
z
=
1
p
y
, (5.8)
w
t
+u
w
x
+v
w
y
+w
w
z
=
1
p
z
.
From Denition (5.2.1) the Euler equations can be transformed. Using:
u
u
x
=

x
x
2
=
1
2
x
(
x
)
2
,
v
u
y
=

y
xy
=
1
2
x
(
y
)
2
,
w
u
z
=

z
xz
=
1
2
x
(
z
)
2
,
5.3 Forces 27
u
v
x
=

x
xy
=
1
2
y
(
x
)
2
,
v
v
y
=

y
y
2
=
1
2
y
(
y
)
2
,
w
v
z
=

z
yz
=
1
2
y
(
z
)
2
,
u
w
x
=

x
xz
=
1
2
z
(
x
)
2
,
v
w
y
=

y
yz
=
1
2
z
(
y
)
2
,
w
w
z
=

z
z
2
=
1
2
z
(
z
)
2
.
the Euler equations (5.8) become:
x
_
t
+
1
2
_
_
x
_
2
+
_
y
_
2
+
_
z
_
2
_
_
= 0,
y
_
t
+
1
2
_
_
x
_
2
+
_
y
_
2
+
_
z
_
2
_
_
= 0, (5.9)
z
_
t
+
1
2
_
_
x
_
2
+
_
y
_
2
+
_
z
_
2
_
_
= 0.
Since the dierentations of the expressions between braces {. . .} with respect to
x, y and z gives zero, which means that these expressions are functions of only
time, C(t), which then provides the Bernoulli equation for an instationary
ow:
t
+
1
2
_
_
x
_
2
+
_
y
_
2
+
_
z
_
2
_
+
p
+gz = C(t), (5.10)

where the term gz is the gravity which was neglected at rst, but returns in this
equation.
5.3 Forces
A ow disturbance results in a drag force (F
D
) -by denition parallel to the
ow direction- on the object. When the ship is seen as a cylinder, as illustrated
in gure 5.1, this force (on a unit length of cylinder) can be written as:
F
D
=
1
2
U
2
C
D
D, (5.11)
with mass density of the uid, U the undisturbed ow velocity, C
D
dimen-
sionless drag coecient and D the cylinder diameter.
1
1
This is from [12] Chapter 4 p8-11
5.3 Forces 28
A lift force (F
l
) is dened as a force component acting perpendicular
to the undisturbed ow velocity, so this force is also perpendicular to the
drag force. There is no lift force when the cylinder is placed in a completely
symmetric ow. When the velocity gets higher and the Reynolds number,
Re =
V d
, with V the velocity of the ow, d the diameter of the cylinder and
the kinematic viscosity of the uid, exceeds 90, then the ow is disturbed
too much and a Von Karman vortex street appears.
2
When vortices are shed
behind the cylinder, there is a resulting force directed toward the vortex from
high velocities in the wake and low local pressures. The lift force can be seen
in gure 5.1. This lift force (on a unit length of the cylinder) is dened as:
F
l
=
1
2
U
2
DC
L
sin(2f
v
t +
F
l
), (5.12)
with mass density of the uid, U the undisturbed ow velocity, D the cylinder
diameter, C
L
dimensionless lift coecient, f
v
vortex shedding frequency, t
time,
F
l
phase shift.
3
Figure 5.1: Drag and lift force are illustrated in this gure, where the ship is
seen as an cylinder and U is the velocity of the ow. F is the resulting force
from drag and lift force.
4
Wind has a stochastic nature which depends on time and location. At
sea the variation in the mean wind velocity is small compared to the wave
period. The wind can be considered as steady, both in magnitude and direction,
which result in constant forces and moment on a xed oating or sailing body.
The wind plays two roles in the behaviour of a oating body. First it exerts
a force on the part of the structure exposed to the air, for the calculation of
this force only local winds are necessary. Secondly, wind generates waves and
currents and through these inuence the ship too. To determine these eect,
wind conditions in a much larger area are needed. The direct wind forces
and moment can be computed as follows:
X
w
=
1
2
air
V
2
rw
C
X
w
(
rw
)A
T
,
2
See for instance: http://en.wikipedia.org/wiki/Von K%C3%A1rm%C3%A1n vortex street
for information on the Von Karman vortex street.
3
4
From [12] Chapter 4 p.13.
5.3 Forces 29
Y
w
=
1
2
air
V
2
rw
C
Y
w
(
rw
)A
L
, (5.13)
N
w
=
1
2
air
V
2
rw
C
N
w
(
rw
)A
L
L,
with X
w
steady longitudinal wind force, Y
w
steady transverse wind force, N
w
steady horizontal wind moment,
air
density of air, V
rw
relative wind velocity,
rw
relative wind direction, A
T
transverse projected wind area, A
L
longitudinal
projected wind area
5
, L length of the ship, C
w
(
rw
)
rw
dependent wind load
coecient.
6
There is actually another approach to calculate the wind force. The ship is then
split into dierent common shapes as circles, rectangles, etc. The total wind
force is then the sum of the forces on the various shapes at dierent locations
at this height. For now this will not be done, but the formulas from equation
(5.13) will be applied.
7
The wind velocity can be calculated using the power law
prole:
V
w
(z) = V
w
(z = 10m)
_
z
10
_1
8
, (5.14)
where z is the height at which the velocity is needed, V
w
(z = 10m) is the
velocity at 10m (which is the typical measurement height) and V
w
(z) is the
velocity at z.
8
There are also dynamic wind eects. The wind velocity is typically dened
as the 1-hour sustained velocity, but there are also 10-minute and 1-minute
velocities, which can be up to 25% and 40% higher than the 1-hour velocities.
There are various spectra used to describe this, namely Harris, Ochi-Shin, DNV
and Wills spectrum.
The forces and moment exerted by a current on a oating object are
composed of two parts. A viscous part, due to friction between the uid and the
structure. For blunt bodies it is small compared to the viscous pressure drag,
so this can be neglected. The second part is a potential part, with a component
due to circulation around the object, and with a component from the free water
surface wave resistance, but this last component is small in comparison with
the rst in most cases and will be neglected. These are the equations for the
current forces and moment:
X
c
=
1
2
V
2
c
C
X
c
(
c
)A
TS
,
Y
c
=
1
2
V
2
c
C
Y
c
(
c
)A
LS
, (5.15)
N
c
=
1
2
V
2
c
C
N
c
(
c
)A
LS
L,
with X
c
steady longitudinal current force, Y
c
steady transverse current force,
N
c
steady yaw current moment, density of water, V
c
current velocity,
c
current direction, A
TS
submerged transverse projected area, A
LS
submerged
longitudinal projected area
5
, L length of the ship, C
c
(
c
)
c
dependent current
5
For transverse and longitudinal projection, see gure (5.2).
6
7
From [14] p71-72
8
This is from [14] p71
5.3 Forces 30
load coecient.
9
Figure 5.2: On the left is the transverse projection of the ship and on the right
is the longitudinal projection of the ship.
There are a lot of books about wave theory. Dierent ways of estimating the
forces and added resistance are explained in them. For the sake of simplicity in
the rst instance only a small part will be studied and applied, but later on this
will be extended.
In DP cyclic wave forces are not important, since they do not result in a net
displacement of the vessel. These waves have short periods (think of 20 seconds
or shorter) so would require also a lot of power from the thrusters, which is
not necessary, because the average position of the vessel is not aected by these
waves.
The other type of wave force is the wave drift force. This force is relatively
recently understood. The forces were reported at rst in 1924, but the eects
were not understood until 1960. A good example of wave drift forces is that
they can be used to retrieve an object from the water for instance, so now it
is clear what is discussed here. The mean wave drift force can be broken down
into four components (ve in shallow water), namely:
contribution due to relative wave height;
contribution due to drop of dynamic pressures due to the submerged hull;
contribution due to pressure due to rst order vessel motions;
contribution due to combined rotation with translational inertia;
contribution due to a second order wave potential (only in shallow water).
The rst of these forces is the dominant one. For high frequency waves, the third
and fourth term disappear, because these waves do not result in signicant rst
order vessel motions. In regular waves, the wave drift forces are constant and
dependent on the wave frequency.
Because the rst term is the dominant wave drift force, the wave drift forces
and wave drift moment are often dened as:
X
wd
= C
X
wd
(
wd
, f
wd
)
1
8
gH
2
1/3
L,
Y
wd
= C
Y
wd
(
wd
, f
wd
)
1
8
gH
2
1/3
L, (5.16)
N
wd
= C
N
wd
(
wd
, f
wd
)
1
8
gH
2
1/3
L
2
,
(5.17)
9
5.4 Ship movements 31
with X
wd
steady longitudinal wave drift force, Y
wd
steady transverse wave
drift force, N
wd
steady yaw wave drift moment, density of water,
wd
wave direction, f
wd
the regular wave frequency, C
wd
(
wd
, f
wd
)
wd
and
f
wd
dependent wave drift coecient, g is the gravity coecient, H
1/3
is the
signicant wave height and L the length of the ship.
10
Reaction forces occur as a result of the vessel moving through the water
and air. Since the reaction forces due to water are much larger than those due
to air, the latter is neglected in commonly. The water reaction forces are hard
to calculate, this will be another extension for the future.
5.4 Ship movements
The coordinate system dened before in Chapter 2 and gure (2.1) was a body-
xed coordinate system, where the origin is in the ships center of gravity (G).
There are also two other coordinate system important when discussing ship
movements, namely the steadily translating coordinate system and the earth-
xed coordinate system. The importance of the steadily translating coordinate
system is found in the fact that the ship carries out oscillations around this
system. For the calculation of the forces the body-xed coordinate system is
important and position reference systems return earth-xed coordinates for the
position of the ship.
The steadily translating coordinate system is moving forward with a constant
ship speed V . If the ship is stationary the direction of this coordinate system are
in the same direction as those of the body-xed system. The origin is at, above,
or under the time-averaged position of the center of gravity. The earth-xed
system has the positive x-axis in the direction of the wave propagation. The
positive z-axis is directed upwards. The (x, y)-plan lies in the still water surface
and can be rotated at a horizontal angle relative to the translating coordinate
system. This can be seen in gure (5.3).
As mentioned before the circular waves are not important in DP situations, but
since this paper handles also situations without DP, here a short introduction
of these waves.
Now all the conventions are know, one can dene the harmonic elevation of the
wave surface in the earth-bound coordinate system by:
=
a
cos (t kx
0
), (5.18)
where
a
is the wave amplitude, k =
2
is the wave number, is the wave

length, is the circular wave frequency, t is the time.
The wave speed is dened as:
c =

k
=

T
.
This wave speed has angle of direction . The steadily translating coordinate
system is moving forward at the ships speed V which leads to:
x
0
= V t cos +xcos +y sin. (5.19)
10
This part about the wave forces is from [14] p75-78
11
5.4 Ship movements 32
Figure 5.3: On the left is the earth-xed coordinate system. On the right are
the steadily translating and the body-xed system.
11
In gure (5.4) a graph can be seen from the ship which encounters the wave.
Figure 5.4: Ship encounters the waves.
12
The period at which this happens is:
T
e
=

c +V cos ( )
=

c V cos
.
From this the circular frequency of encounter,
e
, becomes:
e
=
2
T
e
=
2(c V cos )
= k(c V cos ).
12
5.5 Thrusters 33
Since kc = the relation between and
e
is clear:
e
= kV cos . (5.20)
Substituting equations (5.19) and (5.20) in (5.18) leads to the following equation
for the wave elevation:
=
a
cos ((
e
+kV cos )t k(V t cos +xcos +y sin))
=
a
cos (
e
t kxcos ky sin). (5.21)
The resulting ship movements are:
Surge : x = x
a
cos (
e
t +
x
),
Sway : y = y
a
cos (
e
t +
y
),
Heave : z = z
a
cos (
e
t +
z
),
Roll : =
a
cos (
e
t +
), (5.22)
Pitch : =
a
cos (
e
t +
),
Yaw : =
a
cos (
e
t +
),
in which each of the values is a dierent phase angle.
To nd the motions of a certain point, one needs to transform this point at rst
from the body-bound coordinate system to the steadily translating coordinate
system. The angles of rotation , and are assumed to be small, so one can
apply linearizations: sin and cos 1. With this the transformation
matrix becomes:
_
_
x
y
z
_
_
=
_
_
1
1
1
_
_
_
_
x
b
y
b
z
b
_
_
(5.23)
So with this the motions of a point P are given by:
x
P
= x y
b
+z
b
,
y
P
= y +x
b
z
b
, (5.24)
z
P
= z x
b
+y
b
,
where x, y, z, , and are the motions of and about the center of gravity, G.As
can be seen the motion of the point in x-direction is made up of surge, pitch and
yaw contributions, the motion in y-direction of sway, roll and yaw contributions
and the motion in z-direction of heave, roll and pitch contributions.
5.5 Thrusters
The thrust T
p
and the torque Q
p
, which are illustrated in gure 5.5, of a propeller
are expressed as functions of the rpm n as:
13
T
p
= C
T
n
2
D
4
p
, (5.25)
Q
p
= C
Q
0
n
2
D
5
p
, (5.26)
5.5 Thrusters 34
Figure 5.5: The Thrust and Torque are illustrated here.
14
where C
T
and C
Q
0
are thrust resp. torque coecients, n is the number of rev-
olutions per minute and D
p
is the diameter of the propeller.
Now for the eciency of the propeller, dene the advance number J:
J =
V
A
nD
p
, (5.27)
where V
A
is the average inow speed to the propeller.
The open water eciency
0
is dened as the ratio of the work done by the
propeller in developing a thrust force T
p
V
A
divided by the work required to
overcome the torque 2nQ
p
. With equations (5.25), (5.26) and (5.27) one can
deduct the following formula for the eciency:
15
0
=
T
p
V
A
2nQ
p
(5.28)
=
C
T
n
2
D
4
p
C
Q
0
n
2
D
5
p
V
A
2n
=
C
T
C
Q
0
J
2
(5.29)
5.5.1 Interaction
The performance of a thruster will be inuenced by the position of the thruster,
the shape of the ship and other nearby thrusters. In this section the interaction
between ship and propeller will be discussed. For now the thruster-thruster
interaction will be neglected.
The propulsive eciency is split in three parts: the open water eciency, the
13
From [2, 8]
14
From: http://www.centennialoight.gov/essay/Theories of Flight/props/TH18G2.htm
15
From [2] p.40 and [8] p.26
5.5 Thrusters 35
hull eciency and the relative rotative eciency. The eective power is dened
as:
P
E
= R
T
V
s
, (5.30)
where R
T
is the total calm-water resistance of the ship, excluding resistance
the propellers yield, and V
s
is the speed of the ship, P
E
then denotes the power
needed to tow the ship without propellers.
The power from the thrust is:
P
T
= T
p
V
A
, (5.31)
where T
p
is the thrust, and V
A
the speed of advance of the propeller. The speed
of advance of the propeller is dened as the propeller inow.
The eective power is thus the power to get the ship through the water at a
certain speed, when it is pulled through the water and the power from the thrust
is the power to get the ship through the water at a certain speed with propellers
on the ship. When the ship is tted with a propeller now and is traveling at the
same speed, the pressure eld around the hull changes due to the action of the
propeller. This creates a pressure dierence between the hull surface and the
after part of the hull and thus increases the resistance of the vessel from that
which was measures in the towed resistance case. Now this increase in resistance
can be expressed as:
T
p
= R
T
(1 +a
r
), (5.32)
where a
r
is the resistance augmentation factor.
An alternative way to express this equation is by considering the deduction in
eective thrust, which leads to the following expressing:
R
T
= T
p
(1 t
d
), (5.33)
where t
d
is the thrust deduction factor.
The speed of advance of the propeller is generally slower than the ship speed
due to the ships wake. The retardation of the wake is expressed in the wake
fraction, dened as:
w
r
= 1
V
A
V
s
. (5.34)
Typically, 0 < w
r
< 0.4; the larger value applies for very full hulls. Now a new
eciency can be dened, the hull eciency. This is not really an eciency, since
the values can be bigger than one. The hull eciency, which is the ratio of the
eective power to the thrust power, can be dened as:
H
=
P
E
P
T
=
R
T
V
s
T
p
V
A
=
1 t
d
1 w
r
. (5.35)
The last part of the propulsive eciency is the relative rotative eciency, which
is again not really an eciency, since the values can be bigger than one. This
accounts for the dierences in torque absorption characteristics of a propeller
when operating in mixed wake and open water ows. This value lies close to one
in most cases and is generally within the range 0.96
r
1.04. This eciency
is dened as:
r
=
C
Q
o
C
Q
b
, (5.36)
5.5 Thrusters 36
with C
Q
o
is the torque coecient of the propeller in open water and C
Q
b
is the
torque coecient of the propeller working in the wake of the ship.
Now after all these denition one can dene the propulsive eciency:
D
=
H

0

R
(5.37)
The propulsive eciency is generally greater than the open-water propeller ef-
ciency.
16
16
This section is based on three books which have similar denitions, but all describe the
matter in their own way. From: [2] p.37-40,62-64; [8] p.24-26,32-34; [4] p.85-86,290-292.
37
Chapter 6
Modeling
Two models will be created, one with DP and one without this system. Both
models have the external forces as input and all the information of the ship. In
the models the thruster conguration will be optimized in such a way that the
energy used by the thrusters is minimal. The rst model that is created in this
three month literature study is not very realistic. There are a lot of limitations
in this rst model, but the model is created to develop a working algorithm to
minimize the power the thrusters produce. From this position the model will be
changed and expanded in the last six months of this project.
Which forces are very important and which are less important in dierent situ-
ations can be found in Appendix A gure A.5.
6.1 Power Model
6.1.1 Without DP
Input:
Current, wind, wave forces (direction and size)
Current direction and velocity
Number of Thrusters and position
Important information of the ship (size, weight etc)
Direction and velocity ship (from joystick ?)
In Program:
Optimization least energy thruster conguration
Thrusters can not give more power than their maximum
Boundaries of azimuth thruster to prevent loss of eciency by thrusters
that work against each other
6.1 Power Model 38
Output:
RPM thrusters
direction of azimuth thrusters
6.1.2 With DP
It is needed on beforehand which operating modes
1
are possible in this vessel.
Only those modes are available in the program.
Input:
Current, wind wave forces (direction and size)
Current position vessel
Number of thrusters, position and type (which are azimuth thrusters?)
Mode:
manual/joystick: model without DP
auto-heading: heading needed
auto-position: position needed
auto area position: area vessel needs to be in
auto-track: way-points of course to follow
auto-pilot: course to follow
follow target: position target (every x seconds refreshed)
Type of Position Reference System (best might be (D)GPS. Others need
receivers on seabed etc.)
Optional: Max time to get on position/track
In Program:
Optimization least energy thruster conguration
Thrusters can not give more power than their maximum
Boundaries of azimuth thruster to prevent loss of eciency by thrusters
that work against eachother
Use of PRS
Updates of changes in environment
1
see Chapter 3.2.3
6.2 Programming 39
Output:
RPM thrusters
Direction of Azimuth Thrusters
Optional: time it will take to get on position/track?
6.2 Programming
6.2.1 A rst model
In gure 6.1 the orientation of the axis, the moment and the forces is given. The
moment orientation is taken as clockwise. Using this, one can dene a model for
the ship simulation.
Figure 6.1: The coordinate system.
Input:
Vector with wind force in x direction, wind force in y-direction and moment
of wind.
Vector with current force in x direction, current force in y-direction and
moment of current.
Vector with wave force in x direction, wave force in y-direction and moment
of wave.
Force (x and y direction, or resulting and angle) and moment that are
wanted by the driver of the vessel
Matrix of position of thrusters and forces now.
Vector of maximum force that a thruster can give.
Error tolerance
In program:
1. Calculate the force needed in x-direction:
F
x
= F
x
wanted
F
x
wind
F
x
current
F
x
wave
. (6.1)
6.2 Programming 40
2. Calculate the force needed in y-direction:
F
y
= F
y
wanted
F
y
wind
F
y
current
F
y
wave
. (6.2)
3. Calculate the moment needed:
M = M
wanted
M
wind
M
current
M
wave
. (6.3)
4. Calculate the force the thrusters provide in the current conguration in
x-direction: (F
x
)
now
=
n
i=1
(F
x
)
thruster
i
.
5. Calculate the force the thrusters provide in the current conguration in
y-direction: (F
y
)
now
=
n
i=1
(F
y
)
thruster
i
.
6. Calculate the Moment the thrusters provide in the current conguration:
(M)
now
=
n
i=1
y
i
(F
x
)
thruster
i
+x
i
(F
y
)
thruster
i
.
7. If the forces and moment needed are almost the same as the forces and
moment given, then keep this conguration: if |F
x
(F
x
)
now
| <
1
and
|F
y
(F
y
)
now
| <
1
and |M (M)
now
| <
2
.
8. If they are not almost the same then a new conguration must be found.
6.2.2 Method to nd the new conguration
With F
x
, F
y
and M the forces and moment needed from the thrusters, the
following formulas must hold (with n the number of thrusters, with subscript i
short for thruster i and with x
i
x-coordinate of the position of thruster i and
y
i
y-coordinate of the position of thruster i.)
1
:
F
x
=
n
i=1
(F
x
)
i
, (6.4)
F
y
=
n
i=1
(F
y
)
i
, (6.5)
M =
n
i=1
(y
i
(F
x
)
i
+x
i
(F
y
)
i
). (6.6)
Rewriting equation (6.4) and (6.5) to express (F
x
)
n
and (F
y
)
n
as functions of
the other forces leads to:
(F
x
)
n
= F
x

n1
i=1
(F
x
)
i
, (6.7)
(F
y
)
n
= F
y

n1
i=1
(F
y
)
i
. (6.8)
1
This section is based on [23, 24].
6.2 Programming 41
Substituting (6.7) and (6.8) into (6.6) leads to:
M =
n1
i=1
(y
i
(F
x
)
i
+x
i
(F
y
)
i
) y
n

_
F
x

n1
i=1
(F
x
)
i
_
+x
n

_
F
y

n1
i=1
(F
y
)
i
_
=
n1
i=1
((y
n
y
i
) (F
x
)
i
+ (x
i
x
n
) (F
y
)
i
) y
n
F
x
+x
n
F
y
. (6.9)
Now assume that x
n
= x
n1
.
2
Now rewrite (6.9) to express (F
y
)
n1
as a function
of the other forces:
(F
y
)
n1
=
M +y
n
F
x
x
n
F
y

n1
i=1
((y
n
y
i
) (F
x
)
i
)
n2
i=1
((x
i
x
n
) (F
y
)
i
)
x
n1
x
n
=
M +y
n
F
x
x
n
F
y
x
n1
x
n
. .
s
+
n1
i=1
_
_
_
_
_
y
i
y
n
x
n1
x
n
. .
c
1
(i)
(F
x
)
i
_
_
_
_
_
+
n2
i=1
_
_
_
_
_
x
n
x
i
x
n1
x
n
. .
c
2
(i)
(F
y
)
i
_
_
_
_
_
.
(6.10)
So this means:
(F
y
)
n1
= s +
n1
i=1
c
1
(i)(F
x
)
i
+
n2
i=1
c
2
(i)(F
y
)
i
, (6.11)
(F
y
)
n
= F
y

n2
i=1
(1 +c
2
(i)) (F
y
)
i
s
n1
i=1
c
1
(i)(F
x
)
i
.
Now it is important that a thruster conguration is found in which the total
power is minimized. The formula for the total power is:
g((F
x
)
1
, . . . , (F
x
)
n1
, (F
y
)
1
, . . . , (F
y
)
n2
) =
n
i=1
_
(F
x
)
2
i
+ (F
y
)
2
i
. (6.12)
To minimize this, one needs to nd the forces for which the partial derivatives
of the power are zero, i.e.:
_
_
g
(F
x
)
1
= 0,
.
.
.
.
.
.
.
.
.
g
(F
x
)
n1
= 0,
g
(F
y
)
1
= 0,
.
.
.
.
.
.
.
.
.
g
(F
y
)
n2
= 0.
(6.13)
2
If they are the same, one can also take another thruster and just rewrite the order of the
thrusters!
6.2 Programming 42
From (6.12) it is easy to see that:
g
(F
x
)
i
=
n
j=1
_
_
(F
x
)
j

(F
x
)
j
(F
x
)
i
_
(F
x
)
2
j
+ (F
y
)
2
j
+
(F
y
)
j

(F
y
)
j
(F
x
)
i
_
(F
x
)
2
j
+ (F
y
)
2
j
_
_
, (6.14)
g
(F
y
)
i
=
n
j=1
_
_
(F
x
)
j

(F
x
)
j
(F
y
)
i
_
(F
x
)
2
j
+ (F
y
)
2
j
+
(F
y
)
j

(F
y
)
j
(F
y
)
i
_
(F
x
)
2
j
+ (F
y
)
2
j
_
_
. (6.15)
Equation 6.14 and 6.15 can be simplied since n 3 of the derivatives in the
summation are zero, as can be seen from the following equations:
For j = 1 . . . n 2, j = i :
(F
x
)
j
(F
x
)
i
= 0 and
(F
y
)
j
(F
y
)
i
= 0.
For j = 1 . . . n 2, j = i :
(F
x
)
j
(F
x
)
i
=
(F
x
)
i
(F
x
)
i
= 1 and
(F
y
)
j
(F
y
)
i
=
(F
y
)
i
(F
y
)
i
= 1.
For i = 1 . . . n 2 :
(F
y
)
n1
(F
x
)
i
= c
1
(i),
(F
y
)
n
(F
x
)
i
= c
1
(i),
(F
y
)
n1
(F
y
)
i
= c
2
(i) and
(F
y
)
n
(F
y
)
i
= (1 +c
2
(i)).
For i = 1 . . . n 1(from equation 6.7) :
(F
x
)
n
(F
x
)
i
= 1.
This leads to:
For i = 1 . . . n 1 :
g
(F
x
)
i
=
(F
x
)
i
_
(F
x
)
2
i
+ (F
y
)
2
i
+
c
1
(i)(F
y
)
n1
_
(F
x
)
2
n1
+ (F
y
)
2
n1
(F
x
)
n
+c
1
(i)(F
y
)
n
_
(F
x
)
2
n
+ (F
y
)
2
n
,
(6.16)
For i = 1 . . . n 2 :
g
(F
y
)
i
=
(F
y
)
i
_
(F
x
)
2
i
+ (F
y
)
2
i
+
c
2
(i)(F
y
)
n1
_
(F
x
)
2
n1
+ (F
y
)
2
n1
(1 +c
2
(i))(F
y
)
n
_
(F
x
)
2
n
+ (F
y
)
2
n
.
(6.17)
It is not easy to solve the non-linear system (6.13). How to solve this system
is explained later in this chapter, but when the solution has been found it is
important to compare the forces that are found with the maximum forces of
every thruster. If the maximum force of a thruster is lower than the force needed
in the optimal conguration, one should run the algorithm again, but this time
with the forces of all the thrusters, that should be higher than their maximum
3
in the optimal conguration, held constant at their maximum. In this way one
nds a feasible optimal solution. This can be described as the following iteration
scheme:
3
In a rst implementation the maximum force is dened in an unrealistic way as a maximum
force in x- and y-direction. In reality this maximum is just a maximum rotations per minute
of the thrusters. The method will be extended later on.
6.2 Programming 43
1. Minimize g(F
i
). Say F
is the thruster conguration that gives the least

total force.
2. Compare F
i
to (F
max
)
i
. For all i {1, 2, . . . , 2n 3}: if F
i
> (F
max
)
i
:
F
i
= (F
max
)
i
.
3. If no forces are above max, stop. Else: minimize g(F
k
, F
i
) where F
k
are
the forces that are not at their max yet and F
i
are the forces that are held
constant at their maximum. Now one only nds values for F
k
. The F
i
are
not changed.
4. Repeat step 2.
To solve the non-linear system one can use the BFGS, Broyden-Fletcher-
Goldfarb-Shanno, method. This method will be explained in the next subsec-
tion.
4
6.2.3 Minimization methods
BFGS Method
To minimize the non-linear function f, one needs a starting point x
(0)
and a
starting symmetric positive denite matrix, H
0
. This matrix is mostly chosen
as a positive multiple of the identity matrix.
In fact, this method is a quasi-newton method. In Newtons method one uses the
Jacobian to optimize the search direction. It is expensive to compute and invert
this Jacobian, so to avoid this, one tries in this BFGS Method to approximate
the Jacobian. In every step the approximation gets better. The convergence of
Newtons method is quadratic, so this is a very nice characteristic. That is the
reason why one tries to approximate this method.
5
The iteration for this method is:
1. x
(k+1)
= x
(k)
H
1
k
f(x
(k)
),
2. s
(k)
= x
(k+1)
x
(k)
,
3. y
(k)
= f(x
(k+1)
) f(x
(k)
), (6.18)
4. H
(k+1)
= H
k

H
k
s
(k)
(s
(k)
)
T
(H
k
)
T
(s
(k)
)
T
H
k
s
(k)
+
y
(k)
(y
(k)
)
T
(y
(k)
)
T
s
(k)
In this method it is also possible to write the rst two steps as:
1. H
k
s
(k)
= f(x
(k)
),
2. x
(k+1)
= x
(k)
+s
(k)
. (6.19)
In this way no inverse has to be calculated and a faster method can be applied
to solve the rst equation for s
(k)
.
Convergence can be checked in every step by calculating the norm ||f(x
(k)
||.
The iteration is nished when the norm has become suciently small. One
4
From: http : //www.math.mtu.edu/ msgocken/ma5630spring2003/lectures/global2/global2/node8.html
and http : //en.wikipedia.org/wiki/BFGS method
5
See [13] p.80-82
6.2 Programming 44
should keep in mind that this method only works when (s
(k)
)
T
y
(k)
> 0 holds in
every step. This can be ensured by the use of an appropriate linesearch.
6
In the version of the method described in equation (6.18) no line-search will be
applied. With a linesearch the method changes slighty and one should determine
(k)
in every step. The determination of
(k)
is described later on in this chapter.
The adaptation for a linesearch is changing equation (6.19) by:
2. x
(k+1)
= x
(k)
+
(k)
s
(k)
(6.20)
Gradient descent method
Another method that can be used is the gradient descent method. This method
is also called steepest descent method.
The method has the following scheme:
x
(k+1)
= x
(k)
f(x
(k)
). (6.21)
This can be calculated every step by performing a linesearch, or one can choose
one for every step, but this might mean that the iteration will take a long
time. If is chosen the same in every step then the new x is not guaranteed to
be nearer a solution than the previous one.
Linesearch
To determine the steplength in equation (6.21) is by letting =
m
, where
(0, 1) and m 0 is the smallest nonnegative integer such that the following
inequality, the Armijo rule, holds:
7
f(x
(k)
f(x
(k)
)) f(x
(k)
) < ||f(x
(k)
)||
2
. (6.22)
This parameter is typically set to 10
4
.
A linesearch algorithm is:
1. For k = 1, . . . , k
max
(a) Compute f and f. Test for termination.
(b) Find the least integer m 0 such that equation (6.22) holds for
=
m
.
(c) x = x f.
2. The algorithm stops when the termination criterion is fullled or when
k
max
iterations are done unsuccesfully.
6
From: http://en.wikipedia.org/wiki/Line search
7
This section is from [7] chapter 3.
6.2 Programming 45
Gradient descent with linesearch
A nice combination of the previous two methods is proposed by Vuik in [24].
The algorithm is searching for the
opt
such that x
new
= x
old
f is minimal.
The search for
opt
can be dened by the following algorithm:
1. Start with
min
= 0 and
max
= 1.
Calculate f
i
= f(x
old
i
f(x
old
)) with i = {min, max}.
If f
max
> f
min
STOP. Else set
min
=
max
and multiply
max
with 2
until f
max
> f
min
. Now it holds
min

opt

max
.
2. Split the interval in two such that
opt
remains between
min
and
max
until the distance between
min
and
max
satises the termination crite-
rion.
There are also two other methods possible to nd the zero of a function.
Bisection method
One needs to nd a zero of
().
Input: > 0 accuracy parameter;
(0)
,
(1)
are given such that
(
(0)
) < 0 and
(
(1)
) > 0 (which of course means that a zero is in between if the function is
continuous).
Step 1: If |
(0)
(1)
| < STOP.
Step 2: Let =
1
2
(
(0)
+
(1)
);
Step 3: If
() < 0 then
(0)
:= ; GOTO Step 1.
Step 4: If
() > 0 then
(1)
:= ; GOTO Step 1.
The function
() does not need to be dierentiable, as opposed to the next

method.
8
Newtons method
In this method one needs to nd the zero of
() or minimize the quadratic

function (this yields in most cases the same result): q() = (
(k)
)+
(
(k)
)(
(k)
) +
1
2
(
(k)
)(
(k)
)
2
.
Input: > 0 is the accuracy parameter;
(0)
is the given initial point; k=0;
Step 1: Let
(k+1)
=
(k)
(
(k)
)
(
(k)
)
.
Step 2: If |
(k+1)
(k)
| < STOP.
Step 3: k := k + 1, GOTO Step 1;
The method as presented here might be unstable. The minimum of this approx-
imation can be further from the minimum of than the previous point itself,
but if not too far from the mimimum point, this method converges, as said be-
fore, quadratically.
9
This is actually in one dimension, for multiple dimension
functions the rst step changes to:
8
For Bisection method see [13] p.82-83
9
For Newtons method see [13] p.83
6.2 Programming 46
Step 1: Let
(k+1)
=
(k)
J((
(k)
))(
(k)
),
where J stands for the jacobian matrix of , which is:
_
1
x
1

1
x
n
.
.
.
.
.
.
.
.
.
n
x
1

n
x
n
_
_
In the second step the norm of the dierence is taken.
47
Chapter 7
Results
The model from the previous chapter was implemented in Matlab. The maxi-
mum thruster forces were not taken into account. Only equation 6.13 was solved,
using the method described in Section 6.2.3.
The ship was build considering 20% of the length as the bow. The thrusters
are arranged symmetrically. The moments from the thrusters are calculated in
relation to the center of gravity (CoG). Some resulting gures can be found in
this paragraph, with 3, 4, 5, 6, 10, 20 and 30 thrusters, for a ship with length
100m and width 20m in the gures 7.1-7.7.
The forces working on this ship are:
1. Current -5N in x-direction, 5N in y-direction,
2. Wave: 7N in x-direction, 7N in y-direction,
3. Wind: 3N in x-direction, 8N in y-direction.
The wanted resulting force for the ship is -2N in x-direction and 10N in
y-direction.
The external forces are assumed to work on the CoG, so the moments of the
external forces are 0. The demanded moment for the ship is also 0.
The energy of the optimal congurations can be found in table 7.1.
48
Number of Thrusters Energy in optimal conguration
3 12,34
4 12,21
5 12,22
6 12,44
10 12,37
20 12,22
30 12,91
Table 7.1: Number of thrusters on a ship with length 100m and width 20m versus
the energy in the optimal conguration. The congurations are illustrated in the
gures 7.1-7.7.
Figure 7.1: Conguration of a ship with dimensions 100x20m with 3 thrusters.
49
50
51
52
Chapter 8
Evaluation and Future
Goals
8.1 Evaluation of rst three months
In this 3-month study the formulas for the external forces working on the ship
were found. Furthermore, a rst model was created in which forces, position
of thrusters and ship size where all dened in an articial way. The thruster
conguration was optimized in such a way that the energy of the thrusters was
minimal. All the programming was done in Matlab. The ship was assumed to
be a pointmass, so none of the external forces was given a moment. Since the
thrusters did have a torque, this was a small inconsistency, so the forces were
handled as if they worked on the center of gravity, which meant they had no
moment.
Also a lot of literature was studied. Some very nice articles were also quoted in
the bibliography, but were not cited in this paper; they might be useful in the
future.
The rst three months a little modeling and implementation was done. This
was characterized by ctitious ships with ctitious thruster positions, but the
main goal was to create an optimization program that found the optimum
conguration in a small amount of CPU time. At rst two optimization
algorithms were implemented, but these failed to converge for more than 2
thrusters, so nally another method was used. With this the goal has been
accomplished. From this position on the programs will be made more realistic
in the nal six months of the project.
The project description has as objective:
1. Steering ships in the Ship Simulator game with azimuth thrusters in a
realistic way;
2. Dynamic Positioning of ships with and without azimuth thrusters in the
game and in training applications.
8.1 Evaluation of rst three months 53
The approach is:
1. Study the basics of ship hydrodynamics, azimuth thrusters propulsion,
and DP;
2. Learn Quest3D and Newton Dynamics, as it is implemented in Quest3D
as a plug-in called Newton for Quest (NfQ).
3. Investigate the ship dynamics solutions as implemented in Ship Simulator
by VSTEP.
4. Investigate current methods and tools of calculating forces on ships that
result from current and wind.
5. Design and implement a solution for azimuth thrusters propulsion in Ship
Simulator.
6. Design and implement a solution for DP of vessels in Ship Simulator under
various weather, wave and current conditions.
The approach from the project description was partly done. The basics of ship
hydrodynamics, thruster propulsion and DP were studied. The ship simulator
solutions were investigated and methods for calculating forces on ships as well.
The tutorials of Quest3D were studied and some features of NfQ were explored.
Now it is important to put more realism in the models that are created until
now.
8.2 Future Goals 54
8.2 Future Goals
The last six months of this project will be dedicated to make the model more
realistic.
From the project description, the following subgoals can be extracted:
1. Dene thrust not as thrust in x and y direction, but as a resulting thrust
and angle.
2. Dene a maximum thrust (speed in rpm) per thruster.
3. Optimization with non azimuth thrusters and combinations. This can be
done by dening the angle as a xed value.
4. Go from Matlab to C++. After this it might be necessary to check certain
algorithms with Matlab rst, before implementing in C++.
5. Sailing with and without DP. Take the inertia of the ship into account.
After the forces are calculated realistically, the sailing should be adjusted
of course.
6. Use the formulas for the forces. Find the constants somewhere. These are
the formulas from Chapter 5.
7. Finite length of the ship. This means the forces on the ship should inte-
grated or a vector should be dened with values of the forces for dierent
parts of the ship.
8. With the nite length the forces induce also a moment on the ship.
9. In reality it costs time to speed up or slow down the thrusters and also to
rotate the azimuthing thrusters. This should be taken into account.
10. Penalty for reverse thrust of every thruster.
11. Implement test cases with real ships and known forces.
12. Create with the programmers from VSTEP an interface with information
for the skipper like for instance conning displays (dierent conning displays
can be found in Appendix A gures A.1-A.3).
If there is time enough, the implementation will be extended with more complex
subjects such as circular waves. Furthermore, some other optimization tech-
niques will be used if necessary.
55
Appendix A
Figures
Figure A.1: A picture of a conning display.
1
1
From: http://www.msidesign.se/designs/conning.jpg
56
2
3
2
From: http://www.ami-gfv.com/image2/display1.jpg
3
From: http://www.l-3klein.com/navigation/nautoconning/conning display.jpg
57
Figure A.4: The terminology used often when speaking of ships is illustrated in
this picture.
4
4
From: http://navsci.berkeley.edu/ns12b/Presentations/Ship%20Operations/K%20-
%20Shiphandling.ppt
58
Figure A.5: The importance of the dierent external inuences in specic sailing
conditions and applications.
5
Scan from [3], p. 15
59
Figure A.6: This is an example of a mathematical model behind DP.
5
60
Appendix B
Testproblem
In this testproblem the following forces are working on the ship:
In x-direction: wind = 3, current = -5 and wave = 7.
In y-direction: wind = 8, current = 5 and wave = 7.
The moment for all the external forces is taken 0 for simplicity.
The force demanded by the skipper is -2 in x-direction and 10 in y-direction.
The moment is 0.
There are two thrusters and their positions are: (-6,-50) and (6,-50).
The coordinate system of importance here is the body-xed coordinate system.
Now from equations 6.1, 6.2 and 6.3 one can calculate F
x
, F
y
and M:
F
x
= F
x
wanted
F
x
wind
F
x
current
F
x
wave
= 2 3 (5) 7 = 7,
F
y
= F
y
wanted
F
y
wind
F
y
current
F
y
wave
= 10 8 5 7 = 10,
M = M
wanted
M
wind
M
current
M
wave
= 0 0 0 0 = 0.
For the moment the following formula should hold: M = y
1
(F
x
)
1
+ x
1

(F
y
)
1
y
2
(F
x
)
2
+x
2
(F
y
)
2
= 50 (F
x
)
1
6 (F
y
)
1
+ 50 (F
x
)
2
+ 6 (F
y
)
2
.
With this the following minimization problem can be dened to minimize the
energy of the thrusters:
min
_
(F
x
)
2
1
+ (F
y
)
2
1
+
_
(F
x
)
2
2
+ (F
y
)
2
2
. (B.1)
With :
(F
x
)
1
+ (F
x
)
2
= F
x
= 7, (B.2)
(F
y
)
1
+ (F
y
)
2
= F
y
= 10, (B.3)
50 (F
x
)
1
6 (F
y
)
1
+ 50 (F
x
)
2
+ 6 (F
y
)
2
= M = 0. (B.4)
Multiplying (B.2) with -50 and add this up with (B.4) yields:
50(F
x
)
1
50(F
x
)
2
= 350
50(F
x
)
1
+ 50(F
x
)
2
6(F
y
)
1
+ 6(F
y
)
2
= 0 +
6(F
y
)
1
+ 6(F
y
)
2
= 350. (B.5)
61
Multiplying (B.3) with 6 and add this up with (B.5) yields:
6(F
y
)
1
+ 6(F
y
)
2
= 350
6(F
y
)
1
+ 6(F
y
)
2
= 60 +
12(F
y
)
2
= 290 (F
y
)
2
=
290
12
. (B.6)
Substituting (B.6) into (B.3) yields:
(F
y
)
2
+
290
12
= 10 (F
y
)
1
=
410
12
. (B.7)
One can rewrite equation (B.2):
(F
x
)
1
= 7 (F
x
)
2
. (B.8)
Now substitute (B.6), (B.7) and (B.8) into (B.1) and the minimization problem
is transformed into:
min
_
(7 (F
x
)
2
)
2
+ (
410
12
)
2
+
_
(F
x
)
2
2
+ (
290
12
)
2
min
_
(F
x
)
2
2
+ 14(F
x
)
2
+
43789
36
+
_
(F
x
)
2
2
+
21025
36
(B.9)
To minimize this function, one should search for the point were the derivative
is zero. Taking the derivative of (B.9) leads to:
(F
x
)
2
+ 7
_
(F
x
)
2
2
+ 14(F
x
)
2
+
43789
36
+
(F
x
)
2
_
(F
x
)
2
2
+
21025
36
= 0. (B.10)
To nd this zero one could use algorithms like Newtons method or bisection
method. In this case equation (B.10) was solved by maple, yielding as an answer:
(F
x
)
2
=
29
10
. Using this answer in (B.8) leads to: (F
x
)
1
= 7 +
29
10
=
41
10
.
The optimal conguration of the thrusters is then:
F
1
=
_
41
10
,
410
12
_
(4.1, 34.16666667) ,
F
2
=
_
29
10
,
290
12
_
(2.9, 24.16666667) .
With the minimization methods from Chapter 6.2.3 code from Matlab is created.
From this Matlab calculates:
F
1
= (4.0937, 34.1667) ,
F
2
= (2, 9067, 24.1667) .
As can be seen the congurations are almost the same, the dierence is approx-
imately 0.01.
The conguration can also be seen in the plot from the matlab program in gure
(B.1).
62
Figure B.1: Optimal solution from Matlab program with two thrusters.
BIBLIOGRAPHY 63
Bibliography
[1] Barltrop, N.D.P. (1998). Floating structures: a guide for design and analy-
sis. Volume 2, Ledbury, OPL.
[2] Bertram, V. (1999). Practical Ship Hydrodynamics. 2nd ed. Oxford, But-
terworth Heinemann.
[3] Bray, D. (2003). Dynamic Positioning. 2nd ed. Ledbury, OPL.
[4] Carlton, J.S. (1994). Marine Propellers & Propulsion. Oxford, Butterworth
Heinemann.
[5] Dokkum, K. van (2003). Scheepskennis. 2nd ed. Enkhuizen, Dokmar.
[6] Ghose, J.P. and Gokarn, R.P. (2004). Basic Ship Propulsion. New Delhi,
Allied Publishers Limited.
[7] Kelley, C.T. (1999). Iterative Methods for Optimization.
http://www.siam.org/books/textbooks/download.php
[8] Newman, J.N. (1977). Marine Hydrodynamics. Cambridge, MIT Press.
[9] Rawson, K.J. and Tupper, E.C. (2001). Basic Ship Theory. Volume 1, 5th
ed., Oxford, Buttersworth-Heinemann.
[10] Rawson, K.J. and Tupper, E.C. (2001). Basic Ship Theory. Volume 2, 5th
ed., Oxford, Buttersworth-Heinemann.
[11] Timman, R., Hermans, A.J. and Hsiao, G.C. (1985). Water waves and
ship hydrodynamics: an introduction. Dordrecht, Nijho.
Lecture Notes
[12] Journee, J.M.J. and Massie, W.W. (2001). Oshore Hydromechanics.
Lecture notes.
(http://www.ocp.tudelft.nl/mt/journee/Files/Lectures/OshoreHydromechanics.pdf)
[13] Klerk, E. de, Roos, C. and Terlaky, T. (2003). Nonlinear Optimization.
Lecture notes, Delft University of Technology.
[14] Lagers, G.H.G., Bloemen, P.C.G. and Haakman, N.P. (2004). Dynamic
Positioning. Lecture notes, Delft University of Technology.
BIBLIOGRAPHY 64
Articles
[15] Deter D., (1997). Principal Aspects of Thruster Selection, In: Dynamic
Positioning Conference; 1997 Oct 21-22; Houston USA.
[16] Fossen, T.I., Sagatun, S.I. and Srensen, A.J. (1995). Identication Of Dy-
namically Positioned Ships, In: Control Eng. Practice, Vol.4, No.3, pp.369-
376.
[17] Holvik, J. (1998). Basics of Dynamic Positioning, Dynamic Positioning
Conference; 1998 Oct 13-14; Houston USA, Kongsberg Simrad Inc. (Hous-
ton).
[18] Liang, C.C., Teng, T.L. and Cheng, W.H. (1999). An Eective Control
System of Thrust and Moment Modulation for a Dynamically Positioned
Vessel, In: Marine Geodesy, 22:4, pp 211-227.
[19] Liang, C.C. and Cheng, W.H. (2002). The optimum control of thruster sys-
tem for dynamically positioned vessels, In: Ocean Engineering, 31-(2004),
pp 97-110.
[20] Mahfouz, A.B. (2006). Predicting the capability-polar-plots for dynamic
positioning systems for oshore platforms using articial neural netwerks,
In: Ocean Engineering 34-(2007), pp 1151-1163.
[21] Sinding, P. and Andersen, S.V. (1998). A Force Allocation Strategy for
Dynamic Positioning, In: Proceedings of the Eighth (1998) International
Oshore and Polar Engineering Conference; 1998 May 24-28; Montr`eal
Canada, pp. 346-353.
[22] Srdalen, O.J. (1997). Optimal Thrust Allocation For Marine Vessels,
In: Control Eng. Practice, Vol.5, No.9, pp.1223-1231.
[23] Vuik, C. (2007). REPORT 07-03 The Stadyc model for the movement of
a ship by Multi Thrusters, Delft University of Technology.
[24] Vuik, C. and Benchev, I. (2007). REPORT 07-02 The static model for
the Multi Thruster Control System without constraints, Delft University of
Technology.
INDEX 65
Index
Azimuth, see Thruster, Azimuth
thruster
Coordinate System, 4, 3132, 39
Forces, 2731, 3743, 6061
Current force, 2930, 3739
Drag force, 27
Lift force, 28
Reaction forces, 31
Wave force, 3031, 3739
Cyclic, 3031
Wave drift force, 3031
Wind force, 2829, 3739
Dynamic wind eects, 29
Power law prole, 29
Wind spectra, 29
Hydromechanics, 2533
Archimedes law, 25
Bernouilli equation, 27
Continuity equation, 25
Euler equations, 26
Euler equations with Potential, 27
Hydrostatic pressure, 25
Laplace equation, 26
Law of buoyancy, 25
Navier-Stokes equation, 26
Potential Function, 26
Simplied continuity equation, 26
Modeling, 3746
Power Model, 37
With DP, 3839
Without DP, 3738
Mooring Systems
Anchoring, 9
Dynamic Positioning, 814
Denitions, 8
DP System, 12
Mathematical model, 13
Training, 9
Jack-up barge, 9
Operating modes, 1011, 38
Auto area position mode, 11, 38
Auto track mode, 11, 38
Auto-heading mode, 10, 38
Auto-position mode, 10, 38
Autopilot mode, 11, 38
Follow target mode, 11, 38
Manual/joystick mode, 10, 38
Optimization, 37, 38, 4146, 6061
BFGS Method, 4344
Bisection method, 45, 61
Gradient descent method, 4445
Linesearch, 4445
Newton Method, 43, 4546, 61
Non-Linear System, 42, 43
Position Reference Systems, 1314, 38
(D)GPS, 14
Hydro-acoustic, 13
Taut-wire/rise inclinometers, 13
Ship, 3, 37
Containership, 5
Cruiseships, 5
History, 3
Oshore structures, 5
Ship Moves, 4, 33
Heave, 4, 33
Pitch, 4, 33
Roll, 4, 33
Surge, 4, 33
Sway, 4, 33
Yaw, 4, 33
Tankers, 5
Tugboats, 6
Thrust, 3334
Thrust deduction coecient, 35
Thruster, 67, 3343, 6061
Azimuth thruster, 7, 3739
INDEX 66
Eciency, 3436
Hull eciency, 35
Open water eciency, 34, 36
Propulsive eciency, 3436
Relative rotative eciency, 35
Interaction, 3436
Main propellers, 6
Tunnel thruster, 67
Torque, 3334
Wake fraction, 35
Wave
elevation, 31, 33
encounter, 3233

Dynamic Positioning Simulator

Uploaded by

Copyright:

Available Formats

Dynamic Positioning Simulator

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dynamic Positioning Simulator

Uploaded by

Copyright:

Available Formats

What are the different types of ships discussed in the report?

What are the different thruster types used for dynamic positioning?

Dynamic Positioning Simulator

(nabla): volume of the submerged part of the ship.

: phase angle for roll movement.

: phase angle for pitch movement.

: phase angle for yaw movement.

and since sometimes

+gz = C(t), (5.10)

is the wave number, is the wave

is the thruster conguration that gives the least

() does not need to be dierentiable, as opposed to the next

() or minimize the quadratic

You might also like