Professional Documents
Culture Documents
Performed at Seaware AB
MARTIN ALEXANDERSSON marale@kth.se Master Thesis, KTH Centre for Naval Architecture STOCKHOLM JANUARY 2009
ABSTRACT
This thesis is an overview of added resistance in waves, available methods and how a simplified method can be developed. Added resistance in waves is indeed a very complicated problem. The present methods, most of them developed during the 1970s, give reasonable results for head waves. But for quartering and following waves, the methods have low accuracy, due to uncertainties in strip calculated ship motions for these cases, and the ability to account for roll and yaw motions. In this thesis Gerristma and Beukelmans method and Boeses method to calculate added resistance is implemented and validated against published calculation results, from previous implementations. The implementations are also evaluated against published experimental results in head waves. These evaluations show that added resistance with good accuracy can be calculated for head waves using Gerritsma and Beukelmans method. This thesis also shows how a simplified method to calculate added resistance in waves can be developed. This simplified method only uses the ships main particulars to describe the hull. The method is intended to be used when the entire hull geometry is not available, for instance in the pre-study of a new ship design. The idea with this method is to estimate transfer functions for added resistance without using strip calculations. The transfer functions are parameterized with three parameters, peak value, peak frequency and spreading. Expressions for these parameters are derived with regression analysis, based on analytical results from Gerritsma and Beukelmans method. The simplified method has an accuracy of about 25%, which is about the same amount that usually can be expected of a method to calculate added resistance in waves.
SAMMANFATTNING
Denna rapport r en versikt av adderat vgmotstnd, tillgngliga metoder och en beskrivning av en frenklad metod. Adderat vgmotstnd r ett vldigt komplicerat problem. De metoder som anvnds idag, utvecklades under 70-talet och ger rimliga resultat fr mtande vgor. Metoderna ger dock smre resultat d vgorna faller in snett, frn sidan eller bakifrn. Detta beror p att de strip-berknade fartygsrrelserna r oskra fr dessa vgriktningar och att metoderna inte tar hnsyn till fartygets rullning eller att det girar i vgorna. I denna rapport implementeras Gerritsma & Beukelmans metod och Boeses metod. Implementationerna jmfrts med publicerade resultat frn tidigare implementationer samt publicerade experiment-resultat, fr mtande vgor. Dessa jmfrelser bekrftar att Gerritsma & Beukelmans metod kan anvndas fr att berkna adderat vgmotstnd fr mtande vgor. Detta examensarbete frslr en frenklad metod fr att berkna adderat vgmotstnd. Metoden anvnder endast enkla fartygsparametrar fr att beskriva fartygets skrov. Den frenklade metoden r tnkt att anvndas d ett fartygs exakta skrovgeometri inte r knd, till exempel vid en frprojektering av en ny fartygsdesign. Iden med metoden r att den ska generera transferfunktioner fr adderat vgmotstnd, utan att anvnda stripberkningar. Transferfunktionerna beskrivs med tre parametrar, pikvrde, pikfrekvens och spridning. Matematiska uttryck fr dessa parameterar hrleds med regressionsanalys, baserat p analytiska resultat frn Gerritsma & Beukelmans metod. Den frenklade metoden har en exakthet p ungefr 25%, vilket r den exakthet som normalt kan frvntas av en metod fr att berkna adderat vgmotstnd.
PREFACE
Welcome to this master thesis! This is a Master Thesis at the Royal Institute of Technology, KTH, Centre for Naval Architecture. This master thesis project was performed at the company Seaware AB in Stockholm, Sweden. Seaware AB develops computer programs for weather routing and operational decision support. During my work with this master thesis project I have increased my knowledge about ship hydrodynamics and added resistance in particular. I have understood that this is a very complicated field where a lot of work still can be done. I also realize that I have a lot left to learn, and that there are a lot of aspects to the problem that I havent studied. I have tried to create this thesis as if it would be one that I myself had wanted to read, when I started out the literature study in this project. I would also like to thank: Mikael Palmquist, my supervisor at Seaware AB, and Karl Garme, my supervisor at KTH. I hope you will enjoy this master thesis! Stockholm, January 2009 Martin Alexandersson
2
2.1 3 4 5 5.1
6.3
Approximation of RAW , p ( , RAW , p , Head ) ......................................................................................................44 Approximation of cnorm ( , cnorm, Head ) ..........................................................................................................46
Conslusions.....................................................................................................................................................46 Summary and Discussion ........................................................................................... 47 Suggested reading ..........................................................................................................................................47 References ..................................................................................................................... 48 Appendix .............................................................................................................................. 49 1 Result of linear regression .......................................................................................... 49 2 Distortion of CP and LCG ............................................................................................ 49 CP distortion ..................................................................................................................................................50 2.1 2.2 LCG distortion ................................................................................................................................................50
4 Check of non dimensional parameters ..................................................................... 56 5 The seven ships with their original parameters....................................................... 57 6 Validation of ship motions ......................................................................................... 58
6.1 6.2 7 7.1 7.2 7.3 S.A. van der Stel in Head seas .....................................................................................................................58 S.A. van der Stel in Following waves..........................................................................................................60 Ship geometries ............................................................................................................ 63 S.A. Van der Steel ..........................................................................................................................................63 Series 60 hull model 4210 .............................................................................................................................64 S-175 ................................................................................................................................................................65
NOMENCLATURE
Symbol Quantity Description Linear wave velocity potential. The earth fixed coordinate system. Mean position of ship coordinate system. Ship fixed coordinate system. x coordinate on the ship Surge Sway Heave Roll, port to starboard ( yo to zo ) Pitch, aft to forward ( zo to xo ) Yaw, xo to yo Imaginary unit. Acceleration of gravity. Density of water. Angular wave frequency, in E ( X , Y , Z ) . Wave number. Wave length. Wave celerity. Wave length in diagonal waves. Frequency of encounter, in O( xo , yo , zo ) . Wave amplitude. Heading angle, = 0 following waves = waves from starboard side, = head sea. Ship speed. Ship length in waterline. Ship beam. Draught. Average added resistance in irregular waves. Added resistance in a regular wave. Non dimensional added resistance in a regular wave. Mean response curve. Wave energy spectrum. Added mass per meter. Damping per meter. Sectional damping. Mean depth. Vertical relative velocity, between strip and water. Peak of Raw . 7
2
Wave model
E( X ,Y , Z ) O( xo , yo , zo ) B ( xb , yb , zb ) xb
[ m] [ m] [ m] [ m]
1 2 3 4 5 6
i g
cw
e a
V Lwl b T
[m / s ] [ m] [ m] [ m] [N ]
[N ]
R AW
Raw = R (e )
A g ( b 2 / Lwl )
2 2
R AW
[ ]
[ N / m2 ]
S ( )
a33 b33 b' D Vzb
[m s / rad ] [ Pa s ] [ Pa ] [ Pa ] [ m]
[m / s ] [ ]
Regression analysis
RAW , p
p c p , norm
cnorm RAW ,exp CP
[rad / s ] [rad / s ] [ ]
[ ]
[ ] [ ]
[ m]
ryy
LCG Fn X
[ m] [ ]
Gn yest , Gn ( Gn ) yest , Gn ( Gall ships )
Peak frequency of Raw . Width of Raw peak. Non dimensional p . Non dimensional c . Raw approximated with exponential function. Prismatic coefficient. Radius of pitch gyration. Distance from aft peak to centre of gravity. Froude number Model parameter matrix Coefficient vector in linear regression. Error vector in linear regression. Group n where n = ( Containers , Reefers , RoRos , Tankers and All ships ). Estimation based on linear regression of the data in Gn evaluated for the same group Gn . Estimation based on linear regression of the data in Gn evaluated for all the ships Gall ships . Error vector for linear regression of the data in Gn evaluated for the same group Gn . Standard deviation of G ( Gn ) .
n
G ( Gn )
n
G ( Gn )
n
1 INTRODUCTION
This Master Thesis handles three issues: Added resistance as phenomenon, implementation of three methods and development of a simplified method to calculate added resistance in waves. This thesis thereby contain: a summary of the literature study of this project, a thorough description of three conceptually important methods and a methodology to simplify a method, using regression analysis. 1.1 ADDED RESISTANCE IN WAVES Added resistance in waves is the part of a ships total resistance that is caused by encountering waves. Calculations of added resistance can be used as an addition to the calm water resistance to predict the total resistance of a ship in a seaway. There will always be waves on the sea, so there will always be added resistance. A ship can experience a 15-30% resistance increase in a seaway [12], where the added resistance is the main reason for this increase. Being able to predict added resistance due to waves is therefore a vital part of the prediction of a ships resistance. Prediction of added resistance can for instance be used in the following problems: Weather margin The so called Weather Margin for new ship designs can be decided, where the maximum resistance increase due to weather can be predicted, to decide engine install and so on. Weather Routing Weather Routing is very important due to its economical effect on ship exploitation. It is for instance very important to make good estimations of the time it will take for a ship to travel a route, so the cargo owners know when the ship will arrive in port, minimizing the costs of storage and so on. It is also very important to be able to optimize routes in order to reduce the fuel consumption and emission. A good prediction of Added resistance in waves is important for both these tasks. Performance analysis The previous two problems use the prediction of added resistance to get the total resistance, the reversed problem is however also of interest. Being able to get rid of the influence of the stochastic waves in a seaway, can be used to calculate a ships real calm water resistance. This real calm water resistance can be used as a measurement of the ships performance over time. The ship owners could use this information to determine the value of a ship, and how often it should be docked for antifouling and so on.
1.2 THE NATURE OF ADDED RESISTANCE When a ship is oscillating due to waves, it supplies energy to the surrounding water, energy that will increase the resistance. This energy is primarily transmitted with the waves radiating from the ship (Figure 1). The supplied energy is due to damping of the oscillatory motions. Hydrodynamic damping is dominating for heave- and pitch motions, which are the biggest contributors to added resistance. The viscous damping can therefore be neglected, which means that added resistance can be considered as a non viscous phenomenon [13]. This means that potential theory can be used. The radiation induced resistance is dominating when the ship motions are big. This happens in the region of the resonance frequency of heave and pitch motions (figure 2). The reflection of incident waves is also causing added resistance. The so called diffraction induced resistance is dominating for high wave frequencies (Figure 2), where the ship motions are small.
Energy is also transmitted to the surrounding water by waves generated by the forward speed of the ship. But this is referred to as the calm water resistance, which is not handled in this thesis. The added resistance in a seaway is considered to be independent of the calm water resistance [13].
16
14
12
10
0 0.5
0.6
0.7
0.8
0.9
1.1
1.2
1.3
Figure 2, Radiation induced resistance and diffraction induced resistance, for different wave frequencies.
1.2.1
Usually ship motions and forces are modeled as a so called LTI system (Linear Time Invariant system). This means that a ship is considered as a system which uses a linear sine-wave, representing the water wave, as input signal and delivers a linear sine-wave, representing for instance a motion or a force, as response to this signal. The LTI system is allowed to respond with a phase lag on the input signal and a linear change of the amplitude. These restrictions give a very advantageous property of the LTI system in that the superposition principle can be used. This means that if a signal x ( t ) can be expressed as the sum of sub signals xk ( t ) , the response to this signal y ( t ) can be expressed as the sum of the responses of the sub signals yk ( t ) :
x ( t ) = xk ( t ) y ( t ) = yk ( t )
k k
(1.1)
10
This means that ship motions and forces in irregular waves can be expressed as the sum of the responses in regular waves, which is a very powerful property of a LTI system. In reality ships do not respond linearly to the waves. In order to model the responses as a LTI system, the responses have to be linearized. This linearization gives good accuracy according to [5], since the linear part is dominating the responses. Ship motions are therefore considered to be a first order problem.
1.2.2 Added resistance is a second order problem
The added resistance is the mean force in the heading direction of the ship. Calculating the mean force using a linear force from 1.2.1 will give a zero mean value. This is because the time mean value of an arbitrary sine wave with an arbitrary amplitude A and period time Te is zero:
1 e A cos ( t + ) t = 0 Te 0
A second order sine wave however, will give a non zero time mean value:
2 1 e A2 ( A cos ( t + ) ) t = Te 0 2 T
(1.2)
(1.3)
Therefore the quadratic term in the response has to be included in the problem. The quadratic term is small compared to the linear term but has to be included to obtain a mean value. Strm-Tejsen [13] has shown in experiments that the added resistance in regular waves varies linearly with the wave height squared at a constant wave length, added resistance is therefore considered to be a second order problem. It is unfortunately hard to get good predictions of added resistance, since it is a second order problem. If the motions are predicted with an accuracy of approximately 10-15%, the second order added resistance can not be expected to be of accuracy better than 20-30% [2]. The wave is usually expressed with a velocity potential function. The velocity potential function is derived from boundary conditions that can be linearized. This is referred to as linear wave theory, which will give a linear wave velocity potential. The linear theory is applicable until the wave steepness becomes sufficiently large, that non-linear effects become important. Although added resistance is a second order problem, the linear wave velocity potential is the only one needed. Higher order velocity potentials are not needed, to study the added resistance [5]. 1.3 ADDED RESISTANCE IN IRREGULAR WAVES Added resistance is the time mean value of a second order force. Consider a signal x ( t ) consisting of two signals x1 ( t ) and x2 ( t ) :
x1 ( t ) = A1 cos (1 t + 1 ) x2 ( t ) = A2 cos (2 t + 2 ) x ( t ) = x1 ( t ) + x2 ( t )
The quadratic response to this signal: (1.4)
(1.5)
11
The second order force in an irregular wave can therefore not be expressed with superposition, because of the trigonometric cross terms A1 A2 cos(...) . But added resistance is the time mean value of this second order force, where the trigonometric terms from (1.5) disappears, so that the time mean value of (1.5) can be expressed as:
x (t )
A12 A2 2 = + 2 2
(1.6)
The added resistance in irregular waves can therefore be expressed with superposition of the regular wave responses. Strm-Tejsen [13] has shown this relation in experiments and that the average added resistance RAW in irregular waves with good accuracy can be expressed as:
RAW = 2 R( ) S ( )
0
R ( ) =
R AW ( )
a2
(1.7)
1 S ( ) = a ( ) 2 2
R ( ) is the mean response curve, and S ( ) is the wave energy spectrum. The evaluation of (1.7), made
by Strm-Tejsen, was done by inserting R( ) and S ( ) from regular wave experiments into (1.7), and compare that to the corresponding irregular wave experiment. The usual way to calculate added resistance in irregular waves, is therefore to first calculate the added resistance in regular waves for different wave frequencies and then use (1.7). This is why almost all available methods to calculate added resistance in waves, focus on regular waves. The added resistance for different wave frequencies can be presented in a transfer function like the schematic one in figure 2. It is also important to be aware that the choice of wave energy spectrum S ( ) , will have a big influence on the integrated mean added resistance RAW . The relation between the spectral peaks in the wave energy spectrum S ( ) and the mean response curve R ( ) will have a big impact on the result. So it is reasonable to conclude that to find an accurate wave energy spectrum S ( ) , is as important as to find an accurate prediction of the added resistance in regular waves R( ) . 1.4 NON DIMENSIONAL ADDED RESISTANCE The full scale added resistance R AW in regular waves can be made non dimensional using the following expression:
Raw =
R AW g ( b 2 / Lwl )
(1.8)
This relation has been confirmed by [13] in model tests, using models of the same ship with varying scale. In the rest of this thesis mostly the non dimensional added resistance Raw is considered, unless anything else is stated. 1.5 NON DIMENSIONAL WAVE FREQUENCY The peak of the added resistance transfer function (figure 2) usually occurs at a frequency where the wavelength is about the same size as the ships length. This is due to the big influence of pitch motion, which has its peak here, according to figure 3.
12
Figure 3, wavelengths near the ship length will produce heavy pitching, and added resistance.
This means that the length of the ship will have a big influence on where the peak of the added resistance will be. To capture this relation it is usual to present the transfer functions with a non dimensional frequency, normalized with the ships length in some way. This can be done in a variety of ways, and different authors tend to invent their own way of normalizing the frequency. In the validation part of this thesis, where implementations made in this project are compared to previously published experiments and implementations, the frequency has been made non dimensional in the same way as the compared implementations. The non dimensional frequencies used are:
norm =
(1.9)
norm =
L 2 g
(1.10)
norm = e
L 2 g
(1.11)
1.6 METHODS TO CALCULATE ADDED RESISTANCE IN WAVES Three methods to calculate added resistance in waves have been implemented in this project, Gerritsma and Beukelmans method, Boeses method and Faltinsens asymptotic method. Gerritsma and Beukelmans method is a so called radiated energy method. This problem starts out by trying to describe the energy that the oscillating ship transmits to the surrounding water. It is assumed that to maintain a constant forward ship speed, this energy should be delivered by the ships propulsion plant. Boeses method is a so called pressure integration method, which basically means that the linear pressure in the undisturbed wave is integrated over the ship hull, to obtain a mean force in the heading direction of the ship. It may seem strange that the linear pressure would give a mean force, but it does in this case since the ship hull, where the integration is performed, is moving. Both these methods primarily deal with radiation induced resistance. Faltinsens asymptotic method on the other hand, only deals with diffraction induced resistance, and neglects the ship motions. 1.7 RELATIVE VELOCITY Both Gerritsma and Beukelmans method and Boeses method to calculate added resistance use Relative velocity. The relative velocity is the vertical velocity of the water related to a point on the ship (Figure 4).
13
Relative velocity
Vzb .
1.7.1
Coordinate systems
In the derivation of relative velocity three coordinate systems are used, E ( X , Y , Z ) , O( xo , yo , zo ) and B ( xb , yb , zb ) (Figure 5). The earth fixed coordinate system E ( X , Y , Z ) is fixed in space and has its X-axis aligned with the direction of the wave propagation, it also contains the other two coordinate systems. O( xo , yo , zo ) is positioned at the mean position of the ships centre of gravity. This means that it is traveling with the ship speed along a line with origin in E ( X , Y , Z ) and with an angle to the X-axis. The coordinate system O( xo , yo , zo ) only moves in the X Y plane in E ( X , Y , Z ) with no rotations. B ( xb , yb , zb ) is fixed with the ship. The origin of B ( xb , yb , zb ) situated at CG is expressed by the coordinates 1 (surge), 2 (sway) and 3 (heave) (Figure 6). And if B( xb , yb , zb ) is rotating in O( xo , yo , zo ) around its origin, these rotations are expressed by 4 (roll, yo to zo ), 5 (pitch, zo to xo ) and 6 (yaw, xo to yo ). 14
zb 3 zo xb B(xb,yb,zb) O(xo,yo,zo) 1 xo O( xo , yo , zo ) 5
and
B ( xb , yb , zb ) .
1.7.2
The relative velocity is the velocity of the water particles in the ship fixed coordinate system B( xb , yb , zb ) . The velocities of the water particles in the earth fixed coordinate system E ( X , Y , Z ) can be obtained from the wave velocity potential [5]:
E ( X ,Y , Z , t) =
Particle velocities in E ( X , Y , Z ) :
i g a
e k Z ei( t k X )
(1.12)
uE =
E = a e k Z ei( t k X ) X
(1.13)
wE =
E = i a e k Z ei( t k X ) Z
(1.14)
uE in X -direction, wE in Z -direction. These velocities can be transformed to the ship fixed system B ( xb , yb , zb ) to obtain the relative velocity. First (1.13) and (1.14) are expressed with xb using the
following relations between X in E ( X , Y , Z ) and xb in B( xb , yb , zb ) :
X = (V t + xO ) cos ( ) X = (V t + xb ) cos ( )
Inserted in (1.13) and (1.14):
xO = 1 + xb xb
(1.15)
u E = a e k Z e
i ( t k (V t + xb )cos ( ) )
= a e k Z e
2 i V cos( ) t k xb cos ( ) g
(1.16)
wE = i a e
k Z
2 i V cos( ) t k xb cos( ) g
(1.17)
15
e =
V cos ( ) g
i (e t k xb cos( ) )
(1.18)
u E = a e k Z e
i t k xb cos ( ) ) wE = i a e k Z e ( e
(1.19)
This is still velocities in E ( X , Y , Z ) , but expressed with xb from B( xb , yb , zb ) . The velocities can be transformed to the moving system O( xo , yo , zo ) :
wO = wE
uO = V + u E cos ( )
And transformed to B( xb , yb , zb ) :
(1.20) (1.21)