Theoretical Manual of SEAWAY

(Release 4.19, 12-02-2001)

by: J.M.J. Journée Delft University of Technology Shiphydromechanics Laboratory Mekelweg 2, 2628 CD Delft The Netherlands Report1216a February 2001

2
Abstract This report describes in detail the theoretical backgrounds of the six degrees of freedom ship motions program SEAWAY. SEAWAY is a frequency-domain ship motions computer code, based on the linear strip theory, to calculate the wave-induced loads, motions, added resistance and internal loads for six degrees of freedom of displacement ships, yachts and barges, sailing in regular and irregular waves. When not taking into account interaction e¤ects between the two individual hulls, these calculations can be carried out for twin-hull ships, such as semi-submersibles and catamarans, too. This potential theory program is suitable for deep water as well as for shallow water. Viscous roll damping, bilge keels, anti-roll tanks, free surface e¤ects and linear springs can be added. A dedicated editor takes care for a simple input of data. This report and other information on the strip theory program SEAWAY can be found on the Internet at web site http://dutw189.wbmt.tudelft.nl/~johan, which can also be reached by a link to this site from http://www.shipmotions.nl. Aditional information can be obtained by e-mail (J.M.J.Journee@wbmt.tudelft.nl) from the author. The last revision of this report is dated: 3 July 2001. Remarks and errata are very welcome!

Contents
1 Introduction 2 Strip Theory Method 2.1 De…nitions . . . . . . . . . . . . . . . . . . . . . . . 2.2 Incident Wave Potential . . . . . . . . . . . . . . . 2.2.1 Continuity Condition . . . . . . . . . . . . . 2.2.2 Laplace Equation . . . . . . . . . . . . . . . 2.2.3 Sea Bed Boundary Condition . . . . . . . . 2.2.4 Free Surface Dynamic Boundary Condition . 2.2.5 Free Surface Kinematic Boundary Condition 2.2.6 Dispersion Relationship . . . . . . . . . . . 2.2.7 Relationships in Regular Waves . . . . . . . 2.3 Floating Rigid Body in Waves . . . . . . . . . . . . 2.3.1 Fluid Requirements . . . . . . . . . . . . . . 2.3.2 Forces and Moments . . . . . . . . . . . . . 2.3.3 Hydrodynamic Loads . . . . . . . . . . . . . 2.3.4 Wave and Di¤raction Loads . . . . . . . . . 2.3.5 Hydrostatic Loads . . . . . . . . . . . . . . 2.4 Equations of Motion . . . . . . . . . . . . . . . . . 2.5 Strip Theory Approaches . . . . . . . . . . . . . . . 2.5.1 Zero Forward Speed . . . . . . . . . . . . . 2.5.2 Forward Ship Speed . . . . . . . . . . . . . 2.5.3 End-Terms . . . . . . . . . . . . . . . . . . . 2.6 Hydrodynamic Coe¢cients . . . . . . . . . . . . . . 3 Conformal Mapping 3.1 Lewis Conformal Mapping . . . . . . 3.1.1 Boundaries of Lewis Forms . . 3.1.2 Acceptable Lewis Forms . . . 3.2 Extended Lewis Conformal Mapping 3.3 Close-Fit Conformal Mapping . . . . 3.4 Comparisons . . . . . . . . . . . . . . 4 2-D Potential Coe¢cients 4.1 Theory of Tasai . . . . . 4.1.1 Heave Motions . 4.1.2 Sway Motions . . 4.1.3 Roll Motions . . . . . . . . . . . . . . . . . . . . . . . . . . 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 5 6 9 9 10 11 11 13 14 14 14 15 17 18 22 24 24 27 28 29 30 32 35 36 38 39 40 41 44 47 48 48 57 68

4 4.1.4 Low and High Frequencies . 4.2 Theory of Keil . . . . . . . . . . . . 4.2.1 Notation . . . . . . . . . . . 4.2.2 Basic Assumptions . . . . . 4.2.3 Vertical Motions . . . . . . 4.2.4 Horizontal Motions . . . . . 4.2.5 Appendices . . . . . . . . . 4.3 Theory of Frank . . . . . . . . . . . 4.3.1 Notation . . . . . . . . . . . 4.3.2 Formulation of the problem 4.3.3 Solution of the Problem . . 4.3.4 Low and High Frequencies . 4.3.5 Irregular Frequencies . . . . 4.3.6 Appendices . . . . . . . . . 4.4 Surge Coe¢cients . . . . . . . . . . 4.5 Comparative Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

CONTENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 81 81 82 84 109 119 126 127 128 130 133 135 137 144 146 147 147 147 148 148 150 152 152 153 163 163 165 168 170 172 175 179 181 181 183 185 189 191 191 193 193 194 195 201

5 Viscous Damping 5.1 Surge Damping . . . . . . . . . . . . . . 5.1.1 Total Surge Damping . . . . . . . 5.1.2 Viscous Surge Damping . . . . . 5.2 Roll Damping . . . . . . . . . . . . . . . 5.2.1 Experimental Determination . . . 5.2.2 Empirical Formula for Barges . . 5.2.3 Empirical Method of Miller . . . 5.2.4 Semi-Empirical Method of Ikeda . 6 Hydromechanical Loads 6.1 Hydromechanical Forces for Surge . . 6.2 Hydromechanical Forces for Sway . . 6.3 Hydromechanical Forces for Heave . . 6.4 Hydromechanical Moments for Roll . 6.5 Hydromechanical Moments for Pitch 6.6 Hydromechanical Moments for Yaw . . . . . . . . . . . . .

7 Exciting Wave Loads 7.1 Classical Approach . . . . . . . . . . . . 7.1.1 Exciting Wave Forces for Surge . 7.1.2 Exciting Wave Forces for Sway . 7.1.3 Exciting Wave Forces for Heave . 7.1.4 Exciting Wave Moments for Roll 7.1.5 Exciting Wave Moments for Pitch 7.1.6 Exciting Wave Moments for Yaw 7.2 Equivalent Motions of Water Particles . 7.2.1 Hydromechanical Loads . . . . . 7.2.2 Energy Considerations . . . . . . 7.2.3 Wave Loads . . . . . . . . . . . . 7.3 Numerical Comparison . . . . . . . . . .

CONTENTS 8 Transfer Functions of Motions 8.1 Centre of Gravity Motions . . . . . . . . . . . . . . . . 8.2 Absolute Displacements . . . . . . . . . . . . . . . . . 8.3 Absolute Velocities . . . . . . . . . . . . . . . . . . . . 8.4 Absolute Accelerations . . . . . . . . . . . . . . . . . . 8.4.1 Accelerations in the Earth-Bound Axes System 8.4.2 Accelerations in the Ship-Bound Axes System . 8.5 Vertical Relative Displacements . . . . . . . . . . . . . 8.6 Vertical Relative Velocities . . . . . . . . . . . . . . . . 9 Anti-Rolling Devices 9.1 Bilge Keels . . . . . . . . . . . . . 9.2 Passive Free-Surface Tanks . . . . . 9.2.1 Theoretical Approach . . . . 9.2.2 Experimental Approach . . 9.2.3 E¤ect of Free-Surface Tanks 9.3 Active Fin Stabilisers . . . . . . . . 9.4 Active Rudder Stabilizers . . . . .

5 203 203 204 205 207 207 208 208 209 211 211 212 212 215 218 218 221 223 224 224 225 227 228 229 231 233 238 239 241 244 247 248 248 248 249 249 253 257 258 258 259

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

10 External Linear Springs 10.1 External Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2 Additional Coe¢cients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3 Linearized Mooring Coe¢cients . . . . . . . . . . . . . . . . . . . . . . . . 11 Added Resistances due to Waves 11.1 Radiated Energy Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Integrated Pressure Method . . . . . . . . . . . . . . . . . . . . . . . . . . 11.3 Comparison of Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Bending and Torsional Moments 12.1 Still Water Loads . . . . . . . . 12.2 Lateral Dynamic Loads . . . . . 12.3 Vertical Dynamic Loads . . . . 12.4 Torsional Dynamic Loads . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

13 Statistics in Irregular Waves 13.1 Normalized Wave Energy Spectra . . . . 13.1.1 Neumann Wave Spectrum . . . . 13.1.2 Bretschneider Wave Spectrum . . 13.1.3 Mean JONSWAP Wave Spectrum 13.1.4 De…nition of Parameters . . . . . 13.2 Response Spectra and Statistics . . . . . 13.3 Shipping Green Water . . . . . . . . . . 13.4 Bow Slamming . . . . . . . . . . . . . . 13.4.1 Criterium of Ochi . . . . . . . . . 13.4.2 Criterium of Conolly . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . 15. . . . 15 Numerical Recipes 15. . . . . . . . . . . . . . . . . 15. . . . . . . . . . . . . . . . . 15. . . . . . . . . . . . . . . . . 15. . . .1 First Degree Polynomials . . . . . . . . . 15. . . .2. . 15. . 14. . .4 Curve Lengths . . . .5. . .6 Bending and Torsional Moments .2 Second Degree Polynomials 15. . . . . . . . . . . . . . . . . . . .2 Second Degree Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15. . . . . . . .3. 14. . . . . . . . . . . . . . . . . . . . . . . . . . . . 14. . . . . . . . . . . . . . . . . . . . . . . .2 Second Degree Curves . . . . . .4. . .1. 14.14 Twin-Hull Ships 14. . . . . . . . . . . . . . . . . . . .1 First Degree Derivatives . . . . .2 Integrated Pressure Method .2. . . . . . . 15. . . . . . . . . . . . . . . 14. . . . .4 Exciting Wave Forces and Moments .1 Hydromechanical Coe¢cients . . . . . . . . . . . . . . . . .2 Equations of Motion . . . . . . . . . . . .2. . . . . . . . . . . . . . . .1 Radiated Energy Method . . . . . . .1 First Degree Curves . . . . . . . . . . . . . . . . . . . . . 15. . .2 Second Degree Integrations 15. . .5 Added Resistance due to Waves . . . . . . . .4. . . . . . . . . . . . . . .3 Derivatives . . . .3 Hydromechanical Forces and Moments 14. . . . .1 Polynomials . . . . . . . . . . . . . . . . . . . . . . .1. . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 263 264 264 265 267 268 268 268 271 271 271 272 272 273 273 274 275 276 276 278 278 278 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14. . . . . . . . . . .5. . . . . .3 Integration of Wave Loads . . . . . . . . . . . . 15. . . . . . .1 First Degree Integrations . . . . . . . . . . . . . . . .2 Integrations . . . . . . .

Chapter 1. Chapter 2 gives a general description of the various strip theory approaches. calculates the waveinduced loads and motions with six degrees of freedom of mono-hull ships and barges. Delft University of Technology. this introduction. named SEAWAY. Release 4. The derivations of the hydromechanical forces and moments. based on the ordinary and the modi…ed strip theory. Also an inclusion of the J.shipmotions. ”Theoretical Manual of SEAWAY. this program has been extended and adapted considerably. Since then. as it has been de…ned by [Korvin-Kroukovsky and Jacobs. so a revised report is presented here. This includes for all motions a forward speed e¤ect caused by the potential mass. This approach is called the ”Ordinary Strip Theory Method”.19”. In the past a preliminary description of all algorithms. The present report describes in detail the theoretical backgrounds and the algorithms of a six degrees of freedom ship motions personal computer program. gives a short survey of the contents of all chapters in this report. 1957] for the heave and pitch motions.nl. sailing in a seaway. The principal assumptions are a linear relation between forces and motions and the validity of obtaining the total forces by a simple integration over the ship length of the two-dimensional cross sectional forces. Report 1216a. A general description of the potential ‡ow theory is given.Chapter 1 Introduction This report aims being a guide and an aid for those who want to study the theoretical backgrounds and the algorithms of a ship motions computer program based on the strip theory. Extensive veri…cations and validations of this program have been presented by [Journée. Journée. used in strip theory based ship motions calculations. the wave potential and the wave and di¤raction forces and moments have been described. Ship Hydromechanics Laboratory. This program. A User Manual is given by [Journée.M. The equations of motion are given with solid mass and inertia terms and hydromechanical forces and moments in the left hand side and the wave exciting forces and moments in the right hand side. has been given by the author. 2628 CD Delft. For updates see web site: http://dutw189. The Netherlands. February 2001. 1992]. such as semi-submersibles and catamarans. see [Journée. When not taking into account interaction e¤ects between the two individual hulls.J.nl/~johan or http://www. Mekelweg 2. 2001b]. 2001a].wbmt. too. 0 1 . these calculations can be carried out for twin-hull ships.tudelft.

Chapter 7 describes the wave exciting forces and moments in the right hand side of the six equations of motion of a sailing ship in water with an arbitrarily depth. Chapter 4 describes the determination of the two-dimensional potential mass and damping coe¢cients for the six modes of motions at in…nite and …nite water depths. the principle of the calculation of these potential coe¢cients is based on work of [Ursell. 1949] for circular cylinders and [Frank. the pitch and yaw coe¢cients follow from the moment about the ship’s centre of gravity of the heave and sway coe¢cients. respectively. For shallow water. 1967] for deep water has been described too. 1974] and [Ikeda et al. based on a simple still water resistance curve as published by [Troost. the calculation routines of the two-dimensional hydrodynamic potential coe¢cients of ship-like cross sections are given for the sway. 1959]. In this manner hydrodynamic problems can be solved directly with the coe¢cients of the mapping function. used to determine the conformal mapping coe¢cients. the method of [Keil. 1985]. Two special cases of multi-parameter conformal mapping have been described too: the well known classic Lewis transformation ([Lewis. these cross sections are conformal mapped to the unit circle. 1967] for any arbitrary symmetric cross section. 1978]. has been described. Chapter 6 describes the determination of the hydromechanical forces and moments in the left hand side of the six equations of motion of a sailing ship in deep water for both the ordinary and the modi…ed strip theory method. [Tasai.2 CHAPTER 1.. The surge damping coe¢cient is corrected for viscous e¤ects by an empirical method. 1929]) with two parameters and an Extended-Lewis transformation with three parameters. The pulsating sources method of [Frank. 1973] and using the multi-parameter conformal mapping technique. 1961] and [Jong. Because of using the strip theory approach here. . A very simple and straight on iterative least squares method. The close-…t multi-parameter conformal mapping method is given.based on a variation of the theory of [Ursell. 1955]. Starting from the velocity potentials and the conjugate stream functions of the ‡uid with an in…nite depth as have been given by [Tasai. as given by [Athanassoulis and Loukakis. heave and roll motions of ship-like cross sections. is given. as for instance given by [Tasai. to determine a viscous correction of the roll damping coe¢cients are described in detail. Chapter 5 gives some corrections on the hydrodynamic damping due to viscous e¤ects. For the determination of the two-dimensional hydrodynamic potential coe¢cients for sway. The inclusion of so-called ”End-Terms” has been described. 1974] . [Tasai. The analysis of free-rolling model experiments and two (semi-)empirical methods published by [Miller. Approximations are given for the surge coe¢cients. The advantage of conformal mapping is that the velocity potential of the ‡uid around an arbitrary shape of a cross section in a complex plane can be derived from the more convenient circular section in another complex plane. using the relative motion concept for both the ordinary and the modi…ed strip theory method. At in…nite water depths. INTRODUCTION forward speed e¤ect caused by the potential damping. 1969]. Chapter 3 describes several conformal mapping methods. 1949] . heave and roll motions. 1960]. This approach is called the ”Modi…ed Strip Theory Method”.has been given.

Chapter 11 describes two methods to determine the transfer functions of the added resistances due to waves. Chapter 12 describes the determination of the frequency characteristics of the lateral and vertical shear forces and bending moments and the torsional moments in a way as presented by [Fukuda. Chapter 14 describes the additions to the algorithms in case of twin. The determination of the frequency characteristics of the absolute displacements. A description is given of the calculation procedure of the energy spectra and the statistics of the ship motions for six degrees of freedom. as used in program SEAWAY. A description is given of an inclusion of passive free-surface tanks as de…ned by the experiments of [Bosch and Vugts.3 First. the classical approach has been described. of importance when calculating the probability of shipping green water. rotations. The static and dynamic swell up of the waves. Chapter 8 describes the solution of the equations of motion. For interaction e¤ects between the two individual hulls will not be accounted here. the Rayleigh probability density function has been used. Bow slamming phenomena are de…ned by both the relative bow velocity criterium of [Ochi. 1963]. For the calculation of the probability of exceeding a threshold value by the motions. velocities and accelerations and the vertical relative displacements. The use of a wave potential valid for any arbitrary water depth makes the calculation method. such as semi-submersibles and catamarans. Active …n stabilizers and active rudder stabilizers have been described too. the vertical relative motions and the mechanic loads on the ship in waves coming from any direction. The …rst method is a radiated wave energy method. 1965].hull ships. 1970]. The second method is an integrated pressure method. 1962] for the vertical mode. as published by [Boese. the added resistances. an alternative approach . Chapter 10 describes the inclusion of linear spring terms to simulate the behavior of anchored or moored ships. Chapter 9 describes some anti-rolling devices. 1974]. 1972]. with deep water coe¢cients. as published by [Gerritsma and Beukelman. are de…ned according to [Tasaki.based on di¤raction of waves . an average wave spectrum of Bretschneider and the more narrow Mean JONSWAP wave spectrum. . Chapter 13 describes the statistics in irregular waves. Still water phenomena are described too. suitable for ships sailing with keel clearances down to about 50 percent of the ship’s draft. Chapter 15 shows some typical numerical recipes.with equivalent accelerations an velocities of the water particles has been described. 1966] and by the theory of [Verhagen and van Wijngaarden. Then. Three examples of normalized wave spectra are given: the somewhat wide wave spectrum of Neumann. 1964] and the peak bottom impact pressure criterium of [Conolly. by using the superposition principle.

INTRODUCTION . CHAPTER 1.4 .

irrotational and without viscosity.M. Report 1216a. However. ”Theoretical Manual of SEAWAY. As a result of this. This holds that viscous e¤ects are neglected. Mekelweg 2. the strip theory can deliver less accurate results. Because of the way that the forced motion problems are solved generally in the strip theory. J.0. which can deliver serious problems when predicting roll motions at resonance frequencies. these ”near zero frequency of encounter problems” can be solved by forcing the wave loads to go to zero arti…cially. Then the so-called ”end-terms” can be important too. February 2001. The incorporation of seakeeping theories in ship design has been discussed clearly by [Faltinsen and Svensen.shipmotions. Delft University of Technology. Release 4. The strip theory solves the three-dimensional problem of the hydromechanical and exciting wave forces and moments on the ship by integrating the two-dimensional potential solutions over the ship’s length. as appear in extreme sea states.strip theories are the most successful and practical tools for the calculation of the wave induced motions of the ship.nl/~johan or http://www. or even sometimes lower. In practice. In practice. Ship Hydromechanics Laboratory.19”. experiments showed that the strip theory appears to be remarkably e¤ective for predicting the motions of ships with length to breadth ratios down to about 3. For updates see web site: http://dutw189.tudelft. at least in an early design stage of a ship. 0 5 . 1990].Chapter 2 Strip Theory Method The ship is considered to be a rigid body ‡oating in an ideal ‡uid: homogeneous. So each cross section of the ship is considered to be part of an in…nitely long cylinder. For high-speed vessels and for large ship motions. Interactions between the cross sections are ignored for the zero-speed case.wbmt. An overview of seakeeping theories for ships were presented and it was concluded that .J.nl. free of surface tension. only the external loads on the underwater part of the ship are considered and the e¤ect of the above water part is fully neglected. The strip theory is based on the potential ‡ow theory. The strip theory is a slender body theory. The Netherlands. It is assumed that the problem of the motions of this ‡oating body in waves is linear or can be linearized. 2628 CD Delft. for viscous roll damping e¤ects can be accounted fairly by empirical formulas.nevertheless some limitations . incompressible. so one should expect less accurate predictions for ships with low length to breadth ratios. substantial disagreements can be found between the calculated results and the experimental data of the wave loads at low frequencies of encounter in following waves. Journée.

but time t has replaced x on the horizontal axis. The strip theory is based on linearity. The still water level is the average water level or the level of the water if no waves were present.1 De…nitions Figure 2.6 CHAPTER 2. 2. it can be considered to carry out additional model experiments to investigate for instance added resistance or extreme event phenomena. unsteady divergent wave systems become important. h. The water depth. Figure 2. So when parts of the ship go out of or into the water or when green water is shipped. so that the unsteady waves generated by the ship are propagating in directions perpendicular to the centre plane of the ship. the strip theory does not distinguish between alternative above water hull forms. most relevant values of z will be negative. In reality the wave systems around the ship are far more complex. it looks similar in many ways to the other …gure.1: Harmonic Wave De…nitions Notice that the origin of the coordinate system is at the still water level with the positive z-axis directed upward. At a more detailed design stage.1-a shows what one would observe in a snapshot photo made looking at the side of a (transparent) wave ‡ume. (a positive value) is measured between the sea bed (z = ¡h) and the still water level (z = 0). STRIP THEORY METHOD The strip theory accounts for the interaction with the forward speed in a very simple way. seakeeping prediction methods based upon the strip theory provide a su¢ciently good basis for optimization studies at an early design stage of the ship. Also.1-b is a time record of the water level observed at one location along the ‡ume. Only hydrodynamic e¤ects of the hull below the still water level are accounted for.1 shows a harmonic wave as seen from two di¤erent perspectives. This e¤ect is neglected in the strip theory. its inaccuracy will be gained strongly by inaccuracies in the predicted motions. Because of the added resistance of a ship due to the waves is proportional to the relative motions squared. such as shipping green water and slamming. relative to the cross sectional dimensions of the ship. inaccuracies can be expected. Figure 2. Nevertheless these limitations. . The x-axis is positive in the direction of wave propagation. the wave pro…le is shown as a function of distance x along the ‡ume at a …xed instant in time. For high-speed vessels. This means that the ship motions are supposed to be small. The e¤ect of the steady wave system around the ship is neglected and the free surface conditions are simpli…ed. Figure 2.

or to the trough for that matter. z) coordinate system. the (xb .1. here. In still water.2.the form of the water surface . the wave form moves one wave length during one period so that its speed or phase velocity. The directions of the axes are: x in the direction of the forward speed V . ¸. y0 . the wave length and period are converted to angles using: 2¼ ¸ 2¼ != T k= k¸ = 2¼ !T = 2¼ or: or: (2. c. If the wave is described by a harmonic wave. The subscript ”a” denotes amplitude. the wave pro…le . y0 )-plane lies in the still water surface. T: The ratio of wave height to wave length is often referred to as the dimensionless wave steepness. If the wave moves in the positive x0 -direction (de…ned in a direction with an angle ¹ relative to the ship’s speed vector.3) . with coordinate systems as given in …gure 2. wave lengths and periods are usually actually measured between two consecutive upward (or downward) crossings of the still water level. y. The distance along the time axis is the wave period. A right-handed coordinate system S(x0 . Another right-handed coordinate system O(x. y.2. The origin O lies above or under the time-averaged position of the centre of gravity G.1) in which k is the wave number (rad/m) and ! is the circular wave frequency (rad/s). yb )plane is parallel to the still water surface. V ). and are easier to detect in a wave record. zb ) is connected to the ship with G at the ship’s centre of gravity. Such points are also called zerocrossings. 2³ a =¸: Since the distance between any two corresponding points on successive sine waves is the same. A third right-handed coordinate system G(xb . is given by: c= ¸ ! = T k (2. The directions of the axes are: xb in the longitudinal forward direction. The ship is supposed to carry out oscillations around this moving O(x. The horizontal distance (measured in the direction of wave propagation) between any two successive wave crests is the wave length. The (x. x0 is directed as the wave propagation and z0 is directed upwards. z) is moving forward with a constant ship speed V . then its amplitude ³ a is the distance from the still water level to the crest. z0 ) is …xed in space. y in the lateral port side direction and z upwards.can now be expressed as a function of both x0 and t as follows: ³ = ³ a cos(kx0 ¡ !t) or: ³ = ³ a cos(!t ¡ kx0 ) (2. Obviously. The (x0 . yb . Since sine or cosine waves are expressed in terms of angular arguments. y)-plane lies in the still water surface. yb in the lateral port side direction and zb upwards. DEFINITIONS 7 The highest point of the wave is called its crest and the lowest point on its surface is the trough.2) Suppose now a sailing ship in waves.

y.5) (2.8 CHAPTER 2.4) x0 = V t cos ¹ + x cos ¹ + y sin ¹ From the relation between the frequency of encounter ! e and the wave frequency !: ! e = ! ¡ kV cos ¹ follows: The resulting six ship motions in the O(x. STRIP THEORY METHOD Figure 2.6) (2.and z-axes and three rotations about them: surge sway heave roll pitch yaw : : : : : : x = xa cos(! et + "x³ ) y = ya cos(! e t + "y³ ) z = za cos(! e t + "z³ ) Á = Áa cos(! e t + "Á³ ) µ = µa cos(! e t + "µ³ ) à = à a cos(! e t + "ó ) ³ = ³ a cos(! e t ¡ kx cos ¹ ¡ ky sin ¹) (2. y.7) The phase lags of these motions are related to the harmonic wave elevation at the origin of the O(x. for instance for surge: surge displacement : x = xa cos(! e t + "x³ ) . z) system are de…ned by three translations of the ship’s centre of gravity in the direction of the x-. z) system are found by taking the derivatives of the displacements. y. the average position of the ship’s centre of gravity: wave: ³ = ³ a cos(! e t) (2.8) The harmonic velocities and accelerations in the O(x. which yields: (2.2: Coordinate System The right-handed coordinate system O(x. y. y. z) is moving with the ship’s speed V . x) system.

The relationships presented in these sections are valid for all water depths. it will be necessary to assume that the water surface slope is very small.11) for small arguments. The pro…le of a simple wave with a small steepness looks like a sine or a cosine and the motion of a water particle in a wave depends on the distance below the still water level. The simpli…cations stem from the following approximations for large and very small arguments. t) = P (z) ¢ sin (kx ¡ !t) (2. Engineers .often look for ways to simplify the theory. z. t) of the harmonic waves has to ful…ll four requirements: 1. Suppose a wave moving in the x-z plane. follow from the de…nition of the velocity potential. INCIDENT WAVE POTENTIAL surge velocity : surge acceleration : x = ¡! e xa sin(! e t + "x³ ) _ x = ¡! 2 xa cos(! e t + "x³ ) Ä e 9 (2. v.14) . z. but the fact that they contain so many hyperbolic functions makes them cumbersome to use. The velocity potential ©w (x. Sea bed boundary condition 3.2.2 Incident Wave Potential In order to use the linear potential theory with waves.1 Continuity Condition The velocity of the water particles (u. s (2. These requirements lead to a more complete expression for the velocity potential as will be explained in the following sections. vz ). ©w : @©w @©w @©w (2. This is reason why the wave potential is written as: ©w (x.2.as opposed to (some) scientists .3: for large arguments.12) 2. s. Free surface kinematic boundary condition.13) v = vy = w = vz = @x @y @z Since the ‡uid is homogeneous and incompressible. or alternatively (vx . w) in the three translational directions.2. the continuity condition becomes: u = vx = @u @v @w + + =0 @x @y @z (2.10) in which P (z) is an (as yet) unknown function of z. as shown in …gure 2. Free surface dynamic boundary condition 4. This means that the wave steepness is so small that terms in the equations of motion of the waves with a magnitude in the order of the steepness-squared can be ignored. vy .9) 2. Continuity condition or Laplace equation 2. s sinh(s) t cosh(s) À s tanh(s) t 1 sinh(s) t tanh(s) t s cosh(s) t 1 (2.

z. C2 k t x z ! = = = = = = = = wave potential (m2 /s) base of natural logarithms (-) as yet undetermined constants (m2 /s) wave number (1/m) time (s) horizontal distance (m) vertical distance. t) e C1 .17) Using this result from the …rst boundary condition. so in the equations above: v= @©w = 0 and @y @v @ 2 ©w = =0 @y @y 2 (2. STRIP THEORY METHOD Figure 2.2 Laplace Equation This continuity condition results in the Laplace equation for potential ‡ows: r2 ©w = @ 2 ©w @ 2 ©w @ 2 ©w + + =0 @x2 @y 2 @z 2 (2.15 yields a homogeneous solution of this equation: d2 P (z) ¡ k 2 P (z) = 0 dz 2 with as solution for P (z): P (z) = C1 e+kz + C2 e¡kz (2.16) Taking this into account. the wave potential can be written now with two unknown coe¢cients as: ¢ ¡ ©w (x. a substitution of equation 2. positive upwards (m) wave frequency (1/s) (2.18) (2.10 in equation 2. z. t) = C1 e+kz + C2 e¡kz ¢ sin (kx ¡ !t) ©w (x.10 CHAPTER 2.2.15) Water particles move here in the x-z plane only.19) in which: .3: Hyperbolic Functions Limits 2.

INCIDENT WAVE POTENTIAL 11 2. v = 0 and since the waves have a small steepness (u and w are small).24) and the wave potential with only one unknown becomes: ©w (x.19 provides: kC1 e¡kh ¡ kC2 e+kh = 0 By de…ning: C = 2C1 e¡kh = 2C2 e+kh or: C +kh C and C2 = e¡kh e 2 2 it follows that P (z) in equation 2. at the free surface of the ‡uid.22) (2. This requirement for the pressure is called the dynamic boundary condition at the free surface.2.4 Free Surface Dynamic Boundary Condition The pressure.2.21) P (z) = (2. z. t) = C ¢ cosh k (h + z) ¢ sin (kx ¡ !t) in which C is an (as yet) unknown constant.2.2. this equation becomes: @©w p + + gz = 0 @t ½ At the free surface this condition becomes: @©w p0 + + g³ = 0 @t ½ for: z = ³ (2.3 Sea Bed Boundary Condition The vertical velocity of water particles at the sea bed is zero (no-leak condition): @©w for: z = ¡h =0 @z Substituting this boundary condition in equation 2. p0 . The Bernoulli equation for an instationary irrotational ‡ow (with the velocity given in terms of its three components) is in its general form: ¢ p @©w 1 ¡ 2 + u + v 2 + w2 + + gz = 0 @t 2 ½ (2.18 can be worked out to: C1 = ¢ C ¡ +k(h+z) e + e¡k(h+z) 2 = C cosh k (h + z) (2. z = ³.20) (2. (2.26) In two dimensions. this will not in‡uence the velocities being obtained from the potential ©w .28) (2. With this the equation becomes: .25) 2.23) (2. is equal to the atmospheric pressure.27) The constant value p0 =½ can be included in @©w =@t. p.

the wave pro…le becomes: 1 @©w ³ =¡ ¢ g @t for: z = 0 (2. z. keeping in mind that the vertical displacement ³ is relatively small: ½ ¾ f©w (x.10 in equation 2. as generally will be used in ship motion equations: ©w = ¡³ a g cosh k(h + z) ¢ ¢ sin(!t ¡ kx) ! cosh kh (2. t) @t ¾ z=0 + O("2 ) z=0 which yields for the linearized form of the free surface dynamic boundary condition: @©w + g³ = 0 @t With this. STRIP THEORY METHOD @©w + g³ = 0 @t for: z = ³ (2. z. t)gz=0 + ³ ¢ ½ @©w (x. t) @z + :::::: (2. z.32 yields the wave pro…le: ³= or: ³ = ³ a ¢ cos (kx ¡ !t) with: ³ a = !C ¢ cosh kh g (2.35) or when !t is the …rst of the sine function arguments.12 CHAPTER 2.32) for: z = 0 (2. t) @t ¾ = z=³ @©w (x. is given by the relation: ©w = ³ a g cosh k(h + z) ¢ ¢ sin(kx ¡ !t) ! cosh kh (2.33) With this the corresponding wave potential.37) .36) In deep water. z. the expression for the wave potential reduces to: ©w = ¡³ a g kz ¢ e ¢ sin(!t ¡ kx) ! (deep water) (2.34) !C ¢ cosh kh ¢ cos (kx ¡ !t) g (2.29) The potential at the free surface can be expanded in a Taylor series. depending on the water depth h.30) ½ @©w (x.31) A substitution of equation 2. t)gz=³ = f©w (x. z.

31 this condition is valid for z = 0 too.41) .5 Free Surface Kinematic Boundary Condition So far the relation between the wave period T and the wave length.39 this delivers the free surface kinematic boundary condition or the Cauchy-Poisson condition: for z = 0 (2.40) = @z @t Analogous to equation 2.2. ¸.31) with respect to t provides: @ 2 ©w @³ +g =0 2 @t @t or after re-arranging terms: @³ 1 @ 2 ©w for z = 0 (2. Using the equation of the free surface 2. INCIDENT WAVE POTENTIAL 13 2. A di¤erentiation of the free surface dynamic boundary condition (equation 2.4: Kinematic Boundary Condition This linearization provides the vertical velocity of the wave surface: @³ dz for the wave surface: z = ³ = dt @t The vertical velocity of a water particle in the free surface is then: (2.2. The relation between T and ¸ (or equivalently ! and k) follows from the boundary condition that the vertical velocity of a water particle in the free surface of the ‡uid is identical to the vertical velocity of that free surface itself (no-leak condition). instead of for z = ³ only.34 yields: dz @³ @³ dx = + ¢ dt @t @x dt d³ @³ = +u¢ @t dx for the wave surface: z = ³ (2. this is a kinematic boundary condition.4. see …gure 2. is still unknown.42) + ¢ =0 @t g @t2 Together with equation 2. Figure 2.2.38) The second term in this expression is a product of two values. which are both small because of the assumed small wave steepness.39) @³ @©w for z = 0 (2. This product becomes even smaller (second order) and can be ignored.

Notice the boundaries ¸=h ¼ 2 and ¸=h ¼ 20 in this …gure between short (deep water) and long (shallow water) waves. STRIP THEORY METHOD @z 1 @ 2 ©w =0 + ¢ @t g @t2 for: z = 0 (2. equation 2.2.43 gives the dispersion relation for any arbitrary water depth h: ! 2 = k g ¢ tanh kh (2. The x-axis is coincident with the undisturbed still water free surface and the z-axis and z0 -axis are positive upwards.47) 2. it is assumed that x ¼ xb . The linear ‡uid velocity potential can be split into three parts: ©(x.14 CHAPTER 2. This leads to the following expressions for the wave surface and the …rst order wave potential in the G(xb . z) system is identical to the S(x0 . zb ) system: (2. it is assumed here that the O(x. that equation will generally have to be solved iteratively.44 degenerates to a quite simple form which can be used without di¢culty: !2 = k g (deep water) (2. z. The water depth is assumed to be …nite. y.7 Relationships in Regular Waves Figure 2. the wave frequency ! has to be replaced by the frequency of encounter of the waves ! e .44) In many situations.35) in equation 2.6 Dispersion Relationship The information is now available to establish the relationship between ! and k (or equivalently T and ¸) referred to above. z0 ) system. In case of forward ship speed. T .3 Floating Rigid Body in Waves Consider a rigid body. c and h for a wide variety of conditions. y.43) 2. y ¼ yb and z ¼ zb . yb . A substitution of the expression for the wave potential (equation 2. t) = ©r + ©w + ©d in which: (2. ‡oating in an ideal ‡uid with harmonic waves. In deep water (tanh kh = 1).44. one must determine k or ¸: Since k appears in a nonlinear way in 2.5 shows the relation between ¸. ©w . ! or T will be know. The time-averaged speed of the body is zero in all directions.45) When calculating the hydromechanical forces and the wave exciting forces on a ship.46) ³ = ³ a cos(! e t ¡ kxb cos ¹ ¡ kyb sin ¹) and the expression for the velocity potential of the regular waves.48) . becomes: ©w = ¡³ a g cosh k(h + zb ) ¢ sin(! e t ¡ kxb cos ¹ ¡ kyb sin ¹) ! cosh(kh) (2.2. For the sake of simple notation. y0 . 2.

Of these. have to ful…ll a number of requirements and boundary conditions in the ‡uid. 1.3. Additional boundary conditions are associated with the oscillating ‡oating body. FLOATING RIGID BODY IN WAVES 15 Figure 2. T .49) The velocity potentials. Continuity Condition or Laplace Equation .1 Fluid Requirements From the de…nition of a velocity potential © follows the velocity of the water particles in the three translational directions: vx = @© @x vy = @© @y vz = @© @z (2.3. c and h ©r ©w ©d = the radiation potential for the oscillatory motion of the body in still water = the incident undisturbed wave potential = the di¤raction potential of the waves about the restrained body 2. © = ©r + ©w + ©d .5: Relationships between ¸. the …rst three are identical to those in the incident undisturbed waves.2.

y.56) = vn (x. this can be written as: X @© = vn (x. the continuity condition: @vx @vy @vz + + =0 @x @y @z results into the equation of Laplace: @ 2© @ 2© @ 2© r ©= + 2 + 2 =0 @x2 @y @z 2 (2.50) (2. y. Sea Bed Boundary Condition The boundary condition on the sea bed. y. z) = z ¢ f1 ¡ x ¢ f3 x cos(n. z. x) cos(n. following from the de…nition of the velocity potential. on the surface of the body. y) cos(n. and generalized direction-cosines.58) . Because of the linearization.54) = ¼ dt @z @t Combining these two conditions provides the boundary condition at the free surface: @© @2© +g =0 @t2 @z for: z = 0 (2. z. vn . given by: f1 f2 f3 f4 f5 f6 = = = = = = cos(n. Kinematic Boundary Condition on the Oscillating Body Surface The boundary condition at the surface of the rigid body plays a very important role.52) =0 @z 3. t) @n Because the solution is linearized. z) y cos(n. t).53) At the free surface of the ‡uid. vj . The velocity of a water particle at a point at the surface of the body is equal to the velocity of this (watertight) body point itself. x) = x ¢ f2 ¡ y ¢ f1 (2.16 CHAPTER 2. the pressure p is constant. t) = vj ¢ fj @n j=1 6 (2.55) 4. z) is given by the linearized Bernoulli equation: p = ¡½ @© ¡ ½gz @t or: @© ¡p + g³ = @t ½ (2.57) in terms of oscillatory velocities. y) = y ¢ f3 ¡ z ¢ f2 z cos(n.51) 2. at a point P (x. z) ¡ z cos(n. y) ¡ y cos(n. S. y. y. z. Boundary Condition at the Free Surface The pressure in a point P (x. the vertical velocity of a water particle in the free surface becomes: dz @© @³ (2. so for z = ³(x. The outward normal velocity. STRIP THEORY METHOD As the ‡uid is homogeneous and incompressible. z) at the surface of the body (positive in the direction of the ‡uid) is given by: @© (2. x) ¡ x cos(n. is given by: @© for: z = ¡h (2. fj .

Note: The subscripts 1. p. 2.2. on both sides must have the same directions at any time. FLOATING RIGID BODY IN WAVES 17 The direction cosines are called generalized. The pressure p . these motions are anti-symmetric.2 Forces and Moments ~ ~ The forces F and moments M follow from an integration of the pressure. f2 and f3 have been normalized (the sum of their squares is equal to 1) and used to obtain f4 . this is a symmetric motion. S. the potential value tends to zero: R!1 lim © = 0 (2. y) ©(4) (¡x. y.3. thus the derivative @©=@y.62) . y) ©(3) (¡x.is determined from the velocity potentials by: p = ¡½ @© ¡ ½gz @t ¶ µ @©r @©w @©d + + ¡ ½gz = ¡½ @t @t @t (2. Also displacements are often indicated in literature in the same way: x1 . :::x6 . z) coordinate system. y) = ¡©(4) (+x. at any time on both sides of the body must have the same direction. y) = +©(3) (+x.61) in which ~ is the outward normal vector on surface dS and ~ is the position vector of n r surface dS in the O(x.60) in which ©(i) is the velocity potential for the given direction i. over the submerged surface. For heave oscillations these velocities must be of opposite sign. y) = ¡©(2) (+x. This indicates that for sway and roll oscillations. Symmetric or Anti-symmetric Condition Since ships and many other ‡oating bodies are symmetric with respect to its middle line plane. 5. However. for all three modes of oscillations the vertical velocities. because f1 .59) 6. x2 . f5 and f6 .3. 2.via the linearized Bernoulli equation . thus the derivative @©=@x. :::6 are used here to indicate the mode of the motion. one can make use of this to simplify the potential equations: ©(2) (¡x. the horizontal velocities of the water particles. of the body: Z Z ~ = ¡ F (p ¢ ~ ) ¢ dS n ~ M = ¡ Z Z S S p ¢ (~ £ ~ ) ¢ dS r n (2. Radiation Condition The radiation condition states that when the distance R of a water particle to the oscillating body tends to in…nity. y) for sway for heave for roll (2.

©j . z.63) 2. which is associated with this oscillation in still water. Áj (x. ©j (x. waves are radiated from the body. z) ¢ vj (t) (2. y. y.65) in which the space and time dependent potential term. y. is now written in terms of a separate space dependent potential. z. y. t) = = 6 X j=1 6 X j=1 ©j (x. z. so that the hydromechanical forces ~ ~ F and moments M can be split into four parts too: ¶ ZZ µ @©r @©w @©d ~ = ½ F + + + gz ~ ¢ dS n @t @t @t ~ M = ½ or: ~ ~ ~ ~ ~ F = Fr + Fw + Fd + Fs ~ ~ ~ ~ ~ M = Mr + Mw + Md + Ms (2. y. STRIP THEORY METHOD which can obviously be split into four separate parts.67) (2.3. vj (t) in direction j. t) in direction j.66) and the generalized direction cosines are given by: @Áj fj = @n With this the radiation terms in the hydrodynamic force becomes: ¶ Z Z µ @©r ~ = ½ F ~ ¢ dS n @t S ! Z Z Ã X 6 @ = ½ Á vj ~ ¢ dS n @t j=1 j S (2.64) ZZ µ S S ¶ @©r @©w @©d + + + gz (~ £ ~ ) ¢ dS r n @t @t @t (2. can be written in terms.68) .3 Hydrodynamic Loads The hydrodynamic loads are the dynamic forces and moments caused by the ‡uid on an oscillating body in still water. t) Áj (x. z) in direction j.18 CHAPTER 2. for 6 degrees of freedom as: ©r (x. ©r . The radiation potential. This allows the normal velocity on the surface of the body to be written as: @©r @ X = ©j @n @n j=1 6 X ½ @Áj ¾ ¢vj = @n j=1 6 (2. multiplied by an oscillatory velocity.

Xr2 . This is generally true for all j and k in the range from 1 to 6. Oscillatory Motion Now an oscillatory motion is de…ned. Xr6 ) (2.70) @ X Á vj @t j=1 j @ X Á vj @t j=1 j 6 6 fk ¢ dS @Ák ¢dS @n for: k = 1. :::6 (2. Xr3 ) with: ~ Xrk = ½ Z Z Ã S Z Z Ã S and ! ! ~ Mr = (Xr4 . the force in one direction results from the motion in another direction. FLOATING RIGID BODY IN WAVES and the moment term: ~ M = ½ Z Z µ S 19 @©r @t = ½ Z Z Ã S ¶ 6 (~ £ ~ )dS r n ! (~ £ ~ ) ¢ dS r n (2. the force or moment is caused by a motion in that same direction. suppose a motion (in a complex notation) given by: sj = saj e¡i!t Then the velocity and acceleration of this oscillation are: sj = vj = ¡i!saj e¡i!t _ dvj sj = Ä = ¡i! 2 saj e¡i!t dt (2. When j 6= k.72) dvj ½ = dt Áj @Ák ¢dS @n (2. not everything is solved yet! A solution for this will be found later in this chapter. :::6 (2. The above equation expresses the force and moment components.2. Xr5 .73) This radiation force or moment Xrkj in the direction k is caused by a forced harmonic oscillation of the body in the direction j.75) . This introduces what is called coupling between the forces and moments (or motions).3. Áj .74) (2. When j = k.71) = ½ Since Áj and Ák are not time-dependent in this expression. Xrkj in terms of still unknown potentials.69) @ X Á vj @t j=1 j The components of these radiation forces and moments are de…ned by: ~ Fr = (Xr1 . it reduces to: Xrk = with: Xrkj 6 X j=1 Xrkj Z Z S for: k = 1.

Green’s second theorem. It is valid for any potential function.77) Mkj = ¡< ½ Áj k ¢dS Áj k ¢dS Nkj = ¡= ½! : : .80) Áj Ák ¢dS = ¢dS ¤ @n @n S¤ S¤ The boundary condition at the free surface becomes for © = Á ¢ e¡i!t : ¡! 2 Á + g @Á =0 @z for: z = 0 (2.76) So in case of an oscillation of the body in the direction j with a velocity potential Áj . .79) ¡ Ák ¢ dS ¤ Áj ¢ r Ák ¡ Ák ¢ r Áj ¢ dV = Áj @n @n V¤ S¤ This theorem is generally valid for all kinds of potentials. applied to these potentials. @n @n S S In case of an oscillation of the body in the direction k with a velocity potential Ák . regardless the fact if it ful…lls the Laplace condition or not. it is not necessary that they full…l the Laplace equation.6. In Green’s theorem. So the left hand side of the above equation becomes zero which yields for the right hand side of this equation: Z Z Z Z @Áj @Ák ¤ (2.78) ¢dS Njk = ¡= ½! ¢dS Mjk = ¡< ½ Ák Ák .81) . the hydrodynamic mass and damping (coupling) coe¢cients are de…ned by: 8 9 8 9 < Z Z = < Z Z = @Áj @Áj and (2. S. Consider two separate velocity potentials Áj and Ák . This volume is bounded by the wall of an imaginary vertical circular cylinder with a very large radius R. S ¤ is a closed surface with a volume V ¤ . the sea bottom at z = ¡h. the hydrodynamic mass and damping (coupling) coe¢cients are de…ned by: 8 8 9 9 < Z Z = < Z Z = @Á @Á and (2. Its mathematical background is beyond the scope of this text. see …gure 2. Both of the above radiation potentials Áj and Ák must ful…ll the Laplace equation (r2 Áj = r2 Ák = 0). . STRIP THEORY METHOD The hydrodynamic forces and moments can be split into a load in-phase with the acceleration and a load in-phase with the velocity: Xrkj = ¡Mkj sj ¡ Nkj sj Ä _ ¡ ¢ = saj ! 2 Mkj + isaj !Nkj e¡i!t 0 1 Z Z @Á = @(¡saj ! 2 ½ Áj k dS)A e¡i!t @n S (2. : : @n @n S S Green’s Second Theorem Green’s second theorem transforms a large volume-integral into a much easier to handle surface-integral. is then: ¶ Z Z µ Z Z Z ¡ ¢ @Áj @Ák ¤ 2 2 (2.20 CHAPTER 2. the water surface at z = ³ and the wetted surface of the ‡oating body.

! 2 =g = k tanh kh: k tanh kh ¢ Á = @Á @z for: z = 0 (2. Potential Coe¢cients The previous section provides . FLOATING RIGID BODY IN WAVES 21 Figure 2. at the free surface When taking also the boundary condition at the sea bed and the radiation condition on the wall of the cylinder in …gure 2.symmetry in the coe¢cients matrices with respect to their diagonals so that: Mjk = Mkj and Njk = Nkj (2.84) in which S is the wetted surface of the body only.3.83) R!1 into account. the integral equation over the surface S ¤ reduces to: ZZ S @Á Áj k ¢dS = @n ZZ S Ák @Áj ¢dS @n (2.82) This implies that at the free surface of the ‡uid one can write: k tanh kh ¢ Ák = k tanh kh ¢ Áj = @Ák @z @Áj @z = = @Ák @n @Áj @n ¡! Ák = ¡! Áj = 1 k tanh kh 1 k tanh kh ¢ ¢ @Ák @n @Áj @n 9 = .85) .6: Boundary Conditions or with the dispersion relation.2. Note that the Áj and Ák still have to be evaluated.6: @Á = 0 (for: z = ¡h) and @n lim Á = 0 (2.for the zero forward ship speed case .

y. t) and ©d (x. the symmetry of terms about the diagonal in these matrices (for example that M13 = M31 for zero forward speed) has not been included here. Áw (x. z). y. z. z. @©=@n = 0.91) . y. t). t) = Ád (x. v(t) = 1 ¢ e¡i!t : ©w (x. z. y. The terms on the diagonals (such as Mnn for example) are the primary coe¢cients relating properties such as hydrodynamic mass in one direction to the inertia forces in that same direction.4 Wave and Di¤raction Loads The wave and di¤raction terms in the hydrodynamic force and moment are: ¶ Z Z µ @©w @©d ~ ~w + Fd = ½ + ~ dS n F @t @t S (2.22 CHAPTER 2. 1962]. z.87) For clarity. ©w (x.3. y. multiplied by a normalized oscillatory velocity. y.89) The principles of linear superposition allow the determination of these forces on a restrained body with zero forward speed. y. but is not discussed here. are written now in terms of isolated space dependent potentials.88) and: ~ ~ Mw + Md = ½ Z Z µ S @©w @©d + @t @t ¶ (~ £ ~ )dS r n (2. Forward speed has an e¤ect on the velocity potentials itself. This e¤ect is quite completely explained by [Timman and Newman. O¤diagonal terms (such as M13 ) represent hydrodynamic mass only which is associated with an inertia dependent force in one direction caused by a motion component in another. z) ¢ e¡i!t ©d (x. z) ¢ e¡i!t (2. y. STRIP THEORY METHOD zero and the two matrices with 0 M15 0 M24 0 M26 0 M35 0 M44 0 M46 0 M55 0 M64 0 M66 1 C C C C C C A (2. This simpli…es the boundary condition on the surface of the body to: @©w @©d @© (2. z) and Ád (x. t) = Áw (x. 2.90) = + = 0 @n @n @n The space and time dependent potentials.86) Because of the symmetry of a ship some coe¢cients are hydrodynamic coe¢cients for a ship become: 0 M11 0 M13 B 0 M22 0 B B M31 0 M33 B Hydrodynamic mass matrix: B 0 M42 0 B @ M51 0 M53 0 M62 0 0 B B B B B B @ Hydrodynamic damping matrix: N11 0 N13 0 N15 0 0 N22 0 N24 0 N26 N31 0 N33 0 N35 0 0 N42 0 N44 0 N46 N51 0 N53 0 N55 0 0 N62 0 N64 0 N66 1 C C C C C C A (2.

is known. as given in the previous section. :::6 (2. but the di¤raction potential.92) =¡ d @n @n With this and the expressions for the generalized direction-cosines it is found for the wave forces and moments on the restrained body in waves: Z Z ¡i!t Xwk = ¡i½e (Áw + Ád ) fk dS = ¡i½e¡i!t ZS Z S (Áw + Ád ) @Ák dS @n for: k = 1. The radiation potential has to be determined for the constant forward speed case. Áw . These relations are valid only for a ‡oating body with a zero time-averaged speed in all directions. taking an opposite sign into account. Ád . According to Newman’s approach the wave potential has to be de…ned in the moving O(x. Ák : Z Z Z Z @©k @Á (2.96) Xwk = ¡i½e Áw + Ák dS @n @n S This limiters the problem of the di¤raction potential because the expression for Xwk depends only on the wave potential Áw and the radiation potential Ák . found by [Haskind. has generalized the Haskind relations for a body with a constant forward speed. y.95) This elimination of the di¤raction potential results into the so-called Haskind relations: ¶ Z Z µ @Ák @Áw ¡i!t for: k = 1.97) = ¢e ¢e e ! so that the isolated space dependent term is given by: Áw = ¡i³ a g kz ik(x cos ¹+y sin ¹) ¢e ¢e ! (2. has to be determined.2. FLOATING RIGID BODY IN WAVES This results into: 23 @Áw @Á (2. Green’s second theorem provides a relation between this di¤raction potential. [Newman.3. they underlie the relative motion (displacement . He derived equations which di¤er only slightly from those found by Haskind.98) . 1962] however. These relations.93) in which Ák is the radiation potential. now becomes: ¡³ a g kz ©w = ¢ e ¢ sin(!t ¡ kx cos ¹ ¡ ky sin ¹) ! ¡i³ a g kz ik(x cos ¹+y sin ¹) ¡i!t (2. :::6 (2. and a radiation potential.velocity .acceleration) hypothesis. The corresponding wave potential for deep water. z) system.94) Ád dS = Ák d dS @n @n S S and with @Áw =@n = ¡@Ád =@n one …nds: Z Z Z Z @©k @Á Ád dS = ¡ Ák w dS @n @n S S (2. The potential of the incident waves. are very important. 1957]. as used in strip theory. Ád .

101) or more generally: Xsk = ½g ZZ S zfk ¢ dS for: k = 1. For ships and other ‡oating structures. This holds that small e¤ects such as for instance a decreasing mass due to fuel consumption . the wave loads are given by: Z Z ¡i!t Xwk = ¡ i½e Áw fk dS + i½e¡i!t k Z SZ S Áw Ák ff3 + i(f1 cos ¹ + f2 sin ¹)g dS for:k = 1.102) in which the Xsk are the components of these hydrostatic forces and moments. The second term is caused by the disturbance caused by the presence of the (…xed) body. :::6 (2. this assumption is normally valid during a time which is large relative to the period of the motions.100) The …rst term in this expression for the wave forces and moments is the so-called FroudeKrilov force or moment. which is the wave load caused by the undisturbed incident wave.5 Hydrostatic Loads In the notations used here.3.103) B 0 0 0 Ixx 0 ¡Ixz C B C @ 0 0 0 0 Iyy 0 A 0 0 0 ¡Izx 0 Izz . The velocity of the water particles in the direction of the outward normal n on the surface of the body is: µ ¶¾ ½ @Áw @x @y @z = Áw k +i cos ¹ + sin ¹ @n @n @n @n (2. STRIP THEORY METHOD In these equations is ¹ the wave direction. the buoyancy forces and moments are: ZZ ZZ ~ s = ½g ~s = ½g z~ ¢ dS and M n z (~ x ~ ) ¢ dS r n F S S (2.can be ignored. The solid mass matrix of a ‡oating structure is given below. 0 1 ½r 0 0 0 0 0 B 0 ½r 0 0 0 0 C B C B 0 0 ½r 0 0 0 C C Solid mass matrix: m = B (2. :::6 (2. 2.99) = Áw k ff3 + i(f1 cos ¹ + f2 sin ¹)g With this.4 Equations of Motion The total mass as well as its distribution over the body is considered to be constant with time.24 CHAPTER 2. 2.

EQUATIONS OF MOTION 25 The moments of inertia here are often expressed in terms of the radii of inertia and the solid mass of the structure.105) in which L is the length and B is the breadth of the ship. kyy and kzz .107) in which: resulting external force acting in the center of gravity mass of the rigid body instantaneous velocity of the center of gravity resulting external moment acting about the center of gravity instantaneous angular momentum about the center of gravity time Two important assumptions are made for the loads in the right hand side of these equations: a. are usually smaller in the full load condition than in the ballast condition for normal ships. Thus.4.2. Ixz = Izx . . G. the (generally small) coupling terms.cargo plays an important role. Often. The so-called hydromechanical forces and moments are induced by the harmonic oscillations of the rigid body. The vector equations for the translations of and the rotations about the center of gravity are given respectively by: d ³ ~´ ~ mU F = dt ~ F m ~ U ~ M ~ H t = = = = = = and d ³~´ ~ M= H dt (2.more or less amidships laden . the radii of inertia can be t 0:30 ¢ B to 0:40 ¢ B t 0:22 ¢ L to 0:28 ¢ L t 0:22 ¢ L to 0:28 ¢ L (2. the radii of inertia. the mass is concentrated at the ends (engine room aft and ballast water forward to avoid a large trim). Note that the longitudinal radius of gyration of a long ´ homogeneous rectangular beam with a length L is equal to ³ p about 0:29 ¢ L = 1=12 ¢ L . are simply neglected.106) kxx = 0:289 ¢ B ¢ 1:0 + B in which KG is the height of the center of gravity. above the keel. while for ships with cargo on board (full load condition) the . For many ships without cargo on board (ballast condition). Since Archimedes law is valid for a ‡oating structure: Ixx = kxx2 ¢ ½r Iyy = kyy 2 ¢ ½r Izz = kzz 2 ¢ ½r When the distribution of the solid mass approximated by: 8 < kxx kyy for ships: : kzz (2. moving in the undisturbed surface of the ‡uid. Bureau Veritas proposes for the radius of inertia for roll of the ship’s solid mass: Ã ¶2 ! µ 2 ¢ KG (2.104) of a ship is unknown. The equations of motions of a rigid body in a space …xed coordinate system follow from Newton’s second law.

and z-directions respectively exciting wave moments about the x-. y. It is obvious that the remaining motions sway. STRIP THEORY METHOD b. y. Symmetric and anti-symmetric motions are not coupled. Xw2 . that is to say that a point to starboard has the same motion as the mirrored point to port side. Xh3 Xh4 . y. heave and pitch motions.110) Roll : Ixx ¢ Á ¡Xh4 = Xw4 Ä ¡Izx ¢ Á ¡Xh = Xw . these loads are added to obtain the total loads. y. Ä Pitch : Ixx ¢ µ ¡Xh5 = Xw5 9 Ä Sway : ½r ¢ y ¡Ixz ¢ Ã ¡Xh2 = Xw2 = Ä Ä anti-symmetric motions (2. y. Thus. but will not result in sway. Xh6 Xw1 . The so-called wave exciting forces and moments are produced by waves coming in on the restrained body. Because of this symmetry and anti-symmetry. a ship has a vertical-longitudinal plane of symmetry. after assuming small motions. roll and yaw are anti-symmetric motions. they don’t have any e¤ect on each other. a vertical force acting at the center of gravity can cause surge. heave and pitch motions are symmetric motions.108) in which: ½ r Iij Xh1 . Xw5 .109) . Xw6 = = = = = = = density of water volume of displacement of the ship solid mass moment of inertia of the ship hydromechanical forces in the x-. Surge. Since the system is linear. Ä Yaw : Izz ¢ Ã 6 6 . For instance.and z-axes respectively Generally. one can write the following six motion equations for the ship: Surge: Sway: Heave: Roll: Pitch: Yaw: d (½r ¢ x) _ = ½r ¢ x Ä dt d (½r ¢ y) _ = ½r ¢ y Ä dt d (½r ¢ z) _ = ½r ¢ z Ä dt ´ d ³ _ _ Ä Ä Ixx ¢ Á ¡ Ixz ¢ Ã = Ixx ¢ Á ¡ Ixz ¢ Ã dt ´ d ³ _ Iyy ¢ µ = Ixx ¢ Ä µ dt ´ ³ d _ _ Ä Ä Izz ¢ Ã ¡ Izx ¢ Á = Izz ¢ Ã ¡ Izx ¢ Á dt = Xh1 + Xw1 = Xh2 + Xw2 = Xh3 + Xw3 = Xh4 + Xw4 = Xh5 + Xw5 = Xh6 + Xw6 (2.and z-axes are principal axes. roll or yaw motions. symmetry of the body and that the x-. two sets of three coupled equations of motion can be distinguished for ships: 9 Surge : ½r ¢ x Ä ¡Xh1 = Xw1 = Heave : ½r ¢ z Ä ¡Xh3 = Xw3 symmetric motions (2. Xh2 . so that its motions can be split into symmetric and anti-symmetric components.26 CHAPTER 2. Xh5 . Xw3 Xw4 .and z-directions respectively hydromechanical moments about the x-.and z-axes respectively exciting wave forces in the x-.

Ä +b66 ¢ Ã +c66 ¢ Ã = Xw _ (yaw) +(+Izz + a66 ) ¢ Ã 6 (2. Each slice is treated hydrodynamically as if it is a segment of an in…nitely long ‡oating cylinder. Strip theory considers a ship to be made up of a …nite number of transverse twodimensional slices which are rigidly connected to each other. This means that all waves which are produced by the oscillating ship (hydromechanical loads) and the di¤racted waves (wave loads) are assumed to travel perpendicular to the middle line plane (thus parallel to the y-z plane) of the ship.7.2. the pitch motions can generate roll motions via the anchor lines.112) In many applications.111) (2. 2. Each of these slices will have a form which closely resembles the segment of the ship which it represents. This holds too that the . see …gure 2. Ixz = Izx is not known or small. Ä _ +(+Iyy + a55 ) ¢ µ +b55 ¢ µ +c55 ¢ µ = Xw5 (pitch) 9 (½r + a22 ) ¢ y +b22 ¢ y +c22 ¢ y Ä _ > > > > Ä _ > +a24 ¢ Á +b24 ¢ Á +c24 ¢ Á > > Ä +b26 ¢ Ã +c26 ¢ Ã = Xw2 (sway) > _ > +a26 ¢ Ã > > > > > > > > Ä _ a42 ¢ y +b42 ¢ y +c42 ¢ y = anti-symmetric Ä +b44 ¢ Á +c44 ¢ Á _ +(+Ixx + a44 ) ¢ Á motions > Ä _ > (roll) > +(¡Ixz + a46 ) ¢ Ã +b46 ¢ Ã +c46 ¢ Ã = Xw4 > > > > > > > > Ä _ a62 ¢ y +b62 ¢ y +c62 ¢ y > > > Ä +b64 ¢ Á +c64 ¢ Á _ > +(¡Izx + a64 ) ¢ Á > > .5 Strip Theory Approaches Strip theory is a computational method by which the forces on and motions of a threedimensional ‡oating body can be determined using results from two-dimensional potential theory. for instance. these equations can be solved with a numerical method. hence their terms are often omitted. After the determination of the in and out of phase terms of the hydromechanical and the wave loads. The coupled surge. STRIP THEORY APPROACHES 27 Note that this distinction between symmetric and anti-symmetric motions disappears when the ship is anchored. roll and yaw equations of motion are: a31 ¢ x +b31 ¢ x +c31 ¢ x Ä _ symmetric Ä _ +(½r + a33 ) ¢ z +b33 ¢ z +c33 ¢ z > motions _ +a35 ¢ Ä +b35 ¢ µ +c35 ¢ µ = Xw3 (heave) > > µ > > > > > > > > a51 ¢ x +b51 ¢ x +c51 ¢ x Ä _ > > > > +a53 ¢ z +b53 ¢ z +c53 ¢ z Ä _ > . Then. heave and pitch equations of motion are: (½r + a11 ) ¢ x +b11 ¢ x +c11 ¢ x Ä _ +a13 ¢ z +b13 ¢ z +c13 ¢ z Ä _ Ä +b15 ¢ µ +c15 ¢ µ = Xw1 _ +a15 ¢ µ 9 > > > > > > > (surge) > > > > > > > > > = The coupled sway.5.

the loads on the body are found by an integration of the 2-D loads: Z L Z L Z L Z L surge: sway: heave: roll: pitch: Xh1 = Xh2 = Xh3 = Xh4 = 0 Xh1 ¢ dxb Xw1 = Xw2 = Xw3 = Xw4 = Z L Z L Z L Z L 0 Xw1 ¢ dxb 0 Xw2 ¢ dxb 0 Xw3 ¢ dxb 0 Xw4 ¢ dxb 0 Xw3 ¢ xb ¢ dxb 0 Xh2 ¢ dxb 0 Xh3 ¢ dxb 0 Xh4 ¢ dxb 0 Xh3 Xh5 = ¡ Z L ¢ xb ¢ dxb Xw5 = ¡ Z L .28 CHAPTER 2. In spite of this restriction.5.7: Strip Theory Representation by Cross Sections strip theory supposes that the fore and aft side of the body (such as a pontoon) does not produce waves in the x direction. at least from a practical point of view. STRIP THEORY METHOD Figure 2. For the zero forward speed case. interactions between the cross sections are ignored as well. 2. Fundamentally. strip theory is valid for long and slender bodies only. experiments have shown that strip theory can be applied successfully for ‡oating bodies with a length to breadth ratio larger than three. (L=B = 3).1 Zero Forward Speed When applying the strip theory.

All two-dimensional potential coe¢cients have been de…ned so far in an axis system with the origin.115) is an operator which transforms the potentials ©(x0 . ³ hj .2. the 2-D hydrodynamic sway coe¢cients of this equivalent cross section are translated to 2-D hydrodynamic surge coe¢cients by an empirical method based on theoretical results from three-dimensional calculations and these coe¢cients are used to determine 2-D loads. In this way. the hydromechanical and exciting wave moments have to be corrected for the distance OG. Then. Inaccuracies of the hydromechanical coe¢cients of (slender) ships are of minor importance. t). h in which: n o _ ¤ = D ³¤ ³ hj Dt hj D = Dt ½ and n o Ĥ = D ³ ¤ _ ³ hj Dt hj ¾ (2. ³ ¤ j . y. 2. ³ hj . while the roll moments are obtained directly.114) @ @ ¡V ¢ @t @x (2. because these coe¢cients are relatively small. all sets of six surge loads can be treated in the same numerical way in SEAWAY for the determination of the 3-D loads. velocities. and accelerations. the equivalent j ¤ _¤ components of water particle displacements. and accelerations. Relative to a restrained ship. STRIP THEORY APPROACHES yaw: Xh6 = Z L 0 Xh2 29 ¢ xb ¢ dxb Xw6 = Z L 0 Xw2 ¢ xb ¢ dxb (2. ³ ¤ j . by de…ning an equivalent longitudinal cross section which is swaying. Äwj . t). ³ wj . y0 . velocities. z. The equations of motions are de…ned in the moving axis system with the origin at the average position of the center of gravity.2 Forward Ship Speed Relative to an oscillating ship moving forward in the undisturbed surface of the ‡uid. in ³ w a cross section are de…ned in a similar way by: ³¤ j . the _¤ Ĥ equivalent displacements.113) in which: 0 Xhj 0 Xwj = sectional hydromechanical force or moment in direction j per unit ship length = sectional exciting wave force or moment in direction j per unit ship length The appearance of two-dimensional surge forces seems strange here. respectively.116) . O. in the water plane. G. de…ned in the earth bounded (…xed) coordinate system. Note how in the strip theory the pitch and yaw moments are derived from the 2-D heave and sway forces.5. to the potentials ©(x. z0 . w n o _ ¤ = D ³¤ ³ wj Dt wj and n o Ĥ = D ³ ¤ _ ³ wj Dt wj (2. de…ned in the ship’s steadily translating coordinate system with speed V . at forward ship speed h V in the arbitrary direction j of a water particle in a cross section are de…ned by: ³ ¤j . It is strange! A more or less empirical method is used in SEAWAY for the surge motion.5. moving forward with speed V in waves.

Two di¤erent types of strip theory methods are discussed here: 1. When a ship sails with speed V and a trim angle. the uncoupled two-dimensional potential hydromechanical loads and wave loads in an arbitrary direction j are de…ned by: ¤ Xhj = ¤ Xwj This in the …rst formulation of the strip theory that can be found in the literature. it is obvious that in the equations of motion longitudinal derivatives of 0 0 the two-dimensional potential mass Mij and damping Nij will appear. It contains a more or less intuitive approach to the forward speed problem.3 End-Terms From the previous. Ordinary Strip Theory Method According to this classic method. 1957] and others. 0 Xrsj is the two-dimensional quasi-static restoring spring term. 0 0 In the equations above. these loads become: ¶ ¾ ½µ D i 0 ¤ ¤ 0 0 _ Xhj = Mjj ¡ Njj ¢ ³ hj + Xrsj Dt !e ½µ ¶ ¾ D i 0 ¤ ¤ 0 0 _ Mjj ¡ Njj ¢ ³ wj + Xf kj Xwj = Dt !e D n 0 Mjj Dt D n 0 = Mjj Dt o 0 0 _¤ _¤ ¢ ³ hj + Njj ¢ ³ hj + Xrsj o 0 0 _¤ _¤ ¢ ³ wj + Njj ¢ ³ wj + Xf kj (2.118) This formulation is a more fundamental approach of the forward speed problem. one should prefer the use of the modi…ed method.117) (2. Xf kj is the two-dimensional Froude-Krilov force or moment which is calculated by an integration of the directional pressure gradient in the undisturbed wave over the cross sectional area of the hull. derived from these Froude-Krilov loads. are used to calculate the di¤raction parts of the total wave forces and moments. generally present for heave. Mjj and Njj are the 2-D potential mass and damping coe¢cients.5. as published in detail by [Tasai. but it appeared from user’s experience that for ships with moderate forward speed (F n 5 0:30). through still water. 0 roll and pitch only. the ordinary method provides a better …t with experimental data. STRIP THEORY METHOD The e¤ect of this operator can be understood easily when one realizes that in that earthbound coordinate system the sailing ship penetrates through a ”virtual vertical disk”. These derivatives . 1969] and others. 2. Equivalent directional components of the orbital acceleration and velocity.30 CHAPTER 2. the relative vertical velocity of a water particle with respect to the bottom of the sailing ship becomes V ¢ µ. µ. From a theoretical point of view. Modi…ed Strip Theory Method According to this modi…ed method. Thus from a practical point of view. the use of the ordinary method is advised generally. as published in detail by [Korvin-Kroukovsky and Jacobs. 2.

8.2. while f (xb ) is equal to the local values of Mjj (xb ) or Njj (xb ).8: Integration of Derivatives The numerical integrations of the derivatives are carried out in the region xb (0) · xb · xb (L) only. Because the integration of the derivatives has to be carried out in the region xb (0) ¡ " · xb · xb (L) + ". see …gure 2. Figure 2.119) xb (0) 0 0 with " ¿ L.5. some algebra provides the integral and the …rst and second order moments (with respect to G) over the whole ship length: xbZ (L)+" xbZ (L)+" df (xb ) ¢ dxb = 0 dxb df (xb ) ¢ xb ¢ dxb = ¡ dxb xb (L) Z xb (0)¡" f(xb ) ¢ dxb xb (0)¡" xb (0) . So. the additional so-called ”end terms” are de…ned by f(0) and f (L). STRIP THEORY APPROACHES 31 have to be determined numerically over the whole ship length in such a manner that the following relation is ful…lled: xbZ (L)+" df (xb ) dxb = dxb xb (0)¡" xb (0)¡" xb (0) Z df (xb ) dxb + dxb xb (L) Z xb (0) xb (L) Z df(xb ) dxb + dxb xb (L) xbZ (L)+" df (xb ) dxb dxb = f (0) + = 0 df(xb ) dxb ¡ f (L) dxb (2.

of which the cross sections are simply a region of connected line elements. 2. 1949] derived an analytical solution for solving the problem of calculating the hydrodynamic coe¢cients of an oscillating circular cylinder in the surface of a ‡uid. the theory of [Keil. Generally. 1959] and [Tasai. 1974] . (b) Deep Water Coe¢cients by Close-Fit Conformal Mapping A more accurate mapping has been added by [Tasai. 2. a maximum of 10 coe¢cients are used for de…ning the cross section. these expressions can not be used. Also for the wave loads. Frank’s Pulsating Source Theory for Deep Water Mapping methods require an intersection of the cross section with the water plane and. as a consequence of this.120) xb (0)¡" xb (0) Note that these expressions are valid for the integrations of the potential coe¢cients over the full ship length only. This transformation is carried out by using a scale factor and two mapping coe¢cients. heave and roll coe¢cients can be calculated by several methods: 1. They can not be used for calculating local hydromechanical loads. The accuracy obtained depends on the number of mapping coe¢cients. the draft and the area of the mapped cross section will be equal to that of the actual cross section.which is a very simple and in a lot of cases also more or less realistic method to transform ship-like cross sections to this unit circle . such as are sometimes present in the aft body of a ship.32 xbZ (L)+" CHAPTER 2. they are not suitable for submerged cross sections. [Frank. (a) Deep Water Coe¢cients by Lewis Conformal Mapping [Tasai.6 Hydrodynamic Coe¢cients The two-dimensional hydrodynamic sway. STRIP THEORY METHOD df (xb ) 2 ¢ xb ¢ dxb = ¡2 dxb xb (L) Z f(xb ) ¢ xb ¢ dxb (2. Methods based on Ursell’s Theory and Conformal Mapping [Ursell. Also. These coe¢cients are determined in such a way that the Root Mean Square of the di¤erences between the o¤sets of the mapped and the actual cross section is minimal.to Ursell’s solution . like at a bulbous bow. He developed an integral equation method utilizing the Green’s function which represents a complex potential of a pulsating . by using more than only two mapping coe¢cients.and Lewis conformal mapping can be used. Only the breadth. conformal mapping can fail for cross sections with very low sectional area coe¢cients. 1967] considered a cylinder of constant cross sections with an arbitrarily symmetrical shape. 1961] added the so-called Lewis transformation . This vertical cross section can be fully or partly immersed in a previously undisturbed ‡uid of in…nite depth. because there these derivatives are multiplied with xb -depending orbital motions.based on a variation of Ursell’s potential theory for circular cylinders at deep water . 1960] too. (c) Shallow Water Coe¢cients by Lewis Conformal Mapping For shallow water.

i. The linearized Bernoulli equation provides the pressures after which the potential coe¢cients were obtained from the in-phase and out-of-phase components of the resultant hydrodynamic loads.. The 3-D coe¢cients follow from an integration of these 2-D coe¢cients over the ship’s length. Viscous terms can be added for surge and roll. For each frequency. the two-dimensional potential hydrodynamic sway coe¢cient of this equivalent cross section is translated to two-dimensional potential hydrodynamic surge coe¢cients by an empiric method.6. An equivalent longitudinal cross section has been de…ned. the distance of the cross section to the center of gravity G. .e. HYDRODYNAMIC COEFFICIENTS 33 point source of unit strength at the midpoint of each line element.2. A more or less empiric procedure has followed by the author for the surge motion. The 2-D potential pitch and yaw (moment) coe¢cients follow from the previous heave and sway coe¢cients and the arm. which is based on theoretical results of three-dimensional calculations. Wave systems were de…ned in such a way that all required boundary conditions were ful…lled.

CHAPTER 2.34 . STRIP THEORY METHOD .

M.19”. Mekelweg 2. the cross sections have to be mapped conformally to the unit circle. Ship Hydromechanics Laboratory.Chapter 3 Conformal Mapping Ursell’s derivation of potential coe¢cients is valid for circular cross sections only. N ) number of parameters N X¡ n=0 a¡(2n¡1 ³ 2n¡1 ¢ From this follows the relation between the coordinates in the z-plane (the ship’s cross section) and the variables in the ³-plane (the circular cross section): N X© ª x = ¡Ms (¡1)n a2n¡1 e¡(2n¡1)® sin ((2n ¡ 1)µ) y = +Ms 0 n=0 N X n=0 © ª (¡1)n a2n¡1 e¡(2n¡1)® cos ((2n ¡ 1)µ) J. The advantage of conformal mapping is that the velocity potential of the ‡uid around an arbitrarily shape of a cross section in a complex plane can be derived from the more convenient circular section in another complex plane.wbmt.nl. ”Theoretical Manual of SEAWAY. For the determination of the two-dimensional added mass and damping in the sway. The Netherlands.tudelft. :::. heave and roll mode of the motions of ship-like cross sections by Ursell’s method. In this manner hydrodynamic problems can be solved directly with the coe¢cients of the mapping function.nl/~johan or http://www.J. Journée. 2628 CD Delft. Delft University of Technology. February 2001.shipmotions. Report 1216a. Release 4. 35 . For updates see web site: http://dutw189. The general transformation formula is given by: z = Ms with: z = x + iy ³ = ie® e¡iµ Ms a¡1 a2n¡1 N = = = = = = plane of the ship’s cross section plane of the unit circle scale factor +1 conformal mapping coe¢cients (n = 1.

1969].ship’s cross section follows from putting ® = 0 in the previous relations: x0 = ¡Ms y0 = +Ms N X n=0 f(¡1)n a2n¡1 sin ((2n ¡ 1)µ)g f(¡1)n a2n¡1 cos ((2n ¡ 1)µ)g N X n=0 The breadth on the waterline of the approximate ship’s cross section is de…ned by: b0 = 2Ms ¸a with scale factor: Ms = and the draft is de…ned by: d0 = Ms ¸b with: ¸b = N X n=0 with: ¸a = N X n=0 a2n¡1 b0 2¸a f(¡1)n a2n¡1 g 3. CONFORMAL MAPPING Figure 3.1: Mapping Relation between Two Planes The contour of the . 1929]). A extended description of the representation of ship hull forms by Lewis two-parameter conformal mapping is given by [Kerczek and Tuck. .1 Lewis Conformal Mapping A very simple and in a lot of cases also a more or less realistic transformation of the cross sectional hull form will be obtained with N = 2 in the transformation formula. the well known Lewis transformation ([Lewis.by conformal mapping approximated .36 CHAPTER 3.

Putting a1 .3. LEWIS CONFORMAL MAPPING This two-parameter Lewis transformation of a cross section is de…ned by: ¡ ¢ z = Ms ¢ a¡1 ¢ ³ + a1 ¢ ³ ¡1 + a3 ¢ ³ ¡3 37 In here a¡1 =+1 and the conformal mapping coe¢cients a1 and a3 are called Lewis coe¢cients. derived from the expression for H0 . into the expression for ¾ s yields a quadratic equation in a3 : c1 a2 + c2 a3 + c3 = 0 3 in which: c1 c2 c3 µ ¶ µ ¶2 4¾ s 4¾ s H0 ¡ 1 = 3+ + 1¡ ¢ ¼ ¼ H0 + 1 = 2c1 ¡ 6 = c1 ¡ 4 . The half breadth to draft ratio H0 is given by: H0 = Bs =2 1 + a1 + a3 = Ds 1 ¡ a1 + a3 An integration of the Lewis form delivers the sectional area coe¢cient ¾ s : ¾s = As ¼ 1 ¡ a2 ¡ 3a2 1 3 = ¢ Bs Ds 4 (1 + a3 )2 ¡ a2 1 in which As is the area of the cross section. while Ms is the scale factor.1. Then: ¡ ¢ x = Ms ¢ e® sin µ + a1 e¡® sin µ ¡ a3 e¡3® sin 3µ ¡ ¢ y = Ms ¢ e® cos µ ¡ a1 e¡® cos µ + a3 e¡3® cos 3µ By putting ® = 0 is the contour of this so-called Lewis form expressed as: x0 = Ms ¢ ((1 + a1 ) sin µ ¡ a3 sin 3µ) y0 = Ms ¢ ((1 ¡ a1 ) cos µ + a3 cos 3µ) with the scale factor: Ms = in which: Bs Ds = sectional breadth on the load water line = sectional draft Bs =2 1 + a1 + a3 or: Ms = Ds 1 ¡ a1 + a3 Now the coe¢cients a1 and a3 and the scale factor Ms will be determined in such a manner that the sectional breadth. draft and area of the approximate cross section and of the actual cross section are identical.

38 CHAPTER 3. Figure 3. can be distinguished: ² re-entrant forms. bounded by: for H0 · 1:0 : for H0 ¸ 1:0 : ¾s ¾s < < 3¼ (2 ¡ H0 ) 32 µ ¶ 3¼ 1 2¡ 32 H0 µ ¶ 3¼ 3¼ H0 (2 ¡ H0 ) < ¾ s < 3+ 32 32 4 ¶ ¶ µ µ 3¼ 1 3¼ 1 < ¾s < 2¡ 3+ 32 H0 32 4H0 3¼ 32 µ 1 3+ 4H0 ¶ ² bulbous and not-tunneled forms. these solutions are not considered. a1 = 0 and a3 = 0.2. with the regions of the half breadth to draft ratio H0 and the area coe¢cient ¾ s to match as presented before.1 Boundaries of Lewis Forms In some cases the Lewis transformation can give more or less ridiculous results. The following typical Lewis hull forms. with a minus sign before the square root: p ¡c1 + 3 ¡ 9 ¡ 2c1 a3 = c1 are looped.2: Typical Lewis Forms 3. bounded by: µ ¶ H0 3¼ H0 · 1:0 and 3+ < 32 4 ¾s < . bounded by: for H0 · 1:0 : for H0 ¸ 1:0 : ² conventional forms.1. Some typical and realistic Lewis forms are presented in …gure 3. Since ships are ”better behaved”. CONFORMAL MAPPING The (valid) solutions for a3 and a1 will become: a3 a1 p ¡c1 + 3 + 9 ¡ 2c1 = c1 H0 ¡ 1 = ¢ (a3 + 1) H0 + 1 Lewis forms with the other solution of a3 in the quadratic equation. they intersect themselves at a point within the fourth quadrant. It is obvious that a transformation of a half immersed circle with radius R will result in Ms = R.

bulbous forms and tunneled forms are considered to be valid forms here.1.3. Figure 3. LEWIS CONFORMAL MAPPING ² tunneled and not-bulbous forms. So conventional forms. the area coe¢cient ¾ s is . bounded by: µ µ ¶ ¶ 1 3¼ ¼ 1 for H0 · 1:0 : 3+ < ¾s < 10 + H0 + 32 4H0 32 H0 µ ¶ µ ¶ 3¼ H0 ¼ 1 for H0 ¸ 1:0 : 3+ < ¾s < 10 + H0 + 32 4 32 H0 ² non-symmetric forms. bounded by: 0 < H0 < 1 and ¾s > ¼ 32 µ ¶ 1 1 + H0 + H0 These ranges of the half breadth to draft ratio H0 and the area coe¢cient ¾ s for the di¤erent typical Lewis forms are shown in …gure 3.2 Acceptable Lewis Forms Not-acceptable forms of ships are supposed to be the re-entrant forms and the asymmetric forms. To obtain ship-like Lewis forms.1. see …gure 3.3.3. bounded by: µ ¶ 3¼ 1 H0 ¸ 1:0 and < 3+ 32 4H0 39 ¾s < 3¼ 32 µ ¶ H0 3+ 4 ² combined bulbous and tunneled forms.3: Ranges of H0 and ¾ s of Lewis Forms 3.

to calculate the Lewis coe¢cients. 1985] by taking into account the vertical position of the centroid of the cross section. are avoided when the following requirements are ful…lled: Bs > °Ds 2 with for instance ° = 0:01. the contour of this approximate form is expressed as: x0 = Ms ((1 + a1 ) sin µ ¡ a3 sin 3µ + a5 sin 5µ) y0 = Ms ((1 ¡ a1 ) cos µ + a3 cos 3µ ¡ a5 cos 5µ) with the scale factor: Ms = and: Bs Ds = sectional breadth on the waterline = sectional draft Bs =2 1 + a1 + a3 + a5 or: Ms = Ds 1 ¡ a1 + a3 ¡ a5 . Numerical problems.2 Extended Lewis Conformal Mapping Somewhat better approximations will be obtained by taking into account also the …rst order moments of half the cross section about the x0 .40 CHAPTER 3. 1974] and have been simpli…ed by [Athanassoulis and Loukakis. The three-parameter Extended-Lewis transformation is de…ned by: Z = Ms (a¡1 ³ + a1 ³ ¡1 + a3 ³ ¡3 + a5 ³ ¡5 ) in which a¡1 = +1. CONFORMAL MAPPING bounded by a lower limit to omit re-entrant Lewis forms and by an upper limit to omit non-symmetric Lewis forms: µ ¶ 3¼ ¼ 1 for H0 · 1:0 : ¾s < (2 ¡ H0 ) < 10 + H0 + 32 32 H0 µ ¶ µ ¶ 3¼ 1 ¼ 1 for H0 ¸ 1:0 : < ¾s < 2¡ 10 + H0 + 32 H0 32 H0 If a value of ¾ s is outside of this range it has to be set to the value of the nearest border of this range. So: x = Ms (e® sin µ + a1 e¡® sin µ ¡ a3 e¡3® sin 3µ + a5 e¡5® sin 5µ) y = Ms (e® cos µ ¡ a1 e¡® cos µ + a3 e¡3® cos 3µ ¡ a5 e¡5® cos 5µ) By putting ® = 0. These two additions to the Lewis formulation were proposed by [Reed and Nowacki. for instance with bulbous or aft cross sections of a ship.and y0 -axes. This has been done by extending the Lewis transformation to N = 3 in the general transformation formula. and Ds > ° Bs 2 3.

except the sectional breadth.3 Close-Fit Conformal Mapping A more accurate transformation of the cross sectional hull form can be obtained by using a greater number of parameters N . also the centroids of the approximate cross section and the actual cross section of the ship have an equal position.3. 3. . the equations above can be solved in an iterative manner. a3 and a5 and the scale factor Ms can be determined in such a manner that. A very simple and straight on iterative least squares method of the author to determine the Close-Fit conformal mapping coe¢cients will be described here shortly. The half beam to draft ratio is given by: H0 = Bs =2 Ds = 1 + a1 + a3 + a5 1 ¡ a1 + a3 ¡ a5 An integration of the approximate form results into the sectional area coe¢cient: ¾s = As Bs Ds = ¼ 1 ¡ a2 ¡ 3a2 ¡ 5a2 1 3 5 ¢ 4 (1 + a3 )2 ¡ (a1 + a5 )2 For the relative distance of the centroid to the keel point a more complex expression has been obtained by [Athanassoulis and Loukakis. CLOSE-FIT CONFORMAL MAPPING 41 Now the coe¢cients a1 . 1985]: 3 3 3 PPP · = KB Ds (Aijk a2i¡1 a2j¡1 a2k¡1 ) 3 P = 1¡ i=0 j=0 k=0 H0 ¾ s i=0 a3 2i¡1 in which: Aijk 1 = 4 µ 1 ¡ 2k 1 ¡ 2k 1 ¡ 2k 1 ¡ 2k ¡ + + 3 ¡ 2(i + j + k) 1 ¡ 2(i ¡ j + k) 1 ¡ 2(i + j ¡ k) 1 ¡ 2(¡i + j + k) ¶ The following requirements should be ful…lled when also bulbous cross sections are allowed: ² re-entrant forms are avoided when both the following requirements are ful…lled: 1 ¡ a1 ¡ 3a3 ¡ 5a5 > 0 1 + a1 ¡ 3a3 + 5a5 > 0 ² existence of a point of self-intersection is avoided when both the following requirements are ful…lled: 9a2 + 145a2 + 10a3 a5 + 20H0 a5 > 0 3 5 2 2 9a3 + 145a5 ¡ 10a3 a5 ¡ 20H0 a5 > 0 Taking these restrictions into account.3. draft and area.

have been determined successfully from the o¤sets of various cross sections in such a manner that the mean squares of the deviations of the actual cross section from the approximate described cross section is minimized. yi ) to the normal of the contour to the actual cross section through (x0i .4: Close-Fit Conformal Mapping . With these [Ms ¢a2n¡1 ] values. yi ) lies on the normal of the approximate contour of the cross section in (x0i . Now µi has to be determined. y0i ). while the initial values of a5 until a2N¡1 are set to zero.42 CHAPTER 3. CONFORMAL MAPPING The scale factor Ms and the conformal mapping coe¢cients a2n¡1 . a µi -value is determined for each o¤set in such a manner that the actual o¤set (xi . y0i ).4. will be de…ned by the distance from the o¤set (xi . see …gure 3. The general transformation formula is given by: N o Xn Z = Ms a2n¡1 ³ ¡(2n¡1) n=0 in which: a¡1 = +1. Then the contour of the approximate cross section is given by: x0 = ¡Ms y0 = +Ms with the scale factor: Ms = Bs =2 n=0 N P N X n=0 f(¡1)n a2n¡1 sin ((2n ¡ 1)µ)g f(¡1)n a2n¡1 cos ((2n ¡ 1)µ)g N X n=0 or: Ms = Ds n=0 N P fa2n¡1 g f(¡1)n a2n¡1 g The procedure starts with initial values for [Ms ¢ a2n¡1 ]. with a maximum value of n varying from N = 2 until N = 10. The initial values of Ms . Therefore a function F (µi ). Figure 3. a1 and a3 are obtained with the Lewis method as has been described before.

3. :::N which are rewritten as: N X = n=0 I X i=0 (¡1)n [Ms ¢ a2n¡1 ] I X i=0 fcos ((2j ¡ 2n)µi )g ) = f¡xi sin ((2j ¡ 1)µi ) + yi cos ((2j ¡ 1)µi )g for: j = 0. yi ) from (x0i . the numerical value of he sum of the squares of deviations is known: I X E= fei g i=0 Then new values of [Ms ¢ a2n¡1 ] have to be obtained in such a manner that E is minimized. y0i ) is calculated: ei = (xi ¡ x0i )2 + (yi ¡ y0i )2 After doing this for all I + 1 o¤sets. eventually with somewhat more dense o¤sets near sharp corners. CLOSE-FIT CONFORMAL MAPPING 43 These o¤sets have to be selected at approximately equal mutual circumferential lengths. Then ®i is de…ned by: +xi+1 ¡ xi¡1 cos ®i = p (xi+1 ¡ xi¡1 )2 + (yi+1 ¡ yi¡1 )2 ¡yi+1 + yi¡1 sin ®i = p (xi+1 ¡ xi¡1 )2 + (yi+1 ¡ yi¡1 )2 With this µi -value.3. so: @E =0 @fMs a2j¡1 g for: j = 0. This means that each of the derivatives of this equation with respect to each coe¢cients [Ms ¢ a2n¡1 ] is zero. :::N . :::N The sum of the squares of these deviations can also be expressed as: 8Ã !2 I N X< X E = x + f(¡1)n [Ms ¢ a2n¡1 ] sin ((2n ¡ 1)µi )g : i n=0 i=0 Ã !2 9 N = X n +( yi ¡ f(¡1) [Ms ¢ a2n¡1 ] cos ((2n ¡ 1)µ i )g . the numerical value of the square of the deviation of (xi . n=0 This yields N+1 equations: ( I N X X ¡ sin ((2j ¡ 1)µi ) f(¡1)n[Ms ¢ a2n¡1 ] sin ((2n ¡ 1)µi )g + i=0 n=0 ¡ cos ((2j ¡ 1)µi ) = I X i=0 N X n=0 f(¡1)n [Ms ¢ a2n¡1 ] cos ((2n ¡ 1)µi )g ) = fxi sin ((2j ¡ 1)µi ) ¡ yi cos ((2j ¡ 1)µi )g ( for: j = 0.

0 0 0 4 . CONFORMAL MAPPING To obtain the exact breadth and draft. Reference is also given here to a report of [Jong.9 5 0 9 .1 3 5 0 .0 0 0 2 . so that new values for [Ms ¢ a2n¡1 ] will be obtained.2 7 0 0 .4 Comparisons A …rst example has been given here for the amidships cross section of a container vessel. depending on the dimensions of the cross section. This procedure will be repeated several times and stops when the di¤erence between the numerical E-values of two subsequent calculations becomes less than a certain threshold value ¢E. One can then switch to Lewis coe¢cients with an area coe¢cient of the cross section set to the nearest border of the valid Lewis form area.0 0 0 0 .7 4 0 1 2 .7 0 0 1 2 . etc.0 0 0 9 . with o¤sets as tabled below: i (¡) 0 1 2 3 4 5 6 7 8 Ds ¡ y i (m) 0 . 1973].40 meter and a draft of 9.0 5 0 6 .0 0 0 xi (m) 0 . methods are described to determine the scale factor Ms and the conformal mapping coe¢cients a2n¡1 from the o¤sets of a cross section. These new values are used instead of the initial values to obtain new µi -values of the I + 1 o¤sets again. 3. :::N ¡ 2 f[Ms ¢ a2n¡1 ]g = Bs =2 for: j = N ¡ 1 for: j = N These N + 1 equations can be solved numerically.9 0 0 1 0 .7 0 0 . the last two equations are replaced by the equations for the breadth at the water line and the draft: (N ) I X X f(¡1)n [Ms ¢ a2n¡1 ]g fcos ((2j ¡ 2n)µi )g = n=0 i=0 = I X i=0 f¡xi sin ((2j ¡ 1)µi ) + yi cos ((2j ¡ 1)µi )g N X n=0 for: j = 0.44 CHAPTER 3. In that report several other.5 0 0 1 .9 6 0 1 1 .7 0 0 1 2 .. for instance: ³ ´2 p ¢E = (I + 1) 0:00005 b2 + d2 max max in which: bmax dmax = maximum half breadth of the cross section = maximum height of the cross section N X n=0 f(¡1)n [Ms ¢ a2n¡1 ]g = Ds Because a¡1 = +1 the scale factor Ms is equal to the …nal solution of the …rst coe¢cient (n = 0). The N other coe¢cients a2n¡1 can be found by dividing the …nal solutions of [Ms ¢ a2n¡1 ] by this Ms -value.00 meter. In the most worse case a ”maximum” value of N will be attained without success. In these cases the number N will be increased until the divergence or re-entrance vanish. with a breadth of 25. suitable but more complex.4 4 0 1 2 . Attention has to be paid to divergence in the calculation routines and re-entrant forms.0 0 0 3 .

3 1 9 0 + 1 .0 0 0 0 -0 .1 6 3 2 -0 .0 0 2 4 + 0 .1 6 3 1 -0 .3.0 0 0 0 + 0 .3 1 9 1 + 1 .0 0 3 2 -0 .0 0 0 7 -0 .07 6 0 .0 0 0 0 + 0 .0 00 0 + 0 .24 00 + 1.0 0 2 1 1 2 . L ew is C o n fo rm al M a p p in g N -Pa ra m e ter C lo se F it C o n fo rm a l M a p p in g N 2N -1 (¡) (¡) 2 3 2 3 3 5 4 7 5 9 6 11 7 13 8 15 9 17 10 19 Ms a¡1 a1 a3 a5 a7 a9 a11 a13 a15 a17 a19 RMS (m) (¡) (¡) (¡) (¡) (¡) (¡) (¡) (¡) (¡) (¡) (¡) (m) 1 2.1 2 4 3 -0 .0 0 2 4 1 2 .1 81 0.1 6 3 2 -0 .1 6 3 2 -0 .3 1 8 6 + 1 .0 0 2 6 + 0 .1 2 4 4 -0 .0 0 1 4 + 0 ..0 0 0 0 + 0 .1 2 4 5 -0 .0 0 2 0 + 0 .0 1 0 8 + 0 .0 0 1 6 1 2 . The calculated data of the two-parameter Lewis and the Nparameter Close-Fit conformal mapping of this amidships cross section are tabled below.3 1 8 3 + 1 .4.0 0 1 2 + 0 .3 1 9 4 + 1 .0 0 0 6 0 .0 0 2 3 + 0 .0 0 0 0 + 0 .0 0 2 2 + 0 .0 0 0 0 + 0 .0 1 7 0 .1 2 4 5 -0 .11 36 12 .0 0 3 1 -0 .0 1 0 5 + 0 .0 0 5 4 -0 .3 1 9 3 + 1 . 33 new o¤sets at equidistant length intervals on the contour of this cross section can be determined by a second degree interpolation routine.1 2 4 6 -0 .0 1 0 8 + 0 .0 3 9 0 .0 13 4 1 2 .0 0 3 0 -0 .0 0 2 9 + 0 .1 6 3 3 -0 .0 1 0 7 + 0 .0 0 3 1 -0 .0 0 2 1 + 0 .2 45 7 + 1 .1 64 0 -0 .0 0 0 9 1 2 .00 meter and a draft of 10.0 1 8 0 .1 2 4 5 -0 .1 14 0 12 .0 0 0 0 + 0 .1 2 4 5 -0 .1 6 3 3 -0 .00 meter.0 1 0 8 + 0 . which shows the di¤erences between a Lewis transformation and a 10-parameter close-…t conformal mapping of a rectangular cross section with a breadth of 20.5.3 1 9 5 + 1 .1 51 1 -0 .15 11 -0.0 0 3 0 -0 .5: Lewis and Close-Fit Conformal Mapping of a Rectangle .0 00 0 + 0 .1 16 7 -0 .0 0 2 9 + 0 .00 00 + 0.0 0 1 5 + 0 .0 1 0 7 + 0 .0 0 0 2 1 2 .0 2 7 0 .0 0 0 0 + 0 .0 1 9 0 .0 0 1 5 + 0 .0 1 3 3 + 0 .0 0 8 0 .1 6 3 4 -0 . COMPARISONS 45 For the least squares method in the conformal mapping method. The last line lists the RMS-values for the deviations of the 33 equidistant points on the approximate contour of this cross section.0 0 9 Another example is given in …gure 3.18 0 0.2 84 1 + 1 .0 0 5 3 1 2 . Figure 3.0 0 0 2 + 0 .

46 . CHAPTER 3. CONFORMAL MAPPING .

19”. used in the SEAWAY program. The signs of the uncoupled sway. in this section deviating axes systems are used for the determination of the twodimensional hydrodynamic potential coe¢cients for sway. Ship Hydromechanics Laboratory. February 2001. 0 47 .nl. The two-dimensional hydrodynamic potential coe¢cients have been de…ned here with respect to the O(x.e. an approximation is given for the determination of the surge coe¢cients M11 and 0 N11 . heave and roll motions. will be used. the distance of the cross section to the center of gravity G.wbmt. The Netherlands. the following two-dimensional hydrodynamic coe¢cients have to be obtained: 0 M22 0 M24 0 M33 0 M44 0 M42 and and and and and 0 N22 0 N24 0 N33 0 N44 0 N42 = = = = = 2-D 2-D 2-D 2-D 2-D potential potential potential potential potential mass mass mass mass mass and and and and and damping damping damping damping damping coe¢cients of sway coupling coe¢cients of roll into sway coe¢cients of heave coe¢cients of roll coupling coe¢cients of sway into roll The 2-D potential pitch and yaw (moment) coe¢cients follow from the previous heave and sway coe¢cients and the arm. This holds for the sway and roll coupling coe¢cients a change of sign. Release 4. Report 1216a. based on a variation of Ursell’s potential theory for circular cylinders and Lewis conformal mapping ² the theory of Frank for deep water. i. ”Theoretical Manual of SEAWAY. 0 Finally.. based on Ursell’s potential theory for circular cylinders and N-parameter conformal mapping ² the theory of Keil for shallow and deep water. J. heave and roll coe¢cients do not change. z) coordinate system for the moving ship in still water.J. Journée.M. using pulsating sources on the cross sectional contour. Mekelweg 2.shipmotions. 2628 CD Delft. Delft University of Technology.nl/~johan or http://www. For each cross section. to obtain the 2-D potential coe¢cients: ² the theory of Tasai for deep water. y.2. However.Chapter 4 2-D Potential Coe¢cients This chapter described the various methods.tudelft. For updates see web site: http://dutw189. as shown in …gure 2. During the ship motions calculations di¤erent coordinate systems.

1959] and [Tasai. Figure 4.48 CHAPTER 4. Tasai’s notations have been maintained here as far as possible. 1973]. is based on work published by [Ursell.1 Heave Motions The determination of the hydrodynamic coe¢cients of a heaving cross section of a ship in deep and still water at zero forward speed. Starting points for the derivation these coe¢cients here are the velocity potentials and the conjugate stream functions of the ‡uid as they have been derived by Tasai and also by [Jong. Respectively. 1949].1: Axes System for Heave The cylinder is forced to carry out a simple harmonic vertical motion about its initial position with a frequency of oscillation ! and a small amplitude of displacement ya : y = ya cos(!t + ±) in which ± is a phase angle. 1949]. the determination of the hydrodynamic coe¢cients of a heaving. of which a cross section is given in the next …gure. 4. 2-D POTENTIAL COEFFICIENTS 4. 1973]. 1961] and [Jong. 1960].1 Theory of Tasai In this section. 1959]. [Tasai. [Tasai. 1960]. Suppose an in…nite long cylinder in the surface of a ‡uid. the vertical velocity and acceleration of the cylinder are: y = ¡!ya sin(!t + ±) _ y = ¡! 2 ya sin(!t + ±) Ä This forced vertical oscillation of the cylinder causes a surface disturbance of the ‡uid. [Tasai. as described here. [Tasai. swaying and rolling cross section of a ship in shallow at zero forward speed is based on work published by [Ursell.1. .

² A regular progressive wave system. The amplitudes of these waves decrease strongly with the distance to the cylinder. denoted here by subscript B.1. The two-dimensional velocity potential of the ‡uid has to ful…l the following requirements: 1. Because the heave motion of the ‡uid is symmetrical about the y-axis. Two kinds of waves will be produced: ² A standing wave system. this free surface condition can be written as: N ª @© »b X © ¼ = 0 for: ® ¸ 0 and µ = § © (2n ¡ 1)a2n¡1 e¡(2n¡1)® § ¾ a n=0 @µ 2 in which: »b !2 Ms = ¾a g or: » b = ! 2 b0 (non-dimensional frequency squared) 2g From the de…nition of the velocity potential follows the boundary condition on the surface of the cylinder for ® = 0: @©0 (µ) @y0 =y _ @n @n in which n is the outward normal of the cylinder surface. The linearized free surface condition in deep water is expressed as follows: @© !2 ©+ =0 g @y for: jxj ¸ Bs 2 and y = 0 In consequence of the conformal mapping. At a distance of a few wave lengths from the cylinder. THEORY OF TASAI 49 Because the cylinder is supposed to be in…nitely long.4. y) from which follows: @© =0 @µ for: µ = 0 3. These waves dissipate energy. denoted here by subscript A. the generated waves will be twodimensional. y) = ©(+x. this velocity potential has the following relation: ©(¡x. provided that this amplitude is su¢ciently small compared with the radius of the cylinder and the wave length is not much smaller than the diameter of the cylinder. the waves on each side can be described by a single regular wave train. These waves travel away from the cylinder and a stationary state is rapidly attained.. The wave amplitude at in…nity ´ a is proportional to the amplitude of oscillation of the cylinder ya . The velocity potential must satisfy to the equation of Laplace: r2 © = @2© @2© + 2 =0 @x2 @y 2. .

The following set of velocity potentials. µ) = e¡2m® cos(2mµ) ¾ N ½ »b X 2n ¡ 1 ¡(2m+2n¡1)® n a2n¡1 e cos ((2m + 2n ¡ 1)µ) ¡ (¡1) ¾ a n=0 2m + 2n ¡ 1 The set of conjugate stream functions is expressed as: Ã1 ! 1 ª X© ª g´a X © ªA = P2m à A2m (®. When de…ning: h(µ) = 2x0 b0 = ¡ Integration results into the following requirement for the stream function on the surface of the cylinder: N X ª0 (µ) = yMs _ f(¡1)n a2n¡1 sin ((2n ¡ 1)µ)g n=0 N 1 X f(¡1)n a2n¡1 sin ((2n ¡ 1)µ)g ¾ a n=0 the stream function on the surface of the cylinder is given by: ª0 (µ) = ¡y _ b0 h(µ) + C(t) 2 Because of the symmetry of the ‡uid about the y-axis. as they are given by [Tasai. µ) cos(!t) + Q2m à A2m (®.50 CHAPTER 4. µ) sin(!t) ©a = ¼! m=1 m=1 ª0 (µ) = ¡y _ in which: ÁA2m (®. this boundary condition on the surface of the cylinder (®=0) reduces to: @y0 ¡@ª0 (µ) = y _ @µ @® = ¡yMs _ N X n=0 f(¡1)n (2n ¡ 1)a2n¡1 cos ((2n ¡ 1)µ)g in which C(t) is a function of the time only. [Tasai. µ) sin(!t) ¼! m=1 m=1 . 2-D POTENTIAL COEFFICIENTS Using the stream function ª. µ) cos(!t) + Q2m ÁA2m (®. ful…l these requirements: à 1 ! 1 ª X© ª g´ a X © P2m ÁA2m (®. 1960] and [Jong. 1973]. it is clear that: C(t) = 0 So: b0 h(µ) 2 For the standing wave system a velocity potential and a stream function satisfying to the equation of Laplace. the symmetrical motion of the ‡uid and the free surface condition has to be found. 1959].

It is therefore necessary to add a stream function. In these expressions the magnitudes of the P2m and the Q2m series follow from the boundary conditions as will be explained further on. satisfying the free surface condition and the symmetry about the y-axis. a function describing a source at the origin O is chosen. representing such a train of waves at in…nity. 1973] gave the velocity potential of the progressive wave system as: ¢ g´ ¡ ©B = a ÁBc (x. 1960] and [Jong. 1959]. µ) sin(!t) ¼! When calculating the integrals in the expressions for ÁBs and ÁBc numerically. . Another requirement is a diverging wave train for ® goes to in…nity. [Tasai. y) sin(!t) ¼! in which: ÁBc = ¼e¡ºy cos(ºx) ÁBs = ¼e¡ºy sin(ºjxj) + while: Z1 0 º sin(ky) ¡ k cos(ky) ¡kjxj e dk k2 + º 2 !2 (wave number for deep water) º= g Changing the parameters delivers: ©B = ¢ g´a ¡ ÁBc (®.4. y) cos(!t) + ÁBs (x. µ) cos(!t) + ÁBs (®. THEORY OF TASAI in which: 51 à A2m (®. µ) cos(!t) + à Bs (®. µ) = e¡2m® sin(2mµ) ¾ N ½ »b X 2n ¡ 1 ¡(2m+2n¡1)® n ¡ (¡1) a2n¡1 e sin ((2m + 2n ¡ 1)µ) ¾ a n=0 2m + 2n ¡ 1 These sets of functions tend to zero as ® tends to in…nity. µ) sin(!t) ¼! The conjugate stream function is given by: ªB = in which: ¢ g´ a ¡ à Bc (x. y) sin(!t) ¼! à Bc = ¼e¡ºy sin(ºjxj) à Bs = ¡¼e ¡ºy cos(ºx) + Z1 0 º cos(ky) + k sin(ky) ¡kjxj e dk k2 + º 2 Changing the parameters delivers: ªB = ¢ g´ a ¡ à Bc (®.1. [Tasai. y) cos(!t) + à Bs (x. the convergence is very slowly. For this.

so for ® = 0. This has been shown before for the sway case. µ) When putting ® = 0. µ) cos(!t) m=1 m=1 1 X m=1 + à Bs (®. µ) sin(!t) + ÁBs (®.52 CHAPTER 4. µ) + g´ a ª(®. µ) and ÃBs (®. µ) at the surface of the cylinder. µ) + cos(!t) ¼! m=1 ! ! à 1 X© ª Q2m ÁA2m (®. the stream function is equal to the expression found before from the boundary condition on the surface of the cylinder: Ãà ! 1 X© ª g´ a ª0 (µ) = à B0c (µ) + P2m à A02m (µ) cos(!t) ¼! m=1 à ! ! 1 X© ª + ÃB0s (µ) + Q2m à A02m (µ) sin(!t) m=1 b0 = ¡y h(µ) _ 2 in which: à A02m (µ) = sin(2mµ) ¾ N ½ 2n ¡ 1 »b X n (¡1) a2n¡1 sin ((2m + 2n ¡ 1)µ) ¡ ¾ a n=0 2m + 2n ¡ 1 In this expression à B0c (µ) and à B0s (µ) are the values of ÃBc (®. 2-D POTENTIAL COEFFICIENTS Power series expansions. µ) = P2m ÁA2m (®. µ) + P2m à A2m (®. 1960]. The summation in these expansions converge much faster than the numeric integration procedure. µ) ÁBc (®. as given by [Porter. can be used instead of these last integrals over k. So for each µ. The total velocity potential and stream function to describe the waves generated by a heaving cylinder are: © = ©A + ©B ª = ªA + ªB So the velocity potential and the conjugate stream function are expressed by: Ãà ! 1 X© ª g´ a ©(®. the following equation has been obtained: à ! 1 X© ª à B0c (µ) + P2m à A02m (µ) cos(!t) ! © ª ¼!b0 + à B0s (µ) + Q2m à A02m (µ) sin(!t) = ¡y _ h(µ) 2g´ a m=1 à m=1 1 X . µ) + ! ! ª © sin(!t) Q2m à A2m (®. µ) = ¼! à Ãà ! 1 X© ª à Bc (®.

are required to solve these equations. 1 X ¾ N X ½ 2n ¡ 1 »b m a2n¡1 à A02m (¼=2) = (¡1) ¾a 2m + 2n ¡ 1 n=0 m=1 1 X m=1 ff2m (µ)P2m g ff2m (µ)Q2m g . results for any µ-value less than ¼=2 into a set of two equations with the unknown parameters P2m and Q2m . Then the summations over µ can be replaced by integrations. less than ¼=2. the best …t values of P2m and Q2m are supposed to be those found by means of a least squares method. Another favorable method is to multiply both sides of the equations with ¢µ. so at the intersection of the surface of the cylinder with the free surface of the ‡uid where h(µ)=1.1. For practical reasons the maximum value of m is limited to M. So: à B0c (µ) ¡ h(µ)à B0c (¼=2) = à B0s (µ) ¡ h(µ)à B0s (¼=2) = in which: f2m (µ) = ¡Ã A02m (µ) + h(µ)à A02m (¼=2) The series in these two sets of equations converges uniformly with an increasing value of m.4. Note that at least M values of µ. For a lot of µ-values. THEORY OF TASAI The right hand side of this equation can be written as: ¡y _ in which: A0 = ¼!b0 ya h(µ) = h(µ) ¼» b sin(!t + ±) 2g´ a ´a = h(µ) (A0 cos(!t) + B0 sin(!t)) ya ¼» sin ± ´a b and B0 = ya ¼» cos ± ´a b 53 This results for each µ into a set of two equations: à B0c (µ) + à B0s (µ) + 1 X© ª P2m à A02m (µ) = h(µ)A0 m=1 1 X m=1 ª © Q2m à A02m µ) = h(µ)B0 When putting µ = ¼=2. we obtain the coe¢cients A0 and B0 : A0 1 X© ª = à B0c (¼=2) + P2m à A02m (¼=2) m=1 1 X B0 = à B0s (¼=2) + in which: m=1 ª © Q2m à A02m (¼=2) A substitution of A0 and B0 into the set of two equations for each µ. Each µ-value less than ¼=2 will deliver an equation for the P2m and Q2m series.

one set for P2m and one set for Q2m : 8 9 ¼=2 ¼=2 Z Z M = X< ¢ ¡ P2m f2m (µ)f2n (µ)dµ = à B0c (µ) ¡ h(µ)à B0c (¼=2) f2n (µ)dµ : . m=1 0 0 9 8 ¼=2 ¼=2 Z Z M = X< ¢ ¡ f2m (µ)f2n (µ)dµ = Q2m à B0s (µ) ¡ h(µ)à B0s (¼=2) f2n (µ)dµ . µ) and ÁBs (®. two sets of M equations have been obtained. µ) at the surface of the cylinder. so for ® = 0. :::M m=1 Now the P2m and Q2m series can be solved by a numerical method. Now the hydrodynamic pressure on the surface of the cylinder can be obtained from the linearized equation of Bernoulli: p(µ) = ¡½ @©0 (µ) @t Ãà ! M X© ª ¡½g´ a = ÁB0s (µ) + cos(!t) Q2m ÁA02m (µ) ¼ m=1 à ! ! M X© ª ¡ ÁB0c (µ) + P2m ÁA02m (µ) sin(!t) m=1 It is obvious that this pressure is symmetric in µ. With these P2m and Q2m values. : 0 0 n = 1. :::M n = 1. 2-D POTENTIAL COEFFICIENTS Herewith.54 CHAPTER 4. . the coe¢cients A0 and B0 are known too. From the de…nition of these coe¢cients follows the amplitude ratio of the radiated waves and the forced heave oscillation: ´a ¼» b =p 2 ya A2 + B0 0 With the solved P2m and Q2m values. the velocity potential on the surface of the cylinder (®=0) is known too: Ãà ! M X© ª g´ a ÁB0c (µ) + cos(!t) ©0 (µ) = P2m ÁA02m (µ) ¼! m=1 à ! ! M X© ª + ÁB0s (µ) + sin(!t) Q2m ÁA02m (µ) m=1 in which: ÁA02m (µ) = cos(2mµ) ¾ N ½ 2n ¡ 1 »b X n a2n¡1 cos ((2m + 2n ¡ 1)µ) ¡ (¡1) ¾ a n=0 2m + 2n ¡ 1 In this expression ÁB0c (µ) and ÁB0s (µ) are the values of ÁBc (®.

can be found by integrating the vertical component of the hydrodynamic pressure on the surface of the cylinder: 0 Fy = ¡ +¼=2 Z p(µ) dx0 ds ds ¡¼=2 ¼=2 Z dx0 = ¡2 p(µ) dµ dµ 0 With this the two-dimensional hydrodynamic vertical force due to heave oscillations can be written as follows: ½gb0 ´ a 0 Fy = (M0 cos(!t) ¡ N0 sin(!t)) ¼ in which: M0 1 = ¡ ¾a 1 ¡ ¾a ¼=2 Z N X ÁB0s (µ) f(¡1)n (2n ¡ 1)a2n¡1 cos ((2n ¡ 1)µ)g dµ m=1 ¼» b + 2 4¾ a 0 M X Ã ( n=0 (¡1)m Q2m N X Q2 + m=1 ( N X½ n=0 (2n ¡ 1)2 a2n¡1 (2m)2 ¡ (2n ¡ 1)2 N ¡m X n=0 ¾) (¡1)m Q2m f(2n ¡ 1)a2n¡1 a2m+2n¡1 g )! and N0 as obtained from this expression for M0 . THEORY OF TASAI Heave Coe¢cients 55 The two-dimensional hydrodynamic vertical force. For the determination of M0 and N0 it is advised: M = N . 1960]. With: ½gb0 ´a 0 (M0 cos(!t + ± ¡ ±) ¡ N0 sin(!t + ± ¡ ±)) Fy = ¼ and: ´a ´a A0 cos ± = B0 sin ± = ya ¼» b ya ¼» b the two-dimensional hydrodynamic vertical force can be resolved into components in phase and out phase with the vertical displacement of the cylinder: 0 Fy = ½gb0 ´ 2 a ((M0 B0 + N0 A0 ) cos(!t + ±) + (M0 A0 ¡ N0 B0 ) sin(!t + ±)) ¼ 2 » b ya 0 0 0 Fy = ¡M33 y ¡ N33 y Ä _ This hydrodynamic vertical force can also be written as: 0 0 = M33 ! 2 ya cos(!t + ±) + N33 !ya sin(!t + ±) in which: .1. These expressions coincide with those given by [Tasai. acting on the cylinder in the direction of the y-axis.4. by replacing there ÁB0s (µ) by ÁB0c (µ) and Q2m by P2m .

found before. the following energy balance relation is found: M0 A0 ¡ N0 B0 = ¼2 2 .56 0 M33 0 N33 CHAPTER 4. the axes system of Tasai and the ship motions O(x. The energy delivered by the exciting forces should be equal to the energy radiated by the waves. So: T osc Z ½g´ 2 c 1 a 0 N33 y ¢ ydt = _ _ Tosc 2 0 in which Tosc is the period of oscillation. the two-dimensional hydrodynamic mass and damping coe¢cients of heave are given by: 0 M33 = ½b2 M0 B0 + N0 A0 0 ¢ 2 2 A2 + B0 0 and 0 N33 = ½b2 M0 A0 ¡ N0 B0 0 ¢ ¢! 2 2 A2 + B0 0 The signs of these two coe¢cients are proper in both. With the relation for the wave speed c = g=!. z) coordinate system. y. 2-D POTENTIAL COEFFICIENTS = 2-D hydrodynamic mass coe¢cient of heave = 2-D hydrodynamic damping coe¢cient of heave When using also the amplitude ratio of the radiated waves and the forced heave oscillation. follows the relation between the twodimensional heave damping coe¢cient and the amplitude ratio of the radiated waves and the forced heave oscillation: µ ¶2 ½g 2 ´ a 0 N33 = 3 ! ya With this amplitude ratio the two-dimensional hydrodynamic damping coe¢cient of heave is also given by: ½¼ 2 b2 1 0 0 N33 = ¢ 2 ¢! 2 4 A0 + B0 0 When comparing this expression for N33 with the expression found before.

THEORY OF TASAI 57 4.2 Sway Motions The determination of the hydrodynamic coe¢cients of a swaying cross section of a ship in deep and still water at zero forward speed. . Suppose an in…nite long cylinder in the surface of a ‡uid. 1973]. Figure 4. denoted here by subscript A. Respectively.1. Starting points for the derivation these coe¢cients here are the velocity potentials and the conjugate stream functions of the ‡uid as they have been derived by Tasai and also by [Jong. Two kinds of waves will be produced: ² A standing wave system. the lateral velocity and acceleration of the cylinder are: x = ¡!xa sin(!t + ") _ and x = ¡! 2 xa cos(!t + ") Ä This forced lateral oscillation of the cylinder causes a surface disturbance of the ‡uid. is based here on work published by [Tasai. Because the cylinder is supposed to be in…nitely long. of which a cross section is given in …gure 4.1. the generated waves will be twodimensional.2. 1961] for the Lewis method.2: Axes System for Sway The cylinder is forced to carry out a simple harmonic lateral motion about its initial position with a frequency of oscillation ! and a small amplitude of displacement xa : x = xa cos(!t + ") in which " is a phase angle. The amplitudes of these waves decrease strongly with the distance to the cylinder.4. These waves travel away from the cylinder and a stationary state is rapidly attained.

the waves on each side can be described by a single regular wave train. 2-D POTENTIAL COEFFICIENTS ² A regular progressive wave system. The velocity potential must satisfy to the equation of Laplace: r2 © = @2© @2© + 2 =0 @x2 @y 2. The linearized free surface condition in deep water is expressed as follows: !2 @© ©+ =0 g @y for: jxj ¸ Bs 2 and y = 0 In consequence of the conformal mapping. Because the sway motion of the ‡uid is not symmetrical about the y-axis. At a distance of a few wave lengths from the cylinder. denoted here by subscript B. The wave amplitude at in…nity ´ a is proportional to the amplitude of oscillation of the cylinder xa . y) 3.. The two-dimensional velocity potential of the ‡uid has to ful…l the following three requirements: 1. this velocity potential has the following anti-symmetric relation: ©(¡x. Using the stream function ª this boundary condition on the surface of the cylinder (® = 0) reduces to: @ª0 (µ) @x0 = ¡x _ @µ @® N X = ¡xMs _ f(¡1)n (2n ¡ 1)a2n¡1 sin ((2n ¡ 1)µ)g n=0 . provided that this amplitude is su¢ciently small compared with the radius of the cylinder and the wave length is not much smaller than the diameter of the cylinder. this last equation results into: N ª @© ¼ »b X © = 0 for: ® ¸ 0 and µ = § © (2n ¡ 1)a2n¡1 e¡(2n¡1)® § ¾ a n=0 @µ 2 in which: »b !2 Ms = ¾a g or: » b = ! 2 b0 (non-demensional frequency squared) 2g From the de…nition of the velocity potential follows the boundary condition on the surface S of the cylinder for ® = 0: @©0 (µ) @x0 =x _ @n @n in which n is the outward normal of the cylinder surface S. y) = ¡©(+x.58 CHAPTER 4. These waves dissipate energy.

representing such a train of waves at . the non-symmetrical motion of the ‡uid and the free surface condition. µ) cos(!t) + Q2m à A2m (®. When de…ning: g(µ) = 2y0 b0 = N 1 X f(¡1)n a2n¡1 cos ((2n ¡ 1)µ)g ¾ a n=0 the stream function on the surface of the cylinder is given by: _ ª0 (µ) = x b0 g(µ) + C(t) 2 For the standing wave system a velocity potential and a stream function satisfying to the equation of Laplace. µ) sin(!t) ªA = ¼! m=1 m=1 in which: à A2m (®. Therefore it is necessary to add a stream function. µ) sin(!t) ¼! m=1 m=1 in which: ÁA2m (®. ful…l these requirements: à 1 ! 1 X© ª X© ª g´ ©A = a P2m ÁA2m (®. µ) = e¡(2m+1)® sin ((2m + 1)µ) ¾ N ½ »b X ¡(2m+2n)® n 2n ¡ 1 ¡ (¡1) a2n¡1 e sin ((2m + 2n)µ) ¾ a n=0 2m + 2n The set of conjugate stream functions is expressed as: Ã1 ! 1 ª X© ª g´a X © P2m à A2m (®.4. THEORY OF TASAI 59 Integration results into the following requirement for the stream function on the surface of the cylinder: N X ª0 (µ) = xMs _ f(¡1)n a2n¡1 cos ((2n ¡ 1)µ)g + C(t) n=0 in which C(t) is a function of the time only.1. 1973]. µ) = ¡e¡(2m+1)® cos ((2m + 1)µ) ¾ N ½ »b X ¡(2m+2n)® n 2n ¡ 1 + (¡1) a2n¡1 e cos ((2m + 2n)µ) ¾ a n=0 2m + 2n These sets of functions tend to zero as ® tends to in…nity. The following set of velocity potentials. the non-symmetrical motion of the ‡uid and the free surface condition has to be found. Another requirement is a diverging wave train for ® goes to in…nity. µ) cos(!t) + Q2m ÁA2m (®. In these expressions the magnitudes of the P2m and the Q2m series follow from the boundary conditions. 1961] and [Jong. satisfying the equation of Laplace. as they are given by [Tasai. as will be explained further on.

y) sin(!t) ¼! in which: ÁBc ¢ j = ¡¼e¡ºy sin(ºjxj) ÁBs ¢ j = ¼e¡ºy cos(ºx) ¡ while: j = +1 for x > 0 j = ¡1 for x < 0 !2 (wave number for deep water) º = g Changing the parameters delivers: ¢ g´ ¡ ©B = a ÁBc (®. 1960]. µ) sin(!t) ¼! Z1 0 º cos(ky) + k sin(ky) ¡kjxj jxj e dk + 2 + º2 2 + y2) k º(x The conjugate stream function is given by: ¢ g´ ¡ ªB = a à Bc (x. µ) cos(!t) + ÁBs (®. Power series expansions. y) cos(!t) + ÁBs (x. can be used instead of these last integrals over k: Z1 0 º cos(ky) + k sin(ky) ¡kx e dk = (Q sin(ºx) ¡ (S ¡ ¼) cos(ºx)) e¡ºy 2 + º2 k º sin(ky) ¡ k cos(ky) ¡kx e dk = (Q cos(ºx) + (S ¡ ¼) sin(ºx)) e¡ºy k2 + º 2 Z1 0 . [Tasai. the convergence is very slowly. 1973] gave the velocity potential of the progressive wave system as: ¢ g´ ¡ ©B = a ÁBc (x. 1961] and [Jong. For this. 2-D POTENTIAL COEFFICIENTS in…nity. µ) sin(!t) ¼! When calculating the integrals in the expressions for à Bs and ÁBs numerically. y) cos(!t) + à Bs (x. µ) cos(!t) + à Bs (®.60 CHAPTER 4. a function describing a two-dimensional horizontal doublet at the origin O is chosen. as given by [Porter. y) sin(!t) ¼! in which: à Bc = ¼e¡ºy cos(ºx) à Bs = ¼e¡ºy sin(ºjxj) + Z1 0 º sin(ky) ¡ k cos(ky) ¡kjxj y e dk ¡ 2 + º2 2 + y2) k º(x Changing the parameters delivers: ¢ g´ ¡ ªB = a à Bc (®.

THEORY OF TASAI in which: 1 ´ X ³ p Q = ° + ln º x2 + y 2 + fpn cos(n¯)g n=1 61 S = ¯+ µ ¶ x ¯ = arctan y ´n ³ p 2 + y2 º x pn = n ¢ n! ° = 0:5772156649:::::: (Euler constant) The summation in these expansions converge much faster than the numeric integration procedure. µ) m=1 m=1 1 X m=1 1 X n=1 fpn sin(n¯)g + à Bs (®. µ) = g´ a ¼! à Ãà ! 1 X© ª à Bc (®. µ) cos(!t) ¼! m=1 à ! ! 1 X© ª sin(!t) + ÁBs (®. µ) + cos(!t) P2m à A2m (®. µ) + P2m ÁA2m (®. the stream function is equal to the expression found before from the boundary condition on the surface of the cylinder: Ãà ! 1 X© ª g´ a ª0 (µ) = à B0c (µ) + cos(!t) P2m à A02m (µ) ¼! m=1 à ! ! 1 X© ª + ÃB0s (µ) + Q2m à A02m (µ) sin(!t) m=1 ! ! © ª Q2m à A2m (®. µ) + When putting ® = 0. µ) + Q2m ÁA2m (®. The total velocity potential and stream function to describe the waves generated by a swaying cylinder are: © = ©A + ©B ª = ªA + ªB So the velocity potential and the conjugate stream function are expressed by: Ãà ! 1 X© ª g´ a ©(®.1. A suitable maximum value of n should be chosen. µ) ª(®.4. µ) = ÁBc (®. µ) sin(!t) b0 = x g(µ) + C(t) _ 2 in which: à A02m (µ) = ¡ cos ((2m + 1)µ) ¾ N ½ »b X n 2n ¡ 1 + (¡1) a2n¡1 cos ((2m + 2n)µ) ¾ a n=0 2m + 2n .

2-D POTENTIAL COEFFICIENTS In this expression ªB0c (µ) and ªB0s (µ) are the values of ªBc (®. so at the intersection of the surface of the cylinder with the free surface of the ‡uid where g(µ) = 0. the following equation has been obtained: à ! 1 X© ª à B0c (µ) + cos(!t)+ P2m à A02m (µ) à ! ª © ¼!b0 sin(!t) = x _ g(µ) + C ¤ (t) à B0s (µ) + Q2m à A02m (µ) 2g´ a m=1 m=1 1 X When putting µ = ¼=2. we obtain the constant C ¤ (t): à ! 1 X© ª à B0c (¼=2) + cos(!t) + C ¤ (t) = P2m à A02m (¼=2) à à B0s (¼=2) + m=1 1 X m=1 © ! ª Q2m à A02m (¼=2) sin(!t) in which: A substitution of C ¤ (t) in the equation for each µ-value. So for each µ-value. So: à B0c (µ) ¡ à B0c (¼=2) = g(µ)P0 + à B0s (µ) ¡ à B0s (¼=2) = g(µ)Q0 + 1 X m=1 1 X ff2m (µ)P2m g ff2m (µ)Q2m g m=1 .62 CHAPTER 4. so for ® = 0. µ) at the surface of the cylinder. results for any µ-value less than ¼=2 into the following equation: à ! 1 X© ¢ª ¡ à B0c (µ) ¡ à B0c (¼=2) + cos(!t)+ P2m à A02m (µ) ¡ à A02m (¼=2) à ! © ¡ ¢ª ¼! b0 g(µ) à B0s (µ) ¡ à B0s (¼=2) + Q2m à A02m (µ) ¡ à A02m (¼=2) sin(!t) = x _ g´a 2 m=1 m=1 1 X ¾ N X ½ 2n ¡ 1 »b m ªA02m (¼=2) = (¡1) a2n¡1 ¾a 2m + 2n n=0 The right hand side of this equation can be written as: µ ¶ ¼!b0 xa x _ g(µ) = g(µ) ¡ ¼» b sin(!t + ") 2g´a ´a = g(µ) (P0 cos(!t) + Q0 sin(!t)) in which: P0 = ¡ xa ¼» sin " ´a b and Q0 = ¡ xa ¼» cos " ´a b This delivers for any µ-value less than ¼=2 two sets of equations with the unknown parameters P2m and Q2m . µ) and ªBs (®.

Herewith. THEORY OF TASAI in which: f2m (µ) = ¡Ã A02m (µ) + Ã A02m (¼=2) These equations can also be written as: Ã B0c (µ) ¡ Ã B0c (¼=2) = Ã B0s (µ) ¡ Ã B0s (¼=2) = in which: for m = 0: for m > 0: f0 (µ) = g(µ) f2m (µ) = ¡Ã A02m (µ) + Ã A02m (¼=2) 1 X 63 m=0 1 X m=0 ff2m (µ)P2m g ff2m (µ)Q2m g The series in these two sets of equations converges uniformly with an increasing value of m. two sets of M + 1 equations have been obtained.4. are required to solve these equations. : M X 0 0 n = 0. the velocity potential on the surface of the cylinder (® = 0) is known too: ÃÃ ! 1 X© ª g´ a ( ÁB0c (µ) + ©0 (µ) = P2m ÁA02m (µ) cos(!t) ¼! m=1 Ã ! ! 1 X© ª + ÁB0s (µ) + Q2m ÁA02m (µ) sin(!t) m=1 . Then P0 and Q0 are known now and from the de…nition of these coe¢cients follows the amplitude ratio of the radiated waves and the forced sway oscillation: ´a ¼» b =p 2 xa P0 + Q2 0 With the solved P2m and Q2m values. :::M m=0 Now the P2m and Q2m series can be solved with a numerical method. Note that at least M + 1 values of µ. Then the summations over µ can be replaced by integrations. :::M n = 0. one set for P2m and one set for Q2m : 8 9 ¼=2 ¼=2 Z Z < = ¢ ¡ P2m f2m (µ)f2n (µ)dµ = Ã B0c (µ) ¡ Ã B0c (¼=2) f2n (µ)dµ : .1. m=0 0 0 9 8 ¼=2 ¼=2 Z Z M = X< ¡ ¢ f2m (µ)f2n (µ)dµ = Ã B0s (µ) ¡ Ã B0s (¼=2) f2n(µ)dµ Q2m . Each µ-value less than ¼=2 will deliver an equation for the P2m and Q2m series. For practical reasons the maximum value of m is limited to M. Another favorable method is to multiply both sides of the equations with ¢µ. The best …t values of P2m and Q2m are supposed to be those found by means of a least squares method. less than ¼=2.

Now the hydrodynamic pressure on the surface of the cylinder can be obtained from the linearized equation of Bernoulli: p(µ) = ¡½ @©0 (µ) @t ! ÃÃ 1 X© ª ¡½g´a cos(!t) = ÁB0s (µ) + Q2m ÁA02m (µ) ¼ m=1 Ã ! ! 1 X© ª ¡ ÁB0c (µ) + P2m ÁA02m (µ) sin(!t) m=1 It is obvious that this pressure is skew-symmetric in µ. can be found by integrating the lateral component of the hydrodynamic pressure on the surface S of the cylinder: 0 Fx ¼=2 Z ¡dy0 = ¡ (p(+µ) ¡ p(¡µ)) ds ds 0 ¼=2 Z dy0 dµ = 2 p(µ) dµ 0 With this the two-dimensional hydrodynamic lateral force due to sway oscillations can be written as follows: ¡½gb0 ´ a 0 (M0 cos(!t) ¡ N0 sin(!t)) Fx = ¼ in which: M0 1 = ¡ ¾a + ¼=2 Z N X ÁB0s (µ) f(¡1)n (2n ¡ 1)a2n¡1 sin ((2n ¡ 1)µ)g dµ 0 N¡1 X n=0 ¼ 4¾ a » + b ¾2 a m=1 m=1 ( M X f(¡1)m Q2m (2m + 1)a2m+1 g (¡1)m Q2m N N XX½ n=0 i=0 (2n ¡ 1)(2i ¡ 1) a2n¡1 a2i¡1 (2m + 2i)2 ¡ (2n ¡ 1)2 ¾) . µ) at the surface of the cylinder. so for ® = 0.64 in which: CHAPTER 4. Sway Coe¢cients The two-dimensional hydrodynamic lateral force. µ) and ÁBs (®. acting on the cylinder in the direction of the x¡axis. 2-D POTENTIAL COEFFICIENTS ÁA02m (µ) = sin ((2m + 1)µ) ¾ N ½ »b X n 2n ¡ 1 ¡ (¡1) a2n¡1 sin ((2m + 2n)µ) ¾ a n=0 2m + 2n In this expression ÁB0c (µ) and ÁB0s (µ) are the values of ÁBc (®.

the two-dimensional hydrodynamic mass and damping coe¢cients of sway are given by: 0 M22 ½b2 M0 Q0 + N0 P0 = 0¢ 2 2 P0 + Q2 0 and 0 N22 ½b2 M0 P0 ¡ N0 Q0 = 0¢ ¢! 2 2 P0 + Q2 0 The signs of these two coe¢cients are proper in both.4. by replacing there ÁB0s (µ) by ÁB0c (µ) and Q2m by P2m . found before. z) coordinate system. follows the relation between the twodimensional sway damping coe¢cient and the amplitude ratio of the radiated waves and the forced sway oscillation: µ ¶ ½g 2 ´ a 2 0 N22 = 3 ! xa With this amplitude ratio the two-dimensional hydrodynamic damping coe¢cient of sway is also given by: ½¼2 b2 1 0 0 N22 = ¢ 2 ¢! 4 P0 + Q2 0 . THEORY OF TASAI 65 and N0 as obtained from this expression for M0 . The energy delivered by the exciting forces should be equal to the energy radiated by the waves. With the relation for the wave speed c = g=!. y. So: T osc Z ½g´ 2 c 1 a 0 N22 x ¢ xdt = _ _ Tosc 2 0 in which Tosc is the period of oscillation.1. the axes system of Tasai and the ship motions O(x. For the determination of M0 and N0 it is required: M = N . With: ¡½gb0 ´ a 0 Fx = (M0 cos(!t + " ¡ ") ¡ N0 sin(!t + " ¡ ")) ¼ and: ¡´ a P0 ¡´a Q0 sin " = cos " = xa ¼» b xa ¼» b the two-dimensional hydrodynamic lateral force can be resolved into components in phase and out phase with the lateral displacement of the cylinder: 0 Fx = ½gb0 ´ 2 a ((M0 Q0 + N0 P0 ) cos(!t + ") + (M0 P0 ¡ N0 Q0 ) sin(!t + ")) ¼ 2 » b xa This hydrodynamic lateral force can also be written as: 0 0 0 Ä _ Fx = ¡M22 x ¡ N22 x 0 0 2 = M22 ! xa cos(!t + ") + N22 !xa sin(!t + ") in which: 0 M22 0 N22 = 2-D hydrodynamic mass coe¢cient of sway = 2-D hydrodynamic damping coe¢cient of sway When using also the amplitude ratio of the radiated waves and the forced sway oscillation.

2-D POTENTIAL COEFFICIENTS 0 When comparing this expression for N22 with the expression found before. The two-dimensional hydrodynamic moment acting on the cylinder in the clockwise direction can be found by integrating the roll component of the hydrodynamic pressure on the surface S of the cylinder: 0 MR ¼=2 ¶ µ Z +dx0 ¡dy0 = (p(+µ) ¡ p(¡µ)) ¡x0 + y0 ds ds ds 0 ¼=2 µ ¶ Z dx0 dy0 = ¡2 p(µ) x0 + y0 dµ dµ dµ 0 With this the two-dimensional hydrodynamic roll moment due to sway oscillations can be written as follows: ½gb2 ´ a 0 0 (YR cos(!t) ¡ XR sin(!t)) MR = ¼ in which: YR 1 = 2¾ 2 a ¼=2 Z N N XX© ª ÁB0s (µ) (¡1)n+i (2i ¡ 1) ¢ a2n¡1 a2i¡1 sin ((2n ¡ 2i)µ) dµ 0 N ¼» b X f(¡1)m Q2m ¢ ¡ 3 8¾ a m=1 Ã N n¡m ½ X X (¡2m + 2n ¡ 2i ¡ 1)(2i ¡ 1) ¾ N N M X X ½ (2i ¡ 1)(2n ¡ 2i) 1 X m ¢ a2n¡1 a2i¡1 (¡1) Q2m + 2 2¾ a m=1 (2m + 1)2 ¡ (2n ¡ 2i)2 n=0 i=0 ( n=0 i=0 ) + n=m i=0 N¡m X X N n=0 i=m+n ½ 2n ¡ 2i ¢ a2n¡1 a2i¡1 a¡2m+2n¡2i¡1 ¾ (¡2m ¡ 2n + 2i ¡ 1)(2i ¡ 1) ¢ a2n¡1 a2i¡1 a¡2m¡2n+2i¡1 2n ¡ 2i ¾!) and XR as obtained from this expression for YR .66 CHAPTER 4. by replacing there ÁB0s (µ) by ÁB0c (µ) and Q2m by P2m . The hydrodynamic pressure is skew-symmetric in µ. the following energy balance relation is found: M0 P0 ¡ N0 Q0 = Coupling of Sway into Roll ¼2 2 In the case of a sway oscillation generally a roll moment is produced. With: ½gb2 ´ a 0 0 MR = (YR cos(!t + " ¡ ") ¡ XR sin(!t + " ¡ ")) ¼ and: ¡´ a ¡´ a sin " = P0 cos " = Q0 xa ¼» b xa ¼» b .

the two-dimensional hydrodynamic mass and damping coupling coe¢cients of sway into roll in Tasai’s axes system are given by: 0 M42 = ¡½b3 YR Q0 + XR P0 0 ¢ 2 2 P0 + Q2 0 and 0 N42 = ¡½b3 YR P0 ¡ XR Q0 0 ¢ ¢! 2 2 P0 + Q2 0 In the ship motions O(x.4. y.1. z) coordinate system these two coupling coe¢cients will change sign. found before. . THEORY OF TASAI 67 the two-dimensional hydrodynamic roll moment can be resolved into components in phase and out phase with the lateral displacement of the cylinder: 0 MR ¡½gb2 ´2 = 2 0 a ((YR Q0 + XR P0 ) cos(!t + ") + (YR P0 ¡ XR Q0 ) sin(!t + ")) ¼ » b xa This hydrodynamic roll moment can also be written as: 0 0 0 MR = ¡M42 x ¡ N42 x Ä _ 0 0 2 = M42 ! xa cos(!t + ") + N42 !xa sin(!t + ") in which: 0 M22 0 N22 = 2-D hydrodynamic mass coupling coe¢cient of sway into roll = 2-D hydrodynamic damping coupling coe¢cient of sway into roll When using also the amplitude ratio of the radiated waves and the forced sway oscillation.

The amplitudes of these waves decrease strongly with the distance to the cylinder. Figure 4. is based here on work published by [Tasai. Two kinds of waves will be produced: ² A standing wave system.3. Suppose an in…nite long cylinder in the surface of a ‡uid. 1973]. . Starting points for the derivation these coe¢cients here are the velocity potentials and the conjugate stream functions of the ‡uid as they have been derived by Tasai and also by [Jong. denoted here by subscript A. Respectively. Because the cylinder is supposed to be in…nitely long.3: Axes System for Sway Oscillations The cylinder is forced to carry out a simple harmonic roll motion about the origin O with a frequency of oscillation ! and a small amplitude of displacement ¯ a : ¯ = ¯ a cos(!t + °) in which ° is a phase angle. the angular velocity and acceleration of the cylinder are: _ ¯ = ¡!¯ a sin(!t + °) and Ä ¯ = ¡! 2 ¯ a cos(!t + °) This forced angular oscillation of the cylinder causes a surface disturbance of the ‡uid. of which a cross section is given in …gure 4.1. the generated waves will be twodimensional. These waves travel away from the cylinder and a stationary state is rapidly attained. 2-D POTENTIAL COEFFICIENTS 4.68 CHAPTER 4.3 Roll Motions The determination of the hydrodynamic coe¢cients of a rolling cross section of a ship in deep and still water at zero forward speed. 1961] for the Lewis method.

The linearized free surface condition in deep water is expressed as follows: !2 @© ©+ =0 g @y for: jxj ¸ Bs 2 and y = 0 In consequence of the conformal mapping.4. this velocity potential has the following relation: ©(¡x.. y) 3. denoted here by subscript B.1. provided that this amplitude is su¢ciently small compared with the radius of the cylinder and the wave length is not much smaller than the diameter of the cylinder. THEORY OF TASAI 69 ² A regular progressive wave system. The velocity potential must satisfy to the equation of Laplace: r2 © = @2© @2© + 2 =0 @x2 @y 2. At a distance of a few wave lengths from the cylinder. the waves on each side can be described by a single regular wave train. Using the stream function ª this boundary condition on the surface of the cylinder (® = 0) reduces to: µ ¶ 2 ¡@ª0 (µ) @ x2 + y0 0 _ =¯ @s @s 2 Integration results into the following requirement for the stream function on the surface of the cylinder: _ ¯ 2 ª0 (µ) = ¡ (x2 + y0 ) + C(t) 0 2 . this last equation results into: N ª @© »b X © ¼ © = 0 for: ® ¸ 0 and µ = § (2n ¡ 1)a2n¡1 e¡(2n¡1)® § ¾ a n=0 @µ 2 in which: !2 »b = Ms ¾a g ! 2 b0 or: » b = (non-dimensional frequency squared) 2g From the de…nition of the velocity potential follows the boundary condition on the surface S of the cylinder for ® = 0: @©0 (µ) _ @r0 = r0 ¯ @n @s in which n is the outward normal of the cylinder surface S and r0 is the radius from the origin to the surface of the cylinder. These waves dissipate energy. y) = ¡©(+x. The two-dimensional velocity potential of the ‡uid has to ful…l the following three requirements: 1. Because the roll motion of the ‡uid is not symmetrical about the y-axis. The wave amplitude at in…nity ´ a is proportional to the amplitude of oscillation of the cylinder ¯ a .

as they are given by [Tasai. 1973]. 2-D POTENTIAL COEFFICIENTS in which C(t) is a function of the time only. µ) sin(!t) ¼! m=1 m=1 in which: à A2m (®. µ) cos(!t) + Q2m à A2m (®. µ) cos(!t) + Q2m ÁA2m (®. 1961] and [Jong. µ) = e¡(2m+1)® sin ((2m + 1)µ) ¾ N ½ »b X ¡(2m+2n)® n 2n ¡ 1 a2n¡1 e sin ((2m + 2n)µ) ¡ (¡1) ¾ a n=0 2m + 2n The set of conjugate stream functions is expressed as: Ã1 ! 1 ª X© ª g´a X © ªA = P2m à A2m (®. The following set of velocity potentials. The vertical oscillation at the intersection of the surface of the cylinder and the waterline is de…ned by: b0  = ¯ = Âa sin(!t + °) 2 When de…ning: ¹(µ) = 2 x2 + y0 0 ¡ b 0 ¢2 = à !2 N 1 X n ¡ f(¡1) a2n¡1 sin ((2n ¡ 1)µ)g ¾ a n=0 à !2 N 1 X + + f(¡1)n a2n¡1 cos ((2n ¡ 1)µ)g ¾ a n=0 b0 ¹(µ) + C(t) 4 2 the stream function on the surface of the cylinder is given by: ª0 (µ) =  _ For the standing wave system a velocity potential and a stream function satisfying to the equation of Laplace. the non-symmetrical motion of the ‡uid and the free surface condition has to be found.70 CHAPTER 4. . µ) sin(!t) ¼! m=1 m=1 in which: ÁA2m (®. µ) = ¡e¡(2m+1)® cos ((2m + 1)µ) ¾ N ½ »b X ¡(2m+2n)® n 2n ¡ 1 a2n¡1 e + (¡1) sin ((2m + 2n)µ) ¾ a n=0 2m + 2n These sets of functions tend to zero as ® tends to in…nity. ful…l these requirements: à 1 ! 1 ª X© ª g´ a X © ©A = P2m ÁA2m (®.

as given by [Porter. y) sin(!t) ¼! in which: ÁBc ¢ j = ¡¼e¡ºy sin(ºjxj) ÁBs ¢ j = ¼e¡ºy cos(ºx) ¡ while: j = +1 for x > 0 j = ¡1 for x < 0 !2 (wave number for deep water) º = g Changing the parameters delivers: ¢ g´ ¡ ©B = a ÁBc (®. [Tasai. y) cos(!t) + ÁBs (x. Changing the parameters delivers: ¢ g´ ¡ ªB = a à Bc (®. THEORY OF TASAI 71 In these expressions the magnitudes of the P2m and the Q2m series follow from the boundary conditions.1. For this. as will be explained further on.4. 1973] gave the velocity potential of the progressive wave system as: ¢ g´ ¡ ©B = a ÁBc (x. 1961] and [Jong. a function describing a two-dimensional horizontal doublet at the origin O is chosen. the convergence is very slowly. representing such a train of waves at in…nity. y) cos(!t) + à Bs (x. Another requirement is a diverging wave train for ® goes to in…nity. 1960]. Power series expansions. µ) sin(!t) ¼! . µ) sin(!t) ¼! Z1 0 º cos(ky) + k sin(ky) ¡kjxj jxj e dk + 2 + º2 2 + y2) k º(x The conjugate stream function is given by: ¢ g´ ¡ ªB = a à Bc (x. can be used instead of these last integrals over k: The summation in these expansions converge much faster than the numeric integration procedure. µ) cos(!t) + à Bs (®. y) sin(!t) ¼! in which: à Bc = ¼e¡ºy cos(ºx) à Bs = ¼e¡ºy sin(ºjxj) + Z1 0 y º sin(ky) ¡ k cos(ky) ¡kjxj e dk ¡ 2 + º2 2 + y2) k º(x When calculating the integrals in the expressions for à Bs and ÁBs numerically. µ) cos(!t) + ÁBs (®. Therefore it is necessary to add a stream function. satisfying the equation of Laplace. This has been shown before for the sway case. the non-symmetrical motion of the ‡uid and the free surface condition.

so at the intersection of the surface of the cylinder with the free surface of the ‡uid where ¹(µ) = 1. µ) = ¼! à m=1 ! ! 1 X© ª + à Bs (®. µ) and ªBs (®. µ) cos(!t) m=1 When putting ® = 0. µ) sin(!t) m=1 Ãà ! 1 X© ª à Bc (®. µ) at the surface of the cylinder.72 CHAPTER 4. µ) = cos(!t) ÁBc (®. we obtain the constant C ¤ (t): C (t) = ¤ 1 X© ª à B0c (¼=2) + P2m à A02m (¼=2) cos(!t) m=1 © Q2m à A02m (µ) ! ª sin(!t) = ¡Â _ ¼!b0 ¹(µ) + C ¤ (t) 4g´ a . µ) + Q2m ÁA2m (®. µ) ¼! m=1 à ! ! 1 X© ª + ÁBs (®. 2-D POTENTIAL COEFFICIENTS The total velocity potential and stream function to describe the waves generated by a swaying cylinder are: © = ©A + ©B ª = ªA + ªB So the velocity potential and the conjugate stream function are expressed by: ! Ãà 1 X© ª g´ a ©(®. So for each µ-value. the following equation has been obtained: à ! 1 X© ª à B0c (µ) + P2m à A02m (µ) cos(!t)+ à à B0s (µ) + m=1 m=1 1 X When putting µ = ¼=2. µ) + Q2m à A2m (®. µ) + P2m ÁA2m (®. µ) + P2m à A2m (®. the stream function is equal to the expression found before from the boundary condition on the surface of the cylinder: Ãà ! 1 X© ª g´ a ª0 (µ) = cos(!t) à B0c (µ) + P2m à A02m (µ) ¼! m=1 à ! ! 1 X© ª + ÃB0s (µ) + Q2m à A02m (µ) sin(!t) m=1 b0 = ¡Â ¹(µ) + C(t) _ 4 in which: à A02m (µ) = ¡ cos ((2m + 1)µ) ¾ N ½ »b X n 2n ¡ 1 + a2n¡1 cos ((2m + 2n)µ) (¡1) ¾ a n=0 2m + 2n In this expression ªB0c (µ) and ªB0s (µ) are the values of ªBc (®. µ) sin(!t) g´ a ª(®. so for ® = 0.

THEORY OF TASAI ! 1 X© ª + à B0s (¼=2) + sin(!t) Q2m à A02m (¼=2) m=1 73 à + _ in which: ¼!b0 4g´ a A substitution of C ¤ (t) in the equation for each µ-value.4. So: à B0c (µ) ¡ à B0c (¼=2) = (¹(µ) ¡ 1) P0 + à B0s (µ) ¡ à B0s (¼=2) = (¹(µ) ¡ 1) Q0 + in which: f2m (µ) = ¡Ã A0c (µ) + à A0c (¼=2) These equations can also be written as: à B0c (µ) ¡ à B0c (¼=2) = à B0s (µ) ¡ à B0s (¼=2) = 1 X 1 X m=1 1 X ff2m (µ)P2m g ff2m (µ)Q2m g m=1 m=0 1 X m=0 ff2m (µ)P2m g ff2m (µ)Q2m g .1. results for any µ-value less than ¼=2 into the following equation: à ! 1 X© ¡ ¢ª P2m à A02m (µ) ¡ à A02m (¼=2) à B0c (µ) ¡ à B0c (¼=2) + cos(!t) + à B0s (µ) ¡ à B0s (¼=2) + à m=1 m=1 1 X N X 2n ¡ 1 »b m ªA02m (¼=2) = (¡1) a2n¡1 ¾a 2m + 2n n=0 ! © ¡ ¢ª Q2m à A02m (µ) ¡ à A02m (¼=2) = sin(!t) ¡Â _ ¼!b0 (¹(µ) ¡ 1) 4g´ a The right hand side of this equation can be written as: µ ¶ ¼!b0 ¼Âa ¡Â _ (¹(µ) ¡ 1) = (¹(µ) ¡ 1) ¡ » sin(!t + °) 4g´ a 2´ a b = (¹(µ) ¡ 1) (P0 cos(!t) + Q0 sin(!t)) in which: P0 = ¼Âa » sin ° 2´ a b and Q0 = ¼Âa » cos ° 2´ a b This delivers for any µ-value less than ¼=2 two sets of equations with the unknown parameters P2m and Q2m .

For practical reasons the maximum value of m is limited to M. µ) and ÁBs (®. are required to solve these equations. Another favorable method is to multiply both sides of the equations with ¢µ. so for ® = 0. Then the summations over µ can be replaced by integrations. less than ¼=2. the velocity potential on the surface of the cylinder (®=0) is known too: ÃÃ ! M X© ª g´ a ©0 (µ) = P2m ÁA02m (µ) ÁB0c (µ) + cos(!t) ¼! m=1 ! ! Ã M X© ª Q2m ÁA02m (µ) sin(!t) + ÁB0s (µ) + m=1 in which: ÁA02m (µ) = sin ((2m + 1)µ) ¾ N ½ »b X n 2n ¡ 1 a2n¡1 sin ((2m + 2n)µ) ¡ (¡1) ¾ a n=0 2m + 2n In this expression ÁB0c (µ) and ÁB0s (µ) are the values of ÁBc (®. Note that at least M +1 values of µ. two sets of M + 1 equations have been obtained. :::M m=0 Now the P2m and Q2m series can be solved with a numerical method. Each µ-value less than ¼=2 will deliver an equation for the P2m and Q2m series. m=0 0 0 8 9 ¼=2 ¼=2 Z Z M = X< ¡ ¢ Q2m f2m (µ)f2n (µ)dµ = Ã B0s (µ) ¡ Ã B0s (¼=2) f2n(µ)dµ : . µ) at the surface of the cylinder. one set for P2m and one set for Q2m : 8 9 ¼=2 ¼=2 Z Z M = X< ¡ ¢ P2m f2m (µ)f2n (µ)dµ = Ã B0c (µ) ¡ Ã B0c (¼=2) f2n (µ)dµ : .74 in which: for m = 0: for m > 0: CHAPTER 4. 2-D POTENTIAL COEFFICIENTS f0 (µ) = ¹(µ) ¡ 1 f2m (µ) = ¡Ã A02m (µ) + Ã A02m (¼=2) The series in these two sets of equations converges uniformly with an increasing value of m. . :::M n = 0. The best …t values of P2m and Q2m are supposed to be those found by means of a least squares method. Then P0 and Q0 are known now and from the de…nition of these coe¢cients follows the amplitude ratio of the radiated waves and the forced sway oscillation: ¼» ´a = p b 2 Âa 2 P0 + Q2 0 With the solved P2m and Q2m values. 0 0 n = 0. Herewith.

THEORY OF TASAI 75 Now the hydrodynamic pressure on the surface of the cylinder can be obtained from the linearized equation of Bernoulli: p(µ) = ¡½ @©0 (µ) @t Ãà ! M X© ª ¡½g´ a = ÁB0s (µ) + Q2m ÁA02m (µ) cos(!t) ¼ m=1 à ! ! M X© ª ¡ ÁB0c (µ) + P2m ÁA02m (µ) sin(!t) m=1 It is obvious that this pressure is skew-symmetric in µ. by replacing there ÁB0s (µ) by ÁB0c (µ) and Q2m by P2m .4.1. . Roll Coe¢cients The two-dimensional hydrodynamic moment acting on the cylinder in the clockwise direction can be found by integrating the roll component of the hydrodynamic pressure on the surface S of the cylinder: 0 MR ¼=2 µ ¶ Z +dx0 ¡dy0 + y0 ds = (p(+µ) ¡ p(¡µ)) ¡x0 ds ds 0 ¼=2 µ ¶ Z dx0 dy0 = ¡2 p(µ) x0 + y0 dµ dµ dµ 0 With this the two-dimensional hydrodynamic moment due to roll oscillations can be written as follows: ½gb2 ´ a 0 0 MR = (YR cos(!t) ¡ XR sin(!t)) ¼ in which: YR 1 = 2¾ 2 a ¼=2 Z N N XX© ª (¡1)n+i (2i ¡ 1) ¢ a2n¡1 a2i¡1 sin ((2n ¡ 2i)µ) dµ ÁB0s (µ) 0 M X 1 + 2 2¾ a ¡ ¼» b 8¾ 3 a à ¾ (¡2m + 2n ¡ 2i ¡ 1)(2i ¡ 1) ¢ a2n¡1 a2i¡1 a¡2m+2n¡2i¡1 2n ¡ 2i n=m i=0 ¾!) N¡m X ½ X N (¡2m ¡ 2n + 2i ¡ 1)(2i ¡ 1) ¢ a2n¡1 a2i¡1 a¡2m¡2n+2i¡1 + 2n ¡ 2i n=0 i=m+n m=1 N n¡m ½ XX m=1 N X ( n=0 i=0 (¡1)m Q2m N N XX½ n=0 i=0 (2i ¡ 1)(2n ¡ 2i) ¢ a2n¡1 a2i¡1 (2m + 1)2 ¡ (2n ¡ 2i)2 ¾) f(¡1)m Q2m ¢ and XR as obtained from this expression for YR .

the two-dimensional hydrodynamic mass and damping coe¢cients of roll in Tasai’s axes system are given by: ½b4 YR Q0 + XR P0 ½b4 YR P0 ¡ XR Q0 0 0 0 and N44 = 0 ¢ ¢! M44 = ¢ 2 2 2 8 P0 + Q0 8 P0 + Q2 0 The signs of these two coe¢cients are proper in both. found before. With: ½gb2 ´ a 0 0 MR = (YR cos(!t + ° ¡ °) ¡ XR sin(!t + ° ¡ °)) ¼ and: 2´ a 2´ a sin ° = P0 cos ° = Q0 Âa ¼» b Âa ¼» b the two-dimensional hydrodynamic roll moment can be resolved into components in phase and out phase with the angular displacement of the cylinder: 2½gb2 ´ 2 0 a ((YR Q0 + XR P0 ) cos(!t + °) + (YR P0 ¡ XR Q0 ) sin(!t + °)) ¼ 2 » b Âa This hydrodynamic roll moment can also be written as: Ä _ M 0 = ¡M 0 ¯ ¡ N 0 ¯ 0 MR = R = 44 0 M44 ! 2 ¯ a 44 0 cos(!t + °) + N44 !¯ a sin(!t + °) in which: 0 M44 0 N44 = 2-D hydrodynamic mass moment of inertia coe¢cient of roll = 2-D hydrodynamic damping coe¢cient of roll When using also the amplitude ratio of the radiated waves and the forced roll oscillation. y. So: T osc Z 1 ½g´ 2 c a 0 _ _ N44 ¯ ¢ ¯dt = Tosc 2 0 in which Tosc is the period of oscillation. The energy delivered by the exciting moments should be equal to the energy radiated by the waves. With the relation for the wave speed c = g=!.76 CHAPTER 4. follows the relation between the twodimensional roll damping coe¢cient and the amplitude ratio of the radiated waves and the forced roll oscillation: µ ¶2 ½g 2 ´ a 0 N44 = 3 ! ¯a With this amplitude ratio the two-dimensional hydrodynamic damping coe¢cient of roll is also given by: ½¼2 b4 1 0 0 N44 = ¢! ¢ 2 64 P0 + Q2 0 0 When comparing this expression for N44 with the expression found before. the following energy balance relation is found: ¼2 YR P0 ¡ XR Q0 = 8 . 2-D POTENTIAL COEFFICIENTS These expressions are similar to those found before for the hydrodynamic roll moment due to sway oscillations. the axes system of Tasai and the ship motions O(x. z) coordinate system.

These expressions are similar to those found before for the hydrodynamic lateral force due to sway oscillations.4. can be found by integrating the angular component of the hydrodynamic pressure on the surface S of the cylinder: 0 Fx ¼=2 Z ¡dy0 = ¡ (p(+µ) ¡ p(¡µ)) ds ds 0 ¼=2 Z dy0 = 2 p(µ) dµ dµ 0 With this the two-dimensional hydrodynamic angular force due to roll oscillations can be written as follows: ¡½gb0 ´ a 0 Fx = (M0 cos(!t) ¡ N0 sin(!t)) ¼ in which: M0 1 = ¡ ¾a + ¼=2 Z N X ÁB0s (µ) f(¡1)n (2n ¡ 1)a2n¡1 sin ((2n ¡ 1)µ)g dµ 0 N¡1 X n=0 ¼ 4¾ a » + b ¾2 a m=1 m=1 ( M X f(¡1)m Q2m (2m + 1)a2m+1 g (¡1)m Q2m N N XX½ n=0 i=0 (2n ¡ 1)(2i ¡ 1) a2n¡1 a2i¡1 (2m + 2i)2 ¡ (2n ¡ 1)2 ¾) and N0 as obtained from this expression for M0 . With: ¡½gb0 ´ a 0 (M0 cos(!t + ° ¡ °) ¡ N0 sin(!t + ° ¡ °)) Fx = ¼ and: 2´ a 2´ a P0 cos ° = Q0 sin ° = Âa ¼» b Âa ¼» b the two-dimensional hydrodynamic angular force can be resolved into components in phase and out phase with the angular displacement of the cylinder: 0 Fx = ¡2½gb0 ´ 2 a ((M0 Q0 + N0 P0 ) cos(!t + °) + (M0 P0 ¡ N0 Q0 ) sin(!t + °)) ¼ 2 » b Âa 0 0 Ä 0 _ Fx = ¡M24 ¯ ¡ N24 ¯ 0 0 = M24 ! 2 ¯ a cos(!t + °) + N24 !¯ a sin(!t + °) This hydrodynamic angular force can also be written as: in which: . THEORY OF TASAI Coupling of Roll into Sway 77 The two-dimensional hydrodynamic lateral force. For the determination of M0 and N0 it is required: M = N . acting on the cylinder in the direction of the x-axis. by replacing there ÁB0s (µ) by ÁB0c (µ) and Q2m by P2m .1.

. y.78 0 M24 0 N24 CHAPTER 4. found before. 2-D POTENTIAL COEFFICIENTS = 2-D hydrodynamic mass coupling coe¢cient of roll into sway = 2-D hydrodynamic damping coupling coe¢cient of roll into sway When using also the amplitude ratio of the radiated waves and the forced roll oscillation. the two-dimensional hydrodynamic mass and damping coupling coe¢cients of roll into sway are given by: 0 M24 = ¡½b3 M0 Q0 + N0 P0 0 ¢ 2 8 P0 + Q2 0 and 0 N24 = ¡½b3 M0 P0 ¡ N0 Q0 0 ¢ ¢! 2 8 P0 + Q2 0 In the ship motions O(x. z) coordinate system these two coupling coe¢cients will change sign.

The two-dimensional hydrodynamic mass moment of inertia coe¢cient of roll of a Lewis cross section is given by [Grim. Zero-Frequency Coe¢cients The two-dimensional hydrodynamic mass coe¢cient of sway of a Lewis cross section is given by [Tasai. 1961]: 0 M22 (! ½¼ = 0) = 2 µ Ds 1 ¡ a1 + a3 ¶2 ¡ (1 ¡ a1 )2 + 3a2 3 ¢ The two-dimensional hydrodynamic mass coupling coe¢cient of sway into roll of a Lewis cross section is given by [Grim. 1955]: 0 M44 (! µ ¶4 16½ Bs ¢ = 0) = ¢ ¼ 2(1 + a1 + a3 ) µ ¶ 8 16 2 2 2 a1 (1 + a3 ) + a1 a3 (1 + a3 ) + a3 9 9 The two-dimensional hydrodynamic mass coupling coe¢cient of roll into sway of a Lewis cross section is given by [Grim.4.4 Low and High Frequencies The velocity potentials for very small and very large frequencies are derived and discussed in the next subsections. 1955]: 0 0 M24 (! = 0) = ¡M44 (! = 0) ¢ ¼ 1 ¡ a1 + a3 ¢ ¢ 6 Ds a1 (1 ¡ a1 )(1 + a3 ) + 3 a1 a3 (1 + a3 ) + 4 a3 (1 ¡ a1 ) ¡ 5 5 a2 (1 + a3 )2 + 8 a1 a3 (1 + a3 ) + 16 a2 1 9 9 3 12 2 a 7 3 0 In Tasai’s axes system M24 will change sign. All potential damping values for zero frequency will be zero: 0 N22 (! 0 N42 (! 0 N33 (! 0 N44 (! 0 N24 (! = = = = = 0) = 0 0) = 0 0) = 0 0) = 0 0) = 0 .1. 1955]: 16 Ds 0 0 ¢ M42 (! = 0) = ¡M22 (! = 0) ¢ ¢ 3¼ 1 ¡ a1 + a3 ¡ ¢ a1 1 ¡ a1 + 4 a3 ¡ a1 a3 + 3 a2 + 4 a3 ¡ 3 5 5 5 (1 ¡ a1 )2 + 3a2 3 12 2 a 7 3 0 In Tasai’s axes system M42 will change sign. THEORY OF TASAI 79 4.1.

80 In…nite Frequency Coe¢cients CHAPTER 4. 2-D POTENTIAL COEFFICIENTS The two-dimensional hydrodynamic mass coe¢cient of sway of a Lewis cross section is given by Landweber and de Macagno: µ ¶2 µ ¶ Ds 2½ 16 2 0 2 M22 (! = 1) = (1 ¡ a1 + a3 ) + a3 ¼ 1 ¡ a1 + a3 3 The two-dimensional hydrodynamic mass coe¢cient of heave of a Lewis cross section is given by [Tasai. 1959]: 0 M33 (! ½¼ = 1) = 2 µ Bs 2(1 + a1 + a3 ) ¶2 The two-dimensional hydrodynamic mass moment of inertia coe¢cient of roll of a Lewis cross section is given by [Kumai. 1959]: 0 M44 (! ¡ (1 + a1 )2 + 3a2 3 ¢ = 1) = ½¼ µ Bs 2(1 + a1 + a3 ) ¶4 ¡ 2 ¢ a1 (1 + a3 )2 + 2a2 3 All potential damping values for in…nite frequency will be zero: 0 N22 (! 0 N42 (! 0 N33 (! 0 N44 (! 0 N24 (! = 1) = 1) = 1) = 1) = 1) = = = = = 0 0 0 0 0 .

swaying and rolling cross section of a ship in shallow at zero forward speed is based on work published by [Keil.1 Notation His notations have been maintained here as far as possible and references are given here to the formula numbers in his report.2. a Aindex ¹ Aindex b B c Cindex Eindex Findex g G h H Hindices HT I" kx ky m" Mindices Nindex p p r = y2 + z2 t T ¹ U ¹ V Ax x. 1974].2. the determination of the hydrodynamic coe¢cients of a heaving. y. This method is based on a Lewis conformal mapping of the ships’ cross section to the unit circle 4. THEORY OF KEIL 81 4.2 Theory of Keil In this section.draft ratio hydrodynamic moment of inertia wave number in x-direction wave number in y-direction hydrodynamic mass hydrodynamic moment hydrodynamic damping coe¢cient pressure polar coordinates time or integer value draft velocity amplitude of horizontal oscillation velocity amplitude of vertical oscillation cross sectional area space-bound coordinates transfer functions Euler constant (= 0. z Yindices ° " ¹ ³ µ = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = Lewis coe¢cient source strength amplitude ratio Lewis coe¢cient breadth of body wave velocity non-dimensional force or moment non-dimensional exciting force or moment hydrodynamic force acceleration of gravity function (real part) water depth function (imaginary part) …ctive moment arm water depth .57722) phase lag wave amplitude polar coordinate or pitch amplitude .4.

Velocities are positive if they are directed in the positive coordinate direction: @© = vy @y @© = vz @z The value of the stream function increases when .2.82 ¸ ¹ º = ! 2 =g º 0 = 2¼=¸ ½ ' 'indices ©indices à indices ªindices ! = = = = = = = = = = = CHAPTER 4.going in the positive direction . 2-D POTENTIAL COEFFICIENTS wave length wave direction wave number at deep water wave number density of water roll angle part of potential time-dependent potential part of stream function time-dependent stream function circular frequency of oscillation In here.2 Basic Assumptions The following …gure shows the 2-D coordinate system as used by Keil and maintained here.e.. this yields an assumption of small amplitude motions. the indices . The problem will be linearized.that he used .5-a.are: E H j n Q r R V W = = = = = = = = = related to excitation horizontal or related to horizontal motions imaginary part ordering of potential parts related to transverse motions real part related to roll motions vertical or related to vertical motions related to waves 4. The wave number º 0 = 2¼=¸. The space-bound axes system of the sectional contour is given in …gure 4. Physically. follows from: º= 2¼ 2¼h !2 = tanh = º 0 tanh [º 0 h] g ¸ ¸ The ‡uid is supposed to be incompressible and inviscid. The potentials of the incoming waves are described in Appendix 1.the ‡ow goes in the negative y-direction: @© @ª =+ @y @z ª1 < ª2 ! . The ‡ow caused by the oscillating body in the surface of this ‡uid can be described by a potential ‡ow. i. contributions of second and higher order in the de…nition of the boundary conditions will be ignored.

4: Axes System. as Used by Keil ª3 > ª4 ! @© @ª =¡ @z @y @© @ª =¡ @n @s @© @ª =+ @s @n For the imaginary parts.4. THEORY OF KEIL 83 Figure 4. i and j have been used: i for geometrical variables (potential and stream function) and j for functions of time.5: De…nition of Sectional Contour .2. Figure 4.

so the vertical ‡uid velocity at z = h is zero: @© =0 @z for: z = h (Keil-3) 4.84 CHAPTER 4. travelling away from the cylinder. 2-D POTENTIAL COEFFICIENTS 4. 5. which ful…ls the Sommerfeld Radiation Condition: ½p µ ¶¾ @ (Keil-4) lim jyj Re (©) ¡ º 0 Im (©) =0 y!1 @y In here. This yields that the boundary conditions on the surface of the body are given by: · ¸ @© = [vn ]body @n body · ¸ @© dy @© dz = ¢ ¡ ¢ @z ds @y ds body · ¸ dª (Keil-5) = ¡ ds body Two cases have to be distinguished: . The sea bottom is impervious.3 Vertical Motions Boundary Conditions The two-dimensional velocity potential of the ‡uid has to ful…l the following requirements: 1. see …gure 4. The ‡uid is incompressible and the velocity potential must satisfy to the Continuity Condition and the Equation of Laplace: r2 © = @2© @2© + 2 =0 @y 2 @z (Keil-1) 2. thus at the surface of the body is the ‡uid velocity equal to the body velocity.5-b. The harmonic oscillating cylinder produces a regular progressive wave system. The oscillating cylinder is impervious too. º 0 = 2¼=¸ is the wave number of the radiated wave.2. The linearized free surface condition follows from the condition that the pressure at the free surface is not time-depending: µ ¶ @p @© @ 2 © B for: jyj ¸ and z = 0 (Keil-2/1) =½ g ¡ 2 =0 @t @z @t 2 from which follows: @© !2 ©+ =0 g @z or: º© + @© =0 @z for: jyj ¸ B 2 B 2 and z = 0 for: jyj ¸ and z = 0 (Keil-2/2) 3.

z. The hydromechanical loads. t)]body · ¸ @©W @©W = dy ¡ dz dz dy body (Keil-6/2) The stream function of an incoming wave . t)]body ¹ ³! jwt = ¢ e ¢ [(sinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]) sin(º 0 y)]body º In case of another wave direction. so wave direction ¹ = 900 : (Keil-6a) [ªS (y. so ¹ = 900 . only the in y-symmetric part of the potential and stream functions have to be considered. t)]body = ¡ [ªW (y. t) = ¡V ¢ ejwt ¢ ybody + C (Keil-5a/2) (Keil-5a/1) b.is given in Appendix 1 of this section by: ªW = j ¹ ³! j(wt+º 0 y) ¢e ¢ (sinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]) º (Keil-A1) Because only vertical forces have to be determined.which travels in the negative y-direction. this is the amount of ‡uid which has to come . The wave loads. z.e. this problem becomes three-dimensional and a stream function can not be written. z.2. z. which have to be obtained for the vertically oscillating cylinder in still water with a vertical velocity equal to: ¹ V = V ¢ ejwt The boundary condition on the surface of the body becomes: · ¸ · ¸ dª ¹ dy ¡ =V ¢ ¢ ejwt ds body ds body or: ¹ ªbody (y.. t)]body = ¡ [dªW (y. boundary condition (Keil-6a) provides ~ us a ”quasi stream function” ªs . THEORY OF KEIL 85 a.4. However. i. From this follows the boundary condition on the surface of the body for beam waves. z. which have to be obtained for the restrained cylinder in regular waves from the incoming undisturbed wave potential ©W and the di¤raction potential ©S : · ¸ @©W @©S + = 0 @n @n body · ¸ @©W dy @©W dz @©S dy @©S dz ¢ ¡ ¢ + ¢ ¡ ¢ = @z ds @y ds @z ds @y ds body · ¸ dªW dªS (Keil-6/1) = ¡ + ds ds body or: [dªS (y.

z1 . it is supposed to act . z = z0 . y. 2-D POTENTIAL COEFFICIENTS out of the body per unit of length so that totally no ‡uid of the incoming wave enters into the body. z = z0 ) = 0 ¹ in which z0 is the z-coordinate of the ‡uid surface. y. t) dy ¡ dz =4 = @z @y body 0 z0 body = ¹ ³! jwt ¢ e ¢ cos(º 0 x cos ¹) ¢ º fsin ¹ sin(º 0 y1 sin ¹) (sinh [º 0 z1 ] ¡ tanh [º 0 h] cosh [º 0 z1 ]) Zy1 2 +º 0 (1 ¡ sin ¹) cos(º 0 y sin ¹) (sinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]) dyg 0 Potentials 3-D Radiation Potential Suppose a three-dimensional oscillating cylindrical body in previously still water. z=z0 ) dydx ¹ ¡L=2 ¡¢y=2 = ¹0 P (x)dx < 1 (Keil-7a) ¡L=2 Boundary condition (Keil-7) can be ful…lled by a pressure amplitude p(x. From equation (Keil-7a) follows that the pressure amplitude p(x) can be integrated. so: p(x.86 CHAPTER 4. The resulting force P in the z-direction becomes: +L=2 +¢y=2 Z Z +L=2 Z (Keil-7) ¹ P = p(x. y1 . z=z0 ) which ¹ is found by a superposition of an in…nite number of harmonic pressures. t) = ¡j ¢ p(x.over the full length L of the body . This function can be used as an approximation of the problem: 3 2 y1 Z Zz1 h i @ªW @ªW 5 ~ (Keil-6b) ªS (x1 . To …nd the potential of the resulting ‡uid motions.only over a small distance ¢y=2 to both sides of y = 0. y. y. so a Fourier series ¹ expansion follows from: 1 p(x) = ¹ ¼ +1 Z1 Z p(») cos [kx(x ¡ »)] d» dkx ¹ 0 ¡1 . This pressure is not supposed to act over the full breadth of the body. y. The unknown amplitude p of ¹ this pressure has to follow from the boundary conditions. z=z0 )ej!t ¹ or : for jyj > ¢y=2 and jxj > L=2 p(x. this body will be replaced by an oscillating pressure p at the free surface.

2.with the Bernoulli Equation obtained . This expression (Keil-8) for the radiation potential ful…ls the Equation of Laplace: @ 2© @ 2© @ 2© + 2 + 2 =0 @x2 @y @z Now. y. y. This means that ´ remains small too. THEORY OF KEIL 87 Because the pressure amplitude p depends on two variables. z=z0 ) = 2 ¹ p(». ky ) cos [kx (x ¡ »)] cos(ky y) ¢ ©0r (x. the Fourier series expansion ¹ has to be two-dimensional: +1 +1 Z1 Z1 Z Z 1 p(x. t) = e ¢ cosh 0 ¤ 2 2 kx + ky ¢ (z ¡ h) £p ¤ dkx dky d» 2 2 sinh kx + ky ¢ h 0 0 £p (Keil-8) in which the function C(kx . According to equation (Keil-7). z. ´) cos [ky (y ¡ ´)] cos [kx (x ¡ »)] d´ d» dky dkx ¹ 0 0 ¡L=2 ¡¢y=2 It is assumed that the value of ¢y is small.to the following initial de…nition of the radiation potential: Z1 Z1 Z1 j!t C(kx .as a start . z=z0 ) = 2 ¹ ¼ Z1 Z1 +L=2 +¢y=2 Z Z p(». the pressure at the free surface p1 can be obtained from an integration of the . y.derivative to the time of the pressure: Z1 Z1 Z1 @p1 = ej!t ½ C(kx . the pressure amplitude p disappears for jyj > ¢y=2 and ¹ jxj > L=2. ky ) is still unknown. ´) cos [ky (y ¡ ´)] cos [kx (x ¡ »)] d´ d» dky dkx ¹ ¼ 0 0 ¡1 ¡1 in which kx is the wave number in the x-direction and ky is the wave number in the y-direction. ´) d´ cos [kx(x ¡ »)] d» cos(ky y) dky dkx ¹ 0 0 ¡L=2 ¡¢y=2 Z1 Z1 0 0 cos(ky y) ¡L=2 +L=2 Z ¹0 P (») cos [kx (x ¡ »)] d» dky dkx (Keil-7b) This pressure de…nition leads . thus one can safely suppose that: cos [ky (y ¡ ´)] t cos(ky y) which results in: 1 p(x. z=z0 ) = ¹ ¼2 1 = ¼2 Z1 Z1 +L=2 +¢y=2 Z Z p(». ky ) cos [kx (x ¡ »)] cos(ky y) ¢ @t 0 0 0 £p ¤ p 2 2 2 2 2 ! ¡ g kx + ky ¢ tanh kx + ky ¢ h £p ¤ (Keil-8a) ¢ dkx dky d» 2 2 tanh kx + ky ¢ h .4. so for this pressure expression remains: 1 p(x. y.

z=z0 ) = ¹ ¼2 cos(ky y) ¹0 P (») cos [kx (x ¡ »)] d» dky dkx Comparing the two integrands provides: Z1 0 = p1 (x. z=0) ¢ ej!t ¹ and its amplitude becomes: Z1 Z1 Z1 0 2 0 p1 (x. y. y. then combining ¹ ¹ equations (Keil-7b) and (Keil-8b) provides the unknown function C(kx . ky ): Z1 Z1 0 0 +L=2 Z £p ¤ 2 2 2 2 kx + ky ¢ tanh kx + ky ¢ h £p ¤ dkx dky d» 2 2 tanh kx + ky ¢ h 1 p(x. z=z0 . z=z0 ) = ¹ ½ ! ¢ C(kx. z=z0 ) ¹ Z1 Z1 Z1 ½g cos(ky y) C(kx . y. t) = ¡j ¢ p1 (x. 2-D POTENTIAL COEFFICIENTS The harmonic oscillating pressure is given by: p1 (x. ky ) cos [kx (x ¡ »)] d» = 2 2 2 º ¡ kx + ky ¢ tanh kx + ky ¢ h tanh ! ¢ ½g¼ 2 ¡L=2 +L=2 Z £p ¹0 P (») cos [kx (x ¡ »)] d» (Keil-9) When de…ning: . ky ) cos [kx (x ¡ »)] cos(ky y) ¢ (Keil-8b) ! ¡g p 0 If this pressure amplitude p1 is supposed to be equal to the amplitude p.88 CHAPTER 4. ky ) cos [kx (x ¡ »)] ¢ ! 1 = 2 ¼ or: Z1 0 ¹0 P (») cos [kx (x ¡ »)] d» ¡L=2 ¤ 2 2 kx + ky ¢ h £p ¤¢ p 2 C(kx . y. y. ky ) cos [kx (x ¡ »)] = ! 0 0 0 £p ¤ p 2 2 ¢ tanh 2 2 º ¡ kx + ky kx + ky ¢ h £p ¤ ¢ dkx dky d» 2 2 tanh kx + ky ¢ h p £p ¤ 2 2 2 2 kx + ky ¢ tanh kx + ky ¢ h £p ¤ d» = 2 2 tanh kx + ky ¢ h +L=2 Z ¡L=2 ½g º ¡ C(kx .

t) = e Z1 0 j!t A0 (») ¢ This potential ful…lls both the radiation conditions at in…nity and the boundary conditions at the free surface. 2: j!t A0 (K-11) 0 0 F (k) = cosh [k(z ¡ h)] º cosh [kh] ¡ k sinh [kh] (K-11a) When substituting for k the term u = k + il. use will be made here of the value of the potential given in equation (Keil-11) at a large distance from the body: Z1 ©0r (y.reduces to: ©0r (y. z.4. t) = e with: cosh [k(z ¡ h)] cos(ky) dk º cosh [kh] ¡ k sinh [kh] 0 9 81 Z1 <Z = 1 = ej!t A0 F (k)e+iky dk + F (k)e¡iky dk . z. THEORY OF KEIL 89 ¹ ! P (») A0 (») = ½g¼ 2 0 and substituting equation (Keil-9) in equation (Keil-8) provides the radiation potential: +L=2 Z ©0r (x.given in equation (Keil-10) . no waves are travelling in the x-direction.2. The distribution of A(») is constant over the full length of the body from » = ¡1 through » = +1 and the radiation potential . 2-D Radiation Potential In case of an oscillating two-dimensional body. z. a term has to be added to equation (Keil-11). y. So: . t) = ej!t A0 Z1 0 £p ¤ 2 2 cosh kx + ky ¢ (z ¡ h) cos(ky y) £p ¤ p £p ¤ dkx dky d» 2 2 2 2 2 2 º cosh kx + ky ¢ h ¡ kx + ky ¢ sinh kx + ky ¢ h ¡L=2 Z1 0 cos [kx (x ¡ »)] ¢ (Keil-10) cosh [k(z ¡ h)] cos(ky) dk º cosh [kh] ¡ k sinh [kh] (Keil-11) To ful…l also the Sommerfeld Radiation Condition in (Keil-4). For this. so kx = 0 and ky = º 0 . the …rst integral in equation (K-11) integrates for y > 0 over the closed line I-II-III-IV in the …rst quadrant of the complex domain and the second integral in equation (K-11) integrates for y > 0 over the closed line I-V-VI-VII in the fourth quadrant.

....dk + II Z .du Z1 0 = JI + JII + JIII + JIV = 0 F (k)e+iky dk = ¡ lim [JII + JIII + JIV ] R!1 The location of the singular point follows from the denominator in the expression (K-11a) for F (k): º cosh [º 0 h] ¡ º 0 sinh [º 0 h] = 0 Because limR!1 [JIII ] = 0 and JIV disappears too for a large y. For line I-II-III-IV: I F (u)e +iuy du = ZR 0 ..6: Treatment of Singularities 1..90 CHAPTER 4.du + IV Z . 2-D POTENTIAL COEFFICIENTS Figure 4..du + III Z .. the singular point itself delivers a contribution only: JII = ¡i¼ ¢ Residue(º 0 ) cosh [º 0 (z ¡ h)] = ¡i¼ e+iº 0 y ºh sinh [º 0 h] ¡ sinh [º 0 h] ¡ º 0 h cosh [º 0 h] cosh [º 0 h] cosh [º 0 (z ¡ h)] +iº 0 y e = +i¼ º 0 h + sinh [º 0 h] cosh [º 0 h] and the searched integral becomes for y ! 1: Z1 0 F (k)e+iky dk = ¡i¼ cosh [º 0 h] cosh [º 0 (z ¡ h)] +iº 0 y e º 0 h + sinh [º 0 h] cosh [º 0 h] .

4. z. t)g = 0 º 0 h + sinh [º 0 h] cosh [º 0 h] (Keil-11b) .. z. the singular point itself delivers a contribution only: JV = +i¼ ¢ Residue(º 0 ) cosh [º 0 h] cosh [º 0 (z ¡ h)] ¡iº 0 y e = ¡i¼ º 0 h + sinh [º 0 h] cosh [º 0 h] and the searched integral becomes for y ! 1: Z1 0 F (k)e¡iky dk = +i¼ cosh [º 0 h] cosh [º 0 (z ¡ h)] ¡iº 0 y e º 0 h + sinh [º 0 h] cosh [º 0 h] This provides for the potential in equation (Keil-11) for y ! 1: ©0r (y. For line I-V-VI-VII : I F (u)e ¡iuy 91 du = ZR 0 .du Z1 0 = JI + JV + JV I + JV II = 0 F (k)e¡iky dk = ¡ lim [JV + JV I + JV II ] R!1 Because limR!1 [JV I ] = 0 and JV II disappears too for a large y.2... the radiation potential becomes: (Keil-12) ©0r (y. t) = ej!t A0 ¼ cosh [º 0 h] cosh [º 0 (z ¡ h)] sin(º 0 y) º 0 h + sinh [º 0 h] cosh [º 0 h] (Keil-11a) The Sommerfeld Radiation Condition in equation (Keil-4) will be ful…lled when: ej!t A0 ¼º 0 or: Im f©0 (y. t) = ej!t A0 ¢ 81 <Z cosh [k(z ¡ h)] ¢ cos(ky) dk : º cosh [kh] ¡ k sinh [kh] 0 ¾ cosh [º 0 h] cosh [º 0 (z ¡ h)] cos(º 0 y) +j¼ º 0 h + sinh [º 0 h] cosh [º 0 h] cosh [º 0 h] cosh [º 0 (z ¡ h)] cos(º 0 y) ¡ º 0 Im f©0 (y. z. t)g = ©0j (y.du + VI Z . z. t) cosh [º 0 h] cosh [º 0 (z ¡ h)] = ej!t A0 ¼ cos(º 0 y) º 0 h + sinh [º 0 h] cosh [º 0 h] With equations (Keil-11a) and (Keil-11b)...du + V II Z . t) + j©0j (y. z.dk + V Z . z... THEORY OF KEIL 2..

z. equations (Keil-12) and (Keil-13) reduces to: 81 <Z : : 0 (Keil-12c) ©0r1 (y. the potential and stream functions can be written as: (Keil-12-b) ©0 = ©0r1 + ©0rad + j©0j = ej!t A0 ¢ 81 Z ¡kz Z1 ¡kh < e e º sinh [kz] ¡ k cosh [kz] ¢ cos(ky) dk + cos(ky) dk : º ¡k º¡k º cosh [kh] ¡ k sinh [kh] 0 0 ¾ cosh [º 0 h] cosh [º 0 (z ¡ h)] +j¼ cos(º 0 y) º 0 h + sinh [º 0 h] cosh [º 0 h] ª0 = ª0r1 + ª0rad + jª0j = ej!t A0 ¢ (Keil-13b) . z. z. 2-D POTENTIAL COEFFICIENTS ©0 (y ! 1. From the orthogonality condition: @ª @© =+ @y @z follows the stream function: (Keil-13) ª0r (y. º¡k ¡kz 9 = (Keil-12a) 9 = ¡kz e ¡ºz sin(ky) dk + j¼e sin(ºy) . t) + j©0j1 (y. t) + jª0j (y.92 From this follows for y ! 1: CHAPTER 4. z. z. z. t) + jª0j1 (y. º¡k (Keil-13a) Now. consisting of waves with an amplitude: cosh2 [º 0 h] ¹ = A0 !¼ ³ g º 0 h + sinh [º 0 h] cosh [º 0 h] travelling away from both sides of the cylinder. t) = j ¢ ej!t A0 ¼ cosh [º 0 h] cosh [º 0 (z ¡ h)] ¡jº 0 y e º 0 h + sinh [º 0 h] cosh [º 0 h] This means that equation (Keil-12) describes a ‡ow. t) = ¡ej!t A0 ¢ 81 <Z sinh [k(z ¡ h)] sin(ky) dk ¢ : º cosh [kh] ¡ k sinh [kh] 0 ¾ cosh [º 0 h] sinh [º 0 (z ¡ h)] +j¼ sin(º 0 y) º 0 h + sinh [º 0 h] cosh [º 0 h] For an in…nite water depth. t) = ej!t A0 ª0r1 (y. t) = ej!t A0 81 <Z 0 e cos(ky) dk + j¼e¡ºz cos(ºy) . z.

©0r1 is the potential at deep water and ©0rad is the additional potential due to the …nite water depth.also the boundary condition at the sea bed. t) = e j!t Z1 0 g(») cos(k») d» ¡k C1 (k) º cosh [kh] ¡ k sinh [kh] ½ ¾ A0 Z1 0 k cosh [k(z ¡ h)] C1 (k) sinh [kz] ¡ º cosh [kh] ¡ k sinh [kh] cos(ky) dk . The additional potential for a restricted water depth.2. z. t) = ej!t A0 fC1 (k) sinh [kz] + C2 (k) cosh [k(z ¡ h)]g cos(ky)dk (Keil-14) From the free surface condition (Keil-2) follows for jyj > B=2: ej!t A0 Z1 0 fºC2 (k) cosh [kh] + kC1 (k) ¡ kC2 (k) sinh [kh]g cos(ky)dk = 0 The solution of this Fourier integral equation: Z1 0 f (k) cos(k»)dk = g(») is known: 1 f (k) = ¼ Also will be obtained: f (k) = kC1 (k) + C2 (k) fº cosh [kh] ¡ k sinh [kh]g = 0 from which follows: C2 (k) = With this will be obtained: ©0rad (y. ©j can be written in the same way. will be determined in such a way that it ful…lls the free surface condition and . Alternative Derivation Assuming that the real part of the potential at an in…nite water depth. 1960]. z.4. ©rad . another derivation of the 2-D potential is given by [Porter. THEORY OF KEIL 81 Z1 ¡kh <Z e¡kz e º cosh [kz] ¡ k sinh [kz] ¢ sin(ky) dk ¡ sin(ky) dk : º ¡k º ¡k º cosh [kh] ¡ k sinh [kh] 0 0 ¾ cosh [º 0 h] sinh [º 0 (z ¡ h)] ¡j¼ sin(º 0 y) º 0 h + sinh [º 0 h] cosh [º 0 h] 93 In here. As an extension of the additional real potential will be chosen: Z1 0 ©0rad (y. ©r1 .is known.together with ©r1 .

1956] and [Grim. t) + 0 h 0 i 0 0 = A0 ©0r1(y. of which . z. z. t) + j©nj (y. z.and the stream function (Keil-13b) respectively . z. 2-D Multi-Potential The free surface conditions can not be ful…lled with the potential (Keil-12b) .only.94 CHAPTER 4.the real part of the additional potential ©nrad and the imaginary part potential ©nj will be determined. t) + j©0j (y. t) 0 (Keil-15) i h 0 0 0 An ©nr1(y. as given in equation (Keil-12b). Additional potentials ©n are required which ful…ll the boundary conditions (Keil-1) through (Keil-4) and together with ©0 also ful…ll the boundary conditions (Keil-5) or (Keil-6): ©(y. t) = A0 ©0 (y. z. z. 2-D POTENTIAL COEFFICIENTS The still unknown function C1 (k) follows from the boundary condition at the sea bed: · @©0r1 @©0rad + @z @z ¸ =0= z=h So: 8 1 9 Z1 < Z ke¡kh = cos(ky) dk + kC1 (k) cosh [kh] cos(ky) dk = ej!t A0 ¡ : . t) = +ej!t An (k + º)k 2(n¡1) e¡kz cos(ky) dk 0 . z. t) + ©0rad (y. becomes: Z1 0 ©0rad (y. z. t) = e j!t A0 e¡kh º sinh [kz] ¡ k cosh [kz] ¢ cos(ky) dk º ¡ k º cosh [kh] ¡ k sinh [kh] (Keil-14a) The imaginary part can be obtained as described before. 1957]. t) Use will be made here of multi-potentials given by [Grim. z. the extension of the additional real potential. t) + 1 X n=1 1 X n=1 An ©n (y. t) + ©nrad (y.using the Sommerfeld Radiation Condition . z. This results in: Z1 ©nr1 (y. º¡k 0 0 C1 (k) = C2 (k) = e¡kh (º ¡ k) cosh [kh] ¡ke¡kh (º ¡ k) fº cosh [kh] ¡ k sinh [kh]g cosh [kh] With this. z.

t) + jªnj (y. z. z. t) + ©nrad (y. z. z. t) = +ej!t An sinh [º 0 (z ¡ h)] sin(º 0 y) cosh [º 0 h] º 0 h + sinh [º 0 h] cosh [º 0 h] ¢ (Keil-17b) ¼º 2n 0 The potentials ©nj and ©nrad disappear in deep water.2.4. t) = e 1 X© £ ¤ª Anr ['nr1 + 'nrad ] ¡ Anj 'nj + j Anj ('nr1 + 'nrad ) + Anr 'nj n=0 h 0 i h io 0 0 0 0 0 Anr ©nr1 + ©nrad ¡ Anj ©nj + j Anj (©nr1 + ©nrad ) + Anr ©nj (Keil-18) and 1 X n=0 ª(y. z. z. t) = +e j!t (Keil-17) ªnrad (y. t) = +ej!t An (k + º)k 2(n¡1) e¡kh cos(ky) dk º cosh [kh] ¡ k sinh [kh] 0 (Keil-16a) ©nj (y. t) + j©nj (y. z. z. z. t) . z. t) = = n=0 1 Xn n=0 j!t h 0 i 0 0 [Anr + jAnj ] ©nr1 (y. z. t) = 1 h 0 i h io Xn 0 0 0 0 0 = Anr ªnr1 + ªnrad ¡ Anj ªnj + j Anj (ªnr1 + ªnrad ) + Anr ªnj n=0 h 0 i 0 0 [Anr + jAnj ] ªnr1(y. t) = ¡ej!t An cosh [º 0 (z ¡ h)] cos(º 0 y) cosh [º 0 h] º 0 h + sinh [º 0 h] cosh [º 0 h] ¢ (Keil-16b) The orthogonality condition provides the stream function: Z1 An (k + º)k 2(n¡1) e¡kz sin(ky) dk 0 ¼º 2n 0 ªnr1 (y. z. Total Potentials Only the complex constant An with (0 6 n 6 1) in the potential has to be determined: 1 X ©(y. z. t) + ªnrad (y. THEORY OF KEIL 95 (Keil-16) Z1 º sinh [kz] ¡ k cosh [kz] ©nrad (y. t) = ¡ej!t An Z1 0 (k + º)k 2(n¡1) e¡kh º cosh [kz] ¡ k sinh [kz] sin(ky) dk º cosh [kh] ¡ k sinh [kh] (Keil-17a) ªnj (y.

An 'n and An à n have the dimensions of a length [L]. These coe¢cients are dimensional and it is very practical to determine them for the amplitude of ¹ the ‡ow velocity V . z. the complex total potential can now be written as: ©(y. 2-D POTENTIAL COEFFICIENTS 1 X© £ ¤ª Anr [à nr1 + à nrad ] ¡ Anj à nj + j Anj (à nr1 + à nrad ) + Anr à nj n=0 (Keil-19) Summarized. .96 = e j!t CHAPTER 4. t) = ej!t [A0r + jA0j ] 81 <Z cos (k [y + i(z ¡ h)]) dk : º cosh [kh] ¡ k sinh [kh] 0 (Keil-20) +ej!t 1 X n=1 [Anr + jAnj ] ¾ cosh [º 0 h] cos (º 0 [y + i(z ¡ h)]) +j¼ º 0 h + sinh [º 0 h] cosh [º 0 h] 81 <Z cos (k [y + i(z ¡ h)]) dk (º 2 ¡ k 2 )k 2(n¡1) : º cosh [kh] ¡ k sinh [kh] 0 ¾ ¼º 2n cos fº 0 (y + i(z ¡ h))g 0 ¡j ¢ cosh [º 0 h] º 0 h + sinh [º 0 h] cosh [º 0 h] The coe¢cients Anr and Anj with (0 6 n 6 1) have to be determined in such a way that the instantaneous boundary conditions on the body surface have been ful…lled. z. p For ºr = º y 2 + z 2 ! 0: '0r1 (" 1 X 0 0 à 0r1 # ºm = e¡ºz Re f(z + iy)m g cos(ºy) ° + ln(ºr) + m ¢ m! m=1 # ) " 1 X ºm y (Keil-21) + arctan + Im f(z + iy)m g sin(ºy) z m=1 m ¢ m! (" # 1 X ºm = e¡ºz ° + ln(ºr) + Re f(z + iy)m g sin(ºy) m ¢ m! m=1 " # ) 1 X ºm y (Keil-21a) ¡ arctan + Im f(z + iy)m g cos(ºy) z m=1 m ¢ m! with the Euler constant: ° = 0:57722. Expansion of Potential Parts The expansion of the potential parts at an in…nite water depth is given by Grim. also if they then have the dimension [L2n+1 ]: ³ ´ ¹ Ar + jAj = V A0 + jA0 ¹ Ar + jAj = V r j ¹ V Then. t) + iª(y.

2. THEORY OF KEIL For ºr = º p y 2 + z 2 ! 1: '0r1 = à 0r1 Mind you that (m¡1)! º m r2m M X (m ¡ 1)! y Re f(z + iy)m g + ¼e¡ºz sin(ºy) º m r2m jyj m=1 97 M X (m ¡ 1)! y = Im f(z + iy)m g ¡ ¼e¡ºz cos(ºy) m r 2m º jyj m=1 f(z + iy)m g is semi–convergent.4. © ª 'nr1 = (¡1)n (2n ¡ 1)! Re (y + iz)¡2n + à nr1 © ª º Im (y + iz)¡(2n¡1) 2n ¡ 1 (Keil-22) © ª © ª º = (¡1)n (2n ¡ 1)! Im (y + iz)¡2n ¡ Re (y + iz)¡(2n¡1) 2n ¡ 1 (Keil-22a) For the expansion of the remaining potential parts use has been made of the following relations as derived in Appendix 2: 1 X k 2t © ª Re (z + iy)2t cosh [kz] cos(ky) = (2t)! t=0 sinh [kz] sin(ky) = 1 X k 2t+1 ª © sinh [kz] cos(ky) = Re (z + iy)2t+1 (2t + 1)! t=0 1 X k 2t ª © Im (z + iy)2t (2t)! t=0 1 X k 2t+1 ª © cosh [kz] sin(ky) = Im (z + iy)2t+1 (2t + 1)! t=0 With these relations follows from equation (Keil-14a): '0rad = Z1 0 ) 1 1 X k 2t+1 X k 2t © © ª ª ¢ º Re (z + iy)2t+1 ¡ k Re (z + iy)2t (2t + 1)! (2t)! t=0 t=0 1 Z 1 X 1 k 2t+1 e¡kh = dk ¢ (2t + 1)! (º ¡ k) (º cosh [kh] ¡ k sinh [kh]) t=0 0 ª © ªª © © ¢ º Re (z + iy)2t+1 ¡ (2t + 1) Re (z + iy)2t 1 X G(2t + 1) © © ª © ªª = ¡ º Re (z + iy)2t+1 ¡ (2t + 1) Re (z + iy)2t (2t + 1)! t=0 ( e¡kh dk ¢ (º ¡ k) (º cosh [kh] ¡ k sinh [kh]) (Keil-23) .

98 It is obvious that: 1 X t=0 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS ª0rad = ¡ ª © ªª © G(2t + 1) © (2t + 1) Im (z + iy)2t ¡ º Im (z + iy)2t+1 (2t + 1)! Z1 0 (Keil-23a) The function: G(t) = k t e¡kh dk (k ¡ º) [º cosh [kh] ¡ k sinh [kh]] will be treated in the next section. 0 © © ª © ªª ¢ º Re (z + iy)2t+1 ¡ (2t + 1) Re (z + iy)2t 1 X G(2t + 2n + 1) ¡ º 2 ¢ G(2t + 2n ¡ 1) ¢ = (2t + 1)! t=0 © © ª © ªª (Keil-24) ¢ º Re (z + iy)2t+1 ¡ (2t + 1) Re (z + iy)2t 1 X G(2t + 2n + 1) ¡ º 2 ¢ G(2t + 2n ¡ 1) ( (k 2 ¡ º 2 ) k 2(n¡1) e¡kh dk ¢ (k ¡ º) (º cosh [kh] ¡ k sinh [kh]) à nrad = For the imaginary parts can be written: ¢ (2t + 1)! t=0 © © ª © ªª ¢ (2t + 1) Im (z + iy)2t ¡ º Im (z + iy)2t+1 (Keil-24a) cosh2 (º 0 h) (Keil-25) 'oj = +¼ ¢ º 0 h + sinh [º 0 h] cosh [º 0 h] (1 ) 1 X º 2t X º 2t+1 © ª © ª 0 0 ¢ Re (z + iy)2t ¡ tanh [º 0 h] Re (z + iy)2t+1 (2t)! (2t + 1)! t=0 t=0 . it follows from (Keil-16a): Z1 0 'nrad = It is clear that: ) 1 1 X k 2t+1 X k 2t © © ª ª Re (z + iy)2t+1 ¡ k Re (z + iy)2t ¢ º (2t + 1)! (2t)! t=0 t=0 81 1 <Z X 1 k 2t+2n+1 e¡kh dk = (2t + 1)! : (k ¡ º) (º cosh [kh] ¡ k sinh [kh]) t=0 0 9 Z1 = 2t+2n¡1 ¡kh k e ¡º 2 dk ¢ (k ¡ º) (º cosh [kh] ¡ k sinh [kh]) . Further.

dw + .dw + ..du + .. First.4..dw = 0 I II III IV &p 2 ¢ ei¾ 2 J = I . THEORY OF KEIL 99 Function G(t) The function: º 2n 0 (Keil-26) 'nj = ¡¼ ¢ º 0 h + sinh [º 0 h] cosh [º 0 h] (1 ) 1 X º 2t X º 2t+1 © ª © ª 0 0 ¢ Re (z + iy)2t ¡ tanh [º 0 h] Re (z + iy)2t+1 (2t)! (2t + 1)! t=0 t=0 ¡ 2 ¢ = ¡º 2(n¡1) º 0 ¡ º 2 ¢ 'oj 0 º 2n 0 (Keil-26a) à nj = +¼ ¢ º 0 h + sinh [º 0 h] cosh [º 0 h] (1 ) 1 X º 2t X º 2t+1 ª ª © © 0 0 2t 2t+1 ¢ Im (z + iy) ¡ tanh [º 0 h] Im (z + iy) (2t)! (2t + 1)! t=0 t=0 ¢ 2(n¡1) ¡ 2 = ¡º 0 º 0 ¡ º 2 ¢ à oj cosh2 (º 0 h) (Keil-25a) à oj = ¡¼ ¢ º 0 h + sinh [º 0 h] cosh [º 0 h] (1 ) 1 X º 2t X º 2t+1 © ª © ª 0 0 ¢ Im (z + iy)2t ¡ tanh [º 0 h] Im (z + iy)2t+1 (2t)! (2t + 1)! t=0 t=0 G(t) = Z1 0 k t e¡kh dk (k ¡ º) (º cosh [kh] ¡ k sinh [kh]) with unit: £ 1¡t ¤ L has two singular points: k = º and k = º 0 .....2.. Thus.. it is not possible to solve this integral directly.dw + .. this integral will be normalized: G (t) = G(t) ¢ ht¡1 Z1 ut e¡u du = (u ¡ ºh) (ºh cosh [u] ¡ u sinh [u]) 0 0 A substitution of: w = u + iv = provides: wt e¡w dw (w ¡ ºh) (ºh cosh [w] ¡ w sinh [w]) Z1 Z Z Z Z .

So.du + JI + JII + JIII + JIV (Keil-27a) = 0 From this follows: Z1 0 ut e¡u du = ¡ Re fJI + JII + JIII + JIV g (u ¡ ºh) (ºh cosh [u] ¡ u sinh [u]) JI and JII are imaginary because they are residues and JIII = 0 for R ! 1. it remains: Z1 0 With the complex function: ut e¡u du = ¡ Re fJIV g (u ¡ ºh) (ºh cosh [u] ¡ u sinh [u]) 9 8 = <Z t ¡w we dw = ¡ Re : (w ¡ ºh) (ºh cosh [w] ¡ w sinh [w]) . ~ ~ ~ ~ IV (Keil-28) . IV (w ¡ ºh) (ºh cosh [w] ¡ w sinh [w]) ~ ~ ~ ~ with: w = ~ &p 2 ¢ ei¾ 2 the nominator of this integral will be made real by removing... So: 0 G (t) = 8 9 <Z = t ¡w w e (w ¡ ºh) (ºh cosh [w] ¡ w sinh [w]) ~ ~ ~ ~ ¡ Re dw : (w ¡ ºh) (w ¡ ºh) (ºh cosh [w] ¡ w sinh [w]) (ºh cosh [w] ¡ w sinh [w]) .7: Singularities in G-Function Z1 0 = . 2-D POTENTIAL COEFFICIENTS Figure 4.100 CHAPTER 4.

. However. 7. 11. THEORY OF KEIL t¼ = ¡ cos ¢ 0 4 101 Because t is always odd: t¼ G (t) = ¡ cos 4 0 B Z1 Bµ B & ¶t B B p B 2 0 B @ 2 (ºh)2 ¢ ¡ ¢ ª 1 ¡ tan t¼ © (cos & + e¡& ) + 1 + tan t¼ sin & 4 4 ª ¡4ºh& cos & + tan t¼ ¡sin & 4 ©¡ ¢ ¢ ª +& 2 1 + tan t¼ (cos & ¡ e¡& ) ¡ 1 ¡ tan t¼ sin & 4 4 f& 2 ¡ 2ºh (& ¡ ºh)g ¢ ¢ fcosh [&] (2º 2 h2 + & 2 ¡ 2ºh& tanh [&]) + (2º 2 h2 ¡ & 2 ) cos & + 2ºh& sin &g ©¡ 1 C C C C C d& C C A µ ¶t Z1 0 4 (ºh)2 sin & + 2& 2 (cos & ¡ e¡& ) ¡ 4ºh& (cos & + sin &) denominator & p 2 d& = ¡ cos t¼ 4 Z1 0 for t = 1. .. This will be investigated …rst. µ ¶ 4 (ºh)2 (cos & + e¡& ) ¡ 2& 2 sin & ¡ 4ºh& (cos & ¡ sin &) & t p d& denominator 2 for t = 3. . the integral increases monotone with decreasing ºh. 5... This will be approximated by: lim G1 (1) = ue¡u du (u ¡ ºh) (ºh ¡ u2 ) (Keil-29) ºh!0 This can be written as: 8 < 1 : 1 ¡ ºh Z1 0 ºh!0 lim G1 (1) = lim ºh!0 e¡u du u ¡ ºh Z1 0 1 ´ + p ³p 2 ºh ºh ¡ 1 e¡u p du u ¡ ºh ¡u From: Z1 0 1 ´ + p ³p 2 ºh ºh + 1 ( Z1 0 e p du u + ºh . G (1) = G(1) = 0 Z1 0 ue¡u du (u ¡ ºh) (ºh cosh [u] ¡ u sinh [u]) Z1 0 (Keil-28a) This integral converges fast for small ºh values.2. 9. ) 9 = e¡u du = ¡e¡a ¢ u¡a am ° + ln jaj + m ¢ m! m=1 1 X . 0 0 For t > 1 the function G (t) becomes …nite.4. G (1) does not converge for ºh ! 0.

With a known potential. these pressures can be found from the linear part of the instationary pressures as follows from the Bernoulli equation: @© @t pdyn = p ¡ pstat = ¡½ . Hydrodynamic Loads The hydrodynamic loads can be found from an integration of the pressures on the hull of the oscillating body in (previously) still water.102 follows: CHAPTER 4. 2-D POTENTIAL COEFFICIENTS " # 1 X (ºh)m ¡e¡ºh lim G1 (1) = lim ° + ln (ºh) + ºh!0 ºh!0 1 ¡ ºh m ¢ m! m=1 " # p 1 X (ºh) m 2 e¡ ºh 1 + p ³ p ´ ° + 2 ln (ºh) + m ¢ m! 2 ºh 1 ¡ ºh m=1 p ºh ( e 1 ¡ p ³ p ´ ° + 2 ln (ºh) + 2 ºh 1 + ºh m=1 ( " # 1 X (ºh)m ¡e¡ºh = lim ° + ln (ºh) + ºh!0 1 ¡ ºh m ¢ m! m=1 ³ p ´ ¡pºh " # 1 1 + ºh e X (ºh) m p 2 1 ° + ln (ºh) + ºh + + p 2 m ¢ m! 2 ºh (1 ¡ ºh) m=2 ´ p " ³ p #9 m 1 1 ¡ ºh e ºh X p 1 (ºh) 2 = ° + ln (ºh) ¡ ºh + (¡1)m ¡ p 2 m ¢ m! . m 2 (Keil-29a) The imaginary part of integral (Keil-27a) has been treated in Appendix 3. 2 ºh (1 ¡ ºh) m=2 ( " # 1 X (ºh)m ¡e¡ºh = lim ° + ln (ºh) + ºh!0 1 ¡ ºh m ¢ m! m=1 hp i 3 2 ¸ hp i sinh ºh · 1 1 4 5 ° + ln (ºh) p + cosh ºh ¡ 1 ¡ ºh 2 ºh hp i hp i sinh ºh cosh ºh ¡ + 1 ¡ ºh " 1 ¡ ºh 1 1 p p X (ºh) m¡1 X (ºh) m 2 2 1 ¡ ºh ¡ ºh +e + e 2 (1 ¡ ºh) m ¢ m! m ¢ m! m=2 m=2 #) m¡1 m 1 1 p p X X (ºh) 2 (ºh) 2 ¡e ºh (¡1)m (¡1)m + e ºh m ¢ m! m ¢ m! m=2 m=2 = 1 ¡ ° ¡ ln (ºh) " 1 X #9 (ºh) = (¡1)m m ¢ m! .

THEORY OF KEIL = ¡j!½© 103 The potential is in-phase with the oscillation velocity. This is . which means a multiplication with -j. The energy balance provides: ¹ ¹2 ³ !2 NV ¹ ¢ A2 = 2 = V z ¹ ½g 2 ¢ cgroup !º 0 sinh [2º 0 h] = ¢ NV ½g 2º 0 h + sinh [2º 0 h] º2 cosh2 [º 0 h] ¹ = FV j ¹ ¢ ½! V º 0 h + sinh [º 0 h] cosh [º 0 h] The imaginary part of the force must be equal to the hydrodynamic damping coe¢cient times the oscillatory velocity.a force per unit length.4. The vertical force becomes: Z S FV = FV r + jFV j = pdy h i ¹ F 0 + jF 0 (Keil-30) = ½! V V r Vj Z h 1 ³ 0 ´i X 0 0 0 ¹ ej!t = ¡½! V Anr 'nr ¡ Anj 'nj + j Anr 'nj + Anj 'nr dy n=0 S The real part of this force is equal to the hydrodynamic mass coe¢cient times the oscillatory acceleration. from which the damping coe¢cient follows: ¯ ¯ ¯FV j ¯ ¹ NV = ¹ V (Keil-30b) In deep water becomes the hydrodynamic mass for º ! 0 in…nite.2. To obtain the phase of the pressures with respect to the oscillatory motion a phase shift of -900 is required. from which the hydrodynamic mass coe¢cient follows: m" = or non-dimensional: CV = ¹ m" FV r ¼ 2 = ¼ 2 ¹ ½8B ½ 8 B !V (Keil-30a) ¹ FV r FV r = ¹ b !V Instead of this coe¢cient. generally the ratio between amplitude of the radiated wave ¹ ³ and the oscillatory motion z will be used.in the two-dimensional case . because potential (Keil-21) becomes: . Then the pressure is: pdyn = ¡½!© The hydrodynamic force on the body is equal to the integrated pressure on the body.

just as in deep water .104 CHAPTER 4. which is decisive (in…nite mass) in deep water. 2-D POTENTIAL COEFFICIENTS º!0 lim '0r1 = ° + ln (ºr) (Keil-30c) and the non-dimensional mass of a circle becomes: m = lim ¼ 2 º!0 ½ B " 8 # 1 X 8 1 ¡° ¡ ln (ºr) + = 2 ¼ n (4n2 ¡ 1) n=1 ¹ dAV =B º!0 dº lim The hydrodynamic mass for º ! 0 in shallow water remains …nite.provides an in…nite value. it follows from equation (Keil–12c) for the amplitude ratio in shallow water: ¹ ³ !¼ cosh2 [º 0 h] ¹ = A0 ¢ AV = z ¹ g¹ º 0 h + sinh [º 0 h] cosh [º 0 h] z º¼ cosh2 [º 0 h] = A0 ¢ !¹ º 0 h + sinh [º 0 h] cosh [º 0 h] z A0 cosh2 [º 0 h] = ¹ ¼º º 0 h + sinh [º 0 h] cosh [º 0 h] V º 0 sinh [º 0 h] cosh [º 0 h] 0 = A0 ¼ º 0 h + sinh [º 0 h] cosh [º 0 h] Because: A¼ lim 0 = 1 º!0 B it follows: º 0 sinh [º 0 h] cosh [º 0 h] ¹ lim AV = B º!0 º 0 h + sinh [º 0 h] cosh [º 0 h] 0 .in this deep water case .becomes: º!0 It is obvious that equation (Keil-30d) . the radiation potential has to be discussed only. When the contributions of the multi-potentials (which disappear here for the borderline case º ! 0) are ignored. the borderline change-over ”deep to shallow” water provides for this radiation potential: lim ['0r1 + '0rad ] = ° + ln (ºr) ¡ ln (ºh) ³r´ = ° + ln h 0 º!0 lim CV and the amplitude ratio .provide …nite contributions. Because the multipotentials .just as equation (Keil-30c) .

AV (ºB=2) has at ºB=2 = 0 a vertical tangent. This means that all ‡uid particles are in phase with the motions of the body. Wave Loads The wave forces on the restrained body in waves consists of forces F1 in the undisturbed incoming waves (Froude-Krylov hypothesis). The term FV r = m" =½ has to be in…nite. FV j is …nite: ¹ ¹ FV j NV A2 g2 ¹ ¹0 FV j = = = 4 A2 = V ¹ ½! ! V º2 ½! V ¹ ¹0 ¹0 Because limº 0 !0 AV = BV follows limº 0 !0 FV j = B 2 .4. the longer becomes the radiated wave and the faster travels it away from the cylinder. its maximum velocity is gh In case of long waves at shallow water. However. The fact that the hydrodynamic mass goes to in…nity for zero frequency can be explained physically as follows. and the forces caused by the disturbance of the waves by the body. The smaller the frequency becomes. one part F2 in phase with the accelerations of the water particles and another part F3 in phase with the velocity of the water particles: FE = F1 + F2 + jF3 ¹0 FV i = arctan ¹ 0 6= 0 F Vr . the energy has the same velocity. In the borderline case º = 0 has the wave an in…nite length and it travels away . This means that: "HT 6=1 ¹0 So or FV r = m" =½ has to be …nite.just as the pressure (incompressible ‡uid) with an in…nite velocity. THEORY OF KEIL and: ¹ dAV º!0 dº lim ¹ dAV º 0 !0 dº 0 1 B = lim 2 º 0 !0 º 0 h + sinh [º 0 h] cosh [º 0 h] = 1 = lim 105 ¹ Thus. However. which holds too that: "HT =1 ¹0 ¹ FV j FV i = arctan ¹ = arctan ¹ 0 = 0 FV r FV r ¹0 ¹0 ¹0 This condition is ful…lled only when FV j = 0 or FV r = m" =½ = 1. A full explanation is not given here.2. From that can be concluded that at low decreasing frequencies the damping part in the hydrodynamic force will increase. it has been shown here that the result makes some sense. At shallow water can the wave (even p an incompressible ‡uid) not travel in with an in…nite velocity. The …nite value of the hydrodynamic mass at shallow water is physically hard to interpret. This means that the hydrodynamic force is in phase with the motion of the body.

the theory of [Haskind.106 CHAPTER 4. then: FE = ¡½!e j!t Zh µ 0 @' @' ¡' w 'w @y @y ¶ dz (Keil-32) According to equation (Keil-A4) in Appendix 1 is: 'W = ¹ ³ ¢! fcosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]g cos (º 0 y) º . 2-D POTENTIAL COEFFICIENTS These forces will be determined from the undisturbed wave potential ©W and the disturbance potential ©S . for ¹ 6= 900 only an approximation will be found. When ©W = 'W ej!t is the potential of the incoming wave and ©S = 'S ej!t is the potential of the disturbance by the body at a large ¹ distance from the body with a velocity amplitude V = 1. FE = F1 + F2 + jF3 = Z = ¡½! [©W + ©S ] dy S j!t Z ½¹ ³! ¡jº 0 x cos ¹ = ¡½!e (cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) cos (º 0 y sin ¹) e º S ) 1 ³ 0 ´i Xh 0 0 0 ¹ (Keil-31) +V Anr 'nr ¡ Anj 'nj + j Anr 'nj + Anj 'nr dy n=0 In here: ¹ ½! 2 ³ j(!t¡º 0 x cos ¹) e = ¡ º Z S F1 f(cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) cos (º 0 y sin ¹)g dy F2 1 i Xh 0 0 ¹ ej!t = ¡½! V Anr 'nr ¡ Anj 'nj dy n=0 1 Xh n=0 ¹ F3 = ¡½! V ej!t i 0 0 Anr 'nj + Anj 'nr dy ¹ ¹ Using V = ³!. E2 and E3 . 1957] can be used too to ¹ ¹ ¹ determine the amplitudes E1 . As mentioned before. the non-dimensional amplitudes are: ¹ E1 = ¹ E2 ¹ E3 Z ¹ F1 1 ¹ = ¡ B f(cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) cos (º 0 y sin ¹)g dy ½g ³B S 1 i ¹ º Xh 0 F2 0 =¡ Anr 'nr ¡ Anj 'nj dy = ¹ B ½g ³B n=0 1 i ¹ F3 º Xh 0 0 = Anr 'nj + Anj 'nr dy ¹ = ¡B ½g ³B n=0 In case of ¹ = 900 . so beam waves.

Scale factors are given in the table below. the coordinates of the cross section are: y = (1 + a) cos µ + b cos (3µ) z = (1 ¡ a) sin µ ¡ b sin (3µ) Then: ¡ ¢ z + iy = i (y ¡ iz) = i e¡i¢µ + ae+i¢µ + be+i¢3µ (Keil-33b) (Keil-33) (Keil-33a) All calculations will be carried out in the Lewis domain. THEORY OF KEIL 107 From the previous subsections follows the asymptotic expression for the disturbance po¹ tential in still water with V = 1: ¼ cosh2 [º 0 h] = (cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) sin (º 0 y) ¢ º 0 h + sinh [º 0 h] cosh [º 0 h] # " 1 ´ ¡ 2 ¢X³ 0 0 0 0 2(n¡1) ¢ A0r + jA0j ¡ º 0 ¡ º 2 Anr + jAnj º 0 n=1 'y!1 Substituting this in equation (Keil-32).2. provides: = ¡½g ¹ 0 ej!t ³º ¢ Zh FE 2¼ cosh2 [º 0 h] ¢ º 0 h + sinh [º 0 h] cosh [º 0 h] (cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z])2 dz ¢ 0 0 ¢ A0r + jA0j ¡ º 2 ¡ º 0 = ¡½g ¹ j!t ³¼e Non-dimensional: " 0 " 1 ´ ¢X³ 0 0 0 0 2(n¡1) 2 2 A0r + jA0j ¡ º 0 ¡ º Anr + jAnj º 0 ¡ 1 ¢X³ 2 n=1 Anr + jAnj º 2(n¡1) 0 0 0 ¡ ´ # n=1 # ¹ ¹ E1 + E2 " # © ª 1 ¹ ¡ 2 ¢ X 0 2(n¡1) Re FE ¼ 0 2 = =¡ Anr º 0 A0r ¡ º 0 ¡ º ¹ B ½g ³B n=1 " # © ª 1 ¹ ¡ 2 ¢ X 0 2(n¡1) Im FE ¼ 0 2 =¡ = Anj º 0 A0j ¡ º 0 ¡ º ¹ B ½g ³B n=1 (Keil-32a) ¹ E3 Solution The Lewis transformation of a cross section is given by: y + iz = e+i¢µ + ae¡i¢µ + be¡i¢3µ Then. .4.

All stream function parts and boundary conditions can be given as a Fourier series: Ãn = or with 1 2µ 2 X (2m + 1)2 £ ¤ sin (2kµ) cos [(2m + 1) µ] = 1 ¡ + ¼ ¼ k=1 k 4k 2 ¡ (2m + 1)2 1 X 1 X The yet unknown complex coe¢cients An (0 6 n 6 1). 2-D POTENTIAL COEFFICIENTS Ship Breadth BR Draft TI Water depth HT ¢ T I Wave number º0 Acceleration g Forces FG Lewis Form Ship Lewis Form 2 (1 + a + b) BR= f2 (1 + a + b)g 1¡a+b T I= (1 ¡ a + b) W T = HT ¢ (1 ¡ a + b) T I= (1 ¡ a + b) W F = º 0 ¢ T I= (1 ¡ a + b) (1 ¡ a + b) =T I g 1 F fT I= (1 ¡ a + b)g2 To determine the unknowns An .108 CHAPTER 4. z)body . t m=0 fcnm sin [2mµ] + dnm cos [(2m + 1) µ]g by: Ã n = an0 + anm sin [2mµ] 0 m=1 The solution of the by equating coe¢cients generated equations provide the unknowns An . . an equal number of equations has to be formulated. Because Lewis forms are used only here. i i h h (Keil-34) ª (y. can be determined by substituting the stream function (Keil19) and the coordinates of the cross section in the relevant boundary conditions (Keil-5a) through (Keil-6b). the source strengths of the by the ‡ow generated singularity. z)body . a simple approach is possible. t = ªbody (y.

follows in still water from: · ¸ @© = [vn ]body @n body two boundary conditions: 1.2. However. z. Because two motions (a translation and a rotation) are considered here.4. t) = U ej!t [z]body + C 2. THEORY OF KEIL 109 4.4 Horizontal Motions Boundary Conditions The for the vertical motions made …rst four assumptions are valid for horizontal motions too. z.2. For roll: · ¸ (Keil-35a) @© @n body = [vn ]body · ¸ @© dy @© dz = ¢ ¡ ¢ @z ds @y ds body · ¸ dª = ¡ ds body · ¸ j!t dr = ¡¹ !e ' r ds body or: dªbody = ¡¹ !ej!t [r dr]body ' from which follows: ªbody (y. For sway: · @© @n ¸ = [vn ]body · ¸ @© dy @© dz = ¢ ¡ ¢ @z ds @y ds body · ¸ dª = ¡ ds body · ¸ j!t dz ¹ = ¡U e ds body body or: ¹ dªbody = Uej!t [dz]body from which follows: ¹ ªbody (y. t) = ¡ ¤ '! j!t £ 2 ¹ e y + z 2 body + C 2 (Keil-35b) . the …fth boundary condition needs here a new formulation. The potential must ful…l the motion-dependent boundary conditions which have been substituted in the equations (Keil-1) through (Keil-4).

the three dimensional radiation potential for sway and roll can be derived as: +L=2 Z ©0r (x. t) = ej!t Z1 0 A0 (») ¢ This becomes for the two-dimensional case: ©0r (y. z1 . t) = ej!t A0 Z1 0 £p ¤ p 2 2 2 2 kx + ky ¢ cosh kx + ky ¢ (z ¡ h) £p ¤ p £p ¤ sin(ky y)dkx dky d» 2 2 2 2 2 2 º cosh kx + ky ¢ h ¡ kx + ky ¢ sinh kx + ky ¢ h k cosh [k(z ¡ h)] sin(ky) dk º cosh [kh] ¡ k sinh [kh] ¡L=2 Z1 0 cos [kx (x ¡ »)] ¢ (Keil-37) (Keil-38) With the Sommerfeld radiation Condition (Keil-4) and Appendix 3 the total radiation potential becomes: ©0r (y. z. t)]body = and in oblique waves: h i ~ ªs (x. y1 . t) ¹ ³! jwt (Keil-36b) ¢ e ¢ sin(º 0 x cos ¹) ¢ º body [sin ¹ cos(º 0 y1 sin ¹) ¢ (sinh [º 0 z1 ] ¡ tanh [º 0 h] cosh [º 0 z1 ]) 3 Zy1 ¡º 0 (1 ¡ sin2 ¹) sin(º 0 y sin ¹) ¢ (sinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]) dy 5 = 0 body Potentials 2-D Radiation Potential In a similar way as equation (Keil-10) for heave. z. 2-D POTENTIAL COEFFICIENTS For the restrained body in waves. respectively: ¹ ³! jwt ¢ e ¢ [(sinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]) ¢ cos(º 0 y)]body (Keil-36a) º [ªs (y. z. y. only the force in the horizontal direction and the moment about the longitudinal axis of the body will be calculated.110 CHAPTER 4. One gets in beam waves only the in y point-symmetric part of the potential and the in y symmetric part of the stream function of the wave (see Appendix 1). t) + j©0j (y. z. t) = ej!t A0 ¢ 81 <Z k cosh [k(z ¡ h)] sin(ky) dk ¢ : º cosh [kh] ¡ k sinh [kh] 0 (Keil-39) cosh [º 0 h] cosh [º 0 (z ¡ h)] +j¼º 0 sin(º 0 y) º 0 h + sinh [º 0 h] cosh [º 0 h] ¾ and the stream function becomes: . z.

t) = +ej!t An cosh [º 0 (z ¡ h)] sin(º 0 y) cosh [º 0 h] º 0 h + sinh [º 0 h] cosh [º 0 h] ¢ (Keil-41b) The stream function related to it is: Z1 ªnr1 (y. z.4. t) = +ej!t An (k + º)k 2n¡1 e¡kz cos(ky) dk 0 ¼º 2n+1 0 . z. z. z. THEORY OF KEIL 111 ª0r (y. t) = ¡ej!t An (k + º)k 2n¡1 e¡kh sin(ky) dk º cosh [kh] ¡ k sinh [kh] 0 (Keil-41a) ©nj (y.2. t) = ¡ej!t An (k + º)k 2n¡1 e¡kz sin(ky) dk 0 (Keil-41) Z1 º sinh [kz] ¡ k cosh [kz] ©nrad (y. z. z. t) = ej!t A0 ¢ 81 <Z k sinh [k(z ¡ h)] ¢ cos(ky) dk : º cosh [kh] ¡ k sinh [kh] 0 (Keil-40) sinh [º 0 (z ¡ h)] cosh [º 0 h] +j¼º 0 cos(º 0 y) º 0 h + sinh [º 0 h] cosh [º 0 h] ¾ Potential and stream function are separated by: (Keil-39-b) ©0 = ©0r1 + ©0rad + j©0j = ej!t ¢ 81 1 Z <Z ke¡kz ke¡kh º sinh [kz] ¡ k cosh [kz] ¢ sin(ky) dk + ¢ sin(ky) dk : º¡k º ¡ k º cosh [kh] ¡ k sinh [kh] 0 0 ¾ cosh [º 0 (z ¡ h)] cosh [º 0 h] sin(º 0 y) +j¼º 0 º 0 h + sinh [º 0 h] cosh [º 0 h] (Keil-40b) ª0 = ª0r1 + ª0rad + jª0j = ej!t ¢ 81 Z Z1 ¡kh < ke¡kh º cosh [kz] ¡ k sinh [kz] ke cos(ky) dk + ¢ cos(ky) dk ¢ : º¡k º ¡ k º cosh [kh] ¡ k sinh [kh] 0 0 ¾ sinh [º 0 (z ¡ h)] cosh [º 0 h] +j¼º 0 cos(º 0 y) º 0 h + sinh [º 0 h] cosh [º 0 h] 2-D Multi-Potential The two-dimensional multi-potential becomes: Z1 ©nr1 (y. t) + jª0j (y.

t) + iª(y. z. z. t) = ¡ej!t An (k + º)k 2n¡1 e¡kh cos(ky) dk º cosh [kh] ¡ k sinh [kh] 0 (Keil-42a) ªnj (y. 0 The determination of the coe¢cients An follow from the conditions at the body contour. t) = e 1 X© ª Anr ['nr1 + 'nrad ] ¡ Anj 'nj + j [Anj ('nr1 + 'nrad )] + Anr 'nj n=0 (Keil-43a) and j!t 1 X© ª Anr [à nr1 + à nrad ] ¡ Anj à nj + j [Anj (à nr1 + à nrad )] + Anr à nj n=0 ª(y. z. t) = +ej!t An sinh [º 0 (z ¡ h)] cos(º 0 y) cosh [º 0 h] º 0 h + sinh [º 0 h] cosh [º 0 h] ¢ (Keil-42b) Total Potentials With exception of the complex constant An with (0 6 n 6 1).112 CHAPTER 4. t) = e (Keil-43b) with: h i 0 ¹ 0 Ar + jAj = U Ar + jAj 0 0 0 for sway for roll An has the dimension [L2n+2 ]. 2-D POTENTIAL COEFFICIENTS (Keil-42) Z1 º cosh [kz] ¡ k sinh [kz] ªnrad (y. An 'n and An à n have for sway the dimension [L] and for roll the dimension [L2 ]. the potential is known: ©(y. z. i h 0 0 '! Ar + jAj ¹ . Then. z. t) = +ej!t [A0r + jA0j ] (Keil-43) ¼º 2n+1 0 81 <Z k sin (k [y + i(z ¡ h)]) dk : º cosh [kh] ¡ k sinh [kh] 0 ¡ej!t 1 X n=1 [Anr + jAnj ] Writing in a similar way as for heave provides: j!t ¾ cosh [º 0 h] sin (º 0 [y + i(z ¡ h)]) +j¼º 0 º 0 h + sinh [º 0 h] cosh [º 0 h] 81 <Z sin (k [y + i(z ¡ h)]) dk (º 2 ¡ k 2 )k 2n¡1 : º cosh [kh] ¡ k sinh [kh] 0 ¾ ¼º 2n+1 sin fº 0 (y + i(z ¡ h))g 0 ¡j ¢ cosh [º 0 h] º 0 h + sinh [º 0 h] cosh [º 0 h] ©(y. z.

2.4. p For ºr = º y 2 + z 2 ! 1: '0r1 à 0r1 Mind you that M X (m ¡ 1)! y y = ¡ 2 +º Im f(z + iy)m g ¡ ¼ºe¡ºz cos(ºy) 2 m r 2m y +z º jyj m=1 M X (m ¡ 1)! z y = + 2 ¡º Re f(z + iy)m g ¡ ¼ºe¡ºz sin(ºy) 2 m r 2m y +z º jyj m=1 (m¡1)! º m r2m f(z + iy)m g is semi–convergent. n+1 'nr1 = (¡1) à nr1 (Keil-45) n © ª ªo © º Re (y + iz)¡2n = (¡1)n+1 (2n)! Im (y + iz)¡(2n+1) ¡ 2n (Keil-45a) 1 X ½ G(2t + 3) t=0 1 X½ t=0 n © © ª ªo º ¡(2n+1) ¡2n Im (y + iz) (2n)! Re (y + iz) + 2n '0rad = ¾ ª ª © © G(2t + 1) 2t+1 2t ¡º Im (z + iy) Im (z + iy) (2t + 1)! (2t)! ª0rad = ª © © G(2t + 3) G(2t + 1) Re (z + iy)2t+1 ¡ º Re (z + iy) (2t + 1)! (2t)! (Keil-46) ¾ ª 2t (Keil-46a) 'nrad = 1 X ½ G(2t + 2n + 3) ¡ º 2 ¢ G(2t + 2n + 1) t=0 (2t + 1)! © ª Im (z + iy)2t+1 . THEORY OF KEIL Expansion of Potential Parts p For ºr = º y 2 + z 2 ! 0: '0r1 y = ¡ 2 + ºe¡ºz y + z2 " 113 à 0r1 # ºm sin(ºy) ° + ln(ºr) + Re f(z + iy)m g m ¢ m! m=1 #) 1 y X ºm (Keil-44) ¡ cos(ºy) arctan + Im f(z + iy)m g z m=1 m ¢ m! ( " # 1 X ºm z Re f(z + iy)m g = + 2 ¡ ºe¡ºz cos(ºy) ° + ln(ºr) + y + z2 m ¢ m! m=1 " #) 1 y X ºm (Keil-44a) Im f(z + iy)m g + sin(ºy) arctan + z m=1 m ¢ m! 1 X ( " with the Euler constant: ° = 0:57722.

114 CHAPTER 4. 1957] give for the horizontal motions at zero frequency the complex potential: 1 X ' + ià = n=0 1 X n An (y + iz)¡(2n+1) + (Keil-50) m=1 ) n o (y + iz + i2mh)¡(2n+1) + (y + iz ¡ i2mh)¡(2n+1) ( For Lewis forms this becomes: 1 X n=0 1 X ' + ià = An e ¡i(2n+1)µ +i2 )) µ ¶ µ +iµ ¶ ¡iµ ¡i3µ 2p+1 2p + 2n + 1 e + ae + be ¢ (¡1)p 2p + 1 2mH p=0 m=1 1 X 1 Xn µ ¶ ¢p 2n + p ¡ ¡i2µ (¡1) ae + be¡i4µ p p=0 p (i2mH)¡(2n+1) ¢ . 2-D POTENTIAL COEFFICIENTS ¾ ª © G(2t + 2n + 1) ¡ º 2 ¢ G(2t + 2n ¡ 1) 2t ¡º Im (z + iy) (2t)! à nrad = 1 X½ (Keil-47) © ª G(2t + 2n + 3) ¡ º 2 ¢ G(2t + 2n + 1) Re (z + iy)2t+1 (2t + 1)! t=0 ¾ © ª G(2t + 2n + 1) ¡ º 2 ¢ G(2t + 2n ¡ 1) 2t ¡º Re (z + iy) (2t)! (Keil-47a) cosh2 (º 0 h) (Keil-48) 'oj = +¼º 0 ¢ º 0 h + sinh [º 0 h] cosh [º 0 h] (1 ) 1 X º 2t+1 X º 2t © ª © ª 0 0 ¢ Im (z + iy)2t+1 ¡ tanh [º 0 h] Im (z + iy)2t (2t + 1)! (2t)! t=0 t=0 à oj cosh2 (º 0 h) (Keil-48a) = +¼º 0 º 0 h + sinh [º 0 h] cosh [º 0 h] (1 ) 1 X º 2t+1 X º 2t © © ª ª 0 0 ¢ Re (z + iy)2t+1 ¡ tanh [º 0 h] Re (z + iy)2t (2t + 1)! (2t)! t=0 t=0 'nj = +º 0 à nj Zero-Frequency Potential ¢ º 2 ¡ º 2 ¢ 'oj 0 ¢ 2(n¡1) ¡ 2 = +º 0 º 0 ¡ º 2 ¢ à oj 2(n¡1) ¡ (Keil-49) (Keil-49a) [Grim. 1956] and [Grim.

it follows: 2HT > The potential converges too when: 2HT > breadth 6 Hydrodynamic Loads The hydrodynamic force at sway oscillations in still water becomes: ¹ FQ = FQr + jFQj = ¡½! Uej!t ¢ 1 ³ 0 ´i XZ h 0 0 0 ¢ Anr 'nr ¡ Anj 'nj + j Anr 'nj + Anj 'nr dz n=0 S (Keil-51) 1 + ae¡i2µ + be¡i4µ 1¡a+b 1+a+b B = 1¡a+b 2T 4 x water depth (Keil-51a) (Keil-51b) (Keil-52) and at roll oscillations: FR = FRr + jFRj = ¡½! 2 'ej!t ¢ ¹ 1 Z h ´i ³ 0 X 0 0 0 Anr 'nr ¡ Anj 'nj + j Anr 'nj + Anj 'nr dz ¢ n=0 S (Keil-52a) The hydrodynamic moment at sway oscillations in still water becomes: ¹ (Keil-53) MQ = MQr + jMQj = ¡½! U ej!t ¢ 1 Z h ³ 0 ´i X 0 0 0 ¢ Anr 'nr ¡ Anj 'nj + j Anr 'nj + Anj 'nr [ydy + zdz] n=0 S .2. THEORY OF KEIL 115 (1 ( ) p X X µ2n + p¶µp¶ p = (¡1) An ap¡l bl e¡i(2n+2p+2l+1)µ p l n=0 p=0 l=0 " 2p+1 l 1 1 XX XX (Keil-50a) +2 (¡1)p+n (2mH)¡2(p+n+1) 1 X µ ¶µ ¶µ ¶ ¸¾ 2n + 2p + 1 2p + 1 l l¡k k i(2p¡2l¡2k+1)µ a b e 2p + 1 l º0 m=1 p=0 l=0 k=0 These sums converge as long as: e+iµ + ae¡iµ + be¡i3µ e+iµ + ae¡iµ + be¡i3µ = <1 2m ¢ H 2m ¢ HT ¢ (1 ¡ a + b) Because m > 1.4.

116 and at roll oscillations: CHAPTER 4. Fictive moment levers are de…ned by: ¹ MQr ¹ U !m" I "!2 ' ¹ = ¹Rr F ¹ MQj ¹ HQj = ¹ UNQ NR ! 'B ¹ ¹ HRj = ¹Rj 2F 0 0 ¹ HQr = ¹ HRr Non-dimensional values for the sway motions are: ¹ m" FQr = ¹ ½¼T 2 ½! U ¼ T 2 2 2 2 ¹2 ³ º cosh2 [º 0 h] ¹ FQj = 2 = ¢ ¹ y ¹ ½! U º 0 h + sinh [º 0 h] cosh [º 0 h] ¹ MQr = ¹ T ¢ FQr ¹ MQj = ¹ T ¢ FQj CH = ¹ A2 Q ¹ HQr T ¹ HQj T and for roll motions: (Keil-54) CR ¹ A2 R ¹ HRr T ¹ Rj H T Wave Loads ¹ I" MRr = ¼ 4 = ½8T ½! 2 ' ¼ T 4 ¹8 2 ¹ ³ 4º 2 cosh2 [º 0 h] ¹ = ¢ MRj 2 = ½! 2 'B 2 º 0 h + sinh [º 0 h] cosh [º 0 h] ¹ '2 B ¹ 4 ¹ MRr = ¹ T ¢ FRr ¹ MRj = ¹ T ¢ FRj (Keil-54a) The wave loads are separated in contributions of the undisturbed wave and di¤raction: FE = F1 + F2 + jF3 = . 2-D POTENTIAL COEFFICIENTS (Keil-53a) MR = MRr + jMRj = ¡½! 2 'ej!t ¢ ¹ 1 Z h ³ 0 ´i X 0 0 0 ¢ Anr 'nr ¡ Anj 'nj + j Anr 'nj + Anj 'nr [ydy + zdz] n=0 S Of course. the coe¢cients Anr and Anj of sway and roll will di¤er.

THEORY OF KEIL = ¡½! Z S 117 [©W + ©S ] dz Z ½ ¹ ³! = ¡½!e ¡ e¡jº 0 x cos ¹ (cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) sin (º 0 y sin ¹) º S ) 1 ³ 0 ´i Xh 0 0 0 ¹ (Keil-55) +U Anr 'nr ¡ Anj 'nj + j Anr 'nj + Anj 'nr dz j!t n=0 ME = M1 + M2 + jM3 = Z ½ ¹ ³! j!t ¡ e¡jº 0 x cos ¹ (cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) sin (º 0 y sin ¹) = ¡½!e º S ) 1 ³ 0 ´i Xh 0 0 0 ¹ (Keil-56) +U A 'nr ¡ A 'nj + j A 'nj + A 'nr [ydy + zdz] nr nj nr nj n=0 The separate parts are: ½! 2 ¹ j(!t¡º 0 x cos ¹) ³ e = + º Z S F1 f(cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) sin (º 0 y sin ¹)g dz F2 F2 M1 1 i Xh 0 0 ¹ ej!t = ¡½! U Anr 'nr ¡ Anj 'nj dz ¹ = ¡½! U ej!t ½! 2 ¹ j(!t¡º 0 x cos ¹) ³ = + e º n=0 1 Xh n=0 Anr 'nj + Anj 'nr dz Z S 0 0 i f(cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) sin (º 0 y sin ¹)g ¢ M2 M2 ¢ [ydy + zdz] 1 i Xh 0 0 ¹ ej!t Anr 'nr ¡ Anj 'nj [ydy + zdz] = ¡½! U n=0 1 i Xh 0 0 ¹ ej!t = ¡½! U Anr 'nj + Anj 'nr [ydy + zdz] n=0 Dimensionless: ¹ E1 = ¹ E2 = ¹ E3 Z ¹ F1 tanh [º 0 h] =+ f(cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) sin (º 0 y sin ¹)g dz ºAx ½g ¹ 0 Ax ³º ¹ F2 tanh [º 0 h] =¡ Ax ½g ¹ 0 Ax ³º S 1 Xh n=0 1 i ¹ F3 tanh [º 0 h] X h 0 0 = =¡ Anr 'nj + Anj 'nr dz Ax ½g ¹ 0 Ax ³º n=0 i 0 0 Anr 'nr ¡ Anj 'nj dz (Keil-55a) .4.2.

a simple approach is possible. 2-D POTENTIAL COEFFICIENTS ¹ ¹ ¹ M1 + M2 HW r ¡ ¢ = ¹ ¹ T T ¢ F1 + F2 ¹ ¹ HW j M3 = ¹ T T ¢ F3 (Keil-56a) The Haskind-Newman relations are valid too here: " # © ª 1 ¹ ¡ 2 ¢ X 0 2(n¡1) Re FE ¼ 0 2 A0r + º 0 ¡ º = =+ Anr º 0 ¹ Ax ½g ³º 0 Ax n=1 " # © ª 1 ¹ ¡ 2 ¢ X 0 2(n¡1) Im FE ¼ 0 2 = =+ A0j + º 0 ¡ º Anj º 0 Ax ½g ¹ 0 Ax ³º n=1 ¹ ¹ E1 + E2 (Keil-55b) ¹ E3 Solution To determine the unknowns An . .118 CHAPTER 4. an equal number of equations has to be formulated. All stream function parts and boundary conditions can be given as a Fourier series: Ãn = or with cos [2mµ] = 1 + in: Ãn = 1 X 1 X m=0 fcnm sin [(2m + 1) µ] + dnm cos [2mµ]g 16 X m2 sin [(2k + 1) µ] ¼ k=0 (2k + 2m + 1) (2k ¡ 2m + 1) (2k + 1) 1 anm sin [(2m + 1) µ] 0 m=0 The solution of the by equating coe¢cients generated equations provide the unknowns An . Because Lewis forms are used only here.

¹ ³! j!t e fcosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]g cos (º 0 y) º ¹ ³! = ¡ ej!t fsinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]g sin (º 0 y) º ¹ ³! j!t = ¡ e fcosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]g sin (º 0 y) º ¹ ³! j!t = ¡ e fsinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]g cos (º 0 y) º = + ©W V ªW V ©W H ªW H (Keil-A2) (Keil-A2a) (Keil-A3) (Keil-A3a) When the wave travels in the xw -direction.important. so a phase shift of 900 .2.multiplied with j.2. is: ! cos fº 0 (y + i (z ¡ h)) + !tg ¹ (Keil-A1) with: c = ©W + iªW = ³ ¢ c ¢ sinh(º 0 h) º0 ¹ ³! = fcosh [º 0 (z ¡ h)] cos (º 0 y + !t) º 0 sinh [º 0 h] ¡i sinh [º 0 (z ¡ h)] sin (º 0 y + !t)g ¹ ³! = fcosh [º 0 (z ¡ h)] cos (º 0 y + !t) º cosh [º 0 h] ¡i sinh [º 0 (z ¡ h)] sin (º 0 y + !t)g ¹ ³! cosh [º 0 (z ¡ h)] cos (º 0 y + !t) º cosh [º 0 h] ¹ ³! ej!t ¢ cosh [º 0 (z ¡ h)] ejº 0 y = º cosh [º 0 h] ¹ ³! j!t = e fcosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]g ejº 0 y º ©W = ªW = ¡ ¹ ³! sinh [º 0 (z ¡ h)] sin (º 0 y + !t) º cosh [º 0 h] ¹ ³! ej!t ¢ sinh [º 0 (z ¡ h)] ejº 0 y = j º cosh [º 0 h] ¹ ³! = j ej!t fsinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]g ejº 0 y º For the vertical motions is the in y symmetrical part of the potential signi…cant.5 Appendices Appendix 1: Undisturbed Wave Potential The general expression of the complex potential of a shallow wave.4. travelling in the negative y-direction. For the horizontal motions is the in y point symmetrical part . the potential becomes: . THEORY OF KEIL 119 4.

cosh [kz] cos (ky) = cosh [kz] ¢ cosh [iky] 1 = fcosh [k (z + iy)] + cosh [k (z ¡ iy)]g 2 £ ¤ £ ¤ª 1© cosh kre+i¯ + cosh kre¡i¯ = 2( ) 1 1 1 X (kr)2t +i2t¯ X (kr)2t ¡i2t¯ = e + e 2 t=0 (2t)! (2t)! t=0 = 1 X (kr)2t t=0 (2t)! cos (2t¯) 1 X k 2t © ª = Re (z + iy)2t (2t)! t=0 . 2-D POTENTIAL COEFFICIENTS ©W = With: ¹ ³! j(!t¡º 0 xw ) fcosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]g e º xw = x cos ¹ ¡ y sin ¹ yw = x sin ¹ + y cos ¹ the potential becomes: ©W = ¹ ³! j!t jº 0 (y sin ¹¡x cos ¹) e e fcosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]g º This results for the vertical motions in: ¹ ³! j(!t¡º 0 x cos ¹) e fcosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]g cos (º 0 y sin ¹) º ©W V = + (Keil-A4) and for the horizontal motions in: ¹ ³! j(!t¡º 0 x cos ¹) e fcosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]g sin (º 0 y sin ¹) º ©W H = ¡ (Keil-A5) Appendix 2: Series Expansions of Hyperbolic Functions With: p y y 2 + z 2 ¢ e§i arctan z z § iy = = r ¢ e§i¯ the following series expansions can be found.120 CHAPTER 4.

which can be added to ©0r in equation (Keil-11) with which the by ©0r + j©0j described ‡ow of the waves (travelling from both sides of .2.4. THEORY OF KEIL 121 sinh [kz] cos (ky) = sinh [kz] ¢ cosh [iky] 1 = fsinh [k (z + iy)] + sinh [k (z ¡ iy)]g 2 £ ¤ £ ¤ª 1© = sinh kre+i¯ + sinh kre¡i¯ 2( ) 1 1 1 X (kr)2t+1 +i(2t+1)¯ X (kr)2t+1 ¡i(2t+1)¯ + = e e 2 t=0 (2t + 1)! (2t + 1)! t=0 1 X (kr)2t+1 = cos ((2t + 1) ¯) (2t + 1)! t=0 = 1 X k 2t+1 © ª Re (z + iy)2t+1 (2t + 1)! t=0 sinh [kz] sin (ky) = ¡i sinh [kz] ¢ sinh [iky] i = ¡ fcosh [k (z + iy)] ¡ cosh [k (z ¡ iy)]g 2 £ ¤ £ ¤ª i© = ¡ cosh kre+i¯ ¡ cosh kre¡i¯ 2( ) 1 1 i X (kr)2t +i2t¯ X (kr)2t ¡i2t¯ = ¡ e ¡ e 2 t=0 (2t)! (2t)! t=0 = = 1 X (kr)2t t=0 1 X t=0 (2t)! sin (2t¯) ª © k 2t Im (z + iy)2t (2t)! cosh [kz] sin (ky) = ¡i cosh [kz] ¢ sinh [iky] i = ¡ fsinh [k (z + iy)] ¡ sinh [k (z ¡ iy)]g 2 ¤ £ ¤ª £ i© = ¡ sinh kre+i¯ ¡ sinh kre¡i¯ 2( ) 1 1 i X (kr)2t+1 +i(2t+1)¯ X (kr)2t+1 ¡i(2t+1)¯ = ¡ ¡ e e 2 t=0 (2t + 1)! (2t + 1)! t=0 = 1 X (kr)2t+1 sin ((2t + 1) ¯) (2t + 1)! t=0 1 X k 2t+1 © ª Im (z + iy)2t+1 = (2t + 1)! t=0 Appendix 3: Treatment of Singular Points The determination of ©0j and his terms .

t) +j©0j (y. º ¡ j !¹ ¡ k º ¡ j !¹ cosh [kh] ¡ k sinh [kh] g g 0 ("µ © © ª# ¶ 1 2t+1 ª X Re (z + iy)2t !¹ Re (z + iy) = ¡ lim º¡j ¡ ¢ ¹!0 g (2t + 1)! (2t)! t=0 ¢ [G (2t + 1) + jH (2t + 1)]g The function: . From the Euler equation follows with this viscosity force the with time changing pressure change: · ¸ @© @ 2 © @p @© = ½ lim g ¡¹ ¡ 2 ¹!0 @t @z @t @t From this follows the approach as given in a subsection before for the two-dimensional radiation potential: ©0r (y. This approach is based on work carried out by Rayleigh and is given in the literature by [Lamb. an viscous force ½¹w will be included in the Euler equations. z. t) = Z1 cosh [k (z ¡ h)] ´ cos (ky) dk = ej!t A0 lim ³ ¹!0 º ¡ j !¹ cosh [kh] ¡ k sinh [kh] g 0 81 <Z e¡kz j!t cos(ky) dk = e A0 lim ¹!0 : º ¡ j !¹ ¡ k g 0 ³ ´ 9 Z1 = ¡kh º ¡ j !¹ sinh [kz] ¡ k cosh [kz] g e ´ ¢³ cos(ky) dk + . Because the ‡uid is assumed to be non-viscous.is also possible in another way.122 CHAPTER 4. ¹ is the dynamic viscosity and w is the velocity. 2-D POTENTIAL COEFFICIENTS the body away) is given . in a later stage this dynamic viscosity ¹ will be set to zero. z. º ¡ j !¹ ¡ k g º ¡ j !¹ cosh [kh] ¡ k sinh [kh] g 0 ©¡ ¢ ¡ ¢ª = ej!t A0 '0r1 + j'0j1 + '0rad + j'0jad The …rst integral leads to equation (Keil-12a) and the second integral can be expanded as follows: ©0rad +j©0jad = ³ ´ !¹ Z1 º ¡ j g sinh [kz] ¡ k cosh [kz] e¡kh ´ ¢³ cos(ky) dk = lim ¹!0 º ¡ j !¹ ¡ k g º ¡ j !¹ cosh [kh] ¡ k sinh [kh] g 0 ("µ © © ª# ¶ 1 2t+1 ª X Re (z + iy)2t !¹ Re (z + iy) = lim º¡j ¡ ¢ ¹!0 g (2t + 1)! (2t)! t=0 9 Z1 = ¡kh 2t+1 e k ´ h³ ´ i dk ¢ ³ . 1932] for an in…nite water depth. In this approach.

8-a..2. THEORY OF KEIL 123 G (t) + iH (t) = lim ¹!0 Z1 0 will be normalized as done before: 0 0 e¡kh kt ³ ´ ¢ h³ ´ i dk k ¡ º ¡ i !¹ º ¡ i !¹ cosh [kh] ¡ k sinh [kh] g g This is a complex integral and must be solved in the complex domain with w = u + iv .dw dw = lim ¹!0 : ...dw + .. The integrand has singularity for: !¹h and w2 w1 = ºh ¡ i g where w2 is the solution of the equation: ¶ µ !¹h cosh [w] ¡ w sinh [w] = 0 ºh ¡ i g see …gure 4.4. 0 ¡w t = 0 .8: Treatment of Singularities I 8 R <Z Z Z 9 = ¹!0 lim e w .du + ....du + JI + JII = lim ¹!0 : .. (w ¡ w1 ) (w1 cosh w ¡ w sinh w) 0 I II 8 R 9 <Z = . G (t) + iH (t) = ht¡1 [G (t) + iH (t)] Z1 ut e¡u ³ ´ h³ ´ i du = lim ¹!0 u ¡ ºh ¡ i !¹h ºh ¡ i !¹h cosh [hu] ¡ u sinh [hu] g g 0 Figure 4.

It is also possible to solve this integral independently by using another integral path: I wt e¡w dw = lim fJII + JIII + JIV g ¹!0 (w ¡ w1 ) (w1 cosh [w] ¡ w sinh [w]) = 2¼i lim fResidue (w1 ) + Residue (w2 )g ¹!0 will be calculated as done before. The imaginary part of this integral: 8 9 Z < = t ¡w we 0 H (t) = ¡ Im lim dw :¹!0 (w ¡ w1 ) (w1 cosh [w] ¡ w sinh [w]) .124 Thus: Z1 0 CHAPTER 4. It can be found that: Re fJII g = ¡ Re fJIII g Im fJII g = + Im fJIII g Then it follows: . II ¹!0 lim JIV disappears for R ! 1. II can be calculated numerically in an analog way. see …gure 4.8-b. 2-D POTENTIAL COEFFICIENTS G (t) + iH (t) = lim = Because: 0 0 ¹!0 u ¡ ºh ¡ i !¹h g fJI + JII g ¹!0 R!1 lim ³ ´ h³ ut e¡u ´ i du ºh ¡ i !¹h cosh [hu] ¡ u sinh [hu] g R!1 lim JI = 0 follows: G (t) + iH (t) = ¡ lim 0 0 ¹!0 II Z wt e¡w dw (w ¡ w1 ) (w1 cosh [w] ¡ w sinh [w]) The real part of this integral: 8 9 Z < = t ¡w we 0 G (t) = ¡ Re lim dw :¹!0 (w ¡ w1 ) (w1 cosh [w] ¡ w sinh [w]) . This integral has no singularity and the boundary ¹ ! 0 can be passed before integration.

4.2. THEORY OF KEIL

125

H (t) = ¡ Im lim JII
¹!0

0

= ¡¼ lim fResidue (w1 ) + Residue (w2 )g ¹!0 ½ ¾ cosh2 [º 0 h] t¡1 t¡1 = ¼ (º 0 h) ¡ tanh [º 0 h] º 0 h + sinh [º 0 h] cosh [º 0 h] and the imaginary additional potential becomes: © © ª ª) Re (z + iy)2t Re (z + iy)2t+1 = H (2t + 1) ¡º (2t)! (2t + 1)! t=0 ½ ¾ 1 X º 2t cosh2 [º 0 h] 0 = ¼ ¡ º 2t ¢ º 0 h + sinh [º 0 h] cosh [º 0 h] t=0 ( © ª ª) © Re (z + iy)2t Re (z + iy)2t+1 ¡º ¢ (2t)! (2t + 1)!
1 X

½

¾

©0jad

(

¼ cosh2 [º 0 h] ¢ = º 0 h + sinh [º 0 h] cosh [º 0 h] (1 ) 1 X º 2t X º 2t+1 © ª © ª 0 0 ¢ Re (z + iy)2t ¡ tanh [º 0 h] Re (z + iy)2t+1 (2t)! (2t + 1)! t=0 t=0 ¡¼e¡ºz cos (ºy) The same will be found as a di¤erence between (Keil-25) and (Keil-12a).

126

CHAPTER 4. 2-D POTENTIAL COEFFICIENTS

4.3

Theory of Frank

As a consequence of conformal mapping to the unit circle, the cross sections need to have a certain breadth at the water surface; fully submersed cross sections, such as at the bulbous bow, cannot be mapped. Mapping problems can also appear for cross sections with too high or too low an area coe¢cient. These cases require another approach: the pulsating source method of Frank, also called Frank’s Close-Fit Method. The report of [Frank, 1967] has used here to explain this method. Hydrodynamic research of horizontal cylinders oscillating in or below the free surface of a deep ‡uid has increased in importance in the last decades and has been studied by a number of investigators. The history of this subject began with [Ursell, 1949], who formulated and solved the boundary-value problem for the semi-immersed heaving circular cylinder within the framework of linearized free-surface theory. He represented the velocity potential as the sum of an in…nite set of multi-poles, each satisfying the linear free-surface condition and each being multiplied by a coe¢cient determined by requiring the series to satisfy the kinematic boundary condition at a number of points on the cylinder. [Grim, 1953] used a variation of the Ursell method to solve the problem for two-parameter Lewis form cylinders by conformal mapping onto a circle. [Tasai, 1959] and [Porter, 1960], using the Ursell approach obtained the added mass and damping for oscillating contours mappable onto a circle by the more general Theodorsen transformation. [Ogilvie, 1963] calculated the hydrodynamic forces on completely submerged heaving circular cylinders. Despite the success of the multipole expansion-mapping methods, [Frank, 1967] discusses the problem from a di¤erent view. The velocity potential is represented by a distribution of sources over the submerged cross section. The density of the sources is an unknown function (of position along the contour) to be determined from integral equations found by applying the kinematic boundary condition on the submerged part of the cylinder. The hydrodynamic pressures are obtained from the velocity potential by means of the linearized Bernoulli equation. Integration of these pressures over the immersed portion of the cylinder yields the hydrodynamic forces or moments. A simpler approximation to the solution of the two-dimensional hydrodynamic problem was used in the strip theory of ship motions introduced by Korvin-Kroukovsky. The solution of two-dimensional water-wave problems for ship sections by multipole expansion and mapping techniques have been applied to this strip theory by several authors to predict the motions of surface vessels. The work of [Frank, 1967] has largely been motivated by the desirability of devising a computer program, based on strip theory and independent of mapping techniques, to predict the response of surface ships moving with steady forward speed in oblique as well as head or following seas for all six degrees of freedom.

4.3. THEORY OF FRANK

127

4.3.1

Notation

The notation of Frank is as follows:

A(m) B C0 g (m) Iij Jij
(m)

= = = = = = = =

M(!) N N(!) ni
(m)

(m)

(m)

= = = = = = = = = = = = = = = = = = = = =

PV (m) pa pv
(m)

Qj

(m)

Qj+N s sj T t (m) vi x1 yi y0 z = x + iy zi = xi + iyi ®i ³ ³j ´j

(m)

oscillation amplitude in the m-th mode beam of cross section C0 submerged part of cross sectional contour in rest position acceleration of gravity in‡uence coe¢cient in-phase with displacement on the i-th midpoint due to the j-th segment in the m-th mode of oscillation in‡uence coe¢cient in-phase with velocity on the i-th midpoint due to the j-th segment in the m-th mode of oscillation added mass force or moment for the m-th mode of oscillation at frequency ! number of line segments de…ning submerged portion of half section in rest position damping force or moment for the m-th mode of oscillation at frequency ! direction cosine of the normal velocity at i-th midpoint for the m-th mode of oscillation Cauchy pricipal value of integral hydrodynamic pressure in-phase with displacement for the m-th mode of oscillation hydrodynamic pressure in-phase with velocity for the m-th mode of oscillation source strength in-phase with displacement along the j-th segment for the m-th mode of oscillation source strength in-phase with velocity along the j-th segment for the m-th mode of oscillation length variable along C0 j-th line segment draft of cross section time normal velocity component at the i-th midpoint for the m-th mode of oscillation abscissa of the i-th midpoint ordinate of the i-th midpoint ordinate of the center of roll complex …eld point in region of ‡uid domain complex midpoint of i-th segment angle between i-th segment and positive x-axis complex variable along C0 j-th complex input point along C0 ordinate of j-th input point

128 º = ! 2 =g ºk »j ½ ©(m) ! !k = = = = = = =

CHAPTER 4. 2-D POTENTIAL COEFFICIENTS wave number k-th irregular wave number for ajoint interior problem abscissa of the j-th input point density of ‡uid velocity potential for th m-th mode of oscillation radian frequency of oscillation k-th irregular frequency for adjoint interior problem (k-th eigen-frequency)

4.3.2

Formulation of the problem

Consider a cylinder, whose cross section is a simply connected region, which is fully or partially immersed horizontally in a previously undisturbed ‡uid of in…nite depth. The body is forced into simple harmonic motion and it is assumed that steady state conditions have been attained. The two-dimensional nature of the problem implies three degrees of freedom of motion. Therefore, consider the following three types of oscillatory motions: vertical or heave, horizontal or sway and rotational about a horizontal axis or roll. To use linearized free-surface theory, the following assumptions are made: 1. The ‡uid is incompressible and inviscid. 2. The e¤ects of surface tension are negligible. 3. The ‡uid is irrotational. 4. The motion amplitudes and velocities are small enough that all but the linear terms of the free-surface condition, the kinematic boundary condition on the cylinder and the Bernoulli equation may be neglected. Given the above conditions and assumptions, the problem reduces to the following boundary-value problem of potential theory. The cylinder is forced into simple harmonic motion A(m) ¢cos(!t) with a prescribed radian frequency of oscillation !, where the superscript may take on the values 2, 3 and 4, denoting swaying, heaving and rolling motions, respectively. It is required to …nd a velocity potential: n o (Frank-1) ©(m) (x; y; t) = Re Á(m) (x; y) ¢ e¡i!t 1. The Laplace equation: r2 ©(m) = @ 2 ©(m) @ 2 ©(m) + =0 @x2 @y 2 (Frank-2)

satisfying the following conditions:

in the ‡uid domain, i.e., for y < 0 outside the cylinder; 2. The free surface condition: @ 2 ©(m) @©(m) =0 +g @t2 @y (Frank-3)

on the free surface y = 0 outside the cylinder, while g is the acceleration of gravity.

4.3. THEORY OF FRANK 3. The sea bed boundary condition for deep water: ¯ (m) ¯ ¯ ¯ ¯ ¯ ¯r©(m) ¯ = lim ¯ @© ¯ = 0 lim ¯ @y ¯ y!¡1 y!¡1

129

(Frank-4)

4. The condition of the normal velocity component of the ‡uid at the surface of the oscillating cylinder being equal to the normal component of the forced velocity of the cylinder. i.e., if vn is the component of the forced velocity of the cylinder in the direction of the outgoing unit normal vector ~ , then n ~ ¢ r©(m) = vn n ~ (Frank-5)

this kinematic boundary condition on the oscillating body surface the kinematic boundary condition being satis…ed at the mean (rest) position of the cylindrical surface. 5. The radiation condition that the disturbed surface of the ‡uid takes the form of regular progressive outgoing gravity waves at large distances from the cylinder. According to Wehausen and Laitone, the complex potential at z of a pulsating point source of unit strength at the point ³ in the lower half plane is: 8 9 Z1 ¡ik(z¡b = < ³) 1 e G¤ (z; ³; t) = dk ln(z ¡ ³) ¡ ln(z ¡ b + 2P V ³) ; 2¼ : º¡k 0 n o b ¡ e¡iº(z¡³) so that the real point-source potential is: H(x; y; »; ´; t) = Re fG¤ (z; ³; t)g where: z = x + iy Letting: 8 < ³ = » + i´ b = » ¡ i´ ³ !2 º= g (Frank-7)

cos !t sin !t (Frank-6)

G(z; ³) =

then:

1 Re ln(z ¡ ³) ¡ ln(z ¡ b + 2P V ³) : 2¼ n o ¡iº(z¡b ³) ¡i Re e

Z1
0

e

¡ik(z¡b ³)

º¡k

dk

9 = ; (Frank-8)

© ª H(x; y; »; ´; t) = Re G(z; ³) ¢ e¡i!t

(Frank-9)

t ¡ © ª ¼ ) = Re iG(z. ³) ¢ ds n ~ = 0 : . For partially immersed cylinders. y.130 CHAPTER 4. y. C0 Im 8 < : (~ ¢ r) n ~ C0 Z where A(m) denotes the amplitude of oscillation and n(m) the direction cosine of the normal velocity at z on the cylinder. as will be shown in the following section. yield an approximation to the given contour as shown in …gure 4. Another expression satisfying all these conditions is: H(x.3 Solution of the Problem Since ship sections are symmetrical. ³) ¢ e¡i!t 2! (Frank-10) Since the problem is linear.3. ´. Both A(m) and n(m) depend on the mode of motion of the cylinder. The fact that Q(s) is complex implies that equations (Frank-12) represent a set of coupled integral equations for the real functions RefQ(s)g and ImfQ(s)g. » N+1 = » 1 and ´ N+1 < 0. The solution of these integral equations and the evaluation of the kernel and potential integrals are described in the following section and in Appendices B and C. 9 = Q(s) ¢ G(z. Take the x-axis to be coincident with the undisturbed free surface of a conventional twodimensional Cartesian coordinate system. . ´ 1 ) is located on the negative y-axis. the y-axis being the axis of symmetry of C0 . see …gure 4.9. respectively. Application of the kinematic boundary condition on the oscillating cylinder at z yields: 8 9 Z < = Re (~ ¢ r) Q(s) ¢ G(z. ³) ¢ e ¢ ds : . N straight line segments are obtained which. 2-D POTENTIAL COEFFICIENTS Equation (Frank-9) satis…es the radiation condition and also the equations (Frank-1) through (Frank-4). Connecting these N + 1 points by successive straight lines. ´ i ) of C0 to lie in the fourth quadrant so that (» 1 . ³) ¢ ds = A(m) ¢ ! ¢ n(m) . t) = Re Q(s) ¢ G(z. Select N + 1 points (» i . a superposition of equations (Frank-9) and (Frank-10) results in the velocity potential: 8 9 <Z = ¡i!t (m) (Frank-11) © (x. ». ´ N+1 ) is on the positive x-axis. together with their re‡ected images in the third quadrant. C0 where C0 is the submerged contour of the cylindrical cross section at its mean (rest) position and Q(s) represents the complex source density as a function of the position along C0 . (Frank-12) 4. For fully submerged bodies. this investigation is con…ned to bodies with right and left symmetry. Let the cross sectional contour C0 of the submerged portion of the cylinder be in the lower half plane.9. (» N+1 .

The rolling motions are about an axis through a point (0. respectively. respectively.4. The length of the i-th segment is: q¡ ¢2 ¡ ¢2 jsi j = » i+1 ¡ » i + ´ i+1 ¡ ´ i (Frank-14) while the angle made by the i-th segment with the positive x-axis is given by: ½ ¾ ´ i+1 ¡ ´ i (Frank-15) ®i = arctan » i+1 ¡ » i The outgoing unit vector normal to the cross section at the i-th midpoint (xi . y0 ) in the symmetry plane of the cylinder. The coordinates of the midpoint of the i-th segment are: xi = » i + » i+1 2 and yi = ´i + ´i+1 2 (Frank-13) for 1 6 i 6 N. Potentials and pressures are to be evaluated at the midpoint of each segment. any point on the cylinder moves with the velocity: sway: heave: ~ (2) = ¡~ A(2) ! sin !t v i (3) ~ v = ¡~ A(3) ! sin !t j (Frank-18) (Frank-19) .3.9: Axes System and Notation. heave or roll. according to the displacement equation: (Frank-17) S (m) = A(m) ¢ cos !t for m = 2. whereas the corresponding quantities for the re‡ected image in the third quadrant are denoted by the subscript ¡j. i j The cylinder is forced into simple harmonic motion with radian frequency !. length and angle associated with the j-th segment are identi…ed by the subscript j. 3 or 4. as Used by Frank The coordinates. so that by symmetry » ¡j = ¡» j and ´ ¡j = ¡´j for 1 6 j 6 N + 1. THEORY OF FRANK 131 Figure 4. In the translational modes. yi ) is: ~ i = ~ sin ®i ¡ ~ cos ®i n i j (Frank-16) where ~ and ~ are unit vectors in the directions of increasing x and y. corresponding to sway.

. n¡i (2) = ¡ni (2) and n¡i (4) = ¡ni (4) .. i. Considering a point (xi . n¡i (3) = ni (3) . yi ) on C0 .e. consistent with the previously mentioned notation. the set of coupled integral equations (Frank-12) becomes a set of 2N linear algebraic equations in the unknowns: © ª Re Q(m) ¢ (sj ) = Qj (m) © ª Im Q(m) ¢ (sj ) = QN+j (m) . the direction cosines for the three modes of motion are: ni (m) = sway: heave: roll: ni (2) = ¡ sin ®i ni (3) = + cos ®i ni (4) = + (yi ¡ y0 ) sin ®i + xi cos ®i vi (2) = ¡!A(2) sin ®i sin !t vi (3) = +!A(3) cos ®i sin !t vi (4) = +!A(4) f(yi ¡ y0 ) sin ®i + xi cos ®i g sin !t (Frank-21) (Frank-22) Equations (Frank-22) illustrate that heaving is symmetrical. With these stipulations. are anti-symmetrical modes. an inspection of this …gure yields: ½ ¾ q yi ¡ y0 2 + (y ¡ y )2 and Ri = xi µi = arctan i 0 xi ½ ¾ yi ¡ y0 = arcsin Ri ½ ¾ xi = arccos Ri Therefore. although it varies from segment to segment.e. yi ) are: sway: heave: roll: De…ning: vi (m) A(m) ! sin !t then.9.132 CHAPTER 4. i. by elementary two-dimensional kinematics. Equations (Frank-12) are applied at the midpoints of each of the N segments and it is assumed that over an individual segment the complex source strength Q(s) remains constant. the unit vector in the direction of increasing µ is: i j ~ i = ¡~ sin µi + ~ cos µi ¿ yi ¡ y0~ xi ~ = ¡ i+ j Ri Ri so that: roll: ¿ ~ (4) = Ri S (4)~ i v n o = !A(4) (yi ¡ y0 )~ ¡ xi~ sin !t i j (Frank-20) The normal components of the velocity vi (m) = ~ i ¢~ (m) at the midpoint of the i-th segment n v (xi . however. Swaying and rolling. 2-D POTENTIAL COEFFICIENTS The rolling motion is illustrated in …gure 4.

respectively and ½ denotes the density of the ‡uid. so that the integration reduces to a summation. t) are evaluated in the appendix.. !... !) ¢ ni (m) ¢ jsi j pv (m) (xi . THEORY OF FRANK Thus. yi . whence: M (m) (!) = 2 N X i=1 N X i=1 pa (m) (xi . respectively.3. t) = pa (m) (xi . the pressure as well as the potential is a function of the oscillation frequency !. !. As indicated by the notation of equations (Frank-24) and (Frank-25). t) @t (Frank-24) (Frank-25) where pa (m) and pv (m) are the hydrodynamic pressures in-phase with the displacement and in-phase with the velocity. for i = 1. The hydrodynamic force or moment (when m = 4) per unit length along the cylinder. yi . is the integral of p(m) ¢n(m) over the submerged contour of the cross section C0 . !) ¢ ni (m) ¢ jsi j (Frank-26) N (m) (!) = 2 (Frank-27) for the added mass and damping forces or moments. N : + N N o Xn o Xn Qj (m) ¢ Iij (m) + QN+j (m) ¢ Jij (m) = 0 j=1 j=1 N Xn j=1 133 (Frank-23) ¡ o Qj (m) ¢ Jij (m) + N Xn j=1 QN+j (m) ¢ Iij (m) o = ! ¢ A(m) ¢ ni (m) where the superscript (m) denotes the mode of motion.e. i.. yi . necessary to sustain the oscillations. It is assumed that the pressure at the i-th midpoint is the mean pressure for the i-th segment. t) = ¡½ ¢ as: p(m) (xi . !. yi . yi . yi ) along the cylinder is obtained from the velocity potential by means of the linearized Bernoulli equation: p(m) (xi . !) ¢ sin !t @©(m) (xi . 2. The hydrodynamic pressure at (xi . The resulting velocity potential consists of a term in-phase with the displacement and a term in-phase with the velocity. yi .4 Low and High Frequencies For very small frequencies. !) ¢ cos !t +pv (m) (xi . The velocity potentials for very small and very large frequencies are derived and discussed in the next section.. yi ..4. 4. The ”in‡uence coe¢cients” Iij (m) and Jij (m) and the potential ©(m) (xi . . as ! ! 0. yi . the free-surface condition equation (Frank-3) of the section formulating the problem degenerates into the wall-boundary condition: @© =0 @y (Frank-49) .3.

. G1 (z. where Qa is the expression for the source strength as a function of position along the submerged contour of the cross section C0 . the free-surface condition becomes the ”impulsive” surface condition: ©=0 (Frank-50) on y = 0 outside the cylinder. Therefore. Considering the Neumann problem. ³)g so that the velocity potentials for the m-th mode of motion are: Z (m) ©a (x. . ». However.yields a set of N linear algebraic equations in the N unknown source strengths Qj . ³) ! 0. note that the constant K0 in the Green’s function equation (Frank-51) yields by integration an additive constant K to the potential. For partially submerged bodies C0 is no longer (m) (m) closed. so that the contribution of K in integrating the product of the pressure with the direction cosine of the body-surface velocity vanishes. y. Let: Áa (x. ´) = Re fGa (z. But since n¡i = ¡ni for m being even. which can be inferred from the pulsating source-potential equation (Frank-8) by letting º ! 1. The appropriate complex potentials for a source of unit strength at a point ³ in the lower half plane are: o 1 n (Frank-51) G0 (z. there is a Neumann problem for the case ! ! 0 and a mixed problem when ! ! 1. whether the solutions to these problems lead to unique forces or moments.e. It remains to be shown whether these two problems are. ³) = ln(z ¡ ³) + ln(z ¡ b + K0 ³) 2¼ and: o 1 n (Frank-52) ln(z ¡ ³) ¡ ln(z ¡ b + K1 ³) G1 (z. where K0 and K1 are constants not yet speci…ed. i. for a completely submerged cylinder the cross sectional contour C0 is a simply closed curve.after application at the N segmental midpoint . ³) = 2¼ for the Neuman and mixed problems. ´) ds = A(m) n(m) n a C0 which .e. in the language of potential theory. i. well posed. y) = Q(m) (s) ¢ Áa (x. ». (Frank-4) and (Frank-5) remain valid for both asymptotic cases.134 CHAPTER 4. y. respectively. The radiation condition is replaced by a condition of boundedness at in…nity. Equations (Frank-2). and a = 1. when ! ! 1. 2-D POTENTIAL COEFFICIENTS on the surface of the ‡uid outside the cylinder. since as z ! 1. ´) ds a C0 (m) (Frank-53) for a = 0. The mixed problem raises no di¢culty.. ». In fact K1 = 0. y. whereas for extremely large frequencies. An analysis similar to the one in the section on formulating the problem leads to the integral equation: Z ~ (Frank-54) (~ ¢ r) Q(m) (s) ¢ Áa (x.

Applying the free surface condition (Frank-2) on y = T for 0 < x < B.. 1950] also showed that for a set of discrete ”irregular” frequencies the Green’s function-integral equation method failed to give a solution. k = 1... 2. He demonstrated that the irregular frequencies occurred when the following adjoint interior-potential problem had eigen-frequencies. 3.and twodimensional potential problems pertaining to oscillations of rigid bodies in a free surface. 3. (3) The heaving force on a partially submerged cylinder is not unique for. à = 0 on the surface of the cylinder below the free surface. 2.3. n¡i = (3) ni ... à k = Bk sin sinh B B where Bk are Fourier coe¢cients to be determined from an appropriate boundary condition.. 2. etc. The solutions were subject to the provisions that no point of the immersed surface of the body would be outside a cylinder drawn vertically downward from the intersection of the body with the free surface and that the free surface would be intersected orthogonally by the body in its mean or rest position. 1950] proved the existence and uniqueness of the solutions to the three. Separating variables gives the eigen-frequencies: µ ¶ · ¸ k¼x k¼y for: k = 1.. the irregular wave numbers may be easily obtained by separation of variables in the Laplace equation. in this case.5 Irregular Frequencies [John.3. ¡ º k à = 0 on the extension of the free surface inside the cylinder. [John. .. º k being the wave number corresponding to the irregular frequency ! k . 4.. + @ à = 0 inside the cylinder in the region bounded by the immersed surface of @y2 the body and the extension of the free surface inside the cylinder. so that: Z K n(3) ds 6= 0 C0 The constant K0 may be obtained by letting º ! 0 in the pulsating source-potential equation (Frank-8). For a rectangular cylinder with beam B and draft T . the eigen-wave numbers (or irregular wave numbers): · ¸ k¼T k¼ (Frank-28) coth ºk = B B . THEORY OF FRANK 135 C0 Z K n(m) ds = 0 so that the swaying force and rolling moment are unique.4. @à @y @2à @x2 2 3. . Let Ã(x. y) be such that: 1.

occurs at: s L t 5:09 !1 g which is beyond the range of practical interest for ship-motion analysis. 2. fully submerged bodies do not display these characteristics. Therefore. see for instance [Huijsmans. . In practice. º 1 ! 1=T . then from the Taylor expansion: # 1 bT (bT )3 b coth [bT ] = b + ¡ + . the phenomenon of the …rst irregular frequency of wave encounter is not too important. 2-D POTENTIAL COEFFICIENTS are obtained for k = 1. not only at the irregular frequency but also at an interval about this frequency. In particular. It should be mentioned that irregular frequencies only occur for free surface piercing bodies.in non-dimensional form with L denoting the ship length . the …rst irregular wave encounter frequency . 3..10.. for rectangular cylinders of draft T : º1 = 1 T (Frank-30) " a relation that John proved for general shapes complying with the restrictions previously outlined.. For a ship with a 7:1 length-to-beam and a 5:2 beam-to-draft. so that for most surface ships the …rst irregular frequency ! 1 is less for the midsection than for any other cross section.e.. with …nite N . etc. the lowest such irregular wave number is given by: · ¸ ¼ ¼T k1 = coth B B (Frank-29) Keeping T …xed in equation (Frank-29) but letting B vary and setting b = ¼=B. i. while for B=T = 2: º 1 = 1:71. for slender surface vessels. where the cross section is usually almost rectangular. 1996]. At an irregular frequency the matrix of in‡uence coe¢cients of equations (Frank-23) becomes singular as the number of de…ning points per cross section increases without limit. to ”close” the body by means of discretization of the free surface inside the body (putting a ”lid” on the free surface inside the body). bT 3 45 it is seen that as b ! 0. The solid line in this …gure results from including the ”lid”. This interval can be reduced by increasing the number of de…ning points N for the cross section. among others. Increasing the number of panels does not remove the irregular frequency but tends to restrict the e¤ects to a narrower band around it. see the added mass and damping of a hemisphere in …gure 4. which is equivalent to B ! 1. the determinant of this matrix becomes very small.. Most surface vessels have nearly constant draft over the length of the ship and the maximum beam occurs at or near amidships. Therefore.. For a beam-to-draft ratio of B=T = 2:5: º 1 = 1:48.. as N ! 1. ..136 CHAPTER 4.. An e¤ective method to reduce the e¤ects of irregular frequencies is.

Note that º = ! 2 =g > 0.4. The residue of the integrand at k = º is e¡iº(z¡³) .6 Appendices Appendix A: Principle Value Integrals The real and imaginary parts of the principle value integral: Z1 0 PV e¡ik(z¡³) dk º¡k b are used in evaluating some of the kernel and potential integrals. so that: PV Z1 0 e¡ik(z¡³) dk = º¡k b y0 Z1 e¡ik(z¡³) b dk ¡ i¼e¡iº(z¡³) º ¡k b ) (Frank-31) where the path of integration is the positive real axis indented into the upper half plane about k = º.3.10: E¤ect of ”Lid-Method” on Irregular Frequencies 4. THEORY OF FRANK 2000 137 1250 1000 1500 surge/sway Damping (ton/s) Mass (ton) 750 heave 1000 heave 500 500 surge/sway 250 0 0 1 2 3 0 0 1 2 3 Frequency (rad/s Frequency (rad/s ) Figure 4.3. ³ The transformation ! = i (k ¡ º) z ¡ b converts the contour integral on the right hand ³ side of equation (Frank-31) to: Z1 0 y e¡ik(z¡³) b dk = ¡e¡iº(z¡³) º¡k ¡iº(z¡b ³) b ¶ h i µ 2¼i x¡» > 0 for E1 ¡iº(z ¡ b + = ¡e ³) 0 x¡» < 0 n h i (Frank-32) = ¡e¡iº(z¡³) ° + ln ¡iº(z ¡ b ³) ¡iº(z¡b ³) Z1 e¡w dw w . Im fzg < 0 and´Im f³g 6 0.

µ) cos º (x ¡ »)] º ¡k +S (r. µ) sin º (x ¡ ») (Frank-34) PV Z1 0 ek(y+´) sin k (x ¡ ») dk = eº(y+´) [C (r. µ) = ° + ln r + n ¢ n! ¶ 1 X rn sin (nµ) µ µ x¡» > 0 for S (r. 2-D POTENTIAL COEFFICIENTS 1 X (¡1) n=1 n + h in 9 ¡iº(z ¡ b = ³) n ¢ n! . µ) cos º (x ¡ ») provided that: 1 X rn cos (nµ) n=1 C (r. Setting: ¯ ¯ ¯ b¯ r = ¯¡iº(z ¡ ³)¯ n o3 Im ¡iº(z ¡ b ³) o5 + ¼ µ = arctan 4 n b Re ¡iº(z ¡ ³) 2 and the following expression is obtained for equation (Frank-31): Z1 0 PV e¡ik(z¡³) dk = eº(y+´) [cos º (x ¡ ») ¡ i sin º (x ¡ »)] ¢ º¡k # (" 1 X rn cos (nµ) (Frank-33) ¢ ° + ln r + n ¢ n! n=1 #) "µ ¶ X n 1 µ x¡» > 0 r sin (nµ) for +i + µ ¡ 2¼ x¡» < 0 n ¢ n! n=1 b Separating equation (Frank-33) into its real and imaginary parts yields: Z1 0 PV ek(y+´) cos k (x ¡ ») dk = eº(y+´) [C (r.138 CHAPTER 4. µ) sin º (x ¡ »)] º¡k ¡S (r. µ) = µ + + n ¢ n! µ ¡ 2¼ x¡» < 0 n=1 . + µ 2¼i x¡» > 0 for 0 x¡» < 0 ¶ where ° = 0:5772::: is the well known Euler-Mascheroni constant and the value of E1 is de…ned by Abramowitz and Stegun.

Therefore. 0 1 s¡j 0 z=zi Jij = Re 8 > < > : Note that in the complex plane with zi on si : ½h i ~ (z) Re (~ i ¢ r)F n ¾ = Re (· 6 (~ i ¢ r) 4 n ~ 2 sj Z e¡iº(z¡³) ds ¡ (¡1)m b s¡j Z 7 e¡iº(z+³) ds5 ¸ 3 z=zi 9 > = > .4. ds = e¡i®j d³ and: 8 2 3 > > Z < 6 7 Re (~ j ¢ r) 4 ln (z ¡ ³) ds5 n ~ > > : sj 8 9 2 3 > > ³ j+1 > > Z < = d 6 7 ¡i®j i®j = Re ¡ie e d³ ln (z ¡ ³)5 4 > > > dz > > > . so that the indicated di¤erentiation cannot be performed under the integral sign. in that case one may proceed as follows: since: ³ = » + i´ d³ = d» + id´ = ds cos ®j + ids sin ®j = ei®j ds for ³ along the j-th segment. However. ³j z=zj 9 > > = . (Frank-36) z=z1 dF (z) ¡iei®i dz z=z1 ) Considering the term containing ln (z ¡ ³).3. : . it is evident that the kernel integral is singular when i = j. ³j z=zj z=zj 8 9 2 3 > > ³ j+1 > > Z < d = 6 7 = Re ¡i 4 d³ ln (z ¡ ³) d³ 5 > dz > > > : . THEORY OF FRANK Appendix B: Kernel Integrals The in‡uence coe¢cients of equations (Frank-23) are: 2 0 1 Z1 ¡ik(z¡b Z ³ ´ 1 ³) 1 e 6 = Re (~ i ¢ r) 4 @ n ~ ln(z ¡ ³) ¡ ln(z ¡ b + P V ³) dk A ds > 2¼ ¼ º¡k : sj 0 139 Iij (m) 8 > < and: (m) (Frank-35) 9 3 > Z Z1 ¡ik(z+³) = ³ ´ e m @ 1 ln(z + b ¡ ln(z + ³) + 1 P V A ds7 ¡ (¡1) ³) dk 5 > 2¼ ¼ º¡k .

¡ ¢ ¡ ¢ = arg zj ¡ ³ j ¡ arg zj ¡ ³ j+1 = ¼ (Frank-37) If i 6= j. (L2 ) = Re 8 > < > : 2 3 Z ³ ´ 6 7 (~ i ¢ r) 4 ln z ¡ b ds5 n ~ ³ sj (Frank-38) v ¡ ¢ ¢ u ¡ u xi ¡ » j 2 + yi + ´j 2 t¡ = sin (®i + ®j ) ln ¢2 ¡ ¢2 xi ¡ » j+1 + yi + ´j+1 · ¸ yi + ´j yi + ´ j+1 + cos (®i + ®j ) arctan ¡ arctan xi ¡ » j xi ¡ » j+1 The kernel integral containing the principal value integrals is: 8 > < > : 2 Z Z1 0 z=zi 9 > = > . so that: 8 9 2 3 > > Z = < 6 7 ~ 4 ln (z ¡ ³) ds5 (L1 ) = Re (~ i ¢ r) n > > : .140 0 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS Setting ³ = z ¡ ³. ds = ei®j db so that: For the integral containing the ln z ¡ ³ ³. (Frank-39) (L5) = Re = Re 8 > < > : 6 (~ i ¢ r) 4 n ~ ds ¢ P V Z b ³j sj e¡ik(zi ¡³) 7 dk 5 º ¡k Z1 0 b 3 z=zi ¡iei(®i +®j ) = + sin (®i + ®j ) 4P V b ³ j+1 2 d db P V ³ db ³ Z1 0 e ¡ik(zi ¡b ³) º ¡k dk 9 > = > . ¡ ¢ ek(yi +´j ) cos k xi ¡ » j dk º¡k . 9 > = > . sj z=zi v ¢ ¢ ¡ u ¡ u xi ¡ » j 2 + yi ¡ ´ j 2 t¡ = sin (®i ¡ ®j ) ln ¢2 ¡ ¢2 xi ¡ » j+1 + yi ¡ ´ j+1 · ¸ yi ¡ ´ j yi ¡ ´ j+1 + cos (®i ¡ ®j ) arctan ¡ arctan xi ¡ » j xi ¡ » j+1 ³ ´ b term. the last integral becomes: 8 > > < > > : 2 z¡³ j Re ¡i d 6 4 dz z¡³ j+1 Z 0 07 ln ³ d³ 5 3 z=zj 9 > > = > > . di¤erentiation under the integral sign may be performed.

4. respectively.3. (Frank-41) (Frank-42) The integration of the ln(zi ¡ ³) term is straight forward. THEORY OF FRANK Z1 e k(yi +´ j+1 ) 141 cos k xi ¡ » j+1 dk 5 º¡k (Frank-40) ¢ 3 ¡ ¢ 3 ¡P V ¡ cos (®i + ®j ) 4P V 2 0 1 Z 0 ¢ ¡ ek(yi +´j ) sin k xi ¡ » j dk º ¡k e k(yi +´ j+1 ) ¡P V Z1 0 The …rst integral on the right hand side of equation (Frank-36) becomes: 8 > < > : 2 3 Z b 6 7 (~ i ¢ r) 4 e¡iº(z¡³) ds5 n ~ sj sin k xi ¡ » j+1 dk 5 º¡k ¡ (L7) = Re The kernel integrals over the image segments are obtained from equations (Frank-38) through (Frank-41) by replacing » j . » ¡(j+1) = ¡» j+1 and ®¡j = ¡®j . yielding: 8 >Z < > : sj Re 9 > · = ¢ q¡ ¢2 ¡ ¢2 ¡ xi ¡ » j + yi ¡ ´j + » j ¡ » j¡1 ln(zi ¡ ³)ds = + cos ®j xi ¡ » j ln > . » j+1 and ®j with » ¡j = ¡» j . yi ) is: 8 2 3 >Z Z1 ¡ik(zi ¡b N < ³) 1 X e 4ln(zi ¡ ³) ¡ ln(zi ¡ b + 2P V Qj Re ³) ©(m) (xi . . Appendix C: Potential Integrals The velocity potential of the m-th mode of oscillation at the i-th midpoint (xi . j=1 sj s¡j £ º(yi +´ ) ¡ i ¢ ¡ ¢¤ º(yi +´ j+1 ) j cos º x ¡ » = ¡ sin (®i + ®j ) e cos º xi ¡ » j+1 i j ¡e £ ¡ ¢ ¡ ¢¤ + cos (®i + ®j ) eº(yi +´j ) sin º xi ¡ » j ¡ eº(yi +´j+1 ) sin º xi ¡ » j+1 z=z 9 > = > . t) = dk 5 ds > 2¼ j=1 º¡k : sj 0 2 3 9 > Z Z1 ¡ik(zi +³) = e m 4ln(zi + b ¡ ln(zi + ³) + 2P V 5 ds ³) dk ¡ (¡1) > º¡k . yi . s¡j 0 8 9 >Z > Z N < = cos !t X m ¡iº(zi ¡b ³) ¡iº(zi +³) ¨ QN+j Re e ds ¡ (¡1) e ds ¢ > > sin !t : .

respectively. note that ´ j and ´ j+1 are replaced by ¡´ j and ³) ¡´ j+1 . For an arbitrary z in the ‡uid domain: b ³ j+1 b ³j Z ds ¢ P V Z1 0 e ¡ik(z¡b ³) º¡k dk = e i®j ¢ PV Z1 0 dk º¡k b ³ j+1 b ³j Z = e i®j ¢ PV Z1 0 e¡ikz dk º¡k b ³ j+1 b ³j b e¡ik(z¡³) db ³ Z = ¡e i®j ¢ PV Z1 0 e¡ikz eik³ j+1 ¡ eik³ j dk º ¡k k b eik³ db ³ b b where the change of integration is permissible since only one integral requires a principle value interpretation. 2-D POTENTIAL COEFFICIENTS ¡ ¢ q¡ ¢2 ¡ ¢2 ¡ xi ¡ » j+1 ln xi ¡ » j+1 + yi ¡ ´ j+1 ¸ ¡ ¢ ¢ yi ¡ ´ j ¡ yi ¡ ´j+1 ¡ yi ¡ ´j arctan + yi ¡ ´ j+1 arctan xi ¡ » j xi ¡ » j+1 · q¡ ¢ ¢2 ¡ ¢2 ¡ + sin ®j yi ¡ ´ j ln xi ¡ » j + yi ¡ ´ j + ´j ¡ ´j¡1 ¢ q¡ ¢2 ¡ ¢2 ¡ (Frank-43) xi ¡ » j+1 + yi ¡ ´ j+1 ¡ yi ¡ ´j+1 ln ¸ ¡ ¢ ¢ yi ¡ ´ j ¡ yi ¡ ´ j+1 + xi ¡ » j arctan ¡ xi ¡ » j+1 arctan xi ¡ » j xi ¡ » j+1 In the integration of the ln(z ¡ b term. To evaluate the potential integral containing the principal value integral. . the last expression becomes: 21 3 Z Z1 ¡ik(z¡bj+1 ) Z1 ¡ik(z¡bj ) ³ ³ ikbj+1 ³ ikbj ³ ie 4 e ¡e e e e¡ikz ¡ dk + P V dk ¡ P V dk 5 º k º¡k º¡k i®j 0 0 0 (Frank-44) Regarding the …rst integral in equation (Frank-44) as a function of z: F (z) = Z1 0 e ikbj+1 ³ ¡ikz e ¡ eik³ j dk k b (Frank-45) Di¤erentiating equation (Frank-45) with respect to z gives: 81 <Z : 0 F (z) = ¡i 0 e¡ik(z¡³ j+1 ) dk ¡ b Z1 0 e¡ik(z¡³ j ) dk b 9 = . After dividing by º and multiplying by º ¡ k +k under the integral sign.142 CHAPTER 4. proceed in the following manner.

it follows from equation (Frank-46) that º = 0. limz!¡i1 F (z) = 0. (Frank-48) . Since F (z) is de…ned and analytic for all z in the lower half plane and since by equation (Frank-45). sj 0 ½ iei®j h = Re ¡ ln(zi ¡ bj ) ¡ ln(zi ¡ bj+1 ) ³ ³ º 39 Z1 ¡ik(zi ¡bj+1 ) Z1 ¡ik(zi ¡bj ) = ³ ³ e e 5 +P V dk ¡ P V dk . THEORY OF FRANK = So: 1 z ¡ bj ³ ¡ 1 z ¡b ³ 143 j+1 where º is a constant of integration to be determined presently. º ¡k º¡k 0 0 · cos k xi ¡ » j+1 dk ¡ P V º ¡k ¡ ¢ Z1 0 e k(yi +´j ) (K7) = Re = £ ¡ ¢ ¤ £ ¡ ¢ ¤ª 1 © º(yi +´j ) e sin º xi ¡ » j ¡ ®j ¡ eº(yi +´j+1 ) sin º xi ¡ » j+1 ¡ ®j º 9 > = ¡iº(zi ¡b ³) e ds > .3.4. Therefore: 8 >Z < Z1 9 > = ³ ´ ³ ´ F (z) = ln z ¡ bj ¡ ln z ¡ bj+1 + º ³ ³ (Frank-46) (K5) = Re dk > > º¡k : . º¡k º¡k 0 0 v 8 2 u ¡ ¢ ¡ ¢ < u xi ¡ » j 2 + yi + ´ j 2 1 sin ®j 4ln t ¡ = ¢2 ¡ ¢2 º: xi ¡ » j+1 + yi + ´ j+1 +P V Z1 0 ds ¢ P V e ¡ik(zi ¡b ³) (Frank-47) cos k xi ¡ » j dk 5 º¡k ¡ ¢ 3 e k(yi +´ j+1 ) The integration of the potential component in-phase with the velocity over sj gives: 8 >Z < > : sj yi + ´ j yi + ´j+1 + cos ®j arctan ¡ arctan xi ¡ » j xi ¡ » j+1 1 ¢ ¢ 39 ¡ ¡ Z k(yi +´j ) Z1 k(yi +´j+1 ) = e sin k xi ¡ » j e sin k xi ¡ » j+1 5 dk ¡ P V dk +P V .

1974] as a proportion of the total mass of the ship ½r: M11 (S&K) = ® ¢ ½r The factor ® is depending on the breadth-length ratio B=L of the ship: ®= in which: µ ½ ¾ ¶ 1¡b 1+b a= ln ¡ 2b b3 1¡b 2 a 2¡a with: b= s µ ¶2 B 1¡ L With this hydrodynamic mass value. of which a large three-dimensional e¤ect is expected. ¤ ¤ the two-dimensional potential mass M11 and damping N11 can be calculated in an analog manner as has been described for the two-dimensional potential mass and damping of sway. With these two-dimensional values. a correction factor ¯ for three-dimensional e¤ects has been determined: M11 (S&K) ¯= M11 (! = 0) The three-dimensional e¤ects for the potential damping of surge are ignored. 1961] the zero-frequency potential mass for sway can be expressed in Lewis-coe¢cients: 0 M22 (! = 0) = ¼ ½ 2 µ dx 1 ¡ a1 + a3 ¶2 When using this formula for surge. 0 0 M22 and N22 . These frequency-dependent hydrodynamic coe¢cients do not include three-dimensional e¤ects. Only the hydrodynamic mass coe¢cient. the total potential mass and damping of surge are de…ned by: ¤ M11 = B ¢ M11 ¤ N11 = B ¢ N11 in which B is the breadth of the ship. being constant over the ship’s breadth B.4 Surge Coe¢cients sectional breadth Bx sectional draught dx sectional area coe¢cient CMx = ship length L = midship draught d = block coe¢cient CB An equivalent longitudinal section. . 2-D POTENTIAL COEFFICIENTS 4. will be adapted here empirically. the total potential mass of surge is de…ned by: M11 (! = 0) = ¤ B ¢ M11 (! = 0) ¡ ¢ (1 ¡ a1 )2 + 3a2 3 A frequency-independent total hydrodynamic mass coe¢cient is estimated empirically by [Sargent and Kaplan. is de…ned by: By using a Lewis transformation of this equivalent longitudinal section to the unit circle. According to [Tasai.144 CHAPTER 4.

SURGE COEFFICIENTS 145 Figure 4.4. Based on the results of numerical 3-D studies with a Wigley hull form.11: Surge Hydrodynamic Mass So. a proportionality of both the two-dimensional hydrodynamic mass and damping with the absolute values of the derivatives of the cross sectional areas Ax in the xb -direction is assumed: 0 M11 =R L x j dAb jdxb dx x j dAb j dx ¢ M11 0 N11 =R L x j dAb jdxb dx x j dAb j dx ¢ N11 . the cross sectional twodimensional values of the hydrodynamic mass and damping have to be obtained.4. the potential mass and damping of surge are de…ned by: ¤ M11 = B ¢ M11 ¢ ¯ ¤ N11 = B ¢ N11 To obtain a uniform approach during all ship motions calculations.

5 1. ¾ s .5 0 0.Ursell-Tasai’s method with 2-parameter Lewis conformal mapping.0 1.12 compares the calculated coe¢cients for an amidships cross section of a container vessel with the three previous methods: .5 2.5 2.0 1. .0 2.5 2.5 0 Sway .0 1.0 2.Roll Roll .5 frequency (rad/s) frequency (rad/s) frequency (rad/s) frequency (rad/s) Figure 4.0 2.Roll Roll .0 1.12: Comparison of Calculated Coe¢cients With the exception of the roll motions. 300 Mass Coefficient Sway ' M2 2 500 400 300 200 100 5000 Heave ' M 33 4000 3000 Midship section of a containership 2000 1000 0 Roll ' M 44 4250 4000 3750 3500 3250 200 100 Sway . A disadvantage of Frank’s method can be the large computing time. .5 0 0.5 200 Damping Coefficient 150 100 Sway ' N 22 150 Heave ' N 33 300 Frank Close-fit Roll ' N 44 200 150 100 100 200 50 50 0 100 Lewis 50 0 0 0. The roll motion deviation.5 2.5 1.5 1.5 1. Frank’s pulsating source method should be used.146 CHAPTER 4. less than 0.5 2.0 1.5 3000 0 0. when compared with Ursell-Tasai’s method. Generally.5 0 0 0. the three results are very close.0 1.5 1.5 2.0 1.5 2.5 1.5 Comparative Results Figure 4. 2-D POTENTIAL COEFFICIENTS 4.0 1.0 2.Frank’s pulsating source method.5 2.5 1. For submerged sections.0 2. is caused by the description of the ”bilge” by the simple Lewis transformation.Ursell-Tasai’s method with 10-parameter close-…t conformal mapping. . it is advised to use Ursell-Tasai’s method with 10 parameter close-…t conformal mapping. bulbous sections and sections with an area coe¢cient.Sway ' ' N 24 = N 42 0 0.5.0 2.5 1.5 0 0 0.Sway ' ' M 24 = M 42 0 0 0.0 2. predicted with the Lewis conformal mapping method.0 2.

For updates see web site: http://dutw189. This total resistance coe¢cient Ct is given in …gure 5. Ship Hydromechanics Laboratory. February 2001. Mekelweg 2. can be determined simply from the resistance-speed curve of the ship in still water. and an additional viscous part. 5.M. Report 1216a. in principle valid at the ship’s service speed for hull forms with a block coe¢cient CB between 0. Because of these additional contributions to the damping are from a viscous origin mainly. Rsw (V ): d fRsw (V )g B11t = B11 + B11v = dV 5. viscous roll damping e¤ects can be accounted for by empirical formulas. Delft University of Technology. This holds that viscous e¤ects are neglected. In practice. At forward ship speed V . 2628 CD Delft. J. 1955].19”. 0 147 .60 and 0.1 Total Surge Damping For a rough estimation of the still water resistance use can be made of a modi…ed empiric formula of [Troost.1 Surge Damping The total damping for surge B11t = B11 + B11v consists of a potential part. which can deliver serious problems when predicting roll motions at resonance frequencies. The Netherlands. B11t . B11v .1 as a function of the ship length. ”Theoretical Manual of SEAWAY. B11 . the total damping coe¢cient.1.Chapter 5 Viscous Damping The strip theory is based on the potential ‡ow theory.nl/~johan or http://www. additional damping coe¢cients have to be introduced.80: Rsw = Ct ¢ ½r 3 V 2 with: Ct ¼ 0:0036 + in which: 0:0152 log10 fLg + 0:60 (with L in meter) 2 r = volume of displacement of the ship in m3 L = length of the ship in m V = forward ship speed in m/s. For surge and roll.wbmt.J. Journée. Release 4. it is not possible to calculate the total damping in a pure theoretical way.shipmotions.tudelft.nl.

the uncoupled linear equation of the roll motion about the centre of gravity G is given by: Ä _ (Ixx + A44 ) ¢ Á + (B44 + B44v ) ¢ Á + C44 ¢ Á = 0 . which can be derived from the frictional part of the ship’s resistance.2 Viscous Surge Damping This total damping coe¢cient includes a viscous part.1. d fRf (V )g dV 5. VISCOUS DAMPING Figure 5.148 CHAPTER 5. de…ned by the ITTC-line: 1 0:075 Rf (V ) = ½V 2 S ¢ 2 (lnfRng ¡ 2)2 with: Rn = VL º in which: = kinematic viscosity of seawater º = wetted surface of the hull of the ship S Rn = Reynolds number From this empiric formula follows the pure viscous part of the additional damping coe¢cient at forward ship speed V : B11v = which can be obtained numerically.1: Total Still Water Resistance Coe¢cient of Troost Then the total surge damping coe¢cient at forward ship speed V becomes: B11t = 2Ct ¢ ½r 3 V 2 5.2 Roll Damping In case of pure free rolling in still water (free decay test).

·. the equation of motion can be re-written as: Ä _ Á + 2·! 0 ¢ Á + ! 2 ¢ Á = 0 0 B44 + B44v (quotient of damping and moment of inertia) Ixx + A44 C44 (natural roll frequency squared) !2 = 0 Ixx + A44 The non-dimensional roll damping coe¢cient. one can neglect º 2 here and use ! Á ¼ ! 0 .5.2. ROLL DAMPING with: B44 = b44 + OG ¢ b42 + OG ¢ b24 + OG ¢ b22 B44V = b44V C44 = ½gr ¢ GM a42 = a24 and A44 = a44 + OG ¢ a42 + OG ¢ a24 + OG ¢ a22 2 2 149 (potential mass coe¢cient) (potential damping coe¢ciernt) (viscous damping coe¢cient) (restoring term coe¢cient) b42 = b24 while for zero forward speed: These equations can be rewritten as: Ä _ Á + 2º ¢ Á + ! 2 ¢ Á = 0 0 in which: This damping coe¢cient is written as a fraction between the actual damping coe¢cient. B44cr = 2 (Ixx + A44 ) ¢ C44 . Áa . p B44 + B44v . and the critical damping coe¢cient. so for critical damping: ·cr = 1. is given by: º B44 + B44v ·= = p ! 0 2 (Ixx + A44 ) ¢ C44 2º = Suppose the vessel is de‡ected to an initial heel angle. in still water and then released. this leads to: 0 The non-dimensional total roll damping is given now by: ½ ¾ 1 Á(t) !0 ·= ln = (B44 + B44v ) ¢ 2¼ Á(t + TÁ ) 2C44 The non-potential part of the total roll damping coe¢cient follows from the average value of · by: 2C44 ¡ B44 B44v = · ¢ !0 ! 0 TÁ ¼ ! Á TÁ = 2¼ . Herewith. The solution of the equation of motion of this decay becomes: µ ¶ º ¡ºt cos ! Á t + Á = Áa e sin ! Á t !Á Then the logarithmic decrement of the motion is: ½ ¾ Á(t) ºTÁ = ·! 0 TÁ = ln Á(t + TÁ ) Because ! 2 = ! 2 ¡ º 2 for the natural frequency oscillation and the damping is small so 0 Á that º 2 ¿ ! 2 .

see …gure 5.1 Experimental Determination The ·-values can easily been found when results of free rolling experiments with a model in still water are available.2. double amplitudes can be used instead: ( ) ¯ ¯ ¯ Áai ¡ Áai+1 + Áai+2 ¡ Áai+3 ¯ Áai ¡ Áai+1 1 ¯ ¯ versus Áa = ¯ ¢ ln ·= ¯ 2¼ Áai+2 ¡ Áai+3 4 ² Sometimes the results of free rolling tests are presented by: ¢Áa Áa versus Áa with the absolute value of the average of two successive positive or negative maximum roll angles. given by: ¯ ¯ ¢Áa = ¯Áai ¡ Áai+1 ¯ .2: Time History of a Roll Decay Test The results of free decay tests can be presented in di¤erent ways: ² Generally they are presented by plotting the non-dimensional damping coe¢cient. caused by a possible zero-shift of the measuring signal. by: ( ) ¯ ¯ ¯ Áai + Áai+2 ¯ Áai 1 ¯ versus Áa = ¯ ¢ ln ·= ¯ ¯ 2¼ Á 2 ai+2 ² To avoid spreading in the successively determined ·-values. obtained from two successive positive or negative maximum roll angles Áai and Áai+2 . Figure 5. VISCOUS DAMPING 5.2. given by: ¯ ¯ ¯ Áai + Áai+1 ¯ ¯ ¯ Áa = ¯ ¯ 2 and the absolute value of the di¤erence of the average of two successive positive or negative maximum roll angles.150 CHAPTER 5.

Luckily. ROLL DAMPING Then the total non-dimensional roll damping coe¢cient becomes: 9 8 < 2 + ¢Áa = 1 Áa ·= ¢ ln ¢Áa .2. When the damping is very small and the oscillation decays very slowly. Figure 5. this damping coe¢cient is not valid for the actual coupled motions of a ship which will be moving in all directions simultaneously. this is impossible since decay tests are carried out at only one frequency: the natural frequency. it will be found: · = ·1 + ·2 ¢ Áa . several estimates of the decay can be obtained from a single record. Note that these decay tests provide no information about the relation between the potential coe¢cients and the frequency of oscillation. The method is not really practical when º is much greater than about 0. will often have a non-linear behavior as illustrated in the next …gure. Strictly. Indeed. The successively found values for ·. :2 ¡ 2¼ Áa 151 The decay coe¢cient · can therefore be estimated from the decaying oscillation by determining the ratio between any pair of successive (double) amplitudes.5. Be aware that this damping coe¢cient is determined by assuming an uncoupled roll motion (no other motions involved). plotted on base of the average roll amplitude. It is obvious that for a linear system a constant ·-value should be found in relation to Áa .2 and is in any case strictly valid for small values of º only. this is generally the case.3: Roll Damping Coe¢cient For a behavior like this. These experiments deliver no information on the relation with the frequency of oscillation.

the non-dimensional total roll damping coe¢cient. has to be added to this a formula.it is found by [Journée. 1991]: · = ·1 + ·2 ¢ Áa with: ·1 ·2 µ ¶2 B = 0:0013 ¢ d = 0:50 in which B is the breadth and d is the draft of the barge. relative to water plane) length of ship (m) breadth of ship (m) draft of ship (m) block coe¢cient (-) initial metacentric height (m) Froude number (-) amplitude of roll (rad) correction factor on ·1 for speed e¤ect (-) (in the original formulation of Miller: CV = 1:0) ·2 in which: Abk lbk hbk rb L B d Cb GM Fn Áa CV . 5. ³ a . the damping term is depending on the .152 CHAPTER 5.2.so far unknown . 5.solution for the transfer function of roll: Áa =³ a . This non-linear behavior holds that during frequency domain calculations. can be obtained by: p · = ·1 + ·2 ¢ Áa with: ·1 (µ ¶ µ ¶ r µ ¶3 ) 2 L Fn Fn L Fn = CV ¢ 0:00085 ¢ ¢ ¢ + +2¢ B Cb Cb Cb GM ( ) r 3 rb lbk = 19:25 ¢ Abk ¢ + 0:0024 ¢ L ¢ B ¢ rb L ¢ B 3 ¢ d ¢ Cb = = = = = = = = = = = = lbk ¢ hbk = one sided area of bilge keel (m2 ) length of bilge keel (m) height of bilge keel (m) distance center line of water plane to turn of bilge (m) (…rst point at which turn of bilge starts.2.3 Empirical Method of Miller According to [Miller.with its center of gravity. VISCOUS DAMPING while sometimes even a cubic roll damping coe¢cient. G. With a known wave amplitude. A less accurate method is to use a …xed Áa . in the water line .2 Empirical Formula for Barges From model experiments with rectangular barges . 1974]. ·3 ¢ Á2 . this problem can be solved in an iterative manner. ·.

ROLL DAMPING 153 Generally. This coe¢cient can be obtained by stipulating that an equivalent linear roll damping dissipates an identical amount of energy as the non-linear (2) roll damping. For numerical reasons three restrictions have been made here during the calculations: . B44V . into: (1) T T ZÁ ZÁ (2) _ _ _ _ _ ¢ Á ¢ Ádt = B44V ¢ jÁjÁ ¢ Ádt B44V or: 0 0 8 (2) Áa ! ¢ B44V 3¼ For the estimation of the non-potential parts of the roll damping. an equivalent linear roll damping coe¢(1) cient.5. has to be determined. They conclude that the method can be used safely for ordinary ship forms. which conclusion has been con…rmed by the author too. also experiments showed a non-linear behavior of viscous parts of the roll damping. A few subordinate parts have been modi…ed and this empiric method is called here the ”Ikeda method”. this component will be set to zero.if a calculated component of the viscous roll damping coe¢cient becomes less than zero. for applications in frequency domain.4 Semi-Empirical Method of Ikeda Because the viscous part of the roll damping acts upon the viscosity of the ‡uid signi…cantly. local. use has been made of work published by [Ikeda et al.2. But for unusual ship forms. local.2. Himeno and Tanaka claim fairly good agreements between their prediction method and experimental results. for very full ship forms and for ships with a very large breadth to draft ratio the method is not always accurate su¢ciently. 1978]. a suitable value for CV seems to be: q CV = 4:85 ¡ 3:00 ¢ GM F ull Scale 5. like frigates. This results for a linearized quadratic roll damping coe¢cient.. Besides this. Sometimes.mainly viscous . CV = 1:0 but for slender ships. it is not possible to calculate the total roll damping coe¢cient in a pure theoretical way. B44V . The Ikeda method estimates the following linear components of the roll damping coe¢cient of a ship: B44V = (1) B44S B44F B44E B44L B44K = = = = = a correction on the potential roll damping coe¢cient due to forward speed the frictional roll damping coe¢cient the eddy making roll damping coe¢cient the lift roll damping coe¢cient the bilge keel roll damping coe¢cient So. ¾ s > 0:999 then ¾ s = 0:999 . .roll damping coe¢cient B44V is given by: B44V = B44S + B44F + B44E + B44L + B44K Ikeda. OG < ¡Ds ¾ s then OG = ¡Ds ¾ s .if.if. the additional .

With this. B44S Ikeda obtained an empirical formula for the three-dimensional forward speed correction on the zero speed potential damping by making use of the general characteristics of a doublet ‡ow model.154 Notation CHAPTER 5. VISCOUS DAMPING In this description of the Ikeda method the notation of the authors (Ikeda. The rolling ship has been represented by two doublets: one at the stern and one at the bow of the ship. semi-theoretically the forward speed e¤ect on the linear potential damping coe¢cient has been approximated as a fraction of the potential damping coe¢cient by: n h i o ¡150(­¡0:25)2 ¡ 1:0 B44S = B44 ¢ 0:5 A2 + 1 + (A2 ¡ 1) tanh (20(­ ¡ 0:3)) + (2A1 ¡ A2 ¡ 1)e B44 = ­ = »D = A1 = A2 = potential roll damping coe¢cient of the ship (about G) V ! = non-dimensional circular roll frequency g D ! 2 = non-dimensional circular roll frequency squared g 1:0 + » ¡1:2 e¡2»D = maximum value of B44 at ­ = 0:25 D 0:5 + » ¡1:0 e¡2»D = minimum value of B44 at large ­ D with: . Himeno and Tanaka) is maintained as far as possible: ½ º g V Rn ! Áa L B D CM CB SL Sf OG = = = = = = = = = = = = = = = density of water kinematic viscosity of water acceleration of gravity forward ship speed Reynolds number circular roll frequency roll amplitude length of the ship breadth amidships draft amidships section coe¢cient block coe¢cient L ¢ D ¼ lateral area wetted hull surface area distance of centre of gravity above still water level. so positive upwards (sign convention deviates from the paper of Ikeda) = sectional breadth water line = sectional draft = sectional area ¾s H0 a1 a3 Ms rf LO = = = = = = = sectional area coe¢cient sectional half breadth to draft ratio sectional Lewis coe¢cient sectional Lewis coe¢cient sectional Lewis scale factor average distance between roll axis and hull surface distance point of taking representative angle of attack to roll axis approximated by LO = 0:3D distance of centre of action of lift force in roll motion to roll axis approximated by LR = 0:5D height of the bilge keels length of the bilge keels distance between roll axis and bilge keel correction for increase of ‡ow velocity at the bilge pressure coe¢cient lever of the moment local radius of the bilge circle LR = hk Lk rk fk Cp lm rb = = = = = = = Bs Ds As E¤ect of Forward Speed.

for ship forms the average distance between the roll axis and the hull surface can be approximated by: Sf (0:887 + 0:145 ¢ CB ) L + 2OG rf = ¼ with a wetted hull surface area Sf .4.4: Relation Between Density. wholly immersed in the ‡uid.5. An e¤ective Reynolds number of the roll motion was de…ned by: Rn = 0:512 ¢ (rf Áa )2 ! º In here. B44F 155 Kato deduced semi-empirical formulas for the frictional roll damping from experimental results of circular cylinders.0 Sea Water 1020 Kinematic Viscosity (m s) 1. ROLL DAMPING Frictional Roll Damping.5 0 10 20 0 30 Temperature ( C) Temperature ( C) Figure 5. approximated by: Sf = L(1:7D + CB B) The relation between the density.5.2.0 1000 Fresh Water 990 0 10 20 0 30 0.5 Sea Water Density (kg/m ) 3 1010 2 Fresh Water 1. kinematic viscosity and temperture of fresh water and sea water are given in …gure 5. the kinematic viscosity can be expressed into the density of water by the following relation in the kg-m-s system: º ¢ 106 = 1:442 + 0:3924 ¢ (½ ¡ 1000) + 0:07424 ¢ (½ ¡ 1000)2 m2 /s (fresh water) º ¢ 106 = 1:063 + 0:1039 ¢ (½ ¡ 1025) + 0:02602 ¢ (½ ¡ 1025)2 m2 /s (salt water) as given in …gure 5. . 1030 2. Kinematic Viscosity and Temperature of Water When eliminating the temperature of water.

156 Fresh Water 25 CHAPTER 5. Tamiya deduced a modi…cation coe¢cient for the e¤ect of forward speed on the friction component. Using this. Accurate enough from a practical point of view. the quadratic roll damping coe¢cient due to skin friction at zero forward ship speed is expressed as: 1 3 (2) B44F = ½rf Sf Cf 0 2 This frictional roll damping component increases slightly with forward speed. but has been included here.5: Kinematic Viscosity as a Function of Density Kato expressed the skin friction coe¢cient as: Cf = 1:328 ¢ Rn¡0:5 + 0:014 ¢ Rn¡0:114 The …rst part in this expression represents the laminar ‡ow case. The second part has been ignored by Ikeda. this results into the following formula for the speed dependent frictional damping coe¢cient: ¶ µ 1 3 V (2) B44F = ½rf Sf Cf ¢ 1:0 + 4:1 ¢ 2 !L Then. the equivalent linear roll damping coe¢cient due to skin friction is expressed as: ¶ µ 8 1 3 V B44F = Á ! ¢ ½rf Sf Cf ¢ 1:0 + 4:1 ¢ 3¼ a 2 !L Ikeda con…rmed the use of his formula for the three-dimensional turbulent boundary layer over the hull of an oscillating ellipsoid in roll motion. Semitheoretically. . VISCOUS DAMPING Salt Water 25 20 20 Kinematic Viscosity *10 (m s) 0 Temperature ( C) 15 Kinematic Viscosity *10 (m s) 0 Temperature ( C) 2 7 7 2 15 10 10 5 Viscosity Actual Viscosity Polynomial Temperature 0 997 998 999 3 5 Viscosity Actual Viscosity Polynomial Temperature 1000 0 1023 1024 1025 1026 1027 3 1028 Density Fresh Water (kg/m ) Density Salt Water (kg/m ) Figure 5.

generated by a two-dimensional separation.2. These experimental results are …tted by: Cp = 0:35 ¢ e¡° ¡ 2:0 ¢ e¡0:187¢° + 1:5 The value of ° around a cross section is approximated by the potential ‡ow theory for a rotating Lewis form cylinder in an in…nite ‡uid. ROLL DAMPING Eddy Making Damping. Values of rmax (Ã) have to be calculated for: à = à 1 = 0:0 The values of rmax (Ã) follow from: and à = Ã2 = ³ 0:5 a1 (1+a3 ) 4a3 cos ´ With these two results a value rmax and a value à follow from the conditions: . This means that the corresponding quadratic roll damping coe¢cient does not depend on the period parameter but on the hull form only. has to be made. B44E 157 At zero forward speed the eddy making roll damping for the naked hull is mainly caused by vortices. When using a simple form for the pressure distribution on the hull surface it appears that the pressure coe¢cient Cp is a function of the ratio ° of the maximum relative velocity Umax to the mean velocity Umean on the hull surface: °= Umax Umean The Cp -° relation was obtained from experimental roll damping data of two-dimensional models. An estimation of the local maximum distance between the roll axis and the hull surface. From a number of experiments with two-dimensional cylinders it was found that for a naked hull this component of the roll moment is proportional to the roll frequency squared and the roll amplitude squared. rmax .for rmax (à 1 ) > rmax (à 2 ) : rmax = rmax (à 1 ) and à = à 1 .for rmax (à 1 ) < rmax (à 2 ) : rmax = rmax (à 2 ) and à = à 2 The relative velocity ratio ° on a cross section is obtained by: µ ¶ p f3 ¼ 2Ms p 2 2 ³ ´p °= ¢ rmax + a +b H 2Ds ¾ s + OG H0 ¾ s Ds with: H0 = ¾s Bs 2Ds As = Bs Ds Bs Ds = 2(1 + a1 + a3 ) 1 ¡ a1 + a3 © rmax (Ã) = Ms ((1 + a1 ) sin(Ã) ¡ a3 sin(3Ã))2 + ª1=2 ((1 ¡ a1 ) cos(Ã) + a3 cos(3Ã))2 Ms = .5.

With this the equivalent linear eddy making damping coe¢cient at forward speed is given by: 8 1 (2) B44E = ¢Áa !B44e0 ¢ 3¼ 1 + K2 with: V K= 0:04 ¢ !L Lift Damping.158 CHAPTER 5. VISCOUS DAMPING H = 1 + a2 + 9a2 + 2a1 (1 ¡ 3a3 ) cos(2Ã) ¡ 6a3 cos(4Ã) 1 3 ¡ ¢ a = ¡2a3 cos(5Ã) + a1 (1 ¡ a3 ) cos(3Ã) + (6 ¡ 3a1 )a2 + (a2 ¡ 3a1 )a3 + a2 cos(Ã) 3 1 1 ¡ ¢ b = ¡2a3 sin(5Ã) + a1 (1 ¡ a3 ) sin(3Ã) + (6 + 3a1 )a2 + (a2 + 3a1 )a3 + a2 sin(Ã) 3 1 1 f3 = 1 + 4e¡1:65¢10 5 (1¡¾ 2 s) With this a quadratic sectional eddy making damping coe¢cient for zero forward speed follows from: µ ¶2 1 4 rmax (2) 0 B44E = ½D Cp ¢ 0 2 s Ds õ ¶µ ¶ µ ¶2 ! OG f1 rb f1 rb f1 rb 1¡ 1+ ¡ + f2 H0 ¡ Ds Ds Ds Ds with: f1 = 0:5 ¢ (1 + tanh(20¾ s ¡ 14)) ¡ ¢ f2 = 0:5 ¢ 1 (¡ cos(¼¾ s )) ¡ 1:5 ¢ 1 ¡ e5¡5¾s sin2 (¼¾ s ) The approximations of the local radius of the bilge circle rb are given as: r Bs H0 (¾ s ¡ 1) for: rb < Ds and rb < rb = 2Ds ¼¡4 2 for: H0 > 1 and rb > Ds rb = Ds Bs for: H0 < 1 and rb > H0 Ds rb = 2 For three-dimensional ship forms the zero forward speed eddy making quadratic roll damping coe¢cient is found by an integration over the ship length: Z (2) (2) 0 B44E = B44E dxb 0 0 L This eddy making roll damping decreases rapidly with the forward speed to a non-linear correction for the lift force on a ship with a small angle of attack. Ikeda has analyzed this forward speed e¤ect by experiments and the result has been given in an empirical formula. B44L The roll damping coe¢cient due to the lift force is described by a modi…ed formula of Yumuro: à ! 2 1 OG OG B44L = ½SL V kN LO LR 1:0 + 1:4 ¢ + 0:7 ¢ 2 LR LO LR .

ROLL DAMPING The slope of the lift curve CL =® is de…ned by: kN = CL ® µ ¶ B 2¼D = +  4:1 ¢ ¡ 0:045 L L 159 in which the coe¢cient  is given by Ikeda in relation to the amidships section coe¢cient CM : CM < 0:92: 0:92 < CM < 0:97: 0:97 < CM < 0:99: These data are …tted here by:  = 106 ¢ (CM ¡ 0:91)2 ¡ 700 ¢ (CM ¡ 0:91)3 with the restrictions: -if CM < 0:91 then  = 0:00 -if CM > 1:00 then  = 0:35 Bilge Keel Damping.5. created by the bilge keels. Ikeda measured this non-linear drag also by carrying out free rolling experiments with an ellipsoid with and without bilge keels.a component B44K due to the normal force on the bilge keels N  = 0:00  = 0:10  = 0:30 .a component (2) B44K S due to the pressure on the hull surface.2. The drag coe¢cient CD depends on the period parameter or the Keulegan-Carpenter number. . (2) The coe¢cient of the normal force component B44K of the bilge keel damping can be N deduced from experimental results of oscillating ‡at plates. This resulted in a quadratic sectional damping coe¢cient: B44K with: CD = 22:5 ¢ fk hk + 2:40 ¼rk Áa fk = 1:0 + 0:3 ¢ e¡160¢(1:0¡¾s ) (2) 0 N 3 2 = ½rk hk fk ¢ CD The approximation of the local distance between the roll axis and the bilge keel rk is given as: s µ µ ¶ ¶2 rb 2 OG rb + 1:0 + ¡ 0:293 ¢ rk = Ds H0 ¡ 0:293 ¢ Ds Ds Ds The approximation of the local radius of the bilge circle rb in here is given before. B44K The quadratic bilge keel roll damping coe¢cient is divided into two components: (2) .

Ikeda de…nes an equivalent length of a constant negative pressure region S0 over the height of the bilge keels. a quadratic sectional roll damping coe¢cient can be de…ned: (2) 0 B44K S 1 2 2 = ½rk fk 2 h Zk Cp lm dh 0 Ikeda carried out experiments to measure the pressure on the hull surface created by bilge + keels. and on the back face of ¡ the bilge keel. are given by: + Cp = 1:20 and ¡ Cp = ¡22:5 ¢ hk ¡ 1:20 ¼fk rk Áa and the sectional pressure moment is given by: h Zk 0 ¡ ¢ ¡ 2 + Cp lm ¢ dh = Ds ¡A ¢ Cp + B ¢ Cp with: A = (m3 + m4 )m8 ¡ m2 7 m3 (1 ¡ m1 )2 (2m3 ¡ m2 ) 4 + + m1 (m3 m5 + m4 m6 ) B = 3(H0 ¡ 0:215 ¢ m1 ) 6(1 ¡ 0:215 ¢ m1 ) rb m1 = Ds ¡OG m2 = Ds m3 = 1:0 ¡ m1 ¡ m2 m4 = H0 ¡ m1 0:414 ¢ H0 + 0:0651 ¢ m2 ¡ (0:382 ¢ H0 + 0:0106)m1 1 m5 = (H0 ¡ 0:215 ¢ m1 )(1 ¡ 0:215 ¢ m1 ) 0:414 ¢ H0 + 0:0651 ¢ m2 ¡ (0:382 + 0:0106 ¢ H0 )m1 1 m6 = (H0 ¡ 0:215 ¢ m1 )(1 ¡ 0:215 ¢ m1 ) S0 for: S0 > 0:25 ¢ ¼rb m7 = ¡ 0:25 ¢ ¼m1 Ds for: S0 < 0:25 ¢ ¼rb = 0:0 for: S0 > 0:25 ¢ ¼rb m8 = m7 + 0:414 ¢ m1 µ µ ¶¶ S0 for: S0 < 0:25 ¢ ¼rb = m7 + 1:414 ¢ m1 1 ¡ cos rb . Cp . He found that the coe¢cient Cp the pressure on the front face of the bilge keel does ¡ not depend on the period parameter. Cp . while the coe¢cient Cp of the pressure on the back face of the bilge keel and the length of the negative pressure region depend on the period parameter. VISCOUS DAMPING Assuming a pressure distribution on the hull caused by the bilge keels. which is …tted to the following empirical formula: S0 = 0:3 ¢ ¼fk rk Áa + 1:95 ¢ hk + The pressure coe¢cients on the front face of the bilge keel.160 CHAPTER 5.

the frictional part of the roll damping is expected to be smaller than showed above. Calculated Roll Damping Components In …gure 5. Figure 5.2. . for the S-175 container ship design.6. an example is given of the several roll damping components.6: Roll Damping Coe¢cients of Ikeda It may be noted that for full scale ships. because of the higher Reynolds number. ROLL DAMPING 161 The equivalent linear total bilge keel damping coe¢cient can be obtained now by integrating the two sectional roll damping coe¢cients over the length of the bilge keels and linearizing the result: Z ³ ´ 8 (2) 0 (2) 0 B44K = Áa ! B44K + B44K dxb N S 3¼ Lk Experiments of Ikeda have shown that the e¤ect of forward ship speed on this roll damping coe¢cient can be ignored.5. as derived with Ikeda’s method.

162 . CHAPTER 5. VISCOUS DAMPING .

nl/~johan or http://www. When ignoring these outlined terms the ”Ordinary Strip Theory” is presented. as published for instance by [Tasai. 1957] and others.1) Xhj = Mjj ¢ ³ hj + Njj ¢ ³ hj + XRSj Dt In the ”Modi…ed Strip Theory”. Journée. yb ).shipmotions. equivalent longitudinal motions J.nl. 6.Chapter 6 Hydromechanical Loads With an approach as mentioned before. February 2001.wbmt. a description will be given here of the determination of the hydromechanical forces and moments for all six modes of motions. 2628 CD Delft. 1969] and others. The Netherlands. Mekelweg 2. Mjj and Njj are the two-dimensional potential mass and damping and the non-di¤raction part XRSj is the two-dimensional quasi-static restoring spring term. yb .J. For updates see web site: http://dutw189. zb ) axes system with the centre of gravity G in the still water level. Report 1216a. ³ ¤ j is the harmonic h oscillatory motion. ”Theoretical Manual of SEAWAY. as published by [Korvin-Kroukovsky and Jacobs. parallel to (xb . 0 163 . Some of the terms in the hydromechanical loads are outlined there.tudelft. these loads become: ½µ ¶ ¾ D i 0 ¤ 0 0 0 _ Xhj = Mjj ¡ Njj ¢ ³ hj + XRSj Dt !e In these de…nitions of the two-dimensional hydromechanical loads.M. the uncoupled two-dimensional potential hydromechanical loads in the direction j are de…ned by: D n 0 _¤ o 0 0 0 _¤ (6.19”. The ”Modi…ed Strip Theory” includes these outlined terms. At the following pages. Ship Hydromechanics Laboratory.1 Hydromechanical Forces for Surge The hydromechanical forces for surge are found by an integration over the ship length of the two-dimensional values: Z 0 Xh1 = Xh1 ¢ dxb L When assuming that the cross sectional hydromechanical force hold at a plane through the local centroid of the cross section b. so OG = 0. Delft University of Technology. In the ”Ordinary Strip Theory”. Release 4. the hydromechanical loads are calculated in the G(xb .

164 CHAPTER 6. are de…ned by: ³ ¤ 1 = ¡x + bG ¢ µ h @bG _¤ _ ³ h1 = ¡x + bG ¢ µ ¡ V _ ¢µ @xb _ ¼ ¡x + bG ¢ µ _ @bG _ @ 2 bG Ĥ x µ ¢µ+V2 ¢µ ³ h1 = ¡Ä + bG ¢ Ä ¡ 2V @xb @x2 b Ä ¼ ¡Ä + bG ¢ µ x In here. According to the ”Ordinary Strip Theory” the two-dimensional potential hydromechanical force on a surging cross section in still water is de…ned by: D n 0 _¤ o 0 0 _¤ Xh1 = M11 ¢ ³ h1 + N11 ¢ ³ h1 Dt µ ¶ 0 0 Ĥ + N 0 ¡ V ¢ dM11 ¢ ³ ¤ _ = M11 ¢ ³ h1 11 h1 dxb According to the ”Modi…ed Strip Theory” this hydromechanical force becomes: ½µ ¶ ¾ D i 0 0 0 _¤ M11 ¡ Xh1 = ¢ N11 ¢ ³ h1 Dt !e µ ¶ µ ¶ 0 0 V dN11 0 Ĥ + N 0 ¡ V ¢ dM11 ¢ ³ ¤ _ = M11 + 2 ¢ ¢ ³ h1 11 h1 ! e dxb dxb This results into the following coupled surge equation: ½r ¢ x ¡ Xh1 = Ä (½r + a11 ) ¢ x + b11 ¢ x + c11 ¢ x Ä _ +a13 ¢ z + b13 ¢ z + c13 ¢ z Ä _ Ä + b15 ¢ µ + c15 ¢ µ _ +a15 ¢ µ Xw1 = with: a11 = + Z L 0 M11 ¢ dxb b11 ¯ ¯ ¯ Z ¯ 0 ¯V ¯ dN11 +¯ 2 ¢ dxb ¯ ¯! ¯ dxb ¯ e ¯ L ¶ Z µ 0 dM11 0 = + N11 ¡ V ¢ ¢ dxb + b11V dxb L c11 = 0 a13 = 0 . bG is the vertical distance of the centre of gravity of the ship G above the centroid b of the local submerged sectional area. relative to the cross section of an oscillating ship in still water. HYDROMECHANICAL LOADS of the water particles.

2 Hydromechanical Forces for Sway The hydromechanical forces for sway are found by an integration over the ship length of the two-dimensional values: Z 0 Xh2 = Xh2 ¢ dxb L . the expressions for the total hydromechanical coe¢cients in the coupled surge equation become: Z 0 a11 = + M11 ¢ dxb b11 = + c11 a13 b13 c13 a15 = = = = L Z L 0 N11 ¢ dxb + b11V 0 0 0 0 Z 0 = ¡ M11 ¢ bG ¢ dxb L b15 ¯ ¯ ¯ ¯ Z 0 ¯ ¡V ¯ dN11 +¯ 2 ¢ bG ¢ dxb ¯ ¯! ¯ dxb ¯ e ¯ L ¶ Z µ 0 dM11 0 = ¡ N11 ¡ V ¢ ¢ bG ¢ dxb ¡ b11V ¢ BG dxb L c15 = 0 6. derived from the still water resistance approximation of [Troost.2. HYDROMECHANICAL FORCES FOR SWAY b13 = 0 c13 = 0 Z 0 a15 = ¡ M11 ¢ bG ¢ dxb L 165 b15 ¯ ¯ ¯ ¯ Z 0 ¯ ¡V ¯ dN11 ¯ +¯ 2 ¢ bG ¢ dxb ¯ ¯ dxb ¯ !e ¯ L µ ¶ Z 0 dM11 0 N11 ¡ V ¢ ¢ bG ¢ dxb ¡ b11V ¢ BG = ¡ dxb L c15 = 0 The ”Modi…ed Strip Theory” includes the outlined terms.6. When ignoring the outlined terms the ”Ordinary Strip Theory” is presented. 1955]. has been added here. A small viscous surge damping coe¢cient b11V . After simpli…cation.

166

CHAPTER 6. HYDROMECHANICAL LOADS

The lateral and roll motions of the water particles, relative to the cross section of an oscillating ship in still water, are de…ned by: ³ ¤ 2 = ¡y ¡ xb ¢ à ¡ OG ¢ Á h _ _ _ ¤ = ¡y ¡ xb ¢ à + V ¢ à ¡ OG ¢ Á ³ h2 _ Ĥ Ä _ Ä ³ h2 = ¡Ä ¡ xb ¢ à + 2V ¢ à ¡ OG ¢ Á y ³ ¤ 4 = ¡Á h _ ¤ = ¡Á _ ³ h4 Ĥ Ä ³ = ¡Á
h4

According to the ”Ordinary Strip Theory” the two-dimensional potential hydromechanical force on a swaying cross section in still water is de…ned by: D n 0 _¤ o 0 0 _¤ Xh2 = M22 ¢ ³ h2 + N22 ¢ ³ h2 Dt D n 0 _¤ o 0 _¤ + M24 ¢ ³ h4 + N24 ¢ ³ h4 Dt ¶ µ 0 0 _ Ĥ + N 0 ¡ V ¢ dM22 ¢ ³ ¤ = M22 ¢ ³ h2 22 h2 dxb ¶ µ 0 dM24 ¤ 0 0 _¤ ¢ ³ h4 +M24 ¢ Äh4 + N24 ¡ V ¢ ³ dxb According to the ”Modi…ed Strip Theory” this hydromechanical force becomes: ½µ ¶ ¾ D i 0 0 0 _¤ M22 ¡ ¢ N22 ¢ ³ h2 Xh2 = Dt !e ¶ ¾ ½µ D i ¤ 0 0 _ + ¢ N24 ¢ ³ h4 M24 ¡ Dt !e µ µ ¶ ¶ 0 0 V dN22 dM22 ¤ 0 0 Ä _¤ = M22 + 2 ¢ ¢ ³ h2 + N22 ¡ V ¢ ¢ ³ h2 ! e dxb dxb µ ¶ µ ¶ 0 0 V dN24 dM24 ¤ 0 0 Ä _¤ + M24 + 2 ¢ ¢ ³ h4 + N24 ¡ V ¢ ¢ ³ h4 ! e dxb dxb This results into the following coupled sway equation: ½r ¢ y ¡ Xh2 = Ä (½r + a22 ) ¢ y + b22 ¢ y + c22 ¢ y Ä _ Ä _ +a24 ¢ Á + b24 ¢ Á + c24 ¢ Á Ä _ +a26 ¢ à + b26 ¢ à + c26 ¢ à Xw2

= with: a22 = + Z
L 0 M22 ¢ dxb

b22

¯ ¯ ¯ Z ¯ 0 ¯V ¯ dN22 +¯ 2 ¢ dxb ¯ ¯! ¯ dxb ¯ e ¯ L ¶ Z µ 0 dM22 0 = + N22 ¡ V ¢ ¢ dxb dxb
L

6.2. HYDROMECHANICAL FORCES FOR SWAY c22 = 0 Z Z 0 0 a24 = + M24 ¢ dxb + OG M22 ¢ dxb
L L

167

b24

¯ ¯ ¯ Z ¯ Z 0 0 ¯V ¯ dN24 dN22 V +¯ 2 ¢ dxb + 2 ¢ OG ¢ dxb ¯ ¯! ¯ dxb !e dxb ¯ e ¯ L L ¶ ¶ Z µ Z µ 0 0 dM24 dM22 0 0 = + N24 ¡ V ¢ ¢ dxb + OG N22 ¡ V ¢ ¢ dxb dxb dxb
L L

c24 = 0 a26

b26

c26 = 0

¶ Z µ 0 dM22 0 N22 ¡ V ¢ ¢ dxb = + dxb L L ¯ ¯ ¯ Z ¯ Z 0 ¯V ¯ dN22 V 0 ¯ +¯ 2 N22 ¢ dxb + 2 ¢ xb ¢ dxb ¯ ¯ !e dxb ¯ !e ¯ L L ¶ Z µ Z 0 dM22 0 0 = + N22 ¡ V ¢ ¢ xb ¢ dxb ¡ 2V M22 ¢ dxb dxb L L ¯ ¯ ¯ 2Z ¯ 0 ¯V ¯ dN22 +¯ 2 ¢ dxb ¯ ¯! ¯ dxb ¯ e ¯ Z
0 M22

V ¢ xb ¢ dxb + 2 !e

L

The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined terms the ”Ordinary Strip Theory” is presented. After simpli…cation, the expressions for the total hydromechanical coe¢cients in the coupled sway equation become: Z 0 a22 = + M22 ¢ dxb b22 = +
L Z L 0 N22 ¢ dxb

c22 = 0 Z Z 0 0 a24 = + M24 ¢ dxb + OG M22 ¢ dxb b24 = +
L Z L 0 N24 ¢ dxb + OG L Z L 0 N22 ¢ dxb

c24 = 0 Z Z V 0 0 a26 = + M22 ¢ xb ¢ dxb + 2 N22 ¢ dxb !e L Z ZL 0 0 b26 = + N22 ¢ xb ¢ dxb ¡ V M22 ¢ dxb
L L

168 c26 = 0

CHAPTER 6. HYDROMECHANICAL LOADS

So no terms are added for the ”Modi…ed Strip Theory”.

6.3

Hydromechanical Forces for Heave

The hydromechanical forces for heave are found by an integration over the ship length of the two-dimensional values: Z 0 Xh3 = Xh3 ¢ dxb
L

The vertical motions of the water particles, relative to the cross section of an oscillating ship in still water, are de…ned by: ³ ¤ 3 = ¡z + xb ¢ µ h _ _ ¤ = ¡z + x b ¢ µ ¡ V ¢ µ _ ³ h3 Ĥ = ¡Ä + xb ¢ Ä ¡ 2V ¢ µ _ ³ h3 z µ According to the ”Ordinary Strip Theory” the two-dimensional potential hydromechanical force on a heaving cross section in still water is de…ned by:
0 Xh3

D n 0 _¤ o 0 _¤ M33 ¢ ³ h3 + N33 ¢ ³ h3 + 2½g ¢ yw ¢ ³ ¤ 3 = h Dt µ ¶ 0 dM33 0 0 Ĥ _¤ = M33 ¢ ³ h3 + N33 ¡ V ¢ ¢ ³ h3 + 2½g ¢ yw ¢ ³ ¤ 3 h dxb

According to the ”Modi…ed Strip Theory” this hydromechanical force becomes: ½µ ¶ ¾ D i ¤ 0 0 0 _ M33 ¡ Xh3 = ¢ N33 ¢ ³ h3 + 2½g ¢ yw ¢ ³ ¤ 3 h Dt !e µ ¶ µ ¶ 0 0 V dN33 dM33 ¤ 0 0 Ä _¤ = M33 + 2 ¢ ¢ ³ h3 + N33 ¡ V ¢ ¢ ³ h3 + 2½g ¢ yw ¢ ³ ¤ 3 h ! e dxb dxb This results into the following coupled heave equation: ½r ¢ z ¡ Xh3 = Ä a31 ¢ x + b31 ¢ x + c31 ¢ x Ä _ +(½r + a33 ) ¢ z + b33 ¢ z + c33 ¢ z Ä _ Ä + b35 ¢ µ + c35 ¢ µ _ +a35 ¢ µ Xw3

= with: a31 = 0 b31 = 0 c31 = 0 Z 0 a33 = + M33 ¢ dxb
L

6.3. HYDROMECHANICAL FORCES FOR HEAVE ¯ ¯ ¯ Z ¯ 0 ¯V ¯ dN33 +¯ 2 ¢ dxb ¯ ¯! ¯ dxb ¯ e ¯ L ¶ Z µ 0 dM33 0 = + N33 ¡ V ¢ ¢ dxb dxb L Z = +2½g yw ¢ dxb Z
L 0 M33

169

b33 c33 a35

b35

c35

¶ 0 dM33 ¢ dxb = ¡ ¡V ¢ dxb L L ¯ ¯ ¯ ¯ Z Z 0 ¯ ¡V ¯ dN33 V 0 +¯ 2 N33 ¢ dxb ¡ 2 ¢ xb ¢ dxb ¯ ¯! ¯ !e dxb ¯ e ¯ L L ¶ Z µ Z 0 dM33 0 0 N33 ¡ V ¢ ¢ xb ¢ dxb + 2V = ¡ M33 ¢ dxb dxb L L ¯ ¯ ¯ 2Z ¯ 0 ¯V ¯ dN33 ¯ +¯ 2 ¢ dxb ¯ ¯ dxb ¯ !e ¯ L Z = ¡2½g yw ¢ xb ¢ dxb V ¢ xb ¢ dxb ¡ 2 !e
0 N33 L

Z µ

The ”Modi…ed Strip Theory Method” includes the outlined terms. When ignoring the outlined terms the ”Ordinary Strip Theory” is presented. After simpli…cation, the expressions for the total hydromechanical coe¢cients in the coupled heave equation become: a31 = 0 b31 = 0 c31 = 0 Z 0 a33 = + M33 ¢ dxb b33 = +
L Z L 0 N33 ¢ dxb

c33 = +2½g a35 b35 Z

Z V 0 = ¡ ¢ xb ¢ dxb ¡ 2 N33 ¢ dxb !e L Z ZL 0 0 = ¡ N33 ¢ xb ¢ dxb + V M33 ¢ dxb
0 M33 L L

Z
L

yw ¢ dxb

170 c35 = ¡2½g Z
L

CHAPTER 6. HYDROMECHANICAL LOADS yw ¢ xb ¢ dxb

So no terms are added for the ”Modi…ed Strip Theory”.

6.4

Hydromechanical Moments for Roll

The hydromechanical moments for roll are found by an integration over the ship length of the two-dimensional values: Z 0 Xh4 = Xh4 ¢ dxb
L

The roll and lateral motions of the water particles, relative to the cross section of an oscillating ship in still water, are de…ned by: ³ ¤ 4 = ¡Á h _ _¤ ³ h4 = ¡Á Ĥ Ä ³ = ¡Á
h4

According to the ”Ordinary Strip Theory” the two-dimensional potential hydromechanical moment on a rolling cross section in still water is de…ned by: µ 3 ¶ D n 0 _¤ o yw As ¤ 0 0 _ Xh4 = f M44 ¢ ³ h4 + N44 ¢ ³ h4 + 2½g ¢ ¡ ¢ bG ¢ ³ ¤ 4 h Dt 3 2 D n 0 _¤ o 0 _¤ + M42 ¢ ³ h2 + N42 ¢ ³ h2 Dt µ ¶ µ 3 ¶ 0 0 Ĥ + N 0 ¡ V ¢ dM44 ¢ ³ ¤ + 2½g ¢ yw ¡ As ¢ bG ¢ ³ ¤ _ = M44 ¢ ³ h4 h4 h4 44 dxb 3 2 µ ¶ 0 dM42 ¤ 0 0 _¤ +M42 ¢ Äh2 + N42 ¡ V ¢ ³ ¢ ³ h2 dxb or the ”Modi…ed Strip Theory” this hydromechanical moment becomes: ¶ ¾ µ 3 ½µ ¶ D i 0 0 0 _ ¤ + 2½g ¢ yw ¡ As ¢ bG ¢ ³ ¤ Xh4 = M44 ¡ ¢ N44 ¢ ³ h4 h4 Dt !e 3 2 ½µ ¶ ¾ D i 0 0 _¤ + M42 ¡ ¢ N42 ¢ ³ h2 Dt !e µ ¶ µ ¶ 0 0 V dN44 dM44 ¤ 0 0 _¤ ¢ Äh4 + N44 ¡ V ¢ ³ ¢ ³ h4 = M44 + 2 ¢ ! e dxb dxb µ 3 ¶ yw As ¡ ¢ bG ¢ ³ ¤ 4 +2½g ¢ h 3 2 µ ¶ µ ¶ 0 0 V dN42 0 Ĥ + N 0 ¡ V ¢ dM42 ¢ ³ ¤ _ + M42 + 2 ¢ ¢ ³ h2 42 h2 ! e dxb dxb This results into the following coupled roll equation: Ä Ä Ixx ¢ Á ¡ Ixz ¢ à ¡ Xh4 = a42 ¢ y + b42 ¢ y + c42 ¢ y Ä _ Ä _ +(Ixx + a44 ) ¢ Á + b44 ¢ Á + c44 ¢ Á Ä _ +(¡Ixz + a46 ) ¢ à + b46 ¢ à + c46 ¢ à = Xw4

³ ¤ 2 = ¡y ¡ xb ¢ à ¡ OG ¢ Á h _¤ _ _ ³ h2 = ¡y ¡ xb ¢ à + V ¢ à ¡ OG ¢ Á _ Ĥ Ä _ Ä ³ h2 = ¡Ä ¡ xb ¢ à + 2V ¢ à ¡ OG ¢ Á y

6.4. After simpli…cation.. the expressions for the total hydromechanical coe¢cients in the coupled roll equation become: Z 0 a42 = + M42 ¢ dxb + OG ¢ a22 L . has been added here. A viscous roll damping coe¢cient b44V . When ignoring the outlined terms the ”Ordinary Strip Theory Method” is presented. derived for instance with the empiric method of [Ikeda et al. 1978]. HYDROMECHANICAL MOMENTS FOR ROLL with: a42 = + Z L 171 0 M42 ¢ dxb + OG ¢ a22 b42 ¯ ¯ ¯ Z ¯ 0 ¯V ¯ dN42 ¯ +¯ 2 ¢ dxb ¯ ¯ dxb ¯ !e ¯ L ¶ Z µ 0 dM42 0 N42 ¡ V ¢ ¢ dxb + OG ¢ b22 = + dxb L c42 = 0 + OG ¢ c22 Z Z 0 0 a44 = + M44 ¢ dxb + OG M42 ¢ dxb + OG ¢ a24 L L b44 c44 ¯ ¯ ¯ Z ¯ Z 0 0 ¯V ¯ dN44 dN42 V ¯ +¯ 2 ¢ dxb + 2 ¢ OG ¢ dxb ¯ ¯ dxb !e dxb ¯ !e ¯ L L ¶ ¶ Z µ Z µ 0 0 dM44 dM42 0 0 = + N24 ¡ V ¢ ¢ dxb + OG N42 ¡ V ¢ ¢ dxb + b44V + OG ¢ b24 dxb dxb L L ¶ Z µ 3 yw As ¡ ¢ bG dxb = +2½g 3 2 L a46 b46 c46 = 0 + OG ¢ c26 = +½gr ¢ GM ¶ Z Z µ 0 V dM42 0 0 = + M42 ¢ xb ¢ dxb + 2 N42 ¡ V ¢ ¢ dxb + OG ¢ a26 !e dxb L L ¯ ¯ ¯ Z ¯ Z 0 ¯V ¯ V dN42 0 +¯ 2 N42 ¢ dxb + 2 ¢ xb ¢ dxb ¯ ¯! ¯ !e dxb ¯ e ¯ L L ¶ Z µ Z 0 dM42 0 0 = + M42 ¢ dxb + OG ¢ b26 N42 ¡ V ¢ ¢ xb ¢ dxb ¡ 2V dxb L L ¯ ¯ ¯ 2Z ¯ 0 ¯V ¯ dN42 ¯ +¯ 2 ¢ dxb ¯ ¯ dxb ¯ !e ¯ L The ”Modi…ed Strip Theory” includes the outlined terms.

172 b42 = + Z L CHAPTER 6. HYDROMECHANICAL LOADS 0 N42 ¢ dxb + OG ¢ b22 c42 = 0 Z Z 0 0 a44 = + M44 ¢ dxb + OG M42 ¢ dxb + OG ¢ a24 b44 = + L Z L 0 N44 ¢ dxb + OG L Z L 0 N42 ¢ dxb + b44V + OG ¢ b24 c44 = +½gr ¢ GM Z Z V 0 0 a46 = + M42 ¢ xb ¢ dxb + 2 N42 ¢ dxb + OG ¢ a26 !e L Z ZL 0 0 b46 = + N42 ¢ xb ¢ dxb ¡ V M42 ¢ dxb + OG ¢ b26 L L c46 = 0 So no terms are added for the ”Modi…ed Strip Theory”. 6.5 Hydromechanical Moments for Pitch The hydromechanical moments for pitch are found by an integration over the ship length of the two-dimensional contributions of surge and heave into the pitch moment: Z 0 0 0 0 with: Xh5 = ¡Xh1 ¢ bG ¡ Xh3 ¢ xb Xh5 = Xh5 ¢ dxb L According to the ”Ordinary Strip Theory” the two-dimensional potential hydromechanical moment on a pitching cross section in still water is de…ned by surge and heave contributions: µ ¶ 0 dM11 ¤ 0 0 0 Ä _¤ Xh5 = ¡M11 ¢ bG ¢ ³ h1 ¡ N11 ¡ V ¢ ¢ bG ¢ ³ h1 dxb ¶ µ 0 0 Ĥ ¡ N 0 ¡ V ¢ dM33 ¢ xb ¢ ³ ¤ ¡ 2½g ¢ yw ¢ xb ¢ ³ ¤ _ ¡M33 ¢ xb ¢ ³ h3 33 h3 h3 dxb According to the ”Modi…ed Strip Theory” this hydromechanical moment becomes: µ µ ¶ ¶ 0 0 V dN11 0 0 Ĥ ¡ N 0 ¡ V ¢ dM11 ¢ bG ¢ ³ ¤ _ Xh5 = ¡ M11 + 2 ¢ ¢ bG ¢ ³ h1 11 h1 ! e dxb dxb µ µ ¶ ¶ 0 0 V dN33 dM33 ¤ 0 0 _¤ ¡ M33 + 2 ¢ ³ ¢ xb ¢ Äh3 ¡ N33 ¡ V ¢ ¢ xb ¢ ³ h3 ¡ 2½g ¢ yw ¢ xb ¢ ³ ¤ 3 h ! e dxb dxb This results into the following coupled pitch equation: Iyy ¢ Ä ¡ Xh5 = µ a51 ¢ x + b51 ¢ x + c51 ¢ x Ä _ +a53 ¢ z + b53 ¢ z + c53 ¢ z Ä _ _ µ +(Iyy + a55 ) ¢ Ä + b55 ¢ µ + c55 ¢ µ Xw5 = .

6.5. HYDROMECHANICAL MOMENTS FOR PITCH with: a51 = ¡ Z
L 0 M11 ¢ bG ¢ dxb

173

b51

¯ ¯ ¯ ¯ Z 0 ¯ ¡V ¯ dN11 +¯ 2 ¢ bG ¢ dxb ¯ ¯! ¯ dxb ¯ e ¯ L ¶ Z µ 0 dM11 0 = ¡ N11 ¡ V ¢ ¢ bG ¢ dxb ¡ b11V ¢ BG dxb
L

c51 = 0 Z 0 a53 = ¡ M33 ¢ xb ¢ dxb
L

b53 c53

¯ ¯ ¯ ¯ Z 0 ¯ ¡V ¯ dN33 +¯ 2 ¢ xb ¢ dxb ¯ ¯! ¯ dxb ¯ e ¯ L ¶ Z µ 0 dM33 0 = ¡ N33 ¡ V ¢ ¢ xb ¢ dxb dxb L Z = ¡2½g yw ¢ xb ¢ dxb Z
L L 0 M11 ¢ bG ¢ dxb 2

a55 = +

b55

c55

¯ ¯ ¯ Z ¯ 0 ¯V ¯ dN11 2 +¯ 2 ¢ bG ¢ dxb ¯ ¯! ¯ dxb ¯ e ¯ L ¶ Z Z µ 0 V dM33 0 0 2 + M33 ¢ xb ¢ dxb + 2 N33 ¡ V ¢ ¢ xb ¢ dxb !e dxb L L ¯ ¯ ¯ Z ¯ Z 0 ¯V ¯ dN33 2 V 0 +¯ 2 N33 ¢ xb ¢ dxb + 2 ¢ xb ¢ dxb ¯ ¯! ¯ !e dxb ¯ e ¯ L L ¶ Z µ 0 dM11 2 2 0 = + N11 ¡ V ¢ ¢ bG ¢ dxb + b11V ¢ BG dxb L ¶ Z µ Z 0 dM33 0 0 2 + N33 ¡ V ¢ ¢ xb ¢ dxb ¡ 2V M33 ¢ xb ¢ dxb dxb L L ¯ ¯ ¯ ¯ Z 0 ¯ ¡V 2 ¯ dN33 ¯ +¯ 2 ¢ xb ¢ dxb ¯ ¯ dxb ¯ !e ¯ L Z = +2½g yw ¢ x2 ¢ dxb b
L

174

CHAPTER 6. HYDROMECHANICAL LOADS

The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined terms the ”Ordinary Strip Theory” is presented. After simpli…cation, the expressions for the total hydromechanical coe¢cients in the coupled pitch equation become: Z 0 a51 = ¡ M11 ¢ bG ¢ dxb ¯ ¯ ¯ ¯ Z 0 ¯ ¡V ¯ dN11 +¯ 2 ¢ bG ¢ dxb ¯ ¯! ¯ dxb ¯ e ¯ L ¶ Z µ 0 dM11 0 = ¡ N11 ¡ V ¢ ¢ bG ¢ dxb ¡ b11V ¢ BG dxb
L L

b51

c51 = 0 Z 0 a53 = ¡ M33 ¢ xb ¢ dxb
L

b53

¯ ¯ ¯ Z ¯ ¯V ¯ 0 +¯ 2 N33 ¢ dxb ¯ ¯! ¯ ¯ e ¯ L Z Z 0 0 = ¡ N33 ¢ xb ¢ dxb ¡ V M33 ¢ dxb
L

c53 = ¡2½g a55 = + Z
L

Z
L

L

yw ¢ xb ¢ dxb
2

0 M11 ¢ bG ¢ dxb

b55

c55

¯ ¯ ¯ Z ¯ 0 ¯V ¯ dN11 2 +¯ 2 ¢ bG ¢ dxb ¯ ¯! ¯ dxb ¯ e ¯ L Z Z Z V V2 0 0 0 2 + M33 ¢ xb ¢ dxb + 2 N33 ¢ xb ¢ dxb + 2 M33 ¢ dxb !e !e L L L ¯ ¯ ¯ ¯ Z ¯ ¡V ¯ 0 +¯ 2 N33 ¢ xb ¢ dxb ¯ ¯! ¯ ¯ e ¯ L ¶ Z µ Z 0 dM11 2 2 0 0 = + N11 ¡ V ¢ ¢ bG ¢ dxb + b11V ¢ BG + N33 ¢ x2 ¢ dxb b dxb L L ¯ ¯ ¯ 2Z ¯ ¯V ¯ 0 +¯ 2 N33 ¢ dxb ¯ ¯! ¯ ¯ e ¯ L Z = +2½g yw ¢ x2 ¢ dxb b
L

6.6. HYDROMECHANICAL MOMENTS FOR YAW

175

6.6

Hydromechanical Moments for Yaw

The hydromechanical moments for yaw are found by an integration over the ship length of the two-dimensional contributions of sway into the yaw moment: Z 0 0 0 with: Xh6 = +Xh2 ¢ xb Xh6 = Xh6 ¢ dxb
L

According to the ”Ordinary Strip Theory” the two-dimensional potential hydromechanical force on a yawing cross section in still water is de…ned by sway contributions: o D n 0 0 0 _¤ _¤ Xh6 = M22 ¢ xb ¢ ³ h2 + N22 ¢ xb ¢ ³ h2 Dt o D n 0 _ ¤ + N 0 ¢ xb ¢ ³ ¤ _ M24 ¢ xb ¢ ³ h4 + 24 h4 Dt ¶ µ 0 dM22 ¤ 0 0 _¤ ¢ xb ¢ ³ h2 = M22 ¢ xb ¢ Äh2 + N22 ¡ V ¢ ³ dxb µ ¶ 0 0 Ĥ + N 0 ¡ V ¢ dM24 ¢ xb ¢ ³ ¤ _ +M24 ¢ xb ¢ ³ h4 24 h4 dxb According to the ”Modi…ed Strip Theory” this hydromechanical force becomes: ½µ ¶ ¾ D i 0 0 0 _¤ M22 ¡ ¢ N22 ¢ xb ¢ ³ h2 Xh6 = Dt !e ½µ ¶ ¾ i D 0 0 _¤ M24 ¡ ¢ N24 ¢ xb ¢ ³ h4 + Dt !e µ ¶ µ ¶ 0 0 V dN22 0 Ĥ + N 0 ¡ V ¢ dM22 ¢ xb ¢ ³ ¤ _ = M22 + 2 ¢ ¢ xb ¢ ³ h2 22 h2 ! e dxb dxb µ ¶ µ ¶ 0 0 V dN24 dM24 0 0 Ĥ _¤ + M24 + 2 ¢ ¢ xb ¢ ³ h4 + N24 ¡ V ¢ ¢ xb ¢ ³ h4 ! e dxb dxb This results into the following coupled yaw equation: Ä Ä Izz ¢ à ¡ Izx ¢ Á ¡ Xh6 = a62 ¢ y + b62 ¢ y + c62 ¢ y Ä _ Ä _ +(¡Izx + a64 ) ¢ Á + b64 ¢ Á + c64 ¢ Á Ä _ +(+Izz + a66 ) ¢ à + b66 ¢ à + c66 ¢ à Xw6

= with: a62 = + Z
L

0 M22 ¢ xb ¢ dxb

b62

¯ ¯ ¯ Z ¯ 0 ¯V ¯ dN22 +¯ 2 ¢ xb ¢ dxb ¯ ¯! ¯ dxb ¯ e ¯ L ¶ Z µ 0 dM22 0 = + N22 ¡ V ¢ ¢ xb ¢ dxb dxb
L

176

CHAPTER 6. HYDROMECHANICAL LOADS c62 = 0 Z Z 0 0 a64 = + M24 ¢ xb ¢ dxb + OG M22 ¢ xb ¢ dxb
L L

b64

¯ ¯ ¯ Z ¯ Z 0 0 ¯V ¯ dN24 dN22 V +¯ 2 ¢ xb ¢ dxb + 2 ¢ OG ¢ xb ¢ dxb ¯ ¯! ¯ dxb !e dxb ¯ e ¯ L L ¶ ¶ Z µ Z µ 0 0 dM24 dM22 0 0 = + N24 ¡ V ¢ ¢ xb ¢ dxb + OG N22 ¡ V ¢ ¢ xb ¢ dxb dxb dxb
L L

c64 = 0 a66 = +

b66

c66 = 0

¯ ¯ ¯ Z ¯ Z 0 ¯V ¯ dN22 2 V 0 ¯ +¯ 2 N22 ¢ xb ¢ dxb + 2 ¢ xb ¢ dxb ¯ ¯ !e dxb ¯ !e ¯ L L ¶ Z µ Z 0 dM22 0 0 2 = + N22 ¡ V ¢ ¢ xb ¢ dxb ¡ 2V M22 ¢ xb ¢ dxb dxb L L ¯ ¯ ¯ 2Z ¯ 0 ¯V ¯ dN22 +¯ 2 ¢ xb ¢ dxb ¯ ¯! ¯ dxb ¯ e ¯
L

Z
L

0 M22

¢

x2 b

V ¢ dxb + 2 !e

Z µ
L

0 N22

0 dM22 ¡V ¢ dxb

¢ xb ¢ dxb

The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined terms the ”Ordinary Strip Theory” is presented. After simpli…cation, the expressions for the total hydromechanical coe¢cients in the coupled yaw equation become: a62 = + Z
L 0 M22 ¢ xb ¢ dxb

b62

¯ ¯ ¯ ¯ Z ¯ ¡V ¯ 0 +¯ 2 N22 ¢ dxb ¯ ¯! ¯ ¯ e ¯ L Z Z 0 0 = + N22 ¢ xb ¢ dxb + V M22 ¢ dxb
L L

c62 = 0 Z Z 0 0 a64 = + M24 ¢ xb ¢ dxb + OG M22 ¢ xb ¢ dxb
L L

¯ ¯ ¯ ¯ Z Z ¯ ¡V ¯ V 0 0 +¯ 2 N24 ¢ dxb ¡ 2 ¢ OG N22 ¢ dxb ¯ ¯! ¯ !e ¯ e ¯
L L

6.6. HYDROMECHANICAL MOMENTS FOR YAW b64 = + c64 = 0 a66 Z
L 0 N24

177 ¢ xb ¢ dxb + V ¢ OG Z
L 0 M22 ¢ dxb

¢ xb ¢ dxb + V

Z
L

0 M24

¢ dxb + OG

Z
L

0 N22

b66

V = + ¢ ¢ dxb + 2 !e L ¯ ¯ ¯ ¯ Z ¯ ¡V ¯ 0 +¯ 2 N22 ¢ xb ¢ dxb ¯ ¯! ¯ ¯ e ¯ L Z 0 = + N22 ¢ x2 ¢ dxb b
0 M22

Z

x2 b

Z
L

0 N22

V2 ¢ xb ¢ dxb + 2 !e

Z
L

0 M22 ¢ dxb

c66 = 0

¯ ¯ ¯ ¯ Z ¯ ¡V 2 ¯ 0 +¯ 2 N22 ¢ dxb ¯ ¯ ! ¯ ¯ e ¯
L

L

178 . HYDROMECHANICAL LOADS . CHAPTER 6.

Report 1216a. The velocities and accelerations in the direction j of the water particles have to be de…ned. 2628 CD Delft.wbmt.Chapter 7 Exciting Wave Loads The …rst order wave potential in a ‡uid .is given by: ©w = ¡g cosh k(h + zb ) ¢ ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹) ! cosh(kh) in an axes system with the centre of gravity in the waterline. The Netherlands.M. 4 = ¡V ¢ ³ wj Dt wj Dt @t @xb With this. Journée. 3.nl/~johan or http://www.tudelft. February 2001. Ship Hydromechanics Laboratory. Delft University of Technology. 2. ”Theoretical Manual of SEAWAY.nl. For updates see web site: http://dutw189. The local relative orbital velocities of the water particles in a certain direction follow from the derivative in that direction of the wave potential.19”. Release 4. The orbital accelerations of the water particles can be obtained from these velocities by: ½ ¾ n o @ @ D Ä0 = D ³ 0 _ with: for: j = 1. 179 . the relative velocities and accelerations in the di¤erent directions can be found: ² Surge direction: @©w _0 ³ w1 = @xb = +! ¢ _ @ ³ w1 Ä0 ³ w1 = @t = ¡! 2 ¢ ² Sway direction: @©w _0 ³ w2 = @yb = +! ¢ 0 cosh k(h + zb ) ¢ cos ¹ ¢ ³ a cos(!t ¡ kxb cos ¹ ¡ kyb sin ¹) sinh(kh) 0 cosh k(h + zb ) ¢ cos ¹ ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹) sinh(kh) cosh k(h + zb ) ¢ sin ¹ ¢ ³ a cos(!t ¡ kxb cos ¹ ¡ kyb sin ¹) sinh(kh) J.with any arbitrary water depth h .J.shipmotions. Mekelweg 2.

EXCITING WAVE LOADS _ @ ³ w2 Ä0 ³ w2 = @t cosh k(h + zb ) 0 Ä ³ w2 = ¡! 2 ¢ ¢ sin ¹ ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹) sinh(kh) ² Heave direction: @©w _0 ³ w3 = @zb = ¡! ¢ _ @ ³ w3 Ä0 ³ w3 = @t = ¡! 2 ¢ ² Roll direction: sinh k(h + zb ) ¢ ³ a cos(!t ¡ kxb cos ¹ ¡ kyb sin ¹) sinh(kh) 0 0 sinh k(h + zb ) ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹) sinh(kh) _0 _0 @ ³ w2 @ ³ w3 _0 ³ w4 = ¡ =0 @zb @yb Ä0 ³ = 0 w4 of which the zero solution is obvious.180 CHAPTER 7. The pressure in the ‡uid follows from the linearized equation of Bernoulli: p = ¡½gzb + ½g ¢ cosh k(h + zb ) ¢ ³ a cos(!t ¡ kxb cos ¹ ¡ kyb sin ¹) cosh(kh) @p @p @p ¢ dxb + ¢ dyb + ¢ dzb = p0 + @xb @yb @zb with the following expressions for the pressure gradients: @p cosh k(h + zb ) ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹) = +½kg cos ¹ ¢ @xb cosh(kh) cosh k(h + zb ) = +½! 2 cos ¹ ¢ ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹) sinh(kh) @p cosh k(h + zb ) = +½kg sin ¹ ¢ ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹) @yb cosh(kh) cosh k(h + zb ) ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹) = +½! 2 sin ¹ ¢ sinh(kh) @p sinh k(h + zb ) = ¡½g + ½kg ¢ ¢ ³ a cos(!t ¡ kxb cos ¹ ¡ kyb sin ¹) @zb cosh(kh) sinh k(h + zb ) = ¡½g + ½! ¢ ¢ ³ a cos(!t ¡ kxb cos ¹ ¡ kyb sin ¹) sinh(kh) . because the potential ‡uid is free of rotation.

is given by: +y Z³ Z b XF K1 = ¡ @p ¢ dyb ¢ dzb @xb +y Z³ Z b Ä0 = ½ ³ w1 ¢ dyb ¢ dzb ¡T ¡yb ¡T ¡yb .1 Classical Approach First the classical approach to obtain the wave loads .so the longitudinal force due to the pressure in the undisturbed ‡uid .according to the relative motion principle . CLASSICAL APPROACH These pressure gradients can be expressed in the orbital accelerations too: @p Ä0 = ¡½ ¢ ³ w1 @xb @p Ä0 = ¡½ ¢ ³ w2 @yb @p 0 = ¡½ ¢ (g + Äw3 ) ³ @zb 181 7.is given here.1.1 Exciting Wave Forces for Surge The exciting wave forces for surge on a ship are found by an integration over the ship length of the two-dimensional values: Z 0 Xw1 = Xw1 ¢ dxb L According to the ”Ordinary Strip Theory” the exciting wave forces for surge on a restrained cross section of a ship in waves are de…ned by: 0 Xw1 = D n 0 _¤ o 0 0 _¤ M11 ¢ ³ w1 + N11 ¢ ³ w1 + XF K1 Dt µ ¶ 0 0 Ĥ + N 0 ¡ V ¢ dM11 ¢ ³ ¤ + X 0 _ = M11 ¢ ³ w1 11 F K1 w1 dxb According to the ”Modi…ed Strip Theory” these forces become: ¶ ¾ ½µ D i 0 0 0 0 _¤ Xw1 = M11 ¡ ¢ N11 ¢ ³ w1 + XF K1 Dt !e µ ¶ µ ¶ 0 0 V dN11 dM11 ¤ 0 0 0 _¤ = M11 + 2 ¢ ¢ Äw1 + N11 ¡ V ¢ ³ ¢ ³ w1 + XF K1 ! e dxb dxb The Froude-Krilov force in the surge direction .7. 7.1.

it is found: µ ¶ 1 2 XF K1 = ½ ¢ A + k ¢ Sy + k ¢ Iy + ::: ¢ (¡kg cos ¹) ¢ ³ a sin(! e t ¡ kxb cos ¹) 2 with: A=2 Z0 Z0 Z0 yb ¢ dzb Sy = 2 yb ¢ zb ¢ dzb Iy = 2 2 yb ¢ zb ¢ dzb ¡T ¡T ¡T The acceleration term kg cos ¹ ¢ ³ a in here is the amplitude of the longitudinal component of the relative orbital acceleration in deep water at zb = 0.182 CHAPTER 7. The dominating …rst term in this series consists of a mass and this acceleration. This mass term ½A is used to obtain from the total Froude-Krilov force an equivalent longitudinal component of the orbital acceleration of the water particles: ¤ ³ XF K1 = ½A ¢ Äw1 This holds that the equivalent longitudinal components of the orbital acceleration and velocity are equal to the values at zb = 0 in a wave with a reduced amplitude ³ ¤1 : a Ĥ ³ w1 = ¡kg cos ¹ ¢ ³ ¤1 sin(! e t ¡ kxb cos ¹) a +kg cos ¹ ¤ _¤ ³ w1 = ¢ ³ a1 cos(! e t ¡ kxb cos ¹) ! . EXCITING WAVE LOADS Figure 7. this can be written as: XF K1 = ½Ach ¢ (¡kg cos ¹) ¢ ³ a sin(! e t ¡ kxb cos ¹) with: Ach = 2 Z0 sin(¡kyb sin ¹) cosh k(h + zb ) ¢ ¢ yb ¢ dzb ¡kyb sin ¹ cosh(kh) ¡T When expanding the Froude-Krilov force in deep water with ¸ À 2¼ ¢ yw and ¸ À 2¼ ¢ T in series.1: Wave Pressure Distribution on a Cross Section for Surge After neglecting the second order terms.

2 Exciting Wave Forces for Sway The exciting wave forces for sway on a ship are found by an integration over the ship length of the two-dimensional values: Z 0 Xw2 = Xw2 ¢ dxb L According to the ”Ordinary Strip Theory” the exciting wave forces for sway on a restrained cross section of a ship in waves are de…ned by: D n 0 _¤ o 0 0 0 _¤ M22 ¢ ³ w2 + N22 ¢ ³ w2 + XF K2 Xw2 = Dt ¶ µ 0 dM22 ¤ 0 0 0 _¤ ¢ ³ w2 + XF K2 = M22 ¢ Äw2 + N22 ¡ V ¢ ³ dxb According to the ”Modi…ed Strip Theory” these forces become: ½µ ¶ ¾ D i 0 0 0 0 _¤ Xw2 = M22 ¡ ¢ N22 ¢ ³ w2 + XF K2 Dt !e ¶ ¶ µ µ 0 0 V dN22 dM22 ¤ 0 0 0 _¤ ¢ Äw2 + N22 ¡ V ¢ ³ ¢ ³ w2 + XF K2 = M22 + 2 ¢ ! e dxb dxb The Froude-Krilov force in the sway direction .7. When ignoring the outlined terms the ”Ordinary Strip Theory” is presented.1. From the previous follows the total wave loads for surge: Z 0 Ĥ Xw1 = + M11 ¢ ³ w1 ¢ dxb ³ ¤1 = a ¯ ¯ ¯ ¯ Z 0 ¯ V ¯ dN11 Ĥ ¯ +¯ ¢ ³ w1 ¢ dxb ¯ ¯ dxb ¯ ! ¢ !e ¯ L ! Z ï ¯ 0 ¯!¯ 0 dM11 _¤ ¯ ¯¢N ¡ V ¢ + ¢ ³ w1 ¢ dxb ¯ ! e ¯ 11 dxb L Z 0 + XF K1 ¢ dxb L L The ”Modi…ed Strip Theory” includes the outlined terms. CLASSICAL APPROACH with: 183 Ach ¢ ³a A This equivalent acceleration and velocity will be used in the di¤raction part of the wave force for surge.1.so the lateral force due to the pressure in the undisturbed ‡uid . 7.is given by: XF K2 = ¡ +y Z³ Z b @p ¢ dyb ¢ dzb @yb ¡T ¡yb .

this can be written as: XF K2 = ½Ach ¢ (¡kg sin ¹) ¢ ³ a sin(! e t ¡ kxb cos ¹) with: Ach = 2 Z0 sin(¡kyb sin ¹) cosh k(h + zb ) ¢ ¢ yb ¢ dzb ¡kyb sin ¹ cosh(kh) When expanding the Froude-Krilov force in deep water with ¸ À 2¼ ¢ yw and ¸ À 2¼ ¢ T in series. This mass term ½A is used to obtain from the total Froude-Krilov force an equivalent lateral component of the orbital acceleration of the water particles: ¤ XF K2 = ½A ¢ Äw2 ³ ¡T ¡T ¡T This holds that the equivalent lateral components of the orbital acceleration and velocity are equal to the values at zb = 0 in a wave with a reduced amplitude ³ ¤2 : a Ĥ = ¡kg sin ¹ ¢ ³ ¤ sin(! e t ¡ kxb cos ¹) ³ w2 a2 +kg sin ¹ ¤ _¤ ³ w2 = ¢ ³ a2 cos(! e t ¡ kxb cos ¹) ! . it is found: µ ¶ 1 2 XF K2 = ½ ¢ A + k ¢ Sy + k ¢ Iy + ::: ¢ (¡kg sin ¹) ¢ ³ a sin(! et ¡ kxb cos ¹) 2 with: A=2 Z0 yb ¢ dzb Sy = 2 Z0 yb ¢ zb ¢ dzb Iy = 2 Z0 2 yb ¢ zb ¢ dzb ¡T The acceleration term kg sin ¹ ¢ ³ a in here is the amplitude of the lateral component of the relative orbital acceleration in deep water at zb = 0. EXCITING WAVE LOADS Figure 7. The dominating …rst term in this series consists of a mass and this acceleration.2: Wave Pressure Distribution on a Cross Section for Sway +y Z³ Z b Ä0 = ½ ³ w2 ¢ dyb ¢ dzb ¡T ¡yb After neglecting the second order terms.184 CHAPTER 7.

7. From the previous follows the total wave loads for sway: Xw2 = + Z L 0 Ĥ M22 ¢ ³ w2 ¢ dxb ¯ ¯ ¯ ¯ Z 0 ¯ V ¯ dN22 Ĥ +¯ ¢ ³ w2 ¢ dxb ¯ ¯ ! ¢ !e ¯ dxb ¯ ¯ L ! Z ï ¯ 0 ¯!¯ 0 dM22 _¤ ¯ ¯¢N ¡ V ¢ + ¢ ³ w2 ¢ dxb ¯ ! e ¯ 22 dxb L Z 0 + XF K2 ¢ dxb L The ”Modi…ed Strip Theory” includes the outlined terms.1. CLASSICAL APPROACH with: ³ ¤2 = a Ach ¢ ³a A 185 This equivalent acceleration and velocity will be used in the di¤raction part of the wave force for sway. 7. When ignoring the outlined terms the ”Ordinary Strip Theory” is presented.1.3 Exciting Wave Forces for Heave The exciting wave forces for heave on a ship are found by an integration over the ship length of the two-dimensional values: Xw3 = Z L 0 Xw3 ¢ dxb According to the ”Ordinary Strip Theory” the exciting wave forces for heave on a restrained cross section of a ship in waves are de…ned by: 0 Xw3 = D n 0 _¤ o 0 0 _¤ M33 ¢ ³ w3 + N33 ¢ ³ w3 + XF K3 Dt µ ¶ 0 dM33 ¤ 0 0 0 _¤ = M33 ¢ Äw3 + N33 ¡ V ¢ ³ ¢ ³ w3 + XF K3 dxb According to the ”Modi…ed Strip Theory” these forces become: 0 Xw3 ¶ ¾ ½µ D i ¤ 0 0 0 _ = ¢ N33 ¢ ³ w3 + XF K3 M33 ¡ Dt !e µ ¶ µ ¶ 0 0 V dN33 dM33 ¤ 0 0 0 _¤ = M33 + 2 ¢ ¢ Äw3 + N33 ¡ V ¢ ³ ¢ ³ w3 + XF K3 ! e dxb dxb .

EXCITING WAVE LOADS Figure 7.186 CHAPTER 7. this force can be written as: XF K3 = 2½gyw ¢ C3 ¢ ³ a cos(! e t ¡ kxb cos ¹) with: sin(¡kyw sin ¹) k C3 = ¡ ¡kyw sin ¹ yw Z0 sin(¡kyb sin ¹) sinh k(h + zb ) ¢ ¢ yb ¢ dzb ¡kyb sin ¹ cosh(kh) 0 ¡T When expanding the Froude-Krilov force in deep water with ¸ À 2¼ ¢ T and in long waves with ¸ À 2¼ ¢ yw in series. it is found: C3 sin(¡kyw sin ¹) k ¡ = ¡kyw sin ¹ yw t 1¡ k yw Z0 Z0 sin(¡kyb sin ¹) sinh k(h + zb ) ¢ ¢ yb ¢ dzb ¡kyb sin ¹ cosh(kh) ¡T ekzb ¢ yb ¢ dzb ¡T k = 1¡ yw ¡T ¶ Z0 µ k2 2 1 + k ¢ zb + ¢ zb + :::::: ¢ yb ¢ dzb 2 .is given by: 0 XF K3 = ¡ +y Z³ Z b Ä0 = ½ (g + ³ w3 ) ¢ dyb ¢ dzb ¡T ¡yb ¡T ¡yb +y Z³ Z b @p ¢ dyb ¢ dzb @zb After neglecting the second order terms.so the vertical force due to the pressure in the undisturbed ‡uid .3: Wave Pressure Distriubution on a Cross Section for Heave The Froude-Krilov force in the heave direction .

7.1. CLASSICAL APPROACH = = = t with: A=2 ¤ T3 187 ¶ A Sy k 2 Iy 1 ¡ k¢ +k¢ + ¢ + :::::: 2yw 2yw 2 2yw µ ¶ A 1 ¡ k¢ + :::::: 2yw ¤ 1 ¡ kT3 ¤ exp f¡kT3 g Z0 yb ¢ dzb Sy = 2 Z0 yb ¢ zb ¢ dzb Iy = 2 Z0 2 yb ¢ zb ¢ dzb µ ¡T ¡T ¡T can be considered as the draft at which the pressure in the vertical direction is equal to the average vertical pressure on the cross section in the ‡uid and can be obtained by. it is found: C3 sin(¡kyw sin ¹) k = ¡ ¡kyw sin ¹ yw t 1¡ k yw Z0 Z0 Z0 sin(¡kyb sin ¹) sinh k(h + zb ) ¢ ¢ yb ¢ dzb ¡kyb sin ¹ cosh(kh) ¡T sinh k(h + zb ) ¢ yb ¢ dzb cosh(kh) ¡T k(h + zb ) + k 3 ¢ (::::::) + :::::: ¢ yb ¢ dzb cosh(kh) ¡T 0 0 1 Z Z0 kh 1 k @ = 1¡ ¢ yb ¢ dzb + yb ¢ zb ¢ dzb + ::::::A ¢ cosh(kh) yw h k = 1¡ yw ¡T ¡T . ¤ T3 = ¡ ln C3 k This holds that the equivalent vertical components of the orbital acceleration and velocity ¤ are equal to the values at zb = ¡T3 : ¤ Ĥ ³ w3 = ¡kg ¢ e¡kT3 ¢ ³ a3 ¢ cos(! e t ¡ kxb cos ¹) ¡ ¢ ¤ = ¡! 2 ¢ e¡kT3 ¢ ³ a3 ¢ cos(! e t ¡ kxb cos ¹) ¡kg ¡kT3 ¤ _¤ ¢e ³ w3 = ¢ ³ a3 ¢ sin(! e t ¡ kxb cos ¹) ! ¡ ¢ ¤ = ¡! ¢ e¡kT3 ¢ ³ a3 ¢ sin(! e t ¡ kxb cos ¹) kg ¡kT3 ¤ _¤ ³ w3 = ¢e ¢ ³ a3 ¢ cos(! e t ¡ kxb cos ¹) 2 ! ¡ ¢ ¤ = e¡kT3 ¢ ³ a3 ¢ cos(! e t ¡ kxb cos ¹) When expanding the Froude-Krilov force in shallow water with kh ! 0 and in long waves with ¸ À 2¼ ¢ yw in series.

¤ T3 = kh cosh(kh) This holds that the equivalent vertical components of the orbital acceleration and velocity ¤ are equal to the values at zb = ¡T3 : ¤ sinh k(h ¡ T3 ) Ĥ ¢³ a3 ¢ cos(! et ¡ kxb cos ¹) ³ w3 = ¡kg ¢ cosh(kh) µ ¶ ¤ sinh k(h ¡ T3 ) 2 = ¡! ¢ ¢³ a3 ¢ cos(! e t ¡ kxb cos ¹) sinh(kh) ¤ ¡kg sinh k(h ¡ T3 ) _¤ ³ w3 = ¢ ¢³ a3 ¢ sin(! e t ¡ kxb cos ¹) ! cosh(kh) ¶ µ ¤ sinh k(h ¡ T3 ) ¢³ a3 ¢ sin(! e t ¡ kxb cos ¹) = ¡! ¢ sinh(kh) ¡ ln C3 ¡ ¢ B 1 + zh ¢ k ³¤ 3 = w ¤ kg sinh k(h ¡ T3 ) ¢ ¢³ a3 ¢ cos(! e t ¡ kxb cos ¹) !2 cosh(kh) µ ¶ ¤ sinh k(h ¡ T3 ) ¢³ a3 ¢ cos(! e t ¡ kxb cos ¹) = sinh(kh) ¤ It may be noted that this shallow water de…nition for T3 is valid in deep water too. EXCITING WAVE LOADS µ ¶ kh A Sy 1¡ ¢k¢ + + :::::: cosh(kh) 2yw h ¢ 2yw ³ ´ A kh zB 1¡ ¢k¢ 1+ + :::::: ¢ cosh(kh) h 2yw µ ¶ ³ ´ kh zB A 1¡ + :::::: ¢ 1+ ¢k¢ cosh(kh) h 2yw ³ kh zB ´ ¤ 1¡ 1+ ¢ kT3 cosh(kh) h ¾ ½ ³ zB ´ kh ¤ exp ¡ 1+ ¢ kT3 cosh(kh) h Z0 Z0 yb ¢ dzb Sy = 2 yb ¢ zb ¢ dzb = zB ¢ A ¡T ¡T ¤ So in shallow water. T3 can be obtained by. .188 = = = = t with: A=2 CHAPTER 7. because: kh !1 cosh(kh) and zB !0 h for: h ! 1 These equivalent accelerations and velocities will be used to determine the di¤raction part of the wave forces for heave.

so the roll moment due to the pressure in the undisturbed ‡uid . CLASSICAL APPROACH From the previous follows the total wave loads for heave: Z 0 0 Ĥ Xw3 = + M33 ¢ ³ w3 ¢ dxb L 189 ¯ ¯ ¯ ¯ Z 0 ¯ V ¯ dN33 Ĥ ¯ +¯ ¢ ³ w3 ¢ dxb ¯ ¯ dxb ¯ ! ¢ !e ¯ L ! Z ï ¯ 0 ¯!¯ dM33 0 _¤ ¯ ¯ ¢ N33 ¡ V ¢ + ¢ ³ w3 ¢ dxb ¯ !e ¯ dxb L Z 0 + XF K3 ¢ dxb L The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined terms the ”Ordinary Strip Theory” is presented. 7.1.1.7.is given by: XF K4 +y ¶ Z³ Z bµ @p @p ¡ = ¡ ¢ zb + ¢ yb ¢ dyb ¢ dzb @yb @zb +y Z³ Z b³ ´ Ä Ä0 ¢ zb + (g + ³ 0 ) ¢ yb ¢ dyb ¢ dzb = ½ ¡³ w2 w3 ¡T ¡yb ¡T ¡yb .4 Exciting Wave Moments for Roll The exciting wave moments for roll on a ship are found by an integration over the ship length of the two-dimensional values: Z 0 Xw4 = Xw4 ¢ dxb L According to the ”Ordinary Strip Theory” the exciting wave moments for roll on a restrained cross section of a ship in waves are de…ned by: D n 0 _¤ o 0 0 0 0 _¤ M42 ¢ ³ w2 + N42 ¢ ³ w2 + OG ¢ Xw2 Xw4 = +XF K4 + Dt ¶ µ 0 dM42 ¤ 0 0 0 0 _¤ Ä ¢ ³ w2 + OG ¢ Xw2 = XF K4 + M42 ¢ ³ w2 + N42 ¡ V ¢ dxb According to the ”Modi…ed Strip Theory” these moments become: ½µ ¶ ¾ D i 0 0 0 0 0 _¤ M42 ¡ Xw4 = +XF K4 + ¢ N42 ¢ ³ w2 + OG ¢ Xw2 Dt !e µ ¶ µ ¶ 0 0 V dN42 0 0 Ĥ + N 0 ¡ V ¢ dM42 ¢ ³ ¤ + OG ¢ X 0 _ = XF K4 + M42 + 2 ¢ ¢ ³ w2 42 w2 w2 ! e dxb dxb The Froude-Krilov moment in the roll direction .

From the previous follows the total wave loads for roll: Z 0 XF K4 ¢ dxb Xw4 = L + ¯ ¯ ¯ ¯ Z 0 ¯ V ¯ dN42 Ĥ +¯ ¢ ³ w2 ¢ dxb ¯ ¯ ! ¢ !e ¯ dxb ¯ ¯ L ! Z ï ¯ 0 ¯!¯ 0 dM42 _¤ ¯ ¯¢N42 ¡ V ¢ + ¢ ³ w2 ¢ dxb ¯ !e ¯ dxb L Z L 0 Ĥ M42 ¢ ³ w2 ¢ dxb +OG ¢ Xw2 The ”Modi…ed Strip Theory” includes the outlined terms.190 CHAPTER 7.4: Wave Pressure Distriubution on a Cross Section for Roll After neglecting the second order terms. EXCITING WAVE LOADS Figure 7. . this can be written as: µ ¶ CCy CSy XF K4 = ½ ¢ ¡ ¡ + CIz ¢ (¡k 2 g sin ¹) ¢ ³ a sin(! e t ¡ kxb cos ¹) k k CCy = 2 ¢ CSy = 2 sin(¡kyw sin ¹) ¡kyw sin ¹ with: (¡kyw sin ¹)2 ¡ cos(¡kyw sin ¹) 3 ¢ yw ¡T Z0 Z0 sin(¡kyb sin ¹) cosh k(h + zb ) ¢ ¢ yb ¢ zb ¢ dzb ¡kyb sin ¹ cosh(kh) sin(¡kyb sin ¹) ¡kyb sin ¹ CIz = 2 ¡ cos(¡kyb sin ¹) cosh k(h + zb ) 3 ¢ ¢ yb ¢ dzb (¡kyb sin ¹)2 cosh(kh) ¡T For deep water. the cosine-hyperbolic expressions in here reduce to exponentials. When ignoring the outlined terms the ”Ordinary Strip Theory” is presented.

1.1. 7. bG is the vertical distance of the centre of gravity of the ship G above the centroid b of the local submerged sectional area. From this follows the total wave loads for pitch: Z 0 Ĥ Xw5 = ¡ M11 ¢ bG ¢ ³ w1 ¢ dxb ¯ ¯ ¯ ¯ Z 0 ¯ ¡V ¯ dN11 ¤ Ä +¯ ¢ bG ¢ ³ w1 ¢ dxb ¯ ¯ ! ¢ !e ¯ dxb ¯ ¯ L ï ¯ ! Z 0 ¯!¯ 0 _ ¯ ¯¢N11 ¡ V ¢ dM11 ¢ bG ¢ ³ ¤ ¢ dxb ¡ w1 ¯ !e ¯ dxb L Z 0 ¡ XF K1 ¢ bG ¢ dxb ¡ L Z L 0 Ĥ M33 ¢ xb ¢ ³ w3 ¢ dxb L ¯ ¯ ¯ ¯ Z 0 ¯ ¡V ¯ dN33 ¤ +¯ ¢ xb ¢ Äw3 ¢ dxb ¯ ³ ¯ ! ¢ !e ¯ dxb ¯ ¯ L ! Z ï ¯ 0 ¯!¯ 0 dM33 _¤ ¯ ¯¢N ¡ V ¢ ¡ ¢ xb ¢ ³ w3 ¢ dxb ¯ ! e ¯ 33 dxb L Z 0 ¡ XF K3 ¢ xb ¢ dxb L The ”Modi…ed Strip Theory” includes the outlined terms.6 Exciting Wave Moments for Yaw The exciting wave moments for yaw are found by an integration over the ship length of the two-dimensional contributions of sway into the yaw moment: Z 0 Xw6 = Xw6 ¢ dxb L . CLASSICAL APPROACH 191 7.5 Exciting Wave Moments for Pitch The exciting wave moments for pitch are found by an integration over the ship length of the two-dimensional contributions of surge and heave into the pitch moment: Z 0 Xw5 = Xw5 ¢ dxb L with: 0 0 0 Xw5 = ¡Xw1 ¢ bG ¡ Xw3 ¢ xb In here.1.7. When ignoring the outlined terms the ”Ordinary Strip Theory” is presented.

When ignoring the outlined terms the ”Ordinary Strip Theory” is presented.192 with: CHAPTER 7. EXCITING WAVE LOADS 0 0 Xw6 = +Xw2 ¢ xb From this follows the total wave loads for yaw: Z ¤ 0 Xw6 = + M22 ¢ xb ¢ Äw2 ¢ dxb ³ L ¯ ¯ ¯ ¯ Z 0 ¯ V ¯ dN22 Ĥ ¢ dxb ¯ ¯ +¯ ¢ xb ¢ ³ w2 ¯ dxb ¯ ! ¢ !e ¯ L Z ¯ ¯ 0 ¯!¯ dM22 0 _¤ + ¯ ¯ ¢ N22 ¡ V ¢ ¢ xb ¢ ³ w2 ¢ dxb ¯ !e ¯ dxb L Z 0 + XF K2 ¢ xb ¢ dxb L The ”Modi…ed Strip Theory” includes the outlined terms. .

The radiated damping waves have an amplitude ³ ja and ywl is half the breadth of the cross section at the waterline. a44 = 8.2. 1949] and N-parameter conformal mapping. The phase lag "Áxj is de…ned as the phase lag between the velocity potential of the ‡uid © and the forced motion xj .7. heave and roll directions i. The potential coe¢cients Aij and Bij and the phase lags "Áxj . expressed in terms of conformal mapping coe¢cients. can be obtained from the 2-D velocity potentials and the linearized equations of Bernoulli. The velocity potentials have been obtained by using the work of [Ursell.2 Equivalent Motions of Water Particles In the classic relative motion theory. exercised by the ‡uid on a cross section of the cylinder. as given by [Journée and van ’t Veer. 4 The 2-D hydrodynamic loads Xhi in the sway. The cylinder is forced to carry out a simple harmonic oscillation about its initial position with a frequency of oscillation ! and a small amplitude of displacement xja : xj = xja cos !t 0 for: j = 2. is described now.1 Hydromechanical Loads Suppose an in…nite long cylinder in the still water surface of a ‡uid.2. These hydrodynamic loads are: Xhi = 2½gywl 0 g³ ja [Aij cos (!t + "Áxj ) + Bij sin (!t + "Áxj )] ¼ in where j is the mode of oscillation and i is the direction of the load. the average (or equivalent) motions of the water particles around the cross section are calculated from the pressure distribution in the undisturbed waves on this cross section. An alternative approach . EQUIVALENT MOTIONS OF WATER PARTICLES 193 7.based on di¤raction of waves ..to determine the equivalent accelerations and velocities of the water particles around the cross section. . a33 = 2. a42 = 4ywl and for the terms P0j and Q0j : xja ! 2 ¼ ywl sin "Áxj ³ ja g xja ! 2 = + ¼ ywl cos "Áxj ³ ja g P0j = ¡ Q0j The phase lag "Áxj between he velocity potentials and the forced motion is incorporated in the coe¢cients P0j and Q0j and can be obtained by using: µ ¶ ¡P0j "Áxj = arctan +Q0j This equation will be used further on for obtaining wave load phases. a24 = 4=ywl . 1995]. 0 These loads Xhi can be expressed in terms of in-phase and out-phase components with the harmonic oscillations: µ ¶ g³ ja 2 ½aij 0 Xhi = 2 [(Aij Q0j + Bij P0j ) cos !t + (Aij P0j ¡ Bij Q0j ) sin !t] ! xja ¼ with a22 = 2. 3. are given in a previous chapter. 7.

7. Note that the phase lag information "Áxj is vanished here. b24 = ywl . 3. cwave . 4 xj = xja cos !t ³ ´ 0 0 0 Xhj = Xhja cos !t + "hj 0 0 0 0 = Xhja cos "hj cos !t ¡ Xhja sin "hj sin !t The energy required for this oscillation should be equal to the energy radiated by the damping waves: . of regular waves are de…ned by: cwave = and cgroup = Consider a cross section which is harmonic oscillating with a frequency ! = 2¼=T and an 0 amplitude xja in the direction j in previously still water by an oscillatory force Xhj in the same direction j: for: j = 2.2 Energy Considerations ! k cwave 2kh ¢ 2 sinh [2kh] The wave velocity.2. these hydrodynamic loads are expressed in terms of potential mass and damping coe¢cients: Xhi = ¡Mij xj ¡ Nij xj Ä _ 2 = Mij ! xja cos !t + Nij !xja sin !t with: Aij Q0j + Bij P0j 2 P0j + Q2 0j Aij P0j ¡ Bij Q0j = ½bij ¢! 2 P0j + Q2 0j 0 Mij = ½bij Nij 2 3 2 4 3 with b22 = 2ywl . one may write for the roll damping coe¢cient: ¡ 0 ¢2 ¡ 0 ¢2 N24 N42 0 N44 = = 0 0 N22 N22 This relation . b44 = 2ywl and b42 = 2ywl . 1965] has used the following potential damping coupling coe¢cients in his formulation of the hydrodynamic loads for roll: 0 N42 0 N = 044 lw 0 and N24 = N22 ¢ lw 0 0 0 in which lw is the lever of the rolling moment. and the group velocity. cgroup .will be used further on for obtaining the wave loads for roll from those for sway.194 CHAPTER 7.which has been con…rmed by numerical calculations with SEAWAY . EXCITING WAVE LOADS Generally. 0 0 Because N42 = N24 . b33 = 2ywl . [Tasai.

EQUIVALENT MOTIONS OF WATER PARTICLES 195 1 T or: ZT 0 Xhj 0 1 ¢ xj dt = _ T ZT 0 1 0 Njj xj ¢ xj dt = 2 ¢ ½g³ 2 ¢ cgroup _ _ a 2 1 0 1 0 0 Xhja !xja sin "hj = Njj ! 2 x2 = ½g³ 2 ¢ cgroup a ja 2 2 From the …rst part of this equation follows Xhja sin "hj xja 0 = Njj ! ³a ³a From the second part of this equation follows the amplitude ratio of the oscillatory motions and the radiated waves: s xja 1 2½g ¢ cgroup = 0 ³a ! Njj Combining these two last equations provides for the out-phase part . 4 = 2½g ¢ cgroup ¢ Njj ³a ³ ´ 0 = xwja cos !t + "wj ³ ´ 0 0 = Xwja cos !t + "wj for: j = 2.2. "wj . an approximation has to be found. Then the energy consumed by this oscillation is equal to the energy supplied by the incoming waves.3 Wave Loads Consider now the opposite case: the cross section is restrained and is subject to regular incoming beam waves with an amplitude ³ a . for the phase lag of the wave force. in which "wj is the phase lag with respect to the wave surface elevation at the center of the cross section. 0 7. 3. 4 .so the damping part . xwj Xwj 0 0 which is in principle the same equation as the previous one for the out-phase part of the oscillatory force in still water. which will be in phase with its velocity (damping waves). 0 However. 4 In here. 3.7.2. 3. This leads for the amplitude of the exciting wave force to: 0 Xwja q 0 for: j = 2. Xhja sin "hj is the in-phase with the velocity part of the exciting force or moment.of the oscillatory force: Xhja sin "hj q 0 = 2½g ¢ cgroup ¢ Njj ³a 0 0 0 0 0 for: j = 2. The resulting wave force is caused by these motions. Let xwj represents the equivalent (or average) oscillation of the water particles with respect to the restrained cross section.

5: Vector Diagrams of Wave Components for Sway and Heave Heave Mode The vertical wave force on a restrained cross section in waves is: ³ ´ 0 0 0 Xw3 = Xw3a cos !t + "w3 0 0 = Xw3a cos "w3 cos !t ¡ Xw3a sin "w3 sin !t Xw3a = ³ a 0 0 0 0 of which the amplitude is equal to: q 2½g ¢ cgroup ¢ N33 0 For the phase lag of this wave force. has been approximated by this phase: "w3 = "wR 3 = kywl Then. the in-phase and out-phase parts of the wave loads are: 0 XF K3 0 0 0 0 + Xw31 = +Xw3a cos "w3 Xw32 = ¡Xw3a sin "w3 0 0 0 0 0 0 ¶ µ q 0 0 = + ³ a 2½g ¢ cgroup ¢ N33 cos "w3 µ q ¶ 0 0 = ¡ ³ a 2½g ¢ cgroup ¢ N33 sin "w3 . yb = ywl .196 CHAPTER 7. 0 The phase lag of a radiated wave. "w3 . is "wR 3 = kywl The phase lag of the wave force. EXCITING WAVE LOADS Figure 7. at the intersection of the ship’s hull with the waterline. "w3 . an approximation has to be found. "wR 3 .

an approximation has to be found. is equal to: "w2 (! ! 1) = ¡"wI 2 The acceleration and velocity amplitudes of the water particles in the undisturbed surface of the incoming waves are: ³ 0´ a2 = ¡kg sin ¹ still water surface 0 ³ 0´ ¡a2 kg sin ¹ v2 = = ! ! still water surface 0 0 .so at high wave frequencies ! ! 1 . Xw31 and Xw32 follow. the equivalent acceleration and velocity amplitudes of the water particles are: 0 0 0 a3 ¹ v3 ¹ Sway Mode 0 Xw31 = 0 M33 0 Xw32 = 0 N33 0 The horizontal wave force on a restrained cross section in beam waves is: ³ ´ 0 0 0 Xw2 = Xw2a cos !t + "w2 0 0 = Xw2a cos "w2 cos !t ¡ Xw2a sin "w2 sin !t 0 0 0 of which the amplitude is equal to: Xw2a = ³ a 0 For the phase lag of this wave force. EQUIVALENT MOTIONS OF WATER PARTICLES from which the di¤raction terms. These di¤raction terms can also be written as: Xw31 = M33 ¢ a3 ¹ 0 0 0 Xw32 = M33 ¢ v3 ¹ 0 0 0 0 0 197 in which a3 and v3 are the equivalent amplitudes of the acceleration and the velocity of ¹ ¹ the water particles around the cross section. Herewith. yb = ywl . "wI 2 . the phase lag of the wave force.7. "w2 .2.the ship’s hull behaves like a 0 vertical wall and all waves will be di¤racted. is "wI 2 = ¡kywl sin ¹ 0 0 0 q 2½g ¢ cgroup ¢ N22 0 if sin ¹ < 0 then: "wI 2 = "WI 2 + ¼ In very short waves . "w2 . Then. 0 The phase lag of an incoming undisturbed wave. at the intersection of the ship’s hull with the waterline.

198

CHAPTER 7. EXCITING WAVE LOADS

In very long waves - so at low wave frequencies ! ! 0 - the wave force is dominated by the Froude-Krylov force and the amplitudes of the water particle motions do not change 0 very much over the draft of the section. Apparently, the phase lag of the wave force, "w2 , can be approximated by: " # 0 0 XF K2 + M22 ¢ (¡kg sin ¹) 0 ¡ ¢ "w2 (! ! 0) = ¡ arctan 0 N22 ¢ kg sin ¹ !
0 0

When plotted against !, the two curves "w2 (! ! 0) and "w2 (! ! 1) will intersect each 0 other. The phase lag of the wave force, "w2 , can now be approximated by the lowest of these two values:
0 0

"w2 = "w2 (! ! 1) 0 0 0 0 if "w2 (! ! 0) > "w2 (! ! 1) then: "w2 = "w2 (! ! 0) Because "w2 (! ! 1) goes to zero in the low frequency region and "w2 (! ! 0) can have values between 0 and 2¼, one simple precaution has to be taken: if "w2 (! ! 0) >
0 0 0

¼ 0 0 then: "w2 (! ! 0) = "w2 (! ! 0) ¡ 2¼ 2

Now the in-phase and out-phase terms of the wave force in beam waves are: ¶ µ q 0 0 = ¡ ³ a 2½g ¢ cgroup ¢ N22 sin "w2 µ q ¶ 0 0 = + ³ a 2½g ¢ cgroup ¢ N22 cos "w2
0

0 XF K2

+ Xw21 = ¡Xw2a sin "w2 Xw22 = +Xw2a cos "w2
0 0 0 0

0

0

0

from which the di¤raction terms, Xw21 and Xw22 follow. These terms can also be written as:
0 0 0

Xw21 = M22 ¢ a2 ¹ 0 0 0 Xw22 = N22 ¢ v2 ¹ in which a2 and v2 are the equivalent amplitudes of the acceleration and the velocity of ¹ ¹ the water particles around the cross section. Then - when using an approximation for the in‡uence of the wave direction - the equivalent acceleration and velocity amplitudes of the water particles are: Xw21 = ¢ jsin ¹j 0 M22 0 Xw22 = ¢ jsin ¹j 0 N22
0 0 0

a2 ¹ v2 ¹
0

0

7.2. EQUIVALENT MOTIONS OF WATER PARTICLES Roll Mode

199

The ‡uid is free of rotation; so the wave moment for roll consists of sway contributions only. However, the equivalent amplitudes of the acceleration and the velocity of the water particles will di¤er from those of sway. From a study on potential coe¢cients, the following relation between sway and roll damping coe¢cients has been found: ¡ 0 ¢2 ¡ 0 ¢2 N24 N42 0 N44 = = 0 0 N22 N22 The horizontal wave moment on a restrained cross section in beam waves is: ³ ´ 0 0 0 Xw4 = Xw4a cos !t + "w4
0

of which the amplitude is equal to:

Xw4a = ³ a

The in-phase and out-phase parts of the wave moment in beam waves are:
0 XF K4
0

q 0 2½g ¢ cgroup ¢ N44 s ¡ 0 ¢2 N24 = ³ a 2½g ¢ cgroup ¢ 0 N22 ¯ 0 ¯ q ¯N ¯ 0 = ³ a 2½g ¢ cgroup ¢ N22 ¢ 24 0 N22 ¯ 0 ¯ ¯N ¯ 0 = Xw2a ¢ 24 0 N22 ³

+ Xw41 =
0

0 XF K2
0

Xw42 = Xw22 ¢
0 0

from which the di¤raction terms, Xw41 and Xw42 follow. These terms can also be written as: X41 = M24 ¢ a24 ¹ 0 0 0 X42 = N24 ¢ v24 ¹
0 0 0

´ ¯N 0 ¯ ¯ ¯ + Xw21 ¢ 24 0 N22 ¯ 0 ¯ ¯N ¯
0

24
0

N22

in which a24 and v24 are the equivalent amplitudes of the acceleration and the velocity of ¹ ¹ the water particles around the cross section. Then - when using an approximation for the in‡uence of the wave direction - the equivalent acceleration and velocity amplitudes of the water particles are: a24 ¹ v24 ¹
0 0

0

0

Xw41 = ¢ jsin ¹j 0 M24 0 Xw42 = ¢ jsin ¹j 0 N24

0

200 Surge Mode

CHAPTER 7. EXCITING WAVE LOADS

The equivalent acceleration and velocity amplitudes of the water particles around the cross section for surge have been found from: a2 ¹ tan ¹ 0 ¡¹1 a = !
0

a1 = ¹ v1 ¹
0

0

7.3. NUMERICAL COMPARISON

201

7.3

Numerical Comparison

Figures 7.6 and 7.7 give a comparison between these sway, heave and roll wave loads on a crude oil carrier in oblique waves - obtained by the classic approach and the simple di¤raction approach, respectively - with the 3-D zero speed ship motions program DELFRAC of Pinkster; see [Dimitrieva, 1994].

Figure 7.6: Comparison of Classic Wave Loads with DELFRAC Data

Figure 7.7: Comparison of Simple Di¤raction Wave Loads with DELFRAC Data

202 .

CHAPTER 7. EXCITING WAVE LOADS

Chapter 8 Transfer Functions of Motions After dividing the left and right hand terms by the wave amplitude ³ a . given by [Zwaan.wbmt.shipmotions. which is the motion amplitude to wave amplitude ratio. For updates see web site: http://dutw189. two sets of six coupled equations of motion are available. 8. ”Theoretical Manual of SEAWAY.J. Report 1216a. has been used in the strip theory program SEAWAY-DELFT.tudelft. Delft University of Technology. Mekelweg 2.M. Journée. Ship Hydromechanics Laboratory.nl/~johan or http://www. The variables in the coupled equations for the vertical plane motions are: Surge: Heave: Pitch: xa ³a za ³a µa ³a ¢ cos "x³ ¢ cos "z³ ¢ cos "µ³ ¢ cos "y³ ¢ cos "Á³ ¢ cos "ó and and and xa ³a za ³a µa ³a ¢ sin "x³ ¢ sin "z³ ¢ sin "µ³ ¢ sin "y³ ¢ sin "Á³ ¢ sin "ó The variables in the coupled equations for the horizontal plane motions are: Sway: Roll: Yaw: ya ³a Áa ³a Ãa ³a and and and ya ³a Áa ³a Ãa ³a These sets of motions have to be solved by a numerical method. 1977]. and the phase lags of the motions relative to the wave elevation at the ship’s centre of gravity: xa ³a The associated phase lags are: "x³ 0 ya ³a "y³ za ³a "z³ µa ³a "µ³ Áa ³a "Á³ Ãa ³a "ó J. February 2001. Release 4.1 Centre of Gravity Motions From the solutions of these in and out of phase terms follow the transfer functions of the motions. A method which provides continuous good results. 2628 CD Delft. The Netherlands.19”.nl. 203 .

with motions dominated by the mass term. The harmonic displacements in the ship-bound xb .1: Frequency Regions and Motional Behavior With the six centre of gravity motions.204 CHAPTER 8.2 Absolute Displacements Consider a point P (xb .or in the earth bound x.or in the earth bound . zb ) on the ship in the G(xb . the harmonic motions in the ship-bound xb . with motions dominated by the damping term and ² the high frequency region (! 2 À c=a). The transfer functions of the rotations can be made non-dimensional by dividing the amplitude of the rotations by the amplitude of the wave slope k³ a in lieu of the wave amplitude ³ a: xa ya za µa Ãa Áa ³a ³a ³a k³ a k³ a k³ a For motions with a spring term. y and z directions . yb . 8. yb . Figure 8. zb ) on the ship can be calculated. three frequency regions can be distinguished: ² the low frequency region (! 2 ¿ c=(m + a)).in any point P (xb . zb ) ship-bound axes system. TRANSFER FUNCTIONS OF MOTIONS The transfer functions of the translations are non-dimensional. ² the natural frequency region (! 2 t c=(m + a)). An example for heave is given in …gure 8. with motions dominated by the restoring spring term. yb and zb directions . yb and zb directions . yb .1.

y and z directions .25 0.in a point P (xb . 8. ABSOLUTE VELOCITIES 205 x.50 0. 8.0 0. yb .00 Wave Frequency (rad/s) Wave Frequency (rad/s) Figure 8.75 1. y and z directions .00 -270 -360 -450 -540 -630 -720 0 0.0 1. yb and zb directions .4.75 1.75 1.4.3.50 0.3 and 8.5 Crude Oil Carrier V = 0 kn Pitch µ = 180 0 RAO Heave (-) µ = 90 0 0.5 Crude Oil Carrier V = 0 kn Heave 1.00 Wave Frequency (rad/s) Wave Frequency (rad/s) 0 -90 µ = 90 0 0 -90 -180 µ = 180 µ = 90 0 0 Phase εz ζ (deg) -180 -270 -360 -450 -540 -630 -720 0 0.25 0. zb ) on the ship are obtained by taking the derivative of the three harmonic displacements. The harmonic longitudinal displacement is given by: xP = x ¡ yb ¢ Ã + zb ¢ µ = xPa ¢ cos(! e t + "xP ³ ) The harmonic lateral displacement is given by: yP = y + x b ¢ Ã ¡ z b ¢ Á = yPa ¢ cos(! e t + "yP ³ ) The harmonic vertical displacement is given by: zP = z ¡ xb ¢ µ ¡ yb ¢ Á = zPa ¢ cos(! e t + "zP ³ ) Some examples of calculated transfer functions of a crude oil carrier and a containership are given in the …gures 8.in a point P (xb .8.50 µ = 180 0 Phase εθ ζ 0. zb ) on the ship can be obtained from the six centre of gravity motions as presented below.3 Absolute Velocities The harmonic velocities in the ship-bound xb .2: Heave and Pitch of a Crude Oil Carrier.5 µ = 180 0 RAO Pitch (-) 1.5 µ = 90 0 0 0 0. V = 0 Knots Notify the di¤erent speed e¤ects for roll and pitch in …gure 8.2.25 0. yb .75 1. .50 0.25 0. 1.00 0 0 0.or in the earth bound x.

50 µ = 180 0 Phase εθ ζ -270 -360 -450 -540 -630 µ = 90 0 0.25 0.75 5 0.50 V = 0 knots 1.8 1.25 0.50 0.5 µ = 90 0 0 0 0.0 µ = 180 0 0.0 0 0 Containership Lpp = 175 metre 0.75 1.25 Containership Lpp = 175 metre 0 0 0.0 wave frequency (rad/s) wave frequency (rad/s) Figure 8.50 0.4 0.0 1. V = 16 Knots 15 RAO of roll Beam waves 1.25 0.00 Wave Frequency (rad/s) Wave Frequency (rad/s) 0 -90 µ = 90 0 0 -90 -180 µ = 180 0 Phase εz ζ (deg) -180 -270 -360 -450 -540 -630 -720 0 0.00 -720 0 0. TRANSFER FUNCTIONS OF MOTIONS 1. RAO of pitch (-) V = 10 knots 1.00 V = 10 knots V = 0 knots V = 20 knots 0.00 Wave Frequency (rad/s) Wave Frequency (rad/s) Figure 8.2 0.25 0.3: Heave and Pitch of a Crude Oil Carrier.00 0 0 0.75 1.5 µ = 180 0 0.6 0.50 0.5 Crude Oil Carrier V = 16 kn Heave 0 1.5 Crude Oil Carrier V = 16 kn Pitch RAO Heave (-) µ = 90 RAO Pitch (-) 1.4 0.75 1.4: RAO’s of Roll and Pitch of a Containership .206 CHAPTER 8.25 RAO of pitch Head waves V = 20 knots Non-dim.50 0.6 0.75 1.8 1.2 0. RAO of roll (-) 10 Non-dim.

O(x. the harmonic accelerations in the x.1 Accelerations in the Earth-Bound Axes System In the earth-bound axes system.4. y. Thus: ² Longitudinal acceleration: Ä xP = x ¡ yb ¢ Ã + zb ¢ Ä Ä Ä µ 2 = ¡! e ¢ xPa ¢ cos(! e t + "xP ³ ) = xPa ¢ cos(! e t + "xP ³ ) Ä Ä Ä Ä yP = y + xb ¢ Ã ¡ zb ¢ Á Ä Ä = ¡! 2 ¢ yPa ¢ cos(! e t + "yP ³ ) e = yPa ¢ cos(! et + "yP ³ ) Ä Ä Ä zP = z ¡ xb ¢ Ä + yb ¢ Á Ä Ä µ = ¡! 2 ¢ zPa ¢ cos(! et + "zP ³ ) e = zPa ¢ cos(! e t + "zP ³ ) Ä Ä ² Lateral acceleration: ² Vertical acceleration: . yb . y and z direction in a point P (xb . 8. In the ship-bound axes system. z). zb ) on the ship are obtained by taking the second derivative of the three harmonic displacements.8.4. a component of the acceleration of gravity has to be added to the accelerations in the horizontal plane direction.4 Absolute Accelerations In the earth-bound axes system. the harmonic accelerations on the ship are obtained by taking the second derivative of the displacements. ABSOLUTE ACCELERATIONS The harmonic longitudinal velocity is given by: _ _ xP = x ¡ yb ¢ Ã + zb ¢ µ _ _ = ¡! e ¢ xPa ¢ sin(! e t + "xp ³ ) = xPa ¢ cos(! e t + "xp ³ ) _ _ The harmonic lateral velocity is given by: _ _ yP = y + xb ¢ Ã ¡ zb ¢ Á _ _ = ¡! e ¢ yPa ¢ sin(! et + "yp ³ ) = yPa ¢ cos(! e t + "yp ³ ) _ _ The harmonic vertical velocity is given by: _ _ zP = z ¡ xb ¢ µ + yb ¢ Á _ _ = ¡! e ¢ zPa ¢ sin(! e t + "zp ³ ) = zPa ¢ cos(! et + "zp ³ ) _ _ 207 8.

G(xb . The vertical acceleration does not change. Notify the di¤erent behavior of absolute and relative vertical motioins as given in …gure 8. TRANSFER FUNCTIONS OF MOTIONS 8.4. A dynamical swell up should be taken into account. which is not included in the previous formulation.208 CHAPTER 8.2 Accelerations in the Ship-Bound Axes System In the ship-bound axes system.5 Vertical Relative Displacements The harmonic vertical relative displacement with respect to the wave surface of a point P (xb . zb ) connected to the ship can be obtained too: sP = ³ P ¡ z + xb ¢ µ ¡ yb ¢ Á = sPa ¢ cos(! et + "sP ³ ) with: ³ P = ³ a cos(! e t ¡ kxb cos ¹ ¡ kyb sin ¹) It may be noted that the sign of the relative motion is chosen here in such a way that a positive relative displacement implies a decrease of the freeboard. Thus: ² Longitudinal acceleration: Ä Ä µ xP = x ¡ yb ¢ à + zb ¢ Ä ¡ g ¢ µ Ä 2 = ¡! e ¢ xPa ¢ cos(! e t + "xP ³ ) ¡ g ¢ µa ¢ cos(! et + "µ³ ) = xPa ¢ cos(! e t + "xP ³ ) Ä Ä ² Lateral acceleration: Ä Ä yP = y + xb ¢ à ¡ zb ¢ Á + g ¢ Á Ä Ä = ¡! 2 ¢ yPa ¢ cos(! e t + "yP ³ ) + g ¢ Áa ¢ cos(! e t + "Á³ ) e = yPa ¢ cos(! e t + "yP ³ ) Ä Ä ² Vertical acceleration: Ä zP = z ¡ xb ¢ Ä + yb ¢ Á Ä Ä µ = ¡! 2 ¢ zPa ¢ cos(! et + "zP ³ ) e = zPa ¢ cos(! e t + "zP ³ ) Ä Ä 8. yb .5. . These are the accelerations that will be ”felt” by for instance the cargo on the ship. zb ). a component of the acceleration of gravity g has to be added to the accelerations in the longitudinal and lateral direction in the earthbound axes system. yb . An oscillating ship will produce waves and these phenomena will change the relative motion.

0 1. yb .0 RAO tends to 0.0 1.5 wave frequency (rad/s) wave frequency (rad/s) Figure 8.8.6.6 Vertical Relative Velocities The harmonic vertical relative velocity with respect to the wave surface of a certain point P (xb .0 1.5 0 0 0. can be obtained by: sP = _ D f³ ¡ z + xb ¢ µ ¡ yb ¢ Ág Dt P _ _ _ _ = ³ P ¡ z + xb ¢ µ ¡ V ¢ µ ¡ yb ¢ Á in which for the vertical velocity of the water surface itself: _ ³ P = ¡! ¢ ³ a sin(! e t ¡ kxb cos ¹ ¡ kyb sin ¹) . zb ).5 0 1.0 0 0.5 RAO tends to 1. connected to the ship.5: Absolute and Relative Vertical Motions at the Bow 8. VERTICAL RELATIVE VELOCITIES 5 209 5 Containership Head waves Containership Head waves V = 20 knots RAO of vertical absolute bow motions (m/m) RAO of vertical relative bow motions (m/m) 4 4 V = 10 knots 3 V = 20 knots 3 V = 10 knots 2 V = 0 knots 2 V = 0 knots 1 RAO tends to 1.0 1 RAO tends to 0.

TRANSFER FUNCTIONS OF MOTIONS . CHAPTER 8.210 .

1978]. ² Four types of anti-rolling devices and its contribution to the equations of motion are described here: ² bilge keels ² passive free-surface tanks ² active …n stabilisers ² active rudder stabilisers..J. 2628 CD Delft. Since 1880. For updates see web site: http://dutw189. Report 1216a. The Netherlands. With bilge keels they performed a …rst successful attack on the problem of rolling. a quadratic twodimensional roll damping can be de…ned.tudelft. The drag coe¢cient CD depends on the period parameter or the Keulegan-Carpenter number.19”. Ship Hydromechanics Laboratory. Journée.M. as described before.nl/~johan or http://www. Ikeda divides the two-dimensional quadratic bilge keel roll damping into a component due to the normal force on the bilge keels and a component due to the pressure on the hull surface. but in several cases these bilge keels did not prove to be su¢cient. Ikeda measured the quadratic two-dimensional drag by carrying out free rolling experiments with an ellipsoid with and without bilge keels. 0 211 . The normal force component of the bilge keel damping has been be deduced from experimental results of oscillating ‡at plates. February 2001.Chapter 9 Anti-Rolling Devices Since the disappearence of sails on oceangoing ships. Assuming a pressure distribution on the hull caused by the bilge keels. 9.nl. Delft University of Technology. A reliable method to determine this contribution is given by [Ikeda et al.shipmotions. numerous other more or less successful ideas have been tested and used. Release 4. Ikeda carried out experiments to measure the J.1 Bilge Keels Bilge keels can deliver an important contribution to an increase the damping of the rolling motions of ships. The active …n and rudder stabilisers are not build into the program SEAWAY yet. with their stabilising wind e¤ect on the rolling motions. naval architects have been concerned in reducing the rolling of ships among waves. Mekelweg 2. created by the bilge keels.wbmt. ”Theoretical Manual of SEAWAY.

¡ while the coe¢cient Cp of the pressure on the back face of the bilge keel and the length of the negative pressure region depend on the period parameter. which has been …lled until a water level h with a ‡uid with a mass density ½. Figure 9. caused by a passive free-surface tank. resonance frequencies can be obtained with high wave amplitudes at lower water depths.212 CHAPTER 9.2. 1965]. Observe a rectangular tank with a length l and a breadth b. di¤erent approaches have been used. [Verhagen and van Wijngaarden. Experiments of Ikeda showed that the e¤ect of forward speed on the roll damping due to the bilge keels can be ignored. He found that the coe¢cient Cp of the pressure on the front face of the bilge keel does not depend on the period parameter.1 shows a 2-D sketch of this tank with the axis system and notations. is essentially based on the existence of a hydraulic jump or bore in the tank. as given by [Verhagen and van Wijngaarden. 9. A theory. For low and high frequencies and the frequencies near to the natural frequency.2 Passive Free-Surface Tanks The roll damping.1: Axes System and Notations of an Oscillating Tank . as has been shown in a separate chapter. 9. 1965] give a theoretical approach to determine the counteracting moments by free-surface anti-rolling tanks. The distance of the bottom of the tank above the centre of gravity of the vessel is s. based on gasdynamics for the shock wave in a gas ‡ow under similar resonance circumstances. which travels periodically back and forth between the walls of the tank. 1966] give extended quantitative information on these moments. The equivalent linear total bilge keel damping has been obtained by linearising the result. has been adapted and used to describe the motions of the ‡uid. Figure 9. This hydraulic jump can be a strongly nonlinear phenomenon. ANTI-ROLLING DEVICES + pressure on the hull surface created by bilge keels.[Bosch and Vugts. The total bilge keel damping has been obtained by integrating these two two-dimensional roll damping components over the length of the bilge keels.1 Theoretical Approach When a tank which contains a ‡uid with a free surface is forced to carry out roll oscillations. Under these circumstances a hydraulic jump or bore is formed. Ikeda de…nes an equivalent length of a constant negative pressure region S0 over the height of the bilge keels and a two-dimensional roll damping component can be found.

the surface elevation of the ‡uid is described by: z =s+h+³ Relative to the bottom of the tank. so: ¸0 = 2b. the vertical displacement of the tankbottom is described by: z = s + yÁ and after linearisation. For low and high frequencies and the frequencies near to the natural frequency. introduced by the excitation. Their expressions for the internal wave loads are rewritten and modi…ed to be useful for any arbitrary vertical position of the centre of rotation by [Journée. With the wave number and the dispersion relation: k= 2¼ ¸ and ! = p kg tanh [kh] it follows for the natural frequency of surface waves in the tank: s · ¸ ¼g ¼h !0 = tanh b b [Verhagen and van Wijngaarden. the continuity and momentum equations are: @» @» @v +v +» = 0 @t @y @y @v @» @v +v +g + gÁ = 0 @t @y @y In these formulations. with the centre of rotation at the bottom of the container.2. Therefore. Between the surface of the ‡uid and the bottom of the tank. di¤erent approaches have been used.1 and after linearisation. PASSIVE FREE-SURFACE TANKS 213 The natural frequency of the surface waves in a harmonic rolling tank appears as the wave length ¸ equals twice the breadth b.9. 1997]. the acceleration in the zdirection. A calculation routine has been made to connect these regions. so: Áa ! 2 b ¿ g The boundary conditions for v are determined by the velocity produced in the horizontal direction by the excitation. 1965] have investigated the shallow water wave loads in a rolling rectangular container. the linearised surface elevation of the ‡uid is described by: » = h + ³ ¡ yÁ Using the shallow water theory. must be small with respect to the acceleration of gravity g. Low and High Frequencies The harmonic roll motion of the tank is de…ned by: Á = Áa sin(!t) In the axis-system of …gure 9. v denotes the velocity of the ‡uid in the y-direction and the vertical pressure distribution is assumed to be hydrostatic. .

However. is: n o (s+h)!2 µ ¶ b! 0 1 + g ¼!y ³ ´ sin » =h¡ Á ¼! b! 0 ¼! cos 2!0 Now. Obviously. the expression for the surface elevation of the ‡uid » goes to in…nity. it is required that: b _ v = ¡(s + h)Á at: y = § 2 For small values of Áa . this will result into the static moment: ¶ ¾ ½ µ b3 h + Á MÁ = ½gl bh s + 2 12 The phase lags between the roll moments and the roll motions have not been obtained here. satisfying the boundary values for v. so for the limit value ! ! 0. ANTI-ROLLING DEVICES the velocity of the ‡uid v varies between vs and vs = cosh kh with a mean velocity: vs =kh. Experiments showed the appearance of a hydraulic jump or a bore at these frequencies. in very shallow water v does not vary between the bottom and the surface. However.214 CHAPTER 9. the continuity equation and the momentum equation can be given in a linearised form: @» @v +h = 0 @t @y @v @» +g + gÁ = 0 @t @y The solution of the surface elevation » in these equations. the roll moment follows from the quasi-static moment of the mass of the frozen liquid ½lbh and an integration of » over the breadth of the tank: +b=2 µ ¶ Z h MÁ = ½glbh s + Á + ½gl »y dy 2 ¡b=2 This delivers the roll moment amplitude for low and high frequencies at small water depths: µ ¶ h Áa MaÁ = ½glbh s + 2 ½ µ ¶ ³ ´¾ ¾ ½ ³ ´ (s + h) ! 2 ¼! !0 2 !0 3 3 +½glb 1 + tan ¡ Áa ¢ 2 g ¼! 2! 0 ¼! For very low frequencies. When taking the value at the surface. then the linearised equations are not valid anymore. they can be set to zero for low frequencies and to -¼ for high frequencies: "MÁ Á = 0 "MÁ Á = ¡¼ for: ! ¿ ! 0 for: ! À ! 0 Natural Frequency Region For frequencies near to the natural frequency ! 0 . .

¼ +® 2 ¼ = ¡ ¡® 2 for: ! < ! 0 for: ! > ! 0 9. 1965] is given in …gure 9. the limits for the frequency ! are at least: r r 24gÁa 24gÁa !0 ¡ < ! < !0 + b¼ 2 b¼ 2 Comparison with Experimental Data An example of the results of this theory with experimental data of an oscillating free-surface tank by [Verhagen and van Wijngaarden.g. de…ned by: ( ) µ ¶4 r 2 3 2 2Áa h 4 ¼ b (! ¡ ! 0 ) lb MaÁ = ½g ¢ 1¡ 12 ¼ 3b 32gÁa The phase lags between the roll moment and the roll motion at small water depths are given by: "MÁ Á = ¡ "MÁ Á with: 8s 9 < ¼ 2 b (! ¡ ! )2 = 0 ® = 2 arcsin : . With their symbols.2. 1966] have described the physical behaviour of passive free-surface tanks. they found a roll moment amplitude.2 Experimental Approach [Bosch and Vugts. used as an anti-rolling device. has been provided. e. A calculation routine connects these three regions. PASSIVE FREE-SURFACE TANKS 215 Verhagen and van Wijngaarden solved the problem by using the approach in gas dynamics when a column of gas is oscillated at a small amplitude.2. by a piston. At frequencies near to the natural frequency at small water depths. 24gÁa (s ) ¼ 2 b (! ¡ ! 0 )2 ¡ arcsin 96gÁa ¡ 3¼ 2 b (! ¡ ! 0 )2 Because that the arguments of the square roots in the expression for "MÁ Á have to be positive.2. Extended quantitative information on the counteracting moments. caused by the water transfer in the tank. The roll moments have been calculated here for low and high frequencies and for frequencies near to the natural frequency of the tank. the roll motions and the exciting moments of an oscillating rectangular free-surface tank. are de…ned by: ' = 'a cos(!t) Kt = Kta cos(!t + "t ) and the dimensions of the rectangular free-surface tank are given by: .9.

08 and 0. 0. oscillating free surface tank can be written as: Kt = a4' ¢ ' + b4' ¢ ' + c4' ¢ ' Ä _ with: a4' = 0 b4' = c4' Kta 'a ! Kta = ¢ cos "t 'a ¢ sin "t .1000 radians s=b = -0.0333. with a frequency !.06. 1966] have presented extended experimental data of: Kt and "t ¹a = ¤ a 3 ½ glb for: 'a = 0.04. 0.20.40. 0.3. 0.02. -0.216 CHAPTER 9.0667 and 0. taken from the report of [Bosch and Vugts. [Bosch and Vugts.10 An example of a part of these experimental data has been shown for s=b = ¡0:40 and 'a = 0:1000 radians in …gure 9. 1966]. ANTI-ROLLING DEVICES Figure 9.2: Comparison between Theoretical and Experimental Data l b s h ½¤ = = = = = length of the tank breadth of the tank distance of tank bottom above rotation point water depth in the tank at rest mass density of the ‡uid in the tank A non-dimensional frequency range is de…ned by: s b 0:00 < ! ¢ < 1:60 g In this frequency range.00 and +0. the external roll moment due to an. When using these experimental data.20 h=b = 0. 0.

PASSIVE FREE-SURFACE TANKS 217 Figure 9.3: Experimental Data on Anti-Rolling Free-Surface Tanks It is obvious that for an anti-rolling free-surface tank.9. it holds: Áa = 'a So it can be written: Á = Áa cos(! e t + "Á³ ) Kt = Kta cos(! e t + "Á³ + "t ) Then. an additional moment has to be added to the right hand side of the equations of motion for roll: Ä _ Xtank4 = a44tank ¢ Á + b44tank ¢ Á + c44tank ¢ Á with: a44tank = 0 b44tank = c44tank = Kta Áa and !e = ! !e ¢ sin "t Kta ¢ cos "t Áa . build into a ship.2.

5 1.218 CHAPTER 9. 1978] and [Lloyd.4 shows the signi…cant reduction of the roll transfer functions and the signi…cant roll amplitude of a trawler. b44tank and c44tank have to be subtracted from the coe¢cients a44 .5 2.0 1. 1989].4: E¤ect of Free-Surface Tanks on Roll 9. use has been made here of reports published by [Schmitke.3 Active Fin Stabilisers To determine the e¤ect of active …n stabilisers on ship motions. 40 Trawler L = 23. ANTI-ROLLING DEVICES This holds that the anti-rolling coe¢cients a44tank .0 2. caused by an oscillating …n pair are given by: Ä Xf in2 = a2¯ ¢ ¯ + b2¯ Ä Xf in = a4 ¢ ¯ + b4 4 ¯ ¯ Xf in6 = a6¯ with: Ä ¢ ¯ + b6¯ _ ¢ ¯ + c2¯ ¢ ¯ _ ¢ ¯ + c4¯ ¢ ¯ _ ¢ ¯ + c6¯ ¢ ¯ a2¯ = ¡2 sin ° ¢ a¯ b2¯ = ¡2 sin ° ¢ b¯ ) .90 metre Without tank 30 25 30 Trawler L = 23.2.5 0 0 1 2 3 4 5 6 circular wave frequency (1/s Significant wave height (m ) Figure 9. b44 and c44 in the left hand side of the equations of motion for roll. The ocillatory angle of the portside …n is given by: ¯ = ¯ a cos(! e t + ²¯Á ) The exciting forces and moments. 9.90 metre Without tank Significant roll amplitude (deg) Transfer function roll (deg/m) 20 20 15 With tank With tank 10 10 5 0 0 0. being obtained by a free-surface tank.3 E¤ect of Free-Surface Tanks Figure 9.

so ¤=0 or cos ¤=1. This e¤ect is translated into a reduced lift curve slope of the …n.3. The …n acts in the boundary layer of the ship. .9. which will reduce the lift. the sweep angle of the …n pro…le is zero. ACTIVE FIN STABILISERS c2¯ a4¯ b4¯ c4¯ a6¯ b6¯ c6¯ and: a¯ = b¯ c¯ In here: µ ¶ ° = angle of port …n = lift curve slope of …n f in 219 = = = = = = = ¡2 sin ° ¢ c¯ +2(ybfin cos ° + zbfin sin °) ¢ a¯ +2(ybfin cos ° + zbfin sin °) ¢ b¯ +2(ybfin cos ° + zbfin sin °) ¢ c¯ ¡2xbf in sin ° ¢ a¯ ¡2xbf in sin ° ¢ b¯ ¡2xbf in sin ° ¢ c¯ ³ c ´3 1 f in ¢¼ ½sf in ¢ 2 2 Ã ! µ ¶ cf in @CL 1 = ½V Af in ¢ ¼ + ¢ C(k) 2 2 @® f in ¶ µ 1 2 @CL = ½V Af in ¢ ¢ C(k) 2 @® f in @CL @® C(k) = circulation delay function ! e cr = reduced frequency k = 2V Af in = projected …n area sf in = span of …n cf in = mean chord of …n xbf in = xb -coordinate of the centroid of …n forces ybf in = yb -coordinate of the centroid of …n forces zbf in = zb -coordinate of the centroid of …n forces The nominal lift curve slope of a …n pro…le in a uniform ‡ow is approximated by: @CL = @® with: 1:80 ¢ ¼ ¢ (ARE ) q 2 E) 1:80 + cos ¤ ¢ (AR4 ¤ + 4:0 cos ¤ = sweep angle of …n pro…le (ARE ) = e¤ective aspect ratio of …n pro…le Of normal …ns.

220 CHAPTER 9. Because: ± c(±) = crf in ¡ (crf in ¡ ctf in ) ¢ sf in in which: crf in = root chord of …n ctf in = tip chord of …n crf in + ctfin mean chord of …n cf in = 2 the correction to the lift curve slope is: µ ¶ µ ¶ crf in crf in ¡ ctf in 2±BL ±2 BL EBL = ¢ 1¡ ¡ ¢ 1¡ cf in 9sf in 2cf in 8sf in Then the corrected lift curve slope of the …n is: µ ¶ @CL 1:80 ¢ ¼ ¢ (ARE )f in q = EBL ¢ @® f in 1:80 + (ARE )2 in + 4:0 f Generally a …n is mounted close to the hull. so the e¤ective aspect ratio is about twice the geometric aspect ratio: sf in (ARE )f in = 2 ¢ (AR)f in = 2 ¢ cf in . For rectangular …ns. this is simply an assumption of a uniform loading. ANTI-ROLLING DEVICES The velocity distribution in the hull boundary layer is estimated by the following two equations: r ± with: ± < ± BL V (±) = V ¢ 7 ±BL V ¢x ¡0:2 with: Rx = ± BL = 0:377 ¢ xf in ¢ Rx º in which: V (±) V ± ± BL xf in Rx º = = = = = = = ‡ow velocity inside boundary layer forward ship speed normal distance from hull thickness of boundary layer distance aft of forward perpendicular of …n local Reynolds number kinematic density of ‡uid The kinematic viscosity of seawater can be found from the water temperature T in degrees centigrade by: 1:78 m2 /s º ¢ 106 = 1:0 + 0:0336 ¢ T + 0:000221 ¢ T 2 It is assumed here that the total lift of the …n can be found from: s Zfin 1 1 ½CL V 2 (±) ¢ c(±) ¢ d± = ½CLfin ¢ V 2 ¢ Af in 2 2 0 where c(±) is the chord at spanwise-location ±.

So.4 Active Rudder Stabilizers To determine the e¤ect of rudder stabilizers on ship motions. 1989] and [Schmitke. use has been made of reports published by [Lloyd. ACTIVE RUDDER STABILIZERS 221 9. The exciting forces and moments. 1978]. a positive roll moment and a negative yaw moment. caused by this oscillating rudder are given by: _ Xr2 = a2± ¢ Ä + b2± ¢ ± + c2± ¢ ± ± _ Xr4 = a4± ¢ Ä + b4± ¢ ± + c4± ¢ ± ± _ Xr6 = a6± ¢ Ä + b6± ¢ ± + c6± ¢ ± ± with: a2± b2± c2± a4± b4± c4± a6± b6± c6± and: a± = b± c± In here: Vrudder ¼ 1:125 ¢ V µ ¶ @CL @® rudder C(k) ! e ¢ crudder k= 2V Arudder = equivalent ‡ow velocity at rudder = lift curve slope of rudder = circulation delay function = reduced frequency = projected area of rudder ³c ´3 1 rudder ½srudder ¢ ¢¼ 2 2 µ µ ¶ ¶ 1 crudder @CL ½Vrudder ¢ Arudder ¢ ¢ ¼+ = ¢ C(k) 2 2 @® rudder µ ¶ 1 2 @CL = ½V ¢ Arudder ¢ ¢ C(k) 2 rudder @® rudder = = = = = = = = = +a± +b± +c± ¡zbrudder ¢ a± ¡zbrudder ¢ b± ¡zbrudder ¢ c± +xbrudder ¢ a± +xbrudder ¢ b± +xbrudder ¢ c± .4. a positive ± results in a positive side force. The oscillatory rudder angle is given by: ± = ± a cos(! e t + "±Á ) with ± is positive in a counter-clockwise rotation of the rudder.9.

so the e¤ective aspect ratio is equal to the geometric aspect ratio: (ARE )rudder = (AR)rudder = srudder crudder The lift curve slope of the rudder is approximated by: ¶ µ @CL 1:80 ¢ ¼ ¢ (ARE )rudder p = @® rudder 1:80 + (ARE )2 rudder + 4:0 . ANTI-ROLLING DEVICES span of rudder mean chord of rudder xb -coordinate of centroid of rudder forces zb -coordinate of centroid of rudder forces Generally a rudder is not mounted close to the hull.222 srudder crudder xbrudder zbrudder = = = = CHAPTER 9.

19”. The units of these coe¢cients are N/m or kN/m.J. Cpz ). Mekelweg 2. lateral and vertical displacements of a certain point P on the ship are given by: x(P ) = x ¡ yp ¢ Ã + zp ¢ µ y(P ) = y + xp ¢ Ã ¡ zp ¢ Á z(P ) = z ¡ xp ¢ µ + yp ¢ Á The linear spring coe¢cients in the three directions in a certain point P are de…ned by (Cpx . Cpy . Release 4. Ship Hydromechanics Laboratory. Report 1216a.Chapter 10 External Linear Springs Suppose a linear spring connected to point P on the ship.wbmt. 2628 CD Delft.shipmotions.nl/~johan or http://www. J. Delft University of Technology. Journée.M.1: Coordinate System of Springs The harmonic longitudinal. The Netherlands. February 2001. Figure 10. ”Theoretical Manual of SEAWAY. 0 223 .tudelft. For updates see web site: http://dutw189.nl.

224 CHAPTER 10. EXTERNAL LINEAR SPRINGS 10. caused by these linear springs. this results into the following coe¢cients ¢cij .2 Additional Coe¢cients After a change of sign.1 External Loads The external forces and moments. ² Surge: ¢c11 ¢c12 ¢c13 ¢c14 ¢c15 ¢c16 ² Sway: ¢c21 ¢c22 ¢c23 ¢c24 ¢c25 ¢c26 ² Heave: ¢c31 ¢c32 ¢c33 ¢c34 ¢c35 ¢c36 = = = = = = 0 0 +Cpz +Cpz ¢ yp ¡Cpz ¢ xp 0 = = = = = = 0 +Cpy 0 ¡Cpy ¢ zp 0 +Cpy ¢ xp = = = = = = +Cpx 0 0 0 +Cpx ¢ zp ¡Cpx ¢ yp . which have to be added to the restoring spring coe¢cients cij of the hydromechanical loads in the left hand side of the equations of motions. acting on the ship are given by: Xs1 Xs2 Xs3 Xs4 Xs5 Xs6 = = = = = = ¡Cpx ¢ (x ¡ yp ¢ Ã + zp ¢ µ) ¡Cpy ¢ (y + xp ¢ Ã ¡ zp ¢ Á) ¡Cpz ¢ (z ¡ xp ¢ µ + yp ¢ Á) ¡Xs2 ¢ zp + Xs3 ¢ yp +Xs1 ¢ zp ¡ Xs3 ¢ xp ¡Xs1 ¢ yp + Xs2 ¢ xp 10.

heave and pitch motions will be coupled then with the sway. for an anchored platform. 10. When using linear springs. Finally. a linear superposition of the coe¢cients can be used.3 Linearized Mooring Coe¢cients Figure 10.3.10. see for instance [Korkut and Hebert. …gure 10.2-a shows the platform anchored by two anchor lines of chain at 100 m water depth. LINEARIZED MOORING COEFFICIENTS ² Roll: ¢c41 ¢c42 ¢c43 ¢c44 ¢c45 ¢c46 ² Pitch: ¢c51 ¢c52 ¢c53 ¢c54 ¢c55 ¢c56 ² Yaw: ¢c61 ¢c62 ¢c63 ¢c64 ¢c65 ¢c66 = = = = = = ¡Cpx ¢ yp +Cpy ¢ xp 0 ¡Cpy ¢ xp ¢ zp ¡Cpx ¢ yp ¢ zp 2 +Cpx ¢ yp + Cpy ¢ x2 p = = = = = = +Cpx ¢ zp 0 ¡Cpz ¢ xp ¡Cpz ¢ xp ¢ yp 2 +Cpx ¢ zp + Cpz ¢ x2 p ¡Cpx ¢ yp ¢ zp = = = = = = 0 ¡Cpy ¢ zp +Cpz ¢ yp 2 2 +Cpy ¢ zp + Cpz ¢ yp ¡Cpz ¢ xp ¢ yp ¡Cpy ¢ xp ¢ zp 225 In case of several springs.2 shows an example of results of static catenary line calculations. roll and yaw motions. . 1970].2-b shows the horizontal forces at the suspension points of both anchor lines as a function of the horizontal displacement of the platform. Because of these springs. Figure 10. This …gure shows clearly the non-linear relation between the horizontal force on the platform and its horizontal displacement. Figure 10. generally 12 sets of coupled equations with the in and out of phase terms of the motions have to be solved.2-c shows the relation between the total horizontal force on the platform and its horizontal displacement. the surge.

in the surge displacement region: ½ ¾ Total Force Cpx = M ean Displacement .2-c by determining an average restoring spring coe¢cient.226 CHAPTER 10. EXTERNAL LINEAR SPRINGS Figure 10. Cpx . can be obtained from …gure 10.2: Horizontal Forces on a Floating Structure as a Function of Surge Displacements A linear(ized) spring coe¢cient. to be used in frequency domain computations.

Delft University of Technology.J.M.wbmt. it is logical to conclude that a ship moving through still water will dissipate less energy than one moving through waves. For updates see web site: http://dutw189. Release 4. Ship Hydromechanics Laboratory. The time-averaged part of the increase of resistance is called: the added resistance due to waves. The extra wave-induced loss of energy can be treated as an added propulsion resistance. Since both wave patterns dissipate energy. 2500 Still water resistance RSW + Mean added resistance RAW Resistance 2000 Resistance (kN) 1500 Still water resistance RSW 1000 500 0 0 10 Time (s) 20 30 Figure 11.nl.19”. 2628 CD Delft. Report 1216a. The Netherlands. Figure 11.nl/~johan or http://www.Chapter 11 Added Resistances due to Waves A ship moving forward in a wave …eld will generate ”two sets of waves”: waves associated with forward speed through still water and waves associated with its vertical relative motion response to waves.shipmotions.1 shows the resistance in regular waves as a function of the time: a constant part due the calm water resistance and an oscillating part due to the motions of the ship. 0 227 . Raw .1: Increase of Resistance in Regular Waves Two theoretical methods have been used for the estimation of the time-averaged added resistance of a ship due to the waves and the resulting ship motions: J. February 2001. Journée. Mekelweg 2. ”Theoretical Manual of SEAWAY. relative to the incoming regular waves.tudelft.

Because of the added resistance of a ship due to the waves is proportional to the relative motions squared. 11. suitable for all wave directions. ADDED RESISTANCES DUE TO WAVES ² a radiated wave energy method. relative to the cross sections = period of vertical oscillation of the cross section T Z eZ 0 L b033 ¢ Vz¤ 2 ¢ dxb ¢ dt The speed dependent hydrodynamic damping coe¢cient for the vertical motion of a cross section is de…ned here as shown before: 0 b033 = N33 ¡ V ¢ 0 dM33 dxb The harmonic vertical relative velocity of a point on the ship with respect to the water particles is de…ned by: D _0 Vz = ³ w3 ¡ fz ¡ xb ¢ µ + yb ¢ Ág Dt ³ ´ _ _ _ 0w ¡ z ¡ xb ¢ µ + V ¢ µ + yb ¢ Á _ = ³ 3 . suitable for head to beam waves. its inaccuracy will be gained strongly by inaccuracies in the predicted motions.228 CHAPTER 11. 1972] as: P = in which: b033 Vz¤ Te = hydrodynamic damping coe¢cient of the vertical motion of the cross section = vertical average velocity of the water particles. 1972]. ² an integrated pressure method. 1970]. as introduced by [Gerritsma and Beukelman. The transfer function of the mean added resistance is presented as: 00 Raw = Raw ³2 a In a non-dimensional way the transfer function of the mean added resistance is presented as: Raw 00 Raw = ½g³ 2 B 2 =L a in which: L B = length between perpendiculars = maximum breadth of the waterline Both methods will be described here.1 Radiated Energy Method The radiated wave energy during one period of oscillation of a ship in regular waves is de…ned by [Gerritsma and Beukelman. as introduced by [Boese.

an equivalent harmonic vertical relative velocity has to be found. This equivalent relative velocity is de…ned by: ³ ´ _¤ _ Vz¤ = ³ w3 ¡ z ¡ xb ¢ µ + V ¢ µ _ = Vz¤ ¢ cos(! e t + "Vz¤ ³ ) a With this the radiated energy during one period of oscillation is given by: ¼ P = !e ¶ Z µ 0 dM33 0 N33 ¡ V ¢ ¢ Vz¤ 2 ¢ dxb a dxb L To maintain a constant forward ship speed. A second small contribution of the longitudinal component of the vertical hydrodynamic and wave forces has been added.2 Integrated Pressure Method [Boese.2. However. At these low frequencies. When the wave speed in following waves approaches the ship speed the frequency of encounter in the denominator tends to zero. INTEGRATED PRESSURE METHOD 229 For a cross section of the ship. so a high motion peak can be expected.11. as fast as the frequency of encounter. The energy delivered to the surrounding water is given by: c P = Raw ¢ v ¡ cos ¹ 2¼ = Raw ¢ ¡k cos ¹ µ ¶ ¢ Te From this the transfer function of the mean added resistance according to Gerritsma and Beukelman can be found: ¶ Z µ 0 V¤2 Raw ¡k cos ¹ dM33 0 = ¢ N33 ¡ V ¢ ¢ za ¢ dxb 2! e dxb ³2 ³2 a a L This method gives good results in head to beam waves. this energy should be delivered by the ship’s propulsion plant. the potential sectional mass is very high and the potential sectional damping is almost zero. This is caused by the presence of a natural frequency for heave and pitch at this low ! e . The damping multiplied with the relative velocity squared in the nominator does not tend to zero. 11. This results into extreme positive and negative added resistances. 1970] calculates the added resistance by integrating the longitudinal components of the oscillating pressures on the wetted surface of the hull. The wave elevation is given by: ³ = ³ a cos(! e t ¡ kxb cos ¹ ¡ kyb sin ¹) . A mean added resistance Raw has to be gained. in following waves this method fails.

t) = ½g ¢ 2 tanh [kh] The average horizontal force on a cross section follows from: f ¤ (xb ) = T Ze 0 f ¤ (xb . t) ¢ dt µ ¶ 2 zxa 2sa ¢ cos(¡kxb cos ¹ ¡ "s³ ) ½g³ 2 a ¢ ¡1 + 2 + = 4 ³ a ¢ tanh [kh] ³a The added resistance due to this force is: ¶ µ Z dyw ¤ (x ) ¢ ¡ Raw1 = 2 f b ¢ dxb dxb L ¶ Z µ 2 zxa 2sa ¢ cos(¡kxb cos ¹ ¡ "s³ ) ½g³ 2 dyw a = 1¡ 2 ¡ ¢ ¢ dxb 2 ³ a ¢ tanh [kh] dxb ³a L For deep water. the constant term and the …rst harmonic term can be ignored. this part of the mean added resistance reduces to: Z ¡½g dyw (as given by Boese for deep water) Raw1 = s2 ¢ ¢ dxb a 2 dxb L The integrated vertical hydromechanical and wave forces in the shipborne axis system varies not only in time but also in direction with the pitch angle. So: µ 2 ¶ 2 ¡³ + zx ³ ¢ (³ ¡ zx) ¤ f (xb . t) = Z³ p ¢ dzb à ³ ¡³ 2 + (¡Ds + zx )2 + ¢ (³ + Ds ¡ zx ) 2 tanh [kh] ! ¡Ds +zx = ½g ¢ with: zx = z ¡ xb µ. . t) = ½g ¢ + 2 tanh [kh] The vertical relative motion is de…ned by s = ³ ¡ zx .230 CHAPTER 11. ADDED RESISTANCES DUE TO WAVES The pressure in the undisturbed waves is given by: p = ¡½gz + ½g ¢ cosh k(h + zb ) ¢³ cosh(kh) cosh k(h + zb ) = ¡½gz + ½g ¢ ¢ ³ a cos(! et ¡ kxb cos ¹ ¡ kyb sin ¹) cosh(kh) The horizontal force on an oscillating cross section is given by: f(xb . so: µ 2 ¶ 2 ¡³ + zx ³ ¢s ¤ + f (xb . As the mean added resistance during one period will be calculated.

2 shows an example of a comparison between computed and experimental data.2: Added Resistance of the S-175 Containership Design .11.3. Figure 11.3 Comparison of Results Figure 11. COMPARISON OF RESULTS From this follows a second contribution to the mean added resistance: Raw2 ¡1 = Te = ¡1 Te T Ze 231 (Zh (t) + Zw (t)) ¢ µ(t) ¢ dt ½r ¢ z (t) ¢ µ(t) ¢ dt Ä 0 T Ze 0 For this second contribution can be written: 1 Raw2 = ½r ¢ ! 2 ¢ za ¢ µa ¢ cos("z³ ¡ "µ³ ) e 2 So the transfer function of the total mean added resistance according to Boese is given by: Raw2 ³2 a = 1 za µ a ½r ¢ ! 2 ¢ ¢ ¢ cos("z³ ¡ "µ³ ) e 2 ³a ³a ¶ Z µ 2 zxa 2sa ¢ cos(¡kxb cos ¹ ¡ "s³ ) dyw 1 ¢ ¢ dxb + ½g 1¡ 2 ¡ 2 ³ a ¢ tanh [kh] dxb ³a L 11.

232 . CHAPTER 11. ADDED RESISTANCES DUE TO WAVES .

Chapter 12 Bending and Torsional Moments The axes system (of which the hydrodynamic sign convention di¤ers from that commonly used in structural engineering) and the internal load de…nitions are given in …gure 12.wbmt. 2628 CD Delft. Journée.J. February 2001. 233 .tudelft. The Netherlands. Delft University of Technology.nl/~johan or http://www.19”.1. Figure 12.1: Axis System and Internal Load De…nitions To obtain the vertical and lateral shear forces and bending moments and the torsional moments the following information over a length Lm on the solid mass distribution of the ship including its cargo is required: 0 J.shipmotions. Mekelweg 2. Release 4.nl.M. ”Theoretical Manual of SEAWAY. For updates see web site: http://dutw189. Ship Hydromechanics Laboratory. Report 1216a.

for instance to avoid a remaining calculated bending moment at the forward end of the ship. about a horizontal longitudinal axis through the centre of gravity Figure 12. Then m0 (xb ) will be replaced by: ½r m0 (xb ) ¢ m The longitudinal position of the centre of gravity is found from the distribution of the mass per unit length: Z 1 xG = m0 (xb ) ¢ xb ¢ dxb m Lm . The total mass of the ship is found by an integration of the mass per unit length: Z m0 (xb ) ¢ dxb m= Lm It is obvious that this integrated mass should be equal to the mass of displacement. Mostly small adaptions are necessary.2: Distribution of Solid Mass The input values for the calculation of shear forces and bending and torsional moments are often more or less inaccurate. so small deviations are possible. A proportional correction of the masses per unit length m0 (xb ) can be used.234 CHAPTER 12. calculated from the underwater hull form: m = ½r Both terms will be calculated from independently derived data. BENDING AND TORSIONAL MOMENTS m0 (xb ): distribution over the ship length of the solid mass of the ship per unit length 0 zm (xb ): distribution over the ship length of the vertical zb values of the centre of gravity of the solid mass of the ship per unit length 0 kxx (xb ): distribution over the ship length of the radius of inertia of the solid mass of the ship per unit length.

a small deviation is possible. without changing the total mass or the position of its centre of gravity. because of independently derived data. with: c(xb ) = ¡c1 ¢ (xb ¡ xA ¡ 0) µ ¶ Lm c(xb ) = +c1 ¢ xb ¡ xA ¡ 2 c(xb ) = ¡c1 ¢ (xb ¡ xA ¡ Lm ) with: c1 = In here: xA Lm = xb -coordinate of aftmost part of mass distribution = total length of mass distribution 32 ¢ ½r ¢ (xB ¡ xG ) L3 m for: Lm 4 3Lm Lm for: < xb ¡ xA < 4 4 3Lm for: < xb ¡ xA < Lm 4 0 < xb ¡ xA < For relatively slender bodies. Then. . the longitudinal gyradius of the mass can be found from the distribution of the mass per unit length: 2 kyy 1 = m Lm Z m0 (xb ) ¢ x2 ¢ dxb b It can be desirable to change the mass distribution in such a way that a certain required longitudinal gyradius kyy (new) or kzz (new) will be achieved. m0 (xb ) can be replaced by m0 (xb ) + c(xb ). for instance.3: Mass Correction for Buoyancy An equal longitudinal position of the ship’s centre of buoyancy xB is required.235 Figure 12. so: xG = xB Again.

0 0 So. for instance.4: Mass Correction for Center of Buoyancy Then. this value has to be subtracted from 0 zm (xb ). m0 (xb ) can be replaced by m0 (xb ) + c(xb ). .236 CHAPTER 12. If not so. with: c(xb ) = +c2 ¢ (xb ¡ xA ¡ 0) µ ¶ 2Lm c(xb ) = ¡c2 ¢ xb ¡ xA ¡ 8 µ ¶ 4Lm c(xb ) = +c2 ¢ xb ¡ xA ¡ 8 µ ¶ 4Lm c(xb ) = ¡c2 ¢ xb ¡ xA ¡ 8 µ ¶ 6Lm c(xb ) = +c2 ¢ xb ¡ xA ¡ 8 c(xb ) = ¡c2 ¢ (xb ¡ xA ¡ Lm ) with: for: for: for: for: for: for: Lm 8 3Lm Lm < xb ¡ xA < 8 8 4Lm 3Lm < xb ¡ xA < 8 8 5Lm 4Lm < xb ¡ xA < 8 8 5Lm 7Lm < xb ¡ xA < 8 8 7Lm < xb ¡ xA < Lm 8 0 < xb ¡ xA < ¡ 2 ¢ 2 3204 ¢ ½r ¢ kyy (new) ¡ kyy (old) c2 = 9L3 m xA Lm = xb -coordinate of aftmost part of mass distribution = total length of mass distribution In here: The position in height of the centre of gravity is found from the distribution of the heights of the centre of gravity of the masses per unit length: Z 1 0 zG = m0 (xb ) ¢ zm (xb ) ¢ dxb m Lm It is obvious that this value should be zero. zm (xb ) will be replaced by zm (xb ) ¡ zG . BENDING AND TORSIONAL MOMENTS Figure 12.

Figure 12. a proportional correction of the longitudinal distribution of the radii of inertia can be used: 0 0 kxx (xb .5: Mass Correction for Radius of Inertia The transverse radius of inertia kxx is found from the distribution of the radii of inertia of the masses per unit length: 2 kxx 1 = m Lm Z 0 m0 (xb ) ¢ kxx (xb ) ¢ dxb 2 If this value of kxx di¤ers from a required value kxx (new) of the gyradius. new) = kxx (xb . old) ¢ kxx (new) kxx (old) Consider a section of the ship with a length dxb to calculate the shear forces and the bending and the torsional moments.6: Loads on a Cross Section .237 Figure 12.

12. The ordinates of these terms will di¤er generally. this implies for the disk: q(xb ) ¢ dxb = ¡dQ(xb ) Q(xb ) ¢ dxb = +dM (xb ) in which: Q(xb ) M(xb ) = shear force = bending moment so: so: dQ(xb ) = ¡q(xb ) dxb dM (xb ) = +Q(xb ) dxb The shear force and the bending moment in a cross section x1 follows from an integration of the loads from the aftmost part of the ship x0 to this cross section x1 : Q(x1 ) = ¡ M (x1 ) = ¡ = + x Z1 dQ(xb ) ¢ dxb dxb Q(xb ) ¢ dxb 0 @ Zxb dQ(xb ) ¢ dxb A ¢ dxb dxb 1 x0 x Z1 x0 x Z1 x0 x0 So.1 Still Water Loads Consider the forces acting on a section of the ship with a length dxb. The load q(xb ) consists of solid mass and hydromechanical terms. . BENDING AND TORSIONAL MOMENTS When the disk is loaded by a load q(xb ). so the numerical integrations of these two terms have to be carried out separately.238 CHAPTER 12. the shear force Q(x1 ) and the bending moment M (x1 ) in a cross section can be expressed in the load q(xb ) by the following integrals: Q(x1 ) = ¡ M (x1 ) = + x Z1 q(xb ) ¢ dxb q(xb ) ¢ (x1 ¡ xb ) ¢ dxb q(xb ) ¢ xb ¢ dxb ¡ x1 ¢ x Z1 = + x0 x Z1 x0 x0 x Z1 q(xb ) ¢ dxb x0 For the torsional moment an approach similar to the approach for the shear force can be used.

2 Lateral Dynamic Loads Consider the forces acting on a section of the ship with a length dxb. For the lateral mode and the roll mode a similar procedure can be followed. LATERAL DYNAMIC LOADS 239 Figure 12.12. 1962]. reference is given to [Fukuda. 12. the vertical shear force Q3sw (x1 ) and the bending moment Q5sw (x1 ) in still water in a cross section can be obtained from the vertical load q3sw (xb ) by the following integrals: Q3sw (x1 ) = ¡ Q5sw (x1 ) = + x Z1 q3sw (xb ) ¢ dxb q3sw (xb ) ¢ xb ¢ dxb ¡ x1 x Z1 x0 x0 x Z1 q3sw (xb ) ¢ dxb x0 For obtaining the dynamic parts of the vertical shear forces and the vertical bending moments in regular waves. This will be shown at the following pages.7: Still Water Loads on a Cross Section According to Newton’s second law of dynamics. the vertical forces on the unfastened disk of a ship in still water are given by: (m0 ¢ dxb ) ¢ (¡g) = q3sw (xb ) ¢ dxb with: q3sw (xb ) = ½As g ¡ m0 g So. According to Newton’s second law of dynamics.2. the harmonic lateral dynamic load per unit length on the unfastened disk is given by: q2 (xb ) = 0 0 +Xh2 (xb ) + Xw2 (xb ) +½gAs Á ³ ´ 0 Ä ¡ z0 ¢ Á + g ¢ Á Ä Ä ¡m (xb ) ¢ y + xb ¢ Ã m .

8: Lateral Loads on a Cross Section in Waves The sectional hydromechanical load for sway is given by: 0 Xh2 = ¡a022 ¢ y ¡ b022 ¢ y ¡ c022 ¢ y Ä _ 0 0 Ä _ ¡a24 ¢ Á ¡ b24 ¢ Á ¡ c024 ¢ Á Ä _ ¡a026 ¢ Ã ¡ b026 ¢ Ã ¡ c026 ¢ Ã with: 0 a022 = +M22 ¯ ¯ 0 ¯ V dN22 ¯ ¯ +¯ 2 ¢ ¯! dxb ¯ e 0 b022 = +N22 ¡ V ¢ 0 dM22 dxb c022 = 0 0 0 a024 = +M24 + OG ¢ M22 ¯ ¯ 0 0 ¯ V dN24 V dN22 ¯ ¯ ¢ ¯ +¯ 2 + 2 ¢ OG ¢ ! e dxb !e dxb ¯ µ ¶ 0 0 dM24 dM22 0 0 0 b24 = +N24 ¡ V ¢ + OG ¢ N22 ¡ V ¢ dxb dxb 0 c24 = 0 µ ¶ 0 V dM22 0 0 0 a26 = +M22 ¢ xb + 2 ¢ N22 ¡ V ¢ !e dxb ¯ ¯ 0 ¯V ¯ V dN22 0 +¯ 2 ¢ N22 + 2 ¢ ¢ xb ¯ ¯! ¯ ! e dxb e ¶ µ 0 dM22 0 0 ¢ xb ¡ 2V ¢ M22 b026 = + N22 ¡ V ¢ dxb ¯ 2 ¯ 0 ¯ ¯V dN22 ¯ +¯ 2 ¢ ¯! dxb ¯ e c026 = 0 .240 CHAPTER 12. BENDING AND TORSIONAL MOMENTS Figure 12.

3.12.3 Vertical Dynamic Loads Consider the forces acting on a section of the ship with a length dxb. Then the harmonic lateral shear forces Q2 (x1 ) and the bending moments Q6 (x1 ) in waves in cross section x1 can be obtained from the horizontal load q2 (xb ) by the following integrals: Q2 (x1 ) = Q2a cos(! e t + "Q2 ³ ) x Z1 = ¡ q2 (xb ) ¢ dxb x0 Q6 (x1 ) = Q6a cos(! e t + "Q6 ³ ) x x Z1 Z1 = + q2 (xb ) ¢ xb ¢ dxb ¡ x1 q2 (xb ) ¢ dxb x0 x0 12. When ignoring the outlined terms the ”Ordinary Strip Theory” is presented. the harmonic longitudinal and vertical dynamic loads per unit length on the unfastened disk are given by: q1 (xb ) = 0 0 +Xh1 (xb ) + Xw1 (xb ) ³ ´ 0 Ä ¡m (xb ) ¢ x ¡ bG ¢ µ Ä 0 0 +Xh3 (xb ) + Xw3 (xb ) ³ ´ ¡m0 (xb ) ¢ z ¡ xb ¢ Ä Ä µ q3 (xb ) = The sectional hydromechanical load for surge is given by: 0 Xh1 (xb ) = ¡a011 ¢ x ¡ b011 ¢ x ¡ c011 ¢ x Ä _ Ä _ ¡a013 ¢ z ¡ b013 ¢ z ¡ c013 ¢ z 0 Ä ¡ b0 ¢ µ ¡ c0 ¢ µ _ ¡a15 ¢ µ 15 15 . According to Newton’s second law of dynamics. VERTICAL DYNAMIC LOADS The sectional wave load for sway is given by: 0 0 Ä Xw2 = +M22 ¢ ³ w2 ¯ ¯ 0 ¯ V dN22 Ĥ ¯ ¯ +¯ ¯ ! ¢ ! e ¢ dxb ¢ ³ w2 ¯ ï ¯ ! 0 ¯!¯ dM22 0 _¤ + ¯ ¯ ¢ N22 ¡ V ¢ ¢ ³ w2 ¯ !e ¯ dxb 0 +XF K2 0 Ĥ +M24 ¢ ³ w4 ¯ ¯ 0 ¯ V dN24 Ĥ ¯ ¯ +¯ ¢ ¢ ³ w4 ¯ ¯ ! ¢ ! e dxb ï ¯ ! 0 ¯!¯ dM24 0 _¤ + ¯ ¯ ¢ N24 ¡ V ¢ ¢ ³ w4 ¯ !e ¯ dxb ¤ 241 The ”Modi…ed Strip Theory” includes the outlined terms.

9: Vertical Loads on a Cross Section in Waves with: 0 a011 = +M11 ¯ ¯ 0 ¯ V dN11 ¯ ¯ ¢ ¯ +¯ 2 ! e dxb ¯ b011 = = = = = = 0 +N11 0 dM11 ¡V ¢ + b011V dxb c011 a013 b013 c013 a015 b015 c015 0 0 0 0 0 ¡M11 ¢ bG ¯ ¯ 0 ¯ ¯ ¡V dN11 ¢ bG¯ +¯ 2 ¢ ¯ ¯! dxb e µ ¶ 0 dM11 0 = ¡ N11 ¡ V ¢ ¢ bG ¡ b011V ¢ bG dxb = 0 The sectional hydromechanical load for heave is given by: 0 Xh3 (xb ) = ¡a031 ¢ x ¡ b031 ¢ x ¡ c031 ¢ x Ä _ 0 0 Ä _ ¡a33 ¢ z ¡ b33 ¢ z ¡ c033 ¢ z 0 0 _ µ ¡a35 ¢ Ä ¡ b35 ¢ µ ¡ c035 ¢ µ with: a031 b031 c031 a033 = = = = 0 0 0 0 +M33 .242 CHAPTER 12. BENDING AND TORSIONAL MOMENTS Figure 12.

Then the harmonic vertical shear forces Q3 (x1 ) and the bending moments Q5 (x1 ) in waves in cross section x1 can be obtained from the longitudinal and vertical load q1 (xb ) and q3 (xb ) by the following integrals: Q3 (x1 ) = Q3a cos(! e t + "Q3 ³ ) x Z1 = ¡ q3 (xb ) ¢ dxb x0 Q5 (x1 ) = Q5a cos(! e t + "Q5 ³ ) .3. When ignoring the outlined terms the ”Ordinary Strip Theory” is presented.12. VERTICAL DYNAMIC LOADS ¯ ¯ 0 ¯ V dN33 ¯ ¯ +¯ 2 ¢ ¯! dxb ¯ e 243 0 b033 = +N33 ¡ V ¢ 0 dM33 dxb c033 = +2½g ¢ yw a035 0 ¡M33 b035 c035 = ¡2½g ¢ yw ¢ xb µ ¶ 0 V dM33 0 = ¢ xb ¡ 2 ¢ N33 ¡ V ¢ !e dxb ¯ ¯ 0 ¯ ¡V ¯ V dN33 0 +¯ 2 ¢ N33 ¡ 2 ¢ ¢ xb ¯ ¯! ¯ ! e dxb e µ ¶ 0 dM33 0 0 = ¡ N33 ¡ V ¢ ¢ xb + 2V ¢ M33 dxb ¯ 2 ¯ 0 ¯ ¯V dN33 ¯ +¯ 2 ¢ ¯! dxb ¯ e The sectional wave loads for surge and heave are given by: 0 0 Ä Xw1 (xb ) = +M11 ¢ ³ w1 ¯ ¯ 0 ¯ V dN11 Ĥ ¯ ¯ +¯ ¯ ! ¢ ! e ¢ dxb ¢ ³ w1 ¯ ï ¯ ! 0 ¯!¯ dM11 0 _¤ + ¯ ¯ ¢ N11 ¡ V ¢ ¢ ³ w1 ¯ !e ¯ dxb 0 +XF K1 0 0 Ĥ Xw3 (xb ) = +M33 ¢ ³ w3 ¯ ¯ 0 ¯ V dN33 Ĥ ¯ ¯ +¯ ¢ ¢ ³ w3 ¯ ¯ ! ¢ ! e dxb ï ¯ ! 0 ¯!¯ dM33 0 _¤ + ¯ ¯ ¢ N33 ¡ V ¢ ¢ ³ w3 ¯ !e ¯ dxb 0 +XF K3 ¤ The ”Modi…ed Strip Theory” includes the outlined terms.

10: Distribution of Vertical Bending Moment Amplitudes 12.244 x Z1 CHAPTER 12. BENDING AND TORSIONAL MOMENTS x Z1 x Z1 = + q1 (xb ) ¢ bG(xb ) ¢ dxb + q3 (xb ) ¢ xb ¢ dxb ¡ x1 q3 (xb ) ¢ dxb x0 x0 x0 Figure 12. Figure 12. the harmonic torsional dynamic load per unit length on the unfastened disk about an longitudinal axis at a distance z1 above the ship’s center of gravity is given by: q4 (xb .10 shows a comparison between measured and calculated distributions of the vertical wave bending moment amplitudes over the length of the ship. Figure 12.4 Torsional Dynamic Loads Consider the forces acting on a section of the ship with a length dxb.11: Torsional Loads on a Cross Section in Waves According to Newton’s second law of dynamics. z1 ) = 0 0 +Xh4 (xb ) + Xw4 (xb ) ³ 2 ´ 0 0 Ä Ä ¡m0 (xb ) ¢ kxx ¢ Á ¡ zm ¢ (Ä + xb ¢ Ã + g ¢ Á) y +z1 ¢ q2 (xb ) .

4.12. TORSIONAL DYNAMIC LOADS The sectional hydromechanical load for roll is given by: 0 Xh4 (xb ) = 245 ¡a042 ¢ y ¡ b042 ¢ y ¡ c042 ¢ y Ä _ Ä _ ¡a044 ¢ Á ¡ b044 ¢ Á ¡ c044 ¢ Á Ä _ ¡a0 ¢ à ¡ b0 ¢ à ¡ c0 ¢ à 46 46 46 with: 0 a042 = +M42 + OG ¢ a022 ¯ ¯ 0 ¯ V dN42 ¯ ¯ ¢ ¯ +¯ 2 ! dxb ¯ e 0 b042 = +N42 ¡ V ¢ 0 dM42 + OG ¢ b022 dxb c042 = 0 + OG ¢ c022 0 0 a044 = +M44 + OG ¢ M42 + OG ¢ a024 ¯ ¯ 0 0 ¯ V dN44 V dN42 ¯ ¯ +¯ 2 ¢ + 2 ¢ OG ¢ ¯! dxb !e dxb ¯ e µ ¶ 0 0 dM44 dM42 0 0 0 b44 = +N44 ¡ V ¢ + OG ¢ N42 ¡ V ¢ + b044V + OG ¢ b024 dxb dxb µ 3 ¶ yw As ¡ ¢ bG c044 = +2½g ¢ 3 2 µ ¶ 0 V dM42 0 0 0 a46 = +M42 ¢ xb + 2 N42 ¡ V ¢ + OG ¢ a026 !e dxb ¯ ¯ 0 ¯V ¯ V dN42 0 +¯ 2 ¢ N42 + 2 ¢ ¢ xb ¯ ¯! ¯ ! e dxb e µ ¶ 0 dM42 0 0 0 b46 = + N42 ¡ V ¢ ¢ xb ¡ 2V ¢ M42 + OG ¢ b026 dxb ¯ 2 ¯ 0 ¯ ¯V dN42 ¯ +¯ 2 ¢ ¯! dxb ¯ e c046 = 0 + OG ¢ c026 In here. bG is the vertical distance of the centre of gravity of the ship G above the centroid b of the local submerged sectional area. The sectional wave load for roll is given by: 0 0 Ĥ Xw4 (xb ) = +M44 ¢ ³ w4 ¯ ¯ 0 ¯ V dN44 Ĥ ¯ ¯ +¯ ¯ ! ¢ ! e ¢ dxb ¢ ³ w4 ¯ ! ï ¯ 0 ¯!¯ 0 dM44 _¤ ¢ ³ w4 + ¯ ¯¢N44 ¡ V ¢ ¯ !e ¯ dxb 0 +XF K4 0 Ĥ +M42 ¢ ³ w2 .

BENDING AND TORSIONAL MOMENTS ¯ ¯ 0 ¯ V dN42 ¤ ¯ Ä ¯ +¯ ¯ ! ¢ ! e dxb ¢ ³ w2 ¯ ! ï ¯ 0 ¯!¯ 0 dM42 _¤ ¢ ³ w2 + ¯ ¯¢N42 ¡ V ¢ ¯ !e ¯ dxb 0 +OG ¢ Xw2 The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined terms the ”Ordinary Strip Theory” is presented. z1 ) = Q4a cos(! e t + "Q4 ³ ) x Z1 = ¡ q4 (xb . z1 ) by the following integral: Q4 (x1 . Then the harmonic torsional moments Q4 (x1 . z1 ) ¢ dxb x0 .246 CHAPTER 12. z1 ) in waves in cross section x1 at a distance z1 above the ship’s centre of gravity can be obtained from the torsional load q4 (xb .

2628 CD Delft. frequently used in open sea areas ² the Mean JONSWAP wave spectrum. Delft University of Technology. Report 1216a.nl. Release 4. the response energy spectra can be obtained. Generally the frequency ranges of the energy spectra of the waves and the responses of the ship on these waves are not very wide. a narrow wave spectrum. standard representations of the wave energy distributions are necessary. Ship Hydromechanics Laboratory.19”. February 2001. Then the Rayleigh distribution can be used to obtain a probability density function of the maximum and minimum values of the waves and the responses. a …xed relation with T1 can be used not-truncated spectra. ”Theoretical Manual of SEAWAY. The mathematical formulations of these normalised uni-directional wave energy spectra are based on two parameters: ² the signi…cant wave height H1=3 ² the average wave period T1 . which is sometimes used for open sea areas ² the Bretschneider wave spectrum. Journée. From these wave energy spectra and the transfer functions of the responses.shipmotions.Chapter 13 Statistics in Irregular Waves To compare the calculated behaviour of di¤erent ship designs or to get an impression of the behaviour of a speci…c ship design in a seaway. With this function. 0 247 . an average wave spectrum. Mekelweg 2. a somewhat wide wave spectrum. the probabilities on exceeding threshold values by the ship motions can be calculated.wbmt. For updates see web site: http://dutw189. J. Bow slamming phenomena are de…ned by a relative bow velocity criterium and a peak bottom impact pressure criterium.nl/~johan or http://www. based on the centroid of the spectral area curve. Three well known types of normalised wave energy spectra are described here: ² the Neumann wave spectrum. frequently used in North Sea areas.J. To obtain the average zero-crossing period T2 or the spectral peak period Tp .M. The Netherlands.tudelft.

1 Neumann Wave Spectrum In some cases in literature the Neumann de…nition of a wave spectrum for open sea areas is used: ½ ¾ 2 3832 ¢ H1=3 ¡69:8 ¡2 ¡6 S³ (!) = ¢ ! ¢ exp ¢! 5 2 T1 T1 13. an average spectrum ² the mean JONSWAP wave spectrum.1.1: Comparison of Three Spectral Formulations 13. the Bretschneider and the mean JONSWAP wave spectra is given here for a sea state with a signi…cant wave height of 4 meters and an average wave period of 8 seconds.1.2 Bretschneider Wave Spectrum A very well known two-parameter wave spectrum of open seas is de…ned by Bretschneider as: ½ ¾ 2 172:8 ¢ H1=3 ¡691:2 ¡4 ¡5 S³ (!) = ¢ ! ¢ exp ¢! 4 4 T1 T1 Another name of this wave spectrum is the Modi…ed Two-Parameter Pierson-Moskowitz Wave Spectrum. STATISTICS IN IRREGULAR WAVES 13. Figure 13. . a somewhat wide spectrum ² the Bretschneider wave spectrum.1 Normalized Wave Energy Spectra Three mathematical de…nitions with two parameters of normalized spectra of irregular uni-directional waves have been described: ² the Neumann wave spectrum. a narrow spectrum A comparison of the Neumann.248 CHAPTER 13.

From analysis of the measured spectra.4 De…nition of Parameters The n-th order spectral moments of the wave spectrum. The original One-Parameter Pierson-Moskowitz Wave Spectrum for fully developed seas can be obtained from this de…nition by using a …xed relation between the signi…cant wave p height and the average wave period in this Bretschneider de…nition: T1 = 3:861 ¢ H1=3 .1. or I. 13. known as the Joint North Sea Wave Project (JONSWAP) was carried out along a line extending over 100 miles into the North Sea from Sylt Island.13.3 Mean JONSWAP Wave Spectrum In 1968 and 1969 an extensive wave measurement program.C.T. 13.1. Wave Spectrum. Sometimes. NORMALIZED WAVE ENERGY SPECTRA 249 This formulation is accepted by the 2nd International Ship Structures Congress in 1967 and the 12th International Towing Tank Conference in 1969 as a standard for seakeeping calculations and model experiments. a third free parameter is introduced in the JONSWAP wave spectrum by varying the peakedness factor °.S. are: Z1 mn³ = S³ (!) ¢ ! n ¢ d! 0 The breadth of a wave spectrum is de…ned by: s "= 1¡ m2 2³ m0³ ¢ m4³ . de…ned as a function of the circular wave frequency !.1.S. The following de…nition of a Mean JONSWAP wave spectrum is advised by the 15th ITTC in 1978 for fetch limited situations: ¾ ½ 2 172:8 ¢ H1=3 ¡691:2 ¡4 ¡5 ¢ A ¢ °B S³ (!) = ¢ ! ¢ exp ¢! 4 4 T1 T1 with: A = 0:658 8 Ã !2 9 ! < ¡ 1:0 = !p p B = exp ¡ : .C. a spectral formulation of wind generated seas with a fetch limitation was found. This is reason why this spectrum is also called I. ¾ 2 ° = 3:3 (peakedness factor) 2¼ (circular frequency at spectral peak) !p = Tp ¾ = a step function of !: if ! < ! p then ¾ = 0:07 if ! > ! p then ¾ = 0:09 The JONSWAP expression is equal to the Bretschneider de…nition multiplied by the frequency function A ¢ ° B .T.

STATISTICS IN IRREGULAR WAVES The signi…cant wave height is de…ned by: H1=3 = 4:0 ¢ Tp = T1 = 2¼ ¢ T2 = 2¼ ¢ r m0³ = m1³ m0³ = m2³ p m0³ The several de…nitions of the average wave period are: peak or modal wave period. the theoretical relations between the periods are tabled below: 9 T1 = 1:086 ¢ T2 = 0:772 ¢ Tp = for Bretschneider Wave Spectra 0:921 ¢ T1 = T2 = 0:711 ¢ Tp . corresponding to radius of inertia of spectral curve For not-truncated mathematically de…ned spectra. the wave heights will not di¤er much.250 CHAPTER 13. can cause di¤erences between input and calculated wave periods. 1:199 ¢ T1 = 1:287 ¢ T2 = Tp Truncation of wave spectra during numerical calculations. for ”Open Ocean Areas” and ”North Sea Areas” an indication is given of a possible average relation between the scale of Beaufort or the wind velocity at 19. Figure 13. . corresponding to centroid of spectral curve average zero-crossing wave period. 1:296 ¢ T1 = 1:407 ¢ T2 = Tp 9 T1 = 1:073 ¢ T2 = 0:834 ¢ Tp = 0:932 ¢ T1 = T2 = 0:777 ¢ Tp for JONSWAP Wave Spectra . Generally.5 meters above the sea level and the signi…cant wave height H1=3 and the average wave periods T1 or T2 .2 and the table below. corresponding to peak of spectral curve average wave period.2: Wave Spectra Parameter Estimates In …gure 13.

The modal or central periods in these tables correspond with the peak period Tp . Other open ocean de…nitions for the North Atlantic and the North Paci…c.75 8.5 m above sea Open Ocean Areas (Bretschneider) North Sea Areas (JONSWAP) (kn) H1=3 (m) T1 (s) T2 (s) H1=3 (m) T1 (s) T2 (s) ° (-) 1 2 3 4 5 6 7 8 9 10 11 12 2.30 3.3 3.90 8.5 8.3 These data are an indication only.25 10.8 4.7 6.4 6.2 3.5 19.3 3.00 6.55 0.3 3.5 13.75 2.85 8.5 9.30 9.35 5.85 9.0 9.45 9.9 1.5 30.2 3. 1983] and adopted by the 17th ITTC (1984).13.10 9.60 6.0 24.25 4.15 6.30 8.35 6.5 >64.0 1.20 7.10 8.3 3.45 5.90 7.5 37.60 6.30 10. .8 8.3 3.70 10.0 5.45 9.30 1.65 5.40 6.3 3.20 7.6 4. a generally applicable …xed relation between wave heights and wave periods does not exist.1 10. are given in the tables below.25 10.25 0.3 3.80 4.7 5.50 3.0 1. For not-truncated spectra.6 8.55 5.0 51.90 1.10 4.65 0.9 7.3 3.8 3.5 2. the relations with T1 and T2 are de…ned before.70 10.1 2. NORMALIZED WAVE ENERGY SPECTRA 251 Indication of Wave Spectra Parameters Scale of Wind Speed Beaufort at 19.3 3.10 5.8 1.80 3.65 7.5 5.3 3.5 3.85 7.0 7.50 5.65 3. obtained from [Bales.5 59.3 3.0 44.70 6.20 5.0 8.1 4.1.75 7.

12.0.15.5 5.0 .47 48 .16.55 56 .17.0.15.5 37.16 17 .14 >14 0.47 48 .8 5.5 7.0.5 14.4.50 .27 28 .55 56 .2 11.5 10.0 .5 13.0 North Paci…c 0-1 2 3 4 5 6 7 8 >8 0.0 7.5 59.4 28.1 2.5 >14 0-6 7 .5 13.5 51.9 27.3 0.1.21 22 .5 19 24.10 11 .8 6.4 15.3 0.0 .0 Note: 1) Ambient wind sustained at 19.25 5.7 7.88 1.14 >14 0.8 .7 6.0 7.18.5 51.25 .5 59.7 15.0 4-6 6-9 9 .3 .2 <0.1 .5 9.15.8 18.1 0.4 20.252 CHAPTER 13. To convert to another altitude h2 .88 1.6 18.21 22 .1.0 7.5 9.2 .18.0 4-6 6-9 9 .5 >63 0 4.8 23.3 9.8 .2.25 5.25 .1 0.0 16.5 11.5 7.16.50 2.5 8.0.50 .16 17 .3 .5 .1 .2 8.2.1 3.0 20.15.8 9.8 15.25 1.7 12. 3) Based on periods associated with central frequencies included in Hindcast Climatology.2 .10 11 .23.1 .2 22.5 >14 0-6 7 .9 .0 .0 .16.4.5 .5 m above surface to generate fully-developed seas.2 13.4 20.1 16.8 9. apply V2 = V1 ¢ (h2 =19:5)1=7 .2 .7 13.5 0. .5 19 24.3 .0 5. 2) Minimum is 5 percentile and maximum is 95 percentile for periods given wave height range.25 1.0 .05 0. STATISTICS IN IRREGULAR WAVES Open Ocean Annual Sea State Occurrences of Bales (1983) of the North Atlantic and the North Paci…c Sea State Number (¡) Signi…cant Wave Height H1=3 (m) Sustained Probability Wind Speed 1) of Sea State (kn) (%) Modal Wave Period Tp (s) Range Mean Range Mean Range 2) Most 3) Probable North Atlantic 0-1 2 3 4 5 6 7 8 >8 0.88 3.05 0.18.05 3.0 .5 18.5 7.50 2.5 16.63 >63 3 8.5 0.5 37.5 >63 0 7.5 8.27 28 .15.63 >63 3 8.88 3.5 11.1 1.2 0.14.

Figure 13.13.2 Response Spectra and Statistics The energy spectrum of the responses r(t) of a sailing ship on the irregular waves follows from the transfer function of the response and the wave energy spectrum by: Sr (!) = or: Sr (! e ) = µ ra ³a ¶2 ¢ S³ (! e ) µ ra ³a ¶2 ¢ S³ (!) This has been visualized for a heave motion in …gures 13. ::: From the spectral density function of a response the signi…cant amplitude can be calculated. . 1.2.3: Principle of Transfer of Waves into Responses The moments of the response spectrum are given by: mnr = Z1 0 Sr (! e ) ¢ ! n ¢ d! e e with: n = 0. RESPONSE SPECTRA AND STATISTICS 253 13.4. 2.3 and 13.

5 2.0 1.0 0.00 s 1.5 1. which is equivalent to the average zero-crossing period.00 m T2 = 8.0 Containership L = 175 metre Head waves V = 20 knots 2.0 Spectral density heave (m s) 8 6 4 2 2 za = 1.5 2. STATISTICS IN IRREGULAR WAVES 5 3 2 1 0 0 0.5 heav 1.0 s Wave 4 spectrum H1/3 = 5.74 s 2 1.0 1.5 1.0 0 0 0.254 Spectral density wave (m s) 5 2 CHAPTER 13.4: Heave Spectra in the Wave and Encounter Frequency Domain The signi…cant amplitude is de…ned to be the mean value of the highest one-third part of the highest wave heights.0 1.5 Transfer function heave Containership L = 175 metre Head waves V = 20 knots e 1.0 Transfer function 1. in case of a spectrum with a frequency range which is not too wide.5 0 8 6 4 2 0 0 0.00 Wave 4 spectrum 3 2 1 0 0 0.0 Transfer function heave (m/m) 2.92 m 1/3 Tz = 7.5 1.5 za = 1.5 1.5 2.5 1. is found from the spectral radius of inertia by: r m0r T2r = 2¼ ¢ m2r The probability density function of the maximum and minimum values.5 1.5 1.5 2.00 m T2 = 8.5 2.92 m Tz = 7. is given by the Rayleigh distribution: ¾ ½ 2 ra ¡ra f (ra ) = ¢ exp m0r 2m0r This implies that the probability of exceeding a threshold value a by the response amplitude ra becomes: ¾ ½ 2 ra ¡ra ¢ dra P fra > ag = ¢ exp m0r 2m0r a ½ ¾ ¡a2 = exp 2m0r Z1 .0 wave frequency (rad/s) frequency of encounter (rad/s) Figure 13.0 1.0 0 0 0.74 s 2 1/3 2.0 0 Heave spectrum 0.0 H1/3 = 5.0 0. so: p ra1=3 = 2 m0r A mean period can be found from the centroid of the spectrum by: m0r T1r = 2¼ ¢ m1r An other de…nition.

13. based on ! e . Because of the requirement of an equal amount of energy in the frequency bands ¢! and ¢! e.5: Example of Relation Between ! e and ! From the relation between ! e and ! follows: d! e V ¢ cos ¹ = 1:0 ¡ d! d! dk The derivative d!=dk follows from the relation between ! and k: p ! = kg ¢ tanh(kh) .2. it follows: S³ (! e ) ¢ d! e = S³ (!) ¢ d! From this the following relation is found: S³ (! e ) = S³ (!) d!e d! The relation between the frequency of encounter and the wave frequency. based on !. is not equal to the spectral value S³ (!). RESPONSE SPECTRA AND STATISTICS The number of times per hour that this happens follows from: Nhour = 3600 ¢ P fra > ag T2r 255 The spectral value of the waves S³ (! e ). is given by: ! e = ! ¡ kV ¢ cos ¹ Figure 13.5. of which an example is illustrated in …gure 13.

the transformed spectral values will range from plus in…nite to minus in…nite. these values can be obtained from this Sr (!) and the derivative d! e =d!. the calculated mean added resistance values can be presented by: RAW versus T1 or T2 2 H1=3 . T2 = signi…cant wave height average wave periods The mean added resistance in a seaway follows from: RAW = 2 Z1 0 RAW ¢ S³ (!) ¢ d! ³2 a Because of the linearities in the motions. a positive or a negative side. to zero. As a result of this. the calculated signi…cant values can be presented by: ra1=3 H1=3 versus T1 or T2 H1=3 = T1 . for instance for a comparison of the calculated response spectra with measured response spectra. STATISTICS IN IRREGULAR WAVES As can be seen in …gure 13.5. This implies that numerical problems will arise in the numerical integration routine. This is the reason why the spectral moments have to be written in the following format: m0r = Z1 0 kg g ¢ tanh(kh) + h¢cosh2 (kh) d! p = dk 2 kg ¢ tanh(kh) Sr (! e) ¢ d! e = Z1 0 Sr (!) ¢ d! Sr (!) ¢ ! e ¢ d! Sr (!) ¢ ! 2 ¢ d! e m1r = Z1 0 Sr (! e) ¢ ! e ¢ d! e = Sr (! e) ¢ ! 2 ¢ d! e = e µ ¶2 Z1 0 m2r = Z1 0 Z1 0 with: Sr (!) = ra ³a ¢ S³ (!) If Sr (! e ) has to be known. So an integration of Sr (! e ) over ! e has to be avoided. around a wave speed equal to twice the forward ship speed component in the direction of the wave propagation. Because of the linearities.256 So: CHAPTER 13. in following waves the derivative d! e =d! can approach from both sides.

[Tasaki. When sailing in still water. A static swell up should be taken into account. An empirical formula. shipping green water is de…ned by: s¤ > fe a at the forwarde perpendicular The spectral density of the vertical relative motion at the forward perpendicular is given by: µ ¤ ¶2 sa Ss¤ (!) = ¢ S³ (!) ³a .3.3 Shipping Green Water The e¤ective dynamic freeboard will di¤er from the results obtained from the geometric freeboard at zero forward speed in still water and the calculated vertical relative motions of a sailing ship in waves. calculated from the heave. the pitch and the wave motions. SHIPPING GREEN WATER 257 13. A dynamic swell up should be taken into account. for the static swell up at the forward perpendicular is given by [Tasaki. sinkage. 1963]: fe = f ¡ 0:75 ¢ B ¢ with: fe f L B LE Fn = = = = = = e¤ective freeboard at the forward perpendicular geometric freeboard at the forward perpendicular length of the ship breadth of the ship length of entrance of the waterline Froude number L 2 ¢ Fn LE An oscillating ship will produce waves and these dynamic phenomena will in‡uence the amplitude of the relative motion. trim and the ship’s wave system will e¤ect the local geometric freeboard. Then the actual amplitude of the relative motions becomes: s¤ = sa + ¢sa a Then.13. based on model experiments. 1963] carried out forced oscillation tests with ship models in still water and obtained an empirical formula for the dynamic swell-up at the forward perpendicular in head waves: s CB ¡ 0:45 ¢sa !2 L e = ¢ sa 3 g with the restrictions: block coe¢cient: 0:60 < CB < 0:80 Froude number: 0:16 < Fn < 0:29 In this formula sa is the amplitude of the relative motion at the forward perpendicular as obtained in head waves.

13. 1964] translated slamming phenomena into requirements for the vertical relative motions of the ship. the local ‡exibility of the ship’s bottom plating and the overall ‡exibility of the ship’s structure.4 Bow Slamming Slamming is a two-node vibration of the ship caused by suddenly pushing the ship by the waves.258 CHAPTER 13. without forward speed e¤ect. STATISTICS IN IRREGULAR WAVES The spectral moments are given by: mns¤ = Z1 0 Ss¤ (!) ¢ ! n ¢ d! e with: n = 0. which is beyond the scope of any existing theory. between the wave surface and the bow of the ship Ochi de…nes the vertical relative displacement and velocity of the water particles with respect to the keel point of the ship by: s = ³ xb ¡ z + xb ¢ µ _ _ s = ³ ¡ z + xb ¢ µ _ _ xb . A complete prediction of slamming phenomena is a complex task. 2. He de…ned slamming by: ² an emergence of the bow of the ship at 10 percentile of the length aft of the forward perpendiculars ² an exceeding of a certain critical value at the instance of impact by the vertical relative velocity.4.1 Criterium of Ochi [Ochi. 1. the relative angle between the keel and the water surface. Slamming impact pressures are a¤ected by the local hull section shape. the relative velocity between ship and waves at impact. ::: When using the Rayleigh distribution the probability of shipping green water is given by: ½ ¾ 2 ¡fe ¤ P fsa > fe g = exp 2m0s¤ The average zero-crossing period of the relative motion is found from the spectral radius of inertia by: r m0s¤ T2s¤ = 2¼ m2s¤ The number of times per hour that green water will be shipped follows from: Nhour = 3600 ¢ P fs¤ > fe g a ¤ T2s 13.

The spectral moments of the vertical relative displacements and velocities are de…ned by m0s and m0s . BOW SLAMMING with: ³ xb = ³ a cos(! e t ¡ kxb cos ¹) _ ³ xb = ¡! e ³ a sin(! et ¡ kxb cos ¹) 259 So a forward speed e¤ect is not included in the vertical relative velocity. The peak impact pressure is de…ned by: 1 p = Cp ¢ ½s2 _ 2 cr The coe¢cient Cp has been taken from experimental data of slamming drop tests with wedges and cones. 1974] translated slamming phenomena into requirements for the peak impact pressure of the ship.2 Criterium of Conolly [Conolly. In case of slamming both occurrences have to appear at the same time. are statistically independent. So the probability on a slam is the product of the both independent probabilities: _ _ P fslamg = P fsa > Ds g ¢ P fsa > scr g ½ ¾ 2 2 ¡Ds ¡scr _ = exp + 2m0s 2m0s _ 13. emergence of the bow and exceeding the threshold velocity. The probability of emergence of the bow follows from: ½ ¾ 2 ¡Ds P fsa > Ds g = exp 2m0s The second requirement states that the vertical relative velocity exceeds a threshold value. 12 feet per second can be taken as a threshold value for a ship with a length of 520 feet. . _ Emergence of the bow of the ship happens when the vertical relative displacement amplitude sa at 0:90 ¢ L is larger than the ship’s draft Ds at this location. He de…ned slamming by: ² an emergence of the bow of the ship ² an exceeding of a certain critical value by the peak impact pressure at this location.4.4. According to Ochi. Scaling results into: p scr = 0:0928 ¢ gL _ The probability of exceeding this threshold value is: ½ 2 ¾ ¡scr _ P fsa > scr g = exp _ _ 2m0s _ Both occurrences.13. as given in literature.

the section contour below 0:10 ¢ T has been used to de…ne an equivalent wedge: t = 0:10 ¢ T . STATISTICS IN IRREGULAR WAVES Figure 13. are illustrated in …gure 13. The accessory draught t of the wedge follows from the section contour. In the forebody of the ship. The contour of the cross section inside 10 percentile of the half breadth B=2 of the ship has been used to de…ne an equivalent wedge with a half breadth: b = 0:10 ¢ B=2. The accessory half breadth b of the wedge follows from the section contour. An equivalent deadrise angle ¯ is de…ned here by the determination of an equivalent wedge.260 CHAPTER 13. the local draught has been used. If this draught is larger than the local draught. 1989] as a function of the deadrise angle ¯. If so.6: Peak Impact Pressure Coe¢cients Some of these data. as for instance presented by [Lloyd. this draught can be larger than 10 percentile of the amidships draught T . Figure 13.7: De…nition of an Equivalent Wedge .6.

This lower limit is presented in …gure 13. He gives measured impact pressures at a ship with a length of 112 meter over 30 per cent of the ship length from forward. of the water particles with respect to the keel point of the ship is de…ned by: s = _ ª D © ³ xb ¡ z + xb ¢ µ _ Dt _ = ³ xb ¡ z + xb ¢ µ ¡ V ¢ µ _ with: ³ xb = ³ a cos(! e t ¡ kxb cos ¹) _ ³ xb = ¡!³ a sin(! e t ¡ kxb cos ¹) .13. From this.4.8.8: Measured Impact Pressures of a 112 Meter Ship Then the sectional area As below local draught t has to be calculated. a lower limit of pcr has been assumed. including a forward speed e¤ect. Now the equivalent deadrise angle ¯ follows from: ³a´ 0 · ¯ · ¼=2 ¯ = arctan b 2(b ¢ t ¡ As ) a = b Critical peak impact pressures pcr have been taken from [Conolly. BOW SLAMMING 261 Figure 13. These values have to be scaled to the actual ship size. 1974]. Bow emergence and exceeding of this limit is supposed to cause slamming. This approach can be translated into local hull shape-depending threshold values of the vertical relative velocity too: s 2pcr scr = _ ½Cp The vertical relative velocity.

The average period of the relative displacement is found by: r r m0s m0s T2s = 2¼ = 2¼ m2s m0s _ Then the number of times per hour that a slam will occur follows from: Nhour = 3600 ¢ P fslamg T2s ¾ . STATISTICS IN IRREGULAR WAVES ½ 2 ¡Ds ¡s2 _ cr + 2m0s 2m0s _ P fslamg = exp Note that. the spectral moment of the velocities does not follow from the spectral density of the relative displacement as showed in the de…nition of Ochi. because of including the forward speed e¤ect.262 Then: CHAPTER 13.

February 2001. The Netherlands. the wave loads and motions of twin-hull ships can be calculated easily. ”Theoretical Manual of SEAWAY. bij and cij in this chapter are those of one individual hull. Mekelweg 2. Report 1216a. The coordinate system for the equations of motion of a twin-hull ship is given in …gure 14. Journée.19”.wbmt.Chapter 14 Twin-Hull Ships When not taking into account the interaction e¤ects between the two individual hulls. 0 263 .tudelft. 2628 CD Delft. For updates see web site: http://dutw189. Release 4. Ship Hydromechanics Laboratory.shipmotions. de…ned in the coordinate system of the single hull.1.nl/~johan or http://www.nl. The distance between the two centre planes of the single hulls should be constant.M. as given and discussed before. J.1 Hydromechanical Coe¢cients The hydromechanical coe¢cients aij . Delft University of Technology. Each individual hull has to be symmetric with respect to its centre plane.1: Coordinate System of Twin-Hull Ships 14. Figure 14.J.

XT h5 .and z-direction exciting wave moments about the x-.and z-axis exciting wave forces in the x-. XT w2 .264 CHAPTER 14. XT w3 XT w4 . XT h2 . XT h3 XT h4 . y.3 Hydromechanical Forces and Moments The equations of motion for six degrees of freedom and the hydro. XT w6 = = = = = = volume of displacement of the twin-hull ship solid mass moment of inertia of the twin-hull ship hydromechanical forces in the x-.2 Equations of Motion Surge: Sway: Heave: Roll: Pitch: Yaw: Ä ½rT ¢ x Ä ½rT ¢ y ½rT ¢ z Ä Ä Ä IT xx ¢ Á ¡ IT xz ¢ Ã IT xx ¢ Ä µ Ä Ä IT zz ¢ Ã ¡ IT zx ¢ Á ¡ XT h1 = XT w1 ¡ XT h2 = XT w2 ¡ XT h3 = XT w3 ¡ XT h4 = XT w4 ¡ XT h5 = XT w5 ¡ XT h6 = XT w6 The equations of motion for six degrees of freedom of a twin-hull ship are de…ned by: in which: rT IT ij XT h1 . XT h6 XT w1 . y. are de…ned by: ¡XT h1 = +2a11 ¢ x + 2b11 ¢ x + 2c11 ¢ x Ä _ +2a13 ¢ z + 2b13 ¢ z + 2c13 ¢ z Ä _ Ä + 2b15 ¢ µ + 2c15 ¢ µ _ +2a15 ¢ µ ¡XT h2 = +2a22 ¢ y + 2b22 ¢ y + 2c22 ¢ y Ä _ Ä _ +2a24 ¢ Á + 2b24 ¢ Á + 2c24 ¢ Á Ä _ +2a26 ¢ Ã + 2b26 ¢ Ã + 2c26 ¢ Ã ¡XT h3 = +2a31 ¢ x + 2b31 ¢ x + 2c31 ¢ x Ä _ +2a33 ¢ z + 2b33 ¢ z + 2c33 ¢ z Ä _ Ä + 2b35 ¢ µ + 2c35 ¢ µ _ +2a35 ¢ µ Ä _ ¡XT h4 = +2a42 ¢ y + 2b42 ¢ y + 2c42 ¢ y Ä + 2b44 ¢ Á + 2c44 ¢ Á _ +2a44 ¢ Á 2 2 2 Ä _ +2yT ¢ a33 ¢ Á + 2yT ¢ b33 ¢ Á + 2yT ¢ c33 ¢ Á Ä _ +2a46 ¢ Ã + 2b46 ¢ Ã + 2c46 ¢ Ã ¡XT h5 = +2a51 ¢ x + 2b51 ¢ x + 2c51 ¢ x Ä _ +2a53 ¢ z + 2b53 ¢ z + 2c53 ¢ z Ä _ _ +2a55 ¢ Ä + 2b55 ¢ µ + 2c55 ¢ µ µ ¡XT h6 = +2a62 ¢ y + 2b62 ¢ y + 2c62 ¢ y Ä _ Ä _ +2a64 ¢ Á + 2b64 ¢ Á + 2c64 ¢ Á Ä _ +2a66 ¢ Ã + 2b66 ¢ Ã + 2c66 ¢ Ã 2 2 2 Ä _ +2yT ¢ a11 ¢ Ã + 2yT ¢ b11 ¢ Ã + 2yT ¢ c11 ¢ Ã . y.and z-direction hydromechanical moments about the x-.mechanic forces and moments in here.and z-axis 14. TWIN-HULL SHIPS 14. XT w5 . y.

the ”Modi…ed Strip Theory” includes the outlined terms. When ignoring these outlined terms the ”Ordinary Strip Theory” is presented.4 Exciting Wave Forces and Moments The …rst order wave potential for an arbitrary water depth h is de…ned in the new coordinate system by: ©w = ¡g cosh k(h + zb ) ¢ ¢ ³ a sin(! e t ¡ kxb cos ¹ ¡ kyb sin ¹) ! cosh(kh) This holds that for the port side (ps) and starboard (sb) hulls the equivalent components of the orbital accelerations and velocities in the surge.4. sway. EXCITING WAVE FORCES AND MOMENTS In here. yT is half the distance between the centre planes. The exciting wave forces for surge are: Z ³ ¤ ´ 0 Ĥ XT w1 = + M11 ¢ Äw1 (ps) + ³ w1 (sb) ¢ dxb ³ ¯ ¯ ¯ ¯ Z ³ ¤ ´ 0 ¯ V ¯ dN11 Ä ¤ Ä +¯ ¢ ³ w1 (ps) + ³ w1 (sb) ¢ dxb ¯ ¯ ! ¢ !e ¯ dxb ¯ ¯ L ! Z ï ¯ ´ ³ ¤ 0 ¯!¯ dM11 0 _¤ _ ¯ ¯ ¢ N11 ¡ V ¢ + ¢ ³ w1 (ps) + ³ w1 (sb) ¢ dxb ¯ !e ¯ dxb L L . heave and roll directions are equal to: Ĥ ³ w1 (ps) = ¡kg cos ¹ ¢ ³ ¤1 sin(! e t ¡ kxb cos ¹ ¡ kyT sin ¹) a Ĥ (sb) = ¡kg cos ¹ ¢ ³ ¤ sin(! e t ¡ kxb cos ¹ + kyT sin ¹) ³ w1 a1 +kg cos ¹ ¤ ¤ _ ³ w1 (ps) = ¢ ³ a1 cos(! et ¡ kxb cos ¹ ¡ kyT sin ¹) ! +kg cos ¹ ¤ _¤ ³ w1 (sb) = ¢ ³ a1 cos(! et ¡ kxb cos ¹ + kyT sin ¹) ! Ĥ ³ w2 (ps) = ¡kg sin ¹ ¢ ³ ¤2 sin(! et ¡ kxb cos ¹ ¡ kyT sin ¹) a Ĥ (sb) = ¡kg sin ¹ ¢ ³ ¤ sin(! et ¡ kxb cos ¹ + kyT sin ¹) ³ w2 a2 _¤ ³ w3 (ps) +kg sin ¹ ¤ = ¢ ³ a2 cos(! e t ¡ kxb cos ¹ ¡ kyT sin ¹) ! +kg sin ¹ ¤ _¤ ³ w2 (sb) = ¢ ³ a2 cos(! e t ¡ kxb cos ¹ + kyT sin ¹) ! Ĥ ³ w3 (ps) = ¡kg ¢ ³ ¤3 cos(! e t ¡ kxb cos ¹ ¡ kyT sin ¹) a Ĥ (sb) = ¡kg ¢ ³ ¤ cos(! e t ¡ kxb cos ¹ + kyT sin ¹) ³ _¤ ³ w2 (ps) w3 a3 +kg ¤ = ¢ ³ a3 sin(! e t ¡ kxb cos ¹ ¡ kyT sin ¹) ! +kg ¤ _¤ ³ w3 (sb) = ¢ ³ a3 sin(! e t ¡ kxb cos ¹ + kyT sin ¹) ! From this follows the total wave loads for the degrees of freedom. 265 14. In these loads on the following pages.14.

266 + Z L CHAPTER 14. TWIN-HULL SHIPS ¡ ¢ 0 0 XF K1 (ps) + XF K1 (sb) ¢ dxb The exciting wave forces for sway are: Z ³ ¤ ´ 0 Ä (ps) + ³ ¤ (ps) ¢ dxb Ä XT w2 = + M22 ¢ ³ w2 w2 L ¯ ¯ ¯ ¯ Z ³ ¤ ´ 0 ¯ V ¯ dN22 Ä Ä¤ (sb) ¢ dxb ¯ ¯ +¯ ¢ ³ w2 (ps) + ³ w2 ¯ dxb ¯ ! ¢ !e ¯ L ï ¯ ! Z ¯ ¯ ³ ´ 0 _ _ ¯ ! ¯ ¢ N 0 ¡ V ¢ dM22 ¢ ³ ¤ (ps) + ³ ¤ (sb) ¢ dxb + w2 w2 ¯ ! e ¯ 22 dxb L Z ¡ 0 ¢ 0 XF K2 (ps) + XF K2 (sb) ¢ dxb + L The exciting wave forces for heave are: Z ³ ¤ ´ 0 Ĥ ³ XT w3 = + M33 ¢ Äw3 (ps) + ³ w3 (sb) ¢ dxb L ¯ ¯ ¯ ¯ Z ³ ¤ ´ 0 ¯ V ¯ dN33 Ä ¤ Ä +¯ ¢ ³ w3 (ps) + ³ w3 (sb) ¢ dxb ¯ ¯ ! ¢ !e ¯ dxb ¯ ¯ L ! Z ï ¯ ³ ¤ ´ 0 ¯!¯ dM33 _¤ _ ¯ ¯ ¢ N0 ¡ V ¢ + ¢ ³ w3 (ps) + ³ w3 (sb) ¢ dxb ¯ ! e ¯ 33 dxb L Z ¢ ¡ 0 0 + XF K3 (ps) + XF K3 (sb) ¢ dxb L The exciting wave moments for roll are: Z ¡ 0 ¢ 0 XT w4 = + XF K4 (ps) + XF K4 (sb) ¢ dxb + L Z L ¯ ¯ ¯ ¯ Z ³ ¤ ´ 0 ¯ V ¯ dN42 Ä Ä¤ (sb) ¢ dxb ¯ ¯ +¯ ¢ ³ w2 (ps) + ³ w2 ¯ dxb ¯ ! ¢ !e ¯ L ï ¯ ! Z ³ ´ 0 ¯!¯ 0 _ _ ¯ ¯ ¢ N42 ¡ V ¢ dM42 ¢ ³ ¤ (ps) + ³ ¤ (sb) ¢ dxb + w2 w2 ¯ !e ¯ dxb L ³ ¤ ´ 0 Ĥ M42 ¢ Äw2 (ps) + ³ w2 (sb) ¢ dxb ³ +OG ¢ XT w2 +yT ¢ XT w3 .

14.5. ADDED RESISTANCE DUE TO WAVES The exciting wave moments for pitch are: Z ³ ¤ ´ 0 Ä Ä¤ XT w5 = ¡ M11 ¢ bG ¢ ³ w1 (ps) + ³ w1 (sb) ¢ dxb L 267 ¯ ¯ ¯ ¯ Z ³ ¤ ´ 0 ¯ V ¯ dN11 ¤ Ä +¯ ¢ bG ¢ ³ w1 (ps) + Äw1 (sb) ¢ dxb ¯ ³ ¯ ! ¢ !e ¯ dxb ¯ ¯ L ! Z ï ¯ ³ ¤ ´ 0 ¯!¯ dM11 0 _ _¤ ) ¢ bG ¢ ³ w1 (ps) + ³ w1 (sb) ¢ dxb ¡ ( ¯ ¯ ¢ N11 ¡ V ¢ ¯ !e ¯ dxb L Z ¡ 0 ¢ 0 ¡ XF K1 (ps) + XF K1 (sb) ¢ bG ¢ dxb ¡ ¯ ¯ ¯ ¯ Z ³ ¤ ´ 0 ¯ V ¯ dN33 Ä (ps) + ³ ¤ (sb) ¢ dxb ¯ Ä +¯ ¢ xb ¢ ³ w3 w3 ¯ ! ¢ !e ¯ dxb ¯ ¯ L ! Z ï ¯ ³ ¤ ´ 0 ¯!¯ dM33 0 _ (ps) + ³ ¤ (sb) ¢ dxb _ ¯ ¯ ¢ N33 ¡ V ¢ ¡ ¢ xb ¢ ³ w3 w3 ¯ !e ¯ dxb L Z ¡ 0 ¢ 0 ¡ XF K3 (ps) + XF K3 (sb) ¢ xb ¢ dxb L L Z L 0 M33 ³ ¤ ´ Ä (ps) + Ĥ (sb) ¢ dxb ¢ xb ¢ ³ w3 ³ w3 The exciting wave moments for yaw are: Z ³ ¤ ´ 0 Ä (ps) + ³ ¤ (sb) ¢ dxb Ä XT w6 = + M22 ¢ xb ¢ ³ w2 w2 L ¯ ¯ ¯ ¯ Z ³ ¤ ´ 0 ¯ V ¯ dN22 Ä (ps) + Ĥ (sb) ¢ dxb ¯ ¯ +¯ ¢ xb ¢ ³ w2 ³ w2 ¯ dxb ¯ ! ¢ !e ¯ L ! Z ï ¯ ³ ¤ ´ 0 ¯!¯ dM22 0 _ (ps) + ³ ¤ (sb) ¢ dxb _ ¯ ¯¢N ¡V ¢ + ¢ xb ¢ ³ w2 22 w2 ¯ !e ¯ dxb L Z ¡ 0 ¢ 0 + XF K2 (ps) + XF K2 (sb) ¢ xb ¢ dxb L +yT ¢ XT w1 14.5 Added Resistance due to Waves The added resistances can be found easily from the de…nitions of the mono-hull ship by using the wave elevation at each individual centre line and replacing the heave motion z by: and z(ps) = z + yT ¢ Á z(sb) = z ¡ yT ¢ Á .

1972] becomes: ¶ Z µ 0 Vz¤ 2 (ps) + Vz¤ 2 (sb) Raw ¡k cos ¹ dM33 0 a N33 ¡ V ¢ ¢ a = ¢ ¢ dxb 2! e dxb ³2 ³2 a a L with: +kg ¤ _¤ ¢ ³ a3 sin(! e t ¡ kxb cos ¹ ¡ kyT sin ¹) ³ w3 (ps) = ! +kg ¤ _¤ ³ w3 (sb) = ¢ ³ a3 sin(! e t ¡ kxb cos ¹ + kyT sin ¹) ! ³ ´ _¤ _ _ Vz¤ (ps) = ³ w3 (ps) ¡ z ¡ xb ¢ µ + V ¢ µ ¡ yT ¢ Á _ ³ ´ _¤ _ _ Vz¤ (sb) = ³ w3 (sb) ¡ z ¡ xb ¢ µ + V ¢ µ + yT ¢ Á _ 14.2 Integrated Pressure Method The transfer function of the mean added resistance of twin-hull ships according to the method of [Boese.5.268 CHAPTER 14. TWIN-HULL SHIPS 14.1 Radiated Energy Method The transfer function of the mean added resistance of twin-hull ships according to the method of [Gerritsma and Beukelman.5. 1970] becomes: Raw2 = ³2 a 1 za (ps) µa + ½r! 2 ¢ ¢ cos("z(ps)³(ps) ¡ "µ³(ps) ) e 2 ³a ³a 1 za (sb) µa ¢ ¢ cos("z(sb)³(sb) ¡ "µ³(sb) ) + ½r! 2 e 2 ³a ³a ¶ Z µ 2 zxa (ps) 2sa (ps) ¢ cos(¡kxb cos ¹ ¡ kyT sin ¹ ¡ "s(ps)³(ps) ) dyw 1 + ½g 1¡ ¡ dxb 2 ³ a ¢ tanh(kh) dxb ³2 a L ¶ Z µ 2 zxa (sb) 2sa (sb) ¢ cos(¡kxb cos ¹ ¡ kyT sin ¹ ¡ "s(sb)³(sb) ) dyw 1 1¡ + ½g ¡ dxb 2 ³ a ¢ tanh(kh) dxb ³2 a L with: ³(ps) ³(sb) zx (ps) zx (sb) s(ps) s(sb) = = = = = = ³ a cos(! e t ¡ kxb cos ¹ ¡ kyT sin ¹) ³ a cos(! e t ¡ kxb cos ¹ + kyT sin ¹) z ¡ x b ¢ µ + yT ¢ Á z ¡ xb ¢ µ ¡ yT ¢ Á ³(ps) ¡ z + xb ¢ µ ¡ yT ¢ Á ³(sb) ¡ z + xb ¢ µ + yT ¢ Á 14. the harmonic lateral. vertical and torsional dynamic loads per unit length on the unfastened disk of a twin-hull ship are given by: 0 0 qT 1 (xb ) = +XT h2 (xb ) + XT w2 (xb ) .6 Bending and Torsional Moments According to Newton’s second law of dynamics.

14. BENDING AND TORSIONAL MOMENTS ³ ´ ¡m0T (xb ) ¢ x ¡ bG ¢ Ä Ä µ 269 0 0 qT 2 (xb ) = +XT h2 (xb ) + XT w2 (xb ) + 2½gAs ¢ Á ³ ´ 0 Ä Ä ¡m0T (xb ) ¢ y + xb ¢ Ã ¡ zm ¢ Á + g ¢ Á Ä 0 0 qT 3 (xb ) = +XT h3 (xb ) + XT w3 (xb ) ³ ´ 0 Ä Ä ¡mT (xb ) ¢ z ¡ xb ¢ µ 0 0 qT 4 (xb . z1 ) = +XT h4 (xb ) + XT w4 (xb ) ³ 2 ´ 0 0 Ä Ä ¡m0T (xb ) ¢ kT xx ¢ Á ¡ zm ¢ (Ä + xb ¢ Ã + g ¢ Á) y +z1 ¢ qT 2 (xb ) In here: m0T 0 kT xx As = the mass per unit length of the twin-hull ship = the local sold mass gyradius for roll = sectional area of one hull The calculation procedure of the forces and moments is similar to the procedure given before for mono-hull ships.6. .

270 . CHAPTER 14. TWIN-HULL SHIPS .

1.1-a. are described in more detail here.1 First Degree Polynomials A …rst degree . February 2001. 2628 CD Delft. Ship Hydromechanics Laboratory. Mekelweg 2.nl. is de…ned by: f (x) = ax + b with in the interval xm < x < x0 the following coe¢cients: J. 15. see …gure 15.shipmotions. 0 271 . Figure 15.19”.nl/~johan or http://www. as used in the strip theory program SEAWAY.1: First and Second Order Polynomials Through Discrete Points 15. Journée.Chapter 15 Numerical Recipes Some typical numerical recipes.J. Release 4.M. Delft University of Technology.1-a. The Netherlands.1 Polynomials Discrete points can be connected by a …rst degree or a second degree polynomial.b.polynomial.tudelft.wbmt. For updates see web site: http://dutw189. as given in …gure 15.or linear . Report 1216a. ”Theoretical Manual of SEAWAY.

In those cases the trapezoid rule has to be preferred. NUMERICAL RECIPES a = f (x0 ) ¡ f (xm ) x0 ¡ xm b = f (x0 ) ¡ ax0 and in the interval x0 < x < xp the following coe¢cients: f (xp ) ¡ f (x0 ) xp ¡ x0 b = f (x0 ) ¡ ax0 a = Notify that only one interval is required for obtaining the coe¢cients in that interval. SEAWAY uses Simpson’s general rule as a standard.272 CHAPTER 15. Then. the integrations have to be carried out over a number of sets of two intervals.1. 15.2 Integrations Numerical integrations can be carried out by either the trapezoid rule or Simpson’s general rule. based on the following requirements: x0 ¡ xm x0 ¡ xm < 0:2 or: > 5:0 xp ¡ xm xp ¡ xm x0 ¡ xm < 5:0 0:2 < xp ¡ xm Trapezoid rule if: Simpson’s rule if: .1-b. SEAWAY makes the choice bewteen the use of the trapezoid rule andr Simpson’s rule automatically.2 Second Degree Polynomials A second degree polynomial. as given in …gure 15. Numerical inaccuracies can be expected when x0 ¡xm ¿ xp ¡x0 or xp ¡x0 ¿ x0 ¡xm . valid in both intervals. see …gure 15. see …gure 15.1-b. 15.1-a. is de…ned by: f (x) = ax2 + bx + c with in the interval xm < x < xp the following coe¢cients: f (xp )¡f (x0 ) xp ¡x0 a = xp ¡ xm f (xp ) ¡ f (x0 ) b = ¡ a (xp + x0 ) xp ¡ x0 c = f (x0 ) ¡ ax2 ¡ bx0 0 ¡ f (x0 )¡f (xm ) x0 ¡xm Notify that two intervals are required for obtaining these coe¢cients.

see …gure 15.2. the integrand is described by a second degree polynomial: f (x) = ax2 + bx + c Then the integral becomes: x Zp x Zp f (x) dx = xm xm = with: · ¸xp 1 3 1 2 ax + bx + cx 3 2 xm ¡ ¢ ax2 + bx + c dx a = xp ¡ xm f (xp ) ¡ f (x0 ) b = ¡ a (xp + x0 ) xp ¡ x0 c = f (x0 ) ¡ ax2 ¡ bx0 0 f (xp )¡f (x0 ) xp ¡x0 ¡ f (x0 )¡f (xm ) x0 ¡xm .1-b have to be carried out over a set of two intervals.2 Second Degree Integrations Second degree integrations . At each of the two intervals.2.1-a .2. see …gure 15.15. INTEGRATIONS 273 15.1 First Degree Integrations First degree integrations .means the use of a linear function: f (x) = ax + b The integral over the interval xp ¡ x0 becomes: x Zp f (x) dx = x0 x0 x Zp (ax + b) dx = with: a = · ¸xp 1 2 ax + bx 2 x0 f (xp ) ¡ f (x0 ) xp ¡ x0 b = f (x0 ) ¡ ax0 Integration over two intervals results into: x Zp (x0 ¡ xm ) f (xm ) + (xp ¡ xm ) f (x0 ) + (xp ¡ x0 ) f (xm ) f (x) dx = 2 xm 15.integrations carried out by Simpson’s rule.integrations carried out by the trapezoid rule.

2 in terms of fw1. These functions 0 fw1. ¢xb .2 (xb ) can be approximated by second degree polynomials: f (x) = ax2 + bx + c When making use of the general integral rules: Z cos x dx = + sin x Z and: Z x cos x dx = + cos x + x sin x ¡ ¢ x2 cos x dx = +2x cos x + x2 ¡ 2 sin x . 0 0 0 This can be avoided by writing Fw1.3 Integration of Wave Loads The wave loads can be written as: Fw = Fwa cos (! e t + "Fw ³ ) The in-phase and out-phase parts of the wave loads have to be obtained from longitudinal integrations: Z L Z L Fw1 = Fw2 = Fw1 (xb ) dx = Fw2 (xb ) dx = 0 0 0 0 Z L Z L fw1 (xb ) cos xb dxb fw2 (xb ) sin xb dxb 0 0 Direct numerical integrations of Fw1 and Fw2 over the ship length. L.2 (xb ) cos xb and fw1.2 (xb ) vary very slow over short wave lengths. This means that a large number of cross sections are required. require integration intervals. in 0 which the integrands fw1. which are much smaller than the smallest wave length.2. ¢xb 6 ¸min =10.2 (xb ) sin xb . NUMERICAL RECIPES Some algebra leads for the integration over these two intervals to: x Zp f (x) dx = xm ( x0 ¡ xm ¡ xp ¡x0 2 f (xm ) x0 ¡ xm (xp ¡ xm )2 f (x0 ) + 2 (x0 ¡ xm ) (xp ¡ x0 ) ¾ xp ¡ x0 ¡ x0 ¡xm xp ¡ xm 2 + f (xp ) ¢ xp ¡ x0 3 15.274 CHAPTER 15.

b. of which the derivatives have to be determined.3.2-a. . integrated from xm through xp .15.3 Derivatives First and second degree functions. have been given in …gure 15. so over the two intervals x0 ¡ xm and xp ¡ x0 : x Zp Z Z sin x dx = ¡ cos x ¡ ¢ x2 sin x dx = +2x sin x ¡ x2 ¡ 2 cos x x sin x dx = + sin x ¡ x cos x F (x) dx = xm = xm x Zp xm x Zp f (x) cos x dx ¡ ¢ ax2 + bx + c cos x dx x2 cos x dx + b x Zp = a xm x Zp x cos x dx + c xm xm x Zp cos x dx = [+ (f (x) ¡ 2a) sin x + (2ax + b) cos x]xp xm x Zp F (x) dx = xm = xm x Zp xm x Zp f (x) sin x dx ¡ ¢ ax2 + bx + c sin x dx x sin x dx + b 2 x Zp = a xm x Zp x sin x dx + c xm xm x Zp sin x dx with coe¢cients obtained by: = [¡ (f (x) ¡ 2a) cos x + (2ax + b) sin x]xp xm f (xp )¡f (x0 ) xp ¡x0 f (x0 )¡f (xm ) x0 ¡xm a = xp ¡ xm f (xp ) ¡ f (x0 ) b = ¡ a (xp + x0 ) xp ¡ x0 c = f (x0 ) ¡ ax2 ¡ bx0 0 ¡ 15. DERIVATIVES Z 275 the following expressions can be obtained for the in-phase and out-phase parts of the wave loads.

3.will di¤er: # "µ ¶ # "µ ¶ df df 6= dx x=x0 dx x=x0 m (minus or left of x0 ) p (plus or right of x0 ) A mean derivative at x = x0 can be obtained by: i i h¡ ¢ h¡ ¢ df df µ ¶ (x0 ¡ xm ) ¢ dx x=x0 + (xp ¡ x0 ) ¢ dx x=x0 df m p = dx x=x0 xp ¡ xm df dx 15.2: Determination of Longitudinal Derivatives 15.and the derivative at the right hand side of x0 .with index m (minus) .1 First Degree Derivatives The two polynomials . generally. the derivative at the left hand side of x0 .3. NUMERICAL RECIPES Figure 15.276 CHAPTER 15.are given by: for x < x0 : for x > x0 : The derivative is given by: df (x) = am dx df (x) = ap dx f (x) = am x + bm f (x) = ap x + bp for x < x0 : for x > x0 : It is obvious that.2 Second Degree Derivatives The two polynomials .with index p (plus) .each valid over two intervals below and above x = x0 .are given by: for x < x0 : for x > x0 : f (x) = am x2 + bm x + cm f (x) = ap x2 + bp x + cp .each valid over two intervals below and above x = x0 .

with index m (minus) . DERIVATIVES A derivative of a second degree function: f (x) = ax2 + bx + c is given by: df (x) = 2ax + b dx This leads for x < x0 to: µ ¶ ¶ · 277 µ df dx df dx = x=xm2 x=xm1 µ df dx ¶ x=x0 and for x > x0 to: µ (xm1 ¡ xm2 )2 ff (x0 ) ¡ f (xm1 )g + (x0 ¡ xm1 )2 ff (xm1 ) ¡ f (xm2 )g = (xm1 ¡ xm2 ) (x0 ¡ xm1 ) (x0 ¡ xm2 ) · ¸ +2 (xm1 ¡ xm2 ) (x0 ¡ xm1 ) ff (x0 ) ¡ f (xm1 )g + (xm1 ¡ xm2 )2 ff (x0 ) ¡ f (xm1 )g ¡ (x0 ¡ xm1 )2 ff (xm1 ) ¡ f (xm2 )g = (xm1 ¡ xm2 ) (x0 ¡ xm1 ) (x0 ¡ xm2 ) · ¸ ¡2 (xm1 ¡ xm2 ) (x0 ¡ xm1 ) ff (xm1 ) ¡ f (xm2 )g ¡ (xm1 ¡ xm2 )2 ff (x0 ) ¡ f (xm1 )g + (x0 ¡ xm1 )2 ff (xm1 ) ¡ f (xm2 )g (xm1 ¡ xm2 ) (x0 ¡ xm1 ) (x0 ¡ xm2 ) ¸ ¶ df = dx x=x0 µ ¶ df (xp1 ¡ x0 )2 ff (xp2 ) ¡ f (xp1 )g + (xp2 ¡ xp1 )2 ff (xp1 ) ¡ f (x0 )g = dx x=xp1 (xp1 ¡ x0 ) (xp2 ¡ xp1 ) (xp2 ¡ x0 ) · ¸ +2 (xp1 ¡ x0 ) (xp2 ¡ xp1 ) ff (xp2 ) ¡ f (xp1 )g µ ¶ + (xp1 ¡ x0 )2 ff (xp2 ) ¡ f (xp1 )g ¡ (xp2 ¡ xp1 )2 ff (xp1 ) ¡ f (x0 )g df = dx x=xp2 (xp1 ¡ x0 ) (xp2 ¡ xp1 ) (xp2 ¡ x0 ) Generally.15.with index p (plus) .will di¤er: "µ ¶ "µ ¶ # # df df 6= dx x=x0 dx x=x0 m (minus or left of x0 ) p (plus or right of x0 ) ¡2 (xp1 ¡ x0 ) (xp2 ¡ xp1 ) ff (xp1 ) ¡ f (x0 )g ¡ (xp1 ¡ x0 )2 ff (xp2 ) ¡ f (xp1 )g + (xp2 ¡ xp1 )2 ff (xp1 ) ¡ f (x0 )g (xp1 ¡ x0 ) (xp2 ¡ xp1 ) (xp2 ¡ x0 ) A mean derivative df dx at x = x0 can be obtained by: i i h¡ ¢ h¡ ¢ df df µ ¶ dm ¢ dx x=x0 + dp ¢ dx x=x0 df m p = dx x=x0 dm + dp ¢ x0 ¡ xm1 ¡ xm1 ¡xm2 (x0 ¡ xm2 ) 2 = 3 (x0 ¡ xm1 ) ¡ ¡x ¢ xp1 ¡ x0 ¡ xp2 2 p1 (xp2 ¡ x0 ) = 3 (xp1 ¡ x0 ) ¡ with: dm dp .and the derivative at the right hand side of x0 . the derivative at the left hand side of x0 .3.

3: First and Second Order Curves The curve length follows from: x Zp smp = ds p = xm x Zp xm dx2 + dy 2 15.4. Figure 15. see …gure 15.3-b. are given in …gure 15.1 First Degree Curves The curve length of a …rst degree curve. NUMERICAL RECIPES 15.3-a. in the two intervals in the region xm < x < xp is: ½ q ¸ ¸¾ · · q q q 2 2 2 2 smp = p2 p0 1 + p0 + ln p0 + 1 + p0 ¡ p1 1 + p1 ¡ ln p1 + 1 + p1 with: cos ® = q sin ® = q (xp ¡ xm )2 + (yp ¡ ym )2 yp ¡ ym (xp ¡ xm )2 + (yp ¡ ym )2 xp ¡ xm .b.4.3-a.2 Second Degree Curves The curve length of a second degree curve. connected by a …rst degree or a second degree polynomial.4 Curve Lengths Discrete points.278 CHAPTER 15. see …gure 15. in the two intervals in the region xm < x < xp is: q q smp = (x0 ¡ xm )2 + (y0 ¡ ym )2 + (xp ¡ x0 )2 + (yp ¡ y0 )2 15.

15.4. CURVE LENGTHS p0 = ¼ + 2 ¢ p1 p2 (xp ¡ xm ) cos ® + (yp ¡ ym ) sin ® (xp ¡ x0 ) cos ® + (yp ¡ y0 ) sin ® (y0 ¡ ym ) cos ® ¡ (x0 ¡ xm ) sin ® (x0 ¡ xm ) cos ® + (y0 ¡ ym ) sin ® = ¡ (x0 ¡ xm ) cos ® + (y0 ¡ ym ) sin ® (xp ¡ x0 ) cos ® + (yp ¡ y0 ) sin ® (xp ¡ x0 ) cos ® + (yp ¡ y0 ) sin ® = 4 279 .

CHAPTER 15. NUMERICAL RECIPES .280 .

DELFRAC. W. A. d. A. T. The Netherlands. (1983). (1985). [Dimitrieva. Athens. J. M. [Faltinsen and Svensen. L. Eine Einfache Methode zur Berechnung der Wiederstandserhöhung eines Schi¤es in Seegang. Standards of Good Seakeeping for Destroyers and Frigates in Head Seas. Naval Ship Research and Development Centre. 281 . BRD. Analysis of the Resistance Increase in Waves of a Fast Cargo-ship. Berechnung der durch Schwingungen eines Schi¤skörpers Erzeugten Hydrodynamischen Kräfte. 18(217). 32(366). H. Journal of the Society of Naval Architects of Japan. Shipbuilding Department. Ship Hydromechanics Laboratory. 1983] Bales. 3-D Potential Theory Including Wave Di¤raction and Drift Forces Acting on the Structures. 1967] Frank. International Shipbuilding Progress.Bibliography [Athanassoulis and Loukakis. O. Greece. Washington DC. 1966] Bosch. 1962] Fukuda. Oscillation of Cylinders in or below the Free Surface of Deep Fluids. Coupled Motions and Midship Bending Moments of a Ship in Regular Waves. Technical Report 1017. Incorporation of Seakeeping Theories in CAD. Prepared for Seakeeping Committee ITTC. number 8. [Gerritsma and Beukelman. International Shipbuilding Progress. T. Jahrbuch der Schi¤sbautechnischen Gesellschaft. 1985] Athanassoulis. [Fukuda. J. E. London. [Grim. (1974). 1994] Dimitrieva. Roll Damping by Free Surface Tanks. (1994). 1990] Faltinsen. J. In International Symposium on CFD and CAD in Ship Design. Institüt für Schi¤bau der Universität Hamburg. Delft University of Technology. The Netherlands. An Extended-Lewis Form Family of Ship Sections and Its Applications to Seakeeping Calculations. [Bales. (1953). Wind and Wave Data for Seakeeping Performance Assessment. Technical Report 83-S. J. UK. J. (1972). (1967). I. In International Symposium on the Dynamics of Marine Vehicles and Structures in Waves. 1970] Boese. Delft. 1953] Grim. v. W. 1972] Gerritsma. (1970). and Vugts. and Svensen. G. Technical Report 2375. Technical report. P. Netherlands Ship Research Centre TNO. USA. [Conolly. [Bosch and Vugts. Technical Report 258. O. [Frank. S. J. and Beukelman. and Loukakis. (1966). (1962). [Boese. 47:277–299. 1974] Conolly. (1990).

O. (Internet: http://dutw189... Himeno. (Internet: http://dutw189. (1991). Motions of Rectangular Barges. Veri…cation and Validation of Ship Motions Program SEAWAY. [John. 1973] Jong. Ship Hydromechanics Laboratory. J. pages 45–100.19). J. M. A Prediction Method for Ship Rolling. Y. Part II. The Netherlands. Technical Report HSVA-Bericht 1090. In Proceedings 10th International Conference on O¤shore Mechanics and Arctic Engineering. Delft University of Technology. Die Hydrodynamischen Kräfte beim Rollversuch. (2001a). Die Schwingungen von schwimmenden zweidimensionalen Körper. Netherlands Ship Research Centre TNO. (1957). Delft. (1996).nl/~johan or http://www. Y. Varna. Technical Report 145-S.tudelft. J. and van ’t Veer. Japan. Motions and Drift Forces on Moored Vessels in Current. (1955). Technical Report 912. 3. M. d.wbmt. The Netherlands. (Internet: http://dutw189. Shipbuilding Department. HSVA.tudelft. H. Germany. J.tudelft. Delft University of Technology. O. Technical Report HSVA-Bericht 1171. Hamburg. J. M. 1955] Grim. (1992). PhD thesis. 1996] Huijsmans. J. .nl). In Proceedings 5th International O¤shore and Polar Engineering Conference. Otdelenie Tskhnicheshikh Nauk. 1950] John. (2001b). Revised Report 1992. [Journée and van ’t Veer.nl). 1992] Journée. [Jong.nl/~johan or http://www. Stavanger.tudelft. 1991] Journée. (1957).nl). B.shipmotions. In STAB’97 Conference.nl/~johan or http://www. 1956] Grim. Schi¤stechnik. 1995] Journée. The Exciting Forces and Wetting of Ships in Waves (in Russian). 1957] Haskind. The Netherlands. Delft University of Technology. J. 1997] Journée.nl/~johan or http://www. [Haskind. Computation of Hydrodynamic Coe¢cients of Oscillating Cylinders. (1978). Die Schwingungen von schwimmenden zweidimensionalen Körper. M. R.282 BIBLIOGRAPHY [Grim. J. M. Technical Report 00405. University of Osaka Prefecture. A.wbmt. [Journée. (1997). Germany. 1957] Grim. [Journée.shipmotions. J.wbmt.shipmotions.nl). Bulgaria. and Tanaka. Department of Naval Architecture. [Journée. [Grim. Izvestia Akademii Nauk SSSR. [Huijsmans. [Journée. J. F. 2001a] Journée. (1950). (1973). M.wbmt. On the Motions of Floating Bodies. M. First Order Wave Loads in Beam Waves. Technical Report 1213a.. [Journée. (7):65–79. on Pure and Applied Mathematics. Ship Hydromechanics Laboratory. J. (1995). Comm. 2001b] Journée. HSVA. Technical Report 1212a. J. The Hague. The Netherlands. Delft University of Technology. O. M. P. N. User Manual of SEAWAY (Release 4. (1956). Norway. Strip Theory Algorithms. [Ikeda et al. (Internet: http://dutw189. The Netherlands. [Grim. Liquid Cargo and Its E¤ect on Ship Motions. Ship Hydromechanics Laboratory.shipmotions. Hamburg. 1978] Ikeda.

T. Added Mass and Damping Coe¢cients for Cylinders Oscillating in a Free Surface. Symposium. Die Hydrodynamische Kräfte bei der periodischen Bewegung zweidimensionaler Körper an der Ober‡äche ‡acher Gewasser. Institute of Engineering Research. 1929] Lewis. Number ISBN 0-7458-0230-3. 1959] Kumai. [Ochi. Journal of Ship Research. BRD. Dallas. [Porter. B. Dover Publications. pages 451–472. 1974] Miller. M. 6(4):10–17. R. and Hebert. 1989] Lloyd. Cooper Street. Technical Report 6136-74-280. Journal of Fluid Mechanics. 1957] Korvin-Kroukovsky.R. West Sussex. 1969] Kerczek. [Kumai. (1962). (1929). Prediction of Occurence and Severity of Ship Slamming at Sea. The Exciting Forces on Fixed Bodies in Waves. In Proceedings of 5th O. M. University of California. P. University of Hamburg. 65:590–632. R. [Lewis. [Lloyd. M. Journal of Ship Research. Hydrodynamics. and Kaplan. Ship Behaviour in Rough Weather. and Jacobs. A. (1960). 18(2). (1974). (1974). In Transactions SNAME. [Newman. E. . (1989). [Miller. New York. Pressure Distribution. Ellis Horwood Limited. (1974). Texas. 1974] Sargent. M. Oceanics Inc. M. USA. W. Added Mass Moment of Inertia Induced by Torsional Vibration of Ships. [Korkut and Hebert. USA. (1970). and Nowacki. V. The Representation of Ship Hulls by Conformal Mapping Functions. Technical Report 74103. H.BIBLIOGRAPHY 283 [Keil. 1970. Reports of the Research Institute for Applied Mechanics. (1959). T. VII(28). (1932). (1963). T. (1957).. Modi…cations to Lloyds Register of Shipping Strip Theory Computerprogram (LR 2570). [Sargent and Kaplan. 13(4).N. 1960] Porter. Norway. R. (1974). P. Berkeley. H. (1969). F. Market Cross House. USA. A. Unknown Title of a Report on Roll Damping. H. [Korvin-Kroukovsky and Jacobs. Transactions SNAME. 1964] Ochi. NAVSPEC. [Kerczek and Tuck. Inc. E. 1974] Keil. California. Technical Report 305. First and Second Order Forces on a Cylinder Submerged under a Free Surface. J. Dallas. [Ogilvie. 1962] Newman.. Pitching and Heaving Motions of a Ship in Regular Waves. Technical Report 82. W. 1974] Reed. number OTC 1160. 1963] Ogilvie. v. K. C. In O¤shore Technology Conference. 6 edition. and Tuck. N. [Reed and Nowacki. Bergen. Interactive Creation of Fair Ship Lines. The Inertia of Water Surrounding a Vibrating Ship. Some Notes on Static Anchor Chain Curve. 1932] Lamb. F. Seakeeping. (1964). [Lamb. E. R. O. Texas. 1970] Korkut. P019 1EB England. Chichester. J. Journal of Ship Research.

Journal of Fluid Mechanics. II. Fin and Tank Roll Stabilizers. Kyushu University. Vol. 1965] Tasai. F. Canada. R. No 26. Research Institute for Applied Mechanics. [Verhagen and van Wijngaarden.S. Non-linear oscillations of ‡uid in a container. Japan. Formula for Calculating Hydrodynamic Force on a Cylinder Heaving in the Free Surface. Sway and Yaw Motions in Oblique Waves. 1969] Tasai. (1949). (1969). Researches on Seakeeping Qualities of Ships in Japan. (1962). On the Shipment of Water in Head Waves. New England Section of the Society of Naval Arcitects and Marine Engineers. 1960] Tasai. 1949] Ursell. Japan. [Timman and Newman. (1955). P. [Troost. [Tasaki. On the Damping Force and Added Mass of Ships Heaving and Pitching. d. 1978] Schmitke. Dartmouth N.T. VIII. (1963). Department for Shipbuilding and Shiphandling. Atlantic. Vol. N. XIII. Research Institute for Applied Mechanics. The Netherlands. Technical report. F. F. (N-Parameter Family). IX. Vol. [Zwaan. 1959] Tasai. R. Defence Research Establishment.T. and van Wijngaarden. October Meeting. 1965] Verhagen. 5(4):1–7. Technical report. 1955] Troost. . (1965). [Tasai. Kyushu University. In Proceedings 12th I. Delft University of Technology. A Simpli…ed Method for Preliminary Powering of Single Screw Merchant Ships. J. VII. F.C. Vol. Method for Solving Equations with Constant or Variable Right Hand Terms (in Dutch). [Tasai. F.. L. J. (1959). [Tasai. No 45. A. No 35. The Coupled Damping Coe¢cients of a Symmetric Ship. [Tasai. (1965). Journal of the Society of Naval Architects of Japan. a Fortran Program to Predict Ship Roll. Japan. Research Institute for Applied Mechanics. Improvements in the Theory of Ship Motions in Longitudinal Waves. Journal of Ship Research. Model Experiments in Waves. 1962] Timman. F. Technical report. L. R. On the Heaving Motion of a Circular Cylinder on the Surface of a Fluid. Centrale Werkgroep Wiskunde. No 31. (1960). T. [Tasai.284 BIBLIOGRAPHY [Schmitke. ROLLRFT. 8. 1963] Tasaki. Technical Report 78/G. Technical Report 5. Quarterly Journal of Mechanics and Applied Mathematics. 1977] Zwaan. 22(4):737–751. [Ursell. (1978). Kyushu University. Japan. Hydrodynamic Force and Moment Produced by Swaying and Rolling Oscillation of Cylinders on the Free Surface. Including the E¤ect of Rudder. Technical report. Ship Motions in Beam Waves. (1961). and Newman. Research Institute for Applied Mechanics. (1977). 1961] Tasai. Kyushu University.

Sign up to vote on this title
UsefulNot useful