SEAKEEPING: Ship behaviour in rough weather ARJMLLOYD BSc PhD FEng FRINA©ARIM Lloyd 1998 Published by A RJ M Lloyd, 26 Spithead Avenue, Gosport, Hampshire, United Kingdom First published 1989 by Ellis Horwood, Chichester, Sussex, United Kingdom. Printed and bound in Great Britain by RPM Reprographics, Chichester, Sussex, United Kingdom. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. ISBN 0 9532634 0 1 (Revised edition) First published by Ellis Horwood Ltd (ISBN 0 7458 0230 3) Halstead Press (ISBN 0 470 21232 2) International Book Distributors Ltd (ISBN 0 13798562 2) All rights reserved. ‘No part of this publication may be reproduced, stored in a retrieval system, or transmitted , in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the written permission of the publisher, Cover photograph: The view from the bridge of HMS Hermione as she takes a green sea over the fo'c’s'le during the Comparative Seakeeping Trial (Andrew and Lloyd (1981)). Photo Paul Wilson.Dedicated to the memory of No 1 Ship Tank at the Admiralty Experiment Works, Haslar, Gosport 1886 - 1993CONTENTS FOREWARD 1. REGULAR WAVES 1.1 Introduction 1.2 The velocity potential 1.3 Pressure contours and the surface profile 1.4 Wave slope 1.5 Particle orbits 1.6 Pressure fluctuations under a wave 1.7 Energy in a regular wave 1.8 Energy transmission and group velocity 1.9 Summary of the characteristics of regular waves 1.10 Encounter frequency and heading 2, OCEAN WAVE SPECTRA AND STATISTICS 2.1 Wave generation 2.2 Statistical analysis of time histories of irregular waves 2.3 Fourier analysis 2.4 The wave energy spectrum 2.5 Generating a time history from a spectrum 2.6 Spectral moments 2.7 Mean periods 2.8 Spectrum bandwidth 2.9 Idealised wave spectra 2.10 Wave slope spectra 2.11 Wave spreading and short crested waves 2.12 Ocean wave statistics 3. LINEARISED EQUATIONS FOR SMALL AMPLITUDE SHIP MOTIONS IN REGULAR WAVES 3.1 Introduction 3.2 Axes and ship motion definitions 3.3 General equations for ship motions in regular waves 3.4 — Coefficients in the equations of motion 3.5. Simplified equations of motion for a ship with port/starboard symmetry 3.6 Coupling 12 12 13 14 19 20 23 25 28 29 35 35 37 39 4t 42 47 49 53 55 58 68 68 68 na 83 8344 42 43 44 5. 51 5.2 53 5.4 55 5.6 6.1 62 63 64 65 66 * TL 72 13 14 75 16 1 18 19 7.10 TW a1 82 83 84 85 8.6 87 STRIP THEORY Introduction Strip motions Coefficients in the equations of motion Excitations in regular waves LEWIS FORMS AND THEIR HYDRODYNAMIC. PROPERTIES Introduction Lewis forms ‘Added mass and damping coefficients for a heaving Lewis form Added mass and damping coefficients for a swaying Lewis form Added mass and damping coefficients for a rolling Lewis form ‘Measurements of local hydrodynamic properties ROLL DAMPING Sources of roll damping Non linear roll damping: equivalent linearisation Eddy roll damping Skin friction roll damping Appendage roll damping Total roll damping SHIP MOTIONS IN REGULAR WAVES Introduction Transfer functions ‘Vertical plane motions in regular head waves Vertical plane motions in regular following waves Vertical plane motions in regular oblique waves Alternative transfer function presentations Lateral plane motions in regular beam waves Lateral plane motions in regular oblique waves Absolute motions Relative motions ‘Velocities and accelerations SHIP MOTIONS IN IRREGULAR WAVES ‘The electronic filter analogy ‘The encountered wave spectrum ‘The motion energy spectrum Runs velocities and accelerations Effect of matching the wave spectrum and the transfer function Motions in short crested waves Spectral calculations for non linear motion responses 84, 84, 85 87 93 104 104 106 i 113 17 118 122 122 123 124 126 129 131 132 132 132 133 137 138 141 142 145 148 149 150 152 152 153 154 158 159 160 164.91 92 93 94 95 10. 10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 10.9 10.10 10.11 i. ld 112 3 4 115 11.6 17 118 12. 12.1 12.2 123 12.4 12.5 12.6 12.7 12.8 129 12.10 12.11 12.12 12.13 12.14 12.15 12.16 SEAKEEPING TRIALS Fall scale trials, Wave measurements Ship motion measurements ‘Measurements of other seakeeping responses Run lengths and ship courses MODEL TESTING ‘Model seakeeping experiments Model experiment scaling Open water model experiments Laboratory test facilities Wave makers and beaches Instrumentation ‘Model materials ‘Trimming and ballasting ‘Testing in regular waves ‘Testing in irregular waves ‘Tank wall interference PROBABILITY FORMULAE Introduction Probability analysis Histograms The probability density function The Gaussian or normal probability density function The Rayleigh probability density function Significant wave height and related statistics, Joint probabilities ROLL STABILISATION Motion reduction Bilge keels Active roll stabiliser fins Hydrodynamic characteristics of stabiliser fins Constraints on stabiliser fin outreach Equations of motion for a ship with stabiliser fins Stabiliser fin losses Design recommendations for stabiliser fins and bilge keels Active fin control systems System stability Active roll stabiliser fin performance Passive tanks Theory for a U tube passive tank Passive tank natural frequency and damping Design of passive stabilising tanks Passive tank characteristics and design recommendations 6 165 165 167 168 169 170 172 172 172 183 184 187 189 193 194, 198 203 208 210 210 210 210 215 218 221 226 227 227 228 230 231 234 235 237 244 244 253 258 258 260 270 212.17 1B. 13.1 13.2 13.3 13.4 13.5 13.6 14, 14.1 14.2 143 144 14.5 14.6 147 14.8 15. 15.1 15.2 153 15.4 16. 16.1 16.2 16.3 16.4 16.5 16.6 1. 171 17.2 173 174 175 Performance of passive stabilising tanks ADDED RESISTANCE AND INVOLUNTARY SPEED LOSS IN WAVES Introduction Simple theory for added resistance in regular head waves ‘Added resistance in irregular head waves Added resistance due to wind Propeller characteristics Involuntary speed loss SLAMMING, DECK WETNESS AND PROPELLER EMERGENCE Introduction Probability of occurrence Slamming Design recommendations to avoid slamming Deck wetness Freeboard exceedance Effect of bow shape Design recommendations to minimise deck wetness EEFECTS OF SHIP MOTIONS ON PASSENGERS AND CREW Introduction ‘Motion sickness incidence Subjective magnitude Motion induced interruptions SEAKEEPING CRITERIA AND VOLUNTARY SPEED LOSS IN ROUGH WEATHER Introduction Equipment criteria Questionnaires Analysis of numerical data from questionnaires Voluntary speed loss in rough weather Criteria for voluntary speed loss OPERATIONAL EFFECTIVENESS Introduction Season and sea area ‘Wave height and period Ship's course and speed Calculation of operational effectiveness 286 286 287 290 296 296 297 298 301 303 303 304 306 310 315 315 317 318 320 326 328 333 333 334 335 336 33718. THE EFFECT OF HULL SIZE AND FORM ON SEAKEEPING 18,1 Introduction 18.2 Vertical plane motions 18.3 Lateral plane motions 18.4 The Seakeeping Design Package 18.5 Design recommendations APPENDICES: Appendix 1 Appendix 2 ‘Appendix 3 ‘Appendix 4 ‘Appendix 5 Appendix 6 Index ‘The spring mass system Excitation amplitudes and phases Bibliography Notation ‘Annote on units and numerical values Glossary Biographical notes 340 340 340 353 354 358 359 368 369 373 387 388 389 305FOREWARD Seakeeping There are three things which are too wonderful for me, yea, four which I know not: the way of an eagle in the air; the way of a serpent on a rock; the way of a ship in the midst of the sea; and the way of a man with a maid. : Proverbs Chapter 30 verses 18-19 In the days of sail ships were very much more dependent on the weather than they are today. Square rigged sailing vessels could not sail directly into the wind and were strictly limited in their ability to go where the master wanted. In severe conditions it was necessary to shorten sail and even to ride out a storm under bare poles. Many a ship was lost because she was driven ashore under such circumstances. Economic pressures often demanded that the ship's master spread as much canvas as be dared in order to make the best speed. This is nowhere more graphically illustrated than in the stories of the clipper “races” from China to Europe in the nineteenth century. The first ship home with the newly harvested tea crop could demand a premium price for her cargo. Speed was of the essence and these ships sprouted all sorts of additional sails to make the most of every breath of wind available. A heavily laden over canvassed ship was an unpleasant home for the sailors and passengers in rough weather. With the lee gunwale submerged, the decks continually awash, deckhouses damp and cold, life must have been miserable, Yet even in these circumstances the crew would be expected to continue to navigate and steer the ship and to go aloft in order to shorten sail or spread additional canvas as the master demanded. However, the real problems of seakeeping only came to be recognised with the demise of sail and the advent of steam as the prime motive power. Now, for the first time, ships could stezm directly into the wind and sea with a consequent increase of pitch and heave motions. The damaging effects of shipping heavy seas over the bow began to be experienced. The punishing consequences of high speed in rough weather were not fully understood and at least one ship (HMS Cobra in 1901) is believed to have been lost after her hull broke in two after slamming in rough weather. At the same time the steadying effects of tall masts and a good spread of canvas were lost and the new steam ships were found to roll heavily. Itis ironic that this beneficial effect of sails has only recently been rediscovered with the emerging technology of wind assisted propulsion for low powered merchant ships. At about this time William Froude, an eminent Victorian engineer, proposed to build the world's first purpose built model towing tank at Torquay in the United Kingdom. He had recently developed scaling laws for predicting the resistance of ships from tests on models and he 9intended to use the tank for the required scale model experiments. The British Admiralty accepted Froude's proposal on the condition that he also used the tank to investigate ways of reducing the rolling motion of ships. In due course Froude’s Admiralty Experiment Works was moved to Haslar in Gosport and similar towing tanks were built in many different countries. These were often fitted with wave makers which allowed the behaviour of model ships in waves to be studied at leisure providing for the first time, a rudimentary technique for refining a full scale design to ensure adequate performance in rough weather. These model experiments were usually confined to tests in regular head or following waves with occasional tests at zero speed in beam waves. Tests at other headings or in more realistic irregular waves were impossible because of the long narrow shape of the towing tanks and the simplicity of the wave makers. ‘These early model experiments allowed some limited developments in the study of seakeeping but they could not be used to predict the actual performance of ships at sea because no technique for relating the behaviour of the model in the tank’s regular waves to the behaviour of the ship in the chaotic environment of the real ocean was available. Graphic evidence for this may be Found in contrasting the performance of many of the Royal Navy’s destroyers of the Second World War with that of their successors in the postwar period. The specifications for the wartime ships often called for speeds as high as 35 knots and the shipbuilder was required to demonstrate that this could be achieved before the Navy would accept the ship. . Invariably the trials were done in calm water and service experience soon demonstrated that these “trial” speeds could hardly ever be achieved in practice despite engines capable of developing as much as, 70,000 horse power (50 MW). Commonly occurring moderate wave conditions were usually enough to limit the speed in one way or another and it became obvious that the powerful engines were not cost effective. ‘The early postwar designs had half the installed power of their war time predecessors but much more attention was devoted to their performance in rough weather so that their reduced speed potential of perhaps 30 knots was achievable most of the time. The relationship between hull form and good seakeeping performance was not well understood and the classic design of the period, the Type 12 frigate (which evolved into the Leander class), had a superb rough weather performance as a result of the designer’ intuition rather than any precise science. This situation prevailed for many years and the study of seakeeping remained in effective limbo until the publication of a landmark paper by St Denis and Pierson in 1953'. This showed, for the first time, how the motions experienced in the random waves of the ocean could be calculated using the techniques of spectral analysis borrowed from the field of electromagnetic communications. At about the same time theoretical methods of predicting the behaviour of ships in regular waves were being developed. The breakthrough came with Ursell's (1949a,b) theory for predicting the characteristics of the flow around a circular cylinder oscillating in a free surface. Classical transformation techniques allowed these results to be applied to a wide range of shapes of ship * References are listed in the Bibliography in Appendix 3 10like cross-section and the fundamentals of modern ship motion theory were born, These developments some fifty years ago provided the basic tools required to develop routine techniques for the prediction of ship motions in something approaching the real irregular wave environment of the ocean. It was now possible for the first time to predict the rough weather performance of a ship at the design stage and to allow scakeeping to take its rightful place in the design process. Since that time seakeeping has remained an active field of research, but developments have been in the nature of progressive refinements rather than spectacular advances. Techniques for designing roll stabilisers, criteria determination, prediction of long term motion statistics and operational effectiveness have all been added to the naval architect's tool box: seakeeping performance prediction should now be a routine in any ship design office. Unfortunately these developments have not been accompanied by much readily obtainable literature on the subject outside the specialist papers and publications of the learned societies and research institutions. Although the underlying physical principles of seakeeping theory are not generally difficult to understand, the intimate details are mathematically complicated. It follows that calculations of ship motions and related phenomena require access to suitable computer programs and computers. No real progress can be made without them. Fortunately many suitable programs are available in educational, research and design establishments as well as computer bureaux throughout the world. The PAT-86 suite of seakeeping computer programs (available-at the Defence Evaluation and Research Agency at Haslar in the United Kingdom) ‘was used for the examples of ship motion calculations presented in this book. ‘The first edition of this book (Lloyd, 1989) covered all aspects of the subject in detail, resulting in a comprehensive, if rather long and expensive text. This edition is shorter and private publication: by the author has allowed the book to be marketed at a realistic price within the range of the average student of naval architecture. This has been achieved by revising the entire text and omitting material which can be readily found elsewhere such as the chapter on basic fluid mechanics and the detailed derivation of the dynamics of the spring mass system., The opportunity has been taken to include some new material on the growth of wave heights in the ocean, the effect of ship motions on crew task performance and the influence of hull form on seakeeping. ‘L was privileged to lead the Seakeeping Research Group at the Defence Evaluation and Research Agency (formerly the Admiralty Experiment Works) at Haslar in the United Kingdom for the best part of 25 years and this book represents the accumulated knowledge of that period. During that time Thave been helped by many colleagues both within the DERA and outside. In particular I would like to acknowledge the contributions of Mrs P R Loader, Dr RN Andrew, Dr D Fryer, the late Mr WB Marshfield, Mr P Crossland, Mr M Johnson and Mr D K Brown as well as fellow members of the ITTC Seakeeping Committee and colleagues at the United States Naval Academy. I am also indebted to Mr RE Small and Mrs W E Ball, whose painstaking proof reading of the manuscript was invaluable. Without their constant support, advice and inspiration this book would probably have never been written. ARJM Lloyd Gosport 1998 ulREGULAR WAVES 1.1 INTRODUCTION ‘The waves which influence the behaviour of ships at sea are generally irregular and more or less random in nature. No two waves have exactly the same height and they travel across the surface at different speeds and in different directions, Techniques for coping with the chaotic nature of these waves on the real sea surface are described in Chapter 2 but itis first necessary to discuss the characteristics of ideal regular waves. Such waves never occur in the real ocean environment although they can be produced in laboratory towing tanks and form the basis of many seakeeping model experiments. Of equal importance is the fact that the theory of irregular waves is based on the assumption that they can be represented by superposing or adding together a suitable assembly of regular waves. So itis clear that the characteristics of regular waves have a profound influence on the behaviour of ships in rough weather even though they are never actually encountered at sea: an understanding of their nature is one of the vital tools in the study of seakeeping. Fig 1.1 - Regular waves 12Figure 1.1 shows a train of regular waves advancing across the surface of a body of water of constant depth d. The waves are two dimensional: that is, they advance in the x direction and the crests are perpendicular to the x axis. The crests may be considered as extending to infinity on either side of the x axis; alternatively the waves may be imagined to be advancing down a long narrow tank bounded by vertical walls parallel to the x axis. The salient characteristics of the waves are: ¢ _| The instantaneous depression of the water surface below the mean level metres Cy _ | The wave amplitude or vertical distance from the mean level metres (9=0) to ncrest ora trough Cy is always positive. The wave height: twice the wave amplitude metres ; serene ras ea ¢_| ‘he wave celerity: the velocity ofan individual eeat nthe x direction metres/second T | ‘Mae wave period: the time interval between successive crests (ar troughs) passing | seconds a fixed point. &__| ‘Me instantaneous wave slope: the gradient ofthe surface profile radians ¢y _| The maximum wave slope or wave slope amplinude. Gy is always positive, radians HUA. | The wave steepness iz ‘These waves progress across the surface in a regular orderly fashion. Each wave crest advances at the same steady velocity c so that the waves never overtake each other and the wave length A and period T remain constant, ‘The shape of each wave remains the same and the whole wave train appears to advance like a rigid corrugated sheet. 1.2 THE VELOCITY POTENTIAL, We shall use the potential flow techniques of classical fluid dynamics’ to predict the characteristics and structure of regular waves. ‘The velocity potential ¢) is defined by the equations a a Bau; Bey misec : ax a cic where u and v are the horizontal and vertical velocities and ¢) must satisfy Laplace's equation AX = 0 sec a2) if itis to represent a valid fluid flow. "It is assumed that the reader has a working knowledge of classical fluid dynamics 13It is necessary to find a velocity potential ¢ which describes the fluid flow associated with a regular wave, Lamb (1932) showed that the velocity potential = £h0 cosh (Cd ~y)} msec © ‘cosh (k d) is appropriate to the case of the two dimensional regular (sine) wave of small amplitude 6, advancing across the surface of a body of fluid of any constant depth d as illustrated in Figure 1.1; & and « are constants whose physical meaning will be derived in Section 1.3. Almost all calculations of seakeeping performance assume that the water is deep compared to the wave length. If we assume that the depth of water is infinite? then cosh [k(d cosh Kd - 9) 2 op (-% cosh (kd) Pee and the velocity potential becomes = ep (ky) msec 3) 1.3 PRESSURE CONTOURS AND THE SURFACE PROFILE ‘Bernoulli's equation for the unsteady motion of an ideal fluid is 2 LM gs Pio msec? 2 a pe a4) where u? + yy? msec? (1.5) Bernoulli's equation must apply everywhere and can be used to find the surface profile associated 2 In practice d > 2 is an adequate restriction. 4with the velocity potential given by Equation (1.3). The only force applied externally to any fluid particle is gravity. Hence x-@ r-@ go misec? so that Q=gy msec? and Equation (1.4) becomes 2 F-Beay+ tno m*Isec? (6) Constant pressure contour Fig 1.2 - Constant pressure contour beneath a regular wave In calm water the pressure at depth J, is P=pgy, Nim? and a constant pressure contour is a horizontal straight line. Under regular waves this contour is distorted as shown in Figure 1.2. The depth of a point on this contour is yy, +0, m where C, is the depression of the contour below the depth 7. Since the pressure everywhere along the contour and the depth y, are both constant the quantity 15- 8), Ja msec prusiaoe oly will be a constant on the contour at any given time . Itmay be added to the potential without affecting the velocities in any way (since they are functions of the potential gradients and not of the potential itself). So we may define a new velocity potential ' eos fe ae va msec p ‘0 so that op’. ob, P 21sec? BBE ey, mice and Equation (1.6) becomes 2 ag! g , 3 a sec? ©. Wg, =0 misec Tar eo and the prime may now of course be omitted. Tf we now assume thatthe velocity is small (tantamount to assuming that the wave amplitude 2, is small compared with the wave length) we may neglect 9” so that the depression of the constant pressure surface is 2(2) m gl aly =y, +f, or, since 6, is small, ,-2(2| 7 a \ ay =y, a7 ‘Substituting the expression for the velocity potential (Equation (1.3)) yields the equation for the constant pressure contour at depth ¥p Gp = Sy exp (ky, )sin(kx-wt) m (1.8) 16Ss Oe | iy = 50m Bottom Fig 1.3 - Constant pressure contours beneath a 100 m wave: depth 100m As an illustration of this equation Figure 1.3 shows typical pressure contours beneath a regular wave of length 100 metres in water 100 metres deep. These results have been obtained by setting 1 = 0 in Equation (1.8). ‘The surface profile is one of these constant pressure contours (with the pressure equal to the atmospheric pressure). It is obtained by setting y, = 0 in Equation (1.8) to give f= Cysin(kx-ot) m (9) ‘which is the equation of a regular wave of small amplitude ¢, advancing across the fluid surface. Equation (1.9) is illustrated in Figure 1.4. Consider first the wave shape in the geographical or spatial sense. This is tantamount to fixing time at some instant r as, for example, when taking a photograph. If for simplicity we choose r = 0, Equation (1.9) is reduced to C=Gsin(kx) m which represents a simple sine wave starting at the x origin. The wave length is A and the wave number k is now seen to be m (1.10) A second photograph taken a short while later at t = f, would reveal exactly the same wave profile with the same amplitude and wave length moved along the x axis a distance (49 1, / k) 7Wave profiles at successive Time history of wave depression at x= x, io & Fig 1.4 - Regular waves pictured in space and time ‘metres, This can be seen by recasting Equation (1.9) as c=eam[e( 2-22) m and the term «f,/k can be recognised as a phase ‘lag’ which governs the location of the wave along the x axis. As time passes, the lag increases and the wave advances steadily away from the origin with velocity cs misec qu) ¢ is the wave celerity. An alternative view of events can be obtained by fixing the distance x and allowing time to pass. 18Physically this can be considered as recording the time history of the rise and fall of the water surface at some fixed point x = x,. The resulting sine wave of amplitude ¢, and period T is also illustrated in Figure 1.4 and the frequency @ is related to the period by radisec (1.12) ‘The time history is given by recasting Equation (1.9) as (-2)] w}r-— m (1.13) ° C, sin and the term (k x, / «@ seconds) may now be recognised as a phase ‘lag’ which governs the temporal location of the sine wave along the ¢ axis, 1.4 WAVE SLOPE It is sometimes convenient to quantify the severity of the waves in terms of their slope rather than their height or amplitude. The slope of the pressure contours may be obtained by differentiating Equation (1.8) with respect to x: =k Cy em (-ky, ) cos (kx- at) (1.14) Fig 1.5 - Wave depression and wave slope profiles at t = t, ‘The wave slope at the surface is obtained by setting y, = 0 to give a cos(kx- at) rad (1.15) 19where the wave slope amplitude is ty = kl, rad (16) So the wave slope varies sinusoidally in both time and space in much the same way as the surface depression. Figure 1.5 shows the surface profile and the corresponding wave slope at time 1 = t,. The wave slope is a maximum when the surface depression is zero and vice versa. ‘The time history of the wave slope at a certain location (1.15) as , is obtained by recasting Equation _ kx a= ones} a 6-— | | ra and this is compared with the corresponding surface depression time history (Equation (1.13) in Figure 1.6, The wave slope lags the surface depression by a quarter of a wave period. Fig 1.6 - Wave slope time history at x= x, 1.5 PARTICLE ORBITS According to Equation (1.1) the velocities at any point under the wave can be found by differentiating the velocity potential given by Equation (1.3). This gives w= Be -wsn(ke-or) mise ay 20ab a ve -vycos(kx-wt) misec where the velocity amplitudes are eke @ exp (-ky) mise Hee e eee Deptt: 100m Fig 1.7 - Particle orbits under a 100 m wave ——" Sine wave —— _ Trochoid wave © Particle at surface Fig. 1.8 - Orbits of particles at the water surface Since the wave amplitude is assumed to be small the velocity amplitudes given by Equation (1.19) must also be small and it follows that a particle of water oscillating about some point (2; y ) will never stray very for from that point. The path of the particle can therefore be calculated approximately by assuming that it is always subject to the velocities calculated for the point (2, y ), With this assumption the particle's trajectory is obtained by integrating Equations (1.17) and (1.18) to give 21 (1.18) (1.19)Ax - xy cos(kx- wt) m (1.20) ay " yp sin(kx-@t) m (1.21) where Ax and Ay are the deviations of the particle from its datum position ( x, y ) and the amplitudes of its displacements are %y =o, ep (-ky) m (1.22) Figure 1.7 illustrates these formulae and we see that the water particles follow circular orbits and that the amplitude decays very rapidly with depth. Figure 1.8 shows how the circular orbit of a particle at the surface results in the wave profile’. Pressure amplitude ~ kN/m? oO Os 1.0 ° 0.25 0.50 yd 0.75 1.00 Fig 1.9 - Pressure amplitude under a wave: wave length 100m; wave amplitude 1.0m > Figure 1.8 shows that the wave profile generated by particles following circular orbits at the surface is a trochoid with sharper crests and flatter troughs compared to a sine wave. However, the trochoid wave approaches a sine wave as the amplitude becomes infinitesimal and the assumption of sinusoidal waves is adequate for the linear treatment used in this book. 21.6 PRESSURE FLUCTUATIONS UNDER A WAVE, The pressure at any point under a regular wave may be found from Bernoulli's equation (1.6). If we assume that the velocity is small we obtain P=pgy+P kNim* so the pressure at any depth y oscillates around the steady hydrostatic pressure P & ). ‘The fluctuating part of the pressure is P=-pgl,em(-ky) sin(kx- wt) kNim® (1.23) 0 Figure 1.9 shows the variation of the pressure amplitude p g C, exp ( -k y ) beneath a 100 metre long wave. The pressure amplitude decreases with depth and becomes negligible for depths greater than about half the water depth. 1.7 ENERGY IN A REGULAR WAVE ‘The energy associated with a train of regular waves includes contributions from both potential and Kinetic energy. Consider a small element of length 6x metres and width A metres (perpendicular to the page) of the regular wave shown in Figure. 1.10. ‘The surface depression © is given by Equation (1.9) and the mass of water over the clement is approximately -p ¢ 6x A Fig 1.10 - Potential energy in a regular wave tonnes. The centre of gravity of this mass is approximately ~(/2 metres above the undisturbed surface level and its potential energy relative to the undisturbed (calm water) state is (pg @ & A) / 2 Kd. fwenow allow 6x to become infinitesimally small we may integrate to obtain the total potential energy in a single wave length: 232 ee = tf fon? Cex ory de a2 2 ~PeAGA up 4 Fig 1.11 - Kinetic energy ina regular wave Consider now a small element of fluid of width A metres beneath a wave as shown in Figure 1.11. The mass of the element is p 6x 6y A tonnes and it has a total velocity q given by Equation (1.5). So the kinetic energy of the elementis ( p q? 6x by A) / 2 KI. If we now allow &x and éy to become infinitesimally small we may integrate to obtain the total kinetic energy of the fluid in one wave length between the surface and the bottom: ho eal fatacay AD ao Substituting Equations (1.17) - (1.19) in Equation (1.5) gives 2 2 -2k gt RCFE) antiae? (1.25) oF and the kinetic energy in one wave length is found to be(4.26) So the potential and kinetic energies are equal and the total energy in one wave length is perGa E~E,+Ey= 5 kT (1.27) which leads to the remarkable result that the average energy per square metre of sea surface is independent of the wave frequency and depends only on the wave amplitude: est 2 B= kim? (1.28) 1.8 ENERGY TRANSMISSION AND GROUP VELOCITY ‘The energy associated with a sequence of regular waves is transmitted in the direction of their propagation. ‘The rate of energy transmission can be found by considering the energy flux across the plane AA in Figure 1.12. We begin by calculating the rate at which the fluid on the left of a small element of height dy and width A is doing work on the fluid on the right of the element, Fig 1.12 - Energy transmission in a regular wave Since the element is small the pressure and velocity acting on its face may be regarded as constant (at a given time) and the force exerted by the fluid on the left is P 8y A KN. The work done by the fluid on the leftis u P Sy A ki /sec. If we now allow 6y to become infinitesimal the total rate of transmission of energy across the plane AA is obtained by integrating over the depth of the fluid, 25Neglecting the small contribution due to the portion of fluid above the undisturbed surface level (y = 0 ), the rate of transmission of energy is 4 E=A fuPdy ksec (1.29) ° Using Equations (1.17) and (1.23) we find that for very deep water the energy is transmitted at arate E eae iee ee tees) ye The rate of transmission of energy evidently fluctuates with time but we are concerned with its ‘mean value, Over a long period of time (or an integral number of wave periods) the mean value of sin? (kx - wr) is 0.5. So the mean rate of energy transmission is pecGa (1.30) E kllsec Now the total energy is given by Equation (1.27) and this energy is transmitted at a mean velocity given by . misec a3b ‘We may interpret this result by considering the progression of a group of regular waves down a laboratory tank. If the mean energy associated with each wave length is £ KJ per square metre the amplitude of the waves is, from Equation (1.28), eee (4.32) Ps Each individual wave within the group is propagating forward at the celerity c (Equation (1.11)) but the energy is only propagating at c/2 m/sec (Equation (1.31)). So after one wave period each ‘wave will have moved forward one wave length, taking half of its associated energy with it. The other half of the energy is left behind to be added to the energy brought forward by the next wave. In this way the total energy per square metre within the group is kept constant. ‘At the leading edge of the group the first wave will be propagating into calm water. So the orderly exchange of energy from wave to wave does not happen and after one wave period the 26TABLE 1.1 REGULAR WAVE FORMULAE FOR DEEP WATER 4. a> ¢ > ® T k a € Ug ° 7-28 o r pte cS gee ar? an Can 2n as =| =,|& F Vek ONE | ON Ge c=2u, energy of the leading wave is halved. The wave amplitude is reduced and this process continues as the leading edge of the wave train propagates down the tank at the wave celerity. ‘The leading edge of the group proper (defined as the position of the first wave of full amplitude given by Equation (1.32)) propagates down the tank at velocity u, and this is called the group velocity. Individual waves within the group propagate at the wave celerity c, which is twice the group velocity (see Equation (1.31). 27‘TABLE 1.1 (continued) REGULAR WAVE FORMULAE FOR DEEP WATER, a. (d > 5) @>> ad My Yo Fo Yo o T & Fy=Vo=Fo exP(“) i 850, oy e on 1.9 | SUMMARY OF THE CHARACTERISTICS OF REGULAR WAVES ‘The equations derived above may be combined in various ways to produce a multitude of different formulae which are listed for easy reference in Table 1.1. Figure 1.13 shows how wave frequency and celerity depend on wave length. As might have been expected long waves have very low frequencies and vice versa. As if to compensate, the celerity increases with wave length. For example the celerity of a 1000 metre wave is almost 40 metres/second (over 75 knots) compared with only about 4 metres/second (about 8 knots) for a 10 metre wave. This dependency 28of celerity on wave length distinguishes surface waves from most other types of wave motion (notably electromagnetic radiation) and we shall see that it is responsible for some peculiar effects when the waves are encountered by a moving ship. Wave frequency w (radians/second) Celerity ¢ (metres/second) 12 4 10 20 40 100200 1000 Wave length }. (metres) Fig 1.13 - Wave frequency, length and celerity in deep water 1.10 ENCOUNTER FREQUENCY AND HEADING Although the wave frequency «@ has some influence on ship motions in regular waves, the motions are more closely dependent on the frequency with which the moving ship encounters the waves. Before deriving expressions for the encounter frequency we must adopt the convention for defining the ship's heading 1 shown in Figure 1.14 and Table 1.2. The ship is assumed to be attempting to maintain a straight line track at a constant speed U (metres/second) across the sea surface. The waves will cause deviations from the intended course and track, but a directionally stable ship in the hands of an experienced helmsman (or an autopilot) will usually be able to follow a sensibly straight mean course so that the heading angle jt can be readily defined as the angle between the intended track of the ship and the direction of wave propagation. Since the wave crests are 1 metres apart, a crest will meet the ship every T, seconds, where the encounter period is given by A Pye sea c -Ucosp ae (1.33) ‘The component velocity of the ship in the direction of wave propagation is U cos 4. and the 29Following pea? <— Port beam Starboard beam n= 270° m=90° Ucosp Fig 1.14 - Heading definitions waves will overtake the ship with a relative velocity ¢ - U cos y metres/second. The corresponding encounter frequency «, is defined as Ce - Ucosp) radisee 34) or eu (135) and this is illustrated in Figure 1.15. In head and bow waves (90° < # <270*) the ship is heading into the waves and the encounter frequency is always greater than the wave frequency. In beam waves (jt = 90° or 270°) the encounter frequency is equal to the wave frequency and is unaffected by ship speed. ‘On headings abaft the beam (0° o Encounter frequency «, ° Following and quartering waves 0,<0 Fig 1.15 - Encounter frequency and heading In following and quartering waves a given (absolute) value of encounter frequency may be experienced in three different wave systems. Two of these wave systems will give positive encounter frequencies and the third will give a negative encounter frequency. The corresponding wave frequencies can be obtained by rearranging Equation (1.34) to give Fo, w= 8 J14]1-—vcos pn] radisec (1.36) 20 cos ¢ ‘The physical interpretation of this phenomenon is best illustrated with a numerical example. Consider a ship steaming at 20 knots ( /= 10.3 m/sec) in regular following waves. Suppose that the ship encounters the wave system at a frequency |«,| = 0.2 rad/sec. It is required to find the wave systems which could be responsible. Possible results, obtained from Equation (1.36), are given in Table 1.3 together with corresponding celerities and wave lengths. 31Heading 11 (degrees) TABLE 1.2 HEADING AND ENCOUNTER FREQUENCY ‘Wave direction Encounter frequency Following waves. The waves travel in the same direction as the ship. 0, <0. o, can be negative. ° Long waves with high clerity The same 1, can occur overtake the ship. Te ship overtakes | in three diferent wave Tongths. short waves with low eelerity. , <0. 0-90 Quareeiag waves encountering the |, can be negative ship on the starboard side. The same «, can occur in three different wave lengths. . a-0 Beam waves encounteriti, ti caesceronds ila ieee bea the ship on the starboard side mien Bow waves encountering the stip | ys on the starboard side. 4 ; ee co, corresponds to a single wave length Head waves. The waves travel in he |r > © end waves. The waves travel in a «2, coresponils to a single 180 opposite direction to the ship aa Bow waves encountering the ship on | © i «o, corresponds to a single “2 port side. 2 180 - 270 the port si oan Beam waves encountering the ship | + 7 0 : z «9, comesponds to a single 270 on the por side ate a ) 80, +2 radlseo; (n-2, ND mi Finally the associated sine wave amplitudes are calculated using a version of Equation (2.13a)*: bo =f 2 5k@,) 6@, m (2.130) where 5c, is the frequency interval associated with the nth frequency «,. 2.6 SPECTRAL MOMENTS The definition of variance given in Equation (2.2) can be written as (2.14) if the time history has @ zero mean and the number of observations is very large. A time history represented by Equation (2.8) therefore has a variance my = (3b cos (0, ¢ + 6,)) ah m? (2.15) a nit 0 Since the frequencies are chosen in accordance with Equation (2.5) this reduces to 4 If the spectrum is defined at equally spaced frequencies it will be necessary to interpolate the spectral ordinates at the randomly spaced frequencies. ALMe: my 3 Cy om (2.16) and, from Equation (2.13a), 7 oo m, = DS{o) bo = f $0) dom? (2.17) mi i ‘So the variance of the irregular wave time history is equal to the area under the wave energy Spectrum, ‘The time history given by Equation (2.8) can be differentiated to obtain the vertical velocity and acceleration of the sea surface: C0) = D- Cy &, sin (@, f+ ©) misec (2.18) = CQ = V- bq wr cos (0, t+ €,) — misec? 2.19) a ‘These can be regarded as irregular wave time histories in their own right and can be analysed to obtain statistics of velocity and acceleration in exactly the same way as for the surface depression. The amplitudes of the component sine waves for velocity and acceleration are now (yo ®,) and (C9 Wy ) respectively. ‘These velocity and acceleration time histories can be analysed to produce corresponding velocity and acceleration spectra. By analogy with Equation (2.12) the spectral ordinates are respectively 2 o; Sy, ) = 7 i = @, S,(@,) msec per radlsec (2.20) of Sx, ) = 7 _ = 0) S,(,) m%sec* per radisec (2.21) So the velocity and acceleration spectral ordinates can be obtained by multiplying the displacement spectral ordinates by the second and fourth powers of the frequency. By analogy with Equation (2.17) it is clear that the area under the velocity and acceleration spectra must be equivalent to the variances of velocity and acceleration respectively. ‘The ASvariance of velocity is Mm, = Hf Sew) do = f @? Sa) do msec? (2.22) 0 ° and the variance of acceleration is = f Sho) do = f of $0) doo m7Isec* (2.23) > ° m, and m, are called spectral moments since they can be considered as moments of area of the energy spectrum about the vertical axis. In general = f wo" Sw) do mIsec™ (2.24) ° and n may take any positive integer value (n = 0, 1, 2, 2.7 MEAN PERIODS ‘The mean frequency can be found by taking moments about the spectral ordinate axis and determining the centre of area of the spectrum from st my @ = 7) radisec (2.25) My and the corresponding mean period is -_ 2m, Te a acti (2.26) (The spectral moment m, in these formulae may be obtained by setting n = 1in Equation (2.24). Ochi and Bolton (1973) showed that the mean period of the peaks is m, =2n |— sec (2.27) m,(aq) Narrow band Lo time history 1 (seconds) ¢ (metres) x 4 "3 t (seconds) ¢ (metres) Fig 2.6 - (a) Narrow band and (b) wide band time histories and the mean zero crossing period is Fo-2n | sec (2.28) Strictly speaking Equations (2.27) and (2.28) are valid only if the surface depression measured at equal intervals of time is normally distributed (see Chapter 11). In practice this assumption is invariably true for real ocean waves. 2.8 SPECTRUM BANDWIDTH Figure 2.6 shows two imegular wave time histories, and sketches of the corresponding wave energy spectra are shown in Figure. 2.7. The narrow band time history of Figure 2.6(a) could loosely be described as a sine wave of varying amplitude, and the origin of the terminology is clear from the appearance of the spectrum: the wave energy is concentrated in a narrow band of frequencies and little or no energy is present at other frequencies. A peak is nearly always followed in orderly succession by a downward zero crossing, a trough, an upward zero crossing aTand another peak, Peaks below the datum level are rare and it follows that the average period of the peaks is almost the same as the average zero crossing period. ‘The wide band time history contains energy over a wider band of frequencies as shown in Figure 2.7, In this case there are many peaks and troughs which are not immediately followed by zero crossings and the average period of the peaks is much less than the average zero crossing period. There are many peaks below the datum level and many troughs above the datum level. i 2] sr i & g Wide g 7 3 {. 7 ‘Frequency @(rad/sec) Fig 2.7 - Narrow and wide band spectra ‘The ratio between the average period of the peaks and the average zero crossing period can be regarded as a measure of the “narrow bandedness” of the time history and its wave energy spectrum. The bandwidth parameter is defined as (2.29) and values of € lie in the range Oto 1: € = 0 corresponds to a very narrow banded spectrum. with T sec and ¢ = 1 corresponds to a very wide banded spectrum with agIt is customary to assume that the spectra associated with waves and ship motions are narrow banded with e = 0 Cartwright and Longuet-Higgins (1956) showed that the significant wave height can be related to the area under the wave energy spectrum by the equation Fy = 4.00 fig m 2.30) if € = 0 this reduces to Fig = 4.00 fmm, = 4.00 i Sf) dom 3h ° (see Equation (2.17)) So the significant wave height can be estimated by integrating the wave energy spectrum. 2.9 IDEALISED WAVE SPECTRA 2.9.1 Introduction In general the wave energy spectrum derived from an analysis of an irregular wave record obtained at a particular place and time in the ocean will be a unique result that will never be exactly repeated. Although it may be a useful guide to likely wave conditions, its use for ship design purposes is strictly limited and it is customary to rely instead on families of idealised wave spectra. Current practice is to use different formulae for open ocean and coastal (limited fetch) conditions. 2.9.2 Open ocean conditions The International Towing Tank Conference (ITTC) has adopted the Bretschneider spectrum (1952, 1957) as the standard wave energy spectrum to represent the conditions which occur in the open ocean. It is often called the ITTC two-parameter spectrum to distinguish it from an earlier standard spectrum. It is given by -A B 2 Safe) = 2 ep (3) mI(radlsec) (2.32) where A A= 172.75 8 msec* (2.33) 7 Agand B=) sec (2.34) ‘The “two parameters” are the significant wave height /,,, and the average period 7 Equation (2.26)). ‘The area under the Bretschneider spectrum is, (2.35) m2 (2.36) mIsec? and (2.37) where T’ is the gamma function. ‘The mean zero crossing period (Equation (2.28)) is faaestes = 0927 sec (2.38) (2.39) The modal period 7, of the spectrum corresponds to the frequency «, of the peak, which may be obtained by differentiating Equation (2.32) and setting the result to zero. It is found that, 49 radisee (2.40) see (2.41) (see Equation (2.38)). Equations (2.38) and (2.41) may now be used to define the constants A and B more fully: msec (2.42) (2.43) It should be emphasised that the relationships between the periods (Equation (2.41)) are not general and apply only to the special case of the Bretschneider spectrum. Figure 2.8 shows some specimen Bretschneider wave energy spectra for a significant wave height of 4 metres and various modal periods. As expected from Equation (2.35), the area under each spectrum is the same since the significant wave height is the same in each case. The position and height of the peaks depend on the modal period. eySpectral ordinate S oo (metres */ (rad/sec)) 0 0 05 10 15 20 Frequency @ (radians/second) Fig 2.8 - Bretschneider wave energy spectra: Significant wave height 4.0 metres 2.9.3 Coastal waters In coastal waters with limited fetch the Joint North Sea Wave Project spectrum is used ' (Ewing(1975)). In the form accepted by the ITTC the spectral ordinates are defined by Syp (@) = 0.658 C Spe (@) —mU(radlsec) 44) where Spe (o) is the Bretschneider wave spectral density ordinate (Equation (2.32)). The factor C is given by Cc = 3.37 where where ‘Commonly known as the JONSWAP spectrum 52Figure 2.9 shows a comparison between the two spectra for a significant wave height of 4 metres and modal period of 10 seconds. 2.10 WAVE SLOPE SPECTRA An irregular record of wave slope can be analysed to produce an energy spectrum in the same way as an irregular record of wave depression. There are now an infinite number of component sine waves of the form given in Equation (1.15) and the wave number of the nth sine wave component is > |S, 3 Frequency @ (radians/second) Spectral ordinate 8, (0) (metres! frad/sec)) Fig 2.9 - JONSWAP and Bretschneider wave energy spectra: significant wave height 4.0 metres; ‘modal period 10 seconds. ‘The wave slope amplitude of the nth sine wave is given by Equation (1.16): 2 a i (2.45) By = ky Soo 8 and. the wave slope spectral ordinates are given by _& Gel 2 60 2 ot oe (2.46) = a don) rad*I(rad/sec) (see Equation (2.12) 53JONSWAP 004 003 2 8 Wave slope spectral ordinate Sy (a) (rad)'(radfsee)) ° 0 1 2 3 4 Frequency o (radians/second) Fig 2.10 - Wave slope spectra: significant wave height 4.0 metres; modal period 10 seconds So the wave slope energy spectrum can be obtained by multiplying the wave amplitude spectral ordinates by 5, 1g, ‘The Bretschneider and JONSWAP wave slope spectra are then Sy, (0) = AL exp (2) rad?I(radlsec) an og? oF and Sjq_ (0) = 0.658 C S,, (w) rad*/(rad/sec) (2.48) Figure 2.10 shows examples of these wave slope spectra. The JONSWAP spectrum is very sharply peaked but the most striking comparison with the shapes of the corresponding wave amplitude spectra (Figure 2.9) is the much greater comparative importance of high wave frequencies. This corresponds with practical observations: short high frequency waves are often very steep even though their amplitudes are very small. All the relationships derived for wave amplitude spectra (Equations (2.12)-(2.30)) have 54analogous relationships for wave slope spectra. ‘Thus, for example, the variance of wave slope can be obtained by integrating the wave slope spectrum. 2.11 WAVE SPREADING AND SHORT CRESTED WAVES In ideal conditions in the open ocean all the waves might be expected to travel in the same direction. However, these long crested waves in which the wave crests remain straight and parallel are never experienced outside the artificial confines of the laboratory towing tank, although approximations may occasionally be found at sea. It is much more likely that the real waves in the ocean will be travelling in many different directions, although an easily recognised “primary” direction, often more or less aligned with the local wind, may be discernible. Changes in wind direction, the influence of coastlines and bottom topography and the presence of wave systems originally generated elsewhere will all conspire to ensure that the true long crested wave system is at least a rarity and probably a myth. Limit of spreading Secondary wave Giseotion Ship's course Fig 2.11 - Primary and secondary wave directions. ‘The presence of more than one long crested wave system results in alternate enhancement and cancellation of wave crests and troughs, and this phenomenon gives rise to the term short crested to describe the appearance of a wave system with a spread of wave directions. ‘The wave energy spectrum derived from a record of surface elevations obtained at a particular point in the ocean will invariably contain contributions from several different wave directions. Itis often convenient to ignore this fact and assume that the wave system is long crested, and for many purposes this may give acceptable results. However, the degree of wave spreading does have a profound influence on some ship motions (particularly roll) and its effect cannot always be ignored (see Chapter 8). 55PgS¢(@, v) ba bv= Energy in frequency band de and direction band 6v 01365;(0) 0.167 8,(0) 0.1255, (0) 0.1565, (@) 0.083 S;(@) 0.123 5, (0) 0.042 5, (a) ag so 0083 5, (0) 0.011 5, (0) Ee sor 0.082 5, (0) 25 01 S,(0) Primary wave direction Fig 2.13 - Representation of a directional wave spectrum by thirteen long crested spectra at discrete heading intervals of 15°: cosine squared spreading over + 90°. An infinite number of possibilities exist, but for design purposes it is usual to assume that if the primary wave direction is j1 relative to some fixed datum (usually the ship’s course: see Figure 2.11), the secondary wave directions v are distributed in the range - W/2<(v - 1) < m2. ‘The directional wave spectrum is defined such that the quantity p g S,(w,v) 6w dv is equivalent to the wave energy contained in the frequency band 5 and the direction band Sv as shown in Figure 2.12. Hence the directional spectral density ordinate, by analogy with Equation 56(2.12), is given by 50.) m?/(rad/sec) per rad (2.49) where C9 is now the amplitude of the component sine wave appropriate to the nth frequency and the jth direction. For ship design purposes it is assumed that the directional wave spectral ordinates are related to the ordinates of the equivalent total wave energy spectrum 5,(@) by Seleo,v) = 2 cos Vv - p) S(@) m*(rad/sec) per rad (2.50) ‘Trials evidence (Cummins and Bales (1980)) suggests that this cosine squared spreading is appropriate for typically occurring conditions in the open ocean. TABLE 2.1 WEIGHTING FACTORS FOR CALCULATIONS OF SHIP MOTIONS IN SHORT CRESTED WAVES. Cosine squared spreading over 90° by = 13? vou w = 90? 0.000 £75 0.011 + 60? 0.042 #459 0.083 230° 0.125 £ 15° 0.136 0 0.167 Equation (2.50) is of little direct use in practical computations of ship motions in short crested seas, These calculations (see Chapter 8) require the spread wave spectrum to be represented by a discrete (long crested) contribution from each of a finite number of secondary wave directions within the range of the spreading. Each contribution is essentially a scaled down version of the total wave energy spectrum as shown in Figure 2.13. If the secondary wave directions are spaced at intervals of 5v the appropriate wave energy spectrum at each secondary direction is given by W So) mi(radlsec) where the weighting factor W is 57W= 2 cos? (v-) ov 2.51) Table 2.1 lists the weighting factors for intervals of 6v = 15°. 2.12 OCEAN WAVE STATISTICS: 2.12.1 Introduction We have seen how an idealised wave energy spectrum may be defined in terms of the significant wave height and various measures of the average wave period. This allows representative spectra to be constructed for any point in the ocean provided these quantities are known. Of course many 55 Total 16-20 + fede vas | zi 124 +f le wiz | + | + +]+[4 [+ 9-10 +f+]+]+ [+] so f+ f+ | + f+ f+ ff 2 ve f-+ | + | + fa fa | 2 ds | ss [-+ [1 [1 [os [7 [uw [a as [i [2 [4 [2 [6 [0 [is [a [ 80 a4 [2 [4 | | 28 [37 | [2 | + 15, 23 | s [12 [2% [65 | | wo [+ | + [ 187 12 [7 | a | @ [os [iw | + [+ 259 or [a flow fo ful. Ts 225 + indicates less than one part per thousand. Lee, Bales and Sowby (1985) have published a similar atlas for the Pacific Ocean. ‘The hindcast technique avoids the problems of accuracy and fair weather bias associated with visually observed wave data but depends, of course, on the reliability of the mathematical model used to predict the wave conditions. 65If both visually estimated and hindcast wave data are subject to uncertainties, direct measurements of wave spectra must provide the most reliable data of all. However, measuring ‘wave data over a protracted period (Years) is an expensive and complicated undertaking and few attempts at systematic data collection have been made. Probably the most comprehensive is that ‘organised by the US National Oceanographic and Atmospheric Administration (NOAA) and published by Gilhousen et al (1983). The measurement buoys were deployed for at least three ‘years and some were in continuous operation for as long as nine years at various locations around the United States. They recorded information on air and sea temperature and atmospheric pressure as well as wind and waves. The waves were sampled every three hours and a wave spectrum derived from the recorded time history. The significant wave height and mean zero crossing period were derived using Equations 2.31 and 2.28 and scatter diagrams similar to those shown in Table 2.3 prepared. Many other measurements of wave conditions have been made for specific purposes at various locations throughout the world. Typically these are short term studies intended to provide data on the local environment for use in research or specific projects such as ship seakeeping trials, or the design of offshore or harbour installations. Much of the data have been acquired by commercial organisations who regard them as proprietary information not available to the general public. However, in 1982 the United Kingdom Marine Information and Advisory Service (MIAS) published a catalogue listing the data sources open to general use. Over 1350 entries were catalogued. The majority of the measurements were made in the coastal waters around the British Isles and in the North Sea, but a significant quantity of data are also available for North ‘American and Australian waters. Fig 2.19 - Sea areas used by Bales, Lee and Voelker (1981) 662.12.4 Annual growth of wave heights Bacon and Carter (1991) and Barrat (1991) showed that the wave heights measured and observed ‘on ocean weather ships in the North Atlantic have been increasing for at least the last thirty years, ‘Hogben (1994) summarised the data and showed that the mean significant wave height appears to be growing at a rate of about 1.5% per year, as shown in Figure 2.20, This implies thet the wave heights derived from wave atlas data should be increased by a factor F to take account of the time lapse between data collection and the time when the ship is expected to be in service: (2.58) where A is the elapsed time in years between the time the wave data were collected and the time the ship is to enter service, ows: on OWS Indian OWS Lima * OWS Jute ‘Mean significant wave height (metres) 1950 1960 1970 1980 1990-2000 Year Fig 2.20 - Increase of mean significant wave heights in the North Atlantic (Afier Bacon and Carter (1991) and Barrat (1991)) 67LINEARISED EQUATIONS FOR SMALL AMPLITUDE SHIP MOTIONS IN. REGULAR WAVES 3.1. INTRODUCTION Ships do not, in the normal course of events, experience regular waves at sea. So the study of ship motions in regular waves appears at first sight to be an academic exercise of no practical significance. Yet it is an essential first step in the calculation of ship motions in a realistic irregular seaway; moreover, an appreciation of regular wave motions will give the reader an insight which will prove invaluable in understanding the general nature of the motions of ships in rough weather. 3.2. AXES AND SHIP MOTION DEFINITIONS A ship in rough weather experiences a complex sequence of motions as it twists and turns its way across the ocean surface. The motions seem to defy any rational analysis, particularly by those who suffer from their effects on board the ship. Nevertheless, it is possible to make some observations on the characteristics of ship motions which will help to clarify their nature and will form a basis for the modem theory of seakeeping. Let us suppose that the ship is attempting to maintain a straight course at a constant speed U metres/second. Waves continually cause the ship to deviate from its course and track and the helmsman may find it necessary to take corrective action. In addition the ship will rise and fall in response to the changing water level and the deck will seldom be truly horizontal. The ship will generally follow some kind of spiral path which is more or less aligned with the intended course. Finally the ship's speed will be continually varying around the nominal speed U as the ship surges along its track in response to the waves.” Any particular ship's track and motion time history can be represented by a combination of the time histories of three linear? and three angular displacements. ‘These six displacements are defined using the right handed axis system shown in Figure 3.1. ‘The ship speed will generally be about the same as it would be at the same engine power setting in calm water: in head waves it will be reduced and in following waves it will be increased. 2 Linear” here means a displacement along an axis as opposed to a rotation about an axis. For the time being there is no necessary implication that the motions are linear in the sense that they are directly proportional to a force or moment. 68‘The axis system Exy has its origin fixed at £ at the mean water level and regular waves propagate along the Ex axis. A second axis system E x,, Xz) Xzq also has its origin at E but is rotated through the heading angle 1 so that Ex,, coincides with the mean track of the ship Body axis system 4s Jee Earth axis systems Fig 3.1 - Axes and ship motion definitions A point O, lying at the mean water level, moves along Ex,, at the mean speed of the ship, U metres/second. The mean position of the ship's centre of gravity G, lies vertically above O and is taken as the origin of a third axis system G, x, x, x. At any instant of time the position of the ship's centre of gravity G relative to the moving origin G, is defined by three linear displacements: Motion | Notation | Units _ | Positive surge % | metres | forward sway % | metres_| to starboard heave 4 metres | down The attitude of the ship is defined by three angular rotations about the axes G, x, Gy x, and G, x, 3 See Chapter 1 for the definition of heading angle. 6Units Positive radians _| starboard side down radians | bow up radians _| bow to starboard ‘Most ships have port/starboard symmetry and so surge, heave and pitch, which lie in the plane of symmetry, are called vertical plane or symmetric motions. Sway, roll and yaw are termed lateral plane ot antisymmetric motions. ‘The motions are often referred to as “degrees of freedom”. Another righthanded set of axes G Xp, Zp) pq is fixed in the ship and is used to define locations on (or in) the ship's structure. The origin of this body axis system is at the (moving) centre of gravity G and the axes rotate as the ship rolls, pitches and yaws. Locations are defined as: Location Notation | Units Positive Longitudinal Xa metres | forward Lateral metres | to starboard Vertical %s3__| metres _| down ‘The wave depression at any point x is, according to Equation (1.9), C=Gsin(kx- ot) m where the time 7 is measured from an arbitrary datum. Transforming to the axis system aligned with the ship's track we find that the wave depression at any point ( xp) , Xp, ) is C=C, sin (k xp, coop -kxp sinpp-@t) m Hf we choose a datum time such that the moving origin O is at £ at time x r= ___ sec kU cost the moving and Earth fixed frames of reference are related by 70‘Then the wave depression at ( x, , x ) in the moving frame of reference is Sin (@,t- kx, cosu + kx sin) m en 3.3. GENERAL EQUATIONS FOR SHIP MOTIONS IN REGULAR WAVES 3.3.1 Mass and moments of inertia ‘The ship may be regarded as being composed of a large number of very small masses 5m tonnes. Figure 3.2 shows one of these masses located at (xp) . Xpp py ) (relative to the centre of gravity of the ship). If the ship has linear accelerations ,, %, and #, m/sec and angular accelerations ¥,, £, and +, rad/sec® the mass 6m will have linear accelerations See eee 2 py Xy - Xp ¥y mlsec? forwards Hy — Xpy Hy + Spy mlsec? to starboard see Har eeHE CE 2 = 5, + Xp) i, - Xp %, mlsec? downwards From Newton's second law of motion, the forces and moments necessary to sustain these accelerations are OF, = 6m x/ kN surge force forwards OF, ", = 6m % kN sway force to starboard 8F, = 6m % KN heave force downwards BF = 6m Xy_ 8 - Sm xy, kN m roll moment to starboard BF, = 6m xy 4) - 5m xy, %, KN m pitch moment bow up. oF, 6m Xp, ¥) - bm Xp Xj, kN m yaw moment to starboard nFig 3.2 - Accelerations experienced by the mass ém ‘The forces and moments required to sustain the linear and angular accelerations of the whole ship are obtained by allowing 6m to approach zero and integrating over the volume of the ship. Bearing in mind that by definition of the centre of gravity [ar dm = f xp dm = fgg dim = 0 we obtain =I, -4, lasts = F, ts nRwhere F,, F, and F,are the surge, sway and heave forces and F, , F, and F, are the roll, pitch and yaw moments required to sustain the accelerations of the ship. m is the total mass in tonnes and [y, , Js, and Jg, are the moments of inertia of the ship defined by Iy =f xe + yam tm? about the xp, axis Tes J cx + xo am tm? about the xp) axis Teg = fics + Xm) dm tm? about the x,, axis ‘in passing we note that it is often more convenient for practical engineering purposes to calculate the moments of inertia using the equations Ty =m ky tm? G20) Igg = mike tm? G20) Teg = mk tm? G20) where the radii of gyration for conventional ships may be estimated using the approximate equations: 038 m 0.225L m 0.225L m where B is the maximum waterline beam and L is the waterline length. The product moments of inertia are defined by BB2 se = [ a1 Ym am tm - 2 = [tut dm tm 2 = fxm %ys dm tm For conventional ships the product moments of inertia are usually small and are invariably neglected. ‘The equations of motion then reduce to mi, Fo kW, | (i=1,3) (G.3a) F, Nm; (i=4,6) G.3b) 3.3.2 Motions in regular waves ‘The forces and moments in Equations (3.3) may be applied by any external means, but we are here concerned with the forces and moments applied to the ship by a train of regular waves. For a given hull shape at a particular speed and heading in waves of a particular length the forces and moments F, are assumed to be functions of the displacement, velocity and acceleration of the surface depression and the six possible motions. So we may write FF [GGG aes Ci=43)] (B.4a) FeF [GGG (ma; (=46)] vm G.4b) If the wave amplitude is small compared with the wave and ship lengths the motions will also be small and we may use a Taylor series expansion to obtain linear approximations to Equations G4): Fra b+b +e, 6 + (ray £, - by, - cg 5) RNG (i=1,3) @.5a) fl Fe Vm (i= 4,6) B50) ‘The coefficients a,, b,and c, are functions of the wave length and amplitude, heading angle, ship speed and hull form, and quantify the effect of the waves on the forces and moments. The other coefficients a, , by and , quantify the forces and moments required to sustain the motions of "4the ship: these will be considered in detail in Section 3.4. Substituting Equations (3.5) into the six Equations (3.3) we obtain six general linearised equations for smal] amplitude motions in regular waves: Fi Ai A) = Be kN; (i=1,3) (3.62) D(A; % + by 4 + ey =F, Nm (i= 4,6) G.6b) where Ay = 4, y (ie 16; 7 =L6:j*i) and the virtual masses and moments of inertia for each degree of freedom are Ay=m+ay (j= 1,3; 0=1,33 f=t) 7a) Ay sly + ay (i463 i246; f=4) on) The exciting forces and moments due to the waves are Fy = a FBS oo RNG (i= 1,3) (3.82) Fy =a,0+bf+0,0 kNm;(i= 4,6) (3.8b) The regular wave ship motions are related to the nominal wave depression which would have been experienced at the moving origin O in the absence of the hull (in practice will usually bbe within the hull so that the actual surface depression with the hull present cannot be defined). Setting x, = x, = 0 in Equation (3.1) gives for this nominal surface depression at O C= Gsin(o,t) m 69 and the nominal vertical velocity and acceleration of the sea surface perceived by an observer on the ship at O are f= 0, cos(@,t) msec (3.10) 15C= -02 G sin (@, 1) mfsec? ‘Substituting Equations (3.9)-(3.11) in Equations (3.8) we obtain Fut Fy sin (t+ y) ENG (4 = 1,3) Fy, = F, sin(@,t+y,) kNm;(i= 4,6) where the excitation amplitudes are Foun = GY 67 OF +b, 0,% ENG (i= 1,3) Fog = GV Ce - 4 Oe + (b, 0, Nm; (i= 4,6) and the phases are given by The generalised equations of motion (3.6) may now be written as 6 Pe Avs 2 BA #65) = Fr sit C0, + 1, ING (i= 1,3) 6 DY (Ay ¥ + by + 045) ) = Fyg sin (0, ty, ) RN mj im (i= 4,6) Solutions to these equations have the form X= %y_sin(@,t+6,) m; (i=1,3) = %psin(@,t+8,) rad; (i=4,6) Gu) (B.12a) (3.12b) (3.13a) (3.13b) G14) (@.15a) G.15b) G.16a) G.16b) So small amplitude regular sine waves impose sinusoidal exciting forces and moments on the ship and these result in sinusoidal motion time histories. 16‘The motion amplitudes are directly proportional to the excitation amplitudes, which are in turn proportional to the wave amplitude, The phase angles , and 6, relate the excitation and motion time histories to the time history of the wave depression at O as shown in Figure 3.3. The peak excitation occurs, /«a, seconds before the maximum wave depression. Similarly the peak (Positive) motion occurs 6, /a, seconds before the maximum wave depression. Fig 3.3 - Time histories of wave depression, exciting force and motion in regular waves 3.4 COEFFICIENTS IN THE EQUATIONS OF MOTION 3.4.1 Introduction ‘An understanding of the physical significance of the coefficients in the equations of motion can ‘be gained by considering an experiment in which a model of the ship is forced to oscillate in a single degree of freedom (say heave) while being towed at constant speed in calm water. The ‘model is constrained so that all other motions (pitch, roll, yaw etc) are suppressed. The forces and ‘moments normally applied by the waves are replaced by externally applied forces and moments which are just sufficient to sustain the sinusoidal motion of the model in the chosen degree of freedom. We need not concern ourselves with how these forces and moments are generated, measured or applied, only with their magnitudes and effects. In general (for an arbitrary hull shape) it will be necessary to apply forces and moments in all six degrees of freedom i to sustain the single motion x,. The required forces and moments can be found from six equations analogous to the equatioris of motion (3.15): =1, 3) G.17a) s Pee be heya) = Rosin (0,1) A 76 BN H+ by ¥, + cy x)) = Fy sin (@, t) RN m; (4, 6) G.17b) Where Fg are the amplitudes of the applied forces and moments. All the motions except for x, are set t0 zero and the single motion +, leads the forces and moments by e, radians: 4 sin (@, +6) mor rad G.18) Equations (3.17) then reduce to six much simpler equations: A iy At by A, + cy x) = Fy sin (@, 7) EN; =1, 3) (3.198) Ay 3% By 3; + cy 4, = Fg sin (@,t) kN m; (i=4, 6) (3.19b) and these may be recognised as the equations of motion of six second order linear damped spring mass systems with sinusoidal excitation (see Appendix 1). The equations express the six forces and moments required to sustain the single motion oscillation x,. The in phase and quadrature components of each force and moment are given by Equations (A1.12) and (A1.13): 2 = cose, KNm; (i=1,3) (G.20a) x, ‘0 2 Fo 7 =—2 cose, kNmirad; (i (3.200) % ‘p Fa si byo,=-—8 sine, kNim; ( (3.21a) Ho Fro by @, = - 2 sine; KNmim; (i= 4,6) (3.21b) p ‘The components of the six applied forces and moments which are in phase with the motion are therefore associated with the stiffness and inertia coefficients, while the quadrature components are associated with damping. 3.4.2 Heave coefficients IE we set j = 3 in Equations (3.20) and (3.21) and use Equations (3.7) we may find expressions for the heave coefficients 3 + bs; » C33 1 453 » Dss and 53. The most important coefficients are d3,, bj, and C33 which relate the heave motion ( j = 3 ) to the applied heave force (i = 3 ). Figure 3.4 shows the physical mechanisms responsible for these coefficients. 18‘At zero frequency the vessel has no heave velocity or acceleration and the heave force is related to the heave displacement through the single coefficient c,,. This arises because a steady downward heave force produces a steady downward heave displacement and an additional displaced volume of water. The heave force is required to oppose the resulting additional buoyancy force and sustain the heave displacement (see Figure 3.4(c)). A typical relationship between the heave displacement and the heave force required to sustain it is shown in Figure 3.5(a). Provided that the heave displacement is small, this may be approximated by a straight line whose slope is cys. C33 is specifically defined as the gradient of the curve at the origin. Athigher frequencies the in phase component of the applied heave force includes a contribution from the heave inertia or virtual mass A,, (see Equation (3.7a)).This is made up of contributions from the so called added mass a,, as well as the real mass m of the vessel. Ass 4 (a) Heave Sree \ Bee ‘acceleration Yo 3 YY) ¥ ba 3, | @)Heave é ho | Onietty ~~ eae | oer avy | — — | Ves 4 ) Heave onal dle (© Fiplacemant ¥ x Fig 3.4 - Effects of heave mation ‘The former arises because the accelerating hull causes changes in the fluid velocities adjacent to its surface as shown in Figure 3.4(a). The additional force required to accelerate this water as well as the hull is included in the inertia coefficient and the ship behaves as though it has an increased mass. The heave damping coefficient b,, arises because the oscillating ship generates waves which radiate outward and dissipate energy as shown in Figure 3.4(b). A small amount of energy is also dissipated by friction ‘The pitch moment required to sustain the heave oscillation yields estimates of the coefficients gy bss and Cy, (Set i = 5 and j = 3 inEquation (3.19b)). These coefficients quantify the influence of heave on pitch in the equations of motion (3.15) and they occur because local inertia, damping and stiffness forces everywhere along the hull exert pitching moments about the centre of gravity. 19TABLE 31 ADDED MASS AND DAMPING COEFFICIENTS IN THE EQUATIONS (OF MOTION FOR A SHIP WITH PORT/STARBOARD SYMMETRY Velocity tere pitch and force foment ft/s mone acceleration int Pa eee ‘STIFFNESS COEFFICIENTS IN THE EQUATIONS OF MOTION FOR A SHIP WITH PORT/STARBOARD SYMMETRY Key: : significant value small: small: invariably neglected zero, ; zero by geography zero: Zero by symmetry zero; : zero for small motions 80If the ship has fore and aft symmetry like a canoe, the moments arising from the forces on the forward half of the ship will almost exactly balance those arising from the after half of the ship, and these “coupling” coefficients will be very small. For more orthodox forms residual moments, which may not be negligible, will remain. For arbitrary shaped hull forms forces and moments in the other four degrees of freedom (surge, sway, roll and yaw) will also be required to sustain a pure heave oscillation and the relevant coefficients can be identified in a similar way. However, most practical ship forms have port/starboard symmetry so that the forces and moments generated by the pressure changes on the starboard side of the hull are exactly balanced by those on the port side. So all lateral plane excitations associated with motions in the vertical plane are zero. In other words the relationships between the lateral plane forces and moments and vertical plane motions have the form shown in Figure 3.5(b) and all the associated coefficients are zero. Slope = cs, ta) Heave force F, “a “S—~ peck submerges Keel emerges. Heave displacement x, 2 Lateral plane force or moment 0) Vertical plane motion Vertical plane force ormoment p_ Slope = 0 i) : Lateral piane motion Fig 3.5 - Typical heave motion relationships 813.4.2 Pitch coefficients ‘We now consider an experiment in which the model is forced to undergo a pitch oscillation by setting j = 5 in Equations (3.20). The analysis proceeds along lines exactly similar to those used in the heave oscillation experiment and yields the terms ass , Bgs 5 Cqg + 5p + Bgs aNd C55. ‘Again all the lateral plane forces and moments and associated coefficients are zero if the hull has port/starboard symmetry. ‘The pitch oscillations cause local vertical motions everywhere along the hull so that each section of the hull experiences local inertia, damping and stiffness forces analogous to those experienced by the whole model in the heave oscillation experiment, These forces exert moments about the centre of gravity and are responsible for the coefficients d,s , bys and C55. The local forces distributed over the forward part of the hull oppose those on the after part of the hull so that the residual heave forces associated with the coefficients dgy , by, and c,,are usually small. Indeed they would be zero on a hull with fore and aft symmetry at zero speed. ‘The pitch virtual inertia coefficient A,, includes contributions from the so called “added moment of inertia” as well as the true moment of inertia of the ship's structure. ‘This is analogous to the heave added mass already discussed. The true mass moment of inertia may be calculated from Equation (3.2b). 3.4.3 Lateral plane coefficients Expressions for the lateral plane coefficients may be determined in a similar way by setting j= 2, 4 and 6in Equations (3.20). In general, vertical plane forces and moments will always be required to sustain motions in the lateral plane even for ships with port/starboard symmetry. However, the relationship between the vertical plane excitation and the lateral plane motion for ships with lateral symmetry will have the symmetrical U shaped form shown in Figure 3.5 (c). In other words the vertical plane excitation will have the same magnitude and direction regardless of the direction of the lateral plane motion. Since we are concerned only with small motions and our linearisation requires the coefficients to be determined from the slope at the origin, all such coefficients are zero. 3.4.4 ‘Geographical’ coefficients We have scen that some coefficients are zero or small if the ship has port and starboard symmetry. Another class of coefficients are always zero regardless of the hull form. These are all stiffness coefficients associated with the ship's geographical location with respect to the origin Go. No forces or moments are required to sustain surge and sway displacements x, and x, so that 3.4.5 Summary of zero value coefficients Tables 3.1 and 3.2 list 60 coefficients (out of a total of 108) which are zero for a ship with port/starboard symmetry. A further 12 coefficients are usually negligible and are invariably neglected. This results in six much simpler equations for smal] amplitude motions of a ship with lateral symmetry. These are listed in Section 3.5. 823.5 SIMPLIFIED EQUATIONS OF MOTION FOR A SHIP WITH PORT/STARBOARD SYMMETRY. Surge: (m + a, ) # + by = Fyyy sin (@, t+ 7,) aN @.2a) (mr + ayy ) Hy + Dygiy + Gay Hy + Boy Xy + Ogg Hy + Dog ig + Cag Xe Sway: (3.22) = Faq sin (@,t + Y,) kN Heave; (% * as) Hs + bay Hy + O55 Xy + yg Hy + Days Hy + Cas 5 i: (3.22) = Fygg sin (@,t + ¥3) kN Fan hy + Bag hy + Tag + ag ) Hy + Bay Hy + Cas He Roll: @.20) 4 dag He + Bag Hg + Cyg Xe = Fygy Sin (@, t Y,) KN m gg Hy + By Sy + Cay X5 + (Tyg + Ags ) Hy + Dog Hg + Cag Xs Pitch: (200) = Fygy sin (@, t+ Ys) kN m A gy Ey + Dey hy + Ogg By + Degg Hy Yaw: : : + (Teg + Ogg) ig * Bag iy + Cog % en = Fy sin (0,1 +6) kN m 3.6 COUPLING ‘As we have already seen, the vertical plane equations (heave and pitch) are coupled. In other ‘words the heave equation includes terms dependent on pitch so that heave is influenced by pitch and vice versa. However, the surge equation is uncoupled and independent of the other motions. The lateral plane equations are also coupled so that these motions are affected by each other. ‘There is, however, no coupling between the vertical plane motions and the lateral plane motions. ‘This allows the vertical and lateral plane motions to be considered independently.* “This is not necessarily true if the motions are large. 834 STRIP THEORY 4.1 INTRODUCTION Solving the equations of motion (3.22) requires the evaluation of the coefficients and the excitation amplitudes and phases. ‘These may be determined by experiment but this method is laborious and hardly practical for routine calculations. In any case, if experimental methods are to be used, its more appropriate to measure ship motions directly as described in Chapters 9 and 10. Considerable effort has therefore been devoted to developing theoretical methods of determining the coefficients and excitations to allow ship motions to be calculated without recourse to experiment. Various authors, including Tasai (1959), Gerritsma and Beukelman (1967), Salvesen, Tuck and Faltinsen (1970) and Schmitke (1978), have made significant contributions. ‘Their theories are generally similar, differing only in detail and mathematical rigour. They are complicated and a complete description is beyond the scope of this book. This chapter is therefore intended to give an abbreviated presentation of the main features of strip theory in general and is largely based on the methods proposed by Gerritsma and Beukelman. All the theories assume that (a) The ship is slender (ie the length is much greater than the beam or the draught and the beam is much less than the wave length). ) The hullisri id so that no flexure of the structure occurs. (©) The speed is moderate so there is no appreciable planing lift. (@ The motions are small. () The ship hull sections are wall sided. (© The water depth is much greater than the wave length so that deep water wave approximations may be applied. (g) The presence of the hull has no effect on the waves (the so called Froude-Kriloff hypothesis). ‘The theories are grouped under the general heading of strip theory since they all represent the three dimensional underwater hull form by a series of two dimensional slices or strips as shown in Figure 4.1. Each strip is of length &r,, metres (assumed to be small). 84Fach strip has associated local hydrodynamic properties (added mass, damping and stiffness) which contribute to the coefficients for the complete hull in the equations of motion. Similarly the wave excitations experienced by the hull are composed of contributions from all the strips. Fig 4.1 - Representation of underwater hull section shapes by an infinite cylinder Strip theory assumes that these local hydrodynamic properties are the same as would be experienced if the strip were part of an infinitely long cylinder of the same cross sectional shape, as shown in Figure 4.1. In other words three dimensional effects, such as mutual interference between the strips, flow leakage around the ends of the ship and effects due to changes in the shape of the strip over the length 6x,,, are ignored. 42 STRIP MOTIONS Let us first suppose that the ship is undergoing a generalised forced oscillation in all degrees of freedom except surge. If the pitch and yaw oscillations are small the motions of each strip ‘will be essentially confined to the plane of the strip. If the strip is located x,, metres forward of the centre of gravity the motions of a point on the Gixz,, axis will be : ; fg = 2 * py SiN %q =X * Xp) Xe —m_ t0 starboard (4.1) 85Ayg = Xy ~ Mp Sin xy = xy - Xy, xX, m downwards (4.2) Hig = %, rad starboard side down 43) Consider an observer stationed at some fixed point alongside the Er,, axis in Figure 3.1. ‘The oscillating ship passes him at a steady velocity U metres/second. At some instant of time a certain strip is adjacent to the observer and his perception of its lateral velocity is given by the total differential of Equation (4.1): af D f . i 2 Hho = Dae) =p + Hy He + Sys Helse ‘Now the distance x,, from the strip to the approaching centre of gravity is diminishing at the rate Hp =~ U_ misec co) Hence the lateral velocity perceived by the observer is Big = by t Xypig — Ux misec (4.5a) Waterline Fig 4,2 - Velocities of a strip and by a similar argument the perceived lateral acceleration is + Xp) Hy - 2U xy misec? (4.5) Similarly the perceived vertical velocity and acceleration are Xpps + U xy msec (4.62) 86Hig = Hy - Xy tt 2U 5 milsec” (4.6b) and the roll velocity and acceleration are simply ua 3, radisee (47a) Xe rad/sec? (4.7b) An origin 0 in the waterplane, shown in Figure 4.2, is chosen for calculations of the local hydrodynamic properties of each strip. The velocities and accelerations of this point are ay = yg - OG 4{ msec 48a) 4 = ig - OG #{ misec? (4.8b) Vertical motions and roll motions are unaffected by this change of origin, 4.3 COEFFICIENTS IN THE EQUATIONS OF MOTION 4.3.1 General considerations Consider a strip embedded in the infinitely long cylinder lying in calm water as shown in Figure 4.1. The excitations required to sustain the motions in the three possible degrees of freedom may be obtained from equations analogous to the sway, heave and roll equations for the complete ship (Equations (3.22). Note that we are only concerned with that part of the excitation necessary to oppose the hydrodynamic and hydrostatic reactions to the motions. In other words the effects of the real mass and inertia of the strip are excluded from the calculation. For a strip with port/starboard symmetry we have OF, ayy + by A + ay, H+ by = 2 Nm (4.9) Xp tog tot ieee i oF; 3 Gx, Hy + dy Ay + Cy my = a kNIm 4.96) “a1 Tei reLeee tie ety ee eee eee Aig Hy + by Hy yy HL + by Ht cM 8F, == min Bry, (4.9¢) where all the motions,and excitations are referred to an origin O in the waterplane. The primed. coefficients 42 » 6) etc. are all local values (per metre length of strip) analogous to 87the corresponding coefficients for the complete hull (with due allowance for the change of origin). These local coefficients are functions of the size and shape of the section and may be determined using the equations given in Chapter 5. 4.3.2 Coefficients in the heave and pitch equations Consider a ship undergoing simultaneous forced heave and pitch motions in calm water. ‘The momentum of the surrounding water in the plane of the strip is My = ay 8 t mlsec per metre length of the strip. ‘The force required to oppose the hydrodynamic reactions and sustain the motions of the strip is composed of the rate of change of this momentum together ‘with contributions from the damping and stiffness: D / ae aot oF, ( Baws cba bry, kN As successive strips pass the observer stationed alongside the Ex,, axis he will perceive a changing local added mass a,,. So the rate of change of momentum is D t gi af to gt 2M) = a, + kv pp (M5) = a A + oy per metre length of strip. Now we may write for the rate of change of added mass i d : ad dis = Go Can ia = UE Can) per metre length of strip (see Equation (4.4)). Using Equations (4.5) and (4.6) the downward vertical force on the strip becomes ay, ( ¥- Xp) ¥ + 2U 45) des [se Jos-asste) (4.10) 7 9 Cy my) |B BY ‘The total heave force and pitch moment required to balance the hydrodynamic reactions and 88sustain the heave and pitch motions of the ship are obtained by allowing 6x), to approach zero and integrating over the length of the hull: [os (4.11) ~ [a AF, em (4.12) ‘We note the following identities te / dy = ay, tim dxp, py je diy, Xp Ga ft 7 ea eer rar dy, Where 435 q is the local added mass at the stern in tonnes per metre and yy, is the distance of the stern (negative) from the centre of gravity’. Note that 433 4 will be zero if the ship has no transom.” We also note that the local heave force required to sustain a steady downward heave displacement x, metres is Cam = PBB! Bry EN and it follows that foe Fe =p gAy kNim Jobat on 2 mn fos xp, By, = P81, KN mm where the waterplane area and its first and second moments about a transverse axis under the "itis assumed that the cross sectional area at the stem is zero and 4), 2 It is sometimes argued that the local added mass (and damping) at the stern must be nearly zero because of flow leakage effects whatever the shape of the stern. 89centre of gravity are given by = [ Bi dey, m M, = [Bia dx, m3 [Pl ay? ay ‘Substituting Equation (4.10) in Equations (4.11) and (4.12) and comparing term by term with Equations (3.22c) and (3.22e) enables us to obtain expressions for each coefficient in the equations of motion for pitch and heave. The coefficients, for the complete hull are then obtained in terms of the local two dimensional coefficients (433 » b33) and are listed in Tables 4.1 and 4.2, 4.3.3 Coefficients in the lateral plane equations A similar approach is used to determine the coefficients in the lateral plane equations of motion (G.22b, 3.22d and 3.22f). We now consider a ship undergoing simultaneous sway, roll and yaw motions in calm water and the lateral motions of a strip are, given by Equations (4.7) and (4.8).in terms of the local two dimensional coefficients (433 » 33) and are listed in Tables 4.1 and 4.2. The horizontal momentum of the surrounding water in the plane of the strip is My = (ayy 4 + ay Xf) t misec per metre length of strip. The lateral force required to balance the hydrodynamic reactions and sustain the motions of the strip includes contributions from the rate of change of momentum, the lateral motion damping and the lateral force required to sustain the roll velocity: ar, -( Bath) + bi xf 6h 31) oy kn ‘The rate of change of momentum perceived by our stationary observer is and the lateral force is 90‘The angular momentum of the water in the plane of the strip includes a contribution due to the 7 lateral velocity #2: Mi = ay 8/ + ay #1 t msec per metre length of strip. ‘The roll moment about the axis through G required to balance the hydrodynamic reactions and sustain the motions includes contributions from the rate of change of this angular momentum as well as contributions associated with the roll damping and the moment required to systain the lateral velocity. In addition there are contributions associated with the roll stiffness C4 and the moment due to the lateral force 5F, acting through O: D oF eet 1 — any = (Bcd) + bhal + bh sf + elx!) xq, - OG OF, kN m Expanding as before, we obtain oF, da, i aia u-u it) tt Oly Xy ‘a1 + ay (%, + Xp, #, - 2 a, - OG %,) - 0G 8F, Nm ‘The total sway force, roll and yaw moments are obtained by allowing 8x, to approach zero and OLintegrating over the length of the ship: Fy = f dF, kN (4.13) Fy> [aF, Nm 4.14) Fee f Xy, Fy kN m (4.15) Comparing these equations with Equations (3.22b), (3.22) and (3.22f) allows us to obtain expressions for the lateral plane hydrodynamic coefficients for the complete hull in terms of the local values (ajp, by, ete). Again we note that for conventional ships with zero cross section area at the forward perpendicular can where @22 «is the local added mass for horizontal motions in tonnes/metre atthe stern. Again, 422 « ‘wil be zero for ships with no transom.? Similar expressions are valid for the coefficients 43, and a ‘We also note that the roll moment required to sustain a steady roll angle *, radians to starboard is Cua X= B GM, x, KN m so that Ca =m g GM, RN mirad (4.16) where GM, is the fluid metacentric height (ie the metacentric height corrected for free surface effects: sce any text book on basic naval architecture.) Equating terms in equations (4.13) - (4.15) and (3.22b,), (3.224) and (3.22) we obtain the expressions listed in Tables 4.3 - 4.5. 3 Again itis sometimes argued that the added mass and damping coefficients at the stern must be negligible because of three dimensional effects even when the ship has a transom. 9244 EXCITATIONS IN REGULAR WAVES 44.1, General considerations ‘The linearisation of the equations of motion in Chapter 3 allows the wave excitations to be considered independent of any ship motions and to be expressed as functions of the wave amplitude alone. In other words the wave excitations are assumed to be the same as the ship ‘would experience if it were rigidly restrained and allowed no motions at all. According to Equation (3.1) the wave depression at any point (&, + % ) related to the moving origin O is C= fy sin(o,1- kx, cosp+kx, sin) m Since the ship is allowed no motions, the centre of gravity remains above O and we may write x7 Xpp om ‘The wave depression varies across each strip but we assume the ship to be slender (that is, the waterline beam of all strips is much less than the wavelength) and this allows us to calculate the wave depression with sufficient accuracy by setting x20 m ‘The wave depression experienced at each strip is then C=, sin(@,t-Q) m (4.17a) where O = kx, cosy (4.17%) ‘The excitation experienced by each strip is related to the pressures, velocities and accelerations in the water beneath the wave surface. These quantities vary with depth (see Equations (1.17) - (1.23)) and it is usual to simplify the calculations by taking their values at a mean local draught defined by m (4.18) where A’ is the cross sectional area of the strip in square metres and B! is the waterline beam of the strip in metres.* “Note that D is a function of xp,. 93Referring to Equation (1.23) we find that the pressure fluctuation at the mean draught D is P=-pgtiep(-kD)sin(kx-wr) kim? (4.19) relative to the local hydrostatic pressure p g D and, with the assumptions described above, this can be written as - pel ex (-kD)sin(o,t-Q) kNim® (4.20) Similarly the vertical velocity of the water at the mean draught is, from Equation (1.18), v=aGep(-kD)cos(w,t-Q) msec 4.21) and the corresponding vertical acceleration is wo (exp (- kD) sin (at -Q) mlsec? (4.22) The horizontal velocity at the mean draught is, from Equation (1.17) , u -@ te (- kB) sin(w,1-O) =-a@Cep(-kD) msec along the Ex, axis. The athwartships component of this horizontal velocity is w= 0 Cep(-kD) sin misec (4.23) and the corresponding athwartships component of the horizontal acceleration is @ v sin wp misec” 4.24) ‘The slope of the pressure contour at depth D is, from Equation (1.15), oy “kG ew (- kB) cos(w,t- 9) =~ 22 rad and this has an athwartships component given by am sin rad (4.25) ‘The angular velocity of the water in the plane normal to the ship’s axis is equivalent to the rate of change of this component of the pressure contour slope: 94dg = kw exp (-kD)Csinp radisec (4.26) and the corresponding angular acceleration is ip, = -k wv sin rad/sec® (4.27) 44.2 Vertical excitation If we assume that the beam of the strip at the mean draught D is approximately the same as the local waterline beam, the vertical force due to the pressure fluctuation is - BP bx, kN In addition there are contributions arising from the rate of change of the vertical momentum of the water surrounding the strip and a force associated with the vertical velocity of the water. The total vertical force on the strip is then Dewi 1 a] AF yy = | DC Mig) + bys v > BP) Ory, EN where the vertical momentum is Mi; = ay vt misec per metre length of strip. Our stationary observer perceives the rate of change of momentum as D 1 t a 2 mip = + ay De (Me) = sy ¥ + 035 9 Obtaining P, y and ¥ from Equations (4.20) - (4.22) we find that the vertical force on each strip and the associated pitch moment about the centre of gravity are BF = Gy | Pye sin(ot-Q) + Pye cos(,t-Q) | bx, IN (4.28) BFys = Co [ Pag Sin(o,t-Q)* Pye cos(,t-Q) | x, Nm (4.29) where 95P, =(p ¢ 8! -o* a ) exp(-kD) t/ (mlsec? ) t day A Pye = @| by ~ U—= | exp ( -D) 1 / (mlsec*) az, and Pos = Xp; Pag KN me Pog = Xp) Pag KN Im The total heave force and pitch moment amplitudes Fj, and Fs and phases (relative to the maximum wave depression at 0) are then obtained using the equations given in Appendix 2. 443° Sway and yaw excitation Figure 4.3 shows the hydrostatic force experienced by a restrained strip in waves. The inclined water surface causes a lateral shift of the centre of buoyancy from B to B, and the buoyancy force vector is assumed to act normal to the pressure contour at the mean draught D. The buoyancy force is pg A! bx, kN and ithas a horizontal component ££ A! xy, Gy, kN to starboard In addition there are contributions due to the rate of change of horizontal momentum of the water surrounding the strip and forces associated with the lateral and rotational velocities of the water. argo (2 Chad +B + Oh dap +P 8A! ep) By AN The horizontal momentum is foe : Mi, = ay uy + ay, t misec per metre length of strip.re Lis Pressure contour at depth D Fig 4.3 - Lateral plane hydrostatic force and moment on a restrained strip ‘The total horizontal force on the strip is then ae ec gre ty een elgg heal kv > (Maa) = Fo et te im - Ut Dt ap, ary, per metre length of strip. Obtaining ty ty etc, from Equations (4.23) - (4.27) we find that the total sway force on the strip and the yaw moment it exerts about the centre of gravity are Pag sin (@, t ~ Q) + Pye cos (w@, t - Q) | Oxy, kv w= (4.30) Sin (0, = Q) + Pe 60s (0, ¢~ 9) | Bry kN m oye (4.31) where 97da}, day, P, = 0 exp (-4B) sin [ay - 0 SB - «(04-08% J] KN I m ee ap, ay, Pye = 0? expl-4B) sin u( a +p A ~ kas J INI m “ os = ap Pry KN Im sc = Xpy Pag KN I m The amplitudes F,9, and Fg, and phases , and , of the sway force and yaw moment are found using the Equations given in Appendix 2. 444 Roll excitation If the strip shown in Figure 4.3 is wall sided the two triangular wedges ORS and OTU will have equal volumes and identical shapes. Their centres of buoyancy will be at b, and b, and the strip's centre of buoyancy will move from B to B, such that the line BB, is parallel to the line b,b and where 64 is the area of each wedge in square metres. For small values of the slope of the pressure contour &5_ and the centre of buoyancy of each wedge is = metres from the centreline. Hence B B? ay 12 3 m and 98aw - 3 = >; om (4.32) 124 ‘The buoyancy force exerts a roll moment about O: BO opal P8( a5 - OBA ty, Nm per metre length of strip. ‘The total roll moment exerted on the strip about O also includes contributions from the rate of change of the angular momentum of the water surrounding a strip as well as contributions associated with the rate of rotation and the horizontal velocity of water. The equations of motion (3.22) require the roll moment to be related to an axis through the centre of gravity. ‘The lateral force Fy, acting through O exerts a moment about G and this is included to give iz D +o B > ft oF Bans bet + p0( 22-408) ens + Bam a - 0G 5F,, Wm The roll momentum is eH 1 Moa = Oa yy + Ban ty ‘The rate of change of momentum perceived by the fixed observer is 0 | Pas sin(wt-O) * Pye cos( w,t-Q) | bry, ENm (4.33) wherePye = © sin w exp (KD) - 06 [25+ oa’- 104] EN I m The roll moment amplitude F,,, and phase y, are obtained using the equations given in Appendix 2. TABLE 4.1 COEFFICIENTS IN THE HEAVE EQUATION OF MOTION oy [ay ty) 1 (4.34a) byy = f bj dtp, - U ay, kN(mlsec) (4.34) Cx = fea ty = 2B Ay kNIm (4340) ay =~ fq) Oy by -kNMradisec? ) (4.344) , bye = 2W fabs ayy ~ [Hn Ay (4.342) +U Xpy 4 2354 KNM(rad/sec) uy by dt ~ U? ay, - pg M, kNirad (4.34) 100TABLE 4.2 COEFFICIENTS IN THE PITCH EQUATION OF MOTION 455 =~ far ay, dy, kN mi(mlsec*) (4.35a) bys = - fro bys dg, + U Xp) 435, kNml(mlsec) (4.35b) 55 = [an C33 dtm = 9 8M, KN mlm (4.35¢) Gs, = fos G45 Gy, kN ml(radlsec?) (435d) ’ 2»! 2 ot bg = BU fame diy + [sb body Urb. ae. Guage EN mil(radisec) aU fin Be iy + B81, + U2 Hy, Oe KN mlrad (4.35f) TABLE 4.3 COEFFICIENTS IN THE SWAY EQUATION OF MOTION ay = fay diy 1 (4.36a) by = f bx Bp — Ua, kNMmlsec) (4.36b) x4 = f O34 Ay) — OG f ag, dg, RAM rallsec* ) (4.360) Dag = [Bly dtp, - OG f by, dey, + OG U ay, (4.36d) -Uay, kNMradlsec) a5 = | Xm ay, dey, kNi(radlsec? ) (4.360) byg = f Xp bay Wp, ~ U Xpg Qing ~ WU f ay ae, ee ENKradlsec) Gm =~ Uf by dt, + U? ay kNirad (4.368) 101TABLE 4.4 COEFFICIENTS IN THE ROLL EQUATION OF MOTION f iq dxy, - OG f ayy dx, KN mi(mlsec?) (4.37) bay = [bls dig) -U tng ~ OG [bb dry i +06 Uay,, KN mi(mlsec) ; fh, dy, ~ OG fal, dy, ~ OG f ay dy, (4.37c) 0G" faz, dry, KN mi(radisec? ) [Phe ty) ~ U ay, + OG U agg, ~ OG f by de, [Px bq) + OG? [by dp, (4374) -06G? Uay, +OGU ay, kN mirad!sec) Cy, =m g GM, RN mirad (4376) : aa ' og = fa Oe digg ~ OO f ayy hy plait KN mi{radisec? ) bag =~ BU fy digg — U Spr Olan + fy bb by OG f by dty, + OG X44 U ding + 206 U a, dx,, kN mi(rad/sec) (4.378) tg = Ur aly Uf bh diy, + OG Uf Bb dry ec ere (4.37%) -O0G U* ay, KN mirad 102TABLE 4.5 COEFFICIENTS IN THE YAW EQUATION OF MOTION 4 ‘66 = [x An bq EN milsec)? = [Sa bin Bp ~ U Xp) O22 RN mllmisec) : ox. al. = [Fur 24 Atyy ~ OC yy O22 bey KN ml(radlsec? ) = in be ay ~ OB fy Bh dy (4.38a) (4.386) (4.38) (4.38d) + OG U X74 Ging ~ UXqy 4 Gyyq KN mi(radlsec) al Xp) Gp Ep, KN mi(radlsec® ) 24! 2 = [Ahr bes by, ~ U a1 5 Og ~ QW ff xy 42, diy RN mi(radisee) = AU fxg ba dig) + U? Xp g Mg EN mrad 4.380) (4.388) (4382) 103LEWIS FORMS AND THEIR HYDRODYNAMIC PROPERTIES 5.1 INTRODUCTION Progress with strip theory requires the evaluation of the local (two dimensional) added mass and damping coefficients 4 and b, in Equations (4.34) - (4.38). Methods of solution have been developed by Ursell (1949a, 1949b), Tasai (1959, 1960), Grim (1959), Porter (1960) and others. The techniques used are involved and laborious, requiring devious and intricate methods of solution for even the simplest cases. Their application is quite impractical without the aid of a computer. A Heave oscillation y = 2s <= | | i | away | vy Fig 5.1 - Circular cylinder oscillating in the free surface Fortunately computer routines for the calculation of these properties are widely available and naval architects need not usually concern themselves with the details of the complex mathematics involved. It is, however, important to understand the broad nature of the calculations and to appreciate their limitations. This chapter therefore discusses the properties of the commonly used Lewis form approximations for conventional hull cross- sections and gives the expressions for the added mass and damping coefficients, without proof, taken from a comprehensive paper by de Jong (1973). ‘The methods mentioned above generally begin by examining the properties of an infinitely long 104semicircular cylinder’ of radius a metres floating in the free surface of the water as shown in Figure 5.1. The heaving cylinder is shown in the figure, but sway and roll motions (about the longitudinal axis through the origin Q) are also considered. Small motion amplitudes are assumed. The oscillating cylinder generates surface waves which radiate away in the + x directions . The coefficients are calculated with the usual potential flow assumptions of negligible viscosity and compressibility, no flow separation and no skin friction. Ship's hulls do not, as a general rule, have semicircular cross-sections and conformal transformation techniques’ are used to extend the results for the circle into solutions for more realistic hull shapes. In this technique the circle and the flow around it (stream and potential functions) are calculated in the complex z plane where gexetysiret® om G1) These results are mapped into the flow around a hull section in the complex ¢ plane (the hull cross section plane) as shown in Figure 5.2.> C= fe) > 1 m1 Free surface: ern ° ft eye | ty t hited z plane C plane Fig 5.2 - Conformal transformation from the z (circle) plane to the C (ship) plane C is defined as g cm + yy 62) and the two complex planes are related by the transformation "The calculation relates only to the half cylinder below the free surface. The shape above the surface has no effect on the results. ? See any advanced text book on fluid dynamics. 2 The mapping relates only to the underwater shape of the hull cross section. 105o-F@) The functional form of the transformation equations must be determined for every individual case, depending on the size and shape of the section in the Cplane. 5.2. LEWIS FORMS ‘The transformation C=f@=qa ain will map any point on a semicircle of radius a metres in the z plane into a corresponding point on a given shape in the ¢ plane if appropriate values of the coefficients a, @,, dy, a, etc are chosen. The absence of even terms like @ 4, etc. ensures that the transformed shapes have port/starboard symmetry like conventional ship sections. Itis usual to truncate the transformation series to only three terms: 63) We shall see that this allows a wide variety of ship like cross sections to be generated from the semicircle. These forms will not be exact replicas of any given hull cross section, but the match can usually be made sufficiently close to allow adequate estimates of the hydrodynamic coefficients for ship motion calculation. The resulting family of forms are known as Lewis forms, after F M Lewis who first proposed their use (for ship vibration studies) in 1929. Lewis forms are defined by the values of the section area coefficient and we now derive expressions for the transformation coefficients a, , a and a,in terms of these two quantities. 106Since we are here concerned with points on the semicircle of radius a metres we may set ream Substituting Equations (5.1) and (5.2) into Equation (5.3) and separating real and imaginary parts we obtain a pair of parametric equations in 6 describing the shape of the Lewis form in the ¢ plane: aga[(1+a,)sin @ - a, sin(30)| m (64a) Xpp = ay @[(1 - a.) cos @ + a, cos (30)] m 6.4) Substituting 6 = O in Equations (5.4) we obtain the bottom of the semicircle and the keel of the Lewis form: a, +a) m Substituting @ = 7/2 in Equations (5.4), we obtain the intersections of the semicircle and the ‘Lewis form with the water surface: te De ga(l tata) m ‘The beam/draught ratio of the Lewis form is _2(1 +a +a) B! Fe eeseai areas (6.5) Substituting Equations (5.4) we obtain 242 Raja Al 2 (1 - a) - 303) m and the section area coefficient is 1076.6) Explicit equations for the coefficients a,and a, may be obtained by rearranging Equations (6.5) and (5.6) to give a, (iva) (223) (5.78) (5.70) where 2 c-3+40./1-40){ 2:2 © 7 H+2 0» B05 He2 o=os_| o=09 e=09 Fig 5.3 - Examples of Lewis forms 108(a) Area coefficient too large (b) Area coefficient too small Fig 5.4 - Examples of invalid Lewis forms ‘The Lewis form corresponding to a real hull cross-section may therefore be defined in terms of its beam/draught ratio HT and section area coefficient 0. Figure 5.3 shows a range of Lewis forms for various values of H and o*. There is no limit to the permissible beam/draught ratio but only a limited range of section area coefficients are possible. Clearly the formula for 4 (Equation (5.7b)) becomes invalid when C > 9/2 and this will happen if Eye = (H? + 20H + 4) Heit ) 6.8) Lewis forms having section area coefficients greater than this cannot exist. In practice section area coefficients approaching this limit have rather angular shapes of the type shown in Figure 5.4). They are not representative of conventional hull forms: more to the point is the fact that such section shapes would probably experience flow separation around the sharp bilges, and the potential flow techniques employed for predicting the hydrodynamic coefficients would not be expected to give very reliable results. To avoid forms of this nature it is usual to suggest that the Lewis forms should lie completely within the circumscribizig rectangle so that Xm S53 Xp sD! If the section area coefficient is too small the Lewis form will adopt physically impossible shapes with negative values of x,, and x,, as shown in Figure 5.4(b). We therefore also insist that %pq 20, Xp, 2 0 Applying these limits to Equations (5.4) with * Only the right hand side of each section is shown. 109oso 2 2m Im+2 m+ 5.5 ADDED MASS AND DAMPING COEFFICIENTS FOR A ROLLING LEWIS FORM ‘The added mass and damping coefficients for a rolling Lewis form are 1 _ pB!* Py Xx * % Ye 2 Puta seeeeeitaet eat 6.21) wo mete 14 Dy Yq + aX, bi, = sex Po Te * %7e gy mi(radisec) (6.22) 2, 2 Po * % 1. 9 BY? My qo + No Po ay = 2S S28 ki(radfsec? ) (5.23) Po * 7 _ p @ B!* My Py ~ Ny % pee 22 kNMradlsec) (5.24) 2 Po * % Xq, ¥_ 1 My and Ny are defined in Equations (5.16) - (5.19). However Pan, and q, are now obtained by approximate solution for a finite number of unknowns WV of the simultaneous equations H He (1,8) = We (1s W2)= DY Pay fy msec = 7where x Vs (1,0) - Vy (1, 02) = Yann hy msec mb ay (1,8) +2 (1,0) eee 2c 24 -1 Fan = Won (1s BZ) = Yo, (1,895 me 0 ‘Vo, in these formulae is given by Equation (5.20). 5.6 MEASUREMENTS OF LOCAL HYDRODYNAMIC PROPERTIES Accurate calculation of the hydrodynamic properties of cylinders of ship-like cross-section is clearly of paramount importance in the prediction of ship motions in waves. It is therefore somewhat surprising to find that relatively few experiments to verify these calculations have been carried out. Waves radiate away Heave motion Two dimensional hull section Waves radiate away Fig 5.6 - Experiment to measure hydrodynamic coefficients for heave motion 118‘Vugts published the results of the classic experiments in this field in 1968. He tested a number of cylinders in a towing tank at the Delft Shipbuilding Laboratory. Figure 5.6 shows, in simplified form, the arrangement he used. The cylinders were mounted across the tank and an oscillation mechanism was used to impose sinusoidal heave, sway and roll motions in turn, The waves generated by the cylinder motions radiated away and were absorbed by beaches at each end of the tank, some 70 metres from the cylinder. The forces and moments necessary to sustain the cylinder motions were measured and used to determine the added mass, damping and cross coupling coefficients for each motion. Some of his results are shown in Figures 5.7 - 5.10. Cylinder A was circular with (by definition) H = 2 and o = 1/4 while cylinders B and C had ship like cross-sections but the same values of H and o. Figure 5.7 shows that their heave added mass and damping coefficients were virtually identical, confirming that these are essentially functions only of beam/draught ratio and section area coefficient. Small local differences in the shape have little influence on the results, The heave added mass is generally of the same order as the displaced mass p A and rises towards infinity at zero frequency. The coefficients are predicted quite well by the theory. Cylinders D, B and F were all rectangular (except for a small radius at the bilge corners) with © = 1 andbeam/draughtratios H = 2, 4 and 8. Figure 5.8 shows how the heave added mass and damping increase with beam/draught ratio. The triangular cylinder H generally had the Jowest coefficients. The opposite trend is shown for the sway coefficients in Figure 5.9. Finally, Figure 5.10 shows some results for the rolling motion of rectangular cylinder D (o = 1; H = 2). These experiments were conducted using two roll amplitudes of 0.05 and 0.15 radians respectively. The results show that roll motion coefficients depend on the amplitude of the motion: in other words the roll response is non linear, contrary to the assumptions made in the theory described in Chapter 3. Sao Bim «7 Oem 2 2 ron oO is “asl ¢ SAS] aos as a0 0 0s 4m 1s 20) ‘ot9729) wots) Fig 5.7 - Hydrodynamic coefficients (a) added mass and (b) damping for three heaving cylinders. All cylinders have H=2; o=n/4.(After Vugts (1968). 119oO os 10 1s 20 poate ten: (B28) worl) Key Form Ho Symbol Rectangle 2 1 «6 Rectangle 4019 Rectangle F 8 io Tame G ss 0s Fig 5.8 - Hydrodynamic coefficients (a) added mass and (b) damping for four heaving cylinders. (After Vugis (1968)). a’yhA' a 0 05 101520 0 os 10 1s 20 0 B28) (B28) Key Form Ha Symbol RectangleD 2 LO a RectangleE 8 10 [2 Fig 5.9 - Hydrodynamic coefficients (a) added mass and (b) damping for two swaying cylinders. (After Vugts (1968). 120(a) 99 = 0.05 rad (37D vaeR) aaa 0 05 10 15 20 oo 10 15 20 of Ps) ota 2s) Fig 5.10 - (a) Added moment of inertia and (b) damping for a rolling rectangular cylinder showing the effect of roll amplitude (H=2; 0=1). (After Vugts (1968)) 1216 ROLL DAMPING Signal from one corvette to another in full Atlantic gale: Have just seen down your funnel, Fire is burning brightly. From “Make a Signal” by Jack Broome. Douglas-Boyd Books 1994 6.1 SOURCES OF ROLL DAMPING According to strip theory the motion damping arises because the oscillating hull radiates energy in the form of waves away from the ship. For most motions this constitutes the major mechanism for the dissipation of energy. So strip theory estimates of motion damping are generally adequate and reasonable motion predictions are usually obtained. Rolling is unfortunately an exception to this general rule, The wave making damping Byyy, predicted for the potential flow around most hull forms is only a small fraction of the total roll damping which is experienced in reality. Additional important contributions are illustrated in * Figure 6.1. Fig 6.1 - Sources of roll damping Hull forms with relatively sharp comers at the bilges and/or at the keel will shed eddies as the ship rolls. This absorbs a good deal of energy and is a significant source of additional roll damping. Skin friction forces on the surface of the rolling hull may also be significant and any appendages will generate forces which oppose the rolling motion. Eddy shedding, skin friction and the appendage forces experienced at low forward speed arise because of the influence of viscosity which is neglected in strip theory, 1226.2 | NONLINEAR ROLL DAMPING: EQUIVALENT LINEARISATION ‘Wave making roll damping and the damping due to the appendage forces at high forward speed are linear (that is, the roll damping moment is directly proportional to the roll velocity). Viscous roll damping is, however, non linear and generally proportional to the square of the roll velocity. ‘This means that the pure sinusoidal roll response to a sinusoidal wave input (Equation (3.16b)) given by the linear theory is no longer valid. Moreover, the Tinear spectral calculation for motions in irregular waves, which will be described in Chapter 8, is not applicable. In order to circumvent these unwelcome problems we may calculate an equivalent linear damping coefficient which allows for the effects of the non linearities but is used in a linear way. This allows us to continue using the linear equations of motion and the spectral techniques for imregular wave calculations. The equivalent linear damping coefficient is chosen so that the calculated energy dissipated by this term in the equation of motion is the same as that which is actually dissipated by the non linear damping. In general this means that the equivalent linear damping coefficient depends on the roll motion being experienced and a new value of the damping must be calculated for every situation. Since the predominant rolling motions experienced at sea occur at the natural roll frequency ‘we may simplify the treatment of roll damping non linearities by considering only motions at that frequency. Suppose then that the rolling motion is given by y= Xjq Sin 4) rad (6.1) Then the roll moment exerted about the centre of gravity by the equivalent linear damping term will be b,,-2, KN m. In one roll period the work done and energy E dissipated by this linearised damping term will be the integral of the moment times the angular distance moved: 9 wa 4 J Dyy Hg dtg = 4 Weg Buy Kim Hf cos *(w,4f) at ° ° E 2 K O.4 bay Xm KN and the equivalent linearised roll damping coefficient is therefore related to the dissipated energy by E F045 KN mi(radlsec) (6.2) 12363 EDDY ROLL DAMPING ‘Tanaka (1960) conducted a series of model experiments to determine the eddy shedding roll damping characteristics of a number of different hull section shapes as shown in Figure 6.2, Schmitke (1978) used these results to develop expressions for the eddy making damping coefficients for each type of hull section. He postulated that the force due to the eddy shedding acts at the relevant sharp comer at a radius 7, metres from the centre of gravity. The local force resisting the roll motion is expressed in the form FF oO, 8 8 Cy Bg, kN where $ and &x,, are the girth and length of the hull section and C, is an eddy drag coefficient depending on the hull form, Now for sinusoidal rolling motion (Equation (6.1)) the force F exerts a moment about G given by 1 Fo= Frees 2 3 2 cos? pp 8 Cy x3 wg Cos? (0,41) Oxy, KN m and the energy dissipated by this moment in one roll period is KNm UN or triangle Rectangle or full Round bilge or circular Fig 6.2 - Classification of section shapes and definitions of radius r, for eddy roll damping calculations ‘Using Equation (6.2) we see that the equivalent linearised local damping coefficient attributed to eddy making is Xp Tp 8 Cz Oxy, KN mi(radisec) 124‘The total equivalent linearised roll damping coefficient for the complete hull is obtained by integrating along the length of the hull: 4p 0.5 Xp ue foie = — fe 5 Cy dg, KN mi(radlsec) 6.3) It remains to determine the drag coefficient C, which varies along the hull. Schmitke (1978) gave the following empirical formula: Z, Z, exp (-ur, ID’) (6.4) where Z, and Z, are given as functions of B/KG, ¥ (the inclination of the hull section at the waterline) andr /D’ in Figure 6.3; D’ is the local draght and +r, is the effective radius at the keel given by Bl as - 2.69 FE , m for BE e 21 B B RG 0 jor 5S > 21 wee and 1 = 46.7 X49 + 61.7 2% 6.5) with yp in radians. ‘Tanaka (1960) found that Equations (6.4) and (6.5) also applied to very full almost rectangular sections (typical of the midship sections of merchant vessels) with r, now equal to the radius of curvature of the bilge and Z, = 1. For triangular sections at the aft end of cruiser stern ships Schmitke fitted the following quadratic to Tanaka's data: , = 0.438 - 0.449 2% + 0.236 | 4] RG | Round bilge sections have negligible eddy shedding roll damping and C,=0 125for these forms. Figure 6.2 shows the definition of r,, for each of the classes of section shape considered. Fig 6.3 - Z, and Z, for U/V sections (After Tanaka (1960) and Schmitke (1978)) 64 SKIN FRICTION ROLL DAMPING ‘The water flowing past the ship's hull exerts frictional forces on the hull surface. It is usual to express the force acting on a small element of the hull surface in terms of a non dimensional local skin friction drag coefficient defined as e frictional force on element ; p X (local velocity)? x area of element Consider a girthwise element 8s of length 5x,, metres as shown in Figure 6.4. Let the element be positioned at (%), %,;), adistance r metres from the centre of gravity. If the roll velocity is radians/second the velocity at the element will be r x, metres/second and the component velocity tangential to the surface of the hull will be ra, sin (8, +0) ) msec @, and 6, are the polar location of the element and the slope of the hull surface so that: g 2 2| 8rx, sin (0,40, Fig 64 - Roll damping due to skin friction. Then the frictional force acting on the element 6s will be BF = Cy 4 p ra, sin (8, + 0,)P 85 Bx, kV 6.6) and the moment about the centre of gravity is BF, = Cy 4 pr x2 sin? , + 8) B8 By, RN m 61 ‘The work done by the moment in a complete roll cycle is E=4 [ OF, ox, iNm 0 and using Equations (6.1), (6.6) and (6.7) this becomes 4c, aes ar, B= Epox (te bs &r,,)? Bs Xp Ss bs Oxy, Nm Then the equivalent linearised roll damping coefficient for skin friction is obtained from Equation (6.2) by allowing 6x,, and 6s to approach zero and integrating along the hull and around the girth: 127KN ml(radisec) It remains to determine the local skin friction coefficient C,. Schmitke (1978) suggested that the Schoenherr formula for the average skin friction coefficient for “smooth turbulent’ flow used in calculations of ship resistance is appropriate: C, = 0,0004 + [3.36 logyRy - 5.6)? where the Reynolds number is based on the forward speed and length of the ship: pUL By, R, ‘This is clearly inappropriate if the forward speed is zero, and Kato's (1958) formula may then be used: Cp = 1.328 Ry? + 0.014 Ryo where Ry is now a Reynolds number based on the average rolling velocity and the average distance from the centre of gravity: 0.512 p F xp) o, By R, 2 [( 0.887 + 0.145 C, )( 1.7D + BC, ) +2(KG-D)] m Cy is the block coefficient of the hull defined as, A Lengthx Beamx Draught where Ais the hull displacement in cubic metres. 1286.5 APPENDAGE ROLL DAMPING 6.5.1 Drag forces on appendages At zero forward speed the incidence induced on the appendages by the roll motion is 90° and the resulting drag force provides a contribution to the roll damping as illustrated in Figure 6.5(a). If the roll velocity is, radians/second and the appendage is located at a radius r, metres (measured from the centre of gravity to the mid span of the appendage) the roll motion will impart a transverse velocity 7, %, metres/second to the appendage. ‘The resulting drag force on the appendage will be Fy 1 2, Cp 7 Oa AYA, AN 6.9) is the non dimensional drag coefficient. { fe Fi r : oa Fig 6.5 - Roll damping due to a lifting surface ‘The drag force yields a roll damping moment Fy Cy Fria Ay KN m and the energy dissipated in one roll cycle is E=4 [Fy dey kN im ° Using Equation (6,9) this becomes E -2G0 4, rx oy kN m and Equation (6.2) then gives the equivalent linearised damping coefficient due to appendage drag forces at zero speed as 4 Pusan = 3z CoP Hy 4 D Ay ME RN ml(radisec) 6.10) 129‘where the summation is over all appendages. According to Hoerner (1965) a suitable value for the drag coefficient for appendages like rudders or stabiliser fins is oD dT 6.5.2. Lift forces on appendages If the forward speed is not zero the rolling motion induces an angle of incidence on each appendage as shown in Figure 6.5(b). ‘The angle of incidence is rad if ry %, 0 (6.130) 1317 SHIP MOTIONS IN REGULAR WAVES 7.1 INTRODUCTION ‘The strip theory outlined in previous chapters may be used to estimate the motions a ship would experience in regular sinusoidal waves of small amplitude. For conventional ships at moderate speeds these estimates are usually found to be of adequate accuracy for everyday engineering purposes. As an example, this chapter gives the results of a specimen set of calculations of the motions of a frigate of length 125 metres and explains the physical reasons for their nature. Ship motions are of course functions of hull shape and size, and the results given here should not be used to give numerical estimates of the motions of other hull forms. Nevertheless, the same general characteristics will be found to apply to the motions of all conventional monohull ships. 7.2. TRANSFER FUNCTIONS ‘We define the wave depression at the moving origin O (Equation (3.9)) as C= Gsin( a.) m ay and the resulting ship motions (Equations (3.16)) are taken to be X= Xpsin(@,t +5) om @ = 1,3) (7.2a) X= %qsin(w@,6+8,) rad; (f= 4, 6) (7.2b) ‘The motion amplitudes xj. and the phases 6, are functions of the speed U, heading 1 and encounter frequency «,. The amplitudes are assumed to be proportional to the wave amplitude Gq and it is usual to express them in non dimensional form: linear motion amplitudes. x9, 29 and x) are non dimensionalised by dividing by the wave amplitude ¢, ; angular motion amplitudes Xgy X5y and 6 (in radians) are divided by the wave slope amplitude kG, Graphs of the resulting non dimensional amplitudes plotted as a function of encounter frequency are called transfer functions'; they give the proportion of wave amplitude or wave slope amplitude “transferred” by the ship “system” into the ship motions” ? Transfer functions are often wrongly called response amplitude operators (RAOs). Actually the RAO is the square of the transfer function. * Transfer functions may also be plotted as a function of wave frequency or wave length. Each form of presentation has its merits, as discussed in section 7.6. 132‘The phase angles 6, give the phase relationship between the motion and the wave. A positive value means that the maximum positive motion occurs 8, / , seconds before the maximum wave depression is experienced at O. Negative values imply that the motion lags the wave depression at O. 7.3 VERTICAL PLANE MOTIONS IN REGULAR HEAD WAVES ‘We begin by considering the simplest case of ship motions in regular head waves (1 = 180°), Symmetry ensures that roll, sway and yaw are absent and the motions are confined to surge, heave and pitch, ‘The heave and pitch Equations (3.22c) and (3.22e) are coupled so that heave motions are influenced by pitch and vice versa. However, the coupling is usually fairly weak and to a first approximation we may regard the equations as independent. The heave and pitch motions then approximate to the motions of two independent second order spring mass systems as described in Appendix 1. The analogy is not rigorous because the coefficients in the equations are frequency dependent, in contrast to the constant coefficients assumed in the classical equations. Nevertheless, we may define approximate natural frequencies for heave and pitch using Equation (AL6): 5 = radlsec (7.3) m+ ay, = |S —radisec (7.3b) Tos + ss where the heave added mass aj, and the pitch added inertia ag, are to be evaluated at the respective natural frequencies (see Equation (3.2b) for the definition of J,,). Since the natural frequencies are unknown and @,, and a,,are functions of frequency it is necessary to guess an initial value, estimate ayy and a,,, and compute second estimates of the natural frequencies. ‘These second estimates are then used to compute third estimates of a, and ay, and the natural frequencies and so on until the results of the iteration converge. Alternatively an approximate estimate of the natural frequencies may be obtained by assuming : ; 5 2 age mM ty agg Igy tm ‘The surge equation (3.22a) is independent of all the other equations and has no stiffness term C1, %- Surge motions would therefore be expected to be approximately analogous to those of a damped system with no stiffness and there is no natural surge frequency (see Appendix 1). 133Fig 7.1 - Maximum heave and pitch excitations in very long head waves In very long waves the encounter frequency ©, is very low and dynamic effects associated with added mass and damping are virtually negligible. So the excitations and motion responses experienced by the ship are almost wholly attributable to the buoyancy changes as the waves pass the hull. Maximum pitch moment occurs at the wave nodes and maximum heave force occurs at the wave crests and troughs as shown in Figure. 7.1. 120 Maximum heave down Maximum surge forward; ‘Maximum pitch bow up Maximum heave up st by Maximum surge forward; ‘Maximum pitch bow dove Fig 7.2 - Motions in very long head waves. ‘These large excitations in very long waves result in the large motion amplitudes illustrated in Figure 7.2, For moderate ship speeds the wave celerity is very much greater than the ship speed and the vessel may be regarded as virtually stationary as the wave passes by. The ship will behave more or less like a particle of water at the surface, following a circular orbit of radius ¢, given 134by Equation (1,22), So maximum heave (equal to + ¢,) will occur at the wave crests and troughs and maximum surge (also equal to + ¢,) will be experienced at the wave nodes. The ship surges towards the approaching crests and recedes after the crest has passed by. a) Long waves - >> L E— I. $_ ————_SS_ OS +b) Medium waves - = L PAE SIS EEE ¢) Short waves -k << L Fig 7.3 - Buoyancy forces on a restrained ship in regular head waves An observer will see the ship appearing to craw/ like a tiny ant over a succession of very long shallow hills. The ship will always be aligned with the wave surface so that maximum pitch {equal to the wave slope amplitude + & C,) will occur at the wave nodes. ‘Smith measured the total excitations experienced by a restrained model of the Friesland Class destroyer in regular head waves in 1966. Some of his results are shown in Figure 7.4. These show that the ship only experiences significant excitations in head waves when they are longer than about three quarters of the ship length. In shorter waves the buoyancy forces alternate along the ship's hull as shown in Figure 7.3. This, together with the growing importance of dynamic effects at the higher encounter frequencies, results in a general reduction in excitation in shorter waves. ‘Typical calculated heave and pitch transfer functions are shown for the frigate (see Introduction: Section 7.1) at vatious speeds in head waves in Figures 7.5 and 7.6. As expected, all responses approach unity at zero encounter frequency, corresponding to the long wave case discussed above. Ata given speed the responses are generally reduced at very high encounter frequencies because, as we have seen, short (high frequency) waves. do not excite the ship very much. However, as the speed is increased, the wave lengths which do excite the ship are encountered over an ever widening range of frequencies (see the wave length scales in Figures 7.5 and 7.6). At very high speeds this range of frequencies may be wide enough to include the natural frequencies of heave and pitch given by Equations (7.3). ‘The responses may then exhibit resonant peaks as shown at 30 knots in Figure 7.5. Pitch and heave motions are, however, invariably heavily damped and the resonant peaks are never very pronounced. 135Wave length / ship length : 2. /L = 5 2 1 OS 10. Pitch 02F Have al Frsg MCay Go). Fase Maa k Go) | tO aa ere Pitch pitiirit s Heave | fs 0} —Sgpe—$§| E100 J = 0 04 08 12 16 Encounter frequency -, - (rad/sec) Fig 7.4 - Heave and pitch excitations in regular head waves: Friesland class destroyer. (After Smith (1966) Heave / wave amplitude x3 /Cp 10 i Oknois * go oe 30nots 3-200 [- 4 = 400 eee 0 05 10 15 20 25 Encounter frequency 0, - rad/sec Fig 7.5 - Heave transfer functions for the frigate in regular head waves. 136Weave length / ship length :/L 1 os & Pitch [wave slope amplitude xy /ke Cy Phase 8, (deg) 8 0 05 10 15 20 25 | 0. 0 -200 40 Pr) a | Encounter frequency 1, -radsee | Fig 7.6 - Pitch transfer functions for the frigate in regular head waves Figs 7.5 and 7.6 also show the heave and pitch phases, In very long waves the heave phase 6, = 0° aid this indicates that the heave motion is synchronised with the wave motion and that maximum heave (down) occurs in the wave troughs. ‘The pitch phase 6, = - 90° at zero encounter frequency. This indicates that maximum positive (bow up) pitch motion occurs at the wave node (one quarter of an encounter period after the wave trough has passed the ship's centre of gravity (see Figure 7.2)). At higher frequencies these simple phase relationships are somewhat modified by dynamic effects and coupling with the other motions. Nevertheless, they remain largely true over much of the range of frequencies at which appreciable motions are experienced. 7.4 VERTICAL PLANE MOTIONS IN REGULAR FOLLOWING WAVES In following waves the motions are again confined to surge, heave and pitch. Figures 7.7 and 7.8 show calculated heave and pitch transfer functions in regular following waves. These again approach unity when the waves are very long and the encounter frequency approaches zero. Only a limited range of (positive) frequencies can be encountered at any given speed for the reasons discussed in Chapter 1. The transfer functions consequently adopt the shapes shown with two 1371 — 3% pot m4 = ‘1 gue <0 Sos Boal ase 4 : z : : Bos Sos L 4 i i zo Boab = | g Soa al R a ho his : fg iets ° oft = 2 a 2 aS & le | Sao fF + S00 F tie 1S | 25 mos ¢ 20 knots a | | 3-200 $200 2 | zoho = 25 trot NG 4 300 15 knots ao 1 s_10 note 1 1 400 oa 0020406 tog aa 07 04086 Encounter frequency ©, ~radlvee Encounter frequency @,-radsec Fig 7.7 - Heave transfer functions for the Fig 7.8 - Pitch transfer functions for the frigate frigate in regular following waves. in regular following waves. possible motion responses (corresponding to different wave lengths) at any one positive encounter frequency. ‘A third motion response occurs at the corresponding negative encounter frequency (when the ship overtakes the waves). For moderate ship speeds this will only occur in very short waves and the excitations and resulting responses are usually very small. ‘The heave phase 6, is close to zero over most of the range of encounter frequencies for which the response is significant, indicating that the heave motion is again nearly synchronised with the wave motion. The pitch phase 6, is now approximately - 270° (or + 90°) over most of the significant range of encounter frequencies. Maximum positive (bow up) pitch motion now leads the maximum wave depression at the centre of gravity by approximately one quarter of the encounter period. 7.5 VERTICAL PLANE MOTIONS IN REGULAR OBLIQUE WAVES In oblique waves the ship motions are no longer confined to the vertical plane. Roll, sway and yaw motions also occur. However, the vertical plane equations of motion (3.22a), (3.22c) and {(G.22e) for a symmetrical ship are independent of those for the lateral plane ((3.22b), (3.224) and G.221). So the lateral plane motions in oblique waves of small amplitude wiil have no effect on the vertical plane motions and these may therefore be considered in isolation. 138Fig 7.9 - Heave and pitch motions in very long oblique waves. In very long oblique waves the ship again appears to be crawling over a succession of long shallow hills as shown in Figure 7.9. At the crests and troughs the heave motion will again equal the wave amplitude, exactly as in head and following waves. 12 gi r So Fuk 4 fos dul 4 i ? qos fest 4 ioe fob i ze 4 02 Saal ° Bo 10 m 1 ° ot 22} poh | zo. 200 | -o0 | i: : ae soo |! Fam b adie hos on -4a9 + “a 0 su) as 20 2s *os0 p03 10 15 20 2s Exceeds ncomerfrequeny rade Fig 7.10 - Heave transfer functions for the Fig 7.11 - Pitch transfer functions for the Frigate at 20 knots in regular oblique waves _frigate at 20 knots in regular oblique waves The “effective wave length” measured along the ship's track is 2 sec p metres and the corresponding effective wave slope amplitude in the pitch plane is therefore reduced to 2m 1 sec %o kG, cos rad 4) 139(see Equation (1.16)). In these long oblique waves the ship will again align itself with the wave surface, and the maximum pitch, equal to + k C, cos 41, will occur at the wave nodes. Figures 7.10 and 7.11 show typical oblique wave transfer functions for the frigate at 20 knots. On headings forward of the beam (90° < 1 < 180°) the responses are broadly similar to the head wave responses already discussed, but the pitch responses at zero encounter frequency (very long waves) now approach cos 1 (see Equation (7.4)). Only one response is possible at any given encounter frequency and the motions generally decrease with increasing encounter frequency. The heave responses increase as the heading approaches 90° and the wave excitation becomes synchronised along the entire length of the hull. ‘The pitch response decreases as the heading approaches 90° and is usually negligible in beam waves. On headings abaft the beam (0° < p. < 90°) the responses adopt the general form of those already described for following waves. ‘The range of possible encounter frequencies is reduced, depending on the heading, and more than one response is possible at any given encounter frequency. Heave phase 5, is always zero in very long waves, indicating that heave is synchronised with wave depression at all headings. Pitch phase 6, is - 90° on headings forward of the beam and ~ 270° (or + 90°) on headings abaft the beam. i Sg es af a $f af a 0 0s 1 1s 2 0 (0S Ss Wave frequency wo -radee Wavelength / ship length: A/L Fig 7.12 - Alternative presentations of transfer functions for the frigate at 20 knots in oblique waves. 1407.6 ALTERNATIVE TRANSFER FUNCTION PRESENTATIONS tis sometimes more convenient to present the motion responses as functions of wave frequency! ornon dimensional wave length. Figure 7.12 shows the frigate's oblique wave transfer functions plotted in these forms. They have the singular advantage that the responses are now all single valued and the complications of multi valued responses at a given encounter frequency are avoided. Each form of presentation has its own advantages: Plotting responses as a function of “VL allows a physical picture of the ship and wave to be easily visualised: plotting responses as a function of @, allows an appreciation of the importance of natural frequencies (particularly for roll motions) and gives a true indication of the frequency of the motions experienced by the ‘occupants of the ship; plotting responses as a function of «w is often more convenient for calculations of motions in irregular waves (see Chapter 8). 120 | | Maximum heave down | setyve | i Maximum sway to | erboard: Makin : ral to port | te Ty/2 Loy Maximum heave up iu Maximum sway to port; Maximum roll ta starboard Fig 7.13 - Motions in very long beam waves * Sometimes the wave frequency is made non dimensional as o£ & 1417.1 LATERAL PLANE MOTIONS IN REGULAR BEAM WAVES In beam waves pitch motions are, as we have already seen, usually very small. Yaw is usually also negligible and the ship motions are essentially confined to heave, sway and roll. Figure 7.13 illustrates these motions in very long waves. The ship again follows the circular orbit of a particle of water at the surface. ‘The heave and sway motions are therefore equal to the wave amplitude + ¢,: maximum heave motion occurs at the wave crests end troughs and maximum sway occurs at the wave nodes. The ship sways towards the approaching wave crest and recedes after the crest has passed by. Tf there are no internal free surface effects to reduce the effective metacentric height” the ship's deck will always be aligned with the wave surface. Maximum roll, equal to the wave slope amplitude + & (,,, will occur at the wave nodes, Figure 7.14 shows the sway transfer function in beam waves. The sway equation (3.22b) has no stiffness term ¢,, x, 60 there is no sway resonance (see Appendix 1). Sway amplitudes decrease ‘ith increasing encounter frequency but the phase 6, remains essentially constant at about - 90° indicating that maximum positive sway (to starboard) occurs one quarter of an encounter period after the wave trough has passed by (see Figure 7.13). 4 3 : g < > & 3 z z os 2 ~*t6 Wave frequency « - rad/sec Fig 7.14 - Sway transfer function for the frigate at 20 knots in regular beam waves 2 See any text book on naval architecture 142Figure 7.15 shows roll transfer functions for the frigate in beam waves. Roll motion is affected by the sway and yaw motions (see Equation (3.22d)) but in beam waves the yaw coupling is negligible because there are practically no yaw motions. The sway coupling, though of significant proportions, does not alter the basic second order spring mass system characteristics of the roll motion. The roll motion is usually lightly damped so that there is a pronounced resonance close to the natural roll frequency, given by rad/sec (7.Sa) Roll /wave slope amplitude rg 1k Gy Phase 8, (deg) “30 -100 0 02 0 06 08 10 Weave frequency 0 - rad/sec Fig 7.15 - Roll transfer functions for the frigate in beam waves: no free surface effects on metacentric height. (See Equation (A1.6)). The natural roll frequency may then be determined using Equations B.2a), (4.37), and (4.37e). Alternatively an approximate estimate of the natural roll frequency may be obtained as follows. ‘The well known formula? > See any text book on basic naval architecture 143O=0, | Fig 7.16 - Rolling motion in regular beam waves ee BM=—2 m a may be used to calculate the location of the metacentre M relative to the centre of buoyancy B. In this formula the transverse second moment of the waterplane area is 1 foray ot and A is the displaced volume in cubic metres. If the location of the centre of gravity has been determined the solid metacentric height GM, may be calculated from GM, = BM -BG m and the fluid metacentric height Gif, , allowing for any free surface effects, may then be estimated. For preliminary design purposes we may also assume that the added roll inertia is given approximately by dy, = 0.25 Tyg tm? and the natural roll frequency is then given approximately by 144radisec (7.5b) ‘The roll damping increases with forward speed. This gives a general reduction in the peak roll response and a slight reduction in the frequency at which the peak roll response occurs. At zero frequency the roll phase 6, is + 90? indicating that positive roll (to starboard) leads the maximum wave depression by one quarter of a period as shown in Figures 7.13 and 7.16. Atthe natural roll ‘frequency the roll phase is zero and the maximum roll is then synchronised with wave crests and troughs as shown in Figure 7.16. At very high frequencies 5, approaches -90° and the ship then rolls in opposition to the wave slope. Roll motions are then, however, quite small. ‘The responses shown in Figure 7.15 were calculated with no allowance for free surface effects on the metacentric height. If these effects are significant the reduction in the roll stiffness reduces the natural roll frequency and increases the roll response at low frequencies as shown in Figure 7.17. 7.8 LATERAL PLANE MOTIONS IN REGULAR OBLIQUE WAVES Figure 7.18 shows roll transfer functions for the frigate in regular oblique waves. These are plotted in the alternative wave frequency form and the fluid and solid metacentric heights are assumed to be equal. In bow waves the forward speed of the ship increases the frequency of encounter and the roll resonance is excited at lower wave frequencies. In quartering waves the range of encounter frequencies is limited. At 1 = 30° in the case shown the waves are never encountered at the roll natural frequency and the roll resonance is never excited. However;at p= 60° a very wide range of wave lengths (and wave frequencies) is encountered at frequencies close to the natural roll frequency (see Figure 1.16) and the roll response is significantly increased. For this reason roll motion is often a maximum in quartering seas, particularly at high speed. In very long waves («9 = 0) the roll motion amplitude approaches the effective wave slope amplitude ,) given by Equation (7.4). Figures 7.19 and 7.20 show the sway and yaw transfer functions in oblique waves. In very long waves ( = 0) the sway amplitude approaches the athwartships component of the wave orbit radius ¢, sin p. On headings forward of the beam both sway and yaw amplitudes decrease rapidly with increasing wave frequency. Maximum responses occur in quartering seas and rise to very high values when the encounter frequency approaches zero. Strip theory predictions are likely to be inaccurate in these circumstances. In practice the ship would be steered by a helmsman or an autopilot and this would effectively limit these large motion amplitudes. 145Gai, < 6, Rolt/wave slope amplitude xp EC Phase 8, (des) o 02 04 08 08 10 Wave frequency wo» rad/sec Fig 7.17 - Roll transfer functions for the frigate at 20 knots in regular beam waves: effect of fluid metacentric height. Rall /wave slope amplitude xy /ECy 10 Wave frequency o - race Fig 7.18 - Roll transfer functions for the frigate at 20 knots in regular oblique waves. 146é 3 3 & § = s & Yaw /wave slope amplitude xq /k Cy °° 05 1000S 20 Wave frequency - rad/see Fig 7.20 - Yaw transfer functions for the frigate at 20 knots in regular oblique waves. 1477.9 ABSOLUTE MOTIONS ‘The six motions considered so far completely define the possible movements of a ship in a seaway. However, seakeeping studies often call for assessments of the motions experienced at some particular point on the ship, such as the bridge or the flight deck of a warship or the dining room in a passenger ferry. ‘These can be calculated from a knowledge of the six motions we have already defined with respect to the centre of gravity, The angular motions are the same everywhere in the ship but the local linear motions depend on the location within the ship. Let us consider a location defined by the coordinates (py + Xp + Xap ) With respect to the centre of gravity. The longitudinal displacement of this point includes contributions from the surge of the whole ship as well as the products of the lever arms and the pitch and yaw motions. If the angular motions are small the linear displacements relative to the moving origin O are 4%, ~ Xm %p + %yy Xz m_—(Dositive forward) (76) ~ gy Xt Xq) % mM (Positive to starboard) an by = 3, + Ag Xy— Nyy Zs m_— (ositive d 78) ‘Substituting Equations (7.2) we find that each motion is sinusoidal with 5, = 5p sin (@, 6+ 8.) om @= 1,3) as) where the amplitudes are Si) VP2 + Po om (7.10a) Sm = P35 + Pip om (7.100) Sy = YP + Pip m (7.100) where Smz Xe COS 5g + Xyz Xqy CoS HB, om (7.12) Py = Xp Sin 0, ~ Xpy Xyy Sin Oy + Xyy Xqy sin Bg m (7.12) Py = Xyp C08 0, ~ Xpy Xyq COS By + Xp) Nyy COS Bem (7.120) Pry = Xpz Xyq SIM By + Nyy Xeq Sin 8, om (7.124) 148rp C08 85 + pp X4q COS B, ~ Xpy Ky COS 8, m (7.12) rag Sin By + yp Nyy Sit By ~ Xqy Yep Sin dg m (7.128) ae as L 0 os i 1s 20 Wave frequency» -radiscc Absolute vertical motion / wave amplitude $y /%o Relative vertical motion /wave amplitude ry /p os S28 | Fave frequency -radisec Fig 7.21 - Absolute vertical motion transfer Fig 7.22 - Relative vertical motion transfer functions at the bridge of the frigate in functions at the bow of the frigate in regular head waves regular head waves ‘The form of the absolute motion transfer function depends on the position in the ship. Figure 7.21 shows some typical head wave absolute vertical motion transfer functions for a point on the bridge of the frigate. In very long waves (w = 0) the transfer functions approach unity as the ship contours the waves. At very high frequencies the motions become negligible; but at intermediate frequencies the motion phases are such that the contributions from pitch and heave are synchronised and large absolute motions, considerably greater than the wave amplitude, are the result. 7.10 RELATIVE MOTIONS Slamming and deck wetness (see Chapter 14) are of considerable importance in assessing the seakeeping performance of a ship. They are largely determined by the magnitude of the relative motion between the hull and the adjacent sea surface. The relative vertical motion at a point ( Xp) + Xpp » py ) is given by _ = m (positive for increasing immersion) (7.13) where s, is the absolute vertical motion given by Equation (7.8) and ¢ is the local wave depression given by Equation (4.17). Substituting these equations we find that the relative 149motion is Tq sin (s+ 8) m 7.14) where the relative motion amplitude is Ty) = yl 8m ~ 259 Sy COS (By + Q) + CQ mm (7.15) and the phase is given by o Sin Boy + G sin O Byy 608 Oy, ~ Gy c08 D tan 8, 7.16) Q in these equations is given by Equation (4.17b). In practice the presence of the hull causes a considerable distortion of the waves close to the ship and Equation (7.13) is only likely to be reliable at the forward perpendicular. Further aft the equation may underestimate the relative motion by as much as 50%. Techniques for estimating this distortion or “swell up’ are still the subject of research and no method has yet won universal agreement. Nevertheless, Equation (7.13) is still used to estimate relative motion, sometimes with empirical corrections for the swell up as described in Chapter 14. Figure 7.22 shows some typical calculations (with no correction for swell up) for a point on the forefoot of the frigate in head waves. In very Long waves the relative motions are zero because the ship contours the waves. In short waves the ship is essentially stationary so that the wave motion is the only sizeable contribution to the relative motion and the transfer function approaches unity. At some intermediate frequency the motion phases are such that the upward absolute motion is synchronised with the wave depression at the particular location chosen for the calculation. ‘The relative motion is then a maximum and sharply peaked resonances can occur at high speed. 7.11 VELOCITIES AND ACCELERATIONS Since all the motion displacements are of the form x =a) sin(@,t+8) mor rad the motion velocities and acceleration are given by ©, cos (@, 1+ 8) mlsec or radisec ~ %) WF sin (@, +8) misec? or radisec? 150which may be written E54, sin (@,t+ 8 + w/2) misec or radlsec 4H, sin (@,t + & + %) msec? or rad/sec™ where the velocity and acceleration amplitudes are %y @, msec or radisec 2 4%, —mlsec? or radlsec? So the velocity and acceleration transfer functions for any motion can be obtained by multiplying the displacement amplitude responses by the encounter frequency and the square of the encounter frequency respectively. 1518 SHIP MOTIONS IN IRREGULAR WAVES, 8.1 THEELECTRONIC FILTER ANALOGY For many years the assessment of seakeeping performance at the design stage progressed no further than comparisons of ship motions in regular waves. ‘The shortcomings of this approach were widely recognised but further progress had to await the development of new techniques first proposed by St Denis and Pierson (1953). These methods were based on ideas developed in the electronics and communications fields and it is no exaggeration to suggest that their introduction, together with the development of strip theory, form the two main foundations of the modern theory of seakeeping. St Denis and Pierson suggested that the ship could be treated in much the same way as the “black box” electronic filter shown in Figure 8.1. The input signal received by the filter contains a number of different frequency components and these are amplified or attenuated to produce a modified output signal according to the characteristics of the filter. Out Fig 8.1 - The electronic filter analogy The analogy suggests that the ship can also be regarded as a filter, not of electrical signals, but of the waves. In other words we can think of the ship as a “black box” which receives the waves as input and generates ship motions as output. Of course there are a number of different ship motion outputs so we should really regard the ship as a collection of filters, each with its own individual characteristics. Itis also assumed that the filter is ‘linear’ in the sense that the output signal amplitude (the ship motion) at any given frequency is linearly proportional to the input signal amplitude (the wave). Let us consider the case of heave motion in head waves. Figure 7.5 shows typical heave transfer functions for various speeds at = 180° and we may regard these as defining the characteristics of the “heave filter’ of the black box ship. We can see that this is essentially a low pass filter; at low encounter frequencies the wave motions are translated into corresponding heave motions with little attenuation or phase shift. As the encounter frequency rises, the heave 152motions are reduced and at very high frequencies the input is completely attenuated so that there are no resulting heave motions. ‘These ideas can be formalised and quantified by means of the so called “spectral calculation”, This is mathematically valid and rigorous provided that the ship motion responses are linearly proportional to the wave amplitude at any given speed, heading and frequency. This is nearly always true, provided that the motions and waves are of small amplitude, and the spectral calculation is widely used in seakeeping calculations. 8.2 THE ENCOUNTERED WAVE SPECTRUM ‘The first step is to determine the wave energy spectrum as described in Chapter 2. For the time being we shall assume that the waves are long crested. The spectrum may be measured but it is more usual to employ one of the idealised wave energy spectrum formulae (Equations (2.32) or (2.44). These formulae give the wave energy spectrum for a fixed point in the ocean: we require to transform this to the reference frame of an observer on the moving ship. Wave energy spectrin Fig 8.2 - Transforming the wave energy spectrum into the encounter spectrum. ‘We have already seen that waves are encountered by the ship at the encounter frequency defined in Equation (1.35). So the frequencies with which the waves are encountered are increased in head waves and decreased in following waves. It follows that the wave energy spectrum must be shifted along the frequency axis to cover a different range of frequencies when observed from a moving ship. 153Figure 8.2 illustrates the result obtained in head waves: every wave frequency is transformed into corresponding encounter frequency according to Equation (1.35). The frequency interval 600 centred on the wave frequency « transforms into a corresponding encounter frequency interval dc,. The relationship between the two intervals is obtained by differentiating Equation (1.35): bo, =(1- 228 & cos | 60 radlsec 8.1) Now we have seen in Chapter 2 that the area under the wave energy spectrum within the small frequency interval 8c is proportional to the energy contained within that band of frequencies. ‘Transforming the spectrum to the moving frame of reference of the ship does not change this, energy and it follows that the area within the wave frequency range 6 must be exactly reproduced as an equal area within the corresponding encounter frequency range dw,. Hence the ordinates of the wave spectrum perceived by a stationary observer and its counterpart in the encounter frequency domain must be related by S,(@) ba = S, (@,) 0, m? (8.2) or, if 6w and 60, are allowed to become infinitesimal do do, (8.3) S,(@) —-£____ m(radIsec) g > 2w cos p 5, (0, ) = $C) In head waves the effect is to increase the frequencies, widen their range and reduce the spectral ordinate heights as shown in Figure 8.2. The areas under the two spectra are of course, identical since the total wave energy and the significant wave height are unchanged by the transformation. 8.3 THE MOTION ENERGY SPECTRUM 8.3.1 Linear motion spectra The spectrum for a linear (as opposed to angular) motion is calculated by filtering the encountered wave energy spectrum with the appropriate motion transfer function. This is achieved by multiplying each ordinate of the encountered wave spectrum by the square of the motion transfer function at the corresponding encounter frequency. This approach is valid and appropriate for any ship motion when the transfer function is normalised by dividing by the wave amplitude!, " It is usual to normalise the transfer functions for the linear motions like surge, sway, heave, absolute motion and relative motion by dividing by the wave amplitude. 154(0) Encounter frequency domain (8) Wave frequency domain Be ie é 7 7 i 1 T Encountered wave spectrum: 3. Wave spectrum . 2 eer 4 ith eam oa | nea. | ozal pare | | Zo Hes Ze \ 1s 1 ; : Heave transfer function, 0 knots; = 180° |-Heave transfer function. T= 20 knots; w= 180" eo BS é ae J ; ze. a6 2 “Heave encray spectram | “Heave pseudo spectrum E [area = my 1.2m? 7 3° Area = pb J Sab my 1.2m? | = = 20 1 3° 1 go 0s 1 is Soo “5 10 1s Encounter frequency «, (radl/sec) Wave frequency 0 (rad/sec) Fig 8.3 - Calculation of (a) heave energy spectrum in the encounter frequency domain and (b) heave pseudo spectrum in the wave frequency domain. The areas under the two spectra are the same. ‘The calculation for heave is illustrated in Figure 8.3(a). The motion energy spectrum ordinate at each encounter frequency «, is given by a Sq (@,) = SC, ) = m(radlsec) (8.4) 2 ‘The variance of the heave motion is obtained by integrating the heave motion energy spectrum: my = f Sy (0, ) do, m? @5) 0 and the rms heave motion is 0, = Ymg mm 155‘The heave energy spectrum obtained in the encounter frequency domain in this way can be transformed into a heave motion spectrum in the wave frequency domain by using the inverse of the transformation procedure defined in Equation (8.3):* ce a do, a CO) = $5 (0) 7 15 (0, ) £220 U cos mI(radlsec) ‘The resulting spectrum has no physical significance and we shall call it the pseudo spectrum, Nevertheless the variance and rms value of the motion may be still obtained by integrating the pseudo spectrum because the area under the pseudo spectrum in the wave frequency domain is, the same as the area under the true spectrum in the encounter frequency domain. Now the same variance and rms heave motion could have been obtained by conducting the entire calculation in the wave frequency domain as shown in Figure 8.3(b). Here the heave motion pseudo spectral ordinate is given by 2 S,5 (@) = SC@) [2] m(radIsec) (8.6) 0 and the heave motion varjance and rms heave are given by my = f Sa (@) do m* @7 ° The wave frequency domain procedure is generally preferred because no transformations are required and it avoids the complications of negative encounter frequencies and multiple responses in quartering and following waves. However it must be emphasised that the pseudo spectrum is not an indication of the motion spectrum experienced by the occupants of the ship . This can only be obtained by tranforming the psuedo spectral ordinates for the motion into the encounter frequency domain or by working entirely in the encounter frequency domain. 8.3.2 Angular motion spectra Angular ship motion transfer functions? are usually normalised by dividing by the wave slope amplitude and this calls for a slightly different procedure, illustrated for pitch motion in Figure ? Noting that the transformation is valid for ship motions as well as for waves 2Roll, pitch, yaw, autopilot controlled rudder motions, active fin stabiliser motions etc. 1568.4. The wave frequency domain is again preferred for the calculation and itis first necessary to calculate the wave slope spectrum S,, ( « ) (see Equation (2.46)) . The rest of the calculation follows the procedure described for linear motions above. The pitch energy pseudo spectrum is given by #50. . Ss (@) = 8, (@) |= rad7(radIsec) (8.8) res 3 g : = 2 15 T T 7 lil transfer function 2 0 o Siena ta £ Boost 4 0 ! PEE (eereerrereee Terror eer cerIereerereer 3 Pitch pseudo spectrum S004 4 eS 02 - 4 me BEG 1 n o os 1.0 1s Wave frequency (rad/sec) Fig 84 - Calculation of pitch pseudo spectrum for the frigate at 20 knots in irregular head waves. The variance and rms motion are given by f Ss Ce) do rad® (8.9) 0 0, = ymig rad 15784 RMS VELOCITIES AND ACCELERATIONS The rms velocity and acceleration and the motion periods may also be calculated in the wave frequency domain. Using Equations (1.35) and (2.24) the spectral moments for heave are m, = [ Ol S35 Co, ) do, ° (8.10) [le : 2 su} Sg (0) do msec” & ° (1 =0,24) and for pitch ey m, = [ @2 S,5(@, ) do, . . 1) [fe one ¥ x} Ss (@) do radlsec" & (n=0,2,4) and the rms velocity and acceleration are given by 1, = im, misec or radlsec (8.12) = fm, mlsec or radisec? 8.13) Analogous equations apply for other linear and angular motions. Equations (2.26) - (2.28) may then be used to calculate the various mean periods of the motions. 1588.5 EFFECT OF MATCHING THE WAVE SPECTRUM AND THE TRANSFER FUNCTION High transfer function ordinates occurring at frequencies with a good deal of wave energy will give large contributions to the motion energy spectrum. It follows that the rms motion depends on the extent to which the motion transfer function “matches” the wave spectrum. Figure 8.5 shows as an example the effect of varying the modal period of the wave energy spectrum on the ms relative motion for the forefoot of the frigate in irregular head waves, ‘Figure 8.6 shows these and other motions plotted as a function of modal period. The motions which are most sensitive to modal period are those, like relative motion and roll, which have. distinct transfer function peaks. q lees aed 5 g Bo : 7 as = 20 § 2 & o rr ui s = ls - \z ante an07e af Wave frequency « rad/sec Fig 8.5 - Effect of matching the wave energy spectrum with the transfer function. Relative motions at the bow of the frigate at 20 knots in irregular head waves: Significant wave height 5.5 metres. 159? 3 & Head waves Heave B27 4 5 Absolute. Sy ‘motion at bridge Z| & Relative Hn) ‘motion at bow Ey Tee rear ee | 7 T T T T | Rollin beam waves Ese 4 & | Bab S11 Pitchin head waves : ey os 0 1S 20 25 Modal period T, (seconds) Fig 8.6 - Effect of modal period on the motions of the frigate at 20 knots. Significant wave height 5.5 metres 8.6 MOTIONS IN SHORT CRESTED WAVES ‘The procedures outlined above may be used to calculate motions in irregular long crested waves. ‘These are rare, as we have seen, and itis often necessary to extend these techniques to cope with more realistic short crested waves. Now it was shown in Chapter 2 that the continuous short crested wave spectrum S, (<2, v) could be represented by a finite number of attenuated long. crested spectra distributed around the predominant wave direction (see Figure 2.13). Each long crested spectrum is given by §,(@,v)=WS(o) where W is a weighting factor depending on the secondary wave direction v - pt given in Table 2.1 160TABLE 8.1 SPECIMEN CALCULATION OF ROLL MOTION IN SHORT CRESTED WAVES Cosine squared spreading over + 90°, Heading relative to predominant wave direction: = 45°, Significant wave height 5.5 metres. Modal wave period 12.4 seconds. Bretschneider wave spectrum, Ship speed 20 knots. Secondary | Heading Roll Weighting Contribution wave relative to | motion factor to short direction | secondary | variance | (see Table | crested relative to | wave in long | 2.1) variance primary | direction | crested wave spectrum direction Re My i bry deg deg deg? deg? -90 ~45 36.24* 0.000 0.00 -75 -30 12.11 0.011 0.13 -60 -15 2.92 0.042 0.12 -45 0 0.00 0.083 0.00 ~30 15 2.92 0.125 0.37 -15 30 12.11 0.156 1.89 S05 AS: 36.24" O16 Ia G.05 055 15 60, 73.27 0.156 11.43 30 75 18.23 0.125 2.28 45 90 8.64 0.083 0.72 60, 105 434 0.042 0.20 5S 120 2.72 0.011 0.03 30, 135, 1.42 0.000. 0.00 13.22 deg? Rims roll in short crested waves = "23.22 = 4.82 deg + Rims roll in Jong crested waves = "36.24 = 6.03 deg 161i LT Heave 1 feces are == OS ere ertarseedeseecareerleae iH L£\ — snort crested 7) Rus displacement Ras displacement (degrees) (degrees) ° Heading 1 (degrees) PRCT Te re re re ree reece eee ee eee Cee eee Fig 8.7 - Effect of wave spreading on the motions of the frigate at 20 knots in irregular waves.. Significant wave height 5.5 metres; modal period 12.4 seconds. Jt follows that the contribution to the motion variance from each secondary wave direction is Omg = Wimy m? or deg? and the rms motion in short crested waves is es m or deg 162‘The total motion variance in short crested waves is obtained by summing the contributions from all the reduced long crested wave spectra: Vex na gg = bm, m? or deg? veoma Input sped aang, signcon wave hight and modal pared aaa average raltempine Caan eae Calelate average roll damping j<—_} roll amplitude x y Cala functions nat y Cala wa Diver at Spec of oto accep Y k Callers Conpare wh [> |_ prea ae ifirnss scepiaie Fal ren Fig 8.8 - Iterative calculation of non linear motions in irregular waves 163Table 8.1 gives the results of a specimen calculation of the rolling motion of the frigate at 20 kknots in quartering waves (= 45°). In long crested waves the rms roll at this heading is 6.03° for the particular wave spectrum used in the calculation. Cosine squared wave spreading reduces this to only 4.82°. Figure 8.7 shows the effect of wave spreading on heave and pitch as well as on roll. In general, wave spreading smooths out the more extreme variations of the motion. ‘The effects are small for heave but quite dramatic for roll. Wave spreading results in significant roll motions in following waves and, to a lesser extent, in head waves. It also reduces the roll motions at the ‘worst heading by a considerable amount. In the same way spreading increases the pitch motions in beam waves, but the effects at other headings are less pronounced. 8.7 SPECTRAL CALCULATIONS FOR NON LINEAR MOTION RESPONSES ‘The procedures outlined above rely on the assumptions that the motion responses are linearly proportional to the wave amplitude, We have seen that this is usually the case, but roll motions may be an exception to this general rule. In this case a slightly more involved procedure, illustrated in Figure 8.8, is required, Itis first necessary to estimate or guess the rms roll expected in the particular combination of speed, heading, significant wave height and modal period for which the calculation is being performed. The average roll amplitude may then be estimated using Equation (11.23). This done, the total equivalent linearised roll damping coefficient may be calculated for the chosen roll amplitude using the methods outlined in Chapter 6. The average frequency may be taken as the natural roll frequency of the ship. ‘The calculation then proceeds through the usual stages of determining the motion transfer functions and combining these with the appropriate wave spectra to obtain rms values. Wave spreading should be taken into account. ‘The rms roll results are compared with the initial guess: if the differences are large (as will usually be the case) the calculation is repeated using the new rms value and an improved estimate of the roll damping. This procedure is repeated until the rms roll angle reaches an asymptotic value and the calculation is terminated. The whole calculation must then be repeated for every speed, heading, significant wave height and modal period specified. 1649 SEAKEEPING TRIALS “Believe me my young friends there is nothing - absolutely nothing - half so much worth doing as simply messing about in boats"! From “Wind in the Willows" by Kenneth Graham 1908, 9.1 FULL SCALE TRIALS Fall scale seakeeping trials, in which the motions, deck wetness and other seakeeping phenomena of interest are monitored in a measured wave environment, seem to be an attractive method of assessing and comparing the performance of ships in rough weather. The waves are, of course, imegular and it is necessary to record them and the motions simultaneously and to analyse the results using the spectral analysis techniques discussed in Chapter 2. The rms motions (or other seakeeping responses such as deck wetness frequency) can then be ploited as a function of significant wave height and compared with those obtained from trials in other ships. ? | T T T T T Bal PMieo-gg | Bou ge 4 Dol, 3 ‘aor aol 6 4 geal J 2 Leander 4 S22- Tribal oo 4 olye os ii | om 4 6 Be wo 2 Speed (tots Fig 9.1 - Comparative seakeeping trial results: pitch motions in head waves (After Andrew and Lloyd (1981)) ' There is no substitute for practical experience. Probably the most useful and rewarding times in the author's career were spent engaged in seakeeping trials with the Royal Navy. 165However, we have seen that the rough weather behaviour of a ship is a function not only of the significant wave height, but also of the modal wave period, the shape of the wave spectrum and the degree of wave spreading. So results obtained and compared in this simple way are likely to be scattered and possibly misleading if the wave conditions at the time of the trials were in any ‘way dissimilar. In short, seakeeping trials are difficult to conduct satisfactorily because the wave environment cannot be controlled in the experiment. ‘The only way that these problems can be overcome is by running two or more ships side by side in simultaneous trials in nominally identical wave conditions. Trials of this nature have been reported by Bledsoe, Bussemaker and Curamins (1960) and by Andrew and Lloyd (1981). Figure 9.1 shows the rms pitch motions and the mean zero crossing periods measured on two frigates in the latter trials in severe head waves (significant wave height 6-7 metres). These results form an objective comparison between the motions experienced by two particular ships in a particular rough weather environment: any peculiarities or changes in the wave conditions during the time of the trial must have been experienced in equal measure by both ships (since they were only about 400 metres apart) so any differences in the motions may be attributed to differences in the design of the two ships. Trials with a single ship are infrequent because of the expense involved and comparative trials involving two ships are even more of a rarity. Such trials certainly cannot be regarded as a routine ‘way of ascertaining the performance of a new design. In any case, seakeeping trials, by their very nature, cannot be used as part of the design process since they require the design to be finalised and the ship to be built before they can be conducted. Model experiments and theoretical studies provide the only practical method of estimating the seakeeping qualities of a new ship at the design stage. Trials do, however, offer the definitive method of verifying theoretical calculations or predictions based on model experiments. If the wave spectrum is measured it may be used with estimated ‘motion transfer functions to predict the rms motions and periods experienced during the trial as shown in Figure 9.1. Alternatively the motion transfer functions may be estimated from the ratio of the measured motion energy spectrum to the measured wave energy spectrum. Equation (8.4) may be rearranged to give So _ | Sa) % 5, (@,) @.1) with similar expressions for other linear motion transfer functions. In the same way Equation (8.8) may be rearranged to give Ss (@, ) = eee (9.2) Ro ES fe 166with similar expressions for other angular motion transfer functions. 9.2 WAVE MEASUREMENTS. ‘The heart of any seakeeping trial is the measurement of the waves. Without a proper technique for recording and analysing the waves experienced by the ship the results obtained can only be related to potentially unreliable visual estimates of the sea state at the time of the trial. For this reason the use of relatively simple wave buoys such as the Waverider (Figure 9.2) has become de rigeur in all serious seakeeping trials. The Waverider consists of a stainless steel sphere of 700 mm diameter weighing 106 kg. An accelerometer to monitor vertical accelerations is mounted on a heavily damped pendulum within the sphere. The pendulum keeps the accelerometer aligned with the tue vertical Fig 9.2 - Waverider buoy. (Reproduced by permission of Datawell bv.) and the accelerometer's output is integrated twice by electronic circuits to provide an analogue record of the vertical displacement of the buoy. The resulting signal is transmitted to a receiver ashore or on board the trials ship. The buoy may be moored to the sea bottom or allowed to float freely. The system will give satisfactory measurements of waves covering a frequency range from 0.22 to 4.1 radians/second, corresponding to a wave length range from about 4 metres to 1300 metres. This is sufficient for most practical purposes except where measurements of very short waves are required (where, for example, trials are being conducted on a small boat). ? Registered trade mark, Manufactured by Datawell by, Zomerlustraat 4, 2012 LM Haarlem, The Netherlands 167Simple buoys of this kind can give no information about the directional spreading of the waves (or even the predominant wave direction). More sophisticated buoys which can give information on wave spreading are now becoming available and techniques of hindeasting the waves from ‘measured wind data have also been used. These have not, however, yet became routine and most trials still use simple measurements of point wave spectra, relying on visual observations to determine the predominant wave direction. 9.3 SHIP MOTION MEASUREMENTS: Most seakeeping trials are concerned with the measurement of the ship's displacements in the six degrees of freedom. Angular motions are generally measured using gyros of the type used in aircraft navigation systems. In warships it is often possible to use the ship's own weapon system gyros but on other ships the trials team will usually have to supply its own transducers. Direct measurements of the linear motion displacements are impossible because no suitable fixed datum levels are available. Instead, the usual practice is to measure the surge, sway and heave accelerations using accelerometers mounted on a small platform stabilised by gyros to remain in a horizontal plane (the gyros may also be used to measure the roll and pitch). Stabilisation is necessary because a "strapdown' accelerometer fixed to the ship's deck would measure the lateral force estimator or apparent acceleration in the plane of the deck (see Equation (15.12)) rather than the true horizontal acceleration. In the same way the apparent surge acceleration in the plane of the deck would be affected by pitch. Measurements of heave acceleration are reasonably immune from these effects providing the pitch and roll angles are not too large. ‘The stabilised accelerometers should ideally be located at the centre of gravity of the ship so that true measurements of the surge, sway and heave accelerations are obtained. In practice this is often impossible to achieve. Even if the location of the centre of gravity is known when the transducers are installed it may well turn out to be in some inaccessible or inconvenient location. Itis, in any case, unlikely that the exact location of the centre of gravity will be known at the planning stage because it depends on the loading state of the ship and precise determination of its position will only be possible at the time of the trial. So in practice itis likely that the transducers will be located at some arbitrary position relative to the centre of gravity and it will be necessary to correct their measurements to allow for this error. If the transducers are located at (gj, Xj, 43) they will measure local absolute accelerations ,, 5, and §, given by differentiating Equations (7.6) - (7.8). The tue accelerations of the centre of gravity are then given by + Xp He ~ Xp #5, msec” (9.3) + Xpy Hy - Ap) Ze milsec? @4) ~ Xpy Hy + Xp) Ee mlsec® @5) 1689.4 MEASUREMENTS OF OTHER SEAKEEPING RESPONSES Seakeeping trials are often concerned with the measurement of responses other than motions in waves. Typical examples are deck wetness and slamming. Very simple instrumentation will often suffice for measurements of the frequency of occurrence of these events. Indeed a seasoned observer with a watch, a pencil and a log book is really all that is required. More permanent records of deck wetness, which can be analysed at leisure in the less distracting environment of the shore based laboratory, can be obtained by a video recording of the forecastle. This technique ‘was used successfully in the comparative seakeeping trials reported by Andrew and LLoyd (1981) and their average deck wetness interval results are shown in Figure 9.3, 100 5 Average deck wetness interval (seconds) 12 4 16 18 20 22 Speed (knots) Fig 9.3 - Deck wemness measurements on a frigate in head waves. (After Andrew and Lloyd (1981)) ‘Measurements of deck wetness severity can be obtained by mounting pressure transducers at suitable locations on the ship's upper works. Slamming frequency measurements may be obtained by using strain gauges to monitor the bending moment experienced by the main hull girder. Siams will then be readily detected as short periods of high frequency (typically 1.0 - 2.0 Hz) oscillation in these records. These oscillations are caused by the hull whipping after each slam and are quite distinct from the longer period oscillations in the bending moment experienced at the wave encounter frequencies. Alternatively the high frequency vibrations may be detected using an accelerometer fixed to the ship's structure. Slamming severity may also be monitored by analysing these records. Pressure transducers are sometimes let into the hull surface to measure local slamming impact pressures. It will, of course, be necessary to dock the ship if the transducers are to be fitted below the waterline. This technique requires transducers and a recording system capable of responding 169to the very short rise times (of the order of milliseconds) typical of hydrodynamic impact phenomena. This will often be incompatible with the requirements for the recording system for ship motions and a separate system may be necessary. 9.5 | RUNLENGTHS AND SHIP COURSES. We have seen in Chapter 2 that at least 100 pairs of peaks and troughs are required in an irregular time history to ensure a reasonably reliable estimate of the rms motion. Each trial ran must be of sufficient duration to achieve this minimum standard. The actual length required may be estimated from strip theory calculations of the motions in the wave spectrum expected during the trial. The mean period of the peaks for each motion may then be calculated from Equation (2.27). Figure 9.4 shows the results obtained for a trial planned in a frigate. 25 Minimum run time —> 2 = 5 8 & > s 3 i& i 315 tS = = . g £ 3 8 s Soe a $ & 3 3 gS 5h Zo 0 o 30 60 30 120 150 180 Heading (degrees) Fig 9.4 - Minimum run times for 100 motion peaks. Frigate at 20 knots: modal period 12.4 seconds ‘As expected, the mean periods are longest in following waves where the encounter frequencies are low. The run time required to achieve 100 motion cycles is given by 1007, = 2 tes 9.6) a7 gg minutes (0.6) where 7, is the mean period of the peaks (in seconds) for the chosen motion. The required run time is given by the maximum envelope value of T,, for all the motions and this is also shown, in Figure 9.4. 170Jn practice longer runs than this absolute minimum are advisable, A certain amount of additional time should be allowed for the ship to settle onto its new course and speed at the beginning of each run: more importantly it should be realised that every precaution should be taken to ensure that data of adequate quality are collected. The opportunity of conducting a seakeeping trial occurs so rarely that it would be false economy to shorten the runs because of economic or operational pressures. If time is short it is better to reduce the number of runs rather than their Jengths. Predominant : wave direction H 1 Following Head Wave buoy 10 miles Fig 9.5 - Typical sequence of runs for a seakeeping trial at 20 knots ‘Very long runs are undesirable because they may take the ship too far away from the wave buoy. ‘The wave measurements will then be unrepresentative of the conditions experienced by the ship during at least part of the run. Long runs also increase the risk of the wave conditions changing during the run. A good rule of thumb is to add a contingency of 10 minutes to the minimum calculated run time for each course. Figure 9.5 shows a sequence of courses for a trial planned on this basis in the frigate at 20 knots. The wave buoy is launched at the beginning of the head sea run and the course sequence is chosen to minimise the distance from the ship to the buoy. 17110 MODEL TESTING “The only difference between men and boys is the size and cost of their toys” Source unknown: seen by the author in a vintage car dealer's showroom in Monterey, California, USA in 1986. 10.1 MODEL SEAKEEPING EXPERIMENTS Preparing a ship for a seakeeping trials an expensive and time consuming business. It is usually necessary to select the trials period some time in advance and there can be no guarantee that suitable weather conditions will occur. Many a trial has been postponed or cancelled because there were no appreciable waves on the days allocated for it! Even if waves of suitable severity are experienced, uncertainties about the degree of wave spreading may still limit the utility and general applicability of the results obtained. However, the main disadvantage of full scale trials is that they require the ship to be built before they can be run. As such they are virtually useless as a method of assessing the seakeeping qualities of a ship at the design stage. ‘Model testing provides an attractive alternative. Models are much less expensive than ships and can often be entirely dedicated to the required experiments. Moreover the model can be built before the prototype ship and a number of alternative designs can be tested. Indeed, before the advent of strip theory, model testing provided the only method of assessing the seakeeping qualities of the ship at the design stage. If the model is to be tested in a towing tank or a seakeeping basin the waves can be produced (and reproduced) to order. Suitable measurements are generally easier to accomplish than at full scale. However, scaling problems can never be completely overcome and it must be admitted that model tests in the controlled artificial environment of the laboratory always lack something of the uncertain harsh reality of the real world experienced by the ship. 10.2. MODEL EXPERIMENT SCALING 10.2.1 Dimensional analysis Consider a model ship in a system of regular long crested waves, How should the test conditions be scaled to ensure that the model's motions are an accurate reproduction of the motions which would have been experienced by the ship at full scale? To answer this question we employ the techniques of dimensional analysis. These are discussed in detail by Massey (1986) and in many other text books and we shall not give a general treatment here. Suffice it to say that the technique allows the proper identification of the correct model test conditions in terms of non dimensional groups of the quantities which are relevant. 172TABLE 10.1 DIMENSIONS OF QUANTITIES APPEARING IN EQUATION (10.1) 2 0 1 0 u 0 1 a H ° ° 0 L o 1 0 Cx] 0 1 0 tr) 1 2 0 r 1 3 0 Let us consider as an example the heave motion of a ship and its model in regular waves. For the time being we may assume that we have no detailed knowledge of the physical processes involved: even so we might surmise that the heave amplitude x,, will be a function of the wave amplitude ¢, and wave length 2, the speed U and heading u. and the size, shape and inertias of the hull. In addition, the heave amplitude would be expected to depend on the physical properties of the water (density p and viscosity 1) and the acceleration due to gravity. We might therefore write a general mathematical expression relating these eleven quantities as: i [lo Be Os te ba xy LET Po tye } m 0.1) where f, is some as yet undetermined function which will be the same for both model and ship. [ x, ] represents a sufficient number of coordinates to define the shape of the hull and [ 7 ] represents the moments of inertia of the hull. The dimensions of these quantities are listed in Table 10.1. 173Massey (1986) shows how an expression of this form can be rearranged and written in terms of smaller number of non dimensional parameters. The theory of dimensional analysis allows the required number of non dimensional parameters to be determined and for the case considered here the non dimensional heave amplitude can be expressed as a function of seven such groups. ‘Many different formulations are possible and equally valid but it is convenient to consider the form a _, [bo a U Cx,] [7] purl (102) a ‘L' ven’ L' pL by where f, is some unknown function which is the same for both model and ship." Equation (10.2) tells us that the non dimensional heave amplitude “22 will be the same at both 0 model and full scale provided that all the parameter groups on the right hand side of the equation have the same numerical values at model and full scale. This requirement dictates the conditions required for the model experiment, ‘We define the model scale or dimension ratio as L Re (0.3) uM where Ly, and L, are the lengths of the model and the ship. The dimension ratio generally lies in the range from 10 to 100. "These particularnon dimensional parameters are chosen with the benefit of hindsight and a knowledge of the characteristics of ship motions in regular waves. Other selections are, in principle, equally valid providing that the total number of non dimensional groups is seven and that they are all independent. Any one of the groups may be multiplied or divided by any other non dimensional group (including ones that do not appear in Equation (10.2). For example the non dimensional wave ampliuae may be divided by the non dimensional wave length 4 to ee i & aH A express the non dimensional wave amplitude as z and this is sometimes more convenient. ied by an [E * to give an alternative non dimensional wave frequency parameter w,|— (see Table % 1 for relevant formulae). Similarly the non dimensional wave length may be multi 17410.2.2 Model and wave dimensions 7 Cx] If the dimensionless coordinates are to be identical for model and ship we require each corresponding dimension” to be related by Zam, as ly ls so that Hos py = 10.4) ‘om = Rm 0.4) In other words the model must be geometrically similar to the ship in all respects. ‘The underwater hull shape should be accurately reproduced and it is convenient to mode! the bull up to the weather deck. It is not usually necessary to represent the superstructure as this has little effect on ship motions except possibly in very severe conditions. In the same way the requirement that the non dimensional wave amplitude and wave length must be the same leads to Cou = = a (10.5) and ay ‘s gm (10.6) 10.2.3 Mass and inertia ‘The requirement to represent the underwater hull form accurately demands that the model's waterline be correctly located. This requires that the model's mass be correctly scaled. ‘The ‘model's mass is = Py [By Dy Ody 1 (10.7) where the integration is performed over the length of the huli. Now all the model's dimensions must be scaled according to Equation (10.4). So the model mass may be written as p, Pur Ms mt { Bs Dz 6 drys = 9, Re t (10.82) s 2 Dimensions’ may be taken to include quantities such as the radii of gyration, metacentric height, centre of floatation and the coordinates of the centre of gravity. 175If Pus = Ps this becomes pests (10.8b) mena t So if the water densities are identical the model mass is reduced in proportion to the cube of the dimension ratio. In practice itis usual to test models of ocean going ships in fresh water so that the water densities are not identical and this should be taken into account by using Equation (10.88). Tn this case itis generally assumed that p = 1.025tonnes/metre? for salt water and ‘The model's moments of inertia are scaled by ensuring that In I, Pula Ps Ls and the model's moments of inertia are then given by Puy Ts aR? tm? (10.92) s If Py = Ps this becomes tm? (10.9b) So the model's moments of inertia must be reduced in proportion to the fifth power of the dimension ratio if the water densities are the same at model and full scale, Again it is necessary to allow for differing water densities by using Equation (10.9a.) 17610.2.4 Frequencies and periods ‘The wave frequency is given by 278 padisec (see Table 1.1). Using Equation (10.6) it follows that @y = ©, YR radisec (10.10) Model wave frequencies must therefore be increased in proportion to the square root of the dimension ratio and are higher than the corresponding frequencies in ship scale. Since T= sec (10.11) Ty = 2 see (10.12) Model wave periods must therefore be reduced in proportion to the square root of the dimension ratio. 10.2.5 Heading ‘The heading 41, is of course, a non dimensional group in its own right and it is self evident that, the model and full scale headings must be the same. 10.2.6 Speed: the experiment conundrum, Consider a small particle of water somewhere near the ship. Suppose that there is a corresponding model scale particle at the corresponding location (defined by Equation (10.4)) near to the model and that each particle has dimensions which are some small fraction of the ship's or model's length. Then the mass of the particle in each case will be proportional to p L?. ‘The particle's velocity v_ will be proportional to the velocity of the ship or model and the distance ds it moves in a given time df will be proportional to the ship's length. The particle's acceleration can be written 77which is proportional to U/L. So the inertia force (mass x acceleration) experienced by the particle will be proportional to p L? U*. The gravity force on the particle is (mass x g) which is proportional to p g L?. So we find that the Froude number’, which is one of the non dimensional parameters in Equation (10.2) , is the square root of the ratio of the inertia and gravity forces on the particle: inertia force _ | p L? U? U. (10.13) ~ | sees ‘gravity force pel? fel ‘The surface area of the fluid particle is proportional to L? and the viscous shear stress is proportional to pty times the velocity gradient.‘ The velocity gradient is proportional to U/L and the viscous force on the particle is therefore proportional to Hy U L. So we find that the Reynolds number’, another non dimensional parameter in Equation (10.2), is the ratio of the inertia and viscous forces on the particle : inertia force _ p U?L? _ pUL R, = tnertia force _ epee ‘N ‘viscous force ty U L By (10.14) Referring to Equation (10.2) we see that we require the model and full scale Froude numbers to be identical: Uy Us V Suku y 8s bs which leads to the speed scaling law Uy = US mi =—5 msec (10.15) yR so that the model speed must be reduced in proportion to the square root of the dimension ratio. Tn honour of William Froude who realised its importance as a parameter for determining the required speed for model experiments in the nineteenth century. The French mathematician Reech also recognised its significance and it is sometimes known as the “Nombre de Reech” in France. 4 See any text book on fluid mechanics. In honour of the nineteenth century mathematician Osborn Reynolds. 178