Naval Architecture Second Revisien Volume III « Motions in Waves and Controllability Edward V. Lewis, Editor Published by The Society of Naval Architects and Marine Engineers 601 Pavonia Avenue Jersey City, NJCopyright 1980 by The Secietyof Naval Archit and Marine Raginers 1 i understood and agreed that nothing expen erin i intended or shall be construed to {ry erro corporton ay ih rote ae agin SNAME or ey ft Library of Congres Catalog Card No. 860520 TBHN No oousrro003, Printed in the United Sater of America ‘int Pring, Noemi, 1980Preface ‘The aim of this second revision (third edition) of the Society's successful Principles of Naval Architecture was to bring the subject matter up-to-date through revising or rewriting areas of greatest recent technical advances, which meant that some chapters would require many more changes than others. The basic objective of the book, however, remained unchanged: to provide a timely survey of the basic prin- ciples in the field of naval architecture for the use of both students and active professionals, making clear that research and engineering are continuing in almost all branches of the subject. References to available sources of additional details and to ongoing work to be followed in the future are included. ‘The preparation of this third edition was simplified by an earlier decision to incorporate a number of sections into the companion SNAME publication, Ship Design and Construction, which was revised in 1980. The topics of Load Lines, ‘Tonnage Admeasurement and Launching seemed to be more appropriate for the latter book, and so Chapters V, VI, and XI became IV, V and XVII respectively, in Ship Design and Construction. This left eight chapters, instead of 11, for the revised Principles of Naval Architecture, which has since become nine in three volumes. ‘At the outset of work on the revision, the Control Committee decided that the increasing importance of high-speed computers demanded that their use be dis- ‘cussed in the individual chapters instead of in a separate appendix as before. Tt was also decided that throughout the book more attention should be given to the rapidly developing advanced marine vehicles. In regard to units of measure, it was decided that the basic policy would be to use the International System of Units (S.L). Since this is a transition period, conventional U'S. (or “English”) units would be given in parentheses, where prac tical, throughout the book. This follows the practice adopted for the Society's ‘companion volume, Ship Design and Construction. The U.S. Metric Conversion Act of 1975 (P-L. 94-168) declared a national policy of increasing the use of metric systems of measurement and established the US. Metric Board to coordinate voluntary conversion to SI. The Maritime Administration, assisted by a SNAME fad hoe task group, developed a Metric Practice Guide to “help obtain uniform ‘metric practice in the marine industry,” and this guide was used here as a basic reference. Following this guide, ship displacement in metric tons (1000 kg) rep- resents mass rather than weight. (In this book the familiar symbol, 4, is reserved for the displacement mass). When forces are considered, the corresponding unit is the kilonewton (ki, which applies, for example, to resistance and to displacement weight (symbol W, where W = pAg) or to buoyaney forces. When conventional or English units are used, displacement weight is in the familiar long ton unit (Continued)PREFACE (2240 Ib), which numerically is 1015 x metric ton, Power is usually in kilowatts {2 W ~ 134 hp. A conversion table also is included in the Nomenclature at the end of each volume ‘The first volume of the third edition of Principles of Naval Architecture, com: prising Chapters I through IV, deals with the essentially static principles of naval, architecture, leaving dynamic aspects to the remaining volumes. The second vol- lume consists of Chapters V Resistance, VI Propulsion and VII Vibration, each of which has been extensively revised or rewritten. Volume Ill contains the two final chapters, VIIL Motions in Waves and 1X Controllability. Because of important recent theoretical and experimental devel- ‘opments in these fields, it was necessary to rewrite most of both chapters ané to add much new material. But the state-of-the-art continues to advance, and so extensive references to continuing work are included, November 1989 Edward V. Lewis EditorTable of Contents Volume III Page Page Preface ti Acknowledgments. i Chapter 8 (VII) MOTIONS IN WAVES Rosent F, Beck, Professor, University of Michigan; WitLIaM B. Cusaains,** David Taylor Research Center; Joux F, Datzeut, David Taylor Research Center; PaiLir MaNDEL,* and Witt1a C, WEBSTER, Professor, University of California, Berkeley 1. Inteodeton 1 5 Derived Responses 109 2 Geean Waves a 8 Control of Ship Motions 136 5 Ship Responec to Regular Waves "at 4 Kasesiog Shp Seaway Performanae”” 13? 4 The Shiptn'a Seaway" a Design Abpects 160 References 1m Nomenclature 18 Chapter 9 (1X) CONTROLLABILITY C. Livooux Crawe,"* Exxon Corporation; HaRv20 EDA, Professor, Stevens Institute of Technology; AUBxANDER C, LaNDSBURG, US, Maritime Administration 1. ateouction sa 1st 9, Theoretical Predition of Design 2 The Control Laop and see ‘Goalfiient end Systems ations of oon se Tentitestion ae 8 ‘Stability and 10. acclrtng Sig i aki’ at rations 195, U1 Automate Contrl Systems 2 4. Analjate of Course Keoping IE Bitecta ofthe Environment. oes ‘Controls Fixed Seabiy 19 Id. Vessel Waterway Inersctons 29 & Suabilty and Conta 25 Tk, Hydrodynamic of Control Surfaces.” 231 & Anabel ofrrsig iy 20 TE Manouvering Tras and Performance 1. Bree Hanning Model Fests ‘Requirements a5 tnd Hydraule Modal 25 16. Appleation to Dec at 8 Nonlinear Equations of ition Tt, Detign of Rudder and Other Control fd Captive Model Tests ar Devices 388 References 08 Nomenclature as General Symbols Bi Index a Now retired Ds ceased Note:The ofc afilitions given are thve atthe te of writing the chaptersAcknowledgments In this Volume III, the Editor wishes to thank the authors of Chapter VIII, Robert F. Beck, John F, Dalzell, Philip Mandel and William C. Webster, for stepping in to complete the chapter on Motions in Waves after the untimely death of William E, Cummins. He also acknowledges the valuable assistance of SNAME T&R Panel H7 (eakeeping Characteristics) chaired by David D. Moran, in reviewing and ‘commenting on early drafts of the chapter manuscript. Generous permission was granted by D. C. Murdey and his associates in the National Research Council of Canada for us to publish excerpts from their valuable reports on a series of model tests in calm water and in waves. Drafting services were provided by Keith L. MacPhee. ‘The Baitor also wishes to express his appreciation for John Nachtsheim's val- uuable efforts in guiding the completion of Chapter IX on Controllability, and to Alexander Landsburg for joining in to assist the original two authors. All three ‘authors wish to acknowledge their indebtedness to Philip Mandel, the author of ‘the corresponding chapter in the preceding edition. Extensive use has been made of the original text and figures, ‘The authors also wish to thank the members of Panel H-10 (Ship Controllability) who provided useful comments, especially Abra- ham Taplin, Completion of this chapter was greatly facilitated by Roderick A, Barr, who aasisted in organizing the chapter in its early stages, by the excellent technical review and suggestions given by John A. Youngquist, and by the drafting services of Robinson de la Cruz ‘The Control Committee provided essential guidance, as well as valuable assist- ance in the early stages. Members are: John J. Nachtsheim, Chairman ‘Thomas M. Buermann William A. Cleary, Jr. Richard B. Couch Jerome L. Goldman Jacques B. Hadler Ronald K. Kiss Donald P. Roseman Stanley G. Stiansen Charles Zeien ‘The Editor wishes to thank all of the Volume III authors for their fine work and full cooperation in making suggested revisions. Finally, he acknowledges the in- dispensable efforts of Trevor LewisJones in doing detailed editing and preparing text and figures in proper format for publication. November 1989 EV. Lewis ‘EditorCL. Crane, H. Eda, “A. Landsburg CHAPTER IX Controllability Section 1 Introduction LA. Definition and Scope. Controllability encom- passes all aspects of regulating a ship's trajector Epeed, and orientation at sea as well as in restric ‘waters where positioning and station keeping are of particular concern. Controllability includes starting, Steering a steady course, turning, slowing, stopping, backing, and in the ease of submarines, diving. The study of the complex subject of controllability usu- ally divided into three distinet areas or funetions: (a) "Coursekeeping (or steering)—The mairtenance of'a steady mean course or heading. Interest centers ‘on the ease with which the ship can be held to the ‘). Manewvering—The controlled change in the di- rection of motion (turning or course changing). Inter~ fst centers on the ease with which change ean be accomplished and the radius and distance required to Accomplish the change. @). Speed Changing—The controlled change in speed including stopping and backing. Interest centers fon the ease, rapidity and distance covered in accom plishing changes. Performance varies with water, depth, channel re strictions, and hydrodynamic interference from nearby vessels of obstacles. Coursekeeping and maneuvering characteristics are particularly sensitive to ship trim For conventional ships, the two qualities of course keeping and maneuvering may tend to work against tach other, an easy turning ship may be difleult to ‘Keep on course whereas a ship which maintains course well may be hard to turn, Fortunately a practical com promise is nearly always possible. Since controllability isso important, it isan essential consideration in the design of any floating structure. Controllability is, however, but one of many eonsié- erations facing the naval architect and involves co Promises with other important characteristics, Some Solutions are obtained through eomparioon with the characteristics of earlier successful designs. In other ‘caser, experimental techniques, theoretical analyses, ‘and rational design practices must all come into play to assure adequacy, ‘Three tasks are generally involved in producing @ ship with good controllability: “Ga)_ Establishing realistic specifications and eriteria for coursekeeping, maneuvering, and speed changing. @)_ Designing the hull, control surfaces, appen: ages, steering gear, and control systems to meet these requirements and predicting the resultant per formance, “@) , Conducting full-scale trials to measure perform: ‘ance for comparison with required criteria and predic tions, “This chapter will deal with each of these three tasks in detail 12 Goal and Organization of the Chopter. The goal of this chapter isto introduce the basies of eon- {rollablity analysis and its many facets in a manner that will lead to the use of rational design procedures to assure adequate ship controllability ‘The chapter is organized to provide en understand: ing of controllability and for Influeneing it while in teracting with the design of the vessel's hull, machinery, and other features. The material is ar- ‘ranged to provide a progression of information start ‘ng with general prineples (Sections 2 through 11 on theory and analysis), the influences of factors such as environment and channel restrictions (Sections 12 and 18), the hydrodynamics of control surfaces (Section 14), and trials and performance requirements (Section 18) ‘The final Sections (16 and 17) provide an introduction tothe application of maneuverability analysis tools and methods to the design of the ship and its appendages Tor satisfactory control by the helmsman and autopilot. ‘The design-oriented naval architect should find Sec- tions 18-17 mort help2 Section PRINCIPLES OF NAVAL ARCHITECTURE nd The Control Loop and Basic Equations of Motion 21 The Control Loop. For surface ships, course: keeping, speed changing, and maneuvering involve pri manly’ forees, moments, and motions acting in all directions in the horizontal plane. For submarines, the third dimension also comes into play. Hydrodynamic ‘motion forees and interactions acting on the vessel's Doll, rudder, and other appendages are of first com sideration and difficulty. However, it important to recognize that the responses of large number of other mechanical, electronic, environmental and, most importantly, human factors all influence contrellail ity. The following discussion of controllability based ‘on the concept of control loops illustrates the role of ‘these many factors. Consider first the closed-loop directional control sy tem from Segel (1960) shown in Fig. 1. Starting at the left of the figure, there is a desired path or tra jectory that the ship's conning officer wants to follow either under conditions of steady steaming at sea or in maneuvering. In an idealistic case, the desired path would be displayed for use by the helmsman (or al ternatively by an autopilot). Simultaneously, again for the idealistic ease, the path actually being traversed ‘would also be shown on the display. If these two paths do not coincide, corrective action is taken by the helms: ‘man or autopilot by changing the helm in a direction that will tend to correct the path error, This action activates the steering gear which changes the rudder Position which in turn exerts a control foree on the Ship. This contrl force acts to induce on the hull an angle of attack, an angular velocity, and other motions ‘These motions of the hull introduce the major hydro dynamic forees and moments that effect the change in heading and path, In addition to control and hydrodynamic forees and ‘moments, external disturbances such as the wind, eur- rent, or waves may also be simultaneously acting on the ship. Ideally, che resultant heading and actval path coordinates ean ‘then be fed continuously baek to the ‘helmsman’s display. This last step closes the control loop. Complete references ae listed at the en ofthe chapter, In the real case all of th information concerning the actual instantaneous path i rarely known, On most ships only the heading and sometimes the rate of tam fare continuously determinable, While the cooelinates Of position are available only ecasionally. In spite of the shortcomings of the real cas, the contro loop stl functions as shown in Fig. 1 but with less than com plete information available co the conning officer and Felmeman (or autopilot). Relatively large-scale Changes in postion are determined or deduced in re Strieted water conditions from visual observations, Fadae digplay, or from visual ross beatings. Modern leetronie navigation systems such as Loran C and Global Positioning System (GPS) also offer & means for determining large seale changes tn positon. Consideration of the control loop of Fig. 1 shows that each ofits elements plays a vital roe in the overall controllability of the ship. The ast two elements of the loop, the "ship" and the “steering gear snd rad den” are of the greatest concern to the naval architec although the human factors present must be contin. tally reviewed to develop 4 succesful design ‘Whereas the directional contra loop of Fig. 1 fane tions to determine the path, a second loop of interest, the speed control oop, functions to determine the Speed along the path. The only common link between the two loops is the conning officer, who issues the drders in both cases, Inthe eae of the speed conte Toop, an operator oF engineer receives the orders of the conning officer and manipulates the power outpat and direction of rotation of the main propulsion ma chinery to maintain, accelerate, slow down, sop, oF reverse the speed of the ship. Inthe open sen where decisions ean be made in a more leisurely manner the conning officer has ample time to issue tho necessary ders to both control loops. In restricted snd cor. ested waters, however, orders to both control loops tray have to be issued simultaneously, With develop tment of automation, the integration of these loops and the elimination ofthe intermediate roles of the helm man onthe bridge and the operator inthe engine room i becoming commonpiace. * With today’s sophisticated dil rigs and track keep ing Vessel, the tatomati controller has indeed be- [omen fe fief a eee ae ce one fe i a J ICONTROLLABILITY posmow or 8 oF BaP Ae i. 2 Otani es jn ne a 1948) come quite advanced. Heading and speed, plus transverse position error and fore and aft position er- For are used to compute vector thrasts required of the ‘various foree effectors (propellers, rudders, thrusters, tte), and proper distribution of eovrection forees/mo- nents is automatically ordered, 22 Axes Fixed Relative to the Earth. The basic dynamics of maneuvering and coursekeeping can be described and analyzed using Newton's equations of ‘motion. Basi equations in the horzontal plane ean be considered first with reference to cne set of axes fixed relative to the earth and a second set fixed relative to the ship. Fig. 2 shows typical fixed and moving axes for a surface ship. The path is usually defined as the tra- jectory of the ships center of gravity. Heading refers to the direction (, angle of yaw) of the ship's longi tudinal axis with respect to one of the fixed axes. The difference between the heading ard the actual course {or direction of the velocity vector at the center of gravity) i the drift or leeway angle, 8. When the ship [8 moving along a curved path, the drift angle is thus the difference in direction between the heading and the tangent to the path of the certer of gravity. "There are significant factors that couple the speed of a ship and its path. For example, itis shown later that path changing (turning) and even path keeping (coursekeeping) cause involuntary speed reductions ‘These effects arse from the fact that any misalign ment between the x-axis of the stip as shown in Fig 2'and its velocity vector, V, inereises the drag force ‘acting on the ship. In addition, on multipleserew ves Sels the thrust produced by each propeller can be con trolled individually so as to influence the path as well ‘the speed. On ships with an odd number of propel: lers, or with any aumber of unirotating propellers, the direction of rotation of the uneompensated-for propel: ler() influences the path. Even on ships where the trajectory eontrol and speed contol loops operate in dependently, changes in the direcion of propeller ro- tation are commonly used in conjunetion with rudder setion to control the motion in restricted waters. Fig. 2 shows a righthand orthogonal system of ret 193, ference axes zs and y, whose direetions are fixed with Fespect to the surface of the earth. For surface ships in-caim water, iis obvious that the path of the center fof gravity is restrained at all times to the horizontal plane 249 The positive direction ofthe z-axis is taken to be in the general direction of the motion; ts precise direction s arbitrary, but is fixed with respect to the arth, Positive 2 taken downward or into the plane bf the paper, positive y, is to starboard. The motion Of the ship subsequent to time, £ = to i completely defined by the coordinates yo, Yoo and the angle of yaw, J. Sines in Pig. 2, Zs positive downward into the paper, the sign of shown in Fig. 2 is negative. Referring to these axes fixed in the earth, the New: tonian equations of motion of the ship are Xo= Abe Cured) Y,= Pec Sway) o N=1L% (aw) where the two dots above the symbols yor yoo and indicate the second derivatives of those values with respect to time, f, and, X;and ¥, = total forces in x, and yo directions respee- tively a = mass of ship N= total moment about an axis through the enter af gravity of Ship and para! to ‘mass moment of inertia of ship about the = yaw angle in the horizontal plane meas- tured from the vertical, zy plane to the ‘axis of the ship. 2.9. AxesFixedin the Ship. Inspite of the apparent simplicity of Equations (0), the motion of a ship is more Conveniently expressed when referred to the axes 2° fand y fixed with respect to the moving ship as also Shown in Fig. 2. The moving. axes, like the fixed axes Zeand yo, form a righthand orthogonal system, but ‘withthe difference that the origin stays atthe center of gravity for all time, ¢. The axis is along the cet terplane, eoineident with the longitudinal axis of in trtia which may be assumed, with very small eror, to be parallel to the baseline ‘of the ship. Its positive Gireetion is forward, The direction of the axis is re ferred to as the heading; hence yis the heading angle fas well at the yaw angle. The zaxis is also in the fenterplane of the ship, but is normal to x and is pos itive downward; the y-axis is normal to x and z and is positive to starboard. The instantaneous linear velocity Df the origin of the moving axes is represented by the vector V and the orientation of the moving axes with Fespect tothe direction of motion is given by the angle, B, the so-called drift angle or angle of attack measured from V to as shown in Fig. 2 The velocity V is, of course, always tangent to the path of the ship, The194 orientation ofthe moving axes with repect to the x, sy axes i, as noted earlier, the angle of yaw v. Tn he Particular case shown in Fig. 2 both and W are negative, frorde wo convert Equations (1) fom axes fixed in the'earth to axes fixed in the moving si, the otal forces X and Yin the and gdieetons, respectively, are exprested in terms of Xan ¥-: X= Xycosy + Yosing Y= Yeon y — Xosinw ® Thewise bag = wos y— vin y Yoo = using + veosy ® whore the dot above the symbol signifies che fest derseative ofthe quantity with respect to time, and and'v are the components of V slong £ and, respec tive. Then Bro= Weasy— 6 sin (using + rem Hom Wsing + Fco8b+ (wooed vsinvd — Substituting Eguation (in Equation (1) and inserting the resulting values of X, and Yn Bquaion (2) ye the simple expressions X= Au ~ 09) Yeates tua) ‘These and th third member of Equations (1) comprise the pertinent equations of motion in the hornet Blane assuming sero rol pith, and heave, Rol itch, fue yaw are the rotary ship motions about the z, dnd axes, respectively. Surge, sway and heave af the eranslatory motions slong the same axes. For eam pletenens: X= au — vp) is surge Ye aie + ui) issway ° Na1 is yaw Note the existence of the term Aue in the equation for ¥ and Ay in the equation for X, whereas terms like these were not present in Equations (1). These are the socalled centrifugal-force terms which exist when systems with moving axes are eonsidered, but do not exist when the axes are fixed in the earth, ‘Equations (6) have been developed for the case where the origin of the axes, 0, is at the center of -Rravity of the ship. For many reasons, itis frequently Aesirable to locate the origin not at the center of seravity but rather at the intersection of planes of Symmetry. For example, for body-of revolution subs marines, if Os located at the axis of symmetry rather than at the vertical position of the center of gravity, important simplifications ace achieved. For surface ships, locating the origin at the midlength rather than PRINCIPLES OF NAVAL ARCHITECTURE at the longitudinal position of the center of gravity is Aesirable for two reasons; one is to simplify certain Computations, the other is thatthe location of the een ter of gravity is not constant but changes with the condition of loading. (Throughout this chapter it is ‘assumed that LCG coincides with LCB and with the ‘midship location) 24. Forces Acting on @ Vessel During @ Moneu- ver. The forces and moments (left hand side) of the equations of motion (6) are built up of four types of forces that act on a ship during a maneuver: (0) Hydrodynamic forees acting on the hull and appendages due to ship velocity and acceleration, rud ‘der deflection, and propeller rotation. (©). Inertial reaction forces caused by ship accel eration, (c) Bavironmental forces due to wind, waves and (@) External forees such as tugs or theusters ‘The first two types of forces generally actin the horizontal plane and involve only surge, sway and yay responses, although rolling effects (hee!) occur in the maneuvering of high-speed ships and the Small Wa- {erplane Area Twin Hull (SWATH) vessels. Hydro- Alynamie forees fall into two basic categories, those arising from hull velocity through the water (damping forces) and those arising. from hull accelerations through the water (added mass forces). The ship ae cclerations produced by these and any external forces result in balancing inertial reaction (d'Alembert forees ‘and moments), especially when turning. ‘The effect of a radder on turning is indirect. Moving the rudder produces a moment that causes the ship to change heading s0 as to assume an angle of attack (ceway angle) to the direction of motion of the center of gravity. Consequently, hydrodynami¢ forees on the hull are generated whieh, ater a time, cause a change of lateral movement of the center of gravity. The lat: ‘eral movement is opposed by the inertial reactions. If the rudder remains ata fixed porition, a steady Urn condition will evolve when hydrodynamic and inertial forces and moments come into balance ‘When in shallow or restricted waters, various com plex effects come into play. Interactions between ves sels further complicate hydrodynamie and inertial Toree analysis. Section 18 describes how some ofthese complications have heen treated ‘The ship may also be operating and maneuvering in fn environment where wind, waves, and current are present. The effect of current is usually incorporated with the hydrodynamic forees by eonsidering the re ative velocity between the vessel and the water al though studies in restricted waters require more careful analysis, Wind and wave forces are generally ‘reated as external forees as deseribed in Section 12 ‘Wind velocity i generally unsteady and hence foreesCONTROWABILTY and moments due to wind will be time dependent. ‘These forces are generally proportional to the above water area of the ship and the square of the relative velocity between the ship and the wind. Forces and ‘moments also vary with the direction of the wind ve Tocity relative to the ships axes, “Two distinet types of wave forces act, The steady and slowly varying forees due to secondorder wave ‘drift effects are generally nore important for ship ‘controllability than the firsterder forces, which are of Primary importance for seabeeping, as deserved in Chapter VIII. However, the latter can be important for the case of following seas where frequency of ‘encounter is small, Wave drift forces depend primarily 95 ‘on ship length and on the relative magnitudes of wave length and amplitude. Pitching motion changes the shape of the immersed dru and ean therefore have signifieant effects on the coefficients in the equations of motion, particularly in ‘quartering and following seas. Finally, tugs and thrusters ereate effective forces when utilized at relatively slow speeds. The forees they fevelop are for the most part external to the hydro. ‘dynamies of the maneuver and are normally treated af independent additions “The simple case of controllability, assuming a calm ‘open sea without wind, waves, current, and external forces, will be considered first. Section 3 Motion Stability and Linear Equations 3.1 Definitions of Motion Stability. The concept of path keeping is strongly related to the concept of ‘course stability or stability of direction, A body is said to be stable in any particular state of equilibrium in rest or mation if, when momentarily disturbed by an external force of moment, it tends to return, after lease from the disturbing fere, to the state of equ librium existing before the body was disturbed. In the case of path Keeping, the mast obvious external di turbing foree would be a wave or @ gust of wind. For optimum path Keeping, it wosld be desirable for the ship to resume its original path after passage of the disturbance, with no intervention by the helmsman ‘Whether this wil happen depends on the kindof motion stability possessed ‘The various kinds of motion stability associated with ships are classifed by the attributes of their intial State of equilibrium that are retained inthe final path Of their centers of gravity. For example, in each of the cases in Fig. 3, a ship is initially assumed to be traveling at constant speed along a straight path. In Case I, termed straightline ce dynamie stability, the final path after release from a disturbance retains the Straightline attribute of the initial state of equib- lum, but not its direction. In Case II, directional sta Uility, the final path after release from a disturbance retains not only the straightlie attribute ofthe initial path, but also its direction. Case II is similar to Case Texcept that the ship does not oscillate after the disturbance, but passes smooth tothe same final path 1s Case I. The distinction betaeen these two cases is discussed in Section 4. Finally in Case IV, positional ‘motion stability, the ship returns to the original path ie: the final path not only has the same direction as the original path, but alsa its same transverse position relative to the surface of the earth ‘The foregoing kinds of stability form an ascending hierarchy. Achieving straight-line stability (Case 1) is the designer's usual goal for most ships when steered by hand. The other cases require various degrees of automatic contro 32 Course Stability With Controls Fixed and Controls Working. Allof these kinds of stability have meaning ‘conirl surfaces (rudders) fixed at zero, with con surfaces free to swing, or with controls either lly oF automatically operated. The first two nana arnt SE FO pacers spciry ork SREB tEanona IV rostona. worn saa8iTy SE mocaresmatewrancous onronsance |cases involve only the last-two elements of the control Joop of Fig. 1, whereas the last eaze involves all of the elements of the control loop. In normal marine usage the term stability usually implies controls-fxed sta bility; however, the term ean also have meaning with the controls working. The following examples indicate distinctions a) A surface ship sailing # calm sea possesses positional motion stability in the vertical plane (and therefore directional and straight-line stability in this plane) with controls fixed. This is an example of the kind of stability shown by Case IV of Fig. 8. In this case, hydrostatic forces “and -moments introduce a tunique kind of stability which in the absence of these forces could only be introduced either by very sophis- tieated automatic controls oF by manual eontrol. The fact that the ship operator and designer can take for ‘granted this remarkable kind of stability does not de- {act from sts intrinsic importance, (@) In the horizontal plane in the open sea with stern propulsion, a self-propelled ship cannot possess either positional or directional stability with controls fixed because the changes in buoyancy that stabilize in the vertical plane are nonexistent in the horizontal plane: However, a ship must posses both of these nds of stability with controls working either under ‘manual or automatic guidance. Possible exceptions in ‘clude sailing vessels, some muiti-hull ships, and foil or planing eraft but not other surface effect ships. (0) ‘The only kind of motion stability possible with self-propelled ships in the horizontal plane with eon- ‘ols fixed is straightine stability. In fact, many ships do not possess it. In subsequent sections of this chap- ter, with some exceptions, whenever controls-fxed sta- billy is mentioned, the intended meaning is controls fixed straightine ‘stability. This kind of stability is desirable, but not mandatory. With each of the kinds of controlsfixed stability, there is associated a numerical index whieh by its sign designates whether the body is stable or unstable in that particular sense and hy its magnitude designates the degree of stability or instability. To show how these indexes are determined, one must resort to the difer: ential equations of motion. 33 Assumptions of Linearity and Simple Addable Ports. In order to understand the impact of various ship design charactersties and features on ship con- trollabilty, itis necessary to first become familiar with certain fundamental aspects relating to the concept of stability and to the development and use of the linear ‘equations of motion. The use of non-linear equations for analysis and predietion and the determination of coefficients through captive model tests, use of theo- retical and empirical coefficient determination methods, and systems analysis is introduced in See: tions # and 9 ‘The force components X, Y and the moment com: ponent Vin Equation (5) are assumed to be composed PRINCIPLES OF NAVAL ARCHITECTURE of several parts, some of which are funstions of the Velocities and accelerations of the ship. In the most ‘general ease they also include terms dependent on the Grientation of the ship relative to the axis of the earth as wel as excitation terms such as those arising from the seaway or from use of the rudder, tut these will be introduced later. For the present they are assumed to be composed only of forces and moments arising from motions of the ship which in tum have been excited by disturbances whose details we need not be ‘concerned with here. Expressed functionally X, Y, and Nave: X= Aya a, 4D y oD 6 N= Flynt bw) In order to obtain a numerical index of motion sta bility, the functional expressions shown in Equation (6) must be reduced to useful mathematical form. This ‘ean be done by means of the Taylor expansion of a function of several variables. The Taylor expansion of a function of a single variable states that If the func tion of a variable, 2, Fig. 4, and all its derivatives are continuous ata particular value of x say 2,, then the ‘Value of the funetion ata Value of 2 not far removed from x, ean be expressed as follows: Flu a a, “ee MO), 82" ef) Fie) = fe) + 80 FO 6 | Be afl) Bafa) ade wa where ‘fle) = value of function ata value of x close to ‘fley) = vale of function at 2 = 2, bene a, and afte) “LE nth derivative of funetion evaluated at 2 = 2, If the change in the variable, ar is made sufficiently small, the higher order terms of & in Equation (7a) an be neglected. Equation (7a) then rediees to = fay + eV) Fee) = fle) + 82S It may be seen from Fig. 4 that Equation (70) is & linear approximation to the real function fl) at 2 = x, + Be and that (76) becomes increasingly accurate ‘as Be is reduced in magnitude, Equation (70) i called the linearized form of (a) "The linearized form of the Taylor expansion of & co)CONTROWABIUTY raped al 4 itn Tree eto 9 se vr function of two variables 2 and y is a simple sum of three linear terms as follows: faa MED yoy MEP go where both 6y and By must be small enough so that higher order terms of each can be neglected as well as the produet &rby. ‘The assumption that renders linearization reason- ably accurate, namely, that the admissible change in variables must be very small, is entirely compatible With an investigation of motion stability. Motion st bility determines whether a very small perturbation from an initial equilibrium position is gong to increase with time oF decay with time. Thus, itis consistent ‘vith the physial reality of motion stability to use the Tnearized Taylor expansions in connection with equa tion (6. For example, by analogy with Equation (70), the linearized V'foree of (6) ean now be written as: ay ley yw) + Be Y= Fy, My thy Po by GY +m) x a tee w- mF w-wyF @ ‘where the subscript 1 refers in all cases to the values Of the variables atthe initial equilibrium condition and ‘where all of the partial derivatives are evaluated at the equilibrium condition. Since the intial equilibrium condition for an investigation of motion stability is straightdine motion at constant speed, it fllows that ty = 8, = J, = 9, = O. Furthermore, since most ‘hips are symmetrical bout their ze plane, they travel ina straight line at 2ero angle of attack; therefore 0, is also 20r0 but this is not necessarily true on ships ‘with an odd numberof propellers or with any number of unirotating propellers, Sections 5.3 and 11.0), Also Decause of symmetry A¥/au = d¥/0u ~ 0 since a 17 change in forward velocity or forward acceleration will produce no transverse free with ship frm that re Simmetieal about the plane. Pinal i the ship in Tuc equilibrium nsteaghtline moan there ean ino Voree acting ont n that conan, therefore Fay tet by Gan Vs) alo sera. Onl 8 not zero but is equal to the resultant velocity, V, in the intial cqlibeum condom, Wah hese sinh cations, Equation (9 reduces to vr yahoo and similarly the sur ‘ment ean be written as avy oye Fe O00 Rite oO ng foree and the yawing: mo- AY AX gy 4 A, g OX xe Hue Bore ax + ey eo) an owe an wae Me MENG a an ay aa” a where the cross-coupled derivatives BY, 31), 9¥/04., N70, and N70 usually have small onzero values because most ships are not symmetrical about the yz plane even if that plane is at the midlength of the ship (Gow and stern shapes are normally quite different) However, the cross-coupled derivatives aX 0, 04781, X78, and OX/ 00, like BY/0u and AY 794 are zero because of symmetry about the re-plane. Hence, equa: tion (8b) reduces to: ax + row . roi A Siingee eeeeeet oasis mater at poe jin etc HKfu a) (a Xu =o Every term of the first two equations of (1) has the dimenslons of a force whereas every term in the ied Squat of (0) has the dimensions of a moment. ‘Therefore, to nondimensionaize (10), which s conve hen for several reasons, the force equations are dk ‘ied through by p/2)L'W" and the momentequations Ey @/dU¥*, Note the similarity between p/2).°V" hel asa nondimensionalizerin this case and(p/2.SV" Yet — Ye — (= aur Hr a0 Ne Nr + Nemo198 used to obtain the resistance coefiionts in Chapter V), Further, asin (Nemenelature, 1952) a primed sym bol will be used to designate the nondimensional form of each of the factors appearing in (10), For example: ’ 2 bu bn ‘Thus, for example, te nondimensional forms of the first couple of terms af the last of Equations (10) are: ee If the surge equation is neglected and ifthe previous lation adopted, 10) bese In nondimensionl -Yor += Poe — 0, = ane “YF oN Ni oN +r = Nye = 0 where the main diferonce between (10) and (11), aside from the prime notation, is that 1 has disappeared sinee u,/V = 1 for smal perturbations “Because of the fact that the derivative ¥", enters into Bquation (11) as an addition to the mass'term, it is termed the virtual mass coefficient. (The term J”, is always negative; io, Y acts to oppose positive ‘see Section 42) Itis thus identical to the concept of ‘added mass. (The forve required to accelerate a body ina fluids always larger than the product ofthe actual mass of the body times its acceleration. This fact izven rise to the concept of “entrained” or “added! mass. However, this added foree should be really in an PRINCIPLES OF NAVAL ARCHITECTURE terpreted as the hydrodynamic foree arising because of the acceleration of the body in the uid, This is precisely the definition of the ¥, force in (10). Sim Mary, A, is termed the virtual moment of inertia ‘eoeficint! The derivatives ¥", and N", are termed ‘coupled virtual inertia coefiients. As noted earlier, these derivatives would be zero if ship hulls, including their appendages, were symmetrical about their y2 planes Tis convenient to use a notation that distinguishes the forces and moments according to their origins. For ‘example the notation ¥,v wil be used to denote the {component of the hydrodynamic force acting atthe Senter of gravity of the ship that is developed as a Fesult of an angle of attack, B. As has been shown, Y-v is only a linear approximation to, or linearization of, this Vforee a8 will be further evident from later ‘examination of Fig. 10, Similar symbols and definitions fre included in the nomenclature for other forces and 2.5, Control Forces ond Moments. It is important tonote that all of the terms of (10) or (01) must include the elect of the ship's udder held at zero degrees (on the centerline), On the other hand, f we want to com Sider the path of « ship with controls working, the {equations of moti (10) or (11) must include terms on the righthand side expressing the control forees and ‘moments created by rudder deflection (and any other ‘control devices) as functions of time, The linearized y component of the force created by rudder deflection lating at the center of gravity of the ship is ¥, By (soe Fig. 5) and the linearized component of the moment created by rudder defection about the z-axis of the ship is Ny By whereCONTROWABIITY by = rudderleflection angle, measured from zz ‘plane of ship to plane of rudder; positive ‘efleetion corresponds to a turn to port for rudders) located at stern linearized derivatives of Y and NV with re ‘pect to rudder-dflection angle 6 ‘The lateral force from the devested rudder thus ere- ates a moment t turn the shp. This turning. action fauses the ship to develop an angle of attack with tespeet to its motion through the water. The lateral {forces then generated by the well designed ship (acting. as. foil moving in lquld at anangle of attack) ereste 2 moment, Nt, that greadly augments the rudder moment. The combined moments cause the turning. ‘mation as indicated in Fig. 5. Por the case of simall perturbations, which is the ‘only ease where (10) and (11) apply, only small detec- tions of the rudder are admiscible, With this resric- tion, the derivatives such as Y",,N',, Y",,and.N’ are ‘evaluated at 8, — 0 and are aisumed fot to change at other admissible values of By. Furthermore, for usual ship configurations, ¥", = 0 and N’, 99 With these assumptions the equations of mation in- cluding the rudder force and moment, are as follows: Moment: Moments yy — Nye — Nr = Nabe 2) Force: a3 — Ya" -(¥,— ar (Sway) =P, Be where: nyo, Nw 2h alsa onze It will be shown in the next three sections how the linearized equations developed s0 far can be used to analyze the problem of course stability and steady turning. But to make numerical predictions it is nee essary to obtain values for some or all of the eoefi- tents or derivatives involved. This Is primarily done bby means of captive model tests, as discussed in Sec tions §,9 and 16. Theoretical approaches to estimating: some coeficients and approaches are also described in Section 16, Section 4 Analysis of Coursekeeping and Controls-Fixed Stability 4.1 sib Indes, Using ons inca arms, so lutions to the Sway and yaw equations provide linear transfer functions permitting the review of the st bility of motion. This Section develops the base sta Dility indexes ‘and describes the definitive. spiral ‘maneuver whose numerical measures are indicative of the stability characterises ofa ship. ‘Equations (12) are two simultaneous diferental equations of the fist order in two unknowns, the hor- {Rontal-velocty component, © and the yaw angular velocity component 7". The simultaneous solution of these two equations for» andr yields a secondoder ‘ifferental equation which leads to the concept of Straightline stability. The soluions for 0" and r" cor respond to the standard solutions of seoond-order di ferential equations which areas follows Vert + Vere Reenter where ¢ = 2.718 Vv, Vs Ry and Ry ae constants of Integration; @, and are the stability indexes with dimensions of 1/tand #s time Lis seen from solution (05) that if oth values of o are negative, o” and” will approach zero with increasing time which means ‘thatthe path of the ship will eventually resume a new Straightline direction. This conresponds to Case [ot Fig. 2 If either oy or oy is positive, wand »* wll a3) increase with increasing time, a straight-line path will never be resumed, and the ship may end up ina steady turn with its rudder held fixed at zero. ‘The relationship between the stability indexes, of solution (13) and the stability derivatives of (11) ean ‘he obtained by substituting the solutions (13) back into (1). Tf this is done, a quadratic equation in a is ob- tained: Ao! + Bo + C= 0 0 where An way Ba — way c= rN, 00M, ‘The two roots of Equation (14), both of which must bbe negative for controls-fxed stability are: =B/A * ((B/A¥ ~ C/A} a= eee From practical standpoint, o, alone is usually given for surface ships. This is because the size for (118 algebraically less than oy and thus itis clear {rom Equations (18) that the motion desription by the term is larger than the oy term after the distur. (usa)200 bance has ended. Hence e, alone is rather a good negative quantitative measure of the degree of sta- bi. 42. The Stability Ceitesion. Expression (Ida) re ‘veal thatthe to essential conditins for both er, and 17 to be negative are a) ‘That C/A be positive, ie, C/A > 0.1 C/A ite ais willbe greater than B/A and whether B/A is positive Or negative, one value of o will always be positive. (0) That B/A be positive. If B/A Is negative and C/A is positive, then both er and o will always be positive. As noted before, if both B/A and C/A are hegative, one value of o will be postive, ‘Thus, the conditions for stability are reduced to the requirements that 8/4 and C/A must bath be positive ‘quantities. Since each ofthe terms of 4, B, and Care nondimensionalized by the same quantities, the mag- nitudes and the signs of A, B, C'nay be determined by examination of either the dimensional oF nondi mensional derivatives appearing in the definitions of A,B, and C in Equation (14) Because the nondimen- Sionalizing items, p, V, and L are always taken as Positive, they do not ehange the signs of the deriva tives. “The derivative Y, willbe treated frst. Its the slope of the Y-foree with respect to an seceieration ¥, and ‘appears in the definition of both A and B of (14). The sketch shown in Fig. 6 represents a ship with an ac saute neeanve 7 Tyla Ys enn PRINCIPLES OF NAVAL ARCHITECTURE celeration + 8 withthe origin taken at the midlength ff the ship. Under these eieumstances, both the Bow and stern experience a acceleration in the positive y direction, Therefore, the inertial reaction pressure of the water being accelerated by the hull produces forces in the negative y-direction on both the bow and stern, Hence, the bow and the stern effects add to give a relatively large negative ¥-force resulting from a poe itive ». Ia disturbance of a negative 7 placed on the ship, the inertial pressures on the bow and the stern will add together to give a relatively large ¥- foree in the positive yedirection, Hence, the plot of ¥ versus. would appear as shown in Fig. 7, and the slope ¥, taken at 5 = 0 would be a negative value of relatively large magnitude. Bor shipshaped bodies ‘with large length to beam (7B) ratios, the magnitude of ¥, is approximately that of the ship's displacement, ‘3. Por example, theoretically ealeulated values of Y, {or elipsoids give values of ~0.9 8 for L/B = 5: 095 8 for L/B = 85: and — 10 for L/B = =, ‘Thus, the term (4 — ¥,) which ogeurs in both A and Bis & large positive number with a magnitude of al most 2.4 Equation (12) Like ¥,, the derivative N, appears in both and 8, and is also always negative and relatively large as will bbe shown in Fig. 8. A sample plot of N versus follows the same relationship as ¥ versus # in Fig. 7 Just as the magnitude of Y, is almost as large as & for shipike forms, the magnitude of WN, is almost as large a8 1, Theoretical calealations for elipsoids show, that Ny = 0. for L/B = 5 N, = ~O8 7, for L/ B= 85; and N, =~ 1.01, for L/B = we. 'As waa indicated in the analysis of the derivative Y,, both the bow and the stern add to contribute toa large negative ¥,. However, in the ease of N, again the bow and ‘stern oppose each other and N,, like Y,, is usually a relatively small quantity of uncertain sign provided the origin is taken close to the ship midlength “he sgn td magnitude of A may now be deter Aan ay, WN, a Foo Fa 8 Sip wih angi actionCONTROLLABILITY 201 ‘To evaluate the relative magnitudes of B and C, itis necessary to examine the nature of the derivatives ¥., No, Yo, and N,. In Fig. 9, the nature of the forees acting’ on a body with & velocity » added to a forward velocity w is shown. Its seen that as a recut of the angle of attack, = —1/V on the body, both the bow land the stern experience a lift force oppositely directed to. Hence, ¥ is always negative. However, the bow contribution to the total Y,rforee is usualy larger than that of the stern so that the center of setion of the total force in the y-direction owing to vis consid erably forward of the midlength of the ship. Hence, with the origin at the midlength, Nis also usually a negative quantity for ships without fins or rudders. It is obvious that the addition of a rudder atthe stem ‘of a ship, for example, wll increase the magnitude of (Yst)ung and hence decrease the negative magnitnde of MwvTf the rudder were sufficiently large, it might ‘even cause N, to become positive; however, ‘his isnot usually the case. A typieal plot of Y verus v and possible plots of N versus v with the origin at QD are Shown in Pig. 10 Tn analyzing the effect of an angular velocity v on Vand Ny location B forward and S aft are assumed as shown in Fig. 11. The origin is again taken at the mnidlength. When the ship is moving ahead with a ve locity Vand an angular velocity ++ is added, point Brat the bow has an angle of attack from starboard ‘rd,/V for small r) producing a negative Y-force ‘and a negative N-momenton the bow. Similerly, point ‘Sat the stern experiences an angle of attack from the port side producing a postive ¥-force at the stern and 4 negative Nmorent. Hence, the bow and the stern ‘add to give a large negative V fora positive y, whereas ‘bow and stern oppose each other to give either a small positive or negative Y-force for a postive r, negative IE the bow dominates. For a negative r, the angles of attack change to opposite sides and hence the force and moment contributions change sign. Sample curves of ¥'versus rand N’ versus 7 for O at (} are shown. in Pig 12. Since B like A is always a large positive quantity {or ships, independent of the choice of origin, the con- dition for stability reduces from B > Onl C > 0 toonly C > 0. stem a aoe Vereaoe \e ge Fe 10Tyid Y ve vo atv ap, Hence, Cis considered the discriminant of dynamic stability. From Equation (10), C= FN, 0 ase, YN > NY, ~ a4) (uaa) with the inequality of Equation (Ide) expressed as x i ¥-8 ¥, the solution can be viewed as a relationship between the lever arm of forees due to yawing and sway. The Inequalities of Equation (Ide) and (Idd) provide the haste criterion for dynamic course stability, indicating whether or not the ship is stable but not giving’ ‘Quantitative measure—as do the os of (13) and (1d). Methods for estimating stability (o,) early in a de sign effort based only on major dimensions have been developed by Clark (1982) based on regression anal- yses and are mentioned in Section 16.8. Another closely elated index T (approximately equal to 1/7) of the K and T pair of indexes is introduced later in Section 54. Tand K can be developed from common trials and are useful in comparing vessels. >0 (a4e)202 vy HON Doheares \, ” yen 0 Seas Fig 12 Typ Ysa and ven lan. ‘The stability indexes in the horizontal plane are not speed dependant in the range of low and moderate Froude numbers where the resistance coeficient, Cy, is essentially constant, sinee the nondimensionalized stability derivatives are sensibly constant in this speed range.” Thus, if a ship possesses controle-ixed, straightline stability in the horizontal plane at low speeds, it will alto be stable in this sense at higher speeds, at least tp to the limiting Froude number to ch the nondimensional mation stability derivatives remain constant; or vice versa, if a ship is unstable at Tow speeds, it will also be unsiable at higher specds.* 443. The Dieudenne Spiral Maneuver. The direct or Dieudonne’ spiral maneuver is a defiitive ship trial (Dieudonne’, 1959 which identifies the directional sta bility characteristics of the vessel, The maneuver con: sists of the following: (a) The ship it “steadied” on a straight course at 4 preselected speed and held on this course and speed {or about 1 min, Onee a steady speed is established, the power plant controls are not manipulated for the duration of the maneuver. (@) After aboot 1 min, the rudder is turned to an angle, Bp, of about 18 dey and held until the rate of change of yaw angle maintains a constant value for ‘about L min. (The rudder angle is then decreased by a small amount (about 5 ceg) and held fied again until a new Value of is achieved and is constant for several minutes, (The foregeing procedure is repeated for differ ent rudder angles changed by small increments from, Say, large starboard values to large port values and Dack again to large starboard values. ‘The numerical measures obtained from the preced: Sf nc, hi ay ur conan pera agro {Nats sp'may prove wh inerearng i pee, However ‘be preven of entining datrhanes a presaded nthe pretnk Sele ke Sion way vate fining a tteesw decunted nSecion 2. PRINCIPLES OF NAVAL ARCHITECTURE ‘ng spiral maneuver are the steady yaving rates as a Tunetion of rudder angle, A plot of these values is indicative of the stability characteristics of a ship. For example, if the plot is single line going from star board rdder to port and back again, as shown for ship in Fig. 18, the ship possesses contros-ixed, Straightline stability; that is, it has @ negative direc: tional stability index. If, however, the plot consists of two branches Joined together to form a “hysteresis” loop, a shown in ship Bof Fig. 18, the ship is unstable; thai, it has a postive stability index. In addition, the height and width of the loop are numerical measires of the degree of instability; the larger the loop the ‘more unstable the ship, The slope ofthe yaw-rate curve at zero rudder angle is @ measure of the degree of stability or instability, It may be predicted using Equa tion (26) or measured from plotted ship trial data However, the linear theory used to develop (26) ie tunable to predict the characteristics of the hysteresis loop for unstable ships. For this purpose the ronlineae theory of Section 8 is essential. Nevertheless, itis clear that @ ship with a hysteresis loop must have’ positive controHfixed stability index. The existence of a loop ‘means that with the rudder fixed at zero, u's not necessarily zero; that is, the ship may continue to turn With the rudder amidships. The mere possibility of this, ‘with no external disturbances acting or having acted inthe recent past, testifies to eontrols-fixed instability. Abkowitz (1964) has pointed out the analogy be tween stability in straight-line motion and stability in heel. Fig. 14 is a plot of the conventional righting moment curve versus the angle of heel, @, for both Stable and unstable ships, The resemblance between Figs, 18 and 14 is apparent immediately. For ship A of Fig. 14, which is stable in heel, the slope of the righting-moment versus heel angle curve at. — 0 is positive, indicating stability. Similarly, in Fig: 18, the Positive slope (as drawn) of y versus &» for ship A Indicates that that ship i stable in straight line motion, Likewise for the unstable ships B in Figs. 13 and 1 ‘The slope at the origin of both curves is negative, indicating instability In the case of the stable ship A in Fig, 18 only one Angular velocity (or turning rate) can result from any igiven rudder deflection, Por the unstable ship Bin Fig. 13, there are regions between the lines aa, and 0b, where there is more than one turning angular velocity Tor a given rudder deflection. For the ship unstable ip straight line motion, there isa region where the shi can tuen against its rudder. During a spiral test, no data can be obtained for the unstable ship B on the dotted curve between (a) and (}) in Fig. 13, because these are points of unstable ‘equilibrium for the particular rudder angles. For ex ample, at ero rudder angle, the ship will tend to move to either positions (e) or (e,) depending on the previous time history of the motion because these are positionsCONTROLLABILITY sore (34) y.0 7] Nee. 5 ‘ 8 i ‘\ q ede of stable equim for zero rudder detection since the slopes 1s drawn ofthe 6, versus v curve at these hints ar positive For aspral text conducted from starboard to port tho unstablship B may start with an angular vloeky Aenoted by point (a) i Pg. 13. Aa the rader angle {Ptednced the angular velocity is redueed following Curve Bunt at sero rudder defection an angular ‘loci nested by point (e) i obtained. pon eon Unuing the rudder defection to por, the ship stl om tinue totam to starboard againot the direction of the ‘der defection unt pont fa) on curve Bis reached ‘Any inrese ofthe rudder angle ta port beyond point oF wl ease toe ship to sudlenly assume the large Scgular vebeeity to por indicated by point) ad Terhape temporary even overshot (2) Sims, ‘ies the spa is repeated from prt to starboard, a ineence in the rudder angle beyond point (2) in Pig 1 lease the ship to ring a8 fast as 6 inertia veil tt to the angular rely indicated by point (0). Hence,an unstable ship ean ar against is rudder Upto cen rer angle and then suddenly swine ithe opposite dreton ta « new sable poston Tor that rudder defection ange ‘This beevior of» ship unstable in straight-line mo- tion is exatiy analogous tothe behavior of 4 ship ‘heh transversely Stable. The transversely ust Able ship casnot remain upright even inthe absence 203 7 “8 Ee = 2 BRAN. a a 3 i Fig 1 lation ete being angi) ang monet once od eae nee of a heling moment. I will hes ether to porto Starboard toan angle of heel ndiated by ether ©) or {ep for ship B of fig. 14, These are positions of equ {ttm sce the slope of curve Bia positive a these points fs port heling moment i applied tothe unst {ble ship B whieh is iialy heeled to the starboard ng (in Fig 1, the ange of bee! wil be reduce, but ll sill romain to starboard unt point (2) reached. Any further increase inthe heeing moment {o port will ease sipB of Fig. lurch from poi, {@)to point (a) which sa stale ponton of large het {0 pore [The hee! angle wil overshoot (oy) but wil finaly atte down at (0)} Hence, no ois in. the Unstable region between (a) and () canbe obtained {or ship B during an inclining experiment. Thus, the behavior ofan unstable ship van elning experiment ‘would be completely analogous to that ofa unstable pnts piel tat. TE was noted in Section 42 that the controls xed stably indexes are not speed dependent at low and Inoderate speeds for motions in the horizontal plane Based on this eonclsion, the results of spial manew ters conducted at diferent speeds should not aifer [om one another sgnfeanty Te should also be noted that the results ofthe spiral maneuver shown in ig {Bare esentaly symmetrical about sero rudder angle andzero yaw rate, These sre typical of results obiained ‘wth shipe that are bth dynamtealy and geometrically24 PRINCPLES OF NAVAL ARCHITECTURE di Soy BE # ig | Sau \ el 5 = | # oe ‘a oa cy symmetrical about the 2zplane. However, because of propeller rotation, ships with an odd number of pro- pollers or with any number of unirotating propellers fre not dynamically symmetrical about the zzplane ‘Therefore, the results of spiral mancuvers conducted ‘vith such ships are likely to be displaced to one side ‘or the other depending on the direction of rotation of the tneompensated propeller. For a stable ship, the rudder angle at_y = Os the rudder angle needed to ‘maintain a straight course. This is the initial equilib- tum rudder angle or 8, in the symbology of Section 53. Tt will also usually be associated with nonzero value of 2,. The values of 8, and n) are also called neutral angles. For an Unstable ship, the rudder angle corresponding to the position of haif the loop height shown in Fig. 18 isthe approximate neutral angle. ‘eis essential in conducting & spiral maneuver to allow sufficient time for conditions to “steady” at each rudder angle otherwise spurious results are obtained, Fig, 15 Strom Tejsen, 1968) shows results of nonlinear predictions (based on systems of equations similar to those introduced in Section 8) which are three spiral tests conducted with three different time intervals be: threen eonseentive rudder deflections (a) 60 see. (0) 120 sec. {e) no time limit. Tn the frst (to cases the results show a sloped “hysteresis loop giving the impression that the ship js unstable whereas in reality the last ease shows that the ship is quite stable. “As wil be discussed in Sections 15 and 16, directional instability is not necessarily bad. Large, slow moving ships with directional instability ean be handled in & fquite satisfactory manner. The degree of instability Telative to the type, size, and speed of the ship is important. Tt is obvious that the spiral maneuver eannot be condueted with submarines in the vertical plane. Here a very simple test, called the meander test, suffices. In this test the stern planes are deflected to a prese lected value for a very short time and then returned to their neutral angle. Ifthe subsequent path followed by the submarine in the vertical plane is a decaying osellation at shown in Pig. 8, Case I, the submarine is directionally stable, Ifthe path is an ineeasing 0s tillation, the submarine is directionally unstable. Be- fauise ships do not possess directional stability in the horizontal plane, the meander test is not used in that plane and recourse is made to the spiral maneuver. we oF Tem 44 The Bech Reverse Spiral ond vers, The Bech or reverse spiral maneuver (Beeh, 1968) ie an alternative test tothe direct spiral manew ver. In the reverse spiral test the ship is steered at constant rate of turn and the mean rudder angle re ‘quired to proce this yaw rate is measured. This pro edure is repeated for a range of yaw rates (from 05 dog per see port to 0.5 deg per see starboard, forcONTROLLABIUTY eee 205 instance) until a complete yaw rate versus rudder an fle relationship io eotablished. Resules for a direction. Ey Stable vessel are similar to results from the direct Spiral. For unstable vessels however, a hysteresis loop is identifiable although a definite relationship is ind ‘cated within the loop, Pig. 16, This is because the test condition is no longer eontrls-fixed. The results still provide the shape of the loop for evaluation of the Agree of instability. ‘A properly calibrated rate gyro and an accurate rud- der angle indicator are required although in certain teases the test may be performed with the automatic Steering devices available on board. If manual steering is used, the instantaneous rate of turn must be visually displayed for the helmsman, either on a recorder or ona rateoftur intr. Using the reverse spiral test technique, points on the curve of yaw rate versus rudder angle may be taken in any order. Although not commonly in use, the pullout tests Fig. 17, provide a indication of a ship's stability on a straight course (Bureher, 1972). The ship is fist made to turn with some rate of turn in either direction. The ‘rudder is then returned to midships (neutral position) If the ship is stable, the rate of turn will decay to zero {for turns to both port and starboard. If the ship is only moderately unstable, the rate of turn will reduce to ome residual rate. The pullout tests should be per- formed to both port and starboard to show possible asymmetry. Normally, pullout tests ean easily be per- formed in connection ‘with other tests being run. Section 5 Stability and Control 5.1 General. The controle-fixed stability indexes discussed in the preceding section constitute one of the important elements of path keeping at sea. Because the practial problen of path keeping involves re peated instances of path correction, its basic elements {end to merge with these of path changing. These basic elements are shown on the control loop of Fig. 1. Path keeping and path changing ability of = ship depends (a) The magnitude and frequency of any yawing ‘moments and sway forees acting to disturb the ship from the desired path (@)_ The character of the response of the ship with controls fixed to these disturbances. This response will be reflected in changes in the ship's path shown at the extreme right of Fig. 1 () The rapidity with which the error between the ‘i's path and he desired path can be detected, and with which corrective action ean be initiated (d)”"The rate at which the corrective action is trans lated into movement of the rudder. This is a funetion of the play between the third and fourth elements of the contfol loop and the rate at which the steering ‘gear can deflect the rudder in the fourth box of Fig. 1 (e)_ The magnitude ofthe control force and moment applied to the ship by the rudder. ‘OF these five elements, only the second is dependent fon the controle-fixed stability of the ship. This is an Important element, but so are all the others. Usually, deficiencies in any'single element of the control loop ‘an be compensated for by improvements in other ele iments, Por example, tis shown in Section 112 that the use of properly designed automatic controls in tlement (c) can correct for controls-fixed instability in flement(b) Often i is assumed that inereases in ud fer size, element (¢), arin the rate of rudder deflection, flement (d), can correct for deficiencies in the path keeping or path-changing ability of a ship. The latter view is shown in Section 17 to be incorrect. Although minor degrees of controls-ixed instability are com ‘monplace in ships, the best design is likely to be that206 which possesses minimum deficiencies in each element Of the control loop. 52. Definitive Maneuvers. The naval architect is mainly concerned with elements 0), (2, and () of the path-keoping and pathchanging problems. Therefore, Certain definitive maneuvers have been devised to dem enetrate th eicacy ofthese elements of the control loop and to exclude as much as possible the influence of eloment (c). Essentially, these maneuvers establish the basi stability and control characteristics of a ship Independent of its helmsman or autopilot: (a) Direct or reversed spiral (see Section 4.8) (6) Zigzag, Z, or Kempf overshoot (see Section 5:3) (e) Turning (see Section 6.1) ‘The spiral maneuver as deseribed earlier serves mainly todetermine stability characteristie, whereas the zig: zag maneuver is to determine control characteristics. ‘The turning maneuver denotes turning qualities, All three maneuvers are important for both merchant and naval ships. Specific performance criteria and other elated trale are dteusted in Section 14 53 Tiglag Maneuver. Second to the spiral ma reaver in importance is the aigeag maneuver, also Known as the Kempf overshoot or “Z” maneuver (Kempt, 1944) "The results of this maneuver are indicative of the ability ofa ship's rudder to eonteol the ship. However, just as the results of the spiral mancuver give some indication of eontrolefectiveness (yaw-angle rate ver- sus rudder angle), so do the results of the zigzag test depend somewhat on the stability characteristics of the ship as well as on the effectiveness of the rudder. ‘The typieal procedure for conducting the test is as tallow (Gere, 1959) sway the ship as in step (a) of the spiral ma neuver. (See Section 4.3). (). Deflect the rudder at maximum rate to a pre selected angle, say 20 dog, and hold until a preselected change of heading angle, say 20 deg. is reached, (At this point, deflect the rudder at maximam rate to an opposite (checking) angle of 20 deg and hold Until the execute change of heading angle on the op- posite side is reached. This completes the overshoot test (d) Ifa rigeng test isto be completed, again deflect the rudder at maximum rate to the same angle in the first direction, This eycle can be repeated through the third, fourth, or more exeeutes although characteris ties through the first overshoot are most important as discussed in Section 15. Tig. 18 shows the results of a zigeag maneuver car ried through five exeeutes. The results shown are those that ean be readily obtained with a controlled model ina towing tank or with a wellinsteamented ship at sea, With ordinary ship navigational aids, only the rudder angle and yavrangle curves are readily obtain able ‘The principal numerical measures of control ob- tained from the overshoot maneuver as illustrated in PRINCIPLES OF NAVAL ARCHITECTURE N ‘7 v “aces Fig, 18 are: (a) The time to reach the second execute ‘yaw angle; (9) the overshoot yaw angle; and (e) the ‘Svershoot width of path, All ofthese are important ‘operational parameters. The first i a direct measure fof the ability of & slip to rapidly change course. It improves with increased rudder effectiveness and with decreased controle fixed stability (Arentaen and Man ‘del, 1960), The secon and third are numerical mea- ‘sures of countermaneavering ability and are indicative fof the amount of antijpation required of a helmsman while operating in restricted waters, Itwas shown that the magnitude of the yawangle overshoot decreases ‘with inereased stably but increases with increased rridder effectiveness. On the other hand, the overshoot Width of path decreases with both increased stability ‘and increased rudder effectiveness. "The results of the zigzag maneuver are speed de pendent. In general, for any given ship the time to reach execute decresses with increasing speed, and the overshoot yaw argle and the overshoot width of path inerease with increasing speed. However, the ow ‘imensional time to reach execute, interpreted in Fig 18, as ship lengths of travel to execute, increases with ingreasing speed beesuse of the influence of the rate of rudder deflection, 8. When » is nondimension alized, 8'_ = 5 4L/V, it may be interpreted as de- {grees of rudder deflection per ship length of travel ‘AL low speeds, this nondimensional rate is much higher than at higher speed since 8 , is essentially indepen:CCONTROWABILTY ent of speed. Hence, with respect to the ship length of travel sale in Fig. 18, the rudder would be deflected more rapidly at low speeds than at high. Hence, the rudder exerts its full influence longer at low spceds, ‘whieh Lens wy reduce Une oudiaeustoual Uane to reach execute as speed is reduced. However in spite of this, ‘beneficial efflect as spood is decreased, the time to reach execute usually inereases with decreasing speed. (The effects of the rudder are reviewed further in Section 1? Tn the ease of submarines, the overshoot maneuver is employed in both the horizontal and vertial plane and its results are perhaps even more operationally Signifieant inthe vertical plane than in the horizontal ‘This i true beeause in the vertical plane submarines ‘ust operate within a relatively shallow layer of ‘water, while they usually have ample freedom of mo tion in the horizontal plane except when they are in festrcted or eongested waterways, Henee, to the sub- ‘marine operator, overshoot pitch angle and overshoot Change of depth are very important parameters. 54 The K and T Coursekooping and Turn exes. This section presents the Nomoto simpli ‘alysis of K and 7 indexes which ean be developed from zigzag trial data. These indexes are widely use, implied analysis tools developed from the linear equations of motion. They are useful in comparing oursekeeping as well as turning abilities, which will be presented further in Seetion 6. ‘While Equations (11) and (12) expressed the linear ‘equation a8 a pair of simultaneous first order differ ‘ential equations, where the constant coefficients are the dimensionless acceleration and velocity. deriva tives, itis possible to express these equations in an alternative form It was fist shown by Nomoto (1957) that these equations ean he written as a pair of de coupled second order equations as follows Tig H+ (TY + Te) t+ r= Ky + WT be a5 Ty Ty W + (Ty + TY + vm Ky by + KETY by ‘This expression for the coeficients in terms of the time constants 7’, 72, Ty and 7, as well as a system gain ‘Kis consistent with eontrol engineering practice Since Equation (15) are & linear system as are (11), a solution similar to (18) may be derived and it may bbe seen that the roots of the solutions are related to the time constants follows: Returning to the linear yaw and sway (11) and (12), itean be seen that they are coupled only through the terms Nv" and ¥.’1", which are typically small, par ticulary'for ships with near fore-andaft symmetry. If these cross-coupling terms are neglected and sway 207 ‘velocity or side slip angle thus eliminated, turning de pends only upon yaw rate, % and is defined by the Simplified non-dimensional yaw equation of motion: a6) ‘Nomoto (1957, 1960, and 1966) noted that this equation could be divided by the yaw damping coefficient, N', fand rewritten in the parametric form: Tbr = Kb. an where the non-dimensional parameters or indexes 7” fand "are given by: nL = Nir = Nig be Tee n/N, =~ NIN, h+m-h as) i= MUN, {In dimensional form the equation ie T¥ +r = KB, where the non-dimensional parameters are related (0 the dimensional Nomoto parameters 7 and K by: r= 1V/L) Ke = KLM) ‘The indiees 7" and K" represent ratios of non-dimen sional coefficients from (18): p= _Y80 inertia coefficient ‘Yaw damping coefficient | __ turning moment coefficient A “yaw damping coefficient Dividing A" by 1" shows that the two indexes are related by K’ _ turing moment coefficient Tr ~ yaw damping cooffcient In practice, Equation (15) can be solved by numerical integration. For the simple case where the rudder is put over suddenly to an angle 8, and held there, the Solution for ris given, in terms of and K, by: re Kade) as) ‘This shows thatthe yaw rater inereases exponentially with time but ata declining rate dependent on T™and approaches a steady value KO, (or K'V,/L). A larger thus provides greater steady-state turning ability, and a smaller value of 7 provides a quicker inital response fo the helm. Quik rons implies good ‘ourse-changing ability and good course-checking abil- ity when a turn (or other maneuver) is completed. Since ‘quick responce is obviously valuable in course keeping (Gteering), itis thus consistent with a smaller 7. The above discussion of Equation (19) shows that Thas no effect at allon steady turning rate, buta small would Feduce the time required to reach a steady turn. At ‘the same time, the index T'is a reciprocal measure of208 —| __ i — ol course stability, with stability increasing with decreas ing T. However, a negative value indicates an unstable dynamic character. The steoring quality indexes K and T have an immediate relationship to the conventional ‘measures of ship turning. Ie may be shown that, under the previously stated simplifying assumption that sway ean be neglected: where oy isthe stability index of Section 4 Thus, 7" offers a direct quantitative measure of straight-line sabi Por steady turning a constant rudder angle 8 1 = Kbg = K’ Vig /L eo) Steady turning diameter, Ds, by defiition is p= ey and hence nondimensional turning diameter, D,/L, and &" are related by: Doll 2) ‘This relation can be derived from (16) in Section 5 by neglecting sway (placing N, = 0). This isin accord with the statement in Secton’5.3 that 2/L depends fn the relative magnitudes of 1, and Ni. It shows that with a larger value of A” a smaller rudder angle ray be used in achieving a given turning diameter. ‘The main maneuvering qualities of a ship using in ‘ear analysis ean thus be characterized using only the indexes 7" and K’, where inereasing. values indicate improving performance: T Course stability 1/7 Responsiveness to rudder K Durning ability A highly maneuverable ship (with high responsive ness to rider and both good turning and low course 2Wirb = 21K, PRINCIPLES OF NAVAL ARCHITECTURE stability) will have a small vaue of T and a large Walue of XK’. In other words, a large ratio K'/T, oF Norrbin parameter, P= R20, (Nomoto and Norrbin, 1969) is indicative of good maneuverability Ivis not'a good indicator of course-keeping ability (or ood steering) however, because this ean be achieved fither by high course stability and low responsiveness {high 7) or by low or even negetive stability and high Fesponsiveness (high 1/77) piss superior automatic onttol In short, a large ratio K'/" suggests good ‘overall controllability only if stability is n0_ greater than necessary. ‘Overshoot angle, which i obtsined from the zigzag maneuver (Section 8) has often teen used as a measure of controllability. Nomoto (1866) has shown that ‘overshoot angle is, for given rudder angle, nearly proportional to the produet K’ 7 Overshoot angle thus has the inherent weakness tha: it cannot be used to discriminate between a ship with: (a) good tureing and {ast response or good course stability (arge A” and small 7") and (b) poor turning and slow response or ‘course stability (small K” and large 7"). The former with large X'/T, is clearly a fa: superior ship in over all controllability. But overshoot angle does indicate turnchecking ability. For further guidance Nomoto has suggested that: ‘Turning moment coefcient = Ay/LT and Yaw inertia coefficient © /L°T where Ae is radder area, and is displaced volume Using dhése spproximations: Ko ab Aub pean ‘where cs a constant of proportionality. Fig. 19 sum tare results for various shie and rudder angles, an indicates by the stright lines that e tends to be independent of ship type and rudder angle, I is clear from (28) tht, since ange A/T is favorable, a large Yale of AyL/7 is desirable. This simpli iar anal $s indicates that ship dimensins (particularly and 5)sas well ax adder area, wil havea significant fle. Once overall ship dimensions are established, both as pects of controllability can be sgaiicantly improved Uy inoreasing rudder size o effectiveness. “The indexes Tand A can be alelated numerically using Equation (15) if hydrodynamic and mass coe ‘ents for the ship are known, One advantage of these inde ie that they cam be derived trom the rests atthe standard tras or free running model maneuvers for comparison with ealeuation. They give physical meaning to the standard teas “The applzation of and Ato determining eriteia of controllability i discussed in Section 14.7, The ele Iments of turning. performance as separated from Coureekeeping.and contol are introduced and. a dressed more flyin Section 6 (23)CONTROLLABILITY 209 Section 6 Analysis of Turning Ability 4. characteristic ofthe Tring Path All ship seuvering, except some stopping maneuvers iveles turning The esponee ofthe ahip to defection of the fuider and the resahing forees and moments pro- ced by the udder can be dived into wo prone (eh An inal tricstnt one in whch ipeitens surge, away, and yew aeecleratons goes {O)" steady tdening portion n which rate of urn st forward speed are constants the path of the {Ships ecular (n the absence of signieant external tones. Fi, 20 isa defiton dagen for turns of ay a ameter. General, the turning path of «ship is char Scterized by" fous. sumericl”reasures; advance, transfer, acta Giamater, and sendy toring cam ter, As shown in Pi. 2, al but the lst are rested ‘oheatng positions ofthe hip rather than to tangents to the turning path, The advance the distance om the origin at "execute" tothe wax of the sip when that ais has turned 90 deg. The transfer isthe dis tance from the orginal approach coarve to the ong of the sip when he cata has turned 90 deg. the {acticl dameter is the distance from the approach Course to the axis of the ship when that as has ttened 180 deg These parameters of hi taming lh are uefa for characeriing maneuters ine Spen sea. Seton 1 dincoses the use ofthe turing Chee farther as a definitive and practical est fa SG mancaverabil Fig. 20 also shows the position of the socalled pivot pointina steuy turn, This pointin of interest, beeuse {olan cbserver aboard taening abit appears a theship were pivoting about point esl some tat the bow, AL ths poine, Decase of the comb nation of th deft angle onthe ship an the ration a the ship, the flow of water pat the ship panel tothe swans of the shi. Forward of this pont, for @ Starboard tum the flow approaches from of the star board ie of the ship aa thown in ig 9, and of this pein he fow approaches from the por side. Ths, 2 fied wer finin the plane of symmetry af the ship ‘ould experience no ange of attack at this lostion Ge alto Seton 16. According to Fig. 20, the de tance between the pve point andthe center of avy Of the ship 7 = sin b. Because smal adie arnt fre usually Sssocited with large drift angles and large radius turns wit smal dr angles, he product Bin doesnot vary signifeanty for ere ips or forthe same ship at diferent turning radi For ost ships the pivot point is somewhere between the Sow and about 1/8 aft ofthe bow (ifandl,1953). Besed on emprel data, the ert angle, indepen, fereraly falls within the following range of vues: BSS L7R S145 and = 18 L/ The former relationship yields values of z, from 0.4 to 0.5 L de pending on the L/R ratio, The latter relationship yields Wales of =, ~ 032. During the first phase of the turn, before it becomes steady, there is aso an apparent “pivoting” point. This point near the bow’ of the ship intially follows a path ‘which is straight-line extension of the approach path, ‘while the stern deviates outward of the approach path And the bow deviates inward of i "The turning circle maneuver has traditionally re ceived the most attention in treatises on ship mane verability. One reason for this is that it has been and Stil i an important practical maneuver that ships fre quently perform, Also, because the final phase of the turning path is a steady-state maneuver, it has lent itself more readily to analytical treatment than have transient maneuvers. {62 The Three Phoses of @ Turn. Suppose that ship is advancing on a straight path when its rudder is deflected and held at a fixed angle as in Fig. 20.'The resulting path of the ship may be divided into three distinct phases Fig 20, Tang path a 2 hip210 PRINCIPLES OF NAVAL ARCHITECTURE ———_— o ‘The fist phase starts at the instant that the rudder begins to deflect and may be complete by the time the rudder reaches its full deflection angle. During this, period, the rudder force, ¥, By, and rudder moment, N, By produce accelerations and are opposed solely by ‘the inertial reaetion of the ship because there has not yet been an opportunity for development of hydro- “ynamie forees arising from a substantial drift angle, ‘ora rotation, r, to develop. Hence, in this stage, B p/V = r=. Using the dimensional Equation (10) and introducting the rudder foree and moment, the Tinearized equations of motion in the first phase of ‘turning are: (Ye Me = Voy ee UL —N)R-N B= Ne Be ‘The values of the accelerations, and # that occur {in this phase can be obtained from these equations. It may be noted in Figs. 20 and 21 that the transverse Acceleration, #, is negative or directed to port in this phise, whereas the turn will eventually be to star- board, This is beeause (for & rudder at the stern) the rudder foree ¥, 8, is directed to port for a starboard turn ‘The accelerations # and # can exis in isolation only momentarily, for they quickly give rise to adrift angle, {and a rotation, x, of the ship. With the introduction ‘of these parameters, the ship enters the second phase ‘of turning. Here the accelerations of the ship coexist With the velocities and all of the terms of Equation a. 2) Chri ftom ph (0) along with the excitation terms ¥, By and N, By fre fully operative (12), The crucial event that takes place at the beginning ofthe second phase of the turn Ki'the ereation of a ¥, eforee positively directed to starboard in Fig. 20 towards the center of the turn, Fesulting from the introduction of the drift angle, ‘The magnitude ofthis force soon becomes larger than the ¥, by-force which is directed to port (Fig. 5. As ‘Show in Fig. 21 tis causes the acceleration ¢ to cease fo grow to port and eventually to be reduced to zero fs the inwardly directed Y. foree comes into balance With the outwardly directed centrifugal foree of the ship, However, in the second phase of the turn, the path of the center of gravity of the ship at first re- Sponds to the Y, b,-force and tends to port before the Y-r-force grows large enough to enforce the starboard turn, This port offset although visibly proteayed in Fig 20 is negligible or nonexistent in practice because of the shortness of phase 1, and the quick development of the large Nemoment in ship forms. 162 Stoody Turning Rodivs. Finally, after some os cillation (ome of which is due to the settling down of the main propulsion machinery and is characteristic of the particular type of machinery and its control s}= tem) the second phase of turning ends with the estab lishment of the final equilibrium of forees. When this ‘equilibrium is reached, the ship settles down to a turn ‘of constant radius as shown in Fig. 20. This i the third, or steudy, phase of the turn, Here v and r have nonzero values, but v and rare zero. Thus, using Equ tions (10) the linearized equations of motion ina steaCONTROWABILITY 25) ‘These two simultaneous equations can be solved for ‘rand v provided that the stability derivatives ¥,, Y., NL Nat the control derivatives Y, and N,are known, Noting that r= ~ rL/V and that the steady turning radive = Wr, chen r= LB othe recip rocal of the ratio of the steady turning radius to the ‘hp gt Salving the nondmensonl version of 25) 26) and | an where and 8 are in radians and positive R denotes 2 starboard turn ‘Thus, according to the preceding linear theory, the steady turning radius would be proportional to the ship length, Z, and inversely proportional to the rudder- deflection angle, By, and the drift angle & would be directly proportional tod. Solutions (26) and (27) are useful for estimating the steady turning radii and drift angles of stable ships ‘with fairly large diameter tums of about four ship lengths or more. They are used to estimate the turning. radi of torpedoes, and are useful for estimating the tring rai of spa at es than maximum rudder angles. "The great majority of merchant ships have turning ters of from two to four ship lengths at full rudder angle, and many ships have tuening diameters ‘of to ship lengths or less. Such tight turns introduce Strong nonlinearities that tend to reduce the validity of the linear equations of motion, Procedures for pre- ticting the maneuvers of tightturning ships are dis fussed in Section 8 ‘64. Relationship Between Steady Turning Radius and 1e Hydredynemie Derivatives. Equation (26) devel: tped from linear theory may be used for stable ships to ptedict the effect of changes in the hydrodynamic derivatives on the turning radius. In slightly modified form, Equation (26) is: Roo Lk It is seen that the numerator is identically the sta- bility criterion, C, Equation (14b) of Section 4.2; it was, shown in Seetion 42 thatthe value of the numerator 's independent of the choice of origin. If the relation ships NOD + VD and a w= D+ DS fare substituted in the denominator, it reduces to YUN )09-— Y's(4")09 whieh i also independent of the cholee of origin, thus i the ship is stable the nu rmerator is positive and ifthe ship is unstable the nu rmerator is negative, The sign of the denominator is tlwaye positive for the following reasnns Ta) Povis always negative and N'y is always neg- ative for rudders located atthe stern (Fig. 22). In the figure bis negative following the definition of these BS given in Section 3.5 and following the sign com ‘ention given in the nomenclature at the end of this chapter. The moment N resulting from negative 8, however, is positive according to the same sign con vention, Simfarly, 3 8 were positive, N would be neg itive, Henee, the derivative Nis always negative for rudders at the stern, In Fig, 22, the foree, Y, arising. from the negative Bf also negative, if 8 were postive, Y would be postive; hence, the derivative, ¥,, is always positive henee, their product is positive. @) Yu always postive, Vis almost always neg: ative; hence, subtracting their product will add posi- tively to YN e) if positive (tis rarely so), its magnitude \V/a2 PRINCIPLES OF NAVAL ARCHITECTURE rs (a tt Pras Fa. 22 iption ot {is bound to be small so that the larger postive ¥".N", product will determine the sign of the denominator Temay therefore be concluded that ifa ship is stable and its rudder is located at the stern, a positive (star board) # will always result from a negative 8 and vice versa. However, if the ship is unstable, then the ait Imerator of the righthand side of Equation (26) is neg ative and & will have the same sign as 5. Physically this means that a ship will turn against its rudder, which is in accord with the behavior of an unstable Ship, Sinee Equation (26) deals only with the slope of the & versus 8 curve at 5 = 0, which is a region of ‘unstable equilibrium for unstable ships, it cannot be ‘used to predict the turning radii of unstable ships. For stable ships, Equation (26) may be used to ex: ‘amine the effect of changes in the individual deriva tives on the turning radius. Equation (26) shows that ‘the effect of changes in ¥",on depends onthe relative ‘magnitude of (N",) compared to N's. If N'y has a ‘eater magnitude than (N’) then inereasing the mag- nitude of ¥”, would decrease the radius of turn. On the other hand, ifthe magnitude of (V) is larger than Nj, then an inerease in the magnitide of Y". tustallyinerease the radius ofthe turn. Since for stable ships, (N") is usually much more negative than N the usual effect of inreasing the magnitude of Y", ‘winerease the radius ofthe turn, Thus, while the ¥.- {oree is responsible for the initiation of a turn in the desired direction, an inerease in the magnitude of Y, ‘doesnot necessarily reduce the steady turning radi ‘The effect of 1", on R is readily predictable. If N", is negative, increasing its magnitude will decrease the positive vaiue of the numerator of the righthand side ‘of (26) and inerease the postive value of the denomi "Farrer oat atthe bow, Ny wil araye be one and aay bee ee bya le sige Ys awaye paive when the rade are Kea oat the see ‘SyaRa0ARO TURN, SOR Pow a (org pase iy Hane in nator, hence will decrease on two counts. On the other hand, if N', were positive, increasing its mag: nitude would increase the numerator and decrease the Genominator, and hence would be increased. ‘The effect of N', and N", on R’ is equally clea. According to Equation (18),'an increase in the mag: nitude of N", will increase 2 while an inerease in the magnitude of’, will decrease. This result isin accord with an intuitive examination of the question. ‘The effects ofthe remaining derivatives on R depend fon the sign of N’,, and can be deduced from Equation (25) when the sign ofthe derivative is knoven (see Table ». 65. Heol Angle in e Turn. While use of the rudder is intended to produce motions only in the yaw (zy) plane, motions are also induced by eross coupling into the piteh (zz) and roll (yz) planes. The unwanted mo: Toble 1—tlfet of Changes in the Detvatives on Steady Tring Rodis for Stable Ships with Rudder) ot the Str) Derivative Y x Ne Ns vr ry FS vty 7 a ot ern tnd roe "genio si of he aria EEE co png ea ft ede meer ing es SAC ER SmeeCONTROWABILTY tions in the roll plane, particularly, are likely to be large enough to be of significance The magnitude of the heel angles induced by the rudder can be estimated by considering the heeling moments arising from the ‘Vertical disposition of the forces described in the pre ring section, That disposition for the first phase of starboard turn, is shown in Fig. 2(a). The direction ‘of most of the forces may be ob (of Equations (24) all terms are gathered on one side land equated to zero as follows: Vibe+ ¥.0+¥,¢-ar=0 C8) Since Y, is always positive and 3, is negative for a Starboard tur, YB is negative ox directed to. port Since ¥, is always negative and 0s negative in the first phase of @ starboard tur, Y, 0 is positive or directed to starboard. Sire ¥., may be either postive for negative, te sign of ¥, ris no: predictable from (Ga) in any event, V, #15 very small compared to Yi. Finally, since @ is negative and A is positive, (Cat) is positive or directed to starboard ‘The approximate angle of heel, , may be obtained by equating the resultant heeling moment, which is the sum of the moments of each of the forces in the yeplane, Fig. 28), to the hydrostatic righting) mo- ent. A graphical slution ofthis ecuation is deseribed in Section 7 of Chapter II. For this purpose YY, may be taken to be acting at half draft, sd, atthe vertical center of the rudder and Ai and Az at the center of gravity of the ship. If moments ace taken about the half draft, itis obvious from Fig. 2(a) that the heel Angle, 6, wil be ta the starboard (postive) in the frst ‘hase of a starboard turn "The forces acting in the yeplane taken from Equa- tion (28) for the third phase of a starboard tum are Shown in Fig. 2900), If moments are taken about the center of gravity of the ship, itis seen that the heel langle. 6, i likely to be to port (negative) since ¥.0 + ¥-r mast be much larger than ¥,Bin order to enforce the starboard turn. Thus, between the first and third phase of a turn, the heel angle of a surface ship changes sign, The heel angle time eocord of ship with ‘large turning heel angle is shown for a starboard turn in Fig. 24, Its seen that the amplitude of the initial heel to starboard in the first phase of the turn is small compared to the amplitude of the second heel to port. This second heel favolves a large overshoot tangle beyond the equilibrium value computed in ac- ‘cordance with Fig, 24). However, eventually the port heel seties down to fairly steadr value eorrespond- ing tthe computed value fr th al phase of the ‘From an operational point of view, a potentially dan- “For bierzed sian the itch ne induced in tring ty the pubes well aby hall samme are ao frequently Tage naught be of concer. : ig. 24. Rabanne er oa enor Men rerous situation exists just prior to the completion of fhe frst large heel to port. A helmsman, fearing too large a heel to port, might at this instant decide to return the rudder quickly to amidships. This would Climinate the Yb, force and the hoel to port would be ‘aggravated rather than alleviated. The only safe action to take in such a situation isto immediately, butslowly fand cautiously, reduce the rudder angle and at the ‘Same time reduce speed as quickly as possible. Tn the ease of a submarine turning, submerged, the heel angle is inboard (starboard heel for a starboard turn) throughout all phases of a turn The reason for this is that the positions of both the ¥,0 + ¥,¢ foree of Fig. 2(a) and the ¥.0 + Yr force of Fig. 2300) are fonsiderably higher relative fo the center of gravity n'a submerged submarine than on a surface ship. In particular, the bridge fairwater existing on practically fll submarines is an effective lifting device and con- ‘bates heavily to both the magnitude and the height fof the Yieforee of Fig. 23(0), neneasing rolling: mo- tment, K-v. It is clear that ifthe Yiu + Yor force is ‘aised sufficiently igh (on some submarines iis raised {oa position above the center of gravity, the heel in the third phase of a turn will be in the same direction fs in the first phase of a turn. Thus, the first heel of 4 submerged turning submarine is an inboard heel of very large amplitude called the snap roll with sub- sequent inboard rolls of diminished mean values, Ac. cording to Arentzen and Mandel (1960), the ratio of the snap roll in the frst phase of a turn to the steady hheel in the final phase of @ turn may be as large a5 289 for a submarine with 2 large bridge fairwater and as lange as 6 fora submarine without a fairwater. The latter submarine, however, has a much smaller steady2 PRINCIPLES OF NAVAL ARCHITECTURE heel angle than a submarine with a fairwater. Thus, the fairwater plays a dual role in turning: {a)_Itinereases the roll excitation ina turn because ofits large influenee on the roll moment due to the transverse velocity A. (}). Teinereases the magnitude of the roll damping moment, K,p, and henee, dampens the amplitude of the overshoot of the snap rol See Section 12 and 15.1 for description of yaw-roll coupling effects in waves and further discussion. 166 Reduction of Speed in a Turn. Part of the rea son thatthe inital snap rol of a submarine is so much Jarger than subsequent rolls is that the speed of the submarine is rapidly reduced as soon as it develops a Substantial deft angle. This is also part of the reason ‘why the first roll to port of the surface ship in Fig. 24 Js much larger than subsequent rolls. However, inthe case of the surface ship, its speed is more greatly reduced by the time it experiences its largest heel Angle than in the ease ofthe submarine. This partially ‘ceounts forthe fact that the ratio of the value of the first large rll to the steady heel shown in Fig. 24 is not nearly ag large as the comparable ratios for sub- ‘marines cited from Arentzen and Mandel (1860). ‘The magnitude of the speed reduction in a turn is largely a function of the tightness of the turning circle (Davidson, 1944). Fig, 25 shows the empirical relation- ship between the ratio of the speed in a steady turn ‘wo the approach speed and of the turning diameter to the ship length, developed by Davidson on the basis of a large number of ship trial and model results, The discounting by Davidson of the differences between fullseale and model results has been shown to be er roneous by StromTejsen (1965), among others. Su: perimposed on the Davidson results in Fig. 25 are Shiba’s (1860) results. Davidson and Shiba concluded that the relatively small seatter of data shown could rot be related 19 rudder angle, appronch speed, oF rudder-area ratio, which were, of course, among’ the variables of the tests. The effect of changes in type of ship power plant and in ship configuration on the speed reduction duringy maneuvers is discussed in Sections ST and 16, In spite of the increasingly severe speed loss asso: ciated with tighter tums, Davidson showed that by Aleereasing the tactical diameter to two ship lengths or less, signifeant operational aspects of turning are Improved. For example, Fig. 26 shows that a 122 m (400 ft) long, 20-knot ship with e 7D/L = 2.0 achioves f full course reversal and has almost completely re- frained its approach speed in the 1% min that are re- {Wired for the ship to pass its original execute point Headed in the apposite direction. On the other hand, the same ship witha 7D/L = 4.5 required 2¥-min and [e remem t E | 4 t cog | fo 28 Compton of 189 deg nd peed cry hacen Tsim (et penn WeaCONTROWABILTY much more sea rdom for the same maneuver. "The speed used in the computation of wee angle in the final phase of a turn in accordance with Fig. 2300) ‘Should of course be the reduced speed as determined from Pig. 25 and not the approach speec 2s ‘Transient turning and complex maneuvers eannot be predicted by linear theory. Instead one must resort titer to nonlinear theory in eonjunction with eaptive fnodel tests (Seetion 7) oF to free-running radio-con- trolled model tests (Section 8) Section 7 Free-Running Model Tests and Hydraulic Models 7.1. Freetunning Model Tost. ‘There re to dis vine epee of moral tests toed inthe alulation of Sip cutrotuoity. freed captive model tests and absvnning model tests Captive model tests willbe {ESSeibd in Section B and are ssful n developing sseficlets for use In ship trajectory pticuon equ ‘Gin’ Freeranning move tate are ore divet and make ope of selpropeled scale model of the sip Fund wth all appendages and remote eorrol, 0 that {Stal maneuvers ean be performed and controllability Sralated Simple freeranning models ace ypilly used to value turning performance ad cove Keeping Bit. Such testing ean be used to obtain numerical measures of stably when the mote sted with 9 Sphisteaed sensorcontol and pefton recording ‘tum eased trajectories from the ec slag. She reverse spiral definitive manewers can directly Sind index values fr comparzon wth established eet Ent ‘or similar values fom an aceplable desig. Stein 18 dieusses further the develgment ands Surnent of performance requirements ‘Contrliablty in shallow water can alo be eval ated ut faclity wth variable water depths val UR enig mode twee re et generally sutable for evaluating storing perform fic sinc sale effect are dest to overcome reerunning ship models an also be used In hy- draule nodes of harbor and waterway situations, i Sled remotely, such model provide 1 method of {raluating varius arrangements onthe hull, ad the ‘Hestvenes of varius types of veces ander put “lar shallow and restricted waterway eorditions Some Tage pled modes are also sned for training, and ‘uearhin faite such a the one in Sorreah, France iDemenet, Lewis ta 1987 "7a. ProwRuning Medal Tost Techniques. Prec: rain del tert require «model whose propeller Totatiog and eontolsurface positon may be realy CGntrelled and recorded es fonction af time. They So require lge maneuvering basn, a8 well <5 theuny for determining and record the 2 and ‘ordinates ofthe origin ofthe model, te model head Sng angle J, and if desied, the model heel angle &, aifans fancin of time Using instructions like those for the fullcale ship. the toring, sigeag, andre ‘versed spiral definitive maneuvers may be earried out ‘with a freesrunning model and the resulting charac- {eristies of these maneuvers may be directly deter tnined. For this reason, free-running model ests have been, and still are, extensively used for predicting the maneuvering characteristics of surface ships, and have feven been used to determine the maneuvering. char Sctvns of submerged submarines in the horn plane Martinussen and Linnerud (1987) provide a current stats ofthe use of free-running models for predieting maneuvering characteristics atthe design stage. Some Of the difficulties in developing test techniques are fonnested with the question of viscous seale effects as ‘dressed by many’ ineluding ITTC (1984), Burcher (1975), Okamoto, Tarai, and Oniki (1972), and Nikolaev tnd Lebedev (1980) Some detailed discussion of seale tects, particularly with regard to control surfaces, twill algo be presented in Section 14.2. Other dificulties fare related to the physical execution of the tests ‘On the physieal sie, cost considerations normally result in ting the same model used for towing and propulsion tests to perform free-running maneuvering Tests. Large models are thus often used which help in the reduction of scale effects but require careful test ‘execution to complete maneuvers because of space lim itations even when performed in large basins, "As Martitussen and Linnerud (1987) recount, five conditions should be met at the start of a maneuver; Forward speed equal to approach speed with the cor: responding propeller revolutions, rudder angle at neu fral, and sway and yaw velocities at zero, Itis difficult tosatisty all five when testing large models of unstable ships. Priority is thus given to forward speed, propeller RPM, and yaw velocity over rudder angle. During each ‘maneuver, the controller must carefully optimize use Of the basin to allow full maneuvers’ to be accom plished, ‘in order that free-running model test results may be directly applicable toa fulscale ship the following ‘ditional conditions, added to that of geometrie sim iivude, should be satisfied: (a). The nondimensional mass moment of inertia of the model about the z-axis, /,, should be identical to that of the ship. () The model rudder should be deflected to the26 A = Tet with tition eorrection same maximum angle as the ship rudder at the same nondimensional detection rate as that of the ship, ie, 2) where the subscript refers to model values and the Subscript to ship alues, Thus the model dimensional udder rate 6 ciferent from that ofthe ship 6, (0) Ifthe sip hel in maneuvers to be properly simlated, the F of the model az well a ie mond tmensional taneverme metacentric height mustbe en tlt that of the ship. In pratce, these are iu conde oy a). The model propeller operating slp ratio should te Montiel to the ship. propelir sip ratio. This & arly ipo the adr fet nthe Propeller race {If he speed loss n maneuvers i tobe propery simulated, the response of the motor that drives the ‘model propeller to an augment in model resistance Should duplicate the response of the power plant of {he full-scale sip to a corresponding augment in sip resistance ‘Condition (ds fulflled inthe ordinary sel propul- sion test by having the towing carriage prose Part. af the thrast necessary to drive the model. If not ae Counted for, sore models will maneuver more sharply {han the fltscle ship beease the rade force (and hence rudder moment) larger forthe model as a result of more flow pas the rudder (Ear raoing dls ti condition common alld with by sing an air propeller to provide part ofthe thrust. Tisai force can be measured and dy tamiealy changed as a function of measured model Speed to improve resis. Fig. 27 shows a eomparon of results from similar ig2ag tests run with and with PRINCIPLES OF NAVAL ARCHITECTURE 1 © Open wate tt, turbulence povided by swemeahujtream of rer 120 Behind hl and propeller; bated on mo (eloped i. 28 br conn Co nin fae one onsen out friction correction, showing the effect of a more felfetve rudder in the latter ease. Condition (eis usually ignored witha constant RPM held. Freerunning model tests thus show less speed Jost in & maneuver than do fullscale results. Results fre often considered acceptable for diesel powered hips which normally show only a 10-15 percent re duction in revolutions in a turning eicle with @ 35 deg rudder angle. Use of a thrust and torque transducer ‘vould allow simulation ofthe propulsion engine char- Ecterstics during tests ‘A8 will be presented later in Section 142, a rudder in'& free stream wil encounter flow separation at smaller rudder angles in model seale. There is a gen- feral belief that a model rudder when placed in the propeller race, encounters a high turbulence level Wig 2 Tyla yo malt onl 10,1978)CconrRowaaiury which makes this seale effect neglogble. Fig. 28 sup- ports this conelusion through its comparison of results for a model rudder tested in open water with test resulte with a rudder behind a propeller on a captive model running at the ship self propulsion point. “The general problem remains of having different model ¥s fullseale boundary layer thickness, after- body separation, and erosslow separation. In addition to the traditional bow trip wire Nikolaev and Lebedeva, (1980) have applied turbulence stimtlation on other parts of the underwater hull. Results reported by Dand (1989) indieate that the ‘model is normally more directionally stable than the Ship. Martinuseen and Linnerud (1981), however, in ficate that hy using large models, preditions of sta: bility characteristies ean be reasonably accurate for (quite different bull forms. The free-running: model in ‘pen water has long been a too! in the analysis of maneuverability. The series type test results of Shiba 960) have been very instructive as to the effect of basic dimensions and rudder area on maneuvering characteristics, Because of the numerous scaling and other difeul- ties noted above, and because free-running model tests Yield terminal results only providing Ite insight into the many individual factors, and because large ma- heuvering basins are required, alteriate procedures fmploying computer simulation in conjunction with taptive model test results have beea developed and Will be presented in Section 8 Before leaving the free-running model, its utility when operated ina hydraulic model ofa waterway wil next be introduced. 72. Hydroule Models. Models of harbor and ves- sel waterway systems accurately modeling hydraulic ows ean be constructed and then free running models 27 piloted remotely through them to determine safety and ffieieney of the operation. Fig. 29 shows one such hhydraulie model test seenario, the Port of Bemhaven, Netherlands (Panel 1:10, 1979). in ths particular ease the channel bend was dever mined, after survey, to present the most ertieal prob- Tem in the maneuvering of containerships. While the existing channel was wide enough to accommodate a large containership, the combined effects of the cur ing mancuver and wind and current restricted entry time toa window of only two hours a day. And even ‘with this restraint, entry was stil judged to be risky by the Rotterdam pilots "A hydraulic model was set up at MARIN of the Bembaven waterway configuration and many diferent large proposed containership models were run through the turning maneuver, The rate of groundings was 30 percent of the transits, reinforcing the pilots’ fears When proposed dredging to ease the bend was tried in the hydraulic model, groundings were reduced to zero with proper tug assistance (tug usage was sim lated using small fans mounted on the model. Fig. 29 shows a sample trajectory resulting from the Bemhaven study. As a result of the project, the Rotterdam Port Authority authorized an expensive ceasing of the channel bend entering Femhaven. While the hyraufic modeling of the waterway ean bbe made quite accurate reflecting correctly the flows ‘of water, sealing problems stil exist andthe dificulties mentioned in the previous sections apply. Scales are usually small (1 to 100 is common) beesuse of the expense of building waterway models, Pilot or helmsman control is from a bird's eye po- sition which provides 4 better understanding of what is happening, but in & compensating manner, move ‘ments take place many times faster than in real life. Section 8 Nonlinear Equations of Motion and Captive Model Tests 8.1. Nonlinear Equations of Motion. Captive mode! teats and associated simulation studies using nonlinear fquations are the most powerful and flexible means fvailable today for predicting controlability. This ap- proach may be initially more costly than a free-running ‘model test program, but once the required hydrody- ‘nami coeffients are determined from the model test Gata wide variety of more accurately predicted ma neuvers and ship operations ean be rapidly. and eco- romieally simulated, with the effects of environment, Control systems, and external forees readily studied Linear theory as discussed in prior sections is useful foranalyzing the influence of ship features on controls: fixed stability andon the turning ability of directionally stable ships inthe linear range. Captive model tests ‘can be used to develop coeficients for these equations. However, as previously noted, near theory fails to prediet accurately the characteristics of the ight mar heuvere that most ships are capable of performing, land it cannot predict the maneuvers of directionally tanatable ships "There is no completely analytical procedure available to this date (1989) for predicting the characteristics of these nonlinear maneuvers. As a result, current com pputer-aided techniques utilize the experimental results From eaptive model tests, with the equations of motion expanded to inelude significant nonlinear and coupled terms, ‘A variety of different approaches to developing set of nonlinear equations of motion exist, ranging218 from using wing theory to applying a Taylor's series expansion to force and moment parameters. A "Cat alog of Existing Mathematical Models” (Hagen, 1983) provides a primer on nonlinear models and reporting: Of coefficients and data, The works of Abkowits (1964, 1965), Strom Tejsen (1965), Bda and Crane (1965), No: rin (1971), Goodman and Gertler (1976), and others should also’ be reviewed. Pedyayevsky (1964) devel: ‘oped a modular system based on physical relationships and the use of wing theory, and more recently, the Japanese Mathematical Modeling Group” (MG) ‘through Kose (1982) and others has developed similar modular model (Section 16 addresses these develop. ments and the modular approach) ‘The Abkowitz and Strom Tejsen Taylor's expansion approach is herewith presented to give an understand: ing of the development of this popular non-modular ‘model, The approach is based on a restatement of Equation (6) to include rudder angle a® follows: Fl, 9, Bed (30) It is assumed in Equation (80) that the only impor- tant forees and moments acting’ on the ship induced by the rudder are those due to rudder deflection, By ‘and thatthe forees and moments produced on the ship sa result of By and 5, are negligible. ‘The complete Taylor expansion of Equation (30) wi terms up to the third order is as follows for X with similar expressions for Y and NV. (As noted in Section 6.7, tight maneuvers involve Jarge speed losses; hence, consideration ofthe A-equs- tion constitutes vital part of this section whereas it ‘ould legitimately be neglected in the earlier consi¢: erations of linear maneuvers) Xe ER Kee tet ke ++ XE + Xba) + bout Ht oo + Kabel + 2X, bur BN Bur + oo + BX uhBe] FE abi + Xi tb occ Kady 4 BX bue + BK Bur + oo + AM HOt + 6X,.Burr + 6Xbut Hes + OX a7] on where X°is the force in the a-direction atthe equilib: ium condition, that is, uy = V PRINCIPLES OF NAVAL ARCHITECTURE pyeouae ‘Terms higher than third order are not included in Equation (31) because experience has shown that ac- ‘euraey isnot significantly improved by their inclusion, Furthermore, practical limitations of measurement techniques and the state of refinement of present the: ‘ory do not justify the inclusion of higher order terms ‘AS a eorisequenee of the geometrical symmetry of ships about the z=plane, the relationship between X tnd 1, for example, must correspond in general form tone of the three relationships shown in Fig, 80. The feature common to all three relationships is that they are symmetrical about the ordinate X- If the relation. ship between X and v as depicted by curves 1 or 3 0f ig. 30 is to be expressed as an expansion in powers of » beyond the first power, then only the even powers Of v can appear in the expression and the coefhcients fof the odd powers must be vero, That is, X is an even fanetion of » which takes the form an + at + ath + @ V/24X,,, ete, from Equation (31) Again, as a result of symmetry about the zz-plane, Abkowite (1964) shows that ¥ is also an even function of 7, Be, and F, that i, Xe) = bye? + bart + bye + Xo) = e:be? + cabs! + Bye + a co) It follows from the provious analysis that cross ‘coupled terms in Equation (31) such as X,,v0uCONTROLLABILITY XogrBu, Xy6xBa, Xog"BU, Xoyr*B4, and $0 on, in ‘olving odd powers of, r, and 8, are also zero, How: ter, eros-coupled terms such a5 Xp, 2° Bu, Nn? Bs Nuebs’Bu, and so on, are nonzero because they involve ven powers of v,r, ard 5, Also, terms such as X,,07, Kevae Xa?Bq, Xev0rBu, and X.,, 78,01 are nonzero, ‘cease they involve evenpowered products of vr, and By (eee also Section 86), In contrast to X, the expressions for ¥ and NV are odd functions of 0, 7. By 8, and Fs that is, only the Coefficients of the terms in the expansion’ with odd poveers are non-zero; those with even powers are zero. Bud funetions are lke those shown in Figs. 6, 8,11, tnd 13 where in all eases the graph of the function is fellected about the orn. The expansion of Yor N as Sfametion of ers Bye oF Fi typically as follows: co) 5a) ¥ (0) = dyvt dye! + dv? + Yq) = ep + eb + cbs! + Although superficially it appears that there should ben correspondence tetween the relationship of ¥ to ‘shown in Equation (5) and the relationship of X to icin reality they are rasly different, for several reas ‘Sons, Oneis that the equilibrium value ofr, designated tin Section 3 s taken as zero. (Any asymmetry due {propeller rotation is neglected for this restricted purpose but i taken account of later in Equations (37) and (3) "The equilibrium vaue of u, 1 is not zero but is equal to the ship veleity, V- Another reason is that the force is the component along the axis of the Aifference between to oppositely directed forces, hammely, the ship resistance and the propeller thrust, Whereas the Y-foree isthe component of a direct hy” Urodynamic force. Por these reasons and others, X is helther an odd nor an even funetion of u but rather its expansion includes all powers of 61. “Additional terms of the nonlinear equations can be eliminated by considering the nature of acceleration forces, Abxowits (1964) states that no second or higher trder acceleration terms ean be expected, on the as- Sumption that there s no significant interaction be fhween viseous and inetia properties of the fluid and that acceleration foreesealeulsted from potential the ‘ory when applied to submerged bodies give linear terms, Hence all terme such as Xi, Xi", Xp, Xeati and #9 on, of Equation (81) ae taken as zero Since X, and X, aze aso zero because of symmetry, the only acceleration derivative that is not taken a8 zero in the nonlinear equation of motion for X is X,, ‘which i also the only acceleration derivative that sp- pears in the near equation for X (10) Combining the noslinear Taylor expansion for X, ‘Bquation (1) with the dynamie response terms of the equation, Equation (6), and taking all of the preced> 29 ing considerations into account, the equation for X ‘becomes: (a = Ay = fw 27 Bd 66) where ‘fle 0.7 Bx) XP MB + KBE + Xa Bi 4 LX + Art + Nad? 4 AX giPBn +X tO + AXE £ UX, + Mor + Xy0Be + XarBe + X,atrbe 4 Kitab + Nig 7B eB ‘The relationship between Yor Nand bu corresponds to that shown in curve 2 of Fig. 80 for X versus. ‘That i, beeause of symmetry about the 2=>plane, YW) S'N(u) =O and the derivatives Yo, Foss Vasur Yer Na, News Noes and NY, are all 270, ‘As Stated earlier, ¥ and N are odd functions of 1 by 6, and 7. It follows that all the crosscoupled {erm in the complete Taylor expansion of ¥ and N {nvolving even powers or even-powered products of v, tbe by and Fare zero, Thus Vogt8%, Yor Bu, Yoh, Ygtby, YarBp, YewrBx, You 00xbu, ete, and similar terms’ for'N, are all zero. ‘The Y-forve and Nmoment induced by the rotation ‘of a single propeller or by unirotating multiple pro pellers, a 7 = Be ~ 0, identified as Y* and N* in equation (26) must, of course, be ineluded in the non- linear equations for ¥ and N. in addition, since ¥* and No are likely to be speed dependent, the following {terms are also taken as nonsero in the Taylor expan: ‘The ¥-force and N-moment induced by propeller ro- tation atv # 0, also diseussed in Section 17.9 (see Fig 253), as well ae their speed dependency, are included jn the following nonzero terms inthe Taylor expansion ¥,06u ¥.0(6uF Nobu Nave) Following Abkowite's reasoning as noted inthe ds cussion of Equation (81) the only acceleration deriva fives not taken as zero in the nonlinear expansion of Y and NV are those appearing in the linear equations cof motion. These are ¥,, ¥,, Ny and N,- Combining the third-order Taylor expansions for Y and N [similar to that shown for Xin GI], with the ‘dynamie response terms of the Y'and Nequations of Equation (6) and taking all of the preceding. consi20 erations into account, the nonlinear equations of mo- tion for Vand Ware's follows Equation G-Me- Khun a9 en N-Bquation: SN HU NYE H Alun, rs Be) (88) where Lilt 07, 8p) = YY + Vit6u + YB! + Yow + Mat? +E ni + Vat + Yiye6u + MotB + OF, = an) 4 ar + Et + WarOet + Var6u + Waar + ¥8y + Mabe + YBa + YL Bar? + Y8p8u + IY. 5y5u! + ¥,a0rBy Alt, 2,7, 89) = NY + NOBU + NY Bu! + Nw AN at? + IN OF Mg yl + Nv FIN vou + Nor +N FAM grt + Nig r8* + Nr + Naar Bl + Nby + Naud + INonBevt + IN Bar + NeByb + IN Bgbu! + Neary An equation similar to Equation (38) could also be Aeveloped for the roll moment, K, which could be used to solve for the heel angle, as a function of time. Equations (36), (87) and (8) ean be solved simulta- neously for the accelerations i, i, and f, as follows: Liles 7, Be) @= x) =U = Nifilus 11, Be) + Yo fils 0 Be) @~ FUN MY, (= Vfl wn Ba) + Nila 7 80) = @— FN = NI NY, 9) ‘These solutions can be rewritten in the form: PRINCIPLES OF NAVAL ARCHITECTURE 8 oo 0 000) dy GF = alt we), 1, re, (0) (0) a = alt ule, (0), r@, (01 where 1(f), o(@, r(t), and 6,(0) are the instantaneous values of u,v, 7, and By at any time Enuation (4b) i a set of three fistorder differential equations for which approximate numerical solutions fare readily obtained on a digital eomputer. The key to the numerical solution is that values of w, v, and rat time ¢ + b¢ are obtained from knowielge of the values of t a, 7, and Bg at time fusing a simple firstorder expansion; that is, u(t + Bt) = u(t) + Bue) (e+ Bt) = v(t) + Bei) on rt + 80) = rte) + atte) ‘This method is found to give adequste accuracy for the present type of differential equations because the accelerations ui, , and + vary but sbwly with time, Owing to the large mass or inertia of x ship compared to the relatively small forces and moments produced by its control surface. Any desired accuracy of the solutions ean be obtained with a computer by using smaller time intervals 8¢ ‘The mathematical model has been developed in di: ‘mensional form. The equations are equally valid in hondimensional form with the stipulaion that the ve- locity used for nondimensionalization should be the velocity at any time, ¢ but not the initial velocity. For further simplification, the nondimensionalizing veloc ity in the nonlinear equations is takes as u(@) rather than VO) One reason cited in this chapter for non-dimension- alization is that the nondimensional derivatives are independent of speed. The extent to which this a sumption js not true for nonlinear mareuvers is taken account of in Equations (38), (37), anc (38) by the im clusion of such terms as Yn, Yoru, Your Pra Yer ‘and s0 on, which represent the changes in the imensional derivatives YY, Y; with speed. ‘Assuming that a full set of hydrodynamic. coeff: cients (Xi, Xign Kew, ¥, ete) is available, and that the rudder defoction’ 8,"ts defined as a function of time, the first step in the caleutation of the trajectory fof a ship would be to set the values of ty % rand 8y at time ¢ = 0. In the most usual ease a, and By atCONTROLLABILITY {= O would be eroand v would be equal to 1, Having fone this, 9, and Fean be ealeuated from Equations (89) and the new velocities at time ¢ = B¢ can be ob- tained from Equations (41). The process is then re peated using the new values for 1, 1, r, and By in Biquations (09) and so on. The values of the velocities ata time tare thus obtained from © alr) u(t) = 100) + vig) = v0) +E v1) ot (2) (0) + & +(e) 6 rb where w(0), 1(0), and r(0) are the values of w, 1, and Dat t 0 and’y represents intermediate values of time (between time 0 and time t — 6!) al which the fcvelerations s(¢), 6(0, and #0) are determined. ‘The instantaneous values of the linear velocities of the ship relative to earth axes (which are needed to alc reer) ised af rea ship sees are ‘btained from Equation (@) re-expressed as L(t) = ult) c08 Y(t) — v(t) in ult) = ult) cos vie) — U0) WO gg) alt) = ult) sin Ye) + v(t) os HO) where ie(t) and f(t) are the components of the in Stantanedus resultant velocity of the origin, O, of the thgalong a Bred et of ert nen a, epee tively. "The instantaneous coordinates of tie path of the origin of the ship (0) and y(t) relative to the fixed Set of earth axes and the orientation othe ship, W(t), fan then be obtained by integration of te last of Equa tions (42) and (49) These are as follows: Wo = HO + Fre) ae ws = aid + Zire 0 + (ule) + mG) sinyia) 8¢ (40) aralt) = ses) +S {Eute) + (cos He) — via) sin yin} 8¢ ‘There remains the question of defning the rudder deflection asa function of time, [tis assumed that the 221 rudder moves with a constant rate of deflection, 5, etermined in accordance with condition 2 of Section ‘9, and that there is a fixed time lag between the instant that radder deflection is ordered and the instant that the rudder begins to move (see item (a) of Section Su}.A rudder dolction up toa coreain meximum angle 8, would be simulated in a computer program as fol- Jows: ExG) Blt) unt € > tag + fe then By(t) = By(l) + rate (t= f6~ fag) until 840) = Bon then B,(0) = Bowe [A rudder function of this type gives a very close approximation to the actual time history of a ship's rudder when @ maneuver is ordered from the bridge. 182 Captive Medel Tests. Captive model tests in tanks are now carried out using a planar motion mech Sniam (PMM) or a rotating arm. In either case the model is tested over a suitable range of important Variables such as drift angle, yaw rate, sway accel ‘ration, yaw acceleration, propeller RPM and rudder ‘angle, and the results are analyzed to obtain the hy- Grodynamie coefficients required in the equations of tmotion. Development of the linear coeficients wil be sddressed fist or design of a control surface, knowledge of the lift, drag and center ofpressure location as a function fof angle of attack, velocity, and controlsurface con- figuration as given in Section 1 is adequate for most Dractical problems, Knowledge of the forces and mo- nents generated by eontrobsurface rate of deflection, ’ and angular aceeleration, 8, are only occasionally important to the design of the steering engine that defects the control surface and rarely, if ever, to the ‘motion of the ship vs. the control-surface system a8, & whole, However, it was found in Equation (Léa) that {fo determine whether a ship is stable or unstable in Straightline motion, one must know not only the forees land moments generated by angle of attack on the ship, but also those generated by angular velocity. In ad- ition to these, the forces and moments generated by linear and angular acceleration must be known in order to determine the magnitude of the stability indexes, Equation (11) or to sompite the trajectory of @ ship from the equations of motion. "The experimental techniques necessary to measure the signieant forees and moments generated by a Ship's hull are much more elaborate and sophisticated {than those necessary to measure the significant con ‘rol'surface forces. Only in the ease of the determi. nation of the velocity dependent derivatives of the hull [E'the experimental technique similar in principle to that used to determine controlsurface forees and mo:‘ments as used forthis purpose ina modern wind tunnel (Whieker, 1958), cr by Joeseel in the river Loire in 1873 (Wan Larimeren,et al 1948), 83. Stroight-tine Tests In @ Towing Tank. ‘Tho ve locity dependent derivatives Y. and N-of aship at any draft and trim ca be determined from measurements fon a model of the ship, ballasted to 2 geometrically Similar draft and trim, towed in conventional towing tank ata eonstant velocity, V, corresponding toa given ship Froude nunber, at various angles of attack, 8, to the model path, Fig. 31 indicates the orientation of the model in the ‘owing tank. From this orientation it is seen that a transverse velocity component, vi pro duced along, the yaxis such that y= -Vaing where the negative sign arises because of the sign convention adopted inthis chapter, Fig. 2 ‘A dynamometer atthe origin, O, measures the force Yand the momert,V experienced by the model at each value of tested. These measurements are then plot ted as a function of » (Fig. 10) and the slopes of the curves taken at ¢ = 0, give numerical values for the ‘lerivatives ¥, ani N, for the model. These values can be reduced to'nondimensional form by dividing hy the proper combination, given in Section 3.4, of model length L, model speed V, and towing-tank water den sity, p. The dimessional ship values of the derivatives tan then be obtained by multiplying: the nondimen- Sonal derivatives by the same respective combinations Of ship length, stip speed, and seawater density With reference to Equation (10), itis not realy nec. essary that the origin, and hence the dynamometer, be located at the center of gravity of model. The results are independent of the location of G- Rather itis most fonvenient that the origin and the dynamometer be Toeated at 0H #0 chat rq 0. If the dynamometer is not located at the derivative N, should be corrected so that it applies to O at ‘As described ir Section (7), the propeller will usu: aly exert an important influence on the hydrodynamic derivatives, Therefore, model tests to determine these derivatives shoul! be conducted with propellers. op trating, preferably at the ship propulsion point. Also, PRINCIPLES OF NAVAL ARCHITECTURE since the undeflested rudder contributes signifleantly to the derivatives the model tests should also inelude the rudder in the amidship postion ‘The technique just described ean also be used to determine the control derivatives Y, and 4. If in Pg. 41 the model were oriented with zero angle of attack, {8 to the flow but the model were towed down the tank ft various values of rudder angle ,, the dynamometer measurements would determine the force Y and the ‘moment Nasa funetion of rudder angle, Plots of these fpainst rudder angle would thus indicate the values Of the derivatives Y, and 1. In addition to these im portant data, comparison of the values of ¥, and Nib, ‘obtained by this means at any given rudder angle with the values obtained from Equation (120) at By = 0 Using isolated eontrolsurface lift and drag data would indicate the magnitude of the interaction effects aris. ing beeause ofthe elose promixity of the rudder to the bul ‘Straightline tests in & towing tank ean also be used to determine the cross-coupling effect of v on Y, and N,and of 6, on Y, and W,. While such information is inadmissible within the context ofthe linear theory, it ‘i important for the nonlinear theory. Also for the purpose of this theory, knowledge of the shape of the Y versus 8, ¥ versus 2, N’versus 6 and V versus © ceurves at large values of vand 8, willbe of importance. ‘84. Rototing-Arm Technique. To measure the ro tary derivatives Yrand Non a model, 8 special type fof towing tank and apparatus called a rotating-arm facility is oeeasionally employed. Tn this facility, an angular velocity is imposed on the model by fixing it to the end of a radial arm and rotating the arm about ‘a vertial axis fixed in the tank as seen in Fig. $2, The model is oriented with its axis and zaxis normal to BBv00 EanesCconTRowaeiuTy the radial arm and ite attached to the arm preferably atthe model's midlength, As a result of this particular Orientation, as the model revolves about the tank axis, rotates at rate r while its transverse veloety compo rent vig at all Limes zero (yaw angle of attack = 0}, and its axial veloety component 1 is identical to iis linear speed, The model ie rotated at a constant linear speed at various radi, R, and a dynamometer measures the force Vand the moment N acting on the model, Since the angular velocity ris given by the only way to vary rat constant linear speed is to vary R Typical plots of the resulting measurements {after model inertial effects sire deducted) are shown in Fig. 12 and the derivatives ¥, and N, are obtained by evaluating the slopes at 7 ~ 0. Because of ship 3 the values of Y, and N, at the negative ‘alues of r shown in Fig. 12 are a feflection of their Values at positive r but with opposite sign "The model must be ballasted s0 that it floats at the proper draft and trim. Since the otating-arm tests are Conducted with ¢ ~ 0, the results of rotating-arm tests fre independent of the model radius of gyration Ifthe Tatial arm shown in Fig. 32 8 attached to the model tits midlength, distance 2, from the longitudinal Center of gravity of the model, then the dynamometer measurements will yield values for (N, ~ Azcu,) and 2 (Y, = su), Since values of a, x5 and uy are known for the model, the dynamometer measurements ean be Used to determine the bydrodynamie derwatives Nand Y.. Values of these hydrodynamic derivatives forthe fulbseate ship can be determined from thei nondi mnensional values. For the same reasons as indicated in the model tests for N, and Y., the model used in the rotating-arm tests should have the rudder in the fmidship position and the propellers operating atthe Ship propulsion point. Thoomtract te the results of the towing tank est for'Y and N, the tesults of the rotatingearm tests are dependent on the location of the eenter of gravity of the model. If, however, the procedure described in the preceding paragraph i followed, the model center bf geavity need not be located atthe same geometrical position as the G of the ship. The later postin is, of Course, dependent on the condition of loading of the Ship and needs to be knovn for insertion in the equa: {ons of motion of the full-sale ship. Ifthe radial arm shown in Fig. 22 is not attached to the model at the ‘midlength, the results should be coprected tothe mid Tength position ‘The rotating arm facility ean also be used io deter mine ¥, and N. as well a Y, and N,. This i accom plished by towing the model ata variety of values of B.for each revalie, not just at = 0 as in Fig. 22 ‘These tests should inchude both plus and minus values of r By eross plotting the values of ¥, and N, obtained St cach rvalue agaist 7, the values of Y, and N, at 1 can be obtained, Within the range of values ofcoy tm |x Fo. adel stp fr poem te Anon, 1968 + in which linear theory is applicable, this erass plot Should show that ¥. and NY are substantially indepen- dent of 7: Values of the control derivatives ¥, and N, ‘ean also be obtained from rotating arm tests in a mar her exactly analogous to that just deseribed for Y, and N, Unfortunately values of ¥,.N,, ¥-and N, obtained inthis manner do not always agree’ with values ob- tained from straightiine tests For the purposes of the nonlinear theory diseussed in Sections 8.1 and 86, the rotating-arm facility can provide not only values of the hydrodynamic forces land moments at large values of 7, x, and 5, but also information on the cross coupling between these three parameters. Furthermore, with sufficient components the dynamometer used in the rotating-arm experi ments ean also be used to measure the ‘force and the roll moment, K, 8 a function of , v and 8. These are needed with nonlinear theory to predict speed loss and heel during maneuvers. If the model at the end of the rotating arm is tested at various heel angles, 4, as well a at various values of 7, and By the nonlinear effects of heel and the cross-coupling effects between @, rv and 6, can be obtained. ‘A major drawback’ associated with rotating-arm tests is that they require a specialized faelity of sub- stantial size; they cannot be conducted in the long harrow tank conventionally used for resistance and propulsion testing. There are only a few rotating-arm facilities in the world. The largest in this country is ut the David Taylor Research Center, Carderock, Md, ‘witha diameter of 8 m (260 ft). See Fig. 38 for a view of the apparatus at Davidson Laboratory. Other problems associated with the rotating-arm technique ace: (a) "The model must be accelerated and data ob: tained within a single revolution. Otherwise, the model will be running in its own sake and its velocity with respect to the uid will not be accurately known. (8) In order to obtain values of the derivatives ¥,, N,, ¥and N, at'r = 0, data at small values of r are necessary. This means that the ratio ofthe radius of turn, R, to the model length, Z, should be large. For large models, a large facility’ is required. Smaller models may tse a smaller tank, but models too small will lead to scale effets in the ship prediction, ‘85 Planor Motion Machoniam ("PM") Tech- que. In order to avoid the large expense of ro. PRINCIPLES OF NAVAL ARCHITECTURE {ating arn facility, a device known as a Planar Motion Mechanism (PMM) was developed for use in the com ventional ong and narrow towing tank to measure the Velocty-dependent derwatives, Y, and N, the rotary derivatives ¥, and N,, as well as the acceleration de rivatives N,, ¥,, ¥, and 4, Ths apparatus, deseribed Dy Gertler (1988) and Goodiman (1360), was developed atthe David Taylor Research Center to determine sub- marine derivatives in the horizontal and. vertical planes, Subsequently by similar devices, deseribed by StromTessen (1964), Paulling and Sibyl, (1962), and Van Leeuwen (1964), were developed elsewhere for surface ship models. TBasiealy the PMM consists of two oscillators, one ‘of which produces a transverse oscillation at the bow land the other a tansverse oseillation atthe stern while the model moves down the towing tank at a constant velocity u, 38 measured along the centerline of the towing tank. The subseript zero is used here to des- Jgnate the velocity component measured along the sis fixed in the earth, Le, in the towing tank. There. fore, the magnitude ofthe resultant velocity vector V of the motel is not strictly speaking a constant in the planar motion tests. However, for the small motions ‘admissible within linear theory, V~ 1 ‘The PNM, however, is more than a mechanical means of osillating model in a prescribed manner. ‘The transiucers used to measure the forces on the model, ani the special instrumentation required for the proper resolution of the forees, are vital parts of the PNM, “The setup for planar motion tests is shown in Fig, 84. Near the bow a point 8, located a distance x, for- ‘ward of the origin O, (preferably taken at (E) ib o5- cillated transversely with a small amplitude ay and at angular fequeney w. Point $ near the stern at an ‘Mentical distance, 2, aft of the origin is oscillated traversely with the same amplitude, a, and the same frequency, «2 The phasing of the ostillaion of the stern relative to that of the bow ean be adjusted and is indicated by the phase angle ¢. If e = 0, the model remoo oF oxcutanon. 1+ al cca — Owe ‘tawoon denen tere -p tay tan ant OooCONTROLLABILITY experiences an oscillation in pure sway with zero yav, fas shown in Fig. 25. The sway oscillation is of the form: Bea Y= 0,005 ot dy - x a0 sin wt «) a . a ayo 08 wt ‘Two dynamometers located at B and S in Fig. 34 measure the oselatory Y-forees experienced by the model ag a result of its swaying motion. These are the forces, Y, and Yo Ax shown in Equation (45), the ‘elocity v's 2 sine function, 90 deg out of phase with the displacement y, (ory) and the acceleration, F, both ‘of which are cosine functions. Hence the messure- ments of Y, and Y, taken when the time variable has Yalues 90 deg out of phase with the displacement, yu are forces arising from the effets of v and not from the effects of ¢ since the lator is zero at these times ‘This is shown in Fig. $5. The velocity dependent de rivatives ¥, and N, are then obiained from the follow Ing relationships: a¥ 4 Wada t Waa y= Sh a 2 Uo 5 meme © BN, [eka = Wo Nem a= * ae where the subseript “out” refers to the amplitudes of Yp and Y, taken 90 deg out of phase with the dis- ‘and N, the postive sign in front of the ‘expression should always be associated with the case ‘when Yp is positive and the negative sign when Yy is negative. This is necessary be-ause the amplitude a, fnd the frequency w are always taken as positive va: ns ues and, as was shown in Section 42, Y, for example Is always negative, Because of the oscliatory nature fof both Yad Ys, they can be either positive ot neg- tive, hence the alternate sign designation is essential ‘The sign of N, which ean be either postive or ne tive, mil be determined from Equation (46) according to whether the absolute magnitude of Yo is greater oF fess than Yq. [f Yp 8 greater, N, will be negative and vice versa. With the PMM, the proper signs are ob- tained directly etabishing gten sgn convention {tally in the elestronie moaeurimg eqnipment ‘in order to obtain the coefficients of the linear ac- celeration terms of Bquation (10), the in-phase ampli: fhdes of ¥, and ¥, must be measured, siee, as shown in Fig, 36, these correspond to the times when 9 Is a maximum and 0 0, The relationships for these coef ficients are Wada + We [as = Wale, Fae ‘where the eubseript “in” refers to the amplitudes of Ypand ¥, taken in phase with the displacement, y. ‘The remarks concerning the use ofthe = signs made with regard to Equation (45) apply also to Equation (4) as well as to Equation (49) to be developed sub- sequently Tn order to obtain the rotary derivatives ¥, and N, from planar motion tests, the measurements must be made when # = 0, » = 0, and ¢ = 0. Similarly for Y, and N,, the metsurements must be taken when 0, = 0, and 5 = 0. In order to impose an angular velocity and an angular aeccleration on the body vith ‘and # both equal to zero, the model must be towed down the tank with the centerline of the model always tangent to ils path, Fig. 37. This means that the ‘component of the resultant velocity, Vs always zero, sie sna the cae in the stution shown in Fg 38 However, the velocity, which strictly speaking is hel tonstant in the experiment depicted by Fig. 21, is uy fast as it was in Pig. 85 y-8 “ N, ~ a0 syn afenete ate {ieee6 Whereas the motion shown in Fig. 3 is pure sway with zero jaw, that shown in Fg, 37 is pure yaw with zero sway, since sway is defined asa translation along the y-axis fixed in the model. Since 7 is set equal to zero forall time in Fig. 87, there is no motion along. the paris fixed in the model only along the 8X8 fixed in the towing tank and shown as the ordinate of Fig. 37. Since the y-axis is constantly oscillating in direction in Fig. 87, the ordinate of that figure ean be identified only’ as the yyaxis whereas in Fig, 35, the ordinate could be identified as both the yy and axes, Inorderto achieve the kind of motion shown in Fig 37, thas been shown by Goodman (1960) that the phase angle e between the bow and stern oteillators (see Fig 34) must satisfy the condition: where 2 in both cases isthe distance from the origin Of the model to each oscillator. ‘The yaw oscillation shown in Fig. $7 is of the form: Y= —Wy c05 (at ~ €/2) Ga r= Hyosin (wr 2) 48) B= F = + Y0" cos (at ~ /2) where vs is the amplitude of the yaw oscillation, In Eaquation (48), is out of phase with y and ris in phase with J. Therefore, the amplitudes of Y, and Y. mea- sured $0 deg out of phase with y will determine the foree and moment due to rotation, r, and the ampli tudes of ¥, and ¥, measured in phase with d will determine the force and moment due to angular ac celeration, ¢. The force and moment derivatives with respect to r and Fare then expressed at follows: = Tada + Odo Y= au = 8 He y= Arou, = + Weber + (Pda) wero — ag, = tlle + 0% ¥, = beg = “Wah 2a)s = dad PRINCIPLES OF NAVAL ARCHITECTURE where the amplitude ofthe yaw oxilton vy 38 well {5 the frequetey w are alwage taken as postive, The lat equation shove thatthe results ofthe PMA tests are dependent on the model mass moment of inertia, and, hence ite radius of gyration, a8 Well 3 ‘on the distance 1. The procedare described in Section i for expanding the results of rotatingearm tests to the ful aeale in the event that zo/L. fm no the same between model and chip applies tothe results of the PAM tests nthe event thit neither nor the radius of gyration is scaled properly, Thus, the longitudinal sadits of gyration of the model need not be propor- tional to that of the ship. The model must, ofcourse, be ballasted to the peoper draft and trim and it should be self propelled at the ship propulsion point withthe rudder, n the amidship position, included. ‘Caution is necessary in asng the rerulis of planar smodons tests. Since the ship model iat the water urface, is esclatory motions create waves whose properties dopend onthe frequeney of ave gener {ion By esting at various frequencies withthe PMT the frequency dependence of the derivatives can be determined. Hoseever, for many problems in mane ering, we ate most interested in zero or very low Trequencies, For example, the maneuver of ship foing into a turn fat ero frequeney. To determine {fe values ofthe derivatives at zero frequency from planar moton test, itis necessary to pot the deri tives against frequency and to extrapolate to 70 frequency; henee the desirability of testing at low frojuences, However, for many problems, including transient maneuvers and motions in rough Seas, We fare extremely interested in the extent to which the inydrodynamie derivatives are in fact frequency de pendent For tis information the planar motion tests fre indiapensable. Van Leeuwen (1964) presents ex tensive plots of the hydrodynamic derivatives of a Se fies 00, C,_= 0.70, model_as 1 function of both frequency of oseaton and ship speed. Tt shoul! be noted tha the values ofthe derivatives determined either in straightlin tests in the towing tank or on & rotatingarm facility are the values at dero frequeneyofoxlaton These should correspond to the values obtained Trom planar motion teste ex {tapolated to zero frequency ‘he PHM, lke the rotating arm, ean alo be used to devermine nonlinear and eroxseoupling effects ‘hich re essential forthe pradition of nonlinear mae heuvers, Sample results from PMM experiments re Inting ¥ and to simulaneousy large values of 8 and dy are shown in Fig, 38 and related tothe nonkinear equations of motion in Section 8.6. Similar resulta can tlso be obtained from the PM not only relating,” fand N'to ras wel as and By, but also relating the longitudinal fore to. By, andy. ‘The PMM like the rotating arm ean also be ust to determine the rolling Tmoment, a8 a funetion of, Bp, 7, and. “The use of model text data immediately suggestsCONTROWABIITY a7 1 ome ance a.m 266 2 : 20 30 ‘ stg oe the possibilty of seale effets as introduced in the liseussion of free-running models Section 7.2) If the Froude number is satisfied, the Reynolds number wil not be satisfied, However, in determining the Y-forces and Nimomentsliftandelrclation effects are involved. ‘and itis shown in Section 14.2 that there is very litle Seale effect on the slope of lift coefficient versus angle of attack. However, separation or breakdown of lift thoes occur at lower angles of attackeat lower Reynolds tiumber. For example Fig. 88 does show a decresse in the values of Y and N’at large valuss of 8, This may be, but isnot necessarily, evidence of stall. Ifthe cause is stall, and it suspected that sta in the full seale ‘would occur at larger values of By, the data of Fig 28 could be empirially corrected atlarge values of By before using the data for predicting nonlinear maneu vers. In this way, errors in predictions of maneuvering. characteristigs dueto prematurestalof the model may ‘be minimized. ‘86. Eveluction of the Coefficient of the Nonlinear Equation of Motion. All of the derivative coefficients ‘of Equations (36), (87), and (38) with the exception of Some of those involving 1 are best evaluated on the basis ‘of experimental data obtained from captive ‘model tests. For example, from experimental data like those given in Fig. 38 relating Y" and No 8, and all of the derivatives involving 8 and vin Equations (G7) and (38) may be determined, Similar experimental Gata for X as a function of 8s and 2 can also be used to determine all of the derivatives involvng 8p and in Equation (0) Starting with Yi; and Yi, the fist step is to eval- uate the coefficients ¢, and ¢, of the polynomial Equa tion (85a) so that it forms the best least-squares fit to the curve through the experimental data given in Fig 8{q) for f= 0. The coefhcients,¢, and e,, are related to ¥, and Pigg as follows through Equation (37): Yyand 6) = WY In an exactly analogous manner, the derivatives ¥, Yin Nis Ny, Ny and Ne may be determined ree olléting that v” ="— sin 8. Similarly, the derivatives Xn, Ans and Xy of Equation (86) can be determined by fitting Equations (82), (88), and (84) to experimental data relating X to By and B "The cross-coupled derivatives involving. 3, and v may also be determined from data like those given in Fig. 38 For example, to determine Yj, the first step is to measure (Y),-» from Fig. 38(0) for all given ‘Yalues of v7 = —sin B. The next step is to fia poly- homial to these values of Yj as a function of v. Be- ‘cause all values of Y;, are positive and also because ‘of symmetry port and starboard, Ys an even function vie Yuta) = Ao + halt +INCIPLES OF NAVAL ARCHITECTURE 28 Toble 2—Attesiment ofthe Coeticints in he Xquation Relative Importance of Goethe Value of Coefficients (see note I) Mariner Series 60, | StromTejsen Eda and Crane Variable Coefficient “Gags” Model 8.2/1 "G968) 969) ew Ho tats a B bu Toot a ae 0008 = a 8 ae oot = A * 089 = c e soars c & oe = c B ou = = gE rou = = g ibe - - z oe +0168 = D B or feos > D 7 ° = D trou — = zB toy = = E rendu = = eI a 0 = D B [Noms 1 The symbols used in these two eolumns have the following meanings seat ta ae Ee Lae crue Nowe 2k where ij = 1¥ uty = 0 for a third-order expansion For Vue, the first step ie to measure (V).-g at all given values of 6» on Fig. 88 and then proceed in the Same way as described for Yq, similarly for Ng and Ni The cross-coupled derivatives X', X'g and X, an also be similarly determined from experimental data relating X to 6, and 8. Fig. 38 applies only to r’ = 0. Data like these ean be obiained for all values of r’ from either rotating- farm or PMM tests. Given all these additional expert mental data, all the coefficients of Equations (37) and (G8) involving v, by, and r may be derwed. In addition, (W*Y of Equation 87) isthe value of Y" at 6 = Se O in Fig, 88a) and (Nis the value of NT at B = 5, in Fig. 380). If, in addition, the prediction of heel angle during maneuvers is desired, experimental data like that shown in Fig. 98 are required, relating the rlling mo- tment K to 0, By, rand the heel angle @, The feasibility of obtaining these data was discussed in Section 8.4 and 85. ‘The cocficients of the equation expressing the re- bat open f mportance, id deced of minor importance, ice deemed neglgile ane the coefficient ws rored i'dervativas are nondimensionaied on bare pL, T, snd V. lationship between X and u are best determined from the results of open-water propeller tests and ship re- sistance data rather than from either rotating-arm or PMM tests. The reasons for this are outlined in the following. ‘As long as a ship is moving in a straight line at constant speed, = u,, the following relationship is true: X=™--R=0 where T= propeller thrust thrust deduction coefficient R= total ship resistance ‘This equilibrium condition defines the initial propeller thrust and the corresponding propeller torque and rev lutions. However, as soon as a maneuver is initiated, ‘this equilibrium eondition is disturbed and the X-force varies as a function of speed. While ® and ¢ as func tions of speed may be obtained from ship resistanceCONTROUAE Table 2—Anseament of 29 Hy Coefficients in the Y-Equation Relative Importance ‘offen Value of Coeficints (see note) Mariner Series 60, Strom‘Tesen Blu and Crane Variable Coefficient “Claes” Model 5,2, (963) (ies) 6 @ =) soa +000 a B + @¥e~ Fs) -oos = a 7 OF, son ~0260 a B PW 1m ~215, c B mI ° us D> B ete 0.0008, = D we Ye = = zg vow OY, = = zB Play 0108 are a B Po We ° 00461 D B mt Ym 4328 ~nan0e a B a ° = D - = B = = B +0058 +0050 a B -ooms c +025 = ° ° = D 6 = c - = EB ° = D bones +000016 c B ° Ss D - = EB and selfpropulsion tests, the propeller thrust T'as a funtion of 1 is dependent on the type of power plant and the powerplant eettings that are maintained dur. ing a maneuver. For extmple, a diesel engine is essentially a con stant-torgue machine whereas a turbine isa constant power machine, When a diesel ship enters a maneuver there is therefore a decrease in RPM and hence in thrust. On a turbine-driven ship, any reduction in RPM js accompanied by an increase in torque; hence, the reduetion in thrust in a maneuver is less than on a diesel ship (Fig. 48) These considerations governing the relationship between T and w are best taken ac- count of by means of open-water propeller test data ‘8 shown in StromTejsen (1965) ‘Data from Strom-Tejsen and Chisltt (1964) for the ‘Mariner Clase model and from Eda and Crane (1965) for the Series 60, Model 5,1,1, corresponding, most closely to the Mariner Class model are summarized in ‘Tables 2, 8 and 4. Both models are described in Table 5, Am assestment of the importance of the various coefficients made in StromTejsen (1965) is also com tained in Table 4. The two references do not agree as to the importance of the various derivatives. ‘87, Sample Results of ‘Nenlineor Model-Predic- tions. ‘The overall precision of the nonlinear mode prediction technique ‘at developed by StromTejsen (1965) and utilizing PMM experimental data from Chis Jett and Strom Tejsen (1965) is shown in Figs. 89, 40, and 41, A comparison is made in these figures between the results of full-scale turning tests, zigzag and spiral ‘maneuvers conducted with a Mariner Class ship (Morse ‘and. Price, 1962) and the prediction of the identical maneuvers made in Chet and Strom Teen (965 ‘The overall agreement for all of these maneuvers is reasonably good except for the fact that the model prediction in Fig. 39 shows a small tendeney to turn [etter to starboard whereas the fullscale ship shows230 PRINCIPLES OF Ni Table 4—Aseriment of the Vale of Coetiients Mariner Variable Coefficient "Clas bre) —o00u78 GN)" Soon oN, “00555 mow, +035, oo ° 1} 0.00264 vot = r 0009 * 0 re 1158 me 0 a = by 0.0283 a “0.00482 Bat +01082 Burt ° oat ° yb - oy ° a -+D.o00s0 ox 0 ne = 4 tendency to turn better to port. The results of the Spiral manouvers of Fig. 40 confirm these small op- Fost tedencics sins at 2ro rer the mode! pe- ition inficates a tendency to turn to starboard, ‘whereas che ship tends to turn to port. This latter tendeney isin accord with Section 17.9, Predicted data on the speed loss during the turning and zigiag maneuvers shown in Figs. 29 and 41 are shown in Figs. 42 and 48. Three conditions are shown fon each fgure: that corresponding to constant RPM, as prevais, for example, in the usual free-running ‘model turning test; that corresponding to constant power, as for a turbinedriven ship; and that eorre- ‘ponding to constant torque, as fora diesel-driven ship ‘The RPMgpecd relationships for these three cases are shown in Fig. 44 Ttis clear from these data that fee running model tests seriously underprediet the speed los ina tarn for either a turbine or dieseldriven ship, ‘whereas the prediction method outlined in this section can accortmodate realistically the effects of diferent power plants on the speed loss. ‘The ability of the nonlinear model technique to pre dict the spiral maneuvers of unstable ships is shown in Fig. 45. As noted in Section 43, these maneuvers ‘eannot be predicted by linear theory. The derivatives IAVAL ARCHITECTURE Coefficients inthe N-Equation Relative Importance of Coefients (eee note I) Series 60, StromTejeen Eda and Crane Model sta "(5h 6s) = A +000 a B ooo a B os ce B 0306 D B = D = BE = EB ~ 0.0869 rf B ~o.0t D B 1a A B = D = E = E 0024 x B = c = c = D = ¢ = EB = D ~0.0008 c B D = B ‘of the marginally stable ship were obtained by de- fereasing the magnitude of Y., and Nof the stable ship by 10 pereent and increasing the magnitude of 1’, of the stable ship by 10 percent. Similarly, the Gerivatives of the unstable and very unstable ships correspond to 20 and 30 percent changes, respectively, from the stable ship. Also shown in Fig. 49 are the slopes of the r — By curves taken from Fig. 174 for these same ships. ‘the dimensional slopes, 37/08. shown are obtained from the nondimensional slopes, 9/98, given in Fig. 174 by multiplying the latter by VAL = 15 1.689/528 = 0.478, where 161m (628 10) is the length of the Mariner Class and 15 knots isthe speed at 8, ~ 0 deg. The slopes correlate nicely with the characteristics of the spiral maneuvers of the unst- able ships as predicted by the nonlinear model tee niques. ‘88 Comparison of Experimental Techniques ond Quesisteady Theory. Much testing and correlation work has been accomplished with froe running, and captive model tests over the years. Free running and captive model tests using a model of the Marinerlass fargo ship SS Compass Island for which extensive fullseale data are available (Morse and Price, 1961) have been collected and compared under the auspicesCONTROWABILITY 2 wo “4 rm ° - c e | i a TTT 5 ws oo Fo. 27 Compan © fing Sle am locale Who pi ‘iy ned Stn oppress ped (le nd Shenae, 961 1 OF chant OF HEADING, BECHER SECON f Canam of hn annie am ‘of the International Towing Tank Conference. Results ‘of an investigative program concentrating on turning tireles were reported at the Thirteenth Conference in $972 with favorable comparisons. The measurements of the linear foree derivatives by various model test towing tanks using the rotating arm and PMM testing. facilities have also been studied, ‘Concern for frequency dependence in PMM testing anda desire to provide accurate predictions of extreme Ianeuvers, resulted inthe development of larger am plitude PMMs (LAPMI) such as that used atthe Dar. {sh tank, Smite and Chislett (1974) Marine LAPMM results are compared with data from oblique towing tests and rotating arm tests (Smitt, 1975), Ths paper and a related one by Bia at the 1975 conference con ‘clude that comparable linear force data from carefully ‘conducted rotating arm and PMM tests are essentially ‘dential for the Mariner hull form. ‘Development in the early 1970's of the high block (Gq) tanker forms showing instability of flow around the ship stern caused concern for the effets of scale fon maneuverability predictions. Numerous tests were ‘consequently performed to verify the maneuvering ea pability of such forms. Nomoto and Fujii report on feats performed by several researchers addressing the seale problem (Burcher, 1975). Among the tests were those of Sato, etal (1978) who utilized «30 meter free: rrunning model in addition to 4 and 10 meter free rum ig models to investigate scale effects. ‘Trials of the Esso Osaka discussed in Section 13 have paralleled the Mariner trials as a standard for2m PRINCIPLES OF NAVAL ARCHITECTURE Le AND MEADE ANGLE 3 oeaneES imeerane tea b Lod tmaae Mm ‘Setoctons Bis| SRE Ee ars a naire $m at cy ttn narod it ob pe ee +11 {+ ‘types of power pions (StromTeien, 1965, {arate pu of pone an am toes P68 maneuvering prediction correlation of a full form in both deep and shallow water (Crane, 1979), Numerous model towing tanks have tested models of the Osaka ‘with good correlation. See "Model Test,” (1980), Dand and Hood, (1983) and others. ‘With regard to the validity of captive model tests, juasisteady state theory isthe basis upon which hy: tedynamie conditions of steady motions are sesumed to apply in transient conditions. Bureher (1975) and Nomoto (197) provide discussion ofthe sppliabity the theory paricalriy with reference to the ose latory derivatives obtained by PAM experiments Where che mations induced by rudder defen. eis Uiteait to Justify the quasesteadyaasumption oP purely theo‘atieal grounds but the methad has beer Sey sted in hydedymamiea and aerodynamics 808 in practical terms appears to work ‘revuse of captive model text data to develop coe fcents in motion equation Is assumed to be valid Scoring tothe guns steady state thor? along themotion s"slow.” An expression sch i EustonCONTROWABILITY 233 (22) for the forces due to sway motion, should be write 7§-—-——— ten with each variable as a function of time. Thus, | | Taser 2e| |e th eqns tan eng spi op iol mon indetriniatah. Dian, Durch =e tnd Price Gia) reported on vesigntone into the 83uq| Taegan || | time history effect in eaptive ship model testing. Nom» ‘oto (1975) concludes that, Gansta et ha- the 2025 wl I . 1

Você também pode gostar

• The Subtle Art of Not Giving a F*ck: A Counterintuitive Approach to Living a Good Life

  

  No Everand

  The Subtle Art of Not Giving a F*ck: A Counterintuitive Approach to Living a Good Life

  Mark Manson

  Nota: 4 de 5 estrelas

  4/5 (5794)
• The Gifts of Imperfection: Let Go of Who You Think You're Supposed to Be and Embrace Who You Are

  

  No Everand

  The Gifts of Imperfection: Let Go of Who You Think You're Supposed to Be and Embrace Who You Are

  Brené Brown

  Nota: 4 de 5 estrelas

  4/5 (1090)
• Never Split the Difference: Negotiating As If Your Life Depended On It

  

  No Everand

  Never Split the Difference: Negotiating As If Your Life Depended On It

  Chris Voss

  Nota: 4.5 de 5 estrelas

  4.5/5 (838)
• • Revistas
  • Podcasts
  • Partituras
• Principles: Life and Work

  

  No Everand

  Principles: Life and Work

  Ray Dalio

  Nota: 4 de 5 estrelas

  4/5 (599)
• The Glass Castle: A Memoir

  

  No Everand

  The Glass Castle: A Memoir

  Jeannette Walls

  Nota: 4.5 de 5 estrelas

  4.5/5 (1711)
• Sing, Unburied, Sing: A Novel

  

  No Everand

  Sing, Unburied, Sing: A Novel

  Jesmyn Ward

  Nota: 4 de 5 estrelas

  4/5 (1103)
• Hidden Figures: The American Dream and the Untold Story of the Black Women Mathematicians Who Helped Win the Space Race

  

  No Everand

  Hidden Figures: The American Dream and the Untold Story of the Black Women Mathematicians Who Helped Win the Space Race

  Margot Lee Shetterly

  Nota: 4 de 5 estrelas

  4/5 (894)
• Grit: The Power of Passion and Perseverance

  

  No Everand

  Grit: The Power of Passion and Perseverance

  Angela Duckworth

  Nota: 4 de 5 estrelas

  4/5 (587)
• Shoe Dog: A Memoir by the Creator of Nike

  

  No Everand

  Shoe Dog: A Memoir by the Creator of Nike

  Phil Knight

  Nota: 4.5 de 5 estrelas

  4.5/5 (537)
• The Perks of Being a Wallflower

  

  No Everand

  The Perks of Being a Wallflower

  Stephen Chbosky

  Nota: 4.5 de 5 estrelas

  4.5/5 (2099)
• Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future

  

  No Everand

  Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future

  Ashlee Vance

  Nota: 4.5 de 5 estrelas

  4.5/5 (474)
• The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers

  

  No Everand

  The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers

  Ben Horowitz

  Nota: 4.5 de 5 estrelas

  4.5/5 (344)
• Bad Feminist: Essays

  

  No Everand

  Bad Feminist: Essays

  Roxane Gay

  Nota: 4 de 5 estrelas

  4/5 (1015)
• The Outsider: A Novel

  

  No Everand

  The Outsider: A Novel

  Stephen King

  Nota: 4 de 5 estrelas

  4/5 (1839)
• Her Body and Other Parties: Stories

  

  No Everand

  Her Body and Other Parties: Stories

  Carmen Maria Machado

  Nota: 4 de 5 estrelas

  4/5 (821)
• The Sympathizer: A Novel (Pulitzer Prize for Fiction)

  

  No Everand

  The Sympathizer: A Novel (Pulitzer Prize for Fiction)

  Viet Thanh Nguyen

  Nota: 4.5 de 5 estrelas

  4.5/5 (119)
• The Emperor of All Maladies: A Biography of Cancer

  

  No Everand

  The Emperor of All Maladies: A Biography of Cancer

  Siddhartha Mukherjee

  Nota: 4.5 de 5 estrelas

  4.5/5 (271)
• Angela's Ashes: A Memoir

  

  No Everand

  Angela's Ashes: A Memoir

  Frank McCourt

  Nota: 4.5 de 5 estrelas

  4.5/5 (440)
• The Little Book of Hygge: Danish Secrets to Happy Living

  

  No Everand

  The Little Book of Hygge: Danish Secrets to Happy Living

  Meik Wiking

  Nota: 3.5 de 5 estrelas

  3.5/5 (399)
• The World Is Flat 3.0: A Brief History of the Twenty-first Century

  

  No Everand

  The World Is Flat 3.0: A Brief History of the Twenty-first Century

  Thomas L. Friedman

  Nota: 3.5 de 5 estrelas

  3.5/5 (2219)
• A Man Called Ove: A Novel

  

  No Everand

  A Man Called Ove: A Novel

  Fredrik Backman

  Nota: 4.5 de 5 estrelas

  4.5/5 (4609)
• The Yellow House: A Memoir (2019 National Book Award Winner)

  

  No Everand

  The Yellow House: A Memoir (2019 National Book Award Winner)

  Sarah M. Broom

  Nota: 4 de 5 estrelas

  4/5 (98)
• Brooklyn: A Novel

  

  No Everand

  Brooklyn: A Novel

  Colm Tóibín

  Nota: 3.5 de 5 estrelas

  3.5/5 (1937)
• The Art of Racing in the Rain: A Novel

  

  No Everand

  The Art of Racing in the Rain: A Novel

  Garth Stein

  Nota: 4 de 5 estrelas

  4/5 (4200)
• A Tree Grows in Brooklyn

  

  No Everand

  A Tree Grows in Brooklyn

  Betty Smith

  Nota: 4.5 de 5 estrelas

  4.5/5 (1929)
• Devil in the Grove: Thurgood Marshall, the Groveland Boys, and the Dawn of a New America

  

  No Everand

  Devil in the Grove: Thurgood Marshall, the Groveland Boys, and the Dawn of a New America

  Gilbert King

  Nota: 4.5 de 5 estrelas

  4.5/5 (265)
• Steve Jobs

  

  No Everand

  Steve Jobs

  Walter Isaacson

  Nota: 4.5 de 5 estrelas

  4.5/5 (806)
• The Woman in Cabin 10

  

  No Everand

  The Woman in Cabin 10

  Ruth Ware

  Nota: 3.5 de 5 estrelas

  3.5/5 (2322)
• Yes Please

  

  No Everand

  Yes Please

  Amy Poehler

  Nota: 4 de 5 estrelas

  4/5 (1891)
• A Heartbreaking Work Of Staggering Genius: A Memoir Based on a True Story

  

  No Everand

  A Heartbreaking Work Of Staggering Genius: A Memoir Based on a True Story

  Dave Eggers

  Nota: 3.5 de 5 estrelas

  3.5/5 (231)
• Team of Rivals: The Political Genius of Abraham Lincoln

  

  No Everand

  Team of Rivals: The Political Genius of Abraham Lincoln

  Doris Kearns Goodwin

  Nota: 4.5 de 5 estrelas

  4.5/5 (234)
• Fear: Trump in the White House

  

  No Everand

  Fear: Trump in the White House

  Bob Woodward

  Nota: 3.5 de 5 estrelas

  3.5/5 (738)
• John Adams

  

  No Everand

  John Adams

  David McCullough

  Nota: 4.5 de 5 estrelas

  4.5/5 (2409)
• Wolf Hall: A Novel

  

  No Everand

  Wolf Hall: A Novel

  Hilary Mantel

  Nota: 4 de 5 estrelas

  4/5 (3811)
• On Fire: The (Burning) Case for a Green New Deal

  

  No Everand

  On Fire: The (Burning) Case for a Green New Deal

  Naomi Klein

  Nota: 4 de 5 estrelas

  4/5 (73)
• The Light Between Oceans: A Novel

  

  No Everand

  The Light Between Oceans: A Novel

  M.L. Stedman

  Nota: 4.5 de 5 estrelas

  4.5/5 (789)
• The Unwinding: An Inner History of the New America

  

  No Everand

  The Unwinding: An Inner History of the New America

  George Packer

  Nota: 4 de 5 estrelas

  4/5 (45)
• Manhattan Beach: A Novel

  

  No Everand

  Manhattan Beach: A Novel

  Jennifer Egan

  Nota: 3.5 de 5 estrelas

  3.5/5 (792)
• The Constant Gardener: A Novel

  

  No Everand

  The Constant Gardener: A Novel

  John le Carre

  Nota: 3.5 de 5 estrelas

  3.5/5 (104)
• Rise of ISIS: A Threat We Can't Ignore

  

  No Everand

  Rise of ISIS: A Threat We Can't Ignore

  Jay Sekulow

  Nota: 3.5 de 5 estrelas

  3.5/5 (137)
• Little Women

  

  No Everand

  Little Women

  Louisa May Alcott

  Nota: 4 de 5 estrelas

  4/5 (104)