October 1962 THE IMPULSE RESPONSE FUNCTION AND SHIP MOTIONS. by WE, Commins SKIPSTEKNISK FORSKNINGSINBTITUTT KULERTE KORTFORT:” — paTO MOTTATT: | LISTEFeRT:” MG. 63 Ree HYDROMECHANICS LABORATORY RESEARCH AND DEVELOPMENT REPORT Report 1661THE IMPULSE RESPONSE FUNCTION AND SHIP MOTIONS by W.E, Cummins ‘This paper was prosented at the Symposium on Ship Theory at the Institut fiir Schiffbau der Universitat Hamburg, 25-27 January 1962. October 1962 Report 1661ABSTRACT After a review of the deficiencies of the usual equations of motion for an oscillating ship, two new representations are given. One makes use of the impulse response function and depends only upon the system being linesr. The response is given as a convolution integral over the past history of the exciting force with the impulse response function appearing as the kernel. The second ropresentation is based upon a hydrodynamic study, and new forms for the equations of motion are exhibited. ‘The equations resemble the usual oguations, with the addition of con- volution integrals over the past history of the velocity. However, the coefficients in those new equations are independent of frequency, as are the kernel functions in the convolution integrals. Both representations are quite general and apply to transient motions as well as periodic. ‘The relations between the two representa- tions are given, The treatment considers six degrees of freedom, with linear coupling between the various modes.The Impulse Response Function and Ship Motions W.E. Cummins Introduetion Tuatover « decede ago, Weinblum and St. Denis!) presented « comprehensive review of the:state of knowledge at the end fof whet we may cell the “clasical” period in research on sea ‘keeping. Soon after, St. Denis and Pierson") opened the “modern” period (some would prefer to call it the “tatintial” period). The studies of the former period were primarily con ‘cerned with sinusoidal responses to sinusoidal waves, but the introduction of spectral techniques opened the door for the discussion of responses to random waves, both tong and short crested. The construction of the epeetral theory on regular ‘wave theory as « foundation delighted uo all, as it preseated ‘an apparent justification for the admittedly artificial studi of the “Alassical” period ‘The activity during this last decade has been spectacular, with five major and many minor facilities for seekeeping re- search being opened. Hundteds of models have been tested, many full cale trials have boon run, end there has even been some real growth in our knowledge of the subject. In particu: Jar, the spectral tol has been sharpened end tempered by the ‘empiricist, and the analysts have made important advances with the rather frightfal boundary value problem. In fact, we hheve all been forging ahead co rapidly that we appeer to have forgotten that we are wearing a shoe which doesn't quite fit, ‘The occasional pein from a misplaced toe is ignored in our general enthusiasm for progress. ‘The “shoe” to which I refer is our mathematical model, the forced representation of the ship response by 2 system of second order diferential equations. The shoe is squeezed on, with no regard for the shape of the foot, The inedequecy of the shoe is evident in the distortions it must take if it isto be ‘worn at all, Tam referring, of course, to the frequency de- ppendent coefficients which pormit the mathematical model to fit the physical model (if the excitation is purely sinusoidal, that is). Bat what happens when we don’t have # wall defined frequency? The mathematicel model becomes slmest meaning: Jess. True, « Fourier analysis ofthe exciting force (or encoun- tered wave) permite the model to be retained, but physical reality ia almost lest in the infinity of equations required to represent the motion. Let us consider this mathematical model briefly, and restrict ‘ourselves toa single degree of freedom. To be completely fair, Tet us consider 4 pure, sinusoidal osillation. The forcing func: tion (if the system is linear) will be sinusoidal, and can be broken into two components, one in phase with the displace- ‘ment and one 90° out of phase. We further divide the in-phase ‘component into restoring foree, proportional to the displace- ‘ment, and a remainder. The latter we call the inertial foree, and treat it as if it were proportional to the instantaneous Ty etoronces are tated at fhe en of the paper acceleration. The out-of-phase component, which provides all the damping, we treat ao if it were proportional to the in- ‘mantancous velocity. ‘We can now write an equation, which has the appearance of a diferential equation, relating thete various quentt 8(0)¥ + b(@)x + o(a)x = asia (at + 0) But a differential equation is supposed to relate the instan- taneous values of the fonctions involved. If the periodic ‘motion continues, this condition is satisfied. Of course, it could just as well be satisfied by the equation bet eax =f (9) oF more generally (at det bE + (e+ doh x =f) where d fs arbitrary, These exe all equally valid models. One ‘of them isto be preferred only if it truly relatea the displace- ‘ment and its fist and second derivatives to the excitation in ‘some more general way. But suppose f (0) were to be suddenly doubled. Would the instantaneous acceleration be given by = 21 —b(0)s—c(0) x «(0 In general, no! Or suppose the amplitude of the be suddenly increased. Would the out of phase component of £(0, immediately after the change, be equal to bx? Again, in ‘eneral, no. Thus, at best, b(o) must be considered as a sort of “apparent” damping coefficient, (0) as an “apparent” apparent mass, and the physical significance of both is cobscare. When’ the oscillation consists of several coupled modes, the socalled coupling coefficients are equally con- fased and confusing. If we restrict ourselves to 2 phenomenological Investigation cof how a given ship behaves in a givan wave aystem, these di ficulties do not concern us. We simply measure reeponscs to known waves, Most ofthe work over the past deeade las been. of this nature, and much of it has been excellent. However, sooner or later, we are required to consider not “what” but “why,” and a more anelytial technique is demanded. The phenomenological study cen tell us the effect of « change ja ‘hip loa ft sured it; there is no basis for quantitative prediction given the results for one gyradius. And the effect of a chaage in form is presented as an isolated result, unrelated and unrelatable to the geometric parameters involved. We are driven to the ‘use of the model discussed above in an attempt to clarify the relation of cause and effect. But such « poor mirror of reality is of lite velue, and in feet cen do rauck harm. Tam not the fst to ratse this issue. The diffcultien are well nova and a number of writers have discussed them. In part ‘ular, Tick?) bas vigorously argued against our usual practice and has proposed a model which is very close tothe one which “ill be exhibited here. His case is based solely upon the gene- ral characteristics of linear systems, while-we shall take ed> x‘vantage of the principles of hydrodynamics to tie the mode! tothe phenomens. More recently, Davi') has proposed « retio: nal epproach from the point of view of statistics. This is sug particulary since it was the spectral theory of stati- ties which first gave weight to the investigation of responses to periodic waves. Bry, he pei bess of hi apa re “To exhibit « model which permits the representation of the response of a ship (in six degree of freedom) to fan arbitrary forcing function (with excitation in all six ‘modes). The model will not involve frequeney dependent parameters. 2, To separate the various factors governing the response into clearly identifable units, the effect of each to be 1 determinable. Thus the efect of gyradins will be separable jed mass, The added mess will be related only to inertial forces end moments. The nature of the damping force will be exhibited. The effect of ‘coupling vill he derivable and the effect of “tuning” ‘upon coupling will be determinable. In this paper we shall not consider the complementary pro: bblem of the selation of the exciting force tothe incident wave system, This problem is equally basic, and when it has been adequately treated, we will begin to have a satisfactory frame: ‘work for the interpretation of our empirical studies. ‘The Impulse Response Function ‘The basic tool which will be uted i this study is an elemen- tary one, widely used in other Gelds and well known to all engineers: the impulse response function. It is difficult to ‘understand its neglect in our field. Perhaps as Tick suggests, it is because waves look sinusoidal. For any stable linear system, if RQ, the response to a unit impulse, is known, then the response of the system to an arbi teary force £ (9 is xi) = FRO—Y iGo) de or a x)= FRE fear ‘The only assumption required linearity. In the prevent context ths is, of conree, a very strong ‘assumption, and the purists will argue that it implies a thin ‘hip or the equivalent. However all experimental data indicate {hat the assumption is good working approximation for sal to moderate oscillations of real ship forms. We shall hypo- thesize that the assumption holds absolutely. je from convergence) is Let xi + 6) be displacements in the six modes of response: ‘xj = surge (positive forward) % xy = heave (positive upward) X1= ral (positive, deck to starboard) X= pitch (positive, bow downward) yw (positive, how to port) Ler y(t) be the response in mode j to unit impulse att = 0 in mode §. Note that Ry (ce) does not necessarily equal zero, though in a damped system whichis not unstable, it will ordi- narily be finite, In modes without restoring force (sway, surge, and yaw), the impulse response will asymptotically ‘approach some value, For other modes, Ri (2) Tf the {{()} are an arbitrary et for Forcing fonctions the corresponding responses are ae 3 F Rueie—gde PB) ‘Thus, the matrix {Ry (0} completely characterizes theresponse of the ship to an arbitrary excitation. Before we go on, let us consider the relation of these fune- tions to the usual coefficients, First consider the esse where the modes are uncoupled. Let fi(t) = Fieve (ot +) 8 where & isa phase angle whose value will be assigned later. ni) = FF Rats oetoe—o) ta ae (cco(ot + 6) f Rison onde + sin (ot + 8) f Rusinorde] = Fi (Ri? (o) cos (wt + ) + Ru? (@) sin (wr + €)] (4) vic Rat (o) = F Ru (s) cos orde [5a] 10 = Fri snont cy are the Fourier cosine and sine transforms of Ry (0). We shall ‘all these transforme the frequency response functions. We make the further reduction w(t) = F; (Gye coes, + Ry? sine) cos or “(Rj cost, — Ry! eos) sin ot] ‘Taking ten eg) = Ryt/Rit 9 wehave x () = FLOR, + Ry] con at. a Also F (Rye co ot — Ry? sin wt) 60 : 1 1 + GAN" Nov consider the sua epretnation ai thay hea A. ° Using the anf om (7) and (8) easy een hat ae vale Ay (20) at + we 10) = —t _ [ARF + AIA] [A more useful relationship is obtained by setting «; = 0 in (al: a ‘Thus Ry¢ and Ry? are the amplitades ofthe in-phase and out ovephase components ofthe respons toa unit amplitude fore ing fanction of frequency «The impulse response fanstion ie tlated to these functions by = Ry (0) cos or + Ry (o) sinwt. ay Ro = 2 f Ryt(e) cova do aa sf) sine do a} using the Fourier inversion formulas, Note that Ry* end Ry* fare uniguely related. If one is known, then by (12] end [5], the other is determined. Equation [11] ean also he waitten Ryo} + RyHY"* cos [or—K oj] 13} where Rito tune = RX) | a4 RY} ‘Thus, tho response foliows the excitation by the. phase tan! (Rj#/Rj) and has the amplitude ((Rut}* + (Ry)‘The responce for a given frequency, a dctermined by the paieof functions Ry", Roe, or alternatively, the pair (‘Riel + GRP, ta PZ, is @ mapping in the frequency do- rain of the unit response fanction, whichis defined in the time domain, As equations [4] and [11} permit us to pass from cither domain to the other, the two representations are com: pletely equivalent. Viewed in this wey, the frequency response function isa meaningful, useful concept. 18 only when we try to attribute « deeper meaning to it, by imbedding iin a false time domain model, thet we ereste confusion. ‘Now consider the more general, coupled system, with exci tations in a single mode ofthe seme form as given in equation (3). Then xi (0) = F,IRytcos (ot + a) + Rytsin (ou ei} If we consider the usual representation 05) 15 aks + bet + ona) = where fy) = 0 for k #i, we ean develop o system of equa Hons in the unknowns, sj, bye» (The oy are assumed known from static measurements) All 72 of these unknowns. are present, in principle, except where modes are uncoupled. To determine them, it is necessary to consider the responses to ‘excitations in each of the modes separately. We then have ‘enough equations, if we seps ‘components, to determine the coeffice for thom here, so we defer further diceussion until we face ‘a closely related problem. It is only significant to note th they ean, in principle, be determined from the set of impulse response functions, and therefore they contain no information ‘whichis not derivable from these Functions. Setting «5 = 0 in [15], we have the system (0 FiO. Rys cos ot + Ryt sin wt. 16 a = Ba 16), ‘Thus, Ry? and Ry? are the amplitudes of the inphase and ‘ofiphase responses in the j! mode to unit amplitude ex- citeton in the #¥ mode. As before, me no 2 [ ryrnando ma o2F npomance ony ona (Ry)* + (Ry) cos (ot — e) 118) whee taney = Ry ow ‘We have passed over the question of convergence ofthe inte sale in equations (2] and {5}. Consistont with our hypothesic ‘of linearity, we shall assume [f(0| is bounded. There will then be no difficulty unless [Rj (c}| dx does not exist. Unfortunate- 1y, in three modes there are no restoring forces (or else they are negative), and evidently some care is needed in treating these cases, A nogative restoring force implies an unstable system, whidh would be beyond the scope of this analysis. Howover, 340 = J Ry Q—9 filo) de — J Ry a) fo) de + 2y 0) wi) = FRulo§(t—a) ae + Frere —Ryenheodety Oy — BO) ‘The second integral converges, s0 this expression provides a table definition of x (2). Now let f,@) = cost. After an inte eration by parte, we have x)= Vet Ry sino —o ée +Fimje+9—Ry coleorwrdr +500). Our only concern ie with the oselllatory components of xj. ‘These are ensly determined by considering theasymptotic form ofthe above expression ast becomes large. Ry (-+1}->Ry (°°). and the second integral becomes constant, If we set 2 (0) =~ Tin, f [Ry (+) Ry (e] cos or de then Lay cos ot + Ryesin or) where Ry and Ry are the sine and cosine trensforms of Ry (9. We know that x; (0) is sinusoidal, with frequency 0 “Therefore, this expression holds not only for large t but for all, If we define 40 eu RES Ryo (226) Ry = Ryo (226) then {16} still holds. Note however, that Ry® and Ry are no longer transforms of Ry, because these do not exit, Neverthe: Tess, an inversion is still possible. Consider Fim (01, (eo cos wr de Ju) Ry cosinor — JRutsinorae Ryt/o= Ry. Ry is the cosine transform of (Ry () —Ry (9) and Thet Ry) Rye) +2 [ Bireoona Letting t equ sro, Ryfdo Ryo 31 nuget fryeeu—nae 4 When My (2°) 0, this reduces to [17a Similac, Te Ry (i) sinorae =—Ry(oy/ot | Breoras Est Ry (}] / wo and - Ry (0 = Ry6oo) + J [ro Ful) snesae { Rytsinor do jas anes ° 2 ‘Therefore, [17] holds even when Ry TERE and Ry# are known, it is not ciffiult to determine whether or nat Ris(0°) = 0. Equation [23] gives Ry (C2) in terme of Ry. Also 5 using a well known theorem in Fourier transform theory (Refe- rence 5, page 12) ‘When the matrix of impulse response functions ia known, ‘our first objective of finding a representation of the ship re. sponse which i fee of frequency dependence igachieved. These fonctions, which we shall collectively call the impulse response matrix, can in principle be determined experimentally Equations of Motion ‘The transient response of a ship has been considered by Haskind!), who attempted an explicit solution of the boundary value problem. This, we shall not try, as wo are concerned ‘only with finding an appropriate form for the equations of motion to use as « basis forthe interpretation of experimental results, We do not agzee with certain of Haskind’s hypotheses, and our resulting equations differ from his a several important respects Golovato?) carried out an experimental investigation of the declining, oscillation motion in pitch. However, Golovato was not aware of the equivalence between the transient and steady state responses which we have just discussed, 50 he attempted only to match the coeffeients derived from the transient experiment, at the frequency of the declini secillation, with those from «forced oscillation experi tment at this frequency. He was handicapped because of the anomalous behavior of the curve of declining amplitudes. For ‘simple hermonie oscillator, this curve is ¢atraight line when plotted on semilogarithmic poper. His curves departed radi cally from such « pattern, He recognized that this implied that the mathematical model was faulty, and attempted, with some smecess, to ft hie results with forme beeed on Haskind's study More recently, Tasait) has performed declining oscillation experiments in heave using two dimensional forms. His results fare not significantly different from thote of Golovato. He matched his results at the measured frequency with Ursell's theoretical results for forced oscillation. The agreement is quite good. Case I — No Forward Speed Let the ship be floating ot rest in gull water. We use system of coordinates (Ey, Se Ua Rxed' in space, with origin inthe free surface above the conter of gravity ofthe ship. At time t = 0, we suppose the ship tobe given an impulsive displacement As in the j® mode. The time history ofthis im pulse is not significant, but for purposes of visuslization, it may he considered to consist of a movement at a large, uni form, velocity ¥, for a small time 0%, with the motion termi: nated abruptly at the end of this time interval. Then Ax, = vA Daring the impulse, the ow will have a velocity potential which is proportional to the instantaneous impulsive velocity of the ship, It may, therefore, be written vjQj, where wy is a normalized potential for impulsive low. will satisfy the con ditions Wao on a0 126) —Buj/2n= 5; on S en here i, =138 pe jn 4s.6 peice saasy i steady dete vat sre i= wi ventric += pontion vestor with reapect to eg of ship It in well known) that the above problem is equivalent to that ‘obtained by reflecting $ in ty ~ 0 and taking the eurface con: dition over the reflection to be the negative of that over S. The ‘olution to the Neumann problem for the'flow outside this com posite eurface is elao the elution to che given problem in the lower half-space. For nowpathological aurfacea, the solution exists, and in fact can be computed by means of modern, speed equipment.*) Dating the impulse, the fee surface will be elevated by an n= —y, 2M as, (29) Oe ‘Alter the impulse, this elevation will dissipate in a radiating disturbance of the free wurface, until ultimately the fvid is ‘gain at rest in the neighborhood of the ship. Let the velocity potential of this decaying wave motion be @, (t)Axy. It must ‘satisfy the initial conditions 7 Bae, State "Se (30) and i i By ir sais the usual free surface condition 2, Bay 2 ate 182) and the houndary condition on S 29, a = 9 a = (83) ‘We may take this to hold on the original position of S, only introducing errors of higher order in Ax;. This is @ classical problem ofthe Cauchy Poisson type, and there exists an exten ive literature on the subject. With condition [33], itis more ficult, by an order of magnitude, than the Neumann pro- Diem. We assume that it has solution,[Now let the ship undergo an arbitrary small motion in the 4s mode, x, (0). To the frst order, the velocity potential of the resulting flow will be simply @ = ay, + fy —0 4 (0 at 34) It is evident that the boundary condition on S is satisfied on the equilibrivm position of S, as the first term provides the proper normal velocity and 3g, /8n = 0 on this boundary. But also, the value of 86/8n on the actual position of $ will only differ from ite value on S by terms of second and higher order in xj and ite derivatives, 20 we may consider that [34 holds ‘on the ectual postion of the hull To verify thatthe fre surface condition is setsfed, frst note that 20 _ a ai | 00) 3 8 + gy) SH 4 HO Fen ae wt ey ( a0—2 =? ayy ae +f HE? gee By (26) and (30), on fy = 0 this reduces to a0'_ an , Pa O—s = 809 5 [PML aioar 2 3 Ate, 38 2H (BO gigas a By Ba Substituting these inthe fee surface condition 39, 90 52H) 22 2 5,(20 5 530) ae 88 ( a” ay se , 89 28 4 28) siyar=o 135) + {Gree} by (81) and (82). Thus, this condit ig the required potential ‘The formals [34] isa hydrodynamic analog of). Tei quite ‘general, and can, for instance, be used to find the velocity potential due to a sintsofdal oscillation with arbitrary fre ‘quency. Iti, of course, necessary to know the function 9; tnd this presents unpleasant difficulties. In this study we content that (0) exists, and these difficulties do not concern Of more importance than the velocity potential isthe force acting on the body. The dynamie pressure in our linearized model is simply so satisfied, and © 8 peer “ 2 st m4 + [A seo ar ° ayode. (86) ve force (or moment) acting onthe hall on given by in the Ko mode cn ayer a Ke frnere fae f a = mam ve f stool : Be t—9) Se +f xnemos (ae 7 where me f wads (a) a (391 xnto me f PEE nae 3 We can now write the equatins of motion ofthe ship whick is subjected to an arbitrary set of exiting faces, {6 (0). ‘These willbe + JR 0 669 4 Eom By, + mad % + ene; (0 140) there sm, = inertia of the ship in the j!® mode cjg) = hydrostatic force in the k'* mode, due to displace- ment xin the j* mode Kronecker delta Qj, = Tit} k = 0i5#K). on Case Il — Ship Underway ‘The case of the ship experiencing small oveillations sbout 1 reference position of mean uniform velocity is much more complex. A pair of functions, y, and gy, no lnger suffices although the pattern of our analysis will be similar to that followed in Case I. Wo use a fixed reference aysiem, with ty surface and with the e-g. ofthe ship at & ‘We euppose the ship to be moving with u the fy direction, Consider the Cauchy-Poisson problem defined by {30}, [31], 82], and 83}, except that now [33] is to hold on the moving surface 5. This problem has « solution 93 Gis Sarat. V) of course, identical with the q, of Case I when V= 0. sing this; and the obtained in Case 1, we may write the velocity potential for steady motion, = Vin G—Vetete + fortet ade fs) where fon the free at ime t= 0. form velocity V in HID = HEV te tet "That this satisfies the boundary condition on S is evident, as Voy provides the necessary instantaneous normal velocities, and 8—j/n = 0 on S for all r. The free surface condition is ‘also satisfied, as may he verified by direct eveluation, as in Case I ‘The velocity potent moving with constant = Vin Gi —V 1 for the flow generated by the ship locity, after an impulsive start at time foto + fortt—aded (42) ‘The free surface snd ship surface conditions are, satisfied as before, The surface elevation at t= 0s v Mh = [av +e 00.0) ebYe te‘es required, and th intial conditions are met, Therefore, this, taste the stated potential We shall need the stesdy motion velocity potential for the ‘ese in which the ship is displaced by Ax; from its reference position, We could, of course, consider the displaced ship as completely new hall and write down a potential similar to [41] with new functions and gy. lnsteed, we determine the corrections to the ‘} and @, discussed above, which ore necessary to satisfy the new boundary conditions. We wish ‘yy; such that Wet Any = 0 on by = 0 3 ‘which implies thet Yu (488) eo Zev savy ai on 8 lees) (64) my 77 on ty 2 Bon S (ainplncnd) He) en Ba br Tn tree eases, soltions aro immediatly availabe. Ij = 1 ou x Aa ee Bo) = — Os SH 0 a) wi ard = V1 3 is a solution of (48) and [44], traneation. Therefore 1 in this ease we have simple 5a} ‘Stmilesly wr (480) For 5 8, tere is no such stole solution. Noting thatthe right side of [4] zero on (orginal), io only necesary tovind is cange when Sis displaced. Then Sha, oy ay OL Oa Bask xy (60) Buty HE j = 6, the displacement i Jaton ofa yawed body ise lations parallel and perpendicu the solution to (43) and [44] is Walla + Sefkxas te — ba dbxes tad Axe Gr + todbxesto— 1 Bx 9) mvt dnt ply a rotation in yaw, The alent tosimalteneoustrans- to the body axe. Therefore, (7) wo eae Tj = 4, 005, the fiat term of [444] becomes fe ds Gaal “The second term is Mes Jur on $ (displaced) ‘which may be written, using values of and Vx eveluted on S (original), Int As; Gj -em)) [Vi + As ae) Vd I we drop trms of higher order in Ax; and use £27), con: dition [46a] reduces to Mit FH ane eb Base) (48) (22073), ete Om (4 ye 5 | (2 ees vb} Condon [4 nd (a) ae uclen e determing seen gt nat on $dapacedy but we ony nodes Se ett ti we tke te hip sytae coon o arcing! Stary, 6} ond (8) nb aptied wataterens poston af & “To vy sreoonds «wth ey 3 B81 gM ferty=ormo 8) a By * 7 eit and with conditions corresponding to (30), (92), and (83) hold ing. Again we take the ship surface condition to hold on the reference position of S. ‘We need yet one more pair of functions. The normal deri- vative 39;/8n will difer from zero on S (displaced) to first order in xj. To corzect it, we define a function which satisfies the conditions * bs Ss = 2 [ar tt—v dr om § (dnplaced) Rte (60) and Yai 0 on &y=0 sy ‘Ase do not intend to exhibit oltions for a, we shall not reduce the right side. We alo need 8 Gy wih Boo) =, Hs ae By and the other appropriate conditions ssp holding We now have all the pieces needed to write the velocity I for the flow about the ship when displaced by Ax, from its reference position. It will be O=V itn + fm (t—0 de} 2) + bx [vy + f(t —2) ea] + bx ther + Spas t= 9 de) "The terme V oh + Ax;5) provide the necessary normal velocity inthe displaced positon ‘The normal velocities due to 53) Vass a; and VJ gy (t—2) de feaneel, and none of the other terms contributes normal velocities of frst order in Ax. Therefore, the ship surface ‘condition is satisfied in the displaced position. Further, each pair of terms in brackets eatsfies the free surface condition, ‘as may be verified by direct evaluation. « We alee have al the pi for the flow generated by (0). Teil be = Vitis + Ser Gt—2) al needed to assemble the potential ip experiencing small oscillations +E dy + Su (62) 5 (8 de]+ Etsy wa + fea tio 00 da) +E tiv t Leora e io de sa ‘The ship surlace condition is eatsfied as before, except now the term 33, , provides the additional components required for the ovilatory velociies. And again, the bracketed pits of terms satisy the free surface condition. ‘The dynamic pressure at any point in the uid is given by a; vfs a (me s)) ar a fy + vad VS tance 19 4 Sei) oy ae aon Dae (55) two convolution integrals in [35], one involving the esellatory displacement and one involving the oscillatory velocity. These may be reduced to one by means of an inte- iistion by parts, We can go either vey, but there is some advantage in defining Et (tnd + Gy (ut—a]de=O;t—t) (56) 0 that f on By and Seu, yids eeacsi 3a 220 of B09) ot de 7" 20. Oy - {sags Seu y(t) ér 158) ‘The significance of this function ®; can be seen by rewriting the potential for the uniform flow with the body deflected (Equation ($3). It becomes V Cont Sor to de Bx (5 + ¥) +803}. (59) %) Equation [55] now reduces to » se 4 La S faa tev (ute ok 2 yu + ey +90) (60) concerned with the oscillatory value of the hydro: dynamie force, but not steady components. The leet term [59] oes not involve the {x}. However, when we integrate the pressure over S, the fact that S ie changing its position in a steady fow field implies that even this term contributes to the ‘otcilatory pressure. These pressures will be functions of the displacement only. Integrating the pressure over the surface of the ship, we cen write the equations of motion LY lms Oye + mp) %5 + beds ess + KUO aT fe ton where m and my, are as defined in (40) and (38), and by = ev f (ou tye *) do (62) jax; = Total hydrodynamic and hydrostatic force in the i mode, due to displacement x; in the j* mode. Kptt—9) = fe ate (63) ‘There are symmetries which reduce the number of coeff: cients, For instence mane fade z the space enclosed by S, the free surface, and an wwe can apply Green's thoorem, end ve ma ef wae =m (64) 3 Further, if e conser the transverse symmetry of the ship, ‘he matrix {my} reduces to © me omy 0 Mey fon oe oa ome | Ms 0M Om ome 0 my 0 my | Evidently, the matrix {b,} is ofthe same form, excet thet, In general by, ¥ bys. The matrix oy, even simpler as surge and sway diaplacemente provide no restoring forces, hydro static or hydrodynamic, Therefore cpa] on 9 fm 0 tye 65 fend > eT hes) a er ‘The matrix. (Ky, ()} is of the same form as {by.}[Equations (63), though similar in form to those developed bby Haskind, difer from bia in soveral essentials, Haskind found no hydrodynamic force proportional to the displacement, nor dtd he find the components of b, due to Wj and vj- He algo found thet bg = byy = 0, and by, = —bus. The presence Of Wo; in the definition of by, makes i unlikely that such rela tions hold here, Further, his kernel i the convolution integral ‘must difler from that found here, The reason for these dle renceo is that Haskind neglected terms in satisfying the boun- ary condition on the diaplaced $ which are of fist order in x; With equation {61}, we have advanced « long way toward the second objective of this paper. The dynamics of the body have beon seperated from the dynamice of the fluid. Further, the hydrodynemie effects have heen separated into well defined components, etch of which ean be found (in principle) from the solution of « Neumann problem or « Cauchy-Poisson pro- blem, Specifically, we draw the conclusions: 1, ‘The equetions of motion are universlly valid within the range of validity of our assumption of linearity. That is, ny excitation, periodic or non periodic, continuous or discontinuous, is permissible, just eo it results in small Tncements from a condition of uniform forward velo- city, The ease of motion with a negative restoring force, or at least the early history of suck motion, is not ex: cluded. 2. The inertial properties of the fluid are reflected in the products my ij. The coefficients are independent of frequency and of the past history of the motion, so they are legitimate added mastes. Further, they are inde- pendent of forward velecty. 8. There is an effect proportional to %; which accounts for some of the damping. This effect vanishes when the mean forvard speed is zero. 4. There is « hydrodynamic “restoring” force (it may be negative). It ie qual tothe difference between the hydro: ynamic forces acting on the ship: due to the steady flow in the equilibruim position and in the deflected post 5. The effect of past history is embedded in a convolution integral over %t). For sinusoidal motions, this intogral vill ordinarily have components both in phase with the ‘mation and 90° out of phase. The latter component con- tributes to the damping. Hydrodynamies of the Impulse Response Function ‘We nov have tro systems of relations between the excitation and the response of the ship: the impulse response relations, [2} and the equations of motion, {61}. The former are of gree ter value in describing the response to a given excitation, while the later are useful in analyzing the nature of the re- sponse. Both systems hold for small oscillatory motions, oo there are relations between them. We shall examine these Firat, Iet ue tart ith the equations of motion and derive the functions {Rj ()}. Suppose a ship, moving at constant forward velocity, to be subjected to @ unit impulse in the i mode at time t= 0. During the impulse, the equations of motion reduo t mui Bats = da here Bj, Is the Kronecker delta, Suppose the impulse acts uring time At ‘Then, since ¥ At Ai = Ry (+0) e wehave mya (t 0) + Em Ry +O) Ou. ton ‘Asi and j range independently from 1 to 6, we have 36 equa: tions relating the two sets, {may} and {R(0)}. IF the eaue- tions of motion are known, equations (67) fx the iil cond tions from shich the impolse response functions ean be deier Imined. Conversely, if the impulse reaponse functions are Known, these equations yield the appareat meeses. Immediately ater the impulse, we have 0) ar 40) +00 2-50) +00) FRemae—nde= 00. ‘Therefore, considering only zero order terms in tthe equa: tions of motion yield: « meh (+0) +E my, Ry +0) + ba Ryton =o (68) ee which relates the coelfcfents {bj} to the accelerations {hur}. Now suppose the ship to be acted upon by a constant anit force in the # made (me assume a postive restoring force 19 exist in this mode). Then, alter equiibrum ie reached, Zens — bx ist and y= FRy (ae ra Bex} Rutoer= bx (69) 1m modes without a positive restoring force there is difculty as there ie no guarantee that all of the coupling coefficients are necessarily zero, Thus, og; xq, the sway force due to a yaw langle xg, will not ordinarily be zero, or even negligible. We all return to thie pint a litle later. It we rewrite [61] in the form 2 FRx Ru oo ae= Em By, + mma) Ri (9 + eRe (9 + ca. Ry OO) (0) wwe have a set of 36 equations which ean either be regarded {se set of simultancous integral equations for the kernels {Kje (D). o a set of simultaneous integro-differential equa: tions for the impolse response functions (Ri (t)} We have aleady seen (equation [16)) that if £,(9 = cost then x) (0 = Ryfeoset + Rytsinat Substituting these values in the equations of motion, we got -3 {Com By, + mg) oF Ry? — by OR? — ee Rye int 0 Ry Kyo + Ry Kj'] cos ot ley + mp) OFRy— by ORG? — ene Rat 0 Ry! Kye — Rye} sin or} = Fie covae For any given frequency, thie fe an identity, so the net coefficients of os at and tin at must be zero, This gives us 72 equations relating. the transforms {Ry %Ry"} with the seaaforms (Ky Ky). Wehave 0 (Ryt Kae + Ry (may tt ‘ x + Elm By + my) OF Rye — bye RyP— ene Rul ‘ nd — 0 My? Rye —Ry Ky oF fom ty tomy WARY Hh ORY en RAD Ior, equiv ly 28, + mot on eR WIRY! = je + Ky) @ Ry} = Bu, (72a Zlib + Ky) ORY F leon dy. + mp) OF — ej, — OKA] Ry"} = 0 (726) Thus, instead of the integral and integro-diferental equations relating (R) with (K5q}, Equation (70), we have systema of linear equations relating their tranforms Equations [72] are. particularly revealing, If we were 10 tsbitrarly st the Kjy® and Ky! tobe aero, these are precisely the equations we would get between the frequency rexponse functions, Rye and Rj# and the uel frequency dependent cocfcens, Thus, tin elear how freqvency dependency of the Kae and Ky! fy forced onto thete coefficients in the conven- tonal representation, ‘The transforms of {Rj} alee yield useful variants of the selations already given. For instance if we let © = O, we have a Rye O) = Oy (73) ‘more general form of [69] Also, noting that Ry () 2” an prcnuedn mt Rama? Tent onetde webeve 0) == ("oR do (4) hore oR, do. (740) ‘Therefore, 67] and [68] may be writen 2X tm Forse dy] = 5, rs} and Eferita + my Fur Ryedo—by FoR, do} 0 (76) Conclusion In the foresing, we-have preva two mathemati! sodel for repenning the pone characters of sp. ‘The equations of moton are move gener x hey typ © the inal sages of an uetble ation. Whee eto tyme re eGtaly valid we hav relaon wich part oy to pen a atrial) rom eer uo ote ta The inploe rerpon funcon ia ceinly the beter represetalon for computing. repose It negate all feces mechani hyarstae and bperdyamies erat tticien manner pose for computation. Hwee, for ia ‘ery rnin it «por andy ea for exling why the ip sep the way Ht donor ow the regal be Sfesed any Gunge in conten oom For tance Imodl ate eda teed wih estan in cvsin moder Areata in tay mode vl aft the inpube reopens fancon n any pled tee. Since he sip i ae all rodents nde Improper to ate hte epee finns inp flail havo les hey oe cored for he ch oe rats, ‘The hydrodynamic eauationy do not sues from thi di advantage: Known rst eee fede end Ue tte deena: Ora ange in mas itbtin an Be 9 treated independently of the hydrodynamics. It is not uncom ‘mon in model testing to have “incompatible” parasitic inetias in the different modes. Thus, the towing gear may contribute ‘8 diferent masa in surge from that in heave. By means of the ‘equations of motion, the effect of these inertias upon the motions can be analyzed. Thus, the equations of mation pro- vide a more powerful analytic too for studying the releion- ship of the response to the parameters governing that response. ‘We can conclude, then, that thete two representations com: plement eachother; the one for response caleulation, the other for response analysis. In fac, if It in truly prac from one representation to the oll present themselves: 4) Model experiments may be designed to obisin maximum ‘ccuraey rather than maximum realism, Hydrodynamic cffects chould be emphasized in the design since other effects are separately determinable, Thus, one should test ‘at amall gyradias in order that tho effect of the inertial properties of the body itself will be minimized. by Restrainte are Te if their characteris fully known. ‘Thus, rather than directly find the impulse respons matris, in its complete generality, more elementary exper ‘ments may be conducted to determine specific terms in the equations of motion, We may restrict ourselves 10 one, two, or three degrees of freedom and obtain results which are completely valid when interpreted by meane of the equations of motion, ©) The recurring difficulty of handling modes in which the the restoring foree ie zero or negative can be easly over: come. It is clear that an accurate experimental invest n of these modes would uncover practical difeulties analogous to the theoretical ones we have discussed. How: ‘ever, the problem can easly be aolved by imposing known restraints (ie. springs) which wil restore positive stability. The effect of these restraints is readily insludable in the equations of motion, it can be removed by calculation, end the correct impulse respon determined. free of re aint, can be {(Worgetragen am 25. Januar 1962) Manley: “On the ‘Transactions, ‘The Society of Naval Architect and Marine Engineers, Vo. St, 1950. 2) St Denis, Manley, and Pierson, W. J, Je: “On the Motions of Ships in Confused Seas” Transactions, The Soclety of Naval Architects and Marine Engineers, Vol. 1083. 8) Tlek, Leo J: *Difterential Equations with Frequency Dependent Coetfciente™. Journal of Ship Research, Vol8, No. 2, October 1980 4) Davie, Michael Cz “Analysis and Control of Ship Mo- tion in a Random Seaway” MLS. Thesis, Massachusetts Tnatitute of Technology, June 1861 5) Sneddon, Ian N.: “Fourier Transforms”, MeGraw-Hi Boole Company, Ine, 151 4) Haskind, M.D. “Oscillation of « Ship on a Calm Sea” Bulletin de Academie des Sciences de TURSS, Classe des Sciences Techniques, 1915 no. 1, pp 23-34. 1) Golovato, P.."A Study of the Transient Pitsing Ose latfons of a Ship’. Journal of Ship Researeh, Vol. 2, Ne. 4, ‘Marc 880. ©) Tasal, Fuluzo: “On the Free Heaving of a Cylinger Floating on the Surface of x Fiuld". Reports of Researdh Institute for Applied Mechanics, Val. VILJ, No. 32 180 9) Wein blum, G. Ps "On Hydrodynamic Masses". David ‘Taylor Modal Basie Report 80, Apri 1952. 10) Hess, Jona L, and Smith, AMO, “Calculstion of Non-Litting Potential Flow about Arbitrary Three-Dimen- onal Bodies.” Douglas Atreratt Company, ine. Report BS 4ouza, 15, March 1062.