Dalzell J F - A Note on the Form of Ship Roll Damping




VDSTO ADF A -- VM 1 .S 628 -- v. 13 no. 1 (Mar. 1969)-v. 15 no. 4 (Dec. 1971) , v. 18 no. flNL NU NUN SSUB SUA TU VDSTO VU Journal of ship research mel.library@dsto.defence.gov.au DSTO Research Library - Melbourne 506 Lorimer Street Vic 3207 ATTN: PHONE: FJl.X: E-MAIL: SUBMITTED: PRINTED: REQUEST NO. : SENT VIA: EXTERNAL NO. : Journal Society of Naval Architects and Marine Engineers (U.S.) JOURNAL OF SHIP RESEARCH. Society of Naval Architects and Marine Engineers. New York 22(3) 178-185 1978 Dalzell, J.F. A NOTE ON THE FORM OF SHIP ROLL DAMPING 0022-4502 LCCN: 62051402 flNBD: 000000893539;LCN: 62051402;CODEN JSRHAR;P Libraries Australia/.biball-r20-db01 Fair Dealing - S183 - Commonwealth Per VM 1 .S628 E-mail: mel.library@dsto.defence.gov.au 2011-10-18 16:47:49 2011-10-19 11:21:39 COR-10015190 ISO 8795059

03 9626 7434 03 9626 7082 mel.library@dsto.defence.gov.





WARNING. This material has been provided to you pursuant to section 49 of the Copyright Act 1968 for the purpose of research or study. The content may UNSW@ADFA (CRICOS Provider Number: 00100G)he Act.

This material has been provided to you pursuant to section 49 of the Copyright Act 1968 for the purpose of research or study.edu.library@dsto. This IS With the compliments of the Academy Library UNSW@ADFA Northcott Drive Canberra ACT 2600 Australia. The content may UNSW@ADFA (CRICOS Provider Number: 00100G)he Act. You will be invoiced for $0.au 9 pages.defence.gov. WARNING.au for assistance.Ask4it ILL/DDS UNSW@ADFA COR-10015190 REPLY: This document contains NOT an invoice.01. . ABN 57 195 873 179 Please contact (02) 6268 6158 or ill@adfa. E-mail: mel.

To determine whether such a communication would be an infringement. Further dealings by you with this material may be a copyright infringement. (Copyright Amendment Regulations 2010 (No. 249.. .Commonwealth of Australia Copyright Act 1968 Warning This material has been provided to you under section 49 of the Copyright Act 1968 (the Act) for the purposes of research or study.1). it is necessary to have regard to the criteria set out in Division 3 of Part III of the Act. Schedule 4) . The contents of the material may be subject to copyright protection under the Act.

A number of investigations have been made of the application of the Fokker-Planck equation method [6] for the solution of nonlinear vibration problems. In this approach there is a serious drawback with respect to application to the usual nonlinear rolling equation. rolling equation may be con..StevensInstituteof Technology. The methods outlined by St. An approximation to the quadratic term The single-degree-of-freedom sidered in the form: liP where: <p= roll angle I = roll inertia li(ip) = damping function !i(<p) = restoring function F(t) = excitation t = time In the literature there appear a number of forms of restoring function. !i(<p) = j=1. Oalzell2 The objective of the present work was to develop an approximation to the conventional mixed linear-plusquadratic ship roll damping model so that analytical obstacles could be overcome in the application of the functional series expansion to nonlinear ship rolling. This is not the case for the "quadratic" term ordinarily used to represent the damping nonlinearity. Vol. S Numbers in brackets designate Referencesat end of paper. 22. These assumptions being fulfilled.1977. In this technique. Sept. it appears that the terms in the equation must be analytic for small values of the variables. administered by the DavidTaylorNavalShipResearchand DevelopmentCenter under Contract NOOO14-75-C-0278. ship rolling continues to pose a problem. the spectrum of response may be estimated. Haddara (7) uses a modified Fokker-Planck approach which results in estimates for roll variance for the case when the excitation spectrum is white. SHIP RESEARCH 0022-4502/78/2203-0178$00. A mixed linear-plus-cubic approximation was found to be reasonable for this purpose. This same nonlinearity. nonlinear elements in the equation of motion are replaced with linear elements chosen so as to minimize in some sense the resulting errors for random excitation.. and this note (which is an abridgement of [9]) summarizes some steps taken in this direction. + !l(ip) + !i(<p) = F(t) (1) L Bj<pj (2) In most applications. 1978. In the case of both pitch and heave it has been possible to develop linear theoretical approaches to damping which conform reasonably well to observation. Thus the first problem in any attempt to apply the functional series approach to rolling is to overcome the quadratic damping obstacle. pp. 1977. the problem comes ultimately to that of prediction of the statistics of maxima of ship response in realistic (random or irregular) seas. obtaining solutions for the nonlinear roll spectrum in terms of the sum of the linear spectrum and various convolutions. The foregoing form for roll damping has also been used both explicitly and implicitly in modern multi-degree-of-freedom analyses. F. is to be found in many multi-degree-of-freedom analyses. A few recent instances may be cited.3 •.41/0 . in addition to the linearity assumption. There presently appears to be no theoretically based prediction method for roll which is completely free of empiricism with respect to roll damping. None so far seen are fundamentally different from an odd series in <{7. In current practice the usual approach to prediction is some variant of an equivalent linearization technique (Kaplan [3]. and from this a reasonable estimate of the statistics of maxima and minima can be formed. Vassilopoulos [4]) for the estimation of the variance and spectrum of roll. This was pointed out some time ago by Vassilopoulos [8]. Methods of estimation of the magnitude of oscillatory ship motion statistics are reasonably well in hand so long as the motion being predicted can be assumed to be a linear function of regular wave height. 178-185 DATE OF COPYING 1 9 OCT 2011 COpy MADE IN COMPLIANCE WITH THE COPYRIGHT ACT A Note on the Form of Ship Roll Damping J. 1 This note summarizes work carried out under Naval Sea Systems Command General Hydromechanics Research Program. No. In any event. the assumption that the wave process is Gaussian. There appear to have been extremely few deviations from the form of equation (3) in the last century.revised manuscript received August25. If the functional series and a differential equation are to be related. there were indications that this model may be closer to an "equivalent approach" than to an "approximation. The classical damping function may be written in the form (3) The absolute value in the second term of equation (3) is the issue. Evaluation of roll response statistics according to this approach promises to be extremely difficult when the spectrum of excitation is not white (flat). Other approaches have been investigated. 178 Yet another approach to the nonlinear prediction problem is the functional series model.ACADEMY LIBRARY UNIVERSITY OF NSW@ADFA Journal of Ship Research. In this context." Introduction IN MOST CASES of practical application of seakeeping theory or experiment.3. This has not been the case with rolling. Manuscript received at SNAMEheadquarters May 23. Haddara (10) replaced the second term with a cubic term in ip for illustrative purposes and possibly because the quadratic form was impossible to handle in his derivation. or the recognition of its possible occurrence. Lewison [11] also replaced the quadratic term with a cubic term. the form of equation (2) presents' no fundamental problem. Denis and Pierson [IP involve. 2 DavidsonLaboratory. only the first (linear) term is used. Yamanouchi [5] approached the problem with a perturbation technique for the solution to the nonlinear differential equation. oboken. The often-used single-degree-of-freedom equation for roll (which is traceable to the work of Froude [2]) is nonlinear in damping moment. In the course of analyses. His reason appears to have been largely to facilitate analog comJOURNAL OF.New H Jersey.

the form of equation (7) is suitable. and the process iterated if required. K) (5) 0. the general behavior of the approximation beyond the fitted range is the same as that of the quadratic term. if these assumptions are applied to equation (1).. the presence of the parameter (Pc means in principle that estimates of both the roll and roll velocity variances would have to be made for an arbitrary choice of (Pc. In the sense of making estimates of roll variance utilizing the functional series method.. equation (7) might be thought of as a sort of equivalent non linearization. The data-reduction approach has been basically the same since the time of Froude [2]. though he remarked (without elaboration) that the cubic model better fitted the data. / 0. In fact. equation (3). . these used to estimate a reasonable solution domain of (p.0 /. N((p) '" [N21 0. results in (PcNzzl (P + 35 N_22(p3 = N((p) (7) 16 48 !Pc -. say (-(Pc < (P < (Pc). 1.puter setup. One approach is to make a least-squares fit of an odd series to I (PI (P over some assumed range of (p. In particular. the result for truncation of the series at the cubic term is: (6) Figure 1 indicates the degree of approximation involved in equation (6). In the vast majority of studies where some distinction 'is made between "linear" and "quadratic" components of roll damping.2 After solving for the C<k. and found in their analysis of the roll damping that an improved fit of the data was obtained by adding a cubic term in (P to the standard model. The foregoing was sufficient encouragement to consider the replacement of the quadratic term in equation (3) with an odd series in (p.1'3 _ in ship sallying a. If ~tis assumed that equation (6) is an adequate approximation. 186-188 and pp. . 1 Two-term odd series fit to I <pI <p +~ Theory for analysis of roll extinction data In order to assessfurther the approximation of equation (7). and that angles are small. Because both the fitted coefficients are positive in equation (6).. assuming that damping is small.(ip) = cubic damping function model N 21 = coefficient of I~r ~rt of qusdratic roll damping mo&l N 22 = coefficient of quadratic damping model part of roll T = roll period.3. However. 192-196) derives the method according to the first approximation of Kryloff and Bogoliuboff.-. Only the positive half of the quadratic form is shown and the evaluation is made for normalized I (PI (p. two things seemed in order. Takaki and Tasai [12] performed forced oscillation experiments of a ship model with bilge keels. what are considered realistic estimates of N21 and N22 are almost totally empirical. these seem invariably to come out too low. and the second was to consider what violence to the left-hand side of equation (1) was being done relative to observed behavior of ships.j +3 1 (with j = 1.2 + C<k j + 1) . Though theoretical estimates of N 21 are made in the more sophisticated of modern ship motions algorithms.0 0.m = nondimenslonal t = time time cp = roll angle ipc = assumed range of roll velocity w = frequency.4 <Jl/4>c Fig. The maximum deviations within the range of the fit (± (Pc) amount to 3 or 4 percent of cIe. In this case the (K + 1)/2 normal equations in the unknown coefficients C<k become K k=I~ . . theoretical estimates of N 21 or N 22 appear for practical purposes to be totally lacking. The fitting equation was therefore assumed in the following form: 1(p1(p = k=I. roll frequency 179 1978 .------- ----- . the typical procedure being to add something to account for" viscous" damping effects so as to bring computation into line with observation. (k 0. period Y = roll amplitude Yo >I!: initial roll amplitude experiments 1'1. . substitution in the damping function. Conceptually... then the problem is to relate these to the coefficients of an odd expansion. that excitation is zero. equation (3).3.d = coefficients in cubic fitting equation for roll excitation per half cycle F(t) = excitation I = roll inertia N(Ip) = damping function l:!(ip) = cubic approximation to quadratic roll damping model SEPTEMISER --n1omenclature !:l. As noted in the Introduction.8 I: C<k(pk/((Pc)k-2 _ (4) It was further assumed that the odd series is truncated at the term with exponent K. Minorsky [13] (pp. If it is assumed that the coefficients N 21 and N 22 in equation (3) are given. the numerical results are obtained by analysis of ship or model sallying experiments. The first was to find some realistic magnitudes for N 21 and N 22.b = coefficients = coefficients in cubic fitting equation for roll extinction per cycle in quadratic fitting for roll extinction per N 31 = coefficient of linear part of cubic roll damping model N 33 = coefficient of cubic part of cubic roll damping model 01002 = coefficients equation cycle n. The basic result is the differential equation of the amplitude of response. there results (9) Equation (9) is assumed to be an adequate representation of the _______________ in quadratic fitting equation for roll extinction per half cycle l1(cp) = restoring function B 1 = linear coefficient of restoring function c.

equation (7). Vossers [16]. in equation (10): The question is (apart from validity of empirical constants). equation (21).exp[-am]) a and. as before.' +N21cP to one. The result in terms of the notation of equation (15) may be written: Y= 1 + .dY = aY + by2 (15) dm so that in this system N 21and N 22may be estimated from equation (14) with 01 = 2a and 02 = 2b. of the form + N22101 cP= 0 . For example. the fitting equation analogous to equation (15) might be written ._1_ 2'111w J« . equation (13). It appears that there was no dispute at the time about the representation of real fluid effects upon roll as being quadratic.. Some analyses of ship-sallying data A review of recent texts (Lewis [14]. (11) (12) The typical analysis of the curve of declining angles obtained in a ship sallying experiment amounts to estimating dY / dn (the decrease in roll amplitude per cycle) as a function of mean amplitude and then choosing coefficients to make a best fit of the data to an equation of the form of equation (12). the estimates of NSI and Nss from the experiment would become NS1 = y = . will such a change imply responses which do not reconcile with observation? Accordingly.. estimates of NS1 and NS3 could be formed from equation (22) by letting 1'1 = 2e and 'Ys = 2d.'Ys = -'Ys 3'11"w 3'11"ws (22) In Froud's method of fitting with half periods. the equation of declining angles from the ship sallying experiment in which the damping is assumed to be the cubic model may be derived from equations (20). and the cubic extinction model. The fundamental result from Minorsky (13) for the differential equation of roll amplitudes is . the coefficients 1'1 and 'Yswould be chosen so as to best fit the data to (21) Accordingly. once the numbers in the approximate damping function. Having estimates of 01 and 02 from the data. as with the quadratic model. only about the mixed JOURNAL OF SHIP RESEARCH 1. making the time variable change as was done to produce equation (12): dY = _ 'll"NSl Y _ 3'11"Nssw ys (20) dn lw 41 Imagining that the results of a sallying experiment might be analyzed in a way similar to that for the quadratic model [equation (13)]. the original equation Accordingly. and much of the available data are presented in this way.dY = 01Y dn lw -'II" 1'1 = m» Bl 1'1 41 4Bl Nss = -. should be small relative to the range and accuracy of data.]i Y = roll amplitude in the sense that the rolling is represented in the form. that the initial roll amplitude is Yoat time (m) = 0.-. y = _ NSI Y _ 3Nssw2 ys [N21Ywcos~ + N22Y2w21 cose] cosejcose de (11) = _ N21 Y _ 4N22W y2 21 3'111 It is convenient for analysis of sallying data to make a change in the time variable in equation (11). and Blagoveshchensky [17]. for example) fairly convincingly indicates the originator of the mixed linear-pius-quadratic roll damping representation to be Froude [2].. Y = .2. Similarly. equation (3). Given equations of the form of (12).11iiUJ~~eQ lcp + BI'. the equation of the curve of declining angles from the sallying experiment may be found [13].dY = eY dm + 02y2 _ (13) is the form often used. the form of equation (1) changes. or (15). the differences between the quadratic. Assuming. Korvin-Kroukovsky [15]. for the zero-excitation small-angle case as assumed for the ship sallying experiment. have been chosen. I{-{t) = Y(t) sin(wt + constant) Substituting the quadratic damping function.1 2'111w lip + Bgo + NS1Cp + Nsscp =0 (17) 50 0 2 .Yo (1 . In particular . extinction model. the result in the notation of equation (23) comes out to be Y= {I + ~ Y5 (1 - Yo exp[ -em] exp[-2cm]) r (24) 2 -b Yo exp[-am] (16) where the initial conditions of roll amplitude = Yoat time m = 0 have been assumed. That is. if the curves of declining roll angles are examined. equation (24) for the cubic model should not differ radically from equation (16) for the quadratic model within the range of observation. estimated values of N21 and N22 may be formed N21=-01=-01 'II" lw B1 'll"W + dYs (23) 31 3B1 N22 = -02 = -022 (14) 8 8w Froude found it convenient to estimate the decrease in roll per half cycle (dY / dm). let t =nT where T = roll period = 2'11"/ Making this change in equation w. In this case the data are reduced according to . In this case the damping function would be li_(0) = NS10 + Nsscp (18) Substitution in equation (10) results in the differential-equation of rolling amplitude (19) 21 81 or. it is of interest to make believe that equation (17) represents the ship in the sallying experiment. if an approximation of the nature of equation (7) is reasonable relative to available observations. N(Y w cose) cose d ~ (10) where w = roll frequency = VB. Turning to the second point mentioned at the outset. (21). Finally.ship during the sallying experiment. or (23). (13).

This experiment is notable for the large initial roll amplitude (32 deg). 16 CUBIC MODEL. effort was made to produce fits having comparable root-mean-square (rmsl deviations from the data. 5. which. 24 40 o ____ 20 EXPERIMENT QUADRATIC MODEL. Eq. The fit of equations (16) and (24) to data for HMS Volage is shown in Fig. 220 of reference [2]) involve the presumably faired points on the curves of declining angles for five sallying experiments on four ships. 2 through 5 for the data in reference [2].05 deg from these plots. pp. In none of the cases shown would a purely linear damping function 181 1978 . Eq. The data (which appeared in Naval Science. and that of the cubic model. Eq. The first point of the exercise was to make a comparison between the fits of equations (16) and (24) to the original data. Eq.: 2 0 5 10 15 COUNT OF HALF CYCLES 20 25 3 Fig. 16 __ -. Since some of the experimental data used by Froude in demonstrating his point are currently available [2]. TWO EXPERIMENTS QUADRATIC MODEL. The results are shown in Figs. 24 Vl UJ UJ Vl cc '-' UJ 0 UJ UJ UJ cc '-' o . Each experiment was analyzed separately. excluding data from amplitudes less than 1 deg since the values could not be read much closer than ±0. Figure 2 indicates curves of declining angles for two experiments on HMS Inconstant.2 5 10 20 25 15 COUNT OF HALF CYCLES 35 COUNT OF HALF CYCLES Fig. equation (24). HMS Devastation. One of these results was picked arbitrarily. is unusually long. These were in the form of various declining angle curves plotted to relatively small scale. in Figures 2 through 6.3 Fitof equations (16) and (24) to curveof decliningangles obtained in an experiment with HMS Volage in 1871 2 o 5 20 10 15 40 30 35 COUNT OF HALF CYCLES 25 45 30 model. as in those to follow. owing to the ship's light damping in roll. Gawn [18] presents results of several sallying experiments carried out by Froude on an ironclad warship. Eq.U 0:: UJ 0 Vl o 0 6 5 4 3 EXPERIMENT QUADRATIC MODEL. it was of interest to start with these. (Froude had combined the two graphically in his analysis [2]).10· o.1 deg. 5 Fitof equations (16) and (24) to curveof decliningangles obtained in experimentwith HMS Sultan in 1873 found.: -' c:: 0 '-' '" I. -' 0 10 8 6 5 4 c:: c:: 1~ __ ~ 3 -L ~ L_ __ ~ _L __ ~ 0 5 10 15 o Fig. Eq. 16 CUBIC MODEL. 24 '-' 10 8 7 6 5 4 I.U 0 -' 0 c:: . Within the range of data. The logarithmic scale in this figure. 6. all of which apparently were built with bar keels.: -' 0 3 2 :. On the whole. The resulting data were subjected to the procedure outlined in the foregoing. 1874. Figure 5 involves the declining angle curve for HMS Sultan. The experimental results had been plotted to an enlarged scale and it was possible to measure the roll amplitudes observed at each half cycle to within ±O. and accordingly the results were broken into two parts for plotting purposes in Fig. to represent the original data. Eq . In both cases.U I. an indirect semi-trial-and-error approach was utilized.CUBIC MODEL. Because it was not possible to quickly develop a direct least-squares fitting procedure for equations (16) or (24). Eq. 16 CUBIC MODEL. equation (16).4 Fitof equations (16) and (24) to curveof decliningangles obtained by two experiments with HMS Inconstant in 1871 20 Fitof equations (16) and (24) to curveof decliningangles obtained in an experiment with Elorn in 1872 9 8 7 I.U UJ . 3. either would serve as a reasonable interpolator. Figure 4 involves fits to data obtained by French experimenters on a ship named Elorn. 24 Vl EXPERI MENT QUADRATIC MODEL. tends to accentuate reading errors for the lower ranges of roll amplitude. and it is regretted that the details of the ship or experiment could not be SEPTEMBER Fig. with results given in Fig. and the curve was read at each half cycle. there is not much to choose between the abilities of the quadratic model.

Eq. More modern results from sallying experiments are presented as roll extinction curves. The differences between the quadratic and cubic models are clearly less than the scatter of original data. one was abstracted for analysis and this was the case where their model was bare (without bilge keels or artificial 2 o FROM FAIRED EXPERIMENTAL QUADRATIC MODEL. SHIP RESEARCH 182 . In both cases Gawn [18] presents faired lines representing -dY I dm as a function of Y. a selection of the available model experimental data was made and similar analyses carried out. Just as in Fig. Values of dYI dm were originally derived from each and the results were pooled to form some 50 pairs of dY I dm and Y. Of the many model roll extinction experiments carried out by Martin et al [20]. Figure 10 shows these data and the least-square fits of equations (13) and (21) (which are the equations corresponding to the roll extinction per whole cycle). ROLL AMPLITUDE. values of -dY I dm were taken off at equal intervals of Y and the results treated as data to which equations (15) and (23) were fitted by the method of least squares. the results suggest that the replacement of the classical damping model. equation (7). Figure 11 indicates a similar result using model roll extinction data for one experiment presented by Blagoveshchensky [17] on page 640 of his text. a somewhat surprising result in view of the likelihood that the original fairing of the two sets of data was done to the quadratic model.0 and the results analyzed or faired by the experimenter prior to presentation in the form of a plot of -dY I dm versus Y. There is again not much to' choose between the fits. Least-square fits of equations (15) and (23) were made to the pooled data and the resulting fitted lines are also shown in Fig. ROLL AMPLITUDE. 23 RESULTS dY . 9. 15 CUBIC MODEL. the results in Figs. These points were measured from the chart presented by Gawn and appear as circles in Fig. 66 of his text. Vossers [16] presents a set of unfaired model roll extinction data for one experiment in Fig. 9 for ship data. Eq. The ship was a destroyer of a 1935 design (HMS Nubian). Fig.o EXPER IMENT QUADRATI C MODEL. . The resulting fitted lines are compared in Figs. Figure 12 is a third similar analysis for the pooled results of a number of experiments by Lalangas [19] on a Series 60.60 block parent model at zero speed. and it is strongly suspected that neither an assumed purely quadratic nor an assumed purely cubic damping function would yield comparable fits. Eq.dm 0·5 o Fig. The results presented by Gawn [18] of experiments on a third ship are of the same form but unfaired. There were six sallying experiments carried out. the numerical differentiation which provides the decrease in roll per cycle or half cycle is performed 1. that is. 15 CUBIC MODEL. equation (3). 16 CUBIC MODEL. 24 V) LU W w a: '-' _j' _l '" a: 3 0 2 o Fig. For present purposes. Eq. Relative to the present objectives. 8 DEGREES Fit of equations (15) and (23) to roll extinction curve obtained for HMS King George Vin 1914 Fit of equations (15) and (23) to roll extinction curve obtained for HMS Vivian in 1925 JOURNAL OF. with the cubic approximation. 10 through 12 for model data indicate that the differences between the quadratic and cubic models are less than or the same as data scatter. 23 RESULTS o dY dm FROM FAIRED EXPERIMENTAL QUADRATIC MODEL. Eq. 0. The first two sets to be discussed are from experiments carried out on a 1912vintage battleship (HMS King George V) and a 1918 destroyer (HMS Vivian). Eq. Gawn [18] presents three such sets of ship data. 7 2 3 Y. can result in a quite reasonable approximation to the transient roll response so long as an appropriate choice of coefficients can be made. 7 and 8 with points representing the original curves. 9. For present purposes. 6 5 10 COUNT OF HALF CYCLES 15 20 Fit of equations (16) and (24) to curve of declining angles found by experiment with HMS Devastation in May 1873 result in as good a fit. 4 5 DEGREES 6 Y.

-8"'--. Eq. Eq. This line was read off at even increments of roll amplitude and the values appear as circles in Fig. Fit of equations (13) and (21) to model roll extinction data presented by Lalangas 183 . In this instance the cubic fit appears slightly better than the quadratic. Eq. Eq.y 0 ~ o~ o~ 0 Fig. ROLL AMPLITUDE. ROLL AMPLITUDE. the faired data was read off at roughly uniform intervals of roll angle and the result treated as data in least square fits of equations (13) and (21). The least-square fits of both the quadratic and cubic models to the points representing the faired line are also shown. The cubic model yields a physically believable result.----"--_"___-----' 2 /. 15 CUBIC MODEL. 13. the quadratic model does not.5 o o o ~ 0 0 _. 12 5 10 DEGREES V. 23 -Tm o Fig. DEGREES Fit of equations (15) and (23) to roll extinction data obtained in experiments with HMS Nubian 2 6 0 EXPERI t\ENTAL DATA QUADRATIC MODEL. 11 5 10 V. 13 CUBIC MODEL./cf / 3 '" / 0 (7 . 21 o EXPERIMENTAL DATA QUADRATIC MODEL. Figure 14 indicates the results. SEPTEMBER 1978 2. As in the previous example. Eq. Fig.dn - dY dn . Eg. 116) and is the equivalent of a roll extinction curve. The result was obtained by Motora et al [21] by means of a forced-oscillation technique. Eq. The last example picked was presented as a faired experimental result by Vossers [16J (p. o Fig. and is notable for the extremely large roll angles involved. o 4 6 8 DEGREES 10 12 POOLED EXPERIMENTAL DATA V. ROLL AMPLITUDE. 21 dV . This extinction result was presented as a faired line. 13 CUBIC MODEL. 21 o roughness) and was at zero speed. 10 Fit of equations (13) and (21) to model roll extinction data presented by Vossers QUADRATIC MODEL. 15 DEGREES 20 i Fit of equations (13) and (21) to model roll extinction data presented by Blagoveshchensky I o o L 5·0 ~______J__--. ROLL AMPLITUDE.9 2 3 Y.2 o dY POOLED RESULTS OF SIX EXPERIMENTS QUADRATIC MODEL. 13 CUBIC MODEL. Eq.

0 309. the models may be considered equally good within the limitations of observable data.0 0. This magnitude of error 184 The reason for embarking on the present work was that the quadratic time-domain representation for roll damping that has been in use for the past century is a serious analytical obstacle which. In the absence of such other considerations. A mixed linear-plus-cubic model.94 6.0 42. In fact. must be overcome if improvements in techniques for the prediction of nonlinear rolling in random seas are sought. 14 DEGREES Fit of equations (13) and (21) to model roll extinction data derived from forced oscillation experiments by Motora et al Table 1 indicates some of the particulars of the ships and models involved in Figs." .59 17. Ideally. 2 through 14. The basis for the acceptance of the mixed linear-plusquadratic time-domain roll damping model is almost entirely empirical. Concluding remarks Y.214 0.9 2990 2.7 9.67 Martin Displacement. for the large-amplitude results of Fig. It was found that the maximum error of the approximation might be expected to be 2 or 3 percent of the value of the function at the maximum experimental roll velocity. Eq. 13 CUSIC MODEL. and can be made to be quantitatively reasonable within the range of validity of the quadratic model.0 Nubian 36. ROLL AMPLITUDE. In some of the results exhibited in Figs. The results of analyses indicate that this approximation is both quantitatively and qualitatively reasonable within the range and scatter of available experimental data. 1 ton = 0.8 1.0148 0.51 oL~d'~ 0 2 Metric conversion factors: 1 ft = 0. Denis and Pierson [IJ.5 11. 2 3 5 7 8 9 11 12 13 Length.34 1. 2 through 14 . Eq. ROLL AMPLITUDE. Beam ft ft Name/Source Inconstant 337.9 metric tons . DEGREES 4 Fig.8 2. the data analyzed originates from a variety of ships and models. 21 RESULT / .67 Lalangas 0.0333 0.3 270.2 0 Table 1 -Tn dY FROM FAIRED EXPERIMENTAL QUADRATIC MODEL. it is expected that the damping function will be odd in roll velocity and positive for positive roll velocity.3048 m.0 Sultan 555. The additional result that a linear-plus-cubic roll damping model fits experimental data about as well as (sometimes better than) the linear-plus-quadratic model gives rise to speculation that the cubic model might be closer to an "equivalent approach" than to an "approximation. Fig. the choice between the two must be made on a basis other than the fit itself. numerical comparisons were made in[9] between the values of damping moments implied by the quadratic model. analogous to the linear framework of St./ff if Figure No. 2 through 14. was proposed as an approximation. as are what are taken to be realistic coefficients in this model.22 Blagoveshchensky 5. It appears from the results that the cubic approximation is qualitatively correct.0 59. Taking ships and models together.51 0. and it therefore appears that the approximation. in which the coefficients are functions of those of the linear-plus-quadratic model. JOURNAL OF SHIP RESEARCH .initial angle in the sallying experiments ranged from 5.53 1185 2. equation (3). what would be very useful in the prediction of ship rolling statistics in random seas is a prediction framework. Particulars for two ships and two models could not be identified.3 2541 2. equation (7). Eq. The primary consideration in the long history of preference for the quadratic damping representation seems to be that the drag on a body in a real fluid is proportional to velocity squared if the velocity is high enough.5 9205 25550 5.018 0. The degree of fit to observable data shown in Figs.0 King George V 89.0 29. 14 the fit is much more realistic. both are extrapolations. Eq.65 14. GM (jt) tons 5200 2. Using the values of coefficients derived in making the fits of Figs. 13 Fit of equations (13) and (21) to model roll extinction data presented by Martin 50 o 40 FA IRED EXPER IMENTAL RESULT QUADRATIC MODEL. can be justified. The. Physically.35 1.6 m). As may be noted in Table 1. If there are two analytical models which fit the observable data with roughly the same magnitude of error.J Particulars of some of the ships and models for which comparisons are made in Figs.1 9.0 Volage 325. 2 through 13 the fit of the cubic model is slightly superior. The type of appendages ranged from nothing at all through bilge keels on relatively modern forms to bar keels in warships of a 'century ago. the lengths ranged from 5 to 550 ft (1. it appears.5 Vivian 364.5 to 32 deg.2 2.96 0.091 T s 15. o - dn dY 30 20 'I 'I 'P t tJ 10 appears comparable with the magnitude of experimental uncertainty. equation (7). 12 14 8 10 6 Y. Both the mixed quadratic and cubic models can be made to fit this criterion through a choice of coefficients. 2 through 14 for the cubic model suggests another approximation to roll damping. outside the range of data. 13 CUBIC MODEL. and the approximation.97 11. 21 Ij I /. in which the following could be accomplished: (a) At least weak nonlinearities could be accommodated for the general multidegree-of-freedom situation. (b) Multidirectional seas could be considered as input. which is quite simply to assume that it follows the cubic model [equation (18)J in the first place.5 to 167.0 50.

. Stevens Institute of Technology. 14. (d) For economy. Heft 36. "Ship Rolling at Zero Speed in Random Beam Seas with Non-Linear Damping and Restoration... "Derivation and Application of the Fokker-Planck Equation to Discrete Nonlinear Dynamic Systems Subjected to White Noise Random Excitation. 1955. L. 1971. 230. Davidson Laboratory. 1963. W. T. "Optimum Design of Passive Roll Stabilizer Tanks. West Japan Society of Naval Architects. Vol.. Technical Publishing Company. J. B. N. 15. 18 Gawn. "On the Measuring of the Damping Resistance of Roll Through a Large Angle by a Forced Oscillation Method. No. S... some success was achieved in evaluating the spectrum of roll in a way which is relatively straightforward in the context of predictions of nonlinear random response. 150.. SNAME. ITTC. 20 Martin.1967.. "Lecture Notes on Non-Linear Theory of Ship Roll Motion in a Random Sea Way. 1973. Theory of Seakeeping. H. M... Vol." The Institution of Naval Architects.. F. and Landweber. there remains a host of problems. 13 Minorsky.. L. 1960. 1962. "A Note on the Form of Ship Roll Damping." Journal Zosen Kyokai. No.' International Shipbuilding Progress.1957. Behavior of Ships in Waves. "On the Motions of Ships in Confused Seas.alidated in principle. The Netherlands." SITDL-76-1887. Dover Publications. (e) The statistics of maxima could be estimated with firmly based theory in which the possible effects of nonlinearities are accounted for. 20. 8 Vassilopoulos. AD-A031 048/261. No. F. McLeod. SEPTEMBER 1978 185 . R. R. "A Modified Approach for the Application of Fokker-Planck Equation to the Nonlinear Ship Motions in Random Waves. 1973. 9 Dalzell.. "On the Hydrodynamic Derivative Coefficients for Lateral Motions of Ships. Vol. N. 5 Yamanouchi. May 1976. John P. Oct. L. Starn. Oct. Aug. No. 35. The Institution of Naval Architects. the prediction or a major part of it could be carried out in the frequency domain (as well as to take advantage of the accumulating frequency domain descriptions of real sea waves). M. Propulsion and Steering of Ships. London. 3 Kaplan." Tmns. V. and Nishikido. N. lIC of ReSistance. 1963. New York. Vol. E.. 22 Dalzell. H. 17 Blagoveshchensky. "The Motion of Ships in Waves. G. 1953. 14 Lewis. 61.. 6 Caughey." Trans. 1966. 1961. Vol.4. Tokyo. W." International Shipbuilding Progress. Y. By means of the linear-plus-cubic damping model. Stevens Institute of TechnololIT. Comstock. the most important of which is the possibility that the refinements possible in any of the mathematical prediction frameworks which have been suggested may overreach the validity of the assumed physical model for rolling throughout the practical range..1940." Report 983." Schiffstechnik. 242. "On the Effects of Non-linearity of Response on Calculation of the Spectrum. M. Michigan. ' 10 Haddara. 1967. No. 11 Lewison.(c) The hydromechanic data required could be produced with conventional techniques. Stevens Institute of Technology. G.. SNAME. "Estimation of the Spectrum of Nonlinear Ship Rolling: The Functional Series Approach. Edwards." Chapter IX of Principles of Naval Architecture. Denis. Vol. "The Application of Statistical Theory of Non-Linear Systems to Ship Motion Performance in Random Seas. Band 7.. V. Davidson Laboratory. Jan. 4 Vassilopoulos. 100. and Tasai. SNAME. No." SIT-DL-76-1894. RINA. Effect of Roughness on Ship Rolling. References 1 St. 1962.. ITTC. 19 Lalangas. 1947. 7 Haddara. In practice. Ed. V. 21. 1976. Introduction to Non-Linear Mechanics. and Pierson." The Naval Architect." International Shipbuilding Progress. F. R. 2 The Papers of William Froude. Theory of Ship Motions. L. G. . 12 Takaki. 11. 15 Korvin-Kroukovsky. K. G.. "Rolling Experiments with Ships and Models in Still Water." JOURNALOF SHIP RESEARCH. J.. "On Nonlinear Rolling of Ships in Random Seas. S.. 46. J. 1966. P. 1974. Ann Arbor.. Davidson Laboratory. M. AD-A031 055/761. The practical significance of the present results is thought to be that an additional approach to the ideal prediction goal is v.. J. 82. P. "Application of Linear Superposition Techniques to the Roll Response of a Ship Model in Beam Irregular Seas. R W. 21 Motora. Shimizu." Trans. The present work was the beginning of a modest attempt to make a contribution to this much larger problem -by investigating the applicability of the functional series model to single-degree-of-freedom ship rolling [22]. May 1976. Vol. Oct." Trans. A. Nov. Dec." Journal of the Acoustical Society of America. Vol. This goal does not appear to be in hand. M. T. Jr.Vol. 16 Vossers..

Sign up to vote on this title
UsefulNot useful