Onderwerp: Bezoek-historie

1200 - Interim Guidelines For Alternative Assessment Of The Weather Criterion
Geldigheid:24-05-2006 t/m Status: Geldig vandaag

Dit onderwerp bevat de volgende rubrieken.


Ref. T1/2.04                                                                                                            MSC.1/Circ.1200

24 May 2006



1          TheMaritimeSafetyCommittee,atitseighty-firstsession(10to19May2006),approved Interim  Guidelines  for  alternative  assessment  of  the  weather  criterion,  aiming  at  providing the industry  with  alternative  means  (in  particular,  model  experiments)  for  the  assessment  of severewindandrollingcriterion(weathercriterion),ascontainedintheCodeonIntactStability for  All  Types  of  Ships  covered  by  IMO  Instruments  (resolution A.749(18)).           The  Interim Guidelinesshouldbeappliedwhenthewindheelingleverand/ortheangleofroll(asdefinedin paragraphs  and  of  the  Code)  need  to  be  determined  by  means  of  model experiments.


2          Member  Governments  are  invited  to  bring  the  Interim  Guidelines  to  the  attention  of interested parties asthey deemappropriate.







1 Introduction

1.1 The purpose of these Guidelines is to provide alternative means for the assessment of severe wind and rolling criterion (weather criterion) as reported in paragraph 3.2 of the Code on Intact Stability for All Types of Ships covered by IMO Instruments (resolution A.749(18)). In the following guidelines, the angle of roll is referred as f, while in the Code the angle of roll is referred as ?.

1.3                               The Guidelines provide procedures for the determination of the wind heeling lever lw1, as defined in paragraph of the Code, by means of direct measurements.

1.4                               In addition, the Guidelines are given for the experimental determination of the angle of roll f1 as defined in paragraph of the Code.

1.5                               For quantities used but not defined in the following, the definitions of the Code apply.


2 Application

2.1        The tables and formulae described in paragraph of the Code are based on data from ships having:

.1        B/d smaller than 3.5;

.2        OG/d between -0.3 and 0.5;

.3         T smaller than 20 s.

2.2              For ships with parameters outside the above limits, the angle of roll (f1) may be determined with model experiments of a subject ship, following the procedure described in the Guidelines as the alternative.    In addition, the Administration may accept such alternative determinations for any ship if deemed appropriate.

2.3              The alternative means for determining the wind heeling lever (lw1) may be accepted, to  the satisfaction of the Administration, as an equivalent to calculation in paragraph of the Code. When such alternative tests are carried out, reference should be made to the relevant part
of the Guidelines. The wind speed used in the tests should be 26 m/s in full scale with uniform velocity profile. The value of wind speed used for ships in restricted services may be reduced to the satisfaction of the Administration.

2.4              The test programme should be approved by Administration in advance.

2.5              Tests should be documented by means of a report and a video or other visual records containing all relevant information on the model, the procedure and the test results, which should be approved by the Administration.

2.6              Any procedure different from those provided in the Guidelines should be subject to the approval of the Administration.


3 Guidelines For Experimental Determination Of The Wind Heeling Lever Lw1

3.1 Objectives And Definitions


3.1.1    The purpose of the tests addressed in this section is to ensure uniform applicability of model tests for the determination of the steady wind heeling lever, lw1 (paragraph of the Code). It is allowed by this procedure to consider the steady wind heeling lever as dependent on the heeling angle (see figure 3.1.1).


Figure 3.1.1: Weather criterion when the wind heeling lever is dependent on the heeling angle


3.1.2 The standard model test procedure consists of two parts. The first part is a procedure for estimating the heeling moment Mwind due to steady wind in a wind tunnel. A blower may be used as an alternative as long as the uniformity of wind speed is comparable. The second part addresses the estimation of the heeling moment Mwater due to steady drifting in a towing tank.

3.1.3    The steady wind heeling lever, lw1 , is obtained by means of the following equation:



The equation 3.1.3 assumes that the wind force and the drifting force work as a couple. In that case the heeling moment Mw is independent on the point of reduction of the system of forces. However, due to the unavoidable unbalancing of vertical forces arising from direct measurements, the total heeling moment Mw may depend on the point of reduction. For practical purposes, it is considered sufficient to calculate all moments with respect to the point O given by the intersection of the ship centreplane and the waterplane.


3.1.5 Mwind is obtained at full scale by appropriate scaling of results from wind tests carried out as indicated in paragraph 3.3. Mwater is obtained at full scale by appropriate scaling of results from drifting tests carried out as indicated in paragraph 3.4.


3.2 Model Set-Up

3.2.1    Ship model used for wind tests

The model should copy the above-waterline shape of the actual ship and should comply with the following:

. 1        the overall length should be at least 1.25 m;

.2 all sharp corners in the actual ship should be sharp in the model to simulate separated flow;

.3 main fittings on the exposed decks and superstructures, e.g. cranes, masts, bulwarks, should be modelled and fitted properly;

.4 the size of the model should be determined to make the blockage ratio to the wind tunnel less than 5%, where the blockage ratio is defined as the ratio between the lateral projected area of the model above the waterline divided by the area of the test section of the wind tunnel; and

.5        when a blower is used the ship should be within the area of uniform wind speed.

3.2.2    Ship model used for drifting tests

The model should copy the under-waterline shape of the actual ship and should comply with the following:

. 1         the size of the model should comply with paragraph 4.3.2;

.2         not only underwater fittings (e.g. bilge keels, rudders, etc.) but also potential underwater part when the ship heels (e.g. bulwarks, freeing ports, etc.) should be modelled and fitted properly.

3.3 Wind Tests

3.3.1    Wind characteristics

The wind speed should comply with the following:

.1 The minimum wind speed to perform tests should be over the critical Reynolds' number, after which CD is constant.

.2 The wind speed profile should be as uniform as reasonably possible. Except for the boundary layer in the vicinity of the end plate (figure 3.3.1), spatial deviation1 of the wind speed should be less than 1%.

.3 The effects of end plate (due to its shape, size, roughness, etc.) and of the gap between end plate and model should be minimized.




Figure 3.3.1: Example of an arrangement for tests in wind


3.3.2    Complete test procedure

The lateral horizontal force Fwind (and corresponding drag coefficient CD ) and the heeling moment due to wind Mwind with respect to O are obtained by a wind tunnel test or in wind from a blower. In calculating CD according to equation (3.1.4), the actual value of air density during tests should

be used. An example of model test arrangement is shown in figure 3.3.1. Model tests should be carried out in compliance with the following:

. 1 Before tests are carried out, the vertical and horizontal distribution of the wind speed at the model position should be verified.

.2 Tests should be carried out in upright condition and at some heeling angles with appropriate increment to lee and wind side covering a sufficient range of heeling angles to the satisfaction of the Administration. 

.3 In heeled conditions the model shape exposed to wind should be the same as the above-water shape when the ship is floating freely. The change of trim due to heel can be neglected.

3.3.3   Simplified test procedure

As an alternative simplified procedure, the lateral horizontal force Fwind (and corresponding drag coefficient CD ) and the heeling moment due to wind M^ with respect to O can be obtained for the upright condition only and considered as constants (not depending by heeling angle).



1      Spatial deviation is the variation of wind speed in longitudinal direction referring to the main flow and should be measured for the test section without the model.



3.4 Drifting Tests

3.4.1    Complete test procedure

The heeling moment Mwater due to drift with respect to O is obtained by means of towing tank tests. An example of experimental set-up is shown in figure 3.4.1. Model tests should be carried out in compliance with the following:

     .1     the ship model should be attached to a guidance system, which allows the model's free sinkage (an example of experimental arrangement is shown in figure 3.4.1); towing direction is to be at right angle to the longitudinal direction of the model (heading angle 90°); the towing speed should ensure that the measured drift horizontal force is equal to Fwind scaled with the appropriate scaling law. Fwind should be calculated by equation (3.1.4) using the measured drag coefficient in paragraph 3.3 and the assumed wind speed as prescribed in paragraph 2.3; and tests should be carried out in upright condition, and at some heeling angles with appropriate increment to lee and wind side covering a sufficient range of heeling angles to the satisfaction of the Administration.


  • Figure 3.4.1: Example of an arrangement for drifting tests


    3.4.2   Simplified test procedure

    As an alternative simplified procedure, the moment Mwater due to drift can be considered as given by a force equal and opposite to Fwind (f) (as following from paragraph 3.3.2 or 3.3.3) acting at a depth below waterline equal to 0.5 d (where d is the ship draught in upright position).


3.5 Combined Use Of Simplified And Complete Procedures

The combination of complete procedures and simplified procedures can be used.

3.6 Additional Considerations

The steady wind heeling lever, lw1,  is evaluated by means of equation (3.1.3).  When extrapolation is needed outside the tested range of heeling angles, such extrapolation should be carried out to the satisfaction of the Administration.

4 Guidelines For Experimental Determination Of The Angle

4.1 Objectives And Definitions

4.1.1    The objective  of the tests  addressed in this  chapter is the  determination of (paragraph of the Code).   The "angle of roll to windward due to wave action"  is defined, according to weather criterion, as follows:


where is "regular waves roll-back angle", that is the resonant roll amplitude in beam regular waves (heading 90°) having steepness defined in the following sections. The reduction factor 0.7 takes into account the actual irregular nature of the sea.

4.1.2        The standard procedure for the determination of  is that by means of tests in regular waves. The use of alternative procedures is permitted. Sufficient justification should be provided to the Administration regarding the selected procedure.

4.1.3        As a reference in selecting the more suitable procedure it should be noted that:

.1 The direct measurement of  (see paragraph 4.5) can lead to the need of generating very steep waves close to the breaking limit if the ship roll period is very short (see table 4.5.1). Generation of waves with such steepness and sufficient quality can be sometimes difficult due to breaking close to the wavemaker. In addition, in carrying out roll tests, care should be taken during the transient ship behaviour before steady state is reached, because possible large heeling angles (sometimes eventually leading to capsize) can occur. It should be underlined that  is the steady state maximum roll angle, for this reason capsize during initial transient phase of the test does not necessarily lead to not fulfilment of the criterion. It should be underlined that the methodology does not allow for corrections for scale effects on roll damping, and for this reason large models should be preferred when direct measurements are carried out.

2 The there-step procedure (see paragraph 4.6.1) is the simplest among the two proposed alternative procedures. This method was adopted when original weather criterion was developed. The procedure is sensitive to the quality of execution and analysis of roll decay tests. The procedure allows to execute tests for the determination of the effective wave slope coefficient r , with reasonably small steepnesses, leading to rather simple tests. The methodology allows, in principle, corrections for scale effects on roll damping.

.3 The parameter identification technique (PIT) (see paragraph 4.6.2) is a procedure with a large degree of flexibility, that allows to take into account nonlinearities of both damping and restoring, and that provides means for allowing frequency dependence of the "effective wave slope function". The methodology allows, in principle, corrections for scale effects on roll damping. When used with only one series of tests for one single wave steepness, the number of free parameters should be reduced to guarantee robustness of the methodology. The method can take great advantages (regarding robustness and accuracy) from the execution of more than one series of tests at different wave steepness: for this reason the use of at least two different wave steepnesses is strongly recommended. To guarantee correct application of the method, a sufficient basic training of personnel on the theoretical background on which the method is based is needed.


4.2 Model Basin

The facilities of the model basin should be such as to avoid wave reflections and shallow water effects. The breadth of the basin should be larger than the over all length of the model plus 2 m. The quality of the basin should be subject to the satisfaction of the Administration.

4.3 Model Set-Up

4.3.1   Construction      The model should be built geometrically scaled up to the upper weather deck including forecastle and bulwarks and be sufficiently rigid with a smooth finish. The whole model (excluding free flooded spaces) should be watertight in order to guarantee hydrostatic properties.      All superstructures included in stability calculations or that are submerged during the tests should be reproduced to scale to ensure the model has the correct righting arm curve. Superstructures that do not submerge during the tests described below can be omitted.      Appendages such as bilge keels or rudder should be fitted, properly scaled and the report should state which appendages were fitted during the tests.

4.3.2    Scale

To avoid scale effect on roll damping, the model overall length should be at least 2 m. However, the model should be scaled up, if necessary, to make the breadth of the bilge keels greater

than 7 mm. For monohull ships having neither bilge keels nor sharp bilges2, however, the model overall length should be at least 4 m unless frictional effect on roll damping is corrected with theoretical methods described later, but in any case not less than 2 m or a scale 1:75, whichever is greater.

4.3.3    Ballast and weight distribution      The model should be ballasted to the appropriate displacement and loading condition for the ship.  To ensure correct displacement and attitude, draught marks or suitable gauges should be used. Weights should be adjusted to achieve the correct position of the centre of gravity.      Weight distribution should be such as to guarantee reasonable radius of gyration for pitch. Unsymmetrical weights distribution should be avoided as far as practicable.      Inclining tests should be carried out to verify that the value of ship's metacentric height GM corresponds to that of the actual ship within an error of 2% or 1 mm at model scale, whichever is larger.      In addition, depending on the information provided to the model basin, natural roll period Tf in water or roll radius of gyration in air, should be checked to correspond to that provided within an error of 2%.

4.3.4    Roll period Tf to be tested

The ship natural roll period should be used for tests. In case a sufficiently accurate estimation of Tf is not available at the time of tests, they should be carried out for a series of at least 3 different roll periods, from which the results can be finally interpolated for the actual ship roll period.


2     "Sharp bilges" used here means that bilge radius is smaller than 1% of the ship's breadth and the angle between piece-wise lines representing the bilge is smaller than 120°.

4.4 General Experimental Set-Up

4.4.1    Instrumentation

The instrumentation system should be appropriate to the model and type of test carried out. The use of non-intrusive measurement systems is recommended when feasible. If it becomes necessary to attach cables to the model then care should be taken to minimize interference.

4.4.2    Calibration

In order to ensure accurate operation of instrumentation, calibrations should be carried out and reported.

4.4.3    Measurements

Roll, and yaw if necessary, should be simultaneously measured and recorded as appropriate to the purpose of the test. Wave height measurements should be made for all tests with wave probes fixed in the tank.

4.4.4    Wave quality

Wave generation quality should be assessed for the waves corresponding to the minimum and the maximum frequency used in the tests. The wave elevation should be measured by wave probes positioned at least 3 locations along the length of the basin, spanning the drift range of the model. This should be done without the model because the model can disturb incident waves. When the measured double amplitude of the wave elevation converges to a certain value, this value should be regarded as the wave height, H, for each position. Variations in wave height and wave period should be within ±5% among the different measured positions for the same signal.


4.5 Tests In Regular Waves

Tests in regular waves are the standard procedure for determining the "regular waves roll-back angle" f1r. In some cases the direct determination of f1r is not feasible, as, for example, in case of large models having long natural roll period Tf. In such cases alternative procedures can be used as reported in paragraph 4.6.

4.5.1    Test conditions

The wave steepness (factor "s") should be selected from table 4.5.1.

Table 4.5.1: Wave steepness as a function of the full scale natural roll period


Ship roll period


Wave steepness s = H/?
































4.5.2   Direct measurement procedure Tests in regular waves can be used to directly obtain the "regular waves roll-back angle" f1r. f1r is the peak roll response of the ship in regular waves of steepness according to table 4.5.1. In order to determine f1r, the stationary roll motion amplitude should be measured for a sufficient number of frequencies around the natural roll frequency?0=2pTf. The following minimum set  of test points is recommended ??0 = 0.8, 0.9, 0.95, 0.975, 1.0, 1.05, 1.2, with ? being the frequency of the regular wave in rad/s. Additional measurements in the proximity of the response peak might be necessary to allow for an accurate determination of f1r especially in case of strong influence of righting lever non-linearity. During the tests the ship model should be positioned to be normal to the direction of the waves (90° heading angle). The heading angle of the model is either:

.1 fixed, with a guide attached to the towing carriage keeping the sway-heave-roll motion free from restraints. An example of experimental arrangement is shown in figure The carriage should trace the drift motion of the model induced by the beam wave's action. Draught, GM and Tf should be adjusted taking into account the effect of the guide; or

.2 controlled by guide ropes which are fitted to the model on the centreline at the stem and stern, in a symmetrical fashion and at a vertical height between the waterline and the centre of gravity. These lines can be used to correct the model in yaw while allowing drift and sway, provided the heading during tests does not deviate from beam sea for more than 15°. However, whenever the yaw motion is corrected by means of the ropes, the corresponding part of the measured record should be neglected in the subsequent analysis, unless the effect of correction on the quantities of interest is clearly negligible.



Figure An example of the guide for roll test in beam waves      During the tests, care should be taken to use appropriate time windows for the measurements, so that the steady roll amplitude is measured without the influence of reflected waves between the model and the wave maker or the model and the beach.      Data to be recorded are model motions in all measured degree of freedom (DOF) and wave elevation.






4.6 Alternative Procedures

Whendirectmeasurementof      isnotfeasible,alternativeprocedurescanbeusedtocalculate  theangleofrolltowindwardduetowaveaction

      atthesteepnessspecifiedin 4.5.1,bymeans  ofdataobtainedfromtestsinregularwaveswithdifferentsteepnessesand/orothertypeoftests.


In  view  of  the  strict  interrelation  between  the  many  elements  constituting  present  weather criterionassessment,theevaluationofindividualparametersrelevanttothecalculationformula of  the  angle  of  roll  to  windward  due  to  wave  action is  permitted  only  when  they  are  all evaluated  through  experimental  tests  or  appropriate  calculation  procedures.  In  the  following, procedures are reported as alternatives to the direct measurement of   (refer to paragraph 4.5).

4.6.1 Alternative procedures


4.6.2 Alternative Procedure 2: Parameter Identification Technique (PIT)

The  parameter  identification  technique  (PIT)  approach  is  outlined  below,  taking  into  account linear  and  nonlinear  features  of  the  mathematical  model  describing  the  roll  motion  in  beam waves,withotherforcingsourcesorrolldecays.Thebasicstructureofthemethodconsistsinthe regression  of  the  solution  (exact  or  approximate,  analytical  or  numerical)  of  the  system  of differentialequationsdescribingthetimeevolutionofthesystemunderanalysis,containingas unknowns  the  characteristic  parameters  (coefficients  of  the  mathematical  model  adopted  to describe  damping,  restoring,  forcing  terms).  The  regression  is  considered  to  the  experimental valuesofstationaryrollamplitudeversusfrequencyforforcedroll.Thebasicideaonwhichthe PIT  relies  is  thus  as  follows:  the  solution  of  equation  (,  for  any  consistent  set  of parametersanddifferentwavefrequenciesallowstoobtainapredictionfortherollresponse.The parametersofthemodelaremodifiedsystematicallybytheminimizationprocedureinorderto obtain   the   best   agreement   between   the   predictions   given   by   the   model   and   measured experimental  data.  The  “optimum”  set  of  parameters  is  then  obtained  and  used  in  solving equation(,to obtain,finally,thepeak      oftherollresponsecurve.Theangleofrolltowindwardduetowave


action    is calculated according to paragraph 4.1.



When  PIT  is  used,  at  least  two  response  curvesobtained  for  two  different  wave  steepness  are strongly recommended to be used.   Modelling of roll motion in beamsea and determination of model parameters










Naar boven