Académique Documents
Professionnel Documents
Culture Documents
5-02
-05-01
Procedures Page 1 of 17
Testing and Extrapolation Methods
Effective Date Revision
High Speed Marine Vehicles 2002 01
Resistance Test
CONTENTS
3.5.3 Speed
1 PURPOSE OF PROCEDURE
3.5.4 Measured quantities
2 PARAMETERS 3.6 Data Reduction and Analysis
3.6.1 Analysis of model scale
2.1 Data Reduction Equations results
2.2 Definition of Variables 3.6.2 Extrapolation to full scale
3 DESCRIPTION OF PROCEDURE 3.6.2.1 Form factor
3.6.2.2 Model-Ship correlation
3.1 Model and Installation 3.7 Documentation
3.1.1 Model 3.8 Special Considerations
3.1.2 Installation 3.8.1 Air Resistance
3.2 Measurement Systems 3.8.2 Appendage Effects
3.3 Instrumentation 3.8.3 Wetted Area Estimation
3.3.1 Resistance 3.8.4 Spray Resistance
3.3.2 Speed 3.8.5 Blockage
3.3.3 Sinkage and trim 3.8.6 Captive Resistance Tests
3.3.4 Temperature 3.8.6.1 Fully captured force
3.3.5 Air cushion pressure measurements and
3.3.6 Air cushion flow rate simulation
3.4 Calibrations 3.8.6.2 Partially captured force
3.4.1 General remarks measurements
3.4.2 Resistance dynamometer 3.8.6.3 Automatic attitude control
3.4.3 Sinkage and trim method
transducers
3.4.4 Air cushion pressure 4 VALIDATION
3.4.5 Air cushion flow rate 4.1 Uncertainty Analysis
3.4.6 Speed 4.2 Benchmark Tests
3.5 Test Procedure and Data
Acquisition 5 REFERENCES
3.5.1 Method
3.5.2 Range and interval
Updated by Approved
Specialist Committee: Procedures for Resis-
tance, Propulsion and Propeller Open Water 23rd ITTC 2002
Tests of 23rd ITTC 2002
Date Date
ITTC – Recommended 7.5-02
-05-01
Procedures Page 2 of 17
Testing and Extrapolation Methods
Effective Date Revision
High Speed Marine Vehicles 2002 01
Resistance Test
R App
1 PURPOSE OF PROCEDURE C App =
ρ
SV 2
2
The purpose of the procedure is to ensure
consistency of methodology and the acquisition V
of correct results for the resistance tests of Froude Number Fr =
high-speed marine vehicles (HSMV). High gL
Speed Marine Vehicles are for this purpose
defined to be vessels with a design speed corre- V
sponding to a Froude number above 0.45, Depth Froude Number Frh =
gh
and/or a speed above 3.7 ∇ 1/6 (m/s) and/or
where high trim angles are expected or for dy-
namically supported vessels.
VL
Reynolds number Re =
ν
2 PARAMETERS
ENVIRONMENTAL
CARRIAGE HULL MODEL CONDITIONS
COMPUTER
(ii) the speed of the towing carriage rela- 3.3.6 Air cushion flow rate
tive to the water should be measured by
a current meter far in front of the The air cushion flow rate should be detect-
model. In this case the current meter able to within 10% of the mean (design) air
wake and waves should be minimised. flow rate. The air cushion flow rate is often
detected through the use of a calibration dia-
The speed of the model should be measured to gram from the measured pressure and fan
within 0.1% of the maximum speed or to speed.
within 3 mm/sec, whichever is the larger.
3.4 Calibrations
3.3.3 Sinkage and Trim
3.4.1 General remarks
Sinkage fore and aft may be measured with
mechanical guides, potentiometers, encoders, All devices used for data acquisition should
LDVTs or with remote (laser or ultrasonic) be calibrated regularly. For calibration, the
distance meters; the running trim is then measured quantities should be either substi-
calculated from the measured running sinkage tuted by calibrated weights and pulses or
fore and aft. Alternatively, the running trim checked by already calibrated other measuring
may be measured directly using an angular devices. The range of the calibration should
measuring device. include at least the range of values to be meas-
ured in the experiment. Calibration diagrams,
The sinkage should be measured to within where the measured quantities (output values)
1.0 mm. If the trim is measured directly, rather are plotted versus the calibration units (input
than deduced from a measurement of sinkage units), may be useful to check the calibration
fore and aft, it should be measured to an accu- itself as well as the linearity of the instruments.
racy of 0.1 deg. Calibration should generally be in accordance
ITTC – Recommended 7.5-02
-05-01
Procedures Page 7 of 17
Testing and Extrapolation Methods
Effective Date Revision
High Speed Marine Vehicles 2002 01
Resistance Test
with ITTC Quality Manual Standard Procedure 3.4.5 Air cushion flow rate
7.6-01-01.
The calibrations of the resistance dyna- If the air flow rate in the experiment is go-
mometer and the sinkage and/or trim sensors ing to be found from measurement of cushion
should be checked immediately prior to the pressure and fan rotational speed, then calibra-
testing. The calibrations should preferably in- tion curves for the fan(s) must be determined as
clude as much of the measurement chain as part of the calibration. A calibrated flow rate
possible (e.g. amplifier, filter, A/D converter). meter is needed, or a venturi meter or orifice
If the check indicates that the required accura- type instrument must be constructed. The fan is
cies cannot be met, the calibration should be then run at different rotational speeds and the
renewed or the instrument replaced and the delivered pressure must be varied using a vari-
check repeated. Daily checking of a pulse able aperture or some other method. The deliv-
counter type speed measurement device is usu- ered flow rate is measured for each combina-
ally not required. Instead, the check on this tion of backpressure and fan rate of revolu-
device is covered by calibrations carried out at tions. Two-variable calibration curves may
regular intervals. then be constructed. The rotational speed sen-
sor on the fan should be calibrated, for instance
using a pulse counter with verified accuracy.
3.4.2 Resistance Dynamometer
3.5 Test Procedure and Data Acquisition In some cases it is necessary to modify the
LCG to correct for artificial trim effects from
3.5.1 Method resistance components that influence the trim
but do not follow the Froude scaling laws. Ex-
Before the test begins, zero readings of all amples of such resistance are air resistance,
instruments are taken. Zeros should be checked appendage resistance and viscous resistance,
between runs to ensure no drift has occurred. when the propulsion force is applied far from
The model is towed at speeds giving the same the centre of viscous resistance, such as for a
Froude numbers as for the full scale ship. vessel to be propelled by air propellers. When
the tow-point is not in the extension of the pro-
The model speed is selected and the model pulsor line of thrust, it is then also necessary to
accelerated to that speed. If the model has been modify the LCG for the trim effect of the total
held during initial acceleration, it should be model resistance.
released smoothly as soon as the selected speed
has been reached. It is recommended that the An alternative approach to correcting for ar-
data acquisition should begin not later than tificial trim effects would be to apply the tow-
after releasing the model or a steady speed has ing force in such a way that its lever also pro-
been reached. The mean values are derived duces the correct longitudinal trimming mo-
afterwards from the time series, selecting a ment.
time window with the criterion that, after the
mean measurement values have stabilised, a
period of at least five oscillations should be 3.5.2 Range and Interval
used for the average that is entered into the
result. Maximum and minimum values together The speed range should extend from at least
with mean and standard deviations should be 5% below the lowest speed at which reliable
stored for each run. This process is repeated at data is required to at least 5% above the highest
other selected speeds covering the required speed required. This range should be covered
range, avoiding continuous progression from by a suitable number of speeds. Care should be
one limit to the other. For example, runs at al- taken to ensure that there is sufficient number
ternate speeds from the lowest speed to the of speeds to define humps or hollows and other
highest followed by the highest speed to the rapidly changing features of the curve.
lowest filling in the gaps.
available. This has however not always ments should be recorded at the beginning and
shown to be necessary and can be tested end of each test sequence
with uncertainty analysis. For more infor-
mation see Uncertainty Analysis, Example
for Resistance test, provided in QM 7.5-02- 3.5.5 Data Reduction and Analysis
02-02.
The speed, resistance, sinkage, trim, pres-
• It is essential that the speed of the model sure and other continuously recorded quantities
through the water should be constant of the test should be the mean value derived
throughout that part of the test run during from an integration of the instantaneously
which resistance is measured, and for a sig- measured values over the same measuring in-
nificant distance before measuring begins. terval, with the zero measurement being sub-
Steadiness of carriage speed is an essential tracted from the averaged values.
element in achieving steady model speed,
but is not necessarily sufficient since the Running wetted surface must normally be
rate of change of the initial acceleration and derived manually from underwater or above-
the moment and manner of release of the water photographs, video recordings, paint
model may interact with the model- smear techniques or from visual observations
dynamometer system and cause it to oscil- during the test runs as described in Section
late. 3.8.3.
• During the measuring run, the carriage Total resistance and residual resistance co-
speed should normally not vary by more efficients, together with Froude Number, are
than 0.1% of the mean speed or 3 mm/s, calculated for each speed using the data reduc-
whichever is the larger. The cyclic charac- tion equations given in Section 2.1.
teristics of the carriage speed control sys-
tem should be such as not to synchronise
with the natural frequency of the model dy- 3.5.6 Analysis of model scale results
namometer system.
Resistance RTM measured in the resistance
tests is expressed in the non-dimensional form
3.5.4 Measured quantities
based on the Reynolds number of each compo- the wetted sidewall is used. Underwater pho-
nent. Form factors for cylindrical hulls, struts tography is recommended for estimating wet-
and control surfaces have been derived using ted surface area of the inner sidewalls. Aerody-
theoretical and experimental methods (Gran- namic resistance is best estimated from wind
ville, 1976) which may be used if no other tunnel tests. If that is not a possibility, aerody-
source is available. Correlation allowances for namic resistance can be approximated using a
SWATHs have been proposed over a wide drag coefficient of approximately 0.5 applied
range from 0.0000 to 0.0005. to the entire frontal area of the vehicle. Testing
with a superstructure covering the entire model
Hydrofoils – For hydrofoils, the hull resis- is recommended in order to model the impor-
tance should be analysed like the resistance of tant trim effect of the air lift and drag.
an ordinary HSMV without foils. The foil sys-
tem resistance should be computed for full Air Cushion Vehicles (ACV) – It is com-
scale Reynolds number, or expanded from tests mon practice to Froude scale all of the resis-
at a Reynolds number high enough to ensure tance measured on an ACV model except for
fully turbulent flow. In case the foil system was that of fully wetted appendages. Stevens and
present during the towing tests, the drag of the Prokhorov (Savitsky et al., 1981) defended this
model foil system must then be subtracted from approach with the premise that the unrealisti-
the total resistance to get the bare hull resis- cally high friction resistance of the model’s
tance. In this case it is strongly advised that the wetted skirt would be partially offset by lower
foil system resistance be measured during the spray resistance of the model. It is recom-
towing tests, as uncertainty regarding the ex- mended that fully wetted appendages should be
tent of turbulent flow on the foils in model treated the same as for other HSMVs.
scale will make it difficult to calculate the drag
in model scale. Due to lack of correlation data,
it is recommended that a correlation coefficient 3.5.7.1 Form factor
of zero be used. Alternatively, if the hull is of a The use of the 1978 powering performance
type for which correlation is available, the hull procedure implies the use of a form factor
resistance can be corrected with the applicable (1 + k ) . Particular problems arise with
correlation coefficient, while the foil system
drag should be added without a correction due estimates of (1 + k ) for HSMVs in that low
to correlation. speed tests are not normally reliable or
sufficient. Many HSMVs employ transom
Surface Effect Ships (SES) – For SES craft, sterns, leading to a confused flow aft of the
it is common practice to estimate resistance transom at low speeds and wetted surface area
components due to hull friction and aerody- generally changes with speed, resulting in a
namic forces and then deduce the residual re- change in true (1 + k ) with speed. For this
sistance, which includes the friction and in- reason it is currently recommended that, for
duced drag of the seals. Froude scaling of consistency and for the time being, form
speed is based on the cushion length. For cal- factors for HSMVs continue to be assumed as
culating friction resistance it is recommended (1 + k ) = 1.0 .
that a Reynolds number based on the length of
ITTC – Recommended 7.5-02
-05-01
Procedures Page 12 of 17
Testing and Extrapolation Methods
Effective Date Revision
High Speed Marine Vehicles 2002 01
Resistance Test
3.5.7.2 Model-Ship Correlation • For each speed, the following data should
The proposed extrapolation method re- be given as a minimum:
quires an established model-ship correlation, Resistance of the model
partly because the form factor is set to zero. It Sinkage fore and aft, or sinkage and
is not possible to give general guidance to what trim
this correlation factor should be, but is left in- Dynamic wetted surface area (if consid-
stead to each facility to establish its own corre- ered)
lation factor. The extrapolation method adopted Air cushion pressure (if applicable)
should be documented clearly in the test report. Air cushion flow rate (if applicable)
water surface. Corrections can then be made to difference in bow-up or bow-down moment
account for Reynolds number effects based on between the model and full-scale vehicle by
the wind speed measured under the towing car- assuming centres of aerodynamic pressure and
riage and resistance contributed by the nor- hydrodynamic pressure. These forces are then
mally-submerged portion of the hull. Since balanced against the towing force and the re-
ship superstructures are in general relatively sulting moment converted to an effective shift
bluff, the Reynolds number effects on drag are in longitudinal centre of gravity.
often moderate. This would confirm that the
most straightforward way of minimising the
errors introduced by air resistance is to fit the 3.7.2 Appendage Effects
model with a modelled superstructure during
the tests. It is important to make adequate corrections
Before making air resistance corrections it for appendage effects on HSMV model test
is important to measure the actual airspeed be- results. Two methods are commonly used to
neath the carriage, in the area the model will be account for appendage effects:
tested. These measurements can be made with- (i) Testing the bare hull and then sepa-
out the model in place if the model cross sec- rately accounting for the lift and drag of
tion is small compared with the cross section of individual components using analytical
the air space housing the tank. Air speed meas- methods
urements should be made over the speed range (ii) Testing the hull with and without ap-
of interest with the carriage configured as it pendages and expanding the values
will be when tests are conducted. The air speed based on the local Reynolds number of
measurements and physical features of the each component.
above-water portion of the model should be
well documented in the test report so that users Testing both with and without appendages
of the test data can make their own estimates of has the advantage of providing more
air effects if they wish. When estimates of air information for expanding the test data using
resistance are made by staff members at the test different methods. Trim moments caused by
facility, the method used, including details such appendage forces not correctly represented in
us frontal cross section area and drag coeffi- the experiment should be accounted for using
cient should be documented in the test report. equivalent shifts in centre of gravity location
Drag coefficients typically range from 0.3- 1.0. and displacement. If these corrections are made
Since HSMVs such as planing boats are ex- after the tests are completed, the results can be
tremely sensitive to trim, estimates of the ef- obtained by interpolating between results from
fects of aerodynamic forces on trim should be tests with different centre of gravity locations.
made and documented in the same manner as A method for setting up test programmes with
for air resistance. the intent of making corrections at a later time
was proposed by Hoyt & Dipper (1989).
The recommended method of accounting
for aerodynamic effects on trim, which are not HSMVs with lift-producing appendages
properly taken into account by a modelled su- have the added complication of Reynolds num-
perstructure on the model, is to calculate the ber effects on lift. One approach for addressing
ITTC – Recommended 7.5-02
-05-01
Procedures Page 14 of 17
Testing and Extrapolation Methods
Effective Date Revision
High Speed Marine Vehicles 2002 01
Resistance Test
side hulls is different from that of the main higher speeds, blockage effects can be esti-
hull. mated using three dimensional wave resistance
Based on the need for better accuracy and calculations for the situation with the model in
representation of the correct physics, it is a tank.
recommended that running WSA should be
used for HSMVs instead of static WSA. Any For SES and ACVs blockage effects might
one of the measurement methods listed may be calculated using simple numerical methods
give good results depending on the vehicle type like those summarised by Doctors (1992).
and test facility characteristics, but the method
of measurement and likely level of accuracy For displacement and semi-displacement
should be described and defined in test reports. ships, two-dimensional wave resistance calcu-
lations might be applied. Relatively simple
computer programs for blockage and shallow
3.7.4 Spray Resistance water corrections based on thin-ship theory by
Lunde (1961) have been found useful for this
At present there is no accepted method purpose.
available to account for scale effects in resis-
tance attributable to spray.
3.7.6 Captive Resistance Tests
ordinary towed models. The problem of the 3.7.6.3 Automatic Attitude Control Method
scale effect on running attitude can be avoided.
A more sophisticated experimental method
The effect of appendages can be obtained as a
has been developed on the basis of the same
result of the simulation by adding hydrody-
philosophies. The experimental apparatus is
namic forces acting on them into the equilib-
composed of a force measurement system, a
rium equation. This method can easily cope
system for solving the equilibrium equation of
with design changes, such as the location of the
the forces by a computer in real time and a sys-
centre of gravity, appendages and thrust force
tem for continuously changing the running atti-
direction. The disadvantage of the method is
tude of the model, for instance by stepping mo-
that the hydrodynamic force measurements are
tors. Forces acting on a model craft are meas-
time-consuming compared with a conventional
ured and its attitude changed using these values
resistance test. Also, investigations of porpois-
to satisfy the equilibrium of forces. Additional
ing and chine walking are precluded. It is noted
forces acting on appendages and any predict-
that any standing waves in the towing tank
able scale effects can be taken into account in
should be reduced as much as possible since
the calculation.
they affect the lift force directly. It is also more
important to have well-aligned rails and a
smooth running carriage for this method than 4 VALIDATION
for towing a model free to heave and trim.
Typical practical methods together with results 4.1 Uncertainty Analysis
are described by Ikeda (1992, 1993), Yoko-
mizo (1992), and Katayama & Ikeda (1993,
1995, 1996) for planing craft and by Minsaas Uncertainty analysis should be performed
(1993) for fully submerged hydrofoils. in accordance with ‘Uncertainty Analysis in
EFD, Uncertainty Assessment Methodology’
3.7.6.2 Partially Captured Force Measurements as described in QM 4.9-03-01-01 and ‘Uncer-
tainty Analysis in EFD, Guidelines for Uncer-
To avoid the effect of water surface tainty Assessment’ as described in QM 4.9-03-
fluctuation on lift force, hydrodynamic force 01-02. In addition to the above an example
measurements in the free-to-heave condition ‘Uncertainty Analysis, Example for Resistance
have been developed. Using the measured drag Test’ is provided in QM 7.5-02-02-02.
and moment, an equilibrium equation of two
forces is solved to provide the running attitude
and resistance of a fast craft. A variation of this
method, which is used in cases where the trim 4.2 Benchmark Tests
of the ship is going to be controlled, for
instance by a forward lifting foil system, the Benchmark data are collected and described
model is fixed in the required trim without the in ‘Benchmark Database for CFD, Validation
need for any iterations. The required control for Resistance and Propulsion’, QM 7.5-03-02-
system force is then easily determined from the 02.
trim moment measurement. See also the following reference: Summary and
Conclusions of Co-operative Model Resistance
ITTC – Recommended 7.5-02
-05-01
Procedures Page 17 of 17
Testing and Extrapolation Methods
Effective Date Revision
High Speed Marine Vehicles 2002 01
Resistance Test