1
The Specialist Committee on
Model Test of High Speed Marine Vehicles
Final Report and
nd
Recommendations to the 22 ITTC
GENERAL
1.1 Membership and Meetings
The committee appointed by the 21st ITTC
consisted of the following members:
• Dr. Daniele Ranocchia (Chairman): IN-
SEAN - Italian Towing Tank, Italy.
• Dr. Anthony F. Molland (Secretary): Uni-
versity of Southampton, United Kingdom
• Mrs. Susanne Abrahamsson: SSPA Mari-
time Consulting AB, Sweden
• Prof. Yoshiho Ikeda: Osaka Prefecture
University, Japan
997 (8)
Sydney, Australia, July 1997 (7)
Wageningen, The Netherlands, April 1998 (7)
Annapolis, USA, November 1998 (5).
The numbers in the parentheses indicate
members present.
Contact has been made with the Resis-
tance, Propulsion, Manoeuvring, Loads and
Responses General Committees, and with the
following Specialist Committees: Waterjets,
Stability, Environmental Modelling and
Safety of HSMVs, in order to avoid overlap of
content. For the same reason a liaison has
• Mr. Geert K. Kapsenberg: MARIN -
Maritime Research Institute Netherlands,
The Netherlands
• Dr. Myung-Soo Shin: KRISO - Korea Re-
search Institute of Ships & Ocean Engi-
neering, Korea
• Dr. Sverre Steen: MARINTEK - Norwe-
gian Marine Technology Research Insti-
tute A/S, Norway
• Mr. John J. Zseleczky: U. S. Naval Acad-
emy, United States.
The meetings of the Committee were held
as follows:
Rome, Italy, January 1
been established with the ISSC Loads Com-
mittee.
st
1.2 Recommendations of the 21 ITTC
The Recommendations for the future work
of the Committee made by the 21st ITTC were
as follows:
1. Review the status of hydrodynamic tech-
nology related to model tests of high
speed marine vehicles summarised in the
th
Proceedings of the 16 ITTC (1981) and
recommend codes of practice for carrying
out model tests for high speed marine ve-
hicles.
2
2. Review experimental methods to evaluate
the seakeeping performance of multi-hull
forms and HSMVs including active moti-
on control systems and prepare guidelines.
2 INTRODUCTION
As stated in the Proceedings of the 21st
ITTC (Vol. 1, Appendix 2, Section 4.2), the
complete findings of the 16th ITTC HSMV
Committee were documented in a report titled
“The Status of Hydrodynamic Technology as
Related to Model Tests of High Speed Marine
Vehicles” (Savitsky et al., 1981), which was
reviewed by this Specialist Committee. Since
the report is seventeen years old, the Special-
• Extrapolation method / model-ship corre-
lation.
• Types of facilities.
• Typical model sizes.
• Future research.
To address these issues it was necessary to
obtain input from the ITTC members and to
obtain their opinions concerning the major
problem areas associated with model tests on
HSMVs. Therefore a questionnaire was dis-
tributed which addressed the above issues.
The questionnaire was sent in May 1997 to 84
ITTC members.
200

ist Committee circulated a questionnaire to 180

the ITTC member organisations requesting Number of HSMV Tested

160

information on more current practices for 140

HSMV testing. The codes of practice listed in 120

this report are based on both the findings of 100

the 16th ITTC HSMV Committee and the re- 80

sults of the more recent questionnaire. An ef- 60

fort has been made to identify the areas where 40

there is no clear common practice for model 20

testing HSMVs or where the current practice 0

4. Trimaran

6. Surface Piercing

7. Fully Submerged
5. SWATH

10. WIG
1. Planing Hulls

Displacement

3. Catamarans

11. Hybrids
9. ACV
8. SES
is especially questionable. The codes of prac-
2. Semi-

Hydrofoils.

Hydrofoils
tice recommended here can be debated in the
future and amended as necessary.

3 QUESTIONNAIRE ON MODEL
TESTS OF HIGH SPEED MARINE
VEHICLES
One of the tasks of the Committee was to
review recommended procedures for the ex-
ecution of model tests. In order to do this the
Committee decided, at an early stage, that
there was a need for a survey within the ITTC
members who have significant involvement in
HSMV activities. In particular the Committee
felt that there was a need to define the areas
where the limited resources could be best
utilised.
Important areas to be considered include:
• Types of HSMVs tested.
• Types of tests carried out on HSMVs.
Fig. 3.1 - Number of HSMV tested in the period 1992-
1997, divided by category, summed for all answers.
Steen (1998) has made a detailed analysis
of the questionnaire on the basis of 34 replies.
As was seen from the analysis, many ques-
tions required a reply in terms of comments
and free text. In addition, most of the people
filling out the questionnaire do not have Eng-
lish as their first language. Use of different
terminology and difficulties understanding the
questions have lead to a quite low degree of
accuracy in the responses and thus in the pre-
sent analysis. This is anticipated to be a quite
common problem with questionnaires, but an
increased use of questions that require yes/no
selection between specific alternatives or a
 3
number as the reply would be expected to in- These results are summarised in Figures
crease the accuracy. 3.1, 3.2, and 3.3.
Number of Facilities Carrying out Tests on Different Kinds of HSMV

35 Types of Tests for Different Kinds of HSMV

Number of Facilties Carrying out Test Type

30 35

30
25
25
20
20

15 15

10
10
5

5 0

6. Surface Piercing Hydrofoils.

4. Trimaran

7. Fully Submerged Hydrofoils

5. SWATH

10. WIG
2. Semi-Displacement

8. SES

9. ACV
1. Planing Hulls

3. Catamarans

11. Hybrids
0
4. Trimaran

6. Surface Piercing

7. Fully Submerged
5. SWATH

10. WIG
1. Planing Hulls

Displacement

3. Catamarans

11. Hybrids
9. ACV
8. SES
2. Semi-

Hydrofoils.

Hydrofoils

2.1 Resistance 2.2 Propulsion 2.3 Seakeeping

2.4 Manouevring 2.5 Dynamic Stability 2.6 Structural Loads
Fig. 3.2 - Number of facilities carrying out tests on dif- 2.7 Others
ferent kinds of HSMV.

Fig. 3.3 - The number of facilities carrying out differ-

Analysis of the Questionnaires indicated
ent test types on different kinds of HSMV.
the following findings:
• Catamarans are the most frequently tested 4 SURVEY OF EXISTING TEST
vessels. Almost the same number of TECHNIQUES AND PROCEDU-
planing and semi-displacement monohulls RES
were tested. The other categories consid-
ered are much less frequently tested. 4.1 General
• Resistance tests are carried out for almost The review of HSMV model testing tech-
all vessel types that are tested. Seakeeping nology conducted by the 16th ITTC was or-
tests are a good second, while propulsion ganised in order of HSMV type: SWATH,
tests are less than half as frequent as re- semi-displacement, planing, hydrofoil, SES
sistance tests. and ACV. For each type of craft a summary
• The significant problem areas for the was given for different areas of technology
ITTC community are: such as resistance, seakeeping, manoeuvring,
• estimation of wetted surface; performance and propulsor. Upon reviewing
• change of running attitude between re- this work, the Specialist Committee conclud-
sistance and propulsion tests; ed that the recommended codes of practice
• scaling of wake; would be less repetitive if they were arranged
• scaling of appendage resistance; in order of test type rather than HSMV type.
• long time series required for tests in The following main test types were identified:
waves;
• Resistance
• towing method for seakeeping;
• Seakeeping
• scarcity and low quality of full scale
measurements for correlation purposes. • Propulsion
• Manoeuvring
• Structural Loads
4
• Dynamic Instability.
The test types are listed in order of how
often they are carried out according to the
questionnaire survey. Resistance tests are the
most common by a large margin and are ad-
dressed first. Many of the issues that are im-
portant for resistance tests of HSMVs are also
important for other types of testing. To avoid
repetition, these issues are listed only once,
under the test heading in which they first ap-
pear.
4.2 Resistance Tests
The key issues identified by the Commit-
tee for conducting HSMV resistance tests are
included in this section. Recommended codes
of practice are outlined after each issue.
4.2.1 Air Resistance
This is one of the most important areas to
address for the testing of HSMVs. Air resis-
tance is typically not a concern for lower
speed vehicles but can have significant effects
at high speeds. From the current 1996-99
Questionnaire, it was found that facilities ac-
count for air resistance in different ways, such
as testing the model with and without a wind
screen and accounting for the difference, or
building a superstructure on the model and
measuring the combined hydrodynamic and
aerodynamic resistance, with or without cor-
rections for scale.
Given the difference in physical charac-
teristics of each facility it is impossible to
propose a single testing method that will
provide identical results in each facility. Fac-
tors such as the size of the carriage and per-
meability of its structure are difficult to quan-
tify but can significantly affect the flow of air
above the model as the carriage travels down
the tank.
The speed at which air resistance becomes
significant varies with the facility, towing ar-
rangement and vehicle type. If it is decided
that air resistance is insignificant for a par-
ticular HSMV model test, the justification for
that decision should be documented in the test
report.
When air resistance is considered to be
significant, wind tunnel tests provide the best
source of information since the model can be
tested at the correct Reynolds number. How-
ever, for most test programmes, the expense
of wind tunnel tests can be cost prohibitive. A
practical alternative is to tow the model, fitted
with a superstructure, above the water surface.
Corrections can then be made to account for
Reynolds number effects based on the wind
speed measured under the towing carriage and
resistance contributed by the normally-
submerged portion of the hull.
Before making air resistance corrections it
is important to measure the actual airspeed
beneath the carriage, in the area the model
will be tested. These measurements can be
made without the model in place if the model
cross section is small compared with the cross
section of the air space housing the tank. Air
speed measurements should be made over the
speed range of interest with the carriage con-
figured as it will be when tests are conducted.
The air speed measurements and physical
features of the above-water portion of the
model should be well documented in the test
report so that users of the test data can make
their own estimates of air effects if they wish.
When estimates of air resistance are made by
staff members at the test facility, the method
used, including details such us frontal cross
section area and drag coefficient should be
documented in the report. Müller-Graf cited
drag coefficients ranging from 0.3 to 1.0 for
different types of craft (Savitsky et al., 1981).
Since HSMVs such as planing boats are ex-
tremely sensitive to trim, estimates of the ef-
fects of aerodynamic forces on trim should be
made and documented in the same manner as
for air resistance. The recommended method
of accounting for aerodynamic effects on trim
is to calculate the difference in bow-up or
bow-down moment between the model and

full-scale vehicle by assuming a centre of
aerodynamic pressure and hydrodynamic
pressure. These forces can then be balanced
against the towing force and the resulting
moment can be converted to an effective shift
in longitudinal centre of gravity.
4.2.2 Appendage Effects
Appendage drag represents a larger per-
centage of total resistance for most HSMVs
than for conventional displacement hulls. In
general, the resistance of an HSMV hull is al-
so more sensitive to running trim angle, espe-
cially in the case of planing hulls. It is there-
fore extremely important to make adequate
corrections for appendage effects on HSMV
model test results.
Two methods are commonly used to ac-
count for appendage effects:
1. Testing the bare hull and then accounting
for lift and drag of individual components
using analytical methods (Hadler, 1966 and
Kirkman & Kloetzli, 1981);
2. Testing the hull with and without ap-
pendages and expanding values based on the
local Reynolds number of each component.
The 1996-99 Questionnaire on HSMV
testing indicated that both methods are widely
used. Each method has strengths and weak-
nesses as noted in different sections of the
16th ITTC HSMV Report (Savitsky et al.,
1981). For semi-displacement round bilge
hulls Müller-Graf (p. 43) stated that “the in-
fluence of the appendage forces on running
trim and resistance cannot be estimated theo-
retically. Therefore the results of the bare hull
tests are of limited value”. In the next section
on planing hulls, Savitsky (p. 65) stated that
“the usual Reynolds number for model ap-
pendages is so small that turbulence stimula-
tion is not likely; hence, appendages are not
included on the model and their full-scale
drag is best calculated (Blount & Fox, 1976)”.
At present, the recommendation of the
Committee is that either method can be used
5
provided adequate documentation is included
in the test report so users of the data can re-
expand the data using alternate methods if de-
sired. Testing both with and without ap-
pendages has the advantage of providing more
information for expanding test data using dif-
ferent methods. If the purely analytical meth-
od is used, trim moments caused by ap-
pendage forces should be accounted for using
equivalent shifts in centre of gravity location
and displacement. If these corrections are
made after the tests are completed, the results
can be obtained by interpolating between re-
sults from tests with different displacements
and centre of gravity locations. A method for
setting up test programmes with the intent of
making corrections at a later time was pro-
posed by Hoyt & Dipper (1989).
In addition to corrections for appendage
effects, the forces induced by propellers on
the appendages and hull should be addressed,
as these forces can be considerable on
HSMVs. Ideally, because of the complexity of
propeller effects, propulsion tests should be
conducted in addition to standard resistance
tests. In practice however, the high speeds of
HSMVs lead to smaller models which are dif-
ficult to outfit with self-propulsion gear and
use small propellers with low Reynolds num-
bers over the propeller blades. Therefore it is
more common to conduct test programmes
without propulsion tests and rely on correc-
tions to standard resistance tests to account
for propeller forces, as proposed by Hadler
(1966).
HSMVs with lift-producing appendages
have the added complication of Reynolds
number effects on lift. One approach for ad-
dressing scale effects of lift-producing ap-
pendages is to modify the section shape of the
model appendage so that the lift characteris-
tics of the model appendage better represent
those of the full-scale vehicle.
Particular measures and procedures are
proposed by the Committee in the case of hy-
drofoil vessels. The procedures described ap-
6
ply to vessels with fully submerged hydrofoils
that are designed to carry a significant part of
the weight of the vessel. The procedures are
mainly based on the experiences gained at
MARINTEK in the years 1990 to 1995 (Min-
saas, 1993).
The main difference from other high speed
vessels is that scale effect on lift must be
taken into account. It is seen that for Reynolds
number smaller than 1× 10 6 (Hoerner &
Borst, 1985), unpredictable scale effects
might occur, caused by laminar separation on
the pressure side of the foil.
In Figure 4.2.1 the ratio between theoreti-
cal and experimentally measured lift is pre-
sented as function of Reynolds number. The
theoretical calculation is for non-viscous flow.
The comparison is made at ideal angle of at-
tack. The results were obtained by an exten-
sive series of wind tunnel tests carried out for
BSRA (1958). It is seen from Figure 4.2.1 that
the scale effect on lift is considerable, espe-
cially for Reynolds numbers below 1× 10 6 .
Fig. 4.2.1 - Dependence of Reynolds number of the ra-
tio between measured and computed lift µ (BSRA,
1958).
There are two principally different ways of
doing resistance and propulsion tests with hy-
drofoil vessels:
• Testing the hull and foils together.
• Testing only the hull, and computing (or
separately testing) the lift and drag of the
foil system.
When testing the hull and foils together, one
must consider the scale effect on both lift and
drag of the foil system. If the local Reynolds
number of the main lifting surfaces are below
about 1× 10 6 , scale effect on lift should be
considered. For fully foil-borne vessels, the
resistance in take-off condition is often criti-
cal. Take-off conditions occur at lower speeds
than cruising. In addition, the foils often oper-
ate at near maximum lift in this condition.
The maximum lift is more sensitive to
Reynolds number effects than the linear part
of the lift curve. It is very important that the
results concerning the take-off condition are
correct. If the lift to drag ratio in take-off is
over-estimated from model tests, this might
cause the full scale vessel never to be able to
take off, and it might thus become a complete
failure. If it is decided to test the model with
foils free in heave and pitch, the longitudinal
stability during acceleration of the model
might be a problem. If testing the model free
in heave and pitch, any scale effect on lift
should be corrected for by removing or adding
ballast to the model. By scaling the drag of
the foil system, local Reynolds number, form
drag and viscous interaction drag on each foil
component should be treated individually. By
measuring the drag on the foil system in mod-
el scale, the uncertainty of the conversion
process can be reduced.
Having noted the many difficulties related
to testing models with foil system included,
the recommended procedure is to test the
model without foil system, and assess the lift
and drag of the foil system separately. The
model must then be tested fixed to the car-
riage, with measurement of pitch moment,
longitudinal and vertical forces. The model
must be tested over a range of draughts and
trims. By combining the separately obtained
lift and drag of the foil system with the meas-

ured forces on the model (scaled to full scale
applying the ordinary scaling principles), the
running attitude and resulting resistance can
be found as a function of speed by interpola-
tion of the model test results.
The foil system drag can be decomposed
in the following components:
1. Viscous drag (including form drag). Com-
puted separately for each component of
the system.
2. Viscous interference (junction) drag.
Computed separately for each component
of the system.
3. Spray drag. Computed separately for each
surface piercing component of the system.
4. Self-induced drag. Computed separately
for each lifting surface of the system.
5. Induced drag. In practice computed as a
drag induced by downwash from front
foils on aft foils.
6. Wave resistance.
The foil system lift is characterised by the
following effects:
1. Lift of the foil itself, taking span and
spanwise distribution of lift into account.
2. Effect of downwash from upstream foils
on lift. (Correction to effective angle of
attack).
3. Effect of free surface on lift (important
only for foils close to the surface)
Except for the free surface effects, each of
these components can be computed with rea-
sonable accuracy using simple formulae. See
for instance Hoerner & Borst (1985) and Ho-
erner (1965). Free surface effects (except
spray drag) can be computed numerically in
an efficient manner by potential theory (see
for instance Mørch & Minsaas (1991), Mørch
(1992), and Landrini et al. (1999)). An alter-
native that is already possible, and might be-
come more attractive in the future is CFD
computations of the entire foil system.
An alternative to computing the foil lift
and drag is model testing in a high-Reynolds
number cavitation tunnel. The tunnel should
7
preferably have a free water surface. If the
tunnel does not have a free surface, correc-
tions for the free surface effects must be made.
If the tunnel has a free surface, one should
look out for finite depth effects. In any case,
the possibility of significant blockage effects
must be investigated. Calculation and correc-
tion for blockage effects when testing hydro-
foils in a free surface cavitation tunnel has
been extensively covered by Koushan (1997).
4.2.3 Turbulence Stimulation
Boundary layer turbulence stimulation is
recommended when the Reynolds number is
less than 5 × 10 6 based on hull length. For
HSMVs such as planing boats, where wetted
length is variable, Reynolds number should be
based on mean or effective wetted length. If a
test programme is focused on the high-speed
portion of the operation profile, it may be ac-
ceptable to eliminate tests at speeds below
that which produces the recommended
5 × 10 6 Reynolds number. For example,
given a model with a wetted length of 2 me-
ters, tests at speeds below 2.6 m/s would not
be conducted unless turbulence stimulation
was used. For tests where the wetted length is
relatively constant, as for SWATHs and SESs,
and where Reynolds numbers below
5 × 10 6 are unavoidable, a turbulent boundary
layer can be stimulated with studs, trip wires
or sand strips. Each approach has advantages
and disadvantages. Preliminary tests are re-
quired to find the best location for each of
these stimulation devices and to determine
their contribution to the total resistance
(Savitsky et al., 1981, Section by van Oos-
sanen). Preliminary tests are not required if
similar studies can be referenced for hulls
with similar characteristics.
For many HSMVs the forefoot enters the
water at a location that varies with speed.
Therefore, when turbulence stimulation is re-
quired for these vehicles it is recommended
that a method is used that does not involve
8
devices at a fixed location on the hull. A strut
piercing the water surface ahead of the model
as described by Savitsky & Ross (1952) and
reprinted by Savitsky et al. (1981) is recom-
mended. Another acceptable method is to in-
stall a tensioned wire between struts in a hori-
zontal, transverse orientation ahead of the
model. Whichever method is used, compari-
sons tests should be referenced to show users
of the data the effect of the size and location
of the turbulence stimulation system used.
The size of most HSMV appendages is
usually too small to obtain a Reynolds number
of 5 × 10 6 , and the ideal solution is to use a
very large model. This is generally not possi-
ble however, since the size of the model is
often limited by economic considerations or
facility capabilities. Specific studies such as
those conducted by Lackenby (1955), Clem-
ent (1957) and Lewandowski (1989) show ef-
fects of Reynolds number on specific ap-
pendages but the Committee is not prepared
to offer a recommendation of one particular
method of turbulence stimulation at this time.
4.2.4 Wetted Area Estimation
The wetted surface area (WSA) of most
HSMVs varies with speed. It is generally ac-
cepted that accurate determination of the run-
ning WSA of the model is essential. WSA
estimates affect the calculation of frictional
resistance, form factor (if used) and extrapo-
lation procedures. Replies to the current
1996- 99 Questionnaire indicate ongoing
problems of accurate measurement/prediction
of WSA.
Methods of determining WSA are well de-
scribed in the proceedings of the 17th ITTC
(1984) and include: visual observations from
outside model, visual observations from in-
side model, above water photography, under-
water photography, insoluble paint techniques,
water soluble paint techniques and electrical
wetting probes. The results of the current
1996-99 Questionnaire indicate the continuing
use of different methods for deriving WSA,
including photography, video and direct
measurement.
Surface tension may have an effect on
WSA, as discussed in some detail in the pro-
ceedings of the 18th ITTC (1987). Surface
tension leads to a different form of spray be-
tween model and full scale, the model spray
appearing like a sheet of water rather than
droplets as at full scale. For this reason, sepa-
ration of the spray sheet at model scale is de-
layed and the WSA tends to become relatively
larger with decreasing model size and model
speed. Minimisation of scale effects due to
surface tension can be achieved with the use
of larger models, higher speeds, and the fitting
of model spray rails which correctly simulate
full scale rails and which can lead to the cor-
rect determination of WSA.
When making estimates of WSA and
wetted length, a distinction is made between
the area covered by spray and that covered by
solid water. It is common practice to disregard
the viscous drag of spray-covered areas and to
account for only the viscous drag of the area
wetted by solid flow. This practice is ques-
tionable but the flow in the spray region is
extremely complex and no alternative prac-
tices are known.
Whilst the running WSA for
semi-displacement hulls may increase by
5-10%, a number of organisations use still
water WSA, claiming a small loss in accuracy.
Round bilge standard series data, for example,
tend to be based on still water WSA. Con-
flicting opinion on the use of static or running
WSA, and methods of measuring WSA, are
expressed in the proceedings of the 19th ITTC
(1990), and the problems of uncertainty in the
accurate measurement of spray wetted area on
the inboard sides of multihulls is discussed in
the proceedings of the 20th ITTC (1993).
For SWATHs it is standard practice to
measure wetted area separately for the hulls
and struts. The appropriate Reynolds number
is later used to analyse the viscous resistance

of each component separately. This procedure
is also used for trimarans, where the length of
the side hulls is different from that of the
main hull.
Based on the need for better accuracy, and
representation of the correct physics, the
Committee would recommend that running
WSA should be used for HSMVs instead of
static WSA. Any one of the measurement
methods listed may give good results de-
pending on the vehicle type and test facility
characteristics, but the method of measure-
ment and likely level of accuracy should be
described and defined in test reports.
4.2.5 Spray Resistance
At present there is no accepted method
available to account for scale effects in resis-
tance attributable to spray. This is an area that
needs to be addressed in the future but is dif-
ficult to treat because of the wide variation in
spray-generating features of HSMVs. The
treatment of spray resistance is complicated
by changes in spray direction and extent.
4.2.6 Blockage
For all types of HSMV it is important to
avoid the situation in which the hump speed
of the model coincides with the critical speed
of the tank ( gh / v 2 , where h is the tank
depth and v is the model speed). Blockage
was addressed for different types of HSMVs
by Savitsky et. al. (1981) and are summarised
here. For planing hulls, Savitsky stated that
wall effects are believed to be minimal if the
tank width is at least seven times the model
beam. For semi-displacement hulls and hydro-
foils, Müller-Graf stated that tank depth
should be greater than 0.8 times the model
length and the tank width should be greater
than two times the model length.
For SWATHs, van Oossanen stated that
blockage corrections for conventional ships
can be used at Froude numbers below 0.35. At
higher speeds, if the test budget allows,
blockage effects can be estimated using three
9
dimensional wave resistance calculations (fi-
nite element methods) for the situation with
the model in a tank.
For ACVs, Prokhorov stated that the tank
width should be a minimum of four to six
times the model beam and tank depth should
be at least half the model length.
4.2.7 Extrapolation, Form Factor, Correla-
tion
4.2.7.1 Extrapolation
The 1978 ITTC Powering Performance
Method cannot normally be applied directly to
HSMVs due to the often specific nature of
their behaviour. However, as concluded by
the 16th ITTC (1981), the philosophy of the
method, where predictions are based on sound
analysis of the scale effect of the different
components of resistance or propulsion coef-
ficients, can (and should) be applied. In the
case of high speed craft, particular physical
aspects of the problems must be considered,
which include the following: wake scaling,
validity of the form factor, change of wetted
surface area, change of trim, appendage re-
sistance, spray resistance, cavitation effects on
propeller and effect of shaft inclination.
Comprehensive reviews of and proposals
for power prediction methods for HSMVs are
made in the proceedings of the 19th ITTC
(1990). These methods should be applicable
to semi-displacement and planing types, and
are summarised as three procedures.
A. A modified ITTC 1978/88 procedure.
B. A method based only on self-propul-
sion and open water tests.
C. A method based on a special towing
test for models of small high speed
craft (for which self-propulsion tests
are not feasible) with compensating
forces and moments due to the pro-
peller applied to the resistance model.
Propulsive coefficients may be based
on data available from other sources.
10
The findings of the 19th ITTC (1990) were
not conclusive, and recommended that the
power prediction procedures proposed should
be evaluated and further developed. The re-
sults of a questionnaire on powering perfor-
mance of SWATH and catamarans carried out
by the Powering Performance Committee of
the 20th ITTC (1993) indicated wide varia-
tions in the extrapolation methods employed,
although some organisations were using the
modified ITTC 78 method in parallel with
their own methods. The results of the current
1996-99 Questionnaire are similarly incon-
sistent. The evidence available indicates that
it would be improper at this stage to recom-
mend a particular extrapolation technique for
HSMVs. There is however a distinct move-
ment towards the use of the modified ITTC 78
approach (A), and it is recommended that this
approach be pursued where possible. For tests
with model size constrained by tank size or
required speed, and when method (C) is used,
sources of propeller force estimates and their
application to the resistance tests should be
clearly stated.
The following specific considerations can
be made for SWATH, Hydrofoils, SES, and
ACV.
SWATH – Separate friction coefficients
are determined for the struts and submerged
hulls based on the Reynolds number of each
component. Form factors for cylindrical hulls,
struts and control surfaces have been derived
using theoretical and experimental methods
(Granville, 1976) which may be used if no
other source is available. Correlation allow-
ances for SWATHs have been proposed over
a wide range of 0.0000 to 0.0005.
Hydrofoils – The correlation method re-
commended for hydrofoils is considerably dif-
ferent than for other HSMVs because it has
been found that the residual resistance coeffi-
cient for the model is not equal to that of the
ship. The following method, proposed by the
16th ITTC HSMV Committee (Savitsky et al.,
1981), is recommended for expanding model
scale hydrofoil data to full scale.
RTS = ε M ⋅ Disp S ⋅ (1 + k )
where: RTS - total resistance of ship
ε M - drag-to-lift ratio of model
Disp S - displacement of ship
k - correction factor based on full
scale trials of similar vehicles; typically be-
tween 0.1 – 0.2.
Surface Effect Ships (SES) – For SES
craft, it is common practice to estimate resis-
tance components due to wavemaking, fric-
tion and aerodynamic forces and then deduce
the residual resistance, which includes the
friction and induced drag of the seals. Froude
scaling of speed is based on the cushion
length. For calculating friction resistance it is
recommended that a Reynolds number based
on the length of the wetted sidewall is used.
Underwater photography is recommended for
estimating wetted surface area of the inner
sidewalls. Aerodynamic resistance is best es-
timated from wind tunnel tests. If that is not a
possibility, aerodynamic resistance can be ap-
proximated using a drag coefficient of ap-
proximately 0.5 applied to the entire frontal
area of the vehicle.
Air Cushion Vehicles (ACV) – It is com-
mon practice to Froude scale all of the resis-
tance measured on an ACV model except for
that of fully wetted appendages. Stevens and
Prokhorov (Savitsky et al., 1981) defended
this approach with the premise that the unre-
alistically high friction resistance of the
model’s wetted skirt would be partially offset
by lower spray resistance of the model. It is
recommended that fully wetted appendages
should be treated the same as for other
HSMVs.
4.2.7.2 Form Factors
The use of the 1978 powering performan-
ce procedure implies the use of a form factor
(1 + k ) . Particular problems arise with esti-

mates of (1 + k ) for HSMVs in that low speed
tests are not normally reliable or sufficient.
Many HSMVs employ transom sterns, leading
to a confused flow aft of the transom at low
speeds and wetted surface area generally
changes with speed, resulting in a change in
true (1 + k ) with speed. Possible alternative
methods for obtaining (1 + k ) , suitable for
HSMVs, are summarised as follows:
• Measurement of total viscous drag by wake
survey, yielding CV = (1 + k ) ⋅ C F , hence
(1 + k ) over a range of speeds. Measure-
ment of total viscous drag might include
surface debris due to wave breaking and
spray, hence overestimating (1 + k ) . Also,
the method is unlikely to be developable to
routine commercial applications.
• Measurement of wave pattern drag CWP ,
hence (1 + k ) = (CT − CWP ) / C F over a
range of speeds. This method can generally
be applied on a routine commercial basis,
but tends to neglect influence of wave
breaking and hence overestimates (1 + k ) .
• Bow down tests (at low speeds) have been
advocated. Assuming the (usually small)
change in actual hull form is acceptable,
the techniques can provide a realistic esti-
mate of (1 + k ) . This is only for low speeds,
and this same value would have to be ap-
plied through the whole speed range.
Couser et al. (1997) and Cassella et al.
(1998) address the problems of form factors
applied to high speed catamarans. Much of
the content and philosophy of these papers
can be equally applied to monohulls. Both
references review the situation and incorpo-
rate other important work such as that report-
ed by Tanaka (1991) and Cordier & Durmez
(1993), but conclude that more work is re-
quired on the subject.
Results of the current 1996-99 Question-
naire indicate that few organisations incorpo-
11
rate a form factor in their extrapolation proc-
ess. It is considered that more work is re-
quired before reliable form factors could be
applied to HSMVs in a routine commercial
context. For this reason it is recommended
that, for consistency and for the time being,
form factors for HSMVs continue to be as-
sumed as (1 + k ) = 1.0 .
4.2.7.3 Model-Ship Correlation
An effective model-ship correlation exer-
cise requires the use of reliable trials data.
Due often to the constraints of commercial
confidentiality, there is a scarcity of published
trials data. Replies to the current 1996-99
Questionnaire tend to reflect this situation
since these suggest that only a small propor-
tion of organisations use trial data and
model-ship correlation as part of their routine
power prediction procedure for HSMVs.
Acceptable levels of correlation should be
achievable for vessels such as
semi-displacement, planing and SWATHs
whereas, as mentioned in the proceedings of
the 16th ITTC (1981), due to physical aerody-
namic and hydrodynamic complexities, cor-
relation for hydrofoils, ACVs and SES may
not be as straightforward.
The proceedings of the 18th ITTC (1987)
provides a discussion of model-ship correla-
tion and review of available trials data for
SWATHs, planing, hydrofoils, SES and
ACVs. Suggested data requirements for full
scale trials are given, together with a
model-ship correlation example to illustrate
data interpretation. Correlation is discussed in
the proceedings of the 19th ITTC (1990) for a
semi-displacement high-speed passenger ves-
sel for which trial results were available.
Guidelines for trials procedures for
HSMVs are provided in the proceedings of
the 19th ITTC (1990) and it is considered that
these remain applicable and suitable.
The Committee considers that a more
formal and unified approach to model-ship
12
correlation for HSMVs should be established.
To facilitate this aim it would recommend the
establishment by ITTC of a database of trials
data for HSMVs.
4.2.8 Other Resistance Tests
4.2.8.1 Fully Captured Force Meas-
urements and Simulation
Recently, in some experimental facilities,
a new prediction method for the resistance of
planing craft, hydrofoils and hybrid ships has
been carried out. The method is made up of
force measurements of a fully or partly cap-
tured model and a computer simulation using
the database of the measured hydrodynamic
forces. In this method, any additional forces
acting on appendages and scale effects can be
taken into account. Hydrodynamic forces
(drag, lift and trim moment) acting on a fully
captured model, are measured by systemati-
cally changing trim, rise and speed. By solv-
ing the equilibrium equation of forces using
the measured data, running attitudes and re-
sistance of a craft can be obtained. Multi-
component load cells are used to measure the
hydrodynamic forces. The experimental sys-
tem is very simple compared with a conven-
tional resistance test. Each measured hydro-
dynamic force is divided into frictional and
residual components on the basis of Froude
hypothesis, and can be scaled up to the full-
scale ship. It should be noted that wetted area
for each running attitude at corresponding ad-
vanced speed should be used to predict the
frictional component. The problem of the
scale effect on running attitude can be avoided.
The effect of appendages can be obtained as a
result of the simulation by adding hydrody-
namic forces acting on them into the equation.
This method can easily cope with design
changes, like location of centre of gravity, ap-
pendages, thrust force direction and so on.
The disadvantage of the method is that the
hydrodynamic force measurement is time-
consuming work compared to a conventional
resistance test. It is also noted that standing
waves in the towing tank should be reduced as
much as possible because it affects the lift
force directly. A practical method and the re-
sults are described by Ikeda (1992, 1993),
Yokomizo (1992), and Katayama & Ikeda
(1993, 1995, 1996) for planing craft and by
Minsaas (1993) for fully submerged hydro-
foils.
4.2.8.2 Partially Captured Force
Measurements
To avoid the effect of water surface fluc-
tuation on lift force, hydrodynamic force
measurement in heave-free condition has been
developed. Using the measured drag and mo-
ment, an equilibrium equation of two forces is
solved to get running attitude and resistance
of a fast craft. This experimental method has
been developed in order to do a resistance test
of a high-speed craft in a circulating water
channel.
4.2.8.3 Automatic Attitude Control
Method
A more sophisticated experimental
method was developed on the basis of the
same philosophies. The experimental appara-
tus is composed of a force measurement sys-
tem, a system for solving the equilibrium
equation of the forces by a computer and a
system for changing attitude of a model by
stepping motors. Forces acting on a model
craft are measured, and the attitude of it is
changed using those values to satisfy the
equilibrium of forces. Additional forces acting
on appendages and any predictable scale ef-
fects can be taken into account in the calcula-
tion.
4.3 Seakeeping Tests
4.3.1 Seakeeping Investigations
Seakeeping aspects of HSMVs are of par-
ticular interest since the accelerations are in
general high. High accelerations limit the op-
erability from the point of view of passengers
and/or the crew and often also for the cargo.
Due to high speed, the wave encounter fre-
quency can be very high and, depending on

the construction and the size of the vessel, can
be close to the first mode structural natural
frequency. This aspect can have a large effect
on the loads and the fatigue life of the con-
struction.
One of the main differences between the
seakeeping behaviour of HSMVs and con-
ventional ships is the non-linear relationship
between the local wave elevation and the mo-
tions and accelerations of the craft. Also, time
histories of motions and accelerations usually
have complex characteristics such as sharp
peaks or flat troughs that make them un-
suitable for linear analysis methods. The de-
gree of nonlinearity increases with the speed
of the craft. Since semiplaning monohulls op-
erate at lower speeds, they typically have less
complex responses.
When Froude numbers based on length
exceed approximately 1.0, the degree of non-
linearity is considered significant and proce-
dures involving linear superposition are no
longer considered acceptable for analysis of
the measurements (Bulgarelli et al., 1999).
While response amplitude operators (RAO’s)
obtained from model tests in regular waves
may provide interesting information on the re-
sponse of the hull to different exciting fre-
quencies, regular wave tests are not recom-
mended for predicting the response of a
planing hull in random waves. Suggested
guidelines for conducting tests in random
waves and analysing the measured data are
presented in this section.
4.3.2 Model Selection and Construction
There is no minimum requirement for the
model size used for seakeeping tests based on
Reynolds numbers as there is for resistance
tests. It is generally accepted that viscous ef-
fects play a minor role for seakeeping studies,
although such effects do play an important
role for roll damping and forces on stabilising
fins or rudders.
The actual dimensions of the model are
usually governed by the constraints of the ex-
13
perimental facility. The maximum speed of
the towing carriage and the wave maker capa-
bilities are driving factors for a small model,
the required displacement to carry the con-
structional weight, measurement equipment
and propulsion unit(s) give a lower limit.
When small models are built without care-
ful attention to weight distribution, the inertia
of model may be too high to simulate the in-
ertia of the prototype. Therefore model con-
struction should be light and stiff. Wood is
still a good material for models, plywood be-
ing the preferred choice for hard chine models,
a model built from strips on frames for round
bilge hull forms. Good experience is also ob-
tained with foam with a layer of fibre (glass or
carbon) to bring the structural stiffness up to
the required level. This is especially necessary
for long slender models.
Hull stiffness is an important characteris-
tic of the model if the intent is to measure “ri-
gid body accelerations”. Tradeoffs must be
made between reducing model weight and in-
creasing model stiffness. For solid wood
models, a hull thickness of approximately
0.008 × LPP is typical. Hull thickness for
resin/fibre composite models varies consider-
ably depending on fibre characteristics, fibre
orientation and internal hull structure. If ac-
celerometers are installed in the model, the
natural frequency of the hull/accelerometer
foundation can be investigated by tapping the
hull in that area and recording the response of
the accelerometer. The results of these natural
frequency checks should be presented with
the test data.
4.3.2.1 Special Topics Related to Air
Cushion Supported Vehicles
Fans Systems for Air Cushion Supported
Vehicles:
Because air cushion supported vehicles are
in general very light, it can be a problem to
install fans on a model of such a vehicle. In-
stead fans can be installed on the carriage
leading the pressurised air through hoses to
 14
the model. Such a set-up has been used at
MARIN to measure the wave forces on a cap-
tive model of a SES. Using such a set-up,
hysteresis loops were measured in the pres-
sure flow relation just in the hoses in between
the fans and the model. These loops showed
large differences in the pressure flow relation
if compared to static measurements. The ef-
fect is illustrated in Figure 4.3.1.
This effect can be explained by consider-
ing the air in the hose as a mass – spring sys-
tem (Masset et al., 1995). The illustrated phe-
nomenon has an effect on the pressure oscil-
lations in the cushion and hence on the lift
force which is due to the model set-up and not
to the dynamics of the scaled prototype. It is
recommended to avoid such a set-up
whenever possible. Instead it is recommended
to install the fans directly on the model and to
avoid these dynamic effects.
20
18
16
Pressure at fan oulet [kPa]
14
12
10
8
6 ratio’s were reported (Kapsenberg & Blume,
4 1995), so it can be expected that also good
2
0
-1000 -500 0 500 1000
Air flow [m3/s]
Fig. 4.3.1 – Differences in the pressure flow relation if
compared to static measurements experienced at
MARIN on a captive model of a SES.
Dynamic Calibration of Air Cushions:
It is imperative to use Froude scaling for
the over-pressure in the air cushion of an air
cushion supported vehicle. Due to the fact that
normally the ambient pressure is not scaled,
the absolute value of the pressure in the
cushion of the model is far too high. This
means that the dynamics of the air cushion are
incorrectly modelled with a far too high reso-
nance frequency of the air cushion (Moulijn,
1998). Problems associated with this phe-
nomenon are only apparent for higher scale
ratio models.
Results of oscillation tests in the MARIN
depressurised towing tank (Kapsenberg, 1994)
showed the effect of changing the ambient
pressure on the heave added mass and
damping coefficients of a model of an SES.
For models of larger sized full scale ves-
sels, which are model tested in a seakeeping
basin at normal ambient pressure, it is re-
commended to use a diaphragm for the dy-
namic calibration of the air cushion. This dia-
phragm reduces the pressure gradient as a
function of the volume variations, ∂p ∂V , to
the required value. This technique was inde-
pendently developed in France by Ifremer and
in the Netherlands by MARIN.
The diaphragm is a rubber membrane cov-
ering a part of the cushion volume. The mem-
brane is thin so that inertia effects can be ig-
nored. The size and the stretching character-
istics of the membrane must be calibrated de-
pending on the full scale dimensions and the
Dynamic curve model scale ratio adopted. If this technique is
Static fan curve applied, good results of tests at different scale
predictions for the full size vessel can be
1500 made.
For vessels less than about 50 m it is ac-
ceptable to test the model without a dia-
phragm as long as it is kept in mind that the
passenger comfort at high speed and very low
sea states will be over-estimated by the model
tests.
4.3.3 Ride Control Systems
Ride Control Systems can be very impor-
tant for high speed vessels to improve the
seakeeping characteristics. It is important to
include such systems in the model for tests in
waves. If the control system is active, the re-
quirements for the actuators are high, espe-
cially for the phase lag at higher frequencies.

If the control system consists of fins, tur-
bulence stimulation should be applied. The
actual fin angle should always be measured to
have a check on the accuracy of the actuator
system, and the lift force on the fin gives an
idea of the effective angle of incidence which
is relevant for cavitation limits imposed on
the full size vessel.
4.3.4 Towing the Model
The location of the towing point should be
considered in the design of the model.
Tests in head seas can be carried out with
a towed model and restraining the model
completely in the lateral motions. The model
can also be restrained in surge or towed with a
spring system. If the latter is the case, care
should be taken to avoid a resonance frequen-
cy which is close to the wave encounter fre-
quency, a spring should be selected which re-
sults in a resonance frequency at least a factor
2 lower than the lowest wave encounter fre-
quency. If a spring system is used, the reso-
nance frequency should be mentioned in the
test report. It is realistic to have the tow point
on the line of the propeller axis rather than at
the centre of gravity.
For tests in following seas it is considered
important to allow large surge motions of the
model. This freedom is usually essential to
determine correctly if a model is prone to
nose diving or shipping green water over the
bow or not. It will be difficult to model this
correctly using a spring system, an alternative
might be using a constant tension winch. A
good solution is to have a self-propelled
model for tests in following seas. This method
also models the varying propulsive force due
to the wave orbital velocity. Because of the
low encounter frequencies, this phenomenon
will also contribute to the surge motion. This
set-up can still be realised while restraining
the model in the lateral direction.
It is recommended to use a fully self-
propelled model for seakeeping tests in quar-
tering seas. A set-up with a towed model
15
needs a system to keep the model on course;
such a system needs to be rather stiff to com-
pensate for the hydrodynamic mean loads at
high speed and will therefore have an unac-
ceptable influence on the ship motions.
4.3.4.1 Special Topics Related to
Planing Monohulls
Planing monohulls are most often towed at
the intersection of the longitudinal centre of
gravity and the propulsor thrust line. The pro-
pulsive thrust of the prototype is typically
simulated using an inclined towing link or
towing wire, or by using a combination of
pure horizontal force (supplied by a vertical
tow post riding in low-friction bearings) plus
a pure vertical force (unloading weight). In
hulls where the thrust line is very low, for in-
stance outdrive powered boats, it is impos-
sible to tow the model from a low enough
point. In these cases, the tow point is installed
as low as possible and ballast weights are
shifted to compensate for the bow-down mo-
ment imposed by the high tow point. The ef-
fects of pitching moments created by ap-
pendages and propeller forces should be esti-
mated and accounted for separately. It is stan-
dard practice to tow models in waves with a
non-varying towing thrust angle.
Planing boat models usually require sealed
decks to keep water out. In many cases, mod-
els are built with simplified superstructures to
evaluate spray and solid water impacts. De-
tails of the superstructure should be included
in the test report to allow corrections between
the air drag of the model and of the full scale
prototype.
4.3.5 Course Control
Using a completely free running model re-
quires a control system to keep the model on
course. At full scale as well as in the towing
tank, high speed vessels can experience
problems in keeping a steady course. This is
especially true for waterjet propelled vessels,
since they normally do not have a skeg.
16
The roll motion induced by the rudder or
water jet nozzle is a point of concern. This ef-
fect can be considerable and is controlled by
the control law of the autopilot. Preferably
filtering should be used to avoid rud-
der/nozzle oscillations in the wave encounter
frequency.
A check should always be made on the
delivered torque by the steering gear in a bol-
lard pull condition at maximum RPM to pre-
vent problems with an insufficient sized ac-
tuator. However, this is by no means suffi-
cient to ensure a fully controlled model.
4.3.6 Typical Model Tests
Seakeeping model tests are typically con-
ducted in head or following seas with the
model restrained in surge, sway, roll and yaw
and free in pitch and heave. Fridsma (1971),
in testing planing hulls, showed that in head
seas, with Fn > 1 , constant speed tests pro-
duced essentially the same motions and added
resistance as tests with the model free to surge,
with constant thrust. As noted earlier, it is
standard practice to tow models in waves with
a non-varying towing thrust angle.
Usually, at the lower speeds of displace-
ment and semi-planing boats, tests can be
conducted in a towing tank and measurements
are usually transferred to a towing carriage
that is driven in close proximity to the model
and connected by a wiring harness. For most
higher speed planing boat models, free run-
ning tests must be conducted in the open
water of ponds or lakes because of the dis-
tance covered at high speed. Measurement
signals are typically sent to shore via radio
telemetry for recording. The accelerations of
high speed planing monohulls typically make
onboard digital storage impossible.
Free running models are also used for
seakeeping measurements. For instance, free
running models of HSMVs with fully sub-
merged hydrofoils might be tested. When do-
ing seakeeping and also manoeuvring tests
with a free model, it is considered necessary
to test the model complete with foil system
and actively controlled flaps (if fitted to the
full scale vessel). The scale effect on lift is
considered less critical for seakeeping, since
take-off condition is usually not tested in
seakeeping. The scale effect on lift can be
compensated by the flaps or by ballast. When
testing a towed model, it is not necessary to
correct for scale effect on drag. The scale ef-
fect on drag has importance for the forces
from the propulsors. If the model is run self-
propelled and no correction is made, the pro-
peller forces will be over-estimated. The best
solution is to apply a constant tow rope force
at the correct vertical position (note that using
a weight will introduce an inertia that is not
wanted). If it is impossible to apply a correc-
tion force at the correct vertical position, it is
often the best solution to accept the over-
estimated propulsor thrust.
From the questionnaire it was found from
those who responded that, for irregular wave
testing, 38 percent of the facilities use less
than 100 wave encounters as a standard
minimum for basing statistical representations
of test data. 62 percent of the facilities use
more than 100 encounters. The survey had
wide variations, with a mean of 78 encounters.
Based on a review of the data, a minimum of
100 wave encounters is recommended at this
time for testing in head seas. The modal fre-
quency of the encountered wave spectrum
should be used to estimate the total run time
required. Pierce (1992) proposed a method for
relating run length to statistical error. Alt-
hough the method is based on linear analysis
methods, it should provide a rough approxi-
mation of the uncertainty of the data set based
on the total run time used to obtain 100 wave
encounters.
At the high speeds of planing craft, the
tank length of many facilities limits the steady
speed run time to only a few seconds. There-
fore many runs may be required to obtain 100
wave encounters. For obtaining statistics such
as mean, standard deviation and probability

levels it is standard practice to splice the
measurement time records together and analy-
se the combined set of data.
Occasionally model testing clients are in-
terested in evaluating the relative seakeeping
behaviour of a new hull with an existing hull
(Schleicher, 1997). In this case it may be ad-
vantageous to run comparative tests with the
two models side-by-side. This ensures that
both models experience identical irregular
waves, and if the data are used for compara-
tive purposes rather than absolute predictions,
reduces the effects of analysing nonlinear re-
sponses using conventional statistics. Until
studies of interference effects between the
hulls are performed, the maximum model size
should be limited to one half the size consid-
ered acceptable for a standard seakeeping test.
An alternative method of evaluating the
motions of a hull in a seaway is to subject the
model to forced oscillations and to conduct
free decay tests. The hydrodynamic coeffi-
cients obtained from these experiments can be
used in numerical modelling simulations to
predict the response of the hull to wave ex-
citing forces. In these tests the model is fixed
in six degrees of freedom and up to three
moments may be measured. As for all tests in
which the model is restrained in heave, the
level of the water surface should be monitored
when the model is underway to measure any
change in water level caused by the aerody-
namic pressure of the towing carriage (see
Murakami, 1981 and 17th ITTC HSMV Re-
port).
4.3.7 Quantities Measured During Tests
• Wave height (fixed or encountered).
• Speed.
• Resistance.
• Trim.
• Heave – Heave is typically measured at
the longitudinal position of the centre of
gravity.
• Accelerations – It is standard procedure to
measure accelerations in the vertical body
17
axis at several locations. In order to make
it possible to compare data from different
facilities, three standard locations are pro-
posed:
1. Bow – 10 percent of LPP aft of for-
ward perpendicular;
2. CG – position should be changed for
each condition tested;
3. Stern – 10 percent of LPP forward of
aft perpendicular.
• Pressures – Pressures have been measured
at virtually any location; there is no
guidance on location at this time.
• Relative bow motions – These measure-
ments are difficult to make on HSMVs,
specially on planing hulls because of the
large change in attitude of the model at
high speeds.
• Wetted Surface – Records are typically
kept of the bottom surface wetted by solid
water. In some cases, the area wetted by
spray is also measured and recorded.
4.3.8 Instrumentation Used in Wave Tests
• Wave height – Measurements of the en-
countered wave are more desirable than
measurements of the stationary wave.
However it is considerably more difficult
to measure encountered wave elevation at
high speeds. Surface piercing wave probes
are generally unreliable because of water
run-up on the forward side of the probe
and ventilation of the back side of the
probe. Good success has been reported in
measuring encounter wave using electro-
mechanical servo probes. Acoustic probes
have also been used successfully when the
test data have been post-processed using
computer based methods for fairing curves
through “dropout” points in the time rec-
ords, although frequency response prob-
lems have been noted by some investiga-
tors (Hirayama et al., 1988). MARIN uses
a servo controlled wave gauge consisting
of a needle which follows the wave sur-
face. The Figure 4.3.2 gives results of a
series of experiments in a regular wave
 18
(wave frequency = 4.45 rad/s, wave am-
plitude = 23 mm) with increasing speed of
the carriage. The measured amplitude by
the classical wave gauge is compared to
the value measured by the wave servo.
The figure shows that the differences are
less than 3.5% for this speed range and
that the error does not uniformly increase
with speed. Apparently run-up effects and
ventilation compensate at higher speed.
• Speed – Speed measurement is typically
made using the same equipment used for
resistance tests.
• Added resistance/thrust in waves – For
added resistance measurements, instru-
mentation is needed with approximately
twice the capacity used in calm water re-
sistance tests. The frequency response of
the system should be considered if statis-
tics other than average added resistance
are to be evaluated. Preference is given to
using self-propelled models in waves
rather than towed models. This means that
added resistance measurements are impos-
sible and that the thrust must be measured
instead. For propellers the thrust can be
measured in the axis on the outside of the
aftermost bearing, for waterjets the thrust
can be determined from measuring the
pressure in the nozzle. It is recommended
to measure the torque in the shaft just be-
fore the waterjet to detect air ingestion.
The torque cannot be used to estimate the
required power.
1.04
amplitude servo/probe
1.03
1.02
1.01
1.00
0.99
0 2 4 6
carriage speed [m/s]
Fig. 4.3.2 – Results of a series of experiments carried
out at MARIN with a servo controlled wave gauge.
• Motions measurements – Measurement of
motions is important and should in general
be non-intrusive. Instrumentation used in
resistance tests is often unsuitable for
seakeeping because of the dynamic nature
of measurements (fast turnaround and lar-
ge excursions). For instance, weighted
string pulley systems for measuring rise
and fall at bow and stern can not be used
because model accelerations often exceed
1g. However, if tests in head or following
seas are carried out and the model is re-
strained in the lateral motions, a reliable
measurement is obtained by having a wire-
over-potentiometer measurement of the
vertical motion at the bow and stern. Such
a system cannot be used for free running
tests. Reliable 6 DoF optical systems are
now available which can be used for this
situation. At MARIN a system is being
used with the heavy sensor located on the
carriage and a very light transmitter, con-
sisting of a set of three light sources, lo-
cated on the model. The system has a typi-
cal resolution of 0.1 mm for the transla-
tions and 0.1 deg for the rotations (practi-
cal resolution, the claimed accuracy for
laboratory conditions is higher). The
measurement area for this resolution is 1 x
1.2 meters.
• Accelerations – Ideally, the purpose of the
acceleration measurements should be
known in advance. If for structural re-
sponse, different parts of the full scale will
respond at different frequencies (e.g. a lar-
ge heavy panel will not respond to very
sharp acceleration peak, whilst a small
stiff panel will); if for human response
regarding injury, a slow response time as
well (Gollwitzer & Peterson, 1996); if for
human response regarding motion sick-
ness, use the frequency octave method.
8 For operability purposes the accelerations
are more important than the motions. It is
recommended to measure these directly
instead of deriving them by double differ-
entiation of the motions. Servopendulum

accelerometers used for low speed ship
model testing are very accurate but are
limited to frequencies below 100 Hz; pie-
zoelectric accelerometers are limited to
frequencies above 10 Hz; piezoresistive
and variable capacitance are good alterna-
tives that perform well over a frequency
range of 0 to 1000+ Hz. It is recommend-
ed that accelerometer specifications, in-
cluding: range, natural frequency,
damping ratio and linearity be included
when reporting acceleration data. Accel-
eration time histories for planing hulls are
typically triangular with short rise time,
sharp peak and more gradual decay time.
Peak acceleration varies with the frequen-
cy response of the transducer, sampling
rate, filter rate, etc (Zseleczky & McKee,
1989).
• Pressures – Problems with measuring
pressures are very similar to acceleration
measurement problems because of short
rise time. Another problem is that given
the same impulse, large-face pressure
transducers measure lower peak pressures
over a longer time than small-faced trans-
ducers because the stagnation point is
typically concentrated and fast moving.
Published data should not only list the lo-
cation of transducers, but also the diameter
of the sensor face, the range, frequency re-
sponse and linearity of the transducer. For
measuring panel loadings, strain gauges on
panels with scaled dimensions and flex-
ural properties are recommended rather
than point pressure measurements.
• Relative bow motions – These measure-
ments are difficult to obtain at high speeds
and are rarely attempted; no recommended
practice is available at present.
4.3.9 Test Wave Environment
If the model testing is focussed on deter-
mining the operational limits, and the expect-
ed operational limits are not extreme condi-
tions, it is recommended to use the two
parameter ITTC Standard Spectrum as a de-
19
scription of the wave energy distribution over
the frequency range. The main reason for this
choice is that such a spectrum describes a
fully developed sea state which is relevant for
normal operational conditions.
For tests in extreme sea states it is recom-
mended to use a Jonswap type of representa-
tion for the wave energy spectrum with a
peakedness parameter >> 1. Because of the
relatively short duration of the extreme
weather, it cannot be expected that such a sea
state will be fully developed.
These coastal seastates have low signifi-
cant wave heights (1-3 m) compared with
conventional ship seastates (2-6+ m) so the
wavemaker may be operating outside its des-
ign envelope. This should be considered be-
fore selecting the scale ratio.
The presence of swell is for most high
speed vessels a very important factor. The en-
counter frequency of local wind generated
waves is usually sufficiently high so that it
does not result in any significant motions; the
motions due to the swell can be much higher
and therefore more relevant for both operabil-
ity and safety aspects. It is recommended to
use local wave information as much as possi-
ble for operability studies.
Keeping the model speed is a point of
consideration for free running models. Usu-
ally the propeller is driven by an electrical
motor and the RPM of this motor is kept con-
stant. The characteristics of the motor of the
ship are usually not modelled, but this is not
considered to be a significant disadvantage for
high speed vessels.
4.3.9.1 Tests in Regular Waves
Tests in regular waves are often used for
research projects. A problem is to keep the
model speed constant for the runs in different
wave conditions (frequency and amplitude)
due to the changing added resistance. If the
duration of the run is sufficient, a solution
may be found in a feed-back to the RPM con-
20
trol of the electrical motor. The power to
weight ratio of high speed models is normally
sufficient to have a quick response and to ar-
rive at the required value in a short time.
Motion responses of high speed vessels
can be non-linear to a significant degree. This
means that the motion RAO derived from ex-
periments is not unique and cannot be used to
determine the performance in irregular seas
with confidence.
4.3.9.2 Tests in Irregular Seas
Tests in irregular seas are always required
to study extreme effects like slamming, ship-
ping of green water, nose diving and broach-
ing for any ship. Due to the non-linear relation
between ship motions (and accelerations) and
the wave amplitude for some high speed ves-
sels, tests in irregular seas are also required to
check the normal operational limits. This
makes computer predictions for such condi-
tions quite important since it is not feasible to
carry out tests in many sea states. A reason-
able accurate prediction of the operational
limits must be made so that critical conditions
can be selected with confidence.
4.3.10 Data Sampling Rate
The sample rate of the data collection
system should be sufficiently high to measure
the quantities of interest. If these are only the
ship motions, a sample rate of 5 x the maxi-
mum wave encounter frequency is sufficient.
If loads are being measured in a cross section
of the model, a much higher (minimum 3 x)
sample rate is necessary to measure non-linear
effects accurately.
Slamming measurements require very high
sample rates, the required rate is dependent on
the resonance frequency of the slamming sen-
sor. Piezo-electric pressure gauges have for
instance a very high resonance frequencies re-
quiring very high sample rates; a value of 10
kHz is not uncommon.
Focus is now more on measuring slam-
ming with a sensor with a resonance frequen-
cy related to the construction of the full size
ship. If such a sensor is used, the requirements
on the sample rate are usually lower, and 5 x
the lowest resonance peak of the sensor
should be sufficient.
4.3.11 Data Collection
A short run length is not a problem for a
test in regular waves since harmonic analysis
can be done if 5-10 complete wave encounters
are measured. As discussed in Section 4.3.6, a
large number of wave encounters is necessary
for tests in irregular seas, which might make it
necessary to do the test by combining runs in
different realisations of the irregular sea. The
number of wave encounters required for the
analysis is dependent on the purpose of the
test. If only RMS values of motions and ac-
celerations are required, 75 wave encounters
will give a sufficient accuracy. If parameters
of extreme phenomena need to be measured
like slamming pressures, the required run
length should be such that a number of slams
are encountered. If the extreme slam pressure
needs to be determined a run length sufficient
to record 100 slams is required.
In general it is recommended to define the
run length based on the number of relevant
events rather than just a certain measurement
time.
4.4 Propulsion Tests
The analysis of the questionnaires indicat-
ed important issues relevant to propulsion ex-
periments which should be considered. The
most important are:
• model size;
• the change of running attitude between re-
sistance and propulsion test;
• scaling of wake;
• scaling of propeller efficiency ηo ;
• cavitation and ventilation;
• thrust measurement for waterjet propelled
vehicles.

Recommended codes of practice are out-
lined after each issue when possible.
4.4.1 Model Size
Generally speaking, for propulsion tests
the model should be as large as possible. The
upper limits of the model size are given by the
highest carriage speed and by tank depth and
width. The maximum carriage speed should
be somewhat lower than for resistance tests to
provide time for adjusting the propulsor’s
revolutions.
The lower limit of model size depends on
the weight of the propulsor device, the driving
axis, gears, motor and the measurement de-
vices. To avoid increase in model size the
weights of the propulsion units and dyna-
mometers should be very small.
For craft having screw propellers, the
model size depends mainly on the scale ratio
of the full scale propeller to the available
stock propeller diameter. The scale ratio
should guarantee supercritical Reynolds num-
bers at 0.7R of the blades. In this case, if a
model propeller with a correct pitch ratio is
not available, a stock propeller should be se-
lected that has a geometrically similar sum of
pitch and diameter as the prototype.
4.4.2 Change of Running Attitude Between
Resistance and Self Propulsion Test
The large variety of propulsion devices in
use for HSMVs and their different arrange-
ments have considerable influence on the run-
ning trim and sinkage. Thus they influence the
craft’s performance, such as powering, ma-
noeuvring and high speed dynamic stability.
Experimental results coming from propul-
sion tests, as well as predictions, are subject
to various influences like propulsor/hull inter-
action, cavitation, ventilation and, scale ef-
fects due to inequality of Reynolds numbers
for model and full scale.
21
In designing and testing HSMVs and in
predicting their power performance, the in-
duced effects of appendages and their influ-
ence on the equilibrium running condition
must be taken into account to obtain reliable
predictions of full-scale trim and powering
data. In fact, even if the velocity and pressure
fields on the appendages have relatively small
effect on the hull, their wake fields affect the
propulsor performance directly.
To avoid or minimise errors in model test
data evaluation and in predicting performance,
it is important to evaluate correctly the drag of
the appended hull. That can be done experi-
mentally by making large models with ap-
pendages and propulsors having significant
dimensions (higher Reynolds numbers). If the
facility permits to perform tests at high speed
on a large model, the experiments to deter-
mine the appendage drag can be carried out
with appended model first and then repeated
without appendages, with the model locked at
the same trim and heave.
If that is not possible, as already discussed
in Section 4.2.2, the appendage drag can be
determined analytically with a method such as
that proposed by Hadler, 1966.
4.4.3 Scaling of Wake
Past experience and some full-scale data
indicate that the model wake is essentially the
same as the full-scale wake.
The propellers of a semi-displacement
craft, for instance, operate largely or com-
pletely outside the hull boundary layer; scale
effects can therefore be ignored for most in-
clined shaft propeller arrangements.
Although the conventionally obtained
wake fraction is easier to determine, never-
theless, for more accurate full-scale prediction
of propulsion factors the analyses should be
based on the effective wake fraction.
The main problem seems to be which val-
ues of wake fraction should be taken into ac-
count when making full-scale predictions. The
22
wake fractions are influenced by trim, free-
surface effects, appendages, etc. Oblique in-
flow to the propeller can have noticeable ef-
fects if axial flow propeller data are used.
Hence effective wake fractions analysed by
traditional procedures like ITTC Method
78/88 cannot be regarded as a true measure of
the inflow retardation due to the influence of
the ship’s hull. The traditional wake is largely
a propeller rather than a hull characteristic.
Since the inclined shaft propeller charac-
teristics in the behind condition are different
from those of the axial shaft case, the analyses
based on oblique open water characteristics
can provide more reliable predictions.
It is considered that more full-scale data is
required for the different HSMV types with
different propulsor devices before a code of
practice can be proposed.
4.4.4 Scaling of Propeller Efficiency
The answers to the questionnaire pointed
out that the scaling of propeller efficiency is a
problem. In fact the power and rotation rate
prediction on the basis of model tests requires
an accurate assessment of scale and roughness
effects on propeller characteristics.
The ITTC’78 method incorporates a cor-
rection rule based on scale effects on the drag
of the blade section only. The rule, for pro-
peller scale effect, ignores the difference in
extent of the laminar flow over the propeller
blades between model and full size. The result
is that the method predicts greater scale effect
on KT and ηo coefficients and minor scale ef-
fect on CD.
To overcome this there are two ap-
proaches:
1 to control the flow over the propeller
blades;
2 to develop a more accurate scaling pro-
cedure that accounts in a better way for
the mixed type of flow.
Both approaches require that the scaling
rule takes into account the effect of Reynolds
number on the lift. The influence of Reynolds
number on the lift has been discussed in sev-
eral past ITTC reports but it is far from being
solved.
The method of controlling the flow over
the propeller blades should be treated with
extreme care. The results of model tests car-
ried out in the past indicate that turbulence
tripping on the propeller blades is not recom-
mended for application in routine practice.
A method that seems promising is that
proposed by the HSMV Committee of 19th
ITTC as “Alternative Analysis and Prediction
Procedure” which is especially useful if wake,
thrust deduction and efficiency elements are
influenced by oblique inflow. This method
analyses the product
η H ⋅η R = η D / η0 = const
as a function of the propeller loading. The
product can be considered unaffected by scale
effects. The propeller efficiency of the model
propulsion test may be replaced by that of a
large scale geosim propeller with small scale
effect corrections applied.
The method works well for non cavitating
and partially cavitating propellers. For fully
cavitating propellers additional corrections
should be applied.
4.4.5 Cavitation, Ventilation and Propeller
Shaft Inclination
Cavitation and ventilation effects on pro-
peller driven HSMVs have a large influence
on their dynamic behaviour and on their re-
sulting performances. For that reason alterna-
tive and unconventional means of propulsion
are used in such types of craft. The majority
of high-speed craft in operation utilise alter-
natives such as waterjet propulsion systems,
surface piercing propellers, super-cavitating
propellers.
Cavitation and ventilation effects on
power performance predictions can be divided
into:

• influence on propulsor characteristics;
• influence on thrust demand and running
condition.
To account for these phenomena tests
should be carried out in a vacuum facility.
Because it is very difficult to perform such
tests at high speed, the prediction on HSMVs
is usually made in two steps: by tests in a
towing tank and by tests in cavitation tunnel
or depressurised water circulating channel.
As already said, cavitation and ventilation
influence the performance of the propulsor.
The best method to account for this influence
is to carry out cavitation tests with a model
fitted with all the appendages in a large cavi-
tation channel. The test procedure and treat-
ment of data have been presented extensively
in the proceedings of preceding ITTC Confer-
ences. For those organisations that do not
have at their disposal large cavitation facilities,
different approaches were proposed in the
Report of the HSMV Committee of 19th ITTC.
Furthermore, cavitation and ventilation in-
fluence the forces and moments induced on
the hull by the propulsors, especially for ships
with inclined shaft arrangements. Therefore
the resulting trim is different compared with
that measured during tests without these phe-
nomena. It means that results from model
self-propulsion tests in a towing tank at at-
mospheric pressure are in most cases not rep-
resentative for full scale HSMVs. In that case
the effects of cavitation have to be accounted
for separately.
In this respect in propeller driven craft, the
most severe effects are induced by the shaft
inclination. In these craft inclined shafts are
commonly used to place the propellers well
below the hull to avoid or reduce the risk of
air suction or ventilation at all trim angles.
The oblique flow on the propeller causes a
cyclic variation of the angle of attack on the
propeller blade sections. As a result, thrust
and torque fluctuations become larger and
cavitation phenomena are intensified. In
23
making full scale evaluations starting from
model results, an analysis procedure that uses
open water propeller characteristics measured
in oblique flow conditions should be adopted.
The traditional open water approach has
been suggested in the past. In fact there are
complexities regarding the experiments and
procedures involved and there are uncertainti-
es with respect to the techniques for uncon-
ventional propulsors. More work is required
on this matter.
It is well known that the disadvantages of
using traditional axial flow open water tests
are due to the influence of interaction effects
and to the difficulties in determining the nor-
mal force and its change due to cavitation af-
fecting thrust and efficiency.
For hydrofoil ships, in most cases, the
propulsor-hull interaction effects can be ne-
glected. Thus, ordinary propulsion tests are
usually not required. On the other hand, there
might be important propulsor-foil interaction
effects, where cavitation plays an important
role, since hydrofoils operate at high speeds.
It is therefore recommended to test the pro-
pulsor together with the relevant parts of the
foil system in a cavitation tunnel. Depending
on the maximum obtainable Reynolds number
and available measurement equipment, the
cavitation tunnel experiments might also be
used to determine lift and drag on the foil
system. In the cavitation experiments, it is
important that any flaps are modelled and set
to realistic angles during the experiments.
4.4.6 Thrust Measurement for Waterjet Pro-
pelled Vehicles
The questionnaire indicated that this sub-
ject can be a matter of opinion. The Specialist
Committee on Waterjets has worked on the
development of the waterjet self-propulsion
test standardisation. To avoid overlap any fi-
nal decision on this matter has been left to the
Committee on Waterjets. However, as an il-
lustration of common practice, the method
used at HSVA for predicting the thrust of
24
waterjet propelled ships is described briefly. It
is developed starting from the Momentum
Flux Method proposed by the Waterjet Group
of the 21st ITTC (1996).
According to this method the thrust is not
measured directly, but obtained by assuming
the rate of change of momentum to be equal
to the sum of all forces (with appropriate sim-
plifications) acting on a given control volume.
The method is based both on calculations and
measurements. The main steps are:
• the calculation of the momentum velocity
at nozzle exit;
• the calculation of the momentum velocity
upstream of the inlet;
• the determination of the effective inlet;
• the determination of the flow velocity up-
stream of the inlet.
A detailed description of the method is re-
ported in the HSVA Report “Performance and
Analysis of Ship Powering Tests, Waterjet
Propulsion Test Evaluation Method” given to
the Committee in response to the question-
naire.
4.5 Manoeuvring Tests
Manoeuvrability investigations of high-
speed marine vehicles have grown more im-
portant as safe operations are of major con-
cern today.
At the moment, there is no complete set of
six-degrees-of-freedom equations of motions
and accompanying definitions of various cou-
pled and uncoupled hydrodynamic coeffi-
cients required to simulate the manoeuvring
characteristics of HSMVs. Model tests are
then found more reliable as a prediction
method of manoeuvrability, but even model
tests have their shortcomings such as difficul-
ties to model cavitation and ventilation phe-
nomena.
It is important to understand that the hy-
drodynamic forces and moments on the hull
are highly non-linear, and are strongly depen-
dent upon:
• wetted hull geometry;
• appendages;
• trim angle;
• roll angle;
• yaw angle;
• wetted length;
• Froude number;
all of which vary significantly during a
manoeuvre. Further, there is evidence of
strong yaw-roll coupling which is dependent
upon trim angle and speed. While there may
be some success in analytical or semi-
empirical representations of the static stability
derivatives, there has been no demonstrated
success in the prediction of the dynamic
(damping and added mass) derivatives.
Three basic test methods can be used in
exploratory studies of manoeuvrability of
high-speed marine vehicles. They are:
• Captive model tests on rotating arm and
straight course;
• Captive model tests on planar-motion
mechanisms;
• Free-running model tests in a manoeu-
vring basin or outdoors with a radio-
controlled model.
The rotating arm tests provide data useful
for predicting directional stability and turning
circles, but do not provide trim, roll and heave
damping and added mass terms.
A planar motion mechanism can provide
these quantities, but contains a frequency de-
pendence, which must still be identified. The
free-running models do not provide basic hy-
drodynamic data. All three are deficient in not
representing the cavitation effects on propel-
lers, rudders, support struts, fins, etc.
4.5.1 Captive Tests
By straight course tests and rotating arm
tests the coursekeeping stability and running
performance can be evaluated. The procedure
is to mount on the rotating arm or straight-
course apparatus an appended model, which is

free-to-heave and pitch, but restrained at fixed
values of yaw and roll and locked in surge and
sway. The drag, side force, yaw moment and
roll moment are measured at various speeds,
turning radii and rudder deflection.
For dynamic course stability (neglecting
yaw-roll coupling) the data for zero roll are
plotted as a function of yaw angle and radius
for each test speed and from these values of
Nv, Nr, Yv and Yr are obtained. The course
stability is found from linear differential
equations of motion using these hydrody-
namic derivatives and introducing appendage
and propulsion effects, which have not been
included in the test model. The equilibrium
turning diameter, roll angle and rudder de-
flection at each test speed is found by solving
the simultaneous equations of yaw, roll and
side force equilibrium, including the estimat-
ed propeller and appendage effects which ha-
ve not been included in the test model. The
uncertainty in both procedures stems from the
omission of unknown cavitation, ventilation
and interference effects.
4.5.2 Free-Running Model Tests
The roll and yaw angle variations during a
manoeuvre are significant for high-speed
marine vehicles compared to displacement
ships’ performance. Free-running model tests
are an alternative to captive tests. By carrying
out free-running model tests the total dynam-
ics during a manoeuvre is better modelled
compared to the captive tests and conse-
quently test programmes do not need to be as
comprehensive as for the captive tests.
Free-running tests performed at high
Froude numbers do not in general differ from
free-running tests at low Froude numbers ex-
cept that they usually need a large manoeu-
vring basin to complete standard manoeuvres
such as turning circle and pull-out tests, spiral
and zig-zag tests. Due to the need for large
manoeuvring space at high-speed, free-
running model tests are occasionally per-
formed outdoors.
25
For free-running model tests outdoors the
scale factor of the craft is determined by the
craft’s speed, displacement and propulsion
type. The scale model for outdoor tests is usu-
ally manufactured in GRP or similar material
resistant to environmental impacts. A simpli-
fied superstructure is recommended in order
to take into consideration the aerodynamics
and its effect on the running trim, heel angle
etc. Correct scaled and configured propulsions
and appendages like fins, bilge keels, trim-
ming tabs, struts etc are of vital importance
during the model tests.
The model is radio-controlled and usually
gets its power supply from internal combus-
tion engines. The tests are usually carried out
at constant number of shaft revolutions.
The test types carried out at outdoors tests
do not differ from the conventional free-
running test indoors.
Examples of measured parameters could
be:
• model speed;
• position of the model;
• number of revolutions;
• yaw rate;
• rudder deflection;
• heel and pitch angles;
• propulsion parameters.
There are two ways to analyse the data
collected from the free-running outdoor ma-
noeuvring tests. One way is to straightfor-
wardly scale the measured signals for the dif-
ferent test conditions. The other way is to ob-
tain the numerical value of the hydrodynamic
derivatives from the model tests and use them
to establish a reliable mathematical simulation
model.
4.6 Structural Loads
The dimensioning of large high speed ve-
hicles demands a knowledge and methods to
determine the limiting environmental loads,
26
operational aspects and structural strength. To
achieve good design load predictions, appro-
priate model test techniques must be devel-
oped. Model tests are also required for verifi-
cation and calibration of theoretical methods
and numerical codes.
Structural loads can be divided into hy-
droelastic problems and non-hydroelastic
problems. The former requires the structural
dynamics to be modelled correctly. That
means that not only the mass distribution, but
also the stiffness must be scaled. The scaling
relations are:
Structural mass: ( M ) S = ( M ) M ⋅ λ3
Bending stiffness: ( EI ) S = ( EI ) M ⋅ λ5
Shear stiffness: ( AG ) S = ( AG ) M ⋅ λ3
Structural loads can also be divided into
local and global loads. This division implies
that each of these cases can be treated indi-
vidually. In some cases hydroelasticity is im-
portant, in others it can be neglected.
4.6.1 Local Loads
For the local problem the slamming force
is the most important load contribution.
Firstly, one should not consider point pres-
sures, but rather forces on a sensibly chosen
area, e.g. a single plating field (Carcaterra &
Ciappi, 1998, Carcaterra et al., 1999). The
dynamic behaviour of the elastic plate is gov-
erned by the structural properties of the plat-
ing (including longitudinal stiffeners between
two transverse frames). The global response
of the ship will serve as input to the problem
defining the relative speed and orientation of
the local area of interest during the impact
with the waves. For flat-bottom slamming,
and wet-deck slamming of multihulls, hydro-
elasticity is crucial for the magnitude of the
slamming loads. Testing is done either with
the dynamic behaviour correctly modelled, or
the results are used solely to document the oc-
currence of slamming, not the magnitude of
the forces. For bow slamming, and slamming
on surfaces with a deadrise angle of more than
approximately 15°, hydroelastic effects are
less important and testing can be performed
without modelling the local structural dy-
namics (Faltinsen, 1998). Forces can be
measured by means of a suitably sized panel
mounted on a strain-gauge arrangement. The
panel should be stiffly mounted in order to
avoid artificial hydroelastic effects.
The most simple method for measurement
of slamming with modelled structural dy-
namics is with a strain-gauge mounted stiff
panel where stiffness and mass are correctly
scaled to yield the correct frequency of the
first eigenmode. A more correct method is to
model a suitable part of the hull plating cor-
rectly (mass and stiffness). An even more
complete, but elaborate way is to model for
instance a wet deck completely with stiffeners
and plating. An alternative to modelling the
structural dynamics, at least for wet deck
slamming, is to calculate the magnitude of the
local loads numerically. A practical method is
described by Kvålsvold (1994) and Kvålsvold
et al. (1995, 1996).
3.0
Slam pressure [kPa]
2.0
1.0
0.0
-1.0
0.00 0.05 0.10 0.15 0.20 0.25 0.30
time [s]
Fig. 4.6.1 – Results of a slamming experiment on the
wet deck of a model of a SES, off-cushion mode.
Measurement of slamming is demanding
with respect to instrumentation. Compared
with conventional model testing, a much
higher sampling frequency of the digital re-
cording of the different measurements is re-
quired. The sampling frequency should be in
the order of 100 kHz in most cases. Further,

the amplifiers have to satisfy requirements to
a large linear frequency range. For the differ-
ent transducers, sufficiently low rise time and
high resonance frequency of the transducers
are required to match the time scale of an im-
pact.
Figure 4.6.1 shows a typical duration of a
slam measured with a flexible panel. Values
in figure are given on model scale. They are
collected with a sampling rate equal to 7 kHz.
Pressure is measured on a panel which is
connected with a strain gauge to the model.
Resonance frequency of panel was equal to
700 Hz.
4.6.2 Global Loads
The most important global loads are hull
girder bending moments (vertical bending
moments), and for catamarans and other
multihulls, the forces in the bridge deck
structure (split moment, pitch connecting
moment, vertical shear force) will often be
critical. Hydroelastic effects of importance are
springing and whipping. There are two princi-
pally different ways of constructing the
model:
• Continually Elastic model;
• Segmented model.
The continually elastic model makes it
possible to measure the strain in a very large
number of positions. Theoretically, this is also
the most correct way to model the structural
dynamics, since any number of eigenmodes
can be represented. However, one is usually
interested in forces not strain, and establish-
ment of a reliable relation between measured
strain and global (or local) forces on an elastic
model is very difficult. One might need to do
a very detailed FEM analysis of the model to
establish the relation between strain and
global forces. In addition, this relation might
be non-linear and have hysteresis, depending
on the type of materials that are used.
A segmented model is built in a number of
stiff segments connected with force transduc-
27
ers (usually strain gauge type). If the structural
dynamics is to be modelled, springs are added
between the stiff sections. The spring stiffness
and mass distribution should be selected in
order to represent the first few eigenmodes of
the hull girder correctly. The segmented
model is much easier to calibrate, and the re-
sults are much more readily analysed. The
drawbacks are problems with the weight of
the model, and a limited number of eigen-
modes that can be represented (limited by the
number of sections). For most practical cases,
the segmented model should be preferred.
In the case of an hydroelastic model, the
structural properties of the full scale ship must
be determined previous to the design of the
model, for instance by FEM analysis, regard-
less of the selected modelling technique.
For catamarans and other multihulls, both
demihulls and the connection between the
demihulls (bridging structure) should be seg-
mented and joined with force transducers, in
the same manner as for the hull girder. The
model set-up is shown in Figure 4.6.2.
Fig. 4.6.2 – Set-up of a hydroelastic catamaran model.
4.7 Dynamic Instability Tests
4.7.1 Motions in calm water due to dynamic
instability
High speed stability problems of planing
monohulls in calm water are very important as
pointed out in the HSMV Committee report to
the 18th ITTC.
4.7.1.1 Transverse Stability Loss at
High Speed
At high speed the transverse stability de-
creases, and the stability loss causes sudden
28
large heel and chine walking of planing
monohulls, as stated by Washio et al. (1993),
Katayama & Ikeda (1995), and Ibaragi et al.
(1996). Katayama & Ikeda (1996) pointed out
that the dependency of the transverse stability
on running trim causes unstable parametric
roll motion when the craft has pitching mo-
tion.
4.7.1.2 Longitudinal Instability at High
Speed
Porpoising is a typical longitudinal peri-
odic motion due to longitudinal instability.
Katayama & Ikeda (1996, 1997, 1998) report-
ed that this pitch and heave combined motion
is a kind of self-exciting motion of coupled
oscillation system in which coupled restoring
coefficients have different signs to each other.
Bow drop is also longitudinal non-
periodic motion which can occur for planing
monohulls running at high speed. The longi-
tudinal stability characteristics at high speed
also have an important role in the phenome-
non.
4.7.1.3 Course Keeping Ability
The large heel of planing monohulls
sometimes causes course keeping problems
even in calm water, like well known broach-
ing in following seas (Washio et al., 1993,
Ibaragi et al., 1996).
4.7.2 Tests
4.7.2.1 Motion Measurements
Motion measurements of a towed model
free to move in the relevant degrees of free-
dom in calm water are commonly carried out
to find the inception of, or the region of ad-
vanced velocity, and the conditions of ship
weight and location of CG where ship mo-
tions due to dynamic instability occur. Such
experiments can also be carried out in combi-
nation with usual resistance tests. A free run-
ning radio-controlled propelled model test can
provide not only stability information but also
manoeuvring information.
The influence of the propulsor and ap-
pendages on the ship motion due to dynamic
instabilities can be significant, and the model
used in the experiment should be with scaled
propulsors and appendages or with suitably
adapted towing procedures. Furthermore, the
effects of cavitation and ventilation, which
may be significant for small models, can not
be avoided in tests at atmospheric pressure.
Experiments in cavitation facilities may be
required when the effect is expected to be sig-
nificant.
If such effects can not be avoided in ex-
periments, a computer simulation using
measured or predicted hydrodynamic coeffi-
cients would be more reliable than the ex-
periments.
4.7.2.2 Force Measurements
In order to check the occurrence of the
stability loss at high speed, measurement of
the restoring moments is usually carried out.
In fully captured or semi-captured model tests,
the attitude of the model should be adjusted to
the corresponding attitude at each advance
speed. To measure the roll restoring moments,
inclined tests are carried out. The inclined
tests are divided into two kinds which are to
measure change of inclination angle using a
model free to move in the relevant degree of
freedom, or to measure forces acting on a
model fixed in the relevant degree of freedom.
From the data-file of hydrodynamic forces
obtained for a fully captive model test with
systematically changing attitudes and speed,
(described in the section on resistance tests),
all restoring forces including cross-coupling
restoring forces for arbitrary attitude can also
be determined as well as all other forces
which are needed to simulate the steady run-
ning attitude and resistance. Multi-component
balances, multi-component loadcells, or
equivalent, have been employed. The range of
changing attitudes, rise and trim, should be
carefully selected on the base of expected at-
titude.

In order to find the causes of unstable
periodic motions of a planing hull due to dy-
namic instability, it is necessary to know all
hydrodynamic forces, static (restoring) and
dynamic (damping and added mass) forces,
acting on ships. It may be possible to simulate
the motions by computers if these hydrody-
namic forces are known. To measure dynamic
forces, forced oscillation tests are used. The
test facilities and procedures are almost the
same as that used in seakeeping test, but the
mean attitude of the model should be adjusted
to the corresponding running attitude. It
should be noted that it is very important to
know the cross-coupling restoring coefficients
between heave and pitch motions accurately.
Some dynamic forces like roll damping
can also be measured by simple free decay
tests. However the decay of motions at high
speed is usually too rapid to measure the
forces accurately. Therefore forced oscillation
tests are preferable to get accurate results.
An oblique towing test with inclined
model may provide manoeuvring information
on the model for a course keeping problem in
calm water.
4.8 Recent Experimental Activity on
WIG Ships
The principle of the WIG (Wing-In-
Ground Effect) ship is based on a phenome-
non called “Ground” effect (Greco, 1998). If
the wing approaches the ground, the lifting
force of the wing increases, and transport effi-
ciency becomes better.
Based on the ground effect a lot of new
concepts have been developed. During a peri-
od of more than 30 years of designing and
building in Russia, twelve different experi-
mental prototypes and four serial WIG ships
were designed, built and tested. As a result of
these extensive research and experimental ac-
tivities, the Russian programme successfully
designed and produced the 120 ton class Or-
29
lyonok and 400 ton class Lun (Kirillovikh,
1995).
Recently, many efforts have been made in
several countries to use WIG ships for com-
mercial purposes. A comprehensive survey on
this matter has been made by Shin (1998).
It is not possible at present to propose new
standard experimental procedures for this
kind of vessel, but a brief look at Shin’s re-
port is useful to understand the state of the art.
Limiting attention to the experimental point
of view, some interesting papers can be cited.
With the aim of understanding better the
ground effect, an investigation has been made
to determine the effect of ground proximity on
the aerodynamic characteristics of airfoils
having aspect ratio of 1.0 (Carter, 1961). The
investigation was made with the model mov-
ing over the water in a towing tank in order to
eliminate the effects of wind-tunnel walls.
The results indicated that, as the ground was
approached, the lift-drag ratio was increased.
Wind tunnel investigations have been extend-
ed to determine the effect of ground proximity
on the aerodynamic characteristics of thick
highly cambered rectangular wings with dif-
ferent aspect ratios.
More recently Fuwa et al. (1993) measured
the aerodynamic forces, moments and pres-
sure distribution in towing tank. Rectangular
wings and an inversed-delta wing with the
same cross section were tested. The cross
sectional shape of the wings was NACA-6409.
Chun et al. (1996) measured the lift and drag
coefficient variations with ground clearance
and angle of attack for NACA-6409 wings
with aspect ratios of 1.0 and 2.0. Three differ-
ent endplates were attached to the side ends of
the wings. In the first type the bottom of the
end plate was extended from the trailing edge.
In the second and third types the bottom of the
endplate was lowered from the trailing edge
by 0.05 ⋅ c and 0.1 ⋅ c , where c is the chord
length.
30
Gallington (1992) carried out experiments
in order to find the optimum distance from the
leading edge of the wing of the propulsor in
order to have the maximum lift. Four very
small jets were used. The distance derived
was the greatest required for the given entry
area and trailing edge gap. A parametric
Power-Augmented-Ram (PAR) WIG model
was tested free in heave, but fixed in pitch,
over calm water and in waves. Aspect ratios
of 0.5, 0.75, 1.0, 1.5 and 2.0 were evaluated
along with a considerable range of endplate
geometries, flap angles, fan configurations
and angles of attack. Seakeeping tests were
performed in a random sea state condition.
The linearized system of equations de-
scribing the completely free longitudinal mo-
tion of a WIG craft is of fifth order whether or
not the PAR features. The forces and mo-
ments operating on the craft are changed
drastically by inclusion of the PAR feature.
Each important stability derivative can be
found mainly from the experiments and the
remaining neglected or from the quasi-steady
pipe flow.
In order to grasp the aerodynamic charac-
teristics of PAR, measurements of lift, drag
and moments acting on a low aspect wing
were carried out in towing tank. The wing
with endplates and a pair of model propellers
was attached to the towing carriage (Naga-
matsu et al., 1994).
4.9 Recent Experimental Activity on
Hybrid Ships
4.9.1 General
Compared with conventional monohulls,
hydrofoils, ACVs, SESs and SWATHs, hy-
brid ship concepts are relatively new. A vehi-
cle having more than one source of supporting
force simultaneously over a major portion of
its operational speed envelope has been
referred to as a hybrid ship.
The hybrid ship may be considered as
having three types of lift available as a surface
ship: buoyancy, dynamic lift and powered
static lift. The supporting force of the ship’s
weight is combined in various proportions of
these forces.
In terms of model tests, recently studied
hybrid ships can be divided into three groups.
Firstly, the hydrofoil with submerged body,
where the lift is shared 50% by buoyancy and
the remainder by lift force of hydrofoil. The
second group is the hydrofoil catamaran
where the lift force of attached hydrofoil is
about 20-30% of the displacement. The third
group is the semi-submerged vehicle with
negative lift force by hydrofoil. Although this
semi-submerged vehicle is only at the con-
ceptual design stage, there have been many
experiments in towing tanks.
A comprehensive survey of hybrid ships,
which considers both model and full-scale
experiments, has been made for the Commit-
tee by Shin (1998).
4.9.2 Half Buoyancy and Lift by Hydrofoil
4.9.2.1 Monohull, Submerged Lower
Hull and Submerged Hydrofoil
The fully submerged buoyancy and the
dynamic lift obtained by the fully submerged
foils are introduced and studied by Yamanaka
et al. (1991, 1995), and Ogiwara et al. (1993).
This hybrid hull consists of a main mono hull,
a fully submerged lower hull, fully submerged
foils and struts.
Under the parameters of targeted perfor-
mance characteristics, the most advantageous
combination of buoyancy to lift from the point
of ship resistance is found to be the ratio of
50:50. The seaworthiness of the concept has
been analysed and examined with various
model experiments to measure the hydrody-
namic characteristics, in particular the added
wave resistance and motions in waves, using a
captured model and free model installed with
an automatic motion control system.
A seagoing test with an experimental ship
has been carried out. The test ship is a 1/6-

scale model. The LOA of the experimental
ship is 17.1 m with a maximum speed of 41
knots. Using the experimental ship, ship re-
sistance tests in calm and rough seas, sea-
worthiness tests, manoeuvrability tests,
structural response tests and automatic control
tests were performed.
Yamanaka et al. (1991) carried out model
tests. Resistance, lift and trim-moment acting
on the model were measured by two compo-
nent strain gauges. The resistance, lift and
moment acting on the forward and aft foils
were measured by three component strain
gauges. Flaps on the submerged foils are
driven by harmonic-drive mechanism in-
stalled inside the submerged lower hull, and
hinge moments acting on the flaps were
measured by a strain gauge type torsion sensor
incorporated in the harmonic drive mecha-
nism. Tests included a resistance test in which
the model draught was varied from hull-
borne to foilborne, a hydrodynamic force test
in which trim angle was varied and a control
force test in which the flap angles of both
foils were varied separately.
The measured lift of the aft foil assembled
on the whole model has the same tendency as
the whole model lift. Lift reduction of the aft
foil takes place due to three causes:
1. Effect of wave generated by submerged
lower hull;
2. Wave generated by forward foil;
3. Downwash generated by forward foil.
Tests for motion response (Ogiwara et al.,
1993) in waves were performed using a small
model without flap control in oblique waves
and a 1/20 scale model fitted with the motion
control system at take-off and foilborne con-
dition in bow waves. The hydrodynamic pres-
sure was measured with forty-five pressure
gauges on the surface of model.
The model experiments were extended to
measure the added resistance in irregular head
waves. The motion control system was fitted
in the model.
31
The ships’ safe landing capability was
checked in case of control system failure. The
emergency landing device sets the flap and
rudder angles to predetermined positions.
Through these tests, it is confirmed that this
type of ship can land safely and fast enough in
case of control system failure.
4.9.2.2 Twin Hull, Submerged Lower
Hull and Submerged Hydrofoil
Koh et al. (1998), Ahn et al. (1998) and
Kim et al. (1997) developed the 200 TEU
class Hybrid Super High Speed Cargo Ship.
The LOA is 80 m with maximum speed 50
knots. Half of the weight was supported by
the displacement of the submerged lower hull
and the remainder by submerged hydrofoils.
A catamaran hull with two waterjet propul-
sion systems was adopted for upper hull. A
1/8th scale experimental ship was built and
sea trial tests were performed.
This type of ship has a substantial insta-
bility in heave, pitch and roll modes at the
foil-borne stage due to little restoring force, so
an active control is indispensable to keep the
stability. Four hydraulic actuators with servo
valves were installed to drive the foils and
several sensors were used to measure the mo-
tion of the ship (Kim et al., 1998).
4.9.3 Hydrofoil Catamaran
4.9.3.1 Superjet
Kawaguchi et al. (1991) and Arii et al.
(1993) developed a hydrofoil catamaran. Four
series of experiments with a 1/10-scale model
were performed and the results were present-
ed for the properties of resistance, propulsion,
longitudinal stability, and dynamics and roll
control by ailerons. A 12-meter long proto-
type was constructed and served for experi-
ments at sea. Resistance tests were carried out
in a towing tank with a 1/10-scaled model.
A data acquisition system and control
32
system including the actuators was installed
on ship independent from the traditional jet
angle and propulsion control system. Fore and
aft flaps and each aileron were controlled
independently by the manual controller via an
onboard desktop computer, which was used as
an input signal generator, and data acquisition
input signals and output data are obtained
very concisely. Inertial measurement system is
placed at the centre of gravity. Hydraulic and
servo actuator systems are made on the re-
quirements based on the pitch/roll motion
natural frequencies.
The basic test used a two-dimensional hy-
drofoil in a cavitation tank and the actual test
used a catamaran with submerged hydrofoils.
The aspect ratio of the hydrofoil was about
seven.
Arii et al. (1993) performed motion tests
and simulation studies in condition of regular
head seas and regular following seas. Studies
were conducted at a speed of 34 knots and
wave height of 1.0 m. The simulated and ex-
perimental results were in good agreement.
The pitch and heave transfer functions are
very small in head seas but are larger in fol-
lowing seas. These differences are due to the
greater changes in angle of attack in following
seas.
Comparison of motion characteristics
between the developed hydrofoil catamaran
and conventional high-speed craft were per-
formed.
The cavitation performance of the hydro-
foil was investigated using 1/50-hydrofoil
models and a 1/6-scale sea going experimen-
tal ship. The scale effect on the cavitation per-
formance was investigated over the range of
Reynolds number 9.4 × 10 5 to 2.4 × 10 6 . The
inception and growth of cavitation on the hy-
drofoil of the ship was observed at Reynolds
number 3.1×10 7 by small TV-cameras in-
stalled in her struts.
4.9.3.2 Hydrofoil Catamaran
Special model tests in a wind tunnel,
towing tank and cavitation tunnels were per-
formed in order to develop the hydrofoil
catamaran (Minsaas, 1993). Foils and propul-
sors were tested in the free surface tunnel and
in the cavitation tunnel. Drag and lift were
measured and the extent of cavitation studied.
Special tests were made to study the
foil/propulsor interference and the influence
of immersion on lift and propulsor ventilation.
Seven different winglets were tested and fi-
nally an improvement of the order of 7.5 to 10
percent was achieved at 40 knots.
4.9.3.3 Hypercat
Shimizu et al. (1993) studied and devel-
oped a hydrofoil catamaran (Hypercat). The
design including the propulsive performance,
seakeeping performance and cavitation char-
acteristics were studied.
Radio controlled model tests were carried
out. The model was 1.8 m long and self-
propelled by waterjet thrusters. Pitch and roll
motions were measured with clinometers and
rate sensors while the data were transmitted
onshore by telemeter for recording. The tests
were carried out in a 80 x 90 m square basin.
Cavitation tests with the hydrofoil at-
tached hydrofoil catamaran were carried out.
The model span was 400 mm and the chord
length was 140 mm.
4.9.3.4 Foil Assisted Catamaran
Kim et al. (1993) developed the foil-
assisted catamaran. Two hull forms with pas-
sive hydrofoil, which is attached at LCG to
reduce the resistance and slender long bow to
minimise the longitudinal motions, were
tested. The resistance test results show that
the EHP/∆ of the catamaran with foil is about
18% less than that of a normal catamaran.
Model tests and full scale motion meas-
urements were performed for a foil-assisted
catamaran passenger ship of 42 m (Yum et al.,
1995). The full-scale measurements were car-
ried out for three hull-foil conditions consist-

ing of bare hull condition and foil
with/without flap control conditions. The ef-
fects of flap control in reducing motion am-
plitudes were very well confirmed during full-
scale measurements.
4.9.4 Buoyancy and Downward Lift by Hy-
drofoil
A semi-submersible displacement type
high-speed ship with wings is proposed by
Mori et al. (1991, 1993, and 1995). The high
speed is realised by submerging the main hull
by downward lifting force generated by the
wing. The submersion causes the wavemaking
resistance to be remarkably reduced. It is con-
firmed that all the motions of the new dis-
placement-type high-speed ship can be per-
fectly controlled by flaps, ailerons and eleva-
tors.
A 2 m-long self-propelled model was used
to carry out free running experiments. The
model ship had twin propellers and a free-
surface piercing strut through which wires
were connected to the power supplies and
computers on the towing carriage. The active
control of flaps and tail wing was carried out
by changing their angles. Control was propor-
tional to the sum of angle and angular velocity
of rolling and pitching motions. Two cases,
with and without control of the submergence
depth, were tested.
5 GENERAL TECHNICAL CON-
CLUSIONS
The following conclusions are drawn from
a questionnaire and published literature relat-
ing to the various aspects of model testing of
HSMVs:
• Air resistance can have significant effects
during tank tests at high speeds. Given the
difference in physical characteristics of
each facility it is not possible to identify a
single testing method that will provide
identical results.
• Appendage drag can represent a significant
proportion of the total resistance of
33
HSMVs and, as such, reliable estimates of
its value are important. Two methods are
commonly used to account for appendage
effects: 1) testing the bare hull and then ac-
counting for lift and drag of individual
components using analytical methods; 2)
testing the hull with and without ap-
pendages and expanding the values based
on the local Reynolds number of each
component.
• Different methods are used for deriving
wetted surface area. A distinction is com-
monly made between the area covered by
spray and that covered by solid water. It is
common practice to disregard the viscous
drag of spray-covered areas and to account
only for the viscous drag of the area wetted
by solid flow.
• There are wide variations in the extrapola-
tion methods employed, although some or-
ganisations were using the modified ITTC
‘78 method in parallel with their own
methods.
• Few organisations incorporate a form fac-
tor in their extrapolation process. It is con-
sidered that more work is required before
reliable form factors could be applied to
HSMVs in a routine commercial context.
Little formal model-ship correlation ap-
pears to be applied to HSMVs, generally
due to the lack of reliable full-scale data.
• A reliable method for estimating the scale
effects of resistance due to spray has not
yet been developed.
• A promising method using captive model
force measurements has been developed
for HSMVs, although the time involved for
such tests can be large compared with con-
ventional resistance tests.
• In seakeeping tests, measurements of the
encountered waves are more desirable
than measurements of the stationary waves.
Surface piercing wave probes can be unre-
liable because of water run-up on the for-
ward side of the probe and ventilation on
the back side of the probe. Good success
has been reported using electromechanical
34
servo probes, acoustic probes, and servo
controlled wave gauges.
• Measurement of motions is important and
should in general be non-intrusive. If tests
in head or following seas are carried out
and the model is restrained in the lateral
motions, a reliable measurement is ob-
tained by having a wire-over-potentiometer
measurement. Reliable 6 degrees-of-
freedom optical systems are available for
free running tests.
• The presence of swell is for most high
speed vessels a very important factor. The
motions due to swell can be much higher
and therefore more relevant for both oper-
ability and safety aspects.
• Motions responses of HSMVs can be non-
linear to a significant degree. This means
that the motion RAO derived from experi-
ments is not unique and cannot be used to
determine the performance in irregular seas
with confidence.
• Tests in irregular seas are required to check
the normal operational limits. This makes
computer predictions for such conditions
quite important since it is generally not
feasible to carry out tests in many sea
states.
• The resulting trim of a HSMV in the
presence of cavitation is different com-
pared with that measured during tests
without these phenomena. It means that re-
sults from model self-propulsion tests in a
towing tank at atmospheric pressure are in
most cases not representative of full-scale
HSMVs.
• The manoeuvrability of HSMVs is non-
linear and can be significantly affected by
ventilation and cavitation phenomena
when they occur. Free-running or captive
model tests are still the most commonly
used prediction method for manoeuvrabil-
ity. The way in which ventilation, cavita-
tion and non-linearities have been taken
care of at the evaluation of model tests
should be documented in the test report.
• For measurement of global load hull forces
the most practical way is to use a segment-
ed model where the segments are connect-
ed through force transducers.
• The lift and drag of the aft hydrofoil are af-
fected by the fore foil, hull and submer-
gence depth. The interactions among them
should be investigated more precisely. It is
preferable to measure the lift and drag of
the hydrofoil simultaneously with the hull.
• The correlation of hydrofoils should be
considered. The effect of Reynolds number
can lead to a discrepancy in lift and drag.
Furthermore, the effect of cavitation in-
ception of hydrofoils due to Reynolds
number and depth effect should be investi-
gated.
• WIG effects have been established by
means of model tests in towing tanks and
wind tunnels. There appears to be an effect
of the bottom boundary condition which
leads to different results from these facili-
ties.
• The seakeeping quality of the WIG ship
should be considered carefully. Transport
efficiency in rough water with
H 1 / 3 = 0.1 ⋅ c ( c denotes chord length) is
half that in calm water.
• The dynamic behaviour of the Wing In
Ground (WIG) ship is important to ensure
safety. All the static derivatives associated
with the propulsor-wing combination can
be extracted from the experimental data or
theoretical analysis. More importance
should be given to experimental methods
in this field.
• The ships’ safety should be checked for the
case of control system failure. If the ship
does not have natural stability, control
system failure could lead to hazardous
conditions.
6 RECOMMENDATION TO THE
CONFERENCE

The Committee recommends that the Con-
ference adopts the procedures outlined in the
present report and summarised as follows:
• In making air resistance corrections it is
important to measure the actual air speed
beneath the carriage, in the area that the
model will be tested. Air speed measure-
ments should be made over the speed range
of interest with the carriage configured as
it will be when tests are conducted. The re-
commended method of accounting for
aerodynamic effects on trim is to calculate
the difference in bow-up or bow-down
moment between the model and full-scale
vehicle by assuming a centre of aerody-
namic pressure and hydrodynamic pressure.
The difference in moment should be ap-
plied as a shift in the centre of gravity
(paragraph 4.2.1).
• If it is decided that air resistance is insig-
nificant for a particular HSMV model test,
the justification for that decision should be
documented in the test report (paragraph
4.2.1).
• To take account of the effect of append-
ages, testing both with and without ap-
pendages is recommended. If the purely
analytical method is used, trim moments
caused by appendage forces should be ac-
counted for using equivalent shifts in cen-
tre of gravity location and displacement. If
these corrections are made after the tests
are completed, the results can be obtained
by interpolating between results from tests
with different displacements and centre of
gravity locations (paragraph 4.2.2).
• Boundary layer turbulence stimulation is
recommended when the Reynolds number
is less than 5 ⋅ 10 6 based on hull length
(paragraph 4.2.3).
• For many HSMVs the forefoot enters the
water at a location that varies with speed.
Therefore, when turbulence stimulation is
required for these vehicles it is recom-
mended that a method is used that does not
involve devices at a fixed location on the
35
hull. For example, a strut piercing the wa-
ter surface ahead of the model as described
by Savitsky & Ross (1952) is recom-
mended (paragraph 4.2.3).
• The Committee recommends that running
wetted surface area should be used for
HSMVs instead of static area. The method
of measurement and likely level of accura-
cy should be described and defined in test
reports (paragraph 4.2.4).
• It is important to include a motion control
system in the set-up of seakeeping tests
whenever possible (paragraph 4.3.3).
• In making full-scale evaluations starting
from model results, an analysis procedure
that uses open water propeller characteris-
tics measured in oblique flow conditions
should be adopted (paragraph 4.4.3).
• The Committee recommends that for de-
tection of slamming on models, one should
measure forces on a suitably sized panel
rather than measure pressure directly. For
flat-bottom slamming and slamming on
wet-decks the measurement panel must
have scaled mass and stiffness in order to
detect correctly the slamming force (para-
graph 4.6.1).
7 RECOMMENDATIONS FOR FU-
TURE WORK
• Correlation of manoeuvring between
model and ship should be investigated
more precisely. Although the resistance
performance shows reasonable agreement,
there is room for improvement of manoeu-
vring coefficients. Combination of gain for
the control system and quantitative deter-
mination of ride quality are important and
difficult. It is recommended that the ma-
noeuvring parameters be determined more
precisely by a sea trial test using an ex-
perimental ship.
• Specific methods of turbulence stimulation
suitable to be applied to small specific ap-
pendages should be reviewed.
36
• At present there is no accepted method
available to account for scale effects in re-
sistance attributable to spray. This is an
area that needs to be addressed.
• More studies and experimental work
should be carried out to define a reliable
method for obtaining (1 + k ) suitable for
HSMVs.
• A more formal and unified approach to
model-ship correlation for HSMVs should
be established. To facilitate this aim it
would recommend the establishment by
ITTC of a data base of trials data for
HSMVs.
• It is recommended that the Conference in-
vestigates the usefulness and the accuracy
of the test method based on captured force
measurements and the automatically con-
trolled running attitude of high speed ships
(paragraph 4.2.8.3).
• Studies should be carried out to develop
more accurate scaling procedures for wake
and propeller efficiency.
• Establish a rational mathematical model
for manoeuvring motions for high speed
craft, and experimental procedures to ob-
tain hydrodynamic derivatives for the
model.
• Measurements of hydrodynamic forces
acting on a moving ship due to instability
should be carried out to clarify the mecha-
nism of such motion occurrence.
• For WIGs, correlation for resistance,
seakeeping behaviour and power augment-
ed ram (PAR) between the model test and
the actual ship should be studied. Added
resistance in waves should be investigated
more precisely.
• More precise analysis of dynamic behav-
iour of WIGs to ensure longitudinal stabil-
ity should be studied, together with transit
mode behaviour during take-off and land-
ing.
• The hydrodynamic features of hybrid hull
forms are so complicated that the correla-
tion method between model test and full-
scale ship performance should be estab-
lished. Important items for consideration
are as follows:
i. Correlation of manoeuvring parameters
between towing tank tests, free running
tests by experimental ship and full-
scale ship.
ii. Reynolds number effect on hydrofoils
for lift and drag force, and cavitation.
iii. Correlation between towing tank,
cavitation tank and full-scale ship.
iv. The effect of submergence depth for
lift, drag and cavitation of hydrofoils.
8 ACKNOWLEDGEMENTS
The Committee acknowledges the co-
operation of the members organisations who
supported its activity through the answers to
the Questionnaire, expressing views and
providing relevant material.
9 REFERENCES
Ahn, J.-W., Lee, C.-Y., Park, Y.-H., Chung,
J.-A., and Kim, B.-H., 1998, “Performance
Test and Model-Ship Correlation for a
Waterjet Propulsion System,” Trans. of
the Society of Naval Architects of Korea,
Vol. 35, No. 4, Korea, pp. 11-18 (in Ko-
rean).
Arii, T., Yamato, H., Takai, T., and Shigehiro,
R., 1993, “Development of a Motion
Control System for a Foil-Assisted Cata-
maran “'Superjet-30”,” FAST’93, Japan,
pp. 189-200.
Blount, D., and Fox, D., 1976, “Small Craft
Power Prediction,” Marine Technology,
SNAME, Vol. 14, New York, USA.
British Shipbuilding Research Association,
1958, “Experiments on Marine Propeller
Blade Sections,” part III, London, United
Kingdom.
Bulgarelli, U.P., et al., 1999, “Advanced
Aero-Hydrodynamics of WISES”, 5th
HSMV, Capri, Italy.

Carcaterra, A., and Ciappi, E., 1998. “Predic-
tion of the Compressible Stage Slamming
Force on Rigid and Elastic Systems Im-
pacting on the Water Surface,” Proc. of
In.l Conf. on Nonlinear Dynamics,
Blacksburg, Virginia.
Carcaterra, A., et al., 1999, “Slamming of
elastic structure over the Water Surface: a
Shock Spectral Approach,” ISOPE’99,
Brest, France.
Carter, A.W., 1961, “Effect of Ground
Proximity on the Aerodynamic Character-
istics of Aspect-Ratio-Airfoils with and
without End Plates,” NASA TN D-970,
USA.
Cassella, P., Coppola, C., Lalli, F., Pensa, C.,
Scamardella, A., and Zotti, I., 1998, “Geo-
sim experimental results of high speed
catamaran: co-operative investigation on
resistance model tests methodology and on
ship-model correlation,” 7th PRADS, The
Hague, Netherlands.
Chun, H.H., Park, I.R., and Chung, K.H.,
1996, “Computational and Experimental
Studies on Wings in Ground Effect and A
WIG Effect Craft,” Proc. of Ekranoplans
& Very Fast Craft, Australia, pp. 38-59.
Clement, E., 1957, “Scale Effect on the Drag
of a Typical Set of Planing-Boat Append-
ages,” DTMB Report No. 1165.
Cordier, S. and Durmez, F., 1993, “Scale ef-
fects on the resistance components of a
high speed semi-displacement craft,”
FAST’93, Yokohama, Vol. 1, pp. 409-419.
Couser, P.R., Molland, A.F., Armstrong, N.A.
and Utama, I.K.A.P., 1997, “Calm water
powering predictions for high speed cata-
marans,” FAST’97, Sydney, pp765-772.
Faltinsen, O.M., 1998, ”Water Entry of a
Wedge by Hydroelastic Orthotropic Plate
Theory,” Hydroelasticity’98, Fukuoka, Ja-
pan.
37
Fridsma, G., 1971, “A Systematic Study of
the Rough-Water Performance of Planing
Hulls, Irregular Waves, Part II,” Davidson
Laboratory, Stevens Institute of Technol-
ogy, Report 1495.
Fuwa, T., Hirata, N., Hasegawa, J., and Hori,
T., 1993, “Fundamental Study on Safety
Evaluation of Wing-In-Surface Effect Ship
(WISES),” FAST’93, Japan, pp. 1257-
1267.
Gallington, R.W., 1992, “Power Augmenta-
tion of Wing In Ground Effect Craft,”
Proc. of the Intersociety High Performance
Marine Vehicle Conference and Exhibit,
pp. ws9-ws16.
Gollwitzer, R., and Peterson, R., 1996, “Drop
Tests and Planing Boat Dynamics Model-
ling for Investigating Repeated Water-
Entry Shock and Mitigation,” Small Craft
Marine Engineering Resistance and Pro-
pulsion Symposium, University of Michi-
gan, USA.
Granville, P.S., 1976, “Elements of the Drag
of Underwater Bodies,” DTNSRDC, Re-
port No. SPD 672-01, Bethesda, MD.
Greco, M., et al., 1998, “Simplified Model-
ling of the Aero-Hydrodynamics of ‘Ekra-
noplanes’,” Schiffstechnik Bd. 45, Ship
Technology Research, Vol. 45, Hamburg.
Hadler, J., 1966, “The Prediction of Power
Performance on Planing Craft,” Trans.,
SNAME, Vol. 74, New York, USA.
Hirayama, T., Motohiro, H., and Ryuichi, I.,
1988, “On the Accuracy of Seakeeping
Tank Tests and on a Project of Compara-
tive Seakeeping Experiment,” Symposium
on Ship Hull Forms and Seakeeping
Qualities, Society of Naval Architects of
Japan, Japan.
Hoerner, S. F., and Borst, H. V., 1985, “Fluid
Dynamic Lift,” Published by Author.
Hoerner, S. F., 1965, “Fluid Dynamic Drag,”
Published by the Author.
38
Hoyt, J., and Dipper, M., 1989, “A Matrix
Data Base Approach to Planing Craft Re-
sistance Model Experiments,” 22nd ATTC,
St.John’s, Newfoundland.
Ibaragi, H., et al., 1996, “Study on the Trans-
verse Instability of a High-Speed Craft,”
Trans. West-Japan Society of N. A., No.
91, Japan.
Ikeda, Y., et al., 1992, “Hydrodynamic Forces
Acting on a High-Speed Craft Running
with Constant Speed,” Jour. of Kansai Soc.
N. A., No. 218, Japan.
Ikeda, Y., 1993, “Simulation of Running At-
titude and Resistance of High-Speed Craft
Using a Database of Hydrodynamic
Forces,” Japan.
Ikeda, Y., and Katayama, T., 1997, “Stability
of High Speed Craft,” Proc. of 3rd Int.l
Workshop on Theoretical Advances in
Ship Stability and Practical Impact,
Greece.
Kapsenberg, G. K., 1994, “Added Mass and
Damping Coefficients for a Large SES –
Including an Appreciation of Scale Ef-
fects,” NAV’94 Conference, Rome, Italy.
Kapsenberg, G.K., and Blume, P., 1995,
“Model Tests for a Large Surface Effect
Ship at Different Scale Ratio’s,” FAST’95,
Lübeck-Travemünde, Germany.
Katayama, T., and Ikeda, Y., 1995, “An Ex-
perimental Study on Transverse Stability
Loss of Planing Craft at High Speed in
Calm Water,” Jour. Kansai Soc. N. A., No.
224, Japan.
Katayama, T., and Ikeda, Y., 1995, “A Study
on Transverse Instability of Planing Craft
at High Speed in Calm Water,” Proc. KO-
JAM’96, Korea.
Katayama, T., and Ikeda, Y., 1996, “A Study
on Unstable Rolling Induced by Pitching
of Planing Craft at High Advance Speed,”
Jour. Kansai Soc. N. A., No. 255, Japan.
Katayama, T., and Ikeda, Y., 1997, “A Study
on Coupled Pitch and Heave Porpoising
Instability of Planing Craft,” Jour. Kansai
Soc. N. A., Nos. 226, 227, 228, Japan.
Katayama, T., and Ikeda, Y., 1998, “Hydro-
dynamic Forces Acting on Porpoising
Craft at High-Speed,” Proc. of 3rd Int.
Conf. on Hydrodynamics, Korea.
Kawaguchi, H., Miyata, H., Yamato, H., and
Takai, T., 1991, “Full-Scale Experiments
by the First Hydrofoil Catamaran WING-
STAR12 “Exceller”,” FAST’91, Norway,
pp. 1195-1214.
Kim, B.-S., Yoo, S.-Y., Son, Y.-T., Shin, M.-
S., Koh, C.-D., and Yang, S.-I., 1993,
“Improvement of Hydrodynamic Charac-
teristics of Catamaran with Hydrofoil,”
FAST’93, Japan, pp. 631-642.
Kim, J. W., Kim, Y. G., Lee, C. J., and Lee, C.
Y., 1998, “Development of Motion Con-
trol Techniques and Sea Trials of the Test
Ship “NARAE”,” Trans. of the Society of
Naval Architects of Korea, Vol. 35, No. 3,
Korea, pp. 26-37 (in Korean).
Kim, K.-S., Song, I.-H., Ahn, J.-W., and
Moon, I.-S., 1997, “Performance Test of
Pod-Type Waterjet Propulsion System,”
Trans. of the Society of Naval Architects
of Korea, Vol. 34, No. 4, Korea, pp. 21-30
(in Korean).
Kirkman, K., and Kloetzli, J., 1981, “Scaling
Problems of Model Appendages,” 19th
ATTC, Ann Arbor, Michigan, USA.
Kirillovikh, V. N., 1995, “Russian Ekrano-
plans,” Proc. of 21st Century Flying Ships,
Australia, pp. 71-117.
Koh, C.-D., Cho, Y.-J., Shin, M.-S., and Yang,
S.-I., 1998, “Development of Hybrid Su-
per High Speed Cargo Ship “NARAE”,”
Proc. of Workshop on High Speed Marine
Vehicles, Korea, pp. 157-167 (in Korean).
Koushan, K., 1997, “Beitrag zum Kanalein-
fluss bei Tragfl gelversuchen,” Dr. Ing.
 39
Thesis D83, Technical University of Ber- Mori, K.-H., Doi, Y., Kobayashi, M., and
lin, Germany. Shimada, H., 1993, “High Speed Semi-
Submersible Vehicle with Wing-
Kvålsvold, J., 1994, “Hydroelastic Modelling
Hydrodynamic Characteristics and Free-
of Wetdeck Slamming on Multi-hull Ves-
Running Experiment,” FAST’93, Japan,
sels”, Dr.ing Thesis, Dept. of Marine Hy-
pp. 1257-1267.
drodynamics, Norwegian Institute of
Technology, Trondheim, Norway. Mori, K.-H., Ninomiya, S., Doi, Y., Nagaya,
S., and Fujii, Y., 1995, “Free Running
Kvålsvold, J., Faltinsen, O.M., and Aarsnes,
Turning Tests of a New Displacement-
J.V., 1995, “Effect of Structural Elasticity
Type High Speed Semi-Submersible Ship
on Slaming Against Wetdecks of Multi-
with Wings,” FAST’95, Germany, pp.
hull Vessels,” PRADS’95, Korea.
597-596.
Kvålsvold, J., Faltinsen, O.M., and Aarsnes,
Moulijn, J.C., 1998, “Scaling of Air Cushion
J.V., 1996, “Wave Impact on a Horizontal
Dynamics,” Delft University of Technol-
Elastic Plate,” Jour. of Marine Science and
ogy, Ship Hydromechanics Laboratory,
Technology.
Report No. 1151, The Netherlands.
Lackenby, H., 1955, “BSRA Resistance Ex-
Mørch, H. J. B., 1992, “Aspects of Hydrofoil
periments on the Lucy Ashton, Part III –
Design; With Emphasis on Hydrofoil In-
The Ship-Model Correlation for the
teraction in Calm Water,” Dr. Ing Thesis,
Smart-Appendage Conditions,” Trans. of
Division of Marine Hydrodynamics, The
the RINA.
Norwegian Institute of Technology, Nor-
Landrini, M., et al., 1999, “Numerical Simu- way.
lation of the Unsteady Flow Past a Hydro-
Mørch, H.J.B., and Minsaas, K.J., 1991, “As-
foil,” Schiffstechnik Bd. 46, Ship Tech-
pects of Hydrofoil Design; With Emphasis
nology Research, Vol. 46, Hamburg.
on Hydrofoil Interaction in Calm Water,”
Lewandowski, E., 1989, “The Effects of Rey- FAST’91, Trondheim, Norway.
nolds Number, Section Shape and Turbu-
Murakami, T., 1981, “Measurement of Wave
lence Stimulation on the Lift of a Series of
Profile Generated by the Towing Car-
Model Control Surfaces,” 22nd ATTC, St.
riage,” Vol. II, 16th ITTC, pp. 272, Lenin-
John’s, Newfoundland.
grad, USSR.
Masset, J.F., Morel, J.P., and Kapsenberg,
Nagamatsu, T., and Kure, F., 1994, “Experi-
G.K., 1995, “Large Surface Effect Ship
mental Study on Aerodynamic Character-
(SES) Air Cushion Dynamics: An Innova-
istics of PAR-WIG,” Jour. of Kansai So-
tive Methodology for Theoretical Model-
ciety of Naval Architects of Japan, No.
ling Validation,” FAST’95, Lübeck-
222, Japan, pp. 49-56 (in Japanese).
Travemünde, Germany.
Ogiwara, R., Yamanaka, N., Kobayashi, K.,
Minsaas, K.J., 1993, “Design and Develop-
and Moriyama, A., 1993, “A Submerged
ment of Hydrofoil Catamarans in Nor-
Hull and Foil Hybrid Super High Speed
way,” FAST’93, Yokohama, Japan, pp.
Liner,” FAST’93, Japan, pp. 189-200.
83-99.
Pierce, R., 1992, “Run Length and Statistical
Mori, K.-H., and Yasuaki, D., 1991, “Perfor-
Error Estimation for Seakeeping Tests and
mance and Dynamic Stability of a High
Trials,” 23rd ATTC, New Orleans, Louisi-
Speed Semi-submersible Vehicle with
Wings,” FAST’91, Norway, pp. 915-929. ana.
40
Savitsky, D., and Ross, E., 1952, “Turbulence Yamanaka, N., Tomita, M., Yamagami, Y.,
Stimulation in the Boundary Layer of Itoko, T., Kohono, Y., and Hamamatu, M.,
Planing Surfaces,” Davidson Laboratory, 1995, “The Seagoing Test Ship to Verify
Stevens Institute of Technology, Report the Technologies Developed for a Sub-
No. 444. merged Hull and Foil Hybrid Super-High-
Savitsky, D., et al., 1981, “Status of Hydrody- Speed Liner,” FAST’95, Germany, pp.
namic Technology as Related to Model 691-701.
Tests of High-Speed Marine Vehicles,” Yokomizo, K, et al., 1992, “Simulation of
HSMV Panel 16th ITTC, DTNSRDC, Re- Running Attitude and Resistance of a
port No. 81/026, Bethesda, MD,USA. High-Speed Craft Using Database of
Three-Component Hydrodynamic Forces,”
Schleicher, C., 1997, “Investigation of a Hy-
Jour. of Kansai Soc. N. A., No. 218, Japan.
brid Wave Piercing Planing Hull Form,”
FAST’97, Sydney, Australia. Yum, D.-J., Min, K.-S., Song, K.-J., and Lee,
H.-Y., 1995, “Theoretical Prediction of
Shimizu, K., Masuyama, K., Fukushima, M.,
Seakeeping Performance and Comparison
Takashima, J., and Ishii, N., 1993, “A
with Sea Trial Results for High-Speed Foil
Study on the Hydrodynamics Aspects of
Catamaran Ship,” FAST’95, Germany, pp.
Hybrid Hydrofoil Catamaran,” FAST’93,
1053-1063.
Japan, pp. 951-962.
Zseleczky, J, and McKee, G., 1989, “Analysis
Shin, M.-S., 1998, “Survey Report of WIG
Methods for Evaluating Motions and Ac-
Ship,” KRISO Report.
celerations of Planing Boats in Waves,”
Shin, M.-S., 1998, “Survey Report of Hybrid 22nd ATTC, St. John’s Newfoundland,
Ship,” KRISO Report. Canada.
Steen, S., 1998, “Analysis of a Questionnaire
on Current Practice for Testing of
HSMV,” Report from 22nd ITTC Spe-
cialist Committee for Model Tests of High
Speed Marine Vehicles, MARINTEK,
Trondheim, Norway.
Tanaka, H., Nakatake, K., Araki, S., Nakato,
M. and Ueda, T., 1991, “Co-operative re-
sistance tests with geosim models of a
high-speed semi-displacement craft,” So-
ciety of Naval Architects of Japan, vol.
169.
Washio, Y. et al., 1993, “On the Improvement
of Transverse Stability for High-Speed
Craft,” Trans. West-Japan Society of N.A.,
No. 86, Japan.
Yamanaka, N., Osamu, Y., Ryuichi, S., To-
shihiko, N., Toshihiko, A., and Takeshi, F.,
1991, “A Submerged Hull and Foil Hybrid
Super-High Speed Liner,” FAST’92,
Norway, pp. 163-178.



 


 



 

   

 


 




  % "
 1  $
 $  


  
     
  $$   $  $
   "  
 
  
 
  
 ! 
     
 4
    
  
  
 

     !
6
 

     
$$ $
$


   "  


 
 $$ 
 $  $ 
       $


  "
      
   

  !
     
  !
6


"!  #

$ $ 
$
 
$!"     $ "     $ 

    
  7 

% "
 & $  "   
 '

 $$ !      
()*+,! $" 
""


!


  " 
 $   
$$  $ %
""
 $
  "  "     

$$ $-" 
""
 ""   
 $   
" $    $  
  $  
   / -2

 

""

 $

$   
 8" 
$$
  
""
  "  $  $   / -2

 4
$ -
  .  ! 
""
  #!     $
  
 


 #! 
     
  
 $  
    
  $  


  !

 
        !  
!

     




  
 "  
  $ ! 
)!  !   

    )!    
7 


 $ 
"  $ 
   /
0&12
  ! "
" 
) /
03&2    

 $
 "'
"  $  
 $ $$  !  / -2

      $  $
  $$    !  
4!' "   

 $  
  



""
        
!  
 #!    !
   
 $  $  
   
 $
"


 !  $!
 
 
$
$""5
$ 
9

 !  "    "



 !
# 

" $ $ 
$      $!#

       ! $   



$ !'

"    $
 
 $ "  
$

2   "   $ $
  

 $
  
$$  $ 
 

  !
 "


   "


   
 $ 
/>2=



   "  -"
 
  



"?6@$

 $  "
   
 $ 
 $
    
  "  

  
    
 
 "   
 2 
   !  !
  
$ !

4' 
 ' 693&A6
/=2 $
69&B=B
6
  
   
! !  
 $  

 
    $
 '

 !   
   $ 6?9&A6
&/=2$69&B=B
6
" 5
  ! 
   



  "     
 $ 
   
  
"   $ 
 $ 
     
   
!    $  $   
!
 "
 
 
  
 

" 
    8"  
 $$
"
   ! 
   
 
 $    "   4

 
$
   
 $$  $
 $


     
     )    !  
# ! $



 
  
 "$ 
 
  "  $$   
  $  $ $$  
 $ - 



 "

  
     
    $   !    
 !
           

 $$


   
   
 $  $  $$
/
A#2
 


!
 

$   
  
"""
 
 /
A#2C
! 
   " $   
6$ 
 
#$

""
5 $$ $
 )!  


   $  

 !
 
  ! "  
 


  
     
      " 
 $ $

"" 
  $
 
 
 
 
  !
    
  


      $$
  
$
   
  

$
$$
5
: 
 " 

 " 

  
     
 
7 "

 
 
  
 
 
 

  $ 
 
# 
# $
    
  

$

   
     


" $  "  

      

 D

 *" 


  8
"

 "   
 
 
$$
  8"  
# $  ,
$- "
"$ !#
  !  #  8" 
  $;<   




=
8



 :

!
   -"
 " "

 $  
  !  $$ $
 E

  $ -  " $ )*+  !  # 

# !  
!    
  $  
   
 $
   " 


  
  $  
 $     "
 $ 
$$!!
!
 
J

! 

   ! 
#      "  
   $
! 

    16@ $  
    "  
  
" $
#/
 2!

 8/ 

 2

!  " " 
  /
! !
 "
 !$ $
!
F2 $  
   #
 $ )* 
%
)*+
 
   "
 

$$$!


 
  
     
 
 ! 


 ! 
 
 "  
  
 
 
 

     
  


 "  * K;; L
 ,M

$ 
# !#  $$ $  ! 



$  $)*+
 !  #  8" 
""

   $  8  !# 

  
     
  
 "$ 
  !  # 
# 
 
  
 
  
$$ !"'
    
  


2 $
$$ $
G
 H)*+  


 


 
8 7  $
61  4  $ $ 

 
 "
 
  !
""

)*+   
 $  $ 
    $ 
#" 
"
 
 $ 
! "  


!8"!  
$    "   $
%"
&39 $
 /
  

$
2)! 

  $ "


     
   

" $ $$
$$   
    
$
  
$ 
 
""

 /
 J  $   $
"N2  $  !
! 
!
$$  

$"
$  
 

  

 " 
 
 


   4  
""





  
    
 

         !

  
     
    " 
 
  
 
 
 !
  $ $   4 !
    
  

"
  5
H
% I
 92 % "
 &E6 $  "  

 !  #  

#   $

 $


" "  
"
  ; ' ( 

 N 
$)*+,  "     
  

"


 
"



   
#"    
"  4
N.

  $  


 $ 8
 

  
?

 

""  
 
 "
 " 

$

 /
2 !    
   
   $P*G"

 
  $ $
   /   P*G"
 ""
!
  2
  4  
  
"
J
 ; 
 

  

"  

"$



   
 

 $
  

 
         
  
 
 




 "
"
     
  

! 
  $

  ! ! " #
" ! 
! 
  !
 $

! 
 

 
  $



  $ 
! 

  $
   
 !  #
#+

$
$ !        "$
   

   ) 
 $
;  !  $4 "  4!!! 
!4
$!
 $ $; 
   4  !  

8
"
 J$4 $ 
 $$   $  $   $
   

    "$ !

 $$ $
"  
  
  
  

 " $ $  $
"
$ ! 
""8
 $ 
 $$  $ )!  !
 !

 "

 
 
 $ 
8"$
4
!

"
! 

$ +
 
  $    $

O-O

 
  
# 
    $   
"   $ $$ $ $  
 $
 ! $   
 ! ! 
  


#    
$   



   
     4 !


         " 

E2       $ ! -

       $
(  "
 $ . $ 
 $ 
 $$  $  
  



 ""   " "  $" 

 !
    
!7"


    $ 
$ "" " " 





     8"
 !
"

  5 
     4 $ 

 $

$ ! $     $   


  
     
   
 $ "

 $
 
  
 
  
      
  
    
  

""   $  $    !
;$G
 *
*+ "      $   
 



$$
" 
"
$"
  
 
 
#-$$    

$ "$
 $         $$


"
!
  
"    
  $ 
     
  4

(   ; .   $ )!  -$ "

!

" !


   
    
  " 
 P* 

5    
 ! $ "
 
 $$  
    

)!     $    
""   %
 
    !  
   
 &

  !
 
$$ 
"


!
  ! $      $
 "

 $ $

 ! "  
$$$
 !  "    

 

 ! "   "$

   "  ""   

 "

  
  
 $

!
 
 
 $
 
 
    !  "   
$$$!""
!

  $


 
# $ 
  $

   
 $$ $  
"
$
 "" 
""
$
)!   ""
     !


 !  
 #
" '
"   
    $$
! 4


#!
 $

# $ $$   
  $ $

$$!$
 



 !   
4

!  $  $      


  " 
" ! 
#-$$
 

"$

  "      
 
 

 $$ 
$8       
  

  !  
 $

 

 $
 -$ $   !  
# 
 
  
  "  $   D

 $  "  7 $
  
#        
 $  !
 
    ! 

 
$   $  ! $  
   !  
    !
 $   $    # 

 ! !

    "
   +
 
        
 $$  "   
 ! 

4""  !
 $  $
    $ !
 
 $   
 
 
"
  $   
  $$!


 

#-$$
  $
!
#$
  
"

 "
 $ 
 
 
 $
    8" 
! -$ 
$"
 !

 



""
  +
 
   
   "  !  

 $ 
 
"

  !   $
 )*+  "
 $ 
!
       $
 $

  
 
  "


  $   !
 

    
  
 8"    $    
 !     "
!


 
#    $   

  $ $
 $ 
  -
  
 
 
"
 $ ! "  
 $
    
  

8$"
 ! %
 & 
 !# $      $  $

#H$

 

  $ $

  
  $  
  $ 
$ 
 $$ "    
  +
 

   

!   # $!
    $

$
  
 !
 
" 
$  

 $$  $   $  
            
$'

 $ !  "


  
!  
   
  
 
  -$

 "  $ $



 $$
  "
 /
1


$$ $     !   $ " ! 
  

"
! 
  
 $$  $    
  

$ 2  
  $   
    
 
  
  "   

"

 



    $
6
 $ 
 $
"
 "  "
  $
""
  )*+, !   
$  "
  
  
  $ 

 " 
 
" 
 $
!   
  $  
  
  * ?99 $  " 
         

 $$  $ $   $   

  $   


  $     !
   ! ""  
  $ 
 

  8
"     $
 $  

 
 $$$$
 $ 
$!
$
 
#-$$$
    $   
  "   $
$= 
 $  $    
  ""  

 $ $$  
   
 
  
     $

"" 
 $$  
" /$
$
 ! ""  
  $
  


 
  
    




 "$   


$ 2


 $

#
  
$  
 ""  
   



""
       
! $ 
  

  $
  

  
 
 
"  )*+ 
#" 
    
  

! 
 
 

 "

 ! & '!  (
  $  
 $$ $ 


""
     
 )! 
 !  #  
#   I
 $ 
  "   8"
  
#

  " 
 
   $ )*+    
"   
 8   
 !

 
    "  
#" 
 "     $

  
 

 $ 
""
 $

 !  " $   I

 

  
$
$$
  "  
 

 
    
"
 

$  " $ 
#"  
 

 
 
   $-
 $    
      

 
#" 8"
 " 
   "!


$$)*+,
# 
 
  
   " $

$   $  
    
"      4  
!   
      " $   $
 

%!    $ 8"  ;  ?0-6E-6?-6E   )*+Q

  $  $ "

 

"
 '
"       
 ! 
 

  
  #!  
"""

 E %

   $$
8"
 
   R
  
 
  !  $
""

 "" 

    
  
  
  
 

 ! !
 ) *! * 

  R
!
#  !   
R
 


   ;  $  " $! "



 " 
  
 $$ 
"
 " 
 
$   
 3

    "  

    $ ("
 .  ("

$ . $ 
  $
 $ "
"
$
 
*?ES 
$(;.

  
 
 

    
  

 *! %
  
 !  
# ;$ *
 $ 
"" 
  
 $$    "$
!
   
  
 
   
  "     
, "     
P*




  $ 4    

   !

"" $    ""

  
"""
 $   8
$!
#;$*


$P*G"$ "

! !