7.

5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 1 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

Table of Contents

Seakeeping Model Test Procedure for

2.9 Test Program ................................ 11
Global Loads ........................................... 2
2.10 Data Analysis ................................ 12
1. PURPOSE OF PROCEDURE .............. 2
2.11 Data Presentation ......................... 12
2. GLOBAL LOADS SEAKEEPING
EXPERIMENTS .................................... 2 2.12 Comparison to Predictions and
Numerical Simulation .................. 13
2.1 Objectives of experiment................ 2
3. PARAMETER ...................................... 13
2.2 Types and selection of global load
model type........................................ 3 3.1 Parameters to be Taken into
Account .......................................... 13
2.3 Scaling laws and scale ratio
selection ............................................ 4 3.2 Recommendation of ITTC for
Parameters .................................... 14
2.4 Model design ................................... 5
4. VALIDATION...................................... 14
2.5 Design, fabrication, and
instrumentation of structure for 4.1 Uncertainty Analysis .................... 14
segmented model ............................. 6 4.2 References...................................... 14
2.6 Model Segmentation ....................... 8 4.3 Benchmark Tests .......................... 15
2.7 Powering and Steering ................... 9
2.8 Instrumentation .............................. 9

Updated by Approved

26th ITTC Seakeeping Committee 26th ITTC

Date 05/2011 Date 09/2011

 7.5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 2 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

Seakeeping Model Test Procedure for Global Loads

Application to Extreme Loads Stochastic

Analysis
1. PURPOSE OF PROCEDURE
Fatigue Analysis and Design
This procedure will outline the measure-
Safe Operating Envelope
ment of global wave loads through seakeeping
experiments. The procedure shall describe the The objectives will have to be based upon
design of the experiment, the set-up of the the operational or design information required,
model and instrumentation, the test, and the the type and size of vessel, the wave environ-
analysis. There is already a procedure covering ment, operational variables, and the facility to
seakeeping experiments which outlines the be used. The experiment design will seek to
processes and considerations of those model satisfy the objectives within the constraints of
tests (7.5-02-07-02.1). This procedure will the physical experiment. Global loads seakeep-
elaborate and outline the additional considera- ing experiments can be rather complex depend-
tions for measurement of global loads with ent upon the degree and level to which global
various types of model and experiment designs. loads must be know.

The measurement of primary design loads,

2. GLOBAL LOADS SEAKEEPING requires less complexity than the construction
EXPERIMENTS of a model to measure slamming and whipping
loads. A model test intent on measuring slam-
2.1 Objectives of experiment ming and whipping will have to be more con-
cerned with scaling issues as a result of local
The first step in the experiment design is hydrodynamic pressures and hydro-elastic
the determination of the objectives for under- modeling.
taking an experiment to measure loads and the
available methods that are available to meet Validation of computational methods might
those objectives. Possible areas of interest that require greater fidelity and control. Validation
can be addressed by undertaking experiments of computational methods will require greater
to measure global loads are to provide data that measurement and verification of the force and
can help in the understanding of : control variables which affect the resultant
Primary Design Loads global loads, as these will need to be compared
Slamming, Whipping and Springing Loads against the modeled control forces. The inves-
tigators will need to ensure that the variables
Validation of Computational Methods used in the experiment are adequately modeled
Frequency Domain Application to Lifetime and recorded for modeling with the computa-
Designs tional methods.
 7.5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 3 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

Frequency domain oriented experiments are points within the hull using a structure inde-
intent on deriving the frequency response of pendent of the external hull. The model is
global loads relative to the seaway of concern. segmented so that it provides no continuous
The frequency response can then be used in structural support. The primary strength will be
turn to calculate more thorough load responses provided by either an elastic segmentation or
across the range of seaway and operational a rigid segmentation.
environments expected. The frequency based
response functions can also be compared to Model type is mostly only important for
derived computational solutions. whipping experiments. As the hydro-elasticity
is of greater importance there, the model types
The stochastic objective requires the pro- should either be a hydro-structural model or
duction of extreme seaways and components an elastic segmented model; a rigid segmen-
capable of appropriately modeling the extreme tation model should be avoided.
seas anticipated in nature. This will require
experiment scaling of appropriate nature, based 2.2.1 Hydro-structural model
upon the wavemaking capability, and appropri-
ate seaway modeling techniques to ensure Generally speaking, a hydro-structural
proper energy representation and random proc- model can be made to satisfy geometric simi-
ess modeling. larity of the hull form, hydrodynamic similarity
and structural similarity with regard to the
Testing of the model for the purposes of fa- global vertical bending and shearing forces,
tigue analysis requires a test matrix and ex- and hence it can be used to measure the bend-
periment design which allow suitable verifica- ing moments at any cross section over the
tion of distinct seaway and operational sectors. model length. This is a rather complex model
These results will be combined with other nu- which is difficult to manufacture and such ex-
merical and computational methods to populate periments have only been performed on a lim-
an anticipated lifetime of exposure and design ited basis. Additionally, the types of materials
environments. This, in turn, is then used to required, polymer products, are not usually
determine cumulative lifetime global loading. stable in the long term and can be subject to
structural creep. The design and fabrication of
2.2 Types and selection of global load hydro-structural models is costly and time con-
model type suming.

The global load experiment can be per- 2.2.2 Segmented models

formed with either a segmented or an elastic
body model. For an elastic body the proto- Global loads and the resultant strains are
type is representative of the full scale ship the cumulative forces applied on a part of the
down to the local structural level possible even ship due to internal weight and inertial charac-
including hull plating. A whole body structural teristics, control forces, and external hydro-
response is then obtained as function of the static and hydrodynamic forces. To quantify
hydrodynamic loading. As such the elastic these global loads at discrete locations, a model
body model is referred to as a hydro- can be divided into several independent seg-
structural model. For a segmented model ments. Two types of segmented models exist,
the global loads will be investigated at discrete
 7.5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 4 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

rigid or elastic, and depend on the type of beam 2.3 Scaling laws and scale ratio selection
or interface connecting the various segments.
The performance of global loads experi-
Rigid segmented models ments follow the same Froude scaling laws as
used for traditional seakeeping tests. The addi-
For models segmented on a rigid beam, the tional constraints are the scaling of structural
beam must have a sufficient rigidity to be con- similarity as decided upon in the experiment
sidered as infinite compared to the actual rigid- design. All structural similarity must be done
ity of the ship. The model shape does not within the confines of a geometrically and
change on wave peaks or troughs at studied structurally suitable model. The internal struc-
frequencies and the natural frequency of the tural components of the model must satisfy
structure is much greater than the wave fre- weight and volume restrictions, while trying to
quencies. Loading measurements are evaluated provide the targeted structural rigidity intended
by either measuring the effort by individual for modeling of the prototype ship.
segments or from the direct bending moments
The test facility capabilities, operational
of the beam. For the frequencies where the
environment, and test objectives are the pri-
model can be considered as rigid, results for
mary factors which will determine the scale
loading can be used as input for a numerical
ratio selection. The physical properties will be
analysis of the structure. The computations can
scaled according to the appropriate scaling fac-
either be 2D (representation of hull girder) or
tor. These scaling factors are summarized as
3D.
the first eleven entries of Table 1. Additional
scaling factors are required for structural mod-
Elastic segmented models eling and are presented in the lower portion of
Table 1.
Models segmented on a non-rigid beam al-
low for measurements at multiple locations on The structural rigidity, modulus of elasticity,
the beam, and thus a direct measurement to and section modulus all provide additional
obtain strain at all sections. Elastic segmented challenges with respect to satisfying scaling
models can also employ internal rigid structure requirements. In those cases where structural
with instrumented joints at each segment which scaling is not possible; corrections to measured
model the rigidity of the ship at each segment. strain and associated moments, torsions, and
shears might be required.
For all types of segmented models, each
segment must have the same inertial properties Other decisions as relates to scale selection
as the corresponding segment in the real ship are the weight and ballast challenges, and the
and typically the horizontal gap between the type of propulsion and maneuvering required
segments is around 5 to 10 mm. Ideally, the for the model. The model weight and ballast
neutral axis of the backbone for the induced conditions must be obtainable on a total model
moments under investigation should match or and segment level. Each segment must satisfy
be as close as possible to that of the real ship. its own weight and inertial characteristics. If it
is a self propelled model there must be suitable
scale to allow for propulsion and powering
components.
 7.5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 5 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

to a value estimated numerically from full scale

Table 1. Ideal and Practical Scaling Ratios (Din- data.
senbacher, 2010).
Ideal Practical
Quantity Prototype
Model Model
2.4 Model design
Length L L/ L/
Water Density /c /c Once the type of model is selected a design
Time t t/1/2 t/1/2
which integrates the needs within the model
Mass m m/c3 m/c3
Velocity v v/1/2 v/1/2 package is the most challenging part. Design
Acceleration a a a of an elastic ship model is beyond the scope
Force F F/c3 F/c3
Ship Displacement /c3 /c3
of this procedure. The design involves a de-
Moment M M/c4 M/c4 tailed knowledge of the ship structure to be
Pressure p p/c p/c
Frequency (flexural modes
modeled, understanding of the detailed model-
1/2 1/2
and Rigid body motions) ing and scaling laws, and the ability to design
5
Bending Rigidity EI EI/c EI/c5
Shear Rigidity KAG KAG/c3 KAG/c3
and build a model with thin plastic products
Modulus of Elasticity E E/c E/e and the proper load transference. Examples of
Section Area
an elastic ship model test for the SL-7 is pro-
Moment of Inertia I I/4 Ie/c5
Distance from neutral axis to outermost vided by Rodd (1976).
fiber for hull-girder (prototype) y y/ y/r
or strength bar (model)
Section Modulus Z Z/3 Zer/c5 As discussed earlier, the design of a seg-
Flexure Stress /c /er mented model will have two decidedly differ-
Note:
is the ratio of prototype to model length
ent paths based upon whether it is an elastic or
c is the ratio of prototype to model water density rigid segmented model. In either case the bal-
e is the ratio of prototype to model modulus of elasticity
r is the ratio of distances from neutral axis to outermost fiber
lasting of each segment must satisfy the weight
and inertial properties of concern for that sec-
tion, and the overall hull weight and ballast
For a ship beam representation the fre-
conditions must be satisfied.
quency of resonance is proportional to :
An elastic segmented model will require
EI careful design of a backspline, internal truss, or
=
l 3 connecting structure from knowing the strength
and rigidity properties of the prototype. If an
with E, the Young Modulus of the material, I, internal truss or connecting structures are used
the moment of inertia of the girder, the dis- between segments then the connections should
placement and l, the length of the girder. model, as closely as possible, the anticipated
ship rigidity and damping at that point. The
Usually segmented models on elastic beams natural frequencies and achieved modal shapes
use a beam in the same material as full scale. should also be used to evaluate the correct
Theoretically the scaling law is then 5. For structural modeling.
many reasons (length of the beam, uncertainty
on the Youngs modulus, sprung effect, The backspline must be designed with vari-
Achtarides 1983) it is difficult to obtain cor- able beam properties, to at a minimum, satisfy
rectly scaled natural frequencies between the the variations at the stations. The variation of
model and the real ship. That means that the the backspline rigidity beyond the segment
models natural frequency should be adjusted level allows further strain measurement at in-
 7.5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 6 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

termediate longitudinal locations. Examples of

elastic segmented models are provided in Fig-
ure 1 and Figure 2.

A rigid segmented model will also require a

design which meets geometrical and physical
inertial characteristics. A rigid segmented
model can incorporate either a very rigid back-
spline or internal truss, or instrumented rigid Figure 2. Segmented Elastic Model with
joint connections along the segmented plane. Transducers at Segments (Wu, 2010).

This requires some knowledge of the ship

rigidity along the length of the ship. If the ship
design itself is immature, then structural char-
acteristics typical of the ship class may be used.
The backspline is typically constructed from
aluminum due to weight and ease of fabrication.
However, more rigid metals might be more
suitable for more rigid, heavier ships. The
backspline is often built with varying cross-
section to model the ships varied rigidity with
Figure 1. Elastic Segmented Model with Inter- respect to longitudinal location. This can be
nal Backspline Beam (Miyake, 2009). done by altering the flange thickness of the
beam. Other materials can be used, but alumi-
num typically is easy to work with, is less
costly, and provides a larger cross-section for
model attachment. Whenever possible the
backspline should be designed and located so
that the bending neutral axis of the backspline
corresponds to the neutral axis of the ship.
With respect to torsional vibrations the shear
center is important, however very difficult to
2.5 Design, fabrication, and instrumenta- obtain, since for open sections the shear center
tion of structure for segmented model might be below the hull.
For the segmented model there is an in- At the location of the segments the beam is
strumented structure which is used to provide outfitted with strain gages to monitor the pri-
the hull structure. This structure can be either mary forces and moments of concern. Exam-
elastic or rigid. Elastic segmented struc- ples of possible strain gage instrumentation on
ture modeling will allow hull rotation and lon- an internal elastic H-beam backspline are pro-
gitudinal and transverse bending relative to the vided in Figure 3. The relationship between
wave environment. Hence the design must try backspline strain and global loads is deter-
to model the rigidity of the ship hull. mined by force and moment calibration prior to
testing. If there are cross talk terms this can be
 7.5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 7 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

VERTICAL
resolved with a calibration matrix which takes B
BENDING GAGES
A

into account any cross talk components. DETAIL E

VERTICAL
VERTICAL SHEAR
The other type of elastic segmented SHEAR GAGES
LATERAL
SHEAR GAGES

model involves the use of instrumented flexible LATERAL LATERAL

connections at the segment break as shown in BENDING GAGES BENDING GAGES

Figure 4. If possible the flexible connections TORSION

TORSION
GAGES
GAGES
should have the same structural damping as the
ship hull at that segment break in the hull. In
some cases the damping/rigidity can be ad-
B A
justed dependent upon the mechanical ar- VERTICAL
BENDING GAGES

rangements. At a minimum, the degree of Figure 3. Possible Strain Gage Measurement

damping should be at the same order of magni- for Internal Elastic Strength Bar (Dinsenbacher,
tude as the ship structure. The moment and 2010).
shear can be measured via force transducer at
each connection point between the segments.
With the right solution method and number of
determinant measurements the global forces at
the segment break can be defined.

The other form of a segmented model in-

volves the use of a rigid structure. This in-
volves either the use of a rigid instrumented
connection between rigid hull segments, or a
very rigid beam where load transducers be-
tween the segments and the rigid structure de-
fine the cumulative force and moment acting
from that particular segment.

In all cases the inertia of the structure will

need to be accounted for both as a contributor
to the segment inertia as well as the overall hull
inertia. This will require a very thorough
treatment of weight and locations for the struc-
tural components.

In all cases of load instrumentation you are

not really measuring the load directly, but
rather the reaction to loads. Even if load trans-
ducers are calibrated independently, the overall
force from a segment should be verified in situ
with independent application of force and force
Figure 4. Instrumented Flexible Connections
couples about the segment locations.
for Non-Backspline Elastic Segmented Model
(Drummen, 2007).
 7.5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 8 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

2.6 Model Segmentation When choosing the segment layout, the rep-
resentation of the deformed mode shape of a
For models which are segmented, the seg- real continuous hull structure by a segmented
mentation is selected based upon the primary model composed of a finite number of seg-
modes of structural deformation to be studied. ments should be considered. This point should
If only mid-ship bending is required then only be studied numerically, even with simple 2D
one mid-ship cut would be required. If maxi- girder representation. Comparison between
mum shear were also needed then three segmented mode shape and numerical estima-
cuts/four sections would be required. If a more tion of the full scale mode shape should be per-
thorough understanding of the dynamic load formed in order to check the validity of the
flow along the hull is required, then five or model design and segmentation.
more cuts might be required. Thinner cross-
sections at the bow, and steering and propul-
sive needs at the stern will require longer sec-
tions for the bow and stern as shown in Figure
1 and Figure 5.

STA STA STA STA STA STA

15 12 10 8 6 4

Figure 5. Example Segmentation Spacing

(Dinsenbacher, 2010).
Figure 6. Example of Segment Sealant (Hay-
The segment gap spacing is usually 5 to 10 den, 2009).
mm in width. A dental quality latex is placed
along the segment gap completely around any Elastic segmented experiments are per-
water sealed interface. The latex provides a formed to determine a response on one or more
watertight pliable connection as shown in Fig- specific modes, it is thus mandatory to identify
ure 6. The latex seal is indented slightly in the structural damping of the tested structures.
toward the hull so that the external hull shape is Even if the damping at full scale is unknown,
minimally affected. experimentally the structural damping should
be measured. Kapsenberg (2002) notes that if a
Usually only the first longitudinal mode is succession of impacts are observed, the struc-
experimentally simulated (sometimes the sec- tural damping is important, especially when a
ond). As a matter of scale it is not possible to second impact is considered. The response to
simulate the response of the structure on local the second impact can be increased or de-
modes or on combinations of modes which are creased then by the effect of the first impact
sometimes identified at full scale. which is not totally damped. Damping will
have a major effect on the assessment of whip-
ping and springing responses.
 7.5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 9 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

2.7 Powering and Steering gular displacements, and angular rates between
the segments of the model being tested as com-
The powering of the model should be done pared to a rigid model. Due to model flex, it is
in such a way as to minimize application of possible to have varying angles of absolute roll
thrust and a moment on any one section of the and pitch for the ship. In an ideal world with
segmented model. This can be done through unlimited funds a high precision 6-axis meas-
appropriate use of gear boxes, flexible joints, urement device would be placed in each seg-
timing belts and other mechanical rigging ar- ment. Then the relative pitch and roll between
rangements. The ideal arrangement is to chan- sections could serve as a check on angular hull
nel the thrust into the longitudinal line of the rotation as noted from the structure. In most
strength bar or truss. The goal is to ensure that cases the accuracy and cost required for such a
the thrust does not exert a longitudinal moment comparison is not reasonable or obtainable.
onto the strength bar. In the case of waterjets
this is not possible and at a minimum the wa- Instruments can be attached to the rigid
terjets should have the same geometric location strength beam or rigid hull points. Acceler-
as full scale waterjets. Calm water non-zero ometers should be located at pre-determined
speed runs can determine the thrust effects on points of interest. However the instruments
measured loads. should be mounted away from any flexure
points that might affect movement of the beam.
The steering should be performed with an
autopilot algorithm if possible so that the algo- In addition to the regular seakeeping mo-
rithm can be transferred across to simulation tions any parameters which effect loading on
efforts. If manual steering is required steering the hull should be collected. If reasonable the
should be minimized to just that required to loads from the propulsion and steerage should
keep the model on heading. The steering from be documented. If there are sidehulls the loads
either method should be minimized (or linear) induced from the side hull acting on the main
as much as possible during the collection of hull should be instrumented and collected. A
data. Ultimately the steering forces will intro- more detailed summary of the parameters to be
duce a side force to the affected segment and in collected are provided in Section 3.1.
some cases can introduce some elements of roll
and acceleration. The steering motions and Part of the experiment design will be to en-
forces can be estimated by performing some sure that data are collected at a sufficient rate to
modified small angle zig-zag maneuvers which fully evaluate frequency content and measure
might be typical of the rudder and heading maximum values. This concept is demon-
variations anticipated during a seakeeping strated by Figure 7. For global loads the load-
heading run. ing will typically be cyclical in nature. Hence
if the natural frequency of the elastic struc-
ture is known, then a sample rate adequate to
2.8 Instrumentation
minimize the error when collecting load cycles
The instrumentation should be sufficient should be appropriate. A sample rate which
and appropriate to measure all components of provides a minimum of ten data points per cy-
concern for a global loads experiment. Due to cle is recommended as shown in Figure 7-a.
the nature of a global load elastic model, there
For impulse loads, as might occur during
can be a greater variation in accelerations, an-
slamming events, the collection rate should be
 7.5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 10 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

sufficient to capture rise time and impulse

maxima as approximated by the triangular
loading of Figure 7-b. The collection rate
should be selected to minimize the error when
the data is not collected exactly at the loading
peak. Dinsenbacher (2010) provides guidance
for collection rates in order to collect slam
events at both model and full scale.

The sample rates can be pre-calculate

knowing frequencies of excitation, but in many
cases the collection rates are set based upon
prior knowledge and system capabilities.
Given the high acquisition capabilities of mod-
ern collection systems, many collections are
performed at much higher rates than needed
and then parsed or filtered to create a lower
effective sample rate.

Often times when measuring global loads it

is informative and in some cases necessary to
also measure secondary loads to aid in the in-
terpretation of the global loads. In some cases
the secondary loads are measured with pressure
sensor grids. However this method of secon- Figure 7. Examples of Loading Time Histories
dary load measurement requires a rather large as Relates to Required Sample Rates (Dinsen-
number of transducers and interpolation tech- bacher, 2010).
niques to derive the pressure distribution. The
preferred method for secondary load measure-
ment uses slam panels and grillages as shown
in Figure 8. The panels and grillages are nor-
mally designed to represent hull plate and
stringer geometries present on the hull. The
panel and grillage sensors provide a better
method for getting a more accurate secondary
design load. Dinsenbacher (2010) provides
some rather detailed information for panel and
grillage design.

Figure 8. Example of Panel and Grillage Sen-

sors (Dinsenbacher, 2010).
 7.5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 11 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

2.9 Test Program Irregular wave tests should be performed to

determine unknown resonance, obtain response
The test program is designed based upon amplitude operators, and provide time series
the program objectives. The wave environment data suitable for deriving long term statistics.
and the type of runs will be based on the needs The sea state and spectral shape to be modeled
developed by the test objectives. At a mini- for testing is dependent upon anticipated oper-
mum there is a need for calm water runs at zero ability requirements and load concerns. As
speed and at the speeds to be used for testing. with regular waves a range of speeds and head-
Calm water data can be used to note hog and ing should be investigated to identify response
sag of the model at constant speed. Roll decay and operability concerns. Head, beam, and
and maneuvers required for testing will help to following headings should be performed at a
quantify the measured loads from the model minimum to provide motions which might be
operating without waves. These runs can be singularly based upon a co-linear versus or-
used when interpreting the loads measured thogonal motions input. However oblique
when operating in waves. With respect to (bow) headings tend to provide a more realistic
regular wave and irregular wave testing, the operational heading, and in some cases can
guidance provided by the Seakeeping Experi- provide the greatest loading, particularly with
ments ITTC Procedure 7.5-02-07-02.1 is also respect to torsion.
applicable to global loads testing when defin-
ing the environment for testing. Other tests to consider are short crested seas
where the water surface profile can sometimes
If the test program is designed to support provide a more severe loading on the hull. If
numerical simulations, then particular care short crested seas are to be considered than
should be taken to provide 6-DOF motions or there will need to be sufficient definition of the
accelerations. External propulsive and steering target seaway and measurement of the gener-
forces should also be characterized. Initial test ated seaway to verify proper modeling. This is
conditions should be controlled and docu- normally defined and documented by specify-
mented for best correlation. ing the spreading distribution of the short
crested seaway.
Regular waves provide the easiest compari-
son to simulations and to aid in the develop- Other testing to be considered is the collec-
ment of transfer functions. With proper wave tion of long total run times to verify long term
elevation measurements relative to the model estimated maxima of global loads. This is ac-
the phase angle of the loading relative to the complished by testing for extended periods of
wave can be determined. Wavelength () to time. For most basins this means the assimila-
ship length (L) ratios of 0.5 to 2.0 should be tion of individual basin passes. It is important
investigated. A wave steepness ratio of 1/50 to ensure that wavemaker repeat sequences of
(H/) is recommended for good linear results, irregular waves are not of concern for random
however steeper waves should be considered stochastic testing. In the event that design ex-
for specific resonant and critical loading fre- treme wave conditions are known, then the
quencies. Various headings from head to fol- design of a deterministic extreme seaway for
lowing should be considered, and speeds cho- testing might be more suitable if this can be
sen should be based upon heading and opera- obtained in the test facility.
tional scenarios.
 7.5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 12 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

2.10 Data Analysis

Data analysis for global loads testing cen-

ters about the effort to discern global loads
information from measured strains, loads, and
moments measured with the various instru-
mented structure. In many cases post acquisi-
tion calculations must be performed to calcu-
late forces and moments. In some cases cali-
bration matrices must be applied to address
cross-talk and interrelation amongst measured
structural response. Additionally there is also
the need to separate structurally measured re-
sponses into lower frequency global hull re-
sponses as compared to higher frequency
whipping and structural responses which might
be present. These types of analysis and filter-
ing are typically performed in the time domain
at each time step of collected data. The low
and high frequency responses can typically be
separated with digital filtering, or if planned
appropriately the recording of the analog chan- Figure 9. Example of Analysis to Filter and
nels at various stages of filtering. An example Arrive at First Mode and Whipping Responses
of the results of this approach is presented in (Dinsenbacher, 2010).
Figure 9. All of these methods assume data
collection has been performed at sufficiently 2.11 Data Presentation
high rates to collect all phenomena of interest.
The structural data can be presented either
Once the measured responses are divided as full scale values required for design, or as
into the frequencies of concern, the analysis dimensionless values more suitable for com-
will deviate for the two type of responses. parison to other designs and computational
Primary statistics, histograms, spectral analysis, approaches. Model scale values can be used as
and response amplitude operators can be calcu- a way of visualizing and interpreting the results
late for the low frequency global components. early on in the experiment, but ultimately most
The short duration slamming responses will global loads results are needed at a level be-
need to be analyzed with temporal analysis to yond this early analysis. The global loads data
determine rise time and duration of slam events. must be taken from the measurement level of
The whipping and springing motions will need strain and transducers values to overall hull
to be analyzed to determine resonance and loads such as shear, moment, and torsion which
damping. Weibull and extreme value analysis are required in the hull design. These overall
can be performed on measured response event design loads can then be made non-
distributions and response time histories to dimensional using the representation provided
calculate future probability and magnitudes of by Dinsenbacher (2010) in Table 2.
extremes.
 7.5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 13 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

These results can then be presented as op- frequency domain then the comparison should
erational cyclical loads (i.e. for fatigue analysis) most likely be performed in non-dimensional
or anticipated maximum values anticipated frequency coordinates. If specific time domain
based upon seaway conditions, heading, and seaways are modeled, than either the time do-
speed. Extreme value theory and other lifetime main responses and loads in the time domain
design statistics can also be applied to establish can be compared or statistical evaluation of the
the maximum load which should be used in the time domain results may be compared. In all
ship design. instances the measured wave should be used as
input to the simulation to ensure better com-
Table 2. Dimensionless Representation of Key parison for irregular waves. When comparing
Parameters (Dinsenbacher, 2010). the experimental and computed results the un-
QUANTITY DIMENSIONLESS COEFFICIENT
M RAOM certainties associated with each should be de-
Bending Moment ( M ) =CM =
hW gL2S B gL2S B fined to allow proper comparison.
V RAOV
Shear Force ( V ) =CV =
hW gLS B gLS B
T RAOT
=CT =
Torsion Moment ( T )
hW gL2S B gL2S B 3. PARAMETER
p RAO p
Pressure ( p ) =
CP =
hW g g

Pitch Angle ( ) =C

= =
RAO RAO 3.1 Parameters to be Taken into Account
2hW 2 k
LW
LW

Heave Displacement ( z ) C=
Z
z
= RAOz The following parameters defining the tests
hW
RAO RAO
are to be taken into account and documented
Roll Angle ( ) =C = =
2hW



2
LW
k (as applicable):
LW

C=
y
= RAOy
Scale
Sway Displacement ( y ) y
hW

Surge Displacement ( x )
C=x
x
=
hW
RAOx Model dimensions
Yaw Angle ( ) =C =

2hW
=
RAO
2
RAO
k
Ratios of model to tank dimensions
LW LW

=Ca =
a RAOa Hull configuration (lines, appendages, su-
Acceleration ( a )
hW g g



LS LS perstructures, ...)

Deflection ( )
C=
= RAO
hW Loading conditions
Note that in the table the dimensional response is in all cases assumed to be peak to peak.
In addition to symbols already defined, k is the wave number, 2 / LW . Mass distribution (COG, inertias, ...)
Speeds and headings
2.12 Comparison to Predictions and Nu- Towing and/or restraining device characte-
merical Simulation ristics (specially DOF)
When comparing predictions to numerical Wave characteristics (heights, periods,
simulations, this can be performed by looking spectra, dispersions, ...)
at the magnitude of the interested parameters, Autopilot control law
or by looking at the loads and motions in either
the frequency or the time domain. The mode Speed control characteristics
of comparison is dependent upon the nature Run duration
and output type of the numerical simulation. If
the results of the simulation are output in the Number of runs per test condition
 7.5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 14 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

Positions of sensors (accelerometers, rela- 4.1 Uncertainty Analysis

tive motion, encountered wave, ...)
Uncertainty analysis for the experiment
Resonance frequencies for segmented mod- should be performed per the recommendations
els of ITTC Procedure 7.5-02 07-02.1 (Appendix
Sampling frequency A), following the ISO-GUM 1995 guidelines.
Most of the examples and techniques apply
Sensor calibrations and accuracy equally well to global loads experiments with
Rigidity and damping characteristics of variations as required to accommodate struc-
strength bar, or connections used for seg- tural calibrations and measurements.
mented models
4.2 References
3.2 Recommendation of ITTC for Parame-
ters Achtarides, T.A., 1983, Design of multiseg-
mented models for springing experiments,
In addition to the above listed parameters, a PRADS 83, Tokyo and Seoul.
sufficient definition of the model design should
be provided. The type of global loads testing, Dinsenbacher, A., Engle, A., Hermanski, G.,
and assumptions made during the experiment 2010, Guidelines for Hydroelastic Model
design should be documented. The experiment Design, Testing and Analysis of Loads and
documentation should provide any background Responses, Report No. NSWCCD-65-TR-
relative to the experiment and model design 2010/12, Carderock Div, NSWC.
which will aid in the future interpretation and
Drummen, I., 2007, Experimental and Numeri-
correlation to the experimentally collected data.
cal Investigation of Nonlinear Wave-
Induced Load Effects in Containerships
4. VALIDATION considering Hydroelasticity, thesis Norwe-
gian University of Science and Technology,
The global loads seakeeping test can be- Trondheim, Norway (NTNU).
come rather complex. The best way to main-
tain validity across the whole of the experiment Hayden, D.D., Bishop, R.C., Melendez, M.P.,
is to verify the validity of the intermediate 2009, Seakeeping Performance Evaluation
steps. This is accomplished by maintaining of the High Speed Sealift (HSS) Trimaran
accuracies and controls across the experiment as Represented by Model 5594, Report No.
design, model and instrument design and fabri- NSWCCD-50-TR-2009/048, Carderock
cation, ballasting and geometric definitions, Div, NSWC.
wave environment, experimental performance,
and data collection and analysis techniques. Kapsenberg, G.K., vant Veer, A.P., , Hackett,
These intermediate steps have been briefly de- J.P., . Levadou, M.M.D, 2002, Whipping
scribed in the procedure. To maintain and ver- loads due to aft body slamming, 24th
ify validity of the test, the experimenter must Symposium on Naval Hydrodynamics, Fu-
maintain, define, and document validity at the kuoka, Japon, 8-13 July 2002.
intermediate steps.
Miyake, R, et. al., 2009, Experimental Studies
on the Hydroeleastic Response Using a
 7.5-02
ITTC Recommended 07-02.6
Procedures and Guidelines Page 15 of 15

Effective Date Revision

Global Loads Seakeeping Procedure 2011 00

Flexible Mega-Container Ship Model, Hy-

droelasticity in Marine Technology, South Wu, M.K., Hermundstad, O.A., Zhu, S., 2010,
Hampton, 2009. Comparative Study of Springing and
Whipping Effects in Ultra Large Container
Rodd, J.L., 1976, Verification of the Rigid Vi- Ships, ITTC Workshop on Seakeeping,
nyl Modeling Technique: the SL-7 Struc- Seoul, Korea, 19-21 October 2010.
ture, Special Report SSC-259, Naval Ship
Research and Development Center. 4.3 Benchmark Tests

No readily accessible benchmark data

available.