Proceedings of OMAE’03
22nd International Conference on Offshore Mechanics and Arctic Engineering
June 8-13, 2003, Cancun, Mexico
OMAE2003-37174
DEVELOPMENT OF SAFER SHIPS BY DETERMINISTIC ANALYSIS OF EXTREME
ROLL MOTIONS IN HARSH SEAS
Günther F. Clauss
Ocean Engineering Division
Technical University Berlin
Germany
clauss@ism.tu-berlin.de
Janou Hennig
Ocean Engineering Division
Technical University Berlin
Germany
hennig@ism.tu-berlin.de
Heike Cramer
Flensburg Shipyard
Germany
cramer@fsg-ship.de
Stefan Krüger
Technical University Hamburg-Harburg
Germany
krueger@tu-harburg.de
ABSTRACT
NOMENCLATURE
|F(ω)| Fourier spectrum
GM metacentric height
Hmax maximum wave height
Hs significant wave height
Tp peak period
TR period of roll resonance
T0 zero-upcrossing period
d water depth
t time
v velocity
x, y coordinates in space
α1 main board angle
α2 upper flap angle
ζ(t, x) surface elevation
ϑ pitch
µ course
ϕ roll
ω circular frequency
For modern ship design — as confirmed by many examples
of ship design and operation — the current intact stability code
(IMO (2002)) does not provide a reliable basis for the assessment of ship safety in rough seas. In ship design the application of the IS-Code is not supporting design decisions towards
increased safety in rough seas (Cramer and Krüger (2001)), and
ship operators find their cargo and/or vessels endangered by large
roll angles and accelerations (France et al. (2001)). Thus, there is
a great need for procedures to analyze ship safety in rough seas.
This paper presents innovative deterministic seakeeping test
procedures which are used to identify the physical mechanism
endangering intact ships by evaluating cause-reaction relations of
wave/structure interaction. Rogue wave sequences are embedded
in severe seas for computer controlled capsizing tests of different
vessels at the Hamburg Ship Model Basin. The model test results
are used as a basis for the development of non-linear numerical
methods to simulate ship motions in extreme seas with the target
to design safer ships with reduced capsizing risk. Polar plots
following from the non-linear simulations allow the evaluation
of ship safety in severe seas with reference to course, speed and
trim.
1
Copyright  2003 by ASME
Figure 1. ENCOUNTERING A ROGUE WAVE – BAY OF BISCAY (Nickerson (1993))
Figure 2. CONFIGURATION FOR COMPUTER CONTROLLED SEAKEEPING TESTS
DEMAND FOR INTACT STABILITY CRITERIA
At present, the assessment of ship stability – intact or damaged – is confined to the fulfilment of empirical criteria related to the static lever arm curve for still water conditions only.
The IMO intact stability criteria (IS-Code, IMO (2002)) are prescriptive rules based on experiences with cargo ships quite some
years ago. Thus it is not surprising, that the current criteria fail
to appropriately assess the safety of modern designs in rough sea
conditions as ship design rapidly changes. Unfortunately legal
(according to the rules and regulations) is not necessarily safe.
Some of the large container ships recently suffered
from parametric excitation, loosing and/or damaging cargo
(e. g. France et al. (2001)). Many modern designs are susceptible to parametric excitation, not only container ships, but also
RoRo-, RoPax-, ferry and cruise vessels. Also quite a few vessels are endangered by pure loss of stability or combinations
of parametric excitation and loss of stability. Currently the intact stability rules do not cover these dangerous mechanisms and
other unfavourable seakeeping scenarios. In addition, appropriate guidance for the crew on how to identify and avoid such dangerous conditions is lacking. Furthermore, ship operators often
complain about very short roll periods leading to high accelerations, especially if combined with insufficient roll damping
as well as insufficient course-keeping capabilities of vessels in
rough weather.
Due to these problems with the current IS-Code a revision
process has started at the last IMO-SLF meeting. It has been decided that next to some short term amendments the code should
be completely revised with two major aims:
cal simulations shall be possible.
Consequently the following steps have to be performed in the
revision process:
• Identification of safety related situations/mechanisms endangering intact stability
• Collection of existing knowledge and further research related to physical phenomena endangering the ship in rough
sea conditions and towards procedures to assess ship performance in dangerous situations.
• Development of a framework of performance based intact
stability criteria
• Definition of criteria with appropriate standards
The seakeeping test procedure presented here as well as its evaluation and application lead to a better understanding of safety
related aspects in ship stability – not only in situations which appear as evident as the famous rogue wave encounter in Fig. 1.
And based on numerical motion simulations ship designs can be
systematically optimized – already within the early design stages
where optimization potential is largest.
SETUP FOR DETERMINISTIC SEAKEEPING TESTS
Reliable validation of numerical assessment of ship stability calls
for a special test setup which allows to refer the resulting ship behaviour directly to the corresponding wave excitation. The following section gives a brief overview of the new seakeeping test
setup and procedure which has been used for capsizing investigations and validation of numerical data explained in the following
sections.
1. all new criteria shall be formulated as performance based
criteria
2. alternative direct assessments via model test and/or numeri2
Copyright  2003 by ASME
T0 = 10.8 s, Hs = 11.92 m, Hmax = 25.6 m
ζ [m]
t [s]
Figure 4. DRAUPNER NEW YEAR WAVE (Haver and Anderson
(2000)): FULL SCALE DATA AND WAVE TANK GENERATION (RED
LINE, SCALE 1:81).
Figure 3. INVESTIGATED SHIP MODELS AND THE CORRESPONDING SHIP LINES (C-BOX LEFT, RO-RO VESSEL RIGHT)
When the model reaches the critical safety distance from the
wave maker or the absorbers at the opposite side of the tank, the
ship and the carriage stops automatically.
In Fig. 3, the investigated ship types are presented, i. e. a
C-Box multipurpose vessel and a RO-RO ship.
Fig. 2 shows the test setup for fully computer controlled seakeeping tests (see Kühnlein and Brink (2002)). Three main system components are coordinated:
• wave maker
• towing carriage (including horizontal carriage)
• ship model
ROGUE WAVE GENERATION WITHIN HARSH SEAS
Meteomarine conditions are selected as characteristics of the
assessment process and realized in the wave tank by the wave
generation procedure. Such a procedure allows to generate arbitrary types of model seas such as statistically defined storm seas
which contain critical wave groups occurring at desired positions
in time and space. This allows to analyze capsizing scenarios deterministically.
The key step in the wave generation process is the consideration of nonlinear wave propagation. The empirical procedure
is very flexible and can be adapted to any sea condition (Clauss
(2002), Kühnlein et al. (2002), Clauss and Hennig (2001)). Even
extreme events like the New Year Wave (Trulsen and Dysthe
(1997), Haver (2000), Haver and Anderson (2000)) can be modelled in the wave tank with high precision (Fig. 4).
The parameters of the model seas - transient wave sequences
consisting of random seas or regular wave trains with an embedded deterministic high transient wave - are systematically varied
to investigate the ship model response with regard to metacentric
height, model velocity, and course angle for different ship types
(Clauss and Hennig (2002)) and failure modes.
In head seas, the ship is positioned at the side wall of the
tank end, opposite to the wave maker position. In seas from
astern, the ship model waits close to the wave maker until a defined sequence of the wave train has passed. The two-paddle
wave maker starts to generate the specified wave train which has
been selected for an individual test. Preprocessing includes the
calculation of the encounter point in time and space where the
ship meets the wave train under defined conditions. Thus the
model starts to sail through the tank in such a way that it reaches
the encounter position at the proposed time step. For keeping the
given parameters constant, the propulsion is controlled over the
entire run. The ship moves in parallel with the tank side wall at
a required minimum distance. Registration starts by setting the
desired course.
The ship’s course is controlled by the master computer by
telemetry which commands a Z-manoeuvre at constant course
angle and model velocity. These test parameters as well as the
model sea parameters are varied according to the metacentric
height GM of the model, the expected rolling mode and occurrence of resonance. Ship motions in six degrees of freedom are
registered precisely by computer controlled guidance of both, the
towing and the horizontal carriage: During the entire test run, the
ship model stays in the field of vision of the optical system line
cameras. Additionally, the wave train is measured at several fixed
positions of the wave tank.
CORRELATION OF WAVE EXCITATION AND SHIP RESPONSE
For providing useful data for the analysis of the capsizing
process as well as for the validation of numerical methods, exact
correlation of wave excitation and ship rolling is indispensable.
This correlation is achieved by
3
Copyright  2003 by ASME
stationary wave probe at x = 8.74 m
wave elevation at stationary wave probe
(x = 85.03 m)
stationary wave probe
at x = 85.03 m
wave elevation at ship center (moving frame)
velocity of towing carriage v = 1.65 m/s
(in direction of wave propagation)
roll motion
moving wave probe
Figure 5.
REGISTRATION OF A WAVE PACKET AT STATIONARY
WAVE PROBES AND AT A MOVING PROBE – COMPARISON OF CALCULATION AND REGISTRATION
pitch motion
• technical synchronization (see test setup), and
• knowledge of nonlinear propagation of waves
The wave elevation at the position of the cruising ship model in
time and space is calculated (and controlled by registrations during model tests). Nonlinear methods are used for calculating the
moving frame wave train (Clauss and Hennig (2001)). Fig. 5
confirms the high accuracy using these tools: The wave train
(transient wave packet) is measured at a stationary wave probe
close to the position of the wave maker (x = 8.74 m) as well as
at a down–stream position of the wave tank (x = 85.03 m). Finally, the lower diagram presents the wave train as registered on
board of the towing carriage (mean velocity 1.65 m/s). Transformation of the first wave train to the fixed down-stream position
as well as to the moving wave probe travelling with the carriage
is also shown. Agreement of registration and calculation (dots)
is satisfactory.
ship position
target course
broaching with subsequent capsizing →
Figure 6. EXAMPLE FOR BROACHING OF THE RO-RO VESSEL
(GM=1.36 m, v = 15 kn, Z-MANOEUVRE AT µ = ±10◦ ) WITH SUBSEQUENT CAPSIZING IN HARSH SEAS (Tp = 14.6 s, Hs = 15.3 m).
EVALUATION OF CAPSIZING TEST RESULTS
Fig. 6 presents a model test with a RO-RO vessel
(GM=1.36 m, natural roll period TR =19.2 s, v=15 kn) in extremely high seas from astern (ITTC spectrum with Hs =15.3 m,
TP =14.6 s, Z-manoeuvre: target course µ = ±10◦ ).
4
Copyright  2003 by ASME
The upper diagram presents the registration at a stationary
wave probe. As the waves are quite high the associated crests are
short and steep followed by flat and long troughs. In contrast, the
cruising ship – see wave elevation at ship center (moving frame)–
apparently experiences extremely long crests and short troughs
with periods well above 20 s as the vessel is surfing on top of
the waves. The ship looses stability while situated on a wave
crest, broaches, and finally capsizes as the vessel roll exceeds
40◦ and the course becomes uncontrollable (Fig. 9). Note that
the wave characteristics, i. e. wave elevation and the associated
pressure field as well as the acceleration and velocity distribution in time and space refer to the moving frame at the reference
position of the cruising ship, and can be directly correlated to
the ship motions by magnitude and phase. As a consequence,
the seakeeping behaviour and even the mechanism of capsizing
can be deduced and explained on the basis of non-linear causeeffect chains (Clauss (2002)). In the framework of the German
research project ROLL-S similar dangerous situations have been
simulated which are used for validation of numerical models.
Figure 7. PICTURE TAKEN FROM THE SIMULATION OF THE SEAKEEPING BEHAVIOUR OF THE RO-RO VESSEL IN HIGH SEAS FROM
ASTERN
Within the German research project SinSee different
methodologies for quantitative assessment are under development. First results from test calculations for existing ships are
presented in (Krüger (2002)), and the conclusions from numerical motion simulations regarding seakeeping capabilities of different ships correlate well with operational experience.
DIRECT ASSESSMENT OF INTACT STABILITY
In many areas of ship design and approval direct assessment
is accepted to prove sufficient safety or strength (e.g. stress analysis in structural design or evacuation). With respect to the intact stability approval, there are so far no alternative approaches
established next to the ”simple” fulfilment of current empirical
criteria with its already mentioned deficiencies.
Meanwhile, numerical tools have been developed and successfully tested which allow the evaluation of dangerous and
even fatal scenarios. Söding (1987) investigated the loss of the
”E.L.M.A. Tres”, and France et al. (2001) analyzed parametric
excitation of C11 Class Container ships, while the use of a ”design for safety methodology” – based on direct numerical simulations and a qualitative assessment – allows an increase of ship
safety in severe seas without impeding their economical performance. An example for the latter is given in the following section
of this paper. Despite of the known deficiencies of the presently
available numerical tools they can be applied to assess ship safety
when being used appropriately.
When developing or applying direct approaches it should be
noted that the kind of safety level of existing empirical rules as
well as the safety level of international acceptance are still under discussion. To allow a quantitative direct assessment of ship
intact stability it is therefore necessary to develop methodologies for alternative approaches and to estimate the internationally tolerable safety level. Those developments are indispensible
for future ship approval and design, as empirical formulations
will never provide a sufficiently broad and fair evaluation basis
to cover all possible new design developments. This is also reflected in the IMO decision towards performance based criteria
and alternative direct assessments.
OPTIMIZATION OF SEAKEEPING SAFETY
Fig. 7 presents a picture taken from a simulation of the seakeeping behaviour of a RO-RO vessel in high seas from astern.
Based on the results from such individual motion simulations the
basic physical phenomena which potentially endanger the ship
are analyzed, and the ship design is optimized with respect to
these findings resulting in increased safety at rough sea conditions.
To identify efficiently the potentially dangerous situations
within the early design phase a ”design for seakeeping safety
methodology” has been developed (Cramer and Krüger (2001)).
Large numbers of motion simulations are systematically carried
out, and the results are presented in Fig. 8 as polar plots for two
different existing vessels. Both of them have a length around
180 m and were built recently. The loading conditions are chosen
combining the design draught with the associated maximum KG
according to the current IS-Code. In these polar plots the significant wave length is approximately ship length and all simulations
are carried out in short crested irregular seas according to the
JONSWAP spectrum. Installed devices for additional roll damping such as flume-tanks or stabilizing fins are assumed switched
off. For each combination of speed and encounter angle the limiting significant wave height with respect to the danger of extreme
roll angles and capsizing is estimated and the plots are coloured
5
Copyright  2003 by ASME
accordingly.
It can be seen that the RoPax-Ferry (upper diagram) is susceptible to encounter extreme roll angles in following seas in
general, already in reasonably small wave heights. This corresponds well with the experiences of the owners who are very
worried with the ship performance in waves. The lower diagram
of Fig. 8 presents a polar plot for a recently built RoRo-Ship.
This ship has been designed using the above design for seakeeping safety methodology. The two regions of ship resonance are
small, and in general the limiting significant wave heights are
substantially larger. This example illustrates how ships can be
compared based on polar plots and shows the optimization potential gained in early design stages when the assessment of a ship
seakeeping performance is based on nonlinear numerical motion
simulations. Furthermore guidelines for ship operation can be
based on such polar plots.
in such extreme environments.
• Finally, non-linear numerical methods are developed and
validated by dedicated seakeeping model tests in deterministic wave sequences. By systematic simulations the most
critical conditions are identified.
• Based on these developments methodologies for the quantitative assessment of capsizing risk are proposed which then
provide a basis for the improvement of current intact stability criteria. In conclusion ships are designed with improved
sea keeping characteristics and an increased safety with respect to the danger of extreme roll angles and capsizing.
• As a consequence, evaluation methods for capsizing risks
are developed, and stability criteria improved.
CONCLUSIONS
In conclusion, deterministic wave sequences are applied at
model tests for investigating the wave/ship interaction at extreme
sea conditions. Based on these results numerical procedures for
motion simulations are developed, improved and validated which
are used in ship design to optimize the safety of ships in extreme conditions. The efficient coupling between different disciplines like deterministic wave generation and computer controlled model testing promotes the realization of deterministic
model tests, and thus the development of non-linear numerical
simulation methods for ship design, which is necessary to solve
the current problems related to ship safety at rough and extreme
sea conditions.
• In detail, the method is used as a tool to analyze the mechanism of motion behaviour in waves because the non-linear
cause-effect chains are deduced from deterministic wave
field characteristics like pressure field, particle accelerations
and velocities as well as non-linear wave elevation in space
and time.
• Wave trains are designed individually to investigate a specific structure at a certain tank position, i. e. some dedicated
regular waves can precede an extremely high wave or wave
group for simulating memory effects. By stretching or compressing the peak wave the associated frequency and slope
can be tuned accordingly. Also phase relations between the
existing wave and the resulting structure motions can be selected and varied deterministically. Any test can be repeated
identically if a specific effect is analyzed.
• Observed wave registrations, like the extremely high New
Year Wave Sequence can be generated in a wave tank at a
selected model scale. Thus, the genesis of extreme events
in such wave groups is analyzed in space and time. Also,
the seakeeping behaviour of any structure can be evaluated
6
Copyright  2003 by ASME
head seas
Recently built RoPax−Ferry
(traditional rule−based design)
20 kn
t
ar
qu
se
as
ots
he
ad
g
in
er
16 kn
ots
as
se
qu
ar
t
er
in
g
ad
he
limiting significant
wave height
m
14
13
12 kn
ots
8 kn
ots
12
11
4 kn
10
9
8
7
6
beam
seas
beam
seas
5
4
ng
ri
qu
ar
te
ri
ng
te
ar
qu
se
as
3
as
se
following seas
head seas
he
ots
s
he
qu
ar
as
se
te
ad
ri
ng
g
16 kn
in
ad
r
te
ar
se
a
Recently built RoRo−Ship
(design for sea keeping safety)
qu
22 kn
ots
20 kn
ots
12 kn
ots
limiting significant
wave height
m
14
13
12
11
10
9
8
7
6
5
4
8 kn
ots
4 kn
beam
seas
beam
seas
as
se
qu
a
rt
ng
er
ri
in
te
g
se
ar
qu
as
3
following seas
Figure 8.
POLAR PLOTS WITH LIMITING WAVE HEIGHTS FOR TWO DIFFERENT SHIPS - BASED ON DIFFERENT DESIGN STRATEGIES.
7
Copyright  2003 by ASME
ACKNOWLEDGMENT
The authors are indebted to the German Federal Ministry
of Research and Education (BMBF) for funding the projects
”ROLL-S” and ”SinSee”. Within these frameworks the presented
results have been achieved. We highly appreciate the approval
of the succeeding project ”SinSee” which allows further applications of the achieved results to promote the development of
sophisticated analysis tools like neural networks.
Nickerson. Freak waves! Mariners Weather Log, NOAA, 37(4):
14–19, 1993.
H. Söding. Ermittlung der Kentergefahr aus Bewegungssimulationen. Schiffstechnik, 34, 1987.
K. Trulsen and K. Dysthe. Freak Waves a a three-dimensional
wave simulation. In Proceedings of the 21st Symposium on
Naval Hydrodynamics, pages 550–558, 1997.
REFERENCES
G.F. Clauss. Genesis of design wave groups in extreme seas for
the evaluation of wave/structure interaction. In 24th Symposium on Naval Hydrodynamics, Fukuoka, 2002.
G.F. Clauss and J. Hennig. Tailored Transient Wave Packet Sequences for Computer Controlled Seakeeping Tests. In OMAE
2001 - 20th Conference on Offshore Mechanics and Arctic Engineering, Rio de Janeiro, Brasil, 2001.
G.F. Clauss and J. Hennig. Computer Controlled Capsizing Tests
with Tailored Wave Sequences. In OMAE 2002 - 21st Conference on Offshore Mechanics and Arctic Engineering, Oslo,
Norway, 2002. OMAE2002-28297.
H. Cramer and S. Krüger. Numerical Capsizing Simulations
and Consequences for Ship Design. In STG Summermeeting
Gdansk, Poland, 2001.
W.N. France, M. Levadou, T.W. Treakle, J.R. Paulling, K.
Michel, and C. Moore. An Investigation of Head-Sea Parametric Rolling and its Influence on Container Lashing Systems. In
SNAME Annual Meeting, 2001.
S. Haver. Some Evidences of the Existence of Socalled Freak
Waves. In Rogue Waves 2000, Brest, France, 2000.
S. Haver and O.J. Anderson. Freak Waves: Rare Realization of
a Typical Population or Typical Realization of a Rare Population? In Proceedings of the Tenth International Offshore and
Polar Engineering Conference, pages 123–130. ISOPE, Seattle, USA, 2000.
IMO. Code on Intact Stability for all types of ships covered by
IMO instruments. Resolution A.749(18) as amended by resolution MSC.75(69), IMO, 2002.
S. Krüger. Dynamic Stability of RoRo-Ships in waves. Conference: Design and Safety of Ro-Ro Passenger Ships, Kopenhagen, Danish Society of Naval Architecture and Marine Engineering, 2002.
W.K. Kühnlein and K.E. Brink. Model Tests for the Validation
of Extreme Roll Motion Predictions. In OMAE 2002 - Proceedings of 21st Conference on Offshore Mechanics and Arctic
Engineering, Oslo, Norway, 2002. OMAE2002-28269.
W.L. Kühnlein, G.F. Clauss, and J. Hennig. Tailor made freak
waves within irregular seas. In OMAE 2002 - 21st Conference
on Offshore Mechanics and Arctic Engineering, Oslo, 2002.
OMAE2002-28524.
8
Copyright  2003 by ASME
Figure 9.
CAPSIZING OF THE RO-RO VESSEL IN A SEVERE MODEL STORM.
9
Copyright  2003 by ASME