1 Wave induced motions and loads on conventional ships.

Theory and numerical

methods. 0. Faltinsen, NTH.

2 Breaking waves - environmental forces. S.P. Kjeldsen, Maritime Academy.

3 Survival testing of small vessels. T. Nedrelid, Marintek.

4 Dynamic stability/probabilistic approaches. A. Francescutto,

University of Trieste.

5 Operational stability - future rules and regulations. E. Aall Dahle, DnV.

6 Examples of stability, casualties/non-intact problems. G. Hjort,

Norwegian Marine Directorate.

7 Full scale results. Dr. Sebastiani, CETENA.

8 Ship operation in a seaway. T.E. Svensen, DnV.

9 Use of seakeeping theories in design. T.E. Svensen, DnV.

10 Seakeeping of high speed vessels. 0. Faltinsen, NTH.

11 Ride control systems of foil catamarans. E. Lunde, CAMO.

12 Ride control of SES. A. Sorensen, ABB Indjustries.

13 Safety of high speed craft. K.O. Holden, NTNF (no manuscript).

14 Motion characteristics of offshore floating vessels. I.J. Fylling, Marintek.

15 Dynamics of offshore towing line systems. S. Moxnes, Marintek.

16 Heavy lift operations; methods and experiences. L. Loken, Heerema.

understanding and statistics.

17 The experience of extreme events; physical
C.T. Stansberg, Marintek.

18 Position control - from anchoring to DP systems. J.E.W. Wichers, Marin.

19

20
VETP - SHORT COURSE ON NEW TECHNIQUES FOR ASSESSING AND
QUANTIFYING VESSEL STABILITY AND SEAKEEPING QUALITIES
at
MARINTEK, Trondheim 8 - 11 March 1993

Participants:

Company Name:

Abeking & Rasmussen, Germany Erwin Karabinski, Dipl.Ing.

Brodosplit, Croatia Damir Bezinovic, Naval Architect

Danyard A/S, Denmark Christian Schack, Naval Architect

Gusto Engineering by, Holland Carlo van dee Stoep, Senior Eng. Spec.

Insean, Italy Dr. R. Penna

Karlskronavarvet AB, Sweden Jan Bergholtz, Naval Architect

Mats Olsson, Naval Architect

Kva&rner Masa-Yards. Finland Ismo Lindstrom

Ministry of Defence J.L. Perluka,

Roval Netherlands Navy. The Netherlands R. Brouwer

Netherlands Coast Guard, The Netherlands A. Schaap, Ing./Naval Architect

Norwegian Marine Directorate Dag Liseth, Div. Eng.

Sedco-Forex, France Georges Barreau, Naval engineer

SSPA Maritime Consulting AB. Sweden Jan Lundgren

STN Systemtechnic Nord, Germany Frank Meyer, Dipl. Ing.

 WAVE TINDUCED MOTIONS AND LOADS ON SHIPS.
THEORY ANT) NUMIERICAL METHODS

0. Faltinsen
Department of Marine Hydrodynamics
Norwegian Institute of Technology
N-7034 Trondheim, Norway

To be presented at UETP Course on New Techniques for Assessing and

quantifying vessel stability and seakeeping qualities, Trondheim, March 1993.

INTRODUCTION

Normally ship design is based on still water performance. However, it is possible

to incorporate seakeeping considerations from the beginning of the design either
by experimental or numerical methods. Important seakeeping variables can be
local motions, accelerations, added.resistance, slamming, water on deck, liquid
sloshing in tanks and wave bending moments and shear forces. We will focus our
attention on heave and pitch motions.

We will discuss the state of the art of computer aided ship motion predictions both
for conventional and high-speed vessels. For conventional ships this includes
linear theories like strip theories, unified theory and complete three-dimensional
theories. A high speed theory that accounts for the divergent wave systems, are
presented. Nonlinear theories are also discussed. The importance and possibility
to predict the influence of flow separation on the vertical motions of conventional
ships are studied. Numerical methods that accurately describe slamming on hull-
sections are discussed.

We will discuss in details strip theory calculations of heave and pitch of a ship at
moderate forward speed in head sea. It is not common to present errors in ship
motion calculations. However we feel that it is important to ensure that the errors
are smaller than the maximum variations of the heave and pitch of a realistic
family of hull forms. Errors can be divided into human errors, numerical errors
and physical errors. Human errors mean for instance "bugs" in computer
programs, wrong interpretation of input and output. This error source can be
minimized by documentation of proper verification procedures for the computer
programs and by standards for quality control of use of computer programs.

We will present a procedure to estimate numerical and physical errors in ship

motion calculations. This is difficult because we still do not understand properly
all physical phenomena associated with ship motions. However, we will make an
attempt and present examples on er-ror estimates.
SHIP MOTION THEORIES

Generally speaking strip theories are still the most successfuil theories for wave
induced motions of ships at moderate for-ward speed. However from a theoretical
point of view one can question strip theories. A strip theory is based on linearity.
This means for instance that the ship motions are small relative to the cross-
sectional dimensions of the ship. In practice one 'forgets" the linearity
assumptions and applies strip theory programs when parts of the ship go out and
in of the water or in predicting green water on deck. Due to the linearity
assumption there are only hydrodynamic effects of the hull below the mean free
surface level. A strip theory program Ail] not distinguish between alternative
above-water hull forms.

A strip theory is based on potential flow theory. This means for instance viscous
effects are neglected. The most severe consequence of this is in the prediction of
roll at resonance. In practice viscous roll damping effects are accounted for by
empirical formulas.

The way that the forced motion problems are solved in strip theory, the method
cannot be justified when the frequency of encounter is low like it may be in
following and quartering seas.The Seakeeping Committee of the 16th I'ITC reports
for instance substantial disagreementtbetween calculated results and experimental
investigations of vertical wave loads in following waves.

Strip theories account for the interaction with the for-ward speed in a simplistic
way. The effect of the steady wave system around the ship is neglected. The free
surface conditions are simplified so that the unsteady waves generated by the ship
are propagating in directions perpendicular to the centre plane. In reality the
wave systems may be far more complex. For instance for high Froude numbers
unsteady "divergent" wave systems become important. This effect is neglected in
strip theories.

Strip theory is also questionable to apply for sh-ips with low length to beam ratios.
The reason is that strip theory is a slender body theory. On the other hand the
Seakeeping Committee of the 18th WI'C concludes that strip theory appears to
remarkably effective for predicting the motions of ships with length to beam
ratios as low as 2.5. There exists different types of striptheories. One commonly
used method is the STF-method (Salvesen et al. (1970)).

There have been different attempts to improve strip theories by following a

rational approach. One example is Ogilvie & Tuck's theory (1969) for linear forced
heave and pitch motion of a ship. They make a high frequency assumption and
show that strip theory is consistent in the near-field of a slender ship at zero
forward speed. The story is different for for-ward speed. Ogilvie & Tuck include
interactions with the local steady flow. This effect occurs both in the body
boundary conditions, in the free surface conditions and bow the pressure forces are
integrated. In the body boundary conditions they include the socalled mi-terms
with second derivatives of the steady motion potential. These are di fficul tto
compute in areas on the body surface with high curvature. Special care has to be
shown at sharp corners, where the terms are singular. Ogilvie & Tuck avoided the
problem by using an integral theorem to compute the effect of the ra-terms. The
waves that Ogilvie'& Tuck predict in the near-field of a ship at forward speed are
propagating in directions perpendicular to the centre plane of the ship. However,
their form are different from what strip theory predicts. In solving the steady
forward motion part of the problem Ogilvie & Tuck account only for the transverse
wave system. This is appropriate for moderate forward speed. At high forward
speed the divergent wave system is important. There is no clear borderline
between high and moderate forward speed. One existing definition is that high
forward speed means Froude numbers larger than 0.4.

Newman (1978) has presented a unified theory for the forced motion problem of
a ship. Numerical results for both zero and forward speed have been presented by
Newman & Sclavounos (1980). By unified theory is meant that both high and low
frequencies are covered. The theory is based on linear potential flow and is limited
to moderate forward speed. Their numerical results for added mass and damping
at zero speed show quite similar results as strip theory for we(L/g) 1/ 2 > -2 (me =
circular frequency of oscillation of the ship, L = ship length, g = acceleration of
gravity). For lower frequencies at zero forward speed the unified theory shows a
correct behaviour relative to complete linear 3-D solutions while strip theory gives
unsatisfactory results for added mass and damping in the low frequency range.
The forward speed effect is different in strip theory and unified theory. The
comparative studies with experimental data for added mass and damping
coefficients in a moderate forward speed range are not conclusive about what
theory shows the most correct forward speed effect. For zero-speed problems there
is no need to base the analysis on slender body theories like strip theory, Ogilvie
& Tuck's theory and unified theory. There are several linear 3-D numerical
methods and commercial computer codes available for calculation of wave induced
motions and loads on stationary ships. The theoretical basis of the computer codes
are the same. Also for very low forward speed or combined current-wave effect
there are practical -3-D miethods available (see Zhao & Faltinsen (1989) for
instance). By low forward speed is meant moeUIg < -0.15. (U = forward speed of
the ship).

For analysis at moderate or high forward speed the situation is more difficult.
Several research groups have developed linear 3-D methods based on the classical
linear unsteady free surface condition with forward speed. This free surface
condition is appropriate in the far-field of a ship and accounts properly for all
frequency and Froude number effects. However, one can question if this is the
correct free-surface condition to use in the near-field of a ship irrespective of
Froude number and bluntness of the hull. Inglis & Price (1981) have developed a
linear 3-D method based on the classical linear unsteady free-surface condition
with forward speed. Froude numbers (Fn = U/(Lg) 1/ 2 ) 0 and 0.25 were studied. The
difference with strip theory was most significant in the low-frequency range.
Nakos & Sclavounos (1990) presented a linear three-dimensional frequency
domain solution, where they modified the classical free surface condition close to
the ship to account for local effects. Their results are promising. The method is
limited to moeU/g > 0.25. A complete 3-D theory like Inglis & Price (1981) or Nakos
& Sclavounos (1990) requires significant computational time which makes it
presently questionable for routine calculations of motions and wave loads on ships.

Chapman (1975) has presented a simplified high-speed theory for a vertical

surface-piercing flat plate in unsteady yaw and sway motion. Chapman's method
shows good agreement with experimental results, while strip theory is not able to
satisfactorily predict added mass and damping values for high Froude numbers.
Faltinsen & Zhao (1991a&b) have generalized Chapman's method to any type of
slender high-speed ships in waves. This is described in the next chapter.

HIGH-SPEED THEORY

Details about the theoretical and numerical method used to analyze the steady
and linear unsteady flow about high-speed non-planing multihulls in calm water
and waves are described by Faltinsen & Zhao (1991a&b). It is assumed that the
hulls are hydrodynamically independent of each other. This is a reasonable
assumption at high speed as long as the hulls are not to close. One can understand
this by analyzing the steady and unsteady wave systems generated by a high-
speed monohull. The problem is formulated in terms of linear potential flow
theory. A numerical solution for the flow around one hull is found by starting at
the bow. The free surface conditions are used to step the solutions of the free-
surface elevation and the velocity potential on the mean free surface in the
longitudinal direction of the hull. The velocity potential for each cross-section is
found by a two-dimensional analysis- Transom stern effects are accounted for by
assuming that the flow leaves the transom stern tangentially in the downstream
direction so that there is atmospheric pressure at the transom stern.

The wave resistance, the steady vertical forces and pitch moments are found from
the steady flow analysis. The latter can be used to calculate the vertical position
and trim. The transom stern has an important effect on the steady longitudinal
force on the ship. A reason to this can be seen by integrating the hydrostatic
pressure force over the body surface below the mean free surface level. Since there
is atmospheric pressure at the transom stem, the hydrostatic pressure force
causes a longitudinal force on the -vessel.

The wave excitation forces in regular waves and the frequency-dependent added
mass and damping coefficients are found from the unsteady flow analysis. By
combining this with information about mass distribution and hydrostatic
considerations the equations of motion in six degrees of freedom can be solved.

Fig. 1 shows the steady wave elevation according to linear theory around a
parabolic strut with length 1 m, breadth 0.1 m and draught 0.25 m. The Froude
number was 1. A comparison is made with thin ship theory. The agreement
between the two methods is reasonable. Since the method by Faltinsen & Zhao
(1991 a) neglects the transverse wave system (see Ohkusu & Faltinsen (1990)), it
indicates that the transverse wave system is not important for high-speed ships.
Roughly speaking the high-speed theory is valid for Froude numbers higher than
0.4. Fig 2 shows a comparison between experimental and numerical values for the
wave resistance of a high-speed monohull. The agreement is reasonably good.
Fig. 3 shows computed results of unsteady wave generated by a high-speed ship.
The high-speed theory is compared with a thin ship theory. A strut with parabolic
waterplane area and wedge-shaped cross-sectional area in unsteady heave motion
with for-ward speed is studied. The length of the strut is 1 m, the breadth 0.05 m
and the draught is 0.2 m. The draught is consiant along the whole length of the
strut. The Fraude number is 1.0 and the circular frequency of oscillation co = 8
radls. The thin ship theory calculations by Hoff (1990) are based on distributing
three-dimensional sources over the center plane of the ship. The sources satisfy
the classical free surface condition with for-ward speed. The agreement with the
high-speed theory is reasonable, but not as good as for the steady flow problem.
It should be noted that the thin ship theory is also an approximate theory. What
the comparison indicates is that the neglection of the transverse wave system is
reasonable at high Froude number.

The high-speed theory accounts for the transom stern effects only in an
approximate way. The reason is that the numerical method only accounts for
upstream effects, i.e. the method has no knowledge that the pressure should be
atmospheric at the transom stern. The predicted pressure distribution will
therefore be in error in a close vicinity of the transom stern.

The inability to properly describe the transom stemn flow will have an influence
on the predictions of the vertical motions. This is illustrated by Fig. 4. There are
two types of theoretical results. In one case there are included transom stem
effects. This was done by using the normal approach up to the station next to the
transom stem. At the transom stern it is used that the pressure must be
atmospheric. This value was used for the whole last station. There is no
theoretical justification for doing this for the whole station. The main purpose is
to illustrate a possible effect from the transom stemn on the ship motions and
accelerations. No special treatment of the local transom stem flow was made in
the other case. The ship model is the same as presented in Fig. 5. The pitch radius
of gyration is 25% of the ship length.

The experimental values presented in Fig. 4 were given by Blok & Beukelman
(1984). We note that the theory is in good agreement with experimental values.
The description of the local flow around the transom stem has small effect on the
heave motion, while there are some effect on the pitch and the vertical
accelerations in the bow. Including "transom stemn effects" in the numerical
predictions improve the agreement with experimental results.

The high-speed theory presented above can easily be used for catamarans since no
hydrodyrnamic interaction between the hulls is assumed. Ohkusu & Faltinsen
(1990) showed reasonable agreement with experimental values for heave and pitch
added mass and damping coefficients of a catamaran. Nesteg~rd (1990) has shown
how the effect of the air cushion can be included and Falch (1991) has shown how
foilcatmarans can be dealt with.
NONLINEAR SHIP MOTIONS

All unsteady theories mentioned above are linear theories. Committee 1.2
of the
10th ISSC has reviewed the state of the art in prediction of strong non-linear
wave
loads on ships. Presently there are no rational methods available. Experimental
results of wave bending moments, shear forces and torsional moments in
steep
waves can show strong influence of non-linearities. The same is true in predicting
large relative vertical motions between the ship and waves, bow flare
forces,
slamming loads and effect of green water on deck. Committee 1.2 of the 10th
ISSC
recommends practical methods to calculate non-linear wave loads. Important
parts
of the methods are exact calculations of Froude-Kriloff and hydrostatic pressure
forces on the wetted hull surface. The incident wave field is described by
linear
theory. The non-linear hydrodynamics forces due to ship motions and
the
diffraction of the incident waves corresponds closely to conventional strip theories
in the limit of small ship and wave motions. The method is therefore
not
applicable to high forward speed and to low frequency of encounter. The non-linear
part of the hydrodynamic forces are reasonable formulations in the case of
water
impact loading on the ship due to large vertical relative motions, but questionable
in the general case.

In order to develop physically based numerical tools for nonlinear ship motions,
many fundamental physical problems have to be better understood. We
will
concentrate on two aspects. Those are the effects of flow separation and the
water
entry (or slamming) problem. In addition we may mention that the water
exit
problem and the modelling of steep (including breaking) irregular waves need
to
be addressed.

The effect of flow separation on the motions of conventional ships

It is well accepted that flow separation matters in describing roll of conventional

ships around resonance, but it is common practice to neglect flow separation
in the
prediction of heave and pitch motions of a ship. However, Beukelman (1980,
1983)
presented experimental results that suggest flow separation can have an influence
on vertical ship motions. This was evident in his studies of a ship model
with
rectangular cross-sections (see Fig. 6) in regular head sea waves. The Froude
numbers were 0.16 and 0.26. As a part of his studies he presented experimental
heave damping coefficients as a function of Froude number, Fn = U/O
, non-
dimensional circular frequency of oscillations )eLD7 and amplitude of forced
heave oscillations rIU3. Here L is the ship length and g is the acceleration
of
gravity. Part of the damping coefficient is due to linear wave radiation damping.
The nonlinear effects can be interpreted in terms of a drag coefficient CD.
This
means one write the vertical force due to flow separation on the ship as

()
2pC A,
Fv, = -_L) dT31 drj
c-3
i
(1T
where p = mass density of the water, Aw = waterplane area, dly3/dt = heave
velocity. By equivalent linearization it follows that

FV3 = _4PCD Aw oel3ad713

ITdt (2)

The drag coefficient depends on the geometrical form, the free surface, the
Reynolds number and the Keulegan-Carpenter number KC = m3a/D. Here D is the
draught. The free surface effect is a function of ceVLuj and the Froude number.
For cross-sections like a rectangular cross-section where separation occurs from
sharp corners, it is not expected any important Reynolds number dependence as
long as viscous'shear forces do not matter. The latter may be true for small
Reynolds and Keulegan-Carpenter numbers and can matter for small models and
laminar boundary layer flow.

Fig. 7 shows Beukelman's experimental results when the nonlinear part of the
heave damping coefficient is interpreted in terms of a drag coefficient. There is a
clear frequency and Froude number effect. The experiments were done for
rl3a/D = 1115, 2/15, 3/15. The data&did not show any important KC-number
dependence. The Reynolds number dependence is not known. Fig. 7 shows also
numerical value of CD obtained by the two-dimensional vortex tracking method
presented by Faltinsen & Pettersen (1987) and Braathen & Faltinsen (1988). No
effects of Reynolds nuniber and viscous shear forces are included. The midship
cross-section was used in the calculations and the Froude number was zero. The
two-dimensional vertical drag force was non-dimensionalized by the beam when
the drag coefficient was calculated. In practice three-dimensional end effects
should have been accounted for. This will result in lower CD-values. The numerical
results show a similar frequency dependency as the experimental values. The
frequency dependency implies that the free surface waves influence the vortex
shedding. There are no experimental results for zero Froude number, but the
numerical values for Fn = 0 are reasonable relative to the experimental values for
Fn = 0.16.

The vortex tracking method can be described by means of Fig. 8. The problem is
solved as an initial value problem. The vorticity is concentrated in thin boundary
layers and free shear layers (SV). The separation points are assumed known.
Outside the vorticity domain a potential flow problem is solved at each time
instant by means of Green's second identity. On the instantaneous position of the
body boundary it is required that there are no flow through the body surface. On
the free surface SF inside ly I = b(t) (see Fig. 8) the exact dynamic and kinematic
free surface conditions according to potential theory is used. As long as the body
surface is nearly vertical at the waterline, there are no numerical difficulties in
describing the flow at the intersection between the free surface and the body
surface. For y > b(t) where b(t) is a large number dependent on time, the flow
is approximated by a vertical dipole in infinite fluid with singularity at y = 0, z =
0. This implies that all waves are inside IYI = b(t). Faltinsen (1977) has shown
in details how the free surface problem can be handled.
The viscous forces in the numerical model are due to pressure forces and can be
related to the vorticity distribution in the free shear layers and the motions of the
free shear layers. Both the vorticity distribution and the motions of the free shear
layers depend on the presence of free surface waves. This is why the CD-values
presented in Fig. 7 are dependent on oevLE"- . However the presence of the free
shear layers do not have an important influence on the free surface waves when
the KC-number is small. This implies that the linear wave radiation damping is
not influenced by flow separation.

The method described above has a clear advantage in analysing separation from
sharp corners at small KC-numbers. The effect of the free surface can be included
in an easy way. Any modes of motion can be studied. However the method has
disadvantages in long time simulations and in describing flow separation from
continuously curved surfaces. In the latter case it is better to use a vortex-in cell
method or a Navier-Stokes solver.

Fig. 12 shows the heave damping coefficient for different amplitudes of oscillations
for the midship cross-section of the ship presented in Fig. 6. The damping force
was written as the sum of a linear term and a quadratic drag term (see Eq. 1).
The linear potential flow damping due to wave radiation was found to agree with
a frequency domain solution based on Frank Closefit method. The results for 13aa/D
= 1115 were used to derive the CD-values presented in Fig. 7. However it should
be noted that the numerical results presented in Fig. 9 show KC-number
dependence. The figure also illustrates that the effect of flow separation can be
large relative to wave radiation damping, in particular for high frequencies. In a
practical context it is frequencies around a natural frequency that is of prime
interest. For the two-dimensional body analysed in Fig. 9 the non-dimensionalized
natural circular frequency of heave oscillations con(D/g)'1 2 according to the
numerical method. The ratio between viscous drag damping at TI3a/D = 1/15 and
linear potential flow damping is 0.09 and 0.29 for respectively w)(D/g) 1/ 2 = 0.588
and 0.835. This indicates that this ratio can be large when r1a3 is the order of
magnitude of D. However when 1 3a becomes that large, nonlinear potential flow
effects may also matter. The ratio between linear potential flow damping and the
critical damping is 0.13 and 0.066 for respectively m)(D/g) 1V 2 = 0.588 and 0.835.
This suggests that the effect of flow separation matter in predicting heave (and
pitch) motions of the ship model presented in Fig. 6 when the frequency is close
to resonance. However in analysing the vertical motions of the vessel in head sea
one should also account for the effect of flow separation due to pitch motion and
the incident waves. Beukelman (1980, 1983) showed experimentally that viscous
effects also mattered for pitch damping, heave and pitch motion. An example on
pitch results for Fn = 0.16 are presented in Fig. 10. The influence of viscous effects
is largest around resonance. For instance from the results in Fig. 10 we find that
 1.09 - 0.34- for cW L'/g = 2.85
15a =D (3)

1.15 - 1.35-Ca for L'g 3

D

For ships with cross-sections without sharp corners the effect of flow separation
will be less important, while the presence of bilgekeels can make the effect of flow
separation more important. Weinblum & St. Denis (1950) presented experimental
results for vertical motions, that showed influence of bilgekeels.

In order to numerically describe the influence of viscous effects on heave and pitch
motion it is necessary to generalize the method presented above to include three-
dimensional and forward speed effects.

Slamming loads

When calculating slamming on hull sections a two-dimensional flow in the cross-

sectional plane is normally assumed. For a general presentation of slamming sea
Faltinsen (1990).

Zhao & Faltinsen (1993) have presented a boundary element method applicable
for water entry of a broad class of two-dimensional bodies and relative velocity and
orientation between th&'bbdy and the water. They have been able to satisfy the
exact non-linear free surface conditions without gravity and at the same time
properly describe the local flow at the intersection between the free surface and
the body surface. Including gravity effects does not represent a problem. The
intersection problem represented the biggest challenge. It was concentrated on the
impact between an initially calm free surface and a two-dimensional rigid body of
arbitrary cross-section. The effect of flow separation from knuckles or from other
parts of the body was not incorporated. It was assumed that the air flow had no
influence on the water flow. The latter means that a body with horizontal flat
bottom or a small deadrise angle (< - 2-30) is excluded. At the intersection
between the free surface and the body surface a jet flow is created. As a first
approximation the pressure is constant and equal to atmospheric pressure through
the jet. This enables one to simplify the problem. The problem is solved as an
initial value problem. The method predicts pressure as a function of time and
space.

Zhao & Faltinsen (1993) compared their method with the similarity solution
developed by Dobrovol'skaya (1969) for water entry of wedges that are forced with
a constant downward velocity. There is no effect of gravity in the similarity
solution. Since no numerical similarity solution results existed for deadrise angles
ca lower than 30', Zhao & Faltinsen (1993) presented numerical results for (Xdown
to 4'. The numerical method was based on the analytical formulation by
Dobrovol'skaya (1969), but a different numerical solution technique was used in
order to handle smaller deadrise angles accurately. The agreement between the
boundary element method and the similarity solution was good.

An asymptotic method was used by Zhao & Faltinsen (1993) for bodies with small
dea~drise angles. The method is based on matched asymptotic expansions in a
similar way, as outlined by Cointe (1991). In the jet-flow region Wagner's (1932)
solution was used. A simple composite solution for the pressure distribution was
presented. The body has to be two-dimensional and symnmetric about a vertical
line. The assumptions are otherwise similar to the boundary element method by
Zhao & Faltinsen (1993). The maximum pressure is the same as Wagner's
formula, but the pressure distribution along the body is generally different from
what Wagner predicted. The asymptotic method presented by Zhao & Faltinsen
(1993) seems to converge towards the results predicted by the similarity solution
when a -4 0. The maximum pressure is well predicted by the asymptotic method
(Wagner's formula) even for larger deadrise angles < - 4Q0'. For instance Cpmax is
only 7% larger by Wagner's formula than by the similarity solution when a = 3'

Wagner's formula is often used in practical calculations of maximum slamming

pressure for any value of ax. However it has no rational basis for very large ax-
values where it clearly underpredicts the maximum pressure. For instance at a
= 810 Wagner's formula shows CPmax = 0.08, while the similarity solution gives
1.16 (see Fig. 11). It should be noted that the maximum pressure occurs at the
apex of the wedge when ax > -45', while at smaller values of ax it occurs at the
spray root of the jet flow. This is demonstrated in Fig. 12, which shows similarity
results for wedges with deadrise angles from 200 to 810. When a is very large the
pressure shows a rapid~change around the apex. Also when a is small, there is a
rapid change in the pressure around the maximum pressure.

One should be careful in applying results for wedges to other cross-sections. The
local deadrise angle is not the only important body parameter. For instance the
local curvature does also matter. Further the assumption of constant body velocity,
does not account for that acceleration may have importance, in particular for drop
test experiments. We will discuss this further by studying slamming loads on a
bow flare section, which have been experimentally examined by Yamamoto et a].
(1985). The bow flare section was inclined a constant angle during the drop tests
to account for in an approximate way the rolling of the corresponding vessel.

Fig. 13 presents comparisons between experimental and numerical values of the

pressure for the pressure gauges P-2, P-3 and P-4 as a function of the time. The
vertical velocity during the experiments and in the numerical simulations are
shown in Fig. 13. The difference in the numerical and experimental velocity in the
first part of the time record is of no importance. The large retardation (about 3 g)
that occurs later on, is of importance. The small discontinuity in the numerical
pressure calculated is due to added mass effects connected with the sudden
retardation 3g. The numerical simulations are limited in time relative to the
experiments. In the last time instant the spray root is at the knuckle. Since flow
separation from the knuckle will matter later on and this feature is - not
incorporated in the numerical method, the numerical computation had to stop. We
note from Fig. 13 that the numerical method predicts well when the pressure
starts to deviate from atmospheric pressure at P-2, P-3 and P-4. The magnitude
of the pressure is well predicted for P-2 and P-3, while the numerical predictions
are too large for P-4. Arai & Matsunaga (1989) have also made numerical
comparisons with Yamamoto et al. (1985) drop test experiments. Their results are
also presented in Fig. 13 and show good agreement with the results by Zhao &
Faltinsen. The effect of gravity as well as separation from the knuckle was
incorporated. The finite difference method based on the Volume of Fluid method
by Nichols et al. (1981) was used to solve the time dependent Euler equations.

UNCERTAINTY ANALYSIS OF STRIP THEORY CALCULATIONS

It is not common practice to make error estimates of ship motion calculations.

However, strip theory programs are often used to find hull forms with optimum
seakeeping qualities. One would feel more confident in optimalization studies like
this if one could ensure that the errors in predictions are smaller than the
varations of the ship motions and accelerations of the family of hull forms that
is studied.

It is difficult to make error estimates of strip theory calculations. A reason is that

we still do not understand properly all physical phenomena associated with ship
motions. However, we will make an attempt and realize that our estimates can be
easily criticized.

An uncertainty analysis will includethree steps

1. List of errors .1
2. Sensitivity of final results to each error source
3. Combination of errors in final results

Errors can be classified as

1. Numerical errors
2. Physical errors
3. Human errors

Numerical errors are errors measured relative to the theoretical basis of the
computer program. Physical errors are errors measured relative to physical reality,
but do not contain numerical errors. Human errors are due to misuse of computer
programs in terms of specifying wrong input or in wrongly interpreting output
from the computer program. It can also be due to "bugs" in the computer program.
Comparative computer program calculations performed recently show that human
errors should not be disregarded. However, we will exclude human error sources
in the following discussion. One way to minimize the possibility of human error
sources is to establish standards for verification procedures of computer programs
and for quality control of use of computer programs.

We will consider strip theory predictions of heave and pitch of a ship in head sea
waves. From the heave and pitch motions we can obtain the relative vertical
motions and vertical accelerations along the ship. The relative vertical motions are
important in calculating wave impact loads, green water on deck and added
resistance due to waves. There are two classes of strip theories. One is named
OSM and the other one is named STFM (Salvesen, Tuck, Faltinsen (1970)) or
NSM. We will refer to results based on STFM. We will assume the two-
dimensional added mass and damping coefficient are calculated by Frank Closefit
method, which is based on a distribution of wave sources over the mean wetted
body surface of a cross-section. Lewis-form technique is also sometimes used. This
is a conformal mapping technique that assumes the two-dimensional
hydrodynamic properties of a cross-section is adequately described by the cross-
sectional beam, draught and sectional area. This approximate representation of
the cross-sectional form causes additional error sources that will not be further
discussed.

There can be numerical errors associated with source function calculation, matrix
inversion and the presence of irregular frequencies. In the following text we will
concentrate on numerical errors due to number and placement of strips and
segments used to describe each cross-section (or strip). This will result in errors
in added mass and damping coefficients, volume calculations, hydrostatic restoring
coefficients, and Froude-Kriloff and diffraction forces and moments. The error will
also depend on what numerical integration procedures are used to sum up the
effects from each segment and strip. Presently the only way to obtain this error
estimate is by convergence studies. That means by systematically increasing
number of segments and strips. The rate of convergence is dependent on the
frequency, wave heading, Froude number, response variable, shape of the body,
numerical integration procedure, assumed variation of source density and velocity
potential.over each segment, and specified choice of segments and strips.

There are different ways to present the errors. One way is to present the error of
a response variable in regular waves as
VWN - f- I
- IfN-LI

where fN is the computed value with totally N strips and segments and fZ is the
value one would obtain if there were an infinite number of segments and strips.
An estimate of fi could be obtained by plotting fN as a function of LIN and
extrapolating fN to LIN = 0. This procedure has the drawback that it provides us
with a large sum of an-values that are functions of the frequency of the regular
waves. It does no directly tell us what the error is in prediction of ship motions in
irregular sea. We have therefore decided to use the following measure of the error
in heave and pitch predictions

Gj--cvj-
(4)

where 1=3 and 5 correspond to heave and pitch. oj is the calculated standard
deviation of heave or pitch in longcrested irregular sea with a given set of strips
and segments on each strip. ciy. is the extrapolated value with infinite number of
 strips and segments. Enj is a function of the wave heading,
the Froude number and
a non-dimensionalized mean wave period . We have presented
calculations of En3 and an5 in Table 1. A modified Pierson-Moskowitzan example on
spectrum was
used to represent the wave spectrum. This is uniquely
defined by the two
parameters HV3 and T 2 , where H1/ 3 is the significant wave
height and T2 is the
mean wave period defined by the second moment of the
wave spectrum. En • is
independent of the significant wave height. In the example
presented in Table I
number of strips are either 20 or 25. In the case of 25 strips
we selected additional
strips in the bow and stern region of the ship relative to
the case with 20 strips.
The distance between each strip is never larger than the
distance between the
stations of the ship. (It is assumed that the ship is divided
into 20 stations).
Number of offset points are either 7, 10 or 19. This means
that total number of
segments on each strip was 12, 18 or 36. The results with
25 strips and 19 offset
points were used as 0 jp. Strictly speaking this is incorrect.
However we do not
think this choice influences our conclusions.

We note that the error is most sensitive to number of strips.

The maximum error
is 0.13 for heave and 0.11 for pitch. This occurs for T 2
2 (g/L)1 = 1.5. For ship
lengths L = 100 m, 200 m. 300 m this means respectively
T 2 = 4.8 s, 6.8 s, 8.3 s.
This example provides a warning that we should be aware
of possible numerical
errors due to placements of strips and offset points. However,
we believe that it
is possible to keep the error due this error source on a sufficient
low level by using
enough strips and offset points.

Physical errors

Physical errors in strip theory calculations will be categorized

as errors due to
linear potential flow effects, viscous effects and non-linear
potential flow effects.

1. Linear potential flow effects

In a previous section we discussed that strip theory is

an approximate linear
theory. For instance it does not properly account for all oscillatory
wave systems
generated by the body, the interaction with the local steady
flow around the ship
and the three-dimensionality of the flow. It was stated
that strip theory is
questionable for high speed problems and for low frequency
problems. Our
discussion concentrated on the calculation of added mass
and damping. We will
now discuss the error in heave and pitch predictions. Since
there is no generally
accepted exact linear theory for seakeeping predictions of
ship at forward speed,
we have to rely on experimental results to estimate what
"true" linear values are.
Experimental values have also errors, but we believe that
carefully conducted
experiments give less errors than numerical predictions.

Gerritsma et al. (1974) presented comparisons between

strip theory and
experiments for heave and pitch of a series of ships in regular
head sea waves of
small amplitudes. The models were derived from the standard
Sixty series hull
form with L/B = 7 and Cp = 0.7. The width was multiplied
by a constant so that
L/B varied between 4, 5.5, 7, 10 and 20. One of the strip
theories they used in
their numerical prediction is quite similar to the Salvesen-Tuck-Faltinsen
method.
We will refer to this method when we talk about strip theory. Numerical errors
are assumed negligible. In Table 2 is presented the relative error in strip theory
prediction at the wavelength where maximum non-dimensionalized experimental
heave or pitch amplitude occurs. This wavelength will be a function of L/B, Fn and
type of response. The mean of all relative errors in Table'2 is 0.12. The maximum
and minimum relative errors are 0.27 and 0.02. The relative error is frequency
dependent and will generally be smaller for wavelengths outside the domain where
maximum heave and pitch amplitudes occur. For small frequencies (large
wavelengths) the relative error is negligible. The reason is that the heave and
pitch motions are very much determined by hydrostatic and Froude-Kriloff forces.
It does not matter that strip theory predicts wrong added mass and damping
coefficients. For very high frequencies the heave and pitch are small and it is
irrelevant what, the relative errors are. The relative error of heave and pitch
predictions in irregular sea will obviously be largest if the peak period of the wave
spectrum is in the vicinity of a period where maximum heave and pitch occur.

The relative errors of heave and pitch in a short term sea state will be analyzed
by studying the standard deviations of heave and pitch. We will limit ourselves to
longcrested sea. The errors in heave and pitch are denoted Es3 and E.5 and are
presented in Table 3. The largest calculated value for both E.3 and E.5 are 0.05.
This occurs at T 2 (g/L)W1 2 = 2.2. For ship lengths 100 m, 200 m, 300 m this means
respectively T 2 = 7s, 9.9 s, 12.1 s.

2. Viscous effect

We have actually already given examples on the errors due to viscous effects when
we were discussing Beukelman's (1980, 1983) experiments (see equation (3)).

3. Non-linear potential flow effects

A strip theory program assumes linearized free surface and body boundary
conditions and use linearized force expressions. We will base our estimates of
errors due to non-linear potential flow effects by analyzing experimental r-esults
in regular waves.

If we assume the wave slope (2 ýa/X) of the incident waves is small, a correction
to oscillatory forces oscillating with the fundamental frequency Coe of the incident
waves can be found from a third order approximation. For instance the pitch
response that oscillates with frequency We can be written as T5 = Tj5 cos(wOt + C5 ),
where

A1 2
5= A ( 2a) (
+
2
(5)

A1 and A2 are functions of non-dimensionalized frequency, Froude number, wave

heading and the shape of the ship. A., is a function also of the above-water hull
form. An important factor influencing A2 is the shape of the hull surface that is
part of the time in the water and part of the time out of the water. If this part of
the hull surface is nearly vertical, it is expected that A2 is relatively small. For a
frequency where the relative vertical motion is large and the ship has a
pronounced bow flare A2 may be significant. Another case influencing A 2 may be
water on deck or if the ship has cross-sections that move out and in to the water.
In the latter cases we cannot outrule that even stronger non-linear effects exist.

O'Dea & Walden (1985) studied experimentally the non-linear behaviour of a ship
in regular head sea waves. The Froude number was 0.3 and XIL = 1.2. Different
bow forms with varying flare were investigated. This is one source of non-linear
behaviour. Another reason may be that the ship had shallow draught at the aft
stations. By fitting their experimental data for the parent form to equation (5) we
found that
115a 1.1 - 200 (_Q)2 (6)

The variation with bow form was not very strong. This indicates that large relative
vertical motions at the shallow draught aft sections may be an important source
of non-linearities. The results for the heave transfer functions did not show a
similar strong dependence on (2(a/X) as the pitch. The results for the .different bow
shapes were more scattered than the pitch results. Equation (6) shows that non-
linear effects can be important. On. the other hand if we wanted to study the
occurrence of deck wetness, this occurred at 2Va?' = 0.02 in the experiments that
equation (6) is based on.. The relative error in strip theory calculations caused by
non-linear potential flow effects is then 0.08.

O'Dea & Troesch (1986) studied the non-linear behaviour of the S7-175 hull in
regular head sea waves at Fn = 0.2 and oM(L/g) y 2 = 2.4. The non-linear effect was
stronger in heave amplitude than in the pitch amplitude. It is difficult to conclude
from these experiments that the data fitted to equation (5). This is partly due to
scatter in the experiments. The experimental values from DTNSRDC showed
approximately 15% lower values for 713a,/a at 2ýa/W = 0.02 than for 2ta/X = 0.01.

STRIP THEORY AND OPTIMIZATION OF SEAKEEPING PERFORMANCE

The errors in strip theory predictions that we have discussed in the previous text
should be related to the sensitivity of heave and pitch to hull form and to how
much freedom one has to change hull parameters in practical design. We will
illustrate this by the example presented by Takaki (1989). He studied
systematically a family of container ships. Bales procedure was used to determine
what ship has the best seakeeping qualities. Takaki presented transfer functions
of heave and pitch for the prototype ship and the new ship. The differences in
transfer functions are largest in the vicinity of heave and pitch resonance
frequencies. For smaller and larger frequencies the differences in transfer
functions are unimportant. We have presented in Table 4 the differences in heave
and pitch transfer functions for the periods where the largest transfer functions
of the prototype ship occur. These numbers should be compared with the
numerical and physical errors in strip theory calculations. We will assume the
errors that we have listed earlier are independent of each other and estimate the
total error as the square root of the sum of the squares of each error. A first step
in the comparison would be to assume that motion responses are so small that
linear theory is applicable. This has relevance when we want to compare hull
forms in moderate sea conditions. We should then combine numerical errors and
physical errors due to linear potential flow effect. We have no way to say
accurately what the linear physical errors are for the ship presented by Takaki.
We will base our discussion on the data in Table 2. Generally speaking we may
neglect numerical errors relative to physical errors. A general trend is that there
are larger differences-in the heave and pitch transfer functions for the new ship
and the prototype shrip, than the errors in strip theory calculation due to linear
potential flow effects. (Compare Table 4 with Table 2). We can conclude similarly
if we study Takaki's results in irregular sea and compare them with the results
in Table 2 and possible numerical errors. A next step in the comparison would be
to consider non-linear effects. We then have a problem in how to use the results
for regular waves in irreg-ular sea predictions. There exists no theory for doing
that. One approximate way is to use a design wave approach. If we compare the
possible errors due to viscous effects (see equations (3) and possible errors due to
non-linear potential effects (see equation (6)) with the results in Table 4, we see
that the error in strip theory predictions may become larger than the predicted
difference in heave and pitch of the alternative huilforms.

CONCLUSIONS

An overview over seakeeping theories for ships is given. Generally speaking strip
theories are still the most successful and practical theories for calculations of wave
induced motions of conventional ships. The limitations of seakeeping theories are
discussed. Strip theories cannot be justified for high for-ward speed, for low
frequency of encounter between the ship and the waves and for large relative
vertical motions between the ship and the waves.

For high-speed monohulls and catamarans it is pointed out that the most
important ship generated waves are due to 'divergent" wave systems. A high-
speed theory that accounts for the divergent wave systems, are presented.

In order to develop physically based numerical tools for nonlinear ship motions,
many fundamental physical problem have to be better understood. lIt is
concentrated on the effect of flow separation and the water entry (slamming)
problem.

It is indicated that the effect of flow separation can matter in the description of
vertical motions of conventional ships, in particular for hull forms with sharp
corners like bilgekeels. A numeri 'cal two-dimensional method that accounts for the
interaction between free surface waves and flow separation is presented. It gives
reasonable predictions for Fri = 0.16, but not for Fn = 0.26. The effect of flow
separation are both frequency and Froude number dependent.

Slamming on hull cross-sections are discussed. Verified results by an asymptotic

method for locally small deadrise angles, a similarity solution and a nonlinear
boundary element method with jet flow approximation are presented. The
intersection problem between the free surface and the body surface requires high
accuracy. The nonlinear boundary element method is a useful tool for a broad class
of hull forms. The local deadrise angle is not the only hull parameter influencing
the slamming pressures.

An uncertainty analysis of strip theory calculations of heave and pitch of ships at

moderate for-ward speed in head sea is given. The most important numerical en-or
source is placement of strips. However, the error can be minimized and neglected.
This requires convergence studies with increasing number of strips. Physical error
sources are divided into errors due to linear potential flow effects, viscous effects
and non-linear potential flow effects. These en-or sources cannot always be
neglected. However, if' strip theory program are used to find hull forms with
optimum s-eake~eping qualities in moderate sea conditions, the physical errors in
heave and pitch predictions are expected to be smaller than the maximum
variations in heave and pitch of a realistic family of hull forms. It is difficult to
conclude similarly for extreme sea conditions.
TABLE 1

Example on relative numerical errors due to strips and offset points in irregular
long crested head sea waves on a ship with CB = 0.66, beam-draught ratio = 2.4,
Fn = 0.21. (G3- and o7,, = "true" values of standard deviations of heave and pitch
amplitude, Cn3 and Fn5 = relative error in prediction of standard deviation of heave,
and pitch respectively (see Equation (1)), T2 = mean wave period, L = ship length).

rT Number Number 30 o 5 ,•L Heave Pitch

of strips of offset error: error:
points H113 Hva3 En3 En5
20 10 0.02 0.06
.1.0 25 7 0.01 0.06 0.00 0.00
25 10 0.03 0.08

20 10 0.13 0.11
1.5 25 7 0.08 0.50 0.02 0.01
25 10 0.00 0.01

20 10 0.09 0.08
2.0 25 7 0.15" 0.74 0.00 0.00
25 10 0.01 0.00

20 10 0.06 0.05
2.5 25 7 0.20 0.74 0.01 0.00
25 10 0.01 0.00

TABLE 2.

Example on relative errors in strip theory calculations of heave and pitch in head sea due
to linear potential flow effects. (Eheave = relative error of heave amplitude at the frequency
where maximum value of 1 3 /rla occurs in the experiments by Gerritsma et al. (1974), E tch
= relative error of pitch amplitude at the frequency where maximum value of U5:&/(a)
occurs in the experiments by Gerritsma et al. (1974), T13. = heave amplitude, 1 5a = pitch
amplitude, • = incident wave amplitude, k = wave number of incident waves).

LJB Fn Eheave Epitch

4 0.2 0.05 0.08

5.5 0.2 0.05 0.14
7 0.2 0.11 0.17
10 0.2 0.27 0.05
4 0.3 0.04 0.05
5.5 0.3 0.19 0.02
7 0.3 0.24 0.15
10 0.3 0.17 0.17
TABLE 3

Examples on relative errors in strip theory calculation of heave and pitch in irregular head
sea due to linear potential flow effects. (T 2 = mean wave period, L = ship length, HM =
significant wave height, ao U = 3,5) = "true" values of standard deviations of heave and
pitch amplitudes based on linear potential flow, s Uj= 3,5) = relative errors in strip theory
prediction of standard deviations of heave and pitch amplitudes

T 2 , 7•l--,1,,H 3 , Cs3
Ss5
1(3 C;5 f

1.4 150 5.8 0.002 0.001

2.2 19 1.5 0.05 0.05
3.0 16 1.9 0.03 0.05
4.0 15 3.6 0.01 0.03
6.0 16 11.0 0.002 0.02

TABLE 4

Example on difference in transfer functions for heave and pitch for prototype ship and new
ship in Takaki's (1989) optimalization study of seakeeping performance (flaaWa =
Maximum value of heave transfer function of prototype ship, flanaJa = Heave transfer
function of new ship at the same frequency as 113ap is calculated, lsaP/(kQa) = Maximum
value of pitch transfer function of prototype ship, TI5aJ(kQa) = pitch transfer function of
new ship at the same frequency as 115. is calculated.

Fn 713ap-'1 3an 7l5ap-T15an

"l3an l15an
0.2 0.11 0.4
0.25 0.25 0.62
0.3 0.34 0.99
 REFERENCES

Arai, M., Matsunaga, K., 1989a, A numerical study of water entry of two-dimensional
ship-
shaped bodies, Proceedings, PRADS'89, Varna, Bulgaria, Vol. 2, pp. 75-1 to 75-7.

Arai, M., Matsunaga, K, 1989b, A numerical and experimental study of bow flare slamming,
(in Japanese), J. of Soec. N.A. of Japan, Vol. 166, Dec. pp. 343-353.

Beukelman, W., 1980, Added resistance and vertical hydrodynamic coefficients of oscillating
cylinders at speed, Report No. 510, Ship Hydromechanics Laboratory, Delft University
of
Technology.

Beukelman, W., 1983, Vertical motions and added resistance of a rectangular and
triangular
cylinder in waves, Report No. 594, Ship Hydromechanics Laboratory, Delft University
of
Technology.

Blok, J.J., Beukelman, W., 1984, The high speed displacement ship systematic series
hull
forms, SNAME Transaction, Vol. 92, pp. 125-150.

Braathen, A., Faltinsen, 0., 1988, Application of a vortex tracking method to roll
damping,
International Conference on Technology Common to Aero and Marine Engineering,
London.
Chapman, R.B., 1975, Free surface effects for hydrodynamic forces on a surface-piercing
plate oscillating in yaw and sway, Proc.1lst. Int. Symp. Numer. Hydrodyn., pp.
333-350,
David W. Taylor Naval Ship Center, Bethesda, Maryland.

Cointe, R. 1991, Free surface flow close to a surface-piercing body, "Mathematical

approaches in hydrodynamics", Editor: T. Miloh, SIAM, pp. 319-333.

Dobrovol'skaya, Z.N., 1969, On some problems of fluid with a free surface, J. Fluid
Mech.,
Vol. 36, part 4, pp. 805-829.

Falch, S., 1991, Seakeeping of foilcatamarans, Proceedings FAST'91, Trondheim,

Norway,
Tapir publishers, Vol. 1, pp. 209-221.

Faltinsen, 0., 1977, Numerical solutions of transient nonlinear free-surface motion

outside or
inside moving bodies, Proceedings of the Second International Conference on Numerical
Ship
Hydrodynamics, Berkeley, California, USA.

Faltinsen, 0., Pettersen. B., 1987, Application of a vortex tracking method to separated
flow
around marine structures, Journal of Fluids and Structures, 1, 217-237.

Faltinsen, 0., 1990, Sea loads on ships and offshore structures, Cambridge University
Press,
Cambridge, England.

Faltinsen, 0., Zhao, R.. 1991a, Numerical predictions of ship motions at high forward
speed.
Philosophical Transactions of Royal society, Series A.
 Faltinsen, 0., Zhao, R., 1991b, Flow prediction around high-speed ships in waves,
"Mathematical approaches in hydrodynamics", Editor: T. Miloh, SIAM.

Gerritsma, J., Beukelman, W., Glansdorp, C.C., 1974, The effects of beam on the
hydrodynamics characteristics of ship hulls. In Proc. Tenth Symp. on Naval Hydrodynamics,
Eds. R.D: Cooper and S.W. Doroff, pp. 3-34, Arlington, Va.;-Office of Naval Research -
Department of the Navy.

Hoff, J.R., 1990, Three-dimensional Green function of a vessel with forward speed in waves,
Dr.ing.Thesis 1990-25, Division of Marine Hydrodynamics, Norwegian Institute of
Technology, Trondheim, MTA Report 1990:71.

Inglis, R.B., Price, W.G., 1981, The influence of speed dependent boundary conditions in
three dimensional ship motion problems, ISP, 28(318).

Keuning, J.A., 1988, Distribution of added mass and damping along the length of a ship
model at high forward speed, Report No. 817-P, Ship Hydrodynamics Laboratory Delft
University of Technology.

Nakos, D.E., Sclavounos, P.D., 1990, Ship motions by a three-dimensional Rankine Panel
method, Proceedings 18th Symposium on Naval Hydrodynamics, Univ. of Mich., Ann Arbor,
National Academy Press Washington D.C., pp. 2140.

Nesteg&rd, A., 1990, Motions of surface effect ships, A.S. Veritas Research, Report No. 90-
2011.

Newman, J.N., 1978, The theory of ship motions, Adv. Appl. Mech., 18, 221-83.

Newman, J.N., Sclavounos, P., 1980, The unified theory of ship motions, In Proc. Thirteenth
Symp. on Naval Hydrodynamics, ed. T. Inui, pp. 373-98, Tokyo: The Shipbuilding Research
Association of Japan.

Nichols, B.D., Hin, C.W., 1981, Volume of fluid method (VOF) for dynamic free boundaries,
J. of Computational Physics, No. 39.

O'Dea, J.F., Walden, D.A., 1985, The effect of bow flare and non-linearities on the prediction
of large amplitude motions and deck wetness, In Proc. Fifteenth Symp. on Naval
Hydrodynamics, pp. 163-76, Washington D.C.; National Academy Press.

O'Dea. J.F., Troesch, A.W., 1986, Comparative seakeeping model experiments, 21st ATTC,
Washington.

Ogilvie. T.F., Tuck, E.O., 1969, A rational strip theory for ship motions, Part 1, Report No.
013, Dept. Nav. Arcit. Mar. Eng.. University of Michigan, Ann Arbor.

Ohkusu, M., Falhinsen, 0., 1990, Prediction of radiation forces on a catamaran at high Froude
number. Proc. of 18th Syrnp. on Naval Hydrodynamics, Ann Arbor, Michigan.
Salvesen, N., Tuck, E.O., Faltinsen, O.M., 1970, Ship motions and sea loads, Trans, SNAME,
78 : 250-87.

Takaki, M., 1989, Effect of hull forms on ship motions and optimalization of hull forms for
seakeeping performance, Journal of the Society of Naval Architects of Japan, 166 : 239-49.

Wagner, H., 1932, Uber stoss- und gleitvorginge an der oberfldche von flossigkeiten, Zeitschr.
f. angew. Math un Mech., Band 12, Heft 4, pp. 194-235.

Weinblum, G., St. Denis, M., 1950, On the motions of ships at sea, SNAME Transactions.

Yamamoto, Y., lida, K., Fukasawa, T., Murakami, T., Arai, M., Ando, M., 1985, Structural
damage analysis of a fast ship due to bow flare slamming, International Shipbuilding
Progress, Vol. 32, No. 369, pp. 124-136.

Zhao, R. Faltinsen, O.M., 1989, Interaction between current, waves and marine structures,
International Conference on Numerical Hydrodynamics, Hiroshima.

Zhao, R. Faltinsen, 0., 1993, Water entry of two-dimensional bodies, J. Fluid Mech., 246, pp.
593-612.

.2,."
 Fig. I The waveelevation around a vertical strut with parabolic water plane area in steady
forward motions. Fn = 1.0. Length = I m. The results by Hoff (1990) based on
ship theory are shown in the upper half, and the results in the lower half basedthin
on
the linear high-speed theory by Faltinsen & Zhao ( 19 91a).

5- BODY PLAN
* 2 I/2-DExperiments
Approach

3.0-

FULL SIZE DIMENSIONS

20 25 30 35 L=24.6m B=2.47m
(.661 (.83) (.99) (1.16) T=1.25m CB=0.54

Speed in Knots "sd(Froude Number)

Fig. 2 Wave resistance for a high-speed hull. Comparison between the high-speed theory
by Faltinsen & Zhao (1991a) and experiments.
 0.043

0."
01

Fig. 3 The amplitude of the wave elevation around a vertical strut with parabolic water
plane in unsteady heave motion with forward speed. Fn = 1.0, 0n = 8 rad/s. Unit
heave amplitude. Strut length I m. The results from thin ship calculations by Hoff
(1990) in the upper half and the results by the high-speed theory by Faltinsen &
Zhao (1991a) using the classical free surface conditions with forward speed in the
lower half.
 - Heave

'1.5000o oo -- Accenatjon

1.1250- 0

0.7500 0

0.3750 -

0.0000 - ----T
0.oo 0.750 .

Fig. 4 Heave, pitch and vertical acceleration amplitudes for the model presented in Fig. 5
in head sea regular waves. Fn = 1.14. Trim 1.620. Experiments by Blok &
Beukelman (1984). ýa = wave amplitude of the incident waves, k = 2t/mA wave
number of the incident waves, a3 = vertical acceleration amplitude, L = ship length.

A*- Experiments
Heave: IT13 theory (the flow at the transom stem gives no difference

in results)
000 Experiments

Pitch: T151 theory without transom stem effects

k(, 0
- - - - theory with transom stern effects
Acceleration 000 Experiments
at station 19 theory without transom stem effects
(in the bow) - - - - theory with transom stern effects
a3 L
50g
 Fig. 5 Body plan of the model used in the experiments by Keuning (1988).

Length of test waterline 2.00 m

Beam of test waterline 0.25 m
Draught 0.0624 m
Block coefficient 0.0396

2,50 m

o,25_1_ 2.00 m 0,25

2,50 m

F _1
_ T 0 ,2 5
1'0.25' 2.00 m 140.25

0,25 7 7 05
0.1-

Fig. 6 Form and dimensions of ship model with rectangular cross-sections (Beukelman
(1980, 1983)).
 Fn = 0, r3a/D= 1/15
6, EXPERIMENTS
(BEUKELMAN (1983)) Fn = 0.16
3.0-0 EXPERIMENTS
(BEUKELMAN (1983)) Fn = 0.26

1.0--
A-

-I0 0
1.0

0.0- L
1.0 2.0 3.0 4.0 5.0 6.0
Fig. 7 Drag coefficients CD obtained from forced heave oscillation tests of the ship model
with rectangular cross-sections shown in Fig. 6. The data are presented as a function
of non-dimensionalizedftfireqiuency of:oscillation cneiLk for different Froude
numbers.

y =-COy
= +39
._-y b y S F y = b(t),,,:

DISCRETE VORTEX
FLUID DOMAIN
Sco S39

SB
Fig. 8 Flow situation around a two-dimensional cross-section performing forced heave
motion with effect of free surface waves and vortex shedding.
 _LINEAR POTENTIAL FLOW
(2D) DAMPING
B33
P:B.D f 6 DAMPINGAT u13a
D
1
15
A DAMPING AT 23a
2
D 15
0

0.3-

0.2
A

0.1 00

0.0 1.0

Fig. 9 Numerically calculated two-dimensional heave damping coefficients B(2D) fo the

midship cross-section of the ship model presented in Fig. 6. The
data are presented
as a funct'ion of non-dimensionalized frequency of oscillation u)eV/5 -for different
7
oscillation amplitudes T3a. Fn = 0. B beam, D = draught.

Fn = 0.16
Fn 0a= 0.02
m
k5a L(a =0.03 m
ki 0 a = 0.04 m
aa
1.0E

0.5

0
0 1 2 3 4
Fig. 10 Pitch amplitudes T15 , in regular head sea waves of the ship
model presented in Fig.
6. Modeltests by Beukelman (1983). L' = 2.333m. (a = incident
k = wave number of the incident waves. wave amplitude.
 20

10

Wagner jet flow solution

Similarity solution

100 200 300 40 ° 50 0 7b0 80 °

Fig. 11 Prediction of maxiibbm pressure coefficient CPmax during water entry of a wedge
with constant vertical velocity V by means of similarity solution and Wagner's jet
flow solution, a = deadrise angle.

a a 200
------------ a = 250
......... a- 300
16.0- PP .... a = 40*
2
0.5pV . a= 450
a = 50*
12.0 ------- a = 60 °
...... a - 70 °
a = 810
-------------- ----------------

4.0 - .........i lia..

- ~~L -z-- - - -
0 .01 .............. . .. M-.•

-1.0 -0.5 0.0 0.5 1.0

Fig. 12 Predictions of pressure (p) distribution during water entry of a 2-D wedge with
constant vertical velocity V by means of a similarity solution (Zhao & Fahrinsen
(1993)). p0 = atmospheric pressure, p = mass density of the water.
 - Measurements

•° Theory
Theory (Zhao
(Arai &&Matsunaga)
Faltinsen)
P6

,,. ' ~ ~~P-3.• -"..

0~.

Is))

w) 0

-~~-~a- 0.01
V (ms) .01 ec.TIME
4.0 ec

3.5

Fig. 13 Comparisons between numerical and experimental pressure measurements on bow

flare section. The experiments are drop test results by Yamamoto et al. (1985).
NEW TECHNIQUES FOR ASSESSING AND QUANTIFYING VESSEL STABILITY
AND SEAKEEPING QUALITIES.
Organised by European Community University Enterprise Training
Partnership and the University of Trondheim.
March 8 - 11. 1993. Trondheim, Norway.

BREAKING WAVES, KINEMATICS AND FORCES.

by
Soren Peter Kjeldsen.
The Maritime Academy.
Trondheim, Norway.

ABSTRACT.
The investigation deals with breaking waves in deep
waters of spilling, plunging and surging type. It describes
certain areas that have dangerous waves during certain well known
meteorological conditions. Further criteria for inception of
breaking are given related to the shape of the
steep waves and finally examples of crest kinematics and slamming
forces in steep waves are given.

1. INTRODUCTION.
Breaking waves give the most difficult conditions for
manouvering of smaller ships. Therefore the catamaran was
developed on the Coromandel coast in India in order to make
a small vessel with good stability that could be rowed from the
coast throug the surf and into the ocean. Further the outrigger
canoos on the Pasific islands were developed for the same
purpose. All sailors know that waves can appear With vertical
fronts and even overhanging crests. Further it can easily be seen
that the kinematics in the upper part of the crests exceeds the
phase velocity of the waves when a breaking
event occurs. W.'ater particle velocities are therefore larger
in breaking waves than in similar non-breaking waves with.
comparable dimensions and wave forces are therefore larger in
breaking waves. Further in order to describe wave kinematics in
breaking waves we must to destinquish between 3 types of
breaking waves. These are spilling, plunging and surging breakers
and all types can occur both in deep and shallow waters.

2. DANGEROUS AREAS FOR NAVIGATION.

There are a large number of capsizings and heavy weather
damages recorded worldwide. Capsizings are not restricted to
smaller vessels but also quit large fishing trawlers and
freighters have capsized. The british trawler "GOAL" lost with
all members of the crew in Northern Norway is just one example.
On the exposed Norwegian coast 26 norwegian vessels were lost in
only 9 years. 72 people lost their lives. In addition to that we
have numerous examples of near accidents with sliding cargoes and
damaged stability caused b,'"a steep wave. A norwegian research
program "SHIPS IN ROUGH SEAS" was therefore initiated, and it was
found that accidents
and near accidents occur in certain areas. It was then shown
that certain norwegian coastal areas have wave conditions that
are more severe and more confused than deep water conditions a
few nautical miles further off. it was shown that these
dangerous wave conditions are caused by large scale wave
refraction, leading to focusing of waves, see KJELDSEN,LYSTAD &
MYRHAUG 1981. The refraction is in most cases a combination of
wave-current refraction and topographic wave refraction. 24 areas
with dangerous breaking waves have been identified and these are
shown in Fig 1. Sea states containing crossing wave trains from
several directions simultaneously occur here during certain well
known meteorological conditions. These areas are therefore marked
on norwegian sea charts and described in the Pilot for the
norwegian coast.

2.CRITERIA FOR INCEPTION OF BREAKING.

Certain definitions are needed when we deal with steep and
breaking waves in an irregular 3-dimensional sea. Here we
refer to Fig 2. In our studies we have defined a freak wave as
a wave with a wave height that exceeds 2.0 times the significant
wave height, KJELDSEN 1990 (a). It is nessessary to check both
the zero-downcross wave height and the zero-upcross wave height
in order to determine whether a certain
irregular wave is a freak wave or not. 10 dangerous freak waves
were measured in Area 16 with wave radar.
It has been usual to use the wave steepness H/L in order
to determine whether an irregular deep water wave is breaking or
not. However extensive series of experiments performed- at
MARINTEK showed that irregular waves could break before the
theoretical limit 0.141 was reached, KJELDSEN 1983. In an
irregular sea many waves could break at an intermediate steepness
near 0.05, while in the same sea other waves with the same
steepness H/L not were breaking. Therefore we introduced the
crest front steepnes defined both in time and space as shown in
Fig 2, and we developed a counting technique
for successive trough/crest events as they could be observed in
time series from wave radar measurements and established
joint probability density distributions of crest front
steepnesses and wave heights. An example is shown in Fig 3.
Applying this method it became evident, that most data grouped
in a quite regular pattern, but one cycle representing an unusual
condition suddenly dropped outside of this regular pattern during
some gales. This unusual wave was found to be a freak wave.
Alltogether 10 freak waves were identified in Area
No 16. We believe that most of them are steep irregular non-
breaking waves. However in 2 cases we found two freak waves
following each other in the same wave group. Such an encounter
could of course be dangerous for a small ship in particular if
the period of these two waves are close to the rolling period.
The first freak wave could then cause a large roll and the second
might lead to a capsizing. One of the time series containing 2
freak waves are shown in Fig 4. An example of a single freak wave
is shown in Fig 5.
A non-linear wave generation technique was made in order to
reproduce sush waves in the Ocean Basin, KJELDSEN 1982.
A series of tests were performed with non-i inear transient
wave groups leading to a sincgle elIevated ifre;a k wavw, or to two
freak waves fol lowing each other. An e>:;zmp e ics shown in
,iq G. Frl•om these tests performe(J at. MARINTIEK it was Found tihri t
incept ion ol breaIk i lq OCCIlrl-od whell the synopt: ic crest fIront
:;teepnes off the waves I)assed a i imit: o 0.3 . Then -I spi I I i 1)g
-reakr- (ŽVe IoI)ei. Crel-st [rolit :.:t iel:-;c. e-tr measured Wi th
values up to 0.78 and such high values gave plunging breakers.
A crest front steepness 0.78 corresponds to an angle between
the
synoptic crest slope in front of the wave
and the mean water level near 38 degrees, see Fig 2.
The horizontal asymmetry factor for breaking waves is close
0.77. These results are in agrement with later experiments to
performed in other wave tanks.
When these experiments are compared with wave radar data
observed at sea we came to the conclusion that the first wave
shown in Fig 4 is a wave very close to breaking or actually
allready breaking as shown by the small irregularity at the
top
of the crest.
Stability experiments with a large number of ships have been
made at MARINTEK in deep water plunging breaking waves.
If a capsizing is recorded the next problem is then to quantify
the encounter probability for this particular breaking wave
in
the sea area where the ship shall operate.
This encounter probability computations are made from joint
probability ditributions of wave heigts and crest front
steepnesses measured with wave radar using plots of the
type
shown in Fig 3. The method is outlined by KJELDSEN 1983.
The concept of crest front steepness and wave asymmetry
factors as defined in Fig 2 has been recommended for
international use by I.A.H.R. and it has since been applied
by
several wave tanks and oceanographic institutes in their analysis
of wave data and related responses of floating objects,
"see
I.A.H.R. 1986.

3. WAVE KINEMATICS AND SLAMMING.

Another criteria for inception of breaking that has been
used is that the wave particle velocity in the upper part of
the
crest exceeds the wave phase velocity. Therefore a particular
experiment was performed in which crest kinematics
were measured in the plunging breakers, see KJELDSEN,VINJE,BREVIG
& MYRHAUG 1980. The next problem is then to assess the
wave
forces and moments a steep wave is able to create on a floating
structure. In particular wave slamming
is a well known problem in seakeeping applications. Wave slamming
can in some cases be a problem in the tunnel between the
two
hulls on a catamaran. We investigated wave slamming
in a particular experiment in which a breaking wave interacted
with a tilting plate installed at wave crest level. The crest
front steepnesses were varied in these experiments and we found
that the maximum slamming pressure occurred for a particular
value of the crest front steepness, see Fig 7. In the evaluation
of these results we used an algorithm that predicts wave
crest
kinematics and found that the case that gave maximum slamming
pressure corresponds to a situation in which the interior
particle velocities are nearly perpendicular to the tilting wave
plate, see Fig S.

4.WAVE-CURRENT INTERACTION.
Recent research has shown that wave-current interactionls
is particul a: important in the mechanism that leads to the
genferation ca plunging breakers in deep waters at sea.
Areat 'o. .13 : Jis partiCul ar difficult. Hel-re we have a
topographic "o-rnor ef:fect giving strong winds and here
not less
thlmr 71 :;h ijp:; have capsized and 28 people have beŽeni loýsi c rlce
1956 see Fig 9. All seven ships were lost in a small gale,
very suddenly and unexpected the ships encountered and
a steep
breaking wave. Sometimes this wave occurred in the middle
of the
shipping lane. The pattern was the same in all these accidents.
The last accident was the loss of "SUN COAST". Directional
wave
bouys were then installed on the site and wave refraction
computations were performed. A strong focusing of wave energy
was
observed created by the particular topography in the area.
A new
shipping lane has therefore been made in this area that
brings
the vessels further offshore in
normal wave conditions. However in this particular accident
the
investigations performed by the Court of Inquiry showed
that
large core rings travelling in the current along the
coast
contributed to creation of dangerous wave conditions,
SIEVERAAS,KJELDSEN & NRROY 1988.
A better traffic control along the norwegian coast is
certainly needed also in other areas and it has been discussed
to require that foreign vessels that travels in norwegian
waters
must have a pilot onboard.

5. CONCLUSIONS.

5.1- A new shipping lane was made in Area No. 13.

The 24 dangerous areas are now printed in norwegian
navigational charts and a description of dangerous wave
conditions is given in the Pilot for the norwegian coast.

5.2- Use of the crest front steepness parameter and

related parameters for vertical and horizontal asymmetry
is recommended for international use by the International
Association for Hydraulic Research.

5.3- Further research on algorithms that predicts

kinematics in the crests of steep and breaking non-linear
waves
is needed.

6. REFERENCES.
International Association for Hydraulic Research 1986:
"List of Sea State Parameters." Supplement to Bulletin
No. 52
(1986), I.A.H.R. Brussels, Belgium.
KJELDSEN,LYSTAD,MYRHAUG 1981:"Forecast of Breaking Waves
along the Norwegian Coast." The Norwegian Meteorological
Institute and Norwegian Hydrodynamic Laboratories. Technical
Report.
KJELDSEN 1990a:"Breaking Waves" in Proc. from NATO Advanced
Research Workshop on WaterWave Kinematics. Molde, Norway
22-25 May. Kluwer Academic Publishers, The Netherlands.
KJELDSEN 1990b:"The Practical Value of Directional Ocean
Wave
Spectra" in Johns Hopkins APL Technical Digest. Vol 11 No
3 and
4. July-December. The Johns Hopkins University. U.S.A.
KJELDSIN 1982:" 2- and 3-dimensional Deterministic Freak
Waves"
in Proc. 18th International Confe'mrence on Coata] Engineer-ing.
Cape Town, South Africa.
KJEIDSEN 19S3:"Dete•r:-iination of Severe Wave Conditions for
Ocean Systems in a 3-d imens;ioanal ! regul.-Ir Seaway." in Pro;.
1ram the 8th Conqrezs of can Inst iut:e of
('an-Amei
1jval I
ng.i necŽi :.incj. 12-17 September.
NJ I':l.IS{EN , V [ NJ I',M YRIIAIJC , IHi.:Vi1 ; Waslh i rigLton 0C. U.S.A.
I98•0 : "''Ninemal
In C•::; o1" Jeep[' L,I tear-
Breaking Waves" Paper No. 3714 in Proc. 12th Offshore Technology
Conference. Houston, Texas.
KJELDSEN- 1981:"Shock Pressures from Deep Water Breaking Waves"
From Proc. International Symposium on Hydrodynamics in Ocean
Engineering. Vol. 1. Trondheim, Norway.
SIEVERAAS,KJELDSEN,NgEROy 1988::"A Report from the Court of Inquiry
after the Loss of "SUN COAST" at Stad 2 dec 1984."
Department of Justice. Norway.
VINJE & BREVIG 1980:"Breaking Waves on Finite Water Depths.
A Numerical Study."Report The Norwegian Institute of technology
and Norwegian Hydrodynamic Laboratories.
 3
2

NS6nu

9i Area Area
•number name
10 • 1 Vardo
11 2 Tanafjorden
3
•':.• Nordkinn
12• . 4 Breisundet
I; -' 5 Lopphavet
13. 6 Sveinsgrunnen-
14 I IIMalangsgrunnen
7Lofoten
S8 Vestfjorden
.
-- 9 Folla
10 Gripholen
18 •Nj- 24 11 Hustadvika
22 12 Godo
19 13 Stadt
21 14 Buefjorden
0 23 15 Sognesjten
16 Vikingbanken
17 Sierna
18 Skotamedgrunnen
19 Sirearunnen
20 Listafjorden
21 Lista
22 Ryvingen
23 Skagerak Vest
24 Faerder-Tvistein

Fig 1. Twenty-four areas on the Norwegian coast that are

the sites of steep and dangerous waves during certain
meteorological conditions depending of wave direction, wave
height and the position of coastal eddies- ( From Kjeldsen
1990b. )
 DEFINITION OF WAVE DIRECTION 8BFOR A SINGLE BREAKING
WAVE:

yt

P4
DEFINITION OF CREST FRONT STEEPNESS rx IN
SYNOPTIC DOMAIN: B
a for p.-Position
Inception of breaking
E:X.s " ' MW,• . ..

DEFINITION OF CREST FRONT STEEPNESS ct IN TIME DOMAIN:

TIME MWL

2t T.Tzd
Tzd

DEFINITION OF CREST LENGTH v. AND 3-D CREST SHAPE FACTOR i

IN SYNOPTIC DOMAIN:

MWL

VERTICAL ASYMMETRY FACTOR HORIZONTAL ASYMMETRY FACTOR

X -_ IrI

HH

U V

Fig 2. Definition of direction, wave crest length,

crest front steepnes and wave asymmetry factors for a
3-dimensional breaking wave in a directional sea. ( From Kjeldsen
1990a.)
 '1;

0.00 0.06 0.10 0.16 0.20 0.26 0.60

CREST FRONT srEEPNESS

Fig 3. Joint probability density distribution for

crest front steepness and wave height measured with waveradar
in Area 16 on 85.11.06 at 06 GMT. The significant wave height was
8.49 m. ( From Kjeldsen 1990a.)
 IC]

10 - .

O-

-- S8O O9GO 10 000 1080

SECOC40S
TIIIE

Fig 4. Two freak waves following each other

in the time
series of surface elevation found by the half-cycle
counting technique applied in Fig 3. ( From
Kjeldsen 1990a.)
 9 '

0u C-4JN
ud
001

* ~Un-

4W0~

0%44

C) 4-

N W4J0

0) I

'4 WCHr-4

U) Ojo)

C) jo
o r46J~d'f

E-4

tp u

a)00

CmY C)
DISTANCE FROM
WAVE GENERATOR: TEST NO: 214 B R-2
S.L. BUNN B.S Comm 801
O.

X = 10 M ._-. Z
0
I. 20. 10. 60. 80. too.
rIErCSEC.3

TEST NO: I14q DP-2 S.L. BUNN 8.5M COMM 802

r 2 5 sm.r- 0.3
.-
I eill\ ,
-0.)

0.4
0. ,00. * . go.
8O 0.
TllCESEC.j3

x 26 TEST NO: ]01 08-2 S.L. BUNN 8.5M COMM 802

0.2I
x = 26rn .
""llf]

0. 20. 0. 60. 60. 100.

TEST NO: 114 OM-2 5.L. BUNN 8.SM COMM 801

~ -0.23_ _ _ _ _

x = 35 m.
.

-0.0.I_____
"&

I ________________.]_

lESt No: I114 88-2 S.t. DUNN O .SM COMM 802

x 60 TO. 0.4 _

0.2

0. Ž0. *(0 cc. 00. 00.

M I
T I"( (S I

Fig 6. Example of non-linear focusing of many transient

wave components leading to one single freak wave
in the Ocean Basin comparable to the event measured
by wave radar
in Fig 5. ( From Kjeldsen 1990a.
 0=45'
L" L'7- --

3 1
scm --Rise
z 2 9time
milles•.

•5:O0 RUN £.8

Lewi: IHp-z]K±0.L60

0= 45 degrees

CLSPILLING BREAKER
U '00
1CX)

50 100 ISO 200 250

TIME Ifhllsec I

E)=45* Tzd -
FP',-_pg I Hp - zO) 1.00 sec
W p-.gHp A 1.19 sec
0 1.35 sec

nub3e0 BREAKING WAVES

.- •t.clonumber of/

6 '::served
12.0 _:ecks: 62
shock

01 Uý

7I NON -BREAKING WAVES

of

30 t40 5.0
IFFPINESS RiATIO

Fig 7. Example of shock pressures recorded onl a tilted

plate in the crests of non-breaking and breaking waves.
The shock pressures were found to increase with increasing crest
front :;teepnes;s. ( Yrom Kjcids;en 1981t.)
.6

Ci-

kg

.2 -- •.

0.0
.4 . .0 kx .2 .

Fig 8. Numerical simulation of the kinematics in the

crest of a steep breaking wave. ( From Vinje 1980.)
 -id . umr 93-

L - if.

- - - i
AUM ~~ ~ ~ =)f. i ~ igkr

L~za. 19.bt.

if -ý -W m

1- br. HIesnipt

.- R.. .

t. -~VVk~ig

*~5123 9- sai 4 Trlv9 9bn.

LI2k, .Utt iv SV~kuing.

-- -r

... _ F1..

.LSa~ fSisi Are No 13 s ainc 195ku6-

7 fteelosswr I' t&.tri7 in a. modera efwind

steep~S~l wae.0pol otterlvs -one 3of the~k

cazsig ocure -A atul ae rm a-easKede
& 1998.) t'""y I~SVtjig
 LECTURE 1

The principles of ship stability

T. Nedrelid
MARINTEK. 9 February 1993
LECTURE 1

The principles of ship stability

The following items will be presented

0 The principles of hydrostatic stability,

definitions.

0 Excisting rules and regulations based upon

hydrostatic stability.

* Experience of excisting stability

rules/investigation into earlier accidents.

0 Examples of "Stability research".

* Dynamic-situations.
 STABILITY

General description / Definition

TRlE fltLlT/ OF P9VPS6SFL Tb R764*Wh WULEJ

f~a~tDuo inSL & W1

Static Stability

WI,

mii iii I\ob,(J3/rhI

THE PURPOSE OF STABILITY STANDARDS

TO AVOID HEEL

TO AVOID LARGE ROLLING

TO AVOID CAPSIZING

High freeboard ship/large superstructure are exposed

to forces from wind.

* Rolling in waves

* Rudder forces/manoeuvring

* Loading of cargo

rIo U-
Initial Stability/Definition of terms:

Fig. 2

Definition of terms:

A - GZ Righting moment

GZ Righting lever

GM Metacentric height

For a stable equilibrium situation, GM is positive, ic. M

lies above G.

€.S
COMPLETE STATIC STABILITY

Certain features of the GZ curve are of particular

significance and are useful parameters with which to define
the stability possessed by a given design.

The international stability rules and requirements are based

upon this GZ'-curve and these parameters

C?

A a
FP E
The slope at the origin represents the metasentric
height.
Max. GZ is proportional to the largest steady heeling
moment that the ship can sustain without capsizing.

The GZ-curve shows the range of stability

The area under the curve represents the ability of tile
ship to absorb energy imparted to it by wind, waves or
any other extern agency.
* Beam winds rolling

la

wind heeling arm curve

* Lifting of cargo

C#$&40 fl26uwN4UMoM~kjy

All stability discussions are rotating around GZ-curve areas,

max GZ angles, angle of vanishing stability.

90 % of all stability research has l)een consentrated on how

to apply GZ-curve parameters related to various external
hazard-situations. Some of them have totally been out of the
range of "dynamic".
 03
0-

NJJ a a

LL tn _ _

< 0 w
Ow 0
> 0 0 I

-: 03

Z
-
QZ
4< 0<<n
M
F-U)ZI -
0 Ln .0
t
a 1 0
z~ 0
0 _

~ 00

000. -

0~_ X

LI
_ _ _ _A _

___

LL__ o_ 0 _ 0(

0 0 0 0 0

"4ý
JAPANESE LIST 1980

Table 1 Classification of Flooding and Capsizing Accidents of Ships

Case Causes or Conditions of Casualties Number of Casualties

Fishing Boats Cargo Vessels Total

1 Navigating in Quartering or Follow- 25 (19)* 26 (19) 51 (38)

ing Seas
2 Navigating in Head, Bow and Beam
Seas 49 (37) 101 (64) 150 (101)

3 Navigating in Calm Water 5 ( 5) 15 (15) 20 (20)

4 Working as Fishing or Towing Ship 5 ( 5) 25 (20) 30 (25)

5 Hull Break Down 17 (1) 37 ( 5) 54 ( 6)

6 Mishandling of Piping or Valve System 24 C 3) 22 ( 5) 46 ( 8)

7 Anchoring in Harbour When Storm or 12 C 3) 12 ( 4) 24 ( 7)

Typhoon

8 Misleading of Cargo 3 (3) 47 (45) 50 (48)

9 Icing or Drift Ice 8 ( 8) 0 ( 0) 8 ( 8)

10 Reasons Other Than 1-10 5 ( 2) 10 ( 6) 15 ( 8)

TOTAL 153 (86) 295 (183) 448 (269)

Note: Number in parenthesis indicates the number of capsizing accidents.

Table 2 Classification of Main Factors Caused on Capsize

Factors Caused on Capsize Fishing Cargo Total

Boat Vessel

Over Loaded 9 8 17

Top-Heavy 13 11 24
Insufficient Lashing 1 20 21

Inferior Loading 7 17 24

Cargo Shift 13 41 54

Open Door 9 22 31

Inferior Hatch Cover 2 18 20

Ifull Break Down 2 1 3

Ship~,inq water on 1ecR 29 26 55

B ro;ach inq 1 3 4

TOTAl, 52 87 139
THE PRINCIPLES OF STABILITY REGULATIONS

The main purpose of stability regulations is to maintain the capsizing

probability of a vessel on a low acceptable level.

Capsizing implies that the vessel has been the subject of external
forces which has turned the vessel over to a large angle where the
vessel remains in a new stable position. The outcome may be:

The vessel remains in a stable side position (Fig. c).

Water ingress will often start, and the vessel will turn over
1800, or sink.

The vessel turns directly over to 1800 (Fig. b).

The vessel turns over to a large angle (600 - 900) and returns to
an upright position (Fig. a). If cargo shifts due to the large
heel, the outcome may be:

The vessel obtains a list, or possibly a more critical stable side

position

The vessel (which may otherwise have GZ-curves of the type

shown inFig. a or C)tIirns over to 180".
STPBILITY- CRPSIF_E

00 1Oo
STABLE

a. Self-righting.
Capsize not feasible

STABLE STABLE

b. Capsize to 180 /

STABLE !STABLEý STABLE

c. Capsize to stable side position

Fig. 1. Definition of capsize.

~ MARINTEK
EXPERIENCE / ACCIDENTS IN NORWEGIAN WATERS

Several investigations were performed mapping accidents

along the Norwegian coast caused by "Clean stability and
weather conditions" during various periods.

This example shows losses during the period 1970 - 1977.

New examples are still beeing added, each 1 - 5 examples
per year.
The investigations into the history of accidents
will not always give very clear answers to what
was the actual cause of the capsize.

The weather, the loading condition of the actual

ship and other human factors are interacting in a
way that makes it difficult to point out single
technical answers.

This next figure shows that in five (5) cases the

standard stability requirements seems to be OK,
even if these vessels capsized in waves.
GENERAL CONCLUSIONS FROM INVESTIGATIONS
INTO CAPSIZING ACCIDENTS IN ROUGH WATER:

Capsizing is dependent of:

the vessels stability property

the weather and wave situations
the vessel respons in the actual wave situation

Total loss is dependent on:

do cargo shift ?

is the vessel water tight at great angles of heel?

floating ability in a damaged condition
human reactions in critical situations
FIRST LEVEL CONCLUSIONS

Vessels are capsizing in waves due to large motions and

dynamics

Present criteria for stability are insufficient to prevent

this ?
Dynamics are not reflected in the hydrostatic based
stability criteria ?

Can we find new criteria so strong that capsizing will

never happen in the vessel's total operational life ?

Do we accept a certain probability of capsizing to

happen - if only this figure can be calculated to be a
low figure ?
STABILITY RESEARCH EXAMPLES

Mapping of heeling forces wind / rudder / green water

on deck

Definitions of waves as quasi-static approaches

Design parameters against the various heeling forces

Dynamic approaches
Action of steady rolling
Equation of motion (linear theory) with variation
of righting arm constant / position dependent /
time dependent (parametric resonance) and
damping
Maximum rolling amplitude at resonance
Rolling in various sea conditions, directions of the
sea ie.
The effect of' wave and motion on static stability
(following waves)
Capsizing studies in waves
The influence of rolling motion on staIbility
crit eria
RESEARCH PROGRAM "SHIPS IN ROUGH SEAS"

* BREAKING WAVES, FULL SCALE MEASUREMENTS.

* BREAKING WAVES, THEORETICAL MODELS.

* THEORETICAL MODELS FOR EXTREME MOTIONS

* ANALYSIS OF EARLIER ACCIDENTS.

* CRITERIA FOR STABILITY

 THE NORWEGIAN RESEARCH PROJECT

STABILITY AND SAFETY FOR VESSELS IN ROUGH WEATHER

MAIN SUBJECTS:

§ SHIP AND CARGO

§ ENVIRONMENTAL DESCRIPTION

§ STABILITY CRITERIA

N~MAINTEI
 PROGRAM "STABILITY CRITERIA"
RESEARCH

* DEFINING CRITICAL WAVE/VESSEL-SITUATIONS

VARYING TYPE
* SURVIVAL TESTS IN WAVES WITH
OF VESSELS

* DEFINING THE PRINCIPLES

OF FUTURE
STABILITY CRITERIA

-- MARIN M-
 ON THE RELATIONSHIP OF SHIP SURVIVAL TESTING TO REGULATION

by

Terje Nedrelid

Bibliography

Terje Nedrelid is a Naval Architect with a M.Sc. from the Norwegian Technical University of
Norway of 1971.

Industrial experience as naval architect in the design of small vessels (1971-76).

Project Manager for several large Norwegian Research Programs: Ships in Rough Seas
(1979-81), Stability Criteria (1981), Stability- and Safety of Small Vessels in Extreme
Weather Conditions (1981-82), Offshore Marine Operations (1985-88), FPS-2000 -
Floating Productions Systems, Positioning and Mooring (1989-91).

Since 1986 Division Manager at Marintek, Div. of Offshore Floating Structures and as
of 1991, Division Manager of the Marintek Laboratories as well.

ABSTRACT exploring the sensitivity of stability

The purpose of stability regulations is to
maintain the capsizing probability of a vessel
at a low acceptable level. The present
regulations are based upon a statical
approach with some semi-empirical
adjustments for dynamic effects. In our
attempts to improve regulations, we are
discussing new probabilistic approaches and
even operational "stability" as a possible
future basis. From a theoretical point of
view, the phenomena influencing on a vessel
capsizing in waves often proves to be a
combination of very complex hydro-dynamic
events. This fact has lead to a search for
new and improved stability regulations that
at the same time are more and more
complex in formulation and interpretation.
regulations. This way, we can look for even
simpler regulations.
The total safety margin of a vessel is related
to the sensitivity of the vessel design against
capsizing due to mal-operation. Survival
testing can be used as a tool, by which to
determine this sensitivity. A. number of
vessels, some of these new and modem
ones, have capsized due to maloperation and
the following water ingress through open
doors and hatches. This could be too
complex a situation for theoretical analyses
only. The effective performance of survival
testing in advanced moderne laboratories in
combination with theoretical studies, will in
our opinion bc the tool of tomorrow.

This pap~cr will take up to discussion some REGULATIONS

diffuhrcnccs in
the variOus stability
riguilations, C[4. 11hstate of [ie alt oif the Statical st;ib ilitv criteria
th•olrtical mcthtodoloigy. B3y the use of'
cx;iIll)les we will shlow how sur,'ival testing C(i)sizing implies thiat the vessel has been
Of shl1 ) FiB dels in an advanICeLd la)bortory the stubjet:c ( textCrnlal forces which have
faCilily :;llb use (I aas a11 uluiiii ic t[010l iIll Inllie(d thea vue,,'l over to a large angle where

*':\•ck5 I ;i);ti)i:r lut'l.ln\rhX2' J; $,,U I')' 3

% Iv 1
the vessel remans in a stable position. The Most countries refer to the static stability
outcome may be: regulations proposed by the IMO
organisation. These regulations are founded
The vessel remains in a stable on the balance between the vessel weight
position. Water ingress will often and the hydrostatic forces when inclined in
start and the vessel will turn over still water. We refer these regulations to the
180 deg. or sink curve of statical stability (GZ) (see Fig. 2)
and make various comments around the
The vessel turns directly over to 180 slope, range and form of the curve. We
deg. make a comparison of various areas below
the curves.
The vessel turns over to a large angle
(60 - 90 deg.) and returns to an This type of curve (GZ) has proved to be a
upright position. practical way of judging stability of a
particular vessel. It is used in the designing
If cargo shifts due to the large heel, the phase of a vessel as well as in the operation
outcome may be: of a vessel.

The vessel obtains a list, or possibly In years past, the hydrostatic stability
a more critical stable side position technology implied large costs and great
effort in order to produce this type of
The vessel (which may otherwise stability information. Today the technology
have GZ-curves of the type shown in is computerized, well known and can easily
Fig. la or Ic) turns over to 180 deg. be performed onboard a vessel. This could
be the idealistic world of stability regulation,
using only 'the static righting arm (GZ)
terminology. However, capsizing occurs due
to external forces acting on the hull. The
* STABLE I forces of wind, waves and current are all
a. Self-righting, dynamic, often nonlinear and of a complex
Capsize not feasible statistical nature and they all impose a
dynamic motion upon the vessel. Sometimes
this coincide with weight shifting and water
,Tz •ingress onboard the vessel structure which
ultimately makes the stability situation very
'.
0o o complex to understand.
* STABLE STABLE

b. Capsize to 180' The definition of stability regulations are

consequently influenced by the under-
standing of the external forces. The
traditional statical approach of stability
regulations using GZ curves deals with some
6" j..j of the dynamic phenomena. By using a
•i3 •"scilidynamic approach we call deal with this
S .( -1AOL prl~nlplei. ("The dynamic stability of a ship at
c. C1:apsizet llt~la:;:ide po:fitiorcsmdnmi - given angle of)ntal
heel is defined as the work
done in heeling lthe shipv to athatiaIwittii
angle very
si iwly."). One example is to defirne the work
Iig. 1. eIcliion oil c;l)sizc. (d anl extelrnal force, ic. the wind or the

o~wl:, :51 1i , w\II Nt'('.lll ,\29' JI nll y:.lI 99 3

current, to act rather static over one roll
period (Fig. 2).
To handle the effect of the dynamic wave
force a more complex interpretation is
needed.
o o t* o, ,,7-.,,, '*
Fig. 2. Stability criteria for beam wind and
rolling.
The stability standards based upon this
statical approach and terminology are the
foundation for most stability work all over
the world. The IMO regulations represent the
standard rules.
PROBABILISTIC APPROACH
REGULATIONS
We have stated that the external forces are
of a very dynamic, statistical and often very
non-linear nature. The event of a capsizing
itself may be referred to as a statistical
event. Since we have today the improved,
effective theoretical models to calculate
dynamic motions, we might think that we in
the future can calculate the final probability
of a capsizing event to occur. By using this
type of advanced technology and
crres[pond ing termninology, aidiscussion has
started if future stability standaids should he
based on such risk analysis. So farIth
analysis have been used ill piC-analyses of'
actual accidents. However, they have not as
yet been used in any new design situ;aitiu.
•:".rk5tk l\pa Ile•r\uscr'•.Ini\:h\Z') J;.niin~iry I993t•/
This terminology is fascinating to engineers
and governmental bodies since this type of
calculation, at least theoretically, can give an
absolute answer to the safety level of any
vessel or any new vessel design.
The probabilistic approach of establishing
stability standards have today some
limitations. To begin with, the statistical
nature of the environmental forces has not as
yet been satisfactorily mapped out. The
theoretical dynamic models used to calculate
the large motions of a capsizing are not of
the accuracy standard needed. All critical
events following a large motion situation
"that might influence the capsizing events,
such as structural damage, water ingress,
shifting of cargo and various human factors
have never been mapped in such a way that
we can use empirical figures of the
individual events of probability with
reasonable confidence.
However, this situation is improving and we
feel that future stability standards and
regulations will be reflecting this new
terminology. We should be aware that there
is an ongoing technology development in
related industries as ie. offshore, using
reliability calculations in the design analyses
TO of new structures and structure
elements/systems. This technology is quite
parallel to what we define as the problem of
stability.
FUTURE STABILITY REGULATIONS
BASED UPON OPERATIONAL
STABILITY
Through the probabilistic terms we have
been searching for the absolute stability that
will ensure safety against capsizing. The
opcration of a vess-v will, however, always
address the final and practical stability
situationl. here are often several aspects that
define a non-intact vessel when analysing a
capsizing. Nearly 90)% of1 all accidents
rcitetl to stability. happelns tlic to technical
errors, malpractice and other human factors,
Hatches and openings are not closed
according to regulations and as a
consequence the intact area onboard is not
intact anymore. The real static stability is
reduced. We might have a false safety level,
differing from the definition of the
regulations. Also, other malpractice
situations such as wrong manoeuvring in
waves can influence a critical situation.
on high speed
Through recent years, research
craft safety has been performed in Norway.
In this research much effort has put into
establishing critical performance require-
ments (not necessarily stability). The
requirements are based upon the analyses of
certain operational procedures. Emergency
operations in maloperation situations are also
investigated, Ref./5/. The result of the work
has been presented to IMO, Sub-Committee
of Ship Design and Equipment. It is a
proposal for requirements and compliance
criteria related to operational and safety
performance for all high speed crafts. For
high speed crafts this is a matter of course,
Full scale tests to prove the vessel properties
are proposed.
By using the survival testing technique,
some of these investigations might be
performed early in the design stage, and a
similar philosophy should be reflected in
future stability standards also for traditional
displacement vessels.
THE PHYSICS OF CAPSIZING
The theoretical tools for calculating large
dynamic motiu ns have been improved during
the last decade. However, it still seems to be
a long way tlgo before we can calculate
corirctlv :It,,t
l capsizing event in W;aves.
Even ulsi[nf-, superclllnpliteuls, new dytlainlic
theotrics have nHIt as yet l)l(ovided con1fidetlt
InSwels in such a way thut we c;arn Ilse these
calcilations as a final stability approval. The
new dylyainlic thlicrlies ha:ve so 1Ir
'':Vc•k5 \Ip:nz\I*.ci .t~n~,t,\2') I:uu:,1v I 993
us rational ideas around future stability
regulations. However, modern dynamic
theories have improved the physical
understanding of the capsizing event. We
have also gained empirical experience from
these studies, especially when combined with
investigations into earlier accidents. The
most critical ship/wave situations that might
impose a large rolling followed by a
capsizing event, have been mapped based
upon these analysis.
In Norway we performed several research
programs in the early 80's that specifically
dealt with such investigations. Through the
research program "Ships in Rough Seas" we
improved the statistics and the under-
standing of the physics of breaking waves
(kinematics).
MODEL TEST TECHNIQUE SURVIVAL
TESTING
New materials have put us in position to
build ship models, both cheaper and more
accurate than a few years ago. Advanced
measuring techniques, miniaturised radio
data control systems, and a very accurate
wave generation system in the laboratories
enables us to use free running, remote
controlled models that can be accurately
positioned in all kinds of wave situations.
Using advanced simulation computer
programs running on-line and parallel with
the tests, we can effectively perform a
parametric investigation of any critical
wave/vessel situation. The survival testing
can thus be an alternative to purely
theoretical investigations and a very effective
method to:
1. Improve the physical understanding
of dynamic capsizing events in
various wave situations
2. Qualify any new stability standards
or use the tests and calculations to
ot•luil finbd the scnsitivity of'these standards.
4
3. Define operational procedures that
can help the vessel to escape the
critical situation....
w-

The survival procedure will normally be a r

combination of test technique and theoretical
exercise as follows:
psal A: 'Typica1 )Iorvc.iJ f i.hin . I1
inw . l".5

Define critical environmental events.

Test the vessel structure in the

critical event and observe the results.

Establish the levels of the capsizing

event with respect to the "size of the
waves, wind and current".
in • 1.11 -

Establish the statistical base for the ,; .',O.

occurrence of the environmental
situation and the capsizing events.

Perform a parametric analysis by

theoretical tools of additional model
tests to find the "sensitiveness" of the
design or the criteria system.

Usethe results according to then

given purpose:
- to approve the design 1 r
to adjust stability criteria levels
- to define operational procedures. o •,,

In the following, some results from examples

or survival testing will be presented. Fig. 3. Examples of Norwegian vessels.

Waves on the beam

EXAMPLES OF SMALLER VESSELS
Tliaditional dicplacement vessels:
Sorme of the examples referred to are shown
in Fig. 3.
Ior tile fishing vessels and tile coaster 1i1t:
tests were originally per•iriied with the
ilntcnti)on of
investigating situations
co0ieCeI)ling tile stability critCria. The
following wave-vessel sittuations were
dcfilcd :1s critical:
motions in large irregular waves
motions in breaking waves
motions in tunes combinations of
wave train periods against rolling
periods, also wave trains including a
breaking wave.
In following waves
loss of stability on a wave top
parametric resonance in wave trains
Ioroaching situations

:\NItk5 I\1,;tIlcrr\is "- ,r.i,h\2'.) January I )9•

I
 ,-critical combinations of quartering
seas, with tests performed in long-
crested and shortcrested waves,
Certain combinations with swell were
also tested.
In Fig. 5 we have shown how the individual
examples are sensitive to rolling/capsizing in
breaking waves against the given intact GM
value.

Parameters such as: Wave kinematics, wave SURVIVAL TESTS INBREAKING WAVES 5-10 m
heights, static stability of the vessel and ROLLING ANGLE (deg)

detailed superstructure design have been 200

varied through the tests. The stability of the 150 ............
vessels varied around the minimum stability
given by the existing stability criteria. The 10 "
tests were performed in the Marintek Ocean
Basin Laboratories. Motions, acceleration
and water ingress were measured. Most of 0'.
the tests were referred to the intact vessel. 0 .2 .3 A4 .5 .6 .7 .8 .9 10 1.2
However, some cases of water filling in INITIAL STAB GM (m)n
shelter deck areas were also investigated. MODEL A MODEL B CARGO
£ + *

The results are presented in Ref. /1/. The
general conclusions are illustrated in Fig. 4
showing that any vessel can capsize in
waves (turn around in waves) if the dynamic
stability, ie the stability resistance work, is
low.
Fig. 5. Survival testing in breaking waves.
High speed crafts
In recent years the Norwegian focus has
been put on capsizing tests of high speed
crafts, Ref. /2/ and /3/.

A survival testing technique hase been used

CAPSIZING IN 5-10 m BREAKING WAVES for investigating high speed catamaran
Resuls from tests with 6 tilerent vessels designs. One has learned (quote Ref. /3/):
size 6.- 50. m. legjth
Max rolling angle (deg) "Operation of catamarans in ocean areas
200 critical should be carefully studied with respect to
180 line stability to avoid capsizing.
160
140 -
U
120 - A catamaran has righting moment which
100- is enormeous compared to a monohull,
80 - however, the wave loads/heeling moment
60 - will also be much higher for a catamaran
401. than for a nmonohull.
50 150 250 350 450 550 650 800
100 200 300 400 500 600 700 As capsizing safety of inonohulls, the
GM* displasement ( m'onnes) shape of the GZ-curve beyond 40 degrees
heeling angle is ol great imlportance, as is
also the angle of' vanishing stability.

The probability o" deck diving of existing

designs of high speed cralt with slender
Fig. 4. Rolling ;intzles in bicuking w;Imcs. Irlelomies is high.

*:\':, l"5\l•:lwr ,rpr,.\rh\29 Jiiiiy ]99t1

6
Deck diving might occur in very small
wave heights. If wave length and ship
speed is unfavourable, deck diving
becomes a problem even in 2-3 meter
wave height.
(L x B x H . 3.00 x 5.00 x 2.30.)

Deck diving occurs frequently for wave

length/ship length ratios between 1.0 and
2.0, i.e. similar situations that lead to
broaching. B Iwai,

Deck diving is critical for ship speed equal

to or greater than wave speed.

Longitudinal stability, expressed as deck Q

diving, is a very serious problem and
equally important with respect to safety as
transverse stability for high speed craft. Shetrdeck
Longitudinal and transverse stability/
instability and combinations extensively ,k
investigated to develop international rules
for documentation of such characte-
ristics".

New designs
A recent investigation has been performed at
Marintek (1991 - 1992) concerning a new
type of fishing vessel, see Fig. 6. A
catamaran design has been exposed through
a survival testing program. Critical
wave/vessel situation have been investigated.
Special emphasis were put on studies of the
vessel in breaking waves. The results are
discussed in Ref. /7l, here the general
conclusions:
"tile results are very encouraging and
display that this particular catamaran has
a significantly larger critical wave height
compared to a conventional monohull
vessel with the same dynamic stability.
Compared to nin nolull vessels ol the sarne
Ienth the results indicate that this
cataxmarian has a subsLaiasially higher
safety against capsizing inbreaking waves
fro11 tile side."
Fig. 6. Profile and body plan of the
catamaran.
General comments
Our tests showed that vessels havingat least
the minimum required stability very seldom
capsize in normal wave situations. In
breaking wave situations, most smaller
vessels can capsize, even those with very
high stability (small, self righting vessels can
be turned around 3600). When large angles
of roll occur, only those with positive
righting level arm at large angles of heel,
will righten. The Norwegian Maritime
Directorate has put an adjustment of the
standard I MO regulation into force through
the national regulalions for fishing vessels,
saying that the righting level arm shall be
positive at large angles of heel (40 - 60
degrces).

q•:\s;kSi\l);H).:r\uz')*, l ,hl\29l
J;iaI Iagl IV'199 7
For some types of vessels - offshore supply
vessels and other types of low freeboard
vessels - water on deck might be a critical
parameter. Water trapped on deck can
change the initial stability and may start an
inclination that leads to capsizing. Survival
testing will show how sensitive the vessel
structure is to the trapping of water. A
closed bulwark, open rooms in the back or
poop area should be avoided.
we
By performing effective survival testing
can map some of these design problems and
still rely on the statical stability standards,
I. Survival testing gives the answers to
theoretically complex situations. The
results can be interpreted the empirical
way by mapping the effect of possible
critical parameters.
II. We can use this technique to search
for very simple standards.
III. We can qualify such standards or
find the sensitivity of the standard by
using this technique.
STABILITY AND MALOPERATION
In most investigations that follow after a
capsizing accident we refer to the cause of
the accident as a combination of low
stability, technical errors and human
maloperation. large rolling angles in critical
wa.]ves often end into shifting of cargo or/and
water ingress through open doors and
hatches. The non-intact vessel suffers a total
C;I•sizinrg.
Capsizing is most likely to occur for non-
intact vessel conditionis
TO impove satleCly urlt sMR'
)lld look for thi-
sCnsivi%.itV in the design to turn critic;ally
rrll-int:ct (file to huln;nn ralopcr;rtion or
wAhlulic'l errors. Inilijli s1l;,hility philoh.s" )hics
I:\tp:nprnv;&g.I ,,\un\lh\ J;i;.y 1993
and standards should reflect this fact. One
should note possible human mal-operations
and find out if they are followed by low
non-intact critical stability situations.
Examples are:
- water through open hatches
- critical manoeuvres in waves
- maloperated ballast system
- wrong cargo handling
- wrong handling of fishing gear
Survival testing by use of advanced physical
models can give answers to questions such
as these, even if they are complex and
transient.
We might think that a vessel, even if it
fulfills the standard requirements, should not
be approved if the design implies a very
sensitive operation. We can require a design
adjustment or we can increase the standard
static stability requirements to make it less
sensitiv stabilitywise.
To increase* the total safety by improving
vessel stability in future regulations, one
should make requirements that imply
building a kind of minimum stability
barrier for non-intact conditions.
MODERN DESIGNS
During the winter season of 1992, several
accidents occurred with the loss of
Norwegian vessels. The cause of these
accidents have not as yet been fully
determined. However, the question raised is
if new critical wave/vessel phenomena do
occur or if the design of modern fishing
vessels has changed in a way that makes
them sensitive or difficult to operate.
TJ, e labo)ra te, we can comment ou tile
rllowirng examnplc: A modern fishing vessel
- the design characterized by an increased
licighlt/extra deck, was supposed to be a
oe d vessel design stabilitywise, Fig. 7.
 rolling periods and wave motions, one asked
if the dynamics of waves were not quite
understood.

An investigation involving survival testing

a 0 0would have confirmed or excluded this
-----
.. .......... .... theory. Most likely, we would have found
" .. tthateory.
v..........
the poor dst
i gnly in to h inae
n spite
o un-itc
new model designs, of
fulfilling all stability requirements turn into
. .-- a very poor design in the non-intact
S............_- condition. Open hatches leading water into
IL)! rooms, quickly changes this design into a
negative stability vessel. See Fig. 7.

-l .0 -CONCLUSIONS

From a stability problem point of view of a

complex hydrodynamic nature, one should
not hesitate to get into analysing situations
- . due to lack of accurate theoretical tools. The
effective performance of survival testing in
advanced modern laboratories will in our
opinion be the effective tool of tomorrow.
__ Future stability standards should reflect a
combination- of the various techniques
discussed:

Static stability and IMO standards as

the base.
Fig. 7. Example, new Norwegian fishing,
vessel design. Corrective design action.
Analyse the likeliness of the design
ending in a poor non-intact condition
The vessel met with problems in following by using systematic technical errors/
waves. It became non-intact due to a large maloperation investigations.
wave and water ingress through open (FMEA analysis or similar analysis)
hatches, became unstable - then capsized and
sank. The accident happened in a very waste Using a probabilistic approach to
area (Svalbard) and several lives were lost. define the total sensitivity against
capsizing (risk analysis).
Once again the question was raised about
new stability phenomena not investigated Final operational stability.
earlier. Togethur with a combination of high Operational procedures b;a1sed
duck level, long radii of gyration and loiqý upl)on above investigatrns.

.:\,4:kSl \Jm)a r\ti.cr.h\,hi\2' J.u~ti1ry j 993 9

REFERENCES /5/ Sub-Committee on Ship Design and
Equipment, code of Safety of
/1/ T. Nedrelid, Marintek: "Survival Dynamically Supported Craft.
testing, a tool and technique to Submitted by Norway. Appendix N,
establish the safety level of marine "Definition, requirements and
floating structures; smaller vessels to compliance criteria related to
large offshore structures". Stab. 90, operational and safety performance
Naples. for all high speed craft".

/2/ E. Jullumstro, Marintek: "Stability of /6/ P. Werenskiold: "High-speed craft

high speed vessel". Stab 90, Naples.
/3/ Stability of High Speed Craft. NTNF
program of High Speed Marine
Vehicles.
/7/
/4/ E. Huse, T. Nedrelid: "Stability of
semisubmersibles under extreme
weather conditions". OTC 1985,
Paper 4987.
operational performance and
limitation".
3rd Conference on High Speed
Marine Craft. Kr.sund N, 1992.
B. Enerhaug, Marintek: "Capsizing
Experiments of a Catamaran Fishing
Vessel in Breaking Waves". 2nd
International Symposium "Safety and
Working Conditions Aboard Fishing
Vessels. Spain, Sept. 1992.

~l\lp:
q•:,:,,'v5 ,: I ,,rhI\2'l Jat'I
II\tistpLl ."y I%'''•9 I)
UN! VERSITA DEGLI STUD! DI TRIESTE

DIPARTIMENTO DI INGEGRERIA NA VALE. DEL MARE E PER L'AMI3IENTE

VIA A. VAI.ERJO. IO- 34127ITRIESTE

k4-( AQloert- :sQraA ctcccvrrc-

t ea ry-rs
,e LA &s ea
+týQe+L~crc5~et~

't. {c Ice lneQ '%t-, TFr oLeluAý,\, A. L' 3

vc-L.

net~~~~~ (61 A/l-

...hi....l
WEGEMT Short Course on:
"New Techniques for assessing and Quantifying Vessel Stability and
Seakeeping Qualities"
Trondheim, 8-11 March 1993
"DYNAMIC STABILITY - PROBABILISTIC APPROACHES"

Alberto FRANCESCUTTO
Department of Naval Architecture, Ocean and Environmental Engineering
University of Trieste
Via A. Valerio 10, 34127 TRIESTE (Italy)

ABSTRACT
The hydrodynamic part of ship safety is actually managed by means of
prescriptive stability criteria based on static stability characteristics of the ship.
The analysis reveals that this approach leads to many contradictions whose
final effect is to some extent the delay of the formulation and adoption of higher
safety standards. Only the development of a fully physical approach could
allow a realistic evaluation of the probabilities of ship loss through the different
mechanisms. The final goal will be an effectively unified approach to ship
safety.
In the first part of the paper, the safety problem of fishing vessels is
considered in connection with the possibility of their loss in waves. To this end,
different dynamic mechanisms are studied in detail, ranging from capsizing to
loss of control and/or broaching-to phenomena.
The stability of a family of fishing vessels is investigated in the light of
the existing I.M.O. criteria and of the new method of energy-balance proposed
by Strathclyde University is illustrated. Then the set of linear differential
equations describing the dynamics of the antisymmetric motions of roll, yaw
and sway is solved to look for possible unstable solutions. It results that, in
general, I.M.O. statistical criterion is less restrictive than I:M.O. weather
criterion and both fall into a region of approximately 20-30 percent of net area
positive according to energy-balance method. An analysis .of the stability of the
antisymmetric motions reveals the possibility of dangerous phenomena
depending on the trim condition.
For the same family, an analysis of the threshold for the onset of
parametric rolling in a following sea reveals the extreme sensitivity of this type
of vessels to such a particular capsizing mechanism.
These results are then critically reviewed in the second part of this
paper to look for an answer to the currently unsatisfying state of art of small
ship safety.
The loss of a ship is often a direct or indirect consequence of large
amplitude rolling. In the third part of this paper nonlinear rolling is thus
examined from a probabilistic point of view. The nonlinear rolling in a
stochastic sea is analysed by means of an approximate perturbation method.
The results indicate that the possibility of bifurcations to resonant or
subharmonic large amplitude rolling is possible in a narrow band sea. A
numerical time domain simulation confirms these results and gives an
indication on the probability denstities of states, that in the sinusoidal limit
should recover the domains of attraction. The results can be of great interest in
cargo mechanics modelling, an item relevant to ship safety.

INTRODUCTION
Safety at sea is a complex goal to achieve. This is due to the extreme
variety of ships and structures and the different causes that can give rise to
casualties at sea.
Depending on the dominant aspects and on the level of knowledge,
these causes are divided in different categories. An important role is usually
attributed to the structural aspects and to ship stability.
Stability is the part of Naval Architecture traditionally considered to give
the tools for a correct approach to the safety aspects connected with the
possibility of ship loss through the mechanisms of capsizing, sinking and surf
ridinging/broaching-to.
As will be shown in the following, this approach and the safety rules
based on it, are to some extent inconsistent. The introduction of a correct
approach and consequently of the more appropriate name of "hydrodynamic
part of ship safety", will improve the actually unsatisfying level of safety for life,
cargo and environment at sea. In addition, this goal could probably be
obtained without going further in the way of paying increased safety levels with
unacceptable reductions of payload and/or operational capability. This could
originate a positive feedback making more attractive the improvements of
safety in ship design introducing the concept of "hydrodynamic safety
performance" and a subsequent effective "design for safety".
This program rests on the development of a fully physical approach to
ship safety through which the probabilities of the different dangerous
phenomena can be computed. The concepts and methods of reliability could
then be applied to obtain risk evaluation and control.
In this paper, the contradictions of current approach are highligted
together with their negative effect on developments and progress in the field of
ship safety.
Many aspects of ship hydrodynamics together with the consideration of
ship structural aspects and their mutual interaction are involved in the
previsions about ship safety. In particular, ship stability, manoeuvrability and
their interactions, cargo mechanics, water on deck, loss of hull integrity, etc.,
can play an important role. Since it is difficult to deal with all these items in a
short paper, so that, in the following, main consideration will be given to:
- IMO criteria with a typical application on a family of fishing vessels;
- the method proposed at Strathclyde University regarding the energy
balance in waves to take into account the loss of stability in waves;
- the method proposed at Brunel University regarding the stability of
antisymmetric motions as one of the ways to take into account the possible
interactions between transversal and directional stabilities;
 -the nonlinear rolling in deterministic and stochastic sea to take into
account the possible effects on cargo mechanics with shifting of cargo, loss of
structural integrity, etc.
To have an idea of the importance of this discussion in the case of
fishing vessels, a look to the statistics of casualties reported in Table 1 is
sufficient. In Table 1 the casualties leading to capsizing and broaching (about
60% of casualties) regarding Japanese fishing vessels in the period 1973-77
[1 ] are analysed with respect to the causes.

1. IMO CRITERIA AND OTHER RECENTLY PROPOSED METHODS

2.1. Introduction
The problem of the safety of navigation at sea is very difficult to handle
in particular for ships whose characteristic dimensions render them more
sensitive to the action of the marine enviroment and, among these, a relevant
place is reserved to fishing vessels [2,3]. The hull shape of a fishing boat
varies greatly due to different local conditions, fishing methods, -construction
material, engine weights, distance to the fishing grounds and other factors.
Thus, it is obviously very difficult to design a few standard hulls which are
suitable for all conditions.
In the last decades, different organizations have collected and
published results of theoretical calculations, model tests and full scale trial in
an attempt to indicate the trend in the factors which influence stability,
resistance, powering and seakeeping qualities of the vessels. The final goal is
to reach conlusive indications on how to find the optimum hull shape when
designing a new fishing boat. The statistical analysis of the data available is
also intended for estimating the total performance of an existing design so it
can be investigated if there is still room for any improvement.
At international level, IMO has devoted particular attention to small ships
and in general to ships less than 100 m in length and, among these, a relevant
place has been reserved to fishing vessels. Unfortunately, the stability rules
adopted by the main Classification Societies still belong to a quite old
approach, being mainly based on Rahola's statistical results requiring
prescriptive characteristics of the curve of righting arm curve in calm water.
The new proposals [4,5] suggest the extension of the weather criterion to such
ships and are presently included in the Torremolinos Convention, which is yet
not approved by a sufficient number of State Governments to become an
international rule. It would be interesting to investigate the reason of this
unsatisfactory state of art, even if economical and political choiches seem to
constitute the major problem for further progress.
In the first part of this section, the stability of a family of the BSRA trawler
series is investigated in the light of the existing criteria, i.e. the l.M.O. statistical
criterion and the l.M.O. weather criterion. In the calculations, a group of 15
vessels was considered in order to determine the effect on stability of varying
different design parameters.
In the second part of this section. the effect on stability of a longitudinal
wave is taken into account following the Strathclyde approach and
considering the effect of the parametric resonance in following sea. Finally, the
possibility of other ship loss mechanisms, connected with directional stability
such as broaching-to is examined through an analysis of the stability of
antisymmetric motions.

1.2. The family of vessels

The numerical investigations were carried out on a family of fishing
vessels known as the BSRA trawler series [6,7]. The parent form for the series,
model XF, was chosen to represent a ship 150-ft. LBP x 26-ft. 4 in. moulded
breadth x 13-ft. 2 in. moulded draught with a displacement of 847 tons salt
water.
The hull forms of the series were derived from the parent form according
to the following transformations:
- affine distortion of B/T (B/L, T/L) starting from ship XF for hull forms XG, WO,
WP, 907, while keeping L/V"' 3 constant;
- affine distortion of UV 113 (B/L, T/L) starting from ship XG for hull forms WS,
WR, WO, while keeping B/T and V constant;
- variation of CB (Cp) by conformal transformation of the sectional area curve,
starting from ship XF for hull forms ZP, ZQ, but having main dimensions
constant;
- same as above, starting from ship XF for hull forms 851, 852, while
maintaining B/T and L/V 1 3 constant;
- variation of XCB from ship XG for hull forms 975, 977, 978, by modifying the
sectional area curve only.
All the hulls of the family have been normalized to the same
displacement through geometrical similarity and are fitted with similar
superstructures independently on their length. Their main geometrical
characteristics are given in Table 2. The forecastle has an extension of
approximately 25% of the ship's length and the freeboard is kufficient to avoid
deck immersion until a heeling angle of 12,50.

1.3. I.M.O. Statistical criterion

The I.M.O. "statistical" criterion for fishing vessels of 24 meters in length
and over has been endorsed at the "International Convention for Safety of
Fishing Vessels" held at Torremolinos in 1977 [8,9]. The criterion, that derives
from the pionieristic work of Rahola, is written in terms of stability standards
based both on statistical and other analysis of casualty records and on the
experience of different fishing fleets throughout the world. The Standards are
expressed in terms of prescriptive values for certain key features of the righting
arm curve and can be summarized as follows:
-Standard A: The initial metacentric height GM should be not less than 0.35
metres.
- Standard B: The area under the righting lever curve should be not less than
0.055 metrexradians up to 300.
- Standard C: The area under the righting lever curve should be not less than
0.090 metrexradians up to 400 or up to an angle where the non-watertight
openings come under water (whichever is less).
- Standard D: The area under the righting lever curve should be not less than
0.030 metrexradians between the angles of heel from 300 to 400 or such lesser
angle mentioned under Standard C.
- Standard E: The maximum righting lever should occur at an angle of heel
preferably exceeding 300 but not less than 250.
- Standard F: The righting lever should be at least 0.20 metres at an angle of
heel equal to or greather than 300.
To aquire more information about the relative importance of the
requirements introduced by the I.M.O. statistical criterion, we show in Table 3
[11] the maximum allowable height of the centre of gravity above the keel
according to the different Standards. The results indicate the very different
weight of the different Standards. In particular, Standard A, prescribing a
minimum value for the metacentric heigth GM is by far the less resctrictive.
Standard E, relative to the location of the maximum value of the curve GZ is
generally the most restrictive. These results are in agreement with similar ones
found relatively to the fishing vessels of the Ridgely-Nevitt [10].
In spite of its great simplicity, the statistical approach can be, and indeed
is, criticized on the following basis:
- the statistics refers to a great variety of vessels, loading conditions and
operation areas, so that the concept of "homogeneous specimen", i. e. the
basis itself for the validity of a statistical approach can be questioned;
otherwise, the specimen would belong to so specific situations to be hardly
applicable in general;
- it constitutes an "a posteriori" (hindsight) approach, i. e. in the best
hypothesis, it can only follow the trends indicated in the records of casualties
at sea;
- no physical description of the effective interaction environment-ship is
considered.
We will return on this point in Section 2.2.
1.4. I.M.O. Weather criterion
In the recommended I.M.O. "weather criterion" [8,9] for fishing vessels of
45 meters in length and over in unrestricted service, the ability of a ship to
withstand the combined effects of beam wind and rolling should be
demonstrated for each standard condition of loading, with reference to Fig.1,
by means of the following procedure:
- The ship is subjected to a steady wind pressure acting perpendicular to the
ship's centreline which results in a steady wind heeling lever (lwl).
- From the resultant angle of equilibrium (qo), the ship is assumed to roll owing
to wave action to an angle of roll (ql) to windward. Attention should be paid to
the effect of steady wind so that excessive resultant angles of heel are
avoided. The angle of heel under action of steady wind (qo) should be limited
to a certain angle to the satisfaction of the Administration. As a guide, 160 or
80% of the deck edge immersion, whichever is less is suggested.
- The ship is then subjected to a gust wind pressure which results in a gust
wind heeling lever (Iw2).
 -Under these circumstances, area "B" should be equal to or greather than,
area "A".
- Free surface effects should be accounted for in the standard conditions of
loading.
Also in this case the stability of the vessels is investigated in the light of
the proposed criterion. In order to simulate different loading conditions, in the
calculations the height of the centre of gravity above the keel was varied until
the stability requirement failed. As a result we obtain the maximum allowable
height of the centre of gravity for each vessel according to the weather
criterion.
In Table 4 [11 ], we show the maximum height of the centre of gravity
above keel allowed by the statistical and the weather criterion respectively. It
appears that the statistical criterion is less stringent than the weather criterion.
The weather criterium approach is preferred to the statistical one as it
introduces a description of the interaction environment/ship, so that it appears
more "physical" and not "a posteriori". This is only partly true, since the
description is too much simplified to be sufficiently realistic, and limited to a
particular mode of ship loss, i. e. the capsize mode. Infact, the results shown in
Table 4 indicate that the two methods give comparable results, with a slightly
lower maximum allowable center of gravity heights corresponding to the
second. In the opinion of the authors this should upset definitely the statistical
crite riurn.

1.5. The energy balance in waves

The method recently proposed by Strathclyde University [12] is based
on the consideration of a time-dependent roll restoring moment. To this end, in
addition to the combined effects of beam wind and rolling, also the effect of a
following or quartering sea on the stability of the ship is taken into account. In
the proposed procedure the righting arm is computed for different positions of
a wave with respect to the vessel. The wave is assumed to have the same
length of the ship and the encounter period equal to her natural rolling period.
Then, an ultimate half roll is supposed to occur between a windward angle
and an extreme leeward angle. For each position of the wave relative to the
vessel the minimum righting arm curve during the ultimate half roll is obtained
and the corresponding energy-balance, taking into account damping and
wind, is computed to give the net area, as shown in Fig. 2. The percentage of
time with net area positive during the passage of the wave along the ship is
assumed as an indicator of the safety from capsizing. Futher details can be
found in Ref.12.
In the application presented in this paper, the parameters describing the
ultimate roll and the wind lever are esimated using the l.M.O. weather
prescriptions. In the first application no roll damping was assumed to act. A
rough estimate made on one of the vessels, indicated that the consideration of
an equivalent linear damping realistic for these ships in the absence of
antirolling devices, could increase the percentage of net area positive by
about 20%.
The results relative to a group of hulls of the BSRA family to analyse the
effect of the variation of the parameter B/T (the only that was found to be
relevant) are shown in Fig. 3 [111, where the KGmax/D is plotted against BKT. As
one can see, both I.M.O. criteria lie in a region correspong to about 10-25%
net area positive, whereas the request of 100% net area positive if much more
restictive in terms of the maximum allowable height of the centre of gravity.
In Fig. 4, the KG/D is plotted against the Kempf's [13] nondimensional
rolling period TK=TR(g/B)1' 2 for the considered different hulls. Drawing the
lines connecting the KGmax/D characteristic of the different hulls with respect to
the different stability standards considered, one can see that the request of
greater stability, for exemple a greater percentage of net area positive,
corresponds to pushing the Kempf's period towards lower values.
Actually, it is not recommended to have TK<8 or TK>1 4. In the first case,
the roll behaviour of the ship is too stiff, so that the oscillations have a
moderate amplitude but higher frequency, i.e. great accelerations on the deck,
rendering the operations difficult. On the contrary, a great value of TK gives rise
to a cranck behaviour with large amplitudes and still a loss of operational
capability (and of stability!). We will return on this point in Section 2.2.

1.6. The effect of parametric resonance in following sea

The effect of heave-roll coupling on the stability in waves was

investigated by considering the so called parametric roll excitation. As known
[14,15], theoretical predictions, confirmed by experiments, suggest that large
amplitude rolling can be excited in the case of quartering and following seas
when the encounter wave period is approximately half the rolling period and
moreover a threshold condition correlating relative righting arm variation to roll
damping is fulfilled. In design terms, the threshold condition expresses a
correlation between sea intensity and both hull form geometry and weight
distribution. By lowering the centre of gravity or introducing additional roll
damping, one has the possibility to avoid the phenomenon.
Preliminary computations were done using a simplified, but common approach
based on the assumption of linear time dependent rolling equation, so that the
threshold is given in terms of the comparison of the relative GM variation in a
longitudinal wave and the equivalent linear nondimensional damping. The
isocarenic hypothesis was assumed, corresponding to consider the quasi-
static approach. In the present case, the analysis revealed an extreme
sensitivity of this class of vessels to parametric resonance. In fact, no realistic
values for KG/D were found to prevent the phenomenon also in the case of
vessels having bilge keels of area up to several percent of waterline area.
Thus, the only way to avoid such dangerous situation is linked to the ability of
the master, who must appropriately handle the ship at sea. A more detailed
analysis taking into account the effective heave-roll coupling and the limiting
effect on the roll amplitude played by the righting arm and damping
nonlinearities, could change appreciably these conclusions. Moreover, being
a "resonance" way to large amplitude rolling and eventually to capsizing, it is
necessary, for its build-up, that a sufficient degree of autocorrelation be
present in the wave train, so that it is tied to the appearance of a quite narrow
band sea. The roll of a ship in irregular following sea is analysed by
Dunwoody [16] that describes the effect of GM fluctuations as an equivalent
roll damping reduction. The author discusses also the different stability limits
with respect to the onset of parametric resonance connected with the different
levels of stability of a stochastic dynamical system.
1.7. The stability of antisymmetric motions

Following a method developped at Brunel University [17,18], we

consider a mathematical model suitable for the analysis of the stability of the
antisymmetric motions of the ship that is, in order, sway, roll and yaw. This
method could describe the possibility of ship losses due to broaching-to as it
introduces a coupling between roll motion and directional stability. Since, we
are interested in the onset of the deviations, the system of equations governing
the antisymmetric motions may be written as:

m(v÷Ur) = mgo+Yvv+Yvv+Ypp+YOO+Yrr+Yrr

lxP-lxzr = Kvv+Kvv+Kpp+Kpp-pgVGM@+Krr+Krr

lzr-lxzp = Nvv+Nvv+Npp+Npp+Nrr+Nrr

The system is linear, the effect of the rudder is not considered; attention is
focused mainly on the accelerated motions and the effect of trim variations is
considered through the derivatives expressing yaw and sway dependence on
yaw and sway velocity. Dimensionless parameters are used through a suitable
transformation. In particular, the parameter GM/LFn 2 is introduced to study the
stabilizing effect of increasing the metacentric height and the destabilizing
effect of increasing speed. No frequency dependence is considered, i.e. the
slow motion derivatives approach is used.
The solution of a linear system of homogeneous equations undergoes a
fast decay or gives rise to a non decaying oscillatory or diverging behaviour.
Assumed a solution of the form

v(t)=vo e )t r(t)=ro e xt 0(t)=-o e M p(t)=X(t) P•=,+iw

for a non trivial solution to exist, the determinant of the system of algebraic
equations obtained substituting the above expressions in the system of
differential equations has to be zero. From this, the characteristic fourth degree
equation is derived.
The solutions are the eigenvalues of the system, and generally two of
them, p• and 1-12 are real and two g3 +iL)3 are complex conjugate, corresponding
to the fact that of the three motions only roll can have an oscillatory behaviour.
Dynamic instability prevails when at least one of the real eigenvalues or the
real part of the complex is positive, the last case corresponding to oscillatory
dynamic instability.The second step is to analyse the eigenvectors
trepresentng the amplitude of the motion solutions. Their moduli will be
normalized to the yaw motion (put equal to unity) to have the possibility of
individuating immediately which is the predominant. To each eigenvalue X, we
have thus a set of three eigenvectors that represent respectively sway, roll and
yaw ampliwudes. If the eigenvalue is positive in that rangeof GM/LFn2, the
predominant eigenvector indicates the motion that can give rise to dynamic
instability.
 In computing the stability of antisymmetric motions of the BSRA series of
fishing vessels [19], the hydrodynamic derivatives for yaw and sway terms
depending on velocity were estimated from manoeuvrability [20,21]. Only very
rough estimates were possible for the terms depending on accelerations and
those regarding roll motion. This makes absolutely necessary to consider the
possibility of performing a thorough experimentation in this field.
In Fig. 5, the eigenvalues of the model XF for different values of the trim
cofficient y=T/Tm are reported. As one can see, the eigenvalues are very
sensitive to trim variations. In particular, both Pla and the real part .3of X3 are
positive in a wide range for y<O, that is in the trim by bow conditions. No
instability is detected in the normal trim by stern conditions A look at the
eigenvectors reported in Fig.6 relatively to y=-0.2, indicates that the system
instability is yaw dominated in both senses (diverging and oscillatory) since
the values of the at's and of the 8's (representing respectively normalized sway
and roll eigenvectors) are less than unity (representing normalized yaw) for
the first two eigenvalues, whereas the third indicates the possibility of a strong
roll-dominated instability, particularly at low values of GM/LFn 2 (low static
stability or high speed).
In Fig.7 the same results of Fig.5 are reported for the model 907. Here
the range of dynamic instability is wider then for XF. This could appear strange
if one considers that the model 907 is more stable than XF both following IMO
criteria and Strathclyde method. The question is that stability and safety is a
very complex problem and different approaches measure the resistance to
different dangerous phenomena. Here, not the pure capsizing possibility is
considered (of course only a fully nonlinear model could account for that
phenomenon), but also loss of control and broaching to, although in the
opinion of some authors the phenomena of directional instability and
broaching-to are not so clearly connected [22]. Moreover, sometimes the
instability appears connected with an inherently poor course-keeping ability
[23]. The results are not given in terms of KGmaxID due to the uncertainty in
the evaluation of the hydrodynamc coefficients.

2. CRITICAL ANALYSIS OF CURRENT APPROACH TO THE

HYDRODYNAMIC PART OF SHIP SAFETY

2.1. Introduction
Existing rules for the assessment of safety of ships are contained in the
"stability criteria" that in the following will be considered as known. A general
critical analysis of these criteria, of the misleading use of the term "stability"
and of the sometimes ambiguous role played by different actors interested in
ship safety, is contained in Ref. [24]. Here the analysis is focussed on
particular aspects of this approach that have negative influence on the
development of new concepts and methods for the improvement of ship safety
levels. From here on, except where differently stated, we will refer simply to
ship safety instead of hydrodynamic part of ship safety.
It is customary to separate the case of intact ship from that of damaged
ship. Both are based on a static approach. The most advanced formulation
regarding the damaged ship stability begin to include the description of the
transient phase of flooding, where transient is still intended in a quasi-static
approach.

2.2. The case of intact shin

The first serious approach, in the following indicated as IMO Statistical,
originated from the work of Rahola published in 1939. At that time, the
knowledge of the mechanims leading to capsize and broaching-to was poor
and the availability of efficient tools for direct hydrodynamic calculations
extremely limited. Hence a statistical approach that can be regarded as
scientifically unsatisfactory but practically relevant.
The result is in the form of correlations expressing limiting values
(prescriptions) for some characteristics of the righting arm of the ship with
respect to fixed-trim, calm-sea, isocarenic, transversal inclinations.
Due to its relative simplicity and to its apparent relation to the physical
mechanism of capsizing, the correlation was taken as a basis for a stability
criterion that, with small adjustments (requiring years of discussions!) is still in
use.
In the meantime, the need to reduce the number of casualties stimulated
the research and the analysis of casualties. The result was the proposal of the
adoption of a more "physical" criterion, i.e. the IMO Weather criterion. The
more or less evident opposition of many bodies delayed the application of this
rule and originated a complicate classification of ships in relation to the type,
size and range of operation. In many cases, especially for the smaller ships
(the most sensitive to the meteomarine action!) only very elementary
calculations or no calculations at all is requested. The reasons of the delays
and of the differentiation of the requests depending on ship size appa-rently
based on the difficulty and length of involved computa-tions, really masks the
attempt to avoid excessive reduction of paylod or operational capability.
This point worths a particular attention since it is usually assumed an
obvious the fact that increasing safety means decreasing payload. As we will
show, this connection is, at least partially an intrinsic consequence of wrong
initial assumptions about safety.
The statistical criterion, in fact, is an "a posteriori" one. In addition, it
expresses a correlation between observed ship losses, in a particular class of
ships sailing a very restricted region, and computed static stability characte-
ristics. The main criticisms can be sinthetically expressed as:

a) there is no direct connection between the requests and the needs. The
criterion does not involve seakeeping qualities, the forcing effect of
meteomarine environment, trim variations,coupling among different motions,
course keeping ability, shifting of cargo, etc;

b) the statistical nature of the correlation averages between good and

bad projects (with respect to safety), as a result, the good projects are
penalized;

c) as long as the ship remains intact, the probability of casualty at sea is

usually quite low as regards the mentioned mechanisms;
 d) the request to be "simple".
Since ships are objects with a quite high index of reliability, this means
that a few bad projects can be heavily conditioning. This means that even a
small subset of casualties only explainable as a wrong correlation attempt as
per a), can render the system almost "incompressible" in the sense that a very
strong additional request (i.e. a very low KG reducing drastically the payload)
contributes to a negligibly small increase in the observed safety at sea.
This could be advocated as a partial justification of the negative
consideration paid by the owners to improving stability criteria.
To exemplify, the results of a parametric analysis conducted on a family
of fishing vessels belonging to the BSRA series is discussed in some detail.
Firstly, the maximum allowable KGID to satisfy both IMO criteria has been
computed [11 ]. The ships of the chosen subset are all obtained from the first
one through affine transformation. They have the same value of L./V 113 while
B/T increases from 1.97 to the extreme value 3.44 for the model 907. Reported
as a function of BIT, the results exhibit a regular trend for both IMO criteria. The
value of KGmax/D increases with B/T. In Fig. 4 the same values of the heigth of
the center of gravity are reported -as a function of the Kempf's nondimensional
rolling period TK=TR4-g7B. This is an interesting parameter as regards ship
operational capability, since it is not recommended to have TK<c8 because the
roil behaviour is too stiff, so that the oscillations have moderate amplitude but
high frequency corresponding to excessive deck accelerations. At the opposite
extreme, a value TK>14 corresponds to large amplitude rolling (and loss of
stability in the traditional sense). Good operational capability corresponds to
the condition 8cTKc14. A look at Fig. 4 indicates clearly that the two IMO
criteria give comparable results with an average value of TK=8. A more
stringent request of stability pushes the ship toward a too stiff behaviour. For
example, the use of the Strathclyde method corresponding to the energy
balance taking into account the effect on transversal stability of a train of
waves [12], leads to TK=6 if 100% net area positive is requested.
As regards point c), it is important to note that often the loss of a ship is
the result of a concatenation of causes, among which an important role is
certainly played by the following scenarios:
- loop: large amplitude motions/accelerations =* shifting of cargo =>
stability degradation =*;

- large amplitude motions/accelerations =*shifting of cargo loss of

structural integrity of the hull = damage:

-large amplitude motions/heavy weather =* water on deck stability

degradation.
A ship with movable cargo (solid or liquid) or water on deck cannot be
considered intact in the classical sense unless the proper dynamics of cargo
and of the mutual ship/cargo interaction is taken into account. In the particular
case of liquids with free surface, either in a container, as water on deck or
 water in a compartment as a result of flooding after damage, the consideration
of the classical correction to the different components of GM is contradictory
because it introduces a mixed approach statical/pseudo-dynamical. The
system, even if attention is focussed on roiling motion, is in this case a two
degrees of freedom one. Its behaviour is also qualitatively different from that
foreseable with the classical approach. As an example, the results of an
expenimental study of the behaviour of a fishing vessel with water on deck are
reported in Ref. [25]. According to the classical theory the amount and location
of water would contribute to a reduction of the metacentric heigth of about
50%. This does not account for the sharp change of natural rolling period that
doubles. Moreover, the effect of water on deck is in this case stabilizing. But
one does not neglect the fact that the effect of moving water is to some extent
similar to that of passive antirolling tanks that change the location of the
resonance peak giving generally origin to two peaks instead of one. It is not
easy to conclude that the observed is a general trend without a careful
examination of the possible variations of angle and frequency of encounter.
The effect of sloshing of water in a compartment of a fishing vessel is actually
under study [26].
.As regards point d), it has to be observed that the search to
be simple
implies a generally poor physical description of the mechanisms of ship loss.
This is evident for the statistical approach, but a closer examination reveals
that is a problem shared to a large extent by all approaches. As a result, an
unknown percentage of the correlation between casualties and prescriptions
remains unexplained, with the negative feedbacks already mentioned.

2.3. About "optimization"

The simultaneous analysis of the seakeeping and stability performances
[4] on the same specimen of ships indicates a quite good degree of correlation
between the two characteristics, as shown in Fig. 8. Another subset of ships
belonging to the same family, composed of models XF, WS, WR, WO, ZP, ZO,
851, 852, 975, 977 and 978, once more obtained through affine transformation
or conformal transformation of the sectional area curve has then been
considered. In spite of being ships of the same type and comparable size and
forms, the results are now completely different. A look at Fig. 9 indicates that
the degree of correlation is very poor, if any.
It is not difficult to explain these drawbacks. The correlation is good when,
for similar forms, the relevant parameter to be varied is BIT. This, in fact, is
positively correlated with both the investigated characteristics: linear
seakeeping (vertical motions) and static stability. Increasing B/T on the other
hand, decreases the course keeping ability that has been called as a cause of
broaching-to. In these conditions it is a very abstract thing to try to find an
optimum project out of graphs like those of Figs. 8-9. The difficulties connected
with rank optimization when trying to mix vertical and transversal ship motions
and other characteristics, and thus the drawbacks to include stability in
optimization, is also evident from analysis like that reported Fig. 9a, taken
from Ref. [47).

3. STOCHASTIC ANALYSIS OF NONLINEAR ROLLING IN A NARROW BAND

SEA
3.1. Introduction
All approaches to ship capsizing suffer in different measure of intrinsic
limitations. Experimental studies can be quite realistic in the incorporation of
the physical modelling and of the high nonlinearities, but are generally limited
as regards the possibility of variation of significant parameters and
consequently in the parametric inference. On the other hand, theoretical
approaches allow for good analyses of parametric dependences but generally
fall in a sort of "principle of indetermination", being less accurate in the
forecasting of large amplitude motions as they apparently improve
incorporating more details of the physical modelling. As a consequence, if one
wants to describe large amplitude motions (stability), quite rough mathematical
models have to be used, generally one or "one and a hair degrees of freedom
(i.e. one fully nonlinear and some other quasi-linear). On the contrary, low
amplitude motions (seakeeping) can be approached by good analytico-
numerical schemes incorporating, linear or quasi-linear, hydrodynamic
characteristics of increasing complexity. These difficulties are slowly
disappearing due to the advent of powerful techniques for the numerical
handling of nonlinear and 3-dimensional fluid dynamics.
The practical need of relatively simple guidance rules for the design and
operation of ships and ocean vehicles, generally known under the quite
improper name of "stability criteria", formerly approached in static or quasi-
static way, is now studied by means of nonlinear time domain simulations.
These constitute a sort of "numerical experiment", whereas true experiments
are now mainly used for validation porposes, except the cases where the
physical modelling is so complicate that cannot be actually tackled efficiently
by other means, as for exemple the study of the capsizing of vessels in very
rough weather, including breaking waves, performed in the Norwegian Project
"Stability and Safety for Vessels in Rough Weather".
It is never easy to design and interpret correctly an experiment, also in
the case it is a numerical one. In particular, the recently discovered possibility
of bifurcation scenarios, renders very complicate and possibly meaningless
the results of time domain simulations. This paper is devoted to a statistical
analysis of two of the bifurcation schemes of the nonlinear ship rolling in the
stochastic domain.
Having mainly in mind the Stability problems, it is clear that a major role
is played by the large amplitude rolling motion. Unfortunately, the hydro-
mechanical modelling of ship loss is still incomplete, so that the connection
between large amplitude rolling and capsizing is not very clear. In principle,
the large amplitude rolling motion of a ship can, in fact, be a stable motion,
provided it is included in some stability boundary. Inpractice, the ship is a very
complicate system, so that many dramatic scenarios can appear once large
amplitude rolling is in some way originated. Actually, it is not very simple to
quantify the relative importance of these consequencies of large amplitude
rolling. On the other hand, neither the probability of large amplitude rolling has
been stated in a satisfactory way.
In previous papers [27-29], it has been shown that large amplitude rolling
in non extreme seas can be a consequence of a jump between the
antiresonant state and the resonant one, or between non resonant and
subhiarmonic states, due to the strong deviations from linearity. This originates
the possibily of bifurcations that can be effective as a consequence of some
change in the parameters in the case of a deterministic excitation, and due to
the very nature of the oscillation when a narrow band stochastic excitation is
considered.
Since the phenomenon is quite sudden, and the amplitude differences
can be dramatic also in presence of non very intense excitation, a parametric
research to find the probability of its occurrence was initiated. The goal is to
increase the knowledge of the probability density function of the response in
the following three cases of interest:
-bifurcation between the anti- and the resonant oscillation in the main
resonance region in stochastic beam sea:

-bifurcation between non resonant and subharmonic in the first

subharmonic region in stochastic beam sea;
- bifurcation between zero amplitude (or negligible one) and parametric
suharmonic rolling in stochastic following or quartering sea.
As a consequence of the possibility of bifurcatians, the probability density
functions can be bimodal functions for some values of the parameters. This
allows an evaluation of the probabilities of the considered different oscillation
states, and in particular of the probability of the onset of large amplitude
rolling.
In this paper, the attention will be devoted to the first two.

3.2 The sea description

In structural analysis a failure is usually regarded either as a direct
consequence of an overloading or as a consequence of fatigue. This latter is
considered as the effect of a periodic or in any case time-varying load of
intensity lower that that implied in the first.
Capsizing and sinking are also failures, often improperly regarded as a
pure hydrodynamic problem related to ship motions and stability, and thus of
exclusive pertinence of Naval Architecture. In some limiting cases, and almost
always in safety criteria, attention is mainly devoted to hydrostatic aspects
only.
Actually, in the opinion of the author, the linking between the two aspects
should be considered with greater attention, especially in the light of the fact
that many cases of shi 'p loss can be reconducted to structural failure induced
by motions. This can be due to slamming, sloshing, cargo shifting. In particular
this latter can be caused by overloading or fatigue in the lashings.
A part the slamming, that is connected with vertical motions, the other
phenomena are related to rolling. This is the reason why large amplitude
rolling motion is of interest not only when it determines by itself a stability
overloading. It could be in principle a stable, although very unpleasant,
oscillation regime, mainly affecting the ship's operational capability. Actually,
the ship is not a rigid body, so that sloshing, shifting of cargo, a gust of wind
[30] or water on deck [311 could be sufficient to worsen the stability conditions
sufficiently to reach some non-return point.
Different scenarios can lead to large amplitude rolling. Among them,
worib noting are the effect of extreme beam sea or the effect of a moderate
sea, beam or quartering, of appropriate frequency through one of the
mechanisms described in the Introduction. Whereas the action of an extreme
sea, with breaking waves, can actually be approached only by means of
experimental approaches, in the following the attention will be devoted to the
second case that we will refer as the 'resonance and bifurcation scenario'.
Resonance and bifurcations are a consequence of a certain degree of
autocorrelation in the time record representing the stochastic excitation. This
means, strictly speaking, that consistent parts of the record, or subrecords,
have to exhibit pseudosinusoidal characteristics. Of course, we will not
consider the extremely unrealistic case of pure sinusoidal or monochromatic
sea.
High autocorrelation means small spectral bandwidth, so that attention
will be focussed on the narrow band sea spectra, having in mind the fact that it
can be sufficient that a few consecutive cycles exhibit similar characteristics to
have resonance and possibly bifurcations. In a previous paper [32] it was
shown that such spectra can be well represented by white noise filtered
through cascades of linear filters (one or two can be appropriate). Limiting to
the simpler case, one has the following representation of the spectrum:
Sf = 3)2So/[(f2-02)2+(.2 - ]

obtained shaping Gaussian white noise of level So.(Fig. 10) The filter
"damping" y represents the bandwidth and allows the excitation to recover the
sinusoidal case as y-*O; cof is the centerpeak frequency. The excitation so
represented has zero mean and variance a1f2=-ES o .
In the analytical part of this study, the narrow band sea will be
represented in the Stratonovich's form, by means of a couple of slowly varying
amplitude sinusoidal terms. The amplitudes will be represented by two
independent Gaussian processes. Numerical computations will refer to filtered
white noise. In the very narrow band case, a square box filter was used, as it
simplifies the choice of the cutting frequencies and reduces the number of
harmonics required.
A direct analysis of the local autocorrelation was conducted on long time
records used as excitation in the time domain simulations. The results
indicated that for y=0.005 the record is totally pseudosinusoidal within 10%
accuracy, whereas for the peaked JONSWAP, corresponding to y=0.14 the
pseudosinusoidal aspect seems practically limited to short subrecords.
As regards the realizability of such narrow band spectra, we recall the
long discussion leading to the formulation of the JONSWAP standard
spectrum in the search of a more pronounced sharpness than in Pierson-
Moskowitz, in agreement with many observations at sea. Now there is still
some hint in this sense. Moreover, we have to remember the intrinsic accuracy
limitation that is introduced in the formulation passing from observations to
spectrum [331. Finally, we have not to forget the fact that, when considering
seakeeping or seakindliness, the analysis in terms of extreme and significant
values is in general sufficient. The first relate to long time scales, whereas the
second allow short time forecasts. When dealing with stability and safety from
capsizing, on the other hand, we have to consider all the possible dangerous
phenomena. These are not only connected with extreme values of the
excitation, but also with pseudosinusoidality. This latter is a local (in time)
characteristic whose distribution in practical time records is to great extent
unknown apart the very interesting approach to the description of the statistics
of successive wave periods through a two-dimensional Weibull distribution
[34].

3.3. Analytical results

Large amplitude, non extreme, rolling in beam sea can be computed
through the use of analytical/numerical approaches based on relatively
simplified models. In particular, the rolling motion can be considered
uncoupled and described by a second order nonlinear equation of the type:

2(p
Idd-- d(p
• + D(9, "•) + A• GZ((p) = F(t)

that, dividing by I, representing the nonlinear quantities D and G through

polynomials up to the third order and introducing as usual the natural

frequency coo= I becomes [27]:

d 2x
(2gp+601x2 ) dx
2--+ L- dx +
(2-
(2 X + a3x3 = f(t)

The roll equation was written in the absolute rolling angle, since this is
the only one meaningful in the stochastic case. It allows quite interesting
computations in the frequency domain by means of the use of perturbation
methods.
In particular, the use of the method of multiple scales allowed to obtain
approximate expressions valid for the variances of the stationary stochastic
state. Only the final results will be reported here. For more details on the
analytical procedure, see Ref. [35-37].
In the region of synchronism w coo:

- the roll variance a as a function of the tuning ratio 4 / wo;

2-_o 0 2 - n 3o 2 )2 + (2(wogeq) 2 (1 +Y/21.eq) 2 ]CF 2

Gf2 = 1+[;%2oUeq?((

with an equivalent linear damping leq given by:

=
ý1
=-eq + 0eq 02 6eq = ( 8 i +3(00282).

At the orcer of approximation implied in the used method 0r 2 -Wo2 -

2 0o(wf-o)o) and Lhe above equation
for o2 can be transformed in the following
explicit express;on for the frequency shift of the frequency response curve:
 Ot 3 af 2
1 + 8-o2 ± ('-(1+Y/2 I9eq) - (20wo1Ieq) 2 (1 +Y/29.eq)2 )1 /2

-the maximum value am of the roll variance as a function of excitation variance

af can be obtained imposing the reality condition on the square root of the
above expression:

2
(20))2[(g.+ 146eqOm 2 ) 2 (1+y/ 9eq)] = ot2

A similar analysis conducted in the region of the first subharmonic of =-3%o

gives the expressions reported in Ref. [41].

- the quantity a0 representing the roll variance in linear approximation and is

valid for the frequency region far from resonances.

ao2 =(1 +y/2p) af2 / [((02-402 ) 2 + (2i.i) 2 (1+y/2g) 2] =-f2/(O)o2-"co 2 )2

- the excitation amplitude threshold 0 fth for the onset of subharmonic

oscillations as a function of tuning ratio is also given in Ref. [41].
Having in mind the analysis of rolling at intermediate to large amplitude
(Xmax =- 0.8 radians), the righting arm model was truncated to the cubic term. To
get characteristic values of the nonlinear term coefficient (the natural
frequency was set to the value co = 1, corresponding to an oscillation period of
2n), a parametric research was conducted on a specimen of 13 ships in
different loading conditions [29]. The results of a best fit of the cubic polynomial
up to different heeling angles indicated that in the range of interest the values
a3 = 4.0 and a3 = -0.5 could be representative of different situations. The first is
quite common in modern containerships, whereas the second istypical of
most ships in loaded condition.
As regards the damping, intermediate values were chosen, except for the
study of the subharmonic, where a low damping condition was, for the
moment, assumed. A question a part could be that connected with the
assumed model of damping moment. As a result of the perturbative analysis,
an equivalent linear damping -eq is introduced. It depends on the actual roll
variance G 2,so that it is not convenient to use directly a linear model. More
subtle is the distinction between the different nonlinear contributions that here
too appear strictly mixed as in [38]. Both contributions were retained in the
model used in the theoretical analysis in the light of the fact that there is still
some experimental evidence of an angle dependence of damping; moreover,
going to higher orders, the analysis can indicate a separate contribution.
The excitation intensity represented by al = 0.1414, corresponds to a sea
of moderate intensity, well below the intensity corresponding to the breaking
limit for the equivalent sinusoidal wave.
 The deterministic expressions found in [27] are recovered in the limit 7-t40)
as can be checked by inspection. Due to the particuliarities of the employed
method, using a representation based on two slowly varying amplitude
sinusoidal functions and the multiple scales perturbation, it is not very easy to
compare with other results valid for broader band excitation, such as the
stochastic averaging.
The analytical results indicated that the bifurcation scheme is preserved
also in the narrow band stochastic domain, as shown in Figs. 11-2.
An interesting question is constituted by the meaning of the different
rolling variances, corresponding to the same intensity and tuning ratio, of Figs.
11 and Fig. 12, of the curve of maximum variances in synchronism
In the deterministic scheme, the solution was lead back to an initial
values problem, so that the three solutions existing in some frequency range of
the synchronism region can be divided in stable and unstable in the
asymptotic sense. The Van der Pol plane representing the initial conditions
can be divided in two regions containing the corresponding stable solution,
called domains of attraction [28], separated by a separatrix curve passing
through the unstable solution. In the absence of external perturbations
changing some parameter, once started in one of the two domains, the
solution will remain always in the same domain, leading to the stable solution
as steady state. The curve of maximum amplitudes gives the highest of the two
stable oscillation states. Similar considerations can be drawn for the
oscillation in the region of the first subharmonic.
Giving up the deterministic scheme, the strict dependence on initial
conditions has to be renounced to some extent. All the previous conclusions
have to be reviewed in probabilistic terms, looking to the stochastic as a sort of
continuously changing initial condition transient. On the other hand, one has
also the distinction between transient and stationary stochastic. To clarify this
complex matter, an extensive time domain numerical simulation is in progress.
Only the preliminary results will be reported in the following.

3.4 Numerical results

The procedure consisted in the construction of a long record (tmax =-600

periods) of rolling motion through numerical integration of the differential
equation of rolling motion. Then, this record was analysed in terms of filtered
statistics of maxima (analysis of the envelope) to get the pseudlosinusoidal
probability density function (p.d.f.) [39-41].
The results for the synchronism region shown in Fig. 13 exhibit a good
agreement with the frequency domain analytical values. In particular we can
observe that the distribution is bimodal. This indicates that the two "stable"
solutions are no more separated as in the deterministic case and the system
jumps up and down between the two extreme states. On the other hand, the
deterministic behaviour with domains of attraction has to be recovered in the
sinusoidal limit ( 7--40 ). The effect of realistic different initial conditions on the
p.d.f. can be seen in Fic. 14. It seems that the bimodality is preserved, but the
relative importance of the two peaks changes. In the opinion of the author, this
indicates that the sinusoidal limit is recovered through a displacement of the
p.d. between the two peaks and a simultaneous sharpening of the same In the
 limit y-0 this gives a p.d.f. consisting in one delta function located in
correspondence to one of the two stable solutions.
The results in the region of the first subharmonic are shown in Fig. 15. The
pdf is also in this case bimodal. The onset of the subharmonic is evidenced not
only by the presence of the second bump in the pdf. The analysis of the time
record in terms of the period of rolling (not reported in the figures) indicated wide
intervals where the oscillation was pseudo-synchronous with the natural
oscillations of the ship ( period =_2x/o 0o ) that is characteristic of resonance,
instead than the forcing period (=-2p/" =_2p/3woo ).
The presented numerical results for the subharmonic region, for the moment,
refer to a case with very narrow band and quite low roll damping ability. It seems
that the onset of a subharmonic oscillation is restricted to such limiting cases.
Anyway, it is difficult to reach a definite conclusion, so that the analysis is in
progress.

3.5. The probability density function

The probability density function for the response envelope is given in [39-40]
for a Duffing oscillator (linear damping only) in the form:

p(a) = Const a expf - (1+y/2Geq)ol 2

; 0)f2_(0O 2 - 3 0

j [2 ( 2w 3 -2 + (-y + 21-eq) 2 ] a do)

For particular values of the parameters, this function can exhibit multiple
extrema at the roots of dp/do=O. In the limits of the different approximations
implyed by the schemes adopted (" _coo), these points coincide with the roots of
previous equations. In particular, when this equation has three real solutions, the
pdf has two maxima and a minimum, thus exhibiting a bimodal character. The
probability distribution as shown in Fig. 16 shows a fundamentally different
behaviour with respect to the Rayleigh distribution applicable to the linear and
quasi-linear cases (Fig. 17). The multivaluedness disappears as one moves away
from the multivaluedness frequency window on both sides. It also disappears
when the excitation spectrum level or the nonlinear diameter are reduced. The
number of real solutions is reduced to one only, this case being usually defined
as a mild non-Gaussian behaviour [42-45].
The results can be of particular interest in the evaluation of fatigue behaviour
of hull/hull integrity and cargo securing/shifting as indicated in Section 2.2 [461, an
item of great interest in the consideration of ship safety, as indicated by Table 1.
REFERENCES
[1] Takaishi Y., "Consideration on the Dangerous Situations Leading to
Capsize of Ships in Waves", Proceedings 2nd International Conference on
Stability of Ships and Ocean Vehicles STAB'82, Tokyo, 1982, pp. 243-253.
[2] Morrall, A., "Capsizing of Small Trawlers", Trans. RINA, Vol. 122, 1980, 71-
101.
[3] Duong, C., "Fishing Vessel Stability, a State of the Art Review", Bureau
Veritas, Techn. Paper 88/2, 1988.
[4] Proceedings International Conference on the SAFESHIP Project, Ship
Stability and Safety, RINA, 1986.
[5] Nedrelid, T., Jullumstro, E., "The Norwegian Research Project Stability and
Safety for Vessels in Rough Weather", Proc. 3rd International Conference
STAB'86, Gdansk, 1986, Vol.1, 145-155.
[6] Patullo, R.N.M., "The BSRA Trawler Series (Part I)", TRINA, Vol.107, 215,
1965.
[7] Patullo, R.N.M., "The BSRA Trawler Series (Part II)", TRINA, Vol.110, 151,
1968.
[8] International Conference on Safety of Fishing Vessels 1977, I.M.O., London
1983.
[9] Intact Stability Criteria for Passenger and Cargo Ships, I.M.O., London
1987.
[10] Boccadamo, G., Cassella, P., Mauro, S., Scamardella, A., "A New
Methodology in Order to Verify the Ship's Stability in the Preliminary Design
Stage", Proceedings International Symposium Prads'92, Newcastle upon
Tyne, 1992, Vol. 2, pp. 1173-1186.
[11] Francescutto, A., Nabergoj, R., "Analyse Parametrique de la Stabilit6
d'une Famille de Bateaux de Pdche", Bulletin de I'Association Technique
Maritime et Aeronautique, Vol. 90, 1990, pp.63-81.
[12] Vassalos, D., "A Critical Look into the Development of Ship Stability
Criteria Based on Work/Energy Balance", TRINA, Vol.128, 217, 1986.
[13] Norrby, R., "Stability Problems of Coastal Vessels", Int. Shipb. Progress,
Vol. 11, 1964, pp. 121-132.
[14] Cardo, A., Francescutto, A., Nabergoj, R., "Subharmonic Oscillations in
Nonlinear Rolling", Ocean Engng., Vol.11, 1984, pp.663-669.
[15] Allievi, A. G., qalisal, S. M., Rohling, G. F., "Motions and Stability of a
Fishing Vessel in Transverse and Longitudinal Seaways", SNAME Spring
Meeting/STAR Symposium, New York, 1988, pp. 13-31.
[16] Dunwoody, A. B., "Roll of a Ship in Astern Seas - Response to GM
Fluctuation", J.Ship Res., Vol. 33, 1989, 284-290.
[17] Bishop, R.E.D., Neves, M. de A.S., Price, W.G., "On the Dynamics of Ship
Stability", Trans. RINA, 124, 1982, 285-302.
[18] Bishop, R.E.D., Price, W.G., Temarel, P., "On the Dangers of Trim by the
Bow", TRINA. Vol. 131, 1989, pp. 281-303.
[19] Francescutto, A., Reggente, S., Armenio, V., "On the Possibility of Loss of
Control and Broaching of Fishing Vessels", Symposium Technics and
Technology in Fishing Vessels, Ancona, May 1989.
[20] Inoue, S., Hirano, M, Kijima, K., "Hydrodynamic Derivatives on Ship
Maneouvrability", International Shipbuilding Progress, 28, 1981, 1-14.
 [21] Fujino, M., "Lectures on Ship Manoeuvrability - Prediction of Manoeuvring
Performance", University of Tokyo, Technical Report No. 6013, 1985.
[22] Motora, S., Fujino, M., Konoyagi, M., Ishida, S., Shimada, K., Maki, T., "A
Consideration on the Mechanism of Occurrence of Broaching-to Phenomena",
Naval Architecture and Ocean Engineering (Japan), 20, 1982, 92-107.
[23] Bao-an Y., "Ein Beitrag zur Beurteilung der Stabilitat schneller Schiffe bei
gekoppelter Gier-, Quer- und Rollbewegung", Schiffstechnik, Vol. 31, 1984, pp.
22-42.
[24] Francescutto, A., "Is it Really Impossible to Design Safe Ships", Spring
Meetings of The Royal Institution of Naval Architects, London, 27-29 April
1992, paper n. 3, To appear on Transactions of The Royal Institution of Naval
Architects.
[25] Cardo, A., Francescutto, A., Zotti, I., Mattioli, R., "Experimental Study of the
Effect of Water on Deck on the Stability of a Fishing Vessel", Proceedings
International Symposium NAV'92, Genova, July 1992, Vol. 1, pp. 3.3.1-3.3.14.
[26] Cardo, A., Francescutto, A., Armenio, V., Contento, G., "Dynamic Effects of
Liquids on Board on the Stability of a Fishing Vessel", In Preparation for
OTRADNOYE'93, Kaliningrad, May 1993.
[27] Nabergoj, R., "Small Vessel Optimization for Increased Seakeeping and
Stability Performance", Proceedings 4th International Conference on Stability
of Ships and Ocean Vehicles - STAB'90, Napoli, 1990, Vol. 2, pp. 597-603.
[27] Cardo, A., Francescutto, A., Nabergoj, R., "Ultraharmonics and
Subharmonics in the Rolling Motion of a Ship: Steady-State Solution",
International Shipbuilding Progress, Vol. 28, 1981, pp. 234-251.
[28] Cardo, A., Francescutto, A., Nabergoj, R., "Deterministic Nonlinear Rolling:
A Critical Review", Bulletin de I'Association T~chnique Maritime ed
Aeronautique, Vol. 85, 1985, pp. 119-141.
[29] Cardo, A., Francescutto, A., Nabergoj, R., "The Excitation Threshold and
the Onset of Subharmonic Oscillations in Nonlinear Rolling", International
Shipbuilding Progress,Vol. 32, 1985, pp. 210-214.
[30] Dahle, E. Aa., Myrhaug, D., Dahl, S. J., "The effect of Wind on Small
Vessels", Proc. International Conference STAB'90, Napoli, 1990, Vol. 1, pp.
191-199.
[31] Falzarano, J. M., Troesch, A. W., "Application of Modern Geometric
Methods for Dynamical Systems to Problem of Vessel Capsizing with Water-
on-deck", Proceedings 4th International Conference on Stability of Ships and
Ocean Vehicles - STAB'90, Napoli, September 1990, pp. 565-572.
[32] Francescutto, A., Cardo, A., Contento, G., "On the Representation of Sea
Spectra Through Linear Filters" (in italian), Tecnica Italiana, Vol. 56, 1991, pp.
1-10.
[33) Guedes Soares, C., Trov~o, M. F. S., "Influence of Wave Climate
Modelling on the Long-term Prediction of Wave Induced Responses of Ship
Structures", Proceedings of the IUTAM Symposium on Dynamics of Marine
Vehicles and Structures in Waves, London 1990
[341 Myrhaug, D., Rue, H., "Note on a Joint Distribution of Successive Wave
Periods", To appear.
[35] Rajan, S., Davies, H. G., "Multiple Time Scaling of the Response of a
Duffing Oscillator to Narrow-Band Random Excitation", Journal of Sound and
Vibration, Vol. 123, 1988, pp. 497-506.
[36] Davies, H., G., Rajan, S., "Random Superharmonic and Subharmonic
Response: Multiple Time Scaling of a Duffing Oscillator", Journal of Sound
and Vibration, Vol. 126, 1988, pp. 195-208.
[37] Francescutto, A., Nabergoj, R., "A Stochastic Analysis of Nonlinear Rolling
in a Narrow Band Sea", Proceedings 18th ONR International Symposium on
Naval Hydrodynamics, Ann Arbor, August 1990.
[38] Cardo, A., Francescutto, A., Nabergoj, R., "On Damping Models in Free
and Forced Rolling Motion", Ocean Engineering, Vol. 9, 1982, pp. 171-179.
[39] Davies, H. G., Liu, Q., "The Response Envelope Probability Density
Function of a Duffing Oscillator with Narro Band Excitation", To appear on
Journal of Sound and Vibration.
[40] Davies, H. G., Liu, Q., "On the Narrow Band Random Response pdf of a
Nonlinear Oscillator", To appear.
[41] Francescutto, A., "On the Probability of Large Amplitude Rolling and
Capsizing as a Consequence of Bifurcations", Proceedings 10th International
Conference on Offshore Mechanics and Arctic Engineering 'OMAE',
Stavanger, June 1991, Vol. 2, pp. 91-96.
[42] Ness, 0. B., McHenry, G., Mathisen, J., Winterstein, S. R.(1989).
"Nonlinear Analysis of Ship Rolling in Random Bean Waves", Proc. Research
Workshop on Stochastic Mechanics, Technical University of Denmark,
Technical Report N. 244/R, pp. 49-66.
[43] Juncher Jensen, J. (1989). "On Fatigue Damage due to Non-Gaussian
Responses", Danish Center for Applied Mathematics and Mechanics, Report
N. 389.
[44] Juncher Jensen, J. (1991). "Fatigue Analysis of Ship Hulls under Non-
Gaussian Wave Loads", Marine Structures, Vol. 4, pp. 279-294.
[45] Francescutto, A., "Stochastic Modelling of Nonlinear Motions in the
Presence of Narrow Band Excitation", Proceedings International Symposium
on Offshore and Polar Engineering - ISOPE'92, San Francisco, June 1992,
Vol. 3, pp. 554-558.
[46] Hutchison, B. L., "Cargo Mechanics (Application of Seakeeping-
Revisited)", Marine Technoloqy, Vol. 23, 1986, pp. 230-241.
[47] Kishev, R., Dimitrova, S., Gaberova, M.,."A Generalized Procedure for
Rank Optimization Application to Ship Design", Proceedings International
Symposium PRADS'89, Varna 1989, pp. 33.1-33.9.
Table 1. Main causes of loss for capsize and broaching regarding
Japanese fishing vessels in the period 1973-77

Main Cause of Accident N

Overloading 9
Center of mass too high 13
Insufficient securing of cargo 1
Improper stowing of cargo 7
Cargo shifting 13
Openings not secured 9
Weack hatch covers 2
Hull damage 2
Water trapped on deck 29
Broaching 1
Total 86
 Table 2. Main characteristics of the BSRA trawler series.

MODEL L B T D L/V 1 /3 B/T CB CW XCB%

BSRA I XF 45.72 8.03 4.07 4.95 4.85 1.97 0.562 0.774 0.969
BSRA I XG 45.72 8.50 3.85 4.77 4.85 2.21 0.561 0.776 0.960
BSRA I WO 45.72 8.96 3.64 4.63 4.85 2.46 0.562 0.776 0.962
BSRA I WP 45.72 9.40 3.48 4.51 4.85 2.71 0.562 0.777 0.965
BSRA 1907 45.7210.62 3.09 4.21 4.85 3.44 0.560 0.777 0.949
BSRA I WS 41.01 8.98 4:07 5.04 4.35 2.22 0.559 0.772 0.678
BSRA I WR 43.36 8.73 3.95 4.90 4.60 2.22 0.560 0.765 1.095
BSRAIWQ 48.09 8.30 3.75 4.46 5.10 2.22 0.561 0.775 0.959
BSRA IIZP 45.72 8.03 4.08 4.97 4.92 1.97 0.521 0.742 0.888
BSRA IIZQ 45.72 8.03 4.06 4.95 4.76 1.97 0.593 0.806 0.906
BSRA 11851 45.72 8.32 4.23 5.15 4.85 1.96 0.521 0.743 0.892
BSRA 11852 45.72 7.82 3.96 4.81 4.85 1.97 0.593 0.805 0.903
BSRA 11975 45.72 8.50 3.69 4.77 4.85 2,30 0.584 0.816 4.689
BSRA II977 45.72 8.50 3.72 4.76 4.85 2.29 0.580 0.795 0.693
BSRA II978 45.72 8.50 3.74 4.75 4.85 2.28 0.577 0.790 -0.875
minimum 41.01 7.82 3.09 4.21 4.35 1.96 0.521 0.742 -0.875
maximum 48.0910.62 4.23 5.04 5.10 3.44 0.593 0.816 4.689
 Table 3a. KGmax (m) according to IMO statistical criterion.

MODEL STANDARD
A B C D E F
BSRA I XF 3.488 3.315 3.233 3.172 3.216 3.166
BSRA I XG 3.588 3.423 3.320 3.232 3.209 3.238
BSRA I WO 3.769 3.588 3.455 3.327 3.227 3.357
BSRA I WP 3.984 3.775 3.614 3.449 3.269 3.505
BSRA I 907 4.777 4.413 4.123 4.107 3.440 3.959
BSRA I WS 3.882 3.607 3.509 3.432 3.387 3.156
BSRA I WR 3.696 3.527 3.414 3.315 3.286 3.326
BSRA I WQ 3.490 3.333 3.238 3.165 3.134 3.438
BSRA II ZP 3.456 3.187 3.132 3.113 3.245 3.070
BSRA II ZQ 3.645 3.179 3.124 3.103 3.213 3.058
BSRA II851 3.597 3.306 3.256 3.243 3.363 3.188
BSRA 11852 3.359 3.082 3.034 3.023 3.126 2.970
BSRA 11975 3.577 3.261 3.192 3.153 3.229 3.116
BSRA II977 3.537 3.224 3.158 3.123 3.199 3.084
BSRA 11978 3.522 3.209 3.145 3.113 3.184 3.070
 Table 3b. KGmax/D according to IMO statistical criterion.

MODEL STANDARD
A B C D E F
BSRA I XF 0.704 0.669 0.653 0.640 0.649 0.639
BSRA I XG 0.752 0.717 0.696 0.677 0.672 0.678
BSRA IWO 0.815 0.776 0.747 0.719 0.698 0.726
BSRA I WP 0.884 0.838 0.802 0.765 0.725 0.778
BSRA 1907 1.134 1.048 0.979 0.975 0.817 0.940
BSRA I WS 0.770 0.715 0.696 0.681 0.672 0.626
BSRA I WR 0.754 0.720 0.696 0.676 0.670 0.679
BSRA I WO 0.782 0.747 0.725 0.709 0.702 0.770
BSRA IIZP 0.696 0.641 0.630 0.626 0.653 0.617
BSRA IIZQ 0.700 0.642 0.631 0.627 0.649 0.618
BSRA II851 0.698 0.642 0.632 0.629 0.653 0.619
BSRA 11852 0.698 0.640 0.630 0.628 0.677 0.617
BSRA II975 0.749 0.683 0.669 0.661 0.677 0.653
BSRA II977 0.743 0.677 0.663 0,656 0.672 0.648
BSRA II 978 0.742 0.676 0.662 0.655 0.670 0.646
 Table 4. Maximum allowable height of the centre of gravity above keel.

MODEL IMO STATISTICAL IMO WEATHER

KGmax KGmax/D KGmax KGmax/D
BSRA I XF 3.172 0.640 3.005 0.606
BSRA I XG 3.209 0.672 3.046 0.638
BSRA I WO 3.227 0.698 3.107 0.672
BSRA I WP 3.269 0.725 3.225 0.716
BSRA I 907 3.440 0.817 3.659 0.869
BSRA I WS 3.156 0.626 3.268 0.648
BSRA I WR 3.286 0.670 3.125 0.638
BSRA I WQ 3.134 0.702 2.968 0.665
BSRA IIZP 3.070 0.617 3.058 0.615
BSRA II ZO 3.058 0.618 2.991 0.604
BSRA II851 3.188 0.619 3.180 0.617
BSRA 11852 2.970 0.617 2.904 0.603
BSRA 11975 3.116 0.653 3.039 0.637
BSRA II977 3.084 0.648 3.013 0.633
BSRA II 978 3.070 0.646 3.001 0.632
FIGURE CAPTIONS

Fig.1 - The I.M.O. weather criterion according to Ref.7.

Fig. 2 - The method of energetic balance of Strathclyde.

Fig.3 - Comparison of KGmax/D versus main design parameters for different
stability criteria.
Fig.4 - Comparison of KGmax/D versus Kempf's adimensional rolling period
TR(g/B)lQ2 according to I.M.O. stability criteria.
Fig.5 - Variation of the eigenvalues of the characteristic equation for
GM/LFn 2 =0.1 as a function of trim parameter y for model XF. Positive values of
the eigenvalues indicate instability.
Fig.6 - Eigenvectors for the model XF in the trim by bow condition as a function
of GMILFn 2. Yaw is unitary, a indicates sway and 8 roll.

Fig.7 - Variation of the eigenvalues of the characteristic equation for

GM/LFn 2=0.1 as a function of trim parameter y for model 1907. Positive values
of the eigenvalues indicate instability.
Fig. 8. Stability rank index relative to the application of energy balance method
versus rank index for seakeeping (vertical motions) for the same family of
fishing vessels represented in Fig.4.
Fig. 9. Stability rank index relative to the application of energy balance method
versus rank index for seakeeping (vertical motions) for another family of fishing
vessels of the BSRA series.
Fig. 10. Normalized JONSWAP spectrum corresponding to unitary value of the
peack enhancement factor (solid curve). The dashed curve represents the fit
obtained using the rational representation as a linear filter.

Fig. 11 Roll variance a as a function of tuning ratio "o./ o . The following values
have been used for the parameters: ca3 =4.0, g=0.05, ca=0.1l414 (ew=0.2). The
number on the curves indicates the value ofy.

Fig. 12 Roll variance a as a function of tuning ratio "f/%o . The following values
have been used for the parameters: cx3 =-0.5, p=0.05. a1=0.1414 (ew=0.2). The
number on the curves indicates the value of y.

Fig. 13. Comparison of the analytical results in the synchronism region with the
time domain simulation. The following values have been used for the
parameters: X3=4.0, p=0.05, o1f=0.1414, y=0.02. The value o)f/o)o = 1.55 was
used in the numerical computations. Initial conditions corresponding to xo=-0.3
and X0 =O.
Fig. 14. Dependence of the probability density function on the initial conditions
in the synchronism region. Solid curve refers to xo=O. and *o=0., whereas
dashed curve referes to xo=-0.3 and ýo=0. The following values have been
used for the parameters: a3=4.0, g.=O.05, uf=O.1414, "t=0.005, wh/oo = 1.55.

Fig. 15. probability density function in the region of first subharmonic. The
following values have been used for the parameters: a 3 =4.0, g=0.005,
af=0.1414, y'=0.005, wf/o
1 o = 3.2. Initial conditions corresponding to xo=0.15
and ;o=-0.15.

Fig. 16. Envelope probability density function in the presence of bifurcations as

obtained from the analytical approach. The following values were used for the
parameters: co=1., p.=.05, uf=.141 4 ,y=.02, Mfl)o
0=1.40, a3=4.0.

Fig. 17. Envelope probability density function in absence of bifurcations as

obtained from the analytical approach. The following values were used for the
parameters: coo=1., g=.05, of=.1 4 1 4 , -=.02,"0/c)o=1.40, a3=0. (linear system).
GZ

i 2

~~\T
 STATISTICAL
-- WEATHER
0.9 - STRATHCLYDE 70%

0-8

0.7- 50%

0.6-~. . .

a.5
1.5 2.0 2.5 3.0 3.5 4.0

1.0

KG/D 0% 907

0.69w

50%

0.7 100% .X

XF

0.6-
SSTATISTICAL
WEATHER
STRATHCLYDE
0.5j
4 6 8 10 i-I? 14

vl 4 4
4-

1-0 -

0 GILFý0.1
10 au\

-11.0

5-
XE
4- K;MIF/LE=o.1

-1.0 -50 0.5 1.0

Týý (
 1.0-
907

/l
.5 O

-1. 10 0 1.10

1QI.L2

(0)3

-1.0-

12-

2610 12

Sekein ne
 6
4
C

35
.0

2-

0
0 1 2 3 4 5 6

Seakeeplng Index

00

0-

[]- 22TOTAL RANKING

IM-PRRTIAL RANKING BASED ON SHIP BEHAVIOUR IN BEAMOSERS
rM-PARTIRL RANKING BASED ON SHIP BEHAVIOUR IN HERD SEAS
R' F-]-PARTIRL RANKING BASED ON STILL WATER RESISTANCE

akepn
Se. Ine

01 2 3 4 51 6 in 11 12

UFIRIRNT No

Partial versus general rank evaluations of design variants

LLJ
CL

8- .0 1I I
lo

0.80 1.•60 2.40 3.20 4. CC

OMREL 0.

Cd,4

000.

1.0 1.5 O)f~o~o 2.0

 0.0 0.01

0.5-

05
0.0 ro ~f/.1

0.5-

7. L 2.0 pdf (a~u.)

pdf

o./
0.. 0.5 XM to

pdf
(a.u)

0. 0.7 0.2 xm 0.3

 1.2

1.0

0.8

CL 0.6

0.4

0.2

0.0
u.0 0.2 0.4 0.6 0.8 1.0

1.0

0.8

0.6

0.4

02

0.0IL
u.0 0.2 0.4 0.6 a 0. 8 1.0
 21 (vxirdjscfko., I t- lwrdL,h
/14 1toWajLe

(91) (~

64 ý0 ZD
P K
T / TLA

IVFke-rpc&.[2-w LeeL-2
o
LsQ'~4~go..:L2 7e 6]

5
C5
3
10 a
2 63
1 0

00.1 02 0.3 0.4 &J/ILFg'

0Y-
9t(AIAb1W LŽ T- qt-~ tett&
WEGEMT Short Course on:
"New Techniques for assessing and Quantifying Vessel Stability and
Seakeeping Qualities"
Trondheim, 8-11 March 1993

"DYNAMIC STABILITY - PROBABILISTIC APPROACHES"

Alberto FRANCESCUTTO

Department of Naval Architecture, Ocean and Environmental Engineering

University of Trieste
Via A. Valerio 10, 34127 TRIESTE (Italy)

ABSTRACT

The hydrodynamic part of ship safety is actually managed by means of

prescriptive stability criteria based on static stability characteristics of the ship.
The analysis reveals that this approach leads to many contradictions whose
final effect is to some extent the delay of the formulation and adoption of higher
safety standards. Only the development of a fully physical approach could
allow a realistic evaluation of the probabilities of ship loss through the different
mechanisms. The final goal will be an effectively unified approach to ship
safety.
In the first part of the paper, the safety problem of fishing vessels is
considered in connection with the possibility of their loss in waves. To this end,
different dynamic mechanisms are studied in detail, ranging from capsizing to
loss of control and/or broaching-to phenomena.
The stability of a family of fishing vessels is investigated in the light of
the existing I.M.O. criteria and of the new method of energy-balance proposed
by Strathclyde University is illustrated. Then the set of linear differential
equations describing the dynamics of the antisymmetric motions of roll, yaw
and sway is solved to look for possible unstable solutions. It results that, in
general, I.M.O. statistical criterion is less restrictive than I.M.O. weather
criterion and both fall into a region of approximately 20-30 percent of net area
positive according to energy-balance method. An analysis of the stability of the
antisymmetric motions reveals the possibility of dangerous phenomena
depending on the trim condition.
For the same family, an analysis of the threshold for the onset of
parametric rolling in a following sea reveals the extreme sensitivity of this type
of vessels to such a particular capsizing mechanism.
These results are then critically reviewed in the second part of this
paper to look for an answer to the currently unsatisfying state of art of small
ship safety.
The loss of a ship is often a direct or indirect consequence of large
amplitude rolling. In the third part of this paper nonlinear rolling is thus
examined from a probabilistic point of view. The nonlinear rolling in a
stochastic sea is analysed by means of an approximate perturbation method.
The results indicate that the possibility of bifurcations to resonant or
subharmonic large amplitude rolling is possible in a narrow band sea. A
numerical time domain simulation confirms these results and gives an
 indication on the probability denstities of states, that in the sinusoidal limit
should recover the domains of attraction. The results can be of great interest in
cargo mechanics modelling, an item relevant to ship safety.

INTRODUCTION
Safety at sea is a complex goal to achieve. This is due to the extreme
variety of ships and structures and the different causes that can give rise to
casualties at sea.
Depending on the dominant aspects and on the level of knowledge,
these causes are divided in different categories. An important role is usually
attributed to the structural aspects and to ship stability.
Stability is the part of Naval Architecture traditionally considered to give
the tools for a correct approach to the safety aspects connected with the
possibility of ship loss through the mechanisms of capsizing, sinking and surf
ridlinging/broaching-to.
( As will be shown in the following, this approach and the safety rules
based on it, are to some extent inconsistent. The introduction of a correct
approach and consequently of the more appropriate name of "hydrodynamic
part of ship safety", will improve the actually unsatisfying level of safety for life,
cargo and environment at sea. In addition, this goal could probably be
obtained without going further in the way of paying increased safety levels with
unacceptable reductions of payload and/or operational capability. This could
originate a positive feedback making more attractive the improvements of
safety in ship design introducing the concept of "hydrodynamic safety
performance" and a subsequent effective "design for safety".
This program rests on the development of a fully physical approach to
ship safety through which the probabilities of the different dangerous
phenomena can be computed. The concepts and methods of reliability could
then be applied to obtain risk evaluation and control.
In this paper, the contradictions of current approach are highligted
together with their negative effect on developments and progress in the field of
ship safety.
Many aspects of ship hydrodynamics together with the consideration of
ship structural aspects and their mutual interaction are involved in the
previsions about ship safety. In particular, ship stability, manoeuvrability and
their interactions, cargo mechanics, water on deck, loss of hull integrity, etc.,
can play an important role. Since it is difficult to deal with all these items in a
short paper, so that, in the following, main consideration will be given to:
- MO criteria with a typical application on a family of fishing vessels;

- the method proposed at Strathclyde University regarding the energy

balance in waves to take into account the loss of stability in waves;
-the miethod proposed at Brunel University regarding the stability of
antisymmetric motions as one of the ways to take into account the possible
interactions between transversal and directional stabilities;
 -the nonlinear rolling in deterministic and stochastic sea to take into
account the possible effects on cargo mechanics with shifting of cargo, loss of
structural integrity, etc.
To have an idea of the importance of this discussion in the case of
fishing vessels, a look to the statistics of casualties reported in Table 1 is
sufficient. In Table 1 the casualties leading to capsizing and broaching (about
60% of casualties) regarding Japanese fishing vessels in the period 1973-77
[1] are analysed with respect to the causes.

1. IMO CRITERIA AND OTHER RECENTLY PROPOSED METHODS

2.1. Introduction
The problem of the safety of navigation at sea is very difficult to handle
( in particular for ships whose characteristic dimensions render them more
sensitive to the action of the marine enviroment and, among these, a relevant
place is reserved to fishing vessels [2,3]. The hull shape of a fishing boat
vanies greatly due to different local conditions, fishing methods, construction
material, engine weights, distance to the fishing grounds and other factors.
Thus, it is obviously very difficult to design a few standard hulls which are
suitable for all conditions.
In the last decades, different organizations have collected and
published results of theoretical calculations, model tests and full scale trial in
an attempt to indicate the trend in the factors which influence stability,
resistance, powering and seakeeping qualities of the vessels. The final goal is
to reach conlusive indications on how to find the optimum hull shape when
designing a new fishing boat. The statistical analysis of the data available is
also intended for estimating the total performance of an existing design so it
can be investigated if there is still room for any improvement.
At international level, IMO has devoted particular attention to small ships
and in general to ships less than 100 m in length and, among these, a relevant
place has been reserved to fishing vessels. Unfortunately, the stability rules
adopted by the main Classification Societies still belong to a quite old
approach, being mainly based on Rahola's statistical results requiring
prescriptive characteristics of the curve of righting arm curve in calm water.
The new proposals [4,51 suggest the extension of the weather criterion to such
ships and are presently included in the Torremolinos Convention, which is yet
not approved by a sufficient number of State Governments to become an
international rule. It would be interesting to investigate the reason of this
unsatisfactory state of art, even if economical and political choiches seem to
constitute the major problem for further progress.
In the first part of this section, the stability of a family of the BSRA trawler
series is investigated in the light of the existing criteria, i.e. the l.M.O. statistical
criterion and the .M.O. weather criterion. In the calculations, a group of 15
vessels was considered in order to determine the effect on stability of varying
different design parameters.
In the second part of this section, the effect on stability of a longitudinal
wave is taken into account following the Strathclyde approach arid
considering the effect of the parametric resonance in following sea. Finally, the
possibility of other ship loss mechanisms, connected with directional stability
such as broaching-to is examined through an analysis of the stability of
antisymmetric motions.

1.2. The family of vessels

The numerical investigations were carried out on a family of fishing
vessels known as the BSRA trawler series [6,7]. The parent form for the series,
model XF, was chosen to represent a ship 150-ft. LBP x 26-ft. 4 in. moulded
breadth x 13-ft. 2 in. moulded draught with a displacement of 847 tons salt
water.
The hull forms of the series were derived from the parent form according
to the following transformations:
- affine distortion of B/T (B/L, T/L) starting from ship XF for hull forms XG,
WO,
WP, 907, while keeping L/V 113 constant;
- affine distortion of LJV 113 (B/L, T/L) starting from ship XG for hull forms WS,
WR, WO, while keeping B/T and V constant;
- variation of CB (Cp) by conformal transformation of the sectional area curve,
starting from ship XF for hull forms ZP, ZQ, but having main dimensions
constant;
- same as above, starting from ship XF for hull forms 851, 852, while
maintaining B/T and L/VI/3 constant;
- variation of Xc8 from ship XG for hull forms 975, 977, 978, by modifying
the
sectional area curve only.
All the hulls of the family have been normalized to the same
displacement through geometrical similarity and are fitted with similar
superstructures independently on their length. Their main geometrical
characteristics are given in Table 2. The forecastle has an extension of
approximately 25% of the ship's length and the freeboard is sufficient to avoid
deck immersion until a heeling angle of 12,50.

1.3. I.M.O. Statistical criterion

The I.M.O. "statistical" criterion for fishing vessels of 24 meters in length
and over has been endorsed at the "International Convention for Safety of
Fishing Vessels" held at Torremolinos in 1977 [8,9]. The criterion, that derives
from the pionieristic work of Rahola, is written in terms of stability standards
based both on statistical and other analysis of casualty records and on the
experience of different fishing fleets throughout the world. The Standards are
expressed in terms of prescriptive values for certain key features of the righting
arm curve and can be summarized as follows:
- Standard A: The initial metacentric height GM should be not less than 0.35
metres.
-Standard B: The area under the righting leyer curve should be not less than
0.055 metrexradians up to 300.
* - - Standard C: The area under the righting lever curve should be not less than
0.090 metrexradians up to 400 or up to an angle where the non-watertight
openings come under water (whichever is less).
- Standard D: The area under the righting lever curve should be not less than
0.030 metrexradians between the angles of heel from 300 to 400 or such lesser
angle mentioned under Standard C.
- Standard E: The maximum righting lever should occur at an angle of heel
preferably exceeding 300 but not less than 250.
- Standard F: The righting lever should be at least 0.20 metres at an angle of
heel equal to or greather than 300.
To aquire more information about the relative importance of the
requirements introduced by the I.M.O. statistical criterion, we show in Table 3
[11] the maximum allowable height of the centre of gravity above the keel
according to the different Standards. The results indicate the very different
weight of the different Standards. In particular, Standard A, prescribing a
minimum value for the metacentric heigth GM is by far the less resctrictive.
Standard E, relative to the location of the maximum value of the curve GZ is
generally the most restrictive. These results are in agreement with similar ones
found relatively to the fishing vessels of the Ridgely-Nevitt [10].
Inspite of its great simplicity, the statistical approach can be, and indeed
is, criticized on the following basis:
- the statistics refers to a great variety of vessels, loading conditions and
operation areas, so that the concept of "homogeneous specimen", i. e. the
basis itself for the validity of a statistical approach can be questioned;
otherwise, the specimen would belong to so specific situations to be hardly
applicable in general;
- it constitutes an "a posteriori" (hindsight) approach, i. e. in the best
hypothesis, it can only follow the trends indicated in the records of casualties
at sea;
- no physical description of the effective interaction environment-ship is
considered.
We will return on this point in Section 2.2.
1.4. I.M.O. Weather criterion
In the recommended I.M.O. "weather criterion" [8,9] for fishing vessels of
45 meters in length and over in unrestricted service, the ability of a ship to
withstand the combined effects of beam wind and rolling should be
demonstrated for each standard condition of loading, with reference to Fig.1,
by means of the following procedure:
- The ship is subjected to a steady wind pressure acting perpendicular to the
ship's centreline which results in a steady wind heeling lever (1lw).
- From the resultant angle of equilibrium (qo), the ship is assumed to roll owing
to wave action to an angle of roll (ql) to windward. Attention should be paid to
the effect of steady wind so that excessive resultant angles of heel are
avoided. The angle of heel under action of steady wind (qo) should be limited
to a certain angle to the satisfaction of the Administration. As a guide, 160 or
80% of the deck edge immersion, whichever is less is suggested-
- The ship is then subjected to a gust wind pressure which results in a gust
wind heeling lever (Iw2 ).
 -Under these circumstances, area "B" should be equal to or greather than,
area "A".
- Free surface effects should be accounted for in the standard conditions of
loading.
Also in this case the stability of the vessels is investigated in the light of
the proposed criterion. In order to simulate different loading conditions, in the
calculations the height of the centre of gravity above the keel was varied until
the stability requirement failed. As a result we obtain the maximum allowable
height of the centre of gravity for each vessel according to the weather
criterion.
In Table 4 [111, we show the maximum height of the centre of gravity
above keel allowed by the statistical and the weather criterion respectively. It
appears that the statistical criterion is less stringent than the weather criterion.
The weather criterium approach is preferred to the statistical one as it
introduces a description of the interaction environment/ship, so that it appears
more "physical" and not "a posteriori". This is only partly true, since the
description is too much simplified to be sufficiently realistic, and limited to a
particular mode of ship loss, i. e. the capsize mode. In fact, the results shown in
Table 4 indicate that the two methods give comparable results, with a slightly
lower maximum allowable center of gravity heights corresponding to the
second. In the opinion of the authors this should upset definitely the statistical
criterium.

1.5. The energy balance in waves

SThe method recently proposed by Strathclyde University [12] is based
on the consideration of a time-dependent roll restoring moment. To this end, in
addition to the combined effects of beam wind and rolling, also the effect of a
following or quartering sea on the stability of the ship is taken into account. In
the proposed procedure the righting arm is computed for different positions of
a wave with respect to the vessel. The wave is assumed to have the same
length of the ship and the encounter period equal to her natural rolling period.
Then, an ultimate half roll is supposed to occur between a windward angle
and an extreme leeward angle. For each position of the wave relative to the
vessel the minimum righting arm curve during the ultimate half roll is obtained
and the corresponding energy-balance, taking into account damping and
wind, is computed to give the net area, as shown in Fig. 2. The percentage of
time with net area positive during the passage of the wave along the ship is
assumed as an indicator of the safety from capsizing. Futher details can be
found in Ref. 12.
In the application presented in this paper, the parameters describing the
ultimate roll and the wind lever are esimated using the l.M.O. weather
prescriptions. In the first application no roll damping was assumed to act. A
rough estimate made on one of the vessels, indicated that the consideration of
an equivalent linear damping realistic for these ships in the absence of
antirolling devices, could increase the percentage of net area positive by
about 20%/_
The results relative to a group of hulls of the BSRA family to analyse the
effect of the variation of the parameter B/T (the only that was found to be
relevant) are shown in Fig. 3 [11 ], where the KGniax/D is plotted against B/TF. As
K one can see, both I.M.O. criteria lie in a region correspong to about 10-25%
net area positive, whereas the request of 100% net area positive if much more
restictive in terms of the maximum allowable height of the centre of gravity.
In Fig. 4, the KG/D is plotted against the Kempf's [13] nondimensional
rolling period TK=TR(g/B)1' 2 for the considered different hulls. Drawing the
lines connecting the KGmax/D characteristic of the different hulls with respect to
the different stability standards considered, one can see that the request of
greater stability, for exemple a greater percentage of net area positive,
corresponds to pushing the Kempf s period towards lower values.
Actually, it is not recommended to have TK<8 or TK>1 4. In the first case,
the roll behaviour of the ship is too stiff, so that the oscillations have a
moderate amplitude but higher frequency, i.e. great accelerations on the deck,
rendering the operations difficult. On the contrary, a great value of TK gives rise
to a cranck behaviour with large amplitudes and still a loss of operational
capability (and of stability!). We will return on this point in Section 2.2.

( 1.6. The effect of parametric resonance in following sea

The effect of heave-roll coupling on the stability in waves was

investigated by considering the so called parametric roll excitation. As known
[14,15], theoretical predictions, confirmed by experiments, suggest that large
amplitude rolling can be excited in the case of quartering and following seas
when the encounter wave period is approximately half the rolling period and
moreover a threshold condition correlating relative righting arm variation to roll
damping is fulfilled. In design terms, the threshold condition expresses. a
correlation between sea intensity and both hull form geometry and weight
distribution. By lowering the centre of gravity or introducing additional roll
damping, one has the possibility to avoid the phenomenon.
Preliminary computations were done using a simplified, but common approach
based on the assumption of linear time dependent rolling equation, so that the
threshold is given in terms of the comparison of the relative GM variation in a
longitudinal wave and the equivalent linear nondimensional damping. The
isocarenic hypothesis was assumed, corresponding to consider the quasi-
static approach. In the present case, the analysis revealed an extreme
sensitivity of this class of vessels to parametric resonance. In fact, no realistic
values for KG/D were found to prevent the phenomenon also in the case of
vessels having bilge keels of area up to several percent of waterline area.
Thus, the only way to avoid such dangerous situation is linked to the ability of
the master, who must appropriately handle the ship at sea. A more detailed
analysis taking into account the effective heave-roll coupling and the limiting
effect on the roll amplitude played by the righting arm and damping
nonlinearities, could change appreciably these conclusions. Moreover, being
a "resonance" way to large amplitude rolling and eventually to capsizing, it is
necessary, for its build-up, that a sufficient degree of autocorrelation be
present in the wave train, so that it is tied to the appearance of a quite narrow
band sea. The roll of a ship in irregular following sea is analysed by
Dunwoody [161 that describes the effect of GM fluctuations as an equivalent
roll damping reduction. The author discusses also the difierent stability limits
with respect to the onset of parametric resonance connected with the different
levels of stability of a stochastic dynamical system.
 1.7. The stability of antisymmetric motions

Following a method developped at Brunel University [17,18], we

consider a mathematical model suitable for the analysis of the stability of the
antisymmetric motions of the ship that is, in order, sway, roll and yaw. This
method could describe the possibility of ship losses due to broaching-to as it
introduces a coupling between roll motion and directional stability. Since, we
are interested in the onset of the deviations, the system of equations governing
the antisymmetric motions may be written as:

m(v+Ur) = mg4+Yvv+Yvv+Ypp+y•+Yrr+Yr r

lxp-lxzr = Kvv+Kvv+Kpp+Kpp-pgVGMp+Krr+Krr

Izr-lxzp = Nvv+Nvv+Npp+Npp+Nrr+Nrr

The system is linear, the effect of the rudder is not considered; attention is
focused mainly on the accelerated motions and the effect of trim variations is
considered through the derivatives expressing yaw and sway dependence on
yaw and sway velocity. Dimensionless parameters are used through a suitable
transformation. In particular, the parameter GM/LFn2 is introduced to study the
stabilizing effect of increasing the metacentric height and the destabilizing
effect of increasing speed. No frequency dependence is considered, i.e. the
slow motion derivatives approach is used.
The solution of a linear system of homogeneous equations undergoes a
fast decay or gives rise to a non decaying oscillatory or diverging behaviour.
Assumed a solution of the form

v(t)=v oet r(t)=roe t (*(t)=-oeo t p(t)=X(t) *=g+iko

for a non trivial solution to exist, the determinant of the system of algebraic
equations obtained substituting the above expressions in the system of
differential equations has to be zero. From this, the characteristic fourth degree
equation is derived.
The solutions are the eigenvalues of the system, and generally two of
them, wi and 4-2 are real and two li 3 +io>) 3 are complex conjugate, corresponding
to the fact that of the three motions only roll can have an oscillatory behaviour.
Dynamic instability prevails when at least one of the real eigenvalues or the
real part of the complex is positive, the last case corresponding to oscillatory
dynamic instability.The second step is to analyse the eigenvectors
trepresenting the amplitude of the motion solutions. Their moduli will be
normalized to the yaw motion (put equal to unity) to have the possibility of
individuating immediately which is the predominant. To each eigenvalue L., we
have thus a set of three eigenvectors that represent respectively sway, roll and
yaw amplitudes. If the eigenvalue is positive in that rangeof GM/LFn2, the
predominant eigenvector indicates the motion that can give rise to dynamic
instability.
 In computing the stability of antisymmetric motions of the BSRA series of
fishing vessels [19], the hydrodynamic derivatives for yaw and sway terms
depending on velocity were estimated from manoeuvrability [20,21]. Only very
rough estimates were possible for the terms depending on accelerations and
those regarding roll motion. This makes absolutely necessary to consider the
possibility of performing a thorough experimentation in this field.
In Fig. 5, the eigenvalues of the model XF for different values of the trim
cofficient y1=T/Tn are reported. As one can see, the eigenvalues are very
sensitive to trim variations. In particular, both g, and the real part g3 of X3 are
positive in a wide range for y<O, that is in the trim by bow conditions. No
instability is detected in the normal trim by stem conditions A look at the
eigenvectors reported in Fig.6 relatively to y1=-0.2, indicates that the system
instability is yaw dominated in both senses (diverging and oscillatory) since
the values of the a's and of the 8's (representing respectively normalized sway
and roll eigenvectors) are less than unity (representing normalized yaw) for
the first two eigenvalues, whereas the third indicates the possibility of a strong
roll-dominated instability, particularly at low values of GM/LFn 2 (low static
stability or high speed).
In Fig.7 the same results of Fig.5 are reported for the model 907. Here
the range of dynamic instability is wider then for XF. This could appear strange
if one considers that the model 907 is more stable than XF both following IMO
criteria and Strathclyde method. The question is that stability and safety is a
very complex problem and different approaches measure the resistance to
different dangerous phenomena. Here, not the pure capsizing possibility is
considered (of course only a fully nonlinear model could account for that
phenomenon), but also loss of control and broaching to, although in the
opinion of some authors the phenomena of directional instability and
broaching-to are not so clearly connected [22]. Moreover, sometimes the
instability appears connected with an inherently poor course-keeping ability
[23]. The results are not given in terms of KGmax/D due to the uncertainty in
the evaluation of the hydrodynamc coefficients.

2. CRITICAL ANALYSIS OF CURRENT APPROACH TO THE

HYDRODYNAMIC PART OF SHIP SAFETY

2.1. Introduction
Existing rules for the assessment of safety of ships are contained in the
"stability criteria" that in the following will be considered as known. A general
critical analysis of these criteria, of the misleading use of the term "stability"
and of the sometimes ambiguous role played by different actors interested in
ship safety, is contained in Ref. [24]. Here the analysis is focussed on
particular aspects of this approach that have negative influence on the
development of new concepts and methods for the improvement of ship safety
levels. From here on, except where differently stated, we will refer simply to
ship safety instead of hydrodynamic part of ship safety.
It is customary to separate the case of intact ship from that of damaged
ship. Both are based on a static approach. The most advanced formulation
regarding the damaged ship stability begin to include the description of the
transient phase of flooding, where transient is still intended in a quasi-static
approach.

2.2. The case of intact shin

The first serious approach, in the following indicated as IMO Statistical,
originated from the work of Rahola published in 1939. At that time, the
knowledge of the mechanims leading to capsize and broaching-to was poor
and the availability of efficient tools for direct hydrodynamic calculations
extremely limited. Hence a statistical approach that can be regarded as
scientifically unsatisfactory but practically relevant.
The result is in the form of correlations expressing limiting values
(prescriptions) for some characteristics of the righting arm of the ship with
respect to fixed-trim, calm-sea, isocarenic, transversal inclinations.
Due to its relative simplicity and to its apparent relation to the physical
mechanism of capsizing, the correlation was taken as a basis for a stability
criterion that, with small adjustments (requiring years of discussions!) is still in
use.
In the meantime, the need to reduce the number of casualties stimulated
the research and the analysis of casualties. The result was the proposal of the
adoption of a more "physical" criterion, i.e. the IMO Weather criterion. The
more or less evident opposition of many bodies delayed-the application of this
rule and originated a complicate classification of ships in relation to the type,
size and range of operation. In many cases, especially for the smaller ships
(the most sensitive to the meteomarine action!) only very elementary
calculations or no calculations at all is requested. The reasons of. the delays
and of the differentiation of the requests depending on ship size appa-rently
based on the difficulty and length of involved computa-tions, really masks the
attempt to avoid excessive reduction of paylod or operational capability.
This point worths a particular attention since it is usually assumed an
obvious the fact that increasing safety means decreasing payload. As we will
show, this connection is, at least partially an intrinsic consequence of wrong
initial assumptions about safety.
The statistical criterion, in fact, is an "a posteriori" one. In addition, it
expresses a correlation between observed ship losses, in a particular class of
ships sailing a very restricted region, and computed static stability characte-
ristics. The main criticisms can be sinthetically expressed as:
a) there is no direct connection between the requests and the needs. The
criterion does not involve seakeeping qualities, the forcing effect of
meteomarine environment, trim variations,coupling among different motions,
course keeping ability, shifting of cargo. etc;

b) the statistical nature of the correlation averages between good and

bad projects (wilh respect to safety), as a result, the good projects are
penalized:

c) as long as the ship remains intact, the probability of casualty at sea is

usually quite low as regards the mentioned mechanisms;
 d) the request to be "simple".
Since ships are objects with a quite high index of reliability, this means
that a few bad projects can be heavily conditioning. This means that even a
small subset of casualties only explainable as a wrong correlation attempt as
per a), can render the system almost "incompressible" in the sense that a very
strong additional request (i.e. a very low KG reducing drastically the payload)
contributes to a negligibly small increase in the observed safety at sea.
This could be advocated as a partial justification of the negative
consideration paid by the owners to improving stability criteria.
To exemplify, the results of a parametric analysis conducted on a family
of fishing vessels belonging to the BSRA series is discussed in some detail.
Firstly, the maximum allowable KG/D to satisfy both IMO criteria has been
computed [11]. The ships of the chosen subset are all obtained from the first
one through affine transformation. They have the same value of LIV1/ 3 while
B/T increases from 1.97 to the extreme value 3.44 for the model 907. Reported
as a function of BIT, the results exhibit a regular trend for both IMO criteria. The
value of KGmax/D increases with B/IT. In Fig. 4 the same values of the heigth of
the center of gravity are reported as a function of the Kempf's nondimensional
rolling period TK=TR4g/7. This is an interesting parameter as regards ship
operational capability, since it is not recommended to have TK<8 because the
roll behaviour is too stiff, so that the oscillations have moderate amplitude but
high frequency corresponding to excessive deck accelerations. At the opposite
extreme, a value TK>14 corresponds to large amplitude rolling (and loss of
stability in the traditional sense). Good operational capability corresponds to
the condition 8<TK<14. A look at Fig. 4 indicates clearly that the two IMO
criteria give comparable results with an average value of TK=8. A more
stringent request of stability pushes the ship toward a too stiff behaviour. For
example, the use of the Strathclyde method corresponding to the energy
balance taking into account the effect on transversal stability of a train of
waves [12], leads to TK=6 if 100% net area positive is requested.
As regards point c), it is important to note that often the loss of a ship is
the result of a concatenation of causes, among which an important role is
certainly played by the following scenarios:
- loop: large amplitude motions/accelerations • shifting of cargo
stability degradation •;

- large amplitude motions/accelerations * shifting of cargo = loss of

structural integrity of the hull =* damage;

- large amplitude motions/heavy weather = water on deck = stability

degradation.
A ship with movable cargo (solid or liquid) or water on deck cannot be
considered intact in the classical sense unless the proper dynamics of cargo
and of the mutual ship/cargo interaction is taken into account. In the particular
case of liquids with free surface, either in a container, as water on deck or
 water in a compartment as a result of flooding after damage, the consideration
of the classical correction to the different components of GM is contradictory
because it introduces a mixed approach statical/pseudo-dynamical. The
system, even if attention is focussed on rolling motion, is in this case a two
degrees of freedom one, Its behaviour is also qualitatively different from that
foreseable with the classical approach. As an example, the results of an
expenimental study of the behaviour of a fishing vessel with water on deck are
reported in Ref. [25]. According to the classical theory the amount and location
of water would contribute to a reduction of the metacentric heigth of about
50%. This does not account for the sharp change of natural rolling period that
doubles. Moreover, the effect of water on deck is in this case stabilizing. But
one does not neglect the fact that the effect of moving water is to some extent
similar to that of passive antirolling tanks that change the location of the
resonance peak giving generally origin to two peaks instead of one. It is not
easy to conclude that the observed is a general trend without a careful
examination of the possible variations of angle and frequency of encounter.
The effect of sloshing of water in a compartment of a fishing vessel is actually
( under study [26].
As regards point d), it has to be observed that the search to be simple
implies a generally poor physical description of the mechanisms of ship loss.-
This is evident for the statistical approach, but a closer examination reveals
that is a problem shared to a large extent by all approaches. As a result; an
unknown percentage of the correlation between casualties and prescriptions
remains unexplained, with the negative feedbacks alreadt mentioned.

2.3. About "optimization"

The simultaneous analysis of the seakeeping and stability performances
[4] on the same specimen of ships indicates a quite good degree of correlation
between the two characteristics, as shown in Fig:* 8. Another subset of ships
belonging to the same family, composed of models XE, WS, WR. WO, ZP, ZQ,
851, 852, 975, 977 and 978, once more obtained through affine transformation
or conformal transformation of the sectional area curve has then been
considered. In spite of being ships of the same type and comparable size and
forms, the results are now completely different. A look at Fig. 9 indicates that
the degree of correlation is very poor, if any.
It is not difficult to explain these drawbacks. The correlation is good when,
for similar forms, the relevant parameter to be varied is BIT. This, in fact, is
positively correlated with both the investigated characteristics: linear
seakeeping (vertical motions) and static stability. Increasing B/T on the other
hand, decreases the course keeping ability that has been called as a cause of
broaching-to. In these conditions it is a very abstract thing to try to find an
optimum project out of graphs like those of Figs. 8-9. The difficulties connected
with rank optimization when trying to mix vertical and transversal ship motions
arnd other characteristics, and thus the drawbacks to include stability in
optimization. is also evident from analysis like that reported Fig. 9a, taken
from Ref. [47].

3. STOCHASTIC ANALYSIS OF NONLINEAR ROLLING IN A NARROW BAND

SE A
St1 Introduction

All approaches to ship capsizing suffer in different measure of intrinsic

limitations. Experimental studies can be quite realistic in the incorporation of
the physical modelling and of the high nonlinearities, but are generally limited
as regards the possibility of variation of significant parameters and
consequently in the parametric inference. On the other hand; theoretical
approaches allow for good analyses of parametric dependences but generally
fall in a sort of "principle of indetermination", being less accurate in the
forecasting of large amplitude motions as they apparently improve
incorporating more details of the physical modelling. As a consequence, if one
wants to describe large amplitude motions (stability), quite rough mathematical
models have to be used, generally one or "one and a half" degrees of freedom
(i.e. one fully nonlinear and some other quasi-linear). On the contrary, low
amplitude motions (seakeeping) can be approached by good analytico-
numerical schemes incorporating, linear or quasi-linear, hydrodynamic
characteristics of increasing complexity. These difficulties are slowly
disappearing due to the advent of powerful techniques for the numerical
handling of nonlinear and 3-dimensional fluid dynamics.
The practical need of relatively simple guidance rules for the design. and
operation of ships and ocean vehicles,, generally known under the quite
improper name of "stability criteria", formerly approached in static or quasi-
static way, is now studied by means of nonlinear time domain simulations.
These constitute a sort of "numerical experiment", whereas true experiments
are now mainly used for validation porposes, except the cases where the
physical modelling is so complicate that cannot be actually tackled efficiently
by other means, as for exemple the study of the capsizing of vessels in very
rough weather, including breaking waves, performed in the Norwegian Project
"Stability and Safety for Vessels in Rough Weather".
It is never easy to design and interpret correctly an experiment, also in
the case it is a numerical one. In particular, the recently discovered possibility
of bifurcation scenarios, renders very complicate and possibly meaningless
the results of time domain simulations. This paper is devoted to a statistical*
analysis of two of the bifurcation schemes of the nonlinear ship rolling in the
stochastic domain.
Having mainly in mind the Stability problems, it is clear that a major role
is played by the large amplitude rolling motion. Unfortunately, the hydro-
mechanical modelling of ship loss is still incomplete, so that the connection
between large amplitude rolling and capsizing is not very clear. In principle,
the large amplitude rolling motion ofl a ship can, in fact, be a stable motion,
provided it is included in some stability boundary. In practice, the ship is a very
complicate system, so that many dramatic scenarios can appear once large
amplitude rolling is in some way originated. Actually, it is not very simple to
quantity the relative importance of these consequen'cies of large amplitude
rolling. On the other hand, neither the probability of large amplitude rolling has
been stated in a satisfactory way.
In previous papers [27-291, it has been shown that large amplitude rolling
in non extreme seas can be a consequence of a jump between the
antiresonant state and the resonant one, or between non resonant and
subharmonic states, due to the strong deviations from linearity. This originates
thle possibily of bifurcations that can be effective as a consequence of some
change in the parameters in the case of a deterministic excitation, arid due to
the very nature of the oscillation when a narrow band stochastic excitation is
considered.
Since the phenomenon is quite sudden, and the amplitude differences
can be dramatic also in presence of non very intense excitation, a parametric
research to find the probability of its occurrence was initiated. The goal is to
increase the knowledge of the probability density function of the response in
the following three cases of interest:
-bifurcation between the anti- and the resonant oscillation in the main
resonance region in stochastic beam sea;

-bifurcation between non resonant and subharmonic in the first

subharmonic region in stochastic beam sea;

- bifurcation between zero amplitude (or negligible one) and parametric

suharmonic roiling in stochastic following or quartering sea.
As a consequence of the possibility of bifurcations, the probability density
functions can be bimodal functions for some values of the parameters. This
allows an evaluation of the probabilities of the considered different oscillation
states, and in particular of the probability of the onset of large amplitude
rolling.
In this paper, the attention will be devoted to the first two.

3.2 The sea description

In structural analysis a failure is usually regarded either as a direct
consequence of an overloading or as a consequence of fatigue: This latter is
considered as the effect of a periodic or in any case time-varying load of
intensity lower that that implied in the first.
Capsizing and sinking are also failures, often improperly regarded as a
pure hydrodynamic problem related to ship motions and stability, and thus of
exclusive pertinence of Naval Architecture. In some limiting cases, and almost
always in safety criteria, attention is mainly devoted to hydrostatic aspects
only.
Actually, in the opinion of the author, the linking between the two aspects
should be considered with greater attention, especially in the light of the fact
that many cases of ship loss can be reconducted to structural failure induced
by motions. This can be due to slamming, sloshing, cargo shifting. In particular
this latter can be caused by overloading or fatigue in the lashings.
A part the slamming, that is connected with vertical motions, the other
phenomena are related to rolling. This is the reason why large amplitude
rolling motion is of interest not only when it determines by itself a stability
overloading. It could be in principle a stable, although very unpleasant,
oscillation regime, mainly affecting the ship's operational capability. Actually,
the ship is not a rigid body, so that sloshing, shifting of cargo, a gust of wind
[30] or water on deck [1J.1 could be sufficient to worsen the stability conditions
sufficiently to reach some non-return point.
Different scenarios can lead to large amplitude rolling. Among them,
worth noting are the effect of extreme beam sea or the effect of a moderate
 sea, beam or quartering, of appropriate frequency through one of the
mechanisms described in the Introduction. Whereas the action of an extreme
sea, with breaking waves, can actually be approached only by means of
experimental approaches, in the following the attention will be devoted to the
second case that we will refer as the 'resonance and bifurcation scenario'.
Resonance and bifurcations are a consequence of a certain degree of
autocorrelation in the time record representing the stochastic excitation. This
means, strictly speaking, that consistent parts of the record, or subrecords,
have to exhibit pseudosinusoidal characteristics. Of course, we will not
consider the extremely unrealistic case of pure sinusoidal or monochromatic
sea.
High autocorrelation means small spectral bandwidth, so that attention
will be focussed on the narrow band sea spectra, having in mind the fact that it
can be sufficient that a few consecutive cycles exhibit similar characteristics to
have resonance and possibly bifurcations. In a previous paper [32] it was
shown that such spectra can be well represented by white noise filtered
through cascades of linear filters (one or two can be appropriate). Limiting to
1' the simpler case, one has the following representation of the spectrum:
Sf = MS["2022C21

obtained shaping Gaussian white noise of level So.(Fig. 10) The filter
"damping" y represents the bandwidth and allows the excitation to recover the
sinusoidal case as y--ý0; of is the centerpeak frequency. The excitation so
represented has zero mean and variance crf 2 =ltSo .
In the analytical part of this study, the narrow band sea will be
represented in the Stratonovich's form, by means of a couple of slowly varying
amplitude sinusoidal terms. The amplitudes will be represented by two
independent Gaussian processes. Numerical computations will refer to filtered
white noise. In the very narrow band case, a square box filter was used, as it
simplifies the choice of the cutting frequencies and reduces the number of
harmonics required.
A direct analysis of the local autocorrelation was conducted on long time
records used as excitation in the time domain simulations. The results
indicated that for -'=0.005 the record is totally pseudosinusoidal within 10%
accuracy, whereas for the peaked JONSWAP, corresponding to 'y=0.14 the
pseudosinusoidal aspect seems practically limited to short subrecords.
As regards the realizability of such narrow band spectra, we recall the
long discussion leading to the formulation of the JONSWAP standard
spectrum in the search of a more pronounced sharpness than in Pierson-
Moskowitz, in agreement with many observations at sea. Now there is still
some hint in this sense. Moreover, we have to remember the intrinsic accuracy
limitation that is introduced in the formulation passing from observations to
spectrum [33]. Finally, we have not to forget the fact that, when considering
seakeeping or seakindliness, the analysis in terms of extreme and significant
values is in general sufficient. The first relate to long time scales, whereas the
second allow short time forecasts. When dealing with stability and safety from
capsizing, on the other hand, we have to consider all the possible dangerous
phenomena. These are not only connected with extreme values of the
excitation, but also with pseudosinusoidality. This latter is a local (in time)
 characteristic whose distribution in practical time records is to great extent
unknown apart the very interesting approach to the description of the statistics
of successive wave periods through a two-dimensional Weibull distribution
[34].

3.3. Analyical results

Large amplitude, non extreme, rolling in beam sea can be computed
through the use of analytical/numerical approaches based on relatively
simplified models. In particular, the rolling motion can be considered
uncoupled and described by a second order nonlinear equation of the type:

I 2 + D(qp, E) + A GZ(qp) = F(t)

(" that, dividing by I, representing the nonlinear quantities D and G through

polynomials up to the third order and introducing as usual the natural
f A[6'-GMM
frequency coo= becomes [27]:

d 2x (dx)
2-+ (2g+82 1X dx
L+ \dt 3 + o32 x
(j + c 3 x3 = f(t)

The roll equation was written in the absolute rolling angle, since this is
the only one meaningful in the stochastic case. It allows quite interesting
computations in the frequency domain by means of the use of perturbation
methods.
In particular, the use of the method of multiple scales allowed to obtain
approximate expressions valid for the variances of the stationary stochastic
state. Only the final results will be reported here. For more details on the
analytical procedure, see Ref. [35-37].
In the region of synchronism o _ wo:

- the roll variance a as a function of the tuning ratio o / o;o

1 [(3(0c 2 -32 l)2+
2 2
(•2 _ 1 +y/29eq ( -°
3 2 )2 + (2co°geq) 2 (1 + y/2 geq) 1a

with an equivalent linear damping P-eq given by:

1.
-eq = ± T 2
eqeq 6 q = (1 +30Oo262).

At the order of approximation-implied in the used method 0ot2-OC)o2

2rn)o(w:-woo) and the above equation for Q2 can be transformed in the following
explicit expression for the frequency shift of the frequency response curve:
 -- 3 -+ 2 2 2 2
(00 + 8(002
-
5O21 --('• (1 +y/ leq) - (2wOo1.eq) (1+Y/2iieq) )11

-the maximum value am of the roll variance as a function of excitation variance

of can be obtained imposing the reality condition on the square root of the
above expression:

4eqam
5+ 2 )2(1 +Y/2 Peq)] = 0?

A similar analysis conducted in the region of the first subharmonic (Of = 3(0o
gives the expressions reported in Ref. [41].

-the quantity QCrepresenting the roll variance in linear approximation and is

valid for the frequency region far from resonances.

CQ2 =(1 +I/2P-) aj 2 / [(Coo2_W?)2 + (2p9o") 2 (l +y/2.)2] = 0f 2/(0)O 2 ."f2 )2

- the excitation amplitude threshold ofth for the onset of subharmonic

oscillations as a function of tuning ratio is also given in Ref. [41].

Having in mind the analysis of rolling at intermediate to large amplitude

(Xmax = 0.8 radians), the righting arm model was truncated to the cubic term. To
get characteristic values of the nonlinear term coefficient (the natural
frequency was set to the value mo = 1, corresponding to an oscillation period of
2c), a parametric research was conducted on a specimen of 13 ships in
different loading conditions [29]. The results of a best fit of the cubic polynomial
up to different heeling angles indicated that in the range of interest the values
a3 = 4.0 and a3 = -0.5 could be representative of different situations. The first is
quite common in modern containerships, whereas the second is-typical of
most ships in loaded condition.
As regards the damping, intermediate values were chosen, except for the
study of the subharmonic, where a low damping condition was, for the
moment, assumed. A question a part could be that connected with the
assumed model of damping moment. As a result of the perturbative analysis,
an equivalent linear damping 1-eq is introduced. It depends on the actual roll
variance a 2 , so that it is not convenient to use directly a linear model. More
subtle is the distinction between the different nonlinear contributions that here
too appear strictly mixed as in [38]. Both contributions were retained in the
model used in the theoretical analysis in the light of the fact that there is still
some experimental evidence of an angle dependence of damping; moreover,
going to higher orders, the analysis can indicate a separate contribution.
The excitation intensity represented by of = 0.1414, corresponds to a sea
of moderate intensity, well below the intensity corresponding to the breaking
limit for the equivalent sinusoidal wave.
fitt
The deterministic expressions found in [27] are recovered in the limit 'y--0
as can be checked by inspection. Due to the particuliarities of the employed
method, using a representation based on two slowly varying amplitude
sinusoidal functions and the multiple scales perturbation, it is not very easy to
compare with other results valid for broader band excitation, such as the
stochastic averaging.
The analytical results indicated that the bifurcation scheme is preserved
also in the narrow band stochastic domain, as shown in Figs. 11-2.
An interesting question is constituted by the meaning of the different
rolling variances, corresponding to the same intensity and tuning ratio, of Figs.
11 and Fig. 12, of the curve of maximum variances in synchronism
In the deterministic scheme, the solution was lead back to an initial
values problem, so that the three solutions existing in some frequency range of
the synchronism region can be divided in stable and unstable in the
asymptotic sense. The Van der Pol plane representing the initial conditions
can be divided in two regions containing the corresponding stable solution,
called domains of attraction [28], separated by a separatrix curve passing
through the unstable solution. In the absence of external perturbations
changing some parameter, once started in one of the two domains, the
solution will remain always in the same domain, leading to the stable solution
as steady state. The curve of maximum amplitudes gives the highest of the two
stable oscillation states. Similar considerations can be drawn for the
oscillation in the region of the first subharmonic.
Giving up the deterministic scheme, the strict dependence on initial
conditions has to be renounced to some extent. All the previous conclusions
have to be reviewed in probabilistic terms, looking to the stochastic as a sort of
continuously changing initial condition transient. On the other hand, one has
also the distinction between transient and stationary stochastic. To clarify this
complex matter, an extensive time domain numerical simulation is in progress.
Only the preliminary results will be reported in the following.

3.4 Numerical results

The procedure consisted in the construction of a long record (tmax = 600

periods) of rolling motion through numerical integration of the differential
equation of rolling motion. Then, this record was analysed in terms of filtered
statistics of maxima (analysis of the envelope) to get the pseudosinusoidal
probability density function (p.d.f.) [39-41).
The results for the synchronism region shown in Fig. 13 exhibit a good
agreement with the frequency domain analytical values. In particular we can
observe that the distribution is bimodal. This indicates that the two "stable"
solutions are no more separated as in the deterministic case and the system
jumps up and down between the two extreme states. On the other hand, the
deterministic behaviour with domains of attraction has to be recovered in the
sinusoidal limit ( "..--40 ). The effect of realistic different initial conditions on the
p.d.f. can be seen in Fig. 14. It seems tnal the bimodality is preserved, but the
relative importance of the two peaks changes. In the opinion of the author, this
indicates that the sinusoidal limit is recovered through a displacement of the
p.d. between the two peaks and a simultaneous sharpening of the same In the
 limit 'y-40 this gives a p.d.f. consisting in one delta function located in
correspondence to one of the two stable solutions.
The results in the region of the first subharmonic are shown in Fig. 15. The
pdf is also in this case bimodal. The onset of the subharmonic is evidenced not
only by the presence of the second bump in the pdf. The analysis of the time
record in terms of the period of rolling (not reported in the figures) indicated wide
intervals where the oscillation was pseudo-synchronous with the natural
oscillations of the ship ( period = 2/0)0o ) that is characteristic of resonance,
instead than the forcing period (= 2p/"= 2p/3%o ).
The presented numerical results for the subharmonic region, for the moment,
refer to a case with very narrow band and quite low roll damping ability. It seems
that the onset of a subharmonic oscillation is restricted to such limiting cases.
Anyway, it is difficult to reach a definite conclusion, so that the analysis is in
progress.

3.5. The probability density function

The probability density function for the response envelope is given in [39-40]
for a Duffing oscillator (linear damping only) in the form:

p(a) = Const a exp{ - (1 +yl2geq)if2

J2 (2-°22 r3 )2+ (-y + 2 P9eq) 2 ] a do)

For particular values of the parameters, this function can exhibit multiple
extrema at the roots of dp/da=0. In the limits of the different approximations
implyed by the schemes adopted ("f =_ o), these points coincide with the roots of
previous equations. In particular, when this equation has three real solutions, the
pdf has two maxima and a minimum, thus exhibiting a bimodal character. The
probability distribution as shown in Fig. 16 shows a fundamentally different
behaviour with respect to the Rayleigh distribution applicable to the linear and
quasi-linear cases (Fig. 17). The multivaluedness disappears as one moves away
from the multivaluedness frequency window on both sides. It also disappears
when the excitation spectrum level or the nonlinear diameter are reduced. The
number of real solutions is reduced to one only, this case being usually defined
as a mild non-Gaussian behaviour [42-451.
The results can be of particular interest in the evaluation of fatigue behaviour
of hull/hull integrity and cargo securing/shifting as indicated in Section 2.2 [46], an
item of great interest in the consideration of ship safety, as indicated by Table 1.
REFERENCES
[1] Takaishi Y., "Consideration on the Dangerous Situations Leading to
Capsize of Ships in Waves", Proceedings 2nd International Conference on
Stability of Ships and Ocean Vehicles STAB'82, Tokyo, 1982, pp. 243-253.
[2] Morrall, A., "Capsizing of Small Trawlers", Trans. RINA, Vol. 122, 1980, 71-
101.
[3] Duong, C., "Fishing Vessel Stability, a State of the Art Review", Bureau
Veritas, Techn. Paper 8812, 1988.
[4] Proceedings International Conference on the SAFESHIP Project, Ship
Stability and Safety, RINA, 1986.
[5] Nedrelid, T., Jullumstro, E., "The Norwegian Research Project Stability and
Safety for Vessels in Rough Weather", Proc. 3rd International Conference
STAB'86, Gdansk, 1986, Vol.1, 145-155.
[6] Patullo, R.N.M., "The BSRA Trawler Series (Part I)", TRINA, Vol.107, 215,
1965.
[7] Patullo, R.N.M., "The BSRA Trawler Series (Part II)", TRINA, Vol.110, 151,
1968.
[8] International Conference on Safety of Fishing Vessels 1977, I.M.O., London
1983.
[9] Intact Stability Criteria for Passenger and Cargo Ships, I.M.O., London
1987.
[10] Boccadamo, G., Cassella, P., Mauro, S., Scamardella, A., "A New
Methodology in Order to Verify the Ship's Stability in the Preliminary Design
Stage", Proceedings International Symposium Prads'92, Newcastle upon
Tyne, 1992, Vol. 2, pp. 1173-1186.
[11] Francescutto, A., Nabergoj, R., "Analyse Parametrique de la Stabilit6
d'une Famille de Bateaux de Pdche", Bulletin de I'Association Technique
Maritime et A6ronautique, Vol. 90, 1990, pp.63-81.
[12] Vassalos, D., "A Critical Look into the Development of Ship Stability
Criteria Based on Work/Energy Balance", TRINA, Vol.128, 217, 1986.
[13] Norrby, R., "Stability Problems of Coastal Vessels", Int. Shipb. Progress,
Vol. 11, 1964, pp. 121-132.
[14] Cardo, A., Francescutto, A., Nabergoj, R., "Subharmonic Oscillations in
Nonlinear Rolling", Ocean Engng., Vol.11, 1984, pp.663-669.
[15] Allievi, A. G., galisal, S. M., Rohling, G. F., "Motions and Stability of a
Fishing Vessel in Transverse and Longitudinal Seaways", SNAME Spring
Meeting/STAR Symposium, New York, 1988, pp. 13-31.
[16] Dunwoody, A. B., "Roll of a Ship in Astern Seas - Response to GM
Fluctuation", J. Ship Res., Vol. 33, 1989, 284-290.
[17] Bishop, R.E.D., Neves, M. de A.S., Price. W.G., "On the Dynamics of Ship
Stability", Trans. RINA, 124, 1982, 285-302.
[18] Bishop, R.E.D., Price, W.G., Temarel, P., "On the Dangers of Trim by the
Bow", TRINA, Vol. 131, 1989, pp. 281-303.
[19] Francescutto, A., Reggente, S., Armenio. V., "On the Possibility of Loss of
Control and Broaching of Fishing Vessels", Symposium Technics and
Technology in Fishing Vessels, Ancona, May 1989.
[20] Inoue, S., Hirano, M-, Kijima, K., "Hydrodynamic Derivatives on Ship
Maneouvrability", International Shipbuilding Progress, 28, 1981, 1-14.
 [21] Fujino, M., "Lectures on Ship Manoeuvrability - Prediction of Manoeuvring
Performance", University of Tokyo, Technical Report No. 6013, 1985.
[22] Motora, S., Fujino, M., Konoyagi, M., Ishida, S., Shimada, K., Maki, T.,
Consideration on the Mechanism of Occurrence of Broaching-to Phenomena", "A
Naval Architecture and Ocean Engineering (Japan), 20, 1982, 92-107.
[23] Bao-an Y., "Ein Beitrag zur Beurteilung der Stabilit.t schneller Schiffe
bei
gekoppelter Gier-, Quer- und Rollbewegung", Schiffstechnik, Vol. 31, 1984,
pp.
22-42.
[24] Francescutto, A., "Is it Really Impossible to Design Safe Ships", Spring
Meetings of The Royal Institution of Naval Architects, London, 27-29 April
1992, paper n. 3, To appear on Transactions of The Royal Institution of Naval
Architects.
[25] Cardo, A., Francescutto, A., Zotti, I., Mattioli, R., "Experimental Study of
the
Effect of Water on Deck on the Stability of a Fishing Vessel", Proceedings
International Symposium NAV'92, Genova, July 1992, Vol. 1, pp. 3.3.1-3.3.14.
[26] Cardo, A., Francescutto, A., Armenio, V., Contento, G., "Dynamic Effects
of
Liquids on Board on the Stability of a Fishing Vessel", In Preparation
for
OTRADNOYE'93, Kaliningrad, May 1993.
[27] Nabergoj, R., "Small Vessel Optimization for Increased Seakeeping and
Stability Performance", Proceedings 4th International Conference on Stability
of Ships and Ocean Vehicles - STAB'90, Napoli, 1990, Vol. 2, pp. 597-603.
[27] Cardo, A., Francescutto, A., Nabergoj, R., "Ultraharmonics and
Subharmonics in the Rolling Motion of a Ship: Steady-State Solution",
International Shipbuilding Progress, Vol. 28, 1981, pp. 234-251.
[28] Cardo, A., Francescutto, A., Nabergoj, R., "Deterministic Nonlinear Rolling:
A Critical Review", Bulletin de I'Association T6chnique Maritime
ed
A6ronautique, Vol. 85, 1985, pp. 119-141.
[29] Cardo, A., Francescutto, A., Nabergoj, R., "The Excitation Threshold
and
the Onset of Subharmonic Oscillations in Nonlinear Rolling", International
Shipbuilding Progress,Vol. 32, 1985, pp. 210-214.
[30] Dahle, E. Aa., Myrhaug, D., Dahl, S. J., "The effect of Wind on Small
Vessels", Proc. International Conference STAB'90, Napoli, 1990, Vol. 1,
pp.
191-199.
[31] Falzarano, J. M., Troesch, A. W., "Application of Modern Geometric
Methods for Dynamical Systems to Problem of Vessel Capsizing with Water-
on-deck", Proceedings 4th International Conference on Stability of Ships
and
Ocean Vehicles - STAB'90, Napoli, September 1990, pp. 565-572.
[32] Francescutto, A., Cardo, A., Contento, G., "On the Representation of Sea
Spectra Through Linear Filters" (in italian), Tecnica Italiana, Vol. 56, 1991,
pp.
1-10.
[33] Guedes Soares, C., Trov~o, M. F. S., "Influence of Wave Climate
Modelling on the Long-term Prediction of Wave Induced Responses of Ship
Structures", Proceedings of the IUTAM Symposium on Dynamics of Marine
Vehicles and Structures in Waves, London 1990
[34] Myrhaug, D., Rue, H., "Note on a Joint Distribution of Successive Wave
Periods", To appear.
[35] Rajan, S., Davies, H. G., "Multiple Time Scaling of the Response of
Duffing Oscillator to Narrow-Band Random Excitation", Journal of Sound anda
Vibration, Vol. 123, 1988, pp. 497-506.
 (36] Davies, H., G., Rajan, S., "Random Superharmonic and Subharmonic
Response: Multiple Time Scaling of a Duffing Oscillator". Journal of Sound
and Vibration, Vol. 126, 1988, pp. 195-208.
[37] Francescutto, A., Nabergoj, R., "A Stochastic Analysis of Nonlinear Rolling
in a Narrow Band Sea", Proceedings 18th ONR International Symposium on
Naval Hydrodynamics, Ann Arbor, August 1990.
[38] Cardo, A., Francescutto, A., Nabergoj, R., "On Damping Models in Free
and Forced Rolling Motion", Ocean Engineering, Vol. 9, 1982, pp. 171-179.
[39] Davies, H. G., Uu, Q., "The Response Envelope Probability Density
Function of a Duffing Oscillator with Narro Band Excitation*, To appear on
Journal of Sound and Vibration.
[40] Davies, H. G., Liu, Q., "On the Narrow Band Random Response pdf of a
Nonlinear Oscillatorn, To appear.
[41] Francescutto, A., "On the Probability of Large Amplitude Rolling and
Capsizing as a Consequence of Bifurcations", Proceedings 10th International
Conference on Offshore Mechanics and Arctic Engineering 'OMAE',
Stavanger, June 1991, Vol. 2, pp. 91-96.
(• [42] Ness, 0. B., McHenry, G., Mathisen, J., Winterstein, S. R.(1989).
"Nonlinear Analysis of Ship Rolling in Random Bean Waves", Proc. Research
Workshop on Stochastic Mechanics, Technical University of Denmark,.
Technical Report N. 244/R, pp. 49-66.
(43] Juncher Jensen, J. (1989). "On Fatigue Damage due to Non-Gaussian
Responses", Danish Center for Applied Mathematics and Mechanics, Report
N. 389.
[44] Juncher Jensen, J. (1991). "Fatigue Analysis of Ship Hulls under Non-
Gaussian Wave Loads", Marine Structures, Vol. 4, pp. 279-294.
[45] Francescutto, A., "Stochastic Modelling of Nonlinear Motions in the
Presence of Narrow Band Excitation", Proceedings International Symposium
on Offshore and Polar Engineering - ISOPE'92, San Francisco, June 1992;
Vol. 3, pp. 554-558.
[46] Hutchison, B. L., "Cargo Mechanics (Application of Seakeeping-
Revisited)", Marine Technoloqy. Vol. 23, 1986, pp. 230-241.
[47] Kishev, R., Dimitrova, S., Gaberova, M., ."A Generalized Procedure for
Rank Optimization Application to Ship Design", Proceedings International
Symposium PRADS'89, Varna 1989, pp. 33.1-33.9.
Table 1. Main causes of loss for capsize and broaching regarding
Japanese fishing vessels in the period 1973-77

Main Cause of Accident N

Overloading 9
Center of mass too high 13
Insufficient securing of cargo 1
Improper stowing of cargo 7
Cargo shifting 13
Openings not secured 9
Weack hatch covers 2
Hull damage 2
Water trapped on deck 29
Broaching 1
Total 86
 Table 2. Main characteristics of the BSRA trawler series.

MODEL L B T D L/V 1I/3 B/T CB CW XCB%

BSRA I XF 45.72 8.03 4.07 4.95 4.85 1.97 0.562 0.774 0.969
BSRA I XG 45.72 8.50 3.85 4.77 4.85 2.21 0.561 0.776 0.960
BSRA IWO 45.72 8.96 3.64 4.63 4.85 2.46 0.562 0.776 0.962
BSRA I WP 45.72 9.40 3.48 4.51 4.85 2.71 0.562 0.777 0.965
BSRA 1907 45.7210.62 3.09 4.21 4.85 3.44 0.560 0.777 0.949
BSRA I WS 41.01 8.98 4.07 5.04 4.35 2.22 0.559 0.772 0.678
BSRAIWR 43.36 8.73 3.95 4.90 4.60 2.22 0.560 0.765 1.095
BSRA IWQ 48.09 8.30 3.75 4.46 5.10 2.22 0.561 0.775 0.959
BSRA IIZP 45.72 8.03 4.08 4.97 4.92 1.97 0.521 0.742 0.888
BSRA II ZQ 45.72 8.03 4.06 4.95 4.76 1.97 0.593 0.806 0.906
BSRA 11851 45.72 8.32 4.23 5.15 4.85 1.96 0.521 0.743 0.892
BSRA 11852 45.72 7.82 3.96 4.81 4.85 1.97 0.593 0.805 0.903
BSRA 11975 45.72 8.50 3.69 4.77 4.85 2.30 0.584 0.816 4.689
BSRA 11977 45.72 8.50 3.72 4.76 4.85 2.29 0.580 0.795 0.693
BSRA II978 45.72 8.50 3.74 4.75 4.85 2.28 0.577 0.790 -0.875
minimum 41.01 7.82 3.09 4.21 4.35 1.96 0.521 0.742 -0.875
maximum 48.0910.62 4.23 5.04 5.10 3.44 0.593 0.816 4.689
C...

Table 3a. KGmax (m) according to IMO statistical criterion.

MODEL STANDARD
A B C D E F
BSRAIXF 3.488 3.315 3.233 3.172 3.216 3.166
BSRA I XG 3.588 3.423 3.320 3.232 3.209 3.238
BSRA I WO 3.769 3.588 3.455 3.327 3.227 3.357
BSRA I WP 3.984 3.775 3.614 3.449 3.269 3.505
BSRA 1907 4.777 4.413 4.123 4.107 3.440 3.959
BSRA I WS 3.882 3.607 3.509 3.432 3.387 3.156
BSRA I WR 3.696 3.527 3.414 3.315 3.286 3.326
BSRA I WQ 3.490 3.333 3.238 3.165 3.134 3.438
BSRA II ZP 3.456 3.187 3.132 3.113 3.245 3.070
BSRA IIZO 3.645 3.179 3.124 3.103 3.213 3.058
BSRA II851 3.597 3.306 3.256 3.243 3.363 3.188
BSRA 11852 3.359 3.082 3.034 3.023 3.126 2.970
BSRA 11975 3.577 3.261 3.192 3.153 3.229 3.116
BSRA II 977 3.537 3.224 3.158 3.123 3.199 3.084
BSRA 11978 3.522 3.209 3.145 3.113 3.184 3.070
 Table 3b. KGmax/D according to IMO statistical criterion.

MODEL STANDARD
_ A B C D E F
BSRA I XF 0.704 0.669 0.653 0.640 0.649 0.639
BSRA I XG 0.752 0.717 0.696 0.677 0.672 0.678
BSRA IWO 0.815 0.776 0.747 0.719 0.698 0.726
BSRA I WP 0.884 0.838 0.802 0.765 0.725 0.778
BSRA I 907 1.134 1.048 0.979 0.975 0.817 0.940
BSRA I WS 0.770 0.715 0.696 0.681 0.672 0.626
BSRA I WR 0.754 0.720 0.696 0.676 0.670 0.679
BSRA I WO 0.782 0.747 0.725 0.709 0.702 0.770
BSRA IIZP 0.696 0.641 0.630 0.626 0.653 0.617
BSRA II ZO 0.700 0.642 0.631 0.627 0.649 0.618
BSRA 11851 0.698 0.642 0.632 0.629 0.653 0.619
BSRA 11852 0.698 0.640 0.630 0.628 0.677 0.617
BSRA II975 0.749 0.683 0.669 0.661 0.677 0.653
BSRA II 977 0.743 0.677 0.663 0-656 0.672 0.648
BSRA II 978 0.742 0.676 0.662 0.655 0.670 0.646

)
 Table 4. Maximum allowable height of the centre of gravity above keel.

MODEL IMO STATISTICAL IMO WEATHER

KGmax KGmax/D KGmax KGmax/D
BSRA I XF 3.172 0.640 3.005 0.606
BSRA I XG 3.209 0.672 3.046 0.638
BSRA I WO 3.227 0.698 3.107 0.672
BSRA I WP 3.269 0.725 3.225 0.716
BSRA I 907 3.440 0.817 3.659 0.869
BSRA I WS 3.156 0.626 3.268 0.648
BSRA I WR 3.286 0.670 3.125 0.638
BSRA I WQ 3.134 0.702 2.968 0.665
( BSRA 11ZP 3.070 0.617 3.058 0.615
BSRA IIZQ 3.058 0.618 2.991 0.604
BSRA I1851 3.188 0.619 3.180 0.617
BSRA 11852 2.970 0.617 2.904 0.603
BSRA 11975 3.116 0.653 3.039 0.637
BSRA II977 3.084 0.648 3.013 0.633
BSRA II 978 3.070 0.646 3.001 0.632
FIGURE CAPTIONS

Fig.1 - The I.M.O. weather criterion according to Ref.7.

Fig. 2 - The method of energetic balance of Strathclyde.

Fig.3 - Comparison of KGmax/D versus main design parameters for different
stability criteria.
Fig.4 - Comparison of KGmax/D versus Kempf's adimensional rolling period
TR(g/B)12 according to I.M.O. stability criteria.
Fig.5 - Variation of the eigenvalues of the characteristic equation for
GM/LFn 2=0.1 as a function of trim parameter y for model XF. Positive values of
the eigenvalues indicate instability.
Fig.6 - Eigenvectors for the model XF in the trim by bow condition as a function
of GM/LFn 2 . Yaw is unitary, cc indicates sway and 8 roll.

Fig.7 - Variation of the eigenvalues of the characteristic equation for

GM/LFn 2=0.1 as a function of trim parameter y for model 1907. Positive values
of the eigenvalues indicate instability.
Fig. 8. Stability rank index relative to the application of energy balance method
versus rank index for seakeeping (vertical motions) for the same family of
fishing vessels represented in Fig.4.

Fig. 9. Stability rank index relative to the application of energy balance method
versus rank index for seakeeping (vertical motions) for another family of fishing
vessels of the BSRA series.

Fig. 10. Normalized JONSWAP spectrum corresponding to unitary value of the

peack enhancement factor (solid curve). The dashed curve represents the fit
obtained using the rational representation as a linear filter.

Fig. 11 Roll variance a as a function of tuning ratio 6Y/oo . The following values
have been used for the parameters: o 3 =4.0, it=0.05, uf=O.1414 (ew=0.2). The
number on the curves indicates the value of y.

Fig. 12 Roll variance a as a function of tuning ratio oy/woo .The following values
have been used for the parameters: a 3 =-0.5, 9i=0.05, 91=0.1414 (ew=0. 2 ). The
number on the curves indicates the value of y.

Fig. 13. Comparison of the analytical results in the synchronism region with the
time domain simulation. The following values have been used for the
parameters: aX3 =4.0, [L=0.05, 1=0C. 1 4 14 , -y=0.02. The value )I/)oo = 1.55 was
used in the numerical computations. Initial conditions corresponding to x0 =-0.3
and x0 =O.
Fig. 14. Dependence of the probability density function on the initial conditions
in the synchronism region. Solid curve refers to xo=O. and *o=0., whereas
dashed curve referes to xo=-0.3 and xo=O. The following values have been
used for the parameters: ct3 =4.0, g=0.05, o;=0.1414, y,=0.005, oy/w0 = 1.55.

Fig. 15. probability density function in the region of first subharmonic. The
following values have been used for the parameters: a 3 =4.0, g=0.005,
of=0.1414, y=0.005, wt/w o = 3.2. Initial conditions corresponding to x 0 =0.15
and ;0 =-O.15.

Fig. 16. Envelope probability density function in the presence of bifurcations as

obtained from the analytical approach. The following values were used for the
parameters: (oo=1., g=.05, oj=.1414, y=.02, "f/o)o=1.40, ct3=4.0.

Fig. 17. Envelope probability density function in absence of bifurcations as

obtained from the analytical approach. The following values were used for the
parameters: cn)0=1., g=.05, ar=.1414, y=.02, "/- 0=1.40, .3=0. (linear system).
AGz

.- 10'-
1> ~1.0

STATISTICAL
-- WEATHER
0.9- - STRATHCLYDE /0%

Q, 0.8-

0.7-

'100 %

0.6--

0.5
1.5 2.0- 2.5 3.0 3.5 4.0
BIT.

1.0

KG/B 0%, 907

0.9-

0.8Ivc
50%
1 XG
07100% X

0.6-
?~STATISTICAL
WEATHER61
STRADTICLVDB-

4 c- 8 t0 T, . 14

'F 4 4
 0-5(1k, XF
J-/L F,?= 0. 1

-1r 1.0

0.50) 3

5-
XF

-1.0 -0.5 0 0.5 y 1.O

 1.0-
0-51LI907

2 = 0.1
ZKM/LF
"0
10*5

101-.L3

-1. 0 -10 - .

12-

10-

6-

00

4- o

0-0

2 4 6 8 10 12
Seakeoping Index
 6

4-

(0 4

0 2S

0- RKN IOH

*- PARTIAL RANKING BREED ON SHIP DEHAVIDUR IN BERM05EHS

g-PARTIAL RANKING BRSEED ON SHIP GEBRUIOUR IN HERD SEAS
R 0- PARTIAL RANKING BASED ON STILL WATER RESISTANEE

1 2 3 4 5 5 1 a B Inl11 12
URRIANT No
2L
Partial versus general rank evaluations of design variants
 It

Ot

0-

LL'I

(* I

-S0..

0.5.

oJf(~ 2.0
 0.0 0.0/

0.5-

0o.

1-0 L5 ct&-2.0 pdf (a~u)

 pdf
(a.u.)

0.. 0.5 xm t

(a.u.)

0. 0.1 0.2 Xm 0.3

 1.2

1.0

0.8

S0.6

0.4

0.2

00.0

u.0 0.2 0.4 0.6 0.8 1.0

1.0

0.8

0-6

0-4

0.2

0.0
u.0 0.2 0.4 0.6 a 0. 8 1.0
 ~~N- -5 - A eL. -j.kQ1NcsscuAt

IA~ý
ftl WeajL4
Cs9I/T,
A L Vtý2q C A : -(E92 7-)
I.}~tn+e orF§

4-~ ZF a3a

04 02 0/30. 4 057-/

2' 63z
OPERATIONAL STABILITY-Future Rules and Regulations

Emil Aall Dahie, Professor, Dr.ing., DNV Industry, Oslo, Norway

AB3STRACT

In the paper, recommendation is made to base future ships stability rules and regulations on
the principles of risk analysis. As a first step, a level of acceptance should be discussed. If
the present accident statistic is unsatisfactory, it should be spelled out, otherwise effort could
be spent on other types of ship accidents.

For most large ships, intact stability is considered as acceptable regardless of weather, and
emphasis is on damaged stability and water ingress. Qualification, education and training of
ship officers as outlined in STCW 1978 is considered as adequate to maintain stability.

For some large ships design or high-speed operation may, especially in severe weather, cause
an increase in moments acting on the ship to an extent that calls for qualification, education
and training beyond what is common today in order to obtain an acceptable standard. Data
for calculation of risk is to some extent available, and model experiments can be made where
such data is lacking.

For smaller ships, "severe weather" is only "moderate weather" for a larger ship. This in
itself calls for risk control as risk may otherwise become excessive if the ship is operated
without regard to the weather. To some extent, existing regulations reflect this matter in
terms of restricted area of operation etc., but a systematic and logic approach seems to be
lacking.

Ideally, the concept of "capsize" in relation to intact stability should be eliminated by a

combination of technical improvements in design of smaller ships, and by improved
awareness by those operating the ship with regard to keeping the ship intact and securing
cargo. Such restrictions, and adequate operational advice specific to the ship should be
presented in an operational manual.

As about 45 000 relatively small vessels out of a total of approx. 80 000 are in existence,
the short-term solution is to consider more specific operational restrictions, based upon data
from experience, model tests and other research, and supported by better qualifications,
education and training of naval architects and ship officers.

The most important technical improvement for new ships is considered to be implementation
of measures to extend the intact stability to at least 80 degrees heeling angle as has been
implemented by the Norwegian authorities for certain groups of small ships.
 2
INTRODUCTION

The future approach to stability could be based upon functional requirements in the sence that
acceptable safety against capsize is maintained by a combination of technical capability and
operational measures taken when necessary.
As for other marine accidents, the approach could be through methods common in risk
analysis. This implies that risk management, which is the goal of the industry as well as
governments could follow basic principles as follows;

1. Stote the ideal, long-range goal of the effort.

The obvious goal is to eliminate stability-related accidents.
2. Formulate temporary risk acceptance criteria, to serve as milestones on the way
forward.
For the present, such criteria are expressed in stability regulations and in crew
qualification requirements.
3. Perform risk analysis of typical stabilfly-reloated accidents by cause-consequence
analysis. -Compare result with acceptance criteria.
4. Investigate ways and means to meet the acceptance criteria, if they are not met.
5. Having obtained risk acceptance, investigate further risk-reducing efforts that are
practicable and reasonable i. e. cost-effective).

Care must be taken to differentiate between consequences to the ship, and consequences to
the crew as efforts to reduce one may in some instances impair the other.

The natural approach with regard to ship operation related to stability should be to
concentrate on important, risk-reducing factors. No doubt, this has been and stil is the
intention of those involved. The main reason for this is that the role of the naval architect
at the present is limited to the task of designing a ship that complies with stability criteria,
without any regard to operation of the ship. Further, neither national authorities nor
classflcation-societies have adressed operation of ships in an explicit and clear manner.

In the offshore industry, risk analysis has been used for a number of years, following the
basic principles outlined above. Below, operational stability will be discussed within this
Context.

Stability-related accidents occur when the capability of the ship in terms of stability is
exceeded. Often, the otherwise sufficient stability is severely reduced by ingress of water,
or by shifting of cargo. In such cases, the demand on the reduced stability caused by external
forces from the environment, which the ship would otherwise sustain without problems, may
cause an accident. The matter of survivability of ships after ingress of water in terms of
floatability and stability is important. However, in the following, only operational aspects
related to intact stability is discussed, i.e. when the capability of the ship in terms of stability
is unimpaired.

IDEAL GOAL

The natural goal is to strive to eliminate all stability-related accidents - and still maintain an
economnic shipping industry.
RISK ACCEPTANCE CRITERIA3

General

Accidents are never to be accepted, but risk is, provided that the industry is to prevail. For
intact stability accidents, acceptance must be considered for at least four different aspects;
- loss of human life and health
- material damage to, or loss of the ship
- material damage to, or loss of the cargo
- damage to the environment

Risk can be conceived as the product of frequency of an accident, and its consequence.

When a vessel capsizes, the consequence to the ship and its cargo is normally total loss.

For intact stability, acceptance is presently considered as obtained when a ship complies with
requirements to an intact stability curve. Intact is defined in the Load Line Convention, and
recommendations for minimum a stability curve are given in IMOs guideline A. 167 for cargo
and passenger ships, and more stringent for fishing vessels, for which the Torremolinos
Convention text can be considered as a guideline. In addition, national authorities often have
formulated additional requirements for certain types of ships.

The acceptance criteria for intact stability is based only on capability of the ship, and it is
taken for granted that the demand (e.g. environmental forces and moments) will to a large
extent be less than the capability of the ship. This approach is not common in other
industries, where the demands in the area of operation are normally considered in relation
to the capability. When only considering capability, it is not possible to formulate numerical,
acceptance criteria based on frequency of exceedance of demand over capability.

From accident statistics, it can be seen that stability-related accidents are only significant for
certain types of ships, where the demand in terms of forces and moments may become large.
Examples are container ships and cargo ships with deck cargo of considerable height and
sailing ships, for which the windage area and the wind moment is relatively high, and where
there is little or no contribution to intact stability from the elements that increase the external
forces and moments. Another example is high-speed vessels, where various dynamic
moments related to the speed may act. Small vessels are exposed to breaking waves that may
cause accidents, whereas the same waves may have negligible influence on larger ships.
Some interesting data for ship accidents are presented by the COST 301-project, 1987, see
Table 1, where capsize is included in the foundering casualty rate. The importance of ship
size is evident. From the data, also frequency of occurence per year can be assessed.

Generally speaking, loss of cargo and pollution will always be a consequence of a capsize.

There is no distinction between passenger ships and other ships regarding intact stability.
However, for passenger ships, and for ships carrying hazardous cargo, requirements to
 4

Mean Casualty L•aes

Size Classes
in CRT Collision Collision Collision Mean FWaz Mean mean Kean Overall
(Meeting) (Croasing) (Overtaking) Collision Strandieg Raotng Foundering Caualty
Rate Rate Rtate Ra Le R~a e

100 to 499 0.179 0.111 0.036 0.326 0.7S2 0.065 0.277 1.420
500 to 1 599 0.251 0.165 0.062 0.478 0.473 0.089 0.135 1.175
1 000 to 9 999 0.278 0.160 0.063 0.501 0.383 0.048 0.066 0.998
tO 000 to 59 999 0.371 0.163 0.074 0.608 0.407 0.085 0.011 1.12L
60 000 + 0.426 0.248 0.071 '0.745 1.82 0.139 0 2.705
Overall 0.279 0.LSS 0.061 0.495 0.483 0.070 - 0.105 1.153

Note: Figures exprets 6

the mean number of caaualcia/1O ship-ailes.

Table 1. Mean ship casualty rate in European waters (COST 301, 1987).

damaged stability are stronger than for ordinary cargo ships, even though their survival
capability is also improving. Damage stability measures will limit the consequences so that
evacuation may not necessary. Should nevertheless evacuation be required, the passengers
will have more time to evacuate than seafarers on most ordinary ships.

Risk acceotance criteria for the ship

Risk acceptance is presently restricted to checking that the stability is above the minimum
required in national regulations based upon IMO recommendations.
In the future, risk acceptance could be related to frequency of demand (forces and moments
acting) exceeding capability (intact stability moment). For assessment, exceedanceý -of
capability can be found by systematic model tests, by calculations and from well documented
casualties.

Risk acceptance criteria for humans

Risk acceptance for fatalities and injuries, the frequency is for the present adressed by taking
for granted that saving the ship saves the crew. As for reduction of consequences, this is
implicitly covered in general requirements to lifesaving appliances. However, intact stability
accidents are not explicitly adressed as they should be in order to judge their adequacy in a
capsize accident. In the future, this should be the case. On this basis, specific training could
be structured.
 RISK ANALYSIS5

Accident to the ship

The first part of a risk analysis is concerned with the cause for the ship accident. For
stability-related accidents, the cause is that demand exceeds capability, and this naturally calls
for knowledge of both, and subsequent comparison. Ihe intact stability curve expressed by
GZ, multiplied with the displacement, can to some extent be regulated by the operator e.g.
cargo versus ballast (as long as the prescribed stability minimum is not violated). Within
periods the stability can be considered as deterministic. The cargo must be kept in place for
this to be true.

The demand (forces and moment) from the environment is of a stochastic nature. Engineering
methods for calculation of such forces are developed to some extent, and research is ongoing
in many countries. Perhaps more important is the existence of data from model tests and
from well recorded casualties. Model tests may, by gradually and systematically increasing
the demand (forces and moments), serve to locate the border area in which capability
(righting moment) is exceeded, and capsize occurs, Dalile and Myrhaug, 1988. More
specifically, the identified, hazardous elements in the stochastic demand can be reproduced
in model tests. Typically, such elements are;

- breaking waves from the side, Dahle and Myrhaug 1988, 1992.
- steady wind and sequence of high waves with a period corresponding to the
roll
period of the ship, Dalile and Myrhaug 1993.
- following seas, wave length about twice the length of ship, speed of ship nead
to
wave speed.
- high wind with wind gust hitting when ship is heeling away from the wind

The principle of calculating the probability of exceedance P(Demand > Capability) is

illustrated in-figure 1. The calculation is quasi-dynamic in nature. To obtain the frequency
per year, the number of times that exceedance take place must be found from available
weather statistics, Dahle and Myrhaug, 1992.

Due to the shape of ships (long and narrow) as compared to e.g. drilling platforns (square),
the demand can to a large extent be reduced and sometimes even be eliminated by operation
of the ship. Typical and practical measures are;

- choice of heading
- choice of weather in which to operate (weather forecast, distance to harbour)
- choice of ship speed in following seas

Systematic model tests have been carried out, and reported to IMO, at stability conferences
and in technical papers. Nevertheless, for new ships, no analysis of the relationship between
capability and demand is undertaken. Only calculation of capability i.e. of intact stability for
relevant loading conditions is undertaken and the result is then compared with minimum
requirements. For naval architects, this procedure is well covered in university curriculums.
 Figure 1. Principle of quasi-dynamic calculation of probability 6
of P(Demand> Capability)

f(X)

Demand Capability limit (Xj

P(Demand > Capability) =ff(x~dx

X.

Righting moment
x (Capability)

X, Ovenbming mmomat

For ship officers, calculation of intact stability in actual loading conditions (the capability of
the ship) is covered by the 1978 STCW Convention explained as follcws in IMOs document
"An international maritime training guide";

"The training of deck and engineer officers should be designed to give them a
knowledge of basic principles of ship construction and the theories and factors
affecting trim and stability to maintain a safe ship at all times....
The stowage and handling of cargo have an important bearing on safety and the
training of deck officers should ... include general precaution with, and methods of,
stowage and securing of cargos (including bulk cargo and deck cargo), and practical
knowledge of calculation and maintenance of stability during loading, discharging,
ballasting and bunkering and while making a sea passage, including effects of free
surfaces and icing on stability."

Operational decisions that have influence upon c.psizing forces and moments are left to the
discretion of the captain. Even though the influence is known in principle, virtually no
reflection of this can be found in the quotation from STCW above. Operational aspects are
not covered in qualifications, education and training, neither for naval architects nor for ship
officers. Others involved, such as governmental agencies and classification societies have so
far been extremely careful not to involve themselves in operational aspects. IMO, on the
other hand, seem to have intentions of approaching the problem in the present urge for
investigation into the role of the "human factor" in ship accidents.

An important matter seems to be the future qualifications, education and training of naval
architects. Ideally, a ship should not depend heavily on operation in order to obtain an
acceptable safety against capsize. Consequently, naval architects should attempt not to design
small ships with stability curves as illustrated in Figure 2b below.

In a practical world, such vessels will nevertheless be built, and a large number are in
service. Therefore, those operating ships with stability curves as shown in Figures 2b and
 7
2c should be qualified, and should receive education and training for safe operation.
From
the naval architetcs, they should receive a operational manual in which necessary
restrictions
related to acceptance criteria as outlined above should be clearly stated.
As an example of the points discussed above, frequency of capsize for two areas
(Norwegian
Sea and the Indian Ocean) has been presented by Dahle and Myrhaug, 1992,
whereas the
influence of various operational restrictions is discussed by the same authors,
1993 (to
appear).
As part of future design of ships, naval architects should identify and define safe
operation in order to keep the risk at an acceptable level. Operational instructions shouldship
presented in an operational manual, Dahle and Nedrelid, 1986. This requires general be
from stability research to naval architects, and specific input for each individual ship input
from
the naval achitect to those responsible for ship operation.

Accident to humans
Given that a ship accident has occured, there remains an analysis of consequences to humans,
i.e. fatalities and injuries. First, the conditions of the ship after the accident,
and the
prevailing environmental conditions must be noted.
The condition of the ship after the accident can be one of the three shown in Figure
2.
Figure 2. Possible stable position of a ship after demand (forces and moments) have
exceeded capability (stability)

STABLE .

a. Se-i--rght••g.
Capsize not teasible

b. Cwe, z. to stab&, sie position

S ?A&[ t' SIAUL IE

c. Capsize to 180"
 8

The first situation shown in Figure 2a is for a ship which is self-righting, i.e.
a ship which
has only one stable position. Such a ship has relatively long deck erections extending
to the
side as reflected in the GZ-curve shown. Consequences to humans are restricted
to injuries,
and evacuation is not needed.

Secondly, in Figure 2b, a ship a stability curve with a second stable side-position
between
o and 180 degrees is shown. This is a common situation for ships with relatively short deck
erections extending to the side, or long but narrow deck erections, and a relatively
low center
of gravity. Those onboard will have difficulties in getting out, and those who
get out will
often have to evacuate directly into the sea and must be saved within a short
time. After
some minutes, the ship will capsize to 180 degrees or sink sideways. Lifeboats are
normally
of little or no use in such accidents.

Thirdly, in Figure 2c, the situation for a ship with a second stable position at
180 degrees
is shown. Such ships have small deck erections. During a capsize, a minor part
of the crew
manage to get out, but mostly those onboard remain inside and perish. Lifesaving
appliances
are therefore of little use.

To -illustrate the different consequences (loss of ship, loss of personnel), data

from the
Norwegian fishing fleet are presented in Figure 3. While the relative consequence
in terms
of number of ships is only 10%, corresponding loss of life in capsize accidents
amount to
about 70%.

Figure 3. Number of ship accidents versuis loss of lives (1980-ties, Norwegian fishing
fleet)

Leakage 30) (4)

Fire/explosion 28
Grounding 15
Capsize 11 (20)
Contact 10 (3)
Eng. fail. 7
Collision1
Other 8 (6)

110 33
 9
Choice Of type and location of lifesaving appliances onboard (within regulation
framework)
is a matter for naval architects, but such matters are hardly covered in university
curriculum.
Suitability and location onboard of lifesaving appliances should relate to
the actual accidents
as illustrated in Figure 2 as regards capsize in intact condition. Training
for the crew should
be ensured by mandatory safety courses having this in mind. Also
information to, and
training of passengers can be improved.

Ways and means to meet the acceptance criteria

From the preceeding subchapter, it can be concluded that the risk

for material assets in a
capsize is proportional with the capsize frequency as the consequence
is nearly always the
same, ie. total loss of the ship and cargo, and pollution.

Acceptance criteria are not established, but from Norwegian statistics

for
frequency is about 1 capsize per 10 000 vessels a year , Dahle, Myrhaugsmaller vessels, the
and Dahl, 1988.
Sudden capsize to a stable side position, or to 180 degrees often causes
total loss of the crew.
Basically, smaller ships (below 40m in length) with such charateristics
should not be
acceptable. Since 1980, this has been the stand taken by the Norwegian
government for new
fishing vessels between 15 and 40m in length, and such vessels are
no longer built. This
decision was in line with the principle of reducing the frequency of
accidents primarily by
technical means.

For smaller vessels with stability curves as shown in Figures 2b

and 2c, operational
restrictions should be considered as a secondary means for risk reduction.
Of main concern
should be adjustment of heading and (or) speed in order to reduce the
demand (adverse forces
and moments).
To reduce consequences to humans, lifesaving appliances should
be considered for the
capsize situation. In Norway, experience has shown that individual survival
suits are needed,
both for personnel in lifeboats or rafts, as well as for personnel that have
to evacuate directly
into the sea. Most probably, survival suits should also be available
on passenger ships.
Capsize for larger ships is relevant only for certain types. E.g. container
vessels have been
in focus, and efforts have been made to develop ship forms that are
not prone to capsize
when running at high speed in following seas. It seems that more
relevant efforts could be
made to give clear operational instruction for speed reduction, and
for improved container
fastening. Encompassing model tests have provided a good background
for elaboration of
such instructions. Operation of such vessels with respect to intact stability
could be based on
education and training on simulator and by relevant theory.
 10
Further risk-reducin2 efforts

In the terms of risk analysis, the acceptance criteria provide a tolerable limit. However, when
further improvement is reasonable in terms of e.g. cost, and practicable in terms of
operation, further improvement should be considered. Such efforts are contained by the
ALARIP principle (As far as reasonably practicable) which is prevailing in the UK offshore
industry.

CONCLUSION

in the paper, recommendation is made to base future ships stability rules and regulations on
the principles of risk analysis. As a first step, a level of acceptance should be discussed. If
the present accident statistic is unsatisfactory, it should be spelled out, otherwise effort could
be spent on other types of ship accidents.

For most large ships, intact stability is considered as acceptable regardless of weather, and
emphasis is on damaged stability and water ingress. Qualification, education and training of
ship officers as outlined in STCW 1978 is considered as adequate to maintain stability.

For some large ships design or high-speed operation may, especially in severe weather, cause
an increase in moments acting on the ship to an extent that calls for qualification, education
and training beyond what is common today in order to obtain an acceptable standard. Data
for calculation of risk is to some extent available, and model experiments can be made where
such data is lacking.

For smaller ships, "severe weather" is only "moderate weather" for a larger ship. T'his in
itself calls for risk control as risk may otherwise become excessive if the ship is operated
without regard to the weather. To some extent, existing regulations reflect this matter in
terms of restricted area of operation etc., but a systematic and logic approach seems to be
lacking.

ideally, the concept of "capsize" in relation to intact stability should be eliminated by a

combination of technical improvements in design of smaller ships, and by improved
awareness by those operating the ship with regard to keeping the ship intact and securing
cargo. Such restrictions, and adequate operational advice specific to the ship should be
presented in an operational manual.

As about 45 000 relatively small vessels out of a total of approx. 80 000 are in existence,
the short-term solution is to consider more specific operational restrictions, based upon data
from experience, model tests and other research, and supported by better qualifications,
education and training of naval architects and ship officers.

The most important technical improvement for new ships in the lower size range is
considered'to be extention of the intact stability to at least 80 degrees heeling angle as has
been implemented by the Norwegian authorities sinze 1980 for certain groups of new ships.
 I1
REFERENCES

Dahle, E. and Nedrelid, T. Operation manuals for improved safety in a seaway.

Proceedings. STAB 86.

COST 301. Casualty data for the COST 301 area. CEC 1987.

Dahle, E. Myrhaug, D. and Dahl, S. Probability of capsizing in steep and high waves
from the side in open sea and in coastal waters. Ocean Engineering, Vol. 15, pp.
139-151, 1988.

Anon. Use of environmental and operational data in design for safety of smaller
vessel. Proceedings, PRADS'92, Vol.2, pp. 1430-1442.

Myrhaug, D., Dahle, E., and Rue, H. A two-dimensional Weibull distribution and
its application to rolling. Proceedings, OMAE-93 (to appear).

Dahle, E. and Myrhaug, D. Risk analysis applied to smaller vessels in breaking

waves. Transactions, RINA 1993 (to appear).
 CASUALTY STATISTICS:

During the years 1982 through 1991;

153 Norwegian ships were lost at sea,

58 lives were lost in these accidents.

30 of these ships capsized ( 20 % ),

27 lives were lost due to capsize ( 47 % )

STABILITY IS IMPORTANT FOR

SAVING LIVES AT SEA
"Dangerous areas" along the Norwegian coast

o .c
.....

to
Out of 30 ships capsized while under way

NONE WERE LARGER THAN

500 GROSS TONS
Average percentages of Norwegian ships
capsized during the years 1982 - 1991

.2

% ------------------------------------------------------ --------- -- --

.1

Fleet < 100 >100 Fishing Others

avr. Gross ts Dry cargo-
The casualty rate was 3.2 times higher
than the fleet average for general dry
cargo ships less than 500 gr. tons

10 out of 12 dry cargo ships capsized

because of shifting of cargo
 It can not be documented that the intact
stability criteria used for Norwegian ships
(IMO Res. A.1 67(ES.IV)) are inadequate,

- PROVIDED THAT CERTAIN BASIC

OPERATIONAL PRECAUTIONS
HAVE BEEN TAKENI

These casualties were almost without

exception caused by operational errors
Seakeeping qualities are only part of the
problem ;

- i.e. broaching may (and have) lead to

shifting of cargo.

The primary cause of the accident would

still be operational.
Destruction of stability characteristics through lack
of understanding basic stability assumptions

As~.0o6-3 rn-ra& 4P4w.

IN.
5 0.104
0.4 S 4

0.3 •4

-0,2,

DOOR I
4914ýHl TcH
"Operation Manuals" might be a
good idea, but.......

... don't we all tend to read just

enough of user manuals "to
make the $@+ thing work"?
Future research for improved stability
criteria should take the human factor
into account.

How can new regulations be made to

work in practice ?

Ways and means for reducing operational

errors, and their consequences, must be
developed.
University Enterprise Training Partnership - Short Course on:
NEW TECHNIQUES FOR ASSESSING AND QUANTIFYING VESSEL STABILITY AND
SEAKEEPING QUALITIES
Trondheim 8-11 March 1993
Session : Seakeeping. Large Vessels
Lesson : Full scale results
Lecturer: L.Sebastiani

ABSTRACT

Purpose of the present lesson is to describe the problems posed

by execution and elaboration of full-scale trials on large
vessels at sea through illustration of a case study. In
particular the results of a series of full scale measurements
of
the dynamic behaviour of a Single Point Mooring (SPM) system at
sea are presented. Enphasis is given to the presence of slow-
drift oscillations in the surge motions and mooring stresses.
Linear coherence analysis has revealed a strong correlation
between low-frequency surge and mooring stresses. Cross-
bispectral analysis has been applied to the wave-surge
relationship, showing a strong quadratic coupling between input
wave freauencies and output surge oscillations.

1. INTRODUCTION

It is well known (1] that moored vessels in random seas

experience large amplitude low-frequency (LF) oscillations, often
referred as slow-drift motions. They occur at frequencies much
lower than the usual wave frequency range, where the energy
content of the sea is negligibly small, so that traditional
linear sea-keeping theory fails to explain this phenomenon.

In order to understand the instauration of slow-drift motions it

was necessary [2] to invoke a non-linear mechanism, extending
sea-keeping theory up to second order.
This comes out from the fact that for a non-linear system there
is interaction between the input frequencies. If one considers
the ship, from a mathematical point of view, to be a second
order, or quadratic system, couples of wave frequencies interact
each other giving rise to sum- and difference-frequencies which
are respectively higher or lower than the wave frequencies.

The natural frequency of moored vessels, especially in the

realistic case of large masses and small restoring stiffnesses,
is rather low ( of the order of 0.005 hiz corresponding to a
period of about 200 s ) and wave frequencies cannot normally
excite it but difference-frequencies can do so, generating large
amplitude slowly-varying resonant motions.
This phenomenon has been justified theoretically and
experimentally verified by model testing, unfortunately available
full scale mesurements of slow-drift motions of moored vessels at
sea are rather scarce.

The lesson is addressed to present the methodology adopted to

achieve and subsequently elaborate a series of full scale data on
the dynamic behaviour of a SPM system at sea.

An extensive analysis of the experimental results has been

performed specifically devoted to evidentiate the non-linear
mechanismof slow-drift oscillations.

The usual way to analize the dynamic behaviour of an at-sea

vessels is by means of standard spectral techniques (3]. Being
inherently linear these methods are however unable to deal with
non-linear systems and more advanced theories (4] must be used.
In particular cross-bispectral techniques [5] are addressed to
the analysis of input-output relationship for quadratic systems.

Cross-bispectral analysis has been therefore applied to the

collected experimental results in order to get a deeper insight
into the non-linear neture of slow-drift oscillations. To this
end the relation between wave elevation and surge motion has been
carefully analized....

2. FULL SCALE RESULTS

Aim of the experimental trials was to collect full scale data on

non-linear response of a moored vessel at sea. This required the
solution of several pratical problems among which

- to have such a system available.

- to perform the measurement in corrispondence of significant

sea conditions.

- to realize long periods of continuative data aquisition.

- to find a measurement equipment able, to detect slow-drift

oscillations with the required accuracy.

In the following paragraphs we will present our answers to these

problems.

Z.1 Trials Dlescription

The compliant: struc ture used for full scale trials was a single-
point mooring, (5PM) system of turret-kind consisting in the oil-
carrier "AGIP Firenze" and devoted to the storage and preliminary
exploitation of the oil coming from the Nilde field, 40 miles off
Sicily coast near Trapani.
The main characteristics of the ship are

LBP = 280 m

4
B = 0.78 m

T = 16.38 m

A = 139,000 t

The features of the mooring system are illustrated in Fig. 1.

In order to perform the measurement in corrispondance of

particulary significant sea conditions it would have been
necessary to continuatively monitor the system for a whole winter
season. As this was not feasible for practical budget and
organization reasons, besides the announced next dismounting of
the system, trials duration was restricted to 15 days from May
5th to !9th 1q89. In this period it was however observed quite an
important sea state between May 14th and 15th with a maximum
significant wave height of about 3.5 m.

As the need of long periods of continuative data acquisition gave

us some storage problems it was decide to perform only 2 periods
of con-inuativeacqdisition,of 8 hours each for each trials day.
Ship position and wave elevation were however continuatively
measured for the whole day in order to monitor ship trajectory
and sea conditions.
A sample frequency of 1 hz, sufficient to account for the wave
frequencies. was chosen for all the measured quantities with the
only exceptiohn of ship position which, during the extra-periods,
was sanoled at 1/60 hz.

2.2 Equipement Description

In order to detect the non-linear nature of slow-drift

oscillations, the following quantities needed to be measured

a) Environment conditions

- Wave elevation and direction

- Wind speed and direction
- Current speed and direction

b) Shi: position and heading

In o::er to have more complete information on the dynamic

heilavIc'r of the system, the following additional quantities were
nleasul' : .

c) Wavr inducec motions

d) Stresses on the mooring yoke

We will describe the measurement devices according to the above
subdivision.

a) Environment conditions

- Wave elevation was measured by means of a wave buoy anchored

at about 1.5 miles from the ship, accounting for prevalent sea
direction in that season. A directional wave buoy would have
been necessary to measure also wave direction, but
unfortunately such device was not available.

- Wind speed and direction were measured by means of an

anemometer.

- No device for current measurements was available for the

trials.

b) Ship position and heading

- Ship heading was measured by means of a gyrocompass.

- An effective device for reliable measurements of absolute ship

position within the required accuracy of 0.5 m was found in the
radio-positioning system SYLEDIS, specifically rented for these
trials and based on 4 transmitting stations on the ground
(respectively. on-Marettixniisland, Pantelleria insland, Trapani
and Marsala ) and 1 receiving station on board.

c) Wave induced motions

A stabilized platform was installed on board to measure the 3

components of-linear accelerations ( surge, sway and heave ) and
the 2 angular desplacements ( roll and pitch ).

2Y

0
d) Stresses on the mooring yoke

Four strain gauges were symmetrically installed in the lower left

hand of the mooring yoke, measuring the longitudinal stresses
along yoke axis.

The configuration of the measurement equipment is shown in Fig.

2. It could be noted that the devices for stresses measurements
were located at bow while the remaining devices were located
towards stern. Nevertheless it was decided to locate the
aquisitien system, constituted by 3 PC IBM, on the bridge in
order to make monitoring and maintenance operations easier.
The configuration of the acquisition equipment is shown in Fig.
3. One PC IBM was connected to SYLEDIS system, the second one was
connected via MODEM to a pre-acquisition system for the stresses
signals and the third was connected to the remaining devices.

3. DATA ELABORATION

A preliminary examination of-all the collected data was directly

performed in time domain. To this end all the signals were
subdivided in segments of 1024 s corresponding to about 17 min,
their time hystdties plotted and analyzed to find out minimum and
maximum values, standard deiations and zero-crossing periods.

From this examination it was found out that more significant sea
conditions and dynamic responses appeared in 2 successive
aquisition periods of 8 hours each, between the previously
mentioned days of May 14th and 15th. The enviromental conditions
relevant to the two periods are shown in Fig. 4.

It must be noted that a SPM system at sea freely tends to hold

the bow on the prevalent environment direction, so that the more
relevant dynamic quantities from the point of view of non-linear
behaviour are the slow-drift surge motions along longitudinal
ship axis direction and the slow-drift oscillations of the
longitudinal stresses in the mooring yoke.

While longitudinal stresses in the mooring yoke were directly

measured, no such direct measurement of slow-drift surge motions
was available, so that an undirect evalutation of surge was
performed on the basis of the measured ship absolute position.
To this purpose the cartesian x,y coordinates of ship position
measured by SYLEDIS with respect to a conventional geographic
system were preliminary referred to the mean position of the
turret sy.stem, estimated as follows.

During a continuative series of aquisitions the ship described a

roughly circular trajectory, shown in Fig. 5, the center of which
has been taken as the mean position of the turret system. We have
then expressed these new cartesian coordinates x.y in terms of
the polar coordinates R, 5, denominated respectively radius and
phase.
As it Can seen in the figure below, the radius can be identified
with slow-drift surge provided that ship heading is equal to
270 -

us) j_sb r r si
COWP IIEADI)JC(

EAU" TuQEr POSIT1ed

In Fig. 4 we have compared with 270 - for the two previously

mentioned periods:and it ca6 be noted that they are nearly equal
so in the following elaborations we have identified the radius
with the slow-drift surge motions.

It can be noted from Fig. 4 that ship heading and wind direction
have quite exactly the same trend with a small systematic
difference that could be attributed to the presence of current.

A representative ship trajectory and representative time

hystories of wave elevation, surge motions and mooring stresses
are illustrated in Figs. 6,7,9 and 11. In particular Figs. 9 and
11 show clearly that both surge and stresses signals present
slowly-varying oscillations with a period of about 200 s. In
the following two paragraphs we will describe the results of a
more refined analysis in the frequency domain performed on the
data relevant to the two periods of May 14/15th.

3.1 Linear Analysis

Conventional spectral analysis was applied to the relevant

dynamic quantities by means of a suitable FFT algoritm [6]. Let
M be the number of segments in which the time history of a
generic signal x(tL is divided and N the number of samples in
eachi segment.
In this case the Fourier amplitudes of- x(t) are given by

Xk•*= . _A
=I.,..• L I ...

The spectrum is estimated by

.12 (2)
S,.(k) = i f IX(k'(2

In Figs. E.10 and 12 you can see as an example the power spectra
relative zo the three time histories of Figs. 7,9 and 11. Figs.
10 and 12 show in particular a well defined peak in the power.
spectra c: both surge and stresses signals a: a frequency of
about O.0C5 hz which corresponds to a period of 200 s.
In order to give'a.quantitative evalutation of the correlation
between low frequency surge and stresses oscillations linear
coherence analysis was applied to the surge-stresses
relationship, having previously depured these signals from high
frequencies by means of a moving average filter.
The lines: cospectrum between the input signal x(t) and the
output signal y(t) can be evalutated as

Sr,(k) = hZ X(k)r(u)Y(k)U) (3)

and the ccrrespondent linear coherence squared as

2 (4)
C,()
C,(k) S.,,k)Sý,(k)
IS,,(k)l

In Fig. i3 are shown the superimposed spectra of LF surge and

st ress,-e. nod thr:eir linear coherence. It can be seen that the
coherence attains its maximum values, nearly 1. in the region
where th: t-wo spectra are significant.
3.2 Cross-Bispectral Analysis

As previously mentioned cross-bispectral analysis can give a

quantitative indication of the non-linear relationship between
wave and slow-drift surge (7,8,9,10]. The cross-bispectrum
of the
input signal x(t) and the output signal y(t) can be estimated
by:

U
)
SZZk, ~y>1X(k)-(i)X(1Y-(i)Y(k + 1)(j (5)

and roughly measures the strength of the quadratic coupling

between the input frequencies w,, w, and the output frequency w•,
= ww + W£ .

The corresponding bicoherence spectrum squared is

(k 1) =- S)IX(k)(i)X(I)(jl12Y(k
,tl (6)"
+ 1)

The BCS will take a value close to unity when the output
frequency w, = wk + w 2 is excited by quadratic coupling
of input
frequencies wl and wk , while a value of BCS near zero implies
an absence of interaction.

Application of cross-bispectral analysis to the present

problem
requires a long continuative time record in order to achieve
the
frequency resolution adequate to detect the non-linear low-
frequency oscillations. Furthermore a sufficient number
of time
segments is necessary for a reasonable statistical estimate. On
the other hand the requirement of long time records is generally
in contrast with the need of satisfying the underlying assumption
of stationarity. The selection of the proper time record must
result from a compromise between these opposite needs. Besides
the above formulation requires the wave elevation signal
to be
gaussian.

To this end run and chi-squared tests were performed on the

wave
data relative to the central period from the end of May 14th
and
the beginning of May 15th aquisitions and corresponding
to the
most significant sea conditions.

A continuative time record of wave elevation of 4096 s,

starting
from the beginning of the May 15th aquisiCion. was found to be
satisfactorily stationary and gaussian. To this record, and to
the corresponding time record of slow-drift surge motions, cross
bispectral analysis was hence applied dividing the total time
lenght in 4 segments of 102'. s each.
In Fig. 14 it is shown the computed BCS between wave
elevation
and surge motions. It should be noted that the
BCS is
substanzially different from zero in the region corresponding
to
the frequency band from about 0.1 to 0.2 hz, where most
of the
wave energy content falls. Furthemore the peaks of BCS
are
grouped in a narrow spectral band about the line wk + wt
- 0.005
hz in the difference interaction region. This means
than the
slow-drift resonant motions of about 200 s of period
are
effectivly excited by a quadratic coupling between
various
spectral components of wave elevation.

4. CONCLUSIONS

The results of the present research can be summarized as

follows:
- the presence of slow-drift oscillations in the surge motion
and
longitudinal mooring stresses of a SPM system has been
experffientally detected by means of full scale measurements.

- linear coherence analysis has evidenced a strong correlation

between surge and stresses low-frequency oscillations.

- cross-bispectral analysis has indicated the existence

of a
quadratic coupling betweeh exciting wave frequencies
and
resulting surge motions.

Even iH these results are quite satisfactory we are

conscious
that a mere continuative and complete series of full scale
tests
would *e necessary in order to thoroughly define the non-linear
nature cf slow-drift oscillations of an at-sea moored vessel.
In
particular the forces on the mooring yoke should be investigated
more in detail and direct experimental indications on
current
conditions and wave direction should be acquired.

This research should be however intended as a pilot study which

has allctýed us to get a first experience of the pratical
problems
related to the measurements and analysis of slow-drift phenomena,
which c:uld be of great help for future planning of
similar
trials.
5. REFERENCES

[1) Remery, G.F.M., Hermans, A.J., "The Slow Drift Oscillations

of a Moored Object in Random Seas", Society of Petroleum
Engineers Journal. 191-198, 1972.

[2] Pinkster, J.A., "LF 2nd Order Wave Exciting Forces on

Floating Structures", NSMB publ. N. 650, Wageningen, 1980.

[3] Bendat, J.S., Piersol, A.G., "Random data: Analysis and

Measurement Procedures', Wiley-nterscience, N.Y., 1971.

[4] Schetzen, M., "The Volterra and Wiener Theories of Non-

Linear Systems", Wiley, New York, 1980.

[5] Dalzell, J.F.. "Application of the Functional Polynomial

Model of the Added Resistance Problem " , llth Symposium on
Naval Hydrodinamics, University College, London, 1976.

[6] Brigham, J.F., "The Fast Fourier Trasform", Englewood

Cliffs, Prentice-Hall, NJ, 1974.

[7] Kim, Y.C., Powers, E.J., "Digital Bispectral Analysis and

its Applications to Non-Linear Wave Interaction", TEE Trans.
on Plasma Sc., Vol. PS-7, N. 2, June'1979.

[8] Kim, S.B., Powers, E.J.", Miksad, R.W., Fisher, F.J., Hong,
J.Y., 'Non-Linear System Coherence Analysis of the Surge
Response of TLPs Subject of Non-Gaussian Irregular Seas",
OMAE 89, The Hague. March 1989.

[9] Choi, D.W., Miksad, R.W., Powers. E.J., Fisher, F.J.,

"Application of Cross-Bispectral Analysis Techniques to
Model the Non-Linear Response of a Moored Vessel System in
Random Seas", Journal of Sound and Vibration, 1985, 99(3),
309-326.

[10] Borresen, R., "Cross-Bispectral Analysis via Fourier-

Trasform of Time Series", NSFI Rep. R-136 .82, Trondheim,
1982.

[11J Dogliani, M., Monti, S.. Sebastiani, L., Tedeschi, R.,"Full

Scale Measurements and Analysis on a SPM system", Paper to
be presented at the next OMAE 91 Symposium, Stavenger.
Norway. 1991.
 C

-A

I
I

C
I.' C
-'C

r z r 0~

'I

4~

0J
I&.

CE
t~ ti
-'p
C-
U' b..
C,
* no
C

no 0
*-40

x
I

'I,
La

44
0)

C1j

0)
C)

-1
tX4J
 - g. 5. Ci ccu" a: trate:tory Zeszri bed by
-- e Ship.

:1g. 3. Lay-out of the azqu; Stýofl SVsZCm

F Fto 1
U bo o 11t ,
 0C>

-14C N
00S--r4 a

.0 -.- 0 r0

14

0
--. L4,

-14

S)g 00

- 0
II. 44

C)O
0E o
vac
4Pi

ow t
0 C0

CDN u
C-
CDi

a>
2' ~0

CD,

CO
coa w
N0 0
 .l. 0

.04 00a

H 0 0CC

CL,

0 0

- 1.5.*

a. . oa o

.- . -- ... . .. .....-..-

-4 0 O)

N)C)
to~~ 004 o
_e-- - o

0-4 r o

- - m
0)4-4
2 a 0 0dN
CIO
-- 0

--
-- ) i in.-•_ a)•.

0oU 0
o-a - . 00
____
____-.-4 ____ .4 5- -
(14o o -
 o W wi

0L 0) 0

Q 4J w w u
-- w b .aoc
-0 :)

ts a] 0

CID w
 0
0'
'I

S
*
S

'0 -. C
U

1IA1~& C
L.

sZ a
0*-'
U .C

.0 C
0

* *fl
U. *t

C'
*0

A)

CQ

3
V
( LECTURE t
SHIP OPERATION IN A SEAWAY

LECTURER: Tor Svensen, Det Norske Veritas Classification A/S

1. INTRODUCTION

Seakeeping theories are finding increasing use in connection with ship operation. Rapid
calculation methods combined with onboard computers present new opportunities.
This section will present an overview of the various factors affecting vessel
performance in service. The relative importance of these various sources of loss in
performance are discussed. Examples of voluntary and involuntary speed loss are
presented and practical levels of seakeeping criteria are discussed. The principles of
operational simulation and its role in the operation of ships are dicussed. Advanced
systems for weather routing are presented and the possible benefits are outlined. The
use of strategic weather routing in combination with a system for tactical routing
(operational guidance) for damage avoidance is discussed. An example of the
application of such a system is presented. Future development trends are discussed.

2. SPEED PERFORMANCE IN SERVICE

It is often said that a new vessel will never again achieve the same level of
speed/power performance. The loss of performance in service can be attributed to
three main effects:

I) - Deterioration in the underwater hull and propeller condition

2) - The effects of the environment on the ship, including the wind and
the sea

3) - The effects of increasing ship resistance on the operation of the

propeller and the corresponding effect upon the main engine

Deteriorative effects are attributed to increasing hull and propeller roughness and
possibly periodic fouling attacks. These are essentially a function of ship age, time out
of dock and (lqulity oflmaintenance strategy. lI-vironmenial effects are (lependenl upon
 windspeed and direction and the waveheight/period and direction, all of which are
functions of vessel location and movement of low pressure systems, monsoons and
other climatic factors.

In the following the environmental effects will be discussed in some more detail.

Loss of performance due to the effects of wind and waves can be attributed to
involuntary and voluntary speed loss.

INVOLUNTARY SPEED LOSS includes:

- direct and indirect effects due to wind acting on the upper works of the
(1 vessel

- added resistance due to wave induced ship motions in the seaway

- added resistance due to short waves (wave reflection)

- added resistance due to steering and manoeuvring in the seaway ( principally

yawing and rudder motions)

VOLUNTARY SPEED LOSS is a deliberate reduction of speed and/or alteration of

course by the ship's master to prevent damage due to excessive motions or hull
structure loading. The most common effects are:

- excessive slamming, leading to structural damage

- shipping of green water leading to loss or damage of deck cargo or

equipment and, in extreme cases, loss of watertight integrity

- propeller emergence leading to unacceptable propeller shaft loadings

- extreme accelerations

Considering the operation of a slow speed diesel engine, there will be an acceptable
torque rangchr a given RlM within which the engine can normally operate. This can
be an important limiting factor for involuntary speed reduction
Voluntary speed reduction or alteration of course in service is in practice a highly
subjective action by the ship's master based upon an observed degradation in
habitability or operability. By describing seakeeping in terms of physical parameters
such as absolute or relative motions, accelerations, bow wetness and slamming, it is
possible to quantify performance and to determine limiting values from an operability
point of view. These "seakeeping criteria" forms a furndamental part in design analysis
as well as operational weather routing using environmental data.

The relative importance of the various factors influencing vessel performance in service
can be determined using some form of voyage simulation model. A voyage simulation
model is a computer based model for simulating the various in-service conditions and
their effect upon the technical and economic performance of the vessel. Probably the
most important concept in the total voyage simulation model is that the factors
affecting the vessel performance are modelled in combination rather than on an
individual basis. This ensures that more realistic results are obtained. The degree of
sophistication required in a model will depend upon the type of analysis to be
performed.
(

2. "SEAWAY" - A COMPUTER PROGRAMME FOR OPERATIONAL

SIMULATION

The computer programme SEAWAY has been developed for evaluation of the
seakeeping performance of a ship in a seaway. SEAWAY is a practical tool for
designers as well as operators. When used in combination with other computer
programmes, such as OPTWAY, the computer programme SEAWAY may also be
used as basis for advanced weather routing. This aspect is described in more detail in
the following sections.

SEAWAY calculates:

- added resistance due to waves, wind and current

- power setting
- actual ship speed in a seaway (short and long term predictions)
- probability distribution of ship responses such as : vertical and horizontal
velocities and accelerations, rolling, slamming, deck wetness etc.
- ranges of ship speed and courses to be avoided due to exceedance of
seakeeping criteria for a given environmental condition

In combination with the computer programme OPTWAY the optimum route can be
calculated.

The enclosures provided in Appendix I and Appendix 2 describe the principles and
calculation method employed in SEAWAY in more detail.
(..3. STRATEGIC WEATHER ROUTING

General principles. The avoidance of bad weather is the most effective way of
minimizing both operational delays due to voluntary and involuntary speed losses and
occasional heavy weather damage. Weather routing is a method of strategic route
selection using weather forecasts to calculate the optimum route. Criteria for use in
advanced routing procedures will typically be relative motions, accelerations, time and
fuel consumption.

Routing services have been provided on a commercial basis for many years. These
have been relatively primitive services where the main emphasis has been on
meteorology and with little or no consideration for the ship motions and seakeeping
cniteria for the individual ship in question. Advances in seakeeping theory, computer
technology, meteorology and communications technology now makes it possible to
introduce advanced routing services where criteria for speed loss and voluntary speed
reduction are tailored individually to the ship and route optimtization can be performed
directly by the ship's master on the bridge of the vessel. Such systems are under
development and are in the process of being introduced on a commercial basis in 1993.

In practice, weather routing is only usefuil in ocean areas with changing weather
patterns. It is also neccessary that the voyage is of sufficient length to permit feasible
alternative routes. Probably the most important weather routing decision is the
strategic choice made at the start of the voyage. This will typically involve a decision
about following a northerly, great circle or southerly route. Once a decision is made
and the vessel is underway, the economic penalty associated with an alteration of route
K becomes large. The need for an accurate. prognosis covering the complete length and
duration of the intended voyage is therefore obvious.

Storms have three spacial dimensions and in addition a fourth dimension in time.
Therefore a four dimensional consideration of the atmosphere is necessary when
planning to avoid the worst effects of middle latitude depressions

The need for accurate prognosis cover both wind and waves. Between thre two, waves
arc far more important for ship performance, although accurate wind prognosis is the
basis for accurate wave predictions.
(" Wind and wave predictions.

All weather forecasts extending more than a few hours ahead in time are based upon
numerical models of the atmosphere. The models are run to predict the development of
the atmosphere in space and time. Prior to running the models, all available
meteorological observations are input to the model as basis for setting the initial
conditions. The improvements in numerical models combined with advanced
supercomputer technology has resulted in considerable improvements in both accuracy
and length of forecasts. The latest models are capable of producing forecasts extending
about 10 days out in time and the accuracy of the forecasts has improved considerably
over the past few years.

The atmospheric models provide information about the wind fields at the sea surface.
This information serve as input to the wave models. Global wave models can today
provide forecasts of waveheight as well as wave period and direction for both wind
generated seas and swell. The latest wave model run by the European Centre for
Medium Range Weather Forecasts (ECMWF) is capable of wave forecasts as far
ahead in time as 10 days. The new model only became operational in July 1992 and
represents a major step forward when it comes to weather routing of ships. The level
of accuracy is not yet well documented. A comparison with actual waveheight
measurments made at the Snorre Field showed a prognosis error less than I m on.
significant waveheight for the 1st day in 75% of the cases. For the subsequent days
from 2 to 7 the deviation was less than Im in 67, 58, 59, 48, 46 and 49 percent of the
cases respectively.

The data from ECMWF is today available in a resolution matrix of 3x3 degrees on a
global basis. This is sufficient for weather routing of ships on major routes such as the
North Atlantic and the Pacific.

Weather routing principles.

The role of seakeeping in weather routing is to provide a rational basis for estimating
the voluntary and involuntary speed losses for alternative route selection. Using the
computer programme SEAWAY, a database containing all the relevant information
can be tailored to (he individual vessel. The route optimization programme OPTWAY
canl thus be used in combination with the forecasts for wind and waves to evaluate and
select the opti mum ioutc The b)asic principles of route optirmization can be described
as follows:
In the absence of any form of route optimization, the ship will normally follow a Great
Circle Route, which in principle is the shortest distance route. Many different
optimization methods have been developed. The modified isochrome method is based
upon a least time route. In this method a number of lanes parallel to the great circle
route are defined. The time step dt is defined as the interval between two calculations
in the optimization procedure. This interval may correspond to the time between
successive weather forecasts. The optimization is carried out as follows:

Consider the possible positions of a ship after navigating from point A

following different headings during time dt. These arrival points represent
a line called the isochrome. By selecting one point on the isochrome inside
of each lane we define a set of departure points X(i) for the next time step.
From each departure point X(i) a number of different ship courses may be
chosen. If the ship will follow these headings over the next dt time, the
different arrival points may be computed. They represent the isochrome
after 2 x dt time. For the next step only one point inside of each isochrome
is selected using the shortest great circle distance to the destination point
as the selection criterion. Repeating this procedure until the ship reaches
the destination point, the optimum route with the shortest voyage time will
be determined.

Other criteria may be used instead of the least time criterion explained here. These
may, for example, be lowest fuiel consumption or lowest ship responses. The accuracy
of the route optimization is dependent upon the quality of the wind and wave forecasts
as wll as selection of seakeeping criteria and how well the vessel is described in the
model. A very fundamental parameter in the route optimization process is the
numerical values of the individual seakeeping criteria parameters used. These will
normally be selected based upon subjective experience from previous vessel operation,
measurments or published results. Actual measurments will provide the best basis for
determining relevant seakeeping criteria. This may be combined with a system for
tactical routing as described below.
 4. TACTICAL ROUTING

Background

Even when vessels are using active weather routing, weather patterns can change
unexpectedly and the ship may find itself encountering severe weather necessitating
speed reduction or course alteration. Such tactical weather routing actions are taken by
the ship's master in order to minimize the risk of damage to ship or cargo.

For large vessels, such as containership, potentially dangerous situations are typically
water on deck or bow flare slanning. The identification of dangerous situations from
the bridge can be difficult, partly due to the physical distance from the bow and partly
due to the fact that the view of the forward part of the vessel is often obstructed by
containers. Installation of instruments capable of observing and measuring parameters,
such as motions, accelerations or green water on deck, can significantly improve the
situation by providing relevant and accurate information to the ship's master. Such a
system may be used in combination with a system for strategic routing of ship.

The following describes in detail an example of such a development. In this

development a single sensor in the form of a vertical accelerometer was used. Model
tests in combination with calculations were carried out in order to determine both
vessel motion characteristics and limiting criteria for operation in various seastates. By
measuring the ship response spectrum, this can be compared with the criteria database
to detect possible dangerous situations. The ship's master can from the computer
( display select the optimum speed and heading, thus minimizing the risk of heavy
weather damage to ship or cargo. Shortcomings of this relatively simple system are
that the wave heading have to be input manually and the system is not capable of
handling combined wind generated waves and swell from a different direction.

Model tests. The model tests were carried out using a free running model. The model
was equipped for measurment of 6 d.o.f motions, relative motions at the bow and a
special force panel on the deck for measuring the forces due to green seas on the
breakwater and the first tier of deck containers. Initial tests were performed in regular
waves to determinc the critical periods and wave heights with respect to green water
on the forecastle deck. Another objective of the tests in regular waves was to establish
transftrli nctions and] relative phas': for the actual ship inotions for comparison with the
theoretical calculations. Irregular tests were performed for two dill'erent speeds and for
( headings 0 and 30 degrees. The purpose of the irregular wave tests was to obtain
statistical information on the vessel motion, relative bow motion and forces acting on
the breakwater structure. By gradually increasing the waveheight at critical wave
periods, it was possible to establish limiting criteria for green water on deck.

Numerical calculations. Calculations of vessel motions were performed using a linear

strip theory program. The purpose of the calculations was to obtain transferfuctions
for a wider range of headings and speeds than covered by the model tests. The
calculated responses in regular waves showed good agreement with test results.
However, for the extreme responses in irregular waves, there are clear limitations in
the linear theory assumtion as discussed below.

Criteria development. The model tests showed a clear relationship between the
relative motion and the amount of green water on deck. A simplified relationship
between vertical motion and relative vertical motion in the bow was developed based
upon the results of the model tests. This simplified relationship is based upon the
assumption that there is only one wave spectrum (i.e. not a combined situation with
wind generated waves as well as swell) and that the critical situations.of interest are
the bow to beam headings. Using vertical motios instead of relative vertical motions as
the parameter for limiting the vessel operation makes the issue of instrumentation
considerably simpler.

It is important to note that a significant difference was observed between the calculated
relative bow motion and the measured values from the model tests. The difference was
principally speed dependent and was due to the stationary wave field and dynamic
swell-up. When developing the relationship between absolute and relative vertical
motion, the theoretically calculated values were corrected for this effect.

Database development. Based upon the results of the model tests and calculations, a
comprehensive database was developed containing vertical and relative bow motion for
a range of wave headings, periods and ship speeds.

Instrumentation and analysis proerams. An instrumentation pack consisting of a

vertical accelerometer and a dedicated PC for data analysis was developed- On this
system the analogue signals from the accelerometer measuring the vertical
acceleration in the bow, are sampled and coniverted to vertical motion by two times
integration of the signal. At fixed intervals a spectral analysis of the vertical motion
time trace is performed. The results ofthis analysis are transmitted to a second
 operational PC on the Bridge of the vessel where the operational and advisory
programs are implemented.

The most important part of the advisory programs on PC no. 2, is the theoretically
calculated model of the vertical bow motion and the relative bow motion of the ship
for various speeds and headings to the waves. Based upon the data received from PC
no. 1, the operational PC no.2 will present a trend plot with development of the relative
bow motion. A consequence graph is also presented on a semicircle where the effect of
heading and speed changes with respect to relative motion is presented.

Implementation. The system was first implemented on a single vessel in a fleet of

sisterships and tested during a winter season. The system was run in combination with
( asystem for strategic weather routing based upon the SEAWAY and OPTWAY suite
of programs as described previously. The onboard system was implemented with a
data recording facility and a manual log was maintained of observed seastate and
speed/heading where voluntary speed reduction was put into effect. This procedure
allowed the initial criteria for water on deck and voluntary speed reduction as derived
from the model tests to be evaluated and subsequent corrections made based upon
operational experience. The system was subsequently fitted to the complete fleet of
sistervessels.

Conclusions. The peoject demonstrated the effectiveness of model tests in

combination with theoretical calculations in determining realistic operational criteria
for a particular type of vessel. The project also demonstrated that the information
obtained through a systematic set of model tests and calculations can be used in
C combination with relatively simple onboard instrumentation and data analysis system
for providing operational guidance about possible dangerous situations-
Other parameters such as accelerations or vertical bending moment could be used in a
similar manner for other vessel types. It is believed that this type of operational
guidance systems will gain increasing acceptance in the future as costs are no longer a
prohibitive factor.
 (.A

42 4-o
4-o

cz1ý0 ou C
C.,
0 _t

ci)t

Ct W)

W- W

C.). N
C.C)

0) Ctcz 4
Cu 0 ~E
~~oOQ ~tO
( METEOROLOGICAL FORECAST SKILL

SAMPLES OF DEVELOPMENTS IN FORECAST QUALITY

Anomaly corralolion

•0 .6 .-......... -------- ---

0.4 ...... ............ ....

......... ....... Winter
1990/91

0.2 .1979/80

....
. .....
...... .... 1972
0 .
1 3 4 5 6 7 a 9 io
Days

ECMWF FORECAST SKILL (NORTHERN HEJISPHERE)

January 19w0 - Decenter 1990
Forecast skil

1-

to-~
I..... ....
. ................ ~. AG
.VE......
NHLY.EN .........-.........

. 1982
1980 1981... ...... 19841985
1983... I A.O
.......
.... 1988
GE.. 1987
".. - .- - 1989
" - 1988 1990 1991
r,". .........
.-.----.-- ... i O T O IN V R G •...-
WFFORECAST SKILL (TROPICS)
E MONTH *12 MOVING AVERAGE
ECMWF FORECAST SKILL (SOU'rh:ERN.HEIISPHERE)
Foweca,. day on which the 850 h~a
MEN VERGEMONTHLY
NT1{L vedctr wind carel.•ioo roaches ..0.6 MEAN AVERAGE Ja~nuaiy 1980 - December 19930
Faorsl sidU 2 ..........-
........
Januaiy 1980 - Oecerrter 1990 Foroca.l id
1: 11-

Q~sm.r192 I~ sn10~
...
... l .C ... 2 MONT MOVING AVERGE ..... fl~~ic.

2-
V 1-
1983 1981 1982 1983 1984 19-E' 198C 1987 1988 1989 1993 1991 1980 1981 1982 19,2 1984 2985 1986 1987 2988 2989 1990 1991

EF.V' oor-aist skill J:anuary 19M*l ID (ut, duf Vor-,,•r'igon o'.,1 EZMcs vosi Jatu A3ttUC;. pr v tU nf lI (;uSI[flo du
01-
 C. SEAWVAY

A SYSTEM FOR CALCULATION OF PERFORMANCE IN A SEAWAY

C Seakeeping
tests

charateriticsadded
__________________resistance

Velocity
____
____loss ____ ____ ____

EnginenSiAWA

propeller
Weathe datacharacteristics
(waves. wind HP, 0. RPM
and current) _______

Criteria for
voluntary
speed loss
-- OPTWAY

A SYSTEM FOR VESSEL ROUTE OPTIMIZATION

SHIP ATTAINABLE
I'SPEED

(SEAWAY/ J
Il
WEATHER
FORECASTS

(DMI)
CURRENT
DATA

(DMI)
MARINTEK)

IN ------- 0 P T W A Yl

REULSDATA FILE DIGIALIZED

"TO PAPR FFOR PLOTTING+-- MAPS

R 0 U T E P LOT

PLOT
OOPTIMAL
ROUTE
 SN4 Waves Wind Current

Bow

Stern

COORDINATE SYSTEM and SIGN CONVENTION

6t
. 0.

La) re t e)
-irc o t
width L• oArrivallpoi

Departure pointofln

L~i) Great circle route

between departure and
Isochrone arrival point

MODIFIED ISOCHRONE METHOD

 RoRo LPG ca~rder
00(Lpp r190m) Design speed =0 Loin80) Design speed

*0---- .
---- .
....... -..... ...---.- *-----------*o- ----.-- ----
.. ........... 204@ . .---- --.--- -------.........

im 2m 3m 4m Sm Gm 7m min rn 2in- .3m Alm Sm 6m 7mMH,,

Container Vessel LNG carrier

10(Lpp =185 m~) Design speed lt(Lpp 273m) D esign speed

0........ - ....... ...... .............. ........... ... o.......... .. .. . . ......... .......... ........
...........

Im 2m 3m 4m Sm Gm 7m H,,, urn 2mn 3m Am 5m Sm 7m H,,,

Speed loss curves Jor 4 Ship types

- 0

C)=

CO

onoD
>n_ _ _ _ _ _
K HISTOGRAM OF PROBABILITY OF EXCEEDING
SEAKEEPING CRITERIA FOR SHIP 1 AND SHIP 2
ON NORTH ATLANTIC ROUTE

Probability
10
CONTAINER VESEL SHIP 1 SHIP 2
9 LPP 210 m 202 m
B 30.5 m 32 24 m
Draught (operating) 11.58 m 10.02 m
8 CB 0.616 0.634
Speed (operating) 18 knots 18 knots
7-
z z
6 0
z 0
0 0

5 - c0
LCo

4 CLy
z a.0
Z- ii z -
chi
3 r "'o o D ,a (a Mmz
3 ~ -m

2 kt

0k
DECK VERFICAL PROPELLER
WETNESS BOW ACCELERATION EMERGENCE
 Enc1.

T ShI4T6F flOIJP

POSITION. OF TRANSDUCERS

BREAKWATER 131 m from Lpp/2

C.METER FOR
MEASUREMENT OF VERT. MOTI N

135.2 m from Lpp/2

PROBES FOR REL. NOTION

MEASUREMENTS
 SAMPLE RESULTS FROM MODEL TESTS
AND CALCULATIONS

IS knoLs ISknoLs
0 him. vr-tcot bow .o.a oos. r.,o,,., bo- -o,,on
- Colc. verLtcot bo- eolti -! Lte. remottve b motion
2.55 - -T1

- _ _
-
205.0-

2.5- - -

D 2.S

1.0
C1.0---
I.S

0.0 -- -0.0

0 S Io IS 20 2S so 0 5 tO Is 20 25 s0
£rýow prtod (mE ErLoL prtod (f.e I

Measured and calculated responses in regular waves at 15 knots

VerixcoL bow moton, 12.0 knoLs

2.5

I -a- St. d" I%/-, So. qg

2.0- --- '--"- St. iký I.l 90. de

........ ~~o 90 . "• • / -"'U' -- - -

........
...... - -

0.S .

0.0-
O is 20 2S
Pee' uo prod sIeI

ReLoLtve bow mokton, 12.0 knoLs

S.0 -

Z.5 - - - 1 - St. det 1./.1t 0. deg

-~
I St. oý W.sI SO. mog
-. St. de. i..I £0. dog..
20--St. de, Wa.Ide &g - '0 ~ a

_ - -

10.I ---

0 nt 29 2S
 VEOL.

00
...,.o.-
NUMBER OF GREEN SEAS AND En.
RAPPORT
FX-LOADS ON BREAKWATER ReDrt
CATO

HS = 10 m, T = 14 sec. Date
Sp REF.
"Ref

HEADING = 30 deg

TEST NO 240 Vs - 12 kts TEST NO 230. V S - 17 kts

Fx load Number of Fx load Number of
breakwater green seas breakwater green seas

t1000 3 <1000 3
1.000--2000 6 1.000-2.000 3
3.000-4.000 1 2.000-3.000 1
3.000-4.000 3
TOTAL 10 4.000-5.000 2
Encounter waves: 162 5.000-6.000 1
Percentage green seas: 6.2% 7.000-8.000 1
9.000-10.000 1
14.000-15.000 1

TOTAL 16
Encounter waves: 190
Percentage green seas: 8.4%

VS = 17 knots
6 VS = 12 knots

o
.-- '4-

E .

0 2 4 6 8 10 12 14
FX - 103 (kN)
C

.PC2
lentAquit~on Oper ational
aregnal Guidance
IssonPro~cessing SysIMa

Overview of tactical routing system

 T

-3 i

9C.

* .

Screen display - Heavy weather avoidance system.

6
 LECTURE -9

SEAKEEPING IN DESIGN

LECTURER: Tor Svensen, Det Norske Veritas Classification A/S

1. Introduction.

The primary use of seakeeping theories is in connection with the design of new ships.
Seakeeping theories used in combination with model experiments can play an important role in
the design process. However, it is important for the designer to have a good knowledge of the
limitations with the individual theories and methods employed. It is also important for the
designer to know that possible design changes in order to improve seakeeping performance are
feasible. A common problem in the design of commercial vessels is that seakeeping aspects are
considered too late in the design process, after all the majorparameters influencing-seakeeping
performance have been fixed based upon other considerations such as still water. resistance.

The primary area for use of seakeeping theories in design are:

- Evaluation of basic parameters such as motions and accelerations and comparison "
against criteria for cargo integrity, passenger comfort etc.

- Evaluation of speed performance on a given route or selection of routes for

determination of sevice margins and required main engine power

- Calculation of hydrodynamic loads on the hull structure in a seaway and use of this
information in connestion with the structural design of ships

The latter is becoming an increasingly important application as reliability based structural

design procedures are gaining acceptance.

The following sections will discuss some of the important issues relating to the application of
seakeeping theories in design.

2. Use and limitations ofseakeenin2n models

The most common theoretical method for prediction ofseakeeping performance is the linear
strip theory method. l)espite thre very substantial simplifications of the actual physical realities
that arc made in linear strip theory, this method has gained wide acceptance in connection with
ship design. in a survey of ship model tanks to be published by the 1993 ITTC the following
conclusions are presented:

Three-quarters of the Institutions use 2-D strip theory or slender body theory only. Only
about one-quarter use or have developed 3-D computations. The 3-D methods are used
mainly for zero or low forward speed.

Three-quarters of the Institutions use a frequency domain approach. The remaining one-
quarter have a time domain capability.

It was not possible in this survey to distinguish between the methods that are routinely used
and those that are still under development. It should also be noted that some of the institutions
that are the most active in the field of numerical seakeeping do not own any experimental
facility and are therefore not included in the survey. However, the survey do present an
interesting picture of the present state-of-ant.

When considering seakeeping in connection with design, the most important motions are
absolute vertical acceleration and the relative motions at the bow. Vertcal accelerations are
directly related to crew and passenger comfort and safety of the cargo as well as the global
structural loads imposed. Relative motions at the bow are more related to how well the vessel
will be capable of maintaining speed in heavy weather. The following comments are intended
to point out how well current seakeeping theories are capable of predicting the various~
motions and derived parameters of interest in connection with design.

Vertical motions and accelerations: Good results from 2-D strip theory for most ordinary ship
hull forms. Although the theory is limited to small amplitude motions, results show surprisingly
good results for relative large amplitudes when compared with model experiments and full.
scale measurements.

Relative motions: Relative motions at the bow as well as the stern are not predicted well by
linear 2-D theory. This is mainly due to the fact that the stationary wave field (the bow wave)
and the dynamic swell-up due to the pressure field generated by the water entry of the flared
bow section are not included in the theory. In order to predict the relative motions at the bow
with an acceptable degree of accuracy it is necessary to include the non-linear body- boundary
conditions in the Jprediction method- Such program-s have been developed and arc gaining
acceptance iii use.
 Slamming: Slamming pressures are a function of relative verical velocity and geometry. For
deadrise angles less than 20 - 30 degrees, the maximum pressure will occur in the region of
where the jet is formed. Calculation of slamming pressures necessitates the use of nonlinear
theories. The local pressures generated in slamming will normally not have any significant
influence upon the global motions of the vessel. As a result, the only practical way of handling
slamming predictions today is to first predict the vessel relative motions and subsequently use a
non-linear boundary element or similar program for local investigations of slamming. It should
be noted that the heel angle at impact is an important factor when predicting the impact
pressures. Accurate prediction of rolling motion is therefore a requirement in connection with
slamming investigations.

Global loads - vertical bending moment: Vertical bending moment is primarily a function of
inertial loads and pressure forces acting on the hull. For large amplitude motions, linear
theories will underpredict the pressure forces acting on the hull during entry of the bow section
into the water. In order to account for this it is necessary to include incoming wave as well as
the non-linear conditions on the body boundary. 2-D methods for time-domain modelling of
these effects are standard routines today. 3-D methods are under development, but are not.yet
part of routine design procedures. Differences between linear and non-linear predictions of
vertical bending moment can be as high as 50-60 percent.

2-D vs. 3-D theories: Complete numerical solutions based upon 3-D theories with forward
speed are still in their infancy. The major driving force behind the development of 3-D methods
is the prospect of more accurate predictions of hull pressures. This will give more accurate
predictions of local and global loads as well as added resistance in waves. Developments in
computer power may result in 3-D methods playing an icreasing part in routine applications
*' within 2-3 years from now.

3. Seakeepine criteria

When using seakeeping criteria in connection with a design development or evaluation it is

important to remember the following 3 basic requirements:

1. The criteria and corresponding responses must be relevant to the mission of the vessel.

2. Criteria levels must be-related to the actual task or mission of the vessel.
3. The numerical value of criteria levels must be based upon actual ship performance
assessment.

For commercial ocean going vessels, seakeeping performance is principally addressed in terms
of:

Habitability: The ability of the vessel to carry out a mission with a minimum of
discomfort.

Operability: The ability to carry out a mission under all types of weather

A third aspect of seakeeping performance is survivability or seaworthiness. This aspect is

usually not considered in detail by the designer and is generally assumed to be satisfied by
adherence to appropriate classification rules, load line and stability regulations.

In practice the boundaries between habitability aspects and operability aspects of seakeeping
performance are vague and the two will always be considered together.

By describing seakeeping performance in terms of physical -parameters such as absolute and

relative motions;-accelerations, bow wetness and slamming it is possible to quantify'
performance and subsequently evaluate performance in a rational. and systematic manner,
Limiting values for individual performance criteria have been derived from full scale
operational experience. These are:

Design limits: Absolute limiting values which are not to be exceeded in service

Operational limits: Limiting values beyond which performance degradation or

increasing likelihood of vessel or cargo damage will occur

Limiting values for operability forms the basis for voluntary speed reduction in service.
Voluntary speed reduction or alteration of course in service is in practice a highly subjective
action by one ship's master based upon an observed degradation in habitability or operability.
This subjective action is normally not reflected in the design analysis and limiting criteria are
treated as absolute criteria in most aspects of design analysis. This in itself may be a significant
source of error.

The enclosed figures present limiting values for selection of the most common individual
seakeeping performance criteria applicablc to merchant vessels. The significance of individual
criteria for some principal ship types are also presented. From the data presented, it is
significant to note that criteria limits vary considerably depending upon the source and the ship
type. Only for vertical acceleration at the FP is there a reasonable agreement between the
varous sources.

4. Accuracy requirements

The uncertainty associated with establishment of valid limits for individual seakeeping criteria
are important to consider against the uncertainties in the seakeeping prediction method itself.
It is often argued that uncertainties associated with the establishment of limiting values are not
significant when comparing characteristics of two alternative designs. This argument may hold
for naval vessels where mission effectiveness is primarily a function of seakeeping
performance. For merchant vessels seakeeping performance is usually a trade-off against other.
factors such as still water performance and building costs. The absolute value of each
individual seakeeping criteria limits can be an important factor in establishing a merit rating
between alternative designs in techno-economic terms. Using a more severe limiting criteria
will penalize a poorer seakeeping design more in terms of loss of performance. This may lead
to the incorrect conclusion that a particular design is outside the limit of a stated performance
specification.

In order to illustrate the consequence for the design process of errors in seakeeping
predictions or in criteria limits, a case study is presented for a medium size container vessel
operating on the North Atlantic Route. Details of vessel and route are given in the enclosed
table.

The effects of introducing an error in the estimate of the relative motions at the FP upon the
limiting speeds due to deck wetness have been investigated for a range of wave heights in the
North Atlantic. The results of this investigation for a maximum deck wetness probability of
3% are shown in the enclosed figure. Similarly, the effects of changing the limiting value of
individual seakeeping criteria have been investigated, and the results for vertical accelerations
at FP and deck wetness are shown in the enclosed figure.

The results as presented show that variations in the predicted responses, which are well within
the limits of accuracy for present seakeeping theories, can result in dramatic reductions in the
predicted speed for the vessel. An overestimate in relative motions by I metre in this case
represents an error ofonly 12% with a corresponding reduction in maximum permissible speed
of over 3 knots in head seas in the range of significant wave heights from 6 to 9 m. These
results are clearly of significant value in the design process. If voluntary speed reduction at a
significant wave height of 7.5 m is unacceptable to a particular operator, then the designer will
have to improve the seakeeping abilities of the design by, for example, increasing the freeboard
in the bow region. When using this type of deterministic analysis of extreme events it is clearly
important that the seakeeping prediction method employed is capable of generating accurate
results and that criteria limits are stablished with a high degree of confidence. However, in
reality, a vessel will only meet such extreme event situations during a very small percentage of
the total operational time. The overall impact upon techno-economic performance may
therefore be relatively small. This is in practice best examined by performing an operational
analysis for the proposed design using a voyage simulation model. An example of such a
model is the SEAWAY system as already presented. In a voyage simulation model the vessel
operation i modelled on a realistic trading route and the various factors affecting vessel
performance in service are modelled in combination rather than on an individual basis. The
principal steps in a voyage simulation is shown in the enclosed figure.

The environment is described in terms of wind speed, direction, wave height, waveperiod and
direction for predetermined segments of a given route. This permits both voluntary and
involuntary speed losses to be taken into account in the analysis.
Involuntary speed loss is defined as caused by

- Wind resistance
- Added resistance in waves (short waves and ship motion domain)
-. Added resistance due to steering (rudder motions)

Voluntary speed loss is a deliberate reduction in speed or change in heading due to exceedance
of one or more limiting values for individual seakeeping criteria.

Long term or short term statistical weather data can be used in a voyage simulation analysis.
In a design evaluation long term annual statistical data for the relevant routes are usually
employed. This provides the most realistic basis for assessing the total economic measure of
merit for the design.

The selection of operational profile (route) is clearly important for the results of the analysis.

The probability of encountering severe weather conditions varies considerably between

different sea areas with a corresponding variation in vessel performance. It is therefore
important that the design analysis is performed using a realistic operational profile
corresponding to the expected future service profile of the vessel.
( In order to illustrate the above points in some more detail and to examine the importance of
seakeeping characteristics and voluntary speed reduction against other involuntary speed
losses, the container vessel described in the enclosed table has been modelled on a North
Atlantic trading route with a typical roundtrip schedule of 21 days.

In environmental terms the North Atlantic represents one of the worst possible routes and both
voluntary and involuntary speed losses and derived service margins for a vessel on this route
will normally exceed values derived from other routes, including a world-wide operating
scenario.

Constant speed operation is assumed in the present analysis. When using constant speed as
basis for the analyses a target schedule of port arrivals is derived based upon this target speed,
and the speed and power is adjusted continuously to give the correct arrival time. This method
of simulation reflects the normal mode of operation where speed and power is increased during
good weather parts of a passage if the vessel has previously been delayed by voluntary or
involuntary speed loss during bad weather. The practical limitations on installed engine power
are of course observed during such simulation.

The enclosed table presents results from the analysis of involuntary speed loss showing that the
vessel is not capable of maintaining the required schedule despite having an installed engine
power service margin of 15 percent. For the complete roundtrip voyage the average
involunatry speed loss is 0.9 knots and for the Westbound Trans-Atlantic voyage 1.7 knots.

Also enclosed are results from an anlysis of voluntary speed loss due to deck wetness using a 3
percent probability as the limiting criterion. For the Base Case the relative motions at the F.P.
is marginally below the critical value for voluntary speed reduction. In the subsequent cases.
the freeboard has been reduced, thus simulating an overestimate in the relative vertical motions
at the F.P. The results clearly show that a relatively large error in the seakeeping prediction
will have a small impact upon the total speed and power performance of the vessel.

More important to note is the fact that involuntary speed losses contribute far more
significantly to the total loss in performance compared with voluntary speed losses. On the
other routes that the North Atlantic this difference is even more significant.

It is clear from these results that an error of 10-20 percent in seakeeping prediction can be
accepted at the design stage when considering voluntary speed losses and the total t~chno-
economic performance of a proposed new design. The larger contribution by involuntary
speed losses to the total economic performance clearly demonstrates that it isthe accurate
prediction of added resistance which is more important in the design process.

A further comparison between speed losses due to environmental effects and speed losses due
to deteriorative effects was carried out for the same vessel. An increase in average hull
roughness from 125 pm to 300 pim in this case represents a further speed loss of 0.3 knots.
For a 5-6 year old well maintained vessel an average hull roughness of 300 pm would be a
representative number. Further, more dramatic speed losses would clearly take place if the
vessel also experienced hull fouling. A total speed loss in the range I - 1.5 knots due to
environmental and deteriorative effects would therefore be an expected value for this type of
vessel on the North Atlantic route. In economic terms a speed loss of O.5.knots when
translated directly into lost cargo carried represents a total economic loss.of approximately $ 2
million over one year. This figure can serve as a guide when considering ways of improving
the seakeeping and overall.techno-economic performance of a new design.

5. Application of seakeepinz theories in structural design

The conventional method of using seakeeping prediction in connection with structural design is
to perform an analysis of motions and global loads (vertical bending moment and torsional
moment). The vessel is subsequently "placed" on a regular wave giving the same global loads
and the pressures on the hull resulting from this wave are input to the finite element model.

More recent methods of analysis also permit taking hull pressures directly from the seakeeping
analysis in frequency domain and input to the finite element analysis. These are standard
methods and will not be dicussed fuirther here. Instead it is relevant to show how seakeeping
theories may be combined with modem probabilistic methods in the development of rational
methods for structural design

Probabilistic methods are used to determine the probability of failure for a combination of
several variables and events. Computational methods determining the probability of failure for
a combinations of several variables and events are today available through commercial
software. The software will schematically work as in the enclosed figure.

-Inorder to apply probabilistic methods in assessing the safety against loss due to lack of
stability, the following steps are required to be performed :
1) Choose physical model to describe the relevant failure modes (limit states).

2) Model the uncertainties in a consistent way by probability distributions.

3) Choose statistical distributions types and distribution parameters for all uncertain variables.

4) Integrate probability distributions with the physical model in a limit-state function.

It is seen from the above item list and scheme below that the physical model or limit state,
describing when a failure occurs, is the primary item. This numerical limit state model for use
in probabilistic methods may be both a purely experience based regression analysis including
the error information
or
it may be a physical model describing the dynamic behaviour of ship motions with inherent
uncertainties and model uncertainties. (Requires test results and analytical work).

Presently considerable research effort is spent in this area. However, to reach valuable results it
is not only the computational methods that needs to be improved, but also the data.collection
and data analysis, forming a basis for the probability distributions. Experienced personnel can
then join the physical models and decide on standard probability distributions for use in
reliability analyses.

The above model requires that hazards are identified and formulated in terms of limit state
functions. For different vessels there are different hazards or failure modes that need to be
covered.

Combination of load effects: Within each time scale there may be a set of load components.
For example, at the typical wave frequency it may be relevant to consider the horizontal and
vertical bending moment and the local sea pressure acting on a plate field. For a description of
the distributions of the combined load processes a formulation using crossing statistics is a
convenient representation. The upcrossing rate describes the number of crossings per unit time
of a specific level, and may be the basis for the required response distributions both for the
fatigue and extreme value calculations. For a combination of several simultaneous processes
including nonlinear combinations, the oulcrossing rate into the failure domain (g(Xt) 0)
contains the corresponding information (see enclosed illustration). For different types of
failure, or load combinations, the crossing rate gives the distribution of peaks that is used in the
fatigue calculations and the extreme value ditribution used in ultimate limit state calculations
(ULS). The method can be applied to non-Gaussian ( for example roll motion ) and non-
stationary processes ( as for example lifting operations). Non-linear load combinations may
also be considered, as for example buckling under biaxial loading or the Von-Mises stress
criterion. The exact solution may be approximated by simulation methods.

The numerical solution may be determined using a program such as PROBAN (se enclosed
illustration). The resulting outcrossing frequency is used to determine the probability of failure
during one specified storm, to be further processed when determining the annual probability of
failure in the ultimate limit state. The same out-crossing frequency is used in determining the
long term distribution of stress amplitudes for use in the fatigue limit state.

The process for a typical application will consist of the following steps:

- perform seakeeping calculations to obtain motions and loads transfer fuctions as well
as the covariance matrix for the intended load effects

- define limit state fuction

- define wave environment for the analysis

- perform analysis using a probability integral solver program, such as PROBAN

- perform analysis on the resulting outcrossing frequency for determining ULS or

fatigue compliance with requirements.

Examples of the most relevant application area for this type of analysis is fatigue in side
longitudinals of large tankers, buckling of bottom and deck panels or buckling of LNG
spherical tank shells.
2 E

<<)

= a
a)

0l) U
oInl
C11c >c
ID
EA

0)0
C>

0(

-
o0
a) a) a

(-a 0 M0
0J
won (Jj a
 S-% Cd
I:2 a

'.0 C.w CO
CDCD 0~ CD toa, 0

w~ wcwa
wcj2000 U,0: 0
Cd ) C) 14s-
-C C-, <0 co

4C2C~~4~vCC

D)5 ~ EEMEO
tw ~~ ~ C ~ C C)C'C)L
:,C)CD -

C0 C -) Ci

4, >6 -6 02 C- C )I t > -
o' bD02 bflbn!'
C) C "
.- .02

cU
w w m w~
0)~~1 Z )>oc i

C0
aa a)
w. bnb)
0 0 2h

a)S () ca co04a

a) )(1 00
(Da)C
co ca> , z c

C) 0

045-

0 NOJ"C

n0
 Limiting speeds in head waves due to
( deck wetness for 1700 TEU container vessel.
Maximum probability of deck wetness = 3%.

Ship speed
(knots)
22

20
18 -•

16

14-

12
( 101 1/I
3
6mrn 7m 8m 9n

86s *ITz
8.5s

Limiting speeds in head waves for

different limiting values of seakeeping criteria
1700 TEU container vessel.

Ship speed
(knots) Vert accel. 0.25g at FP
22-

20 - "

16

14-

12 -tCCO

10

8-

4 I H
6n 7 8r 9m
Tz
Os 8.5s
(,

ROUTE : NORTH ATLANTIC - ROUNTRIP DISTANCE 7500 n miles

SHIPTYPE : CONTAINER VESSEL- 1700TEU

PRINCIPAL DIMENSIONS :L = 198 m B = 32.2 m T = 10.5 m

OTHER PRINCIPAL DATA : CB= 0.577 Cw = 0.812

PITCH RAD. GYR. = 0.25 Lpp

PROPELLER DIAMETER = 7.1 m

TARGET SCHEDULED SPEED : 22 knots

CALCULATED VOLUNTARY SPEED LOSS DUE TO DECK WETNESS

FOR VARIATIONS IN ESTIMATED RELATIVE MOTIONS AT F.P.

WESTBOUND EASTBOUND
BASE CASE 0 0
1 m (13 %) OVERESTIMATE < 0.1 knots < 0.1 knots

1.5 m (20 %) OVERESTIMATE 0.2 knots 0.2 knots

2 mn (26 %) OVERESTIMATE 0.6 knots 0.5 knots

00)
C

ROUTE : NORTH ATLANTIC - ROUNTRIP DISTANCE 7500 n miles

SHIPTYPE : CONTAINER VESSEL -1700 TEU

PRINCIPAL DIMENSIONS :L = 198 m B = 32.2 m T = 10.5 m

OTHER PRINCIPAL DATA : CB= 0. 5 7 7 Cwp = 0.812

PITCH RAD. GYR. = 0.25 Lpp
PROPELLER DIAMETER = 7.1 m
TARGET SCHEDULED SPEED: 22 knots

INVOLUNTARY SPEED LOSS AND CALCULATED

ACHIEVED SPEED

WESTBOUND EASTBOUND
ACHIEVED SPEED 20.4 21.1

INVOLUNTARY SPEED LOSS 1.7 0.9

 C m PILOA N PROS

Basic Vanable Basic Variable Basic Variable Basic Vaiable

. - _ __

lbrary Limit States System COaniguration

gt___
g2() __ •PROBAN
_z) I- • _ __ __ __

gn (Z )

Failure Probabilities Sensitivity Measures

_______ * Importance
I Factors

-Sensitivity Factors
• /i pp

tOC •
,1 "The goal"

AS VERTAI ASEA*QI
 Clo
0

Ca
co>
0 U-
C.

0) _0
c U 0 co)

a, 0 0
Cu con- - > --. C --
CD
mv M~
oE z Dn 0

C) z

(n
c
(V

'C
Lii
 COMPUTING THE CROSSING RATE INTO FAILURE DOMAIN.

The out crossing rate

v, -F( d• Gde 0(o 4ý)a

ao ae

G is the limit state function.

G is the velocity of the processes into failure domain.

The limit state in this case :

G- 0, - - - 0(( -- 0"(

Where:

oR = Stress resistance level.

or,, = Still water induced stress.

a, = Vertical wave induced stress.

qH = Horisontal wave induced stress.

a,0O = Bottom slamming induced vertical stress.

o,,0 = Flare induced vertical stress.

Fail
Limit state GIX) =0 On the limit state and
positive velocity
~Safe

•Time
Events describing an out-crossing
 APPENDIX 1

GENERAL DESCRIPTION OF SEAWAY

 .... SEAWAY
A COMPUTER PROGRAM FOR
EVALUATING THE SEAKEEPING
PERFORMANCE OF A SHIP IN
A SEAWAY

A PRACTICAL TOOL FOR THE

SHIP DESIGNERS AND OPERATORS

• Added resistance in a seaway

.* Service margins

* Attainable ship speeds in waves

* Prognosis of ship responses

* Linking to ship weather routing

REPOR T
~REPORT
 SEAWAY

INTRODUCTION.

The economic performance of a vessel in service may be seriously affected

by in-service conditions due to environmental and deteriorative
factors. It
is often said that a new vessel on trials will never again achieve the same
level of speed/power performance in service. The economic consequences
of
performance losses in service can be significant and prior knowledge
may
influence design variables such as bow geometry and installed main
engine
power.

In service, weather routing can be an

efficient method of minimizing the
risk of heavy weather damage and finding the most economic passage
between
two ports.

PRINCIPLES.

The loss of performance in service can principally be attributed to

voluntary and involuntary speed losses due to environmental effects

and
. deterirorative effects such as hull and propeller roughness and fouling.

In SEAWAY a complete operational model of the vessel operation is employed

to determine the vessel speed and power performance in service. By
modell-
ing the various factors in combination rather than on an individual
basis
more realistic results are obtained from the analysis.

SEAWAY calculates the speed losses caused by:

* Wave induced motions

* Wave reflections
* Wind resistance
 2

* Ocean currents
* Reduced propulsion efficiency in waves
* Additional resistance due to steering and yawing motions
* Optionally: Deteriorative effects

Involuntary speed losses are principally due to wind and waves. In the ship
motion domain the wave added resistance is obtained directly from the ship
motions predidtion program.

Asymptotic theori is used to calculate the added resistance due to short

waves outside the ship motions domain;

In the case of wind resistance and added resistance due to rudder and yaw-
ings motions a semi-empirical calculation method is employed in SEAWAY.

Voluntary speed reduction is a deliberate reduction of speed and alteration

of course in order 'to- prevent damage to the vessel or its cargo. In SEAWAY
several criteria for voluntary speed reduction are included and the magni-
tude of each may be defined by the user.

Propulsion system. Ship motions and added resistance also result in a de-
terioration in the propulsive efficiency. SEAWAY includes a semi-empirical
model for the loss of propeller thrust in a seaway with full scale simula-
tion of RPM as well as power and torque regulation within a predefined
operational envelope for the main engine.

INPUT DATA.

SEAWAY requires a set of input data including

" ship. propeller and engine characteristics

* environmental data
- motion transfer functions and added resistance in regular
waves

A purpose written interactive program is available to assist with data

input.
 3

Environmental.

The environmental are described in terms of

* Wind velocity and direction

" Wave height and direction
- Swell height and direction
* Current velocity and direction

Long-term statistical data for the North Pacific and for the North Atlantic
Oceans have been included in the program to facilitate rapid prediction of
long term vessel performance. For the other sea areas the necessary data
have to be supplied prior to the program run.

Vessel, propeller, engine.

The main data describjng the vessel and her propulsion

system are:

* Geometrical data
* Ship still water resistance
* Propulsion factors
- Open water characteristics of the propeller
* Engine parameters

If some of these characteristics are unknown (e.g. at the preliminary de-

sign of a ship) the program will calculate a set of estimated values by the
use of an approximate method.

Motion transfer functions and added resIstance

In regular waves.

The program requires these characteristics as

input data. They may be:

Calculated by used the advance strip theory

(the SEAWAY program may be
linked to such pre-processing programs)

Results of seakeeping model tests

 4

* Estimates obtained by scaling of data for a similar ship form.

OUTPUT.

Output from SEAWAY includes

added resistance due to waves, wind and current

power setting
actual ship speed in a seaway (short- and long-term predictions)
probability distributions of ship responses such as: vertical
and
horizontal velocities and accelerations, rolling, slamming, deck
wetness etc.
ranges of ship speed and course to be avoided due to
exceedance of
seakeeping interia (for the given weather conditions)
optimum ship route

Graphical and tabular output options are available.

 5

OPTWAY

The programme OPTWAY has been developed as a post-processor to SEAWAY.

OPTWAY calculates the optimum route of the ship
in a seaway where the dif-
ferent criteria for optimization may be defined by
the user. The programme
is interactive with two principal calculation options:

* optimization of route between two points

* performance evaiuation along Great Circle route between
two points

In both options a number of intermediate way-points may be introduced. This

will force the program to draw the optimum route through
these points.

Data input requirements:

attainable vessel speeds -under different environmental

conditions,
provided by SEAWAY,
departure and arrival points, region of ship operation
and number
of alternative solutions in the optimization procedure
weather forecasts for the relevant operational area
at fixed time
steps and for several days ahead

Output includes:

distance, total time, average speed, fuel consumption for

the
optimum and any alternative routes, such as Great Circle
route.
graphical plot of optimum and alternative
routes.

The North Atlantic and the Nort Pacific Ocean are included as default
regions of ship operation in OPTWAY. Other geographical
areas will be
included upon customer request. OPTWAY may be used
as a stand-alone program
with pre-prepared data files for individual ship types
or as a direct post-
processing program to SEAWAYS.
 6

,REFERENCES.

1. Faltinsen, 0. M., Minsaas, K. J.: Added Resistance in Waves. Centenary

Conference on Marine Propulsion, Newcastle upon Tyne, 15-17th
May,
1984. pp 8/1-8/24.

2. Minsaas, K.J.: Ship Motion in a Seaway. Development of the SEAWAY

com-
puter program for the Calculation of Added Resistance and
Velocity
Loss. Symposium on the Energy Safe in Ship Exploitation, Trondheim.

3. Kauczynski, W.: SEAWAY - computer program (involved theory and user's

manual), MARINTEK report no. *'''I/89. Trondheim 1989.

WK/TES/il0/rapp16i9 019
. 02
 test
E characteristics methods -- RAW
Propelleob
•r; added
resistance

char~acteriskis

Velocity

Engine ,.
characteristics "
f • Engine/
Weathr
/ dat ipropeller
Weathe datacharacteristics
(waves, wind I HP, Q. RPM
and current)

Criteria for

voluntary _.
speed loss

Numerical calculation of performance in a seaway.

 Container Carrier
Velocity loss for different sea states

Head sea
Design
16- . speed H, = 8.25 m

14 :/(Sea state 7)

_0 12
(D 10"
CD

Actual
6
o speed

ý4

4 6 8 1 21 16 Beam sea

:2r 6 -- I
C')

"0 8 _.Slamming .o

CD
oL 1 0 •_...-- •
.
Water on
12 the deck

14
fl lRoll
16 motion

Following
sea

Calculated speed loss for a container ship at a given sea state in the North Pacific. The ship has the
following parameters C :- 0.82. T - 10.73 in. B - 32.2 m. L = 185 m. The actual speed fot
constant power is given together with rangqes where linitts fot acceptable acceleration, roll and
water on deck are ex(:eeed.
 APPENDIX 2

DESCRIPTION OF METHODS USED IN

SEAWAY
 1. COMPUTATION SCHEME

The main purpose of the programme is to make an

estimation of the resulting speed in a seaway,
where a sea
state is defined by the given significant wave
mean period. By summarizing the results for all height and
sea states
with the probabilities for a particular sea state as
weighting coefficients, a long term prognosis is
obtained.
In order to evaluate the added resistance and the
propeller characteristics in irregular sea the following
approach has been apllied: It was assumed that the waves
are consistent with a narrow banded process and that
they
are longcrested. For a given sea state the wave spectrum
was cut into succesive parts. Furthermore, it was assumed
that each such element of wave spectrum consists
waves set, where each regular wave has the definedof regular
amplitude
"a" and the circular frequency " '. Treating them as a
stationary process the probability density function
component wave can be written according to Sveshnikov for a
ill:
a2
a2
2- (W22 -2wi 1 + w2
2a2(W 2
f(a,w) = 3 a 2 2 2 2
2- i
where "a" varies from 0 to co , and "UY" varies from
-oo
to +oo. Physically a negative "Wo" should be interpreted
the same way as a positive one. in
The circular frequencies
"WI'" and "t 2 " are'defined by the first and second moment
ot the wave spectrum. Finally, 6 is the area under the
wave spectrum related to the significant wave heigth
H as:
2= HA/16

Due to such assumptions a calculation for an irregular

sea case is turned into a set of calculations
waves (in the programme each element of the wave for regular
spectrum is
partitioned on the 400 component (regular) waves).
regular wave with For a
amplitude and circular frequency
determined as above, the added resistance, the relative
motion amplitude at the propeller plane and at the
ship acceleration and vertical velocity at the FP, the
bow and at
the midship - are found by using the transfer
They must be earlier determined in a numerical way functions.
aid of the advanced strip theory) or derived from (e.g. by
the model
tests.
Additional effects which change the ship speed as:
resistance due to wave reflection, the wind resistance, the
and
the current influence are also included in the
calculation
procedure.
The changes of the propulsive factors in waves are
made
accordinc to the proposition given in (1).
 The thrust obtained from the propeller
reduced due will also be
to ship motions in a seaway as the
works in waves at different kinetic propeller
conditions (generally,
at the worse ones) and because
occur. A numerical model for the propeller ventilation may
thrust in a regular wave has been calculation of available
derived and introduced in
the programme.
As main propulsion machinery, the
adopted. diesel engine was
For such an engine typical limitations
the revolutions and the torque were regarding
added to the programme.
For each component wave a balance
between the available
thrust (reduced by the thrust factor)
resistance is found and the resulting and the total
torque, ship speed, machinery
and engine revolutions are
summarizing over all the regular calculated. By
waves and over all the
elements of the wave spectrum the
mean values for irregular
sea are obtained.
The spect'ra for ship responses
vertical velocities are calculated. as accelerations and
determine the probabilities of cases They are used to
(component waves) where
a particular response exceeds the
results for all the wave components critical value. When the
comparison with the limits described are summarized, a
in the given criteria
for voluntary speed reduction is made.
The speed loss-due to use of rudder
the is not included in
programme. Similarly, the speed loss due to
motions is not regarded. the yaw
Some more details about the calculation procedure are
given in'the following subsections.

1.1 Transfer functions

The transfer fuctions for the ship

motions (heave, roll,
and pitch) and for the added resistance
The subroutine TRANS read this data are given as data.
spline coefficients, which allow table and calculates the
to determine the transfer
functions for an arbitrary wave
heading. The asymptotic
values for the very short and very
long waves are added.

1.2 Added resistance due to wave reflection

For short waves (wave length/ship

length ratio WL/L <
0.5) the added resistance due to the
reflection of waves may
be significant and therefore can not
be omitted. A simple
formula for the calculation of
proposed by such resistance has been
Faltinsen et al [5]. In the programme a
non-shadow part of the waterline
is found, which is devided
into a set of straigth-line elements.
For each such element
a force caused by the wave reflection
is calculated by aid
of the equation mentioned above. The total
resistance is obtained through a summarizing additional
waterline length. The subroutine over the
RADIFF calculates this
resistance for three WL/L values (0.1,
0.35, and 0.6) and
for three wave headings (0, 30, and 90 deg).
For the other
values an linear interpolation is used.

1.3 Wind resistance

The wind effect is regarded as a constant

resistance
force.
The wind resistance coefficient Cxo can
data or alternatively is assumed to be equalbe given as a
to 0.7 for the
case when wind comes from the direction
opposite to the
ship's forward movement. For other directions a linear
dependency with the angle of wind direction
was adopted:
Cx (90 -Or) c0
where: CC,- wind direction angle (see definition
given in Fig. 1)

The wind resistance is -calculated according to the

equation:
Rwind = 0.5q.:-.a Awind<. Vww' -
Ox

where: 9a - air density

Awind - ship cross sectional area above water
surface
Vww - resulting wind velocity in relation to
a ship (Vww = Vship- cosor + Vwind)
Vwind - wind velocity

The wind velocity and direction can be given as input

data. There is also an other option: the wind velocity can
be calculated as a function of the significant
wave height
from the two alternative equations:

Vwind = -8.0 + 5.614 (HY3 f'9 [m/sec]

Vwind 6.88 (,•,ýj )" Im/sec]

where: IlA,• - significant wave height [m)
 In Fig. 2 these two relations are shown.
If the wind velocity is calculated by the equations
given above, it is assumed that
the wind direction is the
same as the direction of the incoming
waves.

1.4 Current effect

A current changes the speed of a

ship in reference to
the water, since a correction of
the
The corrected speed is used to calculateship's speed is made.
(in the
the propeller characteristics estimation) advance ratio J
evaluation of the encounter frequency and for an
spectra calculation). When routing (in the response
(see section'7) the ship position of a ship is considered
after each time step is
corrected according to the current
velocity and direction.

1.5 Change of propeller characteristics

in a seaway

An example of propeller characteristica

still water and in waves is shown (KT-curve) in
decrease of the KT coefficient in Fig. 3. A significant
ventilation occurs. is observed if propeller
On basis of test results with
propellers in waves given the
in (6), a criterium for the
propeller ventilation occurence was
Such parameters as: the propeller numerically formulated.
initial immersion, the
propeller pitch, number of. revolutions,
the wave amplitude were the wave period, and
included in this formula.
Similarly, a decrease of the
(without ventilation) was expressedpropeller thrust in waves
numerically by using the
same parameters as these mentioned
above.

1.6 Wake factor

The typical changes of the propulsion

waves are given in Fig. coefficients in
4. The ship motions change the
pressure distribution on the hull
and has an influence on
the boundary layer thickness,
which
reduction of the wake factor in waves. will result in a
these changes of the wake factor In order to find
placed at the propeller plane is the model test with a ring
recommended [7]. As such
tests are not common yet, a change
of the wake fraction can
be estimated on a basis of the previous
[71 an aproximation equation was adoted test results. From
in the programme:
w = a - aL + b S.t'
 where: w - wake factor
a and b - coefficients found from the tests
(7]
(a = 0.579, b = 36.103)
We- encounter frequency
Sa - relative motion double amplitude
Lwl - waterline length

1.7 Thrust deduction factor

The reduction of the thrust deduction factor

often observed. in waves is
For moderate sea states these changes are
small, but if the pitch motion is close
significant ',reduction is obtained. to a maximum a
The reduction of the
thrust factor is mainly caused by changes
load. of the propeller
An equation for the thrust deduction estimation
proposed by Nowacki and Sharma [8) was applied in the
programme:

where: t - thrust deduction factor

wp - potential wake factor
w - wake factor
CT - propeller lpad coefficient

The potential wake factor is estimated

speed of the at the nominal
ship using given input data for the wake
thrust coefficients and assuming that and
this value remains
unchanged in waves.
The relative rotative efficiency will also
change in a
seaway. But since no theoretical basis for this phenomenon
is known, the still water value was also used in
conditions. severe

1.8 Criteria for voluntary speed reduction

The ship responses may reach such values

ship operations can not be continued. that different
In order to decrease
the responses the ship speed may be voluntary
course change can also be considered). reduced (a
A number of criteria
for the voluntary speed reduction were proposed.
them were Some of
presented at the last ITTC meeting 19).
choise of the criteria for a given The
ship is a matter of
opinion and must therefore be carefully done
with reference
to the kind of mission which a ship should
fulfill. In the
programme a set of criteria were introduced.
This list can
be easily extended if a special requirement for
a ship is
defined. Presently, the following criteria are included
in
the programme:
a/ deck wetness should not be longer than 7 %
of time
b/ slamming can not occur more than 3 % of time
c/ roll amplitude larger that 25 deg can not
be longer
than 0.1 % of time
d/ vertical relative velocity at the bow larger
than
0.093 g L can not occur more than 3 % of time
(at ballast load)
e/ vertical bow acceleration larger than 0.4
g can
not happen more than 7 % of time (at full load)
f/ vertical bow acceleration larger than 0.49
g can
not occur more than 3 % of time (at ballast load)
g/ significant vertical acceleration at the midship
can not be larger than 0.2 g
h/ vertical bow acceleration (root mean-square
value)
should not be larger than 0.15 g (fishing boats)
i/ vertical acceleration at the midship (root
mean-
square value) should not be larger than. 0.15
g
(fishing boats)
j/ vertical acceleration at AP (root mean-square
va-
lue) should not be larger than 0.2 g (fishing
boats)
k/ roll amplitude (root mean-square value) should
not exceed 4 or 6 (an option) deg - for fishing
boats
In the criteria listed above the following
responses were applied: ship
th'e motion amplitude at the bow,
the amlitude of roll motion, and the vertical
acceleration at the midship and at the bow. velocity and
The spectra for
these responses in the frequence domain
Assuming that the distribution of the ship are determined.
resonses
treated as the Rayleigh distribution, the probability can be
certain oscillatory response "s" exceeds a that a
constant level
"sl" can be giv'en as follows [10]:

P(s>sl) = exp [-slZ/2 moJ

where: min, variance of a certain response "s"

The probabilities of the vertical velocities

accelerations exceeding a given level are easily and
from this equation. For deck wetness and determined
slamming some
additional definitions must be added.
The deck wetness occurs if the vertical
amplitude at the bow is larger than the effective relative
at the bow. The effictive freebord can be estimated freebord
equation [10j, where a correction due to presence by an
of the bow
wave is made:

Fe = F - 0. 75 B. -Fn
 where: Fe - effective freebord [m)
F - geometrical bow freebord [m]
B - ship breadth [m]
L - ship length [m)
Le - length of entrance of -the waterplane [m]
Fn - Froude number [-]

Additionally, a factor increasing the relative motion

amplitude at the bow is introduced (swell effect at the bow)
[10]:

Sa = Sa (1 + [Cb - 0.45)- -7A we)]

where: Cb - block coefficient
L , - ship length
we - frequency of encounter
g - gravitational acceleration

The slamming phenomenon is defined as an event occuring

when the bow relative vertical velocity is larger than the
critical one simultaneously with the relative motion
amplitude at the bow being larger than the ship draught at
this position. The critical velocity is assumed as:

Vcrit = 0.093 gAL

where: L and g are defined above

Then a jointed probability is given by:

P(slamming) = exp[-(T7 /2 m 0 s)-(Vcrit 2 /2 m ý)]

0

where: T - ship draugth at the bow

mos - area under the spectrum of the vertical
relative motion at the bow
mo£- area under the spectrum of the relative
vertical velocity of the bow

In the calculation for criteria used in the programme it

was assumed that the time of each event is the same (it
means e.g. that the time of each slamming is the same
undependently on the wave period). This makes a significant
simplification of the numerical procedure possible, but this
is not correct. Therefore it would be better to re-define
the criteria in an other way (as this is made by some
authors [9J) and instead of the limitation aiven as a
procentage of time, the defined probability should be
applied.
 1.9 machinery characteristica

As ship propulsion machinery a diesel

adopted. engine was
Fig. 5 shows a typical example of the diesel
engine characteristics, where a region
parameters of admissible engine
is indicated. The limitations defined as the
engine revolutions not higher than nominal
measured on the shaft not higher than and the torque
the continuous work) has been introduced the nominal one (for
in the programme.

1.10 Determination of resulting speed

In order to calculate the ship speed in a particular

regular wave, different sets of
propeller revolutions are assumed. the ship speeds and
For all these values the
torque of the propeller is calculated (see Fig.
comparison with the 6). By a
torque produced by the engine,
points of equilibrium are found (Fig. the
6). For each point a
pair of values - the ship speed and
revolutions - are described. the propeller
Then a comparison between the
total ship resistance and the available-thrust
the propeller produced by
(reduced by the thrust factor) for
ship speed from a given set the each
is made. The point of
equilibrium gives the resulting speed
corresponding propeller revolutions. (Fig. 7) and the
The torque obtained on
the shaft is checked and if it is higher
new pdint of equilibrium 1s: determined than nominal then a
propeller with the reduced
revolC~ions. Similarly, if the
propeller revolutions are higher than resulting
the nominal ones, a
new iteration with the reduced engine
power is made.
 REFERENCES

i. FALTINSEN, O.M., MINSAAS, K.J.: Added Resistance in

Waves. Centenary Conference on Marine Propulsion,
New-
castle upon Tyne, 15-17th May, 1984, pp.
8/1-8/24.
2. SORTLAND, B. : ADREWA - Added Resistance
and Propulsive
Performance in Waves and Wind (Computer Program
Documen-
tation). Report of the Norwegian Institute of Technolo-
gy no. UP-84-3, Trondheim, January 1984, in Norwegian.
3. KAUCZYNSKI, W.: "SEAWAY" - Computer Programme.
Examples of Ship Ped Loss Prediction in Variable Some
Wea-
ther. Symposium on the Energy Safe in Ship Exploita-
tion, Trondheim, 5-6th Nov., 1985, pp. 6/1-6/29.
4. MINSAAS, K.J.: Ship Motion in a Seaway. Development of
the SEAWAY computer programme for Calculation
of Added
Resistance and Velocity Loss. Symposium on the Energy
Savi^ in Ship Exploitation, Trondheim, 5-6th Nov.,
pp. 1985,
5/1-5/22, in Norwegian.
FALTINSEN, O.M., et al.: Prediction of Resistance and
Propulsion of a Ship in a Seaway. Thirteenth Symposium
on Naval Hydrodynamics, Tokyo, 1980, pp.
505-529.
6. FLEISCHNER, K.P.: Untersuchungen uber das Zusammenwir-
ken von Schiff und Propeller bei Teilgetauchten
Pro-
pellern, Report no. 35/1975 of Forschungszentrum des
Deutschen Schiffbaus, Hamburg, 1973.
MINSAAS, K.J.: "ADREWA" -'Verification Tests (Perfor-
mance in Waves), Part I. Report of the Norwegian Hydro-
dynamic Laboratories. Project no. 5150025. Trondheim,
January 1985. '"
8. NOWACKI, H., SHARMA, S.D.: Free Surface Effects in Hull
Propeller Interaction. Proceedings of the
9th ONR,
Paris, 1972.
9. Report of the Seakeepeng Committee. 18th Conferance of
ITTC, 1987.
10. FALTINSEN, O.M.: Sea Loads and Motions of Ship. Summa-
ry of Student's Lectures (subject no. 81527). The Nor-
wegian Institute of Technology, Trondheim,
January
1980, in Norwegian.
11. TAKAISHI, Y. et al.: Winds and Waves of the North Pa-
cific Ocean 1964 - 1973, Statistical Diagrams
and Tab-
les. Ship Research Institute, Tokyo, March 1980.
 WIN

VWhtk11W

C-OOR-IAýkTC Sysýet hub 51GW CCN'Vc7,"7tot-J

 v in 8 5.614 /

20,

1 2 4 6

ýAi [Kv-
Gr~vt&5aCe keitkV
Inf luence of Immersion and Ventilation on the open Water Diagram

from Fleischer [6]

,4 8

00m3

SQF
AJA WbanA-C t

'10SL
P -d
 710-

100 - Nominal Power

90-
CC
LU

~70 e ".: :::. .

. .::
....

/ce

75 B85 90 .. 95 0 /...8

% ENGINE SPEED

A :Optimum range for continuous operotion.

A 1 : Range for engte characteristic mn sea trio' wibh fair weather,
•ship filly loden arnd clean hull.
B :Working
60 A LSP range for restricted
Q'c time
Iuc.only (max. 2000.......
hocrs .
C :Upper speed rangye for sea trial condy.
p Engine choracteris tic on shop tricd,ie., opproxirnoted propeller
curve Through the point of M. C.R.Although the engine is
capable of working in range B for restricted time, the
aim should be to design the propeller in such a way
that curve p is not exceeded in continuous service.
t L imitation of range B.
MOP:R Maximum Continuous Rating, i. Nomninai Power (100% Pe)
at Nroni al Engihe Speed (100% ni
pe : Broke Mear Effective PressureO (-torque)
 PIPS QF. '-"54~

VI7~;is
/ i-M
q 7 /

-T / It

to.
 SEAKEEPING OF HIGH SPEED VESSELS

0. Faltinsen
Department of Marine Hydrodynamics
Norwegian Institute of Technology
N-7034 Trondheim - NTH, Norway

To be presented at UETP course on "New Techniques for Assessing and

Quantifying Vessel Stability and Seakeeping Qualities", Trondheim, March 1993.

INTRODUCTION

Catamarans and SES are well known to the high-speed community, while
foilcatamarans are new concepts. The foils of the two Norwegian concepts are
designed so that the catamaran hulls are out of the water at high speed in both
calm water and small sea states. A consequence of this is that a foilcatamaran has
a much lower resistance than a similar sized catamaran at the same speed. When
the hulls of the foilcatamaran are out of the water, the vessel can easily roll over
to one side. The reason is a small restoring roll moment. To counteract this
undesired behaviour a ride control system is used. The control system is also used
to keep a nearly constant vertical position of the center of gravity and to control
the trim angle of the vessel.

In the following text we will--discuss -the -seakeeping behaviour, of-catamarans,

foilcatamarans and SES. Some of the discussion will be based on numerical
results.

SEAKEEPING CHARACTERISTICS

A catamaran, foilcatamaran and a SES have quite different seakeeping behaviour.

This can partly by exemplified by Fig. 1, which shows operational limits based on
numerical calculations of vertical accelerations at the centre of gravity (CG) of a
40 m long catamaran without foil, a 40 m long SES without ride control and a 36
m long foilcatamaran without ride control. A ride control will have a significant
positive influence on the seakeeping behaviour of a foil catamaran. The same is
true for a SES in low sea states. A RMS-value of 0.2 g is used as a criterion in this
example. The vertical accelerations will of course depend on the details of the hull
and foil design. The intention with Fig. 1 is to illustrate features of the
acceleration level at CG of the three different vessel types in head seas. The
results in Fig. 1 are for longcrested waves described by the two-parameter
JONSWAP spectrum recommended by IThC. T1 is the mean wave period defined
by the first moment of the wave spectrum and HI 3 is the significant wave height.
The vessel speed in calm water is 50 knots for the SES, 40 knots for the
catamaran and 50 knots for the foilcatamaran. The SES will have the highest
involuntary speed loss in a seaway. Involuntary speed loss due to waves and wind
is accounted for in the calculations for the SES and the catamaran (Faltinsen et
al. (1991)).
Fig. 1 illustrates that the SES has the lowest and the foilcatamaran has the
second lowest vertical acceleration level when T1 > -6 s. The results for the
foilcatamaran suggest that foil appendages on the catamaran may improve the
seakeeping qualities of a catamaran. However care must be shown in doing that.
An example will illustrate that foil appendages may not improve the seakeeping
qualities. Three hulls were numerically investigated. Hull 1 was a basis hull
without foils. Hull 2 and 3 were modifications of Hull 1. The displacement was
reduced respectively by 10% and 20% relative to Hull 1. This was done by
reducing the displacement volume in the aft end of the ship. The hulls had
transom stems and the reduction in the displaced volume resulted in a decrease
in the local beam at the transom stemn. The reduction in displacements of Hull 2
and 3 were compensated by lift from foils at the aft end of the hulls. The net
result of this was that the vertical accelerations of the three hulls did not differ
very much. One reason is probably that the increase in damping by the foils was
compensated by a decrease in the damping due to the hulls. The latter is due to
both wave radiation damping and dynamic lifting effects. The magnitude of the lift
force and moment on the hull depends on the local beam at the transom stem.
However adding passive foils to a hull without changing the hull is likely to lower
the vertical accelerations. It has probably most effect if the foils are placed in the
bow part were the relative vertical motions are largest. However, out of water
effects of the foils or cavitation will degrade the damping effect of the foils.

Fig. 1 shows that the operational limits for small wave periods are clearly lowest
for the SES. The reason is the cobblestone effect, which is due to resonances
occurring in the air cushion. No ride control was accounted for in the calculations..
Use of ride control will increase the operational limits for the SES at lower wave
periods. Full scale measurements have shown that there are two resonance
frequencies that are important. One is around 2 Hz for a 35 m long vessel. This
can be analysed by Kaplan et al.'s (1981) procedure. Details may also be found in
Faltinsen (1990). According to their theoretical model the dynamic part of the
excess pressure in the cushion is oscillating with the same amplitude all over the
cushion. It is caused by compressibility effects of the air and excited because the
waves change the air cushion volume. In the past most attention has been focused
on this resonance phenomena and ride control systems have been designed to
increase the damping of those resonance oscillations. However acoustic resonance
can be just as important. This was studied by Sorensen et al. (1992) for a SES
with a rigid panel as a seal in the aft part of the cushion. They showed that a
standing one-dimensional longitudinally varying acoustic pressure system with
nodes midships was excited. The natural frequency is 6 Hz for a 28 m long
cushion. The acoustic pressure distribution causes a pitch moment and pitch
acceleration of the vessel. If the vessel has a bag as an aft seal and there is an air
connection between the bag and the air cushion, full scale measurements show
that the acoustic resonance frequency is lower relative to a vessel with a rigid
panel as an aft seal. This has been theoretically studied by Steen (1993). Both the
uniform pressure resonance and the acoustic resonance cannot be investigated in
model scale. Thle reason is that thc natural periods scale like LM/L where [LM is
2
model length and 1, is full scale length, and that wave periods scale like (LMIL)"1
in model sale. When full scale measurements of cobblestone effiects are performed,
it is necessary with accurate measurements of the wave environment. It is not
 sufficient with visual observations. The shape of the wave spectrum, the peak
period, the significant wave height, the mean wave direction and the directional
spreading of the waves will influence the results. Since the wave periods of
interest are quite low (typically wave peak periods between 1 and 2 s)
conventional wave buoys cannot be used.

When characterizing the seakeeping behaviour of a vessel, it is of course not

sufficient with figures like Fig. 1. Different wave headings have to be studied.
Other important quantities are vertical accelerations along the length of the ship,
rolling, relative vertical motions and velocities between the ship and the waves,
and the influence of other wave directions. The criteria for operational limits are
also important.

Operational limits are set by

- safety, comfort and workability criteria

- structural loading and response
- machinery and propulsion loading and response

Seakeeping criteria for ships at moderate speed have been discussed by the
Seakeeping Committee of the YITTC, see for instance the report of the 18th and
19th ITTG. Those criteria are normally related to slamming, deck wetness, roll
rms-values and nns-values for vertical accelerations. They can be used to
determine voluntary speed loss and operability of vessels in different sea areas.
Faltinsen & Svensen (1990) have pointed out the relative large -variation -in
published criteria. This may lead to quite different predictions of voluntary speed
reduction. For high-speed vessels other criteria are also needed. One example is
operational limits due to the propulsion and engine system in a seaway. Meek-
Hansen (1990, 1991) presented service experience with a 37 m long SES equipped
with diesel engines and water jet propulsion. An example with significant wave
height around 2 m, head sea, 35 knots speed showed significant engine load
fluctuations at intervals of 3 to 5 seconds. These fluctuations result in increased
thermal loads in a certain time period, caused by a very high fuel/air ratio. These
high thermal loads combined with high-rated engines and reduced engine
J condition between major overhauls may lead to engine breakdowns.

Possible reasons to the engine loads fluctuations are believed to be

A. Exposure of the water jet inlet to the free air.

B. Flow separation in front of and inside the inlet.
C. Ventilation and penetration of air from the free surface or from
entrained air in the boundary layer.

The phenomenon mentioned above are often coupled in a complicated way. As an

example separation may be one of the requirements for onset of ventilation.
Cavitation occurs in connection with separation. Under given conditions a cavity
will be penetrated and filled with air. Separation and cavitation are first of all
depending on the pressure distribution in and near the waterjet inlet. For a given
shape this distribution depends mainly on speed and thrust (resist~ance) of the
ship.

Item A, i.e. exposure to free air, is a result of the relative vertical motions of
the craft. An operational limit will be related to the probability of the relative
vertical motions amplitude between the vessel and the waves at the waterjet is
beyond a certain limit. Faltinsen et al. (1991) used a rather strict criterion, i.e.
that the mean submergence d of the inlet relative to the local steady free surface
should be at least 4 ( R, where 3R is the standard deviation of the relative vertical
motion at the waterjet inlet. For a catamaran this did not represent a major
problem. However for a 40 m SES equipped with flush inlet it represented a
problem for all sea states with significant wave heights of 1 m and higher. The
problem can be solved by using a scoop inlet. The consequence is increased power
or drag. For instance if the immersion of the inlet is increased by 1.25 m compared
to a flush inlet, it means the order of 10% increased power or drag. Other
possibilities are to drop the pressure in the air cushion or to change the trim. The
penalty is then also increased drag.

Mishima (1992) presented experimental results of airdrawing of waterjets in

waves. Underwater video camera was used to study the flow. Mishima suggests
that the allowable frequency of air drawing should be less than one minute in full
scale. This is a less strict criterion than Faltinsen et al. (1991) used. Mishima
points out that the allowable frequency for gas turbines should be chosen very low
because gas turbines are very sensitive to overspeed due to air drawing. More
studies are needed to formulate a limiting criterion due to waterjet air ingestion.
This implies both better understanding of the physics and that the performance
of ships at sea should be monitored.

An important limiting criterion for voluntary speed reduction for high-speed

vessel is associated with slamming loads on the wet deck and the hulls. One
cannot simply use the slamming criterion for ship hulls operating at moderate
speed. The seakeeping committee of the 19th ITTC is also questioning this
criterion for conventional ships. This criterion does not distinguish sufficiently
between different hull forms. Many high-speed vessels have very slender
forebodies. Applying the conventional criterion of slamming in terms of threshold
velocity Vcr for slender hulls could mean that slamming was predicted to be a
problem, while it was not in reality. For wetdeck slamming it may be more
appropriate to directly use the criterion associated with slamming loads on ships
operating at moderate speed.

A special "deck wetness" problem that occurs for high speed vessels are "deck
diving". This can occur for a catamaran and a foil catamaran in following sea.
(Jullumstro (1990)). A dangerous situation in regular waves is when the ship
speed is close to the wave speed. One of the reasons why "deck diving" can occur
is that high-speed vessel may have a slender forebody, i.e. not enough buoyancy
in the forepart to avoid "deck diving".

The seasickness criteria according to ISO 2631/3 seems to be common to use

for assessment of passenger comfort. It gives limits of RMS values of the
accelerations as a function of frequency (see Fig. 2). Fig. 2 needs some
explanations. It is referred to the az-component of the acceleration. This refers to
a coordinate-system having its origin in the heart of a man. The az-component is
along a direction in the foot-(or buttocks) to-head axis. For a broad band spectrum
the frequency in Fig. 2 means the average frequency of a 1/3 octave band. A 1/3
octave band is defined as follows. Consider f, and f2 to be the lower and upper
frequency of the 1/3 octave band, then f 2=23f1 . The centre frequency of the 1/3
octave band is (flf) 2 . This means fl=fc/216 and f2=fc2 " 6 . For a broad band
spectrum the spectrum should be divided into 1/3 octave bands. The RMS-value
should be evaluated separately for each 1/3 octave band. The RMS-value should
be compared with the limits given in Fig. 2 for different exposure time. Since the
motion sickness region in Fig. 2 is from 0.1 to 0.63 Hz, it implies that the
cobblestone effect of a SES does not cause motion sickness. In the frequency range
1 to 80 Hz there are other criteria according to ISO 263111. These are related to
workability (fatigue). An example is shown in Fig. 3. The figure expresses the
limits of the RMS value of the az-component of the acceleration as a function of
frequency. Fig. 3 should be interpreted in the same way as Fig. 2. By multiplying
the acceleration values in Fig. 3 by 2, one get boundaries related to health and
safety and by dividing the acceleration values in Fig. 3 by 3.15 one get boundaries
for reduced comfort.

INVOLUNTARY SPEED LOSS IN WAVES

Faltinsen et al. (1991) have presented a theoretical method to predict added

resistance in waves of high-speed mono- and multihulls. It is partly based on a
direct pressure integration method using expressions from a linear unsteady flow
analysis in regular waves. The problem is solved to second order in wave
amplitude. The regular wave expressions can be combined with a sea spectrum in
the normal way to obtain mean wave forces or added resistance in a sea state (see
for instance Faltinsen (1990)). Transom stern effects are included in the ex-
pressions and are of importance. The interaction with the local steady flow is
accounted for in an approximate way. It was demonstrated by Faltinsen et al.
(1991) that the latter effect is important for hulls with non-vertical sides at the
waterline in the bow region. Comparisons with model tests show in general
satisfactory results. The interaction with the local steady flow was evaluated by
a quasi-steady approach where the steady longitudinal force on the vessel was
calculated in different oscillatory positions of the ship. The expressions were then
time-averaged. The difficulties in consistently handling the interaction between
the local steady flow and the unsteady flow, imply that one should investigate the
possibility of using a time-domain solution. An obvious drawback will be the
required CPU-time relative to a frequency domain solution.

The air leakage from the cushion in waves has an important effect on the
added resistance of a SES in waves. The air leakage causes the SES to sink and
the still water resistance components to change. For instance the altered excess
pressure in the cushion changes the wave resistance due to the air cushion.
Further the increased wetted surface area of the hulls changes the frictional and
wavemaking resistance due to the hulls. In addition there is a contribution to the
added resistance in a similar way as described previously for mono- and
multihulls. '[his is due to second order non-linear interaction between the dynamic
vessel oscillations and the incident waves. The air resistance on a SES due to wind
and the vessel's own speed is also important. This is not so much the case for a
catamaran. Reasons for this are the presence of the skirt on a SES and a lower
hull resistance on a SES relative to a catamaran. A method to predict the added
resistance in waves of a SES is presented by Faltinsen et al. (1991). Th~is is based
on finding the mean air leakage in waves. The expected value for the drop in
pressure in the cushion is found by using the characteristics for the cushion fans
in combination with an expression for the expected value of the dynamic change
in the leakage area. The fan characteristic gives a relation between the excess
pressure and the volume flux for constant RPM of the fans. When the pressure
drop in the cushion has been found, an estimate of the sinkage is found by
balancing the weight of the SES with the vertical forces due to the excess pressure
in the cushion and the buoyancy forces on the hulls. Due to the increased sinkage
of the SES, there occurs a change in the still water resistance on the hulls. Due
to the change in the excess pressure in the cushion there occurs also a change in
the still water wave resistance due to the cushion pressure. The results will for
instance depend on the condition of the skirts and how the RPM of the fans are
regulated.

Fig. 4 illustrates computed involuntary speed loss of a 40 m SES and a 40 m

catamaran as a function of significant wave height in head sea waves (Faltinsen
et al. (1991)). The calculations were done for mean wave periods T1 between 3.3
s and 11.7 s. We note that the speed is dependent on T1 for a given value of Hv/3.
For the catamaran this is due to the added resistance in waves acting on the hulls.
It can be explained by the dominant peak in the added resistance curve for regular
waves. For the SES it is also associated with the increased air leakage from the
cushion, which depends strongly on the relative vertical motions at the skirt. It
was assumed in the calculations that the skirt in the front part of the SES just
touched the water surface in calm water and that the RPM of the fans was
constant for all sea states. The figure illustrates a very different behaviour in the
speed loss of a SES and a catamaran. The reason to the more rapid drop in speed
of a SES with increasing significant wave height is due to the air leakage in
waves. The example in Fig. 4 demonstrates that one should not use a sea-state
system where there is only one mean wave period associated with one significant
wave height. It should be realized that the total shaft powers for the catamaran
and the SES are respectively 8300 KW and 5500 KW. Even by allowing for a 20-
25% increase in power due to the fans of the SES it is seen that the SES uses less
power and keeps a higher speed than the catamaran for nearly all sea states of
practical interest.

SLAMMING

Slamming loads are important in the structural design of high speed vessels.
Slamming causes also the ship master to reduce the ship speed. The normal way
to predict the voluntary speed loss due to slamming is to first calculate the
standard deviations of relative vertical velocity and motion in a vessel-fixed
coordinate system at places where slamming is likely to occur. The slamming
probability is found by defining a threshold velocity for slamming to occur. An
often used criterion is that a typical ship master reduces the speed if' slams occur
more than 3 of 100 times that waves pass the ship. The conventional way of
defining a threshold velocity does not reflect the effect of the structural form. In
order to come up with better criteria it is necessary to study theoretical models or
performing experiments for water impact against wet decks and hull shapes
typical for high speed vessel. This is also necessary in order to develop rational
criteria for operational limits due to slamming. The criteria should be related to
slamming loads used in the structural design.

Important parameters characterizing slamming are the position and value of

the maximum pressure, the time duration and the spacial extent of the slamming
pressure. The pressure is heavily dependent on the local geometry relative to the
water surface. When the deadrise angle a is small, one should not put too much
emphasize on the peak pressures. It is the pressure integrated over a given area
that is of interest in structural design. WAhen a is small, the mean pressure over
a plate area will obviously be smaller than the peak pressure. Experiments on
slamming loads should more often realize what the results should be used for.
This means one may avoid using pressure gauges to try to find slammidng loads.
The use of panels mounted on force transducers will give a more appropriate input
for design load calculations. High sampling frequency in order to determine the
correct rise time duration and maximum value of the impact load is important.

There is need for systematic studies that show how. the slamming pressure effects
the global accelerations of the vessel. Det norske Veritas' (DnV).- rules for
slamming pressure on high-speed vessel relates the pressure to the acceleration
of the vessel. Implicitly one assumes that the acceleration is a function of the
water impact. The occurrence and the magnitude of the slamming pressure are
strongly dependent on the relative vertical motions and velocities between the
vessel and the waves. Fig. 5 gives an example on how sensitive the relative
vertical motions and velocities are to the mean wave period T2 and the significant
wave height H1/3 . A two-parameter JONSWAP spectrum recommended by IflC
was used in the calculations.

GLOBAL WAVE LOADS

In the discussion of global loads we will concentrate on catamarans. Global

wave loads are expected to be significant for catamarans of lengths larger than
approximately 50 m. Important global loads are vertical bending moments, vertical
shear forces and pitch connecting moments on the half part of the catamaran
obtained by intersecting along the centre plane. The catamaran can normally be
considered rigid in the determination of the global loads. However, for "steady-
state" response it should be ensured that the natural frequencies of global elastic
modes are sufficiently smaller than encounter frequencies of practical interest.
Morris (1991) reported natural frequencies for a 91 m long wave piercing
catamaran, that has 1.6 Hz natural frequency in torsion and 1.5 Hz natural
frequency in transverse arching. Sufficient information on mass distribution is
sometimes lacking when model test results are presented. It is recommended that
the following data are given when global loads in the centre plane of a multihull
is examined.
 Centre of gravity position
Distance from centerline to centre of gravity of one half part*
Pitch, roll and yaw radius of gyration with respect to axis through centre of
gravity
Coupled iner-tia moments in roll-pitch, roll-yaw and pitch-yaw of one half
part*

*) Half part obtained by intersecting along the centre plane of the multihull

Faltinsen et al. (1992) presented numerical and experimental results of global

wave loads on a catamaran at Froude number 0.49. The numerical method is a
further development of the high-speed theory presented by Faltinsen & Zhao (1991
a&b). "Steady-state' response is assumed. The agreement between theory and
experiments is generally satisfactory except for vertical shear forces. The
experiments were done with a free running model in a basin of length 80 m and
breadth 50 m. Regular incident waves of different wave headings were used.
Examples of experimental error sources are: a) Non-constant wave amplitude
along the track of the model; b) Difficulties in accurate heading control; c)
Insufficient number of response oscillations and transient effects in beam,
quartering and following seas; d) Nonlinear effects. Small deviations in heading
may have a noticeable influence on the global loads. The effect of the autopilot
system and the rudder-propulsion system on the global loads were not investigated
systematically. There is need for good numerical modelling of these effects to scale
the results properly to fullscale conditions without autopilot systems. The
numerical model for high speed shows that vertical shear forces and 'Vertical
bending moments are generally largest in beam sea, while the largest values for
pitch connecting moments occur at wave heading 600 for most wave periods.
Vertical shear force and pitch connecting moment are zero in head and following
sea according to the numerical model.

Fig. 6 gives an example of how sensitive the global loads are to wave heading.
Roll motion is important for vertical shear force and pitch connecting moment,
while heave and pitch acceleration influence the vertical bending moment. It is
sometimes advocated that results for beam sea can be based on zero-speed results.
This is not adviceable due to the for-ward speed effects on the hydrodynamic
pressure distribution.

Faltinsen et al (1992) presented longterm predictions of global wave loads on

a catamaran. The effect of involuntary speed reduction was neglected. The vessel
had a Froude number 0.7. If the catamaran did not satisfy the slamming criterium
often reducing the ship speed to Froude number 0.45, it was decided to exclude the
sea state from the long term prediction of the response. This means that
operational limits are imposed. In practice operational limits have to be decided
in a different way. This will influence the final results. The long term predictions
were based on a standard procedure (Pal tinsen (1990). This means that for each
sea state the proper Rayleigh distribution was multiplied with the probability of
occurrence of the sea state. These products were then added together. It was
decided to select design values corresponding to a probability of 10' Results were
presented for different ship lengths between 50 m and 120 mn. Tlhe predicted
design values for global loads were clearly lower than recommended values by
DnV. However, the design values will depend on the philosophy behind the rules.
It will results in lower values if operational limits are imposed. The design
philosophy is therefore important when considering weight optimization of high
speed crafts.

CONCLUSIONS

The seakeeping characteristics and the different behaviour of a SES,

catamaran and a foil-catamaran in waves are discussed. It is pointed out that
cobblestone effects on a SES are also due to acoustic resonance in the air cushion
and that cobblestone effects are troublesome for a SES in small sea states. A SES
has a good seakeeping behaviour in moderate sea conditions. A ride control system
will have a significant positive influence on the seakeeping behaviour of a
foilcatamaran.

The necessity for establishing rational procedures for obtaining operational

limits for high-speed vessels are discussed. An example is operational limits due
to the propulsion and engine system in a seaway. For instance exposure of the
waterjet inlet of a SES to free air can represent a problem. A special "deck
wetness"problem that can occur for high-speed catamarans in following and
quartering waves are "deck-diving".

Speed loss of catamarans and SES in a seaway are discussed. Even is a SES
looses easily the speed in a seaway and-the catamaran does not, it is possible for
the SES to use less power and to keep a higher speed than a similar sized
catamaran in nearly all sea states of practical interest.

Design rules for slamming and global wave loads on high speed vessels are
discussed. The design philosophy is important when considering weight
optimization of high speed crafts.
REFERENCES

Faltinsen, 0., 1990, Sea loads on ships and offshore structures, Cambridge
University Press.

Faltinsen., Svensen, T., 1991, Incorporation of seakeeping theories in CAD,

Proceedings Int. Symp. on CDF and CAD in Ship Design, Wageningen,
Netherlands, Editor G. van Oortmerssen, Elsevier Science Publishers B.V., pp.
147-164.

Faltinsen, 0., Zhao, R., 1991a, Numerical predictions of ship motions at high
forward speed. Phil. Trans. Royal Society, Series A., Vol. 334, pp. 241-252.

Faltinsen, 0., Zhao, R., 1991b, Flow prediction around high-speed ships in waves,
8
"Mathematical approaches in hydrodynamics", Editor T. Miloh, SIAM, pp. 265-28 .

Faltinsen, O.M., Helmers, J.B., Minsaas, K.J., Zhao, R., 1991, Speed loss and
operability of catamarans and SES in a seaway, Proceedings FAST'91, Trondheim,
Norway, Tapir publishers, Vol. 2, pp. 709-725.

Faltinsen, O.M., Holden, K.O., Minsaas, K.J., 1991, Speed loss and"operational
limits of high-speed marine vehicles, Proceedings IMAS'91 - High Speed Marine
Transportation, Sydney, Australia, pp. 2-1 to 2-9.

Faltinsen, O.M., Hoff, J.R., KvAlsvold, J., Zhao, R., 1992, Global wave loads on
high-speed catamarans, Proceedings PRADS'92, Newcastle, England, Vol. 1, pp.
1.360-1.375.

Hoff, J.R., 1990, Three-dimensional Green function of a vessel with forward speed
in waves, Dr.ing. Thesis 1990-25, Division of Marine Hydrodynamics, Norwegian
Institute of Technology, Trondheim, MTA report 1990:71.

Jullumstro, E., 1990, Stability of high-speed vessels", Proceedings STAB'90,

Napoli, Italy.

Kaplan, R., Bentson, J., Davies, S., 1981, Dynamics and hydrodynamics of surface
effect ships, SNAME, Vol. 89, pp. 211-248.

Damage investigation on diesel engines in high

Meek-Hansen, B., 1990,Fifth speed
vehicles, Pro&6edings Int. Congress on Marine Technology Athens'90,
Hellenic Institute of Marine Technology, Athen, Greece.

Meek-Hansen, B., 1991, Engine running conditions during high speed marine craft
operation, Pore. FAST'91, Trondheim, Tapir Publishers, Norway, Vol. 2, pp. 861-
876.

Mishima, S. 1992, An experimental study on air drawing of waterjet inlet for

suface effect ship, Proceedings HPMV'92, Arlington, VA, USA, pp. SES95-SES105.
Ohkusu, M., Faltinsen, 0., 1990, Prediction of radiation forces on a catamaran at
high Froude number, Proceedings 18th Symposium on Naval Hydrodynamics,
Univ. of Mich., Ann Arbor, National Academy Press, Washington D.C. pp. 5-19.

Sorensen, A., Steen, S., Faltinsen, 0., 1992, Cobblestone effect on SES,
Proceedings HPMV'92, Arlington, VA, USA, pp. SEs17-SES30.

Steen, S., 1993, Vertical plane dynamics of SES with flexible bag, Dr.ing.Thesis,
Department of Marine Hydrodynamics, Norwegian Institute of Technology.
 3.0 5.0 10.0
'A I -•T (s)

1.0

2.0 • •

/ " N, 40 m long catamaran (40 knots)

3.0 "
" 36 m long foilcatamaran (50 knots)

4.0 \40 m long SES (50 knots)

5.0

OPERATIONAL LIMIT 0.2g RMS VERT. ACC. CG.

H V3(m)

Fig. 1 Operational limits for a 40 m long catamaran, 36 m long foilcatamaran,

40 m long SES in head sea longcrested waves. H1/ 3 = significant wave
height, T 1 = mean wave period. Speed in calm water is shown on the
figure. Involuntary speed loss used for the SES and the catamaran. No
ride controls are accounted for.
 a Z (m/s 2 )
10
8.0 I

6.3
5.0
4.0 __

3.15 -Motion sickness region

2.5 ---
2.0
1.6-
1.25-
1.0 30 min.
0.8 -
0.63 i
05 2h
0.5

0.315 _

0.25 8 h(tentative)/ j -

0.2
0.16 -... . . .
0.125
0.1 ---- fc (H z)
0.1 0.315 0.63 1.0

Fig. 2 Severe discomfort boundaries according to ISO 2631/3 for the a

z-
component of the acceleration, as a function of frequency and exposure
time. a z = RMS-value of one-third octave band. fc = centre frequency of
one-third octave band.
 a z (m/s 2 )
20
16 _ _ _
12.5
10 ' ': • '; i'

8.0 1 ea-..4- --
6.3 -. __ __ __ _

5.0 I _
4.0I 16 i

3.15 -
1 ruin-
2.5 I1m
2.0
0.63 5 min
0.25
0.0

0.63 o a band sea

,J 4 h e
0.40
0.315 . i 18 h
0ý25 • i
20 4
0
0.2 S e ra g
0.16 4 m
0.10 ffc (H z)
1.0 2.0 4.0 8.0 20 40

Fig. 3 050.0
Fatigue-decreased profiency boundaries for the az-componen
t of the
acceleration as a function of frequency and exposure
time (ISO 26ea/n).
a z = RMS-value of one-third octave band. f(= centre
third octave band. frequency of one-

Vessel speed (knots)

50.0 ]Head sea

Speed range
,•(involuntary speed loss)
40 mn SES

M0. Spe ed range

(involuntary speed loss)
•"40 m catamaran
H 1/3(m)
1.0 2.0 3.0 4.0

'ig. 4 Numerical calculations Of vessel speed range of a 40 in SES and a 40 in

catamnaran ais function of significant wave height in head sea. Only
involuntary speed loss effiects (Faltinsen et al. (1991)).
 RMS - RELATIVE VELOCITY
5-
..
€<..... -0-F. =0.40]
S '. Fn--.3

- 2-

0
0-I I I I
1.5 2.0 2.5 3.0 3.5 4.0 45
T2g7'L

RMNS - RELATIVE MOTION

1.0-
•"-/•. "..•-G- Fn--0.403
0.9 - -.. F"-.71
0.6-

0.2

0.0 0.11 - I I I i

1.5 2.0 2-5 3.0 35 4.0 4-5

Fig. 5 Examples on RMS values of relative vertical motions (oR) and velocities
(OuR) at FP for a catamaran in head sea longcrested waves. L = ship
length. g = acceleration of gravity. Fn = Froude number.

0.0010 -

1
-

i? H 1r -..,
o
_0.0008 9 Heading: 0.0 deg
--- o Heading: 15.0 deg
0.00. .... •--'Heading: 30.0 deg
0.0 06........ -...Headring: 45.0 dieg
Heading: 65.0 deg
0.0004 Heading: 750 deg
-.-.--- Heading: 90.0 deg
. .......... 0--Heading: 1050 deg
a6--Heading: 120.0 deg
0.0002 -____ _ ----- Heading: 135.0 deg
,. ..-" .... -"Heading: 150.0 deg
........Heading: 165.0deg
0.0000 - E I- "_ .'- _.* Heading: 180.0 deg

2 3 4 5 6, •L

Fig. 6 Examples on RMS values of pitch connecting moments ((75) in the centre
plane of a catamaran in longcrested irregular sea as a function of wave
heading and mean wave period T,. Fn=0.7. (Faltinsen et al (1992)).
 FLIGHT CONTROL SYSTEMS FOR
FOIL CATAMARANS
Erling Lunde

CAMO A/S
Olav aryggvasons gt. 24, 7011 Trondheirn, Norway
TIE +47 7 51 49 66, Fax. +47 7 51 42 57

Course notes for the UETP-MST course on

"New techniques for assesing and quantifying vessel stability and seakeeping qualities".

February 1993

Abstract
This paper discusses the development of flight (ride) control systems for
foil supported catafiarans. The'attention is focused on system properties and
problems that are especially relevant to the control engineer. The main topics
are: hydrodynamic model properties, control system synthesis, sensors and
actuators. The reader of this paper should gain an understanding of the main
elements of a flight control system and the main problems that have to be
resolved during the development.

1 Introduction
The development of foil supported naval vessels has been going on for several decades.
At the same time, Norwegian shipyards have a long tradition in building catamarans.
During the last years, these two concepts have been combined into a new type of
vessel called Foilca by two Norwegian shipbuilders: Westamarin West [15, 17] and
Kvmrner Fjellstrand_,[8]. Recently, the Japanese company Mitsubishi presented their
foil catamaran.
In 1989, Westamarin started the planning of a slender hull catamaran modified
with two fully submerged foils fore and one larger aft. Marintek in Trondheim was
involved at an early stage performing small scale model tests. The foils were designed
by Marintek. The first sea trials with this model clearly showed that the vessel was
unstable when foilborne, aid that it co(hl not op(rate without a flight control system
to ensure stability and comfort. CAMO was engaged to develop the flight control
system for Foilcat. 2900', with assistance from SINTEFP' div. of Antomatic Control.
lFoilcat 2900 was manufactured by Westainarin West AS, Mandal, Norway. Today, the vessel
runs ini regular paClIgcr trallic ini D)eirrnmark.
 I

icraul
yd
TH

Figure 1: Control system, main functional elements

This paper describes the main concerns for the development of flight control
systems for fully submerged hydrofoil catamarans. The presentation is partly based
on available literature and partly on experience from Foilcat 2900.
The advantages of foil catamarans to conventional vessels are

" Higher speed (50 knots or higher).

* Better comfort.

* Increased weather Window..

* Lower fuel consumption (compared to speed).

A modern foil supported vessel requires the solution to several complicated con-
trol system problems. Most relevant to the actual operation of the vessel are:

Flight control (attitude control, ride control, stabilization system) is needed to

stabilize the vessel when foilborne, to ensure passenger comfort by compensat-
ing wave induced disturbances, and to control the vessel attitude. This system
controls the heave, pitch and roll degrees of freedom.

Steering includes rudder control and autopilot design, and takes care of the yaw
.land sw4,) degrees of freedom.

Forward speed, surge, is controlled from the thrust handle. This includes engine
and propeller control.

This presentation deals with the design of the flight control system. There are three
main .--l cenIs ihat must be present to complete the control loop (lFigure 1):

" [here 1m,:s1, he a sufficient number of suflicicntly good sc.sors to iicasure the
.essel ruo:.ion (response) and, if possible, incaui re exte:rnal disturhaiices.

" The vesse tnust be equipped with a set of force producing devices, actuators,
hial. carn ;affect veselcl h:lhavior.
 * There must be a control system that reads sensor signals, compares actual
and desired vessel behavior, and calculates actuator inputs so that the desired
vessel behavior is obtained - as good as possible.

'These are the main themes for this paper.

2 Control system development

The development of a flight control system will typically follow a procedure as illus-
trated inFigure 2. One should expect to "loop" around in the diagram.
For the control engineer, the initial step would be to develop a mathematical
vessel model to be used for system analysis, computer simulation and control sys-
tem synthesis.' For a problem of this complexity, controller development is not an
easy task, and could probably not be done without computer simulations. The time
being consumed and the risk involved in sea trials set so severe limitations on ex-
perimentation that it would be unthinkable to develop the system against the real
vessel prototype.
The development of the control algorithms should run in parallel with the spec-
ification of sensors and actuators. One should be careful not to let the algorithm
rely on measurements that cannot be obtained or that will not have the expected
accuracy. Also, there is little use in the world's best control algorithm if we don't
have actuators that are able to effectuate their input commands.
The effect of every possible system failure can easily be tested on the simulator.
Methods for detecting and handling failures should be developed.
If the simulator controller algorithms were programmed with a software interface
(data structures, function definitions, etc.) equivalent to the one in the real time
system, the implementation on that system should be relatively straightforward.
Dynamic simulation combined with animation is very effective for demonstra-
tion of the vessel performance to future crew, employers, potential customers, etc.
(see Figure 3). It may give answers to "What if ..." questions and create an early
confidence in the system.

3 Hydrodynamics and model properties

The development of dynamic models for high speed vessels is normally based on
general hydrodynamic theory and results from small scale model tests. The final
mathematical model then suffers from inaccuracies from interpolation of test results,
scaling effects, and inevitable physical differences between model and full scale vessel.
The complete model becomes very large and more or less inaccessible to analysis (the
Foilcat 2900 hydrodynamic model was described in approximately 3000 FORTRAN
statements). The control engineer usually prefers to reduce the model complexity
to an appro.imate dynamic model describing the main (structural) input-output
dcpenidencies. main couplings, main nonlinearitics, etc. Whatever information that
is lost in this aplIroximation is often inegligibule from a control point of view.
 Idea
Initial specifications

> Model testing

AI
in towing tank Development of failure
detection system

Development of
mathematical model Failure mode
feffect analysis
Development of
simulation program

Prototype
Verification of implementation
simulation model
• :. . •"---:Sea trials

------ d gControl
system

designi n s sF
ligh t con tro ller

Sensor & actuator finished

specification

Simulation analysis

• ~Performancer

evaluation

Figure 2: Control system devclopiiienit (Foilcat 2900)

4
 -- ~ - - - .5 - -

Figure 3: Foilcat 2900 simulator

3.1 Vessel design

The first hydrofoil vessels to be tried out were monohulls. Two foil concepts were
introduced: surface piercing and fully submerged 110].
Vessels with surface piercing foils (Fig-ure 4) have the property of being self-
stabilizing: When the vessel rolls, the foils become unsymmetricalty submerged and
will produce a straightening moment. Similar arguments apply to longitudinal sta-
bility. Unfortunately, since the foils extend through the water surface, they will
be subjected to wave induced disturbance forces, and the vessel will need a control
system to improve comfort.
Vessels with fully submerged foils are much less exposed to wave forces, but
are unstable when foilborne. However, if equipped with a flight control system,
these vessels have tihe potential of very good,: passenger comfort during heavy sea

Figure ,4: Monohmll with surface picirciug foils- A roll uffscL, gives a scl f-stabi lizinjg
momN•nt, 7'.

5
 .* I ., ,...

Figure 5: Outline of hulls and foils for a fully submerged foil catamaran.

conditions.
No foil catamarans with surface piercing foils have been reported so far, to
the
author's knowledge. The common approach is two small foils forward under
each
hull and one large aft stretching across both hulls. Each foil may have one or
several
control surfaces, flaps. A typical configuration of hulls and foils is shown in
Figure
5. The foils are connected, to the hulls by vertical struts, usually the forward
struts
can be rotated and used-as- rudders:
A related concept is catamafans supported by only one set of foils fore or aft.
The
purpose of this system is to provide roll and pitch motion damping only. No
height
control is included. Consequently, control system design becomes much simpler
and
vessel cost becomes lower than for a foilcat.

3.2 Hydrodynamics
The vessel dynamics can be described in terms of six variables (degrees of freedom),
three describing transversal motion: surge, sway and heave; and three describing
rotations: roll, pitch, yaw. The state of the ship is uniquely defined by the
posi-
tion/angle and velocity for each degree of freedom.

Motion variables: Roll: 6 (radians), € (radians/s)

Pitch: 0 (radians), 0 (radians/s)
Yaw: V) (radians), (radians/s)
Surge: X (m), = uý (m/s)
Sway: y (1n), , = , (m/s)
HIeave: z (ii), = n. (i/s)

The system ii;puts are: the flap angles and(I ist nrbai ices froin waves, Cufrrenit and
wind. Disturbances inay bee seen in Lteris of forces or inonjelits acting on the
vessel,
or in teruis of the orbital velocity of tihe wavcs and cuirrent and wind(
speeds.

I;
 External inputs: Flap angles: 6j, i = 1,2,..., n (radians or degrees)
Disturbances: vi, i = 1,2,...,6 (N or m/s)

The flap angle is normally given by hydraulic actuators. These are also described
'by dynamic models, introducing more system variables. It -is, however, wise to deal
with the actuators separately from the vessel dynamics to avoid too complicated
models, see also Section 5.1.

3.2.1 General nonlinear form of hydrodynamic equations

Let the symbol x denote the six dimensional vector of positions and angles

= [,0,;O,x,y,z] T
Together, x arid 5:define the state of the vessel. The dynamic model can now be
described by a second order nonlinear equation

z = f(x,:i,v)+B(x,.i)u
y =g(X, :
which is linear in the input variable u (flap angles). The sensor outputs are denoted
y, f(.) and g(.) are vector functions, and B(.) is a matrix function.

3.2.2 Linearized models

For control design and analysis of a nonlinear system, one often needs to derive a
linear approximation ofithe dynamics valid for a given operating state, x 0 and Ai:
°.
The model is then written in terms of the deviation x from x0 :
:i- Alz+A 2 i+Bu+Cv (1)
y = D 1 z+D2
The linearized model will be a good approximation to the nonlinear model in a
neighborhood of the point (X0 , &ýo).

3.2.3 Choice of controlled variables

When designing a flight control system, some consideration should be paid at the
choice of which variables that are to be controlled. The roll and pitch angles are
obvious candidates. Heave motion is usually defined as the vertical motion at the
center of gravity (CC), this may not be the best variable for control purposes. The
following example illustrates this problem:

Example: Longitudinal dynamics

Figure 6 shows a highly simplified model of the longitudinal dynamics of a foil
catamaran. Our main purpose is to study input coupling to pitch and heave motion.
The equations of motion, for small pitch angles and neglecting gravi'y, are given
by
l-(F, +F2
171l

L.

7
 h

IC.-G " ....-L --- - F1

F2

Figure 6: Simple model of longitudinal dynamics

wave
motion

positive
flap angle
Figure 7: Single foil and flap model

where m is the mass, J is the moment of inertia and F, and F are vertical
2 forces
resulting from flap motion. On the other hand, if we consider the dynamics
of the
heave motion at a point J/rnL to the right of CC in Figure 6 we get the equations

2
S= m
F,
- f2)
= (r,
Note that the heave mootion, as defined by the new variable h, now is decoupled
from
the force F2, i.e., we are free to use F 2 for pitch control without influencing the
heave
degree of freedom.

3.2.4 Single foil and flap

A major contributor to vessel performance is of course the foil/flap system,
see
Figure 7. The vertical Liftforce, L., from a foil running through water with relative
speed, v tu,, is given by

L.. =(Cl(o +Ca+C6,6) •pv2 (2)

wlwre a is the foil angl, of attack, p is the density of water, and C1.o, C,' C can
6 be
regarded as foil design constants. Note that the lift increases by the squared speed,
ald that the same ajp'iis for lift variations fromi flap motion.

,8
 SIGNIFICA.NT, WAVEHEIGIrT IIH, (i)

03 flatxiiftinehew~ht )

00

411

AANW
0-
- 44l 3.50.

Fiur 8 Heigh itfrhlmtationsti

vs pe o olat20.Tehihmi eie szr
4.l hegh

d o t
n el.0m. e l Sank sheltermdi

caiainlmtdrae
0 ihicesn
k pe.It L isdirbettytodnif

t fi S WITH SIGNIFICANT WAVEHEIGHT > "tk-f' SPEED

FOLLOW THE LEFT LIMITATIONS OF THE CURVE. ('knots)

Figure 8: Height limitations vs. speed for Foilcat 2900. The height isdefined as zero
when the keels at the forward struts are at the mean water surface [15w.

To avoid foil cavitation, flap angles must be constrained (flap saturation). The
cavitation limit decreases with increasing speed. It is desirable to try to identify
the saturation function, 6,,o, = S(u.), this may be an important part of the control
algorithm.

4 Desired vessel behavior

4.1 F light height

Tise captain will expect that the vessel runs steadily at the specified height. However,
depending on ite vessel speed, there Will be maximum and minimum limitations to
tie flight height. At lowv speed, the vessel cannot "take-off,'. At high speeds tie
vessel cannot go hullbornec, i.e., it hxs to go foilborne. Figure S shows the possible
flight height range for l~oilcat 2900 as a functrion of speed.
Several aulthors 1101 discuss thc pIchnoinena contouring (tracking Lthe wave) and
platforming (cutting, through thw Wave). Obviously, We prefer the(. vessel to keel) a
constanlit highit regardless of Lth; wavc:s. C'onsequcslty, contouring shnuid ncctir only
as a result of Lthe systemi's inatbility to achlicve platforining, see also Section 6.1.
4.2 Banked turn
The roll angle should be held constant equal to zero except during turning maneu-
vers. At high speed, the vessel should bank inwards when turning to reduce
the
impact of centrifugal forces on the passengers.

4.3 Constant pitch trim

The pitch angle will normally have an optimum value where the drag forces
reach
a minimum and propulsion forces reach their maximum, or as a trade-off between
these two. This angle will be the constant pitch reference.

5 Control system design

I

'There are two main model properties that must be dealt with when designing
the
control system for a foil catamaran:
Couplings: The system is multivariable with couplings from all control surfaces
to
most degrees of freedom. This means that if we try to optimize a controller
that does not take these couplings into account we may introduce structural
closed loop instabilities.
Nonlinearities: The system behavior is nonlinear in the state variables, especially
in' surge and flight height. That is, the vessel response to flap inputs depends
on the vessel speed and height. If these dependencies are not accounted for,
we may get an .ilfistable or a ývery sub-optimal system for some states away
from the normal operational state.
In addition, the vessel interacts with an environment that may cause large involun-
tary motion highly unpleasant to the passengers:

Disiurbances: The control system has to counteract disturbance forces primarily

from waves, but also from wind and current.

In designing control systems for vessels with such properties there are two possible
approaches:
1. Design a model based algorithm that explicitly tries to compensate couplings,
nonlinearities and disturbances. The drawback of this method is that we need
accurate model knowledge, and we need to measure or calculate disturbances.
If this knowledge is insufficient, the algorithm may not work. On the other
hand, if our system knowledge is good enough, we may achieve very good
control system performance.

2. Design a robust controller. This is a potentially simple algorithm that may

depend very little on detailed ilodel k nowledge- Its drawback is that it
tends
to be subopltimial with respect to jperforIlnluce.

T'e two approaches inay be combinicd, tlh(e control algorithmu then becomesa tradeofF
between performance and robustness. lHowever, one should always utilize whatever
IuiOdcl k nowleige is available.

I(1
5.1 Decentralized flap servo control
The flap angle is obtained by help of a electro-hydraulic servomechanism. The servo
will normally have its own position control loop" which tries to position the flap
according to an incoming desired flap angle. This loop is decoupled from the flight
control system, i.e., we have decentralized control loops at each servo.
As seen from the flight controller, the servo unit (actuator) will ideally act as a
device giving the desired flap angle (a unit gain function), S = 6• . In practice, the
servo may exhibit some dynamic behavior, this is usually described by a first order
lag
6= I +Ts S1
If the time constant, T, is small enough, the dynamics can be neglected.
Additionallý, there are mechanical constraints to the flap angle, and there are
rate constraints due to limitations in hydraulic capacity

&min
: bý b56ma, ýmin <_ a

5.2 Dedicated flap control

There has been several related approaches to control systems design for fully sub-
merged foils on monohulls, see e.g. [16, 14]. They all base their design on the
following assumptions:
" Lateral and longitudinal dynamics can be separated.

" Different flaps are dedicated to different tasks according to common engineering
sense, we get a quasi multi-variable feedback system.

* There is one vertical accelerometer directly above each foil/flap. The acceler-
ation signal is used in a local feedback loop for that flap servo only.
One such system for the Tucumcari hydrofoil gunboat is indicated in Figure 9 for a
vessel with three flaps [16], one forward and two aft.
The assumption on lateral/longitudinal separation is reasonable. However, one
should have in mind that the two subsystems use the same actuators such -that one
subsystem may cause flap saturation that will degrade the performance of the other
subsystem.
The decentralized structure of the acceleration feedback loops is not advised.
This structure is due to a common misconception, e.g., as reported in [13]
"... it is assumed that the least ship motion will be achieved if we mini-
mize the motion at each individual foil..."
'This is not true. Obviously, any force being applied to reduce vertical motion,
e.g. above the starboard aft foil/flap, will influence vessel behavior as a whole due
to dynamic couplings. A sub-optimal dedication of flap action may excite coupled
dynlamics and cause low control performance (low feedback gains) and/or oscillations
or instabilities (high gains). ['urtliermorc, dlirect acceleration feedback may introduce
noise induced oscillations since these loops may become very fast. For the Patrol
Hydrofoil Missile [141 these problems were recognized and solved in a very practical
way:

II
 Inputs Gains & Filrtm Outputs

mmaHegh
Stro~ard

Heim
command

Figure 9: Dedicated flap control system [16J

"The acceleration channels are provided primarily for ride smoothing;

they are the higl-bestarder feedbacks used and most susceptible to structure-
induced, oscillatory instabilities; hence, provisions are included to turn
any one or all to zero." (author's italics)

5.3 Optimal. control

Classical optimal control is based on a linear dynamic model (1) and tries to minimize
a quadratic performance index of the kind [4]

2 J0 (3)
The index weights the (squared) deviation, ( pi). from the operating point, (uni 0 ),
against the consumption
l. of power, The optimal control gives feedback from all
state variables it = Glx + 0 2 i:.
Usually, all state variables cannot be measured directly so that we have
struct an cstimator e bg.
al a ner to con-
duce estimates, •, of the remaining
variables p9J.
The Kalman filter performs optimal filtering of random (gaussian)
noise arid disturbances. The control system then lbecomes ,as illustrated
in Figure
10, with feedback from the estimates of the stat': variables.
The performan ce ihex can de [on, xated so that e.g. vertical acceleration is
ainain ized [7J. Note, however, uhal.
the restirig optimal control
giiv i not fo
ifcadkd
acceleration feedback. Noise terns (waves) can be modeled and included in the

12
 .

Vessel
r

Figure 10: Linear Quadratic Gaussian Control (& denotes the estimate of all state
variables).

y
Ve se
P Df-,,X,:i)

Figure 11: Nonlinear inverse dynamics

performance index. This will give additional feedback from wave measurements or
estimates. This is, unfortunately, rather difficult to achieve.
A major drawback of this approach is that the optimal feedback is valid only
for a neighborhood of the operating point, (z0 , O ). One must make sure that
the system is stable and has a certain performance also in all other possible, or
probable, operating points. If necessary, several optimal controllers for different
operating points have to be calculated, and the flight controller will have to switch
(or interpolate) between the different feedback gains according to some scheduling
scheme.

5.4 Inverse dynamic's'

In contrast to the tradeoff between performance and energy consumption in optimal
control, nonlinear inverse dynamics aims for perfect tracking of the reference regard-
less of the cost. 'Flu: idea is to use tie f]ll nonlinear dynamic model to calculate the
control inputlneeded to niake t lie vessel belhave as desired. The method is developed
in several similar versions and under different naries such as linearization by nonlin-
ear feedback, nonlinear dccoupling 161 and iterative inverse [!]. Several applications
can be found, e.g., in robotics 112J and aircraft control [11].
'['lie method is Collceptulal 1) very simple and is calculated as follows (see also

I3
Figure 11):

it=B(,~ji-f~~~)
zf = U

We see that the resulting system has the form of a set of decoupled double integrators.
This system can easily be controlled by a set of PD-controllers.
There are some problems with the realization of the nonlinear feedback: i) the
dynamic model must be sufficiently accurate, ii) the disturbance vector, v, must be
measured or estimated, and iii) the matrix B(z, i) must be invertible (see Section
5-5). Additionally, the control input may not be within the range of the actuators,
i.e., saturation may occur. If these problems can be resolved, inverse dynamics offers
the possibility of almost perfect performance.

5.5 Actuator redundancy

is described by the matrix
The effect of flap action, u, on vessel acceierations, i:,
B(z, ýi). That is, the matrix describes the input couplings, the distribution of flap
induced forces to the different degrees of freeaom. In some way or another, this
information is used in the controller design:

" For the dedicated flap control method (Section 5.2), the main couplings are
defined from the.-engineer's understanding of the real couplings contained in
B(z,
" In optimal control (Section 5.3), the linearized input matrix, B, is used ex-
plicitly in the calculation of the feedback gain matrix, G.

" Inverse dynarics (Section 5.4) uses the inverse of the nonlinear matrix, B -1(x,),

to achieve decoupling.

The Foilcat typically have four or five flaps that are used to control three degrees
of freedom, i.e., we have flap, or actuator. redundancy. Let us consider a simple
example where four flaps are placed symmetrically around the vessel. The 3 x 4
.input matrix from flap angles to roll, pitch and heave response can then be written
on the general form bi -bj b2 -b2
B 53 b -, b., b4
bs bl- bc, br

where the elements b; may be constants or nonlinear expressions. Now, note that
for any value of AN.the input vector

a
ij; i
I, oI wrrtv t liat
B u= 0 0
Captain's Refe
commandsgerao Control Priority Desired

algorthm SstemFlap angles

Sensors -- AnalysisSy
Transformations

Figure 12: Foilcat flight control system

That is, all inputs on the form uo will not affect any of the three degrees of freedom.
The vector uo is said to lie in the null spaceof B [21. Also be aware that a rectangular
matrix does not have a unique inverse, but we may use e.g. the pseudoinverse,
+
B = BT(BBT) - ,.
The redundancy phenomenon opens new perspectives to control: If we don't give
this problem our full attention, we may end up with control algorithms that every
now and then generate input vectors, u, in (or near) the null space of B, and for
those moments, the vessel will be uncontrolled. On the other hand, redundancy
means that the vessel can still be controlled even if we loose one actuator. Also note
that we have the possibility of manipulating the inputs without degrading the vessel
performance. For exahmple, if one of the flaps is saturated, we can bring it back in
operation with an additional input ito where A is properly chosen.

5.6 Foilcat 2900 Flight Control System

The flight control system on Foilcat 2900 have four main modules, see Figure 12.

Reference generator: The captain's commands for desired flight height and steer-
ing are interpreted and transformed to height and roll references that are within
legal (safe) limits. The pitch trim is held constant at 1° bow up.

Analysis & Transformations: All sensor signals go through a failure detection

system that stops any erroneous signal from being used in the control algo-
rithm. In addition, filtering is performed and necessary transformations are
carried out.

Control algorithm: The controller shall make sure that the vessel maintains sta-
bility and compensates disturbance forces in all foilborne situations. The con-
trol algorithm is based on the principle of nonlinear inverse dynamics.

Priority system: The flight controller can handle all flap or actuator failures. Flap
actuation is optiiiizcd coitiriuouIsl to avoic( ilap saLuration and simnltancous
d(:grad;LtioII of jierforiance.

'l'hflight controllhr givcs a set. of d(lsirerI flap ;tngle:s. These arc oItained by seI)arate,
local hydraulic servo control loops.

15
6 Sensors and signal processing
Several sensors are normally needed on a foil catamaran:

* An inertial platform, or vertical gyros, cangive the roll, pitch and yaw
angles
- and possibly also heave motion.

* Accelerometers are often used to give vertical accelerations at different

posi-
tions on the vessel.
* A height sensor is needed to measure the vessel flight height.

" A speed sensor is needed to calculate speed dependent height and flap
angle
limitations.
Obviously, all'sensor signals need some kind of filtering to remove noise,
offsets and
aliasing effects (when sampling a continuous signal). Estimators may be
needed to
generate signals that are not measured directly, e.g. deriving velocities from
position
measurements (see also Section 5.3).

6.1 The flight height

The flight height is usually measured by some kind of range sensor (ultrasonic,
laser)
mounted in the vessel bow. The problem is that the sensor measures the
distance
to a disturbed sea surface while we actually need the distance to the mean
sea level.
An accurate mean level estimate may be obtained by comparing the disturbed
height sensor signal with an .accurate vertical acceleration measurement,
preferably
corrected for roll and pitch motion: This can be done by a simple estimator
as shown
in Figure 13.
The purpose of the estimator in Figure 13 is to use the information from
the
height sensor for mean level determination and the acceleration measurements
for
high frequency motion estimates. By a proper tuning of the estimator
gains we
get transfer functions from the height sensor and the accelerometer, respectively,
to
height estimate as shown in Figure 14. The low frequency range (e.g. mean
value)
uses height sensor information, while the high frequency estimate is essentially
the
double integral of the acceleration.

hm

Figure 13: Integrating heighL sensor sigital and vertical acceleration

netsirermentL.

16
 Gain

h h s2
hm a
1.0

Contouring Plafforming
-- --- -- ---

CW Frequency
Figure 14: Transfer functions from height sensor signal, h., and vertical acceleration,
a, to height estimate, h.

The interesting frequency range can be found from local wave statistics, from typ-
ical frequencies of encounter: The frequency of encounter is defined as the frequency
of the waves relative to a fixed point on the vessel, and is given by

W" + W&)i/g cos f

60ct,=

where w is the absolute wave frequency, fl is the angle of encounter and f3 = 0 means
head seas. The estimator should be tuned so that the typical frequency of encounter
is above the cross frequency, wC.

6.2 Error detection and isolation

The consequences of component failures in flight control systems may be severe. To
reduce the risk of single sensor errors causing serious accidents, one should design
an on-line error detection and isolation algorithm [3].
Erroneous sensor signals should be detected before the control system governs
the vessel into a dangerous situation. It is equally important to isolate the defect
component(s) so that the flight controller can still operate utilizing the functionality
remaining in the system. A signal may be checked against max/min limits or maxi-
mum rate of change. Dynamic relationships, e.g. positions vs. accelerations, can be
checked by means of a dynamic model in an estimator [5].

7 Conclusions
This paper h;s trcatced flight control-without going into much (letail on any subject.
T'l reader is ad vise(l to consult tie rFrcrucc(d literature for further iniformation.
Several important aspects lhas not been discussed at all:

17
 0.50- '"• 3

0.0 mFoilcat

-0.50 .

-1.50 . . .. .

-2.0 . . . . . . Catamaran. . .

-2.50-7. & do1o~

Foilcat ZF (m) 7 ----

Catamaran ZF (m)

Figure 15: Comparison of heave response for Foilcat 2900 and conventional catama-
ran (simulated) at 37 knots in head seas with significant wave height 1.0 m. (Time
axis in seconds.)

The rudders on a-;foilkat-willmost probably be vertical foils. This means that

the same limitations as for flap'control is relevant for rudder control. For example,
if one turns the rudders too much at high speed the steering ability will be lost - we
get what is called "plow out" in the automobile community, the vessel will continue
straight ahead.
Safety has much more to it than simply designing a stable control system. Hard-
ware (sensors, computers) and software redundancy is important in addition to reli-
able failure detection methods. The vessel must "fail safe" in every critical situation.
Quality control of software should be carried out before implementation and sea
trials. Computer simulation is a key tool for a priori testing of flight control software.
There are several rules and regulations regarding the construction and operation
of a high speed vessel, as well as sensors and electronics. It is wise to have these
rules in mind when designing the control system to avoid "smart" solutions that will
not be allowed.

7.1 Vessel performance

A simulation example comparing the response of Foilcat 2900 and a conventional
(:A;tTII;LranI is shown in Figure 15. The cataniaran model %w;s obtained by removing
;L1I forces front the foil system in the Foilcat modcl. 'TU loilcat is foilbornc with both
hulls above ithe 111can sea level. Comparable INIS values for vertical acceleration are
0).0.5y for 'oilcal. and 0.79 for the catamiaran (I q = 9.81 i/s 2 ).
Figure 16 shows the vertical vibrations exeriienccd 1by pafssengers on board the
 BIODYNAMIC EFFECTS: COMFORT AND FATIGUE
SERVERE DISCOMFORT U10UNDARIM- FATICUIE-DECREASED PROFICIENCY OUNtDA RES

4.0

50

I )6
L 0. .00
S 0 AIN.
0 F -I MODERATE
0.8

0.6 - IS ,;'21R S . •ý H-1,

< 0.5
0./ LIGHT
". •[ 0.3 Is( 8 S. IIl =-3M

m RMS
0a
2 ve vt ey.TEsLsumaLE

u
dbfnt ATMAAN SPEE 35,KOS I ,fl1.S
0.1* 0.2
FOCA 0.3
2 0.4 0.50.6
10, 0.81.0
SED3--1 KNOT 2 3 4 5 6 7 8 10 20

[. ^~~VERAGE FR.EQUEN•CY (11,) ,

0CATAMARUAN, SPEED 35 KNOTS, 11 itj= 1.5 NJi

0 FOILCAT2900, SPEED35-45 KNOTS

Figure 16: Comparison of dynamic load for Foilcat 2900 and a conventional cata-
mnaran: RMS value of vertical acceleration vs. frequency. The curves show human
discomfort boundaries for different exposure times [15].
Foilcat 2900. For comparison, the response on board a conventional catamaran is
included. The diagram indicates that the Foilcat can run for more that eight hours
in heavy seas without having too many cases of sea sickness.

7.2 Recommendations
Designing a flight control system for a Foilcat is not easy. Some pitfalls are men-
tioned in this paper, but there are undoubtedly several others of both practical and
theoretical nature. It is therefore highly recommended to bring together experts
from hydrodynamics and control theory to make possible a profitable cooperation.
The references to literature on flight control and related problems are only a few
examples from an enormous amount of relevant literature. Control system developers
are advised to try to be up to date on different areas such as:

* General hydrodynamics: for model building.

" Control theory: recent advances on nonlinear and robust methods may have a
large impact on future systems.
" Robotics: very much research on advanced control of mechanical systems is
going on.
" Estimation theory and failure detection: relevant for all kinds of advanced
sensor systems.

" Linear algebra: important for the analysis of multi-variable systems and for
dealing with redundancy.

References
[1] Balchen, J.B. (1991). Nonlinear decoupling in process control. Modeling Identi-
fication and Control, vol. 12, 81-94.

[21 Ben-Israel, A., T.N.E. Greville (1974). Generalized inverses: theory and appli-
cations. Wiley, New York.

[3] Clark, R.N. (1978). Instrument fault detection. IEEE Trans. on Aerospace and
Electronic Systems, vol. AES-14, 456-465.

[4] Dixon, M. (1976). Linear multivariable control: Its development, computer

implementation, and application to hydrofoil moition control problems. Proc.
lloovcrcraft conference, 369-391.

[5] Frank, P.M. (1990). Fault diagnosis in dynamic systems using analytical and
knowledge-based redundancy - a survey and some new results. Auvomatica, vol.
26, ,159-47-1.
[6] Ireund, E. (1975). The structure of decouplcd non-linear systems. Int. J_ Con-
h-ol, vol. 21, 443-450.

201
 [7] Hsu, S.K. (1975). Application of optimal control theory to a large htdrofoil craft.
Proc. Fourth Ship Control Systems Syrup., Royal Netherlands Naval College, 4
- 46-57.

[8] Instanes, E., J.T. Pedersen (1991). Safe and comfortable operation of Foilcata-
marans. FAST'91, 1093-1112.

[9] Itoko, T., S. Higashino, Y. Yamagami, T. Ikebuchi (1991). The development of

an automatic control system for a submerged bull and foil hybrid Super-High-
Speed Liner. FAST'91, 997-1012.

[10] Johnson, R.J. (1985). Hydrofoils. Naval Engineers Journal,February, 142-199.

[11] Lane, S.H., R.F. Stengel (1988). Flight control design using non-linear inverse
dynamics. Automatica, vol. 24, 471-483.

[12] Mclnroy, J.E., G.N. Saridis (1990). Acceleration and torque feedback for robotic
control: experimental results. J. of Robotic Systems, vol. 7, 813-832.

[13] O'Neill, W.C. (1991). The maximum attenuation of seaway induced motions,
within a given set of design constraints, possible for hydrofoil supported ships.
FAST'91, 1215-1231.

[14] Stark, D.R. (1974). The PHM hydrofoil automatic control system. Proc. Na-
tional Aerospace and Manufacturing Meeting, 1-15.

[15] Svenneby, E.J., K.J. Minsaas (1992). Foilcat 2900 - Design and performance.
3rd Conf. on High Speed Marine Craft, 20 pages.

[16] Vogt, J.E. (1969). Automatic control of the hydrofoil gunboat, Tucumcari. Proc.
AIAA 2nd Advanced Marine Vehicles, 1-10.

[17] RW. (1992). Foilcat ready for service. Ship & Boat International,July/August,
31-34.

21
 Ride Control of Surface Effect Ships
Asgeir J. Sorensen
Department of Engineering Cybernetics, The Norwegian Institute of Technology.
7034 Trondheim, Norway; email: asgeir@itk.unit.no

Keywords: SES, Ride control, Comfort, Collocation, Full scale results

Abstract
A ride control system for active damping of heave and pitch accelerations of Surface Effect
Ships (SES) is presented. Analysis of the dynamic response of SES advancing forward in
waves at
high speed show dominating vertical accelerations in a frequency range of importance
for the
passenger comfort and crew workability. These vibrations are analyzed by a mathematical model
for
the air cushion. This mathematical model accounts for the motions and accelerations in heave
and pitch
induced by both the dynamic uniform and the spatially varying air cushion pressure. The
spatially
varying pressure is described by an analytical modal representation. High ride quality can be
achieved
by compensating-for these vibrations using a ride control system. The development of the ride
control
system is based on the mathematical model. Optimal location of the fan system and the ride
control
system gives significant improvement in the ride quality.

1. INTRODUCTION

High speed marine vehicles will be of increasing importance in the future in the areas of
passenger and cargo transportation as well Jas naval applications. High speed marine
vehicles
transporting up to five hundred passengers may be in operation in few years. and development
of 50
knots cargo craft with displacement of the order of three thousand tons has been initiated in
several
countries. Many new advanced concepts have been suggested. However, major technological
challenges have to be overcome, and the expected increase in speed and size represent a challenge
to
the expertise available from many diverse disciplines. In 1989 the Royal Norwegian Council
for
Scientific and Industrial Research (NTNF) in cooperation with the Norwegian high speed
marine
vehicle industry initiated the Norwegian high speed vehicle research program to take care of
the most
critical technological challenges. Surface Effect Ships (SFS) was one of the high speed concept
that
was given special attention to.
SES has a catamaran type hull form which contains the air cushion with flexible structures
called seals or skirts at the fore and aft ends of the air cushion. Pressurized air is supplied
into the
cushion by a lift fan system and is retained by rigid side-hulls and flexible skirt systems at
the bow
and the stem. The excess pressure lifts the craft and thereby reduces its calm water resistance.
The
major part of the craft weight (about 80 %) is supported by the excess air cushion pressure, while
the
rest of the weight is supported by the buoyancy of the side-hulls. The most common stem seal
system
is the flexible rear bag system consisting of a loop of flexible material, open at both sides
with one
or two internal webs restraining the aft face of the loop into a two or three loop configuration.
Pressurized air from the aft of the air cushion is supplied into the bag system. The bag pressure
is
about 10 - 15 % higher than the air cushion pressure. A major advantage of SES to hovercraft
is that
the rigid side-hulls permit the use of water propulsion, either waterjets or propellers can be used.
The
small draft of the side-hulls in the water is also sufficient to produce the necessary lateral
forces
affecting the manocuvrability and the stability of craft in the horizontal plane. The side-hulls
are
designed with sufficient buoyancy so that the SE5 can float with airgap between wetdeck
and free
surface when the lift fan system is turned off in the same way as conventional catamarans.
 Surface Effect Ships (SES) are known for offering a high quality ride in heavy sea states
compared to conventional catamarans. However. in low and moderate sea states there are comfort
problems due to high frequency vertical accelerations induced by resonances in the pressurized air
cushion and a high performance ride control system is required to achieve satisfactory human comfort
and crew workability. To develop such a ride control system it is essential to use a rational dynamic
model containing the significant dynamics.
Previous ride control systems have been based on the coupled equations of motion in heave
and pitch as derived by Kaplan and Davis (1974, 1978) and Kaplan et al. (1981). Their work were
based on the assumption that the major part of the wave induced loads from the sea was imparted to
the craft as dynamic uniform air pressure acting on the wetdeck. while a minor part of the wave
induced loads from the sea was imparted to the craft as dynamic water pressure acting on the side-
hulls. This work was further extended by Sorensen at al. (1992. 1993) and Sorensen (1993), who
included the effect of spatial pressure variations in the air cushion. It was found that acoustic
resonances in the air cushion excited by incident sea waves can result in significant vertical vibrations.
A distributed model was derived from a boundary value problem formulation where the air flow was
represented by a velocity potential subject to appropriate boundary conditions on the surfaces enclosing
the air cushion volume. A solution was found using the Helmholtz equation in the air cushion region.
In Sorensen and Egeland (1993) and Sorensen (1993) a ride control system for active damping of the
vertical accelerations induced by resonances of both the dynamic uniform and the spatially varying
pressure in the air cushion has been proposed. The basic design principles of such a ride control
system will be presented in this paper.
The mathematical model as presented in Sorensen et al. (1993) will be slightly modified and
adapted to control system design. This mathematical model is then used to derive the ride control
system, which provides active damping of both the dynamic uniform pressure and the acoustic
resonances in the air cushion- Special attention is given to sensor and actuator location to achieve
robust stability and high performance. The stability of the control system is analyzed using Lyapunov
theory as presented in Desoer and Vidyasagar (1975) and in Vidyasagar (1993).
The paper is organized as follows: In Section 2 the mathematical model describing the
vibration phenomena related to both the dynamic uniform pressure and the spatially varying pressure
in the air cushion is presented. Section 3 includes the controller design and the stability analysis in
addition to the problem of sensor and actuator location. Finally. in Section 4 numerical simulations
and results from full scale trials with a 35 m SES are presented.

2. MATHEMATICAL MODELLING

In this section we present a mathematical model for the heave and pitch motion of SES. The
mathematical model as presented in Sorensen et al. (1993) is slightly modified and adapted to control
system design. This model is similar to the work of Kaplan and Davis (1974, 1978) and Kaplan et al.
(1981) but with the significant extension that the effect of the spatially varying pressure in the air
cushion is included. A moving coordinate frame is defined so that the origin is located in the mean
water plane below the centre of gravity with the x-, y- and z-axes oriented positive forwards, to the
port, and upwards respectively (Fig. 1). This type of coordinate frame is commonly used in marine
hydrodynamics to analyze vertical motions and accelerations (Faltinsen. 1991). The equations of
motion are formulated in this moving frame. Translation along the z-axis is called heave and is
denoted 71,(t). The rotation angle around the y-axis is called pitch and is denoted r-(). Heave is
defined positive upwards, and pitch is defined positive with the bow down- We are mainly concerned
about the high frequency vertical vibrations. In this frequency range the hydrodynamic loads on the
side-hulls arc of minor importance. Nevertheless. since the mathematical model is supposed to be valid
also for lower frequencies, hydrodynamic loads on the side-hulls are included. Furthermore, infinite
water depth is assumed.

2
 z

Fig. 1. Surface Effect Ship (SES) - coordinate frame.

The craft is assumed to be advancing forward in regular head sea waves.

The waves are
assumed to have a small wave slope with circular frequency cor The circular
frequency of encounter
0t, is - -

(k,= c + kU (1)

where k=-27r/X is the wave number. X is the sea wave length and U is the
craft speed. The circular
frequency of encounter co,-is the apparent wave frequency as experienced
on the craft advancing
forward at the speed U in head sea. The incident surface wave elevation
Q(x,t) for regular head sea
is defined as

S(xjr) = , sin(moet + kx) (2)

where C_is the sea wave elevation amplitude. The water waves are assumed
to pass through the air
cushion undisturbed as in Kaplan and Davis (1974, 1978) and Kaplan ct
al. (1981). The beam b and
the equilibrium height h. of the air cushion are assumed to be much less
than the length L. Hence, a
one-dimensional ideal and compressible air flow in the x-direction is assumed.
For simplicity, we will
at the equilibrium condition consider a rectangular cushion of length L.
reaching from x=x,.=-L12 at
the stem (AP) to X=XFP'=L/2 at the bow (FP), with the equilibrium height
h. and beam b. This means
that the longitudinal position of the centre of air cushion pressure is assumed
to coincide with the
origin of the coordinate frame. The air cushion area is then given by A,=Lb.

3
 The total pressure variations p,(xst), inside the air cushion is represented by

p,(xt) = p p.(t) + p,V(xt) (3)

where p. is the atmospheric pressure, pj(t) is the dynamic uniform excess pressure and p,,(xt) is the
spatially varying excess pressure. The basic thermodynamic variations in the air cushion are assumed
to be adiabatic. When neglecting seal dynamics, aerodynamics and viscous effects, the external forces
are given by the water pressure acting on the side-hulls and by the dynamic air cushion pressure
p.(t)+p,(x,t) acting on the wetdeck. We will assume that the dynamic air cushion pressure is excited
by incoming sea wave disturbances. In the absence of sea waves, the stationary excess pressure in the
air cushion is equal to the equilibrium excess pressure p,. It is suitable to define the nondimnensional
uniform pressure variations pQ(t) and the nondimensional spatial pressure variations p;ý,(xt) according
to
PIJ() - p.(t) - ,p. I,•(x~t) = P__x't-----
_ (4)
P.
Po
The volumetric air flow into the air cushion is given by a linearization of the fan characteristic
curve (Fig. 2) about the craft equilibrium operating point. It is assumed that q fans with constant RPM
are feeding the cushion, where fan i is located at the longitudinal position x4;. The fan characteristic
can then be represented by

w,(t)= P L IQ + P.-Qi" (MOt) + P,(xt(5)

i.1p((pap
where Q. is the equilibrium air-flow rate of fan'i when p.(t)=p0 and (aQ1ap)lj is the corresponding
linear fan slope about the craft equilibrium operating point Q,• and p.. The total equilibrium air flow
rate into the cushion is then

Q. Q, (6)
I -I

PC

AQ
PO ......... .A p

Fig. 2 Fan characteristic curve.

The volumetric air flow out of the air cushion is proportional to the leakage area A.(t), which is
defined as

At,(t) = Ao + ARC(t) (7)

4
where Aact() is the controlled variable leakage area, and Ao=A.A+A.7 is the equilibrium leakage area
with Af' defined as the stem equilibrium leakage area at x=-L/2 and A7tr defined as the bow
equilibrium leakage area at x=Lf2. The controlled leakage area AR'3(t) is written

A ft(t) = t (ARa+ (xt))

RdAA (8)

where r is the number of louvers, A,,Ra is defined as the mean operating value or bias of the leakage
area of louver i which is located at the longitudinal position x=xu and &4A;Re(xAr) is defined as the
commanded variable leakage area of louver . A louver is a variable vent valve which changes the area
of opening in the wetdeck for the purpose of leakage control. Pressure sensors are used to measure
the pressure variations in the air cushion. Sensor i is located at the longitudinal position x=xi.

Equations of Motion land Dynamic Cushion Pressure

Using the results as derived in Sorensen et al. (1993), the coupled equations of motion in
heave and pitch and the dynamic cushion pressure are set up. The dynamic uniform pressure equation
is derived from a global continuity equation for the air flow into and out of the air cushion. The
spatially varying pressure equation is derived in the time domain from a boundary value problem
formulation by solving the wave equation in the air cushion region subject to appropriate boundary
conditions on the surfaces enclosing the air cushion volume (Fig. 3). One should notice that Helmholtz
equation is derived from the wave equation by assuming harmonic time variations. A one-dimensional
modal solution in the longitudinal direction is found by assuming that the length of the air cushion is
much larger than the beam and height and that the air flow velocities at the ends of the air cushion
are equal to zero.
. L/2 •~2: L/2

AP 11 F
IIa

XXRC

tit

Fig. 3 Boundary value problem.

The last two equations are the heave and pitch equations of the craft centre of gravity
including hydrodynamic loads on the side-hulls in addition to the loads induced from the dynamic air
cushion pressure acting on the wctdcck.

1.Uniform pressure equation

K,ji'(t) + K~p.(t) + p 0 An 1,(t) = K. L>A;ics(Xit) + P'.V 0(t) (9)

5
where the constants K,. K 2 and K. are defined according to

=~~~~-
P.
~2 P , C-~8
. K 3 = -I
1 ( 0)

p.

where p. is the air density at the atmospheric pressure p., p,. is the density of the air at the
equilibrium pressure p. and y is the ratio of specific heat for air. c. is the orifice coefficient varying
between 0.61 and I depending of the local shape on the edges of the leakage area, see Sullivan et al.
(1992). In the numerical simulations c.=0.61 is used. The time derivative of Vo(t) is the wave volume
pumping of the dynamic uniform pressure and is found in the following way for regular head sea
waves
.kL
L/2 sin U.
V0(t)= b f 4(xt)dx = 2 (11)
-L/2 U
2

2. Spatially varying pressure equation

p,(x,r) = ,(t) r.(x) (12)

j I

where r,(x) is the eigenfunctionxbrtlie mode shape function of mode j, which is found to be

r.x) = cosilKjx + L) .x E [_L. 41 (13)

The time derivatives of p/<r) is the modal amplitude function for mode j. Due to linearity and
orthogonality between the modes, we can consider each mode separately and superpose the
contributions from each of them. Hence. the odd modes around the centre of pressure forj = 1, 3. 5....
is written

(t
) + 2ý,mo. p1(t)
j(t + (0 P ) =

- c21 f 5(t)+ c) costu x'.kI AA;RCS (x"t) + PCO(t (14)

. L 2) (14

where

C, = 2Kc'
-C'. 4poL c( (15 )
PoV o poho(jjr )2

6
The wave volume pumping of the odd acoustic modes for regular head sea is

k cos-kL
-4 C 2 a0.C sinanot (16)

Even modes around the centre of pressure j = 2. 4. 6.... is written

,6,(t ) + 2•spos ps(t ) +(2 " rtCS m

+ in, pi(t) = c:S cos.S._XU+.J AAj
A (xtt)i + Po V.(r) (17)
il 2 )

where the wave volume pumping of the even acoustic modes for regular head sea is

k sin-U
4 C2 2 (O.o• cosco t (18)

The relative damping ratio g for the odd and even modes is found to be

c . •"•-,
Ao + "- A,--
m.
Ax sC 0 3
R19) L + LK -IPoIE 1.;Z
o•CSý---l-- (2.+--
and the corresponding eigenfrequency ow
s for the odd and even modes is found to be

CO Sc L j = 1,2. 3.. (20)

3. Heave equation

(m + A 33 )113 (t) + B 33")3 (t) + C 33 rI3(t) - App.(t) =F(t) (21)

where m is the craft mass.

4. Pitch equation

(155 + A55)t) 5 () + B55 f 5(t) + C!3,n

5(t)- 2pob X P,(t) = F5 (t) (02)

where I.,is the moment of inertia around the y-axis.

7
 For simplicity and without loss of generality hydrodynamic and hydrostatic coupling like Ar-,
B,- and CTterms and the coupling between the dynamic uniform pressure and the spatially varying
pressure is assumed to be negligible and hence set equal to zero in the model. The hydrostatic Ci- 'l
terms are found in the standard way by integration over the water plane area of the side-hulls. The
hydrodynamic added-mass coefficients A., the water wave radiation damping coefficients B,, and the
hydrodynamic excitation force in heave Ff'(t) and moment in pitch F'(t), are derived from
hydrodynamic loads on the side-hulls. The hydrodynamic loads on the side-hulls may be calculated
as presented in Faltinsen & Zhao (1991a, b), Faltinsen et al. (1991, 1992) and Nestegkrd (1990).
However, since the main focus in this paper is on the high frequency range, we have used a simplified
strip theory based on Salvesen et aL (1970) for calculation of the hydrodynamic loads. For simplicity
and without loss of generality the effects of transom stem and radiation damping are assumed to be
negligible in the hydrodynamic excitation forces on the side-hulls in heave and pitch. Hence
SkL
sin-U
= 2.eU kL-;(t) (C 33 - woýO A 33) sine)mt (23)
2

2 k2 L kL2
= ' (4 Cos -2 - _k'Lfsin 2 (C33 - corjm,A33)-
skU (24)
2- A 33
U- 2°kL 1cst
cosWr1

where d is the draft of the side-hulls. In the case studied here, the submerged part of side-hulls are
assumed to have constant cross-section area. Examples of two-dimensional frequency depending added-
mass and wave radiation damping coefficients are found in Faltinsen (1990). Constant two-dimensional
B,-and A, values are assumed. The high frequency limit of the two-dimensional added-mass coefficient
found in Faltinsen (1990) is used. The selected wave radiation damping coefficient in pitch
corresponds to the value at the pitch resonance frequency determined from structural mass forces
acting on the craft and hydrodynamic forces on the side-hulls. For heave we have chosen the wave
radiation damping coefficient at the resonance frequency that will exist without the presence of the
excess air cushion pressure. These simplifications are motivated by fact that the effect of damping is
most pronounced around the corresponding resonance frequency.

Discussion of the Mathematical Model

It is seen from equations (21) and (21) that the heave and pitch motions are coupled to the
dynamic excess pressure in the air cushion region. This is expected since the major part of the SES
mass is supported by the air cushion excess pressure- The dynamic air cushion pressure is expressed
as the sum of the dynamic uniform pressure and the spatially varying pressure.
An important question is how many acoustic modes should be included in the mathematical
model. Higher order acoustic modes with the resonance frequencies will be excited by sea waves
containing very little energy, and hence the dynamic response will be unimportant with respect to
established criteria for human comfort and crew workability according to ISO 2631. One should notice
that the air cushion dimensions and the forward speed are parameters that strongly affect the energy
level of the vertical accelerations caused by the acoustic resonances. The acoustic resonance
 frequencies are inversely proportional to the air cushion length as seen from equation (20). The wave
excitation frequency which is given by the circular frequency of encounter Co, = o), + kU, increase
.A! with the forward speed U. Thus waves of relatively low circular frequency wa. may excite the craft in
the frequency range of the acoustic resonances when the speed U is high. This may result in high
energy in the sea wave excitation around the resonance frequencies since the maximum sea wave
height'will tend to increase when the period of the sea waves increases.
The one-dimensional approximation of the spatially varying pressure is valid only for
frequencies below a certain value, given by min X,' . b and min X,; > h, such that max w, c 2T/clb
and max w), c 2tclh,,. Even if the solution is formally presented by an infinite number of acoustic
modes, the number of modes that is reasonable to include in the simulation model is limited by when
the acoustic modes cease to affect the response in the valid frequency range. For frequencies above
the specified max o.. it is necessary to take into account two- and three-dimensional-effects. Then a
more detailed numerical analysis is required, like for instance a boundary element or a finite element
method. The effect of the angle of the rigid stem seal relative to the wetdeck has been shown
numerically (Ulstein, 1991) by a two-dimensional finite element model of the air cushion to have little
influence on the first odd acoustic resonance when the ratio between the cushion length at the water
plane and the wetdeck is not too large.
The relative damping ratio Ej given by equation (19) is an important parameter. As expected
the leakage terms and the fan inflow term contribute to increased damping. One should notice that the
fan slope (aQ/ap)l,,. is negative. We also observe that the longitudinal location of the fan and the
louver systems strongly affects the relative damping ratio. In the case of a single fan system and a
single louver system, it may seem natural to locate the fan and the louver in the middle of the air
cushion, that is xF=xL=O. However, from equation (19) we observe that the relative damping ratio for
the odd modes will be reduced significantly if xL and xF is equal to 0. Maximum damping of both the
odd and even acoustic resonance modes in the case of a single lift fan system and a single louver
system is obtained for x. and xL equal to -L142 or ./2. The relative damping ratio of the first odd
acoustic mode on a 35 m SES with the ride control system off, will increase from about 0.05 to 0.2
by locating the lift fan system at one of-the ends of the air cushion instead in the middle. This means
that the passive damping of the first odd acoustic mode will increase with about 200 %. This gives
a significant improvement in ride quality even when the ride control system is turned off. In the same
manner the active damping due to the ride control system is maximized by locating the louver system
at one of the ends of the air cushion.
An estimate of the resonance frequency of the dynamic uniform pressure can be found by
neglecting the spatially varying pressure and the hydrodynamic and hydrostatic forces acting on the
craft. These simplifications are motivated by the assumption that the response in the vertical plane is
dominated by hydrodynamic and hydrostatic loads in the low frequency range and spatially varying
pressure in the high frequency range. In the intermediate frequency range the response is dominated
by the dynamic uniform pressure. The simplified equation of motion and dynamic pressure become
(K, s + K 3)p,.(s) + pom s lp3.(s) =
sin , (25)
p- .(s) + p 0 K2 AAj (x,5;s)
2

m s • fl. - App(s) = 0 (26)

From thc coupled heave and dynamic uniform pressure equations a second order equation appears with

9
the resonance frequency in [Hz] given by

fo =
2f S-1 "yA, (p.
mh.
+ p.)
.
(27)

with the relative damping ratio

o K 3K
4%f (28)
0 1

From equation (27) we observe that increasing cushion height h. leads to decreasing resonance
frequency. Increasing the mean cushion pressure p.or the craft mass m give the same result. The
relative damping ratio will increase by increasing the equilibrium air flow Q. and by increasing the
fan slope (aQ1ap)I, The fan slope will increase if the fan characteristic curve illustrated in Fig. 2
becomes less steep.
Data for the 35 m SES are given in Appendix A. In the following we will use a finite number
k acoustic modes in the mathematical model. Higher order modes are neglected.

3. ROBUST DISSIPATIVE CONTROLLER DESIGN

In this section we develop and analyze a fide control system based on the mathematical model
derived in the previous sectionr The objective of the controller is to damp out pressure fluctuations
around the equilibrium pressure p. in the presence of sea wave disturbances. This can be formulated
in terms of the desired value of the nondimensional dynamic uniform pressure p 'et) = 0 and the
nondimensional spatially varying pressure p1,A(x,t) = 0,where the superscript d denotes the desired
value. The number of modes to be damped depends on the requirements related to established criteria
for human comfort and crew workability.
The mathematical model of the craft dynamics is of high order as it contains a high number
of acoustic modes. A practically implementable controller has to be of reduced order. When designing
a controller based on a reduced order model, it may happen that the truncated or residual modes give
a degradation of the performance, and even instability of the closed loop system. This is analogous
with the so-called spillover *effect in active damping of vibrations in mechanical structures (Balas,
1978). The inadvertent excitation of the residual modes has been termed contol-spillover, while the
unwanted contribution of the residual modes to the sensed outputs has been termed observation-
spillover, see Fig. 4. This problem was also discussed by Gevarter (1970) in connection to the control
of flexible vehicles. Mode 0 in Fig. 4 is related to the uniform pressure, while the higher order modes
are related to the spatially varying pressure. The controller must be robust with respect to modelling
errors and parametric and non-parametric uncertainties. nonlinearties in sensors and actuators and
component failure. The use of collocated compatible actuator and sensor pairs pmovides a design
technique to circumvent these problems.

10
 Control Spillover.S d 6 ObServation Spillover

Mode k

C
XS
Input • o rol ,SCnUoc Ourpot

Fig. 4 Observation and control-spillover. where CtXL. cos Ic(XL+Lf2)/L and Cdx, = cos kit(x,+Lf2)/L.

State Space Model..

The dynamic system given by equations (9)-(24) is of third order as opposed

vibration damping problems of large flexible space structures that can be to similar
written as an equivalent
second order mass, damper and spring system. see e.g. Joshi (1989). Stability
of the control system
is analyzed using Lyapunrn theory. The dynamic system given by the equations
(9)-(24) is written in
standard state space form
±c = Ax + flu + Ev (9

where the n dimensional state vector x4t) is

r = [1 3 ,'15 .1 3 ,fl.p .p1 .p,......p,pt' 2 .....*P4~ (30)
where n3(/), rj~t), pj(t,prt), p2(t) ...pjt) arc defined as in the equations (9), (14). (17), (21) and
(22). k is the number of acoustic modes. vQ) is the 3+k dimensional disturbance
vector defined as

v(t) = [F;. r;. %/o.f ". V .... (31)

where F,'(r) and F,'(t) arc de~fined in the equations (23) and (24). The time
derivative of V0(t). V1(t).
V(t) arc defined in thc equations (I11). (16) and (1 8). uQt) is the r dimensional
...
control input vector.

Ii
and r is the number of louvers. The elements of u(t) are for i = 1, 2...r defined as

ui(t) = AAIRO(xi.t) (32) t

where AAkC'(xj,) is defined in equation (8). The louver and sensor pairs may be distributed along the
air cushion, preferentially in the longitudinal direction. The pressure sensor i is located at the
longitudinal position x,, and the louver system i is located at the position xtv. y(Q) is the m dimensional
measurement vector and m is the number of pressure sensors. The symbolic expressions for the nxn
system matrix A, nxr control input matrix B, nx(3+k) disturbance matrix E and mxn measurement
matrix C are found in Appendix B.
We will first consider the case where the sensors and actuators are ideal, that is linear and
instantaneous with no noise. It is assumed that the control input matrix B can be chosen such that

C = Brp (33)

where P is a nxn diagonal positive definite matrix providing correct scaling of the BT matrix to the
C matrix. This is referred to the case when there is perfect collocation between the sensors and the
louvers, i.e. xL; = x,, for all i and r = m.
Let s be the differential operator. It can be shown that the pair (A,B) and is controllable and
the pair (C,A) is observable. Hence, the dynamic system given by equation (29) can be represented
by
y(s) = H,(s)u(s) + Hd(s)v(s) (34)
= y.(s) + y,(s)

where the transfer matrices H,(sY-and _H,(s) are given by

H,(s)" = C(sl - A)-'B (35)

H1(s) = C(sI - A)'E

and 14 is the nxn identity matrix-

Let the the controller be defined as the linear time-invariant operator H, between the input
y(t)=Y(--+y,() and the output uQ). The transfer matrix of H, is denoted H,(s). The controller given
by He(s) stabilize the process H,(s) if the controlled system consisting of stabilizable and detectable
realizations of H,(s) and He(s) in standard feedback configuration (Fig. 5) is asymptotical stable when
v(t)=O.

U Yu
PP

HGC yv V

© pHd
Fig. 5 Feedback system.

12
Proportional Control Law
A proportional pressure feedback controller of dimension rxr is proposed according to
u,(s) = He(s) y(s) (36)
H(s) = GP

where G, = diag[g•j > 0 is a constant diagonal feedback gain matrix of dimension rxr. The control
law provides enhanced damping of the pressure variations around the resonance frequencies. The
diagonal feedback gain matrix G, can be determined for instance by pole placement techniques.
The stability proof of the control system is based on Lyapunov's direct method applied on a
linear time-invariant system. The closed-loop system is bounded input bounded output (BIBO) stable
if the equilibrium point x. = 0 of the autonomous closed-loop system, that is vQt) = 0 in equation (29),
is asymptotical stable. Applying the proportional output feedback control law given by equation (36),
the linear time-invariant autonomous closed-loop system becomes
I = (A -BGPC)x = Ax (37)

where the nxn closed-loop system matrix A&= A - BGC is defined in Appendix C.

We can now state the main result of this section.

Theorem 1
The closed-loop system given by equation (37) is asymptotical stable.

Proof:
Define the Lyapunov function candidate

- V(X) =.. x > 0 (38)

*2

where the nxn diagonal positive definite matrix P, is given in Appendix C. V(x) is positive definite.
The time derivative of V(x) is

V(x) = .xT(ActP
2 xf Qx
+ P tA,)x= -. 239 (39)

where the nxn symmetric positive semidefinite matrix Q,, is given in Appendix C. From the equations
(37) and (39) it is seen that
V'(x) = 0
II (4O)

X = X. = [13' 715 0. 0.0. PP...p,.0.0... 00 r

However, from equation (39)

l= 1 = z= / = P2 = = = 0 (41)

13
only if

= =
13 = 15 = P1 =P2 -- -P, 0 (42)
Hence, by the invariant set theorem (Vidyasagar, 1993) the equilibrium point of the closed-loop system
x. = 0 is asymptotically stable and the result of Theorem I follows.

The property of perfect collocation between the sensors and the actuators do not exist in
practice. Since we do not want the'measurements to be influenced by the local flow characteristics
around the vent areas to the louvers, it is necessary to locate the sensors in some distance from
the
louvers. This means that the mode shape functions at the louver and sensor locatibns will
not be
exactly the same. For the acoustic resonance modes of practical interest this may not be
of any
problem due to the long acoustic wave lengths relative to the imperfection in collocation between
the
sensor and actuator pairs. However, this claims that the sensor and actuator pairs are located
in some
distance from a node (Fig. 6). If the vent valve and the sensor are located close to a node, the
vent
valve and the sensor may be located on each side of the node. This may lead to spillover problems
according to Fig. 6. where the mode shape function associated to the sensor will have opposite
sign
compared to the mode shape function associated to the actuator. This is similar to positive feedback.
In Sorensen (1993) it is shown that some imperfection in the collocation will be tolerated
without
violating the stability properties of the closed-loop system.

Actuator

rW
-L/2 L/2
Sensor
Fig. 6 Noncollocated sensor and actuator pair.

The feedback system illustrated in Fig. 5 is BIBO stable. Hence,stability of the closed-loop
system using collocated sensor and actuator pairs is maintained regardless of the number of
modes,
and regardless of the inaccuracy in the knowledge of the parameters. Thus the spillover problem
is
eliminated and the parameters do not have to be known in advance to obtain stability. Notice that
there
are no restrictions to the location of the collocated sensor and actuator pairs with respect to
stability.
Hlowever. optimizing the performance, the longitudinal location of the sensor actuator pairs is
crucial
as seen in equation (19). Robustness with nzspect to unmodelled dynamics an sector nonlinearities
in
the actuators are shown -in Sorensen (1993) using the theory of passivity.

14
 4. SIMULATION EXAMPLES AND FULL SCALE RESULTS
In this section numerical simulations and results from
advancing forward at high speed in head sea waves full scale trials with a 35 m SES
will be presented. The effect of collocation and
noncollocation of the sensor and actuator pairs for the
35 m SES will be investigated. The SES is
equipped with one single fan system and two louvers
and two pressure sensors. Main dimensions and
data of the SES craft are given in the Appendix A. The
number of acoustic modes considered in the
simulation model is four, i.e. k = 4.

Numerical Simulations
Fig. 7 shows the Bode plot of the first diagonal entry
in Hf(im,) between the pressure sensor
y,(s) and the louver u,(s) when the two sensor and
actuator pairs are fully collocated. Sensor I and
louver I are located at the fore end of the air cushion,
while sensor 2 and louver 2 am located at the
aft end of the air cushion. When the frequency of encounter
goes to zero, the dynamic pressure tends
to a static value proportional to K1K . This indicates
2 that the equilibrium pressure p0 will decrease
when the equilibriurn leakage area increases and vice
versa. Around 0.1 Hz there are almost no
response. This is related to the structural mass forces
acting on the SES and the hydrodynamic forces
acting on the side-hulls. The high value around 2 Hz
is due to the resonance of the dynamic uniform
pressure. The high values around 6 Hz, 12 Hz, 18
Hz and 24 Hz are related to the four acoustic
resonance modes. From the phase plot we observe that
the phase is varying between 900 to -90. in the
whole frequency range. This is expected using
collocated sensor and actuator pairs. The second
diagonal entry in H,(io,) behaves in a similar manner
and is not presented here.
Fig. 8 shows the Bode plots of the two diagonal entries
of Hi(iw,) for the case where the two
pressure sensors are located at the fore end of the air
cushion while the two louver systems are located
at the centre of pressure and at the aft end of the air
cushion respectively. From the magnitude plot
of the first diagonal entry we observe that there are
almost no peak values around 6 Hz and 18 Hz.
Since louver I is located at the centre of pressure, it
will have no influence on the first, third and the
higher order odd acoustic resonance modes. From the
phase plot of the first entry we observe that the
phase is decreasing below -90* around 7.Hz. At the
second acoustic resonance frequency around 12
Hz, the sensed pressure signal at the fore end of the
air cushion is 1800 out of phase compared to the
pressure signal at the centre of pressure where louver
I is located. From the second diagonal entry of
lHý(ico) referring to the sensor at the fore end of the
air cushion and louver 2 at the aft end of the air
cushion, we observe that the sensed pressure signal at
the fore end is 1800 out of phase compared to
the pressure signal at the aft end where louver 2 is located.
This is expected as long as noncollocated
sensor and louver pairs are used. Noncollocated sensor
and actuator pairs introduce loss of phase and
may lead to instabilities in the implemented controller
if no bandwidth limitations are defined.
Fig. 9 shows the transfer function of the nondimensional
spatially varying pressure at the aft
end of the air cushion when the ride control system is
turned on and off, and when the two sensor and
actuator pairs are collocated and noncollocated with the
feedback gains given as g,, = g,2 = 1. In the
latter case louver I at the fore end of the air cushion
is reading the sensor signal 2 at the aft end of
the air cushion, while louver 2 at the aft end is reading
sensor signal I at the fore end. The ride
control system contributes to reduced response around
the resonance frequencies when collocated
sensor and actuator pairs arm used. However. in the case
of noncollocated sensor and actuator pairs the
response around the first odd resonance frequency increase
significantly. This is expected since the
ride control system in this case reduce the relative damping
ratio. see equation (19).
Fig. 10 shows the transfer function of the vertical
acceleration at the stem when the ride
control system is turned on and off, and when the two
sensor and actuator pairs are collocated and
noncollocated with the feedback gains given as g, =
g,2 = I. The same tendency as mentioned above
is also secn in the acceleration amplitude. Notice that
the dynamic uniform pressure is not directly
influenced by using collocated or noncollocated actuator
and sensor pairs.

15
 Fig. 11 shows how the ride control system affects the operability limits on a 35 m SES.
Reduced comfort boundary due to vertical accelerations at the stem related to evaluation of human
comfort and crew workability is used. The criteria used for the operational limit is the International
Standard, ISO 2631/1&3 (1985). This standard considers the effect of vertical acceleration on human
perfornance and comfort. It has established criteria for two different frequency bands. The high
frequency band, which is of most importance in this analysis, is based on three different limit levels.
which are:
1) Exposure Limit. giving the limit which can not be exceeded without reduced safety.
2) Fatigue-decreased Boundary, related to maintaining the working efficiency of the
crew.
3) Reduced Comfort Boundary,.related to evaluation of passenger comfort-

The main motivation for this study is comfort problems, so here the Reduced Comfort
Boundary is used. The lower frequency band (0.1 - 0.63 Hz) in the standard is related to severe
discomfort, such as motion sickness. The lower frequency band is of less importance in our study,
because it occur in a frequency band too low to be considered in the present case. Both frequency
bands are dependent on exposure times. Here 4 hours is used. The limiting significant waveheight H,
as a function of peak period T, is specified. A Pierson-Moskowitz type of spectrum is used for all sea
states. The maximum possible significant wave height is for low peak periods limited by the criterion
for maximum wave steepness, given in the figures as max H,. The maximum value for H, is specified
by

1H2gT (43)
10 2Y

The ride control system configuration is the same as used in Fig. 9 and 10. It can be seen that the ride
control system is most effective in the frequency range dominated by uniform pressure variations. The
limiting significant wave height is almost doubled by activating the ride control system. The response
reduction is also significant in the frequency range dominated by spatial pressure variations. The strong
adverse effect given by the noncollocated louver and sensor pairs in the frequency range dominated
by spatial pressure variations is also shown in Fig. 11. This clearly stresses the importance of using
collocated sensor and actuator pairs in the control system.

.-4

Fig. 7 Num. calc. Bode plot of the first diagonal entry of the 2x2 transfer matrix 11ý(iw,)'.XL,=x,t=12
xo=xo=-l2 t. in x,=6 m, U=50 knots. p,=500 mmWc.

16
 b".~~oo 1111 otot

-r to, 30I -

Hz

30

0o to o o o o 10Dt- t- c o o

Fig. 8 Num. caic. Bode plot of the diagonal entries of the 2x2 transfer matxix H~(icn); x,,=x, =12 m,
2
x~rpO m, xti=-l 2 m, x,=6 m, U=5O knots, p.=500 mmWc.

2 -- - - - .. ... .. .. -- - - .......

0 2 4 6 8 10 12 14
Frequency of Encounter [HZ]

-FIGS oif C-ollocated - o-- icae

Fig. 9 Num. caic. transfcr function of the nondimcnsional spatially varying pressure at the aft with
a) RCS off: b) RCS on: ,,I=8,2=i. x~=.= 2 m. x 0 =x,j=-12 m: dJ)RCS on: g,,=g,, =1.
X,.,=X,2=I2 m, x,,=x,,=-12 m.-

17
 100

1i
7840 ....... ..
. . . . .. . ..... ......... . ........... . . ................
60 j.' . .. ........ ..... ........ . - ....... ...
.................. . . .......
..............

40 , ... . ...... ................. , . .... .f. .. ...... ..i.................i ... . ...........

hf.................

30 - ... . ..............
.. . .. ---- -- , .........
-_- -. ... . . .. ........... .... .............
20 IV• .... ....... :
120

C.
0 _- 4 6 8 10 12 14
Frequency of Encounter [Hz]

- RCS off -collocated - Non-collocated

Fig. 10 Num. calc. transfer funcdon of vertical acceleration at the stem with a) RCS off; b) RCS on::
g,,=g,,y=1, xL,=x,)= 12 m. X,2=x,,=-1 2 m; c) RCS on: g,1=g, 2=1. xL,=x,7=l2 m, x,=x,t=-12 m.

Vertical Acceleration at AP
0.5

04..... . \.-.. .: - .... ............... ....................... ......... ...... .......... ....................

"r
.....• i -- '," "• ............ .. ..... ......:......... ..... •..... .......... -............ ............. .....
So I . " "
...
... i:; ..
.....

*~01

1.5 2 2.5 Tp 3' 4' 0

Peak Period Tp[sec]

rCS onf ROSo - -MW HS -0 csro-W,Ccc:e

m 1t
Fig. I I Operability limits for a 335m SES based on the ISO 2631 "Reduced Comfort Boundary', head
sea; a) RCS off: b) RCS on: g,,=g, 2 =I. x,,=xt,=12 mn. x,,=x.2 =-l 2 in; c) RCS on: 9,92
.x,
 Full Scale Results
The prototype ride control system used in the full scale
experiments is based on the passive
controller as presented in Section 3. The control algorithms
-a' in the ride control system were partly
implemented on a personal computer (PC) and partly
analog hardware devices were used. An outer
feedback loop was implemented on the PC. while a
faster inner feedback loop around the electro-
hydraulic louver system was implemented by analog
hardware devices. The lover system consisted of
two vent valves located at the same longitudinal position
x., = 8 m side by side. The two vent valves
were operated in parallel in the outer feedback loop.
This means that the inner feedback loops around
the two vent valves received the same reference signal.
Two pressure sensors located atx,, = 10 m and
X,2 = -10 m were used to measure the excess
pressure variations in the air cushion. One accelerometer
located about 5 m aft of the centre of gravity were
used to measure the vertical accelerations. The
analog sensor signals were prefiltered by an analog
presampling-filter and subsequently transmitted
through the analog to digital (A/D) converter to the PC.
The digitalized sensor signals were then used
by the outer control algorithm, which calculated commanded
algorithm around the louver system. The commanded reference signals to the inner control
reference signals were transmitted through the
D/A converter to the electro-hydraulic louver system.
The inner analog controller loop around the
louver system provided the necessary opening and closing
actions of the vent valves. The experiment
arrangement is illustrated in Fig. 12.

Acceleration

pap0

Fig- 12 Experiment arrangement.

The full scale measurements were carried out in

estimated to vary between 0.3 - 0.6 m. No measurements sea states with significant wave heights
of the sea wave spectra were done. The
estimated significant sea wave height H, and direction
were based on visual observations. The power
spectra of the vertical accelerations with and without
the ride control system will be presented in the
same figure. The two power spectra am calculated from
two time series each of 60 s, where both time
series were stored within a time interval of 3 minutes
running the craft in open sea in order to have
comparative sea states.
Fig. 13 shows the full scale power spectra of the vertical
accelerations about 5 m aft of the
centre of gravity with and without the ride control
system activated. With the ride control system
turned off, we observed response around 2 Hlz, 5 IHz
and 8 Hz. The response around 2 ttz is related
to the resonance of the dynamic uniform pressure,
while the response around 5 Hz and 8 Hz are
related to the first odd and even resonance modes.
Activating the ride control system the response
around 5 Ilz was significantly amplified, while the
response around 2 IHz was reduced a little biL In

19
this case the pressure signal at x,2 = - 10 m was used in the feedback loop. Hence,
the actuator and
sensor pair were completely noncollocated since the louver was located at xL-., = &m.
This means that
the pressure signal at the senor location was 1800 out of phase compared to the actual
pressure at the
actuator location in the frequency range dominated by the first odd acoustic resonance
mode. The
response around 8 Hz was more or less unchanged. Both time series were taken when
the craft was
advancing forward with the speed U = 45 knots in head sea waves with significant
wave height
estimated to be H, = 0.3 m. The equilibrium air cushion pressure was p.= 450 mmWc.
Fig. 14 shows the full scale power spectra of the vertical accelerations about 5 m
aft of the
centre of gravity with and without the ride control system activated. In this case the
pressure signal
at x,, = 10 m was used in the feedback loop. Hence, the louvers and sensors were "almost"
collocated
since the louver were located at xL, = 8 m. With the ride control system turned off,
we observed
response around 2 Hz, 5 Hz and 8 Hz. Activating the ride control system the response
around all three
resonance frequencies were significantly reduced. These time series were taken when
the craft was
advancing forward with the speed U = 41 knots in head sea waves with significant
wave height
estimated to be H, = 0.6 m. The equilibrium air cushion pressure was p. = 430 mmWc.

[nJ
2 0.6

0.7
___________________________

0.6 Ride control system ON

0.5

0.4

0.3

0.2 ,Ride control system OFF

0.1

S1. 2. 3. A. 5. 6. 7. 8. 9. 10.

[Hz]

Fig. 13 Full scale power spectra of the vertical accelerations at the x=-5 m of a
35 m SES with
flexible stem seal with ride control system on (noncollocation) and off: p,=450 mmWc,
U-=45
knots. H,--0.3 m, head sea waves.

20
 0 .35

Ride control system OFF

0.2
0.25

0.25

Ride control system ON

0.i.//

2. 4. 6. 8.
8. t.
10.

[Hz]
Fig. 14 Full scale power spectra of the vertical accelerations at the x=-5 m of
a 35 m SES with
flexible stem seal with fide control system on (collocation) and off, po=430 mmWc,
U=-44
knots, H,=0.6 m, head sea waves.

S S. CONCLUSIONS
The pressure variations in the pressurized air cushion of a SES have two fundamental
characteristics. The pressure variations were divided into a dynamic uniform and a
spatially varying
pressure term. It was seen that the resonances of the dynamic uniform pressure
and the spatially
varying pressure caused excessive vertical accelerations when the craft was advancing
forward in sea.
states which contained energy in the frequency domains corresponding to the resonance
frequencies.
To achieve a high human comfort and crew workability, it is necessary to reduce these
accelerations
using a ride control system. A distributed ride control system has been developed and
a proportional
pressure feedback controller has been proposed. Full scale experiments of a prototype
ride control
system showed significantly improvement in ride quality using a ride control system
which provided
dissipation of energy around the resonance frequencies. The full scale experiments
also showed the
importance of using collocated sensor and actuator pairs in the acoustic dominated
frequency range.
Spillover effects like unwanted excitation of residual modes was avoided using collocated
sensor and
actuator pairs regardless of the number of modes considered and parameter values.

ACKNOWLEDGEMENT
This work has been sponsored by the Ulstcin Group and the Royal Norwegian Council
for
Scientific and Industrial Research (N-FNF)

21
 REFERENCES

BALAS, M.(1978). Feedback Control of Flexible Systems. IEEE Transaction of Automatic Control,
Vol. AC-23, No. 4, pp 673-679.

DESOER, C.A. and VIDYASAGAR, M.(1975). Feedback Systems: Input-Output Properties.

Academic Press. New York.

GEVARTER, W.B.(1970). Basic Relations for Control of Flexible Vehicles. AIAA Journal, Vol. 8,
No. 4.

FALTINSEN, O.M.(1990). Sea Loads on Ships and Offshore Structures. Cambridge University Press.

FALTINSEN, O.M., ,HELMERS, J.B., MINSAAS, K.J. and ZHAO R.(1991). Speed Loss and
Operability of Catamarans and SES in a Seaway. First International Conference on Fast Sea
Transportation -FAST'91, Trondheim, Norway.

FALTINSEN, O.M. and ZHAO, R.(1991 a). Numerical Predictions of Ship Motions at High
Forward Speed. Philosophical Transaction of the Royal Society, Series A.

FALTINSEN, O.M. and ZHAO, R_(1991 b). Flow Predictions Around High-Speed Ships in
Waves. Mathematical Approaches in Hydrodynamics, SIAM.

FALTINSEN, -O.M., HOFF, J.R., KVALSVOLD, J. and ZHAO, R.(1992). Global Loads on
High-Speed Catamarans. PRADS'92, Newcastle, England.

JOSHI. S.M.(1989). Control of Large Flexible.&Space Structures. Berlin Springer-Verlag.

KAPLAN, P. and DAVIS, S.(1974). A simplified Representation of the Vertical Plane Dynamics of
SES Craft AIAA Paper No. 74-314, AIAA/SNAME Advanced Marine Vehicles Conference,
San Diego, California.

KAPLAN, P. and DAVIS, S.(1978). System Analysis Techniques for Designing Ride Control
System for SES craft in Waves. 5th Ship Control System Symposium, Annapolis, MD.

KAPLAN, P., BENTSON, J. and DAVIS, S.(1981). Dynamics and Hydrodynamics of Surface
Effect Ships. SNAME Transactions Vol. 89.

NESTEGARD, A.(1990). Motions of Surface Effect Ships. A.S. Veritas Research Report No.:
90-2011.

SALVESEN, N., TUCK. E.O. and FALTINSEN, O.M.(1970). Ship Motions and Sea Loads.
Transaction SNAME. 78, pp 345-356.

STRANG, G.(1988). Linear Algebra and its Applications. Third Edition. Harcourt Brace Jovanovich
Publishers. Orlando, Florida.

SULLIVAN. P.J., GOSSELIN. F. and HINCHEY, M.J.(1992). Dynamic Response of an Air

Cushion Lift Fan. Intcrsociety High Pcrformancc Marine Vehicle Conference and Exhibit -
HPMV'92. American Society of Naval FEngincers. Washington D.C.

22
 SORENSEN, A.J.(1993). Modelling and Control of SES Dynamics in the Vertical Plane.
Dr. ing.
Thesis, Department of Engineering Cybernetics, the Norwegien Institute of Technology.

"• SORENSEN, AJ.. STEEN, S. and FALTINSEN. O.M.(1992). Cobblestone Effect on

SES.
Intersociety High Performance Marine Vehicle Conference and Exhibit - HPMV'92. American
Society of Naval Engineers, Washington D.C.

SORENSEN, A.J., STEEN, S. and FALTINSEN, O.M.(1993). SES Dynamics in the

Vertical
Plane. Ship Technology Research, the Journal for Research in Shipbuilding and Related
Subjects, Vol. 40, No. 2.

SORENSEN, A.J. and EGELAND, O.(1993). Ride Control of Surface Effect Ships Using Distributed
Control. IFAC 12th World Congress, Sydney, Australia.

ULSTEIN, T. (1991). A Numerical Analys's of the Acoustic Properties in the Air Cushion
to a SES.
Department dg Marine Hydrodynamics, the Norwegian Institute of Technology.

VIDYASAGAR, M.(1993). Nonlinear System Analysis. Second Edition, Prentice Hall, Englewood
Cliffs, New Jersey.

APPENDIX A. SES Main Dimensions.

Length overall 35 m
Equilibrium fan flow rate : 150 m3/s
Linear fan slope . -140 m 2/s.
Cushion length . 28 m
Nominal cushion pressure 500 mmWc
Cushion beam :s 8 m
Cushion height 2 m
Weight 150 ton
Speed 50 knots

23
 APPENDIX B. Symbolic Model Matrices

We define the n dimensional state vector x(t) as

X= [xtx',TxT]T (B.1)

where
x= [nlql.fl.n,.P].x = [p,.pl .. X
. ..... p,] (B.2)

where the definitions of x,(t), x,(t) and x3(t) are as given in equation (30). The nxn system matrix A
is then given by

[ SI ,6 05 . S25' ] (B.3)
A= Ow O~k 1W
S3ý, S4 W S5,k

where 05, 0,, and 0 . are the 5xk, kx5 and kxk zero matrices respectively. I,' is the kxk identity
matrix. The 5x5 dimensional SlJ5 , matrix is defined as

0 0 1 0 0
0 0 0 1 0
C33 0 B--- 0 A p,
m +A33 m+A 33 5 m+A33 (B.4)
S/s =
0 -IC 55 0 - B0
155 +A55 +A55
1,55

0 0 p,°A, 0 1K
3
Kt KI

The 5xk dimensional S25., matrix is defined as

0 0 0 0 ... 0
0 0 0 0 ... 0
S2 0 0 0 0 0 ... 0 (B.5)
di 0 di 0 d 5 ...
0 0 0 0 0 ... 0

24
 where

di. 2
2pb "pb (B.6)
155 +A 55 It

The kx5 dimensional S3, matrix is defined as

00 0 g
1 0

000 0 0
S3, = 0093 0 (B.7)

000 0
where

gi =
4p,,L c
Pohoj__)_
2
(B.8)
p./h.(jmi )2

The kxk dimensional S4,, and S5, matrices

are defined as forj = 1,2, 3. k

diag[-c
$ 1 ] (B.9)

S5, =diag[ 2c0 (B.10)

The nx(3+k) disturbance matrix E is defined
as

0 0 0 0 0 ..
0 0 0 0 0 .. 0
1 0 0 0 0 0

155+A~ss(
E 0 1 0 0 0 . (.11)
( .1

0 0 1 0 0 0
KI

0"3 W

25
The nxr control input matrix B is given by

0 0 ... 0
0 0 0
0 0 ... 0
o 0 . .. 0
p ,.
K2 pc0 K2 p..K
KI KI KI
0ý (B.12)
I LI 2 J L 2 C

CICOS.L1 1 +.....L CICOs27 X u+& c icos2 ( &+ ŽL

L
(0 2 LL2
0 0L
eut matrix C is given by

I.1 c cs.1 1
.0 0 . .. 0
o 0 . .. 0
o 0 . .0

-T Cos+71(xI + J cos+it(x,2 +A'L . . . Cos it(x, +ŽJ (B.13)

2osnkx L+2 7cs 2 L) . . .co2 +L

L 227. )LI2

26
 APPENDIX C. Symbolic Matrices Used in Stability Proof
The nxn closed-loop system matrix A = A - BGC
4 is defined as

AaO,,
($15,s-C15.) 0,•I•1
05,, (S2 5,-C2 1 9ý ) (C.l1)
(

($3,,-C3,6) S4,k (SS•,-C4,,)

where the nxn system matrix A is defined in equation

(8.3). SJsd. S2SX, S3bs S4•, and S5,, are
defined in [he equations (B.4) - (B.10). The 5>5 dimensional
CJ.5 . matrix isdefined as

00000 0
0000 0
0000 0
C1X 00 0 00 0(
0 0 0 0 -P-"2- g,

The 5xk dimensional C25 ,,matrix is defined as

0 0 0 0 0 ... 0
0 0- 0 0 0 ... 0
=C2 0 o0."0 0 0 -..0 (C.3)
0 0 0 0 ...
hi h 2 h 3 hý h 5 h,

where
h.= p---• g2cos-L TLK (C.4)

7 7

27
The kx5 dimensional C3., matrix is defined as

0 0 0 0 ti
0 0 0 0 t:

C3, 5 = 0 0 0 (C.5)

where

tJ = c, g9cosSx U+A
j7,' (C.6)

The kxk dimensional C4, 1 matrix is defined as

U11 u2 3 .. . u

U21 U-MU23 . ..uý

C4 , = u31 u32 u 33• . u3* (C.7)

LI
. *U .L 3 . .. t
.).2 Uk .

where

u-, = cI F os
CO + cCOSfr x+Li , m, n = 1, 2, k (C.8)
I L 2 L 2)
If the sensor and actuator pairs are fully collocated x. = x,1 and r = rn, then C4., is symmetric and
positive semidefinite. This follows since all the underdeterminants of order 2 are zero according to

P1 2 2 C pt L LLq 7c

ICOS P~r x" + i - s

= o Lt,
p0 LL
2
r
L/
L +rL
2fr
LL
Lk2, )J 1
Hence. by induction all principal submatrices (Strang, 1988) of C4,, with order Ž_ 2 have zero
determinants.

28
 The nxn symmetric Qa, matrix is chosen as

o0 0o.3 02 o0
03,4 QI;,0 03, Q23',A
=
Q Okl a 0•,3 (C.10)
OW "0

ow 0. Q3L

The 3x3 dimensional QI, o " matrix is (the subscripts are in accordance with the dimensions of the full
Q,, matrix)

= diag[q .] i = 3. 4, 5
2 B '3 Pz 2 B SS
q 33 =
p3 2 q• / +.(C. 11)
+A33 1
T 55+A

q=5 P - K3 + P,,K 2 g•

The 3xk dimensional Q2,J

3 matrix is
0 0. 0 . .. 0 ]

Q23' k = 0 •0 0 .. . 0
•qQ(2. qss.&
, 2 q;(s.. 3 ) . q(s.k ) (C.12)

q5( 5-k--) =) - h5 -. P5t k.i ti.

where hi and t, are defined in equations (C.4) and (C.6). The kxk dimensional Q3,,1' matrix is
Jc

q(s5-k.I)(j-k-z)
C
q(•s-k.?)(s-&.l)
C
•q(--ls.*
C

Q3L
(C.13)
C
cC

q(`S.,k •.5.j -- P5C.,.= 2 (2• + u,,)

q!!.t.i)(5.k.o = P5., .iu + P5 .,.i ui . i. j = 1,2,..-.k

If = x,;. Q3,,, is a symmetric positive definite matrix since C4,XI' is positive semidefinite and -S5,
is positive definite. see equation (B.10). IHence. the nxn symmetric Q, matrix is positive semidefinite

29
according to Strang (1988) since

Q, = RJ Rd (C.14)

where Rd is a matrix of dimension (3+k)x(5+2k) found to be

R [0330 ,= '5L o(Q'34" Q23a1 (C.15)

%a Ow 0. 'Q2;, 1(Q2;(Q

It can be shown by inspection that

QI1J," > 0

Q3L. > 0 (C.16)

Q3L_ -(Q23;.,)Q13;3) Q2;, > 0

The nxn diagonal positive definite P,, matrix can. be found from the Lyapunov equation
(AdTPa + PA4 = -Q0) to be

=Xag[P,'] i = 1,2,3.... .,, j = 1.2,3.... k

P, = C 33 p PC 5 P p `° (me+A 33 ) (C. 17)
me+A33 - -. +A5 5 3 = K5 p°

ggP , KC "" ; p, 1

30
 Bode Plot -Collocation

100 Bode plot- lyiu

10-1

10-2-

10-3 10-2 10- 10 10' 102

Hz

200 Bo~de plot - Ph .aseiylu)

100-

w- 0

-00 h

-1001

10310-2 10-1 100 101 102

11Z
Bode Plot -Noncollocation

100Boepo.y/I

10-1

10-2

10-3
10-3 10-2 10-' 100 101 102
Hz

200 ~Bode plot - Phaseu

100-

-200-
-300
10-3 10-2 10-' 100 101 102
Hz
Spatially Varying Pressure

* It

CDC

---------------------------------.----------------------- t............--
21 .5......---_----.......... ..................

01
E 6 0 2 1

RCS .... Jf
olctdN ncloae
 Vertical Acceleration

100-
90....--------------
0
98 - ---------- ------

7E- - - -------------------------------
1------------...-------

------------------------------------ -------
t----
80 - ----------------
*
C
1 ---------------
--------- ;.......... .........
70
ca---------------------- -----------

40 . . . .------ ------------ ....... .....................

.........

Fj~~~~~ ~ ..........-------
50.........

0
30 - 2-------
of..Encounter..H..
Frequency------------

off
ROSFeuec Co no unted Na-cllzae
 Operability Limits
Reduced Comfort Boundary

35 m SES
Vertical Acceleration at AP
0.5
E
.. I. ...
... ...
...
--....
...
...
.......... ... -- ------
---- .. ---
...
...
------
...---
------
...
... --
---
... --...
...
--
... ..
-rA
J--
0)

----
.----------
..--
---
-.--
•:0.2 .....
............
.... ................................
---...
0O.2
r)
C5
0-

1 1.5 2 2.5 3 3.5 4 4.5 5

Peak Period Tp [sec]

RCS offl .-- RCS on --- Max Hs --- RCS non-collocated

 Operability Limits
Reduced Comfort Boundary
35 m SES
Vertical Acceleration at AP
0.4

i...
i....
...
i.....
i.......
.........
------------
-..----
-.---
0.....4
Cz

0) 0

1 1.5 2 2.5 '

3 3.54 5 5
Peak Period Tp [sec]

--- RCS off -" Gain=1.0 + Gain=5.0 - Max Hs

 Root-locus Plot
I I wI
Io I
... aýode l. 40
150 ... . ....

0 ....
1 O........ .................. ................ ................. :......... . ... .. ... ....
................ ...
........

lMode 3 ÷+÷÷+

Mode 2+

5 0 . .. ..... ...... .. ............ .. . .............. ...... . .

:÷
mI Mode 1 ' ÷ ÷ +2 +÷ ÷
m + ++

0 ... :
. . .... ...... ..
. . . . . . . . . . . . . . . . .......
........... ....... . . ........... - , + + + .....
..... .* . .. 1 -
g

:Mode 1M o e l+ +: ++ +
++.. .. ..

- 50 - ..... .....
...
... .. ... .. ..... ........ ....... ..... .... ... .. ... ..

:Mode 2

Mode-4

-40 -35 -30 -25 -20 -15 -10 -5 0

Real Axis
Prototype Ride Control System

Acceleration

PPALouver

P PF
 Power Spectrum - Pressure FP
Noncollocation

250000:

200000.

150000.

100000.

50000.

0.-2
2.4. 6. 8. 10.
Power Spectrum - Vert. Accel.
Noncollocation

0.8"

0.7-

0.6-

0.5-

0.4-

0.3

0.2-

0.11
2
4.
.
1. 2. 3. 4. 5. 6. 7. B. 9. 10.
Power Spectrum - Pressure FP
Collocation
200000.

180000.

160000.

140000.

120000.

100000:
60000.

60000.

40000.

20000.

0: -

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
Power Spectrum - Vert. Accel.
Collocation

0.4

0.35

0.3

0.25

0.2

0.t15

0.1

0.05.L

4. 6. 8. 10.
2.
 CONCLUSIONS

* Pressure variations in the air cushion

dominate the craft behaviour.

* Pressure variations in a pressurized air

cushion appear as:
dynamic uniform pressure.
spatially varying pressure.

* Achieving high comfort it is necessary to

use a ride control system.
* Collocated sensor and actuator pairs
provides robust stability and high
performance.

* Optimal longitudinal location of lift fan

system and louvers improve the ride
quality significantly.
 Ride Control of SES

Asgeir J. Sorensen

Outline

* Background / Motivation.

* Mathematical Modelling.

* Control System Design.

* Simulations and Full Scale Results.

* Conclusions.
Surface Effect Ship

OUTBOARD PROFILE

INBOARD PROFILE
 SES Concept

* Air cushion:
excess pressure, 80 % of buoyancy.
dominate the craft response.

* Side hulls:
similar to conventional catamarans.
20 % of buoyancy.

* Seals:
- flexible (rubber)/ composite material.
bow seal.
stem seal -> bag filled with air.
-> rigid panel.
* Lift fan system provides pressurized air
supply to the air cushion:
centrifugal (radial) fan.
axial fan.
bag booster fan.

* Ride control system:

vent valve control (louvers).
fan inlet (area) control.
fan pitch (blade) control.

* Propulsion system:
- waterjet.
- propeller.
 SES Features

* Reduced calm water resistance:

high speed.
low fuel consumption.

* Reduced motion in low frequency range

(seasickness!).

* High frequency vertical accelerations

(comfort problems?).

* Wave induced forces (80 %) imparted as

dynamic air cushion pressure acting on
wetdeck.
 Problem

* Large vertical accelerations

experienced.

* Reduced passenger and crew comfort.

* How to reduce the vibrations and

what is the drawbacks?

- complexity.
- safety.
- loss of air.
- expenses.
 CC 04
S .COO
0c a cd
- 0>
_____ !34

-g
-~ ~5C4)
 Pressure Characteristics

* Two fundamental pressure characteristics:

uniform pressure in frequency range

[0, 5 Hz].

spatially varying pressure in

frequency range [5, ->].
 Cobblestone Effect

* Dominating vertical accelerations around

two frequencies below 10 Hz:

resonance of the dynamic uniform

pressure (2 Hz).

- resonance of the first odd acoustic

mode (5 Hz).
 Mathematical Modelling

* Linear one-dimensional model in heave

and pitch considering:

dynamic uniform pressure.

spatially varying pressure.

* Focus on high frequency accelerations:

- simplified hydrodynamics.

* Seal dynamics are neglected.

 Coordinate Frame

zy 11

7x
7 V7
115

>
V
* Total pressure variations in air cushion:

P,(x,t) = Pa + Po + P"(t) + P,(xt)

where
Pa - atmospheric pressure

P0o equilibrium excess pressure

pu(t) -. uniform excess pressure

psp(xt) spatially varying pressure

* Non-dimensional uniform pressure
variations:

Sp(t) - po
PO

* Non-dimensional spatially varying

pressure:

(x , t) = P s (X0, 0
• PO
 Modal Solution

* Modal solution of the boundary value

problem:

IIp,(xt) = pj(t)
O rj(x)
j= 1

r,-(x) ,cxs ]
- -2'

where
pj(t) = modal amplitude function
rj(x) = modal shape function
 Spatial Pressure Equation

l0
2

pi + 2 j6j + (ajpj forces

where the acoustic damping ratio is:

ý = ao + a 1COS2J (LI
XRCS + -2L
+ JL 2)
+ a 2 cos2Je (xF+ )

and the eigenfrequency is:

j L
 Control Objective

Obtain high ride comfort.

0 Damp out pressure variations around

equilibrium pressure p0.

* Air flow control:

- vent valves (louver system).

* Robust performance and stability.

 Spillover

k+ 1 _ _____k+1

CL Moe+Sx
--------------------------------------------------- ------

* 0

*C 0

C~~ XL-*Moe

1 ~ModeO 1
---------- --------------------------------------------

Controlled mnodeý

Control Input Sensor Output

Controller
 Feedback System

Process
u [Yu

ucyv V
Hc Distubanc

Controller /Disturbance/
 Control Algorithm

* Proportional pressure feedback controller

provides dissipation of energy around
resonance frequencies:

u(t) = -G, (ji(t) + sps,t(xJ))

* Collocated sensor and actuator pairs

ensure stability.

* Longitudinal location of sensor and

actuator pairs important achieving high
performance and stability.
 Imprecise Collocation

Actuator

-L/2 L/2
Sensor
MOTION CHARACTERISTICS OF FLOATING VESSELS.
MOTION TRANSFER FUNCTION SPECIFICATION, DIAGNOSTICS AND
TRANSFORMATION.

by Ivar J. Fylling, MARINTEK

4'

Wave

Wave

In :\•eok 5 1\p:lpt:r\vWtgtzcn 11.1J I i\iI

1992.09.25
1 INTRODUCTION

Motion transfer functions have proved to give an efficient description of the motion
characteristics of floating vessels. The transfer functions are calculated for the 6
motion degrees of freedom for a specified reference point on the vessel, either from
potential theory or from model test results. In the interpretation of motion charac-
teristics we normally focus on the amplitudes and often neglect the phase angles.

In practical use one will often have to calculate motions of other points on the
vessel, motions relative to other vessels, or motions referring to other coordinate
systems. This will require a combination of two or more degrees of freedom, and the
result will depend on both amplitudes and phase angles.

A lot of trouble has been caused by erroneous interpretation of transfer functions and
erroneous transformation. In many cases a simple checking of asymptotic phase
angles can be useful in verifying the actual interpretation. The present note is an
effort to describe the asymptotic phase angles, the different coordinate systems that
are often used, and to show a recipe for transformation of phase angles.

1992 25t:Ictlit.lJ
\.:9. I
I 992-(1it9.25 2
 2 TERMINOLOGY AND CONVENTIONS

Phase Angle If nothing else is said this means forward phase shift.

Backwards phase shift, delay, may be denoted "phase

lag".

Coordinate systems - All coordinate systems are right-handed cartesian sys-

tems.

Transfer functions - Physical relation between harmonic wave and linear

response:
X(t) = R,ý sin(wt +
The transfer function consists of an amplitude ratio.
R, = X.aC,
and a phase angle, *x-

Naming conventions

Vectorial components x1 x, x3 x4 x5 X6
Names of axes X Y Z
Wave particle motions I 7 ),2 Y3
Names of motions Surge Sway Heave Roll Pitch Yaw

Observe that the naming of motions refers to the coordinate system, regardless of
whether the X-axis is pointing "forwards" or "backwards" in a vessel.

m :;k5 I v rimillpt'\ r
l.l.\r.
199)20.1• 2 5
3
3 WAVE POTENTIAL

The wave potential 4, for a regular wave according to Airy's theory can be ex-
pressed as follows:

C1Cos(-ct + kxcosj3 + kysinp - 4¢) (3.1)

where
Ca is the wave amplitude
g is the acceleration of gravity
k is the wave number and
P is the direction of wave propagation. ( 3 = 0 corresponds to wave propagation
along the positive x-axis)
*; is the wave component phase angle
C1 is given by:

=CI- cohk(z+d)
cosh kd (3.2)

where d is the water depth.

In deep water C, can be approximated by

C1 = ekz (3.3)

We then obtain the following relations for the particle velocities and accelerations in
the undisturbed wave field:

v. = Ca w cosp C2 sina
Vy = w sinp C2 sina
wa
Vz = •C w C3 CoSa
(3.4)
a.= Ca cosP C2 cosa
a,= Qo 2 cosp c2 cosa
a, = w(2 C3 sina

1992 \l09.2
c 1\:.,lil.
I ()92.09(.25 4l
where

a = ((at - kx cos• - ky sinp + ac)

Using the deep water approximation we have:

C1 = C2 = C 3 = ekz (3.5)

Taking into account finite water depth we have:

C, cosh k(z+d)
cosh kd

C2 = cosh k(z+d)
sinh kd (3.6)

sinh k(z+d)
C3 = sinh kd

The surface elevation is given by:

= Ca sina (3.7)

It is important to notice these definitions.

Similarly the linearized dynamic pressure is given by:

Pd = -Pg (a C. sina (3.8)

1 :\,0c k 5i \li pc r5w ege iit I.J L'rh

1992.00.25 5
4 PHASE ANGLES OF WAVE PARTICLE MOTIONS

4.1 Particle motions

Particle motions are obtained by integrating the velocity functions, eq. (3.4).

t
=f v• • + x. = CacosP C 2 .(-cosa)
a

1 f Vd + yo = CaSinfl C.(-cosa)

C = f vzdt + z, = Ca C3 (-sina)

C, and C, are depth dependent functions, according to Eq. (3.6)..

The integration constants x., y., z. have been selected so that the average value is
zero.

Thus, the particle motions are expressed by

= CQ cosP C2 (-cos a)
71 = Ca sinp C2 (-cosa)
C= Ca C3 sina

:\.c9k 5 \pa
19 )cr\wcgcint.li Ir 0
19t.19.25 2 (.
4.2 Phase angles of particle translations

In nearly all contexts the surface elevation is selected as reference process when
describing waves and wave induced responses. According to the foregoing section the
particle translations can be written.

C = C. C3 sina
= C.a cosp C2 (-cosa) Ca cosp C2 sin(a + týý)
= Ca sino C2 (-cosa) -a sinp C2 sin(a + 4o)

where

•/= .

Differentiation with respect to time gives an increase of phase angle of n/2.

4.3 Transfer functions of wave angular motions

nu92.c9k25 I'\p iwt:gcinl.lJ F7ri

I ')_.(19.2$ 7
The wave angular motions of a "wave particle rod" that at rest is parallel with the x-
axis (right handed rotations positive) are defined by

VI = N- C. 03 ksinp(-cosa)
ay
=2 = C3 kcosp (+cosa)

Y3 = 8x C2 ksinO (-sina)

This gives phase angles

$y2 = -•/2
( 3 = +T

This will be the asymptotic phase angles of yaw motion of a slender ship when the
wave length becomes large compared with the ship and the ship is oriented parallel
with the x-axis.

Other motions

All other motions or other responses, r. that are linearly dependent on the waves, are
expressed in the following way:

Z = r a sin (wtL + (ýP + (D)

where
r : harmonic amplituide'
=
: phase angle relative to the surface elevation.

ii :Vsck 1\1p: l\wcg1Ul IJI\rh

1992.09.25 8
4.4 Summary of phase angles of wave particle motions

This overview can be used to check asymptotic behaviour of floating structures (long
wavelength) and can serve as a basis for converting transfer functions from one
reference system to another.

Coordinate system

wave profile at , c .rec--f -)

t = 0 for (ýC = 0

Surface elevation

C= a C3 sin(+wot + (ýC + 4)

where
p = -kx cosp - kysinp
I = defines the wave state at the origin at time t = 0
C = wave amplitude at the surface
C3 = is a depth- and frequency dependent function

Table 4.1 Summary of phase angles, 4,

Derivative 0 1 2
Motion Displacement Velocity Acceleration

Surge t -1i2 0 +n/2

Sway 11 -R/2 0 +n/2
Hleave • 0 +n/2 +R
Roll Y, -rt/2 0 n/2
Pitch y. +W/2 +71 +±3n/2
Yaw yn. +±3n/2 0

i99:\s.k95 I \p;i t:r\wcg, ni. li I

I992.3)9. 25S 9
Table 4.2 Summary of motion "amplitudes".

Derivate 0 1 2
Motion Displacement Velocity Acceleration

Surge • cos[3 C2 ucosp

w'a C, U
W,cos2 C2 •

Sway rl sin[ C2 w
osin[ C 2 t_0 to2 sin[p C,_
Heave C . WC3 2
c,c3 a

Roll yi ksinf3 C, wksinp C,

w_ 0 ksinp C,
t2 ,
Pitch Y2 kcosp C3 tL wkcosp C. t, 0 kcosp C3 •a
Yaw Y3 ksinp cospC, a ksinp cosl C, •a U- ksinp cosp C 2 •a

C, is a depth- and frequency dependent function.

Observe that sinj3 and cosp enter the expression for amplitudes and include a sign.

When changing coordinate system this can be taken care of in the phase angles if [
goes out of the first quadrant and one wishes to have positive amplitude expressions.

I \palcirw,:t:in.I, l:lh
192:\s 9k.5
1992.09.25 I10
5 COMPLEX NOTATION OF HARMONIC WAVE FIELD, TRANSFER
FUNCTIONS AND RESPONSES

We define a complex harmonic wave component by:

Z = C, C3 exp[i(ot + (0 - t)P]
IzI = C, C3
Arg Z = + woC + 4)c + 4%,
exp (iArgZ) = cos (ArgZ) + isin (ArgZ)

Thus, the surface elevation is

= Im {Z) = C. C3 sin((t cýP)+ 4((+

= Re {Z exp(-it/2)} = C8 C3 cos (6)t + 4¢ + 4)p - n/2)

All other responses, r, are related to the surface elevation by complex transfer
functions, H,, and can be derived from a complex harmonic function, R:

R:H,Z
r =ImR
IHI = ra/Ca
ArgH, =

in :Vsck :
1\lt \¼Ivt:lttIi F\rh
1992.(09.25
"
6 CHANGE OF COORDINATE SYSTEMS AND SIGN CONVENTIONS

Basis
A right-handed Cartesian coordinate system is used with x-axis pointing upwards.
The sea surface elevation is:

C= Ca sin(Ct - k(xcosp + Ysinp) + 4c

where
3 is t-e propagation direction
4'. is the phase angle, interpreted as forward phase shift

The transfer function of any response j for a given frequency is defined by an

amplitude ratio, l-) (Q) j Žnd a phase angle, . (1)

Xi = Ca l")( sin() t - k(xcosp + ysinp) + 4z (P)

1. Denoting the base case as case no. 1, the following alternatives are discussed:

2. 180 deg. rotation about the Z-axis.

This is often used in order to obtain a head wave condition with 3 = 0.

3. 180 deg. rotation arout X-axis.

This may ire done in order to get positive rotation with the bow turning
starboard. w.hich is a nautical convention. This requires that the Z-axis points
downwards in a richt-handed system.

4. Change of Jirectio.. convention of the waves, so that 11 is coming-from

direction 1icead 0• propagation direction. This is also according to nautical
practice.

5. ofnen.ists
of ;r.,Cl convention so th;it + denotes phase lag instead of

phsisc tid . T'h, is puirely a imitter ol taste mnd h:abit.

in:\sck5I p;jmr\w;:-crn. iJrh

1992.1)9.25 12
 6. Utilization of X-Z symmetry by mirroring the transfer functions about the X-
Z plane.

Table 6.1 gives an overview of these alternatives. For the coordinate selections, cases
1, 2, 3, directions and phase angles are written fully. For the alternative sign
conventions, cases 4, 5, 6, only the changes are given. These can apply to any one of
the cases 1, 2, 3.

Table 6.1 Wave directions and phase angles

Case 1 2 3 4 5 6
110. Coming-from Phase lag Mirror image

Pa-X sin(ox-o)
meters -,
i

Wave j3-.- n - [ + it no change change sign

dir.

Surge
Sway
Heave
1)
03
02+
+0

43
03
7: 02 +
+ It
1 +
0
Roll 0. 04 + no i
change + 7C
Pitaw 05 0)+ ±, a change sign 0
Yaw 06 0646 -+ .+ 7

Note:
1. Change of + at is equivalen! to - at.
2. All cases except 6 represenT the same physical system and responses.
The cases 4. 5, 6 can be su.cerimposed on any of the coordinate system
alternatives 1. 2, 3.
4. If relative motions are to bc calculated, the same modifications must be
carried out both for vessel irinsfer functions and for wave particle motion
transfer function.

in :\c k 5 1\I);:I )tr\wCti'ilC It. 1.1

,
1992.19.2 13
7 EXAMPLES

Example 1
The NSRDC program has been run (case 2.5). The resulting transfer functions are
to be used as MOSSI input (case 3.4).

According to Table 6.1:

MOSSI (M) NSRDC (N) MOSS] WITH

(Case 3,4) (Case 2,5) ON as basis

PM= -1P+ it PN P+a PM= i•N

1' + -- a *N2

$M 9P + , ° 43 *M N3 a

ý1M -ONI -=

In order to use the result for the same wave directions (nominal values), the result
is mirrored and a are added to
•
all phase angles, giving:
9]54 " -C JE N4 3t

9• 5 + -•bn + itN

4
9.1 N2

PM.
I 9
= N4 +

= ýN + A1

Thut add itio

U H 0Y is dlone in ordter to refer to waive doi wnlwards, sue note 4 to tablei

0..

ni :\ýsk 5 \\1;i pt:\wt:gc l I. I.I F\rh

19)92.09.25 14
Example 2
Results from the NSRDC program (case 2, 5) are to be compared to those of
WADIF (case 2). The difference is simply a sign change of all phases, so the
comparable quantities are:

NSRDC WADIF

PN PW

4N2*W

4*N3 W

4 N4 *W4

4*N6 ýW6

Example 3
WAMOF (case 1) is to be run to create input to MOSSI (case 3, 4) for wave

directions PM = 00, 300, 600, 90'. 1200, 2500, 1800.

WAMOF MOSSI

0 P +I PM +=7

¢6 ¢6 +TE

II\ASC k 5 \p~qc
) I Ir\wCg mtLCIIIJ
I -1II
199)'2.09.L25 15
Corresponding wave directions:

MOSSI O3N 00 300 600 900 1200 1500 1800

WAMOF O3w 1800 1500 1200 900 600 300 00

Conversion of phase angles carried out by For relative motion problems, the 1800
WAMOF when results are written to a phase shift of the wave from case 1 to
MOSSI file: - case 3 has to be included:-

$)M2 =ý W2 -ý2 4)W

$112 =4$w93 + 'M_ 53
W11=

$114 =P W)4 ($114 =C$54 + U

$M15 = $W5 - ý
$115 = $595

$C16 =t $WI +C It $6 =C$56

For problems involving only absolute motions these two sets of phase angles are
equivalent.

in :\scOk 51 \;xtfw''\V t :-; I \rh

I.
19921-9.25 10
 HF-MOTION TZANSFER FUNCTION SURGE

Qvý.j3Z: 950505 015S / PI'ot: 930303 1025 Heodtng .0 (dog)

SYSTEM PARAMETERS:

tdent. TANKER
Aml.-.deratoRAO

REFERENCE POINT:
C.4- XP1 .00
2.2- XP2 : .00
2-I............................................XP3 : .00
0 .2 0.4 0 .6 0 .8 1.0 1.2 1.4
Pnose ongle (degrees) Anigutor Frequency (Rod/T)

AmpiCud roI / eoo ec

0.2 0.4 0.6 0.8 1.0 1.2 1.4

Angutor freq~uencyi (Rod/TI

2. 1).)8.25I
 2.0-

0.2 0.4 0.6 0.8 4.0 1.2 4.4

IOPhose cngto (clog.ee.. Angutor freqzuency (Rod/TI

-so
- 0
- SO.

020.4 0.6 0.8 1.0' 1,2- 4.

Angutor frequency (Rod/T)

0.6-]

0.41
0.21
0
I.
-i
0.2 0.4 0.6 0.8 9.0 1.2 1 .4
Pnzso tngL~e Idegrees) AnguLtor frequency (Rod/Ti

100

-200 -I4
0.2 0.4 0.6 0.8 4.0 1.2 1.4
Angutor Frequency (Rod/Ti

0200

0.2 0.4 0.6 0.0 1.0 4.2 1.4

2CC-P-c- .. IAngutor
C-1.0(cler frequency (Rod/T)

Fiue7Ib rnfe uciostr h

:Vk
I\ 0I 1 1WCC1 i Fr
I9).
O9.5I
 AffptLtude rotto Z-/-I
S4-.g ,

3-

2-

0-.I

0.0 0.2 0.4 0.6 0.8 1. 1.2 1.4

ýge
P~osI og-es)Angutor frequency (Rod/TI

300

0-

2100-

0.0 0.2 -0.4 0.6 0.8 1.0 1.2 1.4

Angu tar Irequency (Rod/Ti

Anp~tttude ratio (I

0.0-

0.0 0.2 0.' 0.5 0.8 9.0 1.2 1.4

Phase onglo boec-ees; Angular Frequency (Rod/Ti

-000

-1001

-200- 1 1 --
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
AnguLar frequency (Rod/Ti

I 2.)'
9' .25

D.19
C 1993 Marine Management (Holdings) Ltd Paper 10, Offshore 93
installation of Major Offshore
Structures and Equipment
17-18 February. 1993

Dynamics of offshore towing line systems

S Moxnes, Msc and Dr Ing I J Fylling
The Norwoegian Marine Technology Research Institute (MARINTEK), Norway

This paper describes amodel for estimating extreme towline tensions. The model incorpo-
rates effects from wave induced motions of tug and tow, dynamic response of the towline,
and dynamic behaviour of shock damping devices, such as rendering winches. The model
is implemented in a PC-program suitable for parameter variation studies in a planning or
onboard situation.
The towline dyrmmics part of the model is extensively tested against the non-linear finite
element prograRIFLEX, and the correspondence with respect to dynamic tension is very
good for a wide range of towline configurations and excitation levels.
The development has been initiated by the Norwegian Maritime Directorate as a part of
the work on reducing the risk related to offshore towing operations.

INTRODUCTION 2. During the towing operation the sea gets rougher. The
tug master wants to know if any action should be taken
to reduce the risk of linebreakage.
Ocean towing operations requireskilled personneland good
equipment to be carried out safely. Still, towline breakages 3. The tow is going to pass an area with shallow water. Is
are not rare occurrences. Improvenent of equipment, such there any risk of the towline hitting the sea floor?
as automatic winches and tension monitoring, is continually
taking place. To obtain experience in the handling of new
equipment, and also in the towing of new types of structures
takes time. Besides, many of the offshore support vessels SYSTEM DEFINITION
have towing assignments only occasionally as a small part of
their work. In order to reduce the learning time, it may be The considered system•consists of four components, see Fig 1:
useful to introduce simple computer programs to predict the
dynamic towline loads. 1. TUG, the tugboat;
This paper describes such a program. The program in- 2. TOW, the towed structure;
cludes effects from shock damping devices, such as the so-
called rendering winches. With this program thetug master 3. LINE, the towline, including elastic properties;
can calculate the consequences of changing towing force, 4. WINCH, the winch, which may include an automatic
line length, winch setting,or achange in weather conditions. rendering function.
The development of the program was initiated by the
Norwegian Maritime Directorate in 1991. Marintek who
have considerable expertise in wave induced vessel motions
and line dynamics, have performed the work. METHOD OF ANALYSIS
Some examples of the use of the program are:
1. Given a required safety factor against breakage, the ForTOWCAPtobeanefficientPC-tOOl,Stringentlimitations
limiting seastate for a certain tow operation can be are set on model complexity and size. However, the system
determined. tobeanalysed iscomplex. The problem involvesboth vessel,

Authors' biographies
Simen Moxnes is Research Engineer at MARINTEK in Trondheim Ih.e received his masters degree in naval architecture from the
Norwegminlnstitiiteof Technolog,y(NTII)in 1990. Sincethespringof 1991 he hasbeenemployed at MARINTEK. I-hs rmin activities have
been related to crnnptter simulation of nrifne o(ptrations.
Riling is Senior Principal Research Enoneerat MARINTEK. Ile isa Naval Architect (1970) and holds a Dr •ng degree front NT-I
Ivary
1970j. Hi( has been working with research prortcts related to n"rinto 1 erations, rn~ring systen"sand stationkeeping since 19'.
10-2 Offshore 93 Installabon of Major Offshore Structures and Equipment

. \% IN 'II

Fig I System definiton

towline and winch dynamics. The response of winch and
towline is strongly non-linear and, in addition, coupled.
Frequency domain methods are superior when it comes
to computational efficiency, but they require a linearised
model of the system. To account for non-linearities a time
domain integration must be used. These methods are much
more demanding both with respect to computing time and
ta storage.
Based on these considerations a hybrid method is chosen.
First order vessel responses are well described by linear
equations. Neither tug nor tow will be significantly influ-
enced byline with respect to firstorder response. A frequency
domain analysisbetwen
canow,
give
ug information
nd isinptontothetim
hic the relativedoain
motion
between tug and tow, which isinput to the time domain
analysis simulation ofthe towline-winch system. A stepwise
description of the solution method is given below.
1. Static line configuration is calculated. The static con-
figuration is the basis for calculation of the dynamic
towline response. Maximum sag depth is also obtained
from the static analysis.
2. The motion transfer functions for the line attachment
points are calculated on tug and tow respectively. The
transfer function describing the relative motion of the
line attachment points, can thereby be established,
3. Bycombiningtherelativetowlinemotion transfer func-
tion with the wave spectrum, the response spectrum for
relative motion in the secant direction of the towline is
obtained.
4. From the response spectrum an extreme representation
of the relative motion can be generated
5. The extreme motion is imposed on a single degree of
freedom system consisting of a simplified model of the
towlineand thewinch. Duetostrong non-linearities this
part of the analysis will be performed in the time do-
main. An integration along the time axis will give time
series for cable tension as well as tile winch action.
conditions at one end in order to satisfy specified boundary
conditions at the other. Using this procedure a fairly accurate
static equilibrium configuration for a multi-segment line can
be obtained with a minimum of computational effort.
Calculation of relative motion between tug and tow
Transfer functions
A motion transfer function describes the linear relationship
between regular wave amplitude and vessel response. The
vessel response is given by the motion amplitude and phase
vessev tonse eleva tion
waveelevation. ample an ,
relative to the wave By using complex notation,
information on both response amplitude and phase can be
contained in the same function.
In TOWCAP the motion transfer functions are given
numericallyforeachdegreeoffreedom (DOF) fora number
of frequendes and directions. Linear interpolation is used
between the given values.
Theinputtransfer functions are referring to theorigin in the
local body fixed co-ordinate system. The transfer function
for relative motion between tug and tow in the towline
secant direction is obtained by the following procedure (all
angles and co-ordinates are defined in Fig 2):
1. The transfer function for the towline terminal point at
the tug is obtained by a linear transformation of the tug
transfer function.
2. The transfer function for motion in the towline secant
T
direction is then obtained by a rotation ac = a - OTr
around the local z-axis.
The procedure (1) to (2) is repeated for the tow with
T h 0p re180 In addition the phase angle for the toww
ao -T-10.Iadto h hs nl o h o
transfer function must be modified according to its position

The different steps of the method are described more
closely in the following subsections.
Special attention is given to the model of the towline-
winch system.
Calculation of static line configuration
relative to the tug (XCTO,YGIO) and the wave direction, P.
The w'ave heading relative to the tug is PTu = P - 0 "TU,
and
to the tow PTO = p - 0o° .
The transfer function for relative towline secant motion,
11,at wave heading, P3,is obtained by adding the two transfer
functions:

.)e procedure for calculating the towline confi gurationi

is (i P) = It (0t) ) 11,0 ( or 10)1
basLed on a 'shooting methlod', or iteration oll botd;ry 1=
 Ofishore 93 Installaton of Major OfIshore Structures and rquipment 10-3

a''P

XlTT

XTO

(• TON 0TO. Y L~

Fig 2 Co-ordinate systems and angle definitions

Relative motion response spectrum a and zero-crossing period T., the expected extreme value in
By combining the relative towline motion transfer function a time interval T cn be estimated by:
with the wave spectrum, the response spectrum for relative ,.- 0.5772 (3)
motion, rl, in the towline direction is obtained: E x[ l In (CTf.)+ 52 3)

S (W)= Il T1, I-(W ) (2) The spectral moments are defined as:

whereS C is the wave spectrum desaribing theactual seastate. r (4)

mn = JoS(0o) U
0
Extreme representation of the relative motion
The standard deviation is given by:
In order to avoid simulation of a long irregular time series.
a representative harmonic extreme cycle is determined as - (5)
follows. The motion amplitude is estimated from the r=- J (0))Wo
sponse spectrim. The frequency is given by the estimated o
maximum motion and velocity amplitudesin a time interval
of length T. For a GaUssian process X, with standard deviation and the zero-crossing period:
* -4 Oflshore 93 Installatbo- r,1Major Offshore Strctures and Equipment

'/•(I)
/

/
Fig 3 Line dynamics model

T =x 2 .- (6) x (12)

02 This is a simplification of the model for mooring lines

The spectrum of the relative velocity is given by: presented in Ref 1.

S (W) = (0 2S (0) (7) Winch model

Thestandarddeviationandzero-arossingperiodforboth The rendering winch is modelled by an equivalent mass,

Therelative motion and velocity can thereby be found. dampingand stiffness,and theequation of motion becomes:

Having estimated maximum amplitude, T,., and veloc- m'i .+c'i. + k'x = T +T (13)
inv, T,, the related oscillation frequency is found: * S
01 = fi ý /Ti (8) when the winch is rendering, and:

The desired extreme motion is thereby determined: x , =0 (14)

1 = 1"1 sin cOt (9) if the winch is stopped.

The winch starts rendering when the total tension (T o +
Td)exceeds F.
Calculation of towline-winch dynamics
Coupled towline-winch model
Towline model
Two different sets of equations are needed to describe the
.ne-towlinemodelisbasedontheassumptionthattheshape dynamic towline-winch system, depending on the winch
c: the dynamic motion, (Ol), is equal to the change in the operational mode.
s:atic line geometry. Mass forces on the towline are ne- In the rendering mode:
zlected. Moment equilibrium with respect to point P yields Equation (13) describes the winch dynamics.
(see Fig 3): By assuming that the towline leaves the winch horizon-
tally, the following relation is obtained: -
JIdlfc xildl =aTDc (10) x = ipX (15)

TUX: is the tension caused by the drag loading on the line, The dynamic line tension can be expressed as:
2_d is given by:
T'DC = k E (x- u)-k Su (11) To = kE(x -- u) (16)

il'ysicaliv equation (lO) expresses that the moment of T1 Equations (IS) and (16) inserted in equation (13) yields:
ir- line end I about point P (line end 2) should balance the
sum of the moments from the drag forces on each tiny mx+CX (k'ekE)x= (17)
:ement along the line. Manipulation with equations (10) kit,± m'n+c'rT4+ k'rl- T-,
nd (1 1) leads to the following equation describing the line
motion, 11: Using tequation (9) in equation (17) one obtains:
 Offshore 93 Installhon of Major Offshore Stnadures and Equipment 10-5

C.

Fig 4 C-oupled towline - winch model

wnh aa
4r.

m'i< + c'i + (k'+ kE) X= kEu + F1- T s (18) 1. motion transfer functions in 6 DO:F f Or the tug;

for the tow (op-

2. motion transfer functions in 6 D:OF:)
where: tional);
-
F = ", • mW2 sin cot + c'c0 cosw~t + k'sin o) t) (19) 3. towline data;

A system of two coupled equations with two unknowns winch d

describing the dynamics of the towline-winch system in the
rendering mode is now established. Equation (12) is of first
order and equation (18) of the second order. Some further
manipulation brings these equations into a form suitable to
be solved by astandard numerical scheme:
I[-t[(k+ k,_)x -c'i+ ku+ F,, -TJ
A data base of motion transfer functions for standard
vesselsand structureswill be included. When data is lacking
for either tug or tow (or both) these can be scaled and used.
as input totheanalysis. If thetow is much larger than the tug,
the relative motion will be dominated by the tug response,
and the tow transfer functions may be omitted.
(20) -

Operational parameters
u - ) k , (21) The following operational parameters can be varied, to

In the stopped mode: reflect actual or forecast weather, and actual or planned
Equation (14) describes the winch dynamics. towin parameters
Since the winch velocity is zero, the line end motion is 1. tug force;
equal to the relative motion, i: 2. wave spectrum parameters;
* = i (22) 3. wave direction;

The equation describing the line motion in equation (21) 4. vessel headings;
is unchanged.
The towline-winch model is illustrated in Fig 4. 5. line length;
6. line direction.
Method for numerical integration
A svstem in this form is well suited for implementation in a
- standard numerical scheme? For solving the equations of Analysis output
motion a fourth order Runge-Kutta (R-K) method is used. R-K
methods are robust and their simplicity makes them prefer- For each analysis (ie,set of operational parameters) the
able to more sophisticated methods for most problems. presented outputs from the program are:

1. maximum line tension;

DESCRIPTION OF THE PROGRAM 2. safety factor against breakage;

3. sag depth;
This section gives a brief description of the TOWCAP pro- 4. vessel response.
grain,

System data User interface

The following system data are rtMlifired to perform a The ieterfacebetween TOWCA', external filesand the user
TOWCAI] analysis: isillustrated in Fig 5.
10-6 Offshore 93 Installation of Major Offshore Structures and Equipn;.:

I)AIA 1FIJNC1
RANSFER
IONS

USER Pe SCREEN

IrTO\NVCAP
P~kAN _11 E15 IELI1

OFI
ATINA I I

Fig 5 TOWCAP user interface

Tablet Tovwine data

Sgment Segntent Numberof Segment Diamneter Elastic Factor of Unit urighl Weight Drag
number type elements length I.) (rn) modulus elasticity in Utter ratio coefficient
(kNl(my) (&NfI) wer/lair nonnail
longitudinal
I Catenszy 30 500/1000 0.072 0.7490E+08 1.00 0.1730 0.810 1500/0.150

Operational parameters, such as tug force, suspended Table 1! Seastates for verification of TOWCAP
line length, winch setting and wave spectrum parameters,
are given interactively by the user. 1, (in)
System-fixed data such as motion transfer functions are T, (s) 3D 6.0 9.0
read from external files. For a given system these can thereby 6.0 1 2
be considered as beinginsid e a 'black-box',ie they require no 8.0 3 4
manipulation by the user. This is desirable for onboard use 10.0 5
of the program.

F-ESTING AND VERIFICATION
Testing of the TOWCAP program is split into several activi-
ties as follows:
1. testing of the calculation of relative motion;
2. testing of the sag depth calculation;
3. testing of the calculation of dynamic line tension;
4. testing of the coupled winch - towlinedynamics calcu-
lation.
5. testing of th overall performance of the program.
The routines for calculation of first order vessel response,
are taken from the MARINTEK program MIMOSA-2.' The
implementation of these routines in TOWCAP has been
tested by comparing the calculated response directly with
"'sUlts from MIMOSA-2 (activity (1 )).
The sag depth calculation has been checked against ana-
lytical expression (activity (2)).
The line dynamics model has been tested against the
finite element program RIFLEX, which again has been exten-
sively verified,' -' (activity (3)).
Activities (4) and (5) will be carried out as part of a full
scale test, which is planned in co-operation with several tug
boat operators.
Results from activities (1), (2)and (3)are presented in this
paper.
Test procedure
A system consisting of a tug boat and a towline is simulated.
Thetow isassumed to be much larger than the tug, so that the
tow response -n be neglected. The tug is W.ring in head seas
and onlythesurge motion isconsidered. The relative motion
between the line ends will thereby be equal to the surge
motion of the tug.
Four parametersare varied in the verification stud y of the
dynamic line model: line length, L, tug force, T , significant
wave helight, H., and zero-crossing period of the wave spec-
 Offshowe 93 Instalation of Major Offshore Structures and Equipment 10-7

Table IIl Comparison of extreme surge motion Table V Companson of dynwJ-n line tension

Sbwc Hs TZ Surgenmpliwde Surgeperod T!t S., S'id Vv.non: nsi&n

number (in) (s) (,n) (s) no no rII)

TOWCAP MIMOSA-2 TOWCAP MIMOSA-2 RIFLEX TOWCAP % dijerence

1 3.0 6.0 1.40 1.40 8.65 8.68 1 1 A 560 683 22.0

2 6.0 6.0 2111 2.81 8.65 8.68 2 2 A 1355 1528 12.8
3 6.0 8.0 3.93 3.93 10.42 10.47 3 3 A 1830 2100 14.8
4 9.0 8.0 5.90 5.90 10.42 10.47 4 4 A 2700 3-104 1.7
5 9.0 10.0 6,78 678 12.21 12.30 5 1 B 730 768 52
6 2 B 1560 1597 2.4
7 3 B 2170 2213 2-0
8 4 B 3350 3374 0.7
9 1 C 380 401
424 5.5
3.0
10 2 C 800
Table IV Comparison of towline sag depth

11 3 C 1100 1153 4B
System Linwle•gth Tug frar Sag dEph
eonflguralin (in) (N) (Ma) 12 4 C 1700 !747 2.8
13 5 C 1900 1998 5.2
Anolvtic.o TOWCAP % dibvrea 14 1 D 410 410 0.0

500 300 1&0 18.1 +0.6 1s 2 D 840 834 -0.7

A
500 500 10.8 10.9 +0.9 16 3 D 1160 1167 0.6
B
C 1000 700 30.9 31.1 +0.6 17 4 D 1740 1761 1.2
D 1000 1000 21.6 21.7 +05 18 5 D 2000 2017 0.9

trum, T. The winch is locked. Consequently this verification Results

covers only the towline response part and not therendering
winch part of the model.
Thecalculated surge motion of the tug for each seastate is
compared to results for the same vessel calculated by Ml-
MOSA-2.
The dynamic line tensions calculated by TOWCAP are
verified against results given by the finite element program
RIFLEX. Each case is first run with TOWCAP. The resulting
regular line end motion is imposed on a non-linear finite
element model of the towline.
For each system configuration, given by line length and
tug force, sag depth at mean position, as calculated by
TOWCAP, is compared to the following analytical expres-
sion giving the sag depth for a single segment taut towline:
1 2
SAG -•wL T8
where w is unit weight per length in water, L is sus-
pended line length and T. is horizontal component of the
line tension.
A total number of 18 test cases are run.
Input data
Motion transfer functions for a 2000t tug vessel have been
used.
A single segment wire rope towline is used in the studv.
Data for the towline are given in Table 1.
The verification study is performed for twoline lengthsat
50mn and 10(X)m. The tug force is varied between 300 kN and
I0(XX) kN.
'T7he wave condition is defilned by a tw -1 )arameter I'M
spectrum. Theconsidered seastatesaresummaristd inTable
II.
The vessel surge motion as calculated by TOWCAP and
MIMOSA-2 for the defined seastates is presented in Table
III.
The calculated sag depths at mean position for the differ-
ent system configurations are presented in Table IV.
Table V summarises the results from the dynamic
simulations.
DISCUSSION
Vessel response
(23) MIMOSA-2 and TOWCAP give identical results for the
calculated surge amplitude. This should be expected, since
the same calculation procedure is followed.
The period estimated by MIMOSA-2 is the relative mo-
tion spectrum zero-crossing period, as calculated by equa-
tion (6). The surge period given by TOWCAP is calculated
from estimated extreme motion and velocity amplitudes,
according to equation (8)_ The difference, however, is insig-
nificant for the present motion spectrum.
It is important to keep in mind that this is only a verifica-
tion of the implementation of the first order vessel response
calculation routines. The quality of the response estimate is
greatly dependent on the accuracy of the transfer functions
that are given as input-
Sag depth
Tlhecorrespond en cebetwecn th e ar.n lyticallycalculated and
TOWCAP estimated sag depth is ve.ry goo(d (within 1.0%).
10-8 Offshore 93 Installation of Major Offshore Structures and Equipment

Dynamic line tension 2. W H Presset aINumetrcal Recipes.Carrindge Universitv l'ress

(1988).
S
•-seen f
from Table e3. between TOWCAP
l IV, thee agreement MIMOSA-2
P ReotNMiIP-35 Version 2,User's Documentation, MARINTEK
Report No MTS1 F 90-0358
and RIFLEX is very satisfactory with respect to the dynamic 4. 'RIFLEX,FlexibleRiserSvsteni Analysis Program.Test Manual'.
tension for all the test series A,BC and D. The largest MARINTEK and SINTEF Division of Structural Engineering.
deviation is observed in test series A. This is the series with Trondheim, Norway 1987.
the shortest towline and lowest tug force. Here TOWCAP 5. 1J FyvUin C T Stansberg and K Mo. 'Extreme motions and
overestimates the dynamic tension by 22.0% (test 1). For the anchor line loads in turret mooring systems', Boss '92, London,
(July 1992).
by.TOWCAP
other series the dynamic tensions predicted
are within 6 % of the RIFLEX results.
The results indicate that the assumptions made as a basis APPENDIX A
for the simplified line model hold for typical towline con-
figurations. By 'typical towline configurations' is meant
relatively shallow sag (= taut) catenary configurations. The Nomenclature for symbols used in text
assumptions made were:
I-The shape of the dynamic motion isequal to the change a - moment arm of towline end tension
instatic line geometryn cm - generalised line damping
c - winch damping coefficient
2. The dynamics are dominated by transverse drag forces, c° - equivalent velocity dependent winch damping
so that mass forces and longitudinal drag forces can be coefficient = c/r'
neglected. f, - drag force per unit length
The resulting towline model is very simple and requires F - winch rendering limit
onlya fraction of the computational effortthat isrequired by H, - significant wave height
a finite element program for solving the same task. H'rU. - wave elevation to motion transfer function in
towline secant direction for the towline terminal
point at the tug
HTO,• - wave elevation to motion transfer function in
CONCLUSIONS towline secant direction for the towline terminal
point at the tow
The computer program presented in this paper isan efficient H - wave elevation to motion transfer function for
tool for predicting extreme towline loads. The results are relative motion between the towline ends
slightly on the conservative side compared with a finite I - winch mass moment of inertia
element program. The program requires small computer time k - winch stiffness coefficient
and parameter variations can easily be carried out on a PC. k' - equivalent position dependent winch stiffness
The selected modelling and analysis method can deal coefficient = kjr'
efficiently with the non-linearities introduced by rendering k- - generalised line stiffness = k,+ kE
winches. kE - axial line stiffness
Further verifications against full scale measurements are k, - static catenary stiffness
planned. L - towline length
m - n'th order spectral moment
m - equivalent winch mass = I/r 2
i - moment arm of drag force
ACKNOWLEDGEMENTS r - winch radius to top layer
u - displacement of the line
The development of the pilot version of TOWCAP was u - velocityoftheline
funded by the Norwegian Maritime Directorate (NMD), the x - tangential motion excitation of the line end
Norwegian Shipowners Association (NSA) and the Norwe- x' - winch pay-out co-ordinate
gian Oil Industry Association. The ongoing workis spon- SC - wave spectrum
sored by NMD and NSA as well as a group consisting of S, - spectrum of the relative motion between the tow-
future users of the program. Participants in this group are: line ends
Brbvig Offshore A/S and Austevoll Management A/S; t - time variable
Eidesvik & Co A/S; Smedvig A/S; Solstad Shipping A/S; T - time interval
Sverre Farstad & Co A/S. TO - dynamic line tension
TD~C - dynamic line tension caused by drag loading
TS - static line tension
T - average zero-crossing period
REFERENCES X TO - global x-position of the tow
°
.T - global y-position of the tow
K -irsen and P'C Sandvik,'Efficient nethods for calculaion of - global towline direction
dynarmic t1ociring line tension', EUROMS '90, Trondheim (Au- art - towline direction relative to loeal tug co-ordinate
gust 1990). system
 Ofishore 93 Instailabon of Major Off shore Structures and Eqiuipnvxm-t 10-9

a0 - towline directior relative toIiocalI tow co-ordinate TI - relative motion between the towline
system ends
3 - global wave direction 0 - tug heading
>jh - wave direction relative to the tug 6 T0 - tow heading
0ýo - wave direction relative to the tow W - angular frequency
 HEAVY LIFT OPERATIONS -
METHODS AND EXPERIENCE -
POSITIONING AND STABILITY CONTROL

Lars Loken - HeereMac

ABSTRACT
This paper describes in broad terms the installation aspects of the Snorre Concrete
Templates (CFT's). The method used to install the CFT's at the correct location is describedFoundation
as
as the ballasting system of the SSCV DB-102. The positioning system configuration for normal well
control is also described and important results of the station keeping capability for vessel
the 12 point
mooring system is discussed. Offshore measurements were recorded on the performance
of the SSCV
positioning capabilit'ies and the line characteristics of the vessel and cranes. It was shown
that the
positioning requirements and ballasting involved in such an operation can be well controlled.

Introduction
In August 1991 the SSCV DB-102 lift installed four Concrete Foundation Templates (CFT's)
Snorre Field in the Norwegian Sector of the North Sea. The installation was performed by at the
HeereMac
v.o.f (HeMc) under subcontract form Norwegian Contractors A/S (NC). The operator
to the field,
SAGA Petroleum A/S, had granted NC an EPCI contract for fabrication and installation
of the CFT's.
Each CFT's had a weight of approximately 6000 tonnes and had to be dual crane lifted
and
to a depth of 310 meters, to function as suction anchors for the Snorre Tension Leg Platform lowered
The application of the CFT's required positioning on the seabed to be within 0.5 meter. tethers.
It was also
required that the penetration ii the seabed was performed sufficiently controlled to avoid
the soil structure. For this reason, the SSCV was used to control the rate of loweringdamage to
inclination of the CFT during its initial penetration. To ensure positioning within and the
the required
tolerances, the CFT was connected to the SSCV down to a penetration of 4 meters in the
seabed after
which the rigging was disconnected. Acoustic transponders were pre-installed in the existing
Template transponder brackets as reference at their predetermined absolute position. Well

The CFT Lifts was monitored by the use of the installed Display, Logging and Simulation
system on the SSCV by which the hook loads and SSCV motions could be monitored and (DLS)
DLS system enables actual display of relevant data and (ballast) simulation of in particular stored. The
heavy
to ensure a safe operation. In order to monitor the CFT's during installation and final set down lifts
seabed, the CFI' was equipped with an instrument package which was capable of logging on the
and
the heading, inclination and overall position of the CFT during the whole installation sequence, storing
was not part of the instrumentation needed for the positioning and control system for the but
SSCV.
instrument packages served the purpose for collecting data for experience and possible computer The
model
calibration.

IFi:Vlc xt\gucnrMa\ .lr [ 18.02.93

PROJECT DESCRIPTION

General Snorre Project

The Snorre Field is situated in 310 meters of water depth near the Norwegian Fields of Gullfaks and
Statfjord. The development constitutes a TLP, a weUl template, four Concrete Foundation Templates
and a Subsea Production System. Several papers have already given a description of the project in
general and of the various components involved.
Installation of the Well Template and the docking piles for the CFT's was performed by the SSCV
DB-102 in the summer of 1990. The CFT's were fabricated on the transport barge at the NC Yard in
Hinna, Stavanger and transported to site to.be installed in June 1991. In 1992 the TLP was towed to
site and connected to the prepared foundation.

Installation Aspects
Installation of the four CFT's were performed by dual crane lifting one by one from the cargo barge
to the seabed. Given that the weieht of around 6000 tonnes was to be set at a depth of 310 meters it
was required to re-reeve the main fall riggina of PS and SB cranes and include additional wire.
Applying this method, each main hoist obtained a capacity of 3500 tonnes up to a radius of 51.6
meters.
The installation sequence for each CFT required about a day and was as follows:

- Bring in and moor cargo barge to the DB-102.

- Install Installation Module on CFT and test.
- Connect grommets to, both blocks. -
- Pretension and, if necessary, ballast cargo barge.
- Lift off the CFT.
- Remove cargo barge.
- Immerse CFT under constant de-airing.
- Lower CFT in preparation for docking.
- Docking of CFT against pre-installed piles.
- Lower CFT to initial penetration of around 4 meters.
- Remove rigging.
Bring CFT to final penetration.
Specific new installation aspects involved the lifting of a heavy concrete structure from a barge, the
de-airing process on this scale, the docking arrangement, the underwater control of the CFT during
initial penetration and the suction operation.
The CFT's had been fabricated on the cargo barge on top of a concrete grillage with plastic sheets and
without any need for additional seafastening. It was found that this connection functioned very well
giving a smooth lift-off.
The de-airing had to be performed together With the lowering at such a rate that the hook load
remained at a level suitable to ensure the stability on the splash zone- In practice the hook load of each
crane was kept above 1200 tonnes by short controlled lowering periods with continuous de-airing. In
this manner full submergence was obtain 35 to 40 minutes after touching the water. The de-airing was
without doubt the most spectacular phase in this generally submerged installation.
The docking arrangement (Figure 1) had been designed such that after docking the guide frames and
a large part of the frame remained on top of the docking piles. This ensured that as little structure as
possible would have to penetrate the soil.

I:Vlcxl\gcncra 1\p •pcr.II 2 18.02.93

During penetration (Figure 2) controlled lowering was required to avoid loss of seal and subsequent
damage of virgin soil. The method used offshore involved pumping out of seawater from the three
cells in line with the lowering of the CFT by the crane vessel. The lowering rate required of the crane
system was in the order of 1.5 m/hour or 2.5 cm/min. This-very low lowering rate could, however,
be obtained with more accuracy by performing controlled ballasting down of the SSCV itself rather
thad by using the cranes. This operation is contrary to what is normally performed during heavy lift
operations where it is more often required to ballast as fast as possible to keep up with load pick-up
in the cranes. In this operation ballasting was performed as slow as possible.
Several simulations were performed in the HeMc office as well as on the DB-102 ballast simulators
(DLS) to check the feasibility of the operation. Furthermore, several real time simulations were
performed with the vessel to check whether the lowering rate could be attained. These preparations
proved successful and it was chosen to adopt this method.

SSCV Control Systems

Description of DLS System

In 1991 HeMc installed a tailor made DLS system on board DB-102. This system can be used in three
modes of operation:
display of on line characteristics of vessel and cranes
static simulation
time domain simulation of lift and ballast operations.
Logging of pertinent variables is possible in all modes of operation for reporting and evaluation
purposes.
Basic aim of the DLS system is three fold:
to assist captain and ballast operator in preparing lift and ballast procedures
to assist the ballast operator in taking adequate ballast actions during an actual lift and ballast
operation
to provide the captain with information to control vessel stability.

Job Preparation:
Personnel on the bridge of the DB-102 make use of the DLS system in simulation mode in the job
preparatory phase of a project. Crane and ballast actions can be simulated in order to obtain the
expected vessel and crane characteristics during the project execution. Alternative lift and ballast
procedures can be simulated to obtain an optimised lift and ballast procedure contemplating with
specific characteristics of a project.

Simulation Modes:
The DLS program comprises two simulation modes, a static and a dynamic simulation mode
respectively. The static simulation mode is used to obtain the vessels behaviour in terms of draft, heel,
trim and GM-values for specified crane loads and filling rates of the ballast water tanks. The static
simulation mode is used to check defined stages of a lift operation and to create initial conditions for
a dynanmic simulation.
The dynamic simulation mode is used to simulate in a realistic manner a complete lift and ballast
operation. Basically, the ballast operator has to perform the same actions as required during a actual
lift or ballast operation. Valves of the ballast system have to be opened and closed, ballast pumps have
to be started and stopped and hoisting, booming and slewing speeds of the cranes have to be controlled
during a dynamic simulation. The simulator calculates these effects on hook loads and filling rate of
ballast water tanks and subsequently on vessels draft, heel, trim and GM-values. The dynamic
simulation mode is a very realistic crane vessel simulator, which is used to train ballast operators for
lift and ballast operations.

I ":\'Icxt\gcuicrat \p:ijc r.II 3 I 8.02.93

Graphical Displays:
The display of on line crane and vessel characteristics is the main function of the DLS system.
Required information is provided via data links with the crane computers and the data network system
of the vessel. The graphical presentation, which is identical to that used in simulation mode, comprises
three lay-outs:
heel - trim - cranes picture (Figure 3)
ballast system picture (Figure 4)
ballast routing picture (Figure 5)
In display mode, a large number of data quantities are measured. In simulation mode, this data is
replaced by user defined input or simulated data dependent on the type of simulation performed.

Heel - Trim Cranes Picture:

The on line position and loading condition of the cranes are represented by the following data
quantities:

Measured:
- slew angle (deg)
- outreach (m)
- load (t)
- side lead (deg)
- hook speed (rnlmin)
Calculated:
filtered load W-t)
load moment " (%)
The filtered load is calculated by applying a first order filter on the measured load. The filtered load
is a steady signal without the wave induced load fluctuations. The load moment is defined as the
percentage of the actual load related to the maximum allowable load being a function of the actual
outreach of the crane.
The vessel characteristics presented in this picture are:

Measured:
- heel (deg)
- trim (deg)
- draught starboard aft (m)
- draught port aft (W)
- draught starboard fore (m)
- draught port fore (W)

Calculated:
average draught (W)
longitudinal stability GoMI (i)
transverse stability GoMt (W)
The measured heel and trim are visualised by a moving cross in a fixed cross which layout is
considered to be very convenient for the captain and ballast operator.
The average draught is calculated using the measured drafts, heel and trim. Unreliable draft readings
can be deactivated for the calculation of the average draught whilst maintaining an acceptable level
of accuracy. The average draught is used to obtain the displacement of the vessel.

E:Vl'cxt\gencrzl\pa pc r. II 4 18.02.93
 The longitudinal and transverse stability is calculated taking into
account the followin2 effects:
vessel modifications
position of cranes
crane loads
deck load distribution
mooring system
filling of miscellaneous tanks
free surface effect.

Ballast System Picture:

The filling of all ballast water tanks as a percentage, in tornes
and by means of a bar graph is
presented in one picture in a layout reflecting the actual location
vessel. Also the capacity of all ballast water tanks is presented. In of the ballast water tanks in the
the same picture the status of the
ballast water tank valves is presented by means of colour changes.
By
a ballast water tanks'is inactive or connected to either the suction or this feature it is clear whether
pressure side of the ballast water
pumps.
Ballast Routing Picture:
The routing of ballast water flow depends on the status of all valves
and can be very complex. This
picture presents the status of all valves by means of colour changes
ballast water flow is visualised by means of colour changes of the and bar graphs. The routing of
lines
valves. The captain and ballast operator are able to monitor the selected connecting (partially) open
routing of the ballast water
flow by means of this picture.

Stability Report - Pertinent Data:

On line an updated stability report is availablje which can be printed.
a summary page and pages with detailed information. The stability report comprises
The summary page with on line information can
be made visible on a text screei to monitor on line the stability
of the vessel. The stability of the
vessel (GoMI and GoMt) is calculated using the following items and
effects:
- vessel light weight (February 1985)
- vessel modifications
- correction cranes
- crane loads
- deck loads and constants
- deck load correction
- anchors and anchor wires
- ballast water tanks
- miscellaneous tanks.
The crane loads are measured by load sensors. The crane loads
are adopted to be weight items
concentrated in the boom tip location. By doing so, the pendulum effect
of the load is included in the
stability calculations.
The weight distribution of the deck load items and other constants
are defined by the captain and the
ballast operators and reflect the actual deck layout.
The weight distribution of the anchors and anchor wires and the loading
vessel is assumed to be constant within the DLS system and is defined of the mooring system on the
by the ballast operator.
The filling of the ballast water tanks is measured. This information
is available within the DLS system.
The filling of the miscellaneous tanks is measured for some tanks and
defined for others by the ballast
operators.
Alarms are built in (o inform the captain and ballast operators in
case the actual vertical location of
vessel's centre of gravity approaches or exceeds the maximum allowable
vertical location. Both values
are presented on the sumnmary page of the Stability Report.

vre x \t~e cr~a lpa pc r. II 5 18.02.93

The hydrostatic data of the vessel is available within the DLS system which is used to calculate the
stability characteristics (GMI and GMt) of the vessel. The longitudinal (GMI) and transverse (GMt) ",
stability values are calculated and presented on line on the teit and a graphical screen. The total free
surface effect is presented by the variables GoGI and GoGt: The stability values including the free
surface effect are denoted as GoMI and GoMt.

Position System Configuration

The DB-102 had two Albatross positioning systems for normal vessel control, an ADP and an APM.
Each system can control the vessel thrusters, while the APM system also uses the anchor winch
tensions as input, in addition to the various sensors and position reference systems. An ADP 311
dynamic position system was used as a back-up system.

The system consist of:,

Albatross Position Mooring System (APM)
Albatross 503 Dynamic Positioning System (ADP)
Albatross 311 Dynamic Positioning System
Four Wind Sensors
Two Gyro Compasses
Two Vertical Reference Units
Printers
Artemis Position Reference System
Simrad HPR Position Reference System
Albatross Taut Wire Position Reference _System
Un-interruptible Power Supply
Stabilised Power Supply
APM/ADP/Manual Changeover Panel
SAUs (Signal Acquisitid.n .Units)
The propulsion and dynamic position-keeping system consist of six electric driven, retractable
controllable pitch thrusters sufficient to keep the vessel on station for normal North Sea operations.
All the thrusters are controlled by the dynamic positioning system supplied by Kongsberg.

:I.:'c6x I\gcicr:I\1 xip-r.II 6 18.02.93

- OFFSHORE EXPERIENCE

Anchoring and Positioning

HeMe had performed mooring analysis for the anchoring of the SSCV. The vessel was
positioned by
a 12' line wire mooring system. The station keeping system was analyzed according to NMD
criteria,
Consequence Class 2.
The analysis determined the limiting criteria, minimum safety factors for the intact and damage
system
and safety factors for the most critical lines during transient motion together with minimum
safety
factor before anchor lift for this line. The weather limitations and tensions at which anchors
lifts occur
corresponded to approximately BF 7-8.

Initially anchors were set according to anchor plan (ref. Figure 6).

Using the various types of pre-installed reference systems (Simrad HPR, taut wires and
Artemis) the
dynamic positioning system is able to keep the vessel in a pre-arranged position. Environmental
forces
such as waves, sea current and wind etc. will be controlled by thruster power with
a calculated
intensity and direction thus maintaining the vessel in its original position.

During the installation the following reference systems were used with the APM computer
system:
Microfix
Taut wire
HPR transponders.
The APM performed very well in thruster control damping mode during installation. Good
reference
systems are very important for the performance of the APM system for successful installation.
The HPR and taut-wire worked moderate, Microfix (brought in through artemis) also worked
moderate.
For better performance of the reference system a "H.P.R. deep water transponders" or
allowing the
position system to use other reference systems'will be beneficial.

When the above systems are performing well, the thruster control positioning is very
reliable. -

The 12 point anchor pattern at 310 meters water depth is rather elastic and the APM thruster
control
damping is recommendable with winds above BF 4-5.

The APM computer system was constantly used in the "monitor mode":

- monitoring
- consequence analyzer
- self diagnostics
- alarms for limiting requirements.
Due to the installation involved the damping mode, heading keeping mode, positioning keeping
mode
or composition of these three was used by the APM/DP operator.

Winch Operations
For control of the winches, the vessel, utilise APM position mooring system, which
is able to
continuously monitor all 12 mooring lines. In case of over-tension of one of more lines,
the computer
is able to propose a new anchor line configuration or, on operator request, to start thruster
in order to
release the over-tension. The system calculates continuously the resulting ship trajectory
in the event
of breakage of any mooring lines and sounds an alarm if the vessel moves out of at pre-set
safety
zone.
Another feature of the system is the possibility to simulate any anchor pattern and, by input
of changes
in environmental conditions to see the effect on mooring lines and try alternative corrective
action in
order to avoid any serious consequences.

The Compuler (lisplays the opltimal lengths and tensions for the individual anchor lines,
based on
wanted iosition/heading and existing cnvironmental conmliions. Winch operations will be
aiumi:lly by opccaling the anchor consoles. perforimed

I-£:VI'cx t\gncnral\lvar-ll 7 18.02.93

For example, when it is required:
to change vessel position
to change vessel heading
tension optimize
relieving tension in individual anchor lines.
Further, the APM computer determined where it was necessary to take in and/or pay out anchor lines
simultaneously to maintain control over the vessel. Tension optimize, the take in and/or pay out, will
not affect the vessels position or heading unit but will achieve that the load can be taken up in other
anchor lines.
This information is used in conjunction with the tension optimize function.

The display shows:

- wanted position coordinates in cartesian and polar forms
- wanted and actual heading
- actual anchor line tension and length data
- proposed anchor line length changes
- North-direction pointer (graphic)
- vessel excursions (graphic)
- vessel heading indicator (graphic).

Consequence Analysis
The consequence analysis calculation takes approximately, with a dependable updated vessels
estimator, about 12 hours computing time to fuLfil the calculation (a minimum of I to 1,5 hour per
anchor is required). :

The analysis is used to predict the conseqluences of anchor line breakage, or power blackout resulting
in total thruster failure. These predictions embraces:
- the maximum tension in a particular anchor line after a total thruster failure
- the maximum tension in the next most critical anchor line if a particular anchor line should
break
- maximum position off-rmn after total thruster failure
- maximum position off-mn after a breakage of an anchor line.

Thruster Control
The thruster capacity is used to reduce the amount of external forces to be taken up by the anchoring
system, to dampen the low frequency motion or to produce additional stiffness. The thruster system
is able to compensate for any single line failure and consist of the following modes:

Damping Mode:
The "damping mode" uses the selected thrusters to reduce vessel low frequency oscillations to a
minimum, and also to reduce position hunting when re-positioning the vessel.
Position Mode:
The "position mode" uses the selected thrusters to either maintain the vessel's position against
environmental forces, or to reposition the vessel.

IHleading Mode:
The "heading mode' uses the selected thrusters to maintain the vessels heading, or to assist in bringing
the vessel onto a new selected heading.

EI:\Text\gcnerad\papcr.Il 8 18.02.93
Conditional Mode:
The "conditional mode" can be adapted to place the selection of the first three modes under automatic
control of the operation computer.

Vessel's Estimator
The APM computer system contains a mathematical model of the dynamics of the DB-102 which
simulates the vessel response to various forces. Deviation from the specified position and heading of
the vessel, detected by the APM position and heading reference system and changes in the
environmental forces are processed by the model to generate appropriate corrective commands to the
thrusters.
Advantages in using a computer model of the unit:
- No loss of control during short term loss of the vessel's position, heading and environmental
references.
- Minimalisation of modulation of the outputs to the thruster, due to wave motion.
- Simplified combination of several reference systems to obtain optimum results.
- Extensi'e error detection in the peripheral sensors.
- Accurate position keeping is obtained by RMS calculations of all reference input signals.

SUMMARY
The installation of the Snorre CFT's by the DB-102 combined several specific aspects such as duel
crane lifting and lowering of a 6000 tonnes concrete item to 310 meters water depth and docking and
placement to within stringent tolerances.
In particular, this paper has describ..d the DLS system used to control the installation of the CFT in
the seabed and the positioning system performance. The installation demonstrated that the SSCV was
well suited to position and lower the CFT into the soil in the required manner.

ti:VI'c xt\g crierdt\;xcr.lII 9 18.02.93

.......
.. .. ...

Fl
LA C
 F - jI L I F 'T 1

.1U

FV L

u' j~

rn. C,,- D

__ __
KMOT_

i t OE UF E
 - P

L- ---
-

H II i iiiif

irziiiI IiIYIirt

!:t - til -ýI`

IV

it l FI

-r -,JI
. "N__ f; 1V 1 iillt I=L '

-L - _ _

)O .fIm -I:
Ii~ili- il Li
7afo
 7, _

I-L -L

FNP

;0-.` c:

F 7-;7 ~F

! _4

f'WT

AlI

a
a' F I

--I" i , -- - - ..........
iS. .. , -.--
 I.
~LK .1.F. I ~-

I
C.~..I A F

2 t 2 ii o .' 8 A'
0~ U Cr lii -
a D ~ a
- C,. E [

......... I. 0

---- a- I 2 -a.'

-I- -" - 0
:1::
LI ii II
-a..--
- - .-
.2I dl
X~
>2
I-; Iii ~
III -~ - C
-~'
5 ~z S *

I 2
k
-
I
/

/ I
------------------------------------
K /

N.
K

N.

N
N
S -

K.

S--.
N

------
/...V
.-. '

--....---.
I

. -.
1-
-N ____
Presented at the Short Course on:
"New Techniques for Assessing and Quantifying Vessel Stability
and Seakeeping Qualities"
Trondeim, 8-11 March 1993
(EC COMETT Program)

THE EXPERIENCE OF EXTREME EVENTS;

PHYSICAL UNDERSTANDINGS AND STATISTICS

C.T. Stansberg

Norwegian Marine Technology Research Institute A/S (MARINTEK),

Trondheim. Norway

Abstract

Physical mechanisms determining extreme responses on moored floating units in offshore

conditions are discussed, on basis of model test results. Large low-frequency horizontal
drift oscillations are seen tc-induce particularly large extreme mooring line loads. Since
the slow-drift motion is non-lifearly excited, and therefore non-Gaussian, the extreme
offset may become significantly larger than estimated by standard methods. Non-linear
characteristics of catenary mooring lines may further increase the extreme loads.

1. INTRODUCTION

Estimating extreme responses of moored floating systems for offshore operations is not
straightforward. The primary reasons are the combined motions made up of wave fre-
quency (linear) as well as low-frequency (non-linear) wave loads, and also possible non-
linear characteristics of mooring lines, risers etc. The low-frequency motions may be
particularly important due to the long eigenperiods (- 1/2 - 3 minutes) of the moored
structures. Special procedures and statistical methods should be applied for analysis of
such cases. It is the intention of the present paper to discuss proper analysis methods seen
in light of the physical mechanisms behind the responses. The study is a review based on
experiences from a project carried out as a part of the Norwegian FPS.2000 research
programme [I]. This project was in particular dealing with the problem of extreme
motions and line tensions of a turret moored production unit.

199zc 5Iklzp '\,.t'vll t

2. AN EXAMPLE.

The motions and mooring line tensions of a turret moored production ship in random
ocean waves have been extensively studied through 1:70 model scale experiments in
MARINTEKs Ocean Basin [2,3,4.5]. This was carried out as a part of the Norwegian
FPS2000 research programme. In particular, the statistical nature of the extreme events
was investigated. For this purpose. tests with very long records, corresponding to 18 hours
full scale, were made. An important point here was the generation of long random wave-
trains with proper, stationary wave statistics. without repeated sequences of the signal. A
steady current and wind was also included. Their main influence was that they determined
most of the mean offset and mean direction of the vessel.

Some details of the test model set-up are shown in figure 1. The mooring system was
modelled by realistic catenary lines, which is important in order to simulate proper motion
and line tension characteristics.

Essential results from the tests are illustrated in figure 2. Time series samples, power
spectra and statistical probability distributions of wave elevation, surge motion and
mooring line tension are included. The time series samples show the most extreme event
observed, occurring as a purely random evefit in an 18-hour storm record. It is seen that
the extreme offset and the corresponding line tension are mainly dominated by a large
low-frequency component, which are inductd by 2nd order non-linear wave drift forces.
The event is generated by a random, particular combination of waves that are not nec-
essarily the largest of the total record, but rather representing a group of successing large
waves.

The power spectra in figure 2 clearly confirm that the measured surge and tension records
are each composed of a low-frequency as well as of a wave-frequency contribution. This
is also reflected in the probability distribution diagrams: Surge and line tension are clearly
deviating from the "standard" Rayleigh distribution, while the wave heights are seen to
follow the Rayleigh model quite well. The deviation for surge and tension is due to the
low-frequency contributions, combined with non-linear characteristics of the catenary
mooring system. The "message" from the distribution plots is that the random, extreme
event shown in the time history plots, for which the measured surge and tension values are
far beyond estimates based on the commonly used Rayleigh model, is not a singular event
falling outside the family of general observations. It is rather a natural part of a statistical
population deviating from the common, simple model. The origin and nature of such
events is the subject of the following chapters.

( ):\Vck5 \
I1939.0(),)
12
3. RESPONSE STATISTICS OF MOORED FLOATING VESSELS IN
RANDOM WAVES - A BRIEF GENERAL FORMULATION.

A general random response y(t) to a dynamic, randoni input wave elevation signal x(t)
may be written as:

y(t) = Y1 (t) + y(t) . + Yt) + .. ()

where

Y(t)= f dt' x(t') h1(t-t') (2)

is the 1st order (linear) response, and Y2 (t) ....

yY(t),.. are the higher-order (non-linear)
response contributions. hl(t) is the linear impulse response of the response system. The
linear term is of particular interest because of its simple relationship to the wave. In the
frequency domain, it is characterized by the one-to-one relationship between wave and
response frequencies:

Y1 (f) = Hi(f) X(f) (3)

while for nth order terms, a given response frequency is coupled to the interaction between
n wave frequencies. In eq.(3), Y1 , H, and X are Fourier amplitudes (transforms of y,(t),
hl(t) and x(t), respectively:

YI(f) = f dt yl(t) exp(-j2rct)

Hi(f) = f dt h1 (t) exp(-j2,tt) -(4)

(the linear transfer function)

X(f) = f dt x(t) exp(-j2nt)

3.1. Linear response statistics.

We now assume that the wave elevation signal x(t) is a Gaussian distributed, stationary
random process. This is most often a good approximation for the description of linear and
low-frequency responses. The statistical behaviour of linear response signals y1 (t) may
then also be shown to be Gaussian, as a result of the simple relationship in eqs.(2-3). As a
result, the largest wave and response amplitudes A may be assumed to be Rayleigh

( )I ck
I. S mtpt'l\wcgcIt i .tI:;
I It'193ft U.'))
3
distributed:

P(A>a) = exp[-a 2/2 cay2] (5)

if the mean value (ymean) is neglected. Here cGy is the standard deviation of the parent
time series y, (t). The expected extreme value in a record with N amplitudes may then be
roughly estimated as:

<ymax> = ymean + oy XZlN (6)

In a typical 3-hours record with wave periods around 10-15 seconds, this gives dynamical
extremes in the range of 3.5 - 4 times the standard deviation cya.

There is, however, a considerable statistical variability connected with the expected
extreme of a given record. This variability may be quantified by a standard deviation
0
ymax around the expected <ymax>. Its magnitude is approximately 30-40% of the record
standard deviation t.

3.2. Second-order resoonse statistics.

In many cases, especially for vertical vessel motion responses like heave and pitch, the
above linear response mode:often gives a good representation of the real world. Incertain
other cases, however, the total response also includes significant higher-order contribution.
In general, for the horizontal motion of large moored floating structures, the 2nd order
term may become the dominating one, since the low-frequency resonant oscillations of
such systemý are easily excited by 2nd order wave drift forces. The 2nd order Fourier
amplitudes Y(f) of the response may often be approximately written as:

Y2 (f) = HR(f) f dfo X* (f0 -f) X(fo) H2 E(fo ,fo-f) (7)

(from which the 2nd order response y2 (t) may be found from a simple Fourier transform
as in eq.(4))

where H2E(fl.f 2 ) is the 2nd order (quadratic) transfer function (QTF) of the hydrodynamic
wave drift force on the vessel, and HR(f) is the dynamic transfer function of the low-
frequency vessel oscillations determined by the mooring system. For more general 2nd
order descriptions, see e.g. ref. [5,6,7].

In contrast to the linear responses, 2nd order (and in general, non-linear) responses may
not be assumed GatIssian even though the input waves are Gaussian. The statistical
behaviour of a response like the one in eq.(7) is determined by the system transfer
functions H1 2E (f 0
1 .f2) and HR( -

( ):\ck5 I4p:.1 w \VA

'22Lct Is
1993.11.(.I) 4•
 With H2 E and HR set equal to I , we simply get the case expressing the square of
the

wave signal:

y (t)= x 2 (t)
(8)
which is illustrated in figure 3. It is seen from this figure that a 2nd order
response is
composed of a low-frequency (difference-frequency) as well as of a high-frequency
(sum-
frequency) component, and that the square process leads to a relative amplification
of the
largest amplitudes. It can be shown that the amplitudes A of the low-frequency
part of
x 2 (t) (actually describing the wave group formation in the time domain) are
exponentially
distributed:

P[A>a] = ext[-a/a y]
(9)
from which it can be found that for a given record standard deviation cay,
the probability
of large responses is significantly increased compared to that for Rayleigh
distributed
amplitudes. The expected extremes are:

<ymax> = ymean + a ylnN

(10)
with a connected statistical variability cy max = 1.3a Y. In practice, the
largest normalized
amplitudes of'a 3-hours record are about 1.5 - 2 times those of the Rayleigh
distribution.
It is also interesting to note that the statistical variability of a sample extreme
is increased
by a factor of 3.
For a general 2nd order system with 2E' HR not equal to 1,the probability
distribution is
falling in between the 2 outer limits represented by the Rayleigh and the exponential
models, see figure 4. Exact solutions for low-frequency as well as for high-frequency
systems may be found in literature [7,8]. A simplified statistical model for
slow-drift,
described in [5.9], is briefly discussed in Chapter 5.

4. WHICH BASIC PHENOMENA DETERMINE THE EXTREME MOTIONS

AND LINE TENSIONS?

As indicated in the example of Chapter 2, the extreme events (mooring line

forces) in a
random sea occur as consequences of large non-linear slow-drift oscillations
of the vessel.
These oscillations may most often be more or less described as 2nd order
responses
described in Chapter 3, for which significant deviations from the commonly
used Rayleigh
statistics should be expected. See figure 5, where results for low-pass as well
as high-pass
filtered records from the model test results in figure 2 are shown. As seen
from this figure,
the wave frequency motion and tension also contributes somewhat to the total
motion, but
the basic underlying process leading to the extreme tension loads normally
turns out to be
tie slow-drift motion. It is especially easily seen in moderate
sea states with shorter
waves, illustrated by figure 6, but it is also still often the case in storm waves.
Therefore
the understanding of the slow-drift motion statistics is essential, and will be
briefly
discussed belov..

: ck5 I k; tw r\": t- lI.c Is

Furthermore, the non-linear characteristics of the mooring lines, such as time-varying
mooring stiffness and line drag forces, are important, especially when combined with the
extreme low-frequency motion. These are also discussed later in this Chapter.

It is evident from the discussion so far that extreme events with moored floaters is not a
straight-forward linear problem. In such cases, a thorough description and understanding of
the phenomena for a given system may sometimes require particular model tests and/or
sophisticated non-linear time domain numerical simulations. However, some general
guidelines describing the main mechanisms may be established for this class of problem,
as will be discussed in the following.

4.1. Slow-Drift Motion Statistics

For the physical understanding of the slow-drift motion statistics, a simplified model is
described as follows. Assume the exciting drift forces to be proportional to the wave group
energy signal E(t) (defined as the low-frequency part of the square wave elevation in
eq.(8)), and the dynamical behaviour of the moored system to be described by a linearly
damped harmonic oscillator. The response Fourier amplitude may then be formulated as:

Y2 (f) = HR(f) HD(fP) f dfoX*f 0of) X(fo ) (11)

where HD(fp) is the hydrodynamic surge drift force coefficient

at the spectral peak wave frequency fp. (Thus the simple model assumes constant drift
force coefficients in the actual wave fequency range). In the time domain, this gives:

y2 (t) = HD(fP) f dt'E(t') hR(t-t') (12)

where hR(t) is the Fourier transform of HR(f).

As shown in Chapter 3, the square wave envelope signal E(t) is exponentially distributed,
with large probability for large values. The statistical behaviour of the response y2 (t) is
then given by the low-frequency dynamical response characteristics hR(t) (or HR(f)), or, in
other words, the slow-drift damping characteristics and eigenperiod. This is illustrated in
figure 7, where time series sequences obtained numerically with the above model, with the
damping as a variable parameter, are shown. A clear gradual change in the statistical
behaviour is observed: With large damping, the response statistics resembles that of the
asymmetric, exponentially distributed exciting signal. With low damping, the response
statistics is more symmetric, with a relatively lower extreme (the absolute magnitude is
increased, though). This is also related to corresponding changes in the response spectra
(figure 8). In general, the slow-drift motion response will follow a distribution in between
the exponential and Rayleigh models, as demonstrated by the model test results (figure 5).

( ):\ck5 Ikp:pc:\wcpciitc is6

1993.(13 .O)c (6
A simple physical picture may be applied to explain the behaviour described here: With
large damping, only one or a few independent wave groups are instantaneously "rememb-
ered" by the oscillating system, which means that the statistical nature of the excitation is
reflected in the response. With low damping, the system memory is longer, i.e. the
instantaneous response is a result of many independent wave groups. Thus with a gradual-
ly decreased damping, the response will gradually and asymptotically approach a Gaussian
process, according to the commonly known Central Limit Theorem of statistics.

The damping of the slow-drift motion is a central point also for other reasons. First, it
should be mentioned that in rough sea, the total damping may be considerably increased
due to drag forces on the catenary mooring lines [10]. Time varying wave-drift damping
may then also become large [11]. For the case referred to in Chapter 2, the total damping
was as high as 20 % due to these factors. Furthermore, non-linear damping contributions
may sometimes reduce the largest extremes.

A more detailed description and discussion of the model referred above is found in ref.[9].

The statistical variability of sample extremes of I-hour records is illustrated in figure 9.

2 independent experimental realisations of the same slow-drift process are shown. The
large variability qualitatively confirms the general prediction noted in connection with
eq.(10). See also ref. [4].

4.2. Non-Linearities of Catenary Moorin, System

Extreme mooring line forces, occuring as a result of the extreme slow-drift motion
described above, may be "amplified" due to the non-linearities in the mooring system.
First, a non-linear quasi-static stiffness, as shown in figure 10, will increase extreme low-
frequency components, as shown in the example in figure 11 taken from the experiments
referred in Chapter 2. Simnilarly, the non-linear stiffness will also enhance the largest
wave-frequency tension amplitudes. The wave-frequency component is, however, not a
quasi-static response to the surge motion, but rather dynamical response to the tangential
upper-line-end motion. An additional, significant contribution to the largest tensions is
therefore due to dynamical effects from inertia and non-linear drag forces on the lines,
which results in an instantaneous, dynamical mooring line stiffness considerably higher
than the quasi-static one. For very high frequencies it will asymptotically be given by the
elastic, axial stiffness of the line.

In practice, the net effect from the non-linear motion and mooring line characteristics may
represent a doubling of the expexted extreme line tension relative to estimates based
directly on the Rayleigh model, i.e. a considerable effect.

More details on the extreme mooring line loads are discussed in ref. 131.

( ):\ck5 I\ :i;•ci\wcgeid .crx

I1993.)•1.0) 7
5. SIMPLIFIED MODELS FOR EXTREME RESPONSE PREDICTION.

The prediction of extreme slow-drift motion may in general be based on the 2nd order
formulation briefly described in Chapter 3. A complete theory on the extreme estimation is
given in ref.[7]. With the simplified model in Chapter 4, the prediction problem is
significantly reduced, and may basically be connected to the average linearized damping
level of the slow-drift motion. A simple statistical model based on this has been estab-
lished [9]. For a system with a given eigenperiod in a given sea state, the expected
extreme may be established as a function of the damping level and the duration of the rec-
ord, as shown in figure 12. The model may formally be written as:

c<ymax> = yrhean + oy [A InN + B] (13)

where A and B are damping-dependent coefficients. See ref. [5] & [9] for more details.

The prediction of the connected extreme tension load is more complicated. In general,
long-duration model tests and/or non-linear time-domain numerical simulations may be
necessary. A simplified procedure might follow the outline given below. See ref.[3] for
more details. First, the extreme quasi-static low-frequency tension may be found directly
from the extreme motion estimate via the tension/offset curve (figure 10). The connected
wave-frequency tension component connected with this event is strongly coupled to the .
low-frequency component, and is physically not separable from it due to the time-varying
quasi-static and dynamical mooring stiffnesi However, a rough estimate of the wave-
frequency contribution may be established if the average dynamical transfer function
(RAO) between the in-line wave-frequency vessel motion and the corresponding tension is
known (e.g. from numerical computations or from model tests).The wave-frequency
motion may be assumed Gaussian distributed. Thus the corresponding extreme tension
may first be estimated by a standard frequency plane analysis using a Rayleigh estimation,
and then amplified by a factor reflecting the "typical" RAO (stiffness) amplification in a
large wave group, relative to the average RAO. (An estimate of this amplification should
therefore be available).

If the low-frequency and high-frequency extreme tension estimates then are added linearly,
a conservative total extreme estimate is normally obtained. In some cases, a reduced corre-
lation between these 2 components should be assumed, resulting in a consequently lower
extreme estimate. It is, however, most often difficult to estimate this correlation reduction
without some records available from model tests and/or numerical simulations.

:t193
A)• ,t.)
 6. CONCLUSIONS.

The experimental examples have demonstrated that extreme motions and mooring
line
loads of a floating production system must be estimated by more advanced methods
than
the commonly used Rayleigh model estimation. Extreme response amplitudes
may in
certain cases be of the order twice the Rayleigh-based estimates. The reason is
the large
influence from non-linear physical mechanisms, such as 2nd order slow-drift motion
oscillations and non-linear mooring characteristics. This may lead to probability
distribu-
tions significantly different from the Rayleigh model. A considerable statistical
variability
is also connected with the extreme of a sample record. It is increased by a factor
of 2 - 3
relative to the "standard" Rayleigh case. At the same time, a considerable damping
(-
20%) of the slow-drift motion is observed, due to drag damping on the mooring
lines and
to wave drift damping. This should also be taken into account when the estimation
of
extreme loads is concerned.

7. ACKNOWLEDGMENT.

The model tests were financed by the Norwegian FPS2000 Research Programme,
sup-
ported by: the Royal Norwegian Council for Scientific and Technical Research
(NTNF),
Elf Aquitaine, Esso Norge, Det Norske Veritas, Norsk Hydro, Petrobras, Statoil
and
Tecnomare.

():\•ck51 "iN•pc r\,,.'c gtr Crni,:.

I l)3.( I 'i.1)l
0
8. REFERENCES.

1. Fylling, I., Stansberg, C.T. and Mo, K., 1992, "Extreme Motions and Anchor Line
Loads in Turret Mooring Systems", Proceedings, BOSS'92 Conference, London,
U.K.

2. Stansberg, C.T., 1991, "Model Tests on Extreme Motions and Mooring Line Loads
of a Turret Moored Vessel, Part 2", MARINTEK Report MT3005.19.01, Oct. 1991,
(Restricted).

Stansberg, C.T.. "On the Estimation of Extreme Mooring Line Forces", Proceed-
ings, the 11 th OMAE Conference, Calgary, Canada, June 1992.

4. Stansberg, C.T.. 1992, "Basic Statistical Uncertainties in Predicting Extreme 2nd

Order Slow-Drift Motion". Proceedings, the 2nd ISOPE Conference, San Fransisco,
Cal., June 1992.

5. Stansberg, C.T., 1992. "Model Scale Experiments on Extreme Slow-Drift Motion in

Irregular Waves", Proceedings the BOSS 92 Conference, London, U.K.

6. Stansberg, C.T. 1983, "Statistical Analysis of Slow-Drift Responses", ASME

Journal of Energy Resources Techn6logy, Vol. 105, pp. 4 6 5 -4 7 4 .

7. Ness, A., 1986, "The Statistical Distribution of Second-Order, Slowly Varying

Forces and Motions", Appl. Ocean Research, Vol. 8.

8. Nass, A. & Ness, G.M. 1992, "Response Statistics of Tethered Platforms in

Random "Waves", Appl. Ocean Research, Vol. 14, pp. 23-32.

9. Stansberg, C.T. 1991, "A Simple Method for Estimation of Extreme Values of
Non-Gaussian Slow-Drift Responses", Proceedings, Vol.Ifl, 1st ISOPE Confernce,
Edinburgh, Scotland, pp. 442-451.

10. Huse, E., 1991, "New Developments in Prediction of Mooring System Damping",
OTC-6593 Paper, Proceedings, 23rd OTC Conference, Houston, Texas.

11. Aanesland, V., Kaasen, K.E., and Krokstad, J.R., 1992, "Wave Drift Damping of a
Turret Moored Ship", Proceedings, BOSS'92 Conference, London, U.K.

( :\,eLk5 Ikmpaet\w rc
gcmt .týl:
 CATENARY MOORING

FPS 2000

A' c)hoJ

Figure 1. Illustrating the vessel with one of 8 catenary lines.

TEST SEUJtES662 - 666 -

POWER SPECTRA PROBtABILITY

DISTRIBUTIONS
T IM E HII
STO RY S K P I S-
I
INLDIGEXTREM I I WAV E-
II H-

L IZ

20SRG
SURGE____

*i
01
ur [.p I t .. jr....
, ;jjiy fv i u..
-s:"!C:T 4 L fi 1 .14j id urn
 ll??, ( i2 AI
W itS,.

I'm 270
M 701 250 90 2150 3000

Rw M27. CO!Q?hl2
?t15 q1

5
270 25) M 28 0 2150 300

'2

• ..
:9

:3

Figure 3. The square wave elevation process.

H [}
mI]3 10-T
o
510
WU 5"10-;

W -

?RayleigV
L distribution
-d 5

2 3 4 5 6 a 10 15

NORMALIZED AMPLITUDE -n•

_-yz-YJ-Q------•A

-igure 4. Thu Rayleigh and exponential anplitude distributions (WcibtlI scale).

r TF5" SERIES 061 .

POWER SPECTRA PROBAIIILITY

TIME HISTORY SAMPLES _ _ _ DISTRIBIUTIONS

INCLUUDIRNG
EXR'ECE SUR GE
5 jI SUGESURGE

20

- ST = 1 s = 13. . + o

1 I I
.A Tm Ai-smles power spcr an prbbiiyditiutosfrm18hu.
is~tor etSeis

Fiz. 5.A Time hislory samples, power spectra and probability distributions from
iS-hours tes series.

Sea Stlae: T, = 15 1, = 13.1 an. Windt Curn.

Low-frequency components..
__ ____ ____ TESIO CSUeRG ENSO SORMLCE

TEST SERIES 661 66

POWER SPECTRA PROBABILITY

DISTRIBUTIONS

___ ___ __, _'______. ___ _ * ... ..... ...

+ • s-- --

II.
........
, .._,___,

lig. > IIIIl'IIt . ItIry '+:IH~l "II',

i• Nike'
r : :,11111
I*ItJlbklhINIly (liMtIbI, horI,I rIrnt IS Iha•ur Ic~I S+tileN'.

Sot':. St;:I':'I " :1 IY'P II, i I f+lm. WVn~l .+ (>Irr~cII.

 0 0

- E._E w

0. c 0
"c 1 0
- -t, ; C oc i
/;>

o " - o s'

ato to
t ) a
o0 -.

> -

" Ix C ' _
-)~ 0j'

• 0

C C
0

0•
< I
•00

0 <. o-

cc L

..... .. . ... ..

a: 0
 4.-eu, U
C
I-..-. C,
*0o.
lEO a
C,

K~0Vh I-
C-

N
C)

tn
H'
A. a a
jJE~
Ct
.3-on
I,
:0 I
I o
* C)
I- C)
L-.
C)

~* Etfl C.)
C)
I,
.3 Ur'i
0 ii CA
SO E I..
Ct C a)
C
0~

a C)
-~
.-. 0'
to
I -4 4~-4

kfluL)ads ~3%Od

CA
C
0
- a)
(5 - C
zi.. I H ~ E
0~2 II -4 -C,, I-
H 0'- C)
0-ri
x-z I - -c
I-
"-4 C) ~O
.;- -, I

[W I - ____

Ct -~~*~-----*--~-- 4 I - - - C 0a)
o
____ ______ ___ ______ ______ 4- 0.

1t
I -

LKS~L - = -

-
_________

~
I ____________________________

___ ___

N
4)

to
 TIME HISTORIES. 2 1-HOUR RECORDS

REC. Al
aY 0. 917:
3~ý .Ta-4 .224i

* ______R-EC a1
a ý0.652m
I - "aax=2. 975m

IV§V9

Fig-ure 9. Statistical variability of slow drift: 2 different 1-hours realisations.

7
-modeltests,
21% damp. 30%
=theory,with
6 damping level indicated

2
50 100 200 300 500 1000
NUMBER Or PEAKS

15 gtire 12. E:x pec ted and ninas ii cd slow -cIft- It110!ion extremnes.
 0 0
o -~~0 -C
2 -

___ ~~~1- . I C -

~0 _ t. C.)

_____ _____

~i
_____ _____ LI..
[2 .................................................................................................
____ ____
____
____
____ ____ 0
o0- 1=2
.
I 0
0~ -o
-~-
~LL
C-)
C.)
C-)
'-S

C.)0
___ - 00 -a
C t4~~
0
I- C-'
U-, -
o 0
__ __ o -~ 0 CL
-, I -2
II -
-c
C)
-L-.I--L± -~---t--k-*~ -A--J-..kJ... -h---A--h--4~ .t.-A--I~1... 0 ~ I 4.t-LJ.
It) 0 '0 0 '0
-
0
-
-, 0
-~
0
0
a
0
-0
0
0
0
0
0
0
0
0
0
_
0
0
It)
'0 --
2
-

C-) r..~
I-,
:3
r~~j 06.Jns ji (MT C3JO3 *UIi
On

~0
0
Cd)

0J

-C
C-)

o '4-

In
0
I C.)
C-)
a, ___________ ___________
c
____________ In 0
-g (4-~
C)

~ C
a
V

a
_____ _____ o~ ~'
C'-4 C
a
~ C
C)
C
a, 0
'~ '~C)
C0 _____ 0 -~
______ _______ ~.n o -
0
- C.)
o
-a 4-. -o
~ 0
Cr
0
0.'
(a
C
In 2
C
______
_________ 0
- :3 0
I-.
C

to
0
0
(0 0 0 0 0 0 0 -
o 0 0 0 0
o C) C' Ci 0 0)
tj~) -r c-j -

:3
On

N~I 3)JOJ DU9

 POSITION CONTROL - FROM ANCHORING TO OP SYSTEMS

by
Dr G.E.W. Wichers
Maritime Research Institute Netherlands

INTRODUCTION
In the last decade the design of offshore mooring
of floading techniques have changed. The changes systems and the
in the develop-
ments have been made possible by a few important achievements
as
for instance:
- the increased reliability of flexible risers;
- better understanding of the low frequency motion behaviour
moored ships in waves, wind and current; of
- prototype experience with the increasing number of moored
Floating (Production) Storage Offloading (F(P)SO) systems;
- application of dynamic positioning (DP) on large
off loading
tankers.
modern designs of F(P)SOs and Offshore Loading Systems
a reduction of the large variety in the past. In the (OLS) are
past
systems designed and ap plied 'were the compliant articulatedtypical
tems. The following concepts apply to the majority of sys-
the present
offshore designs:
- For permanent mooring systems: internal or external turrets,
many cases provided with composite lines. in
- For semi-permanent mooring systems: disconnectable underkeel
buoy or riser turrets and DP systems.
- For off loading: tandem mooring and DP.
For the design of the systems in the pre-design stage
retical empirical mathematical simulation models can semi-theo-
be applied,
* while for the verification of the design model experiments
indispensable. In the design process the physical phenomena are
control the motions of a moored or DP tanker must be which
The physical phenomena understood.
are the exciting loads, the reaction
forces and the motion control system (mooring system or DP con-
trol). A review will be given of the basic approaches to assess
the motion control of a permanently moored and a DP
controlled
tanker exposed to wind, waves and current.

THlE TANKER MOTIONS

A tanker kept in position in irregular waves, wind
will be exposed to wave frequency loads- in the six and current
degrees of
freedom (DOF) and mean loads in combination with slowly varying
loads in the three DOFs in the horizontal plane.
The wave frequency loads are caused by the well-known first
wave forces. The mean loads are due to mean wave drift order
force,

-I-
current and wind loads. The, slowly os~illating loads are caused
by the second order slowly oscillating wave drift forces and
loads as induced by the low frequency components of the wind
speed spectrum. In some cases low frequency excitation can even
occur due to macro vortices in a turbulent current or due to
shedding from the vessel itself.
The result is that the vessel will perform wave frequency motions
and at the same ti~me is subjected to a mean displacement and low
frequency motions. The wave frequency motions introduce the wave
frequency dynamics in the mooring legs of a mooring system. For
oP systems the,' wave frequency motions can affect the effectivity
of the thrusters. The most important part of the power needed for
station keeping of a tanker is caused by the mean displacement
and the low frequency motions. The control, system of a DP system
has to react. to these motions, while the force level in a mooring
system will be dominated by these motions.
In order to describe the low frequency motions the wave drift
forces, the wind and current loads and the low frequency hydro-
dynamic reaction forces have to be known. Each of the mentioned
components will be briefly described in the next sections.

WIND, CURRENT AND WAVE DRIFT FORCES ON A TANKER

Current and wind loads

For a tanker-based FPSO the current loads can be derived from the
updated current coefficients as given by OCIMF. The coefficients
in longitudinal, transverse and yaw direction will be given not
only for the fully loaded draft as function of the ratio
draft/water depth but also for the ballasted condition (even
keel) of tankers with cylindrical and conventional bow.
Wind load data for a tanker with standard superstructure and
without process equipment can be found with the OCIMF data. To
account for the process equipment and possible interaction with
the superstructure computer programs based on wind tunnel tests
are available or wind tunnel tests have to be carried out.

wind spectra
Because of the importance of the low frequency excitation on
moored structureis wind spectra can be applied to the tanker. The
spectral denfsity in the low frequency part of the wind spectra
can be considerable. For various formulations of wind spectra the
spectra based on an hourly mean wind speed of 30.9 rn/s (10 mn
elevation) is given in Fig. 1. In [1] the results are given of
the effect of the application of wind spectra and a 1-minute gust
on a moored 200 kDWT tanker in co-linearly directed irregular
waves and wind.
The results show that due to the 1-hour mean wind speed the mean
displacement is small but that the oscillating wind force causes

-2-
 a most probable surge motion which generally
is lower than the
values found for the steady i-minute gust.
Wave-current interaction on wave drift
forces
The wave drift forces contribute generally
the total low frequency excitation for a dominant part to
tanker. In order to determine the on a moored or DP controlled
mean
wave drift force it is common practice and slowly oscillating
transfer function (QTF). to compute the quadratic
By means of spectral or time domain
techniques the mean and the oscillating
determined. The computations are wave drift force can be
based on 3-Dl potential theory
and are valid for the zero-speed condition,
see for instance [2].
The QTF of the wave drift force on
a
erably when moored in a current field.tanker may increase cons id-
dition the speed dependency of the In the zero-current con-
slowly oscillating moored tanker will wave drift force due to the
dependency of the wave drift force be present also. This speed
QTF of.the wave drift damping. In can be distinguished by the
waves combined with current the wave case the tanker is moored in
by the current speed dependent QTF. drift force can be described
tions of the vessel The slowly oscillating mo-
in the current field are governed by the-
current speed associated QTF and the
of the wave drift damping. For head current speed associated QTF
cedures are given in [.3]. As an* examplewaves the computation pro-
force with and without-current~and the QTF of the wave drift
the wave drift damping for an
LNG carrier is given in Fig.> 2.
Besides the QTF in head waves and
drift forces for waves combined with head current, also the wave
arbitrarily directed current
are necessary. The general computation
opment, see [4]. An example of the procedure is under devel-
QTF for another direction of
wave and current is shown in Fig. 3.

LOW FREQUENCY HYDRODYNAMIC REACTION FORCES

In the aforementioned section the mean
tion forces are briefly explained. and low frequency excita-
Due
moored or DP controlled vessel will performto the excitation the
in the horizontal plane. Besides the low frequency motions
hydrodynamnic forces will react to mooring and DP system also
the
placements, velocities and accelerations. slowly oscillating dis-
drodynamic reaction forces can be split The low frequency by-
velocity dependent terms. The acceleration in the acceleration and
low frequency added mass and can be terms consist of the
computed with 3-D potential
theory. The low frequency velocity dependent
sist of the viscous resistance. terms, however, con-
These terms cannot be computed
accurately and have to be derived from
model tests.
For the derivation of the low frequency
forces and moment distinction has to hydrodynamic reaction
be
and current. Compared to the still water made between still water
that the resistance forces/moment may resistance it is obvious
strongly be influenced by
the presence of the current (lift and
drag forces).

-3-
In the following the description of the low frequency hydrodynam-
ic reaction forces in terms of the equations of motion will be
elucidated.

Still water
For the equations of motion the following formulations are as-
sumed, see [3]:

(M+a1 )5X
1 (M+a2 2 )x2 x 6 + a 2 6x 6 +XIS W

M+a22)M2 + a26M 6 = -(M+a 1 )x1 x 6 + X 2 Sw

66)6+ a 6 2 '2 - -(a 2 2 -a)x 2 - a6 2 1 6 6S W

in which:
XlS w , X 2SW X 6S W = low frequency viscous fluid resistance
forces/moment-components
a*l, a2 2 , a66, a2 6 , a 6 2 = low frequency added mass coefficients
xi x2 , x 6 = surge, sway and yaw motion
M, = inertia properties of the vessel.
-
In order to determine the low frequency viscous resistance
force/moment components in still water caused by sway and yaw
modes of motion, planar motion mechanism tests may be carried
out. For the surge mode of motion extinction tests in still water
can be carried out. The resistance coefficients as derived from
the extinction tests in surge direction are given in Fig. 4. The
resistance coefficients in the sway and yaw direction are given
in Fig. 5. For the sway direction the damping coefficient B 2 2 and
B6 2 were obtained as follows:

FP
X = -½p T B2 2 f x2 Ix
2Idl = -½p T B2 2 (FP-AP)x2 jx 2j
AP
FP2
2
X62= -½p T B 6 2 f 21x211 dl = -½p T B6 2 (FP -AP2)x 2 1x2 1
AP

while in yaw direction the coefficients B 2 6 and B 6 6 can be deter-

mined from:

FP 1 3 3
x -p T B2 6 I x6 11x 6 l1dl = T B 2 6 (FP +AP)x 6 1 1
AP

X = -½p T B 6 6 F x6 11x 6 111 dl = p T B 6 6 (FP 4+AP)x 6 1 6 1

AP

-4--
in which:
X2 2 , X 6 2, X2 6 , X6 6 = measured forces/moment, in-quadrature with
the applied displacements'
AP = ordinate of aft perpendicular
FP = ordinate of fore perpendicular.
Knowing the resistance coefficients B 2 2 ' B 2 6 , B6 2 and B 6 , the
coupled sway and yaw low frequency resistance forces/momenE can
be described assuming a strip type distribution of the transverse
coefficients over the length of the vessel. Decoupled in surge
direction the resistance formulations are as follows:

XlSW = 11 X1

FP
X2SW = -'p T f C(l)(x 2 +x 6 1)1x 2 +x61dl
AP
FP
X6SW = -½p T f C(l)(x 2 +x 6 1)1x
AP 2 +' 6 111 dl

In current
For the equations of motion the following formulations are as-
sumed, see [3]:

(M+all)i•
1 = (M+a 2 2 )x2 x 6 + Xistat + Xldy n

(M+a22) 2 + a 2 60 6 = -(M+a1 1 )x1 x 6 + X2stat + X2dyn

(16+a66) 6 + a 622 X6stat + X6dyn

in which:
Xlstat = ½p L T C c cr)V 2r
statc cr cr
X = ½p L T C O(P)V 2
cstat cr

being the quasi-steady current forces/moment components, where:

_ 2 v2)½
V = (u2 + v = relative current velocity
cr r
'cr = arctan(-v r/-u ) relative current angle of incidence

-5-
and the dynamic current load contribution is assumed to be:

X idyn = -(a2 2 -all)V c sin (c-X6)X6 + XlD

X2dy n = -(a 2 2 -al1 )Vc cos(Pc-x6)X 6 + X2D
x
X6dyn I
potential part
I
+
60
IL+ X 6D
viscous part

To determine the dynamic current contribution as acting on a

tanker, yaw and sway oscillation tests may be carried out in
current. A typical result of a model test to determine the dy-
namic current contribution in surge, sway and yaw direction due
to the motion in yaw direction in 2 knots current is given in
Fig. 6. By means of the dynamic current'load contributions as
function of current speed, oscillating velocities and current
angles a non-dimensional database can be established.
By means of the above given descriptions the low frequency hydro-
dynamic reaction forces can be implemented in the equations of
motion.

PERMANENT MOORING SYSTEM

Introduction
An example of a permanent mooring system can be a turret-chain
system. The turret is a turntable connected to the bow or to the
keel of a tanker. To keep the tanker in position radially spaced
mooring lines are attached to the top of the turntable. The num-
ber of mooring legs can range from 6 up to 12 or more. A sketch
of such a system is given in Fig. 7. In the turntable swivels are
mounted for transfer of the oil. The oil is transported through
flexible risers running from the turntable to subsea wells or
pipeline end manifolds.

In order to design the mooring system pre-studies and model tests

have to be carried out. From the results of the pre-study and the
model tests the size and pattern of the mooring legs (chains or
chain-wire combinations), the position of the turntable (stabil-
ity) and the strength (survival and fatigue loads) have to be
obtained. Besides these aspects also attention has to be paid to
the layout of the flow/control lines.
Further attention has to be paid to the statistics of the input
and output. The input concerns the statistics of the wave, wind
and current climate like the determination of their proper prob-
abilistic joint occurrence. The results of the computer simu-
lations and the model tests are the output. The statistical
evaluation of the results like confidence intervals of the most
probable maximum, joint probability of low and wave frequency
quantities and extreme effects such as slamming and green water
have to be carefully considered. A review of the probabilistics

-6-
 for the input data and the statistics
of the output is given in
[5].
As an introduction a simplified exercise
assess the low frequency motions of will be carried out to
tanker. This exercise is followed by a moored 200 kDWT loaded
method of the complete analysis on a brief description of the
a turret moored tanker in
irregular head waves.
Simplified method to assess the low frequency
tanker motions
For the simplified exercise use is made
kDWT tanker co-linearly moored in irregular of a fully loaded 200
The mooring system consists of a linear waves and current.
surge direction of the tanker. The spring above water in
water depth. The body plan of the tanker is moored in 82.5 m
tanker, the main particulars
and input data like QTF of the wave
drift
drift damping are given in [3]. The exercise forces and the wave
of the current and the stiffness of concerns the effect
quency the mooring on the low fre-
motions. Therefore computations are
2 knots current and without current carried out with
and with two stiffnesses of
the mooring system viz. 6.8 tf/m and
60 tf/m.
Because of, the linearity of the equation
frequency motions in surge direction of motion of the low
carried out in the frequency domain. The the computations can be
given below. In the- frequency domain computation procedure is
the variance of the low
frequency surge motion will be" see [3]:

2
oX, =C b c SF(Pe)

and the most probable maximum displacement

according to Longuet-
Higgins gives:

Xlmax = XI + -2
* ax 1 - N
x1

in which:
b = total displacement
C1 1 spring constant
N number of oscillations =
=--tc duration
1 natural of time
period of system
X, = it /C 11 = mean displacement of the tanker
Xt
i= total mean force.
The mean drift force in irregular waves
becomes:

F =2f S ()T(w,w) du)

0

-7-
The mean wave drift damping coefficient will be:

= B(W)
02 S(W) 2 d
0 _ýa
while the spectral density of the force becomes:

8 f Sj(W+p) S((w) T (&+pw) dw

0
where:
Sz(() = spectral density of the wave elevation
T(w,w) = continuous equivalent of the discrete
Tii = (=i,li)
BI(w) / 2
a= quadratic transfer function of the wave drift damping
T(w+p,w) = continuous equivalent of the discrete quadratic trans-
fer function for the'amplitude of the drift force
Tij = T(wi,w.) with 6i Ž wj
= differenceafrequency" Ž 0
SF(P) = spectral density of the drift force
F = mean value of the drift force
B = mean wave drift damping coefficient.
For the wave spectrum a Pierson-Moskowitz type spectrum with a
significant wave height H = 10 m and a mean period T = 12 s was
chosen. The computed speciral densities are shown in Lig. 8.
The low frequency viscous surge damping coefficient without cur-
rent is derived from Fig. 4, while for the current damping coef-
ficient the following formula is used:

bl c =2Fc/Vc

in which F is the mean current force in longitudinal direction

and V c is t~e current velocity.
The results of the computations are given in Table 1 on the next
page.
For the simplified approach the computation procedure has been
shown to calculate the motions of a linearly moored tanker. In an
irregular sea combined with current both the mean wave drift
force and the slowly oscillating part of the wave drift force
will increase considerably due to the current effect. As a result
the motions are larger in a current field than in the non-current
condition.

-8-
Comparing the results in the non-current condition with different
springs it can be noticed that the spectral density of the wave
drift force considerably decreases with a stiffer mooring system.
The relation between the spectral density and'the natural period
of the system is shown in Fig. 8.

Because of the increased stiffness and the decreased spectral

density the motions will decrease. The mooring forces, however,
will be increased approximately by a factor 2 compared with the
softer mooring system.

Table 1
Without With Without
current current (2 kn) current
c11 = 6.8 tf/m c 1 1 = 6.8 tf/m c 1 1 = 60 tf/m

M+a tf.s 2 /m 26,145.4 26,145.4 26,145.4

Pe rad/s 0.0161 0.0161 0.0479
T s 390 390 131
e
t s 9000 9000 9000
N - 23 23 68.7
F tf _ 85.6, -103.8 -85.6
F -10.4
c
B1 tf/m/s 29.0 25.7 29.0
BI1 tf/m/s 16.0 23.0
bll c tf/m/s - 20.2 -
b tf/m/s 45.0 45.9. 52.0
SF(Pe ) tf 2 . s 27,700.0 43,000.0 15,000.0
a m 11.92 14.71 2.75
i m -12.59 -16.79 -1.43
x lmax m -42.45 -53.63 -9.43
X lmax tf 288.7 364.7 565.8

Complete simulations
For the complete simulation a turret-chain system exposed to
irregular waves is considered. The configuration of the system is
shown in Fig. 7.
In a static sense the reaction force on the tanker moored by a
chain-turret system represents the load displacement curve, being
the restoring force. In reality, due to the low and wave fre-
quency motions of the turret, the chains are moving through the

-9-
fluid. The chain forces will not only be' influenced by the static
load, curve, but also by the forces caused by the velocity and
acceleration dependent hydrodynamic loading (the Morison equa-
tion) and the chain inertia forces along the chain length. The
result is that the static load curve can strongly deviate from
the inomentaneous mooring forces. in the time domain, due to the
low and high frequency motions of the turret, the "static load"
will be stiffened or weakened. Due to the "stiffening' and the
'.weakening" of the static load curve the low frequency
motions of
the tanker can be strongly influenced. In fact, the time average
over the content of the hysteresis of the horizontal dynamic
mooring force versus displacement is called the "low frequency
chain damping".

To take into account the time dependent "low frequency chain

damping", the complete integrated system of chain-turret moored
tanker has to be simulated in the time-domain with the correct
coupling between the high and low frequency motions.
The mooring system is modelled by means of the lumped mass meth-
od, where each of the mooring legs is represented by a large
number of discrete elements. On each nodal point the segment
tension, weight, fluid forces and bottom reaction forces will be
taken into account. Because of the direct time-domain simulation.
method the dynamic b'ehaviour of the mooring chains must be mod-
elled correctly, resulting in small time steps in the integration
method (0.05 s).
The convolution integrals of the first and second order wave
forces are taken at the momentaneous position of the tanker in
the irregular wave field resulting in the exact wave loading on
the tanker at each time step. Further, the hydrodynamic reaction
forces were obtained by the Cummins impulse response functions,
while the viscous surge damping of the tanker was taken from
Fig. 4.
The computations were applied to a 200 kDWT loaded tanker in
82.5 m water depth exposed to a PM spectrum with H =9.8 m and
T = 13.0 s. The layout of the mooring pattern cgnsists of 6
cAains, each 700 m long. The pre-tension amounts to 40.83 tf and
the chain diameter is 5.5 inches.

The results of the model tests and the computations using the
same wave train are given in a sample time history in Fig. 9.
From the results it can be concluded that a good comparison is
achieved. For more information reference is made to [6] and [7].

DYNAMIC POSITIONING

Introduction
Nowadays DP tankers may be used as an alternative for hawser type
mooring, for tandem mooring and for offshore buoy loading sys-
tems. A well-known buoy loading system is the UKOLS system at the
Statfjord field; careful experiments have been carried out for OP

-10-
off loading in tandem configuration (in full scale as well as in
model tests). It can be observed that there is a growing tendency
at new field developments to use dynamically positioned systems
instead of bow hawser moorings.
The conventional method of DP implies position as well as heading
control of the ship, an approach which requires substantial side
thrust capacity at the stern of the ship. The consequence is the
application of a twin screw arrangement with high performance
rudders. Also, the fitting of tunnel or azimuthing thrusters aft
is costly, even oh a single screw tanker, due to the presence of
engine room and shaft arrangement. The application of bow loading
with weather following DP will improve the workability and reli-
abili,ty of the' OP off loading concept, and since there is no need
of side thrust capacity at the stern, reduced capital expenditure
for the OP equipment on the export tanker follows.
The hydrodynamic investigations for a OP system are carried out
in two phases:
- A pre-study in which computer calculations and simulations are
carried out. Herein physical trends can be investigated and a
concept exploration can be made. Although DP simulations are
instructive, the same holds true as for mooring simulations:
the results are not to be used for the final design data.
- For reliable design data, relating the available power and
thruster capabilities to the positioning accuracy and limits of
"workability", it.-.is possible to carry out closed
loop OP
tests.

DP simulations
Computer programs for DP simulations basically carry out a con-
trol loop in which.the ship position is given by the hydrodynamic
mathematical model which is also used for mooring simulations,
see [.3]. The position difference with the required set-point for
dynamic positioning, and the ship horizontal velocities are
inputted to the control algorithm for thruster action, see [8].

For the control algorithm the effective forces generated by the

thrusters, propellers and rudders have to be known. The effective
forces can differ widely from the bollard pull values. This is
caused by both thruster-thruster, thruster-hull, propeller-hull,
rudder-propeller-hull interaction as well as interaction with the
current. The hull and current types of interaction are to be
attributed to the changes in the pressure distribution on the
hull induced by the working propellers, rudders and thrusters as
compared to the pressure distribution due to current only.

Fig. 10 gives an example of typical pressure distribution changes

for a bow thruster operating whilst the vessel is subjected to a
certain head current. The low pressure region downstream of the
jet exit side is the dominant factor and integrates to an appre-
ciable side force opposite to the thrust T, thus reducing the
effective side force of the thruster.
Fig. 11 shows the results of thruster-thruster interaction under
a flat plate for azimuthing thrusters. -In the graphs the inter-
action as function of the distance between two thrusters and as
function of azimuth angle of the forward thruster is shown.
By means of computations, see [9], and/or a set of separate model
tests the effectivity values of the thrusters, rudders and pro-
pellers can be determined. Using the proper effectivity values in
the control algorithm the performance of the ship will be im-
proved.
Usually the sipmulation programs are limited to low frequency
motions only, which is sufficient in most applications. However,
when a ship is equipped with bow tunnels, and its draft (or sea
state) is such that frequent bow tunnel emergence occurs, signif-
icant degradation of thruster effectivity will be the result as
is shown in Fig. 12. Try-out of computer simulations in an irreg-
ular sea with this type of degradation is carried out. The re-
sults are shown in Fig. 13.
on basis of computer simulations it is possible to select an
optimum concept for the DP in terms of:
- thruster arrangement and type;'
- location and power of thrust units;
- mode of DP: full pos -ition and ~heading control or weather DP;
- safety and redundancy- evaluation.

By means of the computations for different DP concepts the limit-

ing sea state can be determined in which -with the selected
thruster configuration- the tanker can stay sufficiently accurate
in position.
Physically, the limit for DP' is the result of a combination of
effects as is shown in Fig. 14. In the figure relating position-
ing accuracy and sea state, four different regions can be dis-
cerned. In the first and second region the positioning is accu-
rate. In the third region the environmental loads on the tanker
are so high that the thrust requirements exceed the available
capacity for relatively long periods of time. In this situation
the positioning accuracy deteriorates although the positioning is
still stable and feasible. In the fourth region, the positioning
error increases rapidly. The maximum excursions are very large
and the risk exists that the position gets lost, either due a
high wave group or due to the impossibility to keep station.
Furthermore, specific events, such as the ship response to ref-
erence failure, sudden wind changes and thruster failure or ap-
plication of wave feed forward can be investigated. By means of
this information a model test program can be optimized.

Closed loop model tests

The main differences with computer simulations are:
- Environmental conditions are applied to maximum accuracy in
laboratory conditions.

-12-
- Low frequency as well as high frequency effects are fully rep-
resented.
- Noise in the form of high frequency motion components are pres-
ent in the position error, which makes it necessary to apply
Kalman filtering before the positioning error can be inputted
into the control system. An example of a control loop is given
in Fig. 15.
- Accurate measurements can be made and the settings of the DP
control system can be optimized in a realistic environment to
achieve the most economic positioning.
Hence, the design information that is obtained from closed loop
model tests is for final power selection of thrusters and opti-
mized control coefficients. As an example the positioning of an
export tanker during model tests in a 4 m sea state is shown for
two concepts of tandem positioning in Fig. 16:
A. Fully stern positioning.
B. Stern following, but neglecting the surge motion of the stor-
age tanker.
Fig. 17 shows that concept A requires about twice as much power
consumption at the main_ propeller (in mean value and standard
deviation) as concept B, while the positioning was similarly
accurate in terms of separation distance between tankers.

FUTURE DEVELOPMENTS
So far a review of the state-of-the-art approaches for the design
of moored and DP controlled tankers is given. The approaches con-
cern both the semi- theoretical empirical mathematical models and
physical model testing.
Further improvements in the semi-theoretical empirical mathemati-
cal models have to be carried out. Some of the developments are
given below:
- wave drift-current interaction for arbitrary directions of
waves and current and water depth;
- test duration for reliable statistical results;
- bow tunnel efficiency in waves;
- the application of adaptive control systems for optimum DP
control routines;
- improvements of wave feed forward techniques;
- the hydrodynamic interaction between tandem moored tankers of
which the initial work has been carried out as described in
[10].

REFERENCES
1. Feikema, G.J. and Wichers, J.E.W.: "The effect of wind spec-
tra on the low frequency motions of a moored tanker in sur-
vival condition", OTC paper No. 6605, 1991.
2. Pinkster, J.A.: "Low frequency second order wave exciting
forces on floating structures", Doctor's thesis, University
of Technology Delft. 1980.

-13-
3. Wichers, J.E.W.: "A simulation model for a single point
moored tanker", Doctor's thesis, University of Technology
Delft, 1988.
4. Wichers, J.E.W. and Huijsmans, R.H.M.: "Wave drift current
interaction on a tanker in oblique waves", Proceedings of
MARIN Workshop Hydrodynamics: Computations, Model Tests and
Reality, MARIN, The Netherlands, May 1992.
5. Aalbers, A.B.: "Mooring and offloading: model testing for
design", Third Offshore Symposium on FPSO Technology, SNAME
Texas Section, Houston, February 1993.
6. Dercksen, A., Huijsmans, R.H.M. and Wichers, J.E.W.: "An
improved method for calculating the contribution of hydro-
dynamic chain damping on low frequency vessel motions", OTC
paper No. 6967, 1992.
7. Kat, J.O. de and Dercksen, A.: "The influence of spectral
wave characteristics on the dynamics of a turret mooring
system", OMAE'93, Edinburgh, 1993.
8. Aalbers, A.B. and Nienhuis, U.: "Dynamic positioning of tank-
ers: Design aspects and recent developments", Symposium
Floating Production Systems, London, 1989.
9. Nienhuis, U: "Analysis of thruster effectivity for dynamic
positioning and low speed manoeuvring" Doctor's thesis, Uni-
versity of Technology Delft, 1992.
10. Feikema, G.J., Huijsmans, R.H.M. and Aalbers, A.B.: "Low
frequency motions of tankers on tandem offloading condition",
OTC paper No. 6945, 1992.

175

TM1Jif1Nori1T)
-TrIrrni

0.00T) or 111
AI?
to - - "'

Sso

(nd/i)l
"FrIqtrtnc
9I"Sftt ran"
(V
= 309 m~0;00zi).
0.001 D.D1 0.1 1.0
Fr~vqtncy (rbd/1)
Fig. 1 -Spectral densities following the for-mulation of Harris-
DnV, Davenport, Ochi-Shin,
-04 Wills and API wind spectra
(vw = 30.9 m/s; 10 m).

-14-
 -15 7.5

Wave or'ht force Wave drift damping

* Ca * 2.45m
o 'sC-ted . 1.05,m Decay test (ref. 11))
*,s Computed1.0 t
C Oflvuted

-30 5.0

( t ~ m
• )
(tf.s .m -3)

-5 2.5

0 ./

o 0.5 1.0 0 0.5 1.0

S(rad/s) (a/s

Fig. 2 - Qiadrattic transfer function of wave drift force and wave

drift damping for an LNG tanker sailing in head waves
(earth-bound<•wave frequencies).
-15 7.5

Wave drift force

o U * 0 r/s comTputed
a ,/l$
U 1.ý
derived jEstimated wave drift damping

from
model tests

-tO S.0

"2 3
(tf.m ) (tf.s.m& )

o 0.5 1.0 0 0.5 1.0

w (ra./s) a (rad/S)
Fig. 3 - Quadratic transfer function of wave drift force and wave
drift damping in bow quartering waves (earth-bound wave
frequencies).

-15-
 Tanker size:
a 200 kDWr (100% loided)
- 250 kDrT (100: loaded)
55 kDVT ( 80% loaded) 3
o Symn trical V - 213,717 w
LNG carrier- 3
x 1250,00 m Ref. (2-21

50

40 - - -- a-- -

= 20

10

0
0 1 2 3 4 5-105
3 2 3 -I
. -S / in M .s

Fig. 4 - Measured viscous damping coefficient for the surge mode

of motion as function of wetted hull area and surge
frequency.

---..-.--
-. -- FULLY LOADED
BALLASTED TANKER
TANKER

SWAY MODE OF MOTION

KC =4 KC -4(,9

(1.333) B62 .
B2.0
B2 2 6

1.0 -.-- - 2 -' "-."- .-

(1.01) (2.28)

0 0
0 0.025 0.050 0.075 0 0.025 0.050 0.075

YAW MODE OF MOTION

(4.09)
2.0 4
B 0(.333)B
B66 jj• 26

1.0 2 -
0(1.0) (2.28)

0 0.025 0.050 0.075 0 0.025 0.050 0.075

Fig. 5 - Low frequency viscous damping coefficients for the sway

and yaw mode of motion in still water.

-16-
 200 kTDW tanker - 82.5 m water depth - V = M.3-
C

& - IA6a 0.002 rad.s- l

I6a I
--- = 0.003 rad.s- 1 HModel
-
1A I
I6a = 0.005 rad.s I tests
-I
1i6a I 0.007 rad.s
100 .....- Fourier approximation

50 LOADED 20 BALLASTED

250
50 -0
• ~~2
5 - "

-250
40,000.-50

-100
30,000 , 0

20,000 10,000

0180 360 0 180 360

Ocr in deg
ýcr in deg

Fig. 6 - Dynamic current contributions in surge, sway

and yaw
direction due to motion in yaw direction in 2 knots
rent, based on negative yaw velocity. cur-

-17-
Fig. 7 - Turret moored tanker.'

V = 2 kn
Vc = 0 kn

4
6*10

4*104
x
N .
in tf 25
2*10 4

0 0.02 0.04
v in rad/s

Fig. 8 Computed spectral densities of the wave drift forces.

-18-
 SarnpS tim NlofilnCof masred adWrrPfd QWanttt - -.-- --- tnnankd
Btetscr~iw wct,
&ctm Ha-g.e m. T-1-3.O0 s ccrp.*ed

X TURRET
(MJ

Z TURRET
IMJ

FX TURRET t

F CHAIN 4 20f

0 1- 240

SECONDS

Fig. 9 - Sample time histories of measured and computed quanti-

ties.

-19-
 V

Fig. 10 - Bow thruster-current interaction.

1.0 *D..

T/To -------

0.4 0.

- T/T Coklotu aed 2

0.2 * T/T o measured 02 -- *
0- /Q 0 calculated • a--,
- - 0/0 OI~jI~~d---- p. X/D *tO0
0 0/0, meosure.d
0 o0
0 4 a 12 16 20 0 20
20 30

X/D

Fig. 11 Thruster-thruster interaction under a flat plate as:

a) a function of distance between the two thrusters;
b) a function of azimuth angle of the forward thruster.

-20-
 Regular waves: amplitude ±1.0 m
Wave frequency 0.78 rad/s
Wave frequency 1.04 rad/s

Regular waves: amplitude ±1.5 m

L=

Relative motion

Fig. 12 - Thrust versus relative

wave height at bow tunnel.

-21-
 X-COG 5. Without With ventilation
0V
Cm) -
-5
5
Y-COG 0
-m)
5
YAW _ _ _ _
(deg)

THRUST 750
MAIN PROP. 0
(kN) -750
THRUST 100
BOW THR.1 0
(kN) -100
THRUST 100
BOW THR.2 0 -
(kN) -100
THRUST 100
STERN THR.1 0
(k.N) -100
THRUST 00
1
STERN THR.2 0
(kN) 10

0 100 200 300 0 100 200 300

(s) (s)
Fig. 13 - DP simulations showing the incident thrust loss after
bow tunnel ventilation.

Position
os n Toouh
thruster /
Large
error Increasing saturation excursions
Easy thrusteafter high
posi- saturation wave groups

Sea state 1- Hs2

Fig. 14 - Limiting sea states for DP.

Ship ] Motions > n

(model)filter controller
I L__._> K&,man
---------
Wind-r---> prediction
Current I model xy Np
or waves I
Thruster ttfeed forward)-> allocation
action <

Fig. 15 - Example of control loop.

-22-
 So

80,

Concept B

40-1

C~Cocept A

Fig. 16 - Tandem offloading for different DP concepts.

EFCFT OF COWmOL IED00 FOR TANDEM DP OFF4OADm

I
----- - ETHOD^A
----- o--- METODB

DP EWCOF(T TANKCER: FR.L LOAD

3000 1500-

02000 1000-

11000L 50 "
o -/ 0i"

0C
0-

A B C A S
8 C
Sea condition Sea condition

Fig. 17 - Main propeller power consumption.

-23-