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:
Gusto Engineering by, Holland Carlo van dee Stoep, Senior Eng. Spec.
0. Faltinsen
Department of Marine Hydrodynamics
Norwegian Institute of Technology
N7034 Trondheim, Norway
INTRODUCTION
We will discuss the state of the art of computer aided ship motion predictions both
for conventional and highspeed vessels. For conventional ships this includes
linear theories like strip theories, unified theory and complete threedimensional
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.
Generally speaking strip theories are still the most successfuil theories for wave
induced motions of ships at moderate forward 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
abovewater 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 forward 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 ships 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 STFmethod (Salvesen et al. (1970)).
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 3D 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 zerospeed 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 3D 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 currentwave effect
there are practical 3D 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 3D methods based on the classical
linear unsteady free surface condition with forward speed. This free surface
condition is appropriate in the farfield of a ship and accounts properly for all
frequency and Froude number effects. However, one can question if this is the
correct freesurface condition to use in the nearfield of a ship irrespective of
Froude number and bluntness of the hull. Inglis & Price (1981) have developed a
linear 3D method based on the classical linear unsteady freesurface 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 lowfrequency range.
Nakos & Sclavounos (1990) presented a linear threedimensional 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 3D 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.
HIGHSPEED THEORY
Details about the theoretical and numerical method used to analyze the steady
and linear unsteady flow about highspeed nonplaning 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 crosssection is
found by a twodimensional 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 frequencydependent 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 highspeed ships.
Roughly speaking the highspeed 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 highspeed monohull. The agreement is reasonably good.
Fig. 3 shows computed results of unsteady wave generated by a highspeed ship.
The highspeed theory is compared with a thin ship theory. A strut with parabolic
waterplane area and wedgeshaped crosssectional area in unsteady heave motion
with forward 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
threedimensional sources over the center plane of the ship. The sources satisfy
the classical free surface condition with forward speed. The agreement with the
highspeed 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 highspeed 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 highspeed 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 nonlinear
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 nonlinearities. 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 nonlinear wave loads. Important
parts
of the methods are exact calculations of FroudeKriloff and hydrostatic pressure
forces on the wetted hull surface. The incident wave field is described by
linear
theory. The nonlinear 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 nonlinear
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.
()
2pC A,
Fv, = _L) dT31 drj
c3
i
(1T
where p = mass density of the water, Aw = waterplane area, dly3/dt = heave
velocity. By equivalent linearization it follows that
The drag coefficient depends on the geometrical form, the free surface, the
Reynolds number and the KeuleganCarpenter number KC = m3a/D. Here D is the
draught. The free surface effect is a function of ceVLuj and the Froude number.
For crosssections like a rectangular crosssection 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 KeuleganCarpenter 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 KCnumber
dependence. The Reynolds number dependence is not known. Fig. 7 shows also
numerical value of CD obtained by the twodimensional 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
crosssection was used in the calculations and the Froude number was zero. The
twodimensional vertical drag force was nondimensionalized by the beam when
the drag coefficient was calculated. In practice threedimensional end effects
should have been accounted for. This will result in lower CDvalues. 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 CDvalues
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 KCnumber 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 KCnumbers. 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 vortexin cell
method or a NavierStokes solver.
Fig. 12 shows the heave damping coefficient for different amplitudes of oscillations
for the midship crosssection 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 CDvalues presented in Fig. 7. However it should
be noted that the numerical results presented in Fig. 9 show KCnumber
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 twodimensional body analysed in Fig. 9 the nondimensionalized
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)
For ships with crosssections 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
Zhao & Faltinsen (1993) have presented a boundary element method applicable
for water entry of a broad class of twodimensional bodies and relative velocity and
orientation between th&'bbdy and the water. They have been able to satisfy the
exact nonlinear 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 twodimensional rigid body of
arbitrary crosssection. 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 (<  230) 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 jetflow region Wagner's (1932)
solution was used. A simple composite solution for the pressure distribution was
presented. The body has to be twodimensional 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'
One should be careful in applying results for wedges to other crosssections. 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.
1. List of errors .1
2. Sensitivity of final results to each error source
3. Combination of errors in final results
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 crosssection. Lewisform technique is also sometimes used. This
is a conformal mapping technique that assumes the twodimensional
hydrodynamic properties of a crosssection is adequately described by the cross
sectional beam, draught and sectional area. This approximate representation of
the crosssectional 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 crosssection (or strip). This will result in errors
in added mass and damping coefficients, volume calculations, hydrostatic restoring
coefficients, and FroudeKriloff 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
 IfNLI
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 anvalues 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
Gjcvj
(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 nondimensionalized mean wave period . We have presented
calculations of En3 and an5 in Table 1. A modified PiersonMoskowitzan 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.
Physical errors
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)).
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 nonlinear potential flow effects by analyzing experimental results
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)
O'Dea & Walden (1985) studied experimentally the nonlinear 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 nonlinear
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 nonlinearities. 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
nonlinear potential flow effects is then 0.08.
O'Dea & Troesch (1986) studied the nonlinear behaviour of the S7175 hull in
regular head sea waves at Fn = 0.2 and oM(L/g) y 2 = 2.4. The nonlinear 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.
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 differencesin 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 nonlinear effects. We then have a problem in how to use the results
for regular waves in irregular 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
nonlinear 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 forward 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 highspeed 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 twodimensional 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.
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, beamdraught 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).
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).
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
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.
"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 twodimensional
ship
shaped bodies, Proceedings, PRADS'89, Varna, Bulgaria, Vol. 2, pp. 751 to 757.
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. 343353.
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. 125150.
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 surfacepiercing
plate oscillating in yaw and sway, Proc.1lst. Int. Symp. Numer. Hydrodyn., pp.
333350,
David W. Taylor Naval Ship Center, Bethesda, Maryland.
Dobrovol'skaya, Z.N., 1969, On some problems of fluid with a free surface, J. Fluid
Mech.,
Vol. 36, part 4, pp. 805829.
Faltinsen, 0., Pettersen. B., 1987, Application of a vortex tracking method to separated
flow
around marine structures, Journal of Fluids and Structures, 1, 217237.
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 highspeed 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. 334, Arlington, Va.;Office of Naval Research 
Department of the Navy.
Hoff, J.R., 1990, Threedimensional Green function of a vessel with forward speed in waves,
Dr.ing.Thesis 199025, 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. 817P, Ship Hydrodynamics Laboratory Delft
University of Technology.
Nakos, D.E., Sclavounos, P.D., 1990, Ship motions by a threedimensional 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, 22183.
Newman, J.N., Sclavounos, P., 1980, The unified theory of ship motions, In Proc. Thirteenth
Symp. on Naval Hydrodynamics, ed. T. Inui, pp. 37398, 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 nonlinearities on the prediction
of large amplitude motions and deck wetness, In Proc. Fifteenth Symp. on Naval
Hydrodynamics, pp. 16376, 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 : 25087.
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 : 23949.
Wagner, H., 1932, Uber stoss und gleitvorginge an der oberfldche von flossigkeiten, Zeitschr.
f. angew. Math un Mech., Band 12, Heft 4, pp. 194235.
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. 124136.
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 twodimensional bodies, J. Fluid Mech., 246, pp.
593612.
.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 highspeed theory by Faltinsen & Zhao ( 19 91a).
5 BODY PLAN
* 2 I/2DExperiments
Approach
3.0
Fig. 2 Wave resistance for a highspeed hull. Comparison between the highspeed 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 highspeed 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
2,50 m
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 crosssections (Beukelman
(1980, 1983)).
Fn = 0, r3a/D= 1/15
6, EXPERIMENTS
(BEUKELMAN (1983)) Fn = 0.16
3.00 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 crosssections shown in Fig. 6. The data are presented as a function
of nondimensionalizedftfireqiuency 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 twodimensional crosssection 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
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
Similarity solution
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
 
Fig. 12 Predictions of pressure (p) distribution during water entry of a 2D 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
Is))
w) 0
~~~a 0.01
V (ms) .01 ec.TIME
4.0 ec
3.5
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 nonbreaking 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.
4.WAVECURRENT INTERACTION.
Recent research has shown that wavecurrent 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. Helre we have a
topographic "ornor 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.
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
2225 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. JulyDecember. The Johns Hopkins University. U.S.A.
KJELDSIN 1982:" 2 and 3dimensional Deterministic Freak
Waves"
in Proc. 18th International Confe'mrence on Coata] Engineering.
Cape Town, South Africa.
KJEIDSEN 19S3:"Dete•r:iination of Severe Wave Conditions for
Ocean Systems in a 3d imens;ioanal ! regul.Ir Seaway." in Pro;.
1ram the 8th Conqrezs of can Inst iut:e of
('anAmei
1jval I
ng.i necŽi :.incj. 1217 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 FaerderTvistein
yt
P4
DEFINITION OF CREST FRONT STEEPNESS rx IN
SYNOPTIC DOMAIN: B
a for p.Position
Inception of breaking
E:X.s " ' MW,• . ..
TIME MWL
2t T.Tzd
Tzd
MWL
HH
U V
10  .
O
0u C4JN
ud
001
* ~Un
4W0~
0%44
C) 4
N W4J0
0) I
'4 WCHr4
U) Ojo)
C) jo
o r46J~d'f
E4
tp u
a)00
CmY C)
DISTANCE FROM
WAVE GENERATOR: TEST NO: 214 B R2
S.L. BUNN B.S Comm 801
O.
X = 10 M ._. Z
0
I. 20. 10. 60. 80. too.
rIErCSEC.3
r 2 5 sm.r 0.3
.
I eill\ ,
0.)
0.4
0. ,00. * . go.
8O 0.
TllCESEC.j3
0.2I
x = 26rn .
""llf]
x = 35 m.
.
0.0.I_____
"&
I ________________.]_
x 60 TO. 0.4 _
0.2
3 1
scm Rise
z 2 9time
milles•.
Lewi: IHpz]K±0.L60
0= 45 degrees
CLSPILLING BREAKER
U '00
1CX)
E)=45* Tzd 
FP',_pg I Hp  zO) 1.00 sec
W p.gHp A 1.19 sec
0 1.35 sec
6 '::served
12.0 _:ecks: 62
shock
01 Uý
of
30 t40 5.0
IFFPINESS RiATIO
Ci
kg
.2  •.
0.0
.4 . .0 kx .2 .
L  if.
   i
AUM ~~ ~ ~ =)f. i ~ igkr
L~za. 19.bt.
if ý W m
1 br. HIesnipt
. R.. .
t. ~VVk~ig
 r
... _ F1..
T. Nedrelid
MARINTEK. 9 February 1993
LECTURE 1
* Dynamicsituations.
STABILITY
Static Stability
WI,
TO AVOID HEEL
TO AVOID CAPSIZING
* 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
lies above G.
€.S
COMPLETE STATIC STABILITY
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.
la
* Lifting of cargo
C#$&40 fl26uwN4UMoM~kjy
NJJ a a
LL tn _ _
< 0 w
Ow 0
> 0 0 I
: 03
Z

QZ
4< 0<<n
M
FU)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
Over Loaded 9 8 17
TopHeavy 13 11 24
Insufficient Lashing 1 20 21
Inferior Loading 7 17 24
Cargo Shift 13 41 54
Open Door 9 22 31
B ro;ach inq 1 3 4
TOTAl, 52 87 139
THE PRINCIPLES OF STABILITY REGULATIONS
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 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:
00 1Oo
STABLE
a. Selfrighting.
Capsize not feasible
STABLE STABLE
b. Capsize to 180 /
~ MARINTEK
EXPERIENCE / ACCIDENTS IN NORWEGIAN WATERS
do cargo shift ?
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"
MAIN SUBJECTS:
§ ENVIRONMENTAL DESCRIPTION
§ STABILITY CRITERIA
N~MAINTEI
PROGRAM "STABILITY CRITERIA"
RESEARCH
 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.
Project Manager for several large Norwegian Research Programs: Ships in Rough Seas
(197981), Stability Criteria (1981), Stability and Safety of Small Vessels in Extreme
Weather Conditions (198182), Offshore Marine Operations (198588), FPS2000 
Floating Productions Systems, Positioning and Mooring (198991).
Since 1986 Division Manager at Marintek, Div. of Offshore Floating Structures and as
of 1991, Division Manager of the Marintek Laboratories as well.
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 GZcurves 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. Selfrighting, 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
The stability standards based upon this However, this situation is improving and we
statical approach and terminology are the feel that future stability standards and
foundation for most stability work all over regulations will be reflecting this new
the world. The IMO regulations represent the terminology. We should be aware that there
standard rules. is an ongoing technology development in
related industries as ie. offshore, using
reliability calculations in the design analyses
PROBABILISTIC APPROACH TO of new structures and structure
REGULATIONS elements/systems. This technology is quite
parallel to what we define as the problem of
We have stated that the external forces are stability.
of a very dynamic, statistical and often very
nonlinear nature. The event of a capsizing
itself may be referred to as a statistical FUTURE STABILITY REGULATIONS
event. Since we have today the improved, BASED UPON OPERATIONAL
effective theoretical models to calculate STABILITY
dynamic motions, we might think that we in
the future can calculate the final probability Through the probabilistic terms we have
of a capsizing event to occur. By using this been searching for the absolute stability that
type of advanced technology and will ensure safety against capsizing. The
crres[pond ing termninology, aidiscussion has opcration of a vessv will, however, always
started if future stability standaids should he address the final and practical stability
based on such risk analysis. So farIth situationl. here are often several aspects that
analysis have been used ill piCanalyses of' define a nonintact vessel when analysing a
actual accidents. However, they have not as capsizing. Nearly 90)% of1 all accidents
yet been used in any new design situ;aitiu. rcitetl to stability. happelns tlic to technical
Parameters such as: Wave kinematics, wave SURVIVAL TESTS INBREAKING WAVES 510 m
heights, static stability of the vessel and ROLLING ANGLE (deg)
The results are presented in Ref. /1/. The Fig. 5. Survival testing in breaking waves.
general conclusions are illustrated in Fig. 4
showing that any vessel can capsize in High speed crafts
waves (turn around in waves) if the dynamic
stability, ie the stability resistance work, is In recent years the Norwegian focus has
low. been put on capsizing tests of high speed
crafts, Ref. /2/ and /3/.
New designs
Fig. 6. Profile and body plan of the
A recent investigation has been performed at catamaran.
Marintek (1991  1992) concerning a new
type of fishing vessel, see Fig. 6. A
catamaran design has been exposed through General comments
a survival testing program. Critical
wave/vessel situation have been investigated. Our tests showed that vessels havingat least
Special emphasis were put on studies of the the minimum required stability very seldom
vessel in breaking waves. The results are capsize in normal wave situations. In
discussed in Ref. /7l, here the general breaking wave situations, most smaller
conclusions: vessels can capsize, even those with very
high stability (small, self righting vessels can
"tile results are very encouraging and be turned around 3600). When large angles
display that this particular catamaran has of roll occur, only those with positive
a significantly larger critical wave height righting level arm at large angles of heel,
compared to a conventional monohull will righten. The Norwegian Maritime
vessel with the same dynamic stability. Directorate has put an adjustment of the
Compared to nin nolull vessels ol the sarne standard I MO regulation into force through
Ienth the results indicate that this the national regulalions for fishing vessels,
cataxmarian has a subsLaiasially higher saying that the righting level arm shall be
safety against capsizing inbreaking waves positive at large angles of heel (40  60
fro11 tile side." degrces).
q•:\s;kSi\l);H).:r\uz')*, l ,hl\29l
J;iaI Iagl IV'199 7
For some types of vessels  offshore supply and standards should reflect this fact. One
vessels and other types of low freeboard should note possible human maloperations
vessels  water on deck might be a critical and find out if they are followed by low
parameter. Water trapped on deck can nonintact critical stability situations.
change the initial stability and may start an Examples are:
inclination that leads to capsizing. Survival
testing will show how sensitive the vessel  water through open hatches
structure is to the trapping of water. A  critical manoeuvres in waves
closed bulwark, open rooms in the back or  maloperated ballast system
poop area should be avoided.  wrong cargo handling
 wrong handling of fishing gear
we
By performing effective survival testing
can map some of these design problems and Survival testing by use of advanced physical
still rely on the statical stability standards, models can give answers to questions such
as these, even if they are complex and
transient.
I. Survival testing gives the answers to
theoretically complex situations. The We might think that a vessel, even if it
results can be interpreted the empirical fulfills the standard requirements, should not
way by mapping the effect of possible be approved if the design implies a very
critical parameters. sensitive operation. We can require a design
adjustment or we can increase the standard
II. We can use this technique to search static stability requirements to make it less
for very simple standards. sensitiv stabilitywise.
III. We can qualify such standards or To increase* the total safety by improving
find the sensitivity of the standard by vessel stability in future regulations, one
using this technique. should make requirements that imply
building a kind of minimum stability
barrier for nonintact conditions.
STABILITY AND MALOPERATION
l .0 CONCLUSIONS
~l\lp:
q•:,:,,'v5 ,: I ,,rhI\2'l Jat'I
II\tistpLl ."y I%'''•9 I)
UN! VERSITA DEGLI STUD! DI TRIESTE
t ea ryrs
,e LA &s ea
+týQe+L~crc5~et~
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 broachingto 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 energybalance 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 2030 percent of net area
positive according to energybalance 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/broachingto.
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 197377
[1 ] are analysed with respect to the causes.
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 broachingto is examined through an analysis of the stability of
antisymmetric motions.
m(v÷Ur) = mgo+Yvv+Yvv+Ypp+YOO+Yrr+Yrr
lxPlxzr = Kvv+Kvv+Kpp+KpppgVGM@+Krr+Krr
lzrlxzp = 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
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 112 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
rolldominated 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
broachingto are not so clearly connected [22]. Moreover, sometimes the
instability appears connected with an inherently poor coursekeeping ability
[23]. The results are not given in terms of KGmaxID due to the uncertainty in
the evaluation of the hydrodynamc coefficients.
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 quasistatic
approach.
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;
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 twodimensional Weibull distribution
[34].
2(p
Idd d(p
• + D(9, "•) + A• GZ((p) = F(t)
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. [3537].
In the region of synchronism w coo:
=
ý1
=eq + 0eq 02 6eq = ( 8 i +3(00282).
2
(20))2[(g.+ 146eqOm 2 ) 2 (1+y/ 9eq)] = ot2
; 0)f2_(0O 2  3 0
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
quasilinear 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 nonGaussian behaviour [4245].
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. 243253.
[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, 145155.
[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. 11731186.
[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.6381.
[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. 121132.
[14] Cardo, A., Francescutto, A., Nabergoj, R., "Subharmonic Oscillations in
Nonlinear Rolling", Ocean Engng., Vol.11, 1984, pp.663669.
[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. 1331.
[16] Dunwoody, A. B., "Roll of a Ship in Astern Seas  Response to GM
Fluctuation", J.Ship Res., Vol. 33, 1989, 284290.
[17] Bishop, R.E.D., Neves, M. de A.S., Price, W.G., "On the Dynamics of Ship
Stability", Trans. RINA, 124, 1982, 285302.
[18] Bishop, R.E.D., Price, W.G., Temarel, P., "On the Dangers of Trim by the
Bow", TRINA. Vol. 131, 1989, pp. 281303.
[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, 114.
[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 Broachingto Phenomena",
Naval Architecture and Ocean Engineering (Japan), 20, 1982, 92107.
[23] Baoan Y., "Ein Beitrag zur Beurteilung der Stabilitat schneller Schiffe bei
gekoppelter Gier, Quer und Rollbewegung", Schiffstechnik, Vol. 31, 1984, pp.
2242.
[24] Francescutto, A., "Is it Really Impossible to Design Safe Ships", Spring
Meetings of The Royal Institution of Naval Architects, London, 2729 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. 3.3.13.3.14.
[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. 597603.
[27] Cardo, A., Francescutto, A., Nabergoj, R., "Ultraharmonics and
Subharmonics in the Rolling Motion of a Ship: SteadyState Solution",
International Shipbuilding Progress, Vol. 28, 1981, pp. 234251.
[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. 119141.
[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. 210214.
[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.
191199.
[31] Falzarano, J. M., Troesch, A. W., "Application of Modern Geometric
Methods for Dynamical Systems to Problem of Vessel Capsizing with Water
ondeck", Proceedings 4th International Conference on Stability of Ships and
Ocean Vehicles  STAB'90, Napoli, September 1990, pp. 565572.
[32] Francescutto, A., Cardo, A., Contento, G., "On the Representation of Sea
Spectra Through Linear Filters" (in italian), Tecnica Italiana, Vol. 56, 1991, pp.
110.
[33) Guedes Soares, C., Trov~o, M. F. S., "Influence of Wave Climate
Modelling on the Longterm 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 NarrowBand Random Excitation", Journal of Sound and
Vibration, Vol. 123, 1988, pp. 497506.
[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. 195208.
[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. 171179.
[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. 9196.
[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. 4966.
[43] Juncher Jensen, J. (1989). "On Fatigue Damage due to NonGaussian
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. 279294.
[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. 554558.
[46] Hutchison, B. L., "Cargo Mechanics (Application of Seakeeping
Revisited)", Marine Technoloqy, Vol. 23, 1986, pp. 230241.
[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. 33.133.9.
Table 1. Main causes of loss for capsize and broaching regarding
Japanese fishing vessels in the period 197377
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.
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.
i 2
~~\T
STATISTICAL
 WEATHER
0.9  STRATHCLYDE 70%
08
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 iI? 14
vl 4 4
4
10 
0 GILFý0.1
10 au\
11.0
5
XE
4 K;MIF/LE=o.1
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
akepn
Se. Ine
01 2 3 4 51 6 in 11 12
UFIRIRNT No
8 .0 1I I
lo
Cd,4
000.
0.5
05
0.0 ro ~f/.1
0.5
o./
0.. 0.5 XM to
pdf
(a.u)
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
IVFkerpc&.[2w LeeL2
o
LsQ'~4~go..:L2 7e 6]
5
C5
3
10 a
2 63
1 0
0Y
9t(AIAb1W LŽ T qt~ tett&
WEGEMT Short Course on:
"New Techniques for assessing and Quantifying Vessel Stability and
Seakeeping Qualities"
Trondheim, 811 March 1993
ABSTRACT
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/broachingto.
( 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;
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 broachingto is examined through an analysis of the stability of
antisymmetric motions.
m(v+Ur) = mg4+Yvv+Yvv+Ypp+y•+Yrr+Yr r
lxplxzr = Kvv+Kvv+Kpp+KpppgVGMp+Krr+Krr
Izrlxzp = 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
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 quasistatic
approach.
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 yý0; 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 twodimensional Weibull distribution
[34].
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. [3537].
In the region of synchronism o _ wo:
1.
eq = ± T 2
eqeq 6 q = (1 +30Oo262).
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].
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
quasilinear 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 nonGaussian behaviour [42451.
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. 243253.
[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, 145155.
[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. 11731186.
[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.6381.
[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. 121132.
[14] Cardo, A., Francescutto, A., Nabergoj, R., "Subharmonic Oscillations in
Nonlinear Rolling", Ocean Engng., Vol.11, 1984, pp.663669.
[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. 1331.
[16] Dunwoody, A. B., "Roll of a Ship in Astern Seas  Response to GM
Fluctuation", J. Ship Res., Vol. 33, 1989, 284290.
[17] Bishop, R.E.D., Neves, M. de A.S., Price. W.G., "On the Dynamics of Ship
Stability", Trans. RINA, 124, 1982, 285302.
[18] Bishop, R.E.D., Price, W.G., Temarel, P., "On the Dangers of Trim by the
Bow", TRINA, Vol. 131, 1989, pp. 281303.
[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, 114.
[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 Broachingto Phenomena", "A
Naval Architecture and Ocean Engineering (Japan), 20, 1982, 92107.
[23] Baoan Y., "Ein Beitrag zur Beurteilung der Stabilit.t schneller Schiffe
bei
gekoppelter Gier, Quer und Rollbewegung", Schiffstechnik, Vol. 31, 1984,
pp.
2242.
[24] Francescutto, A., "Is it Really Impossible to Design Safe Ships", Spring
Meetings of The Royal Institution of Naval Architects, London, 2729 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. 3.3.13.3.14.
[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. 597603.
[27] Cardo, A., Francescutto, A., Nabergoj, R., "Ultraharmonics and
Subharmonics in the Rolling Motion of a Ship: SteadyState Solution",
International Shipbuilding Progress, Vol. 28, 1981, pp. 234251.
[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. 119141.
[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. 210214.
[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.
191199.
[31] Falzarano, J. M., Troesch, A. W., "Application of Modern Geometric
Methods for Dynamical Systems to Problem of Vessel Capsizing with Water
ondeck", Proceedings 4th International Conference on Stability of Ships
and
Ocean Vehicles  STAB'90, Napoli, September 1990, pp. 565572.
[32] Francescutto, A., Cardo, A., Contento, G., "On the Representation of Sea
Spectra Through Linear Filters" (in italian), Tecnica Italiana, Vol. 56, 1991,
pp.
110.
[33] Guedes Soares, C., Trov~o, M. F. S., "Influence of Wave Climate
Modelling on the Longterm 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 NarrowBand Random Excitation", Journal of Sound anda
Vibration, Vol. 123, 1988, pp. 497506.
(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. 195208.
[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. 171179.
[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. 9196.
(• [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. 4966.
(43] Juncher Jensen, J. (1989). "On Fatigue Damage due to NonGaussian
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. 279294.
[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. 554558.
[46] Hutchison, B. L., "Cargo Mechanics (Application of Seakeeping
Revisited)", Marine Technoloqy. Vol. 23, 1986, pp. 230241.
[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. 33.133.9.
Table 1. Main causes of loss for capsize and broaching regarding
Japanese fishing vessels in the period 197377
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...
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 0656 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.
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. 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.
. 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
0.9
0.8Ivc
50%
1 XG
07100% X
0.6
?~STATISTICAL
WEATHER61
STRADTICLVDB
4 c 8 t0 T, . 14
'F 4 4
05(1k, XF
J/L F,?= 0. 1
1r 1.0
0.50) 3
5
XF
2 = 0.1
ZKM/LF
"0
10*5
101.L3
1. 0 10  .
12
10
6
00
4 o
00
2 4 6 8 10 12
Seakeoping Index
6
4
(0 4
0 2S
0 RKN IOH
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.
0.. 0.5 xm t
(a.u.)
1.0
0.8
S0.6
0.4
0.2
00.0
1.0
0.8
06
04
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 STABILITYFuture Rules and Regulations
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 highspeed 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.
As about 45 000 relatively small vessels out of a total of approx. 80 000 are in existence,
the shortterm 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;
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, riskreducing 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
classflcationsocieties 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.
Stabilityrelated 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 stabilityrelated 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 stabilityrelated 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 highspeed 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 301project, 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
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
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 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 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
The first part of a risk analysis is concerned with the cause for the ship accident. For
stabilityrelated 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
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 quasidynamic calculation of probability 6
of P(Demand> Capability)
f(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. Seirght••g.
Capsize not teasible
c. Capsize to 180"
8
The first situation shown in Figure 2a is for a ship which is selfrighting, 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 GZcurve 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 sideposition
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.
Figure 3. Number of ship accidents versuis loss of lives (1980ties, Norwegian fishing
fleet)
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.
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 highspeed 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.
As about 45 000 relatively small vessels out of a total of approx. 80 000 are in existence,
the shortterm 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
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.
139151, 1988.
Anon. Use of environmental and operational data in design for safety of smaller
vessel. Proceedings, PRADS'92, Vol.2, pp. 14301442.
Myrhaug, D., Dahle, E., and Rue, H. A twodimensional Weibull distribution and
its application to rolling. Proceedings, OMAE93 (to appear).
o .c
.....
to
Out of 30 ships capsized while under way
.2
%    
.1
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.......
ABSTRACT
1. INTRODUCTION
The compliant: struc ture used for full scale trials was a single
point mooring, (5PM) system of turretkind 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
a) Environment conditions
a) Environment conditions
2Y
0
d) Stresses on the mooring yoke
3. DATA ELABORATION
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.
us) j_sb r r si
COWP IIEADI)JC(
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.
Xk•*= . _A
=I.,..• L I ...
.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 surgestresses
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
2 (4)
C,()
C,(k) S.,,k)Sý,(k)
IS,,(k)l
U
)
SZZk, ~y>1X(k)(i)X(1Y(i)Y(k + 1)(j (5)
(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.
4. CONCLUSIONS
[8] Kim, S.B., Powers, E.J.", Miksad, R.W., Fisher, F.J., Hong,
J.Y., 'NonLinear System Coherence Analysis of the Surge
Response of TLPs Subject of NonGaussian Irregular Seas",
OMAE 89, The Hague. March 1989.
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.
F Fto 1
U bo o 11t ,
0C>
14C N
00Sr4 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
04 r o
  m
0)44
2 a 0 0dN
CIO
 0

 ) i in.•_ a)•.
0oU 0
oa  . 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
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.
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:
2)  The effects of the environment on the ship, including the wind and
the sea
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. lIvironmenial 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.
 direct and indirect effects due to wind acting on the upper works of the
(1 vessel
 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 inservice 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.
(
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:
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.
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:
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.
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.
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
swellup. 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.
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.
42 4o
4o
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
Anomaly corralolion
0.2 .1979/80
....
. .....
...... .... 1972
0 .
1 3 4 5 6 7 a 9 io
Days
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 19E' 198C 1987 1988 1989 1993 1991 1980 1981 1982 19,2 1984 2985 1986 1987 2988 2989 1990 1991
EF.V' ooraist 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
C Seakeeping
tests
charateriticsadded
__________________resistance
Velocity
____
____loss ____ ____ ____
EnginenSiAWA
propeller
Weathe datacharacteristics
(waves. wind HP, 0. RPM
and current) _______
Criteria for
voluntary
speed loss
 OPTWAY
SHIP ATTAINABLE
I'SPEED
(SEAWAY/ J
Il
WEATHER
FORECASTS
(DMI)
CURRENT
DATA
(DMI)
MARINTEK)
IN  0 P T W A Yl
R 0 U T E P LOT
PLOT
OOPTIMAL
ROUTE
SN4 Waves Wind Current
Bow
Stern
6t
. 0.
La) re t e)
irc o t
width L• oArrivallpoi
Departure pointofln
*0 .
 .
....... ..... .... **o . 
.. ........... 204@ . . . .........
0........  ....... ...... .............. ........... ... o.......... .. .. . . ......... .......... ........
...........
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
C.METER FOR
MEASUREMENT OF VERT. MOTI N
IS knoLs ISknoLs
0 him. vrtcot 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
0.S .
0.0
O is 20 2S
Pee' uo prod sIeI
_  
10.I 
0 nt 29 2S
VEOL.
00
...,.o.
NUMBER OF GREEN SEAS AND En.
RAPPORT
FXLOADS ON BREAKWATER ReDrt
CATO
HS = 10 m, T = 14 sec. Date
Sp REF.
"Ref
HEADING = 30 deg
t1000 3 <1000 3
1.0002000 6 1.0002.000 3
3.0004.000 1 2.0003.000 1
3.0004.000 3
TOTAL 10 4.0005.000 2
Encounter waves: 162 5.0006.000 1
Percentage green seas: 6.2% 7.0008.000 1
9.00010.000 1
14.00015.000 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
3 i
9C.
* .
6
LECTURE 9
SEAKEEPING IN DESIGN
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 influencingseakeeping
performance have been fixed based upon other considerations such as still water. resistance.
 Evaluation of basic parameters such as motions and accelerations and comparison "
against criteria for cargo integrity, passenger comfort etc.
 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 following sections will discuss some of the important issues relating to the application of
seakeeping theories in design.
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:
Threequarters of the Institutions use 2D strip theory or slender body theory only. Only
about onequarter use or have developed 3D computations. The 3D methods are used
mainly for zero or low forward speed.
Threequarters 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 stateofant.
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 2D 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 2D theory. This is mainly due to the fact that the stationary wave field (the bow wave)
and the dynamic swellup 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 nonlinear body boundary
conditions in the Jprediction method Such programs 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
nonlinear 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 nonlinear conditions on the body boundary. 2D methods for timedomain modelling of
these effects are standard routines today. 3D methods are under development, but are not.yet
part of routine design procedures. Differences between linear and nonlinear predictions of
vertical bending moment can be as high as 5060 percent.
2D vs. 3D theories: Complete numerical solutions based upon 3D theories with forward
speed are still in their infancy. The major driving force behind the development of 3D 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 3D methods playing an icreasing part in routine applications
*' within 23 years from now.
3. Seakeepine criteria
1. The criteria and corresponding responses must be relevant to the mission of the vessel.
2. Criteria levels must berelated 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
In practice the boundaries between habitability aspects and operability aspects of seakeeping
performance are vague and the two will always be considered together.
Design limits: Absolute limiting values which are not to be exceeded in service
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 tradeoff 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 technoeconomic 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 technoeconomic 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.
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 worldwide 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 TransAtlantic 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 1020 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 56 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.technoeconomic performance of a new 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).
3) Choose statistical distributions types and distribution parameters for all uncertain variables.
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 nonGaussian ( for example roll motion ) and non
stationary processes ( as for example lifting operations). Nonlinear load combinations may
also be considered, as for example buckling under biaxial loading or the VonMises 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 outcrossing 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
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
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
(,
WESTBOUND EASTBOUND
BASE CASE 0 0
1 m (13 %) OVERESTIMATE < 0.1 knots < 0.1 knots
00)
C
WESTBOUND EASTBOUND
ACHIEVED SPEED 20.4 21.1
.  _ __
gt___
g2() __ •PROBAN
_z) I • _ __ __ __
gn (Z )
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.
G 0,    0((  0"(
Where:
Fail
Limit state GIX) =0 On the limit state and
positive velocity
~Safe
•Time
Events describing an outcrossing
APPENDIX 1
.* Service margins
REPOR T
~REPORT
SEAWAY
INTRODUCTION.
PRINCIPLES.
* 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.
In the case of wind resistance and added resistance due to rudder and yaw
ings motions a semiempirical calculation method is employed in SEAWAY.
Propulsion system. Ship motions and added resistance also result in a de
terioration in the propulsive efficiency. SEAWAY includes a semiempirical
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.
Environmental.
Longterm 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.
* Geometrical data
* Ship still water resistance
* Propulsion factors
 Open water characteristics of the propeller
* Engine parameters
OUTPUT.
OPTWAY
Output includes:
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 standalone program
with preprepared data files for individual ship types
or as a direct post
processing program to SEAWAYS.
6
,REFERENCES.
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
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
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 []
VWhtk11W
20,
1 2 4 6
ýAi [Kv
Gr~vt&5aCe keitkV
Inf luence of Immersion and Ventilation on the open Water Diagram
,4 8
00m3
SQF
AJA WbanAC t
'10SL
P d
710
90
CC
LU
/ce
75 B85 90 .. 95 0 /...8
% ENGINE SPEED
VI7~;is
/ iM
q 7 /
T / It
to.
SEAKEEPING OF HIGH SPEED VESSELS
0. Faltinsen
Department of Marine Hydrodynamics
Norwegian Institute of Technology
N7034 Trondheim  NTH, Norway
INTRODUCTION
Catamarans and SES are well known to the highspeed 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.
SEAKEEPING CHARACTERISTICS
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 onedimensional 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.
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
rmsvalues and nnsvalues 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 highspeed 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 highrated engines and reduced engine
J condition between major overhauls may lead to engine breakdowns.
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.
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 highspeed vessel may have a slender forebody, i.e. not enough buoyancy
in the forepart to avoid "deck diving".
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 nonlinear 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.
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 vesselfixed
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.
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 highspeed 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 twoparameter JONSWAP spectrum recommended by IflC
was used in the calculations.
*) Half part obtained by intersecting along the centre plane of the multihull
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 zerospeed results.
This is not adviceable due to the forward speed effects on the hydrodynamic
pressure distribution.
CONCLUSIONS
Speed loss of catamarans and SES in a seaway are discussed. Even is a SES
looses easily the speed in a seaway andthe 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, 0., Zhao, R., 1991a, Numerical predictions of ship motions at high
forward speed. Phil. Trans. Royal Society, Series A., Vol. 334, pp. 241252.
Faltinsen, 0., Zhao, R., 1991b, Flow prediction around highspeed ships in waves,
8
"Mathematical approaches in hydrodynamics", Editor T. Miloh, SIAM, pp. 26528 .
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. 709725.
Faltinsen, O.M., Holden, K.O., Minsaas, K.J., 1991, Speed loss and"operational
limits of highspeed marine vehicles, Proceedings IMAS'91  High Speed Marine
Transportation, Sydney, Australia, pp. 21 to 29.
Faltinsen, O.M., Hoff, J.R., KvAlsvold, J., Zhao, R., 1992, Global wave loads on
highspeed catamarans, Proceedings PRADS'92, Newcastle, England, Vol. 1, pp.
1.3601.375.
Hoff, J.R., 1990, Threedimensional Green function of a vessel with forward speed
in waves, Dr.ing. Thesis 199025, Division of Marine Hydrodynamics, Norwegian
Institute of Technology, Trondheim, MTA report 1990:71.
Kaplan, R., Bentson, J., Davies, S., 1981, Dynamics and hydrodynamics of surface
effect ships, SNAME, Vol. 89, pp. 211248.
MeekHansen, B., 1991, Engine running conditions during high speed marine craft
operation, Pore. FAST'91, Trondheim, Tapir Publishers, Norway, Vol. 2, pp. 861
876.
Sorensen, A., Steen, S., Faltinsen, 0., 1992, Cobblestone effect on SES,
Proceedings HPMV'92, Arlington, VA, USA, pp. SEs17SES30.
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 • •
5.0
6.3
5.0
4.0 __
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
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
Fig. 3 050.0
Fatiguedecreased profiency boundaries for the azcomponen
t of the
acceleration as a function of frequency and exposure
time (ISO 26ea/n).
a z = RMSvalue of onethird octave band. f(= centre
third octave band. frequency of one
Speed range
,•(involuntary speed loss)
40 mn SES
 2
0
0I I I I
1.5 2.0 2.5 3.0 3.5 4.0 45
T2g7'L
0.2
0.0 0.11  I I I i
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
. .......... 0Heading: 1050 deg
a6Heading: 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
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
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
* Better comfort.
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:
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.
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
Simulation analysis
• ~Performancer
evaluation
4
 ~    .5  
Figure ,4: Monohmll with surface picirciug foils A roll uffscL, gives a scl fstabi 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 usedas 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.
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.
= [,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.
L.
7
h
F2
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.
,8
SIGNIFICA.NT, WAVEHEIGIrT IIH, (i)
03 flatxiiftinehew~ht )
00
411
AANW
0
 44l 3.50.
d o t
n el.0m. e l Sank sheltermdi
caiainlmtdrae
0 ihicesn
k pe.It L isdirbettytodnif
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.
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 "takeoff,'. 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.
'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 suboptimal 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:
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.
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 electrohydraulic 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
" Different flaps are dedicated to different tasks according to common engineering
sense, we get a quasi multivariable 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 suboptimal 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
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)
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.
I3
Figure 11):
it=B(,~jif~~~)
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 PDcontrollers.
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
55). 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.
" 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
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.
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.
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.
" 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).
hm
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
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.
7 Conclusions
This paper h;s trcatced flight controlwithout 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. . .
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.)
4.0
50
I )6
L 0. .00
S 0 AIN.
0 F I MODERATE
0.8
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
SED31 KNOT 2 3 4 5 6 7 8 10 20
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:
" Linear algebra: important for the analysis of multivariable systems and for
dealing with redundancy.
References
[1] Balchen, J.B. (1991). Nonlinear decoupling in process control. Modeling Identi
fication and Control, vol. 12, 8194.
[21 BenIsrael, 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. AES14, 456465.
[5] Frank, P.M. (1990). Fault diagnosis in dynamic systems using analytical and
knowledgebased redundancy  a survey and some new results. Auvomatica, vol.
26, ,159471.
[6] Ireund, E. (1975). The structure of decouplcd nonlinear systems. Int. J_ Con
hol, vol. 21, 443450.
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
 4657.
[8] Instanes, E., J.T. Pedersen (1991). Safe and comfortable operation of Foilcata
marans. FAST'91, 10931112.
[11] Lane, S.H., R.F. Stengel (1988). Flight control design using nonlinear inverse
dynamics. Automatica, vol. 24, 471483.
[12] Mclnroy, J.E., G.N. Saridis (1990). Acceleration and torque feedback for robotic
control: experimental results. J. of Robotic Systems, vol. 7, 813832.
[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, 12151231.
[14] Stark, D.R. (1974). The PHM hydrofoil automatic control system. Proc. Na
tional Aerospace and Manufacturing Meeting, 115.
[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, 110.
[17] RW. (1992). Foilcat ready for service. Ship & Boat International,July/August,
3134.
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
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 compensatingfor 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 sidehulls 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 sidehulls. 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 sidehulls permit the use of water propulsion, either waterjets or propellers can be used.
The
small draft of the sidehulls 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 sidehulls
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 zaxes 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 zaxis is called heave and is
denoted 71,(t). The rotation angle around the yaxis 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
sidehulls arc of minor importance. Nevertheless. since the mathematical model is supposed to be valid
also for lower frequencies, hydrodynamic loads on the sidehulls are included. Furthermore, infinite
water depth is assumed.
2
z
(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
3
The total pressure variations p,(xst), inside the air cushion is represented by
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 sidehulls 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
Q. Q, (6)
I I
PC
AQ
PO ......... .A p
The volumetric air flow out of the air cushion is proportional to the leakage area A.(t), which is
defined as
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
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.
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 onedimensional
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
The last two equations are the heave and pitch equations of the craft centre of gravity
including hydrodynamic loads on the sidehulls in addition to the loads induced from the dynamic air
cushion pressure acting on the wctdcck.
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
where r,(x) is the eigenfunctionxbrtlie mode shape function of mode j, which is found to be
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 ) =
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 coskL
4 C 2 a0.C sinanot (16)
where the wave volume pumping of the even acoustic modes for regular head sea is
k sinU
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•CSýl (2.+
and the corresponding eigenfrequency ow
s for the odd and even modes is found to be
3. Heave equation
4. Pitch equation
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 sidehulls. The
hydrodynamic addedmass 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 sidehulls. The hydrodynamic loads on the sidehulls 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 sidehulls in heave and pitch. Hence
SkL
sinU
= 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 sidehulls. In the case studied here, the submerged part of sidehulls are
assumed to have constant crosssection area. Examples of twodimensional frequency depending added
mass and wave radiation damping coefficients are found in Faltinsen (1990). Constant twodimensional
B,and A, values are assumed. The high frequency limit of the twodimensional addedmass 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 sidehulls. 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.
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 onedimensional 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 threedimensionaleffects. 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 twodimensional 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 ofthe 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
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 S1 "yA, (p.
mh.
+ p.)
.
(27)
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.
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 socalled spillover *effect in active damping of vibrations in mechanical structures (Balas,
1978). The inadvertent excitation of the residual modes has been termed contolspillover, 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 nonparametric 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 controlspillover. where CtXL. cos Ic(XL+Lf2)/L and Cdx, = cos kit(x,+Lf2)/L.
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
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)
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 timeinvariant system. The closedloop system is bounded input bounded output (BIBO) stable
if the equilibrium point x. = 0 of the autonomous closedloop system, that is vQt) = 0 in equation (29),
is asymptotical stable. Applying the proportional output feedback control law given by equation (36),
the linear timeinvariant autonomous closedloop system becomes
I = (A BGPC)x = Ax (37)
where the nxn closedloop system matrix A&= A  BGC is defined in Appendix C.
Theorem 1
The closedloop system given by equation (37) is asymptotical stable.
Proof:
Define the Lyapunov function candidate
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)
13
only if
= =
13 = 15 = P1 =P2  P, 0 (42)
Hence, by the invariant set theorem (Vidyasagar, 1993) the equilibrium point of the closedloop 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 closedloop 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 closedloop
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 sidehulls. 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) Fatiguedecreased 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 PiersonMoskowitz 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]
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.' . .. ........ ..... ........ .  ....... ...
.................. . . .......
..............
30  ... . ..............
.. . ..   , .........
_ . ... . . .. ........... .... .............
20 IV• .... ....... :
120
C.
0 _ 4 6 8 10 12 14
Frequency of Encounter [Hz]
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
"r
.....• i  '," "• ............ .. ..... ......:......... ..... •..... .......... ............ ............. .....
So I . " "
...
... i:; ..
.....
*~01
Acceleration
pap0
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
___________________________
0.5
0.4
0.3
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
0.2
0.25
0.25
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 (NFNF)
21
REFERENCES
BALAS, M.(1978). Feedback Control of Flexible Systems. IEEE Transaction of Automatic Control,
Vol. AC23, No. 4, pp 673679.
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 HighSpeed Ships in
Waves. Mathematical Approaches in Hydrodynamics, SIAM.
FALTINSEN, O.M., HOFF, J.R., KVALSVOLD, J. and ZHAO, R.(1992). Global Loads on
HighSpeed Catamarans. PRADS'92, Newcastle, England.
KAPLAN, P. and DAVIS, S.(1974). A simplified Representation of the Vertical Plane Dynamics of
SES Craft AIAA Paper No. 74314, 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.:
902011.
SALVESEN, N., TUCK. E.O. and FALTINSEN, O.M.(1970). Ship Motions and Sea Loads.
Transaction SNAME. 78, pp 345356.
STRANG, G.(1988). Linear Algebra and its Applications. Third Edition. Harcourt Brace Jovanovich
Publishers. Orlando, Florida.
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, 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.
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
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
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
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
diag[c
$ 1 ] (B.9)
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
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
L 227. )LI2
26
APPENDIX C. Symbolic Matrices Used in Stability Proof
The nxn closedloop system matrix A = A  BGC
4 is defined as
AaO,,
($15,sC15.) 0,•I•1
05,, (S2 5,C2 1 9ý ) (C.l1)
(
00000 0
0000 0
0000 0
C1X 00 0 00 0(
0 0 0 0 P"2 g,
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• g2cosL 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)
U11 u2 3 .. . u
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
= 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•
Q23' k = 0 •0 0 .. . 0
•qQ(2. qss.&
, 2 q;(s.. 3 ) . q(s.k ) (C.12)
where hi and t, are defined in equations (C.4) and (C.6). The kxk dimensional Q3,,1' matrix is
Jc
q(s5k.I)(jkz)
C
q(•sk.?)(s&.l)
C
•q(ls.*
C
Q3L
(C.13)
C
cC
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)
%a Ow 0. 'Q2;, 1(Q2;(Q
QI1J," > 0
The nxn diagonal positive definite P,, matrix can. be found from the Lyapunov equation
(AdTPa + PA4 = Q0) to be
ggP , KC "" ; p, 1
30
Bode Plot Collocation
101
102
100
w 0
00 h
1001
100Boepo.y/I
101
102
103
103 102 10' 100 101 102
Hz
100
200
300
103 102 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 
Fj~~~~~ ~ ..........
50.........
0
30  2
of..Encounter..H..
Frequency
off
ROSFeuec Co no unted Nacllzae
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
i...
i....
...
i.....
i.......
.........

..
.
0.....4
Cz
0) 0
0 ....
1 O........ .................. ................ ................. :......... . ... .. ... ....
................ ...
........
lMode 3 ÷+÷÷+
Mode 2+
:÷
mI Mode 1 ' ÷ ÷ +2 +÷ ÷
m + ++
0 ... :
. . .... ...... ..
. . . . . . . . . . . . . . . . .......
........... ....... . . ...........  , + + + .....
..... .* . .. 1 
g
:Mode 1M o e l+ +: ++ +
++.. .. ..
 50  ..... .....
...
... .. ... .. ..... ........ ....... ..... .... ... .. ... ..
:Mode 2
Mode4
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
Asgeir J. Sorensen
Outline
* Background / Motivation.
* Mathematical Modelling.
* 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.
* Propulsion system:
 waterjet.
 propeller.
SES Features
 complexity.
 safety.
 loss of air.
 expenses.
CC 04
S .COO
0c a cd
 0>
_____ !34
g
~ ~5C4)
Pressure Characteristics
zy 11
7x
7 V7
115
>
V
* Total pressure variations in air cushion:
where
Pa  atmospheric pressure
Sp(t)  po
PO
(x , t) = P s (X0, 0
• PO
Modal Solution
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
ý = ao + a 1COS2J (LI
XRCS + 2L
+ JL 2)
+ a 2 cos2Je (xF+ )
j L
Control Objective
k+ 1 _ _____k+1
CL Moe+Sx
 
* 0
*C 0
C~~ XL*Moe
1 ~ModeO 1
 
Controlled mnodeý
Process
u [Yu
ucyv V
Hc Distubanc
Controller /Disturbance/
Control Algorithm
Actuator
L/2 L/2
Sensor
MOTION CHARACTERISTICS OF FLOATING VESSELS.
MOTION TRANSFER FUNCTION SPECIFICATION, DIAGNOSTICS AND
TRANSFORMATION.
4'
Wave
Wave
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.
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 Xaxis 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:
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 xaxis)
*; is the wave component phase angle
C1 is given by:
=CI cohk(z+d)
cosh kd (3.2)
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
C1 = C2 = C 3 = ekz (3.5)
C, cosh k(z+d)
cosh kd
C2 = cosh k(z+d)
sinh kd (3.6)
sinh k(z+d)
C3 = sinh kd
= Ca sina (3.7)
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)
The integration constants x., y., z. have been selected so that the average value is
zero.
= 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
•/= .
VI = N C. 03 ksinp(cosa)
ay
=2 = C3 kcosp (+cosa)
Y3 = 8x C2 ksinO (sina)
$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 xaxis.
Other motions
All other motions or other responses, r. that are linearly dependent on the waves, are
expressed in the following way:
where
r : harmonic amplituide'
=
: phase angle relative to the surface elevation.
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
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
Derivative 0 1 2
Motion Displacement Velocity Acceleration
Derivate 0 1 2
Motion Displacement Velocity Acceleration
Sway rl sin[ C2 w
osin[ C 2 t_0 to2 sin[p C,_
Heave C . WC3 2
c,c3 a
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
Z = C, C3 exp[i(ot + (0  t)P]
IzI = C, C3
Arg Z = + woC + 4)c + 4%,
exp (iArgZ) = cos (ArgZ) + isin (ArgZ)
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 righthanded Cartesian coordinate system is used with xaxis pointing upwards.
The sea surface elevation is:
where
3 is te propagation direction
4'. is the phase angle, interpreted as forward phase shift
1. Denoting the base case as case no. 1, the following alternatives are discussed:
5. ofnen.ists
of ;r.,Cl convention so th;it + denotes phase lag instead of
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.
Case 1 2 3 4 5 6
110. Comingfrom Phase lag Mirror image
PaX sin(oxo)
meters ,
i
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.
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).
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
0..
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
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:
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:
$M15 = $W5  ý
$115 = $595
For problems involving only absolute motions these two sets of phase angles are
equivalent.
SYSTEM PARAMETERS:
tdent. TANKER
Aml..deratoRAO
REFERENCE POINT:
C.4 XP1 .00
2.2 XP2 : .00
2I............................................XP3 : .00
0 .2 0.4 0 .6 0 .8 1.0 1.2 1.4
Pnose ongle (degrees) Anigutor Frequency (Rod/T)
2. 1).)8.25I
2.0
so
 0
 SO.
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
:Vk
I\ 0I 1 1WCC1 i Fr
I9).
O9.5I
AffptLtude rotto Z/I
S4.g ,
3
2
0.I
ýge
P~osI oges)Angutor frequency (Rod/TI
300
0
2100
Anp~tttude ratio (I
0.0
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
1718 February. 1993
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 PCprogram suitable for parameter variation studies in a planning or
onboard situation.
The towline dyrmmics part of the model is extensively tested against the nonlinear 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 ForTOWCAPtobeanefficientPCtOOl,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. Ihs 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 NTI
Ivary
1970j. Hi( has been working with research prortcts related to n"rinto 1 erations, rn~ring systen"sand stationkeeping since 19'.
102 Offshore 93 Installabon of Major Offshore Structures and Equipment
. \% IN 'II
towline and winch dynamics. The response of winch and conditions at one end in order to satisfy specified boundary
towline is strongly nonlinear and, in addition, coupled. conditions at the other. Using this procedure a fairly accurate
Frequency domain methods are superior when it comes static equilibrium configuration for a multisegment line can
to computational efficiency, but they require a linearised be obtained with a minimum of computational effort.
model of the system. To account for nonlinearities a time
domain integration must be used. These methods are much
more demanding both with respect to computing time and Calculation of relative motion between tug and tow
ta storage.
Based on these considerations a hybrid method is chosen. Transfer functions
First order vessel responses are well described by linear A motion transfer function describes the linear relationship
equations. Neither tug nor tow will be significantly influ between regular wave amplitude and vessel response. The
enced byline with respect to firstorder response. A frequency vessel response is given by the motion amplitude and phase
domain analysisbetwen
canow,
give
ug information
nd isinptontothetim
hic the relativedoain
motion vessev tonse eleva tion
waveelevation. ample an ,
relative to the wave By using complex notation,
between tug and tow, which isinput to the time domain information on both response amplitude and phase can be
analysis simulation ofthe towlinewinch system. A stepwise contained in the same function.
description of the solution method is given below. In TOWCAP the motion transfer functions are given
1. Static line configuration is calculated. The static con numericallyforeachdegreeoffreedom (DOF) fora number
figuration is the basis for calculation of the dynamic of frequendes and directions. Linear interpolation is used
towline response. Maximum sag depth is also obtained between the given values.
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 Theinputtransfer functions are referring to theorigin in the
line attachment points, can thereby be established, local body fixed coordinate system. The transfer function
3. Bycombiningtherelativetowlinemotion transfer func for relative motion between tug and tow in the towline
tion with the wave spectrum, the response spectrum for secant direction is obtained by the following procedure (all
relative motion in the secant direction of the towline is angles and coordinates are defined in Fig 2):
obtained. 1. The transfer function for the towline terminal point at
4. From the response spectrum an extreme representation the tug is obtained by a linear transformation of the tug
of the relative motion can be generated transfer function.
5. The extreme motion is imposed on a single degree of 2. The transfer function for motion in the towline secant
freedom system consisting of a simplified model of the T
direction is then obtained by a rotation ac = a  OTr
towlineand thewinch. Duetostrong nonlinearities this around the local zaxis.
part of the analysis will be performed in the time do The procedure (1) to (2) is repeated for the tow with
main. An integration along the time axis will give time T h 0p re180 In addition the phase angle for the toww
series for cable tension as well as tile winch action. ao T10.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 relative to the tug (XCTO,YGIO) and the wave direction, P.
closely in the following subsections. The w'ave heading relative to the tug is PTu = P  0 "TU,
and
Special attention is given to the model of the towline to the tow PTO = p  0o° .
winch system. The transfer function for relative towline secant motion,
11,at wave heading, P3,is obtained by adding the two transfer
Calculation of static line configuration functions:
a''P
XlTT
XTO
(• TON 0TO. Y L~
Relative motion response spectrum a and zerocrossing 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:
'/•(I)
/
/
Fig 3 Line dynamics model
T =x 2 . (6) x (12)
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:
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 105
C.
m'i< + c'i + (k'+ kE) X= kEu + F1 T s (18) 1. motion transfer functions in 6 DO:F f Or the tug;
describing the dynamics of the towlinewinch system in the A data base of motion transfer functions for standard
rendering mode is now established. Equation (12) is of first vesselsand structureswill be included. When data is lacking
order and equation (18) of the second order. Some further for either tug or tow (or both) these can be scaled and used.
manipulation brings these equations into a form suitable to as input totheanalysis. If thetow is much larger than the tug,
be solved by astandard numerical scheme: the relative motion will be dominated by the tug response,
and the tow transfer functions may be omitted.
I[t[(k+ k,_)x c'i+ ku+ F,, TJ (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 towlinewinch 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 RungeKutta (RK) method is used. RK
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:
The following system data are rtMlifired to perform a The ieterfacebetween TOWCA', external filesand the user
TOWCAI] analysis: isillustrated in Fig 5.
106 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
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)
Systemfixed 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 'blackbox',ie they require no 8.0 3 4
manipulation by the user. This is desirable for onboard use 10.0 5
of the program.
FESTING AND VERIFICATION The line dynamics model has been tested against the
finite element program RIFLEX, which again has been exten
sively verified,' ' (activity (3)).
Testing of the TOWCAP program is split into several activi Activities (4) and (5) will be carried out as part of a full
ties as follows: scale test, which is planned in cooperation with several tug
1. testing of the calculation of relative motion; boat operators.
Results from activities (1), (2)and (3)are presented in this
2. testing of the sag depth calculation; paper.
3. testing of the calculation of dynamic line tension;
4. testing of the coupled winch  towlinedynamics calcu Test procedure
lation.
5. testing of th overall performance of the program. A system consisting of a tug boat and a towline is simulated.
Thetow isassumed to be much larger than the tug, so that the
The routines for calculation of first order vessel response, tow response n be neglected. The tug is W.ring in head seas
are taken from the MARINTEK program MIMOSA2.' The and onlythesurge motion isconsidered. The relative motion
implementation of these routines in TOWCAP has been between the line ends will thereby be equal to the surge
tested by comparing the calculated response directly with motion of the tug.
"'sUlts from MIMOSA2 (activity (1 )). Four parametersare varied in the verification stud y of the
The sag depth calculation has been checked against ana dynamic line model: line length, L, tug force, T , significant
lytical expression (activity (2)). wave helight, H., and zerocrossing period of the wave spec
Offshowe 93 Instalation of Major Offshore Structures and Equipment 107
Table IIl Comparison of extreme surge motion Table V Companson of dynwJn line tension
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
a0  towline directior relative toIiocalI tow coordinate 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
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 DB102. 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 DB102 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 preinstalled 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.
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 rereeve 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:
Job Preparation:
Personnel on the bridge of the DB102 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 GMvalues 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 GMvalues. The dynamic
simulation mode is a very realistic crane vessel simulator, which is used to train ballast operators for
lift and ballast operations.
Measured:
 slew angle (deg)
 outreach (m)
 load (t)
 side lead (deg)
 hook speed (rnlmin)
Calculated:
filtered load Wt)
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.
Initially anchors were set according to anchor plan (ref. Figure 6).
Using the various types of preinstalled reference systems (Simrad HPR, taut wires and
Artemis) the
dynamic positioning system is able to keep the vessel in a prearranged 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 tautwire 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 45.
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 overtension 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 overtension. 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 preset
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
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 offrmn after total thruster failure
 maximum position offmn 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 repositioning 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 DB102 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 DB102 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.
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
IV
it l FI
r ,JI
. "N__ f; 1V 1 iillt I=L '
L  _ _
)O .fIm I:
Ii~ili il Li
7afo
7, _
IL 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, 811 March 1993
(EC COMETT Program)
C.T. Stansberg
Abstract
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 lowfrequency (nonlinear) wave loads, and also possible non
linear characteristics of mooring lines, risers etc. The lowfrequency 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.
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 setup 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 18hour storm record. It is seen that
the extreme offset and the corresponding line tension are mainly dominated by a large
lowfrequency component, which are inductd by 2nd order nonlinear 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 lowfrequency as well as of a wavefrequency 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
lowfrequency contributions, combined with nonlinear 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:
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
lowfrequency 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.(23). 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:
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:
In a typical 3hours record with wave periods around 1015 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 3040% of the record
standard deviation t.
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 higherorder contribution.
In general, for the horizontal motion of large moored floating structures, the 2nd order
term may become the dominating one, since the lowfrequency 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:
(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, nonlinear) 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( 
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 lowfrequency (differencefrequency) as well as of a highfrequency
(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 lowfrequency
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:
It is evident from the discussion so far that extreme events with moored floaters is not a
straightforward 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 nonlinear 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.
For the physical understanding of the slowdrift 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 lowfrequency 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:
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 lowfrequency dynamical response characteristics hR(t) (or HR(f)), or, in
other words, the slowdrift 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 slowdrift motion response will follow a distribution in between
the exponential and Rayleigh models, as demonstrated by the model test results (figure 5).
The damping of the slowdrift 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 wavedrift 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, nonlinear damping contributions
may sometimes reduce the largest extremes.
A more detailed description and discussion of the model referred above is found in ref.[9].
Extreme mooring line forces, occuring as a result of the extreme slowdrift motion
described above, may be "amplified" due to the nonlinearities in the mooring system.
First, a nonlinear quasistatic 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 nonlinear stiffness will also enhance the largest
wavefrequency tension amplitudes. The wavefrequency component is, however, not a
quasistatic response to the surge motion, but rather dynamical response to the tangential
upperlineend motion. An additional, significant contribution to the largest tensions is
therefore due to dynamical effects from inertia and nonlinear drag forces on the lines,
which results in an instantaneous, dynamical mooring line stiffness considerably higher
than the quasistatic 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 nonlinear 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.
The prediction of extreme slowdrift 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 slowdrift 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:
where A and B are dampingdependent coefficients. See ref. [5] & [9] for more details.
The prediction of the connected extreme tension load is more complicated. In general,
longduration model tests and/or nonlinear timedomain numerical simulations may be
necessary. A simplified procedure might follow the outline given below. See ref.[3] for
more details. First, the extreme quasistatic lowfrequency tension may be found directly
from the extreme motion estimate via the tension/offset curve (figure 10). The connected
wavefrequency tension component connected with this event is strongly coupled to the .
lowfrequency component, and is physically not separable from it due to the timevarying
quasistatic 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 inline wavefrequency vessel motion and the corresponding tension is
known (e.g. from numerical computations or from model tests).The wavefrequency
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 lowfrequency and highfrequency 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 Rayleighbased estimates. The reason is
the large
influence from nonlinear physical mechanisms, such as 2nd order slowdrift motion
oscillations and nonlinear 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 slowdrift 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.
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.
9. Stansberg, C.T. 1991, "A Simple Method for Estimation of Extreme Values of
NonGaussian SlowDrift Responses", Proceedings, Vol.Ifl, 1st ISOPE Confernce,
Edinburgh, Scotland, pp. 442451.
10. Huse, E., 1991, "New Developments in Prediction of Mooring System Damping",
OTC6593 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
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
H [}
mI]3 10T
o
510
WU 5"10;
W 
?RayleigV
L distribution
d 5
2 3 4 5 6 a 10 15
INCLUUDIRNG
EXR'ECE SUR GE
5 jI SUGESURGE
20
 ST = 1 s = 13. . + o
1 I I
.A Tm Aismles power spcr an prbbiiyditiutosfrm18hu.
is~tor etSeis
Fiz. 5.A Time hislory samples, power spectra and probability distributions from
iShours tes series.
+ • s 
II.
........
, .._,___,
 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