THE NONLINEAR DYNAMICS OF SHIPS IN BROACHING

Kostas J. Spyrou1, [email protected]

Abstract commonly arising in waves which come from behind and

propagate in a direction forming a small angle, say 10-30
Recent research on ship motion dynamics has led to the
deg, with the longitudinal axis of the ship. Although the
discovery of the underlying causes of broaching. This is a
inception of broaching represents a problem of instability
type of ship motion instability which is manifested on the
on the horizontal plane, capsize may be incurred at the post-
horizontal plane with a sudden divergence from the initial
critical stage due to development of large heel as energy is
course and it may end with a rapid ship capsize. We shall
transferred into the roll direction.
summarise the key dynamical phenomena which are
Broaching could happen to small as well as to larger
responsible for broaching.
ships. It is notable however that the dynamics involved do
not seem to fit always into a single pattern. Frequently,
1. INTRODUCTION broaching is manifested as a sudden divergent yaw, which
Ship survivability against capsize in heavy seas has become peaks within a single wave length. Control is lost when the
one of the areas of primary concern among ship researchers, middle of the ship lies somewhere on the down-slope and
designers and regulators in recent years. When a ship is nearer to the trough. In other cases there is a gradual,
subjected to the effect of large waves it may capsize oscillatory-type build-up of yaw as successive waves
impinge on the ship from behind. In moderate sea states a
according to a number of different scenarios, depending on
ship is more likely to broach-to if it runs with a high speed
the magnitude and direction of the wave excitation and the
and is slowly overtaken by the waves. Broaching may also
ship’s own capability to resist such excitations. Resonant or
occur at lower speeds if the waves are very steep.
breaking waves approaching a ship from the side (“beam-
In this paper we shall present an overview of our
seas”) have a potential to excite large rolling which could
research on broaching which spans a period of five years
result in capsize, especially if the intensive oscillation of the
and has taken place in Japan, the UK and Greece, on the
ship causes shift of cargo or, if a considerable quantity of
basis of fellowships provided by, or linked with, the EU [1-
green water is shipped on the deck.
13]. Key contribution of this research has been the
More dangerous still can be a group of steep and association of specific nonlinear phenomena with broaching
relatively long waves approaching a ship from the stern behaviour. Nonlinearity due to the effect of the wave force
(“following-sea”). Waves of this kind are known to incur is a key factor when the surge dynamics are involved in the
significant reductions in roll restoring capability (i.e. the creation of broaching. Consideration of stiffness
tendency to return to the upright position) for many types of nonlinearity in the roll direction is also essential when
vessels and they may also instigate dangerous coupled studying the possibility of a capsize occurring in the wake
motions. According to a popular classification, in of controllability loss. Simple analogues of the surge and
following-seas a ship may capsize in at least three different roll dynamics are shown in Fig. 2, while some basic
ways: Pure-loss of stability is a sudden, non-oscillatory type concepts about nonlinear systems are presented in the
capsize taking place around a wave crest due to slow Appendix.
passage from a region of the wave where roll restoring has
become negative. Parametric instability is the gradual
build-up of excessively large rolling created by a
mechanism of internal forcing, the result of a fluctuating
restoring that depends on where the ship lies in relation to
the wave.
The third distinctively identified, yet until recently not
fully understood, type of capsize is broaching (Fig. 1).
Broaching is an unintentional change in the horizontal-plane
kinematics of a ship. Broadly, it may be described as the
“loss of heading” by an actively steered ship, that is
accompanied by an uncontrollable build-up of large
deviation from the desired course. Broaching is more

1
Department of Naval Architecture and Marine Engineering,
Ship Design Laboratory, National Technical University of Athens,
9 Iroon Polytechniou, Zographou, Athens 157 73, Greece. Fig. 1: A typical example of broaching behaviour
 2. BACKGROUND understood by observing the motions of a scaled model and
by analysing their time-series. Very notable were the
2.1 The international regulatory regime experiments of Nicholson [27], Fuwa [28], Marshfield [29]
The stability regulations which non-military vessels need to and more recently of Umeda [30] and DeKat & Thomas
satisfy are usually agreed at the International Maritime [31]. Perhaps the most complete investigation of broaching
Organisation (a United Nations Agency usually referred-to took place at the Ship Research Institute of Tokyo in the
as IMO) and then enforced by the Governments of the early eighties. It combined tests on a radio-controlled
Member States. Currently there exist two sets of criteria: model of a fishing boat, independent measurements of
The first rely on a statistical analysis of accidents that was hydrodynamic forces, comparisons with data obtained from
carried out in the thirties and involved only the static roll the real, full-size, boat and theoretical analysis [24, 28].
characteristics. They are supplemented by a second set, Several important issues came out of all these studies;
providing some simplistic account of the roll dynamics due like for example that the instability should occur when a
to side wind and waves (“weather criterion”) [14]. ship lies on the down-slope of the wave and the stern is
Broaching remains basically unaccounted although two “resting” on a crest. Also, that in a broaching situation the
requirements are of some relevance: The limit of a ship is overpowered by an excessive wave yaw moment
maximum 100 heel, due to a prescribed heeling moment which cannot be counterbalanced by the maximum moment
associated with a ship turn on the horizontal plane produced by the rudder. Some suggested also that the
(applicable to passenger ships); and the requirement of a reduced effectiveness of the rudder on the wave’s down-
minimun acceptable turning rate of the rudder (the slope might be an important factor [21] while others
prescribed minimum is rather too low for preventing considered this to be a secondary effect [24].
broaching however) that appears in SOLAS, the IMO
Convention for ship safety [15]. Notable is also a recently
developed at IMO guidance for the ship Master which is
intended to help him to avoid dangerous situations. It is
based on a combination of theoretical considerations with
extensive experimental results and is assumed to cover,
albeit in an empirical way, wave environments where
broaching is likely to occur.

2.2 Summary of earlier research

Fig.2 : The concept of a ball rolling in a potential well offers
In the USA in 1948 Davidson [16] proved that a ship which some simple analogues for the roll (left) and surge
could keep a straight-line course in calm water might be (right) ship dynamics.
unable to achieve this if it encountered following waves. In
1951 in Germany Grim [17] pointed out the possibility of
occurrence of surf-riding if long and steep waves approach 2.3 Possibilities for improvement
a ship for the stern. Surf-riding is a peculiar type of Earlier analyses were mostly content with an ordinary
behaviour where the ship is suddenly captured and then stability examination based on the sway, yaw and possibly
carried forward by a single wave. This has been believed by surge motion equations, with different positions of the ship
many to work as a precursor of broaching. It has been considered on a selected reference wave. A number of
studied in Germany, Russia Japan, Australia and elsewhere, crucial simplifications are innate to these approaches
with most notable a relatively recent study in Japan by Kan however. Firstly, they have relied on linear motion
where were identified the fundamental aspects of this equations. However, nonlinearity can give rise to responses
phenomenon [18]. All these studies concentrated however that are not deducible, even at a qualitative level, from a
on the single-degree surge dynamics and, because of this, an linear low-amplitude analysis. Therefore critical effects may
explanation on how surf-riding is linked with broaching remain unaccounted and terms which could probably
could not be developed. explain broaching were dropped from the equations.
Well known theoretical studies on broaching were Another weakness is their exclusive concentration on static
carried out in the UK by Rydill [19], Du Cane & Goodirch and steady-state behaviour. The ship is usually assumed to
[20] and Renilson & Driscoll [21]; in Holland by Wahab & be at quasi-static equilibrium and to travel at exactly zero
Swaan [22]; in the USA by Eda [23]; in Japan by Motora et frequency of encounter (or momentarily 'frozen' at certain
al [24] and Umeda [25]; and in Russia by Ananiev & positions of the wave). However, broaching is a
Loseva [26]. Broaching has been studied also with phenomenon of dynamic nature and therefore, the study of
experimental methods: Tests with scaled radio-controlled transient behaviour is essential. Steady-state analysis
physical models have taken place in large square ship usually results in overestimation of a system’s safety
model basins of the UK, Japan and the USA, in the hope margin.
that the pattern of a ship’s behaviour in the waves could be
 3. ELEMENTS OF THE NEW APPROACH freedom (surge, sway, yaw and roll) plus the angle of the
rudder when automatic control through an autopilot is
Our research on broaching was comprised of the following present. The controls vector a is comprised of two sets of
stages: parameters: the ship based and the exogenous. In the first
(1) Firstly, a mathematical model of coupled ship motions group belong the desired heading and the constants (gains)
in long and steep regular waves was developed, of the autopilot. These are substituted by the rudder angle
suitable for frequencies-of-encounter2, between the when there is no continuous rudder control (then ship
ship and the waves, near to zero. behaviour is simply examined for specific rudder angles).
(2) The analysis of surf-riding was extended to cover a Also in the ship based controls belong the nominal Froude
quartering sea (i.e. waves meeting the ship from number and the ship’s metacentric height, the latter
behind but with a non-zero angle). This is the wave dictating initial roll stability. The Froude number is a
environment where broaching is most likely to occur. nondimensional version of the ship’s forward speed which
Bifurcation analysis (see Appendix) was carried out. is a state variable. However in the nominal form it is based
(3) The physics behind the connection of surf-riding with on the steady speed obtained in still water, which is a
broaching and capsize was explored. Furthermore, the representative of the propeller’s rate of rotation. Hence, the
range of ship control parameter values (for given wave nominal Froude number is a control variable. The
characteristics) where broaching takes place was exogenous control parameters considered were, the length
identified. and the maximum steepness of the wave. Since the wave
excitations could be expressed as functions of the ship
(4) The occurrence of broaching at lower speeds where
position on the wave, the direct time-dependence could be
surf-riding is not involved in the generation of the
removed from the excitation terms.
horizontal-plane instability was studied.
5. NONLINEARITY IN SURGE DYNAMICS [5, 18]
4. BASICS OF THE MATHEMATICAL MODEL
In a following or quartering sea the ship motion pattern
We applied Newton’s second law for linear and angular may depart from the ordinary periodic response. In Fig. 2
momentum in order to describe a ship’s motion in reference we have considered the changes in the geometry of surge
to an inertial system of axes fixed at a trough of the motion under the gradual increase of the nominal Froude
incoming regular wave system. These axes move with a number, Fn . We have assumed an environment of steep
constant velocity which equals the wave celerity. The waves coming from behind, say with λ L = 2.0 ,
motion equations were transformed subsequently in order to H λ = 1 20 where λ , L are respectively wave, ship
fit a non-inertial axes system, having its origin at the middle
of the ship. length; and H is the wave height. The Froude number
Ship motion is controlled through the magnitudes of the corresponding to a speed equal with the celerity of this
hydrodynamic loads produced by the propeller and the wave is 0.564. For a low Fn there is only a harmonic
rudder. But the type of motion depends crucially on the periodic response [plane (a)] at the encounter frequency.
hydrostatic and hydrodynamic loads acting on the ship hull. However, as the speed approaches the wave celerity the
To facilitate the calculation of the hull forces these were response becomes asymmetric [plane (c)]. The ship stays
separated into the so-called manoeuvring forces due to
hydrodynamic reaction caused by the ship motion, and
forces produced by the existence of the waves. The
hydrodynamic reaction part is influenced by viscosity
especially in the stern area and for this reason it cannot be
predicted accurately. For simplicity we assumed these
forces to be suitable analytic functions whose coefficients
were identified from experiments with a scaled model. With
similar experiments we could measure also the wave
excitations, though for these we could reliably follow also a
theoretical approach.
The derived set of motion equations was brought into
the standard dynamical system’s representation
dz dt = f (z, a ) through suitable transformations. t is time
and z is the state variables’ vector comprised of the
Fig. 3: Transformations of a ship’s forward motion. On each
displacements and velocities for the considered degrees of plane the horizontal axis is the speed while the vertical is
the ship’s relative position on the wave. The horizontal
2
This is the modified wave frequency which the ship actually axis across the planes is a representative of the
experiences due to forward speed, a Doppler-like phenomenon. propeller’s rate of rotation.
longer in the crest region and passes quickly from the
trough. In the literature this referred as asymmetric large-
amplitude surging, or, surfing on a crest.
In parallel an alternative, stationary behaviour starts to
coexist, owed to the fact that the resistance force which
opposes the forward motion of the ship in the water, can be
balanced by the sum of the thrust produced by the propeller
and the wave force along the ship’s longitudinal axis. This
is known as surf-riding and the main feature is that the ship
is forced to advance with a constant speed that equals to the
wave celerity. On a plane having as axes the position and
velocity of the ship along the wave direction, the surf-riding Fig. 4: 2-D projections of the curve of states of
states appear in pairs: one nearer to the wave crest and the surf-riding in a quartering sea.
other nearer to the trough. Nearer to the crest they are
unstable. Stable surf-riding can be realised only in the should become impossible to reach. We have found that
vicinity of the trough. when the heading of the ship does not coincide with the
Surf-riding is characterised by two speed thresholds: the wave direction, the points of surf-riding equilibrium belong
first is where the balance of forces becomes possible [plane to a closed curve. 2-D projections of this curve, identified
(b)]. For a fishing vessel that we examined in detail this by coupling our mathematical model with a continuation
arose at a Fn about 0.32 (for a wave with algorithm, are shown in Fig. 4. Stability is very dependent
λ L = 2.0 , H λ = 1 20 ) which was about 57% of the wave upon the method of rudder control. We investigated
stability with the rudder: (a) fixed at certain angles, and, (b)
celerity value. The second threshold (occurring at a Fn
controlled by an autopilot. For the latter feedbacks based on
slightly higher than 0.4 ) flags the complete disappearance
the yaw angle (the heading error from the desired course)
of the periodic motion. This critical for safety event is
and the yaw rate were used.
manifestation of a global bifurcation phenomenon, known
Very interesting is the occurrence of “Hopf”
as homoclinic saddle connection5. It occurs when a periodic
bifurcations (points B in Fig. 4) creating stable oscillatory
orbit collides in state-space with an unstable equilibrium
surf-riding (Fig. 5). It is notable that such oscillatory motion
(the surf-riding state near to the wave crest). A dangerous
on a single wave had been observed in experiments
transition towards some alternative, nearby or distantly
performed in Japan
located, state is the practical consequence.
[18] and it is
It has been shown that the underlying dynamics of surge
believed that an
resemble those of a pendulum with constant torque. A
explanation has been
generic form of the surge motion equation is as follows:
provided by our
2 3
d 2x dx  dx   dx  analysis. Below will
+ b1 + b2   + b3   + f sin kx = d
dt 2 dt  
dt  dt  be summarised the
key phenomena, as
where x is the relative position of the ship on the wave
identified from an
measured from a trough, t is time, bi , i = 1,2,3 are investigation based
coefficients that depend on the nominal Froude number and on fixed-controls (no
the wave celerity; f is the amplitude of the wave force for autopilot). This case
surge; and d is the difference between thrust and was found to be the
resistance for a speed equal to that of the wave. most interesting in a Fig. 5: Self-sustained oscillations in
The periodic motion of a ship overtaken by waves is dynamical sense. the surf-riding mode
represented by full rotations of the pendulum whereas
surf-riding corresponds to the equilibrium states. The key
nonlinearity is the sinusoidal nature of the stiffness term. 6.2 Chaos and further homoclinic events [4]
We have used the angle of the rudder as the bifurcation
6. THE QUARTERING SEA parameter. The self-sustained oscillations created at a point
near to the middle of the wave’s down slope (point B in
6.1 Continuation studies [1] figures 4 and 5) went through a period doubling cascade as
From a continuity consideration one should expect surf- the rudder angle δ was increased slightly. A projection of a
riding to be realisable also for a range of non-zero angles of chaotic response is shown in Fig. 6 together with the
heading. This range however should not extend too far; corresponding power spectrum showing the frequency
because as a beam sea situation is approached, the wave content of the response. The existence of chaos was
force in the surge direction will diminish and equilibrium confirmed also through calculation of Lyapounov
exponents. These are the trough the wave yaw moment produces a destabilising
basically time averages effect. The homoclinic connection shown in Fig. 3 is
of the logarithms of the therefore the key event behind broaching. Once the periodic
moduli of eigenvalues. response is vanished, the likely types of motions are, either
The important point is surf-riding or broaching. Which of the two should prevail
that they can indicate depends on whether the steering system is capable to check
whether initially close the turn before the yaw angle becomes too large.
by orbits tends to A broaching case is shown in the upper graph of Fig. 7.
diverge with an The closed orbit is the initial periodic motion as the waves
exponential rate. overtake the ship. The periodic motion passes from the
Further increase of vicinity of the unstable surf-riding point (near the crest) and
the angle δ brought the for this reason its shape is distorted. As the ship is set on the
response back to verge of the homoclinic connection any slight increase of
periodicity; but due to the propeller rate by an unaware Master would be enough to
the increasing δ the instigate broaching. Broaching could be avoided simply by
oscillatory response selecting a higher proportional gain for the autopilot Such a
came nearer to the case can be seen in Fig. 7 (lower). After the homoclinic
(unstable) points of Fig. 6: Chaotic attractor and power connection the ship reached a state of surf-riding lying near
surf-riding (see Fig. 5). to the trough.
spectrum.
This had as a result a
new homoclinic connection (unrelated with the one 7. GLOBAL DYNAMICS [7]
discussed earlier). This event ended the oscillations Generally, it is sensible to regard broaching as a change of
performed by the ship while it was carried forward by the state caused by variation of some ship-based or exogenous
wave. The ship was then forced to return to a more usual control parameter. This defines a problem of transient, and
motion pattern where it is overtaken by the waves. in our case also multi-degree, dynamics which we have
tackled by assuming that the change of state was effected
6.3 The link of surf-riding and broaching [5-8] upon a steady initial motion pattern. It is not unreasonable
The key for explaining why surf-riding is conducive to to imagine a ship to be in a nearly steady periodic motion,
broaching lies in the role of the nonlinear surge motion running ahead or before the waves, when suddenly a group
dynamics. As already of steeper waves approach it from behind. Or perhaps that
was explained, in large broaching the Master of a ship caught in surf-riding is attempting to
and relatively long alter the propeller revolutions, or the rudder angle, in order
waves (particularly for to escape and return to the ordinary periodic motion
heading (rad)

a ratio of wave-length pattern.

to ship-length between The variation of a control parameter should produce a
1 and 2 where there is transition towards some other state. What is essential here
evidence that the is that, in certain instances, an erratic type of behaviour,
)

12
ty (m/s

saddle near crest

11
tendency for broaching 10
9
namely broaching and/or capsize, is possible to emerge. In
veloci

is greater) a ship does 8

7 Fig. 8 is shown the result of a systematic exploration of the
6
surge

not approach smoothly 5 outcomes when a ship escapes from surf-riding. The ship is
4 -1 0 1
the condition of zero assumed to be lying initially at a fixed point with its
relative position on the wave, cos(2πx/λ)
frequency of encounter nominal Froude number, Fn , ‘tuned’ on the wave celerity.
but it jumps to it as 0.20 The desired heading, ψ r is taken in the range where stable
soon as a certain speed 0.18
0.16 surf-riding can exist. These control settings determine
threshold is exceeded. If 0.14
uniquely an initial (surf-riding) condition. Then, Fn is
heading (rad)

the ship’s longitudinal 0.12

0.10 suddenly reduced to Fnlow and the emerging type of
axis lies at some small 0.08

angle relatively to the 0.06 behaviour is noted. The procedure is repeated until the
direction of the waves,
0.04
0.02
control plane [ψ r , Fnlow ] is covered with sufficient
0.00
this dynamic effect is 11 resolution. Surf-riding is seen to occupy the high Fn
/s)

10
locity (m

‘imported’ into yaw. At 8

9
region. Immediately below is located the domain of
7
the ensuing stage, the 6
periodic motion which appears to be ‘in competition’ with
surge ve

5
ship passes from a 4 -1 0 1
broaching. At higher headings broaching and surf-riding
trough, where it should relative position on the wave, cos(2πx/λ) tend to interface directly. Most notable is how the broaching
experience a tendency domain becomes ‘eroded’ by capsize [4].
to turn because around Fig. 7: Simulation of broaching.
 During broaching a ship is subjected to combined ζ is the damping ratio of the steered ship in yaw which
excitations which include: a direct roll moment from the receives relatively large values unless there is no differential
waves, an ‘internal’ forcing caused by a time-varying roll gain in the autopilot;
stiffness, and a moment due to hydrodynamic reaction τ , Ω are scaled
arising from the sway and yaw motions. We proposed the with respect to the
following simplified roll model for broaching [10]: natural frequency in
d 2ϕ dϕ
+ (1 + ε cos Ωτ )(ϕ − ϕ 3 ) = f p cos Ωτ + f c (v, r )
yaw. The amplitude
+ 2ζ h of the parametric
dτ 2
dτ
ϕ is a scaled roll angle; ζ is the damping ratio; τ , Ω are, forcing is the ratio of
the wave yaw
scaled time and encounter frequency; ε , f p are the
excitation to the
amplitudes of ‘internal’ and direct moment due to the wave ship’s own static
(both depend on heading); and f c is the roll moment due to gain3. The bias
the sway, v , and yaw, r , velocities. parameter a plays
the role of a constant
external excitation Fig. 9: Jump to resonant yaw.
that depends on the
angle between the desired course and the wave propagation
direction. It is increased with the proportional gain of the
autopilot, with the desired heading and with the ship’s static
gain (a characteristic of a hull). On the other hand, it is
reduced if a hull is not very responsive to the rudder. As the
yaw equation is Mathieu type, parametric instability should
be expected. Due to the large damping however, the internal
forcing required is much higher (i.e. steeper waves) than
that at zero encounter frequency.
We considered this instability further on the basis of the
mathematical model outlined in Section 4. We found that
there is a critical desired heading (given the ship and the
wave) where a bifurcation occurs creating a stable
subharmonic response (Fig. 9). Further increase of the
desired heading caused a rapid increase of the amplitude of
yaw oscillation, leading shortly to a turn backwards of the
Fig. 8: Possible outcomes of an escape from surf-riding steady yaw response curve. This precipitated a sudden and
dangerous jump to resonant yaw. The transient behaviour
was a growing oscillatory yaw which corresponds closely to
8. BROACHING AT LOW SPEED [8] what has been described as cumulative-type broaching.
In a review by Conolly [32] of broaching incidents, those
9. CONCLUDING REMARKS
that had happened at speeds much lower than the wave
celerity were distinguished from the more classical high- In this review article were summarised the most important
speed events. At low to moderate speeds, surf-riding is of our findings regarding the basic nature of the broaching
rather unlikely to happen, unless the water is very shallow. instability of ships. Recently, we have widened the scope of
Therefore, if broaching has occurred in that speed range, our research by considering also other, single or combined,
there should be a different mechanism in action which does types of dynamic instability. Our long term objective is to
not involve surf-riding. develop a rigorous technical basis for the assessment of ship
We considered whether in that case the instability could stability in a seaway with full account of the large-
be an intrinsic feature of the yaw motion of a ship overtaken amplitude ship dynamics. This could lead to more rational
by very large waves. Assuming a nearly following sea and stability regulations, improved design and a better guidance
an autopilot capable to limit large oscillatory course for Masters on how to cope with adverse weather.
deviations, the fundamentals of yaw motion is captured by
ACKNOWLEDGEMENTS
the following equation [8]:
The EU grants with contract no: ERBFMBICT982963
d 2ψ dψ
+ 2ζ + (1 − h cos Ω τ )ψ = a (TMR) and ERBCHBGCT930427 (HCM) are gratefully
dτ 2
dτ acknowledged. Similarly is acknowledged the award of a

3
Proportionality constant between steady yaw and rudder angle.
STA Fellowship (ID no 193065) offered by the Japanese 23. H. Eda (1972) Directional stability and control of ships in
Government for which the author was proposed to the waves. J. Ship Res., 16/3.
Japanese authorities by the EU. 24. S. Motora, M. Fujino, M. Koyonagi, S. Ishida, K. Shimada &
T. Maki (1981) A consideration on the mechanism of
occurrence of broaching-to phenomena. JSNA, 150.
REFERENCES
25. N. Umeda (1998) New remarks and methodologies for intact
1. K.J. Spyrou (1995) Surf-riding, yaw instability and large stability assessment, 4th Int. Ship Stab. Workshop, St. Johns.
heeling of ships in following/quartering waves, Ship Technol. 26. D.M. Ananiev & L. Loseva (1994) Vessel’s heeling and
Res., 42/2. stability in the regime of manoeuvring and broaching in
2. K.J. Spyrou (1995) Surf-riding and oscillations of a ship in following seas, STAB’94, Melbourne, Florida.
quartering waves, J. Mar. Sci. Technol., 1/1. 27. K. Nicholson (1974) Some parametric model experiments to
3. K.J. Spyrou & N. Umeda (1995) From surf-riding to loss of investigate broaching-to. Int. Symp. Dynam. Mar. Veh. and
control and capsize: A model of dynamic behaviour of a ship Struct. in Wav., London.
in following/quartering seas. PRADS’95, Seoul. 28. Fuwa, T., Sugai, K., Yoshino, T. & Yamamoto, T. (1982): An
4. K.J. Spyrou (1996) Homoclinic connections and period experimental study on broaching-to of a small high-speed
doublings of a ship advancing in quartering waves. CHAOS, boat. Pap. Ship Res. Inst. 66, Tokyo.
6/2. 29. W.B. Marshfield (1987) HASLAR model experiments 1977
5. K.J. Spyrou (1996) Dynamic instability in quartering seas: to 1986. Rep. RE TR87315, Ministry of Defence.
The behaviour of a ship during broaching. J. Ship Res., 40/1. 30. N. Umeda (1998) Broaching experiments (still unpublished)
6. K.J. Spyrou (1996) Geometrical aspects of broaching-to 31. J.O. deKat & W.L. Thomas III (1998) Broaching and capsize
instability. 2nd Int. Workshop Stab. Oper. Saf. Ships, Osaka. model tests for validation of numerical ship motion
7. K.J. Spyrou (1997): Dynamic instability in quartering seas:- predictions, 4th Int. Ship Stab. Workshop, St. Johns.
Part II: Analysis of ship roll and capsize for broaching. J. Ship 32. J. E. Conolly (1972) Stability and control in waves: A survey
Res., 40/4. of the problem. J. Mech. Eng. Sci., 14/7.
8. K.J. Spyrou (1997): Dynamic instability in quartering seas:-
Part III: Nonlinear effects on periodic motions. J. Ship Res., APPENDIX: Elementary concepts of nonlinear dynamics
41/3.
9. K.J. Spyrou (1997) On the multi-degree, nonlinear dynamics Nonlinearity: Non-proportional relationship between the
of ship motions with application to the broaching problem. force and the variables describing the system’s response.
IUTAM Appl. Nonl.Chaotic Dyn. Mech., Ithaca. Steady and transient motion: The long resp. short term
10. K.J. Spyrou (1997) A new method to analyze escape response that follows the application of some steady
phenomena in multi-degree ship dynamics applied to the excitation on a dynamical system.
broaching problem. STAB'97, Varna.
State vector: The set of variables that can describe fully
11. K.J. Spyrou (1997) On the nonlinear dynamics of broaching-
to. Int. Conf. Des. Abnormal Waves, Glasgow. the motion. State space (or phase space) is the enlarged
physical space spanned by the state vector.
12. K.J. Spyrou (1999) Similarities in the yaw and roll behaviour
of ships in extreme astern seas. IUTAM NOMES ’99, Hanoi. Control vector: The set of parameters which we can vary
13. K.J. Spyrou ,(1999) On course stability and control delay. during the investigation. They define a control space.
Accepted in Int. Shipbuilding Prog. Continuation (or path- following): The tracing of steady
14. IMO (1995) Code on Intact Stability for All Types of Ships responses under variation of a control parameter.
Covered by IMO Instruments, Res. A.749(18), London.
15. IMO (1997) SOLAS, Consolidated Edition, London.
Bifurcation: A change in the qualitative characteristics of
the response; such as a change of stability and a smooth or
16. Davidson, K.S.M. (1948) A note on the steering of ships in
following seas. 7th Int. Congr. Appl. Mech., London.
abrupt creation/disappearance of a pattern of behaviour.
Bifurcations are distinguished into local and global. Typical
17. O. Grim 1951 Das Schiff in von Achtern Anlaufender See.
JSTG, 45. cases of local bifurcation are, the creation of an oscillatory
response despite the absence of external periodic excitation
18. M. Kan (1990) Surging of large amplitude and surf-riding of
ships in following seas. Sel. Pap. Naval Archit. Ocean Engin., (Hopf bifurcation); the doubling of the period of an
Soc. Naval Archit. Japan, 28. oscillation (flip bifurcation); and the folding backwards of
19. L.J. Rydill, (1959) A linear theory for the steered motion of the steady response curve as a certain influential parameter
ships in waves. Trans. RINA , 101. is varied (fold bifurcation). This bifurcation usually creates
20. P. DuCane & J.R. Goodrich (1962) The following sea, a sudden jump phenomenon.
broaching and surging. Trans. RINA, 104. Chaos: Random output from a regularly forced system.
21. M.R. Renilson & A. Driscoll (1982) Broaching - An Also, sensitivity to initial conditions which leads to
investigation into the loss of directional control in severe unpredictability.
following seas. Trans. RINA, 124.
22. R. Wahab R. & W.A. Swaan (1964) Course-keeping and
broaching of ships in following seas. J. Ship Res., 7/4.