The Damaged Ship, London, UK

CONTENTS

The Damaged Ship The Steepest Learning Curve Yet

D. Vassalos and A. Jasionowski, University of Strathclyde
L. Guarin, Safety at Sea Ltd

A Practical Guide To Damage Stability Assessment Regulation on Damage Stability

J Gullaksen, JG Consultant Engineers, Denmark

Consideration of Damage to Ships from Conceptual Design to Operation: The

Implications of Recent and Potential Future Regulations Regarding Application, Impact
and Education
K W Hutchinson, Babcock International Group, UK
A L Scott, Maritime and Coastguard Agency, UK
P N H Wright, M D Woodward and J Downes, Newcastle University, UK

MoD(UK) Salvage and Marine Operations Team Use and Development of Specialist
Tools and Emergency Preparedness to the Support of a UK Military Maritime Incident
S Quinn and N Hills MoD, UK

Case Study of Damage Stability Criteria of Merchant Vessels and Warships

R Perez and J M Riola, Escuela Tcnica Superior de Ingenieros Navales (Universidad
Politcnica de Madrid), Spain

The Application of High Pressure Water Mist as Part of a Holistic Fire Fighting System
S E Ratcliffe, Atkins, UK

Estimation of Orifice Flow Rates for Flooding of Damaged Ships

C D Wood, A J Sobey, D A Hudson and M Tan, University of Southampton, UK
P James, Lloyds Register, UK

Dealing with the Hinged Ship - A Review of the Analytical and Practical Issues Based
Upon The M.T. Elli Salvage
K Ellam, Swire Salvage (Pte) Ltd, Singapore,
C Moore, Herbert Engineering Corp., USA
R Tagg, Herbert Engineering Shanghai, China

The Rapid Reliability and Risk Assessment of Damaged Ship Structure

M Shahid, Binary Systems and Engineering (UK) Ltd, UK

Progressive Structural Failure and Residual Strength of Damaged Ships

S. Kwon, D. Vassalos and G. Mermiris, The Ship Stability Research Centre (SSRC),
Department of Naval Architecture and Marine Engineering, University of Strathclyde, UK

Global Wave Loads on Damaged Ship Structures: An Experimental Procedure and Some
Preliminary Results
D Fone, T W P Smith and K R Drake, University College London, UK
J Borg, Lloyds Register, London, UK
 The Damaged Ship, London, UK

Simulations of Motions of a Damaged Ship in Regular Waves

A C Gaillard, G X Wu and P Wrobel, University College London, UK

Selecting the Safest Option After Damage A Tool for Decision Support
A. Martin, QinetiQ Ship Structures Team

The Structural and Stability Assessment and Subsequent Recovery of a Damaged

Lifeboat
A Harman, N Chaplin, H Phillips and S Austen, Royal National Lifeboat Institution (RNLI),
UK

A GA Based Decision Support Tool for Stability and Structural Viability Under Damage
T Martins and Sousa Lobo, CINAV Portuguese Navy, Portugal
Refloating and Strength Assessment of Grounded Ships
P. Mangriotis and S. Pollard, London Offshore Consultants

Damage Decision Books for use by Masters on Vessels with no Computer Based Loading
System
M Simpson and I Wallbridge, Hart Fenton & Company Ltd, UK

Expanding the Scope of Rapid Response Damage Assessment (RRDA) A Classification

Society Perspective
G Wang, American Bureau of Shipping, USA

The Damaged Ship - MoD Salvage Response to HMS Endurance in The Magallan Straits
in December 2008
J R Ward, M Watts and D Price, Ministry of Defence, UK

Tolerable Safety of Damaged Naval Ships

S Marshall, Ship Stability Safety Regulator, Ministry of Defence (UK).
The Damaged Ship, London, UK

THE DAMAGED SHIP THE STEEPEST LEARNING CURVE YET

D Vassalos and A Jasionowski University of Strathclyde, UK
L Guarin, Safety at Sea Ltd, UK

SUMMARY

With artefacts on human endeavours at sea dated as far back as 6500 B.C., it is mind boggling to think that it was not
until 250 B.C. when the first recorded steps to establish the foundation of Naval Architecture, floatability and stability,
were made by Archimedes. It is even more astonishing that practical pertinence and function of these two very basic
principles remained dormant for nearly two millennia after this, before the first attempts to convey the meaning of
stability to men of practice took place in the 18th century by Hoste and Bouguer. Pertinent regulations, especially after
accidents involving water ingress and flooding (through subdivision and damage stability), were introduced even much
later. Notably, the first Merchant Shipping Act of 1854 addressing subdivision and leading eventually and after heavy
loss of life to the adoption of the first internationally agreed system of subdivision in SOLAS 1929 whilst the first
specific criterion on residual static stability standards at the 1960 SOLAS Convention. This tortoise pace gave way to
the steepest learning curve in the history of Naval Architecture with the introduction of the probabilistic damage stability
rules in the late sixties rules as an alternative to the deterministic requirements. Prompted and motivated to adopt a more
rational approach to damage stability, necessitated the development of appropriate methods, tools and techniques
capable of meaningfully addressing the physical phenomena involved. Within 50 years, this new impetus has climaxed
to the zero tolerance concept of Safe Return to Port in July 2009 and to an open proclamation by the Secretary of the
International Maritime Organisation (IMO) that deterministic requirements have no future.

The authors, having played a protagonist role in contemporary developments on damage stability and having been
involved closely in implementing such developments to the design of modern safety-critical ships, will attempt to
provide a complete (albeit succinct) account of the The Damaged Ship Learning Curve and demonstrate the
implications of the underlying developments on ship design and operation.

Keywords: damage stability and survivability fundamentals, past, present and future developments

1. INTRODUCTION
From a basic Naval Architecture perspective, the most
fundamental goal to be achieved is for a ship to remain
afloat and upright, especially so after an accident
involving water ingress and flooding. Regulations to
address the former focus on subdivision and the latter on
damage stability, even though more recent regulatory
instruments tend to cater for both issues concurrently.
The first Merchant Shipping Act of 1854 is the first
known legal requirement addressing safety at sea
concerning watertight bulkheads, leading eventually and
after heavy loss of life (notably the sinking of the Titanic
on 14 April 1912 and the loss of some 1,500 lives
providing the catalyst for the adoption on 20 January
1914 of the first International Convention for the Safety
Of Life At Sea), wars and other ills to the adoption of
the first internationally agreed system of subdivision in
SOLAS 1929. The first damage stability requirements,
on the other hand, were introduced following the 1948
SOLAS Convention and the first specific criterion on
residual stability standards at the 1960 SOLAS
Convention with the requirement for a minimum residual
GM of 0.05m. This represented an attempt to introduce a
margin to compensate for the upsetting environmental
forces. It is worth mentioning that a regulation on
"Watertight Integrity above the Margin Line" was also
introduced reflecting the general desire to do all that was
reasonably practical to ensure survival after severe
collision damage by taking all necessary measures to
limit the entry and spread of water above the bulkhead
deck. The first probabilistic damage stability rules for
passenger vessels, deriving from the work of Kurt
Wendel on Subdivision of Ships, [1] were introduced
in the late sixties as an alternative to the deterministic
requirements of SOLAS 60, in the belief that this
represented a more rational approach to addressing
damage stability. Evidence, however, of common
sense driving rule making for damage stability is very
scarce. With accidents providing the main motivation,
emphasis has primarily been placed on reducing
consequences, i.e., on cure rather than prevention. The
prevailing situation could be drastically improved
through understanding of the underlying mechanisms
leading to vessel loss and to identification of governing
design and operation parameters to target risk reduction
cost-effectively. This, in turn, necessitates the
development of appropriate methods, tools and
techniques capable of meaningfully addressing the
physical phenomena involved. Having said this, it was
not until the early 90s when dynamic stability pertaining
to ships in a damaged condition, was addressed by
simplified numerical models, [2]. The subject of dynamic
ship stability in waves with the hull breached received
much attention following the tragic accident of Estonia,
to the extent that led to a step change in the way damage
stability is being addressed, namely by assessing the
performance of a vessel in a given environment and
loading condition on the basis of first principles. In
parallel, motivated by the compelling need to understand

the impact of the then imminent introduction of
probabilistic damage stability regulations on the design
of cargo and passenger ships and the growing
appreciation of deeply embedded problems in both the
rules and the harmonisation process itself, an in-depth
evaluation and re-engineering of the whole probabilistic
framework was launched through the EC-funded project
HARDER, [3]. In this respect, HARDER became an
IMO vehicle carrying a major load of the rule
development process and fostering international
collaboration at its best a major factor contributing to
the eventual success in achieving harmonisation and in
proposing a workable framework for damage stability
calculations in IMO SLF 47. Deriving from
developments at fundamental and applied levels in
project HARDER as well as other EU projects such as
NEREUS [4], ROROPROB [5] and SAFEVSHIP [6] and
other international collaborative efforts (work by the
Stability in Waves Committee at the International
Towing Tank Conference from 1996 onwards, e.g., [7]),
a clearer understanding of damage stability started to
emerge together with a confidence in the available
knowledge and tools to address the subject effectively
and with sufficient engineering accuracy. All these
efforts provided the inspiration and the foundation for
SAFEDOR [8], which provided the opportunity for
consolidating contemporary developments on damage
survivability, thus rendering implementation possible
even at design concept level. The knowledge gained can
now be used to address critically all available regulatory
instruments and to foster new and better methodologies
to safeguard against known design deficiencies in the
first instance, until safer designs evolve to reflect this
knowledge, [9], [10], [11].
However, the cultural shock of adopting probabilistic
rules in the maritime industry has had a more profound
effect, the results of which are yet to be fully appreciated.
Surprisingly, the biggest influence so far is seen at the
birth place of prescription: The future is Risk-Based
was proclaimed at the International Maritime
Organization (IMO) post-HARDER (2002) and goal-
setting-performance-based approaches the new face of
safety. What is known as Safe Return to Port (SRtP) of
SOLAS 2009, enforceable on every passenger
newbuilding vessel and on special primrose ships over
120m in length and those carrying more than 60 non-
crew onboard, will challenge everything we know and do
about safety. The flood gate is now open! The new
regulations represent a step change away from the
current deterministic methods of assessing subdivision
and damage stability. The old concepts of floodable
length, criterion numeral, margin line, 1 and 2
compartment standards and the B/5 line are disappearing
whilst contemporary developments adopt a more holistic
approach to addressing The Damaged Ship. This
entails design and operational measures ranging from
consideration of all conceivable (statistical, experiential,
judgemental) damage scenarios to deal with subdivision,
dynamic damage stability in all feasible loading
The Damaged Ship, London, UK
conditions and environments, residual functionality of
essential systems onboard post-damage, evacuation and
rescue and decision support systems onboard. Moreover,
such considerations encompass the principal hazards
(flooding and fire) over the life-cycle of the vessel,
targeting cost-effective safety as a key design objective
and alongside other conventional design objectives.
Clearly, the obvious need for a holistic approach using
knowledge in all its forms became all too compelling,
driving industry and academia into an unprecedented
frenzy of activity and developments that could only
stop with a credible solution to the damage stability
problems, particularly for passenger ships. With
monumental effort over the past two decades, the quest
of this mountain is within reach as outlined in the
following, starting by providing a brief account of the
theoretical background and development of the requisite
first-principles tools on damage stability, leading to
contemporary developments and implications on ship
design and operation and finishing with an
implementation example in an attempt to demonstrate the
level of consolidation of all these developments.
2. THEORETICAL BACKGROUND
Even provision for initial static stability information,
based upon inclining experiments to measure the
metacentre height (GM), became mandatory for
passenger ships as late as 1931. It is not surprising,
therefore, to note that much more complex problems of
ship stability in the presence of wind and waves attracted
the interest of the profession only after mathematical
models and computer technology capable of handling the
necessary computations became available in the middle
of the 20th century. In particular, following the early
work of Newton, Euler, Froude, Stokes, Green, Rayleigh
and other advances of mostly fundamental sciences, a
new era of modern approaches to dynamic ship response
in random seas started with the advent of spectral
analysis, introduced by St Denis and Pierson in 1953.
This was followed by the first ship motion theories
suitable for numerical computations, introduced in the
50s, notably by Grim in 1953, Korvin-Kroukovsky and
Jacobs in 1957 and Cummings in 1962. Notable
advances contributing to developments of intact ship
motion prediction tools include: 2D free surface ship
hydrodynamics by Ursell in 1949, Kim in 1969, Frank in
1967; pioneering calculations of 3D hydrodynamics of a
semi-submerged heaving sphere by Havelock in 1955,
first successful applications of integral equation
techniques in 3D by Hess & Smith in 1958, application
of this method for arbitrary 3D bodies oscillating at the
free surface by Yeung in 1970; linear motion dynamics
using strip theory in 5 DOF with forward speed by
Salvesen at al. in 1970; other endless variants of strip
theory giving way to 3D panel methods at 1st, 2nd and
higher orders. The many techniques developed facilitated
almost routine analyses for many fluid flow and dynamic
The Damaged Ship, London, UK
phenomena pertinent to ship performance and safety, all
focusing on ships in intact condition.
It was not until the 1980s and 90s that eventually one of
the most difficult problems of dynamic stability
pertaining to ship in a damaged condition, was addressed
initially by simplified numerical modelling techniques,
dealing with water on deck developed by Dillingham in
1981 and the numerical model of damaged Ro-Ro vessel
dynamic stability and survivability, Vassalos and Turan,
[2], Jasionowski, [12].
Assessment of ship performance in terms of damage
survivability, however, is not a straightforward
undertaking, as in addition to the complexity of
predicting ship behaviour in waves utilising techniques
pertinent to intact ships, further intricate phenomena
arise with water ingress-egress through the ship hull and
the ensuing ship-floodwater interaction and water
sloshing, [13]. This, in turn, depending on compartment
geometry, dimensions and position with respect to the
axis of rotation, amount of water, and amplitude and
frequency of motion, [14], displays a behaviour ranging
from small-amplitude short waves formation and non-
linear standing waves to highly non-linear hydraulic
jumps or combinations of all these, [15]. The dynamic
pressures exerted on the compartment walls are also of
non-linear nature as they comprise both non-impulsive
loads related to fluid transfer as well as impulsive
localised loading. Such dynamic effects of fluid motion
on the ship response, and vice-versa, have been
extensively studied since the late 60s mostly from the
viewpoint of roll stabilising tanks, water trapped on deck,
LNG carriers and related problems where the amount of
fluid mass in the tank is constant. However, the problem
of a ship undergoing progressive flooding entails further
degrees of freedom and complexity arising from fluid
mass variation, which also renders all related processes
non-stationary, (see, for example, Figure 1).
Figure 1: Time series from survivability tests of PRR1
Ro-Ro vessel, capsize case, Hs=4.5m, [HARDER]
Published research on the subject exhibits tremendous
variety in levels of sophistication and type of approaches
used to solving these problems. Two classes of
approaches can be broadly distinguished: simplified
numerical methods based on rigid-body theory and using
a Bernoulli-based mechanism for modelling water
ingress-egress (elaborated more fully in Section 3 of this
paper) and techniques employing the latest advances in
Computational Fluid Dynamics (CFD). Studies on
coupled ship motion and water sloshing based on the
latter approach have been reported by Mikelis et al, [16],
Francescutto et al, [17], Bass et al, [18] and Daalen et al,
[19]. In these studies, the excited due to tank/ship
motion internal fluid behaviour is dealt with by coupling
the solution of RANS equations with the simultaneous
time-domain solution of equations of intact ship motions,
treating the fluid forces as external input. Further, de
Veer et al, ref. [20], presented an attempt to predict, in a
similar manner, effects of water ingress with the rate of
flooding itself estimated from Bernoullis equation. In
addition to water sloshing coupled to a 6-DOF ship
motion prediction model, Woodburn et al, [21], led the
way to representing water ingress/egress and damaged
ship dynamics in a more sophisticated (albeit still
simplified) manner, allowing for direct coupling between
external and internal fluid domains via a water column,
Figure 2.
Figure 2: Coupling between sea-ship-floodwater
responses, [21]
These techniques, although addressing the problem of
intact and damaged ship dynamics with water sloshing at
the most fundamental of levels, are plagued with
practical solution setbacks, deriving from two reasons:
the very large fluid domains required and the presence of
free surface. The applied numerical solution schemes
proposed, such as the VOF method, suffer from
notorious inability to conserve the fluid mass with time
marching, due to fluid diffusion near the free surface,
which is severe especially in the presence of wave fields.
Highly refined space discretisation must be used, which
increases grid density thus rendering computation
excessive and unaffordable. Additionally, for the case of
bodies undergoing motions, the grids must be
instantaneously adapted to the new fluid geometry,
which is a non-trivial numerical problem, adding to the
complexity of using even the most advanced general-
purpose CFD tools available today.
Having said this, there does not seem to be much of a
doubt that in the foreseeable future the aforementioned
problems will be overcome. The work of the CFD Group
at the Ship Stability Research Centre (SSRC), for
instance, using Fluid Particle Methods (FPM Figure 3),
VOF and other fundamental approaches [22], [23]. [24],
 The Damaged Ship, London, UK

[25], [28] is likely to become routine procedure for Naval
Architects.
Figure 3: A SPH Approach to Ship-Floodwater
Dynamics, [24]
As is the general consensus at present, however, these
methods require excessive computational as well as
expert effort, preventing their methodological application
to studies on dynamic ship stability. It is envisaged that,
presently, the use of these kind of tools will be applied to
address many basic problems, such as higher order
effects of waves diffraction upon encountering a ship
with a breached hull, highly turbulent (rotational) and
locally 3-dimensional flows at the damage opening or
non-linear floodwater behaviour inside the ship
compartments coupled with effects of instantaneous
water ingress/egress on ship hydrodynamics. More
methodological treatment of such tools, leading to
knowledge intensive models (for example response
level, e.g. Vassalos et al, [28], [29], de Kat, [30], or to be
moving in accordance with some basic motion
mechanism, e.g. Papanikolaou et al, [31], [32] and
Jasionowski, [12]. This section offers a brief overview of
the mathematical model deriving from this approach,
including the generalised ship motions and floodwater
motion mechanism with some discussion on validation
and benchmark studies and focus on issues remaining
unresolved.
3.1 SHIP MOTIONS
Equations describing damaged ship behaviour are
derived from fundamental motion principles: the law of
conservation of linear and angular momentum. The law
normally applied to rigid bodies, is here also extended to
the internal fluid mass, resolved in a body-fixed system
of reference, as shown in Figure 4. Rigorous derivation
leads to a set of 6 scalar equations for linear and angular
motions. Of these, the three equations for angular
motions are presented here in vector form (1). For a
detailed explanation of all the relevant terms and of the
model itself, see Jasionowski, [2].
z

surfaces) is already taking place; refer for example to y'

EC-funded IP project VIRTUE [27]. Gs

Such numerical treatment of damage stability is deemed G

rGs G
ri
to evolve into a viable alternative to physical model Z x'

testing, once fuller understanding of the overly complex Y

O
processes involved in ship foundering is gained and
reliable methods for modelling this become adequately
X
validated and accessible.

The next session sets to discuss various aspects of the Figure 4: Coordinate system fixed to the centre of
modelling of the dynamic behaviour of a damaged ship, gravity of the intact vessel
by summarising the research efforts of the last 15 years,
with a threefold purpose: Firstly, to illustrate the d G G d G
suitability of simplified techniques for representing the ( I 's + I 'w ) '+ M w r 'w v 'Gs +
dt dt
behaviour of damaged ships; secondly, to demonstrate G G G
the practical use of this kind of numerical models; + M w ( ' r 'w ) v 'w +
thirdly, to highlight the difficulties and knowledge gaps G d G G G G
on various aspects remaining unresolved to date and + M w r 'w v 'w + ' ( v 'Gs + v 'w ) +
therefore underline the need for on-going concerted dt
research efforts on damage ship stability. d G G G
+ M w r 'w ( v 'Gs + v 'w ) +
dt
3. DAMAGE SHIP STABILITY d G G G G
+ I 'w '+ ' ( I 's + I 'w ) ' = M 'Gs
FUNDAMENTALS dt (1)
G
The approach that has attracted considerable research The right hand side of the equation, M 'Gs , and
G
interest over the recent past is based on rigid body the respective force vector F of the rectilinear
dynamics, aiming at achieving a balance between motions, represent all the external forces and
simplicity and meaningful representation of physics. moments acting on the vessel, expressed in a
Here, the mass of floodwater in the ship hull is treated as body-fixed system of reference, Gsxyz, located
a pendulum attached to the ship, with its mass located at at the ship centre of mass. These
the centre of the fluid volume, which in turn is found forces/moments are predicted with
from the intersection between tank geometry and fluid conventional Naval Architecture methods. The
free surface, the latter assumed to be flat. The fluid free Froude-Krylov and restoring forces and
surface is either assumed to remain parallel to the sea moments are integrated up to the instantaneous
The Damaged Ship, London, UK

wave elevation; the radiation and diffraction natural frequency, remains a function of roll amplitude,
forces and moments are derived from linear which cannot be known a priori and hence not suitable
potential theory and expressed in time domain for application to time-domain simulation in random
using convolution and spectral techniques, seas. In this respect, an engineering approximation has
respectively. The hull asymmetry due to been proposed in [2], whereby a discrete piece-wise
floodwater is taken into account by a constant treatment of the linearised coefficient is used
database approach, whereby the with the coefficient evaluated at the wave spectrum peak
hydrodynamic coefficients are predicted frequency and for an amplitude corresponding to the
beforehand, and then interpolated during the amplitude of the last half-roll cycle. In this approach the
simulation, as illustrated in 5. viscous roll damping will vary with time, constantly
adjusting to roll amplitude.
PRR1, Fn=0.0
A44 PROTEUS3 Vugts experiments
0.1
Lewis mapping
A44 Numerical (Ha)
0.09 0.04 N-S solution
A44, -1m sink, -20deg heel, -1 deg trim (Ha)
0.08
A44 PROTEUS3, -1m sink, -20deg heel, -1 deg trim
0.07 0.03

0.06

0.05 0.02
b44 B

A B2 2 g
0.04
0.01
0.03

0.02
0
0.01 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

0
1 B
0 1 2 3 4 5 6 7 8 9 10
2 2 g
B [-]

Figure 6: Damping for 2D cross-section; comparison

Figure 5: Dimensionless roll added moment for 3D
between potential flow, experiment and a RANS solver
PRR1 hull form, [HARDER]; effect of 1m sinkage, -
results (ICEM-COMET, unstructured collocated finite
20deg heel and 1deg (aft) trim, derived by strip theory
volume technique, a variation of SIMPLE algorithm for
and 3D panel methods at Fn=0.0
pressure correction, High Resolution Interface Capturing,
HRIC for the free surface tracking).
The second order drift forces, wind and current effects
and other forces of viscous origin are also catered for, at
Some successful attempts of solving this fundamental
present based on parametric formulations. Naturally the
problem with RANS solvers allowing prediction
gravity force and moment vectors correspond to ship and
accuracy as demonstrated in Figure 6 have been reported
floodwater weights.
in [34], which is indicative of what is to be expected in
G G G G G G the near future regarding modelling of ship
F , M 'Gs = FGravity + FF K ,Re storing + FRadiation + FDiffraction hydrodynamics with viscous effects taken into account.
G G
+ FManouvring , Rudder + FDrift ,Current , Wind, Viscous The whole system (1), after re-arranging into a matrix
form as a set of twelve differential equations of the first
Where: order, is solved for position in space of the centre of
G G G G
FRe storing FF K Direct integration of static gravity of the intact ship rGs = vGs dt and three
pressures on actual geometry rotations through a 4th order Runge-Kutta-Feldberg
G integration scheme with variable step size.
FRadiation Convolution techniques
G
FDiffraction Spectral techniques
G 3.2 INTERNAL SLOSHING MODEL
FManoeuvring , Rudder Empirical formulae
G Still undetermined in equations (1), are the relevant
FDrift ,Current , Wind, Viscous Empirical formulae
vectors for floodwater location, velocity and
G G d G
acceleration, r 'w , v 'w and v 'w , respectively. These
A correction for viscous effects on roll motion is applied dt
based on an established empirical method proposed in are the quantities that must be derived from a model
[33], where the viscous damping moment is divided into representing the sloshing water phenomenon. When CFD
several components: friction, eddy shedding, lift, wave techniques are used, these vectors and related forces and
and bilge keel and the total force is obtained by a moments can be derived from integration of pressure due
superposition of all these components. However, the to fluid motion. Here, however, Figure 7, the
proposed method, representing the non-linear viscous simplification mentioned in the foregoing, is adopted.
damping as an equivalent linear coefficient at the roll
 The Damaged Ship, London, UK

d G G G G G
dt r 'w = v 'w ( v 'w n ' ) n '
(1)
d vG ' = aG ' ( aG ' nG ') nG '
dt w f f

S
w G
n

af (af n)n
G G G
(aG f n) n
G G

z
Figure 8: Description of geometric constraints for
G motion of centre of buoyancy of floodwater; symmetrical
af and asymmetrical compartments.
y Gs

Figure 7: Fluid particle w (centre of buoyancy) in Having determined fluid motion, the forces and moments
G due to its displacement can be calculated. For
acceleration field a f moving on the potential surface S;
demonstration purposes, the moment vector extracted
all vectors are resolved in the Gsxyz system of reference. from equation (1) is used and presented in the form of
equation (5), where three components are distinguished
to represent inertial, gravity and non-linear effects and
A lumped mass model, the initial concept was presented presented in equations (6), (7) and (8), respectively. Note
by Papanikolaou et al in [32], has been developed here that the fluid inertia matrix, I 'w , contains only the
treating the floodwater as a free point-mass moving due G
to the acceleration field and restrained geometrically by inertia of a single point-mass located at a position r 'w in
predetermined potential surfaces of the centre of the ship-fixed system of reference at Gs. Since the mass
buoyancy for given amounts of floodwater, FMPS (Free is constant, the terms containing the time derivative of
Mass in Potential Surface), as shown in Figure 7 and the mass, will vanish.
Figure 8.. This model, derived from simple rigid body
motion consideration, similar to that leading to equations G G G G
M 'wat = M 'I + M 'g + M 'N (5)
(1), is presented as a set of equations (2), with graphical
explanation provided in Figure 7.
Where,
The total forcing acceleration vector is:
G G G G G G G d G
a ' f = g ' a 's 2 ' v 'w * v 'w (2) M ' I = I 'w ' (6)
dt
G G G G
Where as , see equation (4), is a ship motion-related M ' g = M w r 'w g ' (7)
acceleration vector expressed in a body-fixed system of
G G G G
reference. M 'N = M w ( ' r 'w ) v 'w +
G d G G G
G d G d G G G G G G
a 'S = v 'Gs + ' r 'w + ' ( v 'Gs + ' r 'w ) (4) + M w r 'w v 'w + ' ( v 'w ) +
dt (8)
dt dt
d G G G
G I 'w '+ ' ( I 'w ) '
n is the instantaneous normal vector to the potential dt
surface of floodwater motion, determined from a damage
compartment geometry database. Note that the vector is a
G 3.3 FLOODWATER INGRESS/EGRESS
function of rw and volume of the fluid. Finally, * is an
artificial coefficient introduced to represent damping of
floodwater motion. This coefficient is assigned an ad hoc Water ingress/egress affects the dynamics of a damaged
value derived from experimental data using a simple ship in two distinct ways: firstly, the influence on ship
box-shaped compartment. With the geometric hydrodynamics and through coupling of floodwater
information concerning the tank stored in a database, dynamics with ship dynamics the overall dynamic
such as shown in 8, the model is complete. Equation (2) behaviour; secondly, and the most considerable
is set up for each flooded compartment within the ship consequence of flooding in a typical Ro-Ro scenario, is
and solved simultaneously with the equations for ship vessel capsize or sinkage as a result of water
motion. accumulation, Figure 9.
The Damaged Ship, London, UK

ROLL MOTION takes for a vessel to capsize from the instant her hull is
breached, is paramount. Therefore, water ingress/egress
70
60 will have to be considered as one of the major factors,
50 the accuracy of prediction of which will determine the
ANGLE (deg)

40
30 robustness of the whole damaged ship dynamics model.
20
10
Again CFD RANS solvers would provide the accuracy
0 warranted.
-10 0 100 200 300 400 500 600 700 800

TIME (sec)
3.4 HYDRODYNAMICS OF A DAMAGED SHIP
WATER ON VEHICLE DECK
How to meaningfully model the influence of floodwater
7000
6000 on ship hydrodynamics and the coupling of floodwater
WATER (tonnes)

5000
4000
dynamics into ship dynamics can only be speculated at
3000 present. In the interim, experimental studies are being
2000
1000
undertaken using scale models of typical modern
0 passenger Ro-Ro vessel [35], aiming to qualify and
-1000 0 100 200 300 400 500 600 700 800
quantify the level of contribution of hull damage and
TIME (sec)
ensuing flooding to the overall ship hydrodynamic
properties.
Figure 9: Typical Ro-Ro vessel capsize mode with water
on deck 800

Therefore modelling this element should be performed to 700 Experimental measurment

the highest accuracy attainable. However, knowledge on Theoretical

600
the mechanics of water flow through the damage opening
SUM( dQ/dt ) [m3/s] kkkkk

of a moving ship in the presence of incident, diffracted 500

and radiated waves, is virtually non-existent. Therefore, 400

the fundamental Bernoulli fluid momentum equation is
used most commonly (as depicted in Figure 10), with all 300

the unknown flow physics encapsulated in a single 200

averaged flow loss coefficient K derived from 100

experiments.
0
0 500 1000 1500 2000 2500
Time [s]

Figure 11: Comparison between experimental

dh measurements and theoretical predictions of water
(t ) accumulation

These entail forced oscillations of the model in damaged

conditions in various degrees of freedom with the use of
purposely-designed mechanical forcing systems, such as
shown in Figure 12.

Figure 10: Simplified modelling of water ingress on a

Ro-Ro car deck

Q = K A 2 g dh dt (9)
t
Where,
K Flow loss coefficient
A Submerged area of damage opening

Although this model can prove accurate in many cases,

as demonstrated in Figure 11, by no means does it
resolve the problem of not knowing the instantaneous
water ingress/egress in case of marginal survivability.
This has direct impact on the accuracy of estimations of
how long it takes a vessel to sustain the action of waves
whilst undergoing progressive flooding that leads to
eventual capsize (time to capsize). Considering the
subtlety of new regulations for SRtP, the accuracy of Figure 12: Cylindrical, freely-floating, section of a Ro-
predictions of ship survivability including the time it Ro vessel subjected to forced oscillations in calm water

The derived results have been compared with predictions
using well-established numerical techniques leading to
the general conclusion that whilst the effect of hull
damage on hydrodynamic forces in heave is negligible,
the corresponding effect in roll is considerable, as shown
in Figure 13.
Figure 13: Measured characteristics (roll damping
moment and phase lag) in intact and damaged conditions
Estimation of the coefficients was carried out based on
time series of the force signals and a least squares fit
assuming that the two pertinent moments (added moment
and damping) are orthogonal and independent. It is
worth emphasising that it is not clear to what extent, if at
all, such orthogonal decomposition of these effects is
valid, as the water ingress and egress and ensuing effects
on ship hydrodynamics and floodwater dynamics are of
strongly non-linear character and, therefore, could be
neither in phase with velocity nor acceleration of ship
motion. The exact nature of the effects of flooding is not
known. The main aim of this investigation was to
qualitatively assess the nature and scale of the effects of
hull damage and ensuing flooding on ship
hydrodynamics and motions, with further rigorous study
to quantify these effects left for future work.
Nevertheless, the work carried out thus far clearly
demonstrates that that floodwater motions have a major
impact on damping. As shown in Figure 13, with the
excitation frequency approaching sloshing natural
frequency, the hull and floodwater oscillations are out of
The Damaged Ship, London, UK
phase (and in the presence of the damage opening a
significant drift, 1.3 m/s full scale, is observed), which
results in a large dip in the roll motion phase
characteristics. Bearing in mind that the amplitude
response is virtually unaffected, this causes a substantial
increase of damping (note also that the frequency shift in
phase angle between intact and damaged condition is a
consequence of the difference in the dry hull KG
resulting from maintaining the same draught in both
conditions). In fact, the tenfold increase of the roll
damping moment cannot be solely associated with the
inappropriate mathematical model used in the analysis
of the experimental results.
4. CONTEMPORARY DEVELOPMENTS
With first-principles tools still struggling to deliver the
requisite accuracy and level of detail within a time frame
that can be routinely utilised by industry, contemporary
regulatory developments are a step ahead for the first
time in Naval Architecture history. This necessitates
concerted effort at global level to ensure safe transition
from deterministic to performance-based safety. More
specifically, in May 2000, the IMO Secretary-General
called for a critical review of the safety of large
passenger ships noting that "what merits due
consideration is whether SOLAS requirements, several of
which were drafted before some of these large ships were
built, duly address all the safety aspects of their operation
in particular, in emergency situations. This visionary
prompt led the IMO Maritime Safety Committee (MSC)
to adopt a new philosophy and a working approach for
developing safety standards for passenger ships. In this
approach, illustrated in Figure 14 (SLF 47/48), modern
safety expectations are expressed as a set of specific
safety goals and objectives, addressing design
(prevention), operation (mitigation) and decision making
in emergency situations with an overarching safety goal,
commensurate with no loss of human life due to ship
related accidents. The term Safe Return to Port (SRtP)
has been widely adopted in discussing this framework,
which addresses all the basic elements pre-requisite to
quantifying the safety level (life-cycle risk) of a ship at
sea.
IMO (SLF 47/48) Passenger Ship Safety
Casualty Threshold Return to Port
100% vessel survivability
(indefinitely)
(RTP)
Abandon Ship
Flooding
100% vessel survivability
Fire for a specified period of time [3h]
Ship functions / systems availability after a casualty
Figure 14: The IMO Framework for Passenger Ship
Safety
The Damaged Ship, London, UK

More specifically the following elements are explicitly
addressed:
Prevention/Protection: Emphasis must be placed on
preventing the casualty from happening in the first
place as well as on safeguards (in-built safety) to limit
explained in [36] and outlining here in Section 5. An
evaluation framework, already being applied to the
design of cruise/RoPax ships, is shown in Figure 15 next.
Safe Return to Port / Casualty Threshold
Accident Causality Analysis Consequence Analysis Mitigation Analysis

consequences. Emphasis must also be paid to the Collision

Collision
Statistics
HAZID
Modelling
Statutory assessment
SOLAS (A-index)
First Principles Analysis
-transient flooding
-cross flooding
Time to flood/
capsize Systems
Availability
Evacuation
Analysis

many international regulations addressing prevention -progressive flooding

of accidents: namely COLREGS, Load Line, STCW Grounding/

Grounding/
Stranding
Stranding
Statistics
HAZID
Modelling
First Principles Analysis
-transient flooding
-cross flooding
Systems
Availability
Evacuation
Analysis

and ILO 68-152, among others.

-progressive flooding

Timeline Development: The new focus here is on the

Statistics Statutory assessment First Principles Analysis Time to untenable
Fire HAZID SOLAS ChII -fire/smoke propagation conditions Systems Evacuation
Fire Modelling Availability Analysis

timeline development of different events. For the first

Safety Level (Total Risk)
time in the history of rule-making, it is not only
important to know whether a vessel will survive a Figure 15: Risk-Based Design Implementation (Safety
given casualty in a given loading condition and Level)
operating environment but also the time the vessel will
remain habitable and the time it takes for safe and With the focus in this paper on damage stability and
orderly abandonment and for recovery of the people survivability analysis, the best demonstration on
onboard. consolidation of all contemporary developments and
Casualty Threshold: This advocates the fact that the knowledge (regulatory, scientific, experiential) thought
ship should be designed for improved survivability so possible, is to consider a design implementation on a
that, in the event of a casualty, persons can stay safely very large passenger ship, using first-principles tools for
on board as the ship proceeds to port. In this respect all the requisite analyses. This is outlined in Section 5
and for design purposes (only), a casualty threshold next. More details can be found in [39], [40]. [41].
needs to be defined whereby a ship suffering a
casualty below the defined threshold is expected to 5. DESIGN IMPLEMENTATION
stay upright and afloat and be habitable for as long as
necessary [currently it constitutes part of the design The specific design highlighted here relates to the largest
work to determine this value rationally, as it greatly cruise ship ever built (the second ship of the series was
influences the design arrangements (capacities)] in launched in November of last year) during the concept
order to return to port under its own power or wait for development phase, under the name Project Genesis,
assistance. having the general particulars depicted in Figure 16 and
Emergency Systems Availability / Evacuation and Table 1 below.
Rescue: Should a casualty threshold be exceeded the
ship must remain stable and afloat for sufficiently long
time [3 hours recommended] to allow safe and orderly
evacuation (assembly, disembarkation and
abandoning) of passengers and crew. Emergency
systems availability to perform all requisite functions
in any of the scenarios considered is, therefore, Figure 16: Project Genesis
implicit in the framework. In addition, the ship should
be crewed, equipped and have arrangements in place Length 361 m
to ensure the health, safety, medical care and security
Breadth 47 m
of persons onboard in the area of operation, taking
Draught 9.15 m
into account climatic conditions and the availability of
SAR functions and until more specialised assistance is Gross Tonnage 225,000
available. Air Draught 72 - 65 m
Number of Guests 5,400
Considering the above, it is worth emphasising that none Number of Crew 2,166
of the questions arising (survival time?; functional LSA Capacity 8,460
availability post-casualty?; time needed for Table 1: Project Genesis, Main particulars
abandonment?) can be addressed in terms of rule
compliance. Nonetheless, achievement of these goals in As this vessel represents a step change in the size of
the proposed holistic, goal-based and proactive approach mega-ships, including some unique, innovative features,
would ensure safety of human life commensurate with focusing on safety and adopting a performance-based
the safety expectations of today, by implicitly addressing (risk-based) design methodology came naturally; in fact
all key elements of risk, for total risk (Safety Level) the initiative was entirely that of the owner. The task in
estimation and for direct use in Risk-Based Design, as hand was no less than ensuring that Genesis is not only
 The Damaged Ship, London, UK

the largest ship ever built but also the safest and do so
during the concept design stage commensurate with all Average historical
other design goals and functional requirements. The Principal hazards, frequency of
j
hz j
framework adopted to achieve this is the IMO framework occurrence, frhz ( hz j )
for passenger ship safety as depicted in Figure 15 and is
briefly described in the following. Flooding due to 1.48E-3, (cruise
1
collision ships), [37]
5.1 TOTAL RISK (SAFETY LEVEL) 0.92E-2, (cruise
2 Fire
ships), [37]
A common way of presenting graphically the chance of a Intact Stability
3 .......................
loss (risk) in terms of fatalities is by using the so-called Loss
F-N diagram, the plot of cumulative frequency of N or 4 Systems Failure ........................
more fatalities together with related criteria, [37], Figure
17. In addition, some form of aggregate information is Table 2: Principal hazards
used, such as the expected number of fatalities E ( N ) ,
often referred to as potential loss of life, PLL. With passenger ships, flooding- and fire-related scenarios
comprise over 90% of the risk (regarding loss of life) and
1E+00 almost 100% of all the events leading to decisions to
abandon ship. Therefore, it would be possible to
Frequency of N or more fatalities per ship year

1E-01
estimate the total risk (safety level) of a passenger vessel
by addressing these two principal hazards in a consistent
1E-02
manner and framework, allowing for their contribution to
risk to be formally combined as implied in (12).
Deriving from this, the following specific issues had to
1E-03
be addressed of which only the first is considered in the
paper.
1E-04

(a) Flooding survivability analysis (collision and

1E-05
1 10 100 1000 10000
grounding)
Fatalities [N] (b) Fire safety analysis
(c) Post-accident systems availability analysis
Figure 17: FSA Cruise Ships Societal Risk (d) Evacuation and rescue analysis

5.1(a) Risk Model 5.1(b) Flooding Survivability Analysis

N max To undertake a complete flooding (damage) survivability

Risk PLL E ( N ) F (i )
i =1
N (10) analysis, involving all the elements portrayed in the
framework of Figure 15, a complex geometric model is
Where N max the maximum number of persons onboard needed, which in this case comprised 717 compartments
and the FN curve is given as: with 1,160 openings, shown here in Figure 18.
N max
FN ( N ) = fr ( i )
i=N
N (11)

The frequency frN ( N ) of occurrence of exactly N

fatalities per ship per year is modelled as follows:
nhz
frN ( N ) = frhz ( hz j ) prN N hz j
j =1
( ) (12)

Where, nhz is the number of loss scenarios considered,

and hz j represents a loss scenario, pertinent to any of the
principal hazards. Furthermore, frhz ( hz j ) is the
Figure 18: Flooding Survivability Analysis Model for
frequency of occurrence of scenario hz j per ship year, Project Genesis, [39], [40]. [41]
and prN N hz j ( ) is the probability of occurrence of Considering all the elements involved, listed here below,
exactly N fatalities, given that loss scenario hz j has is indeed a very comprehensive undertaking but the
occurred. As shown in Table 2, estimates for the latter ensuing benefits could also be immense.
for flooding- and fire-related hazards have been derived
based on statistics. (i) Statutory Assessment
The Damaged Ship, London, UK

Compliance with SOLAS 2009 (probabilistic double hull, and position of staircases, lifts and escapes.
rules) Using the Attained Subdivision Index, payload capacity,
Optimisation of watertight subdivision steel weight and other regulatory requirements as typical
(ii) Flooding Risk Analysis objectives/constraints, the optimisation problem outcome
Frequency typically includes: reduced number of bulkheads,
Consequences reduced deck heights, reduced void volume, reduced
(iii) Time to Capsize number of escape ways and required staircases, reduced
Analytical and performance-based approaches steel weight, reduced complexity in tank arrangements,
Vulnerability assessment (as designed / as increased crew and service areas, improved functionality
operated) and, if required, improved Attained Subdivision Index.
(iv) Time to Abandon Ship In order to make the process effective, participation by
Assembly and evacuation performance all decision-makers (designer, owner and yard) is
(v) Evaluation of casualty threshold / return to port / essential to properly define the optimisation variables,
safety level objectives and constraints as early as possible in the
Probabilistic approach; link to systems
availability post-casualty design stage. Using this approach, known as platform
optimisation, high survivability internal ship layouts have
(vi) Evaluation of casualty threshold / return to port / been developed, without deviating much from the current
safety level SOLAS practice, thus making it easier for ship designers
Probabilistic approach; link to systems to relate to the proposed procedure, see for example,
availability post-casualty Figure 20.
(i) Statutory Assessment
0.955

0.95

With emphasis now ob goal-setting regulations, the 0.945

dilemma of prescriptive SOLAS-minded designers, 0.94

illustrated in Figure 19, in the simplest of levels, must be

0.935

0.93

overcome. It is obvious that internal subdivision 0.925

arrangement is a key issue affecting ship performance, 0.92

functionality and safety, all of which have to date been 0.915

0.91

catered for through the provision of rules and regulations 0 200 400 600
Feasible Designs
800
Pareto-optimal Designs
1000 1200

that reflect, in essence, codification of best practice.
Throwing this away and leaving on the table a blank
sheet, makes ship subdivision a very difficult problem
indeed. This was essentially the problem addressed in the
EU project ROROPROB, [5]. Building on the
understanding of Index A as outlined in [9] [11],
affords a straightforward way of determining the risk
profile of a vessel for all pertinent damage scenarios at
an early design stage and hence devise an effective
means of risk reduction by focusing primarily on the high
risk scenarios.
Aft peak bulkhead Machinery space bulkhead
Figure 20: Platform Optimisation Design Variants
(ii) Flooding Risk Analysis
Frequency Analysis ( frhz ( hz1 ) : Even though analysis
targeted both collision and grounding related flooding,
only collision is addressed here to allow comparisons
between Project Genesis and the rest of the cruise fleet.
As records of 111 ship years of statistics, obtained from
the owner, showed zero occurrences of flooding
incidents, the frequency of 1.48E-3per ship year (1 event
Collision bulkhead
every 571 ship years), deriving from statistics of the
existing cruise ships (Table 2) was used instead.

Consequence Analysis prN ( N hz1 ) : The comprehensive

New requirements for
double bottom
Minor damage concept (still deterministic) for passenger
vessels, but no specific requirements on location of watertight
risk model described in [36] requires two parameters to
subdivision. Required index to be met be estimated: the first is the time required for orderly
evacuation of passengers and crew in any given scenario,
A>R derived from numerical simulations using advanced
evacuation simulation software, [38]; the second is the
Figure 19: Largely Unguided Subdivision time to capsize/sink, which is evaluated by sampling the
(Probabilistic Rules) random variables comprising loading conditions, sea
states and damage characteristics (location, length,
The optimisation process typically produces several height, penetration according to the damage statistics
hundred design alternatives depending on the complexity adopted in the probabilistic rules in SOLAS 2009) using
of the ships layout and the number of variables. Typical Monte Carlo sampling. Each damage scenario is
variables of the optimisation problem include: type of simulated using explicit dynamic flooding simulation by
subdivision, number, location and height of watertight PROTEUS3, [12]. The investigation in Project Genesis
bulkheads, deck heights, tank arrangement, casings, involved a case by case explicit dynamic flooding
 The Damaged Ship, London, UK

simulation accounting for transient- cross- and 500 scenarios

progressive-flooding, impact of multi-free surfaces,
watertight and semi-watertight doors (see e.g., Figure 21
and Figure 22.

SWT doors D2 holding

windows under water Figure 22: Monte Carlo Simulation Set up Collision,
[39], [40]. [41]
1.E+00

Figure 21: Time-Domain Simulation of the Flooding 1.E-01

Frequency of N or more fatalities

Process (windows and SWT doors), [39], [40]. [41] 1.E-02

FSA Cruise Ships (DNV, SAFEDOR)

500 scenarios were used resulting in an absolute 1.E-03

sampling error for the cumulative probability of time to 1.E-04

capsize of the order of 4%-5%). A typical set up of Genesis

1.E-05
Monte Carlo simulations is shown in Figure 23 (generic)
and Figure 24 for collision studies. 1.E-06
1 10 100 1000 1000
N Fatalities

Figure 23: Societal Risk Collision Accidents, [39],

[40]. [41]

6. ONBOARD DECISION SUPPORT (DSS)

D3

D2
Having achieved the goal of designing a safe ship cost-
effectively and to go beyond all new and emerging safety
D1 requirements by utilising all available knowledge and
technology, the question that came naturally to fore
whether this extensive knowledge acquired during the
Figure 22: Time-Domain Simulation of the Flooding design phase could be used to manage operational /
Process (various openings), [39], [40]. [41] accidental risk. More importantly, in the knowledge that
passenger ships are operated in a way that may
A comprehensive experimental programme was also set compromise their watertight integrity (legislation on the
up to verify the numerical simulations, offering watertight doors issue has yet to be approved by MSC
corroborative evidence and hence confidence in the following recommendations by DE 54 in its last session),
derived results, which are presented in Figure 25 as an F- the goal post could be set even higher, namely to target
N curve together with results from the FSA on cruise optimum balance between safety and operational
ships [37]. The results clearly demonstrate the superior efficiency, tackling such questions and concerns led to
flooding survivability characteristics of Project Genesis. the development of iStand, an onboard Decision Support
System installed already on the first ship of the Genesis
Damage case Outcome Implication
Water Case i=1 t(i)
series with the following general features in addition to
ingress? Case i=2 t(2) Vessel unable to survive
for 3h
being a standard onboard loading computer:
yes Numerical
Case i=k simulations t(k)
Vessel survives for
at least 3h (t)
Real time sensors and hardware integration (link to
Collision Case i=342
Model
t(342)
ships SMS): tank levels, draughts, door states, water
tests
ingress alarms, wind and waves
no Minor incident Vulnerability log: global and local ship vulnerability
Performance-
Performance-based
evaluation and verification t = time to capsize
to flooding
Criticality assessment: survival time, escape and
Figure 21: Monte Carlo Simulation Scheme Collision evacuation time (crises management)
The Damaged Ship, London, UK
Corrective action search: evaluation of the impact of
corrective actions. Ballast system availability.
Essential systems availability post-flooding
(verification of compliance of SRtP requirements)
The DSS developed comprises a very powerful computer
with a massive database encapsulating all the features of
SRtP and linked to sensors capable of monitoring all the
elements that could affect ship vulnerability
(vulnerability to fire is currently undergoing
development). As such, it has become clear that in
addition to offering unique capabilities as a decision
support tool for crisis management and emergency
response the DSS potential uses include design,
operation (ballasting/de-ballasting, advanced
maintenance of onboard systems, management of
watertight integrity) and what is very important with
complex ships, crew training.
7. CONCLUDING REMARKS
Based on the work presented in the foregoing, the
following concluding remarks can be drawn:
A painstaking evolutionary development in the
subject of ship stability is giving way to
unprecedented scientific and technological changes
at an ever increasing pace which, despite all the
simplifying assumptions still embedded in the
available tool-set, facilitate step changes in
dealing with ship safety. However, due care is now
required to ensure a smooth transition to managing
this change and the emerging complexity.
With this place, it is hoped that robustness and
reliability of these methods and tools will help
promote their routine use in every-day design and
operation practice within the near future.
The probabilistic framework of the new harmonised
rules for damage stability calculations offer
flexibility and added degrees of freedom for
designers to enhance safety cost-effectively both in
targeting statutory compliance as well as pursuing
specific safety objectives in the strife of the maritime
industry to embrace innovation as a means of
gaining and sustaining competitive advantage.
With a clear trend towards probabilistic and risk-
based frameworks to addressing ship safety in a
holistic manner, it is important to base such
developments on clear understanding of the
underlying principles and of the intention of the
ensuing rules and regulations and/or criteria.
Despite problems faced by the industry to cross the
bridge from Rules-Based to Goals-Based (Risk-
Based) Design, this paper demonstrates that pre-
requisite scientific and technological developments
are in hand for Risk-Based Design to be fully
implemented in the maritime industry.
The need, however, to inculcate all major
stakeholders in these new developments must remain
a priority and clear targets set to facilitate the
transition from prescriptive to goal-setting
regulations, design and operations.
8. ACKNOWLEDGMENTS
The support received over the years by the European
Commission in undertaking part of the research work
presented here is gratefully acknowledged. The authors
would also like to express their appreciation and sincere
thanks to the maritime industry and, in particular, to
RCCL for offering them the unique opportunity being
part of the design team in Project Genesis.
9. REFERENCES
[1] Wendel, K: Subdivision of Ships, Diamond
Jubilee International Meeting, New York, June
1968, pp 12-1 to 12-21.
[2] Vassalos, D and Turan, O: Development of
Survival Criteria for Ro-Ro Passenger Ships - A
Theoretical Approach". Final Report on the S0T
Ro-Ro Damage Stability Programme, University
of Strathclyde, December 1992.
[3] HARDER, Harmonisation of Rules and Design
Rationale: EC Contact No. GDRB-CT-1998-
00028, Final Technical Report, July 2003
[4] NEREUS, First-Principles Design for Damage
Resistance against Capsize: EC Project
CONTRACT No. G3RD-CT 1999-00029, 1999-
2002.
[5] ROROPROB, Probabilistic Rules-Based Optimal
Design for Ro-Ro Passenger Ships: EU FP5 RTD
Project G3RD-CT-2000-00030, 1999-2002.
[6]`Jasionowski, A: Survival Criteria for Large
Passenger Ships, SAFENVSHIP Project,
September 2005, Final Report, Safety at Sea Ltd
[7] Jasionowski, A, Vassalos, D: Benchmark Study
on the Capsizing of a Damaged Ro-Ro Passenger
Ship in Waves, Final Report to the ITTC
Specialist Committee on the Prediction of Extreme
Motions & Capsizing, December 2001.
[8] SAFEDOR (2005-2009): Design, Operation and
Regulation for Safety, Integrated Project, FP6_2
Contract TIP4-CT-2005-516278,
www.safedor.org.
[9] Vassalos, D, York, A, Jasionowski, A, Kanerva, M
and Scott, A: Design Implications of the New
Harmonised Damage Stability Regulations,
STAB 2006, Rio de Janeiro, Brazil, September
2006.
[10] Vassalos, D and Jasionowski, A: SOLAS 2009
Raising the Alarm, 9th International Stability
Workshop, Hamburg, Germany, August 2007.
[11] Vassalos, D, Jasionowski, A, York, A and
Tsakalakis, N: SOLAS 90, Stockholm

Agreement, SOLAS 2009 The False Theory of
Oranges and Lemons, 10th International Stability
Workshop, Daejeon, Korea, March 2008.
[12] Jasionowski, A: An Integrated Approach to
Damage Ship Survivability Assessment, PhD
dissertation, February 2001, University of
Strathclyde.
[13] Vassalos, D, Jasionowski, A: Damaged Ship
Hydrodynamics, Proceedings of the 6th
International Ship Stability Workshop, Webb
Institute, New York, October 2002
[14] Van den Bosh, JJ Vugts, JH: Roll damping by
free surface, Report no. 134, Shipbuilding
Laboratory, Delft.
[15] Hamlin, NA, Lou, YK, Maclean, WM, Seibold, F,
Chandras, LM, Liquid sloshing in slack ship
tanks - theory, observations and experiments,
SNAME Transactions, Vol. 94, 1986, pp. 159-
195.
[16] Mikelis, NE, Miller, JK, Taylor, KV: Sloshing in
partially filed liquid tanks and its effect on ship
motions: numerical simulations and experimental
verification, RINA Spring Meetings 1984, Paper
No. 7.
[17] Francescutto, A, Contento, G: An experimental
study of the coupling between roll motion and
sloshing in a compartment, ISOPE94, Osaka,
10-15 April, 1994, Paper N. 94-YI1-4
[18] Bass, D, Cumming, D: An experimental and
numerical investigation of the effects of water
trapped on the deck, STAB 2000, Launceston,
Tasmania, Australia.
[19] Daalen, EFG van, Kleefsman, KMT, Gerrits, J,
Luth, HR, Veldman, AEP: Anti-roll tank
simulations with a volume of fluid (VOF) based
Navier-Stokes Solver, 23rd Symposium on Naval
Hydrodynamics, 2000.
[20] Veer, R vant, Kat Jan de; Experimental and
numerical investigation on progressive flooding
and sloshing in complex compartment
geometries, STAB2000, Launceston, Tasmania,
Australia.
[21] Woodburn, P, Letizia, L and Gallagher, P:
Fundamentals of Damaged Ship Survivability,
RINA Transactions, June 2002.
[22] Gao, Q, Kara, F, Shigunov, V, and Vassalos, D:
Numerical Simulation of Flooding of a Damaged
Ship, NuTTS 2004, Hamburg.
[23] Gao, Z, Vassalos, D and Gao, Q: A multiphase
CFD method for prediction of floodwater
dynamics, 10th International Conference on
Stability of Ships and Ocean Vehicles, St
Petersburg, 2009.
[24] Shen, L and Vassalos, D: Applications of 3D
parallel SPH for sloshing and flooding, 10th
International Conference on Stability of Ships and
Ocean Vehicles, St Petersburg, 2009.
[25] Gao, Q, Gao, Z and Vassalos, D: Numerical
Study of Damage Ship Flooding, 4th
The Damaged Ship, London, UK
International Design For Safety Conference,
Trieste, 18-20 October 2010.
[26] Gao, Z, Vassalos, D and Gao, Q: Numerical
simulation of water flooding into a damaged
vessels compartment by the volume of fluid
method, Ocean Engineering 37 (2010) 1428
1442.
[27] VIRTUE, The Virtual basin, EC FP6 IP, Project
516201, TIP5-CT-2005-516201, 2005, 2009.
[28] Vassalos, D., Turan, O: "A realistic approach to
assessing the damage survivability of Passenger
ships", Transactions SNAME 1994.
[29] Vassalos, D., Letizia, L: "Formulation of a non-
linear mathematical model for a damaged ship
subject to flooding", Sevastianov Symposium,
Kaliningrad, 1995.
[30] Kat, J de: Dynamics of a ship with partially
flooded compartment, 2nd Stability Workshop,
Japan, 1996.
[31] Zaraphonitis, G, Papanikolaou, A and Spanos, D:
"On a 3D Mathematical Model of the Damage
Stability of Ships in Waves", STAB 1997,
Bulgaria.
[32] Papanikolaou, A, Zaraphonitis, G, Spanos, D,
Boulougouris, E, Eliopoulou, E: Investigation
into the capsizing of damaged Ro-Ro passenger
ships in waves, STAB 2000, Launceston,
Tasmania, Australia.
[33] Himeno, Y: Prediction of ship roll damping state
of the art, Report No 239, University of
Michigan, College of Engineering, September
1981.
[34] Salui, KB, Sarkar, T and Vassalos, D: An
Improved Method for Determining Hydrodynamic
Coefficients in Roll Motion Using CFD
Techniques, Ship Technology Research, Vol47,
No. 4, October 2000.
[35] Cichowicz, J, Jasionowski, A and Vassalos, D:
Uncertainty Assessment in Experiments on a
Floating Body in Forced Roll Motion in Calm
Water, Proceedings of the 10th International
Stability Workshop, 2010, Wageningen,
Netherlands.
[36] Vassalos, D, Jasionowski, A and Guarin, L:
Risk-Based Design: A Bridge too far?, OC
2008
Seakeeping and Stability, Osaka, Japan, March
2008.
[37] IMO MSC 85: FSA Cruise Ships, submitted by
Denmark, May 2009.
[38] Vassalos D, Kim H, Christiansen G and Majumder
J, A Mesoscopic Model for Passenger Evacuation
in a Virtual Ship-Sea Environment and
Performance Based Evaluation, Pedestrian and
Evacuation Dynamics, April 4-6, 2001
Duisburg, www.safety-at-
sea.co.uk/evi/publications.htm.
[39] Jasionowski, A: Survivability Assessment -
GENESIS, SaS-RCGN10, 24th September 2007.
The Damaged Ship, London, UK

[40] Jasionowski, A: Physical Model Experiments on

survivability of Genesis, RCGN14-Experiments-
Report-24, 2nd March 2009.
[41] Chen, Q, Tsakalakis, N and Jasionowski, A:
SAFEDOR Mathematical model of risk
physical experiments, D5.6.3, SAFEDOR-D-
5.6.3-2009-04-30-SSRC-Experimentsrev-1, 17
July 2009.
The Damaged Ship, London, UK

A PRACTICAL GUIDE TO DAMAGE STABILITY ASSESSMENT REGULATION ON DAMAGE

STABILITY

J Gullaksen, JG Consultant Engineers, Denmark

SUMMARY

In order to assess the behaviour of the ship after damage, two different approaches are used for damage stability
assessment: the deterministic approach and the probabilistic approach, which are to be applied depending on the ship
type. The damage stability requirements are more critical than the intact stability requirements, and as a result of new
harmonised probabilistic damage stability rules, IMO Resolution MSC 216(82), which introduced new ship damage
stability regulations into SOLAS from 1 January 2009, discussions about what effects the new rules will have on ship
designs have been held throughout the shipping industry. This paper is intended to introduce the conceptual framework
and to provide clarification for the initial preparation of damage stability assessment and to the use of these regulations.
An Excel Spreadsheet with VBA code has been developed for calculation of the pi factor probability to various extents
of damage and damage case occurring.

NOMENCLATURE cases, the number of which, as well as the number of

compartments involved in each case, depend on the
Symbol Definition Unit ships dimensions and internal subdivision. The
A Attained subdivision index - deterministic approach can be classified in a
b Mean transverse distance m mathematical modeling language, as one in which every
General location of compartment(s) set of variable states is uniquely determined by
G -
factor parameters in the model and by sets of previous states of
GZ Positive righting lever m these variables. Therefore, deterministic models perform
Heigh of the horizontal subdivision the same way for a given set of initial conditions.
H m
above the baseline
Maximum non-dimensional damage The probabilistic approach refers to a method for
Jmax -
length computing a measure of probable damage survivability
Ls Subdivision length m which takes into account not only the vessels stability
Probability of compartment(s) when certain subdivisions of the ship are damaged but
p -
damaged assigns probabilities to various extents of damage as well
R Required subdivision index - as to the survivability. In a probabilistic or stochastic
r Reduction factor - model, randomness is present, and variable states are not
Ra Range deg described by unique values, but rather by probability
si The factor si, for any damage case - distributions.
The probability to survive in the
sfinal -
final equlibrium stage of flooding In SOLAS2009 two versions of the probabilistic method
e Final equlibrium angle of heel deg are prescribed: one for passenger ships and one for cargo
v Submerged/Flooding angle deg vessels. In present paper only cargo vessels are
- considered.
v Reduction factor v
This paper has been prepared to serve as a practical guide
to damage stability assessment for the purpose to give
1. INTRODUCTION guidance on the application of the regulations for damage
stability calculation including an overview of the present
Two different approaches are used for damage stability regulatory framework. An Excel Spreadsheet1 is used to
assessment: calculate the probabilities of a damage case occurring.
Additional commercial software is necessary to calculate
the deterministic approach and the probability of surviving a particular damage case for
the probabilistic approach, calculation of the attained safety index.

which are to be applied depending on the ship type.
The deterministic approach is based on standard
dimensions of damage extending anywhere along the
ships length or between transverse bulkheads depending
on the relevant requirements. The consequence of such
standard of damage is the creation of a group of damage
First a brief introduction to the deterministic damage
stability regulations are considered in section 2.
Subsequently, in section 3, the basic consepts in the
probabilistic damage stability methods are introduced as
set out in SOLAS2009. Section 4 summarize the
1
Available on http://www.jg-consultant.com, Download section.
 The Damaged Ship, London, UK

information needed for damage stability assessment, European Parliament. According to EU Council
followed by preparation of input in section 5, a Directive 2009/45/EC, Article 3, Scope, the Directive
probabilistic damage stability procedure in section 6, to applies to the following passenger ships and craft,
demonstrate the use of the regulation and description of regardless of their flag, when engaged on domestic
minimum required output for probabilistic damage voyages:
stability assessment in section 7.
a. new passenger ships;
The organization of this paper generally follows Part B b. existing passenger ships of 24 metres in length and
of SOLAS 2009 [1]. above;
c. high-speed passenger craft.Damage Stability

2. DETERMINISTIC DAMAGE STABILITY Besides the probabilistic component, the new SOLAS
2009 Part. B-1 also contains a deterministic component
2.1 INSTRUMENTS COVERED BY conserning passenger ships, Regulation 8. The idea
DETERMINISTIC CRITERIA behind this deterministic component is that a minor
damage shall not result in a major consequence.
Different deterministic methods in damage stability have
been developed depending on ship type, on freeboard These deterministic requirements may partly overrule the
reduction, and on the kind of cargo carried. The probabilistic part of SOLAS 2009 Part B-1.
deterministic approach is used for vessels excluded from
the application of SOLAS 2009, Part B-12: The SOLAS 2009 Part B-2 also contains a deterministic
component regarding bottom damages. According to
- International Bulk Chemical (IBC) Code3, SOLAS 2009, Regulation 9.8, compliance is to be
- The BCH Code, achieved by demonstrating that si, when calculated in
- International Gas Carrier (IGC) Code, accordance with regulation 7-2, is not less than 1 for all
- Guidelines for the design and construction of service conditions when subject to a bottom damage
Offshore Supply Vessel (OSV) Guidelines assumed at any position along the ship's bottom and with
(Resolution A.469(XII)), an extent specified in 9.8.2 for the affected part of the
- International Code of Safety for High-Speed Craft ship.
(2000 HSC Code),
- Damage stability requirements of regulation 27 of 2.2 GENERAL CONSIDERATIONS FOR
the 1966 Load Line Convention as applied in DETERMINISTIC DAMAGE STABILITY
compliance with resolution A.320(IX) and CALCULATIONS
A.514(13), provided that in the case of cargo ships
to which regulation 27(9) applies, main transverse Requirements for ship survival capability, or damage
watertight bulkheads, to be considered effective, are stability, are given in the relevant Regulations.
spaced according to paragraph (12)(f) of resolution
A.320(IX), except ships intended for the carriage of Damage survival capability shall be investigated on the
deck cargo, basis of loading information submitted to the
- Damage stability requirements of regulation 27 of Administration for all anticipated conditions of loading
the 1988 Load Line Protocol, except ships intended and variations in draught and trim.
for the carriage of deck cargo,
- MARPOL Annex I specifies oil tanker subdivision 2.2.1 Damage Assumptions
& damage stability requirements including double-
hulls, except combination carrier (as defined in A general case of assumed maximum extent of damage
regulation II-2/3.14) with type B freeboards are not stability is illustrated in Table 2-2 and Table 2-14.
excluded. MARPOL Annex I, Regulation 28.
Table 2-1 Assumed maximum extent of bottom damage.
For Large Commercial Yacht, the damage stability is
given by the UK MCA, Large Commercial Yacht Code MARPOL / IBC
(LY2) - MSN 1792 (M), UK National Authority, MCA.
For 0.3L from FP Else
2/3
Within EU, the damage stability for ro-ro passenger ships Longitudi Min(1/3L or 14.5 Min(1/3L2/3 or 14.5
is given in Directive 2003/25/EC and 2009/45/EC of the nal m) m)
Transverse Min(B/6 or 10 m) Min(B/6 or 5 m)
2
Vertical MinB/15 or 6 m) Min(B/15 or 6 m)
SOLAS Ch.II-1 Part B-1 provides for probabilistic subdivision &
SOLAS2009, Regulation 9-8.2
damage stability for dry cargo ships of L > 80 m unless covered by
deterministic criteria. Vertical Min(B/20 or 2 m) Min(B/20 or 2m)
3
Note that where a chemical carrier can carry petroleum grades in
addition to noxious liquid substances in bulk, MARPOL Annex I also
4
applies. This has to be in accordance with the relevant regulation in question.
The Damaged Ship, London, UK
Table 2-2 Assumed maximum extent of side damage.
Longitudinal Min(1/3L2/3 or 14.5 m)
Transverse Min(B/5 or 11.5 m)
Vertical From the baseline upwards without limit
The B/5-line is used in ship design and damage stability
calculations. No damage extends inside the B/5-line. Due
to this, bilge main lines, fuel tanks, etc. are normally
located inside this B/5. The probabilistic rules have made
this rule obsolete on new ships.
2.2.2 Flooding Assumptions
The survival requirements regarding flooding, shall be
confirmed by calculations which take into consideration
the design characteristics of the ship; the arrangements,
configuration and contents of the damaged
compartments; the distribution, relative densities and the
free surface effects of liquids; and the draught and trim
for all conditions of loading.
The permeabilities of spaces assumed to be damaged
shall be considered in accordance with the Code.
2.2.3 Standard Of Damage
Ship shall be capable of surviving the damage indicated
in the damage assumptions, with the flooding
assumptions to the extent determined by the ships type
according to the standard.
2.2.4 Survival Requirements And Damage Stability
Criteria
Ship subject to the Code shall be capable of surviving the
assumed damage specified in the damage assumptions to
the standard provided in the standard of damage in a
condition of stable equilibrium and shall satisfy the
criteria in the standard.
For each loading condition, each damage case is to be
considered, and the applicable criteria are to be complied
with.
The metacentric heights, GM, stability levers, GZ, and
centre of gravity, KG, positions for judging the final
conditions are to be calculated by the constant
displacement (lost buoyancy) method or the added
weight method. These give different GM values but the
same righting moment.
The above methods consider consequences of flooding a
particular compartment. The floodable length at any
point in the length of the ship is defined as the maximum
length, having its center at the point in question, that can
be symmetrically flooded at the prescribed permeability,
without immersing the margin line.
3. PROBABILISTIC DAMAGE STABILITY
3.1 GENERAL
The requirements for damage stability calculation refer
here to chapter II-1 of SOLAS as amended and the
related Explanatory Notes. The regulations in SOLAS
Chapter II-1, Parts B-1 through B-4, Subdivision and
Damage Stability, shall apply to cargo ships of 80 m in
length and upwards and to all passenger ships regardless
of length but shall exclude those cargo ships which are
shown to comply with subdivision and damage stability
regulations in other instruments developed by the
Organization.
3.2 REQUIRED SUBDIVISION INDEX
The required subdivision index, R, is a comparative
index that is used to set the "level of safety". The formula
in Regulation 6 that is used to calculate R for ships over
80 m in length was first developed by calculating the
Attained Subdivision Index, A, for a representative
sample of existing dry cargo ships. The sample ships all
had some measure of subdivision and stability before
damage and at least met the minimum intact stability
requirements recommended by IMO. Using the
distribution of A values for various ship lengths, a
formula was devised for R based on the average A
values.
For all ships to which the damage stability requirements
apply, the degree of subdivision to be provided shall be
determined by the required subdivision index R, as
follows:
In present paper only cargo ships greater than 100 m in
length Ls are considered5:
Equation 3-1 =
3.3 ATTAINED SUBDIVISION INDEX
The damage stability calculations are performed at three
drafts: the deepest subdivision load line, ds, the light
service load line, dl, and the partial load line, dp, and
relevant GM values in order to draw a minimum GM
curve where the attained subdivision index A achieves
the minimum required level of safety R.
The attained subdivision index A is obtained by the
summation of the partial indices As, Ap and Al, (weighted
as shown) calculated for the draughts ds, dp and dl in
accordance with the following formula:
Equation 3-2 = . + . + .
5
See SOLAS2009 [1], Regulation 6-2, for further details.

where the indices s, p and l represent the three loading
conditions and the factor to be multiplied to the index
indicates how the index A from each loading condition is
weighted.
For each of the three considered draughts the calculated
partial indices As, Ap and Al shall meet a percentage of
the total attained index A. For dry cargo vessels that
percentage shall be not less than 0.5R.
When calculating the attained index A, each partial index
is a summation of contributions from all damage cases
taken in consideration, using the following formula:
Equation 3-3 = dependent on the calculated stability of the ship
The method of calculating the A for a loading condition
is expressed by the formula:
Equation 3-4 =
[ ]
Where
i represents each compartment or group of
compartments under consideration, i.e. each
investigated damage or group of damages,
t represents the number of damages to be investigated
to calculate Ac for the particular loading condition,
pi accounts for the probability that only the
compartment or group of compartments under
consideration may be flooded, disregarding any
horizontal subdivision,
si accounts for the probability of survival after flooding
the compartment or group of compartments under
consideration, and includes the effect of any horizontal
subdivision.
c represents one of the three loading conditions, To
obtain a maximum index A for a given subdivision, t
has to be equal to T, the total number of damages.
In regulation 7-1, the words compartment and group of
compartments should be understood to mean zone and
adjacent zones.
Zone is a longitudinal interval of the ship within the
subdivision length.
Room is a part of the ship, limited by bulkheads and
decks, having a specific permeability.
Space is a combination of rooms.
Compartment is an onboard space within watertight
boundaries.
Damage is the three dimensional extent of the breach in
the ship.
In summary, the attained subdivision value A is based on
the following factors:
1. The probabilistic value of:
The Damaged Ship, London, UK
a. The pi-factor for each compartment. The p factor
is solely dependent on the geometry of the
watertight arrangement of the ship.
b. The reduction factor r. The r factor may be used to
modify the p factor for a compartment if
longitudinal subdivision is present.
c. The reduction factor v. The v factor is dependent
on the geometry of the watertight arrangement
(decks) of the ship and the draught of the initial
loading condition. It represents the probability
that the spaces above the horizontal subdivision
will not be flooded.
2. The factor of survivability
d. The si-factor for each compartment. The s factor is
after damage in a specific initial condition.
The p, r, v and s factors
The factor pi for a compartment or group of
compartments is to be calculated in accordance with
SOLAS 2009, Part B-1, Regulation 7-1.1 and the factor ri
is to be calculated in accordance with Regulation 7-1.1.2.
The factor si is to be determined for each case of assumed
flooding involving a compartment or group of
compartments according to the requirement indicated
inaccordance with SOLAS 2009, Part B-1, Regulation 7-
2.1 to 7-2.5.5 and together with the factor vi is to be
calculated in accordance with Regulation 7-2.6.
For the calculation of p, r, b and v only the damage is
considered. For the calculation of the s-value the flooded
space is considered.
Calculation of the factor pi
As the p-factor is related to the watertight arrangement
by the longitudinal limits of damage zones and the
transverse distance from the ship side to any longitudinal
barrier in the zone, the following indices are introduced:
j: the damage zone number starting with No.1 at the
stern;
n: the number of adjacent damage zones in question
where j is the aft zone;
k: the number of a particular longitudinal bulkhead as a
barrier for transverse penetration in a damage zone
counted from shell towards the centreline. The shell has
No. 0;
K: total number of transverse penetration boundaries;
,, : the p-factor for a damage in zone j and next (n-1)
zones forward of j damaged to the longitudinal bulkhead
k.
Pure longitudinal subdivision
In order to prepare for the calculation of Index A the
ships subdivision length Ls is divided into a fixed
discrete number of damage zones. These damage zones
will determine the damage stability investigation in the
way of specific damages to be calculated.
The Damaged Ship, London, UK

There are no rules for the subdividing except that the where
length Ls defines the extremes for the actual hull.

However, it is important to consider a strategy carefully , =

to obtain a good result (that is a large attained index A.)
All zones and combination of adjacent zones may the total number of s for the adjacent zones in
contribute to the index A. question.

Single damage zone, (1, 2, ) = 0

, = (1 , 2 )
Then the pi factor can be calculated as:
Two adjacent zones,
, = (1 , 2 ) 1 , 2 (1 , 2 ) For single zone damage
,, = (1 , 2 )[1 , 2 , 1 , 2 , ]
i.e. p4&5 = p45 p4 p5 For two zones damage:
,, = (1 , 2 )1 , 2 ,
Three or more adjacent zones, 1 , 2 ,
1 , 2 1 , 2 ,
, = 1 , 2 1 , 2
1 , 2 ,
1 , 2
+ (1 , 2 ) (1 , 2 )1 , 2 ,
1 , 2 ,
i.e. for three zones p4&5&6 = p46 p45 p56 + p55
For three or more zones damage:
Transverse subdivision in a damage zone ,,
The ri factor may be used to modify the pi factor for a = 1 , 2 1 , 2 ,
compartment if longitudinal subdivision is present. The ri 1 , 2 ,
factor for any compartment is calculated using the 1 , 2 1 , 2 ,
equations in Regulation 7-1 once the "b" value is
determined. 1 , 2 ,
1 , 2 1 , 2 ,
The simplistic definition of "b" is that it is the mean 1 , 2 ,
distance between the shell and the longitudinal + (1 , 2 ) 1 , 2 ,
subdivision in question measured at the deepest 1 , 2 ,
subdivision load line. The complete definition of "b"
however, allows wide latitude in how this distance is
To calculate the pi factor for a compartment using the
measured. Regulation 7-1 says that the distance is
equations in Regulation 7-1, the location of the fore and
measured between the shell and a plane parallel to the
aft limits of the compartment must be known. Typically a
longitudinal bulkhead.
ship can be divided longitudinally into zones based on
the transverse subdivision. A pi factor is then calculated
In calculating r-values for a group of two or more
for each zone, which is applicable to all compartments
adjacent compartments, the b-value is common for all
within that zone.
compartments in that group, and equal to the smallest b-
The total accumulated p is determined by
value in that group:

= ( , , , )
= ,
where n is the number of wing compartments in that

group and , , , is the mean values of b for

individual wing compartments contained in the group.
Accumulating p
The accumulated value of p for one zone or a group of
adjacent zones is determined by:
, subdivision draught - is calculated using information
, = ,,
where T is the number of damage zones plus the
maximum number of combined adjacent zones.
Calculation of the factor si
The s factor at each draft - ds at deepest subdivision
draught, dl at light service draught and dp at partial
from the damaged righting arm curve that results from
flooding the compartment or group of compartments
under consideration.
 The Damaged Ship, London, UK

The si-factor given in SOLAS 2009 Reg. B-1/7-2. The permeability depends on the considered draught:
Calculation of the factor si is defined as follows: deepest subdivision, partial subdivision and light service
draught.
The factor Sfinal,i shall be obtained from the formula:
The permeability also depends on the type of cargo
space: dry cargo, container, ro-ro and/or liquid cargo.

, =
0.12 16 3.4 SUBDIVISION, WATERTIGHT AND
where
WEATHERTIGHT INTEGRITY
GZmax is not to be taken as more than 0.12 m;
Range is not to be taken as more than 16;
SOLAS 2009, Part B-2, Regulations 9 to 17, specifies
K=1 if
requirements to subdivision, watertight and weathertight
K=0 if integrity, including:

= otherwise,

- Double Bottom, Regulation 9
where - Watertight Bulkheads and Decks
is the equilibrium heel angle in any stage of flooding, Watertight Subdivision
in degrees; Arrangement of Watertight Bulkheads
is 25 for cargo ships; and Collision bulkhead
is 30 for cargo ships. Stern tube bulkhead and remaining
watertight bulkhead
The vi factor Openings in Watertight Bulkheads
Using the same indices as above this formula expresses - External Openings
the v-factor: - Cross-Flooding Arrangements

,, = ,, , ,, ,
4. DAMAGE STABILITY INFORMATION
where m is the index for horizontal subdivision.
For all ships to which damage stability requirements
If (H-d) is less than, or equal to, 7.8 m apply, documents including damage stability calculations
are to be submitted.
( )
(, ) = 0.8
7.8 The following documents are to be provided:
else
( ) - drawings showing the external openings, watertight
(, ) = 0.8
7.8 subdivision as well as internal openings
where - damage stability calculation in accordance with
,, = ,, , = 1 SOLAS as amended and the related Explanatory
Notes
If Hm coincides with the uppermost watertight boundary - damage control plan and damage control booklet
of the ship within the range (() () ), and containing all data essential for maintaining the
,,, is to be taken as 0. survival capability
- stability information.
In general, each contribution dA to the index A in the
case of horizontal subdivisions is obtained from the The damage stability calculations are to include
formula: following information:
= [ + ( ) +
+ (1 ) ] main dimensions and lines plan
where: hydrostatic data and cross curves of stability
vm is the v-value calculated in accordance with definition of sub-compartments with moulded
Regulation 7-2.6.1, and smin is the least s-factor for all volumes, centres of gravity and permeability
combinations of damages obtained when the assumed list of the characteristics (volume, centre of gravity,
damage extends from the assumed damage height H permeability) of each compartment which can be
downwards. damaged
a table of openings in bulkheads, decks and side
Permeability Regulation 7-3 shell reporting all the information about:
For the purpose of the subdivision and damage stability - identification of the opening
calculations, the permeability of each general - vertical, transverse and horizontal location
compartment or part of a compartment are considered in - type of closure: sliding, hinged or rolling for
accordance with Regulation 7-3 doors
The Damaged Ship, London, UK
- type of tightness: watertight, weathertight, semi-
watertight or unprotected
- operating system: remote control, local
operation, indicators on the bridge, television
surveillance, water leakage detection, audible
alarm, as applicable
- foreseen utilization: open at sea, normally
closed at sea, kept closed at sea
list of all damage cases corresponding to the
applicable requirements
detailed results of damage stability calculations for
all the loading conditions foreseen in the applicable
requirements
the limiting GM/KG curve, if foreseen in the
applicable requirements
capacity plan
arrangement of cross-flooding, pipes showing
location of remote controls for valves, or special
mechanical means to correct list due to flooding, if
any
watertight and weathertight door plan.
For the probabilistic approach, in addition to the
information listed above, the following is to be provided:
subdivision length Ls
initial draughts and the corresponding GM-values
required subdivision index R
attained subdivision index A with a summary table
for all contributions for all damaged zones.
draught, trim, GM in damaged condition
damage extension and definition of damage cases
with probabilistic values p, v and r
righting lever curve (including GZmax and range)
with factor of survivability s
critical weathertight and unprotected openings with
their angle of immersion
details of sub-compartments with amount of in-
flooded water/lost buoyancy with their centres of
gravity.
5. PREPARATION OF INPUT TO
PROBABILISTIC DAMAGE STABILITY
CALCULATIONS
5.1 GENERAL
The first step is to define the division geometry, since
wee needs to know what are considered divisions. It does
not assume that each tank and compartment is an
independent division.
Normally, each double bottom tank should be included
with at least one other compartment, rather than being a
division by itself.
For defining divisions, a unique number n is assigned to
the division which involves those tanks listed by name.
If the division includes a wing tank, then a parameter
should be used along with the wing breadth b as defined
in the IMO rule.
If a horizontal bulkhead is present above the waterline
and capable of limiting the flooding when not damaged,
its height relative to the baseline can be indicated with a
parameter.
The standard does not stipulate how the ship should be
subdivided. Instead the performance of a proposed
arrangement is evaluated for typical damage scenarios.
5.2 SUBDIVISION LENGTH
According to Regulation 2, the subdivision length Ls is
the greatest projected moulded length of that part of the
ship at or below deck or decks limiting the vertical extent
of flooding with the ship at the deepest subdivision
draught.
The maximum possible vertical extent of damage above
the baseline is ds + 12.5 metres.
5.3 COMPARTMENT BOUNDARIES TABLE
AND SUBDIVISION MATRIX
A compartment boundaries table can be used to describe
the watertight subdivision of the ship, that is
longitudinal-, transverse- and vertical watertight
structure.
Having considered the watertight subdivision of the ship,
that is longitudinal-, transverse- and vertical watertight
structure a useful tool to combine the damages to be
examined is a subdivision matrix and a compartment
boundaries table as shown in Table 5-1. See Appnndix A.
In Table 5-1 the subdivision matrix contains a number of
zones and a number up to two transverse barriers in each
zone and up to two vertical barriers (decks) in each zone
(could be expanded if necessary).
Table 5-1 The subdivision matrix.
Zone (j) x1 x2 b1 b2 H1 H2
1
2
.
The matrix contains information of all the single zone
damages to be considered for calculating the attained
index A that is: the longitudinal zones (x1 x2), transverse
barriers (longitudinal bulkheads in the zones) (b) and
vertical barriers (decks) in the same zones (H).
Wing tanks or compartments are effective in limiting the
flooding due to horizontal penetration through the side
shell of the vessel. The extent of such penetration is
linked to its probability. A greater extent of penetration
requires more energy and is therefore less likely. In
 The Damaged Ship, London, UK

concept, a given wing tank design will limit flooding to External weathertight critical points only affect the
the wing tank when the penetration is less than a certain results if after damage they are submersed at equilibrium.
value. In order to get the highest probability out of this
damage case. i.e. to contribute most to the attained index, The SOLAS 2009 regulations also require identification
the designer will want the penetration value used in the of internal critical points. These points all only affect the
calculations to be as large as possible. answer if after damage they are submersed at
equilibrium, similar to a weathertight point. These points
This causes the following questions to arise: include points of progressive flooding, horizontal
evacuation routes, vertical escape hatches, controls for
- What is this penetration value which just misses watertight closures and equalization devices, and piping
rupturing the wing bulkhead? or ventilation ducting carried through watertight
- Is it simply the distance from the side shell to the boundaries without watertight closures.
wing bulkhead?
- What if the side shell is not a nice flat wall? Regulation 7-2.1 External openings incapable of being
- What if the bulkhead is sloping in various directions, closed weathertight (for use in determination).
has knuckles or is stepped?
- What is a fair penetration value to use in those Regulation 7-2.5.2.1 Progressive flooding (internal and
cases? external) openings (air-pipes, ventilators, weatertights
doord or hatches, etc.).
The Explanatory Notes, which are part of the regulations,
say that the plane need only be tangent to the Regulation 7-2.5.2.2 Bulkhead deck horizontal escape
longitudinal bulkhead, as long as the maximum distance routes.
between the plane and the shell does not exceed twice the
least distance between the plane and the shell. The reader Regulation 7-2.5.3.1 Bulkhead deck vertical escape
should refer to the Explanatory Notes for examples that routes.
illustrate this definition [2], [3] and [4].
Regulation 7-2.5.3.2 Controls for watertight doors, cross-
Damage to the hull in a specific damage zone may just flooding valves, watertight valves and dampers intended
penetrate the ships watertight hull or penetrate further to prevent progressive flooding.
towards the centreline. To describe the probability of
penetrating only a wing compartment, a probability Regulation 7-2.5.3.3 Piping and vents carried through
factor r is used, based mainly on the penetration depth b. watertight boundaries (not fitted with watertight closures
The value of r is equal to 1, if the penetration depth is at the boundary).
B/2 where B is the maximum breadth of the ship at the
deepest subdivision draught ds, and r = 0 if b = 0. For the SOLAS 2009 damage analysis all critical points
included in the analysis can be listed in a table like this:
The penetration depth b is measured at level deepest
subdivision draught ds as a transverse distance from the LCG[ TCG[ VCG[ trim
ship side right-angled to the centreline to a longitudinal m] m] m] [deg]
barrier. displ (+
[LT] (+ aft) stbd) (+ up) (+fwd)
Where the actual watertight bulkhead is not a plane
parallel to the shell, b should be determined by means of ds
an assumed line, dividing the zone to the shell in a
relationship b1/b2 with 1/2 b1/b2 2.
dp
After all of the divisions have been defined, the Attained
Subdivision Index in the current loading condition may dl
be obtained by computing and displays the probability of
damage and probability of survival for each damage case
as well as the attained index.
6. PROCEDURE FOR PROBABILISTIC
5.4 CRITICAL POINTS DAMAGE STABILITY CALCULATIONS
The SOLAS 2009 explanatory notes say "openings are 6.1 GENERAL
the most critical factor to preventing an inaccurate
attained index A". A descriptive procedure is provided to illustrate the
application of the regulations described in this paper. The
The SOLAS 2009 regulations use two types of external calculation steps shown generally follow the order in
critical points: weathertight and unprotected.
The Damaged Ship, London, UK
which the regulations are presented in SOLAS Chapter
II-1, Part B-1.
6.2 PROBABILISTIC DAMAGE STABILITY
CALCULATION STEPS
The following steps illustrate how a probabilistic damage
analysis is performed:
1. Model the vessel and define tanks and
compartments.
2. Define other ship data required for the probabilistic
damage analysis. This data includes the vessel type,
which load cases to consider, and the number of
adjacent damage zones to consider.
3. Defines the boundaries of the damage zones.
4. Longitudinal bulkhead and deck locations are
defined for each zone and groups of adjacent zones.
5. Once steps 2 and 3 have been completed, the p-
factors are calculated. With the bulkheads and
decks defined in step 4, the r- and v-factors are also
calculated.
6. When the zones have been defined, which tanks are
damaged in each zone and sub-zone are also
defined.
7. Perform a probabilistic damage analysis. Large
angle stability analysis for each combination of
loadcase and damage and collates the results to
calculate the attained index. This index (A) is then
compared with the required index R to determine a
pass or fail.
The fundamental approach of the probabilistic damage
analysis is to first assume that the vessel has been
damaged, then to assign a probability, p, that the damage
will occur in a certain area of the ship. Given that the
damage has occurred, a probability that the vessel will
survive, s, is then calculated from certain parameters of
the GZ curve calculated for the vessel in that damage
condition. The conditional probability that the vessel will
sustain certain damage and survive is then given by the
product . By summing this probability for a range
of different damage scenarios, the total probability of the
vessel surviving a damage incident is calculated this is
the attained index, A. The attained index can then be
compared with a required index, R, to determine if the
vessel is sufficiently safe.
A numerical method has been used for the determination
of the probability of damage. The method is implemented
in a VBA script in an Excel spreadsheet enclosed in
appendix. The spreadsheet with VBA macros,
SOLAS2009_JGCE.xlsm6, includes the inputs with
main parameters and calculates the p-factor probabilities
of a damage case occurring. Additional commercial
software is necessary to calculate the probability of
6
Available on http://www.jg-consultant.com,
Download section.
surviving a particular damage case for calculation of the
attained safety index.
6.3 A SIMPLIFIED CALCULATION
PROCEDURE
For each loading condition, a procedure can be written as
follows to calculate the attained subdivision index, A:
1. Initialize Attained Index, A, to zero.
2. Determine probability of damage, prv, as follows:
a. Determine size of integration steps in
longitudinal, transverse and vertical directions
- For all longitudinal integration steps
- For all transverse integration steps
- For all vertical integration steps
b. Determine the volumetric element as imposed by
the present combination of integration steps
c. Determine prv, that is the probability of damage
of this volumetric element
d. Determine the damage case (the combination of
affected compartments) implied by this
volumetric element
e. Does this damage case already exist?
- If No: Create this damage case, and set
aggregated probability of damage to prv
- If Yes: Add prv to the aggregated probability
of damage for this damage case
3. Calculate probability of survival, s.
4. The product, p s is added to the attained index,
A = A + prvs.
5. Steps 3 through 5 are repeated for each division.
6. Steps 1 through 6 are repeated for additional load
condition(s) and the attained indices for each are
averaged.
The pi, ri, vi, and si factors are calculated for each
compartment alone and then for groups of adjacent
compartments.
The products for each compartment or
group of compartments are summed to determine the
attained index.
As both vi and si are different for each draft, the
summation is calculated for each draft and then
averaged to get the resulting A value.
When performing the probabilistic damage procedure,
the procedure sorts the divisions into order according to
their forward longitudinal locations. It then proceeds to
use the divisions from bow to stern. After completing the
survivability analysis using the first (forwardmost)
division, it takes the next division which starts at or after
the aft end of the present division. Hence any
overlapping division will be ignored.
For this purpose, the nominal forward and aft ends of the
division are used. The nominal ends are the same as the
actual ends taken from the geometry unless set nominal
ends at other locations. By this means, divisions which
do overlap can be made acceptable.
 The Damaged Ship, London, UK

From the standpoint of the procedure, the great
advantage of the probabilistic method is that it embraces
the problem of damage extents as well as the problem of
survivability after damage, combining the two in one
elegant measure. This formalizes and standardizes the
generation of damage scenarios and makes it possible to
present the results in a compact format. Since the
probabilistic methodologies directly address specific
features of ship subdivision, there are (at least
theoretically) fewer decisions left to the person running
the calculations. It offers the potential of a high degree of
automation and time-saving for the designer.
The deepest subdivision draught (ds) is the waterline
which corresponds to the summer load line draught of the
ship.
The GM values for the three loading conditions could, as
a first attempt, be taken from the intact stability GM limit
curve. If the required index R is not obtained, the GM
values may be increased implying that the intact loading
conditions from the intact stability book must now meet
the GM limit curve from the damage stability
calculations derived by linear interpolation between the
three GMs.

The calculation for the index A is carried out for even

7. OUTPUT FROM PROBABILISTIC keel. This allows for varying of the ships operating trim
DAMAGE STABILITY CALCULATIONS for each draught within the trims of 0.5%Ls without any
change of the GM requirements. In order to show the
The damage calculations results have to include influence of the trim, in excess of 0.5%Ls, the
minimum the following documentation . calculation of the index A is repeated for all draughts at a
Initial data: new trim.

- subdivision length Ls The resulting initial conditions are listed in a table like
- initial draughts and the corresponding GM-values Table 7-1.
- required subdivision index R
- attained subdivision index A with a summary table Table 7-1 Initial Draughts, Trim and GM-values
for all contributions for all damaged zones Draught Trim GM
[m] [m] [m]
Results for each damage case which contributes to the ds 8 -0.118 4.16
index A: dp 7 0.179 3.58
dl 6 0.805 2.34
- draught, trim, GM in damaged condition
- damage extension and definition of damage cases
7.2 MINIMUM REQUIRED SUBDIVISION
with probabilistic values p, v and r
INDEX, R
- righting lever curve (including GZmax and range) with
factor of survivability s
The minimum required subdivision index, R, is then
- critical weathertight and unprotected openings with
calculated based on Ls according to Regulation 6-2.1
their angle of immersion7
- details of sub-compartments with amount of in-
Table 7-2 Summary of Required Subdivision Index
flooded water/lost buoyancy with their centres of
gravity. Summary of Required Subdivision Index:
This is a Cargo Vessel, Ls > 100m
7.1 INITIAL DRAUGHTS, TRIM AND GM-
VALUES Index R = 0.5138
Min Apartial = 0.2569
Three initial loading conditions are to be used for
calculating the index A. The loading conditions are 7.3 MAIN PARAMETERS
defined by their mean draught d, trim and GM.
Main parameters for p(x1, x2) can be calculated
The light service draught (dl) represents the lower according to Regulation 7-1.1.1.
draught limit of the minimum required GM curve. It
corresponds, in general, to the ballast arrival condition
with 10% consumables for cargo ships.

The partial subdivision draught (dp) is the light service

draught plus 60% of the difference between the light
service draught and the deepest subdivision draught.

7
Regulation 7-2.1, 7-2.5.2 and 7-2.5.3
The Damaged Ship, London, UK

Table 7-3 Calculated Main Parameters 7.5 DEVELOP REQUIRED (OR

Parameter Value Unit ALLOWABLE VCG) CURVE
Ls 111.26 m O
Jmax 0.303 nce the attained subdivision index is satisfactory a
Jkn 0.152 required versus draft curve is developed.
pk 0.917 Alternatively, allowable VCG can also be plotted.
lmax 60.00 m
L* 260.00 m Required at ds = - VCG = 4.16 m (at 8.0 m
b0 11.00 draft)
Jm 0.303
Required at dl = 3.58 m (at 6.0 m draft)
Jk 0.152
b12 11.00 Required at dp = 2.34 m (at 7.0 m draft)
Jm* 0.000
Jk* 0.000 A straight line connecting these two points is the required
b11 -65.34 curve for damage stability. For all drafts below the
b21 -7.26 partial draft, the required at the partial draft is
b22 2.20 assumed to apply. If this curve is not acceptable the
ldamage 33.72 m VCGs can be increased as long as the attained index
exceeds the required index.
7.4 CALCULATE THE ATTAINED
SUBDIVISION INDEX, A
8. SUMMARY AND CONCLUSION
Once the pi, ri, vi and si values are calculated for each
zone or subzone, these values are multiplied. All probabilistic methodologies are based on a thorough
analysis of the vessel's response to damage or flooding
The resulting products for each subzone where single and multiple compartments are assumed to
are summed to determine the attained subdivision index be flooded one at a time and in combination. A range of
at the deep, partial and light drafts. The final attained damage extents is considered, where higher probability
subdivision index is the average of the attained indices at of damage is generally assigned to lesser extents of
the three drafts. damage.

= 0.4 + 0.4 + 0.2 The SOLAS 2009 does not stipulate how the ship should
be subdivided. Instead the performance of a proposed
Ac = partial attained subdivision index, arrangement is evaluated for typical damage scenarios.
for c = s, p and l
s = subdivision draft loading condition The core components in the probabilistic methodology
p = partial draft loading condition can be summarized as:
l = light service draft loading condition
- Required overall level of survivability acounting for
=
, for c = s, p and l
any foreseeable situation where the ship has lost some
of its watertight integrity (index R).
, = damage case partial attained subdivision index - Distributions describing the degree of survivability
under a specific damage (s):
, = , o The basis is a common format based on three
characteristics of the GZ-curve at equlibrium after
i = damage case number flooding (range, max, heel),
N = maximum case number o Is based on the weighted sum of survivability at
three different loading conditions (ds, dp, dl) that
In appendix, Damage Case Partial Attained Indices does not necessary follow the actual operational
Tables, show an example tables of how the calculation of profile.
the attained index is performed. - Distributions describing damage position and extent
(p, r, v):
Required, R Attained, A Pass / Fail
The products of these probabilities are summed over the
As = 0.2569 0.628 Pass
various possible combinations of flooding which could
Ap = 0.2569 0.649 Pass
occur from a single breach of the hull, and the result is
Al = 0.2569 0.667 Pass called the Attained Subdivision Index, A. If the attained
A= 0.5138 0.644 Pass subdivision index A is greater or equal to required
subdivision index R, the vessel fulfills stability
requirements. Thus, the probabilistic damage stability
 The Damaged Ship, London, UK

rules can be summarised in the very short formula A > R 11. AUTHORS BIOGRAPHY
where A is the attained index and R is the required index.
Jannes Gullaksen is the owner of JG Consultant
Key differences between probabilistic and deterministic Engineers, http://www.jg-consultant.com, and a
rules derive from localised details and not at conceptual Maritime Consultant. He is specialized in naval
level. Hence direct comparisons at concept level could architecture and marine engineering. His previous
produce conflicting results. experience are within naval architecture and marine
engineering, including ship hydrostatics, stability and
The deterministic approach can be classified in a strength.
mathematical modeling language, as one in which every
set of variable states is uniquely determined by
parameters in the model and by sets of previous states of
these variables. Therefore, deterministic models perform
the same way for a given set of initial conditions.

Conversely, the probabilistic approach can be classified

in a probabilistic or stochastic model, where randomness
is present, and variable states are not described by unique
values, but rather by probability distributions.

9. ACKNOWLEDGMENT

The concept of a paper addressing this subject was

initially inspired by my draft manuscript to a book, titled
Naval Architecture: Geometry, Hydrostatics, Stability,
Dynamics, Strength and Economics - A Computer Based
Approach Using Excel Spreadsheets with VBA.

10. REFERENCES

[1] International Maritime Organization (IMO), SOLAS

Consolidated Edition 2009.

[2] International Maritime Organization (IMO),

Resolution MSC.216 (82), Chapter II-1 Construction-
Structure, Subdivision and Stability, Machinery and
Electrical Installations.

[3] International Maritime Organization (IMO),

Resolution MSC.281 (85), Explanatory Notes to the
SOLAS Chapter II-1 Subdivision and Damage
Stability Regulations.

[4] International Maritime Organization (IMO),

MSC.1/Circ.1226 15 January 2007, Interim
Explanatory Notes to the SOLAS Chapter II-I
Subdivision and Damage Stability Regulations.

[5] Gullaksen, J., JG Consultant Engineers (JGCE).

Naval Architecture: Geometry, Hydrostatics,
Stability, Dynamics, Strength and Economics - A
Computer Based Approach Using Excel Spreadsheets
with VBA, Witherby Seamanship, to be Published
2011/12.
The Damaged Ship, London, UK

12. APPENDIX A Excel Spreadsheet

A numerical method is implemented in a VBA script in an Excel spreadsheet. The spreadsheet with VBA macros,
SOLAS2009_JGCE.xlsm8, includes the inputs with main parameters and calculates the p-factor probabilities of a
damage case occurring.

Calculate
pi factor

Additional commercial software is necessary to calculate the probability of surviving a particular damage case for
calculation of the attained safety index.

8
Available on http://www.jg-consultant.com, Download section.
The Damaged Ship, London, UK
The Damaged Ship, London, UK
CONSIDERATION OF DAMAGE TO SHIPS FROM CONCEPTUAL DESIGN TO
OPERATION: THE IMPLICATIONS OF RECENT AND POTENTIAL FUTURE
REGULATIONS REGARDING APPLICATION, IMPACT AND EDUCATION
K W Hutchinson, Babcock International Group, UK
A L Scott, Maritime and Coastguard Agency, UK
P N H Wright, M D Woodward and J Downes, Newcastle University, UK
SUMMARY
This paper considers the recent and potential future developments of damage stability regulations with respect to safe
ship design, development and operation and hence investigating how such regulations and techniques can be utilised in
developing compliant design solutions. Arguments regarding the application of such regulations and techniques will be
presented based on the view points of designers, regulators and educators.
NOMENCLATURE
A Attained index
Aj One compartment local index
Bj Two or more compartment local index
GMTf Fluid transverse metacentric height (metres)
GZ Righting lever (metres)
KGf Fluid vertical centre of gravity (metres)
pi Probability that only one compartment(s) under
consideration are flooded
R Required index
si (Conditional) Probability of survival after
flooding of the compartment(s) under
consideration
sfinal,i (Conditional) Probability of survival in the final
stage of flooding of the compartment(s) under
consideration
1. INTRODUCTION
This paper, from the combined perspective of designers,
regulators and educators, seeks to provide an outline of
the recent history and development of stability
regulations and the associated regulatory bodies in order
to provide background to how the current international
stability regulations are formulated by the IMO and how
these are in turn ratified and implemented by Flag States.
This makes reference to how such regulations are under
constant review and identifies the current status of these
regulations as well as outlining potential future
developments and their implications. The particular
issues associated with RO-RO passenger ships and tanker
damage stability are discussed in detail.
The application of these statutory rules and regulations
and the implications on design and operation are
discussed, particularly with respect to the usage and
effectiveness of information used onboard ship. Greater
appreciation of the application of these regulations is
required in industry and the need to inform new
graduates is identified. The implications of this in terms
of providing the appropriate teaching outcomes through
curriculum and course development are discussed. This
includes an outline of how the teaching of intact and
damage stability at Newcastle University has been
enhanced and included in the ship design coursework.
The mechanisms for promoting such curriculum
development to ensure appropriate inclusion and
treatment of such new topics are also outlined in order to
suggest how this can be improved upon in the future.
2. STATUTORY RULES AND REGULATIONS
2.1 STABILITY REGULATION DEVELOPMENT
This following outline of International Maritime
Organisation (IMO) is presented in order to indicate
where we are in the current cycle with respect to IMO
requirements for intact and damage stability for a wide
range of ships.
The primary source of regulations and guidelines
covering ship survivability is the IMO which up to 1982
was called the Inter-Governmental Maritime
Consultative Organisation (IMCO). It was established in
Geneva in 1948 and met for the first time in 1959. It has
its headquarters in London, formerly in Piccadilly, but
now in a purpose-built building on the Albert
Embankment on the south bank of the Thames. It is a
specialised agency of the United Nations (UN) with 169
Member States and 3 Associate Members; Figure 1
outlines the structure of IMO.
Insofar as stability is concerned the initiating Committee
within IMO is usually the Maritime Safety Committee
(MSC) which meets for eight days twice a year and
covers a wide range of safety-related issues. Once a
course of action is agreed to by the MSC, usually at the
instigation of a paper submitted by a Member State, all
technical matters are devolved to a range of specialist
Sub-Committees. Responsibility for stability usually
resides with the Stability, Load Lines and Fishing Vessel
Safety (SLF) Sub-Committee which meets on average
once a year for one week in London. MSC prioritises the
work of the SLF Sub-Committee by setting time periods
for completion of agenda items.

Figure 1: Outline structure of IMO
Within SLF the individual papers and issues on the
agenda are first discussed in the opening plenary session
in the main conference room then passed on to more
specialized working groups or drafting groups where
issues are scrutinised and discussed in greater detail. The
main objective is usually to decide whether specific
regulations need to be introduced to address the newly
perceived risks identified by the MSC or whether current
regulations, such as Safety of Life at Sea (SOLAS) for
example, need to be amended. The final output from the
working groups is usually in the form either of draft
amendments for approval by the SLF Sub-Committee,
meeting in plenary at the end of the week, or proposals
for further work, research etc. which is usually carried
out by intercessional correspondence groups consisting
of specialist volunteers from Member States or Non-
Governmental Organisations (NGOs). These
correspondence groups are given terms of reference and
are instructed to report back to the next meeting of SLF.
IMO has itself no remit to fund research projects but
does encourage Member States to initiate investigations
which are then usually undertaken by academic
institutions or various establishments, businesses or
organisations within the maritime community. Once
SLF has agreed on an appropriate course of action,
whether it is new or amended regulations, resolutions,
circulars etc. or further deliberations, a final report is
prepared for consideration at the next MSC meeting
where, if acceptable, it forms the proposals for new
legislation etc. are agreed to then published with a
specific date for their entry into force.
At this stage it used to be the case that the individual
Flag States would then implement the latest IMO
regulations etc. into their own legal systems. In the
United Kingdom (UK) this takes the form of Statutory
Instruments (SIs) and Marine Guidance Notices (MGNs)
with further Instructions for Surveyors etc. as appropriate
issued by the Maritime and Coastguard Agency (MCA).
With the advent of the European Union (EU), however,
European member countries must also now consider the
EU Directives. Some of these are straightforward
transmissions of IMO regulations but many are written to
cover classes or sizes of ship, such as domestic passenger
ships, which are not part of IMOs remit.
The Damaged Ship, London, UK
2.2 CURRENT INTERNATIONAL STABILITY
REGULATIONS
2.2 (a) Intact Stability
A new Code, the International Code on Intact Stability,
2008 (2008 IS Code) [1, 2], is now in force for passenger
and dry cargo ships of length greater than or equal to 24
metres with keels laid on or after 5th December 2008 and
supersedes Resolution A.749(18) [3, 4] which itself
superseded the original IMO standards [5 - 8].
Accompanying the 2008 IS Code is a set of Explanatory
Notes contained in IMO document MSC.1/Circ.1281 [2].
The 2008 IS Code covers the intact stability of: cargo
ships, including those carrying timber deck cargoes and
deck containers: container ships; passenger ships; fishing
vessels; special purpose ships (SPSs); offshore supply
vessels (OSVs); mobile offshore drilling units (MODUs)
and pontoons. It contains a section including mandatory
criteria, now incorporating the so-called weather
criterion [8] and special criteria for certain types of ships
such as passenger ships, oil tankers of more than 5000
tonnes deadweight, timber deck cargo ships, grain
carriers and high speed craft (HSC).
In fact the 2008 IS Code itself only contains specific
stability criteria for passenger ships (covering heeling
due to passenger crowding and high-speed turning) and
cargo ships carrying timber deck cargo. The mandatory
intact stability criteria for the other ships listed are
covered by their own specialized Codes:
Tankers under Regulation 27 of Annex 1 to The
International Convention for the Prevention of
Pollution from Ships (MARPOL) 73/78 [9].
Grain carriers under the 1991 International Code for
the Safe Carriage of Grain in Bulk (International
Grain Code) [10].
High speed craft under the 2000 HSC Code [11].
Fishing vessels, pontoons, container ships greater than
100 metres, offshore supply vessels, special purpose
ships and mobile offshore drilling units are subject to
non-mandatory guidelines reflecting the specific stability
characteristics of these ship types.
The Damaged Ship, London, UK
2.2 (b) Damage Stability
The latest version of SOLAS, contained in the 2009
Consolidated Edition [12 - 14], covers damage stability
for passenger ships of all sizes and dry cargo ships with
length greater than or equal to 80 metres with keels laid
on or after 1st January 2009. Certain cargo ship types
are excluded from complying with the probabilistic
damage stability regulations in Part B-1 if they comply
with the damage stability requirements included in the
following instruments instead:
Tankers (except combination carriers with Type B
freeboards) under Annex 1 to MARPOL 73/78 [9];
Chemical carriers under the 2007 International Bulk
Chemical (IBC) Code [15];
Liquefied gas carriers under the 1993 International
Gas Carrier (IGC) Code [16];
Offshore supply vessels under the 2006 Offshore
Supply Vessel (OSV) Code [17] which superseded
[18];
Special purpose ships under the 2008 Special Purpose
Ship (SPS) Code [19] which superseded [20];
Ships with reduced freeboards under the International
Convention on Load Lines (ICLL) 1966/1988 [21]
Regulation 27, except those carrying deck cargos;
Mobile offshore drilling units under the 2001 Mobile
Offshore Drilling Unit (MODU) Code [22] the
inclusion of this exemption is under review at present
SLF January 2011 meetings.
It should be noted that roll-on/roll-off (RO-RO)
passenger ships constructed on or after 1st January 2009
and operating in EU waters must for the time being
continue to comply with the Stockholm Agreement
originally invoked under a Regional Agreement
facilitated by IMO Resolution 14 of the 1995 SOLAS
Conference and now incorporated EU Directive
2003/25/EC, as amended [23, 24].
This latest version of SOLAS [12 - 14] harmonizes
previous regulations for passenger ships and cargo ships,
SOLAS90 [25, 26] and Resolution A.265(VIII) [27], into
one set of damage stability regulations based principally
on a probabilistic approach.
2.3 PROPOSED UPDATES TO REGULATIONS
2.3 (a) For Discussion at SLF 53
The intact and damage stability regulations are under
continuous review by the IMO through the MSC and
SLF Sub-Committee, as explained in Section 2.1. An
insight as to the scope of this process can be gleamed
from some of the agenda items for the latest meeting of
SLF in January 2011:
Development of new generation intact stability
criteria - The aim is to expand and refine the scope of
the 2008 IS Code [1] by collecting additional
methodologies for vulnerability criteria and direct
stability assessment, trying to apply the appropriate
criteria depending on whether ships are of standard or
novel layout and investigating assessment procedures
for various stability failure situations such as
quartering seas, parametric resonance, dead ship
condition and broaching.
Guidelines to enhance the safety of small fishing
vessels - This details the development of
enhancements to new guidelines governing the safety
of small fishing vessels, including intact stability.
Standards on time-dependent survivability of
passenger ships in damage conditions - The objective
is to examine the behaviour of damaged ships, in
particular the rate of flooding in varying sea states
and damage scenarios, with the ultimate aim of
increasing the ratio of time to evacuate to time to
capsize. The bench-marking of various numerical
analysis techniques for assessing damage and
flooding situations is also an aim of this work.
Stability and seakeeping characteristics of damaged
passenger ships in a seaway when returning to port by
own power or under tow - A new regulation is being
added to SOLAS [12] introducing a requirement for
the provision of guidelines for the master and loading
officers of large passenger ships in the event of
damage.
Guidelines for verification of damage stability
requirements for tankers and bulk carriers - Concern
has been growing that the damage stability of ships
carrying liquids is not being correctly assessed prior
to departure. A new set of guidelines advocating the
use of onboard computers is being produced.
Safety provisions applicable to tenders operating
from passenger ships - Tenders used to carry
passengers from cruise ships on shore trips are
increasing in size and concern is growing that safety
levels may be insufficient as the passenger capacity
increases.
Review of damage stability regulations for RO-RO
passenger ships - Extensive research work is
currently underway to confirm that the latest SOLAS
2009 regulations [12] make sufficient provision for
the water-on-deck (WOD) phenomenon and to
ensure that freight RO-RO and passenger (RO-PAX)
ships fitted with long lower holds (LLHs) have
sufficient reserve buoyancy in the event of deeply
penetrating side damages.
Legal and technical options to facilitate and expedite
the earliest possible entry into force of the 1993
Torremolinos Protocol - Fishing vessel losses are still
a major concern and it is hoped that implementation
of the Torremolinos Protocol [28] will improve safety
levels.
Amendments to SOLAS chapter II-1 subdivision
standards for cargo ships - There is a question as to
whether offshore supply vessels [17] should comply
with the probabilistic damage stability requirements
of SOLAS 2009 [12] Part B-1 or with the
deterministic requirements of 2006 OSV Code [17]
which cover a penetration depth of only 760 mm or
with a combination of both requirements (see Section

1.2 (b) above). It has also been suggested by UK at
MSC 85 that vessels complying with the 2001
MODU Code [22] need not comply with SOLAS
2009.
Revision of SOLAS chapter II-1 subdivision and
damage stability regulations - During the process of
developing the new SOLAS 2009 amendments [14]
and the accompanying Explanatory Notes [13] there
were many loose ends which need tidying up or
clarifying. This is being done over a two year period,
in time for the next scheduled update to SOLAS. It
includes inter alia issues such as improvements to the
presentation of the limiting GMTf curves, possible
provision of an allowance for the formation of
multiple free surfaces during stage-flooding,
particularly in large passenger ships etc. In all some
sixty potential changes are being discussed, however
many are of a minor editorial nature.
Consideration of IACS unified interpretations - This
is an ongoing agenda item where the International
Association of Classification Societies (IACS)
advises IMO on its agreed interpretation of
Regulations or Explanatory Notes which have been
queried by industry.
In the context of this paper the debates at IMO on RO-
RO and tanker damage stability are considered to be the
most relevant and hence some of these issues are
discussed further.
2.3 (b) Issues with RO-RO Passenger Ships
For the purposes of this paper, the full methodology for
calculating damage stability probabilistically for RO-RO
passenger ships is not presented as it is taken as prior
knowledge. The required index R is set at the same level
as for conventional passenger ships, as is the so-called
sfinal,i factor which accounts for the probability of survival
after flooding the compartment or group of
compartments under consideration. The formulae for R
and sfinal,i were determined by regression analysis using
results from calculations and model test experiments for
existing ships with the objective of ensuring that the new
probabilistically based amendments for passenger ships
would offer the same level of safety as SOLAS90 [25,
26]. The level of safety for passenger ships under
SOLAS90 was set by compliance with the floodable
length regulations (Chapter II-1, Part B, Regulations 4-7)
and the deterministic damage stability requirements of
Regulation 8. Specific damage stability requirements for
RO-PAX ships were added to SOLAS90 in the wake of
the losses of the Herald of Free Enterprise and Estonia
in Regulations 8-1 and 8-2 but were only applied
regionally, first in North-West Europe (NWE), then later
in the rest of Europe under EU Directive 2003/25/EC,
commonly known as the Stockholm Agreement [23,
24].
It appears from subsequent research that the new SOLAS
2009 amendments [12 - 14] have not made any explicit
The Damaged Ship, London, UK
allowance for the WOD phenomenon; the equivalence
has been set to that provided by SOLAS90 only and not
SOLAS90 with the requirements of the Stockholm
Agreement. It is not that the regulators were unaware of
this potential anomaly as the Stockholm Agreement
was only ever applied in Europe so presumably the rest
of the world always believed that the dangers of the
WOD phenomenon on RO-PAX ships were peculiar to
the sea conditions and high traffic density of European
sea-routes. Equally, most European nations considered
that one formula for sfinal,i could be used for both
conventional passenger and RO-PAX ships. Various
methods of incorporating some probabilistic allowance
for WOD into si were proposed but test calculations
seemed to show that they made little difference to the
attained index, A. In addition, the damage statistics used
as the basis for the SOLAS 2009 amendments showed
that most collisions have historically occurred in
relatively calm waters, therefore reducing the possibility
of water encroachment onto the vehicle deck through the
damage opening due to wave action.
However, there were concerns expressed in some
quarters that this was not appropriate, hence some
research projects were initiated to design new SOLAS
2009 compliant ships. The outcome was that there may
indeed be a problem with the sfinal,i factor for RO-PAX
ships. A further problem emerged from this research
suggesting that it is possible to design a RO-PAX ship to
meet SOLAS 2009 with a LLH protected with only B/10
longitudinal bulkheads, rendering it particularly
vulnerable to side damage leading to rapid sinkage, as
opposed to capsize. The underlying problem appears to
be a lack of reserve buoyancy, formerly covered by the
floodable length regulations of SOLAS90, and hence
may necessitate revision of both the sfinal,i and R factors.
For obvious reasons, the loss of the margin line non-
immersion criterion is of much more significance for
RO-PAX ships than conventional passenger ships so it
is difficult to understand how in future these two very
different designs of passenger ship can both be covered
by the same probabilistic damage stability formulae. It
may even be necessary to make a distinction in the
regulations between LLH and non-LLH RO-PAX ships if
the critical loss mechanisms are different, namely
sinkage or capsize.
New research is now underway to explore these issues
further and to identify proposals for amending the
regulations for RO-PAX ships. As noted in Section 1.2
(b) above, the EU has for now insisted that new SOLAS
2009 RO-PAX ships operating in EU waters must
continue to comply with the Stockholm Agreement.
For the time being this will almost certainly prevent LLH
ships with B/10 longitudinal bulkheads from being built
for operation in EU waters.
Inherent within the philosophy of the probabilistic
concept [14, 16 and previous 25 27], for both passenger
and dry cargo ships, is that two different ships may have
The Damaged Ship, London, UK

the same overall global index (of safety with respect to
flooding) but may have quite different actual capabilities
for withstanding hull damage over their length.
Therefore, the authors believe that, in addition to the
level of overall subdivision, there must be a basic
requirement regarding distribution of survivability along
the ships length so that no part of the ship is left
unacceptably vulnerable to flooding. It is suggested that
in order to maintain the philosophy of assessment using
the probabilistic concept, rather than a prescriptive
deterministic assessment, that this is facilitated by the
application of so-called local (partial) indices of
subdivision. It is suggested that two compartment local
indices, Bj, are considered as these are representative of a
worst reality, but also that one compartment local
indices, Aj, or minor damage indices are also
considered as such damages are probable.
As mentioned in Section 2.3 (a), in large passenger ships
etc. with multiple decks there is the possibility of the
formation of multiple free surfaces during stage-flooding
which, in some cases, can lead to the loss of the ship.
2.3 (c) Issues with tankers
The problem with tanker damage stability lies with the
complexity of dealing with fluid loss, so-called fluid
deadweight drop-out, from a damaged tank which may
be filled to any level and carrying liquids with a wide
potential range of specific gravity (SG). If the SG of the
liquid in a particular damaged tank exceeds that of
seawater then the ship could list away from the damaged
side in certain circumstances. When this is allied to an
extensive variety of loading patterns with empty, part-
filled or full tanks and multi-compartment side or bottom
damages it is found that demonstration of compliance
with the damage stability criteria for all potential damage
scenarios required in regulations such as MARPOL [9],
SOLAS [12], the IBC [15] and IGC [16] Codes etc. is all
but impossible other than through use of an onboard
computer due to the complexity and subsequent volume
of information required to adequately convey the damage
performance of the ship to the Master.
stability critical limiting KGf or GMTf curves is all that is
required! To some degree this misunderstanding must be
due to the fact that many, if not most, tanker SIBs
include a critical limiting KGf curve that is not a
combined curve but one purely covering intact stability!
The question must be posed as to why such obviously
misleading and only partly correct information is
routinely included in what is arguably the most important
operational manual onboard a ship and how such content
is approved by the relevant authorities?
This demonstrates that there is a general lack of
awareness of the complexity of damage stability
calculations involving fluid loss (fluid deadweight drop-
out), which is surely an issue that needs to be addressed
by informing through ongoing education (continuing
professional development, CPD) of naval architects and,
most importantly, seafarers. Additionally, such insight
needs to be incorporated in to education of new naval
architects and deck cadets / officers.
Even at IMO there is considerable divergence of opinion
on this issue. Some argue that there is no substantive
evidence to support the claim that one third of the
tanker fleet currently poses an unacceptable risk to life at
sea and to the environment [30]. Other delegations feel
that there is an urgent need to develop guidelines for the
verification of damage stability requirements for tankers
since ships have been shown to regularly sail in
conditions of loading significantly different from those in
the approved stability information booklet [30].

For some ships, such as crude oil tankers, the loading

pattern, cargo SG and tank filling levels are often fairly
predictable and mean that the damage stability can be
verified against pre-approved loading conditions in the
Stability Information Book (SIB, commonly referred to
as the Trim and Stability Book) with some confidence.
However, as soon as any tanker loading condition
deviates significantly [29] from these fixed conditions,
which is quite common for parcel tankers, products
carriers and chemical tankers, then strictly speaking a
new damage stability verification calculation is needed
which should be approved by the certifying authority
before the ship sails. In practice this process is typically
not being followed. It is worrying to report that Port Figure 2: Typical intact and damage critical envelope
State Control (PSC) inspections have shown that many KG curves for a (MARPOL) tanker
loading officers believe that compliance with the intact (deadweight drop-out and zone exceptions)

Figure 2 shows a group of critical limiting envelope KGf
curves for a MARPOL [9] crude oil tanker for one
assumed degree of filling (deadweight available for
loss). It can be seen that the figure is complex and
congested as it contains both intact and damage critical
limiting curves covering the operational trim range of the
ship. The damage results are only for one assumed level
of deadweight available for loss, namely both cargo oil
and water ballast tanks empty hence actually no
deadweight available for loss. The results cover all cases
of side, bottom and raking damage together with
exception cases required to allow the developed loading
conditions to be proved compliant against the combined
Figure 3: Typical generic CCKG curves for a somewhat akin to current procedures on large container
(MARPOL) tanker (with dwt. drop-out)
The Damaged Ship, London, UK
(intact and damage) critical limiting KGf curves, rather
than just by direct calculation. The identification of
such exception cases is a complex and time consuming
process and highly dependent upon the range and extent
of loading conditions developed. Other similar groups of
curves would be required to adequately cover the fill
range of the cargo oil and water ballast tanks; a total of
nine for 0% / 50% / maximum fills, and twenty-five for
0% / 25% / 50% / 75% / maximum fills.
Obviously, the use of such curves is impracticable and
hence it is usual to deconstruct them and group them in,
say, damage specific combined critical limiting KGf
curves (see Figure 3). However, splitting the curves in to
such groups can result in up to seventy-five separate
curves. This number will obviously be increased again if
consumables such as diesel and fresh water etc. are also
considered.
The curves described above are generic in nature as the
fill levels in each zone can be can be checked
individually and are hence independent of the loading
conditions, which is one way that deadweight drop-out
curves are commonly produced. However, they are still
dependent upon the SG of the cargo and if this varies
then SG specific curves would have to also be
provided.
Hence, the application of any combined critical limiting
KGf curves obviously requires substantial time for
interrogation, interpolation and calculation to ascertain if
the ship is safely loaded.
Onboard computers (Loading / Stability Instruments), or
computers linked to onshore facilities, appear to offer a
ready solution to this problem. Until recently computing
speeds would have been a deterrent but improvements to
hardware and software mean that a proposed loading
condition can now have all the required statutory damage
scenarios verified for compliance with the appropriate
regulations within a reasonably short time, typically less
than 30 minutes. The difficulty will arise as soon as the
computer indicates that one or more damage scenarios do
not comply. Some systems may be able to offer advice
on suitable corrective measures but this could involve
some delays as alternative cargo distributions are
proposed and the scenarios re-checked. Assuming
compliance is achieved the cargo and / or consumables
may then have to be physically re-adjusted and it may be
this possibility which is deterring implementation of such
systems. If the system does not automate corrective
measures the loading officer may have to use trial and
error to ensure compliance; this is time consuming and
hence commercially impracticable with no guarantee of
ultimate success.
One solution could be to use advanced planning for each
loading scenario, involving communication between
loading planners ashore and the loading officer onboard,
The Damaged Ship, London, UK
ships. Of course last minute alterations may be
necessary but, as experience is gained with loading the
ship, it may prove possible to reduce or eliminate sailing
delays. Strictly speaking the arrival scenario should also
be checked before departure as the damage stability
characteristics may change as the ship uses up
consumables during the voyage. Increased care may also
be necessary at the design stage to avoid ships which can
only achieve compliance with the stability criteria
through the use of impractical loading parameters, such
as excessive and undesirable bow trim.
3. APPLICATION OF STATUTORY RULES
AND REGULATIONS
3.1 DESIGN OF SHIPS TO SOLAS 2009
To assess the impact of the new SOLAS2009 regulations
[14, 16] the MCA commissioned the design six new
ships of varying types, namely; panamax cruise liner,
post-panamax cruise liner, RO-PAX ship, small coaster,
vehicle carrier and a feeder container ship. The
background to this project is described in MCA Report
552 [31]. A more detailed description of the design
methods used for new RO-PAX ferries compliant with
SOLAS 2009 can be found in MCA Report 592 [32].
The current [14, 16] and previous [25 27] probabilistic
stability regulations are goal-setting and non-prescriptive
in nature. Hence, it is easy to appreciate that their
application to the design of the subdivision of a ship is
not intuitive. The regulations are orientated towards
evaluation of the adequacy of the subdivision
arrangement through the calculation of the associated
overall (global) attained, A, index of subdivision value.
Hence, it is not immediately clear what design rule(s)
may be used to configure appropriate longitudinal,
transverse and / or horizontal subdivision arrangement
except by trial and error or for example the application of
the now deleted (for passenger ships) concepts of margin
line and floodable length. Considering this, one solution
is the use of local (partial) indices of subdivision in
order to suggest subdivision and also, importantly,
endeavouring to ensure uniform survival capability along
the length of the ship. Obviously, there can be one, two
or more compartment local indices of subdivision. One
compartment local indices, Aj, or minor damage indices
should be considered as such damages are probable.
Two compartment local indices, Bj, are similar in extent
to the damage cases within previous SOLAS90 [25, 26]
deterministic regulations and are akin to the requirements
of Regulations 8 of SOLAS 2009 [14, 16] for passenger
ships for minor damages, and hence more
representative of a worst reality hence are also deemed
applicable for analysis.
3.2 INFORMATION CURRENTLY PROVIDED
ONBOARD
For intact stability a ship must as a minimum carry an
approved SIB. A loading computer (Loading / Stability
Instrument) is optional and, if fitted, must conform to the
requirements of 2008 IS Code Part B Chapter 4 [1]. In
addition there is now Revised Guidance to the Master
for avoiding dangerous situations in adverse weather and
sea conditions in MSC.1/Circ.1228 [33] with advice on
avoiding broaching, parametric rolling etc.
The current damage stability requirements for SOLAS-
compliant passenger and dry cargo ships are to be found
in Chapter II-1 Part B-1 Regulation 5-1 and Part B-4
Regulation 19 [12]. The latter refers to MSC.1/Circ.1245
Guidelines for Damage Control Plans and Information
to the Master [34]. There are supplements to the
SOLAS requirements in the Explanatory Notes [13],
particularly in its Appendix.
There is currently considerable debate within IMO about
the best way to present limiting combined critical KGf or
GMTf data to make it both easy to use and yet reliable. It
is desirable to include the intact and damage (combined)
critical curves on one graph but the question is how best
to achieve this given the differing draught and trim
requirements. For intact stability it is relatively easy to
produce limiting curves for a wide range of draughts and
trims to allow for easy interpolation. For damage
stability commonly this is not the case as illustrated in
Section 2.3 (c) for tankers (MARPOL). The sheer
volume of calculations required dictate that the examined
draughts are limited to three and if the operating trim
range is low, less than 0.5%L, calculations are required
only for level trim at the mid and deep draughts and for
operating trim at the light draught. This makes
interpolation across the working range of draughts and
trims for intact and damage stability virtually impossible.
It is hoped that a solution to this will soon be found
during the work being carried out under the SLF Agenda
Item dealing with updates to SOLAS discussed
previously in Section 2.3 (a).
Figure 4: Typical intact and damage critical envelope
KG Curves and loading conditions for a
(SOLAS) dry cargo ship (bare-boat)

The problems associated with calculating the limiting
combined critical KGf or GMTf for tankers and hence,
due to their potential complexity, their use have already
been discussed in Section 2.3(c) and highlighted in
Figures 2 and 3 above. Figure 4 shows the critical
limiting envelope KGf curves for a SOLAS90 [25, 26]
dry cargo ship. It shows a number of intact stability
critical limiting curves covering the operational trim
range of the ship together with the one damage curve
required by the regulations. It can be seen that in this
case the effect of trim is, relatively, insignificant. The
loading conditions are also plotted and it can be seen that
these are typically below all the lowest curves. Hence,
because of this and in order to simplify the information
presented to the Master in the SIB, the actual combined
critical limiting KGf curve is, in order to eliminate
interpolation over trims etc., a single bounding curve,
Figure 5.
Figure 5: Typical simple CCKG Curve for a (SOLAS)
dry cargo ship (bare-boat, no drop-out)
The agreed standards covering onboard computers
(Loading / Stability Instruments) for stability
calculations, if fitted, are now incorporated into the non-
mandatory part of the IS 2008 Code, Part B Chapter 4
[1]. Three main types of calculations performed by
stability software are deemed acceptable:
Type 1 Software calculating intact stability only
(for ships not required to meet a damage stability
requirements);
Type 2 Software calculating intact stability and
checking damage stability on the basis of a limit
curve (e.g. for ships complying with old SOLAS Part
B-1 [25, 26] damage stability calculations etc.) on
previously approved loading conditions;
Type 3 Software calculating intact stability and
damage stability by direct application of pre-
programmed damage cases for each loading
condition. Direct calculation means calculation of
pre-determined damage scenarios for the current
loading condition based on the actual hull and
compartment model stored in the computer, with the
ship being balanced to obtain the damage stability
characteristics at a range of heel angles leading to
equilibrium.
The Damaged Ship, London, UK
3.3 REVIEW OF INFORMATION CURRENTLY
PROVIDED ONBOARD
Approved SIBs, including damage stability where
appropriate, are the minimum legal requirement. Use of
calculators or computers to check stability remain an
optional recommendation. IMO are currently debating
whether or not a non-purpose built computer used for
general shipboard purposes, but also containing a
spreadsheet used for checking loading conditions,
constitutes a Stability Instrument under 2008 IS Code
Part B Chapter 4 [1] and therefore requiring official
approval.
3.4 CURRENT UNDERSTANDING AND USAGE
OF INFORMATION ONBOARD
It is generally recognized that approved SIBs are seldom
used during routine loading operations, surprisingly even
for intact stability. For damage stability, especially on
tankers, there is even less likelihood that approved
documentation will be used.
To cover all the loading and damage possibilities in
advance can result in substantial volume of
documentation, much of it quite complex to understand.
It is often worrying for Surveyors during Port State
Inspections (PSI) to discover how little some crews are
aware of the dangers of inadequate damage stability.
Many think that if the ship complies with the intact
stability limiting critical KGf or GMTf curve, the damage
stability is automatically taken care of or even that the
ship is completely fail-safe and can be loaded in any way
possible without reference to the limiting critical KGf or
GMTf curves! Even on passenger ships it is found that
there may be a lack of awareness of the significance of
Cross-Flooding Arrangements and how they should be
operated in an emergency or the need to keep Watertight
(WT) doors closed when not in use. There is clearly an
urgent need to inform seafarers of this and hence give
them better appreciation of the importance of safely
loading and operating the ship.
3.5 MODIFICATIONS TO INFORMATION
PROVIDED ONBOARD
3.5 (a) Information Provided
In order in some way to negate the situation discussed
previously in Section 3.3, it imperative that documents
relating to intact and damage stability performance of a
ship, namely; the SIB, Capacity Plan, Calibration Tables,
Freeboard Plan, Damage Control Plan and Book etc., are
compiled and published in such a manner that they are
approachable. Arguably, these are most important
operational documents onboard a ship and hence must be
clear, consistent, concise and easy to use. Therefore,
substantial thought, and hence effort, is warranted in
producing clear and annotated documents rather than just
The Damaged Ship, London, UK
pasting computer output in to a document. Where
appropriate, colour should be utilised in order to facilitate
enhanced and rapid understanding of information.
Information should be kept to a minimum and hence, for
example, the inclusion of cross-curves of stability is
questioned when simplified stability data such as a
combined critical limiting KGf curve is provided.
However, simplification of information should not
reduce its applicability or ability to accurately inform
the Master as to the condition of a ship and hence not
facilitate decisions to be made based on accurate and
reliable information. Therefore, there is a balance to be
struck in deriving and presenting data in an accessible
but comprehensive manner.
Regarding the production of combined critical limiting
KGf curves, this has been discussed previously in
Sections 2.3 (c) and 3.2.
Figure 6: Damage consequence diagram
The design of ships to SOLAS 2009 [14, 16] has been
briefly discussed previously in Section 3.1. There are a
number of challenges in the production of operational
information required by the Master, such as the Damage
Control Plan and Book. Under SOLAS90 [25, 26] the
development of such information was relatively
straightforward, as the damage assessment was
deterministic and the hence the applicable damage
scenarios and subsequent actions readily identifiable.
However, utilisation of the probabilistic approach in
SOLAS 2009 obviously does not readily provide such
definitive guidance to the naval architect and in fact, as
discussed in Section 3.1 for non-passenger ships,
survivability to a prescribed damage, say in way of a
main transverse WT bulkhead, cannot be guaranteed.
Hence, the probabilistic assessment has to be analysed in
depth and / or an additional deterministic assessment
undertaken in order to produce the required information.
Considering the above uncertainty regarding the
capability of a ship to withstand hull damage over its
length, the presentation of a damage consequence
diagram, akin to carpet plots produced for many years
for military ships, is applicable. Such a diagram is
presented in Figure 6. However, their construction and
applicability is effected in a similar way by trim, SG and
fill levels etc. as described previously for combined
critical limiting KGf curves.
3.5 (b) Regulatory Developments
For tankers, new guidelines showing how to comply with
the damage stability regulations are currently being
produced by the SLF sub-committee to supplement the
approved documentation already carried onboard to
comply with legal requirements. It is the view of the
authors that an onboard computer system capable of
carrying out all the required damage stability calculations
for the specific loading conditions in question, departure
and arrival, is essential. It should use the actual hull and
compartment model and fully account for any free
surface effects in undamaged tanks. It is also felt that
such systems should offer recommendations on how to
change the loading condition if any of the damage cases
fail. It should also have an off-line training mode to
train the crew using what-if scenarios.
For passenger and dry cargo ships, the use of onboard
computers depends on the ship type and size. For large
passenger ships carrying thousands of persons the very
best computer system, possibly linked to a shore support
base, should considered be the norm. The
FLOODSTAND project [35] is working on such a
system and prototypes are currently being tested. Many
onboard systems from the major software vendors are
widely used throughout the industry. As size decreases
and passenger and crew numbers are fewer, simpler
systems should be sufficient. For small dry cargo ships a
computer system is probably not necessary and the
approved SIB etc. should suffice.

4. UNDERGRADUATE TEACHING OF
INTACT AND DAMAGE STABILITY
4.1 AIM OF UNDERGRADUATE TEACHING
The purpose of an accredited undergraduate degree
programme is to ensure students are equipped with
industry relevant skills and learning outcomes. The
teaching of intact and damage stability is therefore of
fundamental importance in the teaching of naval
architecture at Newcastle University as it is in any such
comparable programme. This should be accompanied by
an understanding of the technical changes, socio-
economic and regulatory environment for the general
subject area. The teaching of these specific naval
architectural topics though is only one part of the
delivery of a complete and balanced accredited degree
programme curriculum and the resources to these topics
have to be allocated accordingly while still ensuring
students gain sufficient understanding of these extensive
subjects given the complexity of the current and future
regulatory framework discussed previously.
4.2 OVERVIEW OF NAVAL ARCHITECTURE
TEACHING
As background, and to put in context the described
developments, it is useful to provide a brief overview of
the naval architecture modules at Newcastle University:
Stage 1 - The Naval Architecture I module provides
an introduction to basic concepts and principles to
allow numerical calculation of hydrostatic quantities
and intact stability assessment with respect to basic
IMO criteria. This is accompanied by drawing office
and laboratory sessions to develop drawing, problem
solving and experimental skills.
Stage 2 - The Naval Architecture II module develops
further this understanding of trim and stability to
include more complex ship and operationally
specific stability criteria related to: list and loll; shift
of cargo; effect of waves; wind loading; loss of
suspended loads etc. It introduces the topic of damage
and the added mass and lost buoyancy methods. This
is accompanied by a substantial flooding laboratory
exercise where the influence on stability on flooding
of a model is measured and compared against
calculation. It concludes with margin line and
floodable length calculations. In fact in recent years,
until the introduction of the NA III module
subsequently described this was the extent of
teaching on the subject of damaged stability.
Stage 3 - In the Ship Design module, the naval
architecture knowledge developed is applied in the
context of a substantial ship design exercise where
the intact stability of the design ship in the arrival and
departure conditions is assessed using commercial
ship design software.
The Damaged Ship, London, UK
4.3 RECENT TEACHING ENHANCEMENTS
There have been a number of substantial changes to the
teaching of stability; implemented to improve
understanding of damage stability and introduce for the
first time the topic of probabilistic damage stability. This
has been achieved by: redesigning of the Flooding
Laboratory; the introduction of Naval Architecture III to
provide a more full treatment of damage stability and
subdivision; inclusion of damage stability in the ship
design coursework; additional guest lectures.
4.3 (a) Flooding Laboratory
The Flooding Laboratory uses a 1.2 metre model afloat
in a dedicated free-standing tank, Figure 7. The students
are provided with a range of weights and a suggested
ballasting plan. Once they have loaded the model to an
appropriate draught and trim, they conduct an inclining
experiment to determine the initial stability. This,
together with the hydrostatic curves provided, gives all
necessary information about the initial condition of the
ship. The students are then guided to flood number six
hold starboard, record the equilibrium condition, and
then cross flood the compartment, again recording the
equilibrium heel and trim. A second inclining
experiment after flooding provides a measure of the
change in stability.
Figure 7: Flooding Laboratory model in dedicated free-
standing tank
The students undertake calculations to compare with the
measured results for both the symmetric and asymmetric
damage cases; the required calculations include using the
added mass method to estimate the equilibrium condition
after flooding starboard hold six and the lost buoyancy
method after cross-flooding.
It is important to ensure that the students achieve the
desired learning outcomes to provide appropriate
understanding of these fundamental methods and it was
found when working in teams to provide one report this
was not always the case. Additionally any errors in these
calculations would become cumulative and invalidate
any subsequent calculations resulting in a lack of
correlation between the measured and calculated values
so that again the intended learning outcomes were not
The Damaged Ship, London, UK

achieved. To ensure appropriate understanding by all 4.3 (d) Guest Lectures

students this was changed to students preparing
individual reports but aided by a calculation proforma.
Students are guided to perform the necessary calculations
and to write down their working in the boxes provided.
These are accompanied by certain key values at the end
of individual sections. In this way the students can, as a
minimum, continue with the next section using the
correct value and, for the more conscientious students, go
back and rectify their mistake.
The outcome of this is the belief that students now better
achieve the desired learning outcomes and are able to
calculate the trim and stability in the flooded condition.
4.3 (b) Introduction of Naval Architecture III
The second major change to the curriculum is the
introduction of a new module to address more advanced
stability issues at Stage 3. Specifically, an existing half
module that covered the subject of Surface Generation
and Fairing was enlarged to a whole module to give the
opportunity to include the new material. This module
therefore deals collectively with the representation of
hullform and the related design of the internal
subdivision.
A further opportunity to enhance understanding of these
topics is planned through invited guest lectures as part of
the recent Future Marine Projects module. In particular
to invite naval architects working within the regulatory
framework and in design to share their experiences and
insights to further augment the fundamentals of the
subject taught in other modules.
4.4 FACILITATING CURRICULUM
DEVELOPMENT
As has been discussed, curriculum development to
address the need to ensure that all naval architecture
degree programmes include sufficient content to provide
graduates with the required domain specific learning
outcomes to ensure professional competence is extremely
important. The means of informing this development are
outlined in Figure 8 and include the combined influences
of industry, research, teaching and Professional
Institutions. The principal means of achieving this are
outlined next.

The concepts of damage in terms of loss of watertight

integrity, and possible means of this occurring, and
extent of damage are described. The concepts of margin
line and floodable length are revisited as a reminder to
students, and then the issues of trapped buoyancy and
cross flooding are introduced. Tanker subdivision and
damage stability is outlined discussing regulations,
typical tanker arrangements, limitations on cargo tank
size and the reasons for these requirements. The
hypothetical oil outflow model [9] is developed as a
means to seeing how these concepts are applied in the
design of tankers. The SOLAS 2009 harmonised
probabilistic regulations [12 - 14] are then introduced,
firstly giving an overview of the regulations and their
concepts, then developing the pi factor and the si factor
Figure 8: Influences on Curriculum Development
components before finishing with an outline of how the
regulations are used in practice.
4.4 (a) Industrial Advisory Boards
4.3 (c) Damage Stability and Ship Design
Industrial Advisory Boards, bringing together industrial
practitioners and academics to review programme
Because of the introduction of NA III module it is
content, are an established and important means to
intended to make a third change to the curriculum. It is
facilitate curriculum development. Industrial Advisory
proposed to enhance the Ship Design coursework [36,
Boards require industry representatives to attend and
37] to now cover the actual calculation of damage
with increasing time pressure and less local industry to
stability performance on the students designs. This will
draw upon there is a danger that such boards are not so
be implemented for the beginning of the next academic
frequent or effective as they once were.
year together with an update to the latest 2011 version of
the commercial ship design software used. The authors
4.4 (b) Collaborative Research
hope to present and share with others their experiences of
this major advance at a suitable forum within the next
In research active departments engaged in collaborative
year.
research and / or those engaged with industry through
consultancy to address focussed industry problems, such

collaboration will provides insight and solutions to
contemporary topics and design issues. In turn these
research topics and findings will inform the curriculum
and module content of the particular subject area as well
the overall programme. This has become an increasingly
important mechanism for curriculum development. In
turn these changes influence industries adoption of new
approaches and techniques as understanding of them
matures and they becomes more accepted knowledge
through their inclusion in the teaching of future
graduates.
4.4 (c) Industrial engagement with teaching
Industry engagement with university teaching through
guest lecture programmes to make students aware of
current demands in the industry is another important
means of contributing to programme content. Industries
contribution to teaching content and methods can be
invaluable, as has been demonstrated with the recent
Student Friendly Software Workshops [38]. Industrial
collaboration can also provide more effective
implementation of new concepts and approaches to the
teaching syllabus, such as has been the case with the Ship
Design module over the past decade [36, 37]. To provide
this support the most effective way was found to be to
bring together the requirements of teaching with the
experience of the software developer and an expert
industry user, Figure 9. This synergy has provided the
fundamental basis for the success of using commercial
ship design software in the ship design coursework of the
Ship Design module and has resulted in a number of
teaching strategies being developed as well as changes to
the curriculum.
Figure 9: Newcastle approach to providing Student
Friendly Software Support
In addition to such engagement with teaching
development, industry can also provide invaluable input
by sharing their experiences directly to students, as
The Damaged Ship, London, UK
through the common lecture series Future Marine
Projects previously mentioned that is almost entirely
made up of guest lecturers from industry; including the
authors of this paper. Contributions are invited on any
technical topic that is relevant to their company and
experience that reflects the contemporary demands of
projects in which they are involved and have practical
experience of across all Marine Technology disciplines
(Naval Architecture, Marine Engineering, Small Craft
Technology and Offshore Engineering). This allows
interdisciplinary interaction and sharing of ideas as well
as making students aware of current trends in the
industry that they may be involved with on graduating.
4.4 (d) Accreditation
The most formal mechanism is accreditation of degree
programmes by accrediting professional institutions. In
the case of RINA, to either achieve solely RINA
accreditation or as also meeting the exemplary academic
requirements for either IEng or CEng registration with
the EC(UK) [39]. Accreditation provides a basis for
educational establishments to review their programmes
and ensure excellence in delivery and content. It is
centred on learning outcomes, including teaching and
learning processes, and the required skills, knowledge
and understanding for professional competence that
engineers need to register in a particular category. The
Professional Institutions are licensed to undertake
accreditations to assess whether these requirements are
met and are allowed to interpret them as appropriate to
their own specialisation.
After initial assessment of the documentation prepare in
support of an accreditation submission, an accreditation
panel is formed normally comprising of academics,
industrial members and the institutions secretariat. The
domain specific content of the curriculum is principally
assessed by the academic members of the accreditation
panel who are familiar with the delivery and content of
such courses but there is also opportunity for industrial
members to comment on the content and the programmes
as a whole. This process relies on the panel members
being familiar with industries demands and
contemporary topics if there are to be suggestions for
curriculum development.
Although the current accreditation process does provide
opportunity to share good practice in terms of teaching
methods and course content, further opportunities to
share such best practice within the marine sector could
perhaps also be achieved through further activities of
RINA and sister professional institutions in addition to
the published technical literature and hosting of
conferences. This could provide another way of
responding to the changing demands of the industry in
terms of the knowledge, skills and learning outcomes
graduates require for professional competence; this could
perhaps be done through more pedagogic focused
seminars and conferences to share best practice.
The Damaged Ship, London, UK
5. CONCLUSIONS
This paper has provided an outline of the recent history,
status and potential development of stability regulations.
The application of such statutory rules and regulations
and the implications on design and operation of ships has
also been discussed and some of the complexities
associated with their implementation highlighted.
It is very apparent that continual development of such
rules and regulations poses a challenge for naval
architects and seafarers alike in appreciating their impact
on the design and operation of vessels. Providing up to
date education to include such developments is also
demanding if their implications are to be fully
appreciated.
It is hoped that a number of useful insights reflecting the
experiences of the authors have been presented and that
these will be of benefit to those involved in the practical
application of such intact and damaged regulations to the
design and operation of ships governed by SOLAS 2009.
It is also hoped that the experiences of the authors in
teaching in this subject area are also of value in terms of
sharing best practice with others involved with the
delivery of undergraduate programmes.
6. DISCLAIMER
The views expressed in this paper are those of the
authors and do not necessarily represent those of the
organisations with which they are affiliated and the
professional institutions of which they are members.
7. REFERENCES
1. 2008 IS Code International Code on Intact Stability
2008, 2009 Edition, International Maritime Organization,
London, United Kingdom, 2009. ISBN 978-92-801-
1501-7 (Annex 2 of Resolution MSC.267(85))
2. MSC.1/Circ1281 Explanatory Notes to the
International Code on Intact Stability 2008, International
Maritime Organisation, London, United Kingdom, 2008.
(prepared by the SLF Sub-Committee of the at its 50th
Session 30th April to 4th May , 2007 and adopted by the
Maritime Safety Committee on 9th December 2008)
3. Resolution A.749(18) Code on Intact Stability for all
types of Ships covered by IMO Instruments, Assembly:
Eighteenth Session, 25th October to 5th November 1993,
Resolutions and Other Decisions. International Maritime
Organization, London, United Kingdom, 1994. ISBN 92-
801-1312-7
4. Resolution MSC.75(69) Adoption of Amendments to
the Code of Intact Stability for all types of Ships covered
by IMO Instruments (Resolution A.749(18)), Marine
Safety Committee, International Maritime Organisation,
London, United Kingdom, 14th May, 1998.
5. Resolution A.167(ES.VI) Recommendation on Intact
Stability for Passenger and Cargo Ships under 100
metres in length, Assembly: Fourth Extraordinary
Session, 26th to 28th November 1968, Resolutions and
Other Decisions. Inter-Governmental Maritime
Consultative Organization, London, United Kingdom,
1969.
6. Resolution A.205(VII) Amendments to the
Recommendation on Intact Stability for Passenger and
Cargo Ships under 100 metres in length (Resolution
A.167(ES.IV)) with respect to Ships carrying Deck
Cargos, Assembly: Seventh Session, 5th to 15th
October1971, Resolutions and Other Decisions. Inter-
Governmental Maritime Consultative Organization,
London, United Kingdom, 1972.
7. Resolution A.168(ES.VI) Recommendation on Intact
Stability of Fishing Vessels, Assembly: Fourth
Extraordinary Session, 26th to- 28th November 1968,
Resolutions and Other Decisions. Inter-Governmental
Maritime Consultative Organization, London, United
Kingdom, 1969.
8. Resolution A.562(14) Recommendations on a Severe
Wind and Rolling Criterion (Weather Criterion) for the
Intact Stability of Passenger and Cargo Ships on 24
metres in length and over, Assembly: Fourteenth
Session, 11th to 22nd November 1985, Resolutions and
Other Decisions. International Maritime Organization,
London, United Kingdom, 1986. ISBN 92-801-1204-X
9. MARPOL Consolidated Edition 2006 Articles,
Protocols, Annexes, Unified Interpretations of the
International Convention for the Prevention of Pollution
from Ships, 1973, as modified by the Protocol of 1978
relating thereto, International Maritime Organization,
London, United Kingdom, 2006. ISBN 978-92-801-
4216-7
10. IG Code 1991 Edition International Code for the Safe
Carriage of Grain in Bulk (International Grain Code),
International Maritime Organization, London, United
Kingdom, 1991. ISBN 92-801-1275-9
11. IGC Code 1993 Edition International Code for the
Construction and Equipment of Ships Carrying Liquefied
Gasses in Bulk, International Maritime Organization,
London, United Kingdom, 1993. ISBN 978-92-801-
4226-6

12. IBC Code 2007 Edition International Code for the
Construction and Equipment of Ships Carrying
Dangerous Chemicals in Bulk, International Maritime
Organization, London, United Kingdom, 2007. ISBN
978-92-801-1277-1
13. 2000 HSC Code 2008 Edition Code of Safety for
High-Speed Craft, 2000, International Maritime
Organization, London, United Kingdom, 2008. ISBN
978-92-801-42400 (as amended by Resolutions
MSC.175(79) and MSC.222(82))
14. SOLAS Consolidated Edition 2009 Consolidated text
of the International Convention for Safety of Life at Sea,
1974, and its Protocol of 1988: articles, annexes and
certificates. Incorporating all amendments in effect from
1 July 2009, International Maritime Organization,
London, United Kingdom, 2009. ISBN 978-92-801-
1505-5 (note Erratum for Part B-1: Stability Regulation
7-2 (3) on page 47 re sfinali)
15. Resolution MSC.281(85) Explanatory Notes to the
SOLAS Chapter II-1 Subdivision and Damage Stability
Regulation, Maritime Safety Committee, International
Maritime Organisation, London, United Kingdom, 4th
December, 2008.
16. Resolution MSC.269(85) Adoption of Amendments to
the International Convention for the Safety of Life at Sea,
1974, as amended, Maritime Safety Committee,
International Maritime Organisation, London, United
Kingdom, 4th December 2008.
17. OSV 2006 Guidelines for the Design and
Construction of Offshore Supply Vessels 2006,
International Maritime Organization, London, United
Kingdom, 2007. ISBN 978-92-801-14867
18. Resolution A.469(XII) Guidelines for the Design and
Construction of Offshore Supply Vessels, Assembly:
Twelfth Session, 9th to 20th November 1985, Resolutions
and Other Decisions. International Maritime
Organization, London, United Kingdom, 1982.
19. 2008 SPS Code Code of Safety for Special Purpose
Ships, International Maritime Organization, London,
United Kingdom, 2008. ISBN 978-92-801-1495-9
20. Resolution A.534(13) Code of Safety for Special
Purpose Ships, Assembly: Thirteenth Session, 7th to 18th
November 1983, Resolutions and Other Decisions,
International Maritime Organization, London, United
Kingdom, 1984.
21. Load Lines Consolidated Edition 2005 Edition
International Convention on Load Lines, 1966 and
Protocol of 1988, as amended in 2003, International
Maritime Organization, London, United Kingdom, 2005.
ISBN 92-801-4194-5
The Damaged Ship, London, UK
22. MODU Code Consolidated Edition 2001 Code for the
Construction and Equipment of Mobile Offshore Drilling
Units, International Maritime Organization, London,
United Kingdom, 2001. ISBN 92-801-5109-6
23. Statutory Instrument 2004 No.2884 Merchant
Shipping (Ro-Ro Passenger ships) (Stability) Regulations
2004, The Stationary Office, London, United Kingdom,
29th November, 2004. ISBN 0-11-050076-8
24. Merchant Shipping Notice 1790(M) Agreement
Concerning Specific Stability Requirements for Ro-Ro
Passenger Ships Undertaking Regular Scheduled
International or Domestic Voyages between European
Ports, The Stationary Office, London, United Kingdom,
January, 2005.
25. SOLAS Consolidated Edition 1992 Consolidated text
of the International Convention for Safety of Life at Sea,
1974, and its Protocol of 1988: articles, annexes and
certificates. Incorporating all amendments up to and
including the 1990 amendments (1991amendments as an
appendix), International Maritime Organization, London,
United Kingdom, 1992. ISBN 978-92-801-1294-5
26. SOLAS Consolidated Edition 2001 Consolidated text
of the International Convention for Safety of Life at Sea,
1974, and its Protocol of 1988: articles, annexes and
certificates. Incorporating all amendments in effect from
1 January 2001, International Maritime Organization,
London, United Kingdom, 2001. ISBN 92-801-5100-2
27. Resolution A.265(VIII) Regulations on Subdivision
and Stability of Passenger Ships as an Equivalent of Part
B of Chapter II of The International Convention for the
Safety of Life at Sea, 1960, Assembly: Eighth Session,
13th to 23rd November 1973, Resolutions and Other
Decisions. Inter-Governmental Maritime Consultative
Organization, London, United Kingdom, 1974.
28. 1993 Torremolinos Protocol Final Act of the
International Conference on Safety of Fishing Vessels,
1993, with attachments, including the Torremolinos
Protocol of 1993 relating to the Torremolinos
International Convention for the Safety of Fishing
Vessels, 1977 and Consolidated text of the regulations
annexed to the Torremolinos International Convention
for the Safety of Fishing Vessels, 1977, as modified by
the Torremolinos Protocol of 1993 relating thereto,
International Maritime Organization, London, United
Kingdom, 1995. ISBN 92-801-1317-8
29. Paris CIC Tanker Damage Stability Guidance Notes,
Paris MOU on Port State Control, 28th May, 2010, and
Tanker Damage Stability Questions on the CIC, Paris
MOU on Port State Control, 20th August, 2010.
30. SLF 52/WP.5 Draft Report to the Maritime Safety
Committee, International Maritime Organisation,
London, United Kingdom, 29th January, 2010.
The Damaged Ship, London, UK
31. Final Report Research Project 552 Assessment of the
Impact of the new Harmonized Probabilistic Damage
Stability Regulations (SOLAS 2009) on the Subdivision of
new Passenger and Dry Cargo Ships, Maritime and
Coastguard Agency, United Kingdom, 26th November,
2007.
32. Final Report Research Project 592 Investigation into
the Safety of RO-RO Passenger Ships fitted with Long
Lower Holds Phase II, Maritime Safety Agency,
United Kingdom, March, 2009.
33. MSC.1/Circ1228 Explanatory Revised Guidance to
the Master for avoiding dangerous situations in adverse
weather and sea conditions, International Maritime
Organisation, London, United Kingdom, 11th January,
2007. (approved by the Maritime Safety Committee at its
eighty-second session 29th November to 8th December,
2006)
34. MSC.1/Circ1242 Guidelines for Damage Control
Plans and Information to the Master, International
Maritime Organisation, London, United Kingdom, 29th
October, 2007. (prepared by the SLF Sub-Committee of
the at its fiftieth Session 30th April to 4th May, 2007, and
adopted by the Maritime Safety Committee at its eighty-
third session 3rd to 12th October, 2007)
35. Integrated Flooding Control and Standard for
Stability and Crises Management is an EU-funded FP7
research project (Ref SST-2007-RTD-1) due to complete
in 2012-13.
36. WRIGHT, P.N.H., HUTCHINSON, K.W. and
WHITE, G.D.J., Use of TRIBON Initial Design for
Teaching Ship Design. 9th International Marine Design
Conference (Ed. M.G. Parsons), Ann Arbor, Michigan,
United States of America, 2, pp. 699 to 722, 2006.
37. WRIGHT, P.N.H., HUTCHINSON, K.W. and
WHITE, G.D.J., Approaches Developed to Support the
use of Tribon Initial Design Software in the Teaching of
Ship Design at Newcastle University, Proceedings
ICCAS 07, RINA International Conference on Computer
Applications in Shipbuilding, Portsmouth, United
Kingdom, Paper No.16, Volume III, pp. 155 to 168, 18th
to 20th September, 2007.
38. WRIGHT, P.N.H, and BIRMINGHAM, R.W.
Towards Student Friendly Ship Design Software. 9th
International Marine Design Conference (Ed. M.G.
Parsons), Ann Arbor, Michigan, United States of
America, 2, pp. 723 to 733, 2006.
39. The Accreditation of Higher Education Programmes
UK Standards for Professional Engineering
Competence, EC(UK), 2008.
9. AUTHORS BIOGRAPHIES
Keith Hutchinson is a Principal Naval Architect within
Marine Division of Babcock International Group
undertaking marine consultancy at their Tyneside offices
in North Shields. He is responsible for the concept
design and investment appraisal of a range of novel /
bespoke merchant, offshore and naval vessels, both new-
build and conversion, in addition to hydrostatic and
dynamic evaluations of other offshore, commercial and
military designs. He graduated in Naval Architecture
and Shipbuilding from Sunderland Polytechnic in 1990
and from the University of Newcastle upon Tyne in 1992
with a BEng in Marine Technology with First Class
Honours in Naval Architecture. Having worked in a
number of yards on the Tyne, he joined Swan Hunter
Shipbuilders in 1990 and in 1992 was appointed Senior
Designer within the Naval Architecture Section of the
Basic Design Department. Upon joining his current
employer in 1995 (then Armstrong Technology), he was
seconded to the Newcastle University Engineering
Design Centre for three years as a Research Associate
developing decision support tools for the evaluation,
optimisation and selection of safer RO-RO ship designs
and other marine vehicles. He has published a number of
papers and given many presentations on aspects of ship
design and performance. He is a Chartered Engineer and
corporate member of RINA, IMarEST and SNAME.
Andrew Scott is Policy Lead for Stability, Subdivision
and RO-ROs within the Marine Technology Branch -
Ship Safety Section of the Maritime and Coastguard
Agency based in their Tyne offices in South Shields. He
is responsible for policy matters relating to stability and
is currently joint co-ordinator of the IMO SLF
correspondence group covering the latest updates to
SOLAS 2009 and on issues relating to RO-RO stability.
He has recently led several MCA-sponsored research
programmes into the impact of the 2009 SOLAS
amendments on ship design most specifically on the
issue of Water-On-Deck (Stockholm Agreement). He
graduated from the University of Newcastle upon Tyne
in 1969 with a BSc degree in Naval Architecture and
Shipbuilding and joined Lloyds Register of Shipping as a
trainee ship surveyor, spending seven years in their
London HQ in various departments with a spell of 3
years working in Amsterdam on survey work, principally
on VLCCs. He then joined BSRA (later BMT) in
Wallsend, as a research officer working on a wide range
of consultancy projects, particularly with RO-PAX ships,
and on software development and support. He has
published a joint paper on Ship Vulnerability to
Flooding (IMC on Design and Safety, Berkeley, 2007)
and given a number of presentations on related topics,
most recently at the Stability 2010 Workshop at
Wageningen. He is a Chartered Engineer and corporate
member of RINA.
 The Damaged Ship, London, UK

Peter Wright is Senior Lecturer in Naval Architecture in
the School of Marine Science and Technology at the
University of Newcastle upon Tyne where he has been a
member of the academic staff since 1995. He is
responsible for the teaching of naval architecture and
ship design as both Stream Leader for Naval Architecture
and as undergraduate Degree Programme Director. He is
Chairman of RINA Professional Affairs Committee
concerned with the accreditation of degree programmes.
He is involved in a number of international research
programmes on aspects of warship and merchant ship
design. He graduated in Naval Architecture and
Shipbuilding from Sunderland Polytechnic in 1989 and
from the University of Newcastle upon Tyne in 1991
with a BEng Naval Architecture and Shipbuilding with
First Class Honours. He was sponsored by and worked
for Vosper Thornycroft (UK) Ltd before returning to the
University. He obtained his PhD in 2004. He has been
involved with and led the Newcastle contribution to
several European and national projects concerned with
both naval and merchant vessel design; most notably
EUCLID RTP3.21, INTERMODESHIP, MARSTRUCT,
the recently completed CREATE3S project and the
current DISPRO project with Safinah Ltd. He is a
Chartered Engineer and corporate member of RINA.
Jonathan Downes is a Senior Research Associate in the
School of Marine Science and Technology at Newcastle
University. His principal research interests are in the
development of theoretical methods for predicting the
strength and reliability of a ship's hull girder in both the
intact and damaged conditions. He has undertaken
research into analysis of marine structures funded by
both the EC (ALERT, POP&C, MARSTRUCT,
INTERMODESHIP, MARTOB) and other funding
bodies, and has significant experience of coordinating
multidisciplinary research projects. He graduated from
the University of Newcastle upon Tyne in 2000 with a
BEng in Marine Technology with Upper Second Class
Honours in Naval Architecture. He obtained his PhD in
2003. He has published many papers, particularly on
structural analysis of ships, and regularly undertakes
consultancy for various companies. He is an Associate
member of RINA.

Michael Woodward is a Lecturer in Naval Architecture

within the School of Marine Science and Technology at
the Newcastle University. He is responsible for the
teaching of Naval Architecture, including stability, and is
Stream Leader for Small Craft Technology. He is
Coordinator and Principal Investigator of the EU funded
project AZIPILOT and has led contributions to past
projects including FASTPOD and OPTIPOD. He is
currently secretary of the ITTC Specialist Committee on
Uncertainty Analysis and contributor to the IMO Work
Group on: Subdivision and Damage Stability (SDS)
Correspondence Group; RO-RO Stability. Originally
qualifying as a traditional wooden shipwright in 1987, he
went on to obtain his MEng. in Marine Technology at
Newcastle University in 2000 and PhD. in 2005.
Publishing regularly to the international audience his
general subject is ship maneuverability and ship
handling, with a specialty in azimuthing pod-drives. He
is a Chartered Engineer and corporate member of RINA
and RIN.
The Damaged Ship, London, UK

MOD(UK) SALVAGE AND MARINE OPERATIONS TEAM USE AND DEVELOPMENT

OF SPECIALIST TOOLS AND EMERGENCY PREPAREDNESS TO THE SUPPORT OF A
UK MILITARY MARITIME INCIDENT

S Quinn and N Hills MoD, UK

SUMMARY

This paper presents a brief overview of the organisation of the UK MoD Salvage and Marine Operations Project Team
and highlights some of the detail of the activities behind the front line deployment, exploring how the organisation is
organised, how it will respond and, most importantly, how it is developing both tools, training and experience to provide
an improving response capability. The paper reviews the formal collaboration with other nation salvage responders and
the relationship with commercial organisations.

1. INTRODUCTION organisation with most of its operational personnel

The UK Ministry of Defence Salvage and Marine
Operations Project Team (S&MO) is a support unit to the
UK armed forces operating in the maritime environment.
It is a first responder organisation, with a multi-
disciplinary maritime team with engineering,
seamanship, naval architecture, diving and project
management skill sets.
The team is set up to provide personnel and material to a
maritime incident within defined readiness parameters,
with first elements ready to deploy within 6 hours of the
initiating event. Operations vary from the minor, eg:
HMS GRIMSBY grounding in Norway 2006, HMS
ILLUSTRIOUS stern seal failure 2005, HMS OCEAN
stern ramp failure in Rio 2009) to the major, eg: HMS
NOTTINGHAM grounding in 2002, HMS
ENDURANCE flooding in 2008 and most recently HMS
ASTUTE grounding in 2010. Some of these incidents
have been described in detail in a variety of technical
publications, including RINA1 2.
possessing several key skill sets in addition to their own
core values as a Marine Engineer or Naval Architect for
example. All operational personnel are deployable and
maintain both industry and military training capabilities
and qualifications to support said deployment. A further
initiative is that all operational posts are now badged as
Sponsored Reserve posts and nearly all of the personnel
have subsequently become sponsored reserves. This
offers the flexibility to deploy S&MO personnel in
support of a maritime incident in any scenario from
peacetime evolutions with contractors right through to
hostile military theatres directly alongside military
personnel
The teams responsibilities are enshrined in the MoD
Endorsed Salvage Policy (Table 1), but can be
summarised as the MoD Technical Authority for
Salvage, Blue Water Towing, Heavy Lift, Under Water
Engineering, Specialist Moorings and Wreck
Management.

Whilst delivery of an effective rapid response service to 2. GENERAL RESPONSE TO A MARINE

the scene of the incident is the most visible output of the
team, like all response organisations, there is significant
investment in training, exercising and capability
development within the team, and a careful nurturing of
relationships, both formal and informal, with MoD and
other stakeholders.
The team itself is comprised of nearly 60 MoD civil
servants, working in two marine salvage units (located in
Devonport and Greenock) and a head quarters staff based
in Bath. It is part of the Defence Equipment and Support
organisation and is lead by the Chief Salvage Officer. As
described above the team is very much of a matrix based
1
RINA 2003 Dry Docking HMS NOTTINGHAM; 730 days of
Salvage and Repair Operations N Hills, C Dicks, S Quinn, J R
Ward, MoD UK
2 departments and staff. The organisation also supplies a
RINA 2010 The Damaged Ship - MoD Salvage Response To
HMS ENDURANCE in the Magallan Straits in December 2008 - J
R Ward, M Watts, D Price, MoD UK
INCIDENT
It is well documented that the UK armed services
experience a varying number of maritime incidents.
These incidents may be the result of an accident or as a
result of enemy action, but this is also a natural
consequence of the tempo for training for combat,
Occurrence of such incidents is dramatically mitigated by
rigid training regimens and discipline, but it is a fact of
life that they do occur, and unfortunately shall probably
continue to do so.
Such an environment demands an established and
rehearsed response system, and Navy Command has
developed the Accident and Crisis Response
Organisation (ACRO), which is designed to ensure
effective call out and mobilisation of the appropriate
cradle to grave management profile for the incident and
facilitates the change in overall lead as the situation
matures. The Fleet Incident Response Cell (FIRC), based

in Northwood, will take the initial lead during a response,
then at an appropriate level of stabilisation of the
casualty, shall transfer the lead to the Fleet Coordinating
Authority (FCA), who will predominantly manage the
recovery phase of any casualty. A profile of the call out
hierarchy is presented in Table 2. It should be noted that
this response process goes well beyond that associated
with a purely technical response to a maritime incident.
S&MO has designed its own process to compliment the
ACRO, and will deploy staff as required, both directly to
the incident under the project management of a Salvage
Master, but also ensure representation in the various
responding headquarters, Northwood and Whale Island,
Portsmouth, to act as liaison between the UK based
senior S&MO incident manager and the respective leads
in those head quarters. Additionally, S&MO will stand
up its own Operations Room to coordinate the activity
that it is charged with providing into the overall incident.
3. INTERNATIONAL COLLABORATION
S&MO PT is developing relationships with Other
Government Departments (OGDs) both in the UK and
other nations. This has particularly manifested itself with
the UK Maritime and Coastguard Agency, with a regular
liaison and joint exercises, which in turn has lead to
European exercises under the BONNEX programme.
These exercises include MoD, MCA, UK and French
Emergency Towing vessels, conducting salvage and
towing training serials such as Exercise BLUE WATER
in 2006, and an internationally observed TOWEX in
2009. This kind of liaison and exercise has already borne
fruit, with the UK ETV used in August 2009, being the
same responding tug to the grounding of HMS ASTUTE
in 2010.
The US equivalent of S&MO is NAVSEA 00C,
Superintendant of Salvage (SUPSALV). Whilst
organised along different lines to the UK team, the
outputs and capabilities are broadly similar. Building on
lessons identified from Op TELIC in 2003, where the
UK and US salvage teams worked in close cooperation to
the US Maritime Component Commander, efforts were
made to formalise the relationship under the Master
Information Exchange Agreement between the two
nations3. The particular Memorandum of Understanding
between the two units is primarily focused on Heavy Lift
Transportation, but also reads across into the
development of a joint capability in software use and
management particular to ship salvage. Initially this
agreement focussed on joint training and preparation and
fell short of joint deployment. However, the joint
capability is now maturing, and the teams enjoy a regular
3
ASNE 2005 - Overcoming Interoperability Challenges for Joint
and Coalition Heavy Lift Transportation Operations - Capt. J
Wilkins USN, S Quinn, N Hills, Cdr. P Hudson USNR, R
Wasalaski, E Ofosu-Apeasah
The Damaged Ship, London, UK
liaison, participate in joint exercises, and have supported
each other on operations, either by making available
facilities or advice, and in the case of the repatriation of
HMS ENDURANCE in 2009, enhancing the MoD heavy
lift recovery team with US personnel.
The team have made strides in developing an accepted
commonality of approach to training and qualification,
such that the US Training Instruction for Heavy Lift /
Docking Officer is coherent with the equivalent UK
Functional Competency, thus ensuring that within
defined areas, the teams can act as force multipliers for
one another.
4. SPECIALIST TOOLS AND TRAINING
There are a plethora of tools and training available to
support a vessel operating in a marine environment. They
range from design and build tools, analysis and operating
information type tools, some of which are complimentary
to each other, others of which are completely stand alone.
These tools may well be suited to assessing the ship in
both an intact and also static damaged condition.
However, during the case of a damaged ship, analyses
may need to be carried out quickly, based upon rapidly
developing information. The ability for the responding
authority to perform rapid calculations, not only to
establish the current condition of the vessel, but also to
produce a variety of recovery options, is crucial. Some
existing software packages are more amenable to this
than others, but there is a difference to analysing the
vessel for its stability or structural performance during
design, build and certification, and to that required to
make the right decisions as to what remedial action
should be taken. This may include lightering, ballasting
to stabilise the grounding, shifting fluids to change trim
or even deliberately allowing oil to escape from a
ruptured tank, a highly emotive issue. It should also
consider the impact of tidal cycles and whether the ship
is safe to be moved from one location to another in
possibly hostile sea states.
In order to maintain availability4 of capability, S&MO
maintains a small cadre of officers, with whom it is
investing training effort in the use of salvage support
software. Some of these officers have a strong
background in naval architecture, but others have
alternative primary skill sets, some of which were
described earlier. Notwithstanding, the intention is that
S&MO PT wishes to guarantee a minimum capability in
this aspect, not only for analyses to be completed in UK
based operations rooms, but also as part of the deployed
team. Consequently, the preference is to train and
exercise the officers in one software package. We have
recognised that whilst it is possible for a very limited
4
Past experience has suggested that a minimum of 6 qualified
personnel are required at any one time to cope with post churn,
leave and absence constraints, and to provide an enduring
response to a long duration incident.
The Damaged Ship, London, UK
number of personnel to become familiar and (proficient
to a certain degree) in a variety of systems, the training
burden becomes untenable if we attempt to expand this to
larger numbers of staff who would not necessarily use
these tools as part of their normal daily duties. It
becomes a question of balance of resources, and in
respect to this, the S&MO team has deliberately elected
to primarily engage with one package.
A secondary consideration in deciding to elect upon one
package was that transfer of data from one software
package to another was a challenge in itself unless
dedicated translation facilities existed. It also made the
task of system validation more complex. The usage of
one software suite on a cradle to grave basis clearly
offered one of the potentially easiest solutions for the
user in terms of training burden and independent
validation.
Having taken the decision to concentrate efforts on one
software package it became essential that the breadth and
capability of that package was maximised. Nowhere is
this illustrated more clearly than in the HMS
NOTTINGHAM salvage and recovery option. This task
involved an initial salvage assessment, stabilisation of
the structural condition of the vessel, modification to
enable blue water towage of the casualty, subsequent
heavy lift transportation to UK and final dry docking in
Portsmouth.
The number of tools and analytical techniques available
to the salvor is also increasing, but they currently tend to
be in a mixture of packages, or have limitations as to
what can safely be analysed with confidence by users of
varying background experience and knowledge.
Historically, S&MO have used HECSALV for many
years having specialised on that package since the early
90s. HECSALV has a world wide customer base, and
was also used by Lloyds Ships Emergency Response
Service prior to them changing to NAPA and Seasafe.
Our engagement with the US Navy through the MOU
described above presented the opportunity for us to
utilise the US Navy version of HECSALV which is the
Programme of Ship Salvage Engineering (POSSE). This
then presented the further opportunity to share
development costs jointly and enter a stage of enhanced
development due to the greatly increased availability of
funding.
For the reasons described above POSSE is the system of
choice for S&MO nonetheless the core specialist users of
POSSE are also charged with maintaining familiarity
with other systems to ensure that a catholic knowledge is
maintained of improvements within the industry. One
capability that is maintained in particular is
PARAMARINE due to its widespread usage across UK
naval platforms.
Over recent years, S&MO, in partnership with US
SUPSALV have engaged with Herbert Software Systems
Inc to develop additional modules for inclusion in the
POSSE suite. These modules are specifically aimed at
providing aids for the analysis for dry docking and heavy
lift operations. Both the UK and US have tried and tested
procedures and information for the docking of vessels
including bespoke information for classes of ships or
individual ships. However, this information tends to be
for an intact vessel for a planned docking. Experience
from operations suggests that the ability to rapidly model
and analyse the docking capability of a damaged vessel,
where the standard arrangements are not suitable, is
required. This has been the focus of the UK and US
funded development work, but not merely to produce a
separate tool, but to make it an integrated part of the
system that the operator is well versed in, and also to
import the results of the damaged analysis that they may
have already completed in support of the casualty.
Speed of response is important, but accuracy is essential
in the provision of advice to either the casualty or the
embarked salvage team. Consequently, the UK is
investing considerable time and energy in ensuring that it
has a library of ship models for all Royal Navy and
Royal Fleet Auxiliary vessels, both in service and those
recently decommissioned5, which accurately represent
the configuration of the vessel. S&MO is working with
the various Platform Design Authorities to try and ensure
that the models are regularly updated, with a proactive
engagement of the duty holders to confirm or otherwise
any significantly changes in the ship arrangements and
configuration.
When operating salvage orientated software, it is not a
question of having operators who can produce the results
of analysis and then sit back or become engaged on other
business. Incidents tend to be of long duration and of an
evolving nature depending upon the success or otherwise
of remedial actions. In developing its cadre of POSSE
trained officers, S&MO is also establishing the process
by which POSSE analyses should be conducted. In
extended incidents, there is a possibility that an
alternative picture of events could be handed over to
the relieving naval architect, potentially leading to an
analysis being completed on inaccurate information, and
incorrect advice being provided to the Salvage Master.
The process is important because it enables an analysis to
be maintained throughout the duration of an incident,
with each step being carefully documented so that
relieving operators can either refresh on steps already
taken, or interrogate previous elements to establish
alternative courses of action as required. Consequently,
we are taking steps to try and ensure our analyses are
5
Models of decommissioned units are maintained, as the MoD
cannot fully relinquish its responsibility to the vessel until it is
formally disposed of. Additionally, it is not unusual for ex-RN and
RFA vessels to be operated by other navies. Whilst less effort is
expended in maintaining the currency of the model, a historical
copy is retained electronically.

auditable, with an expectation to formally record it as
part of our ISO accreditation.
Our access to POSSE is laptop based, and at initiation of
an incident, the responding officer shall select the
appropriate model from a secure drive and run a rapid
validation check for any errors or warnings (Figure 1).
Figure 1 POSSE Model Validation
Then, upon receipt of loading information from the
casualty, the most recent pre-damage and post damage
conditions are established. The process is very much akin
to that which would be carried out by a commercial
emergency response service, with efforts being made to
verify the analysis results with those being reported by
the casualty. As would be expected there may be
discrepancies with draft marks, heel angles, trim, etc,
possibly originating from an overly estimated pre-
incident loading condition, or undetected damage, sealed
compartments not being completely flooded, etc. It is
incumbent on the professionalism, expertise and
experience of the responding operators to achieve a
modelled solution that is within acceptable tolerances,
prior to any remedial actions being calculated and advice
passed back to the casualty.
During this time it is probable that the ships company
shall have effected an immediate first aid plan,
containing flooding, reinforcing boundaries and possible
recovering compartments. Hence, continual updates will
be needed to the modelling of the situation (Figure 2).
Figure 2 POSSE Damage Entry, Time Domain
Stranding Analysis
The Damaged Ship, London, UK
With the vessel hopefully established in a stable situation
(whether it be free floating, stranded, unable to propel,
etc), consideration can be given to identifying
recuperative actions for restoring various capabilities in
the ship. With a warship, the military imperative shall
always be there, and may in extremis, be contrary to
action that might be considered logical for recovering the
situation. For example, a refloating plan for a stranded
vessel may include the deliberate counterflooding of
other compartments. Those compartments, whilst being
attractive to the salvage team as candidates for extra
tankage, may have high military value, eg: magazines,
computer spaces, and the command aim could be focused
on retaining a defensive or war fighting capability of the
unit, over and above the basic float and move functions.
S&MO can conduct many what if analyses of the
stricken vessel, identifying the outcomes of different
causes of action. The father/son architecture of POSSE
enables the operator to look at the different sequencing of
pumping and flooding operations, but retain them and
then run alternative solutions from the same starting
point. The time domain function enables us to plan
recovery actions for stranded vessels, with an ability to
insert tide and wave cycles. All of this can be done
within the same software suite. Because the S&MO PT
maintain this capability on laptops, it is possible that
subsequent calculations can take place in theatre with the
deployed team, as well as in the supporting UK
headquarters. There are other software packages which
can do similar functions with different graphical
representations, but as stated before, POSSE is our
current software of choice.
However, where we think we have seen significant
developments is the use of the package is in docking and
heavy lift analyses. As said before, the operator will have
used POSSE to bring the casualty vessel to an acceptable
level of stability and attitude, this may have restored its
full or partial capability, or, as in the case of
NOTTINGHAM or ENDURANCE, further remedial
action is required to repatriate the vessel to a different
location. Should the vessel be considered robust enough
to be towed, it is a simple proposition for the operator to
import the same POSSE model definition into the TOW
element of the package, input the as calculated floating
condition as already analysed, and readily establish the
required bollard pull and type of towing jewellery
required to move the ship at a variety of speed/seastate
combinations. This in turn can be used to establish the
required strength of available towing fittings on the
casualty vessel, or confirm the requirement to
manufacture new ones locally (as we did with HMS
NOTTINGHAM in 2002), and also to generate a tow
plan which can be supplied as guidance to the tug captain
at the start of any towing evolution.
It is possible that the casualty is not suitable for a towing
operation, and drydocking and heavy lift are the
preferred methods of either restoring capability to the
ship, or transporting her as soon as possible. The
The Damaged Ship, London, UK
UK/USN version of POSSE now has the capability for
the operator to seamlessly continue the analysis of the
casualty vessel to enable a damaged docking condition to
be established. The operator can input data on the
available dry docks (eg: dimensions, water depths,
declivity, etc), then set up a docking block layout (Figure
3) which will suit the damaged vessel, missing out blocks
in way of damaged structure and increasing blocks in
way of areas of likely high loads.
Figure 3 POSSE Dock Block Tool
The operator can set up a pumping condition for the
docking sequence (based upon the tide routine already in
the software), and then enable the analysis to interrogate
the loadings on each individual block, identifying areas
of high loads (Figure 4), and where an alternative dock
block arrangement or material may be required.
Figure 4 POSSE Docking Analysis
Similarly, POSSE has been further developed to include
a capability to analyse the Float On Float Off (FO/FO)
heavy lift of the casualty vessel. This is more involved
that the dry docking module, because it involves the
entry into the problem of another ship definition in the
shape of the heavy lift carrier. S&MO PT and USN
SUPSALV have created, and are in the process of
populating, a database of UK and US warships and
available worldwide heavy lift assets, be they barges or
self propelled vessels. The initial capability of database is
to enable the salvor to complete a rapid assessment of the
available world market, with the software matching the
characteristics of the vessel to be lifted with the key
parameters of the lifting vessel, eg: available cargo deck
space, maximum water over deck, deck strength,
availability of services such as fire mains, electrical
power, etc. The natural extension to this is to liaise with
the operators of the heavy lift ship operators and obtain
as much technical detail on the classes of ships as
possible to enable POSSE ship definitions to be
produced. It would then be possible to link this definition
to the database, such that if they become the potential
carrier for the stricken ship, the definition was
immediately available and can be imported into the main
POSSE analysis.
The POSSE software enables the user to establish a
bespoke cribbing arrangement for the casualty vessel,
and then simulate the floating on of the vessel, and the
deballasting operation can be modelled, with stability of
both vessels being monitored. The deck loadings can be
monitored and different cribbing arrangements used
should there be any overloading. The POSSE analysis is
not designed to take the place of the full calculation that
would be completed by the transport contractor and
presented in their transport manual, but is intended to
compliment that, and ensure that we are a properly
informed customer.
The importance to S&MO PT is that all of this can be
done in the same software system, using the same ship
definition throughout, with a user interface that the
operator is familiar with.
As stated before, software tools are becoming
increasingly capable and there are significant underway
developments in providing mechanisms by which the
dynamic performance of a damaged vessel might be
assessed. This is of particular relevance to the recovery
action, this team has been engaged in at least two
operation in relatively recent years where a damaged and
flooded vessel has had to be moved from the area of the
incident to a safe sheltered haven. For example, HMS
NOTTINGHAM in 2002 had to be moved from an
intermediate location off Lord Howe Island, and moved
350 miles to the West under tow with significant
structural damage and flooded compartments.

Figure 5 HMS NOTINGHAM Flooded Forward
Engine Room
Figure 5 shows approximately 800 tonnes of flood water
in the Forward Engine Room (fortunately this space was
recovered by use of an external seal and pumping), but
supposedly the ship had had to be moved over an
extended distance with this flood water still entrained in
the ship? What is the impact of the flood water on the
overall dynamics, what further damage might be
incurred? HMS ENDURANCE was moved several
hundred miles under tow with a flooded engine room,
fortunately in moderating conditions, had conditions
been worse, a comprehensive analysis of the ship
behaviour under tow may have resulted in different
advice and decisions.
However, developing tools, whilst extremely capable
should also consider the requirements of the end user, is
it for use at supporting head quarters, or also in the field?
S&MO PT is very keen for new techniques to be
established and embraced. The likely level of accuracy of
ship description following damage must be considered,
along with damage inputs. The analyses must be prompt,
but also convey a level of accuracy and confidence that
will not lead to prohibitive conservatisms in any recovery
action plans.
5. ROBUST PLANNING AND EXERCISING
The MoD has recognised that a fundamental part of
emergency preparedness is the ability to train and
practice. With respect to physical damage control, the
MoD ensures that ships companies are very well
instructed at shore establishments via dedicated trainers
(eg: the Damage Repair and Instruction Unit), and very
rigorously tested by the staff of Flag Officer Sea
Training, an organisation with a well deserved
worldwide reputation of excellence. Whilst focus is
maintained on exercising the response to a marine
incident, an increasing emphasis is also being placed on
educating operators on the impact of an event, and how
the response will be shaped. S&MO PT invest in a series
of lectures, briefings and presentations to new
Commanding Officers and Marine Engineer Officers and
ships Logistic and Supply Officers, which not only
provide information on the team capability, but also
present case studies of previous incidents and explore in
detail how the assistance will be provided to the stricken
vessel. It is emphasised that initial first aid and
stabilisation is solely in the hands of the crew (Figure 6),
but that all efforts will be made to provide external
advice and guidance, and material help in men and
material within the next few hours (for example, a
response team was with HMS NOTTINGHAM within 48
hours of the initial incident, and detailed discussions held
with MEO on recovering the engine room within 12
hours, despite the casualty being on the other side of the
world in a relatively remote location). It is incumbent on
The Damaged Ship, London, UK
the supporting shore authorities to ensure that the ship is
best equipped with the necessary information and
training to enable the crew to conduct the initial
stabilisation actions, noting that in the case of the
warship, there will possible be competing command
aims. It may not be possible for the damage control
teams to be reacting to instructions that may be more
amenable to the war fighting capability of the vessel,
rather than simply making sure that the ship is floating
and able to move. As already mentioned, the natural
inclination to counterflood a compartment to take off list
or reduce trim might not be possible if the obvious
candidate compartment is a high value space. Thus
information needs to be robust, but also flexible.
Figure 6 The First Line of Defence
With respect to maritime operations, Navy Command
exercises the ACRO described earlier at least once a year
with a full scenario based activation of all of the relevant
departments. The culture of this exercise is to train rather
than test, and ensure that lessons identified are captured
and implemented. Not only is the MoD proving to itself
that the right measures are in place to provide an
effective response, it is also demonstrating that it has a
capability that is at least as good as to what is demanded
of a commercial ship operator.
Lessons Identified from various incidents
(ENDURANCE Engine Room Flood 2008, ASTUTE
grounding in 2010) have highlighted the need for
accurate information flow and management, particularly
technical data from the casualty. During a significant
incident, of the type that is likely to require technical
analysis (eg: grounding, collision, fire, explosion, etc),
ships staff are most likely to be engaged in immediate
first aid and stabilisation, preserving the safety of the
personnel, the vessel and the environment. There are
incredible pressures on their time and activities, with
many competing priorities, which coupled with natural
anxieties, may mean that demands for information from
shore may be afforded a lower priority. Consequently the
supporting authorities have developed the Naval
Architecture Incident Information Form (NAIIF), which
is designed to collate information on the ship condition
The Damaged Ship, London, UK
and enable it to be transmitted ashore as fast as possible.
The form consists of a common format frontispiece, but
with individual annexes for classes of ship, each annex
representing the loading condition as would be found on
the onboard stability computer and/or stateboard in HQ1.
S&MO PT and the platform Duty Holders are engaged in
exercising the use of this form during FOST serials, in
concert with NAVCOM.
Such techniques are already established within a
commercial emergency response service, and we see this
as a reflection of bringing in best practice into the
military response. Indeed, there are many examples of
how emergency response services are organised which
either mirror what we can achieve within the MoD, or
where we feel there are valuable lessons to be learned.
During the response to an incident, lessons identified
have also highlighted the need for accurate and concise
information management between the various technical
cells, and also between the changes of watch. A
developing naval architecture picture must be contained
in one place, with a clear unambiguous expression of
command aims and strategy provided to the incoming
personnel. This is especially prevalent, if the personnel
are new to the incident and are less familiar with the
class of vessel in question. To assist in preparedness for
this, S&MO PT have started to run a series of emergency
response seminars, which are intended to gather the
DE&S (and other) community naval architects, at various
seniorities, to establish the methods of best practice that
we may employ as general emergency support to a
maritime incident.
6. FUTURE
As has hopefully been expressed in this paper, this is an
area that can only improve by continuous development,
and it is incumbent on those who participate in the
support to damaged vessels, and vessels in distress, that
they are as well prepared as possible to respond to such
an incident.
Such preparation should include investment in the
training of personnel, allowing time for personnel to
rehearse and practice their skills in a dedicated fashion in
a proper Continued Professional Development manner. It
should also include the collaboration of different
agencies and organisations, all of whom have something
to contribute to a successful resolution of an incident.
New analytical tools are under development, and such
tools should be robust, intuitive to use, and afford the
operator complete confidence in the results of their
analysis, leading to accurate and effective formulation of
recovery plans.
7. AUTHORS BIOGRAPHIES
Mr Nigel Hills RCNC is currently the lead Naval
Architect on the staff of S&MO PT. Previous to this he
was the Deputy Team Leader and Head of Naval
Architecture for RN Major Warships where he was
responsible for all Naval Architecture and Safety
Management Issues for CVS, Type 42 (including the
project lead for the recovery of HMS NOTTINGHAM
from Australia in 2002) and LPH classes. He has served
in the project teams responsible for SSN Refitting Docks,
the S & T Class Update, Royal Marine Special Forces
and the former Chief Naval Architects department.
During his time with S&MO PT he has been engaged in
the recovery of HMS ENDURANCE, HMS GRIMSBY
and HMS SUPERB to the UK following various
incidents. He was the project manager for the onsite
survey of a sunken Russian Nuclear Submarine and has
acted as Subject Matter Expert to Canada and Norway
for the heavy lift transport of Russian Nuclear
Submarines. He is the project manager for supplying
Marine Warranty Surveyor support to the transport of
hull modules for the QUEEN ELIZABETH Class aircraft
carriers.
Captain Stephen Quinn OBE MIIMS MNI is a
Salvage Master and Deputy Team Leader Marine
Operations in S&MO PT. A Master Mariner, the first 13
years of his career was spent in the Merchant Navy.
During his MOD career he has been regularly involved in
salvage operations. Major tasks have included the
salvage of HMS BRAZEN following her grounding in
Chile, MV HERMES, MV DANIKA GREEN and MT
SEA EMPRESS. He has conducted 17 aircraft recoveries
including the recovery of a virtually intact Nimrod
aircraft. Stephen was the first S&MO Salvage Officer to
gain a HSE diving qualification and was responsible for
the initial establishment of the Plymouth based Marine
Salvage unit. He was the Senior MOD official and
Salvage Master on Lord Howe Island for the recovery of
HMS NOTTINGHAM. He was also the UK lead for the
recovery of HMS SUPERB and HMS ENDURANCE,
and most recently the salvage lead for the recovery of
HMS ASTUTE from grounding off the west coast of
Scotland.
 The Damaged Ship, London, UK

First Aid repairs; to a standard necessary to reach a repair port; re-floating or recovery of disabled or stranded
MoD owned vessels or Commercial Chartered Shipping (CCS) following collision or grounding.

Recovery of ditched MOD aircraft from the sea or inland waters.

Recovery of MOD-owned vessels sunk in harbour or sheltered waters.

Assistance to a bottomed disabled submarine, conventional or nuclear powered.

Object recovery, including nuclear weapons from the seabed.

Recovery of Commercially Operated Military Assets (COMA), Commercially Owned Military Registered
(COMR) and Commercially Owned Military Operated (COMO) aircraft from the Sea or inland waters where
agreements are in place.

Clearance of MOD ports, facilities and approaches, including Marchwood Sea Mounting Centre, DM Crombie
and UK foreign bases.

The management of all MOD ocean and coastal (blue-water) towing and ship transportation by uplift or barge.

Provision of expertise and advice to the Department of Transport and other government departments (on
repayment terms).

Provide expert advice on wreck management and emerging marine salvage issues.

Table 1 MoD Endorsed Salvage Policy

The Damaged Ship, London, UK

Table 2 Fleet Accident and Crisis Response Organisation

The Damaged Ship, London, UK

CASE STUDY OF DAMAGE STABILITY CRITERIA OF MERCHANT VESSELS AND

WARSHIPS
R Perez and J M Riola, Escuela Tcnica Superior de Ingenieros Navales (Universidad Politcnica de Madrid), Spain

SUMMARY

While there is plenty of data to perform a complete assessment of stability after damage in merchant vessels,
unfortunately that is not the case for warships. There are different regulations of damage stability that could be used for
such exercise, for example the Design Data Sheet, the Naval Ship Code, and within the SOLAS, the Stockholm
Agreement, where water on deck is included.

It is well known that regulations ruling stability after damage in warships are more demanding (in general) than
merchant rules in terms of water on deck, however the Stockholm Agreement provides greater reliability for calculated
results in civilian vessels, compared to those used by warships.

Warships are different and more interesting from the stability point of view, as they have a more detailed study than their
civilian counterparts. The calculations are also more exhaustive and additionally they study large groups of loading
conditions. To corroborate this point a comparative study has been made of results between different Navies.

This paper provides a study of the stability criterion for the US and British Navies, a study of the new NSC criterion
(Classification Society-driven approach) and an analysis of SOLAS from its origins until the new probabilistic approach
going through all the amendments that relate to stability.

The paper concludes with the introduction of a series of comparisons between criteria used by the Navies to maintain the
integrity of its basic approach, and increase the similarities with the criterion of IMO, such as the calculation of water on
deck from the Stockholm Agreement. These comparisons lead to interesting conclusions regarding how current criteria
used by the Navies could be enormously improved just with a few minor changes.

NOMENCLATURE Lpp Length between perpendiculars (m).

The nomenclature, in this technical paper, uses SI units.
It is ordered alphabetically.
Displacement (Tons).
Angle of heel ().
1 Limit angle of the damage righting arm
curve ().
c Angle of steady heel ().
r Rolling angle ().
A Attained subdivision index
(dimensionless).
A1 Area under righting arm curve between
the roll back angle and the equilibrium
heel angle. Reserve of dynamic
stability (m2).
A2 Area under righting arm curve between
the angle of equilibrium and the
extreme intersection between righting
arm (m2).
B Beam at the widest point (m).
CB Block coefficient (dimensionless).
CM Midship coefficient (dimensionless).
D Depth (m).
DN Dynamic stability (mmrd).
GM Permissible metacentric height (m).
GZ Righting arm (m).
KG Permissible height of the centre of
gravity (m).
R Required subdivision index
(dimensionless).
RSK Is the sure kill radius (m).
RSS Is the sure save radius (m).
T Middle draft or middle draught (m).
Trim Difference between the forward and
after drafts (m).
1. INTRODUCTION
To avoid duplication, gaps and shortcomings in safety, it
is important for the Navies to work together with the
Classification Societies in the development of effective
and sustainable arrangements. Thus, development of
rules for warships Naval Ships Rules by various
Classification Societies is the most important
contribution to work in this area. The idea of cooperation
to make an International Convention for the Safety Of
Life At Sea (SOLAS) goes back to the nineties. In
September 1998, Classification Societies of the Member
States of the North Atlantic Treaty Organization (NATO)
met to establish links within their own organization. This
meeting established the Naval Ship Classification
Association (NSCA), in May 2002, and the cooperation
was defined according to the following terms of
reference: promote safety standards at sea, promote
measures to protect the marine environment, promote
and develop common operating standards, undertake

R&D to support the above and communicate the views of
the partnership agreements and the NSCA. The
philosophy of the SOLAS is applicable to merchant ships,
and is not fully transferable to a warship. A warship has
requirements for acoustic signature, electromagnetic
signature, signature radar, electronic warfare,
antisubmarine warfare and it demonstrates that a ship is
not civil.
2. DETERMINISTIC VS PROBABILISTIC
In this moment, the community of Naval Architects is
debating between the probabilistic and the deterministic
methods. It is then necessary to define in this technical
paper what a deterministic or a probabilistic method is.
To verify the validity of a model is necessary to deduct
from it a certain number of hypothesis and then to
corroborate it with observations of predicted results.
Deterministic models correspond to mathematical models
designed on the assumption that the result of an
experiment is determined by the conditions under which
it is performed; stochastic models (probabilistic) are
those data for a model is obtained through a sampling of
probability distributions. This sample allows that
uncertainty (which can be reduced if more data is
collected) and variability are propagated from the model
and demonstrated in the results of the model.
In summary, it could be said that a deterministic model
assumes that the actual result is determined by the
conditions under which the experiment, however when a
stochastic model is used, the experimental conditions
determine only probabilistic behaviour (the probability
distribution) of the observable results.
The US Navy and British Royal Navy criteria are based in
deterministic models, so it is necessary not to forget the
importance of these models.
Given the relevancy of probabilistic methods, they have
been evaluated; in particular we have studied the
philosophy and new tools for its calculation, as the
FSUBD module, part of the system FORAN
CAD/CAE/CAM. We must not forget that this paper has
made a detailed study of the FORAN system modules
used for the study of the intact and after damage stability.
It is equally important to point out the pros and cons of
using probabilistic criteria. These characterizations of
criteria risk are aiming to provide estimates of
uncertainty and variability associated with each of the
predicted levels of risk. This is one of the positive
aspects (stochastic) of such configuration models, but
also leads to confusion in the interpretation of the data.
Furthermore, those same estimations are uncertain and
depend on the methods and assumptions used to make
these calculations. This fact is exacerbated frequently
The Damaged Ship, London, UK
given the limited data set available and perhaps
overestimating risk associated with a particular fault.
It is well known that a more effective method for
implementing the proposal of the survival of an
optimized schema of design of vessels, assessment
procedure is an approach that considers the probability of
survival on the basis of survival in a quasi static criteria
such as the of the US Navy and British Royal Navy. The
navies criteria take into account data from actual damage
to ships and they have been accredited to be reliable until
today that seem satisfactory believed they won't switch
from deterministic method to probabilistic method.
Philosophy for the transformation of these deterministic
in a set of rational criteria with a stochastic approach or
probabilistic criteria is based on the Resolution A.265
(VIII) of the Design Data Sheet for passenger ships. For
passenger ships, not military ships, longer than the latter,
as already mentioned, still relying on deterministic
criteria.
From the first of February of 1992 the probabilistic
method was inserted into SOLAS as Part B-1 of the
Chapter II-1, annex Regulation for subdivision and
damage stability of cargo ship over one hundred meters
in length thats applies to dry cargo ships constructed on
or after the first of February of 1992. Later on, ships with
length between eighty and one hundred meters were also
included.
The 8th Assembly of International Maritime Organization
(IMO), by Resolution A265 (VIII) adopted a set of
probabilistic Regulations of subdivisions and damage
stability passenger ships as a equivalent to and as total
alternative to the requirements of Part B of Chapter II of
the SOLAS, 1960 for passengers.
Finally in the 80th session of IMO, Maritime Safety
Committee (MSC) the working group has finalized a
substantial revision of SOLAS CHII pt 1 A, B and B1
aiming at harmonized damage stability requirements for
all ship types except for tankers, performed by means of
a common probabilistic methods. The draft has been
adopted at the MSC 80 without further modification. The
revised Ch II-1 will apply to all new passengers vessel,
Roll on-Roll off and cargo ships built on or after the first
of January of 2009.
Probabilistic concepts address the probability of damage
occurring at any location throughout a ship and adopt a
more rational criterion of subdivision by considering the
likelihood of damage resulting in the flooding of only
one compartment, or any number of adjacent
compartments, either longitudinally, transversely or
vertically. The residual buoyancy and stability of a ship
is calculated for each of such damage cases, and either a
positive or a zero contribution is associated to each case,
depending on, whether or not, the residual buoyancy and
stability are considered sufficient.
The Damaged Ship, London, UK
In probabilistic terms, a ship need not to survive every
possible case of assumed damage, provided that there are
a number of survival cases which allow obtaining, by
addition of their corresponding positive corresponding
positive contributions, a total value A (Attained
Subdivision Index) equal or greater than a reference R
(Required Subdivision Index). In the new SOLAS
revision, in addition, the partial index As, Ap and Al are
not less than 0.9R for passenger ships and 0.5R for
cargo ships.
The A index attained by a ship considered to be measure
its level of safety against both sinking and capsizing. In
this way, two ships that have different main dimensions
but whose A indexes are equal, may be considered as
having the same level of safety.
It is worth mentioning the debate between deterministic
or probabilistic methods continues. As an example,
Spanish Congress 2009 of Naval Engineering held in
Vigo, it was reached as the conclusion that deterministic
methods should still be considered. In addition we must
emphasize that no Navy used stochastic methods or
probabilistic as exhibited during the development of
Naval Ship Code when the President of the International
Association of Classification Societies that it would thus
remain. However Bureau Veritas is already starting to
dispose of that method for other types of ships.
Therefore, since the focus of this paper was of warships,
mainly due to the appearance on the scene of the new
Naval Ship Code (ANEP 77) which will be mandatory
for all military naval constructions this year and
leveraging the use of the FORAN system, the objective
was the comparison of the various deterministic criteria
and whether the new Naval Ship Code code was really
more or less restrictive than the previous version and in
which cases.
3. DESIGN DATA SHEET
The criteria to evaluate adequate damage stability
performance according the Design Data Sheet (DDS) are
based on the Figure 1. A reduction of the righting arm
equal to 0.05cos is included in the righting arm curve
to account for unknown unsymmetrical flooding or
transverse shift of loose material. Beam wind heeling
arm curve is calculated with the same method as used for
intact stability calculations, but considering a beam wind
velocity of around 32-33 (knots) as defined in DDS 079-
1 (Naval Ship Engineering Centre, 1975). The damage
stability is considered satisfactory if the static
equilibrium angle of heel c, point C without wind
rolling effects does not exceed 15. The limit angle 1 of
the damage righting arm curve is 45 or the angle at
which unrestricted flooding into the ship would occur,
whichever is less.
Figure 1. Damage stability criteria
The criterion is considered fulfilled if the reserve of
dynamic stability A1 is not less than 1.4A2, where A2
extends r to windward as shown in the Figure 1. The
tendency during recent decades in surface naval ship
design was to assess and minimize susceptibility through
detailed signature management. For the naval architect it
is usually enough to assess the adequacy of its design
with respect to vulnerability through the use of the
damaged stability requirements introduced by the various
Navies, such as those used by the US Navy and the UK
Ministry of Defence (MoD).
Based on the concept of the damage function used in the
theory of defence analysis, the fraction of the target
assumed to be damaged within a radius r from the impact
point is assumed to follow the well-known log-normal
distribution given by the Equation 1 (Przemieniecki,
1994):
2 r
1 ln (1)
d (r ) = 1
r
exp dr
0 2 r 2
2
Where RSK is the sure kill radius which means that
[d(RSK) = 0.98], RSS is the sure save radius which means
[d(RSS) = 0.02] and zSS constant equal to 1.45222.
= RSS RSK (2)
1 R
= ln SS (3)
2 2 z SS RSK
The damage extent ranges of naval ships may result from
test analysis, analysis of data from actual engagements,
empirical formulas linking the damage range with the
type and the weight of the warhead or from the use of
damage lengths defined in current deterministic damage
stability regulations for naval ships.
Figure 2. Damage extent on aircraft carrier profile

In the later case a first approximation of the RSS can be
taken according to NES-109 and DDS-079 and it would
be 0.15L, see Figure 2 (Boulougouris and Papanikolaou,
2004). The RSK has been assumed equal to 0.02L.
4. NAVAL SHIP CODE
In addition to Navies, Classification Societies through the
NSCA have a standing invitation to attend the meetings
of the specialist team as active participants. The
specialist team is tasked with the development of a Naval
Ship Code (NSC) that will provide a cost-effective
framework for a naval surface ship safety management
system based on and benchmarked against IMO
conventions and resolutions. The Specialist Team has
established a Goal Based Approach to the development
of the NSC and is now developing each chapter in turn.
This folder in the NAS library contains the latest
documents including NSC chapters, related guidance and
records of meetings. The NSC adopts a goal based
approach. The basic principle of a goal based approach is
that the goals should represent the top tiers of the
framework, against which ship is verified both at design
and construction stages, and during ship operation. This
enables the NSC to become prescriptive if appropriate for
the subject, or remain at a high level with reference to
other standards and their assurance processes. The goal
based approach also permits innovation by allowing
alternative arrangements to be justified as complying
with the higher level requirements. The increasing width
of the triangle as the NSC descends through the tiers
implies an increasing level of detail.
Figure 3. Damage Shapes
5. THE STOCKHOLM AGREEMENT
The Stockholm Agreement was established in the context
of resolution of the fourteen SOLAS of the IMO in 1995,
and authorized government contractors to enter into such
commitments if they believe that the predominant sea
conditions and other conditions require specific local
stability in a certain sea area. In short, these rules are
complementary to the rules SOLAS-90, with the addition
of technical specifications to explicitly take into account
the risk of accumulation of water on the car deck. The
introduction of the Stockholm Agreement is closely
The Damaged Ship, London, UK
associated with three unprecedented stops in the history
of damage stability/survivability assessment:
Water on deck was explicitly taken into account for
the first time. This is remarkable in view of the
knowledge that 85% of all deaths with ferry
accidents relate to car deck flooding.
The effect of waves, and this is even more
remarkable, was explicitly taken into account also
for the first time.
It paved the way to the introduction of performance-
based standards for assessing the damage
survivability of ships.
Figure 4. Damage example with two compartments plus
water on deck
All three steps represent gigantic improvements in the
approach to addressing ferry safety but any potential
benefits will have to be balanced against any likely costs
that might be incurred through the introduction of
inappropriate standards.
6. APLICATION IN MERCHANT SHIPS AND
WARSHIPS
Nowadays, in both practical navigation and shipyard
technical offices, stability tests in load and sea
conditions, as in working or damaged conditions, are
performed with software packages that starting from the
ship design are able to quickly computed the required
data. This research focuses on evaluating the
configuration of warships, with empty deck that could be
a warship, using the working and faulty stability FORAN
modules, in particular Architecture-Project Subsystem
property of SENER Ingeniera y Sistemas. The approach
taken to perform the analysis has been the following: In
the FSURF module, shapes, decks and walls are defined.
Then, VOLUME module defines ship volumes and
computes their volumetric capacity. The LOAD and
FLOOD allow visualization of the detailed requirements
generated from the stability requirement chosen, and also
enables data entry to compute minimum GMs. Inside
The Damaged Ship, London, UK
last the modules, it is possible to check the most common
standard stability criteria and a user define criteria
obtaining if necessary the limiting KG values.
6.1 APLICATION IN MERCHANT SHIPS
The chosen merchant ship, for this part of the analysis, is
a Roll on-Roll off with the dimensions shown in Table 1,
with two propeller shafts; which has a double bottom
with a height upper to a tenth of the beam, B/10. To find
these dimensions, a database with other Roll on-Roll off
ships of similar characteristics, has been used.
L 200.00
B 25.00
T 5.50
D 15.00
CB 0.61
CM 0.96
Table 1. Main dimensions of the MSHIP project
The ship, that we have selected, has called MSHIP. It has
a minimum draft of 4.51 (m) and a maximum draft of
7.10 (m). We need to define the increments between both
of the drafts. In this case, the number of the increments
will be two. We have defined trim equal to zero. Both,
ship and configuration have been evaluated with high
degree of detail in order to achieve equivalent
comparisons. The fact of placing a longitudinal bulkhead
below the deck number three has not been random. By
designing the compartment layout in this way, a bigger
number of faulty conditions and higher number of
possible combinations are achieved.
Figure 5. Visualization of the merchant ship studied with
the FDESIGN module and diagram of one of the damage
6.2 APLICATION IN WARSHIPS
The chosen ship for this analysis is a warship with the
dimensions shown in Table 1, with one propeller shaft;
which has also a double bottom with a height upper to a
tenth of the beam. To find these dimensions, a database
with other warship of similar characteristics has been
used.
L 150.00
B 15.00
T 5.00
D 10.00
CB 0.49
CM 0.78
Table 2. Main dimensions of the WSHIP project
Figure 6. Diagram of one of the damage in the warship
studied, WSHIP, and the 3D visualization
The ship, that we have selected and we have called
WSHIP, has a minimum draft of 3.65 (m) and a
maximum draft of 6.21 (m).
7. RESULTS IN MERCHANT SHIPS AND
WARSHIPS
The establishment of an international maritime law,
especially regarding safety, is a long process that is not
without difficulties, it requires a lengthy period of
research and analysis, consensus and ratification by a
sufficient number of countries. Its implementation is not
always possible in older ships. The first result is that
ships can coexist for years, with two standards widely
depending on their seniority or banner, as happens with
the well-known case of oil tankers without double hull as
the Prestige. However, despite the remarkable technical
and legislative effort that are carried out by IMO or the
major advances in the safety convention SOLAS.
Warships are exempt from these rules and do not exist. In
 The Damaged Ship, London, UK

the naval field, there are not organization equivalent to
IMO to understand the international level about the
safety of such vessels. Traditionally, the warships are
taking the rules of IMO exists that do not interfere with
naval objectives and adapting them to the extent as far as
possible.
7.1 RESULTS IN MERCHANT SHIPS
The calculations are made on intact to see if the merchant
ship, MSHIP, complies with the intact stability.
Criteria
Description
number
1 GZ of 0.2 between 30 and 90
2 DN of 55.0 (mm.rd) between 0 and 30
3 DN of 90.0 (mm.rd) between 0 and 40
4 DN of 30.0 (mm.rd) between 30 and 40
5 GM > 0.150
Angle for which a maximum GZ is obtained
6
> 25
7 IMO weather criterion
Table 3. MSHIP project criteria
The worst damage in SOLAS, considering water on deck,
is composed of two compartments. The following is the
graph, Figure 7 with the data obtained with the FLOOD.
7.1 (b) Design Data Sheet
The US Navy stability criteria are documented in the
DDS 079-1 (US Navy, 1975), which is divided into
criteria for damage stability for both sideprotected and
non-protected vessels. The DDS 079-1 states that an
angle of less than fifthteen degrees is required after
damage for operational requirements. There is no
mention of cross-flood systems except for in the side-
protected vessels, which states that the maximum list
shall not exceed twenty degrees and that arrangements
exist for rapidly reducing the list to less than five
degrees. The current stability criteria used by the US
Navy were developed during and shortly after World War
II (Sarchin and Goldberg, 1962). In the following graph,
Figure 7, we have done the comparative between criteria.

T Criteria KGMAX GMMIN

4.51 12245.90 7 12.38 3.70
5.37 15197.30 7 12.92 2.16
6.24 19142.60 7 13.20 1.07
7.10 22521.90 7 13.22 0.80
Table 4. Limit values for the MSHIP project

Criteria
Description
number
1 Arms rate ( wind & rolling ) < 0.6
2 Areas rate ( wind & rolling ) > 1.4
3 Heeling angle in turning < 15
4 Arms rate ( turning ) < 0.6
5 Areas rate ( turning ) > 1.4
Table 5. MSHIP project criteria

T Criteria KGMAX (m) GMMIN Figure 7. Comparative between the criteria after damage
4.51 12245.9 2 9.14 3.01 for the Merchant ship
5.37 15197.3 2 10.46 1.56
7.1 (c) Naval Ship Code
6.24 19142.6 1 11.35 0.90
7.10 22521.9 1 11.36 0.30
The damage categories, in the NSC, are based on defined
Table 6. Limit values for the MSHIP project
shapes:
7.1 (a) Stockholm Agreement
Sphere. To be used for explosions. For explosions
detonating against the outside of the hull, half the
SOLAS implies safety, but by no means applicable to all
sphere to be used.
types of vessel. Mainly because many of its rules are
unworkable or unrealistic for the warships. The criterion Cube. To be used to define the volume directly
SOLAS begins by defining the extent of damage to affected by fire and which may change in shape to fit
consider. These dimensions, based on statistics of failure, the compartment.
are defined as a fault length equal to 3% of the length Raking/grounding. To be used in the appropriate
plus three meters, a penetration of damage equal to B/5 horizontal orientation to describe the extent of raking
and a height of damage that goes from bottom to top or grounding damage, the apex representing the
without limit (Riola and Perez Fernandez, 2009). maximum penetration.
The Damaged Ship, London, UK

Collision. To be used in the correct vertical
orientation to describe the extent of collision damage
from the bow of another ship, the apex representing
the maximum penetration.
The extent of the worst damage category is defined as
damage category C, significant: sphere with 10 (m) of
radius, cube with 20 (m) of sides, raking/grounding with
40 (m) of length and 5 (m) of equal sides and collision
damage with 40 (m) of height and 5 (m) of equal sides.
The temperature is heat caused by initiating event
assuming no other combustion.
mind when assessing roll on-roll off safety. The
Stockholm Agreement was created on the presumption
that a vessel designed, or modified, to SOLAS 90
standards ensures survival at sea states with Hs of only
1.5 (m). This was suggested in the face of uncertainty
and lack of understanding of the phenomena involved.
The evidence amassed so far and presented in the
following suggests that this was a considerable
underestimate. The maximum penalty of 0.5 (m) height
of water on deck is ill based.
In the Figure 9, we have compared the different criteria
after damage for the warships, using the SOLAS
Convention plus Stockholm Agreement.

7.2 (b) Design Data Sheet

The IMO weather criteria considers wind with gusts and

a roll-back angle which is dependent on the ships static
righting arm and other ship roll characteristics
(International Maritime Organisation, 1994). The US
Navy and other Navies have not kept pace with IMO
developments. They continue to rely on the empirical
World War II criteria until the more sophisticated
methods are developed and validated. Validation and
acceptance of these new methods may take some time.
Current naval ship can be greatly improved with a few
Figure 8. Comparative, using the Naval Ship Code,
small changes which maintain the integrity of their basic
between the criteria after damage for the merchant ships
approach, and increase their commonality with the IMO
criteria.
In the Figure 8, we have compared the different criteria
after damage for the merchant ships, using the NCS
In the following graph, Figure 9, it has been done the
explained before.
comparative between criteria.
7.2 RESULTS IN WARSHIPS

We have done the same calculations for the warship. The

calculations are made on intact to see if the WSHIP
complies with the stability intact, and that if not fulfilled,
the values were obtained at the end of the study would be
worthless.

T Criteria KGMAX (m) GMMIN

3.65 4069.5 7 12.43 3.64
4.43 4801.9 7 13.03 2.23
5.37 6364.5 7 13.30 1.19
6.21 7488.1 7 13.40 0.68
Table 7. Limit values for the MSHIP project

T Criteria KGMAX (m) GMMIN

3.65 4069.5 2 11.46 2.59
4.43 4801.9 2 11.01 1.83
5.37 6364.5 1 10.46 1.42
6.21 7488.1 1 9.85 1.23
Table 8. Limit values for the MSHIP project Figure 9. Comparative between the criteria after damage
for the warship
7.2 (a) Stockholm Agreement
7.2 (c) Naval Ship Code
There are certainly some obvious weaknesses in the
requirements of the Agreement and this must be borne in As we have explain before, the extent of the worst
damage category is defined as damage category C, and of

the four types, there is one, raking/grounding, which is
the worst of all, proof that our ship, WSHIP meets all
known criteria, will not tolerate a failure of forty meters
in length in the double bottom. Therefore, for
comparison between the criteria, we will not use the
failure of raking/grounding which was defined in the
NSC. A comparative has been done with the data
obtained with the FLOOD and its corresponding Figure
10.
Figure 10. Comparative, using the Naval Ship Code,
between the criteria after damage for the Warships
In the Figure 10, a comparison of the different criteria
after damage for the warship has been done, using the
NCS explained before.
8. CONCLUSIONS
In this paper, we have proposed a comparative analysis
of the different criteria of stability after damage.
For this research the various studies and calculations
have been carried out on a designed test vessel. We have
created a vessel to comply with different conditions, like
having an empty deck, without pillars, one propeller
shaft and whose forms are as close as possible to a
warships shape. Having created a ship, that by its nature
would be a warship, she might consider that the criteria
are compared in this article, on the same ship. Figure 9
presents a chart that summarizes the behaviour of each
criterion.
A most important conclusion to emphasize, that while the
approach of the British Royal Navy is more restrictive
than the US Navy, if we are considering the Stockholm
Agreement to SOLAS, is that this convention is the most
restrictive of all. If water is seen on deck, no military
approach is more restrictive than the IMO. It means that
the Navies never have considered in their calculations
water on deck. It is true that due to civil Roll on-Roll off
accidents, during the nineties, the IMO started
considering the water on deck as dangerous and it was
incorporated an annex to the SOLAS, to take account this
problem. Although the Navies never have considered this
kind of flood, it seems that this problem has a critic value
in the ship damage stability analysis. In concrete, this
The Damaged Ship, London, UK
damage should be investigated, if the ship has a low
freeboard.
Depending of these damages, we expose a comparison
between the NSC damages in the warship studied, see
Figure 10.
There are many areas where military vessels could
improve safety standards, although not necessarily to be
regarded as less secure than the civil vessels. It is up to
every government and authority the establishment of
naval security level to offer their equipment and how it is
achieved.
In some countries, warships frontline are built in
accordance with rules of the Classification Societies,
private agencies for their requirements ensuring
compliance with building regulations stricter than the
requirements of SOLAS. An example of this is that ninety
percent of the British fleet is classified, in part or in
whole, under Lloyds Register of Shipping or DNV
(Ingram, 2007). However, there are major difficulties in
implementing all the rules of the Classification Societies
at the naval field (Boral, Gurley Tar Becker and
Humphrey, 2005); especially to establish a priority
mission and capacity combat against security. It is
important to distinguish the importance of the new rules
NSC. As we have tried to reflect throughout the paper,
the NSC has become the criterion of stability in damage
than more is acclimating to the standards of the Navies in
the XXI century. For each type of vessel could be a
priority for study in terms of damages of the NSC.
A warship will be more likely to suffer damage by
explosion than a civilian ship, therefore a civil Roll on-
Roll off will have more risk by collision than by
explosion. Instead a landing ship shall be by chance more
risk having grounding or raking in its bottom. Clear
examples are frigates, corvettes and patrol vessels, where
probabilistically, wouldn't water board problems on deck
from the waterline. The grounding is a very dangerous
flood for any type of vessel, but depending on its size
damage and forms can be considered from serious to
very serious.
This technical paper ends giving a method that helps and
supports the naval architect in the analysis of damage
stability. In this way, the naval engineer is able to decide
what of the existing criteria better fits the requirements of
the ships, as a function of a few principles. For example,
depending of the ship typology is possible to know which
possible damage type is more probable and which criteria
is more restricted for each one of them.
9. ACKNOWLEDGEMENTS
We are heartily thankful to our families whose
encouragement, guidance and support from the initial to
the final level of the technical paper.
The Damaged Ship, London, UK

Lastly, we offer our regards and blessings to all of those 11. AUTHORS BIOGRAPHIES
who supported us in any respect during the completion of
the technical paper. Rodrigo Prez Fernndez holds a PhD in Naval
Engineering at the Universidad Politcnica de Madrid
(Spain). He has worked at SENER within their customer
10. REFERENCES support department since 2005. During that time, he has
had the opportunity to partner with a number SENER's
1. ATHANASSOULIS, G. and SKARSOULIS, M., Wind customers, including two years expatriation in the UK as
and Wave Atlas of the North-Eastern Mediterranean Sea, the main liaison for the development of the new fleet
NTUA-SMHL Publications, 1992. carriers. During his time with SENER, he has also been
2. BOULOUGOURIS, E.K. and PAPANIKOLAOU, A.D., in customer engagements in Malaysia, Brazil, China,
Optimisation of the Survivability of Naval Ships by South Korea, Poland and Norway.
Genetic Algorithms, 3rd Int. Euro Conference on
Computer Applications and Information Technologies in Jos M Riola Rodrguez holds a PhD in Naval
the Maritime Industries. COMPIT04, 2004, Saguenay, Engineering at the Universidad Politcnica de Madrid
Spain. (Spain). Currently, he is the director of the Center for
3. BORAL, P.; GURLEY, G.; TAR BAKE, E. and Sistema de Observacin y Prospectiva Tecnolgica,
HUMPHREY, R., Development of a NATO Naval Ship General Subdivision of Technology of the Spanish
Code, Safety regulations and Naval Class II. London: The Ministry of Defense. Dr Riola is also teaching at the
Royal Institution of Naval Architects, 2005. Department of Naval Architecture and Construction in
4. INGRAM, T.J., Application of commercial ship the Escuela Tcnica Superior de Ingenieros Navales at
maintenance philosophy to naval ships, American Bureau the Universidad Politcnica de Madrid.
of Shipping, 2007.
5. PREZ FERNNDEZ, R., Estudio de la adecuacin y
suficiencia de los kgs lmites para el cumplimiento de
los criterios de estabilidad en los diversos campos, para
los buques de carga rodada (mercantes-militares), 48
Congreso de Ingeniera Naval e Industria Martima, 2009,
Vigo, Spain.
6. PRZEMIENIECKI, J.S., Mathematical Methods in
Defense Analyses, 2nd Edition, American Institute of
Aeronautics and Astronautics, 1994.
7. RAHOLA, J., The judging of the Stability of Ships and
the Determination of the Minimum Amount of Stability,
Transactions. Institute of Naval Arquitects, 1935.
8. RIOLA, J. M. and PREZ FERNNDEZ, R., Estudio
comparativo entre los criterios de estabilidad de la Armada
Norteamericana, Britnica y del SOLAS, Revista
Ingeniera Naval, julio-agosto, 2009.
9. RIOLA, J. M. and PREZ FERNNDEZ, R., Criterios
de estabilidad de la armada norteamericana, britnica y
SOLAS, Boletn de observacin tecnolgica en defensa,
4 trimestre, 2009.
10. RIOLA, J. M. and PREZ FERNNDEZ, R.,
Warship damage stability criteria case study, Journal of
Maritime Research, Diciembre, 2009.
11. SARCHIN, T.H. and GOLDBERG, L.L., Stability and
Buoyancy Criteria for the U.S. Naval Surface Ships,
Trans. SNAME, Vol.70, pp. 418-458, 1962.
12. SURKO, S.W., An Assessment of Current Warship
Damaged Stability Criteria, Naval Engineers Journal,
Vol.106, No.2, pp. 120-131, 1994.
13. US NAVY, Naval Ship Engineering Center, Design
Data Sheet-Stability and Buoyancy of US Naval Surface
Ships, DDS 079-1, US Navy, currently Naval Sea Systems
Command, 1975, Washington, DC, EEUU.
The Damaged Ship, London, UK

THE APPLICATION OF HIGH PRESSURE WATER MIST AS PART OF A HOLISTIC

FIRE FIGHTING SYSTEM

S E Ratcliffe, Atkins, UK

SUMMARY

Current accommodation standards consume significant space on board ships. The development of a holistic approach to
fire fighting would enable a reduction in the requirement for trained ships crew and in turn in accommodation. This will
drive savings in a reduced hull size, propulsion requirements, and reduce the number of people exposed to risk in the
event of a fire.

This study aims to explore the use of High Pressure Water Mist (HPWM) to provide ship wide reactive fire fighting by
considering its application to a generic area of the Type 26 Frigate in peace time and battle scenarios. This study also
offers a brief discussion of the surrounding issues associated with implementing such a system, including estimated
Technology Readiness Levels.

ACRONYMS underlying factor prevents ships crews from being

reduced below a minimum level.
BD Battle Damage
BGCV Branch Group Control Valve The development of a fully automated and holistic
CMS Combat Management System approach to fire fighting onboard Royal Navy (RN) ships
DNV Det Norske Veritas would enable a reduction in the requirement for trained
FAR Firefighters Assistance Robot ships crew. With fewer crew to support this would allow
FN Frame Number ships to be designed with a smaller hull size whilst still
FPA Fire Protection Association delivering the same capability. This will see savings in;
FSC Future Surface Combatant
GB Glass Bulb the cost of building the ship
HPWM High Pressure Water Mist reduced propulsion requirements
IMO International Maritime Organisation reduced ships services requirements
IR Infrared lower ships emissions
NDP Naval Design Partnering Team The number of personnel exposed to risk in
O Open the event of a fire.
PD Positive Displacement
PT Peace Time Ultimately, investment in a modern fire fighting solution
RN Royal Navy will bring savings in the Through Life Cost of a platform.
TRL Technology Readiness Level
US United States (of America)
2. SHIPS FIRE FIGHTING PRINCIPLES
1. INTRODUCTION The following section outlines the basic principles for
optimised fire prevention, fighting and protection. The
The cost of a ship relates directly to its size, both in the FPA report [1, Paragraph 3.2] discusses a progressive
concept phase and in-service. Being able to reduce the and scaled approach to fire protection, beginning with
size of a ship enables savings to be made over an prevention in the first instance and leading through to fire
otherwise larger vessel. However, one of the most fighting as a last resort. This approach is summarised
significant drivers for the size of a ship is the manpower below.
needed to crew it. Current crew living standards consume
significant space due to accommodation, stores, black
and grey water requirements etc.

Where advancements in technology, particularly

navigation and combat management systems, have
enabled the number of crew needed to sail the ship to be
reduced, these savings are often not realised. The
sticking point is the manpower required to fight fires
onboard, especially in damage and threat situations.
Current fire fighting techniques and systems require a
certain amount of manpower in order to function. This

Figure 1.1 - Reducing Through Life Costs by an Increase
in Automation
2.1 PREVENTION AND PROTECTION
In any fire management policy, the avoidance of fire
should play as great a role as fire protection. The first
step of fire prevention encompasses removing and
reducing fire hazards and risks to avoid any incident.
Techniques for doing so might include changing
process to reduce failure paths that may lead to fire, or
changing the design of equipment to remove fire hazards
altogether.
Fire prevention also includes the detection and prediction
of fire scenarios long before the outbreak of fire. Fires
and explosions are generally preceded by a series of
events that lead step by step to the fire situation.
Depending on the nature of these events, there may be
opportunities to intervene and detect a fire scenario early.
Detecting these events leading up to a fire is often
outside the boundary of traditional fire detection systems.
For example, monitoring an engines vibration, fuel
consumption and pressure may provide clues as to
whether or not mechanical failure which could lead to
fire is imminent.
2.2 DAMAGE LIMITATION AND RESPONSE
The damage and consequence resulting from a fire can be
characterised into two areas; the primary damage
resulting directly from the fire and its by-products, and
the secondary damage and consequences of the
implemented fire suppression strategy. The extent of the
damage resulting from a fire and its suppression is a
function of the scale of the fire and hence the scale of the
response. The earlier that fires are detected, prevented
and suppressed, the greater the benefits are for reducing
both primary and secondary damage. Early detection also
allows a better balance between a rapid but high impact
response against a measured and proportionate (directed)
response causing less overall damage.
The Damaged Ship, London, UK
Traditionally, ships fire fighting methods are not
predictive by nature and, following the failure of 1st Aid
fire fighting, accept large amounts of damage as
alternative means are readied. The modernising of fire
fighting response requires the replacement of the man
with an extinguisher with automated on the spot
alternatives such as equipment control, local suppression
and aimed systems designed to reduce the impact of
secondary damage.
2.3 REDUNDENCY AND ITEGRITY OF
SYSTEMS
Fire safety systems and the supporting infrastructure
generally embody safety factors to ensure availability
and effectiveness when required. These factors may
include;
System duplication
System over-specification
System protection (resource guarantee)
Manual backup
Assigning worst case threshold and
designing infrastructure to cope with it.
3 HIGH PRESSURE WATER MIST
SYSTEMS
3.1 GENERAL DESCRIPTION
In suppressing a fire, traditional low-pressure sprinkler
and deluge systems often cause significant water damage
that can be greater than the damage caused by the fire
itself.
HPWM offers equivalent or better fire suppression than
traditional systems with minimal water discharge,
minimising damage to property and reducing the time
and cost of clean-up.
High pressure water is dispersed by fixed nozzles which
create an ultra fine mist over the protected area. The mist
fights fires in three main ways;
Cooling Millions of tiny water droplets
produce a very large heat collecting surface
and rapidly reduces the temperature of the air
in the space.
Smothering The vapour displaces the oxygen
volume in the fire itself, rather than in the
entire space. This means that it does not
present an asphyxiation hazard to personnel.
Attenuation The mist absorbs radiant heat.
Ultra fine mist has the advantage that it requires very
little water and consequently does minimal damage to
equipment. If de-ionised water is used HPWM systems
can also be applied to live electrical fires. In machinery
spaces the major benefit of water mist is that since it is
The Damaged Ship, London, UK
harmless to people, the system can be activated the
second a fire is detected, without any need to first
evacuate. Nor is there any need to shut off vents or close
openings before evacuation, as the water mist will not
escape the space, as gases would. This possibility for
immediate activation means that the fire damage is kept
at a minimum.
Once the fire has been extinguished, the water mist will
quickly cool down the space and thus prevent re-ignition.
3.1 (a) Flash Suppression
A fire protection system designed to provide flashover
suppression aims to keep air/gas temperatures in
compartments too low for materials and fuel sources to
ignite. One of the advantages of flashover suppression is
that it can be achieved with less water with a high
pressure mist system than a traditional sprinkler
extinguishing system. Live fire tests conducted by the US
Navy on the ex-USS Shadwell have demonstrated that
flashover suppression can be achieved using fewer water
mist nozzles in each compartment than would be needed
using a conventional marine sprinkler system [2].
3.1 (b) Pre-emptive Action
A networked system of HPWM sprinklers could be used
to pre-emptively cool certain spaces. This might be a
reaction to developing fire conditions, i.e equipment
telemetry reporting increased risk. or as a precautionary
measure during a fire scenario. For example, water mist
could be used to cool an established escape route
reducing the chance of it being blocked by fire.
3.1 (c) Blast Mitigation
The use of water mist has been shown to have benefits
for mitigating the effects of blasts on Navy ships. Navier-
Stokes simulations performed by Ananth et al [3] and the
experimental results found by Thomas et al [4] suggest
that latent heat absorption is the primary mechanism
behind water mist explosion suppression in a confined
space. The shock front that propagates ahead of the
thermal front immediately following a detonation causes
the water mist droplets to break up; this increases the
heat absorbing surface area and results in an increase in
the droplet vapourisation rate. This cools the gasses in
the region between the shock and thermal fronts.
The second mechanism by which water mist mitigates
blast energy is through momentum absorption.
Simulations by Schwer and Kailasanath [5] concluded
that quasi-static pressures produced by small explosions
were suppressed by water mist. The droplets interact with
the front as it is reflected multiple times absorbing
energy and changes of momentum.
In 2009 the US Naval Research Laboratory conducted a
series of detonation experiments designed to establish a
link between the mist density conditions of a sprinkler
system and its ability to reduce blast impulse in a
confined space [6]. The results of these experiments
show that the higher mist density conditions
outperformed lower mist concentrations in suppressing
blast effects. This shows a significant opportunity for
building water mist into a ships defensive suite.
3.2 SYSTEM CONFIGURATION
Water mist systems can be configured in various ways;
Thermally Activated (Wet pipe) - Wet pipe systems are
typically used in accommodation and similar areas where
solid materials are the combustible media. When ambient
temperature exceeds a given limit, the activation bulb of
the sprinkler bursts and water mist is discharged from
that particular sprinkler.
Deluge - A deluge system normally has open spray heads
with water flow controlled by closed valves. When a
valve is opened, water mist is discharged by all spray
heads in the section controlled by that valve. Deluge
systems are typically used in spaces where fuel fires
could occur.
3.2 (a) Class A Fires
Class A, or ordinary combustible fires are those started
from solid organic material, such as wood, paper, cloth
etc. Typically, spaces containing Class A fire hazards
are protected by thermally activated sprinklers. I.e
localised heat sources cause the sprinkler in the
immediate vicinity to activate. Class A fires typically
spread outwards from a single point, so initially only the
closest sprinklers will open, tackling the fire directly. As
the fire spreads, more nozzles will open and the water
demand from the system will increase.
3.2 (b) Class B Fires
Class B, or flammable liquid fires, are those started from
liquid fuels, oils, chemicals etc. Machinery spaces
present numerous Class B hazards. Since highly
flammable liquids like petroleum can spread quickly and
set alight almost instantaneously, it is necessary to cover
the entire area with water mist from the offset. In such
spaces open nozzle deluge systems are used.
4 SYSTEM DESIGN AND APPLICATION
4.1 SUBJECT AREA T26 BASELINE 2v0
For the purpose of this study, an area of T26 concept
design was selected to provide a framework on which to
demonstrate the use of HPWM. The deck area was split
into four Fire Zones separated by water tight bulkheads.
 The Damaged Ship, London, UK

For the purpose of this study, one fire zone across two
decks has been isolated for use as a representative area
for the application of HPWM.
Frame 71 to Frame 103 on 2 Deck T26 Baseline 2v0 was
chosen to act as a single Fire Zone. This zone
represents a typical area on a Surface Combatant. It
houses common functions and utility spaces that
represent a range of potential fire conditions.
4.2 COMMERCIAL WATER MIST SYSTEM
SELECTION
In order to design for and demonstrate the application of
a HPWM system in this study, it was necessary for a
commercially available HPWM System to be selected for
inclusion in the design. In this case the HI-FOG Marine
System by Marioff, Finland was chosen on the basis of
its pedigree as a marine fire fighting system.
4.3 NOZZLE LAYOUT
Nozzle layouts have been mapped for 2 and 3 Decks
between Frame Numbers 103 and 71. See Figure 2 and
Figure 3 below.
1MB or 1MC 1 mist orifice on centre axis,
B = 0.7mm diameter, C = 1.0mm diameter
1B or 4S 1 = 120 degree cone angle,
thermally activated; 4 = 90 degree come angle,
open nozzle; B = Brass; S = Stainless Steel
b) 2mm, Job 57oC (orange code) bulbs
c) Maximum system working pressure is 140 bar, while
minimum initial working pressure at the sprinkler heads
is 120 bar. A minimum of 100 m2 is to be covered at 120
bar pressure, whereas 280 m2 should be covered at
minimum 60 bar (measured at the nozzle).
d). This sprinkler head may also be used at ceiling height
of 3.0m and 3.5m. The sprinkler head spacing should
then be reduced to 3.30m and 3.05m respectively.
e) Spray heads should be installed outside of the
protected area a distance of at least 1/4 of the maximum
nozzle spacing, in this case 1.0 m outside at the periphery
of the protected object (see IMO MSC/Circ.913, annex
3.4.2.1).

The spacing rules are shown in Table 1 Sprinkler

Spacing and Layout Requirements. These were derived
from the Type Approval Certificates for the HI-FOG
System issued by Det Norske Veritas (DNV) [7] [8].

The HI-FOG system meets the fire test protocols for

Accommodation areas, public spaces and service areas
on civilian ships. The International Maritime
Organisation (IMO) terminology of Accommodation
Areas, Corridors, Public Spaces and Service Areas
correspond in RN terminology to Small spaces,
Passageways, Large spaces and Storage areas
respectively.

For the purposes of this report, both open deluge and

closed thermally activated nozzle types are used. It has
been assumed that Branch Group Control Valves
(BGCV) will control the flow to each branch pipe. This
will allow nozzles to be activated remotely without the
need for the local temperature being high enough to
thermally trigger the bulb. Thus, the system will still
maintain the reactive nature of thermally activated bulbs
but also allow a central control system to pre-emptively
open branch groups to protect spaces in advance of fire
or blast situation.

4.3 (a) Notes on Table

a) Nozzle designation code: 1B 1MB 6MB or 4S 1MC

8MB

6MB or 8MB 6 or 8 mist orifices

surrounding the centre jet,
The Damaged Ship, London, UK

Figure 2: Fire Zone 2, 2 Deck Arrangement and Sprinkler Layout

Figure 3: Fire Zone 2, 3 Deck Arrangement and Sprinkler Layout

 The Damaged Ship, London, UK

Public spaces, Public spaces, 2 Corridors and

IMO Compartment Cabins, <16m2 Storage areas Machinery spaces
>16m2 deck height stairways

Large
Cabins/Small Large Passage and
Naval Equivalent compartments, 2 Storage areas Machinery space
Compartments Compartments stairways
deck height

Vertical distance from

Max Deck Height 2.5 m 2.5 m 5.0 m 2.5 m 2.5 m
object 1.5m 5.5m

Nozzle Type a) 1B 1MB 6MB 1B 1MB 6MB 1B 1MC 6MC 1B 1MB 6MB 1B 1MC 6MC 4S 1MC 8MB

Symbol

Glass Bulb b) (GB) or

GB GB GB GB GB O
Open (O)

K-factor [lpm/bar ] 1.45 1.45 2.5 1.45 2.5 1.9

Flow rate at 120 bar

15.9 15.9 27.4 15.9 27.4 20.8
[lpm] c)

Max spacing One per room 3.5 m d) 3.5 m 3.75 m, centred 2.65 m 4.0m e)

Max distance to
2.850 m 1.750 m 1.750 m 1.875 m 1.325 m -
bulkhead

Max coverage area 16 m2 12.3 m2 12.3 m2 14.1m2 7 m2 16 m2

Nominal water
1.0 1.3 2.2 1.1 3.9 1.3
density [lpm/m2]
Table 1: Sprinkler Spacing and Layout Requirements
The Damaged Ship, London, UK
4.4 DISTRIBUTION ARCHITECTURE OPTIONS
In this section, different possible High Pressure Water
Main distribution architectures will be discussed. In
particular, the way in which they might contribute to
enabling a holistic and survivable water mist system
capable of intelligent self diagnosis and damage control
will be considered. Key factors are the degree of
redundancy in pump sources, the number of routes
available for water to reach any portion of the piping
network, the degree of separation between redundant
components and the arrangement of control valves and
sprinklers. Other factors that effect survivability are the
integrity of the power supply to pumps and valves,
communications and logic systems, location, mounting
and armouring of risers and valves etc. all of which are
excluded from this discussion
Three distribution architectures will be considered;
Centre main distribution
Dual main distribution
Sectional-loop architecture
4.4 (a) Centre Main
Centre Main architectures consist of a single main
running down the centre line on each deck of the ship.
Sectional control valves are located at the zone
boundaries and at riser connections, which are spaced so
to bring water from the pumps on the lower decks to each
level. Sprinkler heads are connected in branch groups
which in turn are connected to the centre main via a
BGCV. If a pipe was to be ruptured between risers on
one deck, valves on the mains would close to isolate that
length of pipe. Water would still be able to flow to the
intact portions of the main on that deck through the
adjacent risers. The centre main in this case must be large
enough to accommodate the combined flow of all the
branch groups.
The level of redundancy in the centre main design is low;
the options for re-routing the water around isolated
sections is limited and as a result the number of branch
groups that can be kept active in the event of damage is
low.
4.4 (b) Dual Main
A step further from the Centre Main concept is the use of
a dual main. This features mains running down both the
port and starboard side of each deck with vertical risers
spaced at intervals along the length of the ship supplying
the mains on each side. Branch group lines are connected
to either the port or starboard main via BGCVs. Zonal
valves are placed at intervals along the two mains so that
damaged sections can be isolated. Any undamaged
branch lines within a compartment that are fed from the
undamaged side of the ship will still be functional. This
architecture can be modified to include crossover mains
and valves to create offset loops.
4.4 (c) Sectional Loop
Sectional Loop Architectures are similar to dual main
arrangements in that a high pressure main runs along the
port and starboard side of each deck of the ship, with
vertical risers placed at regular intervals along the length
of the ship. The network of pipes is separated into loops
by valves and crossover mains. Each loop is served by it
sown riser, which can supply that loop with water or any
of the loops adjacent to it. Risers alternate between port
and starboard sides along the length of the ship in order
to minimise vulnerability. Unlike the dual main
architecture however, crossover mains connect port and
starboard mains on the same level, one on either side of
each water tight bulkhead.
The advantage of sectional loop architecture is the ability
it has to recover from damage with minimal loss of
functionality. Valves are placed so that each loop can be
supplied in two independent ways;
With all the supply coming up the riser serving
that particular loop;
The riser can be closed off so that the supply
must come from an adjacent loop.
Sectional Loop Architectures allow for the subdivision of
the main network into small cells that can be individually
isolated. The greater the number of cells the better the
ability to recover from damage while leaving as much of
the network operational as possible.
Sectional Loop Architectures also offer hydraulic
advantages in that the number of pathways for water to
flow to any one demand point is maximised. A large
demand at a particular point will distribute across
multiple mains and risers, thus allowing for lower flow
rates through the mains, requiring smaller diameter
piping [9, Paragraph 2.4].
It is for these reasons that using a Sectional Loop
Architecture offers the greatest potential for providing a
flexible and survivable HPWM system. Sectional Loop
Architecture will be considered from here on.
4.5 BRANCH GROUP LAYOUT
Figure 4 and Figure 5 below show the nozzles and
branch lines connected to the main on each deck.
Typically a branch group connects four nozzles to the
main and is controlled by a branch group control valve.
Where possible, branch lines are arranged so that nozzles
can provide a curtain of water across the beam of the
ship. Each nozzle would be connected to the branch line
by a branch vein (not shown), sized to only need to carry
the flow for one individual sprinkler, regardless of where
it is in the system i.e regardless of how many sprinklers
are in front of it.
 The Damaged Ship, London, UK

Figure 3: Sectional Loop Architecture

The Damaged Ship, London, UK

Figure 4: General Arrangement of 2 Deck showing Sprinklers, Branch Groups and water main

Figure 5: General Arrangement of 3 Deck showing Sprinklers, Branch Groups and water main.
 The Damaged Ship, London, UK

4.6 WATER FLOW DEMANDS and the separation of the units ensures that the risk of
losing both through battle damage is greatly reduced.
A system configured to provide flashover suppression
cover across an entire platform has the advantage that its However, option B may not offer the scalability or
flow demand grows progressively to match the spread design flexibility required when considering much larger
and intensity of the fire. As the fire spreads through the or much smaller ships. As the number and size of fire
ship, more branch groups can be brought online as zones increases the demand flow rate will also increase.
needed. For the purpose of this study, the water flow Sizing a single pump unit to provide 100% of the design
demands will be calculated for two damage scenarios; flowrate in all circumstances may not be sensible.
peace time and battle damage scenarios.
4.7 (b) Option C
Table 3 compiles the nozzle count and estimated water
flow demands for the HPWM system based on the Option C maintains a reasonable level of redundancy in
layouts shown in Figure 4 and Figure 5. For simplicity, it four pumps each providing a third of the design flowrate.
has been assumed that each branch group holds four The use of one pump per fire zone ensures that pump
sprinkler heads with a K-factor of 1.45 lpm / bar, units are adequately separated and distributed throughout
except for the Machinery Space where a K-factor of 1.9 the ship. Options C offers advantages over Option B in
lpm / bar is used. the size and the scalability of the pumps required.
Running several smaller pump units rather than relying
4.7 PUMPING STRATEGY on large pumps sized to meet the entire demand should
enable savings to be made in running costs and enable a
There are several possible approaches for providing the more flexible design. The remainder of this study
pumping capacity for the HPWM system. Some high assumes Option C.
level options are discussed in Table 2 below.
4.9 PUMP TYPE
Option High Level Pumping Comments
Strategy Selecting a pump to supply a high pressure water mist
A One large pump unit Arrangement A has no system presents a challenging problem. Such a system
sized to meet the full redundancy and is not demands a very high pressure, but also requires the
design flow rate, discussed further
flexibility to vary the flow rate as the demand changes.
connected to a
distribution main Potentially, the pumping system could need to go from
supplying multiple providing high pressure fluid at a low flow rate in a
risers. single compartment to very high pressure high flow rate
B Two pump units in Provides for 100% fluid across several fire zones.
parallel, one aft of redundancy and can
midships, and one assume at a reasonable The high pressures required of the water mist system (up
forward serving a cost for connections, to 150bar) mean that the pumps selected need to be of
common distribution power and filtration. high quality and high capability. Traditionally,
main and multiple Pump units are centrifugal pumps are used with sprinkler systems
risers; each pump unit adequately separated so
sized to meet full design that at least one should
because they allow a constantly varying flow rate to be
flow, so that one unit is be functional at all delivered at with relatively simple equipment. However,
redundant. times. at high pressures, multi-stage centrifugal pumps are
C Four pump units in Provides for redundancy required. These have the potential to be more complex
parallel, one for each in a way that permits and require significant maintenance.
fire zone; each pump each pump unit to be
unit sized for 1/3 of the smaller than More reliable high pressure pumps come in the form of
full design flow, so that arrangement C, such piston type positive displacement (PD) pumps. PD
three units will meet full that redundancy can be pumps, by nature of their design, deliver a fixed volume
design flow with the achieved with three of
of fluid and as such are not best suited to variable
largest out of service. the four pumps. Can
assume that smaller demand sprinkler systems. However, PD pumps can be
pump units will be coupled with special design features to match the fixed
cheaper to run. volume output of the PD pump to a system of variable
Table 2 High Level Pumping Strategy demand.

4.7 (a) Option B

Option B affords 100% redundancy while also separating

the two pumping assets. The sizing of both pumps to
provide the total design flow rate guards against the loss
of a single pump through routine mechanical breakdown
The Damaged Ship, London, UK

Branch Flowrate
Scenario Nozzles
Groups (120bar, K=1.45)
Average branch group 1 4 63.6 Lpm
PT1 3 Branch Groups active in immediate fire area. 3 12 190.8 Lpm

PT2 3 Branch Groups active plus 1 Branch Group in corridor 4 16 254.4 Lpm
BD1 Branch Groups in adjacent spaces including 4 branch groups 7 28 445.2 Lpm
immediately above the damage area.
BD2 As BD1 but increasing to a further 2 branch groups on the same 13 48 763.2 Lpm
deck and 4 Branch Groups above damage area.

BD3 As BD2 but including further 2 on deck and 3 above 18 72 1144.8 Lpm

MS1 machinery space drench system activated (Nozzle type 4S 1MC 2 8 166.4 Lpm
8MB, K-factor 1.9)
BP1 nozzles in the outer most compartments of the ship open in the 6 24 381.6
vicinity of an expected weapons strike. This may be on one or more
decks depending on the accuracy of the prediction.
Table 3: Water Flow Demands

4.9 (a) PD Pumps Applied to Water Mist Systems

To design for the BD3 scenario (Table 3) each pumping

unit would need to be sized to provide a flowrate of
381.6 litres/minute.

The use of PD pumps to drive HPWM systems has been

demonstrated on the ex-USS Shadwell by the US Naval
Research Laboratory [9, Paragraph 2.3]. Their approach
is summarised here.

We can assume that in basic terms, each pump unit

consists of two smaller pumps, a primary and a
secondary pump, driven by a single motor. The two
pumps might be connected by a check valve, held closed
by the high pressure form the primary pump and
preventing flow from the secondary pump entering the
system. This excess flow from the secondary pump flow
Figure 6 Diagram of a PD pump pair consisting of two
might be bypassed by a suction line via a flow bypass
PD pumps driven by a single motor
valve and re-circulated to the reservoirs. As long as the
system demand is less than or equal to the capacity of the
For a large HPWM system, several pump pairs could be
primary pump, the system pressure remains high. Thus
assembled in a skid to form one pump unit The
the first nozzles to operate at the early stage of a fire
minimum demand for the system, for example one or two
deliver water mist with maximum velocity and flow rate.
nozzles operating in peacetime conditions could be met
with one pump operating. As more nozzles open, the
As more nozzles open and the volumetric demand of the
system pressure will drop until it falls below the setting
sprinkler system exceeds the capacity of the first pump,
of the next flow bypass valve. That valve then closes and
the system pressure will drop. Once it drops to the setting
the flow from that pump then enters the system to make
of the secondary pump flow bypass valve, this will allow
up the demand. In this way the pump unit self adjusts to
flow from the secondary pump to enter the system, albeit
match the demand from the sprinkler system. Figure 7
at a lower pressure. A schematic of this arrangement is
below shows an assembly of pump pairs to form one of
shown in Figure 6.
four pump units sized to meet one third of the
maximum expected demand.
 The Damaged Ship, London, UK

The basis of an intelligent sensor network would be the

employment of multiple sensor types and analysis to
diagnose whether there is a fire rather than merely
detect one of the symptoms. Multiple sensors working
together can provide more accurate clues as to whether a
fire has started or not and more importantly if a fire is
likely to start. Like the human crew, the system should
be able to make a reasoned judgement as to the fire
situation based on all its data. The advantage over the
human crew is that the ship can be monitored in all
places at once and 24 hours a day.
Figure 7 Diagram of an assembly of pump pairs
connected in parallel to provide a range of flows In order for the system to correctly match sensor readings
to fire scenarios, a database of fire signatures and sensor
patterns would need to be established for different types
5. SENSORS AND CONTROL of fire and typical false alarms. The more sensor patterns
and fire signatures the system has access to the more
Being able to detect fires quickly and accurately is key to accurate it diagnosis would be. However, given that no
ensuring a ships survivability and protecting life at sea. two fires are identical, the system would need to include
Early detection allows the ships systems and crew to deal the capacity to learn.
with the fire and limit its damage. Generally, on most
surface ships there is poor integration between sensors 5.1 NETWORK INTEGRATION
and suppression systems, limited use of multiple sensor
types and a strong reliance on human input, leading to A holistic system will rely on multiple systems talking
slow detection and numerous false alarms [10]. The to each other in order to co ordinate the response to a fire
principle fire detection system on ships is still routine or damage situation. The key elements that might be
human watch keeping; arguably simple, but slow and involved are shown in Figure 8 below. It shows the break
potentially dangerous for the personnel involved. down of the Fire Control system into two main
components; the Sensors and the HPWM system. I.e.
The aim of a fire detection system should be to provide sensing and reacting. Crossovers exist where GB
comprehensive, fully integrated, multi criteria fire sprinklers both sense and react to fire events. It also
detection cover across the entire ship, intelligently shows how the combat management system might feed
interfacing with fire suppression systems, fluid control into the fire and damage control system. Threat
systems and ships crew. Without this advanced information can be used to coordinate blast suppression
functionality, pre-emptively reacting to fire conditions response using water mist. A key challenge is the design
using HPWM systems will be limited. A novel way of of a coordinating management suite that can effectively
thinking about this is to emulate the sensory functions interface with other ships systems and coordinate action.
that human beings use to detect fire. The position of the human operator in the management
system will also present interesting questions regarding
Human Faculty System accountability and autonomy.
Optical Detection Eyes IR cameras
Machinery monitoring The control system will need to digest information
Casualty location regarding the fire itself, the suppression response, the
Real-time situational location and movement of personnel, any damage control
feedback efforts and the health of the fire fighting infrastructure.
Electronic Nose Ionising and photoelectric An ideal system will be able to react instantly to a fire
smoke detectors
situation by activating relevant branch groups and then
CO and CO2 detectors
continuing to monitor the progress of that fire and
Heat sensors
watch for new events. New fire situations might come
Acoustic Monitoring Ears Machinery monitoring
Shock/blast detection
in the form of further sensor readings from multiple
Brain Power Micro processors
sources, a manual activation at an alarm call point or the
Intelligent control bursting of a glass bulb sprinkler.
Pattern recognition
Neural networks
Voice! Alarms
Situational feedback
Personal Address
Table 4 Fire fighting systems approximated to Human
sensory faculties.
The Damaged Ship, London, UK
Figure 8: Fire and Damage Control System Hierarchy
6. SUPPORTING TECHNOLOGIES
The following section discusses some of the novel
technologies that would be necessary to enable a fully
integrated HPWM system that provides a pre-emptive
fire fighting capability
6.1 SMART VALVES
Valves that can autonomously open or close depending
on the flow conditions that they see are essential to
providing a fire fighting system that removes the need for
human input and decision making. Positioned at critical
points in the distribution main, these motorised valves
incorporate pressure sensors and flow meters to monitor
the conditions in the main.
A balance must be achieved between the benefits and the
cost of installing smart valves in the distribution mains.
The spacing between the valves determines the size of
the area that will be non-functional if valves must be
closed to isolate a rupture. When considering the
Sectional Loop Architecture, the maximum level of
control for isolating damaged piping and rerouting flows
could be achieved by installing a valve on each end of
every pipe connecting two separated grid points. So for a
T-intersection this would mean having a valve on all
three branches. This strategy would result in a high valve
count and, depending on the type of valve used, be
prohibitively expensive. An alternative might be to use
valve nodes at each T-intersection; i.e consolidating
actuators and logic circuits into one housing capable of
operating each valve individually. This would take
advantage of the proximity of the valves at each
intersecting node in the pipe network and achieve some
economies of design.
6.2 HEAT SENSORS
Heat sensors will form a key part of an advanced fire
detection system. Rising temperature conditions indicate
the increased likelihood of a fire. When coupled with
other sensor types such as smoke or visual recognition,
heat sensors can provide an accurate picture of a
developing fire and how it is spreading. This information
can be used to coordinate and prioritise the response to
the fire. Heat sensing also has the potential to provide
feedback to the control system on the effectiveness of the
fire suppression being applied.
6.2 (a) Use of IR Cameras
Infra-red (IR) cameras can also be used to spot and
identify fires. In particular, they are able to spot fires in
hot environments, where heat sensors may give false
alarms. IR cameras measure the intensity of the IR
radiation emitted from objects and surfaces. Using image
recognition software the difference between a flame flare
can be distinguished say from the hot casing of an
engine. This sort of image processing and software
recognition is not foolproof however. Flares and
reflexions might give rise for false alarms. However, IR
cameras can provide clear unambiguous information on
the situation as it develops by feeding live pictures back
to an operator. This allows them to question whether the
cause of alarm is false or a real fire and direct actions
accordingly.
7. TECHNOLOGY READINESS LEVELS
This section offers a brief discussion on the maturity of
the technology required to implement a holistic fire
fighting system. It does not seek to provide a definitive

answer with respect to technology readiness levels, but
more so a food for thought.
7.1 HPWM SYSTEM
HPWM technology is well established. It forms a natural
progression from traditional low pressure sprinkler
systems. HPWM systems are made by a variety of
specialist manufacturers and can be found currently in
use across a wide range of civilian vessels and
increasingly land installations. Military Naval use is
known, the US navy investigating its application as early
as the year 2000 [9]. HPWM systems can be found in
small fixed system set ups e.g. machinery spaces, and
also protecting larger spaces providing the primary
means of fire fighting. Class Society Certification for
passenger and cruise ship applications exists for most
systems, including the Marioff products. Estimate at TRL
8-9.
7.2 SENSORS AND CONTROL
The sensor technology required to drive a holistic system
should be widely available and in use. Smoke detectors,
heat detectors, IR cameras etc are all relatively mature
technologies. The oil and gas industry are often at the
forefront of innovation in this area, using fuel mist
detection to spot fire situations early and IR cameras to
identify fires and false alarms. The processing of the data
they gather is where the key technology questions lie.
Creating a management system that can learn the
difference between characteristics of a real fire and a
false alarm is not beyond the realms of modern
computing power but may require investment in specific
examples and programmes to drive forward. TRL 4 7
7.3 SMART VALVES
Smart valves in various forms are common across the
energy and process industries. The drive for greater
efficiencies from deep sea drilling operations has
demanded large investment from the subsea sector.
Schemes have been designed and implemented for use
off-shore, for example Unocal, now part of Chevron,
installed intelligent pump and valve systems to boost the
efficiency of its Monopod platform in the Cook Inlet
Basin, Alaska [11]. The controller technology that
powers such valves is where the key development lies
the logic behind the actuated valves. Since their use in
high pressure mains to provide automated damage repair
and optimisation response is less well documented, valve
solutions will be custom built to meet the specific needs
of their applications. TRL 7
7.4 WHOLE SYSTEM
The key to creating a holistic system is linking the
different technologies together. Sensor suites combining
with suppression systems to fight fires; potential blasts
being detected and mitigated against; intelligent
The Damaged Ship, London, UK
controllers spotting false alarms and rerouting around
damage will require the system to be greater than the
sum of its parts. This will present the greatest challenge.
For manpower savings to be realised it will require a
change of doctrine, process and attitude, away from
current techniques and further towards pre-emptive and
instant response. TRL 4
8. CONCLUSIONS
This report raises the issue that current ships fire fighting
techniques are old fashioned and slow to act failing
initial 1st aid fire fighting efforts. If ships crews are to
be reduced on future vessels, attention will need to be
paid to developing a holistic and intelligent fire fighting
system that can remove the man with an extinguisher as
far as possible.
This report sees that HPWM is an excellent solution for
providing comprehensive cover across a surface ship. It
allows for an instant response to a fire situation, in that it
is non-toxic, does not require spaces to be sealed and can
be deployed in HV and machinery spaces. The mist acts
in several ways to fight fire but can also be used to
prevent flashovers and pre-emptively cool compartments.
Water mist has also been found to offer blast mitigation.
Mist systems in outlaying compartments could be
activated ahead of a weapon strikes to reduce the
potential damage of internal explosions.
In order to provide comprehensive cover however, the
entire ship needs to be covered by high pressure sprinkler
nozzles that can either be triggered locally by rising
temperatures or on command from a control system. It is
this total coverage element that means personnel are
not required in the large numbers that currently operate
in fire fighting onboard surface ships. Installing and
supplying such a large system presents its own
complexities. Possible savings could be made if certain
areas of a platform were prioritised for HPWM cover.
This could be limited to high risk areas or priority escape
routes and passageways for personnel.
A Sectional Loop Architecture was found to provide the
most scope for enabling an effective HPWM system,
both in terms of hydraulic efficiency and protection
against ruptures. The three dimensional grid with
multiple flow paths provides hydraulic advantage in that
it enables a reduced pumping energy requirement or a
reduction in the distribution pipe size.
The use of sectional valves at crossover points either side
of each bulkhead offers protection and flexibility in the
event of blast damage to any particular section. The
subdivision of the network into small cells that can be
individually isolated, or supplied from alternate routes,
offers the advantage for developing a fast recovery from
blast damage.
The Damaged Ship, London, UK
The disadvantage of such a system lies in the expected
extra cost in labour, design space and material needed to
install the necessary crossover mains, and the number of
nozzle heads needed. The requirement of smart valves at
each pipe node may also present a significant cost.
PD pumping technology was highlighted as an
appropriate means to supply the network. PD pump pairs
could be arranged in parallel in a skid to achieve a range
of flow demands. Each fire zone should contain its own
pump unit and each should be sized to meet a third of the
demand. Water would need to be provided from a fresh,
de-ionised source and could not be supplemented from
the Sea Water Main.
A truly holistic system will require advanced control and
management. The integration of smart sensors, a control
system and human operators will be vital to ensuring the
full implementation of an intelligent fire fighting system.
A range of telemetry and joined up sensing processes
need to be combined with a sophisticated control system
that can diagnose fire as well as detect its symptoms.
The more automated this detection and diagnosis process
is, the greater the potential for saving time taken to
respond to a fire and the scale of the response required to
control it. Integration with Combat Management System
(CMS) offers potential for building in reactive
protection.
Aside from the technical challenges this presents, it will
also require comprehensive rethinking of how fires and
damage control are managed presently and how they are
pictured in the future. A full analysis of the costs of such
a system would need to be assessed against the savings
made in manpower and ships size, also taking into
account the reduction in risk to personnel as a result of
the system being implemented. Provisionally however,
the use of HPWM is recommended for meeting the
reduced manning objectives set out by this report.
Coupled with smart sensor and control technology it
enables a swift and intelligent response to fire across the
entire ship with little or no need for the man with an
extinguisher.
9. REFERENCES
1. FIRE PROTECTION AGENCY, NA-FS Firefighting
Improvement Initiatives as part of the Future Surface
Combatant (FSC) Fire Policy, Final Report (Revision 2),
Ref. FPA/001020, 17 March 2009.
2. MAWHINNEY, J.R., DINENNO, P.J., and
WILLIAMS, F.W., Water Mist Flashover Suppression
and Boundary Cooling System for Integration with DC-
ARM: Summary of Testing, NRL Memorandum Report
8400, September 30 1999.
3. R.ANANTH, H.D.LADOUCEUR, H.D.WILLAUER,
J.P.FARLEY, F.W.WILLIAMS, Effect of Water Mist
on a Confined Blast, Presented before the Suppression
and Detection Research and Applications A Technical
Working Conference (SUPDET 2008), March 2008
Orlando Florida.
4. G.O.THOMAS, A.JONES, M.J.EDWARDS,
Influence of Water Sprays on Explosion Development
in Fuel-Air Mixtures, Combustion Science and
Technology, 1991, 80, Pages 47-61
5. D. SCHWER, K. KAILASANATH, Blast Mitigation
by Water Mist (3) Mitigation of Confined and
Unconfined Blasts, Center for Reactive Flow and
Dynamical Systems, Laboratory for Computational
Physics and Fluid Dynamics, NRL/MR/6410--06-8976,
July 14, 2006
6. Heather D. Willauer, Ramagopal Ananth, John P.
Farley, Gerald G. Back, Victor M. Gameiro, Matthew C.
Kennedy, John OConnor Frederick W. Williams, Blast
Mitigation Using Water Mist: Test Series II, Navy
Technology Center for Safety and Survivability,
Chemistry Division, NRL/MR/6180--09-9182, 12 March
2009.
7. Det Norske Veritas Type Approval Certificate,
Certificate number F-18732, 7 August 2008.
8. Det Norske Veritas Type Approval Certificate,
Certificate number F-18536, 16 November 2007.
9. JR MAWHINNEY PJ DINENNO, New Concepts for
Design of an Automated Hydraulic Piping Network for a
Water mist Fire Suppression System on Navy Ships,
Naval Research Laboratory, Ref NRL/MR/6180-01-
8580, September 2001
10. MICHELLE PEATROSS and DR FRED
WILLIAMS An Overview of Advances in Shipboard
Fire Protection, Hughs Associates Inc
11. JIM BANKS Take Control: Smart valves Step
Forward, offshore-technology.com, dated 18 June 2008,
viewed October 2010
10. AUTHORS BIOGRAPHY
Simon Ratcliffe graduated from Bristol University in
2008 with a Masters Degree in Mechanical Engineering.
He joined Atkins Defence business as a graduate
engineer shortly afterwards. As part of the graduate
training programme Simon spent 4 months on
secondment at the Naval Design Partnership where he
had the opportunity to research and write this paper
alongside his project work.
The Damaged Ship, London, UK

ESTIMATION OF ORIFICE FLOW RATES FOR FLOODING OF DAMAGED SHIPS

C D Wood, A J Sobey, D A Hudson and M Tan, University of Southampton, UK
P James, Lloyds Register, UK

SUMMARY

This paper concentrates on the modelling of compartment flooding rates following the occurrence of damage. Typical
state of the art flooding models use Torricellis formula to calculate flooding rates using a constant co-efficient of
discharge (Cd). Based on Bernoullis theorem, turbulence and viscosity effects are included using a Cd independent of
damage shape or size. Previous work indicates that this assumption over-simplifies the problem to an extent where the
flooding rates used for calculation are in error. Simple validation of a tool to assess flow rates for emergency response is
included.

1. INTRODUCTION jet and those of a submerged orifice are negligible [7],

1.1 OVERVIEW
The motions of a damaged ship in a seaway are subject to
a number of driving forces, from a ships response to
waves, floodwater motion in response to ship motions
and ingress or egress of floodwater; all of which are
dependent on each other. [10] demonstrated, using 2D
strip theory coupled with a hydraulic model, that a
resonant effect can occur between the oscillating
floodwater column inside the ship, ship motions and
external wave field. In certain cases this resulted in
structural loads that exceeded those expected during the
intact condition. In order to accurately analyse the effect
of damage on a ship in waves, ingress and egress of
floodwater must be taken into account. Therefore one of
the key quantities we are required to calculate is the
flooding rate.
Most state of the art damaged ship models have assumed
the flow through a damaged side shell can be simplified
to that of an orifice flow model. Research into orifice
flow has been conducted in a variety of fields, from
orifice flow meters in pipes, to controlling flow in flood
relief dams. This research began with Evangelista
Torricelli who in 1643 made the first observations of
velocity through an orifice, noting that it is proportional
to the square root of the head producing the flow. Using
this velocity the flow rate through the orifice can be
calculated from the velocity at the orifice and the area of
the orifice. However, in practice the actual flow rate is
less than this and is accounted for by the co-efficient of
discharge. The reason real flow rates and the theoretical
flow rates differ is due to viscosity, turbulence and
effects of the real geometry, which are not accounted for
by theory derived from Bernoullis equation, due to the
nature of its limitations to inviscid and irrotational flow.
Flow features for a homogenous fluid flowing through a
sharp edged orifice are common in most engineering and
environmental cases so long as the Reynolds number is
of at least O(103)[5], based on the diameter of the orifice
and the average velocity through it. The differences in
discharge co-efficient between orifices producing a free
with the exception being the case of a very small
submerged orifice as the kinetic energy of the submerged
jet is dissipated by turbulence in the receiving fluid. [2]
classifies orifices in two types; small orifices in which
the head of the fluid varies by a very small amount across
the orifice geometry and large orifices in which the
variation in head from one end of the orifice to the other
is substantial, and therefore must be taken into account
this could be of particular importance for large side
damage cases and in cases where the damage is at the
level of the free surface leading to cases similar to flow
through a gate, although no study has been performed to
specify at what point this results in a noticeable effect.
Previous research in this area has extensively covered
steady state flow through circular orifices, thickness of
the orifice plate, ratio of orifice diameter to external
geometry and local changes of geometry within the
circular orifice such as surface roughness and radius of
corner curvature; however a damage orifice in the side
shell of a ship is rarely circular. Few investigations have
been conducted for unsteady orifice flows that dominate
the initial ship flooding stages or for the effects of
different orifice shapes. A preliminary CFD study was
conducted in [14] demonstrating the potential effects of
shape and of compartment area to damage area ratio.
This study will attempt to provide a more thorough
understanding of what happens to the discharge co-
efficient for a variety of different shapes using CFD. A
response surface will then be used to interpolate between
the CFD results to attempt to provide a method of rapid
assessment of discharge co-efficient.
1.2 ORIFICE FLOW
Torricelli was the first to observe the correlation between
the flow rate and the square root of the pressure head [1];
later proved by application of the Bernoullis equation to
be in the form of ,
(1)

Where v is velocity, g is gravitational constant and h is
the difference in hydrostatic head heights. Since
Bernoullis equation assumes inviscid and irrotational
flow conditions, to account for viscous effects, the co-
efficient of discharge (Cd) is introduced. Equation 2
shows the form of the equation valid for a hydrostatic
head venting into atmospheric pressure and for cases
where there is a dynamic head the total pressure
difference can be used (p).
(2)
Where g is the gravitational constant, h is the difference
between water levels and is the density of the fluid.[2]
state that two main flow features are observed which
account for the use of Cd. Firstly the velocity of the jet is
less than the theory as there is a loss of energy accounted
for by the co-efficient of velocity (Cv), which when
multiplied by the theoretical velocity gives the actual
velocity; this can be determined experimentally.
Secondly a vena contracta is formed, causing a reduction
of the area through which the fluid flows; this is
accounted for by the co-efficient of contraction (Cc) and
like Cv is a ratio of orifice area to the area of the
contraction. As shown in Equation 3 and 4, co-efficient
of discharge is the product of these two co-efficients of
velocity and contraction.
(3)
(4)
[5] identified that flow rate is also dependent on other
factors which experimentally are accounted for within
Cd. Factors such as fluid viscosity (), surface roughness
at the opening (), thickness of the opening (L) and the
area ratio of the orifice with that of the surrounding
geometry (); therefore q=f(A,p,,,,L,) for some fraction c equation 9. Both liquid and gas phases are
function f. On a dimensionless basis, we can expect this
to reduce to the following,
. (5)
The higher the Reynolds number the smaller the
influences are resulting from viscosity on fluid motion.
For the case of orifice flow, the Reynolds number is
defined in equation 6, where D is a characteristic length
associated with the damage orifice (e.g. Circle diameter)
and Re is based on the flow rate, characteristic length and
area in the following way,
. (6)
The Damaged Ship, London, UK
2. METHODOLOGY
2.1 CFD METHODOLOGY
2.1 (a) Overview
The commercial solver Star CCM+ was used to solve the
compartment flooding cases which uses the Reynolds
Averaged Navier-Stokes equations discretised using the
finite volume method. The finite volume method for
incompressible viscous flows is derived in detail in [3].
A brief description will be shown here. Starting point is
the conservation equations for mass, momentum and
generic scalar quantities, shown here in their integral
form.
(7)
(8)
Where is fluid density, V is the control volume
bounded by closed surface S, v is the velocity vector, t is
time, is the diffusion co-efficient and b is the
volumetric source of scalar quantity , p is the pressure,
n is the unit normal vector to the surface facing
outwards. Turbulence is modelled using a k- type
turbulence model, which modifies the calculation of the
viscous stress tensor with the addition of turbulent eddy
viscosity (t); calculated locally as a function of
turbulence kinetic energy k and dissipation . Further
detail is also available in [3]. Star CCM+ uses the well
known Volume of Fluid method (VoF) in conjunction
with a High Resolution Interface Capture (HRIC)
scheme, details available in the user manual, which is
used to simulate the free surface. This creates an
additional transport equation to solve for the volume
modelled, c is set to 1 for water and 0 for air. This way
these two fluids are treated as a single fluid and its
properties vary according to the local volume fraction, an
example is shown for density in equation 10 where
subscript l is for the liquid phase and g the gas phase.
(9)
(10)
Surface tension at the interface is treated as a body force
that is a function of volume fraction c. This is through
the addition of a continuum surface force (CSF). The
CSF model uses a smoothed field of c to define a vector
normal to the free surface. The expression of surface
tension force (Fst) is shown in equation 11, where is the
surface tension co-efficient and is the curvature of the
free surface interface.
The Damaged Ship, London, UK
(11)
The domain of interest is divided up into a finite number
of control volumes with the use of an unstructured grid.
In the centre of each of these cells there lies a
computational point for which the known quantities are
specified and unknown properties are computed. The
above equations are applied to each computational point
and then discretised, which results in one algebraic
equation for each control volume; importantly these
equations are dependent on both the properties from
previous time steps and the control volumes in close
proximity. The equation is of the form as shown in
equation 12, where nj is the number of cell faces around
cell of interest P0; bp0 is the source terms and
contributions from boundary faces.
(12)
In order to calculate the pressure field and couple it to the
velocity field, a guess and correct procedure SIMPLE is
used and Rhie Chows interpolation is used to overcome
the checkerboard effect on collocated grids. Equation
12 is used to calculate each of the variables, but due to
the non-linear nature of the equations, the linearised
system of equations is solved by iterative methods; also
known as inner iterations. A segregated algorithm is then
used to achieve the solution for each time step and
advance [13].
2.1 (b) Compartment flooding simulation
The case is a 3D problem, intended to be comparable to
any generic side shell damage. The geometry can be
described as a cylinder placed with an end in water and a
damage orifice placed in the submerged flat end. The
cylinder is 0.5m long and 0.2m in external diameter with
a wall thickness of 0.01m. The external domain is 0.5m
in diameter and 0.35m in height. Outside of the
compartment the water level is effectively 0.26m high,
whilst inside the compartment the water is at 0m. The
damage shape, location and size are varied which will be
described under subsequent headings.
Figure 1 - Transient flooding simulation geometry
The domain geometry is as shown in figure 2.
Calculations for this study have been performed on the
University of Southamptons Iridis 3 supercomputer.
This investigation comprises of a grid and time step
dependency study, subsequently 30 cases are simulated
with varying shape and dimensions of width and height
to assess sensitivity to the parameters. Reynolds number
ranges from 9.5x103 to 1.03x105.
No-slip wall boundary conditions are used for the
immersed cylinder, with a static pressure outlet at the
cylinder end opposite to the damage. The external area is
bounded by pressure outlets where hydrostatic pressure is
described. The external portion of the domain is
purposefully chosen to be fully immersed in order to
prevent the need to simulate the free surface external to
the cylinder.
A volume of fluids method is used to simulate the two
phases, air and water. The density of these fluids was set
as water=998.2 kgm-3 and air=1.225 kgm-3. Atmospheric
pressure was set to 101325 Pa and gravity prescribed in
the vertical direction at -9.81 ms-2. Initialisation of the
domain in Star CCM+ is achieved using field functions
which define volume fraction and hydrostatic pressure in
the 3D domain.
Careful attention has been given to the convergence
levels in the first and subsequent early time steps during
the bursting and accelerating flow stages, where the
continuity residual is required to reach 1x10-4 before
progressing to the next time step, typically required just
over 100 inner iterations. The importance of this is
demonstrated in a 2D flooding rate study [13].
Inadequate convergence results in the formation of a
non-symmetric jet. After 1.5s of simulation time steps
typically require only 10-40 inner iterations to reach
continuity residual convergence targets and are adjusted
in order to reduce computational expense, although again
inadequate convergence shows a noticeable change in the
flooding rate.
2.1 (c) Grid and time dependency survey
In order to be confident in the mesh and time step
selection, a survey was performed using 7 different
trimmed hexahedral grid sizes ranging from 63,000 to
1.8 million cells. Six time step and convergence strategy
surveys were performed using the range of meshes, some
over all seven mesh sizes, some covering a few of the
meshes. All were performed on one randomly chosen
case which did not feature in the subsequent calculation
of the response surface.

Figure 2 - Grid and time step dependency study results
Survey 1 were the initial runs for each mesh, using a
constant time step of 0.001s and continuity convergence
limits were set to 10-4 where 300 iterations were allowed
to achieve this. Problems occurred during steady state
flooding as too few inner iterations are performed per
time step. In survey 2 the convergence criterion were the
same, but the time steps were adjusted for each mesh to
give a Courant number of 1 according to the predicted jet
velocity at the orifice, but grid convergence was still not
achieved. Survey 3 the time step was kept to give a
Courant number of 1, but the convergence strategy was
changed where a constant 15 iterations were used for all
time steps. This appeared to improve solution during the
steady state phase, but time to flood was over-estimated
due to incorrect solution at the bursting phase of the
flood. Surveys 4, 5 and 6 adopted a 3 stage strategy to
combine the good solution of the initial bursting stage
from surveys 1 and 2 with the good solution of the steady
state flooding stage in survey 3. The first 1.5s of
simulation use the strategy from surveys 1 and 2, the
following 2s are solved using 20 iterations and 15
iterations thereafter. It should be noted that if inadequate
inner iterations are performed a noticeable change in the
flooding rate is observed. Time step sizes are varied
between surveys 4, 5 and 6 to give a Courant number of
0.5, 0.25 and 0.125 respectively. The graph of results
from this study show that time step independence has
been achieved and it appears that grid independence is
achieved.
As generation of response surfaces for a number of
variables requires large amount of simulation, therefore
computational cost is a factor. Mesh refinement at and
immediately around the orifice has been defined as that
of the grid with 800,000 cells, however the rest of the
domain is sized at that of the grid with 250,000 cells. The
trimmed hexahedral mesh is refined around the damage
opening and cell sizes gradually increase slowly outward
toward the pressure outlet boundaries. The discretised
domains contain ~290,000 control volumes, the largest
cells being 0.05m and refined to 5x10-4m - 1x10-3m
around the damage orifice, values chosen in order to
ensure a minimum of 10 cells across the damage orifice
in any direction. Error for this mesh sizing is predicted to
be less than 5% from mesh size and time step, however it
is recognised that this is not experimentally validated and
The Damaged Ship, London, UK
additional errors due to boundary proximity and others
intrinsic to CFD could be present.
2.2 RESPONSE SURFACE
2.2 (a) Overview
The process of determining the flow rate for a damaged
ship can be difficult due to the number of unknowns
within the process. Currently this is most accurately
achieved through the use of CFD but this has a high
computational cost associated with it. To create tool to
determine the flow rate a response surface method has
been chosen. This method will allow a rapid but accurate
determination of the flow rate for a given damage.
Response surfaces are becoming an increasingly popular
method to allow the use of computationally intensive
modelling techniques within computationally intensive
analysis techniques to reduce overall computational time.
Kriging is one such method that was first used by
geologists to estimate mineral concentrations, but is now
used in a wide variety of different applications including
aerodynamics, structures and multi-objective problems
[12].
Kriging is extensively covered elsewhere, including [6]
and [3], and therefore only a brief summary is given
here. The basis function used within kriging is given in
equation 14.
(13)
(14)
Where y is the predicted response value at untried x, x
are the sample points, are the unknown correlation
parameters used to fit the model, is the constant
underlying global portion of the kriging model, r is the
correlation vector of length n, R is an nn symmetric
matrix with a value of unity along the diagonal and f is
a column vector of length n.
The method for selecting the sample points from which
to create the response surface was that of a Latin
Hypercube, first used in McKay et al. [8], as this allowed
a uniform spread of sample points without the points
becoming regular. The number of points required to
generate the surface is gradually increased to determine
the difference in time and accuracy to give an
understanding of the minimum number of points that will
be required to achieve suitable accuracy of results.
To find the values of the correlation parameters, a
genetic algorithm was used. This genetic algorithm was
the standard algorithm within the programming language
MATLAB. The number of generations used were sixty
with a population size of fifty. The genetic algorithm was
The Damaged Ship, London, UK
not optimised for the specific case, but allowed a
reasonably fast determination of the correlation
parameters required for suitable accuracy.
2.2 (b) Response Surface Convergence
The response surface was tested using a number of sets
of data to ensure that it was converging acceptably before
modelling the CFD data set. Initially 13 cases were used
to create the response surface, plotting the maximum
stress calculated in the panel against the varied
properties. Assessing the accuracy of the surface
between the plotted points showed that the error in the
surface was 15%. By increasing the number of points,
the error was able to be reduced. Table 2 shows how the
error in the surface reduced as the number of points to
create the curves increased. For this assessment and
demonstration of the methodology, the number of points
was not increased beyond 50, showing a 1% error in the
surface and deemed suitable for this study.
Table 1: Surface Accuracy in Relation to Points Used
Number of Points RMSE %
13 15
25 8
50 1 Damaged structures are variable in their characteristics
The length of time that is required to generate the 50
points and to optimise the RSM could be performed in a
time totaling approximately 1hr on a standard desktop
computer. The current method requires human input and
it is hoped that in the future this process of generating the
points will be entirely automated.
2.2 (c) Reliability and sensitivity assessment
Damaged ship structural calculations, especially at an
early stage, have a large amount of variability associated
with them due to an inability to determine the properties
of the hull in question and the damage that has been
produced. Therefore it has been decided that reliability
analysis would be a most suitable choice in this
stochastic environment. This will allow an estimate for
the importance each factor has on the flow rate
incorporating potential variability. The reliability
analysis is performed using a Monte Carlo simulation
developed by Sobey et al.[11], where the structural
modelling data required by the Monte Carlo simulation is
represented by the response surface. The simulation
method has been developed to allow problems that
require a large number of runs to be compared to each
other. Originally developed for solving problems related
to atomic research, the simulation method involves
running the same problem repeatedly with a variation of
inputs. The Monte Carlo method has three main steps:
1. Generate a random distribution;
2. Model Calculations;
3. Determine probability.
A simulation was run for the set of statistical
distributions shown in Table 1, resulting in a given
reliability for that set. The simulation consisted of
1000000 runs, a value for which the sensitivity values
had converged. During the simulation the values of the
input variables must be determined for each run. The
first step was to generate a random number that could
then be mapped, using the quantile function, to the
distribution function that best represents the input, the
values of which are shown in Table 1. The random
numbers were determined using code from Numerical
Recipes [7].
Having determined the input variables for each
simulation it is then possible to determine the outputs, in
this case deflection and stress from the model being used.
These variables can then be compared to that of a limit
state. The general limit state function is given by
equation 15.
(15)
Where R is capacity and Q is demand. In this case the
capacity was determined as the mean flow rate and the
demand as the flow rate for the given input values.
due to different methods through which the damage can
be caused and also due to the lack of knowledge about
the damage itself. Therefore, it is important to gain an
understanding for how much a change in an input will
affect the final output and the implications these will
have on other outputs. For example in the case of the
damage hull it will be possible to determine the effect of
the damaged shape on the inflow rate of the damage
comparatively with the other variables used. This will
provide guidance into which variables are the most
important to model accurately allowing more in depth
investigation into these values. As such, a sensitivity
index has been used to investigate the impact of each
input as shown in equation 16.
(16)
Where N is the number of failures, H(Xi) is sample
performance and S(k)(U: Xi) is the score function.
Equation 16 calculates the gradient of the surface of the
output. This gradient can be found from the score
functions of each distribution shown in equation 17 for
the Normal distribution and equation 18 for the Weibull
distribution.
(17)
(18)
Where is the mean, is the shape parameter and
is scale parameter.

3. RESULTS
3.1 CFD
A table of results is available in Appendix A where it can
be seen that solutions were not calculated for all 30
cases. 5 cases with very small damage sizes resulted in
flooding times too computationally expensive for
simulation time required and were omitted from the
response surface generation.
The dominant flow features of this simulation are that of
the initial jet and the vena contracta. The shape and
strength of the vena contracta are affected by its
proximity to compartment walls, determined by location
and ratio of damage size to compartment size.
As flooding in each case occurred over different time
scales, the 3 stage strategy was not found to be
completely suitable for all cases as it was developed for a
case which converged at 7.5s to flood. For scenarios with
a large orifice area the strategy as applied could result in
errors similar to those observed in survey 2 of the grid-
time step dependency study if a steady state flooding
regime is encountered. For this type of error the
subsequent solution was found to over-predict the time to
flood resulting in a lower discharge co-efficient than the
converged solution. On the other side of the scale, for
small damage orifices which occur over longer periods of
time require longer periods of the initial convergence
strategy phase, which were not carried out in this study
and could contribute to total error source. Inadequate
levels of convergence can be observed in the orifice
velocity results as a change in flow rate at the change of
convergence criteria.
Figure 3 - Example of maximum orifice velocity against
time
Typical behaviour of orifice velocity is shown in fig. 3
where in this particular case, bursting flow can be seen in
the first few 1/10ths of a second where a vertical jet is
formed. The jet then collapses and there begins the
accelerating flow as the orifice velocity fluctuates above
and around the linear line. At this point air is entrained in
the water inside the compartment and disturbs the
formation of the vena contracta. Once the bubbles of air
escape the water, the vena contracta fully forms and the
quasi steady stage is reached, in this case after about 1s.
The Damaged Ship, London, UK
The results show an appreciable difference in the
averaged discharge co-efficient with respect to both
shape and size, the sensitivity of which will be discussed
in the following section.
3.2 SENSITIVITY ANALYSIS
Having produced a response surface from the data in
Appendix A it was possible to run a simulation to
determine the sensitivity of the inflow rate to the length
and breadth of the damage as well as its shape. The
values for the input variables are shown in Table 2. From
these input variables it was therefore possible to analyse
the sensitivity of the factors on the orifice flow, the
results of which are given in fig. 4.
Table 2: Input Variables
Damage Mean Coefficient Distribution
Variable of
Variation
No. Sides 6 0.6 Normal
Width 45mm 0.5 Normal
Height 45mm 0.5 Normal
The response surface created using CFD was then tested
against a CFD set developed using a coarser mesh. This
developed a RMS error of 17%. While this error appears
to be large it can be seen the differences between the two
data sets incorporated errors up to 68% and it is therefore
determined that the response surface was accurately
representing the CFD.
Figure 4 - Grid and time step dependency study results
From fig. 4 it is possible to see that for all of the input
values a one unit change to the input produces a similar
change to the orifice flow. For the height and the width
of the damage area this is the same as expected from the
model. It also shows that for the change in shape this
shows that the addition of an extra side to the regular
shape that this had the same effect as increasing the
width or height by 1mm.
4. CONCLUSIONS
From the grid-time step dependency study we have
shown an estimate of the levels of modelling fidelity
The Damaged Ship, London, UK
needed to accurately simulate transient flooding through
an orifice using a RANS based method. Further
confidence will be sought in an external boundary
condition proximity survey, turbulence model survey and
comparison to experiment.
CFD has been used to solve a number of cases of varying
shape, damage width and damage height, where
distribution of these values was achieved using a Latin
Hypercube. The results displayed the expected behaviour
and indicate a dependency on the variables chosen. Error
sources are discussed.
From the results that are shown within this paper it can
be seen that a sensitivity value can be determined using a
response surface method and CFD combined with a
Monte Carlo simulation. This allows an analysis of how
the different input factors affect the flow rate. Further to
its use in understanding how these factors might affect
the flow rate it will allow an understanding for the
probability that a compartment will flood in a given time
period allowing emergency response teams to have a
better understanding for this factor. While this process
takes a large period of time to produce a response
surface, once this initial period is finished the final
analysis during the emergency response will be rapid.
This paper therefore shows that further work should be
performed to take into account a larger number of
variables to determine which factors affect the final flow
rate during a damage scenario. This will allow a
reduction of the number of variables for use in a tool
capable of analysing damaged ship hulls.
5. ACKNOWLEDGEMENTS
The authors would like to thank the UK Ministry of
Defence and Lloyds Register EMEA through the centre
of excellence in marine structures for their continued
support and without whom this research would not have
been possible. The authors would also like to thank
Lloyds Register Ship Emergency Response Service for
assisting in developing their knowledge and
understanding of how assistance may be provided to a
stricken vessel.
6. REFERENCES
1. BATCHELOR, G., An Introduction to Fluid
Dynamics, Cambridge, 1967.
2. DOUGLAS, J., GASIOREK, J., SWAFFIELD,
J. & JACK, L., Fluid Mechanics (5th edition), Pearson
Education Ltd, 2005.
3. FERZIGER, J.H. & PERIC, M., Computational
Methods for Fluid Dynamics, Springer-Verlag, 2003.
4. FORRESTER, A, SOBESTER, A. & KEANE,
A., Engineering design via surrogate modelling: a
practical guide, Wiley, 2008.
5. HOLFORD, J. & HUNT, G., The Dependence
of the Discharge Coefficient on Density Contrast
Experimental Measurements, Proceedings of the 14th
Australasian Fluid Mechanics Conference, 2001.
6. JONES, D.R., A taxonomy of global
optimization methods based on response surfaces,
Journal of Global Optimization, 2001.
7. MASSEY, B. & WARD-SMITH, J.,
Mechanics of Fluids (8th Edition), Taylor and Francis,
2006.
8. MCKAY, M.D., BECKMAN, R.J. &
CONOVER, W.J., Comparison of three methods for
selecting values of input variables in the analysis of
output from a computer code, Technometrics, 21(2),
239-245, 1979.
9. PRESS, W.H., Numerical Recipes,
Cambridge, 1986.
10. SMITH, T.W.P., DRAKE, K.R. & RUSLING,
S., Investigation of the variation of loads experienced by
a damaged ship in waves, Advancements in Marine
Structures, 2007.
11. SOBEY, A.J., BLAKE, J.I.R. & SHENOI, R.A.,
Design for production in FRP boats, International
Conference on Composite Materials, 2009.
12. TOAL, D.J.J., BRESSLOFF, N.W. & KEANE,
A.J., Kriging hyperparameter tuning strategies, AIAA
Journal, 46, 1240-1252, 2008.
13. WOOD, C.D., HUDSON, D.A. & TAN, M.,
CFD Simulation of Orifice Flow for the Flooding of
Damaged Ships, Proceedings of the 13th Numerical
Towing Tank Symposium, 2010.
14. WOOD, C.D., HUDSON, D.A. & TAN, M.,
Numerical Simulation of Compartment Flooding for
Damaged Ships, Proceedings of the 12th Numerical
Towing Tank Symposium, 2009.
7. AUTHORS BIOGRAPHIES
Christian Wood is a research engineer currently
studying for an Engineering Doctorate (EngD) at the
University of Southampton. His research activities
include simulation of waves, ship motions, compartment
flooding and violent free surface motions. His previous
experience includes a MEng in aerospace engineering.
 The Damaged Ship, London, UK

Adam Sobey holds the current position of Research

Fellow at the University of Southampton. His research is
aimed at modelling of damaged ship structures. His
previous experience includes a MEng in aerospace
engineering and a PhD in concurrent engineering.

Dominic Hudson is a senior lecturer in Ship Science at

the University of Southampton and primary academic
supervisor of Christians research. His research interests
include all aspects of Ship hydrodynamics, in particular
the prediction of vessel motions in waves.

Mingyi Tan is a lecturer on the Ship Science at the

University of Southampton and secondary supervisor of
Christians research. His research interests include
applications of computational fluid dynamics and
numerical methods. His previous experience includes
supervision of PhDs and numerous published papers.

Paul James currently holds the position of Programme

Manager at Lloyds Register located in Bristol and is an
industrial supervisor for Christian Woods doctoral
research. Past experience includes the position of Head
of First Entry and Transfer of Class for naval vessels at
Lloyds Register and a BEng from the University of
Southampton.
The Damaged Ship, London, UK

APPENDIX A TABLE OF CFD RESULTS

Case Shape (n) Width (mm) Height (mm) Area (m2) Time to flood (s) Avg. flood rate (m3s-1) Avg. Cd
1 8 2 89 6.81E-05 time est.>90s - -
2 4 35 83 2.05E-03 2.38 0.002673004 0.588
3 7 74 80 2.57E-03 1.97 0.003229315 0.568
4 10 59 23 4.19E-04 13.48 0.00047194 0.508
5 6 56 17 4.76E-04 12.34 0.000515539 0.489
6 3 29 44 1.11E-03 4.96 0.001282611 0.524
7 4 17 56 6.73E-04 8.25 0.000771121 0.517
8 9 86 26 7.65E-04 7.15 0.000889755 0.525
9 6 80 53 2.12E-03 2.25 0.002827444 0.602
10 10 23 59 4.19E-04 13.66 0.000465721 0.501
11 9 68 68 1.58E-03 3.54 0.001797105 0.513
12 5 83 29 1.41E-03 3.62 0.00175739 0.561
13 5 44 50 1.29E-03 4.25 0.001496882 0.523
14 6 14 20 1.40E-04 time est.>40s - -
15 5 47 2 5.53E-05 time est. >100s - -
16 3 38 11 3.62E-04 15.77 0.000403408 0.503
17 6 26 77 1.00E-03 5.64 0.00112797 0.509
18 8 89 65 2.21E-03 2.21 0.00287862 0.587
19 5 71 47 1.96E-03 2.74 0.002321807 0.534
20 9 41 32 4.49E-04 13.1 0.00048563 0.489
21 3 65 14 7.88E-04 7.26 0.000876274 0.502
22 4 77 62 3.38E-03 1.41 0.004511879 0.603
23 8 50 74 1.42E-03 4.02 0.001582525 0.505
24 9 5 8 1.37E-05 time est. >400s - -
25 2 53 38 2.01E-03 2.75 0.002313364 0.519
26 7 32 86 1.19E-03 4.52 0.001407467 0.532
27 7 20 41 3.56E-04 16.91 0.000376212 0.477
28 7 62 35 9.42E-04 6.44 0.000987849 0.474
29 2 8 5 4.00E-05 time est. >100s - -
30 3 11 71 6.76E-04 8.14 0.000781542 0.522
The Damaged Ship, London, UK

DEALING WITH THE HINGED SHIP - A REVIEW OF THE ANALYTICAL AND

PRACTICAL ISSUES BASED UPON THE M.T. ELLI SALVAGE
K Ellam, Swire Salvage (Pte) Ltd, Singapore,
C Moore, Herbert Engineering Corp., USA
R Tagg, Herbert Engineering Shanghai, China

SUMMARY

In August 2009, south of Suez, the MT Elli broke her back in calm water while being ballasted in preparation for dry-
docking. The hull girder hinged near amidships and the deformation, about 22m, was sufficient for the bottom of the hull
to clear the water. This paper outlines the subsequent salvage and some of the analytical issues in salvaging such a
hinged ship including; modelling the geometry, estimating the moment distribution and how it got there, problems with
predicting hinge behaviour, estimating residual strength. The practical issues in straightening the ship and subsequent
separation into two parts in preparation for removal are discussed.

1. INTRODUCTION
The MT Elli was an Aframax crude oil carrier with
double sides and single bottom, built in Japan for
Australian owners in 1984. She has 7 center cargo tanks,
2 slop tanks, and 6 pairs of segregated side ballast tanks.
All bunker fuel was carried in the engine room area.
The ship was owned by an international shipping
company and had three months earlier run aground off
Jeddah whilst on passage in the Red Sea, the details of
this incident are not entirely clear but it was reported that
she had grounded on a sandy bottom along 2/3 of her
midship length.
The vessel was subsequently refloated and had been
towed to the Port of Suez where tank cleaning and
preparations for dry docking the vessel were been carried
out when the incident discussed in this paper occurred.
On the morning of the 28 August 2009 the Ships
Officers were checking the vessels drafts from a
workboat when the vessel broke her back, she was
reported to have been in a ballast condition at the time.
The crew abandoned the vessel shortly thereafter for fear
of her sinking.
2. INITIAL RESPONSE
Five Oceans Salvage Consultants (FOS) from Athens
were appointed by the owners and a Lloyd's Open Form
Salvage Agreement (LOF) with SCOPIC invoked, was
agreed on the 28th August 2009.
The Salvage Tug Arabian Sea Fos was mobilized from
Greece and sailed for Suez once salvage and pollution
control equipment had been placed on board.
3. SALVAGE OPERATION
3.1 THE FIRST DAYS
Within the Suez Canal limits all salvage operations fall
under the jurisdiction of the Suez Canal Authority (SCA)
and initially the Salvors were not permitted to access the
casualty for 24 hours as it was deemed too dangerous.
On Sunday 30 August Salvors were only allowed on
board to assist with the cleanup of residue oil and sludge
that remained on the deck as a result of the tank cleaning
operations. Meanwhile the Suez Canal Authority
Salvage team continued with plans to remove some
ballast from the casualty as well as preparations to move
the vessel from the anchorage to a suitable location
where she could be safely grounded.
During this time Salvors continued with pollution control
measures as well as carrying out some preliminary
inspections of the vessel and the engine room space
where some ingress that had been indentified was
successfully stopped, all overboard valves and sea chests
were also closed as a further precaution against flooding.
It was however found that the pump room was also
flooded.

A team of 15 men led by Senior Salvage Master

Nikolaos Pappas, including Salvage Masters, Salvage
Officers and Engineers, Anti Pollution Personnel as well
as Salvage Divers, was mobilized from Greece.

Swire Salvage, a subsidiary of Swire Pacific Offshore

which has a joint venture partnership with FOS was also
requested to assist and a team mobilized from Singapore. Figure 1: Aft freeboard and threat of further flooding

The casualty was finally moved on the 1st of September
by two SCA tugs and grounded at the bow in 18 meters
of water on a gently sloping sandy shore, at this stage the
SCA allowed the Salvors to continue with the Salvage
Operation.
3.2 STAGE 1 REMOVAL OF POLLUTANTS
The Salvors were now able to proceed with their salvage
plans and the first priority was the removal of the fuel
and lube oils remaining on board as well as the removal
of other marine pollutants, paints and foodstuffs. By this
stage the salvage tug Arabian Sea FOS had arrived
from Piraeus bringing with it a large array of Salvage
equipment which included pumps and additional oil
pollution equipment.
Salvors were now also able to fully investigate the
condition of all tanks on board, both cargo and Fuel oil,
and were able to obtain a better understanding of the
quantity of hydrocarbons that remained on board at this
time and to further develop a plan for their removal.
The decision was also made by Salvors that all sea chests
and overboard valves should be blanked off externally to
prevent any further ingress of water into the casualty, this
was carried out by the salvage divers who welded steel
plates externally over all underwater hull penetrations.
At the same time as this was ongoing, efforts were being
made to not only understand the distribution of weight
within the vessel but also to find and provide the required
data to the HECSALV [1] team who were carrying out
the modelling of the vessel offsite.
One of the difficulties faced was that the effective trim of
the vessel was essentially equivalent to 42 meters (21m
trim over half the ship length) which precluded the use of
normal sounding tables and required Salvors to manually
calculate the volume/weight of the wedge of water/ oil
remaining in the various tanks.
After eight days all the fuel, lube oil and any slops
remaining onboard the casualty had been transferred to
local barges for disposal ashore. In total 7280 m3 had
been removed from the vessel and this figure included all
the dirty water from the pump room.
3.3 STAGE 2 PREPARATIONS
With the removal of this weight from the aft end of the
casualty the Salvors had managed to regain
approximately 3meters of freeboard at the transom, the
safety of the vessel had been secured and there was now
no further danger of the stern section flooding. This
increase in buoyancy aft also resulted in a change in the
trim of both the forward and aft sections; a slight change
(opening) in the hinge was also noted at this time.
The Damaged Ship, London, UK
Salvors could now concentrate their efforts on the work
that was required to separate the two sections, numerous
options had been considered but at the end of the day it
was decided that the best option was still to separate the
two halves so that they could be towed independently to
a suitable scrapyard.
Only now did the Salvors consider it safe to commence
the detailed inspection of the damaged tanks in the
midship section. Only after extensive ventilation to
ensure that all tanks were gas free could the Salvors
commence these tank inspections and start to obtain an
understanding of the extent of the damage in the affected
tanks.
It was found that in addition to the structural damage to
the longitudinal bulkheads and shell plating in the
immediate vicinity of the collapse there was also
significant damage to other transverse bulkheads
throughout the length of the vessel. On further
investigation it was found that where ballast and cargo
lines did not collapse significant forces had been
transferred along these pipelines and throughout the
length of the vessel which caused some watertight
bulkhead penetrations to fail. It was also noted with
interest that less damage had occurred along the cargo
lines as these had been fitted with expanding bellow
connections as opposed the less flexible connecting
collars that had been used on the ballast lines.
With this information extensive modelling was now
carried out using HECSALV for the hinged ship as well
as the two half ship models to determine what actions
and repairs would be required. This was not only to
allow the vessel to be straightened but also to ensure that
once separated the half sections retained sufficient
reserve buoyancy during the tow, even in the event that
the watertight integrity of some tanks was lost.
Figure 2: Initial HECSALV model
Repairs involved the cropping and removal of damaged
pipelines as well as blanking off of all pipes to avoid
further communication and cross flooding between tanks.
Where bulkheads and the watertight integrity of tanks
had been compromised doubling plates were fitted and in
some cases large sections of the bulkhead had to be
replaced.
The Damaged Ship, London, UK
3.4 STAGE 3 BALLASTING
On completion of the repairs Salvors could now
concentrate on globally straightening the casualty prior to
separating the two halves. A detailed ballasting plan was
compiled which involved not only the ballasting of the
midship section but also the simultaneous de-ballasting
of the forward most tanks. This was required as it was
planned to separate the vessel in approximately 10
meters of water, to achieve this a tug was connected to
the bow and during the entire ballasting operation tension
was kept on the tow wire to ensure that the casualty was
gradually towed into shallower water as the deflection of
the hull and the forward draft decreased.
This entire ballasting operation was carried out over a
period of 48 hours using submersible pumps and resulted
in the straightening of the casualty.
It was then decided to place the ship into a slight sagged
deformation by de-ballasting the forward section
(Forepeak and #1 Ballast Tanks) completely in an
attempt to ensure that any remaining intact longitudinal
girders and shell plating at the bottom of the vessel were
separated. This was achieved and at high water a 2
degree upward deflection of the fore section was
recorded, a dive survey at the time confirmed that this
plan had been a success.
Prior to cutting it was required that both the fore and aft
sections had to have sufficient grounding force through
all ranges of tide to ensure that there was minimal
sheering force at the hinge during this process, again the
HECSALV models proved an invaluable tool in
calculating the ballast distribution that was required to
ensure that this condition was met, see Section 4.5.
3.5 STAGE 4 SEPARATION
Once it had been ascertained that the two halves were
stable, Salvors were able to concentrate their efforts on
preparing the casualty for separation. Prior to this all the
main deck cargo pipe lines in the vicinity of the hinge
had to be removed and the remaining ends blanked to
minimize any threat of pollution. Once this had been
completed, the Enkaz 1, a large sheer leg with a 1200
tonne capacity was positioned alongside the casualty, it
was planned to use a large diameter wire which would be
constantly drawn though the casualty to cut the
remaining deck longitudinals and deck plating that was
still holding the two halves together. Unfortunately due
to damaging to the wires that were available for this
operation the final remaining intact sections of the deck
had to be manually cut using oxyacetylene cutters. The
Elli was finally separated on the 19 October 2009, 8
weeks after the initial incident.
Figure 3: Aft section after separation
3.6 STAGE 5 PARTING AND REFLOATING
Once the two sections had been physically separated all
that remained to do was to refloat each section
individually and prepare it for towage.
The HECSALV models had been used to predetermine
the distribution of ballast required in each section to
ensure that a suitable trim and draft was obtained for the
planned towage to the breakers yard. Prior to de-
ballasting a Salvage Tug was connected to the towing
bridle at the bow, this was done in order to be prepared to
tow the forward section into deeper water as soon as
sufficient buoyancy had been regained.
It was planned to refloat the forward section first and de-
ballasting was thus planned to coincide with the next
rising tide. This operation was completed without any
difficulties and the forward section was successfully
refloated and preparation for towing were completed as
required.
The aft section proved to be more complicated as before
refloating could be attempted there were still repairs and
blanking that had to be performed to the #5/6 cargo tank
bulkhead where the cargo pipelines had caused
significant damage to the watertight bulkhead
penetrations. It had been decided to wait and to perform
these repairs underwater to reduce the risks associated
with welding in the cargo oil tanks prior to the initial
ballasting and straightening of the Elli. This did add a
degree of complexity to the operation however the dive
team worked tirelessly and completed this task within
two days of the separation. Following this the
deballasting operation commenced and the aft section
was successfully refloated the following morning on the
rising tide as had been predicted. Final preparations then
continued on board to prepare the aft section for her final
voyage under tow to the breakers yard.
However the Elli was not quite ready for this and on
inspection the following morning the aft section had
assumed a 3 meter trim by the head (midships), this was
certainly not as had been planned. Soundings of all tanks
were then taken and again the HECSALV model of the
aft section was used to determine the possible cause of
the current and unexpected trim, it was finally
determined that the #6 COT must still be tidal and the
salvage divers returned to the water to carry out another
inspection of the bulkhead penetrations. It was found that

a small scallop opening had been overlooked and not
properly sealed. However once this had been rectified the
watertight integrity of the tank was verified over the next
24hours, the ballasting program was able to resume and
the stern section finally prepared for towage.
3.7 FINAL VOYAGE
The Salvage operation was finally completed in just over
9 weeks from the initial incident and once the final
clearance was obtained from the Egyptian authorities the
two sections began a month long tow which took them
south through the Gulf of Suez and along the southern
coasts of Yemen and Oman and through the Arabian Sea
towards the breakers yards in India.
4. POST MORTEM ANALYSIS
The buckling of the MT Elli was no doubt surprising to
those onboard, but is it so in retrospect? In this section
the ultimate moment capacity of the hull girder is
assessed and some lessons learned discussed. Once the
hull girder collapsed, key engineering concerns included
predicting the relaxation of the buckle and the forces and
moments carried through the hinge as the salvage
proceeded. The capabilities and limitations of existing
salvage tools are evaluated.
4.1 MIDSHIP SECTION PROPERTIES
The MT Elli is a double-sided, single tank across, oil
carrier, built in 1984. The hull is constructed primarily
of AH-32 steel with mild steel used for the middle
portions of the side shell and longitudinal bulkheads.
Frame spacing was 4.580 m. The bottom plating is 19
mm near centreline, reducing to 17 mm outboard of the
flat of bottom. Deck plating inboard is also 17 mm. The
deck structure is strengthened in the centre cargo tanks to
account for sloshing loads stepping up to 20 mm plate
outboard of the longitudinal deck girder and increasing to
23 mm in way of the longitudinal bulkhead. Transverses
were added midway between frames in way of the deck
reinforcement and bilge radius. As a consequence of the
deck stiffening the section is biased in strength towards
the deck, with a section modulus of 33.3 m3 at the deck
edge and 29.8 m3 at the keel. The corresponding yield
moment of the section is 9.6 x 105 m-MT. An analytical
model of the midship section is shown in Figure 4.
Figure 4: Midship section from HECSALV
The Damaged Ship, London, UK
4.2 COLLAPSE STRENGTH OF THE HULL
GIRDER
Longitudinally framed ship structures of the generation
of the MT Elli have been conventionally designed using
an allowable stress approach that includes margins to
account for stress concentrations, future corrosion and
buckling. In practice this means that the maximum field
stress in the hull girder under the combination of
maximum still water bending and design wave bending
moments is limited to 60% of the yield stress. At the
time of construction, the typical classification society
minimum buckling strength requirement was 10%
greater than the maximum load expected. For these
types of structures buckling failure under increasing load
is expected to occur in the order of:
plate buckling,
frame tripping,
column collapse between transverse frames, and
gross panel collapse involving both longitudinal and
transverse stiffeners
An ultimate strength analysis of the section using the
program ULTSTR [2], which considers the first three
modes above, shows results consistent with the
prevailing design practice at the time of construction.
Figure 5 shows plots of hogging moment versus
curvature of the hull girder under a series of assumptions.
In ULTSTR one has several options for the post collapse
of structural members; two are shown in the figure, one
where the members are removed as they buckle and a
second where the partial residual strength model is used
for the collapsed members. In ULTSTR panel buckling
is not explicitly considered but is included in evaluating
the tripping and column collapse modes. For this section
the assumption on post-failure residual strength has a
small impact on the ultimate moment capacity.
The collapse mode is predicted to be initiated by tripping
of the bottom stiffeners followed by column collapse
between frames of the centre vertical keel and the two
longitudinal bottom girders. The collapse moment for
the intact section ranges from 7.13 x 105 m-MT (dashed
line) to 7.7 x 105 m-MT (uppermost solid line) depending
upon the length between tripping support assumed. The
higher value corresponds to including tripping support
halfway between transverses although this is not
specifically shown in the midship section. Additional
curves are shown for 5% and 10% corrosion, typical for a
ship of this vintage, applied to all structural members,
with collapse moments of 7.2 x 105 m-MT and 6.7 x 105
m-MT respectively. This ultimate moment capability
could also have been further reduced due to plate
deformation caused by the previous grounding incident.
The ultimate moment capacity of the section in hog is
about 73% of the yield moment capacity. This is
consistent with design practice and with other ships of
this vintage. For comparison, the At-Sea and In-Harbour
The Damaged Ship, London, UK

allowable bending moments taken from the onboard
loading computer software were 2.15 x 105 m-MT (22%
of yield moment, or 30% of ultimate moment) and 3.81 x
105 m-MT (40% of yield moment or 53% of ultimate
moment).
The estimated bending moment in still water prior to the
casualty is in the middle of the range of predicted
ultimate moment capacity. Collapse of the hull girder is
therefore not surprising in retrospect with still water
bending moments apparently nearly double the harbor
allowables.

4.3 (b) Post Collapse Bending Moment

The bending moment carried by a ship represents a

Figure 5: Ultimate moment capacity of MT Elli
4.3 MOMENTS CARRIED BY THE HULL
GIRDER
Is it hogging or sagging? While the deformation is
clearly in the hog direction the post collapse ship was
actually in a sagging moment condition. The
combination of the dynamic nature of the collapse and
the midship flooding during the buckle caused the ship
moment to completely reverse to a sagging condition.
Essentially the ship bent beyond what was required to
neutralize the initial bending moment, unlike what would
have happen if the collapse happened slowly and quasi-
statically.
certain amount of potential energy stored in the hull
girder. Once the restraint of the girder is released due to
buckling, a hinge is introduced in the effective beam and
the hull displaces, upwards due to the excess of
buoyancy over weight amidships, gaining momentum as
it goes. Eventually the direction of the forces reverses as
more of the ship comes out of the water and the motion
slows, and may reverse. The buckling of the ship stops
when the initial potential energy is dissipated in
deforming the ship structure, making waves and
deflecting the hull girder in the opposite direction. The
momentum gained by the ship is sufficient to deflect the
ship past the point where the hinge carries no moment.
As the buckled ship structure can still carry some
moment the eventual equilibrium condition has a sagging
moment.
In order to prevent sinking the MT Elli was intentionally
grounded as described earlier. In that condition the
moment distribution is as shown in Figure 7. This
indicates the hinge can support significant shear and
bending moment. Subsequent attempts to straighten the
ship required the introduction of even larger sag
moments.
Shear +Hog
10 3

x
8
x
2
4.3 (a) Bending Moment Prior To Collapse 6

4
1

The estimated loading of the MT Elli prior to collapse 2

resulted in the shear force and bending moment 0

F47 F49 F52 F54 F57 F59
Mx F62 F64 F67 F69 F72 F74 F77 F79 F85 G0A
G0F
0

distribution shown in Figure 6. In the figure the dashed -2

lines are the allowable At-Sea shear forces and bending -4

-1

moments. The estimated ultimate bending moment is -6

7.16 x 105 m-MT (75% yield moment capacity). This is

-2
x
-8 x

the horizontal dashed line in Figure 5. -10

x=interpolated -Sag
-3

Shear (MT)/10^3 AT SEA Shear Allowable (MT)/10^3

Moment (m-MT)/10^5 AT SEA Moment Allowable (m-MT) / 10 ^ 5

Figure 7: Bending moment after collapse

4.4 PREDICITING HINGE BEHAVIOUR

The prediction of the structural behaviour of damaged

Figure 6: Estimated shear force and bending moment
prior to collapse
ships is still largely a research and development process
that is actively being pursued by universities, navies and
classification societies. Analysis of the behaviour of the
accordioned structure would require a large
displacement, nonlinear finite element analysis. For the
simplified MT Elli salvage analysis it was not possible to
predict the stiffness of the hinge and how it would open
 The Damaged Ship, London, UK

up as the sagging moment was increased. In this salvage

Shear +Hog
10 3

the hinge proved to be able to carry considerable 8

x
x
2
moment, in fact more than the sea-going allowable 6

bending moment. 4
1

While it proved impossible to predict the behaviour of 0

F47 F49
Mx F52 F54 F57 F59 F62 F64 F67 F69 F72 F74 F77 F79 F85
0

the hinge, useful information could be gathered by -2

evaluating the vessel with reducing amounts of hog -4

-1

deflection. In Figure 8 the sagging moment versus hog -6

-2
deflection is plotted. As the deflection is reduced, so -8
x
x

does the sag moment. This analysis, performed in the -10

x=interpolated -Sag
-3

field, indicated that the structure would not fail Shear (MT)/10^3 AT SEA Shear Allowable (MT)/10^3

dramatically as the deflection was reduced, as the Moment (m-MT)/10^5 AT SEA Moment Allowable (m-MT) / 10 ^ 5

additional buoyancy amidships compensated for the Figure 9: Shear and bending of straightened ship
loading provided by filling tanks on either side of the
buckle while straightening the ship.

Figure 10: Aft half aground pivoting at forward edge

Figure 8: Reduction in sag bending moment with

4
reducing buckle deflection
3

4.5 ESTIMATING FORCES IN THE HINGE

PRIOR TO SEPARATION 2

1
As noted in the discussion of Stage 3 of the salvage it
was critical to prevent relative movement of the two 17:1500 17:2100 18:0300 18:0900 18:1500 18:2100
Time
halves at the moment they separated. The shear force
and bending moment distributions of the straightened Tide Height (m) Ground Reaction [D] (MT)/10^3
Fi
ship are shown in Figure 9 which shows almost no shear
gure 11: Time history of ground reaction - forward half
or bending moment amidships. This indicates only small
loads are transferred through the hinge and thus the two
halves should float apart independently in approximately
5. CONCLUSIONS
the same positions. However as the separation was done
while grounded analyses of the two halves separated
The salvage and ultimate removal of a large vessel that
were performed to confirm they remained grounded
has broken its back is a challenging process, both in
throughout the separation. An illustration of the aft half
practical and analytical terms. As in many salvage
is shown in Figure 10 and a time history of the ground
events the actual loading of the vessel prior to the
reaction for the forward half over a series of tide cycles
incident was somewhat uncertain. The collapse of the
in Figure 11.
hull girder under the estimated loading is not surprising
in retrospect.

The designed buckling capacity of ships of longitudinally

framed construction may be well below nominal yield
moment capacity. Once the vessel broke its back it
remained stubbornly resistant to straightening and
eventual separation into two parts. Residual strength of
the buckled structure was enhanced by the contributions
of piping and other systems that had sufficient flexibility
not to rupture during the buckle.
The Damaged Ship, London, UK

Engineering tools provide the capability to assess the
loads and stability of the ship as deflected, and ultimately
as separated. This permitted the safe salvage of the
vessel. However the ability to predict the residual load
carrying capacity of the hinge in the hull is currently
more of a research and development level analysis.
Figure 12 shows the MT Elli aft half departing for
eventual scrapping concluding a successful
salvage/wreck removal.
Swire Salvage is based in Singapore and is a subsidiary
of Swire Pacific Offshore which has offices in over 20
countries and a fleet of over 75 modern AHTS, supply
vessels and salvage support vessels.
Herbert Engineering Corp. in Alameda California, and
its subsidiary companies Herbert Software Solutions, Inc
and Herbert Engineering Shanghai are naval architects,
marine engineers, and marine software developers. The
HECSALV suite of casualty analysis software, used
worldwide by Navies, Coast Guards, Class Societies,
salvors, shipyards, and ship operators, was extensively
used in this salvage analysis.

Capt. Ken Ellam holds the position of Salvage Manager

with Swire Salvage and is based in Singapore. He has
being involved in the marine industry for over 20 years
of which 16 were spent at sea. His career started in
Safmarine in 1984 and has since worked for Smit South
Africa as well as Swire Pacific Ship Management where
he sailed as Master on AHTS and IBSVs prior to taking
Figure 12: The aft half on its way with the forward aft in
up a permanent position ashore in 2008.
the background
Colin Moore is Manager of Advanced Analysis and
Salvage Engineering at Herbert Engineering Corp. He
6. ACKNOWLEDGEMENTS
has twenty-five years of experience in general naval
architecture and marine systems research and
We would like to thank the management of Five Oceans
development. He participated in the development of the
Salvage Consultants for allowing us to make use of both
salvage engineering software HECSALV/POSSE. He
data as well as photographs from the Elli salvage
provides salvage engineering and software training to
operation.
clients worldwide.

Robert Tagg is Vice President of Herbert Engineering

7. REFERENCES
and Manager of Herbert Engineering Shanghai. He has
been actively involved in the design and acquisition of
1. HERBERT ENGINEERING CORP., HECSALV,
commercial ships for the over thirty years. As the
1990-2010.
previous President of Herbert Software Solutions Inc. he
has been involved in marine software development,
2. ADAMACHAK, J.C., ULTSTR (1996): The Revised
testing, and training. He is actively involved in ship
Program for Estimating the collapse of Ship Hulls or
stability and damage stability research, and in regulatory
Hull components Under Longitudinal Bending or Axial
development through his participation on the USA
Compression, Survivability, Structures and Materials
delegation at the IMO.
Directorate Research and Development Report,
NSWCCD, 1996.

8. COMPANY & AUTHOR INFORMATION

Swire Salvage and Five Oceans Salvage are both

professional marine salvage operators. Together they
have established a joint venture entity, Swire Ocean
Salvage (Pte) Ltd, This alliance combines the strengths
of both companies in order to provide marine salvage and
wreck removal services to a greater segment of the
maritime community and especially to ship-owners and
insurance interests across the world.

Five Oceans Salvage Consultants main office is in

Athens and also maintains dedicated salvage vessels in
numerous strategic locations around the globe.
The Damaged Ship, London, UK

THE RAPID RELIABILITY AND RISK ASSESSMENT OF DAMAGED SHIP

STRUCTURE

M Shahid, Binary Systems and Engineering (UK) Ltd, UK

SUMMARY

The rapid assessment of residual structural strength of a damaged ship subsequent to major accident such as collision,
grounding etc. is essential for fast response to minimise damage because of consequence in loss of strength. Evaluation
of strength of complex structures such as that of a ship is a hard and time consuming task. Presently used computer aided
structural analysis tools such as Finite Element Analysis and related technologies are neither developed for Emergency
Response Services nor suitable for such application needing quick assessment of structural strength. A computer aided
structural analysis tool based on artificial neural network and structural response parameters can provide rapid strength
assessment and shall be of immense help for Emergency Response Services to make reliability and risk assessment of
residual structural strength and to formulate informed and reliable emergency response actions. This paper is aimed to
presents such techniques for reliability and risk assessment of damaged ship structures.

NOMENCLATURE Tage Parameter, Aging Degradation

P Pressure (N m-2 )
ANN Artificial Neural Network Pf Probability of Failure
C0, Ci, Cij Coefficient of Response Model
X, Xi, Xj Random Variables
D Structural Damage Function x xi System parameters
Dd Depth of Damaged Structural (m), y(x) Response Function
Dh Height of Damaged Structural (m)
Dl Longitudinal Location of Damage (m) Y Estimate of System Response
Dv Vertical Location of Damage (m) wi Neuron Input Weight Parameter
Dw Width of Damaged Structural (m) Reliability Index
E Modulus of Elasticity (N m-2 ) o Flow Stress (N m-2 )
Ed Energy of Structural Deformation (MJ) y Yield Strength of Material (N m-2 )
ERS Emergency Response Services Modelling Error
f(X) Joint Probability Distribution ( ) Neuron Activation Function
Fs Shear Force (N) -1( ) Inverse Standard Normal Distribution
Fsu Ultimate Shear Force (N)
FE Finite Element
FEA Finite Element Analysis 1. INTRODUCTION
FEM Finite Element Methods
Kd Combination Factor for Dynamic The rapid reliability and risk assessment of damaged ship
Bending Moment structural strength is probably the most important
Kw Combination Factor for Wave Induced element of Emergency Response Services (ERS) to
Bending Moment minimise risk of total failure of hull structure and lost of
L Structural Damage Location Function ship as a consequence. At an early stage of incident,
M Moment (Nm) typical information available are those of speeds of ships
MJ Mega Joules involved and prevailing environmental conditions. And
Mh Horizontal Moment (Nm) there, if any, may only be subjective information
Mv Vertical Moment (Nm) available on structural damage. Therefore, in context of
Msw Still Water Bending Moment (Nm) Emergency Response Services regarding structural
Mt Torsional Moment (Nm) strength of ships in incident, there are two important
Mu Ultimate Moment (Nm) aspects of assessment of a ship structural damage, firstly,
Muh Ultimate Horizontal Moment (Nm) ability to predict extent of structural damage that is likely
Muv Ultimate Vertical Moment (Nm) to occur and, secondly, the ability to make assessment of
M*h Normalized Horizontal Moment residual strength of ship structure subsequent to damage
M*V Normalized Vertical Moment incident.
M*t Normalized Torsional Moment
Mw Wave Induced Bending Moment (Nm)
Md Dynamic Bending Moment (Nm) 2. PREDICTION OF SHIP STRUCTURAL
RS Response Surface DAMAGE
RT Volume of Deformed Structure (m3)
T Thickness (m) The extent of damage to ship structure upon a grounding

or collision incident is estimated using basic technique of
energy balance originally proposed by Minorsky in
1959[1]. The kinetic energy of the grounding or collision
process is balanced against the energy needed for
structural deformation and other accompanying
phenomena such as friction, wave shedding, elastic
bounce back of structure etc. The methods for prediction
of structural damage may mainly be classed on the basis
of how the energy absorbed by structural deformation is
evaluated as follows:
The empirical methods,
The simplified analytical methods,
The simple finite element methods (FEM), and
The explicit dynamic FEM.
The empirical methods are based on empirical correlation
of the extent of structural deformation to the energy
absorbed in the structural deformation process. Total
energy for structural deformation is expressed in term of
volume of structure deformed in collision. Minorsky [6]
analysed 26 collision cases of full-scale ship accidents
and developed the following formula (1):
E d = 47.2 RT + 32.7 (1)
where Ed is the total energy of deformation of steel
structure in MJ and RT is the total volume in m3 of the
deformed steel structure. Since energy absorption
capacity of a structure also depends upon geometric
configuration, a simple formula based on the volume of
the structural material in deformation is not likely to
produce accurate results especially for the collision
involving ships having structural design much different
from those ships studied to develop the empirical
correlation. This method developed from high energy
impacts in ships collision is, however, simple and is the
most cost effective as yet to make quick and reasonable
estimate of energy absorbed in ship structural
deformation and extent of structural damage that is likely
to occur in case of collision.
The simplified analytical method [9] aimed to improve
the estimation of structural deformation energy by
dividing the structure into independent components. The
energy of deformation for various mode of damage of
each structural element, so called super-element, is
determined from simplified analytical methods. The
energy so obtained for all the elements is summed up to
get the total energy of deformation for the structural
damage. The failure modes such as tearing of bottom
plates, folding of web frames, stretching of shell plating
etc. based on the observations from ship accidents and
experiments are included in the analysis. The super-
elements include simple shell, beam, T-form intersection
and X-form intersections etc. Details on the method may
be found in the reference [9].
As an example, the energy absorbed in crushing and
folding damage mode of plate is given by the following
The Damaged Ship, London, UK
closed form formula (2):
0.67
t
E d = 3.50 0 RT (2)
b
where t is the average thickness of the crushed plates, b
is the average width of the plates in the crushed cross-
section, o is flow stress and RT is the volume of crushed
structural material. Similarly, expression for all the
relevant mode of failure of structural element have been
developed and used in calculation of total structural
deformation energy.
The super-element approaches only count energy
absorption on element by element basis and don't take
count of energy propagation and absorption in plastic
deformation of adjoining structure that is mostly present
in high energy impacts.
The simple Finite Element Methods (FEM) uses
nonlinear static structural finite element analysis to
estimate the structure deformations for a given energy
level.
The dynamic FEM uses the explicit analysis techniques
to simulate multi-body dynamics to determine the
structural deformation due to impact. The finite element
methods for collision analysis are comparatively accurate
but are expansive needing more time and resources.
3. RESIDUAL STRENGTH OF DAMAGED
SHIP STRUCTURE
The work done on residual strength assessment includes
that of Smith and Dow [11] who carried out residual
strength of damaged steel ship and offshore structures by
finding strength reduction of damaged panel and effect
on ultimate strength of the hull girder. Paik et al [8]
developed methods for quick estimate of residual
strength of ships after collision and grounding. Wang, et
al [12] developed regression model to estimate residual
ultimate strength for different types of damaged ships.
Gordo and Guedes Soares [4], Ziha and Pedisic [13]),
and Fang and Das [3] all in general followed the similar
approach of damage material removal and assessment of
residual strength from remaining structure using the
conventional methods.
The residual stress and deformation induced in the
adjoining structure during collision or grounding impact
also influence the residual ultimate strength of damaged
ship. Further, such influences dependent on the nature,
size and location of damage. For relatively small impact
and consequently small damage area of structure, the
induced deformation and residual stresses in adjoining
structure shall be small and may be considered
negligible. The simple approach of assessment of
damaged structure by material removal is likely to
produce acceptable results. The large damage mostly
The Damaged Ship, London, UK
occur as a consequence of high energy impact that is
likely to propagate deformation in adjoining structure
and resulting in considerable residual stress as well,
leading to reduction in load bearing capacity of these
structural part and reducing overall ultimate strength of
the hull girder. As an example, residual stress field
obtained from a typical simulation of ships collision is
shown in the Figure 1 below [10]. In such cases, the
simple approach of removal of material of damaged part
for estimation of ultimate strength of hull girder is not
likely to produce accurate results.
Figure 1: Residual stress in a damaged ship structure
(evaluated using FEA)
Further, in case of large structural damage, the transverse
and torsional stiffness of hull girder may reduce to extent
making torsion moment critical load in combination with
vertical and horizontal moments. Therefore, the
conventional ultimate strength assessment procedures
based on progressive failure analysis of beam-columns
elements imposing increasing curvature and using load
shorting curves based on vertical and horizontal moment
interaction may not produce accurate results. For use of
such methods, it shall require modification of load
shortening curves and procedure of progressive failure
analysis to include torsional moment along with vertical
and horizontal bending moments, and vertical shear
force.
The static nonlinear finite element analysis may be used
as usual to assess ultimate strength of damaged structure
against combination of torsion, vertical and horizontal
moment, and shear loads. The initial deformation and
residual stresses arising out of high energy impact shall
be difficult to estimate and include in the ultimate
strength analysis.
The explicit dynamic finite element analysis may be used
to simulate impact and structural damage giving
structural deformation and residual stresses for
assessment of ultimate strength of damaged structure
against combination of torsion, vertical and horizontal
moments, and shear loads using static nonlinear finite
element analysis. This method is computationally
expensive. It shall, however, yield more accurate results
if finite element model is developed carefully to arrest
and minimise modelling errors.
4. COMBINED APPROACH - DAMAGE
PREDICTION AND RESIDUAL
STRENGTH ASSESSMENT
The analysis procedure for damage prediction and
residual strength assessment may be combined together
using the finite element methods. The explicit dynamic
FE analysis is to be first used in simulation of impact
process giving energy against level of penetration as
shown in the Figure 2. The model along with residual
stress is then used in evaluation of residual ultimate
strength of the hull structure using static nonlinear FE
analysis.
The parameters related to analysis for damage prediction
includes impacting ships speed, impacting ship shape at
impact location and impact angle that are to be selected
through suitable design of experiment to develop a
statistical parametric model for probabilistic prediction
of damage to ship structure. Similarly, statistical
parametric model for prediction of ultimate strength of
ships hull may be developed as illustrated in the Figure 2.
5. RAPID ASSESSMENT FOR EMERGENCY
RESPONSE SERVICES
In practice, a number of ships are built on a single design
with a life cycle of each ship exceeding 25 years. It may,
therefore, be appropriate to see the implementation,
effort and cost for rapid reliability and risk assessment
tools for Emergency Response Services in this context.
The parametric model developed for a given design of
ship as outlined in the Figure 2 may be used for rapid
prediction of damage to ship structure as illustrated in the
Figure 3. The residual strength assessment may be made
as illustrated in the Figure 4 using the predicted level of
damage of structure along with prevailing ships loading
and environmental conditions in the area of incident and
likely sea route for safe return of the ship.
The important feature of the method is to make requisite
extensive analysis and evaluation a priory to develop a
Response Surface (RS) or Artificial Neural Network
(ANN) model using statistical techniques. These
statistical parametric models are than to be used for rapid
assessment of damage and residual strength of the ship
structure.

Figure 2: Development of RSM or ANN model for
Emergency Response Services using FEM
Figure 3: Structural damage prediction for
Emergency Response Services
The Damaged Ship, London, UK
Figure 4: Evaluation of residual structural strength of
damaged ship for Emergency Response Services
6. MODELLING - RESPONSE SURFACE
AND ARTIFICIAL NEURAL NETWORKS
For large and complex systems such as ship structure
needing extensive resources and effort in evaluation, a
convenient approach is to identify the system through
evaluations for few key points and to determine the
system behaviour for intermediate points through
interpolation and simulation. The model development
process in general consists of following steps:
Step1: Design of experiment to determine
minimum number of FE analysis required to
suitably model the structural response.
Step2: Response modelling from FE results
obtained in step-1 using appropriate system
identification and simulation technique such as
response surface, artificial neural network etc.
Step3: Using response model with reliability
method for determination of risk of damage to
ship structure and reliability of residual strength
of ship after damage.
The statistical techniques used in selection of minimal
sample data points for system response modelling are
called design of experiment [5]. The design of
experiment is mostly based on Central Composite Design
and Box-Behnken Matrix.
Central Composite Design: using this method,
sampling points are located at five different
the levels for each input variable. One point is
located at the central value of each variable and
this point is termed as central point. For each
variable, there are two axis points that is, for N
number of variables there shall be 2N axis
points. In factorial central composite design of
experiment, there are 2N-F factorial points
where F is fraction of the factorial part. There
shall be a total of 1+2N+2N-F points to be
evaluated to determine system response.
The Damaged Ship, London, UK

Box-Behnken Matrix: The Box-Behnken [1]

design of experiment is a three level {-1, 0, 1},
second order design. These designs are formed
by combining 2N factorial with incomplete
block designs.

In fact, design of experiment depends a lot on domain of

application and need careful implementation to capture
sufficient data with minimal effort for identification and
characterisation of system response to model the system.
Since their first introduction by Box et al in 1951 [2], the
Response Surface Methods are in use in many
applications to simulate systems response. The response
surface is basically regression fit of a polynomial to the
Figure 5: Ultimate strength of hull girder using moment
real response data to approximate general system
interaction response model
response. A typical general form of polynomial
commonly used for regression fit is given in equation (3)
The response surface approximation of a linear system is
below.
essentially a linear polynomial only involving first order
n n n
terms in (3) and are straight forward to develop and use
Y = C 0 + Ci X i + Cij X i . X j + (3) in system response simulation. Unfortunately, most of
i =1 i =1 j =1
real systems are non-linear in nature requiring higher
order polynomial to represent their response. Sometime,
it is also possible to use a suitable transformation on
Where Y is an estimate of the fitted function for the n response data in order to improve regression fit.
number of random variables X. The C0 is an independent Nevertheless, accuracy of response surface
regression coefficient, Ci are coefficients of first order approximation depends on the nature of system response
terms and Cij are coefficients of second order and cross that may not be easily described by polynomial function
terms in the approximating polynomial function where i of the form (3) above. For example, see the Figure 6
& j are 1n. And, the is error in the modelled response showing attempt to model ultimate longitudinal strength
from actual response. of a ship on the basis of hull moment interaction using
polynomial response function (RS) (3) for 2nd order and
A priori knowledge of the type of response of a system 3rd order approximation [10]. The plot also shows the
may be used to reduce number of sample data points response model based on equation (4) marked as
needed to generate system response function. The standard RS. The short coming of the method to model
expected response of the system under consideration may response in highly nonlinear region is apparent. The
be known on the basis of other well-known systems of Artificial Neural Network (ANN) can provide an
similar type. In the case, it shall be efficient to use the improvement over the polynomial response model.
known type of response model with calibration of model
parameters. For example, the vertical and horizontal
moment interaction of conventional ship hull girder is
characterised, in general, by the following formula [7]:

c1 c2
Mv Mh
+ = 1 (4)
M uv M uh

Where Mv, Muv, Mh and Muh are vertical moment,

ultimate vertical moment, horizontal moment and
ultimate horizontal moment, respectively. The parameter
c1 and parameter c2 are unique for each type of ship
structure. It may therefore be more efficient to use above
model for response surface generation of ultimate hull
strength of a conventional ship based on moment Figure 6: Response Surface Model ultimate longitudinal
interaction by determination of parameter c1 and c2 strength of a ships hull
through experimental data points or finite element
analysis as shown in the Figure 5 [10]. The artificial neural network consists of multiple layers
of neurons as shown in a simple illustration of the Figure
7. The neurons in first layer are for inputs and that of last
are for outputs. The layers between input layer and out

put layer are called hidden layers. The each neuron in a
layer is connected to all neuron of succeeding layer.
Mathematically, the function of a neuron is as follows:
n
y ( x) = wi xi (5)
i =1
Where xi are n number of inputs to the neuron, wi are
weight associated to xi inputs, and y(x) is the output of
the neuron. The () is activation function that shape
the output of neuron and is selected according to
application of neural network. The choice of activation
function is the most important factor determining the
neural network behaviour, performance and application.
Further, the number of layers in neural network and the
number of neuron in each layer are import design
parameter on which depend the number of data points
required to train the network for simulation.
Figure 7: A simple illustration of artificial neural
network (ANN)
As discussed above, the main short coming of the
polynomial response function (3) is its inability to adapt
to local variations that make it difficult for dynamic
update of local response to improve accuracy around
design point. The ANN are based on neurons with
weighted functions as given in (5) above that adapt well
to local variation. As an example, compare the out put of
ANN model plotted in the Figure 8 for the same data as
of the Figure 6 for polynomial response model [10].
The accuracy of system response simulation on one hand
and minimizing the number of data points needed in
development of appropriate response model on the other
hand are active field of research.
The Damaged Ship, London, UK
Figure 8: ANN model of ultimate longitudinal strength of
a ships hull
7. ULTIMATE STRENGTH OF DAMAGED
SHIP STRUCTURE
In general, the ultimate strength of a damaged ship
essentially depends upon physical characteristics of
structural damage such as location of damage that
include distance of centre of damage from appropriate
ship reference, and size of damaged part of hull structure
that include damaged depth, damaged height and
damaged width as shown in the Figure 9 [10]. And also
variation in properties of material used in construction of
hull and aging effect such as corrosion and accumulation
of stress during ship in active service (fatigue cycles).
Further, the loading pattern for damaged ship may be
different from that of intact ship depending upon the size,
location and nature of damage to the hull. Accordingly,
the residual ultimate strength of damaged ship depicted
by ultimate moment capacity Mu may be given by the
following:
M v , M h , M t , Fs ,
M u = f D h , D w , D d , Dl , D v , (6)
Tage , y , E
Where f(...) is the ultimate strength response function
which depends on:
Mv vertical moment,
Mh horizontal moment,
Mt torsional moment,
Fs shear force
Dh height of damaged structural part,
Dw width of damaged structural part,
Dd depth of damaged structural part,
Dl longitudinal location of structural damage,
Dv vertical location of structural damage,
Tage aging related structural degradation,
y yield strength of material,
E Modulus of elasticity of the material.
The first four parameters are pertaining to loads
The Damaged Ship, London, UK

combination, the next five parameters D() are related to Therefore

location of structural damage and extent of the damage. (1 / c 5 )
FS
c6
The Tage is parameter to count for structural components M
degradation due to ageing effect as already discussed Rvs = V = 1 (13)
above. The yield strength and modulus of elasticity are M VU FSU

properties of material of the structure of the ship.
And for horizontal ultimate moment

M HU = M HU R ht R hs (14)

Where Rht and Rhs are horizontal ultimate moment

reduction factors due to torsional moment and sectional
shear forces, respectively. For

c7 c8
Figure 9: The parameters for modelling of damaged ship MH M
structure. + T = 1 (15)
M HU M TU
If we consider a particular structural damage D of a ship
at location L we can write: We have
(1 / c 7 )
M T
c8
MU = f ( M v , M h , M t , Fs ) M
D , L ,Tage ,y , E
(7) Rht = H = 1 (16)
M HU M TU

We can model response function using the interaction
function as given in equation (4) above as follows. And
c9 c10
MH F
MV
c1
M
c2 + S =1 (17)
+ H =1 M HU FSU
M M (8)
VU HU
Therefore
(1 / c10 )
FS
c9
Where M VU and M HU are ultimate vertical and M
Rhs = H = 1 (18)
FSU
horizontal moment subjected to moment reduction due to
torsional moment and sectional shear force as follows: M HU

For vertical ultimate moment: MV *

And if we represent by symbol Mv and similarly
M VU = M VU R vt R vs (9) M VU
for other ratio, we can rewrite equation (8) with
Where Rvt and Rvs are vertical ultimate moment reduction substitution of results for ultimate moment reduction
factors due to torsional moment and sectional shear factors as follows:
forces, respectively. For
c1
c3 c4 *
MV M M +
+ T =1
V

( ) (1 (F ) )
(10)
1/ c5
M VU M TU ( )
1 MT
* c 4 1/ c3
S
* c6

c2 (19)
We have M *
(1 / c 3 ) H =1
M M
( ) (1 (F ) )
c4
1/ c9
Rvt = V = 1 T

(11) ( )
1 MT
* c8 1 / c 7 * c10

M VU M TU
S

The state of failure of structure is therefore represented
And by
c5 c6
MV F
+ S =1 (12)
M VU FSU
 The Damaged Ship, London, UK

c1
M *
V +

( ( )
1 MT
*
) (1 (F ) )
c 4 1/ c3
S
* c6
1/ c5


c2 (20)
M *
H 1 > 0

( ( )
1 MT
*
) (1 (F ) )
c8 1 / c 7
S
* c10
1/ c9


Accordingly, limit state function g such that g0
representing structural failure state is defined as follows:

c1

M V*
g = 1 + Figure 10: Comparison of standard response model and

(
1 M T* ( ) ) (1 (F ) )
c 4 1/ c3
S
* c6
1 / c5


(21)
ANN model

c2
8. RELIABILITY AND RISK ASSESSMENT
M H*

( ( )
1 MT
*
) (1 (F ) )
c8 1 / c 7
S
* c10
1/ c9





The standard reliability measure is defined in (23).

= 1 ( P f ) (23)
The cn where n=110 in equation (21) are model
parameters to be determined using nonlinear static finite
element analysis.
Where, 1 (..) is inverse of standard normal
The equation (21) defines the failure limit state for a distribution and P f is probability defined by (24)
particular intact or damaged condition. The parameters cn
below.
are to be further developed as a function of damage
parameters L and D pertaining to damage location and
damage extends to completely define a response model
for rapid assessment for emergency response services.
Pf = f ( X )dX
g ( X )0
(24)

For the case when variations in sectional shear force can Where X is vector of random variables that depict the
be ignored, we can simplify equation (21) as follows: random characteristics of structural strength and loads
parameters, and f(X) is joint probability density function

c1 of X. The g(X) =0 is called limit state dividing
MV *
+ performance of the structural component into failure state
g = 1
(
1 M T*

( ) c 4
)
1 / c3


(22)
(i.e. g(X) <0) and safe state (i.e. g(X) > 0). The g(X) 0
defines the failure domain over which integration of (24)
is performed to determine probability of failure.

c2
M *
H The limit state functions g(X) for ultimate hull strength is

(1 ( )
M T
* c6
)
1 / c5
The ship structural response may be modelled using
artificial neural network (ANN) to improve accuracy
instead of response model of equation (21) or equation
(22) by the normalized sectional shear force and moment
components from nonlinear static finite element analysis.
A comparison of the two types of modelling method is
given in Figure 10 as an example [10]. The model was
developed using static nonlinear FEA to determine
ultimate strength of a damaged ship structure. The
structural damage was simulated using explicit dynamic
multi-body FE simulation.
given by equation (21) that may be used for reliability
analysis of intact as well as for damaged ship conditions.
The ultimate moments and ultimate sectional shear are
characteristic properties of structural material and
configuration. The random variation in properties of
material, such as that of yield strength and modulus of
elasticity, and random variation in structural
configuration, such as that of structural members
dimensions, initial deformation, residual stress etc, may
be taken care of using statistical characterisation of
ultimate moments and ultimate section shear stress. For
loads, the conventional approach is to divide bending
moment on the ship in components on the basis of their
source that is still water bending moment (Msw) due to
difference in weight and bouncy, the wave induced
moment (Mw) and dynamic moment (Md) due to ship
The Damaged Ship, London, UK
motion, impact and slamming. The total moment acting
on ship hull is given by the following equation:
M = M sw+ k w ( M w + k d M d ) (25)
The kw and kd are the load combination factors for wave
induced loads and dynamic loads, respectively. The
random characteristic of the ship loads can be derived
from the wave spectra for a given sea condition. The
detailed discussions on the statistical properties of
strength and load parameters are beyond the scope of this
paper.
The statistical parameters pertaining to loads and
structural strength along with limit state of equation (21)
is used with equation (24) and (23) to determine
reliability indices. Since, load depends on sea state and
ship heading, it is, therefore, appropriate to develop polar
plot of reliability indices for various sea states and ships
heading as illustrated in Figure 11 below [10]. Such polar
plot can readily be used during ship operation by the ship
staff to reduce risk of structural failure for high sea
states. The specialized reliability plot for likely damaged
conditions may be developed for emergency response
services. A well implemented such system shall be an
excellent computer aided decision support tool for rapid
structural assessment and to make informed emergency
response action to minimise loss and risk subsequent to
an accident.
Figure 11: The reliability indices for ship heading and
sea states
9. CONCLUSIONS
For Emergency Response Services, the rapid assessment
of reliability and risk of damaged ship structure is
probably the most important element for informed
response action to minimise risk of total lost of a ship as
a consequence. A comprehensive system has been
described in this paper to develop tools for rapid
assessment of risk of structural damage and reliability of
residual structural strength of a ship. Such system may be
configured to provide requisite structural assessment
with minimal information that is only available at an
early stage of an incident. These tools shall also be
helpful for decision support of ship operation for various
environmental conditions.
10. REFERENCES
1. Box G.E.P. and Behnken, D.W., Some New
Three Level Designs for the Study of Quantitative
Variables, Technometrics, 2, 455-475, 1960
2. Box, G.E.P. & Wilson, K.B., On the
experimental attainment of optimum conditions.,
Journal of the Royal Statistical Society, Series B,13:1
45. 1951
3. Fang C & Das PK., Hull girder ultimate strength
of damaged ship, 9th symposium on practical design of
ships and other floating structures, Luebeck
Travemuende, Germany; 2004a., 2004
4. Gordo JM, Guedes Soares C., Residual strength
of damaged ship hulls, Proceedings of the 9th
international congress of international maritime
association of the mediterranean, Ischia, Italy; 2000.
5. JeffWu, C.F. and Hamada, M., Experiments
Planning, Analysis, and Parameter Design
Optimization, Jhon Wiley & Sons, Inc, ISBN 9-81-
412671-3, 2002
6. Minorsky V.U., "An Analysis of Ship Collision
with Reference to Protection of Nuclear Power Ships", J.
of Ship Research, Vol. 3, No. 2, pp. 1-4, 1959
7. Paik J.K., and Thayamballi A.K., 'Ultimate Limit
State Design of Steel-Plated Structures, John Wily &
Sons Inc, ISBN 0-471-48632-9, 2003
8. Paik JK, Thayamballi AK, Yang SH.,'Residual
strength assessment of ships after collision and
grounding., Marine Technol 1998;35:3854, 1998
9. Pedersen P.T. and Zhang, S., Collision Anlysis
for MS DEXTRA SAFER EURORO spring meeting,
NANTES 28 April 1999
10. Shahid M., Development of Structural Reliability
Techniques And Their Application to Marine Structural
Components and Systems, PhD Thesis, Department of
NAME, University of Glasgow and University of
Strathclyde, Glasgow, Scotland, UK, 2008
11. Smith, C. and Dow, R., Residual Strength of
Damaged Steel Ships and Offshore Structures. Journal
of Constructional Steel Research, Vol. 1, No. 4,
September, 1981
12. Wang, Ge; Chen, Yongjun; Zhang, Hanqing;
Peng, Hua., Longitudinal strength of ships with
accidental damages. Marine Structures, v 15, n 2, 2002,
p 119-138., 2002
13. Ziha K, & Pedisic M., Tracing the ultimate
longitudinal strength of a damaged ship hull girder,
International Shipbuilding Progress 2002;49(3):16176.,
2002
 The Damaged Ship, London, UK

11. AUTHORS BIOGRAPHY

Dr M Shahid PhD, holds the current position of Director

at Binary Systems and Engineering (UK) Limited. Detail
of the company may be found at www.bsyse.co.uk. His
previous experience includes working at various
appointments for Pakistan Navy, as General Manager in
Maritime Technologies Complex, Pakistan and as
Research Fellow at the department of Naval Architecture
and Marine Engineering (NAME) that is a joint
department of University of Glasgow and University of
Strathclyde, Glasgow, Scotland, UK.

The author may be contacted by email at

shahid@bsyse.co.uk
The Damaged Ship, London, UK

PROGRESSIVE STRUCTURAL FAILURE AND RESIDUAL STRENGTH OF DAMAGED

SHIPS
S. Kwon, D. Vassalos and G. Mermiris, The Ship Stability Research Centre (SSRC), Department of Naval Architecture
and Marine Engineering, University of Strathclyde, UK

SUMMARY

This research addresses progressive structural failure of a damaged ship. Depending on the size and location of the initial
damage, progressive structural degradation may threaten the safety of the ship by causing unwanted exacerbated
flooding and/or further reduction of its residual strength. A general methodology is proposed to cover the coupled
problem of loss of stability and loss of structural integrity of a damaged ship due to progressive structural failure. The
procedure includes calculation of loading, progressive structural failure analysis, residual strength assessment and
damage stability assessment. Due to the nature of evolution of the damage and its ensuing consequences, it is important
to adopt a time domain approach. The current stage of development of the methodology is demonstrated by the
application for the case of an Aframax tanker with bottom damage.

NOMENCLATURE extensively and to great depth in literature for passenger

ships, [22]. Assessment of residual strength of the
Poissons ratio damaged ship due to loss of damaged parts has been
BM Residual bending strength of damaged carried out adopting the method to calculate the ultimate
ship hull girder strength of the damaged hull section.
BMmax Maximum bending moment allowed Applications can be found in [15], [19] and also in some
CFD Computational fluid dynamics jointly funded research projects by the European
D, m Material constant for crack propagation Commission and the industry, [6], [17], etc.
da/dN Crack propagation rate (mcycle-1)
E Youngs modulus (MPa) Traditionally, these two analyses have been carried out
FEA Finite element analysis independently with a strong assumption that the initial
G Strain energy release rate (MPam) damage extent remains stable. However, the initial
GM Metacentric height (m) structural damage can propagate during the evolution of
Hs Significant wave height (m) flooding due to dynamic loads and the imposed
KIC Toughness of material (MPa m ) environmental loads. The progressive structural failure
Kmax ,Kmin Maximum & minimum SIF (MPa m ) during these complex situations may result in excessive
flooding and ultimately hull girder collapse before the
LEFM Linear Elastic Fracture Mechanics
damaged ship founders or capsizes.
Q Parameter related with the yield
strength of the material
The importance of the progressive structural failure is
R Stress ratio, Kmin/Kmax
amplified by the loss of Prestige, which is the most
SIF Stress intensify factor
recent accident that resulted in total failure of its hull
Tz Zero up crossing period of wave (sec)
girder after 6 days from the initial damage occurrence at
VCCT Virtual Crack Closure Technique
its starboard side. In this particular case, it was not only
VOF Volume of Fluid
the initial damage but also the progressive structural
K Range of SIF (MPa m ) degradation that led the tanker to the seabed.
Keff Range of the effective SIF (MPa m )
Kmax Maximum Range of SIF (MPa m ) In addition, the newly introduced framework of Safe
Return to Port for passenger ship safety has opened the
Kth Range of threshold SIF (MPa m )
way for zero-loss tolerance and has questioned the
Kth,0 Range of threshold SIF at R = 0 presumption of structural stability and emphasis on the
damage stability assessment alone. That is, under the
premise that the ship should be designed to be its own
1. INTRODUCTION life boat, it was recommended that it should remain
afloat, upright and habitable for 5 days until it can return
The emphasis of a damaged ship due to accidental to port under its own power or until assistance has
loading or structural failure in severe environments is arrived [11]. However, depending on the extent of the
currently placed on its survivability in terms of damage initial damage and the prevailing weather conditions, this
stability assessment and residual strength assessment. time period is sufficient for total hull girder collapse to
occur with the progressive structural failure.
The ensuing loss of stability and potential capsizing /
foundering of the damaged ship have been covered
 The Damaged Ship, London, UK

Hence, the structural survivability of damaged ships by focusing on the stress variation and the time to
cannot be sufficiently assessed by the conventional static capsize respectively, [21]. The calculation process will
strength assessment with initial damage (i.e. removal of be repeated for varying loading input, until either the
the damaged parts of the hull and re-calculation of the strength or the stability criterion fails. A high level
section modulus of the ship), but should take into description of the proposed methodology is shown in
account the effect of the progressive structural failure Figure 1.
due to imposed environmental and flooding loads. The
nature of damage evolution and the change of the Initialdamage
environmental conditions clearly indicate that this (collision/grounding)
process should be carried out in the time domain and that
knowledge of the damage propagation will be invaluable
in the course of establishing the appropriate mitigation i=i+1 Timestep
actions. More importantly though, this information will T(i)
allow such extreme conditions to be taken into
consideration in the design stage and therefore heavily
invest in prevention and compliance with the safe return Loadingt(i1)
to port concept.
Flooding Environment
In response to the above requirement, this paper proposes
a methodology that combines both environmental
(global) and flooding (local) loads, damage propagation
estimation and assessment of residual strength and
damage stability. High level process in time domain is Responset(i)
briefly explained followed by explanation of each Progressive
component. Details of recent developments on a first- structural
principles approach for the simulation of progressive failure
structural failure of a damaged ship are reported with a
brief explanation on the other components as the study
on them has matured over the past few decades.
Verification of the proposed methodology is carried out
for the case of an Aframax tanker with bottom damage Residualstrength(BM) Damagestability (GM)
t(i) t(i)
with encouraging results.

2. METHODOLOGY PROPOSED

The key feature of this approach is the modelling of Pass Criterion

(BMBMmax)
damage evolution in the time domain. For a damage to AND(GM0)
propagate (in the form of plate cracking and stiffener
failure), a source of loading should be identified and
Fail
defined properly. The progressive structural failure in a
damaged ship is achieved by crack propagation analysis Termination
under wave loading. Furthermore, when the damaged Hullcollapse
ship is flooded the water that ingresses and egresses from and/orcapsize
the damaged compartment (due to the ship motion and Figure 1: High level process of the proposed
the waves) constitutes another source of loading that methodology
deteriorates the damaged area further and induces crack
propagation. In turn, the damage extension results to
further structural degradation and the cycle repeats itself 3. LOADING
until either stability is totally lost or residual strength
becomes insufficient to sustain the applied loads. The forces that drive the evolution of the initial damage
are environmental and flooding loading, and they are
For every time step of the calculation, the crack briefly discussed in the following lines.
propagation is analysed and the effect of damage
evolution on the survivability of the damaged ship needs 3.1 ENVIRONMENTAL LOADING
to be assessed in terms of criteria pertaining to residual
strength and damage stability. In principle, such criteria The load induced by the waves is the primary source for
can be expressed as the maximum bending moment crack propagation and it is calculated by an in-house 3D
(BMmax) and the metacentric height (GM) respectively, panel code with Greens function implementation for
and they should account for the dynamics of the situation wave load analysis. This code offers the capability of
The Damaged Ship, London, UK
calculating ship motions and wave loads both in time
domain and frequency domain with inclusion of non-
linear wave effects. The required data for wave load
analysis and the output is listed next.
a. Input
- ship geometry
- loading condition of the intact ship
- location and size of the flooded compartment(s)
- wave condition (Hs, Tz)
- draft and trim at the damaged condition
b. Output
- wave and still water bending moments
- ship motions (primarily heave, pitch and roll)
The obtained wave and still water bending moments
form direct input to the crack propagation analysis. The
total bending moment will be compared to the residual
strength of the damaged ship for every time step of the
process. The calculated ship motions are used in flooding
simulation.
3.2 FLOODING LOADING
If flooding in the damaged compartment and water
ingress and egress occur continuously due to ship
motions like rolling, heaving and pitching, the effect of
flooding pressure on the crack propagation should be
considered. Depending on the size and location of the
damage opening, the size of flooded compartment, the
hull motions and etc., the induced flooding pressure and
its extent is defined. The calculation is based on CFD and
in particular the solution of the Navier-Stokes equations
with the free surface capturing scheme, so called VOF
method. The finite volume method is used to discretise
the governing equations [8]. The required input data and
output of the analysis is:
a. Input
- damage size and location
- geometry of the flooded compartment
- wave condition(s) (Hs, Tz)
- draft and trim at the damaged condition
- ship motions
b. Output
- inflow / outflow velocity
- water height in the flooded compartment
- flooding pressure differential
- sloshing pressure
Time dependant data such as damage size, wave
condition and ship motions are updated at every time step
of the simulation.
4. PROGRESSIVE STRUCTURAL FAILURE
The crack growth modelling emanating from or near the
damaged ship parts is based on the LEFM and Paris Law
in which time is included indirectly in the number of
cycles, N, where N=time/period. This approach has
been adopted not only in the aviation industry, [7], but
also in the shipbuilding industry, [4] and [5].
Paris and Erdogan, [16], hypothesized that the range of
the SIF, K, governs the fatigue crack growth. The
empirical expression of Paris Law shows that the crack
growth rates are proportional to K when plotted on a
log-log scale but experimental data generally showed a
sigmoid shape in the same scale. This is attributed to the
range of threshold SIF, Kth, below which crack
propagation does not occur and the maximum range of
SIF, Kmax, at which the maximum SIF, Kmax, approaches
materials toughness limit, KIC, hence crack propagation
would be accelerated and either a ductile tearing or a
brittle fracture would follow.
Kth Kmax
Figure 2: Typical plot of crack growth rate
In this analysis, an equation proposed in [14] to include
both the range of threshold SIF and the fracture
toughness of the material is utilized. In addition, the
effective range of SIF, Keff, is used as proposed in [23]
to include the effect of non-constant stress ratio, R, of the
applied loads. Hence the equation of crack propagation in
this analysis is obtained as follows:
K eff
= D (K eff K th ) 1 +
da m
dN K IC K max
where, K , K = K (1 R )1 / 3
K eff = th th , 0
1 R / Q
The material constants of D and m are chosen so as the
linear region of the above equation with R = 0.0 to fit
with the original Paris Law.
The fracture behaviour is modelled in ABAQUS (ver.
6.8) with the VCCT for the calculation of the strain
energy release rate (G), and, in turn, the calculation of
the SIF (K) as shown below.
K 2
for plane stress
G = E2
E (
K 1 2 ) for plane strain

The ABAQUS VCCT utilises an intrinsic capability that
analyzes brittle interfacial crack propagation due to
delamination or debonding, [18]. Because VCCT can be
calculated at runtime rather than as a post-processing, no
re-meshing or re-analysis is required and the analysis
time is reduced.
According to [13], VCCT is based on LEFM, which can
be used under conditions of little or negligible plastic
deformation around the crack tip. Therefore, it is
justifiable to use VCCT in this research as long as the
crack propagation is considered under the yield strength
limit. Although this would not be entirely accurate for
the early stages of the damage evolution, at later stages,
where accelerated crack propagation is observed, this
assumption would hold true.
As the validity of VCCT for ABAQUS in calculating SIF
of cracks emanating from a circular opening has been
demonstrated in [12], a procedure for evaluating
progressive structural failure using Paris Law and
VCCT can be formulated as shown in Figure 3 for
damaged ships.
ParisLawSIFfromVCCTCrackPropagation
Figure 3: Proposed procedure of progressive structural
failure analysis
The required input and output of the process is
summarized below.
a. Input
- ship geometry
- initial damage location and size
- wave bending moment
- flooding pressure
- material properties (constants D and m) for
crack propagation
b. Output
- crack size in the time domain
The time dependant data, such as loading, damage and
crack size, will be updated for each time step. The
extended crack size shall be used (i) in the residual
strength assessment, and (ii) to update the opening of the
flooded compartment and any adjacent ones if its
boundaries are breached.
5. RESIDUAL STRENGTH
The post-damage residual strength assessment of a ship
is based on a moment-curvature relationship for stiffened
panels, [10].
The Damaged Ship, London, UK
The moment-curvature result is obtained by imposing a
curvature from sagging to hogging on the hull girder,
which is assumed to consist of several beam-column
elements. For each curvature, the average strain of each
element is calculated and the stress imposed on each
element is obtained from the corresponding load-
shortening curve. The moment sustained by the whole
section is obtained by summing up the moments of each
element induced by axial force and distance of each
element from the neutral position of the section. The
ultimate bending strength of the section is the maximum
bending moment in the moment-curvature curve in
hogging and sagging conditions.
The load-shortening curve of each element is obtained
according to [9], which adopted an elastic-perfectly
plastic behaviour of material for each element. Defining
a load-shortening curve includes the effect of plate
induced failure, flexural bucking failure of column and
tripping failure of stiffener. The effects of residual stress
and initial deformation are also considered.
Validation of the developed code was made through
comparison with FEA results in [12].
The summarized input and output of the residual strength
analysis is listed below. As damage evolves the effective
width of plate as well as loss of stiffeners of the damaged
section will be updated for each effective time step so
that the ultimate residual bending strength can be
estimated appropriately.
a. Input
- load shortening curve of each element, which is
composed of a stiffener and associated plate
- neutral axis of the respective cross section for
the damaged condition
- range of curvature from sagging to hogging
b. Output
- ultimate residual bending moment capacity of
the damaged section in hogging and sagging
6. DAMAGE STABILITY
The time to sink and/or capsize following damage is a
critical factor in determining the actual level of safety of
a damaged ship. The dynamic response of a damaged
vessel and the progressive flooding of its compartments
in a random seaway form a highly non-linear dynamic
system, the behaviour of which can only be modelled
with time domain simulation. This is carried out by in-
house software called PROTEUS.
The main elements of PROTEUS can be summarised as
follows:
- Ship hydrodynamics, derived from properties of the
intact hull, are based either on asymmetrical strip theory
formulation with Rankine source distribution or a 3D
The Damaged Ship, London, UK

source code, both accounting for the non-linearities

arising from instantaneous variation of the mean ship
attitude and large amplitude motions.
- Effects of floodwater dynamics described by a full set
of non-linear equations derived from rigid-body theory.
- Floodwater motions modelled as a Free-Mass-on-
Potential-Surface (FMPS) de-coupled system in an
acceleration field.
- Water ingress/egress based on Bernoullis equation.

The ship geometry is defined in sections for both the hull

and the internal compartments. The necessary
environmental conditions for the simulations are
generated according to JONSWAP or Pierson-
Moskowitz spectra. The output of the calculation concern
ship motions, floodwater mass variation and motion,
flow of floodwater through openings, environmental
forces, etc.
Figure 4: Drawing of an ordinary section of the parallel
body
7. VERIFICATION
7.1 DAMAGE DESCRIPTION
An application of the proposed methodology was made
for the case of an Aframax tanker with bottom damage. A case of damage on the bottom structure of the tanker
In this verification process, it is assumed that only still will be considered based on the ABS guidance, [1],
water bending moment and wave loading are applied. which has been used to define damage extents due to
Also, the damage stability calculation is ignored and the collision and grounding accidents.
focus is placed on the residual strength assessment only.
Based on this reference, the location of the damage is
The ship used in the analysis is an 112,700 DWT crude considered to be in the most unfavourable location
oil carrier fitted with double side structures. The ship has anywhere on the flat bottom within the fore part of the
six cargo tanks on each side. The principal dimensions of hull between 0.5L and 0.2L aft from F.P. Also, it is
the ship are summarized in Table 1 and the drawing of considered that the bottom structure is damaged over a
the midship section is shown in Figure 4. considerable length and the damaged members should be
excluded from the hull strength calculation. The extent of
Table 1: Principle dimensions of tanker under a bottom damage is considered to include bottom girders
consideration attached to the bottom shell plating to a certain depth as
well as bottom shell plate. The structural members
Type of dimension Value
assumed to be damaged and excluded completely or
Length O. A. 250.17 m
partially are explained below (Figure 5).
Length B. P. 239.00 m
Breadth MLD. 44.00 m - Bottom shell plating for a width of 4 m or B/6,
Depth MLD. 21.00 m whichever is greater, where B is the ship breadth;
Draught MLD. (Design) 14.60 m - Double bottom girders attached to the damaged shell
Draught MLD. (Scantling) 14.60 m plating are assumed to be damaged and ineffective up to
Frame spacing 3.78 m the following percentage of the girders height:
25% for girders situated within 1 metre marginal
zones of the damaged plating
75% for girders situated between the marginal zones

- All of the bottom longitudinals within the damaged

bottom shell and the longitudinal stiffeners within the
damaged parts of girders.
 The Damaged Ship, London, UK

loss of buoyancy due to the flooded compartments in

double bottom where the damage is initiated.

Wave data shown in Figure 7 is assumed to be the same

as those estimated in [2] for the Prestige accident. For
simplicity, linear regular wave is assumed in head seas.
The dynamic wave bending moment in hogging and
sagging is calculated according to the wave data as
shown in Figure 7.
Figure 5: Extent of damage due to grounding accident,
[1] Waveheight,Hs Zerocrossingperiod,Tz
18

Waveheight,[m],Period,[sec]
In this analysis, the damage is assumed to be located at 16
14
the centre of bottom structure near amidships within the 12
No. 4 cargo tanks (port and starboard). The extent of the 10
8
initial damage is defined below and seen in Figure 6 6
according to the guidance and a circular opening damage 4
2
is assumed for the convenience. 0
0 12 24 36 48 60 72 84 96 108 120 132 144
1day2days3days4days 5days6days

- Width of damage: 7.33 m (= B/6) Time,[hour]

- Height of damage at centre girder: 1.875 m (= 0.75H) Hogging Sagging

1600
Wavebendingmoment,[MNm]

1400
1200
1000
800
600
400
200
0
0 12 24 36 48 60 72 84 96 108 120 132 144
1day2days3days4days 5days6days
Time,[hour]

Figure 7: Wave data from [2] and corresponding wave

induced bending moments

7.3 PROGRESSIVE STRUCTURAL FAILURE

With the damage extent defined above, a finite element

model of the Aframax tanker is prepared for calculation
of stress intensity factors with VCCT for ABAQUS.
1.875m

7.33m The FE model (Figure 8) consists of No. 3, 4 and 5 cargo

Initialcrack,a0 =36.65mm tanks. 4-node shell elements are mainly used to model
hull plating and web of longitudinal stiffeners while 2-
Figure 6: Extent and location of grounding damage node beam elements are adopted for modelling of flange
of longitudinal stiffeners as well as other stiffeners.
Initial cracks are assumed to be symmetrically emanating
toward the side shell plates so that the vertical bending
moment from wave loads allows the cracks to propagate.
The initial length of crack is assumed to be 0.5% of the
width of initial damage or 36.65 mm of crack size at each
side of edge of damage is adopted as the initial condition
(see Figure 6).

7.2 WAVE LOADS

Wave loads are calculated with the in-house panel code.

Fully loaded condition, in which still water bending
moment at the damaged section is calculated as 1329
MN-m with sagging condition, is considered with some
modification of draft and trim taking into account the Figure 8: FE model of the Aframax tanker
The Damaged Ship, London, UK

Implementation of VCCT in ABAQUS requires 3-D
solid mesh between the bonded surfaces. Hence, the hull
structures around the defined damage have been
modelled with 8 nodes solid element. Connection of shell
elements to solid elements near the boundary of solid
meshes is achieved through a surface-based shell-to-solid
coupling constraint, which is an intrinsic technique in
ABAQUS.
7.3 (a) SIF Calculation
Instead of applying the actual wave loadings to the FE
model, the SIFs are calculated using a unit (i.e. 1.0 MN-
m) vertical sagging bending moment on both ends of the
FE model and the actual SIFs corresponding to the actual
loadings are linearly extrapolated by the ratio of the
actual bending moment to the unit bending moment. The
calculated SIF curve with unit bending moment is shown
in Figure 9. The discontinuities of the curve correspond
to the longitudinal stiffener locations (at 850 mm
intervals).
320
270
220
SIF,[MPam]
smaller than the range of threshold SIF, Kth. This would
mean that the crack does not propagate any further,
hence a zero crack propagation rate is applied in this
particular case.
On the other hand, the above equation gives no crack
propagation rate when the maximum SIF, Kmax, exceeds
the material toughness, KIC. However, this does not mean
that the damaged hull girder would collapse immediately
(i.e. by the next load cycle) but implies that very fast
crack propagation is expected. The crack propagation can
slow down again in the next cycle depending on the
applied load. Therefore, it is more realistic to allow a
constant value of crack propagation when such situation
arises. In this analysis, 1mm/cycle of crack propagation
rate is applied to allow the crack to propagate until the
hull girder collapses. Considering the early stage of
development of the proposed methodology, further
elaboration on this rather conservative assumption
(especially for the latter stages of damage evolution) will
take place in future work. For the time being, the
outcome of the crack propagation analysis is shown in
Figure 10. The stable region of the damage followed by
an interval of rapid growth and instability has been
verified by salvage experts, whom the authors have been
in touch with: the combination of the two phases has

170 been repeatedly observed in the past and indicates that

120
the proposed methodology moves in the right direction.
However, quantification of this behaviour and more
70
precise definition of the two regions depicted in Figure
20 10 are not available at the moment.
3.5 5.5 7.5 9.5 11.5 13.5 15.5 17.5
Halfdamagesize(halfinitialopening+initialcracksize),[m]
25 Unstable damagegrowth
Stable damagegrowth
Figure 9: SIF curve for bottom damage case with 1.0
Halfdamagelength,[m]

MN-m bending moment applied. 20

15
7.3 (b) Crack Propagation
10
With the calculated SIF curve and wave loading data,
crack propagation analysis is carried out according to the 5
equation below:
0
K eff
= D (K eff K th ) 1 +
da m
0 12 24 36 48 60 72 84 96 108 120 132 144
1day2days3days4days 5days6days
dN K IC K max
Time,[hour]

Figure 10: Prediction of the extended crack length

The material constants are D = 7.3210-11 and m = 2.37.
The linear region of the above equation (with R = 0.0) 7.4 RESIDUAL STRENGTH
fits the Paris Law [3]:
With the predicted damage propagation results, residual
da
= 9.5 10 12 K 3.0 strength of the damaged ship is calculated by taking into
dN consideration the damage extent for every time step, i.e.
1 hour, including at the initial damage condition. Also,
A typical value of toughness limit of the steel is used as the ultimate strength of the tanker for intact condition is
140 MPa m and the range of threshold SIF at R = 0.0 is calculated for comparison purposes.
selected as 2.45 MPa m [20].

From a mathematical point of view the above crack

propagation equation may result in negative crack
propagation rate when the effective range of SIF, Keff, is
 The Damaged Ship, London, UK

Intact initialdamage 1daylater ABAQUS is promising for the problem at hand and
2dayslater 3dayslater 4dayslater conforms to past experience from the salvage industry.
Bendingmomentcapacity,[MNm]

2.E+04
- Progressive structural failure can be analysed for a
1.E+04 succession of sea states and offer invaluable information
for the residual strength of the ship under consideration.
5.E+03
Such knowledge can be deployed in the design stage
0.E+00
Sagging Hogging (forming a part of the risk-based design methodology) or
the operational stage in emergency situations and in the
5.E+03 safe-return-to port context.
1.E+04
6.E04 4.E04 2.E04 0.E+00 2.E04 4.E04 6.E04 - The resulting timeline of the damage propagation can
Curvature,[1/m] contribute to more robust decision-making and the
Figure11: Moment curvature curves aversion of Prestige-like accidents in the future.

The most evident challenge faced by the authors is the

Hogging Sagging Intact_hog Intact_sag
validation of the proposed methodology. The penultimate
15000
aim is to conduct the necessary simulations (coupled
Bendingcapacity,[MNm]

10000
with flooding) fully in the time domain and to recreate
the Prestige accident, the investigation of which is
5000 covered in a fairly detailed manner. In this way, it is
1day2days3days4days 5days6days hoped that the added value of the methodology for
0
safeguarding human life and environment will be clearly
5000
0 12 24 36 48 60 72 84 96 108 120 132 144 demonstrated to the scientific community and the
industry at large.
10000
Time,[hour]

Figure 12: Change of residual strength due to progressive 9. REFERENCES

structural failure
1. ABS, Guide for Assessing Hull Girder Residual
Table 2: Summary of residual bending strength Strength, American Bureau of Shipping, 1995
Hogging Sagging
Capacity BM Diff. BM Diff. 2. BAHAMAS MARITIME AUTHORITY, Report of
(MN-m) (%) (MN-m) (%) the Investigation into the Loss of the Bahamian
Intact Registered Tanker Prestige off the Northwest Coast of
10817 - 7866 - Spain on 19th November 2002, The Bahamas Maritime
condition
Initial damage 10229 - 5.43 7731 -1.72 Authority, 2004
1 day later 10225 -5.47 7728 -1.75
3. BS PD 6493, Fatigue Design and Assessment for
2 days later 10221 -5.51 7726 -1.78
Steel Structures, British Standards Institute, 1993
2.5 days later 10093 -6.7 7569 -3.78
3 days later 9545 -11.76 7560 -3.89 4. DEXTER, R.J., PILARSKI, P.J., SSC-413: Effect of
3.5 days later 9055 -16.29 7391 -6.04 Welded Stiffeners on Fatigue Crack Growth Rate, Ship
4 days later 8434 -22.03 7169 -8.86 Structure Committee, 2000
4.5 days later 7466 -30.98 6787 -13.72
5. DEXTER, R.J. and MAHMOUD, H.N., SSC-435:
In this example, the BMmax for the strength criterion Predicting Stable Fatigue Crack Propagation in Stiffened
equals 25% reduction in the bending moment capacity of Panels, Ship Structure Committee, 2004
the intact ship.
6. DEXTREMEL, Design for Structural Safety under
Extreme Loads, Final Technical Report of
8. CONCLUSIONS AND FUTURE WORK DEXTREMEL Project, 2001

On the basis of the results obtained by the application of 7. FARAHMAND, B., SAFF, C., XIE, D. and ABDI, F.,
the proposed methodology to the tanker considered in Estimation of Fatigue and Fracture Allowables for
this analysis, the following conclusions can be drawn: Metallic Materials under Cyclic Loading, American
Institute of Aeronautics and Astronautics, 2007
- A methodology for progressive structural failure
analysis comprising Paris Law and VCCT for 8. GAO, Z., VASSALOS, D. and GAO, Q., Numerical
Simulation of Water Flooding into a Damaged Vessels
The Damaged Ship, London, UK
Compartment by the Volume of Fluid Method, Ocean
Engineering, Vol. 37, pp. 1428-1442, 2010
9. GORDO, J.M., SOARES, C.G., Approximate Load
Shortening Curves for Stiffened Plates under Uniaxial
Compression, Integrity of Offshore Structures-5, EMAS,
pp.189-211, 1993
10. GORDO, J.M., SOARES, C.G., FAULKNER, D.,
Approximate Assessment of the Ultimate Longitudinal
Strength of the Hull Girder, Journal of Ship Research,
Vol. 40, No. 1, pp.60-69, 1996
11. IMO, Report of the Maritime Safety Committee on
its Seventy-eighth Session, MSC 78/26, 2004
12. KWON, S., VASSALOS, D., and MERMIRIS, G.,
Understanding Potential Risk from a Coupled Problem
of Flooding and Structural Degradation of a Damage
Ship, Proceedings of the 4th International Maritime
Conference on Design for Safety, 2010
13. LESKI, A., Implementation of the Virtual Crack
Closure Technique in Engineering FE Calculations,
Finite Elements in Analysis and Design Vol. 43, pp.
261~268, 2007
14. MCEVILY, A. J. and GROEGER, J., On the
Threshold for Fatigue-Crack Growth, Proceedings of the
4th International Conference on Fracture, Vol. 2, pp.
1293-1298, 1977
15. PAIK, J. K., THAYAMBALLI, A. K. and YANG, S.
H., Residual Strength Assessment of Ships after
Collision and Grounding, Marine Technology, Vol. 35.
No. 1, pp. 38-54, 1998
16. PARIS, P. and ERDOGAN, F., A Critical Analysis
of Crack Propagation Laws, Journal of Basic
Engineering, Vol. 85, Issue 4, pp. 528-534, 1963
17. SAFEDOR, Probabilistic Models for Collapse Limit
State, SAFEDOR-D-2.2.2-2006-07-31-IST-rev-4, 2006
18. SIMULIA, ABAQUS Analysis Users Manual ver.
6.8, Simulia
19. SOARES, C. G., LUIS, R. M., NIKOLOV, P.,
DOWNES, J., TACZALA, M., MODIGA, M.,
QUESNEL, T., TODERAN, C. and SAMUELIDES, M.,
Benchmark Study on the Use of Simplified Structural
Codes to Predict the Ultimate Strength of a Damaged
Ship Hull, International Shipbuilding Progress, Vol. 55,
pp. 87-107, 2008
20 SUMI Y., Fatigue Crack Propagation and
Computational Remaining Life Assessment of Ship
Structures, Journal of Marine Science and Technology,
Vol. 3, pp. 102-112, 1998
21. VASSALOS, D., Risk-Based Ship Design, In:
PAPANIKOLAOU, A. D. (ed.) Risk-Based Ship Design:
Methods, Tools and Applications, Springer, 2009
22. VASSALOS, D., JASIONOWSKI, A. and GUARIN,
L., Passenger Ship Safety Science Paving the Way,
Proceedings of the 8th International Ship Stability
Workshop, 2005
23. YAZDANI, N. and ALBRECHT, P., Crack Growth
Rates of Structural Steel in Air and Aqueous
Environments, Engineering Fracture Mechanics, Vol.
32, No. 6, pp. 997-1007, 1989
10. AUTHORS BIOGRAPHIES
Seungmin Kwon is conducting a PhD course with
support of Daewoo Shipbuilding and Marine Engineering
(DSME) in the Ship Stability Research Centre (SSRC),
Department of Naval Architecture and Marine
Engineering, University of Strathclyde. He has graduated
with MSc. (2000) from the Department of Naval
Architecture and Ocean Engineering, Seoul National
University, Republic of Korea and worked for DSME.
His study is focusing on implementation of Risk-based
Ship Design taking into account coupled problem of
progressive flooding and progressive structural
degradation in a damaged ship.
Dracos Vassalos is Professor of Maritime Safety in the
Department of Naval Architecture and Marine
Engineering of the University of Strathclyde and the
Director of the Ship Stability Research Centre (SSRC), a
world-leading centre of excellence on ship stability and
safety. His motto is safety enhancement through
innovation, an idea he has pursued single-mindedly in a
career that spans over 30 years in industry and academia,
promoting the use of scientific approaches in dealing
with maritime safety. He has been instrumental in
helping to create a critical mass in the research
community on safety, through a series of initiatives that
made SSRC the focus of active international
collaboration. He travels the world over promoting
maritime safety, lectures and publishes widely, with
some 400 technical publications, 4 patents and 7
books/Conference proceedings to his credit and a string
of prizes and awards, including some 100+ major
research contracts amounting to over 20M. He served
as Head of Department (1997-2007), Chair of the STAB
Conferences and Workshops (1996-2006), Chair of the
ITTC Stability Committee in Waves (1996-2002), Chair
of WEGEMT (the European Association of Universities
in Marine Technology 1999-2001). Currently, Professor
Vassalos is Chairman of the International Standing
Committee of the Design for Safety Conference, a
theme instigated and promulgated by SSRC. He is also a
long-standing member of the UK delegation to IMO for
ship stability and safety.
 The Damaged Ship, London, UK

George Mermiris is a Research Fellow in the Ship

Stability Research Centre (SSRC). He has graduated with
MEng (2003) and PhD (2010) from the Department of
Naval Architecture and Marine Engineering, University
of Strathclyde. The core of his work is on Risk-based
Design methodology and implementation with main
emphasis on structural analysis and collision dynamics,
advanced mathematical modelling and risk analysis. He
has authored 16 journal and conference papers in this
area.
The Damaged Ship, London, UK

GLOBAL WAVE LOADS ON DAMAGED SHIP STRUCTURES: AN EXPERIMENTAL

PROCEDURE AND SOME PRELIMINARY RESULTS
D Fone, University College London, UK
T W P Smith, University College London, UK
J Borg, Lloyds Register, London, UK
K R Drake, University College London, UK

SUMMARY

When a ship is involved in a damage event, a rapid evaluation of residual structural integrity is required in order to
provide quantifications that can aid the decision making process for emergency response. It is suggested that the effect
on global structural loads caused by any flooding of ship compartments should be incorporated into a structural analysis
when determining residual structural integrity.

A series of experiments are performed to investigate and quantify the effect of damage on vertical midship bending
moment. The model, constrained to heave and pitch, is subjected to both regular and irregular wave excitation in a
series of prescribed damage cases, for symmetrical damage at the keel of the ship. The impact of longitudinal damage
location and damage size is primarily investigated, and the effect of slow forward speed when compared with a
stationary model is considered in order to simulate a ship either being towed or moving under its own power to a safe
haven after damage has been experienced. This paper presents the methodology used in the experiments as well as some
initial results.

Results from the experiments will be used to validate tools being developed at University College London and The
University of Southampton.

1. INTRODUCTION Whilst the analysis of the structural loading on a

Ships operate in a harsh environment and experience
damage to their structures from a variety of sources (e.g.
collision, grounding, terrorist activity). Whilst the
occurrence of damage to ships is not a recent
phenomenon, its inclusion within design methodologies
and the development of computational tools that can be
used for decision guidance in emergency response
scenarios (e.g. salvage guidance for a ship in distress)
are. Following structural damage, it is important to be
able to assess a ships new structural integrity. This
involves quantifications of both structural loading and
structural capacity.
Damage can modify the structural capacity of a ship by
rendering structural elements discontinuous or
inefficient, or modifying load paths and stress
distributions. Analysis of the modified capacity of the
ship can be performed using a variety of methods, many
of which are also applicable to the intact ship. Examples
of such analyses can be found in [1], [2].
Damage below the waterline that results in flooding can
modify both local and global structural loads. Whilst it is
important to understand both of these elements of the
problem, especially as it could be the case that local
loads lead to progressive deterioration of structural
integrity over time, it is the global load problem which
has the most serious implication to the survivability of
the ship and the safety of the crew and passengers
onboard.
damaged, flooded ship in a seaway is similar to damaged
stability analysis (examples of which include [3],[4],[5]),
there are comparatively few published examples of
structural load calculation methods for damaged ships
and fewer data sets available to enable comparison of
global structural loads obtained from computational tools
with experiment data, for purposes of validation and
verification.
1.1 LITERATURE
A number of preceding experiments have examined
different aspects of the behaviour of a damaged, flooded,
ship.
Palazzi and de Kat [6] attempted to simulate the transient
response of a damaged vessel in roll. The simulation
used a potential flow code to solve for hydrodynamic
forces, and also incorporated both quasi-steady
floodwater behaviour and the airflow to and from the
damaged compartment. They found poor agreement
between their experiment and their simulation. By trial
and error they found that by halving the value of the
discharge coefficient Cd (from 0.58 to 0.3), a better
agreement was obtained during the initial few seconds of
the transient response. However, this was at the expense
of agreement over the remainder of the response. Their
experience indicated the potential shortcomings of a
quasi-steady approximation for particularly dynamic
flow through an orifice. However, it also indicated the
feasibility of carrying out experiments at scale to
measure model motions and floodwater behaviour.

Another example of comparisons of the computational
predictions to model experiment data for a damaged ship
are, for motions: [7] and [8], and for global loads: [9] and
[10].
Investigations were carried out on a single specific ship
(the Ro-Ro MS Dextra in a specific loading condition).
Model tests were carried out at 1/125 scale with a force
gauge to measure loading. Two damage cases were
applied; engine room damage (near midships) and
forepeak damage. In both cases the damage extent (size
of opening in the hull) was not varied. Nonlinear effects
were significant in horizontal modes of motion and
resonant roll motion but less remarkable in heave and
pitch modes of motion.
Whilst their damaged ship saw greater wave-induced
load magnitudes for engine room damage, the maximum
overall bending moment reduced. This was because for
the specific examined ship, flooding the engine
compartment changed the midship still water bending
moment from a high magnitude hogging moment to a
lower magnitude sagging moment. Forepeak damage did
not appear to have a significant impact on global load.
Both of the investigations described here were limited to
a small number of damage scenarios (sizes and locations
of damage). The computational analysis simplified
floodwater behaviour to the extent that the internal height
was assumed identical to the height of the incident wave
(and with no modification to the hulls hydrodynamic
coefficients). The model tests were done at a small scale
and the results showed a large amount of scatter. Whilst
they did imply that the presence of damage modified the
structural loading on the ship, the scatter meant that they
could not be used to elucidate which parameters and
physical phenomena are most significant in producing an
accurate estimation of a damaged ships structural
loading.
These studies therefore suggest that further work should
be carried out to try and understand what physical
phenomena are important to a damaged ship and how
best these should be approximated in order to simulate
damaged ship behaviour and obtain analysis appropriate
for engineering judgment in the context of ship design
and operation.
2. GLOBAL STRUCTURAL LOADS ON
DAMAGED SHIPS
An expression for the shear force and bending moment,
V, (modified from [11]) is:
Vj = S j + I j Rj E j Dj (1)
where j = 3, represents the component associated with
heave motion (vertical shear force) and j = 5 the
component associated with pitch motion (vertical
bending moment). The term Sj is the still water shear or
The Damaged Ship, London, UK
bending moment, Ij is the inertia term and Rj, Ej and Dj
are the external forces and moments; hydrostatic
restoring term, exciting term and hydrodynamic term
respectively.
One method for considering the force imparted by the
floodwater on the ship is to treat it as a time varying
mass (varying due to floodwater ingress and egress into
and from the damaged compartment). Under this
framework, the terms most effected by the occurrence of
damage are most likely to be S (as the ship attains a new
static equilibrium in the hypothetical still water
condition) and I, as the floodwater modifies the vessels
longitudinal weight distribution.
Under the above theoretical framework, any simulation
method would need to calculate the ingress and egress of
floodwater. This is where a quasi-steady method could be
used, as has been implemented for damaged stability
analysis.
The terms R, E and D will be modified for the length of
the ship that is damaged. And if the above simplification
(treating floodwater effects as a modification to an intact
ships mass distribution) is made then the magnitude of
the effects on R, E and D needs to be quantified to
establish whether they need to be considered within any
simulation method or whether their modification can be
treated as negligible.
To build on the existing literature in this subject, it was
therefore proposed that a series of experiment and
computational investigations could be undertaken in
order to elucidate the forces acting on a damaged ship
and their sensitivity to parameters describing the damage
(size and location).
3. SIMPLIFYING THE DAMAGED SHIP
One method of approaching the need to provide further
elucidation of the physical phenomena occurring on a
damaged ship is to simplify the problem. This can be
done artificially in towing tanks, which allow for
controlled experiments to be undertaken which constrain
some of the unknowns.
An intact ship in a seaway has six degrees of freedom
(DOF). Commonly, for calculating structural loads, these
are treated as two coupled sets of equations and the
degree of freedom surge is omitted. This simplifies the
problem to solving for vertical motions and loads (pitch
and heave) and lateral motions and loads (roll, sway and
yaw).
Relaxing the constraint that the floodwater in a damaged
ship can be thought of as a fixed mass adds additional
DOF. If the floodwater is approximated as behaving
quasi-statically (that the free surface is always at static
The Damaged Ship, London, UK
equilibrium) then this adds a further 3 DOF (floodwater
heave, pitch and roll).
A simplification applied to the intact ship (that vertical
and lateral motions and loads are uncoupled, [11]) could
also be applied to the damaged ship. Incorporating the
additional DOF of quasi-steady floodwater in this
framework would produce two coupled equations:
Vertical motions: vessel pitch and heave motion and
loads and floodwater pitch and heave motion and
loads
Lateral motions: vessel roll, sway and yaw motion
and floodwater roll motion
In strip theory, [11], the terms from (1) corresponding to
heave and pitch motions and loads are both obtained by
calculating the vertical forces on a strip of a ship.
Therefore, the most simplified and fundamental
experiment and computational simulation that could be
performed to look at a damaged ships physical
phemonena are for a single DOF rigid body motion
(heave) and single DOF floodwater motion (heave).
3.1 1 DOF FLOODWATER BEHAVIOUR IN
RESPONSE TO HEAVE EXCITATION
A set of experiments to interrogate the assumption that
the quantity of floodwater in a damaged compartment
can be considered constant over a wave period, is
described in [12]. These UCL experiments were based on
an experiment by Vugts [13].
The apparatus consisted of a prismatic hull spanning
UCLs wave tank, so that the waves generated, as the
body was forced in oscillation, were approximately two-
dimensional. The tank has a depth of 0.7 m and a length
of 20 m (At low frequency, this limits the number of
oscillations that will be unaffected by wave reflections).
linear oscillator
load cells
rose joint
beam
Figure 1. Experiment configuration for the forced
oscillation of a damaged body.
This experiment suggested that dynamic behaviour of
floodwater (ingress and egress) was not negligible when
calculating the force acting on a damaged section of a
ship, and additionally, its magnitude and phase of
oscillation, relative to the ships heave oscillation, was
not well represented by simplified theoretical models
(such as those using quasi-static floodwater behaviour
with ingress and egress found using hydraulic
approximations).
These experiments were limited to single DOF motion
and damage symmetrical about the centreline of the
model, intentionally simplifying the floodwaters modes
of excitation to provide an artificially constrained
experiment. Whilst work could be carried out to refine
the simulation method and assumptions of its
underpinning physics further, to understand the
applicability of the experiments results to real ships, the
assumptions implicit in the experiments simplification
also required justification.
3.2 2 DOF FLOODWATER BEHAVIOUR IN
RESPONSE TO HEAVE AND PITCH
EXCITATION
One of the key assumptions in the experiment in Section
3.1 was that the floodwater behaviour could be
approximated as quasi-steady and that violent free
surface behaviour (e.g. sloshing) could be considered
negligible both in terms of its influence on floodwater
ingress and egress and on the loads exerted on a damaged
ship.
Still focusing on the vertical motions and loads case, an
experiment was therefore designed for the UCL Ocean
Towing Tank to investigate the interaction between the
sloshing dynamics and the ingress and egress of
floodwater in a centrally located, damaged compartment
of a model constrained to heave and pitch motion. A
model was built, which was similar in shape and
construction to that described in Section 3.1, with a
9.6mm thick PVC half-pipe used as the submerged
section of the model and 20mm thick panels of plywood
used above the waterline. The principal dimensions of
the model were length = 2000 mm, beam = 200 mm,
draught = 100 mm.
stern optical connecting bow optical
sensor arm sensor
location of the wave hinges
video camera probes
Figure 2. Experiment configuration for measuring the
heave and pitch motions of the damaged model

Heave and pitch were measured using optical sensors
fixed onto a stationary carriage above the model. These
sensors simultaneously measured the changing vertical
distance of the bow and stern, as the model oscillated in
the wave. The free surface of the floodwater in the model
was recorded using wave probes and its behaviour was
observed using a video camera located inside the model,
adjacent to the damaged compartment. The model was
constrained to 2 DOF motion using a pair of hinges and a
26mm thick connecting arm which was made from an
aluminium honeycomb sheet. One hinge was fixed to the
model and the second was located on the stationary
carriage. The complete arrangement is shown in Figure
2.
The range of wave frequencies used was 3.2 to 7.85
rads-1, with 20 samples being taken at approximately
even intervals between these values for each
experimental scenario. These wave frequencies
correspond to wavelengths of between half the length to
three times the length of the model.
Ballast was used to ensure the draught and trim remained
constant for each of the 8 experimental (intact and
damage scenarios that were tested. This ensured that the
hydrodynamic and hydrostatic force components were,
with the exception of the presence of damage, consistent
with the intact model. The scenarios tested included full
length damage to the midship compartment at 20mm and
60mm width, as well as an intact ship and an intact
damaged ship. Experiments were performed for wave
excitations corresponding to 20mm and 30mm
amplitude.
Observations from the experiments demonstrate that the
ingress and egress of the floodwater through a section of
damage located at the keel significantly reduces the
severity of the floodwater motion when compared with
the floodwater in an intact flooded condition. Figures 3a,
3b and 3c illustrate the difference in the dynamics of the
floodwater for the same excitation wave amplitude
(30mm), at the same frequency of 5.5 rads-1. Each still
Figure 3a. Intact Flooded Scenario Figure 3b. 20mm Damage Width
The Damaged Ship, London, UK
picture showing the maximum free surface elevation for
a different experimental scenario. Figure 3a is the intact
flooded condition which saw extremely violent sloshing
at this frequency, suggesting that the frequency of
oscillation of the floodwater within the compartment had
coincided with the natural sloshing frequency, thereby
exciting the water to jump out of the compartment. The
subsequent figures o3b and 3c are for damage widths of
20mm and 60mm respectively. They showed the calm,
by comparison, ingress and egress of the floodwater with
no violent sloshing. This implies that these damage cases
have a damping effect on the floodwater motion, as no
significant longitudinal sloshing was noticed with either
of the damage scenarios at any of the investigated
frequencies.
These observations favour the assumption that the
floodwater longitudinal sloshing motion produces a
negligible effect on global structural loads (i.e. that
quasi-steady floodwater motion can be used to
approximate the behaviour in the example geometry and
subdivision investigated). Whilst it is qualitatively
apparent that in some damage cases the effect of the
ingress and egress of floodwater is to mitigate the
sloshing motion, analysis of the data was not performed
to examine whether the sloshing motion significantly
modified the ingress and egress of floodwater, relative to
the results obtained in the pure heave experiment on a
damaged strip (Section 3.1). However, this experiments
findings justified further measurement and investigation
into the interaction effects observed (between floodwater
motion within the compartment and floodwater ingress
and egress) and their consequence on motions and global
structural loads. This experiment also established the
credibility of carrying out experimental work into
damaged ship motions and loads at this scale.
4. EXPERIMENT DESIGN FOR DAMAGED
SHIP GLOBAL LOAD MEASUREMENT
Following on from the preliminary experiments, the
Figure 3c. 60mm Damage Width
The Damaged Ship, London, UK
setup described in Section 3.2 was then adapted to form
the basis of a set of experiments that were principally
focused on assessing the effect of a damage orifice on the
global structural loads experienced by a ship in a seaway.
Experiments were again performed in UCLs Ocean
Towing Tank.
The model described in Section 3.2 was cut in half
through a plane located at the longitudinal midpoint,
perpendicular to the models centreline. A hinge system
was fitted with the hinge point just above the intended
waterline, and offset slightly from the midship location.
Load cells (strain gauges) were then connected to ensure
no rigid body motion between the two hinged sections
and so as to provide a statically determinate
measurement of the bending moment at the hinge
position.
To ensure all of the structural load associated with
vertical bending passes through the strain gauges, a small
gap (~3 mm) is maintained between the two halves of the
model. In damage cases in which the midship
compartment is intact, this gap needed to be made
watertight. In cases where the central compartment is of
interest, portions of the gap were made watertight in
order to maintain the required shape of the damage
orifice. Latex strips were used to ensure that these
modifications did not carry any structural loads.
load
cell
wave probes
hinge point hinge arms
Figure 4. Diagram of hinge design showing pivot, load
cells and the wave probes used to measure free surface
behaviour.
By maintaining the same assumption that the floodwater
inside the damaged compartment behaves quasi-
statically, the motions of the free surface can be
simplified into two degrees of freedom; floodwater heave
and pitch. With the pertinent positioning of wave probes
within the damaged compartment, as shown in Figure 4,
the magnitude of these motions can be recorded
accurately.
4.1 PARAMETERS CONSIDERED IN HINGED
SHIP EXPERIMENT
For the initial experiments only pitch and heave motions
were considered due to the constraint provided by the
arm design described in Section 3.2. This meant that at
this stage only scenarios for damage symmetrical about
the centreline of the vessel will be considered.
Asymmetric damage could produce significant excitation
in roll, sway and yaw motions which could not be
quantified in this experimental setup.
The presence of Perspex bulkheads to provide
subdivision in the hinged ship model enable variation in
the longitudinal position of the damaged compartment.
For individual damage cases, damage is applied in the
form of a rectangular slot, symmetrical about the keel of
the model, cut out along the full length of a
compartment; consistent with the previous experimental
investigations.
Building on the results of previous experiments, one
parameter of interest in the damaged ship global load
experiments is the size of the damage orifice. Damage
orifice sizes are defined as ratios of the damage orifice
area to the compartment internal free surface area. Since
the extent of the damage orifice was equal to the length
of the compartment, this percentage is manifested as the
ratio of damage width to the compartment beam. The
percentages chosen for experimentation were 2.6%,
10.9%, 18.2%, 25.5% and 37.5%.
Plan View
subdivision
damage
orifice freely
venting
Section View
Figure 5. Schematic of Compartmentalised Prismatic
Model.
The described model was subjected to excitation from a
series of regular (sinusoidal) wave trains, and was
positioned so as to simulate head seas conditions.
Traditional bending moment calculations for an intact
ship indicate that peak loading is experienced when the
wave excitation frequency is such that its wavelength is
approximately equal to the length of the ship. With this
in mind, a suitable selection of regular wave frequencies
was made to include sufficient data points below and
above this region.
The presence of a flat bow gave rise to limitations with
regards to wave amplitude selection. Over a certain
threshold, generated waves would have a violent
interaction with the model bow.
Regular waves of frequencies ranging between 0.5Hz
(3.14 rads-1) and 1.22Hz (7.66 rads-1) in regular
increments of 0.0625Hz (or 1 16 Hz ) were used
 The Damaged Ship, London, UK

throughout all experiments, both intact and damaged.

Due to the practical limitations mentioned, two wave
amplitudes were selected; 20mm and 30mm.

5. EXPERIMENT METHOD VALIDATION

Experiment data for motions and loads was converted

into heave, pitch and bending moment RAOs, shown
here in Figures 6, 7 and 8 respectively. For the purposes
of validation, the measured motions and bending
moments are plotted alongside a strip theory computation
for a prismatic hullform with a specified weight
distribution (matching the dimensions and weight
distribution of the experiment model). The strip theory
used in these plots is described in [14].

The plotted results show excellent experimental Figure 7: Intact pitch response compared with modified
repeatability for multiple runs of the same intact case. strip theory prediction.
The strip theory prediction for intact motions
demonstrates that there is good agreement between the
theoretical model and experiment results over the range
of frequency and excitation amplitude investigated.

Experimental results for vertical bending moment at

midships also show good correlation with predictions
from the modified strip theory code.

Figure 8: Intact bending moment response compared

with modified strip theory prediction.

6. SOME INITIAL RESULTS

To date, experiments have been carried out for cases

Figure 6: Intact heave response compared with modified
strip theory prediction.
where damage has been applied to only the central
compartment of the prismatic model. This choice of
damage location was favoured for preliminary
investigations because midships damage was estimated
through computational simulation to have significant
effects on the vertical global loads experienced by the
model.

6.1 EXPERIMENTAL RESULTS FOR MOTIONS

AND LOADS

Damage orifice widths of 29.1mm and 60mm

(corresponding to damage orifice areas of 18.2% and
37.5% of the internal free surface area respectively) were
selected to initially provide a broad range of results with
regards to investigating the effect of damage orifice size
The Damaged Ship, London, UK

on floodwater dynamics and, subsequently, vertical

bending moment.

Validated heave, pitch and vertical bending moment

responses for the two initial damage cases were
compared with the obtained intact results, so as to
observe any effect that gave rise to a change in global
loads. Results of intact, 29.1mm and 60mm damage
orifice widths for heave, pitch and vertical bending
moment are displayed in Figures 9, 10 and 11
respectively, in the form of mean responses. Error bars
denote the experimental values variability over repeated
runs. They display a range above and below mean data
points of 2 standard deviations.

With regards to damaged model motions, the overlapping

of the calculated spread of data for the three curves at a
given data point implies that there is no significant
Figure 10. Comparison of damaged and intact
impact on model motions with damage orifice size.
experimental pitch results.
Given the present dataset, the coincidental change in
motion response with an increase in orifice size,
particularly pertinent at low frequencies, cannot be
clearly distinguished or attributed to a hydrodynamic
phenomenon.

The obtained bending moments for different damage

scenarios indicate that the size of the damage orifice, and
the subsequent possible alteration to floodwater
dynamics, has a significant impact on global vertical
loads. With the inclusion of experimental scatter, results
for each damage case show both noteworthy and
differing departures from values observed in the intact
case. At frequencies corresponding to wavelength equal
to ships length, there is an observed peak bending
moment relief of 5% for the case of small (29.1mm)
damage and approximately 20% for the larger damage
orifice size (60mm).

Figure 11. Comparison of damaged and intact

experimental vertical bending moment results.

6.2 PRELIMINARY INVESTIGATION OF

FLOODWATER DYNAMICS

A preliminary investigation has been undertaken to

Figure 9. Comparison of
experimental heave results.
observe the behaviour of the internal free surface of the
models damaged compartment and assess whether a
change in damage orifice size has any significant impact.
If, as previously described, the internal free surface is
assumed to behave like a flat lid (the instantaneously
quasi-static assumption), the appropriately positioned
wave probes fixed to the model inside the damaged
compartment make it possible to ascertain values for
floodwater heave and pitch motions relative to model
damaged and intact heave and pitch, the latter being calculated using two
wave probe readings and simple trigonometry.

Relative floodwater heave and pitch has been recorded

for both damage cases and post-processed to produce

transfer functions, taking into account experimental
variations between runs.
Figure 12. Comparison of experimental relative
floodwater heave for 29.1mm and 60mm damage orifice
widths.
Figure 13. Comparison of experimental relative
floodwater pitch for 29.1mm and 60mm damage orifice
widths.
The plotted results show that the change in damage
orifice size has a considerable effect on floodwater
dynamics, with variations in the magnitudes of response
distinguishable from scatter in the experimental data.
With regards to floodwater heave relative to the heave of
the model shown in Figure 12, a larger damage orifice
size shows a greater excursion from the internal still
water free surface of the floodwater. The relative
floodwater pitch results shown in Figure 13 display an
overall difference in behaviour between damage cases.
In the larger damage case, results suggest that in the
region of wavelength/ships length matching, the
floodwater pitch exceeds the wave slope of the external
free surface in terms of magnitude; a characteristic of
behaviour that is not observed in the smaller damage
case.
The Damaged Ship, London, UK
Preliminary observations have only taken into account
the comparison of the magnitudes of relative floodwater
heave and pitch behaviour. Further analysis will be
undertaken to observe the relationship between
floodwater heave (relative to model heave) and models
relative vertical motion (heave motion relative to the
external free surface elevation), with similar
methodology being applied to floodwater pitch.
7. CONCLUDING REMARKS
The study of damaged ship motions and loads is
complicated by the large number of degrees of freedom
in the problem (DOF for the model and the floodwater)
and the fluid structure interaction that is associated with
floodwater ingress, egress and its motion inside a
damaged compartment. A naval architects conventional
assumptions for calculating hydrodynamic forces acting
on damaged ship need to be revisited before
computational methods that employ simplifications of
the physics to render the problems soluble (appropriate
for engineering judgment) can be used with confidence.
This paper describes a progression of experiments
performed at UCL to elucidate some of the fundamental
physical phenomena that are behind the fluid structure
interaction and ultimately the global structural loads and
motions of a damaged ship in waves.
Experimental results presented in this paper represent the
initial findings from the latest series of experiments.
These findings build on previous research findings and
support the hypothesis that the inclusion of floodwater
dynamics in the calculation of global vertical loads is
important and can have a significant effect on their
magnitude.
Whilst the current results illustrate the effect of midships
damage to a freely vented compartment only, the
experiments will progress to include a change in
longitudinal location of the damaged compartment as
well as the effect of slow forward speed (to simulate a
ship being towed to a safe haven) on global loads.
8. ACKNOWLEDGEMENTS
The authors are continuously grateful to the UK Ministry
of Defence and Lloyds Register of Shipping for their
ongoing support of this research, and to the University of
Southampton for their collaboration.
9. REFERENCES
1. Khan I.A., Das P.K. 2007. Residual strength and
survivability of ships under combined vertical and
horizontal bending. Advancements in Marine Structures,
197-206. Leiden: Balkema
The Damaged Ship, London, UK

10. AUTHORS BIOGRAPHIEIES

2. Arason M., Kent J., Cockman J. 2005. Validation of
ultimate strength assessment methodology applied in
Paramarine / Seagoing Paramarine. QinetiQ/05/00515
3. Sen P., Konstantinidis C. 1987. A time simulation
approach to the assessment of damage survivability of
ro/ro cargo ships. Transactions of the Society of Naval
Architects and Marine Engineers, 95: 337-355.
Daniel Fone is a postgraduate research student in the
Department of Mechanical Engineering at University
College London, and is currently working towards a PhD
in experimental hydrodynamic loading on damaged
ships. He joined the UK MoD damaged ship research
group in 2009 and is currently their principal
experimental researcher.

4. Turan O. 1992. Dynamic Stability Assessment of
damaged ships by time-domain simulation. PhD Thesis,
University of Strathclyde, UK
5. Hearn G.E., Lafforgue D., Perdriset E., Saydan D.
2008. The hydrodynamic and dynamic motion analysis
of a damaged ship. Transactions of Royal Institute of
Naval Architects. 150:75-99
6. Palazzi L., de Kat J. 2002. Model experiments and
simulations of a damaged ship with air-flow taken into
account. Proceedings of the 6th International Ship
Stability Workshop, Webb Institute
Tristan Smith is a Research Associate in the UCL
Energy Institute. His PhD in Naval Architecture was
titled wave loading on damaged ships. He joined UCL
after starting his career working on warship structural
design and analysis in the UK Ministry of Defence. In
2009, he was awarded the RINA-LRET Ship Safety
Award for his work on damaged ship structures.
Joe Borg is a Trainee Marine Surveyor for Lloyd's
Register. He joined their graduate training programme in
2009 after completing his MSc in Naval Architecture at
UCL. His MSc thesis was titled motions of a damaged
ship.

7. Chan H.S., Incecik A., Atlar M. 2002. Large-
amplitude motion responses of a Ro-Ro ship to regular
oblique waves in intact and damaged conditions. Journal
of Marine Science and Technology, 7: 91-99
8. Korkut H.S., Atlar M., Incecik A. 2004. An
experimental study of motion behaviour with an intact
and damaged Ro-Ro ship model. Ocean Engineering, 31:
483-512
Kevin Drake is a Senior Lecturer in the Department of
Mechanical Engineering at University College London.
His industrial experience includes the design, fabrication
and installation of various structures for the offshore oil
and gas industry. In 2000 he was an expert witness for
the Re-Opened Formal Investigation into the loss of the
MV Derbyshire.

9. Chan H.S., Incecik A., Atlar M. 2003. Global wave

loads on intact and damaged Ro-Ro ships in regular
oblique waves. Marine Structures, 16: 323-344

10. Korkut H.S., Atlar M., Incecik A. 2005. An

experimental study of global loads acting on an intact
and damaged Ro-Ro ship model. Ocean Engineering, 32:
1370-1403

11. Salvessen N., Tuck E.O., Faltinsen O. 1970. Ship

motions and sea loads. Transactions of the Society of
Naval Architects and Marine Engineers, 78: 250-287

12. Smith T.W.P., Drake K.R., Wrobel P. 2009.

Experiments on a damaged ship section. Advancements
in Marine Structures 2, Leiden: Balkema

13. Vugts Ir.J.H. 1968. The hydrodynamic coefficients

for swaying, heaving and rolling cylinders in a free
surface. Report 112S. Netherlands Ship Research Centre

14. Smith T.W.P., Drake K.R., Rusling S. 2007.

Investigation of the variation of loads experienced by a
damaged ship in waves. Advancements in Marine
Structures, 89-98. Leiden: Balkema
The Damaged Ship, London, UK

SIMULATIONS OF MOTIONS OF A DAMAGED SHIP IN REGULAR WAVES

A C Gaillard, G X Wu and P Wrobel, University College London, UK

SUMMARY

The present analysis is to investigate the sea-keeping behaviour of a damaged ship under wave loads, based on the
velocity potential theory. Computations are performed in different flooding situations, corresponding to different
draught, trim and heel. For each of them, hull motions are obtained in terms of Response Amplitude Operators, using a
boundary element based three dimensional (3D) code in the frequency domain. The results for front, beam and stern
quartering waves are provided and discussions are made through comparison with the behaviour of the intact ship.
Analysis of the results indicates that the effect of damage on the ship motion responses depends on the direction of the
wave and its frequency. At different draught and trim angle motions of a damaged ship are close to those of an intact
ship, within those ranges over which simulations are made. Heel variation highlights that it has much stronger effect on
the motion responses.

NOMENCLATURE 1. INTRODUCTION

A Wave amplitude (m)
Aij Added-mass (kg for i,j=1,2,3
or kg.m for i,j=4,5,6 and kg.m for all
other cases)
Aij Dimensionless added mass
Bij Damping (kg.s-1 for i,j= 1,2,3 and
kg.m.s-1 for i,j=4,5,6 and kg.m.s-1 for
all other cases)
Bij Dimensionless damping coefficient
Cij Restoration matrix (kg.s-2 for i,j=1,2,3
and kg.m.s-1 for i,j=4,5,6 and kg.ms-1
for all other cases)
G Green function
g Gravitational acceleration (m.s-2)
h Water depth (m)
k Wave number (m-1)
Wave frequency (rad. s-1)
Velocity potential (m2.s-1)
Density of water (kg m-3)
s Hull surface (m2)
V Velocity (m.s-1)
1, 2, 3 Surge, sway, heave amplitudes (m)
4, 5, 6 Roll, pitch, yaw amplitudes (rad.)
i Dimensionless motion amplitude
D Draught in damage scenarios as
detailed in table 2
H Heel in damage scenarios as detailed in
table 4
LCG Longitudinal centre of gravity (m)
LOA Overall length (m)
LWL Waterline length (m)
TCG Transversal centre of gravity (m)
RAO Response Amplitude Operator
T Trim in damage scenarios as detailed in
table 3
When a ship is damaged with a hole open to external
water, it is most likely that she will lose stability, suffer
from poor sea-keeping performance and be more prone
to structural failure. Assessment of these issues for a
damaged ship is therefore considered to be one of the key
issues in the future development of safe ships. There are
several main causes of damage to a ship, including
collision, grounding and excessive loading.
Investigations have been done to understand the nature of
these causes, their consequences, and to develop
measures to reduce the possibility which can lead to
those accidents. A typical example of damage on a ship
is a crack or a hole on the hull. This would lead to
flooding of compartments, changing the hull equilibrium
and affecting the sea-keeping performance. In extreme
cases, this may lead to capsizing of the ship due to loss of
stability along with spread of the structural damage
across the hull.
The major difference between the behaviour of a
damaged ship with a flooded compartment and the sea-
keeping of an intact ship with a liquid filled internal tank,
such as an LNG, is the direct coupling between the
internal and the external flows. Water motion inside and
outside the ship hull are linked through the hole.
Therefore the internal and external flows need to be
solved together.
Earlier reported research has investigated the
survivability of the ship in a damaged condition.
Numerical models intended for predicting intact and
damaged ship behaviour in waves have been developed,
as discussed by a benchmark study in the programme of
Specialist Committee on Stability at the International
Towing Tank Conference [1]. The focus has been on the
effect of transient flooding on roll motions leading to
capsize of the ship [2,3]. These studies are based on time
domain codes using flat horizontal free surface model
inside damaged compartments. In the work of De Kat
and Peters [4] and Corrignan and Arias [6], a time
domain sea-keeping code is used, where diffraction
 The Damaged Ship, London, UK

forces are obtained using the potential flow theory. Some
non linear effects from hydrostatic force and the
incoming wave are taken into account. Progressive
flooding is modelled based on the Bernoulli equation.
The water level inside damaged compartments is
assumed to be horizontal at all time, neglecting the effect
of sloshing, as in Santos and Soares [5]. The significance
of sloshing on the motion of a damaged ship has been
shown by Cho et al. [7] using CFD and quasi-dynamic
method (lump-spring) to model flooding water. In the
work of Fols and Rizzuto [9], a linear sea-keeping code
is used to calculate the RAOs, vertical and horizontal
bending moment of a damaged ship and the numerical
results are compared with those of an intact ship. Korkut
et al. [8] performed systematic motion tests for one
damage scenario in regular waves at various wave
headings and wave lengths. The six degrees of freedom
(DOF) motion responses are measured. The results for
the intact and damaged ship are compared.
Three dimensional (3D) potential flow theory is used in
our research to predict motions of a damaged ship, taking
into account sloshing inside the damaged compartment
and its coupling with the external flow. Systematic
simulations have been undertaken. Parametric changes in
heel, trim and draft are considered and their influence on
hydrodynamic forces and RAOs are analysed. Results are
compared with those of an intact ship, while varying the
damaged position and the internal configuration of the
ship. Various sea states and wave directions are also
considered.
2. METHOD
The 3D potential flow code WAMIT is used to calculate
the RAOs of a containership model under different
conditions of trim, heel and draught in scenarios with
flooded compartment. The sea-keeping performance
analysis is undertaken in the frequency domain.
2.1 MATHEMATICAL TOOLS
A body in the fluid can be represented by suitable
combinations of sources and sinks on each panel of the
discretised wetted surface of the body. When a large
number of panels are used, the complex shape of a ship
hull can be accurately represented. For the periodic
motion with frequency , the total velocity can be
written as
( x, y, z , t ) = Re[ ( x, y, z )e it ]
With the help of the Greens function G , one can write
[11]:
0
G
s nG G nG ds = 2 ( x, y, z )
4 ( x, y, z )
for (x, y, z) being inside, on or outside the hull surface,
where n is the normal of the body surface. In WAMIT
this integral equation is solved by assuming that the
unknown is constant over each panel.
2.1 (c) Free Surface
Both kinematic and dynamic boundary conditions are
used to define the free surface between air and sea. The
linearised form of the free-surface boundary condition
can be written as
z g = 0
on z = 0 , where g is the acceleration due to gravity.
The total velocity can be expanded as
6
= j j + A( 0 + 7 )
j =1
where j , j = 1,...,6 , correspond to the six degrees of
freedom of the body motion, is the motion amplitude,
j

0 and 7 are the incident and diffracted wave potential,

2.1 (a) Reference System
and A is the amplitude of the incoming wave.
A Cartesian coordinate system o xyz is defined with
The incident velocity potential due to unit wave
z-axis pointing vertically upwards and origin of the amplitude can be written as
reference system located at the stern of the ship on the g cosh k ( z + h) ikx
waterline. x-axis is from stern to bow and y-axis from 0 = e
starboard to port. Body motions and forces are defined i cosh kh
relative to this coordinate system (cf. Figure 1) therefore where k is the wave number and is the root of
pitch motions are relative to stern. 2
= k tanh kh
g
2.1 (b) Velocity Potential and Green Function
and h is the water depth.
Under the assumption of an ideal fluid, the irrotational
flow around a ship can be represented by a velocity 2.1 (d) Thin Walls
G G
potential such as V = with V being the fluid
The integral equation is usually converted into the source
velocity. The potential satisfies Laplace equation
distribution method. The method is normally effective.
= 0 in the fluid domain.
The Damaged Ship, London, UK

However, when water is on both sides of a plate, the

panel size should be of at least the same order as the
thickness of the plate, or smaller to ensure the accuracy
of the calculated result. This makes computation
impractical when the plate is too thin. In fact when the
thickness of the plate tends to zero, the source Figure 1. Containership hull
distribution method becomes singular. In this case, the
dipole distribution should be used. 2.2 (b) Damaged Scenarios
2.1 (e) Body Response in Waves The damaged scenarios correspond to water flooding into
a compartment due to bottom damage in the hull at full
Forces and moments on the body surface are given by the load conditions. This could be associated with a
expressions [10]: grounding event. To evaluate the performance of the
d damaged ship at different conditions, three scenarios
F =
dt nds
S
have been set up independently.
G d G G
M = (r n )ds In the first scenario, damage occurs in the bottom of the
dt S hull, aligned with the centre of gravity leading to a
where is the density of water. vertical sinkage of the ship. The size of the compartment
is altered to allow the variation of the amount of flooded
water. No heel or trim angle is created, and only draught
Combining this with Newtons law, the equation for the
is increased. Details are given in table 2.
body motion can be written as
[ ( M ]= A X
6

j ij
+ A ) + i B + C
ij ij ij i
Table 2. Flooding scenarios Draught influence
j =1 Condition Intact D1 D2 D3 D4 D5
Displacement [t] 19816 19916 20316 20816 21316 21816
where M ij is the body mass and inertia matrix, A ij Flooded mass [t] 0 100 500 1000 1500 2000
Draft [m] 8.43 8.47 8.59 8.74 8.88 9.04
is the added mass matrix, B ij is the damping Then a second scenario is set up, where a flooded
matrix, C ij represent the restoring coefficient and X i compartment (equivalent to an additional flooded mass
of 2000 tonnes) is moved forward and backward from the
the exciting forces. centre of gravity such as to create a trim angle. Details of
centre of gravity position and trim are given in Table 3.
The added mass and damping coefficients can be
obtained from Table 3. Flooding scenarios Trim angle
i
Aij Bij = ni j dS Condition Intact T1 T2 T3 T4
S

while the exciting force from LCG [m] 83.527 83.71 83.985 84.44 83.069
Trim [deg] 0 0.06 0.11 0.20 -0.10
X i = ni (0 + 7 )ds
s
G K G Finally in the third scenario the flooded compartment is
where n = ( n1 , n 2 , n3 ) and n r = ( n4 , n5 , n6 ) with moved sideways to port side to study the influence of
G
r being the position vector from the body coordinates heel. Details of variations of heeling angle and the centre
system. of gravity in the transverse direction are provided in
Table 4.
2.2 DAMAGED VESSEL SCENARIOS
Table 4. Flooding scenarios Heel angle
2.2 (a) The damaged Vessel Condition Intact (H0) H1 H2 H3
TCG [m] 0 0.183 0.367 0.55
The case selected for the present analysis is a Heel [deg] 0 -2.3 -4.8 -7.3
containership with main dimensions detailed in Table 1.
2.3 COMPUTATIONAL PROCEDURE
Table 1. Main dimensions of the ship (full load, intact).
LOA 165m 2.3 (a) Description of the Mesh Geometry
LWL 151m
Breadth 24.38m The vessel geometry is discretised by a 3D surface mesh.
Draught 8.30m The number of panels influences both the accuracy of the
Displacement 21143t results and the computational time. Therefore a mesh
LCG 84.6m density influence analysis has been performed with
priority given to accuracy of the results. In each damaged
 The Damaged Ship, London, UK

case the vessel has been discretised with 8000-9000 cells resonant peak magnitude is 10% bigger than in the intact
per semi-hull, which is found to give converged results case D0 as can be seen in figure 2.
and requires 10-15 hour computational time per
simulation. (CPU 2.20 GHz, RAM 2Go).

2.3 (b) Environmental Conditions

Calculations are made in each case under both intact and

damaged conditions at zero speed. Simulations are
carried out for wave frequencies ranging from 0.05 rad./s
to 2.05 rad./s to include the case of wave length close to
the ship length. Three heading angles: 180, 270 and
315 are considered, corresponding to head, beam and
stern quartering seas coming from port side (damaged
side) (cf. Table 5.)

Table 5. Calculation conditions Figure 2. Scenario D Heave RAO Front waves

Speed 0 kts
Wave frequencies 0.05 -2.05 rad./s The curve for the D5 case shows there is an additional
Wave heading 180, 270, 315 smaller resonant peak at a frequency of around 1.5 rad/s.

Pitch RAO also shows a resonant peak at a frequency

3. RESULTS AND COMMENTS close to 0.65 rad/s. Pitch magnitude increases when the
amount of flooded water increases for front wave
Response Amplitude Operators (RAOs) in the six headings as shown in figure 3.
degrees of freedom, added mass and damping coefficient
are obtained and some of them are compared and
analysed for the intact and damaged containership in the
scenarios of three different damaged conditions.

The RAOs of body motions are non-dimensionalised

such as Ln where n=0 for i=1,2,3 corresponding

= i
i A
to translational motions and n=1 for i=4,5,6
corresponding to rotational motions.

Each scenario is analysed independently: first the draught

increase, then the effect of trimming and finally heeling.

3.1 DRAUGHT VARIATION Figure 3. Scenario D Pitch RAO Front waves

From scenarios D1 to D5, the amount of flooded water is
raised progressively from 0 to 2000 tonnes (equivalent to
a 10% increase in weight compared with the intact ship
at full load condition) leading to a maximum of 7%
increase in draught.
Emphasis is on the analysis of heave, pitch and roll
motions where major variations between intact and
damaged conditions are observed.
3.1 (a) Head seas
3.1 (b) Beam seas
In beam seas, roll motions are important. Roll motion
RAOs present a resonant frequency around 0.85 rad/s.
Another peak is located in the frequency range 1.50-1.60
rad/s as can be seen on figure 4.
Around 0.85 rad/s the magnitude of the response
increases with the amount of flooded water (+20%
between D5 and D1).

Heave motion responses of the damaged model display

similar trends as the intact model. A peak in motions
responses is present at a frequency of 0.55 rad/s
corresponding to wavelength close to the ship length.
Magnitudes of the responses in the intact condition
appear to be slightly smaller than those in the damaged
condition around the resonant peak. Damage case D5
The Damaged Ship, London, UK

Figure 4. Scenario D Roll RAO Beam waves Figure 6. Scenario T Roll RAO Beam waves

As expected there is no roll motion when waves are For all trim cases heave amplitude at the resonant peak is
coming in front of the ship, because of port and starboard different, varying from 0.06 for stern down trim (T4) to
symmetry. 1.2 for bow down trim (T3). (cf. Figure 7).

3.2 TRIM VARIATION

3.2 (a) Front Waves

In front waves, trends of damaged ship motions are quite

similar to those of the intact ship. Pitch motion shows a
resonant peak at around = 0.6 rad/s (wavelength close
to the ship length). For a frequency range of 0.8 to 1.0
rad/s, pitch RAOs show small differences from the intact
motion depending on the trimming angle. When
trimming bow down, the pitch RAO magnitude increases
although the pitch resonance maximum value at 0.55
rad/s is almost the same for all cases (see Figure 5.)
Figure 7. Scenario T Heave RAO Beam waves.

3.2 (c) Stern Quartering Waves

In stern quartering waves, trends of heave motions are

similar for all the damaged cases. Small differences can
be seen in the maximum amplitude value reached at
resonant frequencies. Maximum difference is seen for
heave motion at a frequency of 0.85 rad/s where the bow
down trim seems to reduce the heave response (-15%
from T0 to T3) (cf figure 8)

Figure 5. Scenario T Pitch RAO Front waves

3.2 (b) Beam Waves

In beam waves, and for all damage cases, heave and roll
motion RAOs shows a resonant frequency around 0.85-
1.05 rad/s. No significant difference can be seen between
the different cases in the roll RAOs. Small differences
appear around a frequency of 1.45 rad/s. (Figure 6.).

Figure 8. Scenario H Heave RAO Stern quartering

waves.

3.3 HEEL VARIATION
Two types of trends can be seen in the curves for heeling.
Results for 0 and 2.8 of heeling angle (damage cases
H0 and H1) are quite similar in trends. The damaged
cases H2 and H3 behave differently.
3.3 (a) Front Waves
In damage case H3 , the boat is heeled by 7.3 leading to
important changes in the wetted surface shape and more
important motion changes. In roll motions which exist
only for front waves because there is no longer port and
starboard symmetry, for the damage case H3, a resonant
peak is present at a wave frequency of 0.55 rad/s (cf
figure 9) which corresponds to a wave length close the
ship length.
Figure 9: Scenario H Roll RAO Front waves
3.3 (b) Beam Waves
A minimum heeling angle is required to see visible
effects on motions heave RAOs. Two types of response
can be observed. Results for 0 and 2.8 of heeling angle
(Damage cases H0 and H1) are quite similar in trends to
those of an intact ship. A second type of trend can be
seen for damage scenarios H2 and H3. A difference in
resonant magnitude can be observed in heave with values
smaller for the damaged and heeled ship than for the
intact one (Figure 10).
Figure 10. Scenario H Heave RAO Beam waves
The Damaged Ship, London, UK
In figure 10, one can see that in beam waves with large
heeling angle no resonant frequency can be seen for
cases H2 and H3.
3.3 (c) Stern Quartering Waves
Damage scenario H3 (heel angle of 7.3 degree) leads to
peak response with higher amplitude in front waves and
stern quartering waves for roll.
Figure 11: Scenario H Roll RAO Stern quartering
waves
Similar trends of roll motions can be observed in stern
quartering waves as in front waves (cf figure 9 & 11).
4. CONCLUSIONS
Computations for a container ship in various flooding
conditions have been made, corresponding to different
trim, draught and heel relative to the intact ship.
The response amplitude operators have been used to
assess the influence of the three parameters on the
motion of the damaged ship.
The sinkage of the ship with increase of draught leads to
similar trends of heave, pitch and roll motions as
compared to those of the intact ship. The amplitude of
motions is changed with respect to the amount of draught
increase. Roll motions response increases with draught of
the ship. For roll, heave and pitch RAO, the damaged
resonant magnitude is higher than in the intact case. This
varies depending on the wave direction and the wave
frequency.
When encountering front waves, trimming of the ship has
no important effect on pitch and roll RAOs. The same
trends and magnitude can be observed for all trimming
cases as for the intact ship. Small differences can be
observed in heave and pitch on the slope of the curves.
Only heave in beam waves has important differences
between bow down and stern down trim (Figure 6).
Heel seems to have more influence on motions of the
damaged ship. Two types of trends are observable. For
The Damaged Ship, London, UK
small heeling angles (0- 3) heave and roll motions are
quite similar to those of an intact ship. On the other hand
when the heeling angle becomes larger (8) a resonant
peak appears in roll motions in front and stern quartering
waves.
Further research is required to investigate the influence
of these parameters on the motions of a damaged ship.
Refinement around the transitional heeling angle needs to
be done along with realistic internal design of the ship
with more than one compartment.
5. ACKNOWLEDGEMENTS
This work is supported by The Damaged Research Group
funded by the Ministry of Defence and Lloyds Register
of Shipping, to which the authors are most grateful.
6. REFERENCES
1. VASSALOS, D., The specialist committee on
prediction of extreme ships motions and capsizing,
23rd International Towing Tank Conference, 2002.
2. IKEDA, Y., MA, Y., An experimental study on large
roll motion in intermediate stage of flooding due to
sudden ingress of water, Proceedings of 7th
International conference on stability of ships and
ocean vehicles, 2000.
3. HUANQUI, G., WANAN, S,. NAN, X.,
Experimental study on the damage stability of
containership, Proceedings of 7th International
conference on stability of ships and ocean vehicles,
2000.
4. DE KAT, J., PETERS, A., Model experiments and
simulation of a damaged frigate, IMAM Congress,
2002
5. SANTOS, S., GUEDES SOARES, C., Numerical
assessment of factors affecting the survivability of
damaged ro-ro ships in waves, Ocean Engineering
36, 2009.
6. CORRIGNAN, P., ARIAS, A., Flooding simulations
of ITTC and SAFEDOR benchmark tests cases using
CRS Shipsurv software, Proceedings of the 11th
International Ship Stability Workshop, 2010.
7. CHO, S., SUNG, H, HONG, S., NAM, B,
YOUNGSIK, K, Study on the motions and flooding
process of a damaged ship in waves, Proceedings of
the 11th International Ship Stability Workshop, 2010.
8. KORKUT, E., ATLAR, M., INCECIK, A., An
experimental study of motion behaviour with an
intact and damaged Ro-Ro ship model, Ocean
Engineering 31, 2004.
9. FOLS, L., RIZZUTO, E., Wave induced global
loads for a damaged vessel, Advancement in Marine
Structures, 2007.
10. NEWMAN, J., Marine Hydrodynamics, The
Massachusetts Institute of technology Press, 1977.
11. WEHAUSEN, J., LAITONE, E., Surface waves,
Encyclopedia of Physics, IX, 446-636, 1960.
7. AUTHORS BIOGRAPHIES
Anne Gaillard is a postgraduate research student at
University College London, and is currently working
towards a PhD in sea-keeping behaviour of a damaged
ship. She joined the UK MoD damaged ship research
group in 2010.
Guo Xiong Wu holds the current position of Professor at
University College London. He is also the chair of the
joint LRET (Lloyds Register Educational Trust) research
centre involving University College London, Shanghai
Jiao Tong University and Harbin Engineering University.
His research covers a wide range of problems in naval
architecture, offshore, deep water and coastal
engineering.
Paul Wrobel holds the current position of Course
director in the Marine Engineering and Naval
Architecture department at University College London.
He is member of the Marine Engineering research group
and Professor of Naval Architecture.
The Damaged Ship, London, UK

SELECTING THE SAFEST OPTION AFTER DAMAGE A TOOL FOR DECISION

SUPPORT
A A Martin, QinetiQ Rosyth, UK

SUMMARY

When a ship suffers a collision or grounding that results in significant structural damage, the owner must rapidly decide
whether the vessel is structurally sound enough to continue, whether to wait for assistance, or to plot courses to suitable
ports for repair. This was the situation faced by the UK MoD in July 2002 when HMS Nottingham struck a rock off the
coast of Lord Howe Island. This accident spurred research into a tool able to inform the decision making process, under
a programme known as DALAS.

DALAS has produced a capability to take decisions based on the relative risks of a selection of voyage options. It has
been implemented in the GRC code PARAMARINETM and has been issued to the UK MoD for use in emergency
response. It calculates loads derived from seakeeping tools, weather forecasts, vessel routing and flooded conditions.
The remaining structural capacity is calculated using the NS94 code, a Lloyds Register RSA 2 method. DALAS
represents a step change in capability compared to previous static-balance methods, while work is progressing to address
known limitations.

This paper describes the background to, and development of the DALAS tool, together with examples of its use to select
the safest option after a ship has suffered structural damage.

1. INTRODUCTION

Imagine the situation; you are the master of a 5000 tonne
ship, crossing the Atlantic, when, 500 miles from port,
your ship strikes a submerged object, washed overboard
from another vessel. The damage is immediate and
substantial. A large hole is opened in the structure with
torn side shell and bilge plating. Water pours in and the
ship begins to settle lower. Any available crew rush to
limit the progression of damage and pumps are started.
Within the first half hour, the situation becomes clearer
and the flooding does not progress further. However, you
are now faced with one question and many choices; what
is the safest way of proceeding?
days time; what are the relative risks of trying to drive
the damaged hull faster to arrive ahead of the storm?
In order to take the safest option, a way must be found to
link the predicted weather, the course, the speed of the
ship, the damaged condition and the likely wave induced
stresses for each possible solution. Advice should be
available within the shortest possible time, preferably
within a few hours. Advice should be based on quantified
assessment and there should be an appreciation of the
risks involved in any particular action.

If the flooding can be contained, one of the major threats
to the vessel will come from the loss of structural
strength. The ship is not about to break up in the next
hour, but it is a long way to a port and the weather will
change.
This paper intends to describe the results of a research
programme known as DALAS (Damage and Load
Assessment of Ships,) sponsored by the UK MoD, which
aimed to address this situation.
2. THE PROBLEM
3. THE RESEARCH
3.1 AIMS
The problem was approached primarily through an
improved understanding of the fluid loads. Efforts were
made to build on existing research and through continued
development of established tools. The UK MoD has
conducted extensive research over a number of years into
wave loading of its ships. Many are strain gauged and
analysis of the vertical bending stresses over many years
has been used both to monitor individual ships and to
create design rules for subsequent generations.

The key problem is that the master, or the ship owner,
has a large number of different options for proceeding.
Perhaps 5 ports are equally close for repair of the damage,
but each will require a different course. It is possible that
a more distant port is preferable, for security, or facilities,
but this would require a longer transit. Perhaps there is a
weather window; a storm is predicted on the course in 4
The large database of strain measurements has also been
used to benchmark the accuracy of different wave
loading software, one of the aims of the DALAS
programme. Several different software tools were tested
and the Det Norske Veritas (DNV) code WASIM was
selected as the preffered option for future wave loading
research.

This was paralleled by a search for software for use in
emergency response. The QinetiQ-GRC
PARAMARINETM code was selected for this purpose for
several reasons. It has a broad-based MoD user
community, an established library of ship models, a
structural strength module and already forms part of the
MoD emergency response process. In addition there has
been a history of development of the code for structural
assessment within QinetiQ and its potential for further
enhancement was understood . The DALAS programme
then focussed on improving the definition of fluid loads
in PARAMARINETM .
3.2 LOADS
It is traditional to assume that the load on the ship can be
broken down into still water and wave induced loads.
The still water load on an intact vessel should be known
with a reasonable degree of accuracy if the record of
tank-states and stores or cargo distribution is kept up to
date by the crew. This is usually a deterministic
calculation but there is considerable research to indicate
the variability in the still water load calculation [1] for an
intact vessel which can be used if a fully probabilistic
assessment of loads is to be conducted. In many cases the
deterministic approach is acceptable because the
uncertainities in the still water load are relatively small.
However, the situation is not the same when the ship
floods. Smith, Drake and Wrobel [2] have shown how
the internal mass of floodwater that is usually part of the
still water component of the load calculation, changes
with inflow and outflow. Even if this is accounted for,
the uncertainties in the fluid mass in the hull that is
derived from an intact condition can no longer simply be
applied to the flooded condition.
Because the probability distribution of still water load for
an intact vessel cannot readily be applied to a damaged
vessel, the DALAS-PARAMARINE software makes the
simplification that the still water load should account for
compartments fully flooding to a new static waterline
and that this should be a deterministic calculation at the
present time.
3.3 WAVE LOADING
The wave loading component lies at the core of the
DALAS project.
The first item of research was to investigate if a ship
would have a markedly different wave loading response
when flooded to that when intact. WASIM was used to
generate vertical bending moment (vBM) Response
Amplitude Operators (RAOs) for a small set of flooded
conditions, focussing on the midships region. This
showed that different flooded conditions would produce
unique wave-induced vBM responses. It also showed that
it was time consuming to manually produce, run and
analyse each WASIM model and assessment. If WASIM
The Damaged Ship, London, UK
(or similar) software was to be used to provide advice in
an emergency, then it would be useful to pre-calculate
the RAOs based on previously determined flooded
conditions.
To achieve this, a frigate hullform was created in
WASIM and a set of models built corresponding to 1,2,3
and 4 compartment flooding conditions. Symmetrical
flooding was assumed for simplicity. Time domain
simulations were conducted for each flooded condition,
recording the vertical bending moment response at 21
points along the length of the hull. The hull was run at 5
knots with a set of headings from 0-180 degrees in 15
degree bands. Post processing of this produced a set of
vBM RAOs in the frequency domain for each loading
condition, heading and position on the ship. These were
then incorporated within the PARAMARINE TM code.
3.4 VOYAGE PLANNING
A voyage planning module was then built for
PARAMARINETM. This requires a user to consult the
weather forecast (www.rnein.net was found to provide a
suitable description of significant wave height and
direction), before assembling a voyage plan as a number
of legs. Each leg would consist of a duration, a sea state
and a heading to the waves, (Figure 2). It would then
form the basis for calculating the number of vBM
exceeding given values at each leg. The final result
would combine the results for each leg producing a chart
of the probability of exceeding given bending moments
for the whole voyage. This is illustrated in Figure 3
3.5 Probability of Exceedence
Figure 4 shows how the summation of loads on each leg
of the voyage produces a chart of the probability of
exceeding a given wave-induced vertical bending
moment for the whole voyage. This is repeated for each
position along the length of the hull. The DALAS-
PARAMARINE user is then prompted to select a
probability of interest (1%, 0.1% etc). The resulting
values for each position are then re-assembled to produce
a plot of the appropriate WI-vBM along the length of the
hull.
At this point, the load is neither hog, nor sag. It has come
from the post-processed frequency domain RAOs which
do not differentiate between the two. Correction factors
can then be applied to produce hog and sag and the
default values are +/- 1.0. It is known that for sea states
of 4 or less, the factors are close to unity, but start to
diverge as wave height increases. Suggested factors are
shown below in Table 1.
It then remains for the still water component of load to be
added back to produce a total vBM load profile. At this
point the load can be compared to the available strength
The Damaged Ship, London, UK

Sea Mean Mhog Msag Mhog Msag PARAMARINETM screen capture of a frigate/destroyer
State Hs (IACS) (IACS) (WASIM) (WASIM) hull, damaged forward so that flooding occurs in C,D and
3 0.875 1 -1 1 -1 E watertight subdivisions. All structure within the green
4 1.875 0.987 -1.013 0.98 -1 box is deemed to have been removed, or rendered
5 3.25 0.970 -1.030 0.9 -1.05 ineffective and no longer contributes to the strength of
the hull. The user has assembled a 4 day voyage plan
6 5 0.948 -1.052 0.84 -1.06
back to the UK for repairs. In each leg, the sea conditions
7 7.5 0.916 -1.084 0.82 -1.03
are specified from the weather forecast. The user is then
8 11.5 0.866 -1.134 - - prompted to select probabilities of exceedence for
Table 1: Examples of non-linear correction factors for assessment. Figure 6 shows the vertical bending moment
midships position load for a PExceed of 1%, together with the deterministic
strength. The vertical bending moment in this case
4. STRENGTH includes the still water component. It can be seen that if
the user is happy to accept this PExceed, then the
The strength of a damaged ship is arguably the most decision may well be to proceed with the voyage; the
significant unknown. Various levels of detail are possible. load does not match the strength anywhere along the
In the past QinetiQ have used a Lloyds Register vessel. However, if the user is more cautious, a PExceed
Residual Strength Assessment RSA2 approach with the of 0.01% might be selected. In this case Figure 7 shows
code NS94 [3] to estimate intact and damaged strength. that the bending load is greater than the first failure
The latter was performed by simply removing elements strength in the region of the damage and the user may not
from the intact vessel model and re-calculating the choose to continue with the voyage.
ultimate bending moment capacity. Finite element
approaches (RSA3) are also possible, but tend only to be This example was created from an existing
practical if there is a lot of time available. PARAMARINETM model in about 30-40 minutes. The
most time consuming part is assembling the voyage file
It is recognised that, even for intact ships, an RSA2 from the ships speed, track and weather forecast. It can
method will not tell the whole story. Major assumptions be seen however that the voyage file can be readily
are made regarding the uniformity of material properties changed to reflect alternative strategies; increasing speed,
and scantling position and sizes. Through careful changing the repair port, adjusting heading relative to the
analysis it is possible to create a probabilistic estimate of weather etc.
the ships intact strength [4] accounting for these factors.
This initially promised to be a starting point for creating 6. REMAINING ISSUES
a probabilistic approach to the strength of a damaged The DALAS facility in PARAMARINETM has been
ship. developed to a relatively early proof of concept level
and has a number of issues that remain to be addressed.
It was quickly realised that, although this was a major These include:
step forward in understanding the strength of an intact
vessel, it was not really suitable for damaged ships. The The mix of deterministic data and probabilistic
main variable in the strength of a damaged vessel is data means that the risk of structural failure
likely to be the size and location of the damage. This is cannot be truly stated. The best that can be
very difficult to obtain while a ship is at sea; often the achieved is a quantified comparison of
best that can be known is the extent of any flooding, with alternatives, given different voyage strategies,
little knowledge of the geometry of the hole, or buckling or assumptions concerning the extent of damage.
of the surrounding structure. Without good information
on this, all other variables will have a lower order effect It is further recognised that the assumption of
on the ultimate strength. For this reason a probabilistic symmetric flooding may be reasonable for many
strength prediction will need to wait for further research conditions but is not always applicable.
and a deterministic solution for strength must be used, at
least initially. In the DALAS module of The use of correction factors to obtain hog and
PARAMARINETM, the strength is determined using the sag from the basic WI-vBM response is open to
NS94 method, with a conservative estimate of the further development. It may be sensible to use
amount of structure missing, defined by the user as a IACS values for this, or to generate correction
rectangular box-shaped region. factors unique to each hullform.
It is hoped that this weakness will spur development of
techniques to provide a rapid picture of the scale and
The facility has so far only been demonstrated
nature of underwater damage suitable for employment
for a single hull form. PARAMARINE and
from ships at sea.
WASIM models exist for a much wider set of
hullforms but there is no set of vBM RAOs for
5. EXAMPLE
these hulls generated in a suitable format for
PARAMARINE to access.
An example is illustrated in Figure 5 which shows a
 The Damaged Ship, London, UK

6. CONCLUSIONS focussed on ship survivability to weapon attack, but has

also encompassed structural survivability research and
A tool has been developed within the PARAMARINETM emergency response to damage. He is currently the
software suite that will enable a user to assess the relative capability leader for maritime survivability assessment at
risks of different courses of action if a ship is damaged. QinetiQ Rosyth.
This is based on the vertical bending strength of the
damaged vessel and on the wave loading the vessel may
see. The user can then make an informed decision on the
safest course of action, based on the relative probabilities
of structural failure.

This tool has known limitations, many of which can be

addressed through continued research and use.

7. REFERENCES

1. C Guedes Soares, T Moan, Statistical analysis

of still water load effects in ship structures
Transactions of SNAME 1988 vol 96., 129-156

2. T W P Smith, K R Drake, P Wrobel

Experiments on a damaged ship section.
University College London 2007

3. C S Smith, R S Dow, NS94: Ultimate strength

of a ships hull under biaxial bending ARE
report TR86204 Dunfermline December 1986

4. Magns Arason, John S. Kent Probabilistic

Assessment of hull girder strength and stiffness
OMAE 2006

8. ACKNOWLEDGEMENTS

With a multi-year project such as DALAS, there have

been many contributors. The Author would particularly
like to thank Mr John Kent, Miss Sarah Erskine and Mr
Magnus Arason, (formerly of QinetiQ) for their work at
the beginning of the project, Mrs Carol Toole (QinetiQ)
for technical support running WASIM, Mr Arne
Braathen of Det Norske Veritas for his understanding of
the problem and coding expertise and Mr Gareth Draper
(GRC) for bringing the software together. Tristan Smith
(UK MoD, UCL) provided both expert advice and
continuity through the project.

Finally Mr Michael Mogford, Mr Colin Snell, Mr Tom

Grafton and Mr David Manley (UK MoD) for their
continued support for the research.

8. AUTHORS BIOGRAPHY

Andrew Martin has been working in the field of naval

architecture since graduating from the University of
Strathclyde in 1992. In 1995 He joined the UK Defence
Research Agency, which became QinetiQ in 2001. That
year he graduated with an MSc in Naval Architecture
from University College London. His work has mostly
The Damaged Ship, London, UK

Load Strength

Damage is reported to a Paramarine user Damage is reported to a Paramarine user

Paramarine model is modified with Paramarine model is modified with a

flooded compartments damage template approximating reported
extent

Still water load Wave induced load

NS94D module removes damaged structure

Paramarine calc. Paramarine selection of and recalculates ultimate vertical bending
(Deterministic) probabilistic load distribution strength at cuts along the ships length
at user defined probability.
Stiil water BM

BM capacity
WI BM

X/L X/L X/L

BM load hog

X/L
Sag

hog
BM

X/L
Sag

Paramarine display

Figure 1 Overview of the DALAS PARAMARINE approach

Sea State Sig. Wave ht. (Range)
0 0
1 0-0.1
Voyage Leg Sea state Direction (o) Duration hrs
2 0.1-0.5
3 0.5-1.25 Leg 0-6 5 30 6
4 1.25-2.5
5 2.5-4.0 Leg 6-12 5 45 6
6 4.0-6.0 Leg 12-24 5 45 12
7 6.0-9.0
8 9.0-14.0 Leg 24-36 4 0 12
9 14m+
Leg 36-48 5 0 12
Leg 48-60 3 30 12
Leg 60-72 4 45 12
0 180
Leg 72-84 3 60 12
15 165
Leg 84-96 2 75 12
30 150
135
45
120
60 105
75 90

Figure 2 Assembling a voyage plan from forecast of significant wave height and direction
 The Damaged Ship, London, UK

Ship response Wave excitation

A set of pre-determined flooding conditions User picks a suitable sea area

are specified by user (A,AB, CDE etc.) this defines the distribution of wave freq/heights
Ship speed is specified in each sea state.

WASIM (linear) is run in time domain for this Weather forecast for vessel track produced.
condition - with a variety of wave periods Defined as a series of legs each with an
and headings associated sea state, ship speed and relative
heading to waves

Fourier transform of BM response to freq. domain. New module creates a set of sea conditions
Post processing (POSTRESP) to produce RAO (combination of Hs/Tz) within each sea state

Each sea condition is used to generate a wave

energy spectrum (PM)

For every leg

For every sea condition in the leg

BM response spectrum produced

Area under the response spectrum calculated

Used to define a similar Rayleigh distribution

for the probability of exceeding a given BM

From leg duration, ship response frequency

and prob of sea condition; the number of BM
exceeding given values is calculated

Summation of contributions from each condition

Summation of contributions from each leg

Figure 3 Core of the Wave-Induced vBM calculation

Figure 4 Creating a WI-vertical BM plot along the ship for a user-defined probability of exceedence
The Damaged Ship, London, UK

Figure 5 PARAMARINE screen capture: damage box inflicted on watertight zones C,D,E

Figure 6 Plot of vertical bending moments along the length of the damaged ship. Wave induced load is specified
at a PExceed of 1%
 The Damaged Ship, London, UK

Figure 7 Plot of vertical bending moments. Wave induced load is specified at a PExceed of 0.01%
The Damaged Ship, London, UK

THE STRUCTURAL AND STABILITY ASSESSMENT AND SUBSEQUENT RECOVERY

OF A DAMAGED LIFEBOAT
A Harman, N Chaplin, H Phillips and S Austen, Royal National Lifeboat Institution (RNLI), UK

SUMMARY

In a challenging week that saw three separate incidents of vessels grounding in UK waters, an RNLI Severn class
lifeboat on operational service grounded close in shore whilst trying to rescue a small boat aground on Rathlin Island
harbour breakwater. The resulting recovery operation, whilst not delivering an economically viable vessel, was
successful in preventing pollution and harm to a nominated Special Area of Conservation as well as salvaging thousands
of pounds worth of lifeboat equipment. This paper gives a background to the original design and construction of the
Severn class lifeboat and describes the sequence of events that took place from the point at which the vessel grounded.
The paper also reviews the theoretical and practical stability and structural analyses that were carried out to establish the
residual integrity of the lifeboat, and the recovery options in terms of damaged stability and structural integrity that were
investigated.

1. INTRODUCTION

Thanks largely to the high standards of training, the

courage and dedication of the crews and the quality of
lifeboats and equipment, the RNLI rescues an average of
22 people per day from around the coast of the UK and
Republic of Ireland. In the vast majority of cases, rescue
operations are successful and are carried out within the
boundaries of existing experience and training. However,
due to the unpredictable and hazardous nature of rescue
operations sometimes situations can be encountered that
can endanger crews and boats and require the backup of Figure 2: The grounded lifeboat (with the casualty vessel
support services to remedy. to right of picture)

One such occasion occurred on Tuesday 28th January

2008 when, whilst on operational service at night, the 2. VESSEL DETAILS
Portrush Severn class lifeboat RNLB Katie Hannan
grounded in shallow water off Rathlin island Northern 2.1 PRINCIPAL PARTICULARS
Ireland (see Figure 1). What followed was one of the
most challenging recovery operations undertaken in the Length OA 17 m
history of the RNLI (see Figure 2). Beam 5.5 m
Draught 1.8 m
Rathlin
Island Range 250 nm
Speed 25 knots
Load Displacement 42 tonnes
Crew 7
Construction Fibre reinforced composite
Fuel capacity 5,600 litres (marine diesel)

2.2 DESIGN AND CONSTRUCTION

2.2 (a) Structure and Construction Materials

The high level of structural redundancy present in the

design of the lifeboat contributed significantly to the
Figure 1: Aerial image showing Rathlin Island off the success of the recovery operation and a background to
North coast of Northern Ireland. Also marked are the structural design of the Severn class lifeboat is
locations relevant to the operation described below.

The RNLI uses an in-house structural load prediction

method for lifeboats based on theory, trials and in-service
 The Damaged Ship, London, UK

experience [1]. In addition, a number of lifeboats have

been instrumented to measure their structural response
with the resulting data fed back into the design process
and associated suite of computer models [2].

The global design loads adopted by the RNLI have

generally exceeded classification society requirements.
Figure 3 was developed during the 1990s as part of the
Severn class design development [3] that plots the
ultimate design panel pressure (kPa) against the product
of the displacement and operational speed (tonne knots).
The upper line represents the RNLI design curve with the
lower line representing typical classification society
figures.

Figure 3: RNLI Design Pressures

The decks and the deckhouse structure are also designed

to withstand pitchpole and capsize loads.

The Severn class is of longitudinally stiffened single skin

construction below the chine, typically 9mm thick, with
un-stiffened sandwich construction employed for hull
topsides (75-100mm thick), deck and wheelhouse.

Materials used are unidirectional, biaxial and multi-axial

E-Glass and aramid fabrics laminated with Epoxy resins,
and high density, closed cell PVC foam for core material
in areas of sandwich construction.

Key features of the structure are:

High strength to weight ratio
Robust (damage tolerant)
Long service life and low maintenance
Figure 4: Severn Structural Arrangement
Ease and speed of construction
2.2 (b) Water Tight Integrity
Figure 4 shows a plan view representing the Severn
structural arrangement.
Significant consideration was given to the water tight
integrity of the lifeboat in the design stage. The lifeboat
has to be capable of surviving capsize, pitchpole or
moderate impact damage and continuing on its mission.
The design features a watertight double bottom in the
fore store and survivors space, 5 full width watertight
transverse bulkheads below decks and watertight
The Damaged Ship, London, UK
longitudinal bulkheads creating outboard void spaces in
each compartment (excluding the Fore Peak).
The extent of the water tight sub-division and the damage
resilience throughout the lifeboat were key features of
the recovery process, as was the inclusion of a large bow
thruster sited near the bow. The tunnel is a filament
wound, epoxy construction, and is well encapsulated into
the hull structure. The thruster tube is tied into the
longitudinals and is further stiffened by a centreline
stiffener, making for a strong structure in this area.
2.2 (c) Environmental Considerations
Owing to the damage sustained during the grounding,
and the sensitive location, it is relevant to identify the
significant materials which could have represented
environmental hazards during and after the grounding.
Although the composite hull is made from
potentially hazardous raw materials, their
harmful properties are largely nullified during
the manufacturing curing processes. However
broken up glass fibres and PVC foam, whilst
mostly inert, may be harmful if ingested by
living organisms.
Diesel fuel oil stored in two flexible, purpose-
made fuel bags. The bags are placed in
composite tank boundaries located above a
partial double bottom and inboard of the wing
voids in the fuel tank space in the centre of the
vessel. The capacity of each fuel bag is 2,800
litres.
Engine and gearbox oils.
Hydraulic oil for steering system, trim tabs, bow
thruster and crane.
The environmental aspects are discussed further in
Section 3.
2.2 (d) Stability
All RNLI all weather lifeboats are inherently self
righting and have no angle of vanishing stability. As a
result, the lifeboats are very stable vessels.
The RNLI minimum requirement for damaged stability is
for a 2 compartment standard, however virtually all
lifeboats achieve a much higher degree of damaged
stability (although self righting in a damaged condition is
not a requirement).
The size and form of the Severn class, produces a large
internal volume which permits the inclusion of a partial
double bottom and longitudinal wing voids which
provide a high degree of damaged stability.
Additionally, the use of the thick foam sandwich
construction in the top sides provides a high degree of
secondary buoyancy in the event of damage.
The RNLI has undertaken extensive stability analyses of
the Severn class lifeboat and has proven it can be
considered to be a safe haven for her volunteer crew in
the event of almost any damage, except for the case
where damage has been caused throughout the vessel
breaching hull compartments, void spaces as well as the
wheelhouse. As such damage is considered to be
extremely improbable and the Severn is deemed to have
excellent residual buoyancy and stability in the event of
damage.
3. LOCATION / ENVIRONMENT
Rathlin Island is a small inhabited island approximately 3
miles off the North coast of Northern Ireland. The island
is a nominated EU Special Area of Conservation,
featuring specific rare marine life examples.
For the RNLI, as well as wishing to recover the lifeboat
in as good a condition as possible, the environmental
sensitivity of the location made quick and decisive action
essential, emphasising the importance of ready access to
recovery expertise.
However unfortunate the RNLI may have considered the
grounding in such an environmentally sensitive location,
it was however blessed by the fact that a number of
Rathlin Island residents were extremely capable, well
equipped and supportive of the RNLI.
The recovery team were to become extremely thankful
for their support, not least of all to one resident, who not
only offered a room in his house from which to set up a
temporary site office, but who also owned a range of
machinery and equipment that were to prove
indispensable. Also of benefit was his understanding
nature, as the lifeboat was to eventually end up driven up
the breakwater, to lie within the boundaries of his
property!
Although remote, Rathlin Island had good internet
communications which made liaison with support
services possible. Additionally, there is a regular small
RO-RO ferry service run between Rathlin and
Ballycastle, as well as various local water taxi/charter
boats for hire.
4. INITIAL ASSESSMENT & RECOVERY
ATTEMPT
Whilst incidents of this nature are thankfully infrequent,
there have been other occasions where by damage to a
lifeboat has been sustained. With this in mind, the RNLI
undertook to set down a methodology with regard to the
steps to be taken should a lifeboat be damaged [4].
Although additional steps were required in this particular
case, the overall methodology was followed throughout.

Immediately following the incident, the RNLI
Engineering Office was contacted by the RNLI
Divisional Inspector (Ireland) for advice on the damage
tolerance, damaged stability and capability of deck
fittings of the lifeboat. Armed with that advice, on
Wednesday 29th January an attempt was made to re-float
the boat on the high tide. The vessel was inspected and
was found to have substantial hull damage either side of
the keel which allowed free flooding of the machinery
and tank spaces. Arrangements were rapidly made to
place a local tug on scene and a line was made fast
between the boats. However the 30 tonne pull astern on
the lifeboat was insufficient to free her from the rock
ledge where she remained, with the vertical face of the
aft end of the keel and the rudders lodged in a jagged
rock platform.
The failed recovery attempt was conducted in the face of
worsening weather, and whilst the lifeboat is extremely
tough and stable in deeper water, the fact that she was
grounded in shallow water presented a risk that she could
heel over onto her side, and hence the safety of the
volunteer crew could not be guaranteed. Accordingly an
Irish Coast Guard helicopter was called to transfer the
crew the very short, but hazardous, distance to shore.
The lifeboat was then left to battle it out with the
elements on the lee shore of the granite rock armour
breakwater. The wind rose to storm force and the boat
took a severe pounding, being washed further along the
breakwater with every tide.
The poor weather continued for two days prohibiting
access to the site. During this time the unfolding
conditions were monitored via photographs emailed from
a local resident of the island, and plans and preparations
initiated for the subsequent recovery operation.
Once the initial tug recovery attempt had failed, the
RNLI implemented its emergency procedures and gained
access to pre-defined recovery and pollution control
expertise through the Institutions insurers. A
communications and media plan was also instigated.
When the weather abated sufficiently to make access to
the island possible, an on site assessment was conducted
by RNLI staff and the decision was quickly made to strip
out the lifeboat. As well as reducing the pollution
potential and salving valuable items, the removal of
stores and equipment would lighten the 42 tonne vessel,
improving the chances of a successful recovery
operation.
5. SAFETY & REGULATION
5.1 SITE SAFETY
The RNLIs purpose is to save life at sea, and an inherent
aspect of this is a deeply ingrained safety culture
The Damaged Ship, London, UK
throughout the organisation. Accordingly the on-site
team set up site safety and risk assessment protocols,
documenting the risks, controls and procedures. All staff
and visitors to the site were required to read the key risk
assessments and to implement the necessary control
measures (including appropriate personal protective
equipment) before being allowed on site. Safe access to
the lifeboat was set up using locally available scaffolding
5.2 REGULATORY CONSIDERATIONS
As with any major marine incident in the UK, the
incident was reported to the Maritime and Coastguard
Agencys (MCA) Secretary of State representative
(SOSREP) and the Marine Accident Investigation
Bureau (MAIB), and accordingly the RNLI continued to
work closely with the nominated MCA officers to ensure
safety and pollution considerations were acted upon.
Additionally, the RNLI worked closely with other
relevant agencies such as the Northern Ireland
Environment and Heritage Department (now Northern
Ireland Department of Environment).
5.3 POLLUTION PREVENTION
Although the volumes of hydrocarbons on board the
lifeboat were comparatively small the vessel had
grounded in an environmentally sensitive area which, if
polluted, would have caused significant damage to the
voluntarily funded organisations reputation.
The on-site assessment identified that although the
storms and granite breakwater had inflicted extensive
external damage, the vessels fuel and hydraulic oil tanks
were still intact. Therefore the RNLI and its nominated
recovery experts deployed a fuel bowser across to
Rathlin Island to de-store the hydrocarbons. All fuels
and oils were rapidly and successfully removed
preventing immediate environmental damage.
6. EQUIPMENT REMOVAL
Having removed the hydrocarbons, the team of RNLI
staff and volunteers set about a rapid operation to remove
as much weight and valuable equipment from the vessel
as possible. This operation was undertaken using both a
versatile Manitou mobile crane vehicle as well as an
aerial runway erected between the vessel and the shore.
As weight and (if floating off were to be a recovery
option) stability would be key information to the
recovery operation, a comprehensive record of all items
removed was kept, and incorporated into a running
weight and centre spreadsheet. This information was
based on the component weight information recorded
during the original Severn class lifeboat build.
The Damaged Ship, London, UK
Figure 5: Removal of one of the lifeboat engines using
local mobile crane
7. RECOVERY OPTIONS
As each tide had washed the lifeboat further up the
breakwater, it also created further damage to the hull.
The constantly changing condition of the boat meant that
the RNLI worked on a number of recovery options
simultaneously, exploring the possibilities and
implications of each option, before finally acting on the
one with the greatest chance of success.
7.1 LIFTING ONTO A BARGE
The limited size and capability of the Rathlin Island RO-
RO ferry meant that road based recovery was not a
feasible option for the recovery of the lifeboat, and the
potential for lifting the boat onto a barge was explored.
Whilst confident in the design and structure of the Severn
class, initial on-site inspections could not guarantee that
the boat would have sufficient structural integrity to
sustain a lift unsupported, and therefore it was decided to
investigate building a cradle around the boat to support it
during the main lifting operation. Members of the RNLI
Engineering Office then designed a suitable lifting cradle
with the following design criteria in mind:
To support boat during lifting operations
following strip out, with a factor of safety of 1.5
Tie down points to be included
Spreader beams to be used for lifting
Capable of being dragged across the ground if
required, in order to move the boat to a more
accessible location for craning operations
Designed using standard steel sections to speed
fabrication
Practical on site assembly
Subsequent inspections were able to verify the ability of
the structure to withstand a lift but this cradle design
provided a vital contingency plan if the structure had
been deemed unable to do so.
Ultimately it did not prove feasible to recover the lifeboat
through lifting as there proved to be insufficient water
depth and access for a suitable crane barge to get close
enough to the vessel and the proposal was abandoned.
Having made this decision, attention then focussed on the
alternative of recovery to sea, which had already
progressed as far as preliminary residual stability
investigations.
7.2 FLOATING OPTION
Following discussion with the MCA representative,
Recovery Master and Environmental Officers, it was
agreed that an attempt would be made to drag the boat
off the rocks, tow it out to sea and load it onto a barge.
Obviously a key aspect to the recovery was to identify a
suitable and available barge. The recovery experts
suggested contacting R Wynn & Sons Ltd to see if their
unique barge/floating dry dock, Terra Marique (TM),
was available. Meetings were held with Wynn
representatives to ensure that the barge was both
available and suitable for the intended operation, and
having confirmed both aspects, contractual negotiations
were held to establish the terms for the operation.
The TM is a floodable barge capable of accepting large
cargos afloat or ashore and has the following principal
characteristics of note to this operation:
Length OA: 80m
Beam mld: 16.5m
Depth to coaming: 8.5m
Depth to main deck: 6m
Deadweight: 2211Te
Design draught: 1.6 4.8m
Max cargo hold draught: 2.5m
The recovery plan was then drawn up in increasing detail
liaising with all relevant parties at every step, including
regular periodic meetings of key stakeholders.
The key points for the plan of action were as follows:
Complete de-storing and secure sea going
capability
Confirm route to sea (depth, obstructions etc..)
Confirm likely damaged stability, draught, air
draught, trim and list
Confirm sea going limitations
Arrange tug and other support vessels
Arrange shore facility to receive lifeboat
Turn lifeboat to required heading
Prepare and protect shoreline
Secure towing bridle to lifeboat

7.2 (a) On site damage assessment
A clear concern for the recovery team was the structural
integrity of the vessel. Having received considerable
damage on the breakwater, the team had to ensure that
the vessel retained sufficient strength to undertake the
recovery operation, and preferably without the need for
additional support which would have added weight.
The on site assessment of the vessel was lead by an
RNLI hull surveyor and continual monitoring of the key
structural elements was undertaken. The key findings
were that:
Main keel was extensively missing.
Main bottom longitudinals were heavily
damaged but capping was present.
Central box section (wing tank structure, soles
and deck) were intact.
Although the Severn class lifeboat is the largest in the
RNLI fleet, it is still relatively small and has a low L/B
ratio (3.1), and is very strongly reinforced throughout.
This design coupled with a reduced weight (particularly
in the heavily damaged areas aft) meant that the residual
strength was good, particularly due to her central box
section of sole, wing voids and deck, all of which were
still intact. These facts gave the engineers and on site
team a high level of confidence in the strength of the boat
to survive the recovery operation.
7.2 (b) Damaged stability
Compartments were categorised as intact or damaged,
and appropriate permeability figures assigned, although
in some cases these were difficult to estimate and
therefore numerous calculation variations were run to
assess the impact on the damaged stability. Although
most damage was very obvious, in some cases it was not
possible to be absolutely certain that apparently intact
structures were not in fact cracked and could leak when
the vessel went afloat. Therefore in some spaces it was
decided to employ temporary buoyancy to decrease
potential permeability.
Due to the construction of the vessel, the possibilities for
deployment of temporary buoyancy are limited
(particularly due to the size of access hatches into the
wing spaces). However a cheap and viable local solution
was identified utilising 10 litre plastic drinks containers
from a local factory in Northern Ireland. Several
hundred of the containers were brought to the island on
the ferry and packed tightly into the wing void spaces
outboard of the engine room and steering gear spaces.
With the wing voids either intact or with a very low
permeability, the stability calculations showed that the
damaged vessel would achieve a high level of residual
buoyancy with the wing voids acting in effect as
outriggers to the main damaged centre section of the hull.
The Damaged Ship, London, UK
It was to be this key feature of the Severn design that led
to the successful recovery of the vessel.
Also intact were the two fuel tanks due to their location
above the double bottom and in board of the wing tanks.
Additionally the large survivor space was intact above
the double bottom.
The fore store was sound above the double bottom, and
both wing tanks and the anchor locker were intact.
However although the wing tanks appeared to be sound
internally, they had suffered impact damage externally.
All of these findings were communicated back to the
RNLI Engineering Office and damaged stability
assessments were updated using the Institutions
Maxsurf/Hydromax capability.
As well as suitable residual stability, the recovery team
had to ensure that draught and air draught figures would
be compatible with the Terra Mariques capabilities.
Various assessments were therefore run across a range of
possible flooding scenarios and permeabilities to produce
an envelope of likely stability characteristics and metrics
(see Figure 6), with margins on each estimate or assumed
value.
Figure 6: Example damaged stability information
These analyses confirmed that the vessel should be able
to successfully and safely dock into the flooded TM hold,
without the need for external support or buoyancy.
7.2 (c) Shoreline survey and protection
The vessel had come to rest on a shallow rock ledge to
the seaward side of the breakwater. Charted depths and
tide tables indicated sufficient water for the damaged
vessel to float at high tide some 40 50m to the West of
the breakwater, however even at high tide, the vessel
would have to be dragged across the ledge to the deeper
water. Therefore following a closer on site survey of the
immediate area a suitable natural gulley was identified
which made for the shortest suitable lead to deeper water
The Damaged Ship, London, UK
and in line with deeper water for safe tug and
marshalling operations.
The on-site team then set about deploying a series of
large steel plates to act as both protection for the
environmentally sensitive sea bed, and to reduce the
probability of the damaged hull snagging on underwater
protrusions.
Figure 7 below shows an excerpt of the hand written
RNLI master planning document illustrating the planning
for safe and danger areas for the operation. The plan was
presented to the MCA representatives.
Figure 7: Recovery plan master chart
7.2 (d) Lifeboat alignment
The line of the proposed recovery route was
approximately 120 degrees to Starboard of the bow, and
therefore in order to recover the vessel it was necessary
to drag her round to the required heading. Given the
damage sustained and the rocky nature of the sea bed,
this operation required specialist equipment.
The RNLIs smallest all weather lifeboat, the Mersey
class, is designed to be launched and recovered across an
open beach from a carriage. A marinised crawler
vehicle, the Talus tractor, is used for this operation, and
one of these was brought to Rathlin to assist in turning
the Katie Hannan. Fortunately, the weight of the Talus
was just within the 17 tonne limiting payload of the
islands RO RO ferry. These tractors are equipped with a
powerful hydraulic winch and line capable of exerting 15
tonnes pull as standard. The winch was also especially up
rated for the recovery operation by the manufacturers
prior to transportation to the island.
Two hydraulic jacks, assorted steel plates, props, wooden
blocks and wedges were also sourced and used to lift and
support the boat sufficiently to remove rocks from
beneath the starboard side, and then the lifeboat was
lowered back down onto high density plastic skids
(normally used to recover Mersey class lifeboats) which
allowed the boat to slide as it was pulled round.
The windows of opportunity to achieve this part of the
recovery operation were limited by tide as the staff had
to work under the hull.
The operation to rotate the vessel using the tractor and
plastic runners was fortunately aided by the presence of
the bow thruster tunnel, through which the winch line
was passed to provide a low angle of pull. The plastic
runners helped reduce the stress on the hull limiting the
amount of consequential damage and any further
reduction of residual strength that would be required for
the recovery operation.
7.2 (e) Towing operation planning
As part of the decision making process to recover the
lifeboat to sea, a further risk assessment was undertaken
to identify and manage the risks involved in the
operation. Additionally, as the previous attempt to tow
the vessel clear had failed, the reasons for that failure had
to be assessed and overcome. The vessel weight had
now been reduced to 26 tonnes, the orientation changed
to be a more favourable bow to sea aspect, the sea bed
protected and snagging potential reduced.
It was also necessary to consider a suitable method of
attachment of the hauling line to the vessel. Although
lifeboat fittings are very capable, for example the stem
bollard has a design capability of 20 tonnes, however the
estimated pull required to move the vessel was still in
excess of 30 tonnes. Therefore it was not possible to rely
on the bollard alone to secure the line to.
The on-site recovery team therefore created a
circumferential bridle from webbing slings and shackles
to spread the towing load around the vessel.
Liaison was set up via the recovery experts with the
recovery tug master to ensure that there was sufficient
depth of water and sufficient sea room to conduct the
operation within the limit of the tug line. A suitable
location and line of recovery was established and further
communications were set up to confirm a temporary
exclusion zone for the operation through the Maritime
and Coastguard agency. The zone was an essential
measure to ensure that all local marine traffic such as the
numerous anticipated media vessels etc. were kept out of
the potential danger area.
Finally a very close eye was maintained on weather
forecasts day by day to ensure that the operation could be
conducted within the appropriate weather limitations.
7.2 (f) Support vessels
In addition to the barge Terra Marique, her tug and the
recovery tug, a number of other vessels including an
RNLI all weather Trent class lifeboat from Larne were
deployed to assist with various operations such as
 The Damaged Ship, London, UK

passing the tow line and manoeuvring the vessel into the On arrival at Ballycastle the same personnel plus a
Terra Marique. Lloyds Register surveyor made the vessel safe and
certificated for the longer passage to the final destination
7.3 RECOVERY OPERATION port of Plymouth. Additionally, the RNLI staff made
further inspections and assessments of the lifeboat in the
Once all preparations were completed, the recovery dried hold of the Terra Marique and concluded that the
operation got underway at the prescribed time to suit the vessel had retained sufficient structural integrity to be
tide. All personnel were excluded from the immediate able to be lifted out of the hold.
area and the tug began to pull on the line. After initial
attempts at approximately 20 tonnes pull which did not During the 2 day passage to Plymouth, RNLI and
move the vessel, the line pull was increased to 28 tonnes Babcock Marine Limited staff worked closely to develop
and the vessel began to move. She then made rapid plans for offloading the vessel, which included risk
progress to deeper water as planned. The vessel draft assessments and a step by step method statement.
was monitored from shore using the additional marks
painted on her hull for this purpose and she eventually On arrival at the dockyard, the Terra Mariques hold
settled very close to her predicted marks at draughts of hatches were removed to provide access to the hold. The
2.0m (forward) and 2.1m (aft), comfortably within the mobile crane was set up on the dockside and the lifting
maximum draught of the Terra Marique. plan enacted. Again good use was made of the vessels
bow thruster tunnel to raise the bow slightly to facilitate
Once fully afloat the tug stopped the haul and RNLI staff passing of lifting slings under the hull.
boarded the vessel to set up an alongside tow by another
lifeboat and to carry out an on board inspection. Once the slings were located as defined in the plan, the
lift commenced and the vessel was raised onto the
The inspection revealed that all compartment predictions dockside. During this operation it became possible to see
for damage and intact were correct and that inter- the extent of the damage the vessel had sustained, and
compartment hatches and doors were holding fast. Some this is illustrated in Figure 9 below.
leakage around the edges of a double bottom panel in the
fore store was noted but the rate was slow. Additionally
from what was possible to inspect there did not appear to
be further damage to key structural members and
therefore there were no elements that warranted anything
to jeopardise the short passage into the Terra Marique.
The operation awaited final clearance from the MCA to
allow the Terra Marique to open her bow doors in open
water; an operation never previously undertaken as there
was concern that replacing the doors in a swell could
cause jamming and render her unseaworthy. Fortunately,
the operation was blessed with a high pressure and zero
swell condition in Rathlin Sound on the day in question.

Figure 9: Lifeboat lifted onto dockside at Plymouth

A final condition assessment was then conducted with

the vessel blocked off in the dockyard, and it became
clear that it would not be economically viable to repair
the vessel, and therefore disposal became the only
option.
Figure 8: Lifeboat entering barge Terra Marique
The vessel was disposed of in 2009.
The RNLI crews and Wynn staff then berthed the vessel
safely within the Terra Marique (as shown in Figure 8)
and proceeded to close and pump out the hold, fit the
8. CONCLUSIONS
hold hatches and de-ballast for her 10 mile voyage to
Ballycastle.
Given the adversity of the situation as a whole, the
recovery of the Katie Hannan was only successful due to
The Damaged Ship, London, UK
the high degree of structural redundancy, damage
tolerance and damaged stability inherent in the lifeboats
design. Aside from affirming the RNLIs design
philosophy used for its lifeboats, the incident
highlighted a number of areas which are considered key
to ensuring the success of similar operations in future. As
such, the RNLI is developing and maintaining salvage
packs for the all weather lifeboat classes, based on the
lessons learned during the incident. Without such a
salvage pack, successful recovery was only possible due
to the knowledge of the boat and her capabilities by key
RNLI personnel. The salvage packs will contain the
following key information:
Salvage methodology, considering all probable
scenarios
Complete detailed removable weights and
centres information
Alternative lifting points if required
Lifting cradle designs, for build at short notice if
required
Updated key contacts information within and
outside of the RNLI.
9. ACKNOWLEDGEMENTS
The authors would like to thank the following persons,
without whom the success of the operation might have
been jeopardised:
George Rawlinson, RNLI
Andrew Redden, RNLI
James Duncan, Rathlin Island
R Wynn & Sons
Babcock Marine
Kelly Allen, RNLI
10. REFERENCES
1. R. M. Cripps, H. J. Phillips & C. Cain, Development
of Integrated Design Procedures for Lifeboats. SURV 6:
Surveillance, Pilot & Rescue Craft, RINA, 17-18 March
2004, pp 9-15.
2. H. J. Phillips, R. M. Cripps, S. J. Rees & L. F.
Vaughn, The Development and Validation of a
Hydrodynamic and Structural Design Methodology for
High Speed Rescue Craft. High Speed Craft, RINA, 17-
18 November 2004, pp 65-70.
3. Hudson, F. D., Hicks, I. A. & Cripps, R. M., The
Design and Development of Modern Lifeboats, Proc.
Instn. Mech. Engrs, Part A ,Vol 207, No A1, pp 3-22,
1993
4. Robert M. Cripps, Janice M. Dulieu-Barton, Han K.
Jeong, Holly J. Phillips & R. Ajit Shenoi, The
Development of a Generic Ship Repair Methodology,
Journal of Ship Production, SNAME, Volume 22,
Number 1, February 2006, pp 21-32.
11. AUTHORS BIOGRAPHIES
Antony Harman joined the RNLI as a Naval Architect
in 2008. Responsible for the stability and buoyancy
aspects of the lifeboats, he also forms part of the E-Class
replacement team, concerned with the development of a
replacement for the current lifeboats used on the River
Thames. His previous experience includes working as a
Naval Architect in the commercial RIB and motor yacht
industries.
Neil Chaplin is Staff Officer Operations (Technical) in
the RNLI Operations department. Prior to this he spent
10 years in the RNLI Engineering Office, including
appointments as Principal Naval Architect and project
manager for the Tamar class lifeboat. Before joining the
RNLI he worked for the MoD Procurement Executive on
a variety of support vessels including hydrographic and
fleet auxiliary ships.
Holly Phillips is Principal Naval Architect at the RNLI.
She is responsible for naval architectural aspects of
lifeboats and their equipment and is the Project Manager
for the E-Class replacement boats for the River Thames.
Her previous experience includes rudder system design
and finite element analysis of composite structures.
Steve Austen is the Head of Engineering Support at the
RNLI and manages the Engineering Office in the design,
support and development of lifeboats and their
equipment. After a number of years working for the
MoD, he joined the institution as a Naval Architect in
1999. His subsequent roles included Principal Naval
Architect responsible for the project management of the
new D-class inshore lifeboat, the Fast Carriage Boat 2,
Tamar launching equipment and the writing of numerous
lifeboat repair specifications.
The Damaged Ship, London, UK
A GA BASED DECISION SUPPORT TOOL FOR STABILITY AND STRUCTURAL
VIABILITY UNDER DAMAGE
T Martins and Sousa Lobo, CINAV Portuguese Navy, Portugal
SUMMARY
Stability calculators that allow for the ships master to simulate any load condition and ensure that stability criteria are
fulfilled are usually used onboard. On the other hand, whatever stability or structural criteria are applied, it is impossible
to foresee every possible damaged situation at every load condition. Therefore, this paper proposes a decision support
tool to help the commanding officer/ masters decision after damage and to monitor automatically the ship stability
during operation.
NOMENCLATURE
t Time (sec)
lt Trim at time t (m)
Ct Tank/ watertight fluid capacity at time t
(t)
HYt Values of hydrostatics at time t
FSMTt Transverse free surface moment, at
time t (kN.m)
LCt Load condition at time t
LCGt Longitudinal
LKt Values of stability cross curve at time t
Qi Flow of water in compartment i (m3/h)
STi Function that checks Sarchin and
Goldberg intact stability criteria and
strength criteria (if value > 0 it meets
the criteria)
STd Function that checks Sarchin and
Goldberg damage stability criteria (if
value > 0 it meets the criteria)
SKt Solid cargo distribution onboard at
time t
TCGt Transverse position of the Centre of
gravity at time t (m)
TQ Flow capacity of the ballasting
equipment (m3/h)
TK Set of tanks (superscript stand for tank
types: FO, LO, FW, BL, MS and
subscripts identify the tank)
VCGt Vertical position of the centre of
gravity at time t (m)
WC Set of watertight compartments
(subscripts identify the compartment)
WQi Drainage equipment flow capacity
assigned to compartment i (m3/h)
Binary matrix that expresses if the
equipment is used in each compartment
Real numbers matrix that indicates the
embarked seawater in each tank or
watertight compartment
t Displacement at time t (t)
Real number matrix that indicates the
filled capacity for each tank of the
proposed solution
1. INTRODUCTION
A vessels safety and survivability in case of damage is
dependent not only upon her architectural options,
technical solutions, capabilities, and maintenance, but
also upon the crew knowledge of the ship and on how
good their training to respond to an emergency is. For
that purpose, regulatory bodies, navies and most of the
large companies agree on imposing technical and safety
regulations, on doing surveys regularly, and on training
the crews up to a satisfactory level. In several cases, the
vessels systems automation and redundancy have also
been increased to provide better and safer solutions.
On the other hand, amongst shipping companies and
navies all around the world, crew numbers are becoming
smaller, they are less experienced, and crew replacement
occurs more often. In this paper, these facts are
considered reasons for concern, for they can be a cause
for accidents or poor responses in case one occurs.
In order to surpass these obstacles, a decision support
tool (DST) to advise on how to proceed in case of
damage is purposed. This tool has been designed to
support vessel operation and emergency reaction and is
composed of four separate modules with the following
purposes:
monitoring the vessels loading condition by
automatically controlling the tanks loads;
verification at regular intervals, or as demanded,
of stability and structural criteria;
decision support for reaction in case of damage;
estimation of vessels operational performance
and limitations taking into consideration loading
and environment conditions.
This paper is going to discuss the first three modules, and
focus on the implementation of a Genetic Algorithm
(GA) in the third module to obtain a good solution for the
vessels survival under damage, using the stability
damage criteria as a restriction of the optimization
process, which aims to get a rapid response with onboard
capabilities and available crew members.

2. STATE OF THE ART
Before the current level of automation that nowadays
exists aboard warships, crews used information such as
Compartment Flooding Effect Diagrams and
Compartment Liquid Loading Diagrams to analyse the
damage and decide upon damage control actions. These
practises are still in use and can be found on several
manuals of different navies. Some of the problems with
those diagrams, is that they only present data
corresponding to one single load condition.
According to [1], Flooding Casualty-Control Software
(FCCS) has been in service with the United States Navy
(USN) from 2000 onwards, suffering several changes
since then. In its last version, dated from 2008, the
software can:
evaluate the ships attitude, hydrostatics,
stability and strength status;
model the ships loads and track them over time;
evaluate the effects of flooding on stability and
strength;
evaluate longitudinal strength in various sea
states and if structural damage occurs;
provide warnings and recommendations when
the vessel fails stability or strength criteria;
evaluate stability and bottom reaction forces
during vessel stranding.
Parallel to these improvements in flooding emergency
response situations, Genetic Algorithms (GA) have been
used to optimise the ships arrangement and determining
watertight subdivision [2] taking survivability into
account, among other requisites, in the conceptual design
stage.
Our work done intends to follow the FCCS development
as far as providing warnings and recommendations in
case of emergency. However, we use GAs as a method
to solve the vessels survivability problem, optimizing
response time and drainage resources.
3. CONCEPTUAL MODEL
Analysing the procedures practiced in the operation and
damage control training of Ocean Patrol Vessels or
similar ships, there are some common practices that
might be of interest when developing a DST to improve
safety and damage response:
stability is usually checked at the beginning of a
journey, after which a pre-determined tanks
consumption and ballasting order is followed;
when there are deviations from the studied load
conditions, they are checked against load limits
and intact stability criteria;
The Damaged Ship, London, UK
during Replenishment At Sea (RAS), the initial
and final load conditions are verified against
intact stability criteria;
throughout damage control exercises, decisions
are usually based upon the Compartment
Flooding Effect Diagrams taking the best
known load condition into consideration, though
damage calculations are only done if necessary;
during the same exercises, there is a
predisposition to try and save the compartment
draining the water out before any stability
analysis is done.
Figure 1 represents the DST architecture we used, which
is going to be explained next.
Figure 1 Decision support tool flow diagram
3.1 CONDITION MONITORING
The DST that is being developed is able to evaluate the
vessel load condition and the open/ close condition of
watertight doors, hatches and ventilation valves from
time to time. This information, in case of flooding, will
provide an accurate estimate of the ships hydrostatics
before flooding and if and how the water may spread.
The load condition is estimated from the initial condition
calculated from draught marks readings done before the
journey initiates, taking into account the fuel and fresh
water consumption, as well as ballasting. This
information is given by an automatic sounding system
installed on every tank onboard. No accelerators to
measure roll are used due to the inconclusiveness of the
recorded estimates to measure the metacentric height
(GM) from roll period records while in operation,
expressed by reference [3]
3.2 STABILITY CRITERIA VERIFICATION
Until the user specifies that damage has occurred, the
system monitors continuously the vessel condition, and
The Damaged Ship, London, UK
checking if both intact and damage stability criteria are
met, as well as if the bending moment and shear forces
are within the design values.
In our current DST implementation the Sarchin and
Goldberg stability criteria is used [4], taking into account
specific changes implemented by the Portuguese Navy.
The damage criteria are checked for every group of two
continuous watertight spaces with no regards for decks.
In the calculations, the added weight method is used with
allowances for free surface effects caused by partial
flooding and free communication with the sea.
Another DST capability is to make use of these software
routines to simulate any condition easily and quickly,
allowing it to be used as a normal stability calculator.
3.3 RESPONSE TO DAMAGE
The response to damage can be seen as an optimization
problem where there is one cost function and several
constraints. Figure 2, presents the diagram used to
approach the problem with some considerations about the
implementation already done.
Figure 2 Damage calculations optimisation flow
diagram
First, as damage occurs, the user must specify where, and
how big it is. From it, using the current load condition
and information from watertight openings, the correct
extension of the damage and amount of flooding can be
determined.
Once the scenario is well defined, several solutions can
be idealised. The feasibility of each one of them is then
checked against constraints that in this case are
concerned with fluid transfer between tanks (fuel oil),
space flooding (ballast and voids), or discharges to the
sea (fresh water), as well as verifying the damage
stability criteria and if the structure resists the produced
bending moment.
Only when a solution has been verified, will it be
evaluated. The evaluation process takes into
consideration two different issues:
how far the stability and strength criteria
calculations are from the acceptable limits;
how much time is necessary to implement the
solution.
The next two chapters explain the mathematical
formulation and the genetic algorithm implementation.
4. MATHEMATICAL FORMULATION
Analysing ships stability and damage response as an
optimization problem requires a strict mathematical
formulation. On the other hand, we dont focus much on
the actual method implemented to check the stability
criteria, since the reasoning used can be applied to any
given one.
One must initiate the formulation defining the required
input variables which characterise the vessel:
1. light ship displacement () and centre of gravity
position (LCG, TCG, VCG);
2. hull form that allows us to obtain her
hydrostatics (HY) and cross curves (LK);
3. internal subdivision (TK, WC);
4. other weights that may vary with time (SK,
C(TK)).
The first set of variables are brought together to define a
load condition at the beginning of the journey (LC0). As
the journey continues, the fluids within the tanks change
Ct(TK) while the solid cargo is considered to be
unaltered, so that for a given instant in time:
LCt = f (LC0 , Ct (TKi )) for t > 0 (1)
where f is a known function.
Following this reasoning, the intact stability and
structural criteria verification depends upon the current
load condition. On the other hand, damage stability
criteria not only depend on the load condition, but also
on which watertight compartments are flooded.
Expression (2) translates this into mathematics:
 The Damaged Ship, London, UK

STi (LC0 , Ct (TKi )) > 0

(2) where t1 is the time when damage occurs and t2
STd (LC0 , Ct (TKi ), Ct (WCi )) > 0 for t > 0
is some time after.
where STi and STd stand for functions that calculate how 7. Only if drainage flow of the available equipment
far the criteria values are from their limits and add them is bigger than the flow of water ingress, is
up using scale factors to account for differences between equipment used to drain the compartment
their magnitude.
(WQi ) > Qi (9)
Hence, it can be concluded that for the optimization
process, during operation or in damage, the decision 8. Only if damage stability criteria are met after
variables are how filled each tank is (Ct(TKi)) and how drainage, is equipment used to drain the
flooded each watertight compartment (Ct(WCi)) is, since compartment
one can decide if they are drained out or not.
j (WCi ) = 1
(10)
Further, in this formulation, besides the constraint STd (LC0 , Ct (TKi ), Ct (WCk )) > 0
defined in (2) that ensures that the stability criteria are
met, there are a few other constraints: where j identifies the equipment and WCk
corresponds to a set of flooded compartments
1. Tanks and watertight compartments are filled/ that dont include WCi.
flooded from 0 to 98% of their capacity
The fitness function for damage response has been
0% < Ct (TK i ) < 98% developed to take into account:
(3)
0% < Ct (WCi ) < 98%
how far the results of the damage stability
2. Fresh water will not be produced in sufficient criteria calculations were from their limits;
time, neither ballast is going to be considered to how much time does it take to drain the flooded
fill these tanks. Nevertheless, fresh water can be compartments;
discharged overboard how much time does it take to ballast tanks or to
go ahead with the fluid change between tanks;
Ct (TKiFW ) Ct (TKiFW )
2 1
(4)
each one of these are measured by a value which is then
iFO iFO
multiplied by a scale factor, in order to make their
where t1 is the time when damage occurs and t2 magnitudes comparable, and finally weighted according
is some time after. to the importance the user gives to each one.
What has been described can be translated into
3. Fuel can be changed between tanks but it cant expression (11) that is intended to maximize:
be discharged overboard

k1 [STd (LCt )] + k 2
1 ( ) ( )
Ct TKiB Ct 1 TKiB
+ k3
Ct TK ( i
FO
) = constant (5) iWC Qi WQi iTK TQ


iFO (11)

4. Lub oil and miscellaneous tanks are not changed where k1, k2, k3 are scale/ weight factors t is the time
corresponding to the solution implementation and t-1 is
( )
Ct TKiLOM = constant i LO M (6) the time corresponding to damage.

5. When one of two compartments with a cross-

flooding pipe connecting each other is flooded, 5. GA IMPLEMENTATION
so is the other
A Genetic Algorithm (GA) is a programming technique
(
Ct WC j > 0 ) (
Ct WC j* > 0) (7) that imitates biological evolution as a problem-solving
strategy [5]. Given a specific problem to solve, the GA
evaluates potential solutions, a combination of the
where j and j* are two compartments connected
decision variables, encoded in some fashion
by a cross-flooding pipe
(chromosomes), using a user defined fitness function.
The algorithm begins by evaluating a set of candidate
6. No watertight compartments that were not
solutions (initial population), and then applies genetic
flooded due to damage are going to be flooded
operators such as elitism, mutation and crossover to
evolve towards new solutions (first generation). From
Ct (WCi ) = 0
1
Ct (WCi ) = 0
2
(8)
The Damaged Ship, London, UK

these candidate solutions, the ones with the highest (or in
this case lowest highest) fitness function values are kept
and allowed to reproduce new offspring, which then form
the next generation of candidate solutions. Those
candidate solutions which do not improve are not
considered for the final solution. The procedure is
repeated until a satisfactory fitness function value is
obtained or once a limit number of generations is
reached.
One of the GAs characteristics which makes it
appropriate for this kind of applications is that there is no
constraints as far as the complexity of the fitness
function, being as intricate and as large as required to
solve the problem in hand.
5.1 IMPLEMENTATION
values were passed directly to the next generation
(elitism); 10% of the parents suffered a uniform mutation
by randomly changing the values of some alleles of the
chromosome, and the other 80% where generated using
scattered crossover between parents values.
6. WORKED EXAMPLE
This DST architecture has been used to develop a
software for a ship, with length of approximately 80 m,
draught under 4.0 meters and operational displacement
over 1800 tons. The ship is subdivided by several
watertight bulkheads and up to main deck has 4
watertight decks. She has several tanks, of which 23 are
fuel, ballast, voids or fresh water (voids can also work as
ballast tanks if required).

The GA implementation for this specific problem will The software has been developed using MATLAB,
consider each combination of tank loads at time t as making use of the appropriate toolboxes for user
potential solutions (expression (12)), taking into interface, genetic algorithm implementation and OPC
consideration that flooded watertight compartments are standards interface [6]. This last ensures the
going to be dealt afterwards. Each chromosome has the interoperability between the software and the different
following structure: automation systems that read all sensors, following the
open connectivity standards for data access and exchange
[C (TK ) , C (TK ) , ... C (TK ) ]
t 1 t 2 t i (12) that are being developed and are incorporated to
Microsoft's OLE technology in Windows.
where i is the number of tanks considered and the values
Figure 3 shows the main interface window, where there
of the chromosome are real positive values between 0
are three distinct areas to check the ships hydrostatics, to
and 98.
identify stability alarms, and to read data of current
manoeuvring characteristics. As the vessel runs in
The initial population is going to be built independently
normal condition, there is no need for user intervention,
of the load condition considered by changing ballast and
since from time to time the new load condition is
fresh water tank loads, ensuring that they meet
calculated (based upon draught marks readings and tank
constraints of expressions (3) to (8) and their random
monitoring) and afterwards the strength, intact and
nature, which is necessary for the algorithm. In the
damage stability criteria are assessed. When any one of
studied case a population of 20 possible solutions has
those criteria is not met the appropriate check box turns
shown to be enough.
from green to red, as is shown in the figure.
These purposed solutions, or any other, before their
Stability calculation routines have been written and
fitness value is calculated, will go through an evaluation
validated using hydrostatic data generated by Autohydro
process against all constraints including damage stability
[7] and comparing the results from both software tools.
criteria (expressions (2) to (8)). If any is not verified, a
large number (bigpenalty) is added to the fitness value
Additional information can be accessed from the menu
of that solution, turning it unviable.
on the top of the window, and no other commands or edit
fields change is possible.
Only when all checks have been done, drainage
equipment is attributed by the flooded spaces using a
simple heuristic checking first if there are enough
equipment and then draining first the compartments with
better access, taking into consideration the already
mentioned constraints (expressions (9) and (10)). After
this, the possible solution score can be obtained through
the fitness function using the inverse of expression (11),
since GAs are design to find minima instead of maxima.

As far as how other candidates, beside the initial

population, are brought together, this implementation
used three genetic operators: elitist selection, mutation,
and crossover. The 10% chromosomes with the best Figure 3 Main interface window

To run simulations for different load conditions, there is
an appropriate interface window that calls a routine that
works as standalone. The interface represents the vessel
tanks where, intuitively, the user can fill up the edit fields
and, sometime after, the calculated hydrostatic data and
criteria alarms appear in the edit fields (figure 4).
Figure 4 Simulation interface window (watertight
subdivision and tank identification was erased)
6.1 DAMAGE CALCULATIONS
The damage optimization starts when the user selects the
affected area, and specifies the size and location of the
opening to the sea in the damage interface window
shown in figure 5.
Figure 5 Damage interface window (watertight
subdivision and tank identification was erased)
Such an action, as previously mentioned, calls a routine
that estimates the actual load condition immediately
before damage using sensors information, and
determines the compartments that are going to be flooded
taking into account all different openings and
connections between them.
Once the scenario is well defined, the optimization
process explained in paragraph 5 starts. As explained, an
initial population of 20 candidate solutions is generated
by changing the capacity of ballast and fresh water tanks.
The Damaged Ship, London, UK
After three generations, which stand for 80 damaged
stability calculations and takes about 1 minute (in our
non-optimized prototype implementation), the first
results are made available. The blue edit fields are
updated with the desired values, and drainage equipment
attribution is made. This is the immediate action
proposed by the DST.
The genetic algorithm will start again with the previously
defined initial population having replaced one of the
candidates by the result of the immediate action
solution. This can take as many generations as required,
though in our tests only 50 were used.
Within the genetic algorithm, the operators available in
MATLABs toolbox were applied, taking into
consideration the percentages previously defined in
paragraph 5.1. Therefore, the two best solutions were
passed into the next generation, and constrained random
crossover and mutation operators were also used. We
conducted tests with 4 different initial load conditions,
and for each of these 4 different damage areas. In each
case we repeated the procedure several times, using
different number of generations.
Generically, as the number of generations increases the
fitness values are smaller and therefore the solution is
better. Nevertheless there are cases where this did not
occur and the method suffered from premature
convergence into a local minima. The result was that a
fairly good fitness function was found rather quickly (3
generations), but no improvements were obtained after
that. On the other hand, there were runs that took a long
time to converge to good solutions, but there where in
fact better than the first ones.
One interesting fact that shows what has been stated, is
the comparison between the 90 straight generations with
the last 50 generations of the two step process of Load
case 4 when areas B1, B2 and H1 are damaged. As
shown in figure 7, when using straight generations (i.e.,
running the GA only once), the system finds a good
solution very quickly (3 generation), but then evolves
very little. It is a clear case of premature convergence,
because if we stop the process and re-initiate it with new
initial solutions, but keeping the previously obtained best
one (i.e. use the two step process), the system continues
to find ever better solutions for some more time, reaching
a far better solution.
The Damaged Ship, London, UK
Figure 7 Genetic algorithms progress comparison for
LC4 and damaged areas B1, B2 and H1
In our point of view this fast convergence is desirable in
this case, because we need a good solution in very little
time. It proves that good answers can be obtained quickly
and the proposed method works as genetic algorithms
should. On the other hand, this also means that more
effort has to be applied into studying and developing
dedicated genetic operators for this particular application.
In the second step of the two step process, less pressure
can be put on minimizing the cost function, allowing a
better long term solution to be found.
7. CONCLUSIONS
The work done presented a tool that may be used to
support the decision of the crews as far as how to
respond to damage. This tackles a real issue since no
standard response to damage is possible, and so any
decision considering the vessels stability is prone to
human errors.
It was also shown that the response to damage problem
can be interpreted as an optimization problem and that
genetic algorithms are a feasible method to solve it.
Finally, all of this was supported with an analysis of a
real application that has been validated.
8. REFERENCES
1. Plumley, Damage-Control Tactics in Extreme
Stability Situations, Sea and Shore, 2009
2. Evangelos et al, Optimisation of the Survivability of
Naval Ships by Genetic Algorithms, COMPIT04,
Sigunza, Spain, 9-12 May 2004
3. IMO Resolution A.748(18), Code on Intact Stability
for all types of ships covered by IMO Instruments, 1993
4. Sarchin, T.H. and L.L. Goldberg, Stability and
Buoyancy Criteria for U.S. Naval Surface Ships,
SNAME, New York, 1962
5. Fogel, D.B., Evolutionary Computation : Towards a
New Philosophy of Machine Intelligence, IEEE Press,
1999.
6. Schleipen, M., OPC UA supporting the automated
engineering of production monitoring and control
systems, 13th IEEE International Conference on
Emerging Technologies and Factory Automation 2008.
ETFA 2008
7. Autoship systems Corporation, Autoship Systems
Corporation Software: The Advanced Approach to
Vessel Design (Editorial), June 2005.
9. AUTHORS BIOGRAPHIES
Paulo Triunfante Martins is a naval architect and
structural surveyor of the Portuguese Navy inspection
team following the OPV and Pollution Control vessels
program at Viana do Castelo Shipyards. He also
collaborates with the Portuguese Navy Research Centre
(CINAV) in the development of artificial intelligence
applications for naval architecture and shipbuilding.
During his time at sea, he held several posts aboard
Portuguese Navy frigates and corvettes, from bridge
officer to engineering and damage control officer.
Victor Lobo is an Associate Professor at the Portuguese
Naval Academy, deputy director of CINAV (Portuguese
Navy Research Centre) and invited Associate Professor
at ISEGI-UNL (Statistics and Information Management
Institute of the New University of Lisbon). He has been
invited professor at the Faculty of Science and
Technology of the New University of Lisbon, at ISCTE
(Lisbon), and at the French Naval Academy. His main
research interests concern self-organizing maps, neural
networks, unsupervised learning, clustering, mobile
robotics, and geocomputation.
The Damaged Sh@,London, UK

REFLOATING AND STRENGTH ASSESSMENT OF GROUNDED SilrIPS

P Mangriotis, London Offshore Consultants Limited, UK

SUMMARY

In the aftermath of a grounding casualty, naval architects are called upon to assess the ships damaged characteristics in
support of owners, their insurers and/or their salvors. The author reviews the practical steps employed, focussing on the
field work involved and based on typical case studies, primarily of grounded bulk carriers. Data gathered from surveys
and investigations around the casualty, however patchy and conflicting, can be incorporated into a hydrostatic model to
reflect the casualty situation, often with pressing time constraints. The preparation of a model to calculate floatation and
strength and the minimal data requirements to do so will be described. The naval architect must properly account for the
degree of flooding of cargo holds and other spaces. The methods for calculation of ground reactions are outlined and the
effect of assumptions on the distribution of contact points and on seabed bearing capacity and elasticity discussed.

how further damage and flooding will affect the casualty.

1. INTRODUCTION
1.1 MARINE CASUALTY RESPONSE
London Offshore Consultants (LOC) provide techntcal
advice and expertise at all stages of the marine claims
and disputes process including marine accident
investigation. LOC personnel have given expert
evidence at litigation proceedings and provided expert
evidence at formal enquiries.
There are two main routes in which LOC may become
involved in casualty response. Firstly, in the immediate
aftermath of a casualty, it may be necessary to attend the
scene to provide immediate practical support, to the ship
master and/or owner, or to the salvor. Secondly a request
to provide back-up support from the office may be made
by any of the parties. In any event LOC may be tasked
to provide situation analysis on a practical level and on a
technical level.
LOCs involvement on-site in practical casualty response
has been well documented elsewhere [l, for example]
and is not the subject of this paper. The on-site
personnel are responsible for determining the initial
scene description including an initial assessment of the
casualty, identification of interested parties and the
environmental conditions. An early assessment of the
initial information will be prepared either by the on-site
naval architect or the office back-up. This will address
cargo and tank conditions, hull condition, draughts, water
depths, tides and weather.
They need to know how much weight to remove to
enable the ship to refloat and which spaces to pump out.
Will the refloat condition compromise the ships residual
strength and will the ship be stable once refloated?
The refloating operation should avoid worsening the
casualty situation. Calculation of the global longitudinal
strength and the likely local loads together with an
assessment of the residual structural capacity is required
at each stage. The techniques and typical assumptions
involved are outlined below, together with refloating
techniques employed by salvors in recent operations.
2 HYDROSTATIC MODELLING
2.1 GHS PROGRAM
GHS (General Hydrostatic) by Creative Systems is an
industry standard program used extensively in the
salvage industry. The program can analyse any ship
orientation without limitation of trim and heel angles,
from upright to upside down. Hull models may be
constructed rapidly from minimal information.
Approximate dimensions and a general knowledge of the
hull form can be sufficient to produce a useful model at
the outset of a project. Hull models may be imported
from various data formats. Damage limited by
compartmentahsation and damage in the form of hull
deformations are readily simulated. GHS computes the
weights and centres of all contained fluids at any given
ship orientation. Partially flooded spaces are thereby
accurately represented.

As the situation progresses and more data becomes 2.2 DATAREQUIREMENTS

available, the scope of calculations required will become
clearer. The required calculations may address ground
reaction, longitudinal strength and stability. The salvor,
ship mastedowner and the various insurance interests
need to know whether the current condition is stable and
The requirements for creation of a hydrostatic model
suitable for refloating and strength assessment will
depend on the degree of accuracy required. For

example, if no compartments are flooded, tank or hold
contents may be modelled as masses rather than liquids;
or if there is no concern about longitudinal strength then
the lightweight distribution may be omitted and replaced
by a point mass.
The following plans are typically requested as soon as a
requirement for hydrostatic modelling is identified.
0 Body Plan, Lines Plan or Table of Offsets
0 General Arrangement and / or Capacity Plan
0 Trim and Stability Book / Loading Manual includmg
lightweight distribution
Tank Sounding Tables including location of sounding
tubes
If a structural model is also required then the following
plans would be requested.
0 Midship Section, Construction Profile and Decks
plan, Shell Expansion
2.3 INCOMPLETE DATA AVAILABILITY
Often during a grounding casualty it is not possible to
obtain a full set of drawings at short notice. Typically,
lack of data may be due to owners not possessing the hull
form or the lightweight distribution not being included in
the loading manual.
A hull form or complete model may exist for a sister ship
which can be used - LOC hold a library of typical
designs on record for that purpose. If not, a similar ship
may be used for which the dimensions could be scaled to
match. Alternatively the hull form can be built up by
taking offsets from waterlines, profiles or from the shell
expansion. If the hull form is approximated in these
ways it is unlikely that the model will give accurate
longitudinal strength results, however it should be
adequate for calculating trim and stability.
In the absence of a lightweight distribution an
approximation can be built up either by methods such as
spreading the lightship according to a coffin diagram and
adjusting the trapezoidal sections to match the LCG or by
a station coefficient where a first estimate of the parallel
midbody mass per unit length is made and then point
masses are included derived from published initial design
formulae. Manipulation of the data is carried out in a
spreadsheet.
The Damaged Ship, London, UK
2.4 MODEL VALIDATION
Prior to using the model the hull hydrostatics at a number
of waterlines and tank capacities and centres are
compared against the values provided in the Trim and
Stability Book. IACS standards [2] provide values of
convergence which can be used as a guide to assess
whether the model is satisfactory.
A validation case against the last known loading
condition, usually the ships last departure port, is then
run. It is preferable to achieve correlation with the
results of the draught survey on departure if carried out
rather than the output draughts from the ships loading
computer. If tanks are to be loaded based on ullage or
sounding then the validation condition should reflect tank
fill capacities rather than corresponding mass of liquid.
3 CALCULATIONS
3.1 GROUND REACTION CALCULATIONS
Prior to refloating, an assessment of the ground reaction
will be required in order to assess:
0 the amount of towing force that would be required to
pull the ship from its stranded location,
whether the ship would float free in a high tide,
the likelihood of bottom damage having occurred, in
conjunction with motions and sea bed type, and
0 the ships stability and longitudinal strength and
condition on refloating.
In assessing ground reaction it is vital that accurate
draught cuts are taken. Ideally, as many as possible
should be recorded at various states of tide.
An immediate estimate of the ground reaction can be
made without the hydrostatic model. The simplest
method is to compare the displacement prior to
grounding, with the displacement corresponding to the
draughts of the ship aground. If holds or tanks have been
flooded then the lost buoyancy, which can be estimated
using the capacity plan, must be subtracted from the
grounded displacement.
On a large ship, a substantial ground reaction in terms of
the towing force required to refloat it, may only represent
a small proportion of the total displacement. Care should
be taken to correct the displacements for hull girder
deflection according to the measured draughts or
estimated bending moment. The difference in
displacement method indicates the magnitude but not the
location of the ground reaction. If the grounding has
 The DumugedShip, London, UK
resulted in a large change of trim and the location along
the ships length of the grounding point can be assessed,
then a reasonable estimate of the ground reaction can be
made by assessing the force required to produce the
observed change of trim.
An urgent task once aground will be for the ships staff
to take soundings of all tanks and holds in order to
determine flooded compartments. This survey will
reveal which tanks are intact, slowly flooding or open to
the sea. The naval architect will use this survey together
with the departure condition as a basis for calculating the
condition aground. The structural condition of the hull
girder, in particular whether the ship is in a severely
hogging or sagging condition may be ascertained by
inspection noting any abnormalities on the deck such as
buckling of handrails or excessive displacement of
expansion joints on deck piping.
If the ship is carrying dry bulk cargo or containers and
cargo holds are flooded it is important to assess the hold
and cargo permeabllities accurately. Useful data can be
obtained from standard references [3, 41 for dry bulk
cargoes or a test can be performed on the spot using a
sample of dry cargo drawn from the holds with a bucket.
For container holds the permeability must take account
of the gaps between containers and the contents of the
containers themselves which will inevitably become
flooded over time.
On anythmg other than a hard rocky seabed it is likely
that the grounded ship will begin to create a trough and
sink into the seabed. Evidence of thls talung place would
be obtained by taking soundings around the ship using a
plumb line or an echo sounder mounted on a small craft.
An evident difficulty in refloating presents itself in such
cases as not only does the ship need to be refloated
sufficiently to overcome the ground reaction but also the
draught wdl need to be sufficiently shallow to allow the
ship to be manoeuvred over the surrounding seabed.
With the ship lying in a trough the ground reaction can
be assumed to be evenly distributed over its length or
alternatively proportional to the plan area of the flat
bottom. However if there are currents at the grounding
location a muddy or sandy seabed may be scoured
around the fore and aft ends which can lead to a
concentration of grounding force at midships and a
higher hogging bending moment. An accumulation of
silt or sand may build up on one side of the ship with
scouring occurring on the opposite side leading to a
gradual increase in the ships list.
3.2 MODELLING GROUND REACTION
GHS Provides a very and realistic ground
reaction mechanism. It has the option of including an
array of grounding points with each point having its own
stiffness. Therefore, it can model any grounding
scenario or type of soil. It can also easily model
groundings with restricted heel. The locations and
magnitudes of the forces exerted by the ground are easily
determined.
The grounding area can be modelled in GHS by attachmg
a number of grounding points to the hull at the known
location of contact. The points possess a rigidity
representing the elastic properties of the hull bottom and
the compressibility of the seabed. The grounded
condition is modelled by fixing the trim, draught and
heel at the observed draught cuts and then releasing the
model to obtain an equilibrium. Using this method
iteratively and adjusting the initial trim, heel and draught
and the rigidity of the grounding points as required one
can arrive at an equilibrium condition corresponding to
the observed draught cuts. The depth of the seabed is not
an input to this calculation.
Tide can work in favour of refloating if the ship has gone
aground on a low tide but the opposite holds true if the
ship has gone aground near high tide. A falling tide may
leave the ship susceptible to much higher grounding
forces and this effect may be exacerbated by the cycle of
spring and neap tides. If the ship is grounded on a
pinnacle or short length it may trim or heel considerably
as the tide rises and falls and thus the floatation
characteristics cannot be simply ascertained by raising or
lowering the waterline according to freeboard
measurements at a location on the ships side. The tide
has to be adjusted relative to a datum at the centre of
ground reaction to prevent errors. In a situation where
the observed trim or heel is changing during the tidal
cycle the model can be refined to match this by adjusting
grounding point flexibility and location. Confirmation
that an accurate model of the ships hydrostatics and
grounding can be obtained if the calculated draughts
consistently correspond to the observed draughts.
In addition to the effects outlined above, for seabed types
such as mud or silt a suction effect may occur such that
additional buoyancy is required to refloat the ship. The
breakout force required to dislodge the ship will depend
on the area in contact and the cohesion of the mud and
could result in the force required to lift the ship being
more than double what it would be otherwise. Suction
can be kept to a minimum by the application of a lifting
force over a long period of time or by lifting the ship out
from one side or end so as to allow drainage paths for
water to enter the gap between the mud and hull.
3.3 LONGITUDINAL STRENGTH
CALCULATIONS
GHS will output the bending moment and shear force
distribution in the grounded condition including the
 The Damaged Sh@,London, UK

effect of heel. In calculating longitudinal strength the

advantage of using more grounding points will be to
produce a smoother shear force curve, however, a greater
number of grounding points will makes it more difficult
for GHS to converge on a solution. Usually not more
than 20 to 30 points should be sufficient to model a
grounding surface across and along the hull bottom
Case 1 Case 2
By default GHS assumes that each grounding point will
have a stiffness in the region of 1,000 tonnes per cm and
the grounding stiffness can be varied in order to calibrate
to the observed draughts or heel angles at varying states
of tide. The principal effect of grounding stiffness is to
distribute the overall grounding force in different I I

proportions to areas of the bottom. Adopting a softer Case 3 Case 4

grounding stiffness will in general reduce bending
moments and shear forces as the vessel becomes more
evenly supported.
Case
GHS will output the hull grder deflection if an 1 - Intact 9.8 (100%) 16.4 (100%)
appropriate section modulus for the hull midship section
is defined. This can be useful where the change in
buoyancy distribution caused by hogging or sagging
needs to be considered. An example of this was a
capesize bulk carrier grounded at the bow and the stem
across a channel. The deflection at midships was
calculated to be 82cm. When the effect of hull girder In order to ensure an appropriate margin of longitudinal
deflection was taken into account the sagging bending strength, as a rule of thumb, the calculated bending
moment was reduced by 16% and the shear force was moments and shear forces aground should be maintained
reduced by 12%. The effect of hull girder deflection is below seagoing permissible limits. However in sheltered
always to reduce the maximum bending moment. waters the limit may be too onerous and it may be
justified that the harbour permissible limits with a
If the ship begins to respond to waves whilst aground it suitable reduction for damaged structure are more ,

can be expected that increased bottom damage will
quickly begin to occur, especially if the ship is lying on a
hard seabed. It is prudent and common practice in such
cases to increase ground reaction by ballasting the ship
down using all available ballast tanks and holds so as to
prevent or minimise ship motions until the waves have
abated. Small cracks may propagate into the tank top
plating of cargo holds due to grounding pressure on the
double bottom structure leading to ingress into the cargo
holds and complicating the task of refloating.
Once the bottom becomes damaged the ships
longitudinal strength will be compromised to a degree.
Frequently one cannot ascertain the degree of damage to
the bottom shell and internal structure of the double
bottom because these areas are inaccessible, for example
when embedded in mud or the cargo hold flooded and
full of cargo. The examples below show the reduction in
section modulus of the hull girder of a bulk carrier due to
particular assumed extents of damage to the bottom shell,
associated longitudinal structure, and cargo hold
structure. It is noteworthy that a substantial amount of
bottom section modulus remains in Case 2 even
assuming the entire bottom has been damaged.
appropriate as there would be no waves present.
Alternatively the seagoing permissible limits can be
reduced by a service factor of say 0.5 applied to the wave
bending moment contribution.
3.4 REFLOATING SEQUENCE
The sequence of steps to refloat a ship will be determined
taking account of practical factors such as the availability
of pumps and lifting equipment, the degree of damage
and prevention of pollution. Essentially three options
exist to regain buoyancy: intact tanks can be pumped out,
damaged tanks can be pressurised with air and cargo can
be discharged. A suitable sequence would be drawn up
by the naval architect and salvage master.
A variety of techniques have been developed by salvors
for plugging underwater breaches ranging from the
insertion of wooden chocks into cracks and filling the
remaining cracks with epoxy putty, using blanking plates
or conducting weld repairs inside a cofferdam or
underwater.
 The DumugedShip, London, UK
Another approach to regaining lost buoyancy in bottom
damaged tanks is to pressurise the tank through the use
of compressed air, expelling water through the
penetration in the hull bottom. Care needs to be taken in
such an approach so as to prevent over pressurisation of
the deck structure. This technique is particularly
effective on bulk carriers with combined double bottom
and topside tanks although the limit of pressurisation will
usually be dependent on the strengths of the topside
sloping plating which tends to be the weakest stiffened
structure on the tank.
Cargo and/or bunkers may be discharged to lighten the
ship, although this will depend on the availability of
receiving tonnage or shore facilities. The refloating
sequence should preferably ensure that the final step
includes the dewatering of a substantial ballast tank in
order to overcome ground reaction in one step. The final
moment at which ground reaction reaches zero can be
difficult to predict due to the uncertainties and
approximations made initially to calculate the overall
ground reaction.
An external lift such as from a shear leg crane or barge
equipped with lifting gear may be also be used. This
would be modelled as a lifting force at its point of
application. Such a lift can be useful where there is
concern that the vessel may pass through a point of
instability at the moment of refloating.
If a number of cargo holds are flooded then stability
when refloated can be of concern due to the free service
of liquid in holds, although on a bulk carrier most of the
holds need to be flooded before this would become a
concern. As an example a handy size bulk carrier
possessed positive stability even when four out of five
holds were flooded.
4 FORENSIC ANALYSIS
LOC are regularly requested to conduct retrospective
investigations into the risks involved during the salvage
operation to assist in the apportionment of the salvage
award. Such investigations are made without the same
time pressures and afford an opportunity for more
detailed and accurate analysis.
In such a case there would normally be access to all the
salvors records such as draught measurements, soundings
and any actions involving the movement of masses
onboard the vessel.
A number of scenarios would be investigated such as
what may have happened if a particular hold had become
flooded or there had been an inability to dewater a
particular hold. Thus an opinion may be arrived at with
regard to the difficulties and risks associated with
refloating the vessel.
5 CONCLUSIONS
The approximate nature of refloating calculations and the
time constraints under which they need to take place
should be evident. The use of the hydrostatic model
allows the calculation to be refined to a degree where an
entire refloating sequence can be assembled and verified
promptly. The naval architect cannot operate in isolation
from the practical realities experienced by the salvage
personnel on site and his input allows the salvage
personnel to plan a sequence appropriately.
6 REFERENCES
1. HOLLOWAY, P S, Managing Maritime
Accidents, Lloyds Maritime Academy, July
2008 and subsequently
2. IACS, Recommendation No 48 for Loading
Instruments, Unified Requirement URL 5 -
Onboard Computers for Stability
3. Thomas Stowage - The Properties and Stowage
of Cargoes
4. IMO BC Code - Code of Safe Practice for Solid
Bulk Cargoes
7 AUTHORS BIOGRAPHY
Paraskevas Mangriotis is a Director and Head of Naval
Architecture at London Offshore Consultants Limited.
He is a Chartered Engineer, a member of the Royal
Institution of Naval Architects. He holds responsibility
for managing the London based naval architecture
capabilities of LOC and is particularly experienced in
ship structures and classification matters. He has
provided naval architecture consultancy services for P&I
clubs, Iegal fxms, shipowners and huII insurers and has
appeared as an expert witness at arbitration hearings. In
salvage operations he has undertaken assessments of
salvage risks and dangers as part of the salvage
arbitration process. He has provided salvors with on-site
and office based emergency response advice on strength
and stability aspects and has conducted bid reviews of
wreck removal operations. Paris has also worked on
casualty and other technical investigations and carried
out marine warranty surveys including load out approvals
and site surveys for offshore oil production equipment
and verification of transport calculations.
The Damaged Ship, London, UK
DAMAGE DECISION BOOKS FOR USE BY MASTERS ON VESSELS WITH NO
COMPUTER BASED LOADING SYSTEM
M Simpson and I Wallbridge, Hart Fenton & Company Ltd, UK
SUMMARY
This paper describes background thinking and production of damage decision support books to provide ships masters
with the relevant information required to make a sound judgement and decision in the event of an incident. The paper
compares the options currently available to ships and goes on to describe the paper based approach that Hart Fenton
have used for a number of clients vessels.
1. INTRODUCTION
It is hoped that ships undertake their day to day work
without any incidents. However in the event of a vessel
sustaining damage due to an accident the master has to
make a decision on the best course of action in as short a
time as possible to mitigate the risk to life and property.
The decision made will have a direct bearing on the
outcome of the incident and to a lesser extent the time
taken to make the decision can also affect the outcome. It
is important therefore that the master is provided with
accurate and relevant information with which to make
their decision.
There are a number of methods that can be used by a
ships master to evaluate the state of the vessel after
damage. All are valid options depending upon the type of
operation the vessel is engaged upon.
Ultimately however it is the master who will make the
final decision therefore all options are essentially
guides that assist in the decision making process.
2. AVAILABLE OPTIONS
2.1 LOADING COMPUTERS
Many vessels use loading computers to check their
stability prior to departure. These systems are ever more
capable and can be used to evaluate stability and strength
in the event of damage.
2.1 (a) Advantages
These systems can be particularly useful for vessels
that have many variables including a large range of
loading conditions or many compartments.
Information is available very quickly to the master.
It is possible to run what if analysis to check
remedial actions or further flooding scenarios.
Can be set up to interrogate tanks remotely.
2.1 (b) Disadvantages
If not used regularly errors in use may occur.
Requires training of personnel in what can be
complex computer programs
There is a tendency to over-rely on computers when
under pressure.
Potential for errors or failures where linked to the
vessel power supply and systems.
2.2 EMERGENCY RESPONSE SERVICES
A remote emergency service with dedicated personnel
who operate daily with the software used is an attractive
option.
2.2 (a) Advantages
Can run detailed what if scenarios
No software to learn for ships crew.
No risk of failure due to errors/damage on the ship.
2.2 (b) Disadvantages
Can be costly.
On certain services the response time may not be
acceptable.
Remote staff need to have detailed knowledge of
vessel.
Delays due to relaying of information, etc.
Relies on the accurate transfer of information such
as tank state, loading condition, etc. between vessel
and response team.
2.3 DAMAGE DECISION SUPPORT BOOKS
A damage decision support book is a pre-prepared
document where all survivable damage conditions are
laid out in a simple, clear and concise manner such that
the master can make a decision on the best course of
action in the event of an incident. Similar systems are
used by military vessels.
2.2 (a) Advantages
Simple to use.
Information is available very quickly to the master.
Reliable (could be used as a backup to other
systems)

A detailed decision support book may be more
suitable for vessels with limited variables or where a
response is needed quickly, e.g operating in confined
waters, short sea ferries, etc although as mentioned
highly complex naval vessels use a similar approach.
2.2 (b) Disadvantages
Large number of variables leads to much larger and
more complex book.
Cannot easily cover every possible loading
condition.
3 DECISION MAKING PROCESS
In the event of an incident the vessels master will need
to make a decision based upon the available information
at the time. The speed at which the decision is made can
have a significant impact on the outcome of the incident.
Quickly Ascertain Current Situation gather
relevant data
What has happened
What/how much is damaged
Is there water ingress and where.
Speed of water ingress/quantity of water.
Understand Implications process data
Will the vessel survive current situation
What might happen if the situation changes
Can anything be done to improve the situation
Can the vessel progress to a safe haven
Make Informed decision act upon data
Progress under own power.
Undertake remedial actions
Call for outside assistance
Mayday & Abandon Ship
It is in understanding the implications of the situation
that Damage Decision Books can be of great value.
4 DAMAGE STABILITY ASSESSMENT
Standard Damage Stability Books carried on board ships
are produced to meet statutory requirements. These
requirements cover a broad range of conditions but not
all eventualities. They show that the ship is safe to
operate within the expected operational range.
The use of probabilistic damage analysis provides no
meaningful information to the master on any particular
damage scenario.
A Damage Decision Support book aims to fill in the
gaps, covering more unexpected or serious incidents.
In preparing a damage decision support book a much
wider range of damage conditions will be evaluated
including multiple adjacent compartments, certain non
The Damaged Ship, London, UK
adjacent compartments, Pipework, jet duct or
bowthruster tube failures may also be considered.
Even in a small vessel the number of combinations that
can be evaluated is surprisingly high. The fact that the
condition the vessel is in does not comply with any
statutory damage criteria is no longer relevant, whereas
whether the ship is safe or in immediate danger.
4 METHODS
4.1 INITIAL CONDITIONS
Prior to undertaking any analysis it is necessary to define
the most relevant initial condition(s). Too many initial
conditions will result in an unwieldy document. However
it is necessary to cover as many possible scenarios as
possible.
In the work HF have done to date we have tended to use
three initial loading conditions, agreed in advance with
the client. They tend to be a full load condition a light
load and a half load condition. These are likely to vary
depending upon the operation of the vessel.
4.2 DAMAGE CASES
In conventional deterministic damage analysis the
quantity of damage cases evaluated are dictated by the
relevant codes. However, the intention of a damage
decision support book is to expand upon these statutory
cases so as to ascertain the limits of damage that the
vessel can withstand.
It is important to consider conditions that may occur,
however unlikely they may appear. As previously
mentioned this may include machinery or systems
failures, unexpected damage such as leaking manholes
and bulkheads, etc. Analysing symmetric damage to port
and starboard demi-hulls of a catamaran may be of use in
providing guidance to remedial actions.
4.3 ANALYSIS
HF use NAPA to evaluate all the conditions before
exporting this into excel and Autocad to format into a
suitable document.
To date when undertaking the analysis intermediate
stages of flooding have not been considered, neither have
weather or shift of cargo. However it would be easy
enough to include these items if required by the client.
5 PRESENTATION & EXAMPLES
The presentation of the information is critical to the
successful use of damage decision support books. The
The Damaged Ship, London, UK
information must be relevant, accurate and presented in a
clear and concise manner.
5.1 CONDITIONS TO INCLUDE
Whilst the naval architect will have produced many
conditions to arrive at the final book for the sake of
clarity only those that can be survived tend to be shown
in detail. However this is not to say that un-survivable
conditions are considered irrelevant, as it is equally
important that the master knows the full range of
survivability for the vessel. Those conditions that arent
survivable should be noted in the damage decision
support book.
5.2 CONDITION MATRIX & SEVERITY
A solution to the above problem of how best to present
the damage cases is by use of a simple colour coding for
damage severity. To date we have used four levels of
severity, green, amber, red and black where green
indicates a minor incident and black indicates an
unsurvivable incident.
Hart Fenton have produced a number of books where the
damage conditions analysed are presented as a matrix, at
the start of the document using this colour code (see
Figure 1 below). This immediately indicates to the ships
master the severity of the incident at hand. From the
matrix it is possible to identify the page on which the
more detailed information is found.
Figure 1 Typical Damage Severity Matrix
The criteria used to generate the colour code may vary
with ship type, however typically a combination of heel
or trim, freeboard and residual stability would be used.
5.2 CONDITION DETAIL PAGE
On the page detailing the individual condition the
relevant information presented is likely to vary
depending upon the type of craft, the operators
requirements, etc. Ideally the presentation of the relevant
information should be as straightforward as possible. As
much information should be presented graphically as
possible. See Figure 2 below.
Figure 2 Example Condition Detail Page
As a minimum we would suggest that the floating
position should be shown, with the waterline horizontal
to indicate the true attitude of the ship both longitudinally
and athwartships.
A list of the flooded compartments to confirm that
condition is that being experienced.
A list of those critical points (such as downflooding
points, vents, embarkation points, etc.) should be shown
with a height to the waterline.
The actual heel and trim angles may be of use
machinery or equipment limitations may apply.
A comparison between the intact and damaged stability
curve gives a indication of the severity of the situation.
6 RELEVANT INFORMATION REQUIRED
In the event of an incident the master will need certain
information prior to making a decision on the best course
of action.
Firstly they will want to know whether the vessel will
survive the identified damage scenario. They will then
want to know whether the current situation may get
worse with progressive flooding, etc. or whether it is
possible to improve the situation by cross-flooding or
similar.
 The Damaged Ship, London, UK

The fact that the condition that the vessel is in does not 300 tonne heavy lift vessel during cargo operations. He
comply with any particular damage criteria is irrelevant. carried out a damage survey and technical investigation
However knowing that the vessel is safe or otherwise is a of major collision and fire damage for salvage
valuable piece of information to have. arbitration.

Iain Wallbridge is a Senior Naval Architect with more

7 FURTHER DEVELOPMENT & than 18 years marine industry experience. Since joining
REFINEMENT Hart Fenton he has worked on a number of projects
including undertaking structural calculations and writing
Expand to cover more vessel types method of procurement specification for a crane barge, structural
presentation for multiple cargo types in product tankers calculations for a Ro-Ro ferry conversion and detailed
for example. damage decision support books for a new build Ro-Ro
ferry and a large passenger catamaran. He has also
Remedial actions developed the decision support book format now used at
Intermediate flooding Hart Fenton. Previous experience includes extensive
Progressive flooding stability analysis of a number of lifeboat classes and
Weather and other heeling levers undertaking damage inspections and detailed repair
Formatting within NAPA specifications for FRP, steel and aluminium lifeboats and
Probabalistic overseeing of subsequent repairs.

8 CONCLUSION

In an ideal world damage decision support books will sit

and gather dust. However in the event that an incident
occurs then they need to be immediately available and to
contain pertinent information to guide the master at a
stressful time.

Whilst not suitable for all vessel types they have a part to
play in vessel safety and if produced correctly in
conjunction with the ship operator they can provide a
useful reference document for the master.

9. ACKNOWLEDGEMENTS

Matthew Patey of NAPA Ltd and NAPA Technical

Support for their assistance when using NAPA software.

Capt Kane Taha of Isle of Man Steam Packet Company

and Capt Elwyn Dop of Wightlink Limited for their
constructive comments on support book layout.

10. AUTHORS BIOGRAPHIES

Mike Simpson is the Managing Director of Hart Fenton

appointed in February 1999 and a Naval Architect with
33 years experience in the commercial shipping industry.
He is responsible for all company matters, feasibility
studies, design projects, construction supervision, project
management, monitoring of ship trials and advice to
lawyers. Previous experience includes damage stability
analysis of new design and existing vessels and
specifically he provided technical advice as follows. He
provided immediate technical advice to chief engineer
when a high speed craft with 157 passengers on board
suffered a major three compartment collision. Technical
expert to Canal Authority in the alleged grounding of a
The Damaged Ship, London, UK

EXPANDING THE SCOPE OF RAPID RESPONSE DAMAGE ASSESSMENT (RRDA)

A CLASSIFICATION SOCIETY PERSPECTIVE
G Wang, American Bureau of Shipping, USA

SUMMARY

Shore-based rapid response damage assessment (RRDA) programs have been available to industry for almost 20 years
and have become a key element in responding to tanker casualties. Recent high-profile marine incidents have shown
that the pressures placed on an owner in the aftermath of a casualty have grown significantly. In response to these
changes, ABS has recently implemented a significant enhancement to the RRDA service provided by the classification
society.

This paper highlights the major enhancements of the ABS RRDA program which include enrollment of eligible vessels
at the time of vessel delivery, expanded RRDA service to include both initial incident response and subsequent transit
voyage support and enhanced engineering analyses to estimate loads, global and local strength.

A key feature of this expanded service is the introduction of an integrated software system that simplifies RRDA
enrollment at the time of delivery. By introducing vessels into the RRDA program at the outset, the classification
society can readily access a wider range of technical information relevant to the vessel such as loads and strength
calculations that can be used in providing a more detailed response in the event of a casualty.

The application of advanced design analysis tools to the RRDA decision process has made it possible to perform timely
calculation of hull girder ultimate strength and local buckling and the vessels ultimate strength in the damaged
condition. Further, by using the extensive global wave data held by the society, ABS is able to quickly evaluate the hull
girder and local strength for the planned voyage to a repair facility.

LIST OF SYMBOLS AND ABBREVIATIONS casualty have grown significantly. In addition to the
traditional demand for stabilizing the situation, cargo
CSR IACS Common Structural Rules spills and other environmental impacts have also become
ESF Environmental severity factor an important concern to the owner, as well as many other
Mu Hull girder ultimate strength stakeholders.
RRDA Rapid Response Damage Assessment
SEAS ABS Sea Environment Assessment System In response to these changes, ABS has recently
SWM Still-water bending moment implemented a significant enhancement to the RRDA
WBM Wave-induced bending moment service provided by the classification society. This paper
Factor for adjusting predicted sagging capacity highlights the major enhancements which include:
Environmental severity factor
s Partial safety factor for SWBM Enrollment of eligible vessels at the time of
w Partial safety factor for WBM vessel delivery
u Partial safety factor for ultimate strength Expanded RRDA service to include both initial
incident response and subsequent transit voyage
Enhanced engineering analyses to estimate
1. INTRODUCTION loads, global and local strength

Shore-based rapid response damage assessment (RRDA)
programs have been available to industry for almost 20
years and have become a key element in responding to
tanker casualties.
The main focus of RRDA has been to stabilize the vessel
after an incident and provide information for use in
salvage options including lightering the cargo. The
concerns have been vessels trim, stability and
longitudinal strength. Examples of these traditional
RRDA concerns are shown in Figs. 1 and 2.
Recent high-profile marine incidents have shown the
pressures placed on an owner in the aftermath of a
Further, this paper will introduce the technical
background of the enhanced engineering analyses and
describe the new RRDA software that can be quickly
applied to support informed decisions in the aftermath of
an incident.
2. RRDA SERVICE AND ENHANCEMENT
2.1 RRDA SERVICE
An essential element of the ABS RRDA program is that
ABS will provide a response team when the service is
activated by the client. The members of the team

understand their responsibilities and have collected and
reviewed the technical information likely to be needed in
the event of an incident, thereby saving critical hours of
data preparation and testing.
The ABS RRDA team includes engineers who can
support the client in the aftermath of an incident by
performing strength and stability calculations using
information provided by the Client.
The ABS RRDA service is available and can be called
upon 24 hours a day.
Figure 1. The vessels loss of stability may become
critical.
Figure 2. Loss of structure due to explosion may result
in the ship breaking in two.
2.2 ABS RRDA GUIDE
In July 2010, ABS published the Guide for Rapid
Response Damage Assessment to provide
owners/operators with a description of the process for the
issuance of the RRDA notation to an ABS-classed vessel
(ABS 2010).
This Guide contains the technical requirements and
criteria that meet international regulatory requirements
for vessel emergency response. The ABS RRDA
program provides emergency technical services for
owners/operators whose enrolled vessel experiences an
The Damaged Ship, London, UK
incident that may affect the stability or structural strength
or who may require the rapid provision of technical
analytical services.
ABS-classed vessels in compliance with these guidelines
will be assigned a class notation RRDA, Rapid Response
Damage Assessment, and will receive an enrollment
certificate from ABS.
See Appendix A for the table of contents of this ABS
RRDA Guide.
2.3 RRDA ENROLLMENT
A key feature of the enhanced RRDA service is to enroll
every ABS-classed newbuild tanker, bulk carrier,
containership, gas carrier and tank barge at the time of
delivery for a one year period. Vessel owners may opt
out of the program at any time, and may continue
enrollment after the first year for a fee. Other vessel
types, offshore units or existing vessels may also enroll
for a fee provided all required drawings and data are
submitted.
By enrolling vessels into the RRDA program, ABS can
readily access a wider range of technical information
relevant to the vessel such as loads and strength
calculations that can be used in providing a more detailed
response in the event of a casualty.
Vessels classed by other societies may enroll in the ABS
RRDA program provided all enrollment requirements are
met. Verification that a ship is enrolled in the RRDA
program is by issuance of a certificate and via the ABS
O2K system.
2.4 EXPANDED RRDA SERVICE
The enhanced RRDA service brings both class surveyors
and RRDA team under the leadership of the ABS Chief
Surveyor. This streamlines communications between
ABS engineering and survey interests resulting in more
efficient production of suitable recommendations.
Traditionally, the RRDA team was activated to help with
stabilizing the vessel at the time of incident. The RRDA
team is now able to support the decision process of
owner and salvor (if requested) until the ship is delivered
for repair.
Once a casualtys condition is stabilized, the class
surveyor can assess the damage to hull and machinery.
For ABS-classed vessels, surveyor approval remains a
requirement for all subsequent evaluation of damage and
repair. It is the attending surveyor who will issue the
Certificate of Fitness to Proceed once all
recommendations are completed. Class surveyors and/or
owners may request the RRDA team to assess the
vessels stability and structural strength again as
additional information becomes available.
The Damaged Ship, London, UK
3. SOFTWARE FOR ENHANCED RRDA
ENGINEERING ANALYSES
3.1 ENHANCED ENGINEERING ANALYSIS
Another feature of the enhanced service is the upgraded
engineering analyses.
Traditionally, analyses at the time of an incident are
focused on calculations of vessel trim, stability and still-
water loads on the hull girder. Continued improvements
in computer software have provided new tools that can
be used to assess the damage condition in a more holistic
manner. The traditional analyses are expanded to include
evaluations of loads as well as global and local strength.
The application of advanced design analysis tools to the
RRDA decision process has made it possible to perform
timely calculations of hull girder ultimate strength and
local buckling and the vessels ultimate strength in the
damaged condition. Further, by using the extensive
global wave data held by the society, ABS is able to
quickly evaluate the hull girder and local strength for the
planned repositioning voyage of the vessel to a repair
facility.
3.2 ANALYSIS SOFTWARE
The RRDA program is an integrated software program
which covers HECSALV and a suite of load and strength
calculation tools. The software is used for modelling the
hull structures at the time of RRDA enrollment and for
performing engineering analysis at the time of an
incident.
Calculations to assist with decision making during and
after an incident are as follows:
Residual global strength (in terms of hull girder
ultimate strength)
Loads at the site of incident or along the transit
voyage
Strength of local structural members (in terms
of yielding, buckling/ultimate strength, shear
buckling strength)
The focus of the specially developed software was made
as follows:
Analysis that can be performed in a timely
manner
Easy modelling of structures
Capability of utilizing models of ships that have
been created during design analysis process
Improved presentation and reporting
At the time of an incident response, it is crucial to
evaluate the vessel condition in a timely and efficient
manner. Many calculations are needed within the first
hours of an incident. For this reason, preference in the
methodology of software development was given to first
principle-based tools. (FEM calculations are not
considered suitable for response analysis due to the
extended run times required.)
The additional engineering analysis was based upon
well-accepted industry practice. The design rules of
IACS were drawn upon to the extent possible. An
example is the hull girder ultimate strength calculation.
Where IACS does not provide guidance, established
ABS Rules were adopted. Since the basis of IACS rules
considers vessels trading in the North Atlantic, the ABS
Seas Environmental Assessment System (SEAS)
program is used for calculating the environmental loads
in other locations around the world.
When deemed necessary, established industry practice is
also adjusted by incorporating achievements of advanced
research and development. An example is the adjustment
of hull girder ultimate strength for ships under hogging
condition. (See discussion in Section 4.3 concerning the
factor used in Eq. 2.)
Many design programs that ABS has developed have
been modified and incorporated into this RRDA
program, including
HGSA for calculating sectional properties of
ships cross sections
ABS CSR software for calculating hull girder
ultimate strength, local buckling strength
ABS SEAS program for calculating
environmental loads
This helped to expedite the software development
process.
The following sections will introduce and explain the
methodologies and procedures of the engineering
analyses.
4. GLOBAL STRENGTH
A very important feature of the enhanced analysis
capacity is evaluation of residual hull girder strength for
the damaged hull.
4.1 HULL GIRDER STRENGTH
The global strength (also known as hull girder strength or
global hull girder strength) refers to a vessels capability
to resist vertical bending moment. Traditionally, the hull
girder section modulus is used for representing a vessels
hull girder strength. Recently, with the introduction of
IACS Common Structural Rules (CSR), the hull girder
ultimate strength has become an important measure of
global hull girder strength and technically the hull girder

ultimate strength is a much better measure of a hulls
capacity for resisting vertical bending moment.
The hull girder ultimate strength defines the bending
moment value that could cause collapse the hull girder.
An example of hull girder ultimate strength failure is
shown Fig. 3.
Figure 3. Collapse of a bulk carrier during cargo loading
(www.marinetalk.com).
The incremental-iterative procedure described in the
IACS CSR represents the current industry standard. This
procedure and software has been incorporated into the
ABS RRDA program.
4.2 VERTICAL BENDING MOMENTS
The damaged hull is exposed to still-water bending
moment (SWBM) and wave-induced bending moment
(WBM). The SWBM is a result of the vessel loading
condition, breached tanks, loss of buoyancy,
countermeasure, etc., and is calculated using the
HECSALV program.
New technology now makes it possible to calculate the
WBM quickly. To calculate WBM, RRDA utilizes the
ABS SEAS.
The SEAS program adjusts the design loads with a factor
that reflects the site or route-specific sea environment.
For example, the WBM at the geographic location of an
incident or along a transit voyage is defined as:
WBM= WBM0 (1)
where WBM0 is the design wave-induced bending
moment and is the factor that is dependent on the site
of the incident or along the transit voyage.
4.3 EVALUATION OF GLOBAL STRENGTH
The global strength of a troubled ship needs to satisfy the
following condition in order to prevent hull girder
collapse from occurring:
s SWBM+ s ( WBM0 ) < ( Mu) /m (2)
where s , s and m are the partial safety factors for
SWBM, WBM and residual hull girder ultimate strength,
The Damaged Ship, London, UK
respectively and is a factor for adjusting the predicted
hull girder ultimate strength under hogging condition.
This criterion is based on the IACS CSR criteria for hull
girder ultimate strength and the same partial safety
factors are also adopted (1.0, 1.2 and 1.1, respectively).
The CSR approach does not take into account the
additional structural stresses induced by the lateral
pressure (due to the internal cargo loads and/or external
sea loads) or biaxial stresses resulting from grillage
effects. As s result, the hull girder ultimate strength
under hogging condition is overestimated (Wang et al.
2010). Based on review of existing research papers, is
taken as 0.9 for the ultimate strength of tanker structures
under hogging condition. It can be a much lower value
for bulk carriers. The is 1.0 for sagging condition of all
ship types.
5. SEA ENVIRONMENTAL LOADS
One of the most significant enhancements to the ABS
RRDA program is the use of the ABS SEAS program for
estimating the environmental loads on hull structures.
This program makes it possible to calculate within
minutes, the environmental loads on hull structures for
any specific geographical location or the entire transit
route.
5.1 WORLDWIDE WAVE DATABASE
The ABS wave database in SEAS has been derived from
the global wave statistics provided by the National
Oceanic and Atmospheric Administration (NOAA)
Global Spectral Ocean Wave Model (GSOWM) hindcast
dataset. The ABS wave database consists of 1,102 grid
cells as shown in Fig. 4. For each cell, a wave scatter
diagram is stored with its associated directional
probability distribution, i.e., wave rosette.
With location points defined for a specified route, the
SEAS program searches for the wave zones (grids)
crossed by the route. Based on the vessels travel
distance within the wave zone, SEAS determines the
factor for time spent in each wave zone. Also, SEAS can
provide the wave scatter diagram and relative wave
directionality (to the ships course) for each wave zone.
The weighting factors of time, wave scatter diagrams,
and wave rosettes are needed, which reflect the effects of
the specific transit route.
The Damaged Ship, London, UK
Figure 4 ABSWAVE database grids (total of 1,102 wave
grid cells) (ABS 2009, 2004)
5.2 SITE/ROUTE-SPECIFIC LOADS
The site/route specific loads are calculated by using the
Environmental Severity Factor (ESF). This is used to
adjust the wave-induced dynamic loads. An ESF, , is a
ratio of the severity between the site/route environment
and a base environment. The factors apply only to the
dynamic portions of the load components (static load
components are calculated as part of the trim/stability
calculation).
The ESF is defined as follows:
= LS/LU (3)
where LS is the most probable extreme value at the
incident site or along the intended transit route; and LU is
the most probable extreme value based on the North
Atlantic environment for the corresponding dynamic
load.
A of 1.0 corresponds to the unrestricted service
environment condition, or North Atlantic environment.
A value of less than 1.0 indicates a less severe
environment than the North Atlantic environment.
6. LOADS AND STRENGTH OF LOCAL
STRUCTURAL MEMBERS
Traditionally, local strength has not been routinely
evaluated during an incident response. This is because
the traditional focus of RRDA was ship stability and
global strength. Also, at the time RRDA was first
implemented there was no software available for
routinely performing local strength analysis.
The RRDA software has now been enhanced to include
tools for calculating local loads, local strength, local
buckling/ultimate strength and local shear buckling
strength.
6.1 LOADS ON LOCAL MEMBERS
Reducing loads on tank boundary can be important when
the owner/operator plans for temporary repairs or plans
for a loaded or ballast transit voyage. Therefore, a utility
tool was added to calculate pressure differentials.
An example of this load differential calculation is shown
in Fig. 5. In this case a grounded tanker was floated and
the decision was made to empty the damaged tank and
make repairs. However, this would have resulted in high
static pressures on the inner bottom and slopping
bulkhead plating. This was not a prudent option as the
severely deformed support structures in the bilge area
(floors, girders, and web frames) had become less
effective in supporting the tank boundary. An option
considered at the time of the incident was to minimize
the pressure on the tank boundary by adding a certain
amount of water in the ballast tank.
Figure 5. Partially filling the ballast tank was considered
a suitable option for reducing the pressure differential on
the inner bottom of a grounded tanker.
6.2 STRENGTH OF LOCAL MEMBERS
The structural responses are calculated using classical
beam theory. The buckling/ultimate strength, including
shear buckling strength, is calculated using the IACS
CSR procedure and criteria of prescriptive buckling.
The static loads are calculated by trim/stability software,
and the dynamic environmental loads are based on
calculations of the ABS SEAS program.
Acceptance criteria specified in IACS CSR are the basis
for evaluating the potential failure modes of yielding and
buckling.
7. CONCLUSIONS
This paper presents enhancements to the ABS Rapid
Response Damage Assessment program. Major
enhancements include the enrollment of eligible vessels
at the time of delivery, expanded RRDA service to both
initial incident response and subsequent transit voyage
period with enhanced engineering analyses to estimate
loads and global and local strength.

A key feature of this expanded service is the introduction
of an integrated software system that simplifies RRDA
enrollment at the time of delivery. By introducing these
vessels into the RRDA program at the outset, the
classification society can readily access a wider range of
technical information relevant to the vessel, such as loads
and strength calculations that can be used in providing a
more detailed response in the event of a casualty.
The application of advanced design analysis tools to the
RRDA decision process has made it possible to perform
timely calculation of hull girder ultimate strength and
local buckling and the vessels ultimate strength in the
damaged condition. Further, by using the extensive
global wave data held by the society, ABS is able to
quickly evaluate the hull girder and local strength for
planning the voyage to the repair facility.
8. ACKNOWLEDGEMENTS
The author wishes to thank ABS management for its
support and guidance and M. Lee, D. Ghose, N.
Konduru, V. Raghunathan, and Eric VanDerHorn for
their cooperation during the development of the RRDA
software.
9. REFERENCES
1. ABS, GUIDE FOR RAPID RESPONSE DAMAGE
ASSESSMENT, www.eagle.org, July 2010.
2. ABS, GUIDE FOR BUILDING AND CLASSING
FLOATING PRODUCTION INSTALLATIONS,
www.eagle.org, July 2009.
3. ABS, COMMENTARY ON THE CRITERIA TO
ADAPT THE SAFEHULL SYSTEM TO FPSO
APPLICATIONS, www.eagle.org, May 2004.
4. Wang, G., Chen N., Guo J., Guedes Soares, C.,
Application of Structural Reliability Approach to
Assist Hull Integrity Management, PRADS, 2010.
APPENDIX A. TABLE OF CONTENTS OF ABS
RRDA GUIDE [1]
SECTION 1 General
1 Scope and Application
3 Basis of Notation
5 Automatic Enrollment for New Vessels
7 Enrollment of Vessels Classed by Other Societies
SECTION 2 Documentation
1 Definitions
3 Client Duties
5 Plans and Data to be Submitted
7 Database Preparation
9 Types of Response Analyses
The Damaged Ship, London, UK
SECTION 3 RRDA Program
1 Activating/Notifying RRDA Team
3 Procedure for Reporting Incident
5 Survey Recommendations
7 Client Deliverables
APPENDIX 1 RRDA Notification Bridge Card
APPENDIX B. REQUIREMENTS FOR SERVICES
OF RAPID RESPONSE DAMAGE ASSESSMENT
The ABS RRDA Program is designed to fulfill the
requirements of IMO, USCG and ISM regulations.
B.1 IMO
Effective 4 April 1995; MARPOL 73/78 Annex I,
Regulation 26 requires a Shipboard Oil Pollution
Emergency Plan (SOPEP) for all tankers of 150 gross
tons or more and all other vessels of 400 gross tons or
more. The Guidelines for the Development of
Shipboard Oil Pollution Emergency Plans, IMO MEPC
54 (32), advises that detailed guidance to the vessels
master should be provided to take stability and stress
into consideration when taking actions to mitigate the
spillage of oil or to free the vessel from aground.
While it is not required, it is strongly suggested that
when there is excessive damage, consultation with shore-
based technical assistance is appropriate before taking
any action that may jeopardize the vessel.
As of 1 January 2007; MARPOL 73/78 Annex I, in
accordance with resolution MEPC.117 (52), Regulation
37.4 requires that all oil tankers of 5,000 tons deadweight
or more shall have prompt access to computerized, shore-
based damage stability and residual structural strength
calculation programs.
B.2 USCG
In the United States, the ABS RRDA program satisfies
the U.S. Coast Guard requirements of OPA 90 in 33 CFR
155.240 for oil tankers and offshore oil barges. Owners
are required to have prearranged, prompt access to
computerized, shore-based damage stability and residual
structural strength calculation programs. It also requires
that access to the shore-based calculations program
must be available 24 hours a day.
This requirement became effective 21 January 1995 for
vessels operating in U.S. waters.
B.3 ISM
The ISM Code, Section 8, requires the company to
establish procedures to respond to potential emergency
shipboard situations, including the use of drills and
exercises to prepare for those emergencies.
The Damaged Ship, London, UK

AUTHORS BIOGRAPHY

Ge Wang currently holds the position of Manager,

Advanced Analysis at American Bureau of Shipping
Greater China Division. He is responsible for advanced
analysis, pre-contract engineering review and joint
development projects. His previous experience includes
work with the ABS Corporate Technology Department,
ABS Ship Engineering Department and CONVAC
Offshore Consulting Company. He has had over 100
articles published in journals, conferences, book and
book chapters. He serves on the editorial board of three
major international journals (JMST, JSOS, JOAME), and
on various international committees (PRADS Standing
Committee, ISSC Committee Chairs, Executive
Committee of OOAE). He holds a PhD from University
of Tokyo and a MS and BA from Shanghai Jiao Tong
University.
The Damaged Ship, London, UK
THE DAMAGED SHIP -
MOD SALVAGE RESPONSE TO HMS ENDURANCE IN THE MAGALLAN STRAITS IN
DECEMBER 2008
J R Ward, M Watts and D Price, Ministry of Defence, UK
SUMMARY
This paper overviews the history, construct and capability of the MOD Salvage Organisation and briefly details some of
the more recent salvage tasks that the Organisation has been involved with. As a case study it will then go on the
describe in detail the MOD Salvage response to HMS ENDURANCE, when she was disabled and left drifting in the
hostile waters of the Magellan straits in December 2008.
1. INTRODUCTION
The Salvage and Marine Operations Project Team
(S&MO) of the UK Ministry of Defence (MOD) has
been in existence for over 100 years and has a long and
proud history of worldwide salvage activity.
2. HISTORY OF THE DEFENCE SALVAGE
ORGANISATION
The MOD Salvage Organisation was formed in 1906
following the grounding of HMS Montagu, a 14000 Ton
battleship that was built in 1901 as the Montague and
renamed to the HMS Montagu in the same year. It sank
29th May 1906 by running aground in fog on Shutter
Reef, Lundy Island. Massive efforts by the Royal Navy
to re-float her failed. Six months of salvage work
followed. At the subsequent Court Marshall, the
Admiralty directed that a professional Salvage
Organisation should be formed. This Organisation,
which has had a number of guises over the years
survives today in the form of the MOD salvage and
Marine Operations Project Team.
Fig 1 - HMS MONTAGU Aground of Lundy 1906
During the two world war the MOD salvage team
regenerated and made a significant contribution to the
war effort as prescribed by the prime minister at the
time Sir Winston Churchill
To Import Executive
25th Feb 1941
I learn that the Admiralty Salvage Organisation has
recently made as great a contribution to the
maintenance of our shipping capacity as new
construction, about 370,000 Tons having been salved in
the last five months of 1940, as against 340,000 Tons
built, while the number of ships being dealt with by the
Salvage Organisation has increased very rapidly, from
10 in August to about 30 now.
3. CURRENT CONSTRUCT/CAPABILITY
OF THE MOD SALVAGE
ORGANISATION
The current MOD Salvage Organisation has evolved
from the large post war structure with numerous vessels,
tons of equipment hundreds of personnel to a much
leaner and effective construct. The most recent change
was in 1995 when much of the marine services support
to the Military was privatised, the 6 remaining salvage
and mooring vessels, their crews and diving groups,
together with large warehouses of salvage equipment,
whose running cost where in the region of 35M per
annum where disbanded or transferred to other tasks.
The remaining Salvage and Marine Operations Project
Team, consists of approximately 55 personnel of whom
34 are technical specialists in the maritime/salvage field,
who maintain a variety of use relevant skill sets1. The
Organisation holds a reduced amount of high
value/difficult to obtain equipment2 that is mostly air
portable and hires in platforms and low value/easily
obtainable equipment as and when required. The
current operating budget for the Organisation is in the
region of 3M per annum. The Organisation continues
to invest in new technology, and has recently taken
1
Such as explosive cutting, diving, sonar locator and
marine survey systems, underwater welding etc
2
Such as sonar locator beacon detectors, side scan
sonar, remotely operated vehicles, explosive proof high
capacity pumps etc

delivery of a 1000m depth rated small work class ROV.
It also works with industry and academia to develop
innovative tooling and solution to underwater problems.
3.1 RECENT MARINE SALVAGE ACTIVITY
Since 1992 the Organisation has been involved directly
in or provided advice to more than 30 major salvage
tasks involving military and commercial vessels. Most
recently in support of the refloating of HMS ASTUTE
following her grounding off the North coast of Skye in
Oct 2010.
4. CASE STUDY
MOD SALVAGE RESPONSE TO HMS
ENDURANCE IN THE MAGELLAN
STRAITS IN DECEMBER 2008
4.1 INTRODUCTION
This section of the paper details the deployed
operational element of the salvage of HMS
ENDURANCE which took place during the period 17th
to the 30th of Dec 2008, in Punta Arenas Chile.
4.2 BACKGROUND
At approximately 16/1904Z Dec 08, whilst transiting
the Western Magellan Straits at the Southern tip of
South America (see Fig 2) the main engine room (ER)
of HMS ENDURANCE (END), flooded rapidly 3
through the main engine strainer box4. Despite the
crews best efforts to reduce it5, the ships limited
pumping capacity was rapidly overwhelmed, however,
flooding was contained6, by closing the transverse
watertight doors; as a result the ship lost all propulsion
and most services7. The cabins directly above the ER
on C deck were flooded and subsequently damaged by
3
It was reported that the Engine room and adjacent
spaces within the Transverse WT boundaries,
approximately 1400 Cubic meters, flooded in 7 minutes.
4
The strainer had been deliberately removed to clean,
unfortunately the hull valve opened whilst the strainer
cover was off. END carried 3 RN clearance divers,
during the flooding one of theses divers entered the ER
in SABA Mk1on a number of different occasions in a
vain attempt to stem the flow, by finding and securing
the strainer cover.
5
END carried 3 RN clearance divers, during the
flooding one of theses divers entered the ER in SABA
Mk1on a number of different occasions in a vain
attempt to stem the flow, by finding and securing the
strainer cover.
6 9
Some leakage outside the WT boundary was
observed through cable glands, grey water systems and
a perforated deck.
7 10
The ships emergency generator provided emergency
power and lighting throughout the salvage operation
The Damaged Ship, London, UK
water moving around as the ship came beam on to the
swell and rolled violently.
END broadcast a MAYDAY message and alerted the
duty FLEET controller (DFC) to their plight. S&MO
IPT were alerted by DFC at approximately 16/1930Z.
Following clarification of the predicament of END,
S&MO PT personnel were recalled and shortly
thereafter salvage tugs were contracted and sailed from
the port of Punta Arenas (PA).
Meanwhile the Chileans provided a helicopter, Maritime
Patrol Aircraft, patrol vessel (which transferred three
pumps8 to END) and a pilot vessel. Some non-essential
personnel including School Children were evacuated by
helicopter9. The cruise ship Norwegian Sun, which was
the nearest major vessel, stood by END to provide SAR
assistance if required, but was later able to continue on
her way.
Without power END drifted South Eastwards ever
closer to the coastline10 at the mercy of wind and tide.
By good fortune the wind and slowly changed and the
END drifted close to Bajo Magallanies the only
shallow patch in the area. She let out both anchors and
fetched up on this patch at 170245Z.
END remained securely anchored until the first Salvage
tug MT BEAGLE arrived at approximately 171600Z.
The END took the tugs line on her forward bits and cut
both anchor cables on deck. The BEAGLE began to
slowly tow the END towards PA. Shortly afterwards
the Tug MT AGUILA 3 arrived to initially escort and
was later attached to the stern to reduce the yawing
motion of END while under tow.
4.3 SALVAGE RESPONSE - CHRONOLOGY
OF EVENTS
18 Dec 08
Whilst the END was under tow, the S&MO PT prepared
to deploy a 13 man salvage team to PA, to assist Ship
Staff (SS) in restoring the watertight integrity of the
vessel and prepare/tow END to the Falkland Islands.
The team departed from RAF Brize Norton on the
MOD chartered flight at 172300Z, with other members
8
The pumps provided by the Chileans were mainly
small diesel pumps they were used to pump water from
the after port stairwell on C Deck aft of the A60 door
and forward of the aft ER WT bulkhead, in reality due
to the large amount of water flowing in through the
strainer they were doing no more than providing a good
will panacea to SS.
Just prior to the incident a documentary crew had
been landed at PA, they later returned to the END and
filmed the Salvage operations
The coastline in this area is characterised by deep
water, running right up to sheer rocky cliffs.
The Damaged Ship, London, UK
of the ships company who had been on UK leave at the
time11, an investigation team, a welfare team, two
platform PT representatives and a small amount if
salvage equipment12. The team travelled through the
night arriving at Mount Pleasant Airfield in the
Falklands Islands at 181800Z. There after a short stop
they boarded an RAF C130 Hercules for the 3 hours
flight to PA, where they arrived at 182220Z (1920W
local time 3hours).
Shortly after arrival a small team including the CO of
END, members of the investigation team and
DTL/S&O13 boarded one of ENDs Lynx helicopters at
PA and flew out to join END as she approached PA
under tow. Shortly after arrival the CO took back
command of the END from XO, although the XO
retained operational control. The remainder of the party
arriving on the C130 cleared customs and rested for the
night.
END berthed starboard side to the ASMAR14 finger
jetty see Figure 2 at 2300W with the aid of the two
salvage tugs and the tug MT OTWAY. The salvage tug
MT BEAGLE secured alongside the port aft of END to
assist holding her on to the berth in case of adverse
weather. Once secured two 80 Cubic meters per hour
electric submersible pumps15 were passed to END
allowing the small diesel powered pumps provided by
the Chilean navy to be removed.
Fig 2 HMS ENDURANCE Berth Starboard side to the
Finger jetty at the ASMAR Magellan Straits facility
near Punta Arenas Note the 6 list to port and deep
draft condition. The Tug MT BEAGLE sits on the port
quarter maintaining a standby presence in case of bad
11
These included ENDs CO Capt Gavin Pritchard
12
SABA sets, Diving air compressor, MIKO patches,
LBV mini ROV tools, communications etc the amount
of salvage equipment deployed was limited by the
capacity of the aircraft and by previous LI, most notably
those from the HMS NOTTINGHAM recovery which
suggesting that it was more cost effective to source
salvage equipment locally rather than flying UK based
salvage equipment large distances,
13
Deputy Team leader/Salvage and Operations
14
http://www.asmar.cl/ast_mag.html
15
Powered from the MT BEAGLEs onboard
generators
weather outboard of her is the sullage barge
Brecknock.
19 Dec 08
Suitably rested the salvage team proceeded to the
ASMAR shipyard on the northern outskirts of PA.
Once onboard they were given a toolbox safety talk and
a familiarisation tour around END. Command and
Control (C) was agreed between the CO END and the
DTL/S&O, this essentially made the salvage team
responsible for the recovery operations, whilst leaving
overall responsible for END, specifically its crew and
safety with the CO. Early clarification of roles at this
stage was to prove invaluable during the many
challenges that lay ahead.
The initial salvage plan was to patch the high suction
grill immediately adjacent to the leaking strainer using a
magnetic MIKO patch and then to pump the ER
initially via the funnel casing and then through the aft
Starboard WT door, using locally hired pumps.
Pumping needed to be carefully controlled to ensure that
END retained adequate stability throughout the
dewatering operation; the GM was calculated to be at
minimum at certain stages, particularly until C deck was
fully drained, due to the potential negative effect of free
surface effect.
The salvage team rigged for diving16 during the
morning17. 191320W diver enters the waters.
191400W MIKO patches were secured over the high
suction and its adjacent flooding holes.
Whilst the diving was progressing other members of the
salvage team were checking the locally supplied
pumps18, identifying pumps runs, and checking/sourcing
towing equipment19.
16
Although a Chilean diving team was available, it
was assessed that the risks of using an unknown entity
with the attendant language barriers out weighed the
risks associated with using S&MO IPT personnel
operating in their self contained diving system to secure
the patches.
17
This included identify a suitable local
recompression chamber, obtaining the necessary
permissions and obtaining diving tag outs from adjacent
vessels, and deploying ENDs Gemini as a dive platform.
18
The local International Salvage Union Member
supplied a number of pumps both electric submersible
and pneumatic diaphragm pumps with hoses. Of the
two 3 pneumatic pumps supplied one turned out to be
faulty as did one of the electric submersibles and many
of the hoses where oil contaminated, these needed to be
flushed into the sullage barge. Also the fuel tank on the
LP air compressor supplied to run the pneumatic pumps
melted shortly after being supplied.
19
Like salvage equipment, towing gear was in short
supply, END only had limited stocks, local stocks were

Whilst the salvage team were en route to PA, S&MO
IPT HQ had placed a contract with a local contractor to
recover the slipped anchors and cables. During the
morning of the 19th the MT SKYRING which had
successfully recover one of the anchors berthed forward
of END. A portable generator was sourced and wired
into the END forward power pack during the afternoon
of the 19th, by the salvage team electrician. Once the
power pack was running seven lengths of the recovered
cable and the anchor were restowed in the ENDs
starboard chain locker and two lengths where recovered
onboard END for use in the towing rig. MT
SKYRING then departed to locate and recover the
second anchor, returning to PA on the evening of the 22
Dec 08.
20 Dec 08
The pump runs were improved by cutting an access
point through the funnel casing see Figure 3.
Salvage team/SS were advised by the onsite Platform
PT representative that calculations staff at DE&S
Abbeywood (AW) indicated that the ships stability
would be critical during pumping20, it was therefore
recommended that a number of measures21 where taken
to improve stability before pumping began. At this
point local operations were suspended, whilst the
situation was assessed. The salvage team suggested
that to improve stability and reduce the list, a number of
Double Bottom and half full tanks22 should be pressed
up, this idea was modelled by the team at AW and
accepted. Consequently Starboard Fresh water tanks 5
and 623 were pressed up using the ventilator pipes and
the shore FW supply hose; this remediation action took
most of the day. It subsequently transpired that the
Platform NAs were using a wrongly configured model,
thus the GM issue was never actually a problem.
limited and with limited airfreight capacity, re supply
would during normal periods take at least 5 days to
arrive by road, this would be considerably longer over
the Christmas and New year period. The preparation of
the tow was greatly aided by the recovery of ENDs
slipped anchors
20
It subsequently transpired that the Platform NAs
were using a wrongly configured NAPA software
model.
21
Including the removal of a number of ships boats
this would have proved to be very difficult given the
restricted access and lack of local cranage available.
22
The END was built without sounding pipes. During
normal operations soundings were recorded remotely;
however the flood in the ER permanently disabled this
system making it difficult to measure tank contents.
23
This tank subsequently leaked 15m3 back into the
ER, though open valves/damaged pipework.
The Damaged Ship, London, UK
Fig 3 Access hole at B deck level cut through the
portside of funnel casing into the top of the ER This
was later enlarged to allow man access
21 Dec 08
By 210830W, the first pump was running in the ER
discharging overboard24. Throughout the morning other
pumps were run up until the pump capacity reached
some 240 Cu per hour. It soon became apparent that
either the MIKO patch was not working or other
overboard suction valves were open. SS were unable to
advise the valve line up at the time of the flood,
additionally there were concerns that the heavy rolling
shortly after the flooding incident may have damaged
some pipe work in the ER.
Due to weather limitations on diving, the Salvage
teams mini ROV deployed to try and identify other
potential areas of suction. Additionally a previous
MEO on END who was working in the DE&S
operations cell advised that END carried purpose made
high and low level diver friendly blanks. These were
located in the after ends of ENDs funnel. Due to
continuing adverse weather the decision was made to
shut down the dewatering pumps overnight.
22 Dec 08
At 221000W the 1st diver was in the water commencing
rigging/fitting the purpose made blanks over the high
and low suctions (See fig 4), and fitting the recovered
MIKO Patches over the emergency suction Grill. The
24
Intentions were to decant the clean water from the
ER over the side; once the strainer level had been
reached the remaining oily water floating on top of the
ER would be pumped into the 600 m3 capacity sullage
barge Brecknock. Given the limited sullage capacity
and the fact that the patches could not be guaranteed to
be 100% water tight this was the only course of action.
To ensure that Chilean authorities were happy with this
proposal the PA Harbour Master was advised of the
process and verbally (in front of witnesses) confirmed
his acceptance.
The Damaged Ship, London, UK

dewatering pumps were restarted and a diver swim

carried out using a plastic sheet to detect suction.
Further suction was identified to be coming from the
emergency FW generator grill, and from the High level
suction patch, which had not seated squarely. Due to
the lateness of the day the dewatering pumps were shut
down overnight.

Fig 4 S&MO IPT salvage personnel preparing to

position the purpose made salvage patch over the High
level suction. Fig 5 Aft Starboard C deck water tight door opened
for the 1st time in nearly a week to allow access to the
23 Dec 09 top of the ER

At 0900W the diver entered the water to reposition high 24 Dec 08

level patch and fit a locally fabricated plywood blank
over the emergency FW generator suction. At 1100W At 240830W, with the help of SS the access hatch (see
pumping recommenced. At 1130W END Embarked RN fig 6) sited in the ER workshop was opened giving
CDs swam round the hull checking for further suction improved access to the ER. SS also began to run
points in the ships hull. Several potential suction points lighting strings28 into C deck and the ER workshop.
were located and plugged using wooden bungs. With the pumping apparently under control and the
wind having moderated a crane was arranged to offload
At 231600W the salvage team broke into 4 hours the ENDs second anchor and cable onto the jetty.
watches to provide 24 hours cover whilst dewatering
progressed. At 231820W25 with command approval
the starboard aft C deck WT (see fig 5) door was opened
to allow access for pumps to the ER. Pumping of the
ER space through the access from the ER workshop
continued overnight26, although physical access was
limited by debris27 that had accumulated in the access
during the flood and the ENDs lively passage back to
PA.

25
The WT door is opened almost a week after it was Fig 6 Access hatch from ER workshop to ER,
1st closed. showing pump suction and discharge hoses
26
using 3 electrically driven submersible pumps
(although one of the leads shorted out in the night) and
28
one 3 pneumatically driven diaphragm pump The salvage team had been unable to source
27
This included a buckled WT door, an MMA emergency lighting strings locally; SS therefore
welding machine, numerous hoses, and several removed and adapted WT upper-deck lighting from
mattresses. END for this task.
 The Damaged Ship, London, UK

At 0900W it was noticed that the water level in the ER

was rising slowly. With the help of SS a 4 inch
electrical submersible pump was quickly rigged into the
top of the ER through the access hatch and the damaged
power lead on one of the existing 3 Electric
submersible pump was re-terminated, restoring its
pumping capacity. At this stage despite an estimated
pumping capacity of 350 CuM per hour, the water level
continued to rise. It was agreed with command that if
the water level reached the top of the ER workshop
access, the pumps would be withdrawn and the WT
doors closed to preserve the safety of the vessel. At
0945W, this level was reached and with an immense
joint effort by the salvage team and ship staff the pumps
and lighting were disconnected and the WT door closed.

At 1230W the offload of the MT SKYRING was

completed. Shortly afterwards the END clearance
divers deployed to survey the patches they found that
the small MIKO patches covering the drains to the high
level suctions were missing29 and that water was
flowing into the ship through these inlets and valve No
17. Plywood patches were made to cover the drains
and a wooden bung was inserted in opening of valve No
17. Meanwhile the salvage team enlarged the small
access hole in the funnel to allow man entry into the top
of the ER.

With the hull integrity restored the pumps where rigged

through the top of the ER, via the improved access. At
1630W, the pumps were restarted. Pumping continued
overnight.
25 Dec 08 Christmas Day
At 250815W the starboard aft C deck WT (see Fig 5)
door was re-opened to allow access for pumps/personnel
into the ER workshop and C deck. SS rigged30 a water
driven vent fan (to ventilate the lower ER), and lighting
in C deck and the top of the ER. They also began to
clear debris from the accesses to the ER through the ER
workshop (See Fig 7) and through the Engine Control
room (ECR).
Fig 7 Partially cleared access stairway from the ER
workshop to the ER note the typical debris
encountered
At 251240 after testing the atmosphere members of the
investigation team entered the ER and closed the high
suction overboard valve using a large stilson type
wrench (see Fig 8). They then located and secured the
strainer cover. At 1245W pumping was stopped and no
significant inflow was observed. SS shut down a
number of valves that were believed to be draining the
FW tanks into the ER. Pumping of the ER bilges was
restarted at 1500W using two pneumatically driven
diaphragm pumps31, these pumps discharged into the
sullage barge Brecknock. Pumping operations were
stopped at 1830W, when the atmosphere in the ER
became untenable due to SS spraying cleaning solvent.
At this point the salvage team broke out of watches.

29
It is surmised that the MIKO patches were
dislodged by the MT SKYRING manoeuvring in
position ahead of the END to discharge the send anchor
and cable.
30
Using the ships fire main and a C deck overboard
31
discharge One hired 3 Wilden and one brand new 2
Arrow, found onboard END.
The Damaged Ship, London, UK
Stilson type spanner
securing the actuator
closed
Fig 8 High level suction valve with Stilson type
spanner securing the actuator closed
26 Dec 08 Boxing Day
The pneumatic pumps were restarted to re-commence
the discharge of the ER bilges. At 1000W Finnings
the local Caterpillar Agent and preferred Engine
preservation32 specialists paid a second visit to the END
to inspect the engines, gearbox and auxiliary with view
to quoting for their preservation.
It was noted that the High level suction strainer was
weeping slightly; this was due to a flat spanner being
left between the top of the strainer and the lid when it
was secured. Whilst SS rectified this problem, the
salvage pumps and patches were left in place. In the
meantime the salvage team commenced rigging the tow
using primarily END anchor chain, and limited locally
sourced shackles and wires, from the tug MT
LAREDO33 which replaced the MT BEAGLE as
standby alongside tug.
32
At this stage a number of options were being
considered for re-establishing the capability provided by
END, these included repair and replacement. To ensure
that the overall repair cost was minimised it was
important to ensure that key pieces of machinery were
preserved as soon as possible after drying out to prevent
long term damage.
33
It was also to be used as the lead tug for the tow to
the FI.
At 1830W Finnings advised Ultramar34 that they were
not prepared to undertake the preservation work.
Following some pressure from the Salvage Team
Ultramar identified another preservation team overnight.
27 Dec 08
With the Engine Room external boundaries confirmed
watertight the Salvage Team began to withdraw and
stow the electric submersible pumps, pumping of the
ER continued throughout the day using the pneumatic
pumps discharging into the Brecknock. They also
secured the shaft to prevent rotation during the tow. At
0840 the new engineering team arrived onboard END to
inspect the engines, gearbox and auxiliary with view to
quoting for their preservation.
At 0930 SS began removing debris from C deck and the
ER At 1045 and 1145 ENDs two lynx aircraft were
landed onboard and were securely stowed in the hanger
for the tow passage. Once the END had secured from
flying, the salvage team deployed divers to remove the
patches.
Rigging of the tow was delayed due to the shore crane
not being able to operate due to high winds, only a few
lifts being accomplished before the crane departed for
the day at 1500W.
At 1800W following the agreement of terms the
preservation team arrived and began work on preserving
the gearbox and main engines over night.
28 Dec 08
The salvage team performed a number tow preparation
tasks during the morning, including rigging the
tow35,continuing to pump the ER bilges, transferring a
pump to pump the shaft tunnel, securing the steering
gear on the centerline by means of hydraulic lock and
mustering S&MO IPT equipment for dispatch back to
the UK via a carrier. The salvage team also assisted
SS in the preservation of the auxiliary engines. SS
continued to remove debris from C deck (See Fig 9) to
prevent spontaneous combustion of decaying materials.
34
Ultramar a local agency, who where part of the
group that provided the tugs/sullage barge provided
technical support to the salvage team
35
Although this was again hampered by the lack of
shore crane, this time due to it being a Sunday

Fig 9 C Deck cross passage way looking from port to
starboard before SS began debris clearance.
At 1530W the single 2 pneumatic bilge pump
discharge was transferred to ENDs port after ballast
tank, and the pumping operations was handed over to
ENDs duty watch to allow the Brecknock to be moved
to facilitate the rigging of the primary tow onto the MT
LAREDO.
29/30 Dec 08
The majority of the salvage team returned home to the
UK during this period. Two of the salvage team
remained behind to work with SS to complete the final
tow preparations and then accompany the tow to the FI.
END finally departed PA at 301200W towed by the MT
LAREDO (see Fig 10), with the MT OTWAY acting as
escort.
Fig 10 END being towed by the MT LAREDO en
route from PA to the FI
5. CONCLUSION
The deployed operational element of the salvage of HMS
ENDURANCE which took place during the period 17th
to the 30th of Dec 2008, mainly, in Punta Arenas Chile
The Damaged Ship, London, UK
was timely, professional and successful. It overcame a
number of potential difficulties brought about by
extended supply chains, language difficulties and the
inevitable delays occasioned by the Xmas and New Year
period.
The factors which aided success were;
Early clarification of C2 with END command
An enduring and positive working relation with
SS and PT representatives.
A well trained/practiced highly motivated multi-
skilled salvage response team.
Excellent support from S&MO IPT personnel in
the UK.
Good local agency support.
A willingness by SS and the salvage team to
innovate and adapt.
Good communications by both telephone and
email.
Ships Hull markings.
The factors which hindered success were;
The time of year e.g. Shipyard support limited.
Language.
The long limited supply chain and lack of local
resources.
Lack of SS knowledge on the ER valve line ups.
SS unaware that manufactured blanks were
carried on board.
Lack of clarity on the ER preservation
requirement.
MIKO patches coming off at a critical point in
the pumping cycle.
6. AUTHORS BIOGRAPHIES
James Ward MBE Currently holds the position of
Deputy Team Leader Salvage and Operation, with the
Salvage and marine Operations (S&MO) Project Team,
part of the Ministry of Defences, Defence Equipment
and Support. He is based in Bath, England.
Currently he is the Assistant Director responsible for
Salvage and related Marine Operations, where he is
directly responsible for the two MOD marine salvage
units and the MOD in water maintenance and repair
desk, providing tri-service worldwide marine salvage
and in water maintenance and repair support and
expertise.
As a Master Mariner he joined the MOD in 1991,
following 15 years working in the commercial maritime
industry. Since joining the MOD he has a variety of jobs
within the Salvage and Mooring area. He qualified as a
commercial diver in 1999 and was awarded an MBE in
The Damaged Ship, London, UK

2003 for his work on maritime salvage and protection of David Price is 2nd in Command (2iC) of Marine
the maritime environment. Key achievements include: Salvage Unit (South) located at HMNB DEVONPORT,
Project Manager for the recent salvage of HMS Plymouth. He is responsible for the operational
Endurance off Chile management of the unit and its personnel, maintenance
Project Director for the world leading survey of of essential skills and capabilities and for the project
the Russian Submarine B159. management of salvage, mooring and towing
Project Manager for the Heavy lift of HMS operations.
NOTTINGHAM in Sydney harbour
st
Project Manager for the 1 UK full ROV recovery He joined S&MO in 2002 following 11 years in the
of a fragmented fast jet from the sea. Royal Fleet Auxiliary Service where he became a
Project Manager of the preliminary research, Master Mariner.
initial surveys, temporary patching and pilot oil
extraction operation of HMS Royal Oak in Scapa
Flow.
Project Officer during phase 1 of the return to MV
Derbyshire. Sole UK Government Representative,
data manager and leader of the identification team
during the 52-day phase 2 of the DOT sponsored
return to MV Derbyshire.
Project Manager of the environmental baseline
studies, approvals procedure and initial recovery
phases of the multi-million pound Holy Loch debris
clearance project.
Acted as senior Salvage Master and Project
Director, during high profile salvage incidents,
recent examples including the 4 month recovery of
Tornado jet from the sea adjacent to Torness power
station, emergency towing advice to disabled
vessels, the refloating of a Type 23 frigate and the
recovery of a ditched Merlin helicopter.

Martin Watts - Currently holds the post of Staff

Officer Towing with the Salvage and marine Operations
(S&MO) Project Team, based in Bath, England.

Maritime career began with Shell Tankers in ranks from

Deck Cadet to Chief Officer serving on LNG, VLCC
and product tankers, cumulating as Ship vetting/cargo
inspector before joining the MoD in 2000.

Experience in the MoD has included managing contracts

for the installation of the Faslane Security Barrier, MoD
Falkland Islands Shipping Contract, performing pre
charter ship inspections for various task, aircraft
recovery, heavylift and mooring operations.

Assisted the US Navy (NAVSEA) in the Salvage of

sunken Russian Submarine J484 in Providence R.I.

More recently has been actively involved in the marine

transportation of modules for the Queen Elizabeth Class
aircraft carriers as Marine Warranty Surveyor.

Specific to this presentation acted as salvage team

member in Chile, then Tow Master to bring Endurance
from Chile to the Falklands and as Project Manager for
the heavylift operation from the Falkland Islands to
Portsmouth, UK.
The Damaged Ship, London, UK

TOLERABLE SAFETY OF DAMAGED NAVAL SHIPS

S Marshall, Ship Stability Safety Regulator, Ministry of Defence (UK).

SUMMARY

One of the key elements of naval doctrine is the ability to absorb substantial damage before becoming non-operational.
In terms of damage stability this has traditionally been delivered for naval ships through the application of damage
extents relating to hostile threats. It has generally been assumed that this also provides adequate survivability against
merchant ship accidents such as collision & grounding. Military damage stability standards can be tailored by navies to
reflect the ship role and survivability requirements, thus defining the ship capability. This is illustrated by warships
having a high degree of survivability and naval auxiliaries being closer to mercantile standards. It is however, becoming
more common for navies to adopt minimum safety levels that are as least as effective as merchant ships. Whilst naval
ships are exempt from IMO conventions, compliance may compromise their war-fighting capability and may not
adequately protect the asset from minor damage. This paper examines the considerations to define minimum tolerable
damage extents for naval ships such that:

i) safety requirements are to be at least as effective as statute to protect life and the environment and;
ii) the consequences of damage should be proportional to the initiating event to protect the material military
investment.

Merchant ship damage statistics are examined and methodologies explored for defining minimum safety damage
extents. Finally the role of the Naval Ship Code and Naval Class is explored to provide naval ships with adequate
protection against collision and grounding that is demonstrably at least as effective as statutory requirements.

1. INTRODUCTION Combatant: A warship as defined by Article 29 of the

Navies perform a complex array of operations ranging
from conflicts & peacekeeping duties, to providing
oceanographic surveys, humanitarian aid and fishery
protection. This requires a versatile fleet, operating ships
with significant offensive and self defence weapons on
the front line of warfare (Combatants) and dedicated
support and replenishment ships (Auxiliaries). One of the
key elements of naval doctrine is the ability to absorb
substantial damage before becoming non-operational. As
such naval crews are trained to a high degree of
competency to, where practicable, recover from an
abnormal state such as damage. Naval shipping activities
are increasing coming under scrutiny and although
exempt from most aspects of merchant shipping
regulation. In the UK the Ministry of Defence has to
demonstrate that the level of risk of its shipping activities
is tolerable, as low as reasonably practicable and the
safety arrangements are at least as effective as statutory
requirements1. This paper makes a distinction between
capability and safety in terms of surviving damage are occasionally manned by civilians. However, for these
and develops a methodology for deriving minimum
safety damage extents for naval ships that are at least
as effective as IMO conventions.
United Nations Convention on the Law of the Sea Part
II *, and a ship having a significant offensive capability.
*A ship belonging to the armed forces of a State
bearing the external marks distinguishing such ships of
its nationality, under the command of an officer duly
commissioned by the government of the State and whose
name appears in the appropriate service list or its
equivalent, and manned by a crew which is under regular
armed forces discipline.
Auxiliary: An auxiliary ship, which may or may not be a
warship as defined by Article 29 of the United Nations
Convention on the Law of the Sea Part II, is: Not a
combatant, and operated for non-commercial purposes by
the navy, coastguard or other government protection and
security department or agency of a State.
Auxiliary ships are much closer in design to merchant
ships e.g. a replenishment tanker or arctic patrol ship and
ships the military offensive and self-defensive
requirements are be tailored to suit the support role.

2. COMBATANTS AND AUXILIARIES

The Naval Ship Code2 (discussed later in Para 13)

provides a goal based safety regulation framework for
naval ships that is benchmarked against IMO
conventions. This defines ship types as:

Co
Na m a
va m iv
er
lS ci
ta al
n Design Standard (Hostile Damage)
da
rd
s ar
Figure 1: Range of naval combatants and auxiliaries
The wide variety of types of naval ships is illustrated in
Figure 1.
It is traditional though to ensure that all naval ships (both
combatants and auxiliaries) have a high degree of
damage survivability. Whilst designed to withstand
hostile damage, naval ships do perform similar functions
to merchant ships and face similar survivability hazards
such as collision and grounding. Royal Navy ships have
not been without exposure to the consequences of these
hazards as experienced through the incidents for example
with HMS Southampton 1988 (collision), HMS Brazen
1994 & HMS Nottingham 2002 (grounding). The
survival of these ships demonstrates that designing naval
ships to survive hostile damage does provide an
insurance from the consequences of collision and
grounding.
3. IN-SERVICE SURVIVABILITY
MANAGEMENT
Stability is one of the key ship hazards which might
significantly degrade through life due to capability
enhancements. This is common place for naval ships as
new weapons and sensors are fitting during the life of the
ship resulting in corresponding increases in both
displacement and vertical centre of gravity. The key to
success in addressing this issue is to design the ship with
adequate margins. Even with best intentions at the design
stage, the role of the ship may change considerably over
time which may require changes to the ship during
service to restore margins. With defence budgets being
stretched year on year an additional pressure on margins
is the desire to extend the life of ships. Figure 2 shows a
typical reduction in the through life margin to the point
where action was required to maintain the ship to the
original design standard.
The Damaged Ship, London, UK
Design & Build
in
rg
a In-service
M
ty
iil
b Capability Upgrade
Survivability Upgrade
v
r
u
S
St
a nd
ds
Years
Figure 2: Through life margin management.
4. MINIMUM TOLERABLE
SURVIVABILITY
Naval damage standards are generally considered to
provide adequate protection against the consequences of
both weapons and damage typically experienced by
merchant ships. They are widely considered to represent
damage that some WWII shipping survived and are still
relevant in providing a level of survivability against
todays threats. The ability of a ship to survive hostile
damage has a direct cost impact on the design and can be
the subject of a cost capability trade-off between threat
size & risk, susceptibility and self defence. In adopting a
naval standard there is however, no minimum safety
level defined and as such a lack of transparency in
demonstrating the standard is at least as effective as that
required by merchant ships.
For the benefit of this paper the following definitions are
used:
Capability: the ability of a ship to survive a hostile event,
referred to as hostile damage.
Safety: the ability of a ship performing naval duties to
survive a foreseeable event similar to those experienced
by merchant ships, referred to as accidental damage.
Stipulating a safety level will define the minimum
damage extent naval shipping will be expected to
survive. In doing so it is commonly expected that, for
most ships, this level will be less onerous than capability
damage requirements. Introducing a safety limit will
define the lower limit for the survivability margin should
it be degraded (due to unexpected circumstances) below
the design standard (Figure 2). The distinction between
capability and safety in terms of damage stability is
illustrated in Figure 3.
The Damaged Ship, London, UK

Naval Authority. The adoption of statutory requirements,

Combat
Capable for example SOLAS, to provide a minimum safety level
is a feasible option to be considered. From the diagram in
Figure 3 if a ship has no desirable warfighting capability,
Capability loss minimised
Risk of Damage ALARP

Consequences ALARP
then adopting IMO conventions may be an acceptable
Recoverability

Vulnerability
Susceptibly

Stability remain afloat to

preserve capability following solution. However, in doing so, there is an important
hostile damage > foundational difference to note between the philosophy of SOLAS and
Capability

safety levels
key hazard certification:

Military safety
Stability damaged recognised
by ship value & increased risk from SOLAS: Being desirous of promoting safety of
Hazard from military duties ALARP military duties e.g. RAS, life at sea .. to ensure that, from the point of view of
Safety

unrestricted worldwide operation

safety of life, a ship is fit for the service for which it is
Foundational safety Stability static & dynamically
Hazards from environment ALARP stable to provide safe passage and intended
preserve life in event of collision.

Figure 3: Defining Safety and Capability for damage Key Hazard: a significant danger to the lives of
stability several people, loss or severe damage to the platform or
significant damage to the environment
For merchant ships the safety limit is defined by
compliance with IMO Codes & Conventions e.g. The safety case for naval shipping is distinct from
SOLAS, MARPOL, Load Line, High Speed Code, SOLAS in safeguarding the investment inherent in a
Special Purpose Ship Code. In developing a minimum military asset as an inclusive element of Key Hazard
safety threshold for naval ships, all consequences of the safety certification.
hazards must be considered i.e. loss of life, loss of the
platform and damage to the environment. The following There are significant difficulties in applying SOLAS to
principles have been used to define a tolerable naval warships and many regulations would compromise the
safety limit: war-fighting purpose e.g. fitting orange lifeboats along
the upper deck. Applying elements in isolation can also
i) safety requirements are to be at least as effective be fraught with difficulties due to the integrated approach
as statute to protect life and the environment; to safety by the IMO i.e. the link between post damage
ii) the additional risks of foreseeable damage due angle of heel and the distribution of life saving
to naval operations are taken into account; and equipment. The new harmonised SOLAS 2009 damage
iii) the consequences of damage should be stability standard has been applied to an example warship
proportional to the initiating event to protect the material to investigate the implications of adopting such a
military investment. standard. Whilst warships are neither cargo nor
iv) passenger ships the passenger ship requirements were
To illustrate point iii) above, a navy cannot easily replace selected as being the more onerous. The study revealed,
a warship damaged or lost as the result of an accident. that to comply with the Required Index of Subdivision,
Governments do not normally insure their warships and the amount of watertight subdivision could be reduced to
they are not readily available to charter in the same way the minimum required by SOLAS i.e. a collision
merchant ships are. Furthermore, as navies have few bulkhead, bulkheads fore and aft of the machinery space
ships the loss of any single asset has a much wider and an after peak bulkhead as shown in Figure 4. The
impact on the Armed Forces capability. This is ship would achieve the Required Index with any one
increasingly important as the number of naval ships is in main compartment flooded.
decline.
Damage to the environment has been excluded from this
paper however it is expected that the principles can be
applied to this area in due course.

5. ADOPTING IMO CONVENTIONS

Figure 4: Warship subdivision designed to SOLAS
Naval shipping is exempt in the UK from the majority of
the Merchant Shipping Acts and safety is managed using Should such a ship experience collision damage on a
a risk based approach through a safety case1. This bulkhead it would be lost and would not, for a valuable
requires a level of safety that is tolerable, ALARP and at warship, satisfy the principle that the consequences of
least as effective as statutory requirements. One damage are proportional to the event. It is thus
mandatory element of the safety case is the regulation of concluded, for a combatant, adopting SOLAS to define
Key Hazards, for ships this includes stability, structure, the safety limit would not be appropriate, unless the
escape evacuation & rescue, fire, propulsion and military value of the asset is equitable with the value and
manoeuvring systems and explosives. Within the MoD availability of a merchant ship.
this assurance is provided by an independent body, the

6. THE TRADITIONAL CAPABILITY
APPROACH
It has been traditional to apply a naval stability standard
to all ships with a military role, combatants and
auxiliaries alike. Where existing ships have been
designed to merchant standards this approach has
required some tailoring of the standard using the ALARP
principle. Damage extents associated with hostile
weapons are known and the threats and survivability
requirements for each ship type defined i.e. the capability
requirement. Examples of survived hostile damage are
shown in Figure 5.
USS Mansf ield HMS Glamorgan
Figure 5: Examples of survived hostile damage
Damage extents for combatants can be quite large and
generally are proportional to length. The reasons for this
are two fold, firstly it is easier to design a large warship
against relatively large damage than it is for a small
warship and secondly the larger damage extents reflect
the capability value of larger capital ships and the desire
for them to survive a relatively larger threat.
Figure 6: Warship subdivision
Designing to naval standards thus results in greater
subdivision (see Figure 6) than their mercantile
counterparts. As discussed previously in the paper naval
standards traditionally do not directly accidental damage
e.g. collision and grounding. Such damage may on
occasions however be larger than the capability
requirement i.e. as experienced by the HMS Nottingham
grounding (damage to 9 watertight zones) in 2002.
7. FORMAL SAFETY ASSESSMENT
Formal Safety Assessment (FSA) has been described as
"a rational and systematic process for assessing the risks
associated activities and for evaluating the costs and
benefits for reducing these risks. This is a mature
process and can be applied in a number of different ways
to assess hazards and associated mitigating actions. It can
be used to determine the cost benefit of differing risk
control actions and has been applied in constructing IMO
regulations3. The FSA process is shown in Figure 7.
The Damaged Ship, London, UK
Figure 7: Formal Safety Assessment process
In terms of damage stability the capability of a ship to
survive an incident can be expressed in terms of
emergency contingency arrangements when other risk
control measures have failed. Examining the accidents
associated with naval ships, excluding allisions, the
database of accident damage to naval ships is small and
statistically insignificant to use as a basis for determining
a tolerable safety damage extent. The same data used to
derive SOLAS damage extents can however be evaluated
in order to provide a level of safety that is at least as
effective as merchant ships. In addition to this a
mechanism is required that also takes account of the
military asset value and the principle of proportionality.
The FSA framework provides such a methodology and
has been adopted for this paper.
8. ACCIDENTAL DAMAGE EXTENTS
BASED ON MERCHANT SHIP
EXPERIENCE
Naval ships perform more complex roles than their
mercantile counterparts, however hazards such as
collision or grounding exist for both shipping types.
SOLAS provides protection for lives against these
hazards but does not safeguard property in the manner
required by the UK definition of Key Hazards. There is
however direction in IMO to adopt more naval type
approach to survivability in the future in terms of the
ability to remain float and manoeuvre after damage.
The new SOLAS 2009 harmonised passenger and cargo
ship damage extents have been derived from recorded
damage extents to merchant ships albeit using a complex
probabilistic methodology. The data used to derive the
standard emanates from the HARDER4 project. This
same data has been used to investigate the feasibility of
defining safety damage extents for naval shipping.
Evidence from the development of deterministic
regulations and through reverse engineering has shown
that historically SOLAS requirements for Passenger
ships adopted damage extents broadly in the region of a
50th percentile of the HARDER collision statistics. It is
noted the probabilistic approach no longer directly allows
The Damaged Ship, London, UK
this comparison to be easily made. The HARDER
collision data, both damage length and penetration, is
shown in Figure 8. This illustrates that, for example, the
1962 Sarchin & Goldberg5 damage extents of 0.15L &
B/2 adopted by many navies provides a good level of
survivability against accidental damage. At the same
time Figure 8 also shows that structural damage from an
accident may on occasions be greater than that caused by
a hostile event. The extent of damage suffered by HMS
Nottingham is an example of such an incident.
1.0
Non Dimensional Damage Penetration
0.8
0.6
0.4
0.2
0.0
0.00 0.05 0.10
Figure 8: HARDER Collision statistics
Designing a ship against a naval standard to survive
hostile events will result in much more subdivision than
one designed against SOLAS. This is illustrated in the
difference in number of bulkheads between the warships
in Figures 4 & 6. It is intuitive that when an accidental
damage extent standard is derived the number of
bulkheads required would be, for this design, somewhere
between 4 (SOLAS) and 11 (hostile damage).
To address this, FSA methodology has been applied to a
medium sized warship to assess the cost and benefit of a
range of survivability improvements to survive extents of
accidental damage between SOLAS and a traditional
naval standard. This process is illustrated in Figure 9. To
demonstrate the methodology, data from the European
Maritime Safety Agency on the risk of a collision
accident has been combined with the distribution of
HARDER damage extents to derive a risk of loss. A cost
benefit analysis of survivability improvements can be
determined taking account of their associated cost, the
reduced risk of loss and the cost of the ship loss.
Likelihood of
Accident
Improve
survivability
e.g.
subdivision
Figure 9: Accidental damage cost benefit methodology
In performing this process the following assumptions are
used:
The likelihood of a naval ship suffering accidental
damage is similar to that of a merchant ship.
A ship will be lost if extents of damage are greater than
required by SOLAS.
The cost of improved survivability is only associated
with the ship characteristics and subdivision.
A ship life is 30 years and it will take at least 4 years to
replace a naval ship if lost.
Survivability associated with naval standards is ALARP
and greater levels of survivability are not required.
The minimum safety level is defined by SOLAS.
20X
SOLAS Hostile Capability
Cost Benefit
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
10X
Non Dimensional Damage Length
4X
2X
X
Damage Extent
Figure 10: Survivability cost benefit
Survivability improvements come in a range of costs as
do the ships themselves. Figure 10 illustrates the
resultant cost benefit for a range of survivability
improvements in terms of X where Ship Value:X is
5000:1. The adoption of improvements costing 20X or
1/250 of ship value would not provide any benefit over
SOLAS. At the other extreme (X & 2X) the
improvements are wholly reasonable and in fact the
survivability associated with existing naval standards is
an appropriate safety standard for surviving accidental
damage. In between there is a region where the minimum
cost benefit exists between SOLAS and naval standards.
Repeating this process on a range of ships provides a
tailored safety standard for each ship that satisfies the
principles outlined earlier in the paper.
9. STRUCTURAL LOSS TEMPLATES
Assume if
damage >
SOLAS, ship Cost of Loss
The minimum safety damage extents and hostile damage
is lost capability requirements, will together define overall ship
survivability. To ensure there is a coherent approach to
both residual stability and strength, damage templates
Reduced risk
have been developed to define the structural loss
Net Cost associated with both accidental damage and hostile
of loss
Benefit
damage. The templates6 have been incorporated in the
naval architecture design and analysis tool ParamarineTM
to define, in a cost effective manner, floodable extents
Additional
Cost and associated structural loss. A brief description of the
two categories of templates follows.

10. ACCIDENTAL DAMAGE TEMPLATES
Accidental Damage Templates (ADT) define both
collision and grounding events. The templates contain
varying degrees of structural loss e.g. 100% loss of
structure at the centre of damage to 10% loss where
tearing may occur at the edge. Application of the
templates for collision and a number of grounding cases
are illustrated in Figures 11 & 12.
Figure 11: Accidental Damage Template Collision
Figure 12: Accidental Damage Templates - Grounding
11. HOSTILE DAMAGE TEMPLATES
In turn, Hostile Damage Templates (HDT) reflect the
structural loss associated with explosion and blast
damage of above water and underwater weapons. The
application of these templates is illustrated in Figures 13
& 14. Varying degrees of structural loss are incorporated
e.g. 100% loss at the core to 10% loss at the extremity
due to blast fragments. As with ADTs, for the purposes
of damage stability only the outer edge of the template is
used to identify the compartments at risk of flooding.
The Damaged Ship, London, UK
Figure 13: Hostile Damage Template Above water
Figure 14: Hostile Damage Template - Underwater
12. APPLICATION OF DAMAGE
TEMPLATES
The ADTs & HDTs have been applied, on a stability
only basis, in a pilot study to an in-service warship. To
directly compare the traditional approach (determining
the worst case combination of compartments) and the
structural loss templates the same damage length has
been used for both. As expected dependent upon the
arrangement of subdivision there is duplication of some
damage cases. However the greatest benefit of the
approach is the simplified and rational manner in which
damage cases are defined. For example, for a single
loading condition the traditional approach defined 207
damage cases of which 70 failed the residual stability
criteria. The damage template approach identified 110
cases with only 3 minor failing cases. The cases not
identified by the damage template approach are
considered to be unrealistic and of an acceptably low risk
of occurrence.
Warship damage stability assessments have long been
considered by naval architects to be highly complex and
expensive. The ADT & HDT pilot study has rationalised
the approach to naval damage stability providing a cost
effective method for identifying damage cases and
furthermore transparency between accidental and hostile
damage.
The Damaged Ship, London, UK

13. NAVAL SHIP CODE
The overall aim of the Naval Ship Code2 (NSC) is to
provide a goal based framework for a naval surface ship
safety management system based on and benchmarked
against IMO conventions and resolutions that embraces
the majority of ships operated by Navies. To develop the
NSC the performance goals of IMO conventions were
interpreted and where appropriate amended to take
account of naval operations and doctrine as illustrated in
Figure 15. These form the goals of the NSC which is
now maintained by the International Naval Safety
Association (INSA) with membership from 11 navies and
7 Naval Classification societies.
that experienced by weapon damage. The Formal Safety
Assessment approach can be used to provide a cost
benefit framework for defining a naval safety damage
standard. This takes account of the value of the ship and
at the same time uses an approach allowing
benchmarking against IMO conventions without the need
to directly demonstrate compliance. Making a
distinction between accidental and hostile damage
provides a minimum safety standard onto which the
capability to survive hostile damage can be tailored. With
the current direction of the Naval Ship Code to define
optional safety standards for naval ships, the approach
outlined in this paper is suitable as a candidate option to
satisfy the NSC damage extents performance
requirement.

15. ACKNOWLEDGEMENTS

Thanks are expressed to the following for their

contributions to the programme.
0 Aim 0 Aim

1 Goals 1 Goals

2 Functional Areas
1 2 3 4 5 6

2 Functional Ar eas
7 8 9
Mr D N Smith Ministry of Defence
3 Requirements 3 Requirements
Mr A Peters & Mr A King QinetiQ
Mr C Burden- QinetiQ GRC
Cla

Na
va
ss

4 Verification 4 Verification

5 Justification 5 Justification

IMO (SOLAS/HSC) NATO NSC

16. REFERENCES
Figure 15: Naval Ship Code goal based approach
1. JSP430 MoD Ship Safety Management
The NSC provides both the foundation and military
2. Naval Ship Code Allied Naval Engineering Publication
safety level as illustrated in Figure 3.
77 Edition 2
3. IMO MSC Circ 1023.
Compliance with the NSC is voluntary and it is the
4. IMO MSC 79/11/5
responsibility of a navy to specify the standards to
5. Sarchin, T H. and Goldberg, L. L., Stability and
achieve the performance requirements. The current
Buoyancy Criteria for U.S. Naval Surface Ships
strategy of INSA is to define an optional set of naval
SNAME Transactions, Vol. 70, 1962.
safety standards that are solutions to the code with Naval
6. Smith, D Accidental Damage Templates (ADTs), A
Classification a suitable solution for certain chapters. To
basis for the future of Naval Ship Safety Certification
date the NSC does not contain solutions to the
10th International Conference on Stability of Ships and
performance requirements of Chapter III Stability and
Ocean Vehicles
Buoyancy.

The FSA methodology described in this paper would

17. AUTHORS BIOGRAPHY
provide damage extents defining a safety level for naval
ships and thus a candidate solution for NSC Chapter III.
Steve Marshall is the Naval Authority providing
independent regulation for the stability of all MoD
14. CONCLUSIONS
shipping.
Naval ships vary in their role and type from front line
This paper represents the views of the author and not that
combatants designed to remain afloat following
of the Ministry of Defence.
substantial damage to auxiliaries forming a supporting
role that, for some, look very similar to merchant ships. It
has been traditional practice for both types of ships to
comply with a common naval damage stability standard
albeit occasionally tailored for the support ships. This
standard provides a good mitigation against, but does not
directly address, accidental damage.

Naval ships are generally exempt from IMO conventions

however like merchant ships they can also suffer
accidental damage that on occasions, can be greater than
The Damaged Ship, London, UK

AUTHORS CONTACT DETAILS

D. Vassalos D Fone
University of Strathclyde, UK University College London, UK
d.vassalos@strath.ac.uk d_fone@meng.ucl.ac.uk

J Gullaksen A C Gaillard
JG Consultant Engineers, Denmark University College London, UK
jg@jg-consultant.com acgaillard@gmail.com

K W Hutchinson A. Martin
Babcock International Group, UK QinetiQ Ship Structures Team, UK
Keith.Hutchinson@babcock.co.uk AAMARTIN@qinetiq.com

N Hills A Harman
MoD, UK Royal National Lifeboat Institution (RNLI),
Nigel.Hills749@mod.uk UK
Antony_Harman@rnli.org.uk
R Gregory,
Noble Denton, UK T Martins
richard.gregory@nobledenton.com CINAV Portuguese Navy, Portugal
triunfante.martins@npo.webside.pt
R Perez
Escuela Tcnica Superior de Ingenieros P. Mangriotis
Navales (Universidad Politcnica de Madrid), London Offshore Consultants, UK
Spain p.mangriotis@loc-group.com
rodriperfer@hotmail.com
I Wallbridge
S E Ratcliffe Hart Fenton & Company Ltd, UK
Atkins, UK Iain.Wallbridge@hart-fenton.com
Simon.Ratcliffe@atkinsglobal.com
G Wang
C D Wood American Bureau of Shipping, USA
University of Southampton, UK GWang@eagle.org
C.D.Wood@soton.ac.uk
S Marshall
R Tagg Ship Stability Safety Regulator, Ministry of
Herbert Engineering Shanghai, China Defence, UK
rtagg@herbertsoftware.com DESSESea-ShipStab@mod.uk

M Shahid J R Ward
Binary Systems and Engineering (UK) Ltd, UK Ministry of Defence, UK
shahid@bsyse.co.uk DESSALMO-DTLSO@mod.uk

S. Kwon
The Ship Stability Research Centre, UK
seung.kwon@strath.ac.uk
 The Damaged Ship
The Steepest Learning Curve Yet
D. Vassalos, A. Jasionowski and L. Guarin
The Ship Stability Research Centre, Department of Naval Architecture and
Marine Engineering, University of Strathclyde, Glasgow
The Damaged Ship Conference, RINA, London 26-27 January 2011
In Partnership with
Safety at Sea Ltd., Glasgow, Scotland, UK

1
 Presentation Outline

Basic Definitions / Fundamental Concepts

Rule Development Deterministic / Probabilistic
Damage Stability Fundamentals
Contemporary Developments / Safety Level
(Total Risk) Modelling
Design Implementation
Concluding Remarks

2
www.safety-at-sea.co.uk
 Presentation Outline

Basic Definitions / Fundamental Concepts

Rule Development Deterministic / Probabilistic
Damage Stability Fundamentals
Contemporary Developments / Safety Level
(Total Risk) Modelling
Design Implementation
Concluding Remarks

3
www.safety-at-sea.co.uk
 Basic Definitions
Floatability

100,000 tonnes
Heavy, but it will
fall in the same
way as an apple

Sir Isaac Newton, 1643 1727,

Proposed the law of gravity at the age of 24

The weight will sink, unless www.safety-at-sea.co.uk

 Basic Definitions
Floatability

100,000 tonnes
Courtesy of Mr
Archimedes of Syracuse, it is supported by
287 BC c. 212 BC a submerged air
bubble, displacing
equivalent weight
100,000 tonnes

www.safety-at-sea.co.uk
Intact Stability in Calm Water
Zero Speed (Hydro) Static Stability

www.safety-at-sea.co.uk
Zero Speed Beam Waves

www.safety-at-sea.co.uk
Zero Speed Beam Waves
Water on Deck

www.safety-at-sea.co.uk
Zero Speed Longitudinal Seas
Parametric Rolling

www.safety-at-sea.co.uk
 Forward Speed F/Q Waves

Sub-harmonic Rolling Harmonic Rolling

www.safety-at-sea.co.uk
Forward Speed F/Q Waves
Surf-riding and broaching

www.safety-at-sea.co.uk
 Basic Definitions
Floatability / Damage Stability

100,000 tonnes
Courtesy of Mr
Archimedes of Syracuse, it is supported by
287 BC c. 212 BC a submerged air
bubble, displacing
equivalent weight
100,000 tonnes

What would
happen if WT
integrity is lost? www.safety-at-sea.co.uk
 Basic Definitions
Floatability / Damage Stability in Calm Water
Metaphorically
speaking this will
depend on how
big the damage is What would
and how many air happen if WT
bubbles support integrity is lost?
the weight.

Deterministic?
Probabilistic?
www.safety-at-sea.co.uk
Damage Stability in Calm Water
Zero Speed (Hydro) Static Stability
1.1
1
0.9
0.8

0.7
0.6

Righting lever [m]

0.5 Participant 1
Participant 2
0.4
Participant 3
0.3
Participant 4
0.2
Participant 5
0.1
0
0 10 20 30 40 50 60
-0.1
-0.2
-0.3

Heel angle [deg]

GZ curves - Intact Ship

0.4

0.3

0.2

Righting lever [m]

0.1

0
0 5 10 15 20
Participant 1
-0.1 Participant 2
Participant 3
-0.2 Participant 4
Participant 5

-0.3

Heel angle [deg]

Midship Damage Case GZ curves - Damaged Ship

www.safety-at-sea.co.uk
Zero Speed Damage Stability in Beam Seas
Modes of Capsize

Low Freeboard Ro-Ro

Ships

High Freeboard Ships

Low Freeboard
Conventional Ships

Large Passenger Ships

www.safety-at-sea.co.uk
Low freeboard Ro-Ro ships

Side View Deck View

www.safety-at-sea.co.uk
Low Freeboard Conventional Ships

Side View Deck View

www.safety-at-sea.co.uk
Large Passenger Ships
Intermediate Stages of Flooding MFS Effect

www.safety-at-sea.co.uk
 Intact Stability Vs Damage Stability
Mainly a forward speed longitudinal seas Mainly a zero speed beam seas problem
problem (except for small ships)

Sudden death dynamic capsize Slow death quasi-static capsize

Time-invariant system Time-varying system

Principal uncertainties near escape Principal uncertainties away from capsize

boundaries boundaries

Coupling of manoeuvring and seakeeping Coupling of flooding and seakeeping

models models

Main governing factor is roll restoring Main governing factors are roll restoring
and freeboard

With passenger ships main emphasis on
passenger comfort; with all other ships
low interest on intact stability
With passenger ships main emphasis on
passenger survival (time to capsize); with
all other ships strong interest on damage
stability (impact on environment).
www.safety-at-sea.co.uk
 Presentation Outline

Basic Definitions / Fundamental Concepts

Rule Development Deterministic / Probabilistic
Damage Stability Fundamentals
Contemporary Developments / Safety Level
(Total Risk) Modelling
Design Implementation
Concluding Remarks

20
www.safety-at-sea.co.uk
 Passenger Ro-Ro Vessel Disasters
Relevant Statistics 85% of Deaths due to Car Deck Flooding

1953 Princess Victoria, capsize and sank, open stern

door, car deck flooding
1974 Straitsman, capsized and sank with vehicle door
partly open whilst approaching berth, car deck
flooding because of squat
1987 Santa Margarita Dos, car deck flooding
1987 H. of F. E., car deck flooding
1994 Estonia, car deck flooding
2006 Al Salam Boccaccio 98, fire car deck flooding
(1,000+ deaths)
www.safety-at-sea.co.uk
Rule Development
Swiss Cheese

www.safety-at-sea.co.uk
Deterministic Rules
Standards

www.safety-at-sea.co.uk
 Probabilistic Concept of Ship Subdivision

J I ^
A w
j 1 i 1
j . p i .s ij ; A R A E (s)

A/R Attained/Required Index of Subdivision

j loading condition (draught) under consideration
J number of loading conditions considered
i compartment or group of compartments under consideration
for each j
I all feasible flooding scenarios (single compartments or
groups of adjacent compartments) for each j
w probability mass function of the loading conditions (.4, .4, .2)
pi probability mass function of the extent of flooding
si probability of surviving the flooding scenarios under
consideration at a given j www.safety-at-sea.co.uk
 Probabilistic Rules (SOLAS 2009)
Risk Characterisation

Rc Pc Pw / c Pf / w / c Cc
Pc Probability of a collision event dependent on the
PC/S area of operation, geography/topology/bathymetry,
route, traffic density, ship type, loading condition,
etc.
1 A 1 pi si Pw/c Probability of water ingress conditional on collision
iI
and dependent on crashworthiness
Pf/w/c Probability of capsize/sinkage/collapse conditional
on collision and water ingress; expressed as a
function of sea state, structural strength and time
Cc Consequences deriving from the said collision
event, accounting for loss of (or injury to) life and
property and for impact to the environment www.safety-at-sea.co.uk
 Probabilistic Rules (SOLAS 2009)
Risk Characterisation
0.012 2 Comp

p*(1-s) 3 Comp

4 Comp
0.01
5 Comp

0.008

0.006

0.004

0.002

0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250

3 Comp sample

www.safety-at-sea.co.uk
 Presentation Outline

Basic Definitions / Fundamental Concepts

Rule Development Deterministic / Probabilistic
Damage Stability Fundamentals
Contemporary Developments / Safety Level
(Total Risk) Modelling
Design Implementation
Concluding Remarks

27
www.safety-at-sea.co.uk
 Damage Stability Fundamentals
The PROTEUS Suite

Decision Risk-Based Design

support Safety level Evi

High-speed
concepts
Moorings
Proteus3
PROTEUS Cables
Seakeeping
Stability in waves

Damage
Damage survivability
survivability Hydroelastic
Floodwater dynamics
Floodwater dynamics analysis

RANSE
Manoeuvring
qualities
Cargo-shifting www.safety-at-sea.co.uk
 Advances in Damage Stability/Survivability

Time-domain simulators (PROTEUS)

6DOF, non-inertial systems of reference
Geometric non-linearities in hydrodynamics (restoring
/ FK instantaneous free surface); 2nd order drift
forces, wind and current effects, water on deck
Water Ingress/Egress and floodwater dynamics
Survival boundaries and criteria
Regulatory issues (Probabilistic Rules)
Routine application to concept design, design
evaluation and crisis management

www.safety-at-sea.co.uk
 Mathematical Modelling
Damaged Ship

AA MM m u u B B u u CC u u FF F
ijij ij ij
w
ij j j ij ij j j ij ij jj j j
w

d
M 'I I 'w '
dt

M ' g M w r ' AGw g '

M ' N M w ' r ' AGw v ' AGw
d
M w r ' AGw v ' AGw ' v ' AGw
dt
d
I ' w ' 'I ' w '
dt www.safety-at-sea.co.uk
 Modelling Water Ingress/Egress (Bernoulli)
Level 1 model
800

700 Experimental measurment

Theoretical
600

SUM( dQ/dt ) [m3/s] kkkkk

500

400

300

E R
200

100

0
0 500 1000 1500 2000 2500

vw Time [s]

vs
dhR
dh

zw
n

sgn dh K v f dA
dQ
dt
www.safety-at-sea.co.uk
 Free-Mass-on-Potential Surface Floodwater Model
Level 2 model
Prototype Tank length 0.1m, breadth 1.0m, depth 0.5m, fluid height 0.06m,
rotation axis at (y=0,z=0), ampl 0.1rad

rectangular 0.2

compartment

Ka, Moment about x axis [N m]

h
0.15

used for
verification H
b
0.1

studies of
Experiments by van den Bosch and Vugts
water sloshing b
0.05
Proteus3, fluid motion in phase with oscillations
Proteus3, fluid motion based on PTFM, mi=0.10
Proteus3, fluid motion based on PTFM, mi=0.15

0
0 1 2 3 4 5 6
Freq [rad/s]

Tank length 0.1m, breadth 1.0m, depth 0.5m, fluid height 0.06m,
rotation axis at (y=0,z=0), ampl 0.1rad
180

Phase angle of tank moment [deg]

135

90

Experiments by van den Bosch and Vugts

45
Proteus3, fluid motion in phase with oscillations
Proteus3, fluid motion based on PTFM, mi=0.10
Proteus3, fluid motion based on PTFM, mi=0.15

0
0 1 2 3 4 5 6

Comparison of fluid moment

Freq [rad/s]
amplitudes and
Description of geometric constraints for phase angles derived by experiments and in-
motion of centre of buoyancy of floodwater phase FMPS sloshing modelswww.safety-at-sea.co.uk
Modelling Water Ingress/Egress (CFD)
Level 3 model

www.safety-at-sea.co.uk
Modelling Floodwater Dynamics (CFD / SPH)
Level 3 model

www.safety-at-sea.co.uk
Damaged Ship Dynamics
CFD

www.safety-at-sea.co.uk
 Presentation Outline

Basic Definitions / Fundamental Concepts

Rule Development Deterministic / Probabilistic
Damage Stability Fundamentals
Contemporary Developments / Safety Level
(Total Risk) Modelling
Design Implementation
Concluding Remarks

36
www.safety-at-sea.co.uk
 Contemporary Developments
The IMO Framework for Passenger Ship Safety

IMO (SLF 47/48) Passenger Ship Safety

Casualty Threshold Return to Port

100% vessel survivability (RTP)
(indefinitely)

Abandon Ship

Flooding
100% vessel survivability
Fire for a specified period of time [3h]

Ship functions / systems availability after a casualty

37
www.safety-at-sea.co.uk
 Safety Level Evaluation Framework

Casualty Threshold /Safe Return to Port

Accident Causality Analysis Consequence Analysis Mitigation Analysis

Systems Availability
Collision
Collision Scenarios Flooding survivability analysis
Evacuation & Rescue

Grounding/ Systems Availability

Grounding/
Stranding Scenarios Flooding survivability analysis
Stranding Evacuation & Rescue

Systems Availability
Fire
Fire Scenarios Fire safety analysis
Evacuation & Rescue

Safety Level (Total Risk)

38
www.safety-at-sea.co.uk
 Ship Safety Level (Total Risk)

Risk is a chance of loss of life

The chance is measured by statistics

loss scenarios:

flooding
~90% of
fire the risk
loss of life
intact stability loss
(expected number of fatalities per year)

other

39
www.safety-at-sea.co.uk
 Example Loss Scenario
Flooding | Collision

prevention
Navigation failure

mitigation

Water ingress (hull breach)

Loss of stability

Abandonment

40
www.safety-at-sea.co.uk
 Risk Model

N max
Risk PLL E N F i N
i 1

1E+00

Frequency of N or more fatalities per ship year

1E-01

N max
FN N fr i
1E-02

N 1E-03

i N 1E-04

1E-05
1 10 100 1000 10000
Fatalities [N]

41
www.safety-at-sea.co.uk
Risk Model

frN N frhz hz j prN N hz j
nhz

j 1

42
www.safety-at-sea.co.uk
 Flooding Survivability Analysis

Safe Return to Port / Casualty Threshold

grounding collision

Causality Analysis Consequence Analysis Mitigation Analysis

Statistics Statutory Assessment First Principles Analysis Time to Capsize

HAZID SOLAS (A-index) -transient flooding Systems Evacuation
Modelling -cross flooding Availability and Rescue
-progressive flooding

Statistics First Principles Analysis Time to capsize

HAZID -transient flooding Systems Evacuation
Modelling -cross flooding Availability and Rescue
-progressive flooding

Statistics Statutory Assessment First Principles Analysis Time to untenable

fire

HAZID SOLAS ChII -fire/smoke propagation conditions Systems Evacuation

Modelling Availability and Rescue

Safety Level

www.safety-at-sea.co.uk
 Flooding Survivability Analysis
100s of Compartments, 1000s of damage scenarios

Statutory Assessment
Compliance with SOLAS 2009 (probabilistic rules)
Optimisation of watertight subdivision
Flooding Risk Analysis
Frequency
Consequences
Time to Capsize
Analytical and performance-based approaches
Vulnerability assessment (as designed / as operated)
Time to Abandon Ship
Assembly and evacuation performance
Evaluation of casualty threshold / return to port / safety level
Probabilistic approach; link to system availability post-casualty
44
www.safety-at-sea.co.uk
 Flooding Survivability Analysis
Statutory Assessment SOLAS 2009

Aft peak bulkhead Machinery space bulkhead Collision bulkhead

New requirements for Minor damage concept (still deterministic) for passenger
double bottom vessels, but no specific requirements on location of watertight
subdivision. Required index to be met

( pi si ) = A > R
i=1
45
www.safety-at-sea.co.uk
Flooding Survivability Analysis
Statutory Assessment Platfrom Optimisation (param. model)

Bulkhead location fixed

Deck or Bulkhead
location not fixed
(possibly removed)

Cargo Stores area in

the lower hold
Fixed bulkheadMFB

www.safety-at-sea.co.uk
 Flooding Survivability Analysis
Statutory Assessment Platfrom Optimisation (design variants)

0.955

0.95

0.945

0.94

0.935

0.93

0.925

0.92

0.915

0.91
0 200 400 600 800 1000 1200
Feasible Designs Pareto-optimal Designs

www.safety-at-sea.co.uk
 Flooding Risk Analysis
Frequency Analysis (Historical Data)

frhz hz1 Source: DNV

1.148 E-3 1/sy collision

FSA Cruise
Ships grounding
(SAFEDOR,
FSA, 2007):
1 event fire
every 871
ship years 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014

Frequency per Ship Year

Frequency of event occurrence

48
www.safety-at-sea.co.uk
 Flooding Risk Analysis
Consequence Analysis (Impact on Human Life)

N max

N evac t

N fail t

t
tcap
49
www.safety-at-sea.co.uk
Flooding Risk Analysis
Time to Capsize

50
www.safety-at-sea.co.uk
 Time to Capsize
Performance-Based Monte Carlo Simulation

Damage case Outcome Implication

Water Case i=1 t(i)
ingress? Case i=2 t(2) Vessel unable to survive
for 3h
yes Numerical
Case i=k simulations t(k)
Vessel survives for
at least 3h (t)
Collision Case i=342 t(342)
Model
tests

no Minor incident

Performance-based
evaluation and verification t = time to capsize

51
www.safety-at-sea.co.uk
 Time to Capsize
Typical Results (Ro-Pax)
Analytical estimates of time to
capsize based on SOLAS 2009 s-
factor agree reasonably well with
results from numerical simulations 30% of possible collision scenarios
would lead to capsize within 30 min.

14% of possible collision scenarios

would lead to capsize within 30 min.

www.safety-at-sea.co.uk
 Presentation Outline

Basic Definitions / Fundamental Concepts

Rule Development Deterministic / Probabilistic
Damage Stability Fundamentals
Contemporary Developments / Safety Level
(Total Risk) Modelling
Design Implementation
Concluding Remarks

53
www.safety-at-sea.co.uk
 Oasis of the Seas
General particulars

Gross Tonnage 225 000 Length 361 m

Delivery November 2009 Breadth 47 m
DNV 1A1, Passenger Ship: Draught 9,15 m
EC0, RPS, F-M, LCS(DIS), BIS,
Air Draught 72 - 65 m
TMON, CLEAN, COMF(V)1, FUEL
Number of Guests 5400
Design basis: E0, DP notation
AUTR, NAUT-AW Number of Crew 2166
LSA Capacity 8460 54
www.safety-at-sea.co.uk
 Oasis of the Seas
Flooding Survivability Analysis

Case by case explicit dynamic flooding simulation

Transient and progressive flooding
Impact of multi free-surfaces
Impact of watertight and semi-watertight doors and arrangements
www.safety-at-sea.co.uk
 Flooding Survivability Analysis
Evaluation and Verification (Typical Output)

SWT doors D2 holding!

windows under water!

www.safety-at-sea.co.uk
 Flooding Survivability Analysis
Time to Capsize
Analytical estimates of time to capsize based on SOLAS 2009 s-
factor underestimates the level of survivability in relation to results
from numerical simulations and model tests.

SOLAS 2009 (Analytical)

10.5% of possible flooding cases due to collision will
lead to capsize within 30 min.
(attained index A0.895)

Simulations (performance-based)
1.2% of possible collision scenarios would lead to
capsize within 30 min..

Probability distribution of time to capsize

www.safety-at-sea.co.uk
Flooding Survivability Analysis
Evaluation and Verification
Next Page

717 compartments, 1160 openings (1:50 scale)

www.safety-at-sea.co.uk
 Oasis of the Seas
Flooding Survivability Analysis
Previous Page

Societal Risk
Collision Accidents
1.E+00

1.E-01
Frequency of N or more fatalities

1.E-02

FSA Cruise Ships (DNV, SAFEDOR)

1.E-03

1.E-04

Genesis
1.E-05

1.E-06
1 10 100 1000 10000
N Fatalities
www.safety-at-sea.co.uk
Crisis Management
The Oasis of the Seas Case \study

Can the extensive knowledge acquired during the

design development be used to manage
operational / accidental risk?

60
60
www.safety-at-sea.co.uk
Crisis Management
Onboard Decision Support System

Real time sensors and hardware integration (link to

ships SMS): tank levels, draughts, door status,
water ingress alarms, wind and wves
Vulnerability log: global and local ship vulnerability
to flooding
Criticality assessment: survival time, escape and
evacuation time (crises management)
Corrective action search: evaluation of the impact
of corrective actions. Ballast system availability.
Essential systems availability post-flooding
(verification of compliance of SRtP requirements)
61
61
www.safety-at-sea.co.uk
 62

62
www.safety-at-sea.co.uk
 Presentation Outline

Basic Definitions / Fundamental Concepts

Rule Development Deterministic / Probabilistic
Damage Stability Fundamentals
Contemporary Developments / Safety Level
(Total Risk) Modelling
Design Implementation
Concluding Remarks

63
www.safety-at-sea.co.uk
 Concluding Remarks
A painstaking evolutionary development in the
subject of damage stability is giving way to
unprecedented scientific and technological
changes at an ever increasing pace
necessitating due care to ensure smooth
transition in managing this change and the
emerging complexity.
The probabilistic framework for damage stability
offers flexibility and added degrees of
freedom for designers to enhance safety cost-
effectively.
64
www.safety-at-sea.co.uk
 Concluding Remarks

With a clear trend towards probabilistic and risk-

based frameworks to addressing ship safety in a
holistic manner, it is important to base such
developments on clear understanding of the
underlying principles.
Despite problems faced by the industry to cross
the bridge from Rules-Based to Goals-Based
(Risk-Based) Design, this paper demonstrates that
pre-requisite scientific and technological
developments are in hand for Risk-Based Design
to be fully implemented in the maritime industry.
65
www.safety-at-sea.co.uk
 Concluding Remarks

Operational risk management could potentially fill

design gaps and facilitate crisis management in
emergencies most effectively, using Decision
Support Systems onboard that integrate
pertinent contemporary knowledge and
experience to aid decision-making.

66
www.safety-at-sea.co.uk
A Practical Guide to Damage Stability Assessment
Regulation on Damage Stability

International Conference The Damaged Ship

January 26nd & 27rd, 2011
@ RINA HQ, LONDON, UK

Jannes Gullaksen
Naval Architect, MSc (Applied Mathematics)

JG Consultant Engineers, Denmark

My Background
Specialized in naval architecture and marine engineering. Previous experience are
within naval architecture and marine engineering, including ship hydrostatics,
stability and strength.

Education:
M.Sc. in Engineering (Applied Mathematics)
B.Sc. in Naval Architecture (Marine Engineering)
B.Sc. in Business Economics (International Marketing/Trade)

Owner of JG Consultant Engineers, http://www.jg-consultant.com, and a Maritime

Consultant. At present working with a renewable energy project. Developing a
mathematical model for wave load and energy extraction.

The concept of a paper addressing this subject was initially inspired by my draft
manuscript to a book, titled Naval Architecture: Geometry, Hydrostatics, Stability,
Dynamics, Strength and Economics - A Computer Based Approach Using Excel
Spreadsheets with VBA, to be published 2011/2012.
Content of presentation
Introduction
Deterministic Damage Stability
Probabilistic Damage Stability
Damage Stability Information
Preparation of Input to Probabilistic Damage Stability
Assessment
Procedure for Probabilistic Damage Stability Calculations
Output from Probabilistic Damage Stability Calculations
Summary and Conslusion
Introduction
What is damage stability
Purpose of Damage Stability Assessment
Approaches to damage stability assessment
What is damage stability
Damage stability of a ship is its capacity to
resist damage situations caused by flooding due
to the water overflows into bilges while ship
stability is all about ship performance on still
water and waves. Centre of gravity and centre
of buoyancy of ship is the main factors taken
into account in ship stability.
Purpose of Damage Stability Assessment
In addition to what stability information have to be submitted
to the national authority and the society, the following
technical aspects can be mentioned:
Develop a set of min GM or allowable KG curves to provide an overall
assessment of stability
Demonstrate that the ship meets vulnerability or recoverability requirements

Provide technical guidance to ship designers for areas where improvements

can be made
Ensure that damage stability performance is not degraded during ship design
and construction
Approaches to be followed in damage
stability assessment
In order to assess the behavior of the ship after
damage, two different approaches are used for damage
stability assessment:
the deterministic approach and
the probabilistic approach,
which are to be applied depending on the ship
type.
Deterministic Damage Stability Defined
The deterministic approach is based on standard
dimensions of damage extending anywhere along the
ships length or between transverse bulkheads depending
on the relevant requirements. The consequence of such
standard of damage is the creation of a group of damage
cases, the number of which, as well as the number of
compartments involved in each case, depend on the ships
dimensions and internal subdivision.
Application: SOLAS Part B, footnote to Regulation 4.1
Codes (Deterministic)
(SOLAS Part B, footnote to Regulation 4.1)
International Bulk Chemical (IBC) Code. Note that where
a chemical carrier can carry petroleum grades in addition
to noxious liquid substances in bulk, MARPOL Annex I
also applies.
The BCH Code,
International Gas Carrier (IGC) Code,
Guidelines for the design and construction of Offshore
Supply Vessel (OSV) Guidelines (Resolution A.469(XII)),
International Code of Safety for High-Speed Craft (2000
HSC Code),
 continued
MARPOL Annex I specifies oil tanker subdivision &
damage stability requirements including double-hulls,
except combination carrier (as defined in regulation II-
2/3.14) with type B freeboards are not excluded.
MARPOL Annex I, Regulation 28.
Damage stability requirements of regulation 27 of the
1966 Load Line Convention
Damage stability requirements of regulation 27 of the
1988 Load Line Protocol
 continued
(SOLAS probabilistic exceptions)
Within EU, the damage stability for ro-ro passenger ships is
given in Directive 2003/25/EC and 2009/45/EC of the
European Parliament.
Besides the probabilistic component, the new SOLAS 2009
Part. B-1 also contains a deterministic component concerning
passenger ships, Reg 8.
The SOLAS 2009 Part B-2 also contains a deterministic
component regarding bottom damages. According to SOLAS
2009, Regulation 9.8.
For Large Commercial Yacht, the damage stability is given by
the UK MCA, Large Commercial Yacht Code (LY2) - MSN
1792 (M), UK National Authority, MCA.
Deterministic Damage Stability Assessment
Ship survival capability
Damage Assumptions
Flooding Assumptions
Standard of Damage
Survival Requirements - Damage Stability Criteria
Damage Assumptions
Assumed maximum extent of bottom damage

Assumed maximum extent of side damage

Flooding Assumptions
The survival requirements regarding flooding, shall be
confirmed by calculations which take into consideration
the design characteristics of the ship; the arrangements,
configuration and contents of the damaged compartments;
the distribution, relative densities and the free surface
effects of liquids; and the draught and trim for all
conditions of loading.
The permeability of spaces assumed to be damaged shall
be considered in accordance with the Code.
Standard of Damage
Ship shall be capable of surviving the damage indicated in
the damage assumptions, with the flooding assumptions to
the extent determined by the ships type according to the
standard.
Survival Requirements and Damage
Stability Criteria
Ship subject to the Code shall be capable of surviving the
assumed damage specified in the damage assumptions to
the standard provided in the standard of damage in a
condition of stable equilibrium and shall satisfy the
criteria in the standard.
For each loading condition, each damage case is to be
considered, and the applicable criteria are to be complied
with.
Probabilistic Damage Stability
The requirements for damage stability calculation
refer here to SOLAS Chapter II-1 as amended ,
Parts B-1 through B-4 and the related Explanatory
Notes.
Part B: Subdivision and stability
Part B-1: Stability (Probabilistic damage stability)
Part B-2: Subdivision, watertight and weathertight
integrity
Part B-3: Subdivision load line assignment for
passenger ships
Part B-4: Stability management
Application
In SOLAS2009 two versions of the probabilistic method
are prescribed: one for passenger ships and one for cargo
vessels.
All passenger ships regardless of length. Passenger ship
(SOLAS) is ship which carries more than 12 passengers
Cargo ships of 80 m in length and upwards. Cargo ship
(SOLAS) is any ship which is not a passenger ship
cargo ships which are shown to comply with subdivision
and damage stability regulations in other instruments are
excluded
The basic framework for Probabilistic
Damage Stability Calculations
SOLAS Ch.II-1 Part B-1
REQUIRED SUBDIVISION INDEX, Regulation 6
ATTAINED SUBDIVISION INDEX, Regulation 7
The pi factor
The r factor and b factor

The s factor
The v factor

SOLAS 2009, Part B-2

Subdivision, watertight and weathertight integrity
Probabilistic Damage Stability Defined
The probabilistic approach refers to a method for computing a
measure of probable damage survivability which takes into account
not only the vessels stability when certain subdivisions of the ship
are damaged but assigns probabilities to various extents of damage as
well as to the survivability.
The probabilistic concept takes the probability of survival after
collision as a measure of ship safety in the damaged condition,
referred to as the attained subdivision index A.
The damage stability calculations are performed for three draughts
and relevant GM values in order to draw a minimum GM curve
where the attained subdivision index A achieves the minimum
required level of safety R.
Each case of damage is not required to comply with the applicable
criteria, but the attained index A, which is the sum of the
contribution of all damage cases, is to be equal to or greater than R.
Central concepts in calculation of index A
Subdivision Length (Ls) of the ship is the greatest projected
molded length of that part of the ship at or below the deck that
limiting the vertical extent of flooding with the ship at the
deepest subdivision draught.
Deepest Subdivision draught (ds) is the waterline which
corresponds to the Summer Load Line draught of the ship.
Light service draught (dl) is the light service draught
corresponding to the lightest anticipating loading and
associated tankage, including, however, such ballast as may be
necessary for stability and/or immersion. Passenger ships
should include the full complement of passengers and crew on
board.
Partial subdivision draught (dp) is the light service draught
plus 60% of the difference between the light service draught
and the deepest subdivision draught.
REQUIRED SUBDIVISION INDEX, R
For all ships to which the damage stability requirements
apply, the degree of subdivision to be provided shall be
determined by the required subdivision index R, as
follows:
For cargo ships greater than 100 m in length Ls are
considered

SOLAS Regulation 6 for details

ATTAINED SUBDIVISION INDEX, A
The method of calculating the A for a loading condition is
expressed by the formula:

where
i represents each compartment or group of compartments
pi accounts for the probability that only the compartment or group of
compartments under consideration may be flooded, disregarding any horizontal
subdivision,
si accounts for the probability of survival after flooding the compartment or
group of compartments under consideration, and includes the effect of any
horizontal subdivision.
c represents one of the three loading conditions. To obtain a maximum index A
for a given subdivision, t has to be equal to T, the total number of damages.
... Attained index A continued ...
The attained subdivision index A is obtained by the summation
of the partial indices As, Ap and Al, (weighted as shown)
calculated for the draughts ds, dp and dl in accordance with the
following formula:

where the indices s, p and l represent the three loading

conditions and the factor to be multiplied to the index indicates
how the index A from each loading condition is weighted.
In summary index A is based on:
The probabilistic value of:
The pi-factor for each compartment. The p factor is solely dependent on
the geometry of the watertight arrangement of the ship.
The reduction factor r. The r factor may be used to modify the p factor
for a compartment if longitudinal subdivision is present.
The reduction factor v. The v factor is dependent on the geometry of the
watertight arrangement (decks) of the ship and the draught of the initial
loading condition. It represents the probability that the spaces above the
horizontal subdivision will not be flooded.
The factor of survivability
The si-factor for each compartment. The s factor is dependent on the
calculated stability of the ship after damage in a specific initial
condition.
Damage Stability Information
For the probabilistic approach, in addition to the usual deterministic
information, the following is to be provided:
subdivision length Ls
initial draughts and the corresponding GM-values
required subdivision index R
attained subdivision index A with a summary table for all contributions
for all damaged zones.
draught, trim, GM in damaged condition
damage extension and definition of damage cases with probabilistic
values p, v and r
righting lever curve (including GZmax and range) with factor of
survivability s
critical weathertight and unprotected openings with their angle of
immersion
details of sub-compartments with amount of in-flooded water/lost
buoyancy with their centres of gravity.
Preparation of Input to Probabilistic Damage
Stability Assessment
Define division geometry, i.e. what are considered division?
Compartment boundary table and Subdivision matrix
Critical points, Regulations 7-2-1 and 7-2-5
Weathertight critical points
Unprotected critical points
Subdivision length, Regulation 2
Required subdivision index, R
Initial draughts and the corresponding GM-values
Main input parameters for p(x1, x2) can be calculated
according to Regulation 7-1.1.1.
Central subdivision concepts
In regulation 7-1, the words compartment and group of
compartments should be understood to mean zone and
adjacent zones.
Zone is a longitudinal interval of the ship within the subdivision
length.
Room is a part of the ship, limited by bulkheads and decks,
having a specific permeability.
Space is a combination of rooms.
Compartment is an onboard space within watertight boundaries.
Damage is the three dimensional extent of the breach in the ship.
Subdivision, Watertight and Weathertight
Integrity
SOLAS 2009, Part B-2, Regulations 9 to 17, specifies requirements
to subdivision, watertight and weathertight integrity, including:
Double Bottom, Regulation 9
Watertight Bulkheads and Decks
Watertight Subdivision
Arrangement of Watertight Bulkheads
Collision bulkhead
Stern tube bulkhead and remaining watertight bulkhead
Openings in Watertight Bulkheads
External Openings
Cross-Flooding Arrangements
Compartment boundary table and
Subdivision matrix
Having considered the watertight subdivision of the ship,
that is longitudinal-, transverse- and vertical watertight
structure a useful tool to combine the damages to be
examined is a subdivision matrix and a compartment
boundaries table.
A subdivision matrix contains all zones and all transverse
barriers in each zone and all vertical barriers (decks) in
each zone
Central considerations regarding the
reductions factor r and b
Transverse barriers extent of penetration is linked to
its probability
This causes the following questions to arise:
What is this penetration value which just misses rupturing
the wing bulkhead?
Is it simply the distance from the side shell to the wing
bulkhead?
What if the side shell is not a nice flat wall?
What if the bulkhead is sloping in various directions, has
knuckles or is stepped?
What is a fair penetration value to use in those cases?
Transverse subdivision in a damage zone
In calculating r-values for a group of two or more adjacent
compartments, the b-value is common for all compartments in
that group, and equal to the smallest b-value in that group:
b = min(b1, b2, , bn)
where n is the number of wing compartments in that group and
is the mean values of b for individual wing compartments
contained in the group.

The simplistic definition of "b" is that it is the mean distance

between the shell and the longitudinal subdivision in question
measured at the deepest subdivision load line. The complete
definition of "b" however, allows wide latitude in how this
distance is measured. Regulation 7-1 says that the distance is
measured between the shell and a plane parallel to the
longitudinal bulkhead.
Critical points, Regulations 7-2-1 and 7-2-5
The SOLAS 2009 explanatory notes say "openings are the
most critical factor to preventing an inaccurate attained
index A".
Two types of external critical points: weathertight and
unprotected.
only affect the results if after damage they are submersed at
equilibrium.
Internal critical points all only affect the answer if after
damage they are submersed at equilibrium, similar to a
weathertight point.
Procedure for Probabilistic Damage Stability
Calculations
1. Model the vessel and define tanks and compartments
2. Define other ship data required for damage analysis
3. Defines the boundaries of the damage zones
4. Longitudinal bulkhead and deck locations are defined for
each zone and groups of adjacent zones
5. Once steps 2 and 3 have been completed, the p-factors are
calculated. With the bulkheads and decks defined in step 4,
the r- and v-factors are also calculated
6. When the zones have been defined, which tanks are damaged
in each zone and sub-zone are also defined
7. Perform a probabilistic damage analysis
Simplified Verbal Pseudo Algorithm for
Calculation Procedure
1. Initialize Attained Index, A, to zero.
2. Determine probability of damage, prv, as follows:
a. Determine size of integration steps in longitudinal,
transverse and vertical directions
For all longitudinal integration steps
For all transverse integration steps
For all vertical integration steps
b. Determine the volumetric element as imposed by the
present combination of integration steps
c. Determine prv, that is the probability of damage of this
volumetric element
..continued .
d. Determine the damage case (the combination of affected
compartments) implied by this volumetric element
e. Does this damage case already exist?
If No: Create this damage case, and set aggregated
probability of damage to prv
If Yes: Add prv to the aggregated probability of damage for
this damage case
3. Calculate probability of survival, s.
4. The product, prv s is added to the index, A = A + prvs.
5. Steps 3 through 5 are repeated for each division.
6. Steps 1 through 6 are repeated for additional load condition(s)
and the attained indices for each are averaged.
Output from Probabilistic Damage Stability
Calculations
Results for each damage case which contributes to the index A:
draught, trim, GM in damaged condition
damage extension and definition of damage cases with
probabilistic values p, r and v
righting lever curve (including GZmax and range) with factor of
survivability s
critical weathertight and unprotected openings with their angle of
immersion
details of sub-compartments with amount of in-flooded water/lost
buoyancy with their centres of gravity.
Calculation of p factor
The accumulated value of p for one zone or a group of adjacent zones is
determined by:

where
where
j: the damage zone number starting with No.1 at the stern;
n: the number of adjacent damage zones in question where j is the aft zone;
k: the number of a particular longitudinal bulkhead as a barrier for
transverse penetration in a damage zone counted from shell towards the
centreline. The shell has No. 0;
K: total number of transverse penetration boundaries;
pj,k,n: the p-factor for a damage in zone j and next (n-1) zones forward of j
damaged to the longitudinal bulkhead k.

A spreadsheet with VBA macros, SOLAS2009_JGCE.xlsm, will be

available on http://www.jg-consultant.com, Download section.
Calculation of p factor and r and b factor
Calculation of the s factor
The si-factor given in SOLAS 2009 Reg. B-1/7-2.

where
GZmax is not to be taken as more than 0.12 m;
Range is not to be taken as more than 16;
K=1 if e <= min
K=0 if e <= max

otherwise,

where
e is the equilibrium heel angle in any stage of flooding, in degrees;
min is 25 for cargo ships; and
max is 30 for cargo ships.
Damage Case Partial Attained Indices Table
for Calculation of attained index A
Summary and Conclusion
In this presentation the following has been subject for going
through:
What is damage stability,
Why Damage Stability Assessment,
How damage stability assessment
Deterministic damage stability
Probabilistic
Damage stability information and documentation
Preparation of input to probabilistic damage stability calculations
Presentation of a simplified procedure for probabilistic damage stability
calculation
A verbal pseudo algorithm for probabilistic damage stability calculation
An overview of the necessary output from probabilistic damage stability
calculations for assessment
Key differences between probabilistic and deterministic rules
derive from how the mathematical models are classified.
It difficult to compare, as they build on different modelling
framework.

Deterministic Approach: Probabilistic Approach:

The deterministic approach can be
classified in a mathematical
modelling language, as one in
which every set of variable states
is uniquely determined by
parameters in the model and by
sets of previous states of these
variables. Therefore, deterministic
models perform the same way for
a given set of initial conditions.
Conversely, the probabilistic
approach can be classified in a
probabilistic or stochastic
model, where randomness is
present, and variable states are
not described by unique values,
but rather by probability
distributions.
The core components in the probabilistic
methodology can be summarized as:
Required overall level of survivability accounting for any foreseeable situation
where the ship has lost some of its watertight integrity (index R).
Distributions describing the degree of survivability under a specific damage (s):
The basis is a common format based on three characteristics of the GZ-
curve at equilibrium after flooding (range, max, heel),
Is based on the weighted sum of survivability at three different loading
conditions (ds, dp, dl) that does not necessary follow the actual operational
profile.
Distributions describing damage position and extent (p, r, v):
Based on scattered collision statistics from a variety of ship types and
situations,
Generic distributions coupled to length and breadth normalized with respect
to the stroked ship size. Vertical extent described by constant distribution.
Improvement and future development
Harmonization of damage stability assessment for all types of
flooding causes and all types of ships, including a fully
comparable survivability criteria (s) and harmonization of
safety level requirements (R).
Could well be differentiated for different arrangements and
ship types as needed in order to better describe capability as
long as the survivability measure is kept on a comparable level.
A more optimum regulation text would be desirable. The
standard can be refined and expressed more clear in the
regulations in order to avoid the perception of a black-box
methodology.
THANK YOU
Consideration of Damage to ships from
Conceptual Design to Operation:
The Implications of Recent and Potential
Future Regulations regarding Application,
Impact and Education

KW Hutchinson, Babcock International Group, UK

AL Scott, Marine and Coastguard Agency, UK
PNH Wright, MD Woodward and J Downes, Newcastle University, UK

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 1
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Presentation

Development of Statutory Rules and Regulations

Current International (IMO) Damage Stability Regulations
Proposed Updates to IMO Stability Regulations (SLF 53)
Issues with RO-RO Passenger Ships
Issues with Tankers
Onboard Stability Information
Understanding and Usage of Onboard Stability Information
Summary

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 2
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Paper

Current International (IMO) Intact Stability Regulations

Design of Ships to SOLAS 2009
Review of Onboard Stability Information
Regulatory Developments regarding Onboard Stability Information
Aim of Undergraduate Teaching
Teaching of Naval Architecture at Newcastle University
Recent enhancements at Newcastle University
Facilitating Curriculum Development

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 3
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Development of Statutory
Rules and Regulations

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 4
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Outline structure International Maritime Organisation

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 5
The Royal Institution of Naval Architects, London, 26th-27th January 2011
International Maritime Organisation and Stability

Special Agency of UN, established 1948, first met in 1959

169 Member States, 3 Associate Members
MSC responsible for safety issues, meets for 8 days twice a year
Technical matters devolved to Sub-Committees, stability SLF
SLF meets annually or 1 week discussing papers and issues
Plenary Session and Specialized Working / Drafting Groups
Draft Amendments or Proposals for further work, research etc.
SLF reports considered by MSC, proposals new legislation
IMO Regulations implemented by Flag States into their law
Consideration of Damage to Ships from Conceptual Design to Operation
Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 6
The Royal Institution of Naval Architects, London, 26th-27th January 2011
 Current International (IMO)
Damage Stability Regulations

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 7
The Royal Institution of Naval Architects, London, 26th-27th January 2011
IMO Damage Stability Regulations

Safety of Life at Sea (SOLAS 2009)

Annex 1 to MARPOL 73/78
International Bulk Chemical Code (IBC 2007)
International Gas Carrier Code (IGC 1993)
Offshore Supply Vessel Code (OSV 2006)
Special Purpose Ship Code (SPS 2008)
International Convention on Load Lines (ICLL 1966 / 1988)
Mobile Offshore Drilling Unit Code (MODU 2001)
Stockholm Agreement
Consideration of Damage to Ships from Conceptual Design to Operation
Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 8
The Royal Institution of Naval Architects, London, 26th-27th January 2011
 Proposed Updates to
IMO Stability Regulations (SLF 53)

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 9
The Royal Institution of Naval Architects, London, 26th-27th January 2011
SLF 53, 10th to 14th January 2011 (1)

Development of new generation intact stability criteria

Guidelines to enhance the safety of small fishing vessels
Safety provisions applicable to tenders operating from pass. ships
Options to facilitate entry into force of 1993 Torremolinos Protocol
Consideration of IACS unified interpretations

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 10
The Royal Institution of Naval Architects, London, 26th-27th January 2011
SLF 53, 10th to 14th January 2011 (2)

Standards on time-dependent survivability of damaged pass. ships

Stability and seakeeping of damaged pass. ships returning to port
Guidelines for verification of damage stability for tankers and bulkers
Review of damage stability regulations for RO-RO passenger ships
Amendments to SOLAS subdivision standards for cargo ships
Revision of SOLAS subdivision and damage stability regulations

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 11
The Royal Institution of Naval Architects, London, 26th-27th January 2011
 Issues with
RO-RO Passenger Ships

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 12
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Consideration of Damage to Ships from Conceptual Design to Operation
Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 13
The Royal Institution of Naval Architects, London, 26th-27th January 2011
RO-RO Passenger Ship Damage Stability

The original brief for SOLAS 2009 was to produce new harmonized
probabilistic regulations having an equivalent level of safety to
deterministic SOLAS90 regulations
R and s set at same level for conventional Passenger Ships
SOLAS90 was deterministic and included floodable length
Following Herald of Free Enterprise and Estonia losses the
Stockholm Agreement was added to SOLAS90 for RO-PAX
operating in (NWE) Europe but not the rest of the world
Research appears to show SOLAS 2009 does not explicitly allow for
WOD and hence is SOLAS90 without Stockholm Agreement

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 14
The Royal Institution of Naval Architects, London, 26th-27th January 2011
RO-RO Passenger Ship Damage Stability

Many non-EU IMO Member States have taken this to mean that
WOD is NOT to be allowed for in SOLAS 2009 as it is only an EU
regional matter necessitated by the particular sea conditions and
traffic density specific to EU waters.
Such Member States have never had anything equivalent to the
Stockholm Agreement in their own national regulations
Many EU Member States have argued that the current s factor either
takes sufficient account of WOD or that allowing for WOD has little
or no effect on A anyway and can therefore be ignored
Damage statistics used for SOLAS 2009 showed collisions
predominantly occur in calm waters

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 15
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Research on SOLAS 2009 RO-PAX designs

May be a problem with s factor RO-PAX ships

SOLAS 2009 RO-PAX ships with LLH can sink, rather than capsize,
in calm conditions with only B/10 side penetration into the LLH
Problem appears to be lack reserve of buoyancy (floodable length)
Possible revision to R and s may be necessary
SOLAS 2009s lack margin line criterion more significant to RO-PAX
than conventional passenger ships applicability same formulae
Possible requirement for distinction between LLH and non-LLH type
RO-PAX if critical loss mechanisms different eg sinkage or capsize
Research ongoing, EU RO-PAX still require Stockholm Agreement

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 16
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Outstanding issues to be resolved within IMO

Urgent consideration must be given to the possibility that SOLAS90

Ch II-1 Regulations 4 7 (subdivision and floodable length) have
been overlooked during the harmonization process and that
therefore the safety equivalence criterion between SOLAS90 and
SOLAS 2009 has not been completely fulfilled
These issues need to be discussed and resolved within IMO before
any further work is carried out on adjusting the s or changing R etc

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 17
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Issues with
Tankers

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 18
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Consideration of Damage to Ships from Conceptual Design to Operation
Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 19
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Tanker Damage Stability

Complex due to drop out - fluid deadweight loss, not bare boat
Variability in fill levels, loading patterns, SG, damage scenarios
Applicability of / deviation from Loading Conditions in SIB
Belief compliance with intact CKG all that is required (PSI)
1/3 tankers unacceptable risk to life and environment (IMO)
urgent need develop guidelines verification damage stability (IMO)
Lack of awareness complexity of damage stability calculations
Construction / Presentation of usable Operational Information
SIB simplified stability information / Loading Instrument (non SIB)
Consideration of Damage to Ships from Conceptual Design to Operation
Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 20
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Typical Intact and Damage Critical Curves

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 21
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Typical Intact and Damage Critical Curves

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 22
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Typical Intact and Damage Critical Curves

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 23
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Simplified Tanker Stability Information

Intact Critical KG Curves - why?

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 24
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Typical Intact and Damage Critical Curves

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 25
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Simplified Tanker Stability Information

Intact Critical KG Curves - why?

Condition Specific Combined (intact & damage) Critical KG Curves
- production fairly simple, marginally expands operational scenarios
- few figures, many curves on each figure, not envelope
- confined to specific fill levels, loading patterns, SG
Generic Combined (intact and damage) Critical KG Curves
- production complex, potentially expands operational scenarios
- multiple figures, potentially many exceptions, not simple envelope
- confined to SG only
Consideration of Damage to Ships from Conceptual Design to Operation
Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 26
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Typical generic Combined Critical KG Curves

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 27
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Typical generic Combined Critical KG Curves

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 28
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Typical generic Combined Critical KG Curves

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 29
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Non SIB Tanker Stability Information

Loading / Stability Instruments covered in IS 2008

- Type 1: Intact Stability only
- Type 2: Intact and Damaged Stability using CCKGs
- Type 3: Intact and Damaged Stability by direct calculations
Onboard with direct calculation capability or Onshore facilities
Advice on corrective measures to users
- Re-adjustment of loading condition, Trial and Error
Advanced planning akin to that undertaken for Container Ships
Checking of Departure, Arrival and Interim Loading Conditions
Consideration of Damage to Ships from Conceptual Design to Operation
Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 30
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Onboard Information

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 31
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Information currently provided Onboard

SIB, Capacity Plan, Calibrations, Damage Control, Freeboard Plan

Approachable clear, consistent, concise, usable, comprehensive
- simplified stability information - X-Curves, Wind data?
- Damage Control Plan / Book using the Probabilistic Approach
- Damage Consequence Diagrams (T, t)

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 32
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Damage Consequence Diagram

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 33
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Information currently provided Onboard

SIB, Capacity Plan, Calibrations, Damage Control, Freeboard Plan

Approachable clear, consistent, concise, usable, comprehensive
- simplified stability information - X-Curves, Wind data?
- Damage Control Plan / Book using the Probabilistic Approach
- Damage Consequence Diagrams (T, t)
Presentation of CCKG/GM data under consideration at IMO
- must be easy to use but reliable
- cover operational range of vessel (T, t, fills, loading, SG)

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 34
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Typical Intact and Damage Critical Envelope Curves

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 35
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Typical simple Combined Critical KG Curve

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 36
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Information currently provided Onboard

SIB, Capacity Plan, Calibrations, Damage Control, Freeboard Plan

Approachable clear, consistent, concise, usable, comprehensive
- simplified stability information - X-Curves, Wind data?
- Damage Control Plan / Book using the Probabilistic Approach
- Damage Consequence Diagrams (T, t)
Presentation of CCKG/GM data under consideration at IMO
- must be easy to use but reliable
- cover operational range of vessel (T, t, fills, loading, SG)
SLF addressing this for SOLAS Updates
Consideration of Damage to Ships from Conceptual Design to Operation
Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 37
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Understanding and Usage
of Onboard Information

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 38
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Understanding and Usage of Onboard Information

Approved SIBs etc. seldom utilised, substantial volume and complex

During loading possible intact, not to mention damage, not checked
Surveyors observations of crew during PSIs:
- Little awareness of dangers of inadequate damage stability
- Damage stability often thought complied with if intact acceptable
- Reference to CCKG curves unnecessary, ship completely fail-safe
- Lack of awareness of Cross Flooding Arrangements
- Significance of closure of Watertight Doors when not in use
Need to enhance seafarers appreciation of safe loading / operation
Consideration of Damage to Ships from Conceptual Design to Operation
Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 39
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Summary

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 40
The Royal Institution of Naval Architects, London, 26th-27th January 2011
Summary

Recent history, status and potential development of international

stability regulations
Application of statutory rules and regulations
Implications on the design and operation of ships
Highlighted some complexities associated with implementation
Apparent that continual development of regulations poses a
challenge for naval architects and seafarers alike in appreciating
their impact on the design and operation of vessels
Providing up to date education to capture developments is
imperative if implications are to be fully appreciated

Consideration of Damage to Ships from Conceptual Design to Operation

Hutchinson, Scott, Wright, Woodward and Downes, The Damaged Ship 41
The Royal Institution of Naval Architects, London, 26th-27th January 2011
SALVAGE AND MARINE OPERATIONS
PROJECT TEAM

IMPROVEMENT OF EMERGENCY
RESPONSE FOR SUPPORT TO UK
MILITARY MARITIME INCIDENT
Click to edit Master subtitle style

Captain Stephen Quinn OBE MNI MIIMS

Nigel Hills CEng MRINA RCNC

26th January 2011 RINA Damaged Ship Conference 1

 Content

Salvage and Marine Operations PT

Emergency Response Mobilisation
Collaboration
Emergency Response Preparedness
Use of Tools
Future

26th January 2011 RINA Damaged Ship Conference 2

 S&MO PT Responsibilities
UK MoD Authority in:
Salvage
Towing
Heavy Lift
Specialist Moorings and Boom Defence
Chartering of Specialist Ships
S&MO provides SME input to new platform specifications

26th January 2011 RINA Damaged Ship Conference 3

 Endorsed Salvage Policy
Recovery of floating but disabled units. (Towing, Heavy Lift)
First aid to RN/RFA/STUFT collision/grounding.
Recovery of ditched mod aircraft.
Repair/REFLOATING stranded mod vessels/sheltered waters.
Port clearance
Assistance to a bottomed, disabled S/M.
Recovery of a bottomed, disabled evacuated S/M.
Provision of expertise and advice to DTER/OGD.
Recovery of objects, including nuclear weapons, from the
seabed.
Developing policy on MoD IWMAR.

26th January 2011 RINA Damaged Ship Conference 4

 Salvage and Marine Response Matrix

Volume of Response capability defined by

Defence Planning Assumptions
Resource philosophy is ownership of key
personnel available at R0 +6 hrs to deploy
Absolute minimum quantity of assets owned
Fully integrated with Industry codes of
practice and usage of Industry resources

26th January 2011 RINA Damaged Ship Conference 5

Response to a Military Maritime Incident
Incident Stages
Initiating Event Seconds
Minutes
Notification
First Aid Hours

Stabilisation Days

Recovery Weeks
Repatriation
All needs clear Command and Control
26th January 2011 RINA Damaged Ship Conference 6
26th January 2011 RINA Damaged Ship Conference 7
S&MO Response to a Military Maritime Incident
Initial Notification received
Call out of personnel
Establish HQ & Collate Info
Ascertain availability and engage Industry resources
Analysis of incident
Mobilisation of expertise
Bring vessel to a safe condition
Place vessel where operational status can be regained

26th January 2011 RINA Damaged Ship Conference 8

 Collaboration with OGDs and Other Nations
Combined
TOWEX

Same ETV used

for ASTUTE
Grounding

26th January 2011 RINA Damaged Ship Conference 9

 Liaison with Other Nations

Towing Exercise
Royal Netherlands navy

Heavy Lift of Floating Dry Dock

US Navy
VICTOR SSN Transport
Canada DoD

26th January 2011 RINA Damaged Ship Conference 10

 Emergency Response Preparedness
Requirement to have Emergency Response Plans
Requirement to exercise Emergency Response
Plans
S&MO conduct Table Top Exercises for DE&S
Duty Holders as required
S&MO act as DISTAFF on Exercise PURPLE
QUIVERs (SS) and Exercise RAMULARs (SM)

26th January 2011 RINA Damaged Ship Conference 11

 Emergency Response Exercise
Exercise PURPLE QUIVER
Annual FLEET response exercise for RFA
Damage Scenario
Practice deployment
Rapid Analysis Exercise
Media Awareness Exercise
Impact just
LUST starts to fwd 119 Blkhd
open distance

Vessel concluding RAS

Ships disengage
FTVR disabled
Attack LUST pushed to Pt
FTVR veers
by collision
hard to port

26th January 2011 RINA Damaged Ship Conference 12

 No 5 Deck
5 Deck Plating, Side Shell Plating
severely buckled and split, welds ruptured.
119 Bulkhead buckled with split welds.
5H Deck Passage Way destroyed.
5H POs Bathroom moderate damage.
5H Ladder Chain severely damaged.
5H Drinks Store moderate damage.
5J Telephone Booths and EDR, moderate damage.

26th January 2011 RINA Damaged Ship Conference 13

 Training
Ships Companys tested
extensively before
deployment by 3rd party

They are essential 1st Line

of Defence

Only inherent resource at

outset of incident

26th January 2011 RINA Damaged Ship Conference 14

 DE&S Support to FLEET
Training and Instruction to Ships Companys

Provision of Stability Information

Damaged Stability Aids

Mechanisms for Flow of Information (NAIIF)

26th January 2011 RINA Damaged Ship Conference 15

 Collation of Information

Crew may have other

things on their mind than The Nav Arch would like to know our
draft marks when we have a moment
responding to queries
from shore side

26th January 2011 RINA Damaged Ship Conference 16

 Salvage Plan

The aims of the salvor and the responding authorities

may be at variance to the aims of Command
26th January 2011 RINA Damaged Ship Conference 17
A full operation can last for an
extended period of time

There is a fundamental need to

ensure ALCON remain informed of
the mission aim and what can be
expected from each organisation

Experience shows that this is

particularly important for the ships
company

26th January 2011 RINA Damaged Ship Conference 18

 S&MO Activity
Dissemination of best practice across diverse areas of
MOD
Liaison between Industry and MOD on Salvage
improvements
Frequency of involvement enhanced corporate
awareness
Lectures to RN CO/XO/MEOs and Desigs
Lectures to STANAC
Emergency Response Seminars

26th January 2011 RINA Damaged Ship Conference 19

 Software Tools for Salvage
Numerous software tools available to the Naval Architect
PARAMARINE
NAPA
TRIBON
HECSALV/POSSE
GHS
etc
Each have benefits, each have different areas of strength
Also, some have weaknesses depending on what they are
being used for.

26th January 2011 RINA Damaged Ship Conference 20

 Programme of Ship Salvage Engineering

S&MO historically used HECSALV

S&MO elected to specialise in POSSE
Currently meets our requirements
Mixture of Regular and Semi-Regular Users
Developments in concert with US, shared funding
Aids our aim of joint operability with US
Rationalised training burden
Meets our analysis aims in one package

26th January 2011 RINA Damaged Ship Conference 21

 POSSE enables us to use one code for all
stages of the operation
Eg: NOTTINGHAM 2002, ENDURANCE 2008/09
Incident
Salvage/Stabilisation
Towing Heavy Lift and
Dry Docking
One Ship Definition, One Code, ease of management of
supporting analysis throughout a long duration incident
26th January 2011 RINA Damaged Ship Conference 22
 Salvage Operation

26th January 2011 RINA Damaged Ship Conference 23

26th January 2011 RINA Damaged Ship Conference 24
 Towing Operation

26th January 2011 RINA Damaged Ship Conference 25

26th January 2011 RINA Damaged Ship Conference 26
 Heavy Lift Operation

26th January 2011 RINA Damaged Ship Conference 27

 Heavy Lift / Dry Docking

26th January 2011 RINA Damaged Ship Conference 28

26th January 2011 RINA Damaged Ship Conference 29
Initial Design Of Cribbing and Blocking is
responsibility of Transport Contractor

POSSE enables us to be independent Marine

Warranty Surveyor if required.

POSSE enables us to run the what ifs

POSSE enables us to retrospectively analyse

the docking case
26th January 2011 RINA Damaged Ship Conference 30
26th January 2011 RINA Damaged Ship Conference 31
 Blocks in RED are ineffective
Re-analyse to check residual loads

26th January 2011 RINA Damaged Ship Conference 32

 Analytical Tool Developments

Stability Experienced
Structures Tools available, experience building
Flooded Compartments R&D well established
Effectiveness of Structural Reinforcement, tool
available/required?

26th January 2011 RINA Damaged Ship Conference 33

NOTTINGHAM FER Space recovered by pumping prior
to moving vessel.
Can we assess the impact of this entrained water on overall
ship structure and dynamic stability?

26th January 2011 RINA Damaged Ship Conference 34

NOTTINGHAM 3 Deck Shoring
Built as large as possible in situ at
salvage location.
How can it be analysed rapidly
and effectively?
Is it efficient, is it required?

26th January 2011 RINA Damaged Ship Conference 35

 Future
Development of Analytical Tools
Ease of Use

Dissemination of Best Practice

Liaison with Foreign Navies (DCINC)
Liaison with Industry
Learn from other Industries

Practice
Realistic Training
26th January 2011 RINA Damaged Ship Conference 36
 Questions ?

26th January 2011 RINA Damaged Ship Conference 37

WHILE THERE IS PLENTY OF DATA TO PERFORM A COMPLETE STUDY OF STABILITY AFTER DAMAGE IN
MERCHANT VESSELS, UNFORTUNATELY THAT IS NOT THE CASE FOR WARSHIPS. THERE ARE DIFFERENT
REGULATIONS OF DAMAGE STABILITY THAT CAN BE COMPARED, SUCH AS THE DESIGN DATA SHEET, THE
NAVAL SHIP CODE, AND WITHIN THE SOLAS, THE STOCKHOLM AGREEMENT, WHERE WATER ON DECK IS
INCLUDED.

WHILE REGULATIONS RULING STABILITY AFTER DAMAGE IN WARSHIPS ARE MORE DEMANDING (IN
GENERAL) THAN MERCHANT RULES IN TERMS OF WATER ON DECK, THE STOCKHOLM AGREEMENT
PROVIDES GREATER RELIABILITY OF CALCULATED RESULTS IN CIVILIAN VESSELS, COMPARING TO THOSE
USED BY WARSHIPS.
THIS PAPER PROVIDES A STUDY OF THE STABILITY CRITERION OF THE U.S. AND BRITISH NAVY, A STUDY OF
THE NEW NSC CRITERION (CLASSIFICATION SOCIETY-DRIVEN APPROACH) AND AN ANALYSIS OF SOLAS
FROM ITS ORIGIN UNTIL THE NEW PROBABILISTIC APPROACH GOING THROUGH ALL THE AMENDMENTS
THAT RELATE TO STABILITY.

THE PAPER CONCLUDES WITH THE INTRODUCTION A SERIES OF COMPARISONS BETWEEN THE CRITERIA
USED BY THE NAVIES TO MAINTAIN THE INTEGRITY OF ITS BASIC APPROACH, AND INCREASE THE
SIMILARITIES WITH THE CRITERION OF IMO, SUCH AS THE CALCULATION OF WATER ON DECK IN THE
STOCKHOLM AGREEMENT. THESE COMPARISONS LEAD TO INTERESTING CONCLUSIONS REGARDING HOW
CURRENT CRITERIA USED BY THE NAVIES COULD BE ENORMOUSLY IMPROVED WITH JUST A FEW MINOR
CHANGES.
 Displacement (Tons).
Angle of heel ().
1 Limit angle of the damage righting arm curve ().
c Angle of steady heel ().
r Rolling angle ().
A Attained subdivision index (dimensionless).
A1 Area under righting arm curve between the roll back angle and the equilibrium heel angle. Reserve of dynamic
stability (m2).
A2 Area under righting arm curve between the angle of equilibrium and the extreme intersection between righting
arm (m2).
B Beam at the widest point (m).
CB Block coefficient (dimensionless).
CM Midship coefficient (dimensionless).
D Depth (m).
DN Dynamic stability (mmrd).
GM Permissible metacentric height (m).
GZ Righting arm (m).
KG Permissible height of the centre of gravity (m).
Lpp Length between perpendiculars (m).
R Required subdivision index (dimensionless).
RSK Is the sure kill radius (m).
RSS Is the sure save radius (m).
T Middle draft or middle draught (m).
Trim Difference between the forward and after drafts (m).
TO AVOID DUPLICATION, GAPS AND SHORTCOMINGS IN SAFETY, IT IS IMPORTANT FOR THE NAVIES TO
WORK TOGETHER WITH THE CLASSIFICATION SOCIETIES IN THE DEVELOPMENT OF EFFECTIVE AND
SUSTAINABLE ARRANGEMENTS. THUS, DEVELOPMENT OF RULES FOR WARSHIPS NAVAL SHIPS RULES BY
VARIOUS CLASSIFICATION SOCIETIES IS THE MOST IMPORTANT CONTRIBUTION TO WORK IN THIS AREA.
THE IDEA OF COOPERATION TO MAKE AN INTERNATIONAL CONVENTION FOR THE SAFETY OF LIFE AT SEA
(SOLAS) GOES BACK TO THE NINETIES.
IN SEPTEMBER 1998, CLASSIFICATION SOCIETIES OF THE MEMBER STATES OF THE NORTH ATLANTIC
TREATY ORGANIZATION (NATO) MET TO ESTABLISH LINKS WITH THEIR OWN NATO. THIS MEETING
ESTABLISHED THE NAVAL SHIP CLASSIFICATION ASSOCIATION (NSCA), IN MAY 2002, AND THE COOPERATION
WAS DEFINED ACCORDING TO THE FOLLOWING TERMS OF REFERENCE: PROMOTE SAFETY STANDARDS AT
SEA, PROMOTE MEASURES TO PROTECT THE MARINE ENVIRONMENT, PROMOTE AND DEVELOP COMMON
OPERATING STANDARDS, UNDERTAKE R&D TO SUPPORT THE ABOVE AND COMMUNICATE THE VIEWS OF
THE PARTNERSHIP AGREEMENTS AND THE NSCA.
THE PHILOSOPHY OF THE SOLAS IS APPLICABLE TO MERCHANT SHIPS, AND IS NOT FULLY TRANSFERABLE
TO A WARSHIP. A WARSHIP HAS REQUIREMENTS FOR ACOUSTIC SIGNATURE, ELECTROMAGNETIC
SIGNATURE, SIGNATURE RADAR, ELECTRONIC WARFARE, ANTISUBMARINE WARFARE AND IT
DEMONSTRATES THAT A SHIP IS NOT CIVIL.
 Classic FORAN Modules
New Kernel Modules

FSURF
POWER LOAD

FSUBD
HYDROS FLOOD

VOLUME LAUNCH
ORACLE
DATA BASE
 Next Page

FROM FIRST OF FEBRUARY OF 1992 THE PROBABILISTIC METHOD WAS INSERTED INTO SOLAS AS PART B-1
OF THE CHAPTER II-1, ANNEX REGULATION FOR SUBDIVISION AND DAMAGE STABILITY OF CARGO SHIP
OVER ONE HUNDRED METERS IN LENGTH THATS APPLIES TO DRY CARGO SHIPS CONSTRUCTED ON OR
AFTER THE FIRST OF FEBRUARY OF 1992. LATER ON SHIPS WITH LENGTH BETWEEN EIGHTY AND ONE
HUNDRED METERS WERE ALSO INCLUDED.

THE EIGHT ASSEMBLY OF INTERNATIONAL MARITIME ORGANIZATION (IMO), BY RESOLUTION A265 (VIII)
ADOPTED A SET OF PROBABILISTIC REGULATIONS OF SUBDIVISIONS AND DAMAGE STABILITY PASSENGER
SHIPS AS A EQUIVALENT TO AND A TOTAL ALTERNATIVE TO THE REQUIREMENTS OF PART B OF CHAPTER II
OF THE SOLAS, 1960 FOR PASSENGERS.
Previous Page

FINALLY IN EIGHTY SESSION OF IMO, MARITIME SAFETY COMMITTEE (MSC) THE WORKING GROUP HAS
FINALIZED A SUBSTANTIAL REVISION OF SOLAS CHII PT 1 A, B AND B1 AIMING AT HARMONIZED DAMAGE
STABILITY REQUIREMENTS FOR ALL THE TYPE OF SHIPS EXCEPT FOR TANKERS PERFORMED BY MEANS OF
A COMMON PROBABILISTIC METHODS. THE DRAFT HAS BEEN ADOPTED AT THE MSC 80 WITHOUT FURTHER
MODIFICATION. THE REVISED CH II-1 WILL APPLY TO ALL NEW PASSENGERS VESSEL, ROLL ON-ROLL OFF
AND CARGO SHIPS BUILT ON OR AFTER FIRST OF JANUARY OF 2009.
PROBABILISTIC CONCEPTS ADDRESS THE PROBABILITY OF DAMAGE OCCURRING AT ANY LOCATION
THROUGHOUT A SHIP AND ADOPT A MORE RATIONAL CRITERION OF SUBDIVISION BY CONSIDERING THE
LIKELIHOOD OF DAMAGE RESULTING IN THE FLOODING OF ONLY ONE COMPARTMENT, OR ANY NUMBER
OF ADJACENT COMPARTMENTS, EITHER LONGITUDINALLY, TRANSVERSELY OR VERTICALLY. THE
RESIDUAL BUOYANCY AND STABILITY OF A SHIP IS CALCULATED FOR EACH OF SUCH DAMAGE CASES, AND
EITHER A POSITIVE OR A ZERO CONTRIBUTION IS ASSOCIATED TO EACH CASE, DEPENDING ON, WHETHER
OR NOT, THE RESIDUAL BUOYANCY AND STABILITY ARE CONSIDERED SUFFICIENT.
Next Page
Previous Page
EMIGRATION FROM EUROPE TO AMERICA.

IN RESPONSE TO THE SINKING OF THE RMS TITANIC APRIL OF 1912.

LONDON 1914.

TO KEEP THE CONVENTION UP TO DATE BY PERIODIC AMENDMENTS.

INTERNATIONAL CONVENTION FOR THE SAFETY OF LIFE AT SEA (SOLAS).

YEAR 1953, MV PRINCESS VICTORIA, KILLING 132 PEOPLE.

YEAR 1966, SS HERAKLION, WITH OVER 200 DEATHS.

YEAR 1968, WAHINE, KILLING 52 PEOPLE.

YEAR 1987, MS HERALD OF FREE ENTERPRISE, KILLING 193 PEOPLE.

YEAR 1993, MS JAN HEWELIUSZ, KILLING 55 PEOPLE

YEAR 1994, MS ESTONIA, KILLING 852 PEOPLE.

YEAR 2000, MS EXPRESS SAMINA, KILLING 80 PEOPLE.

YEAR 2006, MS AL-SALAM BOCCACCIO 98, KILLING ALMOST 1000 PEOPLE.

YEAR 2006, MV QUEEN OF THE NORTH, 2 PASSENGERS WERE LOST.

THE STOCKHOLM AGREEMENT REPRESENTS A MAJOR MILESTONE IN THE HISTORY OF RULE
DEVELOPMENT FOR ASSESSING SHIP DAMAGE STABILITY BY TAKING EXPLICITLY INTO ACCOUNT THE
EFFECT OF WATER ON THE RO-RO DECK, BY LINKING DAMAGE SURVIVABILITY EXPLICITLY TO
OPERATIONAL SEA STATES AND BY PAVING THE WAY TO PERFORMANCE-BASED STANDARDS AND TO FIRST-
PRINCIPLES APPROACHES TO SHIP SAFETY.
 FSURF HYDROS
Internal parts

Hydrostatics
Hull form, Decks & Bulkheads

VOLUME POWER

Spaces and Capacities Power Prediction

LAUNCH LOAD FLOOD FSUBD

Launching Loading conditions Damage Stability Subdivision

Intact Stability Probabilistic
Damage Stability
Next Page
Previous Page
THERE ARE MANY AREAS WHERE MILITARY VESSELS COULD IMPROVE SAFETY STANDARDS, ALTHOUGH
NOT NECESSARILY TO BE REGARDED AS LESS SECURE THAN THE CIVIL VESSELS. IT IS UP TO EVERY
GOVERNMENT AND AUTHORITY THE ESTABLISHMENT OF NAVAL SECURITY LEVEL TO OFFER THEIR
EQUIPMENT AND HOW IT IS ACHIEVED.
A WARSHIP WILL BE MORE LIKELY TO SUFFER DAMAGE BY EXPLOSION THAN A CIVILIAN SHIP, THEREFORE
A CIVIL ROLL ON-ROLL OFF WILL HAVE MORE RISK BY COLLISION THAN BY EXPLOSION. INSTEAD A
LANDING SHIP SHALL BE BY CHANCE MORE RISK HAVING GROUNDING OR RAKING IN ITS BOTTOM. CLEAR
EXAMPLES ARE FRIGATES, CORVETTES AND PATROL VESSELS, WHERE PROBABILISTICALLY, WOULDN'T
WATER BOARD PROBLEMS ON DECK FROM THE WATERLINE. THE GROUNDING IS A VERY DANGEROUS
FLOOD FOR ANY TYPE OF VESSEL, BUT DEPENDING ON ITS SIZE DAMAGE AND FORMS CAN BE CONSIDERED
FROM SERIOUS TO VERY SERIOUS.
THIS TECHNICAL PAPER ENDS GIVING A METHOD THAT HELPS AND SUPPORTS THE NAVAL ARCHITECT IN
THE ANALYSIS OF DAMAGE STABILITY. IN THIS WAY, THE NAVAL ENGINEER IS ABLE TO DECIDE WHAT OF
THE EXISTING CRITERIA BETTER FITS THE REQUIREMENTS OF THE SHIPS, AS A FUNCTION OF A FEW
PRINCIPLES.

FOR EXAMPLE, DEPENDING OF THE SHIP TYPOLOGY IS POSSIBLE TO KNOW WHICH POSSIBLE DAMAGE TYPE
IS MORE PROBABLE AND WHICH CRITERIA IS MORE RESTRICTED FOR EACH ONE OF THEM.
THE APPLICATION OF HIGH
PRESSURE WATER MIST AS
PART OF A HOLISTIC FIRE
FIGHTING SYSTEM
Simon Ratcliffe
Introduction
My placement the Naval Design Partnership
Background and drivers
Introduction to HPWM applications
Design
Sensors and Control
Summary
Questions
Naval Design Partnering
Need to sustain the UKs ability to design naval vessels from concept
to the point of build.
Important that a single UK shipbuilding (submarine building) entity
does not lead to single source of ideas;
The diversity of all organisations who contributed should be
recognised.
A partnering solution was identified where MOD would retain
ownership and oversight.
Efficiencies gained by working with a steady state collocated Team
Dont reinvent the wheel every time

The Naval Design Partnering Team is a

MOD-led multi-Industry 'rainbow' team within
D Ships delivering innovative concept design
for Naval vessels
Background
The cost of a ship relates Increasing
automation
directly to its size.
Smaller ships crews would Reducing
manpower
allow for smaller ships
Fire fighting is a key activity
that prevent a ships crew Reducing
Decreasing hull size
falling below a certain level risk

Increasing automation to Reducing

propulsion
remove the man with an requirements

extinguisher will make fire Increasing Reducing

safety emissions
fighting safer and allow for
smaller vessels in future. Reducing Through Life Costs
Background

This study aims explore the use of High Pressure Water

Mist (HPWM) to provide ship wide fire fighting cover.

Application to a generic area of the T26 Frigate to

provide fire fighting in peace time and battle scenarios

This could then be scaled to fit any vessel in the future

HPWM
High pressure water is dispersed by fixed nozzles which create an
ultra fine mist over the protected area. The mist fights fires in three
main ways;

Cooling
HPWM
High pressure water is dispersed by fixed nozzles which create an
ultra fine mist over the protected area. The mist fights fires in three
main ways;

Cooling
Smothering
HPWM
High pressure water is dispersed by fixed nozzles which create an
ultra fine mist over the protected area. The mist fights fires in three
main ways;

Cooling
Smothering
Attenuation
HPWM
High pressure water is dispersed by fixed nozzles which create an
ultra fine mist over the protected area. The mist fights fires in three
main ways;

Advantages over traditional low

Cooling pressure systems are
Smothering
Attenuation
Uses very little water
Causes minimal secondary
damage
Suitable for HV and machinery
spaces
HP reduces pumping losses
Flashover Suppression & Pre-emptive
Cooling
Flashover suppression aims to keep air/gas
temperatures in compartments too low for materials and
fuel sources to ignite.
Pre-emptive cooling refers to reducing the risk of fire
spreading into a compartment by cooling before a fire
takes hold. This technique can be used to safeguard
critical spaces, such as machinery spaces of along key
escape routes.
 Blast Mitigation
Water Mist has been proven to reduce the effects of
explosions in confined spaces.
US studies have shown that the greater the mist
concentration, the greater the reduction in blast pressure
up to 45% reduction.
Could be used as part of a ships defensive suite.
Shock wave breaks mist particles down further,
increasing the surface area and vaporisation rate
cooling the gas front.
Design
Subject Area Type 26 Frigate
Subject Area Type 26 Frigate
Sprinkler Layout

Small
Compartments

Large
Compartments

Corridors and
Stairways
Sprinkler Layout

Small
Compartments

Large
Compartments

Corridors and
Stairways
Distribution Architecture Options

Different High Pressure Water Main distribution

architectures investigated;
Centre Main
Dual Main
Sectional Loop Architecture

In particular, the way in which they might contribute to

enabling a holistic and survivable water mist system.
Sectional Loop Architecture
Sectional Loop Architecture
Water Flow Demand

Based on two scenarios; Peace Time and Battle Damage

Assuming 4 nozzles per branch group

Peace Time
Causes likely to be localised equipment failure, cooking and galley operations, human
error and the failure to follow proper process
fire will start small and begin to spread from a single source.
Least demanding scenario of HPWM system.
Water Flow Demand

Based on two scenarios; Peace Time and Battle Damage

Assuming 4 nozzles per branch group

Battle Damage
More challenging fire conditions
Will involve successively greater number of operating branch groups.
Water Flow Demand
Water Flow Demand

Large compartment
suffers damage
Water Flow Demand

Large compartment
suffers damage

Fires potentially
starting in adjacent
spaces and above
the damaged area.
Water Flow Demand

Large compartment
suffers damage

Fires potentially
starting in adjacent
spaces and above
the damaged area.

As fire spreads,
further branch
groups can be
brought online.
Shown here in
Orange, then
Green.
Pump Type
Required to be high pressure and meet a variable demand.
Traditional sprinkler systems rely on rotodynamic centrifugal pumps.
Other option is to use Positive Displacement pumps.
Configuration trialled on ex-USS Shadwell
Pump Type
Required to be high pressure and meet a variable demand.
Traditional sprinkler systems rely on rotodynamic centrifugal pumps.
Other option is to use Positive Displacement pumps.
Configuration trialled on ex-USS Shadwell
Can be connected in parallel

M M
98 L/min
Bypass

M 98 L/min M 98 L/min
Sensors and
Control
 Sensors and Control
A novel approach in considering such a system is to replicate
human sensory functions.

Optical detection Brain power

* IR cameras * Micro processors
* Machinery monitoring * Intelligent control
* Casualty location * Pattern recognition
* Real-time situational feedback * Neural networks

Electronic nose
* Ionising and photoelectric
smoke detectors Acoustic monitoring
* CO and CO2 detectors * Machinery monitoring
* Heat sensors * Shock/blast detection

Voice!
* Alarms
* Situational feedback
* Personal Address
Sensors and Control
System needs to diagnose whether or not there is a fire, rather than
simply detect one of the symptoms.

However, given that no two fires are identical, the system would
need to include the capacity to learn.
Conclusions
HPWM is an excellent solution for providing comprehensive cover
across a surface ship.
Sprinkler layout, Distribution Architecture and Pumping Strategy
explored
Also been found to offer blast mitigation
Ideally, total coverage required this will incur high infrastructure
costs.
Aside from the technical challenges, it will also demand a
comprehensive rethinking of how fires and damage control are
managed presently and how they are pictured in the future.
Surrounding issues of central control and management explored,
including discussion of potential sensors and their system
integration.
Thank you
Any questions?
Estimation of Orifice Flow
Rates for Flooding of
Damaged Ships
C. D. Wood, A. J. Sobey, D. A. Hudson, M. Tan & P. James

26th January 2011 The Damaged Ship

Overview of Presentation
Introduction

Numerical Solver

Compartment flooding

Sensitivity analysis

Conclusions

Future work

2
Introduction

3
Definition of a damaged ship
Side Shell Failure Hatch Cover Failure Corrosion Fatigue Crack Local Dent

Shell damage or
hatch failure Water Ingress Partial Loss of
Structure

Hold Flooding
Water ingress
Transverse Increase of Hull Decrease of Hull
Increase of hull Bulkhead Failure Girder Loads Girder Strength

girder loads
Progressive Flooding
to adjacent holds

Fluid loads and

motions team Loss of Stability Loss of Reserve Buoyancy Hull Girder Collapse

Collaboration
Loss of Vessel
Strength team

_______________________________

Loss of vessel scenarios (Paik & Thayamballi, 1998) 4

Limitations of state of the art

Effect of waves
Flooding rates
and motions on
through an
flooding rates
orifice

Roll decay Effect of internal

of a flooded sloshing on
ship motions

In combination, a highly non-linear problem

5
Limitations of state of the art

Effect of waves
Flooding rates
and motions on
through an
flooding rates
orifice

Roll decay Effect of internal

of a flooded sloshing on
ship motions

In combination, a highly non-linear problem

6
State of the art research
Experimental and Numerical Study on Progressive Flooding in Full-Scale
(Ruponen et al. 2009)
Fast attack craft Turku
Length 45m
Beam 8.8m
Displacement 221 tonnes

Research (experimental and modelling),

Time domain progressive flooding in calm water
Air stiffness
Permeability of compartments
Coefficient of discharge
Stability

Conclusions, 7
Bernoullis theorem applicable given accurate co-efficient of discharge
Orifice flow research
Existing research conducted in;

Pipe flow/orifice plates

Weirs and sluice gates

Turbo-machinery

Building ventilation

Human Larynx

Parameters investigated have not yet included the effect of shape,

Geometry taken from 2D geometry in Wood et al. (2009)

Experiments to take place over the next quarter

8
Aims and Objectives
Aims

To investigate flooding rate sensitivity for realistic scenarios

throughout a generic ship flooding process and investigate
methods for rapid estimations of subsequent orifice flow rates.

Objectives

Assess confidence in CFD (RANS) modelling

Investigate flooding rate sensitivity for orifice geometry

Assess rapid assessment methods for estimation of orifice flow

rates

Detail future work

9
Numerical Solver

10
Commercial CFD model
Software: Star CCM+ v 4.06

Solves RANS equations for

discretised time and space

Governing equations:
Conservation of mass

Conservation of momentum

Conversation of generic scalar quantities

11
Commercial CFD model
Interface capturing:
Void fraction

Fluid properties in each control volume, e.g. density

Surface Tension

Numerical scheme:
Algebraic system for transported variable

12
Confidence in CFD
A methodology for a reduction in the sources of error can be implemented.

Model errors - Boundary conditions and proximity effects, turbulence models

and initial conditions.

Discretisation errors - grid dependency studies

Convergence errors - finding a level of convergence that satisfies accuracy

whilst keeping computational cost to a minimum.

Round off errors - use of double precision.

Application uncertainties and code errors - by validating results against an

alternative solver code and experiments

13
Compartment Flooding:
Setup

14
Compartment flooding overview
Physical setup

Transient flooding

Partially submerged cylinder

Orifice in the endplate

Freely venting

Reynolds number range of 0.95x103 to 1.03x105 based on square root of

the orifice area.

15
Grid dependency
Domain

Case chosen at random

Hydrostatic pressure at external boundaries

No slip walls on cylinder

Atmospheric pressure at top boundary

Grid

Ranging from 6x104 to 2x106 cells

Solver

Various time step and convergence strategies

16
Solver setting dependency
A note on the continuity convergence from Wood et al. (2009)

3 x 10-2 2 x 10-2 8 x 10-3 2 x 10-3 7 x 10-4

2 x 10-4 1 x 10-4 17
Compartment flooding
Grid/Time dependency

S1 Initial survey, time step

of 0.001s, continuity 10-4 , up
to 300 iterations.

S2 Time step adjusted to

give C=1.

S3 Same time steps, 15

iterations per time step.

S4,5,6 3 stage solver

strategy, with C=0.5, 0.25
and 0.1 respectively.

18
Compartment flooding
Domain

Hydrostatic pressure at external boundaries

No slip walls on cylinder

Atmospheric pressure at top boundary

Grid

Approx. 290,000 cells

Refinement at the orifice 5x10-4 m 1x10-3 m

Cells at the external boundary 0.05m

Cell growth rate of 1.1

19
Compartment Flooding:
Cases

20
Compartment flooding
Variables

Shape - distorted regular polygons with 4 20 sides

Height
Variation in aspect ratio
Width

30 cases in total (10 cases per variable) where the input value
distributions were determined using a Latin Hypercube

21
Results

Bursting flow > Accelerating flow > Quasi-steady flow

22
Orifice velocity
Determining time to flood

Time to flood taken when

velocity falls to zero

Flooding rate varies almost

linearly

For this analysis

Flow rate is simplified to an

averaged flow rate taken from
time to flood

Averaged discharge co-efficient

is derived from averaged flow
rate 23
Results

24
Results tables 4

Shape Width Height Time to flood Avg. flood rate

Case
(n) (mm) (mm)
Area (m2)
(s) (m3s-1)
Avg. Cd 10
4 10 59 23 4.19E-04 13.48 0.00047194 0.508

10 10 23 59 4.19E-04 13.66 0.000465721 0.501

12

Shape Width Height Time to flood Avg. flood rate

Case Area (m2) Avg. Cd
(n) (mm) (mm) (s) (m3s-1)

12 5 83 29 1.41E-03 3.62 0.00175739 0.561 13

13 5 44 50 1.29E-03 4.25 0.001496882 0.523

19 5 71 47 1.96E-03 2.74 0.002321807 0.534

19

25
Sensitivity Analysis

26
Response Surface Method
Response Surface Method developed using Kriging

27
Response Surface Method
Response Surface Method developed using Kriging

28
Response Surface Method
Response Surface Method developed using Kriging

29
Monte Carlo Methods
Run simulation of events

Determine likelihood of variability occurring using probability

density function.

Model the event e.g. Orifice flow.

Score functions used for sensitivity analysis.

Sensitivity Normal Score Function

30
Results
Input Data
Damage Mean Coefficient of Distribution
Variable Variation
No. Sides 6 0.6 Normal
Width 45 0.5 Normal
Height 45 0.5 Normal

31
Sensitivity Results
Conclusions

32
Conclusions
Grid-time-solver setting dependency study has been performed
for the case of transient orifice flow solved using a RANS solver

CFD has been used to solve a number of cases of varying shape,

width and height

Sensitivity analysis has been performed on the results using a

response surface method combined with a Monte-Carlo
simulation

The results show that the parameters investigated here influence

the flow rate and therefore this requires further work with
validation to determine its full extent.

33
Future work

34
Future work
This investigation covering shape, height and width was
restricted to coarse mesh resolution

Detailed investigation to create a krigged response surface

for the variables,

Shape, width and height

Location

Depth

Level of pre-fill

Orifice edge roughness

Air ventilation restriction

Selected cases will be used for experimental validation

Test case for response surface method

35
Acknowledgements
Lloyds Register
Ben Cuckson
Ministry of Defence
Chris Cole
Colin Snell & Surface Ship Structures Team
University of Southampton
Prof. R. Shenoi
University College London
Dr. T. Smith
Dr. K. Drake & PhD students on this project

36
 Any Questions?

C. D. Wood, A. J. Sobey, D. Hudson, M. Tan & P. James

Contact: C.D.Wood@soton.ac.uk

26th January 2011 The Damaged Ship

 DEALINGWITHTHEHINGEDSHIP
AReviewoftheAnalyticalandPracticalIssues
basedupontheM.T.ElliSalvage

Capt.KEllam,SwireSalvage(Pte)Ltd,Singapore,ColinMoore,Herbert
EngineeringCorp.,USAandRobertTagg,HerbertEngineeringShanghai,
China

Capt. Ken Ellam, Swire Salvage (Pte) Ltd, Singapore, Colin Moore, Herbert Engineering Corp.,
USA and Robert Tagg, Herbert Engineering Shanghai, China

RINADamagedShipConference,London,January2011
OverviewofPresentation
Descriptionoftheship
SalvageSituation
StrengthoftheMTElli
StepsintheSalvage
Postmortemanalysis
Lessonslearned

RINADamagedShipConference,London,January2011
MTElli
Aframax crudeoilcarrierwithdoublesides
andsinglebottom
Builtin1984asMTCanopus,onesistership
7centercargotanks,2sloptanksand6pairs
ofsegregatedsideballasttanks
Allbunkerfuelwascarriedintheengineroom
area

RINADamagedShipConference,London,January2011
 SalvageSituation
Ranaground3monthsearlier
Refloated
TowedtoPortofSuez

Over21mofhullgirderdeflection
Undergoingtankcleaninginpreparationfor
drydocking

RINADamagedShipConference,London,January2011
Otherviews

RINADamagedShipConference,London,January2011
StructuralDamage

RINADamagedShipConference,London,January2011
StrengthoftheMTElli
Estimatedpre
casualtyloading
showedbending
momentswellin
excessofclass
allowablesforAt
SeaorInHarbour

RINADamagedShipConference,London,January2011
SalvageSteps
InitialResponse
TheFirstDays
Stage1 RemovalofPollutants
Stage2 Preparations
Stage3 Ballasting
Stage4 Separation
Stage5 PartingandRefloating

RINADamagedShipConference,London,January2011
InitialResponse
TheCanalAuthority
intentionallygrounded
theship
Afterbuckling,the
potentialenergystoredin
thebeamleadtotheship InitialHECSALVmodel
deformingsuchthatits 10

8
Shear

x
+Hog
3

finalpositionwaspastthe
x
2
6

pointofzeromomentin 2
1

thehinge
0 0
F47 F49 F52 F54 F57 Mx
F59 F62 F64 F67 F69 F72 F74 F77 F79 F85 G0A
G0F
-2

-1
-4

Theshipendedupina -6

-8
x
x
-2

saggingmoment -10
x=interpolated -Sag
-3

condition
Shear (MT)/10^3 AT SEA Shear Allowable (MT)/10^3
Moment (m-MT)/10^5 AT SEA Moment Allowable (m-MT) / 10 ^ 5

BMaftercollapse
RINADamagedShipConference,London,January2011
Stage2 Preparations&Ballasting
Bulkheadrepairsrequired
atnumerouspenetrations.
Ballastpipelinesblanked.
Detailedballastingplan
calculatedusingHecsalv.
Simultaneoustransferand
deballasting requiredusing
submersiblepumps.
Controlledgroundingto
predeterminedraft.
RINADamagedShipConference,London,January2011
Stage4 Separation
Reverseddeflectionofhull
thoughballastingtosheer
bottomgirders.
Sheerlegsusedtocut
throughdeckplatingand
longitudinalgirders.
Refloatedandsecured
forwardsectionfirst.
Aftsectionrefloatedafter
COTrepairscompleted
underwatertoensure
watertightintegrityofhull
duringtowage.
RINADamagedShipConference,London,January2011
MidshipSectionProperties
AH32andmild
steel
Framespacing
4.58m

Increaseddeck
thicknessoutboard
toaccountfor
sloshingloads
Intermediate
transversesinway
ofsloshing
reinforcement

SMdeck>SMkeel

RINADamagedShipConference,London,January2011
UltimateStrength(ULTSTR)
9.00E+05 Moment(mMT)
Hog
8.00E+05
SWBendingMoment
7.00E+05

6.00E+05

Intact(compressionremoval)
5.00E+05
Intact(noremoval)
Intact(Longertrippingspan)
4.00E+05 5%corrosion
10%corrosion
SWBMEstimated
3.00E+05

2.00E+05

1.00E+05

0.00E+00 Curvature
0.00E+00 5.00E05 1.00E04 1.50E04 2.00E04 2.50E04 3.00E04 3.50E04
RINADamagedShipConference,London,January2011
Predictinghingebehavior
Topredictthehingeangleas
ballastingproceedsrequiresassessing
theresponseofthebuckledstructure
asadditionalsaggingmomentis
applied
Thisisanonlinearstructuralproblem
thatisbeyondpracticalsolutionina
salvageenvironment
Alternativeapproachwastoevaluate
thesaggingmomentversushinge
angle(orequivalentlyhogdeflection)
Analysisshowedthatasthedeflection
decreasedthesaggingmomentwould
alsodecrease,thusthepotentialfor
uncontrolledreactionofthehull
girderwaslimited
RINADamagedShipConference,London,January2011
Forcesinthehingepriortoseparation
Thegoalwasto
ensurethetwo
halvesoftheship
remained Afthalfagroundpivotingatforwardedge

groundedatthe 5

locationofthecut 3

afterseparation
2

17:1500 17:2100 18:0300 18:0900 18:1500 18:2100

Time

Tide Height (m) Ground Reaction [D] (MT)/10^3

Timehistoryofgroundreaction forwardhalf

RINADamagedShipConference,London,January2011
Asuccessfulsalvage

RINADamagedShipConference,London,January2011
Acknowledgements
We would like to thank the management of
Five Oceans Salvage Consultants for allowing
us to make use of both data as well as
photographs from the Elli salvage operation.

RINADamagedShipConference,London,January2011
 The RINA International Conference 26-27Jan 2011
The Damaged Ship

The Rapid Reliability and Risk Assessment of

Damaged Ship Structure

By
Dr Mohammad Shahid PhD
PE, MRINA, CEng

Binary Systems & Engineering (UK) Ltd Company No: 6711246

241 Glencoat Drive, Paisley, SCOTLAND Registered Office:
Tel: +44 (0) 141 5305927 Fax: +44 (0) 141 8896781 788-790 Finchley Road
London, UK
Slide No: 1 Web: www.BsysE.co.uk email: services@bsyse.co.uk
NW11 7TJ
 Presentation Scheme

Introduction
Review of the State of the Art
Ships Structural Damage Assessment
Ships Structure Residual Strength Assessment

Ship Structural Response Model

Modeling for Emergency Response Services
Response Surface (RS)
Artificial Neural Network (ANN)
Comparison of ANN & RS Model
Comparison of FEA based ANN & RS Model
Rapid Structural Assessment for Emergency Response
Services
Rapid Damage Assessment
Rapid Structure Residual Strength Assessment
Slide No: 2
 Presentation Scheme

Ultimate Strength of Hull Girder

Structural Response Model
Reliability & Risk Analysis
Ship Loads & Reliability Index

Conclusion
Discussion

Slide No: 3
 Introduction

For Emergency Response Services (ERS);

The rapid reliability and risk assessment of damaged ships structural

strength is probably the most important element to minimise risk of
total failure of hull structure and total lost of ship as a consequence.

At an early stage of an incident, often limited information are available that

may only includes:

Speeds of ships involved, direction and location of impact

Prevailing environmental conditions.
And there, if any, may only be subjective information available on structural
damage.

The two important aspects for ERS are, therefore,

Ability to predict extent of structural damage that is likely to occur on the

basis of limited information as mentioned above and,
Ability to make assessment of residual strength of ship structure subsequent
to damage incident

Slide No: 4
 Review of the State of the Art
Structural Damage Assessment
Energy Balance
Originally proposed by Minorsky in 1959

Impact Energy = Energy for Structural Deformation

Minorskys Formula
Ships dynamics E d 47.2 RT 32.7

The empirical methods,

The simplified analytical methods,
The simple finite element methods (FEM), and
The explicit dynamic FEM

Super-element Method
Dividing structure into elements
Evaluation of failure mode and energy
Nonlinear finite element analysis to
evaluate total energy of deformation
Slide No: 5
 Review of the State of the Art
Residual Strength Assessment
Material Removal
Nature of Impact &
Method of Evaluation

The general approach is to use

conventional methods removing
damaged material for residual strength
assessment. Simulated Damage

Method suitable for low energy impact

with negligible deformation and residual
Residual Stress Field
stress in adjoining structure
%age of y
Ship collision and grounding are mostly
11.1%
high energy impacts with significant
deformation and residual stresses in 22.28%
adjoining structure as shown in the figure 33.33%
for dynamic simulation of ship collision. 44.38%
Conventional Method to be revised 55.65%
Use of static/dynamic nonlinear FEM
Slide No: 6
 Ship Structure Response Model

Modeling for
Emergency Response Services

Step1: Design of experiment and FE

analysis

Step2: Response modelling using system

identification and simulation technique
including response surface, artificial neural
networks

Step3: Using response model with

reliability method for determination of risk
of damage to ship structure and reliability
of residual strength of ship after damage.

Slide No: 7
 Ship Structure Response Model

Response Surface
General Regression Model
n n n
Y C 0 Ci X i Cij X i . X j
i 1 i 1 j 1
Xi design variables, C0, Ci, Cij - modeling coefficients, modeling error
Computational cost & accuracy depends upon
Number of variables
Data generation; design of experiment
Response type; linear or nonlinear
Improvement targets
Reduce data set for simulation &
Accuracy
Improvement techniques
Piecewise regression method
Stepwise regression/adaptive response function
Model selection method
Model correction factor method
Rigid & non-index-able with global scope of coefficients
Slide No: 8
 Ship Structure Response Model

Response Surface

Selective Regression Model

based on prior knowledge of system response

For ultimate strength

of hull girder
c1 c2
Mv Mh
1
M uv M uh

Slide No: 9
 Ship Structure Response Model

Response Surface
Problem in Nonlinear Response Modeling
n n n
Y C 0 Ci X i Cij X i . X j
c1 c2 i 1 i 1 j 1
Mv M
h 1
M uv M uh

Global in nature dont

adapt well to highly
nonlinear local regions

Difficult dynamic
updating around design
point

Slide No: 10
 Ship Structure Response Model

Artificial Neural Network (ANN)

Based on weighted function as shown below

n
y ( x) wi xi

i 1
Motivation to use ANN
Reduce data set for simulation
Improved accuracy
Do not suffer from curse of dimensionality as of RSF

Techniques used very well in many engineering fields

but a few known application in marine structural simulation
& reliability analysis

Dynamic updating & adaptive training to reduce size of

training data set especially for FE base reliability analysis.

Slide No: 11
 Ship Structure Response Model

Comparison of ANN & Response Surface Model

Slide No: 12
 Ship Structure Response Model

Comparison of FEA based ANN & Response Surface Model

Slide No: 13
 Ship Structure Response Model
Rapid Structural Assessment for ERS

Rapid Damage Assessment M ade el

a dy Mod
Re ponse
R es

Slide No: 14
 Ship Structure Response Model
Residual Strength Assessment

Rapid Residual Strength Assessment

M ade el
a dy Mod
Re ponse
R es

Slide No: 15
 Residual Strength Assessment
Ultimate Strength of Hull Girder

Structural Response Model

Using Ultimate Moment Capacity Mu as parameter
for ultimate strength of hull girder as follows:
M v , M h , M t , Fs ,

M u f D h , D w , D d , Dl , Dv ,

T
age y, , E
Parameters for Ship Loading Parameters for Structural Damage
Mv vertical moment, Dh height of damaged structural part,
Mh horizontal moment, Dw width of damaged structural part,
Mt torsional moment, Dd depth of damaged structural part,
Fs shear force Dl longitudinal location of structural damage,
Dv vertical location of structural damage,
Parameters for Structural Material
y yield strength,
E Modulus of elasticity,
Tage aging related structural degradation.
Slide No: 16
 Residual Strength Assessment
Ultimate Strength of Hull Girder

Structural Response Model

M v , M h , M t , Fs ,

M u f D h , D w , D d , Dl , Dv ,

T age , y , E
For a specific structural damage
D(Dh, Dw, Dd, Dl, Dv)

MU D , L ,Tage ,y , E
f ( M v , M h , M t , Fs )
For hull girder moment interaction, we can write,
c1 c2
MV M
H 1
M M
VU HU
Where
*
*
M VU Reduced Vertical Ultimate Moment M HU Reduced Horizontal Ultimate Moment
because of Mt and Fs because of Mt and Fs

M VU M VU R vt R vs
M HU M HU R ht R hs
Rvt - Reduction factor for Torsion, Rht - Reduction factor for Torsion,
Rvs - Reduction factor for Shear Force Rhs - Reduction factor for Shear Force
Slide No: 17
 Residual Strength Assessment
Ultimate Strength of Hull Girder
c1 c2
MV M
H 1
M VU M HU

M
M VU R vt R vs M HU M HU R ht R hs
VU
c3 c4
MV M MH
c7
M
c8
T 1 T 1
M VU M TU M HU M TU
1 / c 3 1 / c 7
M V M T
c4

M T
c8
Rvs 1 M
Rht H 1
M VU M TU M HU M TU

c5 c6
MV F
S
c9 c10
1 MH F
M VU FSU S 1
1 / c 5
M HU FSU 1/ c 9
FS FS
c6 c10
MV MH
Rvs 1 Rhs 1
M VU FSU
M HU FSU
Slide No: 18

 Residual Strength Assessment
Ultimate Strength of Hull Girder

Structural Response Model

Using normalized parameters e.g.

Normalized vertical moment Mv * = M V

M VU
We can write, a parametric model for ultimate failure of hull structure as follows:

c1
M V*



1 M
*
T 1 F
c 4 1/ c3
S
* c6
1/ c5


c2
M H*
1


1 M
*
T 1 F
c8 1 / c 7
S
* c10
1/ c9

Slide No: 19
 Residual Strength Assessment
Ultimate Strength of Hull Girder

Structural Response Model

The failure state of hull girder is, therefore,
given by:

c1
M V*



1 M
*
T 1 F
c 4 1/ c3
S
* c6
1/ c5


c2
M H*
1 0


1 M
*
T 1 F
c8 1 / c 7
S
* c10
1/ c9

Slide No: 20
 Residual Strength Assessment
Ultimate Strength of Hull Girder

Structural Response Model

The limit state function g such that g0
represents hull girder failure shall be
as follows:

c1

M V*
g 1

1 M T*

1 F
c 4 1/ c3
S
* c6
1/ c5

c2

M *

1 F
H
1/ c9
1 M
*
T
c8 1 / c 7
S
* c10

Slide No: 21
 Residual Strength Assessment
Ultimate Strength of Hull Girder

Reliability and Risk Analysis

Probability of Failure Pf f ( X )dX

g ( X )0

Reliability Index 1 ( P f )

Where
F ( X ) F ( M V , M H , M t , Fs ) Joint probability density function for moments
and shear force
And Limit State
c1

M V*
g M v , M H , M T , Fs 1


1 M T*

c4

1/ c 3

1 FS
* c6
1/ c 5

c2


*
MH


1 MT
* c8

1/ c 7

1 FS* c10

1/ c 9

Slide No: 22
 Rapid Reliability & Risk Assessment for ERS
Ultimate Strength of Hull Girder

Ships Loads & Reliability Index

Ships Load

M M sw k w ( M w k d M d )
M Bending Moment
Msw Stillwater bending moment
Mw Wave induced bending moment
Md Dynamic bending moment
Kw & Kd Load combination factor

Reliability Index Polar Diagram

Intact ship and damaged ship loads Reliability Index Vs Ship Heading

Wave induced loads & Ship speed/heading

Slide No: 23
 Rapid Reliability & Risk Assessment for
Emergency Response Services
Conclusions

Reviewed state of the art for assessment of damage to ship structure

subsequent to collision/grounding & assessment of residual structural strength.

Rapid Response for Emergency Services requires system to make informed

decision at an early stage of incident when little information are available.

A comprehensive system for rapid assessment of structural damage as well as

residual strength based on ready-made Structural Response Model developed
using Finite Element Analysis, Response Surface and Artificial Neural Network
techniques is described.

Structural Response Model can readily be used with probabilistic analysis tools
for Reliability & Risk Analysis

Introduced Reliability Index Polar Diagram for rapid reference to structural

reliability index for varying sea state and ship heading that shall be of immense use
for Emergency Response Services as well as during normal operation to manage
ship in severe sea conditions.

Slide No: 24
 Rapid Reliability & Risk Assessment for
Emergency Response Services

Other Services Available:

Discussion
Conventional Naval
Architectural Evaluations

Structural Analysis including

Static Linear / Non Linear &
Dynamic Finite Element Analysis &
reliability/risk assessment
THANK YOU
Model test in Numerical Towing
Tank
Contact for any further query:
shahid@bsyse.co.uk
Visit our web site for further info:
www.bsyse.co.uk

Slide No: 25
Framework for the Survivability
Assessment of Damaged Ships
(Progressive Structural Failure)

Seungmin Kwon
Prof. Dracos Vassalos, Dr. George Mermiris
Ship Stability Research Centre
Dept. of Naval Architecture & Marine Engineering
University of Strathclyde

Damaged Ship Conference, 26-27 January 2011, RINA

 OUTLINE

Introduction
Framework Proposed
Verification
Conclusions
Future Work
 INTRODUCTION

Loss of lives
Survivability of damaged ships?
Loss of ships
Pollution

Assessment of damage stability and potential capsizing have

been covered in the literature (e.g. Vassalos et al., 2005) Assessment of residual strength has been
conducted in literature (e.g. SAFEDOR, 2006)

INITIAL DAMAGE EXTENT IS FIXED

INTRODUCTION

LOSS OF THE TANKER PRESTIGE (NOV.

2002)

Loss of Prestige signifies the importance of the

progressive structural failure and the quantification
of the residual strength in time domain
 INTRODUCTION
SAFE RETURN TO PORT (IMO,
Prevention 2004)
Time to remain habitable (3hr)
Abandon Time to recover
(30 min) (5 days)
/ Protection

Rescue to a place of safety

Casualty No Functions / Systems
Casualty
threshold
detection availability
met?

Abandon
Yes Prepare to Recovery
ship into
abandon from survival
survival
ship craft
craft

Time
The ship should be designed A ship should remain afloat,
to be her own best life boat upright and habitable for 5days

Progressive structural failure can develop and

lead to exacerbated flooding or total hull collapse
(depending on the initial damage extents, weather
conditions and flooding mechanism)
INTRODUCTION

This presentation proposes

A methodology to assess the survivability of damaged

ships in time domain;

and is focused especially on

The recent developments of a first-principles

methodology for the simulation of progressive
structural failure
FRAMEWORK PROPOSED
HIGH LEVEL PROCESS
Initialdamage
(collision/grounding)

i=i+1 Timestep
t(i)
The framework enables to
Loadingt(i1)
Environmen
t
Flooding Evaluate survivability of damaged
ships and support decision making
Responset(i)
on board
Progressive
structural
failure Estimate time when a damaged ship
fails to survive and help salvage
Residualstrength
(BM)
Damagestability(GM) operation
t(i)
t(i)

Improve safety of ships through

implementation of risk-based design
Pass Criterion
(BMBMmax)
AND(GM0)

Fail
Termination
Hullcollapse
and/orcapsize
 FRAMEWORK PROPOSED
ENVIRONMENTAL LOAD
Initialdamage
Main features
WaveHogging WaveSagging
(collision/grounding)
250
- in-house code

Bendingmoment,[MNm]
200
i=i+1 Timestep
t(i) - 3D panel
150
- Greens function
- 6DOF
100
Loadingt(i1)
- Nonlinear wave 50

Environment Flooding
- Freq. domain 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4

,[rad/sec]

Responset(i)
Progressive Shipgeometry Intactloadingcondition
structural Initialdamage
Shipspeed (draft,trim,TPC,MTC)
failure Weightdistribution

Residualstrength
Damagestability(GM)
(BM) Wavecondition Floodedcompartment
t(i) Newdraft
t(i) (Hs,Tz) (location,size) &trim

Pass Criterion
(BMBMmax)
Seakeeping &
AND(GM0)
WaveLoad Shipmotion
Fail Analysis

Termination
Hullcollapse WaveBM
and/orcapsize StillwaterBM
 FRAMEWORK PROPOSED
FLOODING LOAD
Initialdamage Initialdamage Intactloadingcondition
Shipgeometry
(collision/grounding) (location,size) (draft,trim,TPC,MTC)

i=i+1 Timestep
t(i)
Floodedcompartment Newdraft
Shipmotions Damagesize (location,size)
Loadingt(i1) &trim
Wavecondition
(Hs,Tz)
Environment Flooding

Main features Flooding

Simulation
Responset(i)
- in-house code
Progressive
structural
- Navier-Stokes eq. Flooding
failure - CFD (Volume Of Fluid) pressure

Residualstrength
Damagestability(GM)
(BM)
t(i)
t(i)

Pass Criterion
(BMBMmax)
AND(GM0)

Fail
Termination
Hullcollapse Dynamic pressure distribution
and/orcapsize near the opening
 FRAMEWORK PROPOSED
PROGRESSIVE STRUCTURAL
FAILURE
Initialdamage
(collision/grounding) Main features
- Crack propagation
i=i+1 Timestep
t(i) - Stress Intensity Factor
- FEA (VCCT) Material
constants
Initialdamage
(location,size, Shipdrawings

Loadingt(i1) - Modified Paris Law (C,m,Kc,Kth0 ) cracksize)

Environment Flooding

Wave Flooding WaveBM

Damagesize
condition pressure StillwaterBM
Cracksize
Responset(i)
Progressive
structural
failure FEA
(ABAQUS/VCCT)

Residualstrength Kmax,Kmin,R,
Damagestability(GM)
(BM) Keff,Kth
t(i)
t(i)
Damage
Propagation
Pass Criterion Analysis
(BMBMmax)
AND(GM0) Numberofcycles Crackpropagation
(dN= t/Tz) rate(da/dN)
Fail
Termination Extended
Hullcollapse Damage&cracksize
and/orcapsize
 FRAMEWORK PROPOSED
PROGRESSIVE STRUCTURAL
FAILURE
Crack propagation (LEFM + Paris Law)
Aviation industry (Farahmand et al., 2007)
Shipbuilding industry (Dexter et al., 2000 & 2004)
Modified Paris Law SIF using VCCT Crack Propagation

Time
N
Period K eff
da

D K eff K th 1
m

dN K IC K max
VCCT: Strain energy release rate, G
K2
Stress Intensity Factor : for Plane Stress VCCT : a capability that analyzes
a constant relates global E brittle interfacial crack propagation
G 2
applied stress with local K (1 2 ) for Plane Strain due to delamination/debonding using
stress near crack tip E the Virtual Crack Closure Technique.
 FRAMEWORK PROPOSED
RESIDUAL STRENGTH
Main features
Initialdamage
(collision/grounding)
- Ultimate bending strength
i=i+1 Timestep - Smiths method
t(i)
- Load-shortening curve (location,size,
Initialdamage Shipsectiondrawing
(plate&stiffener
RangeofBM
curvature
cracksize) information) (saggingtohogging)
Loadingt(i1)
Load
Environment Flooding
shortening
curve

Responset(i)
Damagesize
Progressive Cracksize
structural
failure

Residualstrength Neutral
Damagestability(GM) axis
(BM)
t(i)
t(i)

Residual
Pass Criterion Strength
(BMBMmax) Analysis
AND(GM0)

Fail
Ultimateresidual
Termination
bendingstrength
Hullcollapse
(hogging,sagging)
and/orcapsize
FRAMEWORK PROPOSED
RESIDUAL STRENGTH

Analysis method adopted

Simplified moment-curvature relationship

-Plate induced failure

-Flexural buckling failure of column
-Tripping failure of stiffener

With a given curvature,

- Average strain based on elements
location
- Average stress from corresponding
load-shortening curve
BM i Faxial ,i d i
BM total BM i
where, Faxial ,i i Ai i

d i is distanct to the neutral axis

 FRAMEWORK PROPOSED
DAMAGE STABILITY
Main features a) Input
Initialdamage
(collision/grounding)
- PROTEUS - Ship geometry (hull & internal
i=i+1 Timestep (in-house code) compartments)
t(i)
- Strip theory - Damage size & location
- Nonlinear wave - Wave condition
Loadingt(i1)
- Bernoullis equation b) Output
Environment Flooding - etc. - Ship motions
- Floodwater mass & motions
- Flow of floodwater through opening
Responset(i)
- Environmental forces
Progressive
structural
failure

Residualstrength
Damagestability(GM)
(BM)
t(i)
t(i)

Pass Criterion
(BMBMmax)
AND(GM0)

Fail
Termination
Hullcollapse
and/orcapsize
 VERIFICATION
AFRAMAX TANKER

Information of the target ship

112,700 DWT
12 cargo tanks
Double skins
Type of dimension Value
Length O. A. 250.17 m
Length B. P. 239.00 m
Breadth MLD. 44.00 m
Depth MLD. 21.00 m
Draught MLD.
14.60 m
(Design)
Draught MLD.
14.60 m
(Scantling)
Frame spacing 3.78 m
 VERIFICATION
AFRAMAX TANKER

Bottom damage
Extent of damage in section
from ABS Guidance for residual
strength assessment
0.5% of initial crack size

1.875m

7.65m
Initialcrack,a0 =38.25mm
 Next Page

VERIFICATION
AFRAMAX TANKER

Bottom damage
Location of damage in lengthwise from ABS Guidance for residual
strength assessment
2 frame spacing of damage length in No. 4 WBT

No.4C.O.T.(P&S)
Doublebottom

T.BHD Bottomshell T.BHD

7.56m:2framespacing
 Previous Page

VERIFICATION
AFRAMAX TANKER

Wave loads calculation 18

Waveheight,Hs Zerocrossingperiod,Tz

Waveheight,[m],Period,[sec]
Full load condition (1329 MN-m 16
14
of still water sagging BM) 12
10

Wave data from a report of the 8

6
Prestige accident for 6 days 4
2

Linear regular wave 0

0 12 24 36 48 60 72 84 96 108 120 132 144
1day2days3days4days 5days6days
Time,[hour]
WaveHogging WaveSagging
Hogging Sagging
250
BendingmomentRAO,[MNm/m]

1600

Wavebendingmoment,[MNm]
200 1400
1200
150 1000
800
100
600
50 400
200
0 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 12 24 36 48 60 72 84 96 108 120 132 144
30 sec 1day2days3days4days 5days6days
,[rad/sec] Time,[hour]
 VERIFICATION
AFRAMAX TANKER

Progressive structural failure

3 hold FE model
4 nodes shell & 2 nodes beam elements
8 nodes solid element
shell to solid coupling
 VERIFICATION
AFRAMAX TANKER

Progressive structural failure

SIF calculation with VCCT
Unit vertical sagging BM of 1.0 MN-m is applied
x 10-3

320
Initial
270 damage

220
SIF,[MPam]

170

120

70

20
3.5 5.5 7.5 9.5 11.5 13.5 15.5 17.5
Halfdamagesize(halfinitialopening+initialcracksize),[m]
 VERIFICATION
AFRAMAX TANKER

Progressive structural failure

K eff
Crack propagation law dN D K K th
da
1
m

K IC K max
eff

D = 7.3210-11 and m = 2.37

D and m are calculated so as the linear region of the crack propagation law
fits the Paris law recommended by BS PD 6493
K IC 140 MPa m and K th ,0 2.45MPa m
1mm/cycle or 2mm/cycle 25
Stable damagegrowth Unstable damagegrowth
are used above KIC

Halfdamagelength,[m]
20
2mm/cycle cap
15

1mm/cycle cap
10

0
0 12 24 36 48 60 72 84 96 108 120 132 144
1day2days3days4days 5days6days
Kth Kmax
Time,[hour]
 VERIFICATION
AFRAMAX TANKER
Intact initialdamage 1daylater

Residual strength 2dayslater 3dayslater 4dayslater

Bendingmomentcapacity,[MNm]
2.E+04
1mm/cycle case
Failure criterion as 25% reduction 1.E+04

in bending strength of intact ship 5.E+03

Sagging Hogging
0.E+00

5.E+03

1.E+04
6.E04 4.E04 2.E04 0.E+00 2.E04 4.E04 6.E04
Curvature,[1/m]
Hogging(1mm/cycle) Sagging(1mm/cycle)
Hogging(2mm/cycle) Sagging(2mm/cycle)
BM Hogging Sagging
Capacity 1mm Diff. 2mm Diff. 1mm Diff. 2mm Diff.
15000
(MN-m) /cycle (%) /cycle (%) /cycle (%) /cycle (%)
Bendingcapacity,[MNm]

IntactHogging
10000 Intact 10817 NA 10817 NA 7866 NA 7866 NA
Init. dam 10229 -5.44 10229 -5.44 7731 -1.72 7731 -1.72
5000 1 day 10225 -5.47 10225 -5.47 7728 -1.75 7728 -1.75
1day2days3days4days 5days 2 days 10221 -5.51 10221 -5.51 7726 -1.78 7726 -1.78
0 2.5 days 10093 -6.69 9967 -7.86 7569 -3.78 7664 -2.56
0 12 24 36 48 60 72 84 96 108 120 3 days 9545 -11.76 8767 -18.96 7560 -3.89 7305 -7.14
5000
3.5 days 9055 -16.29 7494 -30.72 7391 -6.04 6814 -13.37
10000
IntactSagging 4 days 8434 -22.03 7169 -8.86
Time,[hour] 4.5 days 7466 -30.98 6787 -13.72
 CONCLUSION

A framework for the survivability of damaged ships is proposed

with inclusion of progressive structural failure analysis.

A methodology for progressive structural failure analysis

comprising crack propagation law and VCCT for ABAQUS is
promising for the problem at hand and conforms to past experience
from the salvage industry.

Progressive structural failure can be analysed for a succession of sea

states and offer invaluable information for the residual strength of the
ship under consideration. Such time-varying knowledge can be
deployed in the operational stage in emergency situations and in the
safe-return-to-port context or in the design stage (forming a part of
the risk-based design methodology).

The resulting timeline of the damage propagation can contribute to

more robust decision-making and the aversion of Prestige-like
accidents in the future.
CHALLENGE AND FUTURE
WORK
The most evident challenge is the validation of the
proposed methodology.

The ultimate aim is conducting the necessary simulations

(coupled with flooding) fully in the time domain and
recreating the Prestige accident.

It is hoped that the added value of the methodology for

safeguarding human life and environment will be clearly
demonstrated to the scientific community and the industry at
large.
Q&A
Global Wave Loads on Damaged Ship Structures:

An Experimental Procedure and some Preliminary

Results
TheDamagedShipConference,RINA,London26th 27th January

DFone,UniversityCollegeLondon,UK
TWPSmith,UniversityCollegeLondon,UK
JBorg,LloydsRegister,UK
KRDrake,UniversityCollegeLondon,UK
Contents of Presentation

Motivation for Work

The Damaged Ship Problem

1 DOF and 2 DOF Experiments to Assess

Floodwater Behaviour

2 DOF Experiments to Assess Global Wave

Loads

Conclusions
Motivation for Work

July2002:HMSNottinghamrunsagroundoffAustraliancoast

QinetiQ,Rosyth,UK GRC,Portsmouth,UK
Wasim,Precal,PRETTY SeagoingParamarine
PanelMethod QuasistaticBalance

StaticMassofFloodwater

Towhatextentcanweassumethatfloodwatercanbetreatedasa fixedmass?
Whatshouldthatmassbe?
The Damaged Ship Problem

Damage Event

TimetoFlood

Time
ShipMotions

Transient Progressive Equilibrium/

Flooding Flooding Flooded

The structural integrity of a damaged ship when at a flooded equilibrium

Motivation for Work
Weather
Forecast

ImpendingSea
State

Quantification
Damage
Event
DistressCall
? ofStructural
Integrity
Decision /
Solution

Damage
Location
Computational Assumptions
Known unknowns:
Is ingress/egress flow quasi-steady and can it be estimated
from a hydrostatic pressure field?
Is sloshing significant?
Can trapped air be ignored?
Does ingress/egress affect coupling between degrees of
freedom?
Can results from experiments on a section be applied
successfully to a whole ship?

Unknown unknowns:
?
Quasi-Steady Assumption

v1 = 0
Ad
Awout
s
Awin

h1 w
equilibrium
waterline v
V2 Ad
h2 = 0

Torricellis Formula (1643): Volumetric Flow Rate, Q:

v2 2 g h1 h2 Q Cd Ad 2gh1 h2
Experimental Setup Damaged Section
Wave probes
Load cells

Damage orifice Restraining beam

0.315 m

0.7 m
1.2 m
1 DOF Floodwater Behaviour - Results
Low Oscillation Frequency: 10mm Amp, 0.5Hz

Approximately quasi steady behaviour

Large amounts of floodwater ingress / egress

Intermediate Oscillation Frequency: 40mm Amp, 1Hz

Little floodwater ingress / egress

High Oscillation Frequency: 40mm Amp, 1.66Hz

Unsteady behaviour
Large amounts of floodwater ingress / egress
2 DOF Floodwater Experiments
Stern Optical Connecting Bow Optical
Wave Probes
Sensor Arm Sensor

Video
Camera
Wave
Location
Direction

Is sloshing significant in damage floodwater dynamics?

Does floodwater dynamics affect coupling between DOF?
Motion constrained to 2 DOF (Heave and Pitch)
Intact and 2 centrally located damage cases tested
2 DOF Floodwater Experiments Qualitative Results

Investigating the effect of a damage orifice on floodwater dynamics when

compared with an intact compartment containing an equivalent mass of
floodwater

f=1.06Hz f=1.06Hz
IntactCompartmentwithequivalentvolume DamagedCompartmentexperiencing
offloodwater ingress/egressoffloodwater
Global Wave Loading on a Damaged Ship Model

PlanView
Global Wave Loading on a Damaged Ship Model

PlanView
Global Wave Loading on a Damaged Ship Model

PlanView
Global Wave Loading on a Damaged Ship Model

PlanView
Global Wave Loading on a Damaged Ship Model

PlanView
Global Wave Loading on a Damaged Ship Model

PlanView
Global Wave Loading on a Damaged Ship Model

Experiment to measure the effect

of damage size and location on
global vertical loads

Constrained to Heave and Pitch

Zero and slow forward speed using

fully wireless data acquisition
system.

Internal fluid free surface

behaviour.

PlanView
Global Wave Loading on a Damaged Ship Model

Experiment to measure the effect

of damage size and location on
global vertical loads

Constrained to Heave and Pitch

Zero and slow forward speed using

fully wireless data acquisition
system.

Internal fluid free surface

behaviour.

PlanView
Global Wave Loading on a Damaged Ship Model

Experiment to measure the effect

of damage size and location on
global vertical loads

Constrained to Heave and Pitch

Zero and slow forward speed using

fully wireless data acquisition
system.

Internal fluid free surface

behaviour.

PlanView
Global Wave Loading on a Damaged Ship Model

Experiment to measure the effect

of damage size and location on
global vertical loads

Constrained to Heave and Pitch

Zero and slow forward speed using

fully wireless data acquisition
system.

Internal fluid free surface

behaviour.

PlanView
Global Wave Loading on a Damaged Ship Model

Experiment to measure the effect

of damage size and location on
global vertical loads

Constrained to Heave and Pitch

Zero and slow forward speed using

fully wireless data acquisition
system.

Internal fluid free surface

behaviour.

PlanView
Global Wave Loading on a Damaged Ship Model

Experiment to measure the effect

of damage size and location on
global vertical loads

Constrained to Heave and Pitch

Zero and slow forward speed using

fully wireless data acquisition
system.

Internal fluid free surface

behaviour.

PlanView
Global Wave Loading on a Damaged Ship Model

Experiment to measure the effect

of damage size and location on
global vertical loads

Constrained to Heave and Pitch

Zero and slow forward speed using

fully wireless data acquisition
system.

Internal fluid free surface

behaviour.

PlanView
Global Wave Loading on a Damaged Ship Model

Experiment to measure the effect

of damage size and location on
global vertical loads

Constrained to Heave and Pitch

Zero and slow forward speed using

fully wireless data acquisition
system.

Internal fluid free surface

behaviour.

PlanView
Global Wave Loading on a Damaged Ship Model

Experiment to measure the effect

of damage size and location on
global vertical loads

Constrained to Heave and Pitch

Zero and slow forward speed using

fully wireless data acquisition
system.

Internal fluid free surface

behaviour.

PlanView
Global Wave Loading on a Damaged Ship Model

Experiment to measure the effect

of damage size and location on
global vertical loads

Constrained to Heave and Pitch

Zero and slow forward speed using

fully wireless data acquisition
system.

Internal fluid free surface

behaviour.

PlanView
Validation of Experimental Setup Intact Results

Comparison of experimental heave

response with prediction from
modified strip theory tool currently
being developed at UCL

Results show good agreement over

the majority of the frequency range

Comparison of experimental pitch

response with prediction from
modified strip theory tool

Results show good agreement over

the majority of the frequency range
Validation of Experimental Setup Intact Results

Midship Vertical Bending Moment results compared with modified strip

theory prediction
Damaged Model Motions Heave

Comparison of Heave RAOs with the inclusion of experimental scatter.

Preliminary results suggest there is no significant change in heave motion with

the possible exception of wave frequencies where = L.
Damaged Model Motions Pitch

Comparison of Pitch RAOs with the inclusion of experimental scatter.

Preliminary results suggest there is no significant change in pitch

response within the range of wave frequencies investigated.
Damaged Model Loads Vertical Bending Moment

Preliminary results suggest there is a significant change in BM response

with changing damage orifice size

Change in BM does not appear to be linearly proportional to damage width.

2 DOF Free Surface Investigation Quantitative Results

Preliminary analysis of floodwater behaviour involves the assumption that the

internal free surface behaves like a flat lid (longitudinal heave and pitch)

Initial 2 DOF results for floodwater heave and pitch response suggest that
damage orifice size significantly affects internal free surface behaviour.

Floodwater Heave RAO Floodwater Pitch RAO

Conclusions
Analysisofthedamagedshipisnontrivial (floodwater,sloshing,
coupledfailuremodes)anddecisionmakinginvolvesshorttime
scales.

Whenashipisdamaged,strength isreducedandtheload(localand
global)ischanged.

Recoverabilityatminimumriskrequiresarationalassessmentof both
loadsandstrengthandthecombinationoftheirassessmentto
ascertain:
global structuralintegrity
localstructuralintegrityandthepotentialforfurtherdamage.

Thepresentedworkcandeveloptoolsforassessment andanalysisof
structuralreliabilityandproduce datasetsforvalidationand
comparisonwithothertoolsthatmightusedifferentassumptions
Thank You
The next steps: Investigate the effect of longitudinal damage location on global
vertical loads

AnyQuestions?
WearecontinuouslygratefultoUKMoDforfundingthiswork,Lloyds
RegisterEMEAfortheirsupportandUniversityofSouthampton
fortheircollaboration.
 Simulations of motions of a
damaged ship in regular waves
The damaged ship conference
26th 27th January 2011

A.Gaillard, UCL
Dr. GX Wu, Pr. P Wrobel, UCL
Motivation for work

Damaged ship:
Stability loss
Poor sea-keeping
Structural failure

Damaged ship research project

Structural response tool development

Motions of a damaged ship in regular waves

Content

Background
Main area of research
Methodology
Influence study
Draught
Trim
Heel
Conclusion
Background

Key aspect:
Coupling internal/external flow

Background:
Roll motion leading to capsize
Time domain simulation
Flat internal surface
Comparison intact/damage motions in particular cases
Main area of research
Model:
3D potential flow theory
Internal sloshing
Internal/external flow coupling

Input:
Damage case
Wave frequency and heading

Output:
Hydrodynamic forces and RAOs of motions
Comparison with intact ship motions
Methodology
Methodology

Use of dipole panels:

Equation degeneration
Thin wall approximation

Sources of High density of panels Thin wall approximation:

large strength: dipoles
Equation degeneration
Experimental conditions

Containership
Front, beam and stern quartering waves
Frequency range 0.05 to 2.05 rad/s

Damage conditions:
Damage to hull bottom leading to:
Draught increase
Heeling
Trim Beam waves
Stern quartering
waves
Front waves
Draught influence

Flooding of a compartment aligned with CoG

From 0t to 2000t (leading to a 8 cm sinkage)

Damaged/Intact: Similar trend of RAOs motion

Scenario D Roll RAO Front waves

Draught influence

Depending on wave heading:

In front waves motions of damaged ship are more
important than for intact ship
Draught influence

Depending on wave heading:

In stern quartering waves, motions of damaged
ship are less important than for intact ship
Draught influence

Depending on wave heading:

In front waves motions of damaged ship are more
important than for intact ship
In stern quartering waves, motions of damaged
ship are less important than for intact ship

Heave Pitch Roll

Front Wave I<D I<D I<D

Beam wave I>D I<D I<D

Stern quartering I>D I>D I>D

wave
I : Intact ship, motions magnitude
D : Damaged ship, motions magnitude
Trim influence

Trim: A 2000t flooded compartment is moved forward and

backward of the CoG to create trim.

Damaged/Intact: Similar trend of RAOs motion

Trim influence

Damaged/Intact: Similar trend of RAOs motion

-Bow down trim and stern down trim have opposite
influence
Trim influence

Damaged/Intact: Similar trend of RAOs motion

- In front waves motions of damaged ship are more
important than for intact ship (BDT)
Trim influence

Damaged/Intact: Similar trend of RAOs motion

Bow down trim and stern down trim have opposite
influence
In front waves motions of damaged ship are more
important than for intact ship (BDT)
Roll motions of damaged ship are more important than
for intact ship
Heave Pitch Roll
Front Wave I=D I<D I<D
Beam Wave I>D I>D I<D
Stern Quartering I>D I>D I<D
Wave
I : Intact ship, motions magnitude
D : Damaged ship, motions magnitude
Heel influence

A flooded compartment is moved from the CoG to

starboard.

2 types of trends depending on heeling angle

Scenario H Heave RAO Beam waves

Heel influence

Damaged/Intact: Similar trend of RAOs motion

Roll motions of heeled ship are more important than for
intact ship
Heave Pitch Roll
Front Waves I=D I<D I<D
Beam Waves I>D I=D I<D
Stern Quartering I>D I>D I<D
Waves

I : Intact ship, motions magnitude

D : Damaged ship, motions magnitude

Further research is required to explain differences

in motions due to different angles of heeling.
Intact vs Damaged

In FW, BW and SQW motions of damaged ship

and intact ships are mostly similar in trends and
magnitudes
Intact vs Damaged

In FW, BW and SQW motions of damaged ship

and intact ships are mostly similar in trends and
magnitudes

In FW, heeling creates an important resonant

peak of roll. (To be further studied)
Comparison

Influence of damage depends on wave heading

and motion:
In front wave, damaged ship tend to present higher
peak of motions
Heave Roll Pitch
Front Waves ID I<D I<D
Beam Waves I>D ID N/A
Stern Quartering I<D I<D I>D
Waves
I : Intact ship, motions magnitude
D : Damaged ship, motions magnitude
Conclusion

Modelling of ship under different Damage

scenarios varying sea state and wave heading

Comparison of the RAOs of motions between

intact and damage and over the different scenario

Influence of damages on motions of the ship

depends on wave heading and the motion
considered.
Further work

Further research on the effect of heeling

Run more damage scenarios
Model realistic internal design and flooding on
different hull shape
Incorporate vortex shedding damping
Comparison with experimental data
 Please feel free to ask any questions?

WearecontinuouslygratefultoUKMoDforfundingthis
work,andLloydsRegisterEMEAfortheirsupport
SELECTING THE SAFEST OPTION AFTER DAMAGE A
TOOL FOR DECISION SUPPORT

Andrew Martin
QinetiQ
Damaged Ship Conference 2011
27th January 2011

Copyright QinetiQ Limited 2011

1
Contents

01 The problem
02 The research
03 Procedure
04 Core of the software
05 Example
06 Conclusions

Copyright QinetiQ Limited 2011

2
01 Problem:

You have been damaged 500 miles from port

What are the relative risks of sailing a ship to a port of repair?
Is it reasonable to sail, or should the ship wait for assistance?
What is the safest route back given the weather forecast?
A tool for comparing the risks of different routes by predicting the wave
loading that might be experienced

Copyright QinetiQ Limited 2011

3
02 Research

MoD sponsored program - DALAS

Wave loading
Damaged condition
Structural Capacity
Advice in emergency situations

Copyright QinetiQ Limited 2011

4
03 Procedure

Obtain as much information from the ship as possible concerning extent

of damage and flooding
Determine position of the vessel and possible routes to a port of repair
Go to www.rnein.net for weather forecast
Look up wave height charts for region of voyage and predicted duration
Draw up a voyage plan, noting duration of each leg, sea state and wave
direction relative to the vessel track.
Put all the information in Paramarine
Determine the level of risk that you are willing to run
Examine the results

Copyright QinetiQ Limited 2011

5
Overview Load Strength

Damage is reported to a Paramarine user Damage is reported to a Paramarine user

Paramarine model is modified with Paramarine model is modified with a

flooded compartments damage template approximating reported
extent

Still water load Wave induced load

NS94D module removes damaged structure

Paramarine calc. Paramarine selection of and recalculates ultimate vertical bending
(Deterministic) probabilistic load distribution strength at cuts along the ships length
Stiil water BM at user defined probability.

BM capacity
WI BM
X/L X/L X/L
BM load hog

X/L
Sag

hog
BM

X/L
Sag

Paramarine display

Copyright QinetiQ Limited 2011

6
 Ship response Wave excitation

04 Core A set of pre-determined flooding conditions User picks a suitable sea area
are specified by user (A,AB, CDE etc.) this defines the distribution of wave freq/heights
Ship speed is specified in each sea state.

WASIM (linear) is run in time domain for this Weather forecast for vessel track produced.
condition - with a variety of wave periods Defined as a series of legs each with an
and headings associated sea state, ship speed and relative
heading to waves

Fourier transform of BM response to freq. domain. New module creates a set of sea conditions
Post processing (POSTRESP) to produce RAO (combination of Hs/Tz) within each sea state

Each sea condition is used to generate a wave

energy spectrum (PM)

For every leg

For every sea condition in the leg

BM response spectrum produced

Area under the response spectrum calculated

Used to define a similar Rayleigh distribution

for the probability of exceeding a given BM

From leg duration, ship response frequency

and prob of sea condition; the number of BM
exceeding given values is calculated

Summation of contributions from each condition

Summation of contributions from each leg

Copyright QinetiQ Limited 2011

7
Core -2 Number of BM exceeding given values is calculated for
whole voyageby summing individual values

Exercise is repeated for every X/L cut

1 1 1
P exceed Pexceed Pexc eed

0 0 0
BM BM BM
Cut 3 Cut 2 Cut 1

User selects a level of probability of interest

WI BM at that probability is recorded and plotted

1 1 1
P exce ed Pexceed Pexceed

0 0 0
BM BM BM
Cut 3 Cut 2 Cut 1

BM

Cut 3 Cut 2 Cut 1

X/L

Copyright QinetiQ Limited 2011

8
Core -3
BM

X/L
Linear WI BM response

Apply Non-linear Correction Factors (+/- 1 default) to obtain hog/sag

BM

X/L

Add still water component

Compare load with damaged strength

Copyright QinetiQ Limited 2011

9
Wave height forecast (Rnein.net)

Copyright QinetiQ Limited 2011

10
Assembling a Voyage file

Sea State Sig. Wave ht. (Range)

0 0
1 0-0.1 Voyage Leg Sea state Direction (o) Duration hrs
2 0.1-0.5
3 0.5-1.25 Leg 0-6 5 30 6
4 1.25-2.5
Leg 6-12 5 45 6
5 2.5-4.0
6 4.0-6.0 Leg 12-24 5 45 12
7 6.0-9.0
8 9.0-14.0 Leg 24-36 4 0 12
9 14m+ Leg 36-48 5 0 12
Leg 48-60 3 90 12
Leg 60-72 4 90 12
0 180
Leg 72-84 3 60 12
15 165
Leg 84-96 2 75 12
30 150
135
45
120
60 105
75 90

Copyright QinetiQ Limited 2011

11
05 Example CDE flooding after impact

Copyright QinetiQ Limited 2011

12
Example CDE flooding after impact
1% PExceed (99% P Survival) 0.01% PExceed (99.99% P survival)

Copyright QinetiQ Limited 2011

13
Caution

The answers are only as good as the info in: knowledge of damage is
usually poor.
Not a true probability of failure or survival WI loads are probabilistic
others are deterministic. Best considered as a comparative guide
Simplifications only vBM considered- flood water is considered as a
solid additional mass
Effect of Non-linearity's in maths and hull still under review
However a major advance of previous static wave balance methods

Copyright QinetiQ Limited 2011

14
06 Conclusions

A tool is available that will allow the relative risks of different voyage strategies to be
compared.
Based on the problem of structural failure due to exceeding vertical bending strength

It is based on the established ParamarineTM architecture, already used by UK MoD in

planning for Emergency Response.

This tool has known limitations, many of which can be addressed through continued
research and use.

Copyright QinetiQ Limited 2011

15
The Structural & Stability Assessment &
Subsequent Recovery of a Damaged
Lifeboat

Presentation by:
Antony Haman, Naval Architect
Neil Chaplin CEng MRINA, Staff Officer (Technical)
 Contents

About the RNLI

The incident
Location / Environment
RNLI hull structure design & assessment
Severn class structure & stability
Regulatory considerations
Pollution prevention
Recovery options
Recovery methodology
Lessons learned
 About Us

The RNLI is a charity that saves life at sea

Voluntarily funded
235 lifeboat stations around UK & RoI
4660 Volunteer lifeboat crew
Reach 90% of all casualties within 10 miles in
30 mins
Up to 100 miles offshore
150 seasonal Lifeguard units on beaches
River Thames
Inland waters
Flood Rescue
Over 23,000 people assisted (2009)
 The Incident

Operational service at night

Attempted rescue of small
RIB
Close inshore in shallow
water
 Location / Environment
Rathlin Island
3 miles off North coast of Northern Ireland
EU Special Area of Conservation

Rathlin
Island
 Severn Class Principal Particulars

Length OA 17 m
Beam 5.5 m
Draught 1.8 m
Range 250 nm
Speed 25 knots
Load Displacement 42 tonnes
Crew 7
Construction Fibre Reinforced Composite
Fuel capacity 5,600 litres (Marine Diesel)
RNLI Hull Structure Design & Assessment
Structural load prediction method for lifeboats based on trials
and in-service experience.
Lifeboats instrumented to measure structural response
Global design loads generally exceed classification society
requirements

819 kPa = 126psi

 Severn Class Structural Arrangement
Watertight Bulkheads

S.G. TANK SURVIVOR FORE FORE

ENGINE ROOM
SPACE SPACE SPACE CABIN PEAK

Double bottom in Forward Cabin and Survivors Space

10 separate watertight void spaces outboard of
longitudinal bulkheads
 Severn Class Construction
Thick un-stiffened Epoxy resins
sandwich UD, biaxial and multi-axial E
construction for hull glass and aramid fabrics
topsides (75- High density, closed cell PVC
100mm), deck and foam for core material in
wheelhouse. sandwich construction
Pre-preg outer skin
Longitudinally
stiffened single High strength to weight ratio
skin construction Robust (damage tolerant)
below the chine, Ease of fabrication & low
10mm fabrication time
Long service life and low
maintenance
 All Weather Lifeboat Stability
Inherently self-righting
Capable of withstanding a pitchpole
2 compartment damage stability
Secondary buoyancy

0.5

0.45

0.4

0.35

0.3
GZ(m)

0.25

0.2

0.15

0.1

0.05

0
0 15 30 45 60 75 90 105 120 135 150 165 180

HEEL ANGLE (DEGREES)

 Initial Recovery Attempt
Engineering Office Advice
Attempted to tow vessel off rocks using a tug
Worsening weather
Crew evacuated
Access to site impossible for 2 days
Recovery plan initiated
 Regulatory Considerations

Incident reported to:

Maritime and Coast Guard Agency (MCA)
Marine Accident and Investigation Branch (MAIB)
Northern Ireland Environment and Heritage Department (now
Northern Ireland Department of Environment)
 Pollution Prevention
Ensuring safety and pollution considerations were
identified and acted upon
Environmentally sensitive area
Fuel and hydraulic oil tanks intact
Fuel bowser deployed and all hydrocarbons de-stored to
prevent environmental damage
 Equipment Removal
Aims:
Reduce weight
Salvage valuable equipment
Weight and stability information key to recovery operation
Comprehensive weights and centres record
Removal using:
Manitou crane
Aerial runway between vessel and shore
Next Page
Previous Page

Lifting Cradle

Difficult to guarantee boat

would sustain unsupported lift
through on-site inspections
Designed to support boat
during lifting operations
Local materials & assembly
 Floating Recovery Option
Lifting proved infeasible :
Insufficient water depth
Poor access
Lifting option abandoned in
favour of floating recovery
 Barge Terra Marique
Semi submersible
Floating dry dock

Length OA : 80m
Beam mld : 16.5m
Depth (coaming) : 8.5m
Depth (main deck) : 6m
Deadweight : 2211 tonnes
Draught : 1.6 4.8 m
Cargo draught : 2.5m
On Site Damage Assessment
 Damaged Stability

Compartments categorised as intact / damaged

Permeability assessments difficult
Temporary buoyancy employed (wing voids)
Wing voids provided stability to damaged central main hull
Findings communicated back to Engineering Office
 Damaged Stability

Various flooding scenarios explored

Permeability variations assessed
Envelope of likely stability characteristics produced
Damaged Stability
Recovery Operation
Lifting Off Barge
 Conclusions and Lessons Learned
Recovery successful due to high degree of structural
redundancy, damage tolerance and damaged stability
Create salvage packs and plans for all ALBs
Digestible damaged stability information
Removable weights and centres information
Hydrocarbons list and emergency fuel stripping
arrangement
Essential contacts list
Lifting cradle design
Alternative lifting points, if required
Risk assessments and method statements
Any questions?
A Genetic Algorithms based decision
support tool for stability and structural
viability under damage

DamagedShipConference
26and27.01.2011

Paulo Triunfante Martins

Portuguese Navy OPV program
CINAV Portuguese Navy research center

Victor Lobo
CINAV Portuguese Navy research center
Portuguese Naval Academy (Escola Naval)
Table of contents

Introduction
Stateoftheart
DecisionSupportToolpresentation
GeneticAlgorithmapproach
ExamplesandResults
Conclusions
Futuredevelopment
Introduction

Loadconditionofthevesselinfluences
Stabilityandsurvivability
Manoeuvrability
Operationalperformance

Loadconditioncanbedeterminedbymonitoringdraught
marksandtankscapacity
Thereareoperationsthatrequiretoknowquicklytheload
conditionandtosimulateanewone(e.g.Replenishment)
Responsetodamageshouldtakeintoconsiderationtheships
loadcondition
Responsetodamagedependsonhowcrewreactstodamage
Crewsmustbetrainedandthereshouldbeameantosupport
stressfuldecisions
State of the art
 Bridge officer Decision Support Tool

ConditionMonitoringandload
conditionestimate
Draughtmarks
Tankcapacity
Watertightopeningscontrol

Ifnodamageoccursinregular
intervalsverifies:
Intactstabilitycriteria
Longitudinalstrengthcriteria
Damagestabilityunderfloodingoftwo
continuouswatertightcompartments
Directionalstability
Operationalperformance
 DST Response to damage

1st step determine:

Loadconditionbeforedamage
Waterinlet
Floodspread
2nd step generatepossible
solutions:
Floodballasttanksandvoids
Changefuelbetweentanks
others

3rd step verifyfeasibilityandda

magecriteriaforcandidatesolutions
4thstep drainageequipment
attributiontoconstraintflood
5thstep evaluatethecandidates
Howlongtocarryoutthesolution
Howfarthecalculationsarefrom
criterialimits
 Optimization model

Inputvariables
(vesselscharacteristics)
dimensions
Decisionvariables
light ship
displacement tanks how filled each tank is

load condition how flooded each compartment is

cross curves
solid weights
watertight
hidrostatics compartments

Constraints
Costfunction tanks capacity limits (0 98%)
fluid discharge constraints (fuel, oil, etc)

f1 (stability quantification) +
f2 (time to ballast/ transfer) +
f3 (watertight drainage time )
cross flooding pipes
impossibility to flood some tanks/
watertight compartments
drainage capacity
damage stability criteria
 Genetic algorithms implementation (1)

i. Problemformulation chromosomecreationfromloadcondition
[10%0%60%....0% 98% 98%70% ]
eachgenestandsforhowfilledeachtankis
ii. InitialPopulationisformedbyfillingupballasttanksrandomlytofind
candidatesolution
[98% 0%...60%....0% ...98% ...98%70% ]
[10%98% ...60%....0% ...98% ...98%70% ]
...
[10%0%...98% ....0% ...98% ...98%70% ]

iii. Checkifthepurposedsolutionverifiesall
constraints Verifies
evaluation
constraints
= No
bigpenalty
Yes

evaluation
 Genetic algorithms implementation (2)

iv. Evaluationisdoneusingafitnessfunction(costfunction):
verificationofdamagestabilitycriteria
timetoimplementthecandidatesolution

v. Afterevaluationanewsetofparents(newcandidatesolutions)is
producedbygeneticoperators:
elitism selectionofsomeofthebest
crossover changeofchromosomesbetweenparents
[98% 0%...60%....0% ...98% ...98%70% ] [10%98% ...60%....0% ...98% ...98%70% ]

[98% 98% ...60%....0% ...98% ...98%70% ]

mutation abnormalchangeofsomegenes

[ 15% 0% ... 20% .... 25% ... 50% ... 98% 70% ]

iv. Eachnewchildisgoingtobecheckagainstconstraintsandthen
evaluated
 Genetic algorithms implementation (3)

1st Initialpopulationofcandidatesolutions
2nd Constraintscheck andEvaluation byfitnessfunction
3rd Selection forreproduction(elitism/crossover/
mutation)
4th Evaluation ofchildren
5th Newpopulation forreproduction

STOPaftersomegenerations
 Software Implementation Users interface
MainCharacteristics
Nousersinterfaceinnormaloperation
Dataisupdatedinregularintervals
Dataisexchangedthroughconnectionswiththeshipsautomation
systemsusingOPCstandards
Criteriaisverifiedaftereachupdated(intactanddamagestability)
[stabilityroutineswerevalidatedagainstAutohydro calculations]
Shipscharacteristics
andloadcondition

Maneuvringand Weightdistribution
operationinformation

Hydrostaticdata

Alarms
 Software Implementation Simulation capability
Implementedcapabilities
Simulationusesthesameroutinesasthemainsoftwarebutno
variablesexchangeispossible
Tankcapacityisdoneinthefiguretohelpusersinterface
Solidcargocanbeintroducedseparately
Criteriaisverifiedafterdatainput(intactanddamagestability)

Shipscharacteristics
Tank andloadcondition
capacity
input
Maneuvring
andoperation
information

Solidcargo
input

Alarms
 Software Implementation Damage response
Implementedcapabilities
Usermustidentifydamagedarea,dimensions andverticalposition
Softwarecapturesloadcondition ,findsfloodspreadanddetermines
waterinletflow
Geneticalgorithm iscalledto:
Bringtogetherseveralcandidatetankfluiddistribution
Decideifthedamagecompartmentcanbesaved
Checkcandidatesolutionagainstdamagecriteriaandotherconstraints
Evaluatecandidatesolutions

Damagedimension
andverticalposition

Damagedarea
identification Proposedsolution
 Software Implementation Optimization process
Geneticalgorithmsprocessandevaluation
Beginstogather20candidatesolutions developedfromload
conditionandchangingballasttanksfloodcapacity
Firstresponseispresentedafter3generations
(80calculations 1minute)
Definiteresponseisobtainedsometimeafter(50generations),
repeatingGAs replacingthedamageloadconditionwiththebest
resultfromthe3generationscalculation
 Conclusions

Conceptual
9 Itisimportanttomonitorloadcondition;
9 Responsetodamagedependsupontheloadcondition;

Developedwork
9 Responsetodamagewasformulatedasanoptimization
problem;
9 Geneticalgorithmswereusedtosolvetheoptimization
problem;
9 AsoftwareDSTwasdevelopedandvalidated.
 RINA DAMAGED SHIP CONFERENCE, JANUARY 2011

DAMAGE DECISION BOOKS FOR USE BY MASTERS ON

VESSELS WITH NO COMPUTER BASED LOADING
SYSTEM

hf@hart-fenton.com
Who we are: Consulting Naval Architects and Marine Engineers

Hart, Fenton & Company brings a successful

track record gained over fifty years service to the
marine industry.

As the specialist marine design subsidiary of

Houlder Limited, Hart Fenton & Company can
also utilise the expertise and resources of one of
the UKs largest marine engineering
consultancies.

Houlder provide Naval Architecture and Marine

Engineering services to the Defence, Marine, Oil
& Gas and Renewable Energy sectors.

67 people over 5 sites work to enhance the

capability of maritime assets through the concept
development, design, procurement,
commissioning and through life support of
vessels, associated equipment and systems.

hf@hart-fenton.com
What we do: Breadth of expertise
Concept design
Testing
Procurement
Surveys
Construction supervision
Conversion supervision
Structural design
Damage stability analysis
Systems design
Project Management
Commissioning
In-service support
Expert witness

hf@hart-fenton.com
Introduction

Accidents happen - vessels get damaged

Damage severity dictates the masters response
Decision requires accurate and relevant information
Number of options for providing the information

hf@hart-fenton.com
Available Options
Option Advantages
Loading Computer Multiple variables easily handled.
Information quickly available to master
What if analysis can be performed
Remote tank interrogation
Disadvantages
Users need to be familiar with system
Training in complex software.
Over-reliance on computers
Electronic system water, fire, supply
issues.

Emergency Response Detailed what if analysis Can be costly

Service No Software to learn
No failure due to fire, flood, supply.
Expert assistance available
Response time from support team
Support team vessel knowledge
Delay due to information relay
Accurate transfer of information to support
team required.

Damage Decision Support Simple to use Greater variables cause increased

Books Information quickly available to master complexity of books.
Reliable Not possible to cover all loading
More suitable for limited variables conditions.

hf@hart-fenton.com
Available Options
Option Advantages Disadvantages

Damage Decision Support Simple to use Greater variables cause increased

Books Information quickly available to master complexity of books.
Reliable Not possible to cover all loading
More suitable for limited variables conditions.

hf@hart-fenton.com
Decision Making Process

Ascertain Current Situation

What has happened
What is damaged
Is there water ingress
How quickly?

Understand Implications
Will vessel survive
What if situation changes
Can situation be improved
Can progress to safe haven

Make Informed Decision

Progress under own power
Undertake remedial actions
Call for outside assistance
Mayday & Abandon ship

hf@hart-fenton.com
Decision Making Process

Understand Implications
Will vessel survive
What if situation changes
Can situation be improved
Can progress to safe haven

hf@hart-fenton.com
Damage Stability Assessment

Damage Stability Books

Statutory requirements
Range of conditions
Not all eventualities
Vessel is safe to operate in the expected operating range

Probabilistic Damage Analysis

Covers many more damage scenarios
Broad range of conditions
But Not all eventualities
Output can appear meaningless

hf@hart-fenton.com
Decision Making Process

Typical Probabilistic Output

hf@hart-fenton.com
Decision Making Process

Typical Probabilistic Output

hf@hart-fenton.com
Damage Stability Assessment

Damage Decision Support Book

Fills in gaps not covered by statutory documents

Include unexpected scenarios;

Non Adjacent spaces

Range of conditions

Not all eventualities can be covered

Potentially have a large quantity of conditions to analyse

hf@hart-fenton.com
Methods

Initial Conditions
Too many document becomes unwieldy
Work to date have used 3 full, half and light loads.
Agreed with client/operator in advance

Damage Cases
Analyse as many as practical to identify limit of survivability
Non-adjacent spaces
Asymmetric flooding
Multiple compartments 3, 4, 5?

Analysis
Using standard Naval Architecture software
Not applied heeling levers
Not applied shift of cargo

hf@hart-fenton.com
Presentation & Examples

Presentation is key to successful use of decision support books

Must be;
Relevant
Accurate
Clear
Concise

Non-survivable conditions should be included

Agreed with client/operator in advance

Hart Fenton have used a colour coded matrix at start of book

Includes all analysed conditions
At a glance indication of severity

hf@hart-fenton.com
Presentation & Examples
Damage Matrix
Starboard Compartments
Page Condition Jet Room Engine Room Void 5 Void 4 Void 3 Void 2 Void 1 Fore Peak Jet Room Engine Room
5 1X
6 2 X

7 3 X

8 4 X

9 5 X

10 6 X

11 7 X

12 8 X

13 9X X

14 10 X X

15 11 X X

16 12 X X

17 13 X X

18 14 X X

19 15 X X

20 16 X X X

21 17 X X X

22 18 X X X

23 19 X X X

24 20 X X X

25 21 X X X

26 22 X X X X

27 23 X X X X

28 24 X X X X

29 25 X X X X

30 26 X X X X

31 27 X X X X X

32 28 X X X X X

33 29 X X X X X

34 30 X X X X X

35 31 X X X X X X

32 X X X X X X

36 33 X X X X X X

34 X X X X X X X

35 X X X X X X X

36 X X X X X X X X

37 37 X X

38 38 X X X

39 39 X X X X

40 40 X X X X X

hf@hart-fenton.com
 Half Load Damage Matrix
Page Condition Aft Peak Aft Voith Space No 8 Void Aft MC Space No 7 Void No 6 Void No 3 Ctr Void Aux MC Space No 5 Void No 4 Void No 2 Ctr Void
93 1x
94 2 x

Single Compartment Damage

95 3 x
96 4 x
97 5 x
98 6 x
99 7 x
100 8 x
101 9 x
102 10 x
103 11 x
104 12
105 13
106 14
107 15
108 16
109 17
110 18
111 19 x x
112 20 x x
113 21 x x
114 22 x x
115 23 x x
116 24 x x
117 25 x x
118 26 x x
119 27 x x
Two Compartment Damage

120 28 x x
121 29 x x
122 30 x x
123 31 x x
125 32 x x
126 33 x x
127 34 x x
128 35 x
129 36 x x
130 37 x x
131 38 x x
132 39 x
133 40
134 41
135 42
136 43
137 44
138 45
139 46
140 47
141 48
142 49
143 50 x x x
144 51 x x x
145 52 x x x
146 53 x x x
147 54 x x x
148 55 x x x
149 56 x x
150 57 x x x
151 58 x x x
152 59 x x x
153 60 x x x
Three Compartment Damage

154 61 x x x
155 62 x x x
156 63 x x x
157 64 x x x
158 65 x x x
159 66 x x x
160 67 x x x
161 68 x x x
162 69 x x x
163 70 x x x
164 71 x x x
165 72 x x
166 73 x
167 74 x
168 75 x
169 76
170 77
171 78
172 79
173 80
174 81
175 82
176 83
177 84
178 85

hf@hart-fenton.com
Presentation & Examples

For each survivable condition further information is presented such

as;
Floating position

Damaged Compartments

Critical Points (downflooding, embarkation, etc.)

Stability Information

hf@hart-fenton.com
Presentation & Examples

hf@hart-fenton.com
Relevant Information Required

From presented information Master can see;

Whether the vessel will survive

What angle of heel and trim will occur

Is there any likelihood of further flooding

Is there enough residual stability

Can something be done to improve the situation

They dont need to know

That the vessel doesnt meet stability criteria

hf@hart-fenton.com
Now for real

SEA EXPRESS - February 2007

274 Passengers 57 Cars
Four Compartments Damaged
Salvage company praise accuracy of HF information

hf@hart-fenton.com
Further Development & Refinement

Multiple Cargo Types

Remedial Actions eg cross flooding

Intermediate & Progressive Flooding

Weather and other Heeling Levers

Improved Formatting

Incorporate Probabilistic analysis

hf@hart-fenton.com
Conclusions

Ideal world damage decision books will not be needed

The advantages they offer:-

They are immediately available

They are simple to use

They contain the relevant information

Can be used as first reference prior to other systems being

available

There is no input requirement at a highly stressed time

hf@hart-fenton.com
 Thank you for your time
Hart, Fenton & Company

hf@hart-fenton.com
 RINA Conference
London, UK
27 January 2011

Expanding the Scope of RRDA:

A Classification Society
Perspective

GE (George) Wang
Manager, Advanced Analysis
ABS
1
Outline

z ABS service of Rapid Response Damage

Assessment (RRDA)
z Recent developments and ABS enhancements
z Technical aspects of RRDA enhancements
z Summary

2
ABS RRDA Program

z Organized to support marine and offshore operators

during emergency incidents that may affect stability
or structural strength
Assist ships and floating facilities in distress with
prompt access to engineering personnel trained in the
use of HECSALV (salvage response software) and
other electronic programs
Meets or exceeds the
requirement of US and
international regulation
Offered to ABS and
non-ABS vessels and
vessel types

3
Inventory

z More then 1,800 units enrolled in the RRDA

4
Overview of Enhancements

z Guide for Rapid Response Damage

Assessment, July 2010
RRDA class notation

z Enrollment free to tankers, bulk

carriers, gas carriers, container vessels
and tank barges
z Centralized service center in ABS
headquarters in Houston
24/7, 365 day response service
Full time RRDA engineers supported
by SED and Technology engineers

5
Technical Aspects of RRDA

z Traditionally, the analysis of RRDA has focused on the vessels

trim and stability. HECSALV software is used.
z With the ABS enhancement, the engineering analysis is
enhanced to include evaluation of loads and strength at the time
of an incident and during transit voyage to a repair facility
Site/route environmental loads (ABS SEAS program)
Hull girder ultimate strength
Local loads, strength
Local buckling strength
Software
z Future enhancements are also
planned to incorporate R&D
achievements

6
Geometry/Trim/Stability

z Geometry/hull intact stability

7
Geometry/Trim/Stability

z Compartments damage stability and loading

8
z Development of computer software has rendered feasible
the assessment of the damaged condition in a:
More holistic manner
Timely fashion

z Established in-house information and data management

systems have made it possible to access the data of
design and survey that are needed in an incident
response.

z Enhanced ABS RRDA will include the capability to assess

the ability of the ship to proceed to a repair facility

z Other improvements are planned for the longer term

9
Enhancements to RRDA
Phase Service Existing/Additional
Trim and stability in damaged condition E
HG longitudinal strength in damaged condition (residual
strength) stillwater E

Initial Condition HG longitudinal strength in damaged condition with wave E

HG ultimate strength A
Local strength A
Local buckling and ultimate strength A
Review cargo transfer sequence E
Review extent of temporary repairs E
Trim and stability in temporary repaired condition E
Transit and
HG longitudinal strength in temporary repaired condition
Temporary
(residual strength) stillwater E
Repair
* General HG longitudinal strength in temporary repaired condition with
Environmental wave E
Conditions of
Route HG ultimate strength A
Local strength A
Local buckling and ultimate strength A
Other analysis as needed A

10
Enhanced Analysis Incident Stabilized

Plan for transit

Evaluate vessel condition
z Assist in making Structural damage
Tanks breached
Temp repair
Transfer cargo

decision with
Weather routing
Previous condition
Port of refuge

respect to
Stabilizing Potential loss of
Stability
Yes
Recommend "No Sail"

an incident HECSALV No
Hull Girger Strength Analysis
Assessing Potential loss of
Residual hull girder ultimate
strength
whether the hull strength
Environmental loads at site or
along route
damaged vessel
can proceed for
Local Strength Calculation
a transit voyage Local pressure No Potential breakup of
hull
Buckling and ultimate strength of
local members

z Perform analyses Crack propagation

Yes

of loads and Other

Options
Recommend "No Sail"

strength in Potential loss of

containment
No

Recommend "Sail"
additional to the Yes

analysis of trim Change in

conditions,
and stability Yes
situation, or
plan
No
Recommend "No Sail"

11
 Analysis: Trim/Stability, Loads & Strength
Design strcutural Model
GHS or HECSALV Model
Scantlings
Hull forms
Design loads
Compartment
Structural response
Lightship
Class permissible

RRDA Trim/Stability Model

RRDA Structural Model
Hull forms
3 transverse sections
RRDA Tanks
Class permissible
Weight distribution
enrollment

Structural
damage
Loading
conditions
Residual hull
girder (ultimate)
strength
Breached
tanks
Vessel
location &
route
Local Strength
Plan for local buckling str
transit
ABS SEAS
Response program

analysis at the
time of Actual
Trim & Dynamic
incident and Stability
SWBM Wave loads loads
SWSF
preparation
for a transit
RRDA load and strength
cal program
Global Local
Stability
Strength Strength

Suitability evaluation for transit

(reporting)
ACE/ACS decision about Stability
fit-to-proceed Global Strength
Local Strength & local buckling str

12
Enhanced RRDA

Software
z ABS is developing new software to assist the enhanced
RRDA program
z Load and strength analyses will include:
Hull girder ultimate strength
Local strength
Local buckling strength

13
Enhanced RRDA: Site/Route Loads

Sea Environment Assessment System (SEAS)

Environmental severity factors
for dynamic loads

Seakeeping Analysis

vs

Site, route and transit North Atlantic

environments environment

Environmental Effects

14
Enhanced RRDA: Site/Route Loads

SEAS: ESF -Factor Calculation Procedure

RAO2 Wave Spectrum Response PDF for
Spectrum Extreme
Values

System Input Output

All weather groups and wave directions: long-term (e.g., 100-year return period):

Most probable extreme values at

& Site and Route (Ls) and North Atlantic (Lu)

Wave (route) L
= = S
Wave (NorthAtlantic) L (IACS rule)
Wave Scatter Diagram Wave Directionality (Rosette) U
for Site or Route

15
Enhanced RRDA: Site/Route Loads

SEAS: Vertical Wave-Induced Bending Moment

VBM
M ws = k1 VBM C1 L2 B (Cb + 0.7) 10 3 Sagging Moment
M wh = k 2 VBM C1 L2 BCb 10 3 Hogging Moment
1.0

0
0.0 0.4 0.65 1.0
Aft Forward
end of L Distance from the aft end of L in terms of L end of L

16
Enhanced RRDA

Residual Global Strength

z Hull girder ultimate strength
Methodology and acceptance
criteria IACS CSR
Inputs
SW from HECSALV
Site/route-specific loads from
SEAS
HGUS IACS approach

s Msw + w Mw < Mu/R

17
Enhanced RRDA

Local Strength
z Analysis of local loads and strength help evaluate the risk
of tank breaching during the transit voyage
z Local loads and strength
Static loads based on actual vessel loading condition
Site/route-specific environmental
loading (CSR/ SEAS program)
Simple beam theory
Cargo Oil

Intact condition Damaged condition

18
Enhanced RRDA

Local Buckling Strength

z Analysis of local buckling strength help assess the
possibility of worsening of structural damages during the
transit voyage
z Local loads and strength
Static loads based on actual vessel loading condition
Site/route-specific environmental loading (CSR/SEAS
program)
IACS CSR buckling/ultimate strength (uni-axial buckling for
plates, column and torsional buckling for stiffeners)

19
Long-term Considerations

z Integrating RRDA with weather routing service?

z Adding analysis of crack propagation?

21
Back to Paper
Summary

z ABS Rapid Response Damage Assessment service

z Recent development and ABS initiatives of
enhancement
z Technical aspects of the enhanced RRDA program

22
www.eagle.org

23
Salvage & Marine Operations PT
Ministry Of Defence

The Damaged Ship -

MOD Salvage Response to HMS ENDURANCE in
the Magellan Straits in December 2008

James Ward MBE

Aims & Objectives

Overview of the MoDs Salvage & Marine Operations

Project Team
History of S&MO
S&MO capability
Recent S&MO activity
Case study - MOD Salvage Response to HMS
ENDURANCE in the Magellan Straits in December 2008
Questions

27 Jan 2011 RINA - The damaged ship 3

Salvage & Marine Operations PT- History -
HMS MONTAGUE 1906

27 Jan 2011 RINA - The damaged ship 4

Salvage & Marine Operations PT- Current Capability

24 hour salvage assistance

Approx 30 operational
personnel + equipment
Air portable/difficult to
source equipment
Majority Sponsored
Reserves
ISO 9001/2000
Commercial sector reliance
Regular exercises with
customers Falcon ROV

27 Jan 2011 RINA - The damaged ship 5

Salvage & Marine Operations PT- Future Capability

Comanche ROV due in

service Jan 2011

27 Jan 2011 RINA - The damaged ship 6

V1.1 dated 29 Jul/DTLSO
Salvage & Marine Operations PT Recent activity

Weld repairs to HMS

NOTTINGHAM MV DUBAI MOON

27 Jan 2011 RINA - The damaged ship 7

Salvage & Marine Operations PT HMS ASTUTE

27 Jan 2011 RINA - The damaged ship 8

HMS ENDURANCE Dec 08

This case study will cover

the following points
The operational activity
undertaken to salve HMS
ENDURANCE
The key factors that
aided success
The key factors that
hindered success
Aftermath

27 Jan 2011 RINA - The damaged ship 9

HMS ENDURANCE Dec 08 Incident location

Chartlet of South America

Area of
incident

27 Jan 2011 RINA - The damaged ship 10

HMS ENDURANCE Dec 08 Incident location

Line
of Drift

27 Jan 2011 RINA - The damaged ship 11

HMS ENDURANCE Dec 08 Initial Incident
Response
Background
Transiting the W Magellan
Straits during the afternoon Chart let of South West Chile
of 16 Dec 08 Initial Incident
location 16/1900Z
Suffered catastrophic major
flood in ER Repair Berth ASMAR Ship
yard PUNTA ARENAS
Lost all propulsion and most
services
WT doors were closed
Embarked ships diver made Shoal where
a number of attempts to ENDURANCE
fetched up
stop the leak 17/0245Z

Drifted and snagged on the

Approximate route of
ENDURANCE under tow

only shoal patch in the area

Towed by chartered tugs
(TOWCON) to Punta Arenas

27 Jan 2011 RINA - The damaged ship 12

HMS ENDURANCE Dec 08 Initial Response

Passengers including some

from the British Schools
Exploring Society
evacuated by Chilean Navy
helicopter as a precaution.

The Chilean Navy tug Pudu,

a helicopter and the cruise
ship Norwegian Sun headed
to stand by the HMS
Endurance.

27 Jan 2011 RINA - The damaged ship 13

HMS ENDURANCE Dec 08 Initial Response

Tugs arrived following

afternoon
Simple tow rigged
through forward lead
bull ring and on to
forward bollards
Both anchor cables cut
using portable gas
cutting set

27 Jan 2011 RINA - The damaged ship 14

HMS ENDURANCE Dec 08 Initial Response

13 man S&MO team with air

portable salvage equipment
along with personnel from
the operator and members
of the ships company (on
leave) flew to the Falkland
Islands.
Quick turnaround to Punta
Arenas via RAF Hercules

27 Jan 2011 RINA - The damaged ship 15

HMS ENDURANCE Dec 08 Initial Response

Arrival PA
Team arrived evening
18 Dec shortly before the
Endurance arrived under
tow.
Small advance team flew
out in RN Lynx for a
situation update.

27 Jan 2011 RINA - The damaged ship 16

HMS ENDURANCE Dec 08 Recovery Activity

Once alongside
Electrically driven pumps
provided by tugs
Additional support craft
chartered in support of
operation
Separate charter for the
recovery of anchors &
cables

27 Jan 2011 RINA - The damaged ship 17

HMS ENDURANCE Dec 08 Recovery Activity

Flood containment
Leaking decks
Migration through
bulkhead glands
Migration through grey
and brown water
sewerage systems.
Constant monitoring

27 Jan 2011 RINA - The damaged ship 18

HMS ENDURANCE Dec 08 Recovery Activity

Aims
To contain flooding and
restore watertight
integrity
To prepare vessel for tow
to the Falkland Islands
Preparations
Access to the flooded ER
Local hire of pumps
Rigging pumps
Pre-use equipment
checks
Familiarisation

27 Jan 2011 RINA - The damaged ship 19

HMS ENDURANCE Dec 08 Recovery Activity

External Patching/Blanking
Initial use of MIKO
magnetic patches
Later use of fabricated
ships blanks and wooden
bungs
Due to uncertainty of
valve line ups all hull
opens had to be suction
tested individually

27 Jan 2011 RINA - The damaged ship 20

HMS ENDURANCE Dec 08 Recovery Activity

Initial Pumping- Entering the

flooded compartments
Single access for pumps and
personnel through funnel
and uptake casing
When water level reduced
WT doors reopened to allow
improved access
After initial gains
uncontrollable reflooding
occurred
Personnel evacuated from
ER and C Deck
WT doors reclosed

27 Jan 2011 RINA - The damaged ship 21

HMS ENDURANCE Dec 08 Recovery Activity

The second entry

Cause of re flooding
found
Larger hole in funnel
space to aid entry into
the ER
Watertight doors on C
Deck remained closed
until effectiveness of hull
blanking proven

27 Jan 2011 RINA - The damaged ship 22

HMS ENDURANCE Dec 08 Recovery Activity

Pumping the ER
Electric salvage pumps
rigged through hatch in
engine room workshop
Clean water pumped
directly over-side
Once at bottom plates
oily water pumped into
sullage barge

27 Jan 2011 RINA - The damaged ship 23

HMS ENDURANCE Dec 08 Recovery Activity

Access to the ER
Physical access to ER
restricted by debris

27 Jan 2011 RINA - The damaged ship 24

HMS ENDURANCE Dec 08 The Cause

The strainer where the original

flood started
The causes of the flood
A known poor valve design
Required that pneumatic
actuator pipes to be
removed
Single valve isolation
No method of manually
rapidly closing the valve
Human factors

27 Jan 2011 RINA - The damaged ship 25

HMS ENDURANCE Dec 08 - Damage

Damage
Extensive damage to C
deck cabins
Extensive damage to
loose ER/ER workshop
fittings

27 Jan 2011 RINA - The damaged ship 26

HMS ENDURANCE Dec 08 Tow Preparations

Rigging the tow

Lots of innovation and
adaption greatly added by
the recovery of the anchors
Forward windlass powered
from a shore supply
Secured END for sea
Remaining END company
Removed debris from C
deck to prevent
spontaneous combustion
Disposed of freezer/fridge
contents
Secured END for sea

27 Jan 2011 RINA - The damaged ship 27

HMS ENDURANCE Dec 08 - Tow

Tow to the FI over New Year

08/09
South Atlantic tow
3 day passage
Un manned tow
2 Tugs

27 Jan 2011 RINA - The damaged ship 28

HMS ENDURANCE Dec 08 - CSFs

Critical success factors

A highly motivated and
professional team
Good working relations
with Ships staff and
other deployed personnel
Good support from UK
S&MO IPT personnel

27 Jan 2011 RINA - The damaged ship 29

HMS ENDURANCE Dec 08 - CSFs

Critical hindrance factors

Long logistics chain
Time of year
Language
Poor quality hired
salvage kit

27 Jan 2011 RINA - The damaged ship 30

HMS ENDURANCE Dec 08/Apr 09 - Aftermath

On arrival at the FI
Ships company returned
Survey work undertaken
C Deck bulkheads
removed
Heavy lift preparations
MV Target chartered in
Cape Town
Proceeded to Rio De
Janeiro for heavy lift
structure to be fitted and
then to the FI
END loaded on MV
Target Feb 09
END arrived Portsmouth
Easter 09

27 Jan 2011 RINA - The damaged ship 31

HMS ENDURANCE Dec 08

Questions

27 Jan 2011 RINA - The damaged ship 32

 Tolerable Safety of
Damaged Naval Ships

Mr S Marshall CEng MRINA RCNC,

Ship Stability Safety Regulator, Ministry of Defence (UK)

1
Tolerable Safety of Damaged Naval Ships

1. Background
2. Distinguishing between safety & capability
3. Options for a naval safety standard
4. Structural loss templates
5. Naval Ship Code

2
Range of Naval Combatants and Auxiliaries

Co
Na m
va m
l er
St cia
an lS
d ar ta
ds nd
ar
d s

3
Capability to Survive Hostile Damage

USS Mansfield HMS Glamorgan

Capability:
The ability of a ship to survive a hostile
event, referred to as hostile damage.

4
Surviving Accidents

Safety:
The ability of a ship performing
naval duties to survive a
foreseeable event similar to those
experienced by merchant ships,
referred to as accidental damage.

5
Safety & Capability Combat
Capable

Capability loss minimised

Risk of Damage ALARP

Consequences ALARP
Recoverability

Vulnerability
Susceptibly
Capability

Military safety
Hazard from military duties ALARP
Safety

Foundational safety
Hazards from environment ALARP
6
 Through Life Survivability

Design & Build

Survivability Margin

In-service

Upgrade
Life extension

Naval Standard
?

Years

7
 Through Life Survivability

Design & Build

Survivability Margin

In-service

Upgrade
Life extension

Naval Standard
ALARP?

Years

8
 Through Life Survivability

Design & Build

Survivability Margin

In-service

Upgrade
Life extension

Hostile Capability
ALARP?
Safety Standard

Years

9
Adopting IMO Conventions

SOLAS 2009 PAX

SOLAS: Being desirous of promoting safety of life at sea .. to ensure that,

from the point of view of safety of life, a ship is fit for the service for which it is
intended

Key Hazard: a significant danger to the lives of several people, loss or severe
damage to the platform or significant damage to the environment

10
SOLAS 2009 v Traditional Naval Approach
Ships between 30m and 92m.
Any two adjacent main compartments. A 'main
compartment' is to have a minimum length of 6m.

Ships greater than 92m.

Damage anywhere along its length, extending 15% of
the waterline length, or 21m whichever is greater.

NAVAL

11
SOLAS 2009 v Traditional Naval Approach

SOLAS

NAVAL

12
Examining Accidental Damage Data
1.0
Non Dimensional Damage Penetration

0.8

0.6

0.4

0.2

0.0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
Non Dimensional Damage Length

13
Formal Safety Assessment

Assume if
Likelihood of damage >
Accident SOLAS, ship Cost of Loss
is lost

Reduced risk
Net Cost
of loss
Benefit

Improve
survivability Additional
e.g. Cost
subdivision

14
 Cost Benefit Analysis

20X

SOLAS Hostile Capability

Cost Benefit

10X

4X

2X
X

Damage Extent
15
Structural Loss Templates

Collision Raking
Damage Penetration
10% loss
B B

10% loss
Shell Plating
50% loss 100% loss

A C A Damage Length

75 %
Damage
Penetration

50% Damage Width

100% loss
Damage Width

50% Damage Length

Damage Length

Smith, D Accidental Damage Templates (ADTs), A basis for the future of Naval Ship Safety Certification
10th International Conference on Stability of Ships and Ocean Vehicles
16
Structural Loss Templates

Safety
Accidental
Damage
Extents
Traditional Naval
approach Capability
Hostile
Damage
Extents

17
Structural Loss Templates - Collision

18
Structural Loss Templates - Collision

19
Structural Loss Templates - Raking

20
Structural Loss Templates Hostile Above Water

21
 Next Page

Structural Loss Templates Hostile Under Water

22
Previous Page

Structural Loss Templates

Safety
Accidental To be
Damage decided
Extents
Naval Damage
Traditional Naval Capability
Extents using
approach
Damage Templates Hostile
Damage
Pilot Exercise Extents

Essential
Vulnerability
Characteristics

23
Naval Ship Code

ANEP 77 - Aim

The overall aim of the Naval Ship Code is to provide a framework for
a naval surface ship safety management system based on and
benchmarked against IMO conventions and resolutions that
embraces the majority of ships operated by Navies.

- Initially created by NATO

- Now maintained by the International Naval Safety Association
- Members include 11 navies and 7 naval classification societies.

www.nato.int/docu/stanag/anep77/ANEP-77.pdf

24
Naval Ship Code

0 Aim 0 Aim

1 Goals 1 Goals
1 2 3 4 5 6 7 8 9

2 Functional Areas 2 Functional Areas

3 Requirements 3 Requirements

Cl

Na
as

v
al
s
4 Verification 4 Verification

5 Justification 5 Justification

IMO (SOLAS/HSC) NATO NSC

25
Safety & Capability Combat
Capable

Capability loss minimised

Risk of Damage ALARP

Consequences ALARP
Recoverability

Vulnerability
Susceptibly
Capability

Military safety Naval Ship Code:

Hazard from military duties ALARP ANEP77
Safety

Foundational safety
Hazards from environment ALARP
26
Questions ?

William Shakespeare
"I have a kind of alacrity in sinking." 27

Merry Wives of Windsor (III, v, 13-14)

Bulkheads are the best thing !

Endurance pictures

28
Bulkheads are the best thing !

Endurance pictures

29
Bulkheads are the best thing !

Endurance pictures

30
Bulkheads are the best thing !

Endurance pictures

31
Approaches for determining safety level
95th
1.0

0.8

Point of diminishing returns

0.6
Percentile

0.4

0.2

0.0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
Non Dimensional Damage Length

HARDER Collision Statistics 32

Merchant Ship Damage

Insert XL chart

33