he Court of Inquiry, set up to investigate the
eemons why 192 lives were lost aboard the United
.;jngdom-registered
RoRo passenger ferry
off
:eebrugge
Harbour.
madea
number
of
.ecommendatiefla concerning research into
reeidual
;tability standards for such ferries.
This paper
:utlines the research which was
commissioned by
:he Department of Transport (Marine Directorate).
:n particular.. the research relates to two series
:f 'damaged' model tests carried out in waves of
:aryir.g significant wave height. with the
models
:eam-Ofl to on-coming waves.
The intention was to
jetermifle critical zones where capsize or
non-:apsize was equally probable.
t
the
sametime,
statical calculations were
erforned by computer at the appropriate
ship
:onditiort (determined by draught. trim and KG for
the damaged condition).
In this manner, a link
as
madebetween
the
residual
stability
as
required
by
the
regulations
and
the
dynamic
situation indicated by the test results.
Th.
set of regulations used to make
this comparison are those which came into force
n 29 April this year for new passenger
ships-(Sometimes referred to am the SOL.AS '90
residual
stability standards).
An important conclusion
from
an
examination of
the
results
of
this
research is that in order to provide
reasonable
protection against capsize. assuming side
damage
occurs in the most critical region of the ship.
the
residual
stability standard should be
at
least that of SOLAS
90.
The corollary to this is that Ro Ro ferries
built
before
April
1990
are
unlikely
to
possess
adequate residual stability standards, except in
well-nigh still water conditions.
The
IMO members have been
informed
of
these
important findings:
the topic concerning the
application of these admittedly higher standards
to existing passenger Re Ro ferries is to be
discussed at the relevant meetings next year
-the Sub-Committee Ofl Subdivision. Loadlines and
Fishing Vessels
Safety
in
February
and
the
Maritime Safety Committee in May.
Further research is shortly to be commissioned
by
t?r
epartment of Transport.
This second phase
o
esearch
will
deal
with
the
degree
of
enhincement in survivability that the fitting of
various
devices
or
design
modifications
to
present designs might provide.
Up
until
now,
this
improvement
in
residual
stability was measured by a series of statical
calculations.
Model tests are needed to verify
that
the
variOus
devices would, provide
this
improvement
in
a
seaway
(or
indeed,
whether
dynamically the improvement may be even better
THE ASSESSMENT OF
DAMAGED STABILITY cki'rIku.A
USING MODEL TESTS
Alan Graham'
than
that
indicated
by
the
purely
statical
calculations).
It
is hoped that tentative results should be
available - in time for the IMO discussions next
year.
irr000crI
It is now more than three years since the
"Herald
of Free Enterprise
capsized outside Zeebrugge
Harbour. with heavy loss of life.
The prime
cause of, thi-s tragic event was, as is stated in
the Court of Inquiry report (1). that the
vessel
went to see with both her inner and outer
bowdoors open.
As a result, considerable quantities
of flood water accumulated on the vehicle deck
which caused the vessel to heel very quickly to a
significant angle.
resulting
in
a
very rapid
capsize.
Hopefully. the statutory measures that
have since been introduced should ensure that
there is no !!rren5of this type of incident.
However, it is recognised that the provision of
large. unrestricted spaces within a typical Ro Ro
passenger ship is potentially dangerous. if water
in
considerable
quantities
is
permitted
to
accumulate on the vehicle deck of such vessels-.
It
is considered that
the greatest chance of
flood water gaining entry to the vehicle space is
when a side collision has occurred, either
with
another vessel or a fixed
object-The formal, investigation into the loss
of the
"Kerald.
in
addition
to
establishing
the
circumstances
surrounding
the
casualty.
also
considered what future measures could 'be taken to
contribute to enhanced safety of life at sea in
the future-
As a result of recommendations made
in the Inquiry report. the Marine Directorate
of
the UK Department of Transport commissioned a
comprehensive
research
programme
with
the
objective of enhancing the survivability of Re Ro
passenger ferries,
after specified side damage
has been assumed to occur.
The programme consisted of various elements, with
the basic objective of determining the
standar4.
of residual stability necessary to enable Ro Ro
passenger ferries
to
survive
flooding',
to
a
prescribed extent, and to avoid rapid capsize in
realistic see-going conditions.
A Risk analysis
study was made to establish the level of risk
involved in operating a typical Re Re passenger
ferry between the United Kingdom and the
near-Continent of Europe.
In this way, any proposed
improvement measures arising from the
research
work intended to
improve post-damage survival
characteristics, could be assessed in terms of
level
of
risk
from
other hazards
such
as
fire/explosion and mel-operation.
It should be
noted that the study confirmed that the
primary
hazaro that might lead to a rapid
capsize was.
jnded,
a
major
side-collision
with
another
TEGHNISCHE WUVER$TET
l..aboratorium voor
Scheepshydromechanlca
,*krch ief
Mekelweg 2,2628 CD Deift
eL016-788873.Fax 015-781838
1Principal Surveyor
Department of Transport
(Marine Directorate)
1 Before attempting to decide what was a
suitable survivability standard for Re Re
passenger ferries. it was necessary to establish the current, standards applying to the UK fleet. Accordingly. ten typical designs of Ro Re
passenger - ferry were chosen: statical
calculations (assuming still water) were made to
deter=ire the degree of cothliance with the set
of residual stability criteria. (2). being discussed at the tiae at the International
Maritime Orgahisatiön (I).
This same Set ofcriteria, with only relative minor modifications.
was accepted by IMO members and entered into
force on 29 April 1990. and apply to all new
passenger ships built after that date. These criteria may be referred to as S0LAS '90 residual stability standards. (3) This computer study showed quite clearly that conventional Re Re
passenger ferry designs have little chance of complying with these new residual stability standards: in most cases, radical design charges are indicated. it was recognised from the outset that it was likely that the survivability standards of current Ro Ro ferries would need to be enhanced significantly, to meet the standards
indicated by the research. There were strong
indications of this in the series of model tests
carried out in the early 1970s in the United
Kingdom on a typical Cross-Channel ferry of that time (4). Accordingly, the research included the consideration of various possible design changes and enhancing devices which could be employed on
existing ferries to improve their survivability
characteristics. These improvements say be
achieved in tvo main ways - by the fitting of internal or external arrangements. Zn broad
terms, the former restrjctC the extent of
internal flooding which might occur, whilst the
latter increases the potential to right the vessel after damage has occurred. All the research studies described previously are useful and provide valuable data for the future, but the
essential question - "Jhat stability standards
are required to give a reasonable guarantee that a ferry built to those standards will not capsize
rapidly?" - needs to be answered.
Unforttinately, a study of historical damage data.
whilst useful in indicating the likely position
and extent of any future damage, does not provide much assistance in answering this. question. It
waS decided that it was necessary to conduct a
series of damaged model experiments in controlled conditions of damage location and extent. condition of loading and weather conditions
(measured in terms of sea state). The. data
provided from such tests would then provide the means to decide what minimum residual standards
of stability are needed to avoid rapid capsize in sea-going conditions.
A more complete summary of the h work is contained in an Overview report (5), published by the Department o. Transport and also distributed to IMO members as an information paper (6).
Main
Fig
1 :General Arrangement of one of the two Models
TT PUR
Two organisations were chosen to carry o
the
model tests - British Maritime Technology (BhTJ of Teddington. England and the Danish Marft,5
Institute (DM1) of Lyrigby. Denmark,
Each
organisátion was commissioned to construct
test a 42nd scale CRP model, representing typical cross-Channel Ro Ro ferry. Fig
i
illustrates the principal features of one of the models.
The damaged space below the vehicle deck ws5 located in the midships region, the length and
longitudinal position being arranged such
that pure sinkage occurs without significant trim down
to the level of the vehicle deck, The assumed
damage was fixed at the statutory limit of + 3.0 (metres), with a Vee-shaped penetratjo of
8/5 at the deck level. The ship condition, was adjusted by the use of light. impervious
inserts at both ends of the damaged space, together with appropriate means of adjusting the model KG.
m5
vehicle space and the midships damaged space were essentially free of obstructions which mightrestrict the flow of floodwater during the tests.
The models were tested in irregular beam seas.
generated by a wavemaker. for various significant
wave heights ranging from 0.5 metres to 5.0 metres. .IONSWAP wave spectra were used with modal periods ep'propriate to wave scatter date
gathered on an all-seasons basis from the southern North Sea region.. For. the testi. the
models were placed in the path of the on-coming waves, with the damage opening facing them. for
the majority of cases. (Some tests were carried
out with the damage on the side remote from the
on-coming waves). Readings of the model motions
were taken continuously, whilst simultaneously the behaviour of both the model and the flood
water within the hull were being observed.
For these tests where capsize did not occur
within a spCci'fied test period, the model KG was
raised and the entire procedure repeated until
capsize took place.
Uhere capsize occurred extremely rapidly, the
tests were repeated using lower ICC values to
ensure that the estimated capSize zone was defined within as narrow a band as possible.
LXILL$G TT RLTH WITH SThTIQ.L
.ITER.tAPresent-day statutory stability criteria are
based on static principles and it is unhikaly
that this situation will change within the foreseeable future. These empirical standardS
are essentially based on the characteristics of a Cl curve, calculated assuming still-water
conditions for a specific condition defined in
terms of draught, trim and XC. They do not
Damage Opening
Vehicle Space
Future Damage
Space
therefore take direct account of the complicate4 dynamic -aetions which actually occur in sea-going condition3
In respect of present intact stability standarde. all the available evidence suggests that they are adequate for all sea states likely to be encountered, provided that watertightneas is preserved and shift of cargo does net take place.
However, where residual stability standards are
concerned. empirical data is sparse and there is therefore some doubt as to whether a ship
complying with these standards -would survive
statutory damage in realistic sea-going
conditions.
-It is' reasonable to assume a positive correlation
between the residual stability curve
characteristics (consisting of GE sax, range and area) and the corresponding ability of a ship
having a residual standard capable of resisting
capsize in reasonable sea-going conditions. The tern 'capsize' needs to be defined as far as
the conduct of these model tests is concerned. A rigid interpretation was deliberately not chosen; rather. 'capsize' was considered to -have taken
place when the rate of change in heel shows a distinct increase.
Each teat run was carried -Out over a maximum
timespan. equivalent to at least 60 minutes ship-scale. Non-Capsize was assumed, unless capsize (as defined above) occurred within this tizescale.
S'r*uToal RIDJ. SmRILITY CITW
Intact stability criteria are empirical in nature and based upon the principal characteristics of a
GZ curve - positive area(s) under the curve. GZ max. and range. These criteria have been applied
for many years now and are thoroughly tested in
all sea conditions likely to be encountered. There, is no firm evidence that an ocean-going
ship has been lost
due to inadequate intact stability,always provided of course that no
water ingress
or shift of cargo took place.
Indeed,
there say be some redundancy
in the present criteria; moreover certain ship types and lengths may have criteria which are morestringent than is strictly required. However.
the- present level of intact stability criteria
seems to meet with general agreement.
Righting Lever
GZ (m.)
H
Area ) 0.015 rn.rad.
Heel Angle
Fig 2
:Diagrammatic representatibn
of the SOLAS '90 residual
stability criteria
Note : (
) are the criteria for
In respect of damage stability, the criteria are
similar in format to that which apply to intact
stability: there is en additional requirement
that a minimum residual freeboard is to be retainçdf tar assumed damage - in other words, a
requi'mnt that the jargin line is not to be
immersed. H . because the incidence of
serious flooding caused by aide-collision damage is relatively low, evidence regarding the desired level of residual stability needed to survive is quite sparse.
Fig-2 shows diagrammatically the recently adopted internatiOnal standard of residual stability
which applies to passenger ships built after-
-April 1990. They represent a significant improvemCnt in safety standards compared to those
applied previously, but they remain standards
which
do not
relate directly to actual seaconditions, being based on statical
considerations. It is most likely that, for the
forseeable future. etatütor-y stability
requirements will be based on statical
calculations. Historical data derived from previous damage incidents will, hopefully. continue to be rare; therefore there is a need 'to allow for the dynamics of a flooding Scenario
and the most straightforward way of doing this
seems to be by a
series of controlled tests-involving "damaged" models.
A direct link may then be established between the
results of statical (mtill--water) calculations and the corresponding test conditions, to enable
a judgement to be made as to
the chances ofsurvival of a ship in a given sea-state.
By such means,
the main factors which govern
survival after damage - damage location and extent, ship condition, internal arrangement of watertight divisiona and permeability of damaged
spaces - may be examined in a systematic manner.
ZLYSIS Cf T
TT RULTS
The two ferries modelled for the tests are quite èpi&sentative of
those UK
passenger ferries engaged on the cross-Channel routes. Present-day designs of such ferries do not vary significantly in either size or proportions. Al-i tend to havethe common features of extensive, unrestricted
vehicle spaces - apart from relatively small side
or centre casings. To derive maximum benefit
from the two sets of tests - and, in addition, to
take account of the tests carried out in the UK
some 20 years ago (4) - a logical, coherent theory is needed to extend the analysis to ships of different size and proportions. Such a
theoretical approach needs to be consistent with
the observed behaviour of the models under test. when subjected to various sea states and with the
damage opening facing towards (or away from) the on-coming waves.
Observations at the time of the test runs indicated that in order that capsize should take
place. minkage has to occur. In other words. there has to be a net increase in the amount of flood water taken on to the vehicle deck. In addition, there is a gradual increase in heel
- angle - in most cases, but not all, towards
the
side of the damage openings -- culminating- in a
significantly increased rate of heel angle immediately prior to capsize.
Capsize is unlikely to occur where the minimum
residual freeboard (in still water) is at least as great as the height of the- on-coming waves.
Qualitatively, then, a ferry'. ability to resist capsize is dependent on
the condition of loading.
the area of the vehicle deck capable of being
flooded, and
the minimum freeboard after damage.
Additionally, a mere severe sea state will mean that increasing amounts of flood water will be taken aboard, and thus reduce the ability to
resist capsize. Hence, for a given sea state the
ability to
survive may be regarded as being
represented by some complex function of KG (or
GM). B. f and
Fig 3 illustrates a typical statical residual GB curve calculated for a specified KG position and residual freeboard, the latter values being taken
from the test results at positions where the
model had a roughly equal chance of surviving or capsizing.
Fig 4 shows a plot of such positions, given for a specified damage location, damage orientation and residual freeboard. It is represented as significant wave height against flooded CM.
It is important that the results obtained from both Sets of model tests, and the previous UK
tests mentioned earlier should be presented in a
format capable of a more general analysis. A non-diensianalised form of presentation was
suggested by EM? and tested on the two sets of
results - see Fig 5. More work needs to be done in this respect, but the initial indications are that this line of approach is promising.
Righting Lever GZ 0.1
(m.)
5
Heel Angle
Fig 3: Typical calculated residual GZ
curve (shipscale)
Corresponding Model Test Condn.
Midship Damage : Facing Waves:
0.58m. residual freeboard:
1 .00m. significant wave height.
10' Wave
height
(m.)
1 2 3 4Flooded Ship GM (m.)
Fig 4: Plot of wave height
A.iflooded GM
Note :
The mean line has been
derived from a zone of
possible capsize either side
of this mean.
I
Hs,
/f
40
3.020
1.00.10
10.GM1 A.T/82
Fig.5 : A suggested nondimensional
representation of the test
results.
5.Om.
Damage into waves
0.58m. residual freeboard.
2.9m.
1.Om.0.5m.
Shipscale
N0NC SIZE
REGION Mean Line RTAINAll the tests conducted to-date relate tá a
typical cress-Channel ferry of current design. with few or no obstructions within the vehicle
spaces.
In the research undertaken so far, the
enhancement in survivability through the fitting of various devices/arrangements ha. been
investigated by means Of statical (still water)
computer calculations. There is a need, therefore, to conduct further tests to confirm
that the improvements in residual stability indicated by these calculations are attainable in sea-going conditions. Therefore, it is intended
to carry out a further series of model tests.
using the models as-built, together with a
modelling of the principal devices/arrangements
studied in Phase 1 of the tests.
When the two models were constructed, provision was made for assumed damage in the forward end. and it is intended to run some tests so as to indicate what standards of survivability are required for damage in this region: this will
enable comparisons to be made with the results
obained assuming damage amidships.
In making the damage stability calculations, it
was dear that the choice of the permeability
factor for the vehicle space could affect the GZ
max and range significantly. It is intended to
modal a full load of vehicles on deck and conduct
tests, so that the results might be compared to the corresponding statical calculations - the
letter using an appropriate permeability factor
for the vehicle space.
ACTI
TZ
The Steering Committee, which was formed to
oversee the UK research programme, produced a report which was forwarded to the Minister for Transport, in which they recommended that SOLAS 90 residual stability standards be applied to existing UK Re ro passenger ferries as soon as
possible. They further recommended that these
same standards should also apply to non-UK
ferries using UK ports. The Minister accepted
these recommendations and subsequently the UK has put forward its views at the 58th Session Of MSC. Recently, the near-European Administrations were
approached to explain the UK position and to
establish the degree of support the UK might
receive in applying S0I.AS 90 standards to
existing passenger ferries. A formal
presentation to IMO of the UK position will be
made at the forthcoming SLF Sub-Committee and the MSC next year.
It is expected that some preliminary findings of
the Phase 2 model tests should be available at
that time.
2T AID
The main objective of the reseach programme
described was to determine the standard of residual stability needed to enable a Ro Ro
passenger ferry to survive flooding and avoid
rapid capsize in realistic sea-going conditions. Relating the model results and subsequent analysis to typical present'day Re Ro ferries, it can be acid that this primary objective has been achieved. In particular, the results indicate that a ferry (with a vehicle space free of obstructions) should have a reasonable chance of
survival (ic. will not cpsize rapidly) if built to coaply with SOLAS 90 standards of residual
stability.
The research programme also provided designers
with much useful data concerning various devices/design arrangements which might be fitted
on new or existing ferries in order to enhance
their survivability. There is a need to model these devices in a further series of teats i
order to establish that the apparent enhancement indicated by still-water coeputer calculations j indeed achieved in a dynamic situation.
Complete benefit from the modal test results will
not be achieved until there is
a consistent.logical theory which will enable the rsultsto be applied to other ship forms and sizes. Observing
the models during the tests provides an insight
into the main factors which need to be taken into
account in producing such a theory. The model
results provide a good basis for judging the
intrinsic worth of any theory proposed.
IMO Intergovernmental Maritime Orgariisation SLF Sub-Committee on Sub-division, Leadlines
and Fishing Vessels Safety MSC Maritime Safety Committee
SOLAS Safety of Life at Sea (Convention) SOLAS 90 A set of residual stability criteria
which is to be applied to passenger
ships built after 29 April 1990 MMSO Her Majesty's Stationery Office KG Height of ship centre of gravity
GM Metacentric height
Cl Righting lever
/ Minimum flooded freeboard
B Ship's meulded breadth
T Ship's moulded draught
H Significant wave height
The UK research programme on Ro Re passenger
ferry safety has provided the source material for this paper. The research was commissioned by the
Marine Directorate of the Department of
Transport, and permission by the Directorate's
Surveyor General to publish is duly acknowledged. Essentially, the contents of the paper are
factual and informative: where opinions are expressed, they do not necessarily reflect those of the Department of Transport.
Formal Investigation - my "Herald of Free Enterprise". Report of Court No 8074. HMSO. London. 1987.
IMO document SLF 32/21 (Annex 4) containing proposed draft amendments to chap II-1/8 of 1974 SOLAS, as amended. London, 1987.
IMO Resolution MSC12(56) containing a
revised text to chap 11-1/8 of 1974 SOL.AS. as amended, London, 1989.
Bird H and Browne RP: Demaged Stability
Model Experiments.
Transactions,
RoyalInstitution of, Naval Architects, Vol 116. 1974. pp 69-91. London, 1974.
An Overview Study. Research into
enhancing the stability and survivability of Ro Ro passenger ferries. Produced by BMT (Defence
Services) Limited for the UK Department of Transport. London. April 1990.
IMO information paper MSC/Irif 7. London. April 99O.
Re Ro Safetl - Report of the Steering
Cesittee of the
Re, Re ferry -safety researchprogramme. London. April 1990.
1110