10th International Conference
on Stability of Ships and Ocean Vehicles
191
ANALYSIS OF THE CAPSIZE AND LOSS OF A 12000 DWT BULK CARRIER
Rumen Kishev, Bulgarian Ship Hydrodynamics Centre, [email protected]
Vladimir Rakitin, Bulgarian Ship Hydrodynamics Centre,
[email protected] Silvia Kirilova, Bulgarian Ship Hydrodynamics Centre,
[email protected]ABSTRACT
In early 2004, a 12000 DWT bulk carrier suffered heavy storm weather when sailing in South-
West Black Sea and has been reported to capsize and sunk shortly afterwards. Numerous lives have
been lost.
In the paper, analysis is made of the circumstances of the accident and consequent loss of
stability and sinking. Official data and documentation on ship technical status as well as weather
conditions en route have been solely utilized as supplied by the National Investigation Agency.
Theoretical and experimental investigations on the vessel’s operational stability, damage
stability and seakeeping have been implemented and the most probable scenario of the accident has
been drawn.
Keywords: Stability, Capsize, Seakeeping
1. INTRODUCTION master reported, that the vessel is intensively
taking water in Hold № 1. A few minutes later
1.1. Background the ship capsized and sank, according witnesses
from a ship heading close behind and
The history of accident, recreated after confirmed by the traffic operators in the port of
witness testimonies, documents and recorded Istanbul.
conversations of the ship officer’s staff with the
port officials, describes the event as follows:
1.2. Site — Black Sea, about 8 -10 nm in front
Early morning on the day of disaster the of Bosforus:
vessel left Ukrainian port Yuzhny for Gebze Witness
(Turkey), loaded with 11750 t coal, in spite of
Shipwreck
the storm warning for the next three days. The
wind and seas in the area were Nord-Nordwest,
with increasing severity. At the approaching of
Bosforus the vessel has been warned by the
port authorities, that the channel is closed due
to the bad weather and the master ordered
change of course at adverse angle to waves and
low speed. Later on, he has been given
instructions to assume course to Bosforus. At a
Figure 1. Location of the two ships at the time
position about 10 miles from the shore the
of the accident.
10th International Conference
on Stability of Ships and Ocean Vehicles
192
1.3. Reported Loading 1.7. Ship condition before capsizing
According the cargo plan submitted by the By testimony of the witness ship’s crew,
ship master and approved by the port immediately before capsize and sinking the
authorities, the vessel has been loaded with subject vessel sailed at 5-6 knots with visible
11751,7 t coal, distributed over cargo holds as heel of 8-10 deg to starboard and about 1 deg
follows: trim to bow.
Hold № 1 – 1800,0 t
Hold № 2 – 3200,0 t
Hold № 3 – 3325,0 t 2. ANALYSIS OF LOADING CONDITION
Hold № 4 – 3426.7 t
The vagueness of initial information on
At this, the logged drafts and trim have loading concerned as cargo itself as well as the
been: lightship weight.
Draft astern: 8,375 m
Draft afore: 8,188 m
Trim to stern: 0,200 m 2.1. Lightship
The lightship weight had changed with
1.4. Heading years of operation and numerous modifications
of hull structure which have been not properly
By testimony of the witness ship’s master, documented and approved. In particular, some
at the time of the accident the subject vessel years ago the tweendeck had been cut off, later
sailed in following seas with a course angle of two of the deck cranes had been removed also.
about 20° with reference to wind and wave
propagation direction. For conformity with existing ship
documentation, lightship weight has been
assumed to be 3773 t, as declared by the latest
1.5. Wind speed cargo plan approved by Yuzhny port
authorities, even if the weight assessment
As reported by the shore weather control during the investigation shows a most probable
center, the wind speed at the time of the value of 3657 t.
accident has been about 55 – 57 km/h, with
blows up to 75 – 80 km/h, direction Nord.
2.2. Cargo weight and distribution
1.6. Wave height Loading data given in 1.3 correspond to the
cargo plan submitted by the ship master before
By testimony of the witness ship’s crew, the last departure and approved by the port
the sea severity has been 6-7 Beaufort, which authorities. In its analysis, however, several
corresponds to significant wave height of about inaccuracies have been found out, namely:
6 m and average period of 9 sec. This conforms • The declared cargo quantities for the four
with wind speed data reported above. cargo holds do not correspond to their
capacity, as shown below:
10th International Conference
on Stability of Ships and Ocean Vehicles
193
Hold № 1 2 3 4 initially poured-in water will be absorbed by
Capacity m 3
2827 4701 4710 4642 coal, but after saturation of the mixture to a
liquid state a free surface will appear.
Declared
1800.00 3200.00 3325.00 3426.67
load, t
Max 3. SEAKEEPING ASSESSMENT
allowable
1949.6 3242.1 3248.3 3201.4
load at
3
SF=1.45 m /t The seakeeping qualities of the vessel have
been analyzed by linear strip theory. Serial
under within within over
Comment calculations have been made on ship motion
loaded norms norms loaded
parameters, motion related phenomena and
• Declared values for Holds № 1 – 3 are added resistance at various speeds of advance
rounded numbers, while for Hold № 4 they and various headings.
are given with accuracy up to the third
digit, which gives grounds to conclude, that The results obtained show that motion
the stability calculations in the cargo plan parameters are within expected limits, the
had been “attuned” to match the draft differences between three loading conditions
readings. An additional argument for that is being minor. However, at adverse headings (as
the unusually high correction constant of was the case of forced storming after closing of
195 t virtually situated in the superstruc- the strait), the deck wetness intensity is very
ture, which has no physical explanation and high – even at speed reduced down to 6 kn at
does not exist in earlier stability Bf6 (Н1/3 = 6 m), the number of deck wetness
calculations. occurrence for loading case СН1 is about Ndw
= 50, while good marine practice allows 30.
• The free surface corrections for service
This means that the bow part and the hatch
tanks had been estimated incorrectly.
cover of Hold № 1 in particular had been
subjected to intensive loading by
For the sake of this investigation, several
overwhelming waves. For loading cases СН2
possible loading cases have been thus
and СН3 this intensity increases 1.5 – 2 times,
formulated:
due to effective freeboard decrease after
flooding of the bow compartments.
Loading Description Weight Draft Draft
case displ, t fore, aft,
m m The added power in Bf6 conditions restricts
attainable speed down to 8 kn, which is close to
According the value reported by the witnesses.
last
СН1 15671.0 8.28 8.32
submitted
cargo plan
Added 500 t
water
СН2 16171.0 9.18 7.88
poured in
Hold № 1
Another
350 t water
СН3 16521.0 9.99 7.52
added in the
forepeak
The last two cases have been lately
considered with and without existence of free Figure 2. Illustration of intensive deck wetness
surface in the holds, having in mind, that at head seas.
10th International Conference
on Stability of Ships and Ocean Vehicles
194
4. WAVE LOADS ON HULL STRUCTURE It is well seen, that in head seas the
resulting shear forces at the bow, in vicinity of
The vertical shearing forces and bending Hold № 1, exceed the maximum allowable
moments in waves have been calculated by values in case of chancing over a wave larger
strip theory, integrating wave loads along the than the statistically average one. It means that
ship length and adding those in still water. at the time of taking adverse course to waves in
heavy storm, the occurrence of structural
The allowable limits have been estimated damage at the bow or dislocation of the
according the IACS Unified Requirements [2]. hatchcover was highly probable.
Below, results of the shearing forces
calculations are illustrated for the CH1 loading 5. STRUCTURE STRENGTH
condition. In the figures, ASSESSMENT
Qw, H1/3 - Vertical shearing force acting
in the cross section at passing along of a wave The overall strength calculations followed
with height equal to H1/3 = 6.0 m the vessel operational lifetime and structural
Qw, Hmax - Vertical shearing force acting modifications, as follows:
in the cross section at passing along of a wave • The vessel has been built in 1975.
with height equal to Hmax = 1.75 H1/3
• In 1996, the tweendeck structure has been
Qwp* - Allowable positive shearing force cut off in all four cargo holds.
Qwn* - Allowable negative shearing force • In 2002, ultrasonic thickness measurement
has been implemented and significant wear
Loading Case СН 1 Head Seas Vs = 6 kn (corrosion) has been observed on structural
1500 members and plating. Thickness reduction
Qw, H1/3 Qw, Hmax
has been done within GL (IACS)
Qwp* Qwn*
recommended limits.
1000
Qw • The same year, partial repairs have been
[t]
implemented, replacing sections of main
500
deck and double bottom with plates thinner
than the recommended reduction.
0
0
AP
2 4 6 8 10 12 14 16 18
FP
20 Following this succession, four structural
conditions have been considered, as follows:
Loading Case СН 1 Haed Seas Vs = 10 kn К1 – ship structural state right after
1500 delivery;
Qw, H1/3 Qw, Hmax
Qwp* Qwn* К2 – ship structural state after tweendeck
1000 cut off;
Qw
[t]
К3 – ship structural state with reduced
500
thicknesses of structural members and
plating;
К4 – ship structural state at the time of the
last voyage.
0
0 AP 2 4 6 8 10 12 14 16 18 FP 20
Figure 3. Shearing forces along the ship length The strength estimation for the case K4 in
as compared to the IACS allowable values. particular shows that the reduction of cross
section inertia amounts to 14%, and this of the
10th International Conference
on Stability of Ships and Ocean Vehicles
195
section modulus – to 19%. At this, the absolute operational stability. More specifically, at
value of the section modulus becomes less than significant wave heights over 5 m, after the
the allowable limit according IACS Unified instance of water penetrating the holds and
Requirements [2]. This practically means that forming free surface the ship becomes unstable.
the ship does not possess even small reserve of Loading Case CH 1 Following Seas
strength to be able to withstand severe wave 20
1 - Low stability zone
loads in harsh environment. 18
2 - Main resonance zone
3 - Parametric resonance zone
5.0 m
16 o
4.0 m 7.0 m μ=0 λ=130 m
Vs [kn] 9.0 m
6.0 m 8.0 m
The reduction of inertia and section 14
Vs oper
modulus leads also to increasing of
12
1
10
deformations from overall bending in waves or 8
during loading-discharge operations and can 6
provoke cracks, plastic deformations or 4
destruction. Increased deflections, together 2
3
2
with the direct action of the green water on 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3
GM [m]
deck, often is the reason for dislodging of
hatchcovers and consequent water penetration, Figure 4. Diagram of endangered stability in
as it had been frequently observed on bulk following seas for the intact ship.
carriers.
Loading Case CH 2 Following Seas
20
Vs [kn] 1 - Low stability zone
6. ASSESSMENT OF STABILITY
18 2 - Main resonance zone
3 - Parametric resonance zone
16
o
IN FOLLOWING SEAS μ=0 λ=130 m
4.0 m 6.0 m
7.5 m
14
5.0 m
8.0 m
12 7.0 m Vs oper
1
The assessment of vessel stability when 10
advancing in following seas has been
8
implemented according the recommendations 4
of the Resolution MSC/Circ.707. A diagram of 2 3
2
safe operation regimes has been drawn, 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3
GM [m]
distinguishing three zones of danger, as
follows: Figure 5. Diagram of endangered stability in
• Zone of significant reduction or loss of following seas for the damaged ship – Hold №
stability, when the wave crest passes along 1 flooded, static heel 9°, 1.30 m trim to bow.
the middle; Loading Case CH3 Following Seas
20
• Zone of the main roll resonance, when the Vs [kn]
18
5.0 m
1 - Low stability zone
2 - Main resonance zone
3 - Parametric resonance zone
wave encounter period becomes close to 16
4.0 m
6.0 m 7.5 m
μ=0o λ=130 m
the roll natural period;
7.0 m
14
12 Vs oper
•
1
Zone of parametric resonance, when the 10
wave encounter period becomes close to 8
the half of the roll natural period;
6
2
2
3
The diagrams for the safe operation in 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3
following seas for the three loading conditions GM [m]
are shown in Figs. 4 – 6. It can be seen that
when sailing with heel and trim and taking into Figure 6. Diagram of endangered stability in
account the newly formed free surfaces in the following seas for the damaged ship – Hold №
partially flooded cargo holds, the vessel falls 1 and forepeak flooded, static heel 9°, 2.45 m
into an unfavorable zone which endangers her trim to bow.
10th International Conference
on Stability of Ships and Ocean Vehicles
196
7. MODEL TESTS ON CAPSIZING IN 7.3 Tests on damaged ship
FOLLOWING SEAS
Both loading cases corresponding to the
Conclusions drawn by the numerical damaged ship conditions СН2 and СН3 have
assessment have been verified by specially been inspected in following seas ( 0° and 20°),
tailored model tests. Various scenarios have speeds of advance 10 and 6 kn, black-out
been realized and examined. condition. Damage has been imitated by
various rates of hatchcover № 1 opening.
7.1. Model preparation From the model responses observed it could
be concluded that even at considerable heel and
The model has been manufactured in 1:38 trim in extreme seas the ship retains positive
scale using drawings submitted by the original stability until the moment of hatchcover loose.
shipbuilder. The freeboard, superstructure and The crucial circumstance for the accident
large deck equipment have been exactly happened to be the water ingress in bow
modeled. The model has been intended for compartments, which sharply reduces stability
free-running tests, with watertight due to the newly formed free surface. In that
compartments for on-board instrumentation. condition, passing of an extremely high wave
The cargo holds and the forepeak have been forces the ship to ride-on, restoring forces
made waterproof by epoxy coating. rapidly decrease and the ship capsize.
The model has been dynamically balanced
along the three inertial axes.
7.2. Tests on intact ship
The tests on the intact ship, corresponding
to loading condition CH1, have been carried
out at various headings and speed of advance,
as follows:
• Advancing in head seas at speed 10 and 6 kn
• Change of course at low speed in beam seas
• Advancing in following seas at speed 10 and Figure 7a. Final phase of surf-riding – large
6 kn wave is passing along the ship, water is
pouring into an opening in the hold.
• Black-out (stop engine) in following seas
In all tested cases, the model kept its good
stability. In head seas, however, intensive deck
wetness has been observed even at low speeds
of advance. It could be a reason for excess
loading on hatchcovers and air-vents as well as
for jammer’s loosing, the probable cause for
water ingress into bow compartments at a later
stage of accident progressing.
10th International Conference
on Stability of Ships and Ocean Vehicles
197
• Horizontal slip of the hatchcover with the
edge leading ahead;
• Horizontal slip of the hatchcover parallel to
the board line;
• Vertical entry into water.
The process of hatchcover submergence has
been registered on video and excursions from
the initial position read every second (model
scale). Surprisingly good reproducibility of
tests has been observed. Sample results are
shown in Fig. 8.
Figure 7b. Ship losses stability.
Excursion, m
0 10 20 30 40 50 60 70 80
0
0.5
Average excursion
after 10 tests - 69 m
Minutes after entry in water
1.5 Test 1
Test 2
2.5
Figure 7c. Ship capsizes. Figure 8. Falling trajectory of the hatchcover of
Hold № 1 in case of entering into water with
the edge ahead.
8. MODEL TESTS ON THE “FALLING
LEAF” EFFECT Based on the test results it can be stated that
the hatchcover’s fastening dogs have been
During initial reconnaissance of the ship loosed in advance, maybe by green water
wreck by a ROV, the hatchcover of Hold № 1 chancing on deck during storm ride, and the
has been found lying on the ground in vicinity shock of collision with the ground at the
of the main body, while other hatchcovers have moment of sinking provoked final
been found intact. Having in mind the water disengagement.
depth at the site, it was considered possible to
find a relation between the distance to the
wreck, the moment of disengagement and the 9. CONCLUSIONS: MOST PROBABLE
manner of hatchcover falling into water, SCENARIO OF THE ACCIDENT
considering the “falling leaf” effect.
From the analysis performed it can be
The investigation has been carried out with concluded, that the reason for the ship
a flexible hatchcover model, imitating the capsizing is the simultaneous occurrence of
geometry as well as weight in water. Three four unfavorable circumstances, namely:
possible cases have been tested, as follows:
• Extremely rough storm conditions;
10th International Conference
on Stability of Ships and Ocean Vehicles
198
• Closing of the strait and related forced length comparable to the ship’s length passes
changes in course and speed as well as the along, the stability arm becomes negative and
lack of asylum haven; the ship capsizes. Further, the water penetrates
the nearby compartments due to local structural
• Water chancing in bow compartments due
damages, the ship losses floatability and sinks.
to dislodging of hatchcover or air-vents and
related reduction of initial stability;
• Instant loss of stability due to forced 10. REFERENCES
raiding on an extremely large following
wave. MSC/Circ.707/19.10.1995 – Guidance to the
master for avoiding dangerous situations in
Taking into account these circumstances, following and quartering seas
the most probable scenario of the accident can
be narrated as follows: IACS Unified Requirements, Rev. 2007
The initial course of ship is advancing with HECSALV – User’s Guide, 2004
about 10 kn in following seas from North to
South. After receiving the information of strait SOLAS, Consolidated edition, 2004
closing the ship changes course to adverse at
low speed. About this time she starts to take MSMSL – User’s Manual, BSHC, 1992
water in Hold № 1 and/or the forepeak, due to
one of the following reasons:
• Breaking of air-vents by waves overcoming
the bow;
• Dislocation of hatchcover of Hold № 1 due
to significant bending or torsion of
structure under extreme wave action and
loosing of fastenings;
• Splitting of a local crack by the above
reasons.
(The exact reason of loosing hold
watertightness could be specified after detailed
diver inspection).
At course change combined with heavy
rolling, cargo shifting or sliding over wet coal
layer initially formed at the bottom caused
static heel. At lack of asylum haven nearby, the
ship assumes her initial course in following
seas with increased speed up to 10 kn with the
intention to reach and pass the Bosforus under
emergency conditions. The continuous water
ingress, already for certain through the
dislodged hatchcover, increases the heel and
trim, free surface is formed in both cargo hold
and forepeak, the metacentric height reduces
sharply. At an instant when an extreme wave of
