International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 6
Number 4
December 2012
1 INTRODUCTION
The most commonly an oil accident is considered as
a result of an oil tanker cargo spill. However this
sort of disaster is rising the most the public
awareness, hopefully cargo spills are not the most
common ones (ITOPF, 2011). The most frequent oil
spill accidents are those related to accidental
discharges from bunker tanks of all vessels, not only
tankers.
Contrary to tankers transporting heavy oil as
cargo, ships’ bunker tanks are not required to have a
double hull structure as protection. The statistics on
the oil spills provided by The International Tanker
Owners Pollution Found reveal that the significant
amount of oil spills of a size in a range 7-700 tones
come from accidents like collisions and groundings
(Fig. 1) (ITOPF, 2011).
In the past 40 years there were 1249 minor spill
accidents, whereas 444 cases account for the major
oil spills of a size above 700 tons, and only tankers
are considered. Taking into account all types of
ships this number can be expected significantly
higher. These data shows that the minor size oil
spills should not be neglected in any risk analysis
related to maritime traffic, as they are the most
frequent oil spills to occur.
At present there are two recognized and adopted
methods for the bunker spill size estimation. Both of
them seem to be quite general in nature and rough in
results.
Figure 1. The distribution of causes of minor oil spills (7-700
tons) (ITOPF, 2011)
One approach is based on historical data. This
approach is relatively easy, fast and straightforward,
thus it has gained popularity among researches and
was adopted in some general studies (HELCOM
Modeling of Accidental Bunker Oil Spills as a
Result of Ship’s Bunker Tanks Rupture – a Case
Study
P. Krata & J. Jachowski
Gdynia Maritime University, Poland
J. Montewka
Aalto University School of Engineering, Espoo, Finland
Maritime University of Szczecin, Poland
ABSTRACT: AIS (Automatic Identification System) data analysis is used to define ship domain for
grounding scenarios. The domain has been divided into two areas as inner and outer domains. Inner domain
has clear border, which is based on ship dynamic characteristics. Violation of inner domain makes the
grounding accident unavoidable. Outer domain area is defined with AIS data analyzing. Outer domain shows
the situation of own ship in compare with other similar ships that previously were in the same situation. The
domain can be used as a decision support tool in VTS (Vessel Traffic Service) centers to detect grounding
candidate vessels. In the case study presented in this paper, one type of ship, which is tanker, in a waterway to
Sköldvik in the Gulf of Finland is taken into account.
495
1990, Safetec 1999, Gucma&Przywarty 2007,
Nyman 2009). The limitations of this approach and
its weak areas were pointed out by
Michel&Winslow (1999) and Eide et al (2007),
where the main concern about historical data was
that they are not necessarily representative to today’s
accident scenarios, mostly due to changes in ship
construction or layout of the tanks.
Another method, included in the IMO guidelines
for approval of alternative tanker designs (IMO
1995, IMO 2003), contains a probabilistic-based
procedure for estimating oil outflow performance.
Probability density functions describing the location,
extent and penetration of side and bottom damage
are applied to a vessel's compartmentation,
generating the probability of occurrence and
collection of damaged compartments associated with
each possible damage incident. All oil is assumed to
outflow from tanks penetrated in collisions, whereas
outflow from bottom damage is based on pressure
balance calculations. This method sounds, however,
more reliable than the previous it still lacks the time
component. The method does not provide this vital
information on the rate of the spill nor the time
needed for tank to be released. From the
preparedness and response point of view this
parameter is essential, as the bunker spills occur in a
close vicinity of a shore and the response time is
usually very limited.
Recently a methodology has been introduced
based on the analytical calculations and time domain
simulations in order to calculate the volume of oil
outflow and outflow rate versus time (Tavakoli et al.
2008, Tavakoli et al. 2010). The method addresses
accidental cargo spills from tankers.
In this paper a method for bunker spill estimation
in spatial-temporal domain is presented. The
methodology takes into account the fluid dynamics,
the size of a tank rupture is estimated with the use of
the IMO methodology. However the damaged tank
is assumed not to be a subject to longitudinal and
transverse motions.
2 BUNKER OIL SPILL MODELING BY 3-D
CFD METHODS
The technique for oil spill modeling applied in the
paper makes use of Computational Fluid Dynamics.
Authors propose the methodology aiming at
estimation a quantity of the bunker spill, a rate of
such a spill and time for the bunker to release. The
method can contribute some information to the
probabilistic approach utilized in previously
mentioned IMO methodology. CFD based solution
seems to be useful for better understanding the oil
outflow process and its duration.
The proposed methodology has a wide range of
applications and is free of the constraints typical for
IMO statistical approach. In the paper a model for
bunker spill estimation is put forward and finally a
case study is presented, which is assumed as an
exemplary grounding accident.
The 3-dimentional simulations of oil trickling and
disseminating in water phenomenon were performed
by the use of the commercial code “Fluent”. The
software is an universal and flexible tool designed
for modeling of liquids dynamics. Most commercial
CFD codes use the finite-volume or finite-element
methods which are well suited for modeling flow
past complex geometries (Bhaskaran&Collins). The
Fluent code uses the finite-volume method (FVM),
and uses the volume of fluid (VOF) method for free
surface problems (Dongming&Pengzhi 2008, Fluent
2006).
The numerical simulations of the oil dispersing in
water phenomenon were performed for a number of
damage extend configuration (Fig.4) and tank
geometry corresponding to the relevant parameters
of the selected bulk carrier. The cross section of a
vessel and the location of a damaged double bottom
tank is shown in Figure 2.
Figure 2. The cross section of a ship and her double bottom
tank to be ruptured
In the course of the study a typical double bottom
bunker tank of an exemplary bulk carrier is
considered. The characteristic dimensions of the
damaged tank are as follows:
− length – 40.0 m;
− breadth – 9.6 m;
− double bottom height – 1.9 m.
The shape of the double bottom bunker tank is
presented in Figure 3.
496
Figure 3. The damaged tank shape and dimensions
A leakage of bunker oil might take place due to a
variety of reasons among which collision or
grounding are the most common (Fig.1). There are
available statistical analyses of damage locations in
all three dimensions within ships hulls and damage
extends. Usually collision and grounding are
researched separately and nowadays some widely
accepted distributions of hull damage size are in use.
Nevertheless the diversity of damage location is
noticed, for the purpose of the case study one
exemplary double bottom tank is considered and one
elevation of the damage above the ship keel. The
vertical extend of tank damage and its location is
shown in Fig. 4.
Figure 4. The location and span of the damage in double
bottom tank (an exemplary case study)
The numerical simulations of oil spill were
carried out for a number of tank damage lengths, i.e.
10%, 40%, 70%, 85% and 100% of compartment
length (Fig. 5). This was to estimate how far
nonlinear effects influence the final results,
especially in terms of a rate of the outflow. The
variable damage length was modeled by the use of
removable panels concept which was convenient
from the computational mesh creation point of view.
When the considered geometry of the damaged
tank was established a set of assumptions required
for the numerical simulations needed to be set up.
The assumptions are related to the computational
mesh creation, Courant number range, time step,
fluid viscosity modeling, etc. Then the boundary
conditions were defined.
Figure 5. Removable panels used for modeling of different
length shell damage
A variable computation time step was applied in
the solution of the conservation equations for mass,
momentum, and volume fraction of the both liquids
water and oil. All numerical simulations were based
on a 3D quadrilateral mesh created in GAMBIT. The
setup of computational mesh is shown in Fig. 6.
Figure 6. The computational mesh applied in 3D simulations
performed by use of Fluent code
497
The computational mesh contains the considered
section of the damaged ship (see Fig. 2 & 4) and a
cuboid of surrounding water. The adjustment of
mesh geometry is one of the key points of CFD
modeling.
In the finite-volume method, such a quadrilateral
is commonly referred to as a “cell” and a grid point
as a “node”. In this approach, the integral form of
the conservation equations are applied to the control
volume defined by a cell to get the discrete
equations for the cell (Bhaskaran&Collins 2009).
The Fluent code is to find a solution such that
mass, momentum, energy and other relevant
quantities are being conserved for each cell. Also,
the code directly solves for values of the flow
variables at the cell centers; values at other locations
are obtained by suitable interpolation
(Bhaskaran&Collins 2009).
The Fluent code is designed to solve the
Reynolds Averaged Navier Stokes (RANS)
equations. RANS equations govern the mean
velocity and pressure. These quantities vary
smoothly in space and time, thus they can be
relatively easy to solve; however they require some
additional modeling to “close” the equations and
these models introduce significant error into the
calculation (Bhaskaran&Collins 2009).
In the course of the computation the variable time
steps were applied in order to solve the conservation
equations for mass, momentum, and volume fraction
of the liquid. A concept of a VOF model is based on
the monovalent assignment of liquid density inside
every single computational cell (Fig. 7).
Figure 7. The mixture density concept in VOF model and mesh
formulation for bottom damaged tank - at the time t=0s the oil
starts to flow out of a crack (see Fig. 6 for clarification of the
ruptured tank location)
Due to modeled fluids characteristics and
relatively low value of expected Reynolds number,
the laminar flow is applied in the course of
numerical simulations. The laminar flows are
characterized by pretty smoothly varying velocity
fields in space and time in which individual layers
move past one another without generating noticeable
cross currents. These flows arise when the fluid
viscosity is sufficiently large to damp out any
perturbations to the flow (Bhaskaran&Collins 2009).
3 RESULTS OBTAINED IN THE COURSE OF
CFD COMPUTATION - A CASE STUDY
The most straightforward attitude towards bunker oil
spills consideration is just to carry out the series of
simulations of an oil outflow. The methodology
proposed in the paper is based on a CFD modeling.
The simulations were performed on the basis of
conditions and assumptions described in the
previous sections of the paper.
The results of computations can be analyzed from
variety viewpoints. The first outcome of the
simulations is a visualization of flow patterns during
the spilling process. This provides a general outlook
on the considered phenomenon and helps to imagine
the possible course of action in case of ship hull
damage (Fig. 8).
The next result of CFD computation is a
possibility of the velocity vector field visualization
(Fig. 9). This is to facilitate the description of an oil
outflow consecutive stages and its interpretation. In
the Figure 9 the velocity vectors of the oil outflow
from a tank can be noticed. Moreover, the velocity
vectors of water flooding a tank may be also
indentified.
Figure 8. A typical flow pattern obtained for bunker oil outflow
by the use of CFD simulations
498
Figure 9. The velocity vector field visualization
However the main purpose of the CFD
simulations was to estimate the volume of the oil
spilled to the sea. It is important that the result
obtained are time dependent. Thus the maximum
allowed time for oil combating action may be
assessed. The exemplary progress of the bunker oil
outflow is shown in Figure 10. The percentage given
in the graph refers to the length of hull damage
according to the Figures 3 and 5.
Figure 10. The volume of remaining oil in a damaged bunker
tank (the initial volume of full tank was equal 700 m
3
)
The graphs plotted in Figure 10 seems to be
rather smooth but one should keep in mind that they
have a cumulative character describing the volume
of oil remaining in the damaged tank. However, the
rate of an oil spill is not a steady value while the oil
trickling and disseminating in water phenomenon is
not a stationary process.
The rate of oil discharge was computed and
plotted for a time span of carried out CFD
simulations. A number of graphs present the results
of computations. The oil outflow rate obtained for
the length of damage equals 25% of the section
length is shown in Figure 11. And respectively: for
70% - in Figure 12, 85% - Figure 13 and 100% -
Figure 14.
Figure 11. The rate of an oil outflow for the length of damage
equal to 25% of considered section’s length
Figure 12. The rate of an oil outflow for the length of damage
equal to 70% of considered section’s length
Figure 13. The rate of an oil outflow for the length of damage
equal to 85% of considered section’s length
Figure 14. The rate of an oil outflow for the length of damage
equal to 100% of considered section’s length
499
The most typical feature of the computed oil
discharge rate is its unstable character in terms of
time. The hint to the explanation of this observation
might be the flow pattern presented in Figure 8. The
visualization reveals bubble-like character of the oil
outflow resulting in variable value of the oil
discharge rate.
All the results of performed CFD computations
are rather coarse due to the adopted assumptions at
the preliminary stage of the research. Thus, the
upper limit of the Courant number was accepted
relatively high and the computational mesh was
generated not very dense. Such assumptions are
justified for a feasibility study and obviously they
will be modified for the planned main research
purposes to obtain the expected satisfactory level of
accuracy.
4 SUMMARY
The study presented in the paper is a preliminary
stage of the planned research and should be rather
found as a practical approach to the feasibility study
not the final result. However, a number of remarks
and conclusions may be drawn.
First of all the realistic possibility of an
application of CFD method to the bunker oil spill
problem is revealed. The accuracy of computation
may be improved by generation of larger size of
meshes and lowering the limit of accepted Courant
number.
From the point of view of shipping stakeholders
the key point of the study is a remark, that CFD
application enables estimation of bunker spill
characteristics at the design stage of a ship. A
variety of scenarios (different layouts of bunker
tanks) can be examined and compared against the
expected size of an oil spill.
Contrary to contemporary utilized methods, the
method presented in this paper provides a number of
advantages like time dependent characteristics of
spilled oil volume and a rate of discharge. Such data
might be useful also in the course of planning and
conducting an oil combating action.
The flexibility of presented CFD-based approach
benefits with strictly desirable proactive character of
the method which is a good prospective for future
research in the field of oil spills protection in the
marine industry.
REFERENCES
Bhaskaran R., Collins L., Introduction to CFD Basics, Cornell
University, New York, 2009, http://dragonfly.tam.cornell.
edu/teaching/mae5230-cfd-intro-notes.pdf.
Dongming Liu, Pengzhi Lin, A numerical study of three-
dimensional liquid sloshing in tanks, Journal of
Computational Physics 227 (2008), pp. 3921–3939, 2008.
Eide, Magnus S., Øyvind Endresen, Øyvind Breivik, Odd
Willy Brude, Ingrid H. Ellingsen, Kjell Røang, Jarle Hauge,
and Per Olaf Brett, Prevention of oil spill from shipping by
modelling of dynamic risk. Marine Pollution Bulletin 54,
no. 10 (October), pp. 1619-1633, 2007.
FLUENT Tutorial Guide, Fluent Inc. 2006.
Gucma L., Przywarty M. 2007. Probabilistic method of ship
navigational safety assessment on large sea areas with
consideration of oil spills possibilities. In Taerwe&Proske
(eds) 5th International Probabilistic Workshop, Ghent,
pp.349-362, 2007.
HELCOM. 1990. Study of the Risk for Accidents and the
Related Environmental Hazards from the Transportation of
Chemicals by Tankers in the Baltic Sea Area. Baltic Sea
Environment Proceedings No. 34. Helsinki Commission
1990.
IMO, Interim guidelines for approval of alternative methods of
design and construction of oil tankers. Regulation 13F(5) of
Annex I of MARPOL 73/78, Resolution MEPC.66(37),
1995.
IMO, Revised interim guidelines for the approval of alternative
methods of design and construction of oil tankers.
Regulation 13F (5) of Annex I of MARPOL 73/78
Resolution MEPC.110(49), 1993.
ITOPF, Information services / data and statistics. ITOPF.
http://www.itopf.com/information-services/data-and-
statistics/statistics/index.html#no, 2011.
Michel K., Winslow T., Cargo Ship Bunker Tanks: Designing
to Mitigate Oil Spillage. SNAME Joint California Sections
Meeting, 1999.
Nyman T. Evaluation of methods to estimate the consequence
costs of an oil spill. 7 FP report, SKEMA study, 2009.
Safetec UK. Identification of Marine Environmental High Risk
Areas (MEHRAs) in the UK. Doc. No.: ST-8639-MI-1-Rev
01. December 1999.
Tavakoli M.T., Amdahl J., Ashrafian A., Leira B.J. Analytical
predictions of oil spill from grounded cargo tankers.
Proceedings of the ASME 27
th
International Conference on
Offshore Mechanics and Arctic Engineering, Estoril,
Portugal, 2008.
Tavakoli M.T., Amdahl J., Leira B.J. Analytical and numerical
modelling of oil spill from a side damaged tank. In Ehlers
and Romanoff (eds) 5th International Conference on
Collision and Grounding of Ships, Espoo 2010.
500
