Cargo Ship Bunker Tanks: Designing to Mitigate Oil Spillage

Keith Michel, President, Herbert Engineering Corp., San Francisco, CA

Thomas S. Winslow, PE, Consultant, Oakland, CA

Collision of Alexia and Enif - Gulf of Mexico, 1995

ABSTRACT
Recent collision and grounding accidents have increased public and industry awareness of the risks of
oil spills from bunker tanks. This paper summarizes historical spill data for freighters, and provides
case histories for representative collision, allision, and grounding casualties. Arranging double hull
protection around the bunker tanks is one means for mitigating the risk of spillage. The location and
size of the fuel oil tanks also influence the likelihood and expected volume of oil spills. The relative
effectiveness of these alternatives are explored using probabilistic oil outflow analysis techniques.

May 14, 1999

SNAME Joint California Sections Meeting

INTRODUCTION
In the wake of the Exxon Valdez spill and the
subsequent passage of the Oil Pollution Act of 1990
(OPA90), there has been a dramatic reduction in the
spillage of oil from tankers in U.S. waters. During the
1990s, the annual spill volume from tankers has been less
than one-tenth of the spill volume experienced during the
1980s. At the same time, recent fuel oil spills from
freighters involved in collisions, allisions, and groundings
(Enif, 1995; Kure, 1997; and New Carissa, 1999) have
raised awareness of the risk of oil spills from bunker
tanks.
The assessment of risk involves the evaluation of the
frequency of accidents together with the consequences of
such events. A formal assessment of consequences of oil
spills should incorporate many factors including the
impact on habitat, costs incurred, and consideration of
injuries and loss of life. Such an assessment is beyond the
scope of this study, and therefore the quantity of oil
spilled is used as a surrogate for consequence.
The United States Coast Guard maintains a database
of petroleum spills occurring within the navigable waters
of the US. This database includes information on the
amount of spillage, the type of vessel involved, and
causality. The frequency and volume of spillage from
bunker tanks on freighters is estimated through analysis of
these historical data. The authors also examined six
accidents involving breaching of fuel oil tanks. These
case histories provide insight into the types of accidents
that occur, and the severity of hull damage encountered.
The use of historical statistics for assessing spill
performance does have limitations. Because oil spills
from collisions and groundings are low probability events,
there is insufficient data to compare the effectiveness of
the different bunker tank arrangements currently in use.
Furthermore, new concepts cannot be evaluated on the
basis of historical data alone. However, probabilistic
analysis utilizing historical statistics on damage extents
provides a means for calculating the relative effectiveness
of designs in mitigating the likelihood and volume of oil
spills.
In this study, the probabilistic oil outflow calculation
methodology developed by IMO to assess alternative
tanker designs is applied to various bunker tank
configurations on tankers, containerships, and bulk
carriers. The intent is to provide the designer with a
better understanding of the influence that the arrangement
and location of bunker tanks has on oil outflow from
collision and grounding accidents.

REGULATORY FRAMEWORK
There are currently no requirements for protectively
locating bunker tanks within cargo vessels. Regulations

13F of MARPOL 73/78 requires all new tankers above

5,000 DWT to have a double hull, a mid-deck, or an
alternative arrangement approved by the International
Maritime Organization (IMO). OPA 90 mandates double
hull construction for all new tank vessels calling U.S.
waters. However, both regulations apply only to cargo oil
tanks, and any fuel oil tanks located within the cargo tank
length. The cargo tank length extends from the aft-most
cargo tank boundary to the collision bulkhead. As tankers
typically have their bunker tanks arranged in the engine
room, these tanks can be located adjacent to the shell.
Restrictions on the storage of fuel oil in double
bottom spaces was first proposed at IMO by Finland in
the 1980s and more recently by Norway. Norway is also
considering an indexing system intended to encourage
environmentally friendly operations by differentiating
fees for ships. The risk of oil discharge from bunker
tanks will be factored into this environmental index. In
the wake of a spill from the wood-chip carrier New
Carissa, the US Congress has also initiated debate on the
need for enhanced regulations related to fuel oil carried
on commercial freighters.

HISTORICAL SPILL DATA

The USCG database includes reported oil spills of all
sizes occurring in U.S. navigable waters. Figure 1 shows
the annual oil spill volume from vessels since 1973. The
data is broken down into three categories: tankers, tank
barges, and other vessels. The significant reduction in
spill volumes since 1990 is readily apparent. The
performance of tankers is most impressive, with annual
spill volume less than one-tenth of the pre-1990 level.
Spill volume from tank barges average about one-third of
the pre-1990 level, and now represents the single largest
source of oil spillage. The reductions in oil spillage
realized by tankers and tank barges has not carried over to
other vessels, and in the 1990s other vessels have become
responsible for an increasing percentage of the total oil
spillage.

Millions of Gallons

Other Vessels
10

20

10

1973 1976 1979 1982 1985 1988 1991 1994 1997

Year

Figure 1 Spill Volume from Vessels in

US Waters
(for period 1973-1997)

90,000
80,000

Annual Oil Spill Volume

No. of Spills

TANKSHIPS
5%

SOURCES
7%

70,000
60,000
50,000

40,000
30,000

20,000
10,000
0

0
1992

A more detailed breakdown of spills by source

indicates that, during the period from 1992 to 1997,
vessels have been responsible for about 54% of the oil
spilled into US waters. (see Figure 2). The other vessel
category shown in Figure 1 includes freighters, freight
barges, tow and tugboats, fishing boats, unclassified
vessels, and all other vessels except tankers and tank
barges. In Figure 2, freighters are separated from other
vessels. The category of freighters includes commercial
cargo vessels such as bulk carriers, containerships, ro-ros,
and general cargo ships. During the 1992-1997 period,
freighters were responsible for about 4% of the total oil
spillage.
UNKNOWN or
OTHER
4%

100,000

1993

1994

Year

1995

1996

1997

Figure 3 Spills from Freighters in

Collision and Grounding Accidents (for
period 1992-1997)
The current level of public concern and political
reaction may seem surprising in light of the relatively low
spill rate for freighters in collision and grounding
accidents (1.33 spills per year with an average spill size of
about 112 cubic meters or 30,000 gallons) as compared to
tank barges and other vessels. However, a number of
recent spills have occurred in environmentally sensitive
regions such as the coasts of Alaska and Oregon, and
Humbolt Bay in California. Spills such as the Kure
(about 4,500 gallons spilled in Humbolt Bay) have
demonstrated that even relatively small spills can lead to
significant environmental impact and substantial cleanup
costs.

REPRESENTATIVE CASE HISTORIES

TANKBARGES Case histories for six accidents involving damage to
31% bunker tanks are summarized below. These include two

PIPELINES
13%

high energy collisions (the President Washington - Hanjin

Hong Kong and the Alexia - Enif), two allisions (the Julie
N and the Kure), and two groundings (the Kuroshima and
FREIGHTERSthe New Carissa).

FACILITIES
22%

4%

OTHER VESSELS
(excluding freighters)
14%

Figure 2 - Spill Volume in US Waters by

Source
(for period 1992-1997)

May 14, 1999

SNAME Joint California Sections Meeting

No. of Spills

Tank Barges

Volume of Oil Spillage (gallons)

Tankers

Between 1992 and 1997, freighters experienced a

total of eight oil spill events 1000 gallons or larger
50
originating from allisions, collisions, or groundings (see
Figure 3). The combined spill volume from these eight
40accidents is about 900 cubic meters (237,000 gallons).
This represents 42% of the spillage from freighters and
only about 3% of the spillage from vessels during this six30
year period.
Thousands of Cubic Meters

15

President Washington

The impact and penetration of the Alexia bow caused the

hatch covers to collapse into the hold below causing
collateral damage to a double bottom bunker tank at the
forward end of the hold.
The force of this collision and the extent of bow
penetration was so substantial that double hull protection
afforded by the outboard port bunker tank failed to protect
the diesel oil tank. Similarly, there are no practical design
options that would have prevented the collateral damage
to the port side and double bottom bunker tanks.
Julie N

Figure 4 President Washington

In May 1994, at the entrance to Pusan Harbor, the
containership Hanjin Hong Kong struck the containership
President Washington on the port side near amidships.
The bow of the Hanjin Hong Kong penetrated the side
shell of the President Washington in way of an empty
bunker tank and an adjacent ballast tank, extending about
2.5 meters beyond the longitudinal bulkhead. Two
adjacent cargo holds were flooded. There was only minor
oil pollution to the harbor from the residual HFO in the
damaged bunker tank.
The tank arrangement on the President Washington is
representative of most containerships where the
significant percentage of bunker oil is stored in wing
tanks outboard of the cargo holds. However, it is likely
that the extent of damage from the penetration of the
Hanjin Hong Kong bow would have exceeded any
practical double hull protection of the bunker tank in this
high energy collision.
Enif
In July 1995 the 230-meter bulk carrier Alexia
collided with the 157-meter bulk carrier Enif in the Gulf
of Mexico near the entrance to the Mississippi River (see
cover photo). The Alexias bow imbedded in the port side
of the Enif, just aft of amidships. It extended into No. 3
Hold, approximately half way through her beam. As a
result of the collision three bunker tanks and one diesel oil
tank on the Enif spilled approximately 360 m3 (95,000
gallons) of mixed diesel and IFO 180. There was only
bow structural damage to the Alexia with no oil spillage.
The ships were successfully separated and lightered
without additional spillage. After the third day only
sheens were reported around the Enif, and visible
evidence of the spill disappeared a few days later.
The port bunker tank and centerline diesel oil tank on
the Enif were damaged from direct contact with the
Alexia bow. The starboard bunker tank was damaged
from resultant shifting of the cargo of coiled steel plate.

Figure 5 Julie N
In September 1996 the product tanker Julie N struck
the south side of the Million Dollar Bridge in Portland,
Maine as the ship transited the draw span. Pilot error was
the cause of the accident. The contact with the bridge
buttress resulted in an oil spill of 353 m 3 (93,200 gallons)
of heavy bunker fuel and 327 m 3 (86,400 gallons) of No.
2 home heating fuel, cargo oil. The oil spill covered 13.7
miles of shoreline and led to a massive clean-up response.
Total costs reportedly approached $50 million.
The damage to the Julie N occurred below the
waterline, on the port side of the bow just aft of the
collision bulkhead. The side shell ripped open the hole
measuring approximately 10 meters in length by 4 meters
in depth. A HFO bunker tank located immediately aft of
the collision bulkhead was breached, and the bulkhead at
the forward boundary of the port cargo tank was ruptured.
The transverse penetration into the bunker tank from
the contact with the bridge buttress was limited, and there
is a good possibility that double hull protection would
have prevented this oil spill.

Kure

recovery operation ensued. This oil spill would likely

have been averted if the bunker tanks were located
outside of the double bottom spaces.
New Carissa

Figure 6 Pier at Humbolt Bay

In November 1997 the 195-meter bulk carrier Kure
contacted the pier while shifting berth at the Louisiana
Pacific Dock in Humbolt Bay. Damage to the hull
consisted of a 350mm hole about 3 meters above the
waterline in way of a forward bunker tank. About 17.2 m 3
(4537 gallons) of IFO 180 was discharged into the bay
before the hole could be plugged. The local wetlands and
shoreline were heavily impacted by the oil spill.
Double hull protection would certainly have
prevented the spill from this minor and very localized
puncture through the hull.
Kuroshima

Figure 8 Kuroshima
In November 1997, the 116-meter refrigerator ship
Kuroshima went hard aground at Summer Bay near Dutch
Harbor, Alaska. The grounding resulted in the breeching
of two double bottom bunker tanks and about 174 m 3
(46,000 gallons) of heavy fuel oil spilled. An additional
288 m3 (76,000 gallons) of HFO was pumped from the
ship to holding tanks ashore to prevent further spillage
and to lighten the ship. The salvage effort took three
months to free the ship, and a costly oil cleanup and

May 14, 1999

Figure 7 New Carissa

In early February 1999 the wood chip bulk carrier
New Carissa drifted aground off the central Oregon coast.
Initially, though hard aground on a sand bottom, there was
no known oil spill. As storm seas pounded the ship
against the bottom, oil began to leak from the ship, and
pollute the nearby coastline. Bunker fuel was located in
three centerline double bottom tanks below Cargo Holds
No. 2 to No. 4, and an additional double bottom tank on
the portside below Cargo Hold No.5. Diesel oil was
stored in the starboard double bottom tank across from the
No. 5 DB. At the time of the grounding the ship had
approximately 60% bunkers on board, consisting of about
1,363 m3 (360,000gallons) of HFO and 114 m3 (30,000
gallons) of diesel oil. It is difficult to know how much
HFO escaped from the grounded vessel, and how much
burned-off during the salvage operation. Estimates of
HFO spillage range from 189 m3 (50,000 gallons) to 265
m3 (70,000 gallons). To date, salvage and oil spill cleanup costs exceed $20 million.
The bunker tank arrangement on the New Carissa
was typical of many bulk carriers where bunker oil is
predominantly stored in double bottom tanks below the
cargo holds. However, it is uncertain whether alternative
bunker tank arrangements would have averted this spill.
The structural failure and breaking open of the vessel
would probably have opened up any tanks in the midships
region of the vessel.

STANDARD DESIGN PRACTICES

Design considerations lead to different bunker tank
arrangements for different ship types.

SNAME Joint California Sections Meeting

the capacity of any one tank does not generally exceed

1000 m3 (264,000 gallons). In comparison, VLCCs have
tanks as large as 3,400 m3 (898,000 gallons), as the HFO
is typically allocated to one or two pairs of ER wing
tanks.

Tankers: The HFO tanks are usually

arranged in one or two pairs of wing
tanks (see arrangements T1 and T2 in
Figure 10). This allows for short piping
runs, and avoids passing HFO piping
through ballast and cargo tanks. The
double-hulled spaces forward of the
engine room are dedicated to cargo oil,
maximizing cargo cubic.

OUTFLOW ANALYSIS OF ALTERNATIVE

BUNKER TANK ARRANGEMENTS
To assess the relative effectiveness of alternative
arrangements for protectively locating bunker tanks,
probabilistic oil outflow calculations were carried out for
tankers, containerships, and bulk carriers.

Containerships: Typically, the majority

of HFO is allocated to wing tanks
outboard of the cargo holds. These
tanks are distributed longitudinally
through the midship region, such that
bunkering or consuming fuel oil does
not significantly alter trim or stability
(see C1 of Figure 11). Additionally,
there will be some bunker oil storage in
engine room wing tanks.

Outflow Calculation Methodology

The IMO guidelines for approval of alternative tanker
designs [1] and the draft regulation for evaluating
accidental outflow for new tankers [2] contain a
probabilistic-based procedure for assessing 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.
Outflow parameters are developed by combining results
from all damage cases:

Bulk Carriers: Capesize bulk carriers

usually carry their fuel oil in engine
room wing tanks similar to tankers. For
the smaller Handysize or Panamax
ships, HFO is most commonly allocated
to center double bottom tanks.
Alternatively, bulk carriers may have
HFO in the outboard double
bottom/wing tanks, or arranged in deep
tanks forward together with engine
room tanks (see B1 of Figure 12).
Tankers
Description
DWT (MT)
HFO (m3)
DO (m3)

Panamax
50,000
1700
220

Aframax
90,000
2900
320

Suezmax
150,000
3800
370

Containerships
Description
750 TEU
DWT (MT)
9,000
HFO (m3)
700
DO (m3)
130

1500 TEU
20,000
2000
200

Panamax
45,000
5600
330

Bulk Carriers
Description Handysize
DWT (MT)
30,000
HFO (m3)
1300
DO (m3)
130

Panamax
70,000
2200
270

CapeSIze
160,000
4000
300

Table 1 Typical Bunker Capacities

VLCC
285,000
7500
400

The probability of zero outflow (Po)

represents the likelihood that no oil will
be released into the environment, given
a collision or grounding casualty which
breaches the outer hull.

The mean outflow parameter (Om) is

the non-dimensionalized mean or
expected outflow.
Post-Pmax References [3], [4], and [5] provide further background
75,000
on the calculation procedures and assumptions. Although
7600
originally developed for the purposes of evaluating
430
alternative tanker designs, the calculation methodology
provides a rational means for comparing outflow
performance of alternative bunker tank arrangements.
The methodology described in the IMO draft
regulation on accidental outflow was applied in these
calculations, with the following adjustments to
accommodate the analysis of bunker tanks.

Table 1 summarized typical capacities for HFO and

DO tanks for various sizes of tankers, containerships, and
bulk carriers.
The high-powered post-Panamax
containerships have the largest HFO storage requirements,
with total HFO capacity for recent newbuildings
exceeding 7,600 m3 (2 million gallons). The HFO is
usually distributed in a number of wing tanks, such that

The IMO procedure assumes all cargo

tanks are 98% full. When evaluating
the bunker tank outflow, three
independent sets of calculations were
run, assuming bunker tanks were 98%
full, 54% full, and 10% full. The
outflow results from these three sets of

calculations were then combined in a

ratio of .25:.50:.25, to simulate the
consumption of fuel oil during the
course of the voyage.

Presentation of Outflow Results

The IMO methodology assumes the vessel has been
in a collision or grounding accident of sufficient energy to
penetrate the outer hull. If the probability of zero outflow
(Po) equals 0.80, then 80% of such collisions and
groundings only penetrate into spaces which do not carry
oil, and therefore 80% of the cases have zero outflow.
Since the mean outflow is the weighted average for all
accidents (whether spillage occurs or not), the average
spill size is the mean outflow divided by (1- Po).
If the probability of zero outflow is 0.80, it then
follows that 20% or 20 out of 100 accidents will be spill
events. The combined outflow from these 20 spills is the
mean outflow multiplied by 100.
The calculation results for the probability of zero
outflow and mean outflow are presented for each tank
configuration, as well as spill frequency (number of spills
per 100 accidents) and total outflow per 100 accidents.
The reader may find it easier to refer to the spill
frequency tables when comparing designs.

The IMO procedure specifies the

calculation of bottom damage outflow
based on hydrostatic balance principles.
A minimum outflow equal to 1% of the
tank capacity is assumed for oil tanks
bounding the lower shell.
This
provision accounts for losses from
initial impact and dynamic effects such
as current and ship motions, for designs
like the mid-deck tanker having tanks
initially in
hydrostatic
balance.
However, 1% of tank capacity is not
adequate to cover losses from a double
bottom tank, as studies indicate that a
water bottom of up to 1 meter will be
introduced by a 3 knot current [4]. For
this study, the 1 meter waterbottom is
assumed, resulting in losses of 50%,
4%, and 1% of the tank capacity when
the initial tank filling levels are 98%,
54%, and 10% respectively.

Outflow Analysis for Tankers

Figure 10 illustrates the five bunker tank
arrangements evaluated for tankers. Calculations were
run for a 280,000 tons deadweight VLCC, with a 5 long x
3 wide cargo tank configuration. A heavy fuel oil
capacity of 7,500 m3 (including service and settling tanks)
is assumed for all configurations.

The probability distribution function for

the longitudinal location of side damage
for tankers assumes a homogeneous
distribution over the ships length. IMO
data for cargo and passenger vessels
used for developing the damage
stability regulation indicates an
increased likelihood of damage forward
(see Figure 9). The cargo vessel
distribution is applied for analysis of the
bulk carriers and containerships.

T1 and T2: These are the most common

configurations, with one or two pairs of
engine room wing tanks.

T3: This arrangement was developed

by BP Shipping, and installed on a
number of Suezmax tankers [6]. An oiltight longitudinal bulkhead divides the
bunker tank into two approximately
equal sized compartments. A minimum
3.0 meter clearance is assumed between
the side shell and the inboard tank. By
drawing fuel from the outer tanks as a
first priority, the outer tanks effectively
serve as voids once they are empty. In
this way, double hull protection is
provided to the fuel oil over a
substantial portion of the vessels
operating life.

T4: Inboard fuel oil tanks are arranged

above the pump room spaces, to
supplement the wing tank storage.

T5: A minimum 2.0 meter wide void

space is arranged outboard of the
bunker tanks in this double hull
configuration.

1.2

Probability Density

Tankers
0.8

0.6

Other Vessels

0.4

0.2

0
0

0.2

0.4

0.6

0.8

Longitudinal Location / Ship Length

Figure 9 Probability Distribution

Functions
for Longitudinal Position of Side
Damage

May 14, 1999

SNAME Joint California Sections Meeting

T1
T2
T3
T4
T5

T1

ENGINE RM

Figure 10 Tanker Bunker Tank

Configurations

T1
T2
T3
T4
T5

No. of Oil Spills

per 100 accidents
4.9
4.1
2.1
3.8
1.9

Total Quantity of
Oil Spilled (m3)
per 100 accidents
10,000
8,800
3,700
5,300
4,100

Table 2 VLCC Projected Spill

Frequency
(with Cargo Tanks empty)

T1
T2
T3
T4
T5

T1
T2
T3
T4
T5

Mean Outflow (m3)

Collision Grounding Combined
4,914
2,800
3,646
4,885
2,800
3,634
4,756
2,801
3,583
4,804
2,796
3,599
4,767
2,801
3,587

Probability of Zero Outflow

Collision Grounding Combined
0.877
1.000
0.951
0.898
1.000
0.959
0.948
0.999
0.979
0.907
0.999
0.962
0.952
1.000
0.981

Table 5 VLCC Outflow Parameters

(with Cargo Tanks 98% Full)
Table 2 and Table 3 contain the projected outflow
assuming the cargo oil tanks are empty at the time of the
casualty. Table 4 and Table 5 correspond to the full load
condition, with all cargo tanks 98% full.
As previously discussed, the bunker tanks are
evaluated at three filling levels (10%, 54% and 98% full),
and results combined in order to simulate fuel oil
consumption during the voyage. For configuration T2,
fuel is first drawn from the forward tanks.
For
configurations T3 and T4, fuel is first drawn from the
outer tanks.
Findings related to outflow from tankers:

There is minimal risk of pollution from

grounding, as the lower edge of the
bunker tanks on tankers are typically
located 25% to 30% of the depth above
baseline. This is illustrated by the low
mean outflow figures for the grounding
condition in Table 3.

Splitting the tanks fore and aft (T2

compared to T1) has relatively little
impact on the frequency of spills (4.9 vs
4.1 per 100 events) and the volume of
spillage (12% reduction). This is not
unexpected due to the short length of
these tanks there is a high probability

T1
T2
T3
T4
T5

T1
T2
T3
T4
T5

Probability of Zero Outflow

Collision Grounding Combined
0.699
0.781
0.748
0.699
0.781
0.748
0.770
0.781
0.777
0.726
0.781
0.759
0.774
0.781
0.778

T5

T4

Mean Outflow (m )
Collision Grounding Combined
242
5
100
213
5
88
84
6
37
132
1
53
95
6
41

Table 3 VLCC Outflow Parameters

(with Cargo Tanks empty)

Total Quantity of
Oil Spilled (m3)
per 100 accidents
364,600
363,400
358,300
359,900
358,700

Table 4 VLCC Projected Spill

Frequency
(with Cargo Tanks 98% Full)
T3

T2

No. of Oil Spills

per 100 accidents
25.2
25.2
22.3
24.1
22.2

that both tanks will be damaged in a

collision.

Fitting a longitudinal bulkhead through

the wing tank (configuration T3) has a
significant impact on both the frequency
of spills and the total spillage volume.
It should be recognized that this
improvement is only realized if the
outer tanks are emptied prior to using
the inboard tanks.

Fitting
small
inboard
tanks
(configuration T4) has a lesser effect on
the frequency of spills, but does reduce
spill volume by nearly half. Again, the
outer tanks must be emptied first if this
gain is to be fully realized.

Providing a 2.0 meter void space

outboard of the wing tank (T5
compared to T1) has the greatest impact
on the frequency of spills (4.9 vs 1.9 per
100 events).

distributed longitudinally in alternate

wing tanks as well as engine room wing
tanks.

C2: This configuration is similar to C1,

except that the adjacent wing tanks are
selected for the HFO.

C3: HFO is allocated in transverse deep

tanks located between the cargo holds.
Both the deep tanks and the engine
room wing tanks are segregated from
the shell by either ballast or void
spaces.

A comparison of Table 3 (cargo tanks

empty) and Table 4 (cargo tanks 98%
full) outflow data reveals that bunker
tanks are responsible for less than 3% of
the spill volume from tankers.
Authors observations and comments:

Subdividing the bunker tanks on tankers

is an effective means for mitigating
outflow, particularly when inner and
outer tanks are arranged.

Providing void spaces outboard of the

bunker tanks more than halves both the
frequency of spills and the expected
quantity of oil spillage. However, the
additional cost for this arrangement is
significant (approximately $500,000 on
a VLCC). Recognizing that potential
reduction in outflow represents less than
2% of the expected outflow from the
cargo tanks, it may be more cost
effective to take additional measures to
mitigate cargo oil spillage.

Outflow Analysis for Containerships

Figure 11 illustrates the three bunker tank
arrangements evaluated for containerships. Calculations
were run for a post-Panamax vessel of approximately
75,000 tons deadweight. A heavy fuel oil capacity of
7,600 m3 (including service and settling tanks) is assumed
for all configurations.

C1

C2

C3
Figure 11 Containership Bunker Tank
Configurations
Table 6 and Table 7 contain the projected outflow for
the three containership configurations. The outflow
analysis assumes the fuel oil is comsumed proportionately
from all tanks (i.e. in the 10% arrival condition, each tank
is assumed 10% full).

C1
C2
C3

No. of Oil Spills

per 100 accidents
19.7
16.4
8.3

Total Quantity of
Oil Spilled (m3)
per 100 accidents
11,200
11,400
3,400

Table 6 Containership Projected Spill

Frequency

C1:
This is the most common
configuration, with HFO storage

May 14, 1999

SNAME Joint California Sections Meeting

C1
C2
C3

Probability of Zero Outflow

Collision Grounding Combined
0.573
0.957
0.803
0.640
0.966
0.836
0.920
0.915
0.917

C1
C2
C3

Mean Outflow (m3)

Collision Grounding Combined
263
11
112
269
10
114
42
28
34

Outflow Analysis for Bulk Carriers

Figure 12 illustrates the five bunker tank
arrangements evaluated for bulk carriers. Calculations
were run for a Panamax vessel of approximately 70,000
tons deadweight, with a heavy fuel oil capacity of 2,200
m3 (including service and settling tanks).

B1: HFO is arranged in a pair of deep

tanks forward of No. 1 Hold and a pair
of engine room wing tanks. A double
bottom is arranged under the forward
deep tanks, which is standard practice
for this type of arrangement.

B2: This configuration is similar to B1,

except that 2 meter wide void spaces are
arranged outboard of all fuel tanks.

B3: All HFO is allocated to two pairs

of engine room wing tanks.

B4: HFO is allocated to three centerline

double bottom tanks.

B5: HFO is allocated to three pairs of

DB/wing ballast tanks.

Table 7 Containership Outflow

Parameters
Findings related to outflow from containerships:

Outflow from grounding accidents is

relatively low in all cases, as all bunker
tanks located above the inner bottom.
The probability of zero outflow for side
damage configuration C1 (0.573) is
quite low due to the longitudinal
distribution of the tanks.
Although
configuration C2 has a somewhat higher
probability of zero outflow for side
damage (0.640 vs. 0.573), the mean
outflow is actually higher (269 vs 263
m3).

Allocating the HFO to inboard deep

tanks per configuration C3 significantly
reduces both the likelihood of a spill
and quantity of oil spilled.
Authors observations and comments:

Configuration C1 provides operational

advantages with regard to the control of
trim, shear forces, and bending
moments, and .is representative of
industry practice. Grouping of tanks
(configuration C2) offers no significant
environmental benefits as compared to
C1.

Although configuration C3 provides

improves outflow performance, this is a
very costly solution. To have sufficient
HFO capacity within the deep tanks the
vessel must be lengthened by
approximately 6 meters. A cost-benefit
analysis should be carried out to further
assess this option.

B1

B2

B1
B2
B3
B4
B5

Probability of Zero Outflow

Collision Grounding Combined
0.852
0.914
0.889
0.968
0.912
0.934
0.921
0.990
0.962
0.999
0.514
0.708
0.968
0.514
0.696

B1
B2
B3
B4
B5

Mean Outflow (m3)

Collision Grounding Combined
40
5
19
9
5
7
32
1
13
1
77
46
25
42
35

Table 9 Bulk Carrier Outflow

Parameters
Findings related to outflow from bulk carriers:

B3

Configuration B3, with all tanks located

in the engine room, provides the best
outflow performance. These tanks are
confined to a short length of the ship,
reducing the probability of penetration
in collisions. Breaching the tanks in a
grounding scenario is very unlikely, and
they are located aft and above the inner
bottom.

The forward deep tanks in configuration

B1 are susceptible to damage from both
collisions and groundings. Even when
double hull protection is arranged
outboard of the bunker tanks
(configuration B2), the mean outflow is
higher as compared to configuration B3
with all HFO storage in engine room
wing tanks.

The
double
bottom
tankage
arrangements, configurations B4 and
B5,
had
the
poorest
outflow
performance. It is interesting to note
that configuration B4 with P/S double
bottom tanks has a slightly lower mean
outflow compared to the center double
bottom arrangement B5.
This is
because
the
higher
outflows
experienced by B4 for side damage are
offset by reduced outflows from bottom
damage. The large center DB tanks in
configuration B4 have a high
probability of damage, and because of
their size spill more oil than the small
wing tanks of configuration B5.

B4

B5
Figure 12 Bulk Carrier Bunker Tank
Configurations
Table 8 and Table 9 contain the projected outflow for
the five bulk carrier configurations. The outflow analysis
for bulk carriers assumes the fuel oil is consumed
proportionately from all tanks.

B1
B2
B3
B4
B5

No. of Oil Spills

per 100 accidents
11.1
6.6
3.8
29.2
30.4

Total Quantity of
Oil Spilled (m3)
per 100 accidents
1,900
700
1,300
4,600
3,500

Table 8 Bulk Carrier Projected Spill

Frequency

May 14, 1999

SNAME Joint California Sections Meeting

11

Authors observations and comments:

With regard to the issue of protectively locating

bunker tanks, the following general conclusions and
recommendations are offered:

As compared to double bottom tanks

and forward deep tanks, allocation of
bunkers to engine room tanks offers
significant advantages with regard to oil
outflow performance.
Double bottom tanks generally have
poor outflow characteristics, and should
be avoided when possible.

CONCLUSIONS AND RECOMMENDATIONS

Figure 13 illustrates the steps leading to oil spillage.
Generally speaking, the higher up the tree corrective
measures are implemented, the more effective they will
be. Evidence of this is the dramatic improvement in
tanker spill performance in the 1990s. We can surmise
that this reduction in spillage is primarily a result of
improved operational and management procedures, as
these improvements were realized prior to implementation
of regulations such as the double hull provision of
OPA 90 and escort tug requirements.
Thus, the preferred approach is to eliminate incidents
(1st order effects) and eliminate accidents (2 nd order
effects). However, third order effects such as protectively
locating tanks to prevent spillage or mitigate outflow in
the event of an accident does offer a last line of defense.
OIL SPILL
SOURCE
A (1st order effects)

B (2nd order effects)

ACCIDENT
OCCURS
C (3rd order effects)
NO SPILL
OCCURS

The location and size of bunker tanks

has a significant impact on outflow
performance, and should be carefully
considered during the design process.

Double bottom tanks are particularly

susceptible to damage and, when
practical, should be avoided

Providing double hull protection for

bunker tanks reduces both the number
of spills and the quantity of outflow, but
comes at a cost. This cost is especially
high for containerships and smaller bulk
carriers and tankers, as the size of the
ship must be increased.

Before
implementing
outflow
regulations on bunker tanks, costbenefit analyses should be carried out to
ascertain the relative cost effectiveness
of double hull protection and other
options.

The development of bunker tank

arrangements for new vessels requires
careful
consideration
of
many
operational issues.
Any outflow
regulations pertaining to bunker tanks
should be performance-based, allowing
optimization of the design.

REFERENCES

INCIDENT
OCCURS

NO
ACCIDENTS

SPILL
OCCURS
D (4th order effects)

OIL
RECOVERED

Figure 13 - Spill Event Tree

OIL NOT
RECOVERED

1IMO, Interim Guidelines for Approval of Alternative Methods of Design and Construction of Oil Tankers under
Regulation 13F(5) of Annex I of MARPOL 73/78, Resolution MEPC.66(37), Adopted September 14, 1995.
2 IMO, BLG3/WP3, Report of the working group at BLG 3 on Revision of MARPOL Regulations I/22 to 24 in the Light
of the Probabilistic Methodology for Oil Outflow Analysis, including a draft of proposed MARPOL Regulation 19,
Accidental Outflow Performance, July, 1998.
3 Sirkar et al, A Framework for Assessing the Environmental Performance of Tankers in Accidental Groundings and
Collisions, SNAME TRANSACTIONS, 1997.
4 Michel, Oil Outflow Analysis of Double Hull Tankers (Volumes 1 and 2), Prepared under Contract DTCG23-95-DHMT001, by Herbert Engineering Corp., for the U.S. Coast Guard, January 1997.
5 Michel, Moore, & Tagg, A Simplified Methodology for Evaluating Alternative Tanker Configurations, Journal of
Marine Science and Technology, Volume 1 Number 4, SNAJ, September 1996.
6 Motor Ship, Innovative design for British trio, December, 1997.