415
1 INTRODUCTION
In recent times, marine diesel engines used as the
mainpropulsionsourceofamodernvessels.Marine
dieselengineconsistsofvarioussub‐systems,thefuel
oilsystemisthemostimportantone.However,many
accidents happen due to the failure of a fuel oil
system of the
main propulsion engine (Islam et al.,
2017a, Islam et al., 2016, Islam et al., 2018a).
Inappropriatemaintenanceisoneofthemaincauses
forthefailureofmarineengineson‐board.Thereisa
lack of appropriate failure data that could help
developing informed maintenance plan (Islam and
Yu, 2018).
The engine manufacturers provide
informationforcarryingoutplannedmaintenanceat
specified running hours, without taking into
considerationcurrentcondition.Thisstudyisfocused
on collecting relevant data from various sources to
identify the most appropriate failure model
representingaspecificcomponentoftheengine.This
study aims to use the
structured questionnaire to
collect the FRH data for the fuel oil system of the
main propulsion engine. Moreover, this study
identifiesthefailure‐pronecomponents.Thecollected
data will be extremely valuable for developing the
reliabilityassessmenttechniques andhelptomakea
better maintenance plan of a fuel oil
system of the
main engine. The collected data and developed
Reliability Assessment of a Main Propulsion Engine Fuel
Oil System- What are the Failure-prone Components?
R.Islam&M.Anantharaman
A
ustralianMaritimeCollege,UniversityofTasmania,Launceston,Australia
F.Khan
M
emorialUniversityofNewfoundland,St.John’s,NL,Canada
V.Garaniya
A
ustralianMaritimeCollege,UniversityofTasmania,Launceston,Australia
ABSTRACT:Themainpropulsionengineistheheartofavesselwhichcarriestheentireloadoftheshipand
propels to move ahead. The main engine consists of various sub‐systems, the fuel oil system is the most
importantone.Fueloil
systemprovidesfueltotheengineviaafuelinjectormountedontheenginecylinder
head.Duringthevoyage,themainengineofashipencountersavariationinloadsandstressesduetorough
weathertoharshmanoeuvring,whichsometimesleadstothebreakdownofthemainengine.Fuel
oilsystems
areidentifiedasoneofthemainreasonsforenginebreakdown.Manyaccidentshappenedduetothefailureof
themainenginefueloilsysteminthelasttwodecades.Toensuresafeandreliablemainpropulsionengine
operation, it is required to assess the reliability ofafuel
oil system. However,there is a significant lack of
appropriate data to develop the reliability assessment techniques for fuel oil system. This study proposes
appropriate data collection and analysis procedure for the reliability assessment of a fuel oil system. Data
relatedtoFailureRunningHours(FRH)ofafueloil
systemiscollectedfrom101experiencedmarineengineers
throughaquestionnaire. Thecollected dataprocessedusing abox plotand analysedfor a normalitytest. It
helpstoidentifythegeneralizationofthedata.Moreover,thisstudyidentifiedfailure‐pronecomponentsofa
fuel oil system. The collected data will help
in developing reliability assessment techniques for accurate
reliabilityanalysisofafueloilsystem.Theidentifiedfailure‐pronecomponentswillassistinfuturereliability
analysisandriskmitigationstrategiesforimprovingtheoverallsafetyandreliabilityoftheshippingindustry.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 13
Number 2
June 2019
DOI:10.12716/1001.13.02.20
416
reliability assessment techniques will be very useful
toassessthereliability ofthe marine enginesand to
plantheeffectivemaintenanceactivitieson‐boardthe
ship.Asaresult,thiswillhelpinimprovingthesafety
ofmarineengines.
This paper comprises five sections. Section 1
represents the introductory information
about the
study. Section 2briefly presents the structure of the
questionnaireandtheselectionoftherespondentsfor
conductingthesurvey.Section3presentsastatistical
analysis of thecollected data. Section 4 presents the
mainfindingofthestudywhileSection5presentsthe
conclusions.
2 QUESTIONNAIRESTRUCTURE
AND
RESPONDENTSSELECTION
The structure of the questionnaire discussed in this
section to better understand the responses. The fuel
oilsystem ofthe main propulsionengine consists of
13components,a)FuelOilSuctionFilterb)Fuel Oil
Supply Pump c) Booster Pump d) Fuel Oil Main
Discharge Filter e) Fuel
Oil Bypass Filter f) FuelOil
Heater g) Viscotherm h) Fuel Oil Injection Pump i)
Fuel Oil Injector j) Fuel Oil High Pressure Pipe k)
Buffer Tank l) Service Tank m) Flow Meter. These
componentsplayanimportantroletooperatethefuel
oil system effectively. The reliability of a
fuel oil
system is a product of all the 13 components. The
structureofthequestionnairepresentedinFigure1.
The structured questionnaire used, and an open‐
ended question asked from the respondents. The
developedquestionnaireispresentedinTable1.
There are two questions in the questionnaire,
question 1 seeks
a response to identify the type of
engine and its model. Moreover, question 2 to
pursuesto know the FRHofeach component of the
fueloilsystemofthemainpropulsionengine.
To collect the responses many marine engineers
identified from the marine industry. The potential
respondents selected based
on their experience on‐
board ship as a marine maintenance engineer.
Particularly,chief,asecondandthirdengineerwithat
least 3 years of maintenance experience on‐board
ships are selected as a respondent. As they are the
bestpersonto askthequestion whohasthe real‐life
experienced on
‐board. Moreover, these respondents
have the experience to work in various type of ship
(i.e.oil tanker, containership, passenger ships, ferry
etc).Furthermore,therespondentshaveanexperience
conductingmaintenanceactivitiesinseveraltypesof
marine diesel engines (i.e. two stroke, four stroke,
highspeed,mediumspeed,lowspeed
etc.
Thequestionnairesendsto200experiencedmarine
engineers around the world through e‐mail.
However, 101 responses received from the
respondents. Based on the statistical validation by
Islametal.(2017a),(Islametal.,2018a,IslamandYu,
2018, Islam et al., 2018b) about 96 sample size is
required to
justify the accuracy of the collected
responses.Therefore,thecollectedresponsesfrom101
experiencedmarineengineersconfirmthevalidityof
enoughresponsesforthisstudy.
3 STATISTICALANALYSISOFTHERESPONSES
Statisticalanalysisisrequiredtoprocessthecollected
data to develop the methodology for the reliability
assessment techniques. Moreover,
it is also required
to identify the generalization of the collected data.
Therefore, after collecting the responses, a box plot
wasdrawntoeliminatetheoutlayersoftheFRHdata
forall13componentsofafueloilsystem.Figure2a)
representstheboxplotforthefueloil
suctionfilterof
the main engine fuel oil system. The out layers
eliminateconsideringa95%confidenceintervalforall
thecomponents.
Figure1.StructureoftheQuestionnaire
417
Table1.QuestionnairetocollecttheFRHdata
a)
2000
1500
1000
500
0
Failure Running Hours (FRH)
a)
b)
10000010000100010010
99.5
99
95
80
50
20
5
2
1
0.01
Shape 1.481
Scale 11008
N96
AD 1.019
P-Value <0.010
Failure Running Hours (FRH)
Percent
Weibull - 95%
b)
Figure2:a)BoxplotforFueloilsuctionfilterandb)Weibull
plotforfueloilsupplypump
Figure2b)representstheWeibullplotofafueloil
suction filter. The Weibull plots were drawn for all
thecomponentsforafueloilsystem.Theplotclearly
showsthat mostofthecollecteddatadoesn’tfollow
the normal distribution nor Weibull considering a
95%confidenceinterval.Asthe
“P”valueislessthan
the0.05.
4 RESULTS
The results of this study presented in Figure 3 and
Figure 4. The FRH data presented in Figure 3. It is
very difficult to present a large number of data in
typicalgraphplotwhichmaylookclumsy.Therefore,
FRHdata
presentedinaFrequencyplotratherusing
a typical graph plot for better presentation. The
collectedFRHdataforallthecomponentsofthemain
engine fuel oil system presented in Figure 3. These
data can be used directly to develop reliability
assessmenttechniquesforthefueloilsystem.
1000800600400200
40
30
20
10
0
Failure Running Hours (FRH)
Frequency
a)
2800024000200001600012000800040000
20
15
10
5
0
Failure Running Hours (FRH)
Frequency
b)
280002400020000160001200080004000
25
20
15
10
5
0
Failure Running Hours (FRH)
Frequency
c)
2000160012008004000
20
15
10
5
0
Failure Running Hours (FRH)
Frequency
d)
418
37503000225015007500
25
20
15
10
5
0
Failure Running Hours (FRH)
Frequency
e)
3000024000180001200060000
25
20
15
10
5
0
Failure Running Hours (FRH)
Frequency
f)
400003000020000100000
30
25
20
15
10
5
0
Failure Running Hours (FRH)
Frequency
g)
3000024000180001200060000
20
15
10
5
0
Failure Running Hours (FRH)
Frequency
h)
15000120009000600030000
25
20
15
10
5
0
Failure Running Hours (FRH)
Frequency
i)
6000050000400003000020000100000
50
40
30
20
10
0
Failure Running Hours (FRH)
Frequency
j
)
6000050000400003000020000100000
50
40
30
20
10
0
Buffer Tank
Frequency
k)
400003200024000160008000
40
30
20
10
0
Srvice Tank
Frequency
l)
80006000400020000
20
15
10
5
0
Flowmeter
Frequency
m)
Figure3. Frequency plot of Fuel Oil System: a) Fuel Oil
SuctionFilter,b)FuelOilSupplyPump,(c)BoosterPump,
(d)FuelOilMainDischargeFilter,(e)FuelOilBypassFilter,
(f) Fuel Oil Heater, g) Viscotherm, h) Fuel Oil Injection
Pump,i)FuelOilInjector,j)FuelOilHigh
PressurePipe,k)
BufferTank,l)ServiceTank,m)FlowMeter.
Figure4illustratesthemean,medianandstandard
deviationofFRHforallthecomponentsofafueloil
system.
419
Figure4.Mean,MedianandStandardDeviationofFRHforaFuelOilSystem
TheresultsofthisstudyidentifyFRHforfueloil
filtersislowerthanothercomponentsofthesystem.It
is preferable to have higher FRH for all the
componentsofafueloilsystem,whichleadstohigher
systemreliabilityandviceversa.However,thisstudy
indicates fuel oil filters
are relatively less reliable
comparedtoothercomponentsofthesystem.Dueto
themaintenanceofthefiltersisaverycriticalaspect
of the fuel oil system, (Knowles and Baglee, 2012).
Previous studies by Islam et al. (2016), Islam et al.
(2017b) also found similar results earlier. Moreover,
FRH
for fuel oil supply pumps, fuel oil booster
pumps, viscotherm, fuel oil heater, fuel injection
pump, and fuel injectors are relatively moderate.
Furthermore,FRHforfueloilservicetankandbuffer
tanks arehigh. However, fuel oil servicetank and a
buffer tank are the most reliable components of the
system.
Fueloilservicetankandbuffertankisapart
of the fuel oil system. However, they are static
components and not subject to any dynamic load.
Therefore, they require minimum maintenance.
Hence,theyhavehigherFRH.
5 CONCLUSIONS
Thisstudyprovides usefulFRHdataforthe fueloil
system
ofthemain propulsion engine. The collected
dataisuniqueinthisfieldofstudy.Thesedataenable
generalizationandthewillhelptodevelopreliability
analysistechniques.Theresultsofthisstudyindicate,
notallthedataforacomponentofamainpropulsion
engines’fuel oilsystems follow the
same behaviour.
Thestudyidentifiedfueloilfiltersasamostfailure‐
pronecomponentsandfueloilservicetankandbuffer
tank as a lowestfailure prone component. Based on
theresultsofthisstudymaintenancetaskofthefuel
oilfilterscanprioritizewhichwillhelptoincreasethe
reliabilityoffueloilfilters.Moreover,itisrequiredto
give consideration for installation of an additional
fueloilfilterwhichcouldbeusefulinroughweather
condition.Interns,itwillenhancethereliabilityofthe
fueloilsystemofthemainpropulsionengine.
REFERENCES
ISLAM, R., ABBASSI, R., GARANIYA, V. & KHAN, F.
2017a. Development of a human reliability assessment
420
techniqueforthemaintenanceproceduresofmarineand
offshore operations. Journal of Loss Prevention in the
ProcessIndustries,50,416‐428.
ISLAM, R., ABBASSI, R., GARANIYA, V. & KHAN, F. I.
2016.Determinationofhumanerrorprobabilitiesforthe
maintenance operations of marine engines. Journal of
ShipProductionandDesign,
32,226‐234.
ISLAM, R., KHAN, F., ABBASSI, R. & GARANIYA, V.
2018a. Human error assessment during maintenance
operations of marine systems–What are the effective
environmentalfactors?Safetyscience,107,85‐98.
ISLAM, R., KHAN, F., ABBASSI, R. & GARANIYA, V.
2018b. Human error probability assessment during
maintenance activities of
marine systems. Safety and
healthatwork,9,42‐52.
ISLAM, R. & YU, H. 2018. Human Factors in Marine and
Offshore Systems. Methods in Chemical Process Safety.
Elsevier.
ISLAM,R.,YU,H.,ABBASSI,R.,GARANIYA,V.&KHAN,
F.2017b.Developmentofamonographforhumanerror
likelihood assessment
in marine operations. Safety
science,91,33‐39.
KNOWLES, M. & BAGLEE, D. Condition management of
marinelubeoilandtheroleofintelligentsensorsystems
in diagnostics.25th International Congress on
Condition Monitoring and Diagnostic Engineering,
COMADEM 2012, June 18, 2012‐June 20, 2012, 2012
Huddersfield, United Kingdom. Institute
of Physics
Publishing.