521
1 INTRODUCTION
Fromtheviewpointofearthenvironmentprotection,
the shipping industry is required to develop and
improve energy saving ship operation technologies.
Forexample,theweatherroutingisoneoftheabove
technologies and it is often used for the navigation
planning of oceangoing merchant ships in order
to
minimizethedistancetraveledandfuelconsumption.
Theeffectivenessoftheweatherroutingontheenergy
saving mainly depends on the accuracy of the
weather forecast data and that of the propulsion
performance prediction in the actual sea where the
effectofwindandwaveonship’smotionexists.
The
weatherforecasttechnologyhasbeenimprovedyear
byyearandaneasiermethodtoobtaintheworldwide
accurate weather forecast data has been proposed
(Yokoi 2010). On the other hand, the propulsion
performance in the actual sea is usually predicted
usingtheSelf Propulsion Factors obtained by model
testsdue
tothesmallamountoffull‐scaleexperiment
dataintheactualsea(Sasaki2009,Tsujimoto2000).
In general, ship’s speed in the actual sea is low
comparedtothespeedmeasuredatthespeedtrialin
thestillwater.Inordertoimprovetheaccuracyofa
propulsionperformance
prediction,itisnecessaryto
understandtheeffectofexternaldisturbancessuchas
wind and wave on the propulsion performance
qualitatively.
It is said that the effect of the wave on the Self
Propulsion Factors are small and the propulsion
performanceunderthewaveisoftenpredictedtaking
into account both
an increase of resistance by the
wave and the propulsive efficiency reduction by a
propeller loading increase. Recently, some scholars
A Study on the Propulsion Performance in the Actual
Sea by means of Full-scale Experiments
J
.Kayano&H.Yabuki
TokyoUniversityofMarineScienceandTechnology,Tokyo,Japan
N.Sasaki
NationalMaritimeResearchInstitute,Mitaka,Japan
R.Hiwatashi
NationalInstituteforSeaTraining,Yokohama,Japan
ABSTRACT: The IMO has adopted Energy Efficiency Design Index (EEDI), Ship Energy Efficiency
Management Plan (SEEMP) and Energy Efficiency Operational Indicator (EEOI) in order to reduce GHG
emissions from international shipping. And, the shipping industry is required to develop and improve the
energy
savingshipoperationtechnologiestomeettheaboveIMOguideline.Theweatherroutingisoneofthe
energysavingnavigationtechnologiesandwidelyadoptedbyoceangoingmerchantships.Theeffectivenessof
the weather routing mainly depends on the accuracy of weather forecast data and the ship’s propulsion
performance prediction. The propulsion
performance in the actual sea is usually predicted using the Self
Propulsion Factors obtained by model tests. It is necessary to understand the propulsion performance
characteristicsintheactualseaconditionsfortheimprovementofpropulsionperformanceprediction.Fromthe
abovepointsofview,theauthorsperformedfull‐scaleexperiments
usingatrainingshipinordertoinvestigate
thepropulsionperformancecharacteristicsinthe actualsea.Thispaperdescribestheanalysisresultsonthe
characteristicsofPowerCurvesandSelfPropulsionFactorsundervariousweatherandseaconditions.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 7
Number 4
December 2013
DOI:10.12716/1001.07.04.7
522
have tried to measure the Self Propulsion Factors
under the wave conditions by means of self
propulsion tests, and the examination of the scale
effect is considered to be necessary in the above
model tests. A full‐scale experiment that includes a
thrustmeasurementseemstobenecessaryinorder
to
investigate the characteristics of the Self Propulsion
Factorsintheactualsea.
From the above points of view, the authors
conducted an experimental study using a training
ship in order to investigate the characteristics of
propulsion performance under various weather and
seaconditions.Inthestudy,theeffectsof
windsand
waves on the propulsion performance are analyzed
separatelyaccordingtothewinddirectionandwind
force.ThecharacteristicsofSelfPropulsionFactorsin
theactualseaarealsoexamined.Thispaperdescribes
the characteristics of a Power Curve (speed–BHP
curve)andwakecoefficient(
t
w
1 )intheactualsea
comparedwiththoseobtainedbymodeltests.
2 EXPERIMENTSANDDATAANALYSIS
2.1 Full‐scaleExperiment
The test ship was a 6,720 G.T. training ship Ginga
MaruandherprincipalparticularsareshowninTable
1. A precise shaft horsepower meter with a shaft
thrustload
meterproducedbytheShoyoEngineering
Co.,Ltdhasbeen installed in thetest ship. Theshaft
thrust load meter uses a high sensitivity sensor to
detect minute compressionstrains onthe shaft (SEC
powershipperformancemonitor).
Table1.Principalparticularsofthetestship
_______________________________________________
Hull
_______________________________________________
Length:Loa(m)105.00
Breadth:B(MLD,m) 17.90
Depth:D(MLD,m) 10.80
Cb0.5186
Draft:d(m)5.96
_______________________________________________
MainEngine
_______________________________________________
Diesel1set
MCR6,620kWx167rpm
_______________________________________________
Propeller(CPP)
_______________________________________________
Prop.Brade.No.4
Prop.Dia.:Dp(m) 4.30
P.R.(BradeAngle) 0.9965(24.4
o
)
_______________________________________________
The experiments were performed during her 2‐
month annual training cruises (from the middle of
JulytothemiddleofSeptember)betweenJapanand
Hawaiian Islands conducted years 2008, 2010 and
2011.Thetracksofher3trainingcruisesareshownin
Figure 1. In the experiments, the propulsion
performance data, engine operation data and
navigationdatawererecordedautomaticallyevery10
secondsusingtheLocalAreaNetworkSystem.
And, the following data were used for the
analysis; main engine handle notch, propeller
revolution,BHP (brake horsepower), torque,thrust,
ship’s position, heading, speed, wind direction and
windvelocity.
12 0
Ｅ
140
Ｅ
16 0
Ｅ
180
Ｗ
160
Ｗ
140
Ｗ
40
Ｎ
20
Ｎ
0
Ｎ
Tok y o
Ho no lu l u
Ko b e
Figure1.Trackchart
2.2 DataProcessing
In the propulsion performance analysis, the above
raw data are processed according to the following
procedure.
1 Identified first weresteady partsof theraw data
where the test ship proceeded at the steady
conditioncontinuouslyfortwo hours.Thesteady
condition is judged by the main engine
handle
notch,propellerrevolution,ship’sheading,speed,
winddirectionandvelocity.
2 The basic data are made by calculating average
values of the steady parts mentioned above and
these average values were divided into 5 wind
directions relative to the ship’s head and stern
centerlineasshowninFigure2.
Inthefigure,the
word“ObliqueHeadWind”indicates theportor
starboardbowwind.
3 Foreachofthewinddirections,themeanvaluesof
basicdatawerecalculatedseparatelyaccordingto
theBeaufortscaleandwereusedfortheanalysis.
Head Wind
30°
30°
Follow Wind
Beam Wind
Quarter Wind
Oblique
Head Wind
30°
60°
60°
Figure2.Definitionofwinddirection
2.3 PredictionMethodofSelfPropulsionFactors
Whentheactualvalue ofthrust(
T
)andtorque( Q )
are measured at a ship, the propulsive efficiency
(
D
) can be calculated using equation (1). Self
Propulsion Factors such as wake coefficient (
w
1 )
and relative rotating efficiency (
R
) can be also
calculated using the propeller characteristic curve
obtainedbythepropelleropentest.
523
()/
Ds
T
TV DHP
DHP BHP
(1)
where
s
V
=speed; DH
P
=delivered horse power;
B
H
P
=brake horse power; and
T
=transmission
efficiency.
The equations to obtain Self Propulsion Factors
usingthepropelleropentestresults,speed,propeller
revolution, BHP, thrust and torq ue are shown from
(2)to(7).TheaboveprocedureisshowninFigure3.
When predicting the Self Propulsion Factors, it is
necessary to meas ure the data accurately in
an
appropriate time interval and to know the effect of
ship’s condition such as displacement and trim on
propulsionperformance.
QN
VT
N
S
D
)60(2
(2)
42
)60( DN
T
K
T
(3)
52
)60( DN
Q
K
Q
(4)
5144.0
)60(
1
S
T
T
V
JDN
w
(5)
5144.0
)60(
1
S
Q
Q
V
JDN
w
(6)
Q
Q
R
K
K
0
(7)
where
N
=propellerrevolution.
Figure 3. Calculation procedure to obtain Self Propulsion
Fators
3 CHARACTERISTICSOFTHEPROPULSION
PERFORMANCE
3.1 PowerCurve
In order to investigate the propulsive characteristics
ofthetestship,theobtaineddataarecomparedwith
the Power Curve obtained by the power prediction
using the model ship and the one obtained by the
speedtrialofthetestshipas
showninFigure4.The
PowerCurveobtainedbythespeedtrialagreeswell
withthatbythepowerprediction.ThemeasuredBHP
values were transformed so that they were
comparable with those obtained in the model test
condition (displacement; 5,763 tons) using the 2/3
power rule based on the
idea of the Admiralty
Coefficient. The plotted data indicate mean values
calculatedaccordingtotheproceduredescribedinthe
previoussection.Therelationshipbetweenmeasured
BHP and her speed was found to be in good order
withthatoftrialresultqualitatively.
1500
2000
2500
3000
3500
4000
4500
5000
11 12 13 14 15 16 17 18 19
BHP(kW)
Speed(kts)
Head windBF<3 HeadWindBF=3 HeadWindBF=4
O.HeadwindBF<3 O.HeadWindBF=3 O.HeadWindBF=4
O.HeadWindBF=5 BeamWindBF<3 BeamWindBF=3
BeamWindBF=4 BeamWindBF=5 QuarterWindBF<3
Qu arterWindBF=3 QuarterWindBF=4 QuarterWindBF=5
FollowWindBF<3 FollowWindBF=3 FollowWindBF=4
FollowWindBF=5 SPEEDTRIAL TANKTES T
108 rpm
118 rpm
125 rpm
140 rpm
Figure4.ComparisonofmeasureddataandPowerCurves
Next,the authorsexaminedthe effectofwindon
the propulsive characteristics for each of the wind
directionsshowninFigure2.Thecomparisonresults
ofthePowerCurvesareshowninFigure5to9.Inthe
figures, Power Curves are shown for each of the
Beaufort scale. On each
of the wind directions, the
Power Curves generally agree with those of speed
trialandpowerpredictionqualitatively.
When the test ship proceeds at steady BHP, her
speeddecreases asthewind forceincreases,andthe
degree of speed reduction decreases as the wind
direction changes to afterward. Thedegree
of speed
reduction in the higher BHP region is greater than
524
that in the lower BHP region. When the test ship
proceedsundergentleseacondition(therangeofthe
Beaufort scale less than 3), her speed is slow
comparedwiththatofspeedtrialandthetwoPower
Curvesarenotinagreement.Thesamecharacteristics
of thePower Curves
as the above were observed in
thefollowwind,andtheeffectofwaveisconsidered
tobeoneofthecauses.
As the above propulsive characteristics are
observedintheactualsea,itisimportanttopredicta
Power Curve taking into account the effect of wind
direction,wind
forceandwave.
1500
2000
2500
3000
3500
4000
4500
5000
11 12 13 14 15 16 17 18 19
BHP(kW)
Speed(kts)
BF<3
BF=3
BF=4
BF<3
BF=3
BF=4
SPEEDTRIAL
TANKTEST
Figure5.ComparisonofPowerCurves(Head Wind)
1500
2000
2500
3000
3500
4000
4500
5000
11 12 13 14 15 16 17 18 19
BHP(kW)
Speed(kts
)
BF<3
BF=3
BF=4
BF=5
BF<3
BF=3
BF=4
BF=5
SPEEDTRIAL
TANKTEST
Figure 6. Comparis on of Power Curves (Oblique Head
Wind)
1500
2000
2500
3000
3500
4000
4500
5000
11 12 13 14 15 16 17 18 19
BHP(kW)
Speed(kts)
BF<3
BF=3
BF=4
BF=5
BF<3
BF=3
BF=4
BF=5
SPEED TRIAL
TANKTEST
Figure7.ComparisonofPowerCurves(BeamWind)
1500
2000
2500
3000
3500
4000
4500
5000
11 12 13 14 15 16 17 18 19
BHP(kW)
Speed(kts)
BF<3
BF=3
BF=4
BF=5
BF<3
BF=3
BF=4
BF=5
SPEEDTRIAL
TANKTEST
Figure8.ComparisonofPowerCurves(QuarterWind)
1500
2000
2500
3000
3500
4000
4500
5000
11 12 13 14 15 16 17 18 19
BHP(kW)
Speed(kts)
BF<3
BF=3
BF=4
BF=5
BF<3
BF=3
BF=4
BF=5
SPEED TRIAL
TANKTEST
Figure9.ComparisonofPowerCurves(FollowWind)
525
3.2 SelfPropulsionFactors
Althoughtheeffectofwindandwaveisincludedin
the propulsion performance data shown in the
previous section, obtained Power Curves agree well
with that obtained by the speed trial qualitatively.
The authors examined the characteristics of Self
PropulsionFactorsusingthedatainthe
regionof16
knotswhereenoughdatawereobtained.
Whendiscussingtheeffectofdisplacementonthe
propulsion performance, it is necessary to examine
the effect of both mean draft and trim. The authors
performed the analysis considering the effect of
displacementaloneduetoscarcemodeltestdataon
thetrim.
The Power Curves of the test ship at full load
(6,308tons)and75%load(5,763tons)obtainedbythe
modeltestsareshowninFigure10.Alsoshowninthe
figurearethePowerCurvespredictedunderthesame
conditions using the propulsion performance
prediction program “HOPE
Light,” which was
developed by the National Maritime Research
InstituteJapan(Sasaki2009).Asthepredictedvalues
agreewellwiththeobservedvalues intheregionof
16 knots, the HOPE Light can be used as a tool to
discuss the characteristics of Power Curve in the
wavesunderthe
aboveconditions.
From the Power Curves shown in Figure 10, the
authors considered that the measured horse power
valuecanbetransformedintothehorsepoweratthe
displacementinthemodeltestusingtheequation(8).
0.9
16 5763
()( )
s
BHP BHP
V Displacement
(8)
2000
2500
3000
3500
4000
4500
15 16 17 18 19
75%Load
75%Load(HOPELight)
FullLoad
FullLoad(HOPELig ht)
BHP(kW)
Speed(kts)
Figure10.PowerCurvesobtainedbymodeltestandHOPE
Light
3.2.1 Effectofwindandwave
The authors estimated the DHP under wind and
wavedisturbancesusingHOPELightandtheresults
are shown as the dotted line in Figure 11 together
withmeasuredDHP.MeasuredDHParetransformed
into the value at the standard condition
(displacement;5,763tons,speed;16
knots)usingthe
equation(8).
Asshowninthefigure,measuredDHPdatatend
to increase in proportion to the increase of wind
velocity and wave height, and thisis mainlydue to
the increase of resistance by wind and wave.
However,onthedistributionofmeasureddatainthe
regionofstrongerwind and higher wave, regularity
is difficult to be observed. Figure 12 shows the
comparison results of DHP between measured and
estimated.ItseemsthatestimatedDHPislowerthan
measuredDHPintheregionmentionedabove.
2000
2500
3000
3500
4000
4500
5000
5500
0 5 10 15
DHP
(
kW
)
Wind velocity (m/s)
Wave 4m
Wave 2m
Wave 3m
Wave 1m
◆4m
◇3m
●2m
●1m
Figure 11. Comparison of measured DHP and estimated
DHPforwindvelocity
2500
3000
3500
4000
4500
2500 3000 3500 4000 4500
DHP(kW, measured)
DHP(kW, predicted)
Figure12.ComparisonofmeasuredDHPandestimated
DHPbyHOPELight
In order to investigate the cause of the above
results, the authors calculated propulsive efficiency
)(
D
using measured BHP, thrust and speed
accordingtotheequation(1),andtheobtainedresults
are plotted against the propeller loading
(
)4/(5.0/
2
2
DVTC
aT
) as shown in Figure
13.The propulsive efficiency
)(
D
decreases in
proportion to the propeller loading. In the figure,
estimated propulsive efficiency (
0D
) using the
following simple equation that takes into account
only the change of propeller efficiency (
0
) is also
shownindottedline.
00
()()()
DH R T
const const C
(9)
where
H
=Hull efficiency; and
R
= Relative
rotatingefficiency.
Figure13. Comparison of measured propulsive efficiency
andestimatedpropulsiveefficiency
526
As is obvious from the figure, measured
propulsive efficiency values
)(
D
are generally
lower than estimated values (
0D
). The difference
between measured values and estimated values is
about5%onaverageandthisvalueisdifficulttobe
disregarded. The reduction of the propulsive
efficiency under wind and wave disturbance seems
nottobeonlyduetotheincreaseofpropellerloading.
3.2.2 Wakecoefficientinthe
actualsea
The authors examined the relationship between
wake coefficient
)1(
t
w and propeller loading
(
T
C ) in order to analyze the cause of propulsive
efficiencyreductionshowninFigure14.
Figure 14 shows the relationship between
propeller loading and wake coefficientin the region
of16knots.Althoughsomedispersionisobservedin
the data, it seems that wake coefficient increases in
proportiontothepropeller
loading.Asthepropeller
loading will be changed by wind, wave and
displacement, the difference of the displacement is
consideredtobeoneofthecausesofwakecoefficient
increasing.
0.7
0.725
0.75
0.775
0.8
1 1 .1 1 .2 1 .3 1 .4 1 .5
(1-w
t
)
C
T
Figure14. )1(
t
w andpropellerloading(
T
C )
Theauthorsinvestigatedtherelationshipbetween
displacementandwakecoefficientasshowninFigure
15.Inthefigure,thewakecoefficientsobtainedbythe
modeltestarealsodisplayedasthedotted line.The
wake coefficients obtained by the full‐scale
experiments increase in proportion to the
displacement. On the other
hand, the wake
coefficientsobtainedbythemodeltestsinthisregion
remainalmostconstant.
Therefore,thereisapossibilitytoexplainthecause
of difference between measured
)(
D
and estimated
(
0D
) shown in Figure 13 by the effect of
displacementon the wake coefficient. Further model
testsseemtobenecessaryinordertoclarifytheeffect
ofdisplacementonthewakecoefficients.
0.71
0.72
0.73
0.74
0.75
0.76
0.77
0.78
0.79
5200 5400 5600 5800 6000 6200
(1-w
t
)
Displacement (ton)
Figure15. )1(
t
w anddisplacement
4 CONCLUSION
The authors performed an experimental study in
order to clarify the characteristics of propulsion
performanceintheactualsea.Resultsobtainedinthis
studyaresummarizedasfollows.
1 The power curve ofa shipin the actualsea with
windand wavedisturbancesdecreasescompared
with the power
curves obtained by power
prediction and a speed trial. The degree of
propulsivepowerreductiondependsonthewind
direction,windforceandwaveheight.
2 As the propulsion performance decreases
comparedwiththespeedtrialresultsevenifaship
proceedsunderthefollowwind,theeffectofwave
isconsideredtobegreaterthanthatofwind.
3 Causesof BHP increasing in thewave conditions
can be divided into the resistance increasing and
thechangeofpropulsiveefficiency.
4 Causes of propulsive efficiency reduction in the
actual sea can be divided into the unavoidable
reduction of propeller efficiency
due to the
resistanceincreasebywindandwaveandothers.
5 In order to determine the characteristics of Self
PropulsionFactorsintheactualsea,itisimportant
tomeasurethethrustandcalculatethepropulsive
efficiencydirectly.
6 Thewakecoefficient
)1(
t
w intheactualseacan
be estimated directly by measuring her thrust.
However,astheaccuracyofthrustmeasurementis
generally inferior to that of the torque
measurement, it is necessary to examine their
mutualrelationbeforehand.
7 The effect of propeller loading on the wake
coefficientintheactualsea
issmallandtheship’s
conditionsuchasdisplacementandtrimseemsto
havealargerimpactonthewakecoefficient. The
authors consider that the above results are
necessarytobeexaminedbymodeltests.
8 Inordertoimproveenergysavingshipoperation
technologies, it is important to
predict the Power
Curve more precisely taking into account the
characteristics of the effect of wind and wave
obtainedinthepresentstudy.
REFERENCES
Yokoi,T.etal.2010.Amechanismonparallelprocessingto
numerical weather prediction for weather routing –
Accuracyevaluationandperformancebenchmarkofthe
sea surface wind prediction‐, Proceeding of Asia
NavigationConference2010;33‐40
Sasaki, N. et al. 2009. Development of ship performance
index (10 mode at
sea), Report of National Maritime
ResearchInstitute,Vol.9No.4;1‐46
Tsujimoto, M. et al. 2000. Evaluation of ship propulsive
performance by analyzing full‐scale data and time
history of ship motion on actual seas, Proceeding of 4
th
OsakaColloquium on Seakeeping Performance of Ship 2000;
1‐9
SECshipperformancemonitor.ShoyoRngineeringCo.,Ltd.
HP, http://www.shoyo‐e.co.jp/english/product_a_01_e.
html
Sasaki,N.etal.2010.Developmentofbasicdesigntoolsfor
high performance ships, Report of National Maritime
ResearchInstitute,Vol.10No.3;1‐21
