233
1 INTRODUCTION
Thesuccessfullaunchandclearawayofalifeboatin
highseastates isaffectedby both thecapabilitiesof
the lifeboat and the actions taken by the coxswains.
The effects of coxswain actions on the ability to
completeasuccessfullaunchandsailawayhavenot
beenfullyinvestigated,norhavethelimitationsofthe
launching equipment in high sea states been fully
explored.Thispaperinvestigatesboth.
Previousscalemodelexperimentswereperformed
to evaluate the factors affecting a successful lifeboat
launchandsailaway(SimõesRéetal.,2002,Simões
Ré & Veitch, 2004, Simões
Ré et al., 2008). These
experimentsinvestigatedthelimitationsinlaunching
considering factors related to wave height, launch
configuration, and lowering speeds from the davit.
Theexperimentalstudiesusedregularwaveswhichis
asimplificationofrealconditionswherewaveshapes
are irregular. These studies also did not include the
full range of sea states that are possible in offshore
operations as wave heights were limited to 10 m.
Additional studies used numerical simulations to
studysimilarfactorsandexploredtheeffectoftiming
of hook release and application of propulsion
(Gabrielsonetal.,2011).
Industrystudieshaveidentifiedthatcoxswain
skill
has an impact on a successful lifeboat launch,
although benchmarking of skills is difficult due to
limitationsintraining(Robson,2007).Evaluatingthe
impact of human performance and skills on a
successful launch in high sea states is not practical.
Due to the perilous nature of launching lifeboats in
roughconditions,theroleoftheoperator(coxswain)
isnotsomethingthatcanbeethicallyinvestigatedin
field trials or experiments, nor practiced in realistic
(rough)waveconditions.Duetothenatureofmodel
experiments, specifically the scaling of time, it is
Use of Simulations to Predict Lifeboat Survivability in
Extreme Waves and the Effectiveness of Coxswain
Performed Actions
R.Billard
1
,R.Rees
1
,B.Veitch
2
&A.S.Re
3
1
VirtualMarine,Paradise,Canada
2
MemorialUniversityofNewfoundland,St.John’s,Canada
3
XataanConsultingLtd,StJohnʹs,Canada
ABSTRACT:Simulationswereusedtoinvestigatetheperformanceoflifeboatsinhighseastatesusingavirtual
wavetank.Numericalsimulationswereperformedinregularandirregularwavestostudylaunchperformance
inextremeweatherconditions.Limitationsinlaunchequipmentandtherole
ofthetimingofcoxswains’actions
wereinvestigated.Thestudyindicatedthatthelifeboatmaynotbeabletosuccessfullylaunchwhensignificant
waveheightsareabove8mandthelifeboatislaunchednearthetroughofawave.Highinitialsetbackand
continuous wave forces result in the vessel
being unable to clear away from the launch platform. As wave
heights increase, the amount of setback and time to exit the launch area increases. Over 35% of launches
resultedinthelifeboatbeingunabletoclearfromthelaunchareawhensignificantwaveheightswere10mor
above.
The study also identified that delay in completion of actions performed by the coxswain, such as
releasingthelifeboathooksandapplyingthrottle,canincreasesetbackandtimetoexitthelauncharea.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 18
Number 1
March 2024
DOI:10.12716/1001.18.01.25
234
difficulttousemodeltestsasa meanstoinvestigate
timedependenthumanfactors.
In thisresearchsimulationswere used to explore
the lifeboat performance in wave heights not
previously tested in scale model experiments and
field trials.The simulator is also used to study how
the timing of
actions performed by the coxswain,
including applying the throttle and releasing the
hooks,affectlaunchperformance.
Detailsarefirstpresentedonthelaunchprocedure
andtheperformancemeasuresdiscussedinthepaper.
2 BACKGROUND
2.1 Launchprocedure
Assummarizedinpreviousresearch(SimõesRéetal.,
2002, Gabrielson et al., 2011),
there are multiple
phasesofalifeboatlaunch.Theyareasfollows:
Lowering phase: lowering the lifeboat from the
davitsystemtothewatersurface.
Water entry: starts when the vessel enters the
water and the lifeboat becomes buoyant. During
thisphase,waterfillsthevesselhydrostaticrelease
unit,andhydrostatic pressuremovesacablelink
toallowthehookreleasehandletoopen.
Release:starts when thehookisreleased and the
vesselisfreefromthefallwires.
Sail away: vessel propulsion (throttle) is applied
andtheoperatormanoeuvresthevesseltoasafe
area
awayfromthelaunchplatform.
Thelaunchstartswhenabrakewireispulledand
thevesselbeginsloweringfromthedavit.Thevessel
continuestoloweruntilthevesselisinthewater and
atimecountbeginsonfillingthehydrostaticrelease.
The hydrostatic release activates when the
vessel
remainsbuoyantfor3secondsorlonger.Ifthewave
falls away from the vessel before it is buoyant, the
hydrostaticreleasedrainsandthetimerestarts.Once
the vessel is buoyant for three seconds, the
hydrostaticreleasesystemallowsthehooksystemto
operate. A hydrostatic indicator on the
hook release
systemmovesprovidingavisualcuetothecoxswain.
The standard procedure is to release the hooks and
then apply full throttle as quickly as possible and
driveawayfromthelaunchplatform.
This paper focuses primarily on the water entry,
release,andsailawayandconsidersthe
relationship
betweentheactionsperformed bythecoxswainand
thetimingoftransitionbetweenthephases.
2.2 Setback
The experimental studies (Simões Ré et al., 2002,
Simões Ré & Veitch, 2004, Simões Ré et al., 2008),
identified that the amount of setback, or backwards
trajectoryofthevessel,increasesinhigher
seastates
when the launch position on the wave is near the
troughonthelowerpartofthewaveupslope.Wave
andwindforcesimpactthevesselonwaterentryand
can push the vessel backwards towards the launch
platform if the waves are against the evacuation
direction(a
headsea).Inheadseaswithwaveheights
of 6 m and above, setback can result in the vessel
being pushed back to within critical safety zones of
launch platforms. The amount of setback and
likelihood of occurrence of setback increases with
waveheight.
Total setback can be a result of
a single wave or
multiple wave encounters may cause progressive
setback before the vessel begins to move forward
(Simões Ré et al., 2002). Figure 1 shows a sample
trajectoryplotofalaunchtoillustratethecasewhere
a vessel experiences initial setback (SB), progressive
setback, and is then able
to progress forward. The
verticalpositionisplottedonthe Yaxis,withvessel
loweringtothestillwaterlineaty= 0.Thehorizontal
displacementpositiveisthedistanceinthe direction
awayfromthelaunchplatform.X=0 isthestarting
horizontalpositionofthevesselwhen
lowered.Inthis
sample the vessel is setback (‐ve x direction) on the
firstwaveencounter,andthesecondwaveencounter
results in higher, or progressive wave setback. The
vessel is then able to progress forward. The
subsequent waves create some backwards
displacement with each wave encounter, but the
overallmovementisinthe+vexdirectionawayfrom
the launch platform. Initial setback, progressive
setback, and forward progress are used to describe
resultsinthispaper.
Figure1.ProgressiveSetback
2.3 Impactoflaunchpositiononwave
Studies identified the impact of timing of release of
thelifeboatatdifferentpointsonthewaveinahead
sea(SimõesRéetal.,2002,SimõesRé&Veitch,2004,
SimõesRéetal.,2008).Launchingonthetroughofthe
wave can
result insignificantsetbackandlaunching
on the crest of a wave results in minimal, or no,
setback. The experiments also showed the effect of
“wave shadowing,” whereby the lowering speed of
thevesselandthewavespeedresultedinlauncheson
theleewardsideofthewave.Withreference
toFigure
2, most launches occurred between‐60 and 60
degrees,onthewaveupslope.Ineffect,itisdifficult
to launch on downslopes, which are favourable to
goodlaunches,andlaunchesaremorelikelytooccur
onupslopeswhich resultinlargesetbacks.Aswave
height or wave steepness increases,
the zone of
possiblelaunchpositionstightens.
235
Taken together, these findings indicated that the
timing of the launch relative to the wave is very
important. It should not be left to chance as it is
something that is within the operator’s ability to
control,atleasttosomeextent.
Figure2.LaunchPositionsforLifeboatWaterEntry
2.4 Performancemeasures
Theprimarymeasureofperformanceinthestudyis
setback. Additional measures are defined for this
study based on target operational outcomes. These
newmeasuresareasfollows,withreferencemadeto
Figure3.
The first additional measure identifies launches
whichmayresultincontactwiththelaunch
platform
andconsequentlymayresultindamagetothevessel
orharmtothecrew.Examinationoflaunchplatforms
andlaunchconfigurationsindicatesthatdavitsystems
areplacedtoprovide20to40mofclearancefromthe
baseoftheplatform.Setbackgreaterthan20mmay
resultin
impactwiththelaunchplatformorresultin
vesselbeingwithinazoneofhighriskofimpact.In
Figure 3, X = 0 is the position directly below the
launch area and X =‐20 m is the distance travelled
towardsthe platform,oppositethetarget evacuation
direction.Toevaluate
performanceforagivensetof
launches, the percentage of outcomes with greater
than20minsetback(%Setbacks>20m)wascalculated.
Figure3. Performance Measures: Setback>20m and
ClearanceTime>60s
Another measure was introduced to evaluate
whether the lifeboat is able to evacuate from the
launch platform quickly. Clearing time is defined as
the time required for the vessel to leave the
splashdown area in the evacuation direction and
reachatargetdistance,whichisdefinedas20mfrom
the
launch position (X = 20 m in Figure 8). Timely
clearance of the lifeboat from the launch area is
desiredtoescapeharmfromhazardsnearthelaunch
platform and to permit the launch of other vessels
from adjacent davits. The percentage of occurrences
withgreaterthan60sof
timeneededtoreach20mis
calculated for a set of launches (%Clearance
Times>60s).Theamountofsetbackandclearingtime
toexitthesplashdownareaarerelated,asthevessel
musttravelalongerdistancetoexitthesplashdown
areaifitissetbackfarther.
If no measure was
recorded for this performance
criteria,thevesselwasunabletoreachtheescapeline
at X = 20 m. This outcome is defined as a failed
clearance. A measure of failed clearance identifies a
performancelimitasthevesselisnotabletoprogress
forwardinthewaveenvironmenttested.
Thislimitis
furtherdiscussedintheresults.
3 SCOPE
With the advent of simulator technology, it is now
possibletoexplorelifeboatandoperatorperformance
in weather conditions typical of their location of
operation. Lifeboat simulators are designed with
accurate numerical behaviour of vessel motions and
wave environments. Trainees interact
with realistic
lifeboatequipmentandperformactionsastheywould
in a real vessel. Studies performed with a lifeboat
simulator have evaluated how skills transfer from
simulatortrainingtorealvessels(Mageeetal.,2016)
andhowskillsareacquiredininitialtraining(Billard
et al. 2019). Recent studies have
focused on how
trainingaffectscoxswainskillacquisitionandlaunch
performance in moderate weather environments
(Billardetal.,2018,Billardetal.,2020).Thesestudies
havenotinvestigatedtheimpactofhumanfactorsin
highseastates.
Simulators are increasingly used to train lifeboat
coxswains.Traineescanpracticeandimprovetiming
ofactions,such as releasing the hooks andapplying
the vessel throttle. There is increased knowledge of
thetimestakentocompletetasksinalifeboatlaunch
via data collected through simulator training
programs.Thetimelyordelayedperformanceofthese
actions is expected to impact the amount of lifeboat
setback and the time to clear from the launch area.
Evaluatinghowthetimingoftheseactionsaffectsthe
launch outcomes will help to define training
objectives.
As in other studies, simulators can explore
scenarios where data is scarce or difficult to obtain
(Groth et al. 2014) and can specifically extend
knowledge of coxswain and lifeboat performance to
high sea states. The study of human factors using
simulationtoevaluateperformanceisevidentinother
operations including flight (McClernon et al. 2011),
medical (Stefandis et al. 2007) and marine (Sellberg,
2017) training. This research shows an example of
how simulations can
be used to evaluate how
operator actions impact the ability to successfully
launchalifeboat.
Thepurposeoftheresearchwastousenumerical
simulations to 1) assess the performance and
limitationsoflifeboatlaunchsystemsinextremeseas
236
and to 2) study the impact of the timing of human
actionsonthelaunchandsail‐awayofthelifeboat.
A numerical simulator, a Virtual Wave Tank
(VWT), was designed to emulate the lifeboat and
wave conditions performed in previous research
(Simões Ré et al., 2002, Simões Ré
& Veitch, 2004,
SimõesRé et al.,2008).Validationwasperformedto
ensure the measured setback is comparable between
the numerical simulator and experimental studies
performed ina wave tank. Comparisons were made
formultiplewaveheightsandlaunchpositionsonthe
waves.ThekinematicsofthevesselintheVWT
were
also compared with results from the experimental
studies.
After validation, we performed three
investigationswiththeVirtualWavetanktostudythe
effect of wave height and timing of coxswain‐
performed tasks on launch performance. The first
investigation built on the outcomes of the
experimental tests (Simões Ré et
al., 2002) and
extendedtheregularwaveconditionstowavesupto
16 m. Simulations were then performed in irregular
seastateswithsignificantwaveheightsof6to12mto
investigate launch performance in irregular waves
with 100‐year return period extreme wind speeds
basedonhistoricaldata of
weatherconditionsinthe
NorthAtlantic(C‐Core,2015).Thethirdinvestigation
studiedhowthetimingofhumanactionsaffectedthe
likelihood of a successful launch. The time taken to
apply propulsion (throttle) is varied. The time to
releasethelifeboathooksoncethevesselisbuoyantin
thewater
andableto bereleasedisalsovaried. The
impact of delayed response in applying the throttle
and hook release is studied. Comparison are also
madebetweencaseswherethrottleisappliedpriorto
release of the hook to investigate how applying an
initial propulsion force influenced the ability to
complete
asuccessfulevacuation.
Performance is evaluated using the measures
identifiedintheprevioussection. The investigations
focusedonperformanceinheadseas.
Thefollowingresearchquestionsareinvestigated:
What is the expected setback of a lifeboat in
extremeregularwavesandirregularwaves?
How is the time to
clear the lifeboat from the
launchstructureaffectedbyseastate?
How does delay in lifeboat throttle and hook
releaseaffectlaunchandevacuationofalifeboat?
3.1 VirtualWaveTank(VWT)
TheVWTsimulationenvironmentusedinthestudyis
a 3D physics‐based engine that was specifically
created
tomodelthemotionofsmallcraftsinmarine
environments. Physics models were derived from
studies of vessel motions in waves, including scale
modelandfull‐scaletestingofthevessels(SimõesRé
etal.,2002,SimõesRé&Veitch,2004,SimõesRéetal.,
2008, Magee et, al. 2016).
Numerical models for all
phases discussed in the launch procedure were
includedinthesimulationenvironment.
Numerical models were implemented to provide
physic‐based responses and timings during the
launch phases. The vessel behaviours at water entry
weremodelledtoincludethetensioninthelowering
wires before the hook is
released, the dynamic
behaviour of the lifeboat as it interacted with the
water surface, and the release of the vessel. The
propulsion and hydrodynamic behaviours of the
vesselonceinthewaterwerealsomodelled.Previous
studies have validated the manoeuvring and
performancecharacteristicsofthevesselmodelledin
this
study are representative of the behavior of the
reallifeboat(Billardetal.2020).Modelsareresolved
on the computer GPU to allow for high‐speed and
high‐resolutionwavemeshestocalculatehydrostatic
andhydrodynamicforces.
Thevesselwasmodelledwithdimensions,weight,
propulsionandsteeringtostudythelifeboat’s
ability
to manoeuvre in the environmental conditions
considered.Thevesselmodelledinthesimulatorwas
a fully loaded lifeboat with length, weight, and
displacement parameters closely matched to the
vesselusedinexperimentsperformedbySimõesRéet
al.(2002).
A twin fall davit was modelled with fall wires
attached
totheforeandafthooksofthevesselduring
lowering.Theloweringspeedofthevesselwaskept
constantat1.0m/s.Thelaunchheightofthedavitwas
35 m. The lowering of the lifeboat is normally
controlledbypullingabrakereleasefromwithinthe
lifeboattoextend
thefallwires.Inthesimulations,the
fall wires continued to extend until the hook is
released.Thisisanormalproceduretomakesurethe
vesselbeginstofloat,andtoreducethelikelihoodthe
vessel is only temporarily buoyant if the wave falls
awayfromthelifeboat.
A
virtualagentwasusedtoperformtheactionsof
the coxswain in the simulator. The virtual coxswain
couldbeprogrammedtoreleasethehook,manipulate
thethrottleandattempttosteerthevesseltodesired
headings. Timings could be set to perform actions
instantaneouslyorwithdelays,orindifferentorders
(i.e. applying throttle before hook release). The
resultantbehaviourofthe vesselwas determinedby
thephysicsenginewhichappliesandresolvesforces
dependingontheactionstaken bythecoxswain.As
anexample,adelayinmovingthethrottleaheadafter
the hook was released resulted in a delay
in the
propulsion and the vessel was free to drift until
propulsion was applied. When manoeuvring, the
virtual coxswain attempted to maintain a constant
headingandusedcorrectivesteeringtocomebackto
a heading if the vessel veered off course. The study
assumedsteeringwasmaintainedtotargetaheading
directly into the waves and away from the launch
platform.
4 STUDYMETHODOLOGY
The study included two stages: 1) validation of the
simulator measures with data from experimental
studies, and 2) using the simulator environment to
performnewstudies.Threestudies,orinvestigations,
were performed with the VWT to study
the lifeboat
performanceinhigherseastates andtoconsiderthe
timingofcoxswainactions.Theinvestigationsvaried
wave shape, wind speed, and coxswain timings to
237
study the effect of these variables. Comparisons are
madebetweeneachstudytoillustrateresults.
4.1 Validation–Simulatorandscalemodelexperiment
Comparisonsaremadebetweentheoutcomesofscale
model testing performed in previous research to
validatethesimulation.Datasetswerecreatedusing
the VWT using
a stokes regular wave, with wave
heightsfrom2to10m,andwindspeedsmatchingthe
scalemodelexperimentaltests(SimõesRéetal.2002).
Validation is performed using the following
comparisonsbetweenthesimulatoroutcomesandthe
scalemodeltests:
1. maximumsetbackforeachwaveheight;
2. setback
formultiplelaunchpositionsonthewave;
and,
3. checking the trajectory of the vessel during water
entryandsailaway.
4.2 Investigation1–Studyofindividualwavesetbackin
highseastates,regularwaves
The first set of test cases investigated the impact of
environmental conditions on lifeboat
setback with
testingextendedtohigherseastatesandwindspeeds
representing storm and hurricane conditions. Test
cases were performed with a stokes regular wave
shapewithwaveheightsrangingfrom2mto16 m.
The approximate wave steepness in each case was
1/20. The simulation used wind speeds and
wave
heights similar to the parameters used in the
experimentalstudies(SimõesRéetal.2002)forwave
heights up to 10 m. The wave heights and wind
speedswereextendedtowaveheightsof12,14,and
16m,usingaveragewindspeedsforobservedwave
conditions (C‐Core,
2015). The parameters for each
wavetestedisprovidedinTable1.
Table1.Series1‐RegularWaveParameters
________________________________________________
WaveHeight WavePeriod MeanWindspeed
(H
w)[m](T)[s][m/s]
________________________________________________
2510
4712
5816
6917
7918
81019
101122
121228
141330
161433
________________________________________________
48launcheswereperformedforeachwaveheight.
Foreachlaunch, thestartingtime of the launchwas
variedresultinginadifferentlaunchpositiononthe
wave,withlaunchescoveringafullwavecycleofone
waveperiod.Themaximumtimepermittedwas240s
(4minutes).
4.3 Investigation2
–Studyoflifeboatperformancein
irregular100yrseas
Thesecondsetofsimulationsinvestigatedthelaunch
andsailawayphaseofthelifeboatinirregularshaped
head seas and high wave heights. The lifeboat was
loweredandlaunchedintoaseastatewithadefined
significant wave height (Hs)
and irregular wave
pattern.Theirregularwaveshapeincludeddominant
waves and lower frequency minor waves. Waves
were generated from a fast Fourier transform to
generate the desired Hs, as measured by the mean
wave height of the highest 1/3 of the waves.
Individual wave heights could exceed Hs. The
maximum
waveheightsinthetestcasesarepresented
intable2.Thepeakperiod(Tp)isthedominantwave
withthehighestenergy.Waveheightsof6mto12m
wereselectedtostudyvesselperformancewherehigh
setback is likely. Wind speeds were taken to be
representative of
100‐yr occurrences in the North
Atlantic(C‐Core,2015)andarehigherthanthewinds
usedintheregularwaves.
Foreachwaveheight,simulationswereperformed
withthreedifferentwavepatterns.Eachwaveshape
hadthecharacteristicparametersidentifiedinTable2.
48launcheswereperformedforeachwave
patternto
cover a full cycle of a dominant wave. The data for
each wave pattern was combined for analysis,
resulting in 144 launches for each combination of
waveheightandwindstudied.
Table2.Series2‐IrregularWaveParameters
________________________________________________
Significant MaxWave PeakWave Mean
WaveHeight Height Period WindSpeed
(Hs)[m][m] (Tp)[s] [m/s]
________________________________________________
68.7 920
811.5 1025
1013.2 1130
1215.8 1233
________________________________________________
4.4 Investigation3–Studyofhumanperformanceon
evacuationperformanceinirregular100yrseas
The third set of simulations varied the time to
complete actions performed by the coxswain in the
lifeboatlaunchandclearaway.Thevirtualcoxswain
intheVWTsimulationwasprogrammedtoperform
thehook
releaseandtomovethethrottlefromneutral
tofullpropulsionatcontrolledtimes.Data collected
from training courses performed by Virtual Marine
hasindicatedthatthetimingofreleaseofhookscan
varyfrom1to5secondsfollowinganindicationthat
the hydrostatic bladder has filled, and the hook
release system can be operated. This delay can be
caused by a combination of human reaction time,
difficulty in operating the hook release handle, or
time taken to perform other tasks. Training records
havealsoidentifiedthetimetoapplyfullthrottlecan
vary between coxswains. Delay in application of
throttlefollowinghookreleasemeansthepropulsion
ofthevesselisdelayed,andthevesselisfreetodriftif
thehookshavebeenreleased.
The study first investigated the application of
throttleanddelayinhookreleaseseparately.Throttle
delay cases assumed the hook was immediately
releasedwhenthehydrostatic
bladderhadfilled,and
times presented are relative to the time of hook
release.Thetimetothrottle(TT)istheamountoftime
thevesselis untetheredby thefall wallsandfreeto
drift before throttle is applied. For the hook release
cases, time to hook release (TR)
was relative to the
instantthehydrostatic bladderhasfilled(t =0),and
thevesselremainedtethereduntilreleaseofthehook.
238
In these cases, throttle was applied immediately on
hookrelease.ThetimingsaresummarizedinTable3.
Table3.DelayedThrottleandHookReleaseCases
________________________________________________
Label Hydrostatic Timeto Timeto
Ready Throttle(TT) HookRelease(TR)
________________________________________________
TT2 t=0st=2st=0s
TT4 t=0st=4st=0s
TR2 t=0st=2st=2s
TR4 t=0st=4st=4
s
________________________________________________
These initial cases studied the delayed
performance of actions normally taken in a launch
sequence where the typical launch procedure is 1)
wait until the vessel is buoyant, 2) release the hook
and3)applythrottle.
An additional series of tests was performed to
investigatetheimpactofearlyapplicationof
throttle,
prior to release of the hooks. This emulates an
operator decision to apply propulsion before the
lifeboatisreleasedfromthefallwires.Thisprocedure
has been suggested by experienced operators as a
meanstogivethevesselinitialthrusttocombatwave
forces,albeitnotastandardoperating
procedure.
In these cases, the virtual coxswain applied the
throttle fully when the vessel was buoyant (i.e.
hydrostaticinterlockhadfilled,t=0),andremained
tethered.Four use cases with differentcombinations
oftimetothrottle(TT)andtimetohookrelease(TR)
are identified in Table 4. Early throttle
provided a
propulsion force before the vessel becomes
untethered, and the hook was released at a time
followingthethrottle.
Simulations were performed for the irregular
wavesidentifiedin Table 2, with Hs from6to12m.
Datasetswereagainacquiredforthreewavepatterns
andcombinedforanalysis,
resultingin144 launches
foreachcaseandwavestudied.
Table4.EarlyThrottleCases
________________________________________________
Label Hydrostatic Timeto Timeto
Ready Throttle(TT) Release(TR)
________________________________________________
TT1‐TR2 t=0st=1st=2s
TT1‐TR3 t=0st=1st=3s
TT2‐TR3 t=0st=1st=3s
TT2‐TR4 t=0st=2
st=4s
________________________________________________
5 RESULTS
In this section summarize the outcomes of the
investigations are summarized and discussed.
Comparisons are made between outcomes of the
studiestoillustratetheeffectofthevariablesstudied.
Multiple measures are discussed to provide insights
on how the outcomes are related and to make
comparisonsbetweenthe
individualinvestigations.
5.1 Results–validation,simulatorandscalemodel
experiment
Comparisons were made between the simulator
measuresandtheexperimentalstudiesperformedby
SimõesRé et al. (2002)tovalidate themeasures and
behaviours observed in the simulator are similar to
theexperimentalstudies. A sample of the validation
cases
arediscussed.
Figure5.Setbackvs.WavePhaseAngle,Hw=6m
Figure6.Setbackvs.WavePhaseAngle,Hw=10m
Figure7.Setbackvs.WaveHeight
Figures 4 and 5 show the measured setback for
various launch positions on a regular wave, with 90
degrees being the wave crest and‐90 degrees being
thewavetrough.Thecomparisonsshowtheobserved
behaviour is the same in the simulator (Simulator)
compared to the scale model experiment
(Experiment),with
setbackincreasingasthevesselis
launchedclosertothetroughofthewave.Ofnote,the
setback in the experiment was limited at
approximately 11 m due to the experimental setup,
withthemodelimpactingthelaunchstructureatthis
point.ThedashedlineonFigures4to6
indicatesthis
limit for the experimental trials. The setback in the
simulatortrials was not limited. Some differences in
the setback measures are observed on the upslope
239
nearthetroughofthewave(30to60degrees)when
thewaveheightis6mandthereisaclosematchwith
mostphaseangl eswhenthewaveheightis10m.As
indicated in Figure 5, the measured setback for the
simulator continued to increase above
11 m as the
vessel was launched closer to the trough (0 to 30
degrees)asthelaunceswerenotlimitedbycollisions.
Figure 6 shows the setback vs. wave height (H
w)
for specific waves for both the simulator and
experimental measures. The solid line indicates the
values where setback is double the maximum wave
height.Theexperimentaloutcomesshowedmaximum
setbackis approximatelydouble thewave height up
toa wave height of 6 m(Simões Ré et al., 2002). At
higher sea states this could not be confirmed in the
experimental results due to the impacts of the
evacuation craft with the structure. The increase in
setback from the simulator tests followed a similar
trendline,withsomeoccurrencesofsetbackabovethe
prediction for the 6 m wave height.
The trend of
increasing setback and variability in setback with
increased wave height is consistent between the
simulatorandexperimentalmeasures.
Figure8.SimulatorXYTrajectory‐Launchnearwavecrest:
H
w=7m,T=9m
Figure9. Simulator XY Trajectory – Launch near wave
troughH
w=7m,T=9
Trajectory comparisons were made between the
experimentalcasesandthesimulatortoensurevessel
kinematics were similar. A key focus was the
observed behaviour of the vessel when it was
launched on different positions between the crest or
troughofawave.Asampleofthevalidationcaseis
discussed.Figures
8 and9aresamplerunsfromthe
simulator showing the trajectory of the vessel on
launch and sail away. Figure 8 shows the vessel
setbackwaslowerwhenthevesselwaslaunchednear
the crest of the wave, and the vessel was able to
continueforwardsteadilywitheach
waveencounter.
With large vessel setback, as in Figure 9, the vessel
had to first overcome the backwards motion. The
vesselprogressedmoreslowly,withsomeprogressive
setbackontheinitialwaveencounters,andthenwas
able to continue forward with additional wave
encounters. Similar behaviour was observed in the
experimental
studies.
Insummary,thecomparisonsindicatethe virtual
wavetankprovidesmeasuresthatarerepresentative
of the vessel and wave interactions seen in the
experimental studies. The amount of setback with
waveheightandthechangeinsetbackwithposition
of launch on the wave (crest or trough) were
consistent
withtheexperimentalstudies.Themotion
of the vessel on water entry and sail away in the
simulator was also representative. Differences
betweenthemeasurescanbeattributedtodifferences
in scaling, variability in physical observations
compared to numerical simulations, and differences
inlimitationsbetweentheexperimentaltestsetupand
the
simulator.
5.2 Results:Investigation1–studyofindividualwave
setbackinhighseastates,regularwaves
A summary of the setback measures for each set of
launches is provided for each of the regular waves
studied.Themeasuredsetbackforeachlaunchisalso
related to the launch position
on the wave (phase
angle).Ineffect,thisdatasetprovidesanextensionto
the outcomes presented in the scale model
experiments, with the outcomes extended to higher
waveheights.
Table 5 provides a summary of the setback
measures for launches performed for each wave
height tested, from 2 m to
14 m. Summary data
includestheaveragesetback(Avg.SB),themedianof
themeasuredsetback(Med.),andstandardDeviation
(SD) for each set of 48 launches performed for each
wave height. The 90th percentile (90th PER.) of
measured setback is provided to indicate the higher
measuresinthedata
setforeachseastate.
The outcomes indicate increasing setback with
increase in wave height, with the average setback
increasingfrom3.23mina4mwaveheighttoover
30mina16mwave.Themeasuresindicatesetback
ashighas65.9mina16m
wave.Thereisalsohigher
variability in the setback as wave heights increase,
which is consistent with previous studies. For all
waveheights,themedianwaslowerthantheaverage
setback,indicatingtherewereahighernumberoflow
setback measures for each set of launches. Figure 9
shows a
graphical summary of this informationina
boxplot.
Table5.SetbackSummary‐RegularWaves
________________________________________________
Hw(m) 4 6 8 10 12 14 16
________________________________________________
Avg.SB(m) 3.23 5.82 8.47 10.1 14.1 20.0 30.4
Med. 2.85 3.75 6.05 6.31 9.11 14.0 28.2
SD90
th
2.99 4.68 6.83 8.98 13.1 17.0 22.4
PER. 7.23 13.0 17.8 24.3 36.3 45.6 62.4
MaxSB(m) 9.1 14.9 25.1 28.0 42.7 54.5 65.9
________________________________________________
240
Figure9.VesselSetback,RegularWaves
Figure 10 shows the setback values and phase
anglesforeachsetofsimulatedlaunchesinthehigher
waveheights,from10mto16m.Theresultsindicate
thatthesplashdownoccursmostfrequentlybetween
0 and 90 degrees, with few occurrences of launches
outsideofthisrangewhenthe
waveheightisgreater
than8m.Analysisofthesetbackandwaveanglefor
higher wave heights shows the maximum setback
increasedsignificantlywhenthevesselwaslaunched
closertothetroughofthewave(0to30degrees).Low
setbackvaluesarepossiblewhentheboatisreleased
closertothecrestofthewave(60to90degrees).The
results are similar to the outcomes of previous
research(SimõesRéetal.,2002).
Figure10.Setbackvs.PhaseAngle,RegularWaves
Figure 11 shows the percentage of launches with
greaterthan20msetback.Asindicatedinthisfigure,
in wave heights of 10 m or greater, the number of
occurrences increases with wave height, with over
50% of the launches in a 14 m wave meeting this
criterion.Figure12shows
thepercentageoflaunches
that required greater than 60 s clearance time
(%Cleartimes>60s). For wave heights of 10 m and
above there were observed cases where the vessel
could not exit the launch location in less than 60 s,
withoccurrencesincreasingaswaveheightincreases.
Thevesselwasable
toevacuateandreach20mfrom
thelaunchposition inlessthan60 sforalllaunches
performedin8mwaveheightorless.
Figure11. SetbackOccurrencesGreaterThan20m,Regular
Waves
Theresultsshowthatinwaveheightsof12mor
abovethenumberofcaseswherethevesselwasnot
able to exit the evacuation zone increased, as
indicatedbytheFailedClearancesseriesinFigure12.
For 50% of the launches performed in a 16 m wave
height the
vessel was unable to exit the evacuation
zone.Thisoutcomeindicatesalimitofthelifeboatin
this high sea state. The outcomes again showed an
increaseofoccurrenceswithincreasingwaveheight.
Figure12. Clearance times Greater Than 60 s and Failed
Clearances,RegularWaves
Investigationofthetrajectorieshighlightsthatthe
maximumsetbackinhighseastatescanbearesultof
continued progressive setback. Figures 13 to 15
present samples of the XY trajectory for a vessel
launch in the three highest wave heights. For each
plot,thelaunchpositiononthewaveis
thesamefor
eachwaveheightandisnearthetroughofthewave.
Inthe12mwave,thevesselwassetbackinitiallyand
wasabletoprogressforwardafter2waveencounters.
Ina14mwave,theinitialandprogressivesetbackset
thevesselbackfurtherand
thevesselwasstillableto
startmovingforwardaftertwowaveencounters.For
the16mwave,thewaveandwindforcescontinued
to push the vessel backwards, and the lifeboat was
unable to move forward. This outcome indicates a
limit has been reached, and there is not enough
propulsion
force to overcome the wave and wind
forces.Asnoted,therewerecasesinthedatasetsfor
both12and14mwaveswherethevesselwasnotable
to exit the evacuation zone, indicating that the
241
combination of initial setback and continuous wave
andwind forces resulted in a limit being reachedin
these sea states. These cases relate to the launches
withhighsetbackshowninFigure9,whichoccurred
whenthevesselwaslaunchednearthetroughofthe
wave(0to30
degrees).
Figure13.VesselTrajectoryHw=12m,RegularWaves
Figure14.VesselTrajectoryHw=14m,RegularWaves
Figure15.VesselTrajectoryHw=16m,RegularWaves
5.3 Results:Investigation2–studyoflifeboat
performanceinirregular100yrseas
For irregular seas, setback is again analysed for the
setsoflaunchesforeachwaveheight.Thepercentage
of occurrences with clearance time greater than 60 s
andfailedclearancesisalsodiscussed.
A summary of the setback
measures (Avg. SB,
Med., SD,90thPER.) is provided for each set of the
144launchesperformedforeachwaveheight.Table7
summarizesthesetbackmeasuresofthelifeboatinthe
irregularseastested, with Hs from6mto12 m. The
average measured setback for each set of
launches
increaseswithincreasingseasate.The90thpercentile
isagainprovidedtoindicatethehighermeasuresin
thedataset.
The 90th percentile indicates there were
occurrenceswithsetbackabove20mforan8mwave
height,withsetbackvaluesabove37mand50min10
m
and 12 m waves, respectively. The standard
deviation of the data increased with wave height
indicatinghighervariabilityinthemeasuredsetback
aswaveheightincreases.Themedia nofthemeasured
setbackforeachseastateremainedlowandbelowthe
mean, with a skew towards lower values. This
outcome
indicates that there were still a higher
numberoflowsetbackvaluesforeachsetoflaunches,
similartothetestsperformedinregularseas.
Table6.SetbackSummary‐IrregularSeas
________________________________________________
Hs(m) 6 8 10 12
________________________________________________
Avg.SB(m) 6.36 8.94 12.99 17.16
Med. 4.60 5.80 5.84 7.07
SD5.31 8.40 15.35 21.49
90
th
PER 6.23 21.85 37.01 51.20
________________________________________________
Figure 16 provides a breakdown showing the
percentage of occurrences for measured setback.
Rangesof setback values are grouped to summarize
thedata.Thefigureindicatestheover50%oflaunches
resultedinlessthan10msetbackforeachofthewave
heights tested. Impact with the launch structure is
unlikely in these cases. The percentage of launches
withsetbacklessthan10mdecreasedfrom78%ina6
mwave height to59% ina12mwave height.Over
74%ofalltestcasesresultedinlessthan20msetback.
Above20m,contactwiththe
launchplatformismore
likely,asdiscussedintheperformancemeasures.The
percentageoflauncheswithgreaterthan20msetback
was 16% in a 10 m wave height and 25% in a 12 m
waveheight.In10mwaves,setbackgreaterthan40m
occurredin8%oftest
cases,increasingto16%ina12
m sea. This result indicates high setback values are
possibleintheseextremeseas.
Figure16.SetbackOccurrences,IrregularWaves
Figure 17 shows the breakdown of the times to
reachthetarget20mdistanceforaclearance.Inmost
cases,thevesselwasabletoreachthetargetdistance
inlessthan60s.Theresultsalsoindicatethatin8,10
and12mwaveheightstherewereseveral
caseswhere
thevesselwasunabletoreachthetargetdistanceof20
m required for clearance, as indicated by the Fail
seriesinFigure17.Inan8msignificantwaveheight,
in 13% of the simulations resulted in a failed
clearance. 35% of cases performed in a 10
m sea
resultedinafailedclearance,increasingto41%ina12
mwave.
242
Figure17.TimetoClearance,IrregularWaves
Figure 18 shows the sample trajectory of the
lifeboat in a 10 m Hs where initial launch position
close to a trough results in high initial setback. The
lifeboat was initially setback over 20 m, and
experienced progressive setback for 2 wave
encountersresultinginafurthersetbackof≈8m.
The
vessel was then able to progress forward. An
additional 5 wave encounters occurred before the
vessel could return to the launch position. For this
caseittookapproximately56secondsforthevesselto
progress from the maximum setback point to the
original launch position. Figure 19 shows the
vessel
trajectoryina12mHsandalaunchnearthetrough
ofthewave.Thevesselexperiencedinitialsetbackof
approximately25mandadditionalwaveencounters
setthevesselbackfurthertocloseto50m.Thevessel
was not able to start forward progress. This result
indicates
alimithasbeenreached.Theseexamplesare
provided to show a case where high initial setback
occurreddue to location of launch on the wave and
was then not able to progress forward and another
where the vessel could not overcome the
environmentalforces.Asnoted,caseswereobserved
inboth10mand12mwaveheightswherethevessel
wasunabletoexittheescapezoneduetoprogressive
setback.
Theseoutcomesshowthereisahigherlikelihood
ofencounteringahazardifseastatesarehigherthan
10 m. The combination of high setback and
progressive setback
can result in possible impact of
thevesselwiththelaunchstructureortheinabilityto
exittoasafearea.Inwaveheightsof8morless,the
setbackwasreducedbutnoteliminated.
Theresultsalsoindicatedthatmostofthelaunches
resultedinlowsetbackeven
inhigherseastates.For
allthewaveheightstestedthemedianofthesetback
measures is less than 7 m and most of the launches
resultedinsetbacklessthan10m.Forthehighestsea
statetested(Hs=12m),thetimetoevacuatewasless
than60
sfor48%ofthelaunches.Thispercentagewas
higher for lower sea states. This result shows that
successful launches can occur in the highest waves
tested if the vessel avoids launching on a wave
positionthatresultsinhighinitialsetback.
Figure18.VesselTrajectory,Hs=10m,IrregularWaves
Figure19.VesselTrajectory,Hs=12m,IrregularWaves
5.4 Results:Investigation3–studyofhuman
performanceonevacuationperformanceinirregular
100yrseas
This section discusses the impact of 1) a delay in
throttle,2)adelayinhookreleaseand3)caseswhere
thethrottleisappliedpriortohookrelease.Foreach
ofthesecases,144
launcheswereperformedforeach
wave height tested. A summary of the setback
measuresforeachsetoflaunchesperformedforeach
wave height is provided. The percentage of
occurrenceswithgreaterthan20msetback,clearance
times greater than 60s, and failed clearances are
discussed.Comparisonsaremadeto
thedatafromthe
second investigation where there was no delay in
throttleortimetohookrelease.
5.4.1 DelayinThrottle
Table 7 presents the summary of the setback
measures for sets of launches performed with a 2
seconddelay(TT2)anda4seconddelay(TT4)intime
to
applying throttle after hook release. Figure 20
showsacomparisonofaveragesetbackmeasuresfor
each sea state, with comparison made to no throttle
delay(TT0).
The results show that there was an increase in
averagesetbackofapproximately17%over allwave
heights when time to throttle is delayed
by 2 s,
compared to the set of launches when there was no
throttledelay.Therewasanaveragesetbackincrease
of35%whenthetimetothrottlewasdelayedby4s.
Similar tothepreviousinvestigations,the medianof
thesetbackmeasureswasbelowtheaveragesetback
foreach
waveheight,indicatingahighnumberoflow
setbackcasesforeachsetoflaunches.Theincreasein
the90thpercentileofmeasuredsetbackforeachofthe
wave heights shows the increased throttle delays
resultedinhighersetbackmeasures.
243
Table7:SetbackSummary–DelayedThrottle,Irregular
Waves
________________________________________________
AverageSetback(m)
________________________________________________
6m 8m 10m 12m
________________________________________________
TT0 6.36 8.94 12.99 17.16
TT2 7.12 10.75 15.37 19.91
TT4 7.95 15.77 15.94 20.75
________________________________________________
Median(m)
________________________________________________
6m 8m 10m 12m
________________________________________________
TT0 4.56 5.73 5.73 7.06
TT2 4.59 5.91 5.62 7.39
TT4 4.72 6.03 6.30 8.88
________________________________________________
90
th
Percentile(m)
________________________________________________
6m 8m 10m 12m
________________________________________________
TT0 14.74 21.86 37.01 51.20
TT2 17.18 25.61 38.64 54.91
TT4 19.31 27.24 39.77 54.17
________________________________________________
Figure21 showsthepercentageofoccurrencesof
setbacksgreaterthan20mincreasedanaverageof4%
for TT2 and 7% for TT4, compared to no throttle
delay.Thepercentagesincreasedtoover20%inan8
mwaveandtoover30%ina10or12mwave
when
throttlewasdelayed4s.AsshowninFigure22,witha
delayinthrottleof2s,themeasuredoccurrenceswith
clearancetimesgreaterthan60sincreasedby11%in
10mwavesandby12%in12mwaves.Relatedtothis
outcome, the increased throttle delays
resulted in
moreoccurrencesofthevesselbeingunabletoleave
theclearancezone,asshownbytheFailedClearances
inFigure23.Ina12mwave,delayedthrottleby2or4
s resulted in the vessel not being able to exit the
evacuationzoneinover40%of
thelaunches.
Figure21. Average setback, Delayed throttle, Irregular
Waves
Relatingthisto operational objectives, the results
suggest the target is to apply throttle as quicky as
possible following hook release. It is unrealistic to
assume a coxswain will be able to release the hook
with no delay though applying the throttle in less
than2sisachievable,asobserved
intrainingcourses.
Training should provide sufficient practice for
traineestolearntooperatethehookreleaseasquickly
as possible. Training scenarios can also incorporate
plausibleoutcomesidentifiedinthisstudy.Ifduring
trainingcoxswainsareobservedtotakealongtimeto
applythrottlethenthereisa
possibilityofacollision
or inability to evacuate the launch area. These
outcomes can be built into simulator scenarios to
providefeedback.
Figure21. Setback Occurrences >20 m, Throttle Delays,
IrregularWaves
Figure22. Clearance Times Greater Than 60 s, Delayed
throttle,IrregularWaves
Figure23. Failed Clearances, Delayed Throttle, Irregular
Waves
5.4.2 DelayinHookRelease
Table 7 shows the summary of the setback
measuresforeachofthesets of launchesperformed
witha 2 second delayin time of hook release(TR2)
and a 4 second delay in releasing the hook (TR4).
Comparisons are again made to an instant time
to
hookreleaseandthrottle(TT0‐TR0).
Table8andFigure24indicateaninitialreduction
inaveragesetbackandoccurrencesofsetbackgreater
than20mwhenthehookreleasedelaywas2s(TR2),
andthenanincreaseinthesevalueswhenthethrottle
delaywas4s
(TR4).Theoccurrenceofclearancetimes
greaterthan60 s also changed, with a reductionthe
percentageofoccurrencesforTR2andanincreasein
TR4. A considerable increase in occurrence of
clearance times greater than 60s and failed
evacuationsoccurredina12mwaveheight,with74%
of the
cases resulting in failed clearances. These
outcomesareshowninFigures25and26.
244
Table8.SetbackSummary,DelayedHookRelease,Irregular
Waves
________________________________________________
AverageSetback(m)
________________________________________________
6m 8m 10m 12m
________________________________________________
TT0 6.36 8.94 12.99 17.16
TR2 5.53 6.39 8.02 8.02
TR4 9.49 11.05 17.57 17.57
________________________________________________
Median(m)
________________________________________________
6m 8m 10m 12m
________________________________________________
TT0 4.56 5.73 5.73 7.06
TR2 4.84 5.54 6.94 7.06
TR4 5.58 7.07 9.79 20.1
________________________________________________
90
th
Percentile(m)
________________________________________________
6m 8m 10m 12m
________________________________________________
TT0 14.7 21.9 37.0 51.2
TR2 8.59 10.3 11.4 27.3
TR4 23.3 25.4 43.5 57.0
________________________________________________
While these outcomes seem counterintuitive, the
behaviour can be explained by considering how the
delayinhookreleaseaffectsthepositionofreleaseon
thewave.Giventhewaveshapeandslope,thevessel
islikelytolandontheupslopeoftheofthedominant
wave (0 to 90 degrees),
as indicated in previous
research(SimõesRéetal.,2002)andInvestigation1of
this paper. The delay in hook release keeps the
lifeboatin position, andthefall wires do notextend
enoughforthevesseltodriftbackwardssignificantly.
Asaresult,forasmalldelaythevessel
couldrelease
onthetoporthedownslopeofthewavewherewave
forces were more favourable to reduce setback. Too
long a delay resulted in both the vessel starting to
drift backwards and release occurring closer to a
troughof thewave where wave forces could induce
moresetback.In
effect,delayinhookreleaseprovided
ashortwindowofbenefitandthe“waveshadowing”
wasreducedforashorttime.
This window is expected to be highly dependent
on wave shape (height, period, and steepness). The
results are specific to the wave shapes used in this
study.Further investigation
isrequired todetermine
ifsimilaroutcomesareseenindifferentwaveshapes,
whichisoutsideofthescopeofthispaper.
Figure24. Average Setback‐Delayed Hook Release,
IrregularSeas
Figure25. Setback Occurrences > 20 m, Hook Release
Delays,IrregularWaves
Figure26:ClearanceTimesGreaterThan60s,HookRelease
Delays,IrregularWaves
Figure27.FailedClearances,HookReleaseDelays,Irregular
Waves
5.4.3 ThrottleBeforeHookRelease
Thefollowingcasessummarizethemeasuresfrom
launches where the throttle is applied prior to the
releaseofthehooks.Asnotedinthemethodology,the
timespresentedarerelativetothetimethevesselwas
abletobereleased(t=0)andthetiming
wasvaried
forthetimetothrottle(TT)andtimetorelease(TR)of
the hooks. In these cases, the lifeboat remained
tetheredandpropulsionforcewasappliedbeforethe
hook was released. Comparisons are made with the
case where throttle and hook release were applied
immediately(TT0‐TR0).
As
indicated in Figures 28 to 30, there was an
improvement in most outcomes when throttle was
applied early and prior to the hook release. The
greatest improvement in all performance measures
occurredwhenthe throttle was applied1safter the
vesselcouldbereleasedandthehookwasreleased1
s
later.ThisseriesisnotedasTT1‐TR2.Asindicatedin
245
Figure 28, Average setback was reduced by
approximately 50% and there was a reduction of
setbackoccurrences greaterthan20 m and clearance
times greater than 60 s. Similar outcomes were
observed for cases TT1‐TR3 and TT2‐TR3. For these
cases, the percentage of setback occurrences greater
than
20 m was reduced to 10% or less in all wave
heightstested,asindicatedinFigure29.Theseresults
indicate that application of throttle before hook
release creates enough initial propulsion to improve
launchperformancebasedonthemeasuresdiscussed.
The results indicate that the timing of throttle
before hook
release must still be performed quickly,
andhookreleasecannotbedelayedtoolong.Thisis
shown in the case where the time to throttle was
performed 2 seconds after the vessel is able to be
launchedandtimetohookreleasewasperformedtwo
seconds following (TT2‐TT4). Figure
28 indicates a
smallincreaseinaveragesetbackinhighseastatesfor
this case. Figure 29 shows there were increased
occurrences of setback greater than 20 m in a 12 m
wave height. Figures 30 and 31 show there was an
increaseofoccurrenceofclearancetimesgreaterthan
60s
andfailedclearancesin10and12mwaveheights.
Thisresultagainsuggeststhathighthrottleandhook
releasedelayscanresultinreducedperformance.
Figure28. Average Setback, Throttle before hook release,
IrregularWaves
Figure29.SetbackOccurrencesGreaterThan20m,Throttle
BeforeHookRelease,IrregularWaves
Figure30. Clearance Times Greater Than 60 s, Throttle
BeforeHookRelease,IrregularWaves
Figure31.FailedClearances,ThrottleBeforeHookRelease,
IrregularWaves
These cases show a procedure that can be
performedtogivethevesselinitialpropulsionpriorto
being released. The results show better launch
performancewhenthrottleisappliedbeforethevessel
is released. The results indicate a need for these
actionstobeperformedinatimelymanner.
6 CONCLUSIONS
The goals of the research were to use simulation to
extendtheknowledgeoflifeboatperformanceinhigh
sea states and to evaluate how human performance
canaffectoutcomes.
Theresultsshowastrongrelationshipbetweenthe
performance measures and wave conditions.
Specifically,bothsetbackandtimetoexitthe
launch
area were both dependent on wave height and the
wave phase angle at the launch point. These results
are the same as found previously in experimental
work up to about 10 m (Simões Ré et al., 2002), but
have extended the wave heights up to 16 m in the
simulation
environment.
The position on the incoming wave at which the
lifeboatwaslaunched(i.e.thewavephaseangle)was
foundtobeparticularlyimportant.Whenlaunchedat
orverynearthecrest,lifeboatsavoidedlargesetback
andwereabletomakewayrelativelyquicklytoclear
thelaunchzone.Conversely,
whenlaunchednearthe
trough or the upslope of the incoming wave, the
lifeboatsweresetbackimmediatelybythewave.The
magnitude of the setback was dependent on wave
heightinadditiontowavephaseangle.Consequently,
theinitialsetbackexperiencedbythelifeboatduring
its first wave encounter made clearing
the launch
moredifficultfortworeasons:first,thelifeboathadto
overcomethemomentumassociatedwiththesetback
246
action;second,itseffectivestartingpointwasbehind
thenominallaunchtarget(directlybelowthedavits)
by a distance equal to the setback (or progressive
setback).
In practical terms, one consequence of setback is
thatthelifeboatcancollidewiththelaunchplatformif
there is insufficient clearance between
the launch
target and the platform. While the environmental
conditions at the timeanevacuation are outside the
control of evacuees, the timing of the launch is not.
Timingalaunchrequiresthatthecoxswaincanseeor
otherwise sense the approaching waves and has
enough familiarity with the lifeboat controls
(e.g.
lowering, releasing the hooks, throttling) to perform
thelaunchoperationwithinthenarrowtimewindow
required for a successful launch on a crest. For a
typicallarge wave,the window for a crest launch is
onlyabout5to7seconds.
The studies of time of throttle delay and
time of
hookrelease timingprovide insightson howhuman
actions can affect launch performances. Interpreting
the outcomes of the third investigation, we see a
general trend that a quicker performance of actions
results in better performance outcomes. This result
hasimplicationsfortraining.Delaysinactionscanbe
due to
inability to recognize launch cues (i.e. the
hydrostatic indicator movement), improper
movementofthehookreleasehandle,orperforming
actions out of order. These timings can be further
delayedifthere arefaults inthesystemthatrequire
additional time to remedy, such as performing a
hydrostatic override procedure. The
results of this
research suggest training goals should target the
quick performance of these actions and training to
providepracticetoimprovethesetimings.
The research also indicates that new operational
procedures can improve launch performance.
Applyingthethrottlepriortohookreleasecanreduce
setback and escape times significantly, as
long as
these actions are performed quickly. This procedure
was suggestedbyoperatorswithmarineexperience.
Operational procedures that result in improved
performance can be embedded into curriculum to
traincoxswains.
Considerationsmustbegiventothespecificityof
the wave environments and launch configuration
when interpreting the research outcomes
in this
paper.Asindicatedinpreviousresearch(SimõesRé&
Veitch, 2004), the wave steepness can have a
considerable effect on the amount of measured
setback, although wave steepness was not varied in
the current work. The simulations focused only on
escape from the platform in a head seawhere
wave
directionisdirectlyagainstthedesiredescapepathof
thelaunch vessel.This scenariowas consideredasa
worstcase.Scenarioswithobliquewaves andwinds
wouldpresentadditionaloperationalchallenges(e.g.
maintainingadesiredheading).
ACKNOWLEDGEMENTS
Theauthorsacknowledgewithgratitudethesupportofthe
NSERC/Husky Energy Industrial Research Chair in Safety
atSea.
REFERENCES
[1]Billard,R.,Smith,J.J.E.(2018)Usingsimulationtoassess
performance in emergency lifeboat launches.
Proceedings,Interservice/IndustryTraining,Simulation,
and Education Conference (I/ITSEC). Paper number
19179.
[2]Billard,R.,Smith,J.,VeitchB.,(2019)Assessinglifeboat
coxswain training Alternatives using a simulator. The
Journal of Navigation Published online by Cambridge
University
Press:19September2019.
[3]Billard, R., Musharraf, M., Smith, J., Veitch B., (2020)
Using Bayesian methods and simulator data to model
lifeboat coxswain performance. WMU Journal of
Maritime Affairs. Published May 2020.
https://doi.org/10.1007/s13437‐020‐00204‐0
[4]C‐Core (2015) Metocean climate study offshore
Newfoundland and Labrador, Nalcor Energy
Report.
http://exploration.nalcorenergy.com/wp‐
content/uploads/2016/09/Nalcor‐Metocean‐Study‐Final‐
Report‐Volume‐2‐27‐May‐2015.pdf
[5]Gabrielsen, O., Helland, B, Seines, P.O., Helland, L. R.,
(2011) Study of davit‐launched lifeboats during
lowering, water entry, release and sail away.
NorweigianOilandGasAssociationwebsitepublication
http://www.olf.no.
[6]GrothK.,Smith,C.,Swiler,L.
(2014)ABayesianmethod
for using simulator data to enhance human error
probabilities assigned by existing HRA methods.
ReliabilityandSystemSafety128(2014),32‐40
[7]International Maritime Organization., & International
Conference on Training and Certification of Seafarers
(2010).STCWincluding2010ManilaAmendments,2017
Edition.
[8]Magee, L.E.,
Smith, J.J.E., Billard, R., & Patterson, A.
(2016) Simulator training for lifeboat maneuvers.
Proceedings of the Inteservice/Industry Training,
Simulation,and EducationConference(I/ITSEC).Paper
number16030.
[9]McClernon,C. K.,McCauley, M.E., O’Connor,P.E.,&
Warm,J.S.(2011).Stresstrainingimprovesperformance
duringastressfulflight.HumanFactors:
TheJournalof
theHumanFactorsandErgonomicsSociety,53(3),207‐
218.
[10]Sellberg, C. (2017). Simulators in bridge operations
training and assessment: a systematic review and
qualitativesynthesis.WMUJournalofMaritimeAffairs,
16(2),247‐263.
[11]Stefanidis,D.,Korndorffer,J.R.,Markley,S.,Sierra,R.,
Heniford, B.T., & Scott,
D.J. (2007). Closing the gap in
operative performance between novices and experts:
does harder mean better for laparoscopic simulator
training? Journalof theAmerican Collegeof Surgeons,
205(2),307‐313.
[12]Robson,J.K.,(2007)OverviewofTEMPSCperformance
Standards.HealthandSafetyExecutiveResearchReport
RR599.Norwich.
[13]Simões Ré,
António & Veitch, B., Pelly, D. (2002)
Systematic investigation of lifeboat evacuation
performance. Transactions, Society of Naval Architects
andMarineEngineers.110:341‐360
[14]Simões Ré, António & Veitch, B. (2004) Evacuation
performance of davit launched lifeboats. Proceedings,
OffshoreMechanicsandOffshoreEngineering,June20‐
25,2004.
[15]SimõesRé,
António&Veitch,B.(2008)Acomparisonof
three types of evacuation system. Transactions‐The
SocietyofNavalArchitects andMarineEngineers, 115,
pp.119‐139,2008
