437
1 EFFECTIVENESSOFSARACTION
Ineverysearchandrescue(SAR)operationatseathe
coordinator seek out to maximize the result while
minimizingtheriskofconductingtheaction(Burciu,
2012).Theresultoftheactionisinfluencedbymany
factors, including the quantity and quality of
available situation
data (type of search object, last
known position, etc.), on hydro‐meteorological
conditions, on available search and rescue units
(SRUs), their capabilities and distance to the scene.
Theeffectivenesscanbeassessedaftercompletionthe
taskbymeasuringtheratiooftheresultobtained to
the assumed one. The first
volume of the IAMSAR
Manual (International Aeronautical and Maritime
Search and Rescue ManualʺOrganization and
Managementʺ) recommends a use of performance
indicatorsbymeasuringthenumberofpeoplesaved
inrelationtopeopleindanger.Thisapproachallows
toanalysedeterminationofpreventionandresponse
effortsinSARsystem.Effectiveness
ofanactionorof
a given period can be calculated from the basic
formula(1):
LS
EFF L
LS LLA
(1)
where, EFF(L)=Programme Effectiveness for
PreventingLossofLife,LS =Lives Saved, LLA=Lives
LostAfterNotification.
However, since the result is composed of many
factors,itcanbemodelledbyselectingandmodifying
thesefactors, during theplanning stage.The way in
whichthesearchorrescueactionisprepared
andthe
wayitisconductedhasahugeimpactontheresult.
Therefore, the potential (theoretical) effectiveness of
theactioncanbedeterminedbeforetheactionstarts.
This potential effectiveness will be the expected
capacity to achieve the goal. Potential effectiveness
can be expressed, among others, by using the
indicator
of probability of success (POS) and by
selecting the most useful rescue units for a given
task.(IMO/ICAO,2016)
Assessment of the Potential Effectiveness of the WIG
Craft in Search Action at Sea Using SARMAP Software
M.Małyszko
M
aritimeUniversityofSzczecin,Szczecin,Poland
ABSTRACT:Selectionandchoiceofsearchandrescueunits isanimportantstageinplanningSARoperations
atsea.Thedifferentcharacteristicsofunitsenablethemtobeusedforspecifictasksatseatovaryingranges.
ThemainadvantageofWIGcraftsistheir
speed,buttheyalsohaveotherfeaturesnecessaryformarinerescue.
Theassessmentoftheunitʹspotentialeffectivenessisbased,amongotherthings,ondefininganddetermining
theparametersinfluencingitsabilitytoachievesuccess.ComputersoftwaresuchasSARMAPcanbeusedfor
action planning. The article contains calculations
for simulated actions of searching for several objects,
additionallyagraphicalsolutionofthesituationhasbeenshown.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 13
Number 2
June 2019
DOI:10.12716/1001.13.02.23
438
POSistheprobabilityoflocatingthesearchobject.
Itis the true measure of effectiveness. Theability to
findanobjectdependsonhavingasensorcapableof
detecting the object and placing this sensor close
enoughtothesearchobjectthatitsdetectionwouldbe
probable.Thisvalue
iscalculatedfromthefollowing
relation:
POS POC x POD (2)
where, POC=Probability of Containment,
POD=ProbabilityofDetection.
Probability of containment (POC) – expresses the
probability that the searched object is within the
search area. The probability area is defined as the
smallestareathatcontainsallpossibleobjectlocations
(inthiscasePOC=100%).Inmanysituations,the
area
issolargethatitcannotbesearchedbyavailableSRU.
Ithappensoftenthatsomesub‐areasaremorelikely
tocontainthesearchobjectthanothers.Therefore,the
areacanbedividedintosmallersectors,forwhichthe
SRUs would be deployed. The POC measures the
probability of
achance to detect an object if it is
actually located within the area. The exact way to
determine the POC is described in the IAMSAR
Manual vol. II“Mission Co‐ordination” and will not
bediscussedindetailinthispaper.
Probability of Detection (POD) is the probability
thatthesearchedobjectwillbefound,assumingthat
in fact it is in the search area. POD dependent on
coverage coefficient, detector sensor, search
conditions and accuracy with which rescue units
navigatethedesignatedsearchpattern.Thecoverage
factor(C)comparestheamountofsearchingthatcan
bedoneto
thesizeofnecessarysearch.Thecoverage
factorcanbeexpressedastheratioofsearcheffort(Z)
to the size of the search area (A), or as the ratio of
sweepwidth(W)andthetrackspacing(S).Coverage
factoristobecalculatedasfollows:C=Z/AorC=W/S.
Searcheffort(Z) of a SRU representthearea that
can be effectively searched by particular unit. The
search effort is computed as the product of search
speed (V), search endurance (T), and sweep width
(W)Z=V*T*W.Sweepwidthdefineshowwideisthe
scanningstripforaSRU.It’sa
measureoftheability
to detect a search object. It takes into account
environmentalconditions(visibility,seastate),sizeof
theobjectandtheoperatorfatigue.Itisprovidedasa
guideforsettingthetrackspacing.Thetrackspacing
is the distance between the centres of adjacent SRU
paths.
Atrackspacingthatislessorequaltoone‐half
the corrected sweep width results in POD of 100%.
Thegreaterthesearcheffortcomparedto the search
area, the higher the coverage ratio (analogically
sweepwidthtotrackspacing).
Selection of SRU for a given task bases on
effectiveness
indicators. The SRU perform different
potential effectiveness depending on the speed,
observation altitude, seaworthiness, range and
detectionequipment.
2 WIGCRAFT
In recent years, the development of technology for
vehicle moving in the ground effect has been
resumed.TheWIGCraftmeansWing‐In‐Ground‐Effect
Craft.Thegroundeffect
isaphysicalphenomenonof
creationanaircushionbetweentheflyingvehicleand
the surface over which it moves. As a result air
pressuredifferencebetweentheareaunderthewings
andoverthewingsappears,andanadditionallifting
forcedirectedupwardsiscreated.Consequently,the
craftachieves
relativelyhighspeedwithverylowfuel
consumption.Birds use this physical principle when
overcomeverylargedistancesabovethesurfaceofa
lakewithoutmovingthewings.Thecraftoperatingin
ground effect flies at a low altitude, which is equal
about 1/3 of the span of its wings, usually
a few
meters.ThemainconstraintsfortheoperationofWIG
craft are hydro‐meteorological conditions. A large
wavescan hinder ordestabilizea flightof a craftin
the ground effect, since it needs a flat surface. The
WIGCraftsareclassifiedunderashipinaccordance
to the
IMO (International Maritime Organization).
WIG units are cheaper to build and maintain than
traditionalairborneunits,andcrewtrainingcostsare
alsolower(comparabletomarinestafftrainingcosts).
WIG units do not also require any special aviation
infrastructureastheyarestationingattheports.The
WISECraft(Wing‐In
‐Surface‐EffectCraft)orEkranoplan
arethe common namesfor WIG Crafts. (Górtowska,
2012)
Examples of modern WIG crafts used for search
and rescue tasks at sea are: Rescue Unit Aron‐7
(Figure1)orMulti‐TaskMarineEkranoplanES‐108.
Aron‐7isaIMOtypeB*WIGcraft
forsearchtasks
(*type B means that in addition to a flight in the
groundeffect,theunithastheoptionofleavingitand
elevates). The craft has been registered by
GermanischerLloyd.
The craft aviates within an altitude of 5 m in
groundeffect,atamaximumspeed
of200km/h(108
knots). The cruising speed is 120‐150 km/h (65‐81
knots). The vehicle has a length of 10 meters,
wingspan12 meters, height 2.9meters. The range of
the craft is 800 km (432Nm). The number of
passengersis5(two‐personcrew).
Figure1. WIG Craft Aron‐7 during take‐off phase. Source:
www.cnsamt.com
A motor is placed at the top of the cabin. The
wingsareintheshapeofanelongatedtrapezoid,and
theballastsaremountedtothemainpartofthehull.
439
TheT‐shapedtailismovedawayfromthecabin.The
unit elevates up to 40 m within 30 seconds with a
jump‐upfeature,andelevatesupto100mwithin60
seconds. The maximum flight altitude is limited to
150 m (IMO standard). The limit does not apply
to
militaryunits.Fuelconsumptioninthegroundeffect
is only 25 l/h. The craft is equipped with thermal
vision system, field of view indication via LCD
monitor and operates at various environments with
itswavelengthIRspectrum.Stabilizationoftheflight
allows to take pictures of good quality. Aron‐
7
possessunderwaterpropulsionsystemallowingcraft
to manoeuvre through port and channels. The
maximumdraftis0.5m.Thevehiclehastheoptionof
wing folding. The assembly or disassembly time is
about10minutes.Afterfoldingthewings,theunitis
ready for transport on the carriage to the
launching
site.(C&SAMT,2007)
3 SARMAPSIMULATOR
The SARMAP computer application (Version 6.5
licenseavailableforMaritimeUniversityofSzczecin)
enables modeling of search and rescue operations
based on Geographic Information System (GIS). The
SARMAPmodelsystemconsistofseveralintegrated
components.Themodelitselfpredictsthemovement
of
various floating objects (e.g. person in the water,
raft, boat) on the sea surface in compressed time
mode.Forthesepurpose,themodelreliesmainlyon
inputtedenvironmentaldata(wind,current)andthe
driftcharacteristicsofthefloatingobjectinquestion.
TheapplicationincludestheabilitytodeploySRU,set
their
search patterns, observe the animation of
conductingthesearchincompressedtimemodeand
calculate the various indicators i.e. probability of
containment, probability of detection, probability of
success. (Applied Science Associates, 2008). Two
methodsareavailablefordeterminingthedriftofthe
object:
IAMSARMethod;
MonteCarloMethod.
TheIAMSARsolutiondetermines the mostlikely
positionofthesearchobjectbasedonthetrajectoryof
threeparticlestravelingleft,rightandstraitdown.An
errorestimate,increasingwithtime,iscomputedfor
each particle, and a box that circumscribes the error
radiusaroundeachparticledefinesthemostprobable
locations of the missing object after a given time
period. This method is limited to single Last Know
Position(LKP)initialization,butdoesallowmultiple
searchobjects in scenario. The Monte Carlo solution
determinesaprobabilitygridbasedonthetrajectory
of a large number of representative particles, each
moving
withsomerandomness.Thegriddividesthe
area in sectors based on the probability of
containment. The Monte Carlo solution allows for
multiple search objects in single simulation,
initializationbasedonsinglepointLKPortrackline,
probabilitycells,POCbasedonprobability.
4 SCENARIO
Theanalysisconcernsonthesituation
ofsearchingfor
the 8 crew members of a fishing vessel with whom
thereisnocontactandthereisasuspicionofathreat
to life. The last known position (54°4030N;
0014°4917E)oftheunitispastdueby12hours.An
assumptionwasmade
thatthepositionerrorcouldbe
0,5Nm.Thetypical and random weatherconditions
forsouthBalticwereappliedfortheentirelengthof
the simulation. The values of wind and current
directionandspeed wereenteredinaone‐hourstep
(average wind NNW 5m/s, average current ESE 0,5
kn).
Due to the lack of important information
concerning the distress situation the coordinator
planning the action has high uncertainty and is not
sure what object he is looking for, i.e. whether the
crewisonthedisabledvessel,whetherthecrewhas
left the vessel and stays on the life
raft, or whether
individualpeopleareinthewater.Itisdifficultcase,
because characteristic of those objects vary a lot.
Different size and different sensitivity to the leeway
makesthatsearchmorecomplicated.Thesearcharea
is much larger, and adjusting the parameters for
specificsearchpatternisnotobvious.
The SARMAP program allows to simulate the
movementofseveraldifferentobjects.Fourfollowing
searchobjectsassumedforthescenario:
1 thecrewisonthevessel(reddots);
2 thecrewisonaliferaft(yellowdots);
3 individualcrewmembersareinthewater:
survivors wear immersion suits and are in
apositionontheback(greendots);
deceasedpersons,facedown(bluedots).
The results of the IAMSAR calculation and the
MonteCarlomethoddiffersignificantlyinthesizeof
the calculated search area, as
these methods use
differentcomputationschemes.Theareacontainsall
possiblelocations(donebyMonteCarlo)isfivetimes
larger then the area calculated with the IAMSAR
(Figure2).Forthefurtheranalysis,alargerareawas
assumed.
Figure2. Prediction of searchobjectsdriftwithin12 hours
(dots) and parallel sweep search covering the search area.
(Source:ownstudy,printscreen).
440
The location of deceased people concentrate
around the LKP, their total drift is small, and the
computationmethodspreadstheirpossiblepositions
randomly. Survivors dressed in immersion suit are
more susceptible to wind influence and move over
timeintheaverageSSEdriftdirection.Liferaft’sand
vessel’s possible locations
are furthestfrom theLKP
in the average SE drift direction.The larger the
objectssurfaceareaabovethewaterlevel,thegreater
theinfluenceofwind.Thedivergenceangleforvessel
is more than 45 degrees to the right and left of the
averagedriftline.
The simulator enable
to apply one of four main
searchpatterns(parallelsweepsearch, sector search,
creeping line, expanding square). In the study first
twopatternshasbeenused.Theparallelsweepsearch
isnormallyutilizedwhentheuncertaintyinlocations
is large. That search covers a rectangular area. The
commencesearchpointis
inoneof the corners.The
search legs are parallel to the sides of the rectangle.
The sector search pattern is intended to search
circular area centred on a datum. It consist of tree
triangular sectors, each composed of three legs
intersecting the datum position. It is most effective
when
thedatumismoreaccurateandthesearcharea
isrelativesmall.
4.1 Possiblesolutions
Hypothetically the nearest SRU (WIG Aron‐7) is
located in port of Kołobrzeg away of the scene of
distance 37,21 Nm. With the transit speed of 80 kn
(the maximal speed is not applied for
fuel saving
reasons), the SRU will arrive at the scene in 28
minutes. For the basic computations, the search is
rectangle area EFGH that contains all possible
locations.Thesearchareahasbeencoveredwiththe
parallelsweepsearchpattern(Figure1),adaptedfor
the WIG craft Aron‐7 characteristic,
and taking into
accountfollowingfactors:drifterror0,3,safetyfactor
1,1andfixerror0,25.
Figure3.AnimationofconductingthesearchbyaSRU(red
line);15‐minutesstep(ontheleft),45‐minutesstep(onthe
right).(Source:ownstudy,printscreen)
The SRU coverage area is equal 483 Nm
2
. The
suggest sweep width for that search (search object:
vessel) is W=7,7 Nm, and the track spacing was set
one‐half of that value (S=3,65 Nm), cause that will
effectinincreasingtheprobabilityofdetectionupto
100%.TheSRUwillsweepthesurfaceatdistance of
1,83 Nm
to the starboard and portside. Because all
possiblelocationareinthesearcharea,theprobability
ofcontainmentis100%.
TheSRUbeginsthesearchinthecornerE,heads
north, and follows the planned route (Figure 3). It
takes1hour45minutestoproceedthewholeroute.
Total
enduranceoftheAron‐7is4hours.Subtracting
thetimeforthetransittothesceneandreturndothe
port, the on‐scene endurance is 2 hours 45 minutes.
On‐sceneenduranceis the productiveavailabletime
forSRU.Thisvalueisusuallytakentobe85%(t
SRU=2
h 20 min), cause 15% of time will be dedicated for
navigatingturnsandinvestigatingsightings.
InthissolutiontheSRUperformwell.Searcheffort
Z=680,36 Nm
2
is larger then the search area,
coverage factor C=2 indicates that the detection is
very likely and the final probability of success is
100%.TheWIGisabletocoverallareawithavailable
time,to detect missing fishingvessel,and with high
likelihoodtodetectsmallerobjectstoo.
The
results would be better with the appropriate
adaptationoftheshapeofthearea,toexcludeareas
wheretheprobabilityofcontainmentisloworequal
zero. A probability distribution map serves this
purpose(Figure4).Thesimulationmodeldetermines
a likelihood grid based on trajectory of alarge
numbers of
representative particles, each moving
withsomerandomness.Thegriddividestheareainto
shades of grey cells based of the probability of
containment.Theprobabilityvaluesaredisplayedin
cornerofthecell.
Figure4. Example of probability grid map. (Source: own
study,printscreen)
The probability maps are helpful for optimizing
thesearchbyfragmentationthearea.Conductingthe
search with available SRUs and modelling search
patternsisnextstepforimprovingtheeffectivenessof
thesearch.
The situation will change if the coordinator
decides to search for persons‐in‐water (PIW) only.
The search
area has to modelled that contains the
possible location of survivors in water only. In the
studyassumedtwoprobablescenarios:thecrewhas
abandonedvessel,theyweartheimmersionsuitsand
their positon is on the back (acc. to survival
techniques recommendation); or the crew did not
survive the
disaster, the bodies float on the water
surface.Thedriftvaluesinthistwo cases varyfrom
eachother.The deceased personis more submerged
441
sotheinfluenceofthewindisless(smallervalueof
theleeway).
Forthisanalysisnewsearchsub‐area(OPRS)has
beensetandtwodifferentsearchpatternscompared
(Figure 5). The search sub‐area is placed over the
predictedlocationsofthePIW(blueandgreendotes).
Figure5. Parallel sweep search (on the left) and sector
search(ontheright)overthesubarea.PIWsearch.(Source:
ownstudy,printscreen)
Fortheparallelsweepsearchthesuggestedsweep
widthis0,5Nmandsametrackspacingwasadopted.
Thelinesmustbequiteclosetoeachother,sothatthe
SRUfollowingtheroutecanbeabletodetectassmall
objects as people in the water on its starboard and
portside.Tightroutelayoutreducessearcheffort.The
SRUsearcheffortZ=93,2Nm
2
,whilethesearcharea
is161,8Nm
2
.TheSRUisnotabletocoverthewhole
areaduetoitson‐sceneendurancelimit.Sooneofthe
solutionistoreducetheareatakingintoaccountthe
probability distribution map and excluding areas
which are less likely. In this case the results are as
follows:POD
=47,04%,POC=55%,and eventually
POS=25,85%
Anothersolutionistouseasectorsearchpattern.
The possible locations surround the datum, so this
methodwillbeeffective.Onesearchtakes65minutes
(on‐sceneendurance2h20min),Theresultsforthat
searchare
asfollows:POD=47,04%,POC=68%,and
eventuallyPOS=31,96%.Ifthe search isunsuccessful
the unit has still ability to perform the search again
and therefore increase the chance for success. The
patternistoberotatedbyanangleof30degreesand
theunitcarriesoutthe
searchforasecondtimewithin
the limits but not along the same route. This will
escalate the probability of the location of the object.
The cumulative probability of success (POS
C) is the
sum of probability of success for the first search
(POS
1) and the probability of success for the second
search (POS
2). For that computation the updating
valueapplies:
1
new old
POC POD xPOC
(3)
where,POC
new=probabilityofcontainmentforsecond
serach; POC
old = probability of containment for the
firstsearch.
ForthatsolutionthePOC
new=36,01%andthePOS2
is 16,94%. Therefore the cumulative probability of
successPOS
Cisequal48,09%.
In this case the use of sector search seems to be
abetter solution, as the probability of finding an
objectis almosttwice ashigh as withthe useof the
methodofparallellines.
5 CONCLUSION
SARMAP modelling simulator is a flexible tool for
assessing the
potential effectiveness of search and
rescueunits.Objects’driftprediction,easydataentry,
searchpatterns application, calculation of indicators,
time compression, objects’ drift animation and
executionofsearchpatternsbydeployedSRUsarethe
mostimportantadvantagesoftheprogramSARMAP.
The coordinator planning the action has a huge
impact
on the final result, through the selection of
factorsandcomponents.Itiseasiertomakemistakes
while being under the pressure of time and
consequences of decisions made. Analytical and
graphical presentation of the situation can help by
importantdecisionsinSARactionplanning.
Onemayhavesomedoubtsabout
theMonteCarlo
method results of object’s drift prediction. Some
particlestravelintheoppositedirectiontothedrift.It
seems more practical that dispersion may be useful
whendeterminingadatumwithanerrororarandom
movementofanobjectalongthedrift.Inthiscase,it
has
resultedinasignificantincreaseinthesizeofthe
searcharea,wherethePOCinsomesub‐areasisvery
loworequaltozero.
TheWIG craft may actively carry out search and
rescuetasks.Intheexaminedscenario,itwasableto
findamissingshiporwith
aprobabilitycloseto50%
tofindpeopleinthewater.Usingitshighspeeditcan
beagreatsupportforcoordinatedactions.Ifanother
vesselismorelikelytofindpeopleinthewater,the
WIGcouldatthattime sweepingalargerarea fora
raftor
ashipdetectingpurpose.
ThebiggestadvantageofWIGisthequickarrival
time, which significantly increases the chances of
survival of people for whom the greatest threat is
hypothermia.Flyinginthegroundeffectgiveshigh
speedandreducesfuelconsumption,whichmakesit
economical and less polluting. If it’s
necessary, the
WIG craft (type B) can elevates up to improve
detectioncapability(PODwillincrease,thereforethe
POSwillbegreater).TheWIGmaynavigateoverthe
water surface and provide direct assistance to the
injured persons, if weather conditions make it
possible for the WIG. When completing a
fleet of
SRUs for particular SAR systems, the option of
choosing WIG vehicles may be considered. Their
searchandrescuecapabilitiescoverdifferentaspects
andmayincreasetheeffectivenessnotonlyofspecific
actionbutalsooftheSARsystemasawhole.Inorder
totakefulladvantageofWIG
craftsonalargescalein
searchandrescueandmarinetransport,furtherwork
andresearchiscertainlyneededontheirdesignand
thereductionofhydrometeorologicalconstraints.
442
REFERENCES
Burciu Z. Niezawodność akcji ratowniczej, Oficyna
WydawniczaPolitechnikiWarszawskiej,Warszawa2012
Applied Science Associates, SARMAP User’s Manual,
Version6.5,2008
C&S AMT, Aron‐7 Catalogue, 2007, on‐line access,
www.cnsamt.com.
Górtowska M. Available Search Effort of the WIG Craft,
Scientific Journals of the Maritime University of
Szczecin,Szczecin2012
IMO/ICAO,
IAMSAR Manual, Volume I, Organization and
management (2016 Edition) – IMO‐Vega, on‐line access,
www.vp.imo.org.
IMO/ICAO, IAMSAR Manual, Volume II, Mission Co‐
ordination (2016 Edition) – IMO‐Vega, on‐line access,
www.vp.imo.org.
IMO/ICAO,IAMSARManual,VolumeIII,MobileFacilities
(2016 Edition) – IMO‐Vega, on‐line access,
www.vp.imo.org.
IMO, SAR
Convention, International Convention on
MaritimeSearch and Rescue1979,‐IMO‐Vega, on‐line
access,www.vp.imo.org.
IMO,SOLASConvention,InternationalConventionforthe
Safety of Life at Sea 1974 – IMO‐Vega, on‐line access,
www.vp.imo.org.
