simulation of w dust transport in the kstar tokamak comparison with fast camera data

In Section 2 , the DUMPRO DUst Movie PROcessing routines developedat CEAto extract dust trajectoriesis presented.In the case of intrinsic dust, the experimental data obtained by image pr

ContentslistsavailableatScienceDirect

journalhomepage:www.elsevier.com/locate/nme

A Autricquea ,∗, S.H Hongb , N Fedorczaka , S.H Sonb , H.Y Leec ,d , I Songc ,d , W Choec ,d ,

C Grisoliaa

a CEA, IRFM, Saint-Paul-Lez-Durance,F-13108, France

b National Fusion Research Institute, 113 Gwahangno, YuSung-Gu, Daejeon 305-333, Republic of Korea

c Department of Physics, Korea Advanced Institute of Science and Technology (KAIST), Daejeon 34141, Republic of Korea

d Impurity and Edge plasma Research Center, KAIST, Daejeon 34141, Republic of Korea

a r t i c l e i n f o

Article history:

Received 13 July 2016

Revised 8 November 2016

Accepted 9 November 2016

Available online xxx

a b s t r a c t

Inthispaper, dusttransportin tokamakplasmas isstudied throughbothexperimentaland modeling aspects Image processing routines allowing dust tracking on CCD camera videos are presented The DUMPRO(DUst MoviePROcessing)code featuresadustdetectionmethodand atrajectory reconstruc-tionalgorithm.Inaddition,adusttransportcodenamedDUMBO(DUstMigrationinaplasmaBOundary)

isbrieflydescribed.IthasbeendevelopedatCEAinordertosimulatedustgrainstransportintokamaks andtoevaluatethecontributionofdusttotheimpurityinventoryoftheplasma.Likeotherdust trans-portcodes,DUMBOintegratestheOrbitalMotionLimited(OML)approachfordust/plasmainteractions modeling OMLgivesdirectexpressionsforplasmaionsand electronscurrents,forcesand heatfluxes

onadustgrain.Theequationofmotionis solved,giving accesstothe dusttrajectory.Anattempt of modelvalidationismadethroughcomparisonofsimulatedandmeasuredtrajectoriesonthe2015KSTAR dustinjectionexperiment,whereWdustgrainsweresuccessfullyinjectedintheplasmausinga gun-typeinjector.Thetrajectoriesoftheinjectedparticles,estimatedusingtheDUMPROroutinesappliedon videosfromthefastCCDcamera inKSTAR,showtwo distinctgeneraldustbehaviors,duetodifferent dustsizes SimulationsweremadewithDUMBOtomatchthe measurements.Plasmaparameterswere estimatedusingdifferentdiagnosticsduringthedustinjectionexperimentplasmadischarge.The exper-imentaltrajectoriesshowlongerlifetimesthanthesimulatedones.Thiscanbeduetothesubstitution

ofaboiling/sublimation pointtotheusualvaporization/sublimation cooling,OMLlimitations(eventual potentialbarriersinthevicinityofadustgrain areneglected)and/ortothelackofavaporshielding modelinDUMBO

© 2016TheAuthors.PublishedbyElsevierLtd ThisisanopenaccessarticleundertheCCBYlicense(http://creativecommons.org/licenses/by/4.0/)

1 Introduction

Dust will be a critical issue for futurefusion devicessuch as

ITER Generated throughvarious processesrelatedto plasma/wall

interactions,dustgrainsareanimportantsourceofimpurities

hav-ing well known consequences in terms of radiative losses and

plasmainstabilitiesgeneration[1] DustcanbeobservedusingCCD

camerasastheyinteractwiththeplasma, throughrecycling

pho-tonemissionandthermalemissivity.Camerasprovidewithvideos

onwhichimageprocessingroutinescanbeappliedinorderto

de-tectdusteventsandmeasuredusttrajectories

∗ Corresponding author

E-mail address: adrien.autricque@cea.fr (A Autricque)

In Section 2 , the DUMPRO (DUst Movie PROcessing) routines developedat CEAto extract dust trajectoriesis presented.In the case of intrinsic dust, the experimental data obtained by image processingis delicate to analyze,since a dust trajectory depends

onthedustmaterial,temperature,sizeandelectriccharge,among others.Dustinjectionwasperformedinseveraltokamaksand al-lowsconstrictionofsome oftheseparameters.Severalcodes ded-icatedtothe modelingofdusttransportinplasmasalreadyexist, namelyMIGRAINe[2] ,DUSTT[3] andDTOKS[4] ,amongothers

InSection 3 ,thenewlydevelopedDUMBO(DUstMigrationina plasmaBOundary) codewillbe briefly presented.Theaimofthis workistopreparefortheinstallationofa dustgun-typeinjector

onthe WESTtokamak,aswell asthe image processingand sim-ulationtoolsdevelopedfordataanalysis.Sincetheinjector design

issimilar to that ofthe KSTARdustinjector [5] , the2015 KSTAR http://dx.doi.org/10.1016/j.nme.2016.11.012

2352-1791/© 2016 The Authors Published by Elsevier Ltd This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ )

Pleasecitethisarticleas:A.Autricqueetal.,SimulationofW dusttransportintheKSTARtokamak,comparisonwithfastcameradata,

example,inSection 4

2 DUMPRO: the image processing code

Intokamakoperation,thecommonlyuseddiagnostictoobtain

measurements on in-vessel dust transport is CCD cameras They

providewithRGB(Red-Green-Blue)videosthat need further

pro-cessingfordusteventstobedetected.Inthissection aredetailed

theDUMPRO(DUstMoviePROcessing)routinesusedtodetectdust

eventsonvideosandreconstructdusttrajectories

The firststepistoisolatethedusteventsappearingonframe

Ablackandwhite(BW)videoiscomputedfromtherawRGBdata

usinganoperationsequencesimilartothatdescribedin[6] :

gray-scale conversion, logical filtering (for pixel-size noise reduction),

backgroundremoval,BWconversion.Thelatterpreprocessingstep

differsfromthe usualpixel intensitythresholdingused toisolate

dust eventsin previous works [7,8] A peak detection method is

appliedtoeachpixeltemporalsignal, noted (t, x),wherexisthe

pixellocationonframeandtisthetime.Ashiftedsignal sh(t, x)is

createdasfollows: sh(t , x)= (t+dt , x+dx)+ds ,where dxisof

theorderofafewpixels,dtafewtimeindicesanddsisafraction

ofthepeakintensity.Peaksarelocatedwhereand/orwhen > sh

This method shows better results on movies with varying

back-groundssince onlysudden events are detected,whereas a brutal

thresholdcould keepsome long lastingelementsthe background

suppressionstepcould not deleteproperly, suchashot spots

ap-paritionorplasmaemissionchanges.DUMPROincludesother

fea-tures,suchasframevibrationcompensation,whichwerenotused

fortheresultspresentedinthispaper

The secondstepconsistsinassociatingthepreviouslydetected

dust events together to reconstruct trajectories The algorithm

worksusinga recurrence methodover time:given adust

trajec-toryreconstructeduntiltheframet0,aprobabilityisassociatedto

everydustdetected onthenextframet0+dt tobe thefollowing

point.Ifthemostprobabledustonframe(t0+dt)hasaprobability

overagiventhreshold,the pointis addedtothe trajectory,

mak-ing this method fully automatic Later on, two successive points

on a dust trajectory will be referred to as parent andchild,

re-spectively.InDUMPRO,theprobabilityformulathatdrivesthe

par-ent/childassociationdependsontwo parameters:(i)the distance

betweenpotentialparentandchild:sinceadustmotionismostly

inertia driven, its velocity vector norm and orientation changes

ratherslowlywithrespecttotheframerateofafastCCDcamera

(∼200HzinthecaseoftheTV2camera inKSTAR).Thus the

dis-tancebetweentwoconsecutivepointsonatrajectoryrecordedby

aCCDcamera mustnot changetoo drastically.(ii) Thedifference

inapparentsizeofthepotential parentandchild:similarlytothe

previous point, dust temperature and size evolutions are rather

slow processes compared to the frame rate of a fast CCD

cam-era.Hereiswherethealgorithmdiffersfrompreviousworks[7,8] ,

whichdidnottakeintoaccountthedustapparentsize.Letus

con-sideraBW videocontainingdusteventsandadusttrajectory

re-constructeduntilframet0.Letibethefinalpointofthetrajectory,

onframet0,andjadustlocatedonframet0+dt.Theprobability

forjtobethechildofiiswrittenasfollows:

P(i , j)=α1×



cosθi , j+1

2



× Gdist



d i , j

+α2× Gsize



s i , j

(1)

whereαi are weights, usually set asα1=5/6 andα2=1/6, d i, j

and i, j are the distance and apparent size difference between i

andj, respectively, Gk are Gaussian functionswithparameters to

be chosen (center andwidth), andθi, j is the anglebetweenthe

vectorslinking the parentofi toi andi toj.Thecenters ofGdist

andGsizearetheaveragedistanceandapparentsizedifference

be-tweentwosuccessivepointsofthedusttrajectoryup toframe t0,

Fig. 1 Map of the probability to find a dust position on frame t 0 + dt, given its

position at times t 0 ( i ) and t 0 − dt (Parent of i )

respectively.ThewidthsofGdist andGsizeareparameters depend-ingonthe cameraresolution,usuallyafew pixels.Fig 1 givesan overviewofthe probabilitycomputationinDUMPRO:anexample

oftrajectoryisplottedoverprobabilityvaluesforeachpixelofthe frame

ResultsofDUMPROroutinesare giveninFig 3 (b)foramovie fromthe2015 KSTARdust injectionexperiment Bluecircles rep-resentingthedetecteddusteventsonthewholevideoareplotted overthesuperimposedframeandbluelinesshowthetrajectories reconstructedbythealgorithm

3 DUMBO: the dust transport code

In parallel to the image processingroutines, a dust transport simulationcodehasbeendevelopedatCEA.NamedDUMBO(DUst MigrationinaplasmaBOundary),itisbasedontheOrbitalMotion Limited (OML) approach [9] Like in other dust transport codes, OMLexpressions are implemented fora spherical dustgrain and Maxwellian distributions for plasma particles energy, taking into accountameanflowvelocityinthecaseofions[10] The plasma background necessary to compute plasma/dust interactions is an input to DUMBO It can be obtained either by plasma modeling codessuchasSOLEDGE-2D[11] orbyexperimentalmeasurements [12] Theaimofthepresentsectionistogive aquickoverviewof themodelimplementedinDUMBO.Moredetailswillbepresented elsewhere

3.1 Dust charging

Adust grainimmersedin aplasmachargesup tothe floating potential φd,whichis determinedbysolving thecurrentbalance Plasmaelectronandioncurrents,notedJ iandJ e,aregivenbyOML [10] DUMBOalsotakesintoaccount secondaryelectron emission (SEE)andthermionicemission(TH)effects

The SEE yield δsee is computed using the Young–Dekker for-mula,sinceitwasshowntogive moreaccurateresultsat scrape-off layer (SOL)relevant energies than the Sternglass one [13] ,and

is integratedoverthe electron incidenceangleand aMaxwellian distribution.Thus δsee dependsmainlyon theincomingelectrons energy (i.e the electron temperature T e), the dust material and

φd.Similar expressions tothat ofMIGRAINe are implemented in DUMBO forφd ≤ 0[2] Inthe case φd ≥ 0,secondary electrons areassumedtobereabsorbedbytheattractinggrain,resultingin

δsee=0

THdesignatesthe electronemission generatedbythe temper-ature increase of a material The thermionic current J th depends

on the dust material, temperature and φd and is given by the

Richardson–Dushmanformula[14]

Thecurrentbalancecanthenbesolvedtofindφd:

The dust electriccharge Q d is calculated usingthe expression

Q d=4π0dφd,where0 isthepermittivityofvacuumand dthe

dustradius[15]

3.2 Equation of motion

Amongstthemanyforcesactingonadustgrainimmersedina

plasma,threearekeptintheDUMBOmodel:theLorentzforce,the

gravity andthe iondrag force F id,whose expression comes from

theOMLtheory[2] Theequationofmotioniswritten:

M ddVd

where M d is the dust mass, V d its velocity, E and B are the

lo-calelectricandmagneticfields,respectively,and g isthe

accelera-tionofgravity.Theiondragforceisusuallythemainforce acting

ondustgrains.Nevertheless,thedustradius d playsanimportant

role intheamplitudeoftheseforces:giventhat F id ∼ r2

d , Q d ∼ r d

and M d ∼ r3

d (neglecting thedependance ofφd on d), it appears

thatgravityplaysanimportantroleforlargegrains

3.3 Dust heating

Theheatingequationiswrittenasfollows:

M d c pdT d

where T d is the dust temperature, p is the T d dependent heat

capacityand Q k are thedifferentheat fluxesimpacting the grain

[15] Q i and Q e are the plasma ions and electrons heat fluxes,

respectively Their expressions come from the OML theory, the

latter being generally the main source of dust heating Q see and

Q th are the secondary and thermionicelectron heat fluxes, given

by theYoung–Dekker andRichardson–Dushman formulas,

respec-tively.Q raddesignstheblackbodyradiationandQ recisthe

recom-binationheatflux.Collectedionsareassumedtorecombineonthe

dustsurfaceandformdihydrogenmoleculesbeforebeingreleased

intotheplasma.Heatfluxesduetootherspeciesarenottakeninto

account,sincetheyhavelowerdensitiesandcarrylessenergy

Va-porizationcoolingisneglected andreplacedbyaT dsaturationon

phasetransitions,whilsttheincomingheatingpowerisdirectedto

thisphasechange

3.4 Mass loss

Whilstinteractingwiththeplasma,adustgrainlosesmassdue

tophysicalsputtering,vaporization/sublimationand,insomecases,

chemicalsputtering.InDUMBO,themasslossequationiswritten:

dM d

dt





sput

dt





vap

(5)

where dM d

dt |sput is the mass loss due to sputtering The

expres-sions from Behrisch and Eckstein, considering different

impact-ing ions with different energies on different target materials,

are implemented in DUMBO [16] The variation of the

sputter-ing yield with the angle of incidence and the energy

distribu-tionfunctionoftheincomingparticlesistakenintoaccount,

sim-ilarly to what is done in MIGRAINe [2] dM d

dt |vap is the vapor-ization/sublimation massloss,expressed withtheHertz–Knudsen

formula[17]

Fig 2 KSTAR dust injection setup: (a) Injection point location in a poloidal section;

(b) Gun-type injector design [5] ; (c) MEB image of the injected W powder

4 Application to the 2015 KSTAR dust injection experiment

Comparingintrinsicandsimulated dusttrajectories isdelicate since CCD camerasdo not give access to importantdust param-eters on which the trajectory depends strongly, such as d , T d, the dust material and distance to the camera A way to con-strain some parameters iscontrolled dust injection Such experi-mentshavebeenperformedonvarioustokamaks(DIII-D,TEXTOR, NSTX,MAST,amongothers)duringthelastdecade.Detailsare pre-sentedin [18] andthe references therein.This section will focus

onthe dustinjectionexperimentperformedin KSTARduringthe

2015 campaign, the application of the DUMPRO routines on the video recorded by a fast visible camera and the comparison be-tweenmeasureddusttrajectoriesandsomesimulatedones gener-atedwithDUMBO

4.1 Dust injection experiment in KSTAR

During the 2015 campaign in KSTAR, dust injection was per-formed using a gun-type injector, whose design is shown in Fig 2 (b).Thechosen powder fallsfromthestoragereservoir into the canon by gravity and is propelled into the plasmaby a pis-ton,whichisputintomotionbyapiezo-electric motor.More de-tailsonthe KSTARgun-typeinjector canbe found in[5] The in-jectionpoint waslocated slightly below the outer mid plane, as shownin Fig 2 (a), andthe injectionvelocity wasa few m/s, di-rectedinwards.Theinjectedamountwas∼2mgofWpowderper shot.Thegrainssizedistributioniswide,rangingfrom∼10μ mup

to∼100μ m.AMEBimage inFig 2 (c)showsthat dustgrainsare

Fig. 3 Application of the DUMPRO routines on the KSTAR #13101 TV2 video: (a) Frame at t = 4 836 s with the DUMPRO region of interest in red; (b) Dust trajectories (blue)

reconstructed on the whole video over the superimposed frame, zoomed in the region of interest, with the two dust behaviors, case 1 and case 2 , underlined in green (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

mostlyaccretedintoclustersof∼100μ msizeandirregularshape

Injected dust trajectories were recorded by the fast CCD visible

camerainstalledinKSTAR.Fig 3 (a)showsasnapshotofthevideo

TheseveralmilligramsofWpowderinjected heatupupon

enter-ingtheSOLandstartemittinglightinthevisiblespectrum,

gener-atingthebrightregionlocatedintheredsquareinFig 3 (a)

4.2 Image processing using DUMPRO

TheDUMPROroutineswereappliedtothevideoinordertoget

thedust trajectories.Results found bythe algorithm canbe seen

inFig 3 (b).Detecteddusttrajectoriesareplottedinblueoverthe

superimposedframeofthewhole video.Fromtheimage

process-ingresults,twodistinctdustbehaviorswereobserved.First,alarge

whitecloud fallsfromthe dustinjectionpoint towardsthe

diver-torregion Labeledascase 1,this maintrajectory corresponds to

thepowderthatjustleft theinjectorandfallsdownwardsdueto

gravity.This behavior isconsistent withtheDUMBO modelsince

theinjectedWpowderisaccretedinto∼100μ mclusters,asizefor

whichgravityisthedominantforce.Littletonotoroidalmotionis

seen onthe video At the end ofthe case 1trajectory, dust gets

closerto the wall andcools down enough to stop emitting light

inthevisiblespectrum,andtheydisappearfromthevideo.During

theendofthecase 1trajectory, otherdustgrainsareobservedon

thebottom-left corner ofFig 3 (b),beingmore isolated and

hav-ing a toroidal motion These dust trajectories will be labeled as

case 2 Assuming that they have a radius of ∼10μ m, the

domi-nantforce actingon themwillbe the iondrag,which isroughly

orientedalongthemagneticfieldlines

The dust grainsfromcase 2can eitherbe the resultofgrains

fromcase 1havingexperienced abouncing dust/wallcollision, or

grainsthat wereisolated fromthe dustcluster ofcase 1 atsome pointduringitsfallingtowardsthedivertor.Inbothcasesitisnot illegitimateto considerthat thetrajectoriesincase 2are isolated

∼10μ m dust grainssince the ∼100μ m size clustersfromcase 1

could be broken upon the eventual dust/wall collision or simply duetointernalforces.Sincetheendofthecase 1trajectorycannot

be observed withthe CCDcamera dueto toolow dust tempera-ture,noconclusionscanbemadeonthispoint andcases 1and2

willbetreatedseparatelylateron

4.3 Comparison with DUMBO simulations and discussion

Comparisonof observeddusttrajectorieswithsimulations has alreadybeenperformedonMAST[19] ,LHD[20] andTEXTOR[21] , usingstereoscopicobservationsinthelattercase.Since no binoc-ular view is available in KSTAR, the dust trajectories given by DUMPRO are 2D, result of 3D trajectories projected in the cam-erasensorplane Inordertocomparewithsimulateddust trajec-tories generatedwith DUMBO,3D trajectories are recreatedfrom themeasured 2D onesby assuming the following:(i) ForCase 1, since thedust areheavy ( d ∼100μ m) andhave agravity driven motion,we assume the trajectory to remain ata chosen toroidal angle.(ii) Concerning Case 2,dust grainsare lighter (d ∼10μ m) andhave an ion dragforce driven motion,which is roughly ori-entedalong themagneticfieldlines.Thus weassume thedustto remaininachosenfluxtube.Thetoroidalanglewherethecase 1

trajectorywasplacedwaschosen inawaythatitremainsmostly

intheSOL withoutcrossing thewall surface.The case 2 trajecto-rieswere placedonfluxtubes asfaraspossiblefromthe plasma corewhileensuringtheexistenceofasolution

Fig. 4 T i (red) and n e (green) profiles at t = 6 4 s from charge exchange spec-

troscopy and line integrated density, respectively T e profile (blue) obtained by fit-

ting the T i one (For interpretation of the references to colour in this figure legend,

the reader is referred to the web version of this article.)

Note that in order to make this2D-to-3D extrapolation some

features ofthe CCDcamera must beknown: position inthe

ves-sel,focallength,sensorsize,amongothers.Asimplepinhole

cam-era model wasused,andthe camera parameters were chosen to

matchthe background(wall) frame asaccuratelyas possible

Re-sultsofthe2D-to-3Dextrapolationprocessareshownforthecase

1trajectoryandthreetrajectoriesfromcase 2inFig 5

For each of the four trajectories extrapolated in 3D from

the DUMPRO routines results, simulations were made using the

DUMBOcode.The plasmabackgroundwasdeterminedusing

sev-eral diagnosticson discharge#13101: EFIT data forthe magnetic

equilibrium and poloidal magnetic field, charge exchange

spec-troscopy forthe iontemperature(T i) profile,lineintegrated

den-sityfortheelectrondensity(n e).ProfileswereextendedintheSOL

usingexponentialdecaysrespectingaC1 matchwiththecore

pro-files The n e profile wasdetermined from the integrated density

using a square root profile in the core, andwe assumed T e=T i

Finally,quantitieswereassumedtoremainconstantoverflux sur-faces.Profilesforn eandT iareprovidedinFig 4 Thetoroidal mag-neticfieldwas∼3T,andtheplasmaionsflowvelocitymapcanbe seeninthebackgroundofFig 5 (a).Theionflowispredominantly parallel,eventhough E × Band∇B × Bdriftvelocitiesaretaken intoaccount

Inthe simulations, thedust grainswere initiatedatthe same location and with the same velocity vector asthe first point of eachexperimentaltrajectory.Theinitialdustradiiwere100μ mfor

case 1and10μ mforcase 2.ResultsareplottedinFig 5 alongwith theexperimentaltrajectoriesextrapolatedin3D Onecanseethat theagreementbetweenexperimentalandsimulatedtrajectoriesis satisfyingincase 1,sincetheyarebothdominatedbygravity Dis-crepancycanbeseenonthetoroidaltrajectory, sincetheiondrag force,whichisdominatedbygravityyetnotnegligible,pushesthe simulated dust in the toroidal direction, counter-clockwise Con-cerning case 2, if simulated trajectories seem close at first, they endupto bemuch shorterthantheexperimental ones This dis-crepancycanbeexplainedbyseveraleffects

First, some cooling mechanisms are not yet accounted for in DUMBO,since SEEisneglected for positivelycharged dustgrains andvaporization/sublimation latentheat cooling isreplaced with

aboiling/sublimationpoint.Theimplementationofthese phenom-enaisunderprogress

Second, it is known that the OML approach used in DUMBO (andotherdustsimulationcodes)presentsseverelimitations,since

it assumes the absence of barriers in the effective potential en-ergy Effectivepotential barrierscan trapa nonnegligible partof the slow incoming ions if d gets to the order of the screening length, which is ∼10μ m in our case[22] On the other hand,if theemittedelectronfluxgetsclosetotheincomingone,potential wellscanformandreducetheelectronemissionitself[1] Another OML limitationappears whilst plasma electrons become magne-tizedwithrespectto d:theirgyrationmotioninducesareduction

intheincomingelectron flux[23] Thesethreeeffectsare not ac-countedforinDUMBOandimpactthedustchargingandheating Third,inthe presentversionof thecode,thematerial ablated

orvaporized from thegrain does not affectit nor the surround-ingplasma.Tobeaccurate, theablatedmaterialcanformacloud

Fig 5 KSTAR dust injection experiment – comparison between dust experimental trajectories, reconstructed with DUMPRO, and simulated ones made with DUMBO: (a) in

a poloidal cross-section, above the ion flow velocity map, with the first wall geometry in white; (b) view from the top of the machine, with the first wall geometry at the mid plane in black

shieldingthe grainfromplasmaheat fluxes InRef [24] ,the dust

radiusabovewhichvaporshielding effectsbecomenonnegligible

wasshowntobe∼1μ m(forWandundertheplasmaparameters

relevantinthisstudy),whichis belowthe dustsizesused inour

simulations.Vaporshieldingmodelshaveshownareductionofthe

evaporationrateuptoanorderofmagnitude[25]

Still, this overheating tendency has been reported on other

codesbased on the samemodel, namelyDUSTT, DTOKSand

MI-GRAINe.The MIGRAINe code wasused tocompare withdust

in-jectedinTEXTOR.The dustgrainsweresmallerthaninourstudy

(<5μ m)andinjectedfromthetopofthemachine.Asimilar

over-heating trend was observed [21] Compensation for overheating

wasaccomplishedintheDUSTTcodebyincludingalarge

empiri-calreductioncoefficienttotheincomingheatflux[26] Concerning

DTOKS,largerdustsizeswereusedinthesimulationstoreproduce

accuratelytheexperimentaldustlifetimes[19]

5 Conclusions

Imageprocessingroutines(DUMPRO)havebeendevelopedand

allowdetectionofdusttrajectoriesonaCCDcameravideo.Adust

transport simulation code (DUMBO) for trajectories modeling is

alsoavailable 2D-to-3D extrapolation ofmeasured dust

trajecto-ries was applied in the KSTAR dust injection experiment

exam-ple, confirmingthat lighter W dust grainsare more sensitive to

theiondragforcethanlargerones.Comparisonbetween

measure-ments andsimulations showeddiscrepancies due to OML

limita-tionsand/ortothelackofavaporshieldingmodel.Improvements

ontheseaspectsarecompulsoryifDUMBOistobeusedtopredict

thebehaviorofinjecteddustintheWESTtokamak

Acknowledgment

This research was partially supported by Ministry of Science,

ICT,andFuturePlanningunderKSTARprojectandwaspartly

sup-ported by National Research Council of Science and Technology

(NST) under the international collaboration & research in Asian

countries(PG-1314)

References

[1] S.I Krasheninnikov , R.D Smirnov , D.L Rudakov , Plasma Phys Controlled Fusion

53 (8) (2011) 083001

[2] L Vignitchouk , P Tolias , S Ratynskaia , Plasma Phys Controlled Fusion 56 (9)

(2014) 095005

[3] A.Y Pigarov, S.I Krasheninnikov, T.K Soboleva, T.D Rognlien, Phys Plasmas 12 (2005) 122508 http://dx.doi.org/10.1063/1.2145157

[4] M Bacharis, M Coppins, J.E Allen, Phys Plasmas 17 (2010) 042505 http://dx doi.org/10.1063/1.3383050

[5] H.Y Lee, S.-H Hong, J Hong, S.H Lee, S Jang, J Jang, T Jeon, J.S Park, W Choe, Rev Sci Instrum 85 (2014) 11D862 http://dx.doi.org/10.1063/1.4886958 [6] S.-H Hong, C Grisolia, P Monier-Gabet, T.S Team, J Nucl Mater 390–391 (20 09) 10 0–102 http://dx.doi.org/10.1016/j.jnucmat.20 09.01.120

[7] N Endstrasser, F Brochard, V Rohde, M Balden, T Lunt, S Bardin, J.-L Bri- ançon, R Neu, J Nucl Mater 415 (1) (2011) S1085–S1088 http://dx.doi.org/10 1016/j.jnucmat.2010.07.045 Proceedings of the 19th International Conference

on Plasma-Surface Interactions in Controlled Fusion

[8] W.U Boeglin, A.L Roquemore, R Maqueda, Rev Sci Instrum 79 (10) (2008) http://dx.doi.org/10.1063/1.2965001

[9] H.M Mott-Smith, I Langmuir, Phys Rev 28 (1926) 727–763, doi: 10.1103/ PhysRev.28.727

[10] C.T.N Willis, M Coppins, M Bacharis, J.E Allen, Phys Rev E 85 (2012) 036403, doi: 10.1103/PhysRevE.85.036403

[11] H Bufferand, G Ciraolo, L Isoardi, G Chiavassa, F Schwander, E Serre, N Fe- dorczak, Ph Ghendrih, P Tamain, J Nucl Mater 415 (2011) S589–S592 http: //dx.doi.org/10.1016/j.jnucmat.2010.11.037

[12] J.P Gunn, R Dejarnac, P Devynck, N Fedorczak, V Fuchs, C Gil, A Kocan,

M Komm, M Kubic, T Lunt, A Monier-Garbet, J.Y Pascal, F Saint-Laurent, J Nucl Mater 438 (2013) S184–S188 http://dx.doi.org/10.1016/j.jnucmat.2013.01

055 [13] P Tolias , Plasma Phys Controlled Fusion 56 (12) (2014) 123002 [14] S Dushman, Rev Mod Phys 2 (1930) 381–476, doi: 10.1103/RevModPhys.2.381 [15] G.L Delzanno, X Tang, Phys Plasmas 21 (2014) 022502 http://dx.doi.org/10 1063/1.4864190

[16] R Behrisch, W Eckstein, Sputtering Part Bombardment, 110, Springer Berlin Heidelberg, 2007, doi: 10.1007/978- 3- 540- 44502- 9

[17] R Marek, J Straub, Int J Heat Mass Transfer 44 (2001) 39–53 http://dx.doi org/10.1016/S0017- 9310(00)00086- 7

[18] S Ratynskaia , C Castaldo , H Bergsåker , D Rudakov , Plasma Phys Controlled Fusion 53 (7) (2011) 074009

[19] G.D Temmerman , M Bacharis , J Dowling , S Lisgo , Nucl Fusion 50 (10) (2010)

105012 [20] M Shoji , H Kasahara , M Tokitani , T Seki , K Saito , S Kamio , R Seki , Y Tanaka ,

A Pigarov , R Smirnov , G Kawamura , H Tanaka , S Masuzaki , Y Uesugi , T Mu- toh , T.L.E Group , Nucl Fusion 55 (5) (2015) 053014

[21] A Shalpegin , L Vignitchouk , I Erofeev , F Brochard , A Litnovsky , S Bozhenkov ,

I Bykov , N den Harder , G Sergienko , Plasma Phys Controlled Fusion 57 (12) (2015) 125017

[22] V Fortov, A Ivlev, S Khrapak, A Khrapak, G Morfill, Phys Rep 421 (1–2) (2005) 1–103 http://dx.doi.org/10.1016/j.physrep.20 05.08.0 07

[23] V.I Demidov, S.V Ratynskaia, K Rypdal, Rev Sci Instrum 73 (10) (2002) 3409–3439 http://dx.doi.org/10.1063/1.1505099

[24] B.T Brown, R.D Smirnov, S.I Krasheninnikov, Phys Plasmas 21 (2) (2014) http: //dx.doi.org/10.1063/1.4866599

[25] S.I Krasheninnikov, E.D Marenkov, J Nucl Mater 463 (2015) 869–872 http: //dx.doi.org/10.1016/j.jnucmat.2014.10.077

[26] R.D Smirnov, S.I Krasheninnikov, A.Y Pigarov, A.L Roquemore, D.K Mans- field, J Nichols, J Nucl Mater 415 (1) (2011) S1067–S1072 http://dx.doi.org/10 1016/j.jnucmat.2011.01.004 Proceedings of the 19th International Conference

on Plasma-Surface Interactions in Controlled Fusion