ELM Energy/Particle Plasma Losses and Fluxes on PFCs · Physics of ELM instability and experimental...

A. Loarte 2004 International Sherwood Theory Conference � Missoula � Montana 28 � 4 � 2004 1

ELM Energy/Particle Plasma Losses and Fluxes on PFCs

Alberto LoarteEuropean Fusion Development Agreement

Close Support Unit - Garching

Contributions from : G. Saibene, R. Sartori, T. Eich, M. Kempenaars, M. Becoulet, G. Federici, P. Snyder, P. Ghendrih, G. Huysmans, C. Pérez, R. Koslowki, I. Nunes,B. Gonçalves, C. Silva, W. Fundamenski, T. Petrie, A. Leonard, M. Fenstermacher,J. Boedo, J. Stober, N. Oyama, Y. Kamada, N. Asakura, G. Counsell, A. Kirk,S. Saarelma, J Lönnroth, V. Parail, �..


Outline of the Talk

1. Introduction

2. Physics of ELM instability and experimental evidence

3. Main plasma ELM energy/particle losses

4. ELM energy/particle fluxes to PFCs

5. Regimes with high Pped + small/no ELMs

6. Conclusions


Introduction (I)H ~ 1 at n ~ nGW in Type I ELMy H-modes requires high Wped

0.4 0.6 0.8 1.00.0

0.2

0.4

0.6

Open symbols Type III ELMs

Wpe

d/Wdi

anped/nGW

DOC-L 1.2MA/1.2TDOC-L 1.2MA/1.2TDOC-L 2.0MA/1.8TDOC-L 2.0MA/2.4TDOC-L 2.0MA/2.4T DOC-L 2.0MA/3.0TDOC-L 3.0MA/3.0TDOC-U 2.0MA/2.4T

PINP = 12MWDOC-U 2.0MA/2.4T



PINP = 17MWDOC-U 2.5MA/2.7THT3 1.1MA/1.2THT3 2MA/2.2THT3 2MA/3.2THT3 2.5MA/2.3THT3 2.5MA/2.7THT3 2.5MA/2.7THT3 2.5MA/3.4THT3 2.5MA/3.4THT3 3MA/2.7THLT 3MA/2.5T

! weak Wped/Wdia dependence on :q95 (2.8 < q95 < 5.2), Pinp , δ (δ > 0.27)

! lower Wped/Wdia @ low Ip & Type III ELMs

Wped/Wdia (JET-Type I ELMs) ~ 0. 4 ± 0.1

JET-Loarte APS�03

JET-Saibene EPS�03

general result for Type I ELMyH-modes (ASDEX-U, DIII-D, JET, �)

ITER


Introduction (II)Pedestal plasma experiences quasi-periodic relaxations " ELMs

∆WELM small fraction of Wplasma (<10 %) to divertor/wall in ~ 200 µs " Large Energy Flux

JET � Type I ELM - Saibene Only Pedestal Region of Plasma Affected by ELMs

JET � Type I ELM - Saibene

Main Plasma


Introduction (III)ITER rough estimates

Wdia = 350 MJ " Wped = 90 � 150 MJ, ∆WELM = 10 - 30 MJ(Adiv = 3 m2, τELM= 300 µs)

TmaxELM > Tev,melt

C,W " ELM Erosion

~ 3.5 µm/ELM + C-target = 2 cm " 6000 ELMs !!ΦELMC,W(MJm-2s-1/2 ) ≤ 50 for small ELM erosion

ITER-Federici PPCF�03


Physics of ELM instability (I)Peeling-Ballooning model of the ELMs (Connor PPCF�98)

(a) ballooning: n=20

α↑↑↑↑ jedge↓↓↓↓

1.0.9

(b) peeling: n=3

1.0.9√ψ

α↓↓↓↓ jedge↑↑↑↑ α↑↑↑↑ jedge↑↑↑↑

√ψ

(c) ballooning-peeling: n=12

1.0.9√ψ

0.8

JET-Huysmans �01

! Ballooning modes driven by grad Pped! Kink (peeling) modes driven by edge current (large Jbootstrap)! Coupled peeling-ballooning modes


Physics of ELM instability (II)Peeling-Ballooning calculations-experiment comparison

DIII-D-Snyder NF�04

! Pped matched with model! Change of mode instability analysed! Role of jbootstrap evaluated (+ measured)


Physics of ELM instability (III)Role of edge current on ELM triggering demonstrated

JET-Becoulet EPS�03

0.0 1.0 2.0 3.0 4.0 5.0

0.0 1.0 2.0 3.0 4.0 5.0

0.0

1.0

2.0

3.0

4.0

60.5 61.5 62.5 63.5

0.0 2.0 4.0 6.0 8.0

ms

]2

-1χ

[103

iB

Pγ

s]

-1[1

04_

γs

]-1

[104

_I [

10A]

6

t [s]

(a)

(b)

(c)

(d)

0.0 2.0 4.0 6.0 8.0

[10

J]6

Wth

(e)

current

ballooningmodegrowth rate

peeling modegrowth rate

thermal energy content

JET-JETTO-Lönnroth-Parail �04

seen in COMPASS-D, MAST, JET

Ramp-up

Peeling-ballooningPure peeling


Physics of ELM instability (IV)Scaling of pedestal pressure and peeling-ballooning model

JET-Loarte APS�03

1 2 3 41000

10000

~ Ip2

DOC (δ ~ 0.3) HT3 (δ ~ 0.45)

P ped (

Pa)

Ip (MA)

PpedType I ~ Ip2(ballooning-like)

Pped depends weakly on : q95 and ne,ped/nGWPpedType I ~ Ip

DIII-D-Snyder NF�04

comparison of model with various experiments required to identify physics processes


Physics of ELM instability (VI)precursors seen before ELMs in manyexperiments with n = 1-10

JET-Pérez NF�04

! precursors� n as expected from P-B

! precursors do not grow as linearly unstable modes et/τ

! role of precursor on ELM trigger is unclear


Physics of ELM instability (VII)Filaments are seen to appear in the plasma boundary prior to

ELMs (n = 10, q =4)MAST-Akers PPCF�03

how much energy flows in the filaments?


Physics of ELM instability (VIII)Analyses of energy fluxes far from strike point in ASDEX

Upgrade are consistent with n = 8 � 24 modesASDEX Upgrade-Eich PRL�04

energy content in stripes less than 3% of ∆WELMdiv !!


Physics of ELM instability (X)ELM spatial and temporal ballooning character

DIII-D-Petrie NF�03 ASDEX Upgrade-Nunes EPS�03 subm. NF�04

similar results on divertor fluxes from ASDEX Upgrade, MAST, JET


Physics of ELM instability (XI)Timescale of ELM pedestal collapse/MHD phase timescale :

collapse of pedestal - hot e- impact on divertor - MHDJET-Loarte APS�03


Physics of ELM instability (XII)Timescale of ELM pedestal collapse ~ 200 - 300 µs for JET Type I ELMs

Similar timescales in JT-60U, DIII-D, ASDEX Upgrade MAST

0.3 0.4 0.5 0.6 0.7 0.80

100

200

300

400

500

1.2MA q95 = 3.12MA q95 = 3.73MA q95 = 3.1

τELM

X-ra

y (µs

)

nped/nGreenwald

0.3 0.4 0.5 0.6 0.7 0.80

100

200

300

400

500

1.2MA q95 = 3.12MA q95 = 3.73MA q95 = 3.1

τELM

MH

D (µ

s)nped/nGreenwald

JET-Loarte APS�03


Physics of ELM instability (XIV)Physical mechanisms of ELM growth

! Non-linear explosive ballooning (Cowley PPCF�03)τELM ~ (τE τA

2)1/3

τA ~ qRn1/2/B, τE ?JET : τELM (τE) ~ 50-100 µs but τELM ~ Rα with α ~ 5/3

τELM ~ (Ip-2/3 n1/3) x (P-2/9 n1/6 Ip1/3) ~ Ip �1/18 (if n ~ Ip, P ~Ip) !!!

! Edge Reconnection (Igitkahnov EPS�01)τELM ~ τA(τη/τA)β (β =[1/3,1])

τELM (β=1/2) ~ (R1/2 Ip-1/2 n1/4) x (a1/2 T3/4) ~ R Ip1/2 (if n ~ Ip, T ~Ip)

More experimental and theoretical work is needed!!!


Physics of ELM instability (XV)

how filaments evolveto an edge plasma collapse ?

MAST-Kirk PPCF�04


Main plasma ELM energy/particle losses (I)

ELM energy/particle losses depend on pedestal parameters

ITPA-Loarte PPCF�03

Small ∆WELM/Wped seen at high ne,ped/nGW and/or high ν*ped(neo)


Main plasma ELM energy/particle losses (II)

Simple picture : small ∆WELM #" small volume of plasma affected by ELM

Peeling/Ballooning Picture : increasing nped - ν*ped

n of unstable modes increase #" poloidal width of modes decrease

ELM affected volume (VELM) " ∆WELM decreases

0.0 0.2 0.4 0.60.00

0.05

0.10

0.15

0.20DOC-U 2MA q95 = 3.7 δ = 0.32

∆WEL

M/W

ped

ν*(neo)0.85 0.90 0.95 1.00

0.0

0.2

0.4

0.6

0.8

1.0

Low νped

Medium νpedHigh νped

Mod

e am

plitu

de (a

.u.)

r/a

JET-Loarte APS�03


0.6 0.7 0.8 0.90

1000

2000Pulse No. 55973 DOC-U 2MA q95=3.6 δ=0.32

T e(eV

)

r/a

Before ELMAfter ELM

Main plasma ELM energy/particle losses (III)

0.6 0.7 0.8 0.9

0.0

0.2

0.4

0.6

0.8

1.0

DOC-U 2MA q95=3.6 δ=0.32 ∆WELM = 0.28 MJ (low ne) ∆WELM = 0.17 MJ (medium ne) ∆WELM = 0.11 MJ (high ne) Type III

Nor

mal

ised

∆T e,

ped/T

e,pe

d

r/a

VELM from fast ECE

decrease of ∆TELM not of VELM " ∆WELM

similar results in DIII-D (Leonard) 0.0 0.5 1.00.0

0.1

0.2

0.3

0.4

0.5

∆Te,

ELM/T

e,pe

d

ν*(neo)

Type I ELMs Type III ELMs

∆WELM down by ~ 3 at VELM ~ constant !!∆WELM #"ELM affected Volume

JET-Loarte APS�03


Main plasma ELM energy/particle losses (IV)

Variation of ∆WELM #" ELM energy transport (DIII-D, JET)

∆WELM = ∆WELMcond (∆TELM) + ∆WELM

conv (∆nELM)

∆WELMcond decreases with ne,ped (or νe,ped(neo))

Smaller ∆WELM #" convective ELMs

DIII-D-Leonard PPCF�02


Main plasma ELM energy/particle losses (V)

0.0 0.2 0.4 0.6 0.80.00

0.25

0.50Inc

reasin

g q95

2MA q95 = 3.7 2MA q95 = 4.7 2.5MA q95 = 3.6 2.5MA q95 = 4.5

DOC-L δ = 0.27 HT3 δ = 0.45

∆Te,

ELM

/Te,

ped

ν*(neo)

∆WELM #" ∆TELM

MHD pedestal stability affects ∆WELM through ELM energy conduction

0.0 0.2 0.4 0.6 0.80.00

0.05

0.10

0.15

0.20

0.25 2MA q95 = 2.8 2MA q95 = 3.7 2MA q95 = 4.7 2.5MA q95 = 3.0 2.5MA q95 = 3.6 2.5MA q95 = 4.5

Increa

sing q

95

∆WE

LM/W

ped

ν*ped(neo)

DOC-L δ = 0.27 HT3 δ = 0.45

ITER

0.0 0.1 0.2 0.3 0.4 0.5 0.60.00

0.05

0.10

0.15

0.20

Increa

sing q

95

2MA q95

= 3.7 2MA q95 = 4.7 2.5MA q

95 = 3.6

2.5MA q95 = 4.5

DOC-L δ = 0.27 HT3 δ = 0.45

∆ne,

ELM/n

e,pe

d

ν*(neo)

high q95 + δ " small ∆WELM at low ν*ped - npedJET-Loarte APS�03


Main plasma ELM energy/particle losses (VI)

0.6 0.7 0.8 0.9

0.0

0.2

0.4

0.6

0.8

1.0

Nor

mal

ised

∆T e,

ped/T

e,pe

d

r/a

Pulse No. 59353 HT3 2.5MA/2.7T Pulse No. 58112 HT3 2.5MA/3.4T

0.0 0.2 0.4 0.6 0.80.00

0.25

0.50

Increa

sing q

95

2MA q95 = 3.7 2MA q95 = 4.7 2.5MA q95 = 3.6 2.5MA q95 = 4.5

DOC-L δ = 0.27 HT3 δ = 0.45

∆Te,

ELM

/Te,

ped

ν*(neo)VELM ~ constant with q95!!

effect of high q95 on P-B stability largest at high δ

ASDEX Upgrade-Saarelma NF�03

P-B stability effect on ELMs "non-linear evolution


Main plasma ELM energy/particle losses (VII)

0.6 0.7 0.8 0.9

0.0

0.2

0.4

0.6

0.8

1.0

Nor

mal

ised

∆T e,

ped/T

e,pe

d

r/a

Pulse 55935 DOC-L 2MA/2.4T Pulse 55955 DOC-U 2MA/2.4T Pulse 58409 DOC-U 2.5MA/2.7T Pulse 59353 HT3 2.5MA/2.7T

Plasma shape (δ, κ, �) changes VELM as expected from P-B analysis(JET & DIII-D (TTF�02-Córdoba) not ASDEX Upgrade (Urano PPCF�03))

0.0 0.2 0.4 0.6 0.80.00

0.05

0.10

0.15

0.20

0.25

∆WEL

M/W

ped

ν*(neo)

DOC-L 2MA/2.4T q95=3.7 δ=0.27DOC-U 2MA/2.4T q95=3.7 δ=0.32DOC-U 2.5MA/2.7T q95=3.1 δ=0.32HT3 2MA/2.2T q

95=3.6 δ=0.45

HT3 2.5MA/2.7T q95

=3.6 δ=0.45

lower ∆WELM/Wped for same ν*ped at high δbut

∆WELM is larger for higher δ (higher Wped) !!!

JET- Loarte PPCF�02JET- Loarte APS�03


Main plasma ELM energy/particle losses (VIII)

Collapse of effect of pedestal plasma at ELM is not Θ symmetricASDEX Upgrade-Nunes EPS�03 subm. NF�04

√ψ

0.8

α↑↑↑↑ jedge↑↑↑↑

1.0.9

ballooning structure is maintained to the end of the ELM collapseSimilar results in JT-60U (Oyama NF�03 & NF�04) & MAST (Kirk PPCF�04)

flux mapped profiles


Main plasma ELM energy/particle losses (IX)

Delay scales with LFS-HFS ion transit time!

Collapse of pedestal plasma during ELMs is not simultaneous at all Θ

ASDEX Upgrade-Nunes EPS�03 subm. NF�04


Main plasma ELM energy/particle losses (X)ELMs are triggered at precise values of ne,ped and Te,ped

ELM crash has a much wider scatter for same ne,ped and Te,ped

0.0 0.5 1.0 1.5 2.00

10

20

30

40

Freq

uenc

y

∆WELM/<∆WELM>0.0 0.5 1.0 1.5 2.00

10

20

Freq

uenc

y

∆Tped,ELM/<∆Tped,ELM>0.0 0.5 1.0 1.5 2.00

10

20

30

Freq

uenc

y∆nped,ELM/<∆nped,ELM>

0.0 0.5 1.0 1.5 2.00

10

20

30

40

Freq

uenc

y

nped,ELM/<nped,ELM>

0.0 0.5 1.0 1.5 2.00

10

20

30

Freq

uenc

y

Tped,ELM/<Tped,ELM>

JET-Loarte APS�03ne,ELM

Te,ELM

∆Te,ELM ∆ne,ELM∆WELM


ELM energy/particle fluxes to PFCs (I)

Eich submitted NF 2003

∆WELM " divertor and main chamberParameters to be determined & understood :! timescales and areas! transport mechanisms (extrapolation)

Outerdivertor

Innerdivertor

ELMs lead to fluxes both at the divertor and main chamber PFCsASDEX Upgrade-Eich PRL�04 ASDEX Upgrade-Herrmann EPS�03


ELM energy/particle fluxes to PFCs (II)

divertor ELM fluxes are toroidally symmetric near separatrix (long Lc/high S)ASDEX Upgrade-Eich PRL�04DIII-D Leonard JNM�97 (Loarte PPCF�03)

but ELM currents in DIII-D?


ELM energy/particle fluxes to PFCs (III)

Outerdivertor

Innerdivertor

divertor ELM energy wetted area ≈ divertor wetted area between ELMs

ASDEX Upgrade-Herrmann PPCF�02 (Loarte PPCF�03)

┴ B & II B are enhanced in a similar way during ELMsSimilar results in MAST (Kirk PPCF�04) and DIII-D (Fenstermacher PPCF�03)

JET-Eich PSI�04


ELM energy/particle fluxes to PFCs (IV)

Significant fraction of ∆WELMdiv arrives after Tmax

surf !!!!ELM erosion in ITER determined by Tmax

surf > TevapC

Tmaxsurf

Time history of divertor ELM energy is complexJET-Eich PSI�02, Matthews NF�03


ELM energy/particle fluxes to PFCs (V)

0.1 10.0

0.1

0.2

0.3

0.4

0.5DOC-L δ = 0.27

1.2MA q95 = 3.12MA q95 = 3.72MA q95 = 4.63MA q95 = 3.1

∆WE

LMou

t(0<

t <τ IR

)/∆W

ELM

out t

ot

ν*(neo)

nped or ν*(neo) " ELMs more convective " more of ∆WELMdiv after Tmax

surf

0.40 > ∆WELMdiv (0<t<τIR)/ ∆WELM

divtot > 0.15

conductive ELM convective ELM

ratio

JET-Loarte APS�03, Eich PSI�04


ELM energy/particle fluxes to PFCs (VI)

Timescale of ELM heat flux on divertor correlated with τII ~ L/cs,ped

PIC-Bergmann NF�02τELMMHD

JET- Eich, ASDEX Upgrade-Herrmann,JT-60U Asakura, MAST-Kirk

Experiment " high Γx-ray (electrons ~ keV) during τELMMHD

! Γdiv = Γe,div(τELMMHD) + Γi,div(τII)

! Formation of high energy sheath


ELM energy/particle fluxes to PFCs (VII)

� ∆WELM/Wped decreases with τII

cRq

pedsFront

,952ππππττττ ====||

Physics Model :

1) Pedestal connects to divertor for τELMMHD

2) Energy flow restricted by sheath (τII) " ∆WELM/Wped ~ (1-exp(-τELMMHD/τII))



ELM energy/particle fluxes to PFCs (VIII)separation of e- flux (τELM

MHD) & ion flux inner divertor

Similar results from DIII-D (Fenstermacher PPCF�03, Boedo APS�03)

150 200 250 300 350 4000

200

400

600

800Empty symbols Type IIIDOC-L δ=0.27

~ 2.2 x (ττττII-175)

∆tD

α (µs)

τII (µs) = L/csped

1.2MA q95 = 3.11.2MA q95 = 3.12.0MA q

95 = 2.8

2.0MA q95 = 3.72.0MA q95 = 3.72.0MA q95 = 4.63.0MA q95 = 3.1

JET-Loarte, PPCF�02, PPCF�03, APS�03

∆tDα ~ 0 for finite τII


ELM energy/particle fluxes to PFCs (IX)

∆tDα ~ 0 for finite τII : a) main plasma collapse (τELMMHD ≠ 0) ?

b) change of ELM start X-point " midplane ?DIII-D-Fenstermacher PPCF�03 ASDEX Upgrade-Nunes EPS�03 subm. NF�04

In-out divertor

HFS-LFS main plasma


ELM energy/particle fluxes to PFCs (X)

0.05 0.10 0.15 0.200.0

0.2

0.4

0.6

0.8

1.0

1.2DOC-L δ = 0.27 1.2MA q95 = 3.1

2MA q95 = 3.72MA q95 = 4.63MA q95 = 3.1

∆WEL

MIR

/∆W

ELM

∆WELM/Wped

larger ∆WELM/Wped " smaller ∆WELMdiv /∆WELM

Transient radiation #" ∆WELMdiv < ∆WELM

∆WELMdiv < ∆WELM but λELM

div ~ λinter-ELMdiv

JET-Loarte APS�03, Eich PSI�04ASDEX Upgrade Herrmann EPS�97 (Loarte PPCF�03)


ELM energy/particle fluxes to PFCs (XI)

vr (km/s): JET (0.5-1.2, Gonçalves PPCF�03, Fundamenski�04), MAST (0.2-1.7, Counsell PPCF�02, Kirk PPCF�04), DIII-D (0.2-0.6, Zeng PPCF�03, Boedo APS�3)

large particle fluxes measured far from separatrix at ELMsDIII-D-Boedo+Zeng APS�03/PPCF�04 MAST-Counsell PPCF�02

τperp,ELM~100 µs~τII " large Γmain chamber


ELM energy/particle fluxes to PFCs (XII)

significant energy fluxes measured on main chamber PFCs

Energy on PFCs on large area of limiter (most ELMs)∆WELM

wall ~ 25 % of ∆WELM

Te,ELMPFC << Te,ped " Ti,ELM

SOL

ASDEX UpgradeHerrmann EPS�03


ELM energy/particle fluxes to PFCs (XIII)What do we know about radial ELM fluxes ?

complex filamentary

rotating structure

DIII-D-Boedo-Fenstermacher APS�03 /PPCF�03 DIII-D-Boedo APS�03

Te,ELM decays radially faster than

ne,ELM

τE-e,SOL < τII

vr,ELM decreases with ne (and r ?)


ELM energy/particle fluxes to PFCs (XIV)

decrease of vr with ne (and ∆WELM?) compatible with ∆WELM

div/∆WELM

0.05 0.10 0.15 0.200.0

0.2

0.4

0.6

0.8

1.0

1.2DOC-L δ = 0.27 1.2MA q95 = 3.1

2MA q95 = 3.72MA q95 = 4.63MA q95 = 3.1

∆WEL

MIR

/∆W

ELM

∆WELM/Wped

JET-Loarte APS�03, Eich PSI�04

DIII-D-Zeng PPCF�04


δδ

δ

ELM energy/particle fluxes to PFCs (XV)What determines vr?

1.Non-linear ballooning explosive mode dependence on (ne, Te, ν*ped)

S.Cowley-PPCF�03

JET-Becoulet-H-mode WS�01

ELM Bθ �amplitude� decreases with nped~


ELM energy/particle fluxes to PFCs (XVI)

vr = cs 2ρ2L/Rrb2

! JET : Tped ~ 1.5 keV, if rb ~ 20 cm " vr ~ 1 km/s! DIII-D : Tped ~ 0.5 keV, if rb ~ 13 cm " vr ~ 1 km/s

Scaling of vr ~ Tped3/2 rb

-2 ~ nped-3/2 rb

-2

2. Dynamics of ELM pressure blob radial transport (Krasheninnikov, D�Ippolito,�.)

dependence of rb on nped, Tped , 2 hypothesis :

! Empirical ∆NELM ~ nped (DIII-D, JET, ASDEX Upgrade)

rb ~ constant

! rb ~ P-B mode width " rb ~ nped−α

ASDEX Upgrade-Nunes EPS�03 subm. NF�04


Extrapolation of Type I ELMs to ITER

WdiaITER = 350 MJ " Wped < 0.4 Wdia

ITER =140 MJ (nped~8 1019 m-3 Tped~4.3 keV)

τIRELM ~ 300 µs

AELMITER = A S.S

ITER = 3 m2 (λpowermidplane = 5 mm)

Convective ELMs Conductive ELMs

∆WELMITER (MJ) 10 28

∆WELMITER, div(MJ) 80% 50%

∆WELMITER,div (0 < t < τIR) (MJ) 20% 40%

ΦELMITER (MJm-2s-1/2 ) 20 72

C-ablation in ITER ~ 50 MJm-2s-1/2

∆WELMITER,wall (MJ) 20% 50%

ELM erosion less critical than previously believed even for conservative assumptions


Conclusions (I)

$ Peeling-Ballooning model provides a reasonable description ofpedestal pressure limitation by Type I ELMs

! Detailed multi-machine comparison (& pedestal width scaling) in progress

! Which is the ELM triggering mechanism ?

! Non-linear evolution of ELM and timescales

$ Type I ELM energy losses determined by ne,ped & Te,ped, q95, δSmall ELMs #" Convective ELMs

! What determines VELM ? ! Physical process that produce convective ELMs (ν*ped and high δ/q95) ?

(link between MHD stability #" ELM energy transport ?)! What determines the ELM variability ?


Conclusions (II)

$ ELM energy fluxes determined by determined by ne,ped & Te,ped

! timescale of ELM energy flux on divertor by τII and not τELMMHD

(role of sheath/IIB transport on divertor ELM energy Flux ?)

! ∆WELMdiv < ∆WELM and fluxes to main chamber PFCs

(convective radial ion transport versus IIB losses ?)

! spatial distribution of ELM heat loads

(toroidal symmetry ?, main chamber fluxes ?, in/out divertor balance ?)


Conclusions (II)

$ ELM energy fluxes determined by determined by ne,ped & Te,ped

! timescale of ELM energy flux on divertor by τII and not τELMMHD

(role of sheath/IIB transport on divertor ELM energy Flux ?)

! ∆WELMdiv < ∆WELM and fluxes to main chamber PFCs

(convective radial ion transport versus IIB losses ?)

! spatial distribution of ELM heat loads

(toroidal symmetry ?, main chamber fluxes ?, in/out divertor balance ? )

$ Regimes with small/no Type II ELMs/good confinement exist but:! operational space is narrow and not well characterised

(reproducibility in various experiments ?) ! extrapolability to next step devices/compatibilty with other requirements ?

(low ν*ped, low q95, high Praddivertor, pellet fuelling, He pumping, �.)

! operation close to double null required ?


Regimes with high Pped + small/no ELMs (I)

Regimes with small or no ELMs and high Pped

a) Regimes with high Pped and small ELMs (Type II ELMs : JT-60U, AUG)

modification of edge MHD plasma stability

b) Regimes with no (or infrequent) ELMs and quasi-continuous losses (EDA, QH-mode, �mixed Type I-II ELMs� in JET)

triggering of transport losses �between ELMs� Pped ≤ Ppedlimit


Regimes with high Pped + small/no ELMs (II)

Mixed Type I-II in JET #" increased broadband MHD fluctuation at low f Tped reaches steady state between type I ELMs with increased fluctuations but

nped does NOT saturate in the Type II phases (" end in Type I)

Increased Inter-ELM transport (+ washboard modes)

JET-SaibeneEPS�03


Regimes with high Pped + small/no ELMs (III)

QH (and EDA) no-ELM regimes are associated with coherent MHD modes at the edge which lead to enhanced transport

DIII-D Doyle PPCF�02, Burrell

PPCF�03

potentially interesting for next step devices but triggering of mode?

reproduced in ASDEX Upgrade

Suttrop PPCF�04


Regimes with high Pped + small/no ELMs (IV)

JT-60 Type II ELM (= Grassy ELM) discharges are obtained at high δ, high βp and high q95 (low ν*ped)

JT-60U � Kamada PPCF�02

0.01

0.1

1

10

0.5 1.0 1.5 2.0 2.5

JT60-U

, Type I (high βp and H-mode)

, Type II and mix (high βp)

Type III (H-mode)

βp

ν *

change of ELM Type because of edge stability change due to Shafranov-shift

reproduced at JET Saibene EPS�04


Regimes with high Pped + small/no ELMs (V)

! PpedType I ~ Pped

Type II transition! Regime is robust to Pin ↑ (if fuelling adjusted)! attributed to change in edge P-B stability! compatible with high β (βN , ν*ped interplay?)

Type II (0.35<δ<0.42, 0.8<Ip(MA)<1)Type III

ASDEX-Upgrade-Stober NF�02ASDEX-Upgrade-Sips NF�02

Type II ELMs in ASDEX Upgrade occur at high nped (ν*ped)


Regimes with high Pped + small/no ELMs (VI)

! Type II require medium/high δ and ne

! Proximity to DN configuration is essential (no type II for ∆Xmp>2cm) + Trade-off δ/q95 ? ! High β not required, but compatible with the regime! (βN~3 obtained)

not reproduced at JET Saibene EPS�03

3.0 3.5 4.0 4.5 5.00.25

0.30

0.35

0.40

0.45

0.50

ASDEX-U

q95

vs average δ

<δ >

q95

Type I Type III Type I+II Type II

-3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.00.25

0.30

0.35

0.40

0.45

0.50

ASDEX-U

Type I Type III Type I+II Type II

<δ>

∆X separatrices midplane (cm)

Type II in ASDEX-U : Quasi DN configuration


Regimes with high Pped + small/no ELMs (VII)

ASDEX Upgrade-Saarelma NF�03

proximity to DN (#15863) leads to a sharp decrease of eigenmode width


Physics of ELM instability (V)Growth of linear mode coincident con ELM crash (low ne)

DIII-D-Snyder APS�01

Similar results for giant ELMs in JET - Huysmans


Physics of ELM instability (IX)

JET-Ghendrih JNM�03Between ELMs

Analyses of particle impact patterns on JET main wall are consistent with ballooning mode n = 12, m = 50

At ELMs


Physics of ELM instability (XIII)JT60U-Oyama PPCF�01

τELM ~ 200 µs

DIII-D-Fenstermacher PPCF�03

τELM ~ 200 µs


Main plasma ELM energy/particle losses (XI)

ELM perturbation into edge plasma much broader than P-B modesDIII-D-Leonard APS�02

Relation between ∆WELM

and P-B modes ?


ELM energy/particle fluxes to PFCs (XVII)

� ∆WELM/Wped decreases with τII

cRq

pedsFront

,952ππππττττ ====||

Physics Model :

1) Pedestal connects to divertor for τELMMHD

2) Energy flow restricted by sheath (τII) " ∆WELM/Wped ~ (1-exp(-τELMMHD/τII))



ELM energy/particle fluxes to PFCs (XVII)

In/Out ELM energy balance changes at ELM and with divertor/main plasma conditions

DIII-D-Leonard PSI�02, Fenstermacher PPCF�03

consistent with Ndivin > Ndiv

out + sheath or changes in with energy flow with ne

Energy fluxes at ELMs

Inter-ELMndiv

in > ndivout

Tdivin < Tdiv

out

qdivin < qdiv

out

Γdivin > Γdiv

out


Out

In

ELM energy/particle fluxes to PFCs(XVIII)

0.0 0.5 1.0 1.5 2.00

10

20

30

100 µsτELM 500 µs 1 ms

Carbon Atom

/ELM

µm/E

LM

EnergyELM (MJ/m2)

0.0

5.0x1022

1.0x1023

1.5x1023

2.0x1023

2.5x1023

3.0x1023

3.5x1023

4.0x1023

EELM > 1MJ/m2 can lead to ΓC > 1023 C-atom/ELM

ND-TITER ~ 1023 particles

Poor C Retention in Divertor @ ELM(Plasma flows out of the Divertor)

Large Proportion of ΓC may get into Core Plasma " increase of Zeff at edge


ELM energy/particle fluxes to PFCs (XIX)

JET MkII: 1 MJ ELMs

time (s)

Wdi

a (M

J)Z

eff

11

12

10

1 MJ

10

20

Pin

, Pra

d (M

W)

2.0

3.0

Ha

! EtransientC-radiation (3 keV, 8 1019 m-3) ~ 1 keV/atom

EELMRAD ~ 16 MJ << Wplasma (~ 350 MJ)

! JET Results agree qualitatively : Modelling/Extrapolation to ITER necessary

EELMRAD ~ 0.5 MJ(C. Ingesson)

Wp~ 11 MJ∆WELM ~ 1 MJ

Energy loss by Carbon Transient Radiation " Probably smallADAS � M. O�Mullane

nped ~ 8 1019 m-3

Tped ~ 3 � 5 keV

ELM Energy/Particle Plasma Losses and Fluxes on PFCs · Physics of ELM instability and experimental...

Documents

Transcript of ELM Energy/Particle Plasma Losses and Fluxes on PFCs · Physics of ELM instability and experimental...