Aucun titre de diapositive

1
AssociationEuratom-Cea
Edge Localised Modes Physics and Edge Issues in
Tokamaks. presented by Becoulet M.
G. Huysmans (1), Y. Sarazin (1), X.Garbet(1), Ph.
Ghendrih (1), F. Rimini (1), E. Joffrin (1),
Litaudon X. (1), P. Monier-Garbet (1)  , J.-M.
Ané (1), P. Thomas (1), A. Grosman (1), (1)
Association Euratom-CEA, CE Cadarache, F-13108
St. Paul-lez-Durance, France. V.Parail (2), H.
Wilson (2), P. Lomas (2), P. deVries(2) , K.-D.
Zastrow(2), G.F. Matthews (2), J. Lonnroth (2),
S. Gerasimov(2), S. Sharapov(2), M.
Gryaznevich(2), G. Counsell(2), S.Fielding(2), A.
Kirk(2), M. Valovic(2), R. Buttery(2) , (2)
Euratom/UKAEA Association, Fusion Culham Science
Centre, Abingdon, OX14 3EA, UK. G. Saibene (3),
R. Sartori (3), A. Loarte (3)  (3) EFDA Close
Support Unit (Garching), 2 Boltzmannstrasse,
Garching, DE. A.Leonard (4), P. Snyder (4), L.L.
Lao(4), P. Gohil(4), T.E.Evans(4), (4) General
Atomics, 3550 General Atomics Court,P.O.Box 85608
San Diego,CA,U.S.A. Y Kamada (5), A Chankin (5),
N. Oyama(5), T.Hatae(5) ,N. Asakura(5), (5) Japan
Atomic Energy Research Institute (JAERI),
Japan O. Tudisco (6), E. Giovannozzi(6) , F.
Crisanti(6), (6) Associazione EURATOM-ENEA sulla
Fusione, C.R. Frascati, Frascati , Italy C.
P.Perez (7), H. R. Koslowski(7) , (7) Institut
für Plasmaphysik, Forschungszentrum Julich,
Germany T.Eich(8), A. Sips(8), L. Horton(8) , P.
Lang (8), A. Hermann (8), J. Stober(8), W.
Suttrop(8), (8) Association Euratom-IPP, MPI fur
Plasmaphysik, 2 Boltzmannstrasse, Garching,
D-85748, Germany P. Beyer(9), (9) UMR 6633PIIM
CNRS-Université de Provence,F-13397 Marseille
Cedex 20, France. S. Saarelma(10), (10) Helsinki
University of Technology, Euratom-TEKES
Association, FIN-02015 HUT, Finland R.A. Moyer
(11) (11) University of California, San Diego, La
Jolla CA 92093,U.S.A. and contributors to
JET-EFDA Workprogramme.
2
Outline.

• Introduction.
• -High confinement scenarios for ITER and ELMs.
• 2. H-mode scenarios and ELMs (theory
  experiment).
• -Ballooning-peeling linear MHD model.
• - Pedestal and SOL transport, non-linear models.
• -ELM size role of density, triangularity, high
  q95, high bp.
• -High confinement regimes with Type II ELMs for
  ITER?
• Internal Transport Barrier (ITB) scenario and
  ELMs.
• - Combined ITB ETB scenarios.
• Active control of ELMs.
• -Edge ergodisation, edge current, pellets.

3
ELM cycle periodic loss of confinement
ELM plasma edge MHD instabilities typical for
H-modes in tokamaks gt periodic fast (200ms)
relaxations of edge pressure gt energy to SOL
gtdivertorwall.

DIII-D- A. Leonard PPCF2002
JET Ph.Ghendrih JNM (2003)
JET G. Saibene PPCF2002
Dnegt convective
before
after
DTegt conductive
before
after
radius(m)
time(s)
divertor
4
Experimental scaling for ELMs types
Type I low fELM, high Ppedgt high
confinement,but large energy losses per
ELM. Type II regimes in highly shaped plasmas,
high Pped, (confinement like Type I ELMs),
small edge MHD activity, but for narrow
operational window. Type III (at low power or
at high density) higher fELM, small energy
losses per ELM, but lower Ppedgt low confinement.
JET Sartori R. PPCF2003 submitted
H-modeType I
Type II
H-modeType III
L-mode
L-mode
L/H threshold 0.45ne 0.75 BTR2 (MW)
5
ELMs are problematic for ITER.
ITER reference scenarios QPfusion/Padd.10(Aymar
et al 2001) high confinement (H98ygt1) high
density (gt0.8nGR), high d0.5 and acceptable
(material limitsmelting, erosion,evaporationgtre
asonable divertor life-time) heat loads on the
divertor target plates DWELMpedlt5-10MJ (if 60
goes to the divertor S3m2 )(Federici PSI 2002).

6
Experiment evidence from many tokamaks
ballooning structure
-Ballooning structure of ELMsgt collapse of Te,
ne on the LFS. -Parallel SOL transport gt
divertor (50 T.Eich EPS2001 ) -SOL
perpendicular turbulent transport (tails,
blobs) gt wall
MAST G.Counsell 2002
MAST A. Kirk 2003
Te, ne collapse on LFS
Outboard Da
wall
plasma
plasma edge
inboard
7
Linear ideal MHD theory ELMsballooning-peeling
modes.
Linear MHD stability analysis (codes MISHKA,
GATO, ELITE). -Ballooning modes driven by
pressure gradient gt pedestal, outboard (LFS),
high n. -Peeling (kink) driven by edge current
(bootstrap) gtX-point, low n1-4 -Coupled
peeling-ballooning gt LFS, pedestal, n10-20
(JET).
JET(MISHKA) G.Huysmans 9thEFPW 2001
JET M. Becoulet et al PPCF2002
Peeling component
Pedestal shoulder
1
0.8
??
8
Pressure collapse in ELM non-linear modelling

• MISHKA modes structure, growth rate, egt
  g2n(a-acrit) if agtacrit
• TELM (a)

dBr (ergodisation)dV(convection)
M. Becoulet,G. Huysmans et al 2003
ELM
diffusion
telm200ms
0.
0.6 time (s)
0.2
0.4
9
Turbulence modelling ELMs?
Resistive ballooning turbulence (dB0, dF0 )
modelling periodic energy bursts through ETB.
Estimations for ELM time 250ms! More
development needed both with MHD turbulence
(DIII-D, BOUT-X.Xu et al New J. of Phys. 2002)
(P.Beyer ,to be submitted PoP2003)
SOL
core
10
Particle transport in SOL to the inner and outer
divertor.
ELM collapse on the LFS gt inner/outer delay in
Da Dt delay t// (ions) 2pRq95/Cs, ped .
Increases with the density.
JET A. Loarte et al PPCF2002
JT-60U A.Chankin, N. Oyama et al NF2002, PPCF
2001
ELM collapse
dBq/dt
Outer LFS
Inner HFS
11
Type I ELM time tdivELM gt tMHDELM
tMHDELM 150-300 ms (JET), similar in JT-60U,
DIII-D, AUG1ms. Not identified parametric
dependence. Weak?
Energy into divertor is deposited with ion flux
time t//ion gt tdivELM increases with the
density.
JET A.Loarte PPCF2002
IR data
12
Toroidal rotation of ELM
Toroidal asymmetry of Type I ELM in JET (similar
TCV H. Reimerdes NF1998). Propagation in
electron diamagnetic direction tSOL// . Not
explained by linear MHD.
JET(M.Becoulet, G. Saibene 2003)
toroidal Mirnov coils
Broken coils
F3
Low nped
High nped
13
ELM size decreases with the density.
What are the key factors to decrease ELM size
keeping high confinement? Multi-machine
experimental scaling DWELM/Wped decreases with
with ne, ped (nped,, tFront// ).
What physics? ne gt Te
A. Loarte PPCF 2002
-MHDgtbotstrap current
-Pedestal transport?
t// 2pRq95/Cs, ped
-SOL transport?
Not identified yet
Log scale
14
Edge bootstrap current decreases with density.
MISHKA modelling for JET diffusion of edge
bootstrap current improves stability for low n
peeling modes. Main difficulty sensitivity of
stability diagram to small changes in Te, ne, Jz
profiles, no direct measurements of edge current.
JET(MISHKA) G.Huysmans 9thEFPW 2001
stable
unstable
15
ELM size ELM affected area? Open question.
As density increases pedestal width (less
obvious on JET!), bootstrap current , mainly
conductive losses DT/T with
density Modelling gt Radial width of mode
decreases gtSmaller ELMs?
DIII-D (ELITE P.Snyder et al IAEA 2002)
DIII-D (A. Leonard et al, PPCF-2002)
16
Transport modelling(TELM) smaller affected area
smaller ELMs?
TELM ( M. Becoulet et al 2003)
Narrow ELM area DWELM/Wped1.2
Large ELM area DWELM/Wped3
17
ELM size role of plasma shapinggt improved MHD
stability
High triangularity (d) gt higher pedestal
pressure gt higher confinement (AUG, JT-60U,
DIII-D, JET)
JET G. Saibene et al PPCF2002
JET(MISHKA) G.Huysmans 9th EFPW 2001
Low d
High d
Ballooning unstable
ITER
Pressure gradient
stable
stable
kink unstable
Edge current
Similar results for AUG,JT-60U, DIII-D
18
High triangularity (edge magnetic
shear)gtGrassy ELMs in JT-60U
High confinement regimes with small grassy ELMs
recipe gt high magnetic shear dgt0.5-0.6, high
q953.5- 6, high bp2.
JT-60U Y. Kamada et al PPCF2002
high bp(2 ) helps gtgrassy at q953.6 (in
ITER3)
19
Type II ELMs in ASDEX-Upgrade.
Type II ELMs d0.4 (Double Null is
important!), q95gt4.2, n/nGR0.85-0.95(high
density) , H981. Broadband MHD n3,lt30kHz. Low
heat load into divertor. Advanced scenario with
Type II at high bp. (0.8MA/1.7T, 10MW NBI) d0.4
(Double Null configuration) q953 (q01 to avoid
saw-teeth) n/nGR0.88, H 98-P1.2-1.3, bp1.8,
bN3.5 Effect of high bp? -Core confinement is
improved (turbulence bootstrap gtflat shear)
-ELMs Type II at lower q95 3.
AUG A. Sips 9thEFPW 2001
To Double Null
20
Linear ideal MHD (GATO) ELM area is small for
Type II ELMs
ELM affected area decreases at high q95 high d
for the same pressure profile. Double Null
configuration increases edge shear even more.
GATO (for AUG) S. Saarelma et al, NF(2003)
n3
n3
21
JET mixed Type IType II
d0.5 q953.4, n/nGR0.9-1.1, H971. High
density (n0.6-0.8!) gt smaller Type I Type II
broad band MHD lt30kHz, n8 (Washbroad resistive
modes? Ch.Perez NF2003). SN and DN configurations
were tried. Not enough factors JET to suppress
Type I ELMs on JET? And for ITER? Other regimes
w/o ELMs QH (DIII-D), EDA(C-mod)
JET G. Saibene et al PPCF(2002), see EPS 2003
ne0.8nGR
Da
Da
ne1.1nGR
Wdia
Wdia
ne
ne
22
Conclusions (I) ELMs in H-modes

• Ideal MHD transport models describe many
  experimental observations ballooning structure,
  fast relaxation of Pped, MHD ELM timetELM,
  frequency fELM.
• - Type I ELM MHD time ( typical pedestal crash
  time) is found 150-300ms for many machines.
  Parametric dependence is not identified yet.
• -ELM rise time on divertor target is correlated
  with ion // SOL transport.
• - Key factors to decrease ELM size?
• -high pedestal density(collisionalty?)
• -high d, high q95, high bp
• -Regimes with benign Type II ELMs at high d
  demonstrated ITERlike H971, n/nGR0.8-0.9, but
  not for ITER-like parameters (n0.05, bp1,
  q953)gt Low n, high power, high current

23
Double barriers ETBITB high bp with grassy
ELMs.
High bp2, high q956.9, high d0.5 gt ITBETB
with grassy ELMs gt high performance (HHy21.2,
n/nGR0.6) divertor heat load is reduced by
factor 4-5 as compared to Type I ELMs.
JT-60U Y. Kamada PPCF(2002)
24
QDB
Quiescent Double Barrier ITBQH-mode without
Type I ELMs on DIII-D (bN3.5, wide range of q95,
d). But counter NBI injection, nped0.1nGR.
Interesting from the point of view low n
pedestal.
D-III-D P. Gohil 8th IAEA TCM2001
25
ITBType I ELMs ? Type III ELMs?
Usually Type I ELMs are not compatible with large
(r ITBgt 0.5) ITBs in JET, DIII-D ITB erosion by
Type I ELMs. If no pure Type II regimes gt small
Type III ELMs ITB (improved core confinement to
compensate poor edge confinement). But how to
keep Type III edge?
JET M. Becoulet PPCF(2002)
Type I
Type III
wall
Te (ECE)
plasma centre
26
Perturbation from Type I ELM propagates to ITB
region?
Suggestion from theory perturbation from ELM
propagates inside gt fast avalanche-like
transport after an ELM inward outward turbulent
fluxes. Why ITB is affected? Slow (t confinement)
erosion of ITB, not MHD collapse! Rotation shear
is affected ? Mechanism is unknown.
ITB
pressure
Steeper gradient gt unstable
centre
SOL
27
High triangularity ITB on JET.
Main difficulty for ITB scenario at high d0.5 is
Type I ELMs avoidance. JET 2003 ITBs
(3.4T/1.5MA) with Type III edge with D2
n/nGR0.7, H98y1.3, bN1.8, bp1.5 , q957,
lasts 6s.
JET M. Becoulet , P. Lomas, O. Tudisco, F.
Rimini, K.-D. Zastrow et al
28
High triangularitygthigher density
Higher density 0.7 nGR at HT(d0.5) compared to
0.4nGR LT(d0.2), but H89 2 (at LT)gt1.7(at
HT) . Future gt larger ITB r ITBgt0.5gt
performance, higher bp, lower q95.
JET M. Becoulet et al (2003)
JET F.Rimini et al(2003)
High d ITBs
29
Conclusions(II) ELMs in ITBs
-Combined regimes with ITB and ETB ITB with
grassy ELMs at high triangularity, high q95,
high bp were demonstrated in JT-60U. ITBETB
w/o ELMs QDB in DIII-D (but counter NBI, low
n/nGR0.1) -High triangularity ITBs (d0.5,
n/nGR0.7, H98y1.3, bN1.8, bp1.5, q957 ) with
Type III ELMs were demonstrated on JET.
Active control of ELMs -gas puffing -impurity
(increased Pradgt control Te,ped, but impurity
accumulation?) -edge current (Ip ramp-up, -down
experiments gt support peeling-ballooning picture
of ELMs, but very Pped, dIp/dt dependent, large
tres ITER) -edge ergodisation, -pellets
30
ELMs control by dBrexternal?
External control coils dBr(t) gt edge
ergodisation altacrit, or artificial ELMs.
Compatibility with high confinement regimes?
R. Moyer,T. Evans DIII-D (C-coils) EPS2002
COMPASS-D S. Fielding et al EPS2001 TCV A.
Degeling et al 2003
See G. Jackson , EPS 2003 Friday -P-4.47
More planned in 2003
31
Pellets.
Pellets gt increase of n, artificial ELMs are
similar to natural.
With pellets 20 Hz smaller Type I ELMs
ASDEX-Upgrade P. Lang (EPS2002) see also this
conference
W/o pellets 3Hz large compound ELMs
32
Conclusions
(towards ITER integrated scenario) 1. Key
factors to decrease Type I ELM size -high d,
high q95, high bpgt Type II ELMs for
ITER? -increase pedestal density (n,
t//ion,..?) gt understanding of SOL energy
and particles transport during an ELM is missing
for the definitive predictions for ITER. 2.
H-modes and combined advanced scenarios (with and
w/o ITBs) at high triangularity high density
with small ELMs demonstrated ITER like
performance (H97ygt1, n/nGR0.7-0.9) , but for the
moment not for ITER-like parameters n0.05,
bp1, q953 (H- mode) q954-5(ITB-scenario).
Aim high current, high power, low pedestal
collisionality regimes! 3. Active control of
ELMs is progressing gt should demonstrate the
compatibility with high confinement regimes for
ITER.
33
Transport through ETB increases with
densitygtsmaller ELMs
increases with density gt if a lt
acrit., no ELMs! , (first demonstrated with
JETTO V. Parail EPS2001). But in experiment Type
Igt Type III transition with ne increase, low
confinement.
TELM M. Becoulet et al 2003
before
after
34
ITBType I ELMs ? Type III ELMs?
Usually Type I ELMs are not compatible with large
(r ITBgt 0.5) ITBs in JET, DIII-D ITB erosion by
Type I ELMs. If no pure Type II regimes gt small
Type IIIITB(improved core confinement) ?
JET M. Becoulet PPCF(2002)
Type I
Type III
wall
Te (ECE)
plasma centre
35
Edge current Ip ramps, drive?
JET Becoulet M. et al 2003

• Edge current (Ip ramp-up)
• first improve stability
• 2)then destabilise peeling modes (when kink
  unstable) Type III or dithering L-mode.
• The result is very sensitive to edge Te, ne,
  dIp/dt tres for ITER?

Ip ramp-up
55601
55599
Ip ramp-down
larger Type I ELMs!
Type I
dithering
nped(55601,55599)
Tped
(similar results MAST Gryasnevich M. et al
2002 COMPASS-D, S. Fielding EPS2001 )