Overview of ELM control by low n magnetic perturbation fields on JET

1
Overview of ELM control by low n magnetic
perturbation fields on JET

• Y Liang (FZ Jülich), JET-EFDA contributors
• 13th European Fusion Theory Conference,
• Oct. 12-15, 2009 Riga Latvia

2
Contributors
Y. Liang1, H. R. Koslowski1, S. Jachmich2, A.
Alfier3, C. G. Gimblett4, E. Nardon4, P. K.
Browning5, P. Devoy5, T. Eich6, C. Giroud4, G.
Maddison4, P. T. Lang6, M. P. Gryaznevich3, D
Harting1, S. Saarelma3, Y. Sun1, R. Wenninger6,
C. Wiegmann1 , T. Zhang1 and JET-EFDA
contributors JET-EFDA, Culham Science Centre,
OX14 3DB, Abingdon, UK 1Association EURATOMFZJ,
Forschungszentrum Jülich GmbH, Institute of
Energy Research IEF-4 Plasma Physics, Partner in
the Trilateral Euregio Cluster, 52425 Jülich,
Germany 2Association EURATOM-Belgian State,
Koninklijke Militaire School - Ecole Royale
Militaire, B-1000 Brussels Belgium
3Associazione EURATOM-ENEA sulla Fusione,
Consorzio RFX Padova, Italy 4EURATOM-UKAEA
Fusion Association, Culham Science Centre, OX14
3DB, Abingdon, OXON, UK 5School of Physics and
Astronomy, University of Manchester, Manchester,
UK 6Max-Planck-Institut für Plasmaphysik,
EURATOM-Assoziation, D-85748 Garching,
Germany e-mail contact of the main author
y.liang_at_fz-juelich.de ? See the Appendix of F.
Romanelli et al., Proceedings of the 22nd IAEA
Fusion Energy Conference, Geneva, Switzerland,
2008
3
Excitation of the 2/1 tearing mode using external
DED perturbation fields
Saturated m/n 2/1 island width up to 17 of
plasma minor radius
Y Liang, et al., EPS2004, ECA Vol.28G, P-1.126
(2004)
4
Secondary structures after collapse events
occurring at the q 2 magnetic surface in the
TEXTOR tokamak
Y Liang, et al., Nucl. Fusion 47 (2007) L21L25
5
Introduction
Complete suppression of type-I ELMs in
collisionless H-mode plasmas
DIII-D (n3)
T Evans, PRL, 92, 235003 (2004) T Evans, Nature
physics, Vol. 2, 419 (2006)
In-vessel Coils
JET (n1, 2)
Control (increase of frequency and reduction of
size) of type-I ELMs in H-mode plasmas
Y Liang, PRL, 98, 265004 (2007) Y Liang, PPCF ,
49, B581 (2007)
External Coils
6
Active ELM control with n 1 magnetic
perturbation field on JET
Ip 1.8 MA Bt 2.1 T q95 4.0 dU 0.45
Heat flux onto the outer divertor
JET69557
IEFCC (kA)
Field off
off
On
nel (1020m-2)
Centre
edge
Da
Times (s)
Active ELM control (frequency/size) observed in a
wide q95 window, but no ELM suppression
Y Liang, et al, PRL, 98, 265004 (2007) Y Liang,
et al, PPCF , 49, B581 (2007) Y Liang et al, JNM,
390391, 733739 (2009)
7
Summary of previous experimental results of ELM
control on JET

• Previous experimental results from JET show that
  type-I ELMs can be actively controlled by the
  application of an n 1, 2 magnetic perturbation
  field
• ELM frequency increases by factor 4 to 5
• DW/W reduces below 2
• Reduction in ELM peak heat fluxes and carbon
  erosion
• Pump-out effects electron density
• Electron and ion temperatures increase (core and
  edge)
• Moderate reduction in thermal energy confinement
• H98(y,2) stays constant
• no drop of total thermal energy confinement in
  high beta plasma
• ELMs were successfully mitigated in low and high
  d H-modes
• and at high beta
• Wide range in q95 (3 - 5) where ELM control has
  been demonstrated

8
Outline

• Introduction
• Experiments to understand the physics of of ELM
  control with low n magnetic perturbations
• Dynamics of edge profiles with n 1 field
• Compensation of density pump-out with gas and
  pellet fuelling
• Optimisation of ergodic edge region
• Multi-resonance effect in ELM frequency vs q95
  with n 1, 2 fields
• Plasma rotation braking with n 1 field
• Summary

9
Dynamic of edge profiles with n 1 field
Ped ne (1019m-3)
JET 77329
Ped Te (keV)
Ped pe (kPa)
w/o n1 field with n1 field
time (s)

• Pedestal pressure with n 1 field applied
  recovers at same rate, but the ELM crash occurs
  earlier at lower pe,ped.
• Pedestal ne is reduced by 20 while the edge Te
  is increased. ?pe is 20 smaller.

10
Influence of magnetic perturbation on X-point
Vacuum
3-D Equilibrium calculation by HINT2 Code
JET

• Flattening of j and p at the islands leads to an
  ergodisation at the island X-points
• Strong enhancement of ergodisation at the
  X-point region due to plasma response may explain
  the density pump-out seen already at a small
  amplitude of the pertubation field

C. Wiegmann, et al, EPS2009, P1.132
11
Comparison of different methods for density
pump-out compensation (I)
w/o n1 field with n 1 field
with n 1 field and fuelling
JET 77332
Pellets
Pe (kPa)
Te (keV)
ne (1019m-3)
y
y
y

• Density pump-out effect can be compensated by
  either gas fuelling or pellet injection

12
Comparison of different methods for density
pump-out compensation (II)

• C_SFE_LT
• Ip 2.0 MA Bt 1.85 T fGWL 0.6
• Pellets
• 77332 Pellets 3.5 mm, 10 Hz
• Gas puffing
• 77335 puffing rate 12E1021el/s

IEFCC (kA)
Fuelling
Da
77332
fELM 15 ? 32 Hz
Da
77335
15 ? 35 Hz
18
22
24
20
Time (s)

• ELM control with recovery of density has been
  achieved

13
Comparison of different methods for density
pump-out compensation (III)

• However, no recovery of energy confinement has
  been observed

14
Optimisation of stochastic edge region

• ITER design criterion (M Schaffer, NF 2008)
• Chirikov parameter larger than 1 for ?pol1/2 gt
  0.925
• Derived from pulses with complete ELM suppression
  on DIII-D
• Can JET discharges be tailored to test this
  criterion?

E Nardon SFP2009 workshop
15
Optimisation of stochastic edge region (II)
Ip 0.84 MA Bt 1.0 - 1.06T
PNBI (106 W)
q95
n 2
IEFCC (kA)
nel (1019 m-2)
Center
edge
Wp (MJ)
V?(rad/s)
R3.05m
3.7m

• Clear influence of the n 2 field on plasma
  density (pump-out effect), magnetic rotation
  braking, and ELM frequency
• No complete ELM suppression was obtained by
  application of n 1 or n 2 fields with a
  Chirikov parameter larger than 1 for a Ypol1/2 gt
  0.925

Ti (keV)
R3.05m
3.7m
Te (keV)
R3.01m
3.7m
fELM 50 - 90Hz
D? (a.u.)
Y Liang, et al., ITPA PEP 2009
16
Is any evidence of Edge stochastisation (I)?
Three stages 1) Slow global braking (seems to be
mainly non-resonant) 2) Dramatic braking, due to
the 2/1 mode penetration 3) Fast spin up of the
edge rotation (R3.70m), which could be the sign
of edge stochastisation (seen e.g. on TEXTOR
K-H Finken et al., PRL 94 2005)
75339
Ip 0.8 MA, Bt 1.1 T
PNBI (107 W)
IEFCC (kA)
n 1
R3.49m 3.63m 3.70m 3.78m
Otor (104 rad/s)
Time (s)
E Nardon, et al., SFP2009
17
Is any evidence of Edge stochastisation (II)?
2/1 island
Stochastic region?

• The plasma back-transits to L-mode
• Two flattened regions of Te
• One around q2
• Possibly one at the very edge (but of course the
  main effect is the loss of H-mode)
• These two regions correspond in the ERGOS vacuum
  modelling to
• The 2/1 islands chain
• The stochastic region at the edge

E Nardon, et al., SFP2009
18
Influence of low n fields on the core MHD
L-mode
24
22
26
16
20
18
Time (s)

• Increasing of the sawtooth period has been
  observed during application of n 2 field.
• broaden of the q 1 surface
• Simultaneously, the amplitude of the n 1
  precursor mode decreases.

19
Resonance effect in ELM frequency vs q95
JET 76962,76963

• ELM control with n 1 field is very sensitive
  to the edge safety factor.
• Small change of q95 from 4.5 to 4.8 results in an
  increase of fELM by a factor of 2-3 and a drop of
  nel by 15 while almost no difference is observed
  without n 1 field.
• Plasma rotation braking from the n 1 field
  does not depend on q95.

PNBI (106 W)
q95
IEFCC (kA)
Wp (105 J)
nel (1019 m-3)
no change
Vf (105 rad/s)
fELM 20 ? 40Hz
q95 4.5
Da
Da
fELM 20 ? 90Hz
q95 4.8
20
21
22
23
Time (s)
Y Liang, et al., EPS2009
20
Multi-resonance effect with n 1 and 2 fields
n 1
n 2
w/o n 1 and 2 fields

• Multiple resonances in fELM vs q95 have been
  observed with n 1 and 2 fields
• Possible explanation in terms of ideal peeling
  mode model by Gimblett et al C G Gimblett et
  al., PRL, 96, 035006-1-4(2006) currently being
  investigated

21
A possible explanation of observed resonances

• The model assumes that an unstable ideal
  external peeling mode triggers the ELM
• C G Gimblett et al., PRL, 96, 035006-1-4(2006).
  
• The edge plasma relaxes, and dE is the radial
  extent of relaxed edge plasma required to
  stabilise all peeling modes.
• Assume that the relaxed state diffuses in a
  classical manner back to the initial state, so
  fELM 1/dE2.

Most unstable toroidal mode number, n
n 1 removal
qa
ELM frequency, fELM1/dE2 (a.u.)

• On applying the external nnC field, a
  neighbouring equilibrium forms and the nnC
  peeling mode is no longer active, thus removing
  it as an ELM trigger.

• The peeling ELM model can qualitatively predict
  the multi-resonance effect with low n
  perturbation field

22
Non-resonant magnetic braking
Y. Liang et al, 22nd IAEA, 2008, EX/4-2 Y. Sun et
al, EPS, 2009, O2.019

• Strong global magnetic braking effect has been
  observed on JET
• The obtained TEFCC has a global profile. It is
  about half of the NBI torque. The maximum torque
  is in the plasma central region.
• No dependence on q95 (at fixed Br/Bt) has been
  observed in the range q95 3 - 5
• Similar plasma braking observed with n 1 and n
  2 fields

23
Comparison with Neoclassical Toroidal Viscosity
(NTV) Theory (I)
Displacement contribution
Eulerian component
Y. Sun, et al, to be submitted PPCF, 2009
NTV torque due to magnetic surface displacement
is dominant.
24
Comparison with Neoclassical Toroidal Viscosity
(NTV) Theory (II)
NTV torque from the n1 field on JET
With the assumption that all the resonant
components are screened
With vacuum assumption
Y. Sun et al, to be submitted PPCF, 2009
The NTV torque from the boundary layer
contribution in n regime agrees well with the
observed torque, if the displacement contribution
is included.
25
Summary

• Recently, ELM control experiments have been
  performed on JET aiming at a better understanding
  of the plasma response on the magnetic
  perturbation.
• With n 1 field, pedestal pressure recovers at
  same rate with respect to the case w/o n 1
  field, but the ELM crash occurs earlier.
• Compensation of the density pump-out effect has
  been achieved with either gas fuelling or pellets
  injection. However, no recovery of energy
  confinement has been observed.
• No complete ELM suppression was obtained by
  application of n 1 or n 2 fields with a
  Chirikov parameter larger than 1 at Ypol1/2 gt
  0.925 which is one of the important criterions
  for the design of ITER ELM suppression coil.
• Multi-resonance effect on q95 in ELM control
  with a low n (1 and 2) magnetic perturbation
  field has been observed.
• NTV torque from the boundary layer contribution
  in n regime agrees well with the observed torque,
  if the displacement contribution is included.