Edge Localised Modes control

1
Edge Localised Modes control by resonant
magnetic perturbations
E. Nardon EURATOM / UKAEA Fusion Association

• Many thanks to M. Bécoulet, P. Cahyna, T.E.
  Evans,
• Kirk, R. Hawryluk, S. Saarelma, S. Sabbagh,
• O. Schmitz and all the community of ELM
  controllers

2
Outline

• Introduction
• Experimental results
• Modelling
• A proposed global picture
• Status of the ITER coils
• Upcoming experiments
• Summary and outlook

3
Introduction

• ITER ELMs are estimated to release 20MJ
• but PFCs can only tolerate 1MJ!
• Resonant Magnetic Perturbations (RMPs) have been
  
• shown to mitigate the ELMs on different machines
• ELM control coils under study for ITER
• Other ELM control techniques exist (ELM pace
  making)
• Original idea stochastize the edge to degrade
  dTe/dr in the transport barrier in order to keep
  the plasma stable
• The reality does not entirely conform to this
  picture
• To date, ELM control by RMPs is not fully
  understood

4
Outline

• Introduction
• Experimental results
• Modelling
• A proposed global picture
• Status of the ITER coils
• Upcoming experiments
• Summary and outlook

5
ELM suppression on DIII-D
Evans 06
I-coils 66 coils, n3, 4kAt

• ELM suppression for tens of tE
• Works only in a narrow resonant q95 window
• Drop in density  pump-out 
• dTe/drETB increases rather than drops!

ne
Te
6
EFCCs n 1 or 2, 20kAt
ELM mitigation on JET
Liang 07
n1 experiment
Pump-out (DIII-D)
No drop in dTe/drETB (DIII-D)
fELM ?, ?WELM ? (but not enough for ITER)
7
Different effects observed on NSTX
n 2 / 3 / (23)

• Destabilisation of Type I ELMs in otherwise
  ELM-free plasmas
• Decrease in fELM in ELMing plasmas
• No density pump-out, possibly due to absence of
  pumping

8
MAST EFCCs have a possible effect on ELMs
EFCCs n 1 or 2 12kAt

• n1 perturbations
• Affected L-H transition
• Often caused H-L back transition
• n2 worked better
• Possible effect on ELMs
• Increase in Type IV
• ELM frequency in low
• collisionality discharges
• (Difficulties to obtain
• a good ref. discharge
• with Type I ELMs)

n2 experiment
fELM
EFCC current
9

• Other machines
• JFT-2M triggering of ELMs in ELM-free periods
  Shoji 90
• COMPASS triggering of ELMs in ELM-free plasmas
  / increase in ELM frequency in ELMing plasmas
  Fielding 01
• TEXTOR reduction in ELM size with the DED
  Unterberg 08
• In summary
• Complex phenomenology
• We will focus on the DIII-D and JET results

10
Outline

• Introduction
• Experimental results
• Modelling
• A proposed global picture
• Status of the ITER coils
• Upcoming experiments
• Summary and outlook

11
Magnetic modelling vacuum approximation
ERGOS modelling (Nardon 07) for DIII-D I-coils
12
ELM mitigation/suppression is correlated with
stochastisation in vacuum field modelling (1/2)

• DIII-D (Fenstermacher 08)
• Width of stochastic layer (?Chirgt1) good
  ordering parameter
• for ELM size

ELM size
?Chirgt1

• Critical width for ELM suppression 3 pedestal
  widths

13
ELM mitigation/suppression is correlated with
stochastisation in vacuum field modelling (2/2)

• JET (Liang 07) the ELM mitigation threshold
  depends on q95
• Consistent with a dependence on sChirikov
  (Nardon 08)

• (?Chirgt1 at JET mitigation) lt (?Chirgt1 at DIII-D
  suppression)
• Remark ELM mitigation also seen at DIII-D
• before ELM suppression

14
Evidence for stochastisation strike point
splitting
IR and Da view
DIII-D results (Schmitz 08)
swall cm

• Vacuum modelling of field lines
• impacts on divertor targets

f degrees

• Consistent structure seen on Da

• Less marked on heat flux

15
Ideal MHD stability analysis

• In DIII-D, RMPs maintain the profiles in the
  stable region for peeling-ballooning modes
• The peak value of dp/dr is not affected much
• The region most affected by the RMPs is actually
  inside the pedestal

0kA (ELMing)
4kA (ELMing)
6.3kA (ELM free)
Evans 08
Snyder 07
16
Ideal MHD stability analysis

• In JET, the plasma remains unstable but moves
  towards
• the peeling boundary (Saarelma 08)
• Mode structure less extended radially,
  consistent with smaller ELM size

During EFCCs pulse
Before EFCCs pulse
(jedge,maxjsep)/2 (MA.m-2)
?
?
Toroidal mode number of most unstable mode
17
Density transport modelling pump-out?

• Pump-out increased transport pumping
• Pumping
• Active pump
• Changes in recycling conditions (strike point
  splitting)
• Increased transport
• // transport driven by dp/dr (Tokar 07)
• Stronger transport than // one related to
  complex role of Er
• (Tokar 08)
• Neoclassical prediction
• Stochastic field ? positive Er holding back the
  electrons

See poster 30 by M. Tokar
See poster 37 by A. Kramer-Flecken
18
Heat transport why does dTe/drETB survive?

• Parallel conduction in vacuum field ? large
  transport

Joseph 07
unless one uses a strong flux limiter Tokar
07, Bécoulet 05 ? With screened RMPs?
See poster 30 by M. Tokar
Numerical factor
19
Magnetic modelling plasma response

• RMPs screening by rotation cylindrical
  modelling
• Fluid (Fitzpatrick theory) Bécoulet 08
• Kinetic Heyn 08
• ? Weak screening at the very edge
• ? Strong screening in the core
• 3D MHD equilibrium realistic geometry
• IPEC Park 07
• ? Plasma response can be non negligible
• even without rotation
•  All 
• JOREK (3D non-linear MHD) Nardon 07
• A large resistivity is needed in these
  simulations
• ? Screening predictions only qualitative

20
Magnetic modelling plasma response

• Key physics of RMPs screening by rotation

Induction equation
Origin of screening
Typical vExB and ve H-mode profiles
vA
vExB and ve add up ? expect little reconnection
Small vExB and ve and large ? ? expect
reconnection
ve
vExBve
vExB
21
Outline

• Introduction
• Experimental results
• Modelling
• A proposed global picture
• Status of the ITER coils
• Upcoming experiments
• Summary and outlook

22
A proposed global picture
Instead of one large stochastic layer
23
A proposed global picture
Two stochastic layers isolated by good flux
surfaces
24
A proposed global picture
Two stochastic layers isolated by good flux
surfaces

• First stochastic layer at the very edge
• Easily created
• Island overlap does not require large RMPs
  because of strong magnetic shear (in vacuum
  approximation)
• Vacuum approximation is OK
• Large ?
• Small rotation
• Responsible for
• Strike point splitting
• Density pump-out
• Remark in DIII-D, pump-out is observed in a
  larger q95 window than for ELM suppression in
  DIII-D
• ELM mitigation in JET (and DIII-D)

25
A proposed global picture
Two stochastic layers isolated by good flux
surfaces

• Good flux surfaces in the middle of the pedestal
• Due to strong rotational screening of RMPs
• Small ?
• Strong vExBve
• Helps maintain dTe/dr
• Cuts // connection between core plasma and
  escaping field lines
• ? Strike point splitting seen more clearly on
  recycling than on heat flux

26
A proposed global picture
Two stochastic layers isolated by good flux
surfaces

• Second stochastic layer from top
• of pedestal
• Although ? is small, reconnection occurs because
  ve compensates vExB
• Stochasticity is more difficult to obtain
• Magnetic shear is smaller than at the very edge
• ? Islands overlap requires larger RMPs
• Possibly not reached at JET
• Responsible for ELM suppression on DIII-D
• Consistent with correlation ELM suppression ?
  stochasticity
• Consistent with the fact that RMPs affect the
  pressure gradient profile mainly inside the
  pedestal

27
Outline

• Introduction
• Experimental results
• Modelling
• A proposed global picture
• Status of the ITER coils
• Upcoming experiments
• Summary and outlook

28
Status of coils design on ITER

• Main physics requirements
• sChirikovgt1 for ?N1/2gt0.92 for all scenarios
• Minimise braking due to Neoclassical
• Toroidal Viscosity (NTV)
• NTV arises from broken axisymmetry
• (Shaing)
• NTV produces a global braking
• NTV is strong at low collisionality
• Concern for ITER
• NTV theory needs to be compared
• to experiments
• Only NSTX so far (Zhu 06)

• Current concept
• 3 rows of 9 coils between blanket and VV
• n4
• Coupled with VS coils
• Spectral flexibility

29
Outline

• Introduction
• Experimental results
• Modelling
• A proposed global picture
• Status of the ITER coils
• Upcoming experiments
• Summary and outlook

30
Upcoming experiments a very active program
COMPASS (Cahyna 08)

• Ongoing experiments on already involved machines
• More machines getting involved / new coils
  systems
• COMPASS restarting in IPP Prague
• ELM coils to be implemented on ASDEX-U
• Studies for possible new ELM coils on DIII-D,
• JET, NSTX
• MAST has new ELM control coils
• First results in L-mode
• First H-mode experiments next week

31
Summary and outlook

• ELM control coils could be an essential element
  in ITER
• ELM control is accompanied by density pump-out,
  but dTe/dr survives
• Is it possible to compensate the pump-out?
• Compatibility with pellet fuelling?
• Unified criterion / recipe for ELM control?
• Requires to understand the physics
• Stochasticity seems to play a key role
• Linear MHD stability analysis suggests that ELM
  control results from changes in profiles
• RMP-induced transport is a difficult topic
• Plasma magnetic response being investigated
• Likely to be a screening in the middle of the
  pedestal
• Suggestion two stochastic layers isolated by
  good flux surfaces
• Plasma braking due to NTV in ITER is an open
  question
• Active upcoming experimental program including
  new contributors

32
Back-up slides
33
Screening currents do not change magnetic
footprint envelope
JET modelling (EFCCs n1)
Vacuum field
Nardon 08
With helical currents put by hand on q5 to
screen the local RMPs
34
ELM mitigation/suppression is correlated with
stochastisation in vacuum field modelling

• DIII-D (Fenstermacher 08) the width of the
  resonant q95 window depends on the RMP mix

I 4kAt C 9.6kAt
I 3kAt C 9.6kAt
I-coils n3 C-coils n1
I 3kAt C 4.8kAt
35
References
Experimental results COMPASS S.J. Fielding et
al., 28th EPS conference (Funchal, 2001), ECA
vol. 25A, p. 1825 JFT-2M T. Shoji et al., 17th
EPS conference (Amsterdam, 1990), ECA vol. 14B,
p. 1452 TEXTOR B. Unterberg et al., submitted
to J. Nucl. Mater. DIII-D T.E. Evans et al.,
Nucl. Fusion 48 (2008) 024002
Vacuum modelling and design of ITER coils E.
Nardon et al., J. Nucl. Mater. 363-365 (2007)
1071 M. Bécoulet et al., Nucl. Fusion 48 (2008)
024003 M. Schaffer et al., Nucl. Fusion 48 (2008)
024004
36
Evidence for stochastisation DIII-D M.E.
Fenstermacher et al., Phys. Plasmas 15, 056122
(2008) JET E. Nardon et al., submitted to J.
Nucl. Mater. Strike point splitting at DIII-D
O. Schmitz et al., submitted to Plasma Phys.
Control. Fusion
MHD stability analysis DIII-D P.B. Snyder et
al., Nucl. Fusion 47 (2007) 961 JET S. Saarelma
et al., submitted to Plasma Phys. Control. Fusion
Plasma response E. Nardon et al., Phys. Plasmas
14 (2007) 092501 M. Bécoulet et al., Nucl. Fusion
48 (2008) 024003 M.F. Heyn et al., Nucl. Fusion
48 (2008) 024005
37
Transport modelling M.Z. Tokar et al., Phys.
Rev. Lett. 98 (2007) 095001 M. Bécoulet et al.,
Nucl. Fusion 45 (2005) 1284 I. Joseph et al.,
Nucl. Fusion 48 (2007) 045009 M.Z. Tokar et al.,
Phys. Plasmas 15 (2008) 072515 NTV W. Zhu et
al., Phys. Rev. Lett. 96 (2006) 225002 A. Cole et
al., Phys. Plasmas 15 (2008) 056102 and refs.
therein