Review of ITER Priorities on Edge/Pedestal Physics Area Alberto Loarte ITER Fusion Science and Technology Department

1
Review of ITER Priorities on Edge/Pedestal
Physics AreaAlberto LoarteITER Fusion
Science and Technology Department
with contributions from ITER, Domestic
Agencies, ITPA Pedestal Group and many others
2
Outline of talk

• Introduction Operation in first 10 years key
  features of Q 10 scenario
• 2. High priority ITER RD issues on transport
  barriers and control
• Q 10 scenario
• Access to and exit from H-mode
• Access to H98 1
• Pedestal parameters and gradients
• TF ripple effects on pedestal and required
  ripple correction
• ELM control compatible with Q 10 scenario
  requirements
• H-mode in ramp-up/down phases
• Development from non-active (H/He) plasmas to Q
  10
• Access to H-mode and H98 1 in H/He (and for Bf
  ? 2.65 5.3T)
• Pedestal plasma, ELMs and ELM control in H/He
  vs. DT
• 3. Conclusions

3
Access to and exit from H-mode L-H Threshold

• Significant uncertainties remain regarding PL-H
  in ITER

• Uncertainties ? natural variability well
  known effects not included in scaling (Zx,
  input torque-plasma rotation, Prad, ) with
  major implications for ITER
• ITER H-mode threshold minimum remains uncertain
  Inner Wall with 4 MWm-2 in shine through areas
  allow full PDTNNBI for ltnegt 3.0 1019 m-3
• Progress in this field for ITER requires
• Physics model of H-mode transition and
  experimental validation
• Quantification of well known effects and
  inclusion in scaling
• Development of ITER-applicable strategies for
  H-mode threshold minimisation

4
Access to and exit from H-mode H-mode to Q 10
(I)

• Most efforts on H-mode access focused so far on
  PL-H on assumption that if H-mode achieved in DT
  ? Pa increase will take plasma to Q 10
• Pa depends on ltngt and ltTgt ? Ploss evolution
  after H-mode tied to plasma changes after H-mode
• Evolution of ltngt after L-H transition plays a
  major role on Pa evolution and scenario
  viability
• ltngt 4 10 1019 m-3 in 30 s versus 4 s

Ip 15 MA - DINA ITER V. Lukash Y. Gribov
5
Access to and exit from H-mode H-mode to Q 10
(II)

• Modelling for ITER L-H transition Pa evolution
  over simplistic ?
• detailed experimental measurements (with/without
  core fuelling)
• modelling of pedestal build-up and plasma
  evolution
• Depending on plasma evolution after L-H (ne) ?
  access to H-mode before current flat top may be
  a more controllable way to Q 10 from L-mode

JET - Sartori PPCF 2004
Fully developed H-mode
ne (1019m-3)
H-mode
L-mode
6
Access to and exit from H-mode exit from Q 10
(III)

• Exit from H-mode (if uncontrolled) presents
  plasma position control problems and large power
  fluxes on inner-wall (contact duration depends on
  CS current saturation)
• Critical issue is Wp evolution at and after H-L
  transition ? Pa and ltngt evolution
• Physics understanding and modelling required to
  determine worse/routine case ITER scenario and
  mitigation strategy

7
Access to H98 1

• ITER operation in H-mode with PLoss just above
  PL-H may be complex
• Cyclic transitions between Type I and Type III
  ELMy H-mode or even L-mode ? Wplasma oscillations
  gt 20
• PaWplasma2 ?Pa oscillationsgt 40 ? amplification
  of Wplasma oscillations
• Understanding of processes leading to low
  confinement phases large ELM dynamics at low
  PL-H (probably not relevant for ITER),
  pedestal/confinement physics at low Pinp/PL-H,
  etc.

JET-Horton NF 1999
JET-Sartori PPCF 2004
8
Pedestal parameters and gradients edge pedestal
density

• If edge particle transport described by simple
  diffusion ? ITER edge plasmas are unlike most in
  present devices
• edge plasma dense and hot ? recycled gas ionised
  far in SOL most of fuel gas ionised in far SOL
  and does not reach separatrix

• If simple picture above applicable to ITER and
  present experiments
• importance of ionisation length on pedestal
  (not clearly seen exp.)
• ITER pedestal (pellet fuelled) unlike present
  gas fuelled experiments ? pedestal experiments
  with pellets ?
• edge transport ? diffusive (i.e. significant
  particle pinch in SOL or pedestal region) ?
  simple picture above not valid ? ITER density
  pedestal similar to present experiments ?

ITER-B2-Eirene Kukushkin
9
TF ripple effects on pedestal plasma and required
correction (I)

• Experiments at JET with variable TF ripple have
  shown that ripple has a negative influence on
  pedestal plasma and energy/particle confinement
  beyond fast particle losses ?physics of effect on
  thermal plasma not understood

Low n - nped ? H98 reduced 20 (dBT 1) High n
- nped ? H98 constant with dBT
JET-Saibene- IAEA 2008
nped threshold 0.3
JET-Saibene- EPS 2009
10
TF ripple effects on pedestal plasma and required
correction (II)

• Natural 18 coil TF ripple in ITER at
  separatrix midplane 1.2 ? reduction by
  insertion of ferromagnetic inserts between
  inner/outer vessel shells
• Decrease of ripple magnitude possible to 0.33
  (5.3T) and 0.52 (2.65T) but correction is not
  toroidally symmetric due to NBI ports (0.65.3T)

FI Regular Sector
FI Regular Sector
No FI
11
TF ripple effects on pedestal plasma and required
correction (III)

• Detailed choice for correction still open 0.6
  symmetric correction or 0.3 in 5/6 of torus and
  0.6 in 1/6 of Torus (NNBI ports)
• Physics basis and/or experiments required to
  guide this choice
• which is right value of ripple outer
  separatrix at midplane, average over outer cross
  section, ?
• Is n dependence found at JET of general
  application ? experiments in smaller tokamaks to
  confirm JET JT-60U behaviour (i.e. by radial
  plasma shifts)
• experiments to create local variations of TF
  ripple (TBM-like in DIII-D)

JT-60U Urano NF 2007
ITER
12
ELM control compatible with Q 10 scenario
requirements (I)

• ELM control is required for ITER operation ? ELM
  erosion limits divertor lifetime ELM caused
  impurity influxes can cause H-L transitions
  and/or radiative collapses/disruptions

• Acceptable ELMs (Dt 500 ms)
• DWELM/Adiv lt 0.5 MJm-2
• Ip 15 MA ? DWELM 0.7MJ
• vs. DWELM gt 20 MJ expected
• Ip 7.5 MA? DWELM 2.0MJ
• vs. DWELM gt 5 MJ expected

• ELM control scheme needs to be compatible with
  requirements for Q 10
• Small or no effect on L-H threshold (or
  application after L-H transition)
• H98 1 , ltnegt/nGW 0.9 with Ploss PL-H
• qdiv lt 10 MWm-2 ltnHegt/ltnegt lt 5

13
ELM control compatible with Q 10 scenario
requirements (II)

• ELM control/suppression with internal coils in
  ITER follows DIII-D results ? field aligned n
  3/4 perturbation with sChir gt 1 in

DIII-D Fenstermacher PoP 2008

• Physics basis for suppression criteria
  evaluation of effects on plasma of resonant and
  non-resonant perturbations required

14
ELM control compatible with Q 10 scenario
requirements (III)

• ELM suppression accompanied by density reduction
  (pump-out) ? compatibility issues with Q 10 in
  ITER

DIII-D Evans NF 2008 IAEA 2008

• Understanding of physics avoidance of density
  pump-out
• Recovery of bulk density required for Q 10 by
  shallow pellet injection (avoiding ELMs)
• Increase of edge density required for
  maintaining high Praddiv (qdiv10 MWm-2)
  ltnHegt/ltnegt lt 5 without triggering ELMs

15
ELM control compatible with Q 10 scenario
requirements (IV)

• Application of ELM control coils leads to
  toroidally asymmetric particle/power fluxes at
  divertor target (DIII-D) similar to field line
  connection length pattern

• Particle fluxes in non-toroidally symmetric
  structures of similar magnitude to separatrix ?
  Excessive net erosion away from separatrix ?
• Rotation of perturbation ?
• Similar effects on power fluxes but magnitude
  depends strongly on ltnegt (or n) ?
  Expectations for ITER ?
• Consequences for radiative divertor operation
  required for Q 10 ?

DIII-D O. Schmitz PPCF 2008
DIII-D M. Jakubowski NF 2009
16
ELM control compatible with Q 10 scenario
requirements (V)

• ELM control with pellet pacing represents a
  significant extrapolation from present
  experimental results

• PFC lifetime requirements ? DWELM lt 1 MJ ? fELM gt
  20-40 Hz
• ITER Natural fELM 1-2 Hz
• tpellet tE/(50-100) in ITER
• Experiments with fuelling pellets ? fELM
  increase by 3-4 tpellet tE/(3-7)
• Planned experiments in DIII-D, JET AUG aim to
  get fELM increase by 5-10 tpellet
  tE/(10-30) with negligible change in ltnegt

ASDEX-Upgrade P. Lang NF04
L. Baylor EPS08
17
ELM control compatible with Q 10 scenario
requirements (VI)

• Key Physics issues
• ELM triggering Pellet size velocity required
  to trigger ELMs in ITER

K. Gal IAEA08
L. Baylor EPS08 G. Kocsis NF 07
10 mm3 6 1020 D
ITER pacing requirements based on conservative
assumption of pellet penetration to pedestal to
trigger ELM ? are they valid for pellet pacing
pellets ?
18
ELM control compatible with Q 10 scenario
requirements (VII)

• Key Q 10 compatibility issues
• Plasma fuelling of pellets optimised for ELM
  triggering ltnegt 0.9 ltnGWgt
• Effect on ltPpedgt H98
• Power losses by partially thermalised pellet
  mass ejected by ELMs
• direct extrapolation of (fuelling) pellet
  experiments yields huge convective power losses
  (10s MW)
• model estimates show that pellet particle
  thermalisation is incomplete associated power
  losses are very small (L. Baylor EPS 08)

19
H-mode access in ramp-up/down

• Minimisation of resistive flux consumption and
  control of vertical stability ? H-mode
  operation before and after Ip 15 MA flat top
• If ltnegt changes at L-H and H-L transition
  difficult to control ( Pa) ? access to /exit
  from H-mode at Ip lt 15 MA may be required for Q
  10

A. Kavin
Ip 15 MA
Ip 10 MA

• H-mode access/exit in Ip ramped phases
• Changes in jedge known to affect H-mode plasmas
  ? consequences for ITER scenarios ?
• ELM control during phases with varying Ip, q95
  and plasma shape ?

20
Access to H-mode and H98 1 in H/He (and for Bf
? 2.65 5.3T) (I)

• Operation in non-active phase will provide first
  experimental evidence of H-mode plasmas in ITER
  
• First ITER basis to assess plasma performance in
  DT
• Important plan DT operations and (if needed)
  upgrades
• Essential to develop ELM (and disruption)
  mitigation schemes with a CFC divertor before
  changing to W for DT
• Assumptions for H/He operations
• H-mode access ltnegt 0.45 nGW
• PLHH 2 PLHD PLHHe 1.0-1.5 PLHD (JET AUG)
• Stationary H-mode ltnegt 0.9 nGW Ploss
  (1.0-1.3) PLHM-07(H/He)
• High priorities for ITER
• Check that assumptions for H/He operations are
  valid (or invalid)
• Determine physics basis to relate H/He with D
  (DT) H-modes
• Demonstrate ELM control with same tools as
  foreseen for DT (RMP coils and H pellet pacing
  (transient))

21
Access to H-mode and H98 1 in H/He (and for Bf
? 2.65 5.3T) (II)

• Relation between He and D(H) power thresholds
• do they scale with usual parameters (ltnegt, Bt) in
  the same way (i.e. are they just the same or
  proportional) ?
• is low density threshold limit the same or
  proportional ?
• Relation between He and D(H) Type I ELMy H-modes
• are the power margin ratios to L-H the same for
  He and D(H) ?
• are pedestal gradients described by
  peeling-balooning model ?
• pedestal widths and scaling in He ?
• does fuelling of He H-mode plasmas match
  expectations from lHeo/lDo ?
• is Type I ELM behaviour similar (bulk plasma
  effects and edge fluxes) ?
• are ELM control techniques developed for D valid
  for He ?
• If PLH(He) is on lower range ? Ipgt 7.5 MA/q95 3
  or Ip 7.5 MA/q95 gt 3 viable
• How much off-axis heating is compatible with He
  H-mode access ?
• How much off-axis heating is compatible with
  Type I ELMy H-mode operation ?

22
Conclusions

• A number of key H-mode-related issues remain to
  be understood to with regards to Q 10 operation
  beyond plasma performance, impurity transport,
  etc., in stationary flat-top conditions
• Access to and exit from H-mode
• Access to H98 1
• Pedestal parameters and gradients in edge
  plasmas thick to neutrals
• TF ripple effects on pedestal and required
  ripple correction
• ELM control compatible with Q 10 scenario
  requirements
• H-mode in ramp-up/down phases
• Access to H-mode and H98 1 in H/He (and for Bf
  ? 2.65 5.3T)
• Pedestal plasma, ELMs and ELM control in H/He
  vs. DT
• Progress in these areas is required for
  finalizing some ITER hardware choices and/or plan
  the operation and/or plan for possible upgrades
• Basic physics understanding
• Experiments (multi-machine when appropriate) to
  characterise phenomena
• Development of validated models that can be used
  for ITER prediction

23
ITER Construction Plan

• First technical plasma in 2018 following the
  completion of tokamak core and key systems
• Delayed/phased installation of in-vessel PFCs,
  HCD and diagnostics after core tokamak
  components have been tested

24
ITER Construction Completion Operations Plan

• Two assembly phases after first technical plasma
  interleaved with plasma operation
• Assembly Phase 2 In-vessel components and part
  of HCD and diagnostics
• Assembly Phase 3 Remaining HCD and diagnostics
• Pre-nuclear shutdown before DD operation
  (installation of full W divertor)

Assembly Phase 1
Assembly Phase 1
Assembly Phase 2
Assembly Phase 3
Assembly Phase 2
Assembly Phase 3
25
H98 1 access (IIIb)

• RD is required to determine Ploss for H 1 or
  Type I ELMy H-mode operation and its scaling
  with device size
• Effect of dW/dt vs machine size in changing
  Ploss/Pinput near Type I/III transition ?
• Unfavourable DTELM size scaling with machine
  size causing Type I/III back transition ?
• Control of DTELM at Ploss/Pinput may reduce
  required Ploss for Type I

JET- PPCF04 - R. Sartori
Te,ped (eV)
26
Access to and exit from H-mode L-H Threshold
(II)

• Characterization and quantification of well
  known effects for ITER

X-point height
Input torque plasma rotation
JET-Andrew PPCF04
Zx/a (JET/ITER) lt 0.3/0.4
Multi-machine experiments and modelling required
to determine physics mechanisms and relevance to
ITER H-mode access
27
Access to H98 1 (I) (more in R. Sartoris
presentation)

• Stationary H 1 may require up to Pinput gt 1.3
  PL-H for ITER Q 10

ASDEX-Upgrade-Ryter-H-mode WS2007
JET-Saibene PPCF 2002
ITER QDT 10, 500 MW ? Padd 50 MW, Pa 100
MW, Pradcore 50 MW Psep 100 MW
28
Access to H98 1 Pfusion Q (III)

• Higher Pinput than foreseen for H98 1 access not
  a major threat to the achievement of ITER high Q
  inductive goal but requires upgrades and may
  impose limitations on scenario
• Heating upgrade (Pinput 100 MW Pinstalled
  130 MW)
• High Ploss ? larger Praddiv to keep qdiv lt 10
  MWm-2
• High Pfusion ? larger loads on in-vessel and
  coils (cooling upgrade and/or pulse length
  limitation)
• Q 7 - 8.5 in inductive scenario ? Pa/Paux
  1.4 -1.7

95 confidence interval
29
Pedestal parameters and gradients fusion
performance status (I)

• Importance of achieving a sufficiently high Pped
  for high Q in ITER identified
• Model for pedestal width developed and compared
  with experimental measurements (more in Snyders
  presentation)
• ? nped
  7 1019 m-3 Tped 4.6 keV ? Q 10

PIPB - NF 2007
Snyder - NF 2009
30
Pedestal parameters and gradients fusion
performance status (II)

• Physics mechanism determining pedestal widths
  remain not well understood ? experiments in
  single device and multi-machine comparisons
• DTe (DTi) consistent with dependence on
  transport ( bped)
• Dne physics much less clear (sources
  transport)

JT-60U-Urano NF 2008
JET-DIII-D- Beurskens EPS 2009
Physics for Dne crucial to predict ITER Q 10
behaviour
31
TF ripple effects on pedestal plasma and required
correction (II)

• Experiments at JT-60U have decreased ripple by
  ferromagnetic inserts
• Reduction of TF ripple changes edge rotation and
  increases Pped
• Even for same VTped ? Pped increases by 20 for
  corrected TF ripple

JT-60U Urano NF 2007