Evolution of the pedestal on MAST and the implications for ELM power loadings

1
Evolution of the pedestal on MAST and the
implications for ELM power loadings Presented by
Andrew Kirk EURATOM / UKAEA Fusion Association
UKAEA authors were funded jointly by the United
Kingdom Engineering and Physical Sciences
Research Council and by the European Communities
under the contract of Association between EURATOM
and UKAEA. The views and opinions expressed
herein do not necessarily reflect those of the
European Commission.
2
Size of type I ELM Energy Loss
Understanding the size of ELM energy loss on
future devices such as ITER is essential
Most predictions are based on scaling data from
present devices
There is a large uncertainty due to our lack of
understanding of the processes involved.
3
Outline

• Extension of MAST operational space to lower
  pedestal collisionality
• The filament structures observed during type I
  ELMs and their relevance to the ELM energy loss
  process
• Working towards a model for ELM energy losses

4
Contribution from MAST
Most MAST discharges have a high pedestal
collisionality due to a low temperature pedestal
and the energy losses due to ELMs are up to 5
of the pedestal energy
MAST data
5
Contribution from MAST
A dedicated campaign using optimised fuelling and
vessel conditioning has been used to increase the
temperature pedestal and extend the MAST data set
to lower collisionalities
New data
Agrees well with the scaling observed on other
devices
6
Pedestal profiles
The lowest pedestal collisionality shot (?ped
0.08) produced to date has
7
Stability analysis
Edge stability calculations, using the 2D ELITE
code, have been performed on these profiles,
which were obtained just before an ELM
A parameter scan shows that the point lies close
to the ballooning boundary - Characteristics
associated with type I ELMs
8
Comparison of the pedestal at high and low ?
Low ? High ? (kPa/?N)
29.0 30.3
?(?N) 0.035 0.017
Similar pressure gradient but low ? case is gt
2xbroader
9
Comparison of ELM losses at high and low ?
Low ?
High ?
No ?Te
Large ?Te
But why are ELM losses greater at lower
collisionality ?
10
What happens during the ELM energy loss process
11
Spatial structure during type I ELMs
At the last IAEA it was known that filament
structures exist during ELMs
MAST
NSTX
ASDEX Upgrade
DIII-D
but what role do they play in the ELM loss
mechanism ?
12
Studies of the evolution of the filaments during
type I ELMs on MAST
MAST has installed new imaging systems and a new
edge TS system to study the evolution of the
filaments
10 kHz frame rate
13
Analysis of the images
3-D field lines are mapped onto the 2-D images
taking into account the distortion of the system.
14
Analysis of the images
3-D field lines are mapped onto the 2-D images
taking into account the distortion of the system.
15
Comparison of type I ELMs at different ?
5 ms exposure
10 ms exposure
Very little difference in the size, number or the
behaviour of the filaments
16
Evolution of the filament during an ELM
17
Reduced views allows the evolution of the
filaments to be studied
100 kHz frame rate
18
Evolution of the filaments
The toroidal and radial location of the filaments
are tracked in consecutive frames
The filaments are observed to rotate in the
co-current direction, decelerate toroidally and
accelerate out radially The width of the
filament is constant in time
19
Evolution of the filaments
The filaments exist during the time that
particles are lost from the pedestal
20
Summary of the evolution of ELM filaments

• Filaments exist for the entire ELM loss time
• Individual filaments detach from the LCFS at
  different times
• Follow the local field line, decelerate
  toroidally and accelerate radially.
• The width perpendicular to the field line is
  constant in time

But what is the energy/particle content of the
filaments?
21
The energy/particle content of the filaments
0. Pre-ELM pedestal
22
The energy/particle content of the filaments
0. Pre-ELM pedestal 1. Formation of tail
23
The energy/particle content of the filaments
0. Pre-ELM pedestal 1. Formation of tail 2.
Radial expansion - no local loss in pedestal
density
Energy in largest filament lt 2.5 ?WELM
24
The energy/particle content of the filaments
0. Pre-ELM pedestal 1. Formation of tail 2.
Radial expansion - no local loss in pedestal
density 3. Filament appears to detach at
mid-plane (leaving a depression)
Energy in largest filament lt 2.5 ?WELM
25
The energy/particle content of the filaments
0. Pre-ELM pedestal 1. Formation of tail 2.
Radial expansion - no local loss in pedestal
density 3. Filament appears to detach at
mid-plane (leaving a depression) 4. Accelerates
radially particle content decreases
exponentially with distance
26
A possible mechanism for type I ELM energy losses
27
A possible mechanism for type I ELM energy losses

• Filaments remain close to the LCFS for 50 - 200
  ms - during this time they enhance transport into
  the SOL.
• Individual filaments detach during this period -
  slow toroidally and accelerate radially.
• At the time of detachment each filament contains
  up to 2.5 ?WELM
• As the filaments travel out radially this
  content is reduced by parallel transport to the
  targets

28
Energy lost through the filaments
Assume that during tELM the filament acts as a
conduit for losses from the pedestal region
tELM(MHD) (tA2tE)1/3 100 ms (MAST)
Total particles down the filaments
Area of filament
Number of filaments
29
Energy lost through the filaments
Assume that during tELM the filament acts as a
conduit for losses from the pedestal region
tELM(MHD) (tA2tE)1/3 100 ms (MAST)
Total particles down the filaments
ELM energy loss
sfilament, fil and pedestal T and n measured
30
Results from the simple model
MAST
Given the simplifications and approximations used
there is moderate agreement with the data
But what about other devices?
31
Results from the simple model
Extrapolation of model to JET
Assume fil10 and ?filament ? a ? ? Size
20 cm
The next step should be to use our improved
knowledge of the ELM loss process to build a
real model
32
Summary

• High temperature pedestal plasmas have been
  achieved on MAST with collisionalities one order
  of magnitude lower than in previous discharges.
• The structure and evolution of the filaments
  observed during type I ELMs is similar at both
  high and low collisionalities
• The evolution of the filaments have been tracked
  through individual ELMs showing how they remain
  attached to the LCFS, decelerate toroidally and
  accelerate radially and at the time of detachment
  each filament carries up to 2.5 of ?WELM
• Based on this evolution a mechanism for ELM
  energy loss has been proposed and a simple ELM
  energy loss model derived the next
  step is to build a real predictive model.