Folie 1

1
Kinetic study of parallel transport in the ITER
Scrape-off Layer D. Tskhakaya1, R.A. Pitts2, A.
Loarte3 and S. Kuhn1 1Association EURATOM-ÖAW,
University of Innsbruck, A-6020 Innsbruck,
Austria 2CRPP, Association EURATOM
Conféderation Suisse, 1015 Lausanne, Switzerland
4EFDA Close Support Unit Garching, D-85748
Garching, Germany Permanent address
Andronikashvili Institute of Physics, 380077
Tbilisi, Georgia
D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008
2
OUTLINE

• Introduction
• Description of the model
• Checking of the model on existing tokamaks
• Simulations of the ITER SOL
• conclusions

D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008
3
Introduction
Type I ELM-ing H-mode is baseline ITER scenario
JET 62218

• t 19.05 s,
• before-ELM

Edge Localized Modes (ELMs) are periodic bursts
of particles and energy into the SOL
ELMs curry 10 of the pedestal energy and at
least 60 of it is deposited to the divertor
plates eroding the plate surface
Damage caused by ELMs (Loarte TF-S1, 2006)
R.A. Pitts, EPS 2005
D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008
4
Introduction

• ELMs can reduce the lifetime of the ITER divertor
  (due to melting, ablation, erosion) to
  unacceptably low level 3x103 full power shots
  106 ELMs
• Maximum allowed ELM energy WELM
• 10 -14 MJ, Federici, PPCF 2003
• 1 MJ, Loarte, IAEA 2008
• Maximum WELM depends on the
• history of the power deposition to
• the divertors
• Estimation of ELM power loads to
• the ITER divertor has become a
• critical issue

Precise kinetic modelling of the ELM transport in
the ITER SOL is required
D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008
5
BIT1 1.5D PIC/MC code (serial, parallel
versions)

• Plasma Arbitrary number of ion species
• 1D3V, full resolution of motion,
• nonlinear collision operator for arbitrary
  number of ion species
• Linear PSl model (i.e. recycling coefficients
  R(E, j), g(E,j) are prescribed)
• Neutrals Number of neutral species is limited
  by
• 2d3V, full resolution of motion
• nonlinear collision operator for elastic and
  inelastic collisions
• Linear PSl model
• Impurity nonlinear, or linear model with
  prescribed n, T profiles
• The electric field is calculated
  self-consistently, the magnetic field is fixed

Simulation geometry
Separatrix
Divertor
D
C
D2
C
D.Tskhakaya, 13th EU-US TTF, Copenhagen,
03.09.2008
6
The mechanism of ELM parallel transport
Electron time scales, te L/vTe
Ion time scales (ti L/cs)
Fast ELM electrons absorbed at the divertor plate
increasing the potential drop across the sheath,
Df
Most of the ELM electrons are kept in the ELM
burst by charge separation forces and propagate
with the same speed as the ions (Cs)
Df limits any further increase of electron power
load and accelerates thermal ions
Plasma sheath
Thermal ions
Fast ELM electrons
Df Te
t
Divertor plate
D. Tskhakaya, EPS 2007
D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008
7
Description of the model
What can be obtined from a simple 1.5D model?
WIR
qdiv
tIR
t
D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008
8
Simulations for JET (i)
JET 62221, WELM 0.4 MJ TELM 1.5 keV
Measurements of the power loads to the divertor
(62221, T. Eich) and PIC simulation results (PSI
2008, D.Tskhakaya).
D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008
9
Simulations for JET (ii)
Inner/outer asymmetry of divertor power loads and
impurity radiation. Case with M0.2 (see T.
Eich, PSI 2008).
Input for the kinetic model nped, Tped, tELM, M,
nimpurity
D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008
10
Simulations for TCV
No impurity radiation has been included in this
model
Power loads to the TCV divertor from the
experiment (35036, J. Marki) and PIC
simulations. 1 corresponds to t 0.5871 s, 2
t 0.6226 s.
D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008
11
Simulations for ITER

• What is different in ITER SOL simulations
• tELM lt t different history of power
  deposition to the divertors?
• System is too large L100 m more than 8.000
  (1D) cells are required
• Two types of major ions what is the ELM parallel
  propagation speed?
• More complicated PSI processes

Main Problems
Numenical stability
PSI, atomic data
Significant disrapency in atomic data from
different sources, e.g. for e C(n) Some PSI
data is missing, e.g. D recycling channels D,
D2, DT
New PIC
Conventional PIC
Poloidal profiles of the energy flux.
D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008
12
Inter-ELM ITER SOL
Poloidal profiles of the plasma tempetarure
Poloidal profiles of the impurity (CC) density,
obtained from B2 simulations by A.S. Kukushkin.
Inner divertor
Outer divertor
Assumption impurity profile stays unchanged
during tELM, t (300 µs)
D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008
13
ELMy ITER SOL

• Simulated ELM parameters
• WELM 4 MJ
• TELM 5 keV
• tELM 200 µs

Two simulations have been made with and without
impurity radiation
Results
Power loads to the divertor are identical!
Where is the mistake? It can be shown
analytically that the energy to the divertors is
curried mainly by the particles with E gt
TELM Electrons with Egt5 keV practically do not
interact with C and C Precise modeling of the
impurities C C(iv) is required! But radiation
changes the amplitude and not the time shape of
the power loads
Simplified cross-section for eC, eC excitation
collisons H. Suno, NIFS-Data
D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008
14
ELMy ITER SOL power loads
Power loads to the ITER outer divertor
Tskhakaya, PSI 2008
Existing semi-analytic model well describes power
loads Eich/Funamenski
D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008
15
ELMy ITER SOL particle flux
Particle fluxes to the ITER outer divertor.
Analytic fit
D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008
16
Conclusions

• 1D plasma 2D neutral/impurity PIC code BIT1 has
  been updated for simulation of the plasma
  parallel transport in the ITER SOL.
• The code has been tested for existing tokamaks
  (JET, TCV). Simulated power loads to the divertor
  plates well reproduce the experimentally observed
  values.
• Power loads to the ITER divertor correspond to
  the energy propagation with Cs speed and can be
  described by analytic functions.
• Most part of the power to the divertors is
  carried by ions, WIR 0.35.
• It seems that impurity radiation has a small
  effect on peaking values of the ELM power loads
  to the ITER divertor plates.
• For estimation of the radiated power a precise
  modelling of C-C(iv) and/or seeded impurities is
  required.
• Future plans Update code to full 2D and perform
  new simulations with self-consistent impurity
  model.

D.Tskhakaya, 13th EU-US TTFW, Copenhagen,
03.09.2008