SOLPS modeling of ELMing Hmode

1
SOLPS modeling of ELMing H-mode
Thesis committee II
Candidate Barbora Gulejová Supervisor of
thesis Dr. Richard Pitts Acknowledgements
David Coster, Xavier Bonnin, Roland Behn,
Marc Beurskens, Stefan
Jachmich, Jan Horácek,
Arne Kallenbach, Janos Marki
2
OUTLINE
Research plan change of direction SOLPS 5
code package ELM simulation - theory Simulation
of Type III ELM at TCV Simulation of Type I
ELMing H-mode at JET Code - experiment
benchmark Code - code benchmark Future plans








3
RESEARCH PLAN
ELMing H-mode baseline scenario for plasma
operation on ITER! Edge
localised mode (ELM) H-mode ? Edge MHD
instabilities ?Periodic bursts of particles and
energy into the SOL ELM leaves edge pedestal
region in the form of a helical filamentary
structure localised in the outboard
midplane region of the poloidal cross-section
Danger ? divertor targets and main walls
erosion ? first wall power
deposition Energy stored in ELMs TCV ? 500 J
JET ?
200kJ
ITER? 10 MJ gt unacceptable

Understanding of the ELM from formation to point
of
interaction with plasma facing components


Important research
goal!
4
RESEARCH PLAN

AIM of thesis contribute to
understanding transport in the SOL transient
events gt ELMs interpretative modeling of both
1.) steady state and
2.) transient particle
and heat fluxes during ELMing H-mode
employing the SOLPS5 fluid/Monte Carlo code
rigorous benchmarking seeking the possible
agreement between
1.) experiment and
simulation
2.) code and different code
Initially hoped to use new unique experimental
data from TCV 1.) AXUV ?

2.) IR ?
Twin camera system Bolometry - total radiated
power ? Lyman a edge
and divertor radiation ? gt to be compared with
SOLPS Investigations gt
low peak transmission of the L? absorption
filters (10)
strong angular dependence of the emission (only
1 at incident angle 60)
strong ageing effect due to light
exposure

gt not usable for detection of hydrogenic
recycling emmision now being used for
different purposes
5
Change of thesis title from Particle sources
in the TCV tokamak edge to SOLPS5 modelling of
ELMing H-mode

1.)

• Type III Type I ELMing H-mode
• TCV JET
• benchmark SOLPS EDGE2D/NIMBUS

2.a)
2.b)
6
Scrape-Off Layer Plasma Simulation

• Suite of codes to simulate transport in edge
  plasma of tokamaks
• B2 - solves 2D multi-species fluid equations on a
  grid given from magnetic equilibrium
• EIRENE - kinetic transport code for
  neutrals based on
• Monte - Carlo algorithm
• SOLPS 5 coupled EIRENE B2.5

Mesh
plasma background gt recycling fluxes
72 grid cells poloidally along separatrix 24
cells radially
B2
EIRENE
Sources and sinks due to neutrals and molecules
Main inputs magnetic equilibrium
Psol Pheat Pradcore
upstream separatrix density ne
EELM Free parameters
cross-field transport coefficients (D-, ?-, v-)

measured
systematically adjusted
D0 D1 C0 C1 C2 C3 C4 C5 C6
7
1.)Type III ELMing H-mode on TCV
ELMs - too rapid (frequency 200 Hz) for
comparison on an individual ELM basis gt Many
similar events are coherently averaged inside
interval with reasonably periodic elms
telm 100 µs
tpre 2 ms
tpost 1 ms
Post-ELM phase
Pre-ELM phase
ELM particles and heat are thrown into SOL
( elevated cross-field transport
coefficients)
Time-dependent ELM simulation starting
from time-dependent pre-ELM steady state
simulation equal time-steps for kinetic and
fluid parts of code, dt 10-6 s
8
Pre-ELM and ELM simulation - theory
Cross-field radial transport in the main SOL -
complex phenomena
Ansatz( D-, ? -, v-) variation
radially transport barrier (TB) poloidally no
TB in div.legs
2 approaches
Pure diffusion v-0 everywhere
More appropriate Convection
simulations with D- D-class
Simulation of ELM
Instantaneous increase of the cross-field
transport parameters D-, ? -, v-!
1.) for ELM time from experiment coh.averaged
ELM tELM 10-4s 2.) at poloidal location -gt
expelled from area AELM at LFS Cross-field
radial transport gt approximate estimation of
transport parameters during ELM corresponding to
the given expelled energy WELM, tELM and AELM
AELM 1.5m2 W 600 J
20

D -
? -
40
0
20
9
Edge diagnostics to constrain the code at TCV
downstream Langmuir probes jsat target
profiles
upstream Edge Thomson scattering
ne and Te
upstream profiles
I R
laser beam
J.Horacek
RCP reciprocating probe
jsat A.m-2
outer target
R.Behn
R-Rsep m
pedestal
ne
jsat
inner target
R-Rsep m
R-Rsep m
Strategy Match these experimental profiles with
data from SOLPS simulation runs by changing
cross-field transport parameters D-,?-, v-
IR cameras Perpendicular heat flux
pedestal
Te
J.Marki
Heat flux MW.m-2
outer target
R-Rsep m
R-Rsep m
10
Simulation of pre-ELM steady state of H-mode
presented at PSI 17, Hefei,China and published
in JNM May 2006
upstream
targets
outer
inner
Jsat A.m-2 LP, average SOLPS
D-,?- radially constant in divertor legs
Te eV
Excellent match upstream and very good match
downstream ?
ne m-3
Ansatz
D m2.s-1 ? m2.s-1
r-rsep
r-rsep
Perp.heat flux MW.m-2 LP, average SOLPS
v m.s-1
6 (div. 5)
inner
same solps result!
Vperp0
Good basis for time-dependent ELM simulation
1 (div. 3)
IR
outer
NO DRIFTS yet!
r-rsep
11
ELM is more convective than conductive !
Type III TCV ELM simulation
Upstream
TS measurements (R.Behn et al., PPCF 49 (2007))
gt larger
drop in ne than Te at the pedestal top
gt
ltTepedgt?.neped exceeds ltnepedgt? .Teped
in the contribution to EELM
gt
SOLPS D increased more then ? during the
ELM 100 times 10
times (2 times in SOL)
change of radial shape !

ETB collapse!
AELM1.5m2 Gaussian function of multiplicators
polloidally 1 ELM cycle of total 400 µs, 100 µs
before ELM
ELM duration tELM 100µs
100 points
during ELM event
200 µs after ELM
smaller TB
? -
D -
Very good agreement !
12
Type III TCV ELM simulation
Downstream
SOLPS lt Exp (LP) factor 1.5
SOLPS gtgt Exp (coav LP) (
R.Pitts,Nucl.Fusion 43 (2003))
factor 3
13
Target power fluxes
Type III TCV ELM simulation
Outer target
Inner target
20
Sheat heat transmission coefficient ? 7.5
preELM ELM
preELM ELM
Time-dep. heat flux estimated from LP
measurements at outer target
9 , SOLPS gt LP factor 2
(R.Pitts, Nuc.Fusion 2003)
SOLPS P (outer target) 240 J P
(inner target) 100 J 55 of EELM (620 J)
14
Jsat profiles broadening during ELM
LP close to outer target strike point
coavelm Jsat
Profiles of target jsat during the ELM rise are
steeper
Outer target
LP
SOLPS
?

Agreement with R.Pitts, Nucl. Fusion 43 (2003)
?
?
15
Time evolution of SOLPS profiles during ELM
Little drop of Te along L 18 m (outer
target) 14 m (inner target) from upstream
to targets
Ti decrease 4.5 x ( outer target )
3 x ( inner target )
strong ion cooling
Outer target transit times ?e
L/vthe 2 µs
?i L/cs(ped) 120 µs !!
(sheat heat transmission coeff8)
Similar to 1D kinetic PIC ELM
simulations D. Tskhakaya, 34th EPS(2007)
ELM inherently kinetic event
gt Simulations with kinetic code BIT1 necessary
next step for TCV Already launched by D.
Tskhakaya for this particular TCV ELM gt will
be compared soon
Time-dependence of simulated Te,Ti for 120 kJ
ELM at JET
16
SOLPS5 simulations with DRIFTS
SOL flows
DRIFTS contribute to in/out assymetries TCV
unconventional equilibrium with an extremely
short X point to inner strike points position -gt
might dominate over drifts and divertor physics
effects
R. Pitts

• Poloidal

• SOLPS X.Bonnin (november 2006)
• implementation of drift terms
• Anomalous contribution (ExB)
• Diamagnetic contribution (?pxB)
• Viscous contribution

Number of issues uncovered ?
SOLPS internal treatment of the
direction of magnetic field components !
With D. Coster (visit in IPP Garching, Mai
2007) simulations with both Btor field direction
! Not yet at sufficiently mature stage for
conclusions More work planned in
2008 !

• Parallel

Switching on drifts its likely to decrease
the predicted Te at outer target may have only
small effect at the inner target Simulations gt
early promise, but
17
2.a)Type I ELMing H-mode on JET
baseline scenario for QDT10 burning plasma
operation on ITER !!! To avoid divertor damage
gt maximum ?WELM (ITER) 1MJ (at TCV 0.005 MJ
!!!) this is achievable on JET
70224
Parameters Bt 3 T Ip 3 MA ne6x1019
m-3 P(SOL) 19 MW No GAS PUFF !! ?ped 0.03
-0.08 (expected for ITER!)
Core LIDAR Edge LIDAR Li beam
HRTS ECE
Dalpha
PIN 20 MW PRAD (below Xpoint) PRAD (core)
ELM parameters felm 2 Hz ?WELM 1 MJ
Langmuir probes
Wdia
Bolometry (A.Huber) radiation between ELMs
Simulated for the first time ! with SOLPS5
time
18
Simulation vs. experiment - upstream
ne m-3
First results for pre-ELM
Te eV
D- 0.01 m2.s-1 in pedestal 1 m2.s-1 in
SOL ?- 0.7 m2.s-1 in pedestal 1 m2.s-1
in SOL (TCV pedestal D- 0.007 m2.s-1
?- 0.25 m2.s-1 )
Ti eV
D m2.s-1 ? m2.s-1
Core LIDAR Edge LIDAR Li beam
HRTS ECE
quite promising unlike for TCV inward
pinch radial profile necessary!!
vperp m.s-1 inward pinch!
19
Simulation vs.experiment - targets (LP)
SOLPS
Outer target
Inner target
LP
LP
80
50
50
30
1,2
0.5
r-rsep
r-rsep

r-rsep map2mid

r-rsep map2mid

Inner target good agreement
Outer target SOLPS Te too high!
20
2.b) Benchmarking code-code
SOLPS
5 B2.5
EIRENE fluid
(Braginskii)
kinetic (Monte Carlo), neutrals
vs.
EDGE2D
NIMBUS fluid (Braginskii)
kinetic (Monte
Carlo), neutrals
(less complex
then EIRENE!)
21
Type I ELMing H-mode on JET
Succesfully modelled by EDGE2D NIMBUS by Arne
Kallenbach (PPCF 46,2004)
58569
Parameters Bt 2 T Ip 2 MA ne 4x1019
m-3 P(SOL) 12 MW GAS PUFF from inner
divertor !!
Dalpha
PIN 14 MW PRAD (core)
ELM parameters felm 30 Hz ?WELM 200 kJ
Wdia
Core LIDAR Edge LIDAR Li beam
ECE
time
22
Simulation vs.experiment-upstream
EDGE2DNIMBUS
SOLPS5
2
2
Same Ansatz D,Chi,v
Same ne,Te,Ti upstream profiles
0
0.05
0.1
r-rsep m

23
Simulation vs. experiment-targets(LP)
outer target
EDGE2DNIMBUS
outer target
inner target
SOLPS5
250
300
LP
25
25
5
7
r-rsep map2mid

r-rsep map2mid

24
Future plans for 2007

• After obtaining the trully time-dependent
• pre-ELM solution ! continue in the
• attempts to simulate the small TCV ELM
• properly -use several different approaches
• Planed visit to JET in february march 2007
• simulate the big JET ELMing H-mode
• Continue in the simulation with DRIFTs
• included in SOLPS

25
Future plans for 2008

• Detailed study of drift effects on TCV Type III
• ELMing H-mode with forward and reversed B-field
• for pre-ELM and ELM simulations (also for JET
  Type I)
• Time-dependent TCV ELM simulations compared with
• results of PIC and hopefully dynamic values of
  kinetic
• heat flux limiters and sheat transmission
  coefficients
• used as input to SOLPS
• Simulation of large ELMing H-mode at TCV using
  full
• power third harmonic ECH (if retrieved)

26

• Thank you for attention !

27
First attempt of SOLPS simulation with DRIFTS
targets
upstream
outer
inner
separatrix
Jsat A.m-2 LP SOLPS
ne SOLPS TS RCP
Not yet completely converged solution
Te eV
pedestal
wall
Te SOLPS TS RCP
ne m-3
R-Rsep
Perp.heat flux MW.m-2 LP SOLPS
inner
D- ?-
6
outer
1
IR
R-Rsep
R-Rsep
R-Rsep
28
First attempt of SOLPS simulation with DRIFTS
NO DRIFTS
DRIFTS
outer
inner
outer
inner
NO DRIFTS Overestimation of outer target
Te DRIFTS Decrease of outer target Te as
expected Same effect on jsat and heat flux!
Inner target not significant effect as
expected
Jsat A.m-2 LP SOLPS
Jsat A.m-2 LP SOLPS
Te eV
Te eV
ne m-3
ne m-3
R-Rsep
Perp.heat flux MW.m-2 LP SOLPS
R-Rsep
Perp.heat flux MW.m-2 LP SOLPS
inner
inner
outer
outer
IR
IR
Good early promise !!!
R-Rsep
R-Rsep
29
Parallel Mach Flows
30
preELM time-dependent solution necessary !!!