Achievements and open issues in impurity profile control at JET.

1
Achievements and open issues in impurity profile
control at JET. M. Valisa and Angioni
Carraro Coffey Lauro-Taroni Predebon
Puiatti Alper Belo Corrigan vanEester Garzotti
Giroud Lerche Mantica Naulin Tala Tsala et al

JET E1\/E2 meeting - Culham 7 April 2011
2
Outline

• Background
• What we have learnt at JET of the effect of RF on
  impurity transport
• Open Issues

3
Background 1

• Impurity accumualtion avoidance may require
  Active Control to guarantee stationary plasma
  fusion experiments and optimization of reactor
  efficiency.
• Codes validation is required to include
  impurities and their control in a ITER/ DEMO
  flight simulator
• RF well know empirical means to pump out
  impurities in present day experiments. Underlying
  mechanims still uncertain.

4
Background 2

• For core issues, what really matters is the
  relationship between
• D_impurities , D_fuel and ce,i , since the
  relevant parameter is dilution.
• Used impurity density perturbations ( trace
  impurity )to work out impurity transport
• as with laser ablation. Modelling of the
  transient evolutions of the impurities provides
  an estimate of the transport coefficients.

Used 1D transport model with
accurate atomic physics
At stationarity

Main diagnostics for metal impurities SXR,
emission lines , bolometry
5
What we have learnt at JET
6
Shown that D imp and ce,i in some situations go
together
RF (3He minority) deposition radius
Change RF deposition profile (heat modulation to
work out heat transport P Mantica Gas
modulation or pellet for DD L Garzotti)
See ME Puiatti et al PoP 2006
7
RF power on electrons is effective as a means to
control heavy impurities in JET low
collisionality regimes
n eff 10-14 ltne gt ltTe gt2 Zeff R 0.2
MH 58144
MC 58149
Dominant ITG ? inward v
Subdominant TEM ? outward v
M-E. Puiatti PoP 13 2006 C. Angioni et al PRL
2006
8
RF power as a means to control heavy impurities
also in JET Elmy H mode / HIGH DENSITY and Ar
puffing in JET
Shots 53548 , 53015 WITH ICRH convection may
become OUTward

• Core diffusion decreases
• Core convection also decreases and may become
  outward

NO ICRH Shot 52136 Strong INward convection
M.E. Puiatti et al .Plas. Phys.Contr. Fus.
44(2002)1863
9
LBO ICRH power scan Ni and Mo are expelled
from the centre as power increases
H minority / H mode / low collisionality/ about
12 MW NBI, 1.5MA, 3T
r 0.2
Mo( 42) and Ni ( 28) similar behaviour
RF power scan
Ni and Mo
Open symbols Shots around 58140 He3 minority
68383 and 81 marginal H mode (L-mode, but
PgtgtLH threshold) Low collisonality, Similar
triangularity and elongation
10
GS2 Simulation of the shots with RF power scan
Quasi linear, electrostatic
No sign of flow inversion with increaasing RF
11
Mo( 42) and Ni ( 28) have very similar behaviour
r 0.5
Ni and Mo
12
Discharges of the RF power scan

RF Power increase
Target Plasma Ip1.5 MA B 3T NBI 12 MW Low
triangularity No sawteeth Central ICRH
13
Discharges of the RF power scan
14
RF power scan and LBO injection of Ni
r 0.2
Out of many correlation attempts ( with rotation,
Ti/Te, q and q shear etc ) the best correlation
is with R/LTi Signature of a neoclassical trend?
Nickel
15
Good correlation of v/D with R/LTe
r 0.2
16
Correlation with toridal rotation
r 0.2
17
Correlation with q and q shear
r 0.2
18
Good correlation of v/D with R/LTe
r 0.2
Stationary profiles extrapolation using
evaluated vand Ds
3MW RF
1 MW RF
NO RF
19
neoclassical V and D from NCLASS
Neoclassical transport parameters too small do
not macth the experiment
0.2
0.15
0.05
0
20
Sensitivity study on neoclassical transport
Simulation - normalized
chord integrated central SXR emission during
the injection of Ni in discharge 74360
Experiment
V neo and D from Exp.
v neo and v/D from Exp
LBO
21
Ds and Vs in the ICRH scan database
Impact of RF scan seems to be more on v than on D
22
Open issues
23
Open Issues

1. Understand the pump out effect of ICRH

• - Role q and q shear?
• Required shots with similar settings but
  different timing of LBO during pulse
• or different timing of ICRH and or NBI
• Role of rotation and shear rotation ?
• Counter beam or different share of ICRH and NBI
  keeping total power constant.

ICRH
24
Open Issues

• Is there a direct role of the RF itself ?
• Test different heating schemes ICRH on
  fundamental harmonic ,
• He3 minority heating
• Are neoclassical terms correctly evaluated?
• - In/out asymmetries / role of centrifugal
  forces. Impact on analysis
• Can sawteeth be as efficient as RF ?
• How large and frequent must ST be? Shown in the
  past that if small their
• efficiency is smaller than 2MW RF (Puiatti et al
  PPCF 2003)

2) Issue of poloidal asymmetries
3) Efficiency of RF compared to sawteeth
25
Open Issues

• 4) Effect of ICRF on Zeff
• In 68383 ( 8 MW ICRH ) Zeff increase from 2 to
  4-5 with Zeff from C
• nearly constant ( L Carraro et al EPS Warsaw)
• See also JET works by Czarneka where the problem
  has been investigated
• in some details. Lot of work also on other
  machines.
• 5) Analysis Tools for dealing with Tungsten
• - Do we have reliable tools for detection and
  analysis ?
• - Will W radiation be overwelming to make
  traditional techniques
• ( such as LBO with Ni and Mo) useless?
• 6) How to implement a feedback control system on
  impurity accumulation

26
Detection and analysis of W on JET

• Data available are
• some SXR/VUV spectroscopic lines (KT2 and KT4)
  with a fairly coarse time resolution. Have lines
  been identified ?
• Soft-X rays a vertical camera with 34 l-o-s
  (250µ filter) and a horizontal camera with 17
  channels (350µ filter).

27
Detection and analysis of W on JET
Heavy impurity transport simulations. Available
the predictive impurity transport simulation
JETTO/SANCO The ADAS tables can be used to
calculate the local emissivities that integrated
along the various l-o-s can simulate the
experimental SXR channels Standard treatment
ZI 1 equations with ionisation and
recombination to and from neighbouring ionised
states provided by ADAS/adf11 tables.
Superstages treatment Reduced set of
equations each representing a bundle of
contiguous ionised stages (a superstage) in
coronal equilibrium between each other. W ?from
74 to 35 , or more aggressively down to 10
superstages.
See L Lauro-Taroni H Summers et al presentation
at the General Task Force T Meeting 16 February
2009
28
Tungsten data available at JET Most recent W
LBOs have been performed during C17 (Nov
2006) 68373 3.2T/2.3MA 4.5 MW ICRH, 8.9MW
NBI, 0.5 MW LHCD 68374 6
MW ICRH, 8.9 MW NBI, 0.8MW LHCD 68387
7.7 MW ICRH, 9 MW NBI, 1.1MW LHCD
W LBO data available at JET available for testing
tools
SXR
68373
BOLO
29
Example of SXR after LBO of W shwoing central
peaking
SXR Horizontal cameras - B.Alper
68373. W ablation at t55 s
30
Example of SXR after LBO of W showing polidal
asymmetry
In-Out Asymmetry
SXR Vertical Camera ( B.Alper)
68373 W ablation at t55
31
Effect of ionization stages partitioning has been
tested
Different superstages partitions of W 74 ion (no
bundling) 35 superstages natural bundling 26
superstages natural bundling, with 55 upwards
bundled into 2 SS 10 superstages ions with
ionisation potential gt800 eV bundled into 1
SS ( for edge plasmas).
All partitions yield the same nW(x,t), same total
number of particles, same Power
But the partition into 10 SS yields a slightly
different SXR simulation, with a faster rise in
the initial phase. ONGOING WORK Could be
superstage treatment be included in feedback
controlled system?
32
Example of ongoing work simulation of Prad and
SXR after W LBO
50
?P bolo P sim
D (m2/s)
V (m/s)
W source
time
100
? Jetto
SXR Black V Ch. 3
(peripheral) Red V Ch 12 (central) Light blue
H Ch11 (central)
time
By L Lauro-Taroni
33
Summary and conclusions

• Central ICRH effective on JET to pump out Ni and
  Mo which feature peaked profiles in JET NBI only
  H mode plasmas
• About 3 MW ICRH required in the analysed shots
• W only very Preliminary analysis by Lauro
  Taroni
• Mechanism for impurity pump out? Trends recall
  neoclassical transport ( proportional to R/LTi)
  but absolute values do not fit
• Ni (28) and Mo(42) seem to behave similarly ,
  and W??
• ICRH is accompanied by higher Zeff .

• Possibility of treatment of W in superstages
  successfully implemented in JETTO/SANCO (ADAS
  files for bundled impurities can be generated )
• Simulations of a Tungsten injection in JET
  started

34
Ni injections
SHOT Ip(MA) Bt(T) NBI (MW) ICRH (MW) H/He3
58143 1.8 3.27 13.6 4.7 He3
58149 1.8 3.27 14.6 5.1 He3
66432 1.8 3.35 20 2 He3
66434 1.8 3.35 20 2 He3
68383 2.3 3.2 8.3 8 H
69808 1.8 3.2 11 0 H
74354 1.5 3 12 0 H
74355 1.5 3 12 1 H
74359 1.5 3 12 3 H
74360 1.5 3 10.7 2.9 H
74363 1.5 3 10.5 2.9 H
35
Mo injections
SHOT Ip(MA) Bt(T) NBI (MW) ICRH
(MW) H/He3
68381 2.3 3.2 9 8.45 H
74357 1.5 3 12.5 0 H
74362 1.5 3. 10.5 3 H
36
Soft-X rays a vertical camera with 34 l-o-s
(250µ filter) and a horizontal camera with 17
channels (350µ filter).