Particle transport and density profile behaviour on JET

1
Particle transport and densityprofile behaviour
on JET

• L. Garzotti
• with contributions from
• G. Corrigan, X. Garbet, D. Heading,
• T. T. C. Jones, P. Lang, P. Mantica,
• H. Nordman, V. Parail, B. Pégourié, G. Saibene,
• J. Spence, P. Strand, M. Valovic, J. Weiland
• (work performed within the JET TF-T
• particle transport working group)

2
Motivation of the work

• Study particle transport in different scenarios
• Investigate the mechanisms leading to density
  profile peaking (in particular the existence of
  an anomalous pinch velocity)
• Compare with predictions of existing theories

3
Description of the work

• Many scenarios analysed L-modes, H modes both
  stationary, quasi-stationary and transient
• JETTO transport code with different models
  employed completely empirical, semi-empirical
  (mixed Bohm/gyro-Bohm) and physics based
  (Weilands ITG and TE modes)
• Particular attention paid to the density profile
  peaking and possible effects of an anomalous pinch

4
Discharges analysed
5
Model description (1)mixed Bohm gyro-Bohm
electron/ion thermal conductivity
particle diffusivity
pinch velocity (Ware pinch always present)
(trapped electron modes, turbulence equipartition)
6
Model description (2)Weilands

• ITG and trapped electron instabilities
• growth rate for each point of the radial mesh
• transport matrix obtained by quasi-linear theory
  combined with saturation level estimate
• the presence of terms like ?T/T, ?n/n and ?B/B
  introduces off-diagonals elements in the
  transport matrix

7
L-mode
51034 - Steady state plasma heated
by radio-frequency only 49030 - Multiple deep
pellet injection 51084 - shallow pellet,
optimised shear 55804 - shallow pellet,
non-optimised shear
8
Shot 51034 only RF heating
Mixed Bohm/gyro-Bohm V Ware pinch V cT
D ?T/T cT -0.25 V cq D ?q/q cq
0.75 Weilands off-diagonal elements Similar
results are obtained for shot 49030 (deep
pellet fuelling).
9
Shot 55804 - standard shear
Analytical D ( 0.6 m2/s) V Ware pinch V
-cT D ?T/T cT 0.25 V cq D ?q/q cq
0.5 Weilands off-diagonal elements
Interferometer (KG1) ? Thomson scattering
(LIDAR)
Simulation of shot 51084 (shallow pellet,
optimised shear) doesnt need any pinch velocity.
JPN 55804
10
Comments on the analysis of L-modes

• Equivalence between mixed Bohm gyro-Bohm
  anomalous pinch Weilands off-diagonals
  elements
• Compatibility with pinch velocity V
  -DcT?T/T 0.125 lt cT lt 0.25 V
  Dcq?q/q 0.25 lt cq lt 0.75 (5 to 10
  times larger than the Ware pinch)
• Off-diagonals elements in Weilands model provide
  the correct particle convection

11
H-modes
47744 - steady state sawtoothing plasma,
mainly radio frequency heated 52979 -
steady state sawtoothing plasma,
mainly NBI heated, slow puffing
12
Shot 47744 mainly RF heating
Weilands without off-diagonals elements
with off-diagonals elements
Mixed Bohm/gyro-Bohm V Ware pinch V
-cT D ?T/T cT 0.125 V cq D ?q/q cq
0.125
13
Shot 52979 slow puffing sawteeth
Weilands without off-diagonals elements
with off-diagonals elements
Mixed Bohm/gyro-Bohm V Ware pinch V
-cT D ?T/T cT 0.125 V cq D ?q/q cq
0.25
14
Comments on the analysis of H-modes

• In the mixed Bohm/gyro-Bohm model, due to the
  fact that D is much lower than in L-mode, the
  effect of the pinch velocity is strongly reduced.
  Particle convection becomes compatible with
  neo-classical Ware convection. In some case this
  is in agreement with the experiment in some other
  it is not. Why?
• A similar effect is observed in the Weilands
  model. The effect of the off-diagonal elements is
  not clear. Sometimes they are needed sometimes
  they arent. Why?
• In general Weilands model seems to do better
  than the mixed Bohm/gyro-Bohm in the gradient
  zone (0.5 lt r/a lt 0.9)

15
Conclusions (1)

• Particle transport has been analysed in different
  scenarios.
• The JETTO code has been employed with analytical,
  semi-empirical and physics based transport model.
• The semi-empirical mixed Bohm gyro-Bohm works
  well in L-mode.
• The physics based Weilands model is as good as
  mixed Bohm gyro-Bohm and gives a better
  reproduction of the gradient zone in H-mode
  (although the role of the off-diagonal elements
  is still not clear)

16
Conclusions (2)

• Convective velocity
• L-mode simulations are compatible with an
  anomalous pinch velocity 5 to 10 times greater
  than the neo-classical Ware pinch. (Reasonably
  well established)
• Dependence on q-profile detected in shallow
  pellet simulations
• H-mode simulations are less clear sometimes the
  expected lower convection (comparable with the
  Ware pinch) is in agreement with experiments
  sometimes it is not

17
Shot 49030 pellet injection
Mixed Bohm/gyro-Bohm V Ware pinch V
-cT D ?T/T cT 0.125 V cq D ?q/q cq
0.25 Weilands
18
Shot 51084 - optimised shear

• Simulation with analytical D and without
  anomalous pinch
• D ( 0.3 m2/s) fitted to have a density profile
  evolution compatible
• with the temperature pulse propagation
• with the integral line density measured by the
  interferometers chords

JPN 51034
19
Shot 52961 slow puffing no sawteeth
Mixed Bohm/gyro-Bohm V Ware pinch V
-cT D ?T/T cT 0.125 V cq D ?q/q cq
0.25 Weilands
20
Partial conclusions for cold pulses

• Optimised shear. No anomalous pinch. (As in
  H-mode).
• Standard shear. Anomalous pinch (5-10 times the
  Ware pinch). (As in L-mode).

21
Model description (1)mixed Bohm gyro-Bohm
electron/ion thermal conductivity
particle diffusivity
pinch velocity (Ware pinch always present)
(trapped electron modes, turbulence equipartition)