seminar pppl, 26 january 2011 association euratom-cea 1 interplay between neoclassical and turbulent...

Seminar PPPL, 26 January 20111

AssociationEuratom-Cea

Interplay between neoclassical and turbulent transport

on Tore Supra

Acknowledgements: J. Abiteboul, F. Clairet, C. Fenzi, S. Futatani1, V. Grandgirard, R. Guirlet, O. Gürcan2, P. Hennequin2,

S. Heuraux3, H. Lütjens4, R. Sabot, Y. Sarazin, E. Trier2, L. Vermare2, D. Villegas, and the Tore Supra Team

X. Garbet

IRFM, CEA Cadarache

1Iter Organisation, 2LPP Ecole Polytechnique, 3IJL Nancy, 4CPhT Ecole Polytechnique



Motivation and outline

• Usual statement : perpendicular transport is turbulent, parallel dynamics is collisional – in practice, a competition.

• When axisymmetry is lost, competition between turbulence and collisional processes is exacerbated.

Outline• Large scale flows : competition between turbulence

Reynolds stress and viscous damping• Particle transport : competition between Ware and

turbulent pinch velocities – Perp. diffusion is enhanced in presence of helical perturbations.



Turbulence diagnostics and codes on Tore Supra

Casati 09• 2 Doppler and 2 fast sweeping reflectometers: profile of turbulence intensity, frequency and wavenumber spectra

• Numerical tools: global (GYSELA) local (GYRO, GENE) gyrokinetic codes, integrated modelling (CRONOS)



Large scale flows : interplay between turbulence generation

and collisional damping

• Tore Supra: large (7%) magnetic field ripple

• Leads to magnetic braking in toroidal direction – controls radial electric field.

• Competes with turbulence Reynolds stress



Flow dynamics

• Radial force balance equation

• Reynolds stress + collisional viscous damping (w/o MHD , fast ion losses) Diamond 05

determine V , V , et Er

neoneot~~ VVVVV

Reynolds stress

Viscous damping

0ne

pBVBVE r

r



Rotation measurements

2 2.2 2.4 2.6 2.8 3 3.21.81.6Z

(m

)

R (m)

CXRS

Doppler reflectometry

phr

lab VBE

V

• Scan in ripple : plasma size

• Doppler reflectometry

Assumption : Vph « -Er/B - supported by simulations direct measurement of Er

• CXRS: measurements of V

Hennequin 09



Radial electric field

• Radial electric field is negative : controlled by thermal ripple losses

• Local trapping krip=3.37 Connor 73 . Ripple-plateau krip=1.5 Boozer 80,

Kovrizhnykh 99

Trier 08

drTdT

kdrn

dneT

Ei

irip

i

iir

r/a=0.6



Intrinsic toroidal rotationFenzi 10

ripple

VC

o-ro

tatio

nC

ntr

• Consistent with competition between ripple friction and turbulent Reynolds stress Hahm 07, Peeters 07, Gurcan 08, Diamond 09, …

neoneoVV

• Counter-rotation in Ohmic plasmas for large ripple

• Co-rotation with smaller ripple.



Neoclassical viscous damping

• Strong collisional friction at the passing/trapped boundary

• Poloidal flow scales with T:

eBT

kV irneo

[Kim '91]

Dif-Pradalier 08

Collisionality

kneo



Poloidal rotation agrees with neoclassical theory in global

simulations• Consistent with

poloidal viscous damping

• However: large corrugations due to turbulent Reynolds stress. v

Total

vNeo

Dif-Pradalier 09

neoneoVV



Reynolds stress important for shear flow

• VVneo, but velocity shear is different

• ExB shear

not the same with or w/o turbulence

Sarazin 10

BE

drd r

E



Is the poloidal rotation neoclassical on Tore Supra?

• Doppler reflectometry Er always agrees with ripple=0 in the core

• CXRS V changes from cntr to co-rotation when ripple becomes weaker

• Radial force balance equation VVneo at low ripple

• Not impossible (e.g. JET Crombé 06, DIII-D Kim 94 TFTR transient Bell 98 ) but surprising : turbulence simulations find VVneo + turbulence corrugations



Particle and impurity transport in the plasma core (q<1)

• Usual form of particle transport =-Dn+Vn

• Usually D,V >> Dneo,Vneo

• Not always true in the core, where values get closer to neoclassical (“neoclassical” here means axisymmetric value)



Modelling of transient confirm the diffusion/convection model

Futatani 10V

nn

Dn



Particle pinch is usually turbulent on Tore Supra

• Long pulses (6mn) with LHCD Eind 0 Ware pinch=0

• Ionization source localised in the edge, pinch due to waves is small.

V (electrons) is turbulent

Hoang 04

0.0001

0.001

0.01

0.1

1

10

0 0.2 0.4 0.6 0.8 1Normalized radius

V(m/s)

D(m2/s)

Vneo (m/s)



Transport during sawtoothing ohmic plasmas

• Pinch velocity close to Ware value, diffusion is not neoclassical

Guirlet 08



Fluctuation level is low inside q=1Dturb is small

%nn

R



0 0.2 0.4 0.6-6

-4

-2

0

2

r/a

Vtrans+stat

0 0.2 0.4 0.60

0.5

1

1.5 Dtrans+stat

Guirlet 08What about impurities?

• Supersonic gas injection or laser blow-off.

• Both transient and steady profiles are used to assess fluxes =-Dn+Vn

• Allows to reduce error bars on D and V by a large factor.



Guirlet 10

D and V for electrons and Nickel vs neoclassical

Impurities are neoclassical in the core, electrons are not

• Vnickel and Dnickel are neoclassical

• How is it possible to get

De>>DNiDNi,neo >>De,neo

• Simplest theory

c,Ni<< c,e ???

c2

Eturb v~D



• m=1,n=1 activity part of the time

• Estimation of magnetic field amplitude

a few 10-3

Is the internal kink mode responsible for high De?

Sabot 10

p

//LB

B



Orders of magnitude

exp,Ni123

TZ

2DZ

2

2

RP,Ni Dsm10v

Rvq1

1D

exp,e122

e

2e,D

2

BD,e Dsm10.2v

6.6D

• Ellectrons: banana-drift regime

• Nickel: ripple-plateau regime

• In summary:

- Collisional electron transport is enhanced–still off by a factor 3-5

- Ni neoclassical transport is unchanged



Conclusions

• GK simulations show that VVneo but radial corrugations due to turbulence are not negligible.

• Er and V agree with neoclassical theory in ohmic plasmas, with strong ripple. Situation at low ripple is unclear: looks like V departs from Vneo.

• Some evidence that in the core, collisional transport due to n=1,m=1 helical perturbation can matter

seminar pppl, 26 january 2011 association euratom-cea 1 interplay between neoclassical and turbulent...

Documents