MAST and the Impact of Low Aspect Ratio on Tokamak Physics

Brian Lloyd for the MAST TeamEuratom/UKAEA Fusion Association

This work was jointly funded by the UK Engineering & Physical Sciences Research Council and Euratom

B. Lloyd 1), J-W. Ahn 2), R.J. Akers 1), L.C. Appel 1), D. Applegate 1), K.B. Axon 1), Y. Baranov 1), C. Brickley 1),

C. Bunting 1), R.J. Buttery 1), P.G. Carolan 1), C. Challis 1), D. Ciric 1), N.J. Conway 1), M. Cox 1), G.F. Counsell 1),

G. Cunningham 1), A. Darke 1), A. Dnestrovskij 3), J. Dowling 1), B. Dudson 4), M.R. Dunstan 1), A.R. Field 1), S. Gee 1),

M.P. Gryaznevich 1), P. Helander 1), T.C. Hender 1), M. Hole 1), N. Joiner 2), D. Keeling 1), A. Kirk 1), I.P. Lehane 5),

F. Lott 2), G.P. Maddison 1), S.J. Manhood 1), R. Martin 1), G.J. McArdle 1), K.G. McClements 1), H. Meyer 1),

A.W. Morris 1), M. Nelson 6), M. R. O'Brien 1), A. Patel 1), T. Pinfold 1), J Preinhaelter 7), M.N. Price 1), C.M. Roach 1),

V. Rozhansky 8), S. Saarelma 1), A. Saveliev 9), R. Scannell 5), S. Sharapov 1), V. Shevchenko 1), S. Shibaev 1),

K. Stammers 1), J. Storrs 1), A. Sykes 1), A. Tabasso 1), D. Taylor 1), M.R. Tournianski 1), A. Turner 1), G. Turri 2),

M. Valovic 1), F. Volpe 1), G. Voss 1), M.J. Walsh 10), J.R. Watkins 1), H.R. Wilson 1), M. Wisse 5)

and the MAST, NBI and ECRH Teams.

1)EURATOM/UKAEA Fusion Association, Culham Science Centre, Abingdon, Oxfordshire OX14 3DB, UK

2)Imperial College, Prince Consort Road, London SW7 2BZ, UK

3)Kurchatov Institute, Moscow, Russia

4)Oxford University, UK

5)University College, Cork, Ireland

6)Queens University, Belfast, UK

7)EURATOM/IPP.CR Fusion Association, Institute of Plasma Physics, Prague, Czech Republic

8)St. Petersburg State Politechnical University, Polytechnicheskaya 29, 195251 St. Petersburg, Russia

9)A.F. Ioffe Physico-Technical Institute, St. Petersburg, Russia

10)Walsh Scientific Ltd, Culham Science Centre, Abingdon, OX14 3EB, UK

Outline

Properties of Low Aspect Ratio Plasmas

Summary

MAST - Mega Ampere Spherical Tokamak

- confinement & transport

- stability

- exhaust

Impact of Low A on Tokamak Physics:

Properties of Low A Plasmas

Bp(R+a) ~ Bt

large field line tilt & low parallel power density in the outboard SOL

enhanced trapping

Bt(R-a) / Bt(R+a) ~ 5

Impact on transport,resistivity

Low moment of inertia high flow velocity (V ~ Vi

th)

Large inherent ExB flow shear suppression of micro-instabilities (ITG)

Strong paramagnetism

High shaping (, ) high Ip capabilityHigh performance at low B

Improved stability + good confinement high beta( ~ 40% in START, NSTX)

super-Alfvenic ionsFast particle driven instabilities

Increased decouplingof j(r) & q(r)

-40

-30

-20

-10

0

10

20

30

40

50

0.2 0.4 0.6 0.8 1 1.2

R (m)

an

gle

(

de

gre

es)

field-line angle (from EFIT)

pellet ablation cloud

#4917

sepa

ratr

ix

Large field line tilt - emphatically illustrated during pellet injection

1 2 3 4

5 6 7 8

Open divertor, up-down symmetric - upgraded 2004

Graphite protection on all plasma contacting surfaces

Adaptable fuelling systems - inboard & outboard gas puffing/multi-pellet injection

Digital plasma control (PCS supplied by GA)

Plasma cross-section and current comparableto ASDEX-U and DIII-D.

L-mode (t= 170ms) H-mode (t = 290ms)

#2948

MAST

R/a = 0.85/0.65m, 2Ip 2MA, B = 0.52T@R

ZEBRA 2D image

2D Zeff CCD detector (ZEBRA)128x128 @200Hz, 256x256 @100Hz

Reconstruction of Zeff from visible bremsstrahlung:

Abel invertedemission

Z=1 prediction from TS

Charge-exchange recombination spectroscopy:

Ion temperature & flow velocity measurements

Resolution of original CXRS system (19 chords) marginal for steep ITBs

Upgraded CXRS facilitated by adaptable low A configuration 200+ chord spectrometer spatial resolution ~ i

poloidal and toroidal chords separate views of two NBI beams

co-NBI cntr-NBI

q=1 surface appears(from SXR data)

Strong toroidicity in the ST significant neoclassical enhancement of plasma resistivity:

Ultra-high resolution TS and visible bremsstrahlungmeasurements of Te and Zeff in MAST allow neoclassical resistivity to be assessed with uniqueaccuracy.

neoclassical

Spitzer

Confinement & TransportMAST data significantly extend confinementdatabases e.g. should give greater confidencein and dependencies

Dataset improved e.g. spread in mainly determined by plasmas with conventionalD-shaped cross-section E

MAST ~ EIPB98y2 but MAST data support

somewhat stronger dependence (E 0.8) than IPB98y2 scaling [Valovic IAEA 2004]

MAST data also exert strong leverage on two-term models of confinement:

Wped –2.13±0.28 [Cordey et al NF 2003]

Pscal(MW)P

th(M

W)

H-mode Power Threshold

Pscal |Bout|0.7ne0.7S0.9F(A)

Low A data:

- clearly favour Pth ~ S rather than Pth ~ R2

- favour dependence on |Bout| rather than Bt(0)

|Bout|2 = Bt2 + Bp

2

[Takizuka et al PPCF 2004]

The (non-linear) aspect ratiodependence is not yet well-determined - postulated by Takizuka et al that it may take a form related to fraction of untrapped particles

H-mode Access

Inboard Gas Puff

H-mode access improved with inboard fuelling - also seen in NSTX and in COMPASS-D (effects less pronounced).

Linked to effect on edge rotation

Two effects may contribute (both enhanced at low A) - neutral viscosity contribution to angular momentum transport (Helander et al) - net torque arising from B drift effects (Rozhansky et al)

H-mode access also optimised for a connected DNDconfiguration with |rsep| < i , i (i 6 mm q ) - indications of a similar (but weaker) effect in AUG

Dependence on fuelling location & magnetic geometry appearsto be amplified at low A - helps to provide improved insight intofactors influencing H-mode transition

Microstability in STs

In STs ITG turbulence might be suppressed by intrinsic pressure driven flow shear since SE/m ~ i* [Kotschenreuther et al 2000]

GS2 ITG growth rate ExB shearing rate in typical MAST H-mode discharge [Applegate et al EPS 2004]

Electromagnetic effects stabilising effect on ETGs but in the plasma core can give rise to dominantly unstable tearing parity modes in the ITG wavelength regime.

ETG growth rate much larger - ‘mixing length’ estimates too low to account for observed transport but non-linear effects (e.g. ‘streamers’) may play a role.

At high , tearing parity modes may be important at both ETG and ITG r length

scales - expected to enhance electron transport [Joiner et al EPS 2004]

Re

Im

radians

A||

e ~ i around mid-radius & close to iZ-CH [Chang & Hinton]

[ exact values very sensitive to relative values of Te, Ti ]

H-mode transport coefficients are close to ion neoclassical

TRANSP simulations

Ti

Te x1019

m-3

ne

e ~ i >> iZ-CH

HHIPB98y2 ~ 0.8

High performance in sawtooth-free L-mode

[m2 /

s][keV

]

TRANSP simulations

HHITER96L ~ 1.6

Strong dependence of H-mode access on magnetic configuration facilitates sensitive H-mode access control mechanism

e.g. allows H-mode/L-mode comparison at same engineering parameters

WL-mode/WH-mode ~ 2/3

H-mode/L-mode comparison discharges

High confinement in counter-NBI

HHIPB98y2 ~ 2

Measured neutron rates consistentwith LOCUST modelling - only ~ 1/3fast ion energy absorbed

But stored energy comparable to co-NBI

ne(r) more peaked, Te(r) broader thanco-NBI

High rotation (V0 ~ 340km/s) due to rapid loss of co-moving ions

in-out Zeff asymmetry x2 (dominated by C6+)

cntr-NBI

co-NBI

Sawtooth-free discharges HH 1.5 with co-NBI; HH 2 with cntr-NBI(m

s)

(ms)

HH=2

HH=1

Neoclassical Ware pinch stronger at low A ( 1/2) and augmented by beam-driven pinch for cntr-NBI (Ware pinch dominant)

- further increased due to higher Zeff of cntr-NBI discharges

Density profile peaking strongly correlated with Ware pinch

[Akers et al EPS 2004]

cntr-NBI

co-NBI

Ion & Electron ITBs

Co-NBIStrong ion ITB i ~ i

Z-CH Weaker eITB e ~ 2-3 x i

Z-CH

Cntr-NBIStrong eITB at large radius

ITB existence criteria

Criteria based on critical values of R/LT or s/LT fail - readily satisfied even when no ITB

In these discharges the driven toroidal flow is the dominant contribution to the ExB flow shear

In this case ITB formation may be linked to a critical Mach number[Field et al EPS 2004]

x1019

m-3

[keV

]

TRANSP simulations

Menard calculation:

- q*/q0=3

- q*/q0=1.5

a) b)

unstable

stable

KINX calculations

Stability

Sawtooth triggered NTMs have been observed in MAST - islandevolution confirms strong role of field curvature stabilisation (Glasser) term at low A

By avoidance of NTMs N > 5, (N > 5li) has been achieved in MAST approaching ideal no-wall beta limit.

Taking the Sauter NTM model, benchmarked against MAST it appears that the STPP may be stable to NTMs

fBS ~ 40%, Wfast ~ 15 - 20%

Bootstrap Current

HighBootstrap

Low internalinductance

HighElongation

Steady-state regimes rely on high bootstrap fraction

fBS ~ 40% in a Component Test Facility (CTF) ~ 90% in an ST Power Plant (STPP)

fBS ~ Nh()Irod/Ip

{h() approx.}

High elongation at low A helps

Primary auxiliary current drive scheme - neutral beam current drive (NBCD)- implementation of an off-axis NBCD capability in MAST under investigation

Fast Particle Driven Instabilities

Low Alfven velocity (low B) + wide gaps in Alfven continuum (high toroidicity & ellipticity) TAEs & EAEs can be driven unstable by energetic ions.

Fishbone modes are seen at higher than TAEs or chirping modes - can trigger longer lasting modes

Chirping modes also observed incl. simultaneous up-down chirpingas predicted by hole-clump model(Berk et al, Pinches et al, Vann et al EPS 2004)

Discrete TAE & EAE activityin MAST - relatively benign

f vA/4qR (TAE)

f vA/2qR

(EAE)

t(s)

f (x100kHz)

CAEs having k >> k||, can be expected to be of increasing importance at high beta due to relatively weak Landau damping.Preliminary evidence for CAE mode activity in MAST [Appel et al EPS 2004]

For TAEs and chirping modes both the amplitudesand number of unstable modes decrease with increasing , in accord with theory, due to plasma pressure and thermal ion Landau damping [Gryaznevich & Sharapov PPCF 2004]

x10-1

TAE eigenfunctions at different values of thermal beta theory predicts TAE modesstabilised in MAST at high beta(as observed experimentally)

TAEs stabilised at high MISHKA

Increasing - TAE mode stabilised

ELMs

Linear D

Thomson scattering

ELMs associated with large radial effluxes at outboard side (<vr> ~ 0.75 kms-1)RP observes large jsat out to ~15 cm

Ballooningnature ofELMs

RP data + edge plasma toroidal velocitymeasurements consistent with n ~ 10filamentary structure

Kirk et al Phys. Rev. Lett. 92 (2004) 245002

Structure sometimes seen on TS profiles outboard of separatrix.

Also observed on the mid-plane linear D array but not on divertor target power footprint

ELMs

Edge profile broadening or other structure only observed in 20 - 25% of cases in which TS fires during ELM D rise.

ELM spatial structure (theory+experiment)

Image simulation of an extended structure @q=4, n=10, #8814

Features are consistent with non-linear ballooning mode theory [Cowley/Wilson] ELM composed of several narrow filaments, each extended alonga field line, projecting into SOL at outboard mid-plane and connecting back into the core plasma far along the field line

The spatial and temporal evolution of an ELM

Filament remains attached to core plasma & acts as conduit for enhanced transport into SOL

Wfil << WELM

Filament eventually detachesat outboard mid-plane

Wfil < 0.03WELM

But favourable divertor target power distribution:- large ratio of outboard to inboard separatrix area ( ~ x4) in low A plasmas- equal up-down power distribution in DND

Pout >> Pin

ExhaustSTs - small area of inboard divertor targets may lead to high power densities

High Bp/Bt in outboard SOL leads to low parallel power densities: local target protrusions intercept a small fraction of power efflux so ST is less sensitive to tile mis-alignment for example

Increases practical feasibility of advanced divertor schemes such asthe cascading pebble divertor

The ST SOL

Modelling of the MAST SOL (OSM2/EIRENE) has shown the importance of including the mirror force ||B/B for the ST [Kirk et al PPCF 2003]

- x 10 larger in MAST than in JET

In MAST the magnetic field is x5 greater in the i/b SOL than the o/b SOL, but the connection length is similar

experimental measurements indicate that outboard transport coefficients exceed inboard coefficients by a factor ~ 2 - 10

SummaryProperties of low A plasmas (high beta, low field, strong shaping, high rotation, high ExB flow shear, large mirror ratio, large field line pitch angle etc) have important effects on plasma behaviour.

Strong synergy between STs and conventional aspect ratio tokamaks:- rapid ST development due to tokamak knowledge base- STs are unique testing ground for tokamak physics

STs are providing valuable input to international databases and helping to provideimproved insight into tokamak behaviour (eg. ELM structure, H-mode access…) as well as providing stringent tests of theory (neoclassical resistivity, Ware pinch, etc.)

The new generation of STs are confirming the excellent confinement and stability properties of low A plasmas (i ~ e ~ i

neo, N ~ ideal no-wall limit)

Low A plasmas exhibit several favourable properties (e.g. large outboard/inboard power efflux ratio, low parallel power density in the outboard SOL etc) to aid development of viable divertor solutions for future ST devices.