Rajesh MaingiOak Ridge National Laboratory

M. Bellb, T. Biewerb, C.S. Changg, R. Maquedac, R. Bellb, C. Busha, D. Gatesb, S. Kayeb, H. Kugelb, B. LeBlancb, J. Menard,b D. Muellerb,

R. Ramand, S. Sabbaghe, V. Soukhanovskiib, D. Stutmanf, NSTX Team

a) Oak Ridge National Laboratory b) Princeton Plasma Physics Laboratoryc) Los Alamos National Laboratory d) University of Washingtone) Columbia University f) Johns Hopkins Universityg) Courant Institute, New York University

H-mode WorkshopSan Diego, CA

Sept. 24-26, 2003

Effect of Gas Fueling Location on H-mode Access in NSTX

ResearchSupported by

High-field side (HFS) Gas Fueling Provides Reproducible H-mode Access in ST’s

Fulop, Helander, Catto, PRL #225003-2; 25-Nov-2002

• Center stack midplane fueling during NBI enabled reproducible H-mode access in MAST

• NSTX installed center stack midplane gas injector in FY’02 - also better H-mode access

• Theoretical calculation shows poloidal gas source location affects electric field magnitude, more at small R/a Poloidal AngleOuter Midplane

Inner Midplane

Comparison Of Center Stack Midplane And Low-field Side (LFS) Fueling Observations

• H-mode access most reproducible with HFS gas fueling from center stack midplane during NBI

• H-mode transition inhibited for high puff, with LFS fueling

• H-mode transition delayed for medium puff with LFS, compared w/HFS midplane

• H-mode transition similar at low puff for LFS and HFS mid

• Edge toroidal rotation (C-III) higher (co-) with HFS and becomes negative after L-H

Summary of Center Stack Top and Lower Dome Observations

• Center stack injector has a smaller initial dump and shorter e-folding decay time of flow rate

• Center stack top injector enabled H-mode access in double-null but not lower single-null (gas trapped in DN?)

• Lower dome dumps gas in very quickly

• Lower dome did not enable H-mode access, but prevented locked/tearing mode reconnection in L-mode discharge when timing was ‘optimized’ (lucky)

NSTX Explores Low Aspect Ratio (A=R/a) physics regime

}A ≥ 1.27

Parameters Design Achieved

Major Radius 0.85m

Minor Radius 0.67m

Plasma Current 1MA 1.5MA

Toroidal Field 0.6T 0.6T

Heating and Current Drive

NBI (100keV) 5MW 7 MW

RF (30MHz) 6MW 6 MW

Wall Conditioning:

~350 deg. bakeout of graphite tiles

Regular boronization (~3 weeks)

Helium Glow between discharges

Center stack gas injectionPassive stabilizing plates

Graphite tilesGas Injectors

Load-and-Dump Gas Injectors Have Different Flow Characteristics and Delay Time

0

50

100

150

200

CS TopLower domeCS Midplane

Flow Rate [torr-l/s]

0

5

10

15

20

25

30

35

40

0 0.2 0.4 0.6 0.8 1 1.2 1.4

CS topLower DomeCS Midplane

Total Gas [torr-l]

Time [sec]

Unfiltered Recycling Light Pattern Changes with Gas Injector Location and Plasma Shape

CS mid, LSN109866@0.2s

X-point+LFS top, LSN109884@0.2s

CS top, DN109885@0.2s

CS top, LSN109869@0.2s

LFS top, LSN109875@0.201s

CS top, USN109877@0.2s

LSN=lower single null; USN=upper single null; DN=double null

LFS

HFS

L-H transition(HFS)

LFS

Ip [MA]

PNBI [MW]

Puff Rate [torr-l/s]

Total Gas [torr-l]

Dα

[a ]u

Center-stack MP Gas Injector Fueling Allows Early L-H Transition and Longer H-mode Duration

•Experiment runwhen L-Hpower thresholdwas < 1NBI src(end of run 2002)

•Note LFS shot onlyhad L-H with 2 NBI-> higher PL-H

Center-stack MP Gas Injector Fueling Allows Higher Edge ne and Te even before L-H Transition

0

0.5

1

1.5

2

2.5

3n

e [1019 m-3] 0.167 s

0

0.1

0.2

0.3

0.4

130 135 140 145 150

Te [keV] 0.167 s

Radius [cm]

0

20

40

60

80

100

110 120 130 140 150

HFS: 109054LFS: 109059

v [km/sec]φ

[ ]Radius cm

0.150 s

0

0.5

1

1.5

2

2.5

3n

e [1019 m-3] 0.167 s

0

0.1

0.20.3

0.4

0.5

0.6

0.7T

e [keV] 0.167 s

0

20

40

60

80

100

20 40 60 80 100 120 140 160

HFS: 109054LFS: 109059

v [km/sec]φ

[ ]Radius cm

0.150 s 20 40 60 80 100 120 140 160

Radius [cm]

•Experiment runwhen L-Hpower thresholdwas < 1 NBI src(end of run 2003)

•Note CS Mid and LFS shot had same tL-H , but

repeat LFS had no L-H-> higher PL-H with LFS

More Reproducible H-mode Access with Center-stack MP Gas Injector Fueling, even at Lower Puff Rate

CS MidLFSLFS

L-H

•Experiment runwhen L-Hpower thresholdwas ~ 1-2 NBI src(start of run 2003)

•Note LFS shot and CS top had no L-H-> higher PL-H

Center-stack Midplane Gas Injector Fueling Has Lower PL-H Than LFS and CS Top

CS MidLFSCS Top

L-H

R~1.48m

Edge Toroidal Rotation Marginally Higher for HFS midplane fueling than Others Before L-H Transition

Edge rotation diagnostic: T. Biewer, TTF 2003

CS MidLFSCS TopCS Mid

L-Htimes

•Vtor generally more positive (co-Ip) for CSmidplane fueling thanouter midplane orCS top (NEED MORE DATA!!)

•Vtor changes direction after L-H

•Vpol not available forthis experiment

•Vtor measured at radiusof peak edge emission(Rpeak

mid~1.48m)

•Error bars (statistical) < 5%channel spacing ~ 3 cm

•Experiment runwhen L-Hpower thresholdwas ~ 1-2 NBI

src(start of run

2003)

Lower dome + LFS Gas Injectors Can Match CS Midplane Flow Rate But Do Not Allow H-mode Access

CS MidLdome+LFSLdome+LFS

L-H

Ldome LdomeLFS

R~1.48m

Higher Toroidal Rotation (co-NBI) Predicted for Center Stack Midplane Fueling

In/Out Comparison(before L-H)

L-H comparison: rotation counter, due to -Er and Pe

• Simulated w/XGC code (guiding center, Monte Carlo,code w/X-point and neoclassical transport)

XGC code: C.S. Chang, US TTF meeting, 4/2003

∇

Normalized Poloidal Flux, NNormalized Poloidal Flux,

N

Summary and Conclusions

• High-field side (HFS) gas fueling from midplane during NBI enabled best reproducibility for H-mode access

• HFS fueling from top allows H-mode access in DND, but not yet in LSN

• LFS injection allow H-mode access, but with higher power threshold at medium-high fueling rates

• Toroidal rotation from Monte Carlo calculations predicted to be higher with center stack midplane fueling than outer midplane fueling– consistent with analytic theory of Fulop, Helander, and Catto– first measurement qualitatively consistent with modeling

Poster copies (email address)