Studies of H Mode Plasmas Produced Directly by Pellet Injection in DIII D

Transcription

1 Studies of H Mode Plasmas Produced Directly by Pellet Injection in by P. Gohil in collaboration with L.R. Baylor,* K.H. Burrell, T.C. Jernigan,* G.R. McKee, *Oak Ridge National Laboratory University of Wisconsin Presented at 27th EPS Conference on Controlled Fusion and Plasma Physics Budapest, Hungary A UCL UCLA Y OF WIS SIT MA NSIN CO UNIVER June 12 16, 2 DIS O N 81 /PG/wj

2 OVERVIEW H mode plasmas have been produced by injecting frozen deuterium pellets into L mode plasmas in Pellets injected from the low field, outside edge of the plasma and from the high field, inside plasma edge were both able to produce H mode plasmas The radial extent of pellet deposition is not important. The production of a steep edge density gradient is important The large influx of particles at the plasma edge from the pellet leads to substantial reductions in the edge electron and ion temperatures. The lowered temperatures are still conducive for the formation of the H mode transport barrier A critical edge temperature is not necessary in these H mode transitions Pellet induced H modes have LH transitions at plasma parameters far below theoretical predictions The power threshold for the H mode transition is reduced by about 2.4 MW (about 3%) using pellet injection Pellets produced H mode plasmas at lower input power than reference plasma discharges without pellets, which stayed in L mode throughout beam heating even in the presence of strong sawteeth 81 /PG/wj

3 A key issue for the physics of H mode plasmas is to determine which plasma quantities are critical for the formation of the edge transport barrier One approach is to directly perturb the edge plasma conditions and observe the subsequent changes to key edge parameters at the H mode transition One hypothesis for the H mode transition is that the attainment of a critical edge electron temperature is required for the H mode transition Injection of frozen deuterium pellets: MOTIVATION This can be directly tested using pellet injection Dramatically changes the edge electron density and temperature Can trigger H mode transitions Perturbation to the edge plasma condition by pellet injection provides for quantitative comparisons between experimental conditions and theoretical predictions from H mode transition theories Pellet induced H mode transitions can be accurately preset in time (i.e. the pellet injection time) so that key fluctuation and profile diagnostic systems can be concentrated about that time 81 /jy

4 EXPERIMENTAL SETUP An unbalanced, double-null diverted discharge with the B drift away from the dominant X point was investigated High H mode power threshold Clear, steady-state L mode conditions Pellets were launched from the inside wall, from an upper vertical port and from the outside wall of the vessel Operational parameters Plasma current, I p = 1.6 MA Toroidal magnetic field, B T = T Target electron density, n e = m 3 Auxiliary heating power (NBI) = MW Safety factor: on-axis, q() edge, q 95 = Elongation, κ = Upper triangularity, δ upp =.7.85 Lower triangularity, δ low = /PG/wj

5 PELLETS WERE LAUNCHED FROM THE LOW FIELD SIDE OR HIGH FIELD SIDE OF Outside wall launched pellets (low field side) were shattered prior to entry into plasma in order to minimize pellet penetration predominantly edge density perturbation Shot Top Launch Type: solid deuterium Pellet size: 2.7 mm Rep rate: up to 1 Hz Speed: 1 35 m s 1 Speed for Specific Shots: Shot , 194 and 16 m s 1 (shattered pellets outside launch) m s 1 (inside launch) m s 1 (inside launch) Inside Wall Launch (High field side) Outside Launch Pellet (Shattered) (Low field side) 81-/rs

6 PELLET INDUCED H MODE (PIH MODE) TRANSITION PRODUCED BY A LOW FIELD SIDE PELLET 1 Injected NB Power (MW) 5. Central T e (ρ =.1) (kev) Line Average density ( 1 19 m 3 ) Total Radiated Power (MW) 15. Edge T e (ρ =.9) Central T i (ρ = 5) (kev) Upper Divertor Photodiode (a.u.) Pellet Injection 5 Edge T i (ρ =.9) Central Toroidal Rotation (ρ = 5) (km s 1 ) 25 Edge Rotation (ρ =.9) 7. Central n e (ρ =.1) ( 1 19 m 3 ) 5. H factor (H 89P ) 3.5 Edge n e (ρ =.9) Time (ms) 2.5 Normalized Beta, β N Time (ms) 81-/rs

7 THE EDGE ELECTRON TEMPERATURE IS REDUCED SUBSTANTIALLY AFTER PELLET INJECTION A critical edge temperature is not required for the H mode transition Upper Divertor Photodiode Time (ms) Electron Density Profiles Electron Temperature Profiles 12 Electron Pressure Profiles 5. 1 n e (1 19 m 3 ) Shot ms before pellet (4248 ms) 2 ms after pellet (426 ms) After LH transition (4272 ms) ELM-free phase (4285 ms) Between ELMS (431 ms) T e (kev) ρ ρ P e (kpa) ρ.9 81-/rs

8 THE INCREASE IN THE EDGE ELECTRON PRESSURE PEDESTAL AND GRADIENT CLEARLY SHOWS THE TRANSITION TO H MODE 1 Electron pressure profiles along Thomson scattering laser path Shot ms before pellet (4248 ms) 2 ms after pellet (426 ms) After LH transition (4272 ms) P e (kpa) z (m) /rs

9 BOTH THE EDGE ION TEMPERATURE AND TOROIDAL ROTATION ARE SIGNIFICANTLY REDUCED AFTER PELLET INJECTION Pellet injection time = 4258 ms Integration time = 5 ms Ion Temperature Profiles Shot ms before pellet 4 ms after pellet After L H transition ELM-free phone Between ELMs Toroidal Rotation Profiles T i (kev) Ω ϕ (1 4 rad s 1 ) ρ ρ.9 81-/jy

10 A GRADIENT IN THE EDGE E r IS ESTABLISHED AFTER PELLET INJECTION AND IS MAINTAINED INTO THE H MODE The E r measurement is averaged over 5 ms integration time Need higher time resolution (1-2 ms) to determine fast changes in E r SHOT m m m m SHOT ms before pellet 4 ms after pellet 9 ms after pellet 14 ms after pellet 19 ms after pellet 24 ms after pellet 29 ms after pellet E r (kv/m) 1 E r (kv/m) Pellet Time 2 Separatrix DITHER L H Time (ms) R (m) /jy

11 EDGE ION TEMPERATURE PEDESTAL AND GRADIENT INCREASE INTO THE H MODE PIH mode at P NBI = 6.8 MW (shot 99559) Discharge with no pellet stays in L mode even at higher P NBI = 9.2 MW (shot 99573) Tanhfit analysis is used to determine edge local parameters 1.5 Upper Divertor Photodiode Ion Temperature (Pellet) (No Pellet) Y Pedestal = A + B; Offset = B A Pedestal Width Offset XKNEE XSYM T i at Density Knee (kev) T i Gradient (kev m 1 ).7.8 X Time (ms) /jy

12 EDGE LOCAL PARAMETERS DETERMINED FROM TANHFIT ANALYSIS CLEARLY SHOW THE TRANSITION TO H MODE WITH PELLET INJECTION A critical n e of between m 4 is required for the H mode transition ( m 4 at midplane) 1.5 Upper Divertor Photodiode Electron Density Electron Temperature Electron Pressure (Pellet) (No Pellet) 1.5 Upper Divertor Photodiode 1.5 Upper Divertor Photodiode n e at Density Knee ( 1 19 m 3 ) n e Gradient* ( 1 19 m 4 ) Time (ms) T e at Density Knee (kev) T e Gradient* (kev m 1 ) Time (ms) Pressure at Density Knee (kpa) Pressure Gradient* (kpa m 1 ) Time (ms) *Spatial measurements are along the laser path in the z-direction, not at the midplane 81-/jy 44 45

13 THE EXPERIMENTAL RESULTS WERE COMPARED WITH THREE MODELS OF THE H MODE Rogers et al. (Proc. 17th IAEA Fusion Energy Conf., Yokohama, Japan, 1998, paper IAEA-CN-69/THP2/1). Based on 3 D simulations of the Braginskii equations α MHD 2µ q 95 κ1/2 a r xpt (2 P/L P ) R B T dψ/dr 2 ( ρ s c s )( t ) α DIAM = L pi,e L Transport is suppressed for α > MHD.5 and α > DIAM.5 (for ) Pogutse et al. (Proc. 24th EPS Conf., 1997, paper P3-141). Based on stabilization of Alfven waves parameterized in terms of normalized beta, β N, and the normalized collision frequency, ν n ( m i ) β N = m e 1/2 4π n T e 2 B k χ p 1 ; ν n = ( ) m 1/4 χ1/2 i p m e λ e k 1/2 81 /PG/wj

14 THE EXPERIMENTAL RESULTS WERE COMPARED WITH THREE MODELS OF THE H MODE (Continued) χ p characterizes the pressure gradient scale length, k is the parallel wavenumber, λ e is the mean free path 2/3 Transport is suppressed when β N > β CRIT = 1 + ν n Wilson et al. (Proc. 17th IAEA Fusion Energy Conf., Yokohama, Japan, 1998, paper IAEA-F1-CN-69/TH3/2). Based on stabilization of peeling modes at collisionality > 1 parameterized in terms of α MHD and ν* Transport is reduced when α > MHD.5 and ν* > 1 81 /PG/wj

15 PELLET INDUCED H MODES HAVE L-H TRANSITIONS AT PLASMA PARAMETERS FAR BELOW THEORETICAL PREDICTIONS Rogers et al. Proc. 17th IAEA Fusion Energy Conf. Yokohama, Japan 1998, paper IAEA-CN-69/THP2/1.6 H mode Pogutse et al. Proc. 24th EPS Conf (P3-141) H mode Wilson et al. Proc. 17th IAEA Fusion Energy Conf. Yokohama, Japan 1998, paper IAEA-F1-CN-69/TH3/2.6 H mode α mhd.4 Shot ms before pellet 2 ms after pellet After LH transition ELM-free phase Between ELMs β /µ 1.5 L mode α mhd.4 L mode.2 L mode After L H Transition.5 After L H Transition.2 After L H Transition α diam After pellet After pellet After pellet ν µ.5 ν* 81-/rs

16 THE POWER REQUIRED TO ACCESS H MODE IS REDUCED BY AT LEAST 2.4 MW INJECTED POWER USING PELLET INJECTION 1 5. PIH Mode (Shot 99559): LFS Pellet, P NBI = 6.8 MW No Pellet, L Mode (Shot 99573): P NBI = 9.2 MW Injected NB Power (MW) Edge T e (ρ =.9) (kev) Total Radiated Power (MW) Shattered LFS Pellet n e ( 1 19 m 3 ) Edge T i (kev) (ρ =.9) Upp. Div. PD (a.u.) Pellets 4 2 Neutron Rate ( 1 15 s 1 ) 7. Edge n e (ρ =.9) ( 1 19 m 3 ) 3. Normalized Beta, β N /rs

17 PIH MODE TRANSITION PRODUCED BY HIGH FIELD SIDE LAUNCHED PELLET (P NBI = 6.7 MW) 1 5. Injected NB Power (MW) 4 2 Central T e (ρ =.1) (kev) Total Radiated Power (MW) Edge T e (ρ =.9) Line Average density ( 1 19 m 3 ).5 Edge T i (ρ =.9) (kev) Upper Divertor Photodiode (a.u.) Pellet Injection 6 3 Edge Toroidal Rotation (ρ =.9) (km s 1 ) 1 Central n e (ρ =.1) ( 1 19 m 3 ) Neutron Rate ( 1 15 s 1 ) 5. Edge n e (ρ =.9) /rs

18 THE HFS PELLET PENETRATES MUCH FURTHER INTO THE PLASMA INTERIOR, BUT STILL PRODUCES A SIGNIFICANT DENSITY GRADIENT AT THE PLASMA EDGE (a.u.) Pellet Time (ms) Upper Divertor Photodiode Shot Electron Density Profiles Electron Temperature Profiles 3 Electron Pressure Profiles 1 2 n e (1 19 m 3 ) Shot ms before pellet (3598 ms) 2 1 ms after pellet (361 ms) 22 ms after pellet (3622 ms) After dither phase (3648 ms) before ELM (3698 ms) ρ T e (kev) ρ.6.8 P e (kpa) ρ 81-/rs

19 FAST DITHERING OR BURSTING OF FLUCTUATION APPEAR ~1 ms AFTER PELLET INJECTION Fast dithering develops into ELM free H mode D α (Reflmon) Amplitude (a.u.) f mean, f rms (khz) D α Pellet time Fluctuation amplitude RMS, f >15 khz Mean and std. dev. fluctuation frequency f mean, (f > 1 khz) f rms, (f > 1 khz) Power spectra S (f) versus time Shot khz UCLA EE UCLA Electrical Engineering 361 Time (ms) /rs

20 EDGE TURBULENCE DURING PELLET-INDUCED H MODE TRANSITION Beam Emission Spectroscopy measurements show different stages of transition behavior ( < k < 3 cm 1, 2 f 2 khz, ρ =.93) Power spectra condenses to low frequency after pellet injection Integrated power remains nearly the same H mode phase shows markedly reduced fluctuation level (2 orders of magnitude reduction in power) BES Fluctuation Signal (a.u.) Fluctuation Power ((di/i) 2 /khz) L Mode Dithering Phase ELM-free 2 (after pellet) H mode Shot Time (ms) Frequency (khz) 1) Pre-pellet L mode phase (moderate fluctuations) 2) Post-pellet, L mode dithering phase (lower frequency fluctuations, dithers) 3) H mode (very low fluctuation level) Frequency-filtered time evolution UNIVERSITY OF WISCONSIN M A DIS O N Spectral Power Comparison Before Pellet (ñ/n = 3.7%) After Pellet (ñ/n = 3.7%) ELM-free H mode (ñ/n =.5%) ρ = /rs

21 PIH MODE TRANSITION PRODUCED BY A HIGH FIELD SIDE PELLET AT REDUCED NBI POWER (P NBI = 4.9 MW) 1 Injected NB Power (MW) Total Radiated Power (MW) 4 Central T e (ρ =.1) (kev) Line Average density ( 1 19 m 3 ) 4 Edge T e (ρ =.9) Central T i (ρ = 5) (kev) 4. 2 Upper Divertor Photodiode (a.u.) Pellet Injection 1 Edge T i (ρ =.9) (kev) Central Toroidal Rotation (ρ = 5) (km s 1 ) Edge n e (ρ =.9) Central n e (ρ =.1) ( 1 19 m 3 ).5 Edge Toroidal Rotation (ρ =.9) (km s 1 ) Neutron Rate ( 1 15 s 1 ) /rs

22 THE POWER REQUIRED TO ACCESS H MODE IS REDUCED BY 2.3 MW INJECTED POWER USING HIGH FIELD SIDE PELLET INJECTION 1 Injected NB Power (MW) PIH Mode (Shot 1162): HFS pellet, P NBI = 4.9 MW No Pellet, L Mode (Shot 1161): P NBI = 7.2 MW Edge T e (ρ =.9) (kev) 5. Total Radiated Power (MW).5 8 n e ( 1 19 m 3 ).8.4 Edge T i (kev) (ρ =.9) Upp. Div. PD (a.u.) Pellet Neutron Rate ( 1 15 s 1 ) H factor (H 89P ) Edge n e (ρ =.9) ( 1 19 m 3 ) 3. Normalized Beta, β N Central n e (ρ =.1) ( 1 19 m 3 ) /rs

23 SUMMARY H mode plasmas have been directly produced by injecting frozen deuterium pellets into L mode plasmas Pellets injected from the low toroidal field side and high field side were both able to produce H mode transitions The production of a steep edge density gradient is important, and not the radial extent of pellet deposition The edge electron and ion temperatures are substantially reduced by the large influx of particles from the pellet The H mode transition still occurs at the lowered temperatures A critical edge temperature is not necessary in these H mode transitions Pellet induced H modes have LH transitions at plasma parameters far below theoretical predictions Just after pellet injection, the edge fluctuations exhibit fast dithering or bursting behavior before steady H mode conditions are achieved Similarly, fluctuation bursting is observed in transitions to VH mode plasma and plasmas with internal transport barriers 81 /PG/wj

24 SUMMARY (Continued) The shear in the edge Er increases gradually during the period of fluctuation bursts E r measurement is averaged over bursts so cannot determine fast changes in Er Future experiments will have increased time resolution The power threshold is reduced by about 2.4 MW injected power (about 3%) using pellet injection Pellets produced H mode plasmas at lower input power than reference plasma discharges without pellet Reference plasma discharges without pellets stayed in L mode throughout the applied neutral beam heating even in the presence of strong sawteeth and higher NBI power 81 /PG/wj