ELMs on C-Mod J.W. Hughes for the Alcator C-Mod team C-Mod/NSTX Pedestal Workshop Princeton, NJ September 7 8, 1
EDA H-mode is supplanted by small ELMs at higher beta 7 6 5 T e ped (ev) 4 3 Unstable Stable ELMs EDA 1 1. 1.5..5 3. 3.5 4. α MHD Higher input power ELITE analysis showed that ELMs were consistent with peelingballooning triggers Mossessian et al., Phys. Plasmas 1 (3) 17. Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 1
ELM signatures extremely sensitive to shaping 18348 183414 183417 Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1
More extreme shaping, density reduction crucial to larger ELMy access Line-averaged density ICRF power Pedestal Te 18137 18349 D-alpha Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 3
More extreme shaping, density reduction crucial to larger ELMy access 1 Te ped (ev) 8 6 4 discrete ELMs Steady EDA 18137 18349 1 3 n ped e (1 m -3 ) Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 4
Differently shaped equilibria, magnetic topologies play a big role in edge stability 1 T e,95 (ev) 8 6 4 EDA H-mode (favorable grad-b drift) ELMy H-mode (favorable grad-b drift, JFTM shape) Transient ELM-free H-mode (unfavorable grad-b drift).1.3 1 3 pedestal collisionality ν* 95 Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 5
ELMs have been observed in reversedfield (non-eda) H-modes CY comparison Fwd. B T Rev. B T 4 1 ICRF (MW) Line-integrated density (1 m - ) D-alpha 1 Radiated Power (MW).6.8 1. 1. Time (s) Unfortunately when conditions were matched in 6, ELMs were not reproduced. Hidden variables? (c.f. L-H thresholds) Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 6
ELMs have also been observed during current ramp-up Also not consistently reproduced Mossessian et al., Phys. Plasmas 1 (3) 17. Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 7
Summary of observations Robust ELMy regimes are only accessed in an atypical shape on C-Mod We seem to be on the cusp of ELM instability in many of our more typical configurations. What makes things so stable? Can modeling explain extreme sensitivity of ELM character to X-point position? Are there hidden variables in the ELM problem? Meanwhile, we have sought ways to trigger ELMs, which might prevent radiative collapse of naturally ELM-free H-modes Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 8
Can rapid vertical jogs stimulate ELMs? Would like the ability to stimulate ELMs in normally shaped low collisionality H-modes for density and impurity control Rapid vertical displacement of plasma equilibrium has been observed to trigger and/or pace ELMs on multiple devices TCV [Degeling 3], AUG [Lang 4], JET [Sartori 8, de la Luna 9], NSTX [Gerhardt 9] Experimental approach Get close to ELM stability boundary: High power H-modes with high edge T, grad-p Jog plasma centroid (ZCUR, z magx ) rapidly (dz/dt=1 4m/s ), many times during H-mode In recent (Feb/Mar 1) experiments we successfully implemented Z-modulation scheme At frequencies up to 1Hz in L-modes with q 95 >=3.1 At frequencies up to Hz in H-modes (EDA and ELM-free) with q 95 >=3.5 Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 9
~1cm peak-to-peak displacement, z, possible with 1Hz z magx -oscillation Vertical control stability maintained at f drive up to Hz, at somewhat smaller values of z Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 1
Jogs implemented easily in EDA H-modes Example: Hz oscillation applied to ZCUR controller from.9 1.4s Note ~55Hz oscillation in nonmodulated plasma Minimum jog duration ~ms in this discharge Small blips in D α observed, roughly synchronized to jogs, both up and down. (Probably a diagnostic feature) Pedestal and global confinement properties essentially unchanged P rad (MW) W mhd (MW) D z P ICRF (MW) n e (1 m -3 α (a.u.) magx (cm) ) 4 4 -.5-1. -1.5..1 3 1.6.8 1. 1199: EDA H-mode, no programmed z-modulation 1198: EDA H-mode, programmed Hz z-modulation time (s) 1. 1.4 Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 11
Underlying edge transport in jogged EDA H-mode not substantially changed QCM Time (s) Time (s) Quasi-coherent mode (f c ~7kHz) persists in jogged EDA, though with modulated lab-frame frequency and interrupted by largest edge events QCM seems to continue to govern edge transport at the same level as without jogs Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 1
Can jogs induce ELMs in otherwise ELMfree H-modes? L-mode density lowered to obtain ELM-free target Z-modulation now applied starting at H-mode transition, from.7 1.4s Multiple backtransitions with gradual density reduction until ~1.s cryopump was utilized in the ELMfree case At lowest densities, back-transitions gradually become replaced with shorter ELM-like events -.5-1. -1.5..1 6 4.6.8 1. Note: L-mode P rad in the ELM-free target is higher than the EDA target, due to lack of recent boronization P rad (MW) W mhd (MW) D z P ICRF (MW) n e (1 m -3 α (a.u.) magx (cm) ) 4 4 1131: ELM-free H-mode with programmed Hz z-modulation 1198: EDA H-mode, programmed Hz z-modulation time (s) 1. 1.4 Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 13
Edge fluctuation levels are low between events in ELM-free discharges Edge Te (kev)??????? L-H threshold Numerous H-L H L back- transitions observed early in discharge (unsurprising) These yield several ms periods of L-mode trasnport Also shorter events (<1ms) which are either very short L- phases or ELMs Time (s) Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 14
Most relaxation events occur when plasma z is at or approaching local maximum P rad (MW) W mhd (MW) D z P ICRF (MW) n e (1 m -3 α (a.u.) magx (cm) ) 1.6 1. 4 -.5-1. -1.5.1.1.8 4 1131 Dashed lines: events corresponding to max. z magx Dotted lines: other events.85.9.95 1. 1.5 1.1 1.15 time (s) Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 15
Summary of vertical jogging These preliminary results represent efforts to scope out H- mode regimes where vertical jogs might produce beneficial effects EDA H-modes generally resilient to controlled vertical oscillations ELM-free discharges with cryopumping show some promise of triggering edge relaxation events, controlling impurity accumulation Future work Induce ELMs in ELM-free discharges in a well-boronized machine Systematically test edge event triggering dependence on jog rate, amplitude, phase Characterize pedestal profiles, equilibrium characteristics leading to ELM triggering; perform stability analysis Attempt jogs to trigger/pace ELMs in I-modes Hughes, ELMs on C-Mod C-Mod/NSTX Pedestal Workshop, Sept. 7 8, 1 16