Transport and confinement in the Mega Ampère Spherical Tokamak (MAST) plasma

Autor: Akers, R J, Ahn, J W, Antar, G Y, Appel, L C, Applegate, D, Brickley, C, Bunting, C, Carolan, P G, Challis, C D, Conway, N J, Counsell, G F, Dendy, R O, Dudson, B, Field, A R, Kirk, A, Lloyd, B, Meyer, H F, Morris, A W, Patel, A, Roach, C M, Rohzansky, V, Sykes, A, Taylor, D, Tournianski, M R, Valovic˘, M, Wilson, H R, Axon, K B, Buttery, R J, Ciric, D, Cunningham, G, Dowling, J, Dunstan, M R, Gee, S J, Gryaznevich, M P, Helander, P, Keeling, D L, Knight, P J, Lott, F, Loughlin, M J, Manhood, S J, Martin, R, McArdle, G J, Pric, M N
Zdroj: Plasma Physics and Controlled Fusion; December 1, 2003, Vol. 45 Issue: 12 pA175-A204, 30p
Abstrakt: A combination of recently installed state-of-the-art imaging and profile diagnostics, together with established plasma simulation codes, are providing for the first time on Mega Ampère Spherical Tokamak (MAST) the tools required for studying confinement and transport, from the core through to the plasma edge and scrape-off-layer (SOL). The H-mode edge transport barrier is now routinely turned on and off using a combination of poloidally localized fuelling and fine balancing of the X-points. Theory, supported by experiment, indicates that the edge radial electric field and toroidal flow velocity (thought to play an important role in H-mode access) are largest if gas fuelling is concentrated at the inboard side. H-mode plasmas show predominantly type III ELM characteristics, with confinement HH factor (w.r.t. scaling law IPB98[y, 2]) around ~1.0. Combining MAST H-mode data with the International Tokamak Physics Activities (ITPA) analyses, results in an L–H power threshold scaling proportional to plasma surface area (rather than PLH ~ R2). In addition, MAST favours an inverse aspect ratio scaling PLH ~ ε0.5. Similarly, the introduction of type III ELMing H-mode data to the pedestal energy regression analysis introduces a scaling Wped ~ ε−2.13 and modifies the exponents on R, BT and κ. Preliminary TRANSP simulations indicate that ion and electron thermal diffusivities in ELMing H-mode approach the ion-neoclassical level in the half-radius region of the plasma with momentum diffusivity a few times lower. Linear flux-tube ITG and ETG microstability calculations using GS2 offer explanations for the near-neoclassical ion diffusivity and significantly anomalous electron diffusivity seen on MAST. To complement the baseline quasi-steady-state H-mode, newly developed advanced regimes are being explored. In particular, ‘broad’ internal transport barriers (ITBs) have been formed using techniques developed at conventional aspect ratio. Electron and ion energy diffusivities are reduced towards the ion-neoclassical level in the ITB region of both co- and counter-injection NBI heated plasmas, with momentum diffusivity up to 10 times lower. Moving out to the edge and SOL, OSM2/EIRENE modelling is being used to extract edge perpendicular particle and heat diffusivities, results being consistent with the ballooning nature of power-flow seen in L-mode and reduction in outboard turbulence seen in ELM-free and inter-ELM H-mode. Modelling of parallel SOL transport requires the inclusion of the mirror force (~10 times higher in MAST than at the conventional aspect ratio) and B2SOLPS5.0 simulations show the edge electric field to be well modelled by neoclassical theory. Transient edge transport phenomena are being studied in detail using a variety of techniques (e.g. probability density function (PDF) and power spectrum analysis, differencing and rescaling methods). Intermittent transport is associated with a radial efflux at up to a tenth of the sound speed and up to 30 cm from the separatrix. Arguably, the most dramatic edge events seen in the plasma periphery are the ELMs. Recent results using fast, high-resolution visible imaging confirm the hypothesis that ELMs have both poloidal and toroidal structures (n ~ 10 at q = 4), consistent with recent theories of the non-linear evolution of ballooning modes.
Databáze: Supplemental Index