Testing Gyrokinetics on C-Mod and NSTX

Abstract

Quantitative benchmarks of computational physics codes against experiment are essential for the credible application of such codes. Fluctuation measurements can provide necessary critical tests of nonlinear gyrokinetic simulations, but such require extraordinary computational resources. Linear microstability calculations with the GS2 [1] gyrokinetic code have been carried out for tokamak and ST experiments which exhibit internal transport barriers (ITB) and good plasma confinement. Qualitative correlation is found for improved confinement before and during ITB plasmas on Alcator C-Mod [2] and NSTX [3] with weaker long wavelength microinstabilities in the plasma core regions. Mixing length transport models are discussed. The NSTX L-mode is found to be near marginal stability for kinetic ballooning modes. Fully electromagnetic, linear, gyrokinetic calculations of the Alcator C-Mod ITB during off-axis rf heating, following four plasma species and including the complete electron response show ITG/TEM microturbulence is suppressed in the plasma core and in the barrier region before barrier formation, without recourse to the usual requirements of velocity shear or reversed magnetic shear [4-5]. No strongly growing long or short wavelength drift modes are found in the plasma core but strong ITG/TEM and ETG drift wave turbulence is found outside the barrier region. Linear microstability analysis is qualitatively consistent with the experimental transport analysis, showing low transport inside and high transport outside the ITB region before barrier formation, without consideration of ExB shear stabilization. Calculations of the ITG mode instability threshold and the dependence of instability on temperature and density scaling length have been used to test the GS2 gyrokinetic model for ITG microturbulence on C-Mod [4-5]. Transport analysis at the ITB onset time in Hmode experiments [6] is in rough agreement with the linear mixing length model !/ . The mixing length estimates exceed experimental estimates of transport by less than a factor of two and are in agreement with experiment in the plasma core (see Table I). GS2