Recent work on the discrepancy between L-mode edge gyrokinetic transport predicted from GYRO and experimental L-mode transport has provided a new clue that might eventually solve the problem. The focus is on GYRO simulations of the extreme edge (r/a=0.9) of the well-studied lowρ* DIII-D discharge 101391 at 2790 msec. The predicted gyrokinetic low-k transport is five to ten times lower than experimental levels and appears to have a dissipative trapped electron scaling χ ~ T7/2/nB2 working against cold edge transport. A GYRO simulation run with drift-kinetic ions instead of gyrokinetic ions, but otherwise identical, recovers experimental transport levels. Discharge #101391 has strong gradients, high-q, high-shear, and very high collisionality in the edge region, and the close packing of singular surfaces requires expensive high radial resolution. The traditional remedy for yielding cold edge transport is to increase the collisionality tenfold to obtain the collisional two fluid or resistive-g mode scaling with χ ~ nT-1/2/a2B2. The standard collisional two fluid equations (used in edge codes like BOUT) are truncated moments of the drift kinetic equations. This suggests that the gyro-averaging approximation may be breaking down although the edge turbulence levels (ten to twenty times the core levels) do not seem to be large enough. Work is in progress to quantify the breakdown level.

Linear plasma response calculations with M3D-C1 show helical displacements of the edge temperature and density profiles of several centimeters along the core Thomson System and BES chords when I-coil fields are applied. The goal of this work is to understand edge plasma response to 3D fields, to validate plasma response models, and to gain confidence in modeling predictions regarding phenomena that are not easily observed experimentally, such as island formation. The calculated displacements are in good quantitative agreement with experimental measurements, both in amplitude and phase. The displacements are significantly larger than expected from vacuum modeling, which confirms the importance of the kink response of the plasma. These results were presented at the Sherwood fusion theory conference earlier this month.

In a study coupling non-equilibrium rare gas kinetics with the flow dynamics deep inside a penetrating gas jet used for disruption mitigation, it was found that the degree of ionization displays a bifurcation at a critical electron temperature. Because the core gas flow in the plasma forms a supersonic bubble, the heavy particles (neutrals and ions) remain cryogenically cold while the electrons receive the joule heating after discharge breakdown so they become the warm species. Given the high gas jet density, the electron recombination physics is dominated by the 3-body formation of diatomic molecular ions. These ionic dimers then undergo rapid dissociative recombination, annihilating electrons in the process. A sudden “condensation phase change”, or bifurcation with drastically lower ionization degree, occurs when the electron temperature drops below a critical point, which depends on the gas temperature, but is typically ~ 1-2 eV. As a result, the electrical resistance of the gas jet discharge increases more than ten times; the resistance affects jet heating, flow divergence and penetration, as well as current contraction. The discharge was found to be in the “abnormal glow” regime in the sense that the current-voltage characteristic is positive. The discharge-excited kinetic processes involved also provide a route to excited dimers (“excimers”) for generation of intense Vacuum Ultra-Violet (VUV) radiation from the gaseous medium of lasers and flash lamps. To detect an excimer emitting argon “gas-blob” at a wavelength of 126 nm in DIII-D one would need a special VUV spectrometer above the 121 nm Lyman-alpha emission line to avoid saturation.

Local linear simulations of DIII-D Alfvén eigenmodes in GYRO have reproduced predictions of the linear flux footprint from global simulations that require more than 100 times as many CPU cores to compute. A direct comparison of the predicted energetic neutral beam injected (NBI) ion and thermal ion fluxes under a simple saturation assumption shows that a sequence of local simulations performed at multiple points along the profile traces out the flux footprint created by the superposition of unstable global reversed shear Alfven eigenmodes (RSAEs) and toroidal Alfven eigenmods (TAE). Even though local simulations lack some of the physics needed to describe the RSAE, specifically shear reversal, the NBI ion flux prediction agrees with the global model to within 15%. Agreement for the TAE induced flux is even better, reproducing a series of thermal ion pinches that may be of interest in driving zonal flows.