Session JP20: Poster Session: Magnetic Confinement: Edge & Pedestal Physics (2:00pm - 5:00pm)

Pedestal particle transport in high opacity regimes on DIII-D and C-Mod

Future fusion reactors will operate at high neutral opacity, which will strongly limit the influence of neutrals to penetrate inside the separatrix, whereas in current day devices part of the pedestal density steep gradient can be attributed to direct ionization of neutrals. We will show experimentally that operation at high neutral opacity is compatible with a steep density pedestal structure. Experiments on DIII-D and C-Mod scanning opacity ($n \times a$, where $n$ is the electron density and $a$ is the minor radius) span a range from $1.5 \times 10^{19}$ to $7 \times 10^{19} \; m^{-2}$. The highest reached opacity values in C-Mod are only a factor 2 smaller than those expected on ITER and at these opacity values we did not find a degradation of the pedestal density structure. We find that the neutral concentration as measured by the from the filterscopes in DIII-D and modeled with SOLPS-ITER decreases inside the pedestal structure with increasing opacity and increasing pedestal density. Transport in DIII-D based on the ratio of $D_{\Chi_i}/D_{\Chi_e}$ is dominated by ETG and MTM turbulence in between ELMs, neither of which contribute much to particle transport. In C-Mod edge transport is characterized by a quasi-coherent enhanced $D_{\alpha}$ mode and no ELMs are present.   

Edge Ionization Rate and Neutral Density from Main Chamber Measurements of Lyman alpha on DIII-D

A new one-dimensional absolutely calibrated Lyman-alpha (Ly-$\alpha$) camera has been implemented to study neutral hydrogenic species at the boundary of DIII-D. The Ly-$\alpha$ camera provides two radial profiles consisting of 20 lines of sight covering 22 cm about the separatrix below the midplane on the high field side (HFS) and low field side (LFS) of the tokamak. Each channel provides a line integrated measurement of the Ly-$\alpha$ (121 nm) brightness, which can be combined with electron temperature and density measurements to determine local time resolved neutral densities and ionization rates. Recent experiments commissioned the diagnostic and provided first results: measured Ly-$\alpha$ brightness profiles track rigid plasma shifts, peak Ly-$\alpha$ brightness scales with gas puff rates and shows similar dynamic behavior as midplane Balmer-alpha measurements. Initial measurements of neutral density found magnitudes of $10^{17}$ m$^{-3}$ on the LFS consistent with modeling and measurements on other machines. The diagnostic has measured a robust in-out asymmetry in Ly-$\alpha$ brightness with an order of magnitude larger signals on the HFS. The diagnostic will continue regular operation providing quantitative studies of neutral transport of the plasma edge on DIII-D.   

DIVIMP-R code tests of an analytic model of impurity leakage from the divertor and accumulation in the main scrape-off layer

Edge codes like SOLPS find spatial distributions of impurity density, nz, which are quite non-uniform. Often nz peaks strongly near the targets and on/near the separatrix at the outside midplane, OMP. High nz of low-Z near the targets is desirable for efficient edge radiative dissipation. High nz of high-Z near the OMP is undesirable for confined plasma performance. SOLPS etc. have grown so sophisticated that they can benefit from interpretation in terms of simple conceptual frameworks. A simple analytic 1D impurity fluid model, 1DImpFM, has been developed for the transport along open field lines of impurity ions in a specified fuel-plasma background [1; Stangeby {\&} Moulton, 2020, NF]. An nz peak at/near the OMP occurs at an impurity stagnation point, ISP, where v\textunderscore z$_{\mathrm{\vert \vert }}$ and flux\textunderscore z$_{\mathrm{\vert \vert }}$ -\textgreater 0. 1DImpFM predicts that nz-peaking will only occur if, in approaching the ISP v\textunderscore z$_{\mathrm{\vert \vert }}$ starts to decrease before flux\textunderscore z\textbar \textbar does. It is hypothesized in [1] that the latter happens because of a natural feedback effect between parallel and cross-field impurity fluxes. The 2D DIVIMP-R (Rectilinear) code has been developed to test this hypothesis. The code finds, as hypothesized, that v\textunderscore z$_{\mathrm{\vert \vert \thinspace }}$drops before the flux\textunderscore z$_{\mathrm{\vert \vert }}^{\mathrm{\thinspace }}$ does, resulting in an increase in nz. Additional results are presented.   

Reduced-model turbulence (nSOLT) simulations comparing three fueling scenarios: neutral puffing (SOL), divertor recycling (edge) and injection from the core

The 2D scrape-off-layer turbulence code (nSOLT) includes 1D Boltzmann neutral-plasma interactions [1], a model of divertor recycling (introduced here), and a fixed source of plasma concentrated at the core-side boundary. 1) Neutral injection in the far-SOL is accomplished by specifying the density of Franck-Condon distributed neutrals, n$_{\mathrm{puff}}$, streaming in from the boundary. 2) Divertor recycling is modeled by injecting a fraction of the particle parallel flux in the SOL back into the edge region as a source of plasma, while 3) the fixed source fuels the edge plasma from the core-side boundary, as in neutral beam injection. For machine parameters (B, R, L$_{\mathrm{//}})$ anticipated at MAST-U, and for a deuterium plasma, turbulent equilibria are obtained that share the same plasma fueling rate for each of the three fueling methods, with only one of the sources ``on'' in each case. Equilibrium plasma and neutral (deuterium) profiles, fueling efficiencies, SOL transparencies, heat flux widths and confinement times are compared. Skewness, cross-phase and spectral measurements of the turbulent fluctuations are presented. [1] D.A. Russell, J.R. Myra and D.P. Stotler, Phys. Plasmas \textbf{26}, 022304 (2019).   

Study of neoclassical transport in resonant magnetic perturbation edge localized mode suppressed plasmas in DIII-D

Resonant magnetic perturbations (RMPs) are applied to mitigate or suppress the edge localized modes (ELMs) in H-mode plasmas, leading to a decrease in the plasma density, which is often referred to as density pump-out. It is essential to understand the role of neoclassical transport in density pump-out. In this study, drift kinetic code NEO with the enhanced capability to handle non-axisymmetric flux geometry is used to evaluate the neoclassical transport properties in DIII-D plasmas. The magnetic field given as an input to the NEO code is calculated using extended magnetohydrodynamic code M3D-C1 and includes the linear two-fluid plasma response. The study performed here indicates an increase in neoclassical particle and energy fluxes during the ELM suppressed phase in DIII-D plasmas, which correlates with the density pump-out phenomenon and is on the same order to experimentally observed flux values. Additionally, investigation on the effect of neoclassical torque viscosity (NTV) on density pump-out and its possible role in the braking of rotation profile due to RMP application, has been carried out. A sensitivity study to examine the impact of the different values of RMP on NTV gives a qualitative agreement to the theory that predicts a rise in NTV with the increase in RMP.   

Global-local gyrokinetic simulations of the tokamak pedestal

We develop a novel approach to gyrokinetics where multiple flux-tube simulations are coupled together in a way that consistently incorporates global profile variation while retaining spectral accuracy. By doing so, the need for Dirichlet boundary conditions, where fluctuations are nullified at the simulation boundaries, is obviated. These conditions, which are typically employed in global gyrokinetic simulation, prevent convergence to the local periodic limit unless large simulation domains are utilized. Thus, our method of global-local gyrokinetics is appropriate for simulations of the pedestal region where the generation of intrinsic momentum is expected to commence and the details of boundary physics are important. Preliminary results from simulations with equilibrium flow shear using this approach are compared to simulations using conventional global methods and to local flux-tube simulations with wavenumber-remapped flow shear. Progress is also reported on implementing profile variation in both the plasma pressure and magnetic geometry.   

Reduction of Blob-Filament Radial Propagation by Parallel Variation of Flow Speeds

Data from an XGC1 gyrokinetic simulation is analyzed to understand the three-dimensional spatial structure and the radial propagation of blob-filaments generated by quasi-steady turbulence in the tokamak edge pedestal and scrape-off layer plasma. Spontaneous toroidal flow speeds vary in the poloidal direction and shear the filaments within a flux surface resulting in a structure that varies in the parallel direction. This parallel structure allows the curvature and grad-B induced polarization charge density to be shorted out via parallel electron motion. As a result, it is found that the blob-filament radial velocity is significantly reduced from estimates which neglect parallel electron kinetics, broadly consistent with experimental observations. Conditions for when this charge shorting effect tends to dominate blob dynamics are derived and compared with the simulation.   

Interpretation and application of extracted features of radiative collapse in Large Helical Device with sparse modeling

The features of radiative collapse have been extracted from high-density plasma experiments in Large Helical Device (LHD) with a sparse modeling technique. The extracted features have been used to explore the underlying physics of the radiative collapse and to develop a machine learning predictor of the collapse. The Sudo scaling is well known as a density limit scaling in stellarator-heliotron plasma. It includes only heating power density and magnetic field but it is thought that more operational conditions than those in the Sudo scaling are involved in the physics of radiative collapse. As extracted features, light impurities’ emission and electron temperature are relevant parameters to predict the occurrence of radiative collapse. Therefore, impurity radiation at the plasma edge especially outside the LCFS has been investigated. Also, the operational limit and the collapse predictor have been developed based on the extracted features and over 85\% of collapse discharges in LHD have been predicted successfully at least 30 ms before occurrence.   

Electromagnetic effect on divertor heat-load width in the total-f edge gyrokinetic code XGC

Predictive electrostatic simulation of the divertor heat-load width $\lambda_q$ in the edge gyrokinetic code XGC has been validated against experimental measurement and Eich formula in the three major US tokamaks [1], including a highest field C-Mod result at a later time. The same code, however, predicted, that the full-current (15MA) FPO ITER will have its $\lambda_q$ that is 6X wider than the most optimistic Eich-formula prediction, while agreeing with the Eich value for the pre-fusion power operation at 5MA. XGC also predicted that the highest current NSTX-U plasma will have $\lambda_q$ that is about 2X wider than the Eich value. It has been found that the arousal of the weakly-collisional TEM across the magnetic separatrix is responsible for the $\lambda_q$ enhancement. Here, we report on an ongoing study with XGC of how EM turbulence affects the divertor heat-load width. XGC has two EM schemes installed: explicit [2] and implicit [3]. Present study is based on the explicit scheme that uses the pullback scheme and the control variate method.\\ ~[1] C.S. Chang, S. Ku, A. Loarte et al., Nucl. Fusion 57, 116023 (2017)\\ ~[2] A. Mishchenko, A. Koenies, R. Kleiber, M. Cole, Phys. Plasmas 21, 092110 (2014)\\ ~[3] G. Chen and L. Chacón, Comp. Phys. Com. 197, 73-87 (2015)   

Toroidal rotation at low torque with NTV

Future devices such as ITER will only be able to apply relatively weak NBI torque, but must avoid dangerous low-rotation regimes that can lead to locking and disruption. In such parameter regimes, the nondiffusive part of the turbulent momentum flux tends to drive tokamak edge rotation co-current. However, ITER also plans to apply three-dimensional magnetic perturbations (MPs). These MPs tend to apply a net counter-current NTV torque to the plasma, often comparable in magnitude to the effective co-current "intrinsic torque" in present-day devices. It is difficult to predict rotation in this regime, due to the near-cancellation of two complicated torques, with different dependence on plasma parameters. As a first step towards the development of such a predictive capability, we combine the modulated-transport model for intrinsic rotation with a simplified global model for the NTV torque. For the NTV torque, we take account of local resonances that can shift in radial position as plasma parameters change. Due to the non-monotonic nature of the NTV torque, rolling over to become weak for strong-enough co-current rotation, the joint turbulent-NTV momentum balance allows the possibility of an edge rotation bifurcation even without island penetration.   

Kinetic Ballooning Modes with bootstrap current in High-beta Pedestal Plasma

We present the Linear and nonlinear simulation results of kinetic ballooning modes (KBM) in edge pedestal plasmas based on the gyro-Landau-fluid (GLF) model under the BOUT$++$ framework. The linear growth rate spectrum of KBMs shows the stabilizing effect of the bootstrap current, both the growth rate and the unstable region of the instabilities decrease, and the spectra shift to the low toroidal mode number. The toroidal resonance can drive the KBM unstable under the ideal peeling-ballooning mode threshold and the Landau damping can destabilize the KBM for low-$\beta $ and have stabilizing effect on the KBM for high-$\beta $ In the nonlinear simulation, the energy loss with bootstrap current is higher than those without bootstrap current, because with bootstrap current local magnetic shear decreases and the fluctuation level increases, resulting in a large energy flux. The saturated fluctuation level increases with $\beta $, while the fraction of the bootstrap current has weak impact on the saturated level. The turbulence in low $\beta $ is mainly the ballooning dominant turbulence while the turbulence in high $\beta $ is mainly the peeling dominant.