Comparison of filament properties in real-size GBS simulations and experiments of TCV-X21

TL;DR

This work addresses the validation of first-principles boundary-plasma turbulence by directly comparing 2-D filament properties from full-size GBS simulations with GPI measurements in TCV-X21. A synthetic GPI diagnostic processes GBS outputs to produce brightness patterns, enabling quantitative comparisons of filament sizes, velocities, and density/temperature fluctuations across the outboard midplane, X-point, and divertor leg. The study finds good agreement for filament poloidal and radial velocities, but systematic overestimation of filament sizes by ~2–3×, with filaments predominantly density-perturbation dominated; in the divertor leg, poloidal velocity tracks instantaneous ExB drift rather than the mean, indicating cross-field ExB transport dominates. The results highlight the need to include neutrals, relax the Boussinesq approximation, and adjust resistivity to achieve more predictive simulations, and demonstrate that the extended TCV-X21 dataset with GPI data provides a valuable validation resource.

Abstract

A direct quantitative comparison of Scrape-Off Layer (SOL) filament properties from fluid turbulence simulations using the GBS code and from experiments on the TCV tokamak is performed within the TCV-X21 validation case. This comparison is made possible by extending the open TCV-X21 dataset with 2D turbulence measurements obtained with Gas Puff Imaging (GPI), providing critical information on the size, velocity, and other key characteristics of turbulent filaments at the outboard midplane and in the divertor region. For the comparison, GBS simulations of TCV-X21 are analyzed using a dedicated synthetic GPI diagnostic that models the neutral helium-plasma interaction and emission processes and accounts for line-integration effects. Poloidal and radial filament velocities are found to be in good agreement between simulations and experiments, while the simulations overestimate the filament radial and poloidal sizes and underestimate the relative fluctuation levels. The simulations further indicate that filaments in the SOL are predominantly represented by density perturbations rather than temperature perturbations, consistent with previous assumptions in experimental analyses of cross-field turbulent transport from GPI data. The poloidal velocity direction of the filaments agrees with the time-averaged $\boldsymbol{E}\times\boldsymbol{B}$ direction at the outboard midplane and X-point region, but not in the divertor leg. Possible explanations are proposed and discussed, highlighting the influence of the instantaneous $\boldsymbol{E}\times\boldsymbol{B}$ velocity components in both poloidal and radial directions. This study provides new insights into turbulent filament behavior and contributes to guiding future efforts to improve first-principles simulations of the boundary plasma.

Figures (9)

• Figure 1: GPI diagnostics in TCV. The locations and fields of view of the midplane GPI and the Xpt GPI configurations are shown in the same poloidal cross-section, together with some typical 2-D plasma fluctuation snapshots. Examples of plasma equilibrium flux surfaces and separatrices are plotted. The plasma can be positioned for X-point GPI measurements around the X-point (red solid line) or around the divertor leg region (purple dashed line). Right image adopted from offedduGasPuffImaging2022a.
• Figure 2: Workflow of the synthetic GPI diagnostic.
• Figure 3: Illustration of the geometry and working principle of the synthetic GPI diagnostic. (a) The LCFS at multiple toroidal angles are shown in dark dotted lines. The magnetic field line at one poloidal plane is plotted as green vectors (also shown in c) ). Multiple toroidal planes of simulation results (purple planes) from the turbulence simulation (here GBS) are provided as an input. The GPI focal point (green circle) and the lines of sight (blue lines) are shown for both midplane GPI and X-point GPI. (b)The simulation domain of the neutral gas injection and reaction with the plasma, together with its spherical coordinate system. (c)A zoomed-in view of the field of view and lines of sight of the X-point GPI system. The magnetic field direction is plotted in green. For visualization, only part of the $96\times 128$ lines of sight are plotted.
• Figure 4: Qualitative comparison of the turbulence patterns at the different poloidal locations. Snapshots of the relative fluctuation of electron density $n_e$, electron temperature $T_e$, and brightness from simulations and experiments are plotted, respectively.
• Figure 5: The filament tracking results at three locations for both the simulation and experimental data. The typical trajectories of the filaments are plotted in a), f) and k), with experimental trajectories in blue and simulated trajectories in red. In the other subplots, the probability density function (PDF) of the filament sizes ($\delta_r$ and $\delta_\theta$) and velocities ($v_r$ and $v_\theta$) in the radial (subscript $r$) and poloidal directions (subscript $\theta$) are plotted. The simulation data is plotted in orange and the experimental data is plotted in blue. The average value and the standard deviation of the distributions are shown by markers and bars, respectively.