Collisionality and magnetic geometry effects on tokamak edge turbulent transport. I. A two-region model with application to blobs

Abstract

A two-region model is proposed to study the effect of collisionality and magnetic geometry on electrostatic turbulence and on the propagation of filamentary coherent structures (blobs) in the edge and scrape-off layer. The model invokes coupled vorticity and continuity equations in two different spatial regions along the magnetic field, taking into account the effect of magnetic field fanning and shear, e.g., near magnetic X-points. A linear dispersion relation for unstable modes illustrates the physics of mode disconnection (ballooning) along the magnetic field and its dependence on collisionality and wave number (scale size). Employing an invariant scaling analysis, dimensionless parameters for the nonlinear model are developed and used to describe the regimes of the system. A blob correspondence rule is postulated to relate the linear mode growth rates and regimes to the convective velocity of blobs. Nonlinear numerical simulations of blob convection show good agreement with a blob dispersion relation derived from the correspondence rule. It is found that collisionality increases the convective velocity. The convective velocity also depends on blob scale size, with either positive or negative exponent, depending on the collisionality regime. Finally, the dimensionless scaling analysis is employed to obtain bounds on the convective velocity suitable for experimental tests.

Far scrape-off layer particle and heat fluxes in high density – High power scenarios

“…Previous investigations in L-mode revealed that the density wing formation is related to a change from a conductive to a convective transport regime [3] where also the outer divertor starts to detach and the turbulence characteristic is changing [7,8]. This agrees well with the model by Myra [9] and Krasheninnikov [10]. This lead to the expectation that a similar transport regime transition will occur in high-density H-modes causing a particle and heat cross-field transport in the far SOL above the values observed in low to medium density H-modes [11].…”

“…Investigations in L-mode discharges at ASDEX Upgrade [7] lead to the interpretation that an increase of the effective collisionality Λ = L c v ei Ω i /(c s Ω e ) causes a change in the filamentary transport from conduction to convection dominated for Λ > 1 -caused by different closure conditions for the parallel current [10]. This provokes a high cross-field transport at the outer midplane [9]. Here, L c is the connection length, v ei the electron-ion collision frequency, c s the ion sound speed and Ω i , Ω e the ion and electron gyrofrequency.…”

Blob properties in full-turbulence simulations of the TCV scrape-off layer

“…The ion temperature at the LCFS is assumed to be T T i e 0 0 = , since no ion temperature measurement was available for that discharge. Resistivity can affect substantially blobs dynamics by electrically disconnecting from the target sheaths and subsequently increasing their convective velocity, as predicted in [25] and experimentally measured in [4,27,28]. The influence of resistivity on blob dynamics has been investigated with seeded simulations in [17,29].…”

Convective transport by intermittent blob-filaments: Comparison of theory and experiment

“…Most simulation work on seeded blob properties has been done with 2D codes (in the x-y plane), although seeded blobs have also been studied using 3D models. 44,51 In addition to seeded blob simulations, turbulence codes have been used to study the nonlinear saturation processes for the turbulence and to infer self-consistently the blob generation rate and statistics. 40,50,53,62,74,81,88 The plasma profiles due to the turbulent (blob) convection have also been compared with experiments.…”

“…Seeded blob simulations have been used to compute the blob velocity in more complicated situations. 37,41,49,51,67,69,70 (A seeded blob is an isolated density peak used to initialize a simulation run; its evolution in time and space is calculated by the simulation code.) Various fluid simulation models have been employed, differing in details of the precise form of the vorticity equation used (e.g., Boussinesq approximation and particular parallel current closures), and whether temperature is evolved as a dynamical variable or not.…”