Experimental investigation of marfes and the density limit in the ASDEX upgrade

Abstract

In the ASDEX Upgrade, X-point marfe formation and its behaviour up to the density limit is investigated in gas-fuelled ohmically-heated single-null discharges over a wide range of pameters: Ip = 0.6-1.2 MA, Bw = I .35-2.4 T, a plasma elongation of 1.6 and 2 , ~ c 2. The standard-ion VB drifl is directed towards the X-point. At medium electron densities inevitably a marfe develops in the vicinity of the active X-point The marfe formation is consistent with a model of thermal instabilities in the radiating edge plasma. Moreover, stable steady-state operation is demonstrated with marfes which can extend significmtly into the bulk plasma. The density limit is always connected with quick m d e expansion and movement followed by mode-locking leading to a major disruption. The limit scales linwly with In md is in good agreement with the Greenwald density limit scaling. The resulting experimental Hugill limit is E,Rqv5 / B E = 2.8 x IO2" m-2 T I . Reversal of the ion VB drift direction away from the target has a detrimental effect an the density limit.

Density limits in toroidal plasmas

“…ASDEX, ASDEX-Upgrade, FT, JT-60, DIII-D, and JET have reported the appearance of MARFEs at densities just below the disruptive limit [57,71,9,[72][73][74]. However, while the formation of MARFEs is generally well parameterized by the density normalized to the density limit, n/n G , many experiments report MARFE formation at densities well below the limit.…”

“…The limit was associated instead with the movement and expansion of the MARFE. In this device, the threshold for X-point MARFE formation was linear with power [73], the plasma could always be made to detach if the power were low enough. ASDEX-upgrade also reported the "completely detached H-mode" (CDH) regime which began at about 0.65n G [88].…”

“…In these experiments (and others as discussed above), core plasma densities well above the detachment threshold have been reported. Typically the detachment threshold density scales as a much stronger function of power than the density limit [73,239].…”

Particle Simulation of the Neoclassical Plasmas

“…This reinitialization does not create any physical current in the simulation domain and, being more uniform (in phase space) than the loss process, it simulates well the replacement of charge lost through the separatrix. This method has been successfully applied recently, with the outer boundary at the separatrix, ρ R = ρ s , to calculate [12] the orbit loss, as well as the related source of poloidal rotation and its dynamics for L-H transition conditions in ASDEX Upgrade [15].…”

“…The initialization schemes and the algorithms for resolving the fluxes and advancing the electric field, as well as the stability issues of the method, are discussed. Section IV presents both global and annular simulation results for Physical Tokamak-2 FT-2 [14] and Axially Symmetric Divertor Experiment ASDEX Upgrade [15] tokamaks showing both spatial and temporal numerical convergence for the calculated radial electric field at transport barrier transition conditions. The conclusions are given in Section V.…”