Dimensionless size scaling of intrinsic rotation in DIII-D

Abstract

A dimensionless empirical scaling for intrinsic toroidal rotation is given; M A~βN ρ *, where M A is the toroidal velocity divided by the Alfvén velocity, β N the usual normalized β value, and ρ* is the ion gyroradius divided by the minor radius. This scaling describes well experimental data from DIII-D, and also some published data from C-Mod and JET.The velocity used in this scaling is in an outer location in minor radius, outside of the interior core and inside of the large gradient edge region in H-mode conditions. This scaling establishes the basic magnitude of the intrinsic toroidal rotation and its relation to the rich variety of rotation profiles that can be realized for intrinsic conditions is discussed. This scaling has some similarities to existing dimensioned scalings, both the (2012)]. These relationships are described.

Predicting the rotation profile in ITER

“…Although the indications of this experiment are that the fast-ions may have contributed to the discrepancy between the dimensionless parameter scan and database results, a direct comparison between the ρ * scaling of intrinsic Mach number at the top of the pedestal and database studies shows some discrepancy still remains. While [23,24] found the Alfvén Mach number to scale as ρ 1 * , the present results shows a scaling of ρ 0 * , which agrees with the older result of [14]. Note that in a dimensionless parameter scan, Mach number and Alfvén Mach number have the same scaling to the extent β has been held constant.…”

“…As explained in that work, these results were unexpected based on concepts of residual stress driven intrinsic rotation and results from database studies of intrinsic rotation, which found the intrinsic rotation at the top of the pedestal to scale as ρ 0 * [14] or ρ 1 * [23,24]. The results of [25,26] also show a scaling of ρ 1 * if L φ ∼ a, where L φ is the radial scale length of electrostatic fluctuation intensity.…”

“…These two inescapable uncertainties are similar. While a single factor or ρ * when projecting from DIII-D to ITER amounts to a factor of ≈1/4, measurements of Mach numbers that scalings derived from databases predict to be the same can in fact differ by a factor of ≈2 [23,24,26].…”

“…The rotation predictions of this work are made with a combination of models for the co-current intrinsic rotation feature at the top of the pedestal and modeling of the neutral beam torque and momentum flux. The pedestal top intrinsic rotation can be predicted from a database of measurements in multiple tokamaks that can be described by a small number of parameters [14,23]. It is also possible to form reduced physics models that can predict the rotation and then test the results against a reasonably varied data set [24][25][26].…”

Effects of collisional ion orbit loss on tokamak radial electric field and toroidal rotation in an L-mode plasma

“…Ion orbit loss has also been used to model the toroidal angular momentum (TAM) transport and toroidal rotation at the outboard of diverted tokamaks [6,15,16,17,18,19,20,21,22]. Consider ion orbits at the outboard midplane; ions with parallel velocity in the direction of toroidal plasma current (denoted by "co-current") drift radially inwards, while ions with oppositely directed parallel velocity (denoted by "counter-current") drift radially outwards (figure 1).…”