A comparison of RF heating calculated with the CQL3D fokker-planck solver and with a monte-Carlo code

Abstract

The bounce-average finite-difference Fokker-Planck code CQL3D [1] is an established tool for modeling of RF and NBI heating. Recently it has been augmented with a Finite-Orbit-Width (FOW) treatment of ions [2]. For validation purposes, and also for accessing the high-collisionality regime, and toroidal variation effects, a Monte-Carlo particle code RFMC has been developed that is based on the MCGO code [3]. Among other goals is an accurate calculation of ion and neutral particle losses to a realistic vacuum chamber. Ion orbits from a thermal initial distribution, and/or from an NBI source, are integrated in time including the effects of collisions, charge exchange events and RF kicks. The RF kicks are based on local quasilinear diffusion coefficients that include all components of RF electric field and all-orders terms for finite Larmor radius effects. The RF operator acts on both perpendicular and parallel component of particle momentum. A procedure for accumulation of particle loss distributions to the wall for test ions and neutrals, is included as a function of poloidal and toroidal distance along the wall, particle energy and the incident angle. As a verification study, a comparison is made between CQL3D and the RFMC code for scenarios of minority ICRH in C-Mod. The calculated profiles of RF power deposition are very similar in CQL3D and RFMC runs, the distribution functions also agree in the main features of non-Maxwellian tails, although some details are different. We also find good agreement of the NBI heating in NSTX, including FOW effects.