Numerical modeling of lower hybrid current drive in fully non-inductive plasma start-up experiments on TST-2

Abstract

Noninductive plasma startup is a critical issue for spherical tokamaks since there is not enough room to provide neutron shielding for the center solenoid. Startup using lower hybrid (LH) waves has been studied on the TST2 spherical tokamak. Because of the low magnetic field of a spherical tokamak, the plasma density needs to be kept at a very low value during the plasma current rampup so that the plasma core remains accessible to the LH waves. However, we have found that higher density was required to sustain larger plasma current. The achievable plasma current was limited by the maximum operational toroidal field of TST2. The existence of an optimum density for LH current drive and its toroidal field dependence is explained through a numerical simulation based on a ray tracing code and a Fokker-Planck solver. In order to access higher density at the same magnetic field, a toplaunch antenna was recently installed in addition to the existing outboardlaunch antenna. Increase in the density limit was observed when the power was launched from the top antenna, consistently with the numerical predictions.

A simulation environment to simulate lower-hybrid-wave-driven plasmas efficiently

“…In the past, on the one hand the coupled GENRAY/CQL3D simulation has been proven useful [26,16] to evaluate LH-wave-driven plasmas in spherical Tokamaks. On the other hand, the MHD-kinetic particle code MEGA [22,21,29,23] has been successfully applied to investigate the interaction between energetic particles and the bulk plasma.…”

“…GENRAY and MEGA apply cylindrical coordinates (R, φ, z), where R, z are coordinates in the poloidal plane. Tsujii et al [26] showed that the differences in the ray patterns between GENRAY and a full-wave code such as TORLH [30] has been rather small and for that reason the computationally less expensive (but from a mathematical point of view less correct) option GENRAY has been chosen to be applied in the simulation environment. In GENRAY, the toroidal and poloidal power spectrum parameters of the antenna, i.e., the index of refraction, are chosen as ∆n φ = 2, n φ = 6 and ∆n θ = 1.6, n θ = 0.…”

“…The low resistivity used in the numerical configuration is justified within the intended period of simulation time which is just about ≈ 1 • 10 −5 s to mimic the low collisionality of the plasma, i.e., 1/Λ D ≈ 1•10 −5 ≪ 1, where Λ D denotes the Debeye sphere. The simulation parameters for GENRAY and CQL3D, i.e, bulk temperatures and densities, SOL parameters, antenna and geometrical configurations, are the same as those presented in [26]. Note that the driven current using GENRAY and CQL3D is about one order of order of magnitude larger than the experimentally measured value [26].…”

“…Furthermore, some plasma parameters were tuned artificially to obtain converged simulations. A more sophisticated simulation including density and temperature profiles measured by Thomson scattering diagnostic [24] was presented recently [26], where a ray-tracing solver GENRAY [17] and a bounced-averaged Fokker-Planck solver CQL3D [4] were coupled using EFIT [9] equilibrium field for peaked distributions. It was observed that the current carrying electrons do not penetrate to the core region for a fixed magnetic field geometry such that the force balance of the plasma was not solved self-consistently.…”

“…However, it was shown that for LH-driven, fully noninductive discharges in TST-2, the correlation between the plasma current and density could be explained by using coupled GENRAY and CQL3D simulations. Still, finite orbit width, radial diffusion through collisions, and RF quasilinear diffusion were not included in the CQL3D simulation such that the quantitative analysis of RF plasma remains challenging [26].…”