Measurements and modeling of the plasma response to electrode biasing in the HSX stellarator

The quasi-helically symmetric (QHS) configuration of the HSX stellarator is predicted to have strongly reduced parallel viscosity compared to conventional stellarators. To assess the physics of plasma flow damping, we use a fast switching biased electrode system to induce plasma flows and a set of Mach probes to measure the flows. Toroidal and poloidal arrays of absolutely calibrated Halpha detectors, coupled to the DEGAS neutral gas code, allow a determination of the neutral density and the ion-neutral friction component of the flow damping. Results show that the flow evolution at bias turn-on or turn-off involves two time-scales and directions. At bias turn-on, the floating potential changes on the electrode time scale (∼1musec.), while the decay of the floating potential at bias turn-off (∼40musec.) is much longer than the time for the power supply to break the electrode current. The fast component of the flow rises and decays on the floating potential time scale, while the slow component of the flow evolution is ∼10 times slower. The measured damping rates of the flows are reduced in the QHS configuration by a factor of approximately two compared to configurations with the quasi-symmetry intentionally broken. We model these experiments using neoclassical theory involving parallel viscosity and ion-neutral friction, including a numerical calculation of the Hamada basis vectors for the 3D HSX geometry. The relaxation phase at the end of the pulse is modeled assuming that the electrode current is terminated, based on the modeling of Coronado and Talmadge; this calculation predicts two time scales for the flow to decay. The plasma spin-up and electric field formation is modeled assuming that the electric field formation is the initiating event; this model predicts a third time scale, intermediate to the decay time scales. The results show that the radial conductivity in HSX is higher than the neoclassical predictions, and that the flows decay faster than predicted. The plasma flow spin-up time is roughly consistent with the time scale predicted by the modeling.