Determine an effective vessel-current prior or constraint under severe sensor dropout

Determine which modification most effectively improves the estimate of the 104 vessel filament currents, and thereby reduces the poloidal-flux error of the diffusion reconstruction at high magnetic-sensor dropout, among a smaller vessel-current prior standard deviation, a reduced set of vessel eigenmodes, and additional constraints analogous to those used in LIUQE.

Background

The diffusion reconstruction samples the toroidal current density while marginalizing over the unmeasured vacuum-vessel currents. It subsequently estimates the vessel currents from the surviving magnetic channels to derive the poloidal flux. Although the reconstructed current-density error remains nearly constant as dropout increases, the derived-flux error grows because the 104 vessel filament currents become difficult to estimate from few sensors.

The paper identifies several possible remedies—a smaller prior standard deviation for the vessel currents, a reduced representation using a few vessel eigenmodes, and additional constraints modeled on LIUQE—but does not determine which remedy is effective. This is therefore an explicitly unresolved methodological question concerning accurate flux reconstruction under substantial sensor loss.