Gradient-based calibration of the anomalous electron scattering frequency in a Hall effect thruster simulation

When modeling cross-field plasma discharges, particularly with a drift–diffusion assumption, semi-empirical anomalous electron transport is typically needed for the model results to match experimental observations. For Hall effect thrusters (HETs), the state-of-the-art approach has relied on manual calibration of the anomalous electron scattering frequency profile based on experimental data with the use of Bohm-like cross-field diffusion terms scaled by the electron cyclotron frequency. In this study, four gradient-based optimization algorithms (Adagrad, RMSProp, Adadelta, and Adam) are used to automatically calibrate the anomalous electron scattering frequency in a one-dimensional quasineutral drift–diffusion model of a SPT-100 HET to reproduce ion bulk velocity measurements obtained from laser-induced fluorescence along the discharge channel centerline. All algorithms converge to similar solutions, which agree well with the experimental data. It is shown that the optimization algorithms can further be used to quantify the uncertainties in the spatial variation of the anomalous scattering frequency and unmeasured plasma properties.