Probabalistic evaluation of closure models for the Hall thruster anomalous collision frequency

The development of fully predictive simulations of crossed-field plasma discharges such as Hall effect thrusters is a longstanding goal of the low temperature propulsion community. However, in many cases, there are aspects of the underlying physics such as anomalous cross-field transport that are not sufficiently understood to produce fully predictive models. With this limitation in mind, we apply Bayesian inference to a new two-equation turbulence model for the anomalous transport in a Hall effect thruster. Using the output of validated simulations as surrogate data, we calibrate this model and obtain a posterior distribution of the model coefficients. We then sample from its coefficient distribution and run a high-fidelity multi-fluid Hall discharge simulation at each sampled set of model coefficients. We simulate the H9 magnetically shielded Hall thruster, which was not part of the training dataset. We repeat this at multiple operating conditions, obtaining probabalistic predictions of key thruster performance parameters, including thrust, discharge current, specific impulse and anode efficiency at each. These are then compared to data.