• Bayesian Data Analysis
• Electric Propulsion
• Hall Thruster
• Plasma Diagnostics
• Plasma Plume Modelling

Nowadays, Hall thrusters represent one of the most attractive technology for space propulsion,
mainly due to their high efficiency and operational versatility. For these reasons, ever since their
first development in the 1960s, they have been extensively experimentally characterized, both to
assess their performance and gain deeper insights into their working principles.
Despite the considerable heritage in the investigation methodologies, many difficulties still remain
in the data analysis of the plasma diagnostics, given the many sources of error in the plasma
measurements and the accuracy of the physical models involved. This thesis aims to develop an
integrated data analysis method based on Bayesian inference, in order to effectively measure the
physical properties of the plasma plume of electric propulsion devices.
Plume measurements were taken during the experimental characterization of SITAEL’s 5kWclass
Hall thruster fed with xenon, using a fast-diving triple Langmuir probe in the near-field
region and a set of Faraday probes in the far-field region. The analysis started by individually
post processing the data collected from each diagnostic. In parallel, a semi-analytical self-similar
plume expansion model was developed to couple the plasma properties measured in the near-field
with those measured in the far-field. Eventually, the self-similar plume model was employed for
the data fusion of the diagnostics measurements, where a Bayesian inference of the measured
plasma properties was carried out using a Markov Chain Monte Carlo sampling technique.
The results, despite the simplicity of the semi-analytical plume model and the complexity of the
plasma physics involved, show good agreement and coherence with the indipendent measurements,
demonstrating the strength of the integrated data analysis approach and establishing this
method as a viable alternative to the traditional ones.