A reliable data-driven model for Ablative Pulsed Plasma Thruster

Ablative Pulsed Plasma Thrusters (APPTs) are high specific impulse electric space propulsion system, but a reliable model equivalent of the experimental model is still unavailable. In this paper, a reliable model is developed based on APPT experimental data by using Machine Learning (ML) ecosystem. The goals of this study are to justify the accuracy and reliability of the newly built APPT model with the existing experimental and simulation model. For four sets of operating conditions, 600 experimental and simulation test operations are done. The experimental voltages and currents are measured with a high-voltage probe and a Rogowski coil, respectively. The simulation voltages and currents are gathered by running the respective simulation program. Comparison results show that the newly built APPT model has better accuracy and reliability than the simulated APPT model as compared to real APPT used in the experiment. This data-driven approach provides a novel way of designing a reliable alternative model of physical APPTs.