Predicting Performance of Hall Effect Ion Source Using Machine Learning

Abstract

Hall Effect Ion Source

In article number 2400555, Wonho Choe and co-workers introduce a machine learning-based approach to accurately predicting the performance of Hall thrusters, a critical technology for space propulsion and industrial ion beam sources. By utilizing an ensemble of neural networks trained on 18,000 simulation datasets validated by experiments, the model achieves high accuracy in thrust and discharge current predictions, enabling the rapid development of optimized, high-efficiency thrusters with shorter design cycles.