TY - JOUR
T1 - Data-physics coupled model for Ion propulsion system
AU - Ren, Siyuan
AU - Tang, Haibin
AU - Ren, Junxue
AU - Guo, Haixu
AU - Wang, Yibai
AU - Wen, Zheng
AU - Li, Zongliang
AU - Song, Fei
AU - Liu, Yizhuo
AU - Ding, Fenglin
N1 - Publisher Copyright:
© 2026 IAA. Published by Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
PY - 2026/9
Y1 - 2026/9
N2 - A novel grid acceleration current model for the ion propulsion system is designed, and a complete ion propulsion system model including power processing unit, gas feed system, discharge chamber, and grid system is established. This model enables time-series simulation of the performance parameters of the ion propulsion system. On this basis, combined with actual data and CNN-LSTM, a data-driven error compensation and prediction model is constructed, which can obtain results with higher fidelity and realize the prediction of future evolution trends. Finally, by integrating the system model and the CNN-LSTM error compensation model, a data-physics coupled model for the ion propulsion system is established, which can achieve high-precision prediction of key parameters such as beam current and acceleration current. In the calculation for the 1 kW operational condition of the LIPS-200 ion thruster with a 13 sccm Xe propellant flow rate, the system model calculates a discharge voltage of 36 V, a discharge current of 4 A, a beam current of 0.8 A, and a thrust of 37.7 mN, which are basically consistent with the actual measured results. On this basis, the data-physics coupled model further improves the result accuracy. In the 50-point prediction, the data-physics coupled model achieves RMSEs of approximately 1.6 × 10−3 A and 3.1 × 10−3 mA for the beam current and acceleration current, respectively. At the same time, it realizes the prediction of the variation trend of key parameters of the electric propulsion system, and shows relatively excellent performance compared with various prediction methods. The research results have certain engineering application value for the design optimization, on-orbit state prediction and long-term health management of the ion propulsion system, particularly exhibiting unique advantages in the ground-to-orbit equivalence of ion propulsion and the autonomous fault tolerance for deep space exploration missions.
AB - A novel grid acceleration current model for the ion propulsion system is designed, and a complete ion propulsion system model including power processing unit, gas feed system, discharge chamber, and grid system is established. This model enables time-series simulation of the performance parameters of the ion propulsion system. On this basis, combined with actual data and CNN-LSTM, a data-driven error compensation and prediction model is constructed, which can obtain results with higher fidelity and realize the prediction of future evolution trends. Finally, by integrating the system model and the CNN-LSTM error compensation model, a data-physics coupled model for the ion propulsion system is established, which can achieve high-precision prediction of key parameters such as beam current and acceleration current. In the calculation for the 1 kW operational condition of the LIPS-200 ion thruster with a 13 sccm Xe propellant flow rate, the system model calculates a discharge voltage of 36 V, a discharge current of 4 A, a beam current of 0.8 A, and a thrust of 37.7 mN, which are basically consistent with the actual measured results. On this basis, the data-physics coupled model further improves the result accuracy. In the 50-point prediction, the data-physics coupled model achieves RMSEs of approximately 1.6 × 10−3 A and 3.1 × 10−3 mA for the beam current and acceleration current, respectively. At the same time, it realizes the prediction of the variation trend of key parameters of the electric propulsion system, and shows relatively excellent performance compared with various prediction methods. The research results have certain engineering application value for the design optimization, on-orbit state prediction and long-term health management of the ion propulsion system, particularly exhibiting unique advantages in the ground-to-orbit equivalence of ion propulsion and the autonomous fault tolerance for deep space exploration missions.
KW - CNN-LSTM
KW - Data-driven
KW - Data-physics coupled model
KW - Grid system
KW - Ion propulsion system
KW - System model
UR - https://www.scopus.com/pages/publications/105036824976
U2 - 10.1016/j.actaastro.2026.04.007
DO - 10.1016/j.actaastro.2026.04.007
M3 - 文章
AN - SCOPUS:105036824976
SN - 0094-5765
VL - 246
SP - 664
EP - 679
JO - Acta Astronautica
JF - Acta Astronautica
ER -