On-orbit Engine Thrust Prediction Algorithm for Geosynchronous Satellites based on Neural Network

Abstract

Focused on the problem of complex in-orbit engine thrust prediction algorithm the geosynchronous satellite, a thrust prediction method based on the composition of the propulsion system is proposed in the paper. The method integrates theoretical model and neural network model, and the prediction model is simple and practical. In order to further improve the accuracy of the model, the in-orbit data was used to optimize the modified parameters using the multi-objective genetic algorithm NSGA-II. Finally, the optimized model is applied to different cases. The comparison between the simulation results and flight data shows that the accuracy of the optimized model is significantly improved, and the in-orbit thrust prediction method is correct and effective.