Multi-fidelity and multi-objective optimization of low-thrust transfers with control strategy for all-electric geostationary satellites

Abstract

To reduce the transfer time and radiation damage of all-electric geostationary satellites, this article develops a metamodel based multi-objective optimization scheme for low-thrust geostationary Earth orbit transfers. A multi-stage control strategy is employed to convert the optimal Earth orbital transfer problem into a parameterized multi-objective optimization problem with minimum total transfer time and proton displacement dose objectives. Based on the notion of multi-model fusion, the multi-fidelity system dynamics and radiation damage models are established to reduce the optimization cost. Furthermore, a novel adaptive Co-Kriging based multi-objective optimization method is proposed to effectively solve the multi-objective optimization problem with multi-fidelity models. In this approach, Co-Kriging metamodels are constructed using the multi-fidelity data to approximate the time-consuming low-thrust transfer models for optimization. And the Co-Kriging metamodels are adaptively refined by the infill samples to explore the Pareto frontier efficiently. Finally, a real-world all-electric geostationary satellite transfer example is investigated. The results show that a number of non-dominated low-thrust trajectories are successfully obtained with limited computational budget. Moreover, the total transfer time and displacement damage dose are significantly reduced after optimization, which illustrates the effectiveness of the metamodel based multi-objective low-thrust transfer optimization scheme.