Multidisciplinary System Design Optimization of Interplanetary Vehicles with Solar Electric Propulsion

Exploring Solar System objects such as planets, moons, and small bodies brings us vital information to unravel the origin of our universe. In such deep-space exploration missions using spacecraft, transporting abundant scientific instruments to the destination is of utmost importance. Generally, however, scientific resources tend to be highly restricted because of other essential components such as bus instruments, structure, and propellant. This paper proposes a design optimization method for interplanetary vehicles with solar electric propulsion, in an attempt to maximize the scientific capability in Solar System explorations. There are innumerable solutions for low-thrust interplanetary trajectories, on which the spacecraft system design problem strongly depends. An infinitely broad solution space must be evaluated extensively so as to determine the optimal system design, which is computationally extremely expensive. To achieve this, we develop a machine-learning model for low-thrust trajectories. A shape-based method is used in order to rapidly generate training trajectory data. The feasibility of the spacecraft components such as power management, communication, and thermal control is evaluated using the predicted trajectory. The developed method allows us to obtain an optimal system design solution that simultaneously maximizes the payload mass and the data downlink rate with the minimal time of flight.