Efficient on-orbit-servicing via routing optimization and low-thrust propulsion

On-Orbit Servicing (OOS) is emerging as a very convenient option for the purpose of maintaining and supporting operational satellites. This research addresses routing optimization of a servicing spacecraft capable of performing successive transfers toward multiple satellites, with the use of low-thrust propulsion. The problem at hand is rather challenging, and a two-layer optimization methodology is proposed for its solution: (i) the inner layer estimates the costs of orbital transfers (in terms of final mass ratio), while (ii) the outer layer determines the optimal sequence and timing for servicing multiple satellites. For comparison, the inner layer investigates orbit transfers completed through (a) high-thrust or (b) low-thrust propulsion. Because the operational scenario involves orbits that are subject to differential precession related to the Earth oblateness, either direct transfers or intermediate drift orbits can be employed to reach each satellite. While direct transfers are preferred when two orbit planes (nearly) coincide, temporary injection into intermediate orbits aims at pursuing the natural alignment of different orbit planes in order to minimize the transfer cost. The low-thrust scenario turns out to be particularly challenging because the evolution of the right ascension of the ascending node along the powered transfer arcs (toward or from intermediate drift orbits) must be considered, due to their prolonged duration compared to the high-thrust case. The outer layer implements a dedicated encoding, accompanied by simulated annealing, aimed at the simultaneous selection of the order of satellites to visit and the respective epochs. In this context, operational requirements are also considered, by including suitable waiting times, associated with phasing and servicing operations. The methodology at hand is successfully applied to an existing mega-constellation, while including multiple mission scenarios, with horizon ranging from 2 through 4 years. Numerical simulations with the use of both high- and low-thrust propulsion point out feasibility of OOS with both technologies, together with the respective limitations, as well as the apparent advantage of using low-thrust propulsion in terms of final mass ratio.