Development and analysis of a resource-aware power management system as applied to small spacecraft.

摘要

In this thesis, an overall framework and solution method for managing the limited power resources of a small spacecraft is presented. Analogous to mobile computing technology, a primary limiting factor is the available power resources. In spite of the millions of dollars budgeted for research and development over decades, improvements in battery efficiency remains low. This situation is exacerbated by advances in payload technology that lead to increasingly power-hungry and data-intensive instruments. The challenge for the small spacecraft is to maximize capabilities and performance while meeting difficult design requirements and small project budgets.; Power management is sought as a solution that can be applied with an existing generation of batteries. Ultimately, the power management problem is one of optimizing system performance and lifetime while maintaining safe operating conditions. This problem is formulated as a constrained, multi-objective combinatorial optimization problem. The problem is argued to be computationally intractable, and a formal proof of optimal substructure is presented. A multi-agent solution paradigm is developed that implements Dynamic Programming and Compromise Programming solutions.; A high-level, "black box" software simulation of a typical power system is used to evaluate the developed method. The parameters used in simulation are taken from existing satellite designs. Compared to a traditional spacecraft operations approach, the developed method is shown to be useful in maximizing the utility of the spacecraft. As a battery ages, the method also has an increasing benefit on minimizing the missions risk.