Multiobjective optimization of crewed spacecraft supportability strategies

Abstract

Future crewed missions present a logistical challenge that is unprecedented in human spaceflight. Astronauts will travel farther from Earth than ever before, and stay in space for longer, without access to regular resupply or the option to abort and return home quickly in the event of an emergency. Under these conditions, supportability - that is, the set of characteristics of a system that drive the amount of resources required to enable safe and effective system operations - will be a much more significant driver of system lifecycle properties than it has been in the past. Many of the strategies that are currently used to mitigate risk for human spaceflight will no longer be available, feasible, or effective. To enable the human exploration missions of the future, new supportability strategies must be identified, characterized, developed, and implemented.

This dissertation addresses this problem by developing and presenting a new methodology for modeling and multiobjective optimization of supportability strategies, minimizing mass and maintenance crew time requirements subject to constraints on risk. The supportability strategy optimization problem is encoded as a multiobjective Constraint Optimization Problem (COP), with a set of decision variables defining a range of supportability strategy options related to level of maintenance, On-Demand Manufacturing (ODM), commonality, redundancy, and distributed functionality. A supportability model is developed which enables evaluation of the mass and crew time associated with a given assignment to those decision variables, or the lower bounds on those objective values associated with a partial assignment.

The resulting model, called Mass, Crew time, and Risk-based Optimization of Supportability Strategies (MCROSS), advances the state of the art in space mission supportability analysis by enabling holistic, rapid, multiobjective optimization and evaluation of the tradeoffs between mass, crew time, and risk for future missions. Model outputs are verified against results from Monte Carlo simulation, and validated via comparison to an existing state-of-the-art NASA supportability model and to flight maintenance data from the International Space Station (ISS). MCROSS is then demonstrated using two case studies, one based on a notional system and the other examining the ISS Oxygen Generation Assembly (OGA). The notional case study is used to validate optimization results against the Pareto frontier identified via full enumeration.

The second case study demonstrates the application of this methodology to a real-world system, showing that MCROSS can identify supportability strategies offering lower mass and crew time options than current approaches. A series of sensitivity analyses are also presented to demonstrate the application of MCROSS in an iterative design process. These results, and the associated analysis capability, provide a powerful analysis tool that can help inform system development and mission design by characterizing tradeoffs between mass, crew time, and risk, along with the underlying strategy decisions. The results and implications of this research are discussed, along with assumptions and limitations. Finally, the contributions of this research are summarized along with potential areas of future work.