An analysis of dependent failures in spacecraft: Root causes, coupling factors, defenses, and design implications

This research represents the first comprehensive adaptation and extension of dependent failure (DF) analysis methodology, developed for the nuclear power industry, to the design of aerospace systems. It includes a historical overview of DF methodology; new material with which to characterize and classify DFs in spacecraft; the collection and analysis of Space Shuttle in-flight anomaly data in search of actual and potential DFs; new methods and guidelines for the engineering analysis of a spacecraft design for the presence of, and defenses against DF candidates; a proposed new model of DF based on classes of causes; and finally, a case study applying and extending all that has been developed and learned in this research (from both nuclear power-related methodology and Space Shuttle data) to an example Earth-orbital spacecraft. The research concludes that the frequency of DFs in spacecraft (as characterized by the Space Shuttle) is not significantly different from that found in nuclear power systems; spacecraft reliability cannot be acccurately (or conservatively) estimated without consideration of DFs; DFs can be most efficiently prevented by defending against coupling factors rather than root causes; through the use of component diversity, the reliability of spacecraft systems can be increased without the need for increasing the reliability of individual components.