Analyzing the Interface of Reliability and Economics of Unmanned Satellites

This paper deals primarily with the reliability of unmanned scientific or commercial satellites. In the course of reliability analysis an hierarchy of reliability predictions is generated, reaching from preliminary apportionment by subsystems: to figures of merit, usually obtained by Boolean algebra; to results based on failure modes, effects and criticality analysis (FMECA); to data from an elaborate, though still static, fault tree; and eventually to data depicting the dynamics of mission time line and contingency analyses. The outlined routine has been developed, mainly in the US, to quite some perfection, but it still leaves unanswered the customer's most natural concern: how much reliability, within what time, at what price? Therefore, an attempt is made to schedule the above work elements, so that the amount of reliability as well as its allocation become subject to the classical criterion of marginal utility. It is fully realized that there are considerations, such as design freeze partly due to long-lead items, that put a limit to this endeavor. The postulated analysis consists, in essence, of two models in sequence: first, the fault tree, which exhibits the technical failure modes, their probability of occurence, and their interdependence. Second, the mission failure model, referring to total and partial deficiences in performing utility functions, and to their interactions as regards the degree of performance.