GEOSYNCHRONOUS COMMUNICATION SATELLITE RELIABILITY: STATISTICAL DATA ANALYSIS AND MODELING

摘要

Reliability is a critical design attribute for systems operating in remote or inhospitable environments, such as spacecraft or sub-sea installations. Surprisingly however, limited empirical data and statistical analysis of satellite reliability exist in the technical literature. In this paper, we fill this gap by collecting and analyzing life data of over four hundred geosynchronous communication satellites successfully launched between 1990 and 2008. First, we conduct a nonparametric reliability analysis of our sample using the Kaplan-Meier estimator (given the censored nature of our data). Second, we derive confidence intervals for our nonparametric reliability results. Finally, we conduct a parametric analysis and discuss the need for finite mixture distributions to properly capture the nonparametric failure behavior of the communication satellites in our sample. The results here provided should prove useful to satellite manufacturers and operators, for example in providing an empirical basis for reliability growth programs, or better gauging investment risks in space systems.