Operational reliability calculations for critical systems

Reliability theory deals with the effect of mean time to repair upon overall system failure rates, but for critical systems such calculations are not what is required because an important performance criterion relates to operational failures, which are fundamentally different to unsafe failures: essentially they are the result of the system-level response to avoid unsafe failures. This paper introduces the particular problem for critical systems in general, presents an analysis of some of the relevant conditions and provides some simulation results in the context of a railway active suspension application that illustrate the overall effects and trends.