Serial replacement maintenance philosophies and multiple-failure diagnostic strategies: a marriage of multiple-fault integrity and common cause sensibility

This paper discusses several methodological considerations associated with the incorporation of serial replacement maintenance philosophies into multiple-failure diagnostic strategies. If the serial replacement of items within isolated ambiguity groups is prioritized using isolation probabilities instead of failure probabilities then the resultant diagnostics are capable of exhibiting both the empirical common sense often attributed to single-fault isolation and the integrity and consistency inherent to multiple-failure diagnostics. Isolation probabilities may also be used to calculate the fault resolution statistics upon which many assessments of testability and diagnostic effectiveness are based. If is important, however, that multiple isolations be simulated for each combination of failed components, both in order to prevent the statistics from being skewed toward primary isolations and to allow the group sizes recorded for serially replaced components to properly reflect the topology of the device or system being diagnosed. The simulation of multiple isolations for each failure combination is also essential for the generation of accurate false removal predictions. Statistics based upon the serial replacement of fault groups isolated by a multiple-failure diagnostic strategy can allow design-phase assessments to more accurately predict the performance of fielded diagnostics, thereby allowing timely measures to be taken to reduce the overall cost of ownership.