Relating failure prognostics to system benefits

Failure prognostics offer several potential benefits for operational systems. By predicting failures before they occur, managers can make better choices about continued usage of a system and better decisions about the timing of preventive maintenance. This paper explores the relationship between the potential benefits and the accuracy of the failure prognostics. It uses simple models to show how inaccuracies in the failure model and the stress data can be related to the benefits. Ultimately, such models can be used to establish accuracy requirements for stress monitors and failure models.