Advances in digraph model processing applied to automated monitoring and diagnosis

Reliability models contain a great deal of information about the modeled system that can be used for automated diagnosis of the system. Using such a model for real time monitoring and diagnosis of dynamic systems may require frequent reconfiguration and solution of the model. As such, an efficient algorithm must be used for model solution. aper describes a new algorithm, called SourceDoubls, which efficiently solves for singletons and doubletons of a digraph reliability model. Compared with previous methods, the SourceDoubls algorithm provides up to a second order of magnitude reduction in the amount of time required to solve large digraph models. This significant increase in model solution speed allows complex digraphs containing thousands of nodes to be used as knowledge bases for real time automated monitoring and diagnosis applications. tly an application to provide monitoring and diagnosis of the Space Station Freedom Data Management System (DMS) is under development at NASA/Ames Research Center (ARC) and NASA/Johnson Space Center (JSC). The Control Center Complex (CCC) at JSC is planning to install this application in a console position in the CCC to assist mission controllers responsible for the DMS. This paper contains an overview of this system and provides details of how it will use digraph models processed by the SourceDoubls algorithm to accomplish its task.