Diagnositc Reasoner Model Production from Atlas Test Program Sets

Model-based diagnostic reasoning provides improvements in fault isolation and repair time as well as cost savings provided by a reduction in false pull rates. To support legacy systems that use ATLAS test program sets for fault isolation and repair, Honeywell has developed a method for extracting diagnostic fault models from ATLAS source code. We will describe the content of the fault models and highlight model features that use data collected in a netcentric maintenance environment to optimize reasoner performance. We will then present a description of the method used to create bootstrap fault models, including a configurable ATLAS source code parser that uses regular text expressions to extract test and fault information from Test Program Sets and create fault models.