A model-based methodology for the integration of diagnosis and fault analysis during the entire life cycle

This chapter shows that model-based systems technology, which has been developed in Artificial Intelligence, is a significant contribution to addressing the challenge of diagnosis and fault analysis, and focuses on the methodological aspects. In this approach, a library of generic, reusable component models represents the sharable knowledge. Automated model composition based on this library and a structural description of a specific plant together with domain-independent problems solving algorithms allow for the automated generation of the required plant-specific results, thus reducing or even eliminating the human efforts involved in the analysis and programming. The chapter outlines the foundations of this approach and demonstrates its maturity by describing its application to the automated generation of code for vehicle on-board diagnosis and the generation of FMEA reports. Finally, it discusses the reuse of the models and the algorithmic basis for other tasks during the lifecycle, such as diagnosability analysis and test generation. © 2007