Hypertext-Based Representation of Knowledge and Reasoning for User Guidance in Diagnosis of Complex Systems
Résumé
The development of a hypertext-based diagnosis system requires, on one hand, adapted tools for a task oriented navigation across the documents, and on the other hand, context sensitive browsing capability of the knowledge base. This paper presents a system that possesses both properties. It is based on a formal model ontology for representing a diagnostic domain, a computational scheme to implement and use the ontology, and a human-machine interface to provide users with an efficient access to objects, methods, and relations described in the ontology. We use an example based on the diagnosis of vibrations of rotating machines throughout the paper to illustrate the approach. The example has been built and implemented according to the norm (AFNOR 1998), which defines the protocol and the evaluation procedure of vibration from signals measured and recorded on the rotating machines.