Diagnosis task decomposition for complex equipment remote monitoring
Résumé
It is hard to diagnose the complex equipment's fault depending only on a single diagnosis method or system. The integration of various methods constitutes a solution to this problem of local and remote diagnosis systems. That requires decomposing diagnosis tasks to sub tasks to properly diagnose the system. Therefore, this paper put forward disassembling diagnosis tasks based on the relational model and the Bayes network. All local and remote diagnosis systems are taken as elements of the equipment. The first part is focused on the steps and algorithm to evaluate the degree of closeness of relations among elements in the system. The closer the relationship is, the closer the part's element should be attached to the corresponding element of the diagnosis system. Then the structure modeling method based on the fuzzy cluster is introduced. The relation-based model could be described as a graph and be mapped into the Bayes network. The three element-node computing models are given. According to the relation between the element-node and the prior probability, the diagnosis task could be decomposed based on the element-node algorithm. At last, the FMS is discussed as an application instance.