Bitte wählen Sie ihr Lieferland und ihre Kundengruppe
Several issues concerning applications of artificial intelligence in technical diagnostics of machinery and equipment were presented. Apart from the first application focusing on acquisition of procedural and declarative knowledge from domain experts, all remaining applications are connected with knowledge acquisition from databases containing either examples obtained during simulations, or collected from real machinery and processes. Diagnostic inverse models are means of diagnostic concluding. The models are usually trained on simulation data. Belief network is other means of efficient modelling the diagnostic relations allowing the representation and deal with uncertainty. The diagnostic multimodels and multilevel models make an important contribution to the classification. Recently, approximate models of processes become important with special attention paid to different soft modelling methods. The classical machine learning methods may be efficiently employed for knowledge acquisition from examples. Since many databases accessible from monitoring systems contain unclassified though quite valuable data, applications of a new methodology of knowledge discovery in databases are very promising. Finally, a comprehensive application of a diagnostic expert system was discussed. The role of artificial intelligence in technical diagnostics in the future was briefly outlined.