Reinforcing fuzzy rule-based diagnosis of turbomachines with case-based reasoning

This paper presents an integrated knowledge-based system, which combines fuzzy rule-based reasoning with case-based reasoning, for turbomachinery diagnosis. By incorporating a case-based reasoning sub-system in a fuzzy rule-based system, the integrated system allows past experience to be applied in a more direct way. This helps improve the diagnostic accuracy of the rule-based system. This approach has been implemented for the specific task of identifying possible causes of observed vibrations in rotating machines, based on the initial work presented in [18]. The ability that the case-based sub-system brings to the integrated system in improving the diagnostic efficacy of the original rule-based system is demonstrated with test results on real cases.