Skip to Main Content
This paper proposes an intelligent diagnosis method for condition diagnosis of rotating machinery by using wavelet transform (WT) and ant colony optimization (ACO), in order to detect faults and distinguish fault types at an early stage. The WT is used to extract a feature signal of each machine state from a measured vibration signal for for high-accuracy condition diagnosis. The decision method of optimum frequency area for the extraction of the feature signal is discussed by using real plant data. We convert the state identification for the condition diagnosis of rotating machinery to a clustering problem of the values of the nondimensional symptom parameters (NSPs). ACO is introduced for this purpose. NSPs are calculated with the recomposed signals of each frequency level. These parameters can reflect the characteristics of the signals measured for the condition diagnosis. The synthetic detection index (SDI), on the basis of statistical theory, is defined to evaluate the applicability of the NSPs. The SDI can be used to select better NSPs for the ACO. Practical examples of diagnosis for a bearing used in the centrifugal fan system are shown to verify the effectiveness of the methods proposed in this paper.