A key challenge to successful implementation of Time-Frequency (TF) analysis for machine diagnosis is the development of an accurate and consistent method to interpret the images so that they truly reflect machine condition. In this paper, the time-frequency domain is used to study signals from industrial bearings. Examination results are presented as TF images. A Fuzzy logic approach is developed to classify Fourier Descriptors obtained from the time–frequency images so that the fault can be automatically identified. The analysis and results of experimental data indicate that bearing faults can be correctly classified using the developed method.
This work by European American Journals is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License