An application of Rényi entropy segmentation in fault detection of rotating machinery

The paper presents a new approach for change detection in vibration signals of a rotating machine using time-frequency information content, making use of the short-term time-frequency Rényi entropy and a segmentation algorithm, based on maximum a posteriori probability (MAP) estimator. The segmentation algorithm operates on Rényi entropy, as a new space of decision. This approach enables more robust change detection in vibrating signals. Finally, we present an application of the proposed approach for a rotating machine, a pump, after the blind source separation (BSS) of the main vibration sources has been performed.