Diagnosis of Bearing Faults Using Optimal Teager-Kaiser Energy Concentrated Time-Frequency Transforms

The early detection of machinery faults plays a major role in high-cost and safety critical industrial processes. The nonstationary signal analysis is one of the means of machinery fault diagnosis. Time-frequency analysis (TFA) is an efficient tool to identify fault frequency components that reveal the machinery health information hidden in the nonstationary signals. Teager-Kaiser energy operator (TKEO) is a frequency-weighted energy estimation tool that can extract amplitude, frequency, and phase modulations associated with fault signals. The predominant focus of analysis so far has revolved around utilizing TKEO primarily as a demodulation algorithm rather than emphasizing its role as an energy estimation method. The aim of this article is to demonstrate the efficacy of TKEO as an energy estimation method in creating a compact window that is intended to concurrently optimize the Teager-Kaiser energy concentration (TKEC) within a specific time duration and frequency interval simultaneously. A multiobjective optimization method is used to determine the maximum time- and frequency-domain TKEC for sequences of finite length. A Teager-Kaiser energy concentrated optimal discrete window (TKECODW) is designed, and its characteristics are mathematically analyzed. Furthermore, TKECODW is used for detecting faults in machines using vibration signals.