Second-Order Time-Frequency-Multisqueezing Transform and its Applications in Rotating Machinery Fault Diagnosis

Rotating machinery holds a significant position in modern industry, and its fault diagnosis is crucial. The technique of time-frequency analysis (TFA) is crucial in fault diagnosis, as it adeptly handles nonstationary signals. Time-frequency-multisqueezing transform (TFMST) is an emerging TFA technique that can produce a time-frequency representation (TFR) with concentrated energy, and can effectively capture both harmonic and impulsive elements in nonstationary signals. However, limited by the performance of the estimation operator, TFMST faces challenges in producing a high resolution TFR when dealing with signals that exhibit strong nonstationarity. To overcome this limitation, this paper proposes a new TFA algorithm, the second-order TFMST, which embeds the second-order time-frequency estimation operator into the multisqueezing framework. The effectiveness of the proposed approach is confirmed through both numerical simulations and experimental results. The findings indicate that the proposed method achieves high-resolution TFR and demonstrates strong robustness against noise. In addition, the method can reconstruct the interested components.