Scaling demodulation-based mode decomposition for analyzing nonstationary signal with close-spaced and intersecting frequency trajectories

Mode decomposition is an important tool for processing vibration signals in engineering fields. However, classic methods cannot work for decomposing a signal with close-spaced and intersecting frequency trajectories. For addressing the aforementioned problem, a novel technique, termed scaling demodulation-based mode decomposition (SDMD), is proposed in this paper. The main novelties of this paper are summarized as follows: (a) a scaling demodulation transform (SDT) is developed to estimate instantaneous frequency (IF) curves by selecting optimal demodulation parameters from a series of parameter sets with a criterion for searching maximal spectrum peaks of demodulated signals; and (b) a time-frequency representation (TFR) with high energy concentration is constructed using the Dirac delta function, estimated IFs and instantaneous amplitudes (IAs). The performance of the SDMD is evaluated through simulated and measured signals. Furthermore, comparison results with classic methods show that the developed technique has a much better ability to decompose and characterize nonstationary signals with close-spaced and intersecting frequency trajectories.

Automated summary

This paper proposes a novel technique, called scaling demodulation-based mode decomposition (SDMD), for processing signals with close-spaced and intersecting frequency trajectories. The performance of SDMD is evaluated through simulated and measured signals, and comparison results with classic methods show its superior ability in decomposing and characterizing nonstationary signals.