Time-frequency distribution decomposition with applications to recognize the looseness state of the viscoelastic sandwich structure

Abstract

In general, a vibration signal consists of several frequency modulation (FM) components. Every component contains different information, and can be characterized by its instantaneous amplitude (IA) and instantaneous phase (IP). In engineering applications, conventional time-frequency analysis methods and signal decomposition methods have shown their power in investigating features of the vibration signal. However, they are limited in resolution and it is hard to analyze these FM components individually. To overcome these deficiencies, a novel signal decomposition algorithm, named time-frequency distribution decomposition (TFDD), is proposed in this paper, which reconstructs one FM component of the signal at a time by estimating its IP and IA. The IA and IP are approximated by two polynomial functions respectively. One important advantage of TFDD is that it can directly extract the component we are interested in. Therefore, we can analyze the key component of the signal with little influence from other components. This will help us to characterize the vibration signal more deeply. Furthermore, it is very stable to noise. This is conductive to protecting the information of the vibration signal. The effectiveness of the TFDD is validated by a numerical simulation and the study of the vibration response signal collected from a viscoelastic sandwich structure (VSS). From the value of permutation entropy of the component extracted by TFDD, the looseness state of the VSS is recognized.

Time–frequency representation using IEVDHM–HT with application to classification of epileptic EEG signals

“…Thus, it is beneficial to study a non-stationary signal in TF domain in order to have meaningful and complete information [4]. TFR is widely used across many fields such as biomedical engineering [5,6], mechanical engineering [7,8], seismology [9], and speech signal processing [10] to study nonstationary signals. The TF domain-based features have been shown useful in pattern recognition and communication [11,12].…”

Digital instantaneous frequency measurement with wide bandwidth for real-valued waveforms using multiple sub-Nyquist channels

“…Whereas [10][11][12] are not only computationally intensive but also their resolution is limited by the Heisenberg uncertainty principle. To overcome these drawbacks, Girolami proposed a three-point method [13] and Yan proposed a signal decomposition algorithm named time-frequency distribution decomposition [14]. However, [10][11][12][13][14] are all limited by the Nyquist theorem so that the bandwidth of their DIFM cannot reach half of the sampling frequency.…”

Aging State Detection of Viscoelastic Sandwich Structure Based on ELMD and Sensitive IA Spectrum Entropy

“…Viscoelastic sandwich structure is a kind of composite structure consisting of elastic-solid layers and viscoelastic layers [1]- [3]. Due to the excellent energy dissipation performance of viscoelastic material, viscoelastic sandwich structure is playing an important role in the vibration absorption and noise reduction of mechanical equipment [4]- [6].…”