Abstract
It is a difficult task to analyze the coupling characteristics of rotating machinery fault signals under the influence of complex and nonlinear interference signals. This difficulty is due to the strong noise background of rotating machinery fault feature extraction and weaknesses, such as modal mixing problems, in the existing Ensemble Empirical Mode Decomposition (EEMD) time–frequency analysis methods. To quantitatively study the nonlinear synchronous coupling characteristics and information transfer characteristics of rotating machinery fault signals between different frequency scales under the influence of complex and nonlinear interference signals, a new nonlinear signal processing method—the harmonic assisted multivariate empirical mode decomposition method (HA-MEMD)—is proposed in this paper. By adding additional high-frequency harmonic-assisted channels and reducing them, the decomposing precision of the Intrinsic Mode Function (IMF) can be effectively improved, and the phenomenon of mode aliasing can be mitigated. Analysis results of the simulated signals prove the effectiveness of this method. By combining HA-MEMD with the transfer entropy algorithm and introducing signal processing of the rotating machinery, a fault detection method of rotating machinery based on high-frequency harmonic-assisted multivariate empirical mode decomposition-transfer entropy (HA-MEMD-TE) was established. The main features of the mechanical transmission system were extracted by the high-frequency harmonic-assisted multivariate empirical mode decomposition method, and the signal, after noise reduction, was used for the transfer entropy calculation. The evaluation index of the rotating machinery state based on HA-MEMD-TE was established to quantitatively describe the degree of nonlinear coupling between signals to effectively evaluate and diagnose the operating state of the mechanical system. By adding noise to different signal-to-noise ratios, the fault detection ability of HA-MEMD-TE method in the background of strong noise is investigated, which proves that the method has strong reliability and robustness. In this paper, transfer entropy is applied to the fault diagnosis field of rotating machinery, which provides a new effective method for early fault diagnosis and performance degradation-state recognition of rotating machinery, and leads to relevant research conclusions.
Highlights
Components like rolling bearings and gears are the most extensively used and vulnerable components in a mechanical transmission system, and they may frequently fail under highly variable loads resulting from complex operating conditions
This paper proposes an early fault detection method for rotating machinery based on the harmonic-assisted multivariate empirical mode decomposition method and transfer entropy, and establishes a rotating mechanical state evaluation index based on the harmonic assisted multivariate empirical mode decomposition method (HA-Multivariate Empirical Mode Decomposition (MEMD))-TE to quantitatively describe the nonlinearity between historical state information under bearing fault conditions
A new nonlinear signal processing method—high-frequency harmonic-assisted unknown operation data → failure-free data TX →Y ; (b) Transfer entropy of failure-free data multi-empirical mode decomposition—is proposed to solve the problem of mode mixing in MEMD
Summary
Components like rolling bearings and gears are the most extensively used and vulnerable components in a mechanical transmission system, and they may frequently fail under highly variable loads resulting from complex operating conditions. To quantitatively study the nonlinear synchronous coupling characteristics and information transfer characteristics of rotating machinery fault signals between different frequency scales under the influence of complex and nonlinear interference signals, a rotating machinery fault testing method based on high-frequency harmonic-assisted multivariate empirical mode decomposition-transfer entropy is proposed for the quantitative study of the nonlinear synchronous coupling characteristics and information transfer between the rotating machinery fault signal and zero fault between different time–frequency scales. The running status of the mechanical system was effectively evaluated and diagnosed through the quantitative description of the degree of nonlinear coupling between signals by (1) extracting the principal characteristics of the mechanical transmission system through high-frequency harmonic-assisted multivariate empirical mode decomposition; (2) subjecting the denoised signal to transfer entropy calculation; and (3) establishing an HA-MEMD-TE (High-frequency harmonic-assisted multivariate empirical mode decomposition-transfer entropy)-based rotating machinery status evaluation index. This study employed transfer entropy for rotating machinery fault diagnosis, providing a new effective means and relevant research findings for the early fault diagnosis and performance degradation status identification of rotating machinery
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
