Research of characteristic vector based on HHT in rotary machine failure diagnostics

This paper introduces the theory and method of Hilbert-Huang transform, and the advantages and shortcomings compared with traditional methods. By improving the stopping criterion of IMF components and endpoint problems, the computing time of signal processing has been reduced, and the issue of endpoint flying wing has been effectively resolved. This paper also analyses the physical meaning of Hilbert marginal spectrum, proposes a failure diagnostic method based the characteristic vector of Hilbert marginal spectrum, explains the steps of this method. At last, through an example of motor diagnostics, the paper verifies that the diagnostic method can accurately and efficiently identify the working state of rotary machines, and shows this method is reasonable.