Rotating machine abnormal vibration source identification method based on VT-ANM-SWT

Abstract

The invention relates to a rotating machine fault diagnosis technology, in particular to a rotating machine abnormal vibration source identification method based on VT-ANM-SWT. The method comprises the steps of: obtaining vibration excitation source mixed signals of the surface of a rotating machine through a vibration acceleration sensor; performing independent component analysis on the vibration signals by using an optimized self-adaptive negentropy maximization blind source separation algorithm, and setting the number of results needing to be output by the algorithm according to the pre-estimated number of main vibration sources; and carrying out time-frequency analysis on each separated independent signal by using the SWT steady-state wavelet transform technology. Based on priori knowledge of a rotating machine, the method can effectively identify the vibration sources of the rotating machine, and can also effectively identify the abnormal vibration sources in abnormal operation. The method successfully solves the problem of abnormal vibration sources of a rotating machine, and provides a set of system for analysis of the problems in the future.

Description

Technical field [0001] The present invention belongs to the field of rotating mechanical fault diagnosis, and in particular, the present invention relates to a rotational mechanical heteroscopic source identification method based on VT-ANM-ST. The present invention is suitable for signals such as internal combustion engines, electric motors, and the like. Background technique [0002] Reduce vibration, reduce noise, and reduce the fault is an important research direction in the field of rotating mechanical research. Effective to reduce the noise reduction and fault diagnosis of rotating machinery, accurately identify its main vibration excitation source is key. Rotating machines tend to be complex, internal incentives, and each coupling, passing through different vibration paths, complex response process. Therefore, the excitation source signal of the rotating machine tends to exhibit non-smooth, nonlinear, and multi-rejection system characteristics. [0003] Comparative Traditio...

AI Technical Summary

Problems solved by technology

Among them, the traditional blind source separation algorithm based on negentropy maximization has to rely on experience to select the coefficients of the quadratic function in its negentropy estimation function, so the accuracy of its calculation results will vary due to different signals.

Images