Gearbox fault diagnosis based on VMD-FastICA

In mechanical transmission equipment, gearbox plays an extremely important role, it is widely used in all walks of life. It is not only expensive, but also prone to failure. With the increase of running time, gear box often has broken teeth, pitting, wear and other failures, bearing outer ring, inner ring, rolling element are also prone to failure. Often a gear failure will lead to the shutdown of the whole equipment. Real time monitoring the running state of the gearbox and fault diagnosis of the gear can take reasonable measures to eliminate the fault in advance and ensure the safe operation of the equipment.

The effective extraction of mechanical fault features is the cornerstone of fault diagnosis. A multi feature fusion diagnosis method based on variational mode is proposed. The energy and permutation entropy of high frequency IMF components are used as the input of vector machine to recognize the fault type. EMD method is used to decompose the vibration signal into several single component signals, and the sample entropy of each component is extracted as the eigenvalue. Finally, vpmcd classifier is used for fault recognition and classification. A fault diagnosis method of rolling bearing based on LMD and spectrum correction is proposed. A fault diagnosis method based on global local mean decomposition and minimum entropy deconvolution is proposed. Based on the above analysis, this paper proposes a gearbox fault diagnosis method combining VMD and FastICA. Through the analysis of the actual fault of the gearbox, the results show that it can well avoid the mode aliasing problem in LCD decomposition, and can accurately extract the fault characteristic frequency and find the specific location of the fault.

Aiming at the problem that it is difficult to realize compound fault diagnosis of gearbox in the case of single channel, a fault diagnosis method based on VMD FastICA is proposed. This method uses fast kurtosis map and correlation coefficient index as the criteria for screening IMF components, and overcomes the limitation that the number of observed signals of traditional fast ICA method must be larger than the number of original signals. Compared with the traditional LCD FastICA method, the results show that the method proposed in this paper can achieve better separation of mixed faults of gearbox, highlight the fault characteristics of bearing inner ring and gear, and the effect is obvious. At the same time, the method proposed in this paper can solve the problem that multiple sensors can not be installed in practical engineering, effectively reduce the impact of noise, and has a good effect in practical application.

Scroll to Top