Abstract

The early weak fault characteristics of rolling bearings are extremely weak and are easily overwhelmed by other noises. In order to effectively extract the characteristics of the early weak faults of the rolling bearings and draw on the multilayer wavelet decomposition idea, a method for diagnosing the early weak faults of the rolling bearing based on the multilayer reconstruction filter is proposed. As we all know, empirical wavelet transform (EWT) makes full use of wavelet filter bank, and variational mode decomposition (VMD) uses Wiener filter bank. This paper fully combines the advantages of the above two methods, adaptively determines the number of modes through empirical wavelet decomposition and divides the original signal, extracts the frequency band that contains the fault characteristic information, and effectively eliminates noise interference. These steps are repeated until the optimal component of the condition is obtained. In the output layer, the weak fault impact components are further separated by the strong filtering and signal decomposition capability of VMD. The advantages of the proposed method are proved by the experiment of weak fault of rolling bearing and the accelerated failure experiment of full life. The proposed method has the advantages of reducing noise influence and adaptive estimation of decomposed modes, which can be applied more efficiently in practice.

Highlights

  • And effective identification of early weak faults for rolling bearings has an important significance to ensure the safety of equipment operation [1]. e small impact of early weak faults leads to weak fault characteristics

  • Data-driven fault diagnosis methods for rolling bearings have been the most widely studied, most of which are based on vibration signals and acoustic signals, which are essentially the same, but each has its own advantages and disadvantages [4]. e advantages of vibration signal used in fault diagnosis are as follows: the cost of the sensor is relatively low, the vibration signal is easy to be measured, and the measured vibration signal can contain more fault information

  • Its disadvantages are as follows: this is a kind of contact measurement, and the sensor needs to be installed in the position close to the workpiece. e advantages of using acoustic signals for fault diagnosis are as follows: the requirements for the installation position are not very strict, they do not need to be attached to the workpiece under test, and the measurement cost is relatively low. e disadvantage is that the measured acoustic signal may contain sound other than the target workpiece, so the SNR is relatively small [5]

Read more

Summary

Introduction

And effective identification of early weak faults for rolling bearings has an important significance to ensure the safety of equipment operation [1]. e small impact of early weak faults leads to weak fault characteristics. Data-driven fault diagnosis methods for rolling bearings have been the most widely studied, most of which are based on vibration signals and acoustic signals, which are essentially the same, but each has its own advantages and disadvantages [4]. E advantages of vibration signal used in fault diagnosis are as follows: the cost of the sensor is relatively low, the vibration signal is easy to be measured, and the measured vibration signal can contain more fault information. E advantages of using acoustic signals for fault diagnosis are as follows: the requirements for the installation position are not very strict, they do not need to be attached to the workpiece under test, and the measurement cost is relatively low. Adam Glowacz proposed a fault diagnosis method of single-phase motor based on acoustic signal [5]. Wang et al used timefrequency curves of vibration signals for fault detection under variable speed [11]

Objectives
Methods
Results
Conclusion
Full Text
Published version (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