Journal article
Bearing fault diagnosis based on an improved morphological filter
Measurement : journal of the International Measurement Confederation, Vol.80, pp.163-178
02/01/2016
DOI: 10.1016/j.measurement.2015.11.028
Abstract
•A new construction of SE is proposed for morphological filter.•Harmonic waveform with the resonant frequency is applied to construct SE.•The newMF’s de-noising capability is superior to existing methods.•Proposed new method has been verified by simulation and experimental cases.
The extraction of repetitive impacts from vibration signals plays an essential role in bearing fault detection. Among different signal processing algorithms, morphological filter (MF) has attracted lots of attention because it could directly extract the geometric structure of the impulsive feature and only needs little computation. However, the conventional MF and some current improvements are based on the local optima of the raw signal to de-noise the noisy signal and its faulty feature extracting capability would be greatly affected by the noise. In this paper, a new improved MF algorithm is proposed to overcome such deficiency. Firstly, morphological gradient (MG) operator is selected in this paper due to its capability of picking up both positive and negative impulses. Then, based on the relationship between the defect induced impulse and a harmonic function with the resonant frequency, the harmonic waveform in a period is adopted to instruct the construction of structuring element (SE). The improved MF can obtain the fault feature from low SNR signals. The processing results of a simulation signal and two sets of experimental signals and a set of comparisons verify the effectiveness and robustness of the proposed method.
Details
- Title: Subtitle
- Bearing fault diagnosis based on an improved morphological filter
- Creators
- Zhiyong HuChao WangJun ZhuXingchen LiuFanrang Kong
- Resource Type
- Journal article
- Publication Details
- Measurement : journal of the International Measurement Confederation, Vol.80, pp.163-178
- Publisher
- Elsevier Ltd
- DOI
- 10.1016/j.measurement.2015.11.028
- ISSN
- 0263-2241
- eISSN
- 1873-412X
- Grant note
- DOI: 10.13039/501100001809, name: National Natural Science Foundation of China, award: 51475441
- Language
- English
- Date published
- 02/01/2016
- Academic Unit
- Industrial and Systems Engineering
- Record Identifier
- 9984222055602771
Metrics
1 Record Views