Affiliation:
1. Xichang Satellite Launch Center
Abstract
Aim at some problem in fault diagnose: the characteristic frequency depends on the speed, the spectrum is complex , which are easy to diagnose error when in the variable conditions, and it is often difficult to identify the fault positioning in the frequency domain. the paper puts forward a new method: Variable condition bearing fault diagnosis basing on time-domain and artificial intelligence , not depend on speed and frequency domain. This method use vibration signal, calculates the kurtosis, skewness, rms etc 12 time-domain value, then these character vectors are sent to the neural network classifier to complete fault type pattern recognition, Finally, the same faults are sent to the next neural network for fault positioning and damage extent identification. The experimental result showed that using this method can obtain very good effect.
Publisher
Trans Tech Publications, Ltd.
Reference4 articles.
1. D. Bently，Predictive maintenance through the monitoring and diagnostics of rolling element bearings[R]. Bently Nevada Co，Application Note，1989，44，2-8.
2. C.T. Yiakopoulos et al. ，Rolling element bearing fault detection in industrial environments based on a K-means clustering approach，Expert Systems with Applications 38 (2011) 2888–2911.
3. Arai M. Mapping abilities of three-layered neural networks. Proc. of IJCNN, Seattle, USA, 1991. 419~423.
4. Loparo K.A. Bearings vibration data set. Case Western Reserve University. .
Cited by
4 articles.
订阅此论文施引文献
订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献