The failure process of mechanical equipments usually consists of a series 01 degraded states, correctly recognizing and estimating the current state of the equipment has important meaning for far- ther preventing equipment from degrading and failing. The equipment degradation state recognition method based on wavelet feature scale entropy and HSMM was proposed. It is the procedures of the method that the signal is decomposed by wavelet transform, the wavelet feature scale entropy is ex?tracted, the wavelet feature scale entropy vectors of the signals which are inputted to the HSMM for training are constructed, the running state classified model of the equipment based on HSMM is con?structed to recognizing tne equipment degradation states. A roller bearing was taken as an example and several states of roller with normal state and different fault severity states were recognized by the pro?posed method. Experiment results snow that the proposed method is very effective.
Key words
information processing /
wavelet feature scale entropy /
hidden semi-Markov model (HSMM) /
state recognition /
degradation state
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
References
[1]Jay Lee. Teleservice engineering in manufacturing challenges and opportunities [ J].International Journal of Machine Tools and Manufacture, 1998，38 ： 901 — 910.
[2] Blanco S, Figliosa A, Quian Q R, et al. Time-frequency analysis of electroence-phalograxn series ( HI ) : Information transfer func?tion and wavelets packets[J]. Physical Review E, 1998，57(1)： 932-940.
[3] 张文炬，苏清祖.车辆变速器故障诊断的Shannon熵研究[J]. 农业机械学报，2002，33(1):80-83.
ZHANb Wen-JU，SU Qing-zu. Exploit of shannon entropv for fault diagnosis of vehicle gearbox [ J]. Journal of Agriculture Mechanism, 2002,33(1)： 80-83. (in Chinese)
[4] 桂中华，韩凤琴.小波包特征熵神经网络在尾水管故障诊断中 的应用[J].中国电机工程学报，2005,25(4)：99 - 102.
GUI Zhong-hua, Peng-qing. Neural network based on wavelet pacet- characteristic entropy for fault diagnosis of draft tube[J]. Proceedings of the CSEE, 2005，25(4) ： 99 — 102. (in Chinese)
[5] 何正友,蔡玉梅，钱清泉.小波熵理论及其在电カ系统故障检测中的应用研究[J].中国电机工程学报，2005，25(5):38 - 43.
HE Zheng-you，CAI Yu-mei, QIAN Qing-quan. A study of wavelet entropy theory and its application in electric power system fault detection[J]. Proceedings of the CSEE, 2005, 25(5) ： 38 — 43. (in Chinese)
[6] 印欣运，何永勇，彭志科，等.小波熵及其在状态趋势分析中的 应用[J] ?振动工程学报，2004，17(2) ： 165 - 169.
YIN Xin-yun, HE Yong-yong，PENG Zhi-ke，et al. Study on wavelet entropy and its applications in trend analysis [J ]. Journal of Vibration Engineering, 2004，17(2): 165 — 169. (in Chi?nese)
[7] 封洲燕.应用小波熵分析大鼠脑电信号的动态特性[J].生物 物理学报，2002，18(3)： 325-330.
FENG Zhou-yan. Dynamic analysis of the rat KEG using wavelet entropy [J] . ACTA Biophysical Sinica, 2002，18(3) ： 325 — 330. (in Chinese)
[8] Raomer L R. A tutorial on hidden markov models and selected ap?plications in speech recognition [ J ]. Proceedings of the IEEE, 1989,77(2):257-286.
[9] Kwan C, Zhang X，Xu R，et al. A novel approach to fault diag?nosis and prognostics, proceedings[C]. ICRAJ 03. IEEE Inter?national Conference on Robotics and Automation. 1(3), Septem?ber 2003： 604-609.
[10] Chinnam R B, Banruah P. Autonomous diagnostics and prognos?tics through competitive learning driven HMM-based clustering [C]. Proc. Of the internal joint Conf on Neural Networks, July 2003： 2466-2471.
[11] Hatzipantelis E, Murray A,Penman J. Comparing hidden markov models with artincial neural network architectures for condition monitoring application[C].Fourth International Con?ference on Artificial Neural Networks，Cambridge, U. K,1995.
[12] Dong M，He D. Hidden semi-Markov models for machinery health diagnosis and prognosis[C]. Trans NAMRI/SMEXXXII (2004):199-206.
[13] Dong M，He D. Equipment health diagnosis and prognosis using hidden semi-Markov Models. 2006，30 ： 738 — 749.
[14] The Case Western Reserve University. Bearing data center[EB/ OL] . [ 2007-05 ] . http： Il www. eecs. cwru. edu/laboratory/ bearing/.
[15] Dong M，He D. A segmental hidden semi-markov odel (HSMM)-based diagnostics and prognostics framework and methodology [ J ]. Mechanical System and Signal Processing, 2006:1-19.
[16] Ferguson J D. Variable duration models for speech [ C]. Pro?ceedings of the 1980 Symposium on the Application of Hidden Markov Models to Text and Speech, Princeton, NJ, 1980： 143 一 179.
{{custom_fnGroup.title_en}}
Footnotes
{{custom_fn.content}}
