An Early Fault Feature Extraction Method Based on t-Distribution Stochastic Neighbor Embedding for Large Rotating Machinery
-
摘要: 将t分布随机近邻嵌入(t-SNE)流形学习方法应用于机械振动信号的故障特征提取,实现高维特征信息降维处理。通过小波包分解算法将原始振动信号分解为多层小波子空间,通过计算各层的小波阈值熵构造高维特征数据,然后采用t-SNE方法对构造的高维特征数据进行数据降维,获取低维故障特征信息。采用本特利转子试验台进行故障仿真实验,对采集获得的几种典型故障状态下的振动数据分别基于小波包阈值熵及统计特征构造2组高维数据,并对2组高维特征数据分别采用t-SNE方法进行数据降维处理获得其二维特征数据,通过对比验证了基于小波包阈值熵法构造高维数据后进行t-SNE数据降维的特征提取方法能够更有效的区分故障特征。Abstract: A manifold learning method of t-distributed stochastic neighbor embedding(t-SNE) is used in fault feature extraction of mechanical vibration signal to realize the dimension reduction of high-dimensional feature information. The acquired vibration signal is decomposed into wavelet subspaces by wavelet packet decomposition algorithm, and high dimensional feature vector is construct by the calculation of wavelet threshold entropy of every subspace, and then the stochastic neighbor embedding method based on t distribution is applied to reduce the constructed high-dimensional feature vector to low dimensional vector to acquire the features of fault. Fault simulation experiment on Bently rotor kit has been implemented, and two groups of high dimensional data vectors were been constructed by statistical parameter and wavelet package threshold entropy with the acquired vibration signals, and two groups of high dimensional vectors were reduced to 2 dimension vector by t-SNE method. By comparing, we validate that the extraction method of t-SNE dimension reduction of data based on high dimensional data constructed by Wavelet Packet Threshold Entropy is effective in distinguishing the fault features.
-
Key words:
- wavelet package threshold entropy /
- manifold learning /
- t-SNE /
- feature extraction
-
[1] 蒋玲莉,刘义伦,李学军,等.小波包去噪与改进HHT的微弱信号特征提取[J].振动、测试与诊断,2010,30(5):510-513 Jiang L L, Liu Y L, Li X J, et al. Feature extraction of weak signal using wavelet packet Denoising and improved hilbert-huang transformation[J]. Journal of Vibration, Measurement & Diagnosis, 2010,30(5):510-513 (in Chinese) [2] 冯辅周,司爱威,饶国强,等.基于小波相关排列熵的轴承早期故障诊断技术[J].机械工程学报,2012,48(13):73-79 Feng F Z, Si A W, Rao G Q, et al. Early fault diagnosis technology for bearing based on wavelet correlation permutation entropy[J]. Journal of Mechanical Engineering, 2012,48(13):73-79 (in Chinese) [3] 周智,朱永生,张优云,等.基于EMD间隔阈值消噪与极大似然估计的滚动轴承故障诊断方法[J].振动与冲击,2013,32(9):155-159 Zhou Z, Zhu Y S, Zhang Y Y, et al. Fault diagnosis of rolling bearings based on EMD Interval-Threshold denoising and maximum likelihood estimation[J]. Journal of Vibration and Shock, 2013,32(9):155-159 (in Chinese) [4] Roweis S T, Saul L K. Nonlinear dimensionality reduction by locally linear embedding[J]. Science, 2000,290(5500):2323-2326 [5] Tenenbaum J B, de Silva V, Langfold J C. A global geometric framework for nonlinear dimensionality reduction[J]. Science, 2000,290(5500):2319-2323 [6] Balasubramanian M, Schwartz E L, Tenenbaum J B, et al. The isomap algorithm and topological stability[J]. Science, 2002,295(5552):7 [7] Zhang Z Y, Zha H Y. Principal manifolds and nonlinear dimensionality reduction via tangent space alignment[J]. SIAM Journal of Scientific Computing, 2005,26(1):313-338 [8] Shao R P, Hu W T, Wang Y Y, et al. The fault feature extraction and classification of gear using principal component analysis and kernel principal component analysis based on the wavelet packet transform[J]. Measurement, 2014,54:118-132 [9] Bu Y D, Chen F Q, Pan J C. Stellar spectral subclasses classification based on Isomap and SVM[J]. New Astronomy, 2014,28:35-43 [10] 徐长发,李国宽.实用小波方法[M].武汉:华中科技大学出版社,2001 Xu C F, Li G K. Practical wavelet method[M]. Wuhan: Huazhong University of Science and Technology Press, 2001 (in Chinese) [11] Bae S H, Choi J Y, Qiu, J, et al. Dimension reduction and Visualization of large high-dimensional data via interpolation[C]//Proceedings of the 19th ACM International Symposium on High Performance Distributed Computing. New York, NY, USA: ACM, 2010:203-214 [12] Hinton G E, Roweis S T. Stochastic neighbor embedding[C]//Advances in Neural Information Processing Systems 15. Cambridge, MA: MIT Press, 2002,15:833-840 [13] Bunte K, Haase S, Biehl M, et al. Stochastic neighbor embedding (SNE) for dimension reduction and visualization using arbitrary divergences[J]. Neurocomputing, 2012,90:23-45 [14] Gisbrecht A, Mokbel B, Hammer B. Linear basis-function t-SNE for fast nonlinear dimensionality reduction[C]//Proceedings of the 2012 International Joint Conference on Neural Networks (IJCNN). Brisbane, QLD: IEEE, 2012:1-8 [15] van der Maaten L, Hinton G. Visualizing data using t-SNE[J]. Journal of Machine Learning Research, 2008,9(11):2579-2605
点击查看大图
计量
- 文章访问数: 171
- HTML全文浏览量: 34
- PDF下载量: 5
- 被引次数: 0