Adaptive Blind Extraction of Rolling Bearing Fault Signal Based on Equivariant Adaptive Separation via Independence
-
摘要: 针对复杂工况下滚动轴承故障信号盲提取问题，该文提出一种独立分量分析(ICA)中非线性函数自适应选择方法，解决了等变化自适应源分离算法(EASI)在多类振动源共存的情况下无法分离轴承故障信号的问题。此外，为了解决在线盲分离算法稳态误差与收敛速率的平衡问题，提出基于模糊逻辑的自适应迭代步长选择方法，极大地提高了学习算法的收敛速度，且稳态误差更小。轴承故障数据的盲提取仿真结果验证了算法的性能。Abstract: For the problem of blind extraction of rolling bearing fault signals under complex working conditions, an adaptive selection method of non-linear functions in Independent Component Analysis (ICA) is proposed, which solves the problem that Equivariant Adaptive Separation via Independence(EASI) can not separate bearing fault signals when multiple vibration sources coexist. In addition, in order to balance the steady-state error and convergence rate of the online blind separation algorithm, an adaptive iterative step selection method based on fuzzy logic is proposed, which improves greatly the convergence speed of the learning algorithm and reduces the steady-state error. The simulation results of blind extraction of bearing fault data verify the performance of the proposed algorithm.
-
Key words:
- Blind signal separation /
- Fault detection /
- Super-Gaussian /
- Sub-Gaussian /
- Fuzzy logic
-
表 1 模糊推理规则
$\mu {{ = S1} }$ $\mu {{ = S2} }$ $\mu {{ = M} }$ $\mu {{ = B} }$ ${D_i}{{ = S1} }$ ${{S1} }$ ${{S1} }$ ${{S2} }$ ${{M2} }$ ${D_i}{{ = S2} }$ ${{S1} }$ ${{S2} }$ ${{M1} }$ ${{M2} }$ ${D_i}{{ = M} }$ ${{M1} }$ ${{M1} }$ ${{M2} }$ ${{B1} }$ ${D_i}{{ = B} }$ ${{M2} }$ ${{M2} }$ ${{B1} }$ ${{B2} }$ 
下载: 导出CSV
表 2 算法的成功率比较
算法名称 成功率(%) EASI, $g(x) = {x^3}$ 0 EASI, $g(x) = \tanh (x)$ 12 本文算法，使用固定步长 88 本文算法，使用模糊逻辑步长 97
下载: 导出CSV
表 3 算法的性能比较
算法 ISR SOBI 0.069 FastICA, $g( \cdot ) = \tanh ( \cdot )$ 0.140 FastICA, $g( \cdot ) = {( \cdot )^3}$ 0.170 FastICA, $g( \cdot ) = ( \cdot )\exp ( - {( \cdot )^2}/2)$ 0.160 本文算法，使用模糊逻辑步长 0.110
下载: 导出CSV
-
郝如江, 卢文秀, 褚福磊. 声发射检测技术用于滚动轴承故障诊断的研究综述[J]. 振动与冲击, 2008, 27(3): 75–79. doi: 10.3969/j.issn.1000-3835.2008.03.019HAO Rujiang, LU Wenxiu, and CHU Fulei. Review of diagnosis of rolling element bearings defaults by means of acoustic emission technique[J]. Journal of Vibration and Shock, 2008, 27(3): 75–79. doi: 10.3969/j.issn.1000-3835.2008.03.019 HYVÄRINEN A, KARHUNEN J, and OJA E. Independent Component Analysis[M]. New York: Wiley, 2001: 9–11. doi: 10.1007/978-0-387-73003-5_305. 李扬, 张伟涛, 楼顺天. 基于联合对角化的声信号深度卷积混合盲分离方法[J]. 电子与信息学报, 2019, 41(12): 2951–2956. doi: 10.11999/JEIT190067LI Yang, ZHANG Weitao, and LOU Shuntian. Deep convolution blind separation of acoustic signals based on joint diagonalization[J]. Journal of Electronics &Information Technology, 2019, 41(12): 2951–2956. doi: 10.11999/JEIT190067 陈雷, 韩大伟, 郭艳菊, 等. 基于回溯搜索优化的卷积混合语音盲分离[J]. 计算机工程与应用, 2017, 53(15): 137–143.CHEN Lei, HAN Dawei, GUO Yanju, et al. Speech convolutive blind separation algorithm based on backtracking search optimization[J]. Computer Engineering and Applications, 2017, 53(15): 137–143. 龚晓峰, 毛蕾, 林秋华, 等. 基于四阶累积量张量联合对角化的多数据集联合盲源分离[J]. 电子与信息学报, 2019, 41(3): 509–515. doi: 10.11999/JEIT180414GONG Xiaofeng, MAO Lei, LIN Qiuhua, et al. Joint blind source separation based on joint diagonalization of fourth-order cumulant tensors[J]. Journal of Electronics &Information Technology, 2019, 41(3): 509–515. doi: 10.11999/JEIT180414 刘嘉辉, 董辛旻, 李剑飞. 基于全矢谱时间固有尺度分解和独立分量分析盲源分离降噪的滚动轴承故障特征提取[J]. 中国机械工程, 2018, 29(8): 943–948. doi: 10.3969/j.issn.1004-132X.2018.08.009LIU Jiahui, DONG Xinmin, and LI Jianfei. Fault feature extraction of rolling bearings based on noises reduced by full vector spectrum ITD-ICA blind source separation[J]. China Mechanical Engineering, 2018, 29(8): 943–948. doi: 10.3969/j.issn.1004-132X.2018.08.009 HE Jun, CHEN Yong, ZHANG Qinghua, et al. Blind source separation method for bearing vibration signals[J]. IEEE Access, 2018, 6: 658–664. doi: 10.1109/ACCESS.2017.2773665 HUANG Xiangdong, JIN Xukang, and FU Haipeng. Short-sampled blind source separation of rotating machinery signals based on spectrum correction[J]. Shock and Vibration, 2016, 2016: 9564938. doi: 10.1155/2016/9564938 胡纯直. 风机齿轮箱多故障诊断问题研究[D]. [硕士论文], 浙江大学, 2017.HU Chunzhi. The research on multi-fault diagnosis of wind turbine gearbox[D]. [Master dissertation], Zhejiang University, 2017. 陈恩利, 张玺, 申永军, 等. 基于SVD降噪和盲信号分离的滚动轴承故障诊断[J]. 振动与冲击, 2012, 31(23): 185–190. doi: 10.3969/j.issn.1000-3835.2012.23.034CHEN Enli, ZHANG Xi, SHEN Yongjun, et al. Fault diagnosis of rolling bearings based on SVD denoising and blind signals separation[J]. Journal of Vibration and Shock, 2012, 31(23): 185–190. doi: 10.3969/j.issn.1000-3835.2012.23.034 许同乐, 王营博, 郑店坤, 等. 基于LMD-ICA降噪的滚动轴承故障特征提取方法研究[J]. 北京邮电大学学报, 2017, 40(1): 111–116.XU Tongle, WANG Yingbo, ZHENG Diankun, et al. Research of the rolling bearing fault signal feature extraction Method based on the LMD-ICA noise reduction[J]. Journal of Beijing University of Posts and Telecommunications, 2017, 40(1): 111–116. 席剑辉, 崔健驰, 蒋丽英. 基于JADE-ICA的滚动轴承多故障信号盲源分离[J]. 振动与冲击, 2017, 36(5): 231–237. doi: 10.13465/j.cnki.jvs.2017.05.037XI Jianhui, CUI Jianchi, and JIANG Liying. JADE-ICA-based blind source separation of multi-fault signals of rolling bearings[J]. Journal of Vibration and Shock, 2017, 36(5): 231–237. doi: 10.13465/j.cnki.jvs.2017.05.037 BELL A J and SEJNOWSKI T J. An information-maximization approach to blind separation and blind deconvolution[J]. Neural Computation, 1995, 7(6): 1129–1159. doi: 10.1162/neco.1995.7.6.1129 CARDOSO J F and LAHELD B H. Equivariant adaptive source separation[J]. IEEE Transactions on Signal Processing, 1996, 44(12): 3017–3030. doi: 10.1109/78.553476 ZHANG Weitao, LOU Shuntian, and FENG Dazheng. Adaptive quasi-newton algorithm for source extraction via CCA approach[J]. IEEE Transactions on Neural Networks and Learning Systems, 2014, 25(4): 677–689. doi: 10.1109/TNNLS.2013.2280285 KARHUNEN J, PAJUNEN P, and OJA E. The nonlinear PCA criterion in blind source separation: Relations with other approaches[J]. Neurocomputing, 1998, 22(1/3): 5–20. doi: 10.1016/s0925-2312(98)00046-0 -
图(4) / 表(3)
计量
- 文章访问数: 2289
- HTML全文浏览量: 441
- PDF下载量: 51
- 被引次数: 0
下载:
