SE-KSD Optimized FRFT-VMD Bearing Fault Diagnosis Method
-
摘要: 提出了一种基于样本熵(SE)和峭度均方差(KSD)优化的分数阶变分模态分解(FRFT-VMD)的故障特征提取方法,同时结合随机森林(RF)分类器对故障进行自动识别分类。针对分数阶傅里叶变换中阶次选择对于数据可分性影响较大的问题,提出利用搜寻样本熵最小值来得到分数阶最优阶次,使得数据中重叠交叉部分在分数域内能更好分离。同时利用峭度均方差准则寻找变分模态分解最优参数使变分模态分解效果更好。通过数据库和实测数据的研究结果表明,该方法提取的信号包含更多和更明显的故障特征频率,大幅度提高了滚动轴承不同状态的故障诊断准确率。
-
关键词:
- 分数阶Fourier变换 /
- 变分模态分解 /
- 特征提取 /
- 故障诊断
Abstract: A fault feature extraction method based on sample entropy (SE) and kurtosis standard deviation (KSD) optimization of fractional variational modal decomposition (FRFT-VMD) is proposed, and a random forest (RF) classifier is combined to perform fault detection. Automatic recognition of classification. Aiming at the problem that the choice of order in the fractional Fourier transform has a greater impact on the separability of data, it is proposed to search for the minimum entropy of the sample to obtain the optimal order of the fractional order. Makes the overlapped part of the data better separated in the score domain. At the same time, the kurtosis standard deviation criterion is used to find the optimal parameters of the variational modal decomposition to make the effect of the variational modal decomposition better. The research results based on database data and measured data show that the signals extracted by this method contain more and more obvious fault characteristic frequencies, which greatly improves the fault diagnosis accuracy of rolling bearings in different states. -
表 1 轴承不同状态数据
Table 1. Different status data of bearings
轴承
状态故障直径/
mm电机转速/
(r·min−1)轴承型号
(驱动端)采样频率/
Hz正常状态 - 1797 SKF6205 12 内圈故障 0.1778 1797 SKF6205 12 滚动体故 0.1778 1797 SKF6205 12 外圈故障 0.1778 1797 SKF6205 12 表 2 各状态轴承对应编号
Table 2. Corresponding numbers of bearings in each state
轴承状态 训练样本数 测试样本数 故障编号 正常状态 32 8 1 内圈故障 32 8 2 滚动体故障 32 8 3 外圈故障 32 8 4 表 3 基于SE-KSD优化的FRFT-VMD-RF分类结果汇总
Table 3. Summary of FRFT-VMD-RF classification results based on SE-KSD optimization
轴承
状态缺陷大小/
mm编号 测试
样本数正确
分类数准确率/% 正常状态 - 1 8 8 100 内圈故障 0.1778 2 8 8 100 滚动体 0.1778 3 8 8 100 外圈故障 0.1778 4 8 7 87.5 汇总 - - 32 31 96.875 表 4 基于VMD-RF分类结果汇总
Table 4. Summary of VMD-RF classification results
轴承
状态缺陷大小/
mm编号 测试
样本数正确
分类数准确率/% 正常状态 - 1 8 4 50 内圈故障 0.1778 2 8 8 100 滚动体 0.1778 3 8 8 100 外圈故障 0.1778 4 8 5 62.5 汇总 - - 32 25 78.1 -
[1] 詹君, 程龙生, 彭宅铭. 基于VMD和改进多分类马田系统的滚动轴承故障智能诊断[J]. 振动与冲击, 2020, 39(2): 32-39. doi: 10.13465/j.cnki.jvs.2020.02.005ZHAN J, CHENG L S, PENG Z M. Intelligent fault diagnosis of rolling bearings based on the VMD and improved multi-classification mahalanobis taguchi system[J]. Journal of Vibration and Shock, 2020, 39(2): 32-39. (in Chinese) doi: 10.13465/j.cnki.jvs.2020.02.005 [2] OU L, YU D J, YANG H J. A new rolling bearing fault diagnosis method based on GFT impulse component extraction[J]. Mechanical Systems and Signal Processing, 2016, 81: 162-182. doi: 10.1016/j.ymssp.2016.03.009 [3] HUANG N E, SHEN Z, LONG S R, et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis[J]. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 1998, 454(1971): 903-995. [4] WU Z H, HUANG N E. Ensemble empirical mode decomposition: a noise-assisted data analysis method[J]. Advances in Adaptive Data Analysis, 2009, 1(1): 1-41. doi: 10.1142/S1793536909000047 [5] DRAGOMIRETSKIY K, ZOSSO D. Variational mode decomposition[J]. IEEE Transactions on Signal Processing, 2014, 62(3): 531-544. doi: 10.1109/TSP.2013.2288675 [6] 章蕾, 许述文, 刘韬. 基于FRFT的非线性调频信号检测[J]. 电子科技, 2010, 23(2): 68-71. doi: 10.3969/j.issn.1007-7820.2010.02.021ZHANG L, XU S W, LIU T. Detection of nonlinear signals using FRFT[J]. Electronic Science and Technology, 2010, 23(2): 68-71. (in Chinese) doi: 10.3969/j.issn.1007-7820.2010.02.021 [7] 刘吉顺, 杨丽荣, 罗小燕, 等. CEEMDAN和样本熵相结合的球磨机负荷识别方法[J]. 机械科学与技术, 2021, 40(2): 249-256. doi: 10.13433/j.cnki.1003-8728.20200037LIU J S, YANG L R, LUO X Y, et al. A ball mill load state identification method in combination with CEEMDAN and sample entropy[J]. Mechanical Science and Technology for Aerospace Engineering, 2021, 40(2): 249-256. (in Chinese) doi: 10.13433/j.cnki.1003-8728.20200037 [8] 郭学卫, 申永军, 杨绍普. 基于样本熵和分数阶傅里叶变换的滚动轴承故障特征提取[J]. 振动与冲击, 2017, 36(18): 65-69. doi: 10.13465/j.cnki.jvs.2017.18.010GUO X W, SHEN Y J, YANG S P. Application of sample entropy and fractional Fourier transform in the fault diagnosis of rolling bearings[J]. Journal of Vibration and Shock, 2017, 36(18): 65-69. (in Chinese) doi: 10.13465/j.cnki.jvs.2017.18.010 [9] LIU X Y, WAN H, SHANG Z G, et al. Automatic extracellular spike denoising using wavelet neighbor coefficients and level dependency[J]. Neurocomputing, 2015, 149: 1407-1414. doi: 10.1016/j.neucom.2014.08.055 [10] 宋晓霞, 向北平, 倪磊, 等. 基于样本熵的改进空域相关去噪算法[J]. 机械设计与研究, 2018, 34(4): 31-36. doi: 10.13952/j.cnki.jofmdr.2018.0140SONG X X, XIANG B P, NI L, et al. Research of Improved spatial correlation denoising algorithm based on sample entropy[J]. Machine Design & Research, 2018, 34(4): 31-36. (in Chinese) doi: 10.13952/j.cnki.jofmdr.2018.0140 [11] 张树, 刘德平. BFA优化VMD参数的轴承故障诊断[J]. 组合机床与自动化加工技术, 2020(5): 45-47. doi: 10.13462/j.cnki.mmtamt.2020.05.011ZHANG S, LIU D P. Bearing fault diagnosis based on BFA optimized VMD parameters[J]. Modular Machine Tool & Automatic Manufacturing Technique, 2020(5): 45-47. (in Chinese) doi: 10.13462/j.cnki.mmtamt.2020.05.011 [12] 黄大荣, 柯兰艳, 林梦婷, 等. 一种参数优化VMD多尺度熵的轴承故障诊断新方法[J]. 控制与决策, 2020, 35(7): 1631-1638. doi: 10.13195/j.kzyjc.2018.1598HUANG D R, KE L Y, LIN M T, et al. A new fault diagnosis approach for bearing based on multi-scale entropy of the optimized VMD[J]. Control and Decision, 2020, 35(7): 1631-1638. (in Chinese) doi: 10.13195/j.kzyjc.2018.1598 [13] 王奉涛, 柳晨曦, 张涛, 等. 基于k值优化VMD的滚动轴承故障诊断方法[J]. 振动、测试与诊断, 2018, 38(3): 540-547.WANG F T, LIU C X, ZHANG T, et al. Fault diagnosis of rolling bearing based on k-optimized VMD[J]. Journal of Vibration, Measurement & Diagnosis, 2018, 38(3): 540-547. (in Chinese) [14] 张宸宸. 基于VMD的超声无损检测缺陷识别方法研究[D]. 大连: 大连海洋大学, 2020.ZHANG C C. Research on defect recognition method of ultrasonic nondestructive testing based on VMD[D]. Dalian: Dalian Ocean University, 2020. (in Chinese) [15] 杨伟, 王红军. 基于VMD共振稀疏分解的滚动轴承故障诊断[J]. 电子测量与仪器学报, 2018, 32(9): 20-27. doi: 10.13382/j.jemi.2018.09.004YANG W, WANG H J. Fault diagnosis of rolling dearing based on VMD and resonance sparse decomposition[J]. Journal of Electronic Measurement and Instrumentation, 2018, 32(9): 20-27. (in Chinese) doi: 10.13382/j.jemi.2018.09.004 [16] 张琛, 赵荣珍, 邓林峰, 等. 基于SVD-EEMD和TEO的滚动轴承弱故障特征提取[J]. 振动、测试与诊断, 2019, 39(4): 720-726.ZHANG C, ZHAO R Z, DENG L F, et al. Weak fault feature extraction method for rolling bearings based on SVD-EEMD and TEO energy spectrum[J]. Journal of Vibration, Measurement & Diagnosis, 2019, 39(4): 720-726. (in Chinese) [17] 刘泽锐, 邢济收, 王红军, 等. 基于VMD与快速谱峭度的滚动轴承故障诊断[J]. 电子测量与仪器学报, 2021, 35(2): 73-79. doi: 10.13382/j.jemi.B2003673LIU Z R, XING J S, WANG H J, et al. Fault diagnosis of rolling bearings based on VMD and fast spectral kurtosis[J]. Journal of Electronic Measurement and Instrumentation, 2021, 35(2): 73-79. (in Chinese) doi: 10.13382/j.jemi.B2003673 [18] 吴海滨, 陈寅生, 张庭豪, 等. 改进多尺度幅值感知排列熵与随机森林结合的滚动轴承故障诊断[J]. 光学 精密工程, 2020, 28(3): 621-631. doi: 10.3788/OPE.20202803.0621WU H B, CHEN Y S, ZHANG T H, et al. Rolling bearing fault diagnosis by improved multiscale amplitude-aware permutation entropy and random forest[J]. Optics and Precision Engineering, 2020, 28(3): 621-631. (in Chinese) doi: 10.3788/OPE.20202803.0621 [19] 温江涛, 周熙楠. 模糊粒化非监督学习结合随机森林融合的旋转机械故障诊断[J]. 机械科学与技术, 2018, 37(11): 1722-1730. doi: 10.13433/j.cnki.1003-8728.20180069WEN J T, ZHOU X N. Fault diagnosis of rotating machinery in combination with unsupervised learning of fuzzy granulation and random forest fusion[J]. Mechanical Science and Technology for Aerospace Engineering, 2018, 37(11): 1722-1730. (in Chinese) doi: 10.13433/j.cnki.1003-8728.20180069