Bearing Fault Recognition and Classification Method based on EWT and Multiscale Fuzzy Entropy and VPMCD
-
摘要: 针对振动信号提取轴承故障特征及识别分类的研究方式,提出了一种结合EWT-多尺度模糊熵-VPMCD的方法。首先,运用经验小波变换提取振动信号的模态分量。其次,引入信息论中的模糊熵算法,并加以多尺度粗粒度划分得到多尺度模糊熵特征描述。然后,用VPMCD对特征向量进行自适应选择预测模型训练。最终通过实验表明:模态分量多尺度模糊熵能够有效描述故障特征;VPMCD在少训练样本情况下获得了最低90%的分类准确率,相较一些常用的分类方法有着更好的性能表现。Abstract: Aiming at the research methods of extracting bearing fault characteristics, identifying and classifying vibration signals, a new method combining EWT、multi-scale fuzzy entropy and VPMCD algorithms is proposed in the thesis. Firstly, the modal components of the vibration signal are extracted by the empirical wavelet transform (EWT). Secondly, the fuzzy entropy algorithm in information theory is introduced, and multi-scale coarse-grained partitioning method is used to obtain the multi-scale fuzzy entropy feature description. Then, VPMCD algorithm is used to train the model by adaptively selecting prediction model. The experiments show that multi-scale fuzzy entropy of modal components can effectively describe the fault features. VPMCD achieves a minimum classification accuracy of 90% with a small number of training samples, which has better performance than some common classification methods.
-
Key words:
- vibration signals /
- fault characteristics /
- EWT /
- multi-scale fuzzy entropy /
- VPMCD
-
表 1 模型表达式
模型 预测特征值${f_i}$表达式 QI ${f_i} = {b_0} + \displaystyle \sum\limits_{j = 1}^{p - 1} {b_j^{(1)}{f_n}} + \displaystyle \sum\limits_{t = 1}^{p - 1} {b_j^{(2)}f_{{n_t}}^2 + }$${\displaystyle \sum\limits_{l = 1}^{p - 2} {\displaystyle \sum\limits_{t = l + 1}^{p - 1} {{b_{tl}}{f_{{n_t}}}} } } {f_{{n_l}}}\quad n,{n_t},{n_l} \ne i,n,{n_t},{n_l} \in [1,p] $ Q ${f_i} = {b_0} + \displaystyle \sum\limits_{j = 1}^{p - 1} {b_j^{(1)}{f_n}} + \displaystyle \sum\limits_{t = 1}^{p - 1} {b_j^{(2)}f_{{n_t}}^2} \quad n,{n_t} \ne i,n,{n_t} \in [1,p] $ LI ${f_i} = {b_0} + \displaystyle \sum\limits_{j = 1}^{p - 1} {{b_j}{f_n}} + \displaystyle \sum\limits_{l = 1}^{p - 2} {\displaystyle \sum\limits_{t = l + 1}^{p - 1} {{b_{tl}}{f_{{n_t}}}{f_{{n_l}}}} } \quad n,{n_t},{n_l} \ne i,n,{n_t},{n_l} \in [1,p] $ L ${f_i} = {b_0} + \displaystyle \sum\limits_{j = 1}^{p - 1} {{b_j}{f_n}} \quad n \ne i,n \in [1,p]$ 表 2 不同故障的模态分量模糊熵示例
故障
类型IMF1
模糊熵IMF2
模糊熵IMF3
模糊熵IMF4
模糊熵IMF5
模糊熵正常 0.249 0.273 0.248 0.223 0.077 滚子故障 0.126 0.307 0.212 0.185 0.133 内圈故障 0.171 0.283 0.240 0.215 0.174 外圈故障 0.257 0.268 0.295 0.166 0.157 表 3 内圈故障VPM系数及模型类别
系数(列) 模态分量(行) IMF1 IMF2 IMF3 IMF4 IMF5 $ {C}_{1} $ 0.649 0.200 1.413 −1.570 −0.383 $ {C}_{2} $ 0.363 2.732 12.820 −8.682 −1.238 $ {C}_{3} $ −0.460 −1.585 −8.048 0.576 −1.801 $ {C}_{4} $ 1.721 −0.790 3.207 −6.847 −5.942 $ {C}_{5} $ −0.701 1.237 −1.481 11.351 6.232 $ {C}_{6} $ 0.612 −0.754 4.852 −4.391 0.651 $ {C}_{7} $ 0.228 0.050 −6.364 2.894 0.107 $ {C}_{8} $ 5.891 4.123 −11.020 0.899 2.815 $ {C}_{9} $ −4.092 −0.056 1.278 1.683 −0.510 $ {C}_{10} $ 0.227 3.931 −1.977 −18.970 −2.766 $ {C}_{11} $ −3.874 −1.733 1.922 2.281 8.196 $ {C}_{12} $ 5.367 1.090 25.518 31.895 3.525 $ {C}_{13} $ 1.485 2.157 9.924 −6.138 −7.819 $ {C}_{14} $ −0.923 −7.421 −6.590 −23.886 −6.006 $ {C}_{15} $ 0.771 −0.567 6.054 −1.610 1.795 模型类别 QI QI QI QI QI 表 4 输出结果示例
特征向量(列) 样本类型(行) 正常
样本滚子故障
样本内圈故障
样本外圈故障
样本正常 0.0253 1.5615 1.2281 0.5184 滚子故障 0.2736 0.0045 0.1846 0.4627 内圈故障 1.0799 0.2244 0.0221 0.1114 外圈故障 0.2029 0.0101 0.1063 0.0329 输出结果 正常 滚子故障 内圈故障 外圈故障 分类结果 正确 正确 正确 正确 -
[1] 池永为. 滚动轴承故障的振动特性分析与智能诊断方法研究[D]. 杭州: 浙江大学, 2018: 4CHI Y W. Analysis of vibration characteristic and intelligent fault diagnosis of rolling bearings[D]. Hangzhou: Zhejiang University, 2018: 4 (in Chinese) [2] 李政, 张炜, 明安波, 等. 基于IEWT和MCKD的滚动轴承故障诊断方法[J]. 机械工程学报, 2019, 55(23): 136-146LI Z, ZHANG W, MING A B, et al. A novel fault diagnosis method based on improved empirical wavelet transform and maximum correlated kurtosis deconvolution for rolling element bearing[J]. Journal of Mechanical Engineering, 2019, 55(23): 136-146 (in Chinese) [3] 谭媛, 孙文磊, 温广瑞, 等. 基于EWT-MDS的风力机轴承劣化趋势识别及故障诊断[J]. 太阳能学报, 2018, 39(12): 3511-3518TAN Y, SUN W L, WEN G R, et al. Wind turbine bearing deterioration trend identification and fault diagnosis based on EWT-MDS[J]. Acta Energiae Solaris Sinica, 2018, 39(12): 3511-3518 (in Chinese) [4] 周奇才, 沈鹤鸿, 赵炯, 等. 基于改进堆叠式循环神经网络的轴承故障诊断[J]. 同济大学学报, 2019, 47(10): 1500-1507 doi: 10.11908/j.issn.0253-374x.2019.10.016ZHOU Q C, SHEN H H, ZHAO J, et al. Bearing fault diagnosis based on improved stacked recurrent neural network[J]. Journal of Tongji University, 2019, 47(10): 1500-1507 (in Chinese) doi: 10.11908/j.issn.0253-374x.2019.10.016 [5] 郭伟超, 赵怀山, 李成, 等. 基于小波包能量谱与主成分分析的轴承故障特征增强诊断方法[J]. 兵工学报, 2019, 40(11): 2370-2377 doi: 10.3969/j.issn.1000-1093.2019.11.022GUO W C, ZHAO H S, LI C, et al. Fault feature enhancement method for rolling bearing fault diagnosis based on wavelet packet energy spectrum and principal component analysis[J]. Acta Armamentarii, 2019, 40(11): 2370-2377 (in Chinese) doi: 10.3969/j.issn.1000-1093.2019.11.022 [6] 李晨晨, 韩清鹏, 李天成, 等. EEMD和分形组合技术对ECS涡轮轴承故障特征提取的研究[J]. 机械科学与技术, 2019, 38(1): 37-43LI C C, HAN Q P, LI T C, et al. Study on fault feature extraction of ECS turbine bearing by combination of EEMD and correlation dimension[J]. Mechanical Science and Technology for Aerospace Engineering, 2019, 38(1): 37-43 (in Chinese) [7] GILLES J. Empirical wavelet transform[J]. IEEE Transactions on Signal Processing, 2013, 61(16): 3999-4010 doi: 10.1109/TSP.2013.2265222 [8] 姜保军, 曹浩. 基于小波分解和样本熵的GA-SVM齿轮箱故障诊断[J]. 组合机床与自动化加工技术, 2019(11): 78-82JIANG B J, CAO H. Fault diagnosis of GA-SVM gearbox based on wavelet decomposition and sample entropy[J]. Modular Machine Tool & Automatic Manufacturing Technique, 2019(11): 78-82 (in Chinese) [9] 张建财, 高军伟. 基于变分模态分解和多尺度排列熵的滚动轴承故障诊断[J]. 噪声与振动控制, 2019, 39(6): 181-186 doi: 10.3969/j.issn.1006-1355.2019.06.032ZHANG J C, GAO J W. Fault diagnosis of train rolling bearings based on variational modal decomposition and multi-scale permutation entropy[J]. Noise and Vibration Control, 2019, 39(6): 181-186 (in Chinese) doi: 10.3969/j.issn.1006-1355.2019.06.032 [10] 杨新, 麻哲瑞, 申赫男, 等. 基于多尺度特征能量-核极限学习机的双循环流化床气流堵塞故障智能诊断[J]. 化工学报, 2019, 70(7): 2616-2625YANG X, MA Z R, SHEN H N, et al. Fault diagnosis of airflow jamming fault in double circulating fluidized bed based on multi-scale feature energy and KELM[J]. CIESC Journal, 2019, 70(7): 2616-2625 (in Chinese) [11] 者娜, 杨剑锋, 刘文彬, 等. KPCA和改进SVM在滚动轴承剩余寿命预测中的应用研究[J]. 机械设计与制造, 2019(11): 1-4, 8 doi: 10.3969/j.issn.1001-3997.2019.11.001ZHE N, YANG J F, LIU W B, et al. Research on application of KPCA and improved SVM in residual life prediction of rolling bearings[J]. Machinery Design & Manufacture, 2019(11): 1-4, 8 (in Chinese) doi: 10.3969/j.issn.1001-3997.2019.11.001 [12] 陈博, 陈光雄. 基于MEEMD和GA-SVM的列车车轮多边形故障识别方法[J]. 噪声与振动控制, 2018, 38(3): 157-161, 197 doi: 10.3969/j.issn.1006-1355.2018.03.030CHEN B, CHEN G X. Fault diagnosis method of wheel polygonization of trains based on MEEMD and GA-SVM[J]. Noise and Vibration Control, 2018, 38(3): 157-161, 197 (in Chinese) doi: 10.3969/j.issn.1006-1355.2018.03.030 [13] 陈剑挺, 吴志国, 叶贞成, 等. 基于收缩极限学习机的故障诊断鲁棒方法[J]. 计算机工程与设计, 2020, 41(1): 208-213CHEN J T, WU Z G, YE Z C, et al. Contractive-ELM based robust method for fault diagnosis[J]. Computer Engineering and Design, 2020, 41(1): 208-213 (in Chinese) [14] RAGHURAJ R, LAKSHMINARAYANAN S. VPMCD: variable interaction modeling approach for class discrimination in biological systems[J]. FEBS Letters, 2007, 581(5): 826-830 doi: 10.1016/j.febslet.2007.01.052 [15] WANG L, GE K G, WU J Y, et al. A novel approach for the pattern recognition of hand movements based on EMG and VPMCD[J]. Journal of Mechanics in Medicine and Biology, 2018, 18(1): 1750115 doi: 10.1142/S0219519417501159 [16] LUO S R, CHENG J S. VPMCD based novelty detection method on and its application to fault identification for local characteristic-scale decomposition[J]. Cluster Computing, 2017, 20(4): 2955-2965 doi: 10.1007/s10586-017-0932-2 [17] 郑近德, 陈敏均, 程军圣, 等. 多尺度模糊熵及其在滚动轴承故障诊断中的应用[J]. 振动工程学报, 2014, 27(1): 145-151 doi: 10.3969/j.issn.1004-4523.2014.01.020ZHENG J D, CHEN M J, CHENG J S, et al. Multiscale fuzzy entropy and its application in rolling bearing fault diagnosis[J]. Journal of Vibration Engineering, 2014, 27(1): 145-151 (in Chinese) doi: 10.3969/j.issn.1004-4523.2014.01.020