Adaptive Multiclass Relevance Vector Machines and its Application to Fault Recognition of Rolling Bearing
-
摘要: 提出了一种基于自适应相关向量机(Adaptive multiclass relevance vector machines,A-MRVM)的滚动轴承故障识别方法,该方法利用遗传算法对多分类相关向量机核函数参数进行优化,依据故障样本自身特性自适应地选取最优核参数,克服核参数人为选取的不确定性,从而构建基于自适应多分类相关向量机的故障识别模型。将该故障识别模型应用于滚动轴承故障识别中,分别提取滚动轴承振动信号小波包能量及EEMD(Ensemble empirical mode decomposition)能量作为故障特征进行故障识别,并与其它方法进行实验对比研究。实验结果表明,所提方法不仅能有效识别出故障类型,且具有较高的故障识别模型构建效率,验证了所提方法的可行性及优越性。同时,该方法也能对故障类型发生的可能性进行评估,为分析滚动轴承故障类型提供更多的参考信息。Abstract: A novel method of rolling bearing fault detection based on adaptive multiclass relevance vector machines, (A-MRVM) was proposed. Genetic algorithm (GA) was used to optimize MRVM kernel function parameters adaptively, according to the characteristics of fault samples, to solve the uncertainty of artificial parameter selections. Therefore, a novel fault recognition model was constructed based on adaptive multiclass relevance vector machine. The fault identification model was applied in the rolling bearing fault identification experiment. The wavelet packet energies and ensemble empirical mode decomposition (EEMD) energies of rolling bearing vibration signal were extracted respectively as fault features. The fault recognition contrast experiments were implemented using different identification methods. Experimental results indicated that the proposed method can identify the fault type effectively, and verified the method was feasible and superior. Additionally, the method can give more available data to evaluate the possibility of fault type.
-
Key words:
- fault detection /
- rolling bearing /
- multiclass relevance vector machines /
- adaptive /
- genetic algorithms /
- EEMD /
- experiments /
- wavelet
-
表 1 小波包能量特征下的滚动轴承故障识别实验结果
模型 训练样本数/个 测试样本数/个 训练时间/s 向量个数/个 平均准确率/% 100 100 82.3 16 100 A-MRVM 80 120 81.2 14 99.96 60 140 80.4 12 100 100 100 178.6 27 97.15 CV-MRVM 80 120 177.1 23 96.83 60 140 176.4 22 97.12 100 100 189.3 18 100 A-RVM 80 120 186.8 15 99.91 60 140 186.4 12 99.93 100 100 7.7 66 93.37 CV-SVM 80 120 6.7 54 94.33 60 140 6.2 47 94.71 表 2 EEMD能量特征下的滚动轴承故障识别实验结果
模型 训练样本数/个 测试样本数/个 训练时间/s 向量个数/个 平均准确率/% 100 100 84.3 12 100 A-MRVM 80 120 83.5 11 100 60 140 82.6 9 100 100 100 182.5 20 100 CV-MRVM 80 120 181.7 23 100 60 140 180.4 21 100 100 100 190.6 14 100 A-RVM 80 120 189.7 12 100 60 140 188.6 11 100 100 100 7.9 57 99.91 CV-SVM 80 120 7.1 52 99.93 60 140 6.7 47 99.91 -
[1] 杨武成.阶次跟踪和Hilbert包络解调在滚动轴承故障诊断中应用[J].机械科学与技术, 2017, 36(12):1873-1876 doi: 10.13433/j.cnki.1003-8728.2017.1212Yang W C. Application of order tracking and Hilbert envelope demodulation in rolling bearing fault diagnosis[J]. Mechanical Science and Technology for Aerospace Engineering, 2017, 36(12):1873-1876(in Chinese) doi: 10.13433/j.cnki.1003-8728.2017.1212 [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] Zeng M, Yang Y, Luo S R, et al. One-class classification based on the convex hull for bearing fault detection[J]. Mechanical Systems and Signal Processing, 2016, 81:274-293 doi: 10.1016/j.ymssp.2016.04.001 [4] Chen J, Randall R B. Intelligent diagnosis of bearing knock faults in internal combustion engines using vibration simulation[J]. Mechanism and Machine Theory, 2016, 104:161-176 doi: 10.1016/j.mechmachtheory.2016.05.022 [5] Wang B, Liu S L, Zhang H L, et al. Fault diagnosis of rolling bearing based on relevance vector machine and kernel principal component analysis[J]. Journal of Vibroengineering, 2014, 16(1):57-69 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=WFHYXW575964 [6] 陈法法, 杨晶晶, 肖文荣, 等.Adaboost-SVM集成模型的滚动轴承早期故障诊断[J].机械科学与技术, 2018, 37(2):237-243 doi: 10.13433/j.cnki.1003-8728.2018.0212Chen F F, Yang J J, Xiao W R, et al. Early fault diagnosis of rolling bearing based on ensemble model of Adaboost SVM[J]. Mechanical Science and Technology for Aerospace Engineering, 2018, 37(2):237-243(in Chinese) doi: 10.13433/j.cnki.1003-8728.2018.0212 [7] Zhu K H, Song X G, Xue D X. A roller bearing fault diagnosis method based on hierarchical entropy and support vector machine with particle swarm optimization algorithm[J]. Measurement, 2014, 47:669-675 doi: 10.1016/j.measurement.2013.09.019 [8] Tipping M E. Sparse Bayesian learning and the relevance vector machine[J]. The Journal of Machine Learning Research, 2001, 1:211-244 http://d.old.wanfangdata.com.cn/Periodical/dsjs201802008 [9] 王波, 刘树林, 蒋超, 等.基于量子遗传算法优化RVM的滚动轴承智能故障诊断[J].振动与冲击, 2015, 34(17):207-212 http://d.old.wanfangdata.com.cn/Periodical/zdycj201517034Wang B, Liu S L, Jiang C, et al. Rolling bearings' intelligent fault diagnosis based on RVM optimized with Quantum genetic algorithm[J]. Journal of Vibration and Shock, 2015, 34(17):207-212(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/zdycj201517034 [10] 周勇, 何创新.基于独立特征选择与相关向量机的变载荷轴承故障诊断[J].振动与冲击, 2012, 31(3):157-161 doi: 10.3969/j.issn.1000-3835.2012.03.032Zhou Y, He C X. Bearing fault diagnsis under varying load conditions based on individual feature selection and relevance vector machine[J]. Journal of Vibration and Shock, 2012, 31(3):157-161(in Chinese) doi: 10.3969/j.issn.1000-3835.2012.03.032 [11] He C, Liu C, Li Y, et al. Intelligent fault diagnosis of rotating machinery based on multiple relevance vector machines with variance radial basis function kernel[J]. Proceedings of the Institution of Mechanical Engineers, Part C:Journal of Mechanical Engineering Science, 2011, 225(7):1718-1729 doi: 10.1177/0954406211400526 [12] Psorakis I, Damoulas T, Girolami M A. Multiclass relevance vector machines:sparsity and accuracy[J]. IEEE Transactions on Neural Networks, 2010, 21(10):1588-1598 doi: 10.1109/TNN.2010.2064787 [13] Gholami B, Haddad M M, Tannenbaum A R. Relevance vector machine learning for neonate pain intensity assessment using digital imaging[J]. IEEE transactions on Biomedical Engineering, 2010, 57(6):1457-1466 doi: 10.1109/TBME.2009.2039214 [14] Damoulas T, Girolami M A. Probabilistic multi-class multi-kernel learning:on protein fold recognition and remote homology detection[J]. Bioinformatics, 2008, 24(10):1264-1270 doi: 10.1093/bioinformatics/btn112 [15] Damoulas T, Girolami M A. Combining feature spaces for classification[J]. Pattern Recognition, 2009, 42(11):2671-2683 doi: 10.1016/j.patcog.2009.04.002