Application of LLTSA Feature Dimensionality Reduction and ELM Model in Fault Diagnosis of Check Valve
-
摘要: 信号的单一特征难以全面反映设备运行状态,而利用多域特征表征设备运行状态时,随着特征维数增加,将引发维数灾难,导致分类器性能退化,降低状态监测模型的辨识性能。针对这一问题,提出线性局部切空间排列(Linear local tangent space alignment,LLTSA)特征降维与极限学习机(Extreme learning machine,ELM)模型的故障诊断方法,利用LLTSA从高维特征空间提取低维流形,实现信号特征的维数约简,保证模型分类性能。该方法首先利用完备总体经验模态分解(Complementary ensemble empirical mode decomposition,CEEMD)对振动信号进行分解,采用相关系数与峭度准则筛选分量,重构得到降噪后的振动信号;然后,计算重构信号的多域特征,并利用LLTSA进行特征维数约简;最后,利用其低维本质特征建立ELM故障诊断模型,监测设备运行状态。高压隔膜泵单向阀运行状态监测实验表明,对振动信号进行特征维数约简,降低特征间的冗余性,可提高ELM模型的故障识别精度。Abstract: It is difficult for a single signal feature to fully reflect the running status of the equipment. When multi-domain features are used to characterize the equipment operating status, dimension catastrophe will occur to cause classifier performance degradation and degrading the performance of the state monitoring model as the feature dimension increases. To solve this problem, a new fault diagnosis method based on the linear local tangent space alignment (LLTSA) feature dimensionality reduction and extreme learning machine (ELM) model is proposed. Using LLTSA to extract low-dimensional manifolds from the high-dimensional feature space reduces the dimensionality of signal features, and ensures model classification performance. In this method, firstly, the vibration signal is decomposed by complementary ensemble empirical mode decomposition (CEEMD), and the correlation coefficient and kurtosis criterion are used to select the components to reconstruct the vibration signal after the noise reduction. Then, calculate the multi-domain features of the reconstructed signal, and use LLTSA to perform feature dimension reduction. Finally, the ELM fault diagnosis model is established with its low-dimensional essential features to monitor the equipment operating status. The monitoring experiment of the check valve of the high pressure diaphragm pump shows that the characteristic dimension of the vibration signal is reduced, and the redundancy between the features is reduced, which can improve the accuracy of the fault recognition of the ELM model.
-
Key words:
- CEEMD /
- LLLTSA feature dimensionality reduction /
- ELM model /
- check valve /
- fault diagnosis /
- vibration signal
-
表 1 7个时域特征数学表达式
特征名称 数学表达式 峰峰值 方差 峭度 均方根 波形因子 裕度因子 脉冲因子 表 2 实验器件型号
实验器件 型号 三缸曲轴驱动活塞式隔膜泵 TZPM 加速度传感器 PCB-ICP 加速度校准器 PCB-394C06 高精度8通道动态数据采集卡 PXIe-3342 控制器 PXI-3050EXT 2.7GHZ 工控机 PXI-9108EXT 8槽PXI机箱 表 3 各IMF分量的相关系数与峭度值
IMF1 IMF2 IMF3 IMF4 IMF5 相关系数 0.22 0.45 0.53 0.64 0.55 峭度值 48.41 5.72 4.89 3.31 4.36 表 4 不同情况下ELM模型测试集结果
处理方式 无处理 LLTSA CEEMD CEEMD+LLTSA 测试集分类结果/% 75.28 85.58 78.25 87.66 -
[1] Ma J, Wu J D, Wang X D. Fault diagnosis method of check valve based on multikernel cost-sensitive extreme learning machine[J]. Complexity, 2017, 2017:8395252 [2] 牟竹青, 冯早, 黄国勇, 等.自适应随机共振和DEMD的单向阀早期故障诊断[J].机械科学与技术, 2018, 37(4):537-544 http://jxkx.cbpt.cnki.net/WKA2/WebPublication/paperDigest.aspx?paperID=29d3d547-6a91-403c-bca4-005c68db3f98Mu Z Q, Feng Z, Huang G Y, et al. Early fault diagnosis of check valve with adaptive stochastic resonance and DEMD[J]. Mechanical Science and Technology for Aerospace Engineering, 2018, 37(4):537-544(in Chinese) http://jxkx.cbpt.cnki.net/WKA2/WebPublication/paperDigest.aspx?paperID=29d3d547-6a91-403c-bca4-005c68db3f98 [3] 郭燕平, 颜文俊, 包哲静, 等.基于经验模态分解和散度指标的风力发电机滚动轴承故障诊断方法[J].电力系统保护与控制, 2012, 40(17):83-87 doi: 10.7667/j.issn.1674-3415.2012.17.016Guo Y P, Yan W J, Bao Z J, et al. Fault diagnosis of bearing in wind turbine based on empirical mode decomposition and divergence index[J]. Power System Protection and Control, 2012, 40(17):83-87(in Chinese) doi: 10.7667/j.issn.1674-3415.2012.17.016 [4] 黄建, 胡晓光, 巩玉楠.基于经验模态分解的高压断路器机械故障诊断方法[J].中国电机工程学报, 2011, 31(12):108-113 http://d.old.wanfangdata.com.cn/Periodical/zgdjgcxb201112018Huang J, Hu X G, Gong Y N. Machinery fault diagnosis of high voltage circuit breaker based on empirical mode decomposition[J]. Proceedings of the CSEE, 2011, 31(12):108-113(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/zgdjgcxb201112018 [5] 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 http://d.old.wanfangdata.com.cn/Periodical/dianzixb201305033 [6] Yeh J R, Shieh J S, Huang N E. Complementary ensemble empirical mode decomposition:a novel noise enhanced data analysis method[J]. Advances in Adaptive Data Analysis, 2010, 2(2):135-156 http://d.old.wanfangdata.com.cn/NSTLQK/NSTL_QKJJ0219457365/ [7] 谈超, 关佶红, 周水庚.增量与演化流形学习综述[J].智能系统学报, 2012, 7(5):377-388 doi: 10.3969/j.issn.1673-4785.201204028Tan C, Guan J H, Zhou S G. Incremental and evolutionary manifold learning:a survey[J]. CAAI Transactions on Intelligent Systems, 2012, 7(5):377-388(in Chinese) doi: 10.3969/j.issn.1673-4785.201204028 [8] Zhuang Z, Ding J M, Tan A C, et al. Fault detection of high-speed train wheelset bearing based on impulse-envelope manifold[J]. Shock and Vibration, 2017, 2017:2102720 [9] Wu J, Wu C Y, Lv Y Q, et al. Design a degradation condition monitoring system scheme for rolling bearing using EMD and PCA[J]. Industrial Management & Data Systems, 2017, 117(4):713-728 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=79002940c21c47eb7663f84a65d3936f [10] Kumar A, Kumar R. Manifold learning using linear local tangent space alignment (LLTSA) algorithm for noise removal in wavelet filtered vibration signal[J]. Journal of Nondestructive Evaluation, 2016, 35(3):50 doi: 10.1007/s10921-016-0366-4 [11] 向丹, 葛爽.基于EMD样本熵-LLTSA的故障特征提取方法[J].航空动力学报, 2014, 29(7):1535-1542 http://d.old.wanfangdata.com.cn/Periodical/hkdlxb201407004Xiang D, Ge S. Method of fault feature extraction based on EMD sample entropy and LLTSA[J]. Journal of Aerospace Power, 2014, 29(7):1535-1542(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/hkdlxb201407004 [12] 苏祖强, 汤宝平, 姚金宝.基于敏感特征选择与流形学习维数约简的故障诊断[J].振动与冲击, 2014, 33(3):70-75 doi: 10.3969/j.issn.1000-3835.2014.03.015Su Z Q, Tang B P, Yao J B. Fault diagnosis method based on sensitive feature selection and manifold learning dimension reduction[J]. Journal of Vibration and Shock, 2014, 33(3):70-75(in Chinese) doi: 10.3969/j.issn.1000-3835.2014.03.015 [13] Luo M, Li C S, Zhang X Y, et al. Compound feature selection and parameter optimization of ELM for fault diagnosis of rolling element bearings[J]. ISA Transactions, 2016, 65:556-566 doi: 10.1016/j.isatra.2016.08.022 [14] 彭进, 王维庆, 王海云, 等.基于EEMD峭度-相关系数准则的多特征量风电机组轴承故障诊断[J].可再生能源, 2016, 34(10):1481-1490 http://d.old.wanfangdata.com.cn/Periodical/ncny201610010Peng J, Wang W Q, Wang H Y, et al. Fault diagnosis method of wind turbine's bearing based on EEMD kurtosis-correlation coefficients criterion and multiple features[J]. Renewable Energy Resources, 2016, 34(10):1481-1490(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/ncny201610010 [15] 张小龙, 张氢, 秦仙蓉, 等.基于ITD-形态滤波和Teager能量谱的轴承故障诊断[J].仪器仪表学报, 2016, 37(4):788-795 doi: 10.3969/j.issn.0254-3087.2016.04.010Zhang X L, Zhang Q, Qin X R, et al. Fault diagnosis method for rolling bearing based on ITD-morphological filter and Teager energy spectrum[J]. Chinese Journal of Scientific Instrument, 2016, 37(4):788-795(in Chinese) doi: 10.3969/j.issn.0254-3087.2016.04.010