A Fault Feature Extraction Method Combining Variational Mode Decomposition and Frequency Domain Integration
-
摘要: 在对振动加速度信号进行积分时, 信号中存在的不可避免的直流分量以及积分过程中产生的趋势项和误差行为, 使得积分结果的精度大大降低, 尤其是二次积分后, 信号的频域特性丢失严重。针对此问题, 提出了一种基于变分模态分解(Variational mode decomposition, VMD)和频域积分相结合的故障特征提取方法。首先基于最大峭度准则对加速度信号进行变分模态分解, 根据皮尔逊相关系数法从若干分量中选取相关系数最大的分量作为最优分量, 最后对最优分量进行频域二次积分, 得到振动位移信号。仿真及实验结果表明该方法有效降低了趋势项及噪声等带来的干扰, 提高了故障信息的辨识精度, 具有较好的优越性。Abstract: Due to the inevitable DC component in the signal and the trend term and error behavior generated during the integration process, the accuracy of the integration result is greatly reduced when integrating the signal of vibration acceleration. Especially after the secondary integration, the frequency domain characteristics of the displacement signal are lost seriously. In order to solve this problem, a fault feature extraction method combining variational mode decomposition (VMD) and frequency domain integration is proposed in this paper. Firstly, the acceleration signal is decomposed by VMD, and then the component with the largest correlation coefficient is selected as the optimal component from several components according to the pearson correlation coefficient method. Finally, the optimal component is subjected to frequency domain quadratic integration to obtain the vibration displacement signal. Simulation and experimental results show that the method can effectively reduce the interference caused by trend items and noise, and improve the accuracy of fault information, which has better superiority.
-
表 1 K=7时, 各分量相关系数表
IMF分量 相关系数 IMF1 0.746 8 IMF2 0.538 0 IMF3 0.064 4 IMF4 0.017 2 IMF5 0.004 2 IMF6 0.003 2 IMF7 0.003 4 表 2 误差评价表
名称 平均峰值误差 平均最大相对误差 平方和误差 传统积分 1.013 2 3.001 8 3.033 0 VMD分解积分 0.029 3 1.969 0 0.232 6 表 3 6308轴承参数表
参数及单位 参数值 外径/mm 90 内径/mm 40 滚道节径/mm 65.5 滚珠节径/mm 15 滚珠个数 8 接触角 0 表 4 K=3时, 各分量相关系数表
IMF分量 相关系数 IMF1 0.131 6 IMF2 0.038 2 IMF3 0.799 5 -
[1] 王济, 胡晓. MATLAB在振动信号处理中的应用[M]. 北京: 中国水利水电出版社, 知识产权出版社, 2006WANG J, HU X. Application of MATLAB in vibration signal processing[M]. Beijing: China Water & Power Press, Intellectual Property Publishing House, 2006 (in Chinese) [2] 程启清, 王建生, 王书付. LabVIEW和MATLAB混合编程的时频域积分方法研究[J]. 机械工程与自动化, 2017(2): 69-71 doi: 10.3969/j.issn.1672-6413.2017.02.029CHENG Q Q, WANG J S, WANG S F. Time-domain and frequency-domain integration methods based on LabVIEW and MATLAB mixed-language programming[J]. Mechanical Engineering & Automation, 2017(2): 69-71 (in Chinese) doi: 10.3969/j.issn.1672-6413.2017.02.029 [3] 邹学利, 翟疆, 张生鹏. 随机数据积分的误差分析及改进方法[J]. 环境技术, 2018, 36(6): 87-93 doi: 10.3969/j.issn.1004-7204.2018.06.020ZOU X L, ZHAI J, ZHANG S P. Error analysis and improvement method of random data integration[J]. Environmental Technology, 2018, 36(6): 87-93 (in Chinese) doi: 10.3969/j.issn.1004-7204.2018.06.020 [4] 许晓红, 倪春锋, 颜晓青. 基于组合辛普森积分的振动测试时域积分方法[J]. 计量学报, 2020, 41(6): 704-709 doi: 10.3969/j.issn.1000-1158.2020.06.12XU X H, NI C F, YAN X Q. Time domain integral method for vibration test based on combined simpson integral[J]. Acta Metrologica Sinica, 2020, 41(6): 704-709 (in Chinese) doi: 10.3969/j.issn.1000-1158.2020.06.12 [5] 陈道云, 孙守光, 李强. 基于经验模态分解及时域积分的轨道高低不平顺获得方法[J]. 铁道机车车辆, 2015, 35(5): 9-14 doi: 10.3969/j.issn.1008-7842.2015.05.03CHEN D Y, SUN S G, LI Q. An acquired method of track vertical profile irregularity based on EMD and time domain integral[J]. Railway Locomotive & Car, 2015, 35(5): 9-14 (in Chinese) doi: 10.3969/j.issn.1008-7842.2015.05.03 [6] 王万金, 张志国, 徐洪洲. 含零漂的遥测加速度振动信号时域积分方法研究[J]. 计算机测量与控制, 2018, 26(9): 255-258 doi: 10.16526/j.cnki.11-4762/tp.2018.09.054WANG W J, ZHANG Z G, XU H Z. Research of telemetry acceleration vibration signals with zero drift in time domain integration[J]. Computer Measurement & Control, 2018, 26(9): 255-258 (in Chinese) doi: 10.16526/j.cnki.11-4762/tp.2018.09.054 [7] 杨贵春, 黄树和, 薛大江, 等. 基频对发动机振动信号积分影响研究[J]. 内燃机, 2017(5): 15-18 doi: 10.3969/j.issn.1000-6494.2017.05.005YANG G C, HUANG S H, XUE D J, et al. Effect of fundamental frequency on the integral of the vibration signal of engine[J]. Internal Combustion Engines, 2017(5): 15-18 (in Chinese) doi: 10.3969/j.issn.1000-6494.2017.05.005 [8] MUGˇAN A. Frequency domain analysis of time integration algorithms[J]. Computer Methods in Applied Mechanics and Engineering, 2001, 190(43-44): 5777-5793 doi: 10.1016/S0045-7825(01)00197-9 [9] BRANDT A, BRINCKER R. Integrating time signals in frequency domain-comparison with time domain integration[J]. Measurement, 2014, 58: 511-519 [10] ZHU H, ZHOU Y J, HU Y M. Displacement reconstruction from measured accelerations and accuracy control of integration based on a low-frequency attenuation algorithm[J]. Soil Dynamics and Earthquake Engineering, 2020, 133: 106122 doi: 10.1016/j.soildyn.2020.106122 [11] 方新磊, 郝伟, 陈宏. 基于频域滤波的加速度信号处理[J]. 仪表技术与传感器, 2012(4): 94-96 https://www.cnki.com.cn/Article/CJFDTOTAL-YBJS201204035.htmFANG X L, HAO W, CHEN H. Acceleration signal processing based on frequency domain filtering[J]. Instrument Technique and Sensor, 2012(4): 94-96 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-YBJS201204035.htm [12] 胡玉梅, 周英杰, 朱浩, 等. 基于趋势项误差控制的频域积分算法研究与应用[J]. 振动与冲击, 2015, 34(2): 171-175 https://www.cnki.com.cn/Article/CJFDTOTAL-ZDCJ201502030.htmHU Y M, ZHOU Y J, ZHU H, et al. Integration algorithm based on trend-control of error in frequency domain[J]. Journal of Vibration and Shock, 2015, 34(2): 171-175 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZDCJ201502030.htm [13] 陈太聪, 张奇. 基于频谱能量形态拟合的加速度积分方法研究[J]. 振动与冲击, 2019, 38(13): 7-12+20 https://www.cnki.com.cn/Article/CJFDTOTAL-ZDCJ201913003.htmCHEN T C, ZHANG Q. Acceleration integration method based on frequency spectral energy morphological fitting[J]. Journal of Vibration and Shock, 2019, 38(13): 7-12+20 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZDCJ201913003.htm [14] 成勋, 周铁, 孙侃, 等. 基于小波去噪振动加速度信号处理及时频域积分方法研究[J]. 电力与能源, 2019, 40(6): 633-637 https://www.cnki.com.cn/Article/CJFDTOTAL-LYJI201906001.htmCHENG X, ZHOU T, SUN K, et al. Time/frequency-domain integration method of vibration acceleration signal processed by wavelet denoising[J]. Power and energy, 2019, 40(6): 633-637 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-LYJI201906001.htm [15] 朱勇, 姜万录, 郑直, 等. 基于有效信息重构的故障旋转机械振动加速度信号积分方法[J]. 中国机械工程, 2015, 26(18): 2511-2517 https://www.cnki.com.cn/Article/CJFDTOTAL-ZGJX201518018.htmZHU Y, JIANG W L, ZHENG Z, et al. Useful information reconstruction-based vibration acceleration signal integral method for faulty rotating machinery[J]. China Mechanical Engineering, 2015, 26(18): 2511-2517 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZGJX201518018.htm [16] 温广瑞, 李杨, 廖与禾, 等. 基于精确信息重构的故障转子系统振动加速度信号积分方法[J]. 机械工程学报, 2013, 49(8): 1-9 https://www.cnki.com.cn/Article/CJFDTOTAL-JXXB201308002.htmWEN G R, LI Y, LIAO Y H, et al. Faulty rotor system vibration acceleration signal integration method based on precise information reconstruction[J]. Journal of Mechanical Engineering, 2013, 49(8): 1-9 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JXXB201308002.htm