作者：安邦¹, 潘宏侠^1,2, 张玉学¹, 赵雄鹏¹

作者单位：1. 中北大学机械与动力工程学院, 山西 太原 030051;
2. 中北大学 系统辨识与诊断技术研究所, 山西 太原 030051

关键词：变分模态分解;模态混叠;极限学习机;自动机;故障诊断

摘要：

由于自动机工作环境复杂、各种响应信号相互叠加，为准确、高效地提取自动机信号的故障特征，提出一种应用变分模态分解（VMD）和极限学习机（ELM）的自动机故障诊断方法。首先对自动机信号进行变分模态分解，并与经验模态分解（EMD）结果进行对比；同时提取各模态分量的能量百分比和各工况下不同样本的样本熵作为特征值；将提取到的特征值输入到极限学习机中进行故障诊断，再与传统的双谱分析诊断结果进行比较。最终VMD方法实现信号频域内各分量的自适应剖分，并得出ELM的故障诊断准确率为87.5%。实验结果表明：变分模态分解能够有效避免模态混叠现象，同时验证所提方法的可行性与有效性。

Application of variational mode decomposition in fault diagnosis of automaton

AN Bang¹, PAN Hongxia^1,2, ZHANG Yuxue¹, ZHAO Xiongpeng¹

1. School of Mechanical and Power Engineering, North University of China, Taiyuan 030051, China;
2. System Identification and Diagnosis Technology Research Institute, North University of China, Taiyuan 030051, China

Abstract: Due to the complex working environment of automatons and the superimposition of various corresponding signals, to extract the fault characteristics of the signal accurately and efficiently, an automaton fault diagnosis method based on variational mode decomposition(VMD) and extreme learning machine(ELM) is proposed. Firstly, the VMD of automaton signal is performed and compared with empirical mode decomposition(EMD) results. Meanwhile, the energy percentage of each component and the sample entropy of each sample are extracted and taken as the eigenvalues. Then, the extracted feature parameters are input to the extreme learning machine (ELM) for fault diagnosis, and compared with the traditional bispectrum diagnostic results. Finally, the VMD method achieved the adaptive subdivision of the components in the signal frequency domain, and the accuracy of the ELM is 87.5%. The results showed that the VMD can effectively avoid the phenomenon of modal mixture, and verified the feasibility and effectiveness of the proposed method.

Keywords: variational mode decomposition;modal mixture;extreme learning machine;automaton;fault diagnosis

2017, 43(7): 112-116  收稿日期: 2016-12-17;收到修改稿日期: 2017-02-12

基金项目: 国家自然科学基金项目（51175480，51675491）

作者简介: 安邦(1993-),男,河北石家庄市人,硕士研究生,专业方向为装备系统检测诊断与控制。

参考文献

[1] 孙致远,郑坚,熊超,等. 基于FMEA和模糊贴近度的装备故障维修方法选择[J]. 火炮发射与控制学报,2015,36(3):86-90.
[2] 吕岩,房立青,齐子元,等. 基于LCD关联维数和SVM的自动机故障诊断[J]. 机械设计与研究,2016,32(4):149-153.
[3] 潘龙,潘宏侠,马白雪. 应用HHT时频分析与SVM的自动机故障诊断[J]. 机械设计与制造,2015(4):59-61.
[4] DRAGOMIRETSKIY K, ZOSSO D. Variational mode decomposition[J]. IEEE Tran on Signal Processing,2014, 62(3):531-544.
[5] WANG Y X, MARKERT R, XIANG J W, et al. Research on variational mode decomposition and its application in detecting rub-impact fault of the rotor system[J]. Mechanical Systems and Signal Processing,2015(60-61):43-25.
[6] 刘长良,武英杰,甄成刚. 基于变分模态分解和模糊C均值聚类的滚动轴承故障诊断[J].中国电机工程学报,2015, 35(13):3358-3365.
[7] 岳应娟,孙刚,蔡艳萍,等. 变分模态分解在轴承故障诊断中的应用[J]. 轴承,2016(8):50-54.
[8] 唐贵基,王晓龙. 变分模态分解方法及其在滚动轴承早期故障诊断中的应用[J]. 振动工程学报,2016,29(4):635-648.
[9] 王振威. 基于变分模态分解的故障诊断方法研究[D]. 河北:燕山大学,2015.
[10] 张学清,梁军,张熙,等. 基于样本熵和极端学习机的超短期风电功率组合预测模型[J]. 中国电机工程学报,2013, 33(25):33-40.
[11] HUANG G B, ZHU Q Y, SIEW C K. Extreme learning machine:theory and applications[J]. Neurocomputing,2006, 70(1):489-501.
[12] 王利琴,董永峰,顾军华. 基于改进极限学习机的心律失常分类[J]. 计算机仿真,2014,31(6):352-356.
[13] 高光勇,蒋国平. 采用优化极限学习机的多变量混沌时间序列预测[J]. 物理学报,2012,61(4):37-45.
[14] 邓勇,于晨松.因子分析和ELM在模拟电路故障诊断的应用[J]. 电子测量与仪器学报,2016,30(10):1512-1519.
[15] 潘宏侠,兰海龙,任海峰. 基于局域波降噪和双谱分析的自动机故障诊断研究[J]. 兵工学报,2014,35(7):1077-1082.