作者：吕岩¹, 房立清¹, 褚怡², 赵玉龙¹

作者单位：1. 军械工程学院火炮工程系, 河北 石家庄 050003;
2. 61081部队, 北京 100142

关键词：混沌理论;相关向量机;自动机;特征提取

摘要：

针对自动机振动信号的非线性与短时冲击特性，提出一种基于混沌理论和相关向量机（relevance vector machine，RVM）相结合的自动机故障诊断方法。首先，计算每一组自动机振动信号的最大Lyapunov指数、关联维数、Kolmogorov熵和相对关联距离熵共4个混沌参数并组成特征矩阵，从而表征自动机状态信息。然后，将特征矩阵输入RVM中进行分类识别，判断故障类型。自动机故障诊断实例表明，通过提取自动机振动信号的4个混沌参数可以实现其运行状态信息表征，并且RVM能够较精确地识别自动机的常见故障；此外，通过与支持向量机（support vector machine，SVM）的故障诊断结果进行对比，验证RVM分类模型的优势。

Automaton fault diagnosis based on chaos theory and relevance vector machine

LÜ Yan¹, FANG Liqing¹, CHU Yi², ZHAO Yulong¹

1. Department of Artillery Engineering, Ordnance Engineering College, Shijiazhuang 050003, China;
2. 61081 Troops, PLA, Beijing 100142, China

Abstract: Aiming at the nonlinear and short time impact characteristics of automaton vibration signal, a method that based on the combination of chaos theory and relevance vector machine (RVM) was proposed. Firstly, the characteristic matrix of the four chaotic parameters of the vibration signal of automatic mechanism was calculated for stating the automaton state information. Finally, the characteristic matrix was put into RVM to recognize different fault types. The experiment results of automatic mechanism show that the representation of the running state information can be realized by extracting four chaotic parameters of the vibration signal of the automatic machine, and RVM can classify usual fault types of automatic mechanism exactly. In addition by comparing with the diagnostic results of SVM, the advantage of RVM is verified.

Keywords: chaos theory;RVM;automaton;feature extraction

2017, 43(3): 111-116  收稿日期: 2016-08-07;收到修改稿日期: 2016-10-11

基金项目: 河北省自然科学基金资助项目（E2016506003）

作者简介: 吕岩(1992-),男,吉林松原市人,硕士研究生,专业方向为兵器性能检测与故障诊断。

参考文献

[1] 房立清,徐春晓,李献锋. 基于微加速度计的高炮自动机测试系统设计[C]//第八届全国信息获取与处理学术会议论文集.威海:中国仪器仪表学会,2010:181-185.
[2] PAN H X, CUI Y P. Automaton fault diagnosis based on chaos theory[J]. Applied Mechanics and Materials,2013,(427-429):697-701.
[3] 康瑞霞,王茂林,陈永刚,等. 某导气式双路供弹机空拨弹故障研究[J]. 火炮发射与控制学报,2008(3):83-85.
[4] 袁玉勇,李开宇,刘文波. 光电雷达电子部件故障树分析[J]. 电子测量技术,2016,39(16):149-152.
[5] PAN H X, PAN M Z, ZHAO R P, et al. An automaton fault diagnosis based on shock response analysis[J]. Applied Mechanics and Materials,2012:226-228,745-748.
[6] KUMAR M, YADAV S P. The weakest-norm based intuitionistic fuzzy fault-tree analysis to evaluate system reliability[J]. ISA Transactions,2012,51(4):531-538.
[7] EDSON D F, PABLO G. Real bounds and lyapunov exponents[J]. Discrete and Continuous Dynamical Systems-Series A,2016,36(4):1957-1982.
[8] WANG A Y, PAN H X, LIU H L. Study on fault feature extraction of high-speed automaton[J]. Indonesian Journal of Electrical Engineering and Computer Science,2013,11(10):5867-5875.
[9] 孙宽雷,韩峻. 基于单子带重构改进算法的舰炮自动机故障特征提取方法[J]. 舰船电子工程,2014(10):123-126.
[10] 姜旭刚,何楚瑶. 基于小波变换的自动机故障诊断[J]. 传感器世界,2013(6):13-16.
[11] 胡敏. 分形理论在火炮自动机故障诊断中的应用[J]. 电子制作,2013(2):148.
[12] LIU Y, ZHANG J Z. Predicting traffic flow in local area networks by the largest lyapunov exponent[J]. Entropy,2016,18(1):1-10.
[13] ALEKSANDAR K, ALEKSANDRA V, TIJANA B. Topographic distribution of EEG alpha attractor correlation dimension values in wake and drowsy states in humans[J]. International Journal of Psychophysiology,2015, 95(3):278-291.
[14] 李奎为,胡瑾秋,张来斌,等. 关联维数无标度区确定方法及诊断应用[J]. 石油机械,2007,35(4):43-45.
[15] 赵贵兵,石炎福,段文锋,等. 从混沌时间序列同时计算关联维数和Kolmogorov熵[J]. 计算物理,1999,16(3):309-315.
[16] NILESH G C, BHAVESH R B. A new algorithm for busbar fault zone identification using relevance vector machine[J]. Electric Power Components and Systems,2016,44(2):193-205.
[17] 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.
[18] 饶国强,冯辅周,司爱威,等. 排列熵算法参数的优化确定方法研究[J]. 振动与冲击,2014,33(1):188-193.