基于时序模型和自联想神经网络的齿轮故障程度评估

PDF(1736 KB)

振动与冲击 ›› 2019, Vol. 38 ›› Issue (2) : 18-24.

论文

张龙1,2,3, 成俊良2，杨世锡1,李兴林3

作者信息 +

Fault severity assessment for gears based on AR model and auto-associative neural network

ZHANG Long1,2,3,CHENG Junliang2,YANG Shixi1,LI Xinglin3

Author information +

文章历史 +

摘要

齿轮是机械传动系统中的重要零部件，在役齿轮故障程度评估是机械系统剩余寿命预测和状态维修的基础。目前广泛研究的基于概率相似度量的故障评估方法存在过早饱和等问题，不利于在线监测。本文提出一种基于自回归时序模型（Auto-regressive, AR）和自联想神经网络（Auto-associative neural network, AANN）的齿轮故障程度在线评估方法，其中AR模型用于齿轮振动信号特征提取，AANN用于故障程度评估。首先提取基准阶段（如无故障阶段）振动信号AR模型系数作为AANN的输入和输出向量，得到基准评估模型。将待评估信号AR系数构成的特征向量输入到基准AANN，得到重构AR系数。基于原始AR系数和重构AR系数组成两个自回归模型分别对待评估信号进行时序建模，分别得到各自的模型残差序列。基于残差序列之间的差异，提出了一种基于残差均方根差值的故障程度定量评估指标。离散故障程度的齿轮振动数据分析结果表明，本文方法能有效区分齿轮故障的不同程度；在此基础上利用齿轮全寿命周期实验数据进一步验证本方法的有效性，结果显示提出的方法能够及时发现肉眼没有观测到的早期故障，且随着齿轮性能的不断退化，能直观反映齿轮故障程度的加深。

Abstract

The gear damage severity evaluation underlines the prognostics and condition-based maintenance of mechanical systems.Being motivated by the fact that paradigms based on the probability similarity,like the hidden Markov model (HMM) and Gaussian mixed model (GMM),tend to an early saturation,a new approach for gear damage evaluation was proposed by making use of an autoregressive model (AR) and auto-associative neural network (AANN).The AR model was made to fit with gear vibration signals and the model coefficients were extracted as feature vectors which were then,fed to AANN to obtain reconstructed AR coefficients.A baseline AANN was trained by using the feature vectors from normal condition.The reconstructed AR coefficients by the baseline AANN will deviate from the original AR coefficients,if the gear condition degrade from normal condition.So,the difference between the residuals of AR models using reconstructed and original AR coefficients was exploited to formulate a gear damage indicator.Two experimental data sets involving one discrete damage-degree and one run-to-failure test were utilized to verify the proposed method.The results show the novel indicator is able to track damage progress with a consistent trend and to detect incipient damage in time.

导出引用

张龙1,2,3, 成俊良2，杨世锡1,李兴林3. 基于时序模型和自联想神经网络的齿轮故障程度评估[J]. 振动与冲击, 2019, 38(2): 18-24

ZHANG Long1,2,3,CHENG Junliang2,YANG Shixi1,LI Xinglin3. Fault severity assessment for gears based on AR model and auto-associative neural network[J]. Journal of Vibration and Shock, 2019, 38(2): 18-24

参考文献

[1] Ben Ali J, Chebel-Morello B, Saidi L, et al. Accurate bearing remaining useful life prediction based on Weibull distribution and artificial neural network[J]. Mechanical Systems and Signal Processing, 2014, 56-57: 150-172.
[2] 雷亚国, 陈吴, 李乃鹏, 等. 自适应多核组合相关向量机预测方法及其在机械设备剩余寿命预测中的应用[J]. 机械工程学报, 2016, 56(01): 87-93.
LEI Ya-guo, CHEN Wu, LI Nai-peng, et al. A relevance vector machine prediction method based on adaptive multi-kernel combination and its application to remaining useful life prediction of machinery[J]. Journal of Mechanical Engineering, 2016, 56(01): 87-93.
[3] 张小丽, 陈雪峰, 李兵, 等. 机械重大装备寿命预测综述[J]. 机械工程学报, 2011, 47(11): 100-116.
ZHANG Xiao-li, CHEN Xue-feng, LI Bing, et al. Review of life prediction for mechanical major equipments[J]. Journal of Mechanical Engineering, 2011, 47(11): 100-116.
[4] 王奉涛, 陈旭涛, 柳晨曦, 等. 基于KPCA和WPHM的滚动轴承可靠性评估与寿命预测[J]. 振动.测试与诊断, 2017, 37(03): 476-483.
WANG Feng-tao, CHEN Xu-tao, LIU Chen-xi, et al. Rolling bearing reliability assessment and life prediction based on KPCA and WPHM[J]. Journal of Vibration, Measurement & Diagnosis, 2017, 37(03): 476-483.
[5] 郭艳平, 颜文俊, 包哲静. 风力发电机组在线故障预警与诊断一体化系统设计与应用[J]. 电力系统自动化, 2010, 34(16): 83-86.
GUO Yan-ping, YAN Wen-jun, BAO Zhe-jing. Design and application of online fault warning and diagnosis integrated system for wind turbines[J]. Automation of Electric Power Systems, 2010, 34(16): 83-86.
[6] Wang Y, Liang M. An adaptive SK technique and its application for fault detection of rolling element bearings[J]. Mechanical Systems and Signal Processing, 2011, 25(5): 1750-1764.
[7] 罗毅, 甄立敬. 基于小波包与倒频谱分析的风电机组齿轮箱齿轮裂纹诊断方法[J]. 振动与冲击, 2015, 34(03): 210-214.
LUO Yi, ZHEN Li-jing. Diagnosis method of turbine gearbox gearcrack based on wavelet packet and cepstrum analysis[J]. Journal of Vibration and Shock, 2015, 34(03): 210-214.
[8] Tang B, Liu W, Song T. Wind turbine fault diagnosis based on Morlet wavelet transformation and Wigner-Ville distribution[J]. Renewable Energy, 2010, 35(12): 2862-2866.
[9] 马伦, 康建设, 孟妍, 等. 基于Morlet小波变换的滚动轴承早期故障特征提取研究[J]. 仪器仪表学报, 2013, 34(04): 920-926.
MA Lun, KANG Jian-she, MENG Yan, et al. Research on feature extraction of rolling bearing incipient fault based on Morlet wavelet transform[J]. Chinese Journal of Scientific Instrument, 2013, 34(04): 920-926.
[10] 于德介, 程军圣, 杨宇. 基于EMD和AR模型的滚动轴承故障诊断方法[J]. 振动工程学报, 2004, 17(03): 84-87.
YU De-jie, CHENG Jun-sheng, YANSG Yu. A fault diagnosis approach for roller bearings based on EMD method and AR model[J]. Journal of Vibration Engineering, 2004, 17(03): 84-87.
[11] 从飞云, 陈进, 董广明. 基于谱峭度和AR模型的滚动轴承故障诊断[J]. 振动.测试与诊断, 2012, 32(04): 538-541.
CONG Fei-yun, CHEN Jin, DONG Guang-ming. Spectral Kurtosis and AR Model Based Method for Fault Diagnosis of Rolling Bearings[J]. Journal of Vibration, Measurement & Diagnosis, 2012, 32(04): 538-541.
[12] Cheng J, Yu D, Yang Y. A Fault Diagnosis Approach for Gears Based on IMF AR Model and SVM[J]. EURASIP Journal on Advances in Signal Processing, 2008, 2008(1): 1-7.
[13] Heyns T, Heyns P S, de Villiers J P. Combining synchronous averaging with a Gaussian mixture model novelty detection scheme for vibration-based condition monitoring of a gearbox[J]. Mechanical Systems and Signal Processing, 2012, 32(4): 200-215.
[14] 曹端超, 康建设, 赵劲松, 等. 改进EMD和HMM的齿轮故障诊断方法应用[J]. 噪声与振动控制, 2013, 33(03): 208-211.
CAO Duan-chao, KANG Jian-she, ZHAO Jin-song, et al. Application of Gear Fault Diagnosis Method Based on Improved EMD and HMM[J]. Noise and Vibration Control, 2013, 33(03): 208-211.
[15] 王建利. 滚动轴承性能退化评价与趋势预测研究[D]. 大连理工大学, 2013.
WANG Jian-li. Research on performance degradation assessment and trend prediction of rolling bearing[D]. Dalian University of Technology, 2013.
[16] Kramer M A. Nonlinear Principal Component Analysis Using Autoassociative Neural Networks[J]. Aiche Journal, 1991, 37(37): 233-243.
[17] Wang D, Miao Q, Kang R. Robust health evaluation of gearbox subject to tooth failure with wavelet decomposition[J]. Journal of Sound and Vibration, 2009, 324(3-5): 1141-1157.
[18] Rafiee J, Tse P W. Use of autocorrelation of wavelet coefficients for fault diagnosis[J]. Mechanical Systems and Signal Processing, 2009, 23(5): 1554-1572.
[19] Parey A, El Badaoui M, Guillet F, et al. Dynamic modelling of spur gear pair and application of empirical mode decomposition-based statistical analysis for early detection of localized tooth defect[J]. Journal of Sound and Vibration, 2006, 294(3): 547-561.
[20] 刘韬. 基于隐马尔可夫模型与信息融合的设备故障诊断与性能退化评估研究[D]. 上海交通大学, 2013.
LIU Tao. Study of Hidden Markov Model and information fusion in equipment fault diagnosis and performance degradation assessment[D]. Shanghai Jiao Tong University, 2013.