基于邻域属性重要度与主成分分析的齿轮箱故障特征约简

摘要/Abstract

摘要： 为有效降低齿轮箱故障特征的维数并提高诊断效率，提出了基于邻域属性重要度与主成分分析法相结合的齿轮箱故障特征约简方法，并利用支持向量机和BP神经网络对诊断的准确率进行对比分析。针对齿轮箱中具有不同程度裂纹的齿轮，选取其时域、频域和基于希尔伯特变换的36个特征；将邻域模型引入到特征属性的约简，构造前向贪心算法，以邻域属性重要度较大的9个特征作为特征集，提取累积贡献率达到95%以上的主成分，分别输入支持向量机和BP神经网络分类器中进行分类识别，并与不经过特征优选的主成分特征融合相对比。结果表明，采用基于邻域属性重要度与主成分分析法相结合的特征约简方法，既可以降低齿轮箱故障特征的维数，又不影响对其运行状态的表征，有助于识别不同裂纹水平的齿轮，与不经过特征优选直接进行融合的方法相比，所提出方法诊断准确率更高，训练时间更短。

关键词: 齿轮箱, 特征约简, 邻域决策系统, 主成分分析, 支持向量机

Abstract: A gearbox fault feature reduction method was proposed to reduce the feature dimension and improve the accuracy of diagnosis based on neighborhood attribute importance and PCA. The SVM and BP neural network were used to analyze the diagnosis accuracy. The 36 features of different crack gears in gearbox were selected based on time-domain, frequency-domain and Hilbert transform. A forward-greedy numerical attribute reduction algorithm was established to select the optimal features based on neighborhood model. The 9 features with higher attribute importance were selected as the feature set. The principal components which had more than 95% cumulative contribution rate were extracted from the optimal feature set and input into the SVM and BP neural network classifier for identification. The results of the above method were compared with the PCA method without the optimal feature selections. Results show that the feature dimension may be reduced, the operating status of the gearbox may be indicated and the gear crack levels will be identified by using the proposed feature fusion method. Compared with the fusion method without the optimal feature selections, the diagnosis accuracy of the optimal feature selection is higher and the training time is shorter.

Key words: gearbox, feature reduction, neighborhood decision system, principal component analysis(PCA), support vector machine(SVM)

中图分类号:

古莹奎, 潘高平, 朱繁泷, 承姿辛. 基于邻域属性重要度与主成分分析的齿轮箱故障特征约简[J]. 中国机械工程.

Gu Yingkui, Pan Gaoping, Zhu Fanlong, Cheng Zixin. Gearbox Fault Feature Reduction Based on Neighborhood Attribute Importance and PCA[J]. China Mechanical Engineering.

参考文献

[1]雷亚国, 何正嘉, 林京, 等. 行星齿轮箱故障诊断技术的研究进展[J]. 机械工程学报, 2011, 47(19): 59-67.

Lei Yaguo, He Zhengjia, Lin Jing,et al. Research Advances of Fault Diagnosis Technique for Planetary Gearboxes[J]. Journal of Mechanical Engineering, 2011, 47(19): 59-67.

[2]林近山, 陈前. 基于多重分形去趋势波动分析的齿轮箱故障特征提取方法[J]. 振动与冲击, 2013, 32(2): 97-101.

Lin Jinshan, Chen Qian.Fault Feature Extraction of Gearboxes Based on Multifractal Detrended Fluctuation Analysis[J]. Journal of Vibration and Shock, 2013, 32(2): 97-101.

[3]王卫国, 孙磊. 基于EEMD-CWD的齿轮箱振动信号故障特征提取[J]. 兵工学报, 2014, 35(8): 1288-1294.

Wang Weiguo, Sun Lei. Gearbox Vibration Signal Fault Feature Extraction Based on Ensemble Empirical Mode Decomposition and Choi-Williams Distribution[J]. Acta Armamentarii, 2014, 35(8): 1288-1294.

[4]雷亚国, 孔德同, 李乃鹏,等. 自适应总体平均经验模式分解及其在行星齿轮箱故障检测中的应用[J]. 机械工程学报, 2014, 50(3): 64-70.

Lei Yaguo, Kong Detong, Li Naipeng, et al. Adaptive Ensemble Empirical Mode Decomposition and Its Application to Fault Detection of Planetary Gearboxes[J]. Journal of Mechanical Engineering, 2014, 50(3): 64-70.

[5]曾令男, 丁建伟, 赵炯,等. 基于互信息的复杂装备高维状态监测数据相关性发现与建模[J]. 计算机集成制造系统, 2013, 19(12): 3017-3025.

Zeng Lingnan, Ding Jianwei, Zhao Jiong, et al. Detecting and Modeling for Associations Between High-dimension Condition Monitoring Data of Complex Equipment Based on Mutual Information[J]. Computer Integrated Manufacturing Systems, 2013, 19(12): 3017-3025.

[6]Peng H C, Long F H, Ding C. Feature Selection Based on Mutual Information: Criteria of Max-dependency, Max-relevance, and Min-redundancy[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2005, 27(8): 1226-1238.

[7]Qu J.Support-vector-machine-based Diagnostics and Prognostics for Rotating Systems[D]. Edmonton: University of Alberta, 2011. [8]Otman B, Yuan X H. Engine Fault Diagnosis Based on Multi-sensor Information Fusion Using Dempster-shafer Evidence Theory[J]. Information Fusion, 2007, 8(4): 379-386.

[9]聂磊, 黄圣国, 舒平, 等. 基于支持向量机(SVM)的民用飞机重着陆智能诊断研究[J]. 中国安全科学学报, 2009, 19(7): 149-154.

Nie Lei, Huang Shengguo, Shu Ping, et al. Intelligent Diagnosis for Hard Landing of Aircraft Based on SVM[J]. China Safety Science Journal, 2009, 19(7): 149-154.

[10]丁世飞, 齐丙娟, 谭红艳. 支持向量机理论与算法研究综述[J]. 电子科技大学学报, 2011, 40(1): 2-10.

Ding Shifei, Qi Bingjuan, Tan Hongyan. An Overview on Theory and Algorithm of Support Vector Machines[J]. Journal of University of Electronic Science and Technology of China, 2011, 40(1): 2-10.

[11]Lee M C. Using Support Vector Machine with a Hybrid Feature Selection Method to the Stock Trendprediction[J]. Expert and Systems with Application, 2009, 36(8): 10896-10904.

[12]Lei Y G, Zuo M J. Gear Crack Level Identification Based on Weighted K Nearest Neighbor Classification Algorithm[J]. Mechanical Systems and Signal Processing, 2009, 23(5): 1535-1547.

[13]古莹奎, 杨子茜, 朱繁泷. 基于主成分分析的齿轮箱故障特征融合分析[J]. 中国机械工程, 2015, 26(11): 1532-1537.

Gu Yingkui, Yang Zixi, Zhu Fanlong. Gearbox Fault Feature Fusion Based on Principal Component Analysis[J]. China Mechanical Engineering, 2015, 26(11): 1532-1537.

[14]古莹奎, 孔军廷, 朱繁泷. 基于邻域属性重要度的齿轮箱故障特征优选方法[J]. 煤炭学报, 2015, 40(增刊2): 560-567.

Gu Yingkui, Kong Junting, Zhu Fanlong. Gearbox Fault Feature Optimal Selection Based on Neighborhood Attribute Importance[J]. International Journal of Coal Science & Technology, 2015, 40(S2): 560-567.

[15]Abu-mahfouz I A. A Comparative Study of Three Artificial Neural Networks for the Detection and Classification of Gear Faults[J]. International Journal of General Systems, 2005, 34(3): 261-277.

[16]胡清华, 于达仁, 谢宗霞. 基于邻域粒化和粗糙逼近的数值属性约简[J]. 软件学报, 2008, 19(3): 640-649.

Hu Qinghua, Yu Daren, Xie Zongxia. Numerical Attribute Reduction Based on Neighborhood Granulation and Rough Approximation[J]. Journal of Software, 2008, 19(3): 640-649.