基于不同工况下辅助数据集的齿轮箱故障诊断

段礼祥,谢骏遥,王凯,王金江

振动与冲击 ›› 2017, Vol. 36 ›› Issue (10) : 104-108.

PDF(1520 KB)
PDF(1520 KB)
振动与冲击 ›› 2017, Vol. 36 ›› Issue (10) : 104-108.
论文

基于不同工况下辅助数据集的齿轮箱故障诊断

  • 段礼祥,谢骏遥,王凯,王金江
作者信息 +

Considering the Influence of Kernel Function of Transfer Component Analysis and its Application in Variable Working Condition of Gearbox Fault Diagnosis

  • DUAN Li-xiang,XIE Jun-yao,WANG Kai,WANG Jin-jiang
Author information +
文章历史 +

摘要

针对变工况下齿轮箱监测数据重用性低,受复杂工况影响大和已训练模型经常失效的问题,提出迁移成分分析方法用于设备故障诊断。迁移成分分析(Transfer Component Analysis, TCA)通过核函数将训练样本与测试样本映射到潜在空间,进而减小训练样本与测试样本的分布差异性。重点对比分析四种核函数对迁移成分分析算法性能的影响,并提出了可迁移度指标KL散度(Kullback-Leibler divergence),用来衡量训练样本与测试样本的分布差异度。通过仿真分析和实验验证,迁移成分分析方法相比传统机器学习算法明显地减小了训练样本与测试样本的分布差异,有效提高了齿轮箱变工况故障诊断的准确率和可靠性,结果表明高斯核函数下的迁移成分分析方法性能最优。

Abstract

In view of the low reusability of monitoring data, the influencing of complex working condition and failure of trained model, Transfer Component Analysis (TCA) is introduced to solve the problem of equipment fault diagnosis in variable working conditions. TCA methoddecreases the distribution difference of train sample and test sample, by utilizing kernel function to map the feature of two samples to latent space. Besides, the effect of four different kernel functions is compared in TCA algorithm and transfer degree index KL divergence (Kullback-leibler divergence) is introduced to measure the distribution difference degree of train samples and test sample. According to the simulation analysis and experiment verification, compared with traditional machine learning methods, TCA performs better in reduction the distribution difference of two samples and improving gearbox diagnose accuracy, especially the Gaussian kernel function. 
 

引用本文

导出引用
段礼祥,谢骏遥,王凯,王金江. 基于不同工况下辅助数据集的齿轮箱故障诊断[J]. 振动与冲击, 2017, 36(10): 104-108
DUAN Li-xiang,XIE Jun-yao,WANG Kai,WANG Jin-jiang. Considering the Influence of Kernel Function of Transfer Component Analysis and its Application in Variable Working Condition of Gearbox Fault Diagnosis[J]. Journal of Vibration and Shock, 2017, 36(10): 104-108

参考文献

[1] 邹今春, 沈玉娣.变工况齿轮箱故障诊断方法综述[J].机械传动, 2012, 36(8)124-132.
ZHOU Jin-chun, SHEN Yu-di, A review of gearbox fault diagnosis methods in variable working conditions [J], Journal of Mechanical, 2012, 36(8)124-132.
[2] 徐刚, 骆志高, 李明义, 陈鹏.变工况运行的机械故障诊断研究[J].机械工程学报, 2001, 37(12):104-107.
XU Gang, LUO Zhi-gao, LI Ming-yi, CHEN Peng, mechanical fault diagnosis research in variable working conditions[J], Journal of Mechanical Engineering, 2001, 37(12):104-107.
[3] 程利军,张英堂,李志宁,等.基于时频分析及阶比跟踪的曲轴轴承故障诊断研究[J].振动与冲击,2012,31( 19) :73-78.
CHEN Li-jun,ZHANG Ying-tang,LI zhi-ning. Fault diagnoseof an engine’s main bearing based on time-frequency analysisand order tracing[J]. Journal of Vibration and Shock,2012,31( 19) : 73-78.
[4]孙自强,陈长征,谷艳玲,等.基于混沌和取样积分技术的大型风电增速箱早期故障诊断[J].振动与冲击,2013,32( 9) : 113- 117.
SUN Zi-qiang,CHEN Chang-zheng,GU Yan-ling et al.Incipient fault diagnosis of large scale wind turbine gearboxbased on chaos theory and sampling integral technology[J].Journal of Vibration and Shock,2013, 32(9): 113-117.
[5]李友荣,曾法力,吕勇,等.小波包分析在齿轮故障诊断中的应用[J].振动与冲击, 2005, 24( 5) : 101-103.
LI You-rong,ZENG Fa-li,LYong,et al. Gear faultdiagnosis by using wavelet packet[J]. Journal of Vibrationand Shock, 2005, 24( 5) : 101 - 103.
[6] SinnoJialin Pan, Qiang Yang.A survey on transfer learning[J]. IEEE Transactions on Knowledge and Data Engineering, 2010, 22(10):1345-1359.
[7] Jielu, VahidBehbood, Peng Hao, et al. Transfer learning using computational intelligence[J]. a survey, Knowledge-based Systems, 2015, 80:14-23.
[8] Jose Costa Pereira, NunoVasconcelos.Cross-modal domain adaptation for text-based regularization of image semantics in image retrieval systems[J]. Computer Vision and Image Understanding, 2014, 124:123-135.
[9] Vishal M. Patel, RaghuramanGopalan, Ruonan Li, et al. Visual domain adaptation: a survey of recent advances[J]. IEEE Signal Processing Magazine, 2015, 15:53-69.
[10] Chuanxian Ren, Daoqing Dai, Kekun Huang, et al. Transfer learning of structured representation for face recognition[J]. IEEE Transactions on Image Processing, 2014, 23(12):5440-5454.
[11] Jun Deng, Zixing Zhang, Florian Eyben, et al.Autoencoder-based unsupervised domain adaptation for speech emotion recognition[J]. IEEE Signal Processing Letters, 2014, 21(9):1068-1072.
[12] Lei Zhang, David Zhang.Domain adaptation extreme learning machines for drift compensation in E-Nose systems[J]. IEEE Transactions on Instrumentation and Measurement, 2015, 64(7):1790-1801.
[13] 周书恒,杜文莉. 基于迁移学习的裂解炉产率建模[J]. 化工学报,2014, 65(12):492-498.
ZHOU Shu-heng, DU Wen-li, Modeling of ethylene cracking furnace yields based on transfer learning[J],CIESC Journal, 2014, 65(12):492-498.
[14] Shuai Huang, Jing Li, Lewei Chen, et al.A transfer learning approach for network modeling[J]. IIE Transactions, 2012, 44:915-931.
[15] S.J.Pan,J.T.Kwok, Q. Yang,Transfer learning via dimensionalityreduction[C].in Proc. 23rd AAAI Conf. Artif.Intell. Chicago, IL, Jul. 2008: 677–682.
[16] Pan S J, Tsang I W, Kwok J T, et al. Domain adaptation via transfer component analysis[J]. Neural Networks, IEEE Transactions on, 2011, 22(2): 199-210.
[17] https://www.phmsociety.org/competition/PHM/09/
apparatus
[18] 罗德扬. 时域同步平均原理与应用[J].振动、测试与诊断, 1999, 19(3): 202-207.
LUO De-yang. Time domain synchronization average principle and application[J]. Journal of Vibration, Measurement & Diagnosis, 1999,19(3): 202-207.
[19]P. Večeř, M. Kreidl, R. Šmíd. Condition Indicators or GearboxCondition Monitoring Systems[J].Czech Technical University in Prague, ActaPolytechnica Vol. 2005,45(6) :35-43.

PDF(1520 KB)

Accesses

Citation

Detail

段落导航
相关文章

/