基于多元分解和深度学习的机器故障检测方法的研究Machine Fault Detection Method Based on Multivariate Decomposition and Deep Learning
林香
摘要(Abstract)：
基于声学的机器故障检测方法是一种新颖的研究方法，但是需要依赖采集足够丰富的故障类型数据.因此，提出一种无监督的机器故障检测方法，即使用机器正常运行状态的数据作为训练样本，从而鉴别出机器异常状态.该检测方法的一个关键问题是信号中的噪声，它会影响检测的性能，针对这个问题，提出了一种基于多元张量分解的准非参数光谱数据去噪策略，即非负正则多重分解，这种策略特别适用于发出稳定声音的机器.通过实验证明了这种声学检测方法，可以使得工业过程的故障监控更加准确.
关键词(KeyWords)： 多元分解;深度学习;无监督机器故障检测
基金项目(Foundation): 福建省中青年教师教育科研项目(JZ181049)
作者(Author): 林香
DOI: 10.16393/j.cnki.37-1436/z.2023.02.022
参考文献(References)：
- [1]刘伟峰.基于机器学习的大规模并行计算机系统硬件故障检测方法[J].技术与市场，2022,29(06):51-52+55.
- [2]鲍文霞，刘杨，杨先军，等.基于门控双卷积神经网络的机动车发动机故障检测[J].安徽大学学报(自然科学版),2022,46(02):39-45.
- [3]Michau G,Palm T,Fink O.Deep feature learning network for fault detection and isolation[C]//Annual Conference of the PHM Society.2017,9(1):108-118.
- [4]Tagawa T,Tadokoro Y ,Yairi T.Structured denoising autoencoder for fault detection and analysis[C]//Asian conference on machine learning.PMLR,2015:96-111.
- [5]Purwins H,Li B,Virtanen T,et al.Deep learning for audio signal processing[J].IEEE Journal of Selected Topics in Signal Processing,2019,13(2):206-219.
- [6]Michau G,Hu Y,Palmé T,et al.Feature learning for fault detection in high-dimensional condition monitoring signals[J].Proceedings of the Institution of Mechanical Engineers,Part O:Journal of Risk and Reliability,2020,234(1):104-115.
- [7]Lim L H,Comon P.Nonnegative approximations of nonnegative tensors[J].Journal of Chemometrics:A Journal of the Chemometrics Society,2009,23(7-8):432-441.
- [8]Seung D,Lee L.Algorithms for non-negative matrix factorization[J].Advances in neural information processing systems,2001,13:556-562.
- [9]Aggarwal C C ,Reddy C K .Data Clustering:Algorithms and Applications[M].Chapman & Hall/CRC,2013.
- [10]Qi Y,Comon P ,Lim LH.Uniqueness of nonnegative tensor approximations[J].IEEE Transactions on Information Theory,2016,62(4):2170-2183.
- [11]Cichocki A,Phan A H.Fast local algorithms for large scale nonnegative matrix and tensor factorizations[J].IEICE transactions on fundamentals of electronics,communications and computer sciences,2009,92(3):708-721.