高效的机械设备故障诊断是保证机械系统正常运行的必要条件，针对传统的故障诊断过于依赖复杂的人工特征提取，已经不能满足实际诊断要求的问题，归纳综述了将物联网(internet of things，IoT)与深度学习相结合进行故障诊断这一具有潜力的技术。阐述了 IoT在设备故障诊断中的便利与优势，对当前基于深度学习的故障诊断就卷积神经网络(convolutional neural network，CNN)、生成对抗网络(generative adversurial network，GAN)、自动编码器(autoencoder，AE)和深度置信网络(deep belief networks，DBN)展开，介绍了各种算法以及应用场景，同时列举了近年来国内外学者在机械设备故障诊断方面开展的研究情况，并比较了不同应用场景下的优势与不足。最后，对于将物联网与深度学习相结合应用在机械设备的故障诊断方向上的未来发展做出展望。
A review of IoT and deep learning based fault diagnosis for mechanical equipments
Efficient fault diagnosis of mechanical equipment is essential to ensure the normal operation of mechanical system.In view of the problem that traditional fault diagnosis relies too much on complex artificial feature extraction,which can no longer meet the requirements of actual diagnosis,this paper summarizes the potential technology of fault diagnosis by combining the internet of things(IoT)and deep learning.This paper describes the convenience and advantages of the IoT in equipment fault diagnosis,introduces various algorithms and application scenarios of deep learning based fault diagnosis on convolutional neural network(CNN),generative adverssion network(GAN),autoencoder(AE)and deep belief network(DBN),and lists the research of mechanical equipment fault diagnosis carried out by domestic and foreign scholars in recent years.What's more,the advantages and disadvantages of different application scenarios are compared.Finally,the development of combining the Internet of Things and deep learning in the direction of mechanical equipment fault diagnosis is prospected.
fault diagnosisinternet of thingsreviewdeep learningdata processing
NETL
NSTL
万方数据
