摘要
针对复杂机电装备故障诊断中存在的数据量大、提取故障特征困难等问题,结合深度学习理论强大的感知与自我学习能力,提出一种基于深度信念网络和多信息融合的复杂机电装备故障诊断方法。将多个传感器的原始时域信号数据输入深度信念网络进行训练,通过反向微调学习对深度信念网络进行整体微调,提高分类准确性,同时在训练过程采用ReLu激活函数和加入Batch Normalization,减少过拟合出现概率的同时提高了网络收敛的速度。将此方法运用到复杂数控加工中心刀具的故障诊断中,结果表明该方法相比传统BPNN算法和采用Sigmoid激活函数的深度神经网络算法准确率更高。
In view of the existing problem of complex electromechanical equipment fault diagnosis,such as large amount of data,the fault feature extraction difficulties,combining with strong perception and self-learning ability of the theory of deep learning,a complex electromechanical equipment fault diagnosis method based on deep belief networks and multi information fusion was proposed. The original time domain signal data were input into the deep belief network to train,and the whole network was adjusted by reverse trimming to improve the classification accuracy. At the same time,Batch Normalization and Re Lu activation function were added to the training process to reduce the chance of over fitting and improve the convergence speed of the network. This method was applied to the fault diagnosis of tool of the complex numerically controlled production center. The results show that this method is more accurate than the traditional BP neural network algorithm and deep neural network algorithm using Sigmoid activation function.
引文
[1]LEE J,NI J,DJURDJANOVIC D,et al.Intelligent Prognostics Tools and E-maintenance[J].Computers in Industry,2006(57):476-489.
[2]ANDREW K S,LIN D M,BANJEVIC D.A Review on Machinery Diagnostics and Prognostics Implementing Condition-based Maintenance[J].Mechanical Systems and Signal Processing,2006(20):1483-1510.
[3]曹立军,王兴贵,秦俊奇.融合案例与规则推理的故障预测专家系统[J].计算机工程,2006,32(1):208-210.CAO L J,WANG X G,QIN J Q.Fault Forecast Expert System Based on Integration of Case and Rule Reasoning[J].Computer Engineering,2006,32(1):208-210.
[4]李凌均,张周锁,何正嘉.基于支持向量机的机械设备状态趋势预测研究[J].西安交通大学学报,2004,38(3):230-233.LI L J,ZHANG Z S,HE Z J.Research on Condition Trend Prediction of Mechanical Equipment Based on Support Vector Machine[J].Journal of Xi’an Jiaotong University,2004,38(3):230-233.
[5]许丽佳,龙兵,王厚军.基于LSSVM-HMM的发射机故障预测研究[J].仪器仪表学报,2008,29(1):21-26.XU L J,LONG B,WANG H J.Study on Fault Prediction of Transmitter Based on LSSVM-HMM[J].Chinese Journal of Scientific Instrument,2008,29(1):21-26.
[6]HINTON G E,SALAKHUTDINOV R R.Reducing the Dimensionality of Data with Neural Networks[J].Science,2006,313:504-507.
[7]MOHAMED A,DAHL G.Acoustic Modeling Using Deep Belief Networks[J].IEEE Transaction on Audio,Speech,and Language Processing,2012,20(1):23-29.
[8]JAITLY N,HINTON G.Learning a Better Representation of Speech Sound Waves Using Restricted Boltzmann Machines[C]//Proceedings of IEEE International Conference on Acoustics,Speech and Signal Processing,2011:5884-5887.
[9]DIELEMAN S,SCHRAUWEN B.End-to-end Learning for Music Audio[C]//Proceedings of IEEE International Conference on Acoustics,Speech and Signal Processing,2014:6964-6968.
[10]HOSHEN Y,WEISS R J,WILSON K W.Speech Acoustic Modeling from Raw Multichannel Waveforms[C]//Proceedings of IEEE International Conference on Acoustics,Speech and Signal Processing,2015:4624-4628.
[11]SAINATH T N,WEISS R J,WILSON K W,et al.Learning the Speech Front-end with Raw Waveform CLDNNs[C]//Proceedings of Interspeech,2015:1-5.
[12]TUSKE Z,GOLIK P,SCHLTER R,et al.Acoustic Modeling with Deep Neural Networks Using Raw Time Signal for LVCSR[C]//Proceedings of Interspeech,2014:890-894.
[13]RAHMANI H,MAHMOOD A,HUYNH D Q,et al.HOPC:Histogram of Oriented Principal Components of 3D Pointclouds for Action Recognition[C]//Proceedings of 13th IEEE European Conference on Computer Vision,2014:742-757.
[14]SONG S,LICHTENBERG S P,XIAO J.Sun RGB-D:ARGB-D Scene Understanding Benchmark Suite[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2015:567-576.
[15]KRIZHEVSKY A,SUTSKEVER I,HINTON G E.ImageNet Classification with Deep Convolutional Neural Networks[C]//Proceedings of International Conference on Neural Information Processing Systems.Curran Associates Inc,2012:1097-1105.
[16]雷亚国,贾峰,周昕,等.基于深度学习理论的机械装备大数据健康监测方法[J].机械工程学报,2015,51(21):49-56.LEI Y G,JIA F,ZHOU X,et al.A Deep Learning-based Method for Machinery Health Monitoring with Big Data[J].Journal of Mechanical Engineering,2015,51(21):49-56.
[17]ZHANG Z,ZHU J,PAN F.Fault Detection and Diagnosis for Data Incomplete Industrial Systems with New Bayesian Network Approach[J].Journal of Systems Engineering and Electronics,2013(3):500-511.
[18]王宪保,李洁,姚明海,等.基于深度学习的太阳能电池片表面缺陷检测方法[J].模式识別与人工智能,2014,27(6):517-523.WANG X B,LI J,YAO M H,et al.Solar Cells Surface Defects Detection Based on Deep Learning[J].Pattern Recognition and Artificial Intelligence,2014,27(6):517-523.
[19]谢吉鹏.去平台下基于深度学习的高速列车走行故障诊断技术研究[D].成都:西南交通大学,2015.
[20]张绍辉.基于多路稀疏自编码的轴承状态动态监测[J].振动与冲击,2015,35(19):125-131.ZHANG S H.Bearing Condition Dynamic Monitoring Based on Multi-way Sparse Autocoder[J].Journal of Vibration and Shock,2015,35(19):125-131.
[21]ACKLEY D H,HINTON G E,SEJNOWSKI T J.A Learning Algorithm for Boltzmann Machines[M].Cognitive Science,1985:147-169.
[22]LEE H,GROSSE R,RANGANATH R,et al.Convolutional Deep Belief Networks for Scalable Unsupervised Learning of Hierarchical Representations[C]//Proceedings of International Conference on Machine Learning.Montreal,2009.
[23]BENGIO Y,LAMBLIN P,POPOVICI D,et al.Greedy Layer-wise Training of Deep Networks[J].Advances in Neural Information Processing Systems,2007,19:153.
[24]HINTON G E.Rectified Linear Units Improve Restricted Boltzmann Machines Vinod Nair[C]//Proceedings of International Conference on Machine Learning.Omnipress,2010.
[25]IOFFE S,SZEGEDY C.Batch Normalization:Accelerating Deep Network Training by Reducing Internal Covariate Shift[C]//Proceedings of International Conference on International Conference on Machine Learning.JMLR.org,2015.
[26]TAMILSELVAN P,WANG P.Failure Diagnosis Using Deep Belief Learning Based Health State Classification[J].Reliability Engineering&Systems Safety,2013,115(7):124-135.