Fault classification based on semi-supervised dense ladder network

doi:10.11949/j.issn.0438-1157.20171488

Abstract

Abstract:

To meet the needs of fault identification and imbalanced dataset in actual industrial processes, a semisupervised deep learning method was proposed for fault classification. The enhanced semi-supervised ladder network, i.e., semi-supervised dense ladder network, was developed by improving network architecture and loss function. Dense connection strategy was applied in the network architecture to maximize information flow between layers of the ladder network, such that features among layers could be transmitted and copied. Meanwhile, to ensure consistency between training target and prediction output, prediction output loss of corrupted encoder was added into original loss function. Experiment results illustrated the proposed method could achieve ideal classification in small ratio marked dataset in actual industrial process.

Key words: semi-supervised learning, ladder network, dense connection, fault classification, algorithm, neural networks, optimization, auto encoder

摘要：

针对工业过程故障识别的需要和实际工业数据小比例有标签、大比例无标签的特点，研究了基于深度学习的半监督故障分类方法。在半监督阶梯网络的基础上，通过对网络结构和损失函数的改进，提出了半监督密集阶梯网络算法。该算法改进了原始的网络结构，添加了各层之间的密集连接，尝试最大化阶梯网络内部的数据信息流，使得各编码解码层之间的特征得以传递和复用。针对损失函数的特点，添加了无噪声编码层的预测输出损失，确保训练目标与模型输出一致。实验结果证明了所提出的新方法能在工业过程的小比例有标签数据情况下，获得理想的分类效果。

关键词: 半监督学习, 阶梯网络, 密集连接, 工业故障分类, 算法, 神经网络, 优化, 自编码器

CLC Number:

TP181

SHI Fangyi, WANG Ziyang, LIANG Jun. Fault classification based on semi-supervised dense ladder network[J]. CIESC Journal, 2018, 69(7): 3083-3091.

施方迤, 汪子扬, 梁军. 基于半监督密集阶梯网络的工业故障识别[J]. 化工学报, 2018, 69(7): 3083-3091.

References 20

[1]	ZHU X. Semi-supervised learning literature survey[J]. Computer Science, 2008, 37(1):63-77.
[2]	HINTON G E, OSINDERO S, TEH Y W. A fast learning algorithm for deep belief nets[J]. Neural Computation, 2006, 18(7):1527-1554.
[3]	KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neural networks[C]//Proceedings of the 25th International Conference on Neural Information Processing Systems. 2012:1097-1105.
[4]	SZEGEDY C, LIU W, JIA Y, et al. Going deeper with convolutions[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2015:1-9.
[5]	GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2014:580-587.
[6]	REN S, HE K, GIRSHICK R, et al. Faster R-CNN:towards real-time object detection with region proposal networks.[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2015, 39(6):1137-1149.
[7]	VINCENT P, LAROCHELLE H, LAJOIE I, et al. Stacked denoising autoencoders:learning useful representations in a deep network with a local denoising criterion[J]. Journal of Machine Learning Research, 2010, 11(12):3371-3408.
[8]	RASMUS A, VALPOLA H, HONKALA M, et al. Semi-supervised learning with ladder networks[J]. Computer Science, 2015, 9(Suppl. 1):1-9.
[9]	VALPOLA H. From neural PCA to deep unsupervised learning[M]//Advances in Independent Component Analysis and Learning Machines. Academic Press, 2015:143-171.
[10]	HUANG G, LIU Z, WEINBERGER K Q, et al. Densely connected convolutional networks[J]. arXiv preprint arXiv:1608.06993, 2016.
[11]	SCHÖLKOPF B, PLATT J, HOFMANN T. Efficient learning of sparse representations with an energy-based model[C]//Advances in Neural Information Processing Systems. 2006:1137-1144.
[12]	PEZESHKI M, FAN L, COURVILLE A, et al. Deconstructing the ladder network architecture[C]//International Conference on Machine Learning. JMLR Org., 2016:2368-2376.
[13]	IOFFE S, SZEGEDY C. Batch normalization:accelerating deep network training by reducing internal covariate shift[C]//International Conference on Machine Learning. 2015:448-456.
[14]	NAIR V, HINTON G E. Rectified linear units improve restricted boltzmann machines[C]//Proceedings of the 27th International Conference on Machine Learning (ICML-10). 2010:807-814.
[15]	SRIVASTAVA R K, GREFF K, SCHMIDHUBER J. Training very deep networks[C]//Advances in Neural Information Processing Systems. 2015:2377-2385.
[16]	LARSSON G, MAIRE M, SHAKHNAROVICH G. FractalNet:ultradeep neural networks without residuals[J]. arXiv preprint arXiv:1605.07648, 2016.
[17]	HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2016:770-778.
[18]	DOWNS J J, VOGEL E F. A plant-wide industrial process control problem[J]. Computers & Chemical Engineering, 1993, 17(3):245-255.
[19]	BLAKE C L, MERZ C J. UCI repository of machine learning databases[J]. Department of Information and Computer Science, 1998, doi:10.1234/12345678.
[20]	SHEN Y, DING S X, HAGHANI A, et al. A comparison study of basic data-driven fault diagnosis and process monitoring methods on the benchmark Tennessee Eastman process[J]. Journal of Process Control, 2012, 22(9):1567-1581.