Multivariate Process Monitoring and Fault Identification Using Convolutional Neural Networks
In multivariate process control (MPC), the conventional multivariate quality control charts (e.g., T2) have been shown to be efficient for out-of-control signals detection based upon an overall statistic, whereas do not relieve the need for multivariate process pattern recognition (MPPR). MPPR is very beneficial to locate the assignable causes that lead to the out-of-control situation in multivariate manufacturing process. Both Support Vector Machines (SVM) and Convolutional Neural Networks (CNN) techniques have been widely applied and obtained many successes in image and visual analysis, but both methods have some weakness. Therefore, we propose a hybrid system that composes the mentioned techniques. Firstly, two different structure of CNNs were pre-trained as feature extractor due to the capability of unsupervised feature learning. The feature extracted by two CNNs were combined to train a SVM classifier. Experimental analysis show that the proposed hybrid system presents better performance than the isolated stand-alone systems.
KeywordsMultivariate process pattern recognition Convolutional neural networks Support vector machines Feature learning
This research was supported by the National Natural Science Foundation of China (No. 51375290, 71777173).
- 1.H. Hotelling, Multivariate quality control—illustrated by the air testing of sample bombsights, in Techniques of Statistical Analysis, ed. by C. Eisenhart, M.W. Hastay, W.A. Wallis (McGraw-Hill, NY, 1947), pp. 11–184Google Scholar
- 9.A. Krizhevsky, I. Sutskever, G.E. Hinton, ImageNet classification with deep convolutional neural networks, in Advances in Neural Information Processing Systems (2012)Google Scholar
- 10.X. Han, K. Rasul, R. Vollgraf, Fashion-mnist: a novel image dataset for benchmarking machine learning algorithms, arXiv:1708.07747 (2017)
- 11.A.F. Agarap, A neural network architecture combining gated recurrent unit (GRU) and support vector machine (SVM) for intrusion detection in network traffic data, arXiv:1709.03082 (2017)
- 12.M. Abadi et al., Tensorflow: large-scale machine learning on heterogeneous distributed systems, arXiv:1603.04467 (2016)
- 13.F. Pedregosa et al., Scikit-learn: machine learning in Python. J. Mach. Learn. Res. 12, 2825–2830 (2011)Google Scholar
- 14.L. Maaten, G. Hinton, Visualizing data using t-SNE. J. Mach. Learn. Res. 9, 2579–2605 (2008)Google Scholar