Fault detection in flotation processes based on deep learning and support vector machine

Abstract

Effective fault detection techniques can help flotation plant reduce reagents consumption, increase mineral recovery, and reduce labor intensity. Traditional, online fault detection methods during flotation processes have concentrated on extracting a specific froth feature for segmentation, like color, shape, size and texture, always leading to undesirable accuracy and efficiency since the same segmentation algorithm could not be applied to every case. In this work, a new integrated method based on convolution neural network (CNN) combined with transfer learning approach and support vector machine (SVM) is proposed to automatically recognize the flotation condition. To be more specific, CNN function as a trainable feature extractor to process the froth images and SVM is used as a recognizer to implement fault detection. As compared with the existed recognition methods, it turns out that the CNN-SVM model can automatically retrieve features from the raw froth images and perform fault detection with high accuracy. Hence, a CNN-SVM based, real-time flotation monitoring system is proposed for application in an antimony flotation plant in China.

Keywords

flotation processes / convolutional neural network / support vector machine / froth images / fault detection

References

[1] YangC-h, ZhouK-j, MouX-m, GuiW-hua. Froth color and size measurement method for flotation based on computer vision [J]. Chinese Journal of Scientific Instrument, 2009, 30(4): 717-721

[2] XuC-h, GuiW-h, YangC-h, ZhuH-q, LinY-q, ShiCao. Flotation process fault detection using output PDF of bubble size distribution [J]. Minerals Engineering, 2012, 26: 5-12

[3] ZhaoL, PengT, HanH, CaoW, LouY-g, XieX-tian. Fault condition recognition based on multi-scale co-occurrence matrix for copper flotation process [C]//. Proceedings of the 19th World Congress International Federation of Automatic Control, 2014, South Africa, Cape Town: 70917097

[4] AldrichC, MaraisC, SheanB J, CilliersJ J. Online monitoring and control of froth flotation systems with machine vision: A review [J]. International Journal of Mineral Processing, 2010, 96(1): 1-13

[5] YangC-h, XuC-h, GuiW-h, DuJ-jiang. Nonparametric density estimation of bubble size distribution for monitoring mineral flotation process [C]//. Proceedings of the 28th Chinese Control Conference, 2009, China, Shanghai: 29412945

[6] LiuJ-p, GuiW-h, TangZ-h, YangC-h, ZhuJ-y, LiJ-qi. Recognition of the operational statuses of reagent addition using dynamic bubble size distribution in copper flotation process [J]. Minerals Engineering, 2013, 45: 128-141

[7] LiZ-m, GuiW-hua. The method of reagent control based on time series distribution of bubble size in a gold-antimony flotation process [J]. Asian Journal of Control, 2018, 20(6): 2223-2236

[8] YangC-h, RenH-f, GuiW-h, YanF, TangZ-hui. Performance recognition using texture credit distributed SVM for froth flotation process [J]. Chinese Journal of Scientific Instrument, 2011, 32(10): 2205-2209

[9] HeM-f, YangC-h, WangX-l, GuiW-h, WeiL-jun. Nonparametric density estimation of froth colour texture distribution for monitoring sulphur flotation process [J]. Minerals Engineering, 2013, 53: 203-212

[10] LingY-q, YangC-h, HeM-f, GuiW-hua. Fault condition detection for sulfur flotation process based on texture unit distribution [J]. Computing Technology and Automation, 2013, 32(1): 28-31

[11] GuiW-h, LiuJ-p, YangC-h, ChenN, LiaoXi. Color co-occurrence matrix based froth image texture extraction for mineral flotation [J]. Minerals Engineering, 2013, 46: 60-67

[12] LecunY, HuangF J, BottouL. Learning methods for generic object recognition with invariance to pose and lighting [C]//. Proceedings of the 2004 IEEE Computer Society Conference, 2004, USA, Washington DC: 97104

[13] FuY, AldrichC. Froth image analysis by use of transfer learning and convolutional neural networks [J]. Minerals Engineering, 2018, 115: 68-78

[14] SzegedyC, LiuW J, SermanetP, ReedS, AngurlovD, RabinovichA. Going deeper with convolutions [C]//. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015, USA, Boston, MA: 19

[15] LecunY, BottouL, BengioY, HaffnerP. Gradient-based learning applied to document recognition [J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324

[16] BouvrieJNotes on convolutional neural networks [R], 2006, USA, Cambridge

[17] ByunH, LeeS W. A survey on pattern recognition applications of support vector machines [J]. International Journal of Pattern Recognition and Artificial Intelligence, 2003, 17(3): 459-486

[18] NgH W, NguyenV D, VonikakisV, WinklerS. Deep learning for emotion recognition on small datasets using transfer learning [C]//. Proceedings of the 2015 ACM on International Conference on Multimodal Interaction, 2015, USA, Seattle: 443449

[19] SzegedyC, VanhouckeV, IoffeS, ShlensJ, WojnaZ. Rethinking the inception architecture for computer vision [C]//. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, USA, Las Vegas: 28182826

[20] MaatenL V D, HintonG. Visualizing data using t-SNE [J]. Journal of Machine Learning Research, 2008, 9: 2579-2605

[21] HsuC W, ChangC C, LinC J. A practical guide to support vector classification. [EB/OL], 2003

[22] MaA-l, XuD-g, XieY-f, YangC-h, GuiW-hua. Analysis of dynamic texture features of flotation froth images based on space-time characteristics of complex networks [J]. Journal of Chemical Industry and Engineering, 2016, 68(3): 1023-1031