Deep learning-based classification of production defects in automated-fiber-placement processes

Production Engineering - Tập 13 - Trang 501-509 - 2019
Carsten Schmidt1, Tristan Hocke1, Berend Denkena1
1Institute of Production Engineering and Machine Tools, Leibniz Universität Hannover, Stade, Germany

Tóm tắt

This paper presents a deep learning-based approach for the detection and classification of production defects that complements an existing thermographic online monitoring system for Automated-Fiber-Placement (AFP) processes. The detection and classification procedure is performed in two stages. In the first stage, the system monitors each tow individually and classifies its process status. Furthermore, it detects and classifies production defects that affect individual tows such as a tow-twist. In the second stage, the system monitors the total width of the faultless tows. In this stage, production defects effecting multiple tows, for example gaps or overlaps, are detected and classified. Twelve different deep convolution neural networks (CNN) with three various architectures are learned supervised relating to different data sets. The performance of both identification stages is explored separately before the entire system will be set up. Therefore, the thermal images of the data sets are superimposed by noise to test the performance of the selected CNN.

Tài liệu tham khảo

Rudberg T, Neilson J, Henschied M, Cemenska J (2014) Improving AFP cell performance. SAE Int J Aerosp Manuf Autom Fast Conf Halbritter A, Harper R (2012) Big parts demand big changes to the fiber placement status quo. SME Composites Manufacturing, Mesa Denkena B, Schmidt C, Völtzer K, Hocke T (2016) Thermographic online monitoring system for automated fiber placement processes. Composite Part B. https://doi.org/10.1016/j.compositesb.2016.04.076 Alexandra K, Linb S, Brabandta D, Böhlkeb T, Lanzaa G (2014) Quality control in the production process of SMC lightweight material. In: Proceedings of the 47th CIRP conference on manufacturing, vol. 47, pp 772–777. https://doi.org/10.1016/j.procir.2014.01.138 Schulz M, Goldbach S, Heuer H, Meyendorf N (2011) Ein Methodenvergleich - ZfP an Kohlefaserverbundwerkstoffen mittels wirbelstrom- und ultraschallbasierender Prüfverfahren. DGZfP, Jahrestagung Lukaszewicz D (2011) Optimization of high-speed automated layup of thermoset carbon fiber preimpregnates. Dissertation, University of Bristol Schmidt C, Denkena B, Hocke T, Völtzer K (2017) Thermal imaging as a solution for reliable monitoring of AFP processes, 3. ACM Automated Composites Manufacturing, Montreal rd Völtzer K (2018) Online process monitoring of the Automated Fiber Placement process based on thermal imaging. Dissertation, Leibniz University Hannover Weimer D, Scholz-Reiter B, Shpitalni M (2016) Design of deep convolutional neural network architectures for automated feature extraction in industrial inspection. CIRP Ann Manuf Technol 65:417–420. https://doi.org/10.1016/j.cirp.2016.04.072 Masci J, Meier U, Ciresan D, Schmidhuber J, Fricout G (2012) Steel defect classification with Max-Pooling Convolutional Neural Networks. IJCNN. https://doi.org/10.1109/IJCNN.2012.6252468 Simonyan K, Zisserman A (2015) Very deep convolutional networks for large-scale image recognition. ICLR Mlakic D, Nikolovski S, Majdandzic L (2018) Deep learning method and infrared imaging as a tool for transformer faults detection. J Electr Eng 6:98–106. https://doi.org/10.17265/2328-2223/2018.02.006 Marani R, Palumbo D, Galietti U, Stella E, D’Orazio T (2016) Automatic detection of subsurface defects in composite materials using thermograpgy and unsurpervised machine learning. 8th IS. https://doi.org/10.1109/IS.2016.7737471 Groß L, Herwig A, Berg DC, Schmidt C, Denkena B, Horst P, Meiners D (2018) Production-based design of a hybrid load introduction element for thin-walled CFRP structures. Prod Eng Res Devel 12:113–120. https://doi.org/10.1007/s11740-018-0821-4