A comparative study of classification methods for state recognition in injection molding
-
Kurt Pichler
,
Julian Brunthaler
Abstract
In this paper, different methods for state classification of injection molding cycles without requiring access to the corresponding control signals from the machine control are compared. The states closing, plasticizing, demolding and null state are to be differentiated based on data obtained from sensors retrofitted to the machine. For this purpose, data from a triaxial clamp–on accelerometer is used, which can be easily retrofitted to existing machines. This setup was used to record an extensive data set at several industrial production facilities during the production of different products. Various well known evaluation methods are applied to this problem. For example, features are extracted from the raw data and then classified using logistic regression. The raw data is also classified directly using long short-term memory neural network. Furthermore, the raw data is converted into images using bispectra and then classified using convolutional neural networks or Bag of Features. The results of the individual methods are compared and discussed in terms of accuracy and computational effort.
Zusammenfassung
In dieser Arbeit werden verschiedene Methoden zur Zustandsklassifizierung von Spritzgießzyklen verglichen, ohne dass ein Zugriff auf die entsprechenden Steuersignale der Maschinensteuerung erforderlich ist. Es sollen anhand von Daten, die von an der Maschine nachgerüsteten Sensoren gewonnen werden, die Zustände Schließen, Plastifizieren, Entformen und Nullzustand unterschieden werden. Dazu werden die Daten eines dreiachsigen aufklemmbaren Beschleunigungssensors verwendet, der leicht an bestehenden Maschinen nachgerüstet werden kann. Mit diesem Messaufbau wurde ein umfangreicher Datensatz an mehreren industriellen Produktionsanlagen bei der Herstellung verschiedener Produkte aufgezeichnet. Verschiedene bekannte Methoden werden auf dieses Problem angewandt. So werden beispielsweise Merkmale aus den Rohdaten extrahiert und anschließend mit Hilfe logistischer Regression klassifiziert. Die Rohdaten werden auch direkt mit Hilfe eines Long Short-Term Memory neuronalen Netzes klassifiziert. Darüber hinaus werden die Rohdaten mit Hilfe von Bispektren in Bilder umgewandelt und dann mit Hilfe von Convolutional Neural Networks oder Bag of Features klassifiziert. Die Ergebnisse der einzelnen Methoden werden im Hinblick auf Genauigkeit und Rechenaufwand verglichen und diskutiert.
About the authors
Kurt Pichler received the Dipl-Ing. degree in industrial mathematics from Johannes Kepler University (JKU) Linz, Austria, in 2005, the Dr. techn. degree in technical sciences from JKU in 2014, and the MSc degree in artificial intelligence from JKU in 2023. From 2006 to 2007, he was with the Institute for Design and Control of Mechatronical Systems at the JKU. He joined the Area SENS of the Linz Center of Mechatronics GmbH in 2007 and works currently as a senior engineer and project leader. His research interests include feature engineering, machine learning, artificial intelligence, fault diagnosis and predictive maintenance.
Julian Brunthaler holds a Master degree in Artificial Intelligence from Johannes Kepler University Linz and a bachelor degree in Mathematics from the University of Passau. From 2021 to 2023, he worked as a data scientist at AISEMO GmbH. Since 2023, he has been an AI engineer at Pataky Software GmbH. Julian has (co-)authored several publications, including papers on state classification in injection molding cycles and state recognition in injection molding based on accelerometer data sets.
Veronika Putz received her Dipl.-Ing. (equivalent to MSc.) in Electrical Engineering from Graz University of Technology and her Dr. techn. (equivalent to PhD.) from Johannes Kepler University Linz. She joined the Linz Center of Mechatronics GmbH in 2012 as a senior researcher. Since 2020 she leads the team “Data Analysis & AI” and coordinates related research and development topics in this area. She has raised significant funding and has been actively involved in national and international projects, as well as projects directly funded by industrial partners. Her main research interests include sensors and metrology, imaging technologies in sensing, condition monitoring and predictive maintenance, and various other applications utilizing classical signal processing methods as well as machine learning.
Sandra Schober received her Dipl.-Ing. Degree in industrial mathematics from Johannes Kepler University Linz, Austria, in 2013. From 2012 to 2013, she wrote her master thesis at the Linz Center of Mechatronics GmbH, and afterwards, she also started working there as research & development engineer. Now, she additionally leads projects in the field of feature engineering, machine learning and artificial intelligence, which are also her research interests.
Wolfgang Auer studied business informatics at Johannes Kepler University, Linz. He was working as a software engineer for a farm equipment company. From 2009 to 2019, he has been working to build up his startup Smartbow GmbH, Weibern, Austria, as CEO. Smartbow was awarded multiple times with the DLG Eurotier Gold Medal for the best innovation in precision livestock farming. From 2019 to 2024, he was also CEO and co-founder of AISEMO GmbH. He currently works for Pataky Software GmbH as head of marketing and sales.
Acknowledgments
This work has been supported by the COMET-K2 Center of the Linz Center of Mechatronics (LCM) funded by the Austrian federal government and the federal state of Upper Austria.
-
Research ethics: Not applicable.
-
Informed consent: Not applicable.
-
Author contributions: All authors have accepted responsibility for the entire content of this manuscript and approved its submission.
-
Use of Large Language Models, AI and Machine Learning Tools: None declared.
-
Conflict of interest: The authors state no conflict of interest.
-
Research funding: None declared.
-
Data availability: Not applicable.
References
[1] S. K. Selvaraj, A. Raj, R. R. Mahadevan, U. Chadha, and V. Paramasivam, “A review on machine learning models in injection molding machines,” Adv. Mater. Sci. Eng., vol. 2022, 2022.10.1155/2022/1949061Search in Google Scholar
[2] H. Ha and J. Jeong, “CNN-based defect inspection for injection molding using edge computing and industrial IoT systems,” Appl. Sci., vol. 11, no. 14, p. 6378, 2021. https://doi.org/10.3390/app11146378.Search in Google Scholar
[3] T. Mao, Y. Zhang, H. Zhou, D. Li, Z. Huang, and H. Gao, “Data driven injection molding process monitoring using sparse auto encoder technique,” in 2015 IEEE International Conference on Advanced Intelligent Mechatronics (AIM), 2015, pp. 524–528.10.1109/AIM.2015.7222587Search in Google Scholar
[4] E. Lughofer and K. Pichler, “Data-driven prediction of possible quality deterioration in injection molding processes,” Appl. Soft Comput., vol. 150, 2024, Art. no. 111029, https://doi.org/10.1016/j.asoc.2023.111029.Search in Google Scholar
[5] M.-H. Tsai, et al.., “Development of an online quality control system for injection molding process,” Polymers, vol. 14, no. 8, p. 1607, 2022. https://doi.org/10.3390/polym14081607.Search in Google Scholar PubMed PubMed Central
[6] H. Jung, J. Jeon, D. Choi, and J.-Y. Park, “Application of machine learning techniques in injection molding quality prediction: implications on sustainable manufacturing industry,” Sustainability, vol. 3, no. 4120, 2021.10.3390/su13084120Search in Google Scholar
[7] V. Ketonen and J. O. Blech, “Anomaly detection for injection molding using probabilistic deep learning,” in 2021 4th IEEE International Conference on Industrial Cyber-Physical Systems (ICPS), 2021.10.1109/ICPS49255.2021.9468190Search in Google Scholar
[8] D. N. Chi Nam, T. Van Tung, and E. Y. K. Yee, “Quality monitoring for injection moulding process using a semi-supervised learning approach,” in IECON 2021 – 47th Annual Conference of the IEEE Industrial Electronics Society, 2021.Search in Google Scholar
[9] N. John Punnoose, P. Vadakkepat, A.-P. Loh, and E. K. Y. Yap, “Data-driven quality estimation for production processes with lot-level quality control,” in IECON 2021 – 47th Annual Conference of the IEEE Industrial Electronics Society, 2021.10.1109/IECON48115.2021.9589245Search in Google Scholar
[10] M. Charest, R. Finn, and R. Dubay, “Integration of artificial intelligence in an injection molding process for on-line process parameter adjustment,” in 2018 Annual IEEE International Systems Conference (SysCon), 2018.10.1109/SYSCON.2018.8369500Search in Google Scholar
[11] C.-C. Tsai and C.-H. Lu, “Adaptive decoupling predictive temperature control using neural networks for extrusion barrels in plastic injection molding machines,” in 2015 IEEE International Conference on Systems, Man, and Cybernetics, 2015, pp. 353–358.10.1109/SMC.2015.73Search in Google Scholar
[12] J. Z. Zhang, “Development of an in-process poka-yoke system utilizing accelerometer and logistic regression modeling for monitoring injection molding flash,” Int. J. Adv. Manuf. Technol., vol. 71, no. 9, pp. 1793–1800, 2014. https://doi.org/10.1007/s00170-013-5604-7.Search in Google Scholar
[13] E. Esteves Moreira, et al.., “Industry 4.0: real-time monitoring of an injection molding tool for smart predictive maintenance,” in 2020 25th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA), vol. 1, 2020.10.1109/ETFA46521.2020.9212167Search in Google Scholar
[14] J. Brunthaler, P. Grabski, V. Sturm, W. Lubowski, and D. Efrosinin, “On the problem of state recognition in injection molding based on accelerometer data sets,” Sensors, vol. 22, no. 16, p. 6165, 2022. https://doi.org/10.3390/s22166165.Search in Google Scholar PubMed PubMed Central
[15] K. Pichler, et al.., “State classification in injection molding cycles using transformation of acceleration data into images,” in 2023 IEEE 24th International Conference on Information Reuse and Integration for Data Science (IRI), 2023, pp. 115–120.10.1109/IRI58017.2023.00027Search in Google Scholar
[16] K. Pichler, T. Ooijevaar, C. Hesch, C. Kastl, and F. Hammer, “Data-driven vibration-based bearing fault diagnosis using non-steady-state training data,” J. Sens. Sens. Syst., vol. 9, no. 1, pp. 143–155, 2020. https://doi.org/10.5194/jsss-9-143-2020.Search in Google Scholar
[17] S. Belagoune, N. Bali, A. Bakdi, B. Baadji, and K. Atif, “Deep learning through LSTM classification and regression for transmission line fault detection, diagnosis and location in large-scale multi-machine power systems,” Measurement, vol. 177, 2021, Art. no. 109330, https://doi.org/10.1016/j.measurement.2021.109330.Search in Google Scholar
[18] S. Fleischanderl, “Time series transformation into images for production state recognition,” Master’s thesis, Johannes Kepler University Linz, 2023.Search in Google Scholar
[19] S. Panda and C. Panda, “A review on image classification using bag of features approach,” Int. J. Comput. Sci. Eng., vol. 7, no. 6, pp. 538–542, 2019. https://doi.org/10.26438/ijcse/v7i6.538542.Search in Google Scholar
[20] I. Guyon and A. Elisseeff, “An introduction to variable and feature selection,” J. Mach. Learn. Res., vol. 3, 2003.Search in Google Scholar
[21] J. G. Dy and C. E. Brodley, “Feature selection for unsupervised learning,” J. Mach. Learn. Res., vol. 5, 2004.Search in Google Scholar
[22] S. Marsland, Machine Learning – An Algorithmic Perspective, Boca Raton, FL, Chapman & Hall/CRC, 2009.Search in Google Scholar
[23] B. Lei, et al.., Classification, Parameter Estimation and State Estimation: An Engineering Approach Using MATLAB, 2nd ed Hoboken, NJ, John Wiley & Sons, 2017.Search in Google Scholar
[24] A. R. Webb and K. D. Copsey, Statistical Pattern Recognition, Chichester, West Sussex, John Wiley & Sons, Ltd, 2011.10.1002/9781119952954Search in Google Scholar
[25] S. Hochreiter and J. Schmidhuber, “Long short-term memory,” Neural Comput., vol. 9, no. 8, pp. 1735–1780, 1997. https://doi.org/10.1162/neco.1997.9.8.1735.Search in Google Scholar PubMed
[26] A. Krizhevsky, I. Sutskever, and G. E. Hinton, “Imagenet classification with deep convolutional neural networks,” in Proceedings of the 25th International Conference on Neural Information Processing Systems – Volume 1, NIPS’12, Red Hook, NY, USA, Curran Associates Inc, 2012, pp. 1097–1105.Search in Google Scholar
[27] C. Szegedy, et al.., “Going deeper with convolutions,” in 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015, pp. 1–9.10.1109/CVPR.2015.7298594Search in Google Scholar
[28] K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” International Conference on Learning Representations (ICLR), 2014.Search in Google Scholar
[29] U. Greb and M. G. Rusbridge, “The interpretation of the bispectrum and bicoherence for non-linear interactions of continuous spectra,” Plasma Phys. Controlled Fusion, vol. 30, no. 5, pp. 537–549, 1988. https://doi.org/10.1088/0741-3335/30/5/005.Search in Google Scholar
[30] H. Joseph Weaver, Theory of Discrete and Continuous Fourier Analysis, USA, John Wiley & Sons, Inc., 1989.Search in Google Scholar
[31] The MathWorks Inc., “Matlab version: 9.13.0 (R2022b),” 2022.Search in Google Scholar
[32] The MathWorks Inc., “Image processing and computer vision toolbox,” 2022.Search in Google Scholar
[33] K. Mikolajczyk, B. Leibe, and B. Schiele, “Local features for object class recognition,” in Tenth IEEE International Conference on Computer Vision (ICCV’05) Volume 1, vol. 2, 2005, pp. 1792–1799.10.1109/ICCV.2005.146Search in Google Scholar
[34] T. Tuytelaars and K. Mikolajczyk, Local Invariant Feature Detectors: A Survey, Hanover, MA, Now Foundations and Trends, 2008.10.1561/9781601981394Search in Google Scholar
[35] H. Bay, A. Ess, T. Tuytelaars, and L. Van Gool, “Speeded-up robust features (SURF),” Comput. Vis. Image Understand., vol. 110, no. 3, pp. 346–359, 2008. https://doi.org/10.1016/j.cviu.2007.09.014.Search in Google Scholar
[36] P. Goodarzi, A. Schütze, and T. Schneider, “Comparison of different ML methods concerning prediction quality, domain adaptation and robustness,” TM – Tech. Mess., vol. 89, no. 4, pp. 224–239, 2022. https://doi.org/10.1515/teme-2021-0129.Search in Google Scholar
© 2024 Walter de Gruyter GmbH, Berlin/Boston
Articles in the same Issue
- Frontmatter
- Editorial
- Measurement systems and sensors with cognitive features III
- Research Articles
- Opportunities of artificial intelligence in the field of calibration services
- Neuronale Netze zur Startwertschätzung bei der Identifikation piezoelektrischer Materialparameter
- Spectral reconstruction using neural networks in filter-array-based chip-size spectrometers
- Prediction of water absorption of recycled coarse aggregate based on deep learning image segmentation
- Surface mortar detection and performance evaluation of recycled aggregates based on hyperspectral technology
- A comparative study of classification methods for state recognition in injection molding
Articles in the same Issue
- Frontmatter
- Editorial
- Measurement systems and sensors with cognitive features III
- Research Articles
- Opportunities of artificial intelligence in the field of calibration services
- Neuronale Netze zur Startwertschätzung bei der Identifikation piezoelektrischer Materialparameter
- Spectral reconstruction using neural networks in filter-array-based chip-size spectrometers
- Prediction of water absorption of recycled coarse aggregate based on deep learning image segmentation
- Surface mortar detection and performance evaluation of recycled aggregates based on hyperspectral technology
- A comparative study of classification methods for state recognition in injection molding
