Abstract
Carbon fiber reinforced polymers (CFRP) are lightweight but strong composite materials designed to reduce the weight of aerospace or automotive components – contributing to reduced greenhouse gas emissions. A common manufacturing process for carbon fiber tapes consists of aligning tows (bundles of carbon fiber filaments) side by side to form tapes via a spreading machine. Tows are pulled across metallic spreading bars that are conventionally kept in a fixed position. That can lead to high variations in quality metrics such as tape width or height. Alternatively, one could try to control the spreading bars based on the incoming tows’ profiles. We investigate whether a machine learning approach, consisting of a supervised process model trained on real data and a process control model to choose adequate spreading bar positions is able to improve the tape quality variations. Our results indicate promising tendencies for adaptive tow spreading.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
Notes
- 1.
Data available via https://doi.org/10.6084/m9.figshare.12213746.v1 Code is available here: https://github.com/isse-augsburg/adaptive-spreading.
References
Appel, L., Kerber, A., Abbas, B., Jeschke, S., Gries, T.: Determination of interactions between bar spreading process parameters and spreading quality for the development of an automated quality control of spread high modulus fiber tows. In: Proceeding of the 17th European Conference on Composite Materials, ECCM 2016, 17, pp. 26–30 (2016)
Breiman, L.: Random Forests. Machine Learning 45, 5–32 (2001). https://doi.org/10.1023/A:1010933404324
Floreano, D., Dürr, P., Mattiussi, C.: Neuroevolution: from architectures to learning. Evol. Intel. 1(1), 47–62 (2008). https://doi.org/10.1007/s12065-007-0002-4
Gizik, D., Metzner, C., Weimer, C., Middendorf, P.: Spreading of heavy tow carbon fibers for the use in aircraft structures. In: Proceeding of the 17th European Conference on Composite Materials, ECCM 2016, 17 (2016)
MacKenzie, H.D.J. et al.: On the Road toward 2050: Report Massachusetts Institute of Technology Potential for Substantial Reductions in Light-Duty Vehicle Energy Use and Greenhouse Gas Emissions. Massachusetts Institute of Technology (2015). https://energy.mit.edu/publication/on-the-road-toward-2050/
Irfan, M.S., et al.: Lateral spreading of a fiber bundle via mechanical means. J. Compos. Mater. 46(3), 311–330 (2012)
Mandel, T., Liu, Y.E., Brunskill, E., Popovíc, Z.: Offline evaluation of online reinforcement learning algorithms. In: 30th AAAI Conference on Artificial Intelligence, AAAI 2016, pp. 1926–1933 (2016)
Marissen, R., van der Drift, L.T., Sterk, J.: Technology for rapid impregnation of fibre bundles with a molten thermoplastic polymer. Compos. Sci. Technol. 60(10), 2029–2034 (2000)
Morgan, P.: Carbon Fibers and Their Composites. CRC Press, Boca Raton (2005)
Nagabandi, A., Kahn, G., Fearing, R.S., Levine, S.: Neural network dynamics for model-based deep reinforcement learning with model-free fine-tuning. In: 2018 IEEE International Conference on Robotics and Automation (ICRA), pp. 7559–7566 (2018)
Newell, J.A., Puzianowski, A.A.: Development of a pneumatic spreading system for kevlar-based sic-precursor carbon fibre tows. High Perform. Polym. 11(2), 197–203 (1999)
Peng, X.B., Andrychowicz, M., Zaremba, W., Abbeel, P.: Sim-to-Real Transfer of Robotic Control with Dynamics Randomization. In: Proceedings - IEEE International Conference on Robotics and Automation, pp. 3803–3810 (October 2018)
Savitzky, A., Golay, M.J.E.: Smoothing and differentiation of data by simplified least squares procedures. Anal. Chem. 36(8), 1627–1639 (1964)
Such, F.P., Madhavan, V., Conti, E., Lehman, J., Stanley, K.O., Clune, J.: Deep Neuroevolution: Genetic Algorithms Are a Competitive Alternative for Training Deep Neural Networks for Reinforcement Learning (2017). http://arxiv.org/abs/1712.06567
Sutton, R.S.: Dyna, an integrated architecture for learning, planning, and reacting. ACM SIGART Bull. 2(4), 160–163 (1991)
Thomas, P.S., Theocharous, G., Ghavamzadeh, M.: High confidence off-policy evaluation. In: Proceedings of the National Conference on Artificial Intelligence 4, 3000–3006 (2015)
Acknowledgments
This research is partly funded by the Bavarian Ministry of Economic Affairs, Regional Development and Energy in the project LufPro.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this paper
Cite this paper
Krützmann, J., Schiendorfer, A., Beratz, S., Moosburger-Will, J., Reif, W., Horn, S. (2020). Learning Controllers for Adaptive Spreading of Carbon Fiber Tows. In: Nicosia, G., et al. Machine Learning, Optimization, and Data Science. LOD 2020. Lecture Notes in Computer Science(), vol 12566. Springer, Cham. https://doi.org/10.1007/978-3-030-64580-9_6
Download citation
DOI: https://doi.org/10.1007/978-3-030-64580-9_6
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-64579-3
Online ISBN: 978-3-030-64580-9
eBook Packages: Computer ScienceComputer Science (R0)