Optimization of 3D Printing Modes for Electric Arc Surfacing Using a Digital Twin of the Process

Abstract:

The article proposes a method for choosing the optimal modes of 3D printing by electric arc surfacing on CNC machines using a digital twin of the process. As a digital twin, a neural network model is used that approximates the dependence of the stability of the surfacing process and the geometric parameters of the printed layer on the surfacing modes: voltage, current and minute feed. The possibility of optimizing 3D printing modes using neural network modeling based on a modified backpropagation method is shown.

• Periodical: Materials Science Forum (Volume 1052)
• Pages: 414-419
• DOI: https://doi.org/10.4028/p-v4165r
• Online since: February 2022
• Authors: Dmitrii A. Shatagin*, Maksim S. Anosov, Andrei Galkin, Natalia Klochkova
• Keywords: 3D Printing, Cyber-Physical System, Digital Twin, Hybrid System, Machine Learning, Neural Network, Optimization, WAAM

References

[1] N. Lv, J. Zhong, H. Chen, S. Chen, Real-time control of welding penetration during robotic GTAW dynamical process by audio sensing of arc length, The International Journal of Advanced Manufacturing Technology. 74(1-4) (2014) 235-249.

DOI: 10.1007/s00170-014-5875-7

[2] C. Chen, R. Xiao, H. Chen, N. Lv, Arc sound model for pulsed GTAW and recognition of different penetration states, The International Journal of Advanced Manufacturing Technology. 108 (2020) 3175-3191.

DOI: 10.1007/s00170-020-05462-z

[3] D. Shatagin, A. Galkin, A.N. Osmehin, N. Klochkova, Cyber-physical hybrid processing system digital twin, Materials Science Forum. 1037 (2021) 119-124.

DOI: 10.4028/www.scientific.net/msf.1037.119

[4] Yu.G. Kabaldin, D.A. Shatagin, M.S. Anosov, P.V. Kolchin, A.V. Kiselev, Diagnostics of the 30-print process on a CNC machine using machine learning approaches, Vestnik Mashinostroyeniya, 1 (2021) 55-59.

DOI: 10.36652/0042-4633-2021-1-55-59

[5] I.V. Pentegov, V.N. Sidorets, Comparative analysis of models of dynamic welding arc, Automatic Welding. 2 (1989) 33-36.

[6] E.N. Vereshchago, V.I. Kostyuchenko, Model of the electric arc in Matlab / Simulink, Electrical Engineering and Power Engineering. 2 (2013) 40-46.

[7] A. Savitsky, M. Haltof, Problems of determining the parameters of mathematical models of electric arcs in circuits with current sources, Electricity. 1 (2016) 25-34.

[8] Information on: https://github.com/ranalytics/data-mining.

[9] Information on: https://ru.wikipedia.org/wiki/Random_forest.

[10] J. Gonzalez, W. Yu, Non-linear system modeling using LSTM neural networks, IFAC-PapersOnLine. 51(13) (2018) 485-489.

DOI: 10.1016/j.ifacol.2018.07.326

[11] Information on: https://en.wikipedia.org/wiki/Long_short-term_memory.

[12] D. Eck, J. Schmidhuber, Learning the long-term structure of the blues, in: 12th International Conference on Artifical Neural Networks (ICANN 2002), Lecture Notes in Computer Science, 2415, Springer, Berlin, Heidelberg, pp.284-289.

DOI: 10.1007/3-540-46084-5_47

[13] J. Schmidhuber, F. Gers, D. Eck, Learning nonregular languages: A comparison of simple recurrent networks and LSTM, Neural Computation. 14(9) (2002) 2039-2041.

DOI: 10.1162/089976602320263980

[14] A. Graves, J. Schmidhuber, Framewise phoneme classification with bidirectional LSTM and other neural network architectures, Neural Networks. 18(5-6) (2005) 602-610.

DOI: 10.1016/j.neunet.2005.06.042

[15] P. Malhotra, L. Vig, G. Shroff, P. Agarwal, Long short term memory networks for anomaly detection in time series, 23rd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN, Bruges, Belgium, (2015).