Development of machine learning method of titanium alloy properties identification in additive technologies

Authors

DOI:

https://doi.org/10.15587/1729-4061.2018.134319

Keywords:

titanium alloy powders, microstructure, morphology, granulometric composition, additive technologies, artificial intelligence methods

Abstract

Based on the experimentally established data on the parameters of microstructure, elemental and fractional composition of titanium alloy powders, four classes of their conformity (a material with excellent properties, optimal properties, possible defects in the material and defective material) as source raw materials for the additive technologies are identified. The basic characteristics of the material, which determine its belonging to a certain class, are established. Training and test samples based on 20 features that characterize each of the four classes of titanium alloy powders for the implementation of machine learning procedures were built. The developed method for identification of the class of material, based on the use of the second-order Kolmogorov-Gabor polynomial and the Random Forest algorithm, is described. An experimental comparison of the developed method work results with existing methods: Random Forest, Logistic Regression, and Support Vectors Machines based on the accuracy of their work in the training and application modes was made. The visualization of the results of all the investigated methods was given.

The developed supervised learning method allows constructing models for processing a large number of each input vector characteristics. In this case, the Random Forest algorithm provides satisfactory generalization properties while retaining the advantages of an additional increase of the accuracy based on the Kolmogorov-Gabor polynomial.

The main advantages of the developed method, in particular, regarding the additional increase of the accuracy of the classification task solution, are experimentally determined. The developed method allows increasing the modeling accuracy by 34.38, 33.34 and 3.13 % compared with the methods: Support Vectors Machine, Logistic Regression, and Random Forest respectively.

The obtained results allow one to considerably reduce financial and time expenses during the manufacture of products by additive technologies methods. The use of artificial intelligence tools can reduce the complexity and energy efficiency of experiments to determine the optimum characteristics of powder materials.

Author Biographies

Roman Tkachenko, Lviv Polytechnic National University S. Bandery str., 12, Lviv, Ukraine, 79013

Doctor of Technical Sciences, Professor, Head of Department

Department of Publishing Information Technologies

Zoia Duriagina, Lviv Polytechnic National University S. Bandery str., 12, Lviv, Ukraine, 79013 Katolicki Uniwersytet Lubelski Jana Pawła II Al. Racławickie, 14, Lublin, Poland, 20-950

Doctor of Technical Sciences, Professor

Department of Applied Materials Science and Materials Engineering

Ihor Lemishka, Lviv Polytechnic National University S. Bandery str., 12, Lviv, Ukraine, 79013

Postgraduate student

Department of Applied Material Science and Materials Engineering

Ivan Izonin, Lviv Polytechnic National University S. Bandery str., 12, Lviv, Ukraine, 79013

PhD, Assistant

Department of Publishing Information Technologies

Andriy Trostianchyn, Lviv Polytechnic National University S. Bandery str., 12, Lviv, Ukraine, 79013

PhD, Assistant

Department of Applied Materials Science and Materials Engineering

References

  1. Duriagina, Z. A., Kovbasyuk, T. M., Bespalov, S. A. (2016). The Analysis of Competitive Methods of Improvement of Operational Properties of Functional Layers of Flat Heating Elements. Uspehi Fiziki Metallov, 17 (1), 29–51. doi: 10.15407/ufm.17.01.029
  2. Pidkova, V. (2015). Structure and properties of Mg, Al, Ti oxide and nitride layers formed by ion-plasma sputtering. Functional Materials, 22 (1), 34–39. doi: 10.15407/fm22.01.034
  3. Duryagina, Z. A., Bespalov, S. A., Borysyuk, A. K. Pidkova, V. Ya. (2011). Magnetometric analysis of surface layers of 12Х18Н10Т steel after ion-beam nitriding. Metallofizika i Noveishie Tekhnologii, 33 (5), 615–622.
  4. Ostash, O. P., Andreiko, I. M., Kulyk, V. V., Babachenko, O. I., Vira, V. V. (2009). Influence of the mode of thermal treatment and load ratio on the cyclic crack-growth resistance of wheel steels. Materials Science, 45 (2), 211–219. doi: 10.1007/s11003-009-9177-4
  5. Ostash, O. P., Andreiko, I. M., Kulyk, V. V., Vavrukh, V. I. (2013). Influence of braking on the microstructure and mechanical behavior of railroad wheel steels. Materials Science, 48 (5), 569–574. doi: 10.1007/s11003-013-9539-9
  6. Ostash, O. P., Anofriev, V. H., Andreiko, I. M., Muradyan, L. A., Kulyk, V. V. (2013). On the concept of selection of steels for high-strength railroad wheels. Materials Science, 48 (6), 697–703. doi: 10.1007/s11003-013-9557-7
  7. Ostash, O. P., Kulyk, V. V., Poznyakov, V. D., Haivorons’kyi, O. A., Markashova, L. I., Vira, V. V. et. al. (2017). Fatigue crack growth resistance of welded joints simulating the weld-repaired railway wheels metal. Archives of Materials Science and Engineering, 2 (86), 49–52. doi: 10.5604/01.3001.0010.4885
  8. Mueller, T., Kusne, A. G., Ramprasad, R. (2016). Machine Learning in Materials Science. Reviews in Computational Chemistry, 186–273. doi: 10.1002/9781119148739.ch4
  9. Tsmots, I., Teslyuk, V., Teslyuk, T., Ihnatyev, I. (2017). Basic Components of Neuronetworks with Parallel Vertical Group Data Real-Time Processing. Advances in Intelligent Systems and Computing, 558–576. doi: 10.1007/978-3-319-70581-1_39
  10. Duriagina, Z. A., Tkachenko, R. O., Trostianchyn, A. M., Lemishka, I. A., Kovalchuk, A. M., Kulyk, V. V., Kovbasyuk, T. M. (2018). Determination of the best microstructure and titanium alloy powders properties using neural network. Journal of Achievements in Materials and Manufacturing Engineering, 87 (1), 25–31. doi: 10.5604/01.3001.0012.0736
  11. Tkachenko, R., Doroshenko, A., Izonin, I., Tsymbal, Y., Havrysh, B. (2018). Imbalance Data Classification via Neural-Like Structures of Geometric Transformations Model: Local and Global Approaches. Advances in Computer Science for Engineering and Education, 112–122. doi: 10.1007/978-3-319-91008-6_12
  12. Narushynska, O., Teslyuk, V., Vovchuk, B.-D. (2017). Search model of customer's optimal route in the store based on algorithm of machine learning A. 2017 12th International Scientific and Technical Conference on Computer Sciences and Information Technologies (CSIT). doi: 10.1109/stc-csit.2017.8098787
  13. Shakhovska, N. B., Noha, R. Y. (2015). Methods and Tools for Text Analysis of Publications to Study the Functioning of Scientific Schools. Journal of Automation and Information Sciences, 47 (12), 29–43. doi: 10.1615/jautomatinfscien.v47.i12.30
  14. Rajan, K. (2005). Materials informatics. Materials Today, 8 (10), 38–45. doi: 10.1016/s1369-7021(05)71123-8
  15. Xu, J., Zhu, L., Fang, D., Liu, L., Xu, W., Li, Z. (2012). Prediction of glass transition temperatures for polystyrenes from cyclic dimer structures using artificial neural networks. Fibers and Polymers, 13 (3), 352–357. doi: 10.1007/s12221-012-0352-0
  16. Raccuglia, P., Elbert, K. C., Adler, P. D. F., Falk, C., Wenny, M. B., Mollo, A. et. al. (2016). Machine-learning-assisted materials discovery using failed experiments. Nature, 533 (7601), 73–76. doi: 10.1038/nature17439
  17. Fernandez, M., Boyd, P. G., Daff, T. D., Aghaji, M. Z., Woo, T. K. (2014). Rapid and Accurate Machine Learning Recognition of High Performing Metal Organic Frameworks for CO2 Capture. The Journal of Physical Chemistry Letters, 5 (17), 3056–3060. doi: 10.1021/jz501331m
  18. Zlenko, M., Popovich, A., Mutylina, I. (2013). Additivnye tekhnologii v mashinostroenii. Sankt-Peterburg, 221.
  19. Haznaferov, M. V., Ovchinnikov, A. V., Yanko, T. B. (2015). Tekhnologiya polucheniya «low-cost» poroshkov legirovannogo titana dlya additivnyh processov. Titan, 2, 31–36.
  20. Duriahina, Z. A., Trostianchyn, A. M., Lemishka, I. A., Skrebtsov, A. A., Ovchynnykov, O. V. (2017). Hranulometrychni kharakterystyky poroshku tytanovoho splavu VT20 otrymanoho metodom vidtsentrovoho plazmovoho rozpylennia elektrodu. Metaloznavstvo ta obrobka metaliv, 1, 45–51.
  21. Duriagina, Z., Trostyanchyn, A., Lemishka, I., Skrebtsov, A., Ovchinnikov, O. (2017). The influence of chemical-thermal treatment on granulometric characteristics of titanium sponge powder. Ukrainian Journal of Mechanical Engineering and Materials Science, 3 (1), 73–80.
  22. Ovchinnikov, A. V., Ol'shaneckiy, V. E., Dzhugan, A. A. (2015). Primenenie nesfericheskih gidrirovanyh i degidrirovanyh poroshkov titana dlya polucheniya izdeliy v additivnyh tekhnologiyah. Vestnik dvigatelestroeniya, 1, 114–117.
  23. Petrik, I. A., Ovchinnikov, A. V., Seliverstov, A. G. (2015). Razrabotka poroshkov titanovyh splavov dlya additivnyh tekhnologiy primenitel'no k detalyam GTD. Aviacionno-kosmicheskaya tekhnika i tekhnologiya, 8, 11–16.
  24. Broeke, J., Perez, J. M. M., Pascau, J. (2015). Image Processing with ImageJ. Packt Publishing, 256.
  25. Gavrilova, N. N., Nazarov, V. V., Yarovaya, O. V. (2012). Mikroskopicheskie metody opredeleniya razmerov chastic dispersnyh materialov. Moscow, 52.
  26. Shanyavskiy, A. A., Banov, M. D., Zakharova, T. P. (2010). Principles of physical mesomechanics of nanostructural fatigue of metals. I. Model of subsurface fatigue crack nucleation in VT3-1 titanium alloy. Physical Mesomechanics, 13 (3-4), 133–142. doi: 10.1016/j.physme.2010.07.004
  27. Model' Random Forest dlya klassifikacii, realizaciya na c#. Available at: https://habr.com/post/215453
  28. Random Forest: progulki po zimnemu lesu. Available at: https://habr.com/post/320726
  29. Oleg, R., Yurii, K., Oleksandr, P., Bohdan, B. (2017). Information technologies of optimization of structures of the systems are on the basis of combinatorics methods. 2017 12th International Scientific and Technical Conference on Computer Sciences and Information Technologies (CSIT). doi: 10.1109/stc-csit.2017.8098776
  30. Jordan, M. I., Jacobs, R. A. (1994). Hierarchical Mixtures of Experts and the EM Algorithm. Neural Computation, 6 (2), 181–214. doi: 10.1162/neco.1994.6.2.181
  31. Dronyuk, I., Fedevych, O., Poplavska, Z. (2017). The generalized shift operator and non-harmonic signal analysis. 2017 14th International Conference The Experience of Designing and Application of CAD Systems in Microelectronics (CADSM). doi: 10.1109/cadsm.2017.7916092
  32. Qian, M., Froes, F. H. (Eds.) (2015). Titanium Powder Metallurgy: Science, Technology and Applications. Butterworth-Heinemann, 628. doi: 10.1016/c2013-0-13619-7
  33. Leyens, C., Peters, M. (Eds.) (2003). Titanium and Titanium Alloys: Fundamentals and Applications. Wiley‐VCH Verlag GmbH & Co. KGaA, 532. doi: 10.1002/3527602119
  34. Data Mining Fruitful and Fun. Available at: https://orange.biolab.si
  35. The Python Standard Library. Available at: https://docs.python.org/3/library/index.html

Downloads

Published

2018-06-20

How to Cite

Tkachenko, R., Duriagina, Z., Lemishka, I., Izonin, I., & Trostianchyn, A. (2018). Development of machine learning method of titanium alloy properties identification in additive technologies. Eastern-European Journal of Enterprise Technologies, 3(12 (93), 23–31. https://doi.org/10.15587/1729-4061.2018.134319

Issue

Section

Materials Science