Optimization of neural networks structure selection in modelling spheroidal graphite cast iron for automotive camshafts

Optymalizacja doboru struktury sztucznych sieci neuronowych w modelowaniu zużycia żeliwa sferoidalnego na samochodowe wałki rozrządu

  • Kazimierz Witaszek Silesian University of Technology, Faculty of Transport and Aviation Engineering, Department of Automotive Vehicle Maintenance
  • Krzysztof Garbala AC Spólka Akcyjna
  • Mirosław Witaszek Silesian University of Technology, Faculty of Transport and Aviation Engineering, Department of Automotive Vehicle Maintenance
  • Marcin Rychter Vocational State School of Ignacy Mościcki in Ciechanów, Faulty of Engineering and Economics
Keywords: Artificial neural networks, structure optimization, wear, spheroidal cast iron, Stuttgart Neural Network Simulator, Resilient backPROPagation

Abstract

The present article discusses the process of optimizing the structure of artificial neural networks applied in modelling the wear of spheroidal graphite cast iron (SG cast iron). The networks were trained using the RPROP gradient method with the application of the SNNS package supported by original self-developed software, which enabled automatic creation, training and testing of networks with different sizes of hidden layers. Based on the results of an analysis of learning process and testing a package of 625 networks, the network was selected which – when modelling the process of spheroidal cast iron wear – generates the slightest errors during testing.

References

1. Xu, J. & Yamada, K. & Seikiya, K. & Tanaka, R. & Yamane, Y.: Effect of different features to drill-wear prediction with back propagation neural network. Precision Engineering.2014. Vol. 38, Iss. 4. P. 791-798.
2. Kadua, R.S. & Awarib, G.K. & Sakhalec, C.N. & Modakd, J.P.: Formulation of Mathematical Model for the Investigation of Tool Wears in Boring Machining Operation on Cast Iron Using Carbide and CBN Tools. Procedia Materials Science. 2014. Vol. 6. P. 1710-1724.
3. Veeresh Kumara, G.B. & Pramoda, R. & Raob, C.S.P. & Shiva-kumar Goudac, P.S.: Artificial Neural Network Prediction On Wear Of Al6061 Alloy Metal Matrix Composites Reinforced With -Al2O3.  Materials Today: Proceedings. 2018. Vol. 5, Iss. 5. Part 2. P. 11268–11276.
4. Cios, K.J. & Mamitsuka, H. & Nagashima, T. & Tadeusiewicz, R.: Computational intelligence in solving bioinformatics problems. Artificial Intelligence in Medicine. 2005. Vol. 35, Iss. 1-2. P. 1-8.
5. Tadeusiewicz, R. & Ogiela, L. & Ogiela, M.R.: The automatic understanding approach to systems analysis and design. International Journal of Information Management. 2008. Vol. 28, Iss. 1. P. 38-48.
6. Tadeusiewicz, R.: Neural networks in mining sciences-general overview and some representative examples. Archives of Mining Sciences. 2015. Vol. 60, Iss. 4. P. 971-984.
7. Manieniyan, V. & Vinodhini, G. & Senthilkumar, R. & Si-vaprakasam, S.: Wear element analysis using neural networks of a DI diesel engine using biodiesel with exhaust gas recirculation. Energy. 2016. Vol. 114. P. 603-612.
8. Szaleniec, J. & Wiatr, M. & Szaleniec, M. & Składzień, J. & Tomik, J. & Oleś, K. & Tadeusiewicz, R.: Artificial neural network model-ling of the results of tympanoplasty in chronic suppurative otitis media patients. Computers in Biology and Medicine. 2013. Vol. 43, Iss. 1. P. 16-22.
9. Szaleniec, M. & Tadeusiewicz, R. & Witko, M.: How to select an optimal neural model of chemical reactivity? Neurocomputing. 2008. Vol. 72, Iss. 1-3. P. 241-256.
10. Aleksendrić, D.: Neural network prediction of brake friction mate-rials wear. Wear. 2010. Vol. 268, Iss. 1-2. P. 117-125.
11. Bergmeir, C. & Benítez, J.M.: Neural Networks in R Using the Stuttgart Neural Network Simulator: RSNNS. Journal of Statistical Software. 2012. Vol. 47, Iss. 7. P. 1-26.
12. Olesiuk, D. & Bachmann, M. & Habermeyer, M. & Heldens, W. & Zagajewski, B.: SNNS application for crop classification using hymap data. Proceedings of the Whispers - 2nd Workshop on hyperspectral image and signal processing. 2010. P. 1-4.
13. Riedmiller, M. & Braun, H.: A direct adaptive method for faster backpropagation learning the RPROP algorithm. IEEE International Conference on Neural Networks. 1993. Vol. 1. P. 586-591.
14. Mrówczyńska, M.: Estymacja błędów modelu powierzchni opisa-nych funkcjami kształtu za pomocą sieci neuronowych. Acta Scientinarum Polonorum, Geodesia et Descriptio Terrarum. 2007. Vol. 6, Iss. 1. P. 15-23.
15. Zell, A. & Mamier, G. & Vogt, M. & Mache, N. & Hübner, R. & Döring, S. & Herrmann, K.U. & Soyez, T. & Schmalzl, M. & Som-mer, T. & Hatzigeorgiou, A. & Posselt, D. & Schreiner, T. & Kett, B. & Clemente, G. & Wieland, J.: Stuttgart Neural Network Simu-lator User Manual, Version 4.2. http://www.ra.cs.uni-tuebingen.de/downloads/SNNS/SNNSv4.2.Manual.pdf
16. Yilmazkaya, E. & Dagdelenler, G. & Ozcelik, Y. & Sonmez, H.: Prediction of mono-wire cutting machine performance parameters using artificial neural network and regression models. Engineering Geology. 2018. Vol. 239. P. 96-108.
17. Shebani, A. & Iwnicki, S.: Prediction of wheel and rail wear under different contact conditions using artificial neural networks. Wear. 2018. Vol. 406-407. P. 173-184.
Published
2020-03-31
Section
Eksploatacja i Testy/Exploitation and Tests