PSI - Issue 23

Available online at www.sciencedirect.com Available online at www.sciencedirect.com

ScienceDirect ScienceDirect

Available online at www.sciencedirect.com Structural Integrity Procedia 00 (2019) 000 – 000 Structural Integrity Procedia 00 (2019) 000 – 000

www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia

ScienceDirect

Procedia Structural Integrity 23 (2019) 221–226

© 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the scientific committee of the ICMSMF organizers © 201 9 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the scientific committee of the IC MSMF organizers. In order t stu y he hot deformation behavior of the homoge ized aluminium alloy, the series of uniaxial compression tests has been performed at the temperature range o 723 K – 823 K and the strain rate range of 0.5 s − 1 – 10 s − 1 by means of the Gleeble 3800 simulator. Resulting flow str ss data for the tru train up to 0.6 wer approximat d via Cingara & McQueen's flow stress model and its modification. Param ters of these models (i.e. peak strain, peak tress, steady-stat str ss, hardeni g nd dynamic softening exponents) were described by two differe t intellige t methodologies. In the first case, the description has be n done by utilizing an rtifi i l e ral et r approach. In the second case, genetic algorithms w re used. Comp rison of the results shows tha the artificial neural network approach brings a better approximation fit with the experimental dataset to the complex hot flow stress modelling. © 201 9 The Authors. Published by Elsevier B.V. This is an open acces article under CC BY-NC-ND lic nse (http://creativecommon org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the scientific committee of the IC MSMF organizers. In order to study the hot deformation behavior of the homogenized aluminium alloy, the series of uniaxial compression tests has been performed at the temperature range of 723 K – 823 K and the strain rate range of 0.5 s − 1 – 10 s − 1 by means of the Gleeble 3800 simulator. Resulting flow stress data for the true strain up to 0.6 were approximated via Cingara & McQueen's flow stress model and its modification. Parameters of these models (i.e. peak strain, peak stress, steady-state stress, hardening and dynamic softening exponents) were described by two different intelligent methodologies. In the first case, the description has been done by utilizing an artificial neural network approach. In the second case, genetic algorithms were used. Comparison of the results shows that the artificial neural network approach brings a better approximation fit with the experimental dataset to the complex hot flow stress modelling. 9th International Conference on Materials Structure and Micromechanics of Fracture Modelling the Hot Deformation Behavior of AlSi1MgMn Alloy via 9th International Conference on Materials Structure and Micromechanics of Fracture Modelling the Hot Deformation Behavior of AlSi1MgMn Alloy via Flow Stress Models Utilizing Intelligent Algorithms Petr Opěla a *, Ivo Schindler a , Vladivoj Očenášek b , Petr Kawulok a , Rostislav Kawulok a , Stanislav Rusz a a VSB – Technical University of Ostrava, Faculty of Materials Science and Technology, 17. listopadu 2172/15, 708 00 Ostrava – Poruba, Czech Republic, EU b SVÚM – Non- Ferrous Metal Department, Tovární 2053, 250 88 Čelákovice, Czech Republic, EU Flow Stress Models Utilizing Intelligent Algorithms Petr Opěla a *, Ivo Schindler a , Vladivoj Očenášek b , Petr Kawulok a , Rostislav Kawulok a , Stanislav Rusz a a VSB – Technical University of Ostrava, Faculty of Materials Science and Technology, 17. listopadu 2172/15, 708 00 Ost ava – Poruba, Czech Republic, EU b SVÚM – Non- Ferrous Metal Department, Tovární 2053, 250 88 Čelákovice, Czech Republic, EU Abstract Abstract

Keywords: hot-flow-curve approximation; artificial neural network; genetic algorithm Keywords: hot-flow-curve approximation; artificial neural network; genetic algorithm

* Corresponding author. Tel.: +420-59732-4349. E-mail address: petr.opela@vsb.cz * Correspon ing author. Tel.: +420-59732-4349. E-mail address: petr.opela@vsb.cz

2452-3216 © 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the scientific committee of the IC MSMF organizers. 2452-3216 © 2019 The Authors. Published by Elsevier B.V. This is an ope acces article under CC BY-NC-ND lic nse (http://creativecommon org/licenses/by-nc-nd/4.0/)

Peer-review under responsibility of the scientific committee of the IC MSMF organizers.

2452-3216 © 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the scientific committee of the ICMSMF organizers 10.1016/j.prostr.2020.01.090