Title: Application of neural network to the materials characterisation

Authors: Girolamo Costanza, Maria Elisa Tata, Nadia Ucciardello

Addresses: Department of Mechanical Engineering, University of Rome Tor Vergata, Via del Politecnico 1, 00133 Rome, Italy. ' Department of Mechanical Engineering, University of Rome Tor Vergata, Via del Politecnico 1, 00133 Rome, Italy. ' Department of Mechanical Engineering, University of Rome Tor Vergata, Via del Politecnico 1, 00133 Rome, Italy

Abstract: In this paper, the application of a new technique based on artificial neural network is used for the characterisation of mechanical properties. Three applications are shown in this article: the room temperature superplastic behaviour in PbSn60 alloy, the compressive behaviour of Al foam with different starting powder composition (different pores amount) and the fatigue behaviour in rotating bending tests of Al 6082T6 (as-received and after the fluidised bed treatment). Three neural networks with back-propagation (BP) algorithm have been implemented. After validation, the artificial neural networks are able to predict the output values with considerable efficiency and accuracy.

Keywords: artificial neural networks; ANNs; materials characterisation; mechanical properties; superplastic behaviour; compressive behaviour; fatigue behaviour; rotating bending tests.

DOI: 10.1504/IJCMSSE.2010.033147

International Journal of Computational Materials Science and Surface Engineering, 2010 Vol.3 No.2/3, pp.96 - 113

Published online: 10 May 2010 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article