In:
Materials Science Forum, Trans Tech Publications, Ltd., Vol. 773-774 ( 2013-11), p. 268-274
Kurzfassung:
This paper presents an investigation of the capabilities of artificial neural networks (ANN) in predicting some mechanical properties of Ferrite-Martensite dual-phase steels applicable for different industries like auto-making. Using ANNs instead of different destructive and non-destructive tests to determine the material properties, reduces costs and reduces the need for special testing facilities. Networks were trained with use of a back propagation (BP) error algorithm. In order to provide data for training the ANNs, mechanical properties, inter-critical annealing temperature and information about the microstructures of many specimens were examined. After the ANNs were trained, the four parameters of yield stress, ultimate tensile stress, total elongation and the work hardening exponent were simulated. Finally a comparison of the predicted and experimental values indicates that the results obtained from the given input data reveal a good ability of the well-trained ANN to predict the described mechanical properties.
Materialart:
Online-Ressource
ISSN:
1662-9752
DOI:
10.4028/www.scientific.net/MSF.773-774
DOI:
10.4028/www.scientific.net/MSF.773-774.268
Sprache:
Unbekannt
Verlag:
Trans Tech Publications, Ltd.
Publikationsdatum:
2013
ZDB Id:
2047372-2
Permalink