PREDICTION OF MECHANICAL PROPERTIES OF SOME IRON-BASED POWDER METALLURGY MATERIALS USING ARTIFICIAL NEURAL NETWORK

References: 17th International Multidisciplinary Scientific GeoConference SGEM 2017, www.sgem.org, SGEM2017 Conference Proceedings, ISBN 978-619-7408-12-6 / ISSN 1314-2704, 29 June - 5 July, 2017, Vol. 17, Issue 61, 287-292 pp, DOI: 10.5593/sgem2017/61/S24.038

ABSTRACT

A learning algorithm was used as an artificial neural network (ANN) tool to predict the mechanical properties of some iron-based powder metallurgy (P/M) materials. Specimens prepared from atomized iron powder and from pre-alloyed iron base powders powders were analyzed in this paper. The specimens were fabricated by press and sintering of the mixed powders. The samples were compressed in a universal mechanical testing machine to a pressure of 400 and 600 MPa. The sintering temperature was approximately 1.150 °C. The conventional compaction and sintering were used to obtain the maximum density of 7.01 g/cm 3 in the sintered state. The influence of green porosity, the particle sizes and sintering time on properties of the specimens analyzed were investigated. An increase in the sintered density of the samples was correlated with a lower porosity and smaller pore size. Porosity, particle size and sintering time were defined as the input variables of the model. Hardness was used as the outputs variables of the model. It is observed that the artificial neural network used confirmed the viability of the model and is in good correlation with the experimental values. ANN as an alternative can be used to predict the mechanical properties of some iron-based materials obtained by powder metallurgy route.

Keywords: powder metallurgy, artificial neural network, porosity, density