Scholarly record

ARTIFICIAL NEURAL NETWORK (ANN) APPLICATION TO PREDICT THE MECHANICAL PROPERTIES IN SOME IRON-BASED POWDER METALLURGY ALLOYS

Florin Bogdan Marin

Abstract

In the present study, the mechanical properties of some iron-based sintered alloys prepared by powder metallurgy route were predicted using Artificial Neural Network (ANN) approach. Artificial Neural Networks (ANNs) are important tools used to predict the complex processes with many variables and interactions. In order to prepare the sintered compacts, the powders prepared from pre-alloyed iron base powders produced by atomization (< 45, 45-63, 63-100, 100-150, >150 ?m) were the materials analyzed in this paper. The analyzed powders were consolidated in compacts in a mold using uniaxial pressing at 400 and 600 MPa with the disc dimensions of ? 8 ? 6 mm. Then, the consolidated compacts of prepared powders were sintered in a laboratory furnace at two different temperatures. The sintering temperature was approximately 1.150°C and two sintering times of 60 and 90 minutes. The microhardness measurements were performed to evaluate the mechanical properties of these specimens. Then, to predict the mechanical properties of the sintered specimens, a neural network as multi layer perceptron back propagation type was constructed. It was found that the proposed ANN model studied in this paper can be used as an alternative to predict the mechanical properties of some iron-based materials obtained by powder metallurgy route.

Publication Impact Profile

• Captures
• Mendeley - Readers: 6

PLUMX see details

DOI resolver: doi.org/10.5593/sgem2019/2.1/s07.015

Publication details

Back to publication list