Scholarly record

MODELING OF DIMENSIONAL CHANGES IN SOME IRON-BASED POWDER METALLURGY COMPACTS USING AN ARTIFICIAL NEURAL NETWORK

Florin Bogdan Marin

Abstract

Modeling using artificial neural networks (ANNs) is used in most fields of production. An approach to modeling the dimensional changes in some iron-based powder metallurgy (PM) parts during the sintering process is given. The proposed model used is a multilayer neural network with a backpropagation learning algorithm. This type of neural network model gave the best results in the process of modeling. The specimens used in this study are prealloyed iron-based powders. The particle size of the powders is ranging from 45 to 150 ?m. In the first step, the powders were mixed with zinc stearate 1% for 30 minutes, then were single pressed in a die at a pressure of 600 MPa. Following pressing, the green compacts were subject to sintering in a laboratory furnace at 1150? C for 90 and 120 minutes. During the sintering process, there are a great number of the process factors that appear, with an influence on the dimensional changes and also on the final accuracy of a sample (sintering temperature, sintering time, type of the protection atmosphere). It was concluded that the practical effects of the modeling using an ANN in determination of compact dimensions is in accordance with the data obatined by measurements of the dimensional changes during sintering.

Publication Impact Profile

• Captures
• Mendeley - Readers: 3

PLUMX see details

DOI resolver: doi.org/10.5593/sgem2019/2.1/s07.060

Publication details

Back to publication list