Print Page   |   Sign In   |   Join
Online Store: ET '08 Individual Papers
Main Storefront
        

Use of Adaptive Neural Networks (ANN) in Aluminum Extrusion Process Control

Item Options
Sign in for your pricing!
Price: $20.00
Quantity: *
 
Description

Use of Adaptive Neural Networks (ANN) in Aluminum Extrusion Process Control

Rodrigo Camargo Campana, Companhia Brasileira de Alumínio (CBA) and University of São Paulo, Brazil
Ronald Lesley Plaut, University of São Paulo, Brazil

Track: EXTRUSION/DIE, THEORETICAL - Neural Networks

ABSTRACT --- Adaptive Neural Networks (ANN) can be employed in the
analysis of complex input/output data under industrial conditions, even when
they may present substantial noise. Despite presenting conceptually substantial
mathematical complexity associated with non-linear parametric models that
include transfer equations and respective “training procedures “, ANN are largely
employed under industrial conditions in different engineering areas such as steel
processing, with great success. This work aims at the presentation of the
applicability of ANN in the specific case of the AA6063 and AA6351 extruded
alloys, under different industrial conditions, trying to exemplify its applicability
in process control. Results will be analyzed in terms of the adhesion to the ANN
of the collected industrial data as well as the relevance of each input within the
ANN.

© Extrusion Technology for Aluminum Profiles Foundation (ET Foundation). All rights reserved. No part of The Proceedings may be reproduced in any form without the express written permission of the ET Foundation.

Association Management Software Powered by YourMembership  ::  Legal