Toggle Main Menu Toggle Search

Open Access padlockePrints

Modelling the distortion of long product sections after hot rolling using finite elements and neural networks

Lookup NU author(s): Konstantinos Bezas, Tom Musicka, Dr Jie ZhangORCiD, Professor Elaine Martin


Full text for this publication is not currently held within this repository. Alternative links are provided below where available.


The paper highlights the use of linear and non-linear quadratic Design of Experiments (DoE) for predicting the most influential factors affecting distortion during cooling of steel long products. The outcome of this investigation indicated the need to use advanced non-linear techniques such as Artificial Neural Networks in order to build a global model for predicting distortion, using data generated from Finite Element predictions. The development of this type of hybrid model shows considerable saving in CPU time since a fully trained ANN once learning is complete can make prediction in a fraction of the time required to get a single FE prediction of distortion. This has the potential to produce significant cost benefits and render the application feasible for implementation in advisory control systems.

Publication metadata

Author(s): Bezas K, Farrugia D, Richardson A, Musicka T, Zhang J, Martin EB

Editor(s): McNulty, G.J.

Publication type: Conference Proceedings (inc. Abstract)

Publication status: Published

Conference Name: 5th International Conference on Quality, Reliability, and Maintenance (QRM 2004)

Year of Conference: 2004

Pages: 201-204

Publisher: Professional Engineering Publishing

Library holdings: Search Newcastle University Library for this item

ISBN: 9781860584404