The use of an artificial neural network to estimate tool costs in cold roll-forming processes

The use of an artificial neural network to estimate tool costs in cold roll-forming processes

Anthony Downes, Peter Harley

Department of Mechanical Engineering, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK.

DOI:

https://doi.org/10.7494/cmms.2006.3.0173

Abstract:

The cold roll-forming industry is extremely competitive and the majority of work tenders that are submitted are unsuccessful. There are several issues that influence the tool costs, but the central problem lies in predicting the number of rolls that are required to roll-form the section and therefore determine the forming machine size. This paper discusses a method that assists tool cost estimations in cold roll-forming processes. The objective is to reduce the cost of generating work tenders whilst ensuring that the accuracy of the cost estimation is maintained, or improved. To facilitate this approach a LISP program was developed to process AutoCAD drawings of section geometry and to evaluate selected section features such as the total number of bend regions. The section features were then processed by an artificial neural network that was trained to predict the size of the forming machine that would be required to roll-form the section.

Cite as:

Downes, A., Hartley, P. (2006). The use of an artificial neural network to estimate tool costs in cold roll-forming processes. Computer Methods in Materials Science, 6(3-4), 203 – 212. https://doi.org/10.7494/cmms.2006.3.0173

Article (PDF):

Keywords:

Cold roll forming, Artificial neural networks, LISP

References: