Using an artificial neural network to assist roll design in cold roll-forming processes

Cold roll-forming is a highly efficient and economical method for the production of sheet metal products, compared to rival processes. It has a major difficulty, however, because it frequently experiences quality problems such as longitudinal curvature. The tool design method is largely subjective and often involves a trial and error procedure where rolls are tested in the workshop. This paper develops a technique using an artificial neural network to assist in the design of roll-forming tools.