Prediction of rolling force in cold rolling by using physical models and neural computing

Abstract In this study physical models and a neural network theory have been integrated to a program package in order to predict rolling force in cold rolling. The parameters required by the model such as the friction parameter and the deformation resistance of the materials have been determined from measured rolling parameters and materials alloying elements by applying the Bland-Ford-Ellis (BFE) rolling force model and an artificial neural network model (ANN). Measured data of over 6000 coils have been used in the training of the ANN. The calculated results were in good agreement with measurements.