Neural network model of the profile of hot-rolled strip

The dynamics of shape or profile depends on the several process parameters in a hot strip mill. This research paper presents the prediction of the profile or strip crown generated during rolling in the finishing stands of a hot strip mill with the use of artificial neural network (ANN) by taking into account the operational parameters such as width, thickness, speed, and roll bending, among others. A multi-layer perceptron neural network is trained with a back-propagation algorithm. Verification of the model is made by a comparison of the measured and predicted strip profile. A sensitivity analysis of the contributing factors to the profile variations is also carried out that shows the strip speed as one of the major contributor to the strip profile. A good comparison is found between the measured strip crown with the ANN-predicted crown demonstrating the modeling route as a robust one.