A method for optimizing the band wide distribution at the ends of a continuous strip rolling mill train.

摘要

A method for optimizing the band wide distribution (y) at the ends of a rolling mill train (3) which continuous strip (4) by adjusting the adjustment position of rollers (6) during the passage of the band ends corresponding to a by means of predetermined parameters (s) described path of travel (f), and, on account of process parameters (x_{according to}) the rolling process by means of a neural network (140) to make a prediction concerning the band wide distribution (y) is made, and the error (δy) between the predicted band wide distribution (y) and the measured actual volume wide distribution (y_is) for adaptation of the neural network (140) in the sense of a reduction in the error (δy), and wherein with the adapted neural network (140) iteratively optimum values (s_opt) for the parameters (s) of the path of travel (f) are determined, for which the predicted band wide distribution (y) is at a minimum of a predetermined nominal - band wide distribution y_should) in accordance with which deviates from, and then the setting position of the rollers (6) is adjusted, characterized in that the prediction in addition, owing to the parameter (s_is) the travel curve (f_is) takes place (2, 3, 4).