Rolling model self-learning coefficient determination method

Abstract

The invention provides a rolling model self-learning coefficient determination method, and the method can effectively describe the relation between a rolling model self-learning coefficient and product specifications; the method comprises the following steps: finish rolling the k rolled piece, and storing the product specification and rolling model self-learning coefficient of the k rolled piece into a preset history record data set; determining a thickness grading position and a width grading position of the k rolled piece according to pre-determined thickness and width grading positions; obtaining rolled piece history record from the preset history record data set according to the determined k rolled piece thickness and width grading positions and belonging steel kinds, and carrying out multivariant nonlinear regression so as to obtain a regression equation; using the regression equation to calculate the rolling model self-learning coefficient of the grading position, weighting and averaging the calculated rolling model self-learning coefficient with a corresponding old value in a rolling model self-learning coefficient stratification table corresponding to the steel kind to which the k rolled piece belongs, and updating the old value. The rolling model self-learning coefficient determination method relates to the automation model field.

Description

technical field [0001] The invention relates to the field of automatic models, in particular to a method for determining self-learning coefficients of rolling models. Background technique [0002] In recent years, the automatic model self-learning of strip rolling process is very important to improve the prediction accuracy of the model and the stability of product quality. The model self-learning coefficient is used to correct the model prediction error in the actual production process. Generally speaking, different steel grades and product specifications have different model self-learning coefficients. How to deal with the relationship between the model self-learning coefficient and product specifications directly affects the model The effect of self-learning. At present, the most commonly used method for processing model self-learning coefficients is the model self-learning coefficient layer table. [0003] The model self-learning coefficient layer table stores the corr...

Images