Cold-rolled sheet shape signal online mode recognition method

Abstract

The invention provides a cold-rolled sheet shape signal online mode recognition method. The method includes the steps of receiving sheet shape measurement values of measurement segments in the width direction of a cold-rolled strip steel measured by a sheet shape instrument in an online mode, comparing the measurement values with a set distribution value of a corresponding target sheet shape, obtaining sheet shape deviation values of all measurement segments, determining the largest value of the absolute values of the sheet shape deviation values, carrying out the normalization processing on the sheet shape deviation values of all measurement segments, carrying out the rough-filtering processing on deviation values, after the normalization processing, of the measurement segments, recognizing sheet shape modes through a radial basis function neural network based on the differential evolution intelligent optimization algorithm, and judging a sheet shape mode recognition result. Through the normalization processing and the rough-filtering processing of the sheet shape deviation values of all measurement segments, the adverse influence of bad point data of the sheet shape measurement values on sheet shape mode recognition can be obviously eliminated, and recognition accuracy of the sheet shape mode is improved. Molding accuracy and network training efficiency are obviously improved due to the fact that the differential evolution intelligent optimization algorithm is applied to the sheet shape mode recognition based on the radial basis function neural network.

Description

technical field [0001] The invention belongs to the field of cold-rolled steel strips, in particular to an online pattern recognition method for cold-rolled strip shape signals. Background technique [0002] In the cold-rolled strip production process, cold-rolled flatness pattern recognition is an important part of the cold-rolled flatness control system. Shape pattern recognition is to identify and classify the measured shape signal output by the shape meter, so as to accurately judge the shape defect type of the strip, provide control basis for the control system, and finally produce high-quality cold-rolled strip material products. [0003] The traditional plate shape signal pattern recognition method is based on the polynomial decomposition method of the least square method and the improved orthogonal polynomial regression decomposition method. These methods have poor anti-interference ability and have theoretical defects. High-precision shape control requirements. T...

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to provide an online pattern recognition method for cold-rolled flatness signals, which can effectively solve the problems of unsatisfactory accuracy and real-time performance, complex identification model structure and network training problems that are often encountered when using traditional flatness pattern recognition methods. The technical problems of too long time, poor stability and robustness can provide a reliable control basis for the control system, and provide a strong guarantee for improving the quality of cold-rolled strip shape control

Images