A method for predicting crown of hot-rolled strip steel plate based on deep learning

Abstract

The invention discloses a method for predicting the convexity of hot-rolled strip steel plates based on deep learning, which includes the following steps: S1, collecting and recording strip steel production data, and then preprocessing the collected data, including deleting missing values ​​and abnormal values Deletion, normalization; S2, according to the strip steel production data, use the attribute selection method based on Morisita index to remove redundant and irrelevant attributes in the collected data, and screen out the minimum number of attributes that can represent the variation of the plate crown The attribute constitutes the input variable set of the prediction model; S3, based on the input variable set, establishes the prediction model of the exit plate convexity of the strip steel based on the deep and wide convolutional neural network, so as to obtain the convexity of the exit plate of the hot-rolled strip steel. The invention uses the convolutional layer in the convolutional neural network to extract high-order features and invariant features of data, learns local correlations between variables, and combines the global feature learning ability of the deep neural network to significantly improve the prediction accuracy of the convexity of the plate.

Description

technical field [0001] The invention relates to a method for predicting the convexity of a hot-rolled strip steel plate based on deep learning, and belongs to the field of metallurgy. Background technique [0002] The steel industry is a pillar industry of national production and social development. Among steel products, strip steel is known as general steel, and is an important basic material for high-end products. It is widely used in construction, transportation, national defense and other fields. With the continuous improvement of the manufacturing level, the demand for high-precision plates and strips has increased sharply, and the requirements for the dimensional accuracy of plates and strips have become more stringent. As an intermediate step of strip rolling, hot rolling has a significant impact on the product quality of downstream processes such as cold rolling. Therefore, it is of great significance to achieve precise control of hot-rolled strip shape. The shape o...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a method for predicting the crown of hot-rolled strip steel plate based on deep learning, which can effectively solve the problems in the prior art, especially using traditional machine learning and deep learning models for prediction, the model The problem of low forecasting accuracy and generalization ability

Images