Strip hot continuous rolling multi-sampling rate process monitoring method and device

By combining multi-scale convolutional neural networks and multi-directional kernel principal component analysis, the problem of monitoring uneven sampling rates and nonlinear variables in the hot strip rolling process was solved, achieving efficient and accurate process monitoring and improving the model's flexibility and computational efficiency.

CN118194122BActive Publication Date: 2026-07-21UNIV OF SCI & TECH BEIJING
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
UNIV OF SCI & TECH BEIJING
Filing Date
2024-03-20
Publication Date
2026-07-21

AI Technical Summary

Technical Problem

Existing technologies lack efficient and accurate process monitoring methods that directly utilize data with uneven sampling rates and nonlinear variables, making it difficult to effectively monitor complex hot strip rolling processes.

Method used

A method combining multi-scale convolutional neural networks (MCNN) and multi-directional kernel principal component analysis (MKPCA) is adopted. Through data preprocessing, feature extraction and modeling, an MCNN-MKPCA model is constructed, and fault detection is performed using multi-sampling rate data.

Benefits of technology

It improves the flexibility and robustness of monitoring, effectively handles nonlinear relationships, enhances the model's versatility and computational efficiency, and achieves efficient and accurate process monitoring.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118194122B_ABST
    Figure CN118194122B_ABST
Patent Text Reader

Abstract

The present application relates to the technical field of industrial process monitoring, in particular to a strip hot continuous rolling multi-sampling rate process monitoring method and device. The strip hot continuous rolling multi-sampling rate process monitoring method comprises: collecting multi-sampling rate historical data of a strip hot continuous rolling production line to obtain training data; constructing a model according to a multi-scale convolutional neural network structure and a multi-direction kernel principal component analysis model structure to obtain a to-be-trained MCNN-MKPCA model; training the to-be-trained MCNN-MKPCA model using the training data to obtain an MCNN-MKPCA model; calculating based on the training data according to the MCNN-MKPCA model to obtain a control limit; obtaining online data of the strip hot continuous rolling production line; inputting the online data into the MCNN-MKPCA model to calculate to obtain a statistic; and checking according to the statistic and the control limit to issue a fault alarm. The present application is a high-efficiency and accurate process monitoring method directly using non-uniform data and non-linear variables.
Need to check novelty before this filing date? Find Prior Art