一种面向跨时间尺度采样的工业过程运行指标建模方法
By using disk discretization and a modeling method for industrial process operation indicators with constraints across time scales, the problem of model input-output mismatch for multi-sampling period data is solved. By utilizing data without operation indicators, the modeling accuracy and data utilization rate are improved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA UNIV OF MINING & TECH
- Filing Date
- 2025-01-21
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies suffer from input-output mismatch when dealing with multi-period data sampling, and potential information in a large number of process data samples without operational indicators is ignored, resulting in low data utilization and reduced model accuracy.
An industrial process operation index modeling method oriented towards cross-time scale sampling is adopted. The maximum information coefficient time delay technique of disk discretization is used to identify and align the time delay. The model is constructed by combining cross-time scale constraint terms. The geometric structure information of process data samples without operation indexes is used to establish a soft measurement model.
It corrected the erroneous temporal correspondence between process data, solved the problem of model input-output mismatch, made full use of data without operational indicators, improved modeling accuracy and reduced the risk of overfitting.
Smart Images
Figure QLYQS_11 
Figure QLYQS_13 
Figure QLYQS_21