A method and device for detecting a soft measurement model for float glass product quality
By constructing a soft measurement model based on multiple sensors and deep learning, the problems of accuracy and timeliness in waviness detection in float glass production were solved, enabling real-time online detection of float glass product quality and improving the control capability and efficiency of the production process.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- UNIV OF SCI & TECH BEIJING
- Filing Date
- 2024-04-08
- Publication Date
- 2026-06-12
AI Technical Summary
Existing technologies make it difficult to accurately and promptly detect waviness during float glass production, resulting in significant delays in product quality inspection and limited guidance. Furthermore, traditional methods can damage products and are not suitable for rapid production processes.
A soft measurement model detection method and device for float glass product quality is constructed by employing multi-sensor process variable screening, dual k-means condition identification, multi-scale time-space feature extraction and fusion, and high-confidence pseudo-label sample construction, combined with deep metric learning and genetic algorithms.
It enables real-time online detection of waviness during float glass production, reducing detection delay, improving the accuracy of product quality control and production efficiency, and adapting to the complexity of industrial processes and the characteristics of continuous multi-process production.
Smart Images
Figure CN118332364B_ABST