一种超大光学基板研抛过程自适应闭环控制方法及系统
By constructing a full-process adaptive closed-loop control architecture and integrating multi-source sensing networks and deep reinforcement learning models, real-time acquisition of multi-dimensional process signals and adaptive optimization of parameters in the polishing process of ultra-large optical substrates were achieved. This solved the problems of incomplete monitoring and insufficient adaptability in existing technologies, and improved processing quality and efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- NANYANG NEW PRECISION OPTICS CO LTD
- Filing Date
- 2026-06-17
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies cannot fully perceive multi-dimensional process signals during the polishing of ultra-large optical substrates, making it difficult to establish a complete understanding of complex processing states. Furthermore, offline-trained machine learning models cannot adapt to real-time changes during the processing, resulting in insufficient adaptability of control strategies.
A full-process adaptive closed-loop control architecture is constructed, integrating a multi-source sensing network for multi-dimensional process signal acquisition. Combined with a dynamic relationship model of deep reinforcement learning and multi-model fusion, online parameter learning and optimization decision-making are achieved, and real-time adjustment is realized through the collaborative control of distributed actuators.
It achieves comprehensive monitoring and adaptive control of the polishing process of ultra-large optical substrates, and can dynamically adjust process parameters in real time, thereby improving processing quality and efficiency and solving the problems of incomplete monitoring and insufficient adaptability in existing technologies.
Smart Images
Figure CN122411276A_ABST