The semiconductor display industry faces increasing product varieties, higher quality demands, and a market becoming short-cycle with growing customization requirements. These challenges require enterprises to shift from traditional manufacturing thinking, integrating AI with manufacturing to upgrade smart manufacturing levels for efficient and low-cost operations. CSOT's LCD panel production process is often disturbed by various sources, causing quality deviations. To ensure product quality, key parameters of work-in-progress need detection to meet process standards. However, some inspections cannot be done in real-time on the production line, so only sampling for offline measurement is used. This leads to inability to timely detect defects, easily causing batch rework or scrap, resulting in high resource and cost waste.
Greatech Eastern Intelligence, based on Eastern Intelligence industrial edge gateways and IoT platform, collects process data and sampling measurement data. Then, by analyzing thousands of process factors for panel products, it quickly and accurately establishes the relationship between key process indicators like film thickness and quality measurement characteristics, achieving virtual metrology full inspection. The system identifies key anomaly factors, enabling production engineers to promptly analyze causes and adjust process parameters to prevent large-scale defects. Additionally, Virtual Metrology (VM) configures alarm rules to notify systems like MES and OEE for multi-system coordinated control and parameter adjustment. The model online operation periodically auto-tunes to continuously improve VM prediction accuracy.
