Virtual Metrology Pattern Clustering for Real-Time Tool Maintenance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing virtual metrology systems face challenges in ensuring the accuracy and reliability of quality control in manufacturing processes, particularly in recognizing unrecognizable issues, which can lead to worsened problems and make real-time corrections impossible, and they often rely on data-driven methodologies that require significant labor and resource investment for establishing models for each equipment type.
Innovation Solution
A virtual metrology system that clusters tool condition data according to predetermined patterns, allowing for normal sampling, maintenance, or shutdown measures based on pattern matches, and dynamically updates patterns as new data accumulates, ensuring accurate and efficient quality control without the need for extensive model retraining.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data-driven methodology is used to ensure correctness of VM values, then measurement precision is improved, but device complexity increases due to extensive model creation requirements
Solution Approach 1:
The system performs preliminary actions by continuously collecting and storing process data and metrology data in advance. This allows the system to have pre-established data pools that can be quickly analyzed without requiring extensive real-time model creation, thus improving measurement precision while reducing the complexity of on-demand model development
Solution Approach 2:
The system creates virtual copies of the manufacturing process through computational models that simulate physical metrology measurements. These virtual metrology values are generated by copying and analyzing historical process data patterns, eliminating the need for physical measurement equipment and complex real-time model creation for each measurement
2Measurement precision
If feedback paths are used to correct data issues, then measurement precision is improved, but loss of time increases due to iterative correction processes
Solution Approach 1:
The system performs preliminary data validation and anomaly detection by comparing new data against historically established patterns and ranges. Issues are identified and flagged before they propagate through the system, eliminating the need for time-consuming iterative feedback corrections and enabling immediate corrective actions
Solution Approach 2:
The system performs self-diagnosis by automatically detecting data anomalies and identifying their root causes through pattern recognition. The system self-corrects by adjusting processing parameters or flagging specific issues without requiring external feedback loops, thereby reducing correction time while maintaining data precision
3Measurement precision
If individual models are established for different chambers to ensure conjecture accuracy, then measurement precision is improved, but loss of time increases due to significant labor cost for model establishment
Solution Approach 1:
The system establishes a universal data collection and analysis framework that can be applied across multiple chambers and equipment types. By creating a common infrastructure for data management, pattern recognition, and virtual metrology generation, the system achieves chamber-specific accuracy without requiring separate model establishment processes for each chamber, significantly reducing the time and labor required
Solution Approach 2:
The system maintains a single adaptable model framework that dynamically adjusts its parameters and weighting factors based on chamber-specific data patterns. Rather than creating entirely separate models for each chamber, the system modifies existing model parameters to account for chamber variations, achieving high conjecture accuracy while minimizing model establishment time
Data Source
AI summary
A virtual metrology system at least includes a process apparatus including a set of process data, the process apparatus producing a workpiece according to the set of process data. A virtual metrology server is configured to gather the set of process data, cluster the set of process data to obtain data clusters, and compare the data clusters with patterns. If the data clusters meet the patterns corresponding to the data clusters, performing a corresponding maintenance, repair, and overhaul step on the process apparatus.