Event Processing System for Real-Time Manufacturing Yield Improvement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional Yield Management Systems (YMS) in manufacturing, such as semiconductor production, analyze defects only after the manufacturing process is complete, failing to utilize real-time data for predictive analysis and in-line resolution, leading to delayed identification and remediation of defects.
Innovation Solution
An event processing system that applies rules and heuristics discovered through offline analysis to real-time manufacturing data, allowing for the identification of potential defects during the process and enabling immediate remedial action by storing selected data in memory-resident storage for efficient processing without impacting online transaction processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional Yield Management Software analyzes defects after manufacturing completion, then comprehensive end-of-line yield data can be obtained, but defect identification is delayed and real-time remediation is impossible
Solution Approach 1:
The system performs preliminary analysis of manufacturing data to identify potential defects before the manufacturing process completes. By analyzing data in real-time during production, the system can detect anomalies and predict defects early, enabling preventive actions to be taken before defective products are fully manufactured, thus reducing both time loss and improving detection accuracy
Solution Approach 2:
The system implements continuous feedback loops where manufacturing data is constantly monitored, analyzed, and used to adjust process parameters in real-time. This feedback mechanism enables the system to respond to emerging defects immediately, closing the loop between detection and remediation without waiting for end-of-line results, thereby reducing time loss while maintaining high detection accuracy
2Productivity
If real-time data processing is implemented during manufacturing, then immediate defect detection and remediation are enabled, but system complexity increases
Solution Approach 1:
The system segments the manufacturing data processing into distinct modular components: data collection modules, real-time analysis modules, defect detection modules, and remediation modules. Each module handles specific tasks independently, which reduces overall system complexity while enabling comprehensive real-time processing. This modular architecture allows the system to maintain high productivity through parallel processing without requiring excessive complexity in any single component
Solution Approach 2:
The event processing system is designed with universal components that can handle multiple types of manufacturing data and defect scenarios through a single integrated platform. Rather than implementing separate specialized systems for different defect types, the system uses multi-functional algorithms and data structures that can adapt to various manufacturing processes and defect patterns, thereby achieving high productivity without proportionally increasing system complexity
3Reliability
If comprehensive manufacturing data is collected and analyzed in real-time, then predictive defect identification is improved, but data processing overhead increases
Solution Approach 1:
The system extracts only the most critical and relevant features from comprehensive manufacturing data for real-time analysis, rather than processing all raw data. By identifying and extracting key indicators that are most predictive of defects, the system maintains high reliability in defect identification while significantly reducing computational overhead. Irrelevant or redundant data elements are filtered out before analysis, optimizing the balance between reliability and resource consumption
Solution Approach 2:
The system applies partial analysis to comprehensive data by focusing computational resources on the most critical data elements and time-sensitive parameters. Rather than equally processing all manufacturing data, the system selectively intensifies analysis on high-risk parameters while using lighter processing for stable, low-risk parameters, thereby achieving reliable predictive identification with optimized resource utilization
Data Source
AI summary
An event processing system identifies a process event associated with an identified defect in a manufacturing process. The event processing system selects a plurality of data elements from a manufacturing data source based on the process event. The manufacturing data source is associated with the manufacturing process during execution of the manufacturing process. During execution of the manufacturing process, the event processing system applies an event rule to the plurality of data elements to determine whether the event rule is satisfied. During execution of the manufacturing process, the event processing system performs a predefined action upon determining that the event rule is satisfied and selects additional data elements from the manufacturing data source upon determining that the event rule is not satisfied.