Process Parameter Correlation Monitoring for Early Fault Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing manufacturing systems in the semiconductor and panel industry struggle to efficiently and accurately detect correlation deviations in process parameters, leading to the production of defective products due to equipment abnormalities.
Innovation Solution
A manufacturing system incorporating a modeling calculation unit to establish a control sequence table using historical parameter data and an analysis calculation unit to calculate contribution values based on real-time data, facilitating the detection of parameter correlations and deviations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If traditional measurement stations are used to randomly measure product quality, then measurement coverage is achieved, but detection timing is too late causing large numbers of defective products
Solution Approach 1:
The system performs preliminary detection by continuously monitoring process parameters and establishing control sequence tables before defective products are produced. The Fault Detection and Classification system proactively identifies abnormalities in real-time, enabling early intervention before quality degradation occurs, thus resolving the contradiction between detection timing and defective product rate.
2Reliability
If Fault Detection and Classification system is implemented, then defective product cost is reduced, but parameter correlation detection accuracy is insufficient due to high complexity
Solution Approach 1:
The system segments the complex parameter monitoring task into two distinct components: (1) a control sequence table generation module that processes historical data to establish normal parameter relationships, and (2) a contribution value calculation module that compares real-time parameters against the control sequence. This segmentation enables accurate detection of parameter correlation deviations despite the high dimensionality and complexity of process parameters.
3Productivity
If equipment parameters are monitored in real-time, then process control is improved, but the large number of highly correlated parameters makes abnormality detection difficult
Solution Approach 1:
The system introduces a control sequence table as an intermediary structure that captures the normal correlation relationships among multiple process parameters. By comparing real-time parameter combinations against this intermediary control sequence, the system efficiently detects abnormalities without being overwhelmed by the sheer number of parameters. The control sequence table acts as a reference model that simplifies the detection process while maintaining accuracy.
Data Source
AI summary
A manufacturing system for electronic devices is provided, including: a machine, a modeling calculation unit, and an analysis calculation unit. The machine provides a historical parameter data and a real time parameter data. The modeling calculation unit receives the historical parameter data to establish a control sequence table based on the historical parameter data. The analysis calculation unit calculates a contribution value based on the control sequence table and the real time parameter data.