FDC Sensor Trend Classification for Optimal Interlock Limits
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In the semiconductor manufacturing industry, existing fault detection and classification (FDC) systems struggle to efficiently analyze and classify sensor data in real-time, leading to suboptimal equipment operation and reduced productivity.
Innovation Solution
A process management system equipped with a fault detection and classification (FDC) module that analyzes sensor data trends, classifies data types, calculates specification data, and verifies optimal operating ranges by simulating interlock applications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If sensor data is analyzed and classified in real-time to detect faults, then equipment usage is maximized and productivity is improved, but the complexity of data analysis and classification increases
Solution Approach 1:
The patent segments sensor data into multiple dimensions including time-based trends (increasing, decreasing, constant), magnitude relationships (above mean, below mean, equal to mean), and statistical distributions (normal, skewed left, skewed right, uniform). This segmentation transforms complex raw sensor data into structured, classified categories that are easier to analyze and act upon, resolving the contradiction between real-time analysis capability and data complexity.
2Reliability
If specification data is calculated and verified through simulation, then optimal operating ranges are identified and interlock occurrences are reduced, but the time and computational resources required increase
Solution Approach 1:
The patent performs preliminary simulation and verification of specification data before deploying it to production equipment. By pre-calculating optimal specification data through simulations and verifying its effectiveness in reducing interlock occurrences, the system prepares optimized parameters in advance. This preliminary action ensures that when specification data is applied to equipment, it already has proven reliability, reducing the need for extensive real-time verification and minimizing production downtime.
3Measurement precision
If sensor data is classified by trend and type, then fault detection accuracy is improved, but the processing complexity and data storage requirements increase
Solution Approach 1:
The patent transforms raw sensor data by changing its parameters through classification into standardized categories. Instead of analyzing continuous, unstructured sensor values, the system converts them into discrete classification parameters such as trend direction (increasing/decreasing/constant), relative magnitude (above/below/equal to mean), and distribution type (normal/skewed/uniform). This parameter transformation maintains fault detection accuracy while significantly simplifying processing and storage requirements.
Data Source
AI summary
A process management system includes a fault detection and classification (FDC) module that analyzes a trend of sensor data of an equipment or a facility, classifies the sensor data as a type according to the trend, calculates specification data according to the type, and selects optimal specification data by verifying the specification data, and a user interface module that displays the optimal specification data, a simulation result, and an increase/decrease in an interlock before/after changing to the optimal specification data, and a storage that stores the sensor data, the specification data, and the optimal specification data. The sensor data is state information measured by a sensor mounted on the equipment or the facility, the specification data includes an upper control limit (UCL) and/or a lower control limit (LCL), and when the sensor data exceeds the UCL or the LCL, the interlock is applied to the equipment or the facility.