Measurement Data Change Sensing for Subtle Process Drift Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current sensing techniques in manufacturing processes are inadequate to detect subtle changes in measurement data, leading to quality variability issues.
Innovation Solution
A method and system that extract and compare data using statistical distributions, calculate a target range, and determine changes based on inclusion ratios and distribution tests to identify subtle changes in measurement data, enabling preemptive quality control measures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If statistical process control technique is used to sense data trends, then short-term downward or upward trend can be detected, but subtle changes between pre-measurement data and post-measurement data cannot be identified
Solution Approach 1:
The patent segments the measurement data into reference data (pre-measurement) and comparison data (post-measurement), and further segments the analysis into multiple statistical tests including distribution tests, representative value tests, and target range inclusion ratio tests. This segmentation allows detection of subtle changes by analyzing different aspects of the data separately and comprehensively
Solution Approach 2:
The patent introduces multiple dimensions of analysis beyond simple trend detection: distribution characteristics, representative values, and target range inclusion ratios. By analyzing data across these multiple dimensions, the system can detect subtle changes that would be invisible in traditional single-dimension trend analysis
2Reliability
If traditional sensing methods are used, then simple trend detection is achieved, but quality variability cannot be minimized due to inability to detect subtle changes
Solution Approach 1:
The patent performs preliminary statistical analysis on measurement data by establishing reference data before production changes and comparing post-change comparison data against this reference. This preliminary action framework enables early detection of subtle changes before they manifest as obvious quality defects, allowing preemptive quality control measures
Solution Approach 2:
The system establishes a feedback mechanism by continuously comparing measurement data against statistical reference data and target ranges, then using the results to determine whether quality degradation or enhancement has occurred. This feedback loop enables rapid response to subtle changes, minimizing quality variability through timely corrective actions
Data Source
AI summary
A method and a system for sensing fine changes in processing/equipment measurement data are provided. A data change sensing method according to an embodiment of the present invention extracts a part on the basis of a statistical distribution of reference data and comparison data, calculates a target range on the basis of a specification, and discriminates data, included in the target range, among the extracted reference data and comparison data so as to determine data changes. Therefore, fine changes in measurement data for processing or equipment can be sensed in a manufacturing process, thereby enabling pre-estimation of potential quality variability of products and quick preemptive actions for preventing quality degradation.