Abnormality Portent Detection for Semiconductor Manufacturing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In semiconductor device manufacturing, existing abnormality detection systems fail to effectively monitor the degradation of consumable parts in manufacturing apparatuses, leading to either premature replacement or missed defects, resulting in increased costs and yield loss.
Innovation Solution
An abnormality portent detection system that collects and analyzes sensor data from multiple parameters to calculate contribution rates, shifting the boundary between normal and abnormal periods to identify parameters showing maximum change before an abnormality occurs, allowing for early detection of consumable part degradation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing abnormality detection systems are used to monitor consumable parts, then abnormality detection capability is maintained, but detection precision is insufficient leading to premature replacement or missed defects
Solution Approach 1:
The system changes the parameter being monitored from simple abnormality presence/absence to contribution rates of multiple parameters. By calculating how much each parameter contributes to state transitions and identifying parameters with maximum contribution rates before abnormality timing, the system achieves more precise detection while maintaining reliability.
Solution Approach 2:
The system transitions from single-parameter monitoring to multi-dimensional analysis by examining multiple parameters simultaneously, calculating their contribution rates, and analyzing temporal patterns. This dimensional expansion enables more accurate prediction of consumable part degradation while maintaining system reliability.
2Reliability
If consumable parts are replaced based on fixed schedules, then equipment reliability is maintained, but unnecessary replacement increases operational costs
Solution Approach 1:
The system performs preliminary detection of consumable part degradation by analyzing contribution rates of multiple parameters before actual abnormality occurs. By identifying parameters showing maximum change before abnormality timing, the system enables proactive maintenance scheduling that prevents both premature replacement and missed defects, optimizing operational costs while maintaining equipment reliability.
Solution Approach 2:
The system implements feedback through continuous monitoring of parameter contribution rates and comparing them against thresholds. This feedback mechanism allows dynamic adjustment of maintenance schedules based on actual consumable part condition, replacing parts only when necessary while maintaining equipment reliability.
3Measurement precision
If multiple parameters are monitored to improve detection accuracy, then abnormality detection precision improves, but system complexity increases
Solution Approach 1:
The system extracts only the most critical information by calculating contribution rates for each parameter and identifying those with maximum contribution rates before abnormality timing. This extraction approach allows monitoring of multiple parameters while focusing analysis on the most informative subset, improving detection precision without proportionally increasing system complexity.
Data Source
AI summary
According to one embodiment, in an abnormality portent detection system, a collection unit time-sequentially collects plural kinds of parameters related to a state of an apparatus. A calculation unit calculates, while temporally changing a boundary between a first period and a second period in a time-sequential variation characteristic of each of plural kinds of parameters, a contribution rate of each of the plural kinds of parameters with respect to a transition from a first state of the apparatus before the boundary to a second state of the apparatus after the boundary. An extraction unit extracts, among the plural kinds of parameters, a parameter showing a change in which the contribution rate has a maximum value at a timing before an abnormality occurrence timing of the apparatus with respect to a time of the boundary, based on a result of the calculation.