Production Capacity Boundary Estimation Using Takt Time Outlier Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining production capacity boundaries in semiconductor manufacturing, primarily relying on normal distribution algorithms for Takt Time data, are not accurate enough, leading to deviations in capacity planning and resource preparation.
Innovation Solution
A method that involves acquiring and sorting time intervals between production ending points of adjacent lots, parsing distribution features to determine a data removing condition, and removing outliers based on a mean value to accurately determine production capacity boundaries using minimum and maximum time intervals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If normal distribution algorithm is used to process Takt Time data, then the processing is simple, but the accuracy of production capacity boundary determination is insufficient
Solution Approach 1:
The patent segments the Takt Time data processing into multiple stages: initial sorting of time intervals, identification of boundary regions, iterative removal of outliers from boundaries, and final capacity boundary determination. This segmentation transforms a single complex estimation into a systematic multi-step process that improves accuracy while maintaining manageable complexity
Solution Approach 2:
The patent performs preliminary actions by first sorting all Takt Time intervals and identifying boundary regions before determining the final production capacity boundaries. The iterative removal of outliers from boundaries is performed in advance of the final calculation, ensuring that the mean value used for outlier identification reflects the true central tendency of the data
2Measurement precision
If multiple iterations of outlier removal are performed, then the accuracy of production capacity boundary determination is improved, but the data processing time increases
Solution Approach 1:
The patent applies partial action by performing outlier removal iterations only at the boundary regions of the sorted Takt Time data, rather than processing the entire dataset repeatedly. The extraction step length parameter controls the scope of processing, removing only the necessary portion of data to achieve accurate boundary determination while minimizing unnecessary computational effort
Data Source
AI summary
A method, apparatus and device for determining production capacity boundaries are provided. In the method, related data for producing a run size of elements by a production device lot by lot is acquired and time intervals between production ending time points of adjacent lots are determined according to the related data; the determined time intervals are sorted to obtain a time interval sequence; distribution features of time intervals at two boundaries are parsed respectively to determine whether a data removing condition is satisfied; if Yes, an outlier is determined according to a mean value of the present time interval sequence, the time interval of the extraction step length where the outlier is located is removed, and whether the data removing condition is satisfied is determined; and if No, production capacity boundaries are determined according to minimum and maximum time intervals of the present time interval sequence and the run size.