Manufacturing Control Using Intermediate Quality Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In semiconductor manufacturing, identifying and addressing quality issues in intermediate products is challenging, as poor quality can delay the identification of problematic processes and reduce final product yield, as the quality of intermediate products is often not predetermined.
Innovation Solution
A manufacturing control apparatus that acquires past data to predict output values and quality of intermediate products, using components like an acquiring part, output predicting part, and quality predicting part to control processing conditions and ensure intermediate products meet target values, thereby improving final product quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If quality inspection is performed only on final products, then inspection cost is reduced, but identification time of problematic manufacturing processes increases
Solution Approach 1:
The patent performs quality prediction on intermediate products before they become final products. The quality predicting part uses past manufacturing data to predict the quality of intermediate products at various stages, enabling early identification of problematic processes before they affect final product yield, thus resolving the contradiction between early detection and productivity
2Productivity
If quality prediction is performed on all intermediate products, then final product yield increases, but computational complexity and data processing requirements increase
Solution Approach 1:
The patent replaces complex physical quality inspection mechanisms with a computational quality prediction system. The quality predicting part uses statistical analysis of past manufacturing data to predict intermediate product quality, substituting physical measurement equipment with data processing algorithms, thereby reducing device complexity while maintaining high final product yield
Solution Approach 2:
The patent creates a virtual model of the manufacturing process using past manufacturing data. The acquiring part collects historical data about manufacturing conditions and final product quality, building a database that serves as a virtual copy of the manufacturing system, enabling quality prediction without physical inspection of each intermediate product
Data Source
AI summary
A manufacturing control apparatus has an acquiring part that acquires past manufacturing data in which a target value of an output value of the intermediate product, the output value of the intermediate product, and quality of a final product produced from the manufacturing apparatus are associated with one another, an output predicting part that predicts, based on the target value of the intermediate product and the output value of the intermediate product in the past manufacturing data, the output value of the intermediate product for each of possible target values of the intermediate product, and a quality predicting part that predicts, based on the output value of the intermediate product and the quality of the final product in the past manufacturing data, the quality of the final product from a predicted value of the output value of the intermediate product for each of the predicted possible target values.