Battery Yield Prediction From Foreign Matter Inspection Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for battery production fail to accurately predict yield in the inspection process due to variations in foreign matter states, leading to excessive discard of materials and increased production loss costs.
Innovation Solution
A quality management system that utilizes a relational expression to correlate foreign matter inspection data with battery inspection data to predict yield, incorporating a model storage unit configured to store a relational expression that formulates a relational expression that formulates a relational expression that formulates a relational expression that formulates a relation between foreign matter inspection data and battery inspection data, allowing for yield prediction and display of yield prediction values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all materials or products in process that are mixed with foreign matter are discarded, then the reliability of battery inspection is improved, but the productivity is worsened due to excessive reduction in yield
Solution Approach 1:
The patent changes the parameter of foreign matter evaluation from binary (presence/absence) to continuous (state characteristics including volume, composition, and position). By analyzing these parameter variations, the system can distinguish between foreign matter that causes micro-short circuits and those that do not, enabling selective acceptance or rejection based on specific parameter thresholds rather than universal discard
Solution Approach 2:
The patent implements a feedback mechanism where foreign matter inspection data is correlated with battery inspection results through a relational expression. The system learns from actual battery inspection outcomes (pass/fail) and adjusts the acceptance criteria for foreign matter, creating a closed-loop system that improves both reliability and yield by continuously optimizing the decision boundary based on empirical data
2Manufacturing precision
If foreign matter inspection is performed to detect micro-short circuits, then the quality of battery inspection is improved, but the loss of time is worsened due to additional inspection processes
Solution Approach 1:
The patent merges the foreign matter inspection process with the battery inspection process by establishing a relational expression that correlates foreign matter characteristics with battery inspection results. Instead of treating them as separate sequential processes, the system integrates the evaluation criteria, allowing simultaneous assessment of both foreign matter presence and battery performance in a unified inspection framework
Solution Approach 2:
The patent performs preliminary analysis by pre-establishing the relational expression between foreign matter inspection data and battery inspection results. This relational model is configured in advance based on historical data and theoretical correlations, enabling the system to quickly evaluate new samples without performing time-consuming iterative analysis during actual inspection
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided is a quality management device capable of reducing production loss cost by predicting a yield in a battery inspection process based on foreign matter inspection data of a battery material or a product in process. A quality management device 10 includes: a model storage unit 11 configured to store a relational expression that formulates a relation between foreign matter inspection data from a foreign matter inspection device 21 configured to measure a foreign matter mixed as an impurity in a battery material or a product in process in a battery production process (an electrode production process 31 and a cell production process 32) and battery inspection data from a battery inspection device 22 configured to inspect, in a battery inspection process 33, electrical characteristics of a battery produced in the battery production process; a yield prediction value calculation unit 12 configured to calculate a yield prediction value in the battery inspection process using the relational expression of the model storage unit 11 with the foreign matter inspection data as an input; and a yield prediction value display unit 13 configured to display the yield prediction value on a display device 23.