Battery Separator Black Spot Detection Using Dual Imaging AI
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting black spots in battery separators, which cause fine short circuits, rely on visual inspection, leading to high variability and low detection rates due to human error and the inability to identify small spots.
Innovation Solution
A system utilizing a winding machine, foreign material removal unit, first and second image measuring units, and deep learning to automatically detect and analyze black spots in battery separators, enhancing detection accuracy and reducing variability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If visual inspection with naked eye is used, then inspection simplicity is maintained, but detection precision and reliability deteriorate due to large dispersion among inspectors and inability to detect small black spots
Solution Approach 1:
The patent replaces the mechanical visual inspection method with an automated image processing system consisting of cameras, image processing units, and analysis software. This substitution eliminates human error and subjectivity while achieving consistent, high-precision detection of black spots across all separators.
Solution Approach 2:
The patent creates digital copies (images) of the separator surfaces using cameras, then analyzes these copies through image processing algorithms. This allows multiple inspectors to simultaneously examine the same digital images without the limitations of direct visual inspection, improving both precision and consistency.
2Reliability
If multiple cells are dismantled and analyzed manually, then more black spots may be found, but time consumption and productivity deteriorate significantly
Solution Approach 1:
The patent implements continuous automated inspection where separators are constantly monitored as they pass through the system. The image capture and analysis process runs continuously without interruption, eliminating the need to stop production for manual inspection and dramatically improving productivity while maintaining high detection reliability.
Solution Approach 2:
The system performs self-inspection by automatically capturing images, processing them through algorithms, and identifying black spots without human intervention. The automated analysis unit independently evaluates each separator, eliminating the need for manual dismantling and inspection while maintaining consistent detection standards.
3Measurement precision
If automated image processing is implemented, then detection precision improves, but device complexity and initial cost increase
Solution Approach 1:
The patent divides the inspection system into distinct functional modules: image capture units with cameras, image processing units that analyze the captured images, and control units that coordinate the process. This segmentation allows each component to be optimized independently and simplifies maintenance and troubleshooting while achieving high detection precision.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system for finding black spots in a separator according to an embodiment includes a winding machine (11) and a rewinder (12) that take out and wind a separator from a rechargeable battery cell to produce and supply a sample of the separator; a foreign material removal unit (20) that removes a foreign material from the surface of the separator; a first image measuring unit (31) that obtains a first image of a part where black spots are estimated in the separator passing through the foreign material removal unit (20) with a first camera (C1) and records the position of the first image; a second image measuring unit (32) that first selects a part where black spots are estimated and acquires a second image of black spots and a foreign material other than black spots for the recorded position by a second camera (C2) by using the first image for the separator passing through the first image measuring unit (31); and a black spot sorting unit (40) that displays the position after secondary selecting the black spots by deep learning the first image and the second image with a deep learning software.