Battery Separator Black Spot Inspection with Dual-Camera Deep Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for detecting black spots causing fine short circuits in rechargeable battery separators rely on human visual inspection, which is prone to high dispersion and low detection efficiency, especially for small spots.
Innovation Solution
A method using a combination of cameras and deep learning to automatically identify black spots by obtaining and analyzing images of separator surfaces, employing adhesion and removal rollers to prepare samples, and utilizing X-ray fluorescence for component analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If human visual inspection is used to detect black spots, then the detection method is simple, but the detection precision and reliability are low due to large dispersion among inspectors and inability to detect small spots
Solution Approach 1:
The patent replaces the mechanical human visual inspection system with an automated optical detection system comprising a camera, illumination device, and image processing unit. This substitution eliminates human factors causing dispersion and improves detection precision while maintaining operational simplicity through automated image analysis algorithms.
Solution Approach 2:
The patent creates an optical copy (image) of the separator surface and analyzes it through image processing. By capturing the separator surface as an image and processing it computationally, the system achieves high detection precision for small black spots without requiring direct human observation, thereby eliminating inspector variability.
2Productivity
If multiple cells are dismantled and analyzed to find black spots, then the detection coverage increases, but the productivity decreases due to time-consuming manual inspection
Solution Approach 1:
The patent enables continuous inspection by processing separator images automatically without interruption. The system can continuously capture, process, and analyze images of multiple separators in sequence, maintaining both high productivity through automated operation and high detection precision through consistent image processing algorithms, eliminating the need for time-consuming manual dismantling and inspection.
3Measurement precision
If visual inspection with naked eye is used, then the equipment complexity is low, but the detection ratio of small black spots is very low
Solution Approach 1:
The patent applies local quality enhancement by using targeted illumination devices that specifically light up the separator surface areas where black spots are likely to occur. The illumination is optimized to enhance the contrast and visibility of small black spots in specific regions, improving detection ratio without requiring complex equipment across the entire inspection system.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enhances the detection ratio of black spots and reduces analyst dispersion by providing accurate and efficient identification of black spots, improving inspection speed and reducing over-detection.
Implementation Method 1
a first adhesion roller and a first removal roller are sequentially contacted (with) and rotated on the first surface of the separator, the foreign material on the first surface may be removed using the viscosity difference between the first adhesion roller and the first removal roller
Implementation Method 2
automatically analyzing the components of the black spots by using an equipment, the black spots may be specified twice in the first selection and the second selection, in the automatically analyzing, the components of the black spots may be analyzed with X-rays using X-ray fluorescence analysis (XRF) equipment
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method of finding black spots in a separator according to an embodiment includes taking out (ST1) a separator from a rechargeable battery cell; removing (ST2) a foreign material of the separator surface; obtaining (ST3) a first image for a portion where black spots are estimated in the separator by using a first camera (C1) and recording a position of the first image; first selecting (ST4) the part where black spots are estimated by using the first image, and acquiring a second image for black spots and a foreign material other than black spots in the separator by using a second camera (C2) for the recorded position; and secondary selecting (ST5) black spots by deep learning the first image and the second image with a deep learning software and then displaying the position of the black spots after.