Electrode Stack Corner Detection Using 3D Imaging and Neural Nets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining the placement accuracy of electrode sheets in electrode composite stacks are not fully automated, precise, or robust, often relying on specialist evaluation of 3D image data with poor signal-to-noise ratios, and do not provide effective correction measures for placement errors.
Innovation Solution
A computer-implemented method using 3D imaging and convolutional neural networks to accurately determine the position of corner regions in electrode composite stacks by generating corrected edge profiles and extrapolating corner positions, ensuring precise and automated corner detection without compression or stretching of electrode sheets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If electrode sheet placement accuracy is increased to ensure complete coverage, then manufacturing precision is improved, but production speed decreases
Solution Approach 1:
The patent replaces manual specialist evaluation of 3D image data with an automated neural network system that processes CT scan images to determine electrode placement accuracy. This automation resolves the contradiction by enabling rapid, precise measurement without requiring slow manual analysis, thus maintaining high manufacturing precision while increasing production speed.
Solution Approach 2:
The patent changes the measurement parameters by using 3D CT scan imaging with specific slice thicknesses (0.5-2mm) and automated neural network analysis rather than manual 2D image evaluation. This parameter change enables faster processing while maintaining or improving measurement precision, resolving the speed-accuracy tradeoff.
2Productivity
If 3D image data acquisition speed is increased, then productivity is improved, but signal-to-noise ratio deteriorates
Solution Approach 1:
The patent applies preliminary action by using a trained neural network system that has been pre-trained on大量 CT scan images to recognize and filter noise patterns. This pre-training enables the system to rapidly process high-speed acquisition images while maintaining high signal-to-noise ratio through learned noise discrimination, resolving the contradiction between acquisition speed and measurement precision.
Solution Approach 2:
The patent creates a digital 3D model (copy) of the electrode stack from CT scan data, allowing multiple analyses to be performed on the copied data without requiring additional physical scanning. This copying approach enables rapid repeated measurements of the same high-quality 3D data, maintaining signal-to-noise ratio while achieving high productivity through virtual rather than physical re-scanning.
3Extent of automation
If automated evaluation system is implemented, then extent of automation is improved, but device complexity increases
Solution Approach 1:
The patent implements a universal neural network system that performs multiple functions: it identifies electrode positions, determines placement accuracy, detects rotation, and generates correction recommendations. This multi-functionality reduces overall system complexity by consolidating what would otherwise require multiple separate devices into a single automated evaluation system.
Solution Approach 2:
The patent introduces a coordinate system as an intermediary that standardizes the representation of electrode positions and orientations. This intermediary framework simplifies the automation by providing a common reference system that the neural network can process uniformly, reducing the complexity of handling diverse measurement data types.
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
Enables robust, precise, and fully automated determination of electrode sheet positions, improving manufacturing efficiency and accuracy by correcting placement errors and optimizing the manufacturing process.
Implementation Method 1
3D imaging of the corner region of electrode sheets of the electrode stack in a recording area using an imaging method, in particular a computed tomography imaging method
Implementation Method 2
a computed tomography imaging method, such that a data set is generated which includes 3D positional information of the electrode sheets
Data Source
Figure 1~2
Figure 3~5
Figure 6
AI summary
The invention relates to a method for determining the position of corners of polygonal electrode sheets in at least one corner region of an electrode stack (ESV), comprising the steps of: - 3D imaging of the corner region of the electrode stack (ESV) in a recording area using an imaging method, such that a data set is generated which includes 3D positional information of the electrode sheets (A, K) in the corner region of the electrode stack (ESV) relative to a support (5, SD, SB) or a marker arranged in the recording area, - Determining from the data set a first and a second edge profile (20x, 20y) of the edges (21x, 21y) framing the corner region of each electrode sheet (A, K), wherein a position of the corner (E1, E2, E3, E4) of the respective electrode sheet (A, K) is determined based on the edge profile (20x, 20y). The electrode sheets and their edges are determined using a neural network system.